The FASEB Journal • Research Communication

The hepatic phosphatidylcholine transporter ABCB4 asmodulator of glucose homeostasis

Katrin Hochrath,*,1 Marcin Krawczyk,*,1 Reinhild Goebel,* Miriam Langhirt,*Birgit Rathkolb,†,‡ Kateryna Micklich,†,‡ Jan Rozman,‡,§ Marion Horsch,‡

Johannes Beckers,‡,� Martin Klingenspor,§ Helmut Fuchs,‡ Valérie Gailus-Durner,‡

Eckhard Wolf,† Monica Acalovschi,¶ Dietrich A. Volmer,#

Martin Hrabe de Angelis,‡,¶,** and Frank Lammert*,2

*Department of Medicine II, Saarland University Medical Center, Homburg, Germany; †Instituteof Molecular Animal Breeding and Biotechnology, Ludwig-Maximilians-Universität München,Munich, Germany; ‡German Mouse Clinic, Institute of Experimental Genetics, Helmholtz ZentrumMünchen, German Research Center for Environmental Health GmbH, Neuherberg, Germany;§Department of Molecular Nutritional Medicine, Else Kröner-Fresenius Center and Zentralinstitut furErnährungs- und Lebensmittelforschung (ZIEL) Research Center for Nutrition and Food Science,and �Department of Experimental Genetics, Technische Universität München, Freising-Weihenstephan, Germany; ¶Department of Medicine III, Iuliu Hatieganu University of Medicine andPharmacy, Cluj-Napoca, Romania; #Institute of Bioanalytical Chemistry, Saarland University,Saarbrücken, Germany; and **German Center for Diabetes Research, Neuherberg, Germany

ABSTRACT The hepatic phosphatidylcholine (PC)transporter ATP-binding cassette (ABC) B4 flops PC fromhepatocytes into bile, and its dysfunction causes chroniccholestasis and fibrosis. Because a nuclear receptor-de-pendent PC pathway has been determined to exertantidiabetic effects, we now analyzed the role ofABCB4 in glucose metabolism. We bred congenicAbcb4-knockout (Abcb4�/�) mice on the fibrosis-sus-ceptible BALB/cJ background. Knockout mice andwild-type controls were phenotyped by measuringplasma glucose concentrations, intraperitoneal glu-cose tolerance, hepatic RNA expression profiles, andliver histology. In addition, 4 procholestatic ABCB4gene variants were correlated with blood glucoselevels in 682 individuals from 2 independent Euro-pean cohorts. Systemic glucose levels differ signifi-cantly between Abcb4�/� mice and wild-type controls,and knockout mice display improved glucose toler-ance with significantly lower area under the curvevalues on intraperitoneal glucose challenge. Of note,hepatic expression of the antidiabetic nuclear recep-tor 5A2 (LRH-1) is induced consistently in Abcb4�/�

mice, and its specific rare PC ligands are detected in

liver by mass spectrometry imaging. In humans,serum glucose levels are associated significantly withthe common ABCB4 variant c.711A>T. In summary,ABCB4 might play a critical role in glucose homeo-stasis in mice and humans. We speculate that theeffects could be mediated via LRH-1-dependent PCpathways.—Hochrath, K., Krawczyk, M., Goebel, R.,Langhirt, M., Rathkolb, B., Micklich, K., Rozman, J.,Horsch, M., Beckers, J., Klingenspor, M., Fuchs, H.,Gailus-Durner, V., Wolf, E., Acalovschi, M., Volmer,D. A., Hrabe de Angelis, M., Lammert, F. The hepaticphosphatidylcholine transporter ABCB4 as modulatorof glucose homeostasis. FASEB J. 26, 5081–5091 (2012).www.fasebj.org

Key Words: ATP-binding cassette transporter � chronic cholangitis� liver fibrosis � liver receptor homolog

Bile is composed mainly of water, sterols, bile salts,and phospholipids, in particular, phosphatidylcho-line (PC) (1). These substances, apart from water,are transported from hepatocytes into bile by specifichepatobiliary ATP-binding cassette (ABC) proteins,namely ABCB4 (PC floppase), ABCB11 (bile saltexport pump), and ABCG5/8 (sterol transporter) (2).Once transferred into the lumen of bile ducts, sterols,bile salts, and PC form mixed micelles to reduce the

1 These authors contributed equally to this work.2 Correspondence: Department of Medicine II, Saarland

University Medical Center, Kirrberger Str. 100, 66421 Hom-burg, Germany. E-mail: frank.lammert@uks.eu

doi: 10.1096/fj.12-209379This article includes supplemental data. Please visit http://

www.fasebj.org to obtain this information.

Abbreviations: ABC, ATP-binding cassette; AIC, Akaikeinformation criterion; ALT, alanine aminotransferase; AP,alkaline phosphatase; AUC, area under the curve; BMI,body mass index; DLPC, dilauroyl-phosphatidylcholine;DUPC, diundecanoyl-phosphatidylcholine; FTICR, Fouriertransform ion cyclotron resonance; Hyp, hydroxyproline;LDH, lactate dehydrogenase; IpGTT, intraperitoneal glu-cose tolerance test; LRH-1, liver receptor homolog 1;MALDI, matrix-assisted laser desorption/ionization; MSI,mass spectrometry imaging; PC, phospatidylcholine; SNP,single-nucleotide polymorphism

50810892-6638/12/0026-5081 © FASEB Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from  Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from  Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

toxicity of bile acids and to establish physiological bileflow. Conversely, in the case of dysfunction of the ABCproteins, bile composition is deranged and cholestaticdisorders with liver dysfunction prevail (3). We andothers identified ABCB4 gene variants that increase therisk of cholestasis in pregnancy (4, 5) and showed thatrare mutations in the hepatic PC transporter cause theso-called low phospholipid-associated cholelithiasis syn-drome, characterized by recurrent extra- and intrahe-patic gallstones in young adults (6, 7). In line with theseresults, mice lacking the ABCB4 transporter are char-acterized by the absence of PC in bile and increasedbile toxicity (8, 9). As a result, ABCB4-deficient animalsspontaneously develop chronic cholangitis and second-ary biliary fibrosis (10), resembling sclerosing cholan-gitis in humans (11).

PC flopped from hepatocytes into bile by ABCB4 iscomposed of a variety of fatty acids. As we reportedpreviously (9), the major PC species in murine bileinclude C16:0/C18:2, C16:0/C18:1, and C16:0/C20:4fatty acids. Furthermore, PCs with even shorter satu-rated fatty acid acyl side chains [i.e., diundecanoyl-PC(DUPC), C11:0/C11:0 and dilauroyl-PC (DLPC),C12:0/C12:0] have been recognized as endogenousligands of the nuclear receptor liver receptor ho-molog-1 (LRH-1) (12, 13). Nuclear receptors serve astranscriptional factors, and LRH-1 is expressed pre-dominantly in liver, small intestine, preadipocytes,and adrenal glands, where it is involved in metabolicpathways governing sterol transport, bile salt homeo-stasis, and steroidogenesis (14). Interestingly, activa-tion of LRH-1 by PC has been demonstrated toimprove insulin sensitivity in diabetic mice (13).Besides this, studies of PC transfer protein-knockout(Pctp�/�) mice show a link of specific PC species andthe regulation of hepatic glucose metabolism (15, 16).

To date, Abcb4-knockout (Abcb4�/�) mice havebeen broadly investigated with respect to liver phe-notypes (8 –10, 17), but associations between dys-function of the ABCB4 transporter and glucose me-tabolism or with other metabolic traits have not beenelucidated. Of note, genetic studies in humans havelinked the 7q21.1 locus, harboring the ABCB4 gene,with an increased risk of diabetic nephropathy (18),whereas linkage and association between 7q21.1–21.3loci and non-insulin-dependent diabetes, as well asinsulin resistance, were reported almost 2 decadesago (19, 20).

In our current study, we have performed a system-atic phenotyping (21) of Abcb4�/� mice and found anovel association between glucose levels and defi-ciency of the hepatic PC transporter. Subsequentexpression analyses in Abcb4�/� animals allowed usto dissect metabolic pathways that contribute tolower glucose levels. In addition, we genotyped 4common ABCB4 polymorphisms in 682 adult individ-uals from 2 independent European cohorts andcorrelated serum glucose levels with the frequenciesof procholestatic variants.

