Toxicology and Applied Pharmacology 232 (2008) 327–336

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology

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

Role of vitamin C transporters and biliverdin reductase in the dual pro-oxidant andanti-oxidant effect of biliary compounds on the placental-fetal unit in cholestasisduring pregnancy

Maria J. Perez a, Beatriz Castaño b, Silvia Jimenez b, Maria A. Serrano b,Jose M. Gonzalez-Buitrago a, Jose J.G. Marin b,⁎a Research Unit, University Hospital, Salamanca, Spainb Laboratory of Experimental Hepatology and Drug Targeting, CIBERehd, University of Salamanca, 37007 Salamanca, Spain

Abbreviations: BR, bilirubin; BVRa, biliverdin-IXα redDCFH-DA, 2,7-dichlorofluorescein diacetate; DMSO,glyceraldehyde-3-phosphate dehydrogenase; GCDCA,ICP, intrahepatic cholestasis of pregnancy; MMP, mitocOCP, obstructive cholestasis during pregnancy; DCF, oxidreactive oxygen species; DCF-H, reduced dichlorofluoresvitamin C transport protein; TCA, taurocholic acid; tBOUDCA, ursodeoxycholic acid.⁎ Corresponding author. Department of Physiology

Miguel de Unamuno E.I.D. S-09, 37007 Salamanca, SpainE-mail address: jjgmarin@usal.es (J.J.G. Marin).

0041-008X/$ – see front matter © 2008 Elsevier Inc. Aldoi:10.1016/j.taap.2008.07.013

a b s t r a c t

a r t i c l e i n f o

Article history:

Maternal cholestasis causes Received 11 June 2008Revised 4 July 2008Accepted 8 July 2008Available online 23 July 2008

Keywords:Bile acidBiliverdinMitochondriaOxidative stressPregnancyReactive oxygen speciesVitamin C

oxidative damage to the placental–fetal unit that may challenge the outcome ofpregnancy. This has been associated with the accumulation of biliary compounds able to induce oxidativestress. However, other cholephilic compounds such as ursodeoxycholic acid (UDCA) and bilirubin have directanti-oxidant properties. In the present study we investigated whether these compounds exert a protectiveeffect on cholestasis-induced oxidative stress in placenta as compared to maternal and fetal livers, andwhether this is due in part to the activation of anti-oxidant mechanisms involving vitamin C uptake andbiliverdin/bilirubin recycling. In human placenta (JAr) and liver (HepG2) cells, deoxycholic acid (DCA) similarrates of free radical generation. In JAr (not HepG2), the mitochondrial membrane potential and cell viabilitywere impaired by low DCA concentrations; this was partly prevented by bilirubin and UDCA. In HepG2,taurocholic acid (TCA) and UDCA up-regulated biliverdin-IXα reductase (BVRα) and the vitamin Ctransporter SVCT2 (not SVCT1), whereas bilirubin up-regulated both SVCT1 and SVCT2. In JAr, TCA andUDCA up-regulated BVRα, SVCT1 and SVCT2, whereas bilirubin up-regulated only SVCT2. A differentialresponse to these compounds of nuclear receptor expression (SXR, CAR, FXR and SHP) was found in both celltypes. When cholestasis was induced in pregnant rats, BVRα, SVCT1 and SVCT2 expression in maternal andfetal livers was stimulated, and this was further enhanced by UDCA treatment. In placenta, only BVRα wasup-regulated. In conclusion, bilirubin accumulation and UDCA administration may directly and indirectlyprotect the placental–fetal unit from maternal cholestasis-induced oxidative stress.

© 2008 Elsevier Inc. All rights reserved.

Introduction

Intrahepatic cholestasis of pregnancy (ICP) is usually diagnosedduring the second, andmore often third, trimester of pregnancy and ischaracterized by the accumulation of bile acids (BAs) in maternalserum. ICP usually implies a benign condition for the mother, but itmay have serious repercussions on the conceptus, including increasedfetal distress, premature delivery, and perinatal mortality and

uctase; DCA, deoxycholic acid;dimethyl sulfoxide; GAPDH,glycochenodeoxycholic acid;hondrial membrane potential;ized dichlorofluorescein; ROS,cein; SVCT, sodium-dependentOH, tert-butyl hydroperoxide;

and Pharmacology, Campus. Fax: +34 923 294669.

l rights reserved.

morbidity (Lammert et al., 2003). Under physiological circumstances,the placenta plays a crucial role in protecting the fetus from theadverse effects of potentially toxic endogenous substances, includingBAs (Marin et al., 2003) and xenobiotics that, for different reasons, canreach the maternal circulation (Marin et al., 2004). However, thesesubstances can affect the placenta itself. Thus, alterations in thestructure and function of this organ have been found in animalmodelsof maternal hypercholanemia induced in rats by complete obstructivecholestasis during the last week of pregnancy (OCP), which in thisspecies represents one third of the gestational period. These changesinclude trophoblast atrophy together with a reduced ability to carryout vectorial transfer of BAs (Macias et al., 2000). The latter wasfavored by the creation of inversely directed gradients as comparedwith the physiological situation (Monte et al., 1995).

The cytotoxicity associated with the most hydrophobic BAs in theliver iswell-known (Attili et al.,1986). In hepatocytes,where BAs becomeaccumulated during cholestatic diseases, BAs have been found to disruptcellmembranes through their detergent action (Billington et al.,1980), topromote the generation of reactive oxygen species (ROS) and, in turn, to

328 M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

oxidatively modify lipids, proteins and nucleic acids, and eventuallycause apoptosis (Sokol et al., 2001). BAs are able to trigger apoptosis inhepatocytes by activating Fas death receptors directly (Faubion et al.,1999) or bymitochondrial dysfunction inducedby internal stress, suchasthe generation of ROS (Rodrigues et al., 1998). Additionally, they canactivate Kupffer cells to generate ROS that may contribute to liver tissueinjury (Ljubuncic et al., 1996). An interesting question that has recentlybeen addressed is how the accumulation of BAs in themother can affectfetal organs. We have recently reported that OCP induces oxidativedamage and activation of themitochondrial pathway of apoptosis in therat fetal liver (Perez et al., 2005) and placenta (Perez et al., 2006). Totalglutathione levels and the activities of anti-oxidant enzymes, such asglutathione peroxidase, glutathione-S-transferase and catalase werereduced in this situation (Perez et al., 2006, 2005).However,weobservedamoderate increase in the expression of the sodium-dependent vitaminC transport proteins (SVCT), responsible for the cellular uptake ofascorbic acid or vitamin C (Perez et al., 2007), which suggested theactivation of protective mechanisms.

Bilirubinwas traditionally consideredmerely as awaste compoundand is believed to behave as a toxic compound for the fetus, mainly asregards the developing nervous system. Indeed, neonatal hyperbilir-ubinemia can lead to brain damage (“kernicterus”) (Gourley, 1997).However, at low concentrations, a major role of this biliary pigment isto participate in the bilirubin–biliverdin anti-oxidant cytoprotectantcycle, which is maintained by biliverdin reductase (BVR) activity(Sedlak and Snyder, 2004). This function is mainly dependent on theexpression of BVRα, which has been found to be moderately up-regulated in the maternal liver–placenta-fetal liver trio during OCP(Perez et al., 2007). When biliary secretion is markedly impaired orabsent, bilirubin accumulates in plasma and tissues (Muraca et al.,1988; Van Hootegem et al., 1985), and since it has been suggested thatthe accumulation of bilirubin in cholestatic livers may limit BA-induced oxidative liver injury (Granato et al., 2003) the followinginteresting questions arose: i) Might the accumulation of bilirubin inthe mother during cholestasis of pregnancy play a protective anti-oxidant role for the conceptus? ii) Towhat extent does this occur? iii) Ifthis effect really exists, is it due only to the direct effect of bilirubin orare other indirect mechanisms also involved?

