Abstract
Background/Aim: Genome-wide association studies revealed a link between gastric cancer (GC) and single nucleotide polymorphisms (SNPs) of prostate stem cell antigen (PSCA), phospholipase C epsilon-1 (PLCE1) and mucin-1 (MUC1) genes. Herein, we aimed to evaluate associations between PSCA (C>T, rs2294008; G>A, rs2976392), MUC1 (C>T, rs4072037) and PLCE1 (A>G, rs2274223) SNPs and GC or high-risk gastritis (HRAG). Materials and Methods: Using TaqMan system, SNPs were genotyped in 252 patients with GC, 136 patients with HRAG and 246 controls. Results: PSCA rs2294008 allele T was linked with risk of GC (odds ratio (OR)=1.88, p<0.001) and HRAG (OR=1.49, p=0.009). Allele A of PSCA rs2976392 was associated with development of GC (OR=1.88, p<0.001) and HRAG (OR=1.56, p<0.01). MUC1 rs4072037 allele G was protective against development of GC (OR=0.64, p=0.0005), while no differences were found for PLCE1 rs2274223. Conclusion: Polymorphisms of PSCA (rs2976392, rs2294008) and MUC1 (rs4072037) genes are linked with GC and HRAG.
Gastric cancer (GC) is one of the most common causes of cancer-related death worldwide (1). Most recent reports on cancer reveal a declining incidence of GC in Western countries; nevertheless, the mortality associated with GC remains very high (1). The pathways of GC pathogenesis have merged into a complex picture, which involves Helicobacter pylori infection as well as genetic, epigenetic and other co-founding factors (2). A genome-wide association study (GWAS) approach has been applied in various diseases including cancer. The first GWAS on GC was published in 2008 and identified an association with a single nucleotide polymorphism (SNP) of prostate stem cell antigen (PSCA) gene (3). A few GWAS studies confirmed this association and revealed new susceptibility loci at mucin-1 (MUC1) and phospholipase C epsilon-1 (PLCE1) genes (3, 4).
Gene SNPs detected in the aforementioned GWAS studies have been implicated in various cancer-related molecular pathways. PSCA gene encodes a cell membrane glycoprotein responsible for cellular activation (3). PSCA gene is expressed in the epithelium of the stomach and it is particularly down-regulated in gastric tissue with intestinal metaplasia (3). MUC1 gene encodes a cell membrane protein that plays a role in forming protective mucous barriers on stomach epithelial surfaces and is essential in intracellular signaling (5). Different studies have shown that MUC1 is involved in the regulation of H. pylori-induced chronic gastritis (5). PLCE1 gene encodes an enzyme that catalyzes hydrolysis which generates second messengers affecting the cell cycle (6). The PLCE1-related signaling network has an impact on several critical carcinogenetic processes, including proliferation, cell survival, metabolism, and tumor growth (6). The changes of encoding sequences of these genes have functional roles and may influence the process of carcinogenesis.
Due to the large number of traits analyzed, GWAS results need replication in independent cohorts of patients in different populations for validation of these findings. Most of the replication studies on GWAS results in patients with GC have been carried-out on Asian populations, while representative data on Europeans are lacking.
In the study, we analyzed four SNPs reported in GC GWAS studies: PSCA G>A (rs2976392; NM_005672.4: c.133+80G>A), PSCA C>T (rs2294008; NM_005672.4:c.-26C>T), MUC1 A>G (rs4072037, NM_001204295.1, c.93G>A, pThr31=) and PLCE1 A>G (rs2274223, NM_016341.3:c.5780A>G, p.His1925Arg). The study population consisted of 252 patients with GC, 136 patients with high-risk gastritis (HRAG) and 246 controls from Lithuania and Latvia, countries which are still distinguished by a high prevalence of H. pylori infection and GC (7, 8). Only two studies have previously analyzed the role of PSCA SNPs in patients with GC precursor state, i.e. atrophic gastritis (9, 10). Furthermore, MUC1 A>G (rs4072037) and PLCE1 A>G (rs2274223) gene polymorphisms have not previously been studied in premalignant gastric conditions.
Materials and Methods
Study population. Patients and controls were recruited during 2006-2013 at two gastroenterology Centers in Lithuania (Department of Gastroenterology, Lithuanian University of Health Sciences, Kaunas) and Latvia (Riga East University Hospital). HRAG patients and controls were included from the out-patient departments of these Institutes. The inclusion criterion for controls was no history of previous malignancy. Patients with HRAG were defined as in our previous genotyping articles (11, 12) based on criteria suggested by Uemura et al. (13) and Meining et al. (14). Patients with HRAG had undergone upper endoscopy with thorough histological analysis of stomach mucosa according to Sydney classification (15). All patients with GC had histological verification of gastric adenocarcinoma and were recruited from out-patient and in-patient departments. H. pylori status was determined with anti-H. pylori IgG antibodies in serum.
