Clinicopathologic significance of genetic alterations in hepatocellular carcinoma

Cancer Genetics and Cytogenetics 146 (2003) 8–15

Clinicopathologic significance of genetic alterations in hepatocellular carcinoma A. Pang, I.O. Ng, S.T. Fan, Y.L. Kwong* Departments of Medicine, Pathology, and Surgery, University Department of Medicine Professorial Block, Queen Mary Hospital, Pokfulam Road, Hong Kong Received 23 January 2003; accepted 25 February 2003

Abstract

Hepatocarcinogenesis may involve multiple mutations with distinctive pathogenetic and clinicopathologic significance. To test this hypothesis, 68 cases of hepatocellular carcinoma (HCC) were studied prospectively for genetic–clinicopathologic correlation. Ten pathologic characteristics were evaluated. TP53 (alias p53) gene mutation was studied by a polymerase chain reaction (PCR)– single-strand conformation polymorphism–sequencing; CDKN2B (alias p15) and CDKN2A (alias p16) gene methylation by methylation-specific PCR; and genetic imbalances by comparative genomic hybridization (CGH). TP53 gene mutations occurred in 25% of cases, more than half being codon 249 G to T transversion. Methylation of CDKN2A was frequent (61.7%); of CDKN2B, rare (5.9%). The CGH analysis showed a median of nine aberrations per case, with amplifications more frequent than deletions. Isochromosomes might be involved in about 25% of cases. Amplifications of 1q and 8q were most frequent. Clinicopathologic correlations showed that CDKN2A methylation was significantly associated with tumors arising in cirrhotic livers; amplifications of 17q was significant in multiple parameters of tumor invasiveness (size, venous invasion, poor cellular differentiation, microsatellite formation); other amplifications (1q, 6p, 10p, and 20p) were also significant in tumor invasion; and deletions (at 1p, 11q, 4q, and 14q) were significant in tumor growth. Consistent patterns of genetic alterations were defined in HCC, which might represent distinctive pathways in hepatocarcinogenesis. 쑖 2003 Elsevier Inc. All rights reserved.

1. Introduction Hepatocellular carcinoma (HCC) is one of the most common cancers in the world [1], with a marked geographic variation in incidence, being very prevalent in sub-Saharan Africa and Asia, but uncommon in the West [1]. This is due to different risk factors. Chronic hepatitis B virus (HBV) infection [2] and exposure to the carcinogen aflatoxin [3] are important risk factors for African and Asian populations. Chronic HBV carriers have a relative risk of nearly 200: 1 for developing HCC [2], which may be due to cirrhosis and the oncogenic potential of the HBV genome [4]. The carcinogenic potential of aflatoxin relates to its capacity to induce G:C to T:A transversion in the TP53 gene, with specific clustering at the third base of codon 249 [5]. It also acts synergistically with HBV infection to increase the risk of HCC [6]. Chronic hepatitis C virus (HCV) is a risk factor * Corresponding author. Tel.: ⫹852-2-855-4597; fax: ⫹852-2-974-1165 E-mail address: [email protected] (Y.L. Kwong). 0165-4608/03/$ – see front matter 쑖 2003 Elsevier Inc. All rights reserved. doi: 10.1016/S0165-4608(03)00103-1

in the West [7], although the underlying pathogenetic mechanisms remain undefined. Hepatocellular carcinoma has a poor prognosis, with a 5-year survival of less than 3% in inoperable cases. Genetic mutations also play significant roles in hepatocarcinogenesis. The TP53 gene is pivotal in maintaining genomic stability, and mutations of the gene have been implicated in HCC formation. In addition to the clustering of mutations to codon 249 in aflatoxin prevalent areas [3,5], mutations also occur in other exons in sporadic cases in most parts of the world [8–10]. CDKN2A (alias p16, MTS1) is another candidate gene involved in hepatocarcinogenesis [10]. It is a potent cyclindependent kinase inhibitor [11] that halts cell-cycle progression at the G1–S phase boundary. Loss of CDKN2A function may lead to unregulated cellular proliferation [11]. The localization of the CDKN2A gene to 9p21, which is often deleted in HCC [12], suggests that it may be a tumor-suppressor gene. Interestingly, aberrations of CDKN2A appear to show some heterogeneity in different localities. Several studies in

