Giant Unilamellar Lipid Vesicles

Giant Unilamellar Vesicles (GUVs), aqueous droplets stabilized by lipid bilayers, are useful model systems to study the physical properties of cell membranes. This application, however, depends on our ability to fabricate GUVs that faithfully mimic a target cell membrane. Unfortunately, conventional methods for vesicle production yield vesicles with polydisperse sizes and nonuniform lipid compositions, even within the same batch; this severely limits their use as model cell membranes. We propose a novel microfluidic approach for the production of GUVs using water-in-oil-in-water double emulsion drops with ultrathin shells as templates. We show that our approach results in GUVs with uniform sizes and compositions.

 GUV Bilayers as Model Cell Membranes

GUVs, aqueous droplets stabilized by lipid bilayers, are useful model systems for the study of the physical properties of cell membranes; however, this application depends critically on the ability of GUV bilayers to faithfully mimic target cell membranes. For example, when the bilayer of a GUV is composed of a mixture of different lipids, the GUV membrane exhibit spatial heterogeneities that might bear a certain resemblance with lipid rafts in cell membranes. This phenomenon of lipid phase separation provide a powerful means to both assemble and regulate the function of key proteins, and to tune the mechanical properties of the membrane as a whole. However, the physical characteristics or even the formation of lipid domains in GUVs are highly sensitive to several factors, such as the lipid composition of the membrane, the presence of any residual solvents in the membrane, and the GUV size; these factors must thus be carefully controlled during GUV fabrication. Unfortunately, GUVs produced using conventional approaches, such as electroformation or reverse emulsification, vary widely in both their size and their composition, even within the same batch. Former microfluidic approaches overcome these limitations, but introduce undesirable solvents with the GUV membrane; this is likely the reason why microfluidic-templated GUV bilayers do not exhibit lipid phase separations. It is therefore essential to continue developing new approaches to GUV fabrication that enable control over the GUV size, lipid composition, and the formation of microdomains within the GUV membrane.

Figure 1. (a) Schematic illustration of the glass-capillary microfluidic device used for production of water-in-oil-in-water double emulsion drops with ultra-thin shells. (b) Optical microscope image showing a typical production.

Microfluidic Devices for Double Emulsion Drops with Ultrathin Shells

Currently, we are using glass-capillary microfluidic devices to fabricate water-in-oil-in-water double emulsion drops with ultrathin shells; these emulsions serve as templates for GUV formation upon carefully controlled removal of the oil that forms the shell of the double emulsions. The device consists of two tapered cylindrical capillaries inserted into the opposite ends of a square capillary. We then insert another smaller cylindrical capillary into the left tapered capillary, as illustrated in Figure 1a. We use this smaller capillary to inject the innermost aqueous phase, which ultimately forms the aqueous core of the double emulsions. We treat the left tapered capillary to render its surface hydrophobic; this prevents wetting of the innermost aqueous phase on the capillary wall. We use this tapered capillary to inject the middle oil phase, a ternary lipid mixture of 35 mol.% 1,2-dioleyl-sn-glycero-3-phosphocholine (DOPC), 35 mol.% 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and 30 mol.% cholesterol, dissolved in a mixture of 36 vol.% chloroform and 64 vol.% hexane to a total lipid concentration of 5 mg/mL. We inject both the innermost aqueous phase and the middle oil phase at a flow rate of 500 mL/h; under these conditions the innermost aqueous phase forms large water-in-oil emulsion drops within the left tapered capillary. We then inject the outer aqueous phase, a 10 wt.% PVA solution, through the interstices between the left tapered capillary and the square capillary, at a flow rate of 5000 mL/h. We treat the right tapered capillary to render its surface hydrophilic, thus preventing wetting of the middle oil phase on its wall. This protocol forces the water drops to become re-emulsified, forming monodisperse water-in-oil-in-water double emulsion drops with ultrathin shells at the orifice of the right tapered capillary, as shown in Figure 1b. To prevent osmotic stresses, we collect these drops in a sucrose solution having the same osmolarity as the inner water cores.

