- Email: [email protected]
The effects of phospholipid molecular species on cholesterol crystallization in model biles: the influence of phospholipid head groups
Journal of Hepatology 1998; 28: 1008–1014 Printed in Denmark ¡ All rights reserved Munksgaard ¡ Copenhagen
Copyright C European Association for the Study of the Liver 1998
Journal of Hepatology ISSN 0168-8278
The effects of phospholipid molecular species on cholesterol crystallization in model biles: the influence of phospholipid head groups Yehuda Ringel, Giora J. So¨mjen, Fred M. Konikoff, Ruth Rosenberg, Moshe Michowitz and Tuvia Gilat Department of Gastroenterology, Tel-Aviv Sourasky Medical Center, Ichilov Hospital, Tel-Aviv, and the Minerva Center for Cholesterol Gallstones and Lipid Metabolism in the Liver, Sackler Faculty of Medicine, Tel-Aviv University, Israel
Background/Aims: Variations in the molecular species of biliary phospholipids have been shown to exert major effects on cholesterol solubility and carriers in model and human biles. The aim of this study was to explore systematically the effects of various phospholipid head groups on the cholesterol crystallization process in model biles. Methods: Three different control model biles were prepared using varying proportions of egg lecithin, cholesterol and Na taurocholate. In the test biles, 20% of the egg lecithin was replaced with synthetic phosphatidylserine, phosphatidylethanolamine, phosphatidylglycerol or phosphatidylcholine, keeping the phospholipid acyl chains and other biliary lipids constant in each experiment. Results: Phosphatidylserine and phosphatidylglycerol significantly prolonged the crystal observation time, from 2 days to 10 and 6 days, respectively (p∞0.02), while phosphatidylethanolamine had little and phos-
phatidylcholine no effect. The crystal growth rate was significantly slowed down with 20% phospholipid replacement in the following order: phosphatidylglycerol ±phosphatidylserine ±phosphatidylethanolamine. The total crystal mass after 14 days, as measured by chemical analysis, was reduced by 59% with phosphatidylserine (p∞0.05), and by 73% with phosphatidylglycerol (p∞0.05); while phosphatidylethanolamine had little effect. The precipitable cholesterol crystal fractions after 14 days were significantly reduced with phosphatidylserine (54%) and phosphatidylglycerol (37%), but not with phosphatidylethanolamine or phosphatidylcholine. Conclusions: Variations in the head groups of biliary phospholipids may markedly slow down the cholesterol crystallization process in model biles.
and bile salts are two ampiphilic biliary lipid molecules, which are crucial in the secretion, solubilization and transport of biliary cholesterol. Both are present in bile as a mixture of several molecular species. Biliary research in recent decades has focused mainly on the various molecular species of bile salts. It was shown that minimal changes in bile salt molecules induced the secretion of undersaturated bile (1–4), supersaturated bile (5), affected intestinal cholesterol absorption (6), and caused liver toxicity in mammals (7). The importance of changes in the molecular forms
of biliary phospholipids emerged more recently. The addition of small amounts of specific phospholipids normalized the rapid nucleation time of lithogenic biles, while the addition of equimolar amounts of bile salts had no such effect (8). Increasing amounts of lecithin progressively prolonged the nucleation time of model and human biles (8,9). More importantly, changes in the head group or fatty acid chains of the phospholipid molecule resulted in major alterations in the nucleation time and cholesterol carriers in bile. These occurred without any change in the absolute or relative concentration of phospholipids in the model bile (9–12). As the results of bile salt therapy for gallstone dissolution or prevention of recurrence were shown to be suboptimal, it was thought worthwhile to explore the potential value of biliary phospholipid modulation. In a previous study we investigated the effects of varying
P
Received 9 July 1997; revised 12 January; accepted 15 January 1998
Correspondence: Fred Konikoff, Department of Gastroenterology, Ichilov Hospital, 6 Weizman St., Tel Aviv 64239, Israel. Tel: 972-3-697 4470. Fax: 972-3-697 4622. e-mail: konikoff/post.tau.ac.il
1008
Key words: Cholesterol crystallization; Model bile; Phospholipid head groups.
Phospholipid species and cholesterol crystallization
saturation of the sn-2 fatty acid of the major biliary phospholipid, phosphatidylcholine (12). In the present study we have investigated the effects of various phospholipid head groups while keeping the fatty acid composition constant.
Materials and Methods Cholesterol (Sigma, St. Louis, Mo, USA) was twice recrystallized from hot ethanol and Na taurocholate (Na TC) (Sigma, St. Louis, Mo, USA) was twice recrystallized from ethanol and ether (13). Egg phosphatidylcholine (EPC) and de novo synthesized phosphatidylethanolamine (PE), phosphatidylserine (PS) and phosphatidylglycerol (PG) (Avanti Polar Lipids, Alabaster, Ala, USA) were used without further purification. All lipids used in this study were pure by TLC standard. The radiolabeled lipids, [3H]-cholesterol, and Palmitoyl-2[1-14C] oleoyl-PC, were purchased from Amersham Laboratories (Buckinghamshire, UK). Preparation of model bile solutions Egg phosphatidylcholine (EPC), cholesterol and Nataurocholate (NaTC) mixtures were dissolved in CHCl3/CH3OH (2:1 v/v), dried under N2 at room temperature, lyophilized overnight and kept at ª20æC under argon until used. Model bile solutions were prepared by suspending the dried lipids in 150 mM NaCl, 1.5 mM disodium EDTA, 50 mM Tris-HCl pH 8.0 (NaTE) and incubating the suspension at 55æC for 1 h. The solubilized model biles were incubated in sealed vials under argon at 37æC for the duration of the experiments. Aliquots from the models were examined daily during the first 6 days and subsequently every 2 days. Types of bile models. Three different model biles were used in all experiments. The models were all composed to simulate human gallbladder bile. The three model biles (A, B, & C) were chosen to represent the range of variability occurring in native biles in terms
TABLE 1 Model solutions used Model CHOL. PL mM mM
TC mM
CSI %
Total lipids Crystallization g/dl characteristics
A
18.5
33.6
120
160
9.2
B
15
30
150
134
10.3
C
10
20
100
146
6.8
COT – 2.0 CGR – 0.21 COT – 1.3 CGR – 0.09 COT – 4.0 CGR – 0.09
CHOLΩcholesterol; PLΩphospholipids; TCΩtaurocholate; CSIΩ cholesterol saturation index; COTΩcrystal observation time [days]; CGRΩcrystal growth rate [OD/day].
