Expression and Antitumor Effects of TRAIL in HumanCholangiocarcinoma

SHINJI TANAKA,1 KEISHI SUGIMACHI,1 KEN SHIRABE,1 MITSUO SHIMADA,1 JACK R. WANDS,2 AND KEIZO SUGIMACHI1

Tumor necrosis factor (TNF)-related apoptosis-inducingligand (TRAIL)/Apo2L has been recently identified as im-portant in promoting programmed cell death in breast andcolon adenocarcinomas. In this study, we investigated theexpression and therapeutic potential of TRAIL in cholangio-carcinoma, one of the most devastating human hepatic ma-lignancies. Expression of TRAIL receptors was determinedin 13 patients with resectable intrahepatic cholangiocarci-noma. Cellular effects of TRAIL in promoting apoptosis ofhuman cholangiocarcinoma cells were analyzed after expo-sure to recombinant protein, as well as following transfec-tion with a cDNA expression construct. In vivo effects ofTRAIL on tumor growth were investigated after subcutane-ous injection of cholangiocarcinoma cells into nude mice.Analysis of 13 clinical and tissue samples revealed thatTRAIL receptors containing the death domain were presentin all cholangiocarcinomas as well as paired normal hepatictissues derived from surgically resected margins. In con-trast, 7 tumors did not express the TRAIL decoy receptorslacking the death domain; such receptors were detectable inall of the normal hepatic tissue counterparts. RecombinantTRAIL induced extensive programmed cell death in cholan-giocarcinoma cell lines lacking decoy receptor expression.Transfection of the ectodomain of TRAIL also induced cel-lular apoptosis; this effect was abolished by introduction ofthe generalized lymphoproliferative disease–like mutationin the TRAIL protein. Finally, in vivo administration of re-combinant TRAIL substantially inhibited subcutaneous tu-mor growth of human cholangiocarcinoma cells. Inductionof apoptosis in tumor cells is possible with a biologicallyactive TRAIL, and suggests that this cytokine is a promis-ing antitumor agent against human cholangiocarcinoma.(HEPATOLOGY 2000;32:523-527.)

The tumor necrosis factor (TNF) family of cytokines con-sists of 10 transmembrane (type II) glycoproteins, i.e., TNF-a,lymphotoxin-a, lymphotoxin-b, CD27L, CD30L, CD40L,OX40L, 4-1BBL, FasL/Apo1L/CD95L, and TNF-related apo-ptosis-inducing ligand (TRAIL)/Apo2L,1,2 and the corre-sponding cellular receptors have been identified.3 Of particu-lar interest is the finding that some, but not all, receptors forTNF-a, FasL, and TRAIL contain a homologous cytoplasmicregion (“death domain”) essential for mediating apoptosis.3,4

Effective antitumor therapy with chemotherapeutic drugs andradiation frequently requires p53 tumor-suppressor func-tion5; however, p53 is mutated in more than half of humantumors, leading to therapeutic resistance. On the other hand,death ligands induce apoptosis independent of p53 in a vari-ety of tumor cells, and thus offer a potential complementaryapproach to conventional cancer therapy.

TNF-a was originally identified because of its ability toinduce hemorrhagic tumors in mice.6 Attempts to use TNF-afor systemic anticancer chemotherapy were discontinued as aresult of severe cytotoxic side effects related to induction ofprogrammed cell death in normal cells. In this regard, TNF-aconcentrations have been found to be significantly higher inacute alcoholic6 and fulminant hepatitis,7 in which a substan-tial increase in transaminase and bilirubin levels suggests adirect cytotoxic role on hepatocytes. Although TNF-a triggersliver apoptosis in vivo and in vitro, hepatocytes appear moresensitive to FasL/FasR-mediated apoptosis. Indeed, agonisticanti-Fas antibody treatment in vivo rapidly progresses to ful-minant and lethal apoptotic hepatitis as a result of the highlevel of Fas expression on hepatocytes.8 Thus, despite theability of TNF-a and FasL to induce apoptosis in cancer cells,documented toxic side effects on normal cells preclude theiruse in chemotherapy.