In total, 634 individuals were included in the study (246 controls, 136 HRAG and 252 GC); 285 came from the Latvian group (100 controls, 127 GC and 58 HRAG) and 349 from the Lithuanian group (146 controls, 125 GC and 78 HRAG). All patients in the present study were of European descent. The study was approved by the Ethics Committees of the Lithuanian University of Health Sciences (Protocol Nr. BE-2-10) and Central Medical Ethics Committee of Latvia (Protocol Nr. 01-29.1). All patients and controls gave their informed consent to take part in the study.
DNA extraction and genotyping. Genomic DNA from samples was extracted from peripheral blood mononuclear cells using a salting-out method and stored at −20°C until analysis as described in our previous studies (12, 14). PSCA G>A (rs2976392), PSCA A>G (rs2294008), MUC1 A>G (rs4072037) and PLCE1 A>G (rs2274223) SNPs were genotyped by real-time PCR (RT-PCR), using TaqMan® assays with a 7500™ real-time cycler, in accordance with the manufacturer's instructions (Life Technologies, Carlsbad, California, USA). Dubious samples underwent repetitive genotyping analysis.
Characteristics of the participant groups within the study.
Statistical analysis. Age is shown as means with standard deviations, and was compared using ANOVA and unpaired Student's t-test. Statistical analysis of the genotyping data was performed using PLINK software version 1.07 (16). All SNPs underwent assessment for Hardy–Weinberg equilibrium (HWE). Association of HRAG and GC with gene polymorphisms was calculated using logistic regression analysis with adjustment for age, gender, and country of birth, with 95% confidence intervals (CI). Disease risk related to SNPs was assessed using genotypic, allelic, recessive and dominant models. p-Values were adjusted for multiple testing (significance threshold α=0.0125; 0.05/4).
Results
Characteristics of participants. All participants were classified into three study groups: controls (n=246), HRAG (n=136) and GC (n=252). The characteristics of control, GC and HRAG groups are presented in Table I. The participants differed significantly regarding age and gender distribution between the groups. Males were predominant in GC and accounted for 61.7% in the GC group, while in the control and HRAG groups they constituted 23.8% and 30.9%, respectively. Controls were of similar age to HRAG and GC groups (Table I). To eliminate the potential bias of differences in age and gender distribution among the groups, these parameters were included as co-variates in further logistic regression analysis. A total of 40.9% of patients were positive for H. pylori in the GC group; however, H. pylori IgG status was not available in around 47.6% of individuals in the GC group. Over 78.1% of those in the control group and 77.9% in the HRAG group were H. pylori-positive. Histological classification for diffuse and intestinal GC types was available for 73.4% of patients with GC (Table I).
Genotypic and allelic frequencies of prostate stem cell antigen (PSCA), mucin-1 (MUC1) and phospholipase C epsilon-1 (PLCE1) single nucleotide polymorphisms (SNPs) in controls, and patients with gastric cancer (GC) and high-risk gastritis (HRAG).
Hardy-Weinberg equilibrium. Overall, samples from six individuals failed genotyping analyses for PSCA rs2294008, 21 for PSCA rs2976392, 23 for MUC1 rs4072037 and eight for PLCE1 rs2274223 repeatedly and were excluded from the study. All SNPs within the study were in HWE: PSCA rs2294008, p=0.807; PSCA rs2976392, p=0.511; MUC1 rs4072037, p=0.168; PLCE1 rs2274223, p=0.733.
PSCA, MUC1 and PLCE1 gene polymorphisms in GC and HRAG. The frequencies of alleles and genotypes of PSCA rs2294008 and rs2976392, MUC1 rs4072037, PLCE1 rs2274223 in control, GC and HRAG groups is presented in Table II. Allele T of PSCA rs2294008 was less frequent in controls (48.4%) than in patients with GC (63.7%, odds ratio (OR)=1.88, p=1.1×106) and with HRAG (58.2%, OR=1.49, p<0.01). Similarly, TT genotype of PSCA rs2294008 was more frequent among patients with GC (40.6%, OR=3.70, p<0.001) and HRAG (33.5%, OR=2.21, p=0.012) than in controls (23.1%). PSCA rs2976392 was noted for a higher frequency of allele A in GC (63.6%, p<0.001) and HRAG groups (59.1%, p=0.004) when compared to controls (48.3%). Genotype AA of PSCA rs2976392 was associated with higher risk of GC both in dominant (p=0.0002) and recessive models (p<0.001) (Table II). MUC1 rs4072037 allele G was associated with decreased risk of GC (OR=0.64, p=00005). Genotype GG of MUC1 rs4072037 was also linked with lower risk of GC in AA vs. GG (OR= 0.39, p=0.004), as well as in the dominant model (OR=0.39, p<0.001). A similar link of MUC1 rs4072037 was also detected in patients with HRAG, where GG genotype (23.8%) was less frequent than in controls (OR=0.42%, p=0.007). PLCE1 rs2274223 alleles and genotypes were distributed equally among control, GC and HRAG groups (Table II).
