[CANCERRESEARCH57, 1743—1749.May 1, 1997!

ABSTRACT

Cholangiocarcinoma represents a challenging primary malignancy ofthe liver with no effective medical therapy and a poor prognosis. We haveinvestigated the role of tamoxifen and estrogen receptors (ERa) in theregulation of growth of human cholangiocarcinoma. Two human cholangiocarcinoma cell lines, OZ and SK-ChA-1, were grown in the presence of

graded concentrations of tamoxifen; the effects on cell growth were determined by cell counting or 3-(4,5-dimethylthiazol-2y1)-2,S-diphenyltet

razolium proliferation assay. The presence of ER protein was tested byIndirect hnmunofluorescence and immunoprecipitation. In addition, ceHswere grown in estrogen-depleted media supplemented with exogenous17@-estradiol. ER mRNA was evaluated by reverse transcription-PCRand Northern blotting. Finally, one cholangiocarcinoma cell line wasgrown as a xenograft In athymic nude mice; tamoxifen effects on in vivo

tumor growth were determined with biweekly caliper measurements.Tamoxifen (5—10,.tM)caused dose-dependent in vürogrowth inhibition oftwo human cholangiocarcinoma cell lines. In addition, growth inhibitionof one cell line (SK-ChA-1) grown as a xenograft in nude mice by tamoxifen was observed. The presence of ER protein was suggested by 17@J-estradlol stimulation of tumor cell growth in vitro and confirmed byimmunoprecipitation. Immunofluorescence microscopy was ineffective atdetection of ER protein. Reverse transcriptlon-PCR demonstrated thepresence ofER mRNA in both cell lines. Northern blot analysis confirmedthe presence of full-length 6.5-kb ER mRNA. No ER deletion mutantswere detected. Tainoxifen inhibited the growth of human cholangiocarcinoma in vitro and in vivo. ER protein and mRNA were detected in both cell

lines. The mechanism(s) of tamoxifen-medlated growth inhibition is nit

clear but may occur via ER protein or additional pathways The ability oftamoxifen to inhibit tumor growth may offer an alternative adjunctivetreatment for cholangiocarcinoma.

INTRODUCTION

Cholangiocarcinoma is a malignant tumor of the biliary tree that

originates from the bile duct epithelial cell or cholangiocyte. Cholangiocarcinoma is an increasingly frequent diagnosis worldwide; in theUnited States, approximately 3000 new cases are reported each year.There is currently no effective medical, chemotherapeutic, or radiation therapy available; thus, the disease has a poor prognosis (1—5).

The etiology of cholangiocarcinoma is unknown; however, severalwell-described associations are reported as increased risks for thedevelopment of cholangiocarcinoma. These include chronic liverfluke infestation, exposure to various chemicals, and certain congenit.al diseases of the biliary tree (1, 4, 6—8).The association of PSC3with cholangiocarcinoma is also now well recognized (1 , 4, 9). Thecommon underlying theme for these associations is that chronicinflammation of the biliary tree leads to an increased risk of cholangiocarcinoma development.

Received 10/15/96; accepted 3/1/97.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

I This research was partially funded by Grant DK 50995 from the NIH and Grant

RRG6636 from the NIH Comprehensive Cancer Center.2 To whom requests for reprints should be addressed, at University of Alabama at

Birmingham, 405 Kracke Building, 1922 Seventh Avenue, South, Birmingham, AL35294-0005.

3 The abbreviations used are: PSC, primary sclerosing cholangitis; ER, estrogen

receptor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RT-PCR. reverse transcription-PCR; TGF, transforming growth factor.

Tamoxifen, an anti-estrogen, has clear efficacy in the treatment of

women with ER-positive breast cancer (10, 11). In addition, subgroups of ER-negative breast cancer patients also respond to tamoxifen therapy (1 1, 12). Recent studies have also shown that tamoxifenhas efficacy in the treatment of central nervous system tumors such asmalignant gliomas and craniopharyngiomas (13—15).On the basis ofthe poor survival record and the lack of effective therapy for cholangiocarcinoma and our desire to find an alternative form of therapy for

biliary tract malignancy, we investigated the role of tamoxifen and itspotential efficacy in slowing the growth of human cholangiocarcinoma cell lines. The purpose of this investigation was to addresswhether tamoxifen inhibits the in vitro and in vivo growth of humancholangiocarcinoma and to initiate studies on the mechanism(s) oftumor growth inhibition.

MATERIALS AND METHODS

Cell Lines. Two separatelyderivedhumancholangiocarcinomatumorcelllines, designated OZ and SK-ChA-l, were provided by Dr. N. F. LaRusso(Mayo Clinic, Rochester, MN) and Dr. A. Knuth (Ludwig Institute for CancerResearch, London, United Kingdom). The original characterization of these

cell lines has been described (16, 17). MCF-7 (human breast cancer cell line),HEP-G2 (a human hepatoma cell line), PANC-l (a human pancreatic cancercell line), and COS cells (negative control) were purchased from the AmericanType Culture Collection (Rockville, MD). Cells were routinely grown in RPMI

1640 (Cellgro; Fisher Scientific, Atlanta, GA) supplemented with 2 m@tL-glutamine, 5 units/mI penicillin, 5 @.tg/mlstreptomycin, and 5—10%heatinactivated FCS. Cell lines were screened routinely for Mycoplasma contamination either by DNA staining with the Hoescht dye (Sigma Chemical Co., St.

Louis, MO) or a commercially available Mycoplasma detection kit (Boehringer

Mannheim, Indianapolis, IN).Chemicals/Molecular Probes. Tamoxifen and l7@-estradiol were pur

chased from Sigma and dissolved in 95% ethanol. A human ER cDNA probewas purchased from the American Type Culture Collection. A human GAPDHcDNA probe for Northern blotting was provided by Dr. G. Alpini (Scott andWhite Clinic, Temple, TX). Based upon published sequences ( 18), oligonucleotide primers specific for human ER and (3-actin used in RT-PCR weresynthesized at the Oligonucleotide Core Facility at University of Alabama atBirmingham. A monoclonal antibody to the ER (H226) used for immunoprecipitation studies was purchased from Abbott Laboratories (Abbott, IL). TheDAB Quick Staining antibody kit was purchased from Dako (Carpinteria, CA)to visualize the ER protein bands.

