Learn more: PMC Disclaimer | PMC Copyright Notice
GST polymorphisms are associated with hepatocellular carcinoma risk in Chinese population
Abstract
AIM: To investigate the association between GSTM1 and GSTT1 polymorphisms and the risk of hepatocellular carcinoma (HCC) in Chinese population.
METHODS: Literature databases including PubMed, ISI web of science and other databases were searched. Pooled odds ratio (OR) and 95% CI were calculated using random- or fixed- effects model. Subgroup analysis and sensitivity analysis were also performed.
RESULTS: Nineteen studies of GSTM1 (2660 cases and 4017 controls) and 16 studies of GSTT1 (2410 cases and 3669 controls) were included. The GSTM1/GSTT1 null genotypes were associated with increased risk of HCC in Chinese population (for GSTM1, OR = 1.487, 95% CI: 1.159 to 1.908, P = 0.002; for GSTT1, OR = 1.510, 95% CI: 1.236 to 1.845, P = 0.000). No publication bias was detected. In subgroup analysis, glutathione S-transferases polymorphisms were significantly associated with HCC risk among the subjects living in high-incidence areas, but not among the subjects living in low-incidence areas.
CONCLUSION: The present meta-analysis suggests that GSTM1/GSTT1 null genotypes are associated with increased risk of HCC in Chinese population.
INTRODUCTION
Liver cancer is one of the most common types of cancer and one of the most common causes of cancer-related death[1]. The death rates have increased for both men and women with liver cancer over the past two decades[2]. The incidence and mortality rates of liver cancer vary considerably among racial and ethnic groups[3]. Asians, particularly Chinese, have a high risk of developing liver cancer[4]. About 80%-90% of all cases of primary liver cancer are hepatocellular carcinoma (HCC).
The pathogenesis of HCC may have a genetic and environmental basis[5,6]. Epidemiological studies have shown that HCC is associated with many environmental factors, including alcoholism, chronic infection with hepatitis B virus (HBV), hepatitis C virus (HCV), and dietary exposure to aflatoxin B1 (AFB1)[7]. These hepatocarcinogens result in increased generation of reactive oxygen species and free radicals that cause liver damage and repair. Thus the accumulated multistage genetic mutations may lead to liver carcinogenesis[8-11].
During liver carcinogenesis, cellular defense mechanisms can alleviate the effects of oxidative stress and exogenous toxins. One of the essential antioxidant is the reducing compound glutathione. Reduced glutathione can be conjugated to various xenobiotics and endobiotics by glutathione S-transferases (GST), a superfamily of cytosolic soluble detoxification enzymes. In humans, cytosolic soluble GSTs are encoded by seven distinct genes: Alpha, Mu, Omega, Pi, Sigma, Theta and Zeta[12-14]. GSTs play an important role in cellular protection against oxidative stress and exogenous toxins. Homozygous deletion of GST genes (null genotype) results in decreased enzyme activity, which will impede detoxification and may ultimately increase the risk of many diseases, including HCC[13-16]. GSTM1 and GSTT1 have been the most extensively studied GST genes. While many studies have investigated the relationship between GSTM1 and GSTT1 polymorphisms and HCC risk, so far the results have been inconsistent.
Recently, a meta-analysis result did not suggest a statistically significant increased risk of HCC with the GST null genotypes[17]. However, a large number of studies were not reported in that meta-analysis (which included only 9 studies of GSTM1 and 8 studies of GSTT1 in Chinese population)[17]. Since the ethnic background and the environmental exposures may vary greatly across populations in different geographic regions, the conclusion could not be drawn[17]. During the preparation of this paper, another meta-analysis was published showing that the null genotypes of GSTM1 and GSTT1 are both associated with an increased HCC risk[18]. However, these investigators omitted some important data and introduced some incorrect information[18]. For example, Indians are mainly of Indo-European and Dravidian ancestries, which should be distinguished from East Asian population[18]. Many studies, including our previous studies[16,19-22], have reported on the effects of ethnic differences on genetic predisposition to human diseases. In addition, it is important to note that the allele frequencies and the genotype distributions for GST genes differed significantly across ethnic groups[23]. For example, the frequency of GSTM1 null genotype is about 0.53 in Caucasians/Asians and about 0.28 in Africans; and the frequency of GSTT1 null genotype is about 0.20 in Caucasians and about 0.52 in Asians[23]. Therefore, heterogeneity was introduced in that meta-analysis, which may not accurately assess the effects of GSTM1 and GSTT1 null genotypes on the risk of HCC[18]. In this study, we reinvestigated the relationship between GSTM1/GSTT1 polymorphisms and the risk of HCC. We focused on the association between GST polymorphisms and the risk of HCC in Chinese population, because Chinese are at a much greater risk of developing HCC compared to other ethnic groups. A total of 19 studies of GSTM1 (2660 cases and 4017 controls) and 16 studies of GSTT1 (2410 cases and 3669 controls) were included. This meta-analysis has a much greater number of subjects and thus a much greater statistical power; therefore, it may define the effects of GST gene polymorphisms on the risk of HCC more precisely.
