- Email: [email protected]
Dissolution of human cholesterol gallstones in simulated chenodeoxycholate-rich and ursodeoxycholate-rich biles
GASTROENTEROLOGY
1984:87:150-8
Dissolution of Human Cholesterol Gallstones in Simulated Chenodeoxycholate-Rich and Ursodeoxycholate-Rich Biles An In Vitro Study of Dissolution Mechanisms YONG-HYUN
PARK,
HIROTSUNE
IGIMI,
Rates and
and MARTIN
C. CAREY
Department of Medicine, Harvard Medical School, and Division of Gastroenterology, and Women’s Hospital, Boston, Massachusetts
We have compared the kinetics and physical-chemical mechanisms of human cholesterol gallstone dissolution in simulated normal, chenodeoxycholaterich, and ursodeoxycholate-rich “biles.” Owing to reduced micellar cholesterol solubilizing capacities, dissolution rates in ursodeoxycholate-rich biles were initially slower than in normal or chenodeoxycholate-rich biles. At later time points, dissolution rates in ursodeoxycholate-rich bile became accelerated; this was shown to be associated with the development of a JameJJar liquid-crystalline phase that took place first on the stones’ surfaces and was then followed by dispersion of liquid-crystalline vesicles into the micellar solution. As subsequent dissolution occurred in a two-phase system of micelles and liquid-crystalline vesicles, the quantity of cholesterol solubilized exceeded micellar saturation. In normal and chenodeoxycholate-rich biles, no phase changes Received August 17, 1982. Accepted March 19, 1984. Address requests for reprints to: Martin C. Carey, M.D., Division of Gastroenterology, Department of Medicine, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115. Dr. Y-H. Park’s present address is Department of General Surgery, Seoul National University Hospital, Seoul, South Korea. Dr. H. Igimi’s permanent address is First Department of Surgery, Fukuoka University School of Medicine, Fukuoka, Japan. This work was supported in part by a Research Grant in Digestive Diseases (AM 18559) and by a Research Career Development Award (AM 00195) to Dr. Carey from the National Institutes of Health. This study was presented in part at the Annual Meeting of the American Gastroenterological Association, New York, New York, and published in abstract form in GASTROENTEROLOGY
Brigham
were observed either on the surfaces of the stones or in the micellar solution, and the quantity of cholesterol solubilzed was limited by micellar saturation. These results are consistent with phase equilibria diagrams of the simulated bile systems and suggest that the predominant physical-chemical mechanism of in vivo gallstone dissolution with ursodeoxychoJic acid is via liquid crystalline dispersion of cholesterol. In contrast, micellar dissolution of cholesterol is the only mechanism possible with chenodeoxycholic acid. The clinical efficacies of optimal doses of chenodeoxycholic acid (CDCA) and ursodeoxycholic acid (UDCA) in dissolving cholesterol gallstones are conthe rates at sidered comparable (l-6); nevertheless, which human gallstones dissolve in bile enriched with conjugates of either agent have not been evaluated. Due to the recent emergence of concepts demonstrating that the physical-chemical mechanisms of gallstone dissolution in model and native biles enriched with CDCA and UDCA conjugates may differ (7-g), it has become important to obtain in vitro information on the dissolution of human gallstones with bile compositions that closely simulate those found in vivo. We have, therefore, studied the dissolution rates and mechanisms of human cholesterol gallstones in simulated “normal,” conjugated CDCArich (“cheno-rich”), and conjugated UDCA-rich (“urso-rich”) biles. We carried out dissolution under unstirred conditions to mimic the hydrodynamic
1981;80:1248.
The authors thank Dr. Tetsuo Uchida for performing the silver assays and Professor Bert L. Vallee for use of his atomic absorption spectrometer. 0 1984 by the American Gastroenterological Association 0016-5085/84/$3.00
Abbreviations used in this paper: C, cholate; CDC, chenodeoxycholate; CDCA, chenodeoxycholic acid; DC, deoxycholate; G/T, glycine to taurine; UDC, ursodeoxycholate; UDCA, ursodeoxycholit acid.
July
1984
GALLSTONE DISSOLUTION RATES AND MECHANISMS
situation in the gallbladder. We used matched gallstones from the same gallbladder to control for stone composition and, because dissolution rates depend critically on gallstone surface area (lo), we measured the surface areas of the stones by “silvering” and then normalized dissolution rates per square millimeter of stone surface (see reference 10). We demonstrated that micellar dissolution of cholesterol gallstones occurs initially in all model biles. Because the cholesterol solubilizing capacity of urso-rich micelles is considerably less than that of normal and cheno-rich micelles, initial micellar dissolution was slowest in urso-rich bile. After prolonged incubation, however, liquid-crystalline surface films, and eventually liquid-crystalline dispersion of cholesterol, became important dissolution mechanisms in urso-rich bile. Subsequently, the dissolution rates equaled or even exceeded those in normal or chenorich bile and leveled off at cholesterol contents well above the micellar phase limits.
Materials and Methods Materials Sodium salts of taurine and glycine conjugates of cholate (C), deoxycholate [DC), chenodeoxycholate (CDC), and ursodeoxycholate (UDC) were purchased from Calbiothem-Behring Corp., [San Diego, Calif.) and, after recrystallization. were found to be sa% pure by standard methods (11).Egg-yolk lecithin (grade 1) was obtained from Lipid Products [Surrey, U.K.) and was >99% pure as described (12). Cholesterol (Nu-Chek Preparations, Austin, Minn.) was of identical purity to that used in an earlier study (11). Cholesterol oxidase kits (Determiner TC5) for measuring cholesterol (13) were obtained from Kyowa Hakko Kogya Company [Tokyo, Japan). Reagents for the assay of lecithin and bile salts were the same as used elsewhere (12). Other reagents, chemicals, and buffers (generally American Chemical Society purity) were obtained from Fisher Scientific Company (Medford, Mass.). Gallstones were obtained at elective cholecystectomy from 6 patients with multiple stones, sterile biles, and with no history of bile acid ingestion. Written informed consent and institutional human subjects committee approval were obtained. Chemical analysis of several stones from each gallbladder revealed that they contained >a5% cholesterol by weight (10). The sole exception was one set of stones that contained
Methods Preparation of micellar solutions. Simulated bile salt compositions (details given below and in figure cap-
151
tions) were patterned according to typical gallbladder bile compositions reported for nongallstone individuals and patients chronically ingesting UDCA or CDCA (14-19). To prepare each solution, appropriate amounts of a bile salt mixture (measured gravimetrically) were dissolved in CHC&-MeOH (1: 1 by vol) and added to solutions of lecithin or lecithin plus cholesterol (each measured by dry weight) in the same solvent (12). The multicomponent mixtures were dried under a stream of Nz followed by dessication under reduced pressure. When constant weight was achieved, sufficient Tris-buffer (0.05 M) in 0.15 M NaCl was added to give 20 ml of IO-g/d1 mixed micellar solutions (12). The micellar solutions were buffered to pH 7.4-7.5 in four dissolution studies and to pH 8.0-9.0in one study. Dissolution studies. The dissolution methodology and silvering technique used in estimating the surface area of gallstones have been detailed elsewhere (10). In brief, each gallstone was skewered with stainless-steel surgical wire [size #2) and then suspended by a cross wire from the 20 ml of the top of a glass container that contained simulated bile solution (10).The apparatus was flushed with Nz, hermetically sealed, and then immersed in a water bath at 37°C and left unshaken for the duration of each experiment. At designated time intervals, the gallstone was removed from the dissolution apparatus, the simulated-bile solution was sonicated for exactly 10 s.and a BOO-LOO-~1aliquot was taken for cholesterol [and, when indicated, bile salt and lecithin) analysis. Immediately after sampling, the gallstone was gently reimmersed in the dissolution solution. This procedure was designed to disturb the unstirred hydrodynamic situation as little as possible (10). Upon termination of each experiment, the gallstone was removed, washed with distilled water, and dried under Nz. We then chemically coated each gallstone with atomic silver (10) and measured the dissolution rate of the silver coating in an unstirred thiosulfate-ferricyanide solution. Dissolved silver was assayed by atomic absorption spectrometry. We obtained an estimate to the nearest square millimeter of the surface of the gallstone through comparison of the silver dissolution rate from the gallstone with the silver dissolution rates of a series of silvered plastic spheres of known dimensions (10). Experimental design. Experiments were divided into five groups to allow for evaluation of the effects of bile salt composition, glycine to taurine (G/T) conjugated bile salt ratio, added calcium chloride, percent cholesterol saturation, and time on dissolution rates and mechanisms all at a fixed (-4: 1 molimol) bile salt/lecithin ratio with 10 g/d1 of total lipids at 37°C. Single stones from the same gallbladder were used in each matched experiment. In experiment 1, dissolution rates in normal, cheno-rich, and urso-rich biles were compared. No cholesterol was initially added to the mixed micellar solutions and the molar proportions of conjugated bile salts CDC, UDC, C, and DC in moles per 100 moles were 40: 3: 35:22 [normal), 80:7:10:3 (cheno-rich), and 30:40:20:10 (urso-rich). The G/T conjugated bile salt ratio was 3: 1. The cholesterol concentration in the dissolution medium was measured, without a change of solution, at 1-4 h [termed “early”) and 24-96 h [termed “late”). In experiment 2, dissolution rates
152
GASTROENTEROLOGY
PARK ET AL.
