This article is being submitted as part of Kok-lung Chan's master's thesis.
High-density allelotyping of chromosome 8p in hepatocellular carcinoma and clinicopathologic correlation
Article first published online: 12 JUN 2002
Copyright © 2002 American Cancer Society
Volume 94, Issue 12, pages 3179–3185, 15 June 2002
How to Cite
Chan, K.-L., Man-Fong Lee, J., Guan, X.-Y., Fan, S.-T. and Oi-Lin Ng, I. (2002), High-density allelotyping of chromosome 8p in hepatocellular carcinoma and clinicopathologic correlation. Cancer, 94: 3179–3185. doi: 10.1002/cncr.10612
- Issue published online: 12 JUN 2002
- Article first published online: 12 JUN 2002
- Research Grant Council. Grant Number: HKU 7342/00M
- hepatocellular carcinoma;
- chromosome 8p;
- loss of heterozygosity;
- clinicopathologic correlation
Allelic deletions are frequent genetic alterations in patients with hepatocellular carcinoma (HCC).
To evaluate the allelic losses on chromosome 8p in HCC patients and define their clinicopathologic significance, we performed high-density allelotyping on 8p in 60 patients with HCC and analyzed the clinicopathologic correlation.
Using 24 microsatellite markers, allelic losses on 8p were frequent. Loss of heterozygosity (LOH) at one or more loci was observed in 34 (57%) HCC patients. When the allelic losses were compared between groups categorized by clinicopathologic variables, significant correlation was found between tumors with interstitial losses and larger tumor size (> 5 cm; P = 0.026). In addition, allelic loss at D8S298 at 8p22 was associated closely with venous permeation, tumor microsatellite formation, and larger tumor size (P = 0.019, 0.024, and 0.007, respectively). LOH at locus D8S1721 at 8p23.1 was seen more frequently in nonencapsulated tumors (P = 0.007) and LOH at D8S1771 at 8p21.3 was associated with a larger tumor size and poorer cellular differentiation (P = 0.018 and 0.049, respectively).
Allelic losses on 8p are frequent in HCC patients. Association of allelic losses at specific loci on 8p with a more aggressive tumor behavior suggests that loss/inactivation of putative tumor suppressor gene(s) located at these regions may confer a tumor growth advantage and contribute to the progression of HCC. Cancer 2002;94:3179–85. © 2002 American Cancer Society.
Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide. It has a characteristic geographic distribution with the highest incidence in Southeast Asia and sub-Saharan Africa, and is the second most common cause of death due to cancer in Hong Kong and Southeast Asia. The risk factors are well known, namely, chronic hepatitis B and C virus infections, cirrhosis, and exposure to aflatoxin B1. These etiologic factors predispose hepatocytes to the development of HCC either directly or indirectly.1 Carcinogenesis is a result of accumulation of genetic alterations of both protooncogenes and tumor suppressor genes. Although alterations of protooncogenes are uncommon in patients with HCC,2 chromosomal deletions in regions containing candidate tumor suppressor genes or gene mutations of possible tumor suppressor genes have been reported.
Allelic deletions on chromosome 8p are frequent in patients with solid cancers such as colorectal, lung, prostate, breast, and oral carcinomas.3–7. These findings suggest that chromosome 8p may contain one or more tumor suppressor genes. Moreover, allelic losses on chromosome 8p are associated with a more aggressive clinical behavior with metastasis occurring in patients. For example, deletion of chromosome region 8p22 was associated with an invasive phenotype in patients with urinary bladder carcinoma8 and deletion at 8p22–8p23.1 was associated with an advanced tumor stage and aggressive histologic type in breast carcinoma patients.7
Studies on comparative genomic hybridization (CGH) or loss of heterozygosity have shown that 8p is also one of the most frequently altered chromosomes in HCC patients.9–12 However, most LOH studies focused on genomewide investigations, and reports on detailed deletion mapping as well as detailed clinicopathologic correlation are scant. To identify regions of frequent allelic losses on chromosome 8p and to delineate the relation between allelic losses and clinical tumor behavior, we performed a high-density allelotyping analysis on tumor and nontumorous liver samples along the entire length of chromosome 8p in 60 HCC patients. The clinicopathologic significance of the different patterns of chromosome 8p deletions and of LOH at individual loci was also analyzed in detail.