Treatment with ursodeoxycholic acid (UDCA), which has beneficialeffects in several cholestatic liver diseases (Paumgartner and Beuers,2002), is able to reduce maternal pruritus, restore biochemicalparameters to normal levels, including serum bilirubin and transami-nases, and decrease the risk of premature delivery (Lammert et al., 2003).In previous studies carried out by our group, treatment of ICP patientswith UDCAwas found to have a beneficial effect on the ability of humantrophoblasts to transport BAs (Serrano et al.,1998).Moreover, in ratswithOCP, UDCA treatment partially prevented OCP-induced trophoblastatrophy and impairment in the excretory function of the placenta(Serrano et al., 2003). A UDCA-induced protective effect in rat placentaand fetal liver has been reported to be partly due to its anti-oxidantproperties, preventing decreases in glutathione levels and oxidativedamage, and hindering the activation of apoptosis (Perez et al., 2005,2006). However, whether other indirect mechanisms, such as the up-regulation of BVR and vitamin C transporters are also involved is notknown. Accordingly, these aspects were also investigated.

Compounds such as BAs and bilirubin can activate several nuclearreceptors and transcription factors, including farnesoid X receptor(FXR), the small heterodimer partner (SHP), the pregnane X receptor(PXR) and the constitutive androstane receptor (CAR) which partici-pate in regulating the expression of several proteins involved in themetabolism and transport of several biliary compounds (Handschinand Meyer, 2005). SXR is the human orthologue of rodent PXR.Although these proteins are predominantly expressed in liver, it hasrecently been found that they are also expressed in trophoblast cellsisolated from human term placentas and the choriocarcinoma celllines BeWo, Jeg-3 and JAr (Serrano et al., 2007). FXR, PXR, and CAR

protect against hepatic BA toxicity in a complementary manner,suggesting that they serve as a partially redundant defense mechan-ism to prevent BA-induced hepatic damage during cholestasis (Guo etal., 2003). We also investigated whether the expression of thesenuclear receptors, which may be involved in the control of indirectprotective mechanisms activated by bilirubin and UDCA in thematernal liver–placenta-fetal liver trio, was modified.

In summary, the hypothesis of the present study was that certainbiliary compounds accumulated during maternal cholestasis, such asbilirubin, orused to treat this disease, suchasUDCA,mayexert aprotectiveeffect on cholestasis-induced oxidative stress in the conceptus and thatthis may be due in part to the activation of anti-oxidant mechanismsinvolving vitamin C uptake and biliverdin/bilirubin recycling.

Materials and methods

Animals, hepatocytes and cell lines. Pregnant Wistar CF rats(University of Salamanca, Spain)were used, following the experimentalprotocols approved by the Local Ethical Committee for the Use ofLaboratory Animals. Rats were anesthetized on day 14 of pregnancywith isoflurane and a sham operation (Control group) or completebiliary obstruction (OCP group) was performed (Monte et al., 1996). Inbrief, using a non-absorbable suture, a double ligation of the commonbile duct (separated by 2 mm) was carried out. The common bile ductwas surgically divided between the ligatures. During the followingweek, some of these animals (OCP+UDCA group) received dailyintragastric administration of UDCA (Sigma-Aldrich, Madrid, Spain)(60 µg/100 g b.wt.). The control and OCP groups received vehicle(50 mM NaCl, 50 mM Na2CO3, pH 8.3) alone daily. On day 21 ofpregnancy, the animals were euthanized under sodium pentobarbitalanesthesia and liver samples from the mothers and fetuses as well asplacenta samples were collected.

Male Wistar rats were used as donors to obtain isolatedhepatocytes. Rat hepatocytes were isolated by a two-step collagenaseperfusion, and were cultured in supplemented Williams' medium E(Sigma-Aldrich) (Martinez-Diez et al., 2000).

As model for placenta and liver cells, the human choriocarcinomaJAr (HTB-144) and human hepatoblastoma HepG2 (HB-8065) cell lineswere used. They were obtained from the American Type CultureCollection (ATCC, Rockville, MD, USA). JAr were cultured in RPMI 1640(Sigma-Aldrich) and HepG2 in MEM (Sigma-Aldrich), both containing10% FCS (TDI S.A., Madrid, Spain) and 1% antibiotic–antimycoticsolution (Sigma-Aldrich), and other supplements as reported by thesupplier, in a humidified atmosphere in 5% CO2 at 37 °C.

To carry out the experiments, the cells were plated and incubatedovernight in the absence of any of the tested compounds and thentreated with different concentrations of deoxycholic acid (DCA),taurocholic acid (TCA), UDCA, bilirubin or tert-butyl hydroperoxide(tBOOH), a xenobiotic compound commonly used in in vitro experi-ments as pro-oxidant agent. This compound enhances ROS produc-tion, depletes glutathione levels, and induces lipid peroxidation in rathepatocytes (Martin et al., 2001). Bilirubin, TCA and UDCA were firstdissolved in dimethyl sulfoxide (DMSO) and then added to culturemedium at final DMSO concentration of 0.5% unless otherwise stated.For flow cytometry and gene expression studies, cells were detachedwith trypsin-EDTA solution and then pelleted by centrifugation at250 ×g for 10 min, washed once with PBS, pelleted again, andresuspended in the appropriate medium.

Cell viability, ROS generation and mitochondrial membrane potential.Cell viability was evaluated using the Neutral Red (Sigma-Aldrich)test (Fautz et al., 1991). ROS generation was measured spectro-fluorometrically, using 2.5 µg/ml of the ROS-detecting probe, 2,7-dichlorofluorescein diacetate (DCFH-DA) (Sigma-Aldrich) on aMicroplate Fluorescence Reader (Fluoroskan Ascent FL, Thermo,Electron Corporation, Finland) in adherent cells (Rosenkranz et al.,

329M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

1992). Measurement of ROS production is based on the change ofreduced dichlorofluorescein (DCF-H) into oxidized dichlorofluorescein(DCF) over 48 h incubation. DCF-H is produced inside the cells byesterase-mediated cleavage of DCFH-DA, which is taken up by simplediffusion.

Flow cytometry in a FACSort flow cytometer (BD Biosciences, SanJose, CA) was used to measure ROS generation, mitochondrialmembrane potential (MMP) and cell viability. Cells were seeded intosix-well plates and treated as described above for each experimentaldesign, followed by incubation with medium containing 5 µg/mlrhodamine 123 (Rh123, Sigma-Aldrich) (MMPmeasurement) or 2.5 µg/ml DCFH-DA (ROSmeasurement) for 1 h, trypsinized, and resuspendedin FBS free-medium. Propidium iodide (PI, Sigma-Aldrich) (5 µg/ml)was added 5min before fluorescencewasmeasured by flow cytometry,as described previously (Perez and Cederbaum, 2002).

Measurement of gene expression levels. To determine mRNA levelsby real-time RT-PCR, total RNA was isolated from cell lysates or rattissue samples (placenta or liver homogenates) using RNAeasy spincolumns from Qiagen (Izasa, Barcelona, Spain). After treatment withRNase-free DNase I (Roche, Barcelona, Spain), RNA was quantifiedfluorimetrically with the RiboGreen RNA-Quantitation kit (MolecularProbes, Leiden, The Netherlands). The cDNA was synthesized from2 µg of total RNA using random hexamers and avian myeloblastosisvirus RT (Cloned AMV First-Strand cDNA Synthesis kit, Invitrogen,Barcelona, Spain). Real-time quantitative PCR was then performedusing AmpliTaq Gold polymerase (Applied Biosystems, Madrid, Spain)in an ABI Prism 5700 Sequence Detection System (AppliedBiosystems). The thermal cycling conditions were as follows: a singlecycle at 95 °C for 10 min followed by 45 cycles at 95 °C for 15 s and at60 °C for 60 s. Detection of amplification products was carried outusing SYBR Green I. Non-specific products of PCR, as examined by 2.5%agarose gel electrophoresis or melting temperature curves, were notfound in any case. As a calibrator, total RNA from the human colorectalcarcinoma cell line Caco-2, from human liver or from adult male ratliverwere used. The results ofmRNA abundance for the target genes ineach sample were normalized on the basis of 18S rRNA abundance,which was measured with the TaqMan Ribosomal RNA ControlReagents kit (Applied Biosystems).

