Biochimica et Biophysica Acta 1851 (2015) 19–29

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bba l ip

Bile acid signaling in lipid metabolism: Metabolomic and lipidomicanalysis of lipid and bile acid markers linked to anti-obesity andanti-diabetes in mice☆

Yunpeng Qi a,b, Changtao Jiang a, Jie Cheng a,c, Kristopher W. Krausz a, Tiangang Li c,Jessica M. Ferrell c, Frank J. Gonzalez a, John Y.L. Chiang c,⁎a Laboratory of Metabolism, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USAb Department of Pharmaceutical Analysis, School of Pharmacy, Second Military Medical University, Shanghai 200433, Chinac Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH 44272, USA

Abbreviations: CYP7A1, cholesterol 7α-hydroxylase; CoverexpressingCYP7A1 in the liver;HFD, high-fat diet; UPLCliquid chromatography coupled with electrospray ionizationspectrometry; FXR, farnesoidX receptor; LPC, lysophosphatiline; SM, sphingomyelin; CER, ceramide; CA, cholic acidchenodeoxycholate; T-β-MCA, tauro-β-muricholic acid; TCTDCA, taurodeoxycholic acid; PCA, principal componentsquares-discriminant analysis; OPLS-DA, orthogonal prdiscriminant analysis; α-MCA, α-muricholic acid; UDCA, umuricholic acid; DCA, deoxycholic acid☆ This article is part of a Special Issue entitled Linkinglipodomics.⁎ Corresponding author at: Department of IntegrativeM

Medical University, 4209 State Route 44, Rootstown, OHfax: +1 330 325 5910.

E-mail address: jchiang@neomed.edu (J.Y.L. Chiang).

http://dx.doi.org/10.1016/j.bbalip.2014.04.0081388-1981/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 11 March 2014Received in revised form 17 April 2014Accepted 28 April 2014Available online 4 May 2014

Keywords:CYP7A1lipidomicstauro-β-muricholic acidfarnesoid X receptor (FXR)bile acid metabolism

Bile acid synthesis is the major pathway for catabolism of cholesterol. Cholesterol 7α-hydroxylase (CYP7A1) is therate-limiting enzyme in the bile acidbiosynthetic pathway in the liver andplays an important role in regulating lipid,glucose and energymetabolism. Transgenicmice overexpressing CYP7A1 (CYP7A1-tgmice) were resistant to high-fat diet (HFD)-induced obesity, fatty liver, and diabetes. However themechanism of resistance to HFD-induced obe-sity of CYP7A1-tg mice has not been determined. In this study, metabolomic and lipidomic profiles of CYP7A1-tgmice were analyzed to explore the metabolic alterations in CYP7A1-tg mice that govern the protection againstobesity and insulin resistance by using ultra-performance liquid chromatography-coupledwith electrospray ioniza-tion quadrupole time-of-flight mass spectrometry combinedwithmultivariate analyses. Lipidomics analysis identi-fied seven lipid markers including lysophosphatidylcholines, phosphatidylcholines, sphingomyelins and ceramidesthat were significantly decreased in serum of HFD-fed CYP7A1-tgmice. Metabolomics analysis identified 13metab-olites in bile acid synthesis including taurochenodeoxycholic acid, taurodeoxycholic acid, tauroursodeoxycholic acid,taurocholic acid, and tauro-β-muricholic acid (T-β-MCA) that differed between CYP7A1-tg and wild-type mice.Notably, T-β-MCA, an antagonist of the farnesoid X receptor (FXR) was significantly increased in intestine ofCYP7A1-tg mice. This study suggests that reducing 12α-hydroxylated bile acids and increasing intestinalT-β-MCA may reduce high fat diet-induced increase of phospholipids, sphingomyelins and ceramides,and ameliorate diabetes and obesity. This article is part of a Special Issue entitled Linking transcriptionto physiology in lipodomics.

© 2014 Elsevier B.V. All rights reserved.

YP7A1-tg mice, transgenic mice-ESI-QTOFMS, ultra performancequadrupole time-of-flight mass

dylcholine; PC, phosphatidylcho-; TCA, taurocholic acid; CDCA,DCA, taurochenodeoxycholate;

s analysis; PLS-DA, partial leastojections to latent structures-rsodeoxycholic acid; β-MCA, β-

transcription to physiology in

edical Sciences, Northeast Ohio44272. Tel.: +1 330 325 6694;

1. Introduction

1.1. Bile acid synthesis

Bile acid synthesis is the major pathway for catabolism of cholesterolto bile acids. In the liver, cholesterol 7α-hydroxylase (CYP7A1) is thefirstand rate-limiting enzyme of the bile acid biosynthetic pathway produc-ing two primary bile acids, cholic acid (CA, 3α, 7α, 12α-OH) andchenodeoxycholic acid (CDCA, 3α, 7α-OH) in humans (Fig. 1) [1].Sterol-12α hydroxylase (CYP8B1) catalyzes the synthesis of CA. Inmice, CDCA is converted to α-muricholic acid (α-MCA: 3α, 6β, 7α-OH)andβ-muricholic acid (β-MCA: 3α, 6β, 7β-OH). Bile acids are conjugatedto taurine or glycine, secreted into the bile and stored in the gallbladder.After a meal, bile acids are released into the gastrointestinal tract. In theintestine, conjugated bile acids are first de-conjugated and then 7α-dehydroxylase activity in the gut flora converts CA to deoxycholic acid(DCA: 3α, 12α), and CDCA to lithocholic acid (LCA: 3α), two majorsecondary bile acids in humans. In mice, CDCA is converted to

Fig. 1. Bile acid synthesis. In the classic bile acid synthesis pathway, cholesterol is convertedto cholic acid (CA, 3α, 7α, 12α) and chenodeoxycholic acid (CDCA, 3α, 7α). CYP7A1 is therate-limiting enzyme and CYP8B1 catalyzes the synthesis of CA. In mouse liver, CDCA isconverted to α-muricholic acid (α-MCA, 3α, 6β, 7α) and β-MCA (3α, 6β, 7β). Most bileacids inmice are taurine (T)-conjugated and secreted into bile. In the intestine, gut bacteriade-conjugate bile acids and then remove the 7α-hydroxyl group from CA and CDCA to formsecondary bile acids deoxycholic acid (DCA, 3α, 12α) and lithocholic acid (LCA, 3α),respectively. T-α-MCA and T-β-MCA are converted to T-hyodeoxycholic acid (THDCA,3α, 6α), T-ursodeoxycholic acid (TUDCA, 3α, 7β), T-hyocholic acid (THCA, 3α, 6α, 7α)and T-murideoxycholic acid (TMDCA, 3α, 6β). These secondary bile acids are reabsorbedand circulated to liver to contribute to the bile acid pool. Secondary bile acids ω-MCA(3α, 6α, 7β) and LCA are excreted into feces.

Fig. 2. Bile acid signaling pathways. Bile acids activate FXR, TGR5 and cell signaling path-ways to inhibit CYP7A1 and CYP8B1 gene transcription. 1) Hepatic FXR/SHP pathway: bileacid activated-FXR induces SHP, which inhibits HNF4α and LRH-1 trans-activation ofCYP7A1 and CYP8B1 gene transcription in hepatocytes. Bile acid response element bindsHNF4α and LRH-1. 2) Intestinal FXR/FGF19/FGFR4 pathway: in the intestine, FXR inducesFGF15 (mouse)/FGF19 (human), which is secreted into portal circulation to activate FGFreceptor 4 (FGFR4) in hepatocytes. FGFR4 signaling stimulates JNK1/2 and ERK1/2 path-ways of MAPK signaling to inhibit CYP7A1 gene transcription by phosphorylation and inhi-bition of HNF4α binding activity. 3) FXR-independent signaling pathways: Conjugated bileacids activate PKCs,whichactivate theMAPKpathways to inhibit CYP7A1. Bile acids also ac-tivate insulin receptor (IR) signaling IRS/PI3K/PDK1/AKT, possibly via activation of epider-mal growth factor receptor (EGFR) signaling, MAPKs (MEK, MEKK), to inhibit CYP7A1gene transcription. The secondary bile acid TLCA activates TGR5 signaling in Kupffer cells.TGR5 signaling may regulate CYP7A1 by an unknown mechanism. TCA activatessphingosine-1-phosphate (S1P) receptor 2 (S1PR2), which may activate AKT and ERK1/2to inhibit CYP7A1. S1P kinase 1 (Sphk1) phosphates sphingosine (Sph) to S-1-P, whichactivates S1PR2. On the other hand, nuclear SphK2 interacts with and inhibits histonedeacetylase (HDAC1/2) and may induce CYP7A1. The role of S1P, SphK2, and S1PR2signaling in regulation of bile acid synthesis is not known.