MATERIALS AND METHODS
Patients and Specimens
Sixty patients with primary HCC were selected randomly and and underwent resection at the University of Hong Kong, Queen Mary Hospital, between October 1992 and August 1999. All patients were Chinese. Of the participants, 47 were men and 13 were women, ranging in age from 28 to 82 years (mean: 56 years). Of the 60 patients, 46 (77%) were positive for hepatitis B surface antigen (HBsAg), whereas only 4 (7%) patients were positive for hepatitis C virus antibody (anti-HCV). The criteria for resectability were the absence of distant metastasis or main portal vein thrombosis, anatomically resectable disease, and adequate liver function reserve. None of the patients had received other therapies including chemoembolization or chemotherapy before resection. The tumor and nontumorous liver samples (at least 1 cm away from the corresponding tumors) were snap-frozen in liquid nitrogen immediately after resection and kept at −70 °C. Serum HBsAg and anti-HBs were assayed by an enzyme immunoassay test (Abbott Laboratories, Chicago, IL).
Genomic DNA Extraction
Each frozen sample was embedded in O.C.T. compound (Tissue-Tek, Torrance, CA) and twenty 16-μm sections were cut serially. A 6-μm section from the tumor and nontumorous liver blocks in each case was stained with hematoxylin and eosin to confirm the homogeneity of the tumor and nontumorous cell population. Genomic DNA was extracted from the sections of each sample by standard phenol/chloroform extraction after sodium dodecyl sulfate/proteinase K digestion.
Multiplex Polymerase Chain Reaction
Twenty-four polymorphic microsatellite markers with an average of 2.5 cM genetic distance along the entire length of chromosome 8p and two polymorphic microsatellite markers on 8q were used in this study (Table 1). Two rounds of each multiplex PCR were performed, using three to six of the above primer pairs in each of the multiplex PCR. Each primer pair was labeled with a fluorescent dye in a reaction mixture with a final volume of 10 μl. The mixture consisted of 1 × PCR buffer, 2.5 mM MgCl2, 250 μM dNTP, 5 pmol of each primer, 50 ng of genomic DNA, and 0.75 U AmpliTaq Gold polymerase. The PCR was performed with an initial temperature of 96 °C for 12 minutes, followed by 30 cycles of 94 °C denaturation for 15 seconds, 55 °C annealing for 30 seconds, 72 °C elongation for 1 minute, and a final 72 °C extension for 12 minutes in a GeneAmp PCR system 9600 (Perkin-Elmer, Applied Biosystem, Foster City, CA; BMBIFT, HI). Single PCRs were performed to confirm the multiplex PCR products when necessary.
|Genetic distance from 8p telomere (cM)||Locusa||Size (bp)||Manufacturerb||Heterozygozity||LOH/informative marker||LOH %|
Microsatellite Analysis for LOH
The PCR products were loaded on a 5% polyacrylamide gel containing 6 M urea and run for 2 hours in an ABI PRISM 377 automated sequencer (Perkin-Elmer). The data were collected and quantified with the GeneScan and Genotyper 2.5 software (Applied Biosystems, Foster City, CA). The peak signal intensities of the alleles of tumor DNA were compared with those of the corresponding nontumorous liver DNA in each case. A reduction in marker signal intensity greater than 50%, i.e., when the LOH index calculated from the Genotyper software was less than 0.5 or greater than 2, according the following formula, was defined as LOH: LOH index = (T2/T1) / (N2/N1), where “T” was the tumor sample, “N” was the normal sample, and 1 and 2 were the intensities of smaller and larger alleles, respectively. The experiments were repeated at least once for all LOH-positive loci and some query loci by individual PCR for confirmation.
Ten tumor-related factors were assessed as described previously.13 The size of the largest tumor nodule and the number of tumor nodules were assessed grossly. Microscopically, the presence of a tumor capsule was noted. Spread of the tumor was assessed by evidence of venous invasion, tumor microsatellite formation, and direct invasion into the adjacent liver parenchyma. Venous invasion and tumor microsatellite formation were evidence of intrahepatic tumor metastasis. Cellular differentiation was classified according to Edmondson and Steiner. Presence of cirrhosis and chronic hepatitis in the nontumorous liver samples and tumor at resection margin were assessed histologically. Immunoperoxidase staining on liver tissue was performed using monoclonal antibody against HBsAg (Dako, Glostrup, Copenhagen).
The Fisher exact test or chi-square test was used to analyze the categorical data. Analysis of variance, t –test, and the Kruskal–Wallis test were used to analyze continuous data as appropriate. Tests were considered significant when P values were equal to or less than 0.05.
