Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

miR-33 controls the expression of biliary transporters, and mediates statin- and diet-induced hepatotoxicity

miR-33 controls the expression of biliary transporters, and mediates statin- and diet-induced... Research Article Control of hepatic bile secretion by miR-33 miR-33 controls the expression of biliary transporters, and mediates statin- and diet-induced hepatotoxicity 1 1 1 2 3 Ryan M. Allen , Tyler J. Marquart , Carolyn J. Albert , Frederick J. Suchy , David Q.-H. Wang , 4 1,5 1,5 ´ * Meenakshisundaram Ananthanarayanan , David A. Ford ,Angel Baldan Keywords: ABCB11; ATP8B1; Bile secretion is essential for whole body sterol homeostasis. Loss-of-function cholestasis; miR-33; statins mutations in specific canalicular transporters in the hepatocyte disrupt bile flow and result in cholestasis. We show that two of these transporters, ABCB11 and ATP8B1, are functional targets of miR-33, a micro-RNA that is expressed from DOI 10.1002/emmm.201201228 within an intron of SREBP-2. Consequently, manipulation of miR-33 levels in vivo with adenovirus or with antisense oligonucleotides results in changes in bile Received January 23, 2012 Revised May 16, 2012 secretion and bile recovery from the gallbladder. Using radiolabelled cholesterol, Accepted May 23, 2012 we show that systemic silencing of miR-33 leads to increased sterols in bile and enhanced reverse cholesterol transport in vivo. Finally, we report that simvastatin causes, in a dose-dependent manner, profound hepatotoxicity and lethality in mice fed a lithogenic diet. These latter results are reminiscent of the recurrent cholestasis found in some patients prescribed statins. Importantly, pretreatment of mice with anti-miR-33 oligonucleotides rescues the hepatotoxic phenotype. Therefore, we conclude that miR-33 mediates some of the undesired, hepatotoxic GSee accompanying article http://dx.doi.org/10.1002/emmm.201201565 effects of statins. INTRODUCTION apical (canalicular) membrane of hepatocytes by three distinct transmembrane transporters: ABCB11 (ATP-binding cassette, Bile is a complex mixture of bile acids (BAs), cholesterol, sub-family B, member 11; also known as bile salt export pump, phospholipids, proteins and other organic molecules and ions BSEP), ABCG5/ABCG8 (ATP-binding cassette, sub-family G, that serves two main purposes: the solubilization of dietary member 5/8; an obligate heterodimer that facilitates cholesterol lipids in the intestine, and the removal of waste metabolites efflux) and ABCB4 (ATP-binding cassette, sub-family B, through the faeces. Cholestasis refers to a condition with member 4; also known as Multi-drug resistance gene MDR3/ impairment in bile secretion and/or flow, which leads to hepatic MDR2 in humans/mice; which pumps phospholipids, mostly injury and, in severe cases, organ failure that requires liver phosphatidylcholine, PC) (Esteller, 2008). A fourth transporter, transplantation. The major biliary lipids are secreted across the ATP8B1 (ATPase, aminophospholipid transporter, class I, type 8B, member 1), maintains the asymmetry of phospholipid species to promote the required lipid packing of the canalicular (1) Edward A. Doisy Department of Biochemistry and Molecular Biology, membrane for resistance to hydrophobic bile salts and Saint Louis University School of Medicine, Saint Louis, MO, USA canalicular membrane transport (Paulusma et al, 2006, 2008; (2) The Children’s Hospital Research Institute, University of Colorado School Ujhazy et al, 2001). Inactivating mutations in ATP8B1, ABCB11 of Medicine, Aurora, CO, USA (3) Department of Internal Medicine, Saint Louis University, Saint Louis, MO, or ABCB4 result in progressive familial intrahepatic cholestasis USA (PFIC) type 1, 2 or 3, respectively (Hori et al, 2010; Morotti et al, (4) Department of Internal Medicine-Digestive Diseases, Yale University 2011). Accordingly, these three genes are also known as FIC-1, -2 School of Medicine, New Haven, CT, USA and -3, respectively. Patients with benign recurrent intrahepatic (5) Center for Cardiovascular Research, Saint Louis University, Saint Louis, cholestasis (BRIC) also have mutations in any of the latter genes, MO, USA but the residual activity of the mutant transporter prevents *Corresponding author: Tel: þ1 314 9779227; Fax: þ1 314 9779206; the full PFIC phenotype (Hori et al, 2010; Morotti et al, 2011). E-mail: [email protected] 2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License (CC BY-NC 3.0), which permits use, distribution 882 and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Loss-of-function mutations in ABCG5 or ABCG8 result in Here we test the proposal that miR-33 also modulates hepatic sitosterolemia (Hubacek et al, 2001; Yu et al, 2002), but not bile metabolism by decreasing the expression of specific sterol in cholestasis. Mice deficient in any of these transporters transporters in the canalicular membrane of hepatocytes. Our phenocopy the human syndromes (Lammert et al, 2004; data show that conserved sequences in the 3 UTR of ABCB11 and Pawlikowska et al, 2004; Shah et al, 2010; Wang et al, 2003; ATP8B1 are functional miR-33-responsive elements (RE), and Yu et al, 2002). Nevertheless, PFIC/BRIC are thought to develop that manipulation of miR-33 levels by adenoviral over-expression from the inability to secrete BAs (ABCB11 defect) or or with antisense oligonucleotides results in altered biliary output phospholipids (ABCB4 defect), or from the inability to maintain in vivo. Hence, miR-33 limits sterol efflux in the hepatocyte canalicular membrane lipid structure (ATP8B1 defect). through both the basolateral membrane (via ABCA1) and the We (Marquart et al, 2010) and others (Gerin et al, 2010; Horie apical membrane (via ABCB11 and ATP8B1). The physiological et al, 2010; Najafi-Shoushtari et al, 2010; Rayner et al, 2010) importance of the murine miR-33 pathway is also supported in recently showed that the evolutionarily conserved miR-33 is experiments using radiolabelled cholesterol, showing that miR-33 expressed from within intron 16 of SREBP-2, and that this silencing increases the amount of labelled sterols recovered from miRNA downregulates the expression of the sterol transporter the gallbladder as well as the overall reverse cholesterol ABCA1. Manipulation of miR-33 levels in vivo results in changes transport. We also report that administration of statins, which in circulating levels of high-density lipoproteins (HDL) (Horie induce the expression of miR-33, results in decreased hepatic et al, 2010; Marquart et al, 2010; Najafi-Shoushtari et al, 2010; expression of Abcb11 and Atp8b1, but not other canalicular Rayner et al, 2010, 2011a,b). Consequently, silencing miR-33 transporters. Finally, we show that silencing miR-33 rescues the could be useful as a therapeutic intervention to increase plasma hepatotoxicity and lethality caused by co-administration of HDL in patients with hypercholesterolemia. simvastatin and a cholate-rich diet. We conclude that pharma- cologic manipulation of hepatic miR-33 levels might represent a new approach to manage certain cholestatic syndromes. RESULTS Systemic silencing of miR-33 in mice increases both bile secretion and the expression of hepatic Abcb11 and Atp8b1 In an effort to understand the in vivo physiological importance of miR-33, we injected chow-fed mice with saline, and scrambled or anti-miR-33 oligonucleotides. A week later, we collected different tissue samples. Data show a twofold increase in the volume of bile recovered from the gallbladders of anti-miR-33 animals, compared to controls (Fig 1A). However, the overall concentration of PC, cholesterol and BAs did not change in bile between the three experimental groups (Fig 1B). Next, we tested the expression of several bile-related hepatic canalicular transporters. We found that the messenger RNA (mRNA) levels of both Abcb11 and Atp8b1 were significantly increased in the livers of mice injected with anti-miR-33 oligonucleotides, compared to controls (Fig 1C). The mRNA levels of other transporters (Abcg5, Abcg8, Abcb4) and other genes involved in Figure 1. Increased bile secretion following silencing of miR-33. A. Bile recovered from the gallbladder of mice (n ¼ 5) on chow diet, injected with saline, and scrambled or anti-miR-33 oligonucleotides (5 mpk i.v., for 2 consecutive days). Animals were then kept for a week on chow and fasted overnight before sample collection. B. Levels of phosphatidylcholine (PC), cholesterol (chol) and bile acids present in gallbladder bile in the same mice. C. Relative expression of hepatic canalicular transporters in the same mice. D. A different group of mice (n ¼ 6–8) was injected as described above. A week later, mice were anesthetized, the bile duct cannulated, and hepatic bile collected for 1 h. E. Levels of bile acids, phosphatidylcholine (PC), cholesterol (chol) present in bile from the last group of mice. Data shown as mean  SEM. p< 0.05; p< 0.01. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 883 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 bile homeostasis (i.e. Shp, Cyp7a) remained unchanged in the experiment in HuH7 cells with [ C]-cholesterol, to monitor same livers (Fig 1C and Supporting Information Fig S1). As both conversion to [ C]-BAs and secretion of labelled sterols. expected, previously described miR-33 targets (Abca1 and The cells were incubated in media supplemented with 0.2% Cpt1a) were induced by the antisense treatment (Supporting albumin, but not serum, to minimize loss of the labelled Information Fig S1). To test whether the increase in bile cholesterol via ABCA1 and ABCG1. Data in Supporting recovery from the gallbladder of mice receiving anti-miR-33 Information Fig S2 show that cells overexpressing miR-33 oligonucleotides shown in Fig 1A was due to accelerated bile had a significant decrease in their ability to efflux labelled BAs, secretion, an independent group of animals were treated as but not cholesterol, compared to controls. These results suggest described above, and after an overnight fasting they were that BA secretion is impaired following miR-33 overexpression anesthetized and the common bile duct was cannulated to allow (consistent with decreased ABCB11 levels), likely resulting in the collection of hepatic bile during 60 min. Data in Fig 1D and E intracellular accumulation of BAs, which suppress further show that the secretion rates of total bile, BAs and PC were conversion of cholesterol into BAs. significantly increased in mice injected with anti-miR-33 oligonucleotides, compared to control animals. The secretion Hepatic overexpression of miR-33 decreases biliary output in rate of cholesterol also trended up in these same mice, but did cholate-fed mice not reach statistical significance. The apparently contradictory Mice fed a lithogenic diet (21% fat, 1.25% cholesterol and 0.5% results shown in Fig 1B and E might be explained by changes in cholate) exhibit disrupted bile homeostasis and, after a few water secretion/reabsorption across the canalicular, ductal and weeks, develop cholestasis (Khanuja et al, 1995; Moschetta et al, gallbladder epithelium, which are known to modulate bile 2004; Shah et al, 2010; Yu et al, 2005). Our data show that this composition (Portincasa et al, 2008). Nevertheless, the data in diet significantly decreases hepatic miR-33 levels after a week, Fig 1 demonstrate a functional role for miR-33 on bile secretion. compared to mice fed chow (Fig 3A, lanes 1–2). To test the relevance of hepatic miR-33 under pathological conditions, we ABCB11 and ATP8B1 have functional miR-33 responsive performed a diet-induced cholestasis experiment: we injected sequences in the 3(UTR mice with saline, and empty or miR-33-encoding adenovirus, Based on the previous results, we hypothesized that both switched them to the lithogenic diet the following day, and kept ABCB11 and ATP8B1 are direct targets of miR-33. Analysis of the them on this diet for 7 additional days. Predictably, the 3 UTR of these genes revealed evolutionarily conserved expression of miR-33 was increased in the livers of mice sequences that are partially complementary to miR-33 (Fig 2A receiving the miR-33 vector (Fig 3A, lanes 3–4). As expected, all and B). To test whether these sequences are functional, we the mice fed the lithogenic diet showed an enlarged gallbladder. cloned the 3 UTR of both human and mouse genes, or the However, the volume of bile recovered from mice transduced putative miR-33 responsive sequences, immediately down- with adeno-miR-33 was 45% smaller than that from mice stream of a luciferase reporter. Co-transfection with a miR-33- receiving saline or adeno-empty (Fig 3B), suggesting hepatic bile encoding plasmid confirmed that these genes indeed respond to retention. Additionally, the livers from mice receiving miR-33 miR-33 (Fig 2C and D). Hence, miR-33 expression resulted in vectors appeared slightly enlarged, resulting in a significant 40% decrease in luciferase activity when the reporter is fused increase in the liver to body mass ratio (Fig 3C). Analysis of liver to the 3 UTR or the response elements of human/mouse ATP8B1 contents revealed that overexpression of miR-33 resulted in (Fig 2C; lanes 5–8 and 11–14) or ABCB11 (Fig 2D; lanes 1–4 and significant increases in hepatic BAs, total cholesterol and 7–10). As expected, mutations that prevent the binding of the esterified cholesterol (EC), but no changes in unesterified seed sequence of the miRNA abolished the response to miR-33 cholesterol (UC) or PC (Fig 3D). For reasons that remain (Fig 2C and D; lanes 9–10 and 15–16). obscure, the amount of triglycerides (TGs) was significantly reduced in the livers of mice transduced with miR-33, compared miR-33 downregulates ABCB11 and ATP8B1 in both human to control animals (Fig 3D). The molecular basis of this latter and mouse hepatocytes effect will require further investigation. Previous reports We next determined whether the mouse and human ABCB11 showed that genes involved in fatty acid b-oxidation and TG and ATP8B1 genes are regulated following miR-33 overexpres- metabolism such as CPT1a (carnitine palmitoyltransferase 1a), sion. Hence, mouse primary hepatocytes or human HuH-7 CROT (carnitine O-octanoyltransferase), HADHB (hydroxyacyl- hepatoma cells were transduced with empty or miR-33-encoding CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase, adenovirus. Data show that the expression of human/murine beta subunit), SIRT6 (sirtuin 6) and AMPK1a (AMP-activated ABCB11 and ATP8B1 was significantly reduced in cells protein kinase 1a) are direct targets of miR-33 (Davalos et al, following overexpression of miR-33 (Fig 2E and F), while 2011; Gerin et al, 2010). From these studies it was inferred that mRNA levels of other hepatic transporters (Abcg5, Abcg8, miR-33 might function to limit fatty acid utilization in Abcb4, Abcc2, Slc10A1) remained unchanged (Fig 2E and F). In hepatocytes and other cell types, and that sustained elevated agreement with the mRNA data, protein levels for both ATP8B1 levels of miR-33 could lead to fatty liver. Based on this literature, and FXR-induced ABCB11 were downregulated following we expected to see an increase in hepatic TG contents in mice miR-33 overexpression (Fig 2G). Collectively, data from that overexpressed miR-33, compared to control animals. Figs 1 and 2 identify ATP8B1 and ABCB11 as functional direct However, data showing a significant decrease in liver TGs targets of miR-33. Additionally, we performed a pulse-chase was reproducible in 2 independent experiments (n ¼ 5/group). 