of lipid composition and crystallization kinetics. The models displayed short or long COT and fast or slow CGR, as shown in Table 1. Types of phospholipids tested. Within each model (A, B & C), substitution experiments were performed. One hundred per cent egg lecithin was used for preparation of the control solutions. The other investigated bile solutions were prepared by substituting 20% of the egg lecithin by various synthetic phospholipids (Table 2). Two series of experiments were performed: in series 1 the fatty acids were 16:0 and 18:1, while in series 2 they were 16:0–18:2. Various head groups were tested in each series (Table 2). All models were prepared in triplicate and kept under the same conditions throughout the experimental period. Dose-dependency experiments were performed using model biles A&B. One hundred per cent egg lecithin was used for the control solutions which were compared to the test solutions prepared by substituting the egg lecithin with increasing proportions (1%, 2.5%, 5%, 10%, 20%) of the synthetic phospholipids. Evaluation of cholesterol crystal formation and growth Crystal observation time (COT) assay. Crystal observation time (also called ‘‘Nucleation time’’) was determined as described by Holan et al. (14). Aliquots (5 ml) from each model bile were examined by polarized light microscopy. The COT was determined as the time of first detection of at least three cholesterol monohydrate crystals per microscopic field at 100-fold magnification. Crystal growth rate (CGR) assay. Crystal growth was monitored spectrophotometrically using a SPEC-
TABLE 2 Alterations in phospholipid head group Series 1 (Fatty acids: 16:0–18:1) Phospholipid head Fatty acid chains group Sn-1 Sn-2 PE PS PG EPC (Control)
16:0 16:0 16:0 Mixed
18:1 18:1 18:1
Series 2 (Fatty acids: 16:0–18:2) Phospholipid head Fatty acid chains group Sn-1 Sn-2 PC PE PS EPC (Control)
16:0 16:0 16:0 Mixed
18:2 18:2 18:2
PCΩphosphatidylcholine; PEΩphosphatidylethanolamine; phosphatidylserine; EPCΩegg PC.
PSΩ
1009
Y. Ringel et al.
Fig. 1. A representative crystal growth curve (model A, Table 1). The OD of the solution (after dissolution of noncrystalline compounds) during 14 days of incubation. The squares represent the means of triplicate samples measured in duplicate. The dashed line gives an estimation of the CGR by linear regression analysis, and the arrow shows the OD difference between day 0 and day 14. The error bars represent the SEM (nΩ6).
TRA – microplate reader (SPECTRA – STL, Austria) as described by us previously (15). Aliquots (50 ml) of the lipid solutions were mixed and shaken vigorously with equivalent volumes of sodium taurodeoxycholate (NaTDC) (200 mM) in microplate wells. After 60 min at room temperature, the microplates were shaken again and the absorbance, at 405 nm, in each well was measured. Each model was prepared in triplicate and sampled in duplicate for measurement. The data were collected and analyzed by an IBM-compatible personal computer, and the optical density was calculated. A graph describing the averaged optical density changes for each solution was plotted. The slope in the steepest region of the curve was determined by a linear regression fit to at least three measurements and defined as the crystal growth rate (Fig. 1). Total crystal mass. Chemical analyses of cholesterol and phospholipids were performed on each sample on the last day of the experiment (day 14), as previously described (16). The samples were collected from the micro-wells, centrifuged in an Airfuge (Beckman) at 70 000 rpm for 5 min and the supernatant solutions were separated. Lipid determinations were performed on the total samples (before centrifugation) as well as on the supernatant solutions. The amount of cholesterol in the pellet was calculated by subtracting the amount in the supernatant solution from the total. The crystalline character of the pellet was confirmed by polarized light microscopy. Crystal mass was also estimated spectrophotometrically by measuring the difference in optical density between day 14 and day 0 (Fig. 1). Ultracentrifugation assay. Precipitable cholesterolcontaining lipid aggregates were harvested by sequential ultracentrifugation at predetermined time points
1010
(days 3 and 14) during the crystallization process, as previously described (12,17). In brief, the aggregates were first precipitated from a sample (0.5 ml) of the nucleating model bile by ultracentrifugation (200 000 g, 1 h, 20æC) in 0.15 M Nacl (4.5 ml). This procedure has been previously validated, and shown to separate multilamellar vesicles and all microcrystalline structures from the soluble phase of the model bile (17). The pellet was then placed on a sucrose gradient (0– 20%) in 5.5-ml nitrocellulose tubes, and centrifuged in a Centrikon T-1000 ultracentrifuge, using a TST 55.5 swinging bucket rotor (Kontron Instruments) at 200 000 g for 1 h at 20æC to separate the various cholesterol precipitates (17,18). Fractions (500 ml) were collected from the top and analyzed for lipid content by scintillation counting of the dually radiolabeled (3HCh/14C-PC) samples, and examined by phase contrast light microscopy. The density of the fractions was verified by a refractometer. Native bile Fresh gallbladder bile from a cholesterol gallstone patient was ultracentrifuged (200 000 g, 1 h, 25æC). The supernatant was collected and triplicate samples (0.5 ml) were used as controls. Additional samples of 0.5 ml were prepared by adding 6.72 mM synthetic 16:0– 18:2 phosphatidylserine in triplicate. The bile samples were kept in sealed vials under argon at 37æC throughout the experimental period. Aliquots (3 ml) were examined daily for determination of COT. Statistical analysis Each lipid dispersion was prepared in triplicate, and duplicate aliquots were measured from all samples. Mean values of optical density and standard errors were calculated. Crystal growth rates were calculated from linear regression analysis of the crystal growth curves as explained above. Analysis of variance was performed for the triplicates of each solution. The pvalue was calculated by Student’s t-test.