MATERIALS AND METHODS

Animals

Generation of congenic BALB-Abcb4�/� mice

The BALB-Abcb4�/� mice were generated by backcrossingthe Abcb4tm1Bor-knockout from the fibrosis-resistant FVB/NJstrain (The Jackson Laboratory, Bar Harbor, ME, USA) intothe fibrosis-susceptible BALB/cJ background for 10 genera-tions (22, 23). BALB/cJ control mice were obtained fromCharles River Laboratories (Sulzfeld, Germany). Mice werehoused in individually ventilated cages with a 12-h light-darkcycle; temperature and humidity were regulated to 22 � 1°Cand 55 � 5%, respectively, with water and standard diet(Altromin 1314; Altromin, Lage, Germany) provided ad libitum. Allmice used for experiments were aged between 14 to 20 wk.

Genotypes of mice were confirmed by PCR of tail DNAusing neo (5=-CTTGGGTGGAGAGGCTATTC-3= and 5=-AG-GTGAGATGACAGGAGATC-3=) and Abcb4 (5=-CACTTGGA-CCTGAGGCTGTG-3= and 5=-TCAGGACTCCGCTATAACGG-3=) specific primer pairs. The PCR contained 10� PCR buffer(Applied Biosystems, Darmstadt, Germany), 2 mM MgCl2, 10�M dNTPs, 10 �M primer, 1.25 U TaqDNA polymerase(Invitrogen, Darmstadt, Germany) and 20-100 ng of DNA in25-�l reactions. PCR cycling conditions were 94°C/30 s,55°C/60 s, and 72°C/30 s for 35 cycles, with a final extensionstep of 10 min at 72°C.

Phenotypic characterization of biliary fibrosis

Liver samples for histopathological evaluation were fixed in4% neutral buffered formalin at 4°C for 24 h and embeddedin paraffin. Paraffin sections (1 �m) were stained withhematoxylin-eosin or Sirius red for the detection of collagen.Histomorphometric analysis of hepatic fibrosis was per-formed with semiautomatic microscopic image analysis (Leicaapplication suite software; Leica, Wetzlar, Germany). Relativecollagenous areas were calculated from 10 microscopic fields(�100 view) randomly chosen on each liver section. Inaddition, we quantified collagen in liver by colorimetricmeasurement of the collagen specific amino acid hydroxypro-line (Hyp), as described previously (22, 24).

Blood samples were obtained from isoflurane-anesthetizedmice by puncturing the retro-orbital sinus with capillaries andcollected in heparinized tubes. Plasma alanine aminotransfer-ase and alkaline phosphatase activities, as well as glucoseconcentrations, were measured with an Olympus AU400chemistry analyzer, using adapted reagents provided by Olym-pus (Hamburg, Germany).

Intraperitoneal glucose tolerance test (IpGTT)

After overnight food withdrawal (16–18 h), mice wereweighed and intraperitoneally injected with 2 g glucose/kgbody weight. Blood samples for glucose measurements (Accu-Chek Aviva; Roche Diagnostics, Mannheim, Germany) werecollected from tail vein and analyzed before glucose injectionand at 15, 30, 60, and 120 min afterward. For calculating thearea under the curve (AUC), the trapezoidal rule was used,and the area below the baseline glucose level was excluded.

Genome-wide hepatic expression profiling

For genome-wide expression analysis, liver tissue of 4 maleAbcb4�/� and 4 BALB/cJ control mice was snap-frozen inliquid nitrogen, and total RNA was isolated using an RNeasy

5082 Vol. 26 December 2012 HOCHRATH ET AL.The FASEB Journal � www.fasebj.org Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

Midi kit (Qiagen, Hilden, Germany). The cDNA microarrayswere generated, hybridized, and analyzed as described previ-ously (25, 26). Two-chip hybridizations were performed withtotal RNA for each individual mutant mouse against a refer-ence RNA pool of the same organ. The normalization (27)and the selection of the significantly differentially expressedgenes with reproducible up- or down-regulation (26, 28)includes 1.3% false-positive results in combination with meanfold change �1.3�. Expression data were submitted to theGene Expression Omnibus database (GSE26699; http://www.ncbi.nlm.nih.gov/geo/), where a full description of themicroarray platform is also available (GPL4937). IPA (http://www.ingenuity.com) was used to identify overrepresentedGene Ontology (GO) terms in the categories molecularfunctions and canonical pathways among the differentiallyexpressed genes.

Hepatic expression of genes involved in glucose homeostasis

To determine the hepatic expression of single genes involvedin glucose metabolism, liver samples of mice allowed access tofood overnight were harvested. For expression analysis duringIpGTT, organs were collected from animals denied access tofood overnight that were euthanized at baseline (t0) and after60 min (t60). All samples were immediately snap-frozen inliquid nitrogen and stored at �80°C until analyzed. Tominimize the effects of circadian rhythm, all mice wereeuthanized between 9 and 11 AM.

The steady-state hepatic mRNA expression levels of indi-vidual genes were determined by quantitative real-time PCR(TaqMan; Applied Biosystems, Foster City, CA, USA), using 1�g of RNA for reverse transcription and 18S RNA as theendogenous control, with 1 cycle at 95°C for 10 min, followedby 45 cycles at 95°C for 30 s and 60°C for 60 s. The expressionlevel of each gene was calculated by the ��Ct method (29), inrelation to their counterpart wild-type controls and alsonormalized to wild-type animals. For expression analysis afterIpGTT, ��Ct was calculated relative to the dedicated geno-type-specific controls at t0 (e.g., ��CtLrh-1 � �CtLrh-1 BALB-Abcb4�/� female at t60 � mean �CtLrh-1 BALB-Abcb4�/�

females at t0). The relative quotient (2���Ct) of all data wasnormalized to values of BALB/cJ mice at t0.

Primary hepatocytes

Hepatocytes of BALB/cJ and BALB-Abcb4�/� mice (n�5–7/genotype and treatment group) were isolated using a colla-genase perfusion 2-step system as described previously (30).After purification, hepatocytes were cultured on collagen-coated 6-well plates (300,000 cells/well) in Williams’ E me-dium supplemented with 10% fetal bovine serum and 100 nMdexamethasone for 4 h. Subsequently, medium was changedfor overnight serum starvation. Hepatocytes were incubatedwith 100 �M DLPC (Tocris, Bristol, UK) or ethanol for 24 h.Cells were washed twice with ice-cold phosphate-bufferedsaline and harvested. Total RNA was isolated with the QiagenRNeasy Minikit as described above. Relative mRNA quantifi-cation by real-time PCR was performed for Cyp7a1, Cyp7b1,Cyp8b1, Gck1, G6pc, Lrh-1, Nr0b2, and Pck1 with TaqMan assaysas described above.

Identification of hepatic phospholipids

Mass spectrometry imaging (MSI) experiments for identifica-tion of phospholipid species including the LRH-1 ligandsDUPC (C30H61NO8P) and DLPC (C32H65NO8P) in liver wereperformed on a Bruker SolariX 12-T matrix-assisted laserdesorption/ionization (MALDI) Fourier transform ion cyclo-

tron resonance (FTICR) instrument equipped with an ApolloII MALDI/electrospray ionization source and a Smart BeamII NdYAG 1-kHz high-repetition-rate laser (Bruker, Bremen,Germay). Liver samples were snap-frozen, sliced into 4-�msections using a cryostat, and placed onto microscopy glassslides. Homogeneous MALDI matrix layers were depositedonto the tissue surface by means of a Bruker ImagePrepsystem using �-cyano-4-hydroxycinnamic acid solution at 7mg/ml dissolved in 50% acetonitrile-water (v/v) 0.2%trifluoroacetic acid. The pixel size in imaging experimentswas 75 �m. Positive ions were monitored in experiments inthe m/z range of 150–1500. For elemental formula assign-ments, no atom constraints for C, H, N, O, and P were set. Allcalculations were performed with the MIDAS molecular for-mula calculator (National High Magnetic Field Laboratory,Tallahassee, FL, USA).

