Cancer Research Versailles No Abst Advances in Breast Cancer Research
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Research 67, 4586, May 15, 2007. doi: 10.1158/0008-5472.CAN-06-3464
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Correction (v67,p5998)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nishida, N.
Right arrow Articles by Boland, C. R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Nishida, N.
Right arrow Articles by Boland, C. R.

Molecular Biology, Pathobiology, and Genetics

Extensive Methylation Is Associated with ß-Catenin Mutations in Hepatocellular Carcinoma: Evidence for Two Distinct Pathways of Human Hepatocarcinogenesis

Naoshi Nishida1,2, Takafumi Nishimura2, Takeshi Nagasaka1, Iwao Ikai3, Goel Ajay1 and C. Richard Boland1

1 Department of Internal Medicine, Division of Gastroenterology, Sammons Cancer Center and the Baylor Research Institute, Baylor University Medical Center, Dallas, Texas and Departments of 2 Gastroenterology and Hepatology and 3 Gastroenterological Surgery, Kyoto University Graduate School of Medicine, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto, Japan

Requests for reprints: C. Richard Boland, Gastrointestinal Cancer Research Laboratory (250 Hoblitzelle), Baylor University Medical Center, 3500 Gaston Avenue, Dallas, TX 75246. Phone: 214-820-2692; Fax: 214-818-9292; E-mail: RickBo{at}BaylorHealth.Edu.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Hepatocellular carcinoma (HCC) with p53 mutations is usually characterized by extensive chromosomal instability (CIN), whereas those with ß-catenin mutations have relatively less CIN and the molecular pathogenesis of these tumors is unknown. Methylation of CpG dinucleotides in the promoters of cancer-related genes is another characteristic feature of HCCs. The aim of this study was to determine the contribution of the methylator phenotype to HCC and its relationship to genomic instability. Fractional allelic loss (FAL) was determined using 400 microsatellite markers in 81 HCCs and 77 corresponding noncancerous livers as a measure of CIN. Methylation of 21 genetic loci was quantitated using combined bisulfite restriction analysis. Using hierarchical clustering analysis based upon the quantification of methylation levels, all HCCs were segregated into two groups characterized by either limited or extensive methylation. Mutations in the ß-catenin and p53 genes were determined by DNA sequencing. We found that the methylation levels were significantly higher in the HCCs than in noncancerous livers in 18 of the 21 loci (P values ranged from 0.035 to <0.0001). Among 18 loci, elevated levels of methylation at nine loci were significantly associated with ß-catenin mutations (P values ranged from 0.02 to <0.0001). In addition, the presence of ß-catenin mutations was associated with HCCs in the extensive methylation group (P < 0.0001), whereas p53 mutations correlated with high FAL scores (P = 0.0036). These data suggest that HCCs can be classified into two distinct categories based upon promoter methylation, CIN, and mutations of cancer-related genes. HCCs with extensive methylation harbor frequent ß-catenin mutations, whereas HCCs with high levels of CIN are associated with p53 mutations, suggesting the presence of two independent pathways for the pathogenesis of HCC. [Cancer Res 2007;67(10):4586–94]


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Hepatocellular carcinoma (HCC) is a common malignancy worldwide, and the overall prevalence of the at-risk population is expected to grow with time. HCC is etiologically associated with several distinct risk factors (1), and different forms of genetic and epigenetic alterations have been described in HCC (2). Consequently, we and others have hypothesized that the evolution and progression of this disease may involve more than one molecular pathway of tumorigenesis.

Previous reports suggest that human HCC with p53 mutations tend to have extensive chromosomal instability (CIN), whereas others have mutations in ß-catenin and relatively little CIN (3). These findings are consistent with animals studies in which liver tumors in transgenic mice have been classified based upon CIN and ß-catenin mutations (4).

In addition to genetic alterations, aberrant methylation of CpG dinucleotide sequences in the promoter regions of cancer-related genes is a common feature of HCCs, which may play a role in the evolution of these cancers (5). Thus far, several independent studies have shown abnormal methylation of multiple loci in HCC and have attempted to classify tumors based upon the presence or absence of methylation (6–10). However, the analysis and interpretation of methylation in HCC is more complex than for some other human cancers, limiting attempts to characterize these tumors. One must not only discriminate aberrant methylation detected in cancer cells from the background methylation in aging cells but also distinguish this from the methylation burden contributed by the underlying viral infection and inflammation in these tissues (11, 12). Because both chronic inflammation and viral infections can induce methylation (13), it is likely that the low-level background methylation frequently present in HCC patients may be interpreted inappropriately as relevant to the pathogenesis of the disease when a highly sensitive methylation assay is used. In addition, there is a growing realization that the density of methylation should be taken into consideration to understand biologically meaningful levels of hypermethylation in tumors, a concept that has not been addressed in HCC.

