Effects of Perfluorocarbon Gases on the Size and Stability

Article pubs.acs.org/Langmuir

Effects of Perfluorocarbon Gases on the Size and Stability Characteristics of Phospholipid-Coated Microbubbles: Osmotic Effect versus Interfacial Film Stabilization Csongor Szíjjártó, Simona Rossi, Gilles Waton, and Marie Pierre Krafft* Systèmes Organisés Fluorés à Finalités Thérapeutiques (SOFFT), Institut Charles Sadron (CNRS, UPR 22), Université de Strasbourg, 23 rue du Loess, 67034 Strasbourg Cedex 2, France S Supporting Information *

ABSTRACT: Micrometer-sized bubbles coated with phospholipids are used as contrast agents for ultrasound imaging and have potential for oxygen, drug, and gene delivery and as therapeutic devices. An internal perfluorocarbon (FC) gas is generally used to stabilize them osmotically. We report here on the effects of three relatively heavy FCs, perfluorohexane (F-hexane), perfluorodiglyme (F-diglyme ), and perfluorotriglyme (F-triglyme), on the size and stability characteristics of microbubbles coated with a soft shell of dimyristoylphosphatidylcholine (DMPC) and on the surface tension and compressibility of DMPC monolayers. Monomodal populations of small bubbles (∼1.3 ± 0.2 μm in radius, polydispersivity index ∼8%) were prepared by sonication, followed by centrifugal fractionation. The mean microbubble size, size distribution, and stability were determined by acoustical attenuation measurements, static light scattering, and optical microscopy. The half-lives of F-hexane- and F-diglyme-stabilized bubbles (149 ± 8 and 134 ± 3 min, respectively) were about 2 times longer than with the heavier F-triglyme (76 ± 7 min) and 4−5 times longer than with air (34 ± 3 min). Remarkably, the bubbles are smaller than the minimal size values calculated assuming that the bubbles are stabilized osmotically by the insoluble FC gases. Particularly striking is that bubbles 2 orders of magnitude smaller than the calculated collapse radius can be prepared with F-triglyme, while its very low vapor pressure prohibits any osmotic effect. The interface between an aqueous DMPC dispersion and air, or air (or N2) saturated with the FCs, was investigated by tensiometry and by Langmuir monolayer compressions. Remarkably, after 3 h, the tensions at the interface between an aqueous DMPC dispersion (0.5 mmol L−1) and air were lowered from ∼50 ± 1 to ∼37 ± 1 mN m−1 when F-hexane and F-diglyme were present and to ∼40 ± 1 mN m−1 for Ftriglyme. Also noteworthy, the adsorption kinetics of DMPC at the interface, as obtained by dynamic tensiometry, were accelerated up to 3-fold when the FC gases were present. The compression isotherms show that all these FC gases significantly increase the surface pressure (from ∼0 to ∼10 mN m−1) at large molecular areas (70 Å2), implying their incorporation into the DMPC monolayer. All three FC gases increase the monolayer’s collapse pressures significantly (∼61 ± 2 mN m−1) as compared to air (∼54 ± 2 mN m−1), providing for interfacial tensions as low as ∼11 mN m−1 (vs ∼18 mN m−1 in their absence). The FC gases increase the compressibility of the DMPC monolayer by 20−50%. These results establish that, besides their osmotic effect, FC gases contribute to bubble stabilization by decreasing the DMPC interfacial tension, hence reducing the Laplace pressure. This contribution, although significant, still does not suffice to explain the large discrepancy observed between calculated and experimental bubble half-lives. The case of F-triglyme, which has no osmotic effect, indicates that its effects on the DMPC shell (increased collapse pressure, decreased interfacial tension, and increased compressibility) contribute to bubble stabilization. Fhexane and F-diglyme provided both the smallest mean bubble sizes and the longest bubble half-lives.

■

INTRODUCTION Gaseous microbubbles are being used as blood pool contrast agents for ultrasound imaging.1−5 Intravascular delivery of oxygen in the form of stabilized microbubbles may offer a simple means of oxygenating tissues.6,7 Other applications under investigation include targeted microbubbles for molecular imaging,8,9 ultrasound-triggered drug and gene delivery, and use as clot-breaking agents.10,11 Rational microbubble engineering, and size and property control for such applications, requires a better understanding of the basics of © 2011 American Chemical Society

bubble stabilization in relation to both shell and internal phase components. Stabilization of microbubbles by gases that have low Ostwald coefficients (a low partition coefficient between the gas phase and aqueous solution), yet relatively high saturated vapor pressures, is well established, both theoretically and experimentally.12,13 Fluorocompound gases appear to be the only Received: November 8, 2011 Revised: December 16, 2011 Published: December 16, 2011 1182

dx.doi.org/10.1021/la2043944 | Langmuir 2012, 28, 1182−1189

Langmuir

Article

Table 1. Physicochemical Characteristics of the Microbubble-Stabilizing FCs Investigated: Molecular Weight (MW), Boiling Point (bp), Molar Volume (Vm), Diffusion Coefficient (D), Saturated Vapor Pressure (psatd), Water Solubility of Liquids (cwater), Ostwald Coefficient (L = cwater/cgas, with cgas = psatd/RT) name nitrogen perfluorohexane (F-hexane) perfluorooctane (F-octane) perfluoro(diethylene glycol dimethyl ether) (F-diglyme ) perfluoro(triethylene glycol dimethyl ether) (F-triglyme)

MW (g mol−1)

bp (°C)