Patients

Recruitment of cohorts

We analyzed two independent, previously described cohorts(n�682; 485 women; aged 21-89 yr) from Germany (n�366;ref. 31) and Romania (n�316; refs. 32, 33). Table 1 summa-rizes the baseline clinical characteristics of both cohorts. TheGerman cohort comprised unrelated individuals with gall-stones (n�183) and gallstone-free controls (n�183). TheRomanian cohort consisted of 235 siblings with gallstonesfrom 100 families and 216 unrelated gallstone-free controls.In the analysis, all German individuals, all Romanian controls,and one randomly selected sibling from each Romanianfamily were included. Serum fasting glucose levels from bloodsamples were determined by standard clinical chemical assaysin the central laboratories at both university hospitals. Noneof the individuals was receiving glucose-lowering therapy atinclusion. Informed consent was obtained from all patients,and the study was approved by the independent ethicscommittees of the University of Aachen (Aachen, Denmark)and the University of Cluj-Napoca (Cluj-Napoca, Romania).

Genotyping of ABCB4 polymorphisms

Genomic DNA was isolated from EDTA-anticoagulated bloodaccording to the membrane-based QIAamp DNA extraction

TABLE 1. Clinical characteristics of study cohorts

Characteristic Value

Patients (n)Total 682Romania 316Germany 366

Gender (n)Female 485Male 197

Serum glucose (mM)Median 5.36Range 2.76–21.82

Body mass index (kg/m2)Median 26Range 14–49

Age (yr)a

Median 59Range 21–89

aAge was calculated at the date of inclusion.

5083ABCB4 AS MODULATOR OF GLUCOSE HOMEOSTASIS Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

protocol (Qiagen). The ABCB4 single-nucleotide polymor-phisms (SNPs) in exon 6 (c.504C�T, rs1202283), exon 8(c.711A�T, rs2109505), exon 16 (c.1954A�G, rs2230028),and intron 26 (rs31653) were genotyped using solution-phasehybridization reactions with 5=-nuclease and fluorescencedetection (TaqMan SNP genotyping assays) in a 7500 FASTreal-time PCR system (TaqMan). The PCRs contained 20 to40 ng of DNA, 1� TaqMan genotyping Master Mix, and a 1�TaqMan SNP genotyping assay in 10-�l reactions. Amplificationconditions were 95°C for 10 min, followed by 40 cycles of 95°C for10 s and 60°C for 60 s. Genotyping success rates were �99%.

Statistics

Statistical analysis was performed with SPSS 19 (IBM,Ehningen, Germany), unless stated otherwise. The Kol-mogorov-Smirnov test was used to determine whether datahad a normal distribution. Quantitative data are expressedas means � se or median and ranges, as appropriate.Means and medians were compared with Student’s t tests orMann-Whitney U tests, respectively. Two-way ANOVA wasapplied to assess the influence of mouse gender and Abcb4genotype on glucose levels during IpGTT. For all tests,values of P 0.05 were regarded as significant.

The consistency of genotyping results with Hardy-Weinberg equilibrium was checked by exact test (http://ihg2.helmholtz-muenchen.de/cgi-bin/hw/hwa1.pl) (34).The effects of genotype, age, gender, and body mass index

(BMI) on glucose levels were assessed by linear regressionanalyses. To characterize the influence of the above factorson glucose levels and to obtain the optimal linear model, abackward variable selection using the Akaike informationcriterion (AIC) was computed. The AIC optimal modelincludes genotype, age, and BMI.

RESULTS

Hepatic phenotype of Abcb4�/� mice

Increased serum activities of alkaline phosphatase (AP)and bilirubin levels demonstrate prominent cholestasisin BALB-Abcb4�/� mice lacking the hepatic PC trans-porter compared with that in controls (Fig. 1A). Simul-taneous increases of alanine aminotransferase (ALT)and lactate dehydrogenase (LDH) activities indicatemixed hepatocellular and cholestatic liver injury. Fig-ure 1B demonstrates that the hepatic collagen concen-trations in Abcb4�/� mice, as assessed by Hyp contents,are markedly increased compared with those with con-trols (males: 289 vs. 191 �g of Hyp/g of liver; females:343 vs. 213 �g of Hyp/g of liver). Figure 1C showsrepresentative liver sections of BALB-Abcb4�/� mice

Figure 1. A) Liver enzyme activities and bilirubin concentrations in plasma from male and female BALB-Abcb4�/� and BALB/cJwild-type mice; n � 10/gender and genotype. *P 0.05; ***P 0.001. B) Hyp levels in Abcb4�/� (solid bars) and wild-type mice(open bars); n � 11-26/gender and genotype. **P 0.01. C) Representative liver sections from female (top panels) and male(bottom panels) Abcb4�/� mice (left panels) and wild-type controls (right panels). Sirius red stain. D) Hepatic collagen areasin Abcb4�/� and wild-type mice, measured in paraffin sections stained with Sirius red (see C) and calculated by semiautomaticmicroscopic image analysis; n � 5–7/gender and genotype. **P 0.01.

5084 Vol. 26 December 2012 HOCHRATH ET AL.The FASEB Journal � www.fasebj.org Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

stained with Sirius red. We observed proliferation ofbile ducts and prominent periportal collagen aggrega-tion. Accordingly, Fig. 1D shows that mean collagenareas are significantly (P0.01) enhanced in liversfrom Abcb4�/� mice of both genders compared withthose for wild-type animals (males: 1.67�0.22 vs.0.36�0.09%; females: 2.01�0.29 vs. 0.82�0.13%).

Glucose tolerance in Abcb4�/� mice

Figure 2A illustrates that Abcb4-knockout mice allowedaccess to food overnight display significantly (P0.001)lower plasma glucose concentrations than correspondingwild-type controls (males: 6.40�0.29 vs. 9.05�0.54 mM;females: 4.08�0.52 vs. 6.90�0.39 mM). Of note, glucoseconcentrations in ABCB4-deficient mice after overnightfood withdrawal are significantly (P0.05) higher than

those in wild-type controls (males: 4.10�0.11 vs.3.40�0.05 mM; females: 3.20�0.12 vs. 2.90�0.07 mM;Fig. 2B).

The graphic of the plasma glucose levels at differ-ent time points during IpGTT (Fig. 3A) shows thatAbcb4-knockout mice display significantly (P0.001)lower glucose peaks during IpGTT. Based on theglucose concentrations at different time points, wecalculated AUC values (Fig. 3B), demonstrating thatAbcb4�/� animals have significantly (P0.01) lowerAUC between 30 and 120 min after glucose adminis-tration. Two-way ANOVA (Supplemental Table S1)indicated that each variable (i.e., Abcb4 genotype,gender, and time) affects glucose levels independently(P0.001). For comparison, we performed IpGTT infemale Abcb4�/� mice on the FVB/NJ background,

Figure 2. Plasma glucose levels in male and female BALB-Abcb4�/� mice (solid bars) and BALB/cJ wild-type controls (openbars) in mice allowed access to food overnight (A) and after overnight food withdrawal (B). n � 10/gender and genotype.*P 0.05; ***P 0.001.

Figure 3. IpGTT in male and female BALB-Abcb4�/� mice (solid symbols) and BALB/cJ wild-type controls (open symbols).A) Plasma glucose levels before and 15, 30, 60 and 120 min after glucose challenge. Female mice are indicated by circles andmale animals by squares. ***P 0.001. B) AUC during IpGTT for the time intervals 0–30 min (open and solid bars and 30–120min (light and dark striped bars), as approximated using the trapezoidal rule excluding the area below the baseline glucose level(glucose at t0). **P 0.01.

5085ABCB4 AS MODULATOR OF GLUCOSE HOMEOSTASIS Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

which also showed significantly (P�0.01) lower AUCbetween 30 and 120 min after glucose exposurecompared with FVB wild-type controls.