20 Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

ursodeoxycholic acid (UDCA: 3α, 7β), α-MCA to hyocholic acid (HCA:3α, 6α, 7α) and murideoxycholic acid (MDCA: 3α, 6β), and β-MCA toω-MCA (3α, 6α, 7β) and hyodeoxycholic acid (HDCA: 3α, 6α). The num-ber of hydroxyl groups, their location andα- or β-orientation, and conju-gation determine their hydrophobicity and detergent properties.In humans, most bile acids are glycine or taurine-conjugated and CA,CDCA and DCA are the most abundant bile acids. In mice, most bileacids are taurine-conjugated and CA and α- and β-MCAs are the mostabundant bile acids. Bile acids facilitate absorption of dietary fats,steroids, and lipid soluble vitamins into enterocytes and are transportedvia portal circulation to the liver for metabolism and distribution toother tissues and organs. About 95% of bile acids are reabsorbed in theileum and transported to the liver to inhibit CYP7A1 and bile acid syn-thesis. Enterohepatic circulation of bile acids provides a negative feed-back mechanism to maintain bile acid homeostasis. Alteration of bileacid synthesis, secretion and transport causes cholestatic liver diseases,gallstone diseases, fatty liver disease, diabetes and obesity [1]. Deficiencyof CYP7A1 in humans is associated with hypercholesterolemia andpremature atherosclerosis [2].

1.2. Bile acid signaling

Bile acids are signaling molecules that activate several intracellularsignaling pathways [1,3]. Bile acids are known to activate the farnesoidX receptor (FXR) [4–6] and a membrane G protein coupled-receptorTGR5 (Gpbar1) [7,8]. CA and CDCA are potent endogenous ligands ofFXR, whereas LCA and DCA and their taurine conjugates are more potentagonists of TGR5. FXR plays a critical role in regulation of bile acid synthe-sis and secretion, and lipid and glucosemetabolism in the liver. TGR5 is aGαs protein coupled receptor (GPCR) that activates cAMP signaling inmany cells and plays a role in energy metabolism in brown adiposetissue, relaxing and refilling gallbladder, secreting glucagon-like peptide

1 (GLP-1) in intestinal endocrine cells and controlling GI motility [9,10].TGR5 is expressed in the epithelial cells of the gastrointestinal system,including intestine, spleen, cholangiocytes, gallbladder, hepatic sinusoidalendothelial cells andhepaticmacrophages, andKupffer cells [7,8,11,12]. Arecent report shows that conjugated bile acids are potential agonists ofsphingosine-1-phosphate receptor 2 (S1PR2) in hepatocytes [13]. S1PR2is a Gαi class of GPCR activated by sphingosine-1-phosphate (S1P).

1.3. Mechanisms of bile acid feedback regulation of bile acid synthesis

Two FXR-dependent mechanisms are known to inhibit bile acid syn-thesis. In the liver bile acid-activated FXR induces a negative receptorsmall heterodimer partner (SHP) to inhibit trans-activation activity of he-patic nuclear factor 4α (HNF4α) and liver receptor homologue-1 (LRH-1)that bind to the bile acid response element in the CYP7A1 and CYP8B1gene promoters (Fig. 2, Pathway 1). In the intestine, bile acids activateFXR to induce fibroblast growth factor (mouse FGF15, or humanFGF19), which activates hepatic FGF receptor 4 (FGFR4) and cJun N-terminal kinase 1/2 (JNK1/2) and extracellular-regulated kinase 1/2(ERK1/2) signaling of mitogen-activated protein kinase (MAPK) path-ways to inhibit trans-activation of CYP7A1/CYP8B1 gene by HNF4α [14](Pathway 2). Several FXR-independent cell-signaling pathways havebeen reported and are shown as Pathway 3 (Fig. 2). Conjugated bileacids are known to activate several protein kinase Cs (PKC) and growthfactor receptors, epidermal growth factor receptor (EGFR), and insulin re-ceptor (IR) signaling to inhibit CYP7A1/CYP8B1 and bile acid synthesis viaactivating the ERK1/2, p38 and JNK1/2 pathways [15–18]. Bile acids andpro-inflammatory cytokines TNFα and IL-1β produced in hepatocytes

21Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

or secreted from Kupffer cells activated ERK1/2 and JNK1/2 to phosphor-ylate and inactivated HNF4α, and resulted in inhibiting CYP7A1 andCYP8B1 gene transcription [19,20]. Conjugated bile acids activate S1PR2in hepatocytes [13]. Extracellular signaling activates sphingosine kinase1 (SphK1) to phosphorylate membrane sphingosine to S1P, which acti-vates S1PR2 and downstream ERK1/2 and AKT signaling and may inhibitCYP7A1/CYP8B1 gene transcription. Interestingly, nuclear SphK2 is asso-ciated with histone 3 and S1P inhibits histone deacetylase 1/2 (HDAC1/2) [21]. It is possible that bile acids may activate SphK2 to stimulateCYP7A1 gene transcription by an epigenetic mechanism.

1.4. Bile acids regulation of lipid and glucose metabolism

Recent studies have unveiled that bile acids play a critical role inmaintaining lipid, glucose and energy homeostasis through activationof FXR and TGR5 [1,9]. It has been known for a long time that treatinghumangallstone patientswith CDCAdecreases hepatic VLDLproductionand plasma triglyceride levels [22], while treating hypercholesterolemicpatientswith bile acid binding resins increases plasma triglyceride levels[23,24]. Hepatic lipids and circulating cholesterol and triglycerides areaccumulated in FXR-/- mice, whereas activation of FXR by bile acids orFXR agonists decreases plasma cholesterol and triglycerides in wildtype mice [25]. It was suggested that the FXR/SHP pathway inhibits ste-roid response element binding protein 1c (SREBP-1c)-mediated hepaticlipogenesis [26]. Activation of FXR by bile acid feeding or administrationof the FXR agonist GW4064 lowered fasting plasma glucose and im-proved insulin sensitivity in obese and diabetic db/db mice [25,27],while FXR-deficient mice had insulin resistance and hyperglycemia[27]. Plasma FGF19 increases during the postprandial period in humans,presumably due to increased bile acid signaling [28]. FGF19 transgenicmice were resistant to diet-induced obesity and insulin resistance [29,30]. In addition, FGF19 has been shown to repress hepatic glucose pro-duction [31], promote glycogen synthesis [32], repress lipogenesis [33,34] and increase metabolic rate [29,30].

1.5. Bile acids protect against high fat diet induced obesity and diabetes

Transgenic mice overexpressing CYP7A1 (CYP7A1-tg) have beenused to demonstrate that increasing conversion of cholesterol to bileacids can protect against lithogenic diet-induced atherosclerosis inmice [35]. We used CYP7A1-tg mice as a model to demonstrate thatincreasing bile acid synthesis and pool size may protect against high-fat diet (HFD)-induced obesity, fatty liver, and insulin resistance [36].These mice had lower body fat mass and higher lean mass whenfed HFD than did strain-matched wild-type (WT) mice. In CYP7A1-tgmice, bile acid pool size increased 2.5-fold and CYP8B1 expression wasabolished. The gallbladder bile acid composition was significantlyaltered. The major bile acids in CYP7A1-tg mice are CDCA (55%), α-and β-MCA (27%) and UDCA (15%). Microarray gene profiling analysisshowed markedly increased expression of key genes in de novo choles-terol synthesis in CYP7A1-tg mice [37]. Hepatic cholesterol synthesiswas increased 10-fold but fatty acid synthesis was reduced 60%. Biliarycholesterol, bile acid and phospholipid secretion and fecal cholesteroland bile acid excretion were increased, so that hepatic cholesterolhomeostasis ismaintained [38]. However, the specificmetabolic profilescharacterizing the resistance to HFD-induced obesity of CYP7A1-tgmicehave not been determined. This novel mouse model provides a uniquein vivo system to study the underlying molecular mechanism of theanti-diabetic and anti-obesity functions of bile acids.

