Association of the Glutathione S-transferase T1 Null Genotype with Risk of Gastric Cancer: a Meta-analysis in Asian Populations

Background: A large number of studies have been published to investigate the association between the null genotype of glutathione S-transferase T1 ( GSTT1 ) with gastric cancer. However, the results were inconsistent and conflicting. The aim of this study was to estimate the relationship between this polymorphism in the GSTT1 gene and gastric cancer risk in Asian populations by meta-analysis. Materials and Methods: A literature search was performed in PubMed, Embase, Chinese Biomedical database (CBM), Weipu database, Wanfang database, and China National Knowledge Infrastructure database (CNKI). Statistical analysis was conducted by using Review Manager 5.3. Results: Thirty-nine studies with a total of 7,737 gastric cancer cases and 10,823 controls were included in this meta-analysis. The meta-analysis of total studies showed that the null genotype in GSTT1 was associated with increased risk of gastric cancer in Asians (OR=1.19, 95% CI=1.08-1.31, p =0.0002). Subgroup analysis showed a significant relationship between GSTT1 null genotype and gastric cancer in East-Asians, as well as in subgroup analysis of hospital-based design. On subgroup analysis by smoking status, alcohol status, Helicobacter pylori infection status, and histology type, no significant association of this polymorphism with susceptibility to gastric cancer was found. Conclusions: In conclusion, the results showed that the null genotype of GSTT1 is significantly associated with an increased risk in gastric cancer in Asian populations.


Introduction
It is estimated that by 2020, more than 15 million cases of cancer will occur in the world, with deaths increasing to 12 million (Kanavos, 2006). According to the report, gastric cancer is the fourth most common malignancy among worldwide and the second leading cause of global cancer death, which has a high incidence in Asia, especially in Eastern Asia (Jemal et al., 2011). Identifying risk factors for gastric cancer development is essential to prevent this deadly disease. Numbers of studies have been shown that gastric cancer is a disease of multiple etiologic factors involving infectious, nutritional and environmental factors (Setiawan et al., 2000;Tripathi et al., 2011;Shi et al., 2014). However, the regional differences in the incidence of gastric cancer worldwide might imply that variants in various genetic factors also influence the susceptibility to this disease (Jing et al., 2012).
Human glutathione S-transferases T1 (GSTT1) is phase II metabolizing enzyme protecting against cancer by detoxifying numerous potentially cytotoxic/genotoxic compounds, which has been implicated in the carcinogenic database and Chinese National Knowledge Infrastructure (CNKI) was conducted using combinations of the following keywords: "gastric cancer or gastric carcinoma or gastric neoplasm" and "polymorphism or variant or mutation" and "GSTT1 or glutathione S-transferase T1". The language was limited to Chinese and English. All studies that evaluate the relationship between polymorphisms in GSTT1 gene and gastric cancer risk were retrieved. Studies included in the meta-analysis had to meet all of the following criteria: (1) evaluation of the polymorphisms in GSTT1 gene and gastric cancer risk, (2) use of a case-control design, (3) published in a journal, (4) sufficient genotype data for estimating the odds ratio (OR) with 95% confidence intervals (CI). The exclusion criteria were: (1) abstracts and reviews, (2) studies that did not offer genotype frequency. Additionally, in studies with overlapping or repeating data, the most recent or complete studies with the largest sample size were recruited into the meta-analysis.

Data extraction
Two of the authors extracted all data independently (YZ and LCD), complied with the selection criteria, and must reach a consensus on all items. In case of conflicting evaluations, a third author (JB) assessed the articles. The following information were collected: first author's name, year of publication, country of origin, cancer type, genotyping method, mean age of case, total number of cases and controls, and genotype distributions (both null GSTT1 and non-null GSTT1) in cases and controls. The distributions of GSTT1 genotypes in subgroups (smoking status, alcohol status, Helicobacter pylori (H. pylori) infection status, and histology type) were also elicited.

Statistical analysis
Statistical analysis was conducted by using Review Manager 5.3 and STATA 12.0. The strength of association between GSTT1 null polymorphism and gastric cancer risk was assessed as OR with corresponding 95% CI. The significance of the pooled OR was evaluated by using a Z-test and p<0.05 was considered statistically significant.
Heterogeneity among studies was valued by Chisquare test, which was considered significant for p≤0.10. The fixed-effects and random-effects model were used to pool the results, when p≤0.10, a random-effects model was used; when p﹥0.10, a fixed-effects model was utilized. Six subgroups were analyzed to evaluate subregion-specific, study design-specific, smoking status-specific, alcohol status-specific, H. pylori infection status-specific and histology type-specific effects, which were performed by subregion group (East-Asia, South-Asia, Southeast-Asia and West-Asia), study design (Hospital-based study and Population-based study), smoking status (Smoking and Non-smoking), alcohol status (Alcohol and Non-alcohol), H. pylori infection status (Infection and Non-infection), and histology type (Diffuse and Intestinal). Sensitivity analysis was conducted by sequentially excluding each study to check the stability of the result. Inverted funnel plots was utilized to provide a diagnosis of publication bias and the degree of asymmetry was tested by Egger's test (p<0.05 was considered significant).

Meta-analysis results
The heterogeneity was analyzed for all 39 studies and the value of Chi-square test was 65.85 with 38 degrees of freedom and p=0.003 in a random-effects model. Overall, DOI:http://dx.doi.org/10.7314/APJCP.2016.17.3.1141 The GSTT1 Polymorphism and Gastric Cancer Risk: a Meta-analysis in Asia there was statistical evidence of an association between the null genotype of GSTT1 and gastric cancer risk. OR was 1.19 (95% CI=1.08-1.31) and the test for overall effect Z value was 3.68 (p=0.0002). The results suggested that null genotype in GSTT1 have an increased risk of gastric cancer ( Figure 2 Table 3.

