7430
Langmuir 2004, 20, 7430-7435
Characterization of Fluorocarbon-in-Water Emulsions With Added Triglyceride Jeffry G. Weers,*,† Rebecca A. Arlauskas, Thomas E. Tarara,‡ and Timothy J. Pelura§ Alliance Pharmaceutical Corp., 6175 Lusk Boulevard, San Diego, California 92121 Received March 10, 2004. In Final Form: May 22, 2004 Fluorocarbon-in-water emulsions are being explored clinically as synthetic oxygen carriers in general surgery. Stabilizing fluorocarbon emulsions against coarsening is critical in maintaining the biocompatibility of the formulation following intravenous administration. It has been purported that the addition of a small percentage of long-chain triglyceride results in stabilization of fluorocarbon emulsions via formation of a three-phase emulsion. In a three-phase emulsion, the triglyceride forms a layer around the dispersed fluorocarbon, thereby improving the adhesion of the phospholipid surfactant to the dispersed phase. In the present study, we examined the effect of triglyceride addition on the physicochemical characteristics of the resulting complex dispersion. In particular, we examined the particle composition and stability of the dispersed particles using a method which first fractionates (classifies) the different particles prior to sizing (i.e., sedimentation field-flow fractionation). It was determined that the addition of a long-chain triglyceride (soybean oil) results in oil demixing and two distinct populations of emulsion droplets. The presence of the two types of emulsion droplets is not observed via light scattering techniques, since the triglyceride droplets dominate the scattering due to a large difference in the refractive index between the particles and the medium as compared to fluorocarbon droplets. The growth of the fractionated fluorocarbon emulsion droplets was followed over time, and it was found that there was no difference in growth rates with and without added triglyceride. In contrast, addition of medium-chain-triglyceride (MCT) oils results in a single population of emulsion droplets (i.e., a three-phase emulsion). These emulsions are not stable to droplet coalescence, however, as significant penetration of MCT into the phospholipid lipid interfacial layer results in a negative increment in the monolayer spontaneous curvature, thereby favoring waterin-oil emulsions and resulting in destabilization of the emulsion to the effects of terminal heat sterilization or mechanical stress.
Introduction Due to their enormous capacity for dissolving gases, fluorocarbons (FCs) are being developed as synthetic oxygen carriers (“blood substitutes”).1-3 Additionally, FCbased formulations have been employed in targeted therapeutics and molecular imaging applications.4-7 To prevent formation of fatal emboli following intravascular administration, the FC must first be emulsified in a water continuous phase. Clearance of emulsion droplets from the vascular compartment by monocytes and macrophages is controlled to a large extent by the size of the dispersed droplets.8,9 Rapid clearance of emulsion droplets by the reticuloendothelial system may be accompanied by flulike * Corresponding author. E-mail:
[email protected]. † Current address: Transave, Inc., 11 Deer Park Drive, Suite 117, Monmouth Jct, NJ 08852. ‡ Current address: Nektar Therapeutics, 150 Industrial Road, San Carlos, CA 94070. § Current address: Kereos Inc., 4041 Forest Park Avenue, St. Louis, MO 63108. (1) Krafft, M. P.; Riess, J. G.; Weers, J. G. In Submicron Emulsions in Drug Targeting and Delivery; Benita, S., Ed.; Harwood Academic Publishers: Amsterdam, The Netherlands, 1998; p 235. (2) Riess, J. G. Chem. Rev. 2001, 101, 2797. (3) Krafft, M. P.; Chittofrati, A.; Riess, J. G. Curr. Opin. Colloid Interface Sci. 2003, 8, 251. (4) Lanza, G. M.; Wallace, K. D.; Scott, M. J.; Cacheris, W. P.; Abendschein, D. R.; Christy, D. H.; Sharkey, A. M.; Miller, J. G.; Gaffney, P. J.; Wickline, S. A. Circulation 1996, 94, 3334. (5) Lanza, G. M.; Abendschein, D. R.; Hall, C. S.; Marsh, J. N.; Scott, M. J.; Scherrer, D. E.; Wickline, S. A. Invest. Radiol. 2000, 35, 227. (6) Flacke, S.; Fischer, S.; Scott, M. J.; Fuhrhop, R. J.; Allen, J.; McLean, M.; Winter, P.; Sicard, G. A.; Gaffney, P. J.; Wickline, S. A.; Lanza, G. M. Circulation 2001, 104, 1280. (7) Winter, P. M.; Caruthers, S. D.; Kassner, A.; Harris, T. D.; Chinen, L. K.; Allen, J. S.; Zhang, H.; Robertson, D. J.; Wickline, S. A.; Lanza, G. M. Cancer Res. 2003, 63, 5838.