Association analysis of PSCA, MUC1 and PLCE1 SNPs with diffuse and intestinal-type GC. Allelic and genotypic distribution for the four investigated SNPs among controls and different histological subtypes of GC (intestinal and diffuse) are presented in Table III. This sub-analysis was limited only to those patients for whom histological classification of GC subtype was available. Allele T of PSCA rs2294008 was more frequent in diffuse-type GC (59.6%, p=0.017) than in controls (23.1%), but this tendency was much more profound for intestinal-type GC (66.5%, OR=2.13, p<0.001). A similar observation was made for the TT genotype, which was associated with higher risk of intestinal-type GC both in dominant (OR=2.22, p=0.0049) and recessive (OR=4.17, p=0.001) comparisons. Allele A of PSCA G>A SNP was more common in diffuse-type GC (58.8%) vs. controls (51.7%, p=0.026). However, comparison of allele A frequencies revealed greater differences between controls (51.7%) and intestinal-type GC (66.7%, OR=2.17, p=0.001). Intestinal-type GC was also denoted by higher frequency of AA genotype (42.0%) than controls (26.7%, p=0.0009). MUC1 A>G allele G was associated with reduced risk of diffuse-type GC (OR=0.52, p=0.0006), but not of intestinal-type GC (OR=1.31, p=0.136). Association analysis also showed that GG MUC1 genotype was associated with reduced risk of diffuse-type GC in a dominant model (OR=0.34, p=0.0003). PLCE1 A>G SNP was not associated with the presence of intestinal- or diffuse-type GC in all statistical models that were applied, with similar distributions of alleles and genotypes among cases and controls (Table III).
Association analysis between H. pylori and PSCA, MUC1 and PLCE1 SNPs. The analysis of genotypic and allelic distribution among H. pylori-positive and -negative individuals for four SNPs of PSCA, MUC1 and PLCE1 genes is presented in Table IV. Overall, alleles and genotypes in this sub-analysis were distributed equally among H. pylori-positive and - negative individuals for all polymorphisms that have been investigated in this study. There was a tendency for higher distribution of GG genotype (20.4%) and G allele (46.2%) in H. pylori-positive individuals when compared to controls (14.5% and 42.7%, respectively); however, the significance remained below the required threshold (Table IV).
Discussion
Multiple SNPs have been linked with gastrointestinal malignancies in recent GWAS studies (17), but the validation data in different populations are still scarce. In the current study, we evaluated PSCA C>T rs2294008, PSCA G>A rs2294008, MUC1 A>G rs4072037 and PLCE1 A>G (rs2274223) SNPs in an Eastern European cohort with GC and premalignant gastric conditions. The major finding of the current study is the significant association of PSCA and MUC1 SNPs not only with GC, but also with its precursor states, elucidating potential genetic predisposition that might be involved in early stages of gastric cancer development. To the best of our knowledge, there are only two studies that have assessed the role of these PSCA SNPs in patients with GC precursor state, i.e. atrophic gastritis (9, 10). Furthermore, this is the first article to analyze MUC1 A>G (rs4072037) and PLCE1 A>G (rs2274223) SNPs in pre-malignant gastric conditions.
From all SNPs investigated in our study, PSCA gene polymorphisms rs2294008 and rs2976392 have probably been extensively studied most. It is well-known that these polymorphisms are in linkage disequilibrium within the PSCA gene (3). These two SNPs were linked with diffuse-type GC in GWAS (3). Later studies confirmed this association in other populations (4, 9, 10), and suggested that these genetic alterations are also linked with intestinal-type GC (18). rs2294008 has been linked with the risk of different types of cancer, including bladder cancer (19) and esophageal cancer (20). A recent meta-analysis by Gu et al. including 18,820 GC cases and 35,766 controls from 16 studies confirmed the association for rs2294008 and rs2976392 of PSCA gene (21). Furthermore, a study by Tanikawa et al. showed a link between these PSCA SNPs with the presence of duodenal ulcer (22). Functional studies have shown that rs2294008 influences transcriptional activity and may even have a clinical implication in the treatment of bladder cancer (23). The potential implications of PSCA SNPs in the clinical setting are of interest for further research in GC. Based on the currently available data, these PSCA SNPs are clearly linked with development of GC and the results of our present study support these findings. Our data suggest that TT genotype for rs2294008 and AA genotype for rs2976392 are linked with different histological subtypes of GC; however, the strength of association was higher for intestinal-type GC. It is worth mentioning that our results are in line with previously published studies suggesting that these SNPs are important in the early stages of GC development (9, 10).