1.6. Metabolomics and lipidomics

To investigate the metabolic changes in CYP7A1-tg mice that governprotection against obesity, metabolomics and lipidomics profiling wereemployed.Metabolomics aims to reveal variousmetabolic characteristicsof external or internal perturbations to biological systems by profiling

low-molecular-weight metabolites in bio-samples [39–41]. Lipidomicsis a sub-metabolomics platform that provides a comprehensive analysisof lipid species within a cell or tissue, which plays an essential role indefining the biochemical mechanisms of lipid-related disease processesthrough identifying alterations in cellular lipid metabolism, traffickingand homeostasis [42,43]. The current study of metabolomics andlipidomics profiling of CYP7A1-tg mice using ultra-performance liquidchromatography-coupled with electrospray ionization quadrupole time-of-flightmass spectrometry (UPLC-ESI-QTOFMS) detected and character-ized small organic molecules in biological materials [44–47]. This studyexplores the involvement of metabolites such as phospholipids and bileacids in the regulation of metabolic homeostasis, and provides insightsinto the mechanism for the anti-obesity and anti-diabetic effect ofincreased bile acid pool and altered bile acid composition.

2. Materials and methods

2.1. Chemicals and reagents

The lipid standards1-palmitoyl-sn-glycero-3-phosphocholine(LPC 16:0), 1-stearoyl-sn-glycero-3-phosphocholine (LPC 18:0),1,2-diarachidonoyl-sn-glycero-3-phosphocholine (PC 20:4), N-palmitoyl-D-erythro-sphingosylphosphorylcholine (SM 18:1/16:0), andN-stearoyl-D-erythro-sphingosine (C18 ceramide), etc., were obtainedfrom Avanti Polar Lipids, Inc. (Alabaster, AL). Bile acid standards cholicacid (CA), taurocholic acid (TCA), chenodeoxycholic acid (CDCA), tauro-β-muricholic acid (T-β-MCA), taurochenodeoxycholic acid (TCDCA),taurodeoxycholic acid (TDCA) and the internal standard chlorpropamidewere purchased from Sigma-Aldrich (St. Louis, MO). All solvents andorganic reagents were of the highest obtainable grade.

2.2. Animals

CYP7A1-tg mice overexpressing rat CYP7A1 under the control of theapolipoprotein E3 (Apoe3) promoter (B6.Cg-Tg (APOE-CYP7A1) 1Rjd/Mmcd) were obtained as described previously [48]. The mice were fur-ther bredwithwild-typeC57BL/6Jmice (The Jackson Laboratory, BarHar-bor, ME). Transgenic mice and wild-type littermates were backcrossedbetween 8 and 9 generations to yield ~99% C57BL/6J background. Micewere maintained on a standard chow diet and water ad libitumand housed in a room with a 12-hour light (6 am to 6 pm)/12-hourdark (6 pm to 6 am) cycle. To induce obesity and insulin resistance,16–20-week-old male CYP7A1-tg and wild type mice were fed anHFD (Harlan-Teklad, #88137, 42% fat calories and 0.2% cholesterol) orchowdiet for 4 months. For other experiments, 16–20-week-old femaleCYP7A1-tg and female wild type mice were used. Food intake wasdetermined for a two-day period. All animal manipulations and treat-ments were performed under protocols approved by the InstitutionalAnimal Care and Use Committees at Northeast OhioMedical University.Serum and harvested tissues were frozen at−80 °C for further analysisat the end of the study.

2.3. RNA analysis

RNA was extracted using TRIzol reagent (Invitrogen). Quantitativereal-time PCR (qPCR) was performed using cDNA generated from 1 μgtotal RNA with the SuperScript II Reverse Transcriptase kit (Invitrogen).Primers were designed for qPCR using Primer Express software (AppliedBiosystems, Foster City, CA), and sequences are available upon request.qPCRs were carried out using SYBR green PCR master mix (AppliedBiosystems) in anABI Prism7900HTSequenceDetection System(AppliedBiosystems). Values were quantified using the comparative thresholdcycle method, and samples were normalized to β-actin.

22 Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

2.4. Metabolomics analysis

2.4.1. Sample preparationFor serum lipidomics analysis 25 μl serum was extracted by 4-fold

cold chloroform: methanol (2:1) solution containing 2 μM LPC (17:0),PC (17:0), SM (17:0) and CER (17:0) (Avanti Polar Lipids, Alabaster,AL) as internal standards. The samples were vortexed for 30 s andthen allowed to stand for 5 min at room temperature. The mixturewas centrifuged at 13,000 rpm for 5 min and then the lower organicphase was collected and evaporated at room temperature under vacu-um and the residue was dissolved in chloroform: methanol (1:1),followed by diluting with isopropanol:acetonitrile:H2O (2:1:1) prior toUPLC-MS analysis. For tissue lipidomics analysis, about 50mgaccuratelyweighted tissues were homogenized with 700 μL methanol:H2O (4:3)solution and then extracted using 800 μL chloroform containing 2 μMLPC (17:0), PC (17:0), SM (17:0) and CER (17:0) as internal standards.The homogenate was incubated at 37 °C for 20min followed by centrifu-gation for 20min at 13,000 rpm. The lower organic phasewas transferredto a new tube and dried under vacuum. The residuewas suspendedwith100 μL chloroform: methanol (1:1) solution and then diluted withisopropanol:acetonitrile:H2O (2:1:1) solution before injection.

2.4.2. Metabolomics analysisFormetabolomics analysis, 50mg tissue sampleswere homogenized

in 500mL 50% aqueous acetonitrile containing 5 μM of chlorpropamide(internal standard). The samples were vortexed and centrifuged at13,000 rpm for 20 min at 4 °C to remove particulates and precipitateprotein. The supernatant was transferred to an autosampler vial foranalysis.

For metabolomics discovery, a 5 μl aliquot of supernatant sampleswas injected into the UPLC-ESI-QTOFMS system (Waters, Milford, MA)with a Waters Acquity BEH 1.7 μm C18 (2.1 × 50 mm) column. Thegradient mobile phase comprises 0.1% formic acid in water (A) and0.1% formic acid in acetonitrile (B). The gradient was maintained at ini-tial 95% A for 0.5min, to 40% A at 4min, and then to 1% A at 8min. Flushfor 1min, then equilibrate at initial conditions for 1.5min. Flow ratewas0.5 ml/min. Column temperature was maintained at 60 °C. WatersSynapt HDMS Q-TOF was operated in both positive and negativemodes, scanning 50–850 amu, at a rate of 0.3 scans/sec. The followinginstrument conditions were used: capillary 3 kV, source temp 120 °C,sampling cove 30 V, desolvation gas flow 850 L/h at 400 °C.

2.4.3. Lipidomics analysisFor lipidomics discovery, samples were analyzed by UPLC-ESI-QTOF

MS using a Water Acquity CSH 1.7 μm C18 column (2.1 × 100 mm)under the following conditions: UPLC: A-acetonitrile/water (60/40), B-isopropanol/acetonitrile (90/10), both A and B contain 10mMAmmoni-umacetate and 0.1% formic acid. Gradient: initial 60%A to 57%A at 2min,to 50% A at 2.1 min*, to 46%A at 12min, to 30% A at 12.1 min*, to 1% A at18 min before returning to initial conditions at 18.5 min with equilibra-tion for 2 additional minutes (an * indicates ballistic gradient). Flow ratewas 0.4 ml/min. Column temperature was maintained at 55 °C. MS wasrun under the same conditions as above, except run time is 18 min. Atypical chromatogram using the lipidomics method is shown in Supple-mentary Fig. S1.

2.4.4. Biomarker identification and quantitationBiomarkers were screened by analyzing ions in the loading scatter

plot, andmetabolomics databases (METLIN andMadisonMetabolomicsConsortium Database) were searched to find potential candidates. Toconfirm the identities of the putative markers, the authentic standardswere compared with the metabolites based on MS/MS fragmentationpattern and retention time.

Concentrations of the metabolites were determined by multiplereaction-monitoring mass spectrometry based on standard curvesusing authentic standards.

Quantification of lipid markers was performed by MRM and/orparent ion scanning using a Waters UPLC Acquity system coupled to aWaters Xevo TQ mass spectrometer. A Waters Acquity BEH C18 column(2.1 × 100 mm) was used. UPLC: A-water, B-acetonitrile/IPA (5/2),both A and B contain 10 mM ammonium acetate and 0.1% formic acid.Gradient: initial 70% A for 1 min, then linear gradient to 50% A at 3 min,to 1% A at 8 min, held until 15 min, then returning to initial conditionsover 1 min, and held for an additional 2 min for column equilibration.Flow rate was 0.4 ml/min. Column temperature was maintained at50 °C.Waters XevoTQwas operated inMRMmode. The following instru-ment conditions were used: Capillary 2.2 kV, source temp 150 °C,desolvation gas flow 850 L/h at 450 °C. Total run time is 18 min. Thecone voltage and collision energy for the MRM transitions was optimallydetermined for each transition by the instrument's IntelliStart software.

