Inherited functional single nucleotide polymorphisms (SNPs) in DNA repair genes may alter DNA repair capacity and thus contribute to cancer risk.
Three ERCC1 functional SNPs (rs2298881C>A, rs3212986C>A and rs11615G>A) and two XPF/ERCC4 functional SNPs (rs2276466C>G and rs6498486A>C) were genotyped for 1125 gastric adenocarcinoma cases and 1196 cancer-free controls by Taqman assays. Odds ratios (OR) and 95% confidence intervals (CI) were used to estimate risk associations, and false-positive report probabilities (FPRP) were calculated for assessing significant findings.
ERCC1 rs2298881C and rs11615A variant genotypes were associated with increased gastric cancer risk (adjusted OR = 1.33, 95% CI = 1.05–1.67 for rs2298881 AC/CC and adjusted OR = 1.23, 95% CI = 1.05–1.46 for rs11615 AG/AA, compared with their common genotype AA and GG, respectively). Patients with 2–3 ERCC1 risk genotypes had significant increased risk (adjusted OR = 1.56, 95% CI = 1.27–1.93), compared with those with 0–1 ERCC1 risk genotypes, and this risk was more significantly in subgroups of never drinkers, non-gastric cardia adenocarcinoma (NGCA) and clinical stage I+II. All these risks were not observed for XPF SNPs.
Citation: He J, Xu Y, Qiu L-X, Li J, Zhou X-Y, Sun M-H, et al. (2012) Polymorphisms in ERCC1 and XPF Genes and Risk of Gastric Cancer in an Eastern Chinese Population. PLoS ONE 7(11): e49308. https://doi.org/10.1371/journal.pone.0049308
Editor: Xiao-Ping Miao, Huazhong University of Science and Technology, China
Received: July 18, 2012; Accepted: October 8, 2012; Published: November 15, 2012
Copyright: © 2012 He et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was supported by funds from China’s Thousand Talents Program at Fudan University and the grant from the Ministry of Health (grant number 201002007). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Gastric cancer is the second leading cause of cancer-related deaths worldwide, with an estimated nearly one million new cases, accounting for 10% of all cancer-related deaths occurred in 2008 and approximately 40% of all cases occurred in China , ranking the third most common cancer in China . To date, the etiology of gastric cancer remains unclear, although currently suggested risk factors include smoking and dietary deficiencies , , gastroesophageal reflux and obesity , high body mass index , and Helicobacter pylori (H. pylori) infection . It is likely that genetic variation may also influence susceptibility to gastric cancer .
DNA repair are important in maintaining stability and integrity of human genome. In humans, there are at least five major DNA repair pathways consisting of more than 130 genes, of which the nucleotide excision repair (NER) pathway removes a wide variety of DNA lesions, including bulky adducts, cross links, oxidative DNA damage, alkylating damage and thymidine dimers –. In addition, at least seven xeroderma pigmentosum (XP) complementation groups have been identified, which are the rate-limiting ones in the NER mechanism . For example, the excision repair cross-complimentary group 1 (ERCC1) gene encodes a subunit of the NER complex required for the incision step of NER. More importantly, the ERCC1 protein forms a heterodimer with the XPF endonuclease (also known as ERCC4) to catalyze the 5′ incision in the process of excising the DNA lesion . Since the ERCC1 protein is critical for NER and may influence genomic instability, variations in ERCC1 are likely to play an important role in maintaining genomic stability that is disrupted in carcinogenesis , .
Three common ERCC1 variants, namely rs11615, rs321986 and rs3212961, have been investigated previously, because they are probably functional . Specifically, two case-control studies reported the association between ERCC1 rs11615 variant genotypes and gastric cancer risk, including one study of 126 cases  and another of 314 cases  in Italian populations. For the association between ERCC4 polymorphisms and cancer risk, previous studies have mainly focused on a functional SNP rs1800067G>A, and few case-control studies reported on gastric cancer to date , .
To further assess the associations of ERCC1 and ERCC4 polymorphisms with gastric cancer risk, we conducted a case-control study by genotyping five potential functional SNPs, three in ERCC1 and two in ERCC4, in an Eastern Chinese Han population of 1125 gastric cancer cases and 1196 cancer-free controls.
