(CANCER RESEARCH 52. 6358-6364, November 15, 1992]

p53 Gene Mutation Spectrum in Hepatocellular Carcinoma1

Tatsuya Oda,2 Hitoshi Tsuda, Aldo Scarpa,3 Michiie Sakamoto, and Setsuo Hirohashi4

Pathology Division, National Cancer Center Research Institute, 5-1-1 Tsukiji, Chuo-ku, Tokyo 104 [T. O., H. T., A. S., M. S., S. H.J, and Institute of ClinicalMedicine, Surgery, Tsukuba University, 1-1-1 Tennodai, Tsukuba city, Ibaraki 305 [T. O.], Japan

ABSTRACT

In order to clarify the significance of mutation of the p53 tumorsuppressor gene in the genesis and development of human hepatocellu-lar carcinoma (HCC) in an aflatoxin I!, Ion-exposure area, the spectrum, i.e.. incidence, type, and site, oipS3 gene mutations was examinedin 169 tissue samples resected mainly from Japanese patients usingsingle-strand conformation polymorphism analysis and direct sequencing. Forty-nine tumors (29%) showed a p53 mutation (39 point mutations and 10 frameshifts). The point mutations comprised 18 transitions,only 4 of which occurred at CpG sites, and 21 transversions. Two evo-lutionarily conserved domains, IV and V, contained 65% of all mutations and codon 249 was the most frequent mutation site (7/49). Thespectrum of p53 mutation did not differ among HCCs in relation to thetype of hepatitis virus infection, sex, age, and background liver diseaseof patients, tumor size, or presence of metastasis, but incidence and sitewere significantly associated with the degree of differentiation of cancercells. In poorly differentiated HCC, p53 mutation was frequent (54%)and clustered on domains IV and V, whereas in well or moderatelydifferentiated HCC, the mutation was less frequent (21%) and equallydistributed on domains II to V. Restriction fragment length polymorphism analysis revealed loss of heterozygosity on chromosome 17p in 55(69%) of 80 informative cases and in 34 (95%) of 36 cases with pS3mutation. Therefore, p53 gene mutation is suggested to occur independently of the type of viral infection or status of preexisting liver diseaseand to occur preferentially in moderately and poorly differentiatedHCCs in association with or after loss of another p53 alíeleas a lateevent of HCC progression.

INTRODUCTION

Genetic alterations occurring endogenously or caused by exogenous factors play a fundamental role in the multistage processes of tumorigenesis and progression of malignancy (1).These include integration of viral genomes (2, 3), activation ofprotooncogenes (4), and inactivation of tumor suppressor genes(5).

HCC5 is one of the most common cancers in Asia and Africa

(6). Various studies have confirmed its association with hepatitis B or C viral infection (70% of all cases) (7, 8). Whileactivation of known protooncogenes does not seem to play animportant role (9, 10), frequent allelic loss on specific chromosomal arms (1, 4q, 5q, lip, 13q, 16p, 16q, and 17p) (11-16)indicates that dysfunction of diverse tumor suppressor geneslocated on these chromosomal arms is involved in the development of HCC. Among these, the p53 gene (17), located onchromosome 17pl3, has been well analyzed, and frequent loss

Received 6/22/92: accepted 9/11/92.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 with 18 U.S.C. Section 1734 solely to indicate this fact.

1This study was supported by a Grant-in-Aid for the Comprehensive 10-YearStrategy for Cancer Control from the Ministry of Health and Welfare of Japan.

2 Awardee of a Research Resident Fellowship from the Foundation for Promotion of Cancer Research, Tokyo, Japan.

3 Supported by a Foreign Research Fellowship. Present address: Institute ofPathology. Verona University, Strada Le Grazie, 37134 Verona, Italy.

4 To whom requests for reprints should be addressed.5 The abbreviations used are: HCC, hepatocellular carcinoma; HBV, hepatitis B

virus; HCV, hepatitis C virus; PCR, polymerase chain reaction; SSCP, single-strand conformation polymorphism; LOH, loss of heterozygosity; HBsAg, hepatitis B virus surface antigen.

of one alíeleand mutations in the other alíelehave been reported to occur in diverse human cancer types including HCC(18-21).

Study of the mutational types of tumors seems to be a powerful way of assessing the contribution of endogenous versusexogenous sources of DNA damage if carcinogens leave their"footprints." Specific carcinogens are, in fact, known to pro

duce specific and reproducible types of damage in the DNAmolecule, mainly point mutations (22). For example, N-nitroso-yV-methylurea and dimethylbenzanthracene have beenshown to induce transition from G to A and transversion fromA to T in the c-Ha-ras gene in rat and mouse, respectively(23, 24). Transversion from G to T is caused by both exogenouscarcinogens (25) and endogenous processes such as free radicaldamage arising from normal biochemical reactions in mouseand monkey (26). On the other hand, a number of mutations indifferent genes occur with a high frequency at particular sites;e.g., CpG sites are preferential targets for point mutations inmammalian cells during the process of DNA replication, presumably due to the spontaneous deamination of methylatedcytosine residues (27). Examples of p53 mutations due to eitherexogenous or endogenous mechanisms can be found in differenthuman cancers. In colon carcinoma and Burkitt-type malignantlymphoma, 50-65% of mutations are represented by C to Ttransitions at CpG sites. On the other hand, in non-small celllung cancer, 57% of point mutations are G to T transversions,suggesting a specific form of direct DNA damage due to thepresence of benzo(a)pyrene in tobacco ( 18).

