- Email: [email protected]
Intralipid 10%: physicochemical characterization
BASIC NUTRITIONAL INVESTIGATION
Intralipid 10%: Physicochemical Characterization Jacqueline Fe´re´zou, PhD, Annie Gulik, PhD, Nicole Domingo, PhD, Fabien Milliat, Jean-Claude Dedieu, Suzanne Dunel-Erb, PhD, Claudine Chevalier, and Andre C. Bach, PhD From the Laboratoire de Physiologie de la Nutrition, Universite´ Paris-Sud, Orsay, France; the Centre de Ge´ne´tique Mole´culaire, Gif-sur-Yvette, France; the Nutrition Humaine et Lipides, Marseille, France; and the Centre d’Ecologie et Physiologie Energe´tiques, Strasbourg, France OBJECTIVES: Parenteral fat emulsions contain two populations of particles: artificial chylomicrons rich in triacylglycerols (TAG), and liposomes (bilayer of phospholipids [PL] enveloping an aqueous phase). Centrifugation permits isolating the liposomes in the infranatant called mesophase. The aim of the present work was to better characterize this mesophase chemically and to view the particles it contains by electron microscopy. METHODS: Electron microscopy (Philips 410) was performed after cryofracture on native 10% Intralipid, mesophase (centrifugation for 1 h at 27 000 g), and a liposome-enriched fraction (ring of density 1.010 –1.030 g/l obtained after centrifuging mesophase in a KBr density gradient at 100 000 g for 24 h). The TAG and protein content of the mesophase was analyzed and the proteins partially characterized by immunodetection (Western-blot). RESULTS: This electron microscope study of 10% Intralipid gives evidence for the coexistence of artificial chylomicrons (mean diameter, 260 nm) and liposomes (43 nm), the latter being smaller than expected and containing 8% w/w TAG after purification. The solubilization of TAG in PL bilayers (reported to be ⱕ3.1% w/w) might have been increased in parenteral emulsions by the manufacturing process or/and the high TAG/PL ratio. Minute amounts of proteins have also been detected and partially characterized using a specific antibody raised against the human 7 kDa Anionic Polypeptide Factor (APF), known to strongly interact with PL in bile. CONCLUSIONS: This work has shown that the size (mean diameter, 43 nm) of the liposomes present in 10% Intralipid is smaller than that usually assumed. Traces of hydrophobic proteins in the emulsion may account for certain allergic reactions sometimes observed in infused patients. Nutrition 2001;17: 930 –933. ©Elsevier Science Inc. 2001 KEY WORDS: parenteral fat emulsion, artificial chylomicrons, liposomes, triacylglycerols, phospholipids, electron microscopy, proteins
INTRODUCTION Owing to their insolubility in water, plant oils (or triacylglycerols, TAGs) have to be infused in the form of an oil-in-water emulsion, i.e., a fine suspension of oil particles stabilized in the aqueous phase by an heterogeneous mixture of phospholipids (PLs) as an active surface agent.1,2 For a long time PL-stabilized TAG emulsions were thought to consist of a single kind of TAG-rich particles analogous to chylomicrons,3 which for that reason are called artificial chylomicrons. Since 1980, however, commercially available parenteral emulsions are known to contain PLs in excess of the amount strictly necessary to stabilize the artificial chylomicrons.4,5 On the basis of physicochemical arguments, it has been proposed that this PL excess exists in the form of PL-rich vesicles or liposomes.6 –9 Their dispersion in the continuous aqueous phase of the emulsion constitutes a “ mesophase ” that can be separated by centrifugation of the emulsion.10,11 Because photographs of liposomes from commercially available fat emulsions are rarely found in the literature,7,12 we have performed freeze-fracture electron microscopy (FFEM) on native 10% Intralipid, a mesophase, and a mesophase subfraction enriched with liposomes.9 This emulsion, which has been used extensively and has been studied for
Correspondence to: Andre´ Bach, PhD, Centre d’Ecologie et Physiologie Energe´tiques, 67087 Strasbourg Cedex 2, France. E-mail: [email protected] Nutrition 17:930 –933, 2001 ©Elsevier Science Inc., 2001. Printed in the United States. All rights reserved.
more than 30 y, is recognized for its high stability and reproducible physicochemical characteristics. Nutritionists now recommend infusing the 20% formula (higher TAG:PL ratio) because liposomes decrease the clearance rate of the infused fat.13 However, we used the 10% formula, which is richer in liposomal PLs11 and poorer in TAGs, thus needing no further dilution. We investigated whether liposomes purified by ultracentrifugation have the same ultrastructure as those in native emulsion. The comparison of our results with those obtained previously on the same emulsion by other techniques7,9 prompted us to reevaluate liposome size and composition. We also report some preliminary data that draw attention to trace amounts of proteins in the mesophase of parenteral emulsions.
MATERIALS AND METHODS The starting material for our study was a commercially available emulsion (10% Intralipid, Kabi-Pharmacia, St-Quentin en Yvelines, France) composed of soybean oil and egg PL (PL:TAG ratio of 12:100 [w/w]). The procedures used for emulsion fractionation and the assays have been described in detail.9 In a first step, a 12-mL sample of native emulsion was centrifuged at 15°C (L8-70, Beckman Inc., Palo Alto, CA, USA) in a heat-sealed polyallomer tube (Beckman) at 20 000 rpm (27 000g) for 1 h using an angular rotor. The tube was cut with a tube-slicer (Beckman) just under the 0899-9007/01/$20.00 PII S0899-9007(01)00667-0
Nutrition Volume 17, Number 11/12, 2001
Physicochemical Characterization of Parenteral Fat Emulsions
931
RESULTS AND DISCUSSION Particle Size
FIG. 1. Result of the two-step fractionation of the emulsion. Left: Upper lipid cake constituted of compacted artificial chylomicrons and an infranatant named mesophase after the first step of centrifugation (see text). Right: Upper residual artificial chylomicrons and opalescent ring of liposomes (named the phospholipid-rich fraction) in the middle of the tube after the second step of centrifugation.
