Atorvastatin induces bile acid-synthetic enzyme Cyp7a1 by suppressing FXR signaling in both liver and intestine in mice.

Statins are effective cholesterol-lowering drugs to treat CVDs. Bile acids (BAs), the end products of cholesterol metabolism in the liver, are important nutrient and energy regulators. The present study aims to investigate how statins affect BA homeostasis in the enterohepatic circulation. Male C57BL/6 mice were treated with atorvastatin (100 mg/kg/day po) for 1 week, followed by BA profiling by ultra-performance LC-MS/MS. Atorvastatin decreased BA pool size, mainly due to less BA in the intestine. Surprisingly, atorvastatin did not alter total BAs in the serum or liver. Atorvastatin increased the ratio of 12α-OH/non12α-OH BAs. Atorvastatin increased the mRNAs of the BA-synthetic enzymes cholesterol 7α-hydroxylase (Cyp7a1) (over 10-fold) and cytochrome P450 27a1, the BA uptake transporters Na⁺/taurocholate cotransporting polypeptide and organic anion transporting polypeptide 1b2, and the efflux transporter multidrug resistance-associated protein 2 in the liver. Noticeably, atorvastatin suppressed the expression of BA nuclear receptor farnesoid X receptor (FXR) target genes, namely small heterodimer partner (liver) and fibroblast growth factor 15 (ileum). Furthermore, atorvastatin increased the mRNAs of the organic cation uptake transporter 1 and cholesterol efflux transporters Abcg5 and Abcg8 in the liver. The increased expression of BA-synthetic enzymes and BA transporters appear to be a compensatory response to maintain BA homeostasis after atorvastatin treatment. The Cyp7a1 induction by atorvastatin appears to be due to suppressed FXR signaling in both the liver and intestine.

liver and also represses the transcription of BA-synthetic enzymes ( 20 ).
BA transporters play pivotal roles in maintaining BA homeostasis. Ntcp mediates the uptake of conjugated BAs, whereas Oatp1b2 transports unconjugated BAs ( 21 ). Bsep, which is generally accepted as the main BA effl ux transporter in the liver, mediates the biliary excretion of monovalent conjugated BAs (such as TCA) and has a relatively poor affi nity for unconjugated BAs ( 22 ). The fi nding that Mrp2 mediates the biliary excretion of divalent conjugated BAs (such as tauro-BA-sulfates, TCDCA-S and TLCA-S) ( 23 ) suggests that Mrp2 is a promising candidate to be an alternative canalicular BA transporter. It is debatable whether Bcrp transports BAs, because of contradictory fi ndings ( 24,25 ).
Because cholesterol is the substrate for de novo synthesis of BAs in the liver, we hypothesized that the cholesterollowering drugs could also affect BA homeostasis, and the changes would be mediated by the coordinated regulation of BA-synthetic enzymes and/or transporters. Therefore, the present study utilized a highly sensitive and accurate ultra-performance (UP)LC-MS/MS method recently developed in our laboratory ( 26 ) to characterize BA profi les in various compartments of the BA EHC in mice treated with atorvastatin, and determined the expression of BArelated genes in the liver and intestine to reveal possible regulatory mechanisms. This study also aims to provide the fi rst in vivo evidence for the effect of statins on the expression of transporters in the liver.

Chemicals and reagents
Atorvastatin was purchased from Cayman Chemical (Ann Arbor, MI). The sources of internal standards (ISs) and various taurine-conjugated and unconjugated BA standards were the same as previously reported ( 26 ). All other chemicals, unless indicated, were purchased from Sigma-Aldrich (St. Louis, MO).