Quantification of bile acidswasperformedonaWatersAcquityH-classUPLC system using a Waters Acquity BEH C18 column (2.1 × 100 mm)coupled to a Waters Xevo G2 Q-TOF mass spectrometer. UPLC: A-0.1%formic acid in water, B-0.1% formic acid in acetonitrile. Gradient: initial80% A for 4 min, to 60% A at 15 min, to 40% A at 20 min, to 10% A at21 min. Flush for 1 min, then equilibrate at initial conditions for4 min. Flow rate is 0.4 ml/min. Column temperature was maintainedat 45 °C. Waters Xevo G2 Q-TOF was operated in negative mode, scan-ning 50–850 amu, at a rate of 0.3 scans/sec. The following instrumentconditions were used: Capillary 1.5 kV, source temp 150 °C, samplingcove 30 V, desolvation gas flow 850 L/h at 500 °C.

2.4.5. Data processing and multivariate data analysisChromatographic and spectral data were de-convoluted by

MarkerLynx software (Waters). A multivariate data matrix containinginformation on sample identity, ion identity (retention time and m/z),and ion abundance was generated through centroiding, deisotoping, fil-tering, peak recognition, and integration. The intensity of each ion wascalculated by normalizing the single ion counts vs. the total ion countsin the whole chromatogram. The data matrix was further exported intoSIMCA-P software (Umetrics, Kinnelon, NJ) and transformed by meancentering and pareto scaling, the technique that increases the impor-tance of low abundance ions without significant amplification of noise.Statistical models including principal components analysis (PCA), partialleast squares-discriminant analysis (PLS-DA), orthogonal projections tolatent structures-discriminant analysis (OPLS-DA) were established torepresent the major latent variables in the data matrix.

2.4.6. Data analysisExperimental values were expressed as mean± S.D. Statistical analy-

sis was performed with two-tailed Student’s t-test or Mann-Whitney testusing Prism 6.0 (GraphPad Software, Inc., San Diego, CA). A p-value b 0.05was considered statistically significant.

3. Results

3.1. CYP7A1-tg and wild-type mice challenged with HFD had distinctlipidomic profiles

When challenged with an HFD, CYP7A1-tg mice had lower body fatmass and higher lean mass compared to wild-type mice [38]. As a plat-form for comprehensive and quantitative description of the set of lipidspecies, lipidomics was used to investigate the mechanism of this phe-notype. By use of an unsupervised PCA model with the cumulative R2X0.677 for serum and 0.593 for liver, CYP7A1-tg and wild-type micewere clearly separated based on the scores plot (SupplementaryFig. S2), indicating that these two groups have distinct lipidomic pro-files. Supervised PLS-DA models were then established to maximizethe difference of metabolic profiles between CYP7A1-tg and wild-typegroups as well as to facilitate the screening of lipid marker metabolites(Fig. 3). For serum samples, two latent variables (LV) were calculated,and the cumulative R2Y and Q2 were 0.993 and 0.878, respectively. For

Table 1List of identified lipidomics markers.

No. Metabolites name MF Retention time, M + H

1 LPC 16:0 C24H50NO7P 1.4873, 496.34032 LPC 18:0 C26H54NO7P 2.1969, 524.37113 LPC 18:1 C26H52NO7P 1.5914, 522.35544 LPC 18:2 C26H50NO7P 1.2159, 520.34045 PC 16:0-20:4 C44H80NO8P 9.7434, 782.56956 PC 16:0-22:6 C46H80NO8P 9.0373, 806.56997 SM 16:0 C39H79N2O6P 8.584, 703.5749

Themost significant seven ions based on PLS-DAmodels were identified by tandemmassspectrometry and retention time comparisons with authentic standards. Abbreviations:MF, molecular formula.

23Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

liver samples, two LV were calculated with the cumulative R2Y and Q2

0.937 and 0.495, respectively. No over-fitting of the data was observedaccording to results of the permutation test. Based on the PLS-DAmodel, the most significant ions that led to the separation betweenCYP7A1-tg and WT groups in serum (Fig. 3A) and liver (Fig. 3B) wereobtained from the loading plots shown in Fig. 3C (serum) and Fig. 3D(liver), in which the ions furthest away from the origin contributesignificantly to the clustering of the two groups. The statistically signif-icant thresholds of variable influence on projection (VIP) values largerthan 1.5 and p values less than 0.05 were used [49]. Finally, seven ionswere identified (Table 1) by tandem mass spectrometry and retentiontime comparisons with authentic standards (Suppl. Fig. S3). Theselipid markers (Table 1) include LPC 16:0, LPC 18:0, LPC 18:1, LPC 18:2,PC 16:0-20:4, PC 16:0-22:6, and SM 16:0.

3.2. Lipidomics marker levels decreased in HFD-fed CYP 7A1-tg mice

A number of endogenous lysophosphatidylcholine (LPC), phosphati-dylcholine (PC), sphingomyelin (SM) and ceramide (CER) were deter-mined and quantified. In control diet-treated mice, except PC 16:0/22:6 and SM 16:0, no clear differences were observed between thetwo groups (Fig. 4A and Fig. 4B). In the HFD-fed groups, all the lipidmarkers listed in Table 1 were significantly decreased in CYP7A1-tgmice compared to wild-type mice. Notably, all the detected SMs, andsome CER (C16, C24: 1) were significantly decreased in serum of theCYP7A1-tg group compared to the wild-type group (Fig. 4C and D).CER is produced by hydrolysis of SM and synthesized from serine andpalmitoyl-CoA [50]. CERs are typically found at trace levels in tissues

Fig. 3. PLS-DA analysis of CYP7A1-tg andwild-type (WT)mice challengedwith HFD. Based oshown for serum (A) and liver samples (B). Based on the loading plots (C for serum and D for ligroups were obtained and identified as follows: 1. LPC16:0; 2. LPC18:0; 3. LPC18:1; 4. LPC 18:2

[51,52]. Hence, CER may fail to be screened out as metabolite markersusing statistical methods. Since both SM and CER were implicated in thedevelopment of atherosclerosis, insulin resistance, diabetes, and obesity[53], reduced SM and CER may reduce inflammation and contribute tothe phenotypes of resistant to diabetes and obesity of CYP7A1-tg mice[37]. Interestingly, in ileum of control diet-fed CYP7A1-tg mice, thesedetected lipidmarkers were decreased (Supplementary Fig. S4A), where-as this is not the case for liver (Suppl. Fig. S4B).

3.3. Metabolomics profiling of CYP7A1-tg and wild-type mice highlightedmarkers in bile acid pathway

PCA of the UPLC-ESI-QTOFMS-negative mode data from mousegallbladder, liver, ileum and colon samples revealed distinct metabolic

n the score plots, distinct lipidomic profiles of male CYP7A1-tg andwild-type groups werever) themost significant ions that led to the separation between CYP7A1-tg and wild-type; 5. PC16:0-20:4; 6. PC16:0-22:6; 7. SM16:0.

Fig. 4. Serum lipid marker quantitation results by multiple reaction-monitoringmass spectrometry based on standard curves using authentic standards. Data were expressed asmean± SD. Significant comparisonwas based on two-tailed Student's t-test or Mann-Whitney test. An * indicates p b 0.05 (with respect to the wild-type group). Abbreviations:WT-CTR,control diet-treated male wild-type mice; Tg-CTR, control diet-treated male CYP7A1 transgenic mice; WT-HFD, high-fat diet-treated male wild-type mice; Tg-HFD, high-fat diet-treatedmale CYP7A1 transgenic mice.

Fig. 5. OPLS-DAhighlighted thirteenmarkers in bile acid pathway that contribute significantly to the clustering of CYP7A1-tg andwild-type (WT)mice. Ileum bile acids are shown.(A) In the score plot, female CYP7A1-tg andWTmice werewell separated; (B) using a statistically significant thresholds of variable confidence approximately 0.75 in the S-plot, a numberof ions were screened out as potential markers, which were later identified as 13 bile acid metabolites, including α-MCA, TCA, CDCA, and TCDCA etc.

24 Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

25Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

profiles of the CYP7A1-tg and wild-type mice (SupplementaryFig. S5). Subsequently, OPLS-DA was performed to screen for poten-tial markers that contribute to clustering. The S-plot of OPLS-DA datarevealed covariance and correlation between the variables and themodel, and decreases the risk of false positives in the selection of po-tential biomarkers [54]. Using a statistically significant threshold ofvariable confidence approximately 0.75 in the S-plot, a number of ionswere selected as potential markers. For example, ions m/z 514.2838 at11.02min andm/z 124.0073 at 0.30min had significantly reduced levels,whereas m/z 498.2883 at 8.80 min andm/z 498.2907 at 13.83 min weremarkedly enriched in the CYP7A1-tg mice. Ion identification was thenperformed as described above and identified 13 metabolites (Fig. 5 andTable 2), including taurine, CDCA, UDCA, α-MCA, β-MCA, CA, tauro-murideoxycholic acid (TMDCA), tauro-ursodeoxycholic acid (TUDCA),tauro-hyodeoxycholic acid (THDCA), TCDCA, TDCA, T-β-MCA, and TCA.

