- Email: [email protected]
Development of Inherently Echogenic Liposomes as an Ultrasonic Contrast Agent†
Development of Inherently Echogenic Liposomes as an Ultrasonic Contrast Agent† HAYAT ALKAN-ONYUKSEL‡§X, SASHA M. DEMOS§, GREGORY M. LANZA∇, MICHAEL J. VONESH∇, MELVIN E. KLEGERMAN‡, BONNIE J. KANE∇, JER KUSZAK⊥, AND DAVID D. MCPHERSON∇ Received September 29, 1995, from the ‡Department of Pharmaceutics and Pharmacodynamics and §Bioengineering Program, University of Illinois at Chicago, Chicago, IL 60612, ∇Division of Cardiology, Department of Medicine, Northwestern University Medical School, Chicago, IL 60611, and ⊥Department of Pathology, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, IL 60612. Final revised x manuscript received January 8, 1996. Accepted for publication March 5, 1996X. Present address for correspondence: UIC, College of Pharmacy, 833 South Wood Street M/C 865, Chicago, IL 60612. Abstract 0 Ultrasonic contrast agents have been developed for improved assessment of blood flow and tissue perfusion. Many of these agents are not inherently acoustically reflective (echogenic), and nearly all are not suitable for tissue specific targeting. The purpose of this study was to develop acoustically reflective liposomes, which are suitable for antibody conjugation, without using gas or any other agent entrapment. Echogenic liposomes were prepared from phosphatidylcholine (PC), phophatidylethanolamine (PE), phosphatidylglycerol (PG), and cholesterol (CH), using a dehydration/rehydration method. The formulation was optimized for higher acoustic reflectivity by varying the lipid composition. Liposomes were imaged with a 20 MHz intravascular ultrasonic imaging catheter. Echogenicity levels were expressed using pixel gray scale. The presence of PE and PG at specific concentrations improved echogenicity due to their effects on liposomal morphology as confirmed by freeze-etch electron microscopy. The acoustic reflectivity of liposomes was retained when liposomes were treated with blood at room temperature and 37 °C under in vitro conditions. It was demonstrated that the liposomes were also acoustically reflective in vivo after they were injected into a miniswine model. We have developed echogenic liposomes that are stable and suitable for tissue specific targeting as a novel contrast agent. This new contrast agent can be used for ultrasonic image enhancement and/or treatment of targeted pathologic sites.
Introduction Ultrasound imaging provides structural and functional information including features of tissue and flow characteristics. The use of ultrasound for pathologic tissue characterization has been challenging. The advent of contrast agents for ultrasonic diagnosis has provided a substantial leap in the ability of ultrasound to determine structural morphology. These agents alter or enhance the returning ultrasound signal and provide improved surface or tissue resolution. The physical principles and clinical applications of ultrasound contrast agents have been reviewed.1,2 The action of the contrast agents may be due to three different effects: backscatter, sound attenuation, and changes in sound speed, with most promising agents relying entirely upon increased backscatter. The most useful ultrasound contrast agents include entrapped gas solutions. The first contrast agents to produce entrapped gas used manually agitated solutions which included saline, 5% dextrose, hydrogen peroxide, and even blood.3,4 The bubbles formed using these solutions had variable large size and short half-life. Application of sonica† Supported in part by the National Institutes of Health (HL-46550) and the Feinberg Cardiovascular Research Institute, Northwestern University Medical School. X Abstract published in Advance ACS Abstracts, April 15, 1996.
486 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
tion instead of hand shaking improved the size and stability somewhat but not with consistent clinical utility.5 Presently, several commercially produced contrast agents are undergoing evaluation in clinical and animal laboratory settings. Albunex is a promising ultrasound contrast agent consisting of a 5% air-filled albumin microsphere suspension. Its mean diameter is 4 µm, and evaluation of its safety and efficacy has demonstrated that it is safe and feasible for assessment of myocardial perfusion.6 Two other contrast agents, SHU 508A and SHU 454, are made of galactose microparticles which produce microbubbles when mixed with diluent. The mean size of SHU 454 (Echovist) microbubbles is 3 µm with a relatively large range, whereas SHU 508A (Levovist) microbubbles range from 2 to 8 µm. The size of these agents is important if venous injection and pulmonary transit to the arterial circulation are being considered. No adverse hemodynamic effects were observed following intravenous injections of both agents.7 Perfluorochemicals (Imagent) as colloidal suspensions are also being tested to enhance the ultrasound backscatter in tissue and also Doppler signals.8 The reflectivity of perfluorocarbons is felt to be due to their high density (1.9 g/mL) and low acoustic velocity (600 m/s) resulting in an acoustic impedance difference of 30% between it and the adjacent tissues.1 Perhaps the most promising agent developed so far is Aerosomes which are gas-filled liposomes.9,10 They are used as a contrast agent to evaluate the cardiovascular system and for tumor therapy. A drawback however is their large size above the acceptable limit for vascular infusion. A newly developed ultrasound contrast agent, EchoGen, can undergo a phase change from an emulsion containing a water-immiscible liquid to echogenic gas bubbles upon injection into the blood stream.11 Although all these agents have been used with varying degrees of success, there are still some associated difficulties. One common problem for gas-containing agents is the control of size and stability of gas bubbles after injection. The encapsulated systems, liposomes and microparticles, offer significant improvement, but the in vivo and in vitro stability of these two-phase systems is not yet well determined. Particles greater than 4-5 µm, when injected intravenously, are trapped in pulmonary capillaries and can not pass from the right side to the left side of the heart. Due to reticuloendothelial system (RES) uptake, the half-life of the microspheres when injected into blood has been shown to be very short (<1 min).1 Furthermore, all the contrast agents developed to date are not suitable for targeting by antibody conjugation. A contrast agent that can overcome these problems and be targeted to specific tissue components is attractive because (i) a locally high concentration of the agent will provide better imaging, (ii) the problem of short half-life and RES uptake of traditional contrast agents will be solved since the agent will be attached to the target tissues before reaching the RES organs, hence in vivo stability will be
