Cocarcinogenic Effects of Intrahepatic Bile Acid …...2013/12/30 · and the proliferation and...

DNA Damage and Repair

Cocarcinogenic Effects of Intrahepatic Bile AcidAccumulation in Cholangiocarcinoma Development

Elisa Lozano1, Laura Sanchez-Vicente1, Maria J. Monte1,4, Elisa Herraez1, Oscar Briz1,4,Jesus M. Banales2,3,4, Jose J.G. Marin1,4, and Rocio I.R. Macias1,4

AbstractBile acid accumulation in liver with cholangiolar neoplastic lesions may occur before cholestasis is clinically

detected. Whether this favors intrahepatic cholangiocarcinoma development has been investigated in this study.The E. coli RecA gene promoter was cloned upstream from Luc2 to detect in vitro direct genotoxic ability byactivation of SOS genes. This assay demonstrated that bile acids were not able to induce DNA damage. Thegenotoxic effect of the DNA-damaging agent cisplatin was neither enhanced nor hindered by the hepatotoxic andhepatoprotective glycochenodeoxycholic and glycoursodeoxycholic acids, respectively. In contrast, thioacetamidemetabolites, but not thioacetamide itself, induced DNA damage. Thus, thioacetamide was used to induce livercancer in rats, which resulted in visible tumors after 30 weeks. The effect of bile acid accumulation on initialcarcinogenesis phase (8 weeks) was investigated in bile duct ligated (BDL) animals. Serum bile acid measurementand determination of liver-specific healthy and tumor markers revealed that early thioacetamide treatment inducedhypercholanemia together with upregulation of the tumor marker Neu in bile ducts, which were enhancedby BDL. Bile acid accumulation was associated with increased expression of interleukin (IL)-6 and downregu-lation of farnesoid X receptor (FXR). Bile duct proliferation and apoptosis activation, with inverse pattern(BDL > thioacetamide þ BDL >> thioacetamide vs. thioacetamide > thioacetamide þ BDL > BDL), wereobserved. In conclusion, intrahepatic accumulation of bile acids does not induce carcinogenesis directly butfacilitates a cocarcinogenic effect due to stimulation of bile duct proliferation, enhanced inflammation, andreduction in FXR-dependent chemoprotection.

Implications: This study reveals that bile acids foster cocarcinogenic events that impact cholangiocarcinoma.Mol Cancer Res; 12(1); 1–10. �2013 AACR.

IntroductionThe silent symptoms and the lack of specific markers

account for a frequent late diagnosis of intrahepatic cholan-giocarcinoma, which contributes to a bad prognosis of thesepatients because tumor progression is already advanced andliver resection is usually not recommended. In addition, chol-angiocarcinoma has a poor response to alternative therapeuticoptions, such as radiotherapy and chemotherapy (1, 2).

Although in many cases the etiology of cholangiocarci-noma remains unknown, some risk factors, such as primarysclerosing cholangitis, hepatolithiasis, parasitic infections,choledocal cysts, and carcinogen exposure, have been asso-ciated with its development (1). All these conditions sharethe characteristic of a certain degree of liver injury due tochronic cholestasis and inflammation. Indeed, cholangio-carcinoma is often associated with chronic cholestasis and itis well known that exposure to high concentrations of bileacids might be associated with an increased incidence ofgastrointestinal tumors (3). However, the role of bile acids incholangiocarcinogenesis remains poorly understood. Thereis evidence of the relationship, probably through the inter-action with the nuclear receptor FXR (farnesoid X receptor)or the membrane-bound receptor TGR5, between bile acidsand the proliferation and survival of biliary epithelial cells(4). Moreover, several studies in animals support the rela-tionship between cholestasis and cholangiocarcinoma pro-gression (5, 6). On the other hand, FXR, whose expression isdecreased in cholangiocarcinoma (7), is involved in liver cellschemoprotection against genotoxic compounds (8).To better understand the biology of intrahepatic cholan-

giocarcinoma, several experimental models have been

Authors' Affiliations: 1Laboratory of Experimental Hepatology and DrugTargeting (HEVEFARM), University of Salamanca, IBSAL, Salamanca,Spain; 2Department of Liver and Gastrointestinal Diseases, BiodonostiaResearch Institute, Donostia University Hospital, San Sebastian, Spain;3IKERBASQUE (Basque Foundation for Science), University of BasqueCountry (UPV/EHU), San Sebastian, Spain; and 4National Institute for theStudy of Liver and Gastrointestinal Diseases (CIBERehd, Carlos III HealthInstitute), Madrid, Spain.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Corresponding Authors: Jose J.G. Marin, Department of Physiology andPharmacology, Campus Miguel de Unamuno E.D. 37007-Salamanca,Spain. Phone: 34-923-294674; Fax: 34-923-294669; E-mail:[email protected]; and Rocio I.R. Macias, [email protected].

doi: 10.1158/1541-7786.MCR-13-0503

�2013 American Association for Cancer Research.

MolecularCancer

Research

www.aacrjournals.org OF1

on May 6, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst November 19, 2013; DOI: 10.1158/1541-7786.MCR-13-0503

http://mcr.aacrjournals.org/

developed. Some of them are based on the treatment withthioacetamide, a thiono-sulfur–containing compound,which in rodents causes acute liver injury after a single dose(9), but cirrhosis and cholangiocarcinoma developmentafter chronic administration (10–12). Thioacetamide hasbeen shown to require metabolic activation by hepaticcytochrome P450 CYP2E1 to give thioacetamide sulfoxide(TASO) and subsequently the highly reactive sulfodioxide(TASO2) (13), which covalently binds to cellular macro-molecules and accounts for thioacetamide toxicity.In this study, the cholangiocarcinogenic effect of thioa-

cetamide, alone or in combination with bile acid retentioninduced by bile duct ligation (BDL), has been investigated.Moreover, the ability of bile acids to either induce DNAdamage directly or to enhance or prevent the effect ofgenotoxic agents was determined.