Mice. Athymic (nu/nu) female BALB/c mice, 6—8weeks of age, werepurchased from Charles River Laboratories for tumor inoculation. All animals

were maintained in a sterile environment; cages, bedding, food, and water wereautoclaved, and animals were maintained on a daily 12-h Iight/l2-h dark cycle.

Cell Proliferation Assays. Cell proliferationwas assessed eitherby directcell counting with a Coulter counter (Hialeab, FL) or a commercially availablecolorimetric cell proliferation assay (CellTiter 96AQueous assay; PromegaCorp., Madison, WI). For the colorimetric proliferation assay, cells were platedin quadruplicate into 96-well tissue culture plates (Nunc; Fischer Scientific) at

5—10,000cells/well in a final volume of 200 pi. Cells were allowed to adhereovernight, medium containing various inhibitors was added, and cell countswere performed at various times.

Estradiol Stimulation of Cell Growth. Cell lines were maintained inestrogen-depleted medium (phenol red-free RPMI 1640 and charcoal-treatedFCS) for 7 days before seeding at 5 X l0cells/well (24-well plates; Nunc) inestrogen-depleted medium. Cells were stimulated with l0@, 10@ and 10M 17f3-estradiol for 7 days (performed in triplicate). Controls included a carriercontrol alone without estradiol (methanol) and FCS that had not been treated

1743

Tamoxifen-mediated Growth Inhibition of Human Cholangiocarcinoma1

Lorenzo K. Sampson, Selwyn M. Vickers, Weizhong Ying, and John 0. Phillips2

Department of Surgery IL K. S.. £M. V.1 and Division of Gastroenterology and Hepatology. Department of Medicine fW. Y.. J. 0. P.]. The University of Alabama atBirmingham, Birmingham, Alabama 35294-IXXJ7

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

TAMOXIFEN INHIBITION OF HUMAN CHOLANGI@ARCINOMA

with charcoal. After 1 week, cells were collected by trypsinization and countedon a Coulter counter. Results are expressed as the mean cell number withunstimulated control cells grown in estrogen-depleted medium normalized to100%.

RNA Extraction. Total RNA was extracted from cell lines by the single

step method (19). Briefly, cultured cells were homogenized in a denaturingsolution containing 4 M guanidinium thiocyanate followed by mixing with

0. 1X volume 2 M sodium acetate, pH 4.0. An equal volume of phenol wasadded, vortexed, and followed with 0.2X volume of chloroform/isoamylalcohol. The resulting mixture was centrifuged, and the upper aqueous phasewas precipitated with isopropanol and washed with 70% ethanol. RNA concentration was determined spectrophotometrically.

RT-PCR. Total RNA (1 ,.@g)was reverse-transcribedin the presenceof 1mM dTPs (Promega), 20 mg oligo(dT) primers, 10 mM Dli', and 5 units

reverse transcriptase/ml (Life Technologies, Inc., Grand Island, NY) for 10mm at 65°C,60 mm at 42°C,and 5 mm at 90°C.Three pairs of primers wereselected that spanned either the DNA-binding region (exons 2 and 3, primers

I and 2) or part of the ligand-binding region (exon 5, primers 3 and 4; and exon6, primers 5 and 6) of the ER (I 8). ER cDNA was amplified using Taqpolymerase (Promega) for 40 cycles; each cycle included one cycle at 44°CforI mm, 55°Cfor I mm, and 72°Cfor 1 mm. Amplified DNA fragments werethen analyzed by a 2% agarose gel.

Northern Blot Analysis. Northern blot analysis was performed essentiallyas described (19). Briefly, total RNA (1.0 mg) was electrophoresed through aformaldehyde-agarose gel, blotted onto a nylon membrane (Sehleicher &Schuell, Keene, NH), and UV cross-linked. Hybridization was performed in aTechre Hybridiser HB-!0 for 60 mm at 68°Cusing a [32P]dATP-labeledERprobe in 15 ml of Quik Hyb solution (Stratagene). Plasmid ER cDNA wasisolated with a Qiagen kit (Chatsworth, CA) and digested with EcoRI (Promega); the probe was extracted with the Gene Clean II kit (Bio 101, Inc., LaJolla, CA). 5'-End labeling of probes was performed with 14 kinase (19).

Probes were incubated with b'-32PIATP 10 mCi/ml (Amersham Corp., Arlington Heights, IL) in 25 mt@iTris/HC1 (pH 7.5), 10 mM MgC12, 1 p1 of13-mercaptoethanol (diluted 1:12.6), and 10 units of T4 kinase (Promega) in a

50-@il volume at 37°C for 60 mm. Unincorporated 32P was separated from

incorporated radiolabel with a Nuctrap Probe purification column (Stratagene,

La Jolla, CA). After washing, the membrane was developed by autoradiogra

phy using KOdak XAR-5 film. Membranes were also blotted with a GAPDHprobe, which served as a control for the amount of RNA loaded to each lane(19).

