Langmuir 2009, 25, 3999-4008
3999
Interaction between Lamellar (Vesicles) and Nonlamellar Lipid Liquid-Crystalline Nanoparticles as Studied by Time-Resolved Small-Angle X-ray Diffraction† Pauline Vandoolaeghe,*,‡ Justas Barauskas,§ Markus Johnsson,| Fredrik Tiberg,‡,| and Tommy Nylander‡ Center for Chemistry and Chemical Engineering, Physical Chemistry 1, Lund UniVersity, P.O. Box 124, SE-22100 Lund, Sweden, Department of Bioanalysis, Institute of Biochemistry, Mokslininku 12, LT-08662 Vilnius, Lithuania, and Camurus AB, Ideon Science Park, SE-22370 Lund, Sweden ReceiVed August 26, 2008. ReVised Manuscript ReceiVed NoVember 20, 2008 The kinetics of structure change when dispersions of two different types of lipid-based liquid-crystalline phases, one lamellar and one reversed, are mixed has been investigated using synchrotron small-angle X-ray diffraction and ellipsometry. The systems studied were (i) cubic-phase nanoparticles (CPNPs) based on glycerol monooleate (GMO) stabilized with a nonionic block copolymer, Pluronic F-127; (ii) CPNPs based on phytantriol (PtOH) stabilized with D-R-Tocopheryl polyethylene glycol 1000 succinate (Vitamin E TPGS); and (iii) hexagonal-phase nanoparticles (HPNPs) based on a lipid mixture of diglycerol monooleate/glycerol dioleate, stabilized by Pluronic F-127. Timeresolved small-angle X-ray diffraction was used to track structural changes within nonlamellar nanoparticles when they interact with uni- and multilamellar vesicles of dioleoylphosphatidylcholine and dipalmitoylphatidylcholine. The results are very dependent on the type of nanoparticles under investigation. For GMO-based CPNPs, a strong interaction is observed on mixing with vesicular dispersions that leads to large changes in unit size dimensions as well as a later transition from cubic to lamellar structure. These results are in good agreement with previous studies on the interaction of GMO-based CPNPs with planar bilayers using neutron reflectivity, where the diffraction peak shifted with time upon mixing. The structural changes are much less prominent for the PtOH-based CPNPs and the HPNPs upon mixing with phospholipid vesicles. These results are correlated with those from measurement studying interactions between the liquid-crystalline nanoparticles and supported phospholipid bilayers by ellipsometry. Also, here the GMO-based CPNPs show more pronounced and rapid adsorption and interaction with the supported bilayer surface than do the other types of nonlamellar nanoparticles. The interaction also depends on the bilayer properties, where significantly slower lipid mixing is observed for a bilayer in the gel state compared to a bilayer in the liquid-crystalline phase. This study is not only relevant for drug-delivery applications but also shows the potential of synchrotron small-angle X-ray diffraction in studying time-dependent structural changes as a consequence of the interaction between different lipid self-assembled aggregates in complex systems.
Introduction Lipids and their mixtures exhibit rich phase behavior and have the properties to form a wide range of self-assembled nanostructures. Liquid-crystalline nanoparticles of different internal structure, size, and morphology have been prepared using a range of lipids and lipid mixtures and furthermore have been studied in relation to their application as drug-delivery carriers.1-5 Nonlamellar nanoparticles have, in this respect, a number of potentially advantageous properties as compared with emulsions and liposomes. This includes the favorable solubilization and encapsulation ability of many drugs and also excellent physical stability at high lipid content, enabling large drug loads and smaller dose volumes. The ease of preparation and ability to tune the internal particle structure from being built up by lipidsurrounding discrete aqueous domains to bicontinuous structures † Part of the Neutron Reflectivity special issue. * Corresponding author. Phone: (46) 46 222 3332. Fax: (46) 46 222 4418. E-mail:
[email protected]. ‡ Lund University. § Institute of Biochemistry. | Camurus AB.
(1) Lasic, D. D. Liposomes: From Physics to Applications; Elsevier: New York, 1993. (2) Lasic, D. D. J. Controlled Release 1997, 48, 203–222. (3) Ulrich, A. S. Biosci. Rep. 2002, 22, 129–150. (4) Larsson, K. Lipids—Molecular Organization, Physical Functions and Technical Applications; The Oily Press Ltd: Dundee, 1994. (5) Bicontinuous Liquid Crystals; CRC: Boca Raton, FL, 2005; Vol. 127.
are also interesting in relation to encapsulation and release properties. The fact that it has been shown to be possible to prepare particle dispersions with extremely well-defined interior phase structure, particle shape, and narrow size distributions6-8 has opened up possibilities to obtain a better understanding of structure-function relationships in different applications. In the present study, we have been interested in interactions between nonlamellar liquid-crystalline layers and lipid bilayer structures as a simple model of cell membranes. The kinetics of changes in structure when dispersions of two different types of liquidcrystalline phases, one lamellar and one reversed, are mixed has never before been studied to our knowledge. This study is aimed at determining the correlation between nanoparticle formulations and the impact of this formulation on biological phenomena, such as cell adhesion and hemolysis. The intermediate structures formed when cell membranes fuse can be regarded as between lamellar and reversed structures (e.g., inverted and HII phases).9,10 Phospholipid bilayers have been used extensively as model membranes11 in the form of particle dispersions (i.e., vesicles or (6) Barauskas, J.; Johnsson, M.; Tiberg, F. Nano Lett. 2005, 5, 1615–1619. (7) Barauskas, J.; Johnsson, M.; Joabsson, F.; Tiberg, F. Langmuir 2005, 21, 2569–2577. (8) Johnsson, M.; Tiberg, F. U.S. Patent WO/2006/077362, 2006. (9) Siegel, D. P.; Burns, J. L.; Chestnut, M. H.; Talmon, Y. Biophys. J. 1989, 56, 161–169. (10) Siegel, D. P.; Green, W. J.; Talmon, Y. Biophys. J. 1994, 66, 402–414. (11) Sackmann, E. Science 1996, 271, 43–48.
