UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies on the antiatherogenic properties of liposomes Rodrigueza, Wendi V. 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-893664.pdf [ 6.93MB ]
JSON: 831-1.0099217.json
JSON-LD: 831-1.0099217-ld.json
RDF/XML (Pretty): 831-1.0099217-rdf.xml
RDF/JSON: 831-1.0099217-rdf.json
Turtle: 831-1.0099217-turtle.txt
N-Triples: 831-1.0099217-rdf-ntriples.txt
Original Record: 831-1.0099217-source.json
Full Text

Full Text

STUDIES ON THE ANTIATHEROGENIC PROPERTIES OF LIPOSOMES by WENDI V. RODRIGUEZA BSc. Biochemistry, University of British Columbia, 1988 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF BIOCHEMISTRY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1994 © Wendi V. Rodrigueza, 1994 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. (Signature) Department of T > 0 C/E\£/^0sf»M The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT There has been a wealth of prior literature demonstrating that the infusion of phospholipid causes the reversal of experimentally induced atherosclerosis in animals. However, the development of liposomal therapies to manage atherosclerosis is only now being considered seriously. The studies presented in this thesis were aimed at confirming the antiatherogenic properties of phospholipid using a pharmaceutically acceptable formulation of liposomes capable of pre-clinical development. The work can be divided into three main areas of investigation. In order to test the capacity of liposomes to act as a "sink" for cholesterol, the relative rates of movement of sterol were measured between leaflets and between vesicles. The kinetics of movement of cholesterol and cholesterol sulphate (CS) between vesicles were studied using donor vesicles of various size ranging from 40-250 nm, composed of egg phosphatidylcholine (EPC)/Sterol/N-palmitoyl dihydrolactosylcerebroside (75:10:15 mol ratio). Trace amounts of [3H]-sterol in donor vesicles was used to monitor sterol transfer into a 10 fold excess of EPC large unilamellar vesicles (LUV) of 100 nm diameter. Following the addition of a lectin, the two populations of vesicles were separated by centrifugation. The rate constants for efflux and transbilayer diffusion for both sterol molecules were determined after fitting kinetic data to a three compartment model. The rate of intermembrane exchange for CS was considerably faster than for cholesterol in all liposomes tested. Using the kinetic model, a rate of transbilayer movement for cholesterol and CS was estimated. In both cases, it was found to be slower than the rate of efflux from the surface of vesicles. The results from this chapter suggest that the greatest rate of cholesterol uptake in vivo should be achieved using unilamellar vesicles with greatest surface area and the smallest diameter. The second group of studies examined the capacity of different liposomal preparations (as studied in chapter two), to mobilize cholesterol in vivo. This involved measuring the ability of liposomes to increase plasma cholesterol concentrations in mice. Surprisingly, LUV, composed of EPC, were found to be approximately twice as effective at mobilizing cholesterol than small unilamellar vesicles (SUV) of the same composition. This is because the increase in plasma cholesterol is proportional to the residence time of vesicles in the circulation. LUV with ii a diameter of approximately 100 nm, accumulate the most sterol during 24 h in the animal model tested here. Vesicles composed of phospholipid in a gel state at physiological temperatures, gave rise to a smaller increase in plasma cholesterol compared to liquid-crystalline vesicles. A significant decrease in the ratio of cholesterol-to-phospholipid in the lipoprotein pool was observed following liposome infusion, whereas cholesterol levels and turnover in erythrocytes remained constant. These results indicate that net transfer of cholesterol into liposomes in vivo occurs more extensively from the lipoprotein cholesterol pool than from the erythrocyte cell membrane pool. This is consistent with the hypothesis of Williams et al. (1984) that liposomes enhance reverse cholesterol transport by generating cholesterol-poor HDL particles that can extravasate and promote more sterol efflux from peripheral tissues. In the final groups of studies presented here, the antiatherogenic properties of EPC LUV of 100 nm diameter were tested in an experimental model for atherosclerosis in which rabbits were fed a diet rich in cholesterol. Forty eight rabbits were divided into two diet groups fed standard rabbit chow or fed a diet enriched in cholesterol (0.5% by weight) to induce the formation of atherosclerotic plaques. Prior to the initiation of phospholipid therapy, the cholesterol diet was ceased and all animals were returned to standard rabbit chow. The treatment protocol involved a total of 10 bolus injections of vesicles at a phospholipid dose of 300 mg/kg or an equivalent volume of saline, with one injection given to each animal every 10 days. Liposomal injections brought about a large movement of cholesterol into the blood pool and resulted in a significant reduction in the cholesterol content of aortas as well as the degree of surface plaque involvement in atherosclerotic animals. Most notably, the thoracic aorta exhibited a 48% reduction in tissue cholesterol content per gram of protein compared to tissue from saline-treated controls. Histochemical analyses revealed that aortas from animals receiving the repeated infusions of phospholipid, displayed less cholesterol deposits in lesions, and a moderate reduction in the ratio of intimal/medial thickness. This liposome-mediated regression of atheroma was observed despite persistent elevation of plasma cholesterol levels after removal of the animals from the high cholesterol diet. These results support previous observations and suggest that phospholipid infusion may indeed be a useful therapy in the management of atherosclerosis. iii TABLE OF CONTENTS Abstract ii Table of Contents iv List of Figures vii List of Tables x Abbreviations xi Acknowledgements xiv Dedication xv Chapter 1 Introduction 1.1 Liposomes and their therapeutic potential in atherosclerosis: an overview 1 1.2 Membranes 4 1.2.1 Biological Membranes 4 1.2.2 Chemical and physical properties of phospholipids and cholesterol 5 Phospholipids 6 Cholesterol 8 Cholesterol-phospholipid interactions 8 1.2.3 Liposomes 12 Classification and preparation of liposomes 12 Multilamellar vesicles (MLV) 12 Small unilamellar vesicles (SUV) 15 Large unilamellar vesicles (LUV) 15 1.3 Factors influencing cholesterol movement between membranes 16 1.3.1 Lipid Motions 17 1.3.2 Intermembrane cholesterol exchange 17 Collision complex model 19 Aqueous diffusion model 20 Existence of cholesterol domains in membranes 23 1.4 Lipoprotein classification, structure, metabolism and function 26 1.4.1 Structure and classification 26 1.4.2 Metabolic relationships between lipoproteins 29 Chylomicrons 29 Very low density lipoproteins (VLDL) 31 Low density lipoproteins (LDL) 31 High density lipoproteins (HDL) 31 1.4.3 Reverse cholesterol transport 33 1.4.4 Cholesterol homeostasis 34 iv 1.5 Pathogenesis of atherosclerosis 37 1.6 Factors that influence the fate of liposomes and liposomal lipid in vivo 38 1.6.1 Role microcirculation in liposome clearance 40 1.6.2 Recognition by phagocytic cells 40 1.6.3 Role of lipoproteins and apolipoproteins 42 1.7 Liposome-induced regression of atherosclerosis 43 1.7.1 Vesicle-cell interactions 43 1.7.2 Vesicle-HDL interactions 44 1.7.3 Mechanism of action 47 1.8 Thesis objectives 47 Chapter 2 The Kinetics of Exchange and Net Flux of Cholesterol and Cholesterol Sulphate between Liposomal Membranes 2.1 Introduction 49 2.1.1 An overview 49 2.1.2 Cholesterol exchange and net efflux 50 2.2 Materials and methods 55 2.2.1 Materials 55 2.2.2 Synthesis of [l,2-3H]-cholsterol sulphate from [l,2-3H]-cholesterol 55 2.2.3 Preparation of vesicles 56 2.2.4 Determination of vesicle diameters 57 2.2.5 Measurement of [3H]-sterol exchange 58 2.2.6 Calculation of rate constants for transbilayer diffusion and exchange 59 2.2.7 Estimation of the amount of lipid present in outermonolayer vesicles 62 2.2.8 Ion exchange chromatography of vesicles containing cholesterol sulphate 62 2.3 Results 63 2.3.1 Compartment volumes and lipid distributions 63 2.3.2 Effect of liposome size and lamellarity on cholesterol exchange 67 2.3.3 Effect of liposome size and lamellarity on cholesterol sulphate exchange 70 2.4 Discussion 75 Chapter 3 The Influence of Size and Composition on the Cholesterol Mobilizing Properties of Liposomes In Vivo 3.1 Introduction 80 3.2 Materials and methods 81 3.2.1 Materials 81 3.2.2 Preparation of vesicles 81 3.2.3 Vesicle diameters 82 3.2.4 Vesicle infusions 82 v 3.2.5 Gel filtration 83 3.2.6 Lipid analysis 83 3.2.7 Erythrocyte cholesterol specific activity 84 3.2.8 Cholesterol exchange in vitro 84 3.2.9 Cholesterol excretion 85 3.2.10 Statistical analyses 85 3.3 Results 86 3.3.1 Cholesterol mobilization 86 3.3.2 Liposome size and cholesterol mobilization 91 3.3.3 Liposome composition 96 3.3.4 Source of liposomal cholesterol 98 3.3.5 Cholesterol excretion 101 3.4 Discussion 102 Chapter 4 Cholesterol Mobilization and Regression of Atheroma Induced by Liposomes 4.1 Introduction 106 4.2 Materials and Methods 107 4.2.1 Materials 107 4.2.2 Rabbits 107 4.2.3 Experimental design 108 4.2.4 Vesicle Preparation 110 4.2.5 Collection of blood and tissue samples 110 4.2.6 Preparation aortas for analyses 110 4.2.7 Digitization I l l 4.2.8 Lipid analysis 112 4.2.9 Protein analyses 112 4.2.10 Histochemical analysis of aortic samples 112 4.2.11 Statistical analyses 113 4.3 Results 113 4.3.1 Establishment of lesions 113 4.3.2 Cholesterol mobilization 114 4.3.3 Effect of repeated injections 119 4.3.4 Source of liposomally accumulated cholesterol 121 4.3.5 Assessment of atherosclerotic plaque involvement 123 Aortic lipid content and digitization 123 Histochemical analyses of aortic samples 125 4.3.6. Liver Cholesterol content 126 4.4 Discussion 129 Chapter 5 Summarizing Discussion and Future Directions 134 Bibliography 140 vi LIST OF FIGURES Figure 1. Diagrammatic Structure of a Biological Membrane 5 Figure 2. Structure of Phospholipids and Commonly Occuring Headgroups 7 Figure 3. Structure of Cholesterol, Cholesterol Sulphate and a Cholesterol Ester 9 Figure 4. Model of Cholesterol-Phospholipid Interaction in a Bilayer 10 Figure 5. Schematic Representation of Liposome Size and Lamellarity 13 Figure 6. Freeze-Fracture Electron Micrographs of LUV Produced by Extrusion 14 Figure 7. Various Motions that a Lipid Molecule Undergoes in a Membrane 18 Figure 8. Schematic Representation of the Collision Complex Model 19 Figure 9. Schematic Representation of the Aqueous Diffusion Model 21 Figure 10. Energy of Activation Associated with Cholesterol Exchange 22 Figure 11. Structure of a Lipoprotein Particle 28 Figure 12. Metabolic Relationships Between Lipoproteins 30 Figure 13. Cholesterol Homeostasis within a Cell 36 Figure 14. Postulated Steps Leading to the Pathogenesis of Atherosclerosis 39 Figure 15. Anatomy of a Vessel Wall 41 Figure 16. Proposed Mechanism of Liposome Induced Regression of Atherosclerosis 46 Figure 17. Schematic Representation of Three Categories of Sterol Exchange Observed in the literature 52 Figure 18. Literature Summary of the Kinetics of Cholesterol Exchange 53 Figure 19. Experimental Conditions, Rate Constants and the Three Compartment Model 60 Figure 20. Estimation of Outermonolayer Phospholipid in MLV40o 65 vii Figure 21. Flux of Cholesterol Between Donor and Acceptor Vesicles 66 Figure 22. Relative Rates of Cholesterol Depletion from the IM and OM Estimated by the Kinetic Model 70 Figure 23. Flux of Cholesterol Sulphate Between Donor and Acceptor Vesicles 72 Figure 24. Relative Rates of Cholesterol Sulphate Depletion from the IM and OM Estimated Using the Kinetic Model 73 Figure 25. Elution of Vesicles from DEAE-Sepharose Columns: Measurement of Surface Charge 74 Figure 26. Cholesterol Mobilization by the Intravenous Injection of Phospholipid Liposomes 87 Figure 27. Separation of Vesicles from Plasma by Gel Filtration 89 Figure 28. Rate of Cholesterol Accumulation of Vesicles in vivo 90 Figure 29. Estimation of Cholesterol Mobilized by Different Sized Vesicles 94 Figure 30. Vesicle Lamellarity and the Rate of Cholesterol Equilibration Between Membranes 95 Figure 31. Effect of Lipid Composition on the Rate and Extent of Cholesterol Mobilization 97 Figure 32. Possible Source of Liposomally Accumulated Cholesterol 100 Figure 33. Experimental Design Employed to Test the Efficacy of Liposomal Therapy.. 109 Figure 34. Redistribution of Cholesterol into Plasma Following a Treatment with LUV100 at 300 mg/kg 115 Figure 35. Phospholipid Clearance During a Treatment 116 Figure 36. An Estimate of Cholesterol Removed During an Injection of EPC LUV10o— 118 Figure 37. Consequences of Repeated Injections of EPC LUV100 120 Figure 38. Cholesterol-to-Phospholipid Ratios in Lipoproteins During Vesicle Infusions 122 viii Figure 39. Assessment of Aortic Lipid Content and the Degree of Plaque Involvement 124 Figure 40. Representative Sections Obtained From the Thoracic Aorta 127 Figure 41. Schematic Representation of a Histological Section Used for the Measurement of Intimal/Medial Ratios 128 ix LIST OF TABLES Table 1. Summary of Liposome Induced Regression of Lesions in Animal Models of Atherosclerosis 3 Table 2. Composition of the Major Lipoprotein Classes 27 Table 3. Apolipoprotein Composition of Human Plasma Lipoproteins 27 Table 4. Approximate Half-times for Cholesterol Exchange from Donor Particles 45 Table 5. Vesicle Preparations Used in Sterol Flux Studies 57 Table 6. Summary of Rate Constants and Half-times for Flip-Flop and Exchange in Various Sized Vesicles 67 Table 7. Measurement of Intimal/Medial Ratios in the Different Regions of the Aorta of Vesicle and Saline-Treated Atherosclerotic Animals 128 x ABBREVIATIONS ACAT apo apoA apoB apoC apoD apoE BCA C3 C-C CE Cer CETP CHOL CMC CL C:P CS DSPC DSPG EDTA EPC EPG HBS HDL acyl CoA:choIesterol transferase apolipoproteins apolipoprotein A apolipoprotein B apolipoprotein C apolipoprotein D apolipoprotein E bicinchoninic acid complement protein 3 carbon-carbon bond cholesterol esters cerebroside cholesteryl ester transfer protein cholesterol critical micellar concentration cardiolipin cholesterol-to-phospholipid ratio cholesterol sulphate distearoylphosphatidylcholine distearoylphosphatidylglycerol ethylenediaminetetraacetic acid egg phosphatidylcholine egg phosphatidylglycerol HEPES-buffered saline high density lipoprotein xi H&E HEPES HMG-CoA reductase IgG IM kl k2 LCAT LDL LPL LUV LUV100 LUV50 MLV MLV400 NMR N.S. OM PA PC PE PG PI PL QELS RCT RES haematoxylin and eosin [4-(2-hydroxyethyl)]-piperazine ethane sulfonic acid |3-Hydroxy-|3-methylglytaryl-coenzyme A reductase immunoglobulin G inner monolayer rate constant for transbilayer diffusion of sterol rate constant for intermembrane sterol exchange lecithinrcholesterol acyl transferase low density lipoprotein lipoprotein lipase large unilamellar vesicles LUV extruded through 100 nm pore size polycarbonate filters LUV extruded through 50 nm pore size polycarbonate filters multilamellar vesicles MLV extruded through 400 nm pore size polycarbonate filters nuclear magnetic resonance not significant outer monolayer phosphatidic acid phosphatidylcholine phosphatidylethanolamine phosphatidylglycerol phosphatidylinositol phospholipid quasi-electric light scattering reverse cholesterol transport reticuloendothelial system xii SDS SUV UA tfcflip ^ e x c Tc TG TLC VLDL sodium dodecyl sulphate small unilamellar vesicle half-times half-times of flip-flop half-times of intermembrane exchange gel-to-liquid crystalline phase transition temperature triglycerides thin layer chromatography very low density lipoprotein xiii ACKNOWLEDGEMENTS There are quite a few people who helped me throughout graduate studies and I'd like to say thank you to all of them, especially the following people. Sandy, thank you for the help with the animals and all the last minute things; Sean, thanks for the figures and help also with the animals; Cath, thank you for your endless encouragement; Laval and Jeff thanks for simply being around; The members of the Cullis lab who are too many to mention, but have always been there to lend a helping hand over the years, thank you. Dana Masin, the staff at the Research Centre and George Spurr thank you for the assistance in animal handling; Dr. Haydn Pritchard and the members of his lab thanks for sharing your expertise; Marcel, Neil, Kim, Tom and Roger Brownsey thank you for your support. And of course my supervisors, Drs. Mick Hope and Pieter Cullis, I thank you both for the patience, support, and encouragement. Mick, thank you very much for all the understanding, reassurances and for reading this thesis. Graduate studies has been one of the most memorable times in my life. I am indebted to Mick and Pieter for giving me the opportunity to pursue such an interesting project, to enjoy science and to be a part such a dynamic and fun group. I am also thankful for the friends I've made in the department, although nameless, have in their own way helped me and have made graduate school so enjoyable and worthwhile. Finally, I'd like to thank my family. Thank you Mimi for being here with me; Mommy and Papa thanks for your understanding and encouragement; and to the rest of my family, my brother, my new sister and my aunts, thank you for all that you have done for me over the years, it has been appreciated. xiv To Mommy and Papa and my family xv CHAPTER 1 INTRODUCTION 1.1 LIPOSOMES AND THEIR THERAPEUTIC POTENTIAL IN ATHEROSCLEROSIS: AN OVERVIEW It is now recognized that reverse cholesterol transport (RCT), which involves the efflux of extrahepatic cell cholesterol and its transport to the liver, is an important process in the prevention and possible reversal of atherosclerosis (Franceschini et. al, 1991). High density lipoproteins (HDL) are believed to be the physiological acceptor of tissue free cholesterol and high levels of HDL are correlated with a decreased risk of cardiovascular disease (Gwynne, 1991; Rothblat and Phillips, 1991). These observations have led to searches for the means to increase HDL concentrations and/or stimulate RCT. Several studies have demonstrated that the intravenous infusion of phospholipid vesicles results in the regression of atherosclerosis (reviewed in Williams et al., 1984). This phenomenon is thought to occur because vesicles act as a "sink" for HDL cholesterol thereby generating cholesterol-depleted HDL particles which have an increased capacity to scavenge more sterol from vascular and extravascular tissues thus augmenting the ability of HDL to promote RCT (Williams et al., 1984, Williams and Tall, 1988). In the 1950's, pioneering work by Friedman and Byers demonstrated that lipid infusion caused the regression of atherosclerotic lesions in rabbits (Friedman et al., 1957). However, at the time neither the structure of liposomes nor lipoproteins were known. Their initial investigations were focused on gaining insight into the interrelation of elevated plasma cholesterol, triglycerides and phospholipid levels observed in hyperlipidaemic individuals. It was observed that when rats were given an intravenous infusion of triglycerides or phospholipid in aqueous suspension, the phospholipid, but not triglycerides caused a dramatic increase in the concentration of cholesterol in plasma (Friedman and Byers, 1956; DiLuzio and Zilversmit, 1 1960) and whole blood (Friedman and Byers, 1958). This result was subsequently shown to be independent of changes in the rates of excretion (Byers and Friedman, 1962) and synthesis (Byers et al., 1962) of cholesterol, but rather were caused by the mobilization of extravascular pools of sterol. Indeed a decrease in the cholesterol content of several peripheral tissues was noted (Friedman and Byers, 1958). Thus, the transient hypercholesterolaemia owing to phospholipid infusion appears to result from a redistribution of tissue cholesterol into the bloodstream (Byers et al., 1962; Byers and Friedman, 1969). From this it was hypothesized that repeated infusions of phospholipid might also mobilize cholesterol stores from established atheroma. Indeed, using the cholesterol-fed rabbit, it was demonstrated that weekly infusions of phospholipids at a dose of 400-1200 mg/kg resulted in a dramatic regression of lesions in these animals (Friedman et al., 1957). A number of other studies in a variety of animal models have since supported the observations of Friedman and colleagues (1957). These have demonstrated that experimentally induced atherosclerosis can be significantly reversed following the intravenous administration of liposomes and that even lower doses of 200 mg/kg are also effective (see Table 1 for summary and Williams et al., 1984). For instance, in rabbits made atherosclerotic by a cholesterol-free, semi-synthetic diet rich in saturated fat, repeated infusions of phospholipid suspensions resulted in a regression of lesions (Patelski et al., 1970). In baboons, the incidence and severity of experimental atherosclerosis induced over 6 months, by the combination of cholesterol feeding and the injection of bovine serum albumin to produce vasculitis and intimal injury, was also reduced as a result of concurrent thrice weekly infusions of polyunsaturated phosphatidylcholine (Howard et al., 1971). Furthermore, arterial lesions in Japanese Quail, induced by a cholesterol rich diet, were significantly reduced following 3 subsequent months of phospholipid infusions, even though the animals were maintained on the enriched diet (Stafford and Day, 1975). 2 Table 1. Summary of liposome induced regression of lesions in animal models of atherosclerosis Animal Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Baboon Quail Investigators Friedman et al. (1957) Maurukas et al. (1959) Sachs et al. (1960) Byers and Friedman (1960) Adams et al. (1967) Patelski et al. (1970) Howard et al. (1971) AJtman et al. (1974) Stafford and Day (1975) Induction/diet 1% chol, 3 months or 3% chol, 3 months normal diet 3 % chol 11 weeks 1%, 1 month 2.5 months standard diet 2% chol, 3 mo. then normal diet l%chol, 4 weeks semi-synthetic diet semi-synthetic diet, 6 months 0.75% chol 6 months 2% chol 3 months and during treatment Lipid 90 % animal lecithin crude DMPC soybean PC crude animal brain soybean (2 sources) ovolecithin Lipostabil Lipostabil Lipostabil Lexinol-Cal (IV) Asolectin (oral) Lipostabil Dose 400 mg/kg 400 mg/kg 700 mg/kg 1000 mg/kg 500 mg/kg 100 mg/kg 250 mg/kg 200 mg/kg 400 mg/kg Treatment 4 bolus injections 5 slow infusions 3-5 weekly slow infus. 4 slow infusions 5-16 slow infusions (biweekly) 12 slow infusions (biweekly) 3-7 biweekly injections 35 biweekly injections 30 injections thrice-weekly 26 biweekly injections 12 weekly injections Results 33% TC/g 100% visual 76% TC/g 68% visual 42% TC/g increased lipid deposits (visual) 48% TC/g 31% visual 22% TC/g 42% visual increased lipid deposits 40% TC/g 100% visual 19% CE 15% chol 35% visual visually less no improvement 64% TC/g 80% visual a Abbreviations: cholesterol (chol), total cholesterol (TC), cholesterol esters (CE) b % results represent reductions 3 Despite these seemingly striking results it was not until the mid 1980's that a rational explanation was formulated (Williams et al., 1984). Williams and colleagues proposed that the infusion of phospholipids might be a viable means of inducing the rapid regression of atherosclerotic plaque in humans, but that more studies involving well characterized liposomal systems were needed. The work described in this thesis examines in detail the liposomal properties required to maximize the mobilization of cholesterol in vivo, and investigates the ability of an optimized system to cause regression of plaque in cholesterol-fed rabbits. 1.2 MEMBRANES 1.2.1 Biological membranes Biological membranes are composed of lipid, protein and carbohydrate with virtually all the carbohydrate covalently bound to proteins or lipid (see Figure 1). Lipid molecules are organized in a bilayer which serves as a selectively permeable barrier to separate the cytoplasm of cells from the environment and to compartmentalize intracellular organelles which act as functional units within the cells. The lipid matrix allows considerable lateral and rotational movement of membrane components, thus providing an appropriate milieu for the function of membrane proteins (Singer and Nicholson, 1972). Membrane proteins can be divided into two main categories, integral and peripheral. Integral proteins are embedded within the bilayer matrix and usually span the membrane whereas peripheral proteins are associated with either the intracellular or extracellular membrane surfaces and do not penetrate the hydrophobic core of the bilayer to the same extent. The almost limitless combinations of these molecules give biological membranes their diversity of structure and function. 4 Figure 1 Diagrammatic Structure of a Biological Membrane Most biological membranes consist of a bi layer of lipid with proteins penetrating into either side or spanning the entire membrane. The carbohydrate moeities on lipids or proteins face the outer surface. Reproduced from Gurr and Harwood, 1991. 1.2.2 Chemical and physical properties of phospholipids and cholesterol Lipids are amphipathic molecules that are soluble in organic solvents. This highly diverse group includes: fatty acids, acylglycerols, phospholipids, glycosphingohpids and sterols. The two most common constituents of biological membranes are phospholipids (phosphoglycerides and sphingomyelin) and sterols, particularly cholesterol. In specialized membrane structures such as the skin and nerve tissues, glycosphingohpids, sterols, and fatty acids predominate. The lipids employed in this thesis are mostly phospholipids and cholesterol, therefore the more general aspects of their structure and interactions in aqueous solutions are presented. 5 Phospholipids Phospholipids are phosphorus containing molecules that can be divided into two families, one containing sphingosine and the other glycerol. The most common membrane phospholipids are sphingomyelin and the glycerophospholipids (see Figure 2). Sphingomyelin consists of a fatty acid linked via an amino group to the long chain amino alcohol sphingosine, with phosphorocholine esterified to the C-1 hydroxyl. Glycerophospholipids, on the other hand, contain two fatty acids esterified to the C-1 and C-2 hydroxyls of glycerol with a phosphate esterified to C-3. This molecule is called phosphatidic acid (PA) which is a minor component of most eukaryotic membranes (Figure 2). The bulk of membrane phospholipids are derived from PA by the esterification of different alcohols to the phosphoryl group and can be subdivided into several classes (see Figure 2) which include, phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), phosphatidylglycerol (PG) and phosphatidylinositol (PI). In addition, PG can be esterified to PA to generate diphosphatidylglycerol or cardiolipin (CL), the latter is unique to mitochondria in mammalian cells. Each category of phospholipid can contain a variety of fatty acids varying in length and degree of unsaturation. Typically, mammalian glycerophospholipids contain a saturated chain at C-1 and an unsaturated chain at C-2. The chain length can vary from 16-24 carbons, but the most common saturated fatty acids are palmitic (16 carbons, 0 double bond, written as 16:0) and stearic (18:0) whereas the most abundant unsaturated fatty acids are oleic acid (18 carbons, 1 double bond, written as 18:1) and linoleic acid (18:2). Sphingomyelin contains the hydrocarbon chain of sphingosine and usually a saturated acyl chain attached via an amide linkage. In mammalian plasma membranes, PC and PE are quantitatively the most important phospholipids, followed by sphingomyelin and PS. However, there is considerable variation between membranes, particularly in membranes that define intracellular organelles. 6 Figure 2 Structure of Phospholipids and Commonly Occuring Headgroups Shown below are (A) Sphingomyelin, (B) A glycerophospholipid, phosphatidic acid and (C) Commonly occuring phospholipid headgroups. B Name of X-OH Formula of X Name of phospholipid Water Choline Ethanolamine Serine Glycerol Phosphatidylglycerol myo-lnositol - C H J C M J N I C H J J -CH,CH,NH, H K CH£H{OH)CHfiH o I 0 CH,OCB 1 I RCOCH -CHjCH(OH)-CH: I j -O—P—O—CH, Phosphatidic acid Phoshatidylcholine Phosphatidylethanolamine Phosphatidylserine Phosphatidylglycerol Oiphosphatidylglycerol (cardiolipin) Phosphatdylinositol 7 Cholesterol Cholesterol is the most abundant sterol and is an essential structural component of eukaryotic membranes, where it can comprise up to 45 mol% of the total bilayer lipid. In addition to its structural role, cholesterol influences a variety of membrane associated events such as passive permeability, enzyme activity and possibly fusion (Demel and deKruiff, 1976; Yeagle 1985). Furthermore, cholesterol is the precursor for synthesis of bile acids and steroid hormones. The structure is shown in Figure 3A, it is an amphipathic molecule in which the cyclohexane rings and the aliphatic chain at C-17 comprise the hydrophobic region and the 3|3-OH forms a small polar headgroup. Cholesterol-phospholipid interactions In an aqueous environment most phospholipids spontaneously adopt a bilayer configuration with their hydrophilic headgroups interacting with the aqueous solution, whilst the hydrophobic acyl chains form a tightly packed organization that excludes water. This arrangement is stabilized by hydrogen bonding between headgroups and by van der Waals forces between the hydrophobic acyl chains. In a bilayer, phospholipids can exist in several states, for example a solid, "gel" state or a fluid, "liquid-crystalline" state depending on the temperature. Their acyl chains have motional freedom around each carbon-carbon (C-C) bond and can form kinks due to the formation of gauche isomers. At temperatures below the gel-to-liquid crystalline phase transition temperature (Tc) their acyl chains adopt an average conformation of minimum energy that is all trans, or fully extended, to allow maximum interchain interaction and close packing. Above Tc, the average number of gauche isomers per acyl chain increases; in turn this reduces the van der Waals interactions, shortens the acyl chain lengths and increases the distance between molecules. Thus, in the liquid-crystalline state the lipid bilayer decreases in thickness and undergoes lateral expansion owing to the larger area occupied by each individual phospholipid molecule. 8 Figure 3 Structure of (A) Cholesterol, (B) Cholesterol Sulphate and a ( Q Cholesterol Ester B HO—i 9 Figure 4 Model of Cholesterol-Phospholipid Interaction in a Bilayer A schematic representation of the alignment and interaction of cholesterol with a phospholipid in a bilayer. Interfaclal membrane region Headgroup — o — P — C I O Motion restricted Motion unaffected P-OH group 1.1 nm Rigid sterol ring 0.8 nm Flexible tail Phospholipid Cholesterol Cholesterol is thought to orient within a bilayer such that the 3|3-OH group is exposed to the aqueous surface and is aligned with the fatty acyl carbonyls of the phospholipids, although it is not clear whether hydrogen bonding occurs between these groups (Huang, 1977; Yeagle, 1985). The planar hydrophobic rings and tail interact with the acyl chains (see Figure 4), resulting in a reduction in the number of gauche conformations that occur in the first 10-12 carbons of the acyl chains thus hindering their mobility (Kawato et al., 1978; reviewed in Phillips, 1992). This ability of cholesterol to influence the packing of phospholipids is termed the "condensation effect". It is best illustrated in monolayer studies which determine the surface area occupied per lipid and provides information concerning the molecular packing of 10 the lipids. Using this method, it has been demonstrated that the area occupied in a mixed monolayer of phosphatidylcholine and cholesterol is much less than would be predicted by the sum of their individual areas (Demel and de Kruiff, 1976). Although the hydrocarbon tail only has small effects on the motion of the phospholipid acyl chains, it is essential for the expression of the condensation effect because it orients cholesterol within the bilayer so that the van der Waal interactions between sterol and phospholipid are maximized (Houslay and Stanley, 1982). In contrast to the condensing effect, the insertion of the 3J3-OH group of cholesterol into the interfacial region of the bilayer causes an increase in the distance between phospholipid headgroups and reduces their electrostatic interaction. The result is a slight increase in the freedom of movement of the headgroups as well as an increase in the hydration of the bilayer. The influence of cholesterol on the packing of phospholipid molecules in a bilayer is highly dependent on whether the temperature is above or below the Tc of the pure phospholipid (Gennis, 1989). At temperatures above Tc, cholesterol increases bilayer rigidity by ordering the liquid-crystalline lipid whereas at temperatures below Tc, it increases bilayer fluidity by disordering the solid-state lipid and preventing phospholipid acyl chains from adopting an all trans configuration necessary for close packing. These apparently opposing effects lead to a broadening, and eventual abolition, of the lipid phase transition. At approximately 50 mol % cholesterol, the gel-to-liquid crystalline phospholipid phase transition can not be detected (Ladbrooke et al., 1968). Cholesterol is recognized to preferentially interact with specific lipids and the following affinity sequence has been demonstrated: sphingomyelin»PS, PG>PC»PE (Nakagawa et al., 1979; Demel et al., 1977). In addition, within a phospholipid family with the same headgroups, cholesterol will preferentially associate with the more unsaturated acyl chains. The molecular basis for such specificity is not fully understood although the overall structure and contour of the phospholipid is believed to play an important role in maximizing van der Waals and inter-headgroup interactions. 