limit of the lipid cake formed by the compacted artificial chylomicrons at the top of the tube (Fig. 1, left). A 2-mL mesophase sample of the homogeneous infranatant, or mesophase, was then centrifuged at 15°C in a KBr density gradient for 24 h at 40 000 rpm (100 000g) in a swinging bucket. A white layer at the surface showed the presence of a minor population of residual artificial chylomicrons (density ⬍ 1.008 g/mL). The opalescent ring in the middle of the tube (density range ⫽ 1.010 –1.030 g/mL: Fig. 1, right) corresponds to a PL-rich fraction (PLRF) of the mesophase, which was harvested as a population of purified liposomes.9 Small amounts each of PLRF, mesophase, and native emulsion sample containing 30% glycerol as a cryoprotectant were deposited onto a copper planchet and rapidly frozen in liquid propane. It was subsequently fractured and shadowed in a Balzers freeze etch unit using platinum carbon shadowing. The replicas were examined with a Philips 410 electron microscope. A specific antibody raised against the anionic polypeptide factor (APF), isolated by high-performancce liquid chromatography from human bile, was used to characterize proteins in the mesophase by an enzyme-linked immunosorbent assay, as previously described.14
The only published pictures of liposomes in parenteral fat emulsions we know of concern diluted Intralipid (with a possible effect of dilution on the particle lipid organization) or a noncommercially available fat emulsion. Rotenberg et al.7 published a picture of the mesophase isolated from a five-fold diluted 10% Intralipid that showed liposomes coexisting with residual artificial chylomicrons when viewed by cryoelectron microscopy. Later, Westesen and Wehler8,12 published a micrograph obtained by FFEM from a diluted (1/1 [v/v]) model emulsion (soybean oil and fractionated egg yolk PL; PL:TAG ratio of 18:250 [w/w]) and concluded that most of the particles were smaller than 100 nm in diameter. In neither study was the size of the liposomes accurately assessed. The native emulsion (Fig. 2A) showed a population of large artificial chylomicrons (mean diameter ⫽ 200 ⫾ 10 nm, n ⫽ 59) with the typical cross fracture showing a particle filled with material8 that accounted for about 30% of the total particles. They coexisted with smaller particles (mean diameter ⫽ 33.8 ⫾ 2.1 nm, n ⫽ 140), most of which presented the typical concave or convex fracture surface of lipidic vesicles. We did not detect stacked-layer structures dispersed between artificial chylomicrons, as reported in several studies6,15,16 but likely corresponding to artifacts, especially after negative staining. Our results are consistent with the qualitative observations of Westesen and Wehler8,12 on their model emulsion. In the mesophase (Fig. 2B), the particles had a mean diameter of 39 ⫾ 2 nm (n ⫽ 209). About 15% of them showed a cross fracture, indicating the presence of small artificial chylomicrons. After further centrifugation, the resultant PLRF (Fig. 2C) was not very different from the corresponding mesophase with regard to particle shape. However, the population of particles was more homogeneous: only 1% to 2% of the particles corresponded to residual artificial chylomicrons (mean diameter ⫽ 45 nm). All the others were liposomes with a mean diameter of 32.5 ⫾ 0.8 nm (n ⫽ 229) and thus smaller (P ⬍ 0.0001, Student’s t test) than the mesophase particles (mixture of liposomes and residual artificial chylomicrons) and close to the size of liposomes observed in the native emulsion. Because the fracture does not necessarily propagate through the largest possible cross section of the spherical particles, one has to correct for the underestimated size obtained by FFEM by a factor of about 30%.17 In this way, 10% Intralipid contains artificial chylomicrons with a mean diameter of about 200 ⫻ 1.3, i.e., 260 nm, and liposomes of about 33 ⫻ 1.3, i.e., 43 nm. Regarding the size of the artificial chylomicrons, there was good agreement despite the type of technique used. Indeed, the mean diameter of 10% Intralipid artificial chylomicrons was 250 to 280 nm as estimated by light scattering spectroscopic techniques5,7,9,16,18,19 and electron microscopy.5,7,20 –22 Conversely, the liposome mean diameter of 40 nm determined by FFEM was discrepant with the 70 to 100 nm measured by quasi-elastic light scattering.7,9 This latter value is, however, probably overestimated, as light scattering techniques23 performed on heterogeneously dispersed particles have been known to emphasize the larger ones7 or overlook particles smaller than 100 nm.8,12 Consequently, the new value found by FFEM seems more reliable. With the same FFEM technique, 40-nm liposomes were found in another commercial emulsion of analogous composition (B. Braun, Boulogne, France; results not shown). This size seems more logical because it allows a better understanding of the intravascular metabolism of infused liposomes. It is difficult to understand how 70- to 100-nm particles could be converted into smaller, 40- to 60-nm-diameter24 lipoprotein-X particles after cholesterol loading and acquisition of apolipoproteins.13 Moreover, the reevaluated liposome size improves the interpretation of the 31P-nuclear magnetic resonance
932
Fe´ re´ zou et al.