A specific type of p53 mutation has also been demonstratedin HCCs of subjects exposed to food contaminated with aflatoxin B,, in which most mutations are G to T transversionsoccurring at codon 249 (19, 20). This type of mutation is similar to that caused by the same substance in mutagenesis experiments in Escherichia coli (28). On the other hand, judging fromstudies on the limited number of liver cancers occurring indifferent geographic areas, the types and sites of p53 mutationsseem to show a random distribution (21, 29, 30). However,comprehensive information on the incidence, recessive or dominant-negative role (31), and clinical implication of p53 mutations in primary HCC is lacking. In particular, it would be ofinterest to determine whether any special type of p53 mutationis associated with HBV and HCV infection, or with histológica!features of preexisting liver disease. Moreover, the p53 mutation spectrum in an aflatoxin low-exposure area like Japanshould be analyzed (29).

In the present study, we examined the spectrum, i.e., incidence, type, and site, of p53 gene mutations and loss of heterozygosity on chromosome 17p in a large series of HCCs.Furthermore, we analyzed the association between p53 mutation and clinical parameters including the type of hepatitis virusinfection.

MATERIALS AND METHODS

Patients and Pathological Samples. One hundred sixty-nine frozen samples of HCC were obtained from 140 patients (128 Japanese,6 Korean, 4 Indonesian, and 2 Taiwanese) who underwent surgicaltreatment at the National Cancer Center Hospital, Tokyo, Japan. All of

6358

pii MUTATION SPECTRUM IN HCC

Table 1 Components of HCC nodules analyzed

Single noduleMultiple nodule

235Only 1analyzedRecurrent

TotalPatients7619

313861

3

140Nodules7638

95

3890

3

169

these were primary HCC tumor nodules: 76 were from patients with asingle HCC nodule; 90 were from 61 patients with multiple HCCnodules; and the remaining 3 were recurrent HCC tumors obtained ata second operation in three different patients (Table 1). Tumors diag

nosed pathologically and/or genetically as metastatic in origin wereexcluded from the present series (32, 33). The 169 tumors were classified into 9 early HCCs corresponding to the in situ or microinvasivestage of cancer, 35 well differentiated HCCs, 84 moderately differentiated HCCs, and 41 poorly differentiated HCCs (32, 34).

Serological tests for serum HBsAg in all 140 patients and serumanti-HCV antibody in 128 patients were performed using the reversepassive hemagglutination procedure (SERODlA-HBs, Fujirebio Inc.,Tokyo, Japan), and an enzyme-linked immunosorbent assay test systemusing anti-HCV antibody (Ortho, Raritan, NJ), respectively. SerumHBsAg and anti-HCV antibody were judged as positive in 30 and 67patients, respectively, whereas 31 patients were negative for both.

PCR-SSCP Analysis and Direct DNA Sequencing. DNA was extracted according to standard procedures (35). Inasmuch as 98% ofp53 gene mutations in diverse types of cancer have been found in exons5-8 (18), we focused our study on these exons. The PCR-SSCP methoddescribed by Orila et al. (36) was used to detect the p53 gene mutation.

Table 2 p53 gene mutations in hepatocellular carcinoma and clinicopalhological parameters

ClinicalparametersCasePatient

Background(Sex/Age) LiverVirusPathological

parameters oftumorsDiffer

entiationSize (mm) PVTT0IM ExonpS3

genemutationsCodonNucleotideAmino

acid17pLOHSingle

noduleHCCS-lS-2S-3S-4S-5S-6S-7S-8S-9S-10S-llS-12S-13S-14S-15S-16MultipleD-l-aD-2-bD-3-aD-4-aD-5-aD-6-aD-7-bD-8-aD-9-aD-9-bD-10-aD-10-bT-l-aT-2-aP-l-aP-l-bMultipleM-lM-2M-3M-4M-5M-6M-7M-8M-9M-10M-llM-12M-13M-14M-15M/67M/50M/73M/75M/64M/64M/63M/51M/46M/63M/57M/70F/63M/35M/62M/72nodule

HCC:allM/57F/64F/65M/65F/65M/65M/62M/56M/65M/65M/58M/58M/68M/62M/58M/58LCCHLCCHFibrosisLCLCCHCHCHNSLCLCLCPCLCnodulesanalyzedCHLCCHCHPCCHLCPCCHCHCHCHCHLCPCPCCBC—