11 1.2.3 LIPOSOMES Classification and preparation of liposomes When phospholipids spontaneously adopt bilayers in aqueous solutions they form macroscopic vesicular structures, enclosing an aqueous compartment which are commonly referred to as liposomes. Many procedures have been developed to make liposomes and these are reviewed by Hope et al., 1993 and 1986. The method of preparation influences the permeability properties, size and lamellarity of liposomal preparations. This section briefly examines the standard preparative methods and focuses on the physical characteristics of liposomes considered important for the application examined in this thesis. Multilamellar vesicles (MLV) The first model membrane systems described by Bangham et al. (1965) were generated by the hydration of a dry film of lipid, adhered to the walls of a test tube, followed by gentle agitation. Under these conditions the structures generated possess many concentric bilayers, separated by narrow aqueous compartments and are referred to as multilamellar vesicles (MLV). Liposomes of this type have diameters in excess of 1 [xm and are heterogenous in size (0.5-10 fim) and lamellarity (Figure 5). MLV can also be made by "reverse phase" techniques involving mixing lipid solutions in organic solvent with an aqueous solution to form an emulsion with subsequent removal of the solvent. Not only are MLV heterogenous in size and lamellarity, but less than 10% of their total lipid is usually present in the outermost bilayer thus limiting their use as a model for biological membranes. More lipid can be exposed in the outer bilayer following successive cycles of freezing and thawing, a process which also increases the internal trapped volume (Mayer et al., 1985). Studying membrane properties such as permeability, lipid exchange and fusion are better conducted using liposomes with a single bilayer, a homogenous size distribution and a well defined internal aqueous compartment (Figure 5). Some common unilamellar vesicle preparations are described in the next sections. 12 Figure 5 Schematic Representation of Liposome Size and Lamellarity (A) Multilamellar vesicle. (B) Large unilamellar vesicle. (Q Small unilamellar vesicle (1 ,000- 10,000 nm) AQUEOUS CHANNELS B 1 (50 - 200 nm ) ?& AQUEOUS CORE m * Wh, '&* . ^ 9b, (25 - 50 nm ) # " % # % i AQUEOUS CORE .•tfJlMftft i*i 1 % ^ % K ^ # 13 F i g u r e 6 Freeze-Fracture Electron Micrographs of LUV produced by Extrusion Vesicles were prepared from multilamellar vesicles at a concentration of 50 mg/ml by extrusion through two stacked polycarbonate filters o f various pore size. The bar represents 200 nm. 1 4 Small unilamellar vesicles (SUV) Small unilamellar vesicles (SUV) typically exhibit a size distribution of 25-50 nm and can be prepared from MLV by sonication (Huang, 1969), French press techniques (Barenholz et al., 1979) or high pressure homogenization. The size of SUV is dependent on the lipid composition and can vary from 25-30 nm for EPC systems and up to 50 nm for saturated phospholipids and cholesterol containing systems. As determined by nuclear magnetic resonance (NMR), these vesicles can have up to 70% of their total lipid in the outer monolayer owing to a small radius of curvature. This imposes geometrical constraints on lipid packing which can give rise to vesicle instability and an enhanced tendency to fuse into larger particles (Lichtenberg et al., 1981; Wong et al., 1982). Other disadvantages of SUV are the small trapped volumes and the potential for lipid peroxidation products that result from harsh sonication procedures and which can affect studies of sterol exchange (Thomas and Poznansky, 1988). Large unilamellar Vesicles (LUV) Large unilamellar vesicles (LUV) can be made from lipid solutions in ethanol or ether injected into an aqueous medium (Szoka and Papahadjopoulos, 1980), reverse phase evaporation (Szoka and Papahadjopoulos, 1978), detergent dialysis (Mimms et al., 1981) and medium pressure extrusion (Hope et al., 1993). In the first two methods, lipid monomers present in the organic solvent are hydrated as the organic solvent is evaporated or diluted. The resulting dispersions, however, are often a heterogenous mixture of oligo- and unilamellar vesicles, requiring extrusion through polycarbonate filters to obtain a homogenous preparation of unilamellar vesicles. In addition, column chromatography is usually performed to further remove remaining organic solvents. These methods are restricted by the solubility of various lipids in the organic solvent used, however, their advantages are that a high trapping efficiency can be achieved and vesicles can be prepared at high concentration. 15 For procedures employing detergent dialysis, detergents with high critical micellar concentrations (CMC) such as sodium cholate or octylglucopyranoside, are preferred otherwise removing detergent is very difficult. Lipid and detergent form mixed micelles which coalesce to form stable vesicles as detergent is gradually dialysed away (Mimms et al., 1981). The above methods generate vesicles with diameters that range from 50-200 nm, however, the processes are tedious, may not be applicable to all lipids and may contain residual solvents or detergents which can affect the properties of the bilayer. Many disadvantages of LUV preparative procedures have been overcome by the introduction of medium pressure extrusion. This technique involves forcing MLV through polycarbonate filters of defined pore size. Routinely, homogenous populations of vesicles with diameters ranging from 50-250 nm can be produced in minutes without the need for organic solvents or detergents (Figure 6). Moreover, vesicles can be produced at concentrations as high as 400 mg/ml with almost no limitation on lipid composition. The most commonly used vesicle prepared using extrusion, and favoured by many research groups, is the 100 nm LUV which is an ideal size for in vivo applications and exhibits an approximately equal distribution of lipid between the inner and outer monolayers. This type of vesicle has been used extensively in the work presented here, as it is also well suited to studies measuring sterol exchange. 1.3 FACTORS INFLUENCING CHOLESTEROL MOVEMENT BETWEEN MEMBRANES A fundamental aspect of the Singer and Nicholson Fluid Mosaic model of biological membranes is the dynamic nature of the structure. Lipid and protein components are in a constant state of motion. Phospholipids, for example, experience axial rotation, intrachain motion and lateral diffusion. Some lipids readily diffuse across bilayers and also move between different membranes (see Figure 7). This section describes the various types of motions that membrane lipid molecules participate in, and focuses on the ability of cholesterol to undergo relatively rapid intermembrane diffusion. 16 1.3.1 Lipid motions In liquid-crystalline bilayers phospholipid molecules undergo fast rotation about their long axis. The C-C bonds of the acyl chain rotate producing gauche isomerizations that cause chain fluctuations. These intrachain motions are extremely fast with a life-time of approximately 10"9 seconds. Lipids can also migrate laterally in the plane of the bilayer, with a diffusion coefficient on the order of 10 cm2/s in both biological and model membranes. This translates to a distance of approximately 2 ^m/sec. The exchange of lipid molecules between each half of the bilayer is often referred to as flip-flop or transbilayer diffusion. In biological membranes, this process is likely facilitated for phospholipids by specific translocases (Deveaux, 1991). Cholesterol, on the other hand, is thought to readily diffuse across bilayers via a passive, non-facilitated route, and how this relates to the overall process of cholesterol exchange between membranes is discussed in section 1.3.2 Intermembrane cholesterol exchange A considerable amount of research has been conducted to better understand the molecular mechanisms involved in cholesterol exchange (for review see Phillips et al., 1987). Knowledge of cholesterol dynamics in membranes is key to understanding the flux of sterol into and out of cells. Disrupting this equilibrium apppears to result in the gradual accumulation of lipid during the development of atheroma. Lipid exchange proteins have been identified that exhibit specificity for phospholipids, triacylglycerols and cholesterol esters, but no proteins have been conclusively found to catalyze the transfer of unesterified cholesterol molecules, which appears to occur through passive diffusion. Many groups have used model membranes to provide insights into the process of cholesterol equilibration between model and cell membranes (Phillips et al. 1987 and references therein). It is now well established that cholesterol can undergo rapid, spontaneous exchange between membranes, via a mechanism that does not require metabolic energy. A key question has been whether or not direct contact is required between donor and acceptor particles 17 Figure 7 Various Motions that a Lipid Molecule Undergoes in a Membrane Lipids in a bilayer exhibit several types of motion that include, the flexing of hydrocarbon chains, rotation around their long axis, lateral movement, transbi layer diffusion and efflux from the bilayer. axial rotation Cb acyl chain motion — (fluidity) phospholipid lateral diffusion cholesterol OH I A | A monomer m^ X+ diffusion T I T I (exchange) transbilayer diffusion \ head group mobility 18 or does cholesterol diffuse through the intervening aqueous gap. The two most frequently proposed models for cholesterol exchange, the collision complex model and the aqueous monomer diffusion model are discussed below. Collision complex model The transition collision complex model proposes that cholesterol molecules diffuse through an intermediate structure formed by the transient fusion of two lipid bilayers, following the collision of donor and acceptor particles (Figure 8). This model was put forward to account for the apparent dependence of the kinetics of sterol movement between membranes on the concentration of donor and acceptor particles (Ferreli et al., 1985; Phillips et al., 1987 for details). For example, when there is a low concentration of interacting particles, or when the rate of formation of the transient collision complex is limiting, the initial rate of cholesterol diffusion into the accepting bilayer is reduced. However, where there is an excess concentration of accepting particles, it is found that the rates of sterol efflux are independent of the absolute concentration of donor and acceptor particles. Figure 8 Schematic Representation of the Collision Complex Model I—+OH ffnffofffffoff^ii.fffowwotffff? 19 There are, however, several reasons why the formation of intermediate collision complexes are unlikely (Phillips et al, 1987). First, it is doubtful that a fusion complex can form in a reversible manner and this would mean that all the lipids from the donor vesicles would be transferred stoichiometrically to the acceptor particles. Second, if the reversible complex is indeed formed, then given that the lateral diffusion rates of cholesterol and phospholipids are similar it is difficult to understand why these two molecules transfer at such different rates. Finally, the juxtaposition of bilayers within distances of less than 3 nm is energetically unfavourable, owing to repulsive hydration forces. Consequently, in order for a lipid molecule to be transferred across an intervening gap of >3 nm, it would need to desorb completely from the donating membrane and hence would essentially be the same as the aqueous diffusion model described below. Aqueous diffusion model The diffusion of lipid monomers such as free fatty acids, phospholipids and cholesterol through the intervening water layer has been described by several investigators (Nichols and Pagano, 1981; Storch and Kleinfeld, 1986; Arvinte and Hildenbrand, 1984; Ferrell et al., 1985). Cholesterol molecules are thought to desorb from the donor lipid/water interface and diffuse through the aqueous space until they collide and are absorbed by an acceptor particle (Figure 9). Strong evidence to support this model includes the demonstration that cholesterol exchange between donor and acceptor membranes occurs even when the two populations are physically separated by dialysis membrane or by a two phase partition system (Bruckdorfer et al., 1984). Nichols and Pagano (1981) have derived rate constants and described the important features of the aqueous diffusion model during their investigations into phospholipid exchange. In brief, the rate at which a lipid monomer desorbs from a membrane is proportional to its concentration on the surface, whereas the rate of adsorption into an acceptor bilayer is a function of the product of free monomer concentration and the surface area of the membrane. 20 Thus when acceptor vesicles are not in sufficient excess, the rate of adsorption of cholesterol monomers in the aqueous phase into acceptor vesicles is affected by the concentrations of donor and acceptor vesicles (Phillips et al., 1987), a situation which is analogous to the conditions described for the collision model. Much of the confusion in the literature on cholesterol exchange has arisen from a failure to appreciate this point. The kinetic dependence of cholesterol exchange on both donor and acceptor particle concentration does not necessarily mean that collision between particles is required for cholesterol transfer to occur. Figure 9 Schematic Representation of the Aqueous Diffusion Model mmmmmmmmmmmim 21 When acceptor vesicles are present in excess, the rate limiting step is the desorption of cholesterol monomers from the surface monolayer of the donor (McLean and Phillips, 1981). In the transition state complex, cholesterol desorbs into the aqueous phase attached only to the bilayer via a small segment of the hydrophobic tail (see Figure 10). The formation of this complex is rate limiting owing to the energy required to move the hydrophobic cholesterol molecule into an aqueous environment (Phillips et al., 1987). Furthermore, the rate of cholesterol desorption is strongly dependent on temperature and the physical characteristics of the donor membrane. Figure 10 Energy of Activation Associated with Cholesterol Exchange A / free energy\ / I OH OH Generally the size of the donor vesicle is inversely proportional to the rate of desorption, this is because, to some extent, vesicle size dictates the degree of lipid packing within the bilayer (McLean and Phillips, 1984; Thomas and Poznansky, 1988). For example, in highly curved membranes such as those common to SUV, the molecular packing of 22 phospholipid molecules in the outer leaflet is reduced compared to that seen in larger, more planar membranes. Consequently, SUV exhibit relatively rapid rates of cholesterol exchange. In LUV systems, lipid molecules in the bilayer pack more closely, and cholesterol molecules are thermodynamically more stable leading to reduced rates of cholesterol efflux (McLean and Phillips, 1984; Thomas and Poznansky, 1988). The presence of long chain saturated phospholipid or sphingomyelin also decreases the rate of cholesterol desorption because of the greater van der Waals interactions with these lipids which act to raise the activation energy barrier (Lund-Katz et al., 1988). In addition, it has been determined that for a given lipid concentration, acceptor particle size is an important factor when considering intermembrane cholesterol exchange for the following reasons: vesicle size dictates (1) the relative diffusion coefficients of particles, (2) the number of particles and (3) the outer monolayer surface area at which cholesterol equilibration occurs (Phillips et al., 1987). Existence of cholesterol domains in membranes Membrane lipids are not randomly distributed. Plasma membranes, in particular, exhibit transmembrane lipid asymmetry with PC and sphingomyelin predominantly in the outermonolayer, whereas PE and PS tend to be located at the inner monolayer (Devaux, 1991). There is some evidence that domains of lipid may also exist in the plane of the bilayer, and in mammalian cells, there appear to be several cholesterol domains (reviewed in Schroeder et al., 1991). There is evidence that the transbilayer distribution of cholesterol is also asymmetric in plasma membranes with the sterol apparently enriched in the inner leaflet (Hale and Schroeder, 1981). Although this is inconsistent with the greater interaction between cholesterol and those phospholipids located in the outer leaflet. In addition to this, cholesterol domains have also been found to exist within the lateral plane of the bilayer. For example, segregated cholesterol-rich and cholesterol-poor areas which have been identified by histochemical labelling with filipin, a polyene antibiotic which associates strongly with cholesterol to form membrane channels dectable by freeze-fracture (Robenek et al., 1982). In addition, at least three pools of 23 cholesterol within cells have been detected by analysing the exchange kinetics using of dehydroergosterol, a fluorescent analogue of cholesterol (Nemecz et al., 1988). The recognition that cholesterol exists within lateral and transbilayer domains suggests that the rate and extent of cholesterol efflux will be intimately tied not only to lipid interactions within lateral domains, but also to the rate of cholesterol flip-flop because in order for cholesterol to transfer from one membrane to another, it can only desorb from the outer leaflet of the donating membrane. In biological membranes, it is generally agreed that transbilayer sterol movement, as detected by fluorescence and cholesterol oxidase techniques, is on a time scale of minutes whereas intermembrane exchange occurs over several hours (reviewed in Schroeder and Nemecz, 1990). Thus, it is assumed that in biological membranes, transbilayer movement of cholesterol is unlikely to be limiting in the exchange process. The rate of transbilayer movement for sterols in model membranes, however, is less clear and there is much disagreement in the literature. Comparing published studies is complicated by the fact that donor and acceptor vesicles of different sizes and compositions were used and these parameters will influence kinetics. The data demonstrates how methods and conditions of measurement can dramatically affect the observed rates of movement of sterol (reviewed in Schroeder and Nemecz, 1990). Furthermore, the difficulty in reconciling the physical location of sterol in the membrane lies in interpreting the kinetic data. For example, if it can be shown that the rate of transbilayer diffusion for cholesterol is fast compared to the rate of exchange, then two or more kinetic pools observed in exchange assays would be evidence that lateral sterol domains are involved in the process. At present there is no direct method for measuring the rate of cholesterol transbilayer movement, and it is usually estimated from the kinetics of cholesterol exchange between membranes, assuming equal C:P ratios in the outer and inner leaflet of the model membrane. It has often been rationalized that if transbilayer movement is slow then the inner leaflet sterol (30-50% of the total sterol, depending on vesicle size), should be essentially non-exchangeable. 24 Several groups have reported complete exchangeability of cholesterol using radioactive isotopes (Phillips et. al, 1987; Nakagawa, 1979; Bloj and Zilversmit, 1977; Kan and Bittman, 1991; Kan et al., 1992), chemical labelling (Dawidowicz and Backer, 1981) and cholesterol oxidase methods (Backer and Dawidowicz, 1981) under both equilibrium (exchange, but no net transfer) and non-equilibrium (net transfer) conditions. Others, using equilibrium conditions, reported the existence of a non-exchangeable pool of cholesterol which was thought to reflect sterol at the inner half of the bilayer (Poznansky and Lange, 1976). Subsequently, however, Poznansky and Lange (1978), reported that transbilayer migration rates, under non-equilibrium conditions, are much faster than under equilibrium conditions. However, more recently several investigators have reported the existence of a non-exchangeable pool detected during both equilibrium and net transfer conditions (Bar et al, 1986; Bar et al., 1987; Nemecz et al., 1988). But because of the apparently fast flip-flop of cholesterol, lateral domains have been invoked by these authors in order to explain the observed kinetics for exchange. Domains that vary in cholesterol concentration are thought to exist due to differences in lipid packing within a leaflet. The composition of these areas appears to be governed by a variety of factors including the C:P of the whole membrane, phospholipid composition and the presence of membrane proteins that require either sterol-rich or sterol-poor regions for activity (reviewed in Schroeder et al., 1991 and Rothblat et al., 1992). Given that the rate of desorption of cholesterol from a membrane into the aqueous phase is dictated, in part, by the degree of cholesterol-phospholipid interaction, then it is expected that individual domains might exhibit differences in the rate constants for desorption. For example, on the basis of lipid packing, it is predicted that in cholesterol-poor domains, cholesterol would more likely undergo desorption into the aqueous phase, resulting in a faster rate of efflux than from sterol-rich domains, in which phospholipids would be more densely packed owing to the condensing effect of cholesterol described in section The existence of lateral domains for cholesterol in vesicles was first suggested by McLean and Phillips (1982). These investigators observed that although C:P ratios were varied 25 in donor vesicles, the kinetics and rates of cholesterol exchange between donor and acceptor vesicles remained relatively unchanged. Subsequently, Bar et al. (1986) detected the existence of a non-exchangeable pool of cholesterol that could not be accounted for as inner monolayer cholesterol and suggested instead that this was likely a segregated pool of cholesterol. More recently, Nemecz and co-workers (1988), monitoring cholesterol exchange between donor and acceptor vesicles have detected the appearance of different kinetic pools of cholesterol by subjecting their kinetic data to linear regression analyses. These investigators used dehydroergosterol, a fluorescent analogue of cholesterol, to show that the non-exchangeable pool appears to be a laterally segregated domain of sterol rather than molecules trapped at the inner monolayer (Nemecz and Schroeder, 1991). Overall, one can conclude that the rate of cholesterol flip-flop is probably not rate limiting to the exchange of sterol between membranes. 1.4 LIPOPROTEIN CLASSIFICATION, STRUCTURE, METABOLISM AND FUNCTION 1.4.1 Structure and classification Cholesterol is transported in the plasma as complexes of lipid and protein, known as lipoproteins. Lipoprotein particles are composed of a series of different proteins called apolipoproteins (apo) and four major classes of lipid: unesterified or free cholesterol, esterified cholesterol, triglycerides and phospholipids. Lipoproteins are classified into several major classes, distinguished by their relative flotation densities and particle size. They include, in decreasing order of size and increasing density, chylomicrons, very low density lipoprotein (VLDL), intermediate density lipoproteins (IDL), low density lipoproteins (LDL), high density lipoproteins (HDL) and very high density lipoproteins (VHDL). Each class represents a continuum of related species that share a common apoprotein composition, but are composed of varying proportions of different lipid types thus generating a wide range of particle sizes and densities (see Table 2). With the exception of LDL, which is primarily associated with apoB, the other lipoprotein classes contain a mixture of apolipoproteins and their subsets, designated for example as apoAI, All and AIV (Table 3). 26 Table 2. Composition of the major lipoprotein classes8' Lipoprotein VLDL IDL LDL HDL2 HDL3 VHDL Protein 10.4 17.8 25.0 42.6 54.9 62.4 Cholesterol 5.8 6.5 8.6 5.2 2.6 0.3 Phospholipid 15.2 21.7 20.9 30.1 25.0 28.0 CEC 13.9 22.5 41.9 20.3 16.1 3.2 TG 53.4 31.4 3.5 2.2 1.4 4.6 aValues represent % mass bData obtained from Fielding and Fielding, 1991 Abbreviations used: cholesterol esters (CE), triglycerides (TG) Table 3. Apolipoprotein composition of human plasma lipoproteins3 Apoprotein apo AI apo All apo AIV apo B48 apo B100 apo CI apo CI I apo CIII apo D apo E Molecular weight (kDa)b 28.1 18.6 43.3 240.8 512.7 6.6 8.2 8.8 19.3 34.2 Lipoprotein HDL HDL HDL Chylomicrons VLDL, LDL VLDL, LDL Chylomicrons, VLDL, LDL Chylomicrons, VLDL, LDL HDL Chylomicrons, VLDL, LDL aData from Fielding and Fielding, 1991 bMolecuIar weights of mature proteins computed from the amino acid sequence 27 Structural analyses have demonstrated that most lipoproteins are spherical in shape and are composed of a monolayer organization of phospholipid, free cholesterol and apoproteins surrounding a neutral lipid core of cholesterol esters (CE) and triglycerides. The structure of a typical HDL particle is shown in Figure 11. Apoproteins AI, All, AIV and C on HDL are exchangeable and characterized by the presence of amphipathic helices, which are grouped into a number of distinct classes and those associated with HDL are of the Class A type (reviewed in Segrest et al., 1992). A striking feature of these helices is that they exhibit a clustering of positively charged residues at the polar-nonpolar interface while at the center of the polar face have negative amino acid residues which are thought to interact with the phospholipid headgroups and help maintain the integrity of the lipoprotein particles. Figure 11 Structure of a Lipoprotein Particle A lipoprotein particle generally is composed of a micellar organization of phospholipid, free cholesterol and apoprotein(s) surrounding a core of neutral lipids. Shown here is an HDL particle. Reproduced from Phillips and Rothblat, 1985. 28 1.4.2 Metabolic relationships between lipoproteins This section briefly reviews the transformation and metabolism of different lipoproteins and comments on their role in lipid transport (Gurr and Harwood, 1991). An awareness of the characteristics, function, and the metabolism of lipoprotein particles, as well as the enzymes that mediate these processes is fundamental to our understanding of cholesterol homeostasis in the body and the reasons why atherosclerotic lesions develop. Chylomicrons Chylomicrons function to transport lipids of dietary (exogenous) origin and are synthesized in the intestine. They contain mostly triglycerides along with small amounts of CE in their core, and just enough phospholipid and apolipoproteins (apoA) necessary to cover their surface. They exhibit very short half-lives of only two to three minutes in the circulation, during which time they likely acquire apoC and apoB, with apoCII playing a key role in activating lipoprotein lipase (LPL) which is located on the endothelial surface of capillary beds. LPL catalyzes the hydrolysis of triglycerides, releasing fatty acids which are taken up by adipocytes and muscle cells, transforming this particle into a chylomicron remnant. At this time, phospholipid, apoAs and the remaining apoC from the chylomicron are also transfered to HDL. The liver is the central organ involved in cholesterol and triglyceride metabolism and is known to sequester the remnant by receptor mediated processes and repackages the cholesterol into VLDL which are then secreted into the hepato-peripheral circulation (see Figure 12). 29 Figure 12 Metabolic Relationship Between Lipoproteins There are several major classes of lipoproteins that are found in plasma. These include the chylomicrons, VLDL, IDL, LDL, HDL and VHDL (see text). Lipoproteins undergo a series of catalytic conversions mediated by enzymes in plasma and these interrelations are depicted below. Reproduced from Gurr and Harwood, 1991. r—0-~c\>E Chylomicron synthesi: HDL synthesis A A Nascent Chylomicron Blood capillary Adipose tissue FA Muscle FA Mammary gland, etc. LIVER CELL PERIPHERAL CELL LIVER CELL 30 Very low density lipoproteins (VLDL) VLDL functions to transport triglycerides and CE of endogenous origin and are synthesized mainly in the liver and intestine. Unlike nascent chylomicrons, VLDL receive their full complement of apolipoproteins upon secretion, which include apoCII, such that upon reaching the circulation, they enter a cascade of lipolysis catalyzed by LPL, resulting in the progressive removal of triglycerides, an enrichment of cholesterol and phospholipid, and an overall increase in density until LDL particles are formed. During this transformation, apoC and apoE are lost leaving only apoB associated with LDL particles. Low density lipoproteins (LDL) LDL are the major carriers of plasma cholesterol in man, rabbit, pig and guinea pig. In most other mammals HDL predominates and can account for up to 80% of the plasma cholesterol. Delivery of cholesterol to peripheral cells occurs via the interaction of LDL with cell surface receptors (Brown and Goldstein, 1986) that recognize apoB and apoE (apoB/E or LDL receptor). A small proportion of LDL are also taken up in the liver and other tissues by endocytosis that does not involve a receptor although this is less efficient. Furthermore, there are also other receptors, for example on macrophages, that recognize modified LDL and are responsible for the degradation of LDL particles that are not recognized by normal surface LDL receptors. This pathway is thought to play a pivotal role in the pathogenesis of atherosclerosis. High density lipoproteins (HDL) HDL is believed to be the particle responsible for the removal of excess cholesterol from peripheral cells that are incapable of metabolizing cholesterol, a process which is crucial for maintaining their cholesterol balance (Miller et al., 1985; Johnson et al., 1991). It was originally thought that HDL were secreted as discoidal particles consisting of a bilayer of mainly PC with apoAI and apoE at the margins of the disc. However, unlike triglyceride-rich lipoproteins, nascent HDL have not been identified within cellular compartments. It now seems 31 most likely that HDL proteins are secreted lipid-poor or lipid-free and assume a discoidal shape as they obtain most of the lipid by the transfer from cell membranes and other lipoproteins (Fielding and Fielding, 1991). These "nascent" HDL particles have been detected in intestinal lymph, liver perfusates and in interstitial fluid (Sloop et al., 1987) and have demonstrated to be important in promoting the initial stages of cholesterol removal from cells (Castro and Fielding, 1988). Following their secretion into plasma they acquire additional surface components which include cholesterol, phospholipids and apolipoproteins, as mentioned earlier, from chylomicrons and VLDL and become HDL3 particles. HDL exists as a variety of species in plasma owing to lipid transfers is usually classified according to their apolipoprotein compositon. For example, HDL species containing only apoAI (often termed HDL2) or having both apoAI and All (HDL3) in varying proportions have been identified. In addition to the major apolipoproteins, apoAI and apoAII, a small proportion of HDL also contain other apolipoproteins (see Table 3) such as apoE (found on approximately 2% of HDL) and the apoC proteins (present on approximately 5% of HDL) which are thought to facilitate HDL particle metabolism. In plasma, HDL is often broadly divided into two subclasses, HDL2 and HDL3, which can be separated by density (see Table 2). A subfraction of HDL containing apoAI is recognized to associate with the enzyme lecithin:cholesterol acyltransferase (LCAT). LCAT works to transfer a fatty acid from PC to cholesterol on the surface of the disc-shaped HDL to form CE which then transfer into the core of the particle (Glomset, 1968). As a consequence of this redistribution of lipid, the HDL particle is transformed from discoidal to a spherical HDL2 particle. HDL2 particles are reconverted to HDL3 particles by the action of lipoprotein lipase, and hepatic lipase which decrease both core triglycerides and surface phospholipid. Cholesteryl ester transfer protein (CETP) also acts on HDL2 by reciprocally transferring CE and triglycerides to VLDL, LDL and HDL (Fielding and Fielding, 1991). When triglycerides are brought into HDL it is hydrolysed by surface lipases, resulting in a smaller, denser particle with less lipid. 