Increased Risk of Developing Digestive Tract Cancer in Subjects Carrying the PLCE1 rs2274223 A>G Polymorphism: Evidence from a Meta-Analysis

Background To date, the association between phospholipase C epsilon 1 (PLCE1) rs2274223 A>G and risk of digestive tract cancer (DTC) remains inconclusive. To derive a more precise estimation of the association, we conducted a meta-analysis on all eligible case–control studies involving 8281 cases and 10,532 controls. Methods A comprehensive search was conducted to identify all eligible studies of PLCE1 rs2274223 polymorphism and digestive tract cancer risk. The pooled odds ratio (OR) and the 95% confidence interval (95% CI) were calculated using a fixed or random effect model. Heterogeneity, publication bias, and sensitivity analysis were also explored. Results Overall, the PLCE1 rs2274223 A>G polymorphism was associated with risk of DTC in all genetic models (GA vs. AA: OR = 1.21, 95% CI = 1.14–1.29, P<0.001; GG vs. AA: OR = 1.30, 95% CI = 1.06–1.60, P = 0.012; GG/GA vs. AA: OR = 1.20, 95% CI = 1.10–1.32, P<0.001; GG vs. GA/AA: OR = 1.21, 95% CI = 1.01–1.46, P = 0.040). The recessive model did not reach statistically significance when the P values were Bonferroni corrected to 0.0125. In the stratified analysis by cancer type, ethnicity, and source of controls, significantly increased risk was observed for esophagus cancer, Asians in three genetic models (heterozygote comparison, homozygote comparison and dominant model), population-based studies in all genetic models, and for gastric cancer in the heterozygote comparison and dominant model after Bonferroni correction. However, in the subsite of gastric cancer, no significant association was found either in cardia or non-cardia gastric cancer. Conclusion Our study indicated that PLCE1 rs2274223 A>G polymorphism was significantly associated with increased risk of DTC, especially among Asian populations. Due to some minor limitations, our findings should be confirmed in further studies.


Introduction
Digestive tract cancer (DTC) referring to a group of malignancies (e.g., located in oral cavity, pharynx and larynx, esophagus, stomach, small and large intestines) is the most common cancer worldwide. It has been estimated that there were nearly 316,970 new diagnosed cases and 147,150 deaths caused by DTC in the United States in 2011 [1]. Esophageal, gastric and colorectal cancers are the leading causes of cancer-related death in Eastern Asian countries [2,3]. Therefore, identification of potential risk factors for DTC may contribute to the prevention and early diagnoses of these lethal cancers.
It has been suggested that DTC carcinogenesis is a combined effect of multiple factors, which contains environmental factors, dietary habits and inherited susceptibility [4,5]. Alcohol consumption and tobacco smoke are the well-recognized risk factors for DTC [6]. Despite the high prevalence of Helicobacter pylori (HP) infection in gastric cancer, human papillomavirus (HPV) is recognized as a major risk factor for oropharygeal cancer [7,8]. Recently, accumulative evidence has shown that genetic factors, especially gene polymorphisms, which involve in multiple biological pathways, such as carcinogen metabolism, apoptosis, DNA repair, cell cycle regulation, and other cellular processes, play important roles in the etiology of DTC [9,10,11,12].
PLCE1 has been speculated to be a effector of small GTPases of the Ras, Rap and Rho families, and contains a guanine nucleotide exchange factor domain for Ras-like small GTPases at its Nterminus and two Ras-binding domains at its C-terminus [13,14,15,16]. Recent studies have reported that PLCE1 plays crucial roles in carcinogenesis and progression of several types of cancers, including cancers of the intestine, skin, bladder, colorectal and head and neck [17,18,19,20,21].
Rs2274223 (A.G) is a non-synonymous single nucleotide polymorphism (SNP) located in 26th exon of the PLCE1 gene and result in the amino acid change from histidine (His) to arginine (Arg) at codon 1927 of PLCE1. In 2010, two large-scale genome-wide association studies (GWASs) simultaneously reported that the new and notable low-penetrance susceptibility locus rs2274223 was strongly associated with risk of esophageal squamous cell carcinoma (ESCC) and gastric cardia adenocarcinoma (GCA) in Chinese population [22,23]. Recent studies indicated that rs2274223 (A.G) was associated with an increased risk of most DTC, such as cancers of oral cavity, pharynx and larynx, esophagus, and stomach [24,25,26,27,28,29,30,31]. Interestingly, another study proved that rs2274223 was associated with a protective effect against colorectal cancer (CRC) in a Chinese population [32]. Otherwise, as reported in a Dutch population and a South African population, it was unlikely that the PLCE1 rs2274223 SNP plays a role in esophageal adenocarcinoma (EAC) or ESCC susceptibility [33,34]. Besides, there was a study focused on the associations between the rs2274223-G allele and the prognosis of gastric cancer patients, suggesting that individuals carrying rs2274223 AG/GG genotypes had a higher survival rate than those carrying the AA genotype [35].
To date, the association between the rs2274223 (A.G) and the susceptibility of DTC are inconclusive in different cancer types and ethnicities, partially because of the different effects of the polymorphism on variants of DTC risk and the relatively small sample size in each of published studies. Hence, we conducted a meta-analysis on all eligible case-control studies involving 8281 cases and 10,532 controls to estimate the overall DTC risk of the rs2274223 (A.G) polymorphism.