10.1021/la802768q CCC: $40.75 2009 American Chemical Society Published on Web 01/06/2009
4000 Langmuir, Vol. 25, No. 7, 2009
liposomes12-14) and as supported bilayers on a planar surface.11 Multilamellar vesicles (MLVs) are readily formed by dispersing the lipid in water, whereas the preparation of unilamellar vesicles (ULVs) typically requires freeze-thaw cycles, followed by sonication or extrusion through a defined pore size membrane.15,16 Supported bilayers can be prepared using different techniques: Langmuir-Blodgett deposition,17-19 Langmuir-Schaeffer technique,20 vesicle fusion,21-23 spin coating,24 and surfactant depletion.25 The method chosen in this study is the surfactant depletion method reported by Tiberg et al.25 In previous work, we have studied the interactions of GMObased CPNPs with supported DOPC bilayers by means of ellipsometry, quartz crystal microbalance with dissipation monitoring, and neutron reflectivity.26,27 The results showed an initial rapid adsorption of intact CPNPs at the bilayer interface, followed by an exchange of lipid material, which results in a net release of particles. Studies of the interactions of nanoparticles with bilayers of different coverage show that the structure and composition of the adsorbed layer at the end stage greatly depend on the initial bilayer coverage. Exchange of material between CPNPs and the bilayer was observed in all cases. For a highcoverage bilayer (80%), extensive exchange takes place between the CPNP components and the bilayer, and at steady state the surface layer comprises 72% glycerol monooleate and 8% DOPC, with no change in the solvent content. There may be residual CPNPs (96%), denoted as phytantriol (PtOH), was kindly provided by Kuraray Co., Ltd. (Japan). A triblock copolymer containing ethylene oxide (EO) and propylene oxide (PO) groups, with the trade name Pluronic F-127 and an approximate formula of EO98PO57EO98 (average molecular weight of 12 600 g mol-1), was obtained from BASF Svenska AB (Helsingborg, Sweden). D-R-Tocopheryl polyethylene glycol 1000 succinate (vitamin E TPGS) was provided by Eastman (Workington, U.K.). Dioleoylphosphatidylcholine (DOPC) and dipalmitoylphophatidylcholine (DPPC) were purchased from Avanti Polar Lipids (Alabaster, AL). Sterile water from B. Braun Medical AB (Bromma, Sweden) was used for the preparation of the nanoparticle dispersion stock solution. All other solutions were prepared with Milli-Q water (a resistivity of 18 MΩ cm and a total organic content of 3 to 4 ppb). Lipid Liquid-Crystalline Nanoparticle (LCNP) Preparation. Three dispersions of nanoparticles with different compositions were used in this study: (i) CPNP dispersions based on GMO and Pluronic F-127 (9/1 w/w); (ii) CPNP dispersions based on PtOH and vitamin E TPGS (88/12 w/w); and (iii) HPNP dispersions based on a DGMO/ GDO lipid mixture (1/1 w/w) and F127 (9/1 w/w). All dispersions were prepared at 95 wt % in sterile water. Unrefined dispersions were prepared by adding appropriate amounts of melted (40 °C) GMO, PtOH, or DGMO/GDO to a solution containing Pluronic-F127 or vitamin E TPGS. The samples were immediately sealed, vigorously shaken, and then left to stir via a vortex mixer for 24-48 h on a mechanical mixing table at ∼350 rpm at room temperature. The resulting dispersions were homogenized by passing six times through a Microfluidizer 110S (Microfluidics Corp., Newton, MA) at 345 bar and 25 °C. In the case of the DGMO/GDO/F127-based hexagonal-phase nanoparticles (HPNP), no homogenization was needed. The samples were transferred to Pyrex glass bottles (50-500 mL) and then subjected to heat treatment using a bench-type autoclave (CertoClav CV-EL, Certoclav Sterilizer GmbH, Traun, Austria) operating at 125 °C and 1.4 bar vapor pressure. A period of about 12 min was required to vent the entrapped air and to heat up the autoclave, and the samples were heated for 20 min at 125 °C. After the heat treatment, the
Liquid-Crystalline Nanoparticle Interactions
Langmuir, Vol. 25, No. 7, 2009 4001 Table 1. Ratios (n/n) and (w/w) and Estimated Values of the Ratio of the Number of Vesicles to the Number of LCNPs for the Different Experiments Performeda experiment GMO-based CPNPs + DOPC ULVs
GMO-based CPNPs + DOPC MLVs
Figure 1. Particle size distribution of the cubic-phase nanoparticle dispersions based on GMO and PtOH and the hexagonal-phase nanoparticle dispersions based on DGMO/GDO as measured by a laser diffraction particle size analyzer. The average sizes were 382, 299, and 196 nm with polydispersity indexes (ratio between standard the deviation and mean size) of 0.15, 0.21, and 0.23 for the GMO-based CPNPs (1), PtOH-based CPNPs (2), and DGMO/GDO-based HPNPs (3), respectively.