Animals and treatments
The animal study was approved by the Institutional Animal Care and Use Committee at the University of Kansas Medical Center. Eight-week-old male C57BL/6 mice were purchased from Charles River Laboratories (Wilmington, MA). Mice were acclimated for 1 month and fed Teklad Rodent Diet #5053 (Harlan Laboratories, Madison, WI) ad libitum according to the guidance of the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC). Mice (n = 5) were administered vehicle (5 mg/ml ␤ -cyclodextrin in saline) or atorvastatin (100 mg/kg/day) po for seven consecutive days. The dose was chosen based on previous publications ( 27,28 ). Mice were anesthetized 24 h after the fi nal dose, blood was collected by orbital bleeding, and serum was collected by centrifuging blood at 8,000 rpm for 15 min. The liver and GB (tissue with bile inside) were removed, frozen separately in liquid nitrogen, and stored at Ϫ 80°C. The SI was fl ushed into 10 ml of saline to collect SI contents, and the large intestine (LI) was fl ushed into another 10 ml of saline to collect LI contents. The three SI segments, namely duodenum, jejunum, and ileum, as well as the LI, were stored separately at Ϫ 80°C. Abcg2) ( 9 ), and bile salt export pump (Bsep/Abcb11) ( 10 ). The genetic polymorphisms of uptake (OATP1B1, 1B3) and effl ux transporters (MRP2, BCRP) that are responsible for inter-individual differences in the pharmacokinetics and pharmacodynamics of statins are well-known in humans ( 4 ). The effect of statins on the expression of several transporters has been investigated in cell lines ( 11 ). The development of transporter-defi cient rodent models provides evidence that Oatp1b2 ( 6 ), Mrp2 ( 12 ), and Bcrp ( 13 ) play important roles in statin transport in the livers of intact animals. However, the in vivo data on the effect of statin treatment on the expression of various transporters in the liver are relatively limited in laboratory animals and in humans.
Another knowledge gap is the role of statins in regulating the homeostasis of bile acids (BAs), which are end products from cholesterol metabolism in the liver. In addition to their roles in cholesterol elimination and nutrient absorption, BAs are known to be signaling molecules that regulate nutrient and energy metabolism ( 14 ). BAs are synthesized in the liver (called primary BAs), stored in the gallbladder (GB), secreted into the intestine, reabsorbed at the end of the small intestine (SI), and subsequently returned to the liver. This cycle is called the enterohepatic circulation (EHC) of BAs. The BA reabsorption from the SI is very effi cient, which is more than 95% effi cient in humans. The rest ( ‫ف‬ 5%) of BAs are eliminated into feces, the amount of which is often used to estimate the production of BAs in the liver.
Primary BAs are synthesized in the liver and conjugated with taurine and glycine on the side chain, or sulfate on the steroid nucleus ( 15 ). Cholic acid (CA) and chenodeoxycholic acid (CDCA), as well as ␣ -and ␤ -muricholic acids (MCAs), are the primary BAs in mice. These primary BAs are converted by bacterial enzymes in the intestine through deconjugation, dehydroxylation, epimerization, and oxidation to form secondary BAs, such as deoxycholic acid (DCA), lithocholic acid (LCA), and ursodeoxycholic acid (UDCA), as well as MCA, murideoxycholic acid (MDCA), and hyodeoxycholic acid (HDCA), as previously summarized ( 16 ).
Several enzymes catalyze BA synthesis in the liver, either through the classic or alternative pathways. Cholesterol 7 ␣ -hydroxylase (Cyp7a1) is the rate-limiting enzyme in the classic pathway ( 17 ), which results in the formation of both CA and CDCA. Cytochrome P450 (Cyp)8b1 is responsible for the 12 ␣ -hydroxylation of BAs, or the formation of CA. Mitochondrial Cyp27a1 catalyzes the side-chain oxidation of cholesterol. The alternative pathway is initiated by Cyp27a1 and followed by Cyp7b1 and Cyp39a1 to produce CDCA. The expression of key BA-synthetic enzymes (Cyp7a1 and Cyp8b1) can be regulated transcriptionally by BA-feedback inhibition, mainly through two mechanisms ( 18 ). One is the BA nuclear receptor farnesoid X receptor (FXR/Nr1b4) and its target gene small heterodimer partner (SHP/Nr0b2) pathway in the liver ( 19 ). The other is activation of FXR in the intestine and the induction of its target gene fi broblast growth factor (Fgf)15, which is an intestinal hormone that travels to the

Total RNA isolation
Total RNA was isolated using RNA Bee reagent (Tel-Test Inc., Friendswood, TX) per the manufacturer's protocol. RNA concentrations were quantifi ed using a NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, DE) at a wavelength of 260 nm.