Materials and MethodsAnimals and experimental groupsMale Wistar rats (280–300 g; University of Salamanca

Animal House, Salamanca, Spain), fed on standard rodentchow (Panlab) and water ad libitum, and maintained on 12-hour cycles of light and dark, were divided into four groups:(i) control, (ii) thioacetamide (Sigma-Aldrich Quimica), (iii)BDL, and (iv) thioacetamide þ BDL. All protocols wereapproved by the University of Salamanca Ethical Commit-tee. BDL was carried out as previously described (14). In thethioacetamideþ BDL group, thioacetamide administration(0.05% in the drinking water) started the day after BDL. Forcomparative purposes some animals received thioacetamidefor 30 weeks, the time needed for the development ofcholangiocarcinomas. Samples were collected from sodiumpentobarbital (50 mg/kg body weight, intraperitoneally)anesthetized animals.Small liver samples or dissected tumors were immediately

placed in the RNAlater RNA stabilization reagent (LifeTechnologies). Additional samples were immersed in liquidnitrogen and stored at �80�C until used for Western blotanalysis or histologic analysis. Blood was collected from theinferior cava vein and the serum obtained was stored at�20�C for further use.

Rat hepatocytesHepatocytes were isolated from male Wistar rats (220–

250 g) as reported previously (15). To obtain conditionedmedium containing thioacetamide metabolites, primary cul-tured hepatocytes were incubated for 24 hours with mediumcontaining different concentrations (5–75 mmol/L) of thioa-cetamide. Then, cells were trypsinized and resuspended in100 mL water to obtain cell lysates, which were immediatelyfrozen in liquid nitrogen and stored at �80�C until use.

Cloning procedures, bacterial cultures, and genotoxicityassaysA 110-bp sequence located in the 50-flanking region of

the RecA gene, previously identified as the proximal RecApromoter (RecApr; GenBank accession number X55552),

was selected for cloning using genomic DNA from E. coli.Multisite Gateway cloning ( Life Technologies) was usedto obtain the expression vector containing the fireflyluciferase gene (Luc2; Supplementary Table S1). All plas-mids underwent gel electrophoresis–based sequencing.E. coli cells (BL21-DE3 strain) were transformed with the

RecApr–Luc2 plasmid and cultured in Luria-Bertani brothmedium supplemented with ampicillin (100 mg/mL). Frac-tions (5mL) of exponential-phase culture grown to OD600¼0.3 at 37�C were distributed into ventilated test tubes con-taining cisplatin, thioacetamide, or bile acids at the desiredconcentration. All compounds were from Sigma-AldrichQuimica. Cisplatin and lithocholic acid (LCA)were dissolvedin dimethyl sulfoxide (�0.2% final concentration), thioace-tamide, deoxycholic acid (DCA), chenodeoxycholic acid(CDCA), glycochenodeoxycholic acid (GCDCA), and gly-coursodeoxycholic acid (GUDCA) were dissolved in water,and rat hepatocyte lysates in culture medium. The bacteriawere incubated with the compounds at 37�Cwith shaking for2.5 hours.The promoter luciferase activity wasmeasuredwiththe Steady-Glo Luciferase Assay System (Promega) in a LAS-4000 image reader (FujiFilm; TDI).

Measurement of steady-state mRNA levelsTotal RNA from rat liver samples was isolated according to

a previously described protocol (14). Retrotranscription wascarried out using the SuperScript VILO cDNA Synthesis Kit(Life Technologies). Real-time quantitative PCR (RT-qPCR) was performed using AmpliTaq Gold polymerase(Applied Biosystems) in an ABI Prism 7300 SequenceDetection System (Applied Biosystems; SupplementaryTable S2). The mRNA levels of each target gene werenormalized with the expression levels of b-actin mRNA.Detection of amplified products was carried out using SYBRGreen I. Total RNA from control liver was used as acalibrator. Expression levels were calculated as 2�DDCt , whereDCt was the difference of Ct between the target gene and thenormalizer in each sample. This was used to calculate DDCtas the difference of this value between the control RNA andexperimental groups.

Apoptosis assaysApoptosis was evaluated by measuring caspase-3 and -8

activities using Ac-DEVD-AMC and Ac-IETD-AFC (EnzoLife Sciences), as previously reported (16), which werenormalized by protein concentrations measured usingbovine serum albumin as standard. The mRNA abundancesof the Bax-a proapoptotic gene and the Bcl-2 antiapoptoticgene were determined by RT-qPCR.

Western blot analysesImmunoblotting analyses of liver lysates were carried out in

7.5% to 10%SDS–PAGE, loading 100mg of protein per lane.Primary antibodies were diluted in PBS–Tween (Supplemen-tary Table S3). Immunoreactive protein bands were visualizedby an enhanced chemiluminescent (ECL) detection system(Amersham Pharmacia Biotech) after incubation with appro-priate secondary antibodies (IgG–HRP-linked).

Lozano et al.