*—NCNBCCB—B—CC——C—NCCBBCC—CCCMPMMPPMPWMMPPPMPMMPPWPPMPPPMMPMM45

+r80++3026128

++33++25120

++36+457530

+4090

++3585

+76

+3320431945

+30215565

+365227

+23+138+c

5++67+

77+

77778888+

7+85+

55778+

5+777+

75+

7-5757157220239241246246249249258266272274274250279176-178176179236239282157-159237-243240241249141249130248163249GTC—

GACTAT-TGTAAC-AGCTCC—TTCATG-CTGATG—AGGAGG—ATGAGG—AGTGAA-TAAGGA—TGAGTG-ATGGTT-GCTGTT-GCT1-base

pairdel1-basepairdel1-basepairinsTGC—

AGCCAT—CGTTAC-TGCAAC—AGCCGC-TGG7-base

pairdel18-basepairdel1-base

pairdelTCC-GCCAGG-TGG1-base

pairdelAGG—AGTCTC—CGCCGC-CAGTAC—AACAGG—

ATGVal—

AspTry—CysAsn—SerSer-PheMet-ValMet—ArgArg—MetArg—

SerGlu—StopGly—StopVal-MetVal-AlaVal-AlaFrameshiftFrameshiftFrameshiftCys—

SerHis—ArgTyr—CysAsn—SerArg—TrpFrameshift6

deletionFrameshiftSer—

AlaArg—TrpFrameshiftArg—

SerLeu—ArgArg—GinTyr—AsnArg—

Met+++++—+++++N+—N+++++++N+++NN+NNNnodule

HCC: only 1 noduleanalyzedM/51M/69M/71M/61M/63M/45M/57M/55M/38M/58M/56F/61M/47M/70F/75CHPCLCCHLCLCCHCHCHLCCHLCPCCHLC—NCC-BBBBC—CCCCMMWPMMMPMMPMPMP3865

+285523110

++105+40+15+2080

+5045

+50+65

+555+

67++++

7++777+

88++

8+5++

6131137156193239242245249249250272273278168-178219AAC—

AGCCTG-CAGCGC-CCCCAT—CCTAAC—AGCTGC—TTCCGC—TGCAGG-TGGAGG—ATGCCC-CTCGTG-ATGCGT—TGTCCT-

CGT31-base pairdel1-base pair delAsn—

SerLeu—GinArg—ProHis—ProAsn—SerCys-PheGly-CysArg—TrpArg—MetPro—LeuVal-MetArg—CysPro—

ArgFrameshiftFrameshift+++++N++++NNN++Recurrent

HCCnoduleR-lR

2F/61M/62CHCH-BPM45+18+ 85273156-CGT- TGT5-base

pair delArg—CysFrameshift+N

a pvTT, portal vein tumor thrombi; CH; chronic hepatitis; PC, precirrhosis; LC, liver cirrhosis; NS, not specific; B, HBV (HBeAg) positive; C, HCV (C-100 antibody)

positive; N, HBsAg negative but HCV not tested: W. well differentiated HCC; M, moderately differentiated HCC; P. poorly differentiated HCC; IM. intrahepaticmetastasis; N, not informative.

* -, both HBV and HCV negative.c -, negative; +, microscopically positive: ++, macroscopically positive.

6359

pS3 MUTATION SPECTRUM IN HCC

The SSCP analysis and direct DNA sequencing were performed using the same procedure as that reported previously (33). Briefly, eachexon 5-8 of the p53 gene was amplified by 35 cycles of PCR using5'-end-labeled primers and Taq polymerase (Perkin Elmer/Cetus, Nor-

walk, CT) and analyzed on 6% polyacrylamide gels after denaturationof the PCR products by heating. Then, abnormally shifted bands detected by SSCP analysis were eluted from the gel, and single-strandtemplates for sequencing were obtained by 55 cycles of asymmetrical

(20:1 primer ratio) PCR. The sequencing was performed using a 7-deaza-GTP Sequenase version 2 kit (United States Biochemicals,

Cleveland, OH) and analyzed on 8% polyacrylamide gel containing 5 Murea.

LOH on Chromosome 17p. Ninety-five cases, including all casesshowing p53 gene mutation, were analyzed for LOH on chromosome17p by restriction fragment length polymorphism analysis using apYNZ22 (DÌ7S5)probe (37), as described previously (13, 14). In 68

Fig. 1. PCR-SSCP analysis of p53 mutationin HCC. (A) Representative cases of SSCPanalysis (orp53 gene mutations in HCC. PCR-amplified fragments which encompass exons5-8 of the p53 gene were electrophoresed independently on 6% polyacrylamide gels. Caseswere judged positive for mutation when bandsshowing different mobility shifts from normalcontrols were observed. The bands showingmobility shifts indicated by arrowheads werethen subjected to DNA sequencing analysis.(B) Direct DNA sequencing of PCR-amplifiedfragments. Representative sequences of abnormally shifted bands from each of exons 5-8 inSSCP analysis are shown. The sequences ofthe coding strands are shown for exons 5 and6. and of the non-coding strands for exons 7and 8. JV,normal controls; Rl, D4-a, S4, M15,and Dl-a, case numbers (see Table 2).

7

6

Case No.