In the present study, we quantified the methylation status of 21 loci, including the promoter regions of 18 cancer-related genes, and three tumor-specific methylated in tumor (MINT) loci (5, 14). The methylation alteration data were further compared with mutations in the ß-catenin and p53 genes and CIN status in these tumors. Our data indicate that HCC can be segregated into two distinct categories, in which one group of tumors is characterized by ß-catenin mutations and extensive methylation of multiple gene promoters, whereas a second group has p53 mutations and is associated with extensive CIN.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Patients. Eighty-one HCC tissues and 77 paired corresponding noncancerous liver tissues from these patients were analyzed for this study. The tumors and their surrounding noncancerous liver tissues were frozen immediately after surgical removal and stored at –80°C until DNA isolation. Sixty HCCs and 58 noncancerous liver tissues were obtained from Kyoto University Hospital (Kyoto, Japan), and 21 HCCs and 19 noncancerous tissues were obtained from Okayama University Hospital (Okayama, Japan). The demographics and related information of these patients are listed in Supplementary Table 1. Written informed consent was obtained from all patients, and necessary approvals were obtained from the institutional review boards of all the involved institutions. Histologically normal liver was obtained from resections done for metastatic cancer done at Baylor University Medical Center.

The CIN status was previously determined in 49 of the 81 HCCs using 400 microsatellite markers equally distributed through the 23 chromosomes (22 autosomes and X; ref. 15). Fractional allelic loss (FAL) scores were calculated, which broadly represent an index of CIN. The mean FAL score and 95% confidence interval (95% CI) for CIN in all HCCs were 26.0% (range, 21.4–30.7).

Quantification of methylation levels using combined bisulfite restriction analysis. Genomic DNA was extracted using QIAamp DNA minikits (Qiagen, Inc.) from frozen tissues. DNA samples (0.5–2.0 µg) were subjected to bisulfite treatment as previously described (16, 17). We selected 21 methylation loci for this study, including 18 cancer-related genes (CDH1, 14-3-3{sigma}, SOCS1, CASP8, RUNX3, HIC-1, GSTP1, p16, RASSF1A, RASSF2, APC, RIZ1, COX2, CACNA1G, DCC, Reprimo, SFRP2, DAPK) and three cancer-specific MINT loci (MINT1, MINT2, MINT31) in which methylation had been reported previously (5, 7, 18–31). Primer sequences, PCR conditions, and restriction enzymes used in combined bisulfite restriction analysis (COBRA) for CDH1, p16, CACNA1G, COX2, DCC, DAPK, 14-3-3{sigma}, MINT1, MINT2, and MINT31 have been reported previously (5, 21, 32, 33). For the remaining gene promoters, the primer sequences were designed in our laboratory and the information and assay conditions are available upon request. Each PCR reaction was done in a total volume of 25 µL, which contained 12.5 µL of HotStarTaq Master Mix (Qiagen), 10 to 40 ng of bisulfite-treated DNA template, and 0.2 µmol/L of each primer pair. After PCR amplification, 5 to 10 µL of the amplified products were subjected to five units of restriction enzyme digestion to determine the degree of methylation at each locus. Digested PCR products were electrophoresed on 2.5% agarose gels and visualized by ethidium bromide staining. Each assay included a positive control DNA sample that was treated with CpG methylase (CpGenome universal methylated DNA; Chemicon International, Inc.), as well as a negative control comprised of normal lymphocyte or fibroblast DNA. Band intensities of digested and/or undigested PCR products were determined using a Kodak Gel-Logic 200 imaging system (Eastman Kodak Co.). The band intensities of restriction enzyme digested PCR products (implying a methylated product) were divided by total band intensities, and the data were represented as percentage density of methylation as described previously (17). The sensitivity of COBRA in our study could detect methylated alleles at densities as low as 2%, and we could not distinguish band intensities lower than 2% from background levels. To account for differences of efficiency of digestion by the restriction enzyme, the percentage methylation density of each tumor DNA sample was normalized to that of CpG methylase-treated DNA, which should result in 100% methylation density. Some of the loci showed relatively small degrees of methylation in the corresponding noncancerous liver tissues. In view of these background methylation levels, we needed to define the cutoff values for each of the markers to analyze tumor-specific hypermethylation. We were extremely stringent with our approach in this regard and did this by using the mean methylation densities for each locus in the noncancerous livers plus 1.96 times the SD (or 95% CI) to define the threshold values for scoring a HCC as "positive" for hypermethylation. In the noncancerous liver group, almost all cases (>95%) had methylation levels below this threshold. In addition to this, we further analyzed these data by defining the methylation index for each tumor using the percentage of loci showing hypermethylation at all 18 loci. Using these approaches in conjunction with hierarchical clustering analysis, we were able to clearly classify all HCCs into two different groups based upon methylation profiles.