Results Series 1 (Fatty acids: 16:0–18:1; Phospholipids: PE, PS, PG) Crystal observation time. In model A the COT was minimally prolonged when 20% of the lecithin was replaced by PE. The COT was markedly prolonged with 20% PS and with 20% PG, to 10 days and 6 days respectively, vs 2 days in the 100% lecithin control solution. The findings were similar but less marked in models B and C in which the effect of PG was similar or greater than that of PS (Fig. 2). The differences in the COT between the test solutions and the lecithin
Phospholipid species and cholesterol crystallization
control solution were statistically significant (pΩ0.015) for PS in model A and for PS and PG in models B and C (pΩ0.02). Cholesterol crystal growth rate. The CGR in the 4 solutions of model A are shown in Fig. 3. Substitution of 20% of lecithin with PE decreased the CGR from 0.213 OD/day to 0.135 OD/day (pΩ0.035). This effect was much more marked with PS and PG, which decreased the CGR to 0.053 OD/day (pΩ0.007) and 0.035 OD/day (pΩ0.001), respectively. The results were similar, though less marked in models B and C (NS). Cholesterol crystal mass. The data were analyzed both by using the difference in OD between the first and the last day of the experiment (day 14) and by chemical analysis of the precipitated crystals on the
last day of the experiment. The results for model A are shown in Fig. 4. (The results were similar though less marked in models B&C.) It can be seen that substitution of 20% of the lecithin by PE produced only a minimal and nonsignificant decrease in the cholesterol crystal mass. However, with 20% PS and 20% PG the decreases were very marked and highly significant. Measuring the OD difference, the decrease was of 62.1% with PS and 79.4% with PG (pΩ0.002, pΩ0.001, respectively) compared to the lecithin control. By chemical analysis the decrease was of 59.2% with PS and 73.4% with PG (pΩ0.005, pΩ0.004 respectively) compared to the control solution. Precipitable cholesterol fractions. The fractions were precipitated and separated (see Methods) after 3 and 14 days of incubation. The findings for model A are shown in Fig. 5. On day 3 the differences were minimal, there were no crystals, and cholesterol was found only in supersaturated (cholesterol/phospholipid ratio ±1) multilamellar vesicles irrespective of the phospholipid molecular species in the solution. On day 14 mature crystals were seen in the high density fractions of the control lecithin solution, while the multilamellar vesicles (low density fractions) had become unsaturated (C/P ratio Æ1). The findings were similar with 20% PE, although the multilamellar vesicles remained saturated (C/P ratio ±1). However, with 20% PS there were fewer crystals (54% of control) and more cholesterol remained in the
Fig. 2. Mean crystal observation time (COT) as determined by polarized light microscopy for models A, B & C (see Table 1). The left bars represent the control (100% lecithin) solution. In the other bars 20% of the lecithin was replaced by synthetic phospholipids (PL) with different head groups as marked at the bottom of the bars (see Table 2) (nΩ3). *p∞0.02, compared to control. PEΩphosphatidylethanolamine; PCΩphosphatidylcholine; PSΩphosphatidylserine.
Fig. 3. Mean crystal growth rate (CGR) of model A as determined from the spectrophotometrically measured curves. The left bar represents the control with 100% lecithin. In the other bars 20% of the lecithin was replaced by synthetic PL, with different head groups as marked at the bottom of the bars (see Table 2) (nΩ3). *p∞0.04, **p∞0.001, compared to control. Abbreviations as in Fig. 2.
Fig. 4. Cholesterol crystal mass on day 14 (model A) was studied by two different methods. In panel a, the mean cholesterol crystal mass was estimated spectrophotometrically by the OD difference between days 14 and 0. In panel b, the mean amount of crystalline cholesterol was determined chemically, after centrifugation, on day 14 (see Methods). The left bars represent the controls with 100% lecithin. In the other bars 20% of the lecithin was replaced by synthetic PL with different head groups as marked at the bottom of the bars (see Table 2) (nΩ3). *p∞0.005, compared to control.
1011
Y. Ringel et al.
Fig. 5. Precipitable lipid structures (model A) were separated by density gradient ultracentrifugation from the four different test solutions. The density increases linearly from fraction 1 (1.01 g/ml) to fraction 10 (1.06 g/ml). The upper panels show the density profiles on day 3 and the lower panels on day 14 of incubation. (HΩCholesterol, (SΩphosphatidylcholine (PC)).
saturated multilamellar vesicles. The findings were similar with 20% PG, with even fewer crystals in the high density fractions (37% of control). Series 2 (Fatty acids: 16:0–18:2; Phospholipids: PC, PE, PS) Crystal observation time. The findings were similar to those found with the 16:0–18:1 PL, although the changes were much less marked. With 20% PS the COT was prolonged from 2.7 to 4 days (pΩ0.01). The findings in models B and C were similar with PS significantly prolonging the COT (pΩ0.024, pΩ0.01, respectively) compared to the control solutions. However, the substitution of 20% of lecithin with a synthetic 16:0– 18:2 PE or PC did not significantly change the COT (data not shown). Cholesterol crystal growth rate. In series 2 experiments, 20% PE did not affect the CGR, while 20% PS significantly reduced it (pΩ0.023). Again, 20% PC did not change the CGR as compared to the control lecithin solution (data not shown).
1012
Cholesterol crystal mass. Twenty percent PE reduced the crystal mass, by chemical analysis, only minimally (NS), whereas 20% of PS reduced the crystal mass by 39% (pΩ0.00005). Measurements of the OD difference produced results similar to those obtained by chemical analysis. In all 3 models, 20% of 16:0–18:2 PC had no effect on the crystal mass. Dose-dependency study There was a significant dose-dependent effect of PS on the COT and CGR. In model A the COT was markedly prolonged by 5%, 10% and 20% PS (5 days, 5.7 days and 6 days, respectively vs. 2.7 days in the 100% lecithin control solution; pΩ0.0003). The findings were similar, but less marked in model B, in which only 20% and 10% PS significantly prolonged the COT (2.3 days and 1.7 days, respectively, vs. 1 day in the 100% lecithin control solution; pΩ0.011). When lecithin was substituted with 20% and 10% PS the CGR, in model B, decreased from 0.159 OD/day to 0.07 OD/day and to 0.06 OD/day (pΩ0.0008), re-
Phospholipid species and cholesterol crystallization
spectively. This effect was similar, though less marked (NS) in model A. Native bile The lipid composition of the native bile was 13.4 mM cholesterol, 54.2 mM PL and 139.4 mM bile salts; hence, the added PS amounted to 11% of the total PL. In the samples with the added synthetic PS, the COT was markedly prolonged from 2 days to 7 days.
