Genetic Variants at 6 p 21 . 1 and 7 p 15 . 3 Identified by GWASs of Multiple Cancers and Ovarian Cancer Risk : a Case-control Study in Han Chinese Women

Ovarian cancer is the leading cause of death from gynecologic cancer in women worldwide (Kumar et al., 2011), with an estimation of 225,500 women diagnosed with ovarian cancer and 140,200 deaths in 2008 (Lowe et al., 2013). Several risk factors of ovarian cancer have already been identified, such as parity (Pasalich et al., 2013), never breastfeeding (Luan et al., 2013), endometriosis (Vargas-Hernandez 2013), tubal ligation (Cramer 2012), body mass index and height (BMI) (Olsen et al., 2013) and family history (Moorman et al., 2013). Furthermore, there is growing evidence suggesting that genetic variants may contribute to the susceptibility to ovarian cancer (Mohamed et al., 2013; Shen et al., 2013). Over the last several years, as a powerful method to investigate the genetic determinants of complex diseases, genome-wide association studies (GWASs) have successfully identified many genetic markers for susceptibility to cancers (Yang et al., 2013). Meanwhile,

Recently, Jin et al. (2012) synthesized the data from several published GWASs of multiple cancers in Han Chinese, including lung cancer, non-cardia gastric cancer (NCGC), and esophageal squamous-cell carcinoma (ESCC), and identified two new susceptible loci (rs2494938 at 6p21.1 and rs2285947 at 7p15.3) associated with risks of multiple cancers (Jin et al., 2012), implying the pleiotropy of genetic variants in these two regions.However, the role of genetic variants at 6p21.1 and 7p15.3 is unknown in the development of female-specific cancers, such as ovarian cancer.In this study, we conducted a casecontrol study with 377 cases and 1034 controls to explore the relationship of two loci (rs2494938 A > G at 6p21.1 and rs2285947 A > G at 7p15.3) with ovarian cancer risk in Han Chinese woman.During the same period as the cases were recruited, cancer-free women in the control group were randomly selected from a cohort of more than 30, 000 participants in a community-based screening program for noninfectious diseases conducted in Jiangsu Province (Jiang et al., 2011).All subjects were genetically unrelated, ethnic Han Chinese women.Each individual was interviewed face-to-face by trained interviewers to get demographic information, age at menarche, menstrual and reproduction history and environmental exposure history.Approximately 5ml venous blood sample was obtained from each subject.Eventually, 377 ovarian cancer cases and 1034 cancer-free controls frequency matched by age (5 years age categories) were included in this study.

Genotyping
Genotyping was conducted by using the TaqMan allelic discrimination assay on ABI PRISM 7900 HT platform (Life Technologies.CarIsbad USA).Information of the primers and probes were shown in Table 1.Experimental conductor was not informed the subjects' case or control status.In each 384-well reaction plate, two negative controls were added for quality control.The genotype data were analyzed by using SDS 2.3 Allelic Discrimination Software (Applied Biosystems).To make sure the genotyping results, 50 cases and 50 controls were randomly selected to repeat, yielding a 100% concordant.

Statistical analyses
χ² test (for categorical variables) and student t test (for continuous variables) were used for evaluating the differences in the distributions of demographic characteristics, selected variables, and frequencies of genotypes between the cases and controls.Logistic regression analysis was applied for estimating the association between genotypes and ovarian cancer risk with adjustment for age, age at menarche, menopausal status, and abortion information.Hardy-Weinberg equilibrium was tested using goodness-of-fit χ² test.All of the statistical analyses were performed with Statistical Analysis System software (9.1.3;SAS Institute, Cary, NC).

Results
Characteristics of ovarian cancer cases and cancerfree controls are shown in Table 2.In 377 ovarian cancer patients, 327 were diagnosed as epithelial ovarian cancers, and 55 were reported as other pathological types.Nearly 72.7% (274) patients were diagnosed at stage III or IV and 27.3% (103) were at stage I or II.The age was comparable between cases and controls (P>0.05).However, ovarian cancer cases had earlier menarche age, more abortion history and more premenopausal status compared with cancer-free controls (all P<0.05).Loci and genotyping information of the two loci (rs2494938 and rs2285947) are presented in Table 3.Call rates of the two loci were above 99% and the genotype frequencies of the two loci in the control subjects were consistent with Hardy-Weinberg equilibrium (P>0.05).As shown in Table 4, a significant association was observed between rs2285947 and ovarian cancer risk (dominant model: adjusted OR=1.45, 95%CI 1.11-1.89,P=0.006; additive model: adjusted OR=1.33, 95%CI 1.08-1.64,P=0.008, ).However, no significant association was found between rs2494938 and ovarian cancer risk (dominant model: adjusted OR=0.86, 95%CI 0.66-1.12,P=0.270; additive model: adjusted OR=0.86, 95%CI 0.69-1.08,P=0.190) (Table 4).
Additionally, we conducted stratification analyses to evaluate the potential association between the two SNPs (rs2494938/rs2285947) and ovarian cancer risks stratified by age, age at menarche, abortion history, and menopausal status (Table 5).The results showed that associations between rs2285947 and ovarian cancer risk were significant in the subgroups with age above 52 years old (additive model: OR=1.37, 95%CI 1.03-1.84-1.75,P=0.032), with age at menarche over 15 years old (additive model: OR= 1.43, 95%CI 1.09-1.86,P=0.009) and premenopausal women (additive model: OR=1.57, 95%CI 1.12-2.21,P = 0.009).However, no significant heterogeneity was detected between the stratified subgroups.Furthermore, no significant result was found for the association between rs2494938 and ovarian cancer risk in every stratum (Table 5).

