Langmuir 2009, 25, 3705-3712
3705
Microbubbles Coated with Disaturated Lipids and DSPE-PEG2000: Phase Behavior, Collapse Transitions, and Permeability Monica M. Lozano and Marjorie L. Longo* Department of Chemical Engineering and Materials Science UniVersity of California, DaVis, California 95616 ReceiVed NoVember 14, 2008. ReVised Manuscript ReceiVed January 9, 2009 Saturated diacyl (disaturated) phosphatidylcholine (PC) mixed with the lipopolymer distearoylphosphatidylethanolamine (DSPE)-polyethyleneglycol molecular weight 2000 (PEG2000) self-assemble as a monolayer at the air-water interface of air-in-water micrometer-scale bubbles (microbubbles), similar to coatings (shells) on leading medical ultrasound contrast agents (UCAs). This system is characterized here to study the impact of the DSPE-PEG2000 species and PC chain-length on the monolayer coating phase behavior, collapse, shedding, and air transport resistance and microbubble dissolution rate and surface contour. Using fluorescence microscopy of dissolving microbubbles, we found that film microstructure and collapse behavior for all chain lengths (n ) 14-20) was indicative of primarily condensed phase monolayers, unlike similar coatings containing polyethyleneglycol 40 stearate (PEG40S) that are either expanded phase or coexisting expanded-condensed phase monolayers. Additionally, we observed a new surface buckling type of behavior with all chain lengths, by bright field microscopy, where the air-water interface continuously appears rough (rather than cyclically rough and smooth), with this behavior most frequently observed for n ) 16. In correlating the statistical frequency of this behavior with the monolayer microstructure, we propose that it arises from a slowed nucleation rate of collapse structures at condensed-condensed phase interfaces, not present in systems containing PEG40S. By modeling the dissolution (radius vs time) data, we obtained, for each chain length, the film air transport resistance (Rshell) that was then fit to a chain-length-dependent energy barrier model. Importantly, the pre-exponential factor was ∼10× higher and the microbubbles persisted ∼4× longer (from 15 µm at a fixed dissolved oxygen content) in comparison to previously studied films containing PEG40S. We attribute the unique stability properties of microbubble coatings containing DSPE-PEG2000 to the propensity of this molecule to form a condensedphase monolayer, such that the monolayer coatings approach the properties of one continuous condensed domain.
Introduction Microbubbles (radius ) 0.5-50 µm) consist of a gas-in-liquid core coated by a thin shell composed of proteins, surfactants, lipopolymers, and/or lipids.1 Current medically related investigations on microbubbles include the design of targeted drug delivery vehicles, metabolic gas carriers, and ultrasound contrast agents (UCAs).1-4 The dissolution rates of medical microbubbles and collapse/shedding behaviors of their coatings and surfaces can limit or extend their usefulness and functionality. For UCAs, preformed microbubbles are injected intravenously and circulation time is on the order of tens of minutes.5,6 For many UCAs the circulation time is limited by dissolution of the gas core into the blood pool.7,8 Previous studies have collapsed by ultrasound the lipid shell encapsulating microbubbles for the purpose of DNA or drug release, useful in therapeutics.3,9-15 Other studies * To whom correspondence should be addressed. E-mail: mllongo@ ucdavis.edu. (1) Ferrara, K.; Pollard, R.; Borden, M. Annu. ReV. Biomed. Eng. 2007, 9, 415–447. (2) Klibanov, A. L. AdV. Drug DeliVery ReV. 1999, 37, 139–157. (3) Lindner, J. R.; Kaul, S. Echocardiography 2001, 18, 329–337. (4) Stride, E.; Edirisinghe, M. Soft Matter 2008, 4, 2350–2359. (5) Fritz, T. A.; Unger, E. C.; Sutherland, G.; Sahn, D. InVest. Radiol. 1997, 32, 735–740. (6) Klibanov, A. L. J. Nucl. Cardiol. 2007, 14, 876–884. (7) 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. (8) Quaia, E. In Contrast Media in Ultrasonography: Basic Principles and Clinical Applications (Medical Radiology/Diagnostic Imaging); Quaia, E., Ed.; Springer: Weinheim, Germany, 2005; pp 3-14. (9) Muller, O. J.; Schinkel, S.; Kleinschmidt, J. A.; Katus, H. A.; Bekeredjian, R. Circulation 2007, 116, 65–65. (10) Wang, J. F.; Wang, J. B.; Chen, H.; Zhang, C. M.; Liu, L.; Pan, S. H.; Wu, C. J. AdV. Therapy 2008, 25, 412–421. (11) Borden, M. A.; Caskey, C. F.; Little, E.; Gillies, R. J.; Ferrara, K. W. Langmuir 2007, 23, 9401–9408.
