Association between the Glutathione-S-transferase T1 null genotype and esophageal cancer susceptibility: a meta-analysis involving 11,163 subjects

Background Glutathione-S-Transferase T1 (GSTT1) null genotype has been shown to be associated with the risk of esophageal cancer. However, the results remain inconsistent. Thus a comprehensive meta-analysis was conducted to assess the strength of association between GSTT1 null genotype and the risk of esophageal cancer. Materials and Methods A literature search of PubMed, Embase, China National Knowledge Infrastructure (CNKI) and Wanfang databases up to March 31, 2017 was conducted and 30 eligible articles with 4482 cases and 6681 controls were finally recruited. The strength of correlation between GSTT1 polymorphism and the susceptibility of esophageal cancer was assessed by the crude odds ratios (ORs) with 95% confidence intervals (CIs). Subgroup analyses and sensitivity analyses were performed to further identify the association. Results GSTT1 null genotype significantly increased the risk of esophageal cancer (OR = 1.20; 95% CI 1.04–1.40; P < 0.05). In a subgroup analysis by ethnicity, GSTT1 null genotype was correlated with a significantly increased risk of esophageal cancer among Asians (OR = 1.33; 95% CI 1.12–1.58; P < 0.05), instead of Caucasians or Africans (OR = 0.91; 95% CI 0.65–1.26; P > 0.05 for Caucasians and OR = 1.32; 95% CI 0.98–1.77; P > 0.05 for Africans). In the analysis by histological type, GSTT1 null genotype was correlated with a significantly increased risk of esophageal squamous cell carcinoma (OR = 1.34; 95% CI 1.12–1.61; P < 0.05), particularly among Asians (OR = 1.54; 95% CI 1.30–1.82; P < 0.05), but not among Caucasians or Africans (OR = 0.87; 95% CI 0.48–1.57; P > 0.05 for Caucasians and OR = 1.32; 95% CI 0.98–1.77; P > 0.05 for Africans). In addition, there is no significant correlation between GSTT1 null genotype and the risk of esophageal adenocarcinoma (OR = 0.98; 95% CI 0.71–1.35; P > 0.05). Conclusions Our findings demonstrate that GSTT1 null genotype significantly increases esophageal cancer risk, particularly in Asians.


INTRODUCTION
Esophageal cancer is the sixth leading cause of cancer-related mortality and the eighth most common cancer worldwide [1]. An estimated 455,800 new esophageal cancer cases and 400,200 deaths occurred in 2012 worldwide [2]. The two major types are esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EADC). Smoking and drinking are well-known environmental risk factors for ESCC, whereas

Meta-Analysis
Oncotarget 15112 www.oncotarget.com obesity and chronic gastroesophageal refluxing are main EADC risk factors. However, only a subset of individuals exposed to those environmental risk factors develop EC, suggesting a role of host susceptibility factors. Some studies have suggested that genetic polymorphisms might explain individual differences in susceptibility to esophageal cancer [3,4].
Glutathione-S-transferases (GSTs) are important phase II biotransformation enzymes that catalyzing the nucleophilic addition of glutathione to several hazardous xenobiotics, including phase I electrophilic and carcinogenic metabolites [5]. However, these enzymes can also activate certain chemicals that target cellular proteins and DNA to elicit detrimental carcinogenic effects through genotoxic and non-genotoxic mechanisms. One of important enzymes in GSTs family is GSTT1. GSTT1 is genetically polymorphic, and deletion polymorphism of the GSTT1 loci (null genotype) results in the loss of functional activity. Several studies have found that GSTT1 null genotype is strongly associated with susceptibility to a number of cancers, such as colorectal, renal and oral cancers et al. [6][7][8]. Previous studies have been published to estimate the association between GSTT1 null genotype and the risk of esophageal cancer, but the results are inconsistent [9][10][11].
To date, several meta-analysis studies have reported the association between null GSTT1 genotype and the risk of esophageal cancer. However, the results of these studies remain outdated and incomprehensive [12][13][14]. In the last 4 years, many case-control studies were published to estimate this association. Thus, to obtain a conclusive result about this association, we performed current metaanalysis that includes all recent publications to review and summarize the association between the GSTT1 polymorphism and the risk of esophageal cancer.