32 1.4.3 Reverse cholesterol transport The process whereby unesterified cholesterol, that has accumulated in cell membranes, is removed by HDL and transported to the liver, to be metabolized, is called reverse cholesterol transport (RCT). It is important to realize that cells are bathed in interstitial fluid and much of the exchange of lipids between cell membranes and HDL occurs in interstitial fluid and lymph vessels. It is believed that HDL or HDL-like discoidal particles (rich in phospholipid and containing only apoAI, more recently termed pre-p HDL due to their electrophoretic mobility on an agarose gel), are involved in the initial stages of RCT (Castro and Fielding, 1988; Johnson et al., 1991). These discoidal HDL are excellent substrates for the LCAT reaction. The ability to separate the different species of HDL has offered new insights into the initial processes involved in reverse cholesterol transport. It has recently been demonstrated by following the metabolism of isotopic cholesterol from cell membranes that cholesterol is first detected in the small pre-p HDL species then appears in larger complexes, including particles containing LCAT (Fielding 1990). LCAT plays a key role in this pathway and catalyzes the consumption of cholesterol on the surface of HDL to form CE resulting in a chemical concentration gradient for cholesterol, that induces a further net transfer of cholesterol from non-hepatic cells, and other lipoproteins, to the site of esteriflcation. These reactions occur with a simultaneous acquisition of apoAII. Cholesterol esters were only subsequently found in HDL fractions containing both apoAI and apoAII (Fielding and Fielding, 1991). Cholesterol is readily exchangeable between the plasma lipoproteins, whereas the transfer of CE products is mediated by CETP. Hence, the transfer of CE molecules to lipoproteins containing apoB or apoE and their subsequent removal by the liver, completes part of the process of reverse cholesterol transport. More importantly, HDL is recognized by the apoB/E receptor of hepatocytes so that cholesterol it has accumulated is delivered, subsequently processed, and is either excreted as faecal sterols, utilized for the synthesis of bile acids or repackaged into VLDL for recirculation into the lipoprotein pool (Figure 12). 33 As mentioned earlier, there is accumulating evidence that the transfer of cholesterol from cells to HDL is influenced by the physical characteristics of size, charge and composition (Castro and Fielding, 1988; Fielding and Fielding, 1991) factors which in turn can influence the configuration of apoAI on the surface of HDL (Sparks et al., 1992; Rothblat et al., 1992). On the whole, it is currently believed that the initial events involved in net cellular efflux of cholesterol from cell membranes to HDL is influenced by : (1) the distribution of cholesterol between cholesterol-rich and cholesterol-poor membrane domains, (2) the concentration gradient of free cholesterol, reflected as the difference in C:P ratios between plasma membrane and acceptor particle and (3) the conformation and transient interactions between the amphipathic helices of HDL apolipoprotein and the donor plasma membrane (reviewed in Rothblat et al., 1992). 1.4.4 Cholesterol homeostasis As outlined in the previous sections the transport system of cholesterol is complex and the maintenance of cholesterol homeostasis in the body involves a delicate equilibrium between free sterol in cells and in plasma lipoproteins. Delivery of cholesterol to cells occurs via the receptor mediated LDL pathways, and by passive exchange of sterol between plasma membranes and lipoproteins (reviewed in Brown and Goldstein, 1986). Once bound to the LDL receptors, particles are endocytosed and their lipid components are degraded by lysosomal enzymes (Figure 13). The cholesterol generated after CE hydrolysis accumulates in a pool that activates acyl CoA:cholesterol transferase (ACAT) to esterify excess cholesterol. The resulting cholesterol esters are stored in cytoplasmic inclusions and become substrates for neutral cholesteryl ester hydrolase, and enter into a continual cycling of cholesterol and CE (see Figure 13). Only tissues that produce steroid hormones and bile acids can metabolize cholesterol. In order to prevent the intracellular accumulation of excess free sterol, cholesterol diffuses to the plasma membrane where the cholesterol content is highly regulated and any excess 34 cholesterol is removed by HDL and its subpopulations. In cells that can metabolize cholesterol, such as hepatocytes, cholesterol produced from CE hydrolysis is incorporated into the endoplasmic reticulum where it inhibits p-hydroxy-p-methylglytaryl-coenzyme A reductase (HMG-CoA reductase), the rate limiting enzyme of cholesterol biosynthesis. Cellular levels of cholesterol are tightly controlled and are maintained by changes in the number of cell surface LDL receptors and by regulating HMG-CoA reductase activity. Recently, Spady and co-workers (1985a) have shown that these two complementary processes do not act simultaneously, but instead work independently to exert their control over cellular cholesterol levels. It appears that hepatocytes can respond to changes in cholesterol availability and act primarily by regulating cholesterol synthesis. Only when the capacity of HMG-CoA reductase regulation is exceeded is the LDL pathway altered. For example, it has been demonstrated that excess LDL given to rats results in a suppression of HMG-CoA reductase activity and hence decreased cholesterol synthesis (Spady et al., 1985b). However, if cellular cholesterol levels do not return to normal levels then hepatocytes further compensate by secreting the excess sterol via enhancing the rate of output of apoB containing lipoproteins (Dietschy, 1990). These observations may have important ramifications to the application of liposomes described in this thesis, with respect to how the liver deals with the excess liposomal cholesterol and this will be discussed later. 35 Figure 13 Cholesterol Homeostasis within a Cell cholesterol LDL 36 1.5 PATHOGENESIS OF ATHEROSCLEROSIS Cardiovascular disease is caused by complications that arise from pathological changes in blood vessels particularly those associated with the heart and brain. This disease is the leading cause of death in North America of which most cases are linked to atherosclerosis. Atherosclerosis is the result of an accumulation of lipid and fibrous tissue in the vessel wall leading to the occlusion of arteries. Although it is difficult to point to a single process responsible for the progressive development of atherosclerosis, owing to the multiplicity of events involved, it is increasingly evident that persistent, elevated plasma cholesterol levels, particularly LDL levels, is a major risk factor (Grundy, 1990a). In addition, it is likely that the balance between LDL delivery and HDL removal of cellular cholesterol has been disrupted. The shift in equilibrium is not dramatic and the accumulation of cholesterol, cholesterol esters and phospholipid, which make up the bulk of atherosclerotic plaques, occurs over several decades (Small, 1988). An initial event thought to precipitate the atherosclerotic process, involves an injury to the endothelium which results in an increased trapping of lipoproteins in the damaged area and the appearance of specific adhesive glycoproteins on the surface of endothelial cells (Ross, 1993). This causes a host of events which can lead to thrombus formation and the attachment and migration of monocytes and T-lymphocytes to and in between cells. This rearrangement and migration of cells, is further influenced by growth factors and chemoattractants that are released by the damaged endothelium, adherent leukocytes and possibly the underlying smooth muscle cells (Ross, 1993; Schwartz, 1991). As the process continues, the migrating cells penetrate beneath the arterial surface where the monocytes deferentiate into macrophages and accumulate excess LDL lipid (Figure 14). As a result, the cells become saturated with cholesterol, promoting the conversion of intracellular sterol to cholesterol esters by ACAT and take on a distinct morphology; bloated with lipid droplets they are commonly referred to as foam cells (Small, 1988). These lipid-rich cells, together with the 37 accompanying lymphocytes, take on the appearance of a fatty streak which can even be detected in children in western societies. Over time some foam cells appear to undergo a transition, forming intermediate lesions that involves the reconversion of excess intracellular cholesterol esters to cholesterol. This in turn leads to the formation of cholesterol crystals and the onset of necrosis and plaque formation (Small, 1988). With continued cell influx, connective tissue formation, lipid accumulation, proliferation and necrosis, the lesions increase size as well as complexity and are distinguished by their lipid rich and fibrous nature. The thickening of the intima results in reduced blood flow which is the hallmark of atherosclerotic disease (Small, 1988). It is thought that during the development of coronary atherosclerosis, a thrombosis at the plaque surface, coupled with a vulnerability of soft, lipid-rich plaques to rupture, can lead to the blockage of small coronary arteries and result in myocardial infarction (Falk, 1992). 1.6 FACTORS THAT INFLUENCE THE FATE OF LIPOSOMES AND LIPOSOMAL LIPID IN VIVO Much of our present knowledge concerning the in vivo fate of liposomes has come from research into the use of liposomes as drug delivery vehicles (Poste et al., 1984; Patel, 1992; Roerdink et al., 1981). Gregoriadis and Ryman (1972) first demonstrated that intravenously injected MLV are cleared in a biphasic manner with the liver and spleen being the major sites of deposition. Thereafter, numerous studies in a variety of animal models and in man have demonstrated similar clearance patterns for LUV and SUV of different composition and dose (for reviews see Gregoriadis, 1988 or Hwang and Beaumier, 1988). Although the reasons for the biphasic nature of clearance are as yet unclear, it is generally accepted that the structural anatomy of the microcirculation and the interaction of liposomes with at least two different groups of plasma proteins mediates their clearance. These include the "opsonins" and the plasma lipoproteins (Bonte and Juliano, 1986; Patel, 1992). 38 Figure 14 Postulated Steps Leading to the Pathogenesis of Atherosclerosis / >, -J Reproduced from Grundy, 1990b 39 The purpose of this section is to describe the fate of liposomes and their lipid, since it is apparent that in order to account for losses of arterial wall cholesterol caused by infusions of liposomes, cholesterol redistributed into plasma and sequestered by vesicles must be either excreted or repackaged in a form that is non-atherogenic. 1.6.1 Role of microcirculation in liposome clearance Structural differences in the anatomy of the vessel walls of different tissues and organs plays a crucial role in the clearance of liposomes administered intravenously. There are three types of capillaries, the continuous, fenestrated and sinusoidal and these are shown in Figure 15. From a mechanical standpoint the continuous and fenestrated capillaries are impermeable to even the smallest liposomes (Poste et al., 1984). However vesicles of approximately 0.1 nm or less can escape the vasculature at sinusoidal capillaries of reticuloendothelial system (RES) tissues that have endothelial gaps of 0.1-0.2 urn and also generally either lack a basement membrane as in the liver or is interrupted as in the spleen, and bone marrow. 1.6.2 Recognition by phagocytic cells It is not inherently obvious why liposomes, which are composed of naturally occuring phospholipids and cholesterol, should be recognized as foreign particles by cells of the immune system. Evidence now suggests that their size and the adsorption of opsonins such that fibronectin, immunoglobulins and complement proteins, mark liposomes to be cleared by circulating monocytes (Gregoriadis, 1988). Importantly, tissue fractionation studies have established that the "resident" macrophages of the RES tissues lining the sinusoids of liver, spleen and bone marrow are the primary cell type that mediate liposomal clearance. The Kupffer cells of the liver likely account for approximately 80% of this activity (Patel, 1992). In general, this is true of all liposomes despite different compositions and markedly different clearance rates. 40 Figure 15 Anatomy of a Vessel Wall Schematic representation of the cells that line vessel walls of different types of capillaries. Endothelial cells (e). Basement membrane (b). 41 Factors known to influence the degree of opsonization of liposomes in vivo include, the surface charge and packing of the lipid bilayer as dictated by phospholipid headgroups, vesicle size, acyl chain composition and cholesterol content (Scherphof et al., 1984). Vesicles with loosely packed bilayers are more susceptible to opsonin adsorption, thus mediating the attachment and uptake of liposomes by specific receptors on phagocytic cells (Scherphof et al., 1984; Moghimi and Patel, 1988a). The surface charge of vesicles has also been demonstrated to play a role in mediating the non-specific adsorption of IgG and complement proteins, which may mark vesicles for clearance via macrophages that exhibit specific receptors for the Fc portion of the IgG molecule, for the complement protein C3, and for fibronectin and other extracellular matrix components (reviewed in Senior, 1987; Chonn et al., 1991). The role of scavenger receptors on macrophages in mediating the uptake of liposomes has also been described (Nishikawa et al., 1990). Furthermore, it has been suggested that serum may contain organ-specific opsonins that selectively enhance liposome uptake by macrophages of specific RES organs such as liver and spleen (Moghimi and Patel, 1988b, Moghimi and Patel, 1988c). Although the liver macrophages (Kupffer cells) represent the major site of liposome accumulation in vivo, SUV can access hepatocytes as they are capable of diffusing through the sinusoidal pores (Roerdink et al., 1981; Poste et al., 1982; Roerdink et al., 1984; Spanjer et al., 1986). 1.6 J Role of lipoproteins and apolipoproteins Apolipoproteins apo Al, All, AIV, B, C and E are known to associate with the surface of liposome membranes and may serve as opsonins for their uptake by macrophages and cells of the liver (Tall and Green, 1981; Guo et al., 1980; Williams and Scanu, 1986; Mendez et al., 1988). In particular, liposomes bearing apoE have been shown to compete for (3-VLDL receptor sites in cultured macrophages that recognize apo E (Williams et al., 1987). Furthermore, the adsorption of apolipoproteins have been suggested to play a role in hepatocyte mediated uptake via the apoB/E receptor (Williams et al., 1986; 42 Bisgaier et al., 1989). This pathway however, likely plays a minor role in liposome clearance. Recently, Williams and colleagues have demonstrated that liposome clearance rates are not significantly different in normal and in Watanabe Heritable Hyperlipidemic (WHHL) rabbits which lack apo B/E (LDL) receptors (Williams et al., 1988). These observations that liposomes administered intravenously are cleared by the liver creates a unique situation wherein cholesterol accumulated by liposomes, during their residence in the circulation, can be processed for excretion or redistribution. The question that now remains is how do vesicles induce the regression of atherosclerosis. 1.7 LIPOSOME-INDUCED REGRESSION OF A THEROSCLEROSIS 1.7.1 Vesicle-cell interactions Vesicles have been demonstrated by several groups to promote free cholesterol efflux from a variety of cells (Phillips et al., 1987 and references therein) and generally a C:P <1 in vesicles stimulates cellular cholesterol efflux whereas ratios >1 induce net uptake. Although vesicles are able to promote the efflux of cholesterol from plasma membranes of cells grown in culture, the ability of phospholipid vesicles to induce efflux of intracellular cholesterol and cholesterol ester remains unclear and may be a function of cell type. Ho et al. (1980) were unable to demonstrate vesicle induced mobilization in mouse peritoneal macrophages, although incubations with HDL induced significant reductions of cellular CE. In contrast, Yau-Young and colleagues (1982) using J774 macrophages and Johnson and coworkers (1990) employing rat Fu5AH cells have demonstrated that phospholipid vesicles are able to reduce CE deposits. Furthermore, Aviram and colleagues (1988) have recently shown that Intralipid (an emulsion of triglycerides and phospholipid, used clinically for parenteral feeding), which forms both chylomicron-like particles and liposomes upon dilution in blood, can also induce significant reductions in cellular stores of cholesterol and CE in vitro. These conflicting results raise questions about the type and concentration of vesicles necessary to promote cellular CE clearance, and suggests that other factors influence cholesterol removal from cells. Indeed, 43 HDL promotes the efflux of cholesterol from cells more efficiently than liposomes. The rates of efflux have been measured in the order of minutes for HDL, but several hours for SUV (Johnson et al., 1991). Interestingly, Arbogast and colleagues (1976) in attempting to understand the mechanisms that underlie HDL induced cholesterol efflux have demonstrated that the combination of HDL and PC SUV promotes a dramatic decrease in the cholesterol ester content of cells. This suggests that important interactions between vesicles and HDL are likely occuring such that they act synergistically to promote efflux of sterol from cells. 1.7.2 Vesicle-HDL interactions It is known that vesicle-HDL interactions result in the net loss of HDL apolipoproteins to vesicles and in the accumulation of excess phospholipid from the outer monolayer of vesicles to HDL by the mechanism of phospholipid-apolipoprotein substitution (Tall and Green, 1981). Furthermore, it has been demonstrated in vitro that the more dense subpopulations of HDL (i.e. HDL3) show a greater capacity to incorporate phospholipid (PL) which is reflected as an increase in the PL/protein ratio in these particles (Tall et al., 1986). Vesicle size (Scherphof and Morselt, 1984), phospholipid composition (Scherphof et al., 1979) and PL/HDL ratios are known to influence the degree of structural alteration that vesicles and HDL undergo when these systems are incubated together in vitro (reviewed in Tall et al., 1986). Vesicle size dictates the curvature of the bilayer which in turn influences lipid packing and hence the extent of phospholipid exchange and apolipoliprotein adsorption. In addition, size also influences the surface area exposed for these lipoprotein/vesicle interactions to occur. It has been shown in vitro that liquid-crystalline SUV, composed of phospholipid only, become completely assimilated into the lipoprotein pool in the presence of excess HDL, whereas MLV are more resistant (Scherphof et al., 1979). At higher ratios >2:1 PL/HDL, vesicles with adsorbed apolipoproteins, discoidal particles (PL/apoAI/apoAII), and phospholipid-enriched HDL are generated (Tall et al., 1986). The degree of saturation of the acyl chain and the cholesterol content of vesicles similarly influence lipid packing and 44 consequently the degree of vesicle-HDL interaction. For instance, vesicles containing phospholipids composed of saturated long chain fatty acids are less susceptible to assimilation and phospholipid transfer (Scherphof et al., 1979) although this is highly dependent on temperature (Chobanian et al., 1979; Tall and Small, 1977). In addition, the titration of increasing amounts of cholesterol into vesicles decreases the net transfer of phospholipid into HDL (Tall, 1980, Guo et al., 1980). There are fragmentary reports on the exchange of cholesterol between vesicles and lipoproteins (Phillips et al., 1987 and references therein), but these studies are complicated because of the assimilation and extensive molecular exchange that occurs between these particles. However, the rates of exchange from lipoproteins to LDL have been determined and found to be fastest for HDL (Phillips et al., 1987 and shown below in Table 4). This is most likely attributed to the small size of HDL and the availability of cholesterol, as well as the nature of the phospholipid matrix in these systems. Table 4. Approximate half-times for cholesterol exchange from donor particles8 Particle (Donor) HDL2 HDL3 LDL SUV (EPC)b LUV (EPC)b Particle radius (nm) 3.9 5.9 9.6 approx. 40 approx. 100 Half-times for exchange (min) 2.9 4.2 45 96 240 aData from Phillips et al. (1987), Sterol exchange from donor lipoproteins to LDL acceptors bData from Chapter 2, Sterol exchange measured from donor particles to EPC acceptors 45 Figure 16 Proposed Mechanism of Liposome Induced Regression of Atherosclerosis The ability of repeated infusions of phospholipid vesicles to cause the reversal of atherosclerosis is believed to be due to the following reasons. Liposomes can act as a "sink" for cholesterol from several sources. In addition, their interactions with lipoproteins, particularly HDL, results in the formation of cholesterol-poor phospholipid-rich HDL that further scavenge and "shuttle" more cholesterol to circulating liposomes. Furthermore, owing to the adsorption of apolipoproteins onto the surface of vesicles, these are transformed into lipoprotein-Iike particles that might also enhance cellular cholesterol efflux. Liposomes are cleared mainly by the fixed macrophages located in the liver and eventually this lipid load is transfered to hepatocytes and is metabolized and excreted. excretion Choi-depleted HDL and lipoprotein-Iike particles bile t o PERIPHERAL TISSUE • m ^ 46 1.7.3 Mechanism of action By surveying and building on the existing information available, Williams and coworkers (1984) proposed that these liposome-HDL interactions give rise to structural changes in HDL which enhance the ability of lipoproteins to promote cholesterol efflux in vivo. It has been suggested that phospholipid infusion enhances the mobilization of extrahepatic cell cholesterol and promotes the regression of atherosclerosis for the following reasons (see Figure 16). First, phospholipid liposomes possess an initial C:P of 0:1 and thus, when given intravenously, can act as a thermodynamic sink for cholesterol and reduce the chemical potential of sterol in plasma thereby promoting cellular cholesterol efflux. Secondly, liposomes appear to preferentially interact with lipoproteins, particularly HDL, to deplete HDL of cholesterol while also donating some liposomal phospholipid. These interactions result in cholesterol-poor, phospholipid-rich HDL which are primed to scavenge cholesterol from accessible peripheral tissues, including the arterial wall and to transfer it to liposomes or other targets. Finally, a portion of the infused vesicles may be transformed into discoidal particles with adsorbed apolipoproteins from HDL, and become assimilated into the lipoprotein pool as HDL-like particles, capable of entering the interstitial space to promote cellular cholesterol efflux. 1.8 THESIS OBJECTIVES It was clear from a critical evaluation of the liposomal systems used in earlier regression studies that in order to develop liposomal therapy for use in the management of human atherosclerosis, injection grade, well characterized, vesicle preparations exhibiting optimum cholesterol mobilizing properties needed to be identified and tested. This pursuit has been aided greatly by an increased in recent years of the factors in recent years which influence vesicle clearance, cholesterol equilibration across membranes and lipoprotein metabolism. Furthermore, technological advances in the purification and manufacture of phospholipids in large quantities, the ability to prepare vesicles of varying size and composition, on a large scale, 47 and the widespread clinical application of liposomes as intravenous drug carriers has made this goal achievable. The studies in this thesis were aimed at providing pre-clinical groundwork necessary to pursue the development of liposomal infusions for the treatment and management of atherosclerosis in humans. This task was approached by initially exploring the rate and extent of cholesterol exchange, using well characterized vesicles. The information obtained was used to determine optimal physical properties of vesicles required to promote cholesterol mobilization in vivo, with the goal of identifying a preparation of liposomes with the greatest ability to reverse atherosclerosis. The studies examining the factors that influence sterol exchange between membranes were carried out in donor and acceptor model membranes and Chapter 2 describes a comparison of the kinetics of exchange of cholesterol and cholesterol sulphate as a function of vesicle size and curvature. In chapter 3, the influence of vesicle size and composition on the rate and extent of cholesterol accumulation and mobilization in vivo were studied in a mouse model. Finally, Chapter 4 examines the efficacy of repeated infusions of an optimized liposomal preparation, by measuring the regression of atherosclerotic lesions in treated, cholesterol-fed rabbits. 48 CHAPTER 2 THE KINETICS OF EXCHANGE AND NET FLUX OF CHOLESTEROL AND CHOLESTEROL SULPHATE BETWEEN LIPOSOMAL MEMBRANES 2.1 INTRODUCTION 2.1.1 An overview This thesis examines the therapeutic use of liposomes for the treatment of atherosclerosis. As described in the introduction, the rationale for such an application arises from the tendency for cholesterol to diffuse from cell membranes and lipoproteins into liposomal bilayers. Consequently, this chapter examines some aspects of the cholesterol exchange process and how these relate to the optimization of a model membrane system that will enhance RCT. The movement of cholesterol between various liposomal systems has been measured and compared to cholesterol sulphate, which is less abundant than cholesterol, but widely distributed in plasma membranes, lipoproteins and skin (Bleau et al., 1974). The reasons for comparing the kinetics of cholesterol sulphate exchange and net flux to those of cholesterol are two fold. First, as reviewed in the following section, there are several models describing the exchange kinetics observed for cholesterol and it is not clear whether the transbilayer movement of cholesterol between model membranes is rate limiting when considering the net efflux of sterol from a donor bilayer to excess acceptor. This is because of the confusion in the literature over the rate of cholesterol flip-flop (Schroeder and Nemecz, 1990). Cholesterol sulphate, on the other hand, contains a negatively charged sulphate moiety (see Figure 3B), therefore it was predicted that the presence of the charge would reduce the energy required for this sterol to desorb from a bilayer surface and therefore enhance the flux from donor to acceptor compared to cholesterol. However, the negative charge should inhibit transbilayer movement owing to the increased energy required to move it through the 49 hydrophobic interior of the membrane (Kan et al., 1992). Consequently, one should observe a clear separation between the kinetics of intermembrane exchange and transbilayer diffusion, in contrast to that observed for neutral cholesterol. Second, although quantitatively less important than cholesterol, the sulphate derivative is found in erythrocytes, sperm and myelin membranes (Bleau et al., 1974) as well as the epidermis and stratum corneum (Wertz, 1992). The physiological role of this molecule remains unclear and to our knowledge there are no published studies comparing the exchange kinetics of this sterol to those of cholesterol. Examining the dynamics of cholesterol sulphate in model membranes may help determine a biological function for this interesting molecule as well as help us better understand the exchange kinetics of cholesterol itself. 2.1.2 Cholesterol exchange and net efflux The kinetics of cholesterol exchange has been studied extensively using liposomes, but conflicting results have been obtained (Phillips et al., 1987; Schroeder et al., 1991 and references therein). The kinetic analyses have been used to construct models describing the distribution of cholesterol pools within a membrane and the exchangeability of sterol in each pool. Essentially, the literature data falls into three categories which are depicted schematically in Figure 17. Plot A is observed in experimental systems in which it appears that all membrane cholesterol behaves as a single pool, and complete exchange is observed. Some researchers, however, have found that there appears to be a non-exchangeable pool (plot B) in which a small proportion of cholesterol appears to be thermodynamically "locked" within the bilayer. The third category, plot C, describes exchange and net flux kinetics in which two exchangeable pools of cholesterol are observed. This type of result has been interpreted as representing outer monolayer and inner monolayer sterol pools or laterally segregated areas of the membrane that exhibit different C:P ratios and therefore different rates of exchange. One problem with trying to compare published data is the variety of model membrane systems employed. The size and composition of donor vesicles can greatly influence the rate 50 and extent of cholesterol exchange. In order to minimize these differences we have compared literature data in which only SUV were used. This is because SUV represent "limit size" vesicles, generally made by the same process (sonication) and should represent a vesicle population that exhibits similar physical characteristics, that can be consistently generated by all research groups. Although MLV are just as readily made, they are heterogenous in size and lamellarity, limiting their use for cholesterol flux studies. LUV provide a better model membrane preparation for studying sterol exchange because they exhibit an almost equal distribution of lipid in their inner and outer leaflets and possess a more planar membrane surface than SUV. However, LUV used in early studies were generated by a variety of different procedures including detergent dialysis and reverse phase evaporation. It is possible that residual contaminants could affect the rates of cholesterol exchange (Thomas and Poznansky, 1988). Furthermore, these procedures often produce LUV that exhibit a broad size distribution. Figure 18 represents data from eleven different studies, plus the data generated in this chapter, using SUV composed of phospholipid in the liquid crystalline state at the experimental temperature, and highlights the enormous variation in the kinetics of cholesterol exchange. In all examples, at least a 10 fold excess concentration of acceptor membranes were used. Consequently, if cholesterol exchange is close to equilibrium then only 9% of the total sterol present in donors should remain. Several groups, (Dawidowicz and Backer, 1981; Backer and Dawidowicz, 1979; Kan and Bittman, 1991; McLean and Phillips, 1984; McLean and Phillips, 1981; Nakagawa et al., 1979; Bloj and Zilversmit; 1977) have shown that movement can be modelled by a single exponential and report complete exchangeability of cholesterol from donors. These kinetics were observed under equilibrium (exchange) and non-equilibrium (net transfer) conditions. In contrast, other groups have noted the existence of a non-exchangeable pool of cholesterol representing approximately 20% of the total that could be detected only under equilibrium conditions (Poznansky and Lange, 1976) or under both net transfer and equilibrium conditions (Bar et al., 1986). More recently, it has been reported that exchange 51 data may be best described as a function of two exponentials (Nemecz et al., 1988). It has been argued by Schroeder and colleages (1991) that the biphasic kinetics observed in these studies is not due to a slowly exchanging pool at the inner monolayer, but instead due to laterally segregated pools of cholesterol, the sizes of which are dependent on the sterol content of the donors. Figure 17 Schematic Representation of Three Categories of Sterol Exchange Observed in the Literature 100 CO •_ O c o O) c c CO E o o 1 _ CD +•• CO Arbitrary units of time 52 Figure 18 Literature Summary of the Kinetics of Cholesterol Exchange Data available from the literature as the percent cholesterol remaining in donor membranes following the incubation with excess acceptors is plotted as a function of time. In essence, data was compiled from studies monitoring the rate of cholesterol flux from cholesterol containing PC donor SUV to acceptor vesicles under both equilibrium (no net transfer) and non-equilibrium (net transfer) conditions. O c o •o O) IB E o i_ o © 0) Time (hours) • - 1 Dawidowicz and Backer, 1981 A — 2 Backer and Dawidowicz, 1979 • o — 3 Kan and Bittman, 1991 + - 4 McLean and Phillips, 1984 A - 5 McLean and Phillips, 1981 + — 6 Nakagawa et al., 1979 - v - 7 BIoj and Zilversmit, 1977 — o — 8 Poznansky and Lange, 1978 - • - 9 Bar et al., 1986 — •— 10 Baretal., 1986 - • — 11 Nemecz et al., 1988 — • — 12 Data from this Chapter 53 Since there is no direct method for measuring the rate of cholesterol transbilayer movement, it is usually estimated from the rate of sterol exchange between membranes. It is assumed that cholesterol is equally distributed across the bilayer such that the C:P ratio at the inner and outer monolayers is equal. If the kinetics of exchange can be modelled by a single exponential, then the rate of transbilayer movement must be at least as fast as the rate of cholesterol exchange. However, when more than one exponential is found, the slower rate can be attributed either to an inner monolayer pool, rate limited by the diffusion of cholesterol to the outer leaflet or domains which exhibit different desorption rates from the surface. In this chapter, we have employed well-defined, unilamellar liposomes as well as multilamellar systems to measure intermembrane cholesterol movement under non-equilibrium conditions (net flux), which mimic the situation for the in vivo application being studied. Furthermore, to help understand the kinetics obtained for cholesterol flux we have also studied the movement of cholesterol sulphate, a molecule that would be expected to exhibit rate limiting, transbilayer diffusion. 