Association of the PSCA rs2294008 C>T Polymorphism with Gastric Cancer Risk: Evidence from a Meta-Analysis

GC is the fourth most common malignancy and the second leading cause of cancer associated death worldwide (Ferlay et al., 2010), so it is the major public health problem faced all around the world. The incidence of gastric cancer is regulated by multi-factors, including oncogene tumor suppressor gene in abnormal activation/ inactivation, immune changes, and etc. With the depth of new technologies, many efforts have been invested in identifying sources of genetic susceptibility to cancer in the last few decades (Dong et al., 2008; Lao-Sirieix et al., 2010). PSCA gene encodes a 123-amino acid glycoprotein, which is a cell surface antigen. Its name was inaccurate, because it is not exclusively expressed in prostate cancer; and highly expressed by a large proportion of human tumors, including metastatic and hormone-related cancers (Reiter et al., 1998; Bahrenberg et al., 2000). Recently, scientists found that PSCA gene is either abnormally expressed or shown different genetic variant in several cancers. GWAS have identified a strong association between PSCA gene rs2294008 C > T polymorphism and risk of GC, especially in Asians (Sakamoto et al., 2008; Shi et al., 2011). Moreover, its genetic variant may correlated with other risk factors or clinicopathologic features, such as sex, age, smoking status, tumor size or stage, location


Introduction
GC is the fourth most common malignancy and the second leading cause of cancer associated death worldwide (Ferlay et al., 2010), so it is the major public health problem faced all around the world. The incidence of gastric cancer is regulated by multi-factors, including oncogene tumor suppressor gene in abnormal activation/ inactivation, immune changes, and etc. With the depth of new technologies, many efforts have been invested in identifying sources of genetic susceptibility to cancer in the last few decades (Dong et al., 2008;Lao-Sirieix et al., 2010).
PSCA gene encodes a 123-amino acid glycoprotein, which is a cell surface antigen. Its name was inaccurate, because it is not exclusively expressed in prostate cancer; and highly expressed by a large proportion of human tumors, including metastatic and hormone-related cancers (Reiter et al., 1998;Bahrenberg et al., 2000). Recently, scientists found that PSCA gene is either abnormally expressed or shown different genetic variant in several cancers. GWAS have identified a strong association between PSCA gene rs2294008 C > T polymorphism and risk of GC, especially in Asians (Sakamoto et al., 2008;Shi et al., 2011). Moreover, its genetic variant may correlated with other risk factors or clinicopathologic features, such as sex, age, smoking status, tumor size or stage, location

Association of the PSCA rs2294008 C>T Polymorphism with Gastric Cancer Risk: Evidence from a Meta-Analysis
Qing-Hui Zhang*, Yong-Liang Yao, Tao Gu, Jin-Hua Gu, Ling Chen, Yun Liu and histological type of GC. More recently, several studies have assessed the relationship between the polymorphism of PSCA C > T and the susceptibility to GC. Nevertheless, controversial results exist regarding the association of PSCA with the risk of GC (Sakamoto et al., 2008;Matsuo et al., 2009;Wu et al., 2009;Lu et al., 2010;Ou et al., 2010;Lochhead et al., 2011;Song et al., 2011;Zeng et al., 2011;Sala et al., 2012). Therefore, we conducted this meta-analysis to confirm these associations.

Identification and eligibility of relevant studies
Search terms "PSCA", "gastric cancer", "genotype", "polymorphism" and "variant" were employed to explore publications in PubMed, ISI Web of Knowledge and Embase databases for relevant reports (last search update February 2012). Only studies published in the English language were included. We did not define any minimum number of patients to be included for meta-analysis. When multiple studies of the same patient population were identified, we included the published report with the largest sample size.

Inclusion and exclusion criteria
The following inclusion criteria were used to select literatures for this analysis: (a) evaluation of the PSCA rs2294008 C > T polymorphism and GC, (b) only the casecontrol studies were considered, (c) sufficient published data for estimating an OR with 95% CI. Major exclusion criteria were: (1) no control population, (2) no available genotype frequency, and (3) duplicated studies.

Data extraction
For each study, the following data were collected: first author's surname, year of publication, country of origin, ethnicity, criteria of enrolled patients, genotyping method, total numbers of cases and controls as well as numbers of cases and controls with CC, CT and TT genotypes. The strength of the association between PSCA rs2294008 C > T polymorphism and GC risk was estimated using OR, with the corresponding 95% CI. Disagreement was resolved by discussion until a consensus was reached between the two authors. We did not define any minimum number of patients for inclusion in our meta-analysis.

Statistical methods
The risk of GC associated with PSCA rs2294008 C > T was estimated for each study by OR with 95% CI. For all studies, we estimated the association under five different types of ORs, namely the allele contrast model (C vs T), the homozygote codominant model (CC vs TT), the heterozygote codominant model (CT vs TT), the dominant model (CC+CT vs TT) and the recessive model (CC vs CT+TT). HWE was tested by the Chi-square test. The Q-statistic was used to investigate the degree of heterogeneity between the trials, and a P-value 0.10 for the Q-test indicated a lack of heterogeneity among studies. We used the fixed-effects model and the random-effects model based on the Mantel-Haenszel method (Jose et al., 2008) and the DerSimonian and Laird method (Kjellsson et al., 2008), respectively, to combine values from each of the studies. A sensitivity analysis was also performed by omitting each study in turn to identify potential outliers. All of the statistical analyses were performed with Review Manage version 4.3 and STATA version 12.0 using twosided P-values.