samples were allowed to cool to room temperature before the dispersions were filtered through a 2 µm Acrodisc syringe filter (Pall Corp., Ann Arbor, MI) to remove small amounts of precipitates that sometimes appeared at the air/liquid interface during cooling. The internal structure of the cubic nanoparticles based on GMO and PtOH comprises the body-centered cubic Im3m space group (no. 229), and the DGMO/GDO-based hexagonal particles originate from the reversed hexagonal phase (HII phase), as determined by SAXD in earlier characterization studies of these systems.6,7,28 Particle Size Distribution. The particle size distribution was measured using a Coulter LS230 laser diffraction particle size analyzer (Beckman-Coulter, Inc., Miami, FL), which operates on the principles of Fraunhofer diffraction for large particles (0.4-2000 µm) and uses the polarization intensity differential scattering method for small particles (0.04-0.5 µm). The instrument was fitted with a 125 mL volume module. Data were collected for 90 s. A standard model based on homogeneous oil spheres with a refractive index of 1.46 was used for the particle size calculations. Reasonable estimates of the uncertainty in the refractive index can shift the obtained particle diameter distributions by only a few percent. Note that the model is based on spherical particles and the measured mean particle size is calculated on the basis of this assumption. Figure 1 shows the particle size distributions obtained for the three different nonlamellar particle dispersions used in this study. Vesicle Preparation. The appropriate amount of phospholipid (DOPC or DPPC) was dissolved in chloroform and dried in a first step with a stream of N2. The phospholipids were then transferred to a vacuum oven for at least 24 h to ensure complete removal of the solvent, and the appropriate volume of Milli-Q water was added. The resulting multilamellar vesicle dispersions were then exposed to 20 freeze-thaw cycles with intermittent vortex mixing to reduce the number of bilayers in the formed vesicles.15,16 A thermostatted bath was heated to 50 °C for the heating step, and another bath composed of a mixture of dry ice and ethanol was used for the cooling step. The average particle size was determined using a Zetasizer Nano ZS dynamic light scattering instrument (Malvern Instruments Ltd., Worcestershire, U.K.) at 25 °C. The particle size was measured assuming a refractive index of 1.46, and the average particle size was found to be 350 nm. To form the DOPC unilamellar vesicles, the multilamellar vesicles obtained after the freeze-thaw cycles were extruded at room temperature (far above the gel-to-liquid-crystalline temperature of -19 °C) by means of a LIPEX extruder (Northern Lipids Inc., Burnaby, Canada) equipped with polycarbonate membrane filters with a 200 nm pore size. The vesicle samples were passed 10 times through the membranes. The vesicles are assumed to have an average size of 200 nm. (28) Johnsson, M.; Lam, Y.; Barauskas, J.; Tiberg, F. Langmuir 2005, 21, 5159–5165.
GMO-based CPNPs + DPPC MLVs PtOH-based CPNPs + DOPC ULVs PtOH-based CPNPs + DOPC MLVs HPNPs + DOPC ULVs HPNPs + DOPC MLVs
molar ratio
weight ratio
number ratio
0.01 0.016 0.023 0.18 0.72 0.18 0.017 0.17 0.008 0.027 0.27
0.022 0.035 0.05 0.4 1.6 0.37 0.04 0.4 0.012 0.04 0.4
0.18 0.28 0.4 0.20 0.81 0.24 0.19 0.12 0.02 0.05 0.04
a The number ratio was calculated on the basis of the size obtained by light scattering. For the MLVs, the hydration was assumed to be 40%.
Small-Angle X-ray Diffraction. Instrument. Synchrotron SAXD measurements were performed at beamline I711 at MAX-lab (Lund University, Sweden), using a Marresearch 165 mm CCD detector mounted on a Marresearch desktop beamline base plate.29,30 Appropriate amounts of vesicle dispersions were added to nanoparticle dispersions. The samples were immediately mixed, placed into 1 mm (i.d.) glass capillaries, sealed, and put into the instrument. A period of about 5 min was required between sample mixing and the actual start of the measurement to mix the dispersions, mount the sample, and evacuate the sample chamber. Diffractograms were then recorded under high vacuum at room temperature (22 °C) using a wavelength, λ, of 1.235 Å at a sample-to-detector distance of 1445 mm. The change in liquid-crystalline structure as a result of the interaction between the two types of particles in the dispersion was then continuously measured as a function of time, where each scattering pattern was recorded for 90 s as 2D images with the CCD camera. Data Analysis. The diffractograms (intensity versus scattering vector Q ) 4π sin(θ/2)/λ, where θ is the scattering angle) were obtained by integration using the Fit2D software provided by Dr. A. Hammersley31 and calibrated values for wavelength and detector positions. The unit cell dimensions of the lamellar-phase and cubicphase structures were calculated using the following equation
a)
2π 2 √h + k2 + l2 Q
(1)
where (h, k, l) are the Miller indices of the different liquid-crystalline planes of the structures. Experiments Performed. Table 1 gives the different molar ratios (n/n) used in the experiments, with the corresponding weight ratios (w/w) and number ratios (number of DOPC vesicles to number of LCNPs). Relatively low ratios were chosen to have some consistency with the immobilized bilayer experiments and to be able to follow the kinetics of the structural changes. Ellipsometry. Instrument. An automated Rudolph Research thinfilm null ellipsometer, type 43603-200E, was used to measure the adsorbed amount and thickness of adsorbed layers in situ using the methodology described by Tiberg and Landgren.32 The light source for these measurements was a xenon arc lamp fitted with a filter for a wavelength of 401.5 nm, and the angle of incidence was 68.23°. The ellipsometer was fitted with a 5 mL cuvette, which was thermostated to (25.0 ( 0.1)°C and agitated with a magnetic stirrer at about 300 rpm. (29) Cerenius, Y.; Stahl, K.; Svensson, L. A.; Ursby, T.; Oskarsson, A.; Albertsson, J.; Liljas, A. Journal of Synchrotron Radiation 2000, 7, 203–208. (30) Knaapila, M.; Svensson, C.; Barauskas, J.; Zackrisson, M.; Nielsen, S. S.;Nørgaard Toft, K.; Vestergaard, B.; Arleth, L.; Olsson, U.; Skov Pedersen, J.; Cerenius, Y. Submitted to J. Synchrotron Rad 2008. (31) Hammersley, A.; FIT2D, V12.077; 19872005. (32) Tiberg, F.; Landgren, M. Langmuir 1993, 9, 927–932.
4002 Langmuir, Vol. 25, No. 7, 2009
Vandoolaeghe et al.