54 2.2 MATERIALS AND METHODS 2.2.1 Materials [l,2-3H]-Cholesterol, [Cholesteryl-l,2-3H(N)]-Cholesteryl hexadecyl ether ([3H]-CHE) and [14C]-Cholesteryl hexadecyl ether ([14C]-CHE)were obtained from Du Pont New England Nuclear (Boston, MA). [l,2-3H]-Cho!esterol sulphate was prepared as described in Methods below. Egg phosphatidylcholine was purchased from Princeton Lipids (Princeton, NJ) and distearoyl phosphatidylglycerol (DSPG) from Avanti Polar Lipids (Birmingham, AL). Cholesterol, cholesterol sulphate (CS), iV-palmitoyl dihydrolactocerebroside (Cer), the galactosyl-binding lectin from Ricinus communis (RCA-120), [4-(2-hydroxyethyl)]-piperazine ethane sulfonic acid (HEPES) and DEAE-Sepharose CL-6B were obtained from Sigma Chemical Company. All solvents, chemicals and thin-layer silica chromatography plates were of analytical grade and were from BDH Chemicals (Vancouver, B.C.). 2.2.2 Synthesis of [l,2-3H]-ChoIesteroI sulphate from [l,2-3H]-Cholesterol [l,2-3H]-Cholesterol sulphate was synthesized from [l,2-3H]-cholesterol (specific activity 58 Ci/mmol) according to the method described by Williams et al. (1983) with the following modifications. Briefly, radiolabeled cholesterol dissolved in ethanol was dried down, mixed with a 100 fold excess of sulfamic acid in 2 ml of dry pyridine in a 16x100 mm pyrex tube and heated at 95-100 °C for 2 hours in a boiling water bath. A parallel reaction with unlabelled cholesterol was also carried out. After the reaction mixture had cooled, 10 ml of methanol was added to extract the lipids from unreacted sulfamic acid and was placed in a round bottom flask. Two additional methanol extractions were performed, added to the flask and the solvents were removed by evaporation. Subsequently, approximately 20 ml of toluene was added to the flask to aid the removal of any remaining pyridine. The CS generated was isolated from unreacted cholesterol by preparative thin layer chromatography (TLC) using ether/methanol/ammonium hydroxide (9:1.3:0.4 v/v) as running solvent. The area of the plate which co-migrated with the CS standard was extracted with benzene/methanol (70:30 v/v). 55 The sodium salt of CS was generated by solubilizing the dry lipid film with an acidified (pH 2) distilled water/chloroform/methanol (1:1:2.1 v/v) solution. The pH of this one phase mixture was raised to pH 7 by titration with NaOH. Thereafter CS was extracted from the lower phase of the mixture following the addition of the appropriate amounts 400 mM NaCl and chloroform to generate two phases according to the Bligh and Dyer (1959) procedure. 2.23 Preparation of vesicles Donor vesicles of EPC/Sterol/Cer (75:10:15 mol ratio) with trace [3H]-CS or [3H]-cholesterol and acceptor vesicles of EPC labelled with [14C]-CHE as a non-exchangeable marker, were taken from lipid stock solutions in benzene/methanol (70:30 v/v) and Iyophilized to ensure homogenous mixing of all lipid components. A second set of donor vesicles, of the same composition but labelled with [3H]-CHE were made for use in parallel experiments designed to monitor the efficacy of vesicle separation after each incubation. Yet another preparation labelled with [3H]-CHE was used in ion exchange chromatography. These experiments demonstrate the loss of surface negative charge in donor vesicles during incubations with acceptors (see Methods below). After lyophilization, the lipid mixtures were hydrated to a concentration of 10 mM with 150 mM NaCl, 20 mM HEPES, pH7.4 and vortexed to generate MLV. LUV and oligolamellar vesicles were prepared by extruding the MLV ten times through two stacked polycarbonate filters of 100 nm and 400 nm pore size respectively, using a water-jacketed thermobarrel Extruder (Lipex Biomembranes, Vancouver, B.C.). SUV of approximately 40 nm diameter were prepared by sonicating MLV for two 15 min cycles using a Branson tip sonifier. The SUV were centrifuged at 10,000 g for 30 min in an Eppendorf benchtop centrifuge to remove titanium fragments originating from the sonicator tip. 56 2.2.4 Determination of vesicle diameters The diameters of vesicles generated by sonication and extrusion procedures were determined by quasi-elastic light scattering (QELS) analyses utilizing a Nicomp Model 370 submicron laser particle sizer (Pacific Scientific, MD) equipped with a 5 mW Helium-Neon Laser. The Nicomp QELS measures fluctuations in light scattering intensities due to vesicle diffusion in solution. The calculated diffusion coefficient is used to obtain the average hydrodynamic radius and thus the mean diameter of vesicles. Table 5 shows the diameter of the three vesicle preparations expressed as the mean ± standard deviation. For reasons of clarity, throughout this chapter and the remaining thesis, vesicles prepared by extrusion are referred to by the filter pore size used in their preparation. Table 5. Vesicle Preparations Used in Sterol Flux Studies. Liposome SUV MLV400 LUV100 Filter pore size N.A. 400 nm 100 nm Method of preparation sonication extrusion extrusion Vesicle diameter (nm) 44.2 ± 2 244 ± 42 119 ± 3 57 2.2.5 Measurement of pH]-sterol exchange The rates of exchange of the [3H]-sterol from donor to acceptor vesicles at 37°C were measured by separating the two vesicle populations at different time intervals and after quantitating the level of radioactivity in the acceptors. Separation was achieved using the lectin binding method as described by Kan et al. (1992) with the following modifications. Briefly, donor vesicles containing sterol and iV-palmitoyl dihydrolactocerebroside were mixed with a 10 fold excess concentration of acceptors, 50 \tl aliquots were placed in 1 ml Eppendorf tubes and diluted to 200 ^1 total volume with HBS. At specific time intervals, tubes were removed, 50 pig of lectin in phosphate buffer was added, vortexed and the mixtures were centrifuged for 3 min at room temperature to separate the lectin-aggregated donors from EPC acceptors. Aliquots (100 fil) of the supernatant containing the acceptors were transfered to scintillation vials and 5 ml of Pico-FIuor scintillant was added prior to counting for 2 min in a Beckman LS 3801 liquid scintillation counter equipped with a [3H]/[14C] dual label counting program. The recovery of acceptor vesicles after their separation from donor vesicles was typically > 85% and was monitored by measuring the [14C]-CHE counts present in the labelled EPC acceptor vesicles. Separation of donor and acceptor vesicles was assessed by using, the non-exchangeable markers [3H]-CHE in donor vesicles and [14C]-CHE in acceptors run in parallel experiments. During earlier time points, it was noted that essentially all donor vesicles were precipitated after the addition of lectin and following centrifugation. However, at progressively longer incubation times, for example >16 hours, approximately 10% of the donors (up from <2% at shorter time points) could be detected in the supernatants of [3H]-CHE donors used in parallel experiments, indicating some exchange of cerebroside had taken place. The fraction of labelled cholesterol appearing in the acceptor LUV after each incubation interval was determined by accounting for donor spill over into the supernatant by correcting [3H]-sterol counts in sterol flux experiments with [3H]-CHE counts in parallel experiments. It is important to stress that this type of correction is possible because CHE does not exchange on 58 the timescale of these experiments. Therefore, the fraction of [3H]-sterol transfered to acceptor in the supernatant after centrifugation at any time t was calculated as follows Jmix. [3H]t [14qroix [3H], ['"q, Xa = lW]J( I14c], ))sterol" l(?H]J( [i4q )}CHE (1) where [3H]{ represents [3H]-sterol present in the supernatant (i.e. transferred to the acceptors) or [3H]-CHE associated with the donors used in the parallel experiments at any time t after separation, and [3H]mjx represents [3H]-sterol or [3H]-CHE in the initial donor and acceptor mixture at t=0 prior to separation. Similarly, [14C]t represent counts remaining in the supernatant as acceptors after separation of the incubation mixture at time t during the sterol exchange experiments or as acceptors remaining in the parallel [3H]-CHE-labelled donor experiments. [14C]mjx represents the total [14C] counts in the donor and acceptor mixtures prior to separation. 2.2.6 Calculation of rate constants for transbilayer diffusion and exchange The rate constants for exchange of cholesterol and cholesterol sulphate from donor to acceptor vesicles (Figure 19A) were determined according to a three compartment model (described in Figure 19B). Compartments A and B represent the inner and outer monolayer of the donor vesicles and compartment C, the bilayer of acceptor vesicles. Rate constant kj, is assigned for the transbilayer movement from A to B and the rate constant k2 for exchange from the outer monolayer of the donors (B) to the acceptors (C). 59 Figure 19 Experimental Conditions, Rate constants and the Three Compartment Model (A) Schematic representation of the experimental conditions and rate constants associated with the exchange process, kx (transbilayer migration) and k2 (exchange). (B) Schematic representation of the three compartment model. Compartment A and B represent the inner monolayer (IM) and outer monolayer (OM) of the donor vesicles and compartment C represents the acceptor bilayer. cholesterol donor vesicle acceptor vesicles 10X excess B IM OM Acceptors k 1 A -« »• e - * -k 2 - • c 60 Flux equations that describe the change in the number of cholesterol or cholesterol sulphate molecules in each compartment over time are described below. The number of moles of sterol in each compartment is proportional to concentration. Therefore: dNA N B N A "dT=kl(vS"V (2) dNB N_A % N C % 'A" V B } + k 2 ( v C ' V B at = k i(v7-v^) + k ^ - ; v r r : ) (3> d N c N B N c -dT=k2(vB"-Vp (4) N denotes the number of moles of sterol in each compartment (A, B, C) and V denotes the volume of each compartment (A, B, C). If At that are sufficiently small are used then we can approximate dt and thus the equations can be rewritten as N B NA ANA^ICV^-V^)* (5) N A N B N C N B A N B = k i ( ^ ; - v^) A t + k 2 ( ^ - y£At (6) NB N C A NC = k2(y^- v^At (7) Instead of analytically solving equations 5, 6, and 7, a computer was used to numerically integrate and fit the experimental data of N^ and solve the simultaneous differential equations 61 2, 3 and 4 for each time interval (At). A curve fitting program was used to approximate the rate constants Iq and k2 by numerically integrating values that would best fit the experimental Nc values (% sterol transferred to acceptors). All cholesterol exchange experiments were performed with at least three different sets of donor and acceptor vesicles in separate experiments with duplicate measurements made at the time points indicated. The data presented are expressed as the mean ± standard deviation of data collected from all the experiments. 2.2.7 Estimation of the amount oflipid present in the outermonolayer of vesicles MLV400 of EPC were prepared in HBS as described above. The proton decoupled 81.0 MHz 31P-NMR spectra of vesicles before and after the addition of 5mM (manganese) Mn2+ were recorded employing a Bruker MSL 200 spectrometer. The free induction decay (FID) was collected using a pulse width of 2.9 us (52°C) with an interpulse delay of 1.0s. After 100 scans, the FID was Fourier transformed using a line broadening function of 25 Hz for the MLV400 and 10 Hz for the LUV10o and SUV40. The Mn2+ added was impermeable to vesicles yet sufficient to broaden beyond detection the 31P-NMR signal from phospholipid molecules facing the external medium. Spectra before and after Mn2+ quenching were weighed to give an estimate of the percent phospholipid present in the outer monolayer of donor vesicles. 2.2.8 Ion exchange chromatography of vesicles containing cholesterol sulphate Donor vesicles composed of EPC/CS (90:10) or EPC/DSPG (95:5) and labelled with 1.0 uCi of [3H]-CHE were prepared as described previously and incubated at 37°C with 10 fold excess EPC acceptors labelled with 0.1 i^Ci [14C]-CHE. At the time intervals indicated, 0.5 ml aliquots of the mixtures were removed and the two vesicle populations separated on the basis of charge employing a 1x10 cm DEAE-Sepharose CL-6B column equilibrated with 50 mM NaCl, 20 mM HEPES (low salt buffer). Initially, 1 ml fractions were collected using low salt buffer then a further 15 fractions were collected whilst eluting with high salt, low pH buffer (1M 62 NaCl, 20 mM citrate, pH 2.0). The elution profiles of both [3H]-CHE and [14C]-CHE labelled vesicles from the column are represented as the % recovery of the total counts added. 2.3 RESULTS The kinetics of sterol exchange were measured for the three types of liposomes being considered for enhancing RCT in vivo. SUV, LUV10o and MLV40o exhibit markedly different ratios of surface area to total membrane, consequently, for the kinetic analyses it was important to obtain accurate estimates of the pool sizes described in Figure 19B. 2.3.1 Compartment volumes and lipid distributions Of the three preparations, only SUV and LUV10o a r e known to be unilamellar (Hope et al., 1986), therefore the phospholipid compartment for these vesicles can be divided into inner monolayer (IM) and outermonolayer (OM). MLV400, on the other hand, are known to have an ill-defined number of internal lamellae. However, using 31P-NMR, it was possible to accurately determine the amount of phospholipid in the surface monolayer and therefore also the combined contribution of the IM of the external bilayer plus internal bilayers. The volumes of compartments A and B, which represent the donor vesicle IM and OM respectively, were calculated from the mean vesicle diameter (measured by QELS) and assuming an average phospholipid surface area of 0.6 nm2 and a bilayer thickness of 5 nm. For the SUV preparation used in these studies (44 nm diameter) this yielded a volume for compartment A that was 42% of the total bilayer volume whereas the OM (compartment B) occupied 58% of the total volume. In all experiments, the phospholipid concentration of acceptors vesicles was 10 fold higher than that of the donors. In the SUV(donor), LUV (acceptor) system, the number of donor vesicles is approximately equal to the number of acceptors, because of the small diameter of the SUV, and there were 1.08 times as many SUV than LUV 1 0 Q acceptor particles despite the latter being 10 fold excess. However, under these 63 conditions the volume of compartment C is 8.6 times the volume of B. For the LUV (donor), LUV (acceptor) experiments the volume of C was 10 fold greater than B, and the number of acceptor vesicles was also 10 fold more than donors. MLV sized through filters with a pore size of 400 nm have been shown by freeze-fracture to contain internal bilayers (Hope et al., 1986). Consequently, simple geometric calculations cannot be used to estimate the IM and OM pool sizes. However, 31P-NMR can be used to measure the amount of phospholipid in the external monolayer of the outer bilayer. Figure 20 shows the 31P-NMR spectrum of MLV40o recorded before and after the addition of 5 mM Mn2+. Spectrum A represents the phosphorous signal from all the phospholipid in the sample. The Mn2+ ion broadens the signal obtained from the phospholipid headgroup to such an extent that it does not contribute to the signal intensity measured under these conditions. The MLV4oo are also impermeable to this ion, consequently, in the presence of Mn2+ the signal intensity detected arises from only phospholipid located in the IM of the outerbilayer and any internal bilayers. Therefore, the difference in the integrated areas of spectra A and B, is equivalent to the proportion of total phospholipid in the OM of the MLV 400 preparation, and was determined to be 40% of the total. Therefore, 80% of the total phospholipid is present in the external bilayer of this oligolamellar system with the remaining phospholipid being located in internal lamellae. For the purposes of this kinetic analyses, compartment A was assumed to be 60% of the total donor lipid (IM plus internal bilayers), and compartment B (OM) as 40%. Therefore, in the MLV40o (donor), LUV 1 0 Q (acceptor) experiments, compartment C was calculated to be 10.4 times greater than B, but the ratio of MLV400 to LUV10o vesicles was 1:44. 64 Figure 20 Estimation of Outermonolayer Phospholipid in MLV400 (A) 31P-NMR spectra of MLV400 composed of EPC before and (B) after the addition of manganese. (C) An estimation of the amount of lipid present in the outersurface of the vesicles. A 23 o -as PPM B ^ « f ' ' J. 29 0 - 23 PPM c 40% Figure 21 Flux of Cholesterol Between Donor and Acceptor Vesicles SUV40(B) , LUV1 0 0(«) and MLV400 (A) donors composed of EPC/CHOL/Cer (75:10:15) were incubated with 10 fold excess concentration of EPC LUV100 at 37°C in Eppendorf tubes. At the indicated times, the two vesicle populations were separated by centrifugation after the addition of 50 ug of galactosyl binding lectin. Subsequently, the % cholesterol transfered to the acceptors was determined after quantitating by scintillation counting an aliquot of the supernatant which contains the acceptors. Data points represent the mean ± standard deviation (error bars) of all experiments described in section 2.2.6. 0) o c o •o c •~" c • M c CO E o l _ ^^ o © 4-« CO 55 100 90 80 70 60 50 40 30 20 10 0 12 16 20 24 Time (hours) 66 2.3.2 Effect of liposome size and lamellarity on cholesterol exchange The rates of cholesterol exchange from three types of donor vesicles are shown in Figure 21. Over the 24 hour time course shown, cholesterol is available for exchange from all three liposomal systems, but in agreement with prior literature, the rates of cholesterol exchange appear to decrease significantly with increasing vesicle size. This is illustrated below in Table 6 which summarizes the rates of transbilayer movement (k^ and exchange (k^ for cholesterol and CS. Using the curve fitting procedure described in methods (section 2.2.5), a series of kj and k2 values were tested and compared to values for NQ entered into the program for a given time t, until the best fit of all the eight NQ data points were reached and values for kj and k2 were derived. Table 6. Summary of rate constants and half-times for flip-flop and exchange in various sized vesicles. Donor Liposome SUV40 LUV100 MLV400 Cho kl 0.075 0.058 0.018 UA fliD (hr) 9.2 12.0 38.3 esterol k2 0.44 0.18 0.057 "/£ exc (hr) 1.6 3.9 12.2 Cholesterol Sulphate ki 0.043 0.047 0.046 UA fliD (hr) 16.1 14.8 15.1 k2 3.41 1.12 0.51 *-Vi exc (hr) 0.20 0.62 1.36 67 SUV40 LUVJOO a°d MLV 4 0 Q were found to exhibit an apparent half-time (t^) for cholesterol exchange of 2, 4 and 12 hours respectively. In addition, we estimate the t^ for transbilayer movement of cholesterol to be approximately 9 hours in SUV40 and 12 hours in LUVJQO- The observation that the rate of cholesterol flip-flop in LUV10o is slower than SUV may be due to the decreased rate of desorption of cholesterol from the OM of the larger vesicles, assuming that the IM and OM concentrations of cholesterol are in equilibrium. The rate of transbilayer movement for cholesterol in MLV40o cannot be determined accurately owing to the sterol exchange expected to occur with the internal lamellae. Using the three compartment model, a lq is obtained which gives rise to a U^ for flip-flop of more than 38h, however, this lq value represents a hybrid rate constant that would overestimate the true rate of transbilayer diffusion. This suggests that flip-flop is very slow in large, planar model membrane bilayers. The kinetic model employed predicts that the transbilayer diffusion of sterol is slower than the rate of desorption from the OM for all three liposomal systems examined. Consequently, under non-equilibrium conditions, favouring net cholesterol flux from donor to acceptor, the OM cholesterol concentration will fall more quickly than the IM concentration. This is shown graphically in Figure 22, panels A-C. Using the kj and k2 values calculated, for the three liposomal systems, the cholesterol content of the OM (solid squares), IM (solid triangles) and the total cholesterol remaining in donors (solid circles) are shown. For SUV40, it is estimated that approximately 75% of the cholesterol located at the OM is removed within 4h, whereas only 15% of the IM sterol pool has been reduced. Similarly, for LUV10o, about 50% of the OM cholesterol has been depleted at 4h compared to less than 10% from the IM (Figure 22B). The exchange process is much slower for MLV40o, nevertheless, after 24h, more than 60% of the sterol in the OM has been removed, but only 15% of the sterol from the IM and internal bilayers is depleted (Figure 22C). 68 2.3.3 Effect of liposome size and lamellarity on cholesterol sulphate exchange In the previous section, it was found that the kinetics of cholesterol exchange could be described by a model in which the rate of diffusion of sterol from the IM to OM was slower than the rate of cholesterol exchange from donor to acceptor vesicles. Here the kinetics for CS exchange are described using the same experimental procedures and model. As predicted, the difference between kj and k2 are much greater than those observed for cholesterol. As shown in Figure 23, CS exchange occurs very rapidly during the first four hours of incubation. The rate constants (see Table 6) give rise to an exchange t^ of 0.2h for SUV40, nearly an order of magnitude faster than that observed for cholesterol in the same vesicles. The rate of exchange is similarly enhanced from LUV10o and MLV40o- This data is consistent with the reduced energy required for CS to desorb from the surface of the OM due to the presence of the hydrophilic sulphate moiety. It is also clear from the net flux data shown in Figure 23, that the kinetics for CS are more obviously biphasic than those seen for cholesterol (Figure 21). The three compartment model was used to obtain kj (transbilayer rate) and k2 (exchange rate) for CS (Table 6). It is interesting to note that the rate of CS flip-flop in SUV40 and LUV10o is slower than that estimated for cholesterol, but not as large as the difference in the rate of exchange between vesicles. When concentrations of CS in the IM and OM are plotted against time then the exaggerated difference between kj and k2 is highlighted. Plotted in Figure 24, it can be seen that within lh, >80% of CS in the OM of SUV40 has been removed, whilst <5% of sterol from the IM has been depleted. A slower flux is seen from LUV100, but by 2h, a similar sterol transbilayer asymmetry has been induced. Although the kinetic model predicts that an asymmetric distribution of cholesterol also exists due to the depletion process, this cannot be readily confirmed by other techniques. However, because CS is negatively charged (pka<2; data not shown) at the pH of these experiments, it is possible to detect changes in distribution by monitoring the decrease in surface charge of the vesicles as CS diffuses into acceptors. Using L U V 1 0 Q donors and 10 fold 69 Figure 22 Relative Rates of Cholesterol Depletion from the IM and OM Estimated by the Kinetic Model The % sterol remaining in donor vesicles of various size following their incubation with 10 fold excess acceptors are redrawn as (• ) . The line tracing the points represent the best fit of the data using the 3 compartment model described (Methods) In addition, the estimated contribution of the inner (A) and outer (•) monolayer sterol is shown. The analysis was carried out for (A) SUV40 donors, (B) LUV100 donors and (C) MLV40o donors. Data points ( • ) , represent the mean ± standard deviation (error bars) of all experiments described in section 2.2.6. 100' 90 | 12 16 20 24 Time (hours) 70 excess acceptors, ion exchange chromatography was employed to show that after 4h, the surface charge of donors was greatly reduced. Elution profiles are shown in Figure 25. Neutral EPC LUV10o elute from the DEAE-Sepharose column in the void volume (Figure 25A) whereas vesicles composed of EPC/CS (90:10 mole ratio) adhere to the positively charged Sepharose and can only be eluted at high ionic strength and low pH, as shown in Figure 25B. However, after 4h incubation with acceptor vesicles the LUV10o containing CS now elute in the void volume (Figure 25C), suggesting that they have lost their surface negative charge, which is consistent with the data shown in Figure 24B, indicating that the OM contains <6% of its original concentration of CS, even though at 4h only 50% of the total CS contained in the vesicles has been depleted. In order to demonstrate that simply reducing the negative surface charge by 50% is not sufficient to prevent vesicles from binding to the ion exchange column, a parallel experiment was performed using donor vesicles composed of EPC/DSPG (95:5 mole ratio). As can be seen, in Figure 25D, these vesicles remain bound to DEAE-Sepharose until eluted with high ionic strength and low pH, even after 4 hours of incubation with 10 fold excess of acceptors. It is interesting to note that t^ for transbilayer diffusion of CS in MLV40o is similar to that observed for the unilamellar systems (see Table 6). Although the reasons for this are not clear, it is likely that the MLV40o containing CS have fewer, if any, internal lamellae due to the presence of negative surface charge (Hope et al., 1986). This hypothesis is supported by the observation that cholesterol donors of MLV40o containing 10 mol% EPG also exhibit faster lq than systems that do not contain PG (data not shown). 71 Figure 23 Flux of Cholesterol Sulphate Between Donor and Acceptor Vesicles SUV40(B) , LUV1 0 0(«) and MLV400 (A) donors composed of EPC/CS/Cer (75:10:15) were incubated with 10 fold excess concentration of EPC LUV100 at 37°C in Eppendorf tubes. At the indicated times, the two vesicle populations were separated by centrifugation after the addition of 50 ng of galactosyl binding lectin. Subsequently, the % CS transfered to the acceptors was determined after quantitating by scintillation counting an aliquot of the supernatant which contains the acceptors. Data points represent the mean ± standard deviation (error bars) of all experiments described in section 2.2.6. J 1 I I I L 0 4 8 12 16 20 24 Time (hours) 72 Figure 24 Relative Rates of Cholesterol Sulphate Depletion from the IM and OM Estimated Using the Kinetic Model The kinetics of cholesterol sulphate transfer to acceptor vesicles are redrawn as ( • ) . The line tracing the points represent the best fit of the data using the 3 compartment model described (Methods) In addition, the estimated contribution of the inner (A) and outer (•) monolayer sterol is shown. The analysis was carried out for (A) SUV40, (B) LUV10o and (C) MLV40o-However, the IM cholesterol for MLV400 includes approximately 20% of the total lipid trapped as internal lamellae donors. Data points (•) , represent the mean ± standard deviation (error bars) of all experiments described in section 2.2.6. ! 1 -S £ I 100 (0 80 70 (0 501 40 30 20 10 a i Iv ' \ \ -^ /f"""""-1 — B J . ' ' 100 8 12 18 Time (hours) 20 24 73 Figure 25 Elution of Vesicles from DEAE-Sepharose Columns: Measurement of Surface Charge Elution profiles from DEAE-Sepharose columns of (A) EPC acceptors, ( • ) , labelled with trace [14C]-CHE, (B) EPC/CS (90:10) donors, ( • ) , labelled with trace [3H]-CHE, ( Q a mixture of A («)and B (•) after 4 hours incubation at 37°C, and (D) a mixture of EPC acceptors (•) and EPC/DSPG (95:5) (• ) , after 4 hours incubation at 37°C. The arrow indicates the point at which elution with high salt, low pH buffer was initiated. E a. •a e a. •a E a to •0 TO •0 50 40 30 20 10 n I • H \ 1 1 / L / •J I c E a •o Fraction numbsr 74 2.4 DISCUSSION The three compartment model used to analyse the flux data presented in this chapter assumes that initially there is a transbilayer distribution of cholesterol such that the C:P ratio of the IM and OM of the donors is equal. This is consistent with the known behaviour of cholesterol in model membrane systems. The data did not fit a single exponential, as has been found by others (see introduction to this chapter and Figure 21), but the efflux kinetics were very well described by two rate constants, where k± represented the rate of transbilayer diffusion from the IM to OM of donors, and k2 the diffusion of sterol from the OM to excess acceptor membrane. However, the rates of cholesterol efflux from the vesicles used in this study were comparable to those reported in the literature (see references associated with Figure 18). Furthermore, when data reporting a single pool of cholesterol was drawn from the literature and fitted to the analysis described here, cholesterol efflux rates were very similar to those reported and a kj value for the best fit was not detected. The data presented in this chapter does not exclude the existence of lateral sterol domains and says nothing about the possible influence such domains may have on the kinetics of cholesterol efflux. Phase diagrams of cholesterol and phospholipids in a bilayer have suggested the existence of discrete lateral domains which form by separation when cholesterol concentrations are between 5-30 mol %, the range used in these studies (Darke et al., 1972). On the other hand, the existence of lateral domains for cholesterol in model membranes remains contentious. It has been argued that the rates of rotation and diffusion of lipid molecules within a bilayer are sufficiently fast that restrictive interactions between phospholipids and cholesterol to affect lateral movement are unlikely (Yeagle, 1981). In addition, Hyslop et al. (1990) using fluorescence intensity and lifetime measurements of cholestatrienol (another cholesterol analogue that exhibits behaviour in a bilayer similar to cholesterol), have demonstrated that sterol molecules in the bilayer do not self-associate at any great degree and are homogenously dispersed throughout the membrane. Ideally, in order to elucidate the possible effects of lateral domains, donor vesicles containing different C:P ratios need to be examined. 75 By comparing synthetic analogues of cholesterol it has recently been shown by Kan et al., (1992), that the rates of sterol exchange between membranes is highly dependent on sterol structure, and molecules more hydrophilic than cholesterol generally exhibit a faster rate of exchange. Moreover, in contrast to their observations for cholesterol, biphasic kinetics were observed for zwitterionic sterols and sterols bearing a positive charge at the 3rd carbon position and the authors suggested that the slow phase kinetics reflects the high energy for inner-to-outer leaflet movement of the charged lipid. The results presented here are consistent with their observations. The apparent rates of transbilayer movement calculated for cholesterol in this chapter are considerably slower than those reported by others (see Schroeder and Nemecz, 1990 for review). However, flip-flop rates on the order of several hours could possiby explain the existence of "non-exchangeable" pools of cholesterol observed by several groups who only measured kinetics for 8h (Bar et al, 1986 and 1987). If the rate of cholesterol transbilayer movement is indeed several hours, then in order to examine the exchangeability of cholesterol, data would need to be collected for at least several half times of transbilayer diffusion to approach equilibrium. This point has also been noted by Kan et al. (1992), who showed that if cholesterol efflux data is monitored until equilibrium is approached then the size of the"non-exchangeable" pool is decreased. Nevertheless, it remains unclear why phospholipid vesicles composed of very similar composition and cholesterol content exhibited such different kinetics (rate and extent) of cholesterol equilibration in the hands of different groups. One possible explanation could be that depending on the exact composition, vesicles of slightly different size will be generated thus affecting the rates of exchange. For example, vesicles with higher cholesterol content have been noted to be larger in size while vesicles composed of negatively charged lipids are likely smaller. Another possibility could be due to differences in the degree of separation between donor and acceptor vesicles traditionally used to assay cholesterol exchange between membranes. If the level of separation between these vesicles progressively diminishes over 76 time, due to an exchange of lipids incorporated to distinguish the two populations, then slightly higher counts detected during later time points may skew the results. In order to account for this spill over in our studies, experiments using a non-exchangeable marker to follow donor vesicle separation were run in parallel, and the counts remaining in the supernatant were used to correct for possible donor interference. Without this correction the data may have fitted a single exponential function. This point has been made by others who have noted a "non-exchangeable" pool of cholesterol (Poznansky and Lange, 1976; Bar et al., 1986). The possible presence of residual MLV in SUV preparations may also contribute to the observation of a "non-exchangeable" pool of cholesterol. Some groups have demonstrated this by examining the kinetics of exchange following the fractionation of the vesicles by gel filtration (Backer and Dawidowicz, 1979), yet again, to illustrate inconsistencies in the literature, other groups have not noted any differences (Poznansky and Lange, 1978; Bar et al., 1986). The kinetics of cholesterol net exchange from SUV40, LUV10o and MLV400 into acceptors have been investigated in this chapter. Of course the proposed application in vivo requires that these liposomes act as acceptors rather than donors. However, the data generated from their behaviour as donors is applicable to their role as acceptors as the cholesterol exchange process is reversible (compare Figure 21 with Figure 30, Chapter 3; Bar et al., 1986; Phillips et al., 1987). For example, from the data, one can predict that for equivalent amounts of phospholipid, SUV40 will absorb cholesterol more quickly than the remaining two systems and the rate of cholesterol uptake by the larger MLV40o will be much less than observed for the unilamellar vesicles. Indeed, this is what is found in vivo and is discussed in greater detail in the next chapter. Finally, it is of interest to note the difference in exchange behaviour between cholesterol and CS. Unfortunately, the biological role of CS is unknown and so it is difficult to speculate as to whether the exchange kinetics observed for this sterol is functional in vivo. Clearly in a biological membrane, CS will be much more readily available for exchange into acceptor membranes or lipoprotein particles than cholesterol. However, transbilayer movement of CS is 77 more restricted than for cholesterol, which will tend to anchor the sulphated sterol at adjacent monolayers to the exchange surface. Quantitatively, CS is most abundant in the epidermis, where it is synthesized by cholesterol sulfotransferase (Elias et al., 1984), and is an integral part of the lipid composition of the outer layers of skin. Moreover, CS has been linked to desquamation, the natural process whereby corneocytes are sloughed from the surface of the skin. As terminally differentiated cells approach the skin surface, CS is hydrolysed to cholesterol by cholesterol sulphatase, which appears to promote desquamation, because in a rare genetic disease where cholesterol sulphatase is missing, excessive scaling occurs. Studies of this disease, x-linked ichthyosis, have shown that desquamation is partly regulated by the CS/cholesterol ratio in the stratum corneum, the outermost layer of skin (Williams, 1992). Normally this ratio is 1/500, whereas scaly skin can have ratios as high as 1/5 (Lampe et al., 1983). It is not clear how the differences in thermodynamic behaviour between CS and cholesterol could be involved in maintaining skin structure. The sulphate moiety of CS is thought to play a role in binding the layers of extracellular lipid within the stratum corneum, so that converting CS to cholesterol helps the release of corneocytes at the surface (Elias et al., 1984). However, it is possible that the exchange properties of CS and cholesterol could be used therapeutically to decrease scaling in patients suffering from x-linked ichthyosis. Given the rapid exchange rate for CS, a topical application of unilamellar liposomes containing cholesterol might be expected to change the CS/cholesterol ratio of the stratum corneum, by removing excess CS as well as inserting exogenous cholesterol. There is evidence that scaling can be reduced by cholesterol replacement therapy (Feingold and Elias, 1993), suggesting that such a liposomal application could be therapeutically useful. In summary, the data presented in this Chapter suggest that in model membrane systems the apparent rate of transbilayer movement of cholesterol is not as fast as the rate of exchange. Therefore IM sterol could represent a more slowly exchanging pool of cholesterol than that found in the OM, and thus give rise to the biphasic kinetics observed here and elsewhere in the 78 literature. However, it is not possible to distinguish between this model and one in which there are lateral domains within the plane of the bilayer that give rise to areas with different rates of cholesterol desorption (Nemecz and Schroeder, 1988). Having said this, the results obtained using CS, are less ambiguous because surface charge was monitored to demonstrate that the rapidly exchanging pool of sterol was located in the OM of vesicles and the slow pool at the IM. Furthermore, the kinetic analysis suggests that for equal amounts of phospholipid the rate at which the liposomal systems will absorb cholesterol in vivo is in the order of SUV40 > LUVJQO » MLV4oo- In the following chapter these systems are studied in vivo and other factors contributing to their ability to accumulate cholesterol intravenously are considered. 