Single-nucleotide polymorphisms in PSCA and the risk of breast cancer in a Chinese population

This study explored the associations between common PSCA single-nucleotide polymorphisms (rs2294008, rs2978974, and rs2976392) and breast cancer among 560 breast cancer cases and 583 controls (Chinese Han women). We found rs2294008 was significantly associated with a high risk of breast cancer (homozygote model, odds ratio [OR]: 1.67, 95% confidence interval [CI]: 1.06–2.59; recessive, OR: 1.64, 95% CI: 1.06–2.53). And stratification by menopausal status revealed an association of the minor allele of rs2294008 with breast cancer risk among premenopausal (homozygote model, OR: 2.41, 95% CI: 1.03–5.66; recessive, OR: 2.80, 95 % CI: 1.21–6.47) and postmenopausal women (allele model, OR: 1.29, 95% CI: 1.01–1.65). Rs2978974 influenced the breast cancer risk among postmenopausal women in heterozygote model (OR: 1.47, 95% CI: 1.05–2.07). When stratified by clinicopathologic features, the T allele of rs2294008 was associated with progesterone receptor status (homozygote model, OR: 1.98, 95% CI: 1.08–3.63; recessive, OR: 1.87, 95% CI: 1.04–3.37), and the rs2976392 polymorphism was associated with high lymph node metastasis risk in homozygote model (OR: 2.09, 95%CI: 1.01–4.31). Further haplotype analysis suggested that Trs2294008 Ars2976392 Grs2978974 haplotype enhances breast cancer risk (OR:1.52, 95%CI:1.23-1.89, P<0.001). Therefore, among Chinese Han women, the PSCA rs2294008, rs2978974, and rs2976392 minor alleles are associated with increased breast cancer risk especially in progesterone receptor positive breast cancer patients, with breast cancer risk in postmenopausal women, and with high lymph node metastasis risk, respectively. Moreover, Trs2294008 Ars2976392 Grs2978974 haplotype was associated with significantly increased risk of breast cancer.


INTRODUCTION
Breast cancer is the most common cancer and the principal cause of cancer-related deaths among Chinese women [1], accounted for 248,620 new cases and 60,473 cancer-related deaths during 2011 [2]. It is a multifactorial disease influenced by complex interactions between genetic, environmental, and lifestyle factors [3]. Genetic research provides insight into carcinogenesis, including the development and treatment of breast cancer.
Single-nucleotide polymorphisms (SNPs) are variations in a single base pair in the DNA sequence and have been widely studied in cancer research in recent years. Several genes that affect breast cancer risk, including BRCA1 (breast cancer 1), BRCA2 (breast cancer 2), PTEN (phosphatase and tensin homolog deleted on chromosome ten), and TP53 (tumor protein p53) have been identified [4][5][6][7][8].
PSCA encodes a 123-amino acid immature lymphocyte cell surface maker with 30% homology to stem cell antigen type 2, a member of the Thy-1/Ly-6 family and is located on chromosome 8q24.2 [10]. PSCA was initially identified as a prostate-specific antigen over-expressed in >80% of prostate cancers, including metastatic and hormone-related cancers [10,11]. Recent studies have shown that PSCA is also abnormally expressed in bladder cancer [12][13][14], gastric cancer [15][16][17], renal cell carcinoma [18], oesophageal cancer [19], gallbladder cancer [20][21][22], and pancreatic cancer [23]. Studies in vitro indicated that PSCA being transfected into PSCA-negative cells caused down-regulated cell proliferation, thus affecting survival of gastric cancer cells [24]. And, down-regulation of PSCA in a human bladder cancer cell line led to inhibition of cell growth via activation of several immune signaling pathways [25]. Genome-wide association studies have revealed many PSCA polymorphisms, among which rs2294008 C>T, rs2978974 G>A, and rs2976392 G>A are the most widely studied ones and may influence susceptibility to different types of cancer [22,26,27]. However, few studies have been performed to investigate the associations of these three PSCA SNPs with breast cancer. A single study with small sample sizes (456 patients and 461controls) revealed that the PSCA SNPs were associated with breast cancer susceptibility among Korean women [9]. Therefore, the present study aimed to comprehensively examine the potential association of three SNPs (rs2294008 C>T, rs2978974 G>A, and rs2976392 G>A) in PSCA with the risk of breast cancer among a population of Chinese women.

Associations between PSCA SNPs and the risk of breast cancer
Detailed allele frequencies and genotype distributions of the three polymorphisms are shown in Table 1. The distributions of rs2294008, rs2978974, and rs2976392 in the control group were in accordance with Hardy-Weinberg equilibrium (P = 0.195, P = 0.164, and P = 0.179, respectively). Both the homozygote and recessive models of rs2294008 revealed an associated with a high risk of breast cancer (TT vs. CC, odds ratio [OR]: 1.67, 95% confidence interval [CI]: 1.06-2.59, P = 0.03; TT vs. CC+TC, OR: 1.64, 95% CI: 1.06-2.53, P = 0.02). We further calculated the power of the rs2294008 SNP homozygote and recessive model analyses, and we were able to reject the null hypothesis that the TT frequency for case and controls is equal with probability (power) = 0.896. No significant associations with rs2976392 and rs2978974 were found in any of the models.

