Triggered Release from Liposomes through Magnetic Actuation of Iron

Feb 25, 2011 - Effect of Gold Nanorods on the Remote Decapsulation of Liposomal Capsules Using Ultrashort Electric Pulses. Yu. V. Gulyaev , V. A. ...
0 downloads 4 Views 3MB Size
LETTER pubs.acs.org/NanoLett

Triggered Release from Liposomes through Magnetic Actuation of Iron Oxide Nanoparticle Containing Membranes Esther Amstad,† Joachim Kohlbrecher,‡ Elisabeth M€uller,§ Thomas Schweizer,|| Marcus Textor,† and Erik Reimhult*,†,^ †

Laboratory for Surface Science and Technology, ETH Zurich, Wolfgang-Pauli-Strasse 10, CH-8093 Z€urich, Switzerland Laboratory for Neutron Scattering, Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland § EMEZ, Electron Microscopy ETH Z€urich, Wolfgang-Pauli-Strasse 16, CH-8093 Z€urich, Switzerland Polymer Chemistry, Department of Materials, ETH Z€urich, Wolfgang-Pauli-Strasse 10, CH-8093 Z€urich, Switzerland ^ Department of Nanobiotechnology, University of Natural Resources and Life Sciences (BOKU), Muthgasse 11/II, A-1190 Vienna, Austria

)



bS Supporting Information ABSTRACT: The ideal nanoscale drug delivery vehicle allows control over the released dose in space and time. We demonstrate that this can be achieved by stealth liposomes comprising self-assembled superparamagnetic iron oxide nanoparticles (NPs) individually stabilized with palmityl-nitroDOPA incorporated in the lipid membrane. Alternating magnetic fields were used to control timing and dose of repeatedly released cargo from such vesicles by locally heating the membrane, which changed its permeability without major effects on the environment. KEYWORDS: Stealth liposome, superparamagnetic iron oxide nanoparticle, triggered release, drug delivery vehicle, nanoreactor, palmityl-nitroDOPA

N

anoscale vesicles are crucial not only for drug delivery applications1-3 but also of high interest to perform ex vivo encapsulation, delivery, and nanoscale chemistry.4,5 Irrespective of the application, it is highly beneficial to externally trigger release of a controlled dose of the encapsulated cargo at a specific time and location,6 as opposed to release that merely relies on passive diffusion or in response to global environmental changes.7 The functional versatility of liposomes renders them one of the most intensively investigated delivery vehicles.3 Typically 100 nm in diameter, biocompatible, and possible to engineer for long in vivo circulation times, liposomes can carry hydrophilic cargo in the aqueous lumen and hydrophobic cargo in the lipid membrane interior.3,8 Their permeability is greatly enhanced around the membrane melting temperature (Tm),9 which depends on the lipid composition.10 Cargo can thus be released if the liposome membrane is heated above Tm. For use of liposomes as thermoresponsive drug delivery vehicles, Tm is typically designed to be close to body temperature to release cargo at the few degrees higher temperature of pathological tissue like cancer. This results in leakage during circulation.11 If Tm instead is chosen far above body temperature, release is inefficient.12,13 To circumvent the incompatible requirements of simultaneous release efficiency and low passive leakage, liposomes have been loaded with hydrophilic plasmonic14 and magnetic15,16 r 2011 American Chemical Society

nanoparticles (NPs) to trigger cargo release with light and high frequency alternating magnetic fields (AMF), respectively. Release using Au NP loaded liposomes can only be triggered in optically transparent media with particles with a diameter >50 nm.17 Only these particles have resonances in the optical or near IR regime. No such restrictions are imposed on magnetic NPs.18 Very recently, triggered release was shown for liposomes having oleic acid-coated iron oxide NPs associated with their membranes.19 However, these liposomes agglomerated at room temperature and were inherently leaky. Furthermore, oleic acidcoated NPs agglomerate19 and thereafter are unlikely to incorporate into liposome bilayers due to size constraints.20 In contrast, we report PEGylated liposomes with Tm far higher than body temperature hosting individually stabilized iron oxide NPs in their membranes in a well-defined assembled structure (Figure 1). As a result, these liposomes are colloidally stable and impermeable at body temperature. Stable incorporation of NPs into liposome membranes allowed repeated, controlled release of cargo triggered with AMF. The release efficiency was so high that we could choose AMF settings to release cargo without increasing the bulk temperature to Tm even without external cooling. These properties were shown to relate directly to the structure and Received: January 14, 2011 Revised: February 10, 2011 Published: February 25, 2011 1664