Fig. 1. Spectrofluorometrically determined reactive oxygen species (ROS) production by livhuman choriocarcinoma JAr cells were incubated with 0–464 µM tert-butyl hydroperoxidmeasured at different time points. (D) ROS generation by rat hepatocytes, and the HepG2arbitrary fluorescent units (AFU) per 104 cells, from three different cultures carried out in tricells.

The primer oligonucleotide sequences and conditions used to carryout quantitative PCR of rat BVRα (Briz et al., 2006), SVCT1 and SVCT2(Perez et al., 2007) and the human nuclear receptors FXR, SHP, PXRand CAR (Serrano et al., 2007) have been described previously. Theprimer oligonucleotide sequences for human BVRα were: forwardprimer (position 254–276, GeneBank Accession Number U34877) 5′-AGA GGT GGA GGT CGC CTA TAT CT-3′ and reverse primer (position339–322) 5′-GGA CGT GCT TGC CAG CAT-3′. For human SVCT1 theywere: forward primer (position 514–533, GeneBank AccessionNumber NM_005847) 5′-GGA TAC GGG AGG TCC AGG GT-3′ andreverse primer (position 678–661) 5′-CTC GGT CGC CAG CAG CTT-3′.For human SVCT2 they were: forward primer (position 966–987,GeneBank Accession Number NM_005116) 5′-AAC AGC AGA GCT GTTGCA CAC A-3′ and reverse primer (position 1198–1181) 5′-GCA TGGCAA TGC CCC AGT-3′. All primers were designed with the assistance ofPrimer Express software (Applied Biosystems), and their specificitywas checked using BLAST. They were obtained from Sigma-Genosys.

Western blot analyses. Immunoblotting analyses of liver andplacenta homogenates were carried out in 7.5 or 10% SDS-polyacrylamide gels, using 100 µg of protein loading per lane. Rabbitpolyclonal antibody against BVRα (OSA-400) was from StressgenBioreagents (bioNova Cientifica, Madrid, Spain). SVCT1 (D19) andSVCT2 (G19) goat-anti-rat polyclonal antibodies and mousemonoclonal antibody against glyceraldehyde-3-phosphatedehydrogenase (GAPDH) (6C5) were from Santa Cruz Biotechnology(CA, USA). Mouse monoclonal antibody against α-actinin (MAB-1682)was from Chemicon (Millipore Co., Madrid, Spain). Anti-goat IgGhorseradish peroxidase-linked antibodies were from Santa CruzBiotechnology. Anti-rabbit or anti-mouse IgG horseradish peroxidase-linked antibodies and enhanced chemiluminescence reagents werefrom Amersham Pharmacia Biotech (Freiburg, Germany).

Statistical analyses. Values are expressed as mean±S.E.M. Statisticalanalyses, were performed using the SPSS 10.0.6 software (SPSS Inc.,Chicago, IL, USA) for Windows (Microsoft Co., Seattle, WA, SA). Tocalculate the statistical significance of the differences between groups,the Bonferroni method for multiple range testing, Student's t-test, orpaired t-test, was used as appropriate.

er and placental cells. (A) Rat hepatocytes, (B) human hepatoblastoma HepG2 and (C)e (tBOOH) for up to 48 h, and ROS generation (dichlorofluorescein fluorescence) wasand JAr cell lines incubated with 464 µM tBOOH for 24 h. Values are means±S.E.M. ofplicate per data point. ⁎Pb0.05 on comparing JAr cells and rat hepatocytes with HepG2

330 M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

Results

Reactive oxygen species (ROS) production by liver and placental cells

ROS generation was induced by long- and short-term exposure ofrat hepatocytes (Fig. 1A), human liver HepG2 cells (Fig. 1B) and humanplacenta JAr cells (Fig. 1C) to tBOOH. ROS production was dependenton tBOOH concentrations in the following order of magnitude:JArNHepG2Nrat hepatocytes (Fig. 1D). The time-course of DCFformationwas linear only during the first few hours. A loss of linearityoccurred independently of the amount of tBOOH added to the culturemedium and hence at different rates of DCF-H transformation, whichsuggests that DCF-H depletion did not account for the observedsteady-state condition.

Using flow cytometry, combined measurements of ROS productionby living cells and cell viability after a shorter exposure time (5 h) ofHepG2 and JAr cells to lower tBOOH concentrations (Fig. 2) were carriedout with a view to elucidating three questions: (i) whether tBOOH-induced toxicitymight account for a reduction in cell viability and hencethe observed plateau in DCF-H transformation; (ii) whether ROS aregenerated by living cells; and (iii) whether increased ROS generationmight in part be due to cell death and the subsequent release ofsubstances able to induce oxidative stress in the remaining living cells.

Fig. 2. Effect of tert-butyl hydroperoxide (tBOOH) on reactive oxygen species (ROS)production and cell viability analyzed by flow cytometry in human hepatoblastomaHepG2 and human choriocarcinoma JAr cells. (A) Typical dot-plots showing dichloro-fluorescein (DCF) fluorescence in living cells (dots delimited by solid line) andpropidium iodide (PI) fluorescence in dead cells (dots delimited by dashed line) inthe HepG2 and JAr cell lines treated with 0.25 or 0.05 mM tBOOH respectively for 5 h.(B) Relationship between the tBOOH concentration and cell viability (circles) or ROSgeneration (squares). Values are means±S.E.M. from three different cultures carried outin triplicate. ⁎Pb0.05 on comparing JAr versus HepG2 cells.

Fig. 3. Effect of deoxycholic acid (DCA) on reactive oxygen species (ROS) production andcell viability analyzed by flow cytometry in human hepatoblastoma HepG2 and humanchoriocarcinoma JAr cells. (A) Typical dot-plots showing dichlorofluorescein (DCF)fluorescence in living cells (dots delimited by solid line) and propidium iodide (PI)fluorescence in dead cells (dots delimited by dashed line) in HepG2 and JAr cell linestreated with 0.25 mM DCA for 48 h. (B) Relationship between the DCA concentrationand cell viability (circles) or ROS generation (squares). Values are means±S.E.M. fromthree different cultures carried out in triplicate. ⁎Pb0.05 on comparing JAr versusHepG2 cells.

Both ROS production and cell death were induced by tBOOH.Viability of JAr cells was markedly decreased when ROS productionwas increased 2-fold by tBOOH. In contrast, in HepG2 cells only a smalldecrease in cell viability was observed at tBOOH concentrationsinducing up to 8-fold increase in ROS production (Fig. 2). This highersensitivity of JAr cells to tBOOH probably explains why in long-termcultures these cells produce more ROS than HepG2 cells (Fig. 1).

To test the effect of hydrophobic BAs, based on results from similarpreliminary experiments carried out tofix sub-toxic concentrations andthe appropriate exposure times (data not shown) we incubated HepG2and JAr cells with DCA for 48 h (Fig. 3). DCA-induced ROS production byHepG2 and JAr cells was similar and low enough to only moderatelydecrease cell viability. This resultwasmoremarked in JAr than inHepG2cells (Fig. 3). It is noteworthy that the tBOOH-andDCA-induced increasein ROS production occurred prior to loss of cell viability, and was not aresult but probably the cause of cell death. The ability of DCA to induceROS production was higher than that of other toxic BAs, such asglycochenodeoxycholic (GCDCA). In comparisonwith untreated controlcells, GCDCA induced ROS production only at high concentrations. Nosignificant GCDCA-induced ROS productionwas observed at concentra-tions of 0.2 and 0.5mM (data not shown).When included in the culturemediumat 1mMfor 48 h,GCDCA-induced rates of ROS generationwere

Fig. 4. Effect of deoxycholic acid (DCA) on the mitochondrial membrane potential(MMP) and cell viability analyzed by flow cytometry in human hepatoblastoma HepG2and human choriocarcinoma JAr cells. (A) Typical dot-plots showing two populations ofliving cells with low (dots delimited by solid thick line) and high (dots delimited by solidthin line) Rh123 fluorescence intensity, corresponding to cells with low and high MMPand a population of dead cells with high propidium iodide (PI) fluorescence intensity(dots delimited by dashed line) in HepG2 and JAr cells treated with 0.25 mM DCA for48 h. (B) Relationship between the DCA concentrations and: i) mean Rh123 fluorescenceintensity of healthy cells with higher MMP, (dots delimited by solid thin line) (circles)and ii) the percentage of cells with reducedMMP (dots delimited by solid thick line, lowRh123 fluorescence intensity) (squares). Values are means±S.E.M. from three differentcultures carried out in triplicate. ⁎Pb0.05 on comparing JAr versus HepG2 cells.

