Biliary Sterol Secretion Is Required for Functional In Vivo Reverse Cholesterol Transport in Mice

Transcription

1 GASTROENTEROLOGY 2011;140: Biliary Sterol Secretion Is Required for Functional In Vivo Reverse Cholesterol Transport in Mice NIELS NIJSTAD,* THOMAS GAUTIER,* FRANÇOIS BRIAND, DANIEL J. RADER, and UWE J. F. TIETGE* *Department of Pediatrics, Center for Liver, Digestive and Metabolic Diseases, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands; Institute for Translational Medicine and Therapeutics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania BACKGROUND & AIMS: High-density lipoproteins (HDLs) protect against atherosclerotic cardiovascular disease, mainly by promoting reverse cholesterol transport (RCT). Biliary sterol secretion supposedly represents the final step in RCT, but the relevance of this pathway has not been explored. We tested the dependency of RCT on functional biliary sterol secretion. METHODS: Macrophage-to-feces RCT was studied in mice with abolished (bile duct ligation) or decreased biliary sterol secretion (adenosine triphosphate binding cassette transporter B4 (Abcb4) / mice, with and without administration of a liver X receptor [LXR] agonist) after intraperitoneal injection of 3 H-cholesterol loaded primary macrophage foam cells from mice. Fecal tracer excretion and also fecal mass sterol excretion were measured. Metabolism and tissue uptake of HDL cholesteryl ester was assessed with HDL kinetic studies. RESULTS: Bile-duct ligation completely abolished RCT from 3 H-cholesterol loaded macrophages to feces (P.001). In Abcb4 / mice lacking biliary cholesterol secretion, RCT was decreased markedly; fecal 3 H-tracer excretion was almost absent within neutral sterols (P.001) and reduced within bile acids (P.05). LXR activation stimulated RCT in wild-type (5.5-fold; P.001) but not Abcb4 / mice, whereas mass fecal sterol excretion increased similarly in both models (P.05). Kinetic studies revealed minimal uptake of HDL cholesteryl ester by the intestine, which decreased on LXR activation (P.05). CONCLUSIONS: Functional RCT depends on biliary sterol secretion; there is no compensatory increase in RCT via bile acids. The stimulating effect of LXR agonists on RCT requires biliary cholesterol secretion. These results have implications for therapies against atherosclerotic cardiovascular disease targeting the RCT pathway. Keywords: Liver; Atherosclerosis; Lipid; Intestine; Bile Acids. Plasma levels of high-density lipoprotein (HDL) cholesterol are strongly inversely correlated with morbidity and mortality from atherosclerotic cardiovascular disease. 1,2 An established mechanism of how HDL confers cardiovascular disease protection is by mediating reverse cholesterol transport (RCT). 3,4 Classically, RCT comprises the shuttling of excess cholesterol accumulating within peripheral cells, most importantly in the setting of atherosclerosis macrophage foam cells within the vessel wall, back to the liver for excretion into the feces via the bile. 2 5 Recently, a novel transport route for cholesterol has been reported, challenging this classic view of RCT by showing that the intestine contributes significantly to mass fecal neutral sterol excretion in a liver X receptor (LXR)-dependent fashion independent of biliary cholesterol secretion. 6,7 Subsequently, it has been proposed that intestinal cholesterol secretion might represent a novel therapeutic target to increase RCT and thereby confer protection against atherosclerotic cardiovascular disease. 7,8 However, the relevance of direct intestinal cholesterol secretion for RCT has not been explored thus far. Therefore, the present study was designed to test the significance of the biliary secretion pathway and the respective contribution of the intestine to RCT. We used HDL kinetic studies as well as an established in vivo RCT assay tracing the transport of labeled cholesterol from macrophage foam cells to feces Materials and Methods Experimental Animals and Treatments C57BL/6J wild-type mice were purchased from Charles River (Sulzfeld, Germany). Bile duct ligation (BDL) was performed by tying off the common bile duct proximal to the gallbladder branch after a midline incision. Sham-operated mice, in which the identical surgical procedures except the actual ligation were performed, served as controls for these sets of experiments. The adenosine triphosphate binding cassette transporter B4 (Abcb4) knockout mice 12 were on a FVB genetic background and FVB mice (Harlan, Horst, The Netherlands) were used as the respective controls. Scavenger receptor Abbreviations used in this paper: abcb4, adenosine triphosphate binding cassette transporter B4; BDL, bile duct ligation; FCR, fractional catabolic rate; HDL, high-density lipoproteins; LXR, liver X receptor; RCT, reverse cholesterol transport; SR-BI, scavenger receptor class B type I by the AGA Institute /$36.00 doi: /j.gastro