Characteristics
In total, 100 articles were retrieved. Figure 1 summarized the selecting process. Finally, a total of 30 studies with 4482 cases and 6681 controls met the inclusion criteria [9][10][11]. Among them, 18 were from Asians, 10 were from Caucasians, and 2 were from Africans. There were 21 studies focused on the risk of ESCC with 3272 cases and 5535 controls, and 8 studies focused on the risk of EADC with 646 cases and 1908 controls. Characteristics of included studies and the distribution of GSTT1 polymorphism are summarized in Table 1 and Table 2, respectively. Table 3 showed the main result of the association between GSTT1 null genotype and the risk of esophageal Oncotarget 15113 www.oncotarget.com cancer. Overall, there was a significant correlation of GSTT1 null genotype with esophageal cancer risk (OR = 1.20; 95% CI 1.04-1.40; P < 0.05; Table 3 and Figure 2). In a subgroup analysis by ethnicity, GSTT1 null genotype was correlated with a significantly increased risk of esophageal cancer among Asians (OR = 1.33; 95% CI 1.12-1.58; P < 0.05; Table 3 and Figure 2), but not among Caucasians and Africans (OR = 0.91; 95% CI 0.65-1.26; P > 0.05 for Caucasians and OR = 1.32; 95% CI 0.98-1.77; P > 0.05 for Africans; Table 3 and Figure 2).

Test for publication bias, sensitivity analyses, and heterogeneity
Publication bias was assessed by both the Begg's funnel plot and the Egger's test. The shape of the Begg's funnel plot did not reveal any evidence of obvious asymmetry ( Figure 5). Egger's test further suggested no evidence of publication bias (P = 0.210). Thus, there was no obvious publication bias in this meta-analysis.
The sensitivity analysis was conducted to test the influence of an individual data on the pooled ORs and CIs by eliding a study in turn. Our findings suggested that the present meta-analysis results were relatively robust and stable ( Figure 6).
As showed in Figure 2 and Figure 3, heterogeneity was significant in overall and in some subgroups. Thus, we measured the sources of heterogeneity by subgroup

DISCUSSION
Many studies suggest genetic variants play important roles in individual susceptibility to esophageal cancer [3,42]. In decades, epidemiological studies have been performed to assess the association of GSTT1 null genotype with the risk of esophageal cancer. However, the results were inconsistent [34,36]. Previous meta-analyses also investigated the association [12][13][14], the findings should be interpreted with very cautions. In Weng et al. study, 11 studies included in the meta-analysis were in Chinese Han population. Although the results showed a significant association between GSTT1 null genotype and esophageal cancer risk, the single Chinese Han population limited the power of the statistical analysis [12]. Also In Yi et al. study, 15 studies included in the meta-analysis were in Asian population. Although the results showed a significant association between GSTT1 null genotype and esophageal cancer risk, the single Asian population limited the power of the statistical analysis [13]. In Cai et al. study, a total of 24 studies were used. Adjusted ORs with corresponding 95% CIs were reported in 9 studies. In the overall analysis there was no significant association between GSTT1 null genotype and esophageal cancer risk. However, meta-analysis of adjusted ORs showed a significant association between GSTT1 null genotype and esophageal cancer risk [14]. Because of the lack of available data, subgroup analysis by histological type was not performed in these studies [12][13][14]. Thus we conducted a comprehensive meta-analysis, to investigate not only the strength of association between GSTT1 null genotype and the risk of esophageal cancer, but also the association of GSTT1 null genotype with the risk of different histological types. This meta-analysis, including 30 case-control studies with 4482 cases and 6681 controls, identified the association between GSTT1 null genotype and esophageal cancer risk. GSTT1 null genotype significantly increased overall esophageal cancer risk. In a subgroup analysis by ethnicity, GSTT1 null genotype was correlated with a significantly increased risk of esophageal cancer among Asians. In the analysis by histological type, GSTT1 null genotype was correlated with a significantly increased risk of ESCC, particularly in Asians. However, there was no significant correlation of GSTT1 null genotype with the risk of EADC. To date, this is the first meta-analysis concerning the association of GSTT1 null genotype with the risk of different histological types. GSTT1, encodes an enzyme that plays a crucial role in the detoxification of a variety of endogenous or exogenous carcinogens. It is located on 22q11.23 with      [5]. GSTT1 is genetically polymorphic, and GSTT1 null genotype results in the loss of functional activity [12]. Our results demonstrated that GSTT1 null genotype significantly increased overall esophageal cancer risk.
Since the results from meta-analysis can be affected by histological types, a subgroup analysis was carried out regarding different histological type for the GSTT1 null genotype. GSTT1 null genotype was correlated with a significantly increased risk of ESCC among Asians, but not among Caucasians and Africans. However, there was no significant correlation of GSTT1 null genotype with the risk of EADC. All results should be interpreted with caution. Only two African studies were recruited in the current meta-analysis, which may restrict statistical power to detect a real assessment in Africans. More large scale studies are needed to verify the results. Subgroup analyses were also performed regarding ethnicity for the GSTT1 null genotype. GSTT1 null genotype was correlated with a significantly increased risk of esophageal cancer among Asians, but not among Caucasians and Africans. The results were more robust on histological type of ESCC. This meta-analysis confirmed the mutual effect of GSTT1 null genotype in different populations to the risk of esophageal cancer. Possible explanations include: (1) significance of these enzymes may vary with the ethnicity genetic backgrounds, environmental exposures and histological types. The regional difference in the frequency of esophageal cancer is probably due to genetic polymorphism and variable exposure to environmental factors; (2) GSTs metabolize a variety of overlapping substrates and individuals lacking GSTT1 can also metabolize the carcinogens by other alternative GST enzymes. Furthermore, there was only one study concerning the association between GSTT1 null genotype and EADC on Asians. Thus, we failed to evaluate the potential role of GSTT1 null genotype in EADC risk in Asians due to the lack of available data to date. More case-control studies on the GSTT1 null genotype are encouraged, especially in Asians, for a better understanding the role of GSTT1 null genotype in the EADC development.
Some limitations must be acknowledged in the current meta-analysis. First, significant heterogeneity was observed between publications for GSTT1 null genotype. Potential sources of heterogeneity include the study design, publication year, ethnicity, country, histological type, sample size, and so on. When subgroup analyses were carried out according to ethnicity and histological type, this heterogeneity was reduced or removed in some subgroups, implying different effects on histological types and ethnic populations. These findings should be interpreted with very cautions. Second, our findings were based on unadjusted ORs and CIs, whereas a more precisely investigation could be performed if the sufficient individual data were available. Third, only two African studies were recruited in the current meta-analysis, the results in African population should be interpreted with caution. More large scale studies are needed to verify the results in Africans. Finally, due to lack of uniform individual-level data, further stratified analysis to measure any interactions between gene-gene variation and genemetabolic traits was not performed.
In conclusion, our meta-analysis findings demonstrated that GSTT1 null genotype significantly increased esophageal cancer risk, particularly in Asians. In addition, GSTT1 null genotype was correlated with a significantly increased risk of ESCC, particularly among Asians. However, more studies are warranted to confirm or refute these correlations, particularly with respect to gene-gene and gene-environment interactions.