A. Pang et al. / Cancer Genetics and Cytogenetics 146 (2003) 8–15

Japan, Australia, and Taiwan failed to show any homozygous deletion, mutation, or methylation inactivation of the CDKN2A gene in HCC [12–14]. On the other hand, studies from Hong Kong showed that CDKN2A was often inactivated in HCC, predominantly through methylation of the 5′ CpG island in the promoter region [15]. The reason for this difference is unclear. CDKN2B (alias p15, MTS2) is another inhibitor of the cyclin-dependent kinases CDK4 and CDK6 [16]. Methylation inactivation of CDKN2B occurred predominantly in hematologic malignancies [16]. It also is localized to 9p21 and so may also be a candidate gene in HCC. Hepatocarcinogenesis is likely to involve multiple steps of genetic mutations. To test the hypotheses that the different genetic alterations involved in this multistep process may have distinctive pathogenetic roles; that these aberrations may be interrelated; and that they may be of clinicopathologic importance, we conducted a prospective study of HCC patients for TP53 gene mutation, CDKN2B and CDKN2A gene methylation, and genomic aberrations with comparative genomic hybridization (CGH). We further examined if these genetic lesions were interrelated, and defined their significance with a comprehensive panel of clinicopathologic parameters.

2. Materials and methods 2.1. Patients Sixty-eight consecutive patients with HCC undergoing surgical resection at a single institute from 1998–99 were included in the study. Surgically resected tumor specimens were snap-frozen until analysis. 2.2. Pathologic evaluation Pathologic evaluation was performed as previously described [17] by one of the investigators (I.O.N.) prospectively and without prior knowledge of the results of the genetic analyses. Ten pathologic characteristics were noted, including gross assessment of tumor size and number of tumor nodules, presence of a tumor capsule, intrahepatic spread of the tumor as evidenced by tumor microsatellite formation and venous invasion, direct invasion into the adjacent liver parenchyma, presence of cirrhosis and chronic hepatitis in the nontumorous liver, histologic evidence of tumor involvement of the resection margin, and cellular differentiation according to the criteria of Edmondson and Steiner [18].

9

products were then analyzed by SSCP. The SSCP conditions for detecting single-base mutations were optimized by analysis of positive control samples with specific mutations. Tumor specimens with abnormal band shifts on SSCP were purified and sequenced bidirectionally with an automated DNA sequence analyzer (ABI Prism 377, PE Biosystems, Foster City, CA). 2.3.2. Study of CDKN2B and CDKN2A gene promoter methylation Methylation of CDKN2B and CDKN2A gene promoter regions was studied with methylation-specific PCR (MS-PCR) [20], using a commercially available kit (CpGenome DNA modification kit, Intergen, Purchase, NY) according to the manufacturer’s instructions. The primers for the CDKN2B and CDKN2A gene promoter regions were as reported [20]. Positive and negative controls were included in all the experiments. 2.3.4. Comparative genomic hybridization (CGH) CGH was performed as previously described [21]. Briefly, tumor and normal reference DNA was labeled by nick translation (Vysis, Naperville, IL) with SpectrumGreen dUTP and SpectrumRed dUTP (Vysis), respectively, and then coprecipitated with Cot-1 DNA (Gibco BRL, Gaithersburg, MD), denatured, preannealed, and hybridized to denatured normal metaphase spreads (Vysis). After hybridization and washing, the metaphase chromosomes were counterstained with 4′,6-diamidino-2-phenylindole (DAPI), captured with a charge-coupled device (CCD) camera, and analyzed with the CytoVision digital imaging system (version 3.1, Applied Imaging, Santa Clara, CA). Green-to-red fluorescence ratio values of 1.25 and 0.75 were used as upper and lower thresholds for the identification of chromosomal imbalances. The telomeric and heterochromatic regions of 1p32~pter, 16p, 19, and 22 were interpreted cautiously; centromeric and heterochromatic regions were not analyzed, because the CotI DNA inhibited DNA binding to these regions. Optimization of CGH conditions to detect unequivocally the genetic imbalances of the control tumor cell line MPE-600 (Vysis) was included as previously described as quality assurance of the experiments [21]. 2.4. Statistical analysis Fisher’s exact and chi-square tests were used for the analysis of categorical data and analysis of variance, t-test, and Kruskal–Wallis test were used for continuous data as appropriate. The correlation analysis was performed with Pearson product-moment correlation coefficient. Tests were considered significant when the P values were less than 0.05.

2.3. Genetic analyses 2.3.1. Analysis of TP53 mutation Analysis of TP53 gene mutation was performed by polymerase chain reaction (PCR), single-strand conformational polymorphism (SSCP), and DNA sequencing. Briefly, exons 4 to 9 of the TP53 gene were amplified with PCR, using appropriate primers as previously published [19]. The PCR

3. Results 3.1. Patients There were 52 male and 16 female patients (M :F = 3.3:1), with a median age of 51 (13–79) years. HBV was positive in 49 cases (76.5%) and HCV in 2 cases (3%).