Mol Cancer Res; 12(1) January 2014 Molecular Cancer ResearchOF2




Histologic analysesOf note, 5 mm thick rat liver cryosections (3 pieces from

different lobes) were fixed with neutral buffered formalin for20 minutes and stained with hematoxylin/eosin. Immunos-taining was performed on cryosections air-dried and fixed incold methanol. Primary antibodies were diluted in 2% fetalcalf serum in PBS (Supplementary Table S3) and secondaryanti-mouse Alexa-488 (Invitrogen) 1:1,000. Nuclei werestained with Dapi (10 mmol/L). Confocal laser-scanningimmunofluorescence microscopy was performed using aZeiss LSM 510 apparatus.

Biochemical and statistical analysesRoutine biochemical parameters were determined in

serum as previously described (14). Total serum bile acidconcentrations were assayed by an enzymatic/fluorimetricmethod (14, 17). Data are presented as means� SEM. Afterperforming an ANOVA test, the Bonferroni method ofmultiple-range testing was used to calculate the statisticalsignificance of differences among groups.

ResultsIn vitro genotoxicity testLuciferase activity in bacteria transformedwith the recom-

binant RecApr–Luc2 plasmid was measured after their incu-bation with cisplatin, used here as a typical DNA-damagingagent, for 2.5 hours. This time was chosen based on previousstudies on the time-course of promoter activity versusbacterial growth (data not shown). A concentration-depen-dent stimulation of RecApr activity at nontoxic cisplatinconcentrations was observed (Supplementary Fig. S1A andS1B). To confirm the cisplatin-induced activation of thebacterial SOS response, RecA expression was measured byRT-qPCR. This revealed a 2-fold increase in RecA mRNAlevels when the bacteria were incubated with 50 mmol/Lcisplatin (data not shown).Using this approach, the effect of bile acids at nontoxic

concentrations for these bacteria was investigated. Theresults from one of the concentration-response determi-nation using DCA are shown in Supplementary Fig. S1Cand S1D. Similar experiments were carried out withCDCA, GCDCA, and LCA (Fig. 1A–C). None of thesebile acids were able to induce an increase in RecApr–Luc2activity. Because thioacetamide metabolites are believed tobe involved in genotoxicity, accounting for the carcino-genic effect of thioacetamide treatment, both pure thioa-cetamide and thioacetamide metabolites contained in thelysate of hepatocytes previously incubated with thioace-tamide were evaluated (Fig. 1D and E). Conditionedmedium, but not pure thioacetamide, was able to enhanceRecApr–Luc2 activity. Finally, the ability of a potentiallyhepatotoxic bile acid, such as GCDCA, and a hepatopro-tective bile acid such as GUDCA, to enhance or reduce,respectively, DNA damage induced by a genotoxic agentsuch as cisplatin, was investigated. Neither GCDCA norGUDCA were able to modify the ability of cisplatin toenhance RecApr–Luc2 activity (Fig. 1F).

In vivo cholangiocarcinoma induction modelThe protocol used here to induce cholangiocarcinoma

caused a significant reduction in body weight (Supplemen-tary Fig. S2A). In animals treated with thioacetamide for 8weeks, this loss resulted in 40% lower body weight thancontrols of the same age (Supplementary Fig. S2A). On theother hand, as previously described (14), 8-week BDLslightly reduced weight gain. Body weight was only 15%lower than in controls of the same age. However, BDLinduced higher mortality (50%) than thioacetamide (Sup-plementary Fig. S2B), which precludes longer experimentalprotocols using BDL. In thioacetamideþ BDL group, bodyweight was lower than in thioacetamide group and themortality was similar to that found in BDL group. Tumorswere visible after thioacetamide administration for 30 weeks(Supplementary Fig. S2C). The histology revealed a cavern-ous aspect typical of cholangiocellular adenomas, whichwereadjacent to invasive intestinal-type and mucin-producingcholangiocarcinoma (Supplementary Fig. S2D). In animalsof 8-week thioacetamide group themacroscopic aspect of theliver was normal (data not shown). However, in comparisonwith control livers (Supplementary Fig. S2E) the histologicanalysis revealed areas of bile duct proliferation and leuko-cyte infiltration with important interindividual variabilityamong animals about the extension of the damage (Supple-mentary Fig. S2F and S2G). Bile duct proliferation wasmoreevident in BDL group (Supplementary Fig. S2H and S2I).As previously described (14), 8-week BDL induced a severedisorganization of the liver parenchyma (Supplementary Fig.S2H and S2I). In the present study, �25% of the paren-chyma was replaced by bile duct–like structures surroundedby fibrotic stroma. Leukocyte infiltration and nodules ofdegenerated tissue rich in inflammatory cells were alsoobserved as from 4 weeks after BDL (14). Smaller areasof bile duct proliferation together with areas of cell degen-eration and fibrosis (Supplementary Fig. S2J), and bileaccumulation (Supplementary Fig. S2K) were observed inthioacetamide þ BDL group.

Serum markersThe hepatotoxic effect of thioacetamide was confirmed

by changes in routine serum markers of liver function(Table 1). As expected, markers of cholestasis wereincreased in serum of BDL groups either with or withoutthioacetamide treatment. Bile acid concentrationsincreased 30-fold after 4 weeks of BDL (both in BDLand thioacetamide þ BDL groups), and 9-fold (BDL) or15-fold (thioacetamide þ BDL) after 8 weeks. Onlygamma-glutamyl transferase (GGT) was higher in thethioacetamide þ BDL than in the BDL group. Signs ofrenal damage such as increased levels of serum urea (Table 1)were observed in the BDL group but, surprisingly, not inthioacetamide þ BDL group.

