Et

T*CI1

R

ilhflfwD

ptacHTvac

np

ehchpenblce

8

Biochemical and Biophysical Research Communications 281, 1041–1044 (2001)

doi:10.1006/bbrc.2001.4470, available online at http://www.idealibrary.com on

xpression of the Hoxa-13 Gene Correlateso Hepatitis B and C Virus Associated HCC

ing Huang,* Victor Chesnokov,* Kazunari K. Yokoyama,† Brian I. Carr,‡ and Keiichi Itakura*,1

Department of Molecular Biology, Beckman Research Institute of City of Hope, 1450 East Duarte Road, Duarte,alifornia 91010; †RIKEN (The Institute of Physical and Chemical Research) Tsukuba Institute, Tsukuba,

baraki 305-0074, Japan; ‡Pittsburgh Transplantation Institute, Biomedical Science Tower,5th Floor, 200 Lothrop Street, Pittsburgh, Pennsylvania 15213,

eceived February 6, 2001

demonstrated by in vivo experiments (4). These resultspppctcttpasm

cdkicblpaatcHcHd

M

mBfw

s

To study the Hoxa-13 gene in the liver, we examinedts expression by RT-PCR in various liver cell lines, rativers under different conditions, and human primaryepatocellular carcinomas (HCCs). The gene was

ound to be expressed in cell lines originating fromiver stem-like cells, but not in cell lines originatingrom hepatocytes and bile duct epithelia. Expressionas induced in rat livers after treatment with

-galactosamine, which is known to induce oval cellroliferation, but not after a two-thirds partial hepa-ectomy (2/3 PH) where induction of oval cell prolifer-tion is thought not to occur. Expression of the geneorrelated with human HCC samples associated withepatitis B or C virus infection in this small series.hese results suggest that the Hoxa-13 gene may pro-ide a potentially useful tool for elucidation of mech-nisms involved in lineage-specific differentiation andarcinogenesis of liver stem cells. © 2001 Academic Press

Key Words: hox gene; liver; hepatocellular carci-oma; stem cell; oval cell; hepatitis virus; hepatoctye;artial hepatectomy.

Although somewhat controversial in the past, recentvidence showing the existence of hepatic stem cellsas become over-whelming (1, 2). When hepatocytesannot respond rapidly to a growth stimulus or afterepatocyte loss as in acute liver failure, a novel cellopulation termed “oval cells” proliferates and differ-ntiates to hepatocytes. Oval cells form a heteroge-eous compartment and display antigenic features ofile duct cells and fetal cells (1, 2). The epithelial celline WB-F344 which has characteristics similar to ovalells was established from normal rat liver (3). Differ-ntiation of WB-F344 cells to hepatocytes has been

Abbreviation used: HCC, hepatocellular carcinoma.1 To whom reprint requests should be addressed. Fax: 626/301-

271. E-mail: kitakura@coh.org.

1041

rovide strong evidence that hepatic stem cells areresent in normal adult livers and the oval cell com-artment contains hepatic stem cells and/or progenitorells. Remarkably, it has recently been reported thathe bone marrow is a potential source of hepatic ovalells (5–7). Furthermore, it has been demonstratedhat small hepatocyte-like progenitor cells observed inhe retrosine model give rise to a large number ofrogeny hepatocytes and restore tissue mass in thedult rat (8). It seems therefore that the adult liver haseveral pathways for the repair of a deficit in the hepaticass under different pathophysiological conditions.Although the presence of hepatic stem cells has be-

ome clear, the mechanisms involved in proliferation,ifferentiation, and carcinogenesis are largely un-nown. To examine the mechanisms, it is important todentify genes that control these processes. We haveloned a cDNA encoding Hoxa-13, a homeobox protein,y screening a cDNA library of the rat hepatoma celline McA-RH7777 with a synthetic oligonucleotiderobe (manuscript in preparation). Homeobox proteinsre transcriptional factors that control cell differenti-tion, proliferation, and carcinogenesis (9). We inves-igated expression of the Hoxa-13 gene in various liverells and primary human HCCs, and found that theoxa-13 gene is expressed in liver stem-like cells (oval

ells) and in human primary HCCs associated withepatitis B and C virus infection, but not in fullyifferentiated hepatocytes.

