Scavenger receptor BI and ABCG5/G8 differentially impact biliary sterol secretion and reverse cholesterol transport in mice

Authors

  • Arne Dikkers,

    1. Department of Pediatrics, University Medical Center Groningen, University of Groningen, the Netherlands
    Search for more papers by this author
  • Jan Freak de Boer,

    1. Department of Pediatrics, University Medical Center Groningen, University of Groningen, the Netherlands
    Search for more papers by this author
  • Wijtske Annema,

    1. Department of Pediatrics, University Medical Center Groningen, University of Groningen, the Netherlands
    2. Top Institute Food and Nutrition, Wageningen, the Netherlands
    Search for more papers by this author
  • Albert K. Groen,

    1. Department of Pediatrics, University Medical Center Groningen, University of Groningen, the Netherlands
    2. Department of Laboratory Medicine, University Medical Center Groningen, University of Groningen, the Netherlands
    Search for more papers by this author
  • Uwe J.F. Tietge

    Corresponding author
    1. Department of Pediatrics, University Medical Center Groningen, University of Groningen, the Netherlands
    2. Top Institute Food and Nutrition, Wageningen, the Netherlands
    • Department of Pediatrics, University Medical Center Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands===

    Search for more papers by this author
    • fax +31-50-3611746


  • Potential conflict of interest: Nothing to report.

  • Supported by grants from the Netherlands Organization for Scientific Research (VIDI Grant 917-56-358 to U.J.F.T.) and the Top Institute (TI) Food and Nutrition (to U.J.F.T.).

Abstract

Biliary lipid secretion plays an important role in gallstone disease and reverse cholesterol transport (RCT). Using Sr-bI/Abcg5 double knockout mice (dko), the present study investigated the differential contribution of two of the most relevant transporters: adenosine triphosphate (ATP)-binding cassette subfamily G member 5 and 8 (ABCG5/G8) and scavenger receptor class B type I (SR-BI) to sterol metabolism and RCT. Plasma cholesterol levels increased in the following order, mainly due to differences in high density lipoprotein (HDL): Abcg5 ko < wild type < Sr-bI/Abcg5 dko < Sr-bI ko. Liver cholesterol content was elevated in Sr-bI ko only (P < 0.05). In Sr-bI/Abcg5 dko plasma plant sterols were highest, while hepatic plant sterols were lower compared with Abcg5 ko (P < 0.05). Under baseline conditions, biliary cholesterol secretion rates decreased in the following order: wild type > Sr-bI ko (−16%) > Abcg5 ko (−75%) > Sr-bI/Abcg5 dko (−94%), all at least P < 0.05, while biliary bile acid secretion did not differ between groups. However, under supraphysiological conditions, upon infusion with increasing amounts of the bile salt tauroursodeoxycholic acid, Abcg5 became fully rate-limiting for biliary cholesterol secretion. Additional in vivo macrophage-to-feces RCT studies demonstrated an almost 50% decrease in overall RCT in Sr-bI/Abcg5 dko compared with Abcg5 ko mice (P < 0.01). Conclusion: These data demonstrate that (1) SR-BI contributes to ABCG5/G8-independent biliary cholesterol secretion under basal conditions; (2) biliary cholesterol mass secretion under maximal bile salt-stimulated conditions is fully dependent on ABCG5/G8; and (3) Sr-bI contributes to macrophage-to-feces RCT independent of Abcg5/g8. (Hepatology 2013;)

Sterol elimination from the body can be accomplished by the liver by way of biliary secretion,1 and by the intestine by way of modulation of absorption rates as well as by transintestinal cholesterol efflux.2, 3 By mediating the excretion of cholesterol, bile acids, and plant sterols the biliary secretion pathway has relevant implications for major disease complexes such as gallstone disease4, 5 and atherosclerotic cardiovascular disease (CVD).6-8 In the case of CVD, biliary secretion is a critical step for reverse cholesterol transport (RCT).6, 9 RCT comprises the transport of cholesterol from macrophage foam cells within atherosclerotic lesions on high density lipoproteins (HDLs) to the liver for excretion into bile and by way of the feces out of the body.8