884  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Figure 2. Functional miR-33 responsive elements in the 3(UTR of ATP8B1 and ABCB11. A,B. Conserved sequences in the 3 UTR of ATP8B1 and ABCB11 are partially complementary to miR-33. The element in human ATP8B1 is located 1877–1897 nt after the stop codon. The element in ABCB11 overlaps the stop codon in primates, while rodents show a conserved sequence 732–751 nt after the stop codon. C,D. Luciferase assays in HEK293 cells using the whole 3 UTR of human or murine ATP8B1 and ABCB11, or a single copy of the responsive elements (RE) identified above, or mutant responsive elements (RE ; where AATGCA was mutated to GGGTTG to prevent complementarity to the seed sequence of the miRNA), co-transfected with (closed bars) or without (open bars) a vector to overexpress miR-33. In grey, data from empty (negative control) and R33 (positive control containing a 100% match to miR-33) reporter vectors. E,F. Relative mRNA expression of canalicular transporters in primary murine hepatocytes (n ¼ 4 dishes/condition) and human HuH-7 hepatoma cells (n ¼ 3 dishes/condition) transduced 48 h with empty or miR-33 adenovirus. G. Relative protein levels in HuH7 cells transduced with empty or miR-33 adenovirus. Some cells were incubated for 16 h in the presence of FXR:RXR agonists (2 mmol/L GW4064 : 1 mmol/L 9-cis-retinoic acid) to induce ABCB11. Asterisk indicates a non-specific band. Data shown as mean  SD; p< 0.01. We can only speculate that the impact of miR-33 on lipid regulatory element binding protein 1) did not confer response to metabolism in vivo is yet to be fully elucidated, both under the miRNA (Supporting Information Fig S3B and C), and the normal diet conditions and under dietary challenge (as is the levels of nSREBP-1 were not reduced in the livers of mice case of data presented in Fig 3). Interestingly, the reduced overexpressing miR-33 (Supporting Information Fig S3D). hepatic expression of Fasn in miR-33 overexpressing mice Analysis of the bile revealed a significant decrease in BAs and (Supporting Information Fig S3A) could explain the decline in a modest increase in cholesterol in samples from adeno-miR-33 TGs in these mice. However, the reasons for such reduced Fasn mice, but no change in PC (Fig 3E). These results are consistent expression are not clear, since no candidate miR-33 elements with miR-33 controlling the expression of hepatic transporters are found in human/mouse FASN. Additionally, the potential involved in sterol mobilization. Indeed, hepatic expression of miR-33 RE found in the 3 UTR of human/mouse SREBP-1 (sterol both Abcb11 and Atp8b1 was markedly reduced in mice EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 885 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 Figure 3. Effect of miR-33 overexpression on diet-induced cholestasis. A. C57BL/6 mice (n ¼ 5–7) were kept on chow diet or lithogenic diet for 7 days (lanes 1–2). A different group of animals were transduced i.v. with empty or miR-33 adenovirus (2  10 pfu/mouse), and then switched to the lithogenic diet (lanes 3–4). After 7 days, mice were fasted overnight and killed the following morning. Data show relative levels of hepatic miR-33. B. The volume of bile recovered from the gallbladder is significantly reduced in mice transduced with miR-33. Picture shows pooled bile collected from five mice each in an independent experiment. C. Liver to body mass ratios. D. Hepatic levels of bile acids; total, unesterified (UC) and esterified (EC) cholesterol; phosphatidylcholine (PC) and triglycerides (TG). E. The amounts of bile acids (BA), cholesterol (chol) and phosphatidylcholine (PC) were determined in bile, and expressed as mol% (mol per 100 mol). Compared to mice infused with saline or Adeno-empty, the bile from animals transduced with Adeno-miR-33 showed increased amounts of cholesterol (11.6  1.1 vs.6.6  0.5 vs.5.9  0.6 mol%; miR vs. scrambled vs. saline; p ¼ 0.009), and decreased amounts of bile acids (78.6  1.7 vs. 83.7  1.0 vs. 82.1  1.4 mol%; miR vs. scrambled vs. saline; p ¼ 0.03), but no change in PC contents (9.8  1.3 vs.9.7  0.6 vs. 12.0  0.9 mol%; miR vs. scrambled vs. saline). F. Relative expression of hepatic canalicular transporters in the same mice. Data shown as mean  SD; p< 0.05; p< 0.01. either in mouse primary hepatocytes (Fig 2E), nor in human HuH7 or HepG2 cells (Supporting Information Fig S4), nor in human Hep3B cells (where exogenous miR-33 also failed to block the induction of ABCG5 by an LXR agonist (Marquart et al, 2010)). Analysis of the 3 UTR regions of the human and murine ABCG5/8 genes did not reveal any sequences with perfect complementarity to the seed sequence of miR-33 (Supporting Information Fig S4). However, when cloned downstream of a luciferase reporter, the 3 UTRs of human (but not mouse) ABCG5, and human/mouse ABCG8 conferred a very modest response to miR-33 overexpression (Supporting Information Fig S4). We speculate that these latter results are due to non- conserved, imperfect 6-mer sequences (Supporting Information Fig S4) that bind to miR-33 with low affinity. Hence, the decreased expression of Abcg5 and Abcg8 noted in Fig 3F could be the result of off-target effects due to the supra-physiological levels of miR-33 achieved using adenovirus, thus potentially limiting the interpretation of the results. Intriguingly, mice deficient in ABCB11 show decreased hepatic levels of Abcg5 and transduced with adeno-miR-33, compared to adeno-empty Abcg8 when fed a lithogenic diet (Wang et al, 2003); the (Fig 3F), while the expression of other transcripts did not molecular mechanism of this cross-talk remains unknown, but change between the three groups of mice (Fig 3F and Supporting authors speculate that it is independent on the levels/activity of Information Fig S3A). Importantly, adenoviral-mediated over- LXR (Wang et al, 2003). Whether the drop in Abcg5/8 levels in expression of miR-33 was also able to reduce the expression of the livers in Fig 3F are due to extreme levels of miR-33, or the both Abcb11 and Atp8b1 in mice fed chow (Supporting result of a yet-unknown signalling pathway that links the Information Fig S3E), suggesting that miR-33 can also modulate expression of different bile transporters will require further basal levels of these transporters in vivo under normal diet investigation. Nevertheless, results from Fig 3 and Supporting conditions. Finally, mRNA levels of the sterol transporters Information Fig S3 confirm a role for miR-33 on bile homeostasis Abcg5 and Abcg8 were also significantly reduced in the adeno- both under normal chow conditions and following dietary miR-33 group fed the lithogenic diet (Fig 3F). These latter results challenge. were unexpected, since the human and murine ABCG5 and ABCG8 genes are not direct targets of miR-33 (Supporting Effect of miR-33 on reverse cholesterol transport (RCT) Information Fig S4). Hence, we could not see repression of these RCT mobilizes extrahepatic cholesterol back to the liver for latter genes by adenoviral-mediated miR-33 overexpression subsequent secretion into bile, and final excretion in the faeces 886  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. (Khera & Rader, 2010; Rader et al, 2009; Wang et al, 2007; Wang & Rader, 2007). Some authors showed that hepatobiliary secretion is an essential component of the RCT pathway (Nijstad et al, 2011). Others have disagreed (Temel et al, 2010). Remarkably, some patients with cholestasis develop xanthomas that usually dissolve after normal bile homeostasis is restored (Emerick & Whitington, 2002; Englert et al, 2006), suggesting that impairment in bile flow ultimately leads to accumulation of sterols in patients. The previously described target of miR-33, ABCA1, plays an essential role for RCT (Wang et al, 2007). While this manuscript was in preparation, Rayner et al (2011b) showed that systemic silencing of miR-33 promotes RCT, and speculated this effect was the result of increased ABCA1- dependent cholesterol efflux. While it is imaginable that constitutive secretion of bile could funnel an increase of intrahepatic cholesterol derived from ABCA1-induced choles- terol efflux in peripheral tissues, authors have shown that increased hepatic ABCA1 activity promotes the re-secretion of cholesterol into new nascent HDL, thus preventing the mobilization of intrahepatic cholesterol towards the biliary secretory pathway (Annema et al, 2012; Nijstad et al, 2009; Tietge et al, 2008; Wiersma et al, 2009). Consequently, we hypothesized that miR-33-dependent changes in biliary trans- porters and overall bile secretion (Figs 1–3) contribute, together with ABCA1, to the effect of miR-33 on RCT. To test whether miR-33 expression alters the mobilization of extrahepatic cholesterol into the bile, we injected macrophage foam cells that were radiolabelled with [ H]-cholesterol into the peritoneal cavity of animals treated with scrambled or anti-miR-33 oligonucleotides, and followed the fate of labelled sterols for 48 h. Hepatic miR-33 expression was induced 2.1  0.3-fold in mice injected with anti-miR-33 oligos. We did not observe changes in body, liver or faecal mass (Fig 4A). Data show that the amount of labelled cholesterol in plasma, but not in liver, Figure 4. Reverse cholesterol transport is enhanced after systemic miR- increased in mice receiving anti-miR-33 oligonucleotides, 33 silencing. C57BL/6 mice (n ¼ 7) were infused i.v. with 5 mpk scrambled or compared to control mice (Fig 4B and C). A large increase in anti-miR-33 oligonucleotides for 2 consecutive days, and 5 days later received bile volume was noted, again, in the gallbladder of anti-miR-33 6 1  10 radiolabelled macrophages by i.p. injection. Mice were kept on chow mice (Fig 4A). Importantly, and validating our hypothesis, the for 48 h until sacrifice. amount of labelled sterols in the bile was also increased in these A. Different parameters in mice receiving scrambled (open bars) and anti- miR-33 (closed bars) treatment at the time of sacrifice. Notice the latter mice (Fig 4D). Finally, labelled sterols recovered from increase in bile recovery from the gallbladder. faeces were increased approximately twofold, compared to B. Percentage of total injected dpm recovered in the plasma at different control animals (Fig 4D). Subsequent lipid extraction from time points after injection of radiolabelled cells. faeces revealed that the amount of labelled neutral sterols C-E. Percentage of total injected dpm recovered at the time of sacrifice in was significantly increased in samples from mice receiving liver, bile from the gallbladder and faeces of the same mice. Data shown anti-miR-33 treatment, compared to controls (2.31  0.48 vs. as mean  SD; p< 0.05; p< 0.01. 0.94  0.18% dpm injected; p ¼ 0.025), while the recovery of faecal BAs trended upwards in the same animals, but did not reach statistical significance (3.25  0.47 vs. 2.64  0.39% dpm Statins repress the expression of miR-33 targets injected; p ¼ 0.37). Together, these data are consistent with our From a clinical perspective, it may be important that statins not hypothesis that miR-33 modulates RCT, likely through the only increase SREBP-2 expression but also they increase miR-33. combined regulation of HDL metabolism (via ABCA1) and bile Here we show that both simvastatin and atorvastatin induce metabolism (via ABCB11 and ATP8B1). Collectively, data hepatic miR-33 expression while at the same time decrease the presented in Figs 1–4 strongly suggest a concerted action of mRNA levels of miR-33 targets Abca1, Cpt1a, Abcb11 and these miR-33 targets. Still, further experiments using mice Atp8b1 (Fig 5A and B). Similar results were obtained in HuH7 deficient for each of these transporters will provide definitive cells (Fig 5C). These data strongly suggest that patients taking answers as to which specific transporter(s) are mediating the statins might have sustained, decreased levels of hepatic miR-33 effect of anti-miR-33 on RCT. targets, including transporters linked to PFIC/BRIC. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 887 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 (Supporting Information Fig S5B). Likely both the statin and the diet contributed to changes in the expression of hepatic miR-33. Importantly, we noted a dose-dependent lethality effect of the statin after mice were switched to the lithogenic diet. Thus, mice on 300 mpk simvastatin had to be euthanized on/before day 5; mice on 150 and 50 mpk simvastatin showed 50 and 100% survival rates, respectively (Fig 6A). This was paralleled by a dramatic dose-dependent increase in liver weight (Fig 6B). The livers of mice receiving 150 mpk simvastatin appeared not only enlarged, but also extremely steatotic (i.e. very pale and soft consistency) (Fig 6C); in contrast, livers from mice receiving 50 mpk simvastatin were visually indistinguishable from the livers of control mice (Fig 6C). Electrospray ionization – mass spectrometry (ESI-MS) analysis confirmed the accumulation of lipids, especially TGs, in livers from mice dosed with 150 mpk simvastatin (Fig 6D). Consistent with an intrahepatic cholestatic phenotype, liver BAs levels were also markedly elevated in these latter mice (Fig 6D). Intriguingly, ESI-MS data show a profound remodelling of ceramides in these livers: ceramides containing long-chain fatty acids (16:0, 18:0 and 20:0) were increased, while those containing very long-chain fatty acids (23:0, 24:1 and 24:0) were significantly reduced in the livers of the 150 mpk group (Supporting Information Fig S5D). The (patho)physiolo- gical significance of these changes remain obscure. Statin- induced hepatotoxicity was mirrored by increased levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), bilirubin and BAs in the blood, resulting in bright yellow plasmas (Fig 6E). Although the overall amount of bile recovered from the gallbladders did not differ between groups (Fig 6F), we noted a dose-dependent decrease in PC, small changes in bile salts concentration, and a dramatic decrease in cholesterol in samples from the 150 mpk group (Fig 6F). These changes in bile composition are consistent with abnormal expression/function of canalicular transporters. Finally, we compared the expression Figure 5. Statins induce miR-33 and reduce the mRNA expression of of selected transcripts in the livers of mice gavaged saline or specific hepatic canalicular transporters. 50 mpk simvastatin (mildly cholestatic, based on BA accumula- A, B. C57BL/6 mice (n ¼ 5) were gavaged daily with statins, and kept on chow tion) (Fig 6G). Data show that both Abcb11 and Atp8b1 mRNA diet. Samples were collected after 7 days, following an overnight were significantly reduced (40%) in the simvastatin group; fasting. C. HuH7 cells were cultured in quadruplicate for 48 h in DMEM supple- these changes were specific since the expression of other bile- mented with 2% lipoprotein-deficient serum in the presence or absence related transporters (Abcb4, Abcg5, Abcg8) did not change of simvastatin. Relative expression of specific genes shown as (Fig 6G). mean  SD; p< 0.05; p< 0.01. The results from Fig 6 support a mechanism in which statins induce miR-33, which in turn reduces the levels of both Abcb11 and Atp8b1, resulting in altered bile secretion from hepatocytes, Silencing miR-33 rescues statin- and diet-induced liver damage which ultimately leads to liver malfunction under conditions of We next examined the combined effect of statins and lithogenic lithogenic dietary challenge. To validate this model, we next diet on mice. In a preliminary experiment, we noted that tested whether silencing miR-33 could rescue the phenotype. simvastatin increased miR-33 levels in a dose-dependent Hence, mice (n ¼ 12) were injected with scrambled or anti-miR- manner in chow-fed mice (Supporting Information Fig S5A). 33 oligonucleotides, and then dosed daily with 150 mpk Then, we gavaged chow-fed mice the statin (0, 50, 150 or simvastatin and fed the lithogenic diet (Fig 7). Notably, hepatic 300 mpk) for 2 days prior to switching them to the lithogenic diet levels of miR-33 at the end of the experiment were reduced for an additional 7 days. Simvastatin was administered daily 40% in mice receiving the antisense oligonucleotides, during these latter 7 days (Fig 6A), while we monitored body compared to control animals (Supporting Information weight and food consumption (Supporting Information Fig S6A). Data show that, with one exception, all anti-miR-33 Fig S5C). At the end of the experiment, hepatic miR-33 mice survived for at least a week, losing 5% body weight; in expression was still elevated in the 50 mpk mice, but had contrast, mice injected with scrambled oligonucleotides showed decreased in the 150 mpk mice, compared to control animals 50% survival rate and significant body weight loss (Fig 7A and 888  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Figure 6. Simvastatin and lithogenic diet induce liver toxicity. C57BL/6 mice (n ¼ 6) were gavaged daily with simvastatin and fed a lithogenic diet. Samples were collected after 7 days, or when mice appeared moribund. A. Survival is hampered by simvastatin in a dose-dependent manner. B. Severe hepatomegaly in mice dosed with 150 mpk simvastatin. C. Representative macroscopic appearance (upper and middle panels), and haematoxylin and eosin staining of paraffin-embedded sections (lower panels) from the same livers. Note the abnormally swollen cells and pale (i.e. steatotic) appearance of the livers in the 150 mpk group. D. Specific hepatic lipids as determined by ESI-MS, and normalized to tissue weight. FFA, free fatty acids; DAG, diacylglycerides; TAG, triacylglycerides; PC, phospha- tidylcholine; UC, unesterified cholesterol; CE, cholesterol esters. p< 0.01 versus saline. E. Representative samples of plasma, and levels of circulating alanine aminotransferase (ALT), aspartate aminotransferase (AST), bile acids and bilirubin. F. Bile was recovered from the gallbladder, pooled and the contents of phosphatidylcholine (PC), cholesterol (c) and bile acids (BA) determined with colorimetric kits. G. Relative expression of hepatic canalicular transporters (upper panel) and other genes involved in bile acid and sterol homeostasis (bottom panel) in samples from mice treated with 0 or 50 mpk simvastatin. Data shown as mean  SD. p< 0.01. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 889 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 Figure 7. Silencing miR-33 rescues the liver damage induced by simvastatin and lithogenic diet. C57BL/6 mice (n ¼ 12) were injected i.v. with 5 mpk scrambled or anti-miR-33 oligonucleotides for two consecutive days, and then treated as in Fig 5. A. Survival curves. B. Body weight, expressed as % compared to mass on the day of first injection. C. Representative macroscopic appearance of the livers, and haematoxylin and eosin staining of paraffin-embedded sections. D. Liver to total body mass ratios. p< 0.01 versus moribund mice. E. Representative samples of plasma. Note the normalization in colour in the anti-miR-33 group. F. Amounts of specific hepatic lipids as determined by ESI-MS, normalized to tissue weight (n ¼ 4). Acronyms for lipid classes as defined in Fig 5. p< 0.01 versus scrambled. G. Relative expression of hepatic canalicular transporters (upper panel) and other genes involved in bile acid and sterol homeostasis (bottom panel) (n ¼5for scrambled treatment; n ¼ 10 for anti-miR-33 treatment). Data shown as mean  SEM. p< 0.01 versus moribund scrambled mice; p< 0.05 versus surviving scrambled mice. 890  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. B). Remarkably, the livers from the anti-miR-33 group appeared non-steatotic (Fig 7C), and the liver/body mass ratio was significantly lower than in the scrambled group that succumbed to the treatment (Fig 7D). Additionally, the plasmas from mice receiving anti-miR-33 appeared normal as compared to the yellow-coloured plasma from mice treated with scrambled oligonucleotides (Fig 7E). Rescue of the statin- and diet-induced phenotype was also evident when we analyzed hepatic lipids: anti-miR-33-treated animals showed a significant normalization in lipid contents, compared to control mice (Fig 7F and Supporting Information Fig S6B and C). Finally, we studied the hepatic mRNA levels of selected genes (Fig 7G). Interestingly, the expression of all canalicular transporters, with the exception of Atp8b1, was severely decreased in mice that succumbed to the treatment. Additionally, the levels of Abcg5 and Abcg8 were, for reasons that are not clear, highly variable within each group of mice (Fig 7G). When only considering those mice that survived 7 days on the diet, treatment with anti-miR-33 oligonucleotides resulted in significant increased expression of Atp8b1, but not of Abcb11 or any other canalicular transporter (Fig 7G). Perhaps the expression of Abcb11 is already maximal in these livers, due to the diet-induced activation of FXR. In general, the mRNA levels of bile-related genes in surviving mice in the scrambled group were similar to those in the antisense group (Fig 7G), suggesting that the expression of these genes is critical for survival. We also analyzed the mRNA expression of selected hepatic Phase I and II detoxifying genes (Supporting Information Fig S6D). Again, we found large differences in the expression of most of these genes Figure 8. miR-33 limits the mobilization of sterols in hepatocytes through within each experimental group, making the interpretation of both the sinusoidal and the canalicular membrane. the data difficult. We speculate that the increased survival of A. miR-33 mediates the cross-talk between the SREBP-2, LXR and FXR mice receiving the anti-miR-33 treatment is likely due to pathways by directly modulating the expression of ATP8B1, ABCB11, ABCA1 complex, coordinated changes in the expression of several and ABCG1 ( in mice, but not in humans). Data suggest that miR-33 might genes, which ultimately results in the accelerated clearance of also indirectly affect the expression of ABCG5/8 (see main text for details). Decreased ATP8B1 activity in human, but not mouse, hepatocytes results bile and drug and/or diet-derived toxic metabolites. Never- in PKC-dependent inactivating phosphorylation of FXR. theless, data in Fig 7 show conclusively that miR-33 mediates B. During episodes of low intracellular cholesterol, or following treatment statin- and diet-induced hepatotoxicity. Additional experiments with statin drugs, miR-33 is transcriptionally induced and reduces the using mice deficient in ABCB11 and/or ATP8B1 will provide expression of sterol transporters. clues about the relative contribution of each specific transporter to statin-induced, miR-33-dependent hepatotoxicity. The exact hepatoprotective mechanism of anti-miR-33 on statin- and diet- induced toxicity is yet to be determined. We speculate that data (Marquart et al, 2010 and this report) and that from other increased bile flow due to de-repression of miR-33 targets such laboratories (Gerin et al, 2010; Horie et al, 2010; Najafi- as Abcb11 and/or Atp8b1 contribute to the protective effect, but Shoushtari et al, 2010; Rayner et al, 2010) suggest that miR-33 whether other bile-independent pathways are also necessary for plays a central role in coordinating those three pathways hepatoprotection remains to be established. (Fig 8A). We hypothesize that miR-33 evolved to limit the efflux of hepatocyte sterols, both through the basolateral and apical membranes, in situations of low intracellular cholesterol that DISCUSSION induce SREBP-2 (Fig 8B). Bile acids bind and activate FXR, which in turn promotes Hepatic sterol homeostasis is a complex process that encom- increased bile secretion (by inducing ABCB11 and ABCB4) and passes de novo synthesis of cholesterol, secretion and uptake of reduced BAs synthesis (by indirectly repressing CYP7a) lipoproteins, conversion of cholesterol to BAs, and bile (Gadaleta et al, 2010; Zhang & Edwards, 2008). Interestingly, secretion. Extensive literature shows that three transcription FXR targets were found to be reduced in patients with PFIC-1 or factors (sterol regulatory element binding protein 2, SREBP-2; following siRNA against ATP8B1 in humans hepatocytes liver-X-receptor, LXR; and farnesoid X-activated receptor, FXR) (Alvarez et al, 2004; Chen et al, 2004; Koh et al, 2009; orchestrate complementary aspects of sterol metabolism. Our Martinez-Fernandez et al, 2009), leading to the proposal that EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 891 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 FXR activity can be modulated by ATP8B1, perhaps via protein both transcripts are decreased by fasting and induced by kinase C (Chen et al, 2010; Frankenberg et al, 2008). Other refeeding (but Srebp-1 is more potently induced than -2) (Horton investigators, however, failed to reproduce these observations et al, 1998; Bennett et al, 2008), and only Srebp-2 is induced by (Cai et al, 2009). Nevertheless, multiple studies in patients with statins (Bennett et al, 2008). Additional studies using primates PFIC/BRIC-1 and -2 support a critical role for both ATP8B1 and or humanized mice (in which transgenic miR-33b is expressed ABCB11 on bile secretion (Davit-Spraul et al, 2010; Harris et al, from an intron of Srebp-1) will be necessary to study the impact 2005; Kubitz et al, 2005; Morotti et al, 2011; van der Woerd et al, of SREBP-1-derived miR-33 on bile metabolism. Nevertheless, 2010). we speculate that anti-miR-33 oligonucleotides might be useful Since SREBP-2/miR-33 are transcriptionally induced follow- to manage patients who develop BRIC as a consequence of ing treatment with statins (Marquart et al, 2010; Najafi- partial loss of expression and/or function of ABCB11 or ATP8B1. Shoushtari et al, 2010; Rayner et al, 2010), we hypothesized that miR-33 might account for some of the (side) effects of these drugs. Several reports show that certain patients develop MATERIALS AND METHODS cholestasis following prescription of statins (Batey & Harvey, 2002; de Castro et al, 2006; Merli et al, 2010; Rahier et al, 2008; Materials Ridruejo & Mando, 2002; Torres et al, 2002). While the Vectors encoding miR-33 were previously described (Marquart et al, molecular events in the livers of these patients are unknown, 2010). HEK293 and HuH-7 cells (ATCC) were maintained in DMEM this clinical setting is consistent with our model (Fig 8) in which supplemented with 10% fetal bovine serum (FBS). Luciferase reporter statin-induced miR-33 represses both ABCB11 and ATP8B1, constructs containing miR-33 response elements or the 3 UTR of thus decreasing bile secretion and eventually leading to human/mouse ATP8B1, ABCB11, ABCG5 and ABCG8 were generated intrahepatic cholestasis. We recognize, however, that our by cloning each sequence into an XbaI site in pGL3-Promoter experimental model is extreme due to the relatively high dose (Promega). 3 UTR fragments were amplified from human or mouse of the statin used and the fact that patients do not normally genomic DNA using Platinum Pfx (Invitrogen). Control (5 -TCCTAGAAA- consume lithogenic diets (i.e. a high fat, high cholesterol and GAGTAGA) and anti-miR-33 (5 -TGCAACTACAATGCA), locked nucleic high BA diet). It is intriguing to speculate, however, that these acid (LNA) oligonucleotides were kindly provided by Miragen patients might carry specific polymorphisms in ABCB11 or Therapeutics Inc. (Boulder, CO). Control oligonucleotides are designed ATP8B1 that would make them more susceptible to statin- to not target any mouse RNA annotated in NCBI databases. induced miR-33 effects. In general, most physicians are cautious Mice received 5 mg/kg (in 100 mL saline) via tail vein injection for 3 to prescribe statins to patients with underlying liver disease, and consecutive days, unless otherwise stated. the effect of these drugs on patients with primary biliary cirrhosis is controversial (Abu Rajab & Kaplan, 2010; Stanca Murine studies et al, 2008; Stojakovic et al, 2007, 2010). Nevertheless, our data C57BL/6 mice were obtained from NCI/Charles River, and maintained clearly demonstrate a dose-dependent effect of simvastatin on in a 12 h/12 h light/dark cycle with unlimited access to food and diet-induced hepatotoxicity and cholestasis in mice. It is water. Where indicated, animals (8–10 week old, n ¼ 6 per group, important to stress the fact that mice tolerate 300 mpk statins unless noted otherwise) were infused via tail vein with adenoviral with a concomitant high fat, high cholesterol diet for months vectors (2  10 pfu) or antisense oligonucleotides (5 mpk, 3 con- (reviewed in Zadelaar et al, 2007), but to our knowledge no secutive days). Where indicated, mice were fed a lithogenic diet previous studies addressed the interaction between statins and (Purina 5A8E). Tissues and bile were collected following overnight cholate feeding. Our data offer conclusive evidence that, at least fasting. RCT experiments were performed as described (Wang et al, in mice, statins potentiate diet-induced cholestasis. From a 2007). Briefly, mice were injected with oligonucleotides (5 mpk) on 6 3 mechanistic perspective, our data show that the phenotype is days 1, 2, 8 and 9, and on day 15 injected i.p. with 1–1.5  10 [ H]- largely rescued by anti-miR-33 oligonucleotides. cholesterol-loaded bone marrow-derived macrophages. Blood samples Authors suggested that a therapy that combines statins and were collected 6, 24 and 48 h thereafter. Liver and bile, and the faeces anti-miR-33 oligonucleotides may be useful in hyperlipidemic produced during the last 48 h were collected and flash-frozen until patients by reversing statin-induced, miR-33-mediated repres- used. The amount of radiolabelled sterols was determined by sion of ABCA1, which ultimately would increase plasma HDL scintillation. Where indicated, mice were dosed daily for 7 days with (Horie et al, 2010; Marquart et al, 2010; Najafi-Shoushtari et al, saline or statin by oral gavage, and killed following an overnight 2010; Rayner et al, 2010). Our new data show that miR-33 plays fasting. Bile secretion was measured in anesthetized mice following a key role in the hepatic response to statins by coordinating the common bile duct cannulation, which allowed the continued expression of several sterol transporters, and that disruption of collection of hepatic bile by gravity. Body temperature was maintained the miR-33 pathway prevents statin-induced hepatotoxicity. during surgery and bile collection at 37  0.58C. All studies were Data support the hypothesis that miR-33 controls whole-body approved by the IACUC at SLU. sterol homeostasis by affecting both HDL biogenesis (via ABCA1), and bile secretion (via ABCB11 and ATP8B1). Primary hepatocytes Primates, but not rodents, express a second miR-33 gene Cells were isolated from 8–10-week-old, male C57BL/6 mice fed chow, (miR-33b) from an intron of SREBP-1. Importantly, SREBPs are using Perfusion and Digest buffers (Invitrogen). Cells were resuspended differentially regulated by dietary challenges or statin treatment: in William’s E Medium (Invitrogen) supplemented with Plating 892  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. The paper explained PROBLEM: Patients with hypercholesterolemia are usually prescribed with anti-miR-33 oligonucleotides. These latter results suggest statins. We and others reported on miR-33, an intragenic micro- that miR-33 mediates, at least in part, the hepatotoxic effects of RNA that is induced by statins and regulates HDL-cholesterol by statins. modulating the expression of the sterol transporter ABCA1 in the liver. Hence, it was proposed that anti-miR-33 oligonucleotides IMPACT: might be used as a new therapeutic approach to manage Statins induce the expression of miR-33. Patients taking statins hypercholesterolemic patients. However, whether miR-33 also likely have sustained, elevated levels of hepatic miR-33. These controls the expression of other genes important for sterol drugs may decrease the expression of hepatic sterol transporters metabolism is unknown. Here we test the hypothesis that miR-33 indirectly, via miR-33. Several clinical reports showed patients modulates the expression of transporters involved in hepato- who recurrently develop intrahepatic cholestasis when put on biliary secretion. Additionally, we study the role of miR-33 on statins, but recover after statin withdrawal. Cholestasis results statin-induced liver toxicity. from decreased bile secretion and/or flow. Patients with mutations in ATP8B1 and ABCB11 develop PFIC/BRIC-1 and -2, RESULTS: respectively. The data showing the coordinated regulation of MiR-33 extends the function of its host SREBP-2 by reducing sterol transporters by miR-33 and its impact on both HDL and biliary secretion through direct repression of ABCB11 and bile metabolism, add to an ever-growing collection of studies ATP8B1. We identify conserved sequences in the 3 UTR of these showing that miRNAs function as critical fine-tuners in multiple genes responsible for this control, and show that changes in normal and disease-related biological processes. We hypothesize hepatic levels of miR-33 alter both bile secretion rates and bile that the increased expression in ATP8B1 and ABCB11 mediated recovery from the gallbladder. The combination of statins and a by an anti-miR-33 treatment might functionally overcome the lithogenic diet results in a dramatic phenotype (hepatomegaly, partial loss of activity of these transporters observed in BRIC-1 liver steatosis, cholestasis and lethality) in mice, that is rescued and BRIC-2 patients, respectively. Supplements (Invitrogen), plated in 12- or 6-well BioCoat Collagen I Immune complexes were detected with horseradish peroxidase (HRP)- plates (BD), and incubated at 378C and 5% CO for 6 h. Then, the conjugated secondary antibodies (1:5000; Bio-Rad). Due to lack of media was switched to William’s E supplemented with Maintenance specificity we were not able to use these (or different) antibodies to Supplements (Invitrogen). Where indicated, cells were transduced with detect ABCB11 and ATP8B1 in liver samples. Adeno-empty or Adeno-miR-33 adenovirus (MOI 3). Total RNA were extracted 72 h after transduction. Luciferase reporter assays Transfection of Hek293 cells was performed in triplicate in 24-well RNA analysis plates by the calcium phosphate method. Luciferase activity was RNA was isolated from cells or livers with Trizol. Complementary DNAs measured 48 h later using the Luciferase Assay System (Promega), and (cDNAs) were generated from 2 mg of DNase1-treated RNA using normalized to b-galactosidase activity to correct for small changes in Superscript III (Invitrogen) and random hexamers. Real-time PCR was transfection efficiency. done with Power SybrGreen reagent (Applied Biosystems), using a LightCycler-480 (Roche). Primer sets are available on request. Values Plasma analysis were normalized to 36B4 and calculated using the comparative C Circulating levels of ALT, AST, bilirubin and total BAs were determined method. The expression of miR-33 was normalized to U6, using by Advanced Veterinary Laboratory (Saint Louis, MO). MiRCURY miRNA assays (Exiqon). Lipid analysis Protein analysis Liver (50 mg) was homogenized in 500 mL of PBS, and lipids were Protein extracts were obtained from cells or tissues, as described extracted from 100 mL of the homogenate in the presence of internal (Marquart et al, 2010). Fifty micrograms of protein were resolved in standards for each lipid class (Bligh & Dyer, 1959). Similarly, bile and 4–12% Bis–Tris gels, transferred to PVDF membranes, and probed with plasma were extracted in the presence of internal standards for each antibodies for ABCB11 (1:500; a gift from Dr. Renxue Wang from lipid class. Lipid species (phospholipids; TGs; cholesterol esters, CE and the British Columbia Cancer Research Center), ATP8B1 (1:200; SCBT ceramides) were quantified directly from lipid organic extracts using sc-67712), ABCA1 (1:1000; Novus NB400-105), ABCG5 (1:200; Novus shotgun lipidomics based on class separation by MS/MS specific NBP1-95209), FXR (1:200; SCBT sc-13063), SREBP1 (1:100; SCBT sc- methods (Ford et al, 2008; Han & Gross, 2001, 2005). Bile and liver 13551), b-actin (1:5000; SCBT sc-130656) and a-tubulin (1:1000; concentrations of cholesterol, bile salts and PC were also determined Sigma T5168), in TBS-Tween20 containing 4% non-fat dry milk. using enzymatic kits from Wako Chemicals. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 893 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 Davalos A, Goedeke L, Smibert P, Ramirez CM, Warrier NP, Andreo U, Cirera- Author contributions Salinas D, Rayner K, Suresh U, Pastor-Pareja JC, et al (2011) miR-33a/b RMA conceived the hypothesis, designed and performed contribute to the regulation of fatty acid metabolism and insulin signaling. experiments, analyzed the data and wrote the manuscript. Proc Natl Acad Sci USA 108: 9232-9237 TJM designed, performed and analyzed the RCT experiment. Davit-Spraul A, Fabre M, Branchereau S, Baussan C, Gonzales E, Stieger B, CJA, RMA and DAF performed and analyzed the lipidomics Bernard O, Jacquemin E (2010) ATP8B1 and ABCB11 analysis in 62 children studies. MA and FJS performed mRNA experiments in HuH-7 with normal gamma-glutamyl transferase progressive familial intrahepatic cells. DQ-HW performed common bile duct cannulations for cholestasis (PFIC): phenotypic differences between PFIC1 and PFIC2 and natural history. Hepatology 51: 1645-1655 secretion experiments. AB conceived the hypothesis, designed de Castro ML, Hermo JA, Baz A, de Luaces C, Perez R, Clofent J (2006) Acute experiments and wrote the manuscript. cholestatic hepatitis after atorvastatin reintroduction. Gastroenterol Hepatol 29: 21-24 Emerick KM, Whitington PF (2002) Partial external biliary diversion for Acknowledgements intractable pruritus and xanthomas in Alagille syndrome. Hepatology 35: We thank Erin Touchette for help with i.v. injections, Dr. Eva 1501-1506 van Rooij (miRagen Therapeutics) for providing LNA oligos, Englert C, Grabhorn E, Burdelski M, Ganschow R (2006) Liver transplantation in children with Alagille syndrome: indications and outcome. Pediatr and Dr. Robert Brown (Memorial University of Newfoundland) Transplant 10: 154-158 for discussions for RCT experiments. This work was supported Esteller A (2008) Physiology of bile secretion. World J Gastroenterol 14: 5641- in part by AHA Grant 11SDG5030012 (to A.B.), and NIH Grants HL107794 (to A.B.), HL074214, HL098907 and RR019232 (to Ford DA, Monda JK, Brush RS, Anderson RE, Richards MJ, Fliesler SJ (2008) D.A.F.) and DK084434 (to F.J.S.). R.M.A. is the recipient of an Lipidomic analysis of the retina in a rat model of Smith–Lemli–Opitz AHA Predoctoral Fellowship (11PRE7240026). syndrome: alterations in docosahexaenoic acid content of phospholipid molecular species. J Neurochem 105: 1032-1047 Frankenberg T, Miloh T, Chen FY, Ananthanarayanan M, Sun AQ, Supporting Information is available at EMBO Molecular Balasubramaniyan N, Arias I, Setchell KD, Suchy FJ, Shneider BL (2008) The Medicine online. membrane protein ATPase class I type 8B member 1 signals through protein kinase C zeta to activate the farnesoid X receptor. Hepatology 48: 1896- Conflict of interest statement: RMA, TJM and AB are pursuing a patent related to this work. The other authors declare Gadaleta RM, van Mil SW, Oldenburg B, Siersema PD, Klomp LW, van Erpecum that they have no conflict of interest. KJ (2010) Bile acids and their nuclear receptor FXR: relevance for hepatobiliary and gastrointestinal disease. Biochim Biophys Acta 1801: 683-692 Gerin I, Clerbaux LA, Haumont O, Lanthier N, Das AK, Burant CF, Leclercq IA, References MacDougald OA, Bommer GT (2010) Expression of miR-33 from an SREBP2 Abu Rajab M, Kaplan MM (2010) Statins in primary biliary cirrhosis: are they intron inhibits cholesterol export and fatty acid oxidation. J Biol Chem 285: safe? Dig Dis Sci 55: 2086-2088 33652-33661 Alvarez L, Jara P, Sanchez-Sabate E, Hierro L, Larrauri J, Diaz MC, Camarena C, Han X, Gross RW (2001) Quantitative analysis and molecular species De La Vega A, Frauca E, Lopez-Collazo E et al (2004) Reduced hepatic fingerprinting of triacylglyceride molecular species directly from lipid expression of Farnesoid X Receptor in hereditary cholestasis associated to extracts of biological samples by electrospray ionization tandem mass mutation in ATP8B1. Hum Mol Genet 13: 2451-2460 spectrometry. Anal Biochem 295: 88-100 Annema W, Dikkers A, Freark de Boer J, Gautier T, Rensen PC, Rader DJ, Tietge Han X, Gross RW (2005) Shotgun lipidomics: electrospray ionization mass UJ (2012) ApoE promotes hepatic selective uptake but not RCT due to spectrometric analysis and quantitation of cellular lipidomes directly from increased ABCA1-mediated cholesterol efflux to plasma. J Lipid Res 53: crude extracts of biological samples. Mass Spectrom Rev 24: 367-412 929-940 Harris MJ, Le Couteur DG, Arias IM (2005) Progressive familial intrahepatic Batey RG, Harvey M (2002) Cholestasis associated with the use of pravastatin cholestasis: genetic disorders of biliary transporters. J Gastroenterol sodium. Med J Aust 176: 561 Hepatol 20: 807-817 Bennett MK, Seo YK, Datta S, Shin DJ, Osborne TF (2008) Selective binding of Hori T, Nguyen JH, Uemoto S (2010) Progressive familial intrahepatic sterol regulatory element-binding protein isoforms and co-regulatory cholestasis. Hepatobiliary Pancreat Dis Int 9: 570-578 proteins to promoters for lipid metabolic genes in liver. J Biol Chem 283: Horie T, Ono K, Horiguchi M, Nishi H, Nakamura T, Nagao K, Kinoshita M, 15628-15637 Kuwabara Y, Marusawa H, Iwanaga Y, et al (2010) MicroRNA-33 encoded by Bligh EG, Dyer WJ (1959) A rapid method of total lipid extraction and an intron of sterol regulatory element-binding protein 2 (Srebp2) regulates purification. Can J Med Sci 37: 911-917 HDL in vivo. Proc Natl Acad Sci USA 107: 17321-17326 Cai SY, Gautam S, Nguyen T, Soroka CJ, Rahner C, Boyer JL (2009) ATP8B1 Horton JD, Bashmakov Y, Shimomura I, Shimano H (1998) Regulation of sterol deficiency disrupts the bile canalicular membrane bilayer structure in regulatory element binding proteins in livers of fasted and refed mice. Proc hepatocytes, but FXR expression and activity are maintained. Natl Acad Sci USA 95: 5987-5992 Gastroenterology 136: 1060-1069 Hubacek JA, Berge KE, Cohen JC, Hobbs HH (2001) Mutations in ATP-cassette Chen F, Ananthanarayanan M, Emre S, Neimark E, Bull LN, Knisely AS, binding proteins G5 (ABCG5) and G8 (ABCG8) causing sitosterolemia. Hum Strautnieks SS, Thompson RJ, Magid MS, Gordon R, et al (2004) Mutat 18: 359-360 Progressive familial intrahepatic cholestasis, type 1, is associated Khanuja B, Cheah YC, Hunt M, Nishina PM, Wang DQ, Chen HW, Billheimer JT, with decreased farnesoid X receptor activity. Gastroenterology 126: Carey MC, Paigen B (1995) Lith1, a major gene affecting cholesterol 756-764 gallstone formation among inbred strains of mice. Proc Natl Acad Sci USA Chen F, Ellis E, Strom SC, Shneider BL (2010) ATPase Class I Type 8B Member 1 92: 7729-7733 and protein kinase C zeta induce the expression of the canalicular bile salt Khera AV, Rader DJ (2010) Future therapeutic directions in reverse cholesterol export pump in human hepatocytes. Pediatr Res 67: 183-187 transport. Curr Atheroscler Rep 12: 73-81 894  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Koh S, Takada T, Kukuu I, Suzuki H (2009) FIC1-mediated stimulation of FXR human primates raises plasma HDL and lowers VLDL triglycerides. Nature activity is decreased with PFIC1 mutations in HepG2 cells. J Gastroenterol 478: 404-407 44: 592-600 Rayner KJ, Sheedy FJ, Esau CC, Hussain FN, Temel RE, Parathath S, van Gils JM, Kubitz R, Keitel V, Haussinger D (2005) Inborn errors of biliary canalicular Rayner AJ, Chang AN, Suarez Y, et al (2011b) Antagonism of miR-33 in mice transport systems. Methods Enzymol 400: 558-569 promotes reverse cholesterol transport and regression of atherosclerosis. J Lammert F, Wang DQ, Hillebrandt S, Geier A, Fickert P, Trauner M, Matern S, Clin Invest 121: 2921-2931 Paigen B, Carey MC (2004) Spontaneous cholecysto- and hepatolithiasis in Ridruejo E, Mando OG (2002) Acute cholestatic hepatitis after reinitiating Mdr2-/- mice: a model for low phospholipid-associated cholelithiasis. treatment with atorvastatin. J Hepatol 37: 165-166 Hepatology 39: 117-128 Shah S, Sanford UR, Vargas JC, Xu H, Groen A, Paulusma CC, Grenert JP, Marquart TJ, Allen RM, Ory DS, Baldan A (2010) miR-33 links SREBP-2 Pawlikowska L, Sen S, Elferink RP, et al (2010) Strain background modifies induction to repression of sterol transporters. Proc Natl Acad Sci USA 107: phenotypes in the ATP8B1-deficient mouse. PLoS One 5: e8984 12228-12232 Stanca CM, Bach N, Allina J, Bodian C, Bodenheimer H, Jr., Odin JA (2008) Martinez-Fernandez P, Hierro L, Jara P, Alvarez L (2009) Knockdown of ATP8B1 Atorvastatin does not improve liver biochemistries or Mayo Risk Score in expression leads to specific downregulation of the bile acid sensor FXR in primary biliary cirrhosis. Dig Dis Sci 53: 1988-1993 HepG2 cells: effect of the FXR agonist GW4064. Am J Physiol Gastrointest Stojakovic T, Putz-Bankuti C, Fauler G, Scharnagl H, Wagner M, Stadlbauer V, Liver Physiol 296: G1119-G1129 Gurakuqi G, Stauber RE, Marz W, Trauner M (2007) Atorvastatin in patients Merli M, Bragazzi MC, Giubilo F, Callea F, Attili AF, Alvaro D (2010) with primary biliary cirrhosis and incomplete biochemical response to Atorvastatin-induced prolonged cholestasis with bile duct damage. Clin ursodeoxycholic acid. Hepatology 46: 776-784 Drug Investig 30: 205-209 Stojakovic T, Claudel T, Putz-Bankuti C, Fauler G, Scharnagl H, Wagner M, Morotti RA, Suchy FJ, Magid MS (2011) Progressive familial intrahepatic Sourij H, Stauber RE, Winkler K, Marz W, et al (2010) Low-dose atorvastatin cholestasis (PFIC) type 1, 2, and 3: a review of the liver pathology findings. improves dyslipidemia and vascular function in patients with primary Semin Liver Dis 31: 3-10 biliary cirrhosis after one year of treatment. Atherosclerosis 209: 178-183 Moschetta A, Bookout AL, Mangelsdorf DJ (2004) Prevention of cholesterol Temel RE, Sawyer JK, Yu L, Lord C, Degirolamo C, McDaniel A, Marshall S, Wang gallstone disease by FXR agonists in a mouse model. Nat Med 10: N, Shah R, Rudel LL, et al (2010) Biliary sterol secretion is not required for 1352-1358 macrophage reverse cholesterol transport. Cell Metab 12: 96-102 Najafi-Shoushtari SH, Kristo F, Li Y, Shioda T, Cohen DE, Gerszten RE, Naar AM Tietge UJ, Nijstad N, Havinga R, Baller JF, van der Sluijs FH, Bloks VW, Gautier T, (2010) MicroRNA-33 and the SREBP host genes cooperate to control Kuipers F (2008) Secretory phospholipase A2 increases SR-BI-mediated cholesterol homeostasis. Science 238: 1566-1569 selective uptake from HDL but not biliary cholesterol secretion. J Lipid Res Nijstad N, Wiersma H, Gautier T, van der Giet M, Maugeais C, Tietge UJ (2009) 49: 563-571 Scavenger receptor BI-mediated selective uptake is required for the Torres M, Sobrino J, Asensio C, Lopez D (2002) Acute cholestatic hepatitis remodeling of high density lipoprotein by endothelial lipase. J Biol Chem induced by cerivastatin. Med Clin 118: 717 284: 6093-6100 Ujhazy P, Ortiz D, Misra S, Li S, Moseley J, Jones H, Arias IM (2001) Familial Nijstad N, Gautier T, Briand F, Rader DJ, Tietge UJ (2011) Biliary sterol intrahepatic cholestasis 1: studies of localization and function. Hepatology secretion is required for functional in vivo reverse cholesterol transport in 34: 768-775 mice. Gastroenterology 140: 1043-1051 van der Woerd WL, van Mil SW, Stapelbroek JM, Klomp LW, van de Graaf SF, Paulusma CC, Groen A, Kunne C, Ho-Mok KS, Spijkerboer AL, Rudi de Waart D, Houwen RH (2010) Familial cholestasis: progressive familial intrahepatic Hoek FJ, Vreeling H, Hoeben KA, van Marle J, et al (2006) Atp8b1 deficiency cholestasis, benign recurrent intrahepatic cholestasis and intrahepatic in mice reduces resistance of the canalicular membrane to hydrophobic bile cholestasis of pregnancy. Best Pract Res Clin Gastroenterol 24: 541-553 salts and impairs bile salt transport. Hepatology 44: 195-204 Wang X, Rader DJ (2007) Molecular regulation of macrophage reverse Paulusma CC, Folmer DE, Ho-Mok KS, de Waart DR, Hilarius PM, Verhoeven AJ, cholesterol transport. Curr Opin Cardiol 22: 368-372 Oude Elferink RP (2008) ATP8B1 requires an accessory protein for Wang R, Lam P, Liu L, Forrest D, Yousef IM, Mignault D, Phillips MJ, Ling V endoplasmic reticulum exit and plasma membrane lipid flippase activity. (2003) Severe cholestasis induced by cholic acid feeding in knockout mice Hepatology 47: 268-278 of sister of P-glycoprotein. Hepatology 38: 1489-1499 Pawlikowska L, Groen A, Eppens EF, Kunne C, Ottenhoff R, Looije N, Knisely AS, Wang X, Collins HL, Ranalletta M, Fuki IV, Billheimer JT, Rothblat GH, Tall AR, Killeen NP, Bull LN, Elferink RP, et al (2004) A mouse genetic model for Rader DJ (2007) Macrophage ABCA1 and ABCG1, but not SR-BI, promote familial cholestasis caused by ATP8B1 mutations reveals perturbed bile salt macrophage reverse cholesterol transport in vivo. J Clin Invest 117: 2216- homeostasis but no impairment in bile secretion. Hum Mol Genet 13: 881- 2224 892 Wiersma H, Gatti A, Nijstad N, Kuipers F, Tietge UJ (2009) Hepatic SR-BI, not Portincasa P, Palasciano G, Svelto M, Calamita G (2008) Aquaporins in the endothelial lipase, expression determines biliary cholesterol secretion in hepatobiliary tract. Which, where and what they do in health and disease. mice. J Lipid Res 50: 1571-1580 Eur J Clin Invest 38: 1-10 Yu L, Hammer RE, Li-Hawkins J, Von Bergmann K, Lutjohann D, Cohen JC, Rader DJ, Alexander ET, Weibel GL, Billheimer J, Rothblat GH (2009) The role of Hobbs HH (2002) Disruption of Abcg5 and Abcg8 in mice reveals their reverse cholesterol transport in animals and humans and relationship to crucial role in biliary cholesterol secretion. Proc Natl Acad Sci USA 99: atherosclerosis. J Lipid Res 50 Suppl: S189-S194 16237-16242 Rahier JF, Rahier J, Leclercq I, Geubel AP (2008) Severe acute cholestatic Yu L, Gupta S, Xu F, Liverman AD, Moschetta A, Mangelsdorf DJ, Repa JJ, Hobbs hepatitis with prolonged cholestasis and bile-duct injury following HH, Cohen JC (2005) Expression of ABCG5 and ABCG8 is required for atorvastatin therapy: a case report. Acta Gastroenterol Belg 71: 318-320 regulation of biliary cholesterol secretion. J Biol Chem 280: 8742-8747 Rayner KJ, Suarez Y, Davalos A, Parathath S, Fitzgerald ML, Tamehiro N, Fisher Zadelaar S, Kleemann R, Verschuren L, de Vries-Van der Weij J, van der Hoorn J, EA, Moore KJ, Fernandez-Hernando C (2010) miR-33 contributes to the Princen HM, Kooistra T (2007) Mouse models for atherosclerosis and regulation of cholesterol homeostasis. Science 238: 1570-1573 pharmaceutical modifiers. Arterioscler Thromb Vasc Biol 27: 1706-1721 Rayner KJ, Esau CC, Hussain FN, McDaniel AL, Marshall SM, van Gils JM, Ray TD, Zhang Y, Edwards PA (2008) FXR signaling in metabolic disease. FEBS Lett 582: Sheedy FJ, Goedeke L, Liu X, et al (2011a) Inhibition of miR-33a/b in non- 10-18 EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 895 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png EMBO Molecular Medicine Pubmed Central

miR-33 controls the expression of biliary transporters, and mediates statin- and diet-induced hepatotoxicity

Download PDF
14 pages

Loading next page...