However, TRAIL/Apo2L was recently discovered as a mem-ber of this cytokine family because of its sequence homologyto TNF-a and FasL.2,9 TRAIL mRNA is constitutively ex-pressed in various tissues,2,9 suggesting a possible physiolog-ical mechanism to protect normal cell types from induction ofapoptosis by TRAIL. One such cellular protective mechanismmay involve differential expression of the TRAIL receptorfamily comprised of DR4, DR5, DcR1, and DcR2.4,10-12 Walc-zak et al. observed no toxicity to normal liver tissues followingin vivo administration of TRAIL, while similar experimentswith FasL led to massive hepatocellular apoptosis and hem-orrhage culminating in fulminant hepatitis.13

Cholangiocarcinoma is an aggressive malignancy resistantto virtually all antitumor agents.14,15 Because of TRAIL effectson certain tumor cell lines with respect to induction of pro-grammed cell death, and a general lack of effect on normalcells, we analyzed TRAIL receptor expression in clinical sam-

Abbreviations: TNF, tumor necrosis factor; TRAIL, tumor necrosis factor–relatedapoptosis-inducing ligand; RT-PCR, reverse-transcription polymerase chain reaction;TUNEL, TdT-mediated dUTP nick end labeling; aa, amino acid; gld, generalized lym-phoproliferative disease.

From the 1Department of Surgery II, Faculty of Medicine, Kyushu University,Fukuoka, Japan; and 2Liver Research Center, Rhode Island Hospital, Brown UniversitySchool of Medicine, Providence, RI.

Received April 7, 2000; accepted June 5, 2000.Supported by Sagawa Cancer Research Foundation; Fukuoka Cancer Society; Kanae

Foundation for Life & Socio-Medical Science; Kaibara Morikazu Medical Science Pro-motion Foundation; a Grant-in-Aid from the Ministry of Education, Science, and Cul-ture of Japan; and CA-35711 from the National Institutes of Health. S.T. is a recipient ofthe Japan Cancer Society Incitement Award.

Address reprint requests to: Shinji Tanaka, M.D., Ph.D., Department of Surgery II,Faculty of Medicine, Kyushu University, 3-1-1 Maidashi, Fukuoka 812-8582, Japan.E-mail: shinjit@surg2.med.kyushu-u.ac.jp; fax: 81-92-642-5482.

Copyright © 2000 by the American Association for the Study of Liver Diseases.0270-9139/00/3203-0012$3.00/0doi:10.1053/jhep.2000.9716

523

ples of intrahepatic cholangiocarcinoma and surroundingnormal liver tissue. In addition, we performed in vitro/in vivostudies with human cholangiocarcinoma cell lines. Significantprogrammed cell death was induced by TRAIL, suggesting apotential therapeutic approach for this disease.

MATERIALS AND METHODS

Expression of TRAIL Receptors in Human Cholangiocarcinoma. Wecompared TRAIL receptor gene expression in tumors derived from13 individuals (9 males/4 females; mean age, 64 years) with his-topathologically diagnosed intrahepatic cholangiocarcinoma (with-out cirrhosis) to paired samples derived from surrounding normalliver. Patients with recurrent tumors were excluded from this study.Tumors were classified histologically as papillary adenocarcinomas(N 5 3), well-differentiated adenocarcinomas (N 5 2), moderatelydifferentiated adenocarcinomas (N 5 3), and poorly differentiatedadenocarcinomas (N 5 5). The clinical follow-up period for all pa-tients was 38.5 6 22.0 months. During the follow-up period, 4 of the13 patients (30.8%) died of the disease. Total cellular RNA was ex-tracted according to standard methods, and TRAIL mRNA detectionwas assessed following the reverse-transcription (RT) reaction and25-cycle polymerase chain reaction (PCR).4,10-13 To each RT-PCRreaction, 100 ng of total RNA was applied. The following primerswere used: DR4: 59-CAGAACGTCCTGGAGCCTGTAAC-39, 59-AT-GTCCATTGCCTGATTCTTTGTG-39 (product size 5 299 bp);DR5: 59-CCTTGGAGACGCTGGGAGAGA-39, 59-TGGGTGATGT-TGGATGGGAGAGT-39 (product size 5 252 bp); DcR1: 59-CGT-TAGGGAACTCTGGGGACAG-39, 59-GGAAGCGTTGGTGTAATCC-ACA-39 (product size 5 329 bp); and DcR2: 59-AATTTGCCTTCTTGC-CTGCTATGTA-39, 59-CTCCTCCGCTGCTGGGGTTTTC-39 (productsize 5 261 bp). We performed RT-PCR of glyceraldehyde-3-phosphate