0022-3549/96/3185-0486$12.00/0
© 1996, American Chemical Society and American Pharmaceutical Association
Table 1sEffect of Composition on Liposome Echogenicity Sample
Composition (PC:PE:PG:CH)
Mean Gray Scale
SD
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
65.0:0:5:30 60.0:0:10:30 67.5:0.5:2:30 65.0:1:4:30 62.5:1.5:6:30 60.0:2:8:30 67.0:1:2:30 64.0:2:4:30 61.0:3:6:30 58.0:4:8:30 66.0:2:2:30 62.0:4:4:30 58.0:6:6:30 54.0:8:8:30 64.0:4:2:30 61.0:6:3:30 58.0:8:4:30 55.0:10:5:30 52.0:12:6:30 65.0:4:1:30 60.0:8:2:30 55.0:12:3:30 50.0:16:4:30 65.0:5:0:30 60.0:10:0:30
97.63 90.59 94.52 97.26 96.28 70.53 92.55 73.69 79.70 73.09 86.76 82.97 83.75 78.86 64.27 69.40 63.97 75.50 82.55 120.08 106.12 115.03 97.01 86.95 94.81
8.01 9.09‘ 14.12 13.91 4.39 2.36 2.85 9.73 13.90 10.38 22.01 5.76 23.64 5.33 2.81 5.72 6.80 13.92 7.69 5.99 16.09 20.47 2.15 13.06 7.47
increased, and (iii) lower doses of the contrast agent will be required, decreasing the cost and risk of toxicity. Tissue specific targeting may have another important advantage with cardiovascular ultrasound. Active plaque morphology can be identified by the contrast agent conjugated to antibodies specific to cardiac tissues. For example, targeting with antibodies such as VCAM-1 and ICAM-1 would demonstrate early stages of atheroma development,12 whereas targeting with antifibrin-attached liposomes would delineate areas of thrombosis and plaque rupture. The purpose of this study was to develop an acoustic contrast agent that (i) is inherently echogenic and stable, (ii) has a mean size lower than 1 µm, eliminating pulmonary entrapment, and (iii) can be conjugated to antibodies for specific tissue targeting. This report describes the development of inherently echogenic liposomes composed of phosphatidylcholine, phosphatidylglycerol, phosphatidylethanolamine, and cholesterol. The compositions that provide high echogenicity are identified by varying the components in a systematic manner. The liposomes demonstrating highest acoustic reflectivity are tested for the retention of their echogenic properties under simulated in vivo conditions and in vivo in an animal model.
Experimental Section MaterialssL-R-Phosphatidylcholine type XIII-E (PC), L-R-phosphatidylethanolamine dipalmitoyl (PE), and cholesterol (CH) were obtained from Sigma Chemical Co., St. Louis, MO. 1,2-Dipalmitoylsn-glycero-3-phosphoglycerol was obtained from Fine Chemicals, Liestal, Switzerland. MethodssPreparation of LiposomessPhospholipids and cholesterol were dissolved in chloroform and combined in amounts to achieve the desired lipid composition as shown in Table 1 . The solvent was evaporated by a rotary evaporator (Labconco, Kansas City, MO), rotated at 120 rpm, and immersed in a thermostated water bath, with the temperature at 50 °C under argon. The resulting lipid film was then placed in a dessicator under vacuum for 2 days for complete drying. The dry lipid film was rehydrated with deionized water to yield a concentration of 10 mg of lipid/mL of water. After removing the film from the flask walls, the dispersion was sonicated in a 175.5 W water bath sonicator (Fisher Scientific, Itasca, IL) until the mean
size of the liposomes was ca. 500 nm. Liposome size was determined by quasielastic light scatter as described below. D-Mannitol (0.2 M) was added to the liposome suspension in a 1:1 (v/v) ratio, and this suspension was frozen overnight at -70 °C. The frozen samples were placed in a lyophilizer (Labconco, Kansas City, MO) for 48 h. The dried lipids were then resuspended with 0.1 M phosphate buffer, pH 7.4, to yield a concentration of 10 mg of lipid/mL of buffer. After hand shaking and brief vortexing, the size of the resulting liposomes was again determined. The mean size of the liposomes remained under 1 µm in all cases and in most cases was <800 nm. Size MeasurementssThe mean size of the liposomes was determined by quasielastic light scattering (QELS) measurements using a Nicomp model 270 submicron particle sizer (Pacific Scientific, Menlo Park, CA) equipped with a 5 mW helium-neon laser at an exciting wavelength of 632.8 nm, a 64-channel autocorrelation function, a temperature-controlled scattering cell holder, and an ADM 11 video display terminal computer (Lear Siegler Inc., Anaheim, CA) for analyzing the fluctuations in scattered light intensity generated by the diffusion of particles in solution. The mean hydrodynamic particle diameter, dh, was obtained from the Stokes-Einstein relation using the measured diffusion coefficient obtained from the fit. ImagingsLiposomes were transferred to liquid scintillation vials and imaged with a 20 MHz high-frequency intravascular ultrasound (IVUS) imaging catheter (Boston Scientific Inc., Sunnyvale, CA). The IVUS catheter was passed through the vial cap and secured. Instrument settings for gain, zoom, compression and rejection levels were optimized at the initiation of the experiment and held constant for all samples. Images were recorded onto 1/2” VHS videotape in real time for subsequent playback and image analysis. Videodensitometric Analysis of Liposome “Brightness”sRelative echogenicity (apparent brightness) of all liposome formulations was objectively assessed via computer-assisted videodensitometry. This process involved image acquisition, preprocessing, automated liposome identification, and gray scale quantification. All image processing and analysis were performed with Image Pro Plus Software (Version 1.0, Media Cybernetics, Silver Springs, MD) running on a dedicated computer (486 CPU, 66 MHz). Randomly selected IVUS images were acquired from videotape for each liposome formulation. Images were digitized to 640 × 480 pixel spatial resolution (ca. 0.045 mm/pixel) and 8 bit (256 levels) amplitude resolution. All analyzed IVUS data were collected at a fixed instrument gain level. The distribution of gray scale values within the image was then adjusted to cover the entire range of possible gray levels using a linear transformation algorithm (i.e., dynamic range was maximized). Image brightness was subsequently scaled such that a reference feature, the annotation text, common to each image, retained a constant gray scale value over all images. An automated liposome detection routine was then run to identify liposomes suspended in solution within an annular region of interest set at a constant radial distance from the imaging catheter. The automated liposome detection routine identified all “bright” objects within the analysis annulus having a gray scale level >29, a roundness ratio (i.e., ratio of maximum diameter:minimum