 
/lp/pubmed-central/mir-33-controls-the-expression-of-biliary-transporters-and-mediates-IEXo7kn4eI

References (80)

Publisher
Pubmed Central
Copyright
Copyright © 2012 EMBO Molecular Medicine
ISSN
1757-4676
eISSN
1757-4684
DOI
10.1002/emmm.201201228
Publisher site
See Article on Publisher Site

Abstract

Research Article Control of hepatic bile secretion by miR-33 miR-33 controls the expression of biliary transporters, and mediates statin- and diet-induced hepatotoxicity 1 1 1 2 3 Ryan M. Allen , Tyler J. Marquart , Carolyn J. Albert , Frederick J. Suchy , David Q.-H. Wang , 4 1,5 1,5 ´ * Meenakshisundaram Ananthanarayanan , David A. Ford ,Angel Baldan Keywords: ABCB11; ATP8B1; Bile secretion is essential for whole body sterol homeostasis. Loss-of-function cholestasis; miR-33; statins mutations in specific canalicular transporters in the hepatocyte disrupt bile flow and result in cholestasis. We show that two of these transporters, ABCB11 and ATP8B1, are functional targets of miR-33, a micro-RNA that is expressed from DOI 10.1002/emmm.201201228 within an intron of SREBP-2. Consequently, manipulation of miR-33 levels in vivo with adenovirus or with antisense oligonucleotides results in changes in bile Received January 23, 2012 Revised May 16, 2012 secretion and bile recovery from the gallbladder. Using radiolabelled cholesterol, Accepted May 23, 2012 we show that systemic silencing of miR-33 leads to increased sterols in bile and enhanced reverse cholesterol transport in vivo. Finally, we report that simvastatin causes, in a dose-dependent manner, profound hepatotoxicity and lethality in mice fed a lithogenic diet. These latter results are reminiscent of the recurrent cholestasis found in some patients prescribed statins. Importantly, pretreatment of mice with anti-miR-33 oligonucleotides rescues the hepatotoxic phenotype. Therefore, we conclude that miR-33 mediates some of the undesired, hepatotoxic GSee accompanying article http://dx.doi.org/10.1002/emmm.201201565 effects of statins. INTRODUCTION apical (canalicular) membrane of hepatocytes by three distinct transmembrane transporters: ABCB11 (ATP-binding cassette, Bile is a complex mixture of bile acids (BAs), cholesterol, sub-family B, member 11; also known as bile salt export pump, phospholipids, proteins and other organic molecules and ions BSEP), ABCG5/ABCG8 (ATP-binding cassette, sub-family G, that serves two main purposes: the solubilization of dietary member 5/8; an obligate heterodimer that facilitates cholesterol lipids in the intestine, and the removal of waste metabolites efflux) and ABCB4 (ATP-binding cassette, sub-family B, through the faeces. Cholestasis refers to a condition with member 4; also known as Multi-drug resistance gene MDR3/ impairment in bile secretion and/or flow, which leads to hepatic MDR2 in humans/mice; which pumps phospholipids, mostly injury and, in severe cases, organ failure that requires liver phosphatidylcholine, PC) (Esteller, 2008). A fourth transporter, transplantation. The major biliary lipids are secreted across the ATP8B1 (ATPase, aminophospholipid transporter, class I, type 8B, member 1), maintains the asymmetry of phospholipid species to promote the required lipid packing of the canalicular (1) Edward A. Doisy Department of Biochemistry and Molecular Biology, membrane for resistance to hydrophobic bile salts and Saint Louis University School of Medicine, Saint Louis, MO, USA canalicular membrane transport (Paulusma et al, 2006, 2008; (2) The Children’s Hospital Research Institute, University of Colorado School Ujhazy et al, 2001). Inactivating mutations in ATP8B1, ABCB11 of Medicine, Aurora, CO, USA (3) Department of Internal Medicine, Saint Louis University, Saint Louis, MO, or ABCB4 result in progressive familial intrahepatic cholestasis USA (PFIC) type 1, 2 or 3, respectively (Hori et al, 2010; Morotti et al, (4) Department of Internal Medicine-Digestive Diseases, Yale University 2011). Accordingly, these three genes are also known as FIC-1, -2 School of Medicine, New Haven, CT, USA and -3, respectively. Patients with benign recurrent intrahepatic (5) Center for Cardiovascular Research, Saint Louis University, Saint Louis, cholestasis (BRIC) also have mutations in any of the latter genes, MO, USA but the residual activity of the mutant transporter prevents *Corresponding author: Tel: þ1 314 9779227; Fax: þ1 314 9779206; the full PFIC phenotype (Hori et al, 2010; Morotti et al, 2011). E-mail: [email protected] 2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License (CC BY-NC 3.0), which permits use, distribution 882 and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Loss-of-function mutations in ABCG5 or ABCG8 result in Here we test the proposal that miR-33 also modulates hepatic sitosterolemia (Hubacek et al, 2001; Yu et al, 2002), but not bile metabolism by decreasing the expression of specific sterol in cholestasis. Mice deficient in any of these transporters transporters in the canalicular membrane of hepatocytes. Our phenocopy the human syndromes (Lammert et al, 2004; data show that conserved sequences in the 3 UTR of ABCB11 and Pawlikowska et al, 2004; Shah et al, 2010; Wang et al, 2003; ATP8B1 are functional miR-33-responsive elements (RE), and Yu et al, 2002). Nevertheless, PFIC/BRIC are thought to develop that manipulation of miR-33 levels by adenoviral over-expression from the inability to secrete BAs (ABCB11 defect) or or with antisense oligonucleotides results in altered biliary output phospholipids (ABCB4 defect), or from the inability to maintain in vivo. Hence, miR-33 limits sterol efflux in the hepatocyte canalicular membrane lipid structure (ATP8B1 defect). through both the basolateral membrane (via ABCA1) and the We (Marquart et al, 2010) and others (Gerin et al, 2010; Horie apical membrane (via ABCB11 and ATP8B1). The physiological et al, 2010; Najafi-Shoushtari et al, 2010; Rayner et al, 2010) importance of the murine miR-33 pathway is also supported in recently showed that the evolutionarily conserved miR-33 is experiments using radiolabelled cholesterol, showing that miR-33 expressed from within intron 16 of SREBP-2, and that this silencing increases the amount of labelled sterols recovered from miRNA downregulates the expression of the sterol transporter the gallbladder as well as the overall reverse cholesterol ABCA1. Manipulation of miR-33 levels in vivo results in changes transport. We also report that administration of statins, which in circulating levels of high-density lipoproteins (HDL) (Horie induce the expression of miR-33, results in decreased hepatic et al, 2010; Marquart et al, 2010; Najafi-Shoushtari et al, 2010; expression of Abcb11 and Atp8b1, but not other canalicular Rayner et al, 2010, 2011a,b). Consequently, silencing miR-33 transporters. Finally, we show that silencing miR-33 rescues the could be useful as a therapeutic intervention to increase plasma hepatotoxicity and lethality caused by co-administration of HDL in patients with hypercholesterolemia. simvastatin and a cholate-rich diet. We conclude that pharma- cologic manipulation of hepatic miR-33 levels might represent a new approach to manage certain cholestatic syndromes. RESULTS Systemic silencing of miR-33 in mice increases both bile secretion and the expression of hepatic Abcb11 and Atp8b1 In an effort to understand the in vivo physiological importance of miR-33, we injected chow-fed mice with saline, and scrambled or anti-miR-33 oligonucleotides. A week later, we collected different tissue samples. Data show a twofold increase in the volume of bile recovered from the gallbladders of anti-miR-33 animals, compared to controls (Fig 1A). However, the overall concentration of PC, cholesterol and BAs did not change in bile between the three experimental groups (Fig 1B). Next, we tested the expression of several bile-related hepatic canalicular transporters. We found that the messenger RNA (mRNA) levels of both Abcb11 and Atp8b1 were significantly increased in the livers of mice injected with anti-miR-33 oligonucleotides, compared to controls (Fig 1C). The mRNA levels of other transporters (Abcg5, Abcg8, Abcb4) and other genes involved in Figure 1. Increased bile secretion following silencing of miR-33. A. Bile recovered from the gallbladder of mice (n ¼ 5) on chow diet, injected with saline, and scrambled or anti-miR-33 oligonucleotides (5 mpk i.v., for 2 consecutive days). Animals were then kept for a week on chow and fasted overnight before sample collection. B. Levels of phosphatidylcholine (PC), cholesterol (chol) and bile acids present in gallbladder bile in the same mice. C. Relative expression of hepatic canalicular transporters in the same mice. D. A different group of mice (n ¼ 6–8) was injected as described above. A week later, mice were anesthetized, the bile duct cannulated, and hepatic bile collected for 1 h. E. Levels of bile acids, phosphatidylcholine (PC), cholesterol (chol) present in bile from the last group of mice. Data shown as mean  SEM. p< 0.05; p< 0.01. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 883 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 bile homeostasis (i.e. Shp, Cyp7a) remained unchanged in the experiment in HuH7 cells with [ C]-cholesterol, to monitor same livers (Fig 1C and Supporting Information Fig S1). As both conversion to [ C]-BAs and secretion of labelled sterols. expected, previously described miR-33 targets (Abca1 and The cells were incubated in media supplemented with 0.2% Cpt1a) were induced by the antisense treatment (Supporting albumin, but not serum, to minimize loss of the labelled Information Fig S1). To test whether the increase in bile cholesterol via ABCA1 and ABCG1. Data in Supporting recovery from the gallbladder of mice receiving anti-miR-33 Information Fig S2 show that cells overexpressing miR-33 oligonucleotides shown in Fig 1A was due to accelerated bile had a significant decrease in their ability to efflux labelled BAs, secretion, an independent group of animals were treated as but not cholesterol, compared to controls. These results suggest described above, and after an overnight fasting they were that BA secretion is impaired following miR-33 overexpression anesthetized and the common bile duct was cannulated to allow (consistent with decreased ABCB11 levels), likely resulting in the collection of hepatic bile during 60 min. Data in Fig 1D and E intracellular accumulation of BAs, which suppress further show that the secretion rates of total bile, BAs and PC were conversion of cholesterol into BAs. significantly increased in mice injected with anti-miR-33 oligonucleotides, compared to control animals. The secretion Hepatic overexpression of miR-33 decreases biliary output in rate of cholesterol also trended up in these same mice, but did cholate-fed mice not reach statistical significance. The apparently contradictory Mice fed a lithogenic diet (21% fat, 1.25% cholesterol and 0.5% results shown in Fig 1B and E might be explained by changes in cholate) exhibit disrupted bile homeostasis and, after a few water secretion/reabsorption across the canalicular, ductal and weeks, develop cholestasis (Khanuja et al, 1995; Moschetta et al, gallbladder epithelium, which are known to modulate bile 2004; Shah et al, 2010; Yu et al, 2005). Our data show that this composition (Portincasa et al, 2008). Nevertheless, the data in diet significantly decreases hepatic miR-33 levels after a week, Fig 1 demonstrate a functional role for miR-33 on bile secretion. compared to mice fed chow (Fig 3A, lanes 1–2). To test the relevance of hepatic miR-33 under pathological conditions, we ABCB11 and ATP8B1 have functional miR-33 responsive performed a diet-induced cholestasis experiment: we injected sequences in the 3(UTR mice with saline, and empty or miR-33-encoding adenovirus, Based on the previous results, we hypothesized that both switched them to the lithogenic diet the following day, and kept ABCB11 and ATP8B1 are direct targets of miR-33. Analysis of the them on this diet for 7 additional days. Predictably, the 3 UTR of these genes revealed evolutionarily conserved expression of miR-33 was increased in the livers of mice sequences that are partially complementary to miR-33 (Fig 2A receiving the miR-33 vector (Fig 3A, lanes 3–4). As expected, all and B). To test whether these sequences are functional, we the mice fed the lithogenic diet showed an enlarged gallbladder. cloned the 3 UTR of both human and mouse genes, or the However, the volume of bile recovered from mice transduced putative miR-33 responsive sequences, immediately down- with adeno-miR-33 was 45% smaller than that from mice stream of a luciferase reporter. Co-transfection with a miR-33- receiving saline or adeno-empty (Fig 3B), suggesting hepatic bile encoding plasmid confirmed that these genes indeed respond to retention. Additionally, the livers from mice receiving miR-33 miR-33 (Fig 2C and D). Hence, miR-33 expression resulted in vectors appeared slightly enlarged, resulting in a significant 40% decrease in luciferase activity when the reporter is fused increase in the liver to body mass ratio (Fig 3C). Analysis of liver to the 3 UTR or the response elements of human/mouse ATP8B1 contents revealed that overexpression of miR-33 resulted in (Fig 2C; lanes 5–8 and 11–14) or ABCB11 (Fig 2D; lanes 1–4 and significant increases in hepatic BAs, total cholesterol and 7–10). As expected, mutations that prevent the binding of the esterified cholesterol (EC), but no changes in unesterified seed sequence of the miRNA abolished the response to miR-33 cholesterol (UC) or PC (Fig 3D). For reasons that remain (Fig 2C and D; lanes 9–10 and 15–16). obscure, the amount of triglycerides (TGs) was significantly reduced in the livers of mice transduced with miR-33, compared miR-33 downregulates ABCB11 and ATP8B1 in both human to control animals (Fig 3D). The molecular basis of this latter and mouse hepatocytes effect will require further investigation. Previous reports We next determined whether the mouse and human ABCB11 showed that genes involved in fatty acid b-oxidation and TG and ATP8B1 genes are regulated following miR-33 overexpres- metabolism such as CPT1a (carnitine palmitoyltransferase 1a), sion. Hence, mouse primary hepatocytes or human HuH-7 CROT (carnitine O-octanoyltransferase), HADHB (hydroxyacyl- hepatoma cells were transduced with empty or miR-33-encoding CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase, adenovirus. Data show that the expression of human/murine beta subunit), SIRT6 (sirtuin 6) and AMPK1a (AMP-activated ABCB11 and ATP8B1 was significantly reduced in cells protein kinase 1a) are direct targets of miR-33 (Davalos et al, following overexpression of miR-33 (Fig 2E and F), while 2011; Gerin et al, 2010). From these studies it was inferred that mRNA levels of other hepatic transporters (Abcg5, Abcg8, miR-33 might function to limit fatty acid utilization in Abcb4, Abcc2, Slc10A1) remained unchanged (Fig 2E and F). In hepatocytes and other cell types, and that sustained elevated agreement with the mRNA data, protein levels for both ATP8B1 levels of miR-33 could lead to fatty liver. Based on this literature, and FXR-induced ABCB11 were downregulated following we expected to see an increase in hepatic TG contents in mice miR-33 overexpression (Fig 2G). Collectively, data from that overexpressed miR-33, compared to control animals. Figs 1 and 2 identify ATP8B1 and ABCB11 as functional direct However, data showing a significant decrease in liver TGs targets of miR-33. Additionally, we performed a pulse-chase was reproducible in 2 independent experiments (n ¼ 5/group). 