2 1044 NIJSTAD ET AL GASTROENTEROLOGY Vol. 140, No. 3 class B type I (SR-BI) knockout mice were obtained from Jackson (Bar Harbor, ME) and back-crossed to the C57BL/6J genetic background for a total of 8 generations. The LXR agonist T (Cayman Chemicals, Ann Arbor, MI) was mixed into the food (0.015% wt/wt) and provided for 8 days before and then throughout the whole 48-hour period of the kinetic and the RCT experiments. The animals were kept in animal rooms with alternating 12-hour periods of light (from 7:00 AM to 7:00 PM) and dark (from 7:00 PM to 7:00 AM), with ad libitum access to water and mouse chow diet (Arie Blok, Woerden, The Netherlands). Animal experiments were performed in accordance with the national laws. All protocols were approved by the responsible ethics committees of the University of Groningen or by the Institutional Animal Care and Use Committee at the University of Pennsylvania. HDL Kinetic Studies HDL was isolated by sequential ultracentrifugation (density, d 1.21) from pooled plasma from healthy human donors and was labeled with cholesteryl hexadecyl ether (CE, cholesteryl-1,2-3 H; PerkinElmer, Waltham, MA) essentially as described. 13 Cholesteryl ether behaves kinetically exactly as cholesteryl ester, however, the ether bond cannot be cleaved after cellular uptake, thereby enabling determination of tissue uptake of the tracer over the time course of the experiment. One million disintegrations per minute (dpm) of the 3 H-CE- HDL were injected via the tail vein. Kinetic experiments were performed over a period of 48 hours, taking plasma samples at 5 minutes, 30 minutes, 1 hour, 3 hours, 8 hours, 12 hours, 24 hours, and 48 hours, as described 14,15 and plasma decay curves were generated by dividing the plasma radioactivity at each time point by the radioactivity at the initial 5-minute time point after tracer injection. Fractional catabolic rates (FCRs) were determined from the area under the plasma disappearance curves fitted to a bicompartmental model by use of the SAAM II program (University of Washington, Seattle, WA). 14,15 Organ uptake of HDL 3 H-CE was determined 48 hours after tracer injection (see later) and expressed as a percentage of the injected dose calculated by multiplying the initial plasma counts (5-minute time point) with the estimated plasma volume (3.5% of total body weight). In Vivo RCT Experiments Thioglycollate-elicited macrophages were isolated from either C57BL/6J or FVB mice and allowed for 5 hours to adhere in RPMI medium (Invitrogen, Breda, The Netherlands) supplemented with 10% fetal bovine serum and antibiotics (Invitrogen). Then cells were washed twice with phosphate-buffered saline and loaded using RPMI supplemented with 2% bovine serum albumin (Sigma, St. Louis, MO), 3 H-cholesterol (1 Ci/mL, NEN Life Sciences Products), and 50 g acetylated low-density lipoprotein for 22 hours. Cells were washed again and equilibrated in RPMI with 0.2% bovine serum albumin for 18 hours. At the end of the equilibration, macrophages were harvested carefully from the plate and injected intraperitoneally into the respective groups of mice at a dose of 2 10E6 cells/mouse. The intraperitoneal route of administration was chosen because differentiated macrophages do not gain access to the vessel wall and the peritoneal cavity represents an extravascular compartment comparable with the vascular wall. Blood samples were drawn at the indicated time points and tracer within plasma was determined by liquid scintillation counting (Beckman LS6500; Beckman Instruments, Palo Alto, CA). Mice were killed after 24 hours in case of the BDL experiment and otherwise after 48 hours; organs were harvested and the total feces produced over the whole duration of the experiment were collected. 9,11 Of note, in contrast to the cholesteryl ether tracer used in the HDL kinetic studies, using free 3 H-cholesterol as a tracer in macrophage loading does not allow drawing conclusions on tissue uptake rates at the end of an experiment because this tracer is freely exchangeable, being subject to storage as well as re-secretion and distribution throughout the body. Radioactivity in gallbladder bile was assessed after lipid extraction and separation into cholesterol and bile acids according to the general method described by Bligh and Dyer, 16 counts for 3 H-cholesterol taken up by the liver were determined by incubating a piece of liver with Solvable (Packard, Meriden, CT) according to the manufacturer s instructions to dissolve the tissue as previously published. 14 Counts recovered from a respective piece of liver were back-calculated to total liver mass and expressed as a percentage of injected dose per whole organ. Similarly, a proximal, middle, and distal piece of small intestine was dissolved from each mouse, counts were determined, and related to the whole small intestine. Feces were thoroughly dried, ground, and aliquots were separated into the bile acid and the neutral sterol fractions as previously published. 17 Briefly, samples first were heated for 2 hours at 80 C in alkaline methanol and then extracted 3 times with petroleum ether. In the top layer, counts within the neutral sterol fraction were determined, whereas counts incorporated into bile acids were assessed from the bottom layer. Counts recovered from the respective aliquots were related to the total amount of feces produced within the whole experimental period and expressed as a percentage of the injected radiotracer dose. Plasma Lipid and Lipoprotein Analysis Mice were bled either from the retroorbital plexus using heparinized capillary tubes or by heart puncture at the time of death. Pooled plasma samples were subjected to fast protein liquid chromatography gel filtration using a superose 6 column (GE Healthcare, Uppsala, Sweden) as described. 18 In individual fractions cholesterol concen-

