Fax: (011) 886 287917654
Promoter methylation of the secreted frizzled-related protein 1 gene SFRP1 is frequent in hepatocellular carcinoma
Article first published online: 22 JUN 2006
Copyright © 2006 American Cancer Society
Volume 107, Issue 3, pages 579–590, 1 August 2006
How to Cite
Shih, Y.-L., Shyu, R.-Y., Hsieh, C.-B., Lai, H.-C., Liu, K.-Y., Chu, T.-Y. and Lin, Y.-W. (2006), Promoter methylation of the secreted frizzled-related protein 1 gene SFRP1 is frequent in hepatocellular carcinoma. Cancer, 107: 579–590. doi: 10.1002/cncr.22023
- Issue published online: 18 JUL 2006
- Article first published online: 22 JUN 2006
- Manuscript Accepted: 14 MAR 2006
- Manuscript Revised: 10 MAR 2006
- Manuscript Received: 2 SEP 2005
- National Science Council, Republic of China. Grant Numbers: NSC 92-3112-B-016-006, NSC 93-3112-B-016-002, NSC 93-2311-B-016-002
- Tri-Service General Hospital, Taiwan, Republic of China. Grant Numbers: TSGH-C95-7-S01, TSGH-C95-7-S02, TSGH-C95-7-S03, TSGH-C95-7-S04
- C. Y. Chai Foundation for Advancement of Education, Sciences, and Medicine
- hepatocellular carcinoma;
- secreted frizzled-related protein 1 gene;
- promoter hypermethylation;
- loss of heterozygosity
The secreted frizzled-related protein 1 gene (SFRP1) encodes a Wnt/β-catenin signaling antagonist and frequently is inactivated by promoter methylation in many tumors. However, the role of SFRP1 in hepatocellular carcinoma (HCC) is not clear. Therefore, the authors investigated whether methylation of the SFRP1 promoter is common in HCC and whether it may influence SFRP1 expression.
Four HCC cell lines, 54 HCCs, 42 cirrhotic livers, 21 livers with chronic hepatitis, and 15 normal control tissues were analyzed for 1) SFRP1 promoter methylation by using methylation-specific polymerase chain reaction analysis and bisulfite sequencing, 2) SFRP1 messenger RNA expression by using quantitative reverse transcriptase-polymerase chain reaction analysis, and 3) loss of heterozygosity (LOH) by using microsatellite markers flanking the SFRP1 locus. HCC cells were treated with the demethylating agent 5-aza-2′-deoxycytidine to determine whether it could restore SFRP1 expression.
SFRP1 promoter methylation was observed in 75%, 48.2%, 21.4%, 14.3% and 0% in HCC cell lines, primary HCCs, cirrhotic livers, livers with chronic hepatitis, and normal control tissues, respectively. Methylation of the SFRP1 promoter region in HCCs increased significantly compared with control tissues. All samples with SFRP1 methylation showed down-regulation of SFRP1 expression. Demethylation treatment with 5-aza-2′-deoxycytidine in HCC cells restored SFRP1 expression. Moreover, LOH of markers D8S505 and D8S1722 was found in 25% and 27.6% of the informative samples, respectively.
The current results suggested that promoter hypermethylation of SFRP1 is a common event in HCC and plays an important role in the regulation of SFRP1 expression. In addition to methylation-mediated down-regulation of SFRP1, LOH also may play a role. Cancer 2006. © 2006 American Cancer Society.
Hepatocellular carcinoma (HCC) is the most frequent primary malignancy of the liver and accounts for as many as 1 million deaths annually worldwide.1–5 The major risk factors include chronic hepatitis B virus (HBV) infection; chronic hepatitis C virus (HCV) infection; environmental carcinogens, such as aflatoxin B1 (AFB1); alcoholic cirrhosis; and inherited genetic disorders, such as hemochromatosis, Wilson disease, and tyrosinemia. Among these, HBV, HCV, and AFB1 are responsible for approximately 80% of all HCCs.4, 5 Research on the molecular genetics and pathogenesis of HCC has become a hot spot in cancer study because of its scientific merits and its clinical importance. Despite the rapid expansion of information obtained from this research, the molecular mechanism of hepatocarcinogenesis and molecular genetics of HCC remains elusive.