MATERIALS AND METHODS
Literature and search strategy
We searched the literature databases including PubMed (1950 to 2010), ISI web of science (1975 to 2010), Embase (1966 to 2010), Chinese Biomedical Database (1978 to 2010), China National Knowledge Infrastructure (1979 to 2010, in Chinese) and Wanfang Data (1982 to 2010, in Chinese).
The search strategy to identify all possible studies involved use of combinations of the following key words: (“glutathione S-transferase” or “GST” or “GSTM1” or “GSTT1”) and (“hepatocellular carcinoma” or “liver cancer” or “HCC”) and (“China” or “Chinese”). The reference lists of reviews and retrieved articles were also searched. Supplementary data were searched for missing data points. If more than one article were published using the same case series, only the study with largest sample size was selected. The literature search was updated on Oct. 20th, 2010.
Inclusion criteria and data extraction
The studies included in the meta-analysis must meet all the following inclusion criteria: (1) evaluating the association between GSTM1 or GSTT1 null genotypes and HCC risk; (2) case-control design; (3) in Chinese population; and (4) sufficient data for calculation of odds ratio (OR) with CI. The following information was extracted from each study: (1) name of the first author; (2) year of publication; (3) language of publication; (4) residence area of the subjects; (5) source of control subjects; (6) numbers of cases and controls; (7) numbers of null genotypes for GSTM1 and GSTT1 in cases and controls; and (8) OR and 95% CI. The authors independently assessed the articles for compliance with the inclusion/exclusion criteria, resolved disagreements and reached a consistent decision.
Statistical analysis
The association between GSTM1/GSTT1 polymorphisms and the risk of HCC was estimated by calculating pooled OR and 95% CI. The significance of the pooled OR was determined by Z test with P < 0.05 considered statistically significant. Q test was performed to evaluate whether the variation was due to heterogeneity or by chance. A random- (DerSimonian-Laird method[24]) or fixed- (Mantel-Haenszel method[25]) effects model was used to calculate pooled effect estimates in the presence (P ≤ 0.10) or absence (P > 0.10) of heterogeneity, respectively. Begg’s funnel plot, a scatter plot of effect against a measure of study size, was generated as a visual aid to detecting bias or systematic heterogeneity[26]. An asymmetric funnel plot indicates a relationship between effect and study size, which suggests the possibility of either publication bias or a systematic difference between smaller and larger studies (“small study effects”). Publication bias was assessed by Begg’s test and Egger’s test[27] with P < 0.05 considered statistically significant. Subgroup analyses were performed to examine the effect of heterogeneity on meta-analysis results. The following subgroup comparisons were analyzed: residence area of the subjects (high-incidence area vs low-incidence area), number of cases (< 100 vs ≥ 100), and source of controls (population-based vs hospital-based). To evaluate the stability of results, sensitivity analysis was performed by removing one study at a time and calculating the overall homogeneity and effect size. Data analysis was performed using STATA version 10 (StataCorp LP, College Station, Texas, USA).
RESULTS
Characteristics of the studies
The literature search identified a total of 137 potential relevant papers. The full text articles were retrieved and carefully reviewed to assess the eligibility according to the inclusion criteria. Forty-three papers met the inclusion criteria[28-70]. However, 23 papers were excluded because they were earlier or smaller reports from the same groups[29,30,32-37,43-46,48-50,53,55-58,61,62,64]. Nineteen studies of GSTM1 (2660 cases and 4017 controls) and 16 studies of GSTT1 (2410 cases and 3669 controls) were included in the meta-analysis, respectively. Most of the cases and controls included in this meta-analysis were HBV carriers. The characteristics of the included studies are listed in Tables Tables11 and and22.