Mutation analysis of ß-catenin and the p53 genes. To detect activating mutations of the ß-catenin gene and inactivating mutations of the p53 gene, direct sequencing of exon 3 of ß-catenin (which includes the GSK3ß-targeted phosphorylation sites) and exons 5 through 8 of p53 were done using previously published primers (34–36). Briefly, after PCR amplification, each PCR product was purified using the QIAquick PCR purification kit (Qiagen) and subjected to cycle sequencing using the BigDye Terminator v1.1 cycle sequencing kit (Applied Biosystems). These products were further purified with the Performa DTR gel filtration cartridges (Edge Bio Systems) and sequenced on an ABI Prism 3100-Avant genetic analyzer (Applied Biosystems).

Statistical analyses. The Wilcoxon rank-sum test was used to compare the methylation status of each locus between any two categorical variables. For classification of HCC based upon methylation profiles, hierarchical clustering analysis was applied. To estimate the relative risk for methylation at each locus and the simultaneous presence of ß-catenin or p53 gene mutations, odds ratios and 95% CI were calculated. {chi}2 tests were used to compare the frequencies of mutations or methylation, as well as clinicopathologic factors in the two HCC groups classified using hierarchical clustering analysis. Spearman's rank correlation test was used to examine the relationship between two continuous variables of methylation densities at two different loci. All P values were two-sided and a P < 0.05 was considered statistically significant. All statistical analyses were calculated using JMP version 4.05J software (SAS Institute, Inc.).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Dense methylation is characteristically present in HCC. We quantitated the methylation levels of 81 HCCs and paired 77 corresponding noncancerous liver tissues at 21 different loci using COBRA (Fig. 1A ). Methylated fractions are expressed in terms of percentage methylated (i.e., the proportion of samples in which a digested band was detected) and percentage methylation densities (as described in the Materials and Methods) as shown in Table 1 . Because noncancerous liver tissues are usually accompanied by lymphocytic infiltration which could interfere with the estimates of methylation density, we analyzed the methylation status of five normal lymphocytic populations. We could not find methylation among the five lymphocytic populations for any locus except 14-3-3{sigma}, which showed variable methylation (data not shown; ref. 37). Therefore, for 14-3-3{sigma}, we compared methylation levels between HCC and 22 histologically normal livers.


Figure 1
View larger version (62K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. A, representative images of COBRA for each methylation locus on 10 different specimens at 21 different promoters. The smaller bands beneath the larger ones represent digested (i.e., methylated) alleles. N, noncancerous liver; T, tumor; *, CpG methylase treated DNA for positive control of COBRA. B, schema of distribution of methylation at each locus in each HCC. HCCs were sorted in ascending order based upon the number of hypermethylated loci. Numbers at the top, the total number of loci with hypermethylation in each subset of HCCs; solid triangles, HCC cases with ß-catenin mutations; open triangles, HCCs with p53 mutations; solid square, locus with hypermethylation; open square, locus without hypermethylation.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Comparison of methylation levels between HCCs and corresponding noncancerous liver tissues and frequencies of hypermethylation in HCC at each locus

 
As shown in Table 1, methylation levels for most of the targets were significantly higher in HCC than in noncancerous liver tissues (P < 0.05), with the exceptions of 14-3-3{sigma}, DCC, CDH1 and DAPK. For 14-3-3{sigma}, no significant differences in methylation were observed between HCCs and normal liver tissues. On the other hand, methylation levels at the DCC locus were higher in a subset of HCCs than their corresponding noncancerous liver tissues, although these differences were not statistically significant (P = 0.0617). In addition, we observed only two methylation events at the CDH1 gene promoter and no evidence for methylation at the DAPK gene promoter using COBRA primers published previously (32, 33). Based upon the methylation levels at the remaining 17 loci which were significantly higher in the HCCs than the corresponding noncancerous liver tissues, these 17 markers plus DCC methylation (for a total of 18 markers) were selected for further analysis.

Frequencies of hypermethylation at each locus in HCC tissues. For the detection of tumor-specific hypermethylation at each locus, we defined cutoff values for each methylation marker based upon the background levels of methylation present in noncancerous liver tissues as described earlier in Materials and Methods. HCCs were scored as hypermethylated if their methylation levels in the neoplastic tissues were higher than the individual threshold values mentioned above. The cut off values for methylation levels and the frequencies of hypermethylation at each locus are listed in Table 1. The most frequent hypermethylation events were observed at the APC locus, followed by GSTP1, RIZ1, p16, HIC-1, CACNA1G, and RUNX3, and >50% of HCCs showed hypermethylation at these seven loci.