During the last decade, extensive studies have applied many strategies, including restriction fragment-length polymorphism, comparative genomic hybridization analysis, and genome-wide microsatellite analysis, for the screening of DNA copy number losses or loss of heterozygosity (LOH) in HCC. Frequent DNA copy number losses or LOH on multiple chromosomal regions, including 1p, 4p, 5q, 6q, 8p, 9p, 13q, 16p, 16q, and 17p, were reported in HCC.6–14 These findings suggest that a variety of tumor suppressor genes (TSGs) in these chromosomal arms may be inactivated during carcinogenesis. In addition to HCC, frequent LOH on 8p has been reported in colorectal cancer and lung cancer.15–18 The localization of SFRP1 to 8p11.2 and its function as an antagonist of Wnt signaling led us to investigate this gene as a candidate TSG.
Wnt glycoproteins comprise a family of extracellular signaling ligands that play essential roles in proliferation, patterning, and fate determination during normal developmental processes.19–25 Although this signaling is critical for normal embryonic development, the aberrant activation of the Wnt signal-transduction pathway has been linked closely to tumorigenesis in adults.26–28 When Wnt ligands are present, the ligands bind to the transmembrane receptors Frizzled (Fz), and the signal is transduced to cytoplasmic protein Dishevelled (Dvl) by phosphorylation. Through a series of molecules, finally, β-catenin accumulates and enters the nucleus, where it forms a complex with members of the T-cell factor/lymphocyte enhanced factor (TCF/LEF) and up-regulates TCF/LEF-dependent transcription of target genes, such as c-myc and cyclin D1.26, 27
Aberrant hypermethylation of CpG islands, which are CpG dinucleotide-rich areas located mainly in the promoter regions of many genes, serves as an alternative mechanism for the inactivation of TSGs in cancer.29–34 Hypermethylation of gene promoters has been implicated increasingly as an early event in hepatocellular carcinogenesis.35–38 Secreted frizzled-related proteins (SFRPs), a family of 5 secreted glycoproteins, are extracellular signaling molecules that antagonize the Wnt signaling pathway.39SFRP1 is inactivated by promoter methylation and, infrequently, by mutation in colorectal cancer,40 papillary bladder cancer,41 mesothelioma,42 and ovarian cancer.43 We hypothesized that CpG island methylation of the SFRP1 promoter may play an important role in regulating SFRP1 expression in HCC. To test this hypothesis, we used methylation-specific polymerase chain reaction (MS-PCR) analysis and a bisulfite sequencing method to analyze the SFRP1 methylation pattern in HCCs. Messenger RNA (mRNA) expression was assessed by quantitative reverse transcriptase (RT)-PCR assay. Subsequently, we determined whether the treatment of HCC cell lines with a demethylating agent, 5-aza-2′-deoxycytidine (5-Aza-CdR), could restore or increase expression of the SFRP1 gene.
MATERIALS AND METHODS
Tissue samples were obtained from surgical specimens with the informed consent of patients at the Tri-Service General Hospital. Fifty-four primary HCC samples and adjacent nontumorous liver tissues were collected during surgery and were frozen immediately in liquid nitrogen and stored at −70°C until DNA/RNA extraction. The diagnosis of HCC was confirmed by histology. Experienced pathologists classified the nontumorous liver tissues as normal controls (3 samples), livers with chronic hepatitis (11 samples), and cirrhotic livers (40 samples). The clinicopathologic characteristics of patients and tumors are summarized in Table 1. In addition, liver biopsies were collected from patients without HCC as normal controls. Two patients with cirrhosis were HBV-positive, 10 patients were chronic hepatitis B carriers, and 12 patients had normal livers.
|Patient no.||Age, Years||Gender||Antigen status||Size, cm||Multifocal||TNM stage||αFP, ng/mL||Vascular invasion||Recurrent||Cirrhosis||Methylation status||Loss of heterozygosity|
We obtained 4 human HCC cell lines from the American TypeCulture Collection (Rockville, MD): HepG2, Hep3B, Huh7, and SK-Hep1, which were grown in Dulbecco modified Eagle medium supplemented with 10% (weight/volume) fetal bovine serum, penicillin 100 U/mL, streptomycin 100 μg/mL, and l-glutamine 2 mmol/L (all from Invitrogen, Carlsbad, CA) at 37°C in an atmosphere of 5% (volume/volume) carbon dioxide in air.