3 March 2011 BILIARY STEROL SECRETION DRIVES RCT 1045 Figure 1. BDL almost completely abolishes macrophage RCT in vivo. After BDL (n 5) or sham surgery (Con, n 6) mice received intraperitoneal injections of 3 H-cholesterol labeled macrophage foam cells. Blood samples were taken at 6 and 24 hours after injection and feces were collected continuously for 24 hours. (A) Time course of 3 H-cholesterol recovery within plasma. (B) Cholesterol (left panel) and 3 H-cholesterol (right panel) distribution over the different lipoprotein subclasses determined by fast protein liquid chromatography analysis of pooled plasma samples 24 hours after injection of 3 H-cholesterol loaded macrophages. (C) 3 H-tracer within liver and (D) 3 H-tracer content within fecal neutral sterols (NS, left panel) and bile acids (BA, right panel) of the respective groups of mice 24 hours after injection of 3 H-cholesterol loaded macrophages. The 3 H-cholesterol data are expressed as the percentage of tracer relative to the total cpm injected. All data are given as means standard error of the mean. ***P.001. VLDL, very low density lipoprotein; LDL, low-density lipoprotein. trations were determined with commercially available reagents (Wako Pure Chemical Industries, Neuss, Germany), and counts were assessed by liquid scintillation counting. Determination of Fecal Neutral Sterol and Bile Acid Mass as Well as Biliary Bile Acid and Cholesterol Contents Fecal samples were lyophilized and weighed. Aliquots thereof were used for determination of neutral and acidic sterol content by gas liquid chromatography as described. 17 Within gallbladder bile, bile salt and cholesterol concentrations were determined as described previously. 16 Quantitative Real-Time Polymerase Chain Reaction Analysis Total RNA was extracted with TriReagent (Sigma) and quantified using a UV-Vis Spectrophotometer (NanoDrop Technologies, Wilmington, DE). Complementary DNA synthesis was performed from 1 g of total RNA using reagents from Applied Biosystems (Darmstadt, Germany). Real-time quantitative polymerase chain reaction was performed on an ABI-Prism 7700 (Applied Biosystems) sequence detector with the default settings as published. 16 Primers and probes were synthesized by Eurogentec (Seraing, Belgium). Messenger RNA expression levels presented were calculated relative to the average of the housekeeping gene cyclophilin and further normalized to the relative expression level of the respective controls. Statistics Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS; SPSS Inc, Chicago, IL). Data are presented as means standard error of the mean. Statistical differences between 2 groups were assessed using the Student t test. To compare more than 2 groups, analysis of variance followed by a Bonferroni post test was used. Statistical significance for all comparisons was assigned at a P value of less than.05. Results BDL Markedly Decreases Fecal Excretion of Macrophage-Derived Cholesterol To test the importance of biliary secretion for functional reverse cholesterol transport from macrophages to feces, we first assessed the impact of completely blocking biliary secretion using BDL. As shown in Figure 1A, macrophage-derived 3 H-cholesterol in plasma was 2.5-fold higher at the 6-hour time point in BDL mice compared with sham controls (1.1% 0.2% vs 2.8% 0.3%; P.001), and even further increased by 4.4-fold at

4 1046 NIJSTAD ET AL GASTROENTEROLOGY Vol. 140, No. 3 the 24-hour time point (1.8% 0.3% vs 7.8% 1.1%; P.001). Fast protein liquid chromatography analysis revealed that the severe cholestasis developing in response to BDL resulted in a decreased HDL cholesterol peak (Figure 1B) and the appearance of a large peak within the very low density lipoprotein cholesterol range, conceivably representing Lipoprotein X. The distribution of counts within plasma over the different lipoprotein peaks largely followed the cholesterol distribution profile (Figure 1B). 3 H-tracer within liver at 48 hours was not different between both experimental groups (1.9% 0.6% vs 2.2% 0.7%; P NS; Figure 1C). However, the excretion of the 3 H-tracer into feces either as cholesterol (1.74% 0.49% vs 0.06% 0.02%; P.001; Figure 1D) or within the bile acid fraction (0.76% 0.07% vs 0.03% 0.02%; P.001; Figure 1D) was almost completely abolished in BDL mice. Severely Reduced Biliary Cholesterol Secretion Caused by the Absence of Abcb4 Expression Decreases Reverse Cholesterol Transport From Macrophages to Feces Because the complete absence of bile secretion caused by BDL might interfere with intestinal function, we next used a genetic model in which biliary cholesterol secretion is virtually absent, but biliary bile acid secretion is not affected, the Abcb4 knockout mouse. 