Effect of cytokine genotypes on the hepatitis B virus-hepatocellular carcinoma association†
See related editorial on pages 654–6, this issue.
In Southern Guangxi, China, chronic infection with the hepatitis B virus (HBV) acquired during the perinatal period from carrier mothers is a primary cause of hepatocellular carcinoma. However, only a minority of HBV carriers eventually develop hepatocellular carcinoma. The authors hypothesized that cytokine genotypes may be important codeterminants of the risk of HBV-related hepatocellular carcinoma.
The authors examined the correlation between polymorphisms in T-helper 1 (Th1) and Th2 cytokine genes among a group of 250 patients with incident hepatocellular carcinoma (cases) and a group of 250 hospital controls who were matched individually to the index case by age, gender, ethnicity, residence, and month of hospital admission in the city of Nanning, Guangxi, China.
Relative to the putative high-activity genotypes, each individual low-activity genotype of interferon γ, interleukin 12 (IL12), and IL18 was associated with a statistically nonsignificant increase (40–60%) in the risk of hepatocellular carcinoma. This risk increased with increasing numbers of low-activity Th1 genotypes after adjusting for potential confounders (2-sided P value for trend = 0.04). Conversely, individual Th2 (IL4, IL10) low-activity genotypes were associated with a statistically nonsignificant reduced risk of hepatocellular carcinoma. This risk decreased with increasing number of low-activity Th2 genotypes after adjusting for potential confounders (2-sided P value for trend = 0.01). Individuals who had the maximum number (i.e., 3) of low-activity Th1 genes and the minimum number (i.e., 0) of low-activity Th2 genes showed a relative risk of 20.0 (95% confidence interval, 1.7–235.0).
Diminished cell-mediated immune response, which is controlled genetically, appeared to be an important risk determinant of HBV-related hepatocellular carcinogenesis. Cancer 2005. © 2005 American Cancer Society.
The incidence of hepatocellular carcinoma (HCC) varies considerably worldwide.1 The highest incidence rates are found in East and Southeast Asia and in Sub-Saharan Africa, and the lowest rates are observed in North America and Western Europe. Southern Guangxi, China, possesses one of the highest rates of HCC in the world, and it is known that chronic infection with the hepatitis B virus (HBV), acquired perinatally from carrier mothers, is a primary cause.2 However, it also is known that only a minority of lifelong chronic carriers of HBV eventually will develop HCC,3 thus implicating the importance of cofactors (environmental or genetic) in HBV-related HCC. Dietary aflatoxin is known as one such cofactor.4
Cytokines are a family of proteins that mediate many of the responses of innate and adaptive immunity. T-helper 1 (Th1) cytokines are related to cell-mediated immune responses, which promote viral clearance.5 Th2 cytokines, conversely, can suppress the production of Th1 cytokines.6 Genetic mutations in certain cytokine genes can result in a reduction in the respective protein production.7 Therefore, an HBV-infected individual who has lower activity Th1 genotypes and higher activity Th2 genotypes may experience a heightened susceptibility to the development of HCC.
In this report, we describe the results of a case–control study that was conducted in Southern Guangxi, China. The objective of this study was to examine the possible roles of Th1 and Th2 genotypes in the development of HBV-related HCC.
MATERIALS AND METHODS
Newly diagnosed patients with HCC were identified from four major hospitals in the city of Nanning in Southern Guangxi, China. Participating hospitals were of comparable levels of quality in patient care and diagnosis. Only patients who were diagnosed during the period from September 1995 through September 1998, ages 20–64 years, who were residents of Nanning City or its neighboring townships were asked to participate in the study. We began the study in October, 1995, and closed enrollment in October, 1998, when 250 patients had been recruited into the study.
For each enrolled patient (case), we identified a consenting control patient among all patients who were admitted to the same hospital within 1 month of the index case's hospital admission who were without a history of cancer or clinical liver cirrhosis. The matching criteria were age (within 3 years), gender, ethnicity (Han, Zhuang, Yao, or other), and district (for residents of Nanning City) or township (for residents of neighboring townships) of residence. Permission to conduct this study was obtained from the Institutional Review Boards at the University of Southern California and the Guangxi Cancer Institute.