Multiplex suspension assay
The mRNA expression of BA-synthetic enzymes and transporters in control and statin-treated mouse livers were determined with Panomics 2.0 QuantiGene Plex technology (Panomics/Affymetrix, Fremont, CA), following the manufacturer's protocol. Briefl y, individual bead-based oligonucleotide probe sets specifi c for each gene examined were developed by Panomics Inc. Genes that encode BA-synthetic enzymes and uptake and effl ux transporters in the liver can be found on Panomics website under panel number 21021 . Samples were analyzed using a Bio-Plex 200 system array reader with Luminex 100 X-MAP technology, and data were acquired using Bio-Plex data manager software version 5.0 (Bio-Rad, Hercules, CA). Assays were performed according to the manufacturer's protocol. The mRNA expression of target genes was expressed as relative light units (RLU) and normalized to the housekeeping gene, Gapdh.

Reverse transcription and real-time PCR analysis
Total RNA was transcribed to single-stranded cDNA using high capacity cDNA reverse transcription kit 1001073 (Applied Biosystems, Foster City, CA), and then the cDNA products were amplifi ed by PCR, using Power SYBR Green PCR Master Mix in a 7900HT Fast real-time PCR system (Applied Biosystems). The mRNAs of genes encoding BA conjugating enzymes [bile acid-CoA ligase (BAL) and bile acid CoA:amino acid N -acyl-transferase (BAT)], BA transporters in the ileum [apical Na + -dependent bile salt transporter (Asbt), organic solute tranporter (Ost) ␣ , Ost ␤ , and Mrp3], ileal BA binding protein (Ibabp), and proteins involved in Cyp7a1 regulation (FXR, SHP, Fgf15, and Fgfr4) were quantifi ed, and normalized to ␤ -actin. The sequences of real-time PCR primers (Integrated DNA Technologies, Coralville, IA) are listed in supplementary Table I.

Statistical analysis
Data are presented as mean ± SEM in bar charts, and mean in pie charts. Data were analyzed by Student's t -test (two-tail), differences being considered signifi cant at P < 0.05. Asterisks (*) represent signifi cant differences compared with vehicle-control ( P < 0.05).

BA pool size and BA distribution in atorvastatin-treated mice
Statins lower cholesterol in the blood by inhibiting cholesterol synthesis, and because BAs are the end products of cholesterol metabolism, statins might also decrease BA synthesis and BAs in the body. Therefore, it was of interest to determine whether atorvastatin alters BA homeostasis in BA extraction from serum 2 H 4 -G-CDCA and 2 H 4 -CDCA (40 g/ml and 20 g/ml, respectively) were used as ISs. Simple protein precipitation with methanol (MeOH) was used for BA extraction from serum. One milliliter of MeOH was added to 50 l of serum-spiked with 5 l of IS, vortexed, and centrifuged at 12,000 g for 10 min. The supernatant was aspirated, evaporated under vacuum, and reconstituted in 50 l of 50% MeOH. Samples were centrifuged at 20,000 g for 10 min before injection.