Fig. 5. Effect of bilirubin (BR) on cell viability of human hepatoblastoma HepG2 andhuman choriocarcinoma JAr cells. (A) HepG2 and (B) JAr cells were incubated with 0, 5,10, 25, 50, 100, 200, 300, 400 and 500 µM BR dissolved in dimethyl sulfoxide (DMSO; 0,0.05, 0.1, 0.25, 0.5,1, 2, 3, 4 and 5%, respectively) for 48 h, and cell viability wasmeasuredby Neutral Red uptake. Values are means±S.E.M. from three different cultures carriedout in triplicate. ⁎Pb0.05 on comparing with control cells treated only with vehicle.

331M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

(in % of control) 175±5, 118±10 and 127±9, in rat hepatocytes, HepG2and JAr cells, respectively. Similar concentration (1 mM) of DCA wasused to compare its ability to induce ROS production to that of GCDCA.When cells were exposed to 1 mM DCA values of ROS generation were(also in % of control for 48 h) 234±4, 223±11 and 152±8, in rathepatocytes, HepG2 and JAr cells, respectively.

To checkwhether this effect was due to the detergent properties ofthese BAs, the BA analog zwitterionic detergent 3-[(3-cholamidopro-pyl) dimethylammonio]-1-propanesulfonate (CHAPS) and the non-ionic detergent Triton X-100 were assayed in similar experiments.Neither of these compounds induced significantly different ROSproduction when compared to control conditions (data not shown).

Effect of deoxycholic acid on the mitochondrial membrane potential(MMP)

Using flow cytometry, Rh123 uptake by mitochondria in HepG2and JAr cells cultured in the absence or presence of 0.25 mM DCA for48 h was measured (Fig. 4). In the dot-plots obtained in these studies,

three cell populations were distinguishable: i) cells displaying strongRh123 fluorescence and low uptake of PI (dots delimited by solid thinline in Fig. 4A), which was consistent with the presence of functionalmitochondria in healthy cells able to exclude PI; ii) cells displaying lowRh123 uptake but that were impermeable to PI. These were assumedto be living cells with reduced MMP (dots delimited by solid thick linein Fig. 4A); iii) a cell population, probably formed by dead cells (dotsdelimited by dashed line in Fig. 4A), characterized by low Rh123uptake but high PI fluorescence.

DCA treatment of HepG2 and JAr cells induced an increase in thesize of the population of cells with low MMP as well as in that of thepopulation of dead cells. Moreover, in healthy HepG2 cells, DCAtreatment induced a shift in the mean value of Rh123 fluorescence,indicating a reduction in MMP in this population (Fig. 4).

A dose-dependent DCA-induced increase in the abundance of cellswith reduced MMP (squares) was observed (Fig. 4). In JAr cells, thisincrease was observed at low DCA concentrations, such as 0.05 mM,whereas in HepG2 cells this was not observed up to 0.25 mM DCA. Adramatic drop in mean MMP fluorescence was observed at DCAconcentrations higher than 0.25 mM in HepG2 cells; this was not seenin JAr cells throughout the range of DCA concentrations assayed (Fig. 4).

Effect of ursodeoxycholic acid and bilirubin on DCA-induced changes inMMP, ROS production, and cell viability

Before evaluating the effect of bilirubin on DCA-induced changes,the toxicity of bilirubin was investigated. Appropriate controls in thepresence of the vehicle (DMSO) were included in the assay. In HepG2cells, no toxicity up to 100 µM bilirubin was found (Fig. 5A).However, a decrease in cell viability was observed at bilirubinconcentrations between 200 and 500 µM. This effect was similar to

332 M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

that caused by DMSO alone (Fig. 5A). Since 0.5% DMSO was not toxicin any case, this was used when needed in the rest of experiments.JAr cells, a similar sensitivity to DMSO-induced toxicity was found(Fig. 5B). However, bilirubin was toxic as from very low concentra-tions (Fig. 5B), which means that this toxicity could not be attributedto DMSO. Consequently, in further experiments to test bilirubin-induced protection against DCA toxicity the following non-toxic forthe cell type concentrations were selected: 50 µM for HepG2 cellsand 5 µM for JAr cells. UDCA had no toxic effect on HepG2 or JArcells at the concentrations used in this study (data not shown).

The ability of UDCA and bilirubin to protect against DCA-inducedcell damage was evaluated by quantifying the size of HepG2 and JArcell populations with either reduced MMP or which were dead afterincubating with or without DCA for 48 h. Treatment of HepG2 cellswith 0.25mM and 0.50mMDCA induced an increase in the number ofcells with reducedMMP by 1.7-fold and 5.7-fold, respectively (Fig. 6A).Neither UDCA nor bilirubin prevented this change. In contrast, thiswas favored by UDCA (Fig. 6A). In JAr cells, 0.10 mM and 0.25 mMDCAinduced 2-fold and 3-fold increases in the percentage of cells withreduced MMP, respectively (Fig. 6B). Both UDCA and bilirubin partlyinhibited this change at low, but not at high, DCA concentrations.

When the effect on cell survival was examined, no effect of UDCAor bilirubin on DCA-induced cell death in HepG2 cells was found(Fig. 7A), whereas in JAr cells a protective effect of both compoundsat low, but not at high, DCA concentrations was observed (Fig. 7B).When DCA-induced ROS formation was investigated at doses atwhich UDCA and/or bilirubin afforded some protection, both

Fig. 6. Effect of ursodeoxycholic acid (UDCA) and bilirubin (BR) on deoxycholic acid(DCA)-induced changes in the mitochondrial membrane potential (MMP) analyzed byflowcytometry in human hepatoblastoma HepG2 and human choriocarcinoma JAr cells.(A) HepG2 and (B) JAr cells were incubated with 0.1–0.5 mM DCA in the absence or thepresence of 5 or 50 µM BR or 100 µM UDCA for 48 h. The percentage of cells withreducedMMP corresponds to a cell populationwith low Rh123 fluorescence intensity asmeasured by flow cytometry. Values are means±S.E.M. from three different culturescarried out in triplicate. ⁎Pb0.05 on comparing DCA-treated cells with theircorresponding control; †Pb0.05 on comparing DCA plus UDCA or BR versus DCA-treated cells.

Fig. 7. Effect of ursodeoxycholic acid (UDCA) and bilirubin (BR) on deoxycholic acid(DCA)-induced cell death analyzed by flow cytometry in human hepatoblastoma HepG2and choriocarcinoma JAr cells. (A) HepG2 and (B) JAr cells were incubated with 0.1–0.5 mM DCA in the absence or the presence of 5 or 50 µM BR or 100 µM UDCA for 48 h.Using flow cytometry, cell death was estimated as the percentage of cells with highpropidium iodide fluorescence. Values are means±S.E.M. from three different culturescarried out in triplicate. ⁎Pb0.05 on comparing DCA-treated cells with their correspond-ing control; †Pb0.05 on comparing DCA plus UDCA or BR versus DCA-treated cells.

compounds were found to inhibit DCA-induced ROS generation inHepG2 (Fig. 8A) and JAr (Fig. 8B) cells.

In vitro effect of UDCA and bilirubin on BVRα, SVCT1 and SVCT2expression

To evaluate the possibility that UDCA and bilirubin may have anindirect effect in the anti-oxidant defense system, the expression ofsome of the genes involved in this systemwas measured. TCA is one ofthemajor BAs in the rat and human BA pool, beingmore abundant andless toxic than DCA. Therefore, TCAwas also included in this part of thestudy to simulate BA accumulation during cholestasis. BVRα mRNAlevels were higher in HepG2 than in JAr cells (Fig. 9A). TCA and UDCA,but not bilirubin, up-regulated BVRα both in HepG2 (2-fold increase)and JAr (4-fold increase) cells. The relative abundance of SVCT1 mRNAwas approximately 10-fold higher in HepG2 than in JAr cells (Fig. 9B).TCA and UDCA had a stimulating effect on SVCT1 expression in JAr butnot HepG2 cells, whereas bilirubin markedly up-regulated SVCT1 inHepG2 but not in JAr cells (Fig. 9B). SVCT2 mRNA levels were alsoapproximately 10-fold higher in HepG2 than in JAr cells (Fig. 9C).However, in the case of this transporter TCA, UDCA and bilirubinwereable to increase its expression both in HepG2 and JAr cells (Fig. 9C).