have shown that shell properties (i.e., composition, microstructure, and surface tension) play an important role in the collapse of microbubbles induced by ultrasound.16-19 A synthetic coating (shell) composition that can be used to stabilize microbubbles consists of two species that assemble into a monolayer at the gas-water interface. The major component is a long-chain lipid, usually saturated diacyl (disaturated) phosphatidylcholine (PC) that imparts lateral rigidity, low surface tension, and reduced gas permeability. The minor component is a lipopolymer comprised of a lipid anchor such as distearoylphosphatidylethanolamine (DSPE) covalently linked by the headgroup to a hydrophilic polymer moiety such as polyethyleneglycol 2000 (PEG 2000), i.e., DSPE-PEG2000. The lipopolymer promotes self-assembly of the monolayer shell, prevents microbubble coalescence, and promotes biocompatibility. A current leading FDA-approved ultrasound contrast agent DEFINITY (Bristol-Myers Squibb Medical Imaging) contains a perfluorocarbon gas core coated with a similar mixture.20 We (12) Lum, A. F. H.; Borden, M. A.; Dayton, P. A.; Kruse, D. E.; Simon, S. I.; Ferrara, K. J. Controlled Release 2006, 111, 128–134. (13) Sboros, V. AdV. Drug DeliVery ReV. 2008, 60, 1117–1136. (14) Yeh, C. K.; Su, S. Y. Ultrasound Med. Biol. 2008, 34, 1281–1291. (15) Vancraeynest, D.; Kefer, J.; Hanet, C.; Fillee, C.; Beauloye, C.; Pasquet, A.; Gerber, B. L.; Philippe, M.; Vanoverschelde, J. L. J. Eur. Heart J. 2007, 28, 1236–1241. (16) Borden, M. A.; Kruse, D. E.; Caskey, C. F.; Zhao, S.; Dayton, P. A.; Ferrara, K. IEEE Trans. Ultrason., Ferroelectr., Frequency Control 2005, 52, 1992–2002. (17) Bloch, S. H.; Wan, P.; Dayton, P. A.; Ferrara, K. Appl. Phys. Lett. 2004, 84, 631–633. (18) Chomas, J. E.; Dayton, P. A.; Allen, J.; Morgan, K.; Ferrara, K. IEEE Trans. Ultrason., Ferroelectr., Frequency Control 2001, 48, 232–248. (19) Leong-Poi, H.; Song, J.; Rim, S. J.; Christiansen, J.; Kaul, S.; Lindner, J. R. J. Amer. Soc. Echocardiogr. 2002, 15, 1269–1276. (20) Kee, P. H.; McPherson, D. D. In Nanoparticles in Biomedical Imaging; Springer: New York, 2008; Vol. 102, pp 343-368.
10.1021/la803774q CCC: $40.75 2009 American Chemical Society Published on Web 02/13/2009
3706 Langmuir, Vol. 25, No. 6, 2009
Lozano and Longo
Figure 1. Chemical structure of (A) DSPE-PEG2000 and (B) PEG40 stearate.
and others have performed studies on microbubbles coated with monolayers of disaturated PC/lipopolymer monolayers that range from monodisperse size production and stabilization to applications as UCAs and drug delivery vehicles.16,21-25 However, despite the importance of dissolution and accompanying collapse/ shedding of the microbubble coating, these phenomena have not been studied using microbubbles coated with disaturated PC/ lipopolymer mixtures. The miscibility behavior of disaturated PCs and DSPEPEG2000 in monolayers and microbubbles has been controversial.2,22,26-29 In order to provide clarification, in a recent study, we developed binary phase diagrams for three disaturated PC/ DSPE-PEG2000 mixtures of increasing PC chain-length using complementary techniques and analysis in a Langmuir trough. We found for these binary phospholipid monolayer systems at 298 K: (1) dimyristoyl PC (DMPC, n ) 14)/DSPE-PEG2000 forms a stoichiometric condensed (C) complex, DMPC3DSPEPEG2000 that is miscible with excess lipid at a wide range of compositions and surface pressures; (2) dipalmitoyl PC (DPPC, n ) 16)/DSPE-PEG2000 displays wide compositional and pressure ranges where the two components are miscible in the expanded (E) and C phase; and (3) distearoyl PC (DSPC, n ) 18)/DSPE-PEG2000 displays monotectic behavior, indicating that its components are almost completely immiscible at all relevant pressures and compositions.30 An important determinant of the phase behavior of these mixtures is the ability of pure DSPE-PEG2000 to form a condensed phase.31-33 We have in the past performed experiments characterizing microbubble dissolution rates, monolayer film/microbubble collapse, and shedding on disaturated PC/polyethyleneglycol 40 stearate (PEG40S) shells as models for disaturated PC/DSPEPEG2000 shells (chemical structures for DSPE-PEG2000 and PEG40S are illustrated in Figure 1). These data yielded resistance (Rshell) to mass transfer of oxygen through these films of increasing PC chain-length. We also investigated the monolayer phase behavior of disaturated PC/PEG40S systems containing 10 mol % PEG40S at 298 K which differed significantly from the (21) Talu, E.; Hettiarachchi, K.; Powell, R. L.; Lee, A. P.; Dayton, P. A.; Longo, M. L. Langmuir 2008, 24, 1745–1749. (22) Borden, M. A.; Martinez, G. V.; Ricker, J.; Tsvetkova, N.; Longo, M. L.; Gillies, R. J.; Dayton, P. A.; Ferrara, K. W. Langmuir 2006, 22, 4291–4297. (23) Kim, D. H.; Klibanov, A. L.; Needham, D. Langmuir 2000, 16, 2808– 2817. (24) Talu, E.; Hettiarachchi, K.; Zhao, S.; Powell, R. L.; Lee, A. P.; Longo, M. L.; Dayton, P. A. Mol. Imaging 2007, 6, 384–392. (25) Talu, E.; Powell, R. L.; Longo, M. L.; Dayton, P. A. Ultrasound Med. Biol. 2007, 34, 1182–1185. (26) Chou, T.-H.; Chu, I.-M. Colloids Surf., A 2002, 211, 267–274. (27) Chou, T.-H.; Chu, I.-M. Colloids Surf., B 2003, 27, 333–344. (28) Frank, C. W.; Naumann, C. A.; Knoll, W.; Brooks, C. F.; Fuller, G. G. In Macromol. Symp.; Wiley-VCH 2001; Vol. 166, p 1-12. (29) Jebrail, M.; Schmidt, R.; DeWold, C. E.; Tsoukanova, V. Colloids Surf., A 2008, 321, 168–174. (30) Lozano, M. M.; Longo, M. L. Soft Matter 2008, in press. (31) Ahrens, H.; Baekmark, T. R.; Merkel, R.; Schmitt, J.; Graf, K.; Raiteri, R.; Helm, C. A. Chem. Phys. Chem. (Short Commun.) 2000, 2, 101–106. (32) Baekmark, T. R.; Elender, G.; Lasic, D. D.; Sackman, E. Langmuir 1995, 11, 3975–3987. (33) Wiesenthal, T.; Baekmark, T. R.; Merkel, R. Langmuir 1999, 15, 6837– 6844.