symptoms (e.g., fever, chills) resulting from the release of metabolites of the arachidonic acid cascade (e.g., thromboxanes, prostaglandins, and interleukins).10 The magnitude of the fever is significantly muted for particles 67% of all the measured fractures had diameters below 140 nm, and a bimodal size distribution (mean diameter ) 117 nm) was obtained. According to TEM, the majority of the particles were overlooked by QLS. This was confirmed by the fact that SUVs (48 nm) could not be detected by QLS even in a 1:100 admixture of emulsion/SUVs. Quantitative 31P NMR spectroscopy also indicated that the majority of particles (>50% by number) had diameters 300
droplet diameter (µm) PS
QLS
0.23 0.08 0.08 0.08 0.08
0.21 0.19 0.19 0.20 0.18
SdFFF 0.25 0.27/0.18
increases the volume percentage of the dispersed phase to 51-53%. Significant increases in viscosity and pseudoplastic behavior have been noted for PFOB-in-water emulsions above ∼50%.28 Indeed, significant increases in kinematic viscosity were noted for the emulsions containing added LCT. A good correlation is noted for the median droplet size between all three sizing techniques (PS, QLS, and SdFFF) for the PFOB-in-water emulsion without added LCT (d50 ) 0.23 ( 0.02 µm). Curiously, the median droplet sizes determined via PS and QLS differ significantly for the emulsions containing added LCT. For PS, >50% of the distribution is located within the last bar of the histogram (cutoff ) 0.08 µm). Hence, the median size is presumably 10% v/v free FC on presterile vials). Discussion Dispersions of FC/LCT/EYP contain up to three types of particles, which differ dramatically in their physicochemical characteristics, including size, density, and refractive index (Table 3). Hence, a sizing technique that first fractionates the different types of particles is a prerequisite for understanding these complex dispersions. In the present instance, SdFFF is able to separate the three types of particles and size them independently. In contrast, ensemble methods (e.g., QLS) typically fail, primarily because the complex algorithms used to analyze the data do not lead to a unique analytical solution. Also, the models must assume information about the particle size distribution and often lose information about small particles that do not scatter light as strongly as larger particles. In the present case, the FC emulsion droplets and SUVs do not scatter light nearly as well as the LCT emulsion droplets, as the differences in refractive index between the particles and the medium are small. Hence, these particles are invisible compared to the strongly scattering LCT emulsion droplets when sized by QLS. LCT emulsion droplets also dominate the apparent particle size distributions obtained by PS (optical detection). The FC droplets are observed as a discrete population only at low LCT contents. In the PS measurements, the SBO droplets appear to have a very small size due to the fact that the ∆F value of the FC component was used to calculate particle size (Stokes’ equation). Different run conditions would be needed to accurately size the LCT emulsion droplets via PS. Previous studies utilizing QLS to size PFD/LCT/EYP emulsions have come to the erroneous conclusion that the addition of LCT stabilizes the PFD particles to droplet coarsening via the formation of a three-phase emulsion.15-18 On the basis of the SdFFF results, it is clear that a threephase emulsion is not formed in these emulsions and that the two oils demix in a manner similar to mixtures of