dehydrogenase as a control mRNA16: 59-GTCAACGGATTTGGTCTG-TATT-39, 59-AGTCTTCTGGGTGGCAGTGAT-39 (product size 5 560bp). All PCR primers were selected to span the introns to detectspecific mRNA sequences.

In Vitro Effects of TRAIL on Cellular Apoptosis of Cholangiocarcinoma.The human intrahepatic cholangiocarcinoma cell line, HuH-28,17

was cultured in Dulbecco’s modified Eagle medium (Sigma Co., St.Louis, MO) supplemented with 10% fetal bovine serum, and theeffects of TRAIL were investigated.2,9 Because there are few availablecell lines of human intrahepatic cholangiocarcinoma, we used thehuman extrahepatic cholangiocarcinoma TFK-1 and HuCCT1 celllines for this research.18,19 All cell lines were provided by Cell Re-source for Biomedical Research, Tohoku University. Exponentialgrowth cells were incubated with recombinant human TRAIL (R & DSystems, Minneapolis, MN) for 24 hours. Cellular apoptosis wasanalyzed by the in situ TdT-mediated fluorescein-dUTP nick endlabeling (TUNEL) method20 using the MEBSTAIN Apoptosis KitDirect (Medical & Biological Laboratories, Nagoya, Japan). In brief,cells were fixed with 4% paraformaldehyde in 0.1 mol/L NaH2PO4

(pH 7.4) and permeabilized with 0.5% Tween 20. Washed cells werelabeled with TdT solution/fluorescein-dUTP, and counted per 200cells 3 5 fields under fluorescent microscopy. The results were de-rived from 3 individual experiments.

Construction of TRAIL Expression Plasmids. The expression plasmidencoding for the ectodomain of TRAIL (sTRAIL) was synthesized.21

In brief, the sequence corresponding to the C-terminus extracellularregion of 114-281 amino acid (aa) was amplified by PCR, ligateddownstream to include a signal peptide sequence, and subclonedinto the pcDNA3 vector. A TRAIL mutant expression plasmid(sTRAILmut) analagous to the Fas ligand of generalized lymphopro-liferative disease was created by PCR, resulting in a phenylalanine

FIG. 1. Expression of DR4, DR5, DcR1, and DcR2 TRAIL receptors in 13paired samples of tumor and surrounding normal liver, and 3 cell lines de-rived from human cholangiocarcinoma measured by RT-PCR. The quality ofRNA was ascertained by a parallel amplification of glyceraldehyde-3-phos-phate dehydrogenase mRNA (lower panel). Note that DcR1 was absent in themajority of cholangiocarcinoma tissues. Furthermore, DcR2 was undetect-able in 6 of 13 cholangiocarcinoma tissues.

FIG. 2. Determination of the percent of apoptotic cells induced by in vitroadministration of recombinant TRAIL. Three human cholangiocarcinomacell lines were studied including HuH-28 (F), TFK-1 (Œ), and HuCCT1 (■).Results were derived from 3 individual experiments, and error bars indi-cate SD.