disaturated PC/DSPE-PEG2000 system containing 10 mol % DSPE-PEG2000. Therefore for microbubbles coated with these mixtures we expect differences in rates of dissolution, Rshell values, and dissolution behaviors. Specifically, for DMPC/PEG40S the monolayer remained in an E phase at all surface pressures and appeared to be well mixed evidenced by no apparent loss of PEG40S during film compression.34 DPPC or DSPC mixed with PEG40S displayed ideal immiscibility with the C-phase PC domains coexisting with E-phase PEG40S that was mostly “squeezed-out” at its relatively low collapse pressure (∼35 mN/ m).34 Air microbubbles coated with these mixed monolayers displayed behavior during dissolution that depended on the number of acyl groups per PC chain. For n e 14, the lipid-coated microbubbles dissolved smoothly and symmetrically because the lipid shell was expelled by forming micelles (n ) 12) or vesicles (n ) 14).35 For n g 16, monolayer collapse on the dissolving microbubble took place in two steps. First small attached vesicles appeared, presumably enriched in “squeezedout” PEG40S, then as the bright PEG40S phase was mostly eliminated, lipid folds appeared.35,36 For n g 18, a new behavior was discovered, lipid folding was accompanied by surface buckling and expulsion of excess lipids via folds.35 The influence of chain length and accompanying mechanical stiffening37 on the collapse mechanism was in agreement with previous works by Lee and co-workers that characterized anisotropic monolayer overcompression in a Langmuir trough.38,39 By analyzing the radius vs time data, we found that air transport resistance of coatings composed of disaturated PC mixed with PEG40S is significant for n > 18 and increases exponentially with increasing number of carbons per chain.40 In the first part of our study, we make general observations of the phase behavior and collapse/shedding mechanisms of disaturated PC/DSPE-PEG2000 monolayer films coating dissolving microbubbles using fluorescence microscopy. We statistically characterize the PC chain-length dependence of the monolayer collapse/shedding mechanisms by bright field microscopy during microbubble dissolution. In the second part of our studies, we obtain the mass transport resistance to air of disaturated PC/DSPE-PEG2000 shells of increasing PC chainlength coating dissolving microbubbles under controlled conditions by modeling the dissolution of air microbubbles in an aqueous media. Using the results of these and previous studies, we surmise that acyl chain condensation in DSPE-PEG2000, not present in the PEG40S of the previously studied systems, dramatically impacts the dissolution behavior in this mixed system via modification of phase behavior. (34) Borden, M. A.; Pu, G.; Runner, G.; Longo, M. L. Colloids Surf., B 2004, 35, 209–223. (35) Pu, G.; Borden, M. A.; Longo, M. L. Langmuir 2006, 22, 2993–2999. (36) Borden, M. A.; Longo, M. L. Langmuir 2002, 18, 9225–9233. (37) Kim, D. H.; Costello, M. J.; Duncan, P. B.; Needham, D. Langmuir 2003, 19. (38) Gopal, A.; Belyi, V. A.; Diamant, H.; Witten, T. A.; Lee, K. Y. C. J. Phys. Chem. B 2006, 110, 10220–10223. (39) Gopal, A.; Lee, K. Y. C. J. Phys. Chem. B 2001, 105, 10348–10354. (40) Borden, M. A.; Longo, M. L. J. Phys. Chem. B 2004, 108, 6009–6016.