Vm (cm mol−1)

D (25 °C) −9 2 −1 (10 m s )

N2 CF3(CF2)4CF3

28 338

−196 59a

201a (20 °C)

19a 5.2e

CF3(CF2)6CF3

438

103c

261 (20 °C)

3.7e

CF3(OCF2CF2)2OCF3

386

64

b

245 (25 °C)

5.2

e

CF3(OCF2CF2)3OCF3

502

105c

308d (25 °C)

4.6e

formula

3

d

psatd (25 °C) (105 Pa)

psatd (37 °C) (105 Pa)

0.29a 0.27f 0.05f

0.42b 0.38f 0.08f

0.26

f

0.01f

b

0.33 0.37f

cwater (10−4 mol m−3) 2.7a (25 °C)

L (10−6) 14480a (35 °C) 23a(25 °C)

0.038g (25 °C)

3 (25 °C)

0.43 (37 °C)

4b (37 °C)

b

0.02f

a

From ref 12. bFrom ref 13. cFrom ref 31. dCalculated from the density and molar weight; the densities were obtained by weighing 5 mL of F-glyme in a volumetric flask at 25 °C. eEstimated using the Hayduk−Laudie equation D = (13.26 × 10−5)η−1.14Vm−0.589 at 25 °C, η being the dynamic viscosity. From ref 32. fEstimated form the energies of vaporization of the FCs using a group additivity method.33 gFrom ref 34.

An effect of the FC on the phospholipid’s interfacial film tension, and hence on the Laplace pressure, has so far not been envisaged, likewise for a possible effect of the FC on the compressibility of the phospholipid film. Additionally, the FC could affect the droplet size distribution, which may also influence bubble stability. In the present paper we report (1) on the effects of three FC gases on the size, size distribution, and stability characteristics of dimyristoylphosphatidylcholine (DMPC)-coated microbubbles and (2) on their effects on the surface tension and compressibility of Langmuir monolayers of DMPC. The three FCs investigated were perfluorohexane (C6F14, F-hexane), perfluorodiglyme (CF3(OCF2CF2)2OCF3, F-diglyme ), and perfluorotriglyme (CF3(OCF2CF2)3OCF3, F-triglyme). F-hexane has been found substantially more effective than its lower homologues for microbubble stabilization1,13 and has been used in some of our previous work.26−28 F-diglyme was chosen because its vapor pressure is comparable to that of F-hexane, despite its higher molecular weight (Table 1), and hence is expected to exercise a comparable osmotic effect. F-triglyme, on the other hand, has a much lower vapor pressure and is not expected to exhibit any significant osmotic stabilization at all. Another general objective was to contribute to identifying the causes for the large discrepancy that is observed between calculated and experimental bubble half-lives. DMPC was selected as the bubble wall phospholipid because its liquid/gel phase transition temperature in monolayers is below room and body temperature, thus avoiding interference of a phase transition with data collection, and because it is well documented. We also investigated the compression behavior of Langmuir films of DMPC under an atmosphere of air or of nitrogen saturated with the FC gases. The kinetics of adsorption of DMPC (dispersed as an aqueous suspension of vesicles) at FCsaturated air/water interfaces were determined by measuring dynamic interfacial tensions. The Langmuir films and the rising bubbles used for the tensiometry experiments constitute valid macroscopic models of micrometer-sized bubbles since the curvature of the bubble wall is essentially null at the molecular level. Obtaining meaningful data supposes access to narrowly dispersed populations of microbubbles, as well as to precise and reliable sizing methods. We have recently shown that a polydisperse preparation of DMPC-coated microbubbles

practical compounds that combine these features with biological inertness.7,14 Because of their biological inertness, extremely low solubility in water, and high oxygen solubility, fluorocarbons (FCs) are being investigated for a range of biomedical uses, including intravascular oxygen transport,7 ophthalmology,15,16 drug delivery,17 treatment of lung diseases,18 lung surfactant replacement preparations,19,20 and biomedical research.21 When FCs are used in particles (bubbles, emulsion droplets), the particle size and size distribution proved to have a strong impact on in vivo recognition, intravascular persistence, and biocompatibility.7 The use of FCs as internal gas components allowed the intravascular half-life of micrometer-sized bubbles to increase substantially, thus lifting a major obstacle to the development of microbubbles for intravascular use as contrast agents. Essentially all the injectable soft-shell microbubble contrast agents for diagnosis commercially available or under development are stabilized by fluorocompound gases, including sulfur hexafluoride, perfluoropropane, perfluorobutane, perfluoropentane, and perfluorohexane.1,22 Most of these agents use phospholipids as the main bubble wall component. The FC-driven bubble stabilization mechanism has been assigned to osmotic counteraction of the blood pressure and Laplace pressure, thus retarding bubble dissolution in the blood. Intravascular bubble stability was foretold to increase with decreasing Ostwald coefficient of the gas. Low Ostwald coefficients are usually found for compounds with low vapor pressures and vice versa. A vapor pressure lower limit (∼0.3 × 105 Pa) was deemed necessary, however, below which condensation of the FC occurs under the action of the combined external pressures.13 However, while the predicted general trend was confirmed experimentally both in vitro and in vivo,13 the experimentally measured bubble half-life values were always larger than the predicted ones by several orders of magnitude. A series of studies aiming at determining the role of the shell of phospholipids on microbubble dissolution in the absence of FCs in the internal gas phase have been conducted.23−25 Microbubble deflation was shown to be accompanied by the expulsion of phospholipid from the bubble’s surface in the form of bilayer fragments through complex collapse and shedding mechanisms.23 The existence of a critical radius below which bubbles cannot be stable has been established. 1183