I I I I I I I I I I I I I I I I I I¿•^.<^l5-'%,<:^^%%-%'?>-%<^o,^^^^)o'& 'r? '$ O iPsOifOC'S'iPtr^Çj

B N R1 D4-a S4

codon273cvG^

codon273

codon239

Arg\TlCys

G-

Ser-«-Asn

- ~ codon.——¿�241

AGCTAGCTExon 8

-»T

AGCTAGCT Ser-»Phe

Exon 7

M15 N

codon ^.219c\örc/

Frameshirt

codon219

codon176

ProCys

N•D1-aicodo

176ycr

AA

CGATC

Exon 6GA

I C (G A I C G AExon 5

6360

p53 MUTATION SPECTRUM IN HCC

cases, extensive study with five additional DNA probes, pl44D6(D17S34), pMCT35-l (DI7S3Ì),pHF12-2 (DÌ7S1),plO-3 (MYH2),and p 10-41 (D17S71) (37), was performed.

RESULTS

Mutations of the p53 gene were detected in 46 (33%) of the140 patients and 49 (29%) of the 169 HCC nodules. The resultsof the genetic study and clinicopathological profiles of the p53mutation-positive HCC patients are summarized in Table 2.Representative results of SSCP analysis and direct sequencingare presented in Fig. 1.

Specific Mutation Spectrum of p53 in HCC. The mutationswere represented by point mutation in 39 cases (80%), microde-letion in 9 cases (18%), and one-base insertion in one case(Table 3). Nineteen point mutations were at G:C and 20 at A:Tbase pair sites. Transition (in which a purine is substituted fora purine or a pyrimidine for a pyrimidine) was detected in 18tumors, G:C^A:T in 8 tumors, 4 at CpG dinucleotide sites, andA:T^G:C in 10. Transversion (in which a purine is substitutedfor a pyrimidine or vice versa) was detected in 21.

The sites of p53 mutations are represented schematically inFig. 2. Mutations were clustered at two domains (38). One wasdomain IV, covering codons 234-258 in exon 7, where 23(47%) of 49 mutations were detected: 20 of 36 point mutations;and 3 of 10 frameshift mutations. The other was domain V,covering codons 270-286 in exon 8, where 9 (18%) of themutations (8 point mutations and 1 microdeletion) were detected. The most frequent mutation site in HCC was codon 249,demonstrating 7 mutations accounting for 14% of all mutationsand 33% of the transversions. Four of them were from Japanesepatients (S-7, D-10-b, P-l-b, and M-9; see Table 1) with G to Ttransversion at the second position in 3 cases and at the thirdposition in 1 case. The other three were from non-Japanesepatients, two being from Taiwanese (D-9 and M-8) with A to Ttransversion at the first position and one from an Indonesian(S-8) with G to T transversions at the third position.

Association of p53 Mutation with Cancer Cell Differentiation. The incidence of p53 mutation was significantly associated with cancer cell differentiation (Table 4). In poorly differentiated HCCs, the mutation was detected in more thanone-half of the cases, whereas in well differentiated HCCs, including early HCCs, the incidence of mutation was extremelylow. Other clinicopathological parameters, such as patient sexand age, tumor size, type of associated liver damage, presence ofportal vein tumor thrombi, and intrahepatic metastasis showedno specific correlation with the incidence of p53 mutation. Irrespective of the degree of differentiation of cancer cells, mutations in domains IV and V were the most common, but thepattern of distribution of p53 mutation sites differed betweenpoorly differentiated HCC and well or moderately differenti-

iucodon=mâ„¢=i

II=

point mutation |

1 1 li 1 1lili1501 1 1Ili200250

III IV V300

Tttrf'

= frameshift mutationttTTTT

Fig. 2.p53 mutation spectrum in HCC. Downward arrows, 39 point mutations;upward arrows and sequential horizontal bars, 10 frameshift mutations and resulting damaged areas; hatched box, highly conserved domains II-V (39).

ated HCC (Fig. 3); in poorly differentiated HCC, 14 of 16 cases(88%) carried the point mutation in domain IV or V. On theother hand, in well and moderately differentiated HCC, pointmutations oïp53were detected outside of domain IV or V in 9(39%) of 23 cases. Thus, mutations in the regions outside domains IV and V were almost entirely (9 of 11) confined to thewell and moderately differentiated groups.

Lack of Striking Difference in/753 Mutation Pattern betweenHBV- and HCV-positive Patient Groups. Among the 128 pa

tients for whom data on both HBV and HCV infection wereavailable, the incidence of p53 mutation was 33% (10 of 30) inHBsAg-positive cases, 30% (20 of 67) in anti-HCV antibody-

positive cases, and 39% (12 of 31) in cases negative for both(Table 4). There was no significant difference in the incidenceor type of mutations among these three groups. The sites ofmutation showed that mutations in the HCV-associated group

tended to be spread along the gene, whereas in the HBV groupthe mutations were clustered in domain IV (Fig. 4), althoughthe difference was not significant. Moreover, different mutations were detected in a single patient, i.e., in three patients withmultiple HCCs and chronic viral infection (D9, DIO, and PI),the types and sites of mutations differed between two tumors(Table 2). The mutations in codon 249 numbered 3 (G to Ttransversions) in HCV-positive cases and 3 in HBV-positivecases (two A to T and one G to T) (Table 2).

