Purpose: Genetic polymorphisms of UDP-glucuronosyltransferase 1A7 (UGT1A7), which detoxifies endogenous and environmental carcinogens, have been reported to be associated with hepatocellular carcinoma (HCC) in German populations. On the other hand, we reported that interleukin-1β (IL-1β) gene polymorphisms were associated with hepatitis C virus (HCV)-related HCC. In this study, we evaluated the association of both genes with the risk of HCC in Japanese HCV-infected patients.
Experimental Design: Genetic polymorphisms of UGT1A7 and IL-1β were investigated in 280 Japanese patients (122 with HCC and 158 without HCC) with chronic HCV infections, by use of standard PCR-based genotyping techniques.
Results: We designated the UGT1A7*1 allele (a haplotype conferring higher activity) as H and the *2, *3, and *4 alleles (haplotypes conferring lower activity) as L. The proportions of UGT1A7 L/L and H/L alleles (genotypes) in patients with HCC (25% and 45%, respectively) were higher than those in patients without HCC (15% and 39%, respectively) with odds ratios of 2.73 (95% confidence interval, 1.40–5.35) and 1.80 (95% confidence interval, 1.05–3.09), respectively, compared with the UGT1A7 H/H alleles. Multivariate analyses revealed that UGT1A7 L/L and IL-1β/−31T/T−511C/C genotypes, the presence of cirrhosis, age >60 years, male sex, and α-fetoprotein >20 μg/ml were associated with the presence of HCC (odds ratios, 2.33, 2.67, 4.20, 3.12, 3.09, and 2.90, respectively).
Conclusion: The UGT1A7 polymorphisms together with IL-1β were associated with the presence of HCC in Japanese HCV-infected patients.
It is estimated that more than 170 million people worldwide are chronically infected with the hepatitis C virus (HCV). The most important consequence of chronic HCV infection is progressive liver fibrosis leading to cirrhosis and finally to hepatocellular carcinoma (HCC), which has significant morbidity and mortality (1, 2, 3, 4) . Many factors, such as alcohol intake, older age at time of infection, male gender, and co-infection with the human immunodeficiency virus or hepatitis B virus accelerate disease progression in HCV infection (5, 6, 7, 8) . In addition, host genetic factors have recently been reported to increase the risk of HCC (9, 10, 11, 12, 13, 14, 15) .
The human UDP-glucuronosyltransferases (UGTs) represent an enzyme superfamily that is capable of catalyzing the glucuronidation of diverse compounds, including therapeutic drugs, endogenous metabolites (e.g., bilirubin and steroid hormones), and known human carcinogens, such as heterocyclic and polycyclic hydrocarbons and heterocyclic amines (16, 17, 18, 19, 20, 21) . UGTs can catalyze the conjugation of hydrophobic compounds of divergent chemical classes to form water-soluble β-d-glucopyranosiduronic acids. These metabolites then undergo renal or biliary elimination from the body. Because of this function, UGTs have been regarded as major biochemical factors in cellular defense and detoxification.
The N129K (codon 129: AAT→AAG), R131K (codon 131: CGA→AAA), and W208R (codon 208: TGG→CGG) polymorphisms of the UGT1A7 gene encode enzymes with lower carcinogen detoxification activities and are associated with certain cancers. In Caucasians and African-Americans, the UGT1A7*3 (129AAG-131AAA-208CGG) and UGT1A7*4 (129AAT-131CGA-208CGG) alleles (haplotypes) are related to orolaryngeal cancer (22) . In Germans, the UGT1A7*3 allele has been found to be associated with colorectal (23) and pancreatic cancer (24) . UGTs are also directly involved in liver diseases. Genetic variations in the UGT promoters are associated with mild and more severe forms of Gilbert’s syndrome (25) . The UGT1A7*3 allele has been identified in association with HCC in Germans (13) . On the other hand, we have reported that a single-nucleotide polymorphism (SNP) in the interleukin-1β (IL-1β) gene is associated with HCV-related HCC in a Japanese population (11) .
In this study, we first aimed to assess the association between the reported polymorphisms of the UGT1A7 gene and the presence of HCC in HCV-infected Japanese patients. We then combined UGT1A7 and IL-1β polymorphism information with clinical information in an attempt to improve the diagnosis of the risk for HCC in HCV-infected patients.