BA extraction from the liver
A piece of liver ( ‫ف‬ 120 mg) was homogenized in 5 vol of water, from which 600 l of homogenate was taken and mixed with 10 l of IS. After 10 min equilibration on ice, the homogenate was mixed with 3 ml of ice-cold alkaline acetonitrile (5% ammonia), vortexed vigorously, and shaken for 1 h at room temperature. The mixture was centrifuged at 12,000 g for 10 min, and the supernatant was collected. The pellet was extracted with 1 ml of MeOH, sonicated for 5 min, and centrifuged at 12,000 g for 10 min. The two supernatants were pooled, evaporated under vacuum, and reconstituted in 100 l of 50% MeOH. The suspension was transferred into a 0.2 m Costar Spin-X HPLC microcentrifuge fi lter (purchased from Corning Inc., Corning, NY), and centrifuged at 20,000 g for 10 min. The supernatant was then ready for injection.

BA extraction from the GB
One milliliter of MeOH was added to each GB, which was broken to release the bile inside and premixed with 100 l of IS. After vigorous vortexing and 10 min sonication, the mixture was centrifuged at 16,000 g for 10 min, and the supernatant was collected. The pellet was extracted with another 2 ml of MeOH. The two supernatants were combined, evaporated under vacuum, and reconstituted in 1 ml of 50% MeOH.

BA extraction from intestinal contents
Intestinal contents were mixed with 100 l of IS and centrifuged at 12,000 g for 10 min to collect the supernatant. The pellet was extracted with 3 ml of MeOH twice. After shaking for 30 min at room temperature, the mixture was centrifuged at 12,000 g for 20 min to collect the supernatant. The three supernatants were pooled, evaporated under vacuum, and reconstituted in 1 ml of 50% MeOH. The suspension was fi ltered before injection.

BA extraction from feces
Mice (n = 5) were acclimated to wire-bottomed metabolic cages for 48 h (housed individually), and feces were collected over a 24 h period. Mouse feces were dried under vacuum and ground to powder. Fifty milligrams of feces were mixed with 10 l IS and 3 ml of MeOH was added. After shaking for 1 h at room temperature, the mixture was centrifuged at 20,000 g for 10 min to collect the supernatant. The pellet was extracted with another 2 ml of MeOH. The two supernatants were pooled, evaporated under vacuum, and reconstituted in 100 l of 50% MeOH. The suspension was fi ltered before injection.

BA quantifi cation
BA concentrations were quantifi ed by a highly sensitive and accurate method established in our laboratory using UPLC-MS/ MS ( 26 ). The conditions of LC and MS were the same as previously reported ( 26 ). BAs were quantifi ed in serum, liver, GB, SI, LI, and feces, as shown in Fig. 3 . In serum, atorvastatin did not change concentrations of BAs in the CA or ␣ MCA pathway, did not change most BAs in the CDCA pathway, except increasing CDCA, and decreased several BAs in the ␤ MCA pathway, including T ␤ MCA (69%) and its secondary BA, T MCA (100%), as well as ␤ MCA (96.6%) and MCA (74.6%). In the liver, atorvastatin did not change the mice. In this study, the dose of atorvastatin (100 mg/kg/day for 7 days) did not cause hepatotoxicity, as evidenced by normal levels of alanine aminotransferase in serum (data not shown). To reveal the overall effects of atorvastatin on BA homeostasis, BA pool size was estimated by adding total BAs in the main compartments of the EHC, namely the liver, GB, and intestine (SI and LI), as published previously ( 29 ). As shown in Fig. 1A, B , atorvastatin decreased BA pool size (35.1%), which was primarily due to the decrease in total BAs in the SI (35%) and LI (56.4%). Atorvastatin also tended to increase fecal BA excretion (53.3%) ( Fig. 1C ). The BA changes in the SI, as well as the fecal excretion, were not statistically signifi cant in the control and statin-treated mice. Furthermore, there were no differences in the body weight or the daily fecal excretion between the control and statin-treated mice (data not shown).
Because BAs in serum determine the effects of BAs on peripheral tissues and statins inhibit cholesterol synthesis in the liver, we focused on the changes of BA composition by statins in the serum and liver. Individual BAs have different hydrophobicity, and thus changes in BA composition could lead to altered hydrophobicity of total BAs. The hydrophobicity index (HI) of individual BAs and their proportions in biological samples are used to predict the HI of total BAs at physiological pH, as published previously ( 29 ). Atorvastatin increased the HI of total BAs in the serum and liver ( Fig. 5D ). However, total BAs remained hydrophilic in the serum and liver after atorvastatin treatment.