Table 1
Characteristics of studies included in the meta-analysis of GSTM1
Ref. | Language of publication | Residence area of subjects | Case | Control | Source of control | OR (95% CI) | Earlier or smaller reports | ||||
Null | Positive | Total | Null | Positive | Total | ||||||
Dong et al[28], 1997 | Chinese | Jiangsu, Guangxi and Hebei | 62 | 48 | 110 | 50 | 62 | 112 | Population | 1.602 (0.943-2.721) | [29,30] |
Yu et al[31], 1999 | English | Taiwan | 42 | 42 | 84 | 216 | 159 | 375 | Hospital | 0.736 (0.458-1.183) | [32-37] |
Wu et al[38], 2000 | Chinese | Hu’nan | 38 | 16 | 54 | 62 | 74 | 136 | Population | 2.835 (1.444-5.565) | NA |
Sun et al[39], 2001 | English | Taiwan | 26 | 43 | 69 | 77 | 51 | 128 | Population | 0.400 (0.219-0.731) | NA |
Zhu et al[40], 2001 | Chinese | Guangdong | 34 | 18 | 52 | 41 | 59 | 100 | Hospital | 2.718 (1.354-5.455) | NA |
Chen et al[41], 2002 | English | Taiwan | 60 | 41 | 101 | 19 | 16 | 35 | Hospital | 1.232 (0.568-2.674) | NA |
Liu et al[42], 2002 | Chinese | Jiangsu | 56 | 28 | 84 | 69 | 75 | 144 | Population | 2.174 (1.243-3.803) | [43-46] |
McGlynn et al[47], 2003 | English | Jiangsu | NA | NA | 231 | NA | NA | 256 | Population | 0.830 (0.570-1.210) | [48-50] |
Li et al[51], 2004 | Chinese | Jiangsu | 122 | 85 | 207 | 118 | 89 | 207 | Population | 1.083 (0.733-1.600) | NA |
Chen et al[52], 2005 | English | Taiwan | 322 | 255 | 577 | 231 | 158 | 389 | Population | 0.864 (0.666-1.121) | [53] |
Deng et al[54], 2005 | English | Guangxi | 117 | 64 | 181 | 172 | 188 | 360 | Hospital | 1.998 (1.383-2.888) | [55-58] |
Guo et al[59], 2005 | Chinese | Henan | 67 | 28 | 95 | 52 | 51 | 103 | Population | 2.347 (1.306-4.218) | NA |
He et al[60], 2005 | Chinese | Guangxi | 68 | 37 | 105 | 77 | 74 | 151 | Hospital | 1.766 (1.059-2.947) | [61,62] |
Long et al[63], 2005 | Chinese | Guangxi | 92 | 48 | 140 | 254 | 282 | 536 | Hospital | 2.128 (1.444-3.137) | [64] |
Ma et al[65], 2005 | Chinese | Guangxi | 37 | 25 | 62 | 29 | 44 | 73 | Population | 2.246 (1.125-4.481) | NA |
Zhang et al[66], 2005 | Chinese | Hubei | 37 | 23 | 60 | 28 | 45 | 73 | Hospital | 2.585 (1.281-5.219) | NA |
Zhu et al[67], 2005 | Chinese | Zhejiang | 56 | 35 | 91 | 61 | 69 | 130 | Hospital | 1.810 (1.049-3.121) | NA |
Long et al[68], 2006 | English | Guangxi | 179 | 78 | 257 | 312 | 337 | 649 | Hospital | 2.479 (1.823-3.370) | NA |
Yang et al[69], 2009 | Chinese | Guangxi | 59 | 41 | 100 | 41 | 19 | 60 | Hospital | 0.667 (0.340-1.309) | NA |
NA: Not available; OR: Odds ratio.