Nutrition Volume 17, Number 11/12, 2001 43 nm mean diameter, the out:in PL ratio thus corresponds to 1.50, which is closer to our experimental value. The remaining slight discrepancy can be explained by small residual artificial chylomicrons in the mesophase (Fig. 2B). Their presence is easily revealed by the significant amounts of TAGs found by thin layer chromatographic analysis of the mesophase, whatever the type of emulsion considered. The amount of TAGs has been quantified not only in Intralipid mesophase (30% to 40% of total lipids [w/w], regardless of the formula) but also, although less, in PLRF (8%, 15%, and 30% [w/w] for 10%, 20%, and 30% Intralipid, respectively).9 Such amounts of TAGs in purified liposomes are incompatible with regard to the maximal solubilization capacity of TAGs in PL bilayers, considered to equal 3.1% (w/w; 2.8%, mol/mol).25 However, this value was obtained with liposomes differing strongly from the liposomes of parenteral fat emulsions. The differences were in the chemical composition (triolein versus soybean oil; purified phosphatidylcholine versus fractionated mixture of PLs from egg yolk), the PL:TAG (w/w) ratio (94:0 to 6 in the presence of 6-0 cholesteryl oleate versus 12:100), and the manufacturing technique (sonication versus homogenization at high pressure and sterilization around 120°C). We believe that these factors increase the maximal solubilization capacity of TAGs in PL bilayers, thus explaining the abundance of TAG in liposomes of parenteral fat emulsions. Proteins
FIG. 2. Electron micrographs of the native emulsion (A), the mesophase (B), and the phospholipid-rich fraction (C). Arrow indicates a cross fracture of a small artificial chylomicron.
spectra that showed7,9 the coexistence in native emulsions of artificial chylomicrons (external PLs only) and liposomes (internal and external PLs) and the predominance of unilamellar liposomes in the mesophase. To calculate the distribution of PLs between the two populations of particles, the value of 1.1 was used for the out:in PL ratio of liposomes, assuming a mean diameter of 70 to 100 nm.7 However, the ratio we measured on the mesophase of 10% Intralipid was significantly higher (1.65).9 For liposomes of
Another parameter to consider is the composition of liposomes. Although it is assumed that commercial emulsions do not contain any protein, a trichloroacetic acid precipitation indicates the presence of proteins in their mesophase. Using bovine serum albumin as a standard, we measured26 87 g of protein per milliliter of mesophase of 10% Intralipid (a direct protein measurement in the native emulsion is impeded by the lactescence due to the very high TAG content), for a protein:PL ratio of about 1.47% (w/w; i.e., 0.09:6.11). Where do these proteins come from? By analogy with the phytosteryl esters found in commercial emulsions,9,27 one possibility is that they accompany the soybean oil as oleosins. This family of hydrophobic proteins stabilizes the stored lipids in all seeds, interacts with PLs, and has been found in commercial alimentary oils.28 But the fact that the mesophase protein:PL ratio seems to be little influenced by the formula (1.57% [w/w]; i.e., 0.055:3.49 for the 20% Intralipid) suggests that the proteins of the emulsion are delivered also by the egg yolk PLs (amounting 12 g in both formulas). Interestingly, a PL-binding hydrophobic protein, the APF, from human and animal bile,14,29 showed a strong interaction with amphipathic PLs. The same strong association was observed between egg yolk lecithin (Sigma, Saint Quentin Fallavier, France) and human APF (APF:PL ratio of 5.01% [w/w]) during the preparation of proteoliposomes containing APF.30 In this study, APF promoted the excretion of cholesterol and PLs into the bile. This protein, which is also a minor apolipoprotein of high-density lipoproteins,14 acts as an antinucleating factor in bile via its vesicle-stabilizing properties.31,32 Thus, with the use of the specific antibody (MD94) raised against purified human APF, an APF signal was detected by enzyme-linked immunosorbent assay32 for the mesophase of Intralipid, corresponding to 2.32 g/mL for the 10% formula. The expected molecular weight of APF (7 kDa) was confirmed by Western-blotting (Fig. 3). The significance of APF in commercial emulsions is still unknown, but it is noteworthy that similar or higher values were obtained by enzyme-linked immunosorbent assay on mesophase from other lipid emulsions (APF g/mL mesophase: Intralipid 20%: 3.01; Medialipide from B. Braun: 10%, 5.56; 20%, 0.88, Endolipide from B. Braun: 10%, 4.72; 20%, 5.35), whereas assays with soybean PL-stabilized parenteral emulsions (Lipofundin S, Braun, Melsungen, Germany) did not show any immunoreactive proteins when using the same APF antibody. Therefore, egg yolk
Nutrition Volume 17, Number 11/12, 2001
FIG. 3. Western blot of Intralipid mesophase incubated in the presence of an antibody against anionic polypeptide factor (MD94).
PLs probably deliver not only residual amounts of cholesterol9,13 but also trace amounts of APF or related peptidic components. It can be assumed that the type of treatment and the level of purification of egg yolk lecithin influence the amount of proteins in the emulsifier. The presence of hydrophobic proteins (provided by oil and lecithin) in fat emulsions, which influences the phase diagram, could determine the lipids organization in the different particle populations and also might account for certain allergic reactions observed in infused patients.33 The use of FFEM allowed the size of 10% Intralipid liposomes to be reevaluated (40 nm versus the classic 70 to 100 nm in diameter) in the native emulsion and its mesophase. Our results showed that the liposome structure is not altered by ultracentrifugation. In the mesophase, which is composed mainly of liposomal PLs, we detected minute amounts of proteins. They were partly characterized as small hydrophobic proteins delivered by egg yolk lecithin used as the emulsifier.
ACKNOWLEDGMENT Arthur Pape is gratefully acknowledged for his help in editing the manuscript.