HepG2, Hep3B, Huh7, and SK-Hep1 cells were seeded at a density of 1 × 105 cells per 100-mm dish and were allowed to attach for 24 hours. Cells were incubated in 5 μM or 10 μM 5-Aza-CdR (Sigma Chemical Company, St. Louis, MO) diluted in phosphate-buffered saline or in phosphate-buffered saline alone for 96 hours to analyze the effect of methylation inhibition on SFRP1 mRNA expression. All incubations were performed in duplicate dishes, and cells were harvested directly for RNA and DNA isolation.
Genomic DNA was extracted from cell lines and tissue samples by using a commercial DNA-extraction kit (QIAmp Tissue Kit; Qiagen, Hilden, Germany). DNA was isolated according to the manufacturer's protocol.
Bisulfite Modification and MS-PCR
Genomic DNA isolated from cells and tissue was subjected to bisulfite methylation analysis. We treated DNA with bisulfite by using an EZ DNA methylation kit (Zymo Research, Orange, CA) according to the protocol described in the user manual. Briefly, 1 μg of genomic DNA was denatured by incubation with 0.2 M NaOH. Aliquots of 10 mM hydroquinone and 3 M sodium bisulfite (pH 5.0) were added and the solution was incubated at 50°C for 16 hours. Treated DNA was purified on a Zymo-Spin I column, desulfonated with 0.3 M NaOH, repurified on a Zymo-Spin I column, and resuspended in 20 μL elution buffer. MS-PCR44 was performed in a volume of 25 μL that contained 1 μL of the sodium-bisulfite-treated DNA with Gold Taq DNA polymerase (PE Applied Biosystems, Foster City, CA) as follows: The MS-PCR primer sequences40 were 5′-GAGTTAGTGTTGTGTGTTTGTTGTTTTGT-3′ (forward) and 5′-CCCAACATTACCCAACTCCACAACCA-3′ (reverse) for unmethylated DNA and 5′-GTGTCGCGCGTTCGTCGTTTCGC-3′ (forward) and 5′-AACGTTACCCGACTC CGCGACCG-3′ (reverse) for methylated DNA. After heating at 92°C for 10 minutes, PCR was performed in a thermal cycler (GeneAmp 2400; PE Applied Biosystems) for 35 cycles, each of which consisted of denaturation at 92°C for 30 seconds, annealing at 61°C for 30 seconds, and extension at 72°C for 30 seconds, followed by a final 10-minute extension at 72°C. The PCR products were analyzed by electrophoresis on a 3% agarose gel. The experiments were repeated 3 times to ensure reproducibility.
Bisulfite-treated genomic DNA was amplified by using primers that have been described previously for human SFRP1.45 Amplified PCR product was purified and cloned into pCR4-TOPO vector (Invitrogen). DNA sequencing was performed on at least 5 individual clones using the 377 automatic sequencer (Applied Biosystems).
We isolated total RNA from each samples by using the Qiagen RNeasy kit (Qiagen, Valencia, CA). An additional DNase I digestion procedure was included in the isolation of RNA to remove contaminating DNA according to the manufacturer's protocol. One microgram of total RNA from each sample was subjected to combinational DNA (cDNA) synthesis by using Superscript II reverse transcriptase and random hexamer (Invitrogen). Then, cDNA was amplified by PCR with primers specific for SFRP1: 5′-TCCCTGTGACAACGAG-3′ (forward) and 5′-GTCCTTCTTCTTGATGGGC-3′ (reverse). RT-PCR for SFRP1 and the glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH) were performed with the PCR Master Mix reagents kit (Applied Biosystems). After heating at 92°C for 10 minutes, PCR was performed in a thermal cycler (GeneAmp 2400; PE Applied Biosystems) for 40 cycles, each of which consisted of denaturation at 92°C for 30 seconds, annealing at 61°C for 30 seconds, and extension at 72°C for 30 seconds, followed by a final 10-minute extension at 72°C. The PCR products were analyzed by electrophoresis on a 2% agarose gel.