Immunoprecipitation of ER Protein. Cells were lysed in high salt lysisbuffer (0.4 M KO, 20 mM HEPES (pH 7.4), 1 mM Dli', and 20% glycerol]supplemented with protease inhibitors (10 mg aprotinin/mI, 50 mg benzamindinc/mi, 50 mg leupeptin/mI, and 40 mg phenylmethylsulfonyl fluoride/mi),followed by passing through a 26-gauge needle six times. After centrifugationat 100,000 x g for 20 mm at 4°C,the supematant (cytosol) was removed, andtotal protein was determined (Micro BCA protein assay reagent kit; Pierce,Rockford, IL). The cytosol fraction was incubated with the human ER mono

clonal antibody H226 (12 @zg/ml;Abbott) followed by the addition of 133@g/mlprotein A-Sepharose (Pharmacia Biotech, Inc., Piscataway, NJ). The

pellet was centrifuged and washed with PBS. The pellet was dissolved intoloading buffer, and protein was separated on a 7.5% SDS-polyacrylamide gelfor 2 h at 100 mV. The protein was then electrophoretically transferred to anitrocellulose membrane and incubated with the monoclona! antibody to ER (1

@g/ml)for 2 h. Finally, the membrane was developed with the 3,3'-diaminobenzidine antibody kit to visualize the protein band. Controls included omission of the primary antibody (H226) and using MCF-7 and COS cells aspositive and negative controls, respectively.

Establishment of Tumor Xenografts. SK-ChA-l cells were grown inRPM! 1640media supplemented with 10%heat-inactivated FCS in T75 flasks.Cells were trypsinized, washed with Dulbecco's PBS (Cellgro), and countedwith a Coulter counter (14). Mice were anesthetized with isoflurane inhalation,and 5 X 106cells were inoculated s.c. into the flanks of mice using a 22-gaugeneedle in a total volume of 0.2 mI/site. Two weeks were allowed for tumorengraftment; the tumor engraftment rate was —90%.Animals with similartumor sizes were assigned to the tamoxifen or control group in a blindedfashion. The treatment was performed independently and in a blinded fashion.

Drug therapy was initiated using 0.! mg of tamoxifen injected i.p. three timesper week or peanut oil with solvent alone in the control group. Calipermeasurements of tumors were performed two times per week for 8 weeks in a

blinded fashion as well (20). The mean tumor volume was calculated using the

formula:

Tumor volumeme@,, (IT/6) X (mean diameter)3

The mean diameter is calculated from the average width and length ofeach xenograft in millimeters. Drug therapy and tumor measurementswere performed over a period of 8 weeks, and no animals died duringthe study. The xenografts were resected aseptically, and tumor tissuewas processed for routine histological examination by H&E stainingand for immunohistochemistry to low molecular weight cytokeratinwith a monoclonal antibody (AE-l; BioGenex, San Ramon, CA) asdescribed (21).

Statistical Analysis. Data were analyzed using a nonparametric analysis

with the Wilcoxon Rank Sum Test. P < 0.05 was considered statisticallysignificant.

RESULTS

in Vitro Growth Inhibition. After allowing human cholangiocarcinoma cell lines to adhere overnight in culture, various concentrations of tamoxifen were added. At the indicated times, cells wereeither counted directly or indirectly by the colorimetric assay. Controls included either the addition of carrier alone (methanol) or no

1.5

1.0

0.5

Fig. I . Tamoxifen-mediated growth inhibition of a humancholangiocarcinoma cell line (OZ). Tumor cells were plated andallowed to adhere to dishes overnight. Medium was removed,and graded concentrations of tamoxifen or carrier control wereadded. At various time points, cells were harvested and countedwith a Coulter counter. Each data point represents the mean ofquadruplicate wells; bars, SD. Cell viability was greater than95%, as determined by trypan blue exclusion. Concentrations oftamoxifen are indicated in the legend.

sO0

4)U

—a-— Methanol....-.-.--- TAM-i @tM

--.-....-. TAM-2.5@iM

----i--. TAM-5pM

-.-..D-.-. TAM-iOpM@1

.:“.I@@ -.-.-.-.-.@ -.-.-.-.-. -...-.-.-..D

1 2 j ‘@@@ .@

Time (days)

1744

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

TAMOXIFEN INmBmoN OF HUMAN CHOLANGIOCARCINOMA

1.25

1.00

0.75

0.50

0.25

‘00

—0---

...--.---

-----.@..

--.-@---

-.-..o..-.--@--

TAM—l@tMTAM-2.5@tMTAM—5jiM

TAM-b jiMTAM-25 jiM

Fig. 2. Tamoxifen-mediated growth inhibition of humancho!angiocarcinoma cell line (SK-ChA-l). The experiment wasperformed in a similar manner as in Fig. 1. Tamoxifen at 25 @zsiwas toxic to the cells. Bars, SD.

t----* -. -U

1 2 3 4 5 6 7

Time (days)

compound added. When cells were manually counted, the viability ofcells was always greater than 95% by trypan blue exclusion. Fig. Ishows the in vitro dose-dependent growth inhibition of a humancholangiocarcinoma cell line, OZ, induced by tamoxifen. The variouscompounds tested included methanol, the carrier control, and gradeddoses of tamoxifen from 1 to 10 nmi, a physiologically relevant dose.Data presented in a similar format with another cholangiocarcinomacell line, SK-ChA-!, are shown in Fig. 2. Higher doses of tamoxifen(25 @m)were toxic. At the 5—10@tMconcentrations of tamoxifen,there was significant and reproducible inhibition of tumor cell growth.The growth inhibition mediated by tamoxifen was not universal;specifically, the growth of HEP-G2 (human hepatoma cell line) andPANC-I (human pancreatic cancer cell line) cell lines was not observed with similar concentrations of tamoxifen (results not shown).These results show that tamoxifen caused decreased proliferation ofOz andSK-ChA-!cells.

in Vivo Growth Inhibltion For in vivo growth inhibition studies,SK-ChA-l cells were chosen because this tumor line has been successfully grown as xenografts in athymic (nude) mice (16). Cells wereinoculated onto the flanks of animals; 2 weeks were allowed for tumorengraftment. The tumor engraftment rate was —90%.The growth ofthe SK-ChA-! cholangiocarcinoma was determined by biweekly caliper measurements, and a significant increase in tumor volume wasdemonstrated in the control animal after 8 weeks. Tamoxifen wasadministered at a dose of 0.! mg i.p. three times per week for a totalof 8 weeks. Compared to control animals, tamoxifen caused a statistically significant growth inhibition in vivo of SK-ChA-1 xenograftsgrown in athymic (nude) mice (Fig. 3). There was no statistical

difference in tumor size at day zero (initiation of drug treatment)