Materials and Methods
The subjects were recruited from an ongoing case-control study as described previously . This study included 1125 unrelated ethnic Han Chinese patients with newly diagnosed and histopathologically confirmed primary gastric adenocarcinoma recruited from Fudan University Shanghai Cancer Center (FUSCC) between January 2009 and March 2011. All the patients were from Eastern China, including Shanghai, Zhejiang, Jiangsu and the surrounding regions. Because gastric adenocarcinoma accounts for nearly 98.5% of all gastric cancer patients seen at our hospital, we excluded from participation those patients who had interstitialoma, gastric adenosquamous carcinoma, squamous cell carcinoma, metastasized cancer from other organs or any histopathologic diagnosis other than gastric adenocarcinoma. An additional 1196 controls of age- (±5 years) and sex-matched cancer-free ethnic Han Chinese with written informed consent were recruited from the Taizhou Longitudinal (TZL) study conducted at the same time period in Eastern China which was approved by the Human Ethics Committee of Fudan University as described previously . All blood samples from patients with gastric adenocarcinoma were provided by the tissue bank of FUSCC. All patients had signed a written informed consent for donating their biological samples to the tissue bank of FUSCC. The response rate was approximately 91% for cases and 90% for controls. This research protocol was approved by the Institutional Review Board of FUSCC.
SNP Selection and Genotyping
We selected potentially functional SNPs of interest by using the NCBI dbSNP database and SNPinfo with the following criteria: (1) the minor allele frequency reported in HapMap was ≥5% for Chinese subjects; (2) affecting transcription factor binding site (TFBS) activity in the putative promoter region; (3) affecting the microRNA (miRNA) binding site activity; and (4) not included in the published GWASs. For the ERCC1 gene, we chose rs2298881C>A that may affect the TFBS activity, in addition to other two widely investigated potentially functional SNPs (rs3212986C>A and rs11615G>A, the former may be associated with transcript stability alterations, while the latter may be associated with mRNA levels alterations). For the ERCC4 gene, we chose rs2276466C>G and rs6498486A>C; the former may affect the miRNA binding site activity, while the latter may affect the TFBS activity as predicted by the SNPinfo. These five potentially functional SNPs also captured other 49 SNPs in the nearby genes (Table S1 for ERCC1 and Table S2 for ERCC4) and were genotyped as described previously .
We used χ2 test to compare the differences in the frequencies of alleles and genotypes as well as demographic and other covariates between the cases and controls. The goodness-of-fit χ2 test was employed to calculate the Hardy–Weinberg equilibrium of genotype distributions in the controls. We calculated crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) by univariate and multivariate logistic regression models, respectively, to evaluate associations between the genotypes and risk of gastric adenocarcinoma with and without adjustment for and stratified by age, sex, smoking/drinking status, primary tumor site and clinical stage. We also performed homogeneity tests to detect the difference in risk estimates among subgroups. Because the selected SNPs appear to be in the same block, the haplotype analysis was not performed.
We calculated the false-positive report probability (FPRP)  for all significant findings. We set 0.2 as an FPRP threshold and assigned a prior probability of 0.01 to detect an OR of 1.50 (for risk effects) or 0.67 (for protective effects) for an association with genotypes under investigation. Only significant results with an FPRP value <0.2 were considered as noteworthy associations.
We performed all analyses with SAS software (version 9.1; SAS Institute, Cary, NC), and all statistical tests were two-sided, and P values less than 0.05 were considered statistically significant.
Characteristics of the Patients
The distributions of demographic characteristics of the subjects are presented in Table S3, and were similar to those presented in a previous report . Briefly, the current case-control study included 1196 cancer-free controls and 1125 gastric cancer cases, including 305 (27.1%) gastric cardia adenocarcinoma (GCA) and 820 (72.9%) NGCA, of which 476 (42.3%) were of TNM stage I+II, and 649 (57.7%) were of stage III+IV according to the 7th Edition of the AJCC .
Association between Selected SNPs and Gastric Cancer Risk
The genotype distributions of the five selected SNPs in cases and controls are shown in Table 1. The observed genotype distributions among the controls were agreed with the Hardy–Weinberg equilibrium (P = 0.550 for rs2298881, P = 0.859 for rs3212986, P = 0.990 for rs11615, P = 0.288 for rs2276466, and P = 0.398 for rs6498486). The genotype distributions were significantly different for ERCC1 rs2298881 (P = 0.037) and ERCC1 rs11615 (P = 0.0496) between the cases and controls, but not for the other three SNPs. When the rs2298881AA genotype was used as the reference, the C variant genotypes were associated with an increased risk of gastric adenocarcinoma (adjusted OR = 1.37, 95% CI = 1.08–1.74 for AC, adjusted OR = 1.26, 95% CI = 0.98–1.63 for CC and adjusted OR = 1.33; 95% CI = 1.05–1.67 for AC/CC after adjustment for age, sex, smoking and drinking status); when the rs11615GG genotype was used as the reference, the A variant genotypes were associated with an increased risk of gastric adenocarcinoma (adjusted OR = 1.25, 95% CI = 1.05–1.48 for AG, adjusted OR = 1.13, 95% CI = 0.78–1.63 for AA and adjusted OR = 1.23; 95% CI = 1.05–1.46 for AG/AA). Nevertheless, no associations with risk of gastric adenocarcinoma were found for other three SNPs (Table 1). Patients with 2–3 risk genotypes of ERCC1 had significant increased risk (adjusted OR = 1.56, 95% CI = 1.27–1.93), compared with those with only 0–1 risk genotypes.