diameter) <2.5, and a size >4 pixels. This procedure excluded virtually all imaging artifact from our detection algorithm. Thus, objects identified were considered to be “liposomes”. Each liposome was outlined and numbered by the computer program. The average gray scale and size of each liposome were also quantified. The mean gray scale value of all pixels identified as “liposomes” with a given image was then computed and used to characterize the echogenicity of a given liposome formulation. Freeze-Etch Electron MicroscopysLiposomes were prepared for freeze fracture according to standard techniques as reported previously.13,14 Briefly, drops of the liposome suspension were frozen in liquid-nitrogen-cooled Freon 22. The samples were fractured using a Balzers BAF 301 freeze-etch unit at -115 °C. The fractured samples were coated with platinum and carbon. The replicas were cleansed in a minimum of two changes of bleach (sodium hypochlorite), washed with distilled water, dried, and collected on 200 mesh copper grids. Replicas were examined and photographed with a JEOL 100CX transmission electron microscope at 80 kV. Liposomes in Simulated Physiological ConditionssEchogenic liposomes consisting of phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol (PG), and cholesterol in a molar ratio of 60:8: 2:30 were prepared by the previously described method. The liposome suspension in buffer was imaged at room temperature. One half of the suspension was warmed in a water bath to a temperature of 37
Journal of Pharmaceutical Sciences / 487 Vol. 85, No. 5, May 1996
Figure 1sEffect of phosphatidylethanolamine and phosphatidylglycerol content on the echogenicity of liposomes. °C and then imaged. The other half of the suspension was combined with canine blood in a 1:1 (v/v) ratio and then imaged at room temperature. The suspension temperature was raised to 37 °C and the suspension imaged again. Canine blood mixed with saline in a 1:1 (v/v) ratio was used as a control. In Vivo EvaluationsA standard Yucatan miniswine model of induced atherosclerosis was developed to test in vivo liposomal acoustic properties. Following development of atherosclerosis (1-3 weeks postinitial surgery), the animal was anesthetized and artificially ventilated. Sheaths were inserted into the femoral and carotid arteries for vascular access and imaging catheter insertion. Under stable hemodynamic conditions, the imaging catheters were inserted into the arteries. Base-line imaging of the artery and luminal blood was performed, and images were recorded onto videotape. Following base-line imaging, 2-5 cc of liposomes was injected through the catheter, and similar imaging occurred. Comparisons between gray scale characteristics of in vivo base-line blood and liposomes in blood were made. Data were compared by ANOVA and determined to be statistically different at p < 0.05.
Results Composition StudiessThe effect of lipid composition on the acoustic properties of the liposomes prepared by the same method and with similar mean size is shown in Table 1 and Figure 1. It is clear that with relatively high concentrations of PE (4-12%), low concentrations of PG (1-3%), and constant cholesterol content (30%), highest echogenicity levels are obtained. The structural properties of a highly echogenic and poorly echogenic liposomal formulation (samples 21 and 17 in Table 1) are demonstrated by freeze-etch electron microscopy in Figure 2. The high echogenic formulation (2, top) consists of a multilamellar structure with well-separated bilayers, whereas low echogenicity is associated with a thick unseparated multilayer for a single liposome (2, bottom). Figure 3, top and middle, shows ultrasound images of liposomal suspensions with formulations corresponding to those in Figure 2. Figure 3, bottom, demonstrates an ultrasonic image of 0.1 M phosphate buffer alone. Buffer, like water, is not acoustically reflective and therefore has mean gray scale values of zero. Stability StudiessAfter optimizing the lipid composition of the liposomes to give high acoustic reflectivity, the next step was to determine if this echogenicity would be retained under in vivo conditions. Figure 4 demonstrates the echogenic levels, as mean gray scale values, obtained when liposomes with high acoustic reflectivity were exposed to simulated in vivo conditions, i.e., 37 °C and presence of blood components. Neither the increase in temperature nor the presence of blood 488 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
Figure 2sFreeze fracture micrographs showing the structure of liposomes (12000×, reproduced at 50% of original size): (top) liposomes with high echogenicity composed of a 60:8:2:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer) and (bottom) liposomes with low echogenicity composed of a 58:8:4:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer).
caused a decrease in acoustic reflectivity of the liposomes. For comparison purposes, the echogenicity of blood alone is also shown. The retention of echogenicity of the liposomes was also demonstrated in actual in vivo conditions as seen in Figure 5. Mean gray scale values for liposomes that were injected into the arterial bed of a Yucatan miniswine model were similar to those obtained under in vitro conditions. Factors Influencing ResultssThe resolution of our 2D images is ca. 30-45 µm. Therefore, although smaller particles may have acoustic reflectivity, they will be displayed at this pixel size; whether the reflectivity is due to single acoustically reflective liposomes or aggregates of such remains to be determined. Additionally, particles within the solution may have added to the specular reflectors of our formulation. However, when measuring the gray scale values of the different liposomal formulations, the only components of the suspensions being analyzed were liposomes and buffer. We have shown that buffer is not acoustically reflective (Figure 3, bottom).
Discussion Effect of Lipid Composition on Echogenicitys Liposomes are vesicles composed of phospholipid bilayers surrounding an aqueous space. They have been shown to be safe for injection and suitable for antibody conjugation.15,16 Liposomes also have the advantage that they can easily be formulated below 1 µm in diameter which is desirable for safe
Figure 4sEffect of simulated physiological conditions of the echogenicity of liposomes: (A) liposomes composed of a 60:8:2:30 molar mixture of PC:PE:PG: CH in 0.1 M phosphate buffer (pH 7.4) at room temperature, (B) liposomes of the same composition in 0.1 M phosphate buffer (pH 7.4) warmed to 37 °C, (C) liposomes of the same composition in canine blood at room temperature, (D) liposomes of the same composition in canine blood warmed to 37 °C, and (E) blood at 25 °C.