Effect of UDCA and bilirubin on the expression of nuclear receptors

To investigate the existence of changes in sensitivity to theregulation of gene expression by biliary compounds, the relative

Fig. 8. Effect of ursodeoxycholic acid (UDCA) and bilirubin (BR) on deoxycholic acid(DCA)-induced changes in reactive oxygen species (ROS) production analyzed by flowcytometry in human hepatoblastoma HepG2 and human choriocarcinoma JAr cells. (A)HepG2 and (B) JAr cells were incubated with 0.5 and 0.1 mM DCA, respectively in theabsence or the presence of 5 or 50 µM BR or 100 µM UDCA for 48 h. ROS generationwasevaluated by measuring the dichlorofluorescein (DCF) fluorescence of viable cells byflow cytometry. Values are means±S.E.M. from three different cultures carried out intriplicate. ⁎Pb0.05 on comparing DCA-treated cells with their corresponding control;†Pb0.05 on comparing DCA plus UDCA or BR versus DCA-treated cells.

Fig. 9. Relative abundance of biliverdin-IXα reductase (BVRα) (A), SVCT1 (B) and SVCT2(C) mRNA in human hepatoblastoma HepG2 and human choriocarcinoma JAr cellsincubated with 100 µM taurocholic acid (TCA), 100 µM ursodeoxycholic acid (UDCA) orbilirubin (BR) at 5 or 50 µM for 48 h. Values are means±S.E.M and are expressed aspercentages of the external calibrators. Three different cultures were carried out intriplicate; ⁎Pb0.05 on comparing with untreated cells (control).

333M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

abundances of FXR, SHP, PXR and CAR mRNA in HepG2 and JAr cellstreated with TCA, UDCA or bilirubin for 48 h were determined (Table1). As compared to normal human liver, the expression of thesenuclear receptors was of similar magnitude both in HepG2 and JArcells, except for SHP and SXR. SHP expression was very high in HepG2cells, whereas SXR expression was very low in JAr cells. SXR was up-regulated by TCA or UDCA treatment in both cell types. In HepG2 cells,bilirubin, at a sub-toxic concentration for these cells (50 µM), was ableto enhance expression of SXR. This effect was not seen in JAr cells,which required a lower sub-toxic bilirubin concentration (5 µM). CARwas up-regulated only by bilirubin in HepG2 cells. FXR and SHPexpression was enhanced by TCA and UDCA in HepG2 and JAr cells.

In vivo effect of cholestasis and UDCA treatment on BVRα, SVCT1 andSVCT2 expression

To elucidate whether changes in BVRα, SVCT1 and SVCT2expression also occurred in an in vivo situation, the levels of thethree mRNAs and proteins were evaluated in OCP model. Cholestasiswas assessed by changes in maternal serum biochemical parameters.These were manifested as increased serum bilirubin concentrations,increased activities in serum alkaline phosphatase, gamma-glutamyltranspeptidase and transaminases, and decreased serum albuminconcentrations (data not shown). The values found here (e.g., bilirubinconcentrations close to 100 µM) were in agreement with the resultsobtained in previous studies using this model (Perez et al., 2005,2007).

As previously reported, OCP induced an up-regulation of BVRα(Fig. 10B), SVCT1 (Fig. 11B) and SVCT2 (Fig. 12B) in rat maternal andfetal liver and, although to a lesser extent, in placenta (Perez et al.,2007). A further significant increase in the expression levels of BVRα(Fig. 10B), SVCT1 (Fig. 11B) and SVCT2 (Fig. 12B) was found in thematernal and fetal livers of the OCP group treated with UDCA.

However, in the placenta of these animals UDCA only exerted astimulating effect on BVRα expression. When Western blot wascarried out to determine protein levels, the results were in generalconsistent with the changes observed in the mRNA of BVRα (Fig. 10A),SVCT1 (Fig. 11A) and SVCT2 (Fig. 12A).

Discussion

It is well-known that the accumulation of hydrophobic BAs duringcholestasis causes mitochondrial respiratory dysfunction, leading tothe generation of oxygen free radicals, oxidative stress, and apoptosisin hepatocytes (Patel and Gores, 1995). The intracellular generation ofhydroperoxides by mitochondria is an early event in this process(Sokol et al., 1995). In the present study we have found that the effectof oxidative insult and the response of defense mechanisms in theliver and placenta differ in several ways.

First, we characterized the ability of cultured liver and placentalcells to produce ROS and their ability to become resistant to them byusing the pro-oxidant compound tBOOH. Liver cells were moreresistant to the toxicity induced by tBOOH-stimulated ROS production.In contrast, the difference in sensitivity of these two cell lines to BAs

Table 1Effect of bile acids and bilirubin on the expression of nuclear receptors

Nuclear receptor Control Treated Control Treated

HepG2 cells HepG2 cells (% of control) JAr cells JAr cells (% of control)

Ct % of liver TCA UDCA BR 50 µM Ct % of Liver TCA UDCA BR 5 µM

FXR 24.1 93±18 308±70a 262±18a 173±26 26.8 27±9 462±68a 767±111a 111±25SHP 22.0 1002±3 200±20a 204±53a 147±28 24.3 73±28 350±38a 537±75a 104±27SXR 24.9 27±5 219±42a 217±31a 335±38a 32.7 0.3±0.1 451±55a 534±58a 110±19CAR 26.2 17±3 134±27 160±40 285±46a 26.5 36±11 175±19 204±45 62±13

Relative abundance of mRNA as determined by real-time quantitative RT-PCR in human hepatoblastoma HepG2 and choriocarcinoma JAr cells incubated with 100 µM taurocholicacid (TCA), 100 µM ursodeoxycholic acid (UDCA) or bilirubin (BR) at 5 or 50 µM for 48 h. In all reactions, human liver was used as calibrator. Values are means±S.E.M. from threedifferent cultures carried out in triplicate.

a Pb0.05 on comparing with untreated control cells by the paired t-test.

334 M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

was less pronounced. DCAwas more efficient than GCDCA at inducingROS production, with a similar capability in liver and placental cells.

BA-induced toxicity is partly due to the induction of themitochondrial permeability transition (Bernardi et al., 2001), whichresults in a loss of MMP (Yerushalmi et al., 2001; Payne et al., 2005)leading to apoptosis activation (Martin and Green,1995;Washo-Stultzet al., 2002). However, ROS-induced mitochondrial dysfunction mayalso lead to necrosis (Kowaltowski et al., 1998). Both cell deathpathways may be involved in BA-induced liver injury in vitro: necrosisat high concentrations and apoptosis at lower BA concentrations(Spivey et al., 1993; Patel et al., 1994).

In the present study, we observed that DCA causes an increase inthe number of placental and liver cells with a reduced MMP. Similarobservations have been reported in studies using human HCT-116colon cancer cells (Payne et al., 2005). Moreover, in the population ofcells with high MMP, the mean value of this parameter was decreasedby DCA in liver cells, but not in placental cells. This difference may bedue to the higher resistance of liver cells to ROS, which is consistentwith the reported resistance to DCA-induced apoptosis (Bernstein etal., 1999). In agreement with our results, the high sensitivity of JAr

Fig. 10. Abundance of protein (A) and mRNA (B) of biliverdin-IXα reductase (BVRα) inmaternal liver, fetal liver and placenta of rats on day 21 of pregnancy. On day 14 ofgestation, pregnant rats underwent a sham operation (Control), obstructive cholestasis,or obstructive cholestasis followed by treatment with UDCA. The values of relativemRNA abundance are means±S.E.M and are expressed as percentages of the externalcalibrators. In all groups n=6 mothers and ≥12 fetuses. ⁎Pb0.05 on comparing with thecontrol; †Pb0.05 on comparing OCP+UDCA versus OCP group.

cells to menadione-induced oxidative stress has been reported(Hallmann et al., 2004).