PATIENTS AND METHODS
To assess genetic polymorphisms related to HCC in HCV-infected patients, we studied 280 consecutive Japanese patients (194 men and 86 women; age range, 24–86 years; median age, 60 years) with chronic HCV infections (122 with HCC and 158 without HCC) who visited the outpatient clinic of the University of Tokyo Hospital between November 2000 and November 2001. To obtain an estimate of the genetic distribution of reference alleles in the general Japanese population, we also obtained DNA samples from 60 healthy individuals who visited our hospital (48 men and 12 women; age range, 24–53 years; median age, 31 years) with no history of liver diseases. We obtained written informed consent for DNA genotyping from all subjects. We also obtained approval for this study from the institutional ethics committee.
All of the patients were positive for HCV antibody, as determined by a second-generation enzyme immunoassay (Ortho Diagnostics, Tokyo, Japan) and were negative for hepatitis B surface antigen (Abbott Laboratories, North Chicago, IL). HCV RNA was measured with the Amplicor HCV assay, version 1 (Roche, Tokyo, Japan), and HCV genotypes were determined by a genotyping assay (SRL Laboratory Co., Tokyo, Japan). Any patient with a daily ethanol intake of ≥80 g for a period longer than 10 years was considered to have a positive history of alcohol abuse. The following clinical parameters were obtained for each patient at the time of whole-blood collection: age, gender, alcohol intake, serum albumin level, serum total bilirubin level, serum alanine aminotransferase (ALT) level, serum α-fetoprotein (AFP) level, prothrombin time, platelet count, and HCV serotype and serum viral load, as measured with the Amplicor-HCV monitor assay. Liver biopsies were performed on 194 patients within 6 months, and the diagnosis of liver cirrhosis was made based on liver histology according to the criteria of Desmet (7) and Scheuer et al. (8) . In patients without biopsy specimens, the diagnosis of cirrhosis was based on the presence of clinical manifestations: portal hypertension (e.g., varices, encephalopathy, or ascites), biochemical abnormalities (elevated serum bilirubin, decreased serum albumin, or prolonged prothrombin time), and obvious morphological changes in the liver, as detected by hepatic imaging (e.g., ultrasonography, computed tomography, arteriography, or magnetic resonance imaging). HCC was diagnosed based on several imaging modalities and was confirmed histologically in sonography-guided fine-needle biopsy specimens from all 122 patients. Initial screening examinations confirmed that no other cancers were present in any of the patients.
Genomic DNA was extracted from 100 μl of whole blood by use of the SepaGene kit (Sanko Junyaku, Tokyo, Japan) according to the manufacturer’s instructions. Extracted DNA was dissolved in 20 μl of 10 mm Tris-HCl buffer (pH 8.0) containing 1 mm EDTA, and the DNA samples were stored at −30°C until use. The genetic polymorphisms in UGT1A7 were determined by direct sequencing of amplified gene fragments.
Polymorphisms have been reported in 12 codons of the UGT1A7 gene: P11P, R68K, G115S, N129K, R131K, E139D, A144A, L147L, A169T, G173C, W208R, and T255T (13 , 26 , 27) . Eight of the 12 codon produce amino acid changes. We used the primers F1 (Ref. 13 ; 5′-gcggctcgagccacttactatattataggagct-3′) and R1 (Ref. 13 ; 5′-gcggatatccataggcactggctttccctgatgaca-3′) to amplify the specific UGT1A7 fragment that covers all 12 codons. A 917-bp fragment of UGT1A7 was amplified by PCR with extracted genomic DNA as the template. PCR was performed with Ready-To-Go PCR beads (Amersham Pharmacia Biotech, Uppsala, Sweden). The thermocycling conditions were as follows: 94°C for 10 min, followed by 30 cycles of 94°C for 30 s, 57°C for 60 s, and 72°C for 60 s. To verify the size of the PCR product, the amplicon was visualized on a 12.5% polyacrylamide gel with the appropriate size markers. For polymorphism determinations, direct sequencing was performed bidirectionally with 10 ng of QIAquick Spin-purified (Qiagen, Hilden, Germany) PCR product and the BigDye Terminator Cycle Sequencing Ready Reaction (PE Applied Biosystems, Foster City, CA), together with the primers S129F (5′-accattgcgaagtgc-3′), S129R (5′-acaattaagccacagg-3′), S208F (5′-ccctgctcctctttcctatg-3′), and S208R (5′-tgaaaaaataggggc-3′), followed by detection on an ABI 310 automated sequencer (PE Applied Biosystems; Ref. 11 ). Polymorphisms in the promoter region of the IL-1β gene were determined as described previously (11) .