dx.doi.org/10.1021/nl2001499 | Nano Lett. 2011, 11, 1664–1670

Nano Letters

LETTER

Figure 1. Schematic of liposomes containing iron oxide NPs in their bilayer. NitroDOPA-palmityl stabilized iron oxide NPs are embedded in liposome membranes consisting of PEGylated and unmodified lipids.

stability of the NP-lipid assemblies investigated by small angle neutron scattering (SANS), transmission electron microscopy (TEM), differential scanning calorimetry (DSC), thermogravimetry analysis (TGA), and fluorescence spectroscopy. Iron oxide NPs with average core diameters of 5 nm (denoted small NPs) and 10 nm (denoted large NPs) surrounded by a palmityl-nitroDOPA shell with a dispersant packing density of ∼1.5 palmityl-nitroDOPA/nm2 (see Supporting Information; Figure S1) were dispersed in chloroform and mixed with 2-distearoyl-sn-glycero-3-phosphocholine (DSPC) lipids containing, unless stated otherwise, 5 mol % 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)2000 Da] (PEG(2)-PE). After chloroform was removed under continuous N2 flow, lipids were redispersed in aqueous media and extruded. Individually palmityl-nitroDOPA-stabilized iron oxide NPs spontaneously incorporated into liposome bilayers, whereas oleic acid stabilized NPs agglomerated to form micelles. This was revealed by TEM investigations (Figure 2a-d) and by the much darker color observed with optical inspection (Figure 2) of as-extruded liposomes. TEM micrographs of trehalose-fixed DSPC liposomes comprising small iron oxide NPs clearly revealed the presence of well dispersed palmityl-nitroDOPA stabilized NPs associated with the liposomes (Figure 2d). If these hydrophobic NPs were free in suspension, they would immediately agglomerate. Thus, these NPs must be embedded in phospholipid membranes. Furthermore, energy dispersive X-ray (EDX) analysis performed in situ with scanning TEM (STEM) on freeze-dried liposomes doped with small palmityl-nitroDOPA stabilized iron oxide NPs allowed to colocalize iron and phosphorus (see Supporting Information Figure S2), also indicating iron oxide NPs are embedded in phospholipid membranes. The striking difference between palmityl-nitroDOPA and oleic acid-stabilized NPs can be related to the binding affinity of the dispersant to iron oxide. While nitroDOPA was shown to irreversibly bind dispersants to iron oxide surfaces,21,37 oleic acid is a reversibly adsorbing dispersant.21 Reversible dispersant adsorption renders oleic acid stabilized NPs prone to agglomeration. Note that NP aggregation could have occurred either within the lipid membranes or before incorporation of the NPs in liposome membranes. Oleic acid stabilized 3-15 nm iron oxide NPs mixed with phospholipids in organic solvents and redispersed in aqueous media have been shown to be surrounded by a phospholipid monolayer yielding stable micelles in aqueous media.22,23 Thus, the micelles found by cryo-TEM (Figure 2b)

Figure 2. Liposomes functionalized with iron oxide NPs. Cryo-TEM images of DSPC liposomes containing 5 mol % PEG(2)-PE that (a) were unmodified and incorporated (b) oleic acid coated and (c) palmityl-nitroDOPA stabilized small iron oxide NPs. Insets show photographs of the respective PBS-based liposome dispersions where the lipid concentration was kept constant at 5 mg/mL. A comparison between (a) and (c) reveals no significant change of the spherical shape of liposomes upon loading their membranes with small, individually stabilized, iron oxide NPs. However, agglomerated, oleic acid stabilized NPs seem to significantly distort the liposome shape. (d) TEM image of trehalose fixed DSPC liposomes containing palmityl-nitroDOPA stabilized small NPs in their membranes. Liposomes were fixed with trehalose and air-dried on a carbon-supported Cu TEM grid where the carbon film had 3.5 μm diameter holes. While the large image was taken in a hole that was spanned by trehalose, the inset was imaged on the carbon support. Individually stabilized NPs with core diameters 5.5 nm are seen. The inset indicates a high NP density of liposomes that were collapsed on the carbon support upon drying in air.