Hepatic expression profiles in Abcb4�/� mice

Statistical analyses of the cDNA transcriptome profilesof liver identified a total of 85 genes differentiallyregulated between Abcb4�/� and wild-type mice (Sup-plemental Fig. S1). Among the data set of regulatedgenes, GO analysis (see Materials and Methods) re-vealed overrepresentation of several terms in the cate-gories of molecular functions and canonical pathways(Table 2). The expression profiling (Supplemental Fig.S1) highlighted genes that we and others have previ-ously identified as modifiers of chronic liver inflamma-tion and fibrosis, such as cathepsin B (Ctsb; ref. 35),carbamoyl-phosphate synthetase 1 (Cps1; ref. 36), he-molytic complement factor 5 (Hc; ref. 23), decorin(Dcn; ref. 37), glycine N-methyltransferase (Gnmt; ref.38), and lipocalin 2 (Lcn2; ref. 39). Moreover, as shownin Table 2, we identified induced expression of 11genes involved in carbohydrate metabolism.

Therefore, we analyzed by real-time PCR hepaticexpression of genes encoding key enzymes and tran-scription factors implicated in glucose homeostasis inmice allowed access to food overnight (Fig. 4A) and in

mice denied access to food overnight before and dur-ing the first 60 min of IpGTT (Fig. 4B). Figure 4A showsthat steady-state mRNA levels of Pck1, Gck, and G6pc areall markedly induced in Abcb4�/� compared with levelsin wild-type mice. Of note, the expression of thePC-activated nuclear receptor Lrh-1 (Nr5a2), the bonafide PC receptor (13), shows a 5-fold increase in Abcb4-deficient mice compared with that in controls. Inprimary hepatocytes (which, however, do not expressthe PC transporter ABCB4 properly at a canaliculardomain), we also observed the highest Lrh-1 mRNAlevels in Abcb4-deficient cells treated with DLPC (seeMaterials and Methods). However, we did not detectmajor expression differences for the gene set tested invivo (Fig. 4) under our experimental conditions (datanot shown).

We did also not detect any expression difference forthe transcription factor Srebf1, which activates the ex-pression of genes dedicated to sterol, triglyceride, andPC synthesis (40) and is repressed by PC (13), betweenAbcb4�/� and wild-type mice in vivo. In contrast, theexpression of Ppargc1a and the transcription factorFoxo1, both involved in the regulation of gluconeogen-esis (41), is induced.

Figure 4B compares the gene expression changesduring glucose challenge (IpGTT). After overnight

TABLE 2. GO terms of the categories molecular function and canonical pathways overrepresented in Abcb4�/�compared with those inwild-type mice

GO term Genes

Molecular functionAmino acid metabolism Gnmt, Ass1, Oat, DdAh1, Csad, Tat, Slc6a7, Cps1Cancer Gstp1, Ass1, Gpx1, Clu, Psen2, Orm1, Csf2ra, Anxa5, Plin2, Lsr, Ly6e,

Enpp2, Tmsb4x, Gstm1, Hpx, Dcn, Gstm3, Lcn2, Pold1, Fabp5, Eef1a1,Ctsb, Ubc, Tmem176b, Rap2a, Col3a1

Carbohydrate metabolism Gnmt, Apoa4, Fabp5, Plin2, Anxa5, Gpx1, Lpl, Eef1a1, Ctsb, Enpp2, Cps1Cardiovascular disease Gstm1, Gstp1, Apoa4, Gpx1, Lcn2, Clu, Psen2, Ttpa, Nad, Dlg5, Csf2ra,

Lsr, Anxa5, Cnst, Pros1, Ddah1, Lpl, Col3a1Cell death Gstm1, Gstp1, Nfic, Gsta1, Dcn, Lcn2, Gpx1, Clu, Psen2, Nad, G0s2,

Csf2ra, Anxa5, Ctsp, Eef1a1, Tmsb4xCell signaling Ass1, Ddah1, Eef1a1, Psen2, Cps1, Nad

Drug metabolism Gstm1, Gstp1, Gnmt, Gsta1, Gstm3, Gpx1, Ctsb, Ly6eGastrointestinal disease Dhx58, Gstm1, Gstp1, Hpx, Lcn2, Gpx1, Clu, Arnt2, Pold1, Dlg5, Csf2ra,

Cnst, Ctsb, Enpp, Cps1, Col3a1Hepatic system disease Gstm1, Hpx, Gstp1, Fabp5, Csf2ra, Plin2, Lpl, Clu, Arnt2, Pold1, Col3a1Lipid metabolism Gstm1, Hpx, Gstp1, Apoa4, Gstm3, Gsta1, Gpx1, Clu, Ttpa, Fabp5, Lsr,

Anxa5, Plin2, Lpl, Eef1a1, Enpp2, Cps1Neurological disease Gstp1, Itih3, Apoa4, Zep36l1, Clu, Psen2, Ttpa, Slc6a7, Arnt2, Orm1,

Nad, Ccf2ra, Pros1, Cnst, Anxa5, Lpl, Enpp2, Tmsb4x, Ubb, Gstm1,Hpx, Ush2a, Orm2, Arhgap10, Dcn, Lcn2, Fam49a, Dlg5, Ctsb, Eef1a1

Vitamin and mineral metabolism Gnmt, Plin2, Gsta1, Lpl, Psen2, Ttpa

Canonical pathwaysAcute phase response signaling Hpx, Itih3, Orm2, Itih2, Orm1Aryl hydrocarbon receptor signaling Gstm1, Gstp1, Nfic, Gstm3, Gsta1Glutathione metabolism Gstm1, Gstp1, Gstm3, Gsta1, Gpx1LPS/IL-1 mediated inhibition of retinoid X

receptor functionGstm1, Gstp1, Fabp5, Gstm3, Gsta1

Metabolism of xenobiotics by cytochrome P450 Gstm1, Gstp1, Gstm3, Gsta1NRF2-mediated oxidative stress response Gstm1, Ubb, Gstp1, Gstm3, Gsta1

Data of microarray experiments; genes were classified by IPA (see Materials and Methods).

5086 Vol. 26 December 2012 HOCHRATH ET AL.The FASEB Journal � www.fasebj.org Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

food withdrawal (t0), only the expression of Gck isappreciably enhanced in Abcb4�/� mice comparedwith wild-type mice. At 60 min after glucose admin-istration (t60), the expression levels of all measuredgenes, with the exception of Gck, increase. In partic-ular, the genes encoding phosphoenolpyruvate car-boxykinase 1, glucose 6-phosphatase, and LRH-1 aremarkedly induced in ABCB4-deficient mice com-pared with control mice.

To provide evidence for LRH-1 activation in Abcb4�/�

mice, we measured hepatic gene expression of LRH-1target genes (in particular Cyp8b1), the Lrh-1 corepres-sor Nr0b2 [also known as small heterodimer partner(Shp)], and the bile salt farnesoid X receptor Nr1h4(alias Fxr) in mice challenged with glucose. As pre-sented in Fig. 4C, Abcb4�/� mice demonstrate signifi-cantly (P0.05) increased expression of Cyp8b1 andCyp7b1 compared with basal levels and with those inwild-type animals. In contrast, expression levels ofCyp7a1 and Nr1h4 did not differ between Abcb4�/� andwild-type mice in this setting.

Mass spectrometry imaging experiments identifyLRH-1 agonists in Abcb4�/� liver

Imaging of liver sections using ultra-high-resolution,accurate mass MALDI-MSI proved the presence of

the phospholipids DUPC (C11:0/C11:0) and DLPC(C12:0/C12:0) in murine liver. In these imagingexperiments (Fig. 5), we were able to identify distinctdistribution patterns of these two natural ligands.Notably, the experimental mass uncertainties for the

Figure 4. Relative mRNA expression levels of genes involvedin glucose homeostasis in BALB/Abcb4�/� and BALB/cJwild-type mice. A) Abcb4�/� mice (solid bars) and wild-typemice (open bars) allowed access to food overnight. Allexpression levels are normalized to controls. Black dottedline marks the expression levels in controls at t0; n �5–12/genotype. B) Expression in Abcb4�/� mice relative towild-type mice during IpGTT at baseline (t0; solid bars) andafter 60 min (t60; hashed bars). All expression levels arenormalized to those of corresponding controls at t0. Blackdotted line marks the expression levels in controls at t0. n �3–10/time point and genotype. C) Hepatic expression levelsof LRH-1 target genes and genes involved in bile salthomeostasis in BALB-Abcb4�/� and BALB/cJ mice duringIpGTT. n � 5–9/time point and genotype. Gene symbols:Cyp7a1, cytochrome P450 family 7, subfamily A, polypeptide

1; Cyp7b1, cytochrome P450 family 7, subfamily B, polypeptide 1; Cyp8b1, cytochrome P450, family 8, subfamily B, polypeptide 1;Fasn, fatty acid synthase; Foxo1, forkhead box O1; G6pc, glucose-6-phosphatase; Gck, glucokinase; Lrh-1, liver receptor homolog 1;Nr0b2, nuclear receptor subfamily 0, group B, member 2; Nr1h4, nuclear receptor subfamily 1, group H, member 4; Pck1,phosphoenolpyruvate carboxykinase 1; Ppargc1, peroxisome proliferator-activated receptor � coactivator 1�; Srebf1, sterolregulatory element binding transcription factor 1. *P 0.05; **P 0.01; ***P 0.001.