At the time of recruitment, all patients were interviewed in person by a trained interviewer by means of a structured questionnaire. The questionnaire solicited demographic information, lifetime use of tobacco and alcohol, and family history of liver disease, including cancer. An alcohol drinker was defined as someone who consumed alcoholic beverages at least once per week for ≥ 6 months. One drink was defined as 360 g of beer (12.6 g of ethanol), 103 g of wine (12.3 g of ethanol), or 30 g of spirit (12.9 g of ethanol).8 A tobacco smoker was defined as someone who had ever smoked on a daily basis. Smokers were asked at what age (in years) they began smoking on a daily basis, the average number of cigarettes they smoked per day, and the total years of cigarette smoking. Former smokers were asked about the number of years since smoking cessation.
At the end of the in-person interviews, all patients were asked to donate a blood sample. Upon consent, 5 mL of unheparinized blood plus 5 mL of anticoagulated whole blood with ethylenediamine tetraacetic acid were collected from each study participant. Serum was separated from the unheparinized blood sample within 2–3 hours after blood draw and was stored continuously at − 20 °C until analysis. All study participants had signed separate consent forms for interview and blood donation.
All study participants were tested for the presence of hepatitis B surface antigen (HBsAg) in serum using commercialized kits (AUSRIA, Abbott Laboratories, North Chicago, IL). The presence of antibodies to the hepatitis C virus (anti-HCV) in serum was tested using an enzyme-linked immunosorbent assay kit (Elisa kit, version 2.0; Ortho Diagnostic Systems) with confirmation of positive samples using a recombinant immunoblot assay (RIBA, version 2.0; Chiron, Emeryville, CA). All serum samples were tested blindly and were identified only by codes without regard to their case–control status.
Genomic DNA from whole blood was extracted using the QiaAmp blood kit according to the manufacturer's instructions (Qiagen). Polymerase chain reaction (PCR) amplifications were carried out in a volume of 12 μL in a Mastercycler gradient (Eppendorf). Final concentrations of reagents were as follows: 1 × Thermoprime polymerase buffer, 200 μM of each deoxynucleotide triphosphate, 1.5 mM MgCl2, 0.25 units ThermoprimePLUS DNA polymerase (all from ABgene), 1.0–2.5 μM specific forward primer, 1.0–2.5 μM specific reverse primer, (for allele-specific PCRs; 1 μM internal control primer mix), and ≈ 20 ng genomic DNA.
Allele-specific PCRs and PCR-restriction fragment length polymorphism (PCR-RFLPs) analyses were performed to determine nine single-nucleotide polymorphisms. Details of these genetic polymorphisms and their methods of measurements are provided in Table 1. In each round of genotyping, a negative (i.e., water) control was included. On a routine basis, 10% of the samples selected at random were genotyped twice. Samples that failed to be amplified and/or genotyped in the first run were tested again for up to three repetitions. All genotype measurements were determined independently by two laboratory personnel. In the case of discordant results, genotyping was repeated until a clear decision could be made.