PC-Lipid/DSPE-PEG2000 Microbubbles
Materials and Methods Materials. The lipid components, 1,2-dimyristoyl-sn-glycero3-phosphocholine (DMPC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), 1,2-diarachidoyl-sn-glycero-3-phosphocholine (DAPC), and lipopolymer, 1,2-distearoyl-sn-glycero-3-phosphaethanolamine-N-[methoxy(polyethylene glycol)-2000] (DSPE-PEG 2000), were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). The fluorescent probe, 3H-indolium,2-[3-(1,3-dihydro-3,3-dimethyl-1-octadecyl-2H-indol-2-ylidene)-1-propenyl]-3,3-dimethyl-1-octadecyl-perchorate (DiIC18) was purchased from Invitrogen (Carlsbad, CA). Microbubble Preparation and Dissolution Experiments. Microbubbles were prepared as described elsewhere.36 Briefly, lipid-coated microbubbles were prepared by dissolving a mixture containing 90 mol % PC and 10 mol % DSPE-PEG2000 in chloroform in a 2.5 mL glass scintillation vial. Chloroform was then removed via evaporation using a gentle stream of nitrogen to obtain a thin mixed lipid film. The sample was then placed in a vacuum oven for 1 h. Then, the film was hydrated with 3 mL of ultrapure water (total lipid 3 mg/mL) of resistivity ∼18 MΩ · cm and resuspended via bath sonication (Fisher, Pittsburgh, PA) with heat for 20 min. The solution was first dispersed with a tip sonicator (Branson, Danbury, CT) placed below the air-water interface at low power for 3 min. Then, the fluorescent probe DiIC18 was added to the suspension of small unilamellar vesicles and allowed to incubate for 20 min at room temperature. The amount of DiIC18 added was 1 µg for every 1 mg of total lipid mass. Microbubbles were finally formed by bringing the probe sonifier tip to the liquid-air interface of the suspension and sonicating at room temperature for 10 s at high power. The microbubble solution was immediately cooled to room temperature by running cold tap water over the exterior of the vial. The dissolution experiments were performed similarly to previous experiments.36 Ultrapure water was degassed while undergoing vigorous stirring in a controlled vacuum. The aqueous air concentration following degassing was verified by measuring the percent dissolved oxygen with a multiprobe system (model YSI 556 MPS, Multi Probe System; YSI Incorporated, Yellow Spring, OH). All microbubble experiments were performed within 1 h of microbubble preparation. Microbubbles were injected from a sample syringe into a laminar flow perfusion chamber (Warner Instruments, Hamden, CT). Microbubbles immediately raised to the top coverglass slips after injection into the chamber due to buoyancy. Both the top and bottom coverglass slips were coated with SurfaSil (Pierce, Rockford, IL) prior to experimentation to prevent electrostatic interactions. Several chamber volumes of degassed water were pumped through the chamber using a syringe pump (Kd Scientific, Holliston, MA) to purge the system. Microbubbles were imaged as they dissolved into degassed water using a Nikon eclipse e400 upright microscope (Nikon, Melville, NY) equipped with a high resolution Orca digital camera (Hamamatsu, Japan). Digital images were analyzed and radius as a function of time was determined using PC Image (Scion, Frederick, MD). Further analysis was performed using a Matlab (The MathWorks, Inc., Natick, MA) file developed in a previous study.41
Results and Discussion Phase Behavior and Collapse of Disaturated PC/DSPEPEG2000 Coatings on Dissolving Microbubbles. In this study, fluorescence microscopy was used to observe the phase behavior (41) Lozano, M. M.; Talu, E.; Longo, M. L. J. Geophys. Res. 2007, 112.
Langmuir, Vol. 25, No. 6, 2009 3707 Table 1. Dominant Phase Microstructures and Collapse/ Shedding Behaviors for Each Disaturated PC Chain Length Mixed with DSPE-PEG2000 PC
chain length
DMPC
14
DPPC
16
DSPC DAPC
18 20
a
microstructurea uniform (or nanodomains not visible under FM) or small domains medium and small domains small domains small domains
microscopic shell collapse and shedding behavior suboptical particles, small vesicles, short tubular vesicles, folds dark folds bright folds + buckle-folds-smooth bright folds + severe buckle-foldssmooth
25 µm2 > medium >2 µm2, small 1000 µm2) of the 10 mol % DSPE-PEG2000 mixture. However, higher area compression rates (e.g., ∼2 × 10-2 s-1 at ro ) 15 µm) were achieved here by these dissolving microbubbles in degassed media. We show in Figure 2B using a Langmuir trough that a faster compression rate resulted in much smaller domains comparable to the microbubble in Figure 2A. We nucleated these domains at a surface pressure of 30 mN/m, which is above the liquidus line of 28 mN/m,30 by using a compression rate ∼2 × 10-3 s-1 and then stopping the compression at 30 mN/m. If we did not stop the compression, as shown in Figure 2C, we observed no visible domain nucleation as high as 40 mN/m, which is near the solidus line of 44 mN/m. 60% of the domain-displaying microbubbles had small domains, similar to Figure 2A, and did not form visible collapse structures as they dissolved, indicating that the monolayer collapsed from a 2D film into micelles. The other 40% had (42) Pu, G.; Borden, M. A.; Longo, M. L. Langmuir 2006, 22, 2993–2999. (43) Shen, Y.; Powell, R. L.; Longo, M. L. Langmuir 2008, 24, 10035–10040.
3708 Langmuir, Vol. 25, No. 6, 2009
Lozano and Longo
Figure 2. (A) Fluorescent micrograph illustrating the shell morphology of a microbubble coated with DMPC/DSPE-PEG2000/DiIC18. Fluorescence micrographs of a monolayer composed of 89 mol % DMPC/10 mol % DSPE-PEG2000/1 mol % Texas Red-DHPE: (B) arrested isothermal compression at a surface pressure of 30 mN/m illustrating nucleation of domains and (C) isothermal compression at a rate of 2 × 10-3 s-1. Scale bar represents 20 µm.
