Serum plant sterols and biliary cholesterol secretion in humans: studies with ursodeoxycholic acid

Ratios of cholestanol, campesterol, and sitosterol to cholesterol in serum are known to reﬂect cholesterol absorption efﬁciency. Here, a possible link between these ratios and biliary secretion rates of cholesterol was investi-gated. Biliary lipid secretion rates and serum sterols were determined in 13 patients with gallstones. Seven were treated with ursodeoxycholic acid (UDCA) (1,000 mg/d). Serum cholesterol and non-cholesterol sterols were also measured in a cross over study in 20 healthy volunteers, who received either placebo or UDCA (750 mg/d). Biliary cholesterol secretion was signiﬁcantly lower, whereas the non-cholesterol sterols and their ratio to cholesterol were higher in patients with gallstones treated with UDCA. A highly signiﬁcant negative linear correlation between the ratios of non-cholesterol sterols to cholesterol and biliary cholesterol secretion was observed. In volunteers, administration of UDCA for 4 weeks was followed by a signiﬁcant increase in non-cholesterol sterols and their ratios. Even 4 weeks after discontinuing UDCA administration, campesterol and sitosterol were still signiﬁcantly higher than pretreatment levels, which was also true for the campesterol-cholesterol ratio after 8 weeks. The results suggest that the ratios of cholestanol, campesterol, and sitosterol

Cholestanol and the two plant sterols campesterol and sitosterol are present in low concentrations in human blood (1,2), but their concentrations are 400-to 1,000-fold lower than cholesterol (1). The concentrations in serum are the results of synthesis (for cholestanol only), absorption efficiency (campesterol and sitosterol), lipoprotein metabolism, and rate of hepatic excretion into bile (3). They are absorbed from the intestine to a much lesser extent than cholesterol (4,5). Cholestanol is synthesized in the liver from cholesterol (6), and is almost absent from the diet. Whereas sitosterol is not metabolized to a significant extent in humans (7), a conversion of cholestanol to bile acids has been reported (6). The ratios of cholestanol, campesterol, and sitosterol to cholesterol in serum have been shown to correlate positively with the cholesterol absorption efficiency (2,8), and have therefore been regarded as markers for changes in the absorption of cholesterol (8)(9)(10).
Recently it has been reported that treatment with ursodeoxycholic acid (UDCA), a bile acid known to reduce the hepatic secretion of cholesterol (11)(12)(13)(14), increases serum concentrations of plant sterols and cholestanol in patients with primary biliary cirrhosis (15) and patients with radiolucent gallstones (16,17). Considering that cholesterol absorption is not affected or even reduced by UDCA (13,14,(18)(19)(20)(21), we thought that these results were not consistent with the observation that the ratio of plant sterols to cholesterol are only markers of cholesterol absorption efficiency. The purpose of the present study was to examine whether the ratios of non-cholesterol sterols to cholesterol in serum also reflect biliary secretion of cholesterol and for how long the non-cholesterol sterols remain elevated after discontinuing UDCA administration.
Thus, we conducted two studies. In study I we measured serum cholestanol and plant sterols concentrations and their ratio to cholesterol as well as biliary cholesterol secretion in patients with gallstones. Seven of these patients were treated with UDCA, and six were not. In study II, serum concentrations of the non-cholesterol sterols were measured before, during, and after UDCA administration in healthy volunteers in a randomized, placebo controlled cross-over study.

Study I
Biliary lipid secretion rates were measured in 13 patients with radiolucent gallstones. Seven patients were treated with UDCA (1,000 mg/day) for 4 weeks and six patients were not treated. Measurements of biliary lipid secretion were performed by the method of Grundy and Metzger (22), as described previously (12,23). Briefly, on the evening before the study the patients were admitted to the metabolic ward of the Department of Clinical Pharmacology and swallowed a triple-lumen tube. The next morning the tube was positioned by X-ray guidance in the duodenum with the two most proximal outlets opposite to the ampulla of Vater and the third 10 cm distally. A liquid formula diet (Nutrison, R. Braun, Melsungen, Germany) was infused constantly into the most proximal outlet (1.42 kcal/kg ϫ h). The liquid formula contained sitosterol (7 mg/100 ml), which was used as a marker to calculate biliary secretion rates. When liquid formula infusion was started, patients on therapy swallowed two capsules of UDCA (500 mg). After allowing 4 hours for gallbladder contraction and stabilization of hepatic lipid secretion, hourly samples were collected for the following 6 h from the second proximal and distal outlets as described previously (12,23). Fasting blood samples were obtained after an overnight fast and serum samples were stored at Ϫ 20 Њ C for analysis of cholesterol and non-cholesterol sterols. Characteristics of the patients are given in Table 1 .