Several transporters localized to the bile canaliculus, the apical plasma membrane of hepatocytes, play essential roles in the biliary secretion process.1 Adenosine triphosphate (ATP)-binding cassette transporter B4 (ABCB4) is required for the biliary output of phospholipids and is key for the formation of mixed micelles, which are necessary for the efficient biliary excretion of sterols.1, 6 This is illustrated by Abcb4 knockout (ko) mice having almost unmeasurable biliary cholesterol secretion rates.10 Another important positive regulator of biliary cholesterol secretion are bile acids. Biliary bile acid (BA) secretion is mediated by ABCB11,11, 12 and stimulating biliary BA output directly translates into increased biliary cholesterol secretion rates.13 ATPase class I type 8B member 1 preserves the rigid structure of the outer canalicular leaflet by inward flipping of phosphatidylserine, making it more resistant to detergents.14 The majority of cholesterol is secreted by way of the obligate heterodimeric transport proteins ATP-binding cassette subfamily G member 5 and 8 (ABCG5/G8), initially identified as the molecular defect in sitosterolemia.15, 16 Recently, Niemann-Pick type C protein 2 has been shown to act as a positive regulator of ABCG5/G8-dependent secretion.17 The absence of Abcg5 and/or Abcg8 in mice results in plant sterol accumulation and an ∼75% reduction in biliary cholesterol output.18-20 However, 25% of the total biliary cholesterol secretion still occurs independent of Abcg5/g8.

The proteins and mechanisms contributing to this Abcg5/g8-independent secretion are not well understood. We recently identified scavenger receptor class B type I (SR-BI) as a protein that localizes to the bile canaliculus.20 When overexpressed, SR-BI causes a substantial increase in biliary cholesterol content21 and also in biliary cholesterol secretion rates,20, 22 an effect that was independent of Abcg5/g8.20 Notably, hepatic SR-BI overexpression was able to restore the substantially decreased biliary cholesterol secretion rates of Abcg5-deficient mice to levels of wild-type controls.20 Therefore, SR-BI conceivably represents a prime candidate responsible for the Abcg5/g8-independent part of biliary cholesterol secretion.

The present study tested this hypothesis by generating Abcg5/Sr-bI double-deficient mice. In depth characterization of this novel double knockout (dko) mouse model revealed that (1) under steady-state conditions SR-BI significantly contributes to the Abcg5/g8-independent part of biliary cholesterol secretion; (2) SR-BI is involved in the cellular uptake of plant sterols; (3) under BA-stimulated conditions Abcg5/g8 are rate-limiting for the biliary cholesterol secretion process; and (4) the absence of Sr-bI on the Abcg5 ko background translates also into a substantial decrease in in vivo RCT.

Abbreviations

ABCB4, ATP-binding cassette transporter B4; ABCG5/G8, ATP-binding cassette subfamily G member 5 and 8; ATP, adenosine triphosphate; BA, bile acid; CVD, cardiovascular disease; HDL, high density lipoprotein; RCT, reverse cholesterol transport; SR-BI, scavenger receptor class B type I.

Materials and Methods

Detailed experimental procedures are provided as online Supporting Information.

Animals.

Sr-bI/Abcg5 dko mice were generated by crossing Sr-bI ko mice (Jackson Laboratories, Bar Harbor, ME) with Abcg5 ko.20 Littermate controls were from the breeding colony. Animals were housed in temperature-controlled rooms (21°C) with alternating 12-hour periods of light and dark and ad libitum access to water and chow diet (Arie Blok, Woerden, The Netherlands). Animal experiments were performed in accordance with national laws. The responsible Ethics Committee of the University of Groningen approved all protocols.

Analysis of Plasma Lipids and Lipoproteins.

Blood was obtained by cardiac puncture under anesthesia following a 4-hour fast. Aliquots of plasma were stored at −80°C until analysis. Plasma phospholipids and triglycerides were measured enzymatically (Roche Diagnostics, Mannheim, Germany, and Diasys, Holzheim, Germany, respectively), cholesterol and plant sterols/stanols were measured by gas chromatography-mass spectrometry (GC-MS) as described.20 Pooled plasma samples were subjected to fast protein liquid chromatography (FPLC) gel filtration using a superose 6 column (GE Healthcare, Little Chalfont, UK) as described.23

Analysis of Liver Lipid Composition.

Liver tissue was homogenized, lipids extracted according to the general procedure of Bligh and Dyer, and lipids were redissolved in water containing 2% Triton X-100 (Sigma-Aldrich, Steinheim, Germany).24 Phospholipid content was determined as published24 and sterols were analyzed by GC-MS as detailed above.

Bile Collection and Composition Analysis.

The gallbladder was cannulated under anesthesia, bile was collected for 30 minutes, and production was determined gravimetrically.24 Body temperature was maintained using a humidified incubator.24 Biliary cholesterol, phospholipid, bile salt, and plant sterol/stanol concentrations were determined and the respective biliary secretion rates calculated.24

BA Infusion Experiments.

The BA infusion experiment was performed as described.13 After gallbladder cannulation and two basal bile samples the mice were continuously infused with tauroursodeoxycholate (TUDCA; Calbiochem/Merck Biosciences, Darmstadt, Germany) in phosphate-buffered saline (PBS) by way of the jugular vein. TUDCA infusion rates were increased in a stepwise manner: 150, 300, 450, and 600 nmol/min. Bile samples were collected at 15-minute intervals for 210 minutes (three samples per infusion step).

