Association of MDM2 SNP309 and TP53 Arg72Pro polymorphisms with risk of endometrial cancer

The incidence of endometrial cancer, a common gynecological malignancy, is increasing in Japan. We have previously shown that the ER/MDM2/p53/p21 pathway plays an important role in endometrial carcinogenesis. In the present study, we investigated the effects of germline single nucleotide polymorphisms in murine double minute 2 (MDM2) SNP309, TP53 Arg72Pro, ESR1 PvuII and XbaI, and p21 codon 31 on endometrial cancer risk. We evaluated these polymorphisms in DNA samples from 125 endometrial cancer cases and 200 controls using polymerase chain reaction-based restriction fragment length polymorphism. The association of each genetic polymorphism with endometrial cancer was examined by the odds ratio and 95% confidence interval, which were obtained using logistic regression analysis. The SNP309 GG genotype non-significantly increased the risk of endometrial cancer. The 95% confidence interval for the GG genotype vs. the TT genotype of MDM2 SNP309 was 1.76 (0.93–3.30). Endometrial cancer was not associated with tested SNP genotypes for TP53, ESR1 and p21. The combination of SNP309 GG + TG and TP53 codon 72 Arg/Arg significantly increased endometrial cancer risk. The adjusted OR was 2.53 (95% confidence interval, 1.03–6.21) and P for the interaction was 0.04. This result was supported by in vitro data showing that endometrial cancer cell lines with the SNP309 G allele failed to show growth inhibition by treatment with RITA, which reduces p53-MDM2 binding. The presence of the SNP309 G allele and TP53 codon 72 Arg/Arg genotype is associated with an increased risk of endometrial cancer in Japanese women.


Introduction
The incidence of endometrial cancer is increasing; it is the second most common gynecologic malignancy in Japan.
Increased estrogen exposure, particularly unopposed estrogen, is a major risk factor for endometrial cancer. Early menarche, infertility, obesity and late menopause are also associated with increased risk of the cancer. Endometrial cancer is broadly classified into one of two clinicopathological types: I and II (1). The former is estrogen-related and occurs in both premenopausal and postmenopausal women. Histologically, it is of the endometrioid type, generally of low cellular grade and has a favorable prognosis. It is frequently preceded by endometrial hyperplasia. These tumor cells frequently express the estrogen receptor (ER), particularly ERα. Type I endometrial cancer development is associated with a variety of genetic alterations, including PTEN inactivation, K-ras mutation, β-catenin mutation and microsatellite instability. Type II endometrial cancer is non-estrogen-related and occurs primarily in postmenopausal women. Type II encompasses non-endometrioid histologies and is represented by serous or clear-cell adenocarcinoma. It is commonly associated with an atrophic rather than a hyperplastic endometrium. The cells are negative for ER and the progesterone receptor (PR); type II endometrial cancer exhibits a high cellular grade and is associated with poor prognosis. Genetic alterations in type II carcinomas include p53 mutation, p16 inactivation, HER-2/neu overexpression and reduced E-cadherin expression (1,2.) A human homologue of the murine double minute 2 (MDM2) gene (also known as HDM2 in humans) is frequently overexpressed in several types of human cancer, particularly in breast carcinomas and soft tissue sarcomas (3). The major contribution of MDM2 to cancer development is through tight inhibition of tumor suppressor p53 activity and stability. Biochemically, MDM2 functions as an E3 ubiquitin ligase responsible for p53 ubiquitination and degradation (4)(5)(6). Recent studies have identified numerous additional MDM2interacting proteins and p53-independent functions of MDM2 in the regulation of multiple signaling pathways, including the pRb/E2F complex and the PI3K/Akt pathway (7). MDM2 overexpression is also strongly related to the presence of ER (8,9.) Elevated ER expression in cells lacking ER induces MDM2 overexpression (3). We have shown that overexpression of wild-type ERα in NIH3T3 cells results in a significant increase in MDM2 protein levels. MDM2 gene expression is also regulated by the Ras-driven Raf/MEK/MAP kinase pathway in a p53-independent manner (10). We have also demonstrated that the Ras/ER/MDM2 pathway is critical for NIH3T3 cell transformation (11,12) and that blockage of this pathway by inhibitors or MDM2 siRNA induces p53 and p21 expression and suppresses cell proliferation in estrogendependent cancer such as endometrial and ovarian cancer (13). These results suggest that the ER/MDM2/p53/p21 pathway plays an important role in the development of endometrial cancer.