3.4.Metabolomicsmarker quantitation revealed altered bile-acid pattern inCYP7A1-tg mice

To examine the bile-acid composition and their levels in the tested tis-sues, 13 bile acid markers in gallbladder, liver, ileum and colon sampleswere quantified. A representative heat map of ileum bile acid markersof wild type and CYP7A1-tg mice is shown in Supplementary Fig. S6.Fig. 6 shows quantities of identified bile acid metabolites in ileum(Fig. 6A), colon (Fig. 6B), liver (Fig. 6C) and gallbladder (Fig. 6D) in wildtype and CYP7A1-tg mice. Consistent with and apart from the previousfindings in gallbladders of CYP7A1-tg mice [38], these data show that inmost of the tested tissues, 12α-hydroxylated bile acids (CA, TCA, DCAand TDCA) were decreased due to inhibition of CYP8B1 by increasedbile acid pool in CYP7A1-tg mice [55]. On the other hand, TCDCA andTMDCA were increased in all tissues. Taurine was markedly decreasedin all tissues due to increased taurine-conjugation of the enlargedbile acid pool. These data indicate that the enlarged bile acid poolin CYP7A1-tg mice contains mostly taurine-conjugated CDCA andmuricholic acids and their derivatives. The most notable finding inthe bile acid metabolomics is the significant increase of T-β-MCA, apotent FXR antagonist [56] in ileum and colon, but not in liver orgallbladder of CYP7A1-tg mice compared to wild-type mice (Fig. 6E).

3.5. Intestinal mRNA expression revealed inhibition of FXR signaling andinduced fatty acid β-oxidation in CYP7A1-tg mice

Analysis of intestinal mRNA expression revealed that in chow-fedCYP7A1-tg mice, apical sodium dependent bile salt transporter (ASBT),an intestinal bile acid reabsorption transporterwas significantly induced,while sinusoidal bile acid efflux transporters (OSTα) and OSTβ were

Table 2Metabolomics analysis highlighted 13 markers in bile acid pathway.

No. Metabolites name MF Retention time, M-H

1 Taurine C2H7NO3S 0.30, 124.00732 CDCA C24H40O4 19.07, 391.28583 UDCA C24H40O4 16.16, 391.28484 α-MCA C24H40O5 13.23, 407.28095 β-MCA C24H40O5 13.64, 407.27946 CA C24H40O5 15.96, 407.28157 TMDCA C26H45NO6S 8.80, 498.28838 TUDCA C26H45NO6S 10.57, 498.29049 THDCA C26H45NO6S 10.77, 498.291110 TCDCA C26H45NO6S 13.83, 498.290711 TDCA C26H45NO6S 14.46, 498.291412 T-β-MCA C26H45NO7S 7.73, 514.284813 TCA C26H45NO7S 11.02, 514.2838

The most significant thirteen ions based on OPLS-DA models were identified by tandemmass spectrometry and retention time comparisons with authentic standards. Abbrevia-tions: MF, molecular formula.

somewhat decreased (Fig. 7A). Increasing ASBT indicates reducedfecal bile acid excretion. OSTα and OSTβ are induced by FXR. Intesti-nal mRNA levels of the genes involved in mitochondrial fatty acid β-oxidation, such as peroxisome proliferator-activated receptor α(PPPARα), long-chain specific acyl-CoA dehydrogenase (ACADL),enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase (EHHADH),and acetyl-CoA acyltransferase 1 (ACAA1), and a fatty acid uptake trans-porter CD36were up-regulated (Fig. 7A), indicating increased fatty acidβ-oxidation.

3.6. Hepatic mRNA expression in CYP7A1-tg mice

CYP8B1wasmarkedly suppressed in liver of CYP7A1-tgmice (Fig. 7B),consistent with the perturbed bile acid profiles as revealed in metabolo-mics (Fig. 6C). Still, hepatic expression ofmRNAs encoding the cholesteroltransporters ATP-binding cassette G5 (ABCG5) and ABCG8 was signifi-cantly induced in CYP7A1-tg mice, in accordance with previous results[38]. Although the intestinal expression levels of fatty acid β-oxidationgenes were significantly increased in CYP7A1-tg mice, their hepaticexpression was not changed (Fig. 7B).

Newly synthesized bile acids are conjugated with taurine in a two-step process, an initial step catalyzed by bile acid: CoA synthase (BACS)and a second step requiring bile acid transferase (BAT) [57]; both areexpressed from FXR target genes. Expression of BACS and BAT andtheir regulator, hepatocyte nuclear factor 4α (HNF4α)wereup regulatedin CYP7A1-tg mice. Also, significant increase of the taurine transporter(TAUT) mRNA and decrease of cysteine sulfonic acid decarboxylase(CSD) mRNA encoding the rate-limiting enzyme for synthesis of taurinewere observed (Fig. 7C). These results suggest that the reduced taurine isdue to decreased taurine synthesis and increased taurine transport aswell as increased conjugation of bile acids.

4. Discussion

Our recent study of CYP7A1-tg mice revealed that increased CYP7A1expression and enlarged bile acid pool resulted in significant improve-ment of lipid homeostasis and resistance to high-fed diet-inducedhepatic steatosis, insulin resistance, and obesity in CYP7A1-tg mice[36]. In this study, metabolomics and lipidomics were employed tocharacterize the metabolic profiles of CYP7A1-tg mice and to providenew insights into the critical role of bile acids in regulation of lipid me-tabolism andmetabolic diseases. Lipidomics analysis of serum lipid pro-files of high fat diet-fed CYP7A1-tg identified 7 lipidomic markers thatwere reduced in CYP7A1-tg mice compared to wild type mice. Metabo-lomics analysis identified 13 bile acid metabolites that were altered inCYP7A1-tg mice. In CYP7A1-tg mice, TCA and TDCA were reduced,whereas T-β-MCA was increased in the intestine compared to that ofwild type mice. The decrease of serum LPC, PC, SM and CER, and 12α-hydroxylated bile acids, and increase of T-β-MCA may contribute tothe resistance to diet-induced obesity and diabetes in CYP7A1-tg mice(Fig. 8).

The presentmetabolomics and lipidomics analysis revealed that evenupon challenging with HFD, CYP7A1-tg mice had reduced lipid levels in-cluding LPC, PC, SM and CER. Metabolomics studies of human steatoticliver tissues and HFD-fed mice showed that serum and liver LPC and PCand other lipids levels were increased compared with non-steatoticlivers, suggesting altered lipid metabolism contributes to non-alcoholicfatty liver disease (NAFLD) [58,59]. In HFD-fed CYP7A1-tg mice, reducedserumPC, LPC, SM and CER levels suggest a role for bile acids inmaintain-ing phospholipid homeostasis to prevent NAFLD. SMs are importantmembrane phospholipids that interact with cholesterol in membranerafts and regulate cholesterol distribution and homeostasis. A role forSMandCER in the pathogenesis of insulin resistance, diabetes andobesity[60] and development of atherosclerosis [53] has been reported. CER hasawide range of biological functions in cellular signaling such as activatingprotein kinase C and c-Jun N-terminal kinase (JNK), induction of β-cell

Fig. 6. Quantitation analysis of the bile acid markers. Bile acid marker levels in the ileum (A), colon (B), liver (C), and gallbladder (D) of female CYP7A1-tg and wild-type (WT) mice.(E) T-β-MCA levels in colon, ileum, liver and gallbladder samples (nmol/mg tissue for colon, ileum and liver, andmmol/ml for gallbladder). Datawere expressed asmean± SD. Significantcomparison was based on two tailed Student's t-test or Mann-Whitney test. An * indicates p b 0.05, and a ** indicates p b 0.01 (with respect to the WT group).