Cell proliferation and differentiationThe expression of cholangiocarcinomamarkers cytokeratin

7 (CK-7) and claudin-4 (18, 19) increased in parallel in allexperimental groups (Fig. 2A and B, respectively). A trend

Bile Acids in Cholangiocarcinogenesis

www.aacrjournals.org Mol Cancer Res; 12(1) January 2014 OF3




toward a progressive increase in both markers over time wasobserved in the BDL group. The change was significantly(P < 0.05) less in thioacetamideþ BDL than in BDL group.Consistent with the histologic results, the administration ofthioacetamide alone for 8 weeks induced a mild increase in

the expression of CK-7 and claudin-4, whereas well-estab-lished cholangiocarcinoma presented elevated levels of bothmarkers (Fig. 2A and 2B). These changes in mRNA abun-dance were consistent with changes in protein levels assayedby Western blot analysis (Fig. 2F).

Figure 1. Determination ofgenotoxic effect. Luminescentsignal of E. coli bacteriatransformed with a RecApr-fireflyluciferase (Luc2) fusion plasmid 2.5hours after incubation of bacteriawith CDCA (A), GCDCA (B) or LCA(C), thioacetamide (TAA) (D), orlysates of rat hepatocytes obtainedafter being incubated with theindicatedconcentrationsof TAA for24 h (TAA metabolites) (E). Effect ofGCDCA and GUDCA on thegenotoxic effect of cisplatin (F).Values aremeans�SEM (from fourassays performed in triplicate).�, P < 0.05; ��, P < 0.01, ascompared with controls.

Table 1. Effect of thioacetamide treatment (0.05% in drinking water), BDL, and combination of both(thioacetamide þ BDL) on biochemical markers in rat serum of liver and kidney function

Thioacetamide BDL Thioacetamide þ BDL

Control 4 weeks 8 weeks 4 weeks 8 weeks 4 weeks 8 weeks CCA

Glucose (mg/dL) 67 � 7 80 � 4 94 � 6 82 � 6 84 � 10 75 � 9 87 � 10 107 � 4AST (U/L) 93 � 14 283 � 98a 85 � 17 173 � 79 233 � 22a 195 � 36a 132 � 32 137 � 19ALT (U/L) 55 � 8 167 � 59a 49 � 5 66 � 16 54 � 5 45 � 7 44 � 1 42 � 12GGT (U/L) 5 � 0.5 10 � 1a 14 � 6a 12 � 7a 41 � 7a 203 � 25a 91 � 68a 62 � 12a

AP (U/L) 104 � 9 122 � 11 168 � 45 201 � 44a 299 � 55a 199 � 38a 232 � 10a 104 � 7Bile acids (mmol/L) 11 � 3 87 � 17a 57 � 17a 367 � 32a 94 � 21a 310 � 100a 153 � 40a 94 � 12a

Bilirubin (mg/dL) 0.2 � 0.1 0.4 � 0.2 0.3 � 0.1 2.1 � 2a 5 � 0.2a 5.6 � 2.2a 5 � 2.5a 0.9 � 0.3a

Urea (mg/dL) 38 � 2 48 � 2 43 � 4 35 � 5 103 � 33a 53 � 8 40 � 3 46 � 2Creatinine (mg/dL) 0.5 � 0.05 0.4 � 0.02 0.3 � 0.02 0.6 � 0.1 0.8 � 0.2 0.6 � 0.1 0.3 � 0.03 0.3 � 0.05

NOTE: Well-established CCA was used for comparative purposes. Values are mean � SEM from �5 animals per group.Abbreviations: ALT, alanine aminotransferase; AP, alkaline phosphatase; AST, aspartate aminotransferase; BDL, bile duct ligation;CAA, cholangiocarcinoma; GGT, gamma-glutamyl transferase.aP < 0.05 as compared with controls.

Lozano et al.





Glutathione S-transferase-pi (GSTpi), the predominantisoform in biliary epithelium of the phase II-detoxificationGST enzymes, is overexpressed in cholangiocarcinoma (20),but is also elevated in animal models of hepatocellularcarcinoma (15). The abundance of GSTpi mRNA (Fig.2C) and protein (Fig. 2F) increased in thioacetamide groups.In thioacetamide-induced cholangiocarcinoma, we observedan overexpression of this enzyme, higher than that found inhepatocellular carcinoma (Fig. 2F), used here as a positivecontrol. BDL alone or in combination with thioacetamidewas not able to significantly upregulate GSTpi, although atrend toward a moderate increase in GSTpi mRNA wasobserved. On the other hand, the expression of GSTa, themajor GST isoform in the healthy liver parenchyma pre-

dominantly located in hepatocytes, was slightly reduced bythioacetamide administration, whereas a significant andprogressive downregulation was observed in the BDL group(Fig. 2D and 2F). GSTa mRNA was also significantlyreduced in the thioacetamide þ BDL group. In cholangio-carcinoma, the abundance of GSTa mRNA was similar tothat seen in healthy liver, but protein levels were markedlyreduced (Fig. 2F).a-fetoprotein (AFP), a commonly used marker for hepa-

tocellular carcinoma, was detected in fetal rat liver, used hereas a positive control, but not in cholangiocarcinoma andsamples obtained from any other group (Fig. 2F). Theseresults demonstrated that, under the experimental condi-tions used here, thioacetamide induced the development of