mRNA expression of BA-synthetic enzymes in the livers of atorvastatin-treated mice
The present observation that statins decrease cholesterol (the substrate for BA synthesis) but do not decrease total BAs in either the liver or serum, suggests that there are compensatory mechanisms to maintain BA homeostasis. The mRNA of Cyp7a1, which encodes the rate-limiting enzyme for the classic pathway of BA synthesis in the liver, was markedly increased by atorvastatin (11.1-fold) ( Fig. 6 ), whereas the mRNA of Cyp8b1, which is responsible for CA formation, remained unchanged after atorvastatin treatment. For the alternative pathway of BA synthesis, the mRNA of Cyp27a1 in the liver was increased by atorvastatin (79.9%). The mRNA of Cyp7b1 in the liver was decreased 44.1% by atorvastatin. The mRNA of Cyp39a1 in the liver was increased by atorvastatin (76.8%). In summary, the mRNAs of the rate-limiting enzyme in the classic pathway (Cyp7a1) and the initial enzyme in the alternative pathway (Cyp27a1) were both increased by atorvastatin treatment, whereas other genes for BA synthesis were only moderately altered (Cyp7b1 and 39a1) or remained unchanged (Cyp8b1).

mRNA expression of regulators for Cyp7a1 transcription in atorvastatin-treated mice
In order to determine the possible regulatory mechanisms for the robust increase in Cyp7a1 mRNA expression by statins, the mRNA expression of several key regulators that are involved in the feedback inhibition of Cyp7a1 were quantifi ed in the liver and ileum. Atorvastatin decreased the mRNAs of SHP (53.6%) in the liver and Fgf15 (83.6%) in the ileum ( Fig. 7 ). Atorvastatin tended to increase FXR mRNA in the liver, but it was not statistically signifi cant. The mRNA levels of Fgfr4 in the liver and FXR in the ileum remained relatively constant after atorvastatin treatment.

mRNA expression of various transporters in the liver and ileum of atorvastatin-treated mice
In addition to BA-synthetic enzymes, BA transporters play a crucial role in promoting the EHC of BAs. Thus, the mRNA expression of BA transporters was quantifi ed in the liver and ileum of statin-treated mice, as well as other BAs in the CDCA pathway in feces, except decreasing TUDCA (52.8%) and increasing CDCA (187%). Atorvastatin did not alter the amount of most BAs in the ␣ MCA pathway in feces, except increasing ␣ MCA (300%). Atorvastatin decreased T ␤ MCA (88.2%) and ␤ MCA (53.7%), but did not alter other BAs in the ␤ MCA pathway in feces.