A
Vol.87.No. 1
I9 0 Cheno-Rich
A Normal ?? &so-Rich
Figure
1.
200
400 /
I
24
I
48
in cheno-rich, cheno-rich plus urso-rich, and urso-rich biles were compared. The mixed micellar solutions were prepared with cholesterol to give a calculated 70% saturaTo simplify these systems, conjugated C and DC tion (20). were omitted. Conjugated CDC/UDC ratios were 9O:lO (cheno-rich), 60:40 (cheno + urso-rich), and 40:60 (ursorich). In conformity with the literature, the G/T ratio was 6: 1 for CDC conjugates and 12: 1 for UDC conjugates (19). To simulate gallbladder filling and emptying, the cholesterol concentrations in the dissolution medium were measured over an initial period of 48 h, and were then repeated with the same stones in fresh dissolution solutions of composition identical to that of the initial solutions. In experiment 3, dissolution rates of stones were bile with lipid composition identistudied in cheno-rich cal to that in experiment 2 but in the presence of 4 mM CaClz and with two variations in the molar G/T conjugated bile salt ratio: 6: 1 for CDC plus 12:1 for UDC vs. 1: 1 for both bile salts. In this experiment, the mixed micellar solutions were initially 70% saturated with cholesterol
(20) and were replaced with fresh solutions of identical composition every 48 h. The cholesterol concentrations in the dissolution medium were measured during the first and third 48-h periods. In experiment 4, dissolution rates were studied in urso-rich and cheno-rich simulated biles (bile salt composition as in experiment 2) that were designed to be either 90% or 100% saturated with cholesterol (20). The dissolved cholesterol concentrations in these solutions were monitored continuously for 16 days but, on account of the prolonged dissolution times, no attempt was made to measure gallstone surface area. In experiment 5, four sets of cholesterol and one set of pigment stones were incubated in cheno-rich and ursorich biles for 1 mo; at the end of this period, the relative bile salt, lecithin, and cholesterol compositions of each well-mixed bile sample were measured. The initial solutions were 80% saturated with cholesterol (20) and contained a G/T conjugated bile salt ratio of 3.5 : 1 (pH 8-9). The molar bile salt proportions (moles per 100 moles) of
I
72
Experiment 1. Dissolution of cholesterol gallstones under unstirred conditions in cholesterolfree simulated bile. A. Initial dissolution rates. B. Dissolution rates after 24 h (20 ml of 10 g/d1 mixed micellar solution; pH 7.4-7.6; 0.15 M NaCl + 0.05 M Tris buffer; 37°C; bile salt to lecithin molar ratio, 4: 1;glycine to taurine conjugated bile salt ratio; 3:1, CDC:UDC:C:DC molar ratios, 80: 7: 10: 3 (cheno-rich); 40: 3:35:22 (normal); 30:40:20:10 (urso-rich).
I
96
CDC, UDC, C, and DC were 21: 50: 15: 14 (urso-rich).
80: 2: 9: 9 (cheno-rich)
and
Gross and microscopic observations. During each experiment, the solutions were observed both grossly and by direct and polarized light microscopy. At the end of each dissolution experiment, the stones were photographed and their grazed surfaces were examined microscopically at ~40 magnification to detect the presence of liquid crystals (9). Well-mixed samples of all biles from experiment 5, including control micellar solutions incubated with two pigment gallstones, were frozen in Freon 22 (Racon Inc., Wichita, Kan.) at -150°C. Freeze fractures were performed with a Balzers BAF 301 apparatus (Balzers, Hudson, N.H.) at -115’C in vacua (1 X lO_‘torr). After coating with platinum and carbon, the replicas were cleaned with a 50% (wtivol) chromic acid solution, mounted on copper grids, and observed with a Hitachi HU-12A electron microscope (Hitachi Corp., Tokyo, Japan) at x 20,000 magnification.
Results For consistency with our earlier work ($I--111, the results of each experiment are plotted as concentration of cholesterol solubilized as a function of incubation time. Accordingly, when cholesterol solubilities are linear (as occurred in all studies with short incubation times], the slopes of the curves depict the dissolution rates. Because the surface area of each gallstone was derived by the silvering method (lo), cholesterol dissolution rates were normalized to units of lo4 mg/dl * h - mm2 of stone surface. Henceforth, these normalized dissolution rates are given in square brackets for every stone in each dissolution experiment. Each set of curves in the figures represents the dissolution of two or three stones from the same gallbladder.
July
GALLSTONE
1984
I
5CQ A. 1st 48HOURS Elk o =Cheno-Rich A =ChenotUrso-Rich . =Urso-Rtch
4oo-
5
4oo B. 2nd48 HOURS -#--o
_____--_
0
_9_-----c-
300-
I
0
I
12
J
I
24
40
36
HOURS Figure
2.
Experiment 2. Dissolution of cholesterol gallstones under unstirred conditions in simulated biles that were initially 70% saturated with cholesterol. A. First 48 h. B. Second 48 h after immersion of the stones in fresh micellar solutions. Other conditions are given in legend to Figure 1, except CDC:UDC ratios which were 90: 10 (cheno-rich), 60:40 (cheno + urso-rich). and 40:60 (urso-rich); and glycine to taurine ratios which were 6: 1 for CDC and 12: 1 for UDC.
As shown in Figure 1 (experiment l),dissolution rates of three stones in each micellar solution were linear within the time of the observations. During the first 4 h, the dissolution rates decreased in the order cheno-rich [195] > normal (1591 > urso-rich bile [127]. During the second time period (24-96 h), the dissolution rates became slower and converged; however, the rank ordering was the same: cheno-rich [loo] > normal [89] > urso-rich bile [85]. In Figure 2
DISSOLUTION
RATES
AND
MECHANISMS
153
(experiment 2), micellar solutions that were initially 70% saturated with cholesterol resulted in a pronounced decrease in dissolution rates compared with experiment 1 (Figure 1). During the first 48 h, normalized dissolution rates decreased in the order cheno-rich [72] > cheno + urso-rich 1671 > urso-rich bile [SO]. During the second 48-h period (with fresh solutions but identical stones), the dissolution rates were appreciably less but the ordering of dissolution rates was reversed: urso-rich [39] :I urso + chenorich [36] > cheno-rich [21]. In Figure 3 (experiment 31, added 4 mM CaCl, appeared to retard dissolution rates during the first 48 h in cheno-rich bile compared with solutions of identical composition but without calcium (Figure 2). Moreover, solutions with high G/T conjugated bile salt ratios appeared to give somewhat faster dissolution rates than those with low (equimolar) G/T ratios 135 vs. 221. During the third 48-h period (Figure 3), overall dissolution rates were retarded slightly but the ordering for the high and low G/T conjugated bile salt ratios remained similar [25 vs. 121. In Figure 4 (experiment 4), dissolution rates over the entire 16-day period were extremely slow in cheno-rich bile that was initially 90% saturated with cholesterol. As demonstrated by the upper micellar cholesterol monohydrate saturation line, micellar saturation was not achieved (panel A). In contrast, dissolution rates, as inferred by tangents to the curves, were initially fast in urso-rich bile that was initially 90%) saturated with cholesterol. After -8 days, dissolution terminated at cholesterol concentrations that leveled off above the micellar cholesterol saturation line (loser dashed line, panel A]. No gallstone dissolution occurred in 100% saturated cheno-rich bile as shown by data points that did not deviate appreciably from the line of micellar cholesterol saturation (upper dashed line, panel B). In contrast, dissolution continued in urso-rich bile that was initially 100%
G/T Ratio Figure
3.
Experiment 3. Dissolution of cholesterol gallstones in cheno-rich bile, 70% saturated with cholesterol and containing 4 mM CaCl,. A and B represent dissolution periods in different micellar solutions of identical initial composition. Other conditions are given in legend to Figure 1, except CDC:UDC ratio which was 90: 10. High glycine to taurine (G/T) ratios were 6:l for CDC and 12:l for UDC; low G/T ratios were I:I for both bile salts.
A. 1st 48HOURS a
,
0 /““____------a _--.
B.
\
Low
3rd 48HouRs High
p------
0‘~_~____I:
-o_----- / . \ 24 HOURS
__----
_--0
. Low / 36
48
PARK ET AL.
GASTROENTEROLOGY
Micellor
ChM
Saturation P---
Vol.
