Glutathione S-transferase M1 polymorphism and esophageal cancer risk: An updated meta-analysis based on 37 studies

Abstract

AIM: To evaluate the relationship between glutathione S-transferase M1 (GSTM1) polymorphism and susceptibility to esophageal cancer (EC).

METHODS: A comprehensive search of the United States National Library of Medicine PubMed database and the Elsevier, Springer, and China National Knowledge Infrastructure databases for all relevant studies was conducted using combinations of the following terms: “glutathione S-transferase M1”, “GSTM1”, “polymorphism”, and “EC” (until November 1, 2014). The statistical analysis was performed using the SAS software (v.9.1.3; SAS Institute, Cary, NC, United States) and the Review Manager software (v.5.0; Oxford, England); crude odds ratios (ORs) with 95% confidence intervals (CIs) were used to assess the association between the GSTM1 null genotype and the risk of EC.

RESULTS: A total of 37 studies involving 2236 EC cases and 3243 controls were included in this meta-analysis. We observed that the GSTM1 null genotype was a significant risk factor for EC in most populations (OR = 1.33, 95%CI: 1.12-1.57, P_{heterogeneity} < 0.000001, and I² = 77.0%), particularly in the Asian population (OR = 1.53, 95%CI: 1.26-1.86, P_{heterogeneity} < 0.000001, and I² = 77.0%), but not in the Caucasian population (OR = 1.02, 95%CI: 0.87-1.19, P_{heterogeneity} = 0.97, and I² = 0%).

CONCLUSION: The GSTM1 null polymorphism may be associated with an increased risk for EC in Asian but not Caucasian populations.

Key Words: Meta-analysis; Glutathione S-transferase M1; Polymorphism; Esophageal cancer; Deletions

Core tip: Many previous studies have investigated the association between the glutathione S-transferase M1 (GSTM1) null genotype and the risk of esophageal cancer (EC), but these studies have provided controversial findings. The present study represents the largest meta-analysis to estimate the association between the GSTM1 polymorphism and EC risk. We investigated these two genotypes (GSTM1 null or GSTM1 present) in terms of EC morbidity.

INTRODUCTION

Esophageal cancer (EC), which is the sixth leading cause of cancer-associated death worldwide, has two major histological types: esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EADC)[1]. The five-year survival rate for EC is less than 20%[2]. A growing body of epidemiological evidence suggests that environmental factors together with genetic factors play important roles in the risk of developing EC[3,4]. The major risk factors for EC include alcohol consumption, smoking tobacco, and micronutrient deficiency[5]. Various factors and multiple processes lead to EC development. In addition to the above mentioned factors, genetic factors also account for EC cases.

Previous studies have suggested that glutathione S-transferases (GSTs) are phase II metabolizing enzymes that detoxify free radicals and other carcinogens[6]. Therefore, individual variation in phases II enzyme activity may contribute to varying susceptibility to EC progression. The GST family plays an important role in the detoxification of a variety of electrophilic carcinogens through conjugation with glutathione, and there is a widely variable organ distribution of the four classes of GSTs, although all of these display esophageal expression: GSTA (a), GSTM (m), GSTP (p), and GSTT (h)[7,8]. Homozygous deletions of GSTM1 have been associated with the loss of enzymatic activity for the detoxification of carcinogens, which consequently confers a risk for some cancers, such as colorectal, pancreatic, esophageal, and head and neck cancers[9-12]. Therefore, the null genotype of GSTM1 might be associated with an increased risk of EC[13]. Many previous studies have investigated the association between the GSTM1 null genotype and the risk of EC, but these studies have provided controversial findings[8,14-18]. It remains uncertain whether the GSTM1 polymorphism is a risk factor for EC. Considering these controversial results, we conducted a meta-analysis summarizing reported case-control or prospective studies to assess the risk of EC.

MATERIALS AND METHODS

Search strategy

We conducted a comprehensive search of the US National Library of Medicine PubMed database and the Elsevier, Springer, and China National Knowledge Infrastructure databases for all relevant studies using combinations of the following terms: “glutathione S-transferase M1”, “GSTM1”, “polymorphism”, and “EC” (until November 1, 2014). Additional eligible studies were identified through references that were cited in the relevant articles. The full text of each potentially relevant paper was scrutinized to ensure that the following inclusion criteria were met: (1) the articles clearly described studies concerning the association of EC with GSTM1 polymorphism; (2) The study design should be observational (case-control or prospective); (3) Sufficient data for estimating the odds ratios (ORs) and 95% confidence intervals (CIs) were present; and (4) If more than one publication reported on the same population, we selected the study with the largest sample size.