Study II
Serum samples of 20 healthy, non-smoking volunteers (5 female, 15 male, aged 19-38) had been stored at -20 Њ C from a previous study, in which the effect of UDCA on splanchnic and systemic hemodynamics and gallbladder motility had been investigated (24). All individuals had been randomly assigned to either group I or group II. Group I had received a placebo first for 4 weeks followed by a washout period of 4 weeks, and thereafter UDCA (250 mg tid) for 4 weeks. Group II had received UDCA first for 4 weeks, and after the washout period, placebo for 4 weeks. Before and at the end of each treatment period, fasting blood samples had been obtained after an overnight fast. None of the subjects had a history of gastrointestinal diseases, diabetes, or renal impairment. None of the volunteers were taking any kind of drugs, including oral contraceptives. One volunteer from group I dropped out before administration of UDCA for personal reasons. Serum samples were analyzed for cholesterol and non-cholesterol sterols.
Both study protocols were in accordance with the ethical guidelines of the Declaration of Helsinki and approved by the ethical committee of the University of Bonn. Written informed consent was obtained from each subject.

Statistical analysis
Data are expressed as mean Ϯ SD to show variation within a group. The changes in sterol concentration within the groups treated with UDCA or placebo were analyzed with the Student's t -test for dependent samples. The subgroup analysis in patients with gallstones was performed with a Student's t -test for independent samples. P of less than 0.05 was considered statistically significant. Correlation analysis was calculated as Pearson's product moment using Fisher's test for significance. All other calculations were performed with the statistical software SPSS™ 9.0 package (SPSS Inc., Chicago, IL).

Study I
Biliary lipid secretion rates of cholesterol, phospholipids, and bile acids in patients with radiolucent gallstones with and without UDCA treatment are summarized in Table 2 . As expected, cholesterol output was significantly lower ( Ϫ 28%, P Ͻ 0.05) in patients treated with UDCA than in patients without treatment, whereas total bile acid and phospholipid secretions did not differ. The concentrations of cholesterol and the non-cholesterol sterols together with their ratio to cholesterol are given in Table 3 . In patients treated with UDCA, serum concentrations of cholesterol were 11% lower than in the other group, but this difference did not reach statistical significance. The levels of cholestanol, campesterol, and sitosterol were 57%, 40%, and 53% higher in patients treated with UDCA. This was only significant for sitosterol ( P Ͻ 0.05). However, the ratios of cholestanol, campesterol, and sitosterol to cholesterol were significantly higher than in patients not treated with UDCA (71%, 69%, and 76%). Combining the results of all patients, a strong negative linear correlation between the ratios of cholestanol, campesterol, and sitosterol to cholesterol in serum and biliary cholesterol secretion was observed ( Fig. 1 ). Even after biliary cholesterol secretions were adjusted for body weight, this correlation remained highly significant for all ratios ( P Ͻ 0.002 or less; data not shown).

Study II
In both groups of volunteers, administration of UDCA had no effect on total cholesterol concentrations in serum ( Table 4 ). In group I, all sterol concentrations also remained remarkably constant during the 8 weeks without UDCA acid (4 weeks placebo and 4 weeks washout period), indicating metabolic steady state conditions. Cholestanol increased significantly during UDCA treatment in both groups compared with pretreatment levels. The increase during treatment with UDCA in group I and II averaged 16% and 20%, respectively. In group II, cho- lestanol decreased to pretreatment levels 4 weeks after discontinuing UDCA and remained constant thereafter (Table 4). Four weeks of UDCA administration also significantly increased the ratio of cholestanol to cholesterol in group I and II by 13% and 24%, respectively ( Table 5 ). Campesterol and sitosterol showed an even more pronounced increase during UDCA treatment, and the concentration of campesterol rose by 42% in group I and 55% in group II. Sitosterol increased by 61% in both groups. Four weeks after UDCA was discontinued in group II, campesterol and sitosterol levels were still significantly higher than pretreatment levels (14% and 21%) ( Table 4); even after 8 weeks, the concentrations of campesterol and sitosterol were still 8% and 10% higher but not of statistical significance ( P ϭ 0.07 for campesterol and P ϭ 0.08 for sitosterol). Also, the ratios of campesterol to cholesterol and sitosterol to cholesterol increased significantly after 4 weeks of administration of UDCA by 38% and 67%, respectively (Table 5). Even after 8 weeks, the ratio of both plant sterols to cholesterol did not return to pretreatment values. The ratio of campesterol to cholesterol and sitosterol to cholesterol were still 8% ( P Ͻ 0.05) and 10% ( P ϭ 0.117) higher compared with pretreatment levels.
The ratios of noncholesterol sterols to cholesterol in the six patients with radiolucent gallstones were lower compared with the values in the healthy volunteers during the control period and even after UDCA administration. These results might indicate that patients have higher secretion rates of biliary cholesterol than controls.