Coincidence of pS3 Mutation and LOH on 17p. Of the 95cases examined, LOH on chromosome 17p was positive in 55(69%) of 80 informative cases (Table 5). These 55 tumors wereconsidered to carry only one alíeleof the p53 gene, and 34 ofthem also carried a p53 gene mutation in the remaining alíele.The other 21 one-allele tumors and virtually all of the two-allele

Table 3 Type ofp53 mutations in hepatocellular carcinomas

Point mutations Frameshift mutation

Mutations at G:C Mutations at A:T

- A:T° -T:A" -C:G* . T:A* -G:C" -C:G* Deletion Insertion

10 1

19 20

39 10

Total 49" Transition-type mutation.* Transversion-type mutation.

6361

pS3 MUTATION SPECTRUM IN HCC

Table 4 Incidence of p5i gene mutations

ClinicalparametersType

of hepatitisvirusesHBV(HBsAgpositive)HCV(anti-HCV antibodypositive)Both

negativeUndeterminedBackground

liverdiseaseChronichepatitisPrecirrhosisLiver

cirrhosisFibrosisAge

(yr)>4950-5960<SexMaleFemaleTotalPathological

paramatersDifferentiation

of cancercellsEarlyHCCAdvanced

HCCWelldifferentiatedModeratelydifferentiatedPoorly

differentiatedTumor

size(mm)<2021-4950<Intrahepatic

metastasisorportal vein tumorthrombiPositiveNegativeTotalNo.Analyzed30673112631756420457511921140No.Analyzed93584413187518881169of

patientsp53

mutationpositive(%)10

(33)20(30)12(39)4

(33)22

(35)6(35)17(30)I

(25)5(25)12(27)29

(39)39

(33)7(33)46

(33)of

nodulesp53

mutationpositive(%)0(0)3(9)24

(29)22(54)7(23)24

(28)18(35)28

(32)21(26)49

(29)

enee, types, and sites. These data suggest that various carcinogens are involved in the genesis ofp53 gene mutation, or if oneor more specific carcinogens are associated, their effect is indirect, most probably due to DNA polymerase infidelity linked tomitogenesis. In this regard, hepatitis virus, which increases cellular turnover by inflammation, could be considered an indirectmutagen linked to mitogenesis.

It has been reported that G to T transversion at the third baseof codon 249 is a specific feature of aflatoxin-related endemicHCCs (19, 20, 29). Moreover, mutation at codon 249 has alsobeen preferentially observed in HCCs from HBV-infected patients in aflatoxin high-exposure areas (40), and mutations atcodon 249 have never been identified in non-HBV-relatedHCCs (18). Thus, it is also suggested that, in addition to aflatoxin intake, HBV infection is necessary for its induction. In

Well differentiated HCC

i

Moderately differentiated HCC

11 11 11 il 11

Poorly differentiated HCC1

1 II 11codon 150

t200 250 300

Fig. 3. p53 mutation spectrum in HCC classified according to cancer celldifferentiation. Downward arrows, point mutations; upward arrows and sequentialhorizontal bars, frameshift mutations and damaged areas.

tumors showed no p53 mutation upon SSCP analysis. Conversely, all except two tumors with p53 mutations also carriedLOH on 17p.

DISCUSSION

In total, 33% of HCC patients showed p53 mutations, andfour noteworthy features of the mutation spectrum in HCCwere revealed: (a) almost equal incidence of transitions andtransversions; (b) low incidence of mutations at CpG sites (8%);(c) high frequency of frameshift mutations (20%); (d) preferential (65%) clustering of mutations at two "hot spots," domains

IV and V.Carcinogens can induce mutations by direct adduct-driven

mutagenesis or by triggering increased cellular turnover (=mitogenesis), which increases the mutation rate indirectly (39).In the present study, no specific type of mutation indicative ofan association with particular carcinogens was found in Japanese HCCs, at variance with the specific predominance of G—«Ttransversion in non-small cell lung cancers. Transition of C—»T

at CpG sites, which results from a spontaneous error of DNAreplication observed in cancers such as colon cancer, was alsoinfrequent. The presence and type of hepatitis virus infectionwere unrelated to the spectrum of p53 mutation. Moreover, inmultiple primary cases both with and without viral infection,p53 mutations were different in each tumor in terms of près-

hepatitis C virus

111 1 111

hepatitis B virus

111 1.codon150 200 250 300

Fig. 4. p53 mutation spectrum in HCC classified according to type of associated hepatitis virus. Downward arrows, point mutations; upward arrows and se-quential horizontal bars, frameshift mutations and damaged areas. Hepatitis Bvirus is HBsAg positive; hepatitis C virus is anti-HCV antibody (C-100) positive.

Table 5 p53 mutations and LOH on chromosome 17p in HCC

17pLOHTotal34

2«36p53

mutation21

2344Total552580*

" These two cases were among the 27 cases analyzed with only 1 probe (see"Materials and Methods").

* Among the 95 cases analyzed, 80 were informative and the remaining 15 were

not informative for the presence of LOH.