Figure 2. Determination of tumormarkers. Expression of bile ductproliferation markers in sham-operated rats (Control) and animalswith thioacetamide treatment (TAA)for 4 or 8 weeks, BDL 4 or 8 weeksbefore sample collection, or both(TAA þ BDL). For comparison,cholangiocarcinoma (CCA)induced by thioacetamidetreatment for 30 weeks is alsoincluded. CK-7 (A), claudin-4 (B),GSTpi (C), GSTa (D), and Neu (E)mRNA abundance was measuredby RT-qPCR. Values are means �SEM (n � 6 analyzed in triplicate).Results were normalized to b-actinmRNA. �, P < 0.05, compared withcontrols. F, representativeWesternblot analysis of liver homogenatesfrom the same experimentalgroups. As positive controls, ratplacenta for CK-7, LS174T cells forclaudin-4, rat fetal liver for AFP, ratcontrol liver for GSTa, and rathepatocellular carcinoma forGSTpi were used.






cholangiocarcinoma, but not hepatocellular carcinoma. Thiswas confirmed by the immunolocalization of CK-7, claudin-4, and GSTpi (Fig. 3).The increased expression or activation of the EGF receptor

2 (ErbB2 or Neu) has been associated with cell invasion,

progression, and differentiation of cholangiocarcinoma (21,22). Administration of thioacetamide increasedNeumRNAlevels 2- and 5-fold after 4 and 8 weeks, respectively,although the change was only significant after 8 weeks. Thelevels of this protooncogene in thioacetamide-induced

Figure 3. Tissue localization ofmarkers. Immunofluorescence ofCK-7, claudin-4, and GSTpi inrat liver cryosections of sham-operated rats (Control), or animalsundergoing treatment withthioacetamide (0.05% in drinkingwater for 8 weeks) alone ((TAA),BDL 8 weeks before samplecollection, or TAA þ BDL. Forcomparison, an additional groupwith well-established intestinal-type cholangiocarcinoma (CCA)due to thioacetamide treatment for30 weeks was included. Bar:50 mm.

Lozano et al.





cholangiocarcinoma were increased more than 100-fold.BDL induced a marked overexpression of this gene. Com-bination of BDLwith thioacetamide did not further enhancethe stimulatory effect of BDL alone (Fig. 2E).

Apoptosis, inflammation, and chemoprotectionWhen Bax-a and Bcl-2 mRNA were measured (Fig. 4A

and B), a similar induction of Bax-a and Bcl-2 was observedin animals receiving thioacetamide alone for 4 and 8 weeks,resulting in a non-significant change in theBax-a/Bcl-2 ratio(Fig. 4C). BDL induced a marked overexpression of Bcl-2after 4 weeks, whereas the abundance of Bax-a mRNA wassimilar to that seen in control livers, which resulted in adecrease in the Bax-a/Bcl-2 ratio (Fig. 4C). In the thioace-tamideþ BDL group, a 2.5-fold increase in Bax-a togetherwith a 6-fold increase in Bcl-2 after 4 weeks were observed,whereas the results observed after 8 weeks were similar tothose found in cholangiocarcinoma, i.e., no change in Bax-amRNA but enhanced Bcl-2 levels, which resulted in areduced Bax-a/Bcl-2 ratio (Fig. 4C), suggesting a prevalenceof survival versus apoptosis. This was consisting with cas-pase-3 activity (Fig. 4D), which was significantly enhancedonly in thioacetamide groups. Caspase-8 activity was notchanged in any experimental group (data not shown).

Interleukin (IL)-6 is a multifunctional cytokine pro-duced by a wide variety of cells that mediate inflammatoryreactions. This cytokine has been found increased in thebiliary tract and the systemic circulation of patients withcholangiocarcinoma (23). In vitro studies have revealedthat IL-6 is able to stimulate cholangiolar tumor cellsgrowth (24). In the present study, IL-6 expression in livertissue samples collected from each experimental group wasdetermined to assess whether the cholangiocarcinogeniceffect of bile acid accumulation could be related withincreased levels of this cytokine. Thioacetamide adminis-tration alone for 4 and 8 weeks did not induce changes inthe expression of IL-6 (Fig. 5A). In contrast this wasenhanced in all groups with BDL and also in cholangio-carcinoma (Fig. 5A).To evaluate whether impaired FXR-mediated chemopro-

tection could be involved in enhanced susceptibility tocholangiocarcinoma development, the expression of thisnuclear receptor was also determined. Thioacetamide hadno or moderate inhibitory effect on FXR expression whenwas administered alone for 4 or 8 weeks, respectively(Fig. 5B). In contrast, a dramatic decrease in FXR expressionin all BDL groups and in cholangiocarcinoma was found(Fig. 5B).

Figure 4. Apoptosis activation.Bax-a (A) and Bcl-2 (B) mRNAabundance, and caspase-3 activity(D) were measured in sham-operated rats (Control) and animalswith thioacetamide treatment (TAA)for 4 or 8 weeks, BDL 4 or 8 weeksbefore sample collection, or both(thioacetamide þ BDL). Forcomparison cholangiocarcinoma(CCA) induced by thioacetamidetreatment for 30 weeks wasincluded. The Bax-a to Bcl-2 ratio(C) was calculated from A and B.Values are means � SEM (n � 6analyzed in triplicate). Results arenormalized to b-actin mRNA.�,P<0.05, comparedwith controls.N.D., not determined due to smallsize of tumor sample.