Analysis of Hepatic Gene Expression by Real-Time Polymerase Chain Reaction (PCR).

Real-time quantitative PCR was performed as published on a 7900HT fast real-time PCR system (Applied Biosystems, Foster City, CA).20 Messenger RNA (mRNA) expression levels were calculated relative to the average of the housekeeping gene cyclophilin and further normalized to the expression levels of the respective controls.

Electron Microscopy.

Mice were anesthetized and perfused with 5 mL fixation buffer (4% paraformaldehyde, 2% glutaraldehyde, pH 7.4).20 Ultrathin sections (70-100 nm) were prepared on the basis of Toluidine blue-stained 1-μm sections, contrast-stained with 2% uranyl acetate and lead citrate according to routine procedures, and examined using a Philips CM100 electron microscope operating at 80 kV.

In Vivo Macrophage-to-Feces RCT Studies.

Experiments were essentially performed as described using thioglycollate-elicited primary peritoneal macrophages from C57BL/6J donor mice loaded in vitro with 50 μg/mL acetylated LDL and 3 μCi/mL [3H]cholesterol (Perkin Elmer, Boston, MA) for 24 hours to become foam cells.25 These were injected intraperitoneally into individually housed recipient mice. Counts in plasma (6, 24, 48 hours) were assessed directly by liquid scintillation counting (Packard 1600CA Tri-carb, Packard, Meriden, CT), tracer uptake into the liver was determined at 48 hours following solubilization of the tissue (Solvable, Packard) as described.26 Feces were collected for 48 hours, then dried, weighed, and ground. After incubation in Solvable counts were determined by liquid scintillation counting and related to the total amount of feces produced over the 48-hour experimental period. All obtained counts were expressed relative to the administered tracer dose.

Statistics.

Statistical analyses were performed using GraphPad Prism (San Diego, CA). Data are presented as means ± standard error of the mean (SEM). Statistical differences between two groups were assessed using the Mann-Whitney U test. When comparing multiple groups, statistical analysis was performed using analysis of variance (ANOVA) followed by a Bonferroni posttest or repeated measures ANOVA followed by a Bonferroni posttest (Fig. 4). The coupling of cholesterol secretion to BAs and phospholipids (Fig. 5) was analyzed with linear regression and slopes were compared. Statistical significance for all comparisons was assigned at P < 0.05.

Results

In the Absence of Sr-bI Plant Sterols Accumulate in Plasma, Preferentially in HDL.

Sr-bI/Abcg5 dko mice were characterized in comparison to wild-type controls and the respective Sr-bI and Abcg5 single ko mice. Plasma total cholesterol levels were increased in the Sr-bI ko (P < 0.05) and significantly decreased in the Abcg5 ko model (P < 0.05), but were unchanged in Sr-bI/Abcg5 dko mice compared to wild-type controls (Fig. 1A). However, FPLC profiles showed major changes in the lipoprotein profile attributable to altered Sr-bI as well as Abcg5 expression (Fig. 1B). While Sr-bI ko mice had a higher and broader HDL peak with a clear shift towards larger particles (Fig. 1B), Abcg5 ko mice exhibited a lower HDL peak with no apparent shift in particle size compared with controls. The Sr-bI/Abcg5 dko mice showed a combination of these two effects: the HDL peak was lower than in Sr-bI ko mice but still had an appreciable shift towards larger particles (Fig. 1B). Plasma plant sterol levels increased gradually from wild-type over Sr-bI ko and Abcg5 ko mice towards the dko (Fig. 1C). While the increase in plasma plant sterols in Abcg5 ko mice is consistent with the intestinal plant sterol hyperabsorption already described in this mouse model,18 the further increase in the absence of Sr-bI indicates that Sr-bI might be involved in cellular plant sterol uptake. When the plant sterol content of isolated lipoprotein fractions was investigated, we noticed that these are preferentially contained within HDL (Fig. 1D). Plasma plant stanols largely followed the pattern observed for plant sterols (Fig. 1E,F). A detailed overview of alterations in specific individual sterols in plasma in the different models is presented in Table 1.

Figure 1.