10

A. Pang et al. / Cancer Genetics and Cytogenetics 146 (2003) 8–15

3.2. Histologic evaluation The mean size of the tumors was 7.8 cm (range: 1.5– 25.5). In the surrounding liver parenchyma, cirrhosis was found in 30 cases (44.1%), chronic hepatitis in 33 cases (48.5%), and apparently normal histology in 5 cases (7.4%). 3.3. Genetic findings 3.3.1. TP53 mutation Seventeen (25%) of the cases were found to have TP53 gene mutations, located at exons 5 to 9 (Table 1). Nine of these were codon 249 G to T transversions. Seven point mutations and one deletion were detected in the other coding regions of TP53, all resulting in amino acid substitutions or premature chain termination. 3.3.2. CDKN2B and CDKN2A methylation Methylation of CDKN2A gene promoter region was detected in 42 (61.7%) of the cases, and of the CDKN2B gene in only 4 cases (5.9%). 3.3.3. CGH Multiple recurring chromosomal aberrations were detected by CGH (Fig. 1). The changes were typically complex, with a median of 9 (range 0–26) aberrations per case. Tumors

with or without TP53 gene mutations had comparable number of aberrations (10.6 vs. 8.5, not significant). Chromosomal amplifications were the predominant genomic alterations (Table 1), with commonly involved regions identifiable. Some amplifications, particularly in 1q and 8q, might involve more than two copies of the loci (Fig. 2). Chromosomal deletions were less frequent, with commonly deleted regions also identifiable (Table 1). In 12 cases, a gain of the p arm was associated with a loss of the q arm; the involved chromosomes were 2 (n ⫽ 1), 3 (n ⫽ 1), 4 (n ⫽ 2), 5 (n ⫽ 1), 6 (n ⫽ 3), 10 (n ⫽ 3), and 12 (n ⫽ 1). A loss of the p arm with a gain of the q arm was found in 19 cases; the involved chromosomes were 1 (n ⫽ 4), 2 (n ⫽ 1), 3 (n ⫽ 1), 8 (n ⫽ 6), 9 (n ⫽ 1), 10 (n ⫽ 1), 11 (n ⫽ 1), 17 (n ⫽ 3), and 19 (n ⫽ 1). These changes appear to be equivalent to isochromosomes p or q, respectively (Fig. 2). 3.4. Clinicopathologic correlations Correlation between the genetic alterations and clinicopathologic features revealed several significant associations (Tables 2 and 3). 3.4.1. CDKN2A methylation and cirrhosis of the liver When the liver was cirrhotic, tumors showed a significantly higher frequency (P ⫽ 0.006) of CDKN2A gene promoter methylation as compared with tumors found in livers with normal histology or chronic active hepatitis (Table 3).

Table 1 Frequency of genetic alterations in 68 cases of hepatocellular carcinoma (HCC) Genetic alteration

Chromosomal region

Frequency (%)

Aberration involved

TP53 mutation

—

25

G 2575→T, Arg 249→Ser (n ⫽ 9) A 1575→G, Asn 131→Ser (n ⫽ 2) A 2983→T, Glu 271→Val (n ⫽ 1) C 3015→T, Arg 282→Leu (n ⫽ 1) G 2995→T, Cys 275→Phe (n ⫽ 1) A 3042→T, Lys 291→Term (n ⫽ 1) C 3185→A, Leu 308→Met (n ⫽ 1)a C 3230→A, Leu 323→Met (n ⫽ 1)a del 1910–1919 (TGGTGCCCTA), resulting in a termination codon at position 246

CDKN2A methylation CDKN2B methylation Amplification

— — 1q 8q 7q 19q 5p 5q 20q 13p 9q 17q 20q 2q 6p 15q 4q 16q 1p 13q

61.7 5.9 75.0 67.6 35.3 29.4 29.4 29.4 29.4 27.9 26.4 26.4 25.0 25.0 23.5 22.1 32.3 26.4 20.6 20.6

Deletion

a

Occurring in the same case.