6362

p53 MUTATION SPECTRUM IN HCC

Chr. 16 LOH

p53 damage indomains IV, V Metastasis

p53 damageoutside domains IV,V

Chr. 17p LOH

Poorly differentiated HCC

Moderately differentiated HCC

Well differentiated HCC

Chronic Hepatitis/ Liver Cirrhosis

Normal Hepatocyte

Fig. 5. Genetic model of multistage hepatocarcinogenesis. Chronic effect ofhepatitis virus infection or alcohol intake induce chronic liver damage. Still unknown genetic or chemical alterations cause development of early or well differentiated HCCs. Thereafter, p53 gene mutations and preceding LOH on 17p aresuggested to be involved in tumor progression to less differentiated, more aggressive cancers. Mutations in domains IV and V of the p53 gene seem to confer highmalignancy in comparison with mutations outside these regions. Other geneticabnormalities, such as chromosome 16q losses, may lead the HCC to a moreadvanced stage, offering metastatic potential (13).

our series, we found seven mutations at codon 249, at differentsites (two in the first, three in the second and two in third), ofdifferent types (five G to T and two A to T), and in differenttypes of hepatitis virus infection (three HBV, three HCV, andthe last one not tested). Our data suggest that codon 249 is acommon hot spot for p53 mutations in HCC, occurring notonly in aflatoxin high-exposure areas but also in low-exposureareas, and that HBV infection is not an indispensable factor forG to T transversion at this site. Whereas it is conceivable thatp53 mutations in aflatoxin high-exposure areas are a directresult of aflatoxin B, adduction to guanine residues, this is notthe case for HCCs in low-exposure areas.

p53 gene mutation hot spots in HCC, domains IV and V,were particularly conspicuous in poorly differentiated HCC respect to wide distribution in well and moderately differentiatedHCCs. This trend was emphasized when the frameshift mutations, which also affect the structure of domains IV and V, 6 inthe poorly differentiated group and 4 in the moderately differentiated group (see Fig. 3), were taken into account (P < 0.05).Therefore, we hypothesize that damage to domains IV and Vsurrounding codon 249 largely destroys the tumor-suppressingactivity of p53, providing a considerable growth advantage,whereas mutation outside these regions results in only weaklyaltered function. Because domains IV and V constitute one oftwo binding sites for SV40 T-antigen in transformed cell lines(41), these regions might play an important role as a bindingsite for an unknown protein in vivo.

The result of RFLP analysis suggested that almost all (94%)tumors with mutation of the p53 gene had loss of the otheralíele.This result confirms the idea that inactivation of p53requires alterations of both alíeles,indicating that a recessiverole is predominant in HCC progression rather than a dominant negative role. It is conceivable that mutations in bothalíelesmay inactivate p53 gene; however, we never found suchcases. p53 gene seems to be mostly inactivated by a combinationof loss of one alíeleand mutation in the remaining alíele.On theother hand, 38% of tumors which carried one p53 alíeledid notshow p53 mutations. Furthermore, 35% of well differentiated

HCCs showed LOH on 17p,6 in contrast to the low incidence

(9%) of p53 mutation at this stage. These observations suggestthat in Japanese HCC, at variance with the model proposed forcolon cancer (42) and HBV-related HCCs (40), allelic loss mayprecede the p53 mutation on the other alíele.However, thepossibility that the SSCP method may underestimate some portion of mutations, although this possibility is quite low (<10%)(43), and the possible presence of additional tumor suppressorgenes mapping to 17p (44) should also be taken into consideration.

Most of the p53 mutations (94%) and LOH on chromosome17p (85%)6 were observed in moderately and poorly differenti

ated HCC. Moreover, morphological examination of the welldifferentiated cases with p53 mutation revealed that a significant proportion of the tumor was composed of EdmondsonGrade II cells. Therefore, these genetic abnormalities may occur largely at the stage of progression from well to moderatelyor to poorly differentiated HCC.

Our comprehensive analysis of this large series of HCCssuggests a role for p53 gene mutation and LOH on 17p in thelate stage of the multistage process (45) of hepatocarcinogenesis. A tentative model of progression of HCC is shown in Fig.5. However, the earliest and the latest genetic events in hepatocarcinogenesis remain undetermined, and identification ofsuch genetic alterations seems to be important for further understanding the molecular basis of multistage hepatocarcinogenesis.

ACKNOWLEDGMENTS

The authors are grateful to Drs. T. Sekiya, K. Hayashi, and \.Murakami of the Oncogene Division for useful technical advice and thestaff of the Department of Surgery, National Cancer Center Hospital(Tokyo, Japan) and Dr. \. Shimamura of the Department of Surgery,National Sanatorium Matsudo Hospital, for providing surgically resected specimens. Polymorphic DNA probes were obtained from theAmerican Type Culture Collection (Rockville, MD) and from the Japanese Cancer Research Resources Bank (Tokyo, Japan).

REFERENCES1. Vogelstein. B.. Fearon. E. R.. Hamilton. S. R.. Kern, S. E., Preisinger, A. C,

Leppert. M.. Nakamura, Y.. White, R., Smits, A. M. M., and Bos, J. L.Genetic alterations during colorectal-tumor development. N. Engl. J. Med.,319: 525-532, 1988.

2. Tiollais, P., Pourcel, C., and Dejean, A. The hepatitis B virus. Nature(Lond.).J/7: 489-495, 1985.

3. zur Hausen, H. Human papillomaviruses in the pathogenesis of anogenitalcancer. Virology, 184:9-\3, 1991.