79 CHAPTER3 THE INFLUENCE OF SIZE AND COMPOSITION ON THE CHOLESTEROL MOBILIZING PROPERTIES OF LIPOSOMES IN VIVO 3.1 INTRODUCTION The development of atherosclerosis involves the accumulation of cholesterol, cholesterol esters and phospholipids in the intima of the major arteries (Small, 1988). As the lesion increases in size, blood flow through the vessel is restricted and can eventually result in clinical symptoms associated with tissue ischemia. It has been noted that the core of atherosclerotic plaques can contain 30% to 65% (dry weight) lipid which therefore contributes a major part of the plaque volume (Small, 1988). If this lipid could be removed significant regression would occur. Many studies have demonstrated that the intravenous administration of liposomes induce the reversal of experimentally induced atherosclerosis (reviewed in Williams and Tall, 1988). As a result of this work and advances in our understanding of liposome-lipoprotein interactions Williams and colleagues (1984) suggested that phospholipid therapy might be a viable means of inducing the rapid regression of atherosclerotic lesions in humans and proposed a mechanism for their activity. It is thought that circulating phospholipid liposomes deplete lipoproteins of cholesterol thereby generating sterol-poor, phospholipid-rich particles. In addition, a proportion of vesicles are believed to bind apolipoproteins and assimilate into the lipoprotein pool as HDL-like particles, which are also capable of entering the interstitial space, promoting cholesterol efflux from cells. Thus, liposomes reduce the volume of atherosclerotic lesions due to the mobilization of cholesterol from the arterial wall which promotes the depletion of cholesterol and cholesteryl esters deposited within the tissue. The normal clearance route of liposomes through the RES delivers the majority of excess cholesterol to the liver for excretion and re-distribution. 80 It is well established that liposome clearance rates are dependent upon size and composition (Gregoriadis, 1988), physical characteristics which will also influence the extent of cholesterol equilibration. Previous studies investigating phospholipid-induced mobilization of cholesterol in vivo however have employed only multilamellar or sonicated vesicles (Williams and Tall, 1986; Williams et al., 1988; Mendez et al., 1988; Friedman and Byers, 1958; Byers et al., 1962). In this chapter some of the physical characteristics of liposomal systems that would be expected to influence cholesterol mobilization given the mechanisms discussed above are examined. In particular, we have compared the extent and rate of cholesterol mobilization, in vivo, employing the three preparations studied in Chapter 2, SUV, LUV10o and MLV400, as well as LUV50. 3.2 MATERIALS AND METHODS 3.2.1 Materials Cholesterol and [4-(2-hydroxyethyl)]-piperazine ethane sulfonic acid (HEPES) were obtained from Sigma Chemical Company. [14C]-Cholesteryl hexadecyl ether and [3H]-cholesterol were purchased from New England Nuclear. EPC, distearoylphosphatidylcholine (DSPC), distearoylphosphatidylglycerol (DSPG) and egg phosphatidyl glycerol (EPG) were supplied by Avanti Polar Lipids (Alabama). Bio-Gel A-15m medium was purchased from BioRad. Ail chemicals, thin layer chromatography plates and solvents were of analytical grade and purchased from BDH Chemicals (Vancouver, B.C.). 3.2.2 Preparation of vesicles All liposome preparations were labelled using trace amounts of [14C]-CHE for the following reasons (1) it does not undergo passive exchange between membranes (Derkesen et al., 1987), (2) the mouse does not exhibit cholesterol ester exchange protein activity and (3) the ether linked fatty acid is not cleaved in the plasma. Consequently, in this model system CHE is an excellent liposome marker and vesicle concentrations in the plasma were estimated from the 81 specific activity of this label. A chloroform solution of EPC and [14C]-CHE was vortexed and solvent was removed under a stream of N2. The sample was subsequently dried under high vacuum for 2 h. The dry lipid film was hydrated in HBS (pH 7.4) to generate MLV. Vesicles were prepared from MLV either by sonication, to generate SUV or extrusion to produce LUV. Sonication was performed as described previously in section 2.2.3. Briefly, the MLV suspension was diluted to 30 mg/ml, immersed in an ice bath and subjected to 3 cycles of sonication each of 10 min duration after which time the initial milky suspension became clear and the vesicle size was 30 nm as determined by QELS. The SUV were centrifuged at 10,000 g for 30 min to remove titanium fragments originating from the sonicator tip. Extrusion was carried out using a 10 ml Extruder equipped with a water jacketed thermobarrel. MLV were sized through two stacked polycarbonate filters of defined pore size to generate a variety of LUV and homogeneous MLV as described previously in section 2.2.3. 3.2.3 Vesicle diameters The size of vesicles generated by sonication and extrusion procedures was determined by QELS. The following diameters are expressed as the mean + standard deviation of vesicle preparations prior to injection. Vesicles prepared by sonication were 30 + 7 nm in diameter (SUV30). Vesicles prepared by extrusion through filters with a pore size of 0.05 nm were 70 + 19 nm, 0.1 fxm pore size 125 + 30 nm, and 0.4 u,m pore size 237 + 90 nm. As mentioned earlier in Chapter 2, vesicles prepared by extrusion are referred to by the filter pore size used in their preparation i.e. LUV50, LUV100 and MLV400. 3.2.4 Vesicle infusions Female BDF-1 or CD-I mice, weighing 20-22 g (Sprague Dawley), were used throughout this study. Vesicles were injected via the tail vein at a dose of 300 mg/kg which was typically 6 mg of lipid in 200 pel of buffer injected for each animal. Control mice were injected with an equal volume of buffer and both groups were killed at specified times and 82 blood collected in EDTA microtainer tubes by heart puncture. Plasma was obtained following centrifugation at 2000 g for 10 min, and an aliquot removed for scintillation analysis using a Beckman LS 3801 liquid scintillation counter. Each time point is the average of data from 16 mice (from 4 separate experiments), unless indicated otherwise. 3.2.5 Gel filtration A 27 x 1.5 cm Bio-Gel A-15m gel filtration column, equilibrated with 150 mM NaCl, 10 mM TRIS, 0.1% EDTA, 0.3% NaN3 (pH 7.4) was used to fractionate plasma samples. Columns were eluted at a flow rate of 1 ml/min and 1 ml fractions were collected for radioactivity and lipid analyses. Data on the C:P ratio of vesicles and lipoproteins after infusion was obtained from pooled fractions corresponding to the liposomal and lipoprotein peaks. The Bio-Gel columns were calibrated with respect to lipoprotein elution by preparing purified human lipoprotein fractions using standard ultracentrifugation procedures (Schumaker and Puppione, 1986) labelling each with [3H]-cholesterol and monitoring the elution profile for radioactivity. 3.2.6 Lipid analysis Pooled column fractions and plasma samples were extracted employing the Bligh and Dyer (1959) procedure. The lipid extracts were analyzed for total cholesterol using the assay method of Rudell and Morris (1973), free and esterified cholesterol concentrations were determined following separation by TLC using hexane/ether/acetic acid (70:30:1 v/v). Standards were used to identify the area of the plate corresponding to these two lipids, the silica was aspirated and the lipid eluted for assay using chloroform:methanol (2:1 v/v). Plasma vesicle phospholipid content was determined by dividing [14C]-CHE radioactivity by liposomal specific activity and phospholipid concentrations were determined by the method of Fiske and Subbarow (1924). Erythrocytes were extracted using the method of Rose and Oklander (1965), followed by a Bligh and Dyer wash to remove residual salts. An aliquot of red blood 83 cells was retained for cell number determination using a Coulter cell counter in order to express cholesterol and phospholipid concentrations as nmoles/109 cells. 3.2.7 Erythrocyte cholesterol specific activity Blood was pooled from a group of mice and red cells packed by low speed centrifugation. The serum was labelled with [3H]-cholesterol by incubation for 10 min at 37°C with 100 ^Ci of radioisotope dried from ethanol. The labelled serum was added to the packed cells and the mixture incubated at room temperature for 30 min. The cells were washed and approximately 10" dpm of [3H]-cholesterol labelled cells injected into the experimental groups via the tail vein. Approximately 1 min after the injection of cells, saline or liposomes were administered. 3.2.8 Cholesterol exchange in vitro Donor and acceptor vesicles were separated employing ion exchange chromatography. In contrast to the experimental conditions used in Chapter 2, a ten fold excess of donor vesicles were used. Donor vesicles (100 nm diameter) composed of EPC/EPG/Chol (40:15:45 mol ratio) were incubated with 100 nm or 400 nm EPC acceptors. Donor vesicles were labelled with [3H]-cholesterol at 5 [xCi/100 mg total lipid and acceptors were labelled with [14C]-CHE at 0.5 nCi/100 mg lipid. At the specified time intervals, 50 \i\ of the incubation mixture (1 mg acceptor + 10 mg donor/ml) were removed and passed down a DEAE-Sepharose 6B-CL column prepared in a 1 ml tuberculin syringe equilibrated with 30 mM NaCl, 20 mM HEPES (pH 8.0). Columns were spun at 1000 g for 1 min prior to applying aliquots of the incubation mixture. The vesicle mixture was spun through the column and the eluant (acceptors) eluted with two subsequent wash/spin cycles with 500 u.1 aliquots of buffer. Recovery of [14C]-labelled vesicles (acceptors) was typically >90%. Control experiments in which donors were labelled with a non-exchangeable marker, similar to that described in Chapter 2, indicated that all of the donor vesicles bound to the ion exchange column under the conditions of the 84 experiment. Cholesterol accumulation by acceptors was determined using an LS 3801 Beckman scintillation counter equipped with a [14C]/[3H] dual label program. 3.2.9 Cholesterol excretion Two groups of mice (n=4) were maintained in metabolic cages and faeces collected daily. After 3 days one group was injected with 200 uJ of saline and the second group with approximately 6 mg of EPC LUV10o (dose 300 mg/kg). Faecal material was collected for a further 7 days. Samples were extracted using an isopropanol/chloroform extraction procedure and subsequently assayed for total cholesterol, free cholesterol and cholesterol esters, as described earlier (section 3.2.6). Experiments were carried out on mice maintained on regular, laboratory food for rodents (cholesterol excretion rate 10-12 u,moles/g faeces) and on Teklad low cholesterol, casein based diet (Teklad Premier, Wisconsin) which resulted in an excretion rate of approximately 0.8 u.moles cholesterol/g faeces). 3.2.10 Statistical analyses Unless otherwise indicated, values are presented as mean ± standard deviation . The significance of the difference of the means were assessed by an analyses of variance using the two sample t-test. Only values of P<0.05 were considered significant. 85 3.3 RESULTS Liposome clearance in vivo is chiefly dictated by liposome size, composition and lipid dose (wt lipid/body wt); liposomal half-life in the circulation, for example, increases as the dose increases (Hwang, 1987; Gregoriadis, 1988). In the studies reported here the lipid dose was not varied but maintained at 300 mg/kg in order to ascertain the importance of size and composition on cholesterol mobilization. This dose is high with respect to traditional intravenous drug delivery studies using liposomes which are typically in the range of 50-100 mg/kg in mice. However, it falls well within the tolerated dose range for this type of particle and it approximates the dose that is likely to be practicable for human administration (Williams et al., 1984; Aviram et al., 1989). 3.3.1 Cholesterol mobilization An example of cholesterol mobilization by liposomes is shown in Figure 26. Extrusion was used to prepare a homogeneous population of LUV with a mean diameter of 125 nm as determined by QELS (referred to as LUV100). A dramatic increase in plasma cholesterol is observed for animals receiving vesicles (Figure 26A). Sterol levels peak 4-8 h after injection at a concentration that is nearly double that measured in the control mice injected with an equivalent volume of saline. Plasma cholesterol concentrations gradually return to normal levels after 48 h which correlates well with the vesicle clearance profile shown in Figure IB. Liposomes were labelled with trace amounts of [14C]-CHE which is a non-exchangeable, non-metabolizable marker frequently used to monitor liposome clearance and distribution in vivo (Derksen et al., 1987). 86 Figure 26 Cholesterol Mobilization by the Intravenous Injection of Phospholipid Liposomes A) Total cholesterol concentration in the plasma of mice treated with saline ( • ) or EPC LUVJQO ( • ) at a dose of 300 mg/kg body wt. Note that the first time point is at t=l hour. Inital plasma cholesterol concentrations at t=0 for both groups were the same. Each time point is the mean + the standard deviation of a minimum n=4 experiments. B) Clearance of [14C]-CHE labelled LUV10o expressed as nmoles vesicle phospholipid/ml plasma. Saline and vesicles were administered as a single, 200 ul bolus injection into the tail vein. Whole blood was collected via heart puncture. Lipid analysis was as described in Methods. Each time point represents the mean + standard deviation of four separate experiments. 4000 3500 3000 rr" 2500 -£ 2000 1500 -1000 500 9000 _T 8000 j"i 7000 6000 5000 4000 3000 2000 1000 O E J h • - V 1 1 1 1 1 1 1 1 B i i ~~i — • 12 16 20 24 28 32 36 40 44 48 Time (hours) 87 Using gel filtration (see Methods section 3.2.5), mouse plasma was fractionated and the cholesterol profile determined using a chemical assay procedure. Plasma from control and liposome-treated animals were compared and the results are shown in Figures 27A and 27B. Panel 27A shows a normal cholesterol distribution with the majority of cholesterol associated with combined LDL and HDL peaks (fractions 22-50). The elution volumes of VLDL, LDL and HDL were determined as described in Methods. A minor quantity of sterol is detected in the void volume corresponding to the larger chylomicron and VLDL lipoprotein particles, but quantitatively these fractions represent <5% of the total cholesterol content of the plasma. The elution profile of plasma from liposome-treated animals (4 h time point) is shown in panel 27B. The [14C]-CHE liposome marker is almost exclusively detected in the void volume indicating that the LUV10o are well separated from the fractions containing LDL and HDL (vesicles smaller than 100 nm diameter are included in the gel and cannot be separated from LDL). The absence of radioactivity in the remaining fractions also indicates that little, if any, assimilation of vesicles into the lipoprotein pool has occurred. However, it is possible that small quantities of vesicles have undergone structural transitions to lipoprotein-like particles, but are removed rapidly from the circulation and cannot be detected. It should be noted that assimilation is distinct from phospholipid exchange. During assimilation it is postulated that whole vesicles are structurally re-organized, primarily as a result of binding apolipoproteins, to form smaller, lipoprotein-like particles (Krupp et al., 1976). Phospholipid exchange with lipoproteins is known to occur, but would not be detected in these experiments using the non-exchangeable lipid marker described. The cholesterol content of column fractions shown in Figure 27B clearly shows that the excess sterol in the plasma of treated mice is associated with LUV. The slight frame shift of peaks between 27A and 27B is the result of differences in elution rate and not due to changes in lipoprotein size. Using TLC analysis we determined that >90% of the liposomal cholesterol was free cholesterol, the remainder being cholesterol ester. The excellent separation of 88 Figure 27 Separation of Vesicles from Plasma by Gel Filtration A) The elution profile of normal mouse plasma (•) from a Bio-Gel A-15m gel filtration column. Fractions were extracted and subjected to a total cholesterol assay as described in Methods. Previous calibration of this gel filtration material, using purified human lipoprotein fractions labelled with [3H]-cholesterol are identified with arrows illustrating (1) VLDL (void volume) (2) LDL and (3) HDL peaks. B) A 4 hr plasma time point from mice treated with EPC LUV100. ( • ) cholesterol content of each fraction and ( • ) indicates the concentration of vesicle phospholipid measured from the specific activity of [14C]-CHE as described in Methods. c a 3 "3 a. o £ a a o JC a o E c 900 800 700 600 500 400 300 200 100 0 uUZu B r V 1 / V--Y > V 180 160 140 120 100 80 60 40 20 0 200 180 160 140 120 100 80 60 40 20 C a 3 a> _ E ^" o b a *-> to a> o x: o o E c 4* C (0 3 a> w E M M " O i_ v> o o JZ u o E 10 20 30 40 50 Fraction number (ml) 89 Figure 28 Rate of Cholesterol Accumulation by Vesicles in vivo Cholesterol .-phospholipid molar ratio of EPC LUV10o isolated from mouse plasma. Vesicles were separated from lipoproteins by gel filtration, the cholesterol and phospholipid content of lipid extracts were determined by chemical assay. Each point represents the mean + standard deviation of four separate experiments. O CO a. Q. CO O a. • o o o (0 o o o 1.20 1.00 0.80 " 0.60 " _1 0.40 -0.20 -0.00 12 16 20 24 Time (hours) 90 LUVjoo fr°m the quantitatively most abundant lipoproteins enabled the straightforward isolation and subsequent analysis of the vesicle lipids. Cholesterol accumulation is shown by the increasing C:P ratio of vesicles over a 24 h time course in vivo, presented in Figure 28. The rate of sterol accumulation by the liposomes is consistent with previously published rates of exchange and transfer for membrane cholesterol (Phillips et al., 1987) and with data presented in Chapter 2 using model membranes. The apparent equilibration at an approximately equimolar C:P ratio also fits with our knowledge of cholesterol dynamics between phospholipid bilayers and plasma membranes, and represents the maximum amount of liposome associated sterol expected under these conditions. Consequently, after 24 h the liposomes that remain in the circulation (approximately 10-15% of the initial dose) are in equilibrium with respect to cholesterol and net sterol movement is negligible. 3.3.2 Liposome size and cholesterol mobilization Previous studies investigating phospholipid induced mobilization of cholesterol in vivo have employed multilamellar or sonicated vesicles (Williams and Tall, 1986; Williams et al., 1988; Mendez et al., 1988; Friedman and Byers, 1958; Byers et al., 1962). Liposome size is a key characteristic in clearance kinetics (Gregoriadis, 1988) and is one of several reasons why sonicated vesicles might be expected to represent the bilayer structure best suited to enhance reverse cholesterol transport. Sonication reduces MLV to "limit size" vesicles. These systems exhibit the minimum radius of curvature that can be adopted by the bilayer configuration without disruption (Hope et al., 1986) and for EPC this is typically a 30 nm diameter vesicle. For a given liposome composition, it is assumed that the smaller the particle diameter the greater the circulation half-life (Gregoriadis and Senior, 1986), consequently it is expected that SUV composed of EPC will circulate longer than other types of liposomes, and therefore mobilize more cholesterol. Furthermore, the packing constraints experienced by phospholipids in SUV, (due to the acute radius of curvature) gives rise to (1) the exposure of a large surface 91 area to accept cholesterol and (2) an instability that can result in fusion (Hope et al., 1986) as well as an increased tendency to assimilate with lipoproteins (Tall, et al., 1986; Scherphof et al., 1978; Krupp et al., 1976). Therefore, given the mechanism by which liposomes are thought to enhance reverse cholesterol transport (Williams et al., 1984), it might be expected that SUV would produce a greater number of HDL-like particles thus promoting efflux of sterol from peripheral tissues. Also, the data presented in Chapter 2, indicate that SUV should accumulate cholesterol more rapidly than the LUV or MLV formulations. Plasma cholesterol concentrations were measured over a 48 h period in animals treated with a variety of liposomal preparations varying in diameter from 30-250 nm. Sonicated vesicles were prepared as described in Materials and Methods (section 3.2.3), the remaining vesicles were produced by extrusion of MLV through filters with defined pore sizes to give vesicle populations with the mean diameters given in Methods. In the text, and in Figure 29, vesicles are referred to by the filter pore size. The amount of cholesterol accumulated and removed by liposomes in vivo is a function of both the rate of cholesterol uptake and the rate of liposome clearance. An estimate of the mass of cholesterol removed from the circulation (mostly by the RES) can be made by calculating the C:P ratio of vesicles in vivo from plasma concentration of vesicle phospholipid and cholesterol as the excess plasma concentration above the control at the various experimental time points. From the data it is known that all the cholesterol above control levels is associated with circulating liposomes. It should be noted that the likelihood that control cholesterol levels in treated animals is lower than saline-treated animals has not been taken into account but this point is addressed in the discussion. The plasma volume of mice used in these studies was approximately 1 ml, consequently the total amount of phospholipid cleared from the circulation between time points was known. Using the average C:P ratio measured for vesicles between each assay interval an estimate of the cholesterol removed was obtained. The analysis was not continued beyond the point where less than 5% of the initial phospholipid dose remains in the circulation, because below this level the measurement error was too large to determine 92 accurate C:P ratios. In Figure 29A we show the cummulative level of cholesterol removed by LUVJOO up to the time when approximately 5% of the dose remains. After 40 h it is estimated that 2800 nmoles of cholesterol have been removed from the circulation by the RES, which represents 33 mol% of the injected phospholipid dose. Using this analysis the various liposomal preparations tested were then compared. For each preparation the plasma cholesterol and phospholipid clearance profiles were determined and analysed as described above. The results in Figure 29B show that LUV mobilize the most cholesterol, and the reasons for this are discussed later. It has been demonstrated previously, using freeze-fracture electron microscopy and NMR analysis (Chapter 2), that MLV sized through 400 nm pores retain a number of internal lamellae and therefore cannot be classified as LUV (Hope et al., 1986). It was not clear from the literature how much of a kinetic barrier to complete cholesterol equilibrium would be presented by a multibilayer system. Despite some discrepancies (Phillips et al., 1987; Schroeder et al., 1991), it appears generally accepted that the transbilayer movement (flip-flop) of cholesterol is rapid (Schroeder and Nemecz, 1990), on the order of seconds to minutes in a liquid crystalline bilayer under conditions that promote net sterol flux. Consequently, it might be expected that multilamellar systems would still act as a good sink for cholesterol as sterol should rapidly disperse through the internal lamellae. However, as suggested by the data presented in Chapter 2 and shown in Figure 30 this is not the case. Using an in vitro system in which LUVJQO or MLV40o were incubated with a ten fold excess of donor vesicles containing tritiated cholesterol (see Methods section 3.2.8) the net transfer of sterol from donor to acceptor was monitored. The rate of cholesterol accumulation in the unilamellar preparation was greater than that observed for the oligolamellar vesicles. These results are in agreement with data determined in Chapter 2 which suggested that cholesterol flip-flop is a relatively slow process. It is interesting to note that in the presence of a ten fold excess of donor vesicles the equilibrium C:P ratio of the acceptor should be approximately 0.9:1. The data in Figure 30 show that the 100 nm acceptors only achieve a ratio of 0.35:1 after 8 h at 9 3 Figure 29 Estimation of Cholesterol Mobilized by Different Sized Vesicles A) Cholesterol mobilization by LUV100 expressed as the cummulative nmoles of cholesterol (bar graph) removed with vesicle phospholipid (•) . The clearance kinetics of the liposomes was followed by [14C]-CHE label. B) Cholesterol mobilization by SUV30, LUV50, LUV100, and MLV4Q0- All vesicles were prepared from EPC and administered at 300 mg/kg body weight. Total cholesterol removed by liposomes ± standard deviation was calculated for n=4 separate experiments as described in the text. Significance levels determined were: SUV30 vs. LUV5 0 (P<0.001), SUV30 vs. LUV100 (P<0.025), SUV30 vs. MLV400 (P<0.001), LUV50 vs. LUV100 (N.S.), LUV50 vs. MLV400 (P<0.001), LUV100 vs. MLV400 (P<0.001). 4000 3500 3000 2500 2000 1500 1000 500 •O 0) > 0 E 0) 0 l_ 0) V) 0) 0 0 w V 0 c •o o > o E © 0 4 8 12 16 20 24 28 32 36 40 Time (hours) 4000 3500 -3000 ' — 2500 o 0 » 2000 -0) » O £ U o E c 1500 -1000 -500 SUVMLUV50LUV100MLVWO 94 xr Figure 30 e S t e r 0'E , ) U i , i b r a— „ M e r a b r a n e s The capacity of EPC I T TV , Cholesterol cont^n• , 1Q0 (•) andEPC Ml V„ , ^ S b y , 0 n exchange c o l u m n T'"me (hours) 95 37°C. This ratio is very similar to that seen in Chapter 2 (Figure 21), wherein cholesterol from LUVJQQ donors to a 10 fold excess of acceptors were monitored and confirms the reversibility of the exchange process. This is approximately half the rate of accumulation observed for the same vesicles in vivo (Figure 26). 3.3 J Liposome composition Another physical characteristic known to influence the circulation time of liposomes in vivo is lipid composition. The addition of cholesterol to liposomes has the most dramatic effect on circulation half-life increasing it by an order of magnitude in some cases. However, in the presence of equimolar cholesterol there is a further increase in circulation half-life for long chain, saturated phospholipids compared to liquid crystalline systems (Mayer et al., 1989). It has also been demonstrated that vesicles, composed of saturated lipids in the gel state are still capable of acting as acceptors of cholesterol (Phillips et al., 1987). Consequently, we compared the cholesterol mobilizing properties of two types of LUV10o composed of EPC/EPG (95:5 mol ratio) which is liquid-crystalline at 37°C and DSPC/DSPG (95:5) a gel-state lipid matrix at the body temperature of the mouse. Phosphatidylglycerol (PG) was incorporated to impart a surface negative charge, necessary to prevent the gel-state vesicles from aggregating in the absence of cholesterol (Nayar et al., 1989). In order to properly compare the two systems a negative charge was also added to EPC vesicles. The results, presented in Figure 31A, reveal that the gel-state vesicles produce a delayed increase in plasma cholesterol which does not peak until after 24 h, whereas EPC/EPG vesicles give rise to a cholesterol profile similar to that observed for EPC alone. From the data in Figure 31A it can be shown that the rate of cholesterol accumulation for these two types of vesicle is the same. The different plasma cholesterol profiles result from the fact that approximately 70% of the DSPC/DSPG vesicles are cleared within 4 h compared to less than 30% of the EPC/EPG LUV100 (Figure 31B). The bulk of cholesterol mobilization i i occurs in the first 24 h consequently liquid crystalline EPC/EPG removed more I 96 I Figure 31 Effect of Lipid Composition on the Rate and Extent of Cholesterol Mobilization A) A comparison of the plasma cholesterol profiles of mice following injections of saline (•) , liquid-crystalline EPC/EPG (95:5) LUV100 (•) and crystalline DSPC/DSPG (95:5) LUV100 (A) B) Clearance kinetics of EPC/EPG (•) and DSPC/DSPG (A) labelled with [14C]-CHE (each time point is the mean of n=4 experiments). a E CO a CO o £ O V) o o E c a E (0 CO .