Preparation of the Substrates. The silica surfaces were provided by Bo Thuner (Department of Chemistry, IFM, Linko¨ping University, Sweden). Silicon wafers (p-type, boron-doped, resistivity 1-20 Ω cm) were thermally oxidized in an oxygen atmosphere at 920 °C for about 1 h, followed by annealing and cooling in an argon flow. This procedure yields a SiO2 layer thickness of ∼300 Å. The oxidized wafers were cut into slides with a width of 12 mm and were cleaned in a mixture of 25% NH4OH (pro analysi, Merck), 30% H2O2 (pro analysi, Merck), and H2O (1:1:5 by volume) at 80 °C for 5 min, followed by a mixture of 32% HCl (pro analysi, Merck), 30% H2O2 (pro analysi, Merck), and H2O (1:1:5 by volume) at 80 °C for 10 min.32 The cleaned oxidized wafers were rinsed with water and ethanol and then stored in ethanol. The surfaces were dried under vacuum (0.001 mbar) and then treated in an air plasma cleaner (Harrick Scientific Corp., model PDC-3XG) for 5 min prior to the start of the experiment. Characterization of the Substrates. The optical properties of the silica substrates (i.e., the complex refractive index of silicon (nˆSi) and the real refractive index (nSiO2) and thickness (dSiO2) of the oxide layer) were characterized at the beginning of each experiment. The ellipsometric angles, Ψ and ∆, were measured in two different ambient media, air and solution. Numerical calculations were performed on the four measured parameters using a three-layer stratified optical model.33 A typical substrate characterization was nˆSi ) 5.5 - 0.3i, nSiO2) 1.5, and dSiO2 ) 300 Å. The recorded values of Ψ and ∆ correspond to the relative amplitude change and phase shift upon reflection of polarized light at an interface, respectively. The ellipsometric angles are related to the complex reflectivity coefficients for light components polarized parallel and perpendicular to the plane of incidence, rˆpand rˆs, respectively:
rˆp ) tan Ψ ei∆ rˆs
(2)
Formation of the Supported Bilayer. The supported lipid bilayer was formed on silica by adsorption and deposition from an aqueous solution, where DOPC or DPPC is solubilized by sugar surfactant DDM in a ratio of 1:6.34,35 Bilayer formation was achieved by a series of subsequent additions of mixed PC-DDM solutions of decreasing total concentration, starting with 0.114 mg mL-1. Each addition of the PC-DDM mixture was followed by a rinse with 0.1 mM HCl, which ensured the removal of excess DDM. This method is based on gradually approaching the two-phase region of the PCDDM aqueous system and thereby causing the deposition of a PC bilayer. Three cycles of addition followed by rinsing have been shown to give complete bilayer coverage consisting only of phospholipids.25,36 The concentrations of the mixture in the second and third additions were 10 and 1%, respectively, of the starting value. All additions were made by adding a small aliquot of a concentrated solution of the mixture to the cuvette, and rinsing steps were performed by means of a peristaltic pump (Ole Dich Instrumentmakers ApS, Hvidovre, Denmark) for at least 15 min at a rate of 1.5 mL min-1. Addition of the LCNP. The LCNP samples were added to the cuvette of the ellipsometer as a small aliquot of a concentrated solution, typically less than 150 µL diluted in 5 mL of HCl solution at pH 4. Water purified by a Milli-Q system was used in all measurements. Data EValuation. The ellipsometry program,37 based on McCrackin‘s approach,38,39 was used to evaluate the data. The recorded (33) Landgren, M.; Jonsson, B. J. Phys. Chem. 1993, 97, 1656–1660. (34) Grant, L. M.; Tiberg, F. Biophys. J. 2002, 82, 1373–1385. (35) Vacklin, H. P.; Tiberg, F.; Thomas, R. K. Biochimica Et Biophysica Acta-Biomembranes 2005, 1668, 17–24. (36) Vacklin, H. P.; Tiberg, F.; Fragneto, G.; Thomas, R. K. Langmuir 2005, 21, 2827–2837. (37) Petrov, P.; Ellipsometry, 1.3.1 ed.; 1994-2001. (38) McCrackin, F. L.; Colson, J. P. Natl. Bur. Stand (US) Misc. Publ. 1964, 256, 61–82. (39) McCrackin, F. L.; Passaglia, E.; Stromberg, R. R.; Steinberg, H. L. J. Res. Natl. Bur. Stand. 1963, 67A, 363–377.
ellipsometric angles were evaluated using a four-layer or five-layer optical model (based on the previously determined substrate optical properties) assuming isotropic media and planar interfaces, as described in detail by Landgren et al.33 The procedure involves calculating the mean refractive index, nf, and the thickness, df, of the adsorbed film numerically from the appropriate expression of the reflectivity coefficients in eq 2.40 The adsorbed amount, Γ, was calculated from nf and df using de Feijter’s formula41
Γ)
(nf - n0)df dn/dc
(3)
where n0 is the refractive index of the bulk solution and dn/dc is the refractive index increment of the adsorbed material as a function of its bulk concentration. In all of the calculations of the adsorbed amount, the value dn/dc ) 0.148 cm3 g-1 was used for both DOPC and DPPC.34
Results and Discussion Synchrotron SAXD. Time-resolved SAXD measurements were performed to study structural changes of the lipid-based nanoparticle dispersions as a function of time after mixing with phospholipid-based vesicles. Measurements were performed to study interactions between the different nonlamellar particle structures, CPNPs and HPNPs, and lamellar vesicle structures, ULVs and MLVs, mixed at different ratios, as described in Table 1. GMO-Based CPNPs. Interaction with DOPC-Based Unilamellar Vesicles. The changes in SAXD intensity versus scattering vector, Q, for GMO-based CPNPs mixed with ULVs composed of DOPC for molar DOPC/GMO ratios of 0.010 and 0.016 are shown in Figure 2a,b, respectively, with the first diffractogram recorded 5 min after mixing. The initially recorded diffractogram of the GMO-based CPNP system can be indexed in accordance with an Im3m cubic structure of unit cell size a ) 13.8 nm, as also observed earlier.6,7 The small, broad diffraction peak appearing at Q ) 1.0 nm-1 (d spacing of about 6.3 nm) indicates the presence of a small fraction of MLV contaminants present in the DOPC-based ULV dispersion. This Q value corresponds to the repeat distance of a fully swollen lamellar phase42 and is consistent with the value observed for the MLVs featured in Figure 4. After mixing, the characteristic Im3m cubic structure peaks shift with time toward higher Q. This observation shows that interactions between CPNPs and ULVs result in lipid exchange, which results in a successive decrease in the unit cell dimension of the cubic structure with time after mixing. Moreover, the decrease in intensity of the characteristic cubic-phase peaks shows that the number of CPNPs and/or the volume fraction of the cubic phase within the particles decreases over time. For the highest DOPC/GMO ratio of 0.023, the cubic-phase signal had already disappeared by the first diffraction measurement (i.e., 5 min after mixing, result not shown). For the lower ratios, a new diffraction peak appears around Q ) 1.2 nm-1 after a few repeat measurements. This observation indicates that a structural transformation of GMObased CPNPs occurs during mixing with the DOPC-based ULVs. For a DOPC/GMO ratio of 0.01, the peak (Q ) 1.18 nm-1) appears about 14 min after mixing. With time, the peak is shifted to larger Q (Q ) 1.236 nm-1 at t ) 45.5 min), but the intensity remains approximately constant. For the higher DOPC/GMO ratio of 0.016, the peak appears after about 8 min, as can be (40) Jenkins, T. E. Journal of Physics D-Applied Physics 1999, 32, R45-R56. (41) Defeijter, J. A.; Benjamins, J.; Veer, F. A. Biopolymers 1978, 17, 1759– 1772. (42) Caracciolo, G.; Sadun, C.; Caminiti, R.; Pisani, M.; Bruni, P.; Francescangeli, O. Chem. Phys. Lett. 2004, 397, 138–143.