fluorinated and hydrogenated surfactants.30 Also, no differences in coarsening rates for the PFD particles were observed with or without added LCT. The ensemble QLS method is simply detecting the slow coarsening of the LCT emulsion droplets. Whereas the FC emulsion droplets coarsen over time via Ostwald ripening processes, the LCT droplets are stable, due to the low water solubility of LCT oils, which slows molecular diffusion processes.31 The addition of LCT also has a profound effect on the visual appearance of the emulsion. Control emulsions containing PFD alone appear brown upon storage due to phospholipid oxidation. Addition of LCT effectively masks the brown color and yields a more aesthetically pleasing white colored emulsion. Although the emulsion appearance is improved, the physical stability of the formulation is not. A single population of emulsion droplets is observed, however, when MCT oils are used instead of LCTs. The amount of TG that can be dissolved into the EYP layer is dependent on the chain length and degree of unsaturation of the fatty acid chains on the triglyceride and on the phospholipid composition. MCT oils are typically 4-5 times more soluble in the same phospholipid layer than LCTs.32 It has been hypothesized that improved affinity between the surfactant and the dispersed phase will result in improvements in emulsion stability to Ostwald ripening by either reducing the interfacial tension or limiting FC diffusion through the interfacial layer.33 In practice, the result is much different. The addition of MCT oil destabilizes the emulsion with respect to droplet coalescence. This is reflected by decreases in stability of the emulsion to increases in mechanical stress and by complete phase separation at elevated temperatures during terminal sterilization. Coalescence is the merging of two droplets into one. The static and dynamic properties of the stabilizing surfactant monolayer have been found to be critically important not only in the initial stages of droplet formation but also during long-term storage.34-36 A correspondence between the phase behavior of oil/water/surfactant ternary mixtures and macroemulsion stability has been noted.34,37,38 In general, surfactants that possess large bulky headgroups and small alkyl tails tend to stabilize oil-in-water emulsions, while surfactants with small headgroups and multiple alkyl chains stabilize water-in-oil emulsions. This was first observed by Harkins et al. in 1917 and was termed the “oriented wedge theory”.38 Significant oil penetration into the surfactant tails tends to favor formation of waterin-oil emulsions or, alternatively, the destabilization of oil-in-water emulsions. Despite the excellent correlation between the phase behavior of oil/water/surfactant mixtures and emulsion stability, the oriented wedge theory (29) Weers, J. G.; Liu, J.; Fields, T.; Resch, P.; Clavin, J.; Arlauskas, R. A. Artif. Cells, Blood Substitutes, Immobilization Biotechnol. 1994, 22, 1175. (30) Funasaki, M. In Mixed Surfactant Systems; Ogino, K., Abe, M., Eds.; Surfactant Sci. Ser. 1993, 46, 145. (31) Arlauskas, R. A.; Weers, J. G. Langmuir 1996, 12, 1923. (32) Adolph, M. In Emulsions and Nanosuspensions for the Formulation of Poorly Soluble Drugs; Muller, R. H., Benita, S., Bohm, B., Eds.; MedPharm: Stuttgart, Germany, 1998; p 119. (33) Riess, J. G.; Arlen, C.; Greiner, J.; LeBlanc, M.; Manfredi, A.; Pace, S.; Varescon, C.; Zarif, L. In Blood Substitutes; Chang, T. M. S., Geyer, R. P., Eds.; Dekker: New York, 1989; p 421. (34) Kabalnov, A.; Wennerstro¨m, H. Langmuir 1996, 12, 276. (35) Kabalnov, A.; Tarara, T.; Arlauskas, R.; Weers, J. J. Colloid Interface Sci. 1996, 184, 227. (36) Kabalnov, A. S.; Weers, J. G. Langmuir 1996, 12, 1931. (37) Israelachvili, J. Colloids Surf., A 1994, 91, 1. (38) Harkins, W. D.; Davies, E. C. H.; Clark, G. L. J. Am. Chem. Soc. 1917, 39, 541.