Identification and Eligibility of Relevant Studies
We searched PubMed and Embase (updated to 28th February 2013) using the following search terms: ''rs2274223'', ''PLCE1'', ''10q23'', ''genetic susceptibility'', ''SNP'', ''polymorphism'' or ''variation'', and ''cancer'' or ''carcinoma'' or ''neoplasia''. The search was limited to English-language articles. We also used a hand search of references of original studies on this topic in order to identify additional studies. Studies included in our meta-analysis have to meet the following inclusion criteria: (a) evaluated the PLCE1 rs2274223 polymorphism and cancer risk, (b) used a casecontrol design and (c) contained available genotype frequency.

Data Extraction
Two investigators independently extracted data and reached a consensus on all the items in cases of discordance. For each eligible study, the following data were extracted: the first author's name, year of publication, ethnicity, country of origin, cancer type, value of Hardy-Weinberg equilibrium (HWE), source of controls, genotyping method,and numbers of genotyped cases and controls. Ethnic descents were categorized as European and Asian (the only one African study was excluded). For studies including subjects of different sites of gastric cancer (cardia and non-cardia), data were extracted separately whenever possible.

Statistical Analysis
The strength of the association between the PLCE1 rs2274223 polymorphism and cancer risk was assessed by the odds ratio (OR) and the 95% confidence interval (95% CI). Pooled ORs were obtained from the combination of individual studies by heterozygote comparison (GA vs. AA), homozygote comparison (GG vs. AA), a dominant model (AA/GA vs. GG), and a recessive model (AA vs. GA/GG). The significance of pooled ORs was determined using the Z-test. Bonferroni correction was utilized for multiple testing. Because multiple comparisons were performed 4 times respectively, the P value lesser than 0.05/4 (0.0125) was accepted for statistical significance after Bonferroni correction. Both the Cochran's Q statistic to test for heterogeneity and the I2 statistic to quantify the proportion of the total variation due to heterogeneity were calculated to estimate heterogeneity among the included studies [36,37]. If the P value of the Q test was ,0.05, indicating a lack of heterogeneity across studies, the summary OR estimate of each study was calculated by the fixed effects model (the Mantel-Haenszel method) [38]. Otherwise, random effects model (the DerSimonian and Laird method) was used [39]. Stratified analyses were also performed by cancer type, ethnicity, source of controls and site of gastric cancer. Sensitivity analyses were performed to evaluate the stability of the results by deleting a single study in the meta-analysis each time to show the influence of the individual data set to the pooled OR. Funnel plots and Egger's linear regression test were used to assess the potential publication bias [40]. All analyses were performed using Stata software (version 8.2; StataCorp LP, College Station, TX), using two-sided P values.