LOH on Chromosome 8p
Twenty-four polymorphic microsatellite markers with an average of 2.5-cM genetic distance along the entire length of chromosome 8p and two polymorphic microsatellite markers on 8q were used in this study. Representative examples of LOH results are shown in Figure 1. The overall frequency of informative loci was 62.7%. Most of the 26 microsatellite markers had heterozygozity ranging from 0.52 to 0.91, which was similar to ranges published on the genome database, except four markers which had heterozygosity ranging from 0.32 to 0.45 (Table 1, Fig. 2). There was a high frequency of allelic loss on 8p, and the frequency of LOH at each locus on chromosome 8p ranged from 21.1% to 52.9% (mean: 40.4%). Locus D8S520 had the highest frequency of LOH (52.9%; Table 1, Fig. 2). In contrast, LOH of the two markers on chromosome 8q were only 8.7%and 12.2%, respectively, showing that allelic losses were confined largely to 8p, not extending to the centromeric 8q region.
Among the 60 cases examined, 34 (57%) tumors showed allelic losses at at least one locus on chromosome 8p. Eighteen (53%) of these 34 tumors showed allelic loss at all of the loci on 8p, and were regarded as having hemizygosity, whereas the remaining 16 (47%) tumors exhibited varying frequencies of allelic losses (interstitial deletions). A deletion map was constructed to illustrate the allelic loss pattern according to the distribution of LOH and retention loci (Fig. 3).
The pathologic features of the 60 cases of HCC were characterized. The tumor size ranged from 2 to 25 cm at greatest dimension (mean: 8.3 cm). Forty-nine (81.7%) of the HCC cases had solitary tumor nodules. Twenty-nine (48%) tumors were classified as having better cellular differentiation (Edmondson Grades I–II). Tumor encapsulation and venous permeation were detected in 24 (40%) and 37 (62%) of the 60 tumors, respectively. Thirty-six (60%) tumors had tumor microsatellite formation and 12 (20%) tumors had direct liver invasion. Twenty-nine (48%) and 25 (42%) of the 60 patients had cirrhosis and chronic hepatitis, respectively, in the resected livers. The resection margin was clear histologically in all but two patients.
When the pathologic features were correlated with the findings of the allelic status, we found that tumors with interstitial deletions or hemizygosity on 8p were associated significantly with tumor size greater than 5 cm (P = 0.026; Table 2.) Tumors arising from cirrhotic livers had fewer partial deletions or no deletion on 8p than those arising from chronic hepatitic or normal livers (P = 0.042; Table 2).
|Parameter||Interstitial/complete 8p loss|
|Tumor size (cm)|
|Cellular differentiation (Edmondson's classification)|
|Tumor microsatellite formation|
|Direct liver invasion|
|No. of nodules|
|Histologic resection margin|
We further investigated the clinicopathologic significance of allelic losses at the 26 loci. Allelic loss at locus D8S1771 located at 8p21.3 was seen more frequently in tumors larger than 5 cm (P = 0.018) and more often had poorer cellular differentiation (P = 0.049; Table 3). Tumors having LOH at locus D8S1721 at 8p23.1 were more frequently nonencapsulated (P = 0.007). In addition, allelic loss at D8S298 at 8p22 was seen more often in tumors having venous permeation, tumor microsatellite formation, and larger tumor size (P = 0.019, 0.024, and 0.007, respectively; Table 3). In HCC patients, venous invasion and tumor microsatellite formation are evidence of intrahepatic tumor metastasis.