Table 2
Characteristics of studies included in the meta-analysis of GSTT1
Ref. | Language of publication | Residence area of subjects | Case | Control | Source of control | OR (95% CI) | Earlier or smaller reports | ||||
Null | Positive | Total | Null | Positive | Total | ||||||
Dong et al[28], 1997 | Chinese | Jiangsu, Guangxi and Hebei | 63 | 47 | 110 | 42 | 70 | 112 | Population | 2.234 (1.305-3.825) | [29,30] |
Yu et al[31], 1999 | English | Taiwan | 41 | 42 | 83 | 181 | 194 | 375 | Hospital | 1.046 (0.650-1.683) | [32-37] |
Sun et al[39], 2001 | English | Taiwan | 30 | 37 | 67 | 77 | 51 | 128 | Population | 0.537 (0.295-0.976) | NA |
Liu et al[42], 2002 | Chinese | Jiangsu | 34 | 50 | 84 | 36 | 108 | 144 | Population | 2.040 (1.146-3.630) | [43-46] |
McGlynn et al[47], 2003 | English | Jiangsu | NA | NA | 231 | NA | NA | 256 | Population | 0.880 (0.590-1.310) | [48-50] |
Liu et al[70], 2003 | Chinese | Guangxi | 28 | 23 | 51 | 18 | 35 | 53 | Population | 2.367 (1.072-5.227) | NA |
Li et al[51], 2004 | Chinese | Jiangsu | 108 | 99 | 207 | 97 | 110 | 207 | Population | 1.237 (0.841-1.820) | NA |
Chen et al[52], 2005 | English | Taiwan | 298 | 279 | 577 | 199 | 190 | 389 | Population | 1.020 (0.788-1.319) | [53] |
Deng et al[54], 2005 | English | Guangxi | 108 | 73 | 181 | 154 | 206 | 360 | Hospital | 1.979 (1.377-2.845) | [55-58] |
Guo et al[59], 2005 | Chinese | Henan | 58 | 37 | 95 | 45 | 58 | 103 | Population | 2.020 (1.146-3.562) | NA |
He et al[60], 2005 | Chinese | Guangxi | 43 | 62 | 105 | 50 | 101 | 151 | Hospital | 1.401 (0.836-2.347) | [61,62] |
Long et al[63], 2005 | Chinese | Guangxi | 82 | 58 | 140 | 234 | 302 | 536 | Hospital | 1.825 (1.251-2.660) | [64] |
Ma et al[65], 2005 | Chinese | Guangxi | 35 | 27 | 62 | 21 | 52 | 73 | Population | 3.210 (1.573-6.551) | NA |
Zhang et al[66], 2005 | Chinese | Hubei | 38 | 22 | 60 | 34 | 39 | 73 | Hospital | 1.981 (0.986-3.982) | NA |
Long et al[68], 2006 | English | Guangxi | 146 | 111 | 257 | 297 | 352 | 649 | Hospital | 1.559 (1.165-2.086) | NA |
Yang et al[69], 2009 | Chinese | Guangxi | 33 | 67 | 100 | 11 | 49 | 60 | Hospital | 2.194 (1.010-4.765) | NA |
NA: Not available; OR: Odds ratio.
Meta-analysis results of the association between GSTM1 polymorphisms and HCC
Nineteen studies of GSTM1, including 2660 cases and 4017 controls, were included in the meta-analysis. The relative frequency of GSTM1 null genotype among control groups ranged from 0.384 to 0.683 in Chinese population. Using a random-effects model, the overall meta-analysis result showed that there was a statistically significant association between GSTM1 null genotype and HCC risk in Chinese population (OR = 1.487, 95% CI: 1.159 to 1.908, P = 0.002). The forest plot is shown in Figure Figure11.
Meta-analysis results of the association between GSTT1 polymorphisms and HCC
Sixteen studies of GSTT1, including 2410 cases and 3669 controls, were included in the meta-analysis. The relative frequency of GSTT1 null genotype among control groups ranged from 0.183 to 0.602 in Chinese population. Using a random-effects model, the overall meta-analysis result showed that there was a statistically significant association between GSTT1 null genotype and HCC risk in Chinese population (OR = 1.510, 95% CI: 1.236 to 1.845, P = 0.000). The forest plot is shown in Figure Figure22.
Subgroup analysis
To examine the effect of heterogeneity between studies on meta-analysis results, we conducted subgroup analyses stratified by the following: residence area of the subjects (high-incidence area vs low-incidence area), number of cases (< 100 vs ≥ 100), and source of controls (population-based vs hospital-based). GST polymorphisms were significantly associated with HCC risk among the subjects living in high-incidence areas (Jiangsu, Zhejiang, Guangxi and Guangdong provinces), but not among the studies living in low-incidence areas. The result of subgroup analysis is shown in Tables Tables33 and and44.