Additionally, we are currently developing PDMS microfluidic devices, based on the same principles than those described above for our glass-capillary devices, which will allow for parallelization and thus for many industrial applications. 

Transformation of Double Emulsion Drops into GUVs

We use these double emulsion drops as templates for GUV formation. Importantly, our templates have ultrathin shells; this feature provides the template with an enhanced stability as compared to emulsions with thicker shells and allows for a better control over the lipid concentration. Unlike hexane, the chloroform in the ultrathin double emulsion shell is a good solvent for the lipids, enabling them to remain fully dissolved during double emulsion formation. The lipids adsorb to the interfaces between the inner core, ultra-thin shell, and the outer aqueous phase, thereby reducing the interfacial energy. Crucially, however, the chloroform is more soluble in water and evaporates more rapidly than the hexane, turning the shell into a hexane-rich poor solvent for the lipids. This reduction in the solvent quality induces an attractive interaction between the two lipid monolayers at the interfaces of the ultra-thin shell, forcing them to stick together and form a bilayer membrane when a random fluctuation bring them closer; this ultimately triggers a dewetting transition, that occurs over a duration of just a couple of minutes. Because the remaining solvent is less dense than the inner aqueous core, it forms a minuscule oil pocket at the top of the double emulsion, which continues to evaporate, producing a stable GUV. This dewetting process is shown in Figure 2.

Figure 2. Time series of overlayed confocal images of the DHPE-Rh (red) and naphtopyrene (blue) channels, which selectively associate with the liquid disordered and liquid ordered microdomains, respectively corresponding to the top plane of a double emulsion drop during its transformation into a GUV.

 Lipid Phase Separation within the GUV Membrane

Simultaneously to the above-described dewetting process, we observe the formation of lipid domains within the GUV membranes as shown in Figure 2. The ternary lipid mixture that we use is known to phase separate into two distinct lipid phases: a liquid disordered (ld) phase, which is enriched in DOPC, and a liquid ordered (lo) phase, which is enriched in DPPC and cholesterol. We distinguish this distinct phases by incorporating two different fluorescent dyes into the lipid mixture, 0.25 mol.% DHPE-rhodamine and 0.75 mol.% naphtopyrene; these selectively associate with the ld and lo phases, respectively. As the chloroform evaporates from the double emulsion shell, the lipids adsorbed to the interfaces between the inner core, ultra-thin shell, and the outer phase form microdomains at each of these interfaces. These microdomains are only slightly visible, as shown in the first frame of Figure 2. They are also highly dynamic and often fuse, as shown in the subsequent three frames of Figure 2. The ultrathin shell of the double emulsion then dewets from the innermost aqueous cores, as shown in the last two frames of Figure 2, forming a GUV that initially contains multiple ld and lo microdomains over its entire surface. Three optical sections, taken at three different depths within such a GUV, are shown in Figure 3a. These domains continue to fuse, thereby minimizing the line energy between their boundaries, ultimately forming two distinct ld and lo microdomains, as exemplified by the three optical sections shown in Figure 3b. These microdomains are highly monodisperse, as shown in Figure 3c, reflecting the exquisite control over the lipid compositions and GUV sizes afforded by our approach. These observations thus highlight the utility of forming GUVs from ultrathin shell double emulsion templates.

Figure 3. (a) Different optical sections, taken at different depths, of a GUV immediately after the dewetting process. (b) Different optical sections of the GUV after fusion of the microdomains, yielding a minimum-energy structure, a Janus-GUV. (c) Overlayed confocal fluorescent image of GUVs with liquid disordered and liquid ordered microdomains which shows the exquisite control over GUV size and composition.

 Conclusions So Far

Our microfluidic approach provides a versatile way to produce GUVs having carefully controlled sizes and lipid compositions. We produce these GUVs using double emulsion drops as templates; the drops have ultrathin shells, resulting in GUVs that contain a minimal amount of residual solvent within their membranes. This enables us to produce GUVs whose membranes can, for appropriate mixtures of lipids, phase separate into microdomains. Our work thus describes a route to produce GUVs that provide a useful model system for the study of the physical properties of cell membranes.

