Surface, core, and X genes of hepatitis B virus in hepatocellular carcinoma
An in situ hybridization study
Article first published online: 27 NOV 2002
Copyright © 2003 American Cancer Society
Volume 99, Issue 1, pages 63–67, 25 February 2003
How to Cite
Jayshree, R. S., Sridhar, H. and Devi, G. M. (2003), Surface, core, and X genes of hepatitis B virus in hepatocellular carcinoma. Cancer, 99: 63–67. doi: 10.1002/cncr.10954
- Issue published online: 14 FEB 2003
- Article first published online: 27 NOV 2002
- Manuscript Accepted: 1 AUG 2002
- Manuscript Revised: 30 JUL 2002
- Manuscript Received: 29 MAR 2002
- hepatitis B virus;
- hepatocellular carcinoma;
- in situ hybridization;
- hepatitis B surface gene;
- hepatitis B core gene;
- hepatitis B X gene;
- hepatitis B surface antigen;
- nonisotopic in situ hybridization
The incidence of hepatocellular carcinoma (HCC) and the seroprevalence of hepatitis B virus (HBV) in this disease state are significantly higher in South India than in North India. Because data on serologic studies do not project the actual association between the two parameters, this study was undertaken.
The prevalence of HBV genes in HCC patients was studied using nonisotopic in situ hybridization. Fifty patients from South India were diagnosed with HCC after performing ultrasound-guided fine-needle aspiration biopsies of liver lesions. The diagnosis was confirmed by cell block studies. Sections cut from paraffin-embedded cell blocks made out of the aspirates were probed with digoxigenin-labeled surface, core, and X regions of the viral genome.
Nuclear integration of the surface gene was observed in 100% (50 of 50), the core gene was positive in 94% (47 of 50), and the X gene was present in 98% (49 of 50) of the cases. An episomal form of the virus was not found. Serum hepatitis B surface antigen was positive only in 48% (12 of 25) of the patients screened.
We found molecular evidence that HBV is an important contributing factor in the etiology of HCC in South India. In HCC, the S gene of the virus was the most prevalent followed by the X and C genes. Only integrated forms of the viral DNA were observed. Nonisotopic in situ hybridization using multiple regions of the viral genome is a good technique for studying this association. It has an added advantage over polymerase chain reaction, of localization of signals in a tumor cell. Cell blocks made from fine-needle aspirates are ideal for in situ hybridization. Cancer (Cancer Cytopathol) 2003;99:63–67. © 2003 American Cancer Society.
Hepatocellular carcinoma (HCC) is ranked as one of the five leading human cancers worldwide, causing at least 250,000 deaths annually.1 The geographic distribution of HCC has suggested some risk factors that predispose individuals to the development of this disease, namely, infection with hepatitis viruses, exposure to aflatoxin, alcohol consumption, and oral contraceptive use.1 It is now accepted that hepatitis B virus (HBV) has carcinogenic potential in humans. Direct and indirect pathways of carcinogenesis are probable. There is evidence pointing to both being involved in HBV-induced HCC. The direct effect of HBV infection is mediated through insertional mutagenesis, with the transactivating properties of products of the X gene and perhaps a truncated preS/S gene being a likely component. Indirectly, persistent inflammation, fibrosis, and cirrhosis result. Alternatively, HBx protein inactivates the p53 tumor suppressor gene by forming a complex with it.2 There are 34 million chronic carriers of HBV in India.3 Therefore, the probability of a positive association between HBV infection and HCC is high. Data from population-based cancer registries from across the country suggest that the incidence of HCC is significantly higher in South India than in North India. The average incidence rate in the South vs. the North is 8.3 versus 4.8 per 100,000 males, respectively.4 A case–control study done simultaneously at Chennai (South India) and Delhi (North India) revealed a higher seroprevalence of HBV in Chennai than in Delhi (74% vs. 59%), as well as a higher number of patients with HCC (35 vs. 17).5 Earlier studies from India, using limited serologic markers, showed a 33–60% association between HBV infection and HCC.6, 7 However, other studies using extended serologic markers showed that association to be about 74%.5 Autopsy studies on HCC cases using HBV markers found an association of 80%.3 Recent reports using different molecular methods on patients from North India have shown the association between HBV infection and HCC to vary between 72% and 91%.8, 9 There are no similar published studies on HCC patients of South Indian origin. The current study was undertaken to determine the prevalence of HBV genes in HCC patients in a South Indian population using the nonisotopic in situ hybridization (NISH) technique.