4. Bos, J. L. raj oncogenes in human cancer: a review. Cancer Res., 49: 4682-4689, 1989.

5. Knudson. A. G. Hereditary cancer, oncogenes, and anti-oncogenes. CancerRes., 45: 1437-1443. 1985.

6. Simonctti, R. G., Camma, C., Fiorello, F., Politi. F., D'amico, G., and

Pagliaro, L. Hepatocellular carcinoma: a worldwide problem and the majorrisk factors. Dig. Dis. Sci., 36: 962-972. 1991.

7. Tanaka. K.. Hirohata. T., Koga. S., Sugimachi. K.. Kanematsu, T., Ohryohji.F., Nawata, H., Ishibashi, H., Maeda, Y., Kiyokawa. H., Tokunaga, K., Irita.Y., Takashita. S., Arase. Y., and Nishino, N. Hepatitis C and hepatitis B inthe etiology of hepatocellular carcinoma in the Japanese population. CancerRes., 51: 2842-2847. 1991.

8. Nalpas. B.. Driss, F., Pol, S.. Hamelin. B.. Housset. C.. Brecho!, C., andIta lin-loi. P. Association between HCV and HBV infection in hepatocellularcarcinoma and alcoholic liver disease. J. Hepatol.. 12: 70-74, 1991.

9. Tsuda, H., Hirohashi, S., Shimosato. Y., Yoshida, T., and Terada. M. Lowincidence of point mutation of c-Ki-ras and N-ras oncogenes in human hepatocellular carcinoma. Jpn. J. Cancer Res., 80: 196-199. 1989.

10. Takada. S.. and Koike, K. Activated N-rai gene was found in humanhepatoma tissue but only in a small fraction of the tumor cells. Oncogene, 4189-193, 1989.

11. Wang. H. P., and Rogler. C. E. Deletions in human chromosome arms 1Ip

6 H. Tsuda, T. Oda, and S. Hirohashi. manuscript in preparation.

6363

pS3 MUTATION SPECTRUM IN HOC

and 13q in primary hepaloccllular carcinoma. Cytogenet. Cell. Genet., 48:72-78.1988.

12. Buelow. K. H.. Murray. J. C., Israel. J. L.. London. W. T., Smith, M.. Kew.M.. Bkmquct. V.. Brechot, C., Redeker. A., and Govindarajah. S. Loss ofheterozygosity suggests tumor suppressor gene responsible for primary hepa-tocellular carcinoma. Proc. Nati. Acad. Sci. USA. 86: 8852-8856, 1989.

13. Tsuda, H., Zhang. W., Shimosato, Y., Yokoto, J., Terada, M., Sugimura, T.,Miyamura, T.. and Hirohashi. S. Alíeleloss on chromosome 16 associatedwith progression of human hcpatocellular carcinoma. Proc. Nati. Acad. Sci.USA. «7:6791-6794. 1990.

14. Zhang. W., Hirohashi, S.. Tsuda. H., Shimosato, Y., Yokota. J., Terada. M.,and Sugimura, T. Frequent loss of heterozygosity of chromosomes 16 and 4in human hcpatocellular carcinoma. Jpn. J. Cancer Res., 81: 108-111. 1990.

15. Fujimori. M., Tokino, T., Hiño,O., Kitagawa. T., Imamura, T., Okamoto,E., Mitsunobu, M., Ishikawa, T., Nakagama, H.. Harada, H., Yagura, M.,Matsubara, K., and Nakamura. Y. Allelotype study of primary hepatocellularcarcinoma. Cancer Res.. 51: 89-93, 1991.

16. Simon. D., Knowles. B. B., and Weith, A. Abnormalities of chromosome 1and loss of heterozygosity of l p in primary hepatomas. Oncogene, 6: 765-770, 1991.

17. Levein, A. J., Momand, J., and Finlay, C. A. The p53 tumour suppressorgene. Nature (Lond.), 351: 453-456, 1991.

18. Hollstein. M., Sidransky. D., Vogelstein, B., and Harris, C. C. p53 mutationsin human cancers. Science (Washington DC), 253: 49-53, 1991.

19. Hsu. I. C.. Metcalf. R. A.. Sun, T., Welsh, J. A., Wang. N. J., and Harris, C.C. Mutational hotspot in thÃp̈53 gene in human hepatocellular carcinomas.Nature (Lond.). 350: 427-428, 1991.

20. Bressac. B., Kew. M., Wands. J., and Ozturk, M. Selective G to T mutationof p53 gene in hepatoccllular carcinoma from southern Africa. Nature(Lond.). 350: 429-431, 1991.

21. Murakami. Y., Hayashi, K., Hirohashi. S., and Sekiya, T. Aberrations of thetumor suppressor p53 and retinoblastoma genes in human hepatocellularcarcinomas. Cancer Res., 51: 5520-5525. 1991.

22. Jones. P. A., Buckley. J. D.. Henderson. B. E.. Ross, R. K.. and Pike. M. C.From gene to carcinogen: a rapidly evolving field in molecular epidemiology.Cancer Res., 51: 3617-3620, 1991.