Fig. 3. Effect of tamoxifen on the growth of human cholangiocarcinoma xenograft in nude mice. SK-ChA-l cells(2 X lIP cells/site) were injected into the flanks of nude mice.After 2 weeks, animals were injected i.p. with 0.! mg tamoxifen emulsified in peanut oil three times per week. Controlswere injected with solvent emulsified in peanut oil. Calipermeasurements were performed two times per week, and themean tumor volume was calculated as described in “Materials

and Methods.―The results represent the mean of 38 separatetumors grown as xenografts. A total of 38 separate tumors weregrown as xenografts. Statistically significant in vivo growthinhibition was observed in the tamoxifen-treated group(P = 0.02). Bars, SD.

between the two groups. However, the group of mice treated withtamoxifen for 8 weeks showed a significant reduction in tumor growthrate (P = 0.02) when using 0.05 as the significance level. Thisinhibition was observed in two separate experiments and performedwith 2! mice bearing a total of 38 tumors. Importantly, this inhibitionwas achieved with a physiologically relevant dose of tamoxifen (22).

Microscopic evaluation of tumor xenografts with H&E stainingrevealed a moderately differentiated adenocarcinoma with biliaryductal formation and mitotic figures (Fig. 4A, X62.5). Fig. 4B is ahigher power view of the tumor (X 125). An immunohistochemicalstain with a monoclonal antibody to low molecular weight cytokeratins (AE-l) shows that the cells maintain their biliary phenotype (Fig.4C).

ER mRNA. Molecular approaches were used to evaluate for thepresence of ER mRNA in the cholangiocarcinoma cell lines. Fig. 5shows an ethidium bromide-stained agarose gel following RT-PCRwith RNA isolated from various cell lines. RT-PCR primers includedER primers on the left half of the gel and @3-actinprimers on the righthalf of the gel. Molecular weight standards are on the far right lane.Both human cholangiocarcinoma cell lines express ER transcript, asdetermined by RT-PCR. Equivalent amounts of (3-actin were observedin the samples. MCF-7 (a human breast cancer cell line) and COScells served as positive and negative controls, respectively. AlthoughCOS cells do not contain ER transcript, @-actinmRNA is present.Thus, by RT-PCR, ER mRNA is detectable in both human cholangiocarcinoma cell lines.

Northern blot analysis was performed to confirm the RT-PCRresults and to determine the size of the ER transcript. ER variantmRNAs with deletions of portions of ER exons have been reported

*900

.@ 800E

@ 700

! 600> 500

I@400

@ 300

@ 200

100

-.- Control-*- Tamoxifen

t P=0.31

* P=O.02

t

-7-40 3 7 10141721242831353842454952

TIme (day.)

1745

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

TAMOXIFEN LNHIB@ON OF HUMAN CHOLANGIOCARCINOMA

protein was detected (results not shown). ER containing human breastcancer tissue was used as a positive control (results not shown).Therefore, a more sensitive, functional assay was used to evaluate thepresence of ER protein in these tumor cell lines. After growth inestrogen-depleted media for 1 week, cells were grown in mediasupplemented with increasing concentrations of estradiol. After 1additional week, cells were counted with a Coulter counter. Theresults in Fig. 7 were obtained with the OZ cell line. The first columnrepresents the control cells grown in charcoal-free, phenol red-freemedia and were normalized to 100%. The second column depicts cellsgrown in phenol-free media but supplemented with 5% FCS that wasnot treated with charcoal; these cells had a 135—140%increase in cellnumber compared to the control. The last three columns represent thegrowth of cholangiocarcinoma cells in the presence of increasingconcentrations of l7j3-estradiol. The addition of exogenous l7f3-

estradiol stimulated the growth of OZ cells. Similar results wereobtained with the SK-ChA-l cholangiocarcinoma line (results notshown). Cell viability was greater than 99%. Thus, the increasedgrowth rate of tumors in the presence of exogenously added estradiolsuggested the presence of ER protein in these tumor cell lines.

The presence of ER protein was confirmed by immunoprecipitation..@ in both cholangiocarcinoma cell lines (Fig. 8). The H226 monoclonal

‘ antibody, which recognizes an epitope in the NH2 terminus of the ER,

% was used (24, 25). ER protein was identified in the positive control,

q MCF-7cells,aswellasbothhumancholangiocarcinomacelllines,,@ . .. .@@@ Oz and SK-ChA-l cells. The COS cells did not contain ER protein.

The Rainbow molecular weight markers are shown in the bottom lane.@ Thus, the presence of ER protein in human cholangiocarcinoma was

confirmed by immunoprecipitation.

.A.@ 41 ., @,,.@ e@

.. 4 , - -p. •@ ‘@

.@ .‘ ,@ ,@ ,@... . .-. -@ . — .@. ‘ ‘(# .‘.-,@ ‘

#_ S@@ • C .@: :1@ @•@@

..@ :@:@ _.@ . , ,. ‘:@, 4, ‘.@

* S., ‘ @A@'••@ •.@

‘ . . ‘. . S..@ .@ @..1@@- . . ,_@ . .@. ‘ ,

C

,@(

p.

,s

0

@@ 4

. •‘r@

S

4' , -.

@ ‘@;.--,.•.