Quantitative RT-PCR analysis was performed on an ABI PRISM 7700 Sequence Detector (Applied Biosystems). The match primers and TagMan Probe were obtained from commercial Applied Biosystems Tagman® Assay-on Demand™ Gene Expression products. Hypoxanthine ribosyltransferase (HPRT) was used as an internal control. PCR reaction was carried out using the TaqMan PCR Master Mix reagents kit. Relative gene expression was determined based on the threshold cycles (Ct) of the gene of interest and of the internal reference gene. The mRNA levels of the interest genes were expressed as the ratio of the interest gene to HPRT mRNA for each sample. The level of each interest gene mRNA in each cancer was compared with the level in corresponding nontumorous liver tissues.46 The average Ct value of the HPRT gene was subtracted from the average Ct value of the interest genes for each sample: SFRP1ΔCt =(Avg. SFRP1Ct − Avg. HPRT Ct), SFRP1ΔΔCt = (Avg. SFRP1ΔCttumor − Avg. SFRP1ΔCtnontumor). The fold change (2-SFRP1ΔΔCt) in expression of the target genes (SFRP1) relative to the internal control gene (HPRT) of each analyzed HCC sample was calculated.
The HCC tumors were examined for LOH by using 2 microsatellite markers (D8S505 (forward, 5-AGCCTG CTATTTGTAGATAATGTTT; reverse, 5-AGTGCTAAG TCCCAGACCA) and D8S1722 (forward, 5-CCTTGCTGGGAATTGTG; reverse, 5-AGCTGCCTGGCTAAGAG) that are found on chromosome 8p11.2 flanking the SFRP1 locus. Amplification was done in 25-μL volumes with 50 ng of genomic DNA, 10 pmol of each primer (5′ fluorescent-labeled primer), 100 μM each dinucleotide triphosphate, and 1.25 U Taq polymerase under the following conditions: 10 minutes at 94°C (hot start), followed by 30 cycles at 94°C for 30 seconds, at 55°C for 30 seconds, and at 72°C for 30 seconds, with a final 7-minute extension at 72°C. PCR products were analyzed using a Beckman CEQ™ 8000 Genetic Analysis System. LOH was analyzed by determining the fluorescent intensity of each allele and calculating the ratio. LOH was defined when an allele peak signal from tumor DNA was <0.5 compared with the signal from nontumorous liver tissues.
The software program SPSS for Windows (version 13; SPSS Inc., Chicago, IL) was used for statistical analyses. The associations between methylation of SFRP1 and clinical parameters were analyzed by using chi-square tests (Tables 2 and 3) and Fisher exact tests (Tables 3 and 4) (2 × 2 table when n < 5), as necessary. We correlated the SFRP1 methylation status with liver disease status (control, chronic hepatitis, cirrhosis, and HCC), down-regulation of SFRP1 mRNA expression, and several clinical variables, including gender, age, tumor size, tumor marker, clinical tumor-lymph node-metastasis (TNM) staging classification, vascular invasion, recurrence, and type of virus infection. The significance level was defined as a P value <0.05.
|Diagnosis||No. of samples||No. of methylated SFRP1 samples||No. of unmethylated SFRP1 samples||P*|
|Characteristic||SFRP1 hypermethylation: No. of patients (%)||P*|
|Absent (n = 29)||Present (n = 26)|
|Tumor size, cm||.17|
|No. of tumor nodules||.57|
|Stage I and II||15||15 (50.0)|
|Stage III and IV||13||11 (45.8)|
|AFP level, ng/mL||.79|
|ALT level, IU/L||.58|
|Total bilirubin level, mg/dL||.76|
|Down-regulation of SFRP1 ≥2-fold||No. of samples||P*|
|Methylation of CpG island (no. of cases)||No methylation of CpG island (no. of cases)|
SFRP1 Promoter Methylation and Silencing in HCC Cell Lines
To investigate epigenetic silencing of SFRP1 in HCC, first, we tested for promoter methylation in 4 hepatoma cell lines (HepG2, Hep3B, Huh7, and SK-Hep1) by using MS-PCR and bisulfite sequencing. Among 4 HCC cell lines, the HepG2 and Hep3B cell lines demonstrated SFRP1 hypermethylation, Huh7 cells were partially methylated, and SK-Hep1 cells were unmethylated (Fig. 1A). Bisulfite sequencing results from the 2 cell lines that were positive for methylation in the MS-PCR analysis are summarized in Figures 1B and 1C. The regions from −283 to −261 and from −133 to −108 were the primer regions used in MS-PCR. The CpGs in these regions frequently were methylated, and their frequencies ranged from 56% to 97% (predominantly from 65% to 80%) (Fig. 1B). In contrast, we could not detect promoter hypermethylation in SK-Hep1 cells. RT-PCR analysis of these 2 hepatoma cell lines with SFRP1 hypermethylation (HepG2 and Hep3B) demonstrated complete transcriptional silencing that was expressed weakly in the partially methylated Huh7 cell line and strongly in the unmethylated SK-Hep1 cell line (Fig. 1D). To confirm that the lack of expression of SFRP1 in the hepatoma cell lines was caused by promoter hypermethylation, we assessed various concentrations of 5-Aza-CdR as inhibitors of DNA methylation. After treatment with 5 μM of 5-Aza-CdR, the unmethylated promoter DNA was detected by MS-PCR, and SFRP1 transcripts were reexpressed in the HepG2 and Hep3B cell lines (Figs. 2A and B). These results indicated that hypermethylation of SFRP1 may be responsible for the absence of transcription.