medium-sized domains (>2 µm2) surrounded by a large interdomain area, similar to Figure 3A, and collapsed by visible folding (see arrows in Figure 3A and Movie 1) and detachment of each fold along its length to form a long tubular aggregate. The presence of C phase domains and fold collapse structures for the DMPC/ DSPE-PEG2000 coating vs their absence in the noncondensing DMPC/PEG40S coating studied earlier was the most striking difference in the fluorescence results when comparing to our previous studies.34,42 Fold collapse structures in the presence of C phase domains are expected since binary lipid monolayers of similar chain lengths containing a condensed fraction have been shown to collapse only by fold structures at several degrees below the main phase transition temperature of the condensed component.39 Seventy percent of the DMPC/DSPE-PEG2000 microbubbles had surfaces that appeared uniformly fluorescent during dissolution, as shown in Figure 3B-D. According to the Langmuir trough results shown in Figure 2C, the lack of domains may be indicative of a fast compression rate, where either the C phase domain nucleation rate was very high so that the domains were numerous and very small or the domains did not nucleate (remained in E-phase). The collapse aggregates formed upon dissolution were generally fluorescent and visible under the microscope. Within this population of microbubbles, 70% formed dumbbell-shaped collapse aggregates that were continuously expelled during dissolution (∼5 µm in length and ∼1 µm in diameter at the spherical ends) (Figure 3B, some collapse aggregates are circled, and Movie 2). The dumbbell-shaped vesicles appear to be tubular vesicles that have contracted at their ends and will be discussed below in relation to a similar observation by Gopal and Lee in a binary monolayer.39 The rest of the microbubbles either formed spherical aggregates (Figure 2C) accompanied by the formation of bright narrow (∼1 µm) folds at small radii that spanned the circumference of the bubble or formed suboptical collapse structures (Figure 3D) indicating that the monolayer collapsed from a 2D film into micelles upon compression. In the work of Gopal and Lee,39 it was found that
Figure 3. Fluorescent micrographs illustrating the shell morphology and collapse behavior of dissolving microbubbles coated with a binary phospholipid monolayer. Lipid shell composed of DSPE-PEG2000 and (A-D) DMPC; (E and F) DPPC; (G) DSPC; and (H) DAPC. Fluorescent probe DiIC18. Scale bars represent 20 µm.
the collapse aggregates of a binary monolayer film transitioned from spherical vesicles to dumbbell-shaped (attached) vesicles to folds as the surface pressure was increased at temperatures a couple degrees below the main phase transition temperature of the condensed component. The observation of predominantly these three structures in our DMPC/DSPE-PEG2000-coated microbubbles is consistent with these observations of Gopal and Lee and our recent monolayer observation30 that this film composition forms a C-phase stoichiometric complex that collapses via tubular vesicles followed by folds at very high surface pressure (∼70 mN/m). Therefore, the homogeneously bright coating on most microbubbles, and even the bright interdomain regions on domain-containing coatings probably contained suboptical C phase domains that nucleated rapidly, as we were able to replicate in the Langmuir trough at compression rates similar to those of the microbubble coatings. The range of behaviors that we observed indicates that the monolayers of individual microbubbles existed at slightly different surface pressures, compositions, and rates of compression. The micelle formation from 30% of the microbubbles coated with DMPC/ DSPE-PEG2000 suggests that for this composition, micelle formation should be included in the sequence of possible collapse aggregates, perhaps in the order, micelles, vesicles, dumbbell-
PC-Lipid/DSPE-PEG2000 Microbubbles
Langmuir, Vol. 25, No. 6, 2009 3709
Table 2. Domain Area Changes from Figure 3E Resulting from a Surface Fold domain number
domain area before (Ab) fold [µm2]
domain area after (Aa) fold [µm2]
percent area change ) (Ab - Aa)/ Ab × 100
1 2 3 4 5 total
10 11 12 18 16 67
8 9 8 14 13 52
20% 18% 33% 22% 19% 22%
shaped vesicles, and folds. Indeed, DSPE-PEG2000 in aqueous solution forms micelles of 6-10 nm diameter.44 The primary difference in the microstructure and collapse behavior for dissolving microbubbles coated with higher chain length (n > 14) disaturated PC mixed with DSPE-PEG2000 (here) vs PEG40S (past experiments),42 was the lack of observation here of attached vesicles before larger scale collapse structures were seen, a consequence of differences in phase behavior that we will discuss here. In our previous studies, we did not carefully examine the influence of compression rate, fold shapes, or statistics for the range of behaviors. Therefore, we make these examinations here. For the case of DPPC/DSPEPEG2000, the majority of the microbubble shells contained wellpacked medium and/or small C phase domains, as illustrated in Figure 2E. Interestingly, the domains are the same size as observed in our recent study30 of this mixture as a Langmuir monolayer compressed at a rate of ∼1 × 10-3 s-1, possibly due to microbubble compression rates being within an order of magnitude (1 × 10-2 at ro ) 15 µm) of each other. The phase diagram for this system at XDSPE-PEG2000 ) 0.1 indicates that at surface pressures > ∼15 mN/m, a single C phase rich in DPPC forms,30 accounting for the lack of attached collapse vesicles here, rather than the phase-separated behavior for DPPC/PEG40S that resulted in “squeeze-out” of most of the E-phase PEG-40S via attached vesicles.34 For these microbubbles, the shells collapsed via the formation of nonfluorescent folds that can be observed as a stepping down in size of the large microbubble in Movie 3. The folds appeared to nucleate in the interdomain regions and propagated across several domains. Table 2 summarizes the domain area change resulting from an impending fold as obtained from Figure 3E. The fold was ∼24 µm in length and based on the total area lost or expelled after the fold, the width of this bilayer fold would be ∼0.33 µm (or ∼0.66 µm for a monolayer) assuming its shape was rectangular. However, as shown in Table 2, the domain with the greater percent surface area loss is found midlength of the fold (domain 3). This can be attributed to a compromise between the sphericity of the bubble and the linear propagation of the fold. Therefore, the shape of the fold is similar to a bilayer wedge protruding normal to the microbubble’s surface. The width midlength of this fold is ∼0.6 or ∼1.2 µm if unfolded. For a smaller percentage of the DPPC/DSPE-PEG2000 microbubbles (10%) the large C phase domains are not as closely packed; however, the regions between the domains also appears to be fully or partially condensed, being significantly darker than typical interdomain regions (see arrows in Figure 3F pointing to light, gray, and dark regions and small microbubble in Movie 3). We observed a similar phenomenon recently in Langmuir trough experiments forming large domains of DMPC/DSPEPEG2000. The condensing of regions between domains occurred at the solidus phase boundary and represented the final expanded (44) Edwards, K.; Johnsson, M.; Karlsson, G.; Silvander, M. Biophys. J. 1997, 73, 258–266.