Langmuir

Article

Figure 1. Variation of the attenuation coefficient α as a function of the ultrasound frequency f for DMPC-coated microbubbles separated by centrifugation. The bubbles were stabilized by (a) F-hexane, (b) F-diglyme , and (c) F-triglyme. The solid lines represent the fits obtained by the least-squares method. the bottom of the centrifuged tube, so that the larger bubbles accumulate at the top of the syringe. The lower 7 mL was taken with a second syringe, and the upper 1 mL was discarded. The content of this second syringe was centrifuged at 600 rpm for 1 min. From this, the first 6 mL was taken again with another syringe and the bottom 1 mL was discarded. This third syringe was again centrifuged at 600 rpm for 1 min. The lower 2 mL was injected into the acoustical measuring cell. For the experiments conducted with F-triglyme, the third run was not needed, as the size distribution did not change after two centrifugations. Aliquots (1 or 2 mL to keep the initial values of the attenuation coefficient comparable) were consistently sampled at a depth of 55 mm from the edge of the glass vial. They were injected into the buffer-filled ultrasonic measuring cell (cell volume 140 mL) and thermoregulated at 25 °C. Acoustical Determination of the Microbubble Size Distribution. The method is based on the attenuation of an acoustical pulse that propagates through the aqueous bubble dispersion. Briefly, a method has been developed that fits standard simple-harmonic resonator curves to measured attenuations to infer the size of the bubbles.35−38 A low-power emitter was used that avoids alteration of the bubble characteristics and stability. Experimental details can be found in ref 28. To determine the standard deviation (SD) and polydispersivity index (PI) for the size distributions of bubble dispersions, Gaussian curves were fitted to the size versus volume graphs. The standard deviation is by definition half of the width ω of the curve at half of the mean value (SD = ω/2). The polydispersivity index is the standard deviation divided by the mean diameter 2 r ̅ and multiplied by 100 (PI = SD/2 r ̅ × 100). Each measurement has been repeated three times on three to five different bubble preparations. Optical Microscopy. The samples were observed by optical microscopy (Olympus BH2, Tokyo, Japan). Three to four droplets of bubble dispersion were placed into a concave slide and covered with a glass slide. Rapid image acquisition was achieved using a Lumenera Infinity 2 charge-coupled device (CCD) camera (Lumenera, Ottawa, Canada). Bubble radii were measured using the ImageJ software on 5− 10 slides.39 Quasi-Elastic Light Scattering (QELS). A Malvern Zetasizer Nano ZS was used for dynamic light scattering at a scattering angle of 90°. The temperature was 25 °C. The z-averaged hydrodynamic mean diameters (D̅ ) of the DMPC vesicles used for the tensiometry experiments were determined using the Malvern software. The measurements were achieved on 0.5 mmol L−1 DMPC dispersions. Exploratory experiments with 0.1 and 0.01 mmol L−1 DMPC dispersions required longer equilibration times and resulted in lesser precision. Langmuir Monolayers. The surface tension (π) versus molecular area (A) isotherms were recorded on a Langmuir minitrough (Riegler & Kirstein, Potsdam, Germany) equipped with two movable barriers (initial area 204 × 60 mm, speed 7.3 mm2 s−1, which corresponds to a reduction of the total area of ∼4% min−1). The surface pressure was measured using the Wilhelmy plate (paper) method. The trough was enclosed in a gastight box (volume ∼9 L) and maintained at 25 ± 0.5 °C. This box was flushed for 1 h with N2 saturated with the FC under investigation as described in ref 19. Saturation was obtained by

stabilized by F-hexane and produced by sonication could be fractionated by gravity into narrowly dispersed subpopulations of different sizes.27,28 Centrifugal fractionation of microbubbles coated with distearoylphosphatidylcholine (DSPC) and stabilized by perfluorobutane has also been reported.29 A difficulty encountered in the investigation of soft-shelled bubbles concerns the measurement of their initial size and of their size stability over time. To resolve this difficulty, we have developed a dedicated multifrequency acoustical setup that allows precise size determination of soft-shell microbubbles.28 The major assets of this technique are a low-power acoustical signal that minimizes alteration of bubble characteristics by the sound wave and a precise determination of sound absorption coefficients.