Table 1. Descriptions of Genetic Polymorphisms of the Cytokine Genes Under Investigation and Respective Genotyping Methods
|IL2||− 330 T/Ga||5′-TATTCACATGTTCAGTGTAGTTCT (f); 5′-TGGATTCACACCCGATTACA (r)||54||PCR-RFLP (MaeI)||150 (T, 150; G, 124, 26)|
|IL4||− 589 C/Tb||5′-TAAACTTGGGAGAACATGGT (f); 5′-TGGGGAAAGATAGAGTAATA (r)||52||PCR-RFLP (AvalI)||195 (C, 177, 18; T, 195)|
|IL6||− 174G/Cc||5′-CTTAGCGCTAGCCTCAATGACGAC (f); 5′-GAGCGAGCGCAGGGGTGACTGACA (r)||63||PCR-RFLP (NlaIII)||429 (G, 260, 169; C, 260, 112, 57)|
|IL10||− 1082 G/Ad||5′-CTACTAAGGCTTCTTTGGGAG (f); 5′-ACTACTAAGGCTTCTTTGGGAA (f); 5′-CAGTGCCAACTGAGAATTTGG (r)||64, 59||Allele-specific||258|
| ||− 819 C/Td||5′-CCCTTGTACAGGTGATGTAAC (f); 5′-ACCCTTGTACAGGTGATGTAAT (f); 5′-AGGATGTGTTCCAGGCTCCT (r)|| || ||233|
|IL12 p40||+ 1188 3′ UTR A/Ce||5′-TTCTATCTGATTTGCTTTA (f); 5′-GGATGTATGGAATGTTTCA (r)||50, 45||PCR-RFLP (TaqI)||233 (C, 68, 165; A, 233)|
|IL18||− 137 G/Cf||5′-CCCCAACTTTTACGGAAGAAAAG (f); 5′-CCCCAACTTTTACGGAAGAAAAC (f); 5′-AGGAGGGCAAAATGCACTGG (r)||68,62||Allele-specific||261|
|INFG||+ 847 T/Ag||5′-TTCTTACAACACAAAATCAAATCT (f); 5′-TTCTTACAACACAAAATCAAATCA (f); 5′-TCAACAAAGCTGATACTCCA (r)||62,56||Allele-specific||264|
|TNF||− 307 G/Ad||5′-ATAGGTTTTGAGGGGCATGG (f); 5′-AATAGGTTTTGAGGGGCATGA (f); 5′-TCTCGGTTTCTTCTCCATCG (r)||64, 59||Allele-specific||184|
No study participants possessed the C allele of the interleukin-6 (IL6)-174 G/C polymorphism. Therefore, this polymorphism was not investigated further.
The distributions of cytokine genotypes were tested for Hardy–Weinberg equilibrium among study participants using the chi-square test. The case–control data were analyzed by standard matched-pair methods, including the conditional logistic regression analysis.9 The associations of HCC risk with various exposure and genetic indices of interest were measured by odds ratios (ORs) and their corresponding 95% confidence intervals (95% CIs) and 2-sided P values. When we examined cytokine genotypes in relation to HCC in study participants stratified by their HBsAg status, the matched case–control pairs were broken up, and unconditional logistic regression models were used to analyze the data. The matching factors age (years), gender, and ethnicity (Han, Zhuang, or Yao) were included as covariates in all unconditional logistic regression models. In all calculations of P values for linear trend, the actual values of the covariates were used in the regression models. We and others have shown consistently that only heavy drinking, but not moderate drinking (up to three drinks per day), is a risk factor for HCC.10 Therefore, regular drinkers were stratified into moderate drinkers (less than three drinks per day) and heavy drinkers (three or more drinks per day). Epidemiologic studies on tobacco and cancer have shown that those who smoke one or more packs per day may experience an observably greater level of risk relative to smokers of lesser amounts and that each additional decade of smoking may be associated with an observable increase in risk among smokers. Therefore, regular smokers were stratified into < 20 cigarettes per day smokers versus ≥ 20 cigarettes per day smokers and into smokers of < 20 years versus smokers of ≥ 20 years. Our reason for stratification by 2 decades instead of a single decade of smoking duration was the relatively small numbers of cases and controls among our participants in the latter categories, leading to unstable estimates of ORs.
Individual cytokine genotypes were classified as putatively high-activity and low-activity. We hypothesized that cytokine genotypes that promote viral clearance are associated with a reduced risk of HCC. Therefore, we predicted that lower-activity Th1 genotypes and higher-activity Th2 genotypes are associated with an increased risk of HCC.
Statistical analyses were carried out using the SAS software package (version 8.2; SAS Institute, Cary, NC) and Epilog for Windows (version 1.0; Epicenter Software, Pasadena, CA). Two-sided P values < 0.05 were considered statistically significant.