We further evaluated the association between variant genotypes of three selected SNPs of ERCC1 and gastric cancer risk stratified by subgroups of age, sex, smoking and drinking status, tumor site and clinical stage, assuming a dominant genetic model based on the results from univariate analysis (Table 2). The rs2298881 variant AC/CC genotypes was associated with an increased risk (adjusted ORs unless otherwise specified) that was more evident in females (OR = 1.64, 95% CI = 1.05–2.56), current-smokers (OR = 1.49, 95% CI = 1.03–2.14), never drinkers (OR = 1.31, 95% CI = 1.00–1.71), NGCA (OR = 1.33, 95% CI = 1.03–1.71) and clinical stage I+II (OR = 1.58, 95% CI = 1.15–2.18). Quite similar results were found for rs11615 variant AG/AA genotypes, especially in the younger group (OR = 1.34, 95% CI = 1.07–1.69), females (OR = 1.38, 95% CI = 1.02–1.88), current-smokers (OR = 1.38, 95% CI = 1.06–1.80), ever drinkers (OR = 1.67, 95% CI = 1.20–2.31), NGCA (OR = 1.23, 95% CI = 1.03–1.47) and clinical stage I+II (OR = 1.33, 95% CI = 1.08–1.65). When all risk genotypes were combined into a new variable, we found that the patients carrying 2–3 risk genotypes had a more evident risk in older group (OR = 1.62, 95% CI = 1.19–2.20), females (OR = 1.82, 95% CI = 1.22–2.72), current-smokers (OR = 1.76, 95% CI = 1.26–2.47), ever drinkers (OR = 1.75, 95% CI = 1.16–2.64), NGCA (OR = 1.59, 95% CI = 1.26–2.01) and clinical stage I+II (OR = 1.78, 95% CI = 1.33–2.38). However, further homogeneity tests suggested no differences in the risk estimates among the strata without statistical evidence of interactions between these variables and the variant genotypes on the risk (data not shown).
The FPRP values at different prior probability levels for significant findings are shown in Table 3. For a prior probability of 0.01, assuming that the OR for specific genotype was 1.50 (risk) or 0.67 (protection), with statistical power of 0.361, the FPRP values were 0.007 for an association of the ERCC1 2–3 risk genotypes, with an increased risk of gastric adenocarcinoma in all individuals. We also found a significant association with gastric adenocarcinoma risk from never drinkers, NGCA and clinical stage I+II among those patients with 2–3 risk genotypes. These significant associations were considered as noteworthy, because the probability of a false-positive result was <20%. In contrast, greater FPRP values were observed for other significant associations between ERCC1 variants and gastric adenocarcinoma, suggesting some possible bias in these positive findings.
In this study, we found that ERCC1 rs2298881C and rs11615A variant genotypes were associated with a pronounced increased risk of gastric adenocarcinoma and that the increased risk associated with 2–3 risk genotypes was more evident in never drinkers, NGCA and clinical stage I+II. To our knowledge, this is the first study that ERCC1 intron rs2298881 C/A and coding region rs11615 G/A polymorphisms were found to be associated with gastric adenocarcinoma risk. Because ERCC1 plays a critical role in NER, these findings are biologically plausible.
ERCC1 has been mapped to chromosome 19q13.32, consisting of 10 exons and encoding a 297 acetaldehyde ammonia product, while ERCC4, located on 16p13.12, comprises 11 exons and encodes a 916 acetaldehyde ammonia product. ERCC1 encodes a subunit for the NER complex that was required for the incision step of the NER pathway , . Importantly, a heterodimer of ERCC1/XPF catalyzes the 5′ incision in the process of excising DNA lesions in recombinational DNA repair and in the repair of inter-strand crosslinks –. Thus, it is possible that functional ERCC1 variants may also play a role in cancer risk. For example, the ERCC1 Asn118Asn (rs11615) SNP has been reported to be associated with some specific subtype of lung cancer as well as early onset of lung cancer . Likewise, the ERCC1 19007 C (rs11615) allele was found to be associated with an elevated lung cancer risk in Asian populations . To date, in the published studies on associations between the ERCC1 rs11615T>C, rs3212986C>A and rs3212961A>C polymorphisms and cancer risk , only two studies focused on ERCC1 polymorphisms and risk of gastric cancer, both of which studied the rs11615T>C only in a relatively small Italy population, including one  with only 126 gastric cancer cases and the other of only 314 cases , but none of these two studies included the SNPs under investigation in the present study.