MATERIALS AND METHODS
Fifty HCC cases were included in the current study. The patients were registered cases of Kidwai Memorial Institute of Oncology, Bangalore, who were seen from February 1997 to December 1998 and from February 2000 to September 2000. The patients came from South India: 14% (7 of 50) were residents of the State of Andhra Pradesh and 86% (43 of 50) were residents of the State of Karnataka. The patients ranged in age from 20 years to 80 years with a median age of 54 years. Eighty-four percent of the patients were 40–79 years old. There was a male predominance, the male-to-female ratio being 11.5:1.
Liver lesions were sampled by ultrasound-guided fine-needle aspiration (FNA). Smears made from the aspirated material were fixed in methanol and stained by May-Grunwald Giemsa and Pap stains. The remaining aspirate was fixed in buffered formalin overnight. Subsequently, the fixed material was centrifuged and made into cell blocks.10 The diagnosis of HCC was made cytologically in correlation with the clinical and ultrasound findings and biochemical parameters wherever available. The preliminary diagnosis and grading of HCC on smears made from the aspirates of liver lesions were confirmed by histopathologic examination of paraffin sections made from cell blocks of the aspirates (hematoxylin and eosin stained).
Hepatitis B surface (HBS) DNA (477 bp; position 108 to 585 on the HBV genome map), hepatitis B core (HBC) DNA (258 bp; position 1763 to 2017 on the HBV genome map), hepatitis B X (HBX) DNA (265 bp; position 1434 to 1683 on the HBV genome map), and HIV2 cDNA (500 bp; negative control) were used as probes. These four DNA fragments were labeled with digoxigenin-dUTP by random priming using a DNA labeling kit (Boehringer Mannheim, Mannheim, Germany). Dr. G. Sridharan (Department of Clinical Virology, Christian Medical College and Hospital, Vellore, India) kindly provided amplified polymerase chain reaction (PCR) products of the HBS DNA, HBC DNA, and the HIV-2 cDNA. The amplified PCR product of the HBX DNA was a gift from Dr. C. Schroeder (Deutsches Krebsforschungszentrum, Heidelberg, Germany).
The procedure of Han et al.11 was followed. Briefly, 5-μm sections were cut and placed on silane coated slides. Following deparaffinization and rehydration, the sections were permeabilized with 0.2 N HCl and digested with 40 μg/mL proteinase K in 10 mM Tris HCl, pH 7.5, and 2 mM CaCl2 for 30 minutes at 37 °C. RNA in the tissue was destroyed by incubating the sections in 100 μg/mL of RNAse (in 0.015 M phosphate-buffered saline, pH 7.5, containing 10 mM MgCl2) for 30 minutes at 37 °C. Endogenous alkaline phosphatase was destroyed by incubating the sections in 20% glacial acetic acid at 4 °C for 15 seconds. The sections were denatured by heating at 98–100 °C for 6 minutes followed by immediate cooling on crushed ice for 1 minute. The probe was denatured by heating at 100 °C for 5 minutes, after which it was cooled on ice for 1 minute. The denatured probe was added to the section (4–5 ng probe per slide), covered with parafilm, and hybridized for 12–16 hours at 42–45 °C in 50% formamide buffer. After stringent washes in varying concentrations of salt (2 × saline sodium citrate [SSC], 1 × SSC, 0.5 × SSC, 0.1 × SSC) and at different temperatures (room temperature and 42°C), immunologic detection was performed using antidigoxigenin alkaline phosphatase- conjugated antibody (appropriately diluted in detergent buffer). After the washes in detergent buffer, color development was done using a nitroblue tetrazolium (NBT)-based color reaction (5-bromo-4-chloro-3-indolyl phosphate and NBT salt) for overnight at room temperature, followed by counterstaining with 1% Neutral Red. The slides were then dried and mounted (in dibytyl phthalate xylene [DPX; Qualigens Fine Chemicals, Mumbai, India]). Sections of paraffin-embedded liver biopsies from HBV-positive and negative hepatitis cases were used as controls (a kind gift from the Department of Pathology, Christian Medical College and Hospital, Vellore, India).