Differential impact of Sr-bI and Abcg5 expression on levels and distribution of plasma sterols. Wild-type, Sr-bI ko, Abcg5 ko, and Sr-bI/Abcg5 dko mice were fasted for 4 hours and a blood sample was taken. (A) Plasma total cholesterol levels, (B) FPLC cholesterol profiles of pooled plasma samples, (C) plasma plant sterol levels, (D) distribution of plant sterols over the VLDL, LDL, and HDL fractions, (E) plasma plant stanol levels, (F) distribution of plant stanols over the VLDL, LDL, and HDL fractions. Data in (A,C,E) are presented as means ± SEM. n = 6 for each group. a indicates statistically significant differences from wild-type, b from Sr-bI ko, c from Abcg5 ko, and d from Sr-bI/Abcg5 dko mice (at least P < 0.05).

Table 1. Body Weight, Liver/Body Weight Ratio, and Sterol Species in Plasma, Liver, and Bile of Wild-Type, Sr-bI Knockout, Abcg5 Knockout, and Sr-bI/Abcg5 Double Knockout Mice
 Wild-TypeSr-bI koAbcg5 koSr-bI/Abcg5 dko
  1. Mice were fasted for 4 hours. Blood was collected by cardiac puncture under anesthesia. Livers were removed, weighed and stored at -80°C until analysis. Continuous bile cannulations were performed for 30 minutes. Cholesterol and individual phytosterol concentrations were determined in plasma, liver and bile by GC-MS. Individual phytosterol biliary secretion rates were calculated. Data are presented as means ± SEM. n = 6 for each group. *Statistically significant differences from wild-type; †from Sr-bI knockout; ‡from Abcg5 knockout; and §from Sr-bI/Abcg5 double knockout mice (at least P < 0.05). N.D., not detected; bw, body weight.

Body weight (g)31.0 ± 0.930.5 ± 1.825.2 ± 1.226.8 ± 1.0
Liver/body weight ratio0.048 ± 0.0010.039 ± 0.002*‡§0.060 ± 0.003*†0.056 ± 0.001*†
Sterol species in plasma    
Cholesterol (mmol/l)2.56 ± 0.155.32 ± 0.35*‡§1.22 ± 0.13*†§3.13 ± 0.40†‡
Brassicasterol (μmol/l)0.44 ± 0.040.82 ± 0.11‡§6.23 ± 0.34*†§19.64 ± 2.16*†‡
Campesterol (μmol/l)31.78 ± 2.4062.43 ± 9.09‡§234.92 ± 20.29*†§364.83 ± 37.93*†‡
Campestanol (μmol/l)0.46 ± 0.040.60 ± 0.11‡§64.05 ± 7.56*†§93.41 ± 12.07*†‡
Stigmasterol (μmol/l)0.10 ± 0.000.14 ± 0.03‡§9.16 ± 0.43*†§23.05 ± 2.67*†‡
β-Sitosterol (μmol/l)9.71 ± 0.6821.21 ± 2.88‡§484.40 ± 47.83*†§740.37 ± 90.08*†‡
Sitostanol (μmol/l)0.10 ± 0.000.14 ± 0.04‡§36.63 ± 4.05*†§61.86 ± 8.79*†‡
Sterol species in liver    
Cholesterol (nmol/mg liver)1.07 ± 0.111.70 ± 0.10*‡§0.64 ± 0.08*†0.75 ± 0.05†
Brassicasterol (pmol/mg liver)0.35 ± 0.090.37 ± 0.10‡§6.10 ± 0.50*†6.00 ± 0.74*†
Campesterol (pmol/mg liver)11.58 ± 1.1915.66 ± 1.55‡§108.18 ± 11.01*†§71.47 ± 7.11*†‡
Campestanol (pmol/mg liver)N.D.0.22 ± 0.14‡§38.06 ± 1.38*†§25.51 ± 2.08*†‡
Stigmasterol (pmol/mg liver)0.04 ± 0.030.09 ± 0.5‡§9.18 ± 0.70*†§5.98 ± 0.59*†‡
β-Sitosterol (pmol/mg liver)2.55 ± 0.273.10 ± 0.16‡§174.81 ± 16.22*†§101.82 ± 12.18*†‡
Sitostanol (pmol/mg liver)N.D.N.D.22.03 ± 1.23*†§14.96 ± 1.60*†‡
Total sterols & stanols (pmol/mg liver)1089.02 ± 108.301720.48 ± 97.30*‡§1000.96 ± 98.81†975.41 ± 62.09†
Sterol species in bile    
Brassicasterol (pmol/min/100g bw)9.21 ± 1.035.67 ± 0.307.49 ± 1.019.10 ± 1.52
Campesterol (pmol/min/100g bw)126.41 ± 16.11112.78 ± 12.8595.44 ± 13.8183.59 ± 5.77
Campestanol (pmol/min/100g bw)7.40 ± 1.257.39 ± 0.58‡§20.51 ± 2.16*†15.72 ± 0.79*†
Stigmasterol (pmol/min/100g bw)1.91 ± 0.111.46 ± 0.09‡§5.25 ± 0.51*†§2.69 ± 0.19†‡
β-Sitosterol (pmol/min/100g bw)54.77 ± 6.6440.12 ± 3.15‡97.06 ± 12.37*†71.97 ± 5.68
Sitostanol (pmol/min/100g bw)1.91 ± 0.111.46 ± 0.09‡§5.53 ± 0.30*†§3.74 ± 0.62*†‡

Hepatic Plant Sterol Content Is Decreased in the Absence of Sr-bI.