Frequent concordant methylation among multiple loci in HCC. The methylation states of the 18 target loci in each HCC are illustrated in Fig. 1B. The data have been sorted according to the number of loci demonstrating hypermethylation. A subset of HCCs had hypermethylation at multiple loci, suggesting concordant methylation in these neoplasms. Therefore, we calculated Spearman's rank correlations ({rho}) to better understand the relationships between methylation levels at pairs of loci. A significantly positive correlation was observed for simultaneous methylation of several gene promoters. Segregating the tumors based upon the Spearman's rank correlation coefficient into three categories, we observed that 22 such correlations had {rho} values between 0.53 and 0.40 (P < 0.0001–0.0002), 26 had {rho} values between 0.39 and 0.30 (P = 0.0004–0.0072), and 44 correlations had {rho} values between 0.29 and 0.20 (P = 0.0087–0.0770; Supplementary Figure 1).

Methylation at certain loci correlated with disease progression in HCC. To better understand the association between methylation events and progression of HCC, we compared the methylation levels for each locus with various clinical variables that reflect the progression of disease using two-sided Wilcoxon rank-sum tests. We observed that the methylation levels for HIC-1 and SFRP2 correlated significantly with larger tumors (P = 0.0384 and P = 0.0468, respectively). Similarly, RASSF1A methylation correlated with the presence of vascular invasion (P = 0.0153), RIZ1 methylation with the presence of multiple tumors (P = 0.0498), and increased methylation for RUNX3 associated with moderate-poorly differentiated HCCs (P = 0.05; Supplementary Table 2).

Methylation of specific genes is associated with ß-catenin mutations in HCC. Next, we investigated the relationship between the methylation status of each locus and the presence of ß-catenin and p53 mutations in HCCs. Overall, we observed activating mutations of ß-catenin (at or near the GSK3ß phosphorylation sites) in 15% (13 of 81) and loss-of-function mutations of the p53 genes in 33% (27 of 81) of the HCCs (Supplementary Table 3). Mutations in ß-catenin segregated with HCCs that were hypermethylated at more than seven methylation markers, whereas HCCs harboring p53 mutations were uniformly methylated regardless of methylation status at any specific locus (Fig. 1B). Upon analyzing the associations between the methylation status at each locus and genetic alterations using Wilcoxon rank-sum tests, we observed that nine loci (APC, GSTP1, RIZ1, p16, HIC-1, CACNA1G, RUNX3, SOCS1, and CASP8) showed significantly higher levels of methylation in HCCs with ß-catenin mutations compared with the HCC with wild-type ß-catenin (Table 2 ).


View this table:
[in this window]
[in a new window]

  		Table 2. Comparison of the methylation status at each locus between HCCs with and without ß-catenin and p53 mutations

 
Interestingly, frequencies of hypermethylation at these nine loci were higher compared with the overall methylation levels of the remaining nine loci except for CASP8, suggesting that these might be methylation-specific markers for HCC (Table 1; Fig. 2A ). The odds ratios for methylation of these nine loci and the relative risk for the presence of ß-catenin mutations ranged from 14.5 (methylation of RUNX3; 95% CI, 2.1–118.1) to 960.3 (methylation of HIC-1; 95% CI, 37.9–80,038.1) which were higher than the incidence of methylation at remaining the loci (Fig. 2B). Contrariwise, methylation at only two loci (APC and p16) was significantly correlated with p53 mutations (Table 2). Finally, the odds ratio for methylation of each locus for HCCs with p53 mutations was lower than for HCCs with ß-catenin mutations except for MINT31, COX2, MINT2, and Reprimo (Fig. 2C).


Figure 2
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. A, frequencies (open circles) of hypermethylation and mean percentage methylation densities (open squares) at each locus. Error bars on the squares, 95% CIs. The loci were ordered according to the frequencies of hypermethylation detected in the HCC samples. B, odds ratios of methylation at each locus for the relative risk of detecting simultaneous ß-catenin mutations. Closed squares, odds ratios for methylation that significantly correlated with mutations; open circles, odds ratios in which methylation events did not correlate with mutations. C, odds ratios for methylation and the presence of p53 mutations. Closed squares, odds ratios when methylation correlated with mutations; open circles, odds ratios when such an association was not present.