After characterizing the sensitivity of HepG2 and JAr cells tooxidative stress we investigated the protective effect of twocompounds, UDCA and bilirubin. The reason for this choice and theresults obtained are discussed below. In contrast to GCDCA and DCA,the hydrophilic BA UDCA has beneficial effects in cholestatic liverdiseases (Paumgartner and Beuers, 2002), mainly due to theprotection against other more toxic BAs (Güldütuna et al., 1993), thestimulation of biliary secretion (Dumont et al., 1980), the anti-oxidantactivity due – in part – to an enhancement in glutathione levels(Mitsuyoshi et al., 1999; Serviddio et al., 2004), and the inhibition ofapoptosis (Rodrigues et al., 1998). In previous studies carried out byour group, UDCA was found to have beneficial effects on placentalfunction both in patients with ICP (Serrano et al., 1998) and in ratswith OCP (Serrano et al., 2003).

Bilirubinwas investigated because it accumulates in cholestasis andbecause this pigment is an anti-oxidant agent of possible physiologicalimportance, here we investigated whether it might play a protectiverole in cholestasis during pregnancy. It has beenpreviously shown that

Fig.11. Abundance of protein (A) andmRNA (B) of SVCT1 inmaternal liver, fetal liver andplacenta of rats on day 21 of pregnancy. On day 14 of gestation, pregnant ratsunderwent a sham operation (Control), obstructive cholestasis, or obstructivecholestasis followed by treatment with UDCA. The values of relative mRNA abundanceare means±S.E.M and are expressed as percentages of the external calibrators. In allgroups n=6 mothers and ≥12 fetuses. ⁎Pb0.05 on comparing with control; †Pb0.05 oncomparing OCP+UDCA versus OCP group.

Fig. 12. Abundance of protein (A) and mRNA (B) of SVCT2 in maternal liver, fetal liverand placenta of rats on day 21 of pregnancy. On day 14 of gestation, pregnant ratsunderwent a sham operation (Control), obstructive cholestasis, or obstructivecholestasis followed by treatment with UDCA. The values of relative mRNA abundanceare means±S.E.M and are expressed as percentages of the external calibrators. In allgroups n=6 mothers and ≥12 fetuses. ⁎Pb0.05 on comparing with control; †Pb0.05 oncomparing OCP+UDCA versus OCP group.

335M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

micromolar concentrations of bilirubin efficiently scavenge peroxylradicals in vitro and that, at low oxygen tension, the anti-oxidantactivity of bilirubin surpasses that of α-tocopherol, the best-knownanti-oxidant against lipid peroxidation (Stocker et al.,1987). Moreover,bilirubin has been found to attenuate free radical-induced cell damagein several experimental models (Motterlini et al., 1996; Foresti et al.,1999; Clark et al., 2000; Dore et al., 1999).

In the present study, both UDCA and bilirubinwere found to be ableto inhibit DCA-induced ROS generation in both placenta and liver cells,although these levels were still very high in liver cells. Thesecompounds prevented the decline in the MMP and the cell deathcaused by DCA at low, but already toxic, concentrations in placentalcells. At highDCAconcentrations, noprotectionwas observed for eitherUDCA or bilirubin. The latter inhibited GCDC-induced apoptosis in rathepatocytes, and this effect has been associated with the inhibition ofROS generation (Granato et al., 2003). However, the effect of bilirubin isdual (anti-oxidant or pro-oxidant), depending on the concentrationand the cell type exposed to this pigment (Rao et al., 2006).

Here, two important components of the anti-oxidant defensesystem were investigated, BVRα and vitamin C transporters. TheBVRα-mediated bilirubin/biliverdin redox cycle seems to be involvedin the detoxification of lipophilic ROS (Sedlak and Snyder, 2004).Vitamin C availability for cells is dependent on its uptake via specificmembrane transporters that have only recently been identified(Tsukaguchi et al., 1999). SVCT1 is highly expressed in liver, whereasSVCT2 is highly expressed in placenta (Tsukaguchi et al., 1999), wherethis carrier is required for the transport of ascorbic acid across thisorgan (Sotiriou et al., 2002). Moreover, a deficiency of SVCT2 innewborn mice has been reported to be lethal (Sotiriou et al., 2002).

We have previously reported that BVRα, SVCT1 and SVCT2 are up-regulated in maternal liver, fetal liver and placenta by maternalcholestasis (Perez et al., 2007). To elucidate whether accumulated BAsor bilirubin were involved in the signalling processes accounting forthe stimulation of this expression, HepG2 and JAr cells were incubated

with TCA or bilirubin. Although TCA, but not bilirubin, induces the up-regulation of BVRα in both cell types, both biliary components (TCAand bilirubin) are probably involved in the up-regulation of SVCT2 inboth liver and placenta. Finally, the up-regulation of SVCT1 proved tobe more complex, because liver cells are highly sensitive to bilirubin(but not to BAs), whereas in placental cells only BAs (but not bilirubin)are able to up-regulate this gene.

We have also previously shown that the anti-oxidant melatonin up-regulates the expression of rat BVRα and SVCT2 in these organs and thatof SVCT1 in liver but not in placenta (Perez et al., 2007). Moreover, it waspreviously found thatUDCAalso stimulates other components of the anti-oxidant system such as catalase, glutathione peroxidase and glutathione-S-transferase activities, as well as glutathione levels in the placentas ofanimals with OCP (Perez et al., 2006). In the present study we found thatUDCA also increased BVRα, SVCT1 and SVCT2 expression in whole ratmaternal and fetal liver tissues. Consistentwith the foregoing,UDCAhadapositive effect on the expression of these genes in both liver and placentalcells, except in the caseof SVCT1 inHepG2cells, in spite of themarkedup-regulation in maternal and fetal livers caused by UDCA administration torats with OCP. Surprisingly, in the rat placenta UDCA only exerted astimulating effect on BVRα expression, whereas in JAr cells, both TCA andUDCAwere able to increase BVRα and SVCT1 and SVCT2 expression. Thedifferential regulation of these transporters may be explained by the factthat both SVCT isoforms play different roles in vitamin C handling. SVCT1is involved inwhole-body vitamin C homeostasis, whereas the main roleof SVCT2 is protection against oxidative stress of cells with particularlyhigh metabolic rate (Savini et al., 2008).

As the first step to investigate the mechanism by which biliarycomponents may affect the expression of BVRα, SVCT1 and SVCT2, herewe analyzed the expression in liver and placental cells of several nuclearreceptors and transcription factors sensitive to activation by BAs,bilirubin, and other related compounds (Handschin and Meyer, 2005).The gene encoding SHP is a major FXR target gene in the liver. Theactivation of FXR by BAs induces the expression of SHP, leading tochanges in the expression of several genes involved in BA metabolismand transport (Kullak-Ublick et al., 2004). Although these nuclearreceptors are predominantly expressed in liver, their expression hasalso recently been described in placental cells, which is consistent withtheir potential role as sensors able to participate in the regulation ofgenes involved in preventing the accumulation of potentially toxiccompounds, including BAs, in the fetus (Serrano et al., 2007).

In the present study, we observed that the BA-induced up-regulation of FXR, SHP and SXR was higher in placental than in livercell lines. Bilirubin induced the expression of SXR and CAR only inHepG2 cells, whereas, in this respect, JAr cells seemed to be insensitiveto bilirubin, at least at the low concentration of the pigment that couldbe used without any toxic effect for these cells.

In sum, UDCA and sub-toxic bilirubin concentrations protect againstcholestasis-induced toxicity during pregnancy. These compounds haveboth direct anti-oxidant properties and indirect beneficial effects, suchas enhancing the expression of several elements of the anti-oxidantdefense system, i.e., BVRα, SVCT1 and SVCT2, as well as several nuclearreceptors sensitive to activation by biliary compounds.