The body weight of Abcg5 ko mice was decreased (P < 0.05, Table 1). Livers of the different models showed no apparent differences in size or gross morphology (Fig. 2A). However, liver weight as percentage of body weight was lower in Sr-bI ko mice and higher in Abcg5 ko and Sr-bI/Abcg5 dko mice (P < 0.05, Table 1). As examined by electron microscopy, there were no apparent differences between the groups on the ultrastructural level, in particular the morphology of bile canaliculi was not altered and there was no apparent indication of cholestasis in all of the models (Fig. 2B).

Figure 2.

Effect of Sr-bI and Abcg5 expression on liver morphology and hepatic sterol content. (A) Livers from wild-type, Sr-bI ko, Abcg5 ko, and Sr-bI/Abcg5 dko mice show no apparent differences in weight, size, or gross morphology. (B) Representative electron microscopy (EM) pictures of livers from wild-type, Sr-bI ko, Abcg5 ko, and Sr-bI/Abcg5 dko mice showing no alterations in ultrastructural morphology. Hepatic contents of (C) total cholesterol, (D) cholesterolester, (E) plant sterols, (F) plant stanols. Data in (C-F) are presented as means ± SEM. n = 6 for each group. a indicates statistically significant differences from wild-type, b from Sr-bI ko, c from Abcg5 ko, and d from Sr-bI/Abcg5 dko mice (at least P < 0.05).

Liver total cholesterol content was increased in Sr-bI ko mice compared to the other models (Fig. 2C, P < 0.05). Thereby the hepatic content of cholesterol esters was highest in Sr-bI ko (P < 0.05), while the Sr-bI/Abcg5 dko mice had decreased hepatic cholesterol esters compared to wild-type mice (Fig. 2D, P < 0.05). Liver free cholesterol content was increased in Sr-bI ko mice only (P < 0.05, data not shown). There were no differences in hepatic phospholipid content (data not shown). Total hepatic plant sterol (Fig. 2E) and stanol (Fig. 2F) contents were rather low in mice with intact Abcg5, whereas the Abcg5 ko model exhibited a substantial increase in hepatic plant sterols as well as stanols (P < 0.05). However, in Sr-bI/Abcg5 dko mice hepatic plant sterols and stanols were significantly lower than in the Abcg5 single ko model (P < 0.05, Fig. 2E,F), indicating a potential involvement of SR-BI in the uptake of plant sterols and stanols. Changes in total as well as individual hepatic sterols are summarized in Table 1.

Biliary Cholesterol Secretion Is Decreased in Sr-bI/Abcg5 dko Mice Under Steady-State Conditions.

Next we performed continuous bile cannulation experiments in the four different mouse models. Bile flow was increased in Sr-bI/Abcg5 dko compared with wild-type and Sr-bI ko mice (P < 0.05, Fig. 3A). Phospholipid and BA secretion rates were both unchanged among the models (Fig. 3B,C). Biliary cholesterol secretion was significantly lower in Abcg5 ko mice, consistent with previous reports,13, 20 but was even further decreased in Sr-bI/Abcg5 dko mice (P < 0.05, Fig. 3D). These data indicate that Sr-bI contributes significantly to the Abcg5-independent part of biliary cholesterol secretion. Plant sterol secretion rates were unchanged among the four models (Table 1); however, plant stanol secretion was higher in Abcg5 ko as well as Sr-bI/Abcg5 dko mice (Table 1). Biliary secretion rates for individual sterols are also given in Table 1. Hepatic mRNA expression of Abcb4 and Abcb11 were unchanged among the four models (Table 2). Abcg5 mRNA expression was unaltered in Sr-bI ko mice. Sr-bI mRNA expression was increased by 48% in Abcg5 ko mice (P < 0.05, Table 2); however, Sr-bI protein expression was decreased (Supporting Fig. I). Hepatic expression of Srebp2 and its target genes Ldlr and Hmgcoar was decreased in the absence of Abcg5 (Table 2). Cyp7a1 expression was elevated in Sr-bI and Abcg5 ko mice, while Cyp8b1 was higher in Abcg5 ko and decreased in the models lacking Sr-bI (Table 2). However, these mRNA expression changes in BA synthesis enzymes did not translate into altered plasma BA levels, BA pool size, fecal BA excretion, or composition of the BA pool in the four models studied (Supporting Table I). A detailed analysis of intestinal gene expression is provided in Supporting Table II.