Quantitative Synthesis
The evaluation of the association between the PLCE1 rs2274223 polymorphism and the susceptibility to digestive tract cancer is presented in Table 2. Overall, the variant G allele of However, the recessive model did not reach statistically significance when the P values were Bonferroni corrected. We also  AA: OR = 1.14, 95% CI = 1.02-1.28, P = 0.017, I 2 = 58.10%), but this did not reach statistically significance when the P value was Bonferroni corrected. In the subgroup analysis by tumor sites of gastric cancer, there was no significant association detected in any genetic comparisons either in cardia or non-cardia gastric cancer.

Test of Heterogeneity
When evaluating the association between the PLCE1 rs2274223 polymorphism and the susceptibility to DTC, we found that there was significant heterogeneity for the homozygote comparison (GG vs. AA: P heterogeneity = 0.001, I 2 = 65.60%), dominant model comparison (GG/GA vs. AA: P heterogeneity = 0.009, I 2 = 57.60%) and recessive model comparison (GG vs. GA/AA: P heterogeneity = 0.005, I 2 = 60.00%) but not for the heterozygote comparison (GA vs. AA: P heterogeneity = 0.113, I 2 = 35.70%). Thus, we assessed the source of heterogeneity for the dominant model comparison by examining cancer type, ethnicity, source of controls and genotyping method. Meta-regression analyses showed that none of these concomitant variables could account for the substantial heterogeneity observed (ethnicity: P = 0.215, genotyping method: P = 0.925, cancer site: P = 0.286 and source of controls: P = 0.408).

Sensitivity Analysis
In order to reflect the influence of the individual dataset to the pooled ORs, we deleted a single study involved in the metaanalysis each time, but the corresponding pooled ORs were not altered materially (data not shown), suggesting that our results were statistically robust.

Publication Bias
Begg's funnel plot and Egger's test were performed to assess the publication bias of the literature. The shape of the funnel plots did not reveal any evidence of obvious asymmetry ( Figure 4 shows the funnel plot of the overall GA vs. AA and GG/GA vs. AA