 
Because of the differences in methylation levels at multiple loci in all 81 HCCs, hierarchical clustering analysis allowed us to broadly categorize the HCCs into two subgroups (groups 1 and 2) based upon methylation profiling (Fig. 3A ). By using percentage methylation densities of all 18 loci for these analyses, 48 HCCs were classified into group 1 and 33 were categorized into group 2. Such clustering analysis further permitted us to evaluate whether specific clinical and/or genetic/epigenetic factors correlated with these two subgroups of HCCs. We observed significant differences in the methylation frequencies and methylation levels of each locus between group 1 and group 2 HCCs. More specifically, the frequency of hypermethylation was significantly higher in group 2 cancers compared with group 1 for all markers except MINT31, Reprimo, and DCC (Supplementary Table 4). Similarly, methylation levels of group 2 tumors were significantly higher than group 1 for all loci except Reprimo, SFRP2, and DCC. Additionally, as a validation of the clustering approach, we observed that the overall methylation index of group 2 cancers was significantly higher compared with group 1 HCCs (P < 0.0001 by Wilcoxon rank-sum tests; Table 3 ). Accordingly, HCCs in group 2 were also considered as the ones having "extensive" methylation, whereas group 1 cancers were considered to have "limited" methylation at the markers analyzed in this study. Interestingly, among the various clinical factors analyzed, only infection with a hepatitis virus (HBV or HCV) showed a significant association with the HCCs categorized as group 2 (P = 0.0009 by {chi}2 test). Furthermore, we noted that all 13 HCCs which had ß-catenin mutations were members of group 2 (P < 0.0001 by {chi}2 test; Table 3; Fig. 3A). On the other hand, 13 of 27 (48%) HCCs with p53 mutations belonged to the group 2.


Figure 3
View larger version (32K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. A, HCC groups classified by hierarchical clustering analysis and ß-catenin or p53 mutations. HCCs were classified into two subgroups based upon the methylation signature. The color map (green-black-red): green, degree of decreased methylation density; red, degree of increased density. The loci examined (Y axis) are ordered as in Fig. 2. Solid triangle, tumor with a ß-catenin mutation; open triangle, HCC with a p53 mutation. As depicted, ß-catenin mutations were significantly associated with group 2 (P < 0.0001). B and C, relationships between the methylation index (B) and FAL scores (C) and the presence or absence of p53 and ß-catenin mutations. B, the mean methylation index (95% CI) for all HCCs was 43.3% (48.1–38.5; hatched horizontal line through each of the two boxes). Mean methylation indices (95% CI) for each group: 40.4% (35.1–45.7) in the group without ß-catenin, 59.0% (51.4–66.5) in group with ß-catenin, 41.6% (35.5–47.7) in the group without p53 mutations, and 46.9% (38.9–54.9) in the group with p53 mutations (line in the diamond). Associations were statistically significant for ß-catenin (P = 0.0017) but were not significant for p53 mutations (P = 0.2495). P values were calculated by two-sided Wilcoxon rank-sum tests. C, FAL score (95% CI) for all of the HCCs was 26.0% (30.7–21.4). Mean FALs (95% CI) for each group: 24.6% (19.4–29.8) in the group without ß-catenin mutations, 32.3% (20.7–44.0) in the group with ß-catenin mutations, 21.1% (16.2–26.2) in the group without p53 mutations, and 35.9% (27.3–44.5) in group with p53 mutations. Associations were statistically significant for p53 mutations (P = 0.0036) but were not significant for ß-catenin mutations (P = 0.1243).

 

View this table:
[in this window]
[in a new window]

  		Table 3. Associations between subgroups of HCC classified by hierarchical clustering analysis and genetic alterations (key mutations, methylation index and FAL)

 
Higher FAL scores associated with p53 mutations, whereas high methylation indices correlated with ß-catenin mutations in HCC. Previously, we had assayed for allelic imbalances at 400 microsatellite loci distributed evenly throughout the chromosomes in 49 of the 81 HCCs (15). In the present study, we compared these FAL scores with the methylation status of the HCCs after hierarchical clustering analysis. We did not observe any significant differences in overall FAL scores between groups 1 and 2 HCCs (P = 0.3816 by {chi}2 test; Table 3). However, FAL scores were significantly higher in HCCs with p53 mutations than in those with wild-type sequences (Fig. 3C; P = 0.0036 by Wilcoxon rank-sum test). Contrariwise, no significant differences were found in the FAL scores between HCCs with and without ß-catenin mutations (P = 0.1243). Additionally, higher methylation indices significantly associated with ß-catenin mutations (Fig. 3B; P = 0.0017) but not with p53 mutations (P = 0.2495).


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
For characterization of HCCs according to the methylation status, quantitative methylation profiling is necessary, as methylation densities are not uniform in all tumors and these have biological implications for gene silencing. In the present study, we have analyzed a large subset of HCCs using quantitative methylation analysis for several gene promoters and provide evidence that a significant majority of HCC is characterized by an extensive methylation phenotype. Additionally, we show that the tumors characterized by extensive methylation frequently harbor ß-catenin mutations and are distinct from neoplasms with CIN, which are associated with p53 mutations.