REFERENCES 1. Schuberth O, Wretlind A. Intravenous infusion of fat emulsion, phosphatides and emulsifying agents. Clinical and experimental studies. Arch Chir Scand 1961; 278(suppl):1 2. Rydhag L. The importance of the phase behaviour of phospholipids for emulsion stability. Fette Seifen Anstrichm 1979;4:168 3. Schoefl GI. The ultrastructure of chylomicra and of the particles in an artificial fat emulsion. Proc R Soc B 1968;169:147 4. Carlson LA. Studies on the fat emulsion Intralipid. I. Association of serum proteins to Intralipid triglyceride particles (ITP). Scand J Clin Lab Invest 1980; 40:139
Physicochemical Characterization of Parenteral Fat Emulsions
933
5. Erkelens DW, Chen C, Mitchell CD, Glomset JA. Studies of the interaction between apolipoproteins A and C and triacylglycerol-rich particles. Biochim Biophys Acta 1981;665:221 6. Groves MJ, Wineberg M, Brain APR. The presence of liposomal material in phosphatide stabilized emulsions. J Disper Sci Tech 1985;6:237 7. Rotenberg M, Rubin M, Bor A, et al. Physico-chemical characterization of Intralipid emulsions. Biochim Biophys Acta 1991;1086:265 8. Westesen K, Wehler T. Investigation of the particle size distribution of a model intravenous emulsion. J Pharm Sci 1993;82:1237 9. Fe´ re´ zou J, Lai NT, Leray C, et al. Lipid composition and structure of commercial parenteral emulsions. Biochim Biophys Acta 1994;1213:149 10. Jakoi L, Quarfordt SH. The induction of hepatic cholesterol synthesis in the rat by lecithin mesophase infusions. J Biol Chem 1974;249:5840 11. Lutz O, Meraı¨hi Z, Fe´ re´ zou J, et al. The mesophase of parenteral fat emulsion is both substrate and inhibitor of lipoprotein lipase and hepatic lipase. Metabolism 1990;39:1225 12. Westesen K, Wehler T. Physicochemical characterization of a model intravenous oil-in-water emulsion. J Pharm Sci 1992;81:777 13. Bach AC, Fe´ re´ zou J, Frey A. Phospholipid-rich particles in commercial parenteral fat emulsions. An overview. Prog Lipid Res 1996;35:133 14. Domingo N, Grosclaude J, Bekaert ED, et al. Epitope mapping of the human biliary amphipathic, anionic polypeptide: similarity with a calcium-binding protein isolated from gallstones and bile, and immunologic cross-reactivity with apolipoprotein A-I. J Lipid Res 1992;33:1419 15. Erkelens DW, Mocking JAJ. The effect of apolipoprotein A on in vitro apolipoprotein C binding and in vivo removal of artificial triacylglycerol-rich particles. Metabolism 1985;34:222 16. Li JM, Caldwell KD. Structural studies of commercial fat emulsions used in parenteral nutrition. J Pharm Sci 1994;83:1586 17. Heegaard CW, Le Maire M, Gulik-Krzywicki T, Moller JV. Monomeric state and Ca2⫹ transport by sarcoplasmic reticulum Ca2⫹-ATPase, reconstituted with an excess of phospholpid. J Biol Chem 1990;265:12020 18. Lutz O, Meraı¨hi Z, Mura J-L, et al. Fat emulsion particle size: Influence on the clearance rate and the tissue lipolytic activity. Am J Clin Nutr 1989;50:1370 19. Davis SS. The stability of fat emulsions for intravenous administration. In: Johnston IDA, ed. Advances in clinical nutrition. Lancaster: MTP Press, 1983: 213 20. Hamilton-Attwell VL, Du Plessis J, Van Wyk CJ. A new scanning electron microscope (SEM) method for the determination of particle size in parenteral fat emulsions. J Microsc 1987;145:347 21. Du Plessis J, Tiedt LR, Van Wyk CJ, Ackermann C. A new transmission electron microscope method for the determination of particle size in parenteral fat emulsions. Int J Pharm 1986;34:173 22. Connelly PW, Kuksis A. Effect of core composition and particle size of lipid emulsions on apolipoprotein transfer of plasma lipoproteins in vivo. Biochim Biophys Acta 1981;666:80 23. Marangoni AG, Rousseau D. The influence of chemical interesterification on physicochemical properties of complex fat systems—1. Melting and crystallization. J Am Oil Chem Soc 1998;75:1265 24. Miyahara T, Fujiwara H, Yae Y, et al. Abnormal lipoprotein appearing in plasma of patients who received a ten percent soybean oil emulsion infusion. Surgery 1979;85:566 25. Hamilton JA, Miller KW, Small DM. Solubilization of triolein and cholesteryl oleate in egg phosphatidylcholine vesicles. J Biol Chem 1983;258:12821 26. Lowry OH, Rosebrough AL, Farr RJ. Protein measurement with folin phenol reagent. J Biol Chem 1951;192:265 27. Clayton PT, Bowron A, Mills KA, et al. Phytosterolemia in children with parenteral nutrition–associated cholestatic liver disease. Gastroenterology 1993; 105:1806 28. Olszewski A, Pons L, Moute´ te´ F, et al. Isolation and characterization of proteic allergens in refined peanut oil. Clin Exp Allergy 1998;28:850 29. Lafont H, Domingo N, Groen AK, et al. AFP/CBP, the small amphipathic anionic protein(s) in bile and gallstones, consist of lipid-binding and calcium-bonding forms. Hepatology 1997;34:1054 30. Martigne M, Domingo N, Chanussot F, et al. Effect of bile anionic polypeptidic fraction on the fate of cholesterol carried by liposomes in the rat. Proc Soc Exp Biol Med 1988;187:234 31. Halpern Z, Lafont H, Arad J, et al. The distribution of biliary anionic fraction (APF) between cholesterol carriers in bile and its effects on nucleation. Hepatology 1994;21:979 32. Domingo N, Groen AK, Leche`ne P, et al. L’APF/CBP, une nouvelle prote´ ine capable a` la fois de lier le calcium et d’induire la fusion de ve´ sicules phosphatidylcholine/choleste´ rol dans la bile. Med Chir Digest 1993;22:117 33. Aimone-Gastin I, Gueant JL, Laxenaire MC, Moneret-Vautrin DA. Pathogenesis of allergic reactions to anaesthetic drugs. Int J Immunopathol Pharmacol 1997; 10:193
(For an additional perspective, see Editorial Opinions)