Discussion Our experiments demonstrate that replacing 20% of the lecithin by PS or PG markedly and significantly prolonged the COT, slowed the CGR and diminished the total crystal mass after 14 days of incubation. Under the same conditions PE produced only a minor effect, but with a similar trend. Changing the fatty acids from 16:0– 18:1 to 16:0–18:2 did not significantly modify the above results, and the effects of the head groups remained unchanged. Substituting 20% of the lecithin (a natural mixture of several molecular forms of PC) with a synthetic 16:0–18:2 PC did not have any noticeable effect. Thus, of the three head groups tested (PC, PE, PS), it is the last one that had the main anti-crystallizing effect. Unexpectedly, PG which is devoid of a head group produced a marked similar effect. In this study we measured the crystallization process of cholesterol, employing a combination of techniques. The crystallization-retarding effects of PS and PG were notable in terms of both crystallization kinetics (COT and CGR) and total crystal mass (OD difference, chemical analysis and ultracentrifugal separation). We were unable to find similar experiments in the published literature. Halpern et al. in 1993 (9) performed experiments in model bile solutions containing a single synthetic phospholipid (not natural lecithin as in the present experiments). When all the phospholipid in the model solution was PE, almost all the cholesterol eluted in the vesicular fraction. With 100% PS, almost all the cholesterol was found in the micellar fraction. With PC, cholesterol was found in both carriers. This effect of PS was also demonstrated in human bile. In all these experiments the phospholipid fatty acids, bile salts and cholesterol were kept constant. The nucleation time was also studied by Halpern et al., with all the phospholipid in a given model bile solution represented by a single synthetic phospholipid. The nucleation time was 5 days with PC, 6 days with PE and 20 days with PS, supporting the results found in the present study. However, the experiments of Halpern et al., were less physiologic since the phospholipids in human bile are never composed of a single molecular species. Moreover, Halpern et al. did not study the chol-
esterol crystal growth rate nor the final crystal mass. They showed a progressive prolongation of the nucleation time of human bile following the addition of exogenous lecithin. The dose-dependency study showed that the effect of PS was also evident at a concentration of 10% and even 5% of total phospholipids. PS in a proportion of 5% of total biliary phospholipids is not too far from the physiologic proportions. It is possible that even with crude methods of modulation such as oral supplementation, such levels could be achieved. This however remains to be demonstrated. Our single pilot experiment suggests that PS might be active in vivo in human bile. The crystallization kinetics as measured in model solutions in the present study cannot be performed in native bile. The pigmentation and turbidity of bile (due to non-lipid components, proteins, etc.) preclude these dynamic measurements of optical density. The COT is a very crude measurement that may mask considerable differences in crystallization kinetics. This was demonstrated by Groen’s group when studying the pronucleating effects of biliary glycoproteins (19) and by us when studying the antinucleating effects of APF. These were not demonstrable in human bile (20), but were clearly demonstrable in model solutions (21) using the present techniques. It is, however, very likely that the present findings also apply to human bile. It was previously shown that small amounts of phospholipids markedly prolonged the nucleation time of bile (8,9). It was also shown that this effect was much magnified when the molecular composition of the added phospholipid had strong antinucleating properties (8) (such as disaturated phospholipids). In the present pilot experiment, the COT of human bile was markedly prolonged by PS at a concentration of 11% of biliary phospholipids. The mechanism by which different molecular configurations of phospholipids affect cholesterol crystallization is at present speculative. It is possible that PS and PG exert their effect by virtue of their negative charge that is absent in PC, which has no net charge. The present study demonstrates the considerable effects of variations in the phospholipid head group on the cholesterol crystallization process in model bile. These data add to the growing body of evidence on the importance of biliary phospholipid molecular species for cholesterol homeostasis in bile. They raise the prospect of preventing or retarding cholesterol crystallization in human bile if and when significant modulation of biliary phospholipids becomes feasible.
Acknowledgements This study was supported in part by a grant from the Chief Scientist’s Bureau, The Israel Ministry of Health.
1013
Y. Ringel et al.
References 1. Danzinger RG, Hofmann AF, Schoenfield LJ, Thistle JL. Dissolution of cholesterol gallstones by chenodeoxycholic acid. N Engl J Med 1972; 286: 1–8. 2. Hofmann AF, Thistle JL, Klein PD, Szczepanik PA, Yu PYS. Chemotherapy for gallstone dissolution. II. Induced changes in bile composition and gallstone response. JAMA 1978; 239: 1138–44. 3. Ju¨ngst D, Brenner G, Pratschke E, Paumgartner G. Low-dose ursodeoxycholic acid prolongs cholesterol nucleation time in gallbladder bile of patients with cholesterol gallstones. J Hepatol 1989; 8: 1–6. 4. Loria P, Carulli N, Medici G, Menozzi D, Salvioli G, Bertolotti M, et al. Effect of ursocholic acid on bile lipid secretion and composition. Gastroenterology 1986; 90: 865–74. 5. Marcus SN, Heaton KW. Deoxycholic acid and the pathogenesis of gall stones. Gut 1988; 29: 522–33. 6. Hofmann AF. Pharmacology of ursodeoxycholic acid, an enterohepatic drug. Scand J Gastroenterol 1994; 29: 1–15. 7. Cohen BI, Hofmann AF, Mosbach EH, Stenger RJ, Rotschild MA, Hagey LR, et al. Differing effects of norursodeoxycholic or ursodeoxycholic acid on hepatic histology and bile acid metabolism in the rabbit. Gastroenterology 1986; 91: 189–97. 8. Ju¨ngst D, Lang T, Huber P, Lange V, Paumgartner G. Effect of phospholipids and bile acids on cholesterol nucleation time and vesicular/micellar cholesterol in gallbladder bile of patients with cholesterol stones. J Lipid Res 1993; 34: 1457– 64. 9. Halpern Z, Moshkowitz M, Laufer H, Peled Y, Gilat T. Effect of phospholipids and their molecular species on cholesterol solubility and nucleation in human and model biles. Gut 1993; 34: 110–5. 10. Tao S, Tazuma S, Kajiyama G. Fatty acid composition of lecithin is a key factor in bile metastability in supersaturated model bile systems. Biochim Biophys Acta 1993; 167: 142–6. 11. Tazuma S, Ochi H, Teramen K, Yamashita Y, Horikawa K, Miura H, et al. Degree of fatty acyl chain unsaturation in
1014
12.
13. 14.
15.
16. 17.
18.
19.
20.
21.
biliary lecithin dictates cholesterol nucleation and crystal growth. Biochim Biophys Acta 1994; 1215: 74–8. Ringel Y, So¨mjen GJ, Konikoff FM, Gilat T. Increased saturation of the fatty acids in the sn-2 position of phospholipids reduces cholesterol crystallization in model biles. Biochim Biophys Acta 1998; 1390: 293–300. Pope JL. Crystallization of sodium taurocholate. J Lipid Res 1967; 8: 146–7. Holan KR, Holzbach RT, Hermann RE, Cooperman AM, Claffey WJ. Nucleation time: a key factor in the pathogenesis of cholesterol gallstone disease. Gastroenterology 1979; 77: 611–7. So¨mjen GJ, Ringel Y, Konikoff FM, Rosenberg R, Gilat T. A new method for the rapid measurement of cholesterol crystallization in model biles using a spectrophotometric microplate reader. J Lipid Res 1997; 38: 165–9. So¨mjen GJ, Gilat T. A non-micellar mode of cholesterol transport in human bile. FEBS Lett 1983; 156: 265–8. Konikoff FM, Laufer H, Messer G, Gilat T. Monitoring cholesterol crystallization from lithogenic model bile by timelapse density gradient ultracentrifugation. J Hepatol 1997; 26: 703–10. Konikoff FM, Carey MC. Cholesterol crystallization from a dilute bile salt-rich model bile. J Crystal Growth 1994; 144: 79–86. Groen AK, Noordam C, Drapers JAG, Egbers P, Jansen PLM, Tytgat GNL. Isolation of potent cholesterol nucleating-promoting activity from human gallbladder bile: role in the pathogenesis of gallstone disease. Hepatology 1990; 11: 525–33. Halpern Z, Lafont H, Arad J, Domingo N, Peled Y, Konikoff FM, et al. The distribution of biliary-anionic polypeptide fraction between cholesterol carriers in bile and its effect on nucleation. J Hepatol 1994; 21: 979–83. Konikoff FM, Lechene de la Porte P, Laufer H, Domingo N, Lafont H, et al. Calcium and the anionic polypeptide fraction (APF) have opposing effects on cholesterol crystallization in model bile. J Hepatol 1997; 27: 707–15.
