Lack of Association between Cytotoxic T-lymphocyte Antigen 4 (CTLA-4) -1722T/C (rs733618) Polymorphism and Cancer Risk: From a Case-Control Study to a Meta-Analysis

Background The association between cytotoxic T-lymphocyte antigen 4 (CTLA-4) gene -1722T/C polymorphism (rs733618) and cancer has been widely assessed, and a definitive conclusion remains elusive. We first performed a hospital based case-control study to measure this association of esophageal cancer with CTLA-4 -1722T/C polymorphism in Han Chinese population, and then carried out a meta-analysis to obtain a comprehensive evaluation for this issue. Methodology/Principal Findings This case-control study involved 629 esophageal squamous cell carcinoma (ESCC) cases and 686 age and gender well matched cancer-free controls. PCR-LDR (polymerase chain reaction-ligase detection reactions) method was used to identify genotypes. Meta-analysis was conducted by STATA (v12.0) software. This case-control study showed no significant difference in the genotype and allele distributions of CTLA-4 -1722T/C polymorphism between esophageal cancer cases and control subjects, in accord with the findings of the further meta-analysis in all genetic models. Evidence of large heterogeneity was observed among all eligible studies in the recessive model. Further subgroup analyses by ethnicity, cancer type and system, detected null associations in this meta-analysis. Conclusion This case-control study and the further meta-analysis, failed to identify the association between CTLA-4 -1722T/C polymorphism and cancer risk.


Introduction
It is estimated that about 12.7 million multiple cancer cases and 7.6 million cancer deaths have occurred in 2008 worldwide, with more than half of the cases and about two-thirds of the deaths in the developing countries [1]. The evidence is mounting that cancer is a complex disease results from interactions between multiple genetic backgrounds and environmental factors [2,3]. Of late, a number of studies demonstrate that genetic variants of the genes that regulate the activation and proliferation of T lymphocytes and nature killer (NK) cells may influence cancer risk [4,5]. In the last decade, single nucleotide polymorphisms (SNPs) have been extensively investigated, and many studies have examined the hypothesis that genetic variants of the immune genes may be relevant to the risk of a variety of cancers [6,7].
Cytotoxic T-lymphocyte antigen 4 (CTLA-4), also named CD152, is a member of the immunoglobulin superfamily. CTLA-4 is expressed mainly on activated T cells, acts as a vital restraining regulator of T-cell proliferation and activation, and induces Fas-independent apoptosis of activated T cells to further inhibit immune function of T-cell [6,8]. Blocking CTLA-4 function and enhancing T cell activation, several different types of malignant neoplasms in tumor-transplanted mice were inhibited or cured, and owned long-lasting antitumor immunity [9]. It suggests that CTLA-4 plays an important role in carcinogenesis. CTLA-4 gene is located on chromosome 2q33, and is composed of four exons that encode several functional domains of the CTLA-4 protein and possess several vital SNPs, such as the +49A/G (rs231775), -318C/T (rs5742909), CT60G/A (rs3087243), -1661A/G (rs4553808), and -1722T/C (rs733618) SNPs, etc [6,10].
A meta-analysis showed that CTLA-4 +49A/G polymorphism may be a risk factor for cancer, whereas -318C/T and +6230G/A (CT60) polymorphisms were lack of association with cancer [4]. Of late, Geng and colleagues reported a meta-analysis with a negative result on the association between CTLA-4 -1722T/C polymorphism and cancer risk [11]. Linkage disequilibrium (LD) plot of CTLA-4 (involving rs733618, rs4553808, rs5742909, rs231775 and rs3087243) was generated using Haploview 4.2 program and the results suggest that 21661A/G (rs4553808) and 2318C/T (rs5742909) are in high LD; the others are in low LD [11]. The CTLA-4 -1722T/C polymorphism has not been investigated in esophageal cancer. To further investigate this potential relationship, we decided to evaluate the association of CTLA-4 -1722T/C polymorphism with esophageal cancer risk in a hospital based case-control study, and then performed a comprehensive meta-analysis to derive a more precise result.