524 TANAKA ET AL. HEPATOLOGY September 2000

(F)-to-leucine (L) change at the 274 aa of the TRAIL protein.22 Fol-lowing DNA plasmid transfection with Superfect reagent (QIAGEN,Santa Clarita, CA) for 48 hours, cells were evaluated for apoptosis.We confirmed successfully transfected cells with either sTRAIL orsTRAILmut by RT-PCR of the transfectants, followed by direct se-quence analysis.

Tumor Formation in Nude Mice. To investigate TRAIL’s in vivo effecton tumorigenicity, 1 3 107 HuCCT1 cells were inoculated subcuta-neously into the back of 10 nude mice23; 0.1 mL phosphate-bufferedsaline with/without 100 ng recombinant human TRAIL was similarlyinjected on days 1, 3, and 5. Nude mice were observed daily for 4weeks, and solid tumor formation was determined.

RESULTS AND DISCUSSION

The cytoplasmic region of several TNF receptor families ofproteins contain about 80 amino acids; designated the “deathdomain,” this region is found in DR4/TRAIL-R14 and DR5/TRAIL-R2 proteins,11 and transduces the apoptotic signal fol-lowing ligand stimulation.3 TRAIL receptor expression wasanalyzed in 13 intrahepatic cholangiocarcinoma samples (Fig.1). Both DR4 and DR5 transcripts were detected in all tumorsand paired normal tissue except 1 that was DR5-positive only(case 13). Recent investigations note the existence of 2 antag-onistic receptors that lack the intracellular death domain, andthus act as decoy receptors (DcR1/TRID/TRAIL-R4),10,11

(DcR2/TRUNDD/TRAIL-R5).12 They are expressed in TRAIL-resistant, but not sensitive, cells.24-26 Analysis revealed that

both decoy receptors were expressed in all paired normal liversamples, consistent with Walczak’s findings that demon-strated complete resistance of normal liver tissue to the bio-logic effects of TRAIL.13 However, a recent report has empha-sized that normal human hepatocytes and culture may besusceptible to TRAIL effects27; further studies will be requiredbefore firm conclusions are reached. In contrast, decoy recep-tors were not detected in 7 of 13 cholangiocarcinoma tumors(Fig. 1). It will also be desirable to perform immunohisto-chemical analysis to determine protein expression of TRAILreceptors; however, the current antibodies available are notsuitable for paraffin-embedded specimens.

The apoptotic effects of TRAIL were analyzed in 3 culturedcell lines (HuH-28, TFK-1, HuCCT1) derived from humancholangiocarcinoma. Programmed cell death was observed inTFK-1 and HuCCT1, but not HuH-28, cells (Fig. 2); of note isthat p53 tumor-suppressor protein function was absent in allcell lines.28 TRAIL-induced programmed cell death appearsindependent in cholangiocarcinoma cells, although p53 is re-quired in some apoptotic signals.5 Both DR4 and DR5 tran-scripts were detected in all cell lines. However, decoy recep-tors were present only in HuH-28, and not in TFK-1 orHuCCT1, cholangiocarcinoma cells. Thus, cell susceptibilityto TRAIL-induced apoptosis was dependent on the absence ofdecoy TRAIL receptors (Fig. 2). Interestingly, more pro-

FIG. 3. Biologic effects of synthesized TRAIL (sTRAIL) on HuCCT1 human cholangiocarcinoma cells. (A) Comparison of the structure and amino acidcomposition at the C-terminus between FasL and TRAIL. TM, transmembrane domain. The mutation F-to-L273 of FasL of gld mice (FasLgld11) and theanalogous mutation F-to-L274 of human TRAIL (TRAILmut) are indicated by the asterisks. (B) Detection of apoptotic cells by the fluorescent TUNEL assayusing the HuCCT1 cholangiocarcinoma cell line transfected with sTRAIL and sTRAILmut. (C) Illustration of the percent of apoptotic cells induced bytransfection of sTRAIL and sTRAILmut expression vectors. Results were derived from 3 individual experiments, and error bars indicate SD.