-.4

DISCUSSION

Cholangiocarcinoma is a malignant neoplasm of the biliary epithehum with devastating clinical consequences (1—5).The etiology ofcholangiocarcinoma remains unknown; however, chronic inflammation of the biliary tract leads to an increased risk for development ofcholangiocarcinoma. The disease is resistant to current systemic treatment including medical, chemotherapeutic, or radiation therapy. Surgery can be an effective cure; however, most patients present withunresectable disease or microscopically positive margins after surgerythat eventually result in tumor recurrence via local tumor spread,peritoneal seeding, or direct extension into the liver parenchyma.Therefore, new treatment modalities are necessary to provide effectivetherapy for patients with metastatic as well as unresectable cholangiocarcinoma. Furthermore, the possibility of a prophylactic chemotherapeutic agent for those patients at increased risk for the development of biliary tumors, such as patients with PSC or congenitalcholedochal cysts, is an additional consideration (26). Tamoxifen may

provide such new treatment.Tamoxifen is a competitive inhibitor of estrogen binding to the ER.

We have shown that tamoxifen inhibits the growth of two humancholangiocarcinoma tumor cell lines in vitro. Importantly, this growthinhibition was achieved with a physiological relevant dose of tamoxifen. Moreover, tamoxifen treatment resulted in a statistically significant growth inhibition of SK-ChA-l human cholangiocarcinomaxenografts in nude mice. Finally, we have demonstrated the presenceof ER at both the mRNA and protein level in two separate humancholangiocarcinoma cell lines. Thus, tamoxifen may have clinicalutility and serve as an adjunctive chemotherapeutic agent, possiblyimproving the survival of patients diagnosed with unresectable ormetastatic cholangiocarcinoma.

Whether the growth inhibition of cholangiocarcinoma occurs viathe ER or another mechanism is unclear and will require further study.

1746

I@ 6@'@1r .‘—

.@, ..@.s -.

@. 5.,•1 @p'-@@@ -.@.

:@@

, @. S@@ •@ .@ 4'J@$'@:) ,

,ji@.@@

@ @.:1

_@.lfl.A:@

Fig. 4. Photomicrographs of human cholangiocarcinoma cells (SK-ChA-l). A, microscopic evaluation of tumor senograft removed from mice and processed for routinecytochemistry revealed a moderately differentiated adenocarcinoma by H&E stain atX62.5. B. higher power of tumor demonstrating ductular structures and mitotic figures(H&E stain, X 125). C, immunohistochemical stain with a monoclonal antibody to lowmolecular weight cytokeratins (AE-l) demonstrating biliary phenotype of tumor (X 125).

(12, 18, 23); tumors with deletion mutations in the ER transcript maynot respond as well to tamoxifen (12, 18, 23). Fig. 6 shows theNorthern blot in ER transcript. The top half of the blot is for ER; thebottom half of the gel shows a blot for GAPDH. MCF-7 and COScells were positive and negative controls, respectively. The two

cholangiocarcinoma cell lines are shown; full-length, intact 6.5-kb

RNA is detected in both cell lines. No smaller forms of deletionmutants of ER mRNA were detected by Northern blotting. Therefore,both human cholangiocarcinomas contain full-length ER transcript.

ER Protein. Immunohistochemistry was used to assess expressionof ER protein in both cholangiocarcinoma cell lines; however, no ER

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

TAMOXIFEN INHIBITION OF HUMAN CHOLANGIOCARCINOMA

—IIC@)c@@Cc#@L)Oc@@c#)c'@

Fig. 5. RT-PCR analysis for ER (left) and @-actin(right). RT-PCR demonstrated the presence of ERmRNA in both human cholangiocarcinoma cell lines(OZ and SK-ChA-l). Controls included MCF-7 human breast cancer cells (positive control) and COScells (negative control). @-Actinwas expressed by allcell lines and was used as an internal control for thepresence of mRNA. Stds, molecular weight standards.

ER

Fig. 6. Northern blot analysis for ER mRNA. TotalRNA was isolated from MCF-7 breast cancer cells(Lane 1), both human cholangiocarcinoma celllines [OZ (Lane 2), SK-ChA- 1 (Lane 3)1, and COScells (Lane 4). Top. Northern blots were performedfor ER, and full-length, 6.5-kb ER transcript wasidentified. Bottom, membranes were stripped andrehybridized with GAPDH cDNA probes.

_@.s...'@@

Tumor growth inhibition by tamoxifen in ER-negative breast cancerhas lead to studies on alternative mechanisms of the action of tamoxifen (27). Thus, alternative mechanisms include modulation ofgrowth-inhibitory and -stimulatory factors (28, 29), induction ofTGF-@3(30—32),apoptotic mechanisms via Bcl-2 receptors (33, 34),tamoxifen binding to high-affinity microsomal binding proteins (35),effects of calcium influx (36), and inhibition of protein kinase Cactivity (37).

TGF-@ has a plethora of biological functions important in theregulation of cellular proliferation and differentiation in most humanepitheial cells (30). Investigations on the effects of tamoxifen on thegrowth of human breast cancer indicate that tamoxifen induces theautocrine secretion of TGF-13 in human breast cancer cells, which thenacts as an inhibitor of tumor growth (31). In addition, tamoxifen mayinduce TGF-f3 secretion by stromal cells leading to decreased cellularproliferation (32). Effects of tamoxifen on TGF-/3 levels in both the invitro and in vivo experiments reported here have not been addressedbut may provide additional insight into the mechanism of tarnoxifenmediated growth inhibition of human cholangiocarcinoma cells.

Prior to the establishment of these tumor cell lines of the biliarysystem, information on the biology of these cells was limited. Theavailability of these cell lines as well as new techniques for theirisolation and long-term culture of bile duct cells has allowed a betterunderstanding of various aspects of these cells (1). Other receptors onbiliary epithelial cells include epidermal growth factor receptor (38),

the secretin receptor (19), and the somatostatin receptor (39). Treatment of a cholangiocarcinoma cell line with a somatostatin analoguecaused both in vivo and in vitro growth inhibition of tumor cell growth(39). Combination chemotherapy with somatostatin and tamoxifen

250

// ,/ ///@

E@ControlU FCS

@ Eat 10pM

IIIEat100pM

D EstlnM* P99%. and eachdata point represents the mean ±5.0 of triplicate experiments; bars. SD.