Intralipid 10%: Physicochemical Characterization Jacqueline Fe´re´zou, PhD, Annie Gulik, PhD, Nicole Domingo, PhD, Fabien Milliat, Jean-Claude Dedieu, Suzanne Dunel-Erb, PhD, Claudine Chevalier, and Andre C. Bach, PhD From the Laboratoire de Physiologie de la Nutrition, Universite´ Paris-Sud, Orsay, France; the Centre de Ge´ne´tique Mole´culaire, Gif-sur-Yvette, France; the Nutrition Humaine et Lipides, Marseille, France; and the Centre d’Ecologie et Physiologie Energe´tiques, Strasbourg, France OBJECTIVES: Parenteral fat emulsions contain two populations of particles: artificial chylomicrons rich in triacylglycerols (TAG), and liposomes (bilayer of phospholipids [PL] enveloping an aqueous phase). Centrifugation permits isolating the liposomes in the infranatant called mesophase. The aim of the present work was to better characterize this mesophase chemically and to view the particles it contains by electron microscopy. METHODS: Electron microscopy (Philips 410) was performed after cryofracture on native 10% Intralipid, mesophase (centrifugation for 1 h at 27 000 g), and a liposome-enriched fraction (ring of density 1.010 –1.030 g/l obtained after centrifuging mesophase in a KBr density gradient at 100 000 g for 24 h). The TAG and protein content of the mesophase was analyzed and the proteins partially characterized by immunodetection (Western-blot). RESULTS: This electron microscope study of 10% Intralipid gives evidence for the coexistence of artificial chylomicrons (mean diameter, 260 nm) and liposomes (43 nm), the latter being smaller than expected and containing 8% w/w TAG after purification. The solubilization of TAG in PL bilayers (reported to be ⱕ3.1% w/w) might have been increased in parenteral emulsions by the manufacturing process or/and the high TAG/PL ratio. Minute amounts of proteins have also been detected and partially characterized using a specific antibody raised against the human 7 kDa Anionic Polypeptide Factor (APF), known to strongly interact with PL in bile. CONCLUSIONS: This work has shown that the size (mean diameter, 43 nm) of the liposomes present in 10% Intralipid is smaller than that usually assumed. Traces of hydrophobic proteins in the emulsion may account for certain allergic reactions sometimes observed in infused patients. Nutrition 2001;17: 930 –933. ©Elsevier Science Inc. 2001 KEY WORDS: parenteral fat emulsion, artificial chylomicrons, liposomes, triacylglycerols, phospholipids, electron microscopy, proteins
INTRODUCTION Owing to their insolubility in water, plant oils (or triacylglycerols, TAGs) have to be infused in the form of an oil-in-water emulsion, i.e., a fine suspension of oil particles stabilized in the aqueous phase by an heterogeneous mixture of phospholipids (PLs) as an active surface agent.1,2 For a long time PL-stabilized TAG emulsions were thought to consist of a single kind of TAG-rich particles analogous to chylomicrons,3 which for that reason are called artificial chylomicrons. Since 1980, however, commercially available parenteral emulsions are known to contain PLs in excess of the amount strictly necessary to stabilize the artificial chylomicrons.4,5 On the basis of physicochemical arguments, it has been proposed that this PL excess exists in the form of PL-rich vesicles or liposomes.6 –9 Their dispersion in the continuous aqueous phase of the emulsion constitutes a “ mesophase ” that can be separated by centrifugation of the emulsion.10,11 Because photographs of liposomes from commercially available fat emulsions are rarely found in the literature,7,12 we have performed freeze-fracture electron microscopy (FFEM) on native 10% Intralipid, a mesophase, and a mesophase subfraction enriched with liposomes.9 This emulsion, which has been used extensively and has been studied for
Correspondence to: Andre´ Bach, PhD, Centre d’Ecologie et Physiologie Energe´tiques, 67087 Strasbourg Cedex 2, France. E-mail: [email protected] Nutrition 17:930 –933, 2001 ©Elsevier Science Inc., 2001. Printed in the United States. All rights reserved.
more than 30 y, is recognized for its high stability and reproducible physicochemical characteristics. Nutritionists now recommend infusing the 20% formula (higher TAG:PL ratio) because liposomes decrease the clearance rate of the infused fat.13 However, we used the 10% formula, which is richer in liposomal PLs11 and poorer in TAGs, thus needing no further dilution. We investigated whether liposomes purified by ultracentrifugation have the same ultrastructure as those in native emulsion. The comparison of our results with those obtained previously on the same emulsion by other techniques7,9 prompted us to reevaluate liposome size and composition. We also report some preliminary data that draw attention to trace amounts of proteins in the mesophase of parenteral emulsions.
MATERIALS AND METHODS The starting material for our study was a commercially available emulsion (10% Intralipid, Kabi-Pharmacia, St-Quentin en Yvelines, France) composed of soybean oil and egg PL (PL:TAG ratio of 12:100 [w/w]). The procedures used for emulsion fractionation and the assays have been described in detail.9 In a first step, a 12-mL sample of native emulsion was centrifuged at 15°C (L8-70, Beckman Inc., Palo Alto, CA, USA) in a heat-sealed polyallomer tube (Beckman) at 20 000 rpm (27 000g) for 1 h using an angular rotor. The tube was cut with a tube-slicer (Beckman) just under the 0899-9007/01/$20.00 PII S0899-9007(01)00667-0
Nutrition Volume 17, Number 11/12, 2001
Physicochemical Characterization of Parenteral Fat Emulsions
931
RESULTS AND DISCUSSION Particle Size
FIG. 1. Result of the two-step fractionation of the emulsion. Left: Upper lipid cake constituted of compacted artificial chylomicrons and an infranatant named mesophase after the first step of centrifugation (see text). Right: Upper residual artificial chylomicrons and opalescent ring of liposomes (named the phospholipid-rich fraction) in the middle of the tube after the second step of centrifugation.