23. Zarbel. H., Sukumar, S., Arthur, A. V., Martin-Zanca. D., and Barbacid, M.Direct mutagenesis of Ha-ra.v-/ oncogenes by /V-nitroso-A'-mcthylurea duringinitiation of mammary carcinogenesis in rats. Nature (Lond.). 315: 382-385.1985.

24. Quintanilla. M.. Brown. K.. Ramsden. M.. and Balmain. A. Carcinogen-specific mutation and amplification of Ha-ras during mouse skin carcinogenesis. Nature (Lond.). 322: 78-80. 1986.

25. Chen, R-H., Mäher,V. M., and McCormick, J. J. Effect of excision repair bydiploid human fibroblasts on the kinds and locations of mutations induced by(±)-7tf,8cv-dihydroxy-9iv,1Od-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene in thecoding region of the HPRTgcne. Proc. Nati. Acad. Sci. USA. 87:8680-8684,1990.

26. Adelman. R.. Saul, R. L., and Ames. B. N. Oxidative damage to DNA:relation to species metabolic rate and life span. Proc. Nati. Acad. Sci. USA,85: 2706-2708, 1988.

27. Abadic, V., Lyonnet, S., Maurin, N., Bertheion, M., Caillaud, C., Giraud, F.,

Mattei, J-F., Rey, J., Rey, F., and Munnich, A. CpG dinucleotides are mutation hot spots in phenylketonuria. Genomics. 5: 936-939, 1989.

28. Foster. P. L., Eisenstadt, E., and Miller, J. H. Base substitution mutationsinduced by metabolically activated aflatoxin K¡.Proc. Nati. Acad. Sci. USA,80: 2695-2698. 1983.

29. Ozturk, M., and Collaborators. p53 mutation in hepatocellular carcinomaafter aflatoxin exposure. Lancet, 338: 1356-1359, 1991.

30. Graham. W.. and Cooksley. E. Hepatocellular carcinoma mutation. Nature(Lond.), 352: 764, 1991.

31. Herskowitz, I. Functional inactivation of gene by dominant negative mutations. Nature (Lond.), 329: 219-222, 1987.

32. Kanai. T., Hirohoshi, S., and Shimosato, Y. Pathology of small hepatocellular carcinoma (a proposal for a new gross classification). Cancer (Phila.),60:810-819, 1987.

33. Oda, T., Tsuda, H., Scarpa, A., Sakamoto, M., and Hirohashi, S. Mutationpattern of the p53 gene as a diagnostic marker for multiple hepatocellularcarcinoma. Cancer Res., 52: 3674-3678, 1992.

34. Sakamoto, M., Hirohashi. S., and Shimosato, Y. Early stages of multiplehepatocarcinogenesis: adenomatous hyperplasia and early hepatocellular carcinoma. Hum. Pathol., 22: 172-178, 1991.

35. Sambrook, J., Fritsch, E. F., and Maniatis, T. Molecular Cloning: a Laboratory Manual, pp. E.3-E.4. Cold Spring Harbor, NY: Cold Spring HarborLaboratory. 1989.

36. Orila, M.. Suzuki, Y., and Hayashi, K. Rapid and sensitive detection of pointmutation and DNA polymorphisms using the polymerase chain reaction.Genomics, 5: 874-879, 1989.

37. Kidd, K. K., Klinger, H. P., and Ruddle. F. H. Human gene mapping 10.Cytogenet. Cell Genet., 51: 1-1148. 1989.

38. Soussi, T., Fromental, C. D., and May, P. Structural aspects of the p53protein in relation to gene evolution. Oncogene, 5: 945-952, 1990.

39. Vogelstein, B., and Kinzler, K. W. Carcinogens leave fingerprints. Nature(Lond.), 355: 209-210, 1992.

40. Scorsone, K. A., Yi-Zhou, Butel. J. S., and Slagle. B. L. p53 mutations clusterat codon 249 in hepatitis B virus-positive hepatocellular carcinomas fromChina. Cancer Res.. 52: 1635-1638. 1992.

41. Jenkins, J. R., Chumakov, P., Addison, C., Sturzbecher, H-W., and Wade-Evans, A. Two distinct regions of the murine p53 primary amino acid sequence are implicated in stable complex formation with simian virus 40 Tantigen. J. Virol, 62: 3903-3906. 1988.

42. Baker, S. J.. Preisinger. A. C.. Jessup, J. M., Paraskeva, C.. Markowitz, S.,Willson, J. K. V., Hamilton, S., and Vogelstein, B. p53 gene mutations occurin combination with 17p allelic deletions as late event in coloréela!tumori-genesis. Cancer Res., 50: 7717-7722, 1990.

43. Hayashi, K. PCR-SSCP: a simple and sensitive method for detection ofmutations in the genomic DNA. PCR Methods Applications. /: 34-38, 1991.

44. Sato, T., Tanigami. A., Yamakawa, K., Akiyama, F., Kasumi, F., Sakamoto,G., and Nakamura, Y. Allelotype of breast cancer: cumulative alíelelossespromote tumor progression in primary breast cancer. Cancer Res., 50:7184-7189, 1990.

45. Fearon, E. R., and Vogelstein, B. A genetic model for colorectal tumorigen-esis. Cell, 61: 759-767, 1990.

6364