A single nucleotide polymorphism (SNP309 T> G; rs2279744) has been identified in the MDM2 promoter. The SNP309 G allele has high affinity for the transcriptional activator SP1, which results in a higher level of MDM2 mRNA and MDM2 protein and subsequent attenuation of the p53 pathway. In humans, the SNP309 G allele is associated with accelerated tumor formation in hereditary cancer associated with Li-Fraumeni syndrome and sporadic soft tissue sarcomas (14). SP1 is a well-characterized co-transcriptional activator of multiple hormone receptors, including ER. In an analysis of 3 different types of sporadic cancer (diffuse large B-cell lymphoma, soft-tissue sarcoma and invasive ductal breast carcinoma), Bond et al (15) showed that the SNP309 G allele is associated with gender-specific and hormone-dependent acceleration of tumorigenesis. Several studies, however, found no evidence of an association between SNP309 and cancer risk (16)(17)(18). The results remain controversial.
The p53 signaling pathway is activated by a wide variety of stress signals. Activation of p53 induces growth arrest, cellular senescence and apoptosis (19). Numerous polymorphisms are present in the TP53 locus. A well-known G to C SNP at TP53 codon 72 results in an arginine (Arg; CGC) or proline (Pro; CCC) polymorphism (rs1042522). This polymorphism is of particular interest owing to its functionality, although its biological function is controversial.
ESR1 is the principal ER expressed in the endometrium and is thought to be important in the development of endometrial carcinoma. The ESR1 gene contains several SNPs whose functional significance remains unknown. The two most frequently studied polymorphisms located in ESR1 gene intron 1 are often identified by their restriction endonucleases, PvuII T/C (rs2234693) and XbaI A/G (rs9340799). p21, a cyclin-dependent kinase inhibitor, is the major downstream component of the TP53 tumor suppressor pathway. This protein binds to and inhibits all cyclin-dependent kinase complexes, causing cell cycle arrest in G1. The p21 codon 31 C/A (rs1801270) in exon 2 leads to a serine (Ser)/Arg amino acid substitution and is located in the DNA-binding zincfinger motif of the gene (20). This polymorphism has been implicated in cervical adenocarcinoma (21) and endometrial cancer (22,23).
The MDM2 SNP309 G/G genotype increases the risk of endometrial cancer in Caucasians (24,25) and in Japanese women (26). Furthermore, a combination of the homozygous Arg/Arg genotype of TP53 codon 72 and the homozygous GG genotype of MDM2 SNP309 is significantly associated with the risk of endometrial cancer in Japanese women.
We performed the present case control study to investigate the relationship between endometrial cancer risk and multiple genetic polymorphisms, including MDM2 SNP309, TP53 Arg72Pro, ESR1 PvuII and XbaI, and p21 codon 31, in Japanese women.

Materials and methods
Study subjects. Subjects included 125 endometrial cancer patients who were diagnosed at the Department of Obstetrics and Gynecology of Kyushu University Hospital and Kyushu Cancer Center Hospital between 1993 and 2010. All patients provided informed consent. There was no family history of endometrial cancer in any of these cases. Control subjects were selected from 7,132 women aged 49-76 years, living in the East Ward of Fukuoka City, who had completed a baseline survey between February 2004 and August 2007 of an on-going cohort study regarding lifestyle-related diseases. Out of 2,055 women aged 49-60 years, 2,783 women aged 60-69 years, and 1,147 women aged ≥70 years who had no history of cancer and who had donated a blood sample for a genetic study under signed informed consent, 106, 66 and 28 women, respectively, were randomly selected in proportion to the age distribution of the 125 cases of endometrial cancer accrued as of the end of March 2010.