ATERIALS AND METHODS

Cell lines. The following cell lines were obtained from ATCC andaintained as indicated: McA-RH7777, H4IIE, HepG2, Hep3B,RL-3A, and AML12. LE6 is a gift from Dr. N. Fausto, BDE1.1 is

rom Dr. D. Hixon, and WB-F344 is from Dr. W. Colmann and theyere maintained as described (3, 10, 11).

Animals. Young adult male F344 rats, 180–200 g, were used totart all experiments. Normal and partially hepatectomized rat liv-

0006-291X/01 $35.00Copyright © 2001 by Academic PressAll rights of reproduction in any form reserved.

ewwhd

fUdHaa

tGTToatinps

tfRtp(1p(fTf7tr(CEwMRrT3

To confirm the intactness of the human RNA, PT-PCR was also donefsr

R

eitvnaTtHiModARpfows

itptbHd5rpda

apisLLlar

aP4a1HpWs

Vol. 281, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

rs were prepared as described previously (12). Treatment of ratsith D-galactosamine was carried out as published (13) and animalsere sacrificed 1, 2, and 3 days after the treatment. Separation ofepatocytes from non-parenchymal cells was done as previouslyescribed (14)

Primary HCC samples. All human primary HCCs were obtainedrom Presbyterian Hospital, Department of Liver Transplantation,niversity of Pittsburgh Medical Center. Serological analysis wereone by ELISA kit (Abbott) using second generation antibody forCV, and by ELISA kit (Abbott) using antibody against HBV surfacentigen and ELISA kit (Diasotin) using antibody against HBV Entigen for HBV.

Preparation of the internal standard RNA. The Hoxa-13 cDNA inhe plasmid was amplified with a pair of the primers, 59-CCG-AATTCCGGTATCTGGATATGCCAGTGGT-39, and 59-CGCGGA-CCGCGGACTCCTCTTCAATGGAATC-39 by PCR for 10 cycles.he PCR product was digested with EcoR1 and BamHI at 37°Cvernight, and ligated to the linearized plasmid pUC19. The result-nt plasmid was digested with EcoRV and a 95 bp DNA was ligatedo the EcoRV site. The PCR product plus the 95 bp DNA was insertednto pBluscrip II vector between EcoRI and BamHI sites. The inter-al standard RNA was produced by transcription of the linearizedlasmid (by BamHI) with T7 DNA-dependent RNA polymerase asuggested by the manufacture (Stratagene).

RT-PCR. Total cellular RNA was isolated from various cells byhe Trizol method (Gibco BRL). First, cDNA strands were producedrom a mixture of total cellular RNA (3 mg) and the internal standardNA (1 3 106 molecules) by incubation with AMV reverse transcrip-

ase (0.6 ml), 5 mM dNTP (3 ml), RNasin (1 ml, Promega) and 10 pMrimer 59-GACTCCTCTTCAATGGAATC-39 in the first strand bufferGibco BRL) at 48°C for 30 min. A sample of cDNA corresponding to50 ng of the cellular RNA was amplified. PCR was carried out in theresence of Taq DNA polymerase (2.5 units, Takara) for 30 cycles94°C, 30 s; 50°C, 30 s; and 72°C 1 min) in the same buffer usedor the reverse transcriptase reaction. The upstream primer 59-ATCTGGATAGTCCAGTGGT-39 and the downstream primer used

or the reverse transcriptase reaction delineate 612 bp (Hoxa-13) and07 bp (internal standard) amplification products. To ensure that theotal RNA were intact, RT-PCR for cyclophiline transcript was car-ied out under the same conditions except using different primersdata not shown). Primers were end labeled with [g-32P] ATP (5000i/mmol; ICN Biolabs) and T4 DNA polynucleotide kinase (Newngland Biolabs). To confirm that the 612 bp product correspondedith the Hoxa-13 transcripts, direct sequencing of the product fromcARH-7777 cells was done (data not shown). RT-PCR of the humanNA was carried out under the same conditions as described for theat RNA, except using a different downstream primer 59-TCCTGT-CTGGAACCAGATT-39and annealing was done at 55°C to give the88 bp (Hoxa-13) and 483 bp (internal standard) amplified products.

FIG. 1. RT-PCR analysis of total RNA isolated from various ratnd mouse cell lines. RT-PCR products were analyzed by electro-horesis on 4% acrylamide gels. The upper bands represent thenternal control RNA and the lower ones the Hoxa-13 mRNA. Lane 1hows DNA size markers, upper one (622 bp) and lower one (527 bp).anes 2–9 show RT-PCR products of control and various total RNAs.ane 2, H2O; lane 3, McARH-7777; lane 4, WB-F344; lane 5, LE6;

ane 6, BDE1.1; lane 7, AML12; lane 8, normal adult rat hepatocytes;nd lane 9, rat nonparenchymal liver cells. The data presented areepresentative of three independent experiments.