Subgroup analyses according to age and menopausal status
Stratification analyses according to age revealed no significant associations between the three PSCA SNPs and the risk of breast cancer (all, P > 0.05) ( Table 2). While, stratification analyses according to menopausal status (Table 3) found that the minor allele of rs2294008 was a risk factor among both premenopausal women (homozygote model, OR: 2.41, 95% CI: 1.03-5.66, P = 0.04; recessive model, OR: 2.80, 95% CI: 1.21-6.47, P = 0.01) and postmenopausal women (allele model: OR: 1.29, 95% CI: 1.01-1.65, P = 0.04). For rs2978974, a significant association with high breast cancer risk was found among postmenopausal women in the heterozygote model (OR: 1.47, 95% CI: 1.05-2.07, P = 0.03). There were no significant associations with rs2976392 in any of the subgroups.

Associations between PSCA SNPs and the clinicopathological features of breast cancer
We evaluated the associations of PSCA SNPs with various clinicopathological features including: tumor size, lymph node metastasis, and the expressions of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER-2). The T allele of rs2294008 was associated with positive PR status (homozygote model, OR: 1.98, 95% CI: 1.08-3.63, P = 0.03; recessive model, OR: 1.87, 95% CI: 1.04-3.37, P = 0.03) ( Table 4). The minor allele of rs2976392 was associated with a high risk of lymph node metastasis in the homozygote model (OR: 2.09, 95% CI: 1.01-4.31, P = 0.04). However, rs2978974 was not significantly associated with any of the clinicopathological features.

Association between PSCA haplotypes and breast cancer risk
We analyzed the association between PSCA haplotypes and the risk of breast cancer. Table 5 shows that T rs2294008 A rs2976392 G rs2978974 haplotype was associated with a significantly increased risk of breast cancer (OR: 1.52, 95%CI: 1.23-1.89, P<0.001 ). The "others" (haplotypes with frequency <1% were merged) were broadly distributed in cases at a low level (OR: 0.46, 95%CI: 0.29-0.71, P<0.001). The significance of this result is limited given the naturally low frequencies of these haplotypes. We did not discover any associations with C rs2294008 G rs2976392 A rs2978974 and T rs2294008 A rs2976392 A rs2978974 in breast cancer.

DISCUSSION
Genetic studies have provided insight into various diseases, including cancers. Understanding the associations between different genes and cancers can improve prevention, treatment, and prognosis estimation. Genome-wide association studies have revealed many genetic markers of different cancers. Recently numerous studies have indicated that PSCA may influence a diverse group of cancers, including gastric, bladder, renal, and Rs2294008 is located in exon 1 of PSCA and its C to T transition has been shown to reduce transcriptional activity of an upstream fragment of PSCA [28,29]. Precious meta-analyses discovered that T allele of rs2294008 was a risk factor for cancer, particularly for gastric and bladder cancers [26,27]. The T allele of rs2294008 increased risk for gastric cancer in Asian populations [30,31] and the genetic variant rs2294008 was identified to confer genetic susceptibility for bladder cancer risk in both Caucasian [12] and Asian [14,20] populations. In this study, we found that the minor allele of rs2294008 was associated with a high risk of breast cancer among both premenopausal and postmenopausal women. There was no association between rs2294008 and ER status, although PR-positive tumors were associated with the T allele. In contrast, a study based on Korean women reported that the minor allele of rs2294008 was associated with reduced breast cancer risk among premenopausal women, increased breast cancer risk among postmenopausal women, and that the T allele increased the ER-negative breast cancer risk [9]. Whist similar, our study provides a more robust analysis as it includes more patients as well as more detailed stratified analyses. Given the heterogeneous nature of breast cancer, the discrepancies between our findings and those of Kim et al.
[9] may be explained by various factors, including region, lifestyle, genetic testing methods, and study design.
Rs2976392 is located in the intron 2 of PSCA and has a strong linkage disequilibrium with rs2294008 C > T [24,32]. The association of this SNP and cancer susceptibility has been widely investigated. Recent metaanalysis has revealed the PSCA rs2976392 polymorphism was significantly associated with increased overall cancer risk [27]. Rs2978974 in the promoter region of PSCA showed low linkage disequilibrium with rs2294008 and the A rs2978974 allele was shown to contribute to bladder cancer susceptibility, presumably due to the loss of binding of ELK1 or other ETS proteins to the PSCA promoter [12]. A study based on 405 gallbladder cancer patients and 247 healthy controls showed that the PSCA haplotype T rs2294008 A rs2978974 conferred low risk of gallbladder cancer in males, while in females, the T rs2294008 G rs2978974 haplotype was related to increased gallbladder cancer risk [22]. Kim et al. found that there was no statistically significant relationship between rs2976392 and breast cancer risk, which is concordant with our study. However, we found the rs2976392 SNP was   associated with an increased risk of lymph node metastasis. This study provides the first investigation of associations between rs2978974 and breast cancer risk. We demonstrated that the minor allele of rs2978974 specifically increased the risk of breast cancer among postmenopausal women, while it was not associated with the risk of breast cancer among all patients, and was not associated with patient age or any of the clinicopathological features.
It is believed that haplotypes may be more important than any single SNP analysis in influencing a clinical response [33,34]. To our knowledge, this is the first report of haplotypes in PSCA rs2294008, rs2976392, and rs2978974 polymorphisms. Haplotype analysis indicated that the T rs2294008 A rs2976392 G rs2978974 haplotype was associated with significantly increased risk of breast cancer.
This study has several limitations. First, the singlecenter design may preclude extrapolation of our findings to other patient populations or ethnic groups. Second, we used a hospital-based case-control design, which may involve selection bias. Third, our sample size was relatively small, which may limit the strength of our stratified analyses. Fourth, we did not consider other important risk factors (e.g., high-dose radiation exposure at the chest, alcohol consumption, and other benign breast lesions), as we did not have access to these data. Therefore, a large well-designed prospective study is needed to validate our findings. Furthermore, biological function studies are crucial for elucidating the role of PSCA in breast cancer.
Our study revealed that the PSCA rs2294008 polymorphism influenced the risk of breast cancer among Chinese women and the rs2978974 polymorphism may specifically increase the risk of breast cancer in postmenopausal women. We found that rs2294008 was associated with PR-positive status and rs2976392 was associated with lymph node metastasis among Chinese women with breast cancer. Furthermore, the T rs2294008 A rs2976392 G rs2978974 haplotype may increase the susceptibility to breast cancer.