Cell culture. Endometrial cancer cell lines were maintained in DMEM supplemented with 10% FBS for HHUA and 15% FBS for Hec6 and Sawano cells.
Cell growth assay. Cells were plated in a 6-cm dish at 1x10 5 cells/dish and incubated with a medium supplemented for 24 h. Subsequently, to reactivate p53 and induce tumor cell apoptosis, the compound RITA (Tocris, Ellisville, MO, USA), was added, and the cells were further incubated for 48 and 96 h. Following incubation, floating cells were washed away and adherent cells were detached from the dishes using 0.25% trypsin. Detached cells were then counted using a hemocytometer.
Statistical analysis. Comparisons of means and proportions between cases and controls were performed using the t-test and Chi-square test, respectively. Deviation from the Hardy-Weinberg equilibrium was evaluated using the Chi-square test with 1 degree of freedom. The association of each genetic polymorphism with endometrial cancer was examined by the odds ratio (OR) and the 95% confidence interval (95% CI), which were obtained using logistic regression analysis. The 95% CI was derived from the standard error of the logistic regression coefficient. Statistical adjustment was made for age and body mass index (BMI) with continuous variables for each included as covariates. Trends in ORs according to the number of minor alleles were evaluated using the Wald test. For analyzing the interaction between two polymorphisms, heterozygous and homozygous genotypes of the minor alleles were combined and statistical evaluation was conducted using the Wald test for the product term of the two variables representing the genotypes containing the minor allele. Statistical significance was declared if a two-sided P-value <0.05 or if the 95% CI did not include unity. Statistical analyses were carried out using SAS version 8.2 (SAS Institute Inc., Cary, NC, USA).

Characteristics of endometrial cancer cases and controls.
Selected characteristics of endometrial cancer cases and controls are shown in Table I Table II shows the association between endometrial cancer risk and SNP309, TP53, ESR1 and p21 polymorphisms. The OR for the SNP309 GG genotype when compared with the SNP309 TT genotype non-significantly increased risk. The crude OR was 1.76 (95% CI, 0.93-3.30), and the age-and BMI-adjusted OR was 1.64 (95% CI, 0.81-3.28). There was also no measurable association between endometrial cancer and the TP53 Arg72Pro, ESR1 PvuII and XbaI, or the p21 codon 31 polymorphisms. Adjustment for age and BMI did  the combination of the two polymorphisms (Table IV), the SNP309 TG and GG genotypes as well as the TP53 Arg/Pro and TP53 Pro/Pro genotypes were combined. A significant increase in the adjusted OR associated with the SNP309 G allele was limited to those homozygous for the TP53 Arg allele (OR, 2.53; 95% CI, 1.03-6.21). A statistically significant interaction was observed between the 2 polymorphisms on the risk of endometrial cancer (P for the interaction=0.04). We repeated the analysis to include postmenopausal status and type I endometrial cancer status. Adjusted ORs for postmenopausal women and type I women possessing both polymorphisms when compared with being homozygous for both wild-type alleles were 2.96 (95% CI, 1.04-8.44) and 2.51 (95% CI, 0.97-6.53), respectively. There was a statistically significant interaction between the two polymorphisms regarding the risk of endometrial cancer (P for the interaction=0.03 and 0.01, respectively). Moreover, the corresponding value for type I among postmenopausal women (59 cases) was 3.24 (95% CI, 1.03-10.16), which reflected a statistically significant interaction between the SNP309 and TP53 Arg72Pro polymorphisms (P for the interaction=0.009).