Genotypic and allelic frequencies of prostate stem cell antigen (PSCA), mucin-1 (MUC1) and phospholipase C epsilon-1 (PLCE1) single nucleotide polymorphisms (SNPs) in controls, and patients with diffuse-type and intestinal-type gastric cancer (GC).
MUC1 A>G SNP (rs4072037) has been linked with increased risk of diffuse-type gastric cancer, with an OR of 1.66 in GWAS (24). Several publications have linked this genetic variation of MUC1 with other cancer types, including ovarian (25) and esophageal (26) cancer. A recent large meta-analysis study of almost 6,580 GC cases and 10,324 controls has confirmed this association, both in Asian and European populations (27). Our results clearly confirm those findings. Subgroup analysis of samples according to subtype of GC showed that G allele and GG genotype have been associated with reduced risk of diffuse-type GC, while association with intestinal type-GC was marginal. The most important finding from our analyses is the significant association of MUC1 rs4072037 with advanced precancerous gastric mucosal alterations, HRAG, which has not been assessed previously.
Genotypic and allelic frequencies of prostate stem cell antigen (PSCA), mucin-1 (MUC1) and phospholipase C epsilon-1 (PLCE1) single nucleotide polymorphisms (SNPs) among Helicobacter pylori-positive and -negative individuals.
The studies conducted mostly in Asian populations have linked rs2274223 of PLCE1 gene with GC and esophageal cancer (28, 29); these findings were confirmed in subsequent replication studies (26) and recent meta-analyses (30, 31), as well as being linked with survival in pw GC (32). We did not observe any significant association between PLCE1 rs2274223 and GC or its precursors. The results of our study show a similar distribution of rs2274223 genotypes and alleles between GC, HRAG and control groups. Differences between our findings and those of previous studies might be related, at least partially, to ethnical differences as this is a common phenomenon in genetic association studies, while further highlighting the need for validation analyses in different population cohorts (33).
In a certain genetic background, H. pylori infection may trigger GC development probably at different levels, and following this hypothesis, only a small proportion of individuals infected with H. pylori infection actually do develop GC (2). Since significant associations between PSCA and MUC1 SNPs with risk of premalignant gastric conditions have been observed, we further evaluated whether these genetic alterations are linked to H. pylori infection. One study has shown that MUC1 rs4072037 and PLCE1 rs2274223 SNPs may increase GC risk in combination with H. pylori seropositivity (34). The analysis of genotypic and allelic distribution among H. pylori-positive and H. pylori-negative individuals did not reveal any significant differences between these two groups. A recent GWAS has revealed certain genetic predisposition for H. pylori seropositivity (35). Based on our results, shared association of GC and HRAG with PSCA and MUC1 gene SNPs is not H. pylori-dependent but might be related to other factors or related carcinogenic pathways.
It is worth pointing-out that the frequencies of genotypes observed in our study for all four SNPs are in line with those of previous studies and PubMed SNP database for European populations (9, 26, 33). Our present study carries certain limitations. The number of individuals is not large in our case–control study; therefore, certain tendencies that have been observed in this study could gain significance in a larger group of patients. There were differences in age and gender distribution between GC, HRAG and control groups, but these factors were included as covariates in the statistical logistic regression analysis in order to eliminate potential influence. Because of the low number of participants, we were not able to stratify risk with respect to subgroups of GC in terms of survival, TNM stage, tumor localization and differentiation grade. In addition, here we provided only an insight into the association between SNPs and clinicopathological characteristics, while further studies are needed to evaluate the interplay with gene expression in the background of those genotypes and alleles. Indeed, our results will provide additional data for further analysis on the topic, especially for PLCE1 rs2274223 which has been reported to have more inconsistent results.
Conclusion
SNPs of PSCA (rs2976392, rs2294008) and MUC1 (rs4072037) genes are linked with susceptibility of GC and HRAG, while no significant association was determined for PLCE1 rs2274223.
Acknowledgements
The Authors would like to thank Raminta Grigaliunaite for help with genotyping in the laboratory. This project was supported by a grant from the Research Council of Lithuania No. MIP-14418 and in part by a grant from the BMBF-0315905D in the frame of ERA-Net PathoGenoMics.
- Received July 27, 2014.
- Revision received October 25, 2014.
- Accepted October 30, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.