Sticking the Landing: Enhancing Liposomal Cell Delivery using Reversible Covalent Chemistry and Caged Targeting Groups

Abstract

Liposomes are highly effective nanocarriers for encapsulating and delivering a wide range of therapeutic cargo. While advancements in liposome design have improved several pharmacological characteristics, an important area that would benefit from further progress involves cellular targeting and entry. In this concept article, we will focus on recent advancements utilizing strategies including reversible covalent bonding and caging groups to activate liposomal cell entry. These approaches take advantage of advancements that have been made in complementary fields including molecular sensing and chemical biology and direct this technology toward controlling liposome cell delivery properties. The decoration of liposomes with groups including boronic acids and cyclic disulfides is presented as a means for driving delivery through reaction with functional groups on cell surfaces. Additionally, caging groups can be exploited to activate cell delivery only upon encountering a target stimulus. These approaches provide promising new avenues for controlling cell delivery in the development of next-generation liposomal therapeutic nanocarriers.

Graphical Abstract

We summarize recent advancements in enhancing liposomal cellular delivery. Examples using reversible dynamic chemistry are first discussed, including those driven by boronic acid-carbohydrate binding and cyclic disulfide-thiol exchange. Next, we discuss the development of a boronate-caged guanidine lipid that enables triggered cell entry only in the presence of ROS. These approaches provide promising new avenues for controlling liposomal cell delivery.

Introduction

Liposomes are spherical nanoparticles formed by the self-assembly of lipids in aqueous solution that exhibit superb drug delivery properties due to their ability to encapsulate and transport therapeutics with wide-ranging physicochemical properties.^[1] Several liposomal formulations are in use in the clinic, which is the culmination of improvements that have been made in terms of the circulation time and targeting of liposomal formulations.^[2] However, an area in which continued efforts are necessary involves improvements to the ability of liposomes to infiltrate cells, especially in a manner that is cell-type specific. Prior work has yielded advancements in this area. In a prominent example, the development of ionizable lipids has facilitated the delivery of RNA,^[3] which revolutionized lipid nanocarrier delivery and led to the development of mRNA vaccines. Grafting of cationic moieties (i.e., cell-penetrating peptides (CPPs))^[4] onto liposome surfaces through conjugation to lipids also improves cell entry. Regarding selectivity, ligands including folate,^[5] RGD peptides,^[6] and antibodies^[7] appended onto liposomes act by engaging particular receptors or other targets that exhibit differential expression on diseased cell surfaces. All of these functionalities generally act by maneuvering liposomes into close proximity of cell surfaces through non-covalent binding interactions that enable the crossing of cellular membranes.

However, there are disadvantages to both of these types of approaches that remain as obstacles to overcome. Cationic moieties often suffer from minimal cell-type selectivity since they rely on the binding of negatively charged membranes, which is the predominant state in nature due to the negative charges that dominate in lipid structures.^[8] While this drawback can be addressed using targeting ligands, it is typically challenging and cost-prohibitive to introduce these onto lipid scaffolds and they may exhibit either limited or insufficient selectivity in the cell-types that can be targeted by liposomes.

This concept article focuses on new avenues for enhancing cell infiltration that show promise for addressing convenience and selectivity issues by exploring alternative functionalities presented on liposome surfaces (Figure 1). One approach has entailed platforms designed to undergo reversible covalent chemistry with partner reactive groups presented on cell surfaces in a manner that engages liposomes to encourage cell entry. In this realm, we will discuss boronic acid liposomes for binding to cell-surface glycans^[9] as well as cyclic disulfide liposomes for conjugation to thiol groups presented on membranes.^[10] Additionally, to enhance control over cationic liposomal systems, we have recently reported liposomes that initially present neutral caged guanidine moieties that are thus silent in terms of cell entry.^[11] However, upon encountering a trigger known to be upregulated in diseased cells, reactive oxygen species (ROS), the cationic guanidinium group is unveiled and cell entry is activated. These approaches have proven beneficial for enhancing cell infiltration using simple and economical lipid structures that can be conveniently incorporated within liposomes. While further work is required to determine whether these liposomal platforms will be effective for improving diseased cell selectivity, these strategies exhibit promise toward that goal since they invoke ligands that are differentially expressed on cell surfaces or activating mechanisms that are envisioned to improve control over cell entry.