Figure 3.

Differential impact of Sr-bI and Abcg5 on biliary sterol secretion under steady-state conditions. Continuous bile cannulation experiments for 30 minutes were performed in wild-type, Sr-bI ko, Abcg5 ko, and Sr-bI/Abcg5 dko mice as detailed in Materials and Methods. (A) Bile flow, biliary secretion rates of (B) phospholipids, (C) bile acids, and (D) cholesterol. Data are presented as means ± SEM. n = 6 for each group. a indicates statistically significant differences from wild-type, b from Sr-bI ko, c from Abcg5 ko, and d from Sr-bI/Abcg5 dko mice (at least P < 0.05).

Table 2. Hepatic mRNA Expression in Wild-Type, Sr-bI Knockout, Abcg5 Knockout and Sr-bI/Abcg5 Double Knockout Mice
 Wild-TypeSr-bI koAbcg5 koSr-bI/Abcg5 dko
  1. Mice were fasted for 4 hours, livers were removed, weighed, and stored at -80°C until analysis. Individual genes are expressed as a percentage of the housekeeping gene cyclophilin and further normalized to the expression levels of the respective controls. Data are presented as means ± SEM. n = 6 for each group. *Statistically significant differences from wild-type; †from Sr-bI knockout; ‡from Abcg5 knockout; and §from Sr-bI/Abcg5 double knockout mice (at least P < 0.05). N.D., not detected.

Hepatic mRNA expression    
Sr-bI1.00 ± 0.05N.D.1.48 ± 0.06*N.D.
Ldlr1.00 ± 0.040.96 ± 0.08‡0.54 ± 0.03*‡§0.85 ± 0.04‡
Abca11.00 ± 0.041.00 ± 0.051.11 ± 0.081.02 ± 0.06
Abcg51.00 ± 0.100.89 ± 0.12N.D.N.D.
Abcg81.00 ± 0.070.72 ± 0.07‡§1.09 ± 0.06†1.07 ± 0.11†
Npc21.00 ± 0.031.11 ± 0.051.15 ± 0.111.19 ± 0.16
Abcb41.00 ± 0.060.86 ± 0.090.93 ± 0.060.74 ± 0.05
Abcb111.00 ± 0.050.86 ± 0.091.17 ± 0.061.11 ± 0.08
Hmgcoar1.00 ± 0.151.17 ± 0.12‡§0.35 ± 0.05*†0.69 ± 0.05†
Cyp7a11.00 ± 0.221.91 ± 0.262.66 ± 0.35*2.00 ± 0.28
Cyp8b11.00 ± 0.110.81 ± 0.14‡1.38 ± 0.12†§0.81 ± 0.04‡
Srebp21.00 ± 0.040.87 ± 0.040.67 ± 0.04*0.77 ± 0.07*
Srebp1c1.00 ± 0.071.12 ± 0.140.85 ± 0.13§1.36 ± 0.11‡
Lxra1.00 ± 0.031.15 ± 0.050.94 ± 0.031.04 ± 0.10
Lxrb1.00 ± 0.060.86 ± 0.020.93 ± 0.030.84 ± 0.06

Abcg5 Is Rate-Limiting for Biliary Cholesterol Secretion Under BA-Stimulated Conditions.

Since biliary cholesterol secretion is lower in Sr-bI/Abcg5 dko mice under steady-state conditions, we next aimed to determine the effects of maximally stimulating biliary cholesterol secretion in a BA infusion experiment. TUDCA was infused in a stepwise manner from 150 up to 600 nmol/min. Upon TUDCA infusion bile flow increased by at least 200% without differences among the models (data not shown). Biliary BA secretion was similarly stimulated in all four models according to the BA infusion rate (Fig. 4A). Phospholipid secretion also increased upon TUDCA infusion (Fig. 4B) with Sr-bI ko mice having the highest secretion rates and Abcg5 the lowest (P < 0.01 compared to Sr-bI ko). Sr-bI/Abcg5 dko mice had an intermediate response (Fig. 4B). In wild-type controls, biliary cholesterol secretion increased in response to TUDCA infusion, while surprisingly Sr-bI ko mice even displayed a trend towards an increased responsiveness (Fig. 4C). However, in both models with absent Abcg5 expression, namely, Abcg5 ko and Sr-bI/Abcg5 dko mice, there was no appreciable impact of TUDCA infusion on biliary cholesterol secretion. These data strongly indicate that the BA-stimulated increase in biliary cholesterol secretion is fully dependent on Abcg5 expression. Figure 5 summarizes the results of the BA infusion experiment by depicting the biliary cholesterol secretion as a function of biliary BA (Fig. 5A) as well as phospholipid secretion (Fig. 5B). On the one hand, these data indicate an enhanced response in biliary cholesterol secretion to increasing biliary BA secretion rates in the Sr-bI ko model (at least P < 0.05). On the other hand, the graphs show that biliary cholesterol secretion does not respond to increasing biliary BA and phospholipid secretion rates when Abcg5 is absent (P < 0.001 compared with wild types and Sr-bI ko).