884  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Figure 2. Functional miR-33 responsive elements in the 3(UTR of ATP8B1 and ABCB11. A,B. Conserved sequences in the 3 UTR of ATP8B1 and ABCB11 are partially complementary to miR-33. The element in human ATP8B1 is located 1877–1897 nt after the stop codon. The element in ABCB11 overlaps the stop codon in primates, while rodents show a conserved sequence 732–751 nt after the stop codon. C,D. Luciferase assays in HEK293 cells using the whole 3 UTR of human or murine ATP8B1 and ABCB11, or a single copy of the responsive elements (RE) identified above, or mutant responsive elements (RE ; where AATGCA was mutated to GGGTTG to prevent complementarity to the seed sequence of the miRNA), co-transfected with (closed bars) or without (open bars) a vector to overexpress miR-33. In grey, data from empty (negative control) and R33 (positive control containing a 100% match to miR-33) reporter vectors. E,F. Relative mRNA expression of canalicular transporters in primary murine hepatocytes (n ¼ 4 dishes/condition) and human HuH-7 hepatoma cells (n ¼ 3 dishes/condition) transduced 48 h with empty or miR-33 adenovirus. G. Relative protein levels in HuH7 cells transduced with empty or miR-33 adenovirus. Some cells were incubated for 16 h in the presence of FXR:RXR agonists (2 mmol/L GW4064 : 1 mmol/L 9-cis-retinoic acid) to induce ABCB11. Asterisk indicates a non-specific band. Data shown as mean  SD; p< 0.01. We can only speculate that the impact of miR-33 on lipid regulatory element binding protein 1) did not confer response to metabolism in vivo is yet to be fully elucidated, both under the miRNA (Supporting Information Fig S3B and C), and the normal diet conditions and under dietary challenge (as is the levels of nSREBP-1 were not reduced in the livers of mice case of data presented in Fig 3). Interestingly, the reduced overexpressing miR-33 (Supporting Information Fig S3D). hepatic expression of Fasn in miR-33 overexpressing mice Analysis of the bile revealed a significant decrease in BAs and (Supporting Information Fig S3A) could explain the decline in a modest increase in cholesterol in samples from adeno-miR-33 TGs in these mice. However, the reasons for such reduced Fasn mice, but no change in PC (Fig 3E). These results are consistent expression are not clear, since no candidate miR-33 elements with miR-33 controlling the expression of hepatic transporters are found in human/mouse FASN. Additionally, the potential involved in sterol mobilization. Indeed, hepatic expression of miR-33 RE found in the 3 UTR of human/mouse SREBP-1 (sterol both Abcb11 and Atp8b1 was markedly reduced in mice EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 885 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 Figure 3. Effect of miR-33 overexpression on diet-induced cholestasis. A. C57BL/6 mice (n ¼ 5–7) were kept on chow diet or lithogenic diet for 7 days (lanes 1–2). A different group of animals were transduced i.v. with empty or miR-33 adenovirus (2  10 pfu/mouse), and then switched to the lithogenic diet (lanes 3–4). After 7 days, mice were fasted overnight and killed the following morning. Data show relative levels of hepatic miR-33. B. The volume of bile recovered from the gallbladder is significantly reduced in mice transduced with miR-33. Picture shows pooled bile collected from five mice each in an independent experiment. C. Liver to body mass ratios. D. Hepatic levels of bile acids; total, unesterified (UC) and esterified (EC) cholesterol; phosphatidylcholine (PC) and triglycerides (TG). E. The amounts of bile acids (BA), cholesterol (chol) and phosphatidylcholine (PC) were determined in bile, and expressed as mol% (mol per 100 mol). Compared to mice infused with saline or Adeno-empty, the bile from animals transduced with Adeno-miR-33 showed increased amounts of cholesterol (11.6  1.1 vs.6.6  0.5 vs.5.9  0.6 mol%; miR vs. scrambled vs. saline; p ¼ 0.009), and decreased amounts of bile acids (78.6  1.7 vs. 83.7  1.0 vs. 82.1  1.4 mol%; miR vs. scrambled vs. saline; p ¼ 0.03), but no change in PC contents (9.8  1.3 vs.9.7  0.6 vs. 12.0  0.9 mol%; miR vs. scrambled vs. saline). F. Relative expression of hepatic canalicular transporters in the same mice. Data shown as mean  SD; p< 0.05; p< 0.01. either in mouse primary hepatocytes (Fig 2E), nor in human HuH7 or HepG2 cells (Supporting Information Fig S4), nor in human Hep3B cells (where exogenous miR-33 also failed to block the induction of ABCG5 by an LXR agonist (Marquart et al, 2010)). Analysis of the 3 UTR regions of the human and murine ABCG5/8 genes did not reveal any sequences with perfect complementarity to the seed sequence of miR-33 (Supporting Information Fig S4). However, when cloned downstream of a luciferase reporter, the 3 UTRs of human (but not mouse) ABCG5, and human/mouse ABCG8 conferred a very modest response to miR-33 overexpression (Supporting Information Fig S4). We speculate that these latter results are due to non- conserved, imperfect 6-mer sequences (Supporting Information Fig S4) that bind to miR-33 with low affinity. Hence, the decreased expression of Abcg5 and Abcg8 noted in Fig 3F could be the result of off-target effects due to the supra-physiological levels of miR-33 achieved using adenovirus, thus potentially limiting the interpretation of the results. Intriguingly, mice deficient in ABCB11 show decreased hepatic levels of Abcg5 and transduced with adeno-miR-33, compared to adeno-empty Abcg8 when fed a lithogenic diet (Wang et al, 2003); the (Fig 3F), while the expression of other transcripts did not molecular mechanism of this cross-talk remains unknown, but change between the three groups of mice (Fig 3F and Supporting authors speculate that it is independent on the levels/activity of Information Fig S3A). Importantly, adenoviral-mediated over- LXR (Wang et al, 2003). Whether the drop in Abcg5/8 levels in expression of miR-33 was also able to reduce the expression of the livers in Fig 3F are due to extreme levels of miR-33, or the both Abcb11 and Atp8b1 in mice fed chow (Supporting result of a yet-unknown signalling pathway that links the Information Fig S3E), suggesting that miR-33 can also modulate expression of different bile transporters will require further basal levels of these transporters in vivo under normal diet investigation. Nevertheless, results from Fig 3 and Supporting conditions. Finally, mRNA levels of the sterol transporters Information Fig S3 confirm a role for miR-33 on bile homeostasis Abcg5 and Abcg8 were also significantly reduced in the adeno- both under normal chow conditions and following dietary miR-33 group fed the lithogenic diet (Fig 3F). These latter results challenge. were unexpected, since the human and murine ABCG5 and ABCG8 genes are not direct targets of miR-33 (Supporting Effect of miR-33 on reverse cholesterol transport (RCT) Information Fig S4). Hence, we could not see repression of these RCT mobilizes extrahepatic cholesterol back to the liver for latter genes by adenoviral-mediated miR-33 overexpression subsequent secretion into bile, and final excretion in the faeces 886  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. (Khera & Rader, 2010; Rader et al, 2009; Wang et al, 2007; Wang & Rader, 2007). Some authors showed that hepatobiliary secretion is an essential component of the RCT pathway (Nijstad et al, 2011). Others have disagreed (Temel et al, 2010). Remarkably, some patients with cholestasis develop xanthomas that usually dissolve after normal bile homeostasis is restored (Emerick & Whitington, 2002; Englert et al, 2006), suggesting that impairment in bile flow ultimately leads to accumulation of sterols in patients. The previously described target of miR-33, ABCA1, plays an essential role for RCT (Wang et al, 2007). While this manuscript was in preparation, Rayner et al (2011b) showed that systemic silencing of miR-33 promotes RCT, and speculated this effect was the result of increased ABCA1- dependent cholesterol efflux. While it is imaginable that constitutive secretion of bile could funnel an increase of intrahepatic cholesterol derived from ABCA1-induced choles- terol efflux in peripheral tissues, authors have shown that increased hepatic ABCA1 activity promotes the re-secretion of cholesterol into new nascent HDL, thus preventing the mobilization of intrahepatic cholesterol towards the biliary secretory pathway (Annema et al, 2012; Nijstad et al, 2009; Tietge et al, 2008; Wiersma et al, 2009). Consequently, we hypothesized that miR-33-dependent changes in biliary trans- porters and overall bile secretion (Figs 1–3) contribute, together with ABCA1, to the effect of miR-33 on RCT. To test whether miR-33 expression alters the mobilization of extrahepatic cholesterol into the bile, we injected macrophage foam cells that were radiolabelled with [ H]-cholesterol into the peritoneal cavity of animals treated with scrambled or anti-miR-33 oligonucleotides, and followed the fate of labelled sterols for 48 h. Hepatic miR-33 expression was induced 2.1  0.3-fold in mice injected with anti-miR-33 oligos. We did not observe changes in body, liver or faecal mass (Fig 4A). Data show that the amount of labelled cholesterol in plasma, but not in liver, Figure 4. Reverse cholesterol transport is enhanced after systemic miR- increased in mice receiving anti-miR-33 oligonucleotides, 33 silencing. C57BL/6 mice (n ¼ 7) were infused i.v. with 5 mpk scrambled or compared to control mice (Fig 4B and C). A large increase in anti-miR-33 oligonucleotides for 2 consecutive days, and 5 days later received bile volume was noted, again, in the gallbladder of anti-miR-33 6 1  10 radiolabelled macrophages by i.p. injection. Mice were kept on chow mice (Fig 4A). Importantly, and validating our hypothesis, the for 48 h until sacrifice. amount of labelled sterols in the bile was also increased in these A. Different parameters in mice receiving scrambled (open bars) and anti- miR-33 (closed bars) treatment at the time of sacrifice. Notice the latter mice (Fig 4D). Finally, labelled sterols recovered from increase in bile recovery from the gallbladder. faeces were increased approximately twofold, compared to B. Percentage of total injected dpm recovered in the plasma at different control animals (Fig 4D). Subsequent lipid extraction from time points after injection of radiolabelled cells. faeces revealed that the amount of labelled neutral sterols C-E. Percentage of total injected dpm recovered at the time of sacrifice in was significantly increased in samples from mice receiving liver, bile from the gallbladder and faeces of the same mice. Data shown anti-miR-33 treatment, compared to controls (2.31  0.48 vs. as mean  SD; p< 0.05; p< 0.01. 0.94  0.18% dpm injected; p ¼ 0.025), while the recovery of faecal BAs trended upwards in the same animals, but did not reach statistical significance (3.25  0.47 vs. 2.64  0.39% dpm Statins repress the expression of miR-33 targets injected; p ¼ 0.37). Together, these data are consistent with our From a clinical perspective, it may be important that statins not hypothesis that miR-33 modulates RCT, likely through the only increase SREBP-2 expression but also they increase miR-33. combined regulation of HDL metabolism (via ABCA1) and bile Here we show that both simvastatin and atorvastatin induce metabolism (via ABCB11 and ATP8B1). Collectively, data hepatic miR-33 expression while at the same time decrease the presented in Figs 1–4 strongly suggest a concerted action of mRNA levels of miR-33 targets Abca1, Cpt1a, Abcb11 and these miR-33 targets. Still, further experiments using mice Atp8b1 (Fig 5A and B). Similar results were obtained in HuH7 deficient for each of these transporters will provide definitive cells (Fig 5C). These data strongly suggest that patients taking answers as to which specific transporter(s) are mediating the statins might have sustained, decreased levels of hepatic miR-33 effect of anti-miR-33 on RCT. targets, including transporters linked to PFIC/BRIC. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 887 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 (Supporting Information Fig S5B). Likely both the statin and the diet contributed to changes in the expression of hepatic miR-33. Importantly, we noted a dose-dependent lethality effect of the statin after mice were switched to the lithogenic diet. Thus, mice on 300 mpk simvastatin had to be euthanized on/before day 5; mice on 150 and 50 mpk simvastatin showed 50 and 100% survival rates, respectively (Fig 6A). This was paralleled by a dramatic dose-dependent increase in liver weight (Fig 6B). The livers of mice receiving 150 mpk simvastatin appeared not only enlarged, but also extremely steatotic (i.e. very pale and soft consistency) (Fig 6C); in contrast, livers from mice receiving 50 mpk simvastatin were visually indistinguishable from the livers of control mice (Fig 6C). Electrospray ionization – mass spectrometry (ESI-MS) analysis confirmed the accumulation of lipids, especially TGs, in livers from mice dosed with 150 mpk simvastatin (Fig 6D). Consistent with an intrahepatic cholestatic phenotype, liver BAs levels were also markedly elevated in these latter mice (Fig 6D). Intriguingly, ESI-MS data show a profound remodelling of ceramides in these livers: ceramides containing long-chain fatty acids (16:0, 18:0 and 20:0) were increased, while those containing very long-chain fatty acids (23:0, 24:1 and 24:0) were significantly reduced in the livers of the 150 mpk group (Supporting Information Fig S5D). The (patho)physiolo- gical significance of these changes remain obscure. Statin- induced hepatotoxicity was mirrored by increased levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), bilirubin and BAs in the blood, resulting in bright yellow plasmas (Fig 6E). Although the overall amount of bile recovered from the gallbladders did not differ between groups (Fig 6F), we noted a dose-dependent decrease in PC, small changes in bile salts concentration, and a dramatic decrease in cholesterol in samples from the 150 mpk group (Fig 6F). These changes in bile composition are consistent with abnormal expression/function of canalicular transporters. Finally, we compared the expression Figure 5. Statins induce miR-33 and reduce the mRNA expression of of selected transcripts in the livers of mice gavaged saline or specific hepatic canalicular transporters. 50 mpk simvastatin (mildly cholestatic, based on BA accumula- A, B. C57BL/6 mice (n ¼ 5) were gavaged daily with statins, and kept on chow tion) (Fig 6G). Data show that both Abcb11 and Atp8b1 mRNA diet. Samples were collected after 7 days, following an overnight were significantly reduced (40%) in the simvastatin group; fasting. C. HuH7 cells were cultured in quadruplicate for 48 h in DMEM supple- these changes were specific since the expression of other bile- mented with 2% lipoprotein-deficient serum in the presence or absence related transporters (Abcb4, Abcg5, Abcg8) did not change of simvastatin. Relative expression of specific genes shown as (Fig 6G). mean  SD; p< 0.05; p< 0.01. The results from Fig 6 support a mechanism in which statins induce miR-33, which in turn reduces the levels of both Abcb11 and Atp8b1, resulting in altered bile secretion from hepatocytes, Silencing miR-33 rescues statin- and diet-induced liver damage which ultimately leads to liver malfunction under conditions of We next examined the combined effect of statins and lithogenic lithogenic dietary challenge. To validate this model, we next diet on mice. In a preliminary experiment, we noted that tested whether silencing miR-33 could rescue the phenotype. simvastatin increased miR-33 levels in a dose-dependent Hence, mice (n ¼ 12) were injected with scrambled or anti-miR- manner in chow-fed mice (Supporting Information Fig S5A). 33 oligonucleotides, and then dosed daily with 150 mpk Then, we gavaged chow-fed mice the statin (0, 50, 150 or simvastatin and fed the lithogenic diet (Fig 7). Notably, hepatic 300 mpk) for 2 days prior to switching them to the lithogenic diet levels of miR-33 at the end of the experiment were reduced for an additional 7 days. Simvastatin was administered daily 40% in mice receiving the antisense oligonucleotides, during these latter 7 days (Fig 6A), while we monitored body compared to control animals (Supporting Information weight and food consumption (Supporting Information Fig S6A). Data show that, with one exception, all anti-miR-33 Fig S5C). At the end of the experiment, hepatic miR-33 mice survived for at least a week, losing 5% body weight; in expression was still elevated in the 50 mpk mice, but had contrast, mice injected with scrambled oligonucleotides showed decreased in the 150 mpk mice, compared to control animals 50% survival rate and significant body weight loss (Fig 7A and 888  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Figure 6. Simvastatin and lithogenic diet induce liver toxicity. C57BL/6 mice (n ¼ 6) were gavaged daily with simvastatin and fed a lithogenic diet. Samples were collected after 7 days, or when mice appeared moribund. A. Survival is hampered by simvastatin in a dose-dependent manner. B. Severe hepatomegaly in mice dosed with 150 mpk simvastatin. C. Representative macroscopic appearance (upper and middle panels), and haematoxylin and eosin staining of paraffin-embedded sections (lower panels) from the same livers. Note the abnormally swollen cells and pale (i.e. steatotic) appearance of the livers in the 150 mpk group. D. Specific hepatic lipids as determined by ESI-MS, and normalized to tissue weight. FFA, free fatty acids; DAG, diacylglycerides; TAG, triacylglycerides; PC, phospha- tidylcholine; UC, unesterified cholesterol; CE, cholesterol esters. p< 0.01 versus saline. E. Representative samples of plasma, and levels of circulating alanine aminotransferase (ALT), aspartate aminotransferase (AST), bile acids and bilirubin. F. Bile was recovered from the gallbladder, pooled and the contents of phosphatidylcholine (PC), cholesterol (c) and bile acids (BA) determined with colorimetric kits. G. Relative expression of hepatic canalicular transporters (upper panel) and other genes involved in bile acid and sterol homeostasis (bottom panel) in samples from mice treated with 0 or 50 mpk simvastatin. Data shown as mean  SD. p< 0.01. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 889 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 Figure 7. Silencing miR-33 rescues the liver damage induced by simvastatin and lithogenic diet. C57BL/6 mice (n ¼ 12) were injected i.v. with 5 mpk scrambled or anti-miR-33 oligonucleotides for two consecutive days, and then treated as in Fig 5. A. Survival curves. B. Body weight, expressed as % compared to mass on the day of first injection. C. Representative macroscopic appearance of the livers, and haematoxylin and eosin staining of paraffin-embedded sections. D. Liver to total body mass ratios. p< 0.01 versus moribund mice. E. Representative samples of plasma. Note the normalization in colour in the anti-miR-33 group. F. Amounts of specific hepatic lipids as determined by ESI-MS, normalized to tissue weight (n ¼ 4). Acronyms for lipid classes as defined in Fig 5. p< 0.01 versus scrambled. G. Relative expression of hepatic canalicular transporters (upper panel) and other genes involved in bile acid and sterol homeostasis (bottom panel) (n ¼5for scrambled treatment; n ¼ 10 for anti-miR-33 treatment). Data shown as mean  SEM. p< 0.01 versus moribund scrambled mice; p< 0.05 versus surviving scrambled mice. 890  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. B). Remarkably, the livers from the anti-miR-33 group appeared non-steatotic (Fig 7C), and the liver/body mass ratio was significantly lower than in the scrambled group that succumbed to the treatment (Fig 7D). Additionally, the plasmas from mice receiving anti-miR-33 appeared normal as compared to the yellow-coloured plasma from mice treated with scrambled oligonucleotides (Fig 7E). Rescue of the statin- and diet-induced phenotype was also evident when we analyzed hepatic lipids: anti-miR-33-treated animals showed a significant normalization in lipid contents, compared to control mice (Fig 7F and Supporting Information Fig S6B and C). Finally, we studied the hepatic mRNA levels of selected genes (Fig 7G). Interestingly, the expression of all canalicular transporters, with the exception of Atp8b1, was severely decreased in mice that succumbed to the treatment. Additionally, the levels of Abcg5 and Abcg8 were, for reasons that are not clear, highly variable within each group of mice (Fig 7G). When only considering those mice that survived 7 days on the diet, treatment with anti-miR-33 oligonucleotides resulted in significant increased expression of Atp8b1, but not of Abcb11 or any other canalicular transporter (Fig 7G). Perhaps the expression of Abcb11 is already maximal in these livers, due to the diet-induced activation of FXR. In general, the mRNA levels of bile-related genes in surviving mice in the scrambled group were similar to those in the antisense group (Fig 7G), suggesting that the expression of these genes is critical for survival. We also analyzed the mRNA expression of selected hepatic Phase I and II detoxifying genes (Supporting Information Fig S6D). Again, we found large differences in the expression of most of these genes Figure 8. miR-33 limits the mobilization of sterols in hepatocytes through within each experimental group, making the interpretation of both the sinusoidal and the canalicular membrane. the data difficult. We speculate that the increased survival of A. miR-33 mediates the cross-talk between the SREBP-2, LXR and FXR mice receiving the anti-miR-33 treatment is likely due to pathways by directly modulating the expression of ATP8B1, ABCB11, ABCA1 complex, coordinated changes in the expression of several and ABCG1 ( in mice, but not in humans). Data suggest that miR-33 might genes, which ultimately results in the accelerated clearance of also indirectly affect the expression of ABCG5/8 (see main text for details). Decreased ATP8B1 activity in human, but not mouse, hepatocytes results bile and drug and/or diet-derived toxic metabolites. Never- in PKC-dependent inactivating phosphorylation of FXR. theless, data in Fig 7 show conclusively that miR-33 mediates B. During episodes of low intracellular cholesterol, or following treatment statin- and diet-induced hepatotoxicity. Additional experiments with statin drugs, miR-33 is transcriptionally induced and reduces the using mice deficient in ABCB11 and/or ATP8B1 will provide expression of sterol transporters. clues about the relative contribution of each specific transporter to statin-induced, miR-33-dependent hepatotoxicity. The exact hepatoprotective mechanism of anti-miR-33 on statin- and diet- induced toxicity is yet to be determined. We speculate that data (Marquart et al, 2010 and this report) and that from other increased bile flow due to de-repression of miR-33 targets such laboratories (Gerin et al, 2010; Horie et al, 2010; Najafi- as Abcb11 and/or Atp8b1 contribute to the protective effect, but Shoushtari et al, 2010; Rayner et al, 2010) suggest that miR-33 whether other bile-independent pathways are also necessary for plays a central role in coordinating those three pathways hepatoprotection remains to be established. (Fig 8A). We hypothesize that miR-33 evolved to limit the efflux of hepatocyte sterols, both through the basolateral and apical membranes, in situations of low intracellular cholesterol that DISCUSSION induce SREBP-2 (Fig 8B). Bile acids bind and activate FXR, which in turn promotes Hepatic sterol homeostasis is a complex process that encom- increased bile secretion (by inducing ABCB11 and ABCB4) and passes de novo synthesis of cholesterol, secretion and uptake of reduced BAs synthesis (by indirectly repressing CYP7a) lipoproteins, conversion of cholesterol to BAs, and bile (Gadaleta et al, 2010; Zhang & Edwards, 2008). Interestingly, secretion. Extensive literature shows that three transcription FXR targets were found to be reduced in patients with PFIC-1 or factors (sterol regulatory element binding protein 2, SREBP-2; following siRNA against ATP8B1 in humans hepatocytes liver-X-receptor, LXR; and farnesoid X-activated receptor, FXR) (Alvarez et al, 2004; Chen et al, 2004; Koh et al, 2009; orchestrate complementary aspects of sterol metabolism. Our Martinez-Fernandez et al, 2009), leading to the proposal that EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 891 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 FXR activity can be modulated by ATP8B1, perhaps via protein both transcripts are decreased by fasting and induced by kinase C (Chen et al, 2010; Frankenberg et al, 2008). Other refeeding (but Srebp-1 is more potently induced than -2) (Horton investigators, however, failed to reproduce these observations et al, 1998; Bennett et al, 2008), and only Srebp-2 is induced by (Cai et al, 2009). Nevertheless, multiple studies in patients with statins (Bennett et al, 2008). Additional studies using primates PFIC/BRIC-1 and -2 support a critical role for both ATP8B1 and or humanized mice (in which transgenic miR-33b is expressed ABCB11 on bile secretion (Davit-Spraul et al, 2010; Harris et al, from an intron of Srebp-1) will be necessary to study the impact 2005; Kubitz et al, 2005; Morotti et al, 2011; van der Woerd et al, of SREBP-1-derived miR-33 on bile metabolism. Nevertheless, 2010). we speculate that anti-miR-33 oligonucleotides might be useful Since SREBP-2/miR-33 are transcriptionally induced follow- to manage patients who develop BRIC as a consequence of ing treatment with statins (Marquart et al, 2010; Najafi- partial loss of expression and/or function of ABCB11 or ATP8B1. Shoushtari et al, 2010; Rayner et al, 2010), we hypothesized that miR-33 might account for some of the (side) effects of these drugs. Several reports show that certain patients develop MATERIALS AND METHODS cholestasis following prescription of statins (Batey & Harvey, 2002; de Castro et al, 2006; Merli et al, 2010; Rahier et al, 2008; Materials Ridruejo & Mando, 2002; Torres et al, 2002). While the Vectors encoding miR-33 were previously described (Marquart et al, molecular events in the livers of these patients are unknown, 2010). HEK293 and HuH-7 cells (ATCC) were maintained in DMEM this clinical setting is consistent with our model (Fig 8) in which supplemented with 10% fetal bovine serum (FBS). Luciferase reporter statin-induced miR-33 represses both ABCB11 and ATP8B1, constructs containing miR-33 response elements or the 3 UTR of thus decreasing bile secretion and eventually leading to human/mouse ATP8B1, ABCB11, ABCG5 and ABCG8 were generated intrahepatic cholestasis. We recognize, however, that our by cloning each sequence into an XbaI site in pGL3-Promoter experimental model is extreme due to the relatively high dose (Promega). 3 UTR fragments were amplified from human or mouse of the statin used and the fact that patients do not normally genomic DNA using Platinum Pfx (Invitrogen). Control (5 -TCCTAGAAA- consume lithogenic diets (i.e. a high fat, high cholesterol and GAGTAGA) and anti-miR-33 (5 -TGCAACTACAATGCA), locked nucleic high BA diet). It is intriguing to speculate, however, that these acid (LNA) oligonucleotides were kindly provided by Miragen patients might carry specific polymorphisms in ABCB11 or Therapeutics Inc. (Boulder, CO). Control oligonucleotides are designed ATP8B1 that would make them more susceptible to statin- to not target any mouse RNA annotated in NCBI databases. induced miR-33 effects. In general, most physicians are cautious Mice received 5 mg/kg (in 100 mL saline) via tail vein injection for 3 to prescribe statins to patients with underlying liver disease, and consecutive days, unless otherwise stated. the effect of these drugs on patients with primary biliary cirrhosis is controversial (Abu Rajab & Kaplan, 2010; Stanca Murine studies et al, 2008; Stojakovic et al, 2007, 2010). Nevertheless, our data C57BL/6 mice were obtained from NCI/Charles River, and maintained clearly demonstrate a dose-dependent effect of simvastatin on in a 12 h/12 h light/dark cycle with unlimited access to food and diet-induced hepatotoxicity and cholestasis in mice. It is water. Where indicated, animals (8–10 week old, n ¼ 6 per group, important to stress the fact that mice tolerate 300 mpk statins unless noted otherwise) were infused via tail vein with adenoviral with a concomitant high fat, high cholesterol diet for months vectors (2  10 pfu) or antisense oligonucleotides (5 mpk, 3 con- (reviewed in Zadelaar et al, 2007), but to our knowledge no secutive days). Where indicated, mice were fed a lithogenic diet previous studies addressed the interaction between statins and (Purina 5A8E). Tissues and bile were collected following overnight cholate feeding. Our data offer conclusive evidence that, at least fasting. RCT experiments were performed as described (Wang et al, in mice, statins potentiate diet-induced cholestasis. From a 2007). Briefly, mice were injected with oligonucleotides (5 mpk) on 6 3 mechanistic perspective, our data show that the phenotype is days 1, 2, 8 and 9, and on day 15 injected i.p. with 1–1.5  10 [ H]- largely rescued by anti-miR-33 oligonucleotides. cholesterol-loaded bone marrow-derived macrophages. Blood samples Authors suggested that a therapy that combines statins and were collected 6, 24 and 48 h thereafter. Liver and bile, and the faeces anti-miR-33 oligonucleotides may be useful in hyperlipidemic produced during the last 48 h were collected and flash-frozen until patients by reversing statin-induced, miR-33-mediated repres- used. The amount of radiolabelled sterols was determined by sion of ABCA1, which ultimately would increase plasma HDL scintillation. Where indicated, mice were dosed daily for 7 days with (Horie et al, 2010; Marquart et al, 2010; Najafi-Shoushtari et al, saline or statin by oral gavage, and killed following an overnight 2010; Rayner et al, 2010). Our new data show that miR-33 plays fasting. Bile secretion was measured in anesthetized mice following a key role in the hepatic response to statins by coordinating the common bile duct cannulation, which allowed the continued expression of several sterol transporters, and that disruption of collection of hepatic bile by gravity. Body temperature was maintained the miR-33 pathway prevents statin-induced hepatotoxicity. during surgery and bile collection at 37  0.58C. All studies were Data support the hypothesis that miR-33 controls whole-body approved by the IACUC at SLU. sterol homeostasis by affecting both HDL biogenesis (via ABCA1), and bile secretion (via ABCB11 and ATP8B1). Primary hepatocytes Primates, but not rodents, express a second miR-33 gene Cells were isolated from 8–10-week-old, male C57BL/6 mice fed chow, (miR-33b) from an intron of SREBP-1. Importantly, SREBPs are using Perfusion and Digest buffers (Invitrogen). Cells were resuspended differentially regulated by dietary challenges or statin treatment: in William’s E Medium (Invitrogen) supplemented with Plating 892  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. The paper explained PROBLEM: Patients with hypercholesterolemia are usually prescribed with anti-miR-33 oligonucleotides. These latter results suggest statins. We and others reported on miR-33, an intragenic micro- that miR-33 mediates, at least in part, the hepatotoxic effects of RNA that is induced by statins and regulates HDL-cholesterol by statins. modulating the expression of the sterol transporter ABCA1 in the liver. Hence, it was proposed that anti-miR-33 oligonucleotides IMPACT: might be used as a new therapeutic approach to manage Statins induce the expression of miR-33. Patients taking statins hypercholesterolemic patients. However, whether miR-33 also likely have sustained, elevated levels of hepatic miR-33. These controls the expression of other genes important for sterol drugs may decrease the expression of hepatic sterol transporters metabolism is unknown. Here we test the hypothesis that miR-33 indirectly, via miR-33. Several clinical reports showed patients modulates the expression of transporters involved in hepato- who recurrently develop intrahepatic cholestasis when put on biliary secretion. Additionally, we study the role of miR-33 on statins, but recover after statin withdrawal. Cholestasis results statin-induced liver toxicity. from decreased bile secretion and/or flow. Patients with mutations in ATP8B1 and ABCB11 develop PFIC/BRIC-1 and -2, RESULTS: respectively. The data showing the coordinated regulation of MiR-33 extends the function of its host SREBP-2 by reducing sterol transporters by miR-33 and its impact on both HDL and biliary secretion through direct repression of ABCB11 and bile metabolism, add to an ever-growing collection of studies ATP8B1. We identify conserved sequences in the 3 UTR of these showing that miRNAs function as critical fine-tuners in multiple genes responsible for this control, and show that changes in normal and disease-related biological processes. We hypothesize hepatic levels of miR-33 alter both bile secretion rates and bile that the increased expression in ATP8B1 and ABCB11 mediated recovery from the gallbladder. The combination of statins and a by an anti-miR-33 treatment might functionally overcome the lithogenic diet results in a dramatic phenotype (hepatomegaly, partial loss of activity of these transporters observed in BRIC-1 liver steatosis, cholestasis and lethality) in mice, that is rescued and BRIC-2 patients, respectively. Supplements (Invitrogen), plated in 12- or 6-well BioCoat Collagen I Immune complexes were detected with horseradish peroxidase (HRP)- plates (BD), and incubated at 378C and 5% CO for 6 h. Then, the conjugated secondary antibodies (1:5000; Bio-Rad). Due to lack of media was switched to William’s E supplemented with Maintenance specificity we were not able to use these (or different) antibodies to Supplements (Invitrogen). Where indicated, cells were transduced with detect ABCB11 and ATP8B1 in liver samples. Adeno-empty or Adeno-miR-33 adenovirus (MOI 3). Total RNA were extracted 72 h after transduction. Luciferase reporter assays Transfection of Hek293 cells was performed in triplicate in 24-well RNA analysis plates by the calcium phosphate method. Luciferase activity was RNA was isolated from cells or livers with Trizol. Complementary DNAs measured 48 h later using the Luciferase Assay System (Promega), and (cDNAs) were generated from 2 mg of DNase1-treated RNA using normalized to b-galactosidase activity to correct for small changes in Superscript III (Invitrogen) and