DiscussionIt is well known that livers affected by diseases character-

ized by prolonged cholestasis and inflammation are moreprone to developing cholangiocarcinoma (1, 2), whichsuggests that bile acids could play a role in cholangiocarci-nogenesis, as has also been suggested for other gastrointes-tinal tumors (3). Results of the present study indicate thatbile acids are not primary genotoxic/carcinogenic agents butrather they are promoter/cocarcinogenic compoundsinvolved in favoring cholangiocarcinoma development.To elucidate whether direct carcinogenic effect of bile

acids could be involved, we used a strategy based on bacterialSOS response to DNA damage (25), which consists in acascade of events affecting a number of genes, with LexAplaying a key role as a repressor for all genes of this system,and RecA encoding a protein that promotes the inactivationof the repressor when the SOS-inducing signal is triggered.We observed that none of the molecular species of bile acidsassayed, nor thioacetamide, induced the SOS response inE. coli. In contrast, both cisplatin and thioacetamide meta-bolites were able to activate RecApr. Moreover, bile acids didnot enhance or prevent the genotoxic effect of cisplatin.These results are consistent with previous studies reportingthat sodium taurocholate and sodium deoxycholate arenot able to trigger the bacterial SOS response by them-selves (26). This may be due to the lack of a direct actionof bile acids on DNA. Indeed, DNA–bile acid adductshave not been observed either after in vitro treatment ofhuman cancer cell lines or in in vivo studies in rats (27).Using bacterial tests, bile acids at very high concentrations(1–10 mmol/L) have been found to enhance the effectinduced by fecal mutagens (26). Moreover, bile acid–induced DNA damage was found when colonocytes wereincubated with DCA or CDCA. In another study, this wasattributed not to a direct effect but rather to the generationof oxidative stress due to the high concentration of bileacids (600 mmol/L) used (28). Our results with several bileacid species at levels ranging from subtoxic (50 mmol/L) tomoderately toxic (1,000 mmol/L) concentrations support

the concept that a direct mutagenic effect of these com-pounds is unlikely.About the in vivo experimental model used here to

evaluate the promoter effect of bile acid accumulation incholangiocarcinoma development, we found amarked inter-individual variability in the carcinogenic response to thioa-cetamide, which may account for the discrepancies previ-ously reported on the onset of cholangiocarcinoma inrodents using long-term administration of this carcinogenicagent (10–12, 29). Both cholangiocarcinoma and hepato-cellular carcinoma have been reported to appear in the liverof rodents treated with thioacetamide (30, 31). Our resultsconfirmed the ability of thioacetamide to induce cholangio-carcinoma but, at least in our experimental setting, noinduction of hepatocellular carcinoma was observed.Cholestasis alone induced massive bile duct proliferation

together with upregulation of Neu and resistance to apop-tosis. Although the combination of BDLwith thioacetamidefor 8 weeks also induced a ductal response, this was weakerthan that induced by BDL alone. A reduction in BDL-induced bile duct proliferation has also been reported whencombining cholestasis with other hepatotoxins such ascarbon tetrachloride (32). The administration of thioaceta-mide alone during the same time, however, enhancedGSTpiexpression andmaintained both the apoptosis resistance andthe overexpression of Neu. Interestingly, this has beenobserved in small hyperplastic biliary ducts in some casesof hepatolithiasis, and in most cases of primary sclerosingcholangitis (21). One possible explanation for this observa-tion is that the toxic effect of thioacetamide could be partiallyovercome by the stimulation of cell proliferation induced byBDL.Thioacetamide administration induced both apoptosis

and survival signals, with a trend to increase the Bax-a/Bcl2ratio, which may be interpreted in the sense that apoptosisprevailed over survival. However, in animals treated withthioacetamide for 30 weeks clear signs of overcoming apo-ptosis in tumor cells were observed, an event that was alsoseen much earlier when thioacetamide was combined with

Figure 5. Expression of IL-6 andFXR. IL-6 (A) and FXR (B) mRNAabundance was measured insham-operated rats (Control) andanimals with thioacetamide (TAA)treatment for 4 or 8weeks,BDL4or8 weeks before sample collection,or both (thioacetamideþ BDL). Forcomparison cholangiocarcinoma(CCA) induced by thioacetamidetreatment for 30 weeks wasincluded. Values aremeans�SEM(n � 6 analyzed in triplicate).Results are normalized to b-actinmRNA. �, P < 0.05, compared withcontrols.

Lozano et al.





BDL. Thus, cholangiofibrosis, a lesion that generally pro-gresses toward the formation of cholangiocarcinoma, wasclearly observed only in thioacetamideþBDL combination.Thus, in the light of the classical three-stage model ofchemically induced liver carcinogenesis our results indicatethat bile acids may act as cell growth stimulating agents,which need the presence of a DNA-damaging initiator topromote cancer development. This promoter effect of bileacids may be due to the inhibition of apoptosis and/or tothe activation of cell proliferation, which may increase thesusceptibility to mutagens and the appearance of neoplastictransformation (33). Several mechanisms have been pro-posed to account for enhanced cell proliferation. Theseinclude activation of protooncogenes, such as c-fos andc-jun (34), and the stimulation of cyclooxygenase-2 activity(35), with the consequent increase in prostaglandin E2production, which is involved in inducing the resistance ofmalignant cells to apoptosis and enhancing their invasive-ness. Our results indicated that bile acid accumulationmightfavor cholangiocarcinoma development by stimulation ofproinflammatory mechanisms and by impairing FXR-medi-ated chemoprotection against genotoxic insults.In conclusion, the accumulation of bile acids in liver tissue