Combination of the TP53 72Pro allele and homozygosity for the p21 codon 31 Ser allele is associated with a decreased risk of endometrial cancer. We further examined the effect of SNP309 combined with ESR1 or p21 and SNP309 with both TP53 and p21 polymorphisms on the risk of endometrial cancer (Table V). No significant differences were observed between the combination of the SNP309 and the PvuII or XbaI or p21 polymorphism. The combination of having the TP53 72Pro allele and homozygosity for the p21 codon 31 Ser allele, however, was associated with a decreased risk of endometrial cancer (crude OR, 0.28; 95% CI, 0.09-0.92). A statistically significant interaction was observed between the 2 polymorphisms for a decreased risk of endometrial cancer (P for the interaction=0.04), although there was no significant difference once adjustment was made for age and BMI. The SNP309 GG genotype abrogates the cytostatic effect of RITA on tumor cells. MDM2 overexpression is observed in multiple malignancies. Due to the importance of the p53-MDM2 interaction, restoration of p53 activity by inhibiting MDM2 binding represents a novel antineoplastic strategy. RITA is a low-molecular-weight compound previously identified in a cell-based screen for wild-type p53-reactivating compounds. RITA binds to the amino terminus of p53, inhibiting p53 binding to MDM2 in cultured cells and in human tumor xenografts in vivo. This results in derepression of p53 and highly efficient induction of apoptosis (32).
We demonstrated that a combination of the SNP309 G allele and the homozygous Arg/Arg genotype of TP53 codon 72 was associated with an increased risk of endometrial cancer (Table IV). These results suggest that polymorphisms of MDM2 and TP53 may influence the effect of RITA. We therefore, assessed the growth-suppressive effect of RITA on three endometrial cancer cell lines, Hec6, HHUA and Sawano, in vitro. All three lines express wild-type p53. SNP309 and TP53 Arg72Pro polymorphisms were analyzed in these lines with PCR-RFLP and direct sequencing. Hec6 cells were SNP309 TT homozygous and TP53 Arg72 homozygous, HHUA cells were SNP309 TG heterozygous and TP53 Arg72 homozygous, and Sawano cells were SNP309 GG homozygous and TP53 Arg72Pro heterozygous (Table VI). Treatment with RITA (0.5 and 1.0 µM) suppressed cell growth in all cell lines in a dose-dependent manner (Fig. 1A). Following treatment with 0.5 µM RITA for four days, the rates of inhibition were: Hec6 cells, 90.9±8.7%; HHUA cells, 54.1±6.0%; and Sawano cells, 49.0±15.8%. Following treatment with 1.0 µM RITA, these rates were 94.3±1.2%, 79.0±4.6% and 61.0±6.7%, respectively. The inhibitory effect was significantly less pronounced in HHUA and Sawano cells with the SNP309 G allele compared to that in Hec 6 cells carrying the SNP309 TT genotype (P<0.05) (Fig. 1B).   (33), however, failed to document a similar association in an Australian study.
The association between the TP53 Arg72Pro polymorphism and endometrial cancer risk has been investigated in several studies in Caucasians with inconsistent findings. The following findings have been reported for women of Asian descent: Ueda et al (27) reported an increased risk of endometrial cancer in Japanese patients harboring the Arg/Arg genotype compared to those with Arg/Pro and Pro/Pro genotypes combined. Niwa et al (34), however, found no such association. By contrast, having the Pro allele conferred an increased risk in Korean women (22).
Two studies have reported a potential interaction between the SNP309 G allele and the TP53 Arg72Pro polymorphisms for endometrial cancer. Nunobiki et al (35) showed that the risk with the SNP309 GG vs. TT or TG was greater in women with Arg/Arg (OR, 3.28; 95% CI, 1.13-9.53) than in women with Arg/Pro or Pro/Pro of the TP53 Arg72Pro (OR, 1.48; 95% CI, 0.62-3.52), but the interaction between the two SNPs was not evaluated.