£ 2* a. o £ a CO o •C a JO m CD > CO _0) O E c 8 12 16 20 24 28 32 36 40 44 48 Time (hours) 97 than 3,000 nmoles to the RES whereas DSPC/DSPG vesicles removed 1,700 nmoles. It has been suggested that gel state vesicles are not as good an acceptor of cholesterol as liquid crystalline vesicles (Yeagle and Young, 1986). However, the data obtained in vivo are supported by observations in vitro (data not shown) and those of Phillips et al. (1987) which do not detect significant differences in the rate of cholesterol accumulation between the two. 3.3.4 Source of liposomal cholesterol The final series of experiments were designed to describe the source of the liposomal cholesterol and its fate. Ultimately, cholesterol efflux must occur from atherosclerotic plaque to achieve regression. However, it is known that the cholesterol within cells and atherosclerotic lesions equilibrates more slowly than sterol present in plasma membranes directly exposed to acceptor particles (Small, 1988; Phillips et al., 1987). Movement of this cholesterol will be a secondary event initiated by the primary efflux of outer membrane cholesterol. In a 20 g mouse approximately 35% of the circulating sterol is associated with lipoproteins and about 65% with the plasma membranes of erythrocytes. However, all of the sterol associated with erythrocytes is free cholesterol whereas a large proportion of lipoprotein sterol is esterified. Consequently, the largest pool of free cholesterol in the circulation is in the red blood cell plasma membrane but this source of cholesterol does not change significantly in the presence of liposomes, despite a two fold increase in plasma sterol concentration. This result is shown in Figure 32A and confirms a recent observation made by Mendez et al. (1988) in rabbits. It is well known that erythrocyte membrane cholesterol can be depleted by liposomes in in vitro (Phillips et al., 1987), consequently it is of interest to determine if erythrocytes act as the primary sterol donor, but their cholesterol content is rapidly replenished by lipoproteins which are in turn able to extravasate and scavenge more sterol from peripheral tissues. Erythrocytes from the mouse were isolated and labelled with [3H]-cholesterol in vitro. The labelled cells were injected into a group of mice, half of which were subsequently treated with saline and half with 300 mg/kg of EPC LUV10o- The specific activity of red blood cell 98 cholesterol was determined over an 8 h time course and the two groups compared. As can be seen from Figure 32B the decrease in cholesterol specific activity is the same for both the control and experimental group. Interpretation of this data is limited by the fact that cells labelled in vitro are also removed from the circulation over a similar time course (determined by chromium labelling; data not shown). However, it can be estimated that at least 50% of cell sterol would have to efflux to account for the rise in plasma cholesterol observed after 8 h, which would result in a considerable dilution of erythrocyte cholesterol if this sterol pool were continuously replenished. Clearly this is not the case suggesting that red blood cell cholesterol is not the primary source of the liposomal sterol accumulated in vivo, and extends previous observations that liposomal infusion causes an increase in cholesterol concentration for whole blood and not simply a redistribution between components of the blood (Friedman and Byers, 1958; Byers et al., 1962). C:P ratios of lipoproteins showed a significant decrease over control values in the first 8 h (Figure 32C), however, we did not distinguish between increased phospholipid content of lipoproteins due to net transfer from the liposomes or net transfer of cholesterol from lipoproteins to vesicles. The ratio returns to normal values after 8 h which mirrors the time course of cholesterol accumulation by vesicles. This supports the hypothesis that it is primarily lipoprotein cholesterol in equilibrium with circulating vesicles, and that lipoproteins mediate the transfer of cholesterol from peripheral tissues to liposomes (Williams et al., 1984). The results are also consistent with observations in vitro that indicate cholesterol can undergo desorption from lipoproteins more readily than from erythrocytes (Phillips et al., 1987). Finally, we noted above that the rate of cholesterol accumulation by LUV100 in vivo (Figure 28) is considerably faster than that observed in vitro (Figure 30), which indicates that the rate of cholesterol desorption from sources in vivo is greater than from the 100 nm vesicle donors used to obtain the data in Figure 30 and Figure 21 in Chapter 2. 99 Figure 32 Possible Source of Liposomally Accumulated Cholesterol A) The cholesterol content of erythrocytes from mice treated with EPC LUV100 ( • ) and saline ( • ) . Cells were isolated, washed and counted prior to extraction for lipid analysis (Methods). B) The specific activity of erythrocyte cholesterol from saline (•) and liposome (• ) treated mice (data points are superimposable). Erythrocytes were labelled in vitro by incubation with serum lipoproteins equilibrated with [3H]-cholesterol (see Methods). C) The change in C:P ratio of the pooled lipoproteins obtained by combining fractions 22-50 from the gel filtration column of saline ( • ) and liposome-treated (•) groups. Significance level of P<0.05 (*) and P<0.01 (**) are as indicated. 300 •= 250 -O 200 X 150 -O E c I I o o a a CO o a a o c • o a o a 300 250 -200 150 100 -2.00 1.80 1.50 1.40 1.20 1.00 0.80 0.80 0.40 -0.20 0.00 0 1 -* 2 I I — i — 3 4 5 S ' 7 8 c 1 — • i 12 20 24 Time (hours) 100 3.3.5 Cholesterol excretion At a dose of 300 mg/kg, >70% of the liposomes are removed by the liver, the remaining vesicles are phagocytosed by fixed macrophages in the spleen and likely the bone marrow with a small percentage distributed in skin and lung tissue (data not shown). The biodistribution follows the pattern observed in many studies on LUV in vivo (Mayer et al., 1989) where the liver is responsible for approximately 70% of liposome clearance. The hypothesis of Williams et al. (1984) proposes that vesicles removed by the liver will deliver cholesterol into the bile synthetic pathway for eventual excretion as a mixture of bile salts and free cholesterol. In order to see if an increase in the excretion of free cholesterol could be detected, two groups of mice were maintained (saline-treated and liposome-treated) in metabolic cages for ten days and faeces were assayed for free cholesterol. On a normal laboratory diet for rodents, mice excrete 10-12 (xmoles of total cholesterol per gram of faeces. Following the administration of saline or liposomes a significant change in this level over a seven day period was not detected. This was not unexpected since the dietary sterol was so much greater than the amounts being mobilized by vesicles. However, employing a low cholesterol diet, which reduced background excretion to approximately 1 jimole cholesterol/g faeces a difference between saline and liposome-treated groups were still not detected. Both groups produced approximately 1 g of faeces/day; at the end of the experimental period (1 week) the accumulated excretion rates for total faecal cholesterol in control and treated mice were 0.795 jimoles/g and 0.777 nmoles/g respectively. 101 3.4 DISCUSSION The potential use of phospholipid vesicles to manage atherosclerosis and other membrane associated, degenerative processes of the cardiovascular system is gaining interest (Williams et al., 1984; Aviram et al., 1989; Mendez et al., 1988; Mayer et al., 1989; Stein et al., 1988). Although it has been recognized for many years that liposomal infusion could promote the regression of atherosclerotic plaque (Williams et al., 1984), it is only recently that their use is being considered seriously as a mode of clinical therapy. In part this is due to the increasing use of liposomes in the pharmaceutical industry as topical (Mezei, 1988) and intravenous drug delivery vehicles (Cullis et al, 1989). As a result of this activity many problems associated with the cost and manufacture of liposomes on a large scale have been overcome. For example, there are now many suppliers, worldwide, of injection grade phospholipids in kilogram quantities, that use manufacturing processes that conform to the regulatory requirements for clinical use in humans. The literature available for the use of vesicle systems for intravenous drug delivery indicate the importance of vesicle size and composition on the circulation half-life and in vivo stability of liposomes in blood (Gregoriadis, 1988; Hwang, 1987). From the proposed mechanism of phospholipid induced cholesterol mobilization it is clear that the mass of sterol removed to the liver will be proportional to the concentration of liposomes that can be maintained for the first 24 h, during which time cholesterol equilibrates to an equimolar concentration with phospholipid of unilamellar vesicles (Figure 28). The finding that SUV30 were less effective than LUV100 in the mass of cholesterol moved from peripheral tissues to plasma, is consistent with the observation that EPC SUV30 are cleared more rapidly than EPC LUV. Although previous studies indicate that smaller particles circulate longer than large particles (Gregoriadis and Senior, 1986; Hwang, 1987), we are not aware of studies comparing cholesterol free SUV30 with LUV10o- It is likely that the relatively fast clearance of SUV30 from the circulation is due to several factors. First, with a diameter of 30 nm these vesicles can extravasate from the sinusoidal tissues of liver and spleen and, as a consequence, will 102 experience a greater volume of distribution than the LUV preparations, which cannot leave the vasculature as readily (Poste et al., 1984; Hwang, 1987; Scherphof et al., 1984). This is supported by observations that SUV accumulated in the liver to a greater extent than M L V 4 0 Q liposomes despite the fact that the latter were cleared more rapidly (data not shown). Scherphof and colleagues (1984) have elegantly shown that SUV can access liver hepatocytes whereas larger liposomes are confined to the Kupffer cell population lining sinusoid surfaces. Secondly, SUV are more unstable in plasma (Scherphof and Morselt, 1984), than large liposomes, and it is likely that a proportion of the small vesicles are assimilated into the lipoprotein pool and cleared rapidly by receptor mediated pathways (as well as phagocytosis) in the liver. We did not find any advantage to using vesicles composed of saturated lipids or using multilamellar systems. The latter observation is to be expected considering the decreased surface area of bilayer available for cholesterol absorption and the inability of cholesterol to rapidly equilibrate across the internal lamellae as discussed in Chapter 2. The primary source of liposome associated cholesterol is probably lipoproteins rather than circulating erythrocytes. Not only does total red blood cell cholesterol remain constant throughout the experiment but the rate of cholesterol turnover does not significantly change in the presence of liposomes, indicating that net sterol movement is occurring from other sources. The fact that the C:P ratio of lipoproteins from treated plasma drops sharply at the first time point (1 h) then climbs back to normal values over the first 8 h suggests that this cholesterol pool equilibrates rapidly with vesicles. After 8 h the circulating liposomes are approaching cholesterol equilibrium with their surroundings and net flux is reduced. It is known that cholesterol desorption from lipoproteins occurs more readily than from erythrocytes (Phillips et al., 1987), consequently the data support the hypothesis that liposomes generate cholesterol-poor lipoprotein particles (particularly HDL) which promotes the efflux of cellular cholesterol. In this way, sterol is shuttled from peripheral tissues to liposomes and reverse cholesterol transport is enhanced. Furthermore, several groups have demonstrated that there is a rapid 103 I i exchange of cholesterol between arterial tissue and lipoproteins (Stein and Stein, 1979; Phillips et al., and references therein), and recently it has been demonstrated that, in vivo, rat HDL can take-up cholesterol from endothelial cells and transport it to liver parenchymal cells (Bakkeren et al., 1990) which is the proposed route of reverse cholesterol transport. Consequently, it seems reasonable to predict that there is a net efflux of cholesterol from arterial wall into plasma during liposome infusion. Although it is estimated that the EPC LUV10o, used in these studies, mobilize approximately 2,800 nmoles of cholesterol, this may be a conservative estimate. The rate of increase in C:P ratio for vesicles calculated from the LUV10o a nd cholesterol plasma concentrations (Figure 26) is less than that measured in vesicles recovered from plasma (Figure 28). This apparent discrepancy can be explained if lipoprotein cholesterol decreases (as indicated by the data shown in Figure 32C) so that the "control" plasma cholesterol level (i.e. cholesterol not associated with vesicles) is lower than the 2,000 nmoles/ml plasma found in saline-treated mice. The C:P ratios shown in Figure 28 would result in cholesterol mobilization as high as 4,600 nmoles, approximately 50 moI% of the injected phospholipid. Our studies have been unable to detect the appearance of free cholesterol in the faeces of mice treated with liposomes. Furthermore, using [3H]-cholesterol loaded liposomes that excretion of radioactivity occurs at a constant rate over a 10 day period (data not shown). EPC LUVJQO mobilize approximately 3 nmoles of cholesterol in a typical EPC experiment. Less than half of this would be expected to appear in bile as free cholesterol even if all liposomal sterol entered the bile pathway, consequently a maximum 1.5 umoles of cholesterol might be excreted over a ten day period (less if re-absorption is taken into account), a level that is difficult to measure above background. The extraction and quantification of bile salts was also attempted, however, these results were too varied to detect differences between control and treated mice. In summary, from the observations reported here, it is concluded that liquid crystalline 100 nm LUV systems are the most effective liposomal preparation for cholesterol mobilization 104 in vivo. Although the data from Chapter 2 indicate that SUV are likely to mobilize cholesterol at a faster rate than LUV10o in vivo, there was not a large difference between SUV and LUV in the rate at which plasma cholesterol levels were increased. However, the fact that LUVJQO exhibit a longer circulation half-life than smaller or larger vesicles, results in this type of vesicle mobilizing more cholesterol for a given dose of phospholipid. Consequently, in the following chapter, the antiatherogenic properties of this optimized liposomal preparation are tested in an experimental model of atherosclerosis, achieved by feeding rabbits a diet rich in cholesterol. 105 CHAPTER 4 CHOLESTEROL MOBILIZATION AND REGRESSION OF ATHEROMA INDUCED BY LIPOSOMES 4.1 INTRODUCTION Coronary heart disease is the leading cause of death in North America, of which most cases are linked to atherosclerosis. Current preventive management of atherosclerotic disease has focused on the use of drugs in conjunction with dietary restrictions to regulate plasma cholesterol levels as there is accumulating evidence that the progression and accumulation of lipids in lesions can be halted when plasma LDL concentrations are kept to near normal levels (Reynolds, 1989). Moreover, antioxidant therapies which suppress the formation and the uptake of modified LDL particles by the cells of the arterial wall are also proving beneficial (Chisolm, 1991). However, while hypocholesterolemic drugs induce favourable plasma cholesterol changes which appear to slow the progression of atherosclerosis (Grundy, 1990b), or even promote the regression of atherosclerosis (Brown et al., 1990; Kane et al., 1990), the responses to treatment are varied, presumably because conditions that promote the efflux and removal of atheroma cholesterol are not stimulated. There is on the other hand, increasing evidence that processes which stimulate the efflux of extrahepatic cell cholesterol and transport it to the liver for excretion are important events in the prevention of atherosclerosis (Gwynne, 1991; Moncada et al., 1993). In the previous chapter the rate and extent to which liposomes of different sizes were able to induce cholesterol mobilization in vivo was investigated. Several parameters were important to achieve a liposomal preparation able to induce significant cholesterol mobilization. It was most important to maximize the surface area capable of accepting cholesterol and also to use a lipid composition and size that would increase circulation half-life of the liposomes 106 thereby providing sufficient time for cholesterol accumulation to occur. EPC LUV100 proved most effective in accumulating cholesterol in mice and were chosen as the formulation for further studies. Consequently, in this chapter, we examine the clearance profiles and the ability of repeated infusions of EPC LUV10o to promote cholesterol mobilization and facilitate the regression of atherosclerotic lesions in the cholesterol-fed rabbit. 4.2 MATERIALS AND METHODS 4.2.1 Materials EPC was supplied by Princeton Lipids (Princeton, NJ). The 0.5% cholesterol supplemented diet was obtained from Teklad Premier. Blood collection tubes and butterfly needles (23 gauge) were from Becton-Dickinson (Missisauga, Ontario). Ketamine, xylazine, heparin, Innovar and Euthanyl were supplied by MTC Pharmaceuticals, Janssen Pharmaceutica and Organon Technika (Ontario). Bio-Gel A-15m was purchased from Bio-Rad. Prepacked Solid Phase silica gel columns were acquired from Burdick & Jackson(Mi., MA). All chemical and solvents were of analytical grade and were from BDH Chemicals (Vancouver, B.C.) 4.2.2 Rabbits Forty eight New Zealand White (NZW) rabbits were housed in wire cages at the Animal Unit of the Research Centre (Shaughnessy Site) conforming to guidelines set by the Canadian Council on Animal Care and the University of British Columbia. The animals were maintained in a controlled temperature environment with a 12 hour dark/light cycle. Approximately 150g of food were given per animal per diem. Water was freely given. 107 4.2.3 Experimental design The correlation between hypercholesterolemia and the onset and progression of atherosclerosis in the rabbit is well established (St. Clair, 1983). It is known that the response of rabbits to dietary cholesterol is varied with some animals behaving as hyper-responders (large increase in plasma cholesterol) and others classified as hypo-responders. Consequently, the rabbits used in this study were carefully screened so that an equal distribution of hypo- and hyper-responders were included in each experimental group. The 48 NZW weanlings were screened by measuring their response to the 0.5% cholesterol-enriched diet (Teklad diet 0533). This involved feeding the diet for one week and monitoring plasma cholesterol concentrations until they returned to baseline. Animals were matched by the extent of the rise in their plasma cholesterol levels as well as the rate at which the levels returned to normal. This enabled an equal distribution of animals to be placed into two groups of 24 that were fed either standard rabbit chow or 0.5% cholesterol enriched rabbit chow for 20 weeks to induce atherosclerotic plaque formation (see experimental outline in Figure 33). During this time, plasma lipid levels were monitored on a monthly basis. Two animals were euthanized due to complications probably associated with handling and were not used in the final analyses. After the diet induction period, five animals per group were sacrificed to establish the formation of lesions and serve as the standard or baseline against which the effectiveness of liposomal treatment was assessed. Thereafter, all remaining animals were fed regular rabbit chow until the conclusion of the study. Based again on the pairing of plasma cholesterol concentrations, the 18 rabbits remaining from each diet group were further separated into groups of 9 and were treated with a total often bolus injections of EPC LUV10o a t a dose of 300 mg/kg or the equivalent volume of saline. Treatment was initiated 4 weeks after their return to standard rabbit chow and was given over a 100 day period and consisted of a single bolus injection of phospholipid at 300 mg/kg in a total volume of approximately 10 ml or saline administered into the marginal ear vein every 10 days. 108 Figure 33 Experimental Design Employed to Test the Efficacy of Liposomal Therapy Forty eight NZW rabbits were paired on the basis of their response to the 0.5% cholesterol enriched diet and were divided into two groups of 24 fed either the high cholesterol diet or standard rabbit chow. After 20 weeks of feeding, 5 animals per group were sacrificed to establish baseline atherosclerotic formation. Treatment with liposomes was initiated after all animals were returned to standard chow and consisted of one bolus injection of EPC LUV10o given once every 10 days for a total of 10 injections. 48 NZW RABBITS I PRESCREEN FOR RESPONDERS 24 RABBITS CHOLESTEROL DIET 1 24 RABBITS NORMAL DIET SACRIFICE TO ESTABLISH BASELINE RETURN TO NORMAL DIET I LIPOSOME TREATED 1 SALINE TREATED I LIPOSOME TREATED 1 SALINE TREATED 109 4.2.4 Vesicle preparation The rabbits ranged from 4-6 kg in weight and thus each treatment for the 18 rabbits receiving vesicles required the preparation of approximately 150 mis of LUV100 at a concentration of 200 mg/ml. Typically, 6 gram aliquots of EPC were hydrated with 30 ml of filtered HBS, pH 7.4, in sterile 50 ml conical tubes, vortexed and kept overnight. As previously described (section 2.2.3), the resulting MLV were used to generate LUV10o by extrusion. Vesicle diameters were determined by QELS and the average diameter of vesicles used for the 10 treatments were 114+7 nm. 4.2.5 Collection of blood and tissue samples Aproximately 100 ul of Innovar was given to promote calmness and vessel dilation in animals to ease routine bleedings necessary for plasma lipid analyses. To facilitate the final blood collection, ketamine (40 mg/kg) and xylazine (8 mg/kg) were given intramuscularly to sedate the animals. Approximately 10 min later, 50 units of heparin (Hepalean) followed by a lethal dose of phenobarbitol (Euthanyl) was perfused into the marginal ear vein before laparotomy. Organs were removed, rinsed in saline and immediately frozen in liquid nitrogen. The heart, with the entire aorta attached, was collected and kept in ice cold saline. The animals were sacrificed in groups of 8-10 on alternate days and the organs and aortas were randomized prior to processing and analyses. 4.2.6 Preparation of aortas for analyses Each aorta was separated from the heart at the aortic valve and was very carefully cleaned to remove any adherent adventitial fat. The aortas were then cut along the ventral surface, spread open, and photographed on a black surface. The developed photographs were later used in conjunction with the negatives to aid in the collection of digitization data as well as to facilitate the division of the aortas into three regions: the arch, the thoracic and abdominal segments as described by Rosenfeld et al. (1988). Of the nine animals from the 4 treatment 110 groups: (1) Saline-treated, cholesterol-fed animals, (2) vesicle-treated, cholesterol-fed animals, (3) saline-treated, standard diet animals and (4) vesicle-treated, standard diet animals, six aortas were stored at -20°C prior to lipid analyses and the remaining three samples were fixed in 10% neutral buffered formalin for at least 48 hours and used for gross staining with Sudan IV (Holman et al., 1958) and histology. At the time of lipid analyses, aortas were pat dry, divided into the three segments and, after determination of wet weight and length, they were homogenized (Polytron) in HBS. Two additional washes of the probe with HBS were collected for each segment to ensure quantitative recovery. 4.2.7 Digitization Digitization analyses involved illuminating the negatives obtained from all unstained aortas and generating an image using a Microcomputer Imaging Device (Imaging Systems). The % plaque involvement was calculated by dividing the area occupied by surface plaque over the area determined for the entire aorta which was possible due to distinct differences in the degree of shading of plaques versus uninvolved aortic tissue. Assessments of % atherosclerotic plaque involvement were performed by two observers and their results were averaged. Interobserver variation was within ± 5%. 4.2.8 Lipid analysis Cholesterol and phospholipid content of the aortas and livers were quantified after lipid extraction of the homogenates (Bligh and Dyer, 1959). Total cholesterol, cholesterol, and cholesterol esters were determined according to the method of Rudel and Morris (1973). Cholesterol and cholesterol esters were separated by silica gel chromatography on pre-packed 200 mg Solid Phase Silica Gel columns placed in 16x100 mm test tube carriers. Cholesterol esters were eluted with 1 ml methylene chloride whereas cholesterol was collected following elution with methylene chloride/methanol (95:5) after transferring the columns to a new carrier. I l l Phospholipid content was measured according to Fiske and Subbarow (1924). Lipoprotein lipid profiles were quantified by enzymatic procedures after phosphotungstic acid precipitation. 4.2.9 Protein analysis Aliquots of aorta or liver homogenates were incubated overnight at 37°C with 1 ml of IN NaOH. Thereafter, sodium dodecylsulphate (SDS) was added to the mixture to make a 1% (v/v) solution necessary to solubilize any remaining particulate matter. Protein content was quantified using the bicinchoninic acid (BCA) protein assay (Pierce Chemical Co., Rockford, IL) after incubation for 1 hour at 60°C and read at A^2 against an albumin standard. 4.2.10 Histochemical analysis of aortic samples Typically, 2-3 mm segments from the arch, thoracic, and abdominal aorta of each of the three animals within each treatment group were cut, further divided into left and right halves and embedded in paraffin. Depending on the length of the aorta, at least 8 segments from each region were prepared as blocks. Alternate sections of 5 fim were adhered to gelatin coated slides from paraffin blocks and visualized with haematoxylin and eosin (H&E) or Weigart's-van Gieson's stains. Intimal and medial widths of the different regions were measured and an intimal/medial ratio calculated for 3 photographs generated from each section then from these values a final mean ± standard deviation was determined for all the sections made from the three animals of each group. The morphology of plaques from animals sacrificed after the diet induction period (but prior to any treatment) was examined after sections were made from segments held into place with tissue mount on wooden stages and quick frozen, first in isopentane then in liquid nitrogen. Subsequently, alternate sections of 5 ^m were adhered to polylysine coated slides and visualized with Sudan IV differentiated with Harris' haematoxylin, H&E or van Gieson's stains to highlight lipids and collagen. 112 4.2.11 Statistical analyses Unless otherwise indicated values are presented as mean ± standard deviation. The significance of the difference of the means was assessed by an analysis of variance using the two-sample t test. Only values of P< 0.05 were considered significant. 4.3 RESULTS 4.3.1 Establishment of lesions Lesions induced in rabbits as a result of maintaining the animals on cholesterol-enriched diets for more than two months, do not regress even when animals are returned to normal rabbit chow for as long as two years (Prichard, 1974). In fact, even after the cessation of the diet, lesions have been noted to progress and increase in complexity (St. Clair, 1983). Moreover, in cases where intermittent feeding schedules were administered or a low cholesterol-enriched diet was given over a period of years, lesions similar to the calcified, ulcerated lesions observed in humans have been produced (Constantinides et al., 1960). As a consequence of this, and because of their easy maintainence plus reliability of lesion formation, the cholesterol-fed rabbit has become an ideal model for testing therapies that promote plaque reversal or regression. The animals in this study were fed a 0.5% cholesterol enriched diet for 20 weeks in order to induce intermediate lesions and not fatty streaks (M. E. Rosenfeld, personal communication). Indeed, chemical and histological analyses of the aortas obtained from animals after the diet induction period but prior to treatment, revealed that the plaques formed were rich in lipid and surrounded by fibrous tissue and consisted of almost equivalent amounts of cholesterol and cholesterol ester. The aortic phopholipid measured in these animals was 15 ± 4 pimol/g wet tissue and total cholesterol was 114 ± 28 (j.mol/g wet tissue of which there was 61 ± 13 \imo\lg cholesterol and 53 ± 15 ujnol/g cholesterol ester. Animals maintained on the standard diet had aortic phospholipid levels of 4 ± 0.3 nmol/g wet tissue and total cholesterol levels of 10 ± 1 nmol/g which was predominantly cholesterol. The degree of surface plaque involvement in cholesterol-fed animals was 78 ± 14%. 113 4.3.2 Cholesterol mobilization Despite the cessation of the high cholesterol diet, rabbits remain hypercholesterolemic for extended periods of time depending upon the duration of cholesterol feeding (Daugherty et al., 1986; Kovanen et al., 1981). During the course of this study, animals on the atherosclerotic diet exhibited plasma total cholesterol concentrations ranging from 5-10 times that of the control animals fed the standard diet. Moreover, plasma cholesterol remained elevated (2-5 times higher) until the conclusion of the study even though standard rabbit chow was given during the treatment period. This point is illustrated in a typical time course of cholesterol mobilization resulting from the infusion of 300 mg/kg EPC LUVJQQ or an equivalent volume of saline and shown in Figure 34. Animals previously fed the high cholesterol diet (panel A) still maintain plasma cholesterol concentrations three times higher than animals maintained on the standard diet throughout the study (panel B) even though the cholesterol diet was ceased 10 weeks earlier. Despite the excess plasma cholesterol in atherosclerotic animals, an injection of LUV 1 0 Q still resulted in a dramatic 2.5 fold increase in plasma cholesterol concentrations in both cholesterol-fed and normal animals when compared to the saline-treated controls. Plasma cholesterol levels peak at 24 hours before returning to baseline levels after 5 days which correlates with the removal of vesicles from the circulation determined by measuring total plasma phospholipid concentration and is illustrated in the clearance profiles shown in Figure 35. Atherosclerotic animals had slightly higher total phospholipid concentrations, nevertheless, similar clearance kinetics of the injected vesicles were seen for both groups. 114 Figure 34 Redistribution of Cholesterol into Plasma Following a Treatment with LUV100 at 300 mg/kg Cholesterol mobilization during a treatment. Animals were injected with EPC LUV10o at a dose of 300 mg/kg or an equal volume of saline via the marginal ear vein. At the indicated time intervals blood was drawn from the medial ear artery. Total cholesterol of plasma was determined for (A) saline-treated, cholesterol diet (O), vesicle-treated, cholesterol diet (•) ; (B) saline-treated, normal diet (D) and vesicle-treated, normal diet (•) fed animals. CB E w a CO o O CO o E 3 CB E to m a. o l a o ** CO CO o E 3 10 9 8 7 6 5 4 3 2 1 0 10 9 8 7 6 5 4 3 2 1 0 20 40 60 80 100 120 Time (hours) 115 Figure 35 Phospholipid Clearance During a Treatment Animals were injected with EPC LUV100 at a dose of 300 mg/kg or an equal volume of saline via the marginal ear vein and at the indicated time intervals blood was drawn from the medial ear artery. Total phospholipid concentrations of plasma were determined for saline-treated, cholesterol diet (O), vesicle-treated, cholesterol diet ( • ) , saline-treated, standard diet ( • ) and vesicle-treated, standard diet ( • ) fed animals as described in Methods. a E «» 0 TO a o £ a. CO o o E 3 120 Time (hours) 116 In Chapter 3, it was shown that the amount of cholesterol accumulated and removed by liposomes with each treatment is a function of the rate of cholesterol uptake and the rate of vesicle clearance. It was also shown that all the plasma cholesterol above baseline levels of saline-treated controls is associated with the circulating liposomes. This was determined by separating vesicles from plasma by gel filtration and analysing them to show that the excess plasma cholesterol was associated with the vesicles and that >90% of the cholesterol was free cholesterol and the remainder CE (data not shown). Consequently, as described in the previous chapter, an estimate of the mass of cholesterol removed from the circulation by the LUV10rj c a n be made by calculating the C:P ratios of vesicles at intervals during the time course following each injection using plasma phospholipid concentrations (vesicle-treated minus saline-treated concentrations) and cholesterol as the excess plasma concentration above the control at the various experimental time points. The plasma volume of the rabbits was approximately 150 ml, consequently, the total amount of phospholipid cleared from the circulation between time points was known. Using the average C:P ratio measured for vesicles between each assay interval an estimate of the cholesterol removed was obtained and is shown in Figure 36. The data represents an average ± standard deviation expressed as the mmoles of cholesterol removed by each treatment in the hypercholesterolemic animals and was calculated from measurements obtained during treatments 1, 4 and 10. The analysis was not continued beyond the point where less than 10% of the initial phospholipid dose remains in the circulation, because below this level the measurement error was too large to determine accurate C:P ratios. After 104 hours we estimate that approximately 1 mmol of cholesterol (387 mg) is removed from the circulation by the RES, which represents approximately 50 mol % of the injected phospholipid dose. Furthermore, we can estimate that based on plasma cholesterol concentrations measured in animals 24 h post-injections that each of the 10 infusions of liposomes caused dramatic cholesterol mobilization (data not shown). 117 Figure 36 An Estimate of Cholesterol Removed During an Injection of EPC LUV100 The amount of cholesterol removed by the RES resulting from an injection of 300 mg/kg EPC LUV100 was estimated the rates of cholesterol redistributed into plasma and phospholipid clearance. See section 4.3.2 for details. 