seems to have no direct carcinogenic effect but may favorcholangiocarcinoma development by playing a cocarcino-genic role due to stimulation of bile duct proliferation,enhanced inflammation and reduction in FXR-dependentchemoprotection.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: E. Lozano, L. Sanchez-Vicente, M.J. Monte, E. Herraez,O. Briz, J.M. Banales, J.J.G. Marin, R.I.R. MaciasDevelopment of methodology: E. Lozano, L. Sanchez-Vicente, M.J. Monte,E. Herraez, O. Briz, J.J.G. Marin, R.I.R. MaciasAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): E. Lozano, L. Sanchez-Vicente, M.J. Monte, E. Herraez, O. Briz,R.I.R. MaciasAnalysis and interpretation of data (e.g., statistical analysis, biostatistics, compu-tational analysis): E. Lozano, L. Sanchez-Vicente, M.J. Monte, E. Herraez, O. Briz,J.M. Banales, R.I.R. MaciasWriting, review, and/or revision of the manuscript: E. Lozano, L. Sanchez-Vicente,M.J. Monte, E. Herraez, O. Briz, J.M. Banales, J.J.G. Marin, R.I.R. MaciasAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): E. Lozano, L. Sanchez-Vicente, M.J. Monte, E. Herraez,O. Briz, R.I.R. MaciasStudy supervision: E. Lozano, L. Sanchez-Vicente, E. Herraez, O. Briz, J.M. Banales,J.J.G. Marin, R.I.R. Macias

AcknowledgmentsThe authors thank L. Mu~noz, J.F. Martin, and J. Villoria for care of the animals.

Technical help byN.Gonzalez and revision of the English spelling, grammar, and styleof the article by N. Skinner are also gratefully acknowledged.

Grant SupportThis study was supported by the Spanish "Instituto de Salud Carlos III" (grants

FIS PI11/00337 and PI12/00380), Ministerio de Ciencia e Innovaci�on (grantSAF2010-15517), "Junta de Castilla y Le�on" (grants SAN673SA07/08, BIO39/SA27/10, SA070A11-2, SA023A11-2, BIO/03/SA23/11, SA015U13, BIO/SA64/13and BIO/SA65/13), "Fundaci�on Investigaci�on Mutua Madrile~na" (Call 2009), and"Asociaci�on Espa~nola Contra el C�ancer" (AECC). The group belongs to the SpanishNetwork for Cooperative Research on membrane Transport Proteins (REIT) andCIBERehd. E. Lozano andL. Sanchez-Vicente were supported by PhD grants from the"Fondo Social Europeo/Junta de Castilla y Le�on."

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be herebymarked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Received September 24, 2013; revised October 29, 2013; accepted October 30,2013; published OnlineFirst November 19, 2013.

References1. de Groen PC, Gores GJ, LaRusso NF, Gunderson LL, Nagorney DM.

Biliary tract cancers. N Engl J Med 1999;341:1368–78.2. Sirica AE, Nathanson MH, Gores GJ, Larusso NF. Pathobiology of

biliary epithelia and cholangiocarcinoma: proceedings of the Henry M.and Lillian Stratton basic research single-topic conference. Hepatol-ogy 2008;48:2040–6.

3. Bernstein H, Bernstein C, Payne CM, Dvorak K. Bile acids as endog-enous etiologic agents in gastrointestinal cancer. World J Gastroen-terol 2009;15:3329–40.

4. Xia X, Francis H, Glaser S, Alpini G, LeSage G. Bile acid interactionswith cholangiocytes. World J Gastroenterol 2006;12:3553–63.

5. Thamavit W, Pairojkul C, Tiwawech D, Itoh M, Shirai T, Ito N.Promotion of cholangiocarcinogenesis in the hamster liver by bileduct ligation after dimethylnitrosamine initiation. Carcinogenesis1993;14:2415–7.

6. Yang H, Li TW, Peng J, Tang X, Ko KS, Xia M, et al. A mouse model ofcholestasis-associated cholangiocarcinoma and transcription factorsinvolved in progression. Gastroenterology 2011;141:378–88.

7. Martinez-Becerra P, Vaquero J, Romero MR, Lozano E, Anadon C,Macias RI, et al. No correlation between the expression of FXR andgenes involved in multidrug resistance phenotype of primary livertumors. Mol Pharm 2012;9:1693–704.

8. Vaquero J, Briz O, Herraez E, Muntane J, Marin JJ. Activation of thenuclear receptor FXR enhances hepatocyte chemoprotection and livertumor chemoresistance against genotoxic compounds. Biochim Bio-phys Acta 2013;1833:2212–9.

9. Chen TM, Subeq YM, Lee RP, Chiou TW, Hsu BG. Single doseintravenous thioacetamide administration as a model of acute liverdamage in rats. Int J Exp Pathol 2008;89:223–31.

10. Cruz A, Padillo FJ, Torres E, Navarrete CM, Mu~noz-Casta~neda JR,Caballero FJ, et al. Melatonin prevents experimental livercirrhosis induced by thioacetamide in rats. J Pineal Res 2005;39:143–50.

11. Praet MM, Roels HJ. Histogenesis of cholangiomas and cholangio-carcinomas in thioacetamide fed rats. Exp Pathol 1984;26:3–14.

12. Al-Bader A, Mathew TC, Abul H, Al-Sayer H, Singal PK, Dashti HM.Cholangiocarcinoma and liver cirrhosis in relation to changes due tothioacetamide. Mol Cell Biochem 2000;208:1–10.

13. Hunter AL, Holscher MA, Neal RA. Thioacetamide-induced hepaticnecrosis. I. Involvement of the mixed-function oxidase enzyme sys-tem. J Pharmacol Exp Ther 1977;200:439–48.

14. Macias RI, Hierro C, de Juan SC, Jimenez F, Gonzalez-San Martin F,Marin JJ. Hepatic expression of sodium-dependent vitamin C trans-porters: ontogeny, subtissular distribution and effect of chronic liverdiseases. Br J Nutr 2011;106:1814–25.

15. Monte MJ, Dominguez S, Palomero MF, Macias RI, Marin JJ.Further evidence of the usefulness of bile acids as molecules forshuttling cytostatic drugs toward liver tumors. J Hepatol 1999;31:521–8.