Liquid-Crystalline Nanoparticle Interactions
Langmuir, Vol. 25, No. 7, 2009 4003
Figure 2. Diffractograms obtained as a function of time when mixing GMO-based CPNPs with DOPC ULVs at DOPC/GMO molar ratios of (a) 0.01 and (b) 0.016. The thick lines correspond to the diffractograms for pure DOPC vesicles (dashed line) and pure GMO-based CPNPs (solid line), and the thin lines correspond to the diffractograms recorded as a function of time after mixing the CPNPs and ULVs. The interval between the measurements of each was 90 s, with the first diffractogram recorded 5 min after mixing. An offset in intensity (y axis) of the baseline (0 intensity) was introduced between each of the diffractograms for clarity.
Figure 3. Tentative equilibrium phase diagram of the D2O-GMO-DOPC system adapted from ref 43, with the rectangle corresponding to the area of interest in this study. Reproduced with kind permission from Springer Science+Business Media: Gutman, H.; Arvidson, G.; Fontell, K.; Lindblom, G. 31P and 2H NMR studies of phase equilibria in the three component system monoolein-dioleoylphosphatidylcholine-water. Surfactants in Solution; Plenum Press: New York, 1984; Vol. 1, p 149, Figure 3.
expected from the increased ULV content. Again, the position of this peak shifts to higher Q values with increased mixing time (Q ) 1.284 nm-1 at t ) 23 min), whereas the intensity of the peak increases somewhat until t ) 14 min (Q ) 1.19 nm-1), after which it decreases. Because only a single diffraction peak appears, it is not possible to identify the new structure unambiguously from only SAXD data. However, bearing in mind that DOPC should drive the spontaneous curvature of the system toward zero, it seems likely that the emergence of the new peak represents
a transition from cubic to lamellar structure within the particles. This notion is supported by the tentative equilibrium phase diagram of the ternary GMO-DOPC-D2O system reported by Gutman et al.43 and reproduced in Figure 3. The rectangle shows the phase diagram area of interest in the present study. It is clear that only cubic and lamellar structures are formed when mixing DOPC and GMO in excess water. At intermediate DOPC/GMO ratios, a three-phase region with cubic and lamellar phases with excess water appears, and at a higher DOPC/GMO ratio, a lamellar phase in excess water (i.e., vesicles) appears. As compared with the tentative equilibrium phase behavior by Gutman, our results indicate a phase transition at rather low DOPC content, which is a bit surprising. Note that there are clear differences between our study and Gutman’s (i.e., nonequilibrium vs equilibrium conditions and the presence of stabilizing polymer Pluronic F-127), and the attribution of the appearing diffraction peak to a lamellar phase is expected but not proven because we observe only one peak. If our assumption is correct that the emergence of the new peak represents a transition to the lamellar phase, then there is clearly a difference in the phase behavior of the GMO dispersed into CPNPs and that of the corresponding bulk system. This may be partially due to the presence of polymeric stabilizing agent F127. For the ternary GMO/F127/ water system at a water content of 70%, a phase transition from the cubic to lamellar phase occurs at a GMO/F127 ratio of about 0.3 (w/w) (i.e., a substantially higher ratio than the 0.1 w/w ratio used in the study).44 In the dispersed colloidal state, it appears that the GMO-based CPNPs are close to a structural phase transition, which in our case is triggered by additions of small (43) Gutman, H. A., G.; Fontell, K. and Lindblom, G. In Surfactants in Solution, Vol. 1, Plenum Press: New York, 1984; 143-152. (44) Landh, T. J. Phys. Chem. 1994, 98, 8453–8467.
4004 Langmuir, Vol. 25, No. 7, 2009
Vandoolaeghe et al.
Figure 4. Diffractograms obtained as a function of time when mixing GMO-based CPNPs with (a) DOPC MLVs at a DOPC/GMO molar ratio of 0.18 and (b) DPPC MLVs at a DPPC/GMO molar ratio of 0.18. The thick lines correspond to the diffractograms for pure DOPC or DPPC vesicles (dashed line) and pure GMO-based CPNPs (solid line), and the thin lines correspond to the diffractograms recorded as a function of time after mixing the CPNPs and MLVs. The interval between measurements was 90 s, with the first diffractogram recorded 5 min after mixing. An offset in intensity (y axis) of the baseline (0 intensity) was introduced between each of the diffractograms for clarity.