Serum hepatitis B surface antigen (HBsAg) was tested using at least two commercial third-generation sandwich enzyme-linked immunosorbent assay kits (Pathozyme HBsAg, Scotland, United Kingdom; Surase B-96, Taiwan, Republic of China; Heprofile HBsAg, Ontario, Canada; Enzygnost HBsAg, Marburg, Germany).
Cytologic Diagnostic Features
The cytologic features of individual aspirates were variable depending on the differentiation of tumor. Fine-needle aspiration smears were very cellular, yielding fragments of tumor tissue and cell clusters. In well differentiated tumors, cells were arranged in a trabecular pattern, in thick cords, papillae, cell balls, and acinae. Some of the cell clusters were separated by sinusoids lined with endothelial cells. The cells resembled hepatocytes, i.e., they had polygonal outlines and a granular or vacuolated cytoplasm. Intracytoplasmic inclusions (hyaline globules) were noted in some cells. The nuclei were centrally located with granular chromatin often with prominent macronuclei. Some of the nuclei had intranuclear–cytoplasmic inclusions. Some of the cells showed intracytoplasmic bile pigments and bile plugs were seen occasionally between the cells. Many naked or stripped malignant nuclei were seen scattered in the background. In less differentiated tumors, smaller sheets or single cells were seen with poorly defined outlines, nuclear pleomorphism, anisocytosis, increased nuclear/cytoplasmic ratios, and prominent nucleoli. Binucleated and multinucleated tumor giant cells were also seen focally. Areas of necrosis and hemorrhage were noted. Moderately differentiated HCC had cytologic features intermediate to those of well and poorly differentiated HCC. Twenty-three cases were well differentiated, 24 were moderately differentiated, and 3 were poorly differentiated HCCs. Figure 1a shows the morphology of HCC in a cell block preparation.
Nuclear integration of the surface gene was observed in all 100% (50 of 50), the core gene was positive in 94% (47 of 50), and the X gene was present in 98% (49 of 50) of the cases studied. The distribution of the signals was localized to the nucleus only in all positive cells (Fig. 1b). No signals were found in the cytoplasm of the tumor cells in any of the cases studied. The signals were not seen uniformly in all cells. Some cells did not have any signals whereas others showed prominent signals. Serum HBsAg was positive in 48% (12 of 25) of the patients. The remaining 25 patients could not be tested for serum HBsAg status because they were lost to follow-up. However, the average HBsAg positivity in HCCs over a period of 10 years at our center has been about 40% (unpublished observation).
In the current study, we investigated the presence of the HBV genes in the liver of HCC patients by NISH. We found 100% of the HCCs were positive for the surface gene, even in the absence of circulating HBsAg. The high positivity obtained in our study was comparable to some reports from high HBV endemic areas like China and Korea. In those two countries, a positivity of 100% has been reported using multiprimered PCRs.12, 13 However, other studies from the same countries have reported prevalence rates ranging from 65% to 82%.14–16 This difference in prevalence in the same country could either be due to variation in the prevalence of the virus across the country, to varying degrees of sensitivities of the molecular techniques used, or to the differences in the regions of the viral genome used as probes and primers. In a large study on HCC patients from North India, Sarin et al.9 showed HBV positivity in 72% of cases using dot blot hybridization. However, the same group of workers in an earlier study from Delhi had shown a positivity of 91% by PCR.8 The relatively lower prevalence observed by Sarin et al.9 may have been due to the lower sensitivity of dot blot hybridization compared with PCR, which was used in the earlier study.8 The difference between their study9 and our current study could be due to a difference in the prevalence of the virus in North India versus South India. Alternatively, the difference could also be ascribed to the differences in the sensitivities of the dot blot and in situ hybridization tests. It is noteworthy that there was no significant difference in the region of the HBV genome used as probes by both the studies. Although Katiyar et al,8 showed a high prevalence (91%) in the North Indian population using PCR, they may have had a 100% prevalence had they used a combination of primers directed against various regions of the viral genome. This point has been highlighted in a previous study.17 The C gene of the HBV is the most conserved region regardless of subtypes of the virus.18, 19 It is, therefore, prudent to use primers/probes directed against this region while studying HBV infection. However, in any study encompassing HBV and HCC, the possibility of deletions of portions of the viral genome as a result of insertional mutagenesis should be borne in mind. In addition, there may be immune selection of cells with integrated viral DNA that are incapable of expressing hepatitis B core and e antigens.20 In the present study, we found 6% of the HCCs to be negative for the C gene. To prevent false-negative results, primers or probes should encompass more than any one region of the viral genome. Our results are in contrast to Western and Japanese studies, which report that the hepatitis C virus (HCV) is associated more frequently and is etiologically implicated with HCC than with HBV.14, 21, 22 Active HBV vaccination programs are probably responsible for this change in viral prevalence in Japan both in HCC patients and in the general population.2