12,19 Macrophage-derived 3 H-cholesterol in plasma (Figure 2A) was reduced in Abcb4 knockout mice at 6 hours (P.001), 24 hours (P.01), and 48 hours (P.001) compared with wild-type FVB controls. Abcb4 knockout mice had decreased plasma HDL cholesterol levels and increased levels of apolipoprotein B-containing lipoproteins compared with wild-type controls (Figure 2B) and the distribution of the 3 H-cholesterol tracer in plasma largely followed the cholesterol distribution (data not shown). On the other hand, the amount of 3 H-tracer recovered within liver at 48 hours was increased significantly in the Abcb4 knockouts (5.6% 0.6% vs 4.3% 0.2%; P.01; Figure 2C). Gallbladder bile from Abcb4 knockout mice at 48 hours contained significantly less 3 H-tracer within bile acids (0.95% 0.05% vs 0.57% 0.11%/100 L; P.01; Figure 2D) and more than 10-fold decreased 3 H- tracer within cholesterol (0.400% 0.017% vs 0.038% 0.008%/100 L; P.01; Figure 2D) compared with wild-type mice. Consistent with the tracer data, cholesterol mass within gallbladder bile of Abcb4 knockouts was reduced significantly, whereas bile acid concentrations were comparable with wild-type controls (Table 1). 3 H-tracer recovered from the small intestine was low in wild-type controls and even lower in Abcb4 knockout mice (0.93% 0.09% vs 0.65% 0.06%; P.05; Figure 2E). Finally, Abcb4 knockout mice had markedly reduced fecal excretion of macrophage-derived 3 H-tracer in the bile acid fraction (1.03% 0.12% vs 0.57% 0.12%; P.05; Figure 2F) and even more pronounced in the neutral sterol fraction (1.14% 0.09% vs 0.26% 0.02%; P.001; Figure 2F). On the other hand, steady-state mass excretion of neither neutral sterols nor bile acids into the feces was significantly different between FVB wild-type controls and Abcb4 knockout mice (Table 1). We then asked whether the intestinal uptake of HDL-CE is enhanced in Abcb4 knockout mice. Kinetic studies using HDL labeled with 3 H-cholesteryl ether revealed that the HDL-CE FCR was unchanged in the Abcb4 knockout mice compared with FVB controls ( vs pools/h; P NS; Figure 3A). Furthermore, tracer uptake into the liver was not different between Abcb4 knockout mice and controls (41% 2% vs 47% 4%; P NS; Figure 3B). Surprisingly, compared with liver, intestinal 3 H-CE uptake was low in control mice and remained unchanged in Abcb4 knockouts (6.2% 0.4% vs 6.5% 0.7%; P NS; Figure 3C), thus absence of Abcb4 did not cause increased intestinal HDL-CE uptake. LXR Activation Is Insufficient to Stimulate RCT From Macrophages to Feces in Abcb4 Knockout Mice It has been shown previously that in Abcb4 knockout mice LXR activation results in increased fecal neutral sterol excretion. 20 Therefore, we also tested whether LXR activation might increase in vivo macrophage RCT in the absence of Abcb4 expression. Efficient induction of LXR target genes in liver and intestine in both animal models used was demonstrated by real-time quantitative polymerase chain reaction analysis as shown in Table 2. Macrophage-derived 3 H-cholesterol in plasma increased significantly in response to LXR activation at the 24-hour and 48-hour time points in both models (each P.001; Figure 2A), however, was significantly lower at these time points in Abcb4 knockouts compared with wild-type controls (P.05; Figure 2A). Fast protein liquid chromatography profiles revealed increased plasma HDL cholesterol levels in LXR agonist-treated FVB controls and to a lesser extent Abcb4 knockout mice (Figure 2B), and counts within plasma at the 48-hour time point largely followed the cholesterol distribution over the lipoprotein subclasses (data not shown). The amount of 3 H-tracer recovered within liver was not affected by LXR activation in each of the respective models (Figure 2C), and was significantly higher in Abcb4 knockout mice receiving the LXR agonist compared with controls (4.1% 0.2% vs 5.8% 0.6%; P.01; Figure 2C). In gallbladder bile, macrophage-derived 3 H-tracer within bile acids remained unchanged in FVB controls (1.01% 0.10%/100 L; P NS; Figure 2D) and increased significantly in Abcb4 knockout mice (1.11% 0.08%/100 L; P.05; Figure 2D) in response to the LXR agonist treatment. On the other hand, 3 H-cholesterol counts in bile increased in both wild-type (0.640% 0.032%/100 L; P.001, Figure 2D) and Abcb4 knockout mice

5 March 2011 BILIARY STEROL SECRETION DRIVES RCT 1047 Figure 2. Severely reduced biliary cholesterol secretion owing to Abcb4 deficiency decreases macrophage RCT in vivo and functional biliary cholesterol secretion is required to mediate the increasing effect of LXR activation on RCT. Respective groups of wild-type FVB or Abcb4 knockout (KO) mice were administered for 8 days either chow (Con) or chow containing the LXR agonist T (T09) and then injected intraperitoneally with 3 H-cholesterol labeled macrophage foam cells. Treatments were continued for 48 hours, and during this period blood samples were taken at 6, 24, and 48 hours. Feces were collected continuously for 48 hours. (A) Time course of 3 H-cholesterol recovery within plasma. Filled triangles, control wild-type mice; open triangles, LXR agonist-treated wild-type mice; filled circles, control Abcb4 knockout mice; open circles, LXR agonist-treated Abcb4 knockout mice. (B) Cholesterol distribution over the different lipoprotein subclasses determined by fast protein liquid chromatography analysis of pooled plasma samples in wild-type (left panel) and Abcb4 knockout (right panel) mice without (diamonds) or