Study population
We included the cases with pathologicallyconfirmed breast cancer, without history of any cancer, were treated at the Department of Oncology (Second Affiliated Hospital of Xi'an Jiaotong University) between January 2013 and October 2014. The healthy individuals who had visited the medical examination center at the Second Affiliated Hospital of Xi'an Jiaotong University for a check-up during the study period were included as controls. All individuals were Chinese Han women, and the controls were frequencymatched to the cases according to age (±5 years) and menopausal status. Finally, 560 eligible patients with an average age 49.09 ± 11.02 years and 583 healthy age-matched controls were included in the study ( Table  6). The cases and controls exhibited similar clinical characteristics with the exception of body mass index (BMI) (P = 0.038).
A standardized epidemiological questionnaire was used to collect demographic and personal information.
Clinical information was collected from medical records and pathological reports. All participants were informed regarding the study's purpose and experimental procedures, and provided their written informed consent. The Human Research Committee at our institution approved the use of blood samples.

SNP selection and genotyping
Peripheral blood samples were collected in a standard tube and stored at -80°C. Genomic DNA was extracted from the peripheral whole blood samples using the Universal Genomic DNA Extraction Kit (version 3.0; TaKaRa, Japan). To achieve a power of at least 50%, only SNPs with a minor allele frequency of >0.01 were included. Three primers were designed to amplify fragments of rs2294008, rs2978974, and rs2976392. Primers and PCR product sequences are shown in Table  7. DNA concentrations were measured by spectrometry (DU530 UV/VIS spectrophotometer; Beckman Instruments, Fullerton, CA, USA), Sequenom MassARRAY RS1000 was used for genotyping, and the related data were managed using Sequenom Typer 4.0 Software [35].

Statistical analysis
Microsoft Excel and SPSS software (version 21.0; SPSS Inc., Chicago, IL, USA) were used for all analyses. P-values were calculated using the χ 2 test, and all tests were two-tailed; a P-value of <0.05 was considered statistically significant. The exact test was used to examine the distribution of each SNP among the controls, and their accordance with the Hardy-Weinberg equilibrium. Five different genetic models were used to evaluate the risk of breast cancer, with "A" used to indicate the major allele and "a" used to indicate the minor allele: the allele model (a vs. A); the co-dominant model (homozygote model: aa vs. AA; heterozygote model: Aa vs. AA); the recessive model (aa vs. AA+Aa); the dominant model (AA vs. Aa+aa); and the over-dominant model (AA+aa vs. Aa). The allelic frequencies for each SNP were compared between cases and controls in each model using the χ 2 test and SNPStats software [36,37]. Power calculations were made by PS software (Power and Sample Size Calculation, which was downloaded online: http://biostat.mc.vanderbilt. edu/wiki/Main/PowerSampleSize). Phase2.1 software was used to conduct all common haplotypes [34] and SPSS software was used to estimate the ORs and 95 % CIs for each haplotype. As shown in Table 6, there was a significant difference in BMI between breast cancer cases and controls (P = 0.038). BMI may be a confounder in the development of breast cancer. Therefore, to control for its effects, we used stratification analyses. First, we calculated OR1 and OR2 by stratification for BMI. Then, the ratio of the OR (unadjusted OR) and OR1/OR2 was calculated. If the ratio was close to 1, the results did not need to be adjusted, indicating that BMI was not a confounder. Otherwise, we would need to adjust results for BMI.