When we combined these three selected ERCC1 SNPs, we found that patients with 2–3 risk genotypes had significantly increased risk of gastric cancer compared with those with 0–1 risk genotypes. In the stratified analysis, we also found that the effect of the rs2298881 AC/CC genotypes on cancer risk was more evident in subgroups of females, current-smokers, never drinkers, NGCA and clinical stage I+II and that quite similar results were for rs11615 variant AG/AA genotypes, especially in younger subjects, females, current-smokers, ever drinkers, NGCA and clinical stage I+II. However, after FPRP was calculated for these significant findings, only the findings for never drinkers, NGCA and clinical stage I+II remained significant for the patients with 2–3 risk genotypes. Therefore, some of our findings from the stratified analysis may be chance findings because of reduced sample sizes in the subgroups. For example, we found that the ever-drinkers had a more evident risk (OR = 1.75, 95% CI = 1.16–2.64) than the never-drinkers (OR = 1.51, 95% CI = 1.18–1.93); however, based on the FPRP calculation, the risk found for the ever-drinkers was not significant for a prior probability of 0.01. This may be ascribed to the reduction of sample sizes, for there were more never-drinkers than ever-drinkers in our current study. Tobacco smoke-related carcinogens may induce various kinds of DNA damage that mainly removed by the NER pathway . If unrepaired, such DNA damage may lead to mutations and thus the initiation of carcinogenesis. For example, DNA adducts identified in fetal blood may increase subsequent risk of developing cancer in the adulthood . For current smokers, the effect of genetic instability on cancer risk may be augmented by accumulated DNA damage caused by continued cigarette smoking. Females may be more susceptible to second-hand smoke or affected by indoor air pollution from unventilated coal-fueled stoves and from cooking fumes . Gastric cancer includes NGCA and GCA, and GCA is localized to the gastroesophageal junction and differs from NGCA in epidemiological characteristics and clinical features , . Finally, the finding that patients with 2–3 ERCC1 risk genotypes had increased risk of having early-stage cancer may suggest that the observed risk could be genetic susceptibility as the cause for carcinogenesis of the target tissue rather than tumor progression that could be driven by additional mutational events .
In summary, this large hospital-based case-control study provided statistical evidence that ERCC1 rs2298881 and rs11615 SNPs, but not ERCC4 SNPs, were associated with gastric cancer risk in an Eastern Chinese population, particularly for never-drinkers, patients with NGCA and early clinical stages. However, the present study had several limitations. First, the patients were selected from FUSCC that did not have a well-defined catchment area for the cases in Eastern China as the TZL study for the controls, which may have selection bias and information bias. Second, only three potential functional SNPs of ERCC1 and two of ERCC4 were investigated in the present study, which did not cover all SNPs of the ERCC1/XPF complex. Finally, we did not have reliable and sufficient information on other environmental exposures, such as dietary intake, occupation and H. pylori infection, due to the nature of the retrospective study design. Because our patients were mainly from Eastern China, it would be ideal to have a multi-center based replication to validate our findings. Without such replication, our findings should be considered preliminary. With such preliminary findings from the present study, we hope more cancer research centers or laboratories in China and other regions of the world with a high incidence of gastric cancer to validate our findings with larger sample sizes, more complete information on dietary intake, occupation and H. pylori infection.
SNPs captured by the selected three ERCC1 functional SNPs as predicted by SNPinfo software.
SNPs captured by the selected two ERCC4 functional SNPs as predicted by SNPinfo software.
Conceived and designed the experiments: Q-YW Y-NW. Performed the experiments: JH YX. Analyzed the data: JH L-XQ. Contributed reagents/materials/analysis tools: JL X-YZ M-HS J-CW Y-JY LJ. Wrote the paper: JH Q-YW Y-NW.
- 1. Jemal A, Bray F (2011) Center MM, Ferlay J, Ward E, et al (2011) Global cancer statistics. CA Cancer J Clin 61: 69–90.
- 2. Hu Z, Ajani JA, Wei Q (2007) Molecular epidemiology of gastric cancer: current status and future prospects. Gastrointest Cancer Res 1: 12–19.