Figure 5sIn vivo liposome echogenicity. Liposomes composed of a 60:8:2:30 molar mixture of PC:PE:PG:CH in 0.1 M phosphate buffer (pH 7.4) at a concentration of 10 mg of lipid/mL of buffer were injected into the arteries of a Yucatan miniswine. The echogenicity of the liposomes in vivo was compared to that of the blood components. Figure 3sUltrasound images of liposomes: (top) intravascular ultrasound catheter in 10 mL glass vial containing liposomes with high echogenicity composed of a 60:8:2:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer), (middle) intravascular ultrasound catheter in 10 mL glass vial containing liposomes with low echogenicity composed of a 58:8:4:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer), and (bottom) intravascular ultrasound catheter in 10 mL glass vial containing 0.1 M phosphate buffer.
pulmonary capillary passage. Gas-filled liposomes have previously been reported to be highly acoustically reflective.17,18 However, trapping gas in liposomes not only requires a sophisticated preparation process but also may cause instability of the system under both in vivo and in vitro conditions. Our study demonstrates that acoustically reflective liposomes can be produced solely by choosing the optimal lipid composition and applying a dehydration/rehydration process. Four different lipids were used in the liposome formulation. Phosphatidylcholine, being the most common
phospholipid in the cell membrane, was chosen as the main lipid. Cholesterol was added to the formulation to add rigidity to the lipid bilayer. Our preliminary data had shown an increase in acoustic reflectivity in the presence of cholesterol.19 Phosphatidylethanolamine and phosphatidylglycerol were included in the liposomal formulation to allow for conjugation and add negative charge, respectively. Separation of the phospholipid bilayers within a liposome appears to affect its acoustic reflectivity. In the presence of PE and PG at optimized concentrations for highest echogenicity, we obtained oligolamellar structures, represented by vesicles within vesicles, whereas for liposomes with low acoustic properties the structure of liposomes consisted of unilamellar or unseparated multilamellar bilayers (Figure 2). This finding is in close agreement with previous data evaluating air-entrapped liposomes, where it was demonstrated that oligolamellar liposomes had higher acoustic reflectivity than unilamellar
Journal of Pharmaceutical Sciences / 489 Vol. 85, No. 5, May 1996
liposomes.18 We believe that PG, because of its negative charge, causes separation of the bilayers as well as prevents the aggregation of liposomes. Phosphatidylethanolamine on the other hand, with its small head group, may affect the curvature of the bilayers and formation of vesicles within vesicles. However, there is an optimal composition where all these effects maximize to give the best structural changes in the liposome for the highest echogenicity. Retention of Echogenicity under in Vivo ConditionssIn order to develop a successful contrast agent, the acoustic reflectivity of the liposomes demonstrated under in vitro conditions should be retained in vivo. First we tested them under simulated in vivo conditions by introducing them into media at body temperature and in the presence and absence of blood components. Temperature elevations up to 37 °C did not show any significant change in echogenicity. This is most likely due to the fact that the gel state of the lipid bilayers which were formulated at room temperature in the presence of cholesterol was still maintained at 37 °C. Similarly, due to the rigidity of the bilayers, the blood components did not interfere with the phospholipid bilayers causing structural changes within the liposomes. Previous studies have shown the improvement of in vivo stability of phosphatidylcholine liposomes in the presence of cholesterol.20 Since the simulated in vivo conditions did not fully represent the real in vivo conditions, such as the presence of certain ions and enzymes, the acoustic behavior of the liposomes was also tested after injection into the arteries of the Yucatan miniswine model. It was confirmed that the echogenic liposomes developed in this study were stable to all the factors originating from in vivo conditions.
Conclusions Liposomes that are acoustically reflective without inclusion of gas or other agents can be prepared by choosing an optimal lipid composition and applying a dehydration/rehydration method. The echogenicity of these liposomes is not lost or decreased at body temperature or in the presence of blood components. The novel contrast agent developed in this study
490 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
can be used de novo or following conjugation to tissue specific antibodies for local image enhancement, diagnosis, and drug delivery.
References and Notes 1. Goldberg, B. B.; Liu, J. B.; Forsberg, F, Ultrasound Med. Biol. 1994, 20, 319-333. 2. Winkelmann, J. W.; Kenner, M. D.; Dave, R.; Chandwaney, R. H.; Feinstein, S. B. Ultrasound Med. Biol., 1994, 20, 507-515. 3. Feigenbaum, H.; Stone, J. M.; Lee, D. A. Circulation 1970, 41, 613-620. 4. Ophir, J.; Parker, K. J. Ultrasound Med. Biol. 1989, 15, 319325. 5. Feinstein, S. B.; Keller, M. W.; Kerber, R. E. J. Am. Coll. Cardiol. 1989, 2, 125-131. 6. Feinstien, S. B.; Cheirif, J. J. Am. Coll. Cardiol. 1990, 16, 316324. 7. Smith, M. D.; Elion, J. L.; McClure, R. R.; Kwan, O. L.; DeMaria, A. N. J. Am. Coll. Cardiol. 1989, 12, 622-1628. 8. Andre, M. P.; Steinbach, G.; Mattrey, R. F. Invest. Radiol. 1993, 28, 502-506. 9. Unger, E.; Shen, D.; Fritz, T.; Kulik, B.; Lund, P.; Wu, G.; Yellowhair, D.; Ramaswami, R.; Matsunaga, T. Invest. Radiol. 1994, 29, S134-S136. 10. Simon, R. H.; Ho, Sy.; Lang, S. C.; Uphoff, D. G.; D’Arrigo, J. S. Ultrasound Med. Biol. 1993, 19, 123-125. 11. Quay, S. C. J. Ultrasound Med. 1994, 13, S9. 12. Cybulsky, M. I.; Gimbrone, M. A. Science 1991, 251, 788-791. 13. Moor, H.; Muhlethaler, K. J. Cell. Biol. 1963, 17, 609-628. 14. Kuszak, J. R.; Alcala, J.; Maisel, H. Exp. Eye Res. 1980, 33, 157166. 15. Alving, C. R. In Liposomes: From Biophysics to Therapeutics; Ostro, M. J., Ed.; Marcel Dekker, Inc.: New York, 1987; pp 202204. 16. Maruyama, K.; Takizawa, T.; Yuda, T.; Kennel, S. J.; Huang, L.; Iwatsuru, M. Biochim. Biophys. Acta Biomembr. 1995, 1234, 74-80. 17. Ryan, J. R.; Davis, A.; Melchior, D. L. U.S. Patent 4,900,540, 1990. 18. Unger, E. C. U.S. Patent 5,123,414, 1992. 19. Lanza, G. M.; Alkan, M. H.; Vonesh, M. J.; Kleggerman, M. E.; Frazin, L. J.; Mehlman, D. J.; Talano, J. V.; McPherson, D. D. J. Am. Coll. Cardiol. 1992, 19, 114A. 20. Senior, J. H. Crit. Rev. Ther. Drug Carrier Syst. 1987, 3, 123.