Subjects
This hospital-based case-control study included 629 sporadic esophageal squamous cell carcinoma (ESCC) cases and 686 cancer-free subjects consecutively recruiting from the Affiliated People's Hospital of Jiangsu University and Affiliated Hospital of Jiangsu University (Zhenjiang City, Jiangsu Province, China), between October 2008 and December 2010. All recruited subjects were local residents of Han Chinese population, and all ESCC subjects were diagnosed by surgical resection and pathologic examination. The ESCC subjects who had a history of personal

DNA extraction, SNP selection, and genotyping
Blood samples were collected with ethylenediamine tetra-acetic acid (EDTA) anticoagulant vacutainer tubes (BD Franklin Lakes NJ, USA). Genomic DNA was extracted from lymphocytes using the QIAamp DNA Blood Mini Kit (Qiagen, Berlin, Germany) and DNA samples were frozen at 280uC. Genotyping of CTLA-4 -1722T/C polymorphism was carried out using the polymerase chain reaction-ligase detection reactions (PCR-LDR) method [12]. The Shanghai Biowing Applied Biotechnology Company provides technical support for genotyping. One hundred and sixty samples were randomly selected and reciprocally tested with directly sequencing for quality control, and the reproducibility were 100%. The primers of directly sequencing used for CTLA-4 -1722T/C genotyping were as follows: F: 5' GCAATAACAACCTAAT-GGGCAC 3'; R: 5' ACTTCCACAGGCTGAACCACT 3' ( Figure S1).

Statistical analysis
Chi-square test (x 2 ) was conducted to measure the differences in the distributions of genotypes, demographic characteristics and selected variables between esophageal cancer cases and controls. Genotype frequencies of CTLA-4 -1722T/C polymorphism among the controls were tested for Hardy-Weinberg equilibrium (HWE) using an internet-based calculator (http://ihg.gsf.de/cgibin/hw/hwa1.pl). The associations between CTLA-4 -1722T/C locus and the risk of ESCC were analyzed by unconditional logistic regression for crude ORs and adjusted ORs when it was appropriate. Statistical analyses were implemented in SAS 9.1.3 software (SAS Institute, Cary, NC). A P,0.05 (two-tailed) was defined as the criterion of statistical significance.

Meta analysis
The meta-analysis is reported on the basis of the Preferred Reporting Items for Meta-analyses (PRISMA) guideline (Checklist S1) [13].
Embase, PubMed, and CBM (Chinese BioMedical Disc), as well as CNKI (China National Knowledge Infrastructure) database were searched up to August 1st, 2013 for publications investigating the association of CTLA-4 -1722T/C polymorphism with cancer risk. The combination terms were 'cancer' or 'tumor' or 'carcinoma' or 'neoplasm' and 'cytotoxic Tlymphocyte antigen 49 or 'CTLA-49 or 'CD152', annexed with 'mutation' or 'variant' or 'SNP' or 'polymorphism'. In addition, the publication language was restricted to English and Chinese, and all studies performed in human subjects were identified. The search results were supplemented by checking all references listed in these studies and published reviews. Included studies were qualified if they met the major included criteria: (1) designed as a retrospective or nested case-control study, (2) evaluated the CTLA-4 -1722T/C polymorphism and Table 3. Logistic regression analyses of associations between CTLA4 -1722T/C (rs733618) polymorphisms and risk of ESCC.    In this meta-analysis, the crude odds ratio (OR) with the corresponding 95% confidence intervals (95% CI) was used to assess the strength of association between the CTLA-4 -1722T/C polymorphism and cancer risk. The Z-test and Pvalue (two-tailed) was used to measure the significance of the pooled OR, and statistical significance was defined as P, 0.05 (two-tailed). Heterogeneity among studies was evaluated by a Chi-square-based I 2 test, I 2 ,25% indicated low heterogeneity, 25%#I 2 #50% indicated moderate heterogeneity, and I 2 .50% indicated large heterogeneity [14]. If I 2 . 50% or P,0.10, the pooled ORs were calculated by the random-effects model (the DerSimonian-Laird method), otherwise the fixed-effects model was implemented (the Mantel-Haenszel method). Subgroup analyses were implemented to measure ethnicity-specific, cancer type-specific and system-specific effects according to ethnicity, cancer type (if any cancer type evaluated by less than three individual investigations, it was combined into "other cancers") and system. The funnel plot and Egger's test were carried out to measure publication bias, which was evaluated by visual inspection of an asymmetric plot. For heterogeneity, funnel plot and Egger's test, statistical significance was considered at P,0.1. In this meta-analysis, all statistical analyses were conducted by STATA software (version 12.0).