DISCUSSION
The present study was carried out to evaluate the effect of UDCA administration on serum concentrations of cholestanol, campesterol, and sitosterol in patients with gallstones and normolipemic, healthy volunteers, and to elucidate a possible link between serum concentrations of these sterols and biliary secretion of cholesterol. After UDCA treatment, the concentrations of these sterols, as well as their ratios to cholesterol, increased significantly. The results are in agreement with those recently reported in patients with gallstones (16,17) and primary biliary cirrhosis (15). Miettinen et al. (15)(16)(17) found a negative relationship between the change of the ratio of campesterol to cholesterol in serum before and during UDCA therapy and the change in cholesterol saturation in gallbladder bile. From the results, the authors suggest that under these circumstances levels of plant sterols might be affected by their biliary elimination and therefore reflect changes in biliary cholesterol secretion. However, measurements of biliary cholesterol secretion were not performed. Therefore, a direct relationship between serum concentrations of these sterols and biliary cholesterol secretion could not be proven.
Administration of UDCA is known to reduce hepatic secretion of cholesterol into bile (11)(12)(13)(14). Some, but not all, studies have shown that UDCA also reduces intestinal cholesterol absorption efficiency (13,14,(18)(19)(20)(21), which one would also expect to cause a reduction of the ratio of plant sterols to cholesterol in serum. Therefore, the findings of the present study in patients with radiolucent gallstones and normolipemic volunteers cannot be related to reduced intestinal absorption of cholesterol and plant sterols by UDCA. The reduction in biliary cholesterol secre-  Values are mean Ϯ SD. Serum sterols and their ratios to cholesterol in serum in patients with radiolucent gallstones with and without treatment with ursodeoxycholic acid (UDCA 1000 mg/day). a Significantly different from control ( P Ͻ 0.05). tion has been attributed to the lower micellar capacity of UDCA to dissolve cholesterol in bile acid-lecithin mixtures (26,27). Indeed, as we have previously shown, UDCA has no influence on the mobilization of hepatic cholesterol into bile in patients with gallstones (12). Therefore, it is likely that UDCA also reduces biliary output of plant sterols and cholestanol. Since UDCA has no effects on lipoprotein metabolism (16,17,28), retained plant sterols in the liver probably lead to higher absolute serum concentrations and their ratios to cholesterol. This led us to consider whether the ratio of cholestanol and plant sterols to cholesterol might be markers of cholesterol secretion into bile. In support of this hypothesis, we showed a close and highly significant negative linear relationship between the ratio of the non-cholesterol sterols to cholesterol with biliary cholesterol secretion in patients with radiolucent gallstones (Fig. 1). This is in line with the results of Miettinen et al. (17). These authors confirmed that UDCA decreases biliary cholesterol secretion by 26%, from 16.4 to 12.1 mg/kg ϫ h Ϫ 1 . However, the ratio of campesterol and sitosterol to cholesterol in bile did not change, indicating that hepatic secretion of campesterol and sitosterol were also reduced, whereas the ratios in serum increased. From the present results, it might also be considered that UDCA affects the expression of the ATPbinding-cassette transporters ABCG5 and/or ABCG8. These sterol transporters have recently been identified in the rare inherited disease of phytosterolemia (29,30). In this disease, one of these two transporters is defective, leading to hyperabsorption and diminished biliary secretion of cholesterol and plant sterols. However, more specific investigations are necessary to prove this speculation. Furthermore, the ratios of noncholesterol sterols to cholesterol in the six patients with radiolucent gallstones were lower compared with the values in the healthy volunteers during the control period and even after UDCA administration. These results might indicate that patients with cholesterol gallstones have higher secretion rates of biliary cholesterol than controls.
It is notable that in healthy volunteers cholestanol had returned to pretreatment levels 4 weeks after UDCA administration was stopped, whereas campesterol and sito-  sterol were still significantly elevated at this time. The halflife of UDCA is short, and it is eliminated from the body in only 1 week (31,32). Thus, biliary cholesterol secretion returned to pretreatment rates. Therefore, the continuing higher ratio of campesterol and sitosterol to cholesterol must be due to other mechanisms. One explanation may be the half-life of these sterols. The terminal half-life of sitosterol after intravenous injection in five normal subjects was 15.8 Ϯ 2.4 (SD) days (1). This half-life could indeed explain that sitosterol and the ratio of sitosterol to cholesterol is still elevated 56 days after discontinuing UDCA (3.5 half-lives). Since the structure of campesterol (24methyl-⌬5-cholesten-3␤-ol) is more closely related to cholesterol than sitosterol (24-ethyl-⌬5-cholesten-3␤-ol), the half-life of campesterol would be between that of sitosterol and cholesterol. This would explain why the ratio of campesterol to cholesterol is still significantly higher at the end of the study. From the present study, the half-life of cholestanol (5␣-cholestan-3␤-ol) must be markedly shorter than that of the two plant sterols. Indeed, earlier studies by Salen and Grundy (33) revealed a terminal halflife of 7.0 Ϯ 2.8 (SD) days in five subjects. For this reason, cholestanol might be considered to be the better marker of changes in biliary cholesterol secretion as demonstrated in the present study. On the other hand, cholestanol is an endogenous product of cholesterol, and other mechanisms might influence its metabolism and serum concentration. From our results, it can be concluded that the ratios of the sterols to cholesterol are not always accurate markers for intestinal cholesterol absorption efficiency, but can under special conditions be indicators for changes of biliary cholesterol secretion rates. However, when non-cholesterol sterols are used as indicators for these purposes, it is essential that metabolic steady state conditions have been stable for at least several weeks.