limit of the lipid cake formed by the compacted artificial chylomicrons at the top of the tube (Fig. 1, left). A 2-mL mesophase sample of the homogeneous infranatant, or mesophase, was then centrifuged at 15°C in a KBr density gradient for 24 h at 40 000 rpm (100 000g) in a swinging bucket. A white layer at the surface showed the presence of a minor population of residual artificial chylomicrons (density ⬍ 1.008 g/mL). The opalescent ring in the middle of the tube (density range ⫽ 1.010 –1.030 g/mL: Fig. 1, right) corresponds to a PL-rich fraction (PLRF) of the mesophase, which was harvested as a population of purified liposomes.9 Small amounts each of PLRF, mesophase, and native emulsion sample containing 30% glycerol as a cryoprotectant were deposited onto a copper planchet and rapidly frozen in liquid propane. It was subsequently fractured and shadowed in a Balzers freeze etch unit using platinum carbon shadowing. The replicas were examined with a Philips 410 electron microscope. A specific antibody raised against the anionic polypeptide factor (APF), isolated by high-performancce liquid chromatography from human bile, was used to characterize proteins in the mesophase by an enzyme-linked immunosorbent assay, as previously described.14
The only published pictures of liposomes in parenteral fat emulsions we know of concern diluted Intralipid (with a possible effect of dilution on the particle lipid organization) or a noncommercially available fat emulsion. Rotenberg et al.7 published a picture of the mesophase isolated from a five-fold diluted 10% Intralipid that showed liposomes coexisting with residual artificial chylomicrons when viewed by cryoelectron microscopy. Later, Westesen and Wehler8,12 published a micrograph obtained by FFEM from a diluted (1/1 [v/v]) model emulsion (soybean oil and fractionated egg yolk PL; PL:TAG ratio of 18:250 [w/w]) and concluded that most of the particles were smaller than 100 nm in diameter. In neither study was the size of the liposomes accurately assessed. The native emulsion (Fig. 2A) showed a population of large artificial chylomicrons (mean diameter ⫽ 200 ⫾ 10 nm, n ⫽ 59) with the typical cross fracture showing a particle filled with material8 that accounted for about 30% of the total particles. They coexisted with smaller particles (mean diameter ⫽ 33.8 ⫾ 2.1 nm, n ⫽ 140), most of which presented the typical concave or convex fracture surface of lipidic vesicles. We did not detect stacked-layer structures dispersed between artificial chylomicrons, as reported in several studies6,15,16 but likely corresponding to artifacts, especially after negative staining. Our results are consistent with the qualitative observations of Westesen and Wehler8,12 on their model emulsion. In the mesophase (Fig. 2B), the particles had a mean diameter of 39 ⫾ 2 nm (n ⫽ 209). About 15% of them showed a cross fracture, indicating the presence of small artificial chylomicrons. After further centrifugation, the resultant PLRF (Fig. 2C) was not very different from the corresponding mesophase with regard to particle shape. However, the population of particles was more homogeneous: only 1% to 2% of the particles corresponded to residual artificial chylomicrons (mean diameter ⫽ 45 nm). All the others were liposomes with a mean diameter of 32.5 ⫾ 0.8 nm (n ⫽ 229) and thus smaller (P ⬍ 0.0001, Student’s t test) than the mesophase particles (mixture of liposomes and residual artificial chylomicrons) and close to the size of liposomes observed in the native emulsion. Because the fracture does not necessarily propagate through the largest possible cross section of the spherical particles, one has to correct for the underestimated size obtained by FFEM by a factor of about 30%.17 In this way, 10% Intralipid contains artificial chylomicrons with a mean diameter of about 200 ⫻ 1.3, i.e., 260 nm, and liposomes of about 33 ⫻ 1.3, i.e., 43 nm. Regarding the size of the artificial chylomicrons, there was good agreement despite the type of technique used. Indeed, the mean diameter of 10% Intralipid artificial chylomicrons was 250 to 280 nm as estimated by light scattering spectroscopic techniques5,7,9,16,18,19 and electron microscopy.5,7,20 –22 Conversely, the liposome mean diameter of 40 nm determined by FFEM was discrepant with the 70 to 100 nm measured by quasi-elastic light scattering.7,9 This latter value is, however, probably overestimated, as light scattering techniques23 performed on heterogeneously dispersed particles have been known to emphasize the larger ones7 or overlook particles smaller than 100 nm.8,12 Consequently, the new value found by FFEM seems more reliable. With the same FFEM technique, 40-nm liposomes were found in another commercial emulsion of analogous composition (B. Braun, Boulogne, France; results not shown). This size seems more logical because it allows a better understanding of the intravascular metabolism of infused liposomes. It is difficult to understand how 70- to 100-nm particles could be converted into smaller, 40- to 60-nm-diameter24 lipoprotein-X particles after cholesterol loading and acquisition of apolipoproteins.13 Moreover, the reevaluated liposome size improves the interpretation of the 31P-nuclear magnetic resonance
932
Fe´ re´ zou et al.