Copyright C European Association for the Study of the Liver 1998
Journal of Hepatology ISSN 0168-8278
The effects of phospholipid molecular species on cholesterol crystallization in model biles: the influence of phospholipid head groups Yehuda Ringel, Giora J. So¨mjen, Fred M. Konikoff, Ruth Rosenberg, Moshe Michowitz and Tuvia Gilat Department of Gastroenterology, Tel-Aviv Sourasky Medical Center, Ichilov Hospital, Tel-Aviv, and the Minerva Center for Cholesterol Gallstones and Lipid Metabolism in the Liver, Sackler Faculty of Medicine, Tel-Aviv University, Israel
Background/Aims: Variations in the molecular species of biliary phospholipids have been shown to exert major effects on cholesterol solubility and carriers in model and human biles. The aim of this study was to explore systematically the effects of various phospholipid head groups on the cholesterol crystallization process in model biles. Methods: Three different control model biles were prepared using varying proportions of egg lecithin, cholesterol and Na taurocholate. In the test biles, 20% of the egg lecithin was replaced with synthetic phosphatidylserine, phosphatidylethanolamine, phosphatidylglycerol or phosphatidylcholine, keeping the phospholipid acyl chains and other biliary lipids constant in each experiment. Results: Phosphatidylserine and phosphatidylglycerol significantly prolonged the crystal observation time, from 2 days to 10 and 6 days, respectively (p∞0.02), while phosphatidylethanolamine had little and phos-
phatidylcholine no effect. The crystal growth rate was significantly slowed down with 20% phospholipid replacement in the following order: phosphatidylglycerol ±phosphatidylserine ±phosphatidylethanolamine. The total crystal mass after 14 days, as measured by chemical analysis, was reduced by 59% with phosphatidylserine (p∞0.05), and by 73% with phosphatidylglycerol (p∞0.05); while phosphatidylethanolamine had little effect. The precipitable cholesterol crystal fractions after 14 days were significantly reduced with phosphatidylserine (54%) and phosphatidylglycerol (37%), but not with phosphatidylethanolamine or phosphatidylcholine. Conclusions: Variations in the head groups of biliary phospholipids may markedly slow down the cholesterol crystallization process in model biles.
and bile salts are two ampiphilic biliary lipid molecules, which are crucial in the secretion, solubilization and transport of biliary cholesterol. Both are present in bile as a mixture of several molecular species. Biliary research in recent decades has focused mainly on the various molecular species of bile salts. It was shown that minimal changes in bile salt molecules induced the secretion of undersaturated bile (1–4), supersaturated bile (5), affected intestinal cholesterol absorption (6), and caused liver toxicity in mammals (7). The importance of changes in the molecular forms
of biliary phospholipids emerged more recently. The addition of small amounts of specific phospholipids normalized the rapid nucleation time of lithogenic biles, while the addition of equimolar amounts of bile salts had no such effect (8). Increasing amounts of lecithin progressively prolonged the nucleation time of model and human biles (8,9). More importantly, changes in the head group or fatty acid chains of the phospholipid molecule resulted in major alterations in the nucleation time and cholesterol carriers in bile. These occurred without any change in the absolute or relative concentration of phospholipids in the model bile (9–12). As the results of bile salt therapy for gallstone dissolution or prevention of recurrence were shown to be suboptimal, it was thought worthwhile to explore the potential value of biliary phospholipid modulation. In a previous study we investigated the effects of varying
P
Received 9 July 1997; revised 12 January; accepted 15 January 1998
Correspondence: Fred Konikoff, Department of Gastroenterology, Ichilov Hospital, 6 Weizman St., Tel Aviv 64239, Israel. Tel: 972-3-697 4470. Fax: 972-3-697 4622. e-mail: konikoff/post.tau.ac.il
1008
Key words: Cholesterol crystallization; Model bile; Phospholipid head groups.
Phospholipid species and cholesterol crystallization
saturation of the sn-2 fatty acid of the major biliary phospholipid, phosphatidylcholine (12). In the present study we have investigated the effects of various phospholipid head groups while keeping the fatty acid composition constant.
Materials and Methods Cholesterol (Sigma, St. Louis, Mo, USA) was twice recrystallized from hot ethanol and Na taurocholate (Na TC) (Sigma, St. Louis, Mo, USA) was twice recrystallized from ethanol and ether (13). Egg phosphatidylcholine (EPC) and de novo synthesized phosphatidylethanolamine (PE), phosphatidylserine (PS) and phosphatidylglycerol (PG) (Avanti Polar Lipids, Alabaster, Ala, USA) were used without further purification. All lipids used in this study were pure by TLC standard. The radiolabeled lipids, [3H]-cholesterol, and Palmitoyl-2[1-14C] oleoyl-PC, were purchased from Amersham Laboratories (Buckinghamshire, UK). Preparation of model bile solutions Egg phosphatidylcholine (EPC), cholesterol and Nataurocholate (NaTC) mixtures were dissolved in CHCl3/CH3OH (2:1 v/v), dried under N2 at room temperature, lyophilized overnight and kept at ª20æC under argon until used. Model bile solutions were prepared by suspending the dried lipids in 150 mM NaCl, 1.5 mM disodium EDTA, 50 mM Tris-HCl pH 8.0 (NaTE) and incubating the suspension at 55æC for 1 h. The solubilized model biles were incubated in sealed vials under argon at 37æC for the duration of the experiments. Aliquots from the models were examined daily during the first 6 days and subsequently every 2 days. Types of bile models. Three different model biles were used in all experiments. The models were all composed to simulate human gallbladder bile. The three model biles (A, B, & C) were chosen to represent the range of variability occurring in native biles in terms
TABLE 1 Model solutions used Model CHOL. PL mM mM
TC mM
CSI %
Total lipids Crystallization g/dl characteristics
A
18.5
33.6
120
160
9.2
B
15
30
150
134
10.3
C
10
20
100
146
6.8
COT – 2.0 CGR – 0.21 COT – 1.3 CGR – 0.09 COT – 4.0 CGR – 0.09
CHOLΩcholesterol; PLΩphospholipids; TCΩtaurocholate; CSIΩ cholesterol saturation index; COTΩcrystal observation time [days]; CGRΩcrystal growth rate [OD/day].