were likely hydrophobic NPs stabilized by a phospholipid monolayer. That agglomerated NPs cannot be incorporated into liposome membranes is related to their size. Theoretical studies suggest a threshold maximum NP diameter of 6.5 nm20 for incorporation of neutral24 NPs into lipid bilayers. Larger or charged cores preferentially form micelles. This size limit is well in agreement with the TEM images where no cores >5.5 nm were seen in liposomes (Figure 2d and Supporting Information Figure S3). Because of the broad core size distribution, even large NPs (with average core diameter of 10 nm) contain a fraction with core sizes 73 to 131 did not influence Tm if embedded in zwitterionic membranes. In contrast, oleic acid coated iron oxide NPs embedded in zwitterionic liposome membranes have been reported to increase Tm19 whereas TOPO stabilized CdSe/ ZnS QDs decreased and broadened Tm of cationic liposomes.32 Reversibly adsorbed dispersants, such as oleic acid, that dissociate from the NP surface diffuse into the membrane. Organic solvents and amphiphiles are well-known to influence membrane properties such as membrane fluidity and Tm. A reason for the contradictory literature on the influence of NPs on membrane fluidity and Tm might thus be a varying degree of ligand dissociation in addition to differences in NP loading and core sizes. Thus, irreversibly adhering hydrophobic dispersants are most likely not only crucial to prevent NP agglomeration but also

to avoid that membrane properties are altered in an uncontrolled way. To use liposomes as delivery vehicles, they have to be colloidally stable at 25-37 C. However, as can be seen in the continued increase of the SANS intensity at small q-values (Figure 4), DSPC liposomes agglomerated also without incorporated NPs if stored at T < Tm in agreement with literature reports.31 Incorporation of 5 mol % PEG(2)-PE into the liposome bilayer lead to an approximately 2.1-2.2 nm thick PEG shell (Figure 1 and Supporting Information Table S2). PEG was shown to be in the mushroom to brush transition regime under these conditions.33 PEGylated liposomes were stable for at least 4 weeks even though they were stored in the gel phase, as was indicated by DLS and SANS measurements (Figure 3). Thus, incorporation of PEGylated lipids into liposomes not only renders these vesicles stealth34 and thus prolongs their circulation time in vivo but also colloidally stable at 25 C. However, the likelihood that micelles are formed increases with increasing amount of PEGylated lipids in the liposomes.35 Thus, the concentration of PEGylated lipids in liposomes should be sufficient to provide good liposome stability but low enough to prevent micelle formation. To probe permeation and actuation of liposomes containing iron oxide NPs in their membranes, they were loaded with calcein, a self-quenching and membrane impermeable dye. PBS-based dispersions containing 0.5 mg/mL calcein-loaded DSPC/PEG(2)-PE liposomes were subjected to AMF pulses. Dispersions of liposomes functionalized with palmityl-nitroDOPA stabilized iron oxide NPs contained ∼0.05 mgiron oxide/ml. As can be seen in Figure 5a, fluorescence rapidly increased upon subjecting liposomes containing palmityl-nitroDOPA-stabilized NPs in their membrane to 5 subsequent, 5 min lasting, 230 kHz, AMF pulses. The system was equilibrated for 1 min between the individual AMF pulses (denoted as AMF sequence 1). If calcein penetrates the liposomal membrane, the calcein concentration in bulk water is below the self-quenching concentration. Therefore, fluorescence 1667