Baseline characteristics
The demographics and risk factors of all subjects are presented in Table 1. The results indicated that cases and controls were fully matched by age and gender. However, there was significant difference on drinking status and smoking between patients and controls (P,0.001). The primary information of CTLA-4 -1722T/C polymorphism was showed in Table 2. For this SNP, the genotyping success rate was 96.43% in all samples. Minor allele frequency (MAF) of controls in our study, was similar to the database of Chinese for this SNP ( Table 2). The genotypic frequencies for CTLA-4 -1722T/C polymorphism among controls were used to evaluated deviation from the HWE, and the result was in HWE (P = 0.284) ( Table 2).

Single-locus analysis
In the single locus analyses, the genotype frequencies of CTLA-4 -1722T/C were 16.53% (CC), 49.10% (TC) and 34.37% (TT) in the patients, and 17.50% (CC), 47.79% (TC) and 34.70% (TT) in the controls, and the difference was no statistically significant (P = 0.862) ( Table 3). In this case-control study, logistic regression analyses showed that the CTLA-4 -1722T/C SNP was not associated with the risk of ESCC. Tobacco use and alcohol consumption are two strong environmental factors, we examined the association in a stratified Table 6. Distribution of CTLA-4 -1722T/C (rs733618 T/C) polymorphisms genotype and allele among multiple cancer patients and controls. analysis by these two factors and the results were null association ( Table 4).

Eligible articles for meta-analysis
The initial search yielded a total of 345 potentially relevant publications. After applying additional filters, 12 case-control studies in 11 publications and our study were eligible for inclusion. The detailed process of selecting and excluding articles is presented in Figure 1.

Meta-analysis results
After combining all qualified studies, a total of 3420 cancer cases and 3675 controls from 13 eligible case-control studies were included for meta-analysis of the association between the CTLA-4 -1722T/C polymorphism and cancer risk. There was null association of CTLA-4 -1722T/C polymorphism with overall cancer risk in all genetic models ( Table 7, Table 8,  Table 9, Figure 2, and Figure 3). In a stratified analysis by ethnicity, the similar results were observed in both Asians and Caucasians ( Table 7). In a stratified analysis by cancer type, there was a decreased risk of gastric cancer in two genetic models: CC vs. TC+TT (OR, 0.36; 95% CI, 0.19-0.66; P = 0.001) and CC vs. TT (OR, 0.45; 95% CI, 0.23-0.86; P = 0.016) ( Table 8). In a stratified analysis by system, null association was also observed ( Table 9). Tests for publication bias, sensitivity analyses, and heterogeneity In this meta-analysis, potential publication bias was detected by Begg's Funnel plot and Egger's test (Figure 4), and the shape of funnel was symmetry in all genetic model. It suggested that there were no publication bias for overall cancer in this meta-analysis ( Sensitivity analyses were carried out to detect the influence of each individual dataset on the pooled OR, with each study dataset set dropped at a time. The outcomes did not change when any individual study was omitted, suggesting the stability of our results ( Figure 5) (data not shown).
Large heterogeneities among the studies were indentified in the recessive model and homozygous model. Since tumor origin, ethnicity and system can influence the results from meta-analyses, we carried out subgroup analyses and the results were presented in Table 7, Table 8 and Table 9. The results indicated that breast cancer, digestive system cancer and Asian population subgroup may contribute to the major heterogeneity. As shown in Table 7, heterogeneity was significant in the recessive model. Further analysis was conducted by Galbraith radial plot in the recessive model ( Figure 6), and the result showed one outlier might contribute to the major sources of heterogeneity. From the forest plot in the recessive model (Figure 2), one can identify that a casecontrol study conducted by Erfani et al. [16] contributes the main heterogeneity.