Data extraction

Two researchers independently extracted the following data from each study that met the inclusion criteria: first author’s surname, year of publication, country, ethnicity of the subjects (stratified into Asian, Caucasian, and African populations), sources of the controls (categorized as population-based studies and hospital-based studies), histological type (adenocarcinoma and squamous cell carcinoma), number of different genotypes in cases and controls, smoking status, and the frequency of different genotypes in the cases and controls. Individuals with “present” genotype were defined as carriers with at least one of the functional alleles in accordance with the definition used in most studies, whereas individuals carrying none of the alleles were classified as the “null” genotype.

Statistical analysis

Crude ORs with 95%CIs were used to estimate the strength of the relationship between the GSTM1 polymorphism and EC risk. The pooled ORs were evaluated for null vs present genotypes. The heterogeneity was assessed using a χ² analysis based on the Q-test[19]. The heterogeneity was considered significant for P < 0.05. In the presence of significant heterogeneity, a random-effects model (the DerSimonian and Laird method)[20] was used to calculate pooled estimates; otherwise, a fixed-effects model (the Mantel-Haenszel method) was used[21]. These two models provided similar results in the absence of heterogeneity. The potential publication bias was assessed using a funnel plot and linear regression asymmetry test[22].The statistical analyses were performed using the SAS (v.9.1.3; SAS Institute, Cary, NC, United States) and Review Manager software (v.5.0; Oxford, England) with two-sided P values and a 0.05 significance level.

RESULTS

Eligible studies

A total of 37 studies involving 2236 EC cases and 3243 controls were finally included in this meta-analysis[8,12-18,23-51]. The main characteristics of these studies are presented in Table 1. Among these studies, one case-control study was nested within a cohort study[51], and 25 studies provided data of the histological type of the EC cases. The smoking statuses of the cases and controls were recorded in six studies.

Table 1 Characteristics of the studies included in this meta-analysis.

Ref.	Year	Ethnicity	Country	Source of controls	Genotype distribution
					Case		Control
					Null	Present	Null	Present
Morita et al[49]	1997	Asian	Japan	PB	23	30	55	77
Nimura et al[43]	1997	Asian	China	HB	47	42	63	74
Hori et al[31]	1997	Asian	Japan	PB	41	53	196	232
Lin et al[27]	1998	Asian	China	PB	20	25	21	24
Shao et al[36]	1999	Asian	China	HB	68	40	55	57
van Lieshout et al[30]	1999	Caucasian	The Netherland	PB	17	17	128	119
Tan et al[14]	2000	Asian	China	PB	46	104	76	74
Shi et al[40]	2002	Asian	China	HB	67	31	51	69
Yokoyama et al[15]	2002	Asian	Japan	HB	103	131	321	313
Gao et al[42]	2002	Asian	China	PB	106	35	133	90
Casson et al[48]	2003	Caucasian	Canada	PB	26	19	25	20
Wang et al[50]	2003	Asian	China	PB	27	35	19	19
Wang et al[44]	2004	Asian	China	HB	74	53	44	57
Abbas et al[33]	2004	Caucasian	French	PB	39	29	59	61
Roth et al[51]	2004	Asian	China	Nest	41	90	145	309
Han et al[39]	2005	Asian	China	HB	46	43	48	51
Lu et al[28]	2005	Asian	China	PB	36	68	4	100
Yin et al[35]	2005	Asian	China	HB	69	37	61	45
Casson et al[45]	2006	Caucasian	Canada	HB	34	22	54	41
Jain et al[32]	2006	Asian	India	HB	39	61	51	86
Dong et al[47]	2007	Asian	China	HB	76	44	51	69
Wideroff et al[41]	2007	Caucasian	United States	PB	37	30	121	87
Rossini et al[29]	2007	Caucasian	Brazil	HB	51	74	99	153
Li et al[34]	2008	Asian	China	PB	77	48	55	70
Zendehdel et al[23]	2009	Caucasian	Sweden	PB	85	85	230	239
Ji et al[38]	2010	Asian	China	PB	111	78	98	127
Malik et al[26]	2010	Asian	India	HB	68	67	79	116
Liu et al[24]	2010	Asian	China	PB	54	43	32	65
Moaven et al[12]	2010	Asian	Iran	HB	65	83	58	78
Li et al[16]	2010	Black	Africa	HB	33	206	80	200
Gao et al[37]	2012	Asian	China	HB	22	18	45	35
Chen et al[17]	2012	Asian	China	HB	68	31	90	96
Liu et al[46]	2013	Asian	China	HB	47	63	74	146
Talukdar et al[25]	2013	Asian	India	PB	44	68	40	90
Sharma et al[13]	2013	Asian	India	PB	129	186	139	297
Dura et al[8]	2013	Caucasian	The Netherland	PB	228	204	318	273
Djansugurova et al[18]	2013	Asian	Kazakhstan	PB	72	43	24	76