Table 3
Subgroup analysis of the association between GSTM1 polymorphism and hepatocellular carcinoma risk
Group | No. of studies (cases/controls) | Statistical method | OR (95% CI) | P |
All studies | 19 (2660/4017) | Random | 1.487 (1.159-1.908) | 0.002 |
Residence area of the subjects | ||||
High-incidence area | 11 (1510/2666) | Random | 1.659 (1.264-2.177) | 0.000 |
Low-incidence area | 7 (1040/1239) | Random | 1.235 (0.753-2.026) | 0.402 |
Mixed areas | 1 (110/112) | - | 1.602 (0.943-2.721) | 0.081 |
No. of cases | ||||
< 100 | 9 (651/1262) | Random | 1.676 (1.061-2.649) | 0.027 |
≥ 100 | 10 (2009/2755) | Random | 1.365 (1.005-1.853) | 0.046 |
Source of controls | ||||
Population-based | 9 (1489/1548) | Random | 1.316 (0.915-1.892) | 0.139 |
Hospital-based | 10 (1171/2469) | Random | 1.675 (1.251-2.243) | 0.001 |
OR: Odds ratio.
Table 4
Subgroup analysis of the association between GSTT1 polymorphism and hepatocellular carcinoma risk
Group | No. of studies (cases/controls) | Statistical method | OR (95% CI) | P |
All studies | 16 (2410/3669) | Random | 1.510 (1.236-1.845) | 0.000 |
Residence area of the subjects | ||||
High-incidence area | 10 (1418/2489) | Random | 1.641 (1.328-2.027) | 0.000 |
Low-incidence area | 5 (882/1068) | Random | 1.152 (0.777-1.707) | 0.483 |
Mixed areas | 1 (110/112) | - | 2.234 (1.305-3.825) | 0.003 |
Number of cases | ||||
< 100 | 7 (502/949) | Random | 1.617 (1.035-2.528) | 0.035 |
≥ 100 | 9 (1908/2720) | Random | 1.457 (1.173-1.810) | 0.001 |
Source of controls | ||||
Population-based | 9 (1484/1465) | Random | 1.441 (1.039-1.997) | 0.028 |
Hospital-based | 7 (926/2204) | Fixed | 1.635 (1.391-1.921) | 0.000 |
OR: Odds ratio.
Sensitivity analysis
Sensitivity analysis was performed by excluding each study at a time. The analysis confirmed the stability of the association between GSTM1 and GSTT1 polymorphisms and HCC risk (data not shown).
Potential publication bias
Begg’s funnel plots were generated to assess potential publication bias (Figure (Figure33 for GSTM1 and Figure Figure44 for GSTT1). No publication bias was detected (Egger’s test, P = 0.542 for GSTM1 and P = 0.136 for GSTT1; Begg’s test, P = 0.677 for GSTM1 and P = 0.299 for GSTT1).
Funnel plot of the meta-analysis of the association between GSTM1 polymorphism and hepatocellular carcinoma risk.
DISCUSSION
GST polymorphisms are implicated in the development of HCC. In the present study, we investigated the relationship between GSTM1 and GSTT1 polymorphisms and the risk of HCC in Chinese population. To minimize language bias, all available studies published in both English and Chinese languages were assessed, including a total of 19 studies of GSTM1 (2660 cases and 4017 controls) and 16 studies of GSTT1 (2410 cases and 3669 controls). The present meta-analysis has a much greater number of subjects and thus a much greater statistical power; therefore, it can define the effect of GST genes polymorphisms on HCC risk more precisely. The pooled analysis of two genes produced similar risk estimates (for GSTM1, OR = 1.487, 95% CI: 1.159 to 1.908, P = 0.002; for GSTT1, OR = 1.510, 95% CI: 1.236 to 1.845, P = 0.000), suggesting that both GSTM1 and GSTT1 null genotypes are associated with an increased risk of HCC in Chinese population.