Characterization of Multiparticulate Dispersions
was dismissed for nearly a century, due to the fact that the surface of an emulsion droplet is flat when viewed on the scale of an emulsifier molecule.39 The link between phase behavior and emulsion stability was solved by Kabalnov and Wennerstro¨m, who hypothesized that it was not the curvature of the droplet that was important but rather the curvature at a nucleation hole! Kabalnov and Wennerstro¨m related the stability of emulsions to the bending properties of the surfactant monolayer. In particular, surfactant monolayers that have positive values of the monolayer spontaneous curvature (H0) stabilize oil-in-water emulsions, while systems with negative values of the spontaneous curvature favor waterin-oil emulsions. Phospholipids are nearly balanced surfactants (H0 ≈ 0).35 As a result, small changes in spontaneous curvature can lead to dramatic changes in emulsion stability. Applied to the current problem, penetration of MCT into the EYP monolayer would be expected to decrease the monolayer spontaneous curvature, favoring inversion to a water-in-oil emulsion and leading to destabilization of the emulsion with respect to droplet coalescence, ultimately leading to phase separation. This is enhanced at elevated temperatures due to increased disorder (increased gauche conformer content) in the lipid acyl chains, which also provides an increase in tail volume, or decrease in H0. In a similar fashion, introduction of 10% w/w cholesterol to 90% w/v PFOB emulsions results in a single population of emulsion droplets, which break upon sterilization. In contrast, cholesterol esters form a distinct population of emulsion droplets, similar to LCTs. Monoglycerides and diglycerides are also soluble in the phospholipid acyl chains, leading to destabilization of FC emulsions with respect to coalescence-mediated processes (data not shown). The results are also in agreement with the results of Shchukin and colleagues who demonstrated that increased affinity between the dispersed phase and the emulsifier leads to enhanced droplet coalescence.40 Johnson et al.41,42 have shown that addition of LCT may aid in the long-term stability of PFD emulsions when the stabilizing surfactant is Pluronic F-68. The enhanced stability is thought to be due to an increase in the cloud point temperature of the polymer due to the presence of small concentrations of SBO at the PFD/water interface. (39) Shinoda, K.; Friberg, S. Emulsions and Solubilization; John Wiley & Sons: New York, 1986; p 174. (40) Shchukin, E. D.; Amelina, E. A.; Parfenova, A. M. Colloids Surf., A 2001, 176, 35. (41) Johnson, O. L.; Washington, C.; Davis, S. S. Int. J. Pharm. 1990, 59, 131. (42) Johnson, O. L.; Washington, C.; Davis, S. S. Int. J. Pharm. 1990, 63, 65.
Langmuir, Vol. 20, No. 18, 2004 7435
In the absence of SBO, the cloud point of Pluronic F-68 is 110 °C. At typical sterilization temperatures (i.e., 121 °C), the emulsion droplets coalesce and the components phase separate. The addition of 2% SBO raises the cloud point to 128 °C, making terminal sterilization possible. Emulsions containing high concentrations of Pluronic F-68 are typically sterile-filtered to avoid cloud point issues. We hypothesize that minor components in the LCT are responsible for the increases in cloud point. The addition of small amounts (∼1% or less) of charged surfactants (e.g., fatty acid soaps) have been shown to have a profound effect on the cloud point of Pluronic surfactants.43 SBO is typically a complex mixture of materials. These include minor components (e.g., fatty acids, monoglycerides, diglycerides, sterols, sterol esters, isoprenoids, and tocopherols) that are surface active and hence more likely to impact the cloud point of the nonionic poloxamer. Conclusions We studied the complex dispersions of particles formed in emulsions containing two immiscible dispersed phases, FC and TG. The following conclusions can be made. (1) The addition of LCT to FC emulsions stabilized by PL results in three types of particles: FC emulsion droplets, LCT emulsion droplets, and SUVs. (2) The addition of LCT has no impact on the coarsening of FC emulsion droplets by either Ostwald ripening or droplet coalescence. (3) The addition of MCT to FC emulsions stabilized by PL leads to mixing of the MCT within the PL acyl chains (i.e., formation of a three-phase emulsion). (4) Formation of a three-phase emulsion leads to a negative increment in the spontaneous curvature of the PL monolayer, thereby destabilizing the FC emulsion with respect to coalescence-mediated growth and ultimately to emulsion breakage during terminal sterilization. (5) The concept of improving emulsion stability by improving the adhesion between the oil and the emulsifier is inconsistent with current models regarding the impact of curvature on emulsion stability to droplet coalescence. Acknowledgment. The authors wish to thank Jean Riess, Alexei Kabalnov, and David Klein for many stimulating discussions on this topic. We would also like to thank Ms. Jin Liu for characterization of the perfluorodecalin emulsions. LA049375E (43) Flore, S. G.; Dellamary, L. A.; Reeve, L. E.; Weers, J. G. U.S Patent 6,280,745 B1, 2001.