Among 250 patients with HCC, 40 patients (16%) were diagnosed histologically, 162 patients (65%) were diagnosed according to positive serum α fetoprotein levels with supportive imaging/clinical evidence, and 48 patients (19%) were diagnosed with imaging/clinical evidence only. Two hundred twenty patients (88%) were men. One hundred ninety-eight patients (79%) were ethnic Han, and the remaining patients were Zhuang (n = 51 patients) or Yao (n = 1 patient). The mean age (± standard deviation) at the time all patients were diagnosed with HCC was 49.3 ± 9.6 years. The level of education was comparable between cases and controls. The highest levels of education among cases were as follows: primary school or below, 14.0%; secondary school, 79.2%; and some college or above, 6.8%. The corresponding figures in control participants were 16.0%, 74.8%, and 9.2%, respectively.
Table 2 presents the distributions of known or suspected risk factors for HCC in the study for the case group and the control group. Consistent with our earlier findings in this population,4 positive HBsAg status was a strong risk factor, whereas positive anti-HCV status played a negligible role. Alcohol consumption and family history of HCC were additional independent risk factors for HCC in this population.
Table 2. Distributions of Known or Suspected Risk Factors for Hepatocellular Carcinoma in the Study Cases and Controls
|HBsAg|| || || || |
| Positive||205||35||43.50 (16.14–117.21)||41.41 (15.24–112.54)|
|Anti-HCV|| || || || |
| Positivec||9||3||3.00 (0.81–11.08)||0.49 (0.06–4.33)|
|Regular alcohol intake|
| Everd||92||40||3.60 (2.19–5.91)||3.22 (1.53–6.79)|
|No. of alcoholic drinks/day|
| < 3||60||31||3.04 (1.79–5.16)||2.79 (1.22–6.40)|
| ≥ 3||32||9||6.09 (2.64–14.05)||4.52 (1.36–15.09)|
| P for trend|| || ||< 0.001||0.008|
| Evere||113||86||1.63 (1.11–2.38)||0.83 (0.40–1.70)|
|No. of cigarettes smoked per day|
| < 20/day||34||29||1.47 (0.85–2.55)||0.80 (0.31–2.03)|
| ≥ 20/day||79||57||1.71 (1.12–2.60)||0.85 (0.37–1.97)|
| P for trend|| || ||0.04||0.72|
|No. of yrs of smoking|
| < 20 yrs||39||34||1.34 (0.74–2.42)||0.59 (0.20–1.73)|
| ≥ 20 yrs||74||52||1.82 (1.14–2.90)||1.01 (0.43–2.40)|
| P for trend|| || ||0.01||0.85|
|First-degree relative had hepatocellular carcinoma|
| Yes||30||2||15.00 (3.59–62.77)||2.48 (0.36–16.90)|
For the case group and the control group separately, the genotypic distributions of interferon γ (IFNG), IL12, IL4, IL10, IL12, and IL18 were all in Hardy–Weinberg equilibrium (all P values > 0.2). Only the genotypic distribution of TNF in control participants deviated from the Hardy–Weinberg equilibrium (P = 0.03).
Table 3 presents the distributions of Th1 genes in the study case and control groups. Relative to the putative high-activity genotypes, each individual low-activity genotype of IFNG, IL12, and IL18 was associated with a statistically nonsignificant (40–60%) increase in the risk of HCC. When we summed the low-activity genotypes across these 3 Th1 genes, the risk of HCC increased with increasing numbers of low-activity genotypes after adjusting for potential confounders (P for trend = 0.04).
Table 3. Distributions of Cytokine Genotypes in the Study Cases and Controls
|Th1 genes|| || || || |
| IFNG|| || || || |
| AT/AAc||94||86||1.15 (0.80–1.66)||1.63 (0.82–3.22)|
| IL12|| || || || |
| AC/CCc||193||178||1.46 (0.93–2.27)||1.37 (0.68–2.75)|
| IL18|| || || || |
| GC/CCc||77||58||1.48 (0.99–2.20)||1.60 (0.77–3.35)|
| Total low-activity Th1 genotypesd|
| 1||104||119||1.17 (0.67–2.04)||1.81 (0.69–4.73)|
| 2||94||82||1.58 (0.86–2.91)||1.86 (0.66–5.21)|
| 3||24||13||2.54 (1.08–5.99)||8.94 (1.59–50.23)|
| P for trend|| || ||0.01||0.04|
|Th2 genes|| || || || |
| IL4|| || || || |
| CCc||6||12||0.46 (0.16–1.31)||0.05 (0.01–0.40)|
| IL10|| || || || |
| AA and TT||130||115||1.00||1.00|
| GG/GA and/or CC/CTc||119||135||0.77 (0.54–1.10)||0.56 (0.30–1.05)|
| Total low-activity Th2 genotypese|
| 1||119||131||0.78 (0.54–1.11)||0.48 (0.25–0.92)|
| 2||3||8||0.31 (0.08–1.21)||0.10 (0.01–1.40)|
| P for trend|| || ||0.07||0.01|
The distributions of the other 2 Th1 genes (IL2 and TNF) were comparable between the case group and the control group. The distribution of the IL2 genotypes in the case group (TT, 45%; TG, 42%; and GG, 13%) was similar to that in the control group (TT, 40%; TG, 48%; and GG, 13%). The distribution of the TNF genotypes also were comparable between the case group (GG, 78%; GA, 20%; and AA, 2%) and the control group (GG, 82%; GA, 16%; and AA, 2%).