The difference between our meta-analysis and the previous metaanalysis may be due to language bias introduced in the previous metaanalysis, which was based primarily on reports in English[17]. Therefore, all available studies published in both English and Chinese languages were assessed in the current meta-analysis. Another recent meta-analysis suggested that null genotypes of GSTM1 and GSTT1 were both associated with increased risk of HCC[18]. However, these investigators omitted some data and included some studies with incorrect information[52,53,59,65,69]. Therefore, the effects of GSTM1 and GSTT1 null genotypes on HCC was not assessed accurately[18]. Considering that Chinese people are at a much greater risk of developing HCC, we focused on the relationship between GST polymorphisms and HCC risk in Chinese population, thereby minimizing ethnic/racial differences. In the subgroup analysis, GST polymorphisms were significantly associated with HCC risk among the subjects living in high-incidence areas, but not among the subjects living in low-incidence areas. This result suggests that the effect of GST polymorphisms on HCC risk may be enhanced by environmental risk factors, such as dietary exposure to AFB1. Since sample size could influence the results, we also performed subgroup analysis stratified by sample size of cases. The results showed that the studies with either large (number of cases ≥ 100) or small sample size (number of cases < 100) had similar risk estimates, indicating that small study effects may not exist in this meta-analysis. In addition, we generated Begg’s funnel plots and found no publication bias among the studies included in this meta-analysis (Begg’s test and Egger’s test, P > 0.1).
The current meta-analysis has vital advantages compared to other studies; however, it does have some limitations. First, the present meta-analysis was based on unadjusted effect estimates and confidence intervals due to insufficient data available for most of the studies. Although the cases and controls were matched on age, sex and residence in all studies, these confounding factors might slightly modify the effect estimates. Second, the effect of gene-environment interactions was not studied in this meta-analysis. Alcoholism, HBV/HCV infections, and dietary exposure to AFB1 may be environmental risk factors that modify the effect estimates. Third, although most primary liver cancer cases are HCC, some of the included studies did not state whether the primary liver cancer patients were histologically confirmed to be HCC. Fourth, the heterogeneity between studies was not well addressed by subgroup analysis, suggesting there were other potential confounding factors in the included studies. Fifth, the results of subgroup analysis should be interpreted with caution because of limited statistical power. We anticipate these issues will be addressed in future studies.
In summary, our research suggests that GSTM1/GSTT1 null genotypes are associated with increased risk of HCC in Chinese population. Considering the increasing prevalence of HCC in China and other countries worldwide, our finding may have important clinical and public health implications. More epidemiological and mechanistic studies are needed to further elucidate the role of GST polymorphisms in HCC and other liver cancers.
COMMENTS
Background
Asians, particularly Chinese, have a high risk of developing liver cancer. About 80%-90% of all cases of primary liver cancer diagnosed are hepatocellular carcinoma (HCC). Previous studies suggest that glutathione S-transferase (GST) polymorphisms (GSTM1 and GSTT1) may be risk factors for HCC. However, recent findings have been inconsistent.
Research frontiers
Meta-analysis was performed to assess the association between GSTM1 and GSTT1 polymorphisms and HCC risk in Chinese population.
Innovations and breakthroughs
The meta-analysis provided new evidence for the association between GSTM1 and GSTT1 polymorphisms and HCC risk in Chinese population. The results of this meta-analysis show that the GSTM1 and GSTT1 null genotypes are both associated with increased risk of HCC in Chinese population, suggesting that GSTM1/GSTT1 null genotype carriers have 1.5 fold higher risk of developing HCC.
Applications
Since GSTM1/GSTT1 polymorphisms are implicated in the pathogenesis of HCC, population-based genetic screening in future may help to identify the individuals at high risk of developing liver cancers.
Terminology
Meta-analysis, which combines the results of several studies that address a set of related research hypotheses, is an important component of a systematic review procedure.
Peer review
This meta-analysis provided new insights into liver cancer research.
Footnotes
Supported by (partially) The Heilongjiang Provincial Health Department, No. 2009-201; and the Administration of Traditional Chinese Medicine of Heilongjiang Province, No. ZHY10-293
Peer reviewer: Wen Zhuang, Professor, Department of Gastrointestinal Surgery, West China Hospital/West China Medical Center, Sichuan University, Chengdu 610041, Sichuan Province, China
S- Editor Sun H L- Editor Wang XL E- Editor Zheng XM