JS950407F
Introduction Ultrasound imaging provides structural and functional information including features of tissue and flow characteristics. The use of ultrasound for pathologic tissue characterization has been challenging. The advent of contrast agents for ultrasonic diagnosis has provided a substantial leap in the ability of ultrasound to determine structural morphology. These agents alter or enhance the returning ultrasound signal and provide improved surface or tissue resolution. The physical principles and clinical applications of ultrasound contrast agents have been reviewed.1,2 The action of the contrast agents may be due to three different effects: backscatter, sound attenuation, and changes in sound speed, with most promising agents relying entirely upon increased backscatter. The most useful ultrasound contrast agents include entrapped gas solutions. The first contrast agents to produce entrapped gas used manually agitated solutions which included saline, 5% dextrose, hydrogen peroxide, and even blood.3,4 The bubbles formed using these solutions had variable large size and short half-life. Application of sonica† Supported in part by the National Institutes of Health (HL-46550) and the Feinberg Cardiovascular Research Institute, Northwestern University Medical School. X Abstract published in Advance ACS Abstracts, April 15, 1996.
486 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
tion instead of hand shaking improved the size and stability somewhat but not with consistent clinical utility.5 Presently, several commercially produced contrast agents are undergoing evaluation in clinical and animal laboratory settings. Albunex is a promising ultrasound contrast agent consisting of a 5% air-filled albumin microsphere suspension. Its mean diameter is 4 µm, and evaluation of its safety and efficacy has demonstrated that it is safe and feasible for assessment of myocardial perfusion.6 Two other contrast agents, SHU 508A and SHU 454, are made of galactose microparticles which produce microbubbles when mixed with diluent. The mean size of SHU 454 (Echovist) microbubbles is 3 µm with a relatively large range, whereas SHU 508A (Levovist) microbubbles range from 2 to 8 µm. The size of these agents is important if venous injection and pulmonary transit to the arterial circulation are being considered. No adverse hemodynamic effects were observed following intravenous injections of both agents.7 Perfluorochemicals (Imagent) as colloidal suspensions are also being tested to enhance the ultrasound backscatter in tissue and also Doppler signals.8 The reflectivity of perfluorocarbons is felt to be due to their high density (1.9 g/mL) and low acoustic velocity (600 m/s) resulting in an acoustic impedance difference of 30% between it and the adjacent tissues.1 Perhaps the most promising agent developed so far is Aerosomes which are gas-filled liposomes.9,10 They are used as a contrast agent to evaluate the cardiovascular system and for tumor therapy. A drawback however is their large size above the acceptable limit for vascular infusion. A newly developed ultrasound contrast agent, EchoGen, can undergo a phase change from an emulsion containing a water-immiscible liquid to echogenic gas bubbles upon injection into the blood stream.11 Although all these agents have been used with varying degrees of success, there are still some associated difficulties. One common problem for gas-containing agents is the control of size and stability of gas bubbles after injection. The encapsulated systems, liposomes and microparticles, offer significant improvement, but the in vivo and in vitro stability of these two-phase systems is not yet well determined. Particles greater than 4-5 µm, when injected intravenously, are trapped in pulmonary capillaries and can not pass from the right side to the left side of the heart. Due to reticuloendothelial system (RES) uptake, the half-life of the microspheres when injected into blood has been shown to be very short (<1 min).1 Furthermore, all the contrast agents developed to date are not suitable for targeting by antibody conjugation. A contrast agent that can overcome these problems and be targeted to specific tissue components is attractive because (i) a locally high concentration of the agent will provide better imaging, (ii) the problem of short half-life and RES uptake of traditional contrast agents will be solved since the agent will be attached to the target tissues before reaching the RES organs, hence in vivo stability will be
0022-3549/96/3185-0486$12.00/0
© 1996, American Chemical Society and American Pharmaceutical Association
Table 1sEffect of Composition on Liposome Echogenicity Sample
Composition (PC:PE:PG:CH)
Mean Gray Scale
SD
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
65.0:0:5:30 60.0:0:10:30 67.5:0.5:2:30 65.0:1:4:30 62.5:1.5:6:30 60.0:2:8:30 67.0:1:2:30 64.0:2:4:30 61.0:3:6:30 58.0:4:8:30 66.0:2:2:30 62.0:4:4:30 58.0:6:6:30 54.0:8:8:30 64.0:4:2:30 61.0:6:3:30 58.0:8:4:30 55.0:10:5:30 52.0:12:6:30 65.0:4:1:30 60.0:8:2:30 55.0:12:3:30 50.0:16:4:30 65.0:5:0:30 60.0:10:0:30
97.63 90.59 94.52 97.26 96.28 70.53 92.55 73.69 79.70 73.09 86.76 82.97 83.75 78.86 64.27 69.40 63.97 75.50 82.55 120.08 106.12 115.03 97.01 86.95 94.81
8.01 9.09‘ 14.12 13.91 4.39 2.36 2.85 9.73 13.90 10.38 22.01 5.76 23.64 5.33 2.81 5.72 6.80 13.92 7.69 5.99 16.09 20.47 2.15 13.06 7.47
increased, and (iii) lower doses of the contrast agent will be required, decreasing the cost and risk of toxicity. Tissue specific targeting may have another important advantage with cardiovascular ultrasound. Active plaque morphology can be identified by the contrast agent conjugated to antibodies specific to cardiac tissues. For example, targeting with antibodies such as VCAM-1 and ICAM-1 would demonstrate early stages of atheroma development,12 whereas targeting with antifibrin-attached liposomes would delineate areas of thrombosis and plaque rupture. The purpose of this study was to develop an acoustic contrast agent that (i) is inherently echogenic and stable, (ii) has a mean size lower than 1 µm, eliminating pulmonary entrapment, and (iii) can be conjugated to antibodies for specific tissue targeting. This report describes the development of inherently echogenic liposomes composed of phosphatidylcholine, phosphatidylglycerol, phosphatidylethanolamine, and cholesterol. The compositions that provide high echogenicity are identified by varying the components in a systematic manner. The liposomes demonstrating highest acoustic reflectivity are tested for the retention of their echogenic properties under simulated in vivo conditions and in vivo in an animal model.