1747

@— 3-actin

1 2 3 4

Estrogen receptor6.5 Kb —@

GAPDH

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

TAMOXIFEN INHIBITION OF HUMAN CHOLANGIOCARCINOMA

1 2345

@ —

Fig. 8. Immunoprecipitation for ER protein. Cell lysates from human cholangiocarcinoma cell lines (OZ and SK-ChA-l) were immunoprecipitated with a monoclonal antibody to the human ER. Precipitates were electrophoresed in 10% SDS-PAGE. Proteinswereblottedontonitrocellulosemembraneanddevelopedas describedin “MaterialsandMethods.―Positive and negative controls included MCF-7 and COS-l cells, respectively.Lane 1, MCF-7;Lane 2, SK-Cha- I;Lane 3, OZ; Lane 4, COS-l; Lane 5, molecularweightstandards.

may enhance tumor growth inhibition, but this will require experimental confirmation.

ER-negative human breast cancer cells have been stably transfectedwith the ER gene (12, 23). These transfected tumor cells regainresponsiveness to l7@-estradiol and may provide an approach for

controlling the growth of ER-negative breast cancer (12, 23). We haverecently transfected human biliary epithelium with an adenoviralvector containing the Escherichia coli @-galactosidase reporter gene(40).Theuniqueaccessibilityof thehepatopancreaticobiliarysystemto percutaneous or endoscopic therapy for cholangiocarcinoma andpancreatic cancer, coupled with results from this study, may allownovel therapeutic approaches for controlling the growth of thesetumors.

The role of oncogenes and tumor suppressor genes in cholangiocarcinoma is under investigation. Yoshida et. al. (41) have reportedthat deletion or inactivation of p16m@@4gene, a cyclin-dependent kinaseinhibitor important in regulation of the cell cycle, is altered in 64% of25 primary cholangiocarcinomas (41). Whether cholangiocarcinomadevelops following progressive genetic insults by a mechanism postulated for other malignancies remains to be determined.

ERs have been demonstrated in both rat and human liver, but littledata are available on ER protein expression by the biliary epithelium.Quantitation of estrogen and androgen receptors in hepatocellularcarcinoma tissue has revealed higher levels of ER protein in tumorthan in adjacent normal human liver (42, 43). In addition, ER variantmutants have been described in some hepatocellular carcinomas (44).Northern blot analysis in this study confirmed that the two cholangiocarcinoma cell lines contained full-length, intact 6.5-kb ERmRNA. ER variant mRNA with deletion of ER exons have beenreported; breast cancer with deletion mutations in the ER transcriptmay not respond as well to tamoxifen (18, 45). Moreover, thesemutant forms of ER transcript have been reported in hepatocellularcancer (44). The report suggests that patients with advanced hepatocellular cancer have improved survival following treatment with tamoxifen (46, 47). Finally, tamoxifen has been demonstrated to inhibitthe growth of human hepatoma cell lines in vitro (48).

Tamoxifen may prove useful as a prophylactic agent in preventingthe occurrence of cholangiocarcinoma in patients diagnosed with

this study is necessary.

REFERENCES

1. Alpini, G., Phillips, J. 0, and taRusso N. F. The biology of biliary epithelia. In: I.Arias, J. Boyer, N. Fausto, W. Jakoby, D. Schachter, and D. Shafritz (eds.), The Liver:Biology and Pathobiology. Ed. 3, pp. 623-653. New York: Raven Press, 1994.

2. Blumgart, L. H. Cancer of the bile ducts. in: Surgery of the Liver and Biliary Tract.in: L. M. Blumgart (ed), pp. 829—853.Edinburgh: Churchill Livingstone, 1998.

3. Farley, D. R., Weaver, A. L., and Nargorney, D. M. Natural history of unresectedcholangiocarcinoma. Patient outcome after noncurative intervention. Mayo Clin.Proc., 70: 425-429, 1995.

4. Lake, J. R. Benign and malignant neoplasms of the gallbladder, bile ducts, andampulla. in: M. Sleisenger and J. Fordtran (eds.), Gastrointestinal Diseases: Pathophysiology/Diagnosis/Management, Ed. 5 pp. 1891—1902. Philadelphia: W. B.Saunders, 1993.

5. Van Leeuwen, D. J., Huibregtse, K., and Tytgat, G. N. J. Carcinoma of the hepaticconfluence 25 years after Klatskin's description: diagnosis and endoscopic management. Semin. Liver Dis., 10: 102—113, 1990.

6. Gallagher, P. J., Mi!lis, R. R., and Mitchinson, M. J. Congenital dilatation of theintrahepatic bile ducts With cholangiocarcinoma. J. Clin. Pathol. (Lond.), 25: 804—808, 1972.

7. Okuda, K., Kubo, Y., Okuzaki, K., Arishima, T., Hashimoto, M., Jinnouchi, S., Sawa,

Y., Shimokawa, Y., Nakajima, Y., Noguchi, T., Nakano, M., Kojiro, M., andNokashima, T. Clinical aspects of intrahepatic bile duct carcinoma including hilarcarcinoma: study of 57 autopsy-proven cases. Cancer (Phula.), 39: 232—246,1977.

8. Purtile, D. T. C!onorchiasis and hepatic neoplasms. Trop. Geogr. Med., 28: 21—27,1976.

9. Phillips, J. 0., Weisner, R., and LaRusso, N. F. Sclerosing cholangitis. in: J. Kirsner(ed), The Growth of Gastroenterologic Knowledge in the Twentieth Century, pp.309—327.Malvern, PA: Lea and Febiger, 1994.

10. Santen, R. J., Manni, A., Harvey, H., and Redmond, C. Endocrine treatment of breastcancer in women. Endocr. Res., II: 221—265,1990.