26 Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

apoptosis and insulin resistance [61,62]. CER increases reactive oxi-dizing species and activates the NF-κB pathway, which induces pro-inflammatory cytokines, diabetes and insulin resistance [63]. CER issynthesized from serine and palmitoyl-CoA or hydrolysis of SM byacid sphingomyelinase (ASM). HFD is known to increase CER and SMin liver [64]. The present observation of decreased SM and CER levelsin HFD-fed CYP7A1-tg mice indicated that bile acids might reduceHFD-induced increase of SM and CER. A previous report revealed that

DCA activates an ASM to convert SM to CER, and Asm−/− hepatocytesare resistant to DCA induction of CER and activation of the JNK pathway[65]. In CYP7A1-tg mice, enlarged bile acid pool inhibits CYP8B1 andreduces CA and DCA levels. Thus, decreasing DCA may reduce ASMactivity and SM and CER levels, and contribute to reducing inflammationand improving insulin sensitivity in CYP7A1-tgmice. It has been reportedrecently that in diabetic patients, serum 12α-hydroxylated bile acids areincreased and correlated to insulin resistance [66]. Reduction of CA is

Fig. 7. Quantitative real timePCR analysis of ileumand livermRNAexpression in CYP7A1-tg andwild-typemice. (A) IleummRNAexpression of genes involved in bile acidmetabolism andtransport (Asbt, Ostα, Ostβ) and fatty acid β-oxidation (Pparα, Acad1, Ehhadh, Acaa1), and fatty acid transporter Cd36 in female chow-fedmice. (B) Hepatic mRNA expression of genes in bileacidmetabolism (Cyp8b1), cholesterol efflux transporters (Abcg5, Abcg8) and fatty acid oxidation (Ehhadh, Acaa1) and fatty acid transport (Cd36) in female chow-fedmice. (C) HepaticmRNAexpression of enzymes in bile acid conjugation (Bacs and Bat), HNF4α, and taurine biosynthesis (Taut and Csd) in chow fed female CYP7A1-tg mice. Data were expressed as mean ± SD.Significance comparison was based on two-tailed Student's t-test or Mann-Whitney test. An * indicates p b 0.05 and an ** indicates p b 0.01 with respect to the wild-type (WT) group. Abbre-viations: Tg, transgenic mice overexpressing CYP7A1 in the liver; WT, wild-type mice; Asbt, apical sodium-dependent bile acid transporter; Ostα, organic solute transporter-α; Ostβ, organicsolute transporter-β; Pparα, peroxisome proliferator-activated receptorα; Acadl, long-chain specific acyl-CoA dehydrogenase; Ehhadh, enoyl-CoA hydratase/liter-3-hydroxyacyl-CoA dehydro-genase; Acaa1, acetyl-CoA acyltransferase 1; Abcg5, ATP-binding cassette G5; Abcg8, ATP-binding cassette G8; Cd36, fatty acid transporter; Taut, taurine transporter; Csd, cysteine sulfonic aciddecarboxylase.

27Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

known to reduce intestinal cholesterol and fat absorption, and improveinsulin resistance and obesity in CYP7A1-tg mice [38].

Several recent studies show that gutmicrobiota are altered in obesi-ty and NAFLD [67]. Gut microbiota regulate bile acid metabolism andreduce T-β-MCA [56,68]. In germ free mice and antibiotic-treatedmice, tauro-conjugated bile acids, especially T-β-MCA, are increased ascompared to conventionally raised mice [68]. Ampicillin increasedhepatic primary bile acid synthesis and suppressed ileal FGF15 expres-sion [69]. It has been suggested that T-β-MCA and T-α-MCA are FXRantagonists that reduce FXR induction of FGF15 resulting in increasedhepatic CYP7A1 expression and bile acid synthesis [56]. A study ofCyp8b1-/- mice showed increased bile acid synthesis and pool size,reduced fecal bile acid excretion and increased bile acid reabsorptionin the intestine [70]. The present metabolomics analysis revealed asignificant increase of T-β-MCA in ileum and colon of CYP7A1-tg mice,which had increased taurine synthesis and bile acid conjugation. Bileacids also have been implicated in modulation of gut microbiota [71].The lack of CA and DCA in the intestine may alter gut microbiota todecrease bile acid de-conjugation and increase T-β-MCA in the intestine.

Inhibition of CYP8B1 increases synthesis of CDCA, which is converted toα- and β-MCAs in the liver. Thus, reducing CA synthesis may increaseT-β-MCA and resulting in increasing bile acid synthesis, intestinal bileacid reabsorption and bile acid pool size [70]. On the other hand, highsaturated fats increase TCA to promote the expansion of the low abun-dant Bilophila wadsworthia, which is associated with increased pro-inflammatory cytokines and colitis in IL10-/- mice [72].

In conclusion,metabolomics and lipidomicswere employed to char-acterize the metabolic profiles of CYP7A1-tg mice, aiming to providenew insights into the mechanism of bile acid signaling in regulation oflipid metabolism and maintain lipid homeostasis. A number of lipidand bile acid markers were unveiled in this study. Decreasing of lipidmarkers, especially SM and CER may explain the improved insulin sen-sitivity and obesity in CYP7A1-tg mice. Furthermore, this study uncov-ered that enlarged bile acid pool size and altered bile acid compositionmay reduce de-conjugation by gut microbiota and increase tauro-conjugated muricholic acids, which partially inhibit intestinal FXR sig-naling without affecting hepatic FXR signaling. This study is significantin applying metabolomics for diagnosis of lipid biomarkers for fatty

Fig. 8. Mechanisms of anti-diabetic and anti-obesity function of bile acids in CYP7A1-tgmice. In CYP7A1-tg mice, overexpressing CYP7A1 increases bile acid pool size and reducescholic acid by inhibiting CYP8B1. Lipidomics analysis revealed decreased serum LPC, PC, SMand CER. These lipidomic markers are increased in hepatic steatosis and NAFLD. Bile acidsmay reduce LPC, PC, SM and CER levels and protect against high fat diet-induced insulin re-sistance and obesity in CYP7A1-tg mice. Metabolomics analysis showed decreased intestinalTCAand TDCAand increased intestinal T-β-MCA in CYP7A1-tgmice.High fat diets are knownto increase CA synthesis and intestinal inflammation. It is proposed that decreasing CA andDCA synthesismay increase intestinal T-β-MCA,which antagonizes FXR signaling to increasebile acid synthesis and prevent high fat diet-induced insulin resistance and obesity.

28 Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

liver diseases, obesity and diabetes. Increasing CYP7A1 activity and bileacid synthesis coupled to decreasing CYP8B1 and 12α-hydroxylated-bile acids may be a therapeutic strategy for treating diabetes and obesity.

Acknowledgements

This study was supported by the National Cancer Institute IntramuralResearch Program and by R37DK058379 and R01DK044442 to JYLC fromNIDDK, NIH.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbalip.2014.04.008.

References

[1] J.Y. Chiang, Bile acids: regulation of synthesis, J. Lipid Res. 50 (2009) 1955–1966.[2] C.R. Pullinger, C. Eng, G. Salen, S. Shefer, A.K. Batta, S.K. Erickson, et al., Human

cholesterol 7α-hydroxylase (CYP7A1) deficiency has a hypercholesterolemicphenotype, J. Clin. Invest. 110 (2002) 109–117.

[3] P.B. Hylemon, H. Zhou, W.M. Pandak, S. Ren, G. Gil, P. Dent, Bile acids as regulatorymolecules, J. Lipid Res. 50 (2009) 1509–1520.

[4] M.Makishima, A.Y. Okamoto, J.J. Repa, H. Tu, R.M. Learned, A. Luk, et al., Identificationof a nuclear receptor for bile acids, Science 284 (1999) 1362–1365.

[5] D.J. Parks, S.G. Blanchard, R.K. Bledsoe, G. Chandra, T.G. Consler, S.A. Kliewer, et al., Bileacids: natural ligands for an orphan nuclear receptor, Science 284 (1999) 1365–1368.

[6] H. Wang, J. Chen, K. Hollister, L.C. Sowers, B.M. Forman, Endogenous bile acids areligands for the nuclear receptor FXR/BAR, Mol. Cell 3 (1999) 543–553.

[7] Y. Kawamata, R. Fujii, M. Hosoya,M. Harada, H. Yoshida, M. Miwa, et al., A G protein-coupled receptor responsive to bile acids, J. Biol. Chem. 278 (2003) 9435–9440.

[8] T. Maruyama, Y. Miyamoto, T. Nakamura, Y. Tamai, H. Okada, E. Sugiyama, et al.,Identification of membrane-type receptor for bile acids (M-BAR), Biochem. Biophys.Res. Commun. 298 (2002) 714–719.

[9] C. Thomas, A. Gioiello, L. Noriega, A. Strehle, J. Oury, G. Rizzo, et al., TGR5-mediatedbile acid sensing controls glucose homeostasis, Cell Metab. 10 (2009) 167–177.

[10] F. Alemi, D.P. Poole, J. Chiu, K. Schoonjans, F. Cattaruzza, J.R. Grider, et al., The receptorTGR5mediates the prokinetic actions of intestinal bile acids and is required for normaldefecation in mice, Gastroenterology 144 (2013) 145–154.