Experimental Section MaterialssL-R-Phosphatidylcholine type XIII-E (PC), L-R-phosphatidylethanolamine dipalmitoyl (PE), and cholesterol (CH) were obtained from Sigma Chemical Co., St. Louis, MO. 1,2-Dipalmitoylsn-glycero-3-phosphoglycerol was obtained from Fine Chemicals, Liestal, Switzerland. MethodssPreparation of LiposomessPhospholipids and cholesterol were dissolved in chloroform and combined in amounts to achieve the desired lipid composition as shown in Table 1 . The solvent was evaporated by a rotary evaporator (Labconco, Kansas City, MO), rotated at 120 rpm, and immersed in a thermostated water bath, with the temperature at 50 °C under argon. The resulting lipid film was then placed in a dessicator under vacuum for 2 days for complete drying. The dry lipid film was rehydrated with deionized water to yield a concentration of 10 mg of lipid/mL of water. After removing the film from the flask walls, the dispersion was sonicated in a 175.5 W water bath sonicator (Fisher Scientific, Itasca, IL) until the mean
size of the liposomes was ca. 500 nm. Liposome size was determined by quasielastic light scatter as described below. D-Mannitol (0.2 M) was added to the liposome suspension in a 1:1 (v/v) ratio, and this suspension was frozen overnight at -70 °C. The frozen samples were placed in a lyophilizer (Labconco, Kansas City, MO) for 48 h. The dried lipids were then resuspended with 0.1 M phosphate buffer, pH 7.4, to yield a concentration of 10 mg of lipid/mL of buffer. After hand shaking and brief vortexing, the size of the resulting liposomes was again determined. The mean size of the liposomes remained under 1 µm in all cases and in most cases was <800 nm. Size MeasurementssThe mean size of the liposomes was determined by quasielastic light scattering (QELS) measurements using a Nicomp model 270 submicron particle sizer (Pacific Scientific, Menlo Park, CA) equipped with a 5 mW helium-neon laser at an exciting wavelength of 632.8 nm, a 64-channel autocorrelation function, a temperature-controlled scattering cell holder, and an ADM 11 video display terminal computer (Lear Siegler Inc., Anaheim, CA) for analyzing the fluctuations in scattered light intensity generated by the diffusion of particles in solution. The mean hydrodynamic particle diameter, dh, was obtained from the Stokes-Einstein relation using the measured diffusion coefficient obtained from the fit. ImagingsLiposomes were transferred to liquid scintillation vials and imaged with a 20 MHz high-frequency intravascular ultrasound (IVUS) imaging catheter (Boston Scientific Inc., Sunnyvale, CA). The IVUS catheter was passed through the vial cap and secured. Instrument settings for gain, zoom, compression and rejection levels were optimized at the initiation of the experiment and held constant for all samples. Images were recorded onto 1/2” VHS videotape in real time for subsequent playback and image analysis. Videodensitometric Analysis of Liposome “Brightness”sRelative echogenicity (apparent brightness) of all liposome formulations was objectively assessed via computer-assisted videodensitometry. This process involved image acquisition, preprocessing, automated liposome identification, and gray scale quantification. All image processing and analysis were performed with Image Pro Plus Software (Version 1.0, Media Cybernetics, Silver Springs, MD) running on a dedicated computer (486 CPU, 66 MHz). Randomly selected IVUS images were acquired from videotape for each liposome formulation. Images were digitized to 640 × 480 pixel spatial resolution (ca. 0.045 mm/pixel) and 8 bit (256 levels) amplitude resolution. All analyzed IVUS data were collected at a fixed instrument gain level. The distribution of gray scale values within the image was then adjusted to cover the entire range of possible gray levels using a linear transformation algorithm (i.e., dynamic range was maximized). Image brightness was subsequently scaled such that a reference feature, the annotation text, common to each image, retained a constant gray scale value over all images. An automated liposome detection routine was then run to identify liposomes suspended in solution within an annular region of interest set at a constant radial distance from the imaging catheter. The automated liposome detection routine identified all “bright” objects within the analysis annulus having a gray scale level >29, a roundness ratio (i.e., ratio of maximum diameter:minimum diameter) <2.5, and a size >4 pixels. This procedure excluded virtually all imaging artifact from our detection algorithm. Thus, objects identified were considered to be “liposomes”. Each liposome was outlined and numbered by the computer program. The average gray scale and size of each liposome were also quantified. The mean gray scale value of all pixels identified as “liposomes” with a given image was then computed and used to characterize the echogenicity of a given liposome formulation. Freeze-Etch Electron MicroscopysLiposomes were prepared for freeze fracture according to standard techniques as reported previously.13,14 Briefly, drops of the liposome suspension were frozen in liquid-nitrogen-cooled Freon 22. The samples were fractured using a Balzers BAF 301 freeze-etch unit at -115 °C. The fractured samples were coated with platinum and carbon. The replicas were cleansed in a minimum of two changes of bleach (sodium hypochlorite), washed with distilled water, dried, and collected on 200 mesh copper grids. Replicas were examined and photographed with a JEOL 100CX transmission electron microscope at 80 kV. Liposomes in Simulated Physiological ConditionssEchogenic liposomes consisting of phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol (PG), and cholesterol in a molar ratio of 60:8: 2:30 were prepared by the previously described method. The liposome suspension in buffer was imaged at room temperature. One half of the suspension was warmed in a water bath to a temperature of 37
Journal of Pharmaceutical Sciences / 487 Vol. 85, No. 5, May 1996
Figure 1sEffect of phosphatidylethanolamine and phosphatidylglycerol content on the echogenicity of liposomes. °C and then imaged. The other half of the suspension was combined with canine blood in a 1:1 (v/v) ratio and then imaged at room temperature. The suspension temperature was raised to 37 °C and the suspension imaged again. Canine blood mixed with saline in a 1:1 (v/v) ratio was used as a control. In Vivo EvaluationsA standard Yucatan miniswine model of induced atherosclerosis was developed to test in vivo liposomal acoustic properties. Following development of atherosclerosis (1-3 weeks postinitial surgery), the animal was anesthetized and artificially ventilated. Sheaths were inserted into the femoral and carotid arteries for vascular access and imaging catheter insertion. Under stable hemodynamic conditions, the imaging catheters were inserted into the arteries. Base-line imaging of the artery and luminal blood was performed, and images were recorded onto videotape. Following base-line imaging, 2-5 cc of liposomes was injected through the catheter, and similar imaging occurred. Comparisons between gray scale characteristics of in vivo base-line blood and liposomes in blood were made. Data were compared by ANOVA and determined to be statistically different at p < 0.05.