Methylation of the SFRP1 Promoter in Primary HCCs
We investigated the frequency of SFRP1 promoter methylation in 54 primary HCCs using MS-PCR (Table 1) (Fig. 3A). Aberrant promoter methylation of SFRP1 was found in 26 of 54 HCCs (48.2%), in 9 of 42 cirrhotic livers (21.4%), in 3 of 21 livers with chronic hepatitis (14.3%), and in 0 of 8 normal control tissues (0%) (Fig. 3B). Bisulfite sequencing results in 2 primary HCCs and 2 corresponding nontumorous liver tissues are summarized in Figure 3C. The CpGs in those regions were methylated with a frequency that ranged from 2% to 59% (Fig. 3C). Because of the heterogeneity of primary HCCs, promoter methylation of SFRP1 was not found in some clones. In contrast, we could not detect promoter hypermethylation in nontumorous controls. These data demonstrated that hypermethylation of the SFRP1 promoter region is a frequent event in HCC and is increased significantly compared with cirrhotic livers, livers with chronic hepatitis, and normal control tissues (Table 2) (P = 0.0001).
Frequent Down-Regulation of SFRP1 mRNA in Primary HCCs
To determine the relation between SFRP1 promoter methylation status and SFRP1 mRNA expression, we used quantitative RT-PCR to analyze SFRP1 gene expression in 54 primary HCCs and their corresponding nontumorous liver tissues. Seven primary HCCs were excluded from this assay because of poor RNA quality. The down-regulation of SFRP1 was observed in 43 of 47 HCCs (91.5%) compared with nontumorous liver tissues (Table 5). In 39 of 43 HCCs (90.6%) SFRP1 was down-regulated significantly (by >2-fold) compared with the corresponding nontumorous liver tissues. The 6 tumors (from Patients 4, 5, 10, 14, 24, and 37) that had both methylation-specific and unmethylation-specific PCR products in nontumorous liver tissues were excluded from the correlation analysis; because, in this study, the level of expression in HCCs was compared with the level in nontumorous liver tissues. The correlation between the down-regulation of SFRP1 expression and the methylation status of SFRP1 in HCCs was statistically significant (Table 4) (P = 0.021). There were HCCs without methylation; however, their SFRP1 mRNA levels were down-regulated. This suggests that other mechanisms, such as LOH and/or other epigenetic alterations, may play a role in the down-regulation of SFRP1 in HCCs.
|Patient no.||SFRP1 methylation status||ΔCt = (Avg. SFRP1Ct −Avg. HPRTCt)||ΔΔCt = ΔCtT-ΔCtN||SFRP1 tumor part relative to nontumor part|
LOH at the SFRP1 Locus in Primary HCCs
We also investigated the status of LOH at the SFRP1 locus in primary HCCs. Two microsatellite markers were used, D8S505 and D8S1722, which were found on chromosome 8p11.2 flanking the SFRP1 locus. In the 54 pairs of HCCs examined, heterozygosity exceed 50% at each of the 2 microsatellite markers, and LOH of markers D8S505 and D8S1722 was 25% (9 of 36 HCCs) and 27.6% (8 of 29 HCCs) of the informative cases, respectively (Fig. 4,) (Table 1). In addition to Patient 7, who had a sample poor RNA quality, in all of the other samples (Patients 8, 9, 19, 25, 29, 37, 38, 44, 46, 48, and 53) who had tumors with LOH, at least 1 marker showed the down-regulation of SFRP1 mRNA expression (Table 5). Although LOH was not too frequent, this finding supports our suggestion that LOH may play a role as another mechanism in the down-regulation of SFRP1 in HCCs.