of lipid composition and crystallization kinetics. The models displayed short or long COT and fast or slow CGR, as shown in Table 1. Types of phospholipids tested. Within each model (A, B & C), substitution experiments were performed. One hundred per cent egg lecithin was used for preparation of the control solutions. The other investigated bile solutions were prepared by substituting 20% of the egg lecithin by various synthetic phospholipids (Table 2). Two series of experiments were performed: in series 1 the fatty acids were 16:0 and 18:1, while in series 2 they were 16:0–18:2. Various head groups were tested in each series (Table 2). All models were prepared in triplicate and kept under the same conditions throughout the experimental period. Dose-dependency experiments were performed using model biles A&B. One hundred per cent egg lecithin was used for the control solutions which were compared to the test solutions prepared by substituting the egg lecithin with increasing proportions (1%, 2.5%, 5%, 10%, 20%) of the synthetic phospholipids. Evaluation of cholesterol crystal formation and growth Crystal observation time (COT) assay. Crystal observation time (also called ‘‘Nucleation time’’) was determined as described by Holan et al. (14). Aliquots (5 ml) from each model bile were examined by polarized light microscopy. The COT was determined as the time of first detection of at least three cholesterol monohydrate crystals per microscopic field at 100-fold magnification. Crystal growth rate (CGR) assay. Crystal growth was monitored spectrophotometrically using a SPEC-
TABLE 2 Alterations in phospholipid head group Series 1 (Fatty acids: 16:0–18:1) Phospholipid head Fatty acid chains group Sn-1 Sn-2 PE PS PG EPC (Control)
16:0 16:0 16:0 Mixed
18:1 18:1 18:1
Series 2 (Fatty acids: 16:0–18:2) Phospholipid head Fatty acid chains group Sn-1 Sn-2 PC PE PS EPC (Control)
16:0 16:0 16:0 Mixed
18:2 18:2 18:2
PCΩphosphatidylcholine; PEΩphosphatidylethanolamine; phosphatidylserine; EPCΩegg PC.
PSΩ
1009
Y. Ringel et al.
Fig. 1. A representative crystal growth curve (model A, Table 1). The OD of the solution (after dissolution of noncrystalline compounds) during 14 days of incubation. The squares represent the means of triplicate samples measured in duplicate. The dashed line gives an estimation of the CGR by linear regression analysis, and the arrow shows the OD difference between day 0 and day 14. The error bars represent the SEM (nΩ6).
TRA – microplate reader (SPECTRA – STL, Austria) as described by us previously (15). Aliquots (50 ml) of the lipid solutions were mixed and shaken vigorously with equivalent volumes of sodium taurodeoxycholate (NaTDC) (200 mM) in microplate wells. After 60 min at room temperature, the microplates were shaken again and the absorbance, at 405 nm, in each well was measured. Each model was prepared in triplicate and sampled in duplicate for measurement. The data were collected and analyzed by an IBM-compatible personal computer, and the optical density was calculated. A graph describing the averaged optical density changes for each solution was plotted. The slope in the steepest region of the curve was determined by a linear regression fit to at least three measurements and defined as the crystal growth rate (Fig. 1). Total crystal mass. Chemical analyses of cholesterol and phospholipids were performed on each sample on the last day of the experiment (day 14), as previously described (16). The samples were collected from the micro-wells, centrifuged in an Airfuge (Beckman) at 70 000 rpm for 5 min and the supernatant solutions were separated. Lipid determinations were performed on the total samples (before centrifugation) as well as on the supernatant solutions. The amount of cholesterol in the pellet was calculated by subtracting the amount in the supernatant solution from the total. The crystalline character of the pellet was confirmed by polarized light microscopy. Crystal mass was also estimated spectrophotometrically by measuring the difference in optical density between day 14 and day 0 (Fig. 1). Ultracentrifugation assay. Precipitable cholesterolcontaining lipid aggregates were harvested by sequential ultracentrifugation at predetermined time points
1010
(days 3 and 14) during the crystallization process, as previously described (12,17). In brief, the aggregates were first precipitated from a sample (0.5 ml) of the nucleating model bile by ultracentrifugation (200 000 g, 1 h, 20æC) in 0.15 M Nacl (4.5 ml). This procedure has been previously validated, and shown to separate multilamellar vesicles and all microcrystalline structures from the soluble phase of the model bile (17). The pellet was then placed on a sucrose gradient (0– 20%) in 5.5-ml nitrocellulose tubes, and centrifuged in a Centrikon T-1000 ultracentrifuge, using a TST 55.5 swinging bucket rotor (Kontron Instruments) at 200 000 g for 1 h at 20æC to separate the various cholesterol precipitates (17,18). Fractions (500 ml) were collected from the top and analyzed for lipid content by scintillation counting of the dually radiolabeled (3HCh/14C-PC) samples, and examined by phase contrast light microscopy. The density of the fractions was verified by a refractometer. Native bile Fresh gallbladder bile from a cholesterol gallstone patient was ultracentrifuged (200 000 g, 1 h, 25æC). The supernatant was collected and triplicate samples (0.5 ml) were used as controls. Additional samples of 0.5 ml were prepared by adding 6.72 mM synthetic 16:0– 18:2 phosphatidylserine in triplicate. The bile samples were kept in sealed vials under argon at 37æC throughout the experimental period. Aliquots (3 ml) were examined daily for determination of COT. Statistical analysis Each lipid dispersion was prepared in triplicate, and duplicate aliquots were measured from all samples. Mean values of optical density and standard errors were calculated. Crystal growth rates were calculated from linear regression analysis of the crystal growth curves as explained above. Analysis of variance was performed for the triplicates of each solution. The pvalue was calculated by Student’s t-test.
Results Series 1 (Fatty acids: 16:0–18:1; Phospholipids: PE, PS, PG) Crystal observation time. In model A the COT was minimally prolonged when 20% of the lecithin was replaced by PE. The COT was markedly prolonged with 20% PS and with 20% PG, to 10 days and 6 days respectively, vs 2 days in the 100% lecithin control solution. The findings were similar but less marked in models B and C in which the effect of PG was similar or greater than that of PS (Fig. 2). The differences in the COT between the test solutions and the lecithin
Phospholipid species and cholesterol crystallization
control solution were statistically significant (pΩ0.015) for PS in model A and for PS and PG in models B and C (pΩ0.02). Cholesterol crystal growth rate. The CGR in the 4 solutions of model A are shown in Fig. 3. Substitution of 20% of lecithin with PE decreased the CGR from 0.213 OD/day to 0.135 OD/day (pΩ0.035). This effect was much more marked with PS and PG, which decreased the CGR to 0.053 OD/day (pΩ0.007) and 0.035 OD/day (pΩ0.001), respectively. The results were similar, though less marked in models B and C (NS). Cholesterol crystal mass. The data were analyzed both by using the difference in OD between the first and the last day of the experiment (day 14) and by chemical analysis of the precipitated crystals on the
last day of the experiment. The results for model A are shown in Fig. 4. (The results were similar though less marked in models B&C.) It can be seen that substitution of 20% of the lecithin by PE produced only a minimal and nonsignificant decrease in the cholesterol crystal mass. However, with 20% PS and 20% PG the decreases were very marked and highly significant. Measuring the OD difference, the decrease was of 62.1% with PS and 79.4% with PG (pΩ0.002, pΩ0.001, respectively) compared to the lecithin control. By chemical analysis the decrease was of 59.2% with PS and 73.4% with PG (pΩ0.005, pΩ0.004 respectively) compared to the control solution. Precipitable cholesterol fractions. The fractions were precipitated and separated (see Methods) after 3 and 14 days of incubation. The findings for model A are shown in Fig. 5. On day 3 the differences were minimal, there were no crystals, and cholesterol was found only in supersaturated (cholesterol/phospholipid ratio ±1) multilamellar vesicles irrespective of the phospholipid molecular species in the solution. On day 14 mature crystals were seen in the high density fractions of the control lecithin solution, while the multilamellar vesicles (low density fractions) had become unsaturated (C/P ratio Æ1). The findings were similar with 20% PE, although the multilamellar vesicles remained saturated (C/P ratio ±1). However, with 20% PS there were fewer crystals (54% of control) and more cholesterol remained in the
Fig. 2. Mean crystal observation time (COT) as determined by polarized light microscopy for models A, B & C (see Table 1). The left bars represent the control (100% lecithin) solution. In the other bars 20% of the lecithin was replaced by synthetic phospholipids (PL) with different head groups as marked at the bottom of the bars (see Table 2) (nΩ3). *p∞0.02, compared to control. PEΩphosphatidylethanolamine; PCΩphosphatidylcholine; PSΩphosphatidylserine.