We used a multivariate logistic regression model to calculate the statistical power required to detect the contribution of a SNP to the risk of harboring HCC, while including other known risk factors for HCC. SNP status was assigned continuous values of X = 0, 1, or 2, which represent homozygous for an allele, heterozygous, or homozygous for a different allele, respectively. The required sample size, n, for the multivariate logistic regression analysis was calculated using the following formula (28) : where Zu is the upper uth percentile of the standard normal distribution; P is the proportion of patients with HCC when X = 1; β* is the size of the effect of an SNP; and R2 is the multiple correlation coefficient (we used 0.009 from our data) relating X with the following four covariates: the presence of cirrhosis, male gender, high serum AFP level, and advanced age (29, 30, 31) . When the effect size of a SNP was assumed to be 0.69, which corresponded to an odds ratio (OR) of 2, the required sample size was calculated to be 75 or 100 for a statistical power of 80 or 90%, respectively. When the effect size of an SNP was assumed to be 1.39, which corresponded to an OR of 4, the required calculated sample size was 18 or 25 for a statistical power of 80 or 90%, respectively. On the basis of these calculations, our sample size was sufficient for conditions under which the OR of an SNP exceeded 2.
The associations between the clinical parameters (age, gender, alcohol intake, cirrhosis, serum albumin, serum total bilirubin, serum ALT, serum AFP, prothrombin time, platelet count, HCV serotype, serum viral load, IL-1β genotype, and UGT1A7 genotype) and the presence of HCC were evaluated using Student’s t test, the Mann–Whitney U test, one-way ANOVA, and the χ2 test. The association between the genotype of each locus and the presence of HCC was evaluated using the χ2 test. The Cochrane–Armitage test was used to test for trends. Possible confounding effects among these variables were adjusted using a multivariate logistic regression model, and the ORs and 95% confidence intervals (CIs) were calculated. P < 0.05 was considered significant in the two-tailed test. The Hardy–Weinberg equilibria of alleles at individual loci were evaluated using the HWE program.3 . The haplotype frequencies for pairs of alleles were estimated using the EH software for estimating haplotype frequencies.3 Linkage disequilibrium coefficients (D′ = D/Dmin or max) were calculated using the 2BY2 program.3 All other data analyses were performed using SPSS, version 10.0 (SPSS Inc., Chicago, IL) and the R statistical package.4
There were no significant differences in alcohol intake, HCV serotype, or viral load between the HCV-infected patients with and without HCC. In patients with HCC, the factors age, male sex, the proportion of patients with cirrhosis, and serum total bilirubin, serum ALT, and serum AFP levels were higher than in patients without HCC; serum albumin levels, prothrombin time, and platelet count were lower than in patients without HCC (Table 1)⇓ .
UGT1A7 Genetic Polymorphisms in the Japanese Population.
We examined the 12 reported polymorphisms harboring codons in exon 1 of the UGT1A7 gene in all 340 subjects (280 patients and 60 healthy individuals) by direct sequencing of the amplified 917-bp fragments. We found polymorphisms in only four codons: P11P, N129K, R131K, and W208R. The polymorphism in codon 11 was a silent change, and polymorphisms of the three other UGT1A7 codons defined the alleles (haplotypes) UGT1A7*2 (129K-131K-W208), UGT1A7*3 (129K-131K-208R), and UGT1A7*4 (N129-R131–208R). The alleles at the individual loci were all in Hardy–Weinberg equilibrium in both the patients and healthy individuals (P > 0.05).
Codons 129 and 131 were in complete linkage disequilibrium (D′ = 1). The 129–131/AAG(K)-AAA(K) polymorphisms (38%) were observed more frequently than was the 208CGG(R) polymorphism (24%). The proportion of patients with UGT1A7*4 was only ∼1.8% in our study. The UGT1A7*2/*4 and *4/*4 alleles were absent in our 340 individuals (Table 2)⇓ . The allele frequencies in the patients without HCC were similar to those of healthy individuals (χ2 = 4.925; P = 0.669).
Association between UGT1A7 Genetic Polymorphisms and HCC.