1042

or the cytokeratin 8 transcript from each human sample (nothown). Radioactive amplified products were subjected to electropho-esis in 4% acrylamide gels and autoradiographed with X-ray film.

ESULTS

Expression of the Hoxa-13 gene in liver cells. Toxamine the expression in different cell types originat-ng from the liver, we carried out RT-PCR analysis ofhe total RNA from several cell lines derived fromarious hepatic lineages, including parenchymal andonparenchymal cells. The RT-PCR products were an-lyzed by electrophoresis on acrylamide gels (Fig. 1).he upper bands are the amplified signals for the in-ernal RNA controls and the lower bands are for theoxa-13 transcripts. The expression was detected only

n the total RNA from cell lines of rat hepatomacARH-7777 (lane 3), epithelia WB-F344 (lane 4), and

val cells LE6 (lane 5), but not in those from rat bileuct epithelia BDE1.1 (lane 6) or mouse hepatocytesML-12 (lane 7). The transcripts were not detected inNA from rat normal hepatocytes (lane 8) or non-arenchymal cells (lane 9) under the conditions usedor RT-PCR. The cell line LE6 was established fromval cells that proliferate in the liver after treatmentith a choline-deficient carcinogenic diet (15) and has

ome characteristics of hepatic stem cells (10).We further analyzed expression of the Hoxa-13 gene

n rat livers under two different conditions, D-galac-osamine treatment that induces specifically oval cellroliferation (13), and a 2/3 partial hepatectomy (PH)hat stimulates proliferation of hepatocytes followed byile duct epithelia, but not oval cells (2). An increase ofoxa-13 gene expression was observed two and threeays after the D-galactosamine treatment (Fig. 2, lanes–7), but not after 2/3 PH (Fig. 2, lanes 8–10). Theseesults strongly suggest that the Hoxa-13 gene is ex-ressed in the liver stem cell compartment, but not inifferentiated liver epithelial cells such as hepatocytesnd bile duct epithelia.

FIG. 2. RT-PCR analysis of total RNA isolated from rat liversfter D-galastosamin treatment and 2/3 partial hepatectomy. RT-CR products were analyzed as described for Fig. 1. 24 h (lane 5),8 h (lane 6), and 72 h (lane 7) after D-galactososamine treatment,nd 24 h (lane 8), 48 h (lane 9), and 72 h (lane 10) after 2/3 PH. Lane, the same DNA markers as used in Fig. 1; lane 2, negative control,2O; lane 3, positive control, WB-F344; lane 4, normal liver. For theositive control experiment, 10 times less amount of the RNA fromB-F344 than the experiment in Fig. 1 was used. The data pre-

ented are representative of two independent experiments.

SdwstrhtdehiaesbHHnifHivHsns

D

entiao

2/3 PH, in which proliferation of oval cells is not ob-sHpsate

ctoFcictvnf

ttfdtsfsaiwaH

escadshopvisatr

A

BDb

pwdicsbHaC

Vol. 281, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Expression of the Hoxa-13 gene in human HCCs.ince it was found that Hoxa-13 gene expression isependent on the liver cell type, we investigatedhether human HCC cells express the Hox gene and if

o, whether the expression correlates with any factorshat could affect hepatocellular carcinogenesis. We car-ied out RT-PCR of the total RNA isolated from tenuman primary HCCs, one normal human liver, andwo human hepatoma cell lines by the same method asescribed for the rat samples, except for using a differ-nt primer (see Materials and Methods). Five of the tenuman HCC samples showed Hoxa-13 gene expression

n various amounts (Fig. 3, lanes 7, 8, 10, 12, and 13)nd the sample in lane 10 demonstrated the highestxpression. In lane 10 an amplified product movinglower than the internal standard RNA was detected,ut the origin of the product is currently unknown. Theoxa-13 transcripts were not detected in the other fiveCC samples (Fig. 3, lanes 4, 5, 6, 9, and 11) or theormal liver (lane 14) under our RT-PCR conditions. It

s noteworthy that expression of the Hoxa-13 gene wasound in five out of six HCC samples associated withepatitis B or C virus infection, but was not detected

n HCC samples from patients without evidence of theirus infection. The human hepatoblastoma cell lineepG2 (Fig. 3, lane 2) expressed the Hoxa-13 tran-

cripts, but the HCC cell line Hep3B (Fig. 3, lane 3) didot express, although the latter expressed Hepatitis Burface antigen (16).