It is known that CIN and epigenetic instability are two major mechanisms for generating genetic diversity in many human cancers (38, 39). However, this concept has not been definitively investigated in HCC. In primary liver cancers, the two major pathways of tumorigenesis have been shown to be mutations in the ß-catenin gene (leading to abrogation of Wnt signaling) and p53 mutations (which relate to extensive CIN; ref. 3). Although there have been suggestions that aberrant methylation of gene promoters may constitute an important epigenetic mechanism of genomic instability in HCC (2), a detailed investigation into aberrant methylation, epigenetic instability, and mutations of the ß-catenin and p53 genes, which are frequently found in human HCC, has not been done previously.

In the present study, we initially assayed 21 promoter loci for methylation and later selected 18 of these in which dense methylation was a distinctive feature in HCCs. Among these 18 markers, several were concordantly methylated in a subgroup of HCCs. This suggested the presence of a methylator phenotype in those tumors with consequent transcriptional silencing of specific cancer-related genes.

It has long been appreciated that HCC typically emerges from a background of chronic hepatic inflammation (1), and it has been suggested that this may be an important trigger for methylation (11). We were not surprised to find a significant association between hepatitis virus infection and increased methylation, which adds mechanistically to our concept of how chronic viral infections may lead to HCC (5, 10). To take this a step further, we compared the methylation status of various genes with clinical factors related to tumor progression. However, no clear relationship was observed for most methylation loci and tumor progression, suggesting that methylation is probably an early event in the evolution of HCC. We would speculate that CIN may come later in the course of HCC and that an accumulation of allelic imbalances may favor tumor progression and metastasis (6, 39, 40).

Our comprehensive analyses of genetic and epigenetic alterations in the HCC cohort revealed highly significant associations between ß-catenin mutations and high levels of methylation at 9 of the 18 target loci. Interestingly, methylation densities and frequencies at these nine loci were especially high in HCC. On the other hand, methylation at the remaining loci was not as prominent, which contrasts with methylation frequencies reported for these loci in colon and gastric cancers, thus distinguishing the pathogenesis of HCC from other gastrointestinal tumors (22, 27, 30, 31). These data support the concept that a combination of extensive methylation and ß-catenin mutations could be a unique characteristic in HCC, either reflecting a common provenance, or alternatively, reflecting synergy in the evolution of these tumors. Methylation profiles of HCC determined by hierarchical clustering analysis of the multiple markers revealed that methylation levels, and not merely methylation frequencies, should be considered when analyzing a putative methylator phenotype in cancer. To date, only one prior study of HCC has used quantitative methylation profiling, and in this instance, discrimination between methylation of liver adenoma and HCC was reported (28). Although this work confirms an association between p53 mutations and higher FAL scores (which correlates with aneuploidy; ref. 3), the association between ß-catenin mutations and extensive methylation is novel. Taken together, these data suggest that HCCs characterized by extensive methylation may develop through a distinct mutational pathway. Some proportion of HCCs showed extensive methylation as well as high FAL scores. And because the majority of tumors investigated in the present study were in advanced stages, it is conceivable that there may be some crossover or convergence between these two pathways in the later stages of tumor development (10).

In conclusion, our data add to our growing appreciation that cancers may evolve through multiple pathways of genomic instability, that epigenetic and genetic alterations may collaborate in the evolution of human HCCs, and that the mutational signatures in the DNA from these tumors may be useful to characterize these tumors. Perhaps more importantly, because epigenetic alterations are theoretically reversible, our findings suggest the possibility of identifying specific methylation markers in HCC which have prognostic implications in the management of early cancers or premalignant liver disease.


   Acknowledgments
 
Grant support: Grants RO1 CA72851 and RO1 CA98572 from the National Cancer Institute, NIH and funds from Baylor Research Institute.

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.