1q23→q32 8q22→qter 7q11→q31 19q10→qter 5p13→pter 5q15→q22 13q14→qter 20p10→pter 9q12→q22 17q10→q25 20q10→qter 2q22→q31 6p21→pter 15q14→qter 4q12→q31 16q13→qter 1q34→pter 13q14→q22

A. Pang et al. / Cancer Genetics and Cytogenetics 146 (2003) 8–15

3.4.2. Tumor size Chromosomal deletions were the predominant aberrations associated with large tumors. Deletion 11q was most significant (P ⬍ 0.001) and was associated with the largest mean tumor size in this series. It was, however, only an infrequent aberration (7.3% of cases). Deletions of 14q and 4q were more frequent aberrations (32.3% and 17.6%) associated with large tumors. Amplification of 17q was the only amplification that was related to the larger tumor size. 3.4.3. Venous invasion In contrast to tumor size, chromosomal amplifications were the predominant aberration related to venous invasion. Although venous invasion associated with 17q amplification might also be due to the larger tumor size, amplifications at 1q and 6p appeared to be independent factors. Deletion of 11q was the only deletion associated with venous invasion, which might in part be due to the associated larger tumor size. 3.4.4. Encapsulation Amplification of 10q was the only aberration significantly related to the absence of tumor encapsulation. 3.4.5. Microsatellite formation Amplifications of 17q and 20p were significantly associated with tumor microsatellite formation, which was a parameter of intrahepatic tumor metastasis. 3.4.6. Cellular differentiation Amplification of 17q was the only aberration associated with poorer cellular differentiation (Edmondson grades 3–4). It was particularly impressive that all the tumors with 17q amplifications in this series (n ⫽ 17) were classified as poorly differentiated. 3.4.7. Number of tumor nodules Deletion of chromosome 1p was related to tumors with more than one primary nodules.

4. Discussion The study of genetic alterations in HCC may give important information on hepatocarcinogenesis. Karyotyping of HCC has limited success, however, owing to the low mitotic index and poor metaphases. Genetic studies have therefore focused on alterations of single genes, including TP53 and CDKN2A [8–12,14–16]. Comparative genomic hybridization, a technique that assesses the overall genetic lesions in a tumor, has recently been used in the study of HCC [22–26]. In this report, we have prospectively studied a cohort of HCC with a comprehensive panel of pathologic parameters for correlation with genetic aberrations. This genetic– clinicopathologic correlation has enabled us to examine the significance of the various genetic alterations in greater detail

11

than other previous studies [14,15,22–26], and to speculate on their possible roles in hepatocarcinogenesis. The frequency of TP53 gene mutation was 25% (17/68), of which 13% (9/68) were at codon 249. This frequency was similar to that seen in previous studies of Chinese and Japanese patients [8,9,27]. Local foodstuff aflatoxin levels were comparable to Western countries [9] with a lower TP53 gene mutation rate (12%–18%) [28]. The slightly higher TP53 gene mutation rate in our population may therefore be due to a synergistic interaction of HBV and aflatoxin in increasing TP53 gene mutation [8]. In the present study, however, we have not been able to detect any clinicopathologic significance of TP53 gene mutations. A high frequency of methylation of the CDKN2A gene was observed, which was higher than in other Asian (Japanese and Taiwanese) patients [12,14]. A novel finding was the highly significant association of tumor CDKN2A methylation with cirrhosis of the liver. Cirrhosis is an important step in hepatocarcinogenesis. During regeneration, the chances of genetic mutations are increased in hepatocytes undergoing active cell division [10]. DNA methylation plays an important regulatory role in gene transcription, particularly during cell division [29]. It is catalyzed by DNA methyltransferase (DNA-MCMT), which targets CpG islands. The coordination of DNA methylation with DNA replication appears to be mediated through differential binding of DNA-MCMT to the protein proliferating cell nuclear antigen (PCNA) during cell cycle [29,30], it being decreased prior to cell division but upregulated late in the cell cycle. During the regeneration of hepatocytes characteristic of cirrhosis, it is possible that the increased cell divisions enhance the chances of perturbation of the repeated up- and downregulation of DNA methylation, leading to abnormal methylation of critical genes such as CDKN2A. This hypothesis will have to be tested by further studies of the control of DNA methylation in cirrhosis. In the present study, genetic amplifications appeared to be the predominant lesion, in contrast to other solid tumors, for which genetic deletions are often prevalent [31]. Another new observation was the apparent formation of isochromosomes involving many chromosomes. Isochromosomes arise from mitotic nondisjunction followed by centromere misdivision. Recent molecular studies, however, showed that in many cases distinct breakpoints might occur outside the centromere [32]. Our results therefore provide novel sites for investigation of possible relevant breakpoints in the involved chromosomes. Amplification of chromosome 1q was the most frequent (75%) genetic imbalance in the present study. It was also a consistent abnormality in other CGH studies of HCC, occurring in 20%–78% of cases [22–26]. Neither the pathogenetic role nor the pathologic significance of 1q amplification was defined in previous studies. Here we have shown, to our knowledge for the first time, that 1q amplification was significantly associated with venous invasion. Interestingly, 1q amplification appears to be a very common aberration