87,
No.
I
showed no change from their initial compositions (not plotted). During the initial dissolution periods in all experiments, the solutions remained grossly and microscopically clear, i.e., they were apparently micellar. However, during the latter dissolution periods, stones bathed in the urso-rich systems but not in the cheno-rich systems became covered with a white
A. DLSSOLU T./ON/A’ CHE.0 - ff/Cff B/L If250L
o
0
.
.
MicelIar ChM
Saturationb____
Ti-----
.
350
300
.
&so-Rich .
I/--_~______________________ ________________ f --. Micellar ChM Saturation I I I I I I I I 4 6’ 0 10 12 14 0 2 DAYS
Figure 4.
Experiment 4. Dissolution of cholesterol gallstones in A. Cheno-rich and urso-rich biles that were initially 90% saturated with micellar cholesterol. B. Same as in A but with micellar solutions that were initially 100% saturated with micellar cholesterol. The dashed horizontal lines labeled micellar cholesterol monohydrate (ChM) saturation were derived from critical tables (20) with corrections for percent UDC conjugates. In each panel, the upper dashed horizontal line corresponds to cheno-rich bile and the lower line to urso-rich bile. Other conditions as in legend to Figure 1 with the exception that molar CDC:UDC ratios were 90: 10 (cheno-rich) and 49: 60 (urso-rich). Glycine to taurine ratio was 6: 1 for CDC, 12 : 1 for UDC.
6,
60
D/SSOLUT/ON /N URSO-R/CU B/L E
100
00
60
40
20
100
h!ZRCE/VTB/LE S4L T Figure 5.
saturated with (micellar) cholesterol. After 6 days of incubation, the amount of cholesterol solubilized leveled off above the line of micellar cholesterol saturation (lower dashed line, panel B). In each experiment with urso-rich bile, the final concentrations of solubilized cholesterol were virtually identical (-350-360 mg/dl). This was further confirmed in experiment 5 (Figure 5) where, after 1 mo of incubation with cholesterol stones, the relative lipid compositions of urso-rich bile (n = 4) gave a percent saturation index (20)of 143 * 4 SD, whereas the relative lipid compositions of cheno-rich bile (n = 4) gave a percent saturation index of 107 k 7 SD. Analysis of the biles incubated with pigment stones
60
loo
Experiment 5. Condensed phase diagrams of chenorich (A] and urso-rich (B) biles based on unpublished phase equilibria experiments of Carey MC and Ko G and Reference 9, according to the lipid compositions in legend to Figure 7. The open and closed circles display initial and final relative lipid compositions of four urso-rich and four cheno-rich biles, each incubated for 1 mo with separate cholesterol gallstones. The size of the symbols encompasses 21 SD for all three lipid components. The semiinterrupted lines [labeled dissolution path) that connect the initial lipid compositions to the cholesterol apex of each triangle represent the theoretical “path” of cholesterol dissolution. The initial cholesterol saturation index (20) in each case was 80%. In the case of cheno-rich bile, the final cholesterol saturation index was 197 2 7 SD and, in the case of urso-rich bile, was 143 t 4 SD (see Discussion for further details).
GALLSTONE DISSOLUTION RATES AND MECHANISMS
July 1984
Figure 6.
155
Gross appearances of two cholesterol gallstones after a 14-days incubation in urso-rich bile (left] and cheno-rich bile (right) (bile compositions listed in legend to Figure 4). The stone on the left, incubated in urso-rich bile, was covered with lamellar liquid crystals (see Results].
film (Figure 6, left). This was shown microscopically to be a lamellar liquid-crystalline phase (12). In dissolution experiment 5, freeze-fracture electron microscopy of the urso-rich biles revealed liquidcrystalline vesicles of varying sizes (Figure 7B). In contrast, cheno-rich biles, after a similar incubation period with cholesterol stones, showed a nonspecific granular pattern and no liquid-crystalline vesicles (Figure 7A). Further, neither urso-rich nor chenorich biles, after incubation with pigment stones, showed any evidence of liquid-crystalline liposomes (not displayed). Discussion In each set of experiments in this study, stones from the same patient’s gallbladder were compared. Not only do such stones have a similar chemical composition (lo), but in an earlier study we demonstrated that the normalized cholesterol dissolution rates (milligrams of cholesterol per deciliter per hour per square millimeter of stone surface) from gallstones in chenodeoxycholate solutions were virtually identical for all stones from the same patient’s gallbladder but differed slightly for stones obtained from the gallbladders of different patients. Hence, we believe that the experimental design used in the present study gives a valid assessment of dissolution rates and mechanisms when different dissolution solutions and gallstones are compared. Traditionally, gallstone dissolution in an unstirred micellar bile salt system is considered to be controlled by the kinetics of diffusion and convection of unsaturated micelles and by the properties of the stone-solution interface, often referred to as the interfacial resistance (11,211. With pure solutions of the common bile salts, interfacial resistance is rather
small (11) but increases markedly in the presence of lecithin (9). As shown in recent model studies with compressed cholesterol monohydrate disks, powder cholesterol slurries, and human gallstones, the kinetics of dissolution in pure conjugated UDC-lecithin systems (7-9) or urso-rich model bile (9) are much more complex; it has been established that in the presence of solid crystalline cholesterol and cholesterol stones, a phase change progressively takes place both on the stone’s surface and in the micellar solution; nevertheless, neither dissolution rates nor mechanisms of human gallstones with simulated human gallbladder biles have heretofore been measured (8,9). Our present experiments demonstrate that, with human gallstones and simulated urso-rich biles, phenomena similar to those observed in pure model systems of tauroursodeoxycholate or glycoursodeoxycholate plus lecithin take place. As noted earlier (7-9) using pure conjugated UDC-lecithin systems, micellar mechanisms predominate during the initial periods of dissolution. This is also shown in this work for simulated biles as exemplified by the fact that cholesterol stones dissolve initially in urso-rich bile at a slower rate than in cheno-rich, cheno plus urso-rich, or normal biles (Figures 1 and 2). These differences are a function of the maximal micellar solubilities of cholesterol that are distinctly lower in the urso-rich systems (9). Furthermore, the kinetics of micellar dissolution are faster with cheno-rich bile than with the corresponding normal compositions as noted previously (9,21). When gallstones or cholesterol crystals are bathed in a mixed micellar solution of the common bile salts and lecithin, some lecithin adsorbs to the cholesterol crystals on the stone’s surface but a micro- or macroscopic phase change does not take place (9,22). In
156
PARK ET AL
GASTROENTEROLOGY
Vol. 87. No
Figure 7. Freeze-fracture electron microscopy of cheno-rich bile (A] and urso-rich bile (B) after a l-mo incubation with cholesterol gallstones, each from the same gallbladder. Initial simulated bile mixtures were 80% saturated with cholesterol (see Figure 5). Bile-salt to lecithin ratio was 4: 1; pH 8-9; 0.15 M NaCl + 0.05 M Tris; total lipid concentration, 10 gidl; CDC:UDC:C:DC molar ratio was 80: 2: 9: 9 (cheno-rich bile) and 21: 50: 15 : 14 (urso-rich bile]; glycine to taurine conjugated bile salt ratio was 3.5 : 1 (see Results).