Figure 1.

Strategies designed to enhance liposomal cell infiltration. Cartoon illustration depicts approaches discussed in this concept, including lipids exploiting boronic acid moieties for carbohydrate binding, cyclic disulfide groups for thiol exchange and boronate-caged guanidines that are activated for cell entry by ROS.

Discussion

The select recent examples that will be highlighted in this concept article build off of extensive prior work in fields such as molecular sensing,^[12] programmed membrane fusion,^[13] and drug delivery exploiting dynamic covalent chemistry.^[14] One approach that has been taken for invoking reversible covalent chemistry to drive cell entry has exploited carriers decorated with boronic acids. These recognition groups have long been exploited for the binding and recognition of glycans through the reversible formation of boronate esters with carbohydrates containing functional groups such as cis diols.^[15] More recently, boronic acids have been used to develop pro-drug and carrier strategies by enhancing delivery of therapeutics across membranes through the binding of cell-surface carbohydrates. In pioneering work, Raines and co-workers developed boronic acid caging moiety 1 (Figure 2) containing an ester functionality linked to a trimethyl lock (TML) self-immolating group.^[16] This was exploited to deliver green fluorescent protein (GFP) across membranes, after which esterase enzymes were harnessed to remove the cage, thereby generating native protein. Additionally, boronic acids have proven invaluable for enhancing polymer entry into cells.^[17]

Figure 2.

Structure and delivery mechanism for boronic acid-containing caging moiety 1 reported by Raines and coworkers. Binding interactions between the boronic acid groups and cell-surface carbohydrates drives cell entry, while subsequent cellular esterase cleavage releases the appended cargo inside cells, in this case the protein GFP. Adapted with permission from ref.^[16] Copyright 2016 American Chemical Society.

Our group set out to develop boronic acid liposomes as an opportunity to expand the toolbox of cell surface ligands that could be exploited for liposomal delivery,^[18] particularly since cell- surface carbohydrates are important targets due to their common dysregulation in diseased cells in terms of both abundance and sugar structure.^[19] To this end, we designed and synthesized boronic acid-lipid conjugates including 2 (Figure 3A) to present this recognition group on membrane surfaces upon their incorporation into liposomal formulations.^[9a] To confirm that liposomes containing 2 bind to carbohydrates, we performed microplate experiments to track the binding of fluorescently labeled liposomes to heparin-biotin conjugates immobilized onto streptavidin-coated microplates, which showed that liposomes otherwise composed of phosphatidylcholine (PC) only bound when they were doped with lipid 2. We also determined that the binding of heparin to 2/PC liposomes was capable of driving the release of both hydrophobic (Nile red) and hydrophilic (sulforhodamine B) cargo, which is important for applications hinging upon triggered release of therapeutic contents.^[20] Finally, through cellular fluorescence microscopy imaging, we showed that the inclusion of 2 within PC liposomes significantly increased delivery of fluorescently labeled liposomes to A375 melanoma cells (Figures 3B–D). This work confirmed that the boronic acid moiety provided a valuable expansion of the toolbox for enhancing liposomal cell delivery.

Figure 3.

A. Structure for boronic acid lipid 2, containing an ortho-(alkylaminomethyl)phenylboronic acid binding unit appended onto a lipid scaffold. B. Cartoon depiction for liposomal cellular entry driven by boronic acid binding to cell surface glycans. C. Confocal fluorescence image for A375 cells treated with rhodamine-phosphatidylethanolamine (Rd-PE) labeled PC liposomes. Minimal cellular entry was observed. D. Cells treated with 10% 2/PC liposomes exhibited significant Rd-PE fluorescence, indicating cell infiltration. Scale bars denote 20 μm. DAPI is shown in blue, Rd-PE is shown in red. Adapted with permission from ref.^[9a] Copyright 2018 Royal Society of Chemistry.