0 1 2 3 4 5 6 7

Distance, x (mm)

Dis

tanc

e, y

(mm

)

0

1

2

25000

0

Figure 5. MALDI-MSI analysis of mouse liver section usinghigh-resolution, accurate mass FTICR-mass spectrometry.Shown is a composite image illustrating spatial distribution ofDLPC (C12:0/C12:0; red) and DUPC (C11:0/C11:0; green).Orange indicates overlapping areas of DUPC and DLPC.Intensity scale of the heatmap corresponds to color contrastunits (25,000 arbitrary units for maximum brightness; red/green); it is not calibrated for concentration units. Extracted[M H] species: DUPC, m/z 594.4129; DLPC, m/z622.4442; extraction window, m/z � 2 mDa; matrix, �-cyano-4-hydroxycinnamic acid; pixel (laser) size, 75 �m.

5087ABCB4 AS MODULATOR OF GLUCOSE HOMEOSTASIS Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

measured m/z ratios of the lipid species on the tissueslice were �0.27 ppm compared with calculated,exact m/z ratios, with no alternative formula propos-als within �2 ppm centered on the measured [M H] ions, thus providing adequate confidence forpositive identification of the two lipid species. Thesebona fide LRH-1 agonists were not previously detectedin liver.

Serum glucose levels in humans are associated with acommon ABCB4 variant

We genotyped 4 common ABCB4 variants that werepreviously shown by us and other groups to be associ-ated with hepatic phenotypes (4, 5). Consistency withHardy-Weinberg equilibrium confirmed robust geno-typing (all P�0.05), and the genotyping results were inline with the frequencies reported in the Entrez SNPdatabase and previous publications (4, 5).

Figure 6 illustrates that carriers of specific ABCB4c.711 variants display significantly (P0.05) differentserum glucose levels. The genotype [AA], [AT], and[T] variants were carried by 478, 184, and 20 individu-als, respectively. The homozygous carriers of theprocholestatic genotype [AA] present with significantlylower median glucose levels (5.33 mM, range 3.09–18.78 mM) than carriers of variants [AT] and [TT](5.47 mM, range 2.76-21.82 mM). As shown in Table 3,the linear model after backward variable selectiondemonstrated that both the ABCB4 variant and ageindependently affect serum glucose levels.

DISCUSSION

To date, genetic polymorphisms in the ABCB4 locus inhumans and the deficiency of the PC transporter inAbcb4�/� mice have been associated solely with chole-static disease. This study is first to underscore anassociation between the ABCB4 transporter and serumglucose levels both in mice and humans. Our results

show that individuals carrying an ABCB4 procholestaticrisk variant have decreased glucose levels in general,and mice lacking the ABCB4 protein present with lowerglucose levels in the fed state than wild-type controls.

In our previous studies we observed marked differ-ences in fibrosis susceptibility in inbred mouse strains,with BALB/cJ inbred mice being most susceptible,whereas FVB/NJ mice were more fibrosis-resistant (22).By introgressing the Abcb4 knockout from the resistantFVB/NJ strain into the fibrosis-susceptible BALB/cJbackground, we created a unique congenic Abcb4mouse model with a more severe liver phenotype.Although we observed similar alterations of glucosehomeostasis in both knockout lines, all subsequentexperiments were performed in BALB-Abcb4�/� mice,which showed more pronounced alterations than theFVB-Abcb4�/� line. Abcb4�/� mice demonstrate abnor-mal hepatic expression of genes involved in glycolysisand gluconeogenesis. This simultaneous induction ofboth pathways is noteworthy and is indicative for path-ological processes resulting in decreased serum glucoselevels. The expression levels of regulators of insulin-dependent glucose metabolism (e.g., Foxo1 and Srebf1)(42, 43) differ only slightly between Abcb4�/� mice andcontrols after glucose challenge; hence, lower glucoselevels do not seem to be related to insulin effects.Because hepatic expression profiles after overnightfood withdrawal do not differ between knockout andwild-type mice (except for Gck), ABCB4 deficiencyappears to confer an improvement in glucose homeo-stasis that becomes apparent in the setting of increasedsystemic glucose levels. Our findings are supported byrecent findings placing Abcb4 in an expression networkof murine genes associated with AUC of glucose inIpGTT, with Abcb4 showing a positive association withglucose intolerance (44).

Because ABCB4 represents the single PC exportpump of liver, one of the possible explanations of ourresults relates glucose levels to distorted hepatic PCmetabolism. Based on reports that PC acts as ligand forLRH-1 (45), it can be suspected that this nuclearreceptor represents a critical regulator of glucose me-tabolism in ABCB4-deficient mice. Indeed, we foundincreased expression of the Lrh-1 gene in Abcb4�/�

mice, which might be related to altered dynamics ofhepatic PC metabolism. As shown recently (13), theactivation of LRH-1 by specific PC species leads todecreased serum glucose levels in diabetic mice. In fact,although the overall composition of major hepaticphospholipids has been reported to be stable inAbcb4�/� mice (46), we were able to demonstrate the

Figure 6. Median and range of serum glucose levels inpatients with different genotypes of ABCB4 variant c.711A�T.Glucose concentrations in carriers of the genotype [AA] aresignificantly lower than in carriers of genotypes [AT] and[TT]. *P 0.05.

TABLE 3. Linear model for serum glucose levels after backwardvariable selection using the AIC

Coefficient Estimate P

ABCB4 [AA] ↔ [AT] [TT] 0.484 � 0.199 0.015Age (yr) 0.024 � 0.005 0.001BMI (kg/m2) 0.264 � 0.018 0.144

5088 Vol. 26 December 2012 HOCHRATH ET AL.The FASEB Journal � www.fasebj.org Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

presence of the bona fide LRH-1 ligands DUPC andDLPC in murine liver by MALDI-MSI. In search forLRH-1 target proteins that directly regulate glucosehomeostasis, a recent article reported that mice withconditional deletion of Lrh-1 in liver show reducedtranscriptional activation of Gck by LRH-1, resultingin reduced glycolysis and glycogen synthesis in re-sponse to glucose exposure (47). These findings arein line with our view of LRH-1 as regulator ofpostprandial glucose homeostasis. Expression analy-ses during IpGTT in our study clearly demonstratedinduction of LRH-1 targets, particularly Cyp8b1, inlivers from Abcb4�/� mice after acute glucose chal-lenge compared with control mice, whereas Cyp7a1expression remained constantly low, in line withindependent regulation of the rate-limiting step ofbile salt synthesis in the setting of high systemic bilesalt levels of cholestatic Abcb4�/� mice (48).

In humans, the composition of hepatic PC in carriersof ABCB4 variants would be difficult to predict, butbased on our findings in the mouse model a shifttoward molecular species that activate LRH-1 and in-duce the expression of genes that modify glucosemetabolism might be expected. On the other hand,given the complexity of glucose homeostasis, one can-not exclude other, non-LRH-1-mediated regulatorypathways that contribute to lower serum glucose levelsin Abcb4�/� mice. Because Abcb4�/� mice display intra-hepatic cholestasis and increased serum bile salt con-centrations (10), bile salt-dependent modulation ofglucose homeostasis via the membrane receptor TGR5could play a role (49, 50), and, in fact, our systematicphenotypic characterization of metabolic rates (O2consumption) points to increased energy expenditurein Abcb4�/� mice (data not shown), albeit additionalfunctional studies are needed. Furthermore, in ourtranslational study, none of the patients carryingABCB4 variants presented with cholestasis, indicatingthat the observed effects on glucose homeostasis arenot related primarily to altered bile salt signaling.