dx.doi.org/10.1021/nl2001499 |Nano Lett. 2011, 11, 1664–1670

Nano Letters

LETTER

Figure 6. Release efficiency of liposomes functionalized with small iron oxide NPs. (a) The fluorescence of DSPC/PEG(2)-PE liposomes functionalized with palmityl-nitroDOPA-stabilized small iron oxide NPs and loaded with calcein at a self-quenching concentration was monitored as a function of time liposomes were exposed to different AMF sequences. The fluorescence increased slower if the system was equilibrated for 5 min in between each 5 min long AMF cycle (sequence 2) compared to the change in fluorescence of liposome dispersions equilibrated only for 1 min (sequence 1) (-9-). Fluorescence increased linearly with exposure time also if liposomes were treated with AMF sequence 1 and 2 in stark contrast to the constant fluorescence measured for liposomes treated with AMF sequence 1 that did not contain any iron oxide NPs (-b-). Fluorescence of dispersions of liposomes containing palmityl-nitroDOPA-stabilized small iron oxide NPs in the membrane remained constant if these dispersions were subjected to 10 cycles of 1 min AMF pulses followed by 1 min equilibration (sequence 3) (-0-). An increase in fluorescence directly translates into release of calcein. (b) Bulk temperatures of liposome dispersions subjected to the respective AMF sequence used in (a). (c) The fluorescence of dispersions of DSPC/ PEG(2)-PE liposomes functionalized with palmityl-nitroDOPA stabilized small NPs and loaded with calcein at a self-quenching concentration was measured as a function of temperature. The fluorescence only increased significantly at T > 50 C if the entire dispersion was heated with a thermostat (-0-). As comparison, the change in fluorescence of liposomes that do not contain any iron oxide NPs (-b-) and liposomes containing small iron oxide NPs in their membranes (-9-) exposed to an AMF is shown as function of temperature. The comparison demonstrates that the cargo release seen in (a) cannot be attributed to bulk water heating and has to be caused by local membrane heating by iron oxide NPs if subjected to an AMF. The lipid concentration in all experiments was 0.5 mg/mL while the iron oxide NP concentration for liposomes loaded with palmityl-nitroDOPA-stabilized iron oxide NPs ≈ 0.05 mgiron oxide/ml.

increases with increasing diffusion of calcein through the liposome membrane. Thus, the increased fluorescence of dispersions of calcein-loaded liposomes containing palmityl-nitroDOPA-stabilized NPs in their membrane upon subjection to an AMF demonstrated AMF triggered calcein release. The increase in fluorescence leveled off after these solutions had been exposed to 5 subsequent AMF pulses (Figure 5a). This was interpreted as complete calcein release. The increased permeability of liposomes upon subjection to AMF pulses is a result of the heat locally generated by iron oxide NPs through Neel relaxations.36 This heat is directly dissipated into the surrounding liposome bilayer bringing the temperature across the Tm of the lipid membrane. As described in the introduction, temperature T in the bilayer lipid membranes have an increased permeability at temperatures close to Tm.9 The unchanged hydrodynamic diameter of liposomes loaded with iron oxide NPs before and after AMF treatment and the absence of iron oxide NP precipitation after AMF treatment demonstrates that the liposomes remained intact and at constant size during this treatment (Figure 5b). This observation confirms that the calcein release is dominated by a change in membrane permeability and not liposome rupture or fusion. If liposomes ruptured and fragmented during AMF treatment, palmityl-nitroDOPA stabilized iron oxide NPs would be exposed to water, which leads to aggregation and precipitation as control experiments showed. However, no precipitation of NPs was seen upon AMF treatments. Alternatively, liposome rupture could lead to reformation of vesicular structures or bicelles incorporating NPs, however, no indication of the corresponding size change was observed by DLS (Figure 5b). Hence, cargo release relies on the increased liposome permeability around Tm as mechanical distortion of the membrane at these frequencies is unlikely for small NPs heating through Neels relaxation.36 Because the liposome structure was retained during AMF exposure, calcein release