A recent expansion of this boronic acid liposome work was inspired by prior reports that bisboronic acids can exhibit enhanced selectivity in carbohydrate binding properties based on the geometric display of binding groups. In early work, Shinkai and co-workers developed bisboronic acid sensors including 3 (Figure 4) that can distinguish between d- and l-forms of sugars.^[21] Additionally, Drueckhammer et al. used a computer-guided approach to develop receptor 4, which exhibited 400-fold selectivity for glucose compared to galactose, mannose, and fructose.^[22] Finally, Wang and co-workers generated a library of bisboronic acid sensors to identify compounds including 5, which was effective for selective labeling of hepatocellular carcinoma cells that express the carbohydrate motif sialyl Lewis X.^[23]

Figure 4.

Structures of representative bisboronic acid sensors that selectively bind carbohydrates. Adapted with permission from ref.^[21–23] Copyright 1995 Springer Nature, 2001 Wiley-VCH and 2004 Elsevier.

In an effort aimed toward eventually enhancing cell-type selectivity of liposomes, we sought to build upon this prior recognition work by developing bisboronic acid liposomes (BBALs).^[9b] We first developed a small library of compounds (6a-d, Figure 5A) containing two boronic acids appended to a lipid scaffold. These structures contain a variable linker in between the two boronic acid moieties, which we hypothesized could lead to differential carbohydrate binding properties in line with prior work. To evaluate the binding activity of liposomes containing these compounds, we developed a competitive microplate binding assay in which rhodamine-labeled dextran (Rd-dextran) binding was first analyzed based on fluorescence enhancements, after which displacement by unlabeled mucin glycoprotein was tested based on subsequent decreases in fluorescence. These studies showcased that these compounds exhibited variable binding affinity toward dextran and mucin (Figure 5E–F). Additionally, BBALs were shown to improve liposomal delivery to vascular smooth muscle cells (VSMCs), which are important targets for treating peripheral vascular disease (PVD) (Figure 5B–D). While we have not yet evaluated cell-type selectivity, carbohydrate binding results yield promise that this platform could exhibit cell-type selectivity.

Figure 5.

A. Structures of bisboronic acid lipids 6a-d designed to enhance carbohydrate selectivity of liposomes. B-D. Images for vascular smooth muscle cell delivery studies show that PC liposomes (B) show minimal fluorescence resulting from liposome entry, while liposomes containing 10% (C) or 20% (D) 6c exhibit enhanced cell labeling, indicated by Rd-PE fluorescence. Scale bars denote 50 μm. DAPI is shown in blue, Rd-PE is shown in red. E-F. Results from a competitive microplate binding assay immobilizing increased amounts of liposomes containing 6a-d resulted in enhanced Rd-dextran binding (E). Subsequent mucin addition could displace Rd-dextran, as indicated by decreases in fluorescence intensities (F). Adapted with permission from ref.^[9b] Copyright 2022 Wiley-VCH.

An additional functional handle that is effective for driving cellular delivery through reversible covalent chemistry is the cyclic disulfide. Matile and co-workers showed that the ring tension in moieties including asparagusic acid (AA) and lipoic acid (LA) leads to enhanced reactivity toward thiols for disulfide exchange that can drive cell delivery.^[24] Therefore, attaching such groups onto cationic amphiphiles has been shown to enhance delivery of carriers including polymersomes and liposomes.^[25] This approach extends upon the use of the disulfide exchange reaction employing acyclic disulfides as an effective means for driving cellular delivery.^[26] This reaction has also been harnessed for applications including liposome triggered cargo release, which is distinct from the work described herein that focuses on driving cellular delivery. For example, Du and co-workers developed lipids containing disulfide groups within their acyl chains that were shown to release cargo upon treatment with dithiothreitol.^[27]

We have explored cyclic disulfide lipids (CDLs) 7a-b (Figure 6A)^[10] as a means for exploiting thiols presented on cell surfaces^[28] as well as for facilitating the decoration of liposome surfaces with functional groups. To confirm thiol reactivity and liposome derivatization, we developed an assay in which reaction with biotin-thiol conjugates enforced the immobilization of 7a-b onto streptavidin-coated microplates, which validated dose-dependent anchoring based on the percentage of 7a-b included within liposomes (Figure 6B). Having confirmed thiol reactivity, we then conducted cellular fluorescence microscopy experiments, which demonstrated that increased percentages of 7a-b incorporated within liposomes enhanced cellular uptake (Figures 6C–E). Therefore, this platform provides a promising avenue for exploiting cell-surface thiol groups to improve liposomal delivery.