16. PerezMJ,Macias RI, Marin JJ. Maternal cholestasis induces placentaloxidative stress and apoptosis. Protective effect of ursodeoxycholicacid. Placenta 2006;27:34–41.

17. MashigeF, ImaiK,OsugaT.A simple andsensitive assayof total serumbile acids. Clin Chim Acta 1976;70:79–86.

18. Shimonishi T, Miyazaki K, Nakanuma Y. Cytokeratin profile relates tohistological subtypes and intrahepatic location of intrahepatic cho-langiocarcinoma and primary sites of metastatic adenocarcinoma ofliver. Histopathology 2000;37:55–63.






19. L�odi C, Szab�o E, Holczbauer A, Batmunkh E, Szíj�art�o A, Kupcsulik P,et al. Claudin-4 differentiates biliary tract cancers from hepatocellularcarcinomas. Mod Pathol 2006;19:460–9.

20. Hayes PC, May L, Hayes JD, Harrison DJ. Glutathione S-transferasesin human liver cancer. Gut 1991;32:1546–9.

21. Endo K, Yoon BI, Pairojkul C, Demetris AJ, Sirica AE. ERBB-2 over-expression and cyclooxygenase-2 upregulation in human cholangio-carcinoma and risk conditions. Hepatology 2002;36:439–50.

22. Treekitkarnmongkol W, Suthiphongchai T. High expression of ErbB2contributes to cholangiocarcinoma cell invasion and proliferationthrough AKT/p70S6K. World J Gastroenterol 2010;16:4047–54.

23. Cheon YK, Cho YD, Moon JH, Jang JY, Kim YS, Kim YS, et al.Diagnostic utility of interleukin-6 (IL-6) for primary bile duct cancerand changes in serum IL-6 levels following photodynamic therapy. AmJ Gastroenterol 2007;102:2164–70.

24. Braconi C, Swenson E, Kogure T, Huang N, Patel T. Targeting the IL-6dependent phenotype can identify novel therapies for cholangiocar-cinoma. PLoS ONE 2010;5:e15195.

25. Walker GC.Mutagenesis and inducible responses to deoxyribonucleicacid damage in Escherichia coli. Microbiol Rev 1984;48:60–93.

26. Nair PP, Davis KE, Shami S, Lagerholm S. The induction of SOSfunction in Escherichia coli K-12/PQ37 by 4-nitroquinoline oxide (4-NQO) and fecapentaenes-12 and -14 is bile salt sensitive: implicationsfor colon carcinogenesis. Mutat Res 2000;447:179–85.

27. Hamada K, Umemoto A, Kajikawa A, Seraj MJ, Monden Y. In vitroformation of DNA adducts with bile acids. Carcinogenesis 1994;15:1911–5.

28. Rosignoli P, Fabiani R, De Bartolomeo A, Fuccelli R, Pelli MA, MorozziG. Genotoxic effect of bile acids on human normal and tumour coloncells and protection by dietary antioxidants and butyrate. Eur J Nutr2008;47:301–9.

29. Yeh CN, Maitra A, Lee KF, Jan YY, Chen MF. Thioacetamide-inducedintestinal-type cholangiocarcinoma in rat: an animal model recapitu-lating the multistage progression of human cholangiocarcinoma. Car-cinogenesis 2004;25:631–6.

30. Newell P, Villanueva A, Friedman SL, Koike K, Llovet JM. Experimentalmodels of hepatocellular carcinoma. J Hepatol 2008;48:858–79.

31. Heindryckx F, Colle I, Van Vlierberghe H. Experimental mouse modelsfor hepatocellular carcinoma research. Int J Exp Pathol 2009;90:367–86.

32. Marzioni M, Alpini G, Saccomanno S, Candelaresi C, Venter J,Rychlicki C, et al. Exendin-4, a glucagon-like peptide 1 receptoragonist, protects cholangiocytes from apoptosis. Gut 2009;58:990–7.

33. de Kok TM, van Maanen JM. Evaluation of fecal mutagenicity andcolorectal cancer risk. Mutat Res 2000;463:53–101.

34. Vel�azquezOC,ZhouD,SetoRW, JabbarA,Choi J, LedererHM, et al. Invivo crypt surface hyperproliferation is decreased by butyrate andincreased by deoxycholate in normal rat colon: associated in vivoeffects on c-Fos and c-Jun expression. JPEN J Parenter Enteral Nutr1996;20:243–50.

35. Glinghammar B, Inoue H, Rafter JJ. Deoxycholic acid causes DNAdamage in colonic cells with subsequent induction of caspases, COX-2 promoter activity and the transcription factors NF-kB and AP-1.Carcinogenesis 2002;23:839–45.

Lozano et al.





Published OnlineFirst November 19, 2013.Mol Cancer Res Elisa Lozano, Laura Sanchez-Vicente, Maria J. Monte, et al. Accumulation in Cholangiocarcinoma DevelopmentCocarcinogenic Effects of Intrahepatic Bile Acid

Updated version

10.1158/1541-7786.MCR-13-0503doi:

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://mcr.aacrjournals.org/content/early/2013/12/30/1541-7786.MCR-13-0503To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-13-0503

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mcr.aacrjournals.org/content/early/2013/12/30/1541-7786.MCR-13-0503


Cocarcinogenic Effects of Intrahepatic Bile Acid …...2013/12/30 · and the proliferation and...

Documents

Transcript of Cocarcinogenic Effects of Intrahepatic Bile Acid …...2013/12/30 · and the proliferation and...