PB: Population-based study; HB: Hospital-based study.

Meta-analysis

Considering the obvious heterogeneity among the 37 included studies (P < 0.001, I² = 77%), the random-effects model (DerSimonian-Laird method) was used to calculate the pooled ORs for the GSTM1 null vs GSTM1 present genotypes. Individuals with GSTM1 null genotypes were significantly associated with an increased risk for EC compared those carrying the GSTM1 present genotype (OR = 1.33, 95%CI: 1.12-1.57, Figure 1). In the sensitivity analysis, individual studies were sequentially removed. The results indicated that no individual study significantly affected the pooled OR, suggesting that these results were statistically robust.

Figure 1 Forest plot for the association between the glutathione S-transferase M1 polymorphism and esophageal cancer risk.

In the subgroup analysis based on ethnicity, a positive correlation was observed between the GSTM1 null genotype and the EC risk in the Asian population (OR = 1.53, 95%CI: 1.26-1.86) but not in the Caucasian population (OR = 1.02, 95%CI: 0.87-11.19). However, the results of the stratified analysis based on histological type showed that the GSTM1 null genotype increased the risk of EC in patients whose histological type was unknown, but no statistically significant association was observed for either the ESCC patients or the EADC patients. Moreover, the heterogeneity was significantly reduced among Caucasian populations and studies based on the histological type of adenocarcinoma. Because only one study (14EADC, 137 ESCC) reported an association between the GSTM1 polymorphism and EADC in Asian populations, we only analyzed the data according to ESCC and EADC in Caucasian populations, and the results showed no statistically significant association between the GSTM1 polymorphism and ESCC or EADC. The main results of this meta-analysis and the heterogeneity test are shown in Table 2.

Table 2 Main results of the pooled odds ratios in this meta-analysis.

	Null vs present
	No. of studies	OR	95%CI	P value
Total	37	1.33	1.12-1.57	0.00001
Ethnicity
Asian	27	1.53	1.26-1.86	0.00001
Caucasian	8	1.02	0.87-1.19	0.97
Histological type
ESCC	22	1.15	0.91-1.45	0.00001
EADC	8	0.98	0.81-1.18	0.93
NR	12	1.82	1.58-2.09	0.007
Smoking status
Smokers	6	0.97	0.53-1.77	0.00001
Nonsmokers	6	0.97	0.57-1.64	0.001
Histological type of Caucasian
ESCC	5	1.15	0.91-1.45	0.33
EADC	8	0.98	0.81-1.18	0.93

The P value for heterogeneity. ESCC: Esophageal squamous cell carcinoma; EADC: Esophageal adenocarcinoma; NR: Not reported.

Publication bias

A funnel plot was used to graphically estimate the publication bias of the literature. As shown in Figure 2, the shape of the funnel plot was symmetrical in the overall population, suggesting the absence of publication bias. The results of Egger’s test showed statistical evidence for funnel plot symmetry (t = 1.76, P = 0.0873).

Figure 2 Funnel plot evaluating the risk of publication bias in this meta-analysis.