Association between SFRP1 Hypermethylation and Clinicopathologic Parameters in Patients with HCC
The correlations between promoter methylation status and clinicopathologic characteristics are shown in Table 3. We did not observe any significant correlation between SFRP1 methylation status and clinicopathologic characteristics among the patients with HCC.
The Wnt/β-catenin signal-transduction pathway is important in tumorigenesis and embryogenesis.19–27, 39, 40, 47–50 Genetic and epigenetic inactivation of APC and E-cadherin is observed frequently in human cancers.49–57 Although Wnt signaling involvement in carcinogenesis has been studied extensively, an exploration of the associations of the SFRP family with tumorigenesis has begun only recently. Previous studies have shown SFRP down-regulation in many cancers, including colorectal cancer, gastric cancer, etc.40, 42, 43, 45, 58 In the current study, first, we demonstrated that SFRP1, which is an antagonist of the Wnt/β-catenin signal pathway, was hypermethylated and down-regulated significantly in HCCs compared with nontumorous livers (normal liver tissues, livers with chronic hepatitis, and cirrhotic livers; P = 0.0001) (Table 2). The frequency of SFRP1 methylation in primary HCCs (48.2%) observed in this study was lower than that found in primary colorectal cancer (82%)45 but higher than that found in ovarian cancer (12%).43 Treatment with a demethylation agent, 5-Aza-CdR, was capable of restoring SFRP1 gene expression in HCC cell lines (Fig. 2). There was a significant correlation between methylation and transcription in primary tissues (P = 0.021) (Table 4). Taken together, these data suggest that SFRP1 down-regulation is associated with methylation-mediated gene silencing.
SFRP1 was down-regulated >2-fold in the absence of promoter hypermethylation in 66.7% of HCCs (12 of 18). The decreased SFRP1 mRNA level may have been caused by genetic alteration (such as LOH) (Tables 1 and, 5) (Fig. 4), mutation, or homozygous deletion, such as what occurs in colorectal cancer.59 In accordance with our data, DNA methylation was detected in liver samples with chronic hepatitis and with cirrhosis, indicating that DNA methylation may be an early event in the pathogenesis of HCC.60, 61 These data suggest that persistent hepatitis virus infection may be associated with SFRP1 promoter methylation in hepatocarcinogenesis. This finding needs to be confirmed by studies in larger numbers of liver cancers.
Dysregulation of the Wnt-β-catenin pathway frequently is activated because of mutations in any components (CTNNB1, AXIN, or FZD7) of this signaling pathway in colon cancers, HCCs, and other cancers.49–51 Suzuki et al.45, 58 have reported on experiments in which they expressed SFRPs in colon cancer cell lines that carried mutations in CTNNB1 or APC. SFRP1, SFRP2, and SFRP5 suppressed Wnt-dependent transcription by −60%. Those results established that Wnt signal activation by mutant β-catenin or APC can be suppressed in part by upstream ligand competitors. Accordingly, it may be possible to suppress the tumor phenotype in Wnt-activated cancer cells by inhibiting the Fz receptor through competition with antagonists. In fact, Suzuki et al. demonstrated that colon cancer cells that overexpress SFRPs have less colony formation and a higher rate of apoptosis. In the current study, we demonstrated that SFRP1 mRNA expression was down-regulated frequently in HCC, and this down-regulation often involved methylation-mediated gene silencing. However, we did not study the biologic function of SFRP1 in HCC. Determining whether the function of SFRP1 in HCC is similar to that in colon cancer will require further investigation.
In conclusion, promoter hypermethylation of SFRP1 is a common event in HCC and plays an important role in the regulation of SFRP1 expression. In addition to methylation-mediated down-regulation of SFRP1, LOH also may play a role. Further investigations will be required to elucidate the importance of SFRP1 in the development of HCC.
- 15Allelic imbalance regions on chromosomes 8p, 17p and 19p related to metastasis of hepatocellular carcinoma: comparison between matched primary and metastatic lesions in 22 patients by genome-wide microsatellite analysis. J Cancer Res Clin Oncol. 2003; 129: 279–286., , , et al.
- 54Epigenetic and genetic alterations of APC and CDH1 genes in lobular breast cancer: relationships with abnormal E-cadherin and catenin expression and microsatellite instability. Int J Cancer. 2003; 106: 208–215., , , et al.
- 60Genetic 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: 970–979., , , , , .