with (squares) LXR agonist administration. (C) 3 H-tracer within liver. (D) 3 H-tracer recovery within the bile acids (BA, left panel) and the cholesterol fractions (Chol, right panel)of gallbladder bile obtained at 48 hours. (E) 3 H-cholesterol within small intestine. (F) 3 H-tracer content within fecal neutral sterols (NS, left panel) and bile acids (BA, right panel) of the respective groups of mice 48 hours after injection of 3 H-cholesterol loaded macrophages. The 3 H-cholesterol data are expressed as the percentage of tracer relative to the total cpm injected. All data are given as means standard error of the mean. *P.05, **P.01, ***P.001 compared with the respective wild-type group, # P.05, ## P.01, ### P.001 compared with the respective chow diet group. VLDL, very low density lipoprotein; LDL, low-density lipoprotein. (0.081% 0.008%/100 L; P.001, Figure 2D), however, remained substantially lower in the Abcb4 knockout model compared with LXR agonist-treated controls (P.001; Figure 2D). Bile acid levels within gallbladder bile did not change in response to LXR activation in both groups, while biliary cholesterol content increased significantly in FVB controls, however, not in the Abcb4 knockout group (Table 1). The 3 H-tracer content of the small intestine did not change in response to LXR agonist treatment in both wild-type controls (0.79% 0.07%; P NS; Figure 2E) and Abcb4 knockouts (0.67% 0.08%; P NS; Figure 2E). Consistent with previous reports, 9,21 LXR activation increased the fecal secretion of macrophage-derived 3 H-cholesterol in wild-type controls by 5.5-fold ( ; P.001; Figure 2F). However, in the Abcb4 knockout model, the LXR agonist resulted

6 1048 NIJSTAD ET AL GASTROENTEROLOGY Vol. 140, No. 3 Table 1. Gallbladder Bile Sterol Contents and Fecal Sterol Excretion in Response to LXR Agonist Treatment FVB Abcb4 knockout Chow T Chow T Biliary cholesterol (mmol/l) a b b Biliary bile acids (mmol/l) Fecal neutral sterols ( mol/d ) a a Fecal bile acids ( mol/d) NOTE. Values are means standard error of the mean. a Significant differences from the respective chow-treated control groups of at least P.05. b Significant differences from the respective FVB wild-type control groups of at least P.05. in an only 1.8-fold ( ; P.01) increase in fecal 3 H-cholesterol tracer excretion and was still by far lower than in the controls (13.5-fold; P.001; Figure 2F). 3 H-tracer recovered within the fecal bile acid fraction remained unaltered in response to LXR activation and was still lower in Abcb4 knockouts compared with controls (1.14% 0.05% vs 0.40% 0.01%; P.001; Figure 2F). In both models, administration of the LXR agonist resulted in significantly increased mass fecal excretion of neutral sterols, whereas bile acid output remained unaltered (Table 1). If the intestine compensates for the lack of biliary cholesterol excretion by up-regulating direct uptake of HDL-CE, treatment of the Abcb4 knockout mice with an LXR agonist might be expected to increase the intestinal uptake of HDL-CE. In wild-type FVB mice, administration of the LXR agonist T did not change the FCR of HDL-CE ( pools/h; P NS; Figure 3A) and also had no effect on tracer uptake into the liver (44% 4%; P NS; Figure 3B). However, intestinal 3 H-CE uptake was decreased significantly in response to LXR activation (4.5% 0.4%; P.05; Figure 3C). Quantitatively similar results were obtained in C57BL/6 wildtype mice (see later for comparison with Scarb1 knockouts). Similar to results in wild-type mice, treatment of the Abcb4 / mice with the LXR agonist left the HDL-CE FCR ( pools/h; P NS; Figure 3A) and hepatic tracer uptake (38% 3%; P NS; Figure 3B) unchanged, whereas intestinal 3 H-CE uptake was significantly lower in response to LXR activation (4.7% 0.3%; P.05; Figure 3C). SR-BI Deficient Mice Have Impaired RCT Owing to Reduced Hepatic HDL-CE Uptake, but the Intestine Does Not Compensate SR-BI (gene name Scarb1) deficient mice have markedly impaired macrophage RCT, 11 at least in part owing to reduced hepatic uptake of HDL-CE. 14,22 The results described earlier indicate that the intestine takes up a small amount of injected HDL-CE, although the pathway for this uptake is unknown. If intestinal SR-BI is responsible, SR-BI knockout mice would be expected to have lower intestinal uptake of HDL-CE than wild-type mice. Compared with wild-type controls and as expected, Scarb1 knockout mice had a reduced HDL-CE FCR ( vs pools/h; P.001; Figure 3A) and reduced hepatic uptake of HDL-CE (23% 2% vs 49% 5%; P.001; Figure 3B), consistent with previous reports. 