Table 3 also presents the distributions of Th2 genes in the study cases and controls. Because the 2 polymorphisms (− 1082 G/A and − 819 C/T) of the IL-10 gene studied were in linkage disequilibrium (all patients with the TT genotype of the − 819 C/T polymorphism had the AA genotype of the − 1082 G/T polymorphism), and the G allele of the − 1082 G/A polymorphism was rare in this study population (4%), we examined the association between the combined genotypes of the 2 IL10 genetic polymorphisms and HCC risk. Both IL4 and IL10 showed reduced risk of HCC in association with their respective, putatively low-activity genotypes. When the numbers of low-activity genotypes were summed across these 2 Th2 genes, the risk of HCC decreased with increasing numbers of low-activity Th2 genotypes after adjustment for potential confounders (P for trend = 0.01).
Table 4 shows the combined effect of Th1 and Th2 genotypes on the risk of HCC. Individuals with the maximum number (i.e., 3) of putatively low-activity Th1 genes and the minimum number (i.e., zero) of low-activity Th2 genes exhibited an OR of 20 (95% CI, 1.7–235.0) relative to individuals who had 0 low-activity Th1 genes and at least 1 low-activity Th2 gene. The interaction effect between the Th1 and Th2 genotypes on the risk of HCC was not statistically significant (P = 0.55).
Table 4. The Combined Effect of T-Helper 1 and T-Helper 2 Genotypes on the Risk of Hepatocellular Carcinoma
|1||55/58||1.75 (0.48–6.46)||49/61||2.44 (0.75–7.98)|
|2||41/53||1.16 (0.30–4.49)||53/29||5.13 (1.21–21.81)|
|3||10/7||7.55 (0.79–72.53)||14/6||19.97 (1.70–234.97)|
We further examined the associations of the Th1 and Th2 genotypes with HCC in HBsAg-positive patients versus HBsAg-negative patients separately. The associations were consistently stronger in HBsAg-positive patients compared with HBsAg-negative patients, but the difference was not statistically significant (P > 0.4) (Table 5).
Table 5. The Effects of T-Helper 1 and 2 Genotypes on the Risk of Hepatocellular Carcinoma in Hepatitis B Surface Antigen (HBsAg)-Positive versus HBsAg-Negative Individuals Separately
|Total no. of low-activity Th1 genotypesb|
| 1–2||38/175||1.65 (0.46–5.92)||160/26||2.13 (0.87–5.23)|
| 3||4/13||2.15 (0.37–12.46)||20/0||— (1.84)c|
| P for trend|| ||0.38|| ||0.01|
|Total no. of low-activity Th2 genotypesb|
| 1–2||20/115||0.69 (0.35–1.36)||102/24||0.44 (0.20–0.95)|
Forty-eight patients in the case group were diagnosed based on imaging/clinical evidence only. We repeated all analyses after excluding these 48 patients and their matched patients in the control group. Results from this subset were comparable to those based on all patients. The OR for HCC was 11 (95% CI, 1.1–108.0) for patients who carried 3 (maximum number) low-activity Th1 genotypes compared with patients who carried 0 low-activity Th1 genotypes. For patients who carried the maximum number (i.e., 3) of low-activity Th1 genotypes but 0 low-activity Th2 genotypes, the OR was 63 (95% CI, 1.7–2297.0) relative to patients who had ≥ 1 low-activity Th2 genotypes but 0 low-activity Th1 genotypes.