73 O > O E © o ft. o w JS o o o E E 16 24 32 40 48 56 64 72 80 88 96 104 Time (hours) 118 4.3 J Effects of repeated injections The goal of this research is to obtain useful, pre-clinical data that can be used to support the development of repeated infusions of LUV10o as a clinical therapy in the management of atherosclerosis. Therefore, it was important to ascertain whether the RES was compromised in its capacity to remove liposomes from the circulation after multiple injections of phospholipid. It is estimated that each injection of 300 mg/kg EPC LUV10o induces a transient 100 fold increase in plasma phospholipid concentrations and on the average, at the end of liposomal therapy (10 injections) each animal received a total of 12-20 mmol (10-15g) of phospholipid. Although, there is evidence that long term, large doses of phospholipid are non-toxic in mice (Allen et al., 1984), we are unaware of studies which have monitored the effects of repeated phospholipid infusions on normal clearance function in rabbits. The clearance profiles of several injections of EPC LUV10o in cholesterol-fed rabbits is shown in Figure 37A. As illustrated, no significant differences in the rates of vesicle clearance between the first, fourth and tenth injections were detected. The data in Figure 37B also supports this observation. It shows that the concentrations of vesicle phospholipid remaining in the circulation, 24 h post-injection in both normal and cholesterol-fed animals is the same over the course of 10 injections. Furthermore, 5 days post-injections, the dose of vesicle phospholipid is completely removed from the circulation and thereafter plasma phospholipid and cholesterol concentrations always return to baseline levels. Consequently, there is no indication that there is accumulative damage to the RES during the treatment regime tested. 119 Figure 37 Consequences of Repeated Injections of EPC LUV100 (A) Clearance profiles of vesicles measured as plasma phospholipid concentration over time. Treatment 1 ( • ) , treatment 4 (•) , treatment 7 (A), and treatment 10 ( • ) . (B) Vesicle phospholipid concentrations remaining in the circulation 24 h-post injections in cholesterol-fed animals ( • ) and standard diet-fed animals (O). CO E CO « Q. JE xT a o JE O . CO o JC o. o E 3 120 144 Time (hours) 16 14 12 10 a E CO m CL I a. = 8 O JC a e CO o o E 3 -< N: I I —•— 1 M i Mr ^ i i B T 1 1 I 3 4 5 6 7 8 9 1 0 Injection number 120 At the conclusion of the study, saline-treated, cholesterol-fed animals still maintained elevated plasma cholesterol levels whereas vesicle-treated animals had levels comparable to animals maintained on the standard diet (data not shown). The reduction seen in plasma cholesterol concentrations of atherosclerotic animals treated with vesicles was a result of a reduction in both plasma LDL and HDL cholesterol concentrations, although the relative proportions of HDL/LDL cholesterol were not affected. No changes in the plasma lipid profiles (cholesterol, phospholipid or triglycerides) were detected in animals maintained on standard rabbit chow throughout the study (data not shown). Plasma phospholipid levels in treated animals were similar to their saline counterparts despite a total injection of approximately 15 grams of phosphatidylcholine per animal during liposomal therapy. These results, suggest that the repeated administration of LUV10o given at 10 day intervals does not appear to compromise RES function or normal plasma lipid homeostasis. 4.3.4 Source of Iiposomally accumulated cholesterol Total erythrocyte cholesterol remained constant throughout the infusions. In contrast, the C:P ratio of lipoproteins decreased over the first 24 hours and then gradually returned to normal levels after more than 48 hours (see Figure 38). This latter rise mirrors the time course of cholesterol accumulation by vesicles in the circulation (see Figure 34). These results are consistent with our earlier observations with mice (Chapter 3), and suggest that the lipoprotein pool of cholesterol in rabbits rapidly equilibrates with vesicles. In addition, these results support the hypothesis that treatment with liposomes leads to the generation of cholesterol-poor lipoprotein particles that may access peripheral tissues and promote cholesterol efflux from cells (Williams and Scanu, 1986 and result from Chapter 3 ). 121 Figure 38 ChoIesteroI-to-PhosphoIipid Ratios Measured in Lipoproteins During Vesicle Infusion The plasma lipoproteins were separated from EPC LUV10o by gel filtration and pooled. Their lipid content was measure following a Bligh and Dyer extraction. C:P ratios in saline controls ( • ) and vesicle-treated animals (•) are shown. CO o • a O o o Q. O Q. 4.00 0.00 8 12 16 20 24 28 32 36 40 44 48 Time (hours) 122 4.3.5 Assessment of atherosclerotic plaque involvement The extent of progression or regression of atherosclerotic lesions was assessed by three complementary methods: (1) chemical lipid and protein assays to determine lesion bulk, (2) assessment of the extent of plaque formation on the surface of aortas by image analyses, (3) histochemistry to examine the morphology and depth of the lesions. Aortic lipid content and digitization Results presented in Figure 39 summarize the extent of lipid infiltration of the whole aorta (WH) and of the three aortic segments designated as arch (AR), thoracic (TH), and abdominal (AB). The lipid content is expressed per gram of protein. It was found that, on average the protein content of aortas from treated and untreated animals did not differ significantly (saline-treated, 0.41 g protein/g wet weight and vesicle-treated, 0.43 g protein/g wet weight). The results show a 25% reduction in the cholesterol content of the whole aorta from vesicle-treated animals. However, the bulk of this loss occurred in the thoracic region which exhibited nearly a 50% loss of cholesterol compared to saline controls (Figues 39A). Results for the abdominal segment were very variable and although mean values for liposome-treated animals were 60% of those for saline-treated controls, these differences were not significant (P <0.1). A power analysis would suggest that increasing the number (though not feasible in this study) might reveal a true difference. The arch region was the least affected by vesicle treatment, but nevertheless the cholesterol content of treated aortas appeared reduced by approximately 10% at a significance level P<0.05. Similar trends were seen for CE content (Figure 39B) and for surface involvement (Figure 39C). In all cases the most significant and extensive reductions occurred for the thoracic region of vesicle-treated aortas. It is important to note that the cholesterol content of plaque after treatment with saline for 3 months did not change significantly from levels measured before treatment began. Prior to therapy, but after cessation of the cholesterol enriched diet, aortas were found to have 61 ± 13 jimoles cholesterol/g wet weight tissue whereas after saline treatment and a normal diet for 12 weeks 123 Figure 39 Assessment of Aortic Lipid Content and the Degree of Plaque Involvement (A) Cholesterol content of the whole (WH), arch (AR), thoracic (TH) and abdominal (AB) aorta expressed per g of protein weight. (B) Cholesterol ester content. (C) Extent of plaque involvement quantitated by digitization. The percent plaque involvement was determined from the area of surface plaque observed and the total area of the aorta being analysed. ^ - K o> ^» o E 3 "* C o o fca a O) i -o w ** M 0> O a 1100 1000 BOO 800 700 600 500 400 300 200 100 0 P<0.05 P< 0.1 800 WH AR TH AB SALINE-TREATED, CHOLESTEROL DIET VESICLE-TREATED, CHOLESTEROL DIET SALINE-TREATED, STANDARD DIET E223 VESICLE-TREATED, STANDARD DIET P<0.05 WH AB TH Region of Aorta AB 124 the level was essentially unchanged at 58 ± 6 nmoles cholesterol/g wet weight tissue. This demonstrates that there was no progression or regression of disease during the treatment period in the absence of vesicle therapy. Histochemical analysis of aortic samples Digitization of images was used to quantitate the extent of plaque involvement whereas histochemical analyses was used to determine the depth and nature of the lesions to be assessed. Extensive raised plaques (intimal thickening) were observed in the cholesterol-fed animals as expected from gross surface morphology inspection. Generally, the plaques exhibited extensive intimal thickening due to stratified lipid deposits that were surrounded by a collagenous network when compared to aortas obtained from normal control animals shown in Figure 40A. The arch region was noted to display more advanced lesions and contained what appeared to be deposits of crystalline cholesterol and showed a few isolated necrotic foci as detected by H&E staining (data not shown). Representative sections of the thoracic aorta of saline-treated and vesicle-treated animals are illustrated in Figure 40B and 40C, respectively. It can be seen that the lesions present in animals treated with vesicles (panel C) manifest less lipid deposits and show moderately reduced plaque thickening when compared to saline-treated, atherosclerotic animals (panel B). This is quantified in Table 7 which summarizes the data obtained from the analysis of pictures taken from multiple sections used to assess the severity of lesions present in the arch, thoracic or abdominal aorta of atherosclerotic animals. As can be seen, a decrease in the intima/medial ratios in the arch and thoracic regions of liposome-treated animals was measured whereas no changes were detected in the abdominal aorta. This analysis implies that, liposomally induced lipid depletion can be expected to alleviate lumen occlusion. The degree of vessel occlusion in the coronary arteries was not determined, however morphological changes seen in the aorta often reflect what occurs in these vessels also (Besterman, 1970). No apparent differences were detected in the aortas from vesicle-treated and saline-treated animals maintained on the standard diet throughout the study. 125 4.3.6 Liver cholesterol content Cholesterol feeding of rabbits often leads to the accumulation of cholesterol in a number of tissues including the liver. However, upon returning animals to regular rabbit chow, tissue cholesterol levels usually revert to normal within a month, with the exception of those found in arterial lesions (Prior and Ziegler, 1965). In this study liver cholesterol was measured in order to gain insight into whether (1) increased biliary excretion of cholesterol might be occuring in vesicle-treated animals owing to the increase levels of phospholipid being cleared through this organ which might result in reduced liver cholesterol levels or (2) whether accumulation of cholesterol could be detected as result of mobilized sterol being delivered with the liposomes (Adams et al., 1967). We found that in atherosclerotic animals, vesicle-treated rabbits demonstrated a slight reduction in liver cholesterol content at levels of 8 ± 3 nmol/g which compared to control animals fed the standard diet, whereas saline-treated cholesterol-fed animals exhibited levels of 11 ±4 u.moI/g. These differences, however, were not found to be statistically significant. Nevertheless, out results are consistent with the observations that the infusion of polyunsaturated phospholipids alleviate fatty livers in the cholesterol-fed rabbit (Adams et al., 1967). 126 Figure 40 Representative Sections Obtained from the Thoracic Aorta Animals fed the standard diet only do not exhibit any sign of plaque formation regardless of whether they were treated with vesicles or saline (Panel A). In contrast, cholesterol-fed animals had extensive intimal thickening and lesions were rich in lipid deposits that were surrounded by collagenous tissue (Panel B). However, in animals that received vesicle treatments the severity of the lesions appears slightly reduced in terms of both a depletion of lipid and a reduction in plaque thickening (Panel C). Sections were stained with Haematoxylin-Eosin, magnification X 10. mmtm " . - } v ' * f, I \<~\ ' ' i ' * " , . , . 7 * * - ^ • r J . . ' >l • f ?TT~ • inlimo w ^ ^ c »<fr-127 Figure 41 Schematic Representation of a Histological Section Used for the Measurement of Intimal/Medial Ratios FIBROUS (collagen) LIPID RICH VESSEL WALL (elastin) INTIMAL MEDIAL Table 7. Measurement of Intimal/Medial Ratios in the Different Regions of the Aorta of Vesicle and Saline-Treated Atherosclerotic Animals. Intimal/Medial Ratios Portion of aorta Arch Thoracic Abdominal Liposome-treated 1.51±0.55 1.34±0.73 1.84±0.95 Saline-treated 1.76±0.94 1.93+1.12 1.81±1.25 Significance (P value) N.S. P< 0.01 N.S. 128 4.4 DISCUSSION The results in this chapter show that repeated injections of EPC LUV100 results in a reduction of aortic cholesterol content, the degree of surface plaque involvement and lesion thickness in cholesterol-fed rabbits. In this respect, the data is in general agreement with the studies listed in Table 1, Chapter 1. Prior work (see summary in Table 1, section 1.1) demonstrating the ability of liposomal systems to promote the reversal of atheroma can be broadly categorized into two types of treatment protocol. One involves fewer treatments, but large doses of phospholipid per animal are infused over a period of several hours whereas the second type employs intensive, often bi-weekly, bolus injections of smaller phospholipid doses. The protocol employed here falls into the second category and is most similar to two independent studies by Altman et al. (1974) and Howard and co-workers (1971). These groups administered doses in the 200-300 mg/kg range and observed dramatic reversals of lesions. However, their liposomes were injected bi-weekly and by the end of the treatment regimen more total phospholipid was administered than the amounts received by the animals in this study. Early work, such as that reported by Byers and Friedman (1957) or Maraukas et al. (1959) employed massive doses, in excess of 1 g/kg, infused over many hours. The EPC LUVJOO formulation used in this rabbit study was administered as a bolus injection without visible discomfort to the animals. The half-life for circulation was approximately 40 h (see Figure 35), consequently, treating bi-weekly would have been possible and under these circumstances might lead to even more regression. In atheroma, most of the lipid found in monocytes/macrophages and smooth muscle cells of the arterial wall is localized in lysosomes and cytoplasmic inclusions, as well, lipid is found in extracellularly owing to rupture of cells (Schwartz, 1991). Studies with cell monolayers suggest that liposomes with C:P ratios lower than those of cells derived from the arterial wall, can induce the efflux of membrane cholesterol and the depletion of intracellular stores of sterol (reviewed in Phillips et al., 1987). However, at least two factors suggest it is unlikely that they do so directly in vivo to a significant extent. Firstly, circulating liposomes 129 interact most readily with plasma components and it is well documented that HDL rapidly gives up its cholesterol to phospholipid acceptors. Secondly, the lipids accumulated in advanced lesions are inacessible to direct contact with liposomes because not only are they localized deep within plaque, but they are also enclosed by collagenous tissue. On the basis of our current understanding, it is reasonable to hypothesize that liposomal infusion leads to a shift in the balance of free cholesterol in vivo. Liposomes promote efflux of tissue cholesterol into cholesterol-poor, phospholipid-rich HDL and its subpopulations and these in turn mediate the efflux of cholesterol from the arterial wall. In support of this Stein and Stein (1979) have shown that HDL can access lesions and is localized thoughout the atherosclerotic plaque. Moreover, Badimon et al. (1990) have recently claimed that the intravenous injection of HDL induces a regression of atherosclerosis in the cholesterol-fed rabbit. Besides these, there are a wealth of studies in vitro that show HDL can efficiently induce the efflux of cell membranes and intracellular cholesterol (reviewed in Johnson et al., 1991). Indeed, HDL has been demonstrated to provoke cholesterol loss from cholesterol loaded mouse peritoneal macrophages resulting in cholesterol ester hydrolysis and a net reduction of intracellular lipid deposits (Yau-Young et al., 1982) therefore reversing the foam cell morphology. It is generally agreed that removal of the lipid from lesions occurs more readily from fatty streaks and soft lesions whereas fibrous, stable lesions are more resistant. In this study, it is particularly encouraging that liposome therapy induced significant losses of cholesterol from the arterial wall despite the fact that lesions were visibly fibrous. Moreover, there was histological evidence that deposits of crystalline cholesterol were present throughout the lesions. Cholesterol crystals though more resistant to dissolution (Small, 1988) can still be solubilized in vitro with liposomes of HDL (Adams et al., 1978a and 1978b). Although macrophages are heavily involved in the formation of lesions, it has been suggested that perhaps they could also play a vital antiatherogenic role by stimulating their intrinsic ester hydrolysis 130 activity as well as collagenase and elastase activities to dissolve the fibrous network of lesions therefore enabling true sclerotic regression to occur (Koren et al., 1991). The data presented indicate that the most extensive regression occurred for lesions in the thoracic region of the aorta, which were observed to be softer and more maleable to the touch than lesions in the arch. In this regard it is interesting to note that >80% of fatal heart attacks are associated with plaque ruptures, where a "gruel" composed of extracellular deposits of cholesterol and CE is extruded into the vessel lumen through tears in the fibrous cap covering the lesion (Falk, 1992). Lipid that spills into the vessel lumen appears to block critical coronary arterioles and the damaged lesions promote the formation of blood clots. Only soft, lipid-rich plaque are rupture-prone and it is this type of lesion that is most amenable to liposome induced lipid regression. The hypothesis of Williams et al. (1984) suggests an efflux of aortic wall cholesterol via HDL into liposomes, which subsequently deliver their lipid load to the liver, where it can be metabolized and eventually excreted. Normally, RCT is believed to occur as HDL absorb cholesterol from peripheral tissues and deliver it to the hepatocytes, and in support of this, it has recently been demonstrated in vivo that HDL can mobilize cholesterol from blood vessel endothelial cells and transports it to parenchymal cells within the liver (Bakkeren et al., 1990). It is well documented, on the other hand, that liposomes are mostly sequestered by the fixed macrophages of the liver and spleen and to some extent by hepatocytes through endocytosis (Poste et al., 1984). Following extensive clearance by cells of the RES, it has been shown that lipid from vesicles is subsequently delivered to the hepatocytes in a time dependent manner, (Roerdink et al., 1981) and eventually excreted in bile (Esnault-Dupuy et al., 1987). Consequently, it can be seen that liposome infusions can augment RCT. This is further supported by a recent study from Mindham and Mayes (1991) in which an in vivo perfusion system showed that macrophages can efficiently deliver cholesterol from senescent erythrocytes to the liver for subsequent excretion in bile. Thus it seems reasonable to predict that the lipid load of vesicles sequestered by Kupffer cells undergoes similar processing and subsequent 131 excretion. Attempts by Byers and Friedman to label body cholesterol pools with [14C]-cholesterol in order to demonstrate RCT were unsuccessful (Byers and Friedman, 1956). However, they only sampled bile 9 hours after infusions and thus may not have allowed sufficient time to elapse for processing to occur. Similarly, during this reasearch numerous attempts have failed to show an increase in sterol or bile acid excretion in faeces resulting from liposomal infusion. It is evident that the future management and possible reversal of atherosclerotic disease, lies in pursuing therapies which enhance RCT whilst suppressing LDL levels (Moncada et al., 1993). Numerous experimental and clinical studies document of the benefits of decreased LDL concentrations and the prevention of oxidized LDL formation (Chisolm, 1991). Furthermore, there is strong epidemiological evidence that inherited traits or drug therapies which increase HDL activity reduce the risk of developing atherosclerosis (Arntzenius, 1991). Although the phospholipid dose of 300 mg/kg employed in this regression study is considerably higher than is presently used in liposomal drug delivery formulations, this lipid level is well below the amounts given to patients receiving daily parenteral nutrition in the form of Intralipid, for example. Moreover, the bolus injection of EPC LUV10o given to rabbits in this study was very well tolerated and thus could forseeably be administered in the clinic using rapid infusion times thus avoiding the need for costly hospitalization. The accumulation of cholesterol and CE in lesions is a gradual process that occurs over many years. It is therefore especially notable that studies in the literature and the research discussed in this thesis, suggest that liposome infusions can induce transient changes in cholesterol dynamics that are so dramatic that years of accumulation can be reversed druing just weeks of therapy. Individuals suffering from homozygous familial hypercholesterolemia (FH), lack function LDL receptors and exhibit excessively high concentration of plasma LDL cholesterol which predispose them to develop premature atherosclerosis (Illingworth and Bacon, 1989). To slow the progression of atherosclerosis, these patients are placed on drug therapy involving one or more hypocholesterolemic agents which have additive and complementary effects to 132 lower plasma cholesterol levels (Reynolds, 1989; Illingworth and Bacon, 1989). However, this is still often insufficient to bring levels into a low risk, normal range to prevent progression of disease. The ability of LUV10o to bring about a reduction of the cholesterol content of atheroma in spite of persisting elevated plasma cholesterol levels, suggests that vesicles could provide therapeutic benefits to this patient population via a mechanism that is also complementary and not competetive with currently available therapy. In summary, the repeated infusions of large unilamellar vesicles composed of EPC promotes a significant reduction in the lipid content of atheroma resulting in the debulking of lesions in the cholesterol-fed rabbit model of atherosclerosis. This regression was observed in spite of persisting elevated plasma cholesterol concentrations during the treatment period. 133 CHAPTER 5 SUMMARIZING DISCUSSION AND FUTURE DIRECTIONS The studies presented in this thesis were aimed at identifying a liposomal preparation capable of exhibiting optimal cholesterol mobilizing properties in vivo and to test its antiatherogenic effects. The rationale behind liposomal therapy has been the recognition that cholesterol readily undergoes exchange between membranes. Consequently, the injection of phospholipid vesicles into the circulation results in the formation of a chemical concentration gradient for cholesterol inducing it to diffuse from various pools and accumulate in circulating liposomes. Presumably, the effectiveness of a given liposomal preparation will be related to its ability to maintain the maximum amount ofsurface area into which cholesterol can be absorbed into. Experiments in Chapter 2 were designed to examine the kinetics of cholesterol exchange in three vesicle types considered for in vivo applications. SUV, LUV10o a nd MLV40o were used to determine the ease at which cholesterol could undergo transbilayer diffusion and its kinetics were compared with CS. CS intermembrane exchange was considerably faster than those observed for cholesterol for a given vesicle size, and in addition for both sterols, the rate of sterol exchange between vesicles was fastest from donor vesicles of smaller size. By subjecting the kinetic data to a three compartment model, the apparent rate of cholesterol transbilayer diffusion of cholesterol was estimated and was found to be slower than the rate of exchange. These results suggested that for equal amounts of phospholipid, the rate at which liposomal systems will absorb cholesterol in vivo would be on the order of SUV> L U V » MLV. These results were later confirmed in Chapter 3, which describes the influence of liposome composition and size on the rate and extent of cholesterol mobilization in mice. 134 SUV30 composed of EPC were found to rapidly increase plasma cholesterol concentrations, consistent with what was predicted from their kinetic behaviour in Chapter 2. However, these systems were cleared more rapidly than LUV100 of the same composition, which accumulate cholesterol more slowly. Consequently, when the rates of cholesterol influx and the rates of vesicle clearance were combined it was shown that LUVKJO were able to mobilize twice as much cholesterol than SUV30. The source of liposomal cholesterol appears to be mainly from the lipoprotein pool, rather than from erythrocytes. Moreover since mice are mainly HDL animals, our results support the observations of Williams and Scanu (1986), that liposomal infusions result in compositional changes in HDL that likely enhances more sterol efflux from peripheral tissues. This rapid efflux of cholesterol from HDL has also been observed in vitro. It seems that cholesterol efflux rates are dependent in part on lipoprotein particle size and the nature of the HDL surface, such that cholesterol exhibits much faster rates of efflux from HDL than from LDL and half-life values of roughly 3 and 45 minutes respectively have been reported (Phillips et al., 1987 and Table 4). The capacity of LUV10o to cause a dramatic mobilization of cholesterol is due to their large surface area and long circulation time which enables maximum equilibration of cholesterol to occur. Both SUV30 and LUV50 were cleared more quickly than LUVJQQ, reducing their capacity to mobilize cholesterol despite exhibiting rapid absorption rates. On the other hand, MLV400 exhibit slow absorption of cholesterol as well as rapid clearance and therefore were not comparable to LUV100 in their capacity to remove sterol to the RES. Interestingly, LUV100 composed of gel-state lipids accumulate cholesterol at an identical rate to EPC LUVjog, but their rapid clearance dramatically reduces their ability to raise plasma cholesterol levels. As a result of these studies in mice, EPC LUV100 were chosen as the liposomal system most likely to exhibit the greatest antiatherogenic properties, and in Chapter 4 their efficacy is evaluated in the cholesterol-fed rabbit model. The cholesterol feeding regimen chosen in Chapter 4, induced the formation of advanced lesions with massive accumulations of cholesterol and CE in the walls of the aorta. 135 Vesicle infusions caused a massive mobilization of cholesterol which promoted significant reductions in aortic cholesterol content, surface plaque involvement and lesion bulk as determined from intimal/medial ratios in treated animals compared to saline controls. Moreover, since the lipid content and surface plaque involvement of saline-treated animals were not significantly different from animals sacrificed prior to the start of any treatment, the differences observed between liposome and saline-treated animals represents regression of atherosclerosis and not simply reduced progression. Furthermore, vesicle infusions do not appear to adversely affect RES function or the long term homeostasis of plasma lipid as the rate of phospholipid clearance following each infusion appeared similar and plasma lipid levels were always found to return to baseline levels after 5 days. On the whole, the regression studies employing liposomal infusions are impressive. Dramatic results have also been reported following LDL apheresis (Keller, 1991). This involves the ex vivo removal of LDL from plasma as blood from a patient is washed through an affinity column that selectively binds the lipoprotein. The removal of LDL from the circulation, radically reduces plasma cholesterol levels and apparently promotes the efflux of cholesterol from peripheral tissues. Although LDL apheresis can promote significant regression, the procedure is highly invasive and requires long and costly dialysis times. The repeated infusion of liposomes should be as effective as apheresis and more straightforward to administer. Overall, this thesis has provided evidence for the potential benefits of liposome induced cholesterol mobilization and strongly supports the clinical application of liposomes. However, the cholesterol-fed rabbits develop atherosclerosis as a result of hypercholesterolemia associated with increase levels of VLDL. A better model of the human disease is the Watanabe Heritable Hyperlipidemic (WHHL) rabbit, because animals lack LDL receptors and, like humans exhibit high plasma concentrations of LDL and slowly develop atherosclerotic lesions throughout life. Furthermore, results from WHHL rabbit models are acceptable as pre-clinical data to support submissions for clinical trials. Consequently, the next phase of pre-clinical development of LUV100 therapy might target this model. 136 Current treatment for hypercholesterolaemic patients exhibiting signs of atherosclerosis involves the use of a variety of cholesterol lowering drugs. For example, bile acid binding resins, promote the excretion of bile salts, which in turn directs more cholesterol to bile synthesis and stimulates expression of LDL receptors. Another class of drugs common in this area are the statins, which inhibit HMG-CoA reductase, the rate limiting enzyme in cholesterol biosynthesis. Recent clinical trials have demonstrated that, in combination, these two types of drugs act additively in reducing plasma cholesterol levels in many patients (reviewed in Moncada et al., 1993). Because liposomal therapy acts via a separate mechanism to both of these drugs, it is reasonable to expect that its effect will also be additive, giving rise to significant improvements in therapeutic benefit. Consequently, a study involving WHHL rabbits should also include data that not only compares efficacy to that produced by current therapy, but efficacy from combination therapies. Moreover, given that liposomes carry cholesterol from peripheral tissues to the liver for excretion, it would be particularly interesting to determine the combined efficacy of LUV10o infusions in conjunction with bile acid resins. Future work that would also build on the studies presented in this thesis include determining the in vivo fate, processing and excretion or redistribution of liposomal phospholipid and cholesterol. This is particularly important in the light of recent observations that infusions of SUV cause a significant rise in both plasma cholesterol and CE levels (Williams et al., 1988). Increases in plasma CE imply that either the secretion of LDL is increased or the expression of LDL receptors is downregulated. In either case, this would not be desirable, as increases in LDL concentrations are linked to atherosclerosis. It should be noted that the administration of EPC LUV 1 0 Q at 300 mg/kg resulted only in a transient increases in plasma cholesterol only and not CE. The reasons for this potential differences in the physiological response to SUV and LUV10o is not clear at this time, however, a possibility is discussed below. Approximately 70% of EPC liposomes are removed by Kupffer cells in the liver. However SUV are sufficiently small to escape the sinusoidal network and access hepatocytes 137 (see section 1.6.1, Introduction). Therefore a significant proportion of small vesicles may deliver cholesterol directly to the parenchymal cells of the liver, which are responsible for cholesterol metabolism. This may stimulate the cells to make apoB-rich lipoproteins (such as LDL) so that excess sterol can be excreted into the lipoprotein pool (Dietschy, 1990). LUV10o> on the other hand, would not give rise to this response because they would be metabolized through Kupffer cells and lipid processed by the RES may be directed to parenchymal cells via a pathway that is linked to excretion rather than recycling, though this is purely speculative. Another possibility would be the rate at which cholesterol is delivered to the liver. SUV absorb cholesterol more rapidly than LUV10o and are also cleared by the liver more quickly. It is possible that the normal cellular control mechanisms for cholesterol metabolism are overwhelmed by this influx and the response is to shut down HMG-CoA reductase, reduce LDL receptor numbers and increase apoB secretion (Dietschy, 1990). Whereas LUV100 deliver cholesterol to the liver at a slower rate (cholesterol is steadily delivered steadily over 3 days, Figure 36, Chapter 4), and decreasing HMG-CoA reductase is sufficient for the cell to control the influx (Kevin J. Williams, personal communication). The next phase of development must include experiments in which animals that closely reflect human lipoprotein metabolism such as the hamster (Meddings et al., 1987), are treated with SUV and LUVjoo and the plasma analysed direcctly for increased apoB, LDL and CE. It will also be important to determine the fate of liposomal cholesterol. As described in Chapters 3 and 4 excess cholesterol could not be detected in the faeces of mice and rabbits following liposomal therapy. Consequently, the direct analyses of bile following the intravenous administration of liposomes may enable the detection of secreted cholesterol. The use of rats would be convenient for these studies because they lack gall bladders and hence remove the complication of having to account for bile storage. Finally, questions surrounding the mechanisms by which liposomes enhance RCT still remain. Although the decrease in cholesterol content of lipoproteins measured following infusion suggests that they are depleted of sterol by liposomes, it is unclear how these 138 cholesterol-poor, phospholipid-rich particles work to promote cholesterol removal from tissues. Since liposomes accumulate mainly free cholesterol, it is likely that HDL particles still retain their CE core and thus remain spherical (see section, Introduction). There is, however, increasing evidence that discoidal HDL particles are the most efficient at promoting the efflux of cholesterol from cell surfaces (Castro and Fielding, 1988; Rothblat et al., 1992). Furthermore, if studies could be directed to determine the identity and relative contribution of "lipoprotein-like" particles, generated during vesicle-HDL interactions (Williams and Scanu, 1986; Krupp et al, 1976), the efficacy of a liposomal preparation could be improved. Of considerable interest is the structural and compositional changes that occur to HDL following their interaction with vesicles. Do HDL particles, depleted of cholesterol, act to shuttle sterol between cells and liposomes? Can they extravasate or exhibit a higher affinity for the cell surface and therefore act more efficiently as cholesterol acceptors? Are prep"-HDL particles generated or maintained? If liposomes can make HDL particles more effective, or generate "HDL-like" particles capable of full participation in RCT, then they could represent a more general form of therapy for the large number of patients at risk of developing atherosclerosis due to low levels of HDL (Khachadurian, 1992). In summary, the research presented in this thesis confirms the antiatherogenic properties of liposomes. A well characterized liposomal system was developed which can be prepared cost effectively and on a large scale, using pharmaceutical grade phospholipids. Consequently, following positive efficacy data from WHHL rabbits, there is no reason to believe this formulation should not be clinically tested. 