 Outlook - Membrane Mechanics

We are interested in the mechanical properties of GUV membranes composed of several lipid phases. Using the shapes of our GUVs, if subjected to osmotic stresses, or their characteristic thermal shape fluctuations we want to determine the membrane tension and bending modulus of each lipid phase. The exquisite control over the lipid composition and GUV radius will allow for accurate determinations of these parameters.

In addition, our technique allows for controlled encapsulation of polymers inside the GUV cores. We are interested on how the presence of crowded polymer mixtures, even exhibiting phase separation, within the GUV cores, affects the mechanics of the lipid membrane.  

Perfluorocarbon Shelled Double Emulsions

The ability of low boiling point liquid perfluorocarbons (PFCs) to phase change from liquid to gas upon ultrasound irradiation makes PFC-based emulsions promising vehicles for triggered delivery of payloads. However, loading hydrophilic agents into PFC-based emulsions is difficult due to their insolubility in PFC. Using microfluidic technologies, we fabricate perfluorohexane (PFH)-shelled double emulsions with large aqueous cores and thus high loading capacities for hydrophilic agents. Additionally, we study the response of these emulsions to ultrasound irradiation. Using a combination of optical and acoustic imaging methods, we observe payload release upon acoustic vaporization of PFH. Our studies demonstrate the utility of our microfluidic technique for controllably loading hydrophilic agents into PFH-based emulsions, which have great potential for acoustically-triggered release.

 Perfluorocarbon Emulsions for Advanced Delivery of Payloads

This work is a collaboration with Prof. Wynter Duncanson, Department of Chemical Engineering at Nazarbayev University and Prof. Tyrone Porter, Department of Mechanical Engineering at Boston University.

An important goal of advanced drug delivery is to controllably supply drugs to specific sites in the body, which often requires the use of carrier vehicles that efficiently encapsulate payloads and release them in response to an external trigger. Ultrasound is an excellent external trigger because it provides both spatial and temporal control over the transmission of thermal and mechanical energy; this enables highly-localized heating or mechanical disruption of carrier vehicles, and hence rapid release of entrapped payloads. Carrier vehicles for ultrasound drug delivery often contain small gas bubbles; these serve as cavitating bodies that concentrate acoustic pressure waves to facilitate disruption of the carrier vehicles. Unfortunately, such vehicles have limited shelf-lives due to the inherent instability of gas bubbles. A promising alternative is to utilize emulsion drops composed of low boiling point liquid perfluorocarbons (PFCs), which have longer shelf-lives; these undergo a liquid-to-gas phase transition when insonified that can be utilized to trigger the release on-demand. Conventionally, PFC emulsion drops are coated using polymer or lipids; these not only provide stability to the emulsion drops but also allow for drug loading. Therefore, payloads can either be dissolved in the emulsion drop or embedded in its coatings. However, due to the poor solubility of hydrophilic agents in both amphiphilic coatings and PFCs, the utility of these emulsion drops as carrier vehicles is restricted to hydrophobic or amphiphilic payloads. To address this limitation, hydrophilic agents are pre-dissolved in water and subsequently emulsified with PFCs through high shear mixing; unfortunately, this strategy leads to wide distributions in both loading capacities and drop sizes. These issues severely limit the utility of PFC-based emulsions as acoustically-activated vehicles for controlled delivery of hydrophilic payloads. Therefore, we consider essential to develop a microfluidic approach for the production of PFC-based emulsions with uniform sizes and controlled loading capacity for hydrophilic agents.