model urso-rich bile, this phenomenon is marked (9) and becomes progressive with time (8,9) and eventually a new macroscopic phase appears (7-9). As demonstrated in Figures 1 and 2, the slower dissolution rates during the later periods of incubation occur because of progressive enrichment of the micelles with cholesterol (9,11). However, at these time points it is apparent that dissolution rates in ursorich biles begin to “catch up” and approach, or even exceed, those in the other bile systems. This occurred irrespective of whether dissolution was continuous, i.e., in the same solution (Figure 1) or whether the dissolution solution [but not the stones) was replaced every 48 h (Figure 2). Cheno-rich bile did not exhibit any acceleration at later time periods (Figure 3) irrespective of the G/T conjugated bile salt ratios or the presence of CaC12. The most likely explanation for the “catch up” effect in urso-rich bile is the development of a liquidcrystalline phase that forms first at the stone-solution interface (Figure 6). It is probable that the dispersion of these liquid crystals into the bulk micellar solution as a separate liquid-crystalline phase (Figure 7) enhances solubility and accelerates cholesterol dissolution. As a consequence, choles-
terol contents in the urso-rich systems exceed equilibrium micellar solubility. An explanation of these phenomena is forthcoming from inspection of the condensed three-component diagrams in Figure 5 [(9) and Carey MC, Ko G, unpublished observations] on which the initial and final relative compositions of all the biles from experiment 5 are plotted. The theoretical dissolution path is an imaginary tie-line connecting the initial composition of the model bile solutions (marked “initial”) to the cholesterol apex of each triangle (representing cholesterol crystals in stones]. On account of the different phase relations above the micellar phase in cheno-rich vs. urso-rich bile (see Reference 9), this dissolution path transects a twophase region in cheno-rich bile where cholesterol monohydrate crystals and saturated micelles coexist as equilibrium phases-hence, dissolution terminates at micellar saturation. However, in the ursorich systems a liquid-crystalline phase also coexists as a precipitate phase and equilibrates along this path with solid cholesterol crystals; hence, dissolution continues above the micellar zone because of liquid-crystalline dispersion. For these reasons, the final relative lipid compositions of four cheno-rich
July 1984
biles (Figure 5) are not appreciably different from 100% micellar saturation, whereas in the urso-rich systems, the final mean cholesterol saturation (143%) greatly exceeds equilibrium cholesterol solubility in the micellar phase. We believe that these observations (Figures 4-7) decisively establish that nonmicellar mechanisms are important in dissolution of human gallstones with simulated urso-rich bile under hydrodynamic conditions that might exist in the human gallbladder. This interpretation critically depends on the validity of the micellar phase limits plotted in Figures 4 and 5. In our judgment, these are extremely precise because (a) calculated cholesterol contents of the micellar solutions from our critical tables (20) have been validated (9) against micellar cholesterol solubilities measured directly by dissolution in a variety of normal, cheno-rich, and urso-rich biles (9); (b) the limits of the micellar phase in Figure 5 and the phase relations above the micellar zone were derived from phase equilibria experiments carried out with each bile salt studied here, both individually and as mixtures (Carey MC, Ko G, unpublished observations); and (c) our freeze-fracture experiments (Figure 6) directly confirm that dissolution occurs in urso-rich bile in a two-phase (micellar plus liquid-crystalline) medium. We conclude, therefore, that in urso-rich bile, a liquid-crystalline mechanism is indeed responsible for accelerating the dissolution of cholesterol from stones, and for bringing the final cholesterol content of the systems above the equilibrium micellar phase boundary. From a clinical perspective, these studies allow for several tentative predictions: (a) under pathophysiological conditions in vivo, gallstone dissolution should be somewhat faster in urso-rich bile than in cheno-rich bile, especially in biles with high cholesterol saturations; (b) gallstone dissolution should occur in urso-rich bile even though the micellar phase may be saturated or supersaturated with cholesterol; and (c) a liquid-crystalline phase should be capable of detection in, and isolation from, gallbladder and duodenal biles of patients undergoing successful gallstone dissolution with UDCA. The first and second of these suggestions are consistent with recent clinical observations (23,24) and the third has been confirmed with urso-rich human gallbladder bile samples obtained from patients undergoing cholesterol gallstone dissolution (25). Nevertheless, caution should be exercised in extrapolating the present results to the clinical situation. During UDCA therapy, the phase relations of bile or the hardness of stones may not be conducive to the formation of liquid-crystalline vesicles. This may give rise to very slow dissolution into an unsaturated micellar phase or failure of dissolution if the
GALLSTONE DISSOLUTION RATES AND MECHANISMS
157
micellar phase is saturated. Furthermore, recent physical-chemical studies (26) on the hepatic biles of tauroursodeoxycholate-infused laboratory-dietfed prairie dogs suggest that a vesicle phase coexists with a saturated micellar phase in the secreted biles. If this also accompanies UDCA therapy in human will be successful gallstone patients, dissolution only when the vesicle phase is unsaturated with cholesterol.
References 1. Nakayama F. Oral cholelitholysis-cheno
versus urso. Japanese experience. Dig Dis Sci 1980;25:129-34. 2. Bouchier IAD. The medical treatment of gallstones. Annu Rev Med 1980;31:59-77. 3. Maton PN, Iser JH, Reuben A, Saxton HM, Murphy GM, Dowling RI-I. Outcome of chenodeoxycholic acid (CDCA) treatment in 125 patients with radiolucent gallstones. Factors influencing efficacy, withdrawal, symptoms and side effects and post-dissolution recurrence. Medicine (Baltimore) 1982;61:86-97. 4. Okumura M, Tanikawa K, Chuman Y, et al. Clinical studies on dissolution of gallstones using ursodeoxycholic acid. Gastroenterol Jpn 1977;12:469-75. acid in the 5 Bachrach WJ, Hofmann AF. Ursodeoxycholic treatment of cholesterol cholelithiasis. Part 1. Dig Dis Sci 1982:27:737-61. acid in the 6. Bachrach WJ, Hofmann AF. Ursodeoxycholic treatment of cholesterol cholelithiasis. Part 2. Dig Dis Sci 1982;27:833-56. 7. Corrigan 01, Su CC, Higuchi WI, Hofmann AF. Mesophase formation during cholesterol dissolution in ursodeoxycholate-lecithin solution: new mechanism for gallstone dissolution in humans. J Pharmacol Sci 1980;69:869-70. formation 8. Su CC, Park JY, Higuchi WI, et al. Mesophase during in vitro cholesterol gallstone dissolution: a specific effect of ursodeoxycholic acid. J Pharmacol Sci 1981;70:7135. 9. Salvioli G, Igimi H, Carey MC. Cholesterol gallstone dissolution in bile: dissolution kinetics of crystalline cholesterol monohydrate by conjugated chenodeoxycholate-lecithin and ursodeoxycholate-lecithin mixtures. Dissimilar phase equilibria and dissolution mechanisms. J Lipid Res 1983;24:70120. 10. Park Y-H, Igimi H, Carey MC. The “mirroring” of gallstones: description of a novel silvering method to determine the surface area of an irregular object. In vitro demonstration that multiple gallstones from the same gallbladder dissolve in unsaturated “bile” at the same rate. Gastroenterology 1982;83:1071-8. 11. Igimi H, Carey MC. Cholesterol gallstone dissolution in bile: dissolution kinetics of crystalline (anhydrate and monohydrate) cholesterol with chenodeoxycholate, ursodeoxycholate and their glycine and taurine conjugates. J Lipid Res 1981;22:254-70. 12. Carey MC, Small DM. Physical-chemistry of cholesterol solubility in bile. Relationship to gallstone formation and dissolution in man. J Clin Invest 1978;61:998-1026. 13 Fromm H, Amin P, Klein H, Kupke I. Use of a simple enzymatic assay for cholesterol analysis in human bile. J Lipid Res 1980;21:259-61. 14. Tokyo Cooperative Gallstone Study Group. Efficacy and indications of ursodeoxycholic acid treatment for dissolving gallstones. Gastroenterology 1980;78:542-8.
158 PARK ET AL.
15. Igimi H, Tamesue N, Ikejiri Y, Shimura H. Ursodeoxycholate-in vitro cholesterol solubility and changes in composition of human gallbladder bile after oral treatment. Life Sci 1977;21:1373-80. 16. Adler RD, Bennion LJ, Duane WC, Grundy SM. Effects of low dose CDCA feeding on biliary lipid metabolism. Gastroenterology 1975;68:326-34. 17. Ponz de Leon M, Carulli N, Loria P, Iori R, Zironi F. Cholesterol absorption during bile acid feeding. Effect of ursodeoxycholic acid administration. Gastroenterology 1980;78:214-9. 18. Salen G, Colalillo A, Verga D, Bagan E, Tint GS, Shefer S. Effect of high and low doses of ursodeoxycholic acid on gallstone dissolution in humans. Gastroenterology 1980;78:1412-8. 19. Stiehl A, Czygan P, Kommerell B, Weis HJ, Hollermtiller KH. Ursodeoxycholic acid versus chenodeoxycholic acid: comparison of their effects on bile acid and bile lipid composition in patients with cholesterol gallstones. Gastroenterology 1978;75:1016-20. 20. Carey MC. Critical tables for calculating the cholesterol saturation of native bile. J Lipid Res 1978;19:945-55.
GASTROENTEROLOGY
Vol. 87, No. 1
21. Kwan KH, Higuchi WI, Hofmann AF. Dissolution kinetics of cholesterol in simulated bile II: influence of simulated bile composition. J Pharmacol Sci 1978;67:1711-4. 22. Hoelgaard A, Frekjaer S. Adsorption of lecithin by cholesterol. J Pharmacol Sci 1980;69:413-5. 23. Salvioli G, Salati R, Lugli R, Zanni I. Medical treatment of biliary duct stones: effect of ursodeoxycholic acid administration. Gut 1983;24:609-14. 24. Fromm H, Roat JW, Gonzalez V, Sarva RP, Farivar S. Comparative efficacy and side effects of ursodeoxycholic and chenodeoxycholic acids in dissolving gallstones: a double-blind controlled study. Gastroenterology 1983;85:1257-64. 25. Igimi H, Asakawa S, Watanabe D, Shirmura H. Liquid crystal formation in ursodeoxycholic-rich human gallbladder bile. Gastroenterol Jpn 1983;18:93-7. 26. Leighton LS, Carey MC. Bile salt hydrophilicity determines whether the physical state of native hepatic bile is predominantly vesicular or micellar (abstr). Gastroenterology 1984;86:1157.