Discussion
In the present case-control study, we investigated the relationship between two loci at 6p21.1 and 7p15.3 identified by the GWAS of multiple cancers and ovarian cancer risk and found that rs2285947 at 7p15.3 was significantly associated with ovarian cancer risk in Han Chinese women.
Rs2285947 is located at 7p15.3 in the intron of DNAH11 gene (Dynein, axonemal, heavy chain)[MIM603339] and SNPs (Single Nucleotide Polymorphisms) in strong linkage disequilibriumLD with rs2285947 are located in the initial parts of both SP4 (Specificity Protein 4[MIM 600540]) and DNAH11.By using publicly available data of cis-expression quantitative trait loci (eQTLs), we found that a SNP rs7788515 in strong LD with rs2285947 (r 2 =0.92) plays a role in the expression of DNAH11 (P=5.40×10 - ) (Ge et al., 2009).Dyneins are microtubuleassociated motor protein complexes and has been reported that in vivo it is necessary for the activation of MAPK (mitogen-activated protein kinase) kinase3/6 and p38 (Cheung et al., 2004).The p38-MAPK pathway plays important roles in the regulation of several biological progresses including cell survival, differentiation, migration (Cuenda et al., 2007;Cuadrado et al., 2010), immune and inflammatory responses (Han et al., 1994;Lee et al., 1994).Our study showed that rs2285947 A allele was associated with an increased risk of ovarian cancer risk in Han Chinese women, consistent with the findings in lung cancer, non-cardia gastric cancer, and esophageal squamous-cell carcinoma by Jin et al. (2012).Furthermore, it has been reported that several other loci at 7p15.3 were associated with many kinds of cancers, such as multiple myeloma (rs4487645) (Broderick et al., 2012;Martino et al., 2012) and lung cancer (rs2710994) (Xun et al., 2011), and DNA copy number aberrations (DCNAs) at 7p15.3 increased the risk of invasive cervical carcinoma (Oh et al., 2012).All these findings provided evidence that genetic variants at 7p15.3 are potentially important susceptibility loci associated with the development of multiple cancers.Further investigations are necessary to identify the specific mechanism of such loci/gene in the development of human cancer.
Rs2494938 is located at 6p21.1 in the intron one of LRFN2 [MM612808] (encoding leucine-rich repeat and fibronectin type III domain-containing protein 2).It has been reported that the LRFN2 in cooperation with MYC lead to erythroblastosis (Castellanos et al., 2007).LRFN2 is a member of synaptic adhesion-like molecule (SALMs) and the family is associated with the N-methyl D-aspartate (NMDA) receptors (Wang et al., 2006).Some evidences had shown that the NMDA receptor 2B is methylated in ESCC and non-small cell lung cancer, exhibiting a tumorsuppressive activity (Kim et al., 2006;Tamura et al., 2011).Jin et al. (2012) reported that A allele significantly increased the risk of multiple cancers including lung cancer, non-cardia gastric cancer, and esophageal squamous-cell carcinoma.However, in the current study, we could not validate the association between rs2494938 at 6p21.1 and ovarian cancer risk, indicating that these loci may have no effect on the development of ovarian cancer.
In this study, several limitations need to be addressed.First, the statistical power was relatively low due to the small sample size, which may miss the real association between loci at 6p21.1 and ovarian cancer risk.Second, the biological mechanism of genetic variants at 7p15.3 was not further investigated in this study.In general, our results indicated that the positive associations between genetic variants at 7p15.3 and multiple types of cancer are also reduplicative in ovarian cancer in Chinese women.The functional relevance of genetic variants in 7p15.3 and the role in susceptibility to ovarian cancer remain to be confirmed in larger epidemiological studies and in-depth mechanistic studies.

Table 5 . Stratified Analysis on the Associations Between Rs2494938/Rs2285947 and Risk of Ovarian Cancer
a Per allele Odds ratio (OR) and 95% confidence interval (CI) adjusted for age, age at menarche, abortion and menopausal status where appropriate; b P for heterogeneity test