12

A. Pang et al. / Cancer Genetics and Cytogenetics 146 (2003) 8–15

A. Pang et al. / Cancer Genetics and Cytogenetics 146 (2003) 8–15 䉳 Fig. 1. Summary of genetic imbalances detected by comparative genomic hybridization (CGH) in 68 cases of hepatocellular carcinoma (HCC). Chromosome losses are indicated by red lines on the left of each ideogram; green lines on the right represent chromosome gains. Each line represents an individual tumor in which the aberration was identified.

in other types of mesenchymal tumors, including osteosarcoma, fibrous histiocytoma, leiomyosarcoma, synovial sarcoma, and alveolar rhabdomyosarcoma [33]. This suggests the amplification of one or more critical protooncogenes in this region, although candidate genes have yet to be identified. Amplification of 8q was the second most common aberration (67.6%), comparable to previous studies showing a frequency of 30%–66% [22–26]. The commonly amplified region was 8q22~8qter, and contained the protooncogene c-myc. In murine models, overexpression of c-myc in the liver caused HCC [34]. Data from human studies were limited, but one study showed amplification of MYC (alias c-MYC) in ~50% of HCC, which predicted a poor prognosis [35]. In the present study, no significant pathologic features were associated with 8q amplification. The high frequencies of 1q and 8q amplifications, however, suggest that they play an important role in hepatocarcinogenesis. Amplification at 17q was the most important aberration in the correlation with the pathologic features examined, being significantly associated with large tumor size, tumor venous invasion, poorer cellular differentiation, and intrahepatic tumor microsatellite formation. The association with cellular differentiation was particularly impressive, with 17q amplification occurring exclusively in poorly differentiated tumors. The results suggest that amplification of a gene or genes at 17q may be critical in HCC progression. Amplification of 17q is frequent in many tumor types, including lung, stomach, breast, and testicular cancers [33]. A possible candidate gene is ERBB2 at 17q11~q12, which has been shown by fluorescence in situ hybridization (FISH) and microarray experiments to be amplified in breast and gastric cancers [33,36]. ERBB2 amplification or overexpression is an important indicator of poor prognosis in many cancers. As no data of ERBB2 in HCC are currently available, our observations call for further studies to define if ERBB2 is involved 䉳 Fig. 2. Representative partial fluorescence ratio profiles of genetic imbalances in HCC. The central line represents a ratio value of 1.0. Green lines to the right indicate ratio values of 1.25 and 1.5; red lines to the left indicate ratio values of 0.75 and 0.5; and n represents the number of homologs examined. (A) Chromosome 1p deletion and 1q amplification, respectively. In the amplification profile, the ratio was ⬎1.5, suggesting that the amplification might exceed two copies. (B ) Chromosome 4q deletion. (C ) Chromosome 6p amplification with a ratio ⬎1.5. (D ) 8p deletion and 8q amplification. (E ) Concomitant 10p amplification and 10q deletion. The profile suggested the cytogenetic equivalence of an isochromosome 10p. (F ) Two profiles of chromosome 11, showing 11q deletion. In the second profile, concomitant amplification of 11p occurred. (G ) Chromosome 14q deletion. (H ) two profiles of chromosome 17, showing varying degrees of 17q amplification.

13

Table 2 Correlation between genetic aberrations and clinicopathologic features Feature

Chromosomal region

Tumor size (cm) 14q

Value

Normal (no.)

Amp (no.)

Del (no.)