   Footnotes
 
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Received 9/18/06. Revised 12/18/06. Accepted 3/ 2/07.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Bruix J, Boix L, Sala M, Llovet JM. Focus on hepatocellular carcinoma. Cancer Cell 2004;5:215–9.[CrossRef][Medline]
  2. Thorgeirsson SS, Grisham JW. Molecular pathogenesis of human hepatocellular carcinoma. Nat Genet 2002;31:339–46.[CrossRef][Medline]
  3. Laurent-Puig P, Legoix P, Bluteau O, et al. Genetic alterations associated with hepatocellular carcinomas define distinct pathways of hepatocarcinogenesis. Gastroenterology 2001;120:1763–73.[CrossRef][Medline]
  4. Calvisi DF, Factor VM, Ladu S, Conner EA, Thorgeirsson SS. Disruption of ß-catenin pathway or genomic instability define two distinct categories of liver cancer in transgenic mice. Gastroenterology 2004;126:1374–86.[CrossRef][Medline]
  5. Shen L, Ahuja N, Shen Y, et al. DNA methylation and environmental exposures in human hepatocellular carcinoma. J Natl Cancer Inst 2002;94:755–61.[Abstract/Free Full Text]
  6. Kondo Y, Kanai Y, Sakamoto M, Mizokami M, Ueda R, Hirohashi S. Genetic instability and aberrant DNA methylation in chronic hepatitis and cirrhosis-A comprehensive study of loss of heterozygosity and microsatellite instability at 39 loci and DNA hypermethylation on 8 CpG islands in microdissected specimens from patients with hepatocellular carcinoma. Hepatology 2000;32:970–9.[CrossRef]
  7. Yu J, Ni M, Xu J, et al. Methylation profiling of twenty promoter-CpG islands of genes which may contribute to hepatocellular carcinogenesis. BMC Cancer 2002;2:29.[CrossRef][Medline]
  8. Lee S, Lee HJ, Kim JH, Lee HS, Jang JJ, Kang GH. Aberrant CpG island hypermethylation along multistep hepatocarcinogenesis. Am J Pathol 2003;163:1371–8.[Abstract/Free Full Text]
  9. Yang B, Guo M, Herman JG, Clark DP. Aberrant promoter methylation profiles of tumor suppressor genes in hepatocellular carcinoma. Am J Pathol 2003;163:1101–7.[Abstract/Free Full Text]
  10. Katoh H, Shibata T, Kokubu A, et al. Epigenetic instability and chromosomal instability in hepatocellular carcinoma. Am J Pathol 2006;168:1375–84.[Abstract/Free Full Text]
  11. Issa JP. CpG island methylator phenotype in cancer. Nat Rev Cancer 2004;4:988–93.[CrossRef][Medline]
  12. Ushijima T, Okochi-Takada E. Aberrant methylations in cancer cells: where do they come from? Cancer Sci 2005;96:206–11.[CrossRef][Medline]
  13. Issa JP. Epigenetic variation and human disease. J Nutr 2002;132:2388–92S.
  14. Toyota M, Ho C, Ahuja N, et al. Identification of differentially methylated sequences in colorectal cancer by methylated CpG island amplification. Cancer Res 1999;59:2307–12.[Abstract/Free Full Text]
  15. Nishimura T, Nishida N, Komeda T, et al. Genome-wide semiquantitative microsatellite analysis of human hepatocellular carcinoma: discrete mapping of smallest region of overlap of recurrent chromosomal gains and losses. Cancer Genet Cytogenet 2006;167:57–65.[CrossRef][Medline]
  16. Frommer M, McDonald LE, Millar DS, et al. A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc Natl Acad Sci U S A 1992;89:1827–31.[Abstract/Free Full Text]
  17. Xiong Z, Laird PW. COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res 1997;25:2532–4.[Abstract/Free Full Text]
  18. Wales MM, Biel MA, el Deiry W, et al. p53 activates expression of HIC-1, a new candidate tumour suppressor gene on 17p13.3. Nat Med 1995;1:570–7.[CrossRef][Medline]
  19. Matsuda Y, Ichida T, Matsuzawa J, Sugimura K, Asakura H. p16(INK4) is inactivated by extensive CpG methylation in human hepatocellular carcinoma. Gastroenterology 1999;116:394–400.[CrossRef][Medline]
  20. Tsuchiya T, Tamura G, Sato K, et al. Distinct methylation patterns of two APC gene promoters in normal and cancerous gastric epithelia. Oncogene 2000;19:3642–6.[CrossRef][Medline]
  21. Iwata N, Yamamoto H, Sasaki S, et al. Frequent hypermethylation of CpG islands and loss of expression of the 14–3-3 {sigma} gene in human hepatocellular carcinoma. Oncogene 2000;19:5298–302.[CrossRef][Medline]
  22. Toyota M, Shen L, Ohe-Toyota M, Hamilton SR, Sinicrope FA, Issa JP. Aberrant methylation of the Cyclooxygenase 2 CpG island in colorectal tumors. Cancer Res 2000;60:4044–8.[Abstract/Free Full Text]