139 BIBLIOGRAPHY Adams, C.W.M., Abdulla, Y.H., Bayliss, O.B. and Morgan, R.S. (1967) Modification of aortic atheroma and fatty liver in cholesterol-fed rabbits by intravenous injection of saturated and polyunsaturated lecithins. J. Pathol. Bacteriol. 94, 77-87. Adams, C.W.M. and Abdulla, Y.H. (1978a) The action of human high density lipoprotein on cholesterol crystals, Part 1: Light-microscopic Observations. Atherosclerosis 31, 464-471. Adams, C.W.M. and Abdulla, Y.H. (1978b) The action of human high density lipoprotein on cholesterol crystals, Part 2: Biochemical Observations. Atherosclerosis 31, 473-480 Allen, T.M., Murray, L., MacKeigan, S. and Shah, M. (1984) Chronic liposome administration in mice: effects on reticuloendothelial function and tissue distribution. J. Pharm. Exp. Ther. 229, 267-275. Altman, R.F.A., Mendon§a, J.M., Schaeffer, G.M.V., Ramos de Souza, J., Bandoli, J.G., da Silva, D.J. and Lopes, C.R.N. (1974) Phospholipids in experimental atherosclerosis, Arzneim.-Forsch. 24, 11-16. Arbogast, L.Y., Rothblat, G.H., Leslie, M.H. and Cooper, R.A. (1976) Cellular cholesterol ester accumulation induced by free cholesterol-rich dispersions. Proc. Natl. Acad. Sci. (USA) 73, 3680-3684. Arntzenius, A.C (1991) Regression of atherosclerosis. Benefit can be expected from low LDL -C and high HDL -C levels (Editorial). Acta Cardiologica 22, 441-442. Arvinte, T. and Hildenbrand, K. (1984) JV-NBD-L-ct-dilauroylphosphatidylethanolamine. A new fluorescent probe to study spontaneous lipid transfer. Biochim. Biophys. Acta 775, 86-94. Aviram, M., Williams, K.J. and Deckelbaum, R.J. (1988) Macrophage cholesterol removal by triglyceride-phospholipid emulsions. Biochem. Biophys. Res. Comm. 156, 709-713. Aviram, M., Williams, K.J., Mcintosh, R.A., Carpentier, Y.A., Tall, A.R. and Deckelbaum, R.J. (1989) Intralipid infusion abolishes ability human serum to cholesterol-load cultured macrophages. Arteriosclerosis 9, 67-75 Backer, J.M. and Dawidowicz, E.A. (1979) The rapid transmembrane movement of cholesterol in small unilamellar vesicles. Biochim. Biophys. Acta 551, 260-270. Backer, J.M. and Dawidowicz, E.A. (1981) Transmembrane movement of cholesterol in small unilamellar vesicles detected by cholesterol oxidase. / . Biol. Chem. 256, 586-588. 140 Badimon, J.J., Badimon, L. and Fuster, V. (1990) Regression of atherosclerotic lesions by high density lipoprotein plasma fraction in the cholesterol-fed rabbit. J. Clin. Invest. 85, 1234-1241. Bakkeren, H.F., Kuipers, F., Vonk, R.J. and van Berkel, T.J.C. (1990) Evidence of reverse cholesterol transport in vivo from liver endothelial cells to parenchymal cells and bile by high density lipoprotein. Biochem.J. 268, 685-691 Bangham, A. D. (1968) Membrane models with phospholipids. Progr. Biophys. Mol. Biol. 18, 29-95. Bar, L.K., Barenholz, Y. and Thompson, T.E. (1987) Dependence on phospholipid composition of the fraction of cholesterol undergoing spontaneous exchange between unilamellar vesicles. Biochemistry 26, 5460-5465. Bar, L.K., Barenholz, Y. and Thompson, T.E. (1986) Fraction of cholesterol underoing spontaneous exchange between small unilamellar vesicles phosphatidylcholine vesicles. Biochemistry 25, 6701-6705. Barenholz, Y., Amselem, S. and Lichtenberg, D. (1979) A new method for preparation of phospholipid vesicles (liposomes) by French press. FEBS Lett. 99, 210-214. Besterman, E.M.M. (1970) Experimental Coronary Atherosclerosis in Rabbits. Atherosclerosis 12, 75-83 Bisgaier, C.L. Siebenkas, M.V. and Williams, K.J. (1989) Effects of apolipoprotein A-IV and A-I on the uptake of phospholipid liposomes by hepatocytes. / . Biol. Chem. 264, 862-866. Bleau, G., Lalumiere, G., Chapdelaine, A., and Roberts, K.D. (1974) Cholesterol sulfate I. Occurence and possible biological function as an amphipathic lipid in the membrane of the human erythrocyte. Biochim. Biophys. Acta 352, 1-9 Bligh, E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37, 911-917 Bloj, B. and Zilversmit, D.B. (1977) Complete exchangeability of cholesterol in phosphatidylcholine/cholesterol vesicles of different degrees of unsaturation. Biochemistry 16, 3943-3948. Bonte, F. and Juliano, R.L. (1986) Interactions of liposomes with serum proteins. Chem. Phys. Lipids 40, 359-372. Brown, G.B., Albers, J.J., Fisher, L.D., Schaefer, S.M., Lin, J.T., Kaplan, C.A., Zhao, X.Q., Bisson, B., Fitzpatrick, V.F. and Dodge, H.T. (1990) Regression of coronary artery disease as a result of intensive lipid-lowering therapy in men with elevated levels of apolipoprotein B. New England J. Med. 323,1289-1298. 141 Brown, M.S. and Goldstein, J.L (1980) The cholesteryl ester cycle in macrophage foam cells. J. Biol. Chem. 255, 9344-9352. Brown, M.S. and Goldstein, J.L (1986) A receptor-mediated pathway for cholesterol homeostasis. Science 232, 34-47. Bruckdorfer, K.R., Crowe, J. and Sherry, M.K. (1984) Evidence for a water-soluble intermediate in exchange of cholesterol between membranes. Biochim. Biophys. Acta 778, 489-496. Byers, S.O. and Friedmain M. (1956) Independence of phosphatide induced hypercholesteremia and hepatic function. Proc. Soc. Exp. Biol. Med. 92, 456-462. Byers, S.O. and Friedman, M. (1969) Probable sources of plasma cholesterol during phosphatide induced hypercholesterolemia. Lipids 4, 123-128. Byers, S.O. and Friedman, M. (1960) Effect of infusions of phosphatides upon the atherosclerotic aorta in situ and as an ocular aortic implant. J. Lipid Res. 1, 343-348. Byers, S. O., Friedman, M. and Sugiyama, T. (1962) Mechanism underlying phosphatide induced hypercholesterolemia. J. Biol. Chem. 237, 3375-3380. Castro, G.R. and Fielding, C.J. (1988) Early incorporation of cell-derived cholesterol into prebeta-migrating high density lipoprotein. Biochemistry 27, 25-29. Chisolm, G.M., (1991) Antioxidants and Atherosclerosis: A current assessment. Clin. Cardiol. 14, 25-30 Chobanian, J.V., Tall, A.R., Brecher, P.I. (1979) Interaction between unilamellar egg yolk lecithin vesicles and human high density lipoprotein. Biochemistry 18, 180-187. Chonn, A., Semple, S.C. and Cullis, P.R. (1991) Separation of large unilamellar liposomes from blood components by a spin column procedure: towards identifying plasma proteins which mediate liposome clearance in vivo. Biochim. Biophys. Acta 1070, 215-222. Constantinides, P., Booth, J. and Carlson, G. (1966) Production of advanced cholesterol atherosclerosis in the rabbit. Arch. Pathol. 70, 81-92. Cullis, P.R., Mayer, L.D., Bally, M.B., Madden, T.D. and Hope, M.J. (1989) Generating and loading of liposomal systems for drug delivery applications. Adv. Drug Delivery Rev. 3, 267-282 Darke, A., Finer, E.G., Flook, A.G. and Phillips, M.C. (1972) Nuclear magnetic resonance study of lecithin-cholesterol interactions. J. Mol. Biol. 63, 265-279. 142 Daugherty, A., Schonfeld,G., Sobel, B.E. and Lange, L.G. (1986) Metabolism of very low density lipoproteins after cessation of cholesterol feeding in rabbits. A factor potentially contributing to the slow regression of atheromatous plaques. / . Clin. Invest. 77, 1108-1115. Dawidowicz, E.A. and Backer, J.M. (1981) The rapid transbilayer movement of thiocholesterol in small unilamellar phospholipid vesicles. Biochim. Biophys. Acta 644, 373-375. Demel, R.A., Jansen, J.W.C.M., Van Dijck, P.W.M. and Van Deenen, L.L. (1977) The preferential interaction of cholesterol with different classes of phospholipids. Biochim. Biophys. Acta 465, 1-10 Derksen J.T.P, Morselt, H.W.M, and Scherphof, G.L. (1987) Processing of different liposome markers after in in vitro uptake of immunoglobulin-coated liposomes by rat liver macrophages. Biochim. Biophys. Acta 931, 33-40. Devaux, P.F. (1991) Static and dynamic lipid asymmetry in cell membranes. Biochemistry 30, 1163-1179. Dietschy, J.M. (1990) LDL cholesterol: its regulation and manipulation. Hospital Practice (office edition) 25, 67-78. DiLuzio, N.R. and Zilversmit, D.B. (1960) Effect of intravenous phospholipid and Triton on lipids of normal and ethanol treated rats. Am. J. Physiol. 199, 991-994. Elias, P.M., Williams, M.L., Maloney, M.E., Bonifas, J.A., Brown, B.E., Grayson, S. and Epstein Jr., E.H. (1984) Stratum corneum lipids in disorders of comification. Steroid sulfatase and cholesterol sulfate in normal desquamation and the pathogenesis or recessive x-linked ichthyosis. / . Clin. Invest. 74, 1414-1421. Esnault-Dupuy,C, Chanusssot, F. Lafont, H., Chabert, C, and Hauton, J. (1987) The relationship between HDL-, LDL-, liposomes-free cholesterol, biliary cholesterol and bile salts in the rat. Biochimie 69, 45-52. Falk, E. (1992) Why do plaques rupture? Circulation Supplement 3, 30-42 Feingold, K.R. and Elias, P.M. (1993) Exogenous lipids influence permeability barrier recovery in aceton treated murine skin. Arch. Dermatol. 129, 728-738. Ferrell, J.E., Lee, K.J. and Huetis, W.H. (1985) Lipid transfer between phosphatidylcholine vesicles and human erythrocytes: exponential decrease in rate with increasing acyl chain length. Biochemistry 24, 2857-2864. 143 Fielding, C.J. (1990) Lecithimcholesterol acyltransferase. In Advances in cholesterol research. Ed. M. Esfahani and J.B. Swaney. pp. 271-314, Telford, Caldwell, NJ Fielding, P.E and Fielding, C.J. (1991) Dynamics of lipoprotein transport in the circulatory system, in Biochemistry of lipids, lipoproteins and membranes. Ed. D.E. Vance and J. Vance JVew Comprehensive Biochemistry Vol. 20 Elvesier, Amsterdam, pp. 427-459. Fiske, C.H. and Subbarow, Y., The colorimetric determination of phosphorus (1924)/. Biol. Chem. 66, 375-400. Franceschini, G., Maderna, Paola and Sirtori, C.R. (1991) Reverse cholesterol transport:physiology and pharmacology. Atherosclerosis 88, 99-107. Friedman, M., Byers, S.O. and Rosenman, R.H. (1957) Resolution of aortic atherosclerotic infiltration in the rabbit by phosphatide infusion. Proc. Soc. Exp. Biol. Med. 95, 586-588. Friedman, M., and Byers, S.O. (1956) Role of hyperphospholipidemia and neutral fat increase in plasma in the pathogenesis of hypercholesterolemia. Am. J. Physiol. 186, 13-18. Friedman, M and Byers, S. O. (1958) Source of excess plasma cholesterol in phosphatide-induced hypercholesterolemia. Am. J. Physiol. 195, 185-188. Gennis, R.B. (1989) In Biomembranes molecular structure and function. Springer-Verlag, New York Glomset, J.A. (1968) The plasma lecithimcholesterol acyltransferase reaction. / . Lipid Res. 25, 714-720. Gregoriadis, G. and Ryman, B.E. (1972) Fate of protein-containing liposomes injected into rates: an approach to the treatment of storage diseases. Eur. J. Biochem. 24, 485-491. Gregoriadis, G (1988) Fate of injected liposomes: observations on entrapped solute retention, vesicle clearance and tissue distribution in vivo. In Liposomes as drug carrier: recent trends and progress. Gregoriadis, G., ed. John Wiley and Sons Ltd., pp. 3-18. Gregoriadis, G. and Senior, J. (1986) In: Targeting of Drugs with Synthetic Systems. Ed. Gregoriadis, G., Senior, J. and Poste, G. Life Sciences 113, 183-192 Grundy, S.M. (1990a) In Cholesterol and atherosclerosis. Diagnosis and treatment. Ed. W.J. Gabello, Gower Medical Publishing, New York. p. 1.29 Grundy, S.M. (1990b) Cholesterol and coronary heart disease; future directions. J. Am. Med. Assoc. 264, 3053-3059. 144 Guo, L.S.S., Hamilton, R.L. Goerke, J., Weinstein, J.N. and Havel, R.J. (1980) Interaction of unilamellar liposomes with serum lipoproteins and apolipoproteins. / . Lipid Res. 21, 993-1003. Gurr, M.I. and Harwood, J.L. (1991) In Lipid Biochemistry. An introduction. 4th Ed. Chapman and Hall, New York Gwynne, J.T., (1991) HDL and Atherosclerosis: An update. Clin. Cardiol. 14,1-17-24. Gwynne, J.T. (1989) High-density lipoprotein cholesterol levels as a marker of reverse cholesterol transport. Am. J. Cardiol. 64, 10G-17G. Hale, J.E. and Schroeder, F. (1981) Asymmetric transbilayer distribution of sterol across plasma membranes determined by fluorescence quenching of dehydroergosterol. Eur. J. Biochem. 122, 649-661. Havel, R.J., Yamada, N and Shames, D.M., Watanabe Heritable Hyperlipidemic Rabbit: An animal model for Familial Hypercholesterolemia. (1989) Arteriosclerosis Supplement 1, 33-38 Ho, Y.K., Brown, M.S. and Goldstein, J.L. (1980) Hydrolysis and excrestion of cytoplasmic cholesteryl esters by macrophages: stimulation by high density lipoprotein and other agents. / . Lipid Res. 21, 391-398. Holman, R.L., McGill, H.C., Strong, J.P. and Gear, J.C., (1958) Technics for studying atherosclerotic lesions. Lab. Invest. 7, 42-47. Hope, M.J., Nayar, R., Mayer, L.D. and Cullis, P.R. (1993) Reduction of liposome size and preparation of unilamellar vesicles by extrusion techniques, in Liposome Technology Vol I Ed. Gregory Gregoriadis. CRC Press Hope, M.J., Bally, M.B., Webb, G. and Cullis, P.R. (1985) Production of large unilamellar vesicles by a rapid extrusion procedure: characterization of size, trapped volume and ability to maintain a membrane potential. Biochim. Biophys. Acta 55-65 Hope, M.J., Bally, M.B., Mayer, L.D., Janoff, A.S and Cullis, P.R. (1986) Generation of multilamellar and unilamellar phospholipid vesicles. Chem. Phys. Lipids 40, 89-107 Houslay, M.D and Stanley, K.K., (1982) in Dynamics of biological membranes. Influence of synthesis, struction and function. John Wiley & Sons Ltd. Howard, A.N., Patelski, J. Bowyer, D.E. and Gresham, G.A., (1971) Atherosclerosis induced in hypercholesterolaemic baboons by immunological injury: and the effects of intravenous polyunsaturated phosphatidyl choline. Atherosclerosis 14, 17-29. 145 Huang, C.H. (1969) Studies on phosphatidylcholine vesicles: formation and physical characteristics. Biochemistry 8, 344-352. Huang, C.H. (1977) A structural model for the cholesterol-phosphatidylcholine complexes in bilayer membranes. Lipids 12, 348-356. Hwang, K.J. (1987) In: Liposomes: From Biophysics to Therapeutics, M.J. Ostro Ed. pp. 109-156, Marcel Dekker, Inc. New York Hyslop, P.A., Morel, B. and Sauerheber, R.D. (1990) Organization and interaction of cholesterol and phosphatidylcholine in model bilayer membranes. Biochemistry 29, 1025-1038. Illingworth, D.R. and Bacon, S. (1989) Treatment of heterozygous Familial Hypercholesterolemia with lipid-lowering drugs, Arteriosclerosis Supplement 1,121-134 Intralipid reference Johnson, W.J., Chacko, G.K., Phillips, M.C. and Rothblat, G.H. (1990) The efflux of lysosomal cholesterol from cells. J. Biol. Chem. 265, 5546-5553. Johnson, W.J., Mahlberg, F.H., Rothblat, G.H., and Phillips, M.C, (1991) Cholesterol transport between cells and high-density lipoproteins Biochim. Biophys. Acta 1085, 273-298. Jonas, A. and Maine, G.R. (1979) Kinetics and mechanism of phosphatidylcholine and cholesterol exchange between single bilayer vesicles and bovine serum high density lipoprotein. Biochemistry 18, 1722-1728. Kan, C.C. and Bittman, R. (1991) Spontaneous rate of sitosterol and cholesterol exchange between phospholipid vesicles and between lysophospholipid dispersions: evidence that desorption rate is impeded by the 24a-ethyl group of sitosterol. / . Am. Chem. Soc. 113, 6650-6656. Kan, C.C, Yan, J. and Bittman, R. (1992) Rates of spontaneous exchange of synthetic radiolabeled sterols between lipid vesicles. Biochemistry 31, 1866-1874 Kane, J.P., Mallow, M.J., Ports, T.A., Phillips, N.R., Diehl, J.C and Havel, R.J. (1990) Regression of coronary atherosclerosis during treatment of familial hypercholesterolemia with combined drug treatment regimens. JAMA 264, 3007-3012. Kawato et al. (1978) Effect of cholesterol on the molecular motion in the hydrocarbon region of lecithin bilayers studied by nanosecond fluorescence techniques. Biochemistry 17, 5026-5031. Keller, C (1991) LDL-apheresis: results of longterm treatment and vascular outcome. Atherosclerosis 86,1-8. 146 Khachadurian, A. (1991) Role of cholesterol in atherosclerosis, in Structural and dynamic properties of lipids and membranes (ed. R.J. Cherry and P.J. Quinn), Portland Press, London, pp. 385-419 Koren, E., Koscec, M. McConathy, W.J. and Fugate, R.D. (1991) Possible role of macrophages in the regression of atherosclerosis. Prog. Lipid Res. 30, 237-243. Kovanen, P.T., Brown, M.S., Basu, S.K., Bilheimer, D.W. and Goldstein, J.L., Saturation and suppression of hepatic lipoprotein receptors: A mechanism for the hypercholesterolemia of cholesterol-fed rabbits. (1981) Proc. Natl. Acad. Sci., (USA) 78,1396-1400. Krupp, L., Chobanian, A.V. and Brecker, P.I. (1976) The in vivo transformation of phospholipid vesicles to a particle resembling HDL in the rat. Biochem. Biophys. Res. Comm. 72, 1251-1258 Ladbrooke, B.D., Williams, R.M. and Chapman, D. (1968) Studies on lecithin-cholesterol-water interactions by differential scanning calorimetry and x-ray diffraction. Biochim. Biophys. Acta. 150, 333-340. Lampe, M.A., Burligame, A.L., Whitney, J., Williams, M.L., Brown, B.E., Roitman, E. and Elias, P.M. (1983) Human stratum corneum lipids: characterization and regional variations. J. Lipid Res. 24, 120-130. Lichtenberg, D., Freire, E., Schmidt, C.F., Barenholz, Y., Feigner, P.L. and Thompson, T.E. (1981) Effect of surface curvature on stability, thermodynamic behaviour and osmotic activity of dipalmitoylphosphatidylcholine single lamellar vesicles. Biochemistry 20, 7700-7707. Lund-Katz, Sissel, Laboda, H.M., McLean, L.R. and Phillips, M.C. (1988) Influence of molecular packing and phospholipid type on rates of cholesterol exchange. Biochemistry 27, 3416-3423. Maurukas, J., and Thomas, R.G., (1959) Treatment of experimental atherosclerosis in the rabbits with L, D. alpha (dimyristoyl) lecithin. J. Lab. Clin. Med. 1, 30-37. Mayer, L.D., Hope, M.J. and Cullis, P.R. (1986) Vesicles of variable sizes produced by a rapid extrusion procedure. Biochim. Biophys. Acta 858, 161-168. Mayer, L.D., Tai, L.C., Ko, D.S.C., Masin, D., Ginsberg, R.S., Cullis, P.R. and Bally, M.B. (1989) Influence of vesicle size, lipid composition and drug-to-lipid ratio on the biological activity of liposomal doxorubicin in mice. Cancer Res. 49, 5922-5930 McLean, L.R. and Phillips, M.C. (1982) Cholesterol desorption from clusters of phosphatidylcholine and cholesterol in unilamellar vesicle bilayers during lipid transfer or exchange. Biochemistry 21, 4053-4059. 147 McLean, L.R. and Phillips, M.C. (1981) Mechanism of cholesterol and phosphatidylcholine exchange or transfer between unilamellar vesicles. Biochemistry 20, 2893-2900. McLean, L.R. and Phillips, M.C. (1984) Cholesterol transfer from small and large unilamellar vesicles. Biochim. Biophys. Acta 116, 21-26. Meddings, J.B., Spady, D.K. and Dietschy, J.M. (1987) Kinetic characteristics and mechanisms of regulation of receptor-dependent and receptor-independent LDL transport in the liver of different animal species and humans. Am. Heart J. 113, 475481. Mendez, A.J., He, J.L., Huang, H.S., Wen, S.R. and Hsia, S.L. (1988) Interaction of rabbit lipoproteins with liposomes of egg yold phospholipids. Lipids 23, 961-967. Mezei, M. (1988) Liposomes in the topical application of drugs: a review. In Liposomes as Drug Carriers Ed. G. Gregoriadis pp. 663-677 Miller, N.E., La Ville, A. and Crook, D. (1985) Direct evidence that reverse cholesterol transport is mediated by high-density lipoprotein in rabbit. Nature 314, 109-110. Mimms, L.T., Zampighi, G., Nozaki, Y., Tanford. C. and Reynolds, J.A. (1981) Phospholipid vesicle formation and transmembrane protein incorporation using octyl glucoside. Biochemistry 20, 833-840. Mindham, M.A. and Mayes, P.A. (1991) Reverse cholesterol transport in the rat. Biochem. J. 219, 503-508 Moghimi, S.M. and Patel, H.M. (1988a) Serum opsonins and phagocytosis of saturated and unsaturated phospholipid liposomes. Biochim. Biophys. Acta 984, 384-387. Moghimi, S.M. and Patel, H.M. (1988b) Tissue specific opsonins for phagocytic cells and their different affinity for cholesterol-rich liposomes. FEBS Lett. 233, 43-147. Moghimi, S.M. and Patel, H.M. (1988c) Differential properties of organ-specific serum opsonins for liver and spleen macrophages. Biochim. Biophys. Acta 984, 379-384. Moncada, S., Martin, J.F. and Higgs, A. (1993) Symposium on regression of Atherosclerosis. Eur. J. Clin. Invest. 23, 385-398. Nakagawa, Y., Inoue, K. and Nojima, S. (1979) Transfer of cholesterol between liposomal membranes. Biochim. Biophys. Acta 553, 307-319. Nayar, R., Hope, M.J. and Cullis, P.R. (1989) Generation of large unilamellar vesicles from long chain saturated phosphatidylcholine by extrusion techniques. Biochim. Biophys. Acta, 9S6, 200-206. 148 Neraecz, G. and Schroeder, F. (1988) Time-resolved fluorescence investigation of membrane cholesterol heterogeniety and exchange. Biochemistry 27, 7740-7749. Nemecz, G., Fontaine, R.N, and Schroeder, F. (1988) A fluorescence and radiolabel study of sterol exchange between membranes. Biochim. Biophys. Acta 943, 511-521. Nishikawa, K. Arai, H. and Inoue, K. (1990) Scavenger receptor-mediated uptake and metabolism of lipid vesicles containing acidic phospholipids by mouse peritoneal macrophages. J. Biol. Chem. 265, 5226-5231. Patel, H.M. (1992) Serum opsonins and liposomes: their interaction and opsonophagocytosis. Crit. Rev. Ther. Drug Carrier Systems 9, 39-90. Phillips, M.C. (1992) Cholesterol-phospholipid interactions and the exchangeability of cholesterol in membranes, in Structural and dynamic properties of lipids and membranes (ed. R.J. Cherry and P.J. Quinn), Portland Press, London, pp. 103-118. Phillips, M.C, Johnson, W.J. and Rothblat, G.H. (1987) Mechanisms and consequences of cellular cholesterol exchange and transfer. Biochim. Biophys. Acta 906, 223-276. Phillips, M.C and Rothblat, G.H. (1985) Cellular cholesterol efflux and influx. In Lipoprotein and cholesterol metabolism in steroidogenic tissues. Ed. J.F. Strauss and K.M.J. Menon. George Stickey Company, Philadelphia, pp. 9-20. Poste, G., Kirsh, R. and Kustler, T. (1984) The challenge of liposome targeting in vivo. In Liposome Technology, Vol HI. Gregoriadis, G., ed. CRC Press Inc., Boca Raton, Fla, pp. 1-28. Poste, G. Bucana, C, Raz, A., Bugelski, P., Kirsh, R. and Fidler, I.J. (1982) Analysis of the fate of sytemically administered liposomes and implications for their use in drug delivery. Cancer Res. 42, 1412-1422. Poznansky, M.J. and Lange, Y. (1978) Transbilayer movement of cholesterol in phospholipid vesicles under equilibrium and non-equilibrium conditions. Biochim. Biophys. Acta 506, 256-264. Poznansky, M.J. and Lange, Y. (1976) Tansbilayer movement of cholesterol in dipalmitoyllecithin-cholesterol vesicles. Nature 259, 420-421. Prichard, R.W. (1974) Regression of atherosclerosis-A perspective. Exp. Mol. Pathol. 20, 407-411. Prior, J.T. and Ziegler, D.D. (1965) Regression of experimental atherosclerosis. Arch. Pathol. 80, 50-57. 149 Rahman, Y.E., Cerny, E.A., Patel, K.R., Lau, E.H. and Wright, B.J. (1982) Differential uptake of liposomes varying in size and lipid composition by parenchymal and Kupffer cells of mouse liver. Life ScL, 31, 2061-2071. Reynolds, G.A., (1989) Rational therapy of Familial Hypercholesterolemia. Circulation 79, 1146-1148 Robenek, H., Jung, W., Gebhardt, R. (1982) The topography of filipin-cholesterol complexes in the plasma membrane of cultured hepatocytes and their relation to cell junction formation. J. Ultrastruct. Res. 78, 95-106. Roerdink, F., Dijkstra, J., Hartman, G., Bolscher, B. and Scherphof, G. (1981) The involvement of parenchymal and kupffer and endothelial liver cells in the hepatic uptake of intravenously injected liposomes; effects of lanthanum and gadolinium salts. Biochim. Biophys. Acta 677, 79-89. Roerdink, F. Regts, J., Van Leewen, B. and Scherphof, G. (1984) Intrahepatic uptake and processing of intravenously injected small unilamellar phospholipid vesicles in rats. Biochim. Biophys. Acta 770, 195-202. Rose, G.H. and Oklander M. (1965) J. Lipid Res. 6, 428-431 Rosenfeld, M.E., Chait, A, Bierman, E.L., King, W., Goodwin, P., Walden, C.E. and Ross, R. (1988) Lipid composition of aorta of Watanabe Heritable Hyperlipemic and comparably hypercholesterolemic fat-fed rabbits: Plasma lipid composition determines aortic lipid composition of hypercholesterolemic rabbits. Arteriosclerosis 8, 338-347. Rothblat, G.H. and Phillips, M.C. (1991) Cholesterol efflux from arterial wall cells. Current Opinion in Lipidology 2, 288-294. Rothblat, G.H., Mahlberg, F.H., Johnson, W.J. and Phillips, M.C. (1992) Apolipoproteins, membrane cholesterol domains and the regulation of cholesterol efflux. J. Lip. Res. 33, 1091-1097. Rudel, L. and Morris, M.D. (1973) / . Lipid Res. 14, 364-366 Scherphof, G. and Morselt, H. (1984) On the size-dependent disintegration of small unilamellar phosphatidylcholine vesicles in rat plasma. Biochem. J. 221, 423-429. Scherphof, G.L., J. Damen and J. Wilschut. (1984) Interactions of liposomes with plasma proteins. In Liposome Technology, Vol III. Gregoriadis, G., ed. CRC Press Inc., Boca Raton, Fla, pp. 205-224. Scherphof, G., Roerdink, F., Waite, M. and Parks, J. (1978) Disintegration of phosphatidylcholine liposomes in plasma as a result of interaction with high-density 150 lipoproteins. Biochim. Biophys. Acta 542, 296-307 Scherphof, G., Damen, J., Dijkstra, J., Roerdink, F. and Spanjer, H. (1984) In: Receptor Mediated Targeting of Drugs. Ed. Gregoriadis, G., Poste, G., Senior, J. and Trouet, A. Life Sci. 82, 267-295 Scherphof, F. Morselt, H., Regts, J., and Wilschut, J.C. (1979) The involvement of the lipid phase transition in the plasma-induced dissolution of multilamellar phosphatidylcholine vesicles. Biochm. Biophysic. Acta 542, 296-307. Schroeder, F., Jefferson, J.R., Kier, A.B., Knittel, J., Scallen, T.J., Wood, W.G. and Hapala, I. (1991) Membrane cholesterol dynamics: cholesterol domains and kinetic pools. Exp. Biol. Med. 196, 235-252 Schroeder, F and Nemecz, G. (1990) Transmembrane cholesterol distribution, in Advances in cholesterol research. M. Esfahani and J. Swaney Eds. Telford Press, West Caldwell, NJ. pp. 47-87. Schumaker, V.N. and Puppione, D.L. (1986) Sequential flotation ultracentrifugation. Methods in Enzymology 128, 155-181 Schwartz, C.J., Valente, A.J., Sprague, E.A, Kelly, J.L. and Nerem, R.M., (1991) The pathogenesis of Atherosclerosis: An Overview. Clinical Cardiology 14, 1-16 Segrest, J., Jone, M.K., De Loof, H. Brouilette, C.G. Venkatachalapathi, Y.V. and Anantharamaiah, G.M. (1992) The amphipathic helix in the exchangeable apolipoproteins: a review of secondary s t ruc tu re and function. J. LipidRes. 33, 141-166. Senior, J. (1987) Fate and behaviour of liposomes in vivo: a review of controlling factors. Crit. Rev. Ther. Drug Carrier Syst. 3, 123-193. Singer, S.J. and Nicholson, G.L. (1972) The fluid mosaic model of the structure of cell membranes. Science 175, 720-73. Small, D.M. (1988) Progression and regression of atherosclerotic lesions: Insights from lipid physical biochemistry. Arteriosclerosis 8, 103-129 Spady, D.K., Turley, S.D and Dietschy, J.M. (1985a) Rate of low density lipoprotein uptake and cholesterol synthesis are regulated independently in the liver. / . Lipid. Res. 26, 465-472. Spady, D.K., Turley, S.D and Dietschy, J.M. (1985b) Receptor-independent LDL transport in rats in vivo. Quantitation, characterization and metabolic consequences. J. Clin. Invest. 76, 1113-1122. 151 Spanjer, H.H. van Galen, M., Roerdink, F.H. Regts, J. and Scherphof, G.L. (1986) Intrahepatic distribution of small unilamellar liposomes as a function of liposomal lipid composition. Biochim. Biophys. Acta 863, 224-230. Sparks, D.L., Lund-Katz, S. and Phillips, M.C. (1992) The conformation of apolipoprotein A-I in discoidal and spherical recombinant high density lipoprotein particles. J. Biol. Chem. 267, 25830-25835. St. Clair, R.W., Atherosclerotic regression in animal models: current concepts of cellular and biochemical mechanisms (1983) Prog. Cardiovasc. Dis. 26,109-132 Stafford, W.W. and Day, C.E. (1975) Regression of atherosclerosis effected by intravenous phospholipid. Artery 1, 106-114. Stein, O. Oette, K., Haratz, D., Halperin, G. and Stein, Y. (1988) Sphingomyelin liposomes with defined fatty acids: metabolism and effects on reverse cholesterol transport. Biochim. Biophys. Acta 960, 322-333 Stein, Y., and Stein, O. (1979) Interaction between serum lipoproteins and cellular components of the arterial wall. In Biochemistry of Atherosclerosis (Eds. A.M. Scanu, R.W. Wissler and G,S, Getz) Dekker, New York, pp. 313-344. Storch, J. and Kleinfeld, A.M. (1986) Transfer of long-chain fluorescent free fatty acids between unilamellar vesicles. Biochemistry 25, 1717-1726. Szoka, F. and Papahadjopoulos, D. (1978) Procedure for preparation of liposomes with large internal aqueous space and high capture by reverse-phase evaporation. Proc. Natl. Acad. Sci. USA 75, 4194-4198. Szoka, F. and Papahadjopoulos, D. (1980) Comparative properties and methods of preparation of lipid vesicles (liposomes). Ann. Rev. Biophys. Bioeng. 9, 467-508. Tall, A.R. and Green, P.H.R. (1981) Incorporation of phosphatidylcholine into spherical and discoidal lipoproteins during incubation of egg phosphatidyline vesicles with isolated high density lipoproteins or with plasma. / . Biol. Chem. 256, 20355-20440. Tall, A.R., Tabas, I. and Williams, K.J. (1986) Lipoprotein-liposome interactions. Methods Enzymol. 128, 647-657. Tall, A.R. (1980) Studies on the transfer of phosphatidylcholine from unilamellar vesicles into plasma high density lipoproteins in the rat. / . Lipid Res. 21, 361-367. Tall, A.R. and Small, D.M. (1977) Solubilisation of phospholipid membranes by human plaasma high density lipoproteins. Nature 265, 163-164. 152 Tashiro, T., Mashima, Y., Yamamori, H., Horibe, K., Nishizawa, M., Sanada, M. and Okui, K. (1991) Increased lipoprotein-X causes hyperlipidemia during intravenous administration of 10% fat emulsion in man. J. Parenteral Enteral Nut. 15, 546-550. Thomas, D.P. and Poznansky, M.J. (1988) Effect of suface curvature on the rate of cholesterol transfer between lipid vesicles. Biochem. J. 254, 155-160. Wertz, P. (1988) Epidermal lipids. Seminars in Dermatol. 11, 106-118. Williams, K.J. and Tall, A.R. (1988) Interactions of liposomes with lipoproteins: relevance to drug delivery systems and to the treatment of atherosclerosis. In Liposomes as Drug Carriers Ed. G. Gregoriadis pp.93-111. Williams, K.J., Vallabhajosula, S., Rahman, I.U., Donnelly, T.M., Parker, T.S., Weinrauch, M. and Goldsmith, S.J. (1988) Low density lipoprotein receptor-independent hepatic uptake of a synthetic, cholesterol-scavenging lipoprotein: Implications for the treatment of receptor-deficient atherosclerosis. Proc. Natl Acad. Sci. (USA) 85, 242-246 Williams, K.J. and Scanu, A.M. (1986) Uptake of endogenous cholesterol by a synthetic lipoprotein. Biochim. Biophys. Acta, 875, 183-194. Williams, K.J., Tall, A.R., Bisgaier, C. and Brocia, R. (1987) Phospholipid liposomes acquire apolipoprotein E in atherogenic plasma and block cholesterol loading of cultured macrophages. / . Clin. Invest. 19, 1466-1472. Williams, K.J., Werth, V.P. and Wolff, J.A. (1984) Intravenously administered lecithin liposomes: a synthetic antiatherogenic lipid particle. Perspectives Biol. Med. 27(3), 417-431. Williams, K.J., and Tall, A.R., Interactions of liposomes with lipoproteins: relevance to drug delivery systems and to the treatment of atherosclerosis (1988) Liposomes as Drug Carriers Edited by G. Gregoriadis 7, 93-111 Williams, K.J., Tall, A.R., Tabas, I., and Blum, C. (1986) Recognition of vesicular lipoproteins by the apolipoprotein B,E receptor of cultured fibroblasts. J. Lipid Res. 27, 892-900. Williams, M.L. (1992) Epidermal lipids and scaling diseases of the skin. Seminars in Dermatol. Wong, M., Anthony, F.H., Tillack, T.W. and Thompson, T.E. (1982) Fusion of dipalmitoylphosphatidylcholine vesicles at 4°C. Biochemistry 21, 4126-4132. Yau-Young, A.O., Rothblat, G.H., and Small, D.M. (1982) Mobilization of cholesterol from cholesterol ester-enriched tissue culture cells by phospholipid dispersions. Biochim. Biophys. Acta 710, 181-187. 153 Yeagle, P.L. and Young, J.E. (1986) Factors contributing to the distribution of cholesterol among phospholipid vesicles. / . Biol. Chem. 261, 8175-8181 Yeagle, P.L. (1985) Cholesterol and the cell membrane. Biochim. Biophys. Acta 822, 267-287. Yeagle, P.L. (1981) Cholesterol rotation in phospholipid vesicles as observed by 13C-NMR. Biochim. Biophys. Acta 640, 263-273. 154 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items