DISCUSSION

GSTM1 is a member of the family of cytosolic GSTs, which are phase II xenobiotic-metabolizing enzymes. These enzymes play a crucial role in the detoxification and elimination of electrophilic carcinogens through conjugation with glutathione[12]. Many studies have investigated the association between the GSTM1 null genotype and various types of cancer, such as colorectal carcinoma, lung cancer, liver cancer, and EC, but the findings are controversial, particularly those obtained for EC[52,53]. The results of this meta-analysis showed that the GSTM1 null genotype is significantly associated with an increased risk of EC in the overall population. Furthermore, in the subgroup analysis by ethnicity, we detected a significant association between the GSTM1 polymorphism and EC risk in Asians but not in Caucasians, suggesting that the GSTM1 null polymorphism might contribute to increased susceptibility to EC in Asians. Similar results have been obtained in several previous meta-analyses[54,55]. However, other studies have shown conflicting results. A pooled analysis of 20 studies from the Archives of Medical Research revealed that there was no evidence of increased risk of EC associated with the GSTM1 null genotype[56]. The result might reflect a relatively small sample size and, to a lesser extent, different ethnicities, different histological types and the source of the controls.

In the present meta-analysis, most of the included studies concerned Asian populations. This phenomenon might be attributed to the occurrence of EC, which displays a remarkable geographical difference. Specifically, the ‘‘EC belt’’, which stretches from North Central China westward through Central Asia and northern Iran, exhibits a particularly high EC incidence in Asian populations[57], which explains why many of the studies were conducted in Asian countries.

In the subgroup analysis based on histological type, no significant association was detected between the GSTM1 polymorphism and ESCC or EADC risk, indicating that histological type might affect the statistical correlation between the GSTM1 polymorphism and EC. Similar results have been reported in previous studies[23,55,58], indicating that further clarification of the histological type might avoid the interference of some confounding factors.

Several potential limitations of the present meta-analysis should also be acknowledged. Only one of the included studies was conducted in Africa, and it did not provide sufficient data for the subgroup analysis based on ethnicity. Therefore, we could not include the African population in the subgroup analysis based on ethnicity. Moreover, only published studies were included in the present meta-analysis, which might have biased the results.

In conclusion, this meta-analysis demonstrated that the GSTM1 null polymorphism might be associated with an increased risk for EC in Asian populations but not in Caucasian populations. Larger well-designed epidemiological studies are warranted to verify these findings.

COMMENTS

Background

Esophageal cancer (EC), which is the sixth leading cause of cancer-associated death worldwide, has two major histological types: esophageal squamous cell carcinoma and esophageal adenocarcinoma. The five-year survival rate for EC is less than 20%. Previous studies have suggested that glutathione S-transferase (GSTs) are phase II metabolizing enzymes that detoxify free radicals and other carcinogens. Therefore, individuals with low phase II activity might have a higher risk of developing cancer. The GST family plays an important role in the detoxification of a variety of electrophilic carcinogens through conjugation with glutathione, and there is a widely variable organ distribution of the four classes of GSTs, namely, GSTA (a), GSTM (m), GSTP (p), and GSTT (h), although all show esophageal expression.

Research frontiers

A growing body of epidemiological evidence suggests that environmental factors together with genetic factors play important roles in the risk of developing esophageal carcinoma: alcohol consumption, smoking tobacco, and micronutrient deficiency are considered the major risk factors for EC. The GSTM1 null genotype has been associated with an increased risk of EC. Many previous studies have investigated the association between the GSTM1 null genotype and the risk of esophageal carcinoma, but these studies provide controversial findings.

Innovations and breakthroughs

The results of the present study indicated that the GSTM1 null polymorphism might be associated with an increased risk of EC in Asian populations but not in Caucasian populations, which would be helpful for the identification of individuals at an increased risk of developing EC.

Applications

The present study enhances the current understanding of the effects of GSTM1 on EC. Larger well-designed epidemiological studies are warranted to confirm the precise mechanism underlying the involvement of the GSTM1 gene in EC progression.

Terminology

GSTM1 is a primary member of the GST family, which comprises enzymes that play important roles in the detoxification of a variety of electrophilic carcinogens through conjugation with glutathione. Homozygous deletions of GSTM1 might disrupt enzymatic detoxification of carcinogens and consequently confer risk for some cancers, such as colorectal, pancreatic, esophageal, and head and neck cancers.

Peer-review

The present study analyzed the effect of the GST1 polymorphism on EC risk. The meta-analysis of 37 studies showed that the GSTM1 null polymorphism is associated with a significantly increased risk of EC.