Results Composition StudiessThe effect of lipid composition on the acoustic properties of the liposomes prepared by the same method and with similar mean size is shown in Table 1 and Figure 1. It is clear that with relatively high concentrations of PE (4-12%), low concentrations of PG (1-3%), and constant cholesterol content (30%), highest echogenicity levels are obtained. The structural properties of a highly echogenic and poorly echogenic liposomal formulation (samples 21 and 17 in Table 1) are demonstrated by freeze-etch electron microscopy in Figure 2. The high echogenic formulation (2, top) consists of a multilamellar structure with well-separated bilayers, whereas low echogenicity is associated with a thick unseparated multilayer for a single liposome (2, bottom). Figure 3, top and middle, shows ultrasound images of liposomal suspensions with formulations corresponding to those in Figure 2. Figure 3, bottom, demonstrates an ultrasonic image of 0.1 M phosphate buffer alone. Buffer, like water, is not acoustically reflective and therefore has mean gray scale values of zero. Stability StudiessAfter optimizing the lipid composition of the liposomes to give high acoustic reflectivity, the next step was to determine if this echogenicity would be retained under in vivo conditions. Figure 4 demonstrates the echogenic levels, as mean gray scale values, obtained when liposomes with high acoustic reflectivity were exposed to simulated in vivo conditions, i.e., 37 °C and presence of blood components. Neither the increase in temperature nor the presence of blood 488 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
Figure 2sFreeze fracture micrographs showing the structure of liposomes (12000×, reproduced at 50% of original size): (top) liposomes with high echogenicity composed of a 60:8:2:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer) and (bottom) liposomes with low echogenicity composed of a 58:8:4:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer).
caused a decrease in acoustic reflectivity of the liposomes. For comparison purposes, the echogenicity of blood alone is also shown. The retention of echogenicity of the liposomes was also demonstrated in actual in vivo conditions as seen in Figure 5. Mean gray scale values for liposomes that were injected into the arterial bed of a Yucatan miniswine model were similar to those obtained under in vitro conditions. Factors Influencing ResultssThe resolution of our 2D images is ca. 30-45 µm. Therefore, although smaller particles may have acoustic reflectivity, they will be displayed at this pixel size; whether the reflectivity is due to single acoustically reflective liposomes or aggregates of such remains to be determined. Additionally, particles within the solution may have added to the specular reflectors of our formulation. However, when measuring the gray scale values of the different liposomal formulations, the only components of the suspensions being analyzed were liposomes and buffer. We have shown that buffer is not acoustically reflective (Figure 3, bottom).
Discussion Effect of Lipid Composition on Echogenicitys Liposomes are vesicles composed of phospholipid bilayers surrounding an aqueous space. They have been shown to be safe for injection and suitable for antibody conjugation.15,16 Liposomes also have the advantage that they can easily be formulated below 1 µm in diameter which is desirable for safe
Figure 4sEffect of simulated physiological conditions of the echogenicity of liposomes: (A) liposomes composed of a 60:8:2:30 molar mixture of PC:PE:PG: CH in 0.1 M phosphate buffer (pH 7.4) at room temperature, (B) liposomes of the same composition in 0.1 M phosphate buffer (pH 7.4) warmed to 37 °C, (C) liposomes of the same composition in canine blood at room temperature, (D) liposomes of the same composition in canine blood warmed to 37 °C, and (E) blood at 25 °C.
Figure 5sIn vivo liposome echogenicity. Liposomes composed of a 60:8:2:30 molar mixture of PC:PE:PG:CH in 0.1 M phosphate buffer (pH 7.4) at a concentration of 10 mg of lipid/mL of buffer were injected into the arteries of a Yucatan miniswine. The echogenicity of the liposomes in vivo was compared to that of the blood components. Figure 3sUltrasound images of liposomes: (top) intravascular ultrasound catheter in 10 mL glass vial containing liposomes with high echogenicity composed of a 60:8:2:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer), (middle) intravascular ultrasound catheter in 10 mL glass vial containing liposomes with low echogenicity composed of a 58:8:4:30 molar mixture of PC:PE:PG:CH (10 mg of lipid/mL of buffer), and (bottom) intravascular ultrasound catheter in 10 mL glass vial containing 0.1 M phosphate buffer.