Publication bias
Inverted funnel plot and Egger's test were used to evaluate the potential publication bias in this metaanalysis. As shown in Figure 5, the shape of the funnel plots seemed approximately symmetrical for total studies analysis. Additionally, in the subgroup analyses by subregion the funnel plots also did not show any evidence of publication bias. In Egger's test the result also revealed the absence of publication bias in the GSTT1 (p>0.05 for Null vs. Non-null model). Briefly, there was no risk of publication bias in this meta-analysis.

Sensitivity analysis
Sensitivity analysis was analyzed by sequentially excluding individual case-control study. The statistically similar results were obtained, which suggest the results of this meta-analysis are stable (Data were not shown

Discussion
Gastric cancer is a multistage process caused by    multifarious factors that environmental and genetic factors may all contribute to the etiology of this carcinoma (Al-Moundhri et al., 2009). The relationship between polymorphisms in genes which encode proteins involving in carcinogens metabolism and individual susceptibility to carcinogenic impacts of the specific chemical compound is a new and concern field of research. GSTT1 is one class of glutathione S-transferases (GSTs)one type of biotransformation enzymes, which is appear in most epithelial tissues of the human gastrointestinal tract. The presumptive function of GSTT1 is to protect tissues resist toxic or carcinogenic compounds that may enter to human body through the digestive tract or respiratory system (Qian et al., 2001). At present, the association between the GSTT1 genetic polymorphisms and susceptibility to gastric cancer has been conflicting. Accordingly, we performed a meta-analysis, a quantitative research method increasing sample size and statistical significance, to estimate whether null genotype in GSTT1 gene associated with gastric cancer risk. Finally, we conducted 39 published studies including 7737 cases and 10823 controls, which met the inclusion criteria for the meta-analyses. In overall combination studies, the results indicated that null polymorphism of GSTT1 is associated with a significant increase in the risk of gastric cancer. In the subgroup analysis by subregion, the GSTT1 null genotype was associated with an increased risk of gastric cancer in East Asian population but not in South Asian, Southeast Asian and West Asian population, which consistent with the statistical result performed by Jemal in 2011 (Jemal et al., 2011). Possible explanation to these different results may be that different genetic backgrounds and environmental exposures might play a key role in the development of gastric cancer. Furthermore, significantly increased gastric cancer risk was found in hospital-based design but not in population-based design. Such result may due to the biases in hospital-based design, the genotype distributions in hospital-based design could not represent the general population very well. In the subgroup analysis of smoking status and alcohol status, it is worthy to mention the inconsonant results which differ with a previous meta-analysis performed by Wang et al (Wang et al., 2014). In Wang's study, GSTT1 null genotype has a significantly increased gastric cancer risk among smokers and alcohol drinkers. Inversely, results of our study indicated the lack of association between GSTT1 null genotype and gastric cancer among smoker and alcohol drinker. The controversial results might due to the huge difference in simple size mainly. The other explanation to these different results may be that the difference or uniform definition of smoking and alcohol drinking in each study suggesting uncertain influence among smoker and alcohol drinker. Previous study suggested that H. pylori infection plays a central role in the gastric cancer carcinogenesis (Trajkov et al., 2007). Although H. pylori infection is independent risk factor for gastric cancer, there was no evidence to suggest an interaction between GSTT1 null polymorphism and this risk factor in the present study. Possible explanation to such result may be the limitation of studies because studies with small sample had insufficient statistical power to detect a slight effect.
In this meta-analysis, only three studies involving 259 cases and 720 controls examined the interaction of the GSTT1 null genotype and the H. pylori infection status in gastric cancer patients. Considering the limited studies, our result should be interpreted with caution. Therefore, more relevant data are needed to confirm this finding.
Some issues which may affect the results should be addressed when performing meta-analysis, such as heterogeneity, sensitivity analysis and publication bias. In our study, heterogeneity existed in overall comparisons. After carrying out sensitivity analysis by sequentially excluding individual studies, we found that the results of our meta-analysis are stable. Hence a random effect model was performed to pooling data. Publication bias is another important issue that may influence the results of meta-analysis, which also be discussed in the present study. In this meta-analysis, both examination results of Inverted funnel plot and Egger's test were indicated that there was no risk of publication bias in this meta-analysis (Inverted funnel plot seemed approximately symmetrical and p=0.675 for Egger's test).
To our knowledge, there was a recently published study by Zhang in 2013(Zhang et al., 2013. They also focus on the correlation of GSTT1 genetic polymorphisms with gastric cancer risk in the Asian population. But there are differences between these two studies. First, the present meta-analysis updates the recent data for this polymorphism and gastric cancer risk and includes more studies than Zhang's study, which possibly providing more reliable conclusions; Second, more subgroups of metaanalysis were addressed in our study, such as subregion group, smoking status, alcohol status, H. pylori infection status and histology type. Although these differences, our study also indicated the null genotype of GSTT1 might contribute to the risk of gastric cancer, which is consistent with Zhang's study results. In summary, the present meta-analysis suggested that null variant in the GSTT1 gene may contribute to gastric cancer risk. However, we have to mention several limitations of this study. First, a literature search just carried out in the selected databases; second, only published studies in Chinese and English were included for data analysis. For these reasons, some potential studies included by other databases or published in other languages or unpublished could be missed. In addition, due to lack of available data, the possible interaction of gene-environment could not be evaluated accurately. Accordingly, larger well-designed studies are warranted to verify these results. Moreover, gastric cancer carcinogenesis undergoes a complex progression from the development of chronic inflammation to acute neoplasias (Macarthur et al., 2004). Thus, further studies investigating GSTT1 polymorphisms for gastritis should be performed.