- 3. Tran GD, Sun XD, Abnet CC, Fan JH, Dawsey SM, et al. (2005) Prospective study of risk factors for esophageal and gastric cancers in the Linxian general population trial cohort in China. Int J Cancer 113: 456–463.
- 4. Gammon MD, Schoenberg JB, Ahsan H, Risch HA, Vaughan TL, et al. (1997) Tobacco, alcohol, and socioeconomic status and adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 89: 1277–1284.
- 5. Mayne ST, Navarro SA (2002) Diet, obesity and reflux in the etiology of adenocarcinomas of the esophagus and gastric cardia in humans. J Nutr 132: 3467S–3470S.
- 6. Chow WH, Blot WJ, Vaughan TL, Risch HA, Gammon MD, et al. (1998) Body mass index and risk of adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 90: 150–155.
- 7. Abnet CC, Freedman ND, Hu N, Wang Z, Yu K, et al. (2010) A shared susceptibility locus in PLCE1 at 10q23 for gastric adenocarcinoma and esophageal squamous cell carcinoma. Nat Genet 42: 764–767.
- 8. Sun X, Li F, Sun N, Shukui Q, Baoan C, et al. (2009) Polymorphisms in XRCC1 and XPG and response to platinum-based chemotherapy in advanced non-small cell lung cancer patients. Lung Cancer 65: 230–236.
- 9. Wood RD, Mitchell M, Sgouros J, Lindahl T (2001) Human DNA repair genes. Science 291: 1284–1289.
- 10. De Silva IU, McHugh PJ, Clingen PH, Hartley JA (2000) Defining the roles of nucleotide excision repair and recombination in the repair of DNA interstrand cross-links in mammalian cells. Mol Cell Biol 20: 7980–7990.
- 11. Friedberg EC (2001) How nucleotide excision repair protects against cancer. Nat Rev Cancer 1: 22–33.
- 12. Cleaver JE (2000) Common pathways for ultraviolet skin carcinogenesis in the repair and replication defective groups of xeroderma pigmentosum. J Dermatol Sci 23: 1–11.
- 13. Wang AT, Sengerova B, Cattell E, Inagawa T, Hartley JM, et al. (2011) Human SNM1A and XPF-ERCC1 collaborate to initiate DNA interstrand cross-link repair. Genes Dev 25: 1859–1870.
- 14. Wood RD (1997) Nucleotide excision repair in mammalian cells. J Biol Chem 272: 23465–23468.
- 15. van Duin M, de Wit J, Odijk H, Westerveld A, Yasui A, et al. (1986) Molecular characterization of the human excision repair gene ERCC-1: cDNA cloning and amino acid homology with the yeast DNA repair gene RAD10. Cell 44: 913–923.
- 16. Zhang L, Wang J, Xu L, Zhou J, Guan X, et al. (2012) Nucleotide excision repair gene ERCC1 polymorphisms contribute to cancer susceptibility: a meta-analysis. Mutagenesis 27: 67–76.
- 17. Ruzzo A, Canestrari E, Maltese P, Pizzagalli F, Graziano F, et al. (2007) Polymorphisms in genes involved in DNA repair and metabolism of xenobiotics in individual susceptibility to sporadic diffuse gastric cancer. Clin Chem Lab Med 45: 822–828.
- 18. Palli D, Polidoro S, D’Errico M, Saieva C, Guarrera S, et al. (2010) Polymorphic DNA repair and metabolic genes: a multigenic study on gastric cancer. Mutagenesis 25: 569–575.
- 19. Vineis P, Manuguerra M, Kavvoura FK, Guarrera S, Allione A, et al. (2009) A field synopsis on low-penetrance variants in DNA repair genes and cancer susceptibility. J Natl Cancer Inst 101: 24–36.
- 20. Shi TY, He J, Qiu LX, Zhu ML, Wang MY, et al. (2012) Association between XPF Polymorphisms and Cancer Risk: A Meta-Analysis. PLoS One 7: e38606.
- 21. He J, Qiu LX, Wang MY, Hua RX, Zhang RX, et al. (2012) Polymorphisms in the XPG gene and risk of gastric cancer in Chinese populations. Hum Genet 131: 1235–44.
- 22. Wang X, Lu M, Qian J, Yang Y, Li S, et al. (2009) Rationales, design and recruitment of the Taizhou Longitudinal Study. BMC Public Health 9: 223.
- 23. Wacholder S, Chanock S, Garcia-Closas M, El Ghormli L, Rothman N (2004) Assessing the probability that a positive report is