■

MATERIALS AND METHODS

Materials. 1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC) was purchased as a dry powder from Avanti Polar Lipids (Alabaster, AL) and used as received. F-hexane (purity >99%) and Pluronic F68 (a poly(oxyethylene)−poly(oxypropylene) triblock copolymer, MW ≈ 8300, purity >99%) were from Sigma-Aldrich (Lyon, France), and Fdiglyme and F-triglyme (purity 98%) were from Exfluor Research Corp. (Round Rock, TX). A HEPES (N-(2-hydroxyethyl)piperazineN′-2-ethanesulfonic acid) buffer solution30 (from Sigma) (20 mmol L−1) in a 150 mmol L−1 NaCl solution was prepared, and its pH was adjusted to 7.0 with 1 N NaOH. Water was purified using a Millipore system (surface tension 72.1 mN m−1 at 20 °C, resistivity 18.2 MΩ cm). The surface tension of the HEPES buffer is 70.0 ± 0.2 mN m−1 at 25 °C. The characteristics of the FCs discussed are collected in Table 1. Preparation of Bubble Dispersions. DMPC (50 mmol L−1) was dispersed by magnetic stirring in the HEPES/NaCl buffer solution for one night at room temperature. Pluronic F68 was added to facilitate phospholipid dispersion and foam formation (DMPC:F68 molar ratio 10:1). A 1 mL aliquot of the DMPC/F68 dispersion was transferred to a glass tube (inner diameter 18 mm, length 63 mm) and presonicated under air at low power (setting 2) for 30 s at 25 °C. The sonicator (Vibracell, Bioblock Scientific, Illkirch, France) was equipped with a 3 mm titanium probe and operated at 20 kHz with an output power of ∼600 W (duty cycle 40%). The DMPC dispersion was sonicated for 15 s (setting 2, ∼200 W) at 25 °C under a N2 atmosphere saturated with F-hexane, F-diglyme , or F-triglyme. N2 was allowed to bubble through three successive vials containing the FC (all three FCs are liquid at room temperature) prior to being flushed into the DMPCcontaining glass tube. The sonicator probe was consistently placed 5 mm below the surface of the dispersion. The resulting aggregated microbubble suspension (foam) was immediately diluted with 14 mL of buffer. Size fractionation was achieved by flotation under gravity (30 min) or centrifugation (Sigma 4K10 centrifuge, Bioblock Scientific). In the latter case, an 8 mL aliquot of the suspension, taken from the glass vial with a 10 mL syringe, was centrifuged at 300 rpm for 1 min (rotor radius ∼6 cm). During centrifugation, the tip of the syringe was facing 1184


Langmuir

Article

allowing N2 to bubble successively through three glass vials containing the FC. A constant gas flow was maintained throughout the experiments. A 15 μL volume of a 1 mmol DMPC solution in chloroform was spread on the surface of the HEPES/NaCl buffer solution (pH 7.0). Bubble Profile Analysis Tensiometry. DMPC (0.05 mmol) was dispersed in 100 mL of the buffer (pH 7.0) using a magnetic stirrer for one night at 25 °C, which is above the phospholipid’s transition temperature (23 °C). Vesicles were prepared by sonicating samples (6 mL) of this dispersion at room temperature using a probe sonicator, first at power 3 (on a scale of 10) for 2 min and then at power 5 for 30 min. Interfacial tension (γ) measurements were performed immediately after sonication on a rising bubble tensiometer (Tracker, Teclis, Longessaigne, France). A rising bubble (∼6 μL) of air, or of FCsaturated air, was formed at the tip of the needle of a microliter syringe inside a quartz cell (∼10 mL) containing the 0.5 mmol L−1 DMPC dispersion. The setup was thermoregulated at ∼25 ± 0.5 °C.

moreover, a significant number of large bubbles were present. On the other hand, centrifugation provided narrowly dispersed, essentially monomodal bubble populations (∼1.2−1.3 ± 0.2 μm). F-hexane and F-diglyme led to very similar size and size distribution characteristics. It should also be noted that the presence of FCs in the filling gas provided significantly smaller centrifuged bubbles than with air alone. The above results were confirmed by optical microscopy (Figure 3). Bubble Stability. The obtainment of samples of bubbles having similar mean sizes and size distributions allowed meaningful comparison of bubble stability. The samples investigated all consisted of essentially monomodal populations of small bubbles (radius ∼1.2−1.3 ± 0.2 μm). The temporal evolution of the attenuation coefficient α was measured for a range of frequencies (Figure 4). The evolution of the radius distributions, as determined from the attenuation curves, is shown in Figure 5. It is seen that F-triglyme, although this FC produced the narrowest distributions, was less effective at stabilizing the bubbles over time than the other two FCs. It is also noteworthy that the bubble mean radii remained essentially constant over time in all three cases. This phenomenon was confirmed by optical microscopy. The number of bubbles in a given sample volume was the same with F-diglyme and F-triglyme and was lower than with F-hexane (see α values in Figure 4). The bubble volume fraction in the dispersions decreased significantly more rapidly over time when stabilization was achieved with F-triglyme as compared to F-hexane or F-diglyme (Figure 6). The half-lives of the F-hexane- and F-diglyme -stabilized bubbles were 149 ± 8 and 134 ± 3 min, respectively, but the half-life was only 76 ± 7 min for the F-triglymestabilized bubbles. For comparison, bubbles that did not contain an FC gas lasted 34 ± 3 min, but their mean radius and polydispersity were larger. A previous study reported half-lives of nonstabilized DMPCcoated air microbubbles below ∼100 s.23 However, the bubble population was highly polydisperse (2−10 μm). In this case, the probability of fusion of two bubbles is high because the internal pressure difference of differently sized bubbles is high and gas diffusion between bubbles is fast. In our case, the measurements were made on monomodal, narrowly dispersed small bubble fractions, which led to longer half-lives. Compression of Langmuir Monolayers of DMPC under an Atmosphere Saturated with Fluorocarbon Gases. Langmuir balance studies were undertaken to help identify possible contributions from the DMPC monolayer to bubble stabilization. When compressed at an air/water interface