To our knowledge, this is the first study examining the role of genetic polymorphisms of major Th1 and Th2 cytokine genes in the development of HCC. The liver is rich in lymphocytes with various cytotoxic activities and cytokine secretion patterns, which facilitate the rapid elimination of virally infected cells.5 Our data suggest that HBV-infected individuals who carry the low-activity genotypes of major Th1 cytokines and/or the high-activity genotypes of major Th2 cytokines may be at extraordinarily high risk for HCC.
IFN-γ is one of the major cytokines secreted by Th1 cells and is crucial for antiviral immune response.7 IFN-γ exerts its pleiotropic effects by transcriptional regulation of numerous genes whose products play distinct roles in antigen presentation pathways.11 It directs an antiviral response by the up-regulation of major histocompatibility complex antigens on the hepatocytes12 and breakdown of viral RNA.13 A T/A polymorphism of the IFNG gene at position + 874 located next to a polymorphic CA repeat and coinciding with a putative neuronal factor κB-binding site has been identified.14 It was shown that the T allele was in complete linkage with 12 CA repeats. The TT homozygosity was associated with elevated IFN-γ production,14 whereas AA homozygotes showed reduced production of IFN-γ by peripheral blood mononuclear cells in in vitro stimulation experiments.15 Furthermore, AA homozygotes showed an increased risk of developing tuberculosis.15 Therefore, the overall data support the notion of impaired antiviral immune response in individuals who have the IFNG-AA genotype. It is noteworthy that the Chinese population may have a greater prevalence of the AA genotype compared with other ethnic groups.16
IL-12 and IL-18 are key regulatory cytokines for the secretion of IFN-γ by Th1 cells and natural killer cells. IL-12 is a heterodimer composed of a light chain (p35) and a heavy chain (p40) and is produced in hepatic cells chronically infected with HBV.17 At the onset of inflammation, IL-12 promotes the development of Th1 cells and IFN-γ production and inhibits the differentiation of Th2 cells.18 An A/C polymorphism at position + 1188 in the 3′-untranslated region of the gene encoding the p40 of the IL-12 cytokine has been identified.19 The A allele was associated with significantly elevated expression of IL-12 in Epstein–Barr virus-transformed human cell lines.20 Similar results were observed in peripheral lymphocytes derived from individuals with the A allele of the IL12 genotype.21 It also was found that IL-12 plays an important role in viral clearance of HBV in vivo. Clearance of HBV and seroconversion in chronic HBV carriers after treatment with IFN-α depended on a substantial increase in IL-12 production.22 Dendritic cells from patients with chronic HBV infection produced significantly lower levels of IL-12 in cultures compared with the levels produced in normal control participants,23 and it was shown that IL-12 inhibited HBV replication in transgenic mice experiments.24
IL-18 is a pleiotropic cytokine with an important function in amplification of Th1 and natural killer cell response to viral infections.25 It induces the production of IFN-γ by T and B lymphocytes and natural killer cells.25 In the presence of IL-12, IL-18 exerts a synergistic effect on IFN-γ production and induces important antiviral activities.24, 26 It was found that IL-18 inhibited HBV replication in the livers of transgenic mice, and the same experiment showed a strong synergistic, antiviral effect of IL-12 and IL-18.27 IL-18 immunoreactivity is lower in HCC cells than in normal liver cells.28 The G allele of the − 137 promoter polymorphism of the IL18 was associated with elevated IL-18 expression and increased levels of IFN-γ mRNA29 in in vitro experiments.