The distribution of the UGT1A7 alleles (genotype) and the IL-1β genotype with regard to the presence of HCC is shown in Table 2⇓ . The frequencies of the UGT1A7*2/*3 and *1/*3 alleles (genotypes) in patients with HCC (15 and 30%, respectively) were higher than those in patients without HCC (4 and 20%, respectively), with ORs of 6.08 (95% CI, 2.22–16.64) and 2.35 (95% CI, 1.26–4.35), respectively, compared with the *1/*1 alleles (genotypes). Interestingly, the frequency of the *2/*3 allele had a higher OR (>6) than the homozygous *2 or *3 allele.
The estimated frequencies of the UGT1A7*3 and *2 alleles (haplotypes) were higher in patients with HCC (28 and 19%, respectively) than in patients without HCC (17 and 14%, respectively), with ORs of 2.09 (95% CI, 1.38–3.18) and 1.71 (95% CI, 1.07–2.74), respectively, compared with the UGT1A7*1 allele, which suggests a positive correlation between these alleles and the presence of HCC (Table 2)⇓ .
The associations between the UGT1A7 polymorphisms and enzymatic activities are known. The UGT1A7*1 allele is related to higher activity, and the remaining alleles (*2, *3, and *4) are related to lower activity. On the basis of these associations, for clarity we designated the former allele as “H,” which represents higher UGT1A7 activity, and the latter alleles as “L,” which represents lower activity. The proportions of UGT1A7 L/L (genotype, homozygous for the lower-activity alleles *2, *3, and *4) and UGT1A7 H/L (genotype, heterozygous for higher- and lower-activity alleles) were higher in patients with HCC (25 and 45%, respectively) than in patients without HCC (15 and 39%, respectively) with ORs of 2.73 (95% CI, 1.40–5.35) and 1.80 (95% CI, 1.05–3.09) compared with UGT1A7 H/H (genotype, homozygous for the higher-activity allele *1). This suggests a positive correlation between the L alleles and the presence of HCC (Table 2)⇓ . Among patients with L/L, H/L, and H/H alleles, 57, 47, and 33%, respectively, had HCC. When we tested for trends, the L allele (haplotype) showed a slightly increased risk for the presence of HCC (P = 0.0019).
Association between IL-1β Genetic Polymorphisms and HCC.
Consistent with our previous report, polymorphisms in the IL-1β gene were associated with HCC. The IL-1β −31 and −511 loci were in complete linkage disequilibrium (D′ = 1). The IL-1β/−31T/T (i.e., −511C/C) genotype showed a significant association with the presence of HCC compared with the C/C genotype, consistent with our previous report (OR, 2.85; 95% CI, 1.37–5.94). The IL-1β/−31 T allele also showed a significant association with the presence of HCC (OR, 1.61; 95% CI, 1.15–2.26; Table 2⇓ ).
Although the IL-1β (2q14) and the UGT1A7 (2q37) genes are located on the same arm of chromosome 2, our analyses found no significant linkage disequilibrium between them (D′ < 0.2). We tried combining UGT1A7 and IL-1β polymorphisms with clinical information to improve the diagnosis of the risk for HCC in HCV-infected patients but could not find an SNP map that predicted high susceptibility to HCC. It therefore appears that the genetic polymorphisms of UGT1A7 and IL-1β are independently associated with the presence of HCC.
Factors Associated with Presence of HCC in HCV-Infected Patients.
The following factors were significantly associated with the presence of HCC according to univariate analyses: the UGT1A7 L/L alleles (P = 0.0225); the IL-1β/−31T/T(−511C/C) genotype (P = 0.0368); age >60 years (P = 0.007); male sex (P = 0.002); the presence of cirrhosis (P < 0.001); serum albumin <3.9 g/dl (P < 0.001); total bilirubin >0.7 mg/dl (P = 0.04); ALT >80 units/l (P = 0.043); AFP >20 ng/ml (P < 0.001); prothrombin time <70% (P < 0.001); and platelet count <12.5 × 104/μl (P < 0.001). Stepwise multivariate logistic regression analysis was performed with these 11 variables. Six variables (UGT1A7 L/L and IL-1β/−31T/T genotypes, presence of cirrhosis, age >60 years, male sex, and serum AFP >20 μg/l) were included in the final model with ORs of 2.33 (compared with H/H), 2.67 (compared with C/C), 4.20, 3.12, 3.09, and 2.90, with 95% CIs of 1.32–6.99, 1.22–4.76, 3.21–6.59, 1.42–6.37, 1.69–4.72, and 1.30–4.32, respectively (Table 3)⇓ .