■

RESULTS Initial Bubble Size Characteristics. The variation of the attenuation coefficient, α, of the ultrasound wave as a function of the ultrasound frequency, f, was determined at the initial measuring time, t0, which is 32 s after injection of the microbubbles in the acoustical cell. Figure 1 displays the results collected for microbubble populations obtained from a 50 mmol concentrated DMPC dispersion exposed to F-hexane, Fdiglyme, or F-triglyme and separated by centrifugation. The bubble size distributions (in volume) determined from the ultrasound absorption spectra are collected in Table 2. Table 2. Size Characteristics of the FC-Stabilized, DMPCCoated Microbubbles As Measured by the Acoustical Method Immediately after Preparation mean radius (μm ± SD) (polydispersivity index, %) bubble gas phase

prepared by centrifugation

air F-hexane F-diglyme F-triglyme

2.1 1.3 1.2 1.2

± ± ± ±

0.4 (9.5) 0.2 (7.7) 0.2 (8.3) 0.2 (8.3)

prepared by separation by gravity polydisperse 1.5 ± 0.2 (6.6) 1.5 ± 0.5 (16.7) 1.6 ± 0.3 (9.4)

Figure 2 depicts the size distributions of the FC-stabilized microbubbles fractioned by centrifugation (Figure 2a) or gravity (Figure 2b). It shows that centrifugation is more efficient than separation by gravity in terms of access to narrowly dispersed bubble populations. For F-diglyme , gravity separation yielded a main bubble population (centered at ∼1.5 μm) that was broad (PI = 16.7%) and had a shoulder at ∼2 μm;

Figure 2. Initial size distributions of microbubbles fractionated by (a) centrifugation or (b) flotation under gravity as determined acoustically. The bubbles contained air only (black) or were stabilized by F-hexane (red), F-diglyme (blue), or F-triglyme (green). 1185


Langmuir

Article

Figure 3. Initial size distributions of microbubbles fractionated by centrifugation as assessed by optical microscopy. The bubbles contained air only (a) or were stabilized by F-hexane (b), F-diglyme , (c) or F-triglyme (d).

Figure 4. Variation of the attenuation coefficient α as a function of time for DMPC-coated microbubbles separated by centrifugation and stabilized by (a) F-hexane, (b) F-diglyme , and (c) F-triglyme.

Figure 5. Time evolutions at 25 °C of the size distributions of DMPC-coated microbubbles stabilized by (a) F-hexane, (b) F-diglyme , and (c) Ftriglyme.

isotherms in the presence and absence of FC gases are shown in Figure 7. Under air, the extrapolated area, A∞, for DMPC was 70 ± 2 Å2 and the collapse pressure, πc, was 54 ± 2 mN m−1, in agreement with earlier reports.40 Under FC-saturated nitrogen, πc increased to 61 ± 2 mN m−1 for F-hexane and F-diglyme and to 58 ± 2 mN m−1 for F-triglyme. The FCs thus caused a significant decrease in surface tension from 17 to 10 mN m−1 (F-hexane and F-diglyme ) and 12 mN m−1 (F-triglyme). The larger surface pressures observed at large molecular areas (e.g., 120 Å2) likely reflect the insertion of FC molecules into the monolayer. However, upon compression, this pressure difference decreases, indicating that the FC molecules are progressively expelled from the DMPC monolayer. The pressure required to expel the FC from the monolayer is higher for F-diglyme (45 mN m−1) than for F-triglyme (28 mN m−1). In the case of F-hexane, the FC is never completely expelled from the monolayer, even at high pressures.

Figure 6. Time evolution, at 25 °C, of the bubble volume for small, DMPC-coated bubbles separated by centrifugation and stabilized by Fhexane (red), F-diglyme (blue), or F-triglyme (green) or containing air only (black).

at 25 °C, the DMPC monolayer is in the liquid expanded state. Its surface pressure, π, versus molecular area, A, compression 1186


Langmuir

Article

DMPC at the interface were quantified by measuring the time, t30%, after which the initial value, γ0, of the interfacial tension of the DMPC dispersion had been lowered by 30%. The kinetics of adsorption of DMPC was found to be accelerated up to 3fold when the FC gases were present.

■

DISCUSSION Contribution of the Fluorocarbon Gases to Microbubble Stabilization: Osmotic Effect. According to current theory,12 the deflation of a bubble in an aqueous medium involves first the rapid exchange of the diffusive gases (air and N2), leading to equilibration of the gas pressures in the bubble and in the surrounding medium. A second step corresponds to the deflation of the bubble itself, which is hindered by the slow diffusion of the poorly soluble FC gas. Eventually, when the saturation pressure reaches the Laplace pressure, the FC gas turns into a liquid and can no longer play its role, causing the bubble to collapse. The collapse radius values Rcoll can be calculated using Rcoll = 2γ/psatd, with γ being the bubble tension. γ was taken as 50 mN m−1 for DMPC-shelled bubbles, as determined by the bubble profile analysis tensiometry experiment (knowing that the equilibrium surface tension does not depend on the concentration). The Rcoll values obtained are ∼3.5, ∼4, and ∼100 μm for F-hexane, F-diglyme , and Ftriglyme, respectively. In this study, however, we were able to prepare narrowly sized bubble populations with much smaller radii of ∼1.3 ± 0.2 μm for all three FC gases investigated. Although the experimental values of the collapse radii cannot be measured exactly for such small bubbles due to the limit of the resolution of optical microscopy, these values are necessarily equal to or lower than those measured. This is particularly striking in the case of F-triglyme, with which it is possible to form stable bubbles with an actual radius smaller by almost 2 orders of magnitude than the calculated collapse radius. Likewise, according to theory, the calculated decrease rate of the square radius of an air bubble stabilized by F-hexane should be ∼0.36 μm2 s−1 (∼0.06 μm2 s−1 for F-diglyme ). This would mean that F-hexane-stabilized microbubbles with a radius of 1.5 μm should last only for ∼1.3 min (∼7.8 min for Fdiglyme ). This value is grossly at variance, by over 3 orders of magnitude, with the experimental data (149 ± 8 min for Fhexane) collected here. Therefore, the actual obtainment of such small and stable bubbles leads to questioning of the validity of the theoretical model. Adding a Contribution of the Fluorocarbon Gases to Bubble Tension Lowering. A first parameter that had not been taken into account in the theoretical treatment is a possible cosurfactant effect of FC gases on the bubble’s surface tension. We actually found that the presence of the FC in the bubble’s gas phase significantly decreases the interfacial tension of the DMPC monolayer. The bubble profile analysis tensiometry experiments show indeed that the interfacial tension of DMPC was decreased from 50 to 37 mN m−1 for both F-hexane and F-diglyme and to 40 mN m−1 in the case of F-triglyme. Using these values, the calculated collapse radii are lowered to ∼2.5 μm for F-hexane, ∼3 μm for F-diglyme , and ∼80 μm for F-triglyme. For F-hexane and F-diglyme these values are already more realistic than those calculated in the absence of the FC-induced tension lowering effect. The comparable effects of these two FCs can be explained by the fact that they have comparable psatd values and Ostwald coefficients (considering that the accuracy on these values is about 1 order of magnitude and that the temperature