HEPATOLOGY Vol. 32, No. 3, 2000 TANAKA ET AL. 525

nounced biologic effects of TRAIL were observed in HuCCT1cells expressing DR4 and DR5, as compared with TFK-1 cellsexpressing only DR4; there was no expression of TRAIL decoyreceptors in these 2 cell lines.

Among the TNF family members, the extracellular regionof TRAIL is the most conserved as compared with FasL(23.2%). In this context, Takahashi et al. reported that a mis-sense mutation in the FasL gene interfered with Fas-mediatedapoptosis in generalized lymphoproliferative disease (gld) ofmice.22 (This mutation involves one of the conserved aminoacids located in the b-strand of the TNF molecule resulting ina FFG substitution to LFG at the 273-275 aa region of FasL.)The same conserved amino acid sequence is present in theTRAIL protein (Fig. 3A). Therefore, the analogous mutationwas created by substituting FFG to LFG at 274-276 aa inTRAIL, and the subsequent biologic effect was examined.Transient transfection of the TRAIL ectodomain (sTRAIL)induced apoptosis in HuCCT1 cells, and the gld-like mutantconstruct (sTRAILmut) abolished the TRAIL’s ability to pro-duce apoptotic change (Fig. 3B, 3C). These results suggestthat this C-terminal conserved sequence was essential to ini-tiate cytotoxic effects in HuCCT1 human cholangiocarci-noma cells.

The in vivo effects of TRAIL on HuCCT1 cholangiocarci-noma tumor cell growth rate was investigated in nude mice.Following subcutaneous injection of 1 3 107 HuCCT1 cells,there was progressive tumor formation in all animals (n 5 10)at the injection site, reaching a mean size of 10 mm within 24to 28 days (Fig. 4). In contrast, no detectable tumor formationwas observed in 10 nude mice following subcutaneous TRAILadministration. Therefore, some of the biologic properties ofTRAIL are characterized by total inhibition of human cholan-

giocarcinoma growth in nude mice, and induction of apopto-sis in the same tumor cells in vitro.

A key mechanism for ligand-mediated transmission of in-tracellular signals is through the interaction of transmem-brane receptors. While caspases are essential mediators ofTRAIL-induced apoptosis and FasL/FasR-induced death sig-nals,29 divergent results stimulate further attempts to identifyregulatory molecules in TRAIL-mediated programmed celldeath. A previous report suggested that the Fas-binding pro-tein was not involved in TRAIL-mediated death signals.29

However, Walczak et al. reported that DR5-mediated apopto-sis was FADD protein–dependent, whereas DR4 was not.30

Chaudhary et al. observed that both DR4 and DR5 may trans-mit apoptotic signals via FADD, and possibly TRADD, inter-action that directly binds to receptors.31 Because DR4 andDR5 transcripts are present in all of our cholangiocarcinomacell lines including the prototype HuCCT1 cells, it is possiblethat dual signaling pathways may be induced following TRAILexposure to cholangiocarcinoma cells. Further studies arenecessary to clarify the role of such transduction moleculeswith respect to TRAIL effects in cholangiocarcinoma, sincethe possibility remains for disabled transduction in TRAIL-positive, but apoptosis-resistant, cholangiocarcinomas.

Human cholangiocarcinomas, resistant to chemotherapyand radiotherapy, are aggressive tumors14 that result in a poorprognosis following surgical resection. Our investigation sug-gests the TRAIL protein has striking antitumor effects thatmay translate to this disease. While FasL is a potent inducer ofapoptosis, reports suggest it has diminished biologic activityfollowing cleavage to the soluble form32; in contrast, TRAILmay be active as a soluble protein, as shown both here and inprior studies, which does not damage normal cells.2,9 Finally,

FIG. 4. Antitumor effects of TRAIL on cholangiocarcinoma growth in nude mice. (A) Demonstration of subcutaneous tumor formation produced byHuCCT1 human cholangiocarcinoma in nude mice with (right) or without (left) recombinant TRAIL administration. (B) Determination of tumor sizefollowing TRAIL administration. Tumor size is expressed in (mm) 6 SD.