Figure 4.

Differential impact of Sr-bI and Abcg5 on biliary sterol secretion under BA-stimulated conditions. Continuous bile cannulation experiments were performed in wild-type, Sr-bI ko, Abcg5 ko, and Sr-bI/Abcg5 dko mice that were infused with TUDCA. The infusion rate was increased in a stepwise manner from 150 to 600 nmol/min as indicated; bile samples were collected for 15 minutes each as detailed in Materials and Methods. Biliary secretion rates of (A) bile acids, (B) phospholipids, and (C) cholesterol in response to the TUDCA infusion. Data are presented as means ± SEM. n = 6 for each group. a indicates statistically significant differences from wild-type, b from Sr-bI ko, c from Abcg5 ko, and d from Sr-bI/Abcg5 dko mice (at least P < 0.05).

Figure 5.

Differential impact of Sr-bI and Abcg5 on the dependency of biliary cholesterol secretion on biliary BA as well as phospholipid secretion. The combined data from the TUDCA infusion experiment shown in Fig. 4 were related to each other to show (A) biliary cholesterol secretion as a function of bile salt secretion and (B) biliary cholesterol secretion as a function of phospholipid secretion. n = 6 for each group. a indicates statistically significant differences from wildtype, b from Sr-bI ko, c from Abcg5 ko, and d from Sr-bI/Abcg5 dko mice (at least P < 0.05).

Sr-bI Contributes to RCT Independent of Abcg5/g8 Expression.

Macrophage-to-feces RCT has previously been demonstrated to be decreased in Sr-bI ko mice.27 Biliary cholesterol secretion plays a critical role in RCT.6 Since biliary cholesterol secretion was further significantly decreased in Sr-bI/Abcg5 dko compared to Abcg5 single ko, we next performed an RCT experiment comparing Abcg5 ko with Sr-bI/Abcg5 dko mice to determine the impact of Sr-bI independent of Abcg5. Plasma 3H-cholesterol tracer levels were higher in Sr-bI/Abcg5 dko mice compared to Abcg5 ko (P < 0.05 at 6 hours and 24, P = 0.059 at 48 hours, Fig. 6A). In addition, hepatic 3H-cholesterol tracer recovery at 48 hours was significantly decreased in Sr-bI/Abcg5 dko mice (Fig. 6B). Most important, however, there was a substantial reduction in total macrophage-to-feces RCT by almost 50% in Sr-bI/Abcg5 dko mice compared to the Abcg5 single ko model. These data demonstrate that the additional independent impact of SR-BI deficiency on biliary cholesterol secretion in mice with absent Abcg5 expression translates functionally into decreased RCT.

Figure 6.

Deficiency of Sr-bI decreases in vivo macrophage-to-feces reverse cholesterol transport independent of Abcg5 expression. An in vivo macrophage-to-feces RCT experiment was performed in Abcg5 ko and Sr-bI/Abcg5 dko mice following intraperitoneal injections with 3H-cholesterol-loaded primary mouse macrophage foam cells as detailed in Materials and Methods. (A) 3H-cholesterol tracer appearance in plasma 6, 24, and 48 hours after macrophage administration. (B) 3H-cholesterol tracer recovery within liver 48 hours after macrophage injection. (C) 3H-cholesterol tracer appearance in feces collected continuously from 0 to 48 hours after macrophage administration. Data are presented as means ± SEM. n = 8 for each group. Statistically significant differences from Abcg5 ko mice are indicated as *P < 0.05, **P < 0.01, ***P < 0.001.

Discussion

The present study primarily investigated the hypothesis that SR-BI contributes to the ABCG5/G8-independent part of biliary cholesterol secretion. To address this question, a novel dko mouse model for Abcg5 and Sr-bI was generated and characterized. Our data demonstrate that (1) SR-BI contributes to the ABCG5-independent part of biliary cholesterol secretion under steady-state conditions; (2) SR-BI is involved in the cellular uptake of plant sterols; (3) the increase in the BA-stimulated secretion of cholesterol into the bile is fully dependent on ABCG5 without further contribution by SR-BI; and (4) consistent with the effects on biliary cholesterol secretion, Sr-bI deficiency in Abcg5 ko mice substantially decreases in vivo RCT.