Study selection
Pubmed, Embase, China National Knowledge Infrastructure (CNKI) and Wanfang databases (the search was updated in March 31, 2017) were searched using the following terms: 'glutathione S-transferase T1' or 'GSTT1', 'polymorphism' or 'variant', and 'esophageal' or 'esophagus', and 'cancer' or 'carcinoma' or 'tumor' or 'malignancy'. The literature search was limited to English or Chinese articles. Additional publications were identified by a manual search based on references of retrieved studied or reviews.

Inclusion and exclusion criteria
The selection criteria were: (1) in a case-control study design, (2) studies that evaluated the relationship between the GSTT1 null genotype polymorphism and esophageal cancer, (3) usable data on genotype frequency. Thus, reports without usable data, reviews, comments and duplicated publications were excluded.

Data extraction
The data were collected by two independent reviewers. The extracted information contained: first author, year of publication, country of origin, ethnicity, number of cases and controls, genotyping method and characteristics of cases and controls. When come to conflicting assessment, disagreements were settled through a discussion among all authors.

Statistical analysis
The strength of correlation between GSTT1 null genotype and the susceptibility of esophageal cancer was assessed by the crude odds ratios (ORs) with 95% confidence intervals (CIs). A P < 0.05 (two-tailed) was considered as statistical significance. A Chi-squarebased I 2 test was used to detect heterogeneity [43] and an I 2 < 25% indicates low heterogeneity, 25% ≤ I 2 ≤ 50% indicates moderate heterogeneity, and I 2 > 50% indicates large heterogeneity [44]. When I 2 > 50% or P < 0.10 (twosided), the random-effects model (the DerSimonian-Laird method) was utilized to pool the data [45], otherwise the fixed-effects model (the Mantel-Haenszel method) was used [46]. Subgourp analyses were conducted according to different ethnicity to identify the specific effects of heterogeneity. Publication bias was assessed by Begg's funnel plot and Egger's test [47]. Sensitivity analyses were conducted by one-way method. All statistical analyses were performed using STATA version 12.0 software (Stata Corporation, College Station, TX, USA).

Author contributions
FH, NZ and LZ designed the study. All authors were involved in article search, extraction, analysis and interpretation of data. FH drafted the manuscript, which was critically revised by NZ and LZ. All authors read and approved the final manuscript.