Acknowledgments

This study was supported in part by the Instituto de Salud Carlos III,FIS (Grant PI051547), Fundacion InvestigacionMedica,MutuaMadrileña(Conv-III, 2006), and Junta deCastilla y León (Grants SA021B06 and SAN/191/2006), Spain. The group belongs to the CIBERehd (Centro deInvestigacion Biomedica en Red) for Hepatology and GastroenterologyResearch funded by the Instituto de Salud Carlos III, Spain.

The authors thank Mrs. M.I. Hernandez Rodriguez for hersecretarial help, and Mr. L. Muñoz de la Pascua and Mr. J.F. MartinMartin for caring for the animals. Thanks are also due to E.J. Keck andN. Skinner for revision of the English text of the manuscript.

336 M.J. Perez et al. / Toxicology and Applied Pharmacology 232 (2008) 327–336

References

Attili, A.F., Angelico, M., Cantafora, A., Alvaro, D., Capocaccia, L., 1986. Bile acid inducedliver toxicity: relation to the hydrophobic-hydrophilic balance of bile acids. Med.Hypotheses 19, 57–69.

Bernardi, P., Petronilli, V., Di Lisa, F., Forte, M., 2001. A mitochondrial perspective on celldeath. Trends Biochem. Sci. 26, 112–117.

Bernstein, H., Payne, C.M., Bernstein, C., Schneider, J., Beard, S.E., Crowley, C.L., 1999.Activation of the promoters of genes associatedwith DNAdamage, oxidative stress, ERstress and protein malfolding by the bile salt, deoxycholate. Toxicol. Lett. 108, 37–46.

Billington, D., Evans, C.E., Godfrey, P.P., Coleman, R., 1980. Effects of bile salts on theplasma membranes of isolated rat hepatocytes. Biochem. J. 188, 321–327.

Briz, O., Macias, R.I., Perez, M.J., Serrano, M.A., Marin, J.J.G.,, 2006. Excretion of fetalbiliverdin by the rat placenta-maternal liver tandem. Am. J. Physiol. Regul. Integr.Comp. Physiol. 290, R749–R756.

Clark, J.E., Foresti, R., Sarathchandra, P., Kaur, H., Green, C.J., Motterlini, R., 2000. Hemeoxygenase-1-derived bilirubin ameliorates post-ischemic myocardial dysfunction.Am. J. Physiol. Heart Circ. Physiol. 278, 643–651.

Dore, S., Takahashi, M., Ferris, C.D., Hester, L.D., Guastella, D., Snyder, S.H., 1999.Bilirubin, formed by activation of heme oxygenase-2, protects neurons againstoxidative stress injury. Proc. Natl. Acad. Sci. USA. 96, 2445–2450.

Dumont, M., Erlinger, S., Uchman, S., 1980. Hypercholeresis induced by ursodeoxycholicacid and 7-ketolithocholic acid in the rat: possible role of bicarbonate transport.Gastroenterology 79, 82–89.

Faubion, W.A., Guicciardi, M.E., Miyoshi, H., Bronk, S.F., Roberts, P.J., Svingen, P.A.,Kaufmann, S.H., Gores, G.J., 1999. Toxic bile salts induce rodent hepatocyteapoptosis via direct activation of Fas. J. Clin. Invest. 103, 137–145.

Fautz, R., Husein, B., Hechenberger, C.,1991. Application of theNeutral Red assay (NR assay)to monolayer cultures of primary hepatocytes: rapid colorimetric viability determina-tion for the unscheduled DNA synthesis test (UDS). Mutat. Res. 253, 173–179.

Foresti, R., Sarathchandra, P., Clark, J.E., Green, C.J.,Motterlini, R.,1999. Peroxynitrite induceshaem oxygenase-1 in vascular endothelial cells: a link to apoptosis. Biochem. J. 339,729–736.

Gourley, G.R., 1997. Bilirubin metabolism and kernicterus. Adv. Pediatr. 44, 173–229.Granato, A., Gores, G., Vilei, M.T., Tolando, R., Ferraresso, C., Muraca, M., 2003. Bilirubin

inhibits bile acid induced apoptosis in rat hepatocytes. Gut 52, 1774–1778.Güldütuna, S., Leuschner, M., Wunderlich, N., Nickel, A., Bhatti, S., Hübner, K., Leuschner,

U., 1993. Cholic acid and ursodeoxycholic acid therapy in primary biliary cirrhosis.Changes in bile acid patterns and their correlation with liver function. Eur. J. Clin.Pharmacol. 45, 221–225.

Guo, G.L., Lambert, G., Negishi, M., Ward, J.M., Brewer Jr., H.B., Kliewer, S.A., Gonzalez, F.J.,Sinal, C.J., 2003. Complementary roles of farnesoidX receptor, pregnaneX receptor, andconstitutive androstane receptor in protection against bile acid toxicity. J. Biol. Chem.278, 45062–45071.

Hallmann, A., Klimek, J., Masaoka, M., Kamiński, M., Kedzior, J., Majczak, A., Niemczyk, E.,Woźniak,M., Trzonkowski, P.,Wakabayashi, T., 2004. Partial characterization of humanchoriocarcinoma cell line JAR cells in regard to oxidative stress. Acta Biochim. Pol. 51,1023–1038.

Handschin, C., Meyer, U.A., 2005. Regulatory network of lipid-sensing nuclear receptors:roles for CAR, PXR, LXR, and FXR. Arch. Biochem. Biophys. 433, 387–396.

Kowaltowski, A.J., Netto, L.E., Vercesi, A.E., 1998. The thiol-specific antioxidant enzymeprevents mitochondrial permeability transition. Evidence for the participation ofreactive oxygen species in this mechanism. J. Biol. Chem. 273, 12766–12769.

Kullak-Ublick, G.A., Stieger, B., Meier, P.J., 2004. Enterohepatic bile salt transporters innormal physiology and liver disease. Gastroenterology 1261, 322–342.

Lammert, F., Marschall, H.U., Matern, S., 2003. Intrahepatic cholestasis of pregnancy.Curr. Treat. Options Gastroenterol. 6, 123–132.

Ljubuncic, P., Fuhrman, B., Oiknine, J., Aviram,M., Bomzon, A.,1996. Effect of deoxycholicacid and ursodeoxycholic acid on lipid peroxidation in cultured macrophages. Gut39, 475–478.

Macias, R.I., Pascual, M.J., Bravo, A., Alcalde, M.P., Larena, M.G., St-Pierre, M.V., Serrano,M.A., Marin, J.J.G., 2000. Effect of maternal cholestasis on bile acid transfer acrossthe rat placenta-maternal liver tandem. Hepatology 31, 975–983.

Marin, J.J.G., Briz, O., Serrano, M.A., 2004. A review on the molecular mechanismsinvolved in the placental barrier for drugs. Curr. Drug Deliv. 1, 275–289.

Marin, J.J.G., Macias, R.I.R., Serrano, M.A., 2003. The hepatobiliary-like excretoryfunction of the placenta. A review. Placenta 24, 431–438.

Martin, C., Martinez, R., Navarro, R., Ruiz-Sanz, J.I., Lacort, M., Ruiz-Larrea, M.B., 2001.tert-Butyl hydroperoxide-induced lipid signaling in hepatocytes: involvement ofglutathione and free radicals. Biochem. Pharmacol. 62, 705–712.

Martin, S.J., Green, D.R., 1995. Protease activation during apoptosis: death by a thousandcuts. Cell 82, 349–352.

Martinez-Diez, M.C., Serrano, M.A., Monte, M.J., Marin, J.J.G., 2000. Comparison of theeffects of bile acids on cell viability and DNA synthesis by rat hepatocytes in primaryculture. Biochim. Biophys. Acta 1500, 153–160.

Mitsuyoshi, H., Nakashima, T., Sumida, Y., Yoh, T., Nakajima, Y., Ishikawa, H., Inaba, K.,Sakamoto, Y., Okanoue, T., Kashima, K., 1999. Ursodeoxycholic acid protectshepatocytes against oxidative injury via induction of antioxidants. Biochem.Biophys. Res. Commun. 263, 537–542.

Monte,M.J., Morales, A.I., Arevalo, M., Alvaro, I., Macias, R.I., Marin, J.J.G.,1996. Reversibleimpairment of neonatal hepatobiliary function bymaternal cholestasis. Hepatology23, 1208–1217.