11. Jordan, V. C. Long term adjuvant therapy for breast cancer. Breast Cancer Res. Treat.,15: 125—136, 1990.

12. Jiang, S. Y., and Jordan, V. C. Growth regulation ofestrogen receptor-negative breastcancer cells transfected with complementary DNAs for estrogen receptor. J. Natl.Cancer Inst., 84: 580—59!, 1992.

13. Pollack, I. F., Randall, M. S., Kristofsk, M. P., Kelly, R. H., Se!ker, R. G., andVertosicky, F. T. Effects of tamoxifen on DNA synthesis and proliferation of humanmalignant glioma lines in vitro. Cancer Res., 50: 7134—7138, 1990.

14. Cou!dwell, W. T., Weiss, M. H., DiGiorgio, L. M., Weiner, L. P., Hinton, D. R.,Ehsesmann, L. R., Corti, P. S., and Apuzzo, M. L. J. Clinical and radiographicresponse in a minority of patients with recurrent malignant gliomas treated withhigh-dose tamoxifen. Neurosurgery, 32: 485—490,1993.

15. Thapar, K., Stetaneana, L., Kovacs, K., Scheithauer, B. W., Lloyed, R. V., Muller,P. J., and Laws, E. A. Estrogen receptor gene expression in craniopharyngiomas: anin situ hybridization study. Neurosurgery, 35: 1012—1017,1994.

16. Knuth, A., Gabbert, H., Dippold, W., Klein, 0., Sachsse, W., Bitter-Suermann, D.,Prellwitz, W., and Zum Buschenfelde, K. H. M. Biliary adenocarcinoma: characterisation of three new human tumor cell lines. J. Hepatol., I: 579—596, 1985.

17. Homma, S., Nagamori, S., and Fujisc, K. Human bile duct carcinoma cell lineproducing abundant mucin in vitro. Gastroenterol. Jpn., 22: 74—79,1987.

18. Daffada, A. A. I., Johnston, S. R. D., Nicholls, J., and Dowsett, M. Detection of wildtype and exon 5-deleted splice variant oestrogen receptor (ER) mRNA in ER-positiveand -negative breast cancer cell lines by reverse transcription/polymerase chalnreaction. J. Mol. Endocrinol., 13: 265—273,1994.

19. Alpini, 0., Uh'ich, C., Phillips, J. 0., Pham, L., Miller, L. J., and LaRusso, N. F.Upregulation of secretin receptor gene expression in rat cholangiocytes after bile ductligation. Am. J. Physiol., 266: G922—G928,1994.

20. Houghton, J. A., Moroda, S. J., Phillips, J. 0., and Houghton, P. J. Biochemicaldeterminants of responsiveness to 5-fluorouraci! and its derivatives in xenografts ofhuman colorectal adenocarcinomas in mice. Cancer Res., 41: 144—149,1982.

21. Ueno, Y., Phillips, J. 0., Ludwig, J., and LaRusso, N. F. Development and characterization of an experimental model of immune-mediated cho!angitis. Proc. Nat!.Aced. Set. USA, 93: 216—220,1996.

22. Brunner, N., Bronzert, D., Vindelov, L. L., Rygarrd, K., Spang-Thomsen, M., andLippman, M. E. Effect of growth and cell cycle kinetics of estradiol and tamoxifen onMCF-7 human breast cancer cells grown in vitro and in nude mice. Cancer Ret., 49:1515—1520,1989.

23. Castles, C. G., Fuqua, S. A. W., Kiotz, D. M., and Hill, S. M. Expression of aconstitutively active estrogen receptor variant in the estrogen receptor-negative BT-20human breast cancer cell line. Cancer Res., 53: 5934—5939,1993.

24. Greene, G. L., Sobel, N. B., King, W. J., and Jensen, E. V. Immunochemical studiesof estrogen receptors. J. Steroid Biochem., 20: 51—56,1984.

25. Greene, G., and Press, M. F. I. Steroid receptor structure (including monoclonalantibodies and new methods of determination). Structure and dynamics of the estrogen receptor. J. Steroid Biochem., 24: 1—7,1986.

26. Jordan, V. C. An overview of consideration for the testing of tamoxifen as a

1748

chronic inflammatory diseases of the biliary tract such as PSC orcongenital biliary cysts. Therefore, a prospective, randomized trial toevaluate the clinical relevance of the tumor growth inhibition ofhuman cholangiocarcinoma mediated by tamoxifen demonstrated in

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

TAMOXIFEN INHIBITION OF HUMAN CHOLANGIOCARCINOMA

preventative for breast cancer. Ann. NY Acad. Sci., 768: 141—147,1995.27. Henderson, I. C., Harris, J. R., Kinne, D. W., and Hellman, S. Cancer of the breast.

in: V. T. De Vita, Jr., S. Hellman, and S. A. Rosenberg (edt.), Cancer, Principles andPractice of Oncology, pp. 1197—1268.Philadelphia: J. B. Lippincott, 1989.

28. Colletti, R. B., Roberts, J. D., Devlin, J. T., and Copeland, K. C. Effect of tamoxifenon plasma insulin-like growth factor I in patients with breast cancer. Cancer Res., 49:1882—1884,1989.

29. Noguchi, S., Motomura, K., Inaji, H., Imaoka, S., and Koyama, H. Down-regulationof transforming growth factor-a by tamoxifen in human breast cancer. Cancer(Phila.), 72: 131—136,1993.

30. Roberts. A. B., and Sporn, M. B. The transforming growth factor-a. in: M. B. Spornand A. B. Roberts (eds.), Handbook of Experimental Pharmacology, vol. 95. pp.419—472.Berlin: Springer-Verlag, 1990.