Discussion
In the current meta-analysis, we ascertained that the PLCE1 rs2274223 A.G polymorphism was significantly associated with increased DTC risk, especially with gastric cancer and esophageal cancer. To our knowledge, this is the first study to investigate the association between the PLCE1 rs2274223 A.G polymorphism and the risk of DTC across different ancestries.
The PLCE1 protein is a member of the phospholipase family and functions as an effector of guanosine triphosphatases (GTPases) such as Ras, Rap1 and Rap2 involving regulation of cell growth, differentiation and apoptosis [14,21]. PLCE1 also interacts with IQ-domain GTPase-activating protein 1 (IQGAP1) [41], which plays an important role in angiogenesis and is expressed in the endothelial cells, where it binds VEGFR2, an important factor for the endothelial cell rearrangement and migration [42]. This polymorphism rs2274223 causes an amino acid change from histidine to arginine in the PLCE1 protein calcium-dependent lipid-binding (C2) domain.
Wang et al. and Abnet et al. simultaneously reported the strong association of the new and notable low-penetrance susceptibility locus rs2274223 with the increased risk of ESCC and GCA in the Chinese population by two large-scale GWASs [22,23]. However, it was unlikely that the PLCE1 rs2274223 SNP played a role in EAC or ESCC susceptibility in American and Dutch [26,33]. Further observation showed that the rs2274223-G allele had a stronger effect on female and GCA than male and non-cardia GC [25]. Luo et al. found that individuals carrying PLCE1 rs2274223 AG/GG genotypes had a higher survival rate than those carrying the AA genotype, which suggested that the rs2274223-G allele may be associated with the prognosis of the gastric cancer patients [35]. Yuan et al. and Ma et al. also performed GAWSs and confirmed that rs2274223 was associated with a significantly increased risk of head and neck cancer [24,31]. Furthermore, Li et al. proved that rs2274223 was associated with a decrease risk of CRC in a Chinese population [32].
Although many epidemiological studies regarding the PLCE1 rs2274223 polymorphism on the risk of variants of DTC had been conducted, the results were conflicting and inconclusive because of various reasons, such as different ethnicities, resident areas, sample size, environmental factors and diet habits. To provide a more comprehensive analysis on the association, we carried out this meta-analysis based on eleven case-control studies with 18,813 participants and indicated that G allele of the PLCE1 rs2274223 A.G polymorphism was associated with increased risk of DTC.
In the analysis stratified by cancer type, we observed increased susceptibility to esophageal cancer in the three genetic comparisons (heterozygote comparison, homozygote comparison and dominant model) and to gastric cancer in the heterozygote comparison (GA vs. AA) and the dominant model comparison (GG/GA vs. AA) after Bonferroni correction. The results were consistent with the conclusions of most previous studies [22,23,25,26,27,28,29,30], except for three studies about esophageal cancer in American, Dutch and South African [26,33,34], partially because of the different ethnicities and the relatively small sample size in the three studies.
When compared by ethnicity, statistically significantly increased risks were found among Asians for three genetic comparison (heterozygote comparison, homozygote comparison and dominant model) but not Europeans after Bonferroni correction. Although the exact mechanism for these ethnic differences is still unknown, one possible reason is due to differences in genetic backgrounds and in the environmental and lifestyle context (such as dietary habits, alcohol consumption and tobacco smoke) [43]. In addition, because of the gene-gene interaction, the influence of the PLCE1 rs2274223 A.G polymorphism might be masked or magnified by the presence of other genes which were unidentified yet in the development of cancer. Other factors such as selection bias, different matching criteria and limited number of studies with available data may have insufficient statistical power to detect a slight difference and may also generate a fluctuated risk estimate.
We also observed a significantly increased DTC risk among studies using population-based controls in every genetic model, but not using hospital-based controls. Some biases may exist in hospital-based studies, for such controls may represent a sample of an ill-defined reference population instead of the general population, particularly when the genotypes investigated were associated with the disease that the hospital-based controls may have. Thus, a proper and representative cancer-free control subject is very important in reducing biases in such genotype association studies.
Several previous studies showed that the association between rs2274223 and the risk of gastric cancer was stronger in the cardia than in the non-cardia gastric cancer [22,23,25,28]. Our study also explored the association with gastric cancer differed by anatomic subsite. Unfortunately, no significant association was detected in any genetic comparisons either in cardia or non-cardia gastric cancer. The null results might be due to the limited number of studies with available data on these characteristics and the cases of the eligible studies were from the different ethnicities, which had insufficient statistical power to detect a slight effect.
Some limitations of this meta-analysis may have affected the objectivity of the conclusions and should be considered in interpreting the results. First, the quantity of published studies was not sufficiently large for a comprehensive analysis, and the lack of original data in some studies limited our further evaluation of potential interactions, such as ethnicity, anatomic subsite and pathological subtype of cancer. Second, although perfect searching strategy was used to identify eligible studies for current metaanalysis, it was still possible that a few studies meeting inclusion criteria were not included. Third, the overall outcomes were based on unadjusted estimates, while a more precise evaluation should be conducted if more detailed individual data were available, such as age, sex, histological types and HP infection. Lacking of information may cause serious confounding bias.
In spite of these limitations, our present meta-analysis also had some advantages. First, we estimated the association conclusively between the PLCE1 rs2274223 A.G polymorphism and DTC risk, and further showed the significant association especially among Asians rather than Europeans. This study may also provide a potential genetic marker and a new insight into the etiology of DTC. Second, we pooled a substantial number of cases and controls from different studies, which greatly increased the statistical power of the analysis. Third, no publication biases were detected, which indicated that the results were likely unbiased.
In conclusion, our meta-analysis suggests that the PLCE1 rs2274223 A.G polymorphism is associated with DTC risk, especially with gastric cancer and esophageal cancer. The PLCE1 rs2274223 A.G polymorphism is an independent risk factor for the development of DTC, and will probably be a potential therapeutic target for new drugs. Nevertheless, larger and welldesigned multicentric studies including other potential DTC risks (such as gene-gene, gene-environmental interactions) should be carried out to validate our findings. Furthermore, research based on the non-DTCs (such as hepatic cellular cancer, lung cancer, breast cancer etc.) should be performed to explore the association between PLCE1 gene polymorphisms and cancer risks.

Supporting Information
Checklist S1 A PRISMA checklist for this meta-analysis.