Figure 7. Compression isotherms of monolayers of DMPC under air (black) and under N2 saturated with FC gases (red, F-hexane; blue, Fdiglyme ; green, F-triglyme).

The isothermal compressibility (C) was calculated from the π/A isotherms according to C = −1/A(dA/dπ). Compared to air, FC gases cause the compressibility of a DMPC monolayer to increase by 20−50% from 1.0 × 10−2 to (1.2−1.5) × 10−2 m mN−1. The reciprocal quantity of C, denoted as the surface compressional modulus, κ (or elasticity), is provided in the Supporting Information (Figure S1). The maximum κ values are ∼100 mN m−1 in the absence of FC, ∼80 mN m−1 in the presence of F-hexane or F-diglyme , and ∼60 mN m−1 with Ftriglyme. Adsorption of DMPC at the Interface between Water and FC-Saturated Air. Interfacial tension measurements are usually conducted with solutions. DMPC, however, has a very low cmc (6 × 10−9 mol L−1).40 Therefore, we have measured the adsorption at the air/water interface of DMPC molecules coming from dispersions of DMPC vesicles, which ensures that the solution is saturated with phospholipid. To ensure reproducibility, the size of the vesicles had to be the same in all the experiments, since this factor contributes to the rate of adsorption of the phospholipid at the interface.41 A dispersion of DMPC (0.5 mmol L−1) was used that had vesicles with a mean diameter of ∼29 ± 6 nm (a typical size distribution is provided in the Supporting Information, Figure S2). The presence of FC gases in the rising bubbles used for tensiometry measurements was found to substantially decrease the interfacial tension, γeq, of the aqueous DMPC dispersions (Figure 8). With F-hexane and F-diglyme , the surface tension

Figure 8. Variation of tension versus time at the interface between an aqueous DMPC vesicle dispersion (0.5 mmol L−1) and N2 saturated with F-hexane (red), F-diglyme (blue), or F-triglyme (green) as compared to air (black).

decreased from ∼50 to ∼37 mN m−1 after 3 h and, in the case of F-triglyme, to ∼40 mN m−1. The rates of adsorption of 1187


Langmuir

■

dependence can be neglected) (Table 1).12 It is therefore not surprising that F-hexane and F-diglyme lead to comparable bubble characteristics, in terms of both size and stability. However, adding this contribution still does not explain the discrepancy between the actual (∼2 h) and calculated (∼10 min) half-lives. In the case of F-triglyme, it is clear that the FCinduced tension lowering effect does not allow explanation of the fact that small bubbles can be obtained. Stabilizing the DMPC Monolayer against Collapse by Fluorocarbons. Another limitation of the current theory is that it assumes that the rate of exchange of the surfactant molecules between the bubble wall and the aqueous medium is far larger than the rate of deflation of the bubble. This implies that the surfactant’s solubility in the aqueous phase is sufficient to allow desorption from the interface to the water phase when pressure is exercised on the bubble and that the bubble’s wall has no mechanical effect. This is definitely not the case with DMPC, which was shown to be expelled from the bubble’s surface in the form of bilayer fragments through collapse and shedding mechanisms.23 It is likely that the insertion of FC molecules into the DMPC monolayer, as demonstrated by the compression and tensiometry experiments, modifies its physical state, which may stabilize the bubble wall against collapse and hence increase the bubble’s half-life. First, the Langmuir monolayer study shows that the FC gases significantly increase the collapse pressure of the DMPC monolayer (from 54 to 60 mN m−1 in the case of F-hexane and F-diglyme and 58 mN m−1 for F-triglyme). This means that, conversely, the actual interfacial tension of the bubble can be decreased from 17 to 10 mN m−1 (F-hexane and F-diglyme ) and 12 mN m−1 (Ftriglyme). Taking these values, the calculated collapse radii are lowered to ∼0.8 μm for F-hexane) and ∼0.9 μm for F-diglyme , values that are close to the experimental values (1.2 μm). One should also note that the compressibility of the DMPC monolayer is increased by 20−50% when the FC gases are present. This higher compressibility may increase the monolayer collapse pressure and endow the microbubbles with higher resilience during shrinkage caused by compression. F-triglyme provides the case of an FC for which the osmotic effect cannot contribute to bubble stabilization due to its very low psatd. F-triglyme has a much lower (by 2 orders of magnitude) psatd value than F-hexane and F-diglyme . The cwater and L values of F-triglyme have not been reported. It is likely that these parameters are somewhat comparable to those of perfluorooctane (Table 1), paralleling the F-diglyme /F-hexane case. It has been predicted that, with such a low psatd, the Ftriglyme would condense rapidly, bringing about the dissolution of the bubbles.12 This is definitely not the case. The dispersion still displays the echogenicity characteristic of a dispersion of microbubbles after 30 min. Although less stable than those stabilized by F-hexane or F-diglyme , F-triglyme-containing bubbles are definitely and significantly more stable than air bubbles (with a half-life of 76 min vs 34 min). As a consequence, the formation of small bubbles, as well as their stability, must have another origin, such as the modification of mechanical properties of the DMPC monolayer. We therefore conclude that the observed increase in bubble stability is due to the bubble shell robustness induced by the presence of Ftriglyme that makes the bubbles capable to withstand the Laplace pressure.