[11] V. Keitel, K. Cupisti, C. Ullmer,W.T. Knoefel, R. Kubitz, D. Haussinger, Themembrane-bound bile acid receptor TGR5 is localized in the epithelium of human gallbladders,Hepatology 50 (2009) 861–870.

[12] V. Keitel, C. Ullmer, D. Haussinger, The membrane-bound bile acid receptor TGR5(Gpbar-1) is localized in the primary cilium of cholangiocytes, Biol. Chem. 391(2010) 785–789.

[13] E. Studer, X. Zhou, R. Zhao, Y. Wang, K. Takabe, M. Nagahashi, et al., Conjugated bileacids activate the sphingosine-1-phosphate receptor 2 in primary rodent hepato-cytes, Hepatology 55 (2012) 267–276.

[14] T. Inagaki, M. Choi, A. Moschetta, L. Peng, C.L. Cummins, J.G. McDonald, et al.,Fibroblast growth factor 15 functions as an enterohepatic signal to regulate bileacid homeostasis, Cell Metab. 2 (2005) 217–225.

[15] P. Dent, Y. Fang, S. Gupta, E. Studer, C. Mitchell, S. Spiegel, et al., Conjugated bile acidspromote ERK1/2 and AKT activation via a pertussis toxin-sensitive mechanism inmurine and human hepatocytes, Hepatology 42 (2005) 1291–1299.

[16] R.T. Stravitz, Y.P. Rao, Z.R. Vlahcevic, E.C. Gurley, W.D. Jarvis, P.B. Hylemon,Hepatocellular protein kinase C activation by bile acids: implications for regulationof cholesterol 7α-hydroxylase, Am. J. Physiol. 34 (1996) G293–G303.

[17] K.H. Song, E. Ellis, S. Strom, J.Y. Chiang, Hepatocyte growth factor signaling pathwayinhibits cholesterol 7α-hydroxylase and bile acid synthesis in human hepatocytes,Hepatology 46 (2007) 1993–2002.

[18] S. Seok, D. Kanamaluru, Z. Xiao, D. Ryerson, S.E. Choi, K. Suino-Powell, et al., Bile acidsignal-induced phosphorylation of small heterodimer partner by protein kinaseczeta is critical for epigenomic regulation of liver metabolic genes, J. Biol. Chem.288 (2013) 23252–23263.

[19] A. Jahan, J.Y. Chiang, Cytokine regulation of human sterol 12α-hydroxylase(CYP8B1) gene, Am. J. Physiol. Gastrointest. Liver Physiol. 288 (2005) G685–G695.

[20] T. Li, A. Jahan, J.Y. Chiang, Bile acids and cytokines inhibit the human cholesterol7alpha-hydroxylase gene via the JNK/c-jun pathway in human liver cells, Hepatology43 (2006) 1202–1210.

[21] N.C. Hait, J. Allegood, M. Maceyka, G.M. Strub, K.B. Harikumar, S.K. Singh, et al.,Regulation of histone acetylation in the nucleus by sphingosine-1-phosphate,Science 325 (2009) 1254–1257.

[22] L.J. Schoenfield, J.M. Lachin, Chenodiol (chenodeoxycholic acid) for dissolution ofgallstones: the National Cooperative Gallstone Study. A controlled trial of efficacyand safety, Ann. Intern. Med. 95 (1981) 257–282.

[23] A. Garg, S.M. Grundy, Cholestyramine therapy for dyslipidemia in non-insulin-dependent diabetes mellitus. A short-term, double-blind, crossover trial, Ann. Intern.Med. 121 (1994) 416–422.

[24] B. Angelin, K. Einarsson, K. Hellstrom, B. Leijd, Effects of cholestyramine andchenodeoxycholic acid on the metabolism of endogenous triglyceride inhyperlipoproteinemia, J. Lipid Res. 19 (1978) 1017–1024.

[25] Y. Zhang, F.Y. Lee, G. Barrera, H. Lee, C. Vales, F.J. Gonzalez, et al., Activation of thenuclear receptor FXR improves hyperglycemia and hyperlipidemia in diabeticmice, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 1006–1011.

[26] M. Watanabe, S.M. Houten, L. Wang, A. Moschetta, D.J. Mangelsdorf, R.A. Heyman,et al., Bile acids lower triglyceride levels via a pathway involving FXR, SHP, andSREBP-1c, J. Clin. Invest. 113 (2004) 1408–1418.

[27] K. Ma, P.K. Saha, L. Chan, D.D. Moore, Farnesoid X receptor is essential for normalglucose homeostasis, J. Clin. Invest. 116 (2006) 1102–1109.

[28] T. Lundasen, C. Galman, B. Angelin, M. Rudling, Circulating intestinal fibroblastgrowth factor 19 has a pronounced diurnal variation and modulates hepatic bileacid synthesis in man, J. Intern. Med. 260 (2006) 530–536.

[29] E. Tomlinson, L. Fu, L. John, B. Hultgren, X. Huang, M. Renz, et al., Transgenic miceexpressing human fibroblast growth factor-19 display increased metabolic rateand decreased adiposity, Endocrinology 143 (2002) 1741–1747.

[30] L. Fu, L.M. John, S.H. Adams, X.X. Yu, E. Tomlinson, M. Renz, et al., Fibroblast growthfactor 19 increases metabolic rate and reverses dietary and leptin-deficient diabetes,Endocrinology 145 (2004) 2594–2603.

[31] M.J. Potthoff, J. Boney-Montoya, M. Choi, T. He, N.E. Sunny, S. Satapati, et al., FGF15/19 regulates hepatic glucose metabolism by inhibiting the CREB-PGC-1α pathway,Cell Metab. 13 (2011) 729–738.

[32] S. Kir, S.A. Beddow, V.T. Samuel, P. Miller, S.F. Previs, K. Suino-Powell, et al., FGF19 asa postprandial, insulin-independent activator of hepatic protein and glycogensynthesis, Science 331 (2011) 1621–1624.

[33] S. Bhatnagar, H.A. Damron, F.B. Hillgartner, Fibroblast growth factor-19, a novelfactor that inhibits hepatic fatty acid synthesis, J. Biol. Chem. 284 (2009) 10023–10033.

[34] M. Miyata, Y. Sakaida, H. Matsuzawa, K. Yoshinari, Y. Yamazoe, Fibroblast growthfactor 19 treatment ameliorates disruption of hepatic lipid metabolism in farnesoidX receptor (Fxr)-null mice, Biol. Pharm. Bull. 34 (2011) 1885–1889.

[35] J.H. Miyake, X.T. Duong-Polk, J.M. Taylor, E.Z. Du, L.W. Castellani, A.J. Lusis, et al.,Transgenic expression of cholesterol-7α-hydroxylase prevents atherosclerosis inC57BL/6 J mice, Arterioscler. Thromb. Vasc. Biol. 22 (2002) 121–126.

[36] T. Li, E. Owsley, M. Matozel, P. Hsu, C.M. Novak, J.Y. Chiang, Transgenic expression ofcholesterol 7α-hydroxylase in the liver prevents high-fat diet-induced obesity andinsulin resistance in mice, Hepatology 52 (2010) 678–690.

[37] T. Li, J.M. Francl, S. Boehme, J.Y. Chiang, Regulation of cholesterol and bile acidhomeostasis by the cholesterol 7α-hydroxylase/steroid response element-bindingprotein 2/microRNA-33a axis in mice, Hepatology 58 (2013) 1111–1121.

[38] T. Li, M. Matozel, S. Boehme, B. Kong, L.M. Nilsson, G. Guo, et al., Overexpression ofcholesterol 7alpha-hydroxylase promotes hepatic bile acid synthesis and secretionand maintains cholesterol homeostasis, Hepatology 53 (2011) 996–1006.

[39] J.K. Nicholson, J.C. Lindon, E. Holmes, 'Metabonomics': understanding the met-abolic responses of living systems to pathophysiological stimuli via multivari-ate statistical analysis of biological NMR spectroscopic data, Xenobiotica 29(1999) 1181–1189.

[40] J.K. Nicholson, J. Connelly, J.C. Lindon, E. Holmes, Metabonomics: a platformfor studying drug toxicity and gene function, Nat. Rev. Drug Discov. 1 (2002)153–161.

29Y. Qi et al. / Biochimica et Biophysica Acta 1851 (2015) 19–29

[41] M. Oresic, A. Vidal-Puig, V. Hanninen, Metabolomic approaches to phenotypecharacterization and applications to complex diseases, Expert. Rev. Mol. Diagn. 6(2006) 575–585.