In this study, we evaluated the relationships between polymorphisms of the UGT1A7 gene and the outcome of chronic HCV infection. The UGT1A7 L alleles, i.e., alleles that confer lower activity, were found to be associated with HCC. Our multivariate model confirmed the association between UGT1A7 L/L alleles and the presence of HCC. Genetic polymorphisms in the IL-1β gene were also included in the final model.
UGT1A7*3 was previously shown to be associated with HCC in a German population (13) . Although our study focused only on HCV-infected patients, our results were similar to those of the German study, which indicates that endogenous or environmental carcinogens facilitate the development of HCC in patients with HCV infection. One difference between our results and those of the German study is the frequency of the H/H (*1/*1) alleles, which was only 7% in German patients with HCC but was as high as 30% in Japanese patients with HCC. This may be explained in part by ethnic differences. It is also possible that factors other than UGT1A7 polymorphisms contribute to HCC in HCV-infected patients.
The UGTs are expressed in the gastrointestinal tract and in the liver in a tissue-specific manner (32) . UGT1A7 plays a critical role in the detoxification of carcinogens. Specifically, it was shown previously that polymorphisms in codons 129, 131, or 208 markedly decreased the carcinogen detoxification activity of UGT1A7 (23 , 33) . In our current study, patients harboring the UGT1A7 L/L alleles showed higher risk for HCC although the UGT1A7 gene is known to be expressed predominantly in the lung and gastrointestinal tract and not in the liver and biliary epithelium. We speculate that the UGT1A7 gene plays a critical role in the detoxification of hepatocarcinogens at the epithelia of the lung and gastrointestinal tract, which are thought to be entry sites for mutagens. Alternatively, there may be an unknown gene that is directly associated with HCC and that acts together with the UGT1A7 polymorphisms, or perhaps an unknown function of UGT1A7 affects the liver through actions at a different site.
We do not yet have a clear explanation for the fact that the UGT1A7 *2/*3 alleles had a higher OR for HCC than did *2/*2 or *3/*3. An evaluation of the functional impacts of the amino acid sequence differences may clarify one of the steps in the development of hepatocarcinogenesis.
To examine whether these UGT1A7 alleles were related to cirrhosis, we calculated the allele frequencies in the presence and absence of liver cirrhosis and found no statistically significant relationship between the two conditions. Thus, the association between the UGT1A7 polymorphisms and HCC was independent of cirrhosis. In our multivariate logistic regression model, the ORs of covariates other than polymorphisms in UGT1A7 and IL-1β, the presence of cirrhosis, age, gender, and AFP were similar to those reported in previous studies (29, 30, 31) . Our study population may have been biased toward patients with HCC or cirrhosis because many patients are referred to our hospital for treatment of HCC. However, our multivariate model included most of the previously reported risk factors for HCC with polymorphisms in the UGT1A7 and IL-1β genes. This suggests that our results can be generalized to the Japanese population.
Recent reports of genetic polymorphisms associated with HCC in patients with chronic HCV infection include the −31−511/T-C haplotype in the promoter region of the gene for the proinflammatory cytokine IL-1β in the Japanese population (11) , polymorphisms in the microsomal epoxide hydrolase gene in Italians (14) , and polymorphisms in the enzyme cytochrome P in Caucasians (12) . Although we failed to find a SNP map that predicted high susceptibility to HCC in this study, accumulating information on HCC-related SNPs may facilitate the identification of patients with high susceptibility to HCC in the future.
Despite the limitations of a case–control study, our analyses showed a prominent effect of the UGT1A7 lower-activity alleles on the risk of developing HCC. The uncertainty of the ORs that arises from the study design might be resolved in a subsequent controlled trial or in a large-scale screening association study. The UGT1A7 L/L alleles, as well as the IL-1β/−31T/T (−511C/C) genotype, might be used as markers of host factors associated with higher risk of HCC in Japanese patients with chronic HCV infection.
Grant support: Grants-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and Health Science Research Grants for Medical Frontier Strategy Research from the Ministry of Health, Labor, and Welfare, Japan.
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.
Requests for reprints: Naoya Kato, Department of Gastroenterology, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. Phone: 81-3-3815-5411, ext. 37198; Fax: 81-3-3814-0021; E-mail:
- Received August 6, 2003.
- Revision received December 29, 2003.
- Accepted December 29, 2003.