Polymorphisms in the 5′ region of TNF (− 307 G/A) and IL2 (− 330 T/G), for which a functional relevance has been postulated,30, 31 did not appear to modify HCC risk in our Chinese population. TNF-α is an important proinflammatory cytokine and a critical mediator of host defense against infection.32 Although TNF-α has antiviral activities that often are synergistic with IFN-γ, very little is known regarding the direct, intracellular, antiviral mechanisms activated by this cytokine.5 Consistent with our findings, a recent Japanese study reported a null association between the TNF-α genotype and HCV-related HCC.33 IL-2 is secreted by Th1 cells and is involved in T-cell and B-cell proliferation. The observed null association of the IL2 polymorphism and HCC risk may be related to the up-regulatory effect of IL-2 on both IFN-γ and IL-4.34
IL-4 and IL-10 are secreted by Th2 cells, and both have overall suppressive effects on the generation of Th1 response.35, 36 IL-4 is the hallmark Th2 cytokine and antagonizes a number of IFN-γ-induced functions on Th1 differentiation and stability.35 The IL4 gene is located within a cytokine gene cluster in the chromosomal region, 5q22–q32.37 In vitro and in vivo experimental data demonstrated that the T allele of the polymorphism at position − 589 (referred as the IL4 C-590T polymorphism in other reports), which has been found to be in linkage disequilibrium with − 33T,38 is associated with an increased expression of IL-4.39 In addition, the T allele was associated with severe respiratory syncytial virus infection,40 suggesting that individuals with the T allele of the IL4 genetic polymorphism may have diminished immune response to viral infection.
IL-10 is a multifunctional cytokine that is expressed by a number of different immune cells, and it has a major role in striking a balance between pathology and protection from infectious diseases.41 IL-10 is an inhibitor of activated macrophages, which play a crucial role in the homeostatic control of innate immune reactions and cell-mediated immunity. IL-10 potently down-regulates the production of IFN-γ by Th1 cells and inhibits the immune response to viral infection.41 Abnormally elevated expression of IL-10 repeatedly has been reported in individuals with chronic HBV infection.42, 43 It has been shown that interindividual differences in IL-10 production are determined genetically.44 Several genetic polymorphisms of the gene in the promoter region (− 1082 G/A, − 819 T/C, and − 592 A/C) have been identified.45 We studied 2 polymorphisms of the IL10 promoter: − 1082 G/A and − 819 T/C. We did not determine the − 592 A/C genotype, because this polymorphism is in virtually complete linkage disequilibrium with the − 819 T/C polymorphism.46 The − 1082 G/A polymorphism is rare in the Chinese population; the G allele frequency was 4% in the current study population, consistent with previous findings in other Asian populations.47 When we examined the combined genotypes of the 2 IL10 genetic polymorphisms, a decreased risk of HCC was associated with the presence of both the − 1082 “G” allele and the − 819 “C” allele, the genotype that is related to reduced plasma IL-10 levels.48
It was found that the − 174 G/C IL6 polymorphism, which appeared to be functional,49 was rare in the Chinese population.50 Our current results are consistent with that earlier finding. We failed to detect any C alleles in our study participants.
The limitations of the current study included a hospital-based study design and a relatively small sample size. The expected study power to detect the observed effect of IFNG genotype on the risk of HCC was 0.76. The corresponding values for the IL12 and IL18 genotypes were 0.34 and 0.65, respectively. Therefore, it is not surprising that none of the three individual genotype-HCC associations noted in this study reached statistical significance. Conversely, the statistical power in this study to detect the combined effect of all three Th1 genotypes (i.e., summation of low-activity genotypes with scores ranging from 0 to 3) was high (100%); indeed, a statistically significant, gene-dose dependent association was observed. Nonetheless, it should be emphasized that this was a hypothesis-generating study that will require confirmation from another study with a larger sample size.
The results of the current study suggest a major role for genetically regulated immune response to hepatitis B infection in determining an infected individual's level of HCC risk. The difference in risk between the 2 extreme categories of cytokine genotypes can be as great as 20-fold. Our results, if confirmed, may lead to better identification of HBV carriers who are at very high risk of HCC and to targeting aggressive antiviral therapy in this subgroup of infected individuals. Conversely, it may be argued that antiviral therapies with relatively severe side effects should be discouraged among HBV carriers who have the most favorable combined cytokine genotypes, given these individuals' considerably lower risk for HCC.
The authors thank Ina Koegel and Aljoscha Schultze for technical assistance with genotyping.