The ABCB4 c.711A allele, which associates with lowerserum glucose levels in our cohorts, is estimated to becarried by more than 80% of Europeans. The effect onglucose levels is underscored by the analysis that in-cluded nongenetic variables, demonstrating that thedifferences between carriers of distinct ABCB4 geno-types remain significant even after inclusion of addi-tional factors in the model (Table 3). This findingsupports the potential role of variant ABCB4 in glucosehomeostasis. According to our previous study (5), thepresence of the common c.711A allele increases therisk of cholestasis in pregnancy. For the patients fromour current study, we do not possess data on thefrequency of obstetric cholestasis, but, in general, indi-viduals carrying the risk allele c.711A do not display anyspecific hepatic phenotype, unless additional triggersinduce liver disease. Data on associations of this ABCB4polymorphism with other disease traits are limited. Inthe previous analysis (51), we demonstrated that gall-stone carriers of the common allele present with lower

total serum cholesterol levels than gallstone-free con-trols; moreover, the same variant affects HDL choles-terol levels but not serum triglycerides. Apart fromintrahepatic cholestasis of pregnancy, dysfunction ofthe hepatic PC transporter ABCB4 has been reportedto underlie other cholestatic disorders, namely progres-sive familial intrahepatic cholestasis type 3 (52), tran-sient neonatal cholestasis (53), drug-induced cholesta-sis (54), and low phospholipid-associated cholelithiasis(3, 6, 7). Nevertheless these phenotypes are confinedsolely to the liver, and metabolic traits in carriers of thec.711A risk variants are most likely not associated withhepatic dysfunction.

In summary, this study shows that genetic variation inABCB4 substantially influences glucose metabolism.Moreover, our results further support the role of phos-pholipids in the regulation of glucose metabolism. Inthis respect, Abcb4�/� mice represent an experimentalframework to further study effects of disturbed PChomeostasis on glycemia. Because future antidiabetictherapies with LRH-1 agonists may be envisioned (55),we speculate that the genotyping of ABCB4 polymor-phisms could be helpful for the stratification ofpatients.

The authors thank all patients for participating in this studyand providing blood samples. The authors also thank thetechnicians, especially Annika Bohner for setting up cell cultureexperiments, and Sandra Geissler, Elfi Holupirek, Ann-ElisabethSchwarz, and Reinhard Seeliger, as well as the animal caretakerteam of the German Mouse Clinic (Michael Gerstlauer, Manu-ela Huber, Renate Huber, Heidi Marr, Annica Miedl, TinaReichelt, Boris Schön, and Horst Wenig), for technical assis-tance in mouse phenotyping. The authors also thank ChristineGau for help with statistical analyses. This study was supported byDeutsche Forschungsgemeinschaft (SFB/TRR57 TP01 to F.L.),by a grant from the German Federal Ministry of Education andResearch to the German Center for Diabetes Research, NGFN-Plus (01GS0851 to E.W., 01GS0850 to M.H.A., and 01GS0869 toM.K.), and Infrafrontier (01KX1012). The German MouseClinic received funding from the European Union (EUMODICLSHG-2006-037188 and Infrafrontier Contract No. 211404).D.A.V. acknowledges research support by the Alfried Krupp vonBohlen and Halbach-Stiftung. The results of this article wereincluded, in part, in an oral presentation at the 60th AnnualMeeting of the American Association for the Study of LiverDiseases in Boston, November 3, 2009, and were published inabstract form in Hepatology [2009; 50(Suppl. 4), A232].

REFERENCES

1. Carey, M. C., and LaMont, J. T. (1992) Cholesterol gallstoneformation. 1. Physical-chemistry of bile and biliary lipid secre-tion. Prog. Liver Dis. 10, 139–163

2. Oude Elferink, R. P., Paulusma, C. C., and Groen, A. K. (2006)Hepatocanalicular transport defects: pathophysiologic mecha-nisms of rare diseases. Gastroenterology 130, 908–925

3. Krawczyk, M., Wang, D. Q., Portincasa, P., and Lammert, F.(2011) Dissecting the genetic heterogeneity of gallbladder stoneformation. Semin. Liver Dis. 31, 157–172

4. Müllenbach, R., Linton, K. J., Wiltshire, S., Weerasekera, N.,Chambers, J., Elias, E., Higgins, C. F., Johnston, D. G., McCar-thy, M. I., and Williamson, C. (2003) ABCB4 gene sequencevariation in women with intrahepatic cholestasis of pregnancy. J.Med. Genet. 40, e70

5089ABCB4 AS MODULATOR OF GLUCOSE HOMEOSTASIS Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

5. Wasmuth, H. E., Glantz, A., Keppeler, H., Simon, E., Bartz, C.,Rath, W., Mattsson, L. A., Marschall, H. U., and Lammert, F.(2007) Intrahepatic cholestasis of pregnancy: the severe form isassociated with common variants of the hepatobiliary phospho-lipid transporter ABCB4 gene. Gut 56, 265–270

6. Rosmorduc, O., Hermelin, B., and Poupon, R. (2001) MDR3gene defect in adults with symptomatic intrahepatic and gall-bladder cholesterol cholelithiasis. Gastroenterology 120, 1459–1467

7. Rosmorduc, O., Hermelin, B., Boelle, P. Y., Parc, R., Taboury, J.,and Poupon, R. (2003) ABCB4 gene mutation-associated chole-lithiasis in adults. Gastroenterology 125, 452–459

8. Smit, J. J., Schinkel, A. H., Oude Elferink, R. P., Groen, A. K.,Wagenaar, E., van Deemter, L., Mol, C. A., Ottenhoff, R., vander Lugt, N. M., and van Roon, M. A. (1993) Homozygousdisruption of the murine mdr2 P-glycoprotein gene leads to acomplete absence of phospholipid from bile and to liver disease.Cell 75, 451–462

9. Lammert, F., Wang, D. Q., Hillebrandt, S., Geier, A., Fickert, P.,Trauner, M., Matern, S., Paigen, B., and Carey, M. C. (2004)Spontaneous cholecysto- and hepatolithiasis in Mdr2�/� mice: amodel for low phospholipid-associated cholelithiasis. Hepatology39, 117–128

10. Fickert, P., Fuchsbichler, A., Wagner, M., Zollner, G., Kaser, A.,Tilg, H., Krause, R., Lammert, F., Langner, C., Zatloukal, K.,Marschall, H. U., Denk, H., and Trauner, M. (2004) Regurgita-tion of bile acids from leaky bile ducts causes sclerosing cholan-gitis in Mdr2 (Abcb4) knockout mice. Gastroenterology 127,261–274

11. Vacca, M., Krawczyk, M., Petruzzelli, M., Sasso, R. C., vanErpecum, K. J., Palasciano, G., van Berge-Henegouwen, G. P.,Moschetta, A., and Portincasa, P. (2007) Current treatments ofprimary sclerosing cholangitis. Curr. Med. Chem. 14, 2081–2094

12. Forman, B. M. (2005) Are those phospholipids in your pocket?Cell Metab. 1, 153–155

13. Lee, J. M., Lee, Y. K., Mamrosh, J. L., Busby, S. A., Griffin, P. R.,Pathak, M. C., Ortlund, E. A., and Moore, D. D. (2011) Anuclear-receptor-dependent phosphatidylcholine pathway withantidiabetic effects. Nature 474, 506–510

14. Fayard, E., Auwerx, J., and Schoonjans, K. (2004) LRH-1: anorphan nuclear receptor involved in development, metabolismand steroidogenesis. Trends Cell Biol. 14, 250–260

15. Roderick, S. L., Chan, W. W., Agate, D. S., Olsen, L. R., Vetting,M. W., Rajashankar, K. R., and Cohen, D. E. (2002) Structure ofhuman phosphatidylcholine transfer protein in complex with itsligand. Nat. Struct. Biol. 9, 507–511