526 TANAKA ET AL. HEPATOLOGY September 2000

the TRAIL-induced death signals do not involve a p53-depen-dent pathway, making them an attractive antitumor approachin human cholangiocarcinoma.

Acknowledgment: The authors thank Immunex Co. for thehelpful information on TRAIL, and Dr. H. Chang (Massachu-setts Institute of Technology) for helpful discussions; K.Ogata, T. Shimooka, and M. Hirata for technical assistance.The editorial assistance of Sharon Chiott is gratefully ac-knowledged.

REFERENCES

1. Cosman D. A family of ligands for the TNF receptor superfamily. StemCells 1994;12:440-455.

2. Wiley SR, Schooley K, Smolak PJ, Din WS, Huang CP, Nicholl JK, Suth-erland GR, et al. Identification and characterization of a new member ofthe TNF family that induces apoptosis. Immunity 1995;3:673-682.

3. Smith CA, Farrah T, Goodwin RG. The TNF receptor superfamily ofcellular and viral proteins: activation, costimulation, and death. Cell1994;76:959-962.

4. Pan G, O’Rourke K, Chinnaiyan AM, Gentz R, Ebner R, Ni J, Dixit VM.The receptor for the cytotoxic ligand TRAIL. Science 1997;276:111-113.

5. Ruley HE. p53 and response to chemotherapy and radiotherapy. Impor-tant Adv Oncol 1996:37-56.

6. Bradham CA, Plumpe J, Manns MP, Brenner DA, Trautwein C. Mecha-nisms of hepatic toxicity. I. TNF-induced liver injury. Am J Physiol1998;275:G387-392.

7. Bird GL, Sheron N, Goka AK, Alexander GJ, Williams RS. Increasedplasma tumor necrosis factor in severe alcoholic hepatitis. Ann InternMed 1990;112:917-920.

8. Ogasawara J, Watanabe-Fukunaga R, Adachi M, Matsuzawa A, KasugaiT, Kitamura Y, Itoh N, et al. Lethal effect of the anti-Fas antibody in mice[published erratum appears in Nature 1993 Oct 7;365(6446):568]. Na-ture 1993;364:806-809.

9. Pitti RM, Marsters SA, Ruppert S, Donahue CJ, Moore A, Ashkenazi A.Induction of apoptosis by Apo-2 ligand, a new member of the tumornecrosis factor cytokine family. J Biol Chem 1996;271:12687-12690.

10. Pan G, Ni J, Wei YF, Yu G, Gentz R, Dixit VM. An antagonist decoyreceptor and a death domain-containing receptor for TRAIL [Com-ments]. Science 1997;277:815-818.

11. Sheridan JP, Marsters SA, Pitti RM, Gurney A, Skubatch M, Baldwin D, Ra-makrishnan L, et al. Control of TRAIL-induced apoptosis by a family of sig-naling and decoy receptors [Comments]. Science 1997;277:818-821.

12. Marsters SA, Sheridan JP, Pitti RM, Huang A, Skubatch M, Baldwin D,Yuan J, et al. A novel receptor for Apo2L/TRAIL contains a truncateddeath domain. Curr Biol 1997;7:1003-1006.

13. Walczak H, Miller RE, Ariail K, Gliniak B, Griffith TS, Kubin M, Chin W,et al. Tumoricidal activity of tumor necrosis factor-related apoptosis-inducing ligand in vivo [Comments]. Nat Med 1999;5:157-163.

14. Shimada M, Takenaka K, Kawahara N, Yamamoto K, Shirabe K, MaeharaY, Sugimachi K. Chemosensitivity in primary liver cancers: evaluation ofthe correlation between chemosensitivity and clinicopathological fac-tors. Hepatogastroenterology 1996;43:1159-1164.