14,22 Intestinal HDL-CE uptake was not reduced and rather slightly increased in Scarb1 knockout mice compared with wild-type mice (7.5% 0.3% vs 6.0% 0.4%; P.05; Figure 3C). Neither the FCR of HDL-CE ( pools/h; P NS; Figure 3A) nor its hepatic uptake (22% 3%; P NS; Figure 3B) changed in Scarb1 knockout mice in response to LXR agonist treatment, whereas the intestinal uptake of HDL-CE actually decreased (5.2% 0.7%; P.05; Figure 3C). Discussion The results of this study show that biliary cholesterol secretion represents the major relevant route for the irreversible elimination of sterols from the body by the RCT pathway. Two different mouse models of substantially impaired biliary cholesterol excretion both displayed markedly reduced fecal excretion of macrophagederived cholesterol, indicating that neither intestinal secretion nor bile acid synthesis are functionally relevant compensating mechanisms when biliary cholesterol secretion is impaired. Our data further indicate that also the effects of LXR agonists in promoting RCT depend on the biliary secretion pathway. The classic view on RCT 2,3,5 as excess cholesterol being transported from the periphery within HDL particles through the plasma compartment to the liver for excretion into bile and subsequently feces has been challenged recently. The intestine was shown to possess the capacity to contribute significantly to mass cholesterol excretion via the feces in a manner susceptible toward activation by LXR agonists. 6,7,20 Because based on these findings the intestine was suggested to represent a novel and promising target for the stimulation of RCT, 7,8 the purpose of our present study was to test the significance of the intestine for in vivo RCT from macrophages to feces. First, we performed a series of HDL kinetic studies to assess the uptake of HDL cholesteryl ester by the intestine under different relevant experimental conditions. Compared with liver, HDL CE uptake by the intestine is

7 March 2011 BILIARY STEROL SECRETION DRIVES RCT 1049 Figure 3. HDL kinetic experiments performed in wild-type, Sr-bI knockout (KO), and Abcb4 knockout mice receiving either chow (Con) or chow supplemented with the LXR agonist T (T09). HDL was labeled with 3 H-cholesteryl ether as described in the Materials and Methods section and injected into the different mouse models (n 4 6 mice per group). Blood samples were taken at different time points over 48 hours, plasma counts were determined and tracer disappearance curves were constructed and modeled using the SAAM II program. At 48 hours tracer uptake into tissues was assessed. (A) FCRs calculated from the respective plasma disappearance curves. (B) Liver uptake of 3 H-cholesteryl ether. (C) Uptake of 3 H-cholesteryl ether by the small intestine. Data are given as means standard error of the mean. # P.05, ### P.001 compared with the respective wild-type group. *P.05 compared with the respective chow diet group. rather low. In addition, based on previous data showing that LXR activation increases intestinal cholesterol secretion 6,20 we were surprised to note that treatment with LXR agonists resulted in decreased HDL CE uptake rates by the intestine in all mouse models investigated. Furthermore, intestinal HDL CE uptake was increased significantly in SR-BI knockout mice directly opposed to what would be predicted if the intestine would play a significant role in RCT. Of note, in intestinal perfusion experiments SR-BI knockout mice have been found to have increased intestinal cholesterol secretion rates, 8 whereas in contrast RCT is decreased in SR-BI knockouts and increased upon hepatic SR-BI overexpression. 11 However, taken together, in our interpretation the results of our kinetic studies would argue rather against than in favor of a major role for the intestine in HDL-driven RCT. For a more direct assessment of the relevance of the biliary route for RCT we chose to mechanically disrupt biliary secretion by performing BDLs. This almost completely abolished RCT, indicating that under these experimental conditions there is no compensation by the intestinal route. However, because BDL is a rather harsh procedure and bile acids within the intestinal lumen might not only be required to mediate efficient cholesterol absorption, but also to elicit intestinal cholesterol secretion, 7,8 we next investigated a genetic model of severely deficient biliary cholesterol secretion, Abcb4 knockout mice. Abcb4 knockouts have virtually absent biliary cholesterol secretion rates secondary to impaired phospholipid secretion into the bile via Abcb4. 12,19 The total delivery of label from macrophages to feces was reduced significantly in this model, indicative of decreased in vivo RCT. Corresponding to the decreased biliary cholesterol secretion rates of these mice we observed that transport of labeled cholesterol from macrophage foam cells to the feces was almost abolished, whereas label recovered within bile acids was reduced slightly. These data indicate that almost all cholesterol RCT occurs via the biliary route and that bile acid synthesis under these conditions is not compensating for decreased biliary cholesterol excretion. It has been shown previously that general 9,21 as well as tissue-specific 23 LXR activation enhances in vivo RCT, and LXR activation also increases direct intestinal cholesterol secretion. 6 When we administered a LXR agonist to Abcb4 knockout mice the plasma steady-state HDL cholesterol levels did not increase to the same extent as in wild-type controls. The HDL cholesteryl ester catabolic Table 2. Hepatic and Intestinal Gene Expression Levels in Response to LXR Agonist Treatment FVB Abcb4 knockout Chow T a Chow T a Abca1 Liver Intestine Abcg5 Liver Intestine Abcg8 Liver Intestine Abcg1 Liver Intestine NOTE. Values are means standard error of the mean. a Significantly different from the respective chow-treated control groups (at least P.05).