phase fraction to form a C phase. The microbubble shell collapsed via the formation of narrow fluorescent folds followed by the detachment of these folds along their length, similar to the fold that will be discussed below for Figure 3G (between arrows). Our results in general for the mixture containing DPPC indicate that the majority of the film is in or entering into a miscible C phase, i.e., generally lacks an E phase, in agreement with recent Langmuir monolayer results and collapses primarily by folding. The film comprised of DSPC/DSPE-PEG2000 coating a dissolving microbubble contains tightly packed small C phase domains, giving it a mottled appearance that can more clearly be seen in the second image of Figure 3G. The phase diagram of this mixture30 indicates that at this composition, the two components are both in the C phase and are almost entirely immiscible at high surface pressure. Seventy five percent of the microbubbles had shells that collapsed via the formation of fluorescent folds (Figure 3G, panel 1 with arrows pointing to the beginning and end of a fold) followed by the detachment of these folds along their length (Figure 3G, panel 2 with arrow pointing at loose tubular aggregate). The expelled folds appeared to form tubular aggregates or elongated bilayer sheets that remained attached at one or both ends to the lipid shell. Most of the folds were expelled (Figure 3G, panel three with arrows pointing at attached and detached ends of a tubular aggregate) after several folds formed and propagated in random directions oftentimes intersecting one another. This randomness may be attributed to the isotropic compression induced by dissolution allowing the film to collapse from every direction. Similar to the coordination of fold detachment observed here, it has been previously shown through Langmuir monolayer studies that upon compression above the collapse pressure, a monolayer containing condensed phase components might form a single fold that can “topple” several next generation folds.38 The other 25% of the microbubble shells collapsed similarly to the majority of DPPC/DSPEPEG2000 microbubbles or dissolved similarly to the majority of the DAPC/DSPE-PEG2000 microbubbles. For dissolving DAPC/DSPE-PEG2000 microbubbles as seen in Figure 3H, the surface appeared dark, indicating a C phase monolayer, and the collapse and shedding process resembled that for DSPC/DSPE-PEG2000, but for 80%, larger-scale changes on the surface took place. During the folding stage, thick fluorescent lines appeared (see panel 2 in Figure 3H and Movie 4), that according to the images presented in the next section probably correspond to severe surface buckling, i.e., rows of elongated dents in the microbubble surface where the threedimensional nature of these gives a high density of fluorescent probe. While for 20% of these microbubbles, their collapse resembled the majority of DPPC/DSPE-PEG2000 microbubbles or the majority of the DSPC/DSPE-PEG2000 microbubbles. Statistical Analysis of Dissolution Surface Contours. We have found bright field imaging useful for statistically characterizing the surface contours of dissolving microbubbles, since the surface contour tends to be more easily categorized in comparison to collapse structures observed by fluorescence microscopy, although the observations are closely related. In addition, surface buckling can more easily be identified in bright-field compared to fluorescence microscopy. Air-filled, disaturated PC/DSPEPEG2000-coated microbubbles dissolving in degassed water (f ) 0.85 ( 0.02) were observed using bright field microscopy to characterize the surface contours during dissolution. Table 3 presents statistics for the microbubble surface contours observed for each disaturated PC chain length mixed with DSPE-PEG2000. As we have shown before,36,42,43 the surface contour depends on both the chain length (n) of the main component and the area
3710 Langmuir, Vol. 25, No. 6, 2009
Lozano and Longo
Table 3. Surface Contour Statistics for Microbubble Shells Composed of Disaturated PC and DSPE-PEG2000 disaturated PC lipid mixed with DSPE-PEG2000 surface contour smooth continuous shell roughness buckle-smooth severe buckle-smooth
DMPC
DPPC
67% (Figure 4A) 33%
14% 86% (Figure 4B)
compression rate (determined by radius and f). Therefore, we have chosen to survey the surface contours of microbubbles within a certain initial radius, ro, range (ro ) 7-25 µm). Even within this range, differences in ro, dissolution rate, and minor variations in composition generally result in two or three behaviors being observed for a single mixture as we showed in the section above. However, one behavior is observed more than the others for each mixture. For the case of DMPC/DSPE-PEG2000 (Figure 4A), 67% of the microbubbles dissolved symmetrically and the surface remained smooth, nearly matching the percentage of microbubble coatings that contained no visible condensed domains. While 33% of the microbubbles dissolved with a surface that appeared rough throughout the dissolution, matching the percentage of microbubble coatings that contained microscopic domains. For DPPC/DSPE-PEG2000 (Figure 4B), the surface on 86% of dissolving microbubbles appeared slightly deformed and remained rough throughout the dissolution while 14% appeared smooth. The 14% may correspond to the 10% that we observed at this
Figure 4. Bright field micrographs illustrating the most common surface contours of dissolving microbubbles each coated with a mixed lipid monolayer. Lipid shell was composed of DSPE-PEG2000 and (A) DMPC, (B) DPPC, (C) DSPC, and (D) DAPC. Scale bars represent 20 µm.