Microfluidic Fabrication of Perfluorohexane-Shelled Double Emulsion Drops

We propose the use of a flow-focusing glass capillary device to fabricate water-in-perfluorohexane-in-water (W/PFH/W) double emulsion drops. This device consists of two tapered cylindrical capillaries inserted into opposite ends of a square capillary; this configuration aligns the axes of the cylindrical capillaries. The injection capillary, on the left, is inserted into the collection capillary, on the right, as illustrated schematically in Figure 4a. To direct the flow of the fluids and controllably emulsify them, it is essential to appropriately modify the capillary surfaces. Therefore, the injection capillary is treated to make its surface fluorophilic; this favors the contact of the PFH with the outer wall of the injection capillary and focuses the PFH flow toward the collection capillary. In addition, the collection capillary is treated hydrophilic to prevent PFH from wetting the inner wall of the collection capillary. The inner water phase  is injected through the injection capillary, whereas the middle PFH phase, is injected through the interstices between the injection and square capillaries. The outer water phase, is injected through the interstices between the collection and square capillaries, and consequently flows in the opposing direction of the inner and middle phases; this focuses the inner, middle and outer phases at the orifice of the injection capillary. This configuration yields monodisperse W/PFH/W double emulsion drops, as shown in Figures 4b and 4c.

Figure 4. (a) Schematic illustration of the microfluidic device used to fabricate water-in-perfluorohexane-in-water double emulsion drops. (b) Optical microscope images showing the fabrication for typical flow rates of the inner, middle and outer phases of 500, 1000 and 2000 mL/h or (c) 750, 750 and 2000 mL/h, respectively, and the monodispersity of the resultant drops upon collection. Scale bars are 200 mm.

 Exposure of Perfluorohexane-Shelled Double Emulsion Drops to Ultrasound

Using this technology, we fabricate double emulsion drops with diameters on the order of several tens of microns; this allows us to easily monitor their response to ultrasound irradiation using optical microscopy. To conduct these studies, we isolate individual double emulsion drops in sealed chambers that contain de-ionized water, and place the chambers in a water bath. We expose the drops to continuous ultrasound irradiation using a tip sonicator. Prior to ultrasound excitation, a uniform distribution of sulforhodamine within the core of the double emulsion is observed as shown in the first frame of Figure 5a. Approximately 35s after ultrasound exposure, the red-colored core begins to darken, and immediately after we observe the sudden ejection of the red-colored hydrophilic dye from the double emulsion drop into the outer aqueous phase as shown in the last frame of Figure 5a.

 Triggered Release by Acoustic Vaporization of PFH

We hypothesize that the release of encapsulated material is triggered by the acoustic vaporization of the PFH shell. To test this hypothesis, we use an acoustic setup that combines high frequency ultrasound with a 5-10 MHz imaging transducer. We immobilize the PFH-shelled double emulsion drops into an acoustically-transparent hydrogel. Vapor PFH is more compressible and more echogenic than liquid PFH and thus appears brighter in a diagnostic ultrasound image. We focus high-amplitude ultrasound pulses into the hydrogels, and we increase the pressure until bright spots with an ellipsoidal shape appear in the diagnostic ultrasound images, as shown in Figure 5b; therefore, diagnostic ultrasound imaging confirms PFH vaporization.

Figure 5. (a) Time series of optical microscope images acquired with a color camera during ultrasound irradiation. (b) B-mode ultrasound image showing the presence of vaporized PFH-shelled double emulsions as bright ellipsoidal spots.

 Conclusions so Far

The microfluidic technique that we describe above allows for the production of monodisperse perfluorohexane-shelled double emulsions with controlled loading capacity for hydrophilic agents. We show the release of a model hydrophilic agent upon application of ultrasound due to perfluorohexane vaporization. These observations highlight the utility of using ultrasound to precisely control the vaporization of PFH to induce payload release; this enhances the potential of our approach for acoustically trigger and monitor drug release.   

Outlook - Physical Mechanism of Acoustic Droplet Vaporization

The control provided by our microfluidic technology over the thickness of the perfluorohexane shell will allow to study the appearance of unique acoustic signals during vaporization of perfluorohexane. We expect to shed light on the physical mechanisms responsible for acoustic vaporization by analyzing this unique signals.