Nutrition Volume 17, Number 11/12, 2001 43 nm mean diameter, the out:in PL ratio thus corresponds to 1.50, which is closer to our experimental value. The remaining slight discrepancy can be explained by small residual artificial chylomicrons in the mesophase (Fig. 2B). Their presence is easily revealed by the significant amounts of TAGs found by thin layer chromatographic analysis of the mesophase, whatever the type of emulsion considered. The amount of TAGs has been quantified not only in Intralipid mesophase (30% to 40% of total lipids [w/w], regardless of the formula) but also, although less, in PLRF (8%, 15%, and 30% [w/w] for 10%, 20%, and 30% Intralipid, respectively).9 Such amounts of TAGs in purified liposomes are incompatible with regard to the maximal solubilization capacity of TAGs in PL bilayers, considered to equal 3.1% (w/w; 2.8%, mol/mol).25 However, this value was obtained with liposomes differing strongly from the liposomes of parenteral fat emulsions. The differences were in the chemical composition (triolein versus soybean oil; purified phosphatidylcholine versus fractionated mixture of PLs from egg yolk), the PL:TAG (w/w) ratio (94:0 to 6 in the presence of 6-0 cholesteryl oleate versus 12:100), and the manufacturing technique (sonication versus homogenization at high pressure and sterilization around 120°C). We believe that these factors increase the maximal solubilization capacity of TAGs in PL bilayers, thus explaining the abundance of TAG in liposomes of parenteral fat emulsions. Proteins
FIG. 2. Electron micrographs of the native emulsion (A), the mesophase (B), and the phospholipid-rich fraction (C). Arrow indicates a cross fracture of a small artificial chylomicron.
spectra that showed7,9 the coexistence in native emulsions of artificial chylomicrons (external PLs only) and liposomes (internal and external PLs) and the predominance of unilamellar liposomes in the mesophase. To calculate the distribution of PLs between the two populations of particles, the value of 1.1 was used for the out:in PL ratio of liposomes, assuming a mean diameter of 70 to 100 nm.7 However, the ratio we measured on the mesophase of 10% Intralipid was significantly higher (1.65).9 For liposomes of
Another parameter to consider is the composition of liposomes. Although it is assumed that commercial emulsions do not contain any protein, a trichloroacetic acid precipitation indicates the presence of proteins in their mesophase. Using bovine serum albumin as a standard, we measured26 87 g of protein per milliliter of mesophase of 10% Intralipid (a direct protein measurement in the native emulsion is impeded by the lactescence due to the very high TAG content), for a protein:PL ratio of about 1.47% (w/w; i.e., 0.09:6.11). Where do these proteins come from? By analogy with the phytosteryl esters found in commercial emulsions,9,27 one possibility is that they accompany the soybean oil as oleosins. This family of hydrophobic proteins stabilizes the stored lipids in all seeds, interacts with PLs, and has been found in commercial alimentary oils.28 But the fact that the mesophase protein:PL ratio seems to be little influenced by the formula (1.57% [w/w]; i.e., 0.055:3.49 for the 20% Intralipid) suggests that the proteins of the emulsion are delivered also by the egg yolk PLs (amounting 12 g in both formulas). Interestingly, a PL-binding hydrophobic protein, the APF, from human and animal bile,14,29 showed a strong interaction with amphipathic PLs. The same strong association was observed between egg yolk lecithin (Sigma, Saint Quentin Fallavier, France) and human APF (APF:PL ratio of 5.01% [w/w]) during the preparation of proteoliposomes containing APF.30 In this study, APF promoted the excretion of cholesterol and PLs into the bile. This protein, which is also a minor apolipoprotein of high-density lipoproteins,14 acts as an antinucleating factor in bile via its vesicle-stabilizing properties.31,32 Thus, with the use of the specific antibody (MD94) raised against purified human APF, an APF signal was detected by enzyme-linked immunosorbent assay32 for the mesophase of Intralipid, corresponding to 2.32 g/mL for the 10% formula. The expected molecular weight of APF (7 kDa) was confirmed by Western-blotting (Fig. 3). The significance of APF in commercial emulsions is still unknown, but it is noteworthy that similar or higher values were obtained by enzyme-linked immunosorbent assay on mesophase from other lipid emulsions (APF g/mL mesophase: Intralipid 20%: 3.01; Medialipide from B. Braun: 10%, 5.56; 20%, 0.88, Endolipide from B. Braun: 10%, 4.72; 20%, 5.35), whereas assays with soybean PL-stabilized parenteral emulsions (Lipofundin S, Braun, Melsungen, Germany) did not show any immunoreactive proteins when using the same APF antibody. Therefore, egg yolk
Nutrition Volume 17, Number 11/12, 2001
FIG. 3. Western blot of Intralipid mesophase incubated in the presence of an antibody against anionic polypeptide factor (MD94).
PLs probably deliver not only residual amounts of cholesterol9,13 but also trace amounts of APF or related peptidic components. It can be assumed that the type of treatment and the level of purification of egg yolk lecithin influence the amount of proteins in the emulsifier. The presence of hydrophobic proteins (provided by oil and lecithin) in fat emulsions, which influences the phase diagram, could determine the lipids organization in the different particle populations and also might account for certain allergic reactions observed in infused patients.33 The use of FFEM allowed the size of 10% Intralipid liposomes to be reevaluated (40 nm versus the classic 70 to 100 nm in diameter) in the native emulsion and its mesophase. Our results showed that the liposome structure is not altered by ultracentrifugation. In the mesophase, which is composed mainly of liposomal PLs, we detected minute amounts of proteins. They were partly characterized as small hydrophobic proteins delivered by egg yolk lecithin used as the emulsifier.
ACKNOWLEDGMENT Arthur Pape is gratefully acknowledged for his help in editing the manuscript.