8 1050 NIJSTAD ET AL GASTROENTEROLOGY Vol. 140, No. 3 rate remained unaltered in LXR agonist-treated Abcb4 knockout mice, whereas the increase in hepatic as well as intestinal expression of Abca1 and Abcg1 in response to the LXR agonist was lower compared to the response of wild-type controls. These results point toward decreased HDL formation in the Abcb4-deficient model. The molecular basis for this phenotype is currently not clear, however, we do not consider this observation to have a major impact on the results of the present study because plasma HDL cholesterol levels do not reliably predict changes in RCT. Interestingly, however, mass fecal excretion of cholesterol increased substantially in Abcb4 / mice, but, in contrast to wild-type controls, not biliary cholesterol content and correspondingly also not counts recovered within neutral sterols in the feces. These results indicate that functional biliary secretion also is required for the stimulating effects of LXR agonists on RCT. These data further suggest that RCT and direct intestinal cholesterol secretion are distinct pathways that are not functionally connected. However, because intestinal cholesterol secretion is still not characterized sufficiently, a definitive assessment has to await identification of the receptors/transporters involved in this process. While our present work was under review, another group reported that decreasing biliary cholesterol secretion by transgenic overexpression of human NPC1L1 and acute bile diversion for 8 hours did not result in significantly decreased rates of macrophage RCT. 24 The mechanistic basis for the different conclusions compared with our study is not apparently clear. Decreasing biliary cholesterol secretion might not be sufficient to reduce RCT significantly because as a result of metabolic compartmentalization there might be a preference for favoring the excretion of cholesterol derived from the RCT pathway. The results of the bile diversion experiment 24 are unexpected. In this model no bile acids are present in the intestine, however, previous intestinal perfusion studies established that the presence of intestinal bile acids is mandatory for intestinal cholesterol secretion. 7 Interestingly, patients with cholestasis regularly present with xanthomas similar to patients with familial hypercholesterolemia who suffer from a severe increase in plasma low-density lipoprotein cholesterol levels owing to defective catabolism. 28 Histologically, these lipid deposits consist of lipid-laden macrophages that closely resemble macrophage foam cells. 29,30 Of note, xanthomas in patients with cholestasis are dynamic, that is, they also disappear over time, when the continuity of bile flow is reestablished. 25,26 These findings might indicate that defective RCT in cholestatic conditions is relevant for human patients and that also in these patients the intestine cannot compensate for reduced or absent biliary cholesterol secretion. In summary, our results indicate that at least under the experimental conditions used in this study biliary cholesterol secretion represents the major pathway relevant for RCT. Put in perspective, based on our data, focusing therapeutic strategies aiming to enhance in vivo RCT on increasing biliary sterol excretion might be relevant for prevention and treatment of atherosclerotic cardiovascular disease. References 1. Assmann G, Gotto AM Jr. HDL cholesterol and protective factors in atherosclerosis. Circulation 2004;109:III8 III Linsel-Nitschke P, Tall AR. HDL as a target in the treatment of atherosclerotic cardiovascular disease. Nat Rev Drug Discov 2005;4: Cuchel M, Rader DJ. Macrophage reverse cholesterol transport: key to the regression of atherosclerosis? Circulation 2006;113: Rader DJ. Molecular regulation of HDL metabolism and function: implications for novel therapies. J Clin Invest 2006;116: Lewis GF. Determinants of plasma HDL concentrations and reverse cholesterol transport. Curr Opin Cardiol 2006;21: van der Veen JN, van Dijk TH, Vrins CL, et al. Activation of the liver X receptor stimulates trans-intestinal excretion of plasma cholesterol. J Biol Chem 2009;284: van der Velde AE, Vrins CL, van den Oever K, et al. Direct intestinal cholesterol secretion contributes significantly to total fecal neutral sterol excretion in mice. Gastroenterology 2007; 133: van der Velde AE, Vrins CL, van den Oever K, et al. Regulation of direct transintestinal cholesterol