Discussion
Of late, several studies have investigated the association between CTLA-4 -1722T/C polymorphism and multiple cancers, a decisive answer is lacking. In this study, a case-control study in Han Chinese population, along with a meta-analysis on overall cancer, attempted to derive a comprehensive evaluation and the results were non-significance. To the best of our knowledge, this is the first case-control study investigating the association between CTLA-4 -1722T/C polymorphism and esophageal cancer risk.
Cancer and autoimmune disease are both multifactorial disorders that results from complex interactions between genetic backgrounds and environmental factors. The CTLA-4 -1722T/C  Table 7. Summary of results of the meta-analysis from different comparative genetic models in the subgroup analysis by ethnicity.  polymorphism (TRC) would reduce a transcription factor binding site for nuclear factor 1 and weaken the expression of cell surface CTLA-4 [11,25], which might play an important role in cancer and autoimmune disease susceptibility. Several meta-analyses showed that CTLA-4 -1722T/C polymorphism might be a risk factor for systemic lupus erythematosus susceptibility [26][27][28][29]. However, the association between this locus and cancer risk was inconclusive. With a growing interest in the associations of genetic Table 9. Summary of results of the meta-analysis from different comparative genetic models in the subgroup analysis by system.  polymorphisms and cancer, several studies have examined the hypothesis that CTLA-4 -1722T/C polymorphism is relevant to the risk of a number of cancers; however, the results remain elusive. Considering the fact that most common SNPs usually make low penetrance cancer susceptibility, this study includes 13 case-control studies with relatively large sample sizes to obtain a precise evaluation between CTLA-4 -1722T/C genetic variation and cancer risk. One individual study has reported positive signal of CTLA-4 -1722T/C polymorphism with cancer [18]; the other individual study has reported negative signal [20]; however, as demonstrated in our overall genetic model results among 7098 subjects, there were non-significance, even in different population subgroups and different system. In a stratified analysis by cancer type, the protective effect conferred by the recessive model and homozygous model was appreciably obvious in gastric cancer subgroup. Considering only three case-control studies were conducted in gastric cancer subgroup and these studies were small sample sizes, which might restrict power to confirm a real influence or generate a fluctuated assessment. All results should be interpreted with very caution. It is also possible that the potential function of this polymorphism is diluted or covered by other genetic background or environment factors, and these important factors should not be ignored. Considering only 13 case-control studies were recruited in this meta-analysis and most of these  CTLA-4 -1722T/C Polymorphism and Cancer Risk studies were small sample sizes, in the future, further investigations with large sample sizes should be carried out to confirm or refute these results. Some merit of current study should be adequate consideration. First, this is to date the first case-control study detecting the association of CTLA-4 gene -1722T/C polymorphism with esophageal cancer. Second, the findings of our case-control study conform to that of the subsequent meta-analysis. Third, in our case-control study, control genotype distributions were consistent with HWE showed our results were less prone to selection bias, the shape of funnel plot indicated that there were no publication bias in current meta-analysis. Fourth, relatively low heterogeneity was observed between publications for CTLA-4 -1722T/C polymorphism.
In addition, some limitations in current study should be acknowledged when interpreting our results. First, in this casecontrol study, all cases and controls were recruited from two hospitals and might not fully represent the general Chinese populations. Second, all included case-control studies for metaanalysis were from Asians and Caucasians; thus, our findings might only be suitable for these two populations. Third, only published studies were recruited in this meta-analysis, publication bias might have inevitably occurred. Fourth, due to the lack of uniform background data for recruited studies, data were not further stratified by other factors (such as, age, gender, smoking, alcohol consumption, and other lifestyle factors). Fifth, in this study, we focused on only -1722T/C polymorphism in CTLA-4, and did not consider other susceptibility genes or polymorphisms. For the low penetrance cancer susceptibility gene effects from SNP, these important genetic and environmental factors should be adequately considered.
In summary, this case-control study along with a meta-analysis, failed to confirm the association between CTLA-4 -1722T/C polymorphism and cancer risk, even across different ethnic subgroups and different systems. In the future, further investigations with large sample sizes and detailed gene-environment data, should be carried out to confirm or refute these results. Figure S1 Direct sequencing analyses for genotypes of CTLA-4 -1722T/C SNP (The three charts represent three genotypes).

(TIF)
Checklist S1 PRISMA checklist, Checklist of items to include when reporting a systematic review or meta-analysis (diagnostic review consisting of cohort studies).