random hexamers. Real-time PCR was transfection efficiency. done with Power SybrGreen reagent (Applied Biosystems), using a LightCycler-480 (Roche). Primer sets are available on request. Values Plasma analysis were normalized to 36B4 and calculated using the comparative C Circulating levels of ALT, AST, bilirubin and total BAs were determined method. The expression of miR-33 was normalized to U6, using by Advanced Veterinary Laboratory (Saint Louis, MO). MiRCURY miRNA assays (Exiqon). Lipid analysis Protein analysis Liver (50 mg) was homogenized in 500 mL of PBS, and lipids were Protein extracts were obtained from cells or tissues, as described extracted from 100 mL of the homogenate in the presence of internal (Marquart et al, 2010). Fifty micrograms of protein were resolved in standards for each lipid class (Bligh & Dyer, 1959). Similarly, bile and 4–12% Bis–Tris gels, transferred to PVDF membranes, and probed with plasma were extracted in the presence of internal standards for each antibodies for ABCB11 (1:500; a gift from Dr. Renxue Wang from lipid class. Lipid species (phospholipids; TGs; cholesterol esters, CE and the British Columbia Cancer Research Center), ATP8B1 (1:200; SCBT ceramides) were quantified directly from lipid organic extracts using sc-67712), ABCA1 (1:1000; Novus NB400-105), ABCG5 (1:200; Novus shotgun lipidomics based on class separation by MS/MS specific NBP1-95209), FXR (1:200; SCBT sc-13063), SREBP1 (1:100; SCBT sc- methods (Ford et al, 2008; Han & Gross, 2001, 2005). Bile and liver 13551), b-actin (1:5000; SCBT sc-130656) and a-tubulin (1:1000; concentrations of cholesterol, bile salts and PC were also determined Sigma T5168), in TBS-Tween20 containing 4% non-fat dry milk. using enzymatic kits from Wako Chemicals. EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 893 Research Article www.embomolmed.org Control of hepatic bile secretion by miR-33 Davalos A, Goedeke L, Smibert P, Ramirez CM, Warrier NP, Andreo U, Cirera- Author contributions Salinas D, Rayner K, Suresh U, Pastor-Pareja JC, et al (2011) miR-33a/b RMA conceived the hypothesis, designed and performed contribute to the regulation of fatty acid metabolism and insulin signaling. experiments, analyzed the data and wrote the manuscript. Proc Natl Acad Sci USA 108: 9232-9237 TJM designed, performed and analyzed the RCT experiment. Davit-Spraul A, Fabre M, Branchereau S, Baussan C, Gonzales E, Stieger B, CJA, RMA and DAF performed and analyzed the lipidomics Bernard O, Jacquemin E (2010) ATP8B1 and ABCB11 analysis in 62 children studies. MA and FJS performed mRNA experiments in HuH-7 with normal gamma-glutamyl transferase progressive familial intrahepatic cells. DQ-HW performed common bile duct cannulations for cholestasis (PFIC): phenotypic differences between PFIC1 and PFIC2 and natural history. Hepatology 51: 1645-1655 secretion experiments. AB conceived the hypothesis, designed de Castro ML, Hermo JA, Baz A, de Luaces C, Perez R, Clofent J (2006) Acute experiments and wrote the manuscript. cholestatic hepatitis after atorvastatin reintroduction. Gastroenterol Hepatol 29: 21-24 Emerick KM, Whitington PF (2002) Partial external biliary diversion for Acknowledgements intractable pruritus and xanthomas in Alagille syndrome. Hepatology 35: We thank Erin Touchette for help with i.v. injections, Dr. Eva 1501-1506 van Rooij (miRagen Therapeutics) for providing LNA oligos, Englert C, Grabhorn E, Burdelski M, Ganschow R (2006) Liver transplantation in children with Alagille syndrome: indications and outcome. Pediatr and Dr. Robert Brown (Memorial University of Newfoundland) Transplant 10: 154-158 for discussions for RCT experiments. This work was supported Esteller A (2008) Physiology of bile secretion. World J Gastroenterol 14: 5641- in part by AHA Grant 11SDG5030012 (to A.B.), and NIH Grants HL107794 (to A.B.), HL074214, HL098907 and RR019232 (to Ford DA, Monda JK, Brush RS, Anderson RE, Richards MJ, Fliesler SJ (2008) D.A.F.) and DK084434 (to F.J.S.). R.M.A. is the recipient of an Lipidomic analysis of the retina in a rat model of Smith–Lemli–Opitz AHA Predoctoral Fellowship (11PRE7240026). syndrome: alterations in docosahexaenoic acid content of phospholipid molecular species. J Neurochem 105: 1032-1047 Frankenberg T, Miloh T, Chen FY, Ananthanarayanan M, Sun AQ, Supporting Information is available at EMBO Molecular Balasubramaniyan N, Arias I, Setchell KD, Suchy FJ, Shneider BL (2008) The Medicine online. membrane protein ATPase class I type 8B member 1 signals through protein kinase C zeta to activate the farnesoid X receptor. Hepatology 48: 1896- Conflict of interest statement: RMA, TJM and AB are pursuing a patent related to this work. The other authors declare Gadaleta RM, van Mil SW, Oldenburg B, Siersema PD, Klomp LW, van Erpecum that they have no conflict of interest. KJ (2010) Bile acids and their nuclear receptor FXR: relevance for hepatobiliary and gastrointestinal disease. Biochim Biophys Acta 1801: 683-692 Gerin I, Clerbaux LA, Haumont O, Lanthier N, Das AK, Burant CF, Leclercq IA, References MacDougald OA, Bommer GT (2010) Expression of miR-33 from an SREBP2 Abu Rajab M, Kaplan MM (2010) Statins in primary biliary cirrhosis: are they intron inhibits cholesterol export and fatty acid oxidation. J Biol Chem 285: safe? Dig Dis Sci 55: 2086-2088 33652-33661 Alvarez L, Jara P, Sanchez-Sabate E, Hierro L, Larrauri J, Diaz MC, Camarena C, Han X, Gross RW (2001) Quantitative analysis and molecular species De La Vega A, Frauca E, Lopez-Collazo E et al (2004) Reduced hepatic fingerprinting of triacylglyceride molecular species directly from lipid expression of Farnesoid X Receptor in hereditary cholestasis associated to extracts of biological samples by electrospray ionization tandem mass mutation in ATP8B1. Hum Mol Genet 13: 2451-2460 spectrometry. Anal Biochem 295: 88-100 Annema W, Dikkers A, Freark de Boer J, Gautier T, Rensen PC, Rader DJ, Tietge Han X, Gross RW (2005) Shotgun lipidomics: electrospray ionization mass UJ (2012) ApoE promotes hepatic selective uptake but not RCT due to spectrometric analysis and quantitation of cellular lipidomes directly from increased ABCA1-mediated cholesterol efflux to plasma. J Lipid Res 53: crude extracts of biological samples. Mass Spectrom Rev 24: 367-412 929-940 Harris MJ, Le Couteur DG, Arias IM (2005) Progressive familial intrahepatic Batey RG, Harvey M (2002) Cholestasis associated with the use of pravastatin cholestasis: genetic disorders of biliary transporters. J Gastroenterol sodium. Med J Aust 176: 561 Hepatol 20: 807-817 Bennett MK, Seo YK, Datta S, Shin DJ, Osborne TF (2008) Selective binding of Hori T, Nguyen JH, Uemoto S (2010) Progressive familial intrahepatic sterol regulatory element-binding protein isoforms and co-regulatory cholestasis. Hepatobiliary Pancreat Dis Int 9: 570-578 proteins to promoters for lipid metabolic genes in liver. J Biol Chem 283: Horie T, Ono K, Horiguchi M, Nishi H, Nakamura T, Nagao K, Kinoshita M, 15628-15637 Kuwabara Y, Marusawa H, Iwanaga Y, et al (2010) MicroRNA-33 encoded by Bligh EG, Dyer WJ (1959) A rapid method of total lipid extraction and an intron of sterol regulatory element-binding protein 2 (Srebp2) regulates purification. Can J Med Sci 37: 911-917 HDL in vivo. Proc Natl Acad Sci USA 107: 17321-17326 Cai SY, Gautam S, Nguyen T, Soroka CJ, Rahner C, Boyer JL (2009) ATP8B1 Horton JD, Bashmakov Y, Shimomura I, Shimano H (1998) Regulation of sterol deficiency disrupts the bile canalicular membrane bilayer structure in regulatory element binding proteins in livers of fasted and refed mice. Proc hepatocytes, but FXR expression and activity are maintained. Natl Acad Sci USA 95: 5987-5992 Gastroenterology 136: 1060-1069 Hubacek JA, Berge KE, Cohen JC, Hobbs HH (2001) Mutations in ATP-cassette Chen F, Ananthanarayanan M, Emre S, Neimark E, Bull LN, Knisely AS, binding proteins G5 (ABCG5) and G8 (ABCG8) causing sitosterolemia. Hum Strautnieks SS, Thompson RJ, Magid MS, Gordon R, et al (2004) Mutat 18: 359-360 Progressive familial intrahepatic cholestasis, type 1, is associated Khanuja B, Cheah YC, Hunt M, Nishina PM, Wang DQ, Chen HW, Billheimer JT, with decreased farnesoid X receptor activity. Gastroenterology 126: Carey MC, Paigen B (1995) Lith1, a major gene affecting cholesterol 756-764 gallstone formation among inbred strains of mice. Proc Natl Acad Sci USA Chen F, Ellis E, Strom SC, Shneider BL (2010) ATPase Class I Type 8B Member 1 92: 7729-7733 and protein kinase C zeta induce the expression of the canalicular bile salt Khera AV, Rader DJ (2010) Future therapeutic directions in reverse cholesterol export pump in human hepatocytes. Pediatr Res 67: 183-187 transport. Curr Atheroscler Rep 12: 73-81 894  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2012) 4, 882–895 www.embomolmed.org Research Article Ryan M. Allen et al. Koh S, Takada T, Kukuu I, Suzuki H (2009) FIC1-mediated stimulation of FXR human primates raises plasma HDL and lowers VLDL triglycerides. Nature activity is decreased with PFIC1 mutations in HepG2 cells. J Gastroenterol 478: 404-407 44: 592-600 Rayner KJ, Sheedy FJ, Esau CC, Hussain FN, Temel RE, Parathath S, van Gils JM, Kubitz R, Keitel V, Haussinger D (2005) Inborn errors of biliary canalicular Rayner AJ, Chang AN, Suarez Y, et al (2011b) Antagonism of miR-33 in mice transport systems. Methods Enzymol 400: 558-569 promotes reverse cholesterol transport and regression of atherosclerosis. J Lammert F, Wang DQ, Hillebrandt S, Geier A, Fickert P, Trauner M, Matern S, Clin Invest 121: 2921-2931 Paigen B, Carey MC (2004) Spontaneous cholecysto- and hepatolithiasis in Ridruejo E, Mando OG (2002) Acute cholestatic hepatitis after reinitiating Mdr2-/- mice: a model for low phospholipid-associated cholelithiasis. treatment with atorvastatin. J Hepatol 37: 165-166 Hepatology 39: 117-128 Shah S, Sanford UR, Vargas JC, Xu H, Groen A, Paulusma CC, Grenert JP, Marquart TJ, Allen RM, Ory DS, Baldan A (2010) miR-33 links SREBP-2 Pawlikowska L, Sen S, Elferink RP, et al (2010) Strain background modifies induction to repression of sterol transporters. Proc Natl Acad Sci USA 107: phenotypes in the ATP8B1-deficient mouse. PLoS One 5: e8984 12228-12232 Stanca CM, Bach N, Allina J, Bodian C, Bodenheimer H, Jr., Odin JA (2008) Martinez-Fernandez P, Hierro L, Jara P, Alvarez L (2009) Knockdown of ATP8B1 Atorvastatin does not improve liver biochemistries or Mayo Risk Score in expression leads to specific downregulation of the bile acid sensor FXR in primary biliary cirrhosis. Dig Dis Sci 53: 1988-1993 HepG2 cells: effect of the FXR agonist GW4064. Am J Physiol Gastrointest Stojakovic T, Putz-Bankuti C, Fauler G, Scharnagl H, Wagner M, Stadlbauer V, Liver Physiol 296: G1119-G1129 Gurakuqi G, Stauber RE, Marz W, Trauner M (2007) Atorvastatin in patients Merli M, Bragazzi MC, Giubilo F, Callea F, Attili AF, Alvaro D (2010) with primary biliary cirrhosis and incomplete biochemical response to Atorvastatin-induced prolonged cholestasis with bile duct damage. Clin ursodeoxycholic acid. Hepatology 46: 776-784 Drug Investig 30: 205-209 Stojakovic T, Claudel T, Putz-Bankuti C, Fauler G, Scharnagl H, Wagner M, Morotti RA, Suchy FJ, Magid MS (2011) Progressive familial intrahepatic Sourij H, Stauber RE, Winkler K, Marz W, et al (2010) Low-dose atorvastatin cholestasis (PFIC) type 1, 2, and 3: a review of the liver pathology findings. improves dyslipidemia and vascular function in patients with primary Semin Liver Dis 31: 3-10 biliary cirrhosis after one year of treatment. Atherosclerosis 209: 178-183 Moschetta A, Bookout AL, Mangelsdorf DJ (2004) Prevention of cholesterol Temel RE, Sawyer JK, Yu L, Lord C, Degirolamo C, McDaniel A, Marshall S, Wang gallstone disease by FXR agonists in a mouse model. Nat Med 10: N, Shah R, Rudel LL, et al (2010) Biliary sterol secretion is not required for 1352-1358 macrophage reverse cholesterol transport. Cell Metab 12: 96-102 Najafi-Shoushtari SH, Kristo F, Li Y, Shioda T, Cohen DE, Gerszten RE, Naar AM Tietge UJ, Nijstad N, Havinga R, Baller JF, van der Sluijs FH, Bloks VW, Gautier T, (2010) MicroRNA-33 and the SREBP host genes cooperate to control Kuipers F (2008) Secretory phospholipase A2 increases SR-BI-mediated cholesterol homeostasis. Science 238: 1566-1569 selective uptake from HDL but not biliary cholesterol secretion. J Lipid Res Nijstad N, Wiersma H, Gautier T, van der Giet M, Maugeais C, Tietge UJ (2009) 49: 563-571 Scavenger receptor BI-mediated selective uptake is required for the Torres M, Sobrino J, Asensio C, Lopez D (2002) Acute cholestatic hepatitis remodeling of high density lipoprotein by endothelial lipase. J Biol Chem induced by cerivastatin. Med Clin 118: 717 284: 6093-6100 Ujhazy P, Ortiz D, Misra S, Li S, Moseley J, Jones H, Arias IM (2001) Familial Nijstad N, Gautier T, Briand F, Rader DJ, Tietge UJ (2011) Biliary sterol intrahepatic cholestasis 1: studies of localization and function. Hepatology secretion is required for functional in vivo reverse cholesterol transport in 34: 768-775 mice. Gastroenterology 140: 1043-1051 van der Woerd WL, van Mil SW, Stapelbroek JM, Klomp LW, van de Graaf SF, Paulusma CC, Groen A, Kunne C, Ho-Mok KS, Spijkerboer AL, Rudi de Waart D, Houwen RH (2010) Familial cholestasis: progressive familial intrahepatic Hoek FJ, Vreeling H, Hoeben KA, van Marle J, et al (2006) Atp8b1 deficiency cholestasis, benign recurrent intrahepatic cholestasis and intrahepatic in mice reduces resistance of the canalicular membrane to hydrophobic bile cholestasis of pregnancy. Best Pract Res Clin Gastroenterol 24: 541-553 salts and impairs bile salt transport. Hepatology 44: 195-204 Wang X, Rader DJ (2007) Molecular regulation of macrophage reverse Paulusma CC, Folmer DE, Ho-Mok KS, de Waart DR, Hilarius PM, Verhoeven AJ, cholesterol transport. Curr Opin Cardiol 22: 368-372 Oude Elferink RP (2008) ATP8B1 requires an accessory protein for Wang R, Lam P, Liu L, Forrest D, Yousef IM, Mignault D, Phillips MJ, Ling V endoplasmic reticulum exit and plasma membrane lipid flippase activity. (2003) Severe cholestasis induced by cholic acid feeding in knockout mice Hepatology 47: 268-278 of sister of P-glycoprotein. Hepatology 38: 1489-1499 Pawlikowska L, Groen A, Eppens EF, Kunne C, Ottenhoff R, Looije N, Knisely AS, Wang X, Collins HL, Ranalletta M, Fuki IV, Billheimer JT, Rothblat GH, Tall AR, Killeen NP, Bull LN, Elferink RP, et al (2004) A mouse genetic model for Rader DJ (2007) Macrophage ABCA1 and ABCG1, but not SR-BI, promote familial cholestasis caused by ATP8B1 mutations reveals perturbed bile salt macrophage reverse cholesterol transport in vivo. J Clin Invest 117: 2216- homeostasis but no impairment in bile secretion. Hum Mol Genet 13: 881- 2224 892 Wiersma H, Gatti A, Nijstad N, Kuipers F, Tietge UJ (2009) Hepatic SR-BI, not Portincasa P, Palasciano G, Svelto M, Calamita G (2008) Aquaporins in the endothelial lipase, expression determines biliary cholesterol secretion in hepatobiliary tract. Which, where and what they do in health and disease. mice. J Lipid Res 50: 1571-1580 Eur J Clin Invest 38: 1-10 Yu L, Hammer RE, Li-Hawkins J, Von Bergmann K, Lutjohann D, Cohen JC, Rader DJ, Alexander ET, Weibel GL, Billheimer J, Rothblat GH (2009) The role of Hobbs HH (2002) Disruption of Abcg5 and Abcg8 in mice reveals their reverse cholesterol transport in animals and humans and relationship to crucial role in biliary cholesterol secretion. Proc Natl Acad Sci USA 99: atherosclerosis. J Lipid Res 50 Suppl: S189-S194 16237-16242 Rahier JF, Rahier J, Leclercq I, Geubel AP (2008) Severe acute cholestatic Yu L, Gupta S, Xu F, Liverman AD, Moschetta A, Mangelsdorf DJ, Repa JJ, Hobbs hepatitis with prolonged cholestasis and bile-duct injury following HH, Cohen JC (2005) Expression of ABCG5 and ABCG8 is required for atorvastatin therapy: a case report. Acta Gastroenterol Belg 71: 318-320 regulation of biliary cholesterol secretion. J Biol Chem 280: 8742-8747 Rayner KJ, Suarez Y, Davalos A, Parathath S, Fitzgerald ML, Tamehiro N, Fisher Zadelaar S, Kleemann R, Verschuren L, de Vries-Van der Weij J, van der Hoorn J, EA, Moore KJ, Fernandez-Hernando C (2010) miR-33 contributes to the Princen HM, Kooistra T (2007) Mouse models for atherosclerosis and regulation of cholesterol homeostasis. Science 238: 1570-1573 pharmaceutical modifiers. Arterioscler Thromb Vasc Biol 27: 1706-1721 Rayner KJ, Esau CC, Hussain FN, McDaniel AL, Marshall SM, van Gils JM, Ray TD, Zhang Y, Edwards PA (2008) FXR signaling in metabolic disease. FEBS Lett 582: Sheedy FJ, Goedeke L, Liu X, et al (2011a) Inhibition of miR-33a/b in non- 10-18 EMBO Mol Med (2012) 4, 882–895  2012 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 895

Journal

EMBO Molecular Medicine – Pubmed Central

Published: Jul 5, 2012

There are no references for this article.