Significancea

23

P ⫽ 0.028

7.2 6.4 11.1 6.9 10.6 6.6 5.4 10.9 7.0 7.7 16.3

47

⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹

13 4 32 19 31 18 30 23

24 26 5 11 5 12 7 2

— — — — — — 0 5

⫺ ⫹

31 27

8 0

1 1

Cellular differentiation (Edmondson’s grade) 17q 1–2 21 0 3–4 29 17

1 0

17q 4q

11q

Venous invasion 1q 6p 17q 11q Encapsulation 10p

Microsatellite 17q 20p Tumor nodules (no.) 1p

8 50 P ⫽ 0.025

16 38 7 22

P ⫽ 0.002

5

P ⬍ 0.001

53 9

⫺ ⫹ ⫺ ⫹

29 20 28 19

5 12 6 13

1 ⬎1

23 3

3 0

P ⫽ 0.038 P ⫽ 0.027 P ⫽ 0.036 P ⫽ 0.018 P ⫽ 0.042

P ⫽ 0.002

P ⫽ 0.033 P ⫽ 0.036 8 6

P ⫽ 0.044

Abbreviations: Amp, amplification; Del, deletion; ⫹, present; ⫺, absent. a Refers to the correlation between features (column 1) and chromosomal changes (columns 4–6).

in hepatocarcinogenesis. Other studies have shown, however, that the amplified regions in 17q are more telomeric to ERBB2 [33], suggesting that different protooncogenes may be involved. Chromosomal amplifications at other regions were also associated with tumor invasiveness, including venous invasion, microsatellite formation, and tumor encapsulation. Amplification at 6p was associated with venous invasion. This region is amplified most often in uveal melanoma, but also in lung, breast, and ovarian cancers [33]. Amplification at Table 3 Correlation between CDKN2A methylation and cirrhosis of the liver Coexisting cirrhosis of liver

CDKN2A gene, methylated

Significance

Present (n ⫽ 20) Absent (n ⫽ 6)

Present (n ⫽ 18) Absent (n ⫽ 24)

— P ⫽ 0.006

14

A. Pang et al. / Cancer Genetics and Cytogenetics 146 (2003) 8–15

related to large tumors. Loss of heterozygosity studies have shown consistent 4q deletions in esophageal and bladder cancer [31]. Deletions at 14q appear to be particularly common in renal cancers [31]. Therefore, these regions might also contain genes that are involved in controlling tumor growth in HCC. In conclusion, the correlation between genetic and clinicopathologic features enables us to address our hypothesis by constructing a model of a multistep oncogenic process of HCC in our patients (Fig. 3). This model forms a conceptual framework for further investigations to localize genes important in the initiation and progression of HCC.

Acknowledgments The authors thank C.M. Lo, R.T.P. Poon for clinical care of the patients; Chris Tam and Eunice Chan for technical help; and M Ng for assistance in statistical analysis. This study was supported by grant no. HKU 7289/97M from the Research Grant Council.

References Fig. 3. A model of the genetic alterations in hepatocarcinogenesis. Hepatitis B virus (HBV) is the predominant predisposing factor, which may synergize with aflatoxin to increase TP53 gene mutation. Chronic HBV infection leads to cirrhosis, and the associated liver cell regeneration enhances abnormal methylation of the CDKN2A gene, resulting in deregulation of cell cycle control. Amplifications of chromosomes 1q and 8q are most common in tumors, and may lead to overexpression of putative protooncogenes that contribute to tumorigenesis. Amplifications of 17q, 1q, 6p, 10p, and 20p result in enhanced tumor invasiveness, and deletions of 1p, 11q, 4q, and 14q lead to increased tumor growth rate.

20p was associated with microsatellite formation. The gene angiopoietin 4 (ANGPT4) is located on 20p13 and may be relevant in tumor angiogenesis [37] and, therefore, in microsatellite formation. Chromosomal deletions were less frequently observed in the present study. Interestingly, they were the predominant aberrations associated with large tumor size. Deletion at 11q predicted the largest tumors in this series, and was also significantly related to venous invasion. The occurrence of 11q deletion in a small number of cases suggested that it might be a secondary aberration related to tumor progression. At 11q, two candidate genes, ATM (11q22.3) and PPP2R1B (11q22~q24), have been associated with leukemia and with lung and colon cancers, respectively [31]. Deletion of 1p was associated with the presence of more than one tumor nodule. Of the putative tumor-suppressor genes located on 1p [31], the TP73 gene (alias p73) gene at 1p36 has been recently shown to be relevant in HCC; it was expressed in ~30% of cases and was a poor prognostic indicator [38]. Whether the overexpression was due to a mutant protein was not been defined. Deletions at 4q and 14q were also