Figure 5. (a) Evolution of the unit cell dimension, a, with time (a) of the cubic structure of the GMO-based CPNPs and (b) of the lamellar structure of the MLVs (open symbols) and of the additional assumed lamellar structure (filled symbols) when GMO-based CPNPs were mixed with DOPC MLVs (circles) and DPPC MLVs (diamonds) in a PC/GMO molar ratio of 0.18 and with DOPC ULVs at a DOPC/GMO molar ratio of 0.01 (inverse triangles) and 0.016 (triangles). As a reference, the expected values for a lamellar phase of GMO and DOPC are indicated by arrows in Figure 5b. The lines are a guide to the eyes only.
amounts of DOPC ULVs. Assuming that the new peak comes from a lamellar phase structure, the expected second peak is not visible within the Q range studied. From the position of the primary peak, a lamellar spacing of about 5.2 nm is obtained, which can be compared with the spacing of the lamellar phase in a binary GMO water system of about 4.0 nm45 and with the 6.3 nm spacing observed for the fully swollen DOPC lamellar phase (cf Figure 5b). The lamellar phase in the pure GMO system (45) Borne, J.; Nylander, T.; Khan, A. Langmuir 2000, 16, 10044–10054. (46) Barauskas, J.; Johnsson, M.; Johnson, F.; Tiberg, F. Langmuir 2005, 21, 2569–2577. (47) Rand, R. P.; Parsegian, V. A. Biochimica Et Biophysica Acta 1989, 988, 351–376. (48) Barauskas, J.; Landh, T. Langmuir 2003, 19, 9562–9565.
has a very low degree of hydration (water layer of about 0.4 nm), and it is apparent that in the mixture of DOPC and GMO the hydration is intermediate with respect to that of the pure structures. Note that preliminary cryogenic transmission electron microscopy experiments have been performed on the mixture of GMO-based CPNPs with DOPC vesicles in a DOPC/GMO molar ratio of 0.24. The images obtained are shown in Figure S1 in the Supporting Information. The results indicate that the interactions of GMO-based CPNPs with DOPC vesicles lead to the formation of larger aggregates with a change in internal structure upon mixing and possibly the disappearance of any order in the mixed solution at long equilibrium times, which is consistent with the SAXD data.
Liquid-Crystalline Nanoparticle Interactions
Langmuir, Vol. 25, No. 7, 2009 4005
Figure 6. Diffractograms obtained as a function of time when mixing PtOH-based CPNPs with (a) DOPC MLVs at a DOPC/PtOH molar ratio of 0.17 and (b) DOPC ULVs at a DOPC/PtOH molar ratio of 0.017. The thick lines correspond to the diffractograms for pure DOPC vesicles (dashed line) and pure PtOH-based CPNP (solid line), and the thin lines correspond to the diffractogram recorded as a function of time after mixing. The interval between measurements was 90 s, with the first diffractogram recorded 5 min after mixing in the case of the MLVs and 8 min after mixing in the case of the ULVs. An offset in the intensity (y axis) of the baseline (0 intensity) was introduced between each of the diffractograms for clarity.
Interaction with DOPC- and DPPC-Based Multilamellar Vesicles. The effects of the lamellar dispersion on the nanostructural changes of nonlamellar nanoparticles were studied by following the interaction of GMO-based CPNPs with DOPC MLVs at a DOPC/GMO ratio of 0.18 (Figure 4a). The initial SAXD profile for the DOPC MLVs gives a unit cell dimension of a ) 6.3 nm, which is consistent with previous studies.6,42,46 When the GMO-based CPNPs are mixed with DOPC MLVs, the cubic-phase peaks again shift to higher Q (decreased a) whereas the lamellar peaks of the MLVs are shifted toward lower Q (increased a) with time. Both shifts clearly show that there is an interaction between the two types of liquidcrystalline nanoparticles. The decrease in intensity of the characteristic peaks corresponding to both types of structures indicates that the number of structured particles decreases and/ or the internal structured phase domains decrease in size. For instance, when the GMO-based CPNPs interact with the DOPC vesicles one could imagine that the outer bilayer of the MLVs is successively “peeled off”, leading to a decrease in the number of bilayers and hence a decrease in the corresponding diffraction peak intensity. Similarly the transfer of DOPC to the CPNPs would move the cubic phase closer to a lamellar phase transition, where this effect will be most pronounced at the particle surface region. As in the case of the ULV interaction, the emergence of a weak diffraction peak (Q ≈ 1.2 nm-1) corresponding to a structural transition to a lamellar phase occurs. However, the structure that formed appears to be less defined than in the case of the ULVs. The reason for this could be that material transfer in the MLV case is less homogeneous due to slower mass-transfer. From an application point of view, it is also important to understand and monitor the effect of the lamellar phase properties
on particle interactions. Here we compared the interaction of GMO-based CPNPs with bilayers formed from two phospholipids with different chain melting temperatures. DPPC (16:0 PC) is in the gel phase at the temperature used for the experiments (room temperature around 25 °C). Conversely, DOPC (18:1 PC cis) is in the fluid state at the same temperature. The change in diffraction pattern with time when GMO-based CPNPs were mixed with DPPC MLVs at a DPPC/GMO molar ratio of 0.18 is shown in Figure 4b. As expected, in this case, the kinetics of change in unit cell dimensions for both the cubic and lamellar structures are slower than in the case of mixing with DOPC MLVs. The gel state of the DPPC MLVs renders the lipid exchange between the two liquid-crystalline structures slower. Furthermore, the area per molecule is smaller for the DPPC layer (about 52 Å2) in the gel phase than for DOPC in the lamellar phase (about 70 Å2).47 This means that more of the acyl chain region is exposed in the case of the DOPC lamellar phase, which facilitates lipid exchange compared to DPPC. Therefore, for the DPPC bilayer we observe that the cubic structure is still present 47 min after mixing (time when the experiment was stopped). GMO and DPPC are expected to be less miscible than GMO and DOPC because of the similarity in acyl chain length between GMO and DOPC. No lamellar structure peak appears in this case at intermediate times. However, the kinetics of exchange are slower in this case, and we cannot rule out that a lamellar structure would form at longer times. Figure 5 shows the time evolution of the unit cell observed for the cubic structure (Figure 5a) as well as for the assumed lamellar phase structure appearing at intermediate times when mixing with DOPC vesicles and for the DOPC and DPPC lamellar structures (Figure 5b) for the mixing of GMO-based CPNPs with DOPC-based ULVs and DOPC- and DPPC-based MLVs.
4006 Langmuir, Vol. 25, No. 7, 2009
Vandoolaeghe et al.