excretion in mice. Am J Physiol Gastrointest Liver Physiol 2008;295:G203 G Naik SU, Wang X, Da Silva JS, et al. Pharmacological activation of liver X receptors promotes reverse cholesterol transport in vivo. Circulation 2006;113: Wang X, Collins HL, Ranalletta M, et al. Macrophage ABCA1 and ABCG1, but not SR-BI, promote macrophage reverse cholesterol transport in vivo. J Clin Invest 2007;117: Zhang Y, Da Silva JR, Reilly M, et al. Hepatic expression of scavenger receptor class B type I (SR-BI) is a positive regulator of macrophage reverse cholesterol transport in vivo. J Clin Invest 2005;115: Oude Elferink RPJ, Ottenhoff R, van Wijland M, et al. Regulation of biliary lipid secretion by mdr2 P-glycoprotein in the mouse. J Clin Invest 1995;95: Tietge UJF, Maugeais C, Cain W, et al. Overexpression of secretory phospholipase A(2) causes rapid catabolism and altered tissue uptake of high density lipoprotein cholesteryl ester and apolipoprotein A-I. J Biol Chem 2000;275: Nijstad N, Wiersma H, Gautier T, et al. Scavenger receptor BImediated selective uptake is required for the remodeling of high density lipoprotein by endothelial lipase. J Biol Chem 2009;284: Tietge UJF, Maugeais C, Cain W, et al. Acute inflammation increases selective uptake of HDL cholesteryl esters into adrenals of mice overexpressing human spla2. Am J Physiol Endocrinol Metab 2003;285:E403 E Wiersma H, Gatti A, Nijstad N, et al. Hepatic SR-BI but not endothelial lipase (EL) expression determines biliary cholesterol secretion in mice. J Lipid Res 2009;50: Tietge UJF, Nijstad N, Havinga R, et al. Secretory phospholipase A2 increases SR-BI-mediated selective uptake from HDL but not biliary cholesterol secretion. J Lipid Res 2008;49: Wiersma H, Nijstad N, de Boer JF, et al. Lack of Abcg1 results in decreased plasma HDL cholesterol levels and increased biliary cholesterol secretion in mice fed a high cholesterol diet. Atherosclerosis 2009;206:

9 March 2011 BILIARY STEROL SECRETION DRIVES RCT Wiersma H, Gatti A, Nijstad N, et al. Scavenger receptor BI (SR-BI) mediates biliary cholesterol secretion independent of Abcg5/g8 in mice. Hepatology 2009;50: Kruit JK, Plosch T, Havinga R, et al. Increased fecal neutral sterol loss upon liver X receptor activation is independent of biliary sterol secretion in mice. Gastroenterology 2005;128: Zanotti I, Poti F, Pedrelli M, et al. The LXR agonist T promotes the reverse cholesterol transport from macrophages by increasing plasma efflux potential. J Lipid Res 2008;49: Brundert M, Ewert A, Heeren J, et al. Scavenger receptor class B type I mediates the selective uptake of high-density lipoproteinassociated cholesteryl ester by the liver in mice. Arterioscler Thromb Vasc Biol 2005;25: Yasuda T, Grillot D, Billheimer JT, et al. Tissue-specific liver X receptor activation promotes macrophage reverse cholesterol transport in vivo. Arterioscler Thromb Vasc Biol 2010;30: Temel RE, Sawyer JK, Yu L, et al. Biliary sterol secretion is not required for macrophage reverse cholesterol transport. Cell Metab 2010;12: Emerick KM, Whitington PF. Partial external biliary diversion for intractable pruritus and xanthomas in Alagille syndrome. Hepatology 2002;35: Englert C, Grabhorn E, Burdelski M, et al. Liver transplantation in children with Alagille syndrome: indications and outcome. Pediatr Transplant 2006;10: McIntyre N, Harry DS, Pearson AJ. The hypercholesterolaemia of obstructive jaundice. Gut 1975;16: Civeira F. Guidelines for the diagnosis and management of heterozygous familial hypercholesterolemia. Atherosclerosis 2004;173: Scott PJ, Winterbourn CC. Low-density lipoprotein accumulation in actively growing xanthomas. J Atheroscler Res 1967;7: Tietge UJF, Maschek H, Schneider A, et al. [Xanthoma disseminatum with marked mucocutaneous involvement]. Dtsch Med Wochenschr 1998;123: Received July 8, Accepted November 24, Reprint requests Address requests for reprints to: Uwe Tietge, MD, Department of Pediatrics, Center for Liver, Digestive and Metabolic Diseases, University Medical Center Groningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands. u_tietge@yahoo.com; fax: (31) Acknowledgments N.N., T.G., D.J.R., and U.J.F.T contributed equally to this study. Conflicts of interest The authors disclose no conflicts. Funding This work was supported by grants from The Netherlands Organization for Scientific Research (VIDI grant to U.J.F.T.) and the National Institutes of Health (P01-HL to D.J.R.).