BA composition in atorvastatin-treated mice
In addition to BA concentration, BA composition is another important factor that leads to different biological effects. For instance, individual BAs have various potencies to activate their receptors, FXR (CDCA > LCA, DCA > CA) and TGR5 (LCA > DCA > CDCA > CA ) ( 30 ). Therefore, it is important to analyze BA composition in statin-treated mice. Atorvastatin did not alter the proportions of total taurine-conjugated or total unconjugated BAs in serum ( Fig. 4 ). In regard to taurine-conjugated BAs in serum, atorvastatin increased the proportion of TCA (43.5 → 68.6%), and decreased the proportions of T ␤ MCA (28.8 → 9.8%) and T MCA (8.1 → 2.1%). In regard to unconjugated BAs in serum, atorvastatin decreased the proportion of ␤ MCA (24.0 → 7.3%), but not other unconjugated BAs ( Fig. 5A ).
In the liver, atorvastatin increased the proportion of total taurine-conjugated BAs (96 → 97.8%) and decreased the proportion of total unconjugated BAs (4 → 2.2%) ( Fig. 4 ). Atorvastatin increased the proportion of TCA and decreased the proportions of T ␤ MCA and T MCA in total taurine-conjugated BAs in the liver. Atorvastatin also increased the proportions of CDCA and ␣ MCA and decreased the proportions of DCA and ␤ MCA in total unconjugated BAs in liver ( Fig. 5B ).
The ratio of 12 ␣ -OH versus non12 ␣ -OH BAs, which is an important aspect of BA composition, was analyzed in this study. Atorvastatin increased the ratio of 12 ␣ -OH versus synthesis in the liver, total BA concentrations in the serum and liver were not decreased by atorvastatin ( Figs. 1, 2 ). However, atorvastatin decreases the BA pool size, which is primarily due to lower BAs in the intestine. The fecal BA excretion usually equals the amount of newly biosynthesized BAs in the liver. The total amount of BAs synthesized in the liver is thought to be regulated by Cyp7a1. In this study, atorvastatin increased fecal BA excretion, which is consistent with the upregulation of Cyp7a1 expression in the liver. A marked decrease (over 50%) in unconjugated BAs was observed in the large intestinal contents of mice treated by atorvastatin ( Fig. 2 ), which correspondingly leads to an increased ratio of conjugated versus unconjugated BAs. Conjugated BAs are more hydrophilic, and thus less ready to cross cell membranes by passive diffusion.

mRNA expression of BA-conjugating enzymes in the liver, as well as proteins responsible for BA reabsorption in the ileum of atorvastatin-treated mice
The mRNA expression of BA-conjugating enzymes (BAL and BAT) was not altered by atorvastatin. In the ileum, none of the transporters or proteins involved in BA reabsorption (Asbt, Ibabp, Ost ␣ , Ost ␤ , or Mrp3) had altered expression after atorvastatin treatment ( Fig. 9 ).