Article

CONCLUSIONS We have studied the effects of three FC gases (F-hexane, Fdiglyme , and F-triglyme) on the size and stability of DMPCshelled microbubbles using acoustical measurements and optical microscopy. Therefore, monomodal, initially similarly sized small bubble populations were prepared, which is mandatory for meaningful stability studies. In parallel, bubble profile analysis tensiometry experiments and compression isotherms of Langmuir monolayers have established that the FC gases exercise a cosurfactant effect by which they decrease the interfacial tension of the DMPC monolayer significantly and that their presence increases the compressibility of the DMPC bubble shell significantly. Although substantial, the decrease in interfacial tension, and consequent increased resistance to Laplace pressure, still does not suffice to explain the discrepancy that exists between the theoretical bubble collapse sizes and the actual obtainment of stable bubbles having much smaller sizes. The increase in compressibility of the DMPC monolayer caused by the FC gases may also contribute to prolonging the bubbles’ half-lives by increasing the bubbles' resilience. Stabilization of small-sized bubbles by Ftriglyme, for which no osmotic stabilization is expected, demonstrates that the latter mechanism is not the only one to operate. Therefore, we conclude that the role of the FCs, which until now was thought to be restricted to osmotic stabilization, also involves membrane property modifications, likely due to the insertion of FC molecules in the DMPC bubble wall, despite the fact that these molecules are lipophobic and not amphiphilic. The effects of FCs on phospholipid-based interfaces is also relevant to the design and properties of emulsions42 and vesicles43 and to applications in which FCs come in contact with biological membranes, for example, treatment of lung diseases,44 organ and cell preservation,45 and cell aggregation control.46

■

ASSOCIATED CONTENT

S Supporting Information *

Figures showing the variation of the compressional modulus of a DMPC monolayer in contact with a FC gas and mean diameter and size distribution of the DMPC vesicle dispersion used for adsorption rate determinations. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Fax: (+33)3 88 41 40 99. E-mail: marie-pierre.krafft@ics-cnrs. unistra.fr.

■

ACKNOWLEDGMENTS We thank the French Research Agency (ANR; Contract 06BLAN-305-01), the European Commission (Nanomagdye, Grant NMP3-SL-2008-214032), and the University of Strasbourg (UdS) for financial support. C.S. and S.R. acknowledge the ANR for a research fellowship.

■

REFERENCES

(1) Schutt, E. S.; Klein, D. H.; Mattrey, R. M.; Riess, J. G. Angew. Chem., Int. Ed. 2003, 42, 3218−3235. (2) Riess, J. G. Curr. Opin. Colloid Interface Sci. 2003, 8, 259−266. (3) Feinstein, S. B. Am. J. Physiol. Heart Circ. Physiol. 2004, 287, H450−H457. (4) Lindner, J. R. Nat. Rev. Drug Discovery 2004, 3, 527−532. 1188


Langmuir

Article

(39) Rasband, W. S. ImageJ; U.S. National Institutes of Health: Bethesda, MD, 1997−2005; http://rsb.info.nih.gov/ij/. (40) Cevc, G. Phospholipids Handbook; Marcel Dekker: New York, 1993. (41) Nguyen, P. N.; Trinh Dang, T. T.; Waton, G.; Vandamme, T.; Krafft, M. P. ChemPhysChem 2011, 12, 2646−2652. (42) Marie Bertilla, S.; Thomas, J.-L.; Marie, P.; Krafft, M. P. Langmuir 2004, 20, 3920−3924. (43) Schmutz, M.; Michels, B.; Marie, P.; Krafft, M. P. Langmuir 2003, 19, 4889−4894. (44) Gerber, F.; Krafft, M. P.; Vandamme, T. F.; Goldmann, M.; Fontaine, P. Angew. Chem., Int. Ed. 2005, 44, 2749−2752. (45) Maillard, E.; Sanchez-Dominguez, M.; Kleiss, C.; Langlois, A.; Sencier, M. C.; Vohouhe, C.; Beitigier, W.; Krafft, M. P.; Pinget, M.; Belcourt, A.; Sigrist, S. Proc. Transplant. 2008, 40, 372−374. (46) Sanchez-Dominguez, M.; Krafft, M. P.; Maillard, E.; Siegrist, S.; Belcourt, A. ChemBioChem 2006, 7, 1160−1163.