DSPC
DAPC
8% 86% (Figure 4C) 6%
5% 10% 85% (Figure 4D)
composition in fluorescence with narrow folds that quickly detached. For the majority of the microbubbles coated with DSPC (n ) 18) or DAPC (n ) 20)/DSPE-PEG2000 (Figure 4C and D, respectively), a cyclic process took place where the surface was noticeably buckled, and in one step (less than 1/32 s) the surface became noticeably smoother and more spherical. This majority behavior correlates with the cyclic mechanism that we observed in fluorescence where folds and lipid material (presumably much of the excess lipid area) appeared to be sloughed off in a cyclic manner. The main difference between DSPC and DAPC/DSPEPEG2000 is that the latter surface is rougher and microbubble sphericity is often not restored completely. For the case of DSPC, each cycle allowed for the expulsion of an average of 19% of the shell’s surface area while for the case of DAPC it allowed for an average of 28%. Interestingly, these values were consistent regardless of the microbubble’s initial size. These values were obtained from radius vs time curves by calculating the average area lost between steps or discontinuities (see Figure 5A) in the curve. The main difference with the acyl chain-dependence of these surface contours in comparison to previous work is the presence
Figure 5. (A) Typical modeling to determine the shell resistance to air transport by comparing the experimental (data points) and theoretical dissolution (solid lines) of a microbubble. The mixed lipid shell was composed of 10 mol % DSPE-PEG 2000 and 90 mol % DMPC (blue circles), DPPC (red squares), DSPC (purple diamonds), and DAPC (green triangles). (B) Shell resistance modeled as an exponential function of the number of carbons per acyl chain for disaturated PC/DSPE-PEG2000 (red, Rshell ) 6.13 exp134n/BT) and disaturated PC/PEG40S (blue, Rshell ) 0.438 exp165n/BT).
PC-Lipid/DSPE-PEG2000 Microbubbles
Langmuir, Vol. 25, No. 6, 2009 3711
of a new behavior, continuous shell roughness. We did not observe this behavior previously in disaturated PC/PEG40S-coated microbubbles or microbubbles coated with an inexpensive food emulsifier (primarily monostearin and monopalmitin) mixed with PEG40S.36,43 We can draw the conclusion, based upon comparing statistics, that this behavior appears to be associated with microbubbles that contain condensed domains, but are not shedding excess lipid in a cyclic manner. Previously, we observed a smooth surface in these cases, indicating that the rate of loss of material by collapse aggregates was able to match the area compression rate caused by the shrinking radius of the microbubble during dissolution. The presence of a rough surface indicates that these rates are not matched, i.e., monolayer collapse is slower, and therefore, the surface buckles slightly giving it a rough appearance. An important difference between this system and the previous systems is that the interdomain regions are also capable of forming a condensed phase, where in comparison, some E phase PEG40S was always trapped between the domains. Nucleation of collapse aggregates occurs at these interfaces.39,42 This work indicates that the nucleation rate is lower if the interface is between two condensed phases, possibly because of a lower hydrophobic mismatch. Shell Resistance During Dissolution. The shell resistances to air transport for the various chain length mixtures were determined by fitting the experimental radius vs time data to a modified version of the Epstein and Plesset model. Epstein and Plesset (EP) derived a model45 that describes the diffusion-limited dissolution of a “clean” bubble at infinite dilution that includes the effect of surface tension at the air-water interface. This model was modified by Borden and Longo36,40 to incorporate a parameter that accounts for the gas mass transfer resistance, denoted Rshell, due to the monolayer coating a microbubble. Included in this model is the effect of an impermeable wall46 to hold a microbubble stationary during observation. As a result, the equation for a dissolving coated microbubble is given by
-
[
1 - f + 2σ ⁄ Par H 1 + 4σ ⁄ 3Par 0.693Dw + Rshell
dr ) dt r⁄
]
(1)
where r is the radius of the microbubble, H is the partition coefficient (0.019), f is the degassing factor (0.85 ( 0.02) of the aqueous medium, σ is the surface tension at the air-water interface, Pa is the atmospheric pressure (1 × 105 Pa), and Dw is the diffusion coefficient of air in water (2 × 10-5 cm2/s). Data analysis involved obtaining the theoretical radius vs time curves by integrating the EP model numerically where the surface tension parameter is determined experimentally. Rshell in this equation is allowed to vary such that the square of the difference between the theoretical and experimental radii scaled by the experimental radius is minimized using Matlab i)N
y)
∑ i)0
(
i riexpt - rth i rth
)
2
(2)
where rexpt and rth corresponds to the experimental and theoretical radius, respectively, at a given time denoted by i. The model is considered a good fit to the experimental data when the value for y is less than 0.1 (where y ) 0 indicates a perfect fit). Microbubble radius vs time data were obtained from bright field images for r < 15 µm. These data were then modeled with the modified EP equation to determine Rshell for air transport (45) Epstein, P. S.; Plesset, M. S. J. Chem. Phys. 1950, 18, 1505–1509. (46) Wise, D. L.; Houghton, G. Chem. Eng. Sci. 1968, 23, 1501–1503.