Fig. 3. Mean crystal growth rate (CGR) of model A as determined from the spectrophotometrically measured curves. The left bar represents the control with 100% lecithin. In the other bars 20% of the lecithin was replaced by synthetic PL, with different head groups as marked at the bottom of the bars (see Table 2) (nΩ3). *p∞0.04, **p∞0.001, compared to control. Abbreviations as in Fig. 2.
Fig. 4. Cholesterol crystal mass on day 14 (model A) was studied by two different methods. In panel a, the mean cholesterol crystal mass was estimated spectrophotometrically by the OD difference between days 14 and 0. In panel b, the mean amount of crystalline cholesterol was determined chemically, after centrifugation, on day 14 (see Methods). The left bars represent the controls with 100% lecithin. In the other bars 20% of the lecithin was replaced by synthetic PL with different head groups as marked at the bottom of the bars (see Table 2) (nΩ3). *p∞0.005, compared to control.
1011
Y. Ringel et al.
Fig. 5. Precipitable lipid structures (model A) were separated by density gradient ultracentrifugation from the four different test solutions. The density increases linearly from fraction 1 (1.01 g/ml) to fraction 10 (1.06 g/ml). The upper panels show the density profiles on day 3 and the lower panels on day 14 of incubation. (HΩCholesterol, (SΩphosphatidylcholine (PC)).
saturated multilamellar vesicles. The findings were similar with 20% PG, with even fewer crystals in the high density fractions (37% of control). Series 2 (Fatty acids: 16:0–18:2; Phospholipids: PC, PE, PS) Crystal observation time. The findings were similar to those found with the 16:0–18:1 PL, although the changes were much less marked. With 20% PS the COT was prolonged from 2.7 to 4 days (pΩ0.01). The findings in models B and C were similar with PS significantly prolonging the COT (pΩ0.024, pΩ0.01, respectively) compared to the control solutions. However, the substitution of 20% of lecithin with a synthetic 16:0– 18:2 PE or PC did not significantly change the COT (data not shown). Cholesterol crystal growth rate. In series 2 experiments, 20% PE did not affect the CGR, while 20% PS significantly reduced it (pΩ0.023). Again, 20% PC did not change the CGR as compared to the control lecithin solution (data not shown).
1012
Cholesterol crystal mass. Twenty percent PE reduced the crystal mass, by chemical analysis, only minimally (NS), whereas 20% of PS reduced the crystal mass by 39% (pΩ0.00005). Measurements of the OD difference produced results similar to those obtained by chemical analysis. In all 3 models, 20% of 16:0–18:2 PC had no effect on the crystal mass. Dose-dependency study There was a significant dose-dependent effect of PS on the COT and CGR. In model A the COT was markedly prolonged by 5%, 10% and 20% PS (5 days, 5.7 days and 6 days, respectively vs. 2.7 days in the 100% lecithin control solution; pΩ0.0003). The findings were similar, but less marked in model B, in which only 20% and 10% PS significantly prolonged the COT (2.3 days and 1.7 days, respectively, vs. 1 day in the 100% lecithin control solution; pΩ0.011). When lecithin was substituted with 20% and 10% PS the CGR, in model B, decreased from 0.159 OD/day to 0.07 OD/day and to 0.06 OD/day (pΩ0.0008), re-
Phospholipid species and cholesterol crystallization
spectively. This effect was similar, though less marked (NS) in model A. Native bile The lipid composition of the native bile was 13.4 mM cholesterol, 54.2 mM PL and 139.4 mM bile salts; hence, the added PS amounted to 11% of the total PL. In the samples with the added synthetic PS, the COT was markedly prolonged from 2 days to 7 days.