16. Shishova, E. Y., Stoll, J. M., Ersoy, B. A., Shrestha, S., Scapa, E. F.,Li, Y., Niepel, M. W., Su, Y., Jelicks, L. A., Stahl, G. L., Glicksman,M. A., Gutierrez-Juarez, R., Cuny, G. D., and Cohen, D. E.(2011) Genetic ablation or chemical inhibition of phosphatidyl-choline transfer protein attenuates diet-induced hepatic glucoseproduction. Hepatology 54, 664–674

17. Popov, Y., Patsenker, E., Fickert, P., Trauner, M., and Schuppan,D. (2005) Mdr2 (Abcb4)�/� mice spontaneously develop severebiliary fibrosis via massive dysregulation of pro- and antifibro-genic genes. J. Hepatol. 43, 1045–1054

18. Iyengar, S. K., Abboud, H. E., Goddard, K. A., Saad, M. F., Adler,S. G., Arar, N. H., Bowden, D. W., Duggirala, R., Elston, R. C.,Hanson, R. L., Ipp, E., Kao, W. H., Kimmel, P. L., Klag, M. J.,Knowler, W. C., Meoni, L. A., Nelson, R. G., Nicholas, S. B.,Pahl, M. V., Parekh, R. S., Quade, S. R., Rich, S. S., Rotter, J. I.,Scavini, M., Schelling, J. R., Sedor, J. R., Sehgal, A. R., Shah,V. O., Smith, M. W., Taylor, K. D., Winkler, C. A., Zager, P. G.,and Freedman, B. I. (2007) Genome-wide scans for diabeticnephropathy and albuminuria in multiethnic populations: thefamily investigation of nephropathy and diabetes (FIND). Dia-betes 56, 1577–1585

19. Prochazka, M., Mochizuki, H., Baier, L. J., Cohen, P. T., andBogardus, C. (1995) Molecular and linkage analysis of type-1protein phosphatase catalytic beta-subunit gene: lack of evi-dence for its major role in insulin resistance in Pima Indians.Diabetologia 38, 461–466

20. Mochizuki, H., Scherer, S. W., Xi, T., Nickle, D. C., Majer, M.,Huizenga, J. J., Tsui, L. C., and Prochazka, M. (1998) HumanPON2 gene at 7q21.3: cloning, multiple mRNA forms, andmissense polymorphisms in the coding sequence. Gene 213,149–157

21. Gailus-Durner, V., Fuchs, H., Becker, L., Bolle, I., Brielmeier,M., Calzada-Wack, J., Elvert, R., Ehrhardt, N., Dalke, C., Franz,T. J., Grundner-Culemann, E., Hammelbacher, S., Hölter, S. M.,Hölzlwimmer, G., Horsch, M., Javaheri, A., Kalaydjiev, S. V.,Klempt, M., Kling, E., Kunder, S., Lengger, C., Lisse, T., Mijalski,T., Naton, B., Pedersen, V., Prehn, C., Przemeck, G., Racz, I.,Reinhard, C., Reitmeir, P., Schneider, I., Schrewe, A.,Steinkamp, R., Zybill, C., Adamski, J., Beckers, J., Behrendt, H.,Favor, J., Graw, J., Heldmaier, G., Höfler, H., Ivandic, B., Katus,H., Kirchhof, P., Klingenspor, M., Klopstock, T., Lengeling, A.,Muller, W., Ohl, F., Ollert, M., Quintanilla-Martinez, L.,Schmidt, J., Schulz, H., Wolf, E., Wurst, W., Zimmer, A., Busch,D. H., and Hrabe de Angelis, M. (2005) Introducing theGerman Mouse Clinic: open access platform for standardizedphenotyping. Nat. Methods 2, 403–404

22. Hillebrandt, S., Goos, C., Matern, S., and Lammert, F. (2002)Genome-wide analysis of hepatic fibrosis in inbred mice identi-fies the susceptibility locus Hfib1 on chromosome 15. Gastroen-terology 123, 2041–2051

23. Hillebrandt, S., Wasmuth, H. E., Weiskirchen, R., Hellerbrand,C., Keppeler, H., Werth, A., Schirin-Sokhan, R., Wilkens, G.,Geier, A., Lorenzen, J., Kohl, J., Gressner, A. M., Matern, S., andLammert, F. (2005) Complement factor 5 is a quantitative traitgene that modifies liver fibrogenesis in mice and humans. Nat.Genet. 37, 835–843

24. Jamall, I. S., Finelli, V. N., and Que Hee, S. S. (1981) A simplemethod to determine nanogram levels of 4-hdydroxyproline inbiological tissues. Anal. Biochem. 112, 70–75

25. Hegde, P., Qi, R., Abernathy, K., Gay, C., Dharap, S., Gaspard,R., Hughes, J. E., Snesrud, E., Lee, N., and Quackenbush, J.(2000) A concise guide to cDNA microarray analysis. BioTech-niques 29, 548–550, 552–544

26. Horsch, M., Schädler, S., Gailus-Durner, V., Fuchs, H., Meyer,H., Hrabe de Angelis, M., and Beckers, J. (2008) Systematicgene expression profiling of mouse model series reveals coex-pressed genes. Proteomics 8, 1248–1256

27. Quackenbush, J. (2002) Microarray data normalization andtransformation. Nat. Genet. 32(Suppl.), 496–501

28. Tusher, V. G., Tibshirani, R., and Chu, G. (2001) Significanceanalysis of microarrays applied to the ionizing radiation re-sponse. Proc. Natl. Acad. Sci. U. S. A. 98, 5116–5121

29. Livak, K. J., and Schmittgen, T. D. (2001) Analysis of relativegene expression data using real-time quantitative PCR and the2���CT method. Methods 25, 402–408

30. Dooley, S., Hamzavi, J., Ciuclan, L., Godoy, P., Ilkavets, I.,Ehnert, S., Ueberham, E., Gebhardt, R., Kanzler, S., Geier, A.,Breitkopf, K., Weng, H., and Mertens, P. R. (2008) Hepatocyte-specific Smad7 expression attenuates TGF-�-mediated fibrogen-esis and protects against liver damage. Gastroenterology 135,642–659

31. Kovacs, P., Kress, R., Rocha, J., Kurtz, U., Miquel, J. F., Nervi, F.,Méndez-Sánchez, N., Uribe, M., Bock, H. H., Schirin-Sokhan,R., Stumvoll, M., Mössner, J., Lammert, F., and Wittenburg, H.(2008) Variation of the gene encoding the nuclear bile saltreceptor FXR and gallstone susceptibility in mice and humans.J. Hepatol. 48, 116–124

32. Grünhage, F., Acalovschi, M., Tirziu, S., Walier, M., Wienker,T. F., Ciocan, A., Mosteanu, O., Sauerbruch, T., and Lammert,F. (2007) Increased gallstone risk in humans conferred bycommon variant of hepatic ATP-binding cassette transporter forcholesterol. Hepatology 46, 793–801

33. Krawczyk, M., Rusticeanu, M., Grünhage, F., Mahler, M., Port-incasa, P., Acalovschi, M., and Lammert, F. (2010) Analysis ofthe common vasoactive intestinal peptide receptor 1 polymor-phism in gallstone patients. J. Gastrointestin. Liver Dis. 19, 273–277

34. Krawczyk, M., Grünhage, F., Mahler, M., Tirziu, S., Acalovschi,M., and Lammert, F. (2011) The common adiponutrin variantp.I148M does not confer gallstone risk but affects fastingglucose and triglyceride levels. J. Physiol. Pharmacol. 62, 369–375

35. Moles, A., Tarrats, N., Fernandez-Checa, J. C., and Mari, M.(2011) Cathepsin B overexpression due to acid sphingomyeli-nase ablation promotes liver fibrosis in Niemann Pick disease. J.Biol. Chem. 287, 1178–1188

36. Perez-Carreon, J. I., Martinez-Perez, L., Loredo, M. L., Yanez-Maldonado, L., Velasco-Loyden, G., Vidrio-Gomez, S., Ramirez-Salcedo, J., Hernandez-Luis, F., Velazquez-Martinez, I., Suarez-

5090 Vol. 26 December 2012 HOCHRATH ET AL.The FASEB Journal � www.fasebj.org Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