Well differentiated HCCs are permissive to viral growth and replication as substantiated by a Korean study.13 However, we found that irrespective of the degree of differentiation, the tumor cells were nonpermissive. Our finding of 100% nuclear integration is in contrast to the finding of Sarin et al.,9 who found integration in only 59% of the cases by genomic Southern blot hybridization. The low percentage of integration observed in their study could be attributed to the technique used, as shown by the titration studies of Shintani et al.23 An in situ hybridization study in HBsAg-positive Chinese patients with HCC revealed the integrated virus in 72% and the episomal state in 19% of patients.15 A French study using genomic Southern blot hybridization had reported integration in 85% of patients.24
The distribution of positive signals was not seen uniformly in all cells. There were cells without any signals and some with prominent signals, as observed by Zhou et al.16 This variation may have been caused by HBV DNA integration in the early stages of hepatocarcinogenesis that was lost during clonal evolution.23 Expression of HBV genes, although presumably involved in early transformation, is dispensable at later stages of tumor progression.
The prevalence of S, precore, and X genes of HBV by nested PCR in HCC patients in a Korean study was 100%, 90%, and 80% respectively.13 Our results on the integration of S and C genes (100% and 94%) compared well with their findings. However, the prevalence of the X gene integrant in our study (98%) was higher. This difference could have been due to the variations in the regions of the viral genome used as probes/primers. A Japanese study has reported a prevalence of the HBX integrant in HCC patients as 88% and HBS/preS2 as 38%.25 Another study in HCV-positive HCC patients showed a varying degree of positivity of various HBV genes: X and C being 100% each and S gene being 77%.23 The lower positivity of the S gene in the study of Shintani et al.23 has been attributed to either a lower sensitivity of PCR using the S primer pair or to the absence of the S gene during the process of oncogenesis.
Serum HBsAg was positive only in 48% of our HCC cases whereas S gene positivity was seen in 100%. The absence of serum HBsAg in the remaining 52% could be due to infection with the surface mutant form of the virus. In India, 49% of the patients with HBV-related HCC are infected with surface mutants.9 Alternatively, serum HBsAg negativity in these cases could be caused by a disrupted integrated S gene because rearrangements and deletions of parts of the viral genome are common during the process of integration.17 However, occult HBV infection may also be responsible for cryptogenic HCC cases in the United States,26 for 59% of serologically undefined cases of HCC in Japan,27 and for 85% of cases in France.24
We found molecular evidence that HBV is an important contributing factor in the etiology of HCC in South India. The prevalence of HBV genes in this association was S is greater than X is geater than C. Only integrated forms of the viral DNA were observed. NISH using multiple regions of the viral genome as probes is a good technique for studying this association. It has an added advantage over PCR of localization of signals in a tumor cell. Cell blocks made out of fine-needle aspirates provide a good material for in situ hybridization. It would be interesting to know whether HCV and dietary aflatoxins play any role as contributory factors in the molecular pathogenesis of HCC associated with HBV in this part of the country.
The authors acknowledge the guidance of Dr. Sudhir Krishna, National Center for Biological Sciences, Bangalore, in helping Dr. R.S. Jayshree in standardizing nonisotopic in situ hybridization.
- 3Hepatitis B: who should be vaccinated. In: SarinSK, SinghalAK. Hepatitis B in India. Problems and prevention. New Delhi: CBS Publishers, 1996: 123–131., .
- 4National Cancer Registry Programme Biennial Report. 1988-1989. Indian Council of Medical Research, New Delhi, July 1992.
- 5Clinical spectrum of hepatitis B in India. In: SarinSK, SinghalAK. Hepatitis B in India. Problems and prevention. New Delhi: CBS publishers, 1996: 51–63..
- 6Hepatitis B vaccination and India. Natl Med J India. 1991; 4: 103–104..
- 10Cytologic techniques. In: KossLG. Diagnostic cytology Philadelphia: Lippincott, 1992: 1451–1515., .
- 20Viral hepatitis — the alphabet so far. Virus Life. 1993; 6: 2–8., , .