ISCUSSION

We have demonstrated that the Hoxa-13 gene isxpressed in liver stem-like cell lines, originating fromonparenchymal epithelial cells (WB-F344) and fromhe oval cell compartment (LE6), and expression isnduced in rat livers after treatment with D-galactos-mine, an agent that is known to induce proliferationf oval cells. However, expression is not induced by a

FIG. 3. RT-PCR analysis of total RNA isolated from humanrimary HCCs. RT-PCR of the total RNA from human primary HCCsas carried out and the products were analyzed in the same way asescribed for Fig. 1. The data presented are representative of twondependent experiments. The upper bands represent the internalontrol RNA and the lower ones the Hoxa-13 transcript. Lane 1hows DNA size markers, upper band (527 bp) and lower band (407p); lane 2, HepG2; lane 3, Hep3B; lanes 4 to 13, human primaryCCs; lane 14, normal human liver. Samples of lanes 5, 7, 8, 10, 12,nd 13 are from human primary HCCs associated with hepatitis B orinfection.

1043

erved. Based on these data, we believe that theoxa-13 gene is expressed in the hepatic stem and/orrogenitor cells. Further study suggests that expres-ion of the Hoxa-13 gene correlates with Hepatitis Bnd C virus associated human primary HCCs, and thathere may be two types of HCC cells with respect toxpression of the gene.Hox proteins play important roles in lineage specific-

ell differentiation of hematopoietic stem cells andhere is evidence that translocation or over expressionf Hox genes in hematopoietic cells is oncogenic (17).urther, it was suggested that hematopoietic stemells may have a prehepatocyte potential (7). Our find-ng of expression of the Hoxa-13 gene in hepatic stemells and/or progenitor cells, therefore, provides a po-entially useful tool for elucidation of mechanisms in-olved in the lineage-specific differentiation and carci-ogenesis of hepatic stem cells, which may originaterom bone marrow cells.

The cellular origins of human HCC are unclear. Inhe adult rat liver, there are at least three potential cellypes, which can contribute to liver repair under dif-erent pathophysiological conditions. These are fullyifferentiated hepatocytes (18), incompletely differen-iated hepatocyte-like cells (8) and oval cells (1, 2). Weuppose that all of the cell types are potential targetsor transformation to HCC, since their longevity as-ures continuous exposure to carcinogens. Addition-lly, the liver is exposed continuously to xenobiotics,ncluding toxins and hepatocarcinogens, and infectionsith various hepatitis viruses including hepatitis Bnd C viruses, that are highly associated with humanCC incidence (19).However, there is no clear lineage marker to differ-

ntiate resultant HCCs. We have found in our smallample that expression of the Hoxa-13 gene correlateslosely with hepatitis B or C virus associated HCCs,nd that no HCC samples from patients without evi-ence of the virus infection showed the expression. Wepeculate that many human HCCs associated withepatitis B or C virus infection would originate fromval cells, since oval cell proliferation has been re-orted in individuals with chronic hepatitis B and Cirus infections (20, 21) and Hoxa-13 gene is expressedn rat oval cells as described above. To more fullyupport our hypothesis, examination of the functionsnd regulation of the Hoxa-13 gene in differentia-ion and carcinogenesis of hepatic stem cells will beequired.

CKNOWLEDGMENTS

We thank N. Fausto for a gift of LE6, D. C. Hixon for a gift ofDE1.1, and W. B. Coleman for a gift of WB-F344. We also thank N.ay for critical reading of the manuscript. This work was supportedy IDEN foundation to K.I. and NIH Grant CA 82723 to BIC.

REFERENCES

1

1

12. Higgins, C. M., and Anderson, R. M. (1931) Experimental pa-

1

1

1

1

1

1

1

2

2

Vol. 281, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

1. Sell, S. (1990) Is there a liver stem? Cancer Res. 50, 3811–3815.2. Grisham, J. W., and Thorgeirsson, S. S. (1997) in Stem Cells

(Potten, D. S., Ed.), pp. 233–282, Academic Press, New York, NY.3. Tsao, M-S., Smith, J. D., Nelson, K. G., and Grisham, J. W.