|Parameter||Allelic loss on D8S1771|
|Tumor size (cm)|
|Cellular differentiation (Edmondson's classification)|
|Allelic loss on D8S1721|
|Allelic loss on D8S298|
|Tumor size (cm)|
|Tumor microsatellite formation|
In this study of HCC in a Chinese population using high-density allelotyping, allelic deletions on chromosome 8p were frequent (57%). Among the 34 tumors with allelic losses on 8p, 53% showed 8p hemizygosity and 47% had interstitial deletions. Our result is in agreement with those reported in previous studies on HCC in which allelic losses on chromosome 8p ranged from 37% to 65%, using CGH or LOH assay.9–12, 14
Tumors with interstitial or complete 8p deletions had a significantly larger tumor size (P = 0.026). It is well established that larger tumor size is a risk factor for tumor recurrence and a poorer prognosis.15 Allelic deletions at several specific loci were associated with a more aggressive tumor phenotype. Tumors with allelic loss at D8S1771 at 8p21.3 frequently had greater tumor size and poorer cellular differentiation (P = 0.018 and 0.049, respectively). Both larger tumor size and poorer cellular differentiation are risk factors associated with a poorer prognosis.15, 16 These pathologic features usually occur late in hepatocarcinogenesis, raising the possibility that deletion in this region is a late event in hepatocarcinogenesis. Deletion on 8p21.3-p22 in HCC patients isassociated with a larger tumor size and advanced stage.17 In breast carcinoma patients, deletion on 8p12-p21 is detected more frequently in tumors with larger size and infiltration.7 In this study, allelic deletion at D8S1721 at 8p23.1 was correlated strongly with the absence of tumor encapsulation (P = 0.007). In HCC patients, the absence of tumor encapsulation is associated with a more aggressive tumor behavior and unfavorable prognosis.18 These results supported that 8p deletions at specific loci were associated with tumor progression and an aggressive behavior. Deletion at locus D8S298 mapped at 8p22 was associated closely with venous permeation, tumor microsatellite formation, and larger tumor size (P = 0.019, 0.024, and 0.007, respectively). In HCC patients, venous permeation and tumor microsatellite formation represent intrahepatic metastasis. These features represent a more aggressive tumor behavior with metastasis in HCC patients and are well established significant risk factors for tumor recurrence and poor prognosis. In patients with urinary bladder carcinoma, 8p22 deletion identified by fluorescent in situ hybridization technique was associated strongly with an invasive tumor phenotype.8 Based on the significant clinicopathologic correlations, an important tumor suppressor conferring suppression of tumor invasion or a metastasis suppressor gene might be harbored at or around this locus. This finding agrees with a previous report using CGH analysis, which suggested that the 8p deletion was associated with tumor metastasis in HCC patients.19
In HCC patients, a candidate tumor suppressor gene named the liver-related putative tumor suppressor (LPTS) gene20 was identified recently and is located at 8p23.1, which shares the same region of location on 8p with D8S1721 in the current study. The expression of the LPTS gene was reduced significantly or was undetectable in HCC samples and cell lines. In addition, functional characterization showed that suppression of the LPTS gene by antisense oligodeoxynucleotides in normal liver cells could enhance cell growth.
Allelic losses at loci D8S1771 (mapped at 8p21.3) and D8S298 (at 8p22) were associated with a more aggressive tumor behavior. At 8p22, there are several putative tumor suppressor genes reported. DLC-1 (deleted in the liver cancer gene)21, 22 had 80% homology to rat RhoGAP. RhoGAP is a GTPase-activating protein for the Rho family proteins and is important in the regulation and organization of actin cytoskeleton and suppression of Ras-mediated tumorigenesis. The DLC-1 gene is situated at 8p21.3–22 and shares the region of location with D8S1771 (at 8p21.3) and D8S298 (at 8p22). The DLC-1 gene was not expressed frequently (20%) or was deleted in HCC patients.22 It also exerted cell inhibition effects in HCC cell lines with deleted DLC-1.22DLC-1 is a putative tumor suppressor gene in HCC patients. Two other candidate tumor suppressor genes, PRLTS (PDGF-receptor beta-like tumor suppressor)23 and FEZ1,24 have also been identified. PRLTS, mapped at 8p22, is commonly deleted in patients with sporadic HCC, colorectal carcinomas, and non–small cell lung carcinomsa,23 whereas FEZ1 gene expression is undetectable in more than 60% of epithelial tumors.24 Therefore, the data suggest that there may be a number of putative tumor suppressor gene loci on chromosome 8p.
We also observed that most of the tumors arising from cirrhotic livers had no allelic losses on 8p, compared with those arising from noncirrhotic livers (P = 0.042). Allelic deletions are frequent in cirrhotic foci in patients with HCC.25 Some of the cirrhotic foci may even harbor 8p losses identical to those found in the corresponding HCCs.26 Our reference control DNA was obtained from the corresponding nontumorous liver and not from normal DNA. If the cirrhotic livers (nontumorous livers) had already had LOH, then allelic deletions in the corresponding tumor tissues might not be detectable by LOH assay. As a result, the incidence of LOH in the HCCs arising from cirrhotic livers might be lower than that of HCCs arising from chronic hepatitic and normal livers. Further studies comparing normal DNA, nontumorous livers, and HCC simultaneously are needed to address this issue.
To conclude, allelic losses on 8p are frequent in predominantly hepatitis B virus-related HCCs in a Chinese population. Association of allelic losses at specific loci on 8p with a more aggressive tumor behavior suggests that loss/inactivation of putative tumor suppressor gene(s) located at these regions may confer a tumor growth advantage and contribute to the progression of HCC.
The authors thank Dr. Matthew Ng for his expert advice on statistical analysis.