(5) Unger, E. C.; Porter, T.; Culp, W.; Labell, R.; Matsunaga, T.; Zutshi, R. Adv. Drug Delivery Rev. 2004, 56, 1291−1314. (6) van Liew, H. D.; Burkard, M. E. Adv. Exp. Med. Biol. 1997, 411, 395−401. (7) Riess, J. G. Chem. Rev. 2001, 101, 2797−2920. (8) Kaufmann, B. A.; Lindner, J. R. Curr. Opin. Biotechnol. 2007, 18, 11−16. (9) Kaufmann, B. A. Curr. Cardiovasc. Imaging Rep. 2010, 3, 18−25. (10) Hernot, S.; Klibanov, A. L. Adv. Drug Delivery Rev. 2008, 60, 1153−1166. (11) Ferrara, K.; Pollard, R.; Borden, M. Annu. Rev. Biomed. Eng. 2007, 9, 415−447. (12) Kabalnov, A.; Klein, D.; Pelura, T.; Schutt, E.; Weers, J. Ultrasound Med. Biol. 1998, 24, 739−749. (13) Kabalnov, A.; Bradley, J.; Flaim, S.; Klein, D.; Pelura, T.; Peters, B.; Otto, S.; Reynolds, J.; Schutt, E.; Weers, J. Ultrasound Med. Biol. 1998, 24, 751−760. (14) Flaim, S. F. Artif. Cells, Blood Substitutes, Immobilization Biotechnol. 1994, 22, 1043−1054. (15) Rico-Lattes, I.; Quintyn, J. C.; Pagot-Mathis, V.; Benouaich, X.; Mathis, A. In Advances in Fluorine ScienceFluorine and Health. Molecular Imaging, Biomedical Materials and Pharmaceuticals, Tressaud, A., Haufe, G., Eds.; Elsevier: Amsterdam, 2008; Chapter 9, pp 407− 420. (16) Kobuch, K.; Menz, D. H.; Hoerauf, H.; Dresp, J. H.; Gabel, V.-P. Graefe’s Arch. Clin. Exp. Ophthalmol. 2001, 239, 635−642. (17) Krafft, M. P.; Riess, J. G. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 1185−1198. (18) Bleyl, J. U.; Ragaller, M.; Tscho, U.; Regner, M.; Kanzow, M.; Hubler, M.; Rasche, S.; Albrecht, M. Anesthesiology 1999, 91, 461−469. (19) Gerber, F.; Krafft, M. P.; Vandamme, T. F.; Goldmann, M.; Fontaine, P. Biophys. J. 2006, 90, 3184−3192. (20) Nakahara, H.; Lee, S.; Krafft, M. P.; Shibata, O. Langmuir 2010, 26, 18256−18265. (21) Haiss, F.; Jolivet, R.; Wyss, M. T.; Reichold, J.; Braham, N. B.; Scheffold, F.; Krafft, M. P.; Weber, B. J. Physiol. 2009, 587, 3153− 3158. (22) Stride, E.; Edirisinghe, M. J. Med. Biol. Eng. Comput. 2009, 47, 809−811. (23) Pu, G.; Borden, M. A.; Longo, M. L. Langmuir 2006, 22, 2993− 2999. (24) Borden, M.; Longo, M. Langmuir 2002, 18, 9225−9233. (25) May, D. J.; Allen, J. S.; Ferrara, K. W. IEEE Trans. Ultrason., Ferroelectr., Freq. Control 2002, 49, 1400−1410. (26) Gerber, F.; Krafft, M. P.; Waton, G.; Vandamme, T. F. New J. Chem. 2006, 30, 524−527. (27) Rossi, S.; Waton, G.; Krafft, M. P. ChemPhysChem 2008, 9, 1982−1985. (28) Rossi, S.; Waton, G.; Krafft, M. P. Langmuir 2010, 26, 1649− 1655. (29) Feshitan, J. A.; Chen, C. C.; Kwan, J. J.; Borden, M. A. J. Colloid Interface Sci. 2009, 329, 316−324. (30) Luo, S.; Pal, D.; Shah, S. J.; Kwatra, D.; Paturi, K. D.; Mitra, A. K. Mol. Pharmaceutics 2010, 7, 412−420. (31) Marchionni, G.; Ajroldi, G.; Righetti, M. C.; Pezzia, G. Macromolecules 1993, 26, 1751−1757. (32) Hayduk, W.; Laudie, H. AIChE J. 1974, 20, 611. (33) Lawson, D. D.; Moacanin, J.; Scherer, K. V.; Terranova, T. F.; Ingham, J. D. J. Fluorine Chem. 1978, 12, 221−236. (34) Kabalnov, A. S.; Makarov, K. N.; Shcherbakova, O. V. J. Fluorine Chem. 1990, 50, 271−284. (35) De Jong, N.; Hoff, L.; Skotland, T.; Bom, N. Ultrasonics 1992, 30, 95−103. (36) Goertz, D. E.; de Jong, N.; van der Steen, A. F. W. Ultrasound Med. Biol. 2007, 33, 1376−1388. (37) Medwin, H. J. Acoust. Soc. Am. 1977, 62, 1041−1044. (38) Leighton, T. G. The Acoustic Bubble; Academic Press: San Diego, CA, 1994. 1189


Effects of Perfluorocarbon Gases on the Size and Stability

Recommend Documents