1984:87:150-8
Dissolution of Human Cholesterol Gallstones in Simulated Chenodeoxycholate-Rich and Ursodeoxycholate-Rich Biles An In Vitro Study of Dissolution Mechanisms YONG-HYUN
PARK,
HIROTSUNE
IGIMI,
Rates and
and MARTIN
C. CAREY
Department of Medicine, Harvard Medical School, and Division of Gastroenterology, and Women’s Hospital, Boston, Massachusetts
We have compared the kinetics and physical-chemical mechanisms of human cholesterol gallstone dissolution in simulated normal, chenodeoxycholaterich, and ursodeoxycholate-rich “biles.” Owing to reduced micellar cholesterol solubilizing capacities, dissolution rates in ursodeoxycholate-rich biles were initially slower than in normal or chenodeoxycholate-rich biles. At later time points, dissolution rates in ursodeoxycholate-rich bile became accelerated; this was shown to be associated with the development of a JameJJar liquid-crystalline phase that took place first on the stones’ surfaces and was then followed by dispersion of liquid-crystalline vesicles into the micellar solution. As subsequent dissolution occurred in a two-phase system of micelles and liquid-crystalline vesicles, the quantity of cholesterol solubilized exceeded micellar saturation. In normal and chenodeoxycholate-rich biles, no phase changes Received August 17, 1982. Accepted March 19, 1984. Address requests for reprints to: Martin C. Carey, M.D., Division of Gastroenterology, Department of Medicine, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115. Dr. Y-H. Park’s present address is Department of General Surgery, Seoul National University Hospital, Seoul, South Korea. Dr. H. Igimi’s permanent address is First Department of Surgery, Fukuoka University School of Medicine, Fukuoka, Japan. This work was supported in part by a Research Grant in Digestive Diseases (AM 18559) and by a Research Career Development Award (AM 00195) to Dr. Carey from the National Institutes of Health. This study was presented in part at the Annual Meeting of the American Gastroenterological Association, New York, New York, and published in abstract form in GASTROENTEROLOGY
Brigham
were observed either on the surfaces of the stones or in the micellar solution, and the quantity of cholesterol solubilzed was limited by micellar saturation. These results are consistent with phase equilibria diagrams of the simulated bile systems and suggest that the predominant physical-chemical mechanism of in vivo gallstone dissolution with ursodeoxychoJic acid is via liquid crystalline dispersion of cholesterol. In contrast, micellar dissolution of cholesterol is the only mechanism possible with chenodeoxycholic acid. The clinical efficacies of optimal doses of chenodeoxycholic acid (CDCA) and ursodeoxycholic acid (UDCA) in dissolving cholesterol gallstones are conthe rates at sidered comparable (l-6); nevertheless, which human gallstones dissolve in bile enriched with conjugates of either agent have not been evaluated. Due to the recent emergence of concepts demonstrating that the physical-chemical mechanisms of gallstone dissolution in model and native biles enriched with CDCA and UDCA conjugates may differ (7-g), it has become important to obtain in vitro information on the dissolution of human gallstones with bile compositions that closely simulate those found in vivo. We have, therefore, studied the dissolution rates and mechanisms of human cholesterol gallstones in simulated “normal,” conjugated CDCArich (“cheno-rich”), and conjugated UDCA-rich (“urso-rich”) biles. We carried out dissolution under unstirred conditions to mimic the hydrodynamic
1981;80:1248.
The authors thank Dr. Tetsuo Uchida for performing the silver assays and Professor Bert L. Vallee for use of his atomic absorption spectrometer. 0 1984 by the American Gastroenterological Association 0016-5085/84/$3.00
Abbreviations used in this paper: C, cholate; CDC, chenodeoxycholate; CDCA, chenodeoxycholic acid; DC, deoxycholate; G/T, glycine to taurine; UDC, ursodeoxycholate; UDCA, ursodeoxycholit acid.
July
1984
GALLSTONE DISSOLUTION RATES AND MECHANISMS
situation in the gallbladder. We used matched gallstones from the same gallbladder to control for stone composition and, because dissolution rates depend critically on gallstone surface area (lo), we measured the surface areas of the stones by “silvering” and then normalized dissolution rates per square millimeter of stone surface (see reference 10). We demonstrated that micellar dissolution of cholesterol gallstones occurs initially in all model biles. Because the cholesterol solubilizing capacity of urso-rich micelles is considerably less than that of normal and cheno-rich micelles, initial micellar dissolution was slowest in urso-rich bile. After prolonged incubation, however, liquid-crystalline surface films, and eventually liquid-crystalline dispersion of cholesterol, became important dissolution mechanisms in urso-rich bile. Subsequently, the dissolution rates equaled or even exceeded those in normal or chenorich bile and leveled off at cholesterol contents well above the micellar phase limits.
Materials and Methods Materials Sodium salts of taurine and glycine conjugates of cholate (C), deoxycholate [DC), chenodeoxycholate (CDC), and ursodeoxycholate (UDC) were purchased from Calbiothem-Behring Corp., [San Diego, Calif.) and, after recrystallization. were found to be sa% pure by standard methods (11).Egg-yolk lecithin (grade 1) was obtained from Lipid Products [Surrey, U.K.) and was >99% pure as described (12). Cholesterol (Nu-Chek Preparations, Austin, Minn.) was of identical purity to that used in an earlier study (11). Cholesterol oxidase kits (Determiner TC5) for measuring cholesterol (13) were obtained from Kyowa Hakko Kogya Company [Tokyo, Japan). Reagents for the assay of lecithin and bile salts were the same as used elsewhere (12). Other reagents, chemicals, and buffers (generally American Chemical Society purity) were obtained from Fisher Scientific Company (Medford, Mass.). Gallstones were obtained at elective cholecystectomy from 6 patients with multiple stones, sterile biles, and with no history of bile acid ingestion. Written informed consent and institutional human subjects committee approval were obtained. Chemical analysis of several stones from each gallbladder revealed that they contained >a5% cholesterol by weight (10). The sole exception was one set of stones that contained
Methods Preparation of micellar solutions. Simulated bile salt compositions (details given below and in figure cap-
151
tions) were patterned according to typical gallbladder bile compositions reported for nongallstone individuals and patients chronically ingesting UDCA or CDCA (14-19). To prepare each solution, appropriate amounts of a bile salt mixture (measured gravimetrically) were dissolved in CHC&-MeOH (1: 1 by vol) and added to solutions of lecithin or lecithin plus cholesterol (each measured by dry weight) in the same solvent (12). The multicomponent mixtures were dried under a stream of Nz followed by dessication under reduced pressure. When constant weight was achieved, sufficient Tris-buffer (0.05 M) in 0.15 M NaCl was added to give 20 ml of IO-g/d1 mixed micellar solutions (12). The micellar solutions were buffered to pH 7.4-7.5 in four dissolution studies and to pH 8.0-9.0in one study. Dissolution studies. The dissolution methodology and silvering technique used in estimating the surface area of gallstones have been detailed elsewhere (10). In brief, each gallstone was skewered with stainless-steel surgical wire [size #2) and then suspended by a cross wire from the 20 ml of the top of a glass container that contained simulated bile solution (10).The apparatus was flushed with Nz, hermetically sealed, and then immersed in a water bath at 37°C and left unshaken for the duration of each experiment. At designated time intervals, the gallstone was removed from the dissolution apparatus, the simulated-bile solution was sonicated for exactly 10 s.and a BOO-LOO-~1aliquot was taken for cholesterol [and, when indicated, bile salt and lecithin) analysis. Immediately after sampling, the gallstone was gently reimmersed in the dissolution solution. This procedure was designed to disturb the unstirred hydrodynamic situation as little as possible (10). Upon termination of each experiment, the gallstone was removed, washed with distilled water, and dried under Nz. We then chemically coated each gallstone with atomic silver (10) and measured the dissolution rate of the silver coating in an unstirred thiosulfate-ferricyanide solution. Dissolved silver was assayed by atomic absorption spectrometry. We obtained an estimate to the nearest square millimeter of the surface of the gallstone through comparison of the silver dissolution rate from the gallstone with the silver dissolution rates of a series of silvered plastic spheres of known dimensions (10). Experimental design. Experiments were divided into five groups to allow for evaluation of the effects of bile salt composition, glycine to taurine (G/T) conjugated bile salt ratio, added calcium chloride, percent cholesterol saturation, and time on dissolution rates and mechanisms all at a fixed (-4: 1 molimol) bile salt/lecithin ratio with 10 g/d1 of total lipids at 37°C. Single stones from the same gallbladder were used in each matched experiment. In experiment 1, dissolution rates in normal, cheno-rich, and urso-rich biles were compared. No cholesterol was initially added to the mixed micellar solutions and the molar proportions of conjugated bile salts CDC, UDC, C, and DC in moles per 100 moles were 40: 3: 35:22 [normal), 80:7:10:3 (cheno-rich), and 30:40:20:10 (urso-rich). The G/T conjugated bile salt ratio was 3: 1. The cholesterol concentration in the dissolution medium was measured, without a change of solution, at 1-4 h [termed “early”) and 24-96 h [termed “late”). In experiment 2, dissolution rates
152
GASTROENTEROLOGY
PARK ET AL.