(1984) A diploid epithelial cell line from adult rat liver withphenotypic properties of oval cells. Exp. Cell Res. 154, 38–52.

4. Coleman, W. B., Wennerberg, A. E., Smith, G. J., and Grisham,J. W. (1993) Regulation of the differentiation of diploid and someaneuploid rat liver epithelial (stemlike) cells by the hepatic mi-croenvironment. Am. J. Pathol. 142, 1373–1382.

5. Petersen, B. E., Bowen, W. C., Patrene, K. D., Mars, W. M.,Sullivan, A. K., Murase, N., Boggs, S. S., Greengerger, J. S., andGoff, J. P. (1999) Bone marrow as a potential source of hepaticoval cells. Science 284, 1168–1170.

6. Theise, N. D., Badve, S., Saxena, R., Henegariu, O., Sell, S.,Crawford, J. M., and Krause, D. S. (2000) Derivatation of hepa-tocytes from bone marrow cells in mice after radiation-inducedmyeloablation. Hepatology 31, 235–240.

7. Theise, N. D., Nimmakayalu, M., Gardner, R., Illei, P. B., Mor-gan, G., Teperman, L., Henegariu, O., and Krause, D. S. (2000)Liver from bone marrow in humans. Hepatology 32, 11–16.

8. Gordon, G. J., Coleman, W. B., Hixson, D. C., and Grisham, J. W.(2000) Liver regeneration in rats with retrosine-induced hepato-cellular injury proceeds through a novel cellular response. Am. J.Pathol. 156, 607–607.

9. Levine, M., and Hoey, T, (1988) Homeobox proteins as sequence-specific transcription factors. Cell 55, 537–540.

0. Lazaro, C. A., Rhim, J. A., Yamada, Y., and Fausto, N. (1998)Generation of hepatocytes from oval cell precursors in culture.Cancer Res. 58, 5514–5522.

1. Yang, L., Faris, R. A, and Hixon, D. C. (1993) Long-term cultureand characteristics of normal rat liver bile duct epithelial cells.Gastroenterology 104, 840–852.

1044

thology of the liver. 1. Restoration of the liver of the white ratfollowing partial surgical removal. Arch. Pathol. 12, 186–202.

3. Lemire, J. M., Shiojiri, N., and Fausto, N. (1991) Oval cell pro-liferation and the origin of small hepatocytes in liver injuryinduced by D-galactosamine. Am. J. Pathol. 139, 535–552.

4. Walpole, H. W., Lee, W. M., Walle, T., Walle, U. K., Willson,M. J., and Kennedy, J. W. (1990) Rabbit hepatocytes primaryculture: Preparation, viability, and use in studies of proparanololmethod. Hepatology 11, 394–400.

5. Yaswen, P., Hayner, N. T., and Fausto, N. (1984) Isolation of ovalcells by centrifugal elutriation and comparison with other celltypes purified from normal and preneoplastic livers. Cancer Res.44, 324–331.

6. Aden, D. P., Fogel, A., Plotkin, S., Damjanov, I., and Knowles,B. B. (1979) Controlled synthesis of HbsAg in a differentiatedhuman liver carcinoma-derived cell line. Nature 282, 615–616.

7. Lawrence H. J., and Largman, C. (1992) Homeobox gene innormal and hematopoiesis and leukemia. Blood 80, 2445–2453.

8. Overturf, K., Al-Dhalimy, M., Ou, C-N., Finegold, M., andGrompe, M. (1997) Serial transplantation reveals the stem-cell-like regenerative potential of adult mouse hepatocytes. Am.J. Pathol. 151, 1273–1280.

9. Gupta, S., and Shafritz, D. A. (1994) in The Liver: Biology andPathology (Arias, I. M., Boyer, J. L., Fausto, N., Jakoby, W. B.,Schachter, D. A., and Shafritz, D. A., Eds.), pp. 1429–1453,Raven Press, New York, NY.

0. Hsia, C. C., Evarts, R. P., Nakatsukasa, H., Marsden, E. R., andThorgeirsson, S. S. (1992) Occurrence of oval-type cells in hepa-titis B virus-associated human hepatocarcinogenesis. Hepatol-ogy 16, 1327–1333.

1. Lowes, K. N., Brenna, B. A., Yeoh, G. C., and Olynyk, J. K. (1999)Oval cell numbers in human chronic liver diseases are directlyrelated to disease severity. Am. J. Pathol. 154, 537–541.