pulmonary capillary passage. Gas-filled liposomes have previously been reported to be highly acoustically reflective.17,18 However, trapping gas in liposomes not only requires a sophisticated preparation process but also may cause instability of the system under both in vivo and in vitro conditions. Our study demonstrates that acoustically reflective liposomes can be produced solely by choosing the optimal lipid composition and applying a dehydration/rehydration process. Four different lipids were used in the liposome formulation. Phosphatidylcholine, being the most common
phospholipid in the cell membrane, was chosen as the main lipid. Cholesterol was added to the formulation to add rigidity to the lipid bilayer. Our preliminary data had shown an increase in acoustic reflectivity in the presence of cholesterol.19 Phosphatidylethanolamine and phosphatidylglycerol were included in the liposomal formulation to allow for conjugation and add negative charge, respectively. Separation of the phospholipid bilayers within a liposome appears to affect its acoustic reflectivity. In the presence of PE and PG at optimized concentrations for highest echogenicity, we obtained oligolamellar structures, represented by vesicles within vesicles, whereas for liposomes with low acoustic properties the structure of liposomes consisted of unilamellar or unseparated multilamellar bilayers (Figure 2). This finding is in close agreement with previous data evaluating air-entrapped liposomes, where it was demonstrated that oligolamellar liposomes had higher acoustic reflectivity than unilamellar
Journal of Pharmaceutical Sciences / 489 Vol. 85, No. 5, May 1996
liposomes.18 We believe that PG, because of its negative charge, causes separation of the bilayers as well as prevents the aggregation of liposomes. Phosphatidylethanolamine on the other hand, with its small head group, may affect the curvature of the bilayers and formation of vesicles within vesicles. However, there is an optimal composition where all these effects maximize to give the best structural changes in the liposome for the highest echogenicity. Retention of Echogenicity under in Vivo ConditionssIn order to develop a successful contrast agent, the acoustic reflectivity of the liposomes demonstrated under in vitro conditions should be retained in vivo. First we tested them under simulated in vivo conditions by introducing them into media at body temperature and in the presence and absence of blood components. Temperature elevations up to 37 °C did not show any significant change in echogenicity. This is most likely due to the fact that the gel state of the lipid bilayers which were formulated at room temperature in the presence of cholesterol was still maintained at 37 °C. Similarly, due to the rigidity of the bilayers, the blood components did not interfere with the phospholipid bilayers causing structural changes within the liposomes. Previous studies have shown the improvement of in vivo stability of phosphatidylcholine liposomes in the presence of cholesterol.20 Since the simulated in vivo conditions did not fully represent the real in vivo conditions, such as the presence of certain ions and enzymes, the acoustic behavior of the liposomes was also tested after injection into the arteries of the Yucatan miniswine model. It was confirmed that the echogenic liposomes developed in this study were stable to all the factors originating from in vivo conditions.
Conclusions Liposomes that are acoustically reflective without inclusion of gas or other agents can be prepared by choosing an optimal lipid composition and applying a dehydration/rehydration method. The echogenicity of these liposomes is not lost or decreased at body temperature or in the presence of blood components. The novel contrast agent developed in this study
490 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
can be used de novo or following conjugation to tissue specific antibodies for local image enhancement, diagnosis, and drug delivery.
References and Notes 1. Goldberg, B. B.; Liu, J. B.; Forsberg, F, Ultrasound Med. Biol. 1994, 20, 319-333. 2. Winkelmann, J. W.; Kenner, M. D.; Dave, R.; Chandwaney, R. H.; Feinstein, S. B. Ultrasound Med. Biol., 1994, 20, 507-515. 3. Feigenbaum, H.; Stone, J. M.; Lee, D. A. Circulation 1970, 41, 613-620. 4. Ophir, J.; Parker, K. J. Ultrasound Med. Biol. 1989, 15, 319325. 5. Feinstein, S. B.; Keller, M. W.; Kerber, R. E. J. Am. Coll. Cardiol. 1989, 2, 125-131. 6. Feinstien, S. B.; Cheirif, J. J. Am. Coll. Cardiol. 1990, 16, 316324. 7. Smith, M. D.; Elion, J. L.; McClure, R. R.; Kwan, O. L.; DeMaria, A. N. J. Am. Coll. Cardiol. 1989, 12, 622-1628. 8. Andre, M. P.; Steinbach, G.; Mattrey, R. F. Invest. Radiol. 1993, 28, 502-506. 9. Unger, E.; Shen, D.; Fritz, T.; Kulik, B.; Lund, P.; Wu, G.; Yellowhair, D.; Ramaswami, R.; Matsunaga, T. Invest. Radiol. 1994, 29, S134-S136. 10. Simon, R. H.; Ho, Sy.; Lang, S. C.; Uphoff, D. G.; D’Arrigo, J. S. Ultrasound Med. Biol. 1993, 19, 123-125. 11. Quay, S. C. J. Ultrasound Med. 1994, 13, S9. 12. Cybulsky, M. I.; Gimbrone, M. A. Science 1991, 251, 788-791. 13. Moor, H.; Muhlethaler, K. J. Cell. Biol. 1963, 17, 609-628. 14. Kuszak, J. R.; Alcala, J.; Maisel, H. Exp. Eye Res. 1980, 33, 157166. 15. Alving, C. R. In Liposomes: From Biophysics to Therapeutics; Ostro, M. J., Ed.; Marcel Dekker, Inc.: New York, 1987; pp 202204. 16. Maruyama, K.; Takizawa, T.; Yuda, T.; Kennel, S. J.; Huang, L.; Iwatsuru, M. Biochim. Biophys. Acta Biomembr. 1995, 1234, 74-80. 17. Ryan, J. R.; Davis, A.; Melchior, D. L. U.S. Patent 4,900,540, 1990. 18. Unger, E. C. U.S. Patent 5,123,414, 1992. 19. Lanza, G. M.; Alkan, M. H.; Vonesh, M. J.; Kleggerman, M. E.; Frazin, L. J.; Mehlman, D. J.; Talano, J. V.; McPherson, D. D. J. Am. Coll. Cardiol. 1992, 19, 114A. 20. Senior, J. H. Crit. Rev. Ther. Drug Carrier Syst. 1987, 3, 123.
JS950407F