A
Vol.87.No. 1
I9 0 Cheno-Rich
A Normal ?? &so-Rich
Figure
1.
200
400 /
I
24
I
48
in cheno-rich, cheno-rich plus urso-rich, and urso-rich biles were compared. The mixed micellar solutions were prepared with cholesterol to give a calculated 70% saturaTo simplify these systems, conjugated C and DC tion (20). were omitted. Conjugated CDC/UDC ratios were 9O:lO (cheno-rich), 60:40 (cheno + urso-rich), and 40:60 (ursorich). In conformity with the literature, the G/T ratio was 6: 1 for CDC conjugates and 12: 1 for UDC conjugates (19). To simulate gallbladder filling and emptying, the cholesterol concentrations in the dissolution medium were measured over an initial period of 48 h, and were then repeated with the same stones in fresh dissolution solutions of composition identical to that of the initial solutions. In experiment 3, dissolution rates of stones were bile with lipid composition identistudied in cheno-rich cal to that in experiment 2 but in the presence of 4 mM CaClz and with two variations in the molar G/T conjugated bile salt ratio: 6: 1 for CDC plus 12:1 for UDC vs. 1: 1 for both bile salts. In this experiment, the mixed micellar solutions were initially 70% saturated with cholesterol
(20) and were replaced with fresh solutions of identical composition every 48 h. The cholesterol concentrations in the dissolution medium were measured during the first and third 48-h periods. In experiment 4, dissolution rates were studied in urso-rich and cheno-rich simulated biles (bile salt composition as in experiment 2) that were designed to be either 90% or 100% saturated with cholesterol (20). The dissolved cholesterol concentrations in these solutions were monitored continuously for 16 days but, on account of the prolonged dissolution times, no attempt was made to measure gallstone surface area. In experiment 5, four sets of cholesterol and one set of pigment stones were incubated in cheno-rich and ursorich biles for 1 mo; at the end of this period, the relative bile salt, lecithin, and cholesterol compositions of each well-mixed bile sample were measured. The initial solutions were 80% saturated with cholesterol (20) and contained a G/T conjugated bile salt ratio of 3.5 : 1 (pH 8-9). The molar bile salt proportions (moles per 100 moles) of
I
72
Experiment 1. Dissolution of cholesterol gallstones under unstirred conditions in cholesterolfree simulated bile. A. Initial dissolution rates. B. Dissolution rates after 24 h (20 ml of 10 g/d1 mixed micellar solution; pH 7.4-7.6; 0.15 M NaCl + 0.05 M Tris buffer; 37°C; bile salt to lecithin molar ratio, 4: 1;glycine to taurine conjugated bile salt ratio; 3:1, CDC:UDC:C:DC molar ratios, 80: 7: 10: 3 (cheno-rich); 40: 3:35:22 (normal); 30:40:20:10 (urso-rich).
I
96
CDC, UDC, C, and DC were 21: 50: 15: 14 (urso-rich).
80: 2: 9: 9 (cheno-rich)
and
Gross and microscopic observations. During each experiment, the solutions were observed both grossly and by direct and polarized light microscopy. At the end of each dissolution experiment, the stones were photographed and their grazed surfaces were examined microscopically at ~40 magnification to detect the presence of liquid crystals (9). Well-mixed samples of all biles from experiment 5, including control micellar solutions incubated with two pigment gallstones, were frozen in Freon 22 (Racon Inc., Wichita, Kan.) at -150°C. Freeze fractures were performed with a Balzers BAF 301 apparatus (Balzers, Hudson, N.H.) at -115’C in vacua (1 X lO_‘torr). After coating with platinum and carbon, the replicas were cleaned with a 50% (wtivol) chromic acid solution, mounted on copper grids, and observed with a Hitachi HU-12A electron microscope (Hitachi Corp., Tokyo, Japan) at x 20,000 magnification.
Results For consistency with our earlier work ($I--111, the results of each experiment are plotted as concentration of cholesterol solubilized as a function of incubation time. Accordingly, when cholesterol solubilities are linear (as occurred in all studies with short incubation times], the slopes of the curves depict the dissolution rates. Because the surface area of each gallstone was derived by the silvering method (lo), cholesterol dissolution rates were normalized to units of lo4 mg/dl * h - mm2 of stone surface. Henceforth, these normalized dissolution rates are given in square brackets for every stone in each dissolution experiment. Each set of curves in the figures represents the dissolution of two or three stones from the same gallbladder.
July
GALLSTONE
1984
I
5CQ A. 1st 48HOURS Elk o =Cheno-Rich A =ChenotUrso-Rich . =Urso-Rtch
4oo-
5
4oo B. 2nd48 HOURS -#--o
_____--_
0
_9_-----c-
300-
I
0
I
12
J
I
24
40
36
HOURS Figure
2.
Experiment 2. Dissolution of cholesterol gallstones under unstirred conditions in simulated biles that were initially 70% saturated with cholesterol. A. First 48 h. B. Second 48 h after immersion of the stones in fresh micellar solutions. Other conditions are given in legend to Figure 1, except CDC:UDC ratios which were 90: 10 (cheno-rich), 60:40 (cheno + urso-rich). and 40:60 (urso-rich); and glycine to taurine ratios which were 6: 1 for CDC and 12: 1 for UDC.
As shown in Figure 1 (experiment l),dissolution rates of three stones in each micellar solution were linear within the time of the observations. During the first 4 h, the dissolution rates decreased in the order cheno-rich [195] > normal (1591 > urso-rich bile [127]. During the second time period (24-96 h), the dissolution rates became slower and converged; however, the rank ordering was the same: cheno-rich [loo] > normal [89] > urso-rich bile [85]. In Figure 2
DISSOLUTION
RATES
AND
MECHANISMS
153
(experiment 2), micellar solutions that were initially 70% saturated with cholesterol resulted in a pronounced decrease in dissolution rates compared with experiment 1 (Figure 1). During the first 48 h, normalized dissolution rates decreased in the order cheno-rich [72] > cheno + urso-rich 1671 > urso-rich bile [SO]. During the second 48-h period (with fresh solutions but identical stones), the dissolution rates were appreciably less but the ordering of dissolution rates was reversed: urso-rich [39] :I urso + chenorich [36] > cheno-rich [21]. In Figure 3 (experiment 31, added 4 mM CaCl, appeared to retard dissolution rates during the first 48 h in cheno-rich bile compared with solutions of identical composition but without calcium (Figure 2). Moreover, solutions with high G/T conjugated bile salt ratios appeared to give somewhat faster dissolution rates than those with low (equimolar) G/T ratios 135 vs. 221. During the third 48-h period (Figure 3), overall dissolution rates were retarded slightly but the ordering for the high and low G/T conjugated bile salt ratios remained similar [25 vs. 121. In Figure 4 (experiment 4), dissolution rates over the entire 16-day period were extremely slow in cheno-rich bile that was initially 90% saturated with cholesterol. As demonstrated by the upper micellar cholesterol monohydrate saturation line, micellar saturation was not achieved (panel A). In contrast, dissolution rates, as inferred by tangents to the curves, were initially fast in urso-rich bile that was initially 90%) saturated with cholesterol. After -8 days, dissolution terminated at cholesterol concentrations that leveled off above the micellar cholesterol saturation line (loser dashed line, panel A]. No gallstone dissolution occurred in 100% saturated cheno-rich bile as shown by data points that did not deviate appreciably from the line of micellar cholesterol saturation (upper dashed line, panel B). In contrast, dissolution continued in urso-rich bile that was initially 100%
G/T Ratio Figure
3.
Experiment 3. Dissolution of cholesterol gallstones in cheno-rich bile, 70% saturated with cholesterol and containing 4 mM CaCl,. A and B represent dissolution periods in different micellar solutions of identical initial composition. Other conditions are given in legend to Figure 1, except CDC:UDC ratio which was 90: 10. High glycine to taurine (G/T) ratios were 6:l for CDC and 12:l for UDC; low G/T ratios were I:I for both bile salts.
A. 1st 48HOURS a
,
0 /““____------a _--.
B.
\
Low
3rd 48HouRs High
p------
0‘~_~____I:
-o_----- / . \ 24 HOURS
__----
_--0
. Low / 36
48
PARK ET AL.
GASTROENTEROLOGY
Micellor
ChM
Saturation P---
Vol.
87,
No.