Discussion Our experiments demonstrate that replacing 20% of the lecithin by PS or PG markedly and significantly prolonged the COT, slowed the CGR and diminished the total crystal mass after 14 days of incubation. Under the same conditions PE produced only a minor effect, but with a similar trend. Changing the fatty acids from 16:0– 18:1 to 16:0–18:2 did not significantly modify the above results, and the effects of the head groups remained unchanged. Substituting 20% of the lecithin (a natural mixture of several molecular forms of PC) with a synthetic 16:0–18:2 PC did not have any noticeable effect. Thus, of the three head groups tested (PC, PE, PS), it is the last one that had the main anti-crystallizing effect. Unexpectedly, PG which is devoid of a head group produced a marked similar effect. In this study we measured the crystallization process of cholesterol, employing a combination of techniques. The crystallization-retarding effects of PS and PG were notable in terms of both crystallization kinetics (COT and CGR) and total crystal mass (OD difference, chemical analysis and ultracentrifugal separation). We were unable to find similar experiments in the published literature. Halpern et al. in 1993 (9) performed experiments in model bile solutions containing a single synthetic phospholipid (not natural lecithin as in the present experiments). When all the phospholipid in the model solution was PE, almost all the cholesterol eluted in the vesicular fraction. With 100% PS, almost all the cholesterol was found in the micellar fraction. With PC, cholesterol was found in both carriers. This effect of PS was also demonstrated in human bile. In all these experiments the phospholipid fatty acids, bile salts and cholesterol were kept constant. The nucleation time was also studied by Halpern et al., with all the phospholipid in a given model bile solution represented by a single synthetic phospholipid. The nucleation time was 5 days with PC, 6 days with PE and 20 days with PS, supporting the results found in the present study. However, the experiments of Halpern et al., were less physiologic since the phospholipids in human bile are never composed of a single molecular species. Moreover, Halpern et al. did not study the chol-
esterol crystal growth rate nor the final crystal mass. They showed a progressive prolongation of the nucleation time of human bile following the addition of exogenous lecithin. The dose-dependency study showed that the effect of PS was also evident at a concentration of 10% and even 5% of total phospholipids. PS in a proportion of 5% of total biliary phospholipids is not too far from the physiologic proportions. It is possible that even with crude methods of modulation such as oral supplementation, such levels could be achieved. This however remains to be demonstrated. Our single pilot experiment suggests that PS might be active in vivo in human bile. The crystallization kinetics as measured in model solutions in the present study cannot be performed in native bile. The pigmentation and turbidity of bile (due to non-lipid components, proteins, etc.) preclude these dynamic measurements of optical density. The COT is a very crude measurement that may mask considerable differences in crystallization kinetics. This was demonstrated by Groen’s group when studying the pronucleating effects of biliary glycoproteins (19) and by us when studying the antinucleating effects of APF. These were not demonstrable in human bile (20), but were clearly demonstrable in model solutions (21) using the present techniques. It is, however, very likely that the present findings also apply to human bile. It was previously shown that small amounts of phospholipids markedly prolonged the nucleation time of bile (8,9). It was also shown that this effect was much magnified when the molecular composition of the added phospholipid had strong antinucleating properties (8) (such as disaturated phospholipids). In the present pilot experiment, the COT of human bile was markedly prolonged by PS at a concentration of 11% of biliary phospholipids. The mechanism by which different molecular configurations of phospholipids affect cholesterol crystallization is at present speculative. It is possible that PS and PG exert their effect by virtue of their negative charge that is absent in PC, which has no net charge. The present study demonstrates the considerable effects of variations in the phospholipid head group on the cholesterol crystallization process in model bile. These data add to the growing body of evidence on the importance of biliary phospholipid molecular species for cholesterol homeostasis in bile. They raise the prospect of preventing or retarding cholesterol crystallization in human bile if and when significant modulation of biliary phospholipids becomes feasible.
Acknowledgements This study was supported in part by a grant from the Chief Scientist’s Bureau, The Israel Ministry of Health.
1013
Y. Ringel et al.
References 1. Danzinger RG, Hofmann AF, Schoenfield LJ, Thistle JL. Dissolution of cholesterol gallstones by chenodeoxycholic acid. N Engl J Med 1972; 286: 1–8. 2. Hofmann AF, Thistle JL, Klein PD, Szczepanik PA, Yu PYS. Chemotherapy for gallstone dissolution. II. Induced changes in bile composition and gallstone response. JAMA 1978; 239: 1138–44. 3. Ju¨ngst D, Brenner G, Pratschke E, Paumgartner G. Low-dose ursodeoxycholic acid prolongs cholesterol nucleation time in gallbladder bile of patients with cholesterol gallstones. J Hepatol 1989; 8: 1–6. 4. Loria P, Carulli N, Medici G, Menozzi D, Salvioli G, Bertolotti M, et al. Effect of ursocholic acid on bile lipid secretion and composition. Gastroenterology 1986; 90: 865–74. 5. Marcus SN, Heaton KW. Deoxycholic acid and the pathogenesis of gall stones. Gut 1988; 29: 522–33. 6. Hofmann AF. Pharmacology of ursodeoxycholic acid, an enterohepatic drug. Scand J Gastroenterol 1994; 29: 1–15. 7. Cohen BI, Hofmann AF, Mosbach EH, Stenger RJ, Rotschild MA, Hagey LR, et al. Differing effects of norursodeoxycholic or ursodeoxycholic acid on hepatic histology and bile acid metabolism in the rabbit. Gastroenterology 1986; 91: 189–97. 8. Ju¨ngst D, Lang T, Huber P, Lange V, Paumgartner G. Effect of phospholipids and bile acids on cholesterol nucleation time and vesicular/micellar cholesterol in gallbladder bile of patients with cholesterol stones. J Lipid Res 1993; 34: 1457– 64. 9. Halpern Z, Moshkowitz M, Laufer H, Peled Y, Gilat T. Effect of phospholipids and their molecular species on cholesterol solubility and nucleation in human and model biles. Gut 1993; 34: 110–5. 10. Tao S, Tazuma S, Kajiyama G. Fatty acid composition of lecithin is a key factor in bile metastability in supersaturated model bile systems. Biochim Biophys Acta 1993; 167: 142–6. 11. Tazuma S, Ochi H, Teramen K, Yamashita Y, Horikawa K, Miura H, et al. Degree of fatty acyl chain unsaturation in
1014
12.
13. 14.
15.
16. 17.
18.
19.
20.
21.
biliary lecithin dictates cholesterol nucleation and crystal growth. Biochim Biophys Acta 1994; 1215: 74–8. Ringel Y, So¨mjen GJ, Konikoff FM, Gilat T. Increased saturation of the fatty acids in the sn-2 position of phospholipids reduces cholesterol crystallization in model biles. Biochim Biophys Acta 1998; 1390: 293–300. Pope JL. Crystallization of sodium taurocholate. J Lipid Res 1967; 8: 146–7. Holan KR, Holzbach RT, Hermann RE, Cooperman AM, Claffey WJ. Nucleation time: a key factor in the pathogenesis of cholesterol gallstone disease. Gastroenterology 1979; 77: 611–7. So¨mjen GJ, Ringel Y, Konikoff FM, Rosenberg R, Gilat T. A new method for the rapid measurement of cholesterol crystallization in model biles using a spectrophotometric microplate reader. J Lipid Res 1997; 38: 165–9. So¨mjen GJ, Gilat T. A non-micellar mode of cholesterol transport in human bile. FEBS Lett 1983; 156: 265–8. Konikoff FM, Laufer H, Messer G, Gilat T. Monitoring cholesterol crystallization from lithogenic model bile by timelapse density gradient ultracentrifugation. J Hepatol 1997; 26: 703–10. Konikoff FM, Carey MC. Cholesterol crystallization from a dilute bile salt-rich model bile. J Crystal Growth 1994; 144: 79–86. Groen AK, Noordam C, Drapers JAG, Egbers P, Jansen PLM, Tytgat GNL. Isolation of potent cholesterol nucleating-promoting activity from human gallbladder bile: role in the pathogenesis of gallstone disease. Hepatology 1990; 11: 525–33. Halpern Z, Lafont H, Arad J, Domingo N, Peled Y, Konikoff FM, et al. The distribution of biliary-anionic polypeptide fraction between cholesterol carriers in bile and its effect on nucleation. J Hepatol 1994; 21: 979–83. Konikoff FM, Lechene de la Porte P, Laufer H, Domingo N, Lafont H, et al. Calcium and the anionic polypeptide fraction (APF) have opposing effects on cholesterol crystallization in model bile. J Hepatol 1997; 27: 707–15.