Figure 6.

A. Structures of cyclic disulfide lipids 7a-b containing LA or AA moieties, respectively, designed to react with cell surface thiols through disulfide exchange, thereby enhancing liposome cell entry. B. Microplate binding results for immobilization of Rd-PE labeled PC liposomes containing various percentages (0–99%) of 7a-b post biotin-thiol incubation. Dose-dependent increases in Rd-PE fluorescence were observed, indicating successful surface modification. C-E. Fluorescence images for MDA-MB-231 cells treated with Rd-PE labeled 99.92% 7a (C), 99.92% 7b (D), or 99.92% PC (E), each doped with 0.08% Rd-PE. While PC control liposomes did not show liposome cell entry, both 7a and 7b liposomes exhibited a significant enhancement in Rd-PE fluorescence, indicating cell entry. Scale bars denote 100 μm. DAPI is shown in blue, Rd-PE is shown in red. Adapted with permission from ref.^[10] Copyright 2022 Wiley-VCH.

Beyond reversible covalent chemistry, caging groups that have long been used in chemical biology also show strong prospects for the activation of liposomal cellular delivery. While a variety of triggers have been explored for stimuli-responsive carriers,^[20,29] reactive oxygen species (ROS) deserve increased scrutiny as agents that are upregulated by diseased cells.^[30] The development of ROS-responsive agents benefits from the identification of arylboronate oxidation as a means for ROS detection by Chang and co-workers.^[31] Tsien et al. exploited this approach to develop caged CPPs for which cell delivery properties were activated upon oxidative removal of boronate cages by ROS.^[32] It should also be noted that liposome triggered release platforms driven by ROS have been reported using thioether-containing lipids by Li and co-workers^[33] as well as boronate-caged lipids by our group.^[34]

We have explored boronate caging as an avenue for controlling cell delivery using cationic liposomes, which interact through electrostatic attraction since mammalian cell membranes primarily present negative charge. Unfortunately, as a result of this charge uniformity, liposomes and other lipid nanoparticles are typically capable of inserting into most cells they encounter. To address this issue, we developed guanidine-lipid 8 bearing boronate caging groups (Figure 7A), which initially generate neutral liposomes that are not effective for cell entry.^[11] The encountering of ROS triggers the oxidative removal of the boronate caging group, loss of a self-immolating linker, and unveiling of a cationic guanidinium group to enhance cell entry. In this manner, initially inactive neutral liposomes are transformed through reaction with ROS into activated cationic liposomes. This approach is expected to provide an added level of control for directing liposomes to enter cells present within diseased tissue.

Figure 7.

A. Structure of ROS-responsive caged guanidine lipid 8. The boronate ester caging groups are designed to be oxidatively cleaved in the presence of ROS, resulting in quinone-methide immolation that unveils the cationic guanidine groups to activate cell entry. B. Zeta potential analysis for PC liposomes containing 0% or 20% lipid 8 treated with 0.1 mM, 0.5 mM, or 1 mM H₂O₂ over time. 8/PC liposomes showed gradual increases in ZP values, while PC control liposomes showed minimal changes. C. Results for a fluorescence-based microplate assay to evaluate binding/fusion between immobilized anionic PS membranes and uncaged 8/PC liposomes. Dose-dependent increases in Rd-PE fluorescence were observed for liposomes containing 8 after H₂O₂ incubation. D-E. Fluorescence images for MDA-MB-231 cells treated with Rd-PE labeled 20% 3/PC liposomes before (D) and after (E) H₂O₂ induced uncaging. Significant increases in Rd-PE fluorescence were seen for uncaged 8 liposomes. Scale bars denote 100 μm. DAPI is shown in blue, Rd-PE is shown in red. Adapted with permission from ref.^[11] Copyright 2022 Wiley-VCH.