DISCUSSION
The present study revealed a novel role of statins in regulating BA homeostasis. Although statins are wellknown to decrease cholesterol, which is the substrate for BA  primary BAs, and thus passive absorption of BAs in the intestine.
Because individual BAs have different physiochemical properties and biological effects, such as the capability to activate BA receptors, it is pivotal to analyze the concentrations of individual BAs as well as the proportion of each BA. In general, the majority of BAs that have altered concentrations after atorvastatin treatment are ␤ MCA and its secondary BAs ( Fig. 3 ). For instance, atorvastatin decreases T ␤ MCA, T MCA, and MCA in serum and the liver, T ␤ MCA and T MCA in the SI, and ␤ MCA and MCA in the LI, as well as T ␤ MCA and ␤ MCA in feces. Moreover, the ratio of 12 ␣ -OH versus non12 ␣ -OH BAs is increased by atorvastatin in the liver, SI, LI, and feces. This is consistent with a previous report on the correlation of 12 ␣ -OH BAs and improved glucose and lipid parameters ( 33 ), considering the lipid-lowering effects of statins.
Interestingly, atorvastatin increases the proportion of (T)CA and decreases the proportion of (T)MCAs in both serum and the liver ( Fig. 5A, B ). Correspondingly, atorvastatin increases the hydrophobicity of total BAs in serum and the liver ( Fig. 5C ). However, the BAs whose levels are altered by statins are mainly trihydroxy-BAs, which are considered hydrophilic. In contrast, DCA and LCA are considered hydrophobic BAs and can disrupt cell membranes ( 34 ) and mediate mitochondria-derived cell death ( 35 ). Therefore, the increase in BA hydrophobicity after atorvastatin treatment is not of concern, which is confi rmed by the normal activity of alanine aminotransferase after statin treatment (data not shown).
Atorvastatin has robust effects on regulating the expression of BA-synthetic enzymes in the liver. Interestingly, the mRNA of Cyp7a1, the rate-limiting BA-synthetic enzyme in the classic pathway, is increased over 10-fold by atorvastatin. The mRNA of the initial enzyme in the alternative pathway, Cyp27a1, is also increased by atorvastatin ( Fig. 6 ). These results suggest that the rate of BA de novo synthesis may be enhanced by atorvastatin. The present fi nding is consistent with a number of previous reports that Cyp7a1 was stimulated by pitavastatin in guinea pigs ( 36 ) and crilvastatin in rats ( 37 ); 7 ␣ -hydroxylase activity was increased by pravastatin in gallstone patients ( 38 ); and CYP7A1 promoter polymorphism was associated with poor response to atorvastatin ( 39 ). However, there are also some contradicting data indicating that Cyp7a1 mRNA was decreased by atorvastatin and simvastatin in guinea pigs ( 36 ) and by pravastatin in bile duct-obstructed rats ( 40 ).
The abundant induction of Cyp7a1 transcription by atorvastatin is likely due to decreased FXR signaling in both the liver and intestine. Statins inhibit cholesterol synthesis, thus leading to a decrease in substrates for BA synthesis. This would possibly lead to decreased amounts of BAs in the EHC. Thus, BA-feedback inhibition of Cyp7a1 transcription is removed. Specifi cally, the FXR signaling activated by BAs in both the liver and intestine is lower after atorvastatin treatment, evidenced by decreased mRNA levels of the FXR target gene, SHP, in the liver, as well as the FXR target gene, Fgf15, in the ileum ( Fig. 7 ). The In contrast to the transporter-mediated active reabsorption that is restricted to the ileum, passive diffusion occurs in both the SI and LI ( 31 ). Primary BAs undergo deconjugation catalyzed by bacterial enzymes, which are mainly colonized in the LI lumen. Furthermore, it has been reported that diet-induced alterations of host cholesterol metabolism are likely to affect the gut microbiome ( 32 ). Therefore, it is possible that atorvastatin may alter the intestinal microbiota composition and decrease the bacterial enzyme-catalyzed deconjugation of  and Bcrp ( Fig. 8 ). The upregulation of Mrp2 and Bcrp is consistent with previous reports in the HepG2 cell line with atorvastatin ( 11 ). In addition, the expression of Oct1 is increased by atorvastatin ( Fig. 8A ), indicating a possible increase in the uptake of organic cations, such as its model substrate tetraethylammonium ( 45 ), as well as the antidiabetic drug metformin ( 46 ). Furthermore, statins may increase the effl ux of phospholipids and cholesterol out of the liver. Atorvastatin increases the mRNA of Mdr2 ( Fig. 8B ), which encodes a phospholipid effl ux transporter to facilitate micelle formation and the function of BAs. In addition, atorvastatin increases the expression of Abca1, which is the basolateral effl ux transporter for cholesterol, as well as Abcg5 and Abcg8, which function as dimers to mediate the canalicular effl ux of cholesterol. This fi nding is consistent with previous reports that pitavastatin increased the mRNA of Abca1 in rat hepatoma cell lines ( 31 ) and simvastatin promoted Abca1 expression in mice with atherogenesis ( 47 ). A synergistic hypocholesterolemic effect could be potentially achieved by simultaneously inhibiting cholesterol biosynthesis and promoting Abcg5/Abcg8-mediated cholesterol excretion ( 48 ). Therefore, enhanced cholesterol excretion is likely an additional mechanism for atorvastatin to lower cholesterol. Further studies on protein expression and transporter membrane localization, as well as transporting activity, are necessary to draw a defi nite conclusion.