[1] Bosch FX, Munoz N. Hepatocellular carcinoma in the world: epidemiologic questions. Adv Appl Biotechnol 1991;13:35–56. [2] Beasley RP, Hwang LY, Lin CC, Chien CS. Hepatocellular carcinoma and HBV: a prospective study of 22,707 men in Taiwan. Lancet 1981;2(8256):1129–32. [3] Wogan GN. Aflatoxin as a human carcinogen. Hepatology 1999; 30:573–5. [4] Feitelson MA, Duan LX. Hepatitis B virus x antigen in the pathogenesis of chronic infections and the development of hepatocellular carcinoma. Am J Pathol 1997;150:1141–57. [5] Hsu IC, Metcalf RA, Sun T, Welsh JA, Wang NJ, Harris CC. Mutational hotspot in the p53 gene in human hepatocellular carcinoma. Nature 1991;350:429–31. [6] Sun Z, Lu P, Gail MH, Pee D, Zhang D, Ming L, Wang J, Wu Y, Liu G, Wu Y, Zhu Y. Increased risk of hepatocellular carcinoma in male hepatitis B surface antigen carriers with chronic hepatitis who have detectable aflatoxin metabolite M1. Hepatology 1999;30:379–83. [7] Colombo M. Hepatitis C virus and hepatocellular carcinoma. Semin Liver Dis 1999;19:263–9. [8] Lunn RM, Zhang YJ, Wang LY, Chen CJ, Lee PH, Lee CS, Tsai WY, Santella RM. p53 mutations, chronic hepatitis B virus infection, and aflatoxin exposure in hepatocellular carcinoma in Taiwan. Cancer Res 1997;57:3471–7. [9] Ng IOL, Chung LP, Tsang S, Lam CL, Lai ECS, Fan ST, Ng M. p53 gene mutation spectrum in hepatocellular carcinomas in Hong Kong Chinese. Oncogene 1994;9:985–90. [10] Ozturk M. Genetic aspects of hepatocellular carcinogenesis. Semin Liver Dis 1999;19:235–42. [11] Serrano M, Hannon GJ, Beach D. A new regulatory motif in cellcycle control causing specific inhibition of cyclin D/CDK4. Nature 1993;366:704–7. [12] Hui AM, Sakamoto M, Kanai Y, Ino Y, Gotoh M, Yokota J, Hirohashi S. Inactivation of p16INK4 in hepatocellular carcinoma. Hepatology 1996; 24:575–9. [13] Biden K, Young J, Buttenshaw R, Searle J, Cooksley G, Xu DB, Leggett B. Frequency of mutation and deletion of the tumor suppressor gene CDKN2A (MTS1/p16) in hepatocellular carcinoma from an Australian population. Hepatology 1997;25:593–7.

A. Pang et al. / Cancer Genetics and Cytogenetics 146 (2003) 8–15 [14] Lin YW, Chen CH, Huang GT, Lee PH, Wang JT, Chen DS, Lu FJ, Sheu JC. Infrequent mutations and no methylation of CDKN2A (p16/ MTS1) and CDKN2B (p15/MTS2) in hepatocellular carcinoma in Taiwan. Eur J Cancer 1998;34:1789–95. [15] Liew CT, Li HM, Lo KW, Leow CK, Chan JY, Hin LY, Lau WY, Lai PB, Lim BK, Huang J, Leung WT, Wu S, Lee JC. High frequency of p16INK4A gene alterations in hepatocellular carcinoma. Oncogene 1999;18:789–95. [16] Singal R, Ginder GD. DNA methylation. Blood 1999;93:4059–70. [17] Ng IO, Lai EC, Fan ST, Ng MM, So MK. Prognostic significance of pathologic features of hepatocellular carcinoma: a multivariate analysis of 278 patients. Cancer 1995;76:2443–8. [18] Edmondson HA, Steiner PE. Primary carcinoma of the liver: a study of 100 cases among 48,900 necropsies. Cancer 1954;7:462–503. [19] Murakami Y, Hayashi K, Hirohashi S, Sekiya T. Aberrations of the tumor suppressor p53 and retinoblastoma genes in human hepatocellular carcinomas. Cancer Res 1991;5:5520–5. [20] Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci USA 1996;93:9821–6. [21] Siu LPL, Wong KF, Chan JK, Kwong YL. Comparative genomic hybridization analysis of natural killer cell lymphoma/leukemia: recognition of consistent patterns of genetic alterations. Am J Pathol 1999;155:1419–25. [22] Marchio A, Meddeb M, Pineau P, Danglot G, Tiollais P, Bernheim A, Dejean A. Recurrent chromosomal abnormalities in hepatocellular carcinoma detected by comparative genomic hybridization. Genes Chromosomes Cancer 1997;18:59–65. [23] Wong N, Lai P, Lee SW, Fan S, Pang E, Liew CT, Sheng Z, Lau JWY, Johnson PJ. Assessment of genetic changes in hepatocellular carcinoma by comparative genomic hybridization analysis. Am J Pathol 1999; 154:37–43. [24] Zimonjic DB, Keck CL, Thorgeirsson SS, Popescu NC. Novel recurrent genetic imbalances in human hepatocellular carcinoma cell lines identified by comparative genomic hybridization. Hepatology 1999; 29:1208–14. [25] Kusano N, Shiraishi K, Kubo K, Oga A, Okita K, Sasaki K. Genetic aberrations detected by comparative genomic hybridization in hepatocellular carcinomas: their relationship to clinicopathological features. Hepatology 1999;29:1858–62. [26] Lin YW, Sheu JC, Huang GT, Lee HS, Chen CH, Wang JT, Lee PH, Lu FJ. Chromosomal abnormality in hepatocellular carcinoma by comparative genomic hybridisation in Taiwan. Eur J Cancer 1999; 35:652–8. [27] Hayashi H, Sugio K, Matsumata T, Adachi E, Takenaka K, Sugimachi K. The clinical significance of p53 gene mutation in hepatocellular carcinomas from Japan. Hepatology 1995;22:1702–7.