Figure 7. Diffractograms obtained as a function of time after mixing DGMO/GDO-based HPNPs with (a) DOPC MLVs in a DOPC/(DGMO/GDO) molar ratio of 0.27 and (b) DOPC ULVs in a molar ratio of 0.008. The thick lines correspond to the diffractograms for pure DOPC vesicles (dashed line) and pure DGMO/GDO-based HPNPs (solid line), and the thin lines correspond to the diffractograms recorded for the mixture. The interval between measurements was 90 s, with the first diffractogram recorded 5 min after mixing. An offset in the intensity (y axis) of the baseline (0 intensity) was introduced between each of the diffractograms for clarity.
As can be seen in Figure 5a, the decrease in the cubic-phase unit cell dimension is significantly slower for the MLVs than for the ULVs, even at 10 times greater DOPC content. However, here we note that the number ratio (vesicles/CPNPs) is similar for the experiments with ULVs or MLVs. Hence, it appears that the structural changes are the consequence of particle-particle interactions rather than monomer transport. Aside from more rapid diffusion, the higher bilayer stress in the smaller ULVs may contribute to lower stability, which in turn could facilitate interactions with the CPNPs. We also note that exchange between CPNPs and MLVs where the chains are in the gel state is significantly slower than where the chains are in the liquidcrystalline state. PtOH-Based CPNPs. The phase behavior of PtOH in aqueous solution is quite similar to that of GMO, and the sequence of phases that occurs with increasing water content is exactly the same.48 Figure 6 shows the change in the diffraction pattern with time when PtOH-based CPNPs were mixed with DOPC MLVs at a DOPC/PtOH molar ratio of 0.17 (Figure 6a) and with DOPC ULV (Figure 6b) at a molar ratio of 0.017. Figure 6a,b shows only a small shift in peak Q values of the cubic structure when PtOH-based CPNPs interact with DOPC. In contrast to the GMO-based CPNPs, the shift is toward lower Q, corresponding to an increase in the unit cell dimensions of the cubic phase. Both the DOPC lamellar phase and the GMO cubic phase swell to larger water content than does the PtOH cubic phase. Hence, the inclusion of DOPC in the PtOH cubic phase is likely to lead to swelling in water. As observed for the interaction of GMO-based CPNPs, the characteristic lamellar peaks corresponding to the MLV structure shift to lower Q with time. Importantly, the interaction between PtOH-based CPNPs does not show the emergence of the new assumed lamellar phase that is seen for the GMO-based CPNPs. Unfortunately, no phase diagram of the PtOH-DOPC-water system is available. Here we
note again that contrary to GMO the hydrocarbon chain of PtOH is considerably shorter than acyl chains of DOPC. Moreover, PtOH is branched and does not contain a double bond. For instance, the bilayer thickness of the PtOH lamellar phase is about 2.8 nm,48 compared to about 3.6 nm for GMO.45 PtOH and DOPC are therefore expected to be less miscible than GMO and DOPC, which may explain this difference. The use of different stabilizers complicates the picture, but this issue needs to be addressed in a separate investigation. DGMO/GDO-Based HPNPs. It is also interesting to determine how the internal structure of the LCNPs affects the interaction with DOPC vesicles. We therefore investigated the DGMO/ GDO-based HPNPs’ interaction with DOPC MLVs and ULVs. Figure 7 shows the diffraction profiles recorded as a function of time after mixing HPNPs with DOPC MLVs in a DOPC/(DGMO/ GDO) molar ratio of 0.27 (Figure 7a) and with DOPC ULVs in a ratio of 0.008 (Figure 7b). The SAXD pattern obtained for the HPNPs is consistent with previous studies of this system (a ) 4.9 nm).6,28 Upon mixing HPNPs with MLVs, the intensity from the hexagonal phase is lost faster than for the cubic structure in the CPNPs, but only a minor shift in Q of the hexagonal peaks is observed with time. The lamellar structure of the MLVs, however, is clearly more stable than in the presence of GMO- or PtOH-based CPNPs, as judged from the slower decay in the intensity of the diffraction peaks and the small shift in spacing with time. Similar findings were obtained for the interactions with ULVs (DOPC/(DGMO/ GDO) ratios of 0.008 and 0.027), where the (10) diffraction peak from the hexagonal phase is still present at the end of the experiments (t ) 22 min) with only minor shifts in peak positions, indicating that the HPNP structure is less sensitive to mixing with the lamellar-forming DOPC lipids.
Liquid-Crystalline Nanoparticle Interactions
Figure 8. Adsorbed amount Γ (filled symbols) and layer thickness d (open symbols) as a function of time determined by ellipsometry after the addition of 0.05 mg mL-1 GMO-based CPNPs (circles), PtOHbased CPNPs (triangles), and DGMO/GDO-based HPNPs (inverse triangles) to a solution at pH 4 in which a supported DOPC bilayer is submerged. The time t ) 0 corresponds to the time of addition of the LCNPs.
Ellipsometry Effect of the Composition and Internal Structure of the Nonlamellar LCNPs. To shed further light on the interactions and structural changes observed on mixing nonlamellar LCNPs with bilayer structures, the surface interactions (adsorption) of CPNP and HPNP systems at a supported DOPC bilayer were studied. The results for the interaction between GMO-based CPNPs and DOPC bilayers have been discussed in detail elsewhere.26 Figure 8 presents the data obtained for the interaction of GMO- and PtOH-based CPNPs as well as DGMO/GDObased HPNPs with DOPC bilayers (the GMO data are adapted from ref 26). In all cases, an initial adsorption is observed, followed by a desorption process, suggesting particle attachment and interfacial lipid exchange leading to the subsequent desorption of LCNP components or attached particles as the surface interaction changes from favorable to unfavorable.26 Although the general process is similar for all particles, the kinetics and interfacial properties recorded at the end of the experiment are quite different. For GMO-based CPNPs, adsorption and subsequent desorption processes are relatively rapid. However, the resulting final layer structure is thicker than the original supported DOPC bilayer, most likely indicating that a small fraction of structured particles (cubic or/and lamellar phase)27 remains attached at the surface.26 In the case of the PtOH-based CPNPs, surface adsorption and desorption is faster than for the corresponding GMO-based CPNPs. The time at which the desorption starts to dominate (i.e., at the adsorption maximum) occurs earlier and at a lower adsorbed maximum amount than for the other LCNPs. Notably, the final thickness is larger but the adsorption value is smaller, again suggesting that a small number of particles exist at the interface and influence the mean optical thickness measured in the experiment. The lower adsorbed amount compared to the initial bilayer suggests that DOPC is removed from the surface during the desorption process. For the HPNP system, adsorption, interfacial exchange, and desorption processes are much slower than observed for the two types of CPNPs. The maximum value of the adsorbed amount is similar to that obtained in the case of the GMO-based CPNPs. Also, the desorption process is much slower, and the adsorbed amount and thickness at steady state are similar to the values recorded for the initial bilayer.