Ashton et al (33) showed no effect of the TP53 polymorphism and MDM2 SNP309 alone or in combination on endometrial cancer risk. They did, however, show that the combination of SNP309 and TP53 was associated with high-grade endometrial cancer (G2+G3). These observations are incompatible with the findings of the present study showing that the combination of SNP309 and TP53 was associated with type I endometrial

A B
cancer (G1+G2). This discrepancy may have arisen due to the fact that G2 and G3 are classified as high-grade tumors. G3 tumors are a distinct entity. For example, in Japanese women, p53 mutations are significantly more common in G3 (43%) than in G1 (11%) tumors (36). p53 mutations are also detected more frequently in type II endometrial cancer. Tumors harboring p53 mutations are often more chemoresistant and have lower apoptotic rates. The status of the TP53 gene (wild-type or mutant type) is critical when determining the relationship between grade and a TP53 polymorphism. In the present study, an analysis of 102 cases that were not type II demonstrated a significant interaction between the MDM2 SNP309 and TP53 Arg72Pro polymorphism with endometrial cancer risk. We did not, however, ascertain whether the combination of MDM2 and TP53 Arg72Pro was associated with high-grade or nonendometrioid tumors as the overall number of these cases (G3, 10 cases; non-endometrioid, 13 cases) was small. A major limitation of the present study was its small size. As most of the women with endometrial cancer are early stage and have positive outcomes, it was difficult to analyze the association between studied polymorphisms and disease outcome in the present study.
Five meta-analyses have been published addressing the MDM2 SNP309 polymorphism and the risk of various types of tumor. Wilkening (41) conducted a meta-analysis of 1,001 cases and 1,889 controls from six published case-control studies (24)(25)(26)33,35) to estimate the effect of SNP309 on endometrial cancer risk for GG vs. TT (OR, 1.54; 95% CI, 1.21-1.95). These analyses indicate that MDM2 SNP309 serves as a tumor susceptibility marker, and that there is an association between MDM2 SNP309 and TP53 Arg/Pro regarding tumor susceptibility.
The other notable finding of the present study was that the combination of the TP53 72Pro allele and homozygosity for the p21 codon 31 Ser allele was associated with a decreased risk of endometrial cancer. This is in contrast to a previous study showing that the combination of the TP53 Pro allele and p21 Ser/Ser genotype significantly increased endometrial cancer risk (OR, 9.55; 95% CI, 4.40-21.24) (22). The result requires further investigation.
Finally, we investigated whether the SNP309 subtype modified the effect of RITA, an agent that blocks the p53-MDM2 interaction. RITA suppresses the proliferation of endometrial cancer cell lines. The effect was less pronounced in HHUA and Sawano cells containing the SNP309 G allele compared with that in Hec6 cells, which harbor the TT genotype. Arva et al (42) showed that cells carrying the SNP309 GG genotype exhibited a compromised TP53 response pathway and formed transcriptionally inactive p53-MDM2 complexes in response to stress. These results support our findings. Additional functional analyses are required to fully elucidate the effects of these polymorphisms on the effect of agents such as RITA on endometrial cancer cell lines.
In conclusion, endometrial cancer is characterized by numerous genetic alterations, including those in p53, K-ras, PTEN and β-catenin. We previously demonstrated that the Ras/ER/MDM2 pathway is important for proliferation of endometrial cancer cells in vitro. We demonstrated in this case-control study that polymorphisms of TP53 and MDM2 modify the effect of this signaling pathway and thus increase the risk of endometrial cancer.
The increased incidence of endometrial cancer in Japan may reflect changes in lifestyle. The interplay between genetic and environmental factors is now being investigated in the context of gynecologic cancer. It is well known that endometrial cancer risk includes an environmental component. Further studies are necessary to elucidate the genetic components of risk. Understanding this relationship is the first step towards developing new methods of endometrial cancer prevention and treatment.