15

[28] Buetow KH, Sheffield VC, Zhu M, Zhou T, Shen FM, Hino O, Smith M, McMahon BJ, Lanier AP, London WT, Redeker AG, Govindarajan S. Low frequency of p53 mutations observed in a diverse collection of primary hepatocellular carcinomas. Proc Natl Acad Sci USA 1992;89:9622–6. [29] Baylin SB. Tying it all together: epigenetics, genetics, cell cycle, and cancer. Science 1997;277:1948–9. [30] Chuang LSH, Ian HI, Koh TW, Ng HH, Xu G, Li BF. Human DNA(cytosine-5) methyltransferase-PCNA complex as a target for p21WAF1. Science 1997;277:1996–2000. [31] Knuutila S, Aalto Y, Autio K, Bjorkqvist AM, El-Rifai W, Hemmer S, Huhta T, Kettunen E, Kiuru-Kuhlefelt S, Larramendy ML, Lushnikova T, Monni O, Pere H, Tapper J, Tarkkanen M, Varis A, Wasenius VM, Wolf M, Zhu Y. DNA copy number losses in human neoplasms. Am J Pathol 1999;155:683–94. [32] Fioretos T, Strombeck B, Sandberg T, Johansson B, Billstrom R, Borg A, Nilsson PG, van Den Berghe H, Hagemeijer A, Mitelman F, Hoglund M. Isochromosome 17q in blast crisis of chronic myeloid leukemia and in other hematologic malignancies is the result of clustered breakpoints in 17p11 and is not associated with coding TP53 mutations. Blood 1999;94:225–32. [33] Knuutila S, Bjorkqvist AM, Autio K, Tarkkanen M, Wolf M, Monni O, Szymanska J, Larramendy ML, Tapper J, Pere H, El-Rifai W, Hemmer S, Wasenius VM, Vidgren V, Zhu Y. DNA copy number amplifications in human neoplasms. Am J Pathol 1998;152:1107–23. [34] Santoni Rugiu E, Nagy P, Jensen MR, Factor VM, Thorgeirsson SS. Evolution of neoplastic development in the liver of transgenic mice co-expressing c-myc and transforming growth factor-alpha. Am J Pathol 1996;149:407–28. [35] Abou Elella A, Gramlich T, Fritsch C, Gansler T. c-Myc amplification in hepatocellular carcinoma predicts unfavorable prognosis. Mod Pathol 1996;9:95–8. [36] Schraml P, Kononen J, Bubendorf L, Moch H, Bissig H, Nocito A, Mihatsch MJ, Kallioniemi OP, Sauter G. Tissue microarrays for gene amplification surveys in many different tumor types. Clin Cancer Res 1999;5:1966–75. [37] Valenzuela DN, Griffiths JA, Rojas J, Aldrich TH, Jones PF, Zhou H, McClain J, Copeland NG, Gilbert DJ, Jenkins NA, Huang T, Papadopoulos N, Maisonpierre PC, Davis S, Yancopoulos GD. Angiopoietin 3 and 4: diverging gene counterparts in mice and humans. Proc Natl Acad Sci U S A 1999;96:1904–9. [38] Tannapfel A, Wasner M, Krause K, Geissler F, Katalinic A, Hauss J, Mossner J, Engeland K, Wittekind C. Expression of p73 and its relation to histopathology and prognosis in hepatocellular carcinoma. J Natl Cancer Inst 1999;91:1154–8.