Study characteristics
We obtained 14 studies about the association between PSCA rs2294008 polymorphism. Following the above    inclusion and exclusion criteria, 9 publications were included in the final meta-analysis (Sakamoto et al., 2008;Matsuo et al., 2009;Wu et al., 2009;Lu et al., 2010;Ou et al., 2010;Lochhead et al., 2011;Song et al., 2011;Zeng et al., 2011;Sala et al., 2012). Characteristics of studies focusing on PSCA rs2294008 C > T are summarized in Table 1. Genotypes and separate P values for PSCA rs2294008 polymorphism are list in Table 2. When we conducted the above analysis, Lochhead's study about Polish was excluded due to its HWE disequilibrium (P = 0.01) in controls. Table 3 lists the main results of pooled ORs for PSCA rs2294008 C > T polymorphism and gastric cancer risk. We observed that the PSCA rs2294008 C > T polymorphism was significantly correlated with gastric cancer risk when all studies were pooled into the meta-analysis (the homozygote codominant model: OR = 0.70, 95% CI = 0.50-0.97, P = 0.28 for heterogeneity; the recessive model: OR = 0.70, 95% CI = 0.55-0.89, P = 0.13 for heterogeneity, Figure 1). Further subgroup analysis showed that rs2294008 C > T polymorphism is

> T and Gastric Cancer Risks In Asian And Chinese Population (for the recessive model).
The study is shown by the point estimate of the OR (the size of the square is proportional to the weight of each study) and 95% CI for OR; pooled OR and 95% CI have been appropriately derived from the random and fixed effects model, respectively 95% CI = 0.71-0.87, P = 0.35 for heterogeneity, Figure  4).

Publication bias
We performed Begg's funnel plot and Egger's test to assess the publication bias of literatures. The results did not show any evidence of publication bias in all the comparisons. Also, the results of Egger's test still did not suggest any evidence of publication bias (P = 0.679 for T vs G; P = 0.134 for GT vs GG; P = 0.104 for TC vs. TT, respectively).

Discussion
PSCA was initially identified as a prostate-specific cell-surface antigen (Argani et al., 2001). PSCA is largely expressed in normal prostate and overexpressed in a majority of prostate cancer. Meanwhile, PSCA is highly expressed in other non-prostatic malignancies, including bladder cancer, pancreatic cancer, and gastric cancer Grubbs et al., 2006). Recently studies have reported that its expression is different in these tumor and adjacent normal tissues, but its gene polymorphisms are significant discrepancies Feng et al., 2008). Among these variants, rs2294008 C > T and rs2976392 G > A are explicit, while these significant relationships were observed in several but not all studies, especially in gastric cancer (Ou et al., 2010). The rs2294008 was shown to be functionally associated with downregulation of PSCA and to increase the risk of gastric cancer, The aim of our study was to demonstrate the role of PSCA rs2294008 C > T polymorphism in the relationship with gastric cancer risk using the meta-analysis. In present study, we observed significant differences among the two major types GC. The effect of the rs2294008 C > T polymorphism was greater in diffuse than in intestinaltype gastric cancer, consistent with previous reports. The genetic susceptibility factors may contribute more to diffuse-type gastric cancer development compared to intestinal-type gastric cancer (Lao-Sirieix et al., 2010), so genetic variations appear to be associated with an increased risk for diffuse-type gastric cancer.
Interestingly, we also observed that PSCA rs2294008 C > T was significantly correlated with risk for noncardia gastric cancer rather than cardia gastric cancer. As well known, H. pylori is the explicit risk factor for gastric cancer, and play an important role in the process of gastritis, intestinal metaplasia, gastric ulcers, and gastric cancer (Tan et al., 2011). Previous researches have reported that rs2294008 C > T polymorphism is a predisposing genetic variant for H. pylori positive gastric cancer. H. pylori infection was strongly associated with the risk of noncardia gastric cancer, but was inversely associated with the risk of cardia gastric cancer (Kamangar et al., 2006). Similar to these studies, our study is shown that PSCA rs2294008 C > T polymorphism is correlated with noncardia gastric cancer, but is not interrelated with cardia gastric cancer, for several genetic model, such as for allele contrast model, homozygote codominant model, dominant model and recessive model.
Simultaneously there was an evidence to indicate that PSCA rs2294008 C > T polymorphism was associated with increased risk of gastric cancer in Asian and Chinese  DOI:http://dx.doi.org/10.7314/APJCP.2012.13.6.2867 Association of the PSCA rs2294008 C>T Polymorphism with GC Risk -a Meta-analysis population only for recessive genetic model.
In summary, our study shows a genetic association between PSCA rs2294008 C > T variant increases susceptibility to GC, whereas we need to be further warranted to verify these findings in large populationbased prospective studies with ethnically diverse populations, as well as biologically functional studies.