ceased when the AMF was turned off (Figure 5a). This allowed close control of the released dose. During repeated exposure of the liposome solution to AMF sequence 1, the bulk water temperature increased to 45 C without cooling (Figure 6a,b). The largest release occurred already during the first pulse for which the temperature increased to 37 C. Control experiments where liposomes containing iron oxide NPs were externally heated using a thermostat only resulted in significantly increased fluorescence if the temperature was 50 C, which is ∼5 C below Tm (Figure 6c). In contrast to what has been reported previously,19 no significant change in fluorescence of dispersions of liposomes without iron oxide NPs in the membrane was observed upon subjecting them to AMF sequence 1 (Figure 5a). This indicates low passive leakage of liposomes. The much reduced leakiness of the liposomes investigated here can be explained by the higher stability of liposomes sterically shielded with PEG(2)-PE and the higher stability of the NPs incorporated in the membranes than what was reported for dipalmitylphosphocholine (DPPC) liposomes incorporating oleic acid coated NPs.19 The change in fluorescence and therefore release efficiency of liposomes doped with palmityl-nitroDOPA-stabilized small and large iron oxide cores was comparable. However, it exceeded that of liposome dispersions containing oleic acid coated iron oxide NPs at complete release by 180% (Figure 5a). The difference in release at the initial pulses close to body temperature was even more dramatic; since it was negligible for oleic acid coated NPs while it was largest for nitroDOPA-stabilized NPs. The release efficiency of liposomes containing oleic acid-stabilized NPs, however, was comparable to that of magnetoliposomes incorporating hydrophilic, PEG(1.5 kDa)-nitroDOPA stabilized iron oxide NPs in their lumen (Figure 5a). This indicates that only few oleic acid stabilized NPs were incorporated into the liposome membranes in good agreement with cryo-TEM images. 1668

dx.doi.org/10.1021/nl2001499 |Nano Lett. 2011, 11, 1664–1670

Nano Letters The much smaller change in fluorescence of dispersions of liposomes containing PEG(1.5 kDa)-nitroDOPA stabilized NPs in their lumen compared to liposomes hosting palmityl-nitroDOPA stabilized NPs in the membrane for the same AMF sequence points to a many times higher release efficiency of the latter liposomes. The greatly improved release efficiency of liposomes containing individually stabilized hydrophobic iron oxide NPs in their membranes can be assigned to a direct transfer of heat into the liposome membrane, as generated by the iron oxide NPs upon subjection to an AMF.18 In contrast, AMFinduced heating through PEG(1.5 kDa)-nitroDOPA-stabilized hydrophilic iron oxide NPs localized in the lumen of the liposomes has to heat bulk water to temperatures approaching Tm to release cargo. The requirement to strongly heat bulk water significantly reduces release efficiency. Furthermore, it prevents release of thermally sensitive chemicals, drugs, and proteins from such magnetoliposomes due to thermal degradation and loss of functionality of the cargo during release. It would also preclude the use in cell cultures and tissue where heating of the bulk liquid will kill surrounding cells. If the AMF sequence was altered such that the system was exposed to 5 min AMF pulses but equilibrated for 5 min between each AMF pulse (AMF sequence 2), the bulk temperature increase stabilized at 38 C (Figure 6b). Although changes in the fluorescence of dispersions of PEGylated liposomes containing small iron oxide NPs in their membrane were smaller if liposomes were treated with AMF sequence 2 instead of sequence 1, continued increase in fluorescence that can be translated into continued cargo release was demonstrated over 12 subsequent AMF cycles (Figure 6a). The slower but continued release is likely related to the effective time a sufficiently large area of the membrane has a temperature close to Tm. No significant change in fluorescence of any liposome dispersion was observed for exposure to 10 subsequent AMF pulses lasting 1 min followed by 1 min equilibration (AMF sequence 3). This indicates that cargo cannot be released with AMF sequence 3. This likely is because the local heat generated by the iron oxide NPs during the first minute of AMF treatment is insufficient to increase the temperature inside the bilayer to close to the Tm ≈ 55 C of DSPC, since the heat is rapidly dissipated in the aqueous environment. Thus, the total exposure time required to release all encapsulated cargo increases with decreasing length of AMF pulses. However, the possibility to release cargo with pulsed AMF sequences paves the way to not only trigger release and control the dose but also to slowly release cargo over prolonged times at arbitrary release profiles and at bulk temperatures close to body temperature. In summary, iron oxide NPs with core diameters