To evaluate this strategy, we initially conducted zeta potential (ZP) analysis to demonstrate that treatment of initially neutral 8/PC liposomes with hydrogen peroxide gave way to strongly positively charged nanoparticles. We were then able to exploit ZP to follow the uncaging process over time and as a function of hydrogen peroxide concentration (Figure 7B). These experiments showed that uncaging was mostly complete within ~6 hours and H₂O₂ concentrations of 0.5 – 1 mM were effective at driving this process. Next, we developed a fusion/binding assay employing immobilized liposomes to probe whether the uncaging of 8 would drive interactions between liposomes (Figure 7C). Here, we found that biotinylated liposomes immobilized onto streptavidin plates^[35] containing the anionic lipid phosphatidylserine (PS) only interacted with fluorescently tagged liposomes containing 8 when the latter were treated with H₂O₂ to drive uncaging. With this verification of vesicle interactions, we next conducted fluorescence cellular microscopy experiments, which showed that fluorescently labeled 8/PC liposomes initially exhibited minimal delivery to MDA-MB-231 epithelial breast cancer cells, but that treatment with 1 mM H₂O₂ activated cell entry (Figures 7D–E). A series of control experiments showed expected negative results, including that delivery was not enhanced for liposomes lacking 8 treated with H₂O₂. This approach shows potential for overcoming the promiscuous nature of cationic lipids by employing caging groups to direct the activation of cationic liposomal carriers.

Outlook

The work described in this concept article showcases reversible covalent chemistry using recognition/reactive groups as well as caged cell-penetrating motifs as promising avenues for enhancing cellular delivery. These strategies exhibit many benefits, particularly since synthetic organic chemistry enables control over structure such that minimalist lipid analogs can be designed for cost-effective and precision control over liposome cell delivery characteristics. Nevertheless, many open questions remain that will determine the efficacy of this technology. For example, it remains to be seen how effective these approaches will be for achieving selectivity that enables differentiation of cell types, and particularly diseased versus healthy cells. Phrased another way, it will be key to determine the extent to which the structures and abundance of cell-surface glycans and thiol motifs can be leveraged to enhance liposomal delivery to specific target cells.

Another important point is whether the targeting moieties effectively respond to their stimuli at appropriate concentration ranges for efficacy in biological systems. For example, while liposomes containing ROS-responsive caged guanidine 8 begin to exhibit uncaging when treated with 100 μM H₂O₂, higher concentrations (~ 500 μM) are needed for significant uncaging. However, the increased concentrations of H₂O₂ associated with cancer cells have been cited as being in the ~100 μM range,^[30] indicating that liposomal platforms that respond to lower concentrations of ROS would be necessary. Nevertheless, this needs to be carefully managed since liposomes that respond to too low of a concentration of analyte target may be too reactive to achieve selective delivery. It should also be noted that the mechanisms for how cell delivery is enhanced have not yet been determined for these platforms, and therefore experiments to determine those would be invaluable for further understanding the driving force for entry. Finally, it should be noted that further work is needed to evaluate the toxicity of these platforms in different biological environments to further evaluate prospects.

Otherwise, there are many opportunities for expanding the toolbox for activating liposomal cell delivery by utilizing different recognition/reactive moieties to target distinct cell-surface biomolecules and functional groups or implementing different cages that respond to different stimuli. As presented in this concept article, these advancements can often be inspired by work in complementary fields such as chemical biology and molecular sensing. It is also plausible that certain targets/triggers will be more effective for activating cell delivery in potential clinical applications, based on considerations such as variations in concentration between diseased and healthy cells, accessibility, reactivity, etc. Nevertheless, this work provides a promising first step toward harnessing chemical differences between cell-surface and other metabolite abundances as a means for directing liposomes to enter cells. The ability to control liposomal activity by manipulating lipid conjugates that can be conveniently exchanged within liposome formulations provides a powerful tool for maximizing the potential for this approach. While clinical liposome delivery brings numerous obstacles to overcome, these strategies show promise for making contributions to circumvent these challenges.