The fi ndings of the present study are in general agreement with the previous report in mice that atorvastatin increased fecal BA excretion and expression of Cyp7a1, but did not affect BA conjugating enzymes ( 49 ). Several reports showed that statins at therapeutic doses had little effect on hepatic BA synthesis or turnover in humans. Specifi cally, atorvastatin did not alter BA concentrations or composition in patients with primary biliary cirrhosis ( 50 ); lovastatin did not alter the synthesis or turn-over of BAs ( 51 ); fl uvastatin did not alter BA composition in bile ( 52 ); and pravastatin did not alter the cholesterol 7 ␣hydroxylase activity ( 53 ) or fecal BAs ( 54 ), but was shown to decrease 7 ␣ -hydroxy-4-cholesten-3-one level, which refl ects BA synthesis ( 55 ). In addition, pravastatin was shown to decrease BA synthesis ( 56 ) and decrease BA concentrations in the plasma of bile duct-obstructed rats ( 40 ). The discrepancies in the literature about BA regulation by statins may be due to species differences, different disease/toxicity models, or different statins used. Further studies are needed to determine the effect of atorvastatin on BA homeostasis in disease status, and whether atorvastatin could be used as a combined treatment with UDCA to treat hepatobilliary diseases.
In conclusion, the present study demonstrated that the cholesterol-lowering drug atorvastatin maintains total BAs in the liver and serum of mice, and this appears to be due to the compensatory induction of some critical BAsynthetic enzymes (Cyp7a1 and Cyp27a1) and BA uptake transporters (Ntcp and Oatp1b2) in the liver. The marked induction of Cyp7a1 by atorvastatin is likely due to the repression of FXR signaling in both the liver and intestine.
inhibited FXR signaling in the liver and intestine probably contributes to the abundant induction of Cyp7a1 transcription in atorvastatin-treated mice. The elevated expression of Cyp7a1, the rate-limiting enzyme for BA synthesis, possibly compensates the prior BA changes by suppressed cholesterol synthesis, and helps to maintain BA homeostasis in liver and serum.
In addition to increasing the expression of BA-synthetic enzymes, another pathway to maintain BA concentrations in the liver after statin treatment might be through increasing BA uptake transporters in the liver. Indeed, atorvastatin increases the expression of the unconjugated BA uptake transporter Oatp1b2. Moreover, atorvastatin also increases the expression of the conjugated BA uptake transporter Ntcp ( Fig. 8A ). The increase in BA uptake transporters in the liver could be a direct effect by atorvastatin or an indirect effect due to decrease in BA recycling from the intestine. Although atorvastatin does not alter the expression of the BA effl ux transporter Bsep ( Fig. 8B ), it increases the mRNA of Mrp2, which mediates the biliary excretion of tauro-BA-sulfates ( 23 ) as well as glutathione, which provides the osmotic force that drives BA-independent bile fl ow ( 41 ). This suggests that atorvastatin might increase bile fl ow. Several reports have shown that BAs seem to cause more changes in SHP and Fgf15 than the BA transporters, which are also FXR targets (42)(43)(44). Comparing germ-free and conventional mice, Sayin et al. ( 43 ) revealed that gut microbiota primarily affect FXR targets in the ileum and not the liver, and despite the marked increase in Fgf15 and SHP, the expression of other FXR target genes were slightly increased (Ost ␣ ), not altered (Ost ␤ , Ibabp, Ntcp, and Bsep), or even decreased (Mrp2) in the presence of gut microbiota. Consistent with these fi ndings, the present study showed that the expression of SHP and Fgf15 were decreased by atorvastatin, whereas other FXR targets, such as Bsep, Asbt, and Ost ␣ / ␤ , remained relatively constant.
In addition to BA transporters, the present study is the fi rst report on the effect of statins in regulating the expression of many transporters (such as Oatp1a1, Oatp1a4, Oct1, Ent1, Mate1, etc.) in the liver of mice. Atorvastatin increases the hepatic expression of the statin uptake transporter Oatp1b2, as well as statin effl ux transporters Mrp2 Fig. 9. mRNA expression of BA-conjugating enzymes in the liver as well as the proteins responsible for BA reabsorption in ileum of atorvastatin-treated mice. Data are presented as mean ± SEM. VEH, vehicle; ATO, atorvastatin.