SR-BI has previously been characterized as the selective uptake receptor for HDL cholesterol into liver and steroidogenic tissues.21, 28 Consequently, deficiency in SR-BI is associated with higher HDL cholesterol levels in plasma and a shift towards larger HDL particles.29 The results of our present study demonstrate that plant sterols also accumulate in plasma in the absence of Sr-bI, suggesting that in addition to cholesterol uptake, Sr-bI is important for cellular uptake of plant sterols and stanols. Consistent with this conclusion is the significant decrease in hepatic plant sterol and stanol content when Sr-bI is not expressed.

Regarding biliary cholesterol secretion, two independent reports thus far suggested a slight but consistent decrease in biliary cholesterol secretion rates in Sr-bI ko mice.22, 30 Overall, these changes translated into a decrease in RCT in the Sr-bI ko model.27 RCT is a process that comprises the movement of cholesterol from macrophage foam cells within atherosclerotic lesions by way of HDL to the liver for final excretion into bile and feces and is thereby highly relevant for the prevention as well as potentially the treatment of atherosclerotic CVD.31 Hepatic overexpression of SR-BI, on the other hand, decreases plasma HDL cholesterol levels and increases hepatic cholesterol content due to enhanced mass cholesterol uptake from HDL.22, 23, 32 In addition, hepatic SR-BI overexpression results in a significant increase in bile cholesterol content21 as well as biliary cholesterol secretion rates.20, 22 These changes translate directly into a significant increase in in vivo RCT.27 Our results indicate that the impact of SR-BI on RCT is apparently independent of ABCG5, since a substantial reduction in the completed RCT was noted in Sr-bI/Abcg5 dko mice compared with the Abcg5 single ko model. These data suggest that, although Abcg5/g8 transport the bulk of cholesterol into bile, Sr-bI has differential functionalities independent of Abcg5/g8 that relate to biliary cholesterol secretion from a pool with a high relevance for RCT.

Importantly, we previously demonstrated that adenovirus-mediated overexpression of Sr-bI increases biliary cholesterol secretion independent of Abcg5 expression and was able to restore the reduced biliary cholesterol secretion rates of Abcg5 ko mice to the levels of normal wild-type controls.20 Our present study, on the other hand, demonstrates that under steady-state conditions there is a further significant decrease in biliary cholesterol secretion when Sr-bI is knocked out in Abcg5-deficient mice. Interestingly, though, these differences are no longer apparent under BA-stimulated conditions. On the contrary, Sr-bI ko mice even exhibit a trend towards increased biliary cholesterol secretion. The mechanism underlying this observation is currently unclear and further experimentation would be required to delineate the functional interplay between intrahepatic cholesterol pools and basolaterally and apically localized Sr-bI. In the dko mice, on the other hand, Abcg5 is clearly rate-limiting, with its absence preventing any significant increase in biliary cholesterol secretion in response to BA infusion.

ABCG5 and ABCG8 are ABC half transporters that form an obligate heterodimer for full functionality.16, 33 Mutations in either ABCG5 or ABCG8 have been characterized as the genetic substrate for sitosterolemia, a disease that is characterized by an accumulation of plant sterols within the body due to increased intestinal absorption and diminished biliary excretion in the absence of functional ABCG5/G8 expression.34 In addition, patients with sitosterolemia display accelerated atherosclerotic lesion formation.35 Interestingly, certain mutations in ABCG5/G8 have also reproducibly been associated with genetically conferred susceptibility to cholesterol gallstone formation, demonstrating that understanding the functionality of this heterodimeric transporter pair is also of prime relevance for gallstone disease.36-38

In summary, the characterization of Sr-bI/Abcg5 double-deficient mice in the present study demonstrates that under steady-state but not under BA-stimulated conditions SR-BI contributes significantly to the Abcg5/g8-independent part of biliary cholesterol secretion. Pathophysiologically, an additional deficiency in Sr-bI also translates into a substantial decrease in in vivo RCT in Abcg5 ko mice. However, when biliary cholesterol output was maximally stimulated by BA infusion, Abcg5/g8 were rate-limiting, stressing the important role of this transporter heterodimer in mass cholesterol secretion into bile. These data add to current concepts on the differential roles and functionalities of SR-BI and ABCG5/G8 in the biliary sterol secretion process, a key mechanism relevant for atherosclerotic CVD as well as gallstone disease.

Acknowledgements

We thank Dr. Han van der Want for invaluable help with electron microscopy and Rick Havinga for excellent technical assistance.

Ancillary