REFERENCES 1. Schuberth O, Wretlind A. Intravenous infusion of fat emulsion, phosphatides and emulsifying agents. Clinical and experimental studies. Arch Chir Scand 1961; 278(suppl):1 2. Rydhag L. The importance of the phase behaviour of phospholipids for emulsion stability. Fette Seifen Anstrichm 1979;4:168 3. Schoefl GI. The ultrastructure of chylomicra and of the particles in an artificial fat emulsion. Proc R Soc B 1968;169:147 4. Carlson LA. Studies on the fat emulsion Intralipid. I. Association of serum proteins to Intralipid triglyceride particles (ITP). Scand J Clin Lab Invest 1980; 40:139
Physicochemical Characterization of Parenteral Fat Emulsions
933
5. Erkelens DW, Chen C, Mitchell CD, Glomset JA. Studies of the interaction between apolipoproteins A and C and triacylglycerol-rich particles. Biochim Biophys Acta 1981;665:221 6. Groves MJ, Wineberg M, Brain APR. The presence of liposomal material in phosphatide stabilized emulsions. J Disper Sci Tech 1985;6:237 7. Rotenberg M, Rubin M, Bor A, et al. Physico-chemical characterization of Intralipid emulsions. Biochim Biophys Acta 1991;1086:265 8. Westesen K, Wehler T. Investigation of the particle size distribution of a model intravenous emulsion. J Pharm Sci 1993;82:1237 9. Fe´ re´ zou J, Lai NT, Leray C, et al. Lipid composition and structure of commercial parenteral emulsions. Biochim Biophys Acta 1994;1213:149 10. Jakoi L, Quarfordt SH. The induction of hepatic cholesterol synthesis in the rat by lecithin mesophase infusions. J Biol Chem 1974;249:5840 11. Lutz O, Meraı¨hi Z, Fe´ re´ zou J, et al. The mesophase of parenteral fat emulsion is both substrate and inhibitor of lipoprotein lipase and hepatic lipase. Metabolism 1990;39:1225 12. Westesen K, Wehler T. Physicochemical characterization of a model intravenous oil-in-water emulsion. J Pharm Sci 1992;81:777 13. Bach AC, Fe´ re´ zou J, Frey A. Phospholipid-rich particles in commercial parenteral fat emulsions. An overview. Prog Lipid Res 1996;35:133 14. Domingo N, Grosclaude J, Bekaert ED, et al. Epitope mapping of the human biliary amphipathic, anionic polypeptide: similarity with a calcium-binding protein isolated from gallstones and bile, and immunologic cross-reactivity with apolipoprotein A-I. J Lipid Res 1992;33:1419 15. Erkelens DW, Mocking JAJ. The effect of apolipoprotein A on in vitro apolipoprotein C binding and in vivo removal of artificial triacylglycerol-rich particles. Metabolism 1985;34:222 16. Li JM, Caldwell KD. Structural studies of commercial fat emulsions used in parenteral nutrition. J Pharm Sci 1994;83:1586 17. Heegaard CW, Le Maire M, Gulik-Krzywicki T, Moller JV. Monomeric state and Ca2⫹ transport by sarcoplasmic reticulum Ca2⫹-ATPase, reconstituted with an excess of phospholpid. J Biol Chem 1990;265:12020 18. Lutz O, Meraı¨hi Z, Mura J-L, et al. Fat emulsion particle size: Influence on the clearance rate and the tissue lipolytic activity. Am J Clin Nutr 1989;50:1370 19. Davis SS. The stability of fat emulsions for intravenous administration. In: Johnston IDA, ed. Advances in clinical nutrition. Lancaster: MTP Press, 1983: 213 20. Hamilton-Attwell VL, Du Plessis J, Van Wyk CJ. A new scanning electron microscope (SEM) method for the determination of particle size in parenteral fat emulsions. J Microsc 1987;145:347 21. Du Plessis J, Tiedt LR, Van Wyk CJ, Ackermann C. A new transmission electron microscope method for the determination of particle size in parenteral fat emulsions. Int J Pharm 1986;34:173 22. Connelly PW, Kuksis A. Effect of core composition and particle size of lipid emulsions on apolipoprotein transfer of plasma lipoproteins in vivo. Biochim Biophys Acta 1981;666:80 23. Marangoni AG, Rousseau D. The influence of chemical interesterification on physicochemical properties of complex fat systems—1. Melting and crystallization. J Am Oil Chem Soc 1998;75:1265 24. Miyahara T, Fujiwara H, Yae Y, et al. Abnormal lipoprotein appearing in plasma of patients who received a ten percent soybean oil emulsion infusion. Surgery 1979;85:566 25. Hamilton JA, Miller KW, Small DM. Solubilization of triolein and cholesteryl oleate in egg phosphatidylcholine vesicles. J Biol Chem 1983;258:12821 26. Lowry OH, Rosebrough AL, Farr RJ. Protein measurement with folin phenol reagent. J Biol Chem 1951;192:265 27. Clayton PT, Bowron A, Mills KA, et al. Phytosterolemia in children with parenteral nutrition–associated cholestatic liver disease. Gastroenterology 1993; 105:1806 28. Olszewski A, Pons L, Moute´ te´ F, et al. Isolation and characterization of proteic allergens in refined peanut oil. Clin Exp Allergy 1998;28:850 29. Lafont H, Domingo N, Groen AK, et al. AFP/CBP, the small amphipathic anionic protein(s) in bile and gallstones, consist of lipid-binding and calcium-bonding forms. Hepatology 1997;34:1054 30. Martigne M, Domingo N, Chanussot F, et al. Effect of bile anionic polypeptidic fraction on the fate of cholesterol carried by liposomes in the rat. Proc Soc Exp Biol Med 1988;187:234 31. Halpern Z, Lafont H, Arad J, et al. The distribution of biliary anionic fraction (APF) between cholesterol carriers in bile and its effects on nucleation. Hepatology 1994;21:979 32. Domingo N, Groen AK, Leche`ne P, et al. L’APF/CBP, une nouvelle prote´ ine capable a` la fois de lier le calcium et d’induire la fusion de ve´ sicules phosphatidylcholine/choleste´ rol dans la bile. Med Chir Digest 1993;22:117 33. Aimone-Gastin I, Gueant JL, Laxenaire MC, Moneret-Vautrin DA. Pathogenesis of allergic reactions to anaesthetic drugs. Int J Immunopathol Pharmacol 1997; 10:193
(For an additional perspective, see Editorial Opinions)