I
showed no change from their initial compositions (not plotted). During the initial dissolution periods in all experiments, the solutions remained grossly and microscopically clear, i.e., they were apparently micellar. However, during the latter dissolution periods, stones bathed in the urso-rich systems but not in the cheno-rich systems became covered with a white
A. DLSSOLU T./ON/A’ CHE.0 - ff/Cff B/L If250L
o
0
.
.
MicelIar ChM
Saturationb____
Ti-----
.
350
300
.
&so-Rich .
I/--_~______________________ ________________ f --. Micellar ChM Saturation I I I I I I I I 4 6’ 0 10 12 14 0 2 DAYS
Figure 4.
Experiment 4. Dissolution of cholesterol gallstones in A. Cheno-rich and urso-rich biles that were initially 90% saturated with micellar cholesterol. B. Same as in A but with micellar solutions that were initially 100% saturated with micellar cholesterol. The dashed horizontal lines labeled micellar cholesterol monohydrate (ChM) saturation were derived from critical tables (20) with corrections for percent UDC conjugates. In each panel, the upper dashed horizontal line corresponds to cheno-rich bile and the lower line to urso-rich bile. Other conditions as in legend to Figure 1 with the exception that molar CDC:UDC ratios were 90: 10 (cheno-rich) and 49: 60 (urso-rich). Glycine to taurine ratio was 6: 1 for CDC, 12 : 1 for UDC.
6,
60
D/SSOLUT/ON /N URSO-R/CU B/L E
100
00
60
40
20
100
h!ZRCE/VTB/LE S4L T Figure 5.
saturated with (micellar) cholesterol. After 6 days of incubation, the amount of cholesterol solubilized leveled off above the line of micellar cholesterol saturation (lower dashed line, panel B). In each experiment with urso-rich bile, the final concentrations of solubilized cholesterol were virtually identical (-350-360 mg/dl). This was further confirmed in experiment 5 (Figure 5) where, after 1 mo of incubation with cholesterol stones, the relative lipid compositions of urso-rich bile (n = 4) gave a percent saturation index (20)of 143 * 4 SD, whereas the relative lipid compositions of cheno-rich bile (n = 4) gave a percent saturation index of 107 k 7 SD. Analysis of the biles incubated with pigment stones
60
loo
Experiment 5. Condensed phase diagrams of chenorich (A] and urso-rich (B) biles based on unpublished phase equilibria experiments of Carey MC and Ko G and Reference 9, according to the lipid compositions in legend to Figure 7. The open and closed circles display initial and final relative lipid compositions of four urso-rich and four cheno-rich biles, each incubated for 1 mo with separate cholesterol gallstones. The size of the symbols encompasses 21 SD for all three lipid components. The semiinterrupted lines [labeled dissolution path) that connect the initial lipid compositions to the cholesterol apex of each triangle represent the theoretical “path” of cholesterol dissolution. The initial cholesterol saturation index (20) in each case was 80%. In the case of cheno-rich bile, the final cholesterol saturation index was 197 2 7 SD and, in the case of urso-rich bile, was 143 t 4 SD (see Discussion for further details).
GALLSTONE DISSOLUTION RATES AND MECHANISMS
July 1984
Figure 6.
155
Gross appearances of two cholesterol gallstones after a 14-days incubation in urso-rich bile (left] and cheno-rich bile (right) (bile compositions listed in legend to Figure 4). The stone on the left, incubated in urso-rich bile, was covered with lamellar liquid crystals (see Results].
film (Figure 6, left). This was shown microscopically to be a lamellar liquid-crystalline phase (12). In dissolution experiment 5, freeze-fracture electron microscopy of the urso-rich biles revealed liquidcrystalline vesicles of varying sizes (Figure 7B). In contrast, cheno-rich biles, after a similar incubation period with cholesterol stones, showed a nonspecific granular pattern and no liquid-crystalline vesicles (Figure 7A). Further, neither urso-rich nor chenorich biles, after incubation with pigment stones, showed any evidence of liquid-crystalline liposomes (not displayed). Discussion In each set of experiments in this study, stones from the same patient’s gallbladder were compared. Not only do such stones have a similar chemical composition (lo), but in an earlier study we demonstrated that the normalized cholesterol dissolution rates (milligrams of cholesterol per deciliter per hour per square millimeter of stone surface) from gallstones in chenodeoxycholate solutions were virtually identical for all stones from the same patient’s gallbladder but differed slightly for stones obtained from the gallbladders of different patients. Hence, we believe that the experimental design used in the present study gives a valid assessment of dissolution rates and mechanisms when different dissolution solutions and gallstones are compared. Traditionally, gallstone dissolution in an unstirred micellar bile salt system is considered to be controlled by the kinetics of diffusion and convection of unsaturated micelles and by the properties of the stone-solution interface, often referred to as the interfacial resistance (11,211. With pure solutions of the common bile salts, interfacial resistance is rather
small (11) but increases markedly in the presence of lecithin (9). As shown in recent model studies with compressed cholesterol monohydrate disks, powder cholesterol slurries, and human gallstones, the kinetics of dissolution in pure conjugated UDC-lecithin systems (7-9) or urso-rich model bile (9) are much more complex; it has been established that in the presence of solid crystalline cholesterol and cholesterol stones, a phase change progressively takes place both on the stone’s surface and in the micellar solution; nevertheless, neither dissolution rates nor mechanisms of human gallstones with simulated human gallbladder biles have heretofore been measured (8,9). Our present experiments demonstrate that, with human gallstones and simulated urso-rich biles, phenomena similar to those observed in pure model systems of tauroursodeoxycholate or glycoursodeoxycholate plus lecithin take place. As noted earlier (7-9) using pure conjugated UDC-lecithin systems, micellar mechanisms predominate during the initial periods of dissolution. This is also shown in this work for simulated biles as exemplified by the fact that cholesterol stones dissolve initially in urso-rich bile at a slower rate than in cheno-rich, cheno plus urso-rich, or normal biles (Figures 1 and 2). These differences are a function of the maximal micellar solubilities of cholesterol that are distinctly lower in the urso-rich systems (9). Furthermore, the kinetics of micellar dissolution are faster with cheno-rich bile than with the corresponding normal compositions as noted previously (9,21). When gallstones or cholesterol crystals are bathed in a mixed micellar solution of the common bile salts and lecithin, some lecithin adsorbs to the cholesterol crystals on the stone’s surface but a micro- or macroscopic phase change does not take place (9,22). In
156
PARK ET AL
GASTROENTEROLOGY
Vol. 87. No
Figure 7. Freeze-fracture electron microscopy of cheno-rich bile (A] and urso-rich bile (B) after a l-mo incubation with cholesterol gallstones, each from the same gallbladder. Initial simulated bile mixtures were 80% saturated with cholesterol (see Figure 5). Bile-salt to lecithin ratio was 4: 1; pH 8-9; 0.15 M NaCl + 0.05 M Tris; total lipid concentration, 10 gidl; CDC:UDC:C:DC molar ratio was 80: 2: 9: 9 (cheno-rich bile) and 21: 50: 15 : 14 (urso-rich bile]; glycine to taurine conjugated bile salt ratio was 3.5 : 1 (see Results).