  23. Yoshikawa H, Matsubara K, Qian GS, et al. SOCS-1, a negative regulator of the JAK/STAT pathway, is silenced by methylation in human hepatocellular carcinoma and shows growth-suppression activity. Nat Genet 2001;28:29–35.[CrossRef][Medline]
  24. Zhong S, Tang MW, Yeo W, Liu C, Lo YM, Johnson PJ. Silencing of GSTP1 gene by CpG island DNA hypermethylation in HBV-associated hepatocellular carcinomas. Clin Cancer Res 2002;8:1087–92.[Abstract/Free Full Text]
  25. Schagdarsurengin U, Wilkens L, Steinemann D, et al. Frequent epigenetic inactivation of the RASSF1A gene in hepatocellular carcinoma. Oncogene 2003;22:1866–71.[CrossRef][Medline]
  26. Xiao WH, Liu WW. Hemizygous deletion and hypermethylation of RUNX3 gene in hepatocellular carcinoma. World J Gastroenterol 2004;10:376–80.[Medline]
  27. Suzuki H, Watkins DN, Jair KW, et al. Epigenetic inactivation of SFRP genes allows constitutive WNT signaling in colorectal cancer. Nat Genet 2004;36:417–22.[CrossRef][Medline]
  28. Lehmann U, Berg-Ribbe I, Wingen LU, et al. Distinct methylation patterns of benign and malignant liver tumors revealed by quantitative methylation profiling. Clin Cancer Res 2005;11:3654–60.[Abstract/Free Full Text]
  29. Akino K, Toyota M, Suzuki H, et al. The Ras effector RASSF2 is a novel tumor-suppressor gene in human colorectal cancer. Gastroenterology 2005;129:156–69.[CrossRef]
  30. Hesson LB, Wilson R, Morton D, et al. CpG island promoter hypermethylation of a novel Ras-effector gene RASSF2A is an early event in colon carcinogenesis and correlates inversely with K-ras mutations. Oncogene 2005;24:3987–94.[CrossRef][Medline]
  31. Takahashi T, Suzuki M, Shigematsu H, et al. Aberrant methylation of Reprimo in human malignancies. Int J Cancer 2005;115:503–10.[CrossRef][Medline]
  32. Suzuki H, Itoh F, Toyota M, et al. Distinct methylation pattern and microsatellite instability in sporadic gastric cancer. Int J Cancer 1999;83:309–13.[CrossRef][Medline]
  33. Ogi K, Toyota M, Ohe-Toyota M, et al. Aberrant methylation of multiple genes and clinicopathological features in oral squamous cell carcinoma. Clin Cancer Res 2002;8:3164–71.[Abstract/Free Full Text]
  34. Toguchida J, Yamaguchi T, Ritchie B, et al. Mutation spectrum of the p53 gene in bone and soft tissue sarcomas. Cancer Res 1992;52:6194–9.[Abstract/Free Full Text]
  35. Huang H, Fujii H, Sankila A, et al. ß-catenin mutations are frequent in human hepatocellular carcinomas associated with hepatitis C virus infection. Am J Pathol 1999;155:1795–801.[Abstract/Free Full Text]
  36. Nishida N, Fukuda Y, Kokuryu H, et al. Role and mutational heterogeneity of the p53 gene in hepatocellular carcinoma. Cancer Res 1993;53:368–72.[Abstract/Free Full Text]
  37. Bhatia K, Siraj AK, Hussain A, Bu R, Gutierrez MI. The tumor suppressor gene 14–3-3 {sigma} is commonly methylated in normal and malignant lymphoid cells. Cancer Epidemiol Biomarkers Prev 2003;12:165–9.[Abstract/Free Full Text]
  38. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature 1998;396:643–9.[CrossRef][Medline]
  39. Baylin SB, Ohm JE. Epigenetic gene silencing in cancer - a mechanism for early oncogenic pathway addiction? Nat Rev Cancer 2006;6:107–16.[CrossRef][Medline]
  40. Nishida N, Fukuda Y, Kokuryu H, et al. Accumulation of allelic loss on arms of chromosomes 13q, 16q and 17p in the advanced stages of human hepatocellular carcinoma. Int J Cancer 1992;51:862–8.[Medline]



This article has been cited by other articles:


Home page
 Clin. Cancer Res.Home page
H. S. Lee, B.-H. Kim, N.-Y. Cho, E. J. Yoo, M. Choi, S.-H. Shin, J.-J. Jang, K.-S. Suh, Y. S. Kim, and G. H. Kang
Prognostic Implications of and Relationship Between CpG Island Hypermethylation and Repetitive DNA Hypomethylation in Hepatocellular Carcinoma
Clin. Cancer Res., February 1, 2009; 15(3): 812 - 820.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
Correction: Genetic and Epigenetic Pathways in HCC
Cancer Res., June 15, 2007; 67(12): 5998 - 5998.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Correction (v67,p5998)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nishida, N.
Right arrow Articles by Boland, C. R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Nishida, N.
Right arrow Articles by Boland, C. R.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online