Cuenca, J. A., Hernandez-Munoz, R., and de Sanchez, V. C.(2010) An adenosine derivative compound, IFC305, reversesfibrosis and alters gene expression in a pre-established CCl4-induced rat cirrhosis. Int. J. Biochem. Cell Biol. 42, 287–296

37. Baghy, K., Dezso, K., Laszlo, V., Fullar, A., Peterfia, B., Paku, S.,Nagy, P., Schaff, Z., Iozzo, R. V., and Kovalszky, I. (2011)Ablation of the decorin gene enhances experimental hepaticfibrosis and impairs hepatic healing in mice. Lab. Invest. 91,439–451

38. Varela-Rey, M., Martinez-Lopez, N., Fernandez-Ramos, D., Em-bade, N., Calvisi, D. F., Woodhoo, A., Rodriguez, J., Fraga, M. F.,Julve, J., Rodriguez-Millan, E., Frades, I., Torres, L., Luka, Z.,Wagner, C., Esteller, M., Lu, S. C., Martinez-Chantar, M. L., andMato, J. M. (2010) Fatty liver and fibrosis in glycine N-methyl-transferase knockout mice is prevented by nicotinamide. Hepa-tology 52, 105–114

39. Borkham-Kamphorst, E., Drews, F., and Weiskirchen, R. (2011)Induction of lipocalin-2 expression in acute and chronic exper-imental liver injury moderated by pro-inflammatory cytokinesinterleukin-1� through nuclear factor- B activation. Liver Int.31, 656–665

40. Horton, J. D., Goldstein, J. L., and Brown, M. S. (2002) SREBPs:activators of the complete program of cholesterol and fatty acidsynthesis in the liver. J. Clin. Invest. 109, 1125–1131

41. Müllenbach, R., Weber, N. S., and Lammert, F. (2011) Nuclearreceptor variants in liver disease. J. Lipids 2012, ID934707

42. Biddinger, S. B., Haas, J. T., Yu, B. B., Bezy, O., Jing, E., Zhang,W., Unterman, T. G., Carey, M. C., and Kahn, C. R. (2008)Hepatic insulin resistance directly promotes formation of cho-lesterol gallstones. Nat. Med. 14, 778–782

43. Scapa, E. F., Pocai, A., Wu, M. K., Gutierrez-Juarez, R., Glenz, L.,Kanno, K., Li, H., Biddinger, S., Jelicks, L. A., Rossetti, L., andCohen, D. E. (2008) Regulation of energy substrate utilizationand hepatic insulin sensitivity by phosphatidylcholine transferprotein/StarD2. FASEB J. 22, 2579–2590

44. Wopereis, S., Radonjic, M., Rubingh, C., van Erk, M., Smilde, A.,van Duyvenvoorde, W., Cnubben, N., Kooistra, T., van Ommen,B., and Kleemann, R. (2012) Identification of prognostic anddiagnostic biomarkers of glucose intolerance in ApoE3Leidenmice. Physiol. Genomics 44, 293–304

45. Ortlund, E. A., Lee, Y., Solomon, I. H., Hager, J. M., Safi, R.,Choi, Y., Guan, Z., Tripathy, A., Raetz, C. R., McDonnell, D. P.,Moore, D. D., and Redinbo, M. R. (2005) Modulation of humannuclear receptor LRH-1 activity by phospholipids and SHP. Nat.Struct. Mol. Biol. 12, 357–363

46. Baghdasaryan, A., Fickert, P., Fuchsbichler, A., Silbert, D.,Gumhold, J., Horl, G., Langner, C., Moustafa, T., Halilbasic, E.,

Claudel, T., and Trauner, M. (2008) Role of hepatic phospho-lipids in development of liver injury in Mdr2 (Abcb4) knockoutmice. Liver Int. 28, 948–958

47. Oosterveer, M. H., Mataki, C., Yamamoto, H., Harach, T.,Moullan, N., van Dijk, T. H., Ayuso, E., Bosch, F., Postic, C.,Groen, A. K., Auwerx, J., and Schoonjans, K. (2012) LRH-1-dependent glucose sensing determines intermediary metabo-lism in liver. J. Clin. Invest. 122, 2817–2826

48. Lee, Y. K., Schmidt, D. R., Cummins, C. L., Choi, M., Peng, L.,Zhang, Y., Goodwin, B., Hammer, R. E., Mangelsdorf, D. J., andKliewer, S. A. (2008) Liver receptor homolog-1 regulates bileacid homeostasis but is not essential for feedback regulation ofbile acid synthesis. Mol. Endocrinol. 22, 1345–1356

49. Keitel, V. (2012) Bile acids as extrahepatic and interorgansignaling molecules. In Hepatobiliary Transport in Health andDisease (Häussinger, D., Keitel, V., and Kubitz, R., eds) pp.117–129, De Gruyter, Berlin

50. Watanabe, M., Houten, S. M., Mataki, C., Christoffolete, M. A.,Kim, B. W., Sato, H., Messaddeq, N., Harney, J. W., Ezaki, O.,Kodama, T., Schoonjans, K., Bianco, A. C., and Auwerx, J.(2006) Bile acids induce energy expenditure by promotingintracellular thyroid hormone activation. Nature 439, 484–489

51. Acalovschi, M., Tirziu, S., Chiorean, E., Krawczyk, M., Grün-hage, F., and Lammert, F. (2009) Common variants of ABCB4and ABCB11 and plasma lipid levels: a study in sib pairs withgallstones, and controls. Lipids 44, 521–526

52. De Vree, J. M., Jacquemin, E., Sturm, E., Cresteil, D., Bosma,P. J., Aten, J., Deleuze, J. F., Desrochers, M., Burdelski, M.,Bernard, O., Oude Elferink, R. P., and Hadchouel, M. (1998)Mutations in the MDR3 gene cause progressive familial intrahe-patic cholestasis. Proc. Natl. Acad. Sci. U. S. A. 95, 282–287

53. Jung, C., Driancourt, C., Baussan, C., Zater, M., Hadchouel, M.,Meunier-Rotival, M., Guiochon-Mantel, A., and Jacquemin, E.(2007) Prenatal molecular diagnosis of inherited cholestaticdiseases. J. Pediatr. Gastroenterol. Nutr. 44, 453–458

54. Lang, C., Meier, Y., Stieger, B., Beuers, U., Lang, T., Kerb, R.,Kullak-Ublick, G. A., Meier, P. J., and Pauli-Magnus, C. (2007)Mutations and polymorphisms in the bile salt export pump andthe multidrug resistance protein 3 associated with drug-inducedliver injury. Pharmacogenet. Genomics 17, 47–60

55. Hohenester, S., and Beuers, U. (2011) Phosphatidylcholines asregulators of glucose and lipid homeostasis: promises andpotential risks. Hepatology 54, 2266–2268

Received for publication May 29, 2012.Accepted for publication August 27, 2012.

5091ABCB4 AS MODULATOR OF GLUCOSE HOMEOSTASIS Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

10.1096/fj.12-209379Access the most recent version at doi:2012 26: 5081-5091 originally published online September 14, 2012FASEB J 

  Katrin Hochrath, Marcin Krawczyk, Reinhild Goebel, et al.   modulator of glucose homeostasisThe hepatic phosphatidylcholine transporter ABCB4 as

  Material

Supplemental 

http://www.fasebj.org/content/suppl/2012/09/28/fj.12-209379.DC1.html

  References

  http://www.fasebj.org/content/26/12/5081.full.html#ref-list-1

This article cites 54 articles, 7 of which can be accessed free at:

Subscriptions

  http://www.faseb.org/The-FASEB-Journal/Librarian-s-Resources.aspx

is online at The FASEB JournalInformation about subscribing to

Permissions

  http://www.fasebj.org/site/misc/copyright.xhtmlSubmit copyright permission requests at:

Email Alerts

  http://www.fasebj.org/cgi/alertsReceive free email alerts when new an article cites this article - sign up at

© FASEB

 Vol.26,  No.12 , pp:5081-5091, August, 2016The FASEB Journal. 54.147.243.86 to IP www.fasebj.orgDownloaded from

Supplementary  figure  1