model urso-rich bile, this phenomenon is marked (9) and becomes progressive with time (8,9) and eventually a new macroscopic phase appears (7-9). As demonstrated in Figures 1 and 2, the slower dissolution rates during the later periods of incubation occur because of progressive enrichment of the micelles with cholesterol (9,11). However, at these time points it is apparent that dissolution rates in ursorich biles begin to “catch up” and approach, or even exceed, those in the other bile systems. This occurred irrespective of whether dissolution was continuous, i.e., in the same solution (Figure 1) or whether the dissolution solution [but not the stones) was replaced every 48 h (Figure 2). Cheno-rich bile did not exhibit any acceleration at later time periods (Figure 3) irrespective of the G/T conjugated bile salt ratios or the presence of CaC12. The most likely explanation for the “catch up” effect in urso-rich bile is the development of a liquidcrystalline phase that forms first at the stone-solution interface (Figure 6). It is probable that the dispersion of these liquid crystals into the bulk micellar solution as a separate liquid-crystalline phase (Figure 7) enhances solubility and accelerates cholesterol dissolution. As a consequence, choles-
terol contents in the urso-rich systems exceed equilibrium micellar solubility. An explanation of these phenomena is forthcoming from inspection of the condensed three-component diagrams in Figure 5 [(9) and Carey MC, Ko G, unpublished observations] on which the initial and final relative compositions of all the biles from experiment 5 are plotted. The theoretical dissolution path is an imaginary tie-line connecting the initial composition of the model bile solutions (marked “initial”) to the cholesterol apex of each triangle (representing cholesterol crystals in stones]. On account of the different phase relations above the micellar phase in cheno-rich vs. urso-rich bile (see Reference 9), this dissolution path transects a twophase region in cheno-rich bile where cholesterol monohydrate crystals and saturated micelles coexist as equilibrium phases-hence, dissolution terminates at micellar saturation. However, in the ursorich systems a liquid-crystalline phase also coexists as a precipitate phase and equilibrates along this path with solid cholesterol crystals; hence, dissolution continues above the micellar zone because of liquid-crystalline dispersion. For these reasons, the final relative lipid compositions of four cheno-rich
July 1984
biles (Figure 5) are not appreciably different from 100% micellar saturation, whereas in the urso-rich systems, the final mean cholesterol saturation (143%) greatly exceeds equilibrium cholesterol solubility in the micellar phase. We believe that these observations (Figures 4-7) decisively establish that nonmicellar mechanisms are important in dissolution of human gallstones with simulated urso-rich bile under hydrodynamic conditions that might exist in the human gallbladder. This interpretation critically depends on the validity of the micellar phase limits plotted in Figures 4 and 5. In our judgment, these are extremely precise because (a) calculated cholesterol contents of the micellar solutions from our critical tables (20) have been validated (9) against micellar cholesterol solubilities measured directly by dissolution in a variety of normal, cheno-rich, and urso-rich biles (9); (b) the limits of the micellar phase in Figure 5 and the phase relations above the micellar zone were derived from phase equilibria experiments carried out with each bile salt studied here, both individually and as mixtures (Carey MC, Ko G, unpublished observations); and (c) our freeze-fracture experiments (Figure 6) directly confirm that dissolution occurs in urso-rich bile in a two-phase (micellar plus liquid-crystalline) medium. We conclude, therefore, that in urso-rich bile, a liquid-crystalline mechanism is indeed responsible for accelerating the dissolution of cholesterol from stones, and for bringing the final cholesterol content of the systems above the equilibrium micellar phase boundary. From a clinical perspective, these studies allow for several tentative predictions: (a) under pathophysiological conditions in vivo, gallstone dissolution should be somewhat faster in urso-rich bile than in cheno-rich bile, especially in biles with high cholesterol saturations; (b) gallstone dissolution should occur in urso-rich bile even though the micellar phase may be saturated or supersaturated with cholesterol; and (c) a liquid-crystalline phase should be capable of detection in, and isolation from, gallbladder and duodenal biles of patients undergoing successful gallstone dissolution with UDCA. The first and second of these suggestions are consistent with recent clinical observations (23,24) and the third has been confirmed with urso-rich human gallbladder bile samples obtained from patients undergoing cholesterol gallstone dissolution (25). Nevertheless, caution should be exercised in extrapolating the present results to the clinical situation. During UDCA therapy, the phase relations of bile or the hardness of stones may not be conducive to the formation of liquid-crystalline vesicles. This may give rise to very slow dissolution into an unsaturated micellar phase or failure of dissolution if the
GALLSTONE DISSOLUTION RATES AND MECHANISMS
157
micellar phase is saturated. Furthermore, recent physical-chemical studies (26) on the hepatic biles of tauroursodeoxycholate-infused laboratory-dietfed prairie dogs suggest that a vesicle phase coexists with a saturated micellar phase in the secreted biles. If this also accompanies UDCA therapy in human will be successful gallstone patients, dissolution only when the vesicle phase is unsaturated with cholesterol.
References 1. Nakayama F. Oral cholelitholysis-cheno
versus urso. Japanese experience. Dig Dis Sci 1980;25:129-34. 2. Bouchier IAD. The medical treatment of gallstones. Annu Rev Med 1980;31:59-77. 3. Maton PN, Iser JH, Reuben A, Saxton HM, Murphy GM, Dowling RI-I. Outcome of chenodeoxycholic acid (CDCA) treatment in 125 patients with radiolucent gallstones. Factors influencing efficacy, withdrawal, symptoms and side effects and post-dissolution recurrence. Medicine (Baltimore) 1982;61:86-97. 4. Okumura M, Tanikawa K, Chuman Y, et al. Clinical studies on dissolution of gallstones using ursodeoxycholic acid. Gastroenterol Jpn 1977;12:469-75. acid in the 5 Bachrach WJ, Hofmann AF. Ursodeoxycholic treatment of cholesterol cholelithiasis. Part 1. Dig Dis Sci 1982:27:737-61. acid in the 6. Bachrach WJ, Hofmann AF. Ursodeoxycholic treatment of cholesterol cholelithiasis. Part 2. Dig Dis Sci 1982;27:833-56. 7. Corrigan 01, Su CC, Higuchi WI, Hofmann AF. Mesophase formation during cholesterol dissolution in ursodeoxycholate-lecithin solution: new mechanism for gallstone dissolution in humans. J Pharmacol Sci 1980;69:869-70. formation 8. Su CC, Park JY, Higuchi WI, et al. Mesophase during in vitro cholesterol gallstone dissolution: a specific effect of ursodeoxycholic acid. J Pharmacol Sci 1981;70:7135. 9. Salvioli G, Igimi H, Carey MC. Cholesterol gallstone dissolution in bile: dissolution kinetics of crystalline cholesterol monohydrate by conjugated chenodeoxycholate-lecithin and ursodeoxycholate-lecithin mixtures. Dissimilar phase equilibria and dissolution mechanisms. J Lipid Res 1983;24:70120. 10. Park Y-H, Igimi H, Carey MC. The “mirroring” of gallstones: description of a novel silvering method to determine the surface area of an irregular object. In vitro demonstration that multiple gallstones from the same gallbladder dissolve in unsaturated “bile” at the same rate. Gastroenterology 1982;83:1071-8. 11. Igimi H, Carey MC. Cholesterol gallstone dissolution in bile: dissolution kinetics of crystalline (anhydrate and monohydrate) cholesterol with chenodeoxycholate, ursodeoxycholate and their glycine and taurine conjugates. J Lipid Res 1981;22:254-70. 12. Carey MC, Small DM. Physical-chemistry of cholesterol solubility in bile. Relationship to gallstone formation and dissolution in man. J Clin Invest 1978;61:998-1026. 13 Fromm H, Amin P, Klein H, Kupke I. Use of a simple enzymatic assay for cholesterol analysis in human bile. J Lipid Res 1980;21:259-61. 14. Tokyo Cooperative Gallstone Study Group. Efficacy and indications of ursodeoxycholic acid treatment for dissolving gallstones. Gastroenterology 1980;78:542-8.
158 PARK ET AL.
15. Igimi H, Tamesue N, Ikejiri Y, Shimura H. Ursodeoxycholate-in vitro cholesterol solubility and changes in composition of human gallbladder bile after oral treatment. Life Sci 1977;21:1373-80. 16. Adler RD, Bennion LJ, Duane WC, Grundy SM. Effects of low dose CDCA feeding on biliary lipid metabolism. Gastroenterology 1975;68:326-34. 17. Ponz de Leon M, Carulli N, Loria P, Iori R, Zironi F. Cholesterol absorption during bile acid feeding. Effect of ursodeoxycholic acid administration. Gastroenterology 1980;78:214-9. 18. Salen G, Colalillo A, Verga D, Bagan E, Tint GS, Shefer S. Effect of high and low doses of ursodeoxycholic acid on gallstone dissolution in humans. Gastroenterology 1980;78:1412-8. 19. Stiehl A, Czygan P, Kommerell B, Weis HJ, Hollermtiller KH. Ursodeoxycholic acid versus chenodeoxycholic acid: comparison of their effects on bile acid and bile lipid composition in patients with cholesterol gallstones. Gastroenterology 1978;75:1016-20. 20. Carey MC. Critical tables for calculating the cholesterol saturation of native bile. J Lipid Res 1978;19:945-55.
GASTROENTEROLOGY
Vol. 87, No. 1
21. Kwan KH, Higuchi WI, Hofmann AF. Dissolution kinetics of cholesterol in simulated bile II: influence of simulated bile composition. J Pharmacol Sci 1978;67:1711-4. 22. Hoelgaard A, Frekjaer S. Adsorption of lecithin by cholesterol. J Pharmacol Sci 1980;69:413-5. 23. Salvioli G, Salati R, Lugli R, Zanni I. Medical treatment of biliary duct stones: effect of ursodeoxycholic acid administration. Gut 1983;24:609-14. 24. Fromm H, Roat JW, Gonzalez V, Sarva RP, Farivar S. Comparative efficacy and side effects of ursodeoxycholic and chenodeoxycholic acids in dissolving gallstones: a double-blind controlled study. Gastroenterology 1983;85:1257-64. 25. Igimi H, Asakawa S, Watanabe D, Shirmura H. Liquid crystal formation in ursodeoxycholic-rich human gallbladder bile. Gastroenterol Jpn 1983;18:93-7. 26. Leighton LS, Carey MC. Bile salt hydrophilicity determines whether the physical state of native hepatic bile is predominantly vesicular or micellar (abstr). Gastroenterology 1984;86:1157.