Monte, M.J., Rodriguez-Bravo, T., Macias, R.I., Bravo, P., El-Mir, M.Y., Serrano, M.A., Lopez-Salva, A., Marin, J.J.G., 1995. Relationship between bile acid transplacental gradientsand transport across the fetal-facing plasma membrane of the human trophoblast.Pediatr. Res. 38, 156–163.

Motterlini, R., Foresti, R., Intaglietta, M., Winslow, R.M., 1996. NO-mediated activation ofheme oxygenase: endogenous cytoprotection against oxidative stress to endothe-lium. Am. J. Physiol. Heart Circ. Physiol. 270, 107–114.

Muraca, M., Fevery, J., Blanckaert, N., 1988. Analytic aspects and clinical interpretation ofserum bilirubins. Semin. Liver Dis. 8, 137–147.

Patel, T., Bronk, S.F., Gores, G.J., 1994. Increases of intracellular magnesium promoteglycodeoxycholate-induced apoptosis in rat hepatocytes. J. Clin. Invest. 94,2183–2192.

Patel, T., Gores, G.J., 1995. Apoptosis and hepatobiliary disease. Hepatology 21,1725–17241.

Paumgartner, G., Beuers, U., 2002. Ursodeoxycholic acid in cholestatic liver disease:mechanisms of action and therapeutic use revisited. Hepatology 36, 525–531.

Payne, C.M., Crowley-Weber, C.L., Dvorak, K., Bernstein, C., Bernstein, H., Holubec, H.,Crowley, C., Garewal, H., 2005. Mitochondrial perturbation attenuates bile acid-induced cytotoxicity. Cell Biol. Toxicol. 21, 215–231.

Perez, M.J., Castaño, B., Gonzalez-Buitrago, J.M., Marin, J.J.G., 2007. Multiple protectiveeffects of melatonin against maternal cholestasis-induced oxidative stress andapoptosis in the rat fetal liver–placenta-maternal liver trio. J. Pineal Res. 43,130–139.

Perez, M.J., Cederbaum, A.I., 2002. Antioxidant and pro-oxidant effects of a manganeseporphyrin complex against CYP2E1-dependent toxicity. Free Radic. Biol. Med. 33,111–127.

Perez, M.J., Macias, R.I., Duran, C., Monte, M.J., Gonzalez-Buitrago, J.M., Marin, J.J.G.,2005. Oxidative stress and apoptosis in fetal rat liver induced by maternalcholestasis. Protective effect of ursodeoxycholic acid. J. Hepatol. 43, 324–332.

Perez, M.J., Macias, R.I., Marin, J.J.G., 2006. Maternal cholestasis induces placentaloxidative stress and apoptosis. Protective effect of ursodeoxycholic acid. Placenta27, 34–41.

Rao, P., Suzuki, R., Mizobuchi, S., Yamaguchi, T., Sasaguri, S., 2006. Bilirubin exhibits anovel anti-cancer effect on human adenocarcinoma. Biochem. Biophys. Res.Commun. 342, 1279–1283.

Rodrigues, C.M., Fan, G., Wong, P.Y., Kren, B.T., Steer, C.J., 1998. Ursodeoxycholic acid mayinhibitdeoxycholicacid-inducedapoptosisbymodulatingmitochondrial transmem-branepotential and reactiveoxygenspecies production.Mol.Med. 4,165–178.

Rosenkranz, A.R., Schmaldienst, S., Stuhlmeier, K.M., Chen,W., Knapp,W., Zlabinger, G.J.,1992. A microplate assay for the detection of oxidative products using 2′,7′-dichlorofluorescin-diacetate. J. Immunol. Meth. 156, 39–45.

Savini, I., Rossi, A., Pierro, C., Avigliano, L., Catani, M.V., 2008. SVCT1 and SVCT2: keyproteins for vitamin C uptake. Amino Acids 34, 347–355.

Sedlak, T.W., Snyder, S.H., 2004. Bilirubin benefits: cellular protection by a biliverdinreductase antioxidant cycle. Pediatrics 113, 1776–1782.

Serrano, M.A., Brites, D., Larena, M.G., Monte, M.J., Bravo, M.P., Oliveira, N., Marin, J.J.G.,1998. Beneficial effect of ursodeoxycholic acid on alterations induced by cholestasisof pregnancy in bile acid transport across the human placenta. J. Hepatol. 28,829–839.

Serrano, M.A., Macias, R.I., Briz, O., Monte, M.J., Blazquez, A.G., Williamson, C., Kubitz, R.,Marin, J.J.G., 2007. Expression in human trophoblast and choriocarcinoma cell lines,BeWo, Jeg-3 and JAr of genes involved in the hepatobiliary-like excretory functionof the placenta. Placenta 28, 107–117.

Serrano, M.A., Macias, R.I., Vallejo, M., Briz, O., Bravo, A., Pascual, M.J., St-Pierre, M.V.,Stieger, B., Meier, P.J., Marin, J.J.G., 2003. Effect of ursodeoxycholic acid on theimpairment induced by maternal cholestasis in the rat placenta-maternal livertandem excretory pathway. J. Pharmacol. Exp. Ther. 305, 515–524.

Serviddio, G., Pereda, J., Pallardó, F.V., Carretero, J., Borras, C., Cutrin, J., Vendemiale, G.,Poli, G., Viña, J., Sastre, J., 2004. Ursodeoxycholic acid protects against secondarybiliary cirrhosis in rats by preventing mitochondrial oxidative stress. Hepatology39, 711–720.

Sokol, R.J., Straka, M.S., Dahl, R., Devereaux, M.W., Yerushalmi, B., Gumpricht, E.,Elkins, N., Everson, G., 2001. Role of oxidant stress in the permeability transitioninduced in rat hepatic mitochondria by hydrophobic bile acids. Pediatr. Res. 49,519–531.

Sokol, R.J., Winklhofer-Roob, B.M., Devereaux, M.W., McKim, J.M., 1995. Generation ofhydroperoxides in isolated rat hepatocytes and hepatic mitochondria exposed tohydrophobic bile acids. Gastroenterology 109, 1249–1256.

Sotiriou, S., Gispert, S., Cheng, J., Wang, Y., Chen, A., Hoogstraten-Miller, S., Miller, G.F.,Kwon, O., Levine, M., Guttentag, S.H., Nussbaum, R.L., 2002. Ascorbic-acidtransporter Slc23a1 is essential for vitamin C transport into the brain and forperinatal survival. Nat. Med. 8, 514–517.

Spivey, J.R., Bronk, S.F., Gores, G.J., 1993. Glycochenodeoxycholate-induced lethalhepatocellular injury in rat hepatocytes. Role of ATP depletion and cytosolic freecalcium. J. Clin. Invest. 92, 17–24.

Stocker, R., Yamamoto, Y., McDonagh, A.F., Glazer, A.N., Ames, B.N., 1987. Bilirubin is anantioxidant of possible physiological importance. Science 235, 1043–1046.

Tsukaguchi, H., Tokui, T., Mackenzie, B., Berger, U.V., Chen, X.Z., Wang, Y., Brubaker, R.F.,Hediger, M.A., 1999. A family of mammalian Na+-dependent L-ascorbic acidtransporters. Nature 399, 70–75.

Van Hootegem, P., Fevery, J., Blanckaert, N., 1985. Serum bilirubins in hepatobiliarydisease: comparison with other liver function tests and changes in thepostobstructive period. Hepatology 5, 112–117.

Washo-Stultz, D., Crowley-Weber, C.L., Dvorakova, K., Bernstein, C., Bernstein, H., Kunke,K., Waltmire, C.N., Garewal, H., Payne, C.M., 2002. A family of mammalian Na+-dependent L-ascorbic acid transporters and arachidonic acid metabolism indeoxycholate-induced apoptosis. Cancer Lett. 177, 129–144.

Yerushalmi, B., Dahl, R., Devereaux, M.W., Gumpricht, E., Sokol, R.J., 2001. Bile acid-induced rat hepatocyte apoptosis is inhibited by antioxidants and blockers of themitochondrial permeability transition. Hepatology 33, 616–626.