Langmuir, Vol. 25, No. 7, 2009 4007
Figure 9. Adsorbed amount Γ (filled symbols) and layer thickness d (open symbols) as a function of time determined by ellipsometry after the addition of GMO-based CPNPs at a concentration of 0.05 mg mL-1 on a DOPC bilayer (circles) and on a DPPC bilayer (diamonds) at pH 4. The CPNPs were added at time t ) 0.
Note that the kinetics of changes observed with ellipsometry is different than that observed with SAXD measurements. The reason is that the total lipid concentration used is almost 2 orders of magnitude higher in the SAXD experiments and than in ellipsometry experiments. The nature of the phospholipid bilayer is also different, one curved and the other flat and supported by a solid surface, which can lead to different kinetics. Effect of the State of the Phospholipids Constituting the Bilayer. In a similar way, as in the SAXD experiments, the influence of the state of the phospholipid used to form the bilayer (i.e., gel state or liquid-crystalline state) on the interaction with GMO-based CPNPs has been studied by ellipsometry. Figure 9 shows the ellipsometry data obtained in the case of the interaction of 0.05 mg mL-1 GMO-based CPNPs with supported DOPC and DPPC bilayers. We observe the same phenomena in both cases, which is fast adsorption until a maximum is reached, followed by a decrease in the adsorbed amount. However, in the case of the interaction with the DPPC bilayer, the maximum value is larger, and the initial adsorption is faster. The desorption is much slower in this case. The desorption is attributed to a change in the layer properties, which makes the interaction unfavorable after mixing has taken place between the two systems. The slower kinetics can thus be rationalized by the fact that the DPPC bilayer is in the gel phase. In the initial stage, GMO-based CPNPs adsorb at the bilayer interface, and mixing starts to occur between the phospholipid molecules and the CPNP components. Because DPPC is in the gel phase, the exchange is slower than in the case of the DOPC bilayer, which explains that more molecules have time to adsorb before the nature of the bilayer is changed and thus before the desorption process occurs. At steady state, the values of adsorbed amounts and thicknesses are very similar in both cases, which suggests that only the kinetics of the interaction differs for the different systems. These results are in good agreement with the results obtained in bulk with SAXD, where slower kinetics is observed in the case of the phospholipid in the gel phase at the investigated temperature. In conclusion, the slower release observed with DPPC bilayers, as correlated with slower changes in the internal structures as shown in the SAXD experiments compared with the DOPC bilayers, verifies our hypothesis that the release of LCNP components from the supported bilayer is triggered by the exchange of lipids.
4008 Langmuir, Vol. 25, No. 7, 2009
Conclusions The interactions of multicomponent liquid-crystalline nanoparticles with lipid bilayers constitute a very complex process involving particle-particle interactions (e.g. fusion) as well as lipid mixing. From a practical standpoint., it is important to improve the understanding of how chemical and structural properties of structured nanoparticles affect interactions and functions in biological systems. This could have important implications in the design of drug-delivery vehicles of active pharmaceutical substances. Here we chose to study three different liquid-crystalline nanoparticle (LCNP) systems with regard to their interactions with three different bilayer structures as simple models of biological membrane structures (i.e., unilamellar vesicles (ULVs), multilamellar vesicles (MLVs), and surfacesupported bilayers). Despite being sterically stabilized and having long-term stability with respect to storage, all LCNP systems interact readily with the phospholipid bilayers. This is evidenced by the rapid decay with time of the intensity of the peak related to the initial internal phase structure of the particles, with the faster signal decay observed for interactions with ULVs compared to MLVs. GMO-based CPNPs show the most pronounced structural changes on mixing with vesicles, which is attributed to the internal structure being less stable and closer to the phase transition. For CPNPs interacting with vesicles and a supported bilayer, more extensive lipid mixing is observed for GMO than for PtOH, which may contribute to the observed phase transition. In terms of effects on the hexagonal phase structure of the HPNPs, negligible changes occur in the unit cell dimension, as can be expected for this phase, and no transition to another structured
Vandoolaeghe et al.
phase is observed upon mixing with lamellar particles. The HPNPs also showed the slowest adsorption and mixing with the supported bilayer surface, showing that the results from the different techniques used are correlated. We also found interesting differences in the interaction of CPNPs depending on the bilayer properties, namely, whether the lipid is in the gel state or in the liquid-crystalline phase. The exchange of lipids is significantly slower if the bilayer is in the gel state, which at a given time leads to less extensive changes within the CPNP internal structure. In conclusion, we have shown that it is possible to study interactions in complex colloidal systems at the mesoscopic level and that structure-function relationships can be established. Future plans include extending the present investigations to the study of more biologically relevant phenomena such as hemolysis and cell fusion. Acknowledgment. The Swedish Foundation for Strategic Research, Vinnova, and the Camurus Lipid Research Foundation financially supported this project, as well as the Swedish Research council and its Linnaeus grant “Organizing Molecular Matter”, OMM. We are grateful to MAX-lab, Lund, Sweden, for allocations of beam time for the synchrotron SAXD experiments. We are grateful to Yngve Cerenius and Matti Knaapila for help with the SAXD experiments. Supporting Information Available: A description of cryogenic transmission electron microscopy experiments as well as some images obtained for the mixture of DOPC vesicles with GMO-based CPNPs. This material is available free of charge via the Internet at http://pubs.acs.org. LA802768Q