Table 4. Comparison of Shell Resistances to Air Transport and Dissolution Times (Initial r ) 15 µm) PC chain +DSPE-PEG2000 dissolution length resistance [s/cm] time [s] 14 16 18 20 22 24 DSPEPEG2000
156 ( 23 205 ( 53 386 ( 44 567 ( 83 198 ( 35
110 136 230 327 133
+ PEG40S resistance dissolution [s/cm] time [s] 1(1 3(1 23 ( 3 118 ( 9 237 ( 20 333 ( 18 -
29 30 41 91 153 204 -
assuming that the microbubble shells containing DSPE-PEG2000 can reach surface tensions close to 0 mN/m at these small radii (deduced from Langmuir isotherms30). The individual radius vs time curves (shown in Figure 5A) exhibit the cyclic steppingdown in size mechanism during dissolution for the cases of n ) 18 and 20. For n ) 16 the almost smooth curve indicates the shell’s slight roughness through the dissolution, while for n ) 14 the smooth curve indicates a symmetrical dissolution. The average Rshell values and their respective standard deviation of the mean values are summarized in Table 4. In that table we include for comparison, Rshell from microbubbles coated with disaturated PC/PEG40S of increasing PC chain-length. In order to obtain these values, we re-evaluated our old data, which was obtained by the same method used here, by fitting the radius vs time data from 15 µm, the initial radius used here. Dissolution times to dissolve completely from r ) 15 µm for each system were obtained by integrating eq 1 numerically at the average Rshell values. Note that dissolution times for microbubbles containing DSPE-PEG2000 are generally ∼4× larger than those containing PEG40S (see Table 4). These dissolution times are specific to the degassing factor experimentally imposed here. The degassing factor of the media surrounding freshly prepared microbubbles, which can be greater than 1 causing individual microbubbles to grow, will depend upon the preparation method. Analysis of these data indicates that the shell resistance for a mixed C phase monolayer increases exponentially, according to the energy barrier model, with increasing fatty acid chain length of the main component in agreement with our previous findings for disaturated PC/PEG40S.40,47 In Figure 5B, we successfully modeled the shell resistance as a function of the number of carbons per chain (n) according to the energy barrier model Rshell ) RoexpEn/BT, where Ro is the pre-exponential factor, E is the additional activation energy per CH2 group per chain, B ) 1.987 cal/mol · K, and T ) 298 K. We obtained E values for the disaturated PC/DSPE-PEG2000 and disaturated PC/ PEG40S of 134 and 165 cal/mol, respectively, which are in the same order of magnitude as the value of 190 cal/mol determined by La Mer for the permeation of water vapor through condensed monolayers of fatty acids.48 Furthermore, as reported by us47 for the disaturated PC/PEG40S, the pre-exponential factor, Ro, was inversely proportional to the domain boundary density of the monolayer, β, which is equal to the ratio between the average domain perimeter and area. In the case of that system, significant air permeation could take place through low surface density E-phase interdomain regions of remaining PEG40S rather than through the highly ordered crystalline domains. By comparing the pre-exponential factors (6.13 vs 0.438) we estimate that 14 times less highly permeable boundary exists per area for the disaturated PC/DSPE-PEG2000 vs disaturated PC/PEG40S. This (47) Pu, G.; Longo, M. L.; Borden, M. A. J. Am. Chem. Soc. 2005, 127, 6524–6525. (48) Archer, R. J.; La Mer, V. K. J. Phys. Chem. B 1954, 59, 200.
3712 Langmuir, Vol. 25, No. 6, 2009
factor drops to 9 if we assume E ) 150 cal/mol (average E between systems) for both systems and then compare their preexponential factors. Taken together, this work strongly suggests that the high gas transport resistance of microbubbles coated with disaturated PC and DSPE-PEG2000 is primarily a consequence of the phase behavior of this system, i.e., the entire film can form a C phase, including the interdomain regions. As an example, for the case of phase separated DSPC and DSPEPEG2000, DSPC-rich C phase domains are separated by a C phase rich in DSPE-PEG2000 with a high resistance of ∼200 s/cm, that can be compared to a much lower resistance of ∼0 s/cm for PEG40S. The shell resistance behavior approaches closely that of a single condensed phase domain (single-crystal) with resultant 4× increase in microbubble persistence.
Conclusion The dissolution rates of medical microbubbles and collapse/ shedding behaviors of their monolayer coatings (shells) and surfaces can, respectively, limit their circulation time in the blood pool and be related to their ultrasound-induced collapse behavior for drug and DNA targeting. We found here that the preexponential factor that controls the shell air transport resistance is ∼10× higher when 10 mol % DSPE-PEG2000 is included in the 90 mol % disaturated PC shell compared to 10 mol % PEG40S, which we interpret as a 10× decrease in the fraction of low
Lozano and Longo
surface density interdomain regions. This conclusion is consistent with the observations in the phase behavior of the monolayer coating and coating/microbubble collapse that suggest a trend toward complete monolayer condensation for all PC chain lengths (n ) 14-20), including the interdomain regions. We found that disaturated PC/DSPE-PEG2000-coated microbubbles displayed a new surface contour, continuous surface roughness (majority structure for n ) 16), which we interpret as a slowing of the rate of film collapse by folds and therefore longer retention time of the coating of the dissolving microbubble. Overall, these results suggest that the lipopolymer component should be carefully chosen to promote condensation of the monolayer film, as achieved using DSPE-PEG2000, and the saturated PC component can be quantitatively chosen, using the energy barrier model, to give the desired film gas transport resistance and desired microbubble collapse behavior. Acknowledgment. Funding provided by a GAANN fellowship from the U.S. Department of Education and the NSF MRSEC Program CPIMA (NSF DMR 0213618). Supporting Information Available: Movies (4) showing collapse and shedding of fluorescently labeled microbubble shells. This material is available free of charge via the Internet at http://pubs.acs.org. LA803774Q