31. Butts, A., MacLennan, K., Flanders, K. C., Sacks, N. P. M., Smith. I., McKinna, A.,Dowseu, M., Wakefield, L. M., Spom, M. B., Baum, M., and Colletta, A. A.Induction of transforming growth factor (3@in human breast cancer in vito followingtamoxifen treatment. Cancer Res.. 52: 4261—4264, 1992.

32. Colletta, A. A., Wakefield, L. M., Howell, F. V., van Roozendaal, K. E. P.,Danielpour, D., Ebbs, S. R., Sporn, M. B., and Baum, M. Anti-oestrogens induce thesecretion of active transforming growth factor f3from human fetal fibroblasts. Br. J.Cancer,62: 405—409,1990.

33. Perry, R. R., Kang, Y., and Greaves, B. Effects of tamoxifen on growth and apoptosisofestrogen-dependent and -independent human breast cancer cells. Ann Surg. Oncol.,2: 238—245,1995.

34. Teixeira, C., Reed, J. C., and Pratt, M. A. C. Estrogen promotes chemotherapeuticdrug resistance by a mechanism involving Bcl-2 proto-oncogene expression in humanbreast cancer cells. Cancer Res., 55: 3902—3907,1995.

35. Sutherland, R. L., Murphy, L. C., Foo, M. S., Green, M. D., Whybourne, A. M., andKrozowski, Z. S. High-affinity anti-oestrogen binding site distinct from the oestrogenreceptor. Nature (Land.), 288: 273—275,1980.

36. Greenberg, D. A., Carpenter, C. L., and Messing, R. 0. Calcium channel antagonistproperties of the antineoplastic antiestrogen tamoxifen in the PCI2 neurosecretorycell line. Cancer Res., 47: 70—74,1987.

37. O'Brian, C. A., Liskamp, R. M., Solomon, D. H., and Weinstein, I. B. Inhibition ofprotein kinase C by tamoxifen. Cancer Res., 45: 2462—2465,1985.

38. Ishii, M., Vroman, B. T., and LaRusso, N. F. Morphological demonstration of

receptor mediated endocytosis of epidermal growth factor by isolated bile ductepithelial cells. Gastroenterology, 98: 1284—1291, 1990.

39. Tan, C. K., Podila, P. V., Taylor, J. E, Nagonsey, D. M., Wiseman, G. A., Gores, G. J.,and LaRusso, N. F. Human cholangiocarcinomasexpress somatostatin receptors andrespond to sornatostatin with growth inhibition. Gastroenterology, 108: 1908—1916,1995.

40. Vickers, S. M., Phillips, J. 0., Kerby, J. D., Bynon, J. S., Thompson, J. A., and Curiel,D. 1. in vito gene transfer to the human biliary tract. Gene Ther., 3: 825—828,1996.

41. Yoshida, S., Todoroki, T.. Ichikawa, Y., Hanai, S., Suzuki, H., Hori, M., Fukao. K.,Missa, M., and Uchida, K. Mutations of p'61@4/CDKN2 and pt3t@k4B/MTS2 genes inbiliary tract cancer. Cancer Res., 55: 2756—2760,1995.

42. Eagon, P. K., Francavilla, A., DiLeo, A., Elm, M. S., German, L., Mazzaferro, V..Colella, G. A., Van Thiel, D. H., and Starzl, 1. E. Quantitation of estrogen andandrogen receptors in hepatocellular carcinoma and adjacent normal human liver.Dig. Dis. Sci., 36: 1303-1308, 1991.

43. Francavilla, A.. Panella, C., Amoruto, A., Giangaspero, A.. Gennari. L.. Mazzaferro,V., Colella, G., Van Thiel, D. H., and Starzl, T. E. Role of estrogens and epidermalgrowth factor in hepatocellular carcinoma (HCC). Dig. Dis. Sci., 36.@1299—1302,1991.

44. Villa, E., Camellini,L., Dugani,A., Zucchi, F., Grottola, A., Merighi, A., Buttafoco,P.,Losi, L., and Manenti, F. Variant estrogen receptor messenger RNA species detected inhuman primary hepatocellular carcinoma. Cancer Res.. 55: 498—500,1995.

45. Daffada, A. A. I., Johnston, S. R. D., Smith, I. E.. Detre, S.. King. N., and Dowsett.

M. Exon 5 deletion variant estrogen receptor messenger RNA: expression in relationto tamoxifen resistance and progesterone receptor Ip52status in human breast cancer.Cancer Res., 55: 288—293,1995.

46. Farinati, F., Salvagnini, M., DeMaria, N., Fomasiero, A., Chiaramonte, M., Rossaro,L., and Naccaroto, R. Unresectable hepatocellular carcinoma: a prospective controlledtrial with tamoxifen. J. Hepatol.. II: 297—301,1990.

47. Martinez-Cerezo, F. J. M., Tomas, A., Donoso. L., Enroquez, J., Guarner, C.,

Balanzo, J.. Martinez-Nogueras, A., and Vilardell. F. Controlled trial of tamoxifen inpatients with advanced hepatocellular carcinoma. J. Hepatol.. 20: 702—706.1994.

48. Jiang, S. Y., Shyu, R. Y., Yeh, M. Y., and Jordan, V. C. Tamoxifen inhibits hepatoma

cell growth through an estrogen receptor independent mechanism. J. Hepatology., 23:712—719, 1995.

I749

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

1997;57:1743-1749. Cancer Res Lorenzo K. Sampson, Selwyn M. Vickers, Weizhong Ying, et al. CholangiocarcinomaTamoxifen-mediated Growth Inhibition of Human

Updated version

http://cancerres.aacrjournals.org/content/57/9/1743

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

.pubs@aacr.orgDepartment at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/57/9/1743To request permission to re-use all or part of this article, use this link

on June 16, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/content/57/9/1743http://cancerres.aacrjournals.org/cgi/alertsmailto:pubs@aacr.orghttp://cancerres.aacrjournals.org/content/57/9/1743http://cancerres.aacrjournals.org/