15. Moradpour D, Wands JR. Hepatic oncogenesis, IVD. Tumors of the liver.In: Zakim D, Boyer ID, eds. Hepatology: A Textbook of Liver Disease.Volume 2. Philadelphia: Saunders, 1996:1490-1512.

16. Tanaka S, Akiyoshi T, Mori M, Wands JR, Sugimachi K. A novel frizzledgene identified in human esophageal carcinoma mediates APC/beta-cate-nin signals. Proc Natl Acad Sci U S A 1998;95:10164-10169.

17. Kusaka Y, Tokiwa T, Sato J. Establishment and characterization of a cellline from a human cholangiocellular carcinoma. Res Exp Med 1988;188:367-375.

18. Saijyo S, Kudo T, Suzuki M, Katayose Y, Shinoda M, Muto T, Fukuhara K,et al. Establishment of a new extrahepatic bile duct carcinoma cell line,TFK-1. Tohoku J Exp Med 1995;177:61-71.

19. Miyagiwa M, Ichida T, Tokiwa T, Sato J, Sasaki H. A new human cholan-giocellular carcinoma cell line (HuCC-T1) producing carbohydrate an-tigen 19/9 in serum-free medium. In Vitro Cell Dev Biol 1989;25:503-510.

20. Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed celldeath in situ via specific labeling of nuclear DNA fragmentation. J CellBiol 1992;119:493-501.

21. Ashkenazi A, Pai RC, Fong S, Leung S, Lawrence DA, Marsters SA,Blackie C, et al. Safety and antitumor activity of recombinant solubleApo2 ligand. J Clin Invest 1999;104:155-162.

22. Takahashi T, Tanaka M, Brannan CI, Jenkins NA, Copeland NG, Suda T,Nagata S. Generalized lymphoproliferative disease in mice, caused by apoint mutation in the Fas ligand. Cell 1994;76:969-976.

23. Tanaka S, Wands JR. A carboxy-terminal truncated insulin receptor sub-strate-1 dominant negative protein reverses the human hepatocellularcarcinoma malignant phenotype. J Clin Invest 1996;98:2100-2108.

24. Gura T. How TRAIL kills cancer cells, but not normal cells [News; com-ment]. Science 1997;277:768.

25. French LE, Tschopp J. The TRAIL to selective tumor death [News; com-ment]. Nat Med 1999;5:146-147.

26. Ashkenazi A, Dixit VM. Apoptosis control by death and decoy receptors.Curr Opin Cell Biol 1999;11:255-260.

27. Jo M, Kim TH, Seol DW, Esplen JE, Dorko K, Billiar TR, Strom SC.Apoptosis induced in normal human hepatocytes by tumor necrosis fac-tor-related apoptosis-inducing ligand [In process citation]. Nat Med2000;6:564-567.

28. Jia LQ, Osada M, Ishioka C, Gamo M, Ikawa S, Suzuki T, Shimodaira H,et al. Screening the p53 status of human cell lines using a yeast functionalassay. Mol Carcinog 1997;19:243-253.

29. Marsters SA, Pitti RM, Donahue CJ, Ruppert S, Bauer KD, Ashkenazi A.Activation of apoptosis by Apo-2 ligand is independent of FADD butblocked by CrmA. Curr Biol 1996;6:750-752.

30. Walczak H, Degli-Esposti MA, Johnson RS, Smolak PJ, Waugh JY, BoianiN, Timour MS, et al. TRAIL-R2: a novel apoptosis-mediating receptor forTRAIL. EMBO J 1997;16:5386-5397.

31. Chaudhary PM, Eby M, Jasmin A, Bookwalter A, Murray J, Hood L. Deathreceptor 5, a new member of the TNFR family, and DR4 induce FADD-dependent apoptosis and activate the NF-kB pathway. Immunity 1997;7:821-830.

32. Tanaka M, Itai T, Adachi M, Nagata S. Downregulation of Fas ligand byshedding [Comments]. Nat Med 1998;4:31-36.

HEPATOLOGY Vol. 32, No. 3, 2000 TANAKA ET AL. 527