[42] J. Hartler, R. Tharakan, H.C. Kofeler, D.R. Graham, G.G. Thallinger, Bioinformatics toolsand challenges in structural analysis of lipidomics MS/MS data, Brief. Bioinform. 14(2013) 375–390.

[43] S.A. Murphy, A. Nicolaou, Lipidomics applications in health, disease and nutritionresearch, Mol. Nutr. Food Res. 57 (2013) 1336–1346.

[44] F. Li, A.D. Patterson, K.W. Krausz, C. Jiang, H. Bi, A.L. Sowers, et al., Metabolomicsreveals that tumor xenografts induce liver dysfunction, Mol. Cell. Proteomics 12(2013) 2126–2135.

[45] C.H. Johnson, O. Slanar, K.W. Krausz, D.W. Kang, A.D. Patterson, J.H. Kim, et al., Novelmetabolites and roles for alpha-tocopherol in humans and mice discovered by massspectrometry-based metabolomics, Am. J. Clin. Nutr. 96 (2012) 818–830.

[46] T. Matsubara, N. Tanaka, K.W. Krausz, S.K. Manna, D.W. Kang, E.R. Anderson,et al., Metabolomics identifies an inflammatory cascade involved in dioxin-and diet-induced steatohepatitis, Cell Metab. 16 (2012) 634–644.

[47] A.D. Patterson, O. Maurhofer, D. Beyoglu, C. Lanz, K.W. Krausz, T. Pabst, et al.,Aberrant lipid metabolism in hepatocellular carcinoma revealed by plasmametabo-lomics and lipid profiling, Cancer Res. 71 (2011) 6590–6600.

[48] J.H. Miyake, X.D. Doung, W. Strauss, G.L. Moore, L.W. Castellani, L.K. Curtiss, et al.,Increased production of Apo B100-containing lipoproteins in the absence ofhyperlipidemia in transgenic mice expressing cholesterol 7α- hydroxylase, J. Biol.Chem. 276 (2001) 23304–23311.

[49] P. Jonsson, S.J. Bruce, T. Moritz, J. Trygg, M. Sjostrom, R. Plumb, et al., Extraction,interpretation and validation of information for comparing samples in metabolicLC/MS data sets, Analyst 130 (2005) 701–707.

[50] Y.A. Hannun, The sphingomyelin cycle and the second messenger function ofceramide, J. Biol. Chem. 269 (1994) 3125–3128.

[51] S.M. Hammad, J.S. Pierce, F. Soodavar, K.J. Smith, M.M. Al Gadban, B. Rembiesa, et al.,Blood sphingolipidomics in healthy humans: impact of sample collection methodol-ogy, J. Lipid Res. 51 (2010) 3074–3087.

[52] J. Li, C. Hu, X. Zhao, W. Dai, S. Chen, X. Lu, et al., Large-scaled human serumsphingolipid profiling by using reversed-phase liquid chromatography coupled withdynamic multiple reaction monitoring of mass spectrometry: method developmentand application in hepatocellular carcinoma, J. Chromatogr. A 1320 (2013) 103–110.

[53] A.E. Alewijnse, S.L. Peters, Sphingolipid signalling in the cardiovascular system:good, bad or both? Eur. J. Pharmacol. 585 (2008) 292–302.

[54] S.Wiklund, E. Johansson, L. Sjostrom, E.J. Mellerowicz, U. Edlund, J.P. Shockcor, et al.,Visualization of GC/TOF-MS-basedmetabolomics data for identification of biochem-ically interesting compounds using OPLS class models, Anal. Chem. 80 (2008)115–122.

[55] B. Kong, L. Wang, J.Y. Chiang, Y. Zhang, C.D. Klaassen, G.L. Guo, Mechanism of tissue-specific farnesoid X receptor in suppressing the expression of genes in bile-acidsynthesis in mice, Hepatology 56 (2012) 1034–1043.

[56] S.I. Sayin, A. Wahlstrom, J. Felin, S. Jantti, H.U. Marschall, K. Bamberg, et al., Gut micro-biota regulates bile acid metabolism by reducing the levels of tauro-beta-muricholicacid, a naturally occurring FXR antagonist, Cell Metab. 17 (2013) 225–235.

[57] D. Chanda, C.H. Lee, Y.H. Kim, J.R. Noh, D.K. Kim, J.H. Park, et al., Fenofibrate differen-tially regulates plasminogen activator inhibitor-1 gene expression via adenosinemonophosphate-activated protein kinase-dependent induction of orphan nuclearreceptor small heterodimer partner, Hepatology 50 (2009) 880–892.

[58] J.C. Garcia-Canaveras, M.T. Donato, J.V. Castell, A. Lahoz, A comprehensive untargetedmetabonomic analysis of human steatotic liver tissue by RP andHILIC chromatographycoupled to mass spectrometry reveals important metabolic alterations, J. ProteomeRes. 10 (2011) 4825–4834.

[59] D. Beyoglu, J.R. Idle, The metabolomic window into hepatobiliary disease, J. Hepatol.59 (2013) 842–858.

[60] J.A. Chavez, S.A. Summers, A ceramide-centric view of insulin resistance, Cell Metab.15 (2012) 585–594.

[61] A.U. Blachnio-Zabielska, M. Pulka, M. Baranowski, A. Nikolajuk, P. Zabielski, M.Gorska, et al., Ceramide metabolism is affected by obesity and diabetes in humanadipose tissue, J. Cell. Physiol. 227 (2012) 550–557.

[62] S. Galadari, A. Rahman, S. Pallichankandy, A. Galadari, F. Thayyullathil, Role ofceramide in diabetes mellitus: evidence and mechanisms, Lipids Health Dis. 12(2013) 98.

[63] J. Jin, X. Zhang, Z. Lu, D.M. Perry, Y. Li, S.B. Russo, et al., Acid sphingomyelinase playsa key role in palmitic acid-amplified inflammatory signaling triggered by lipopoly-saccharide at low concentrations in macrophages, Am. J. Physiol. Endocrinol.Metab. 305 (2013) E853–E867.

[64] G. Chocian, A. Chabowski, M. Zendzian-Piotrowska, E. Harasim, B. Lukaszuk, J.Gorski, High fat diet induces ceramide and sphingomyelin formation in rat's livernuclei, Mol. Cell. Biochem. 340 (2010) 125–131.

[65] S. Gupta, R. Natarajan, S.G. Payne, E.J. Studer, S. Spiegel, P. Dent, et al., Deoxycholicacid activates the c-Jun N-terminal kinase pathway via FAS receptor activation inprimary hepatocytes: role of acidic sphingomyelinase-mediated ceramide genera-tion in FAS receptor activation, J. Biol. Chem. 279 (2004) 5821–5828.

[66] R.A. Haeusler, B. Astiarraga, S. Camastra, D. Accili, E. Ferrannini, Human insulinresistance is associated with increased plasma levels of 12α-hydroxylated bileacids, Diabetes 62 (2013) 4184–4191.

[67] J. Aron-Wisnewsky, B. Gaborit, A. Dutour, K. Clement, Gut microbiota and non-alcoholic fatty liver disease: new insights, Clin. Microbiol. Infect. 19 (2013) 338–348.

[68] J.R. Swann, E.J. Want, F.M. Geier, K. Spagou, I.D. Wilson, J.E. Sidaway, et al., Systemicgut microbial modulation of bile acidmetabolism in host tissue compartments, Proc.Natl. Acad. Sci. U. S. A. 108 (Suppl. 1) (2011) 4523–4530.

[69] M. Miyata, Y. Takamatsu, H. Kuribayashi, Y. Yamazoe, Administration of ampicillinelevates hepatic primary bile acid synthesis through suppression of ileal fibroblastgrowth factor 15 expression, J. Pharmacol. Exp.Ther. 331 (2009) 1079–1085.

[70] X. Hu, Y. Bonde, G. Eggertsen, M. Rudling, Muricholic bile acids are potent regulators ofbile acid synthesis via a positive feedbackmechanism, J. Intern.Med. 275 (2014) 27–38.

[71] L.A. David, C.F. Maurice, R.N. Carmody, D.B. Gootenberg, J.E. Button, B.E. Wolfe, et al.,Diet rapidly and reproducibly alters the human gut microbiome, Nature 505 (2013)559–563.

[72] S. Devkota, Y. Wang, M.W. Musch, V. Leone, H. Fehlner-Peach, A. Nadimpalli, et al.,Dietary-fat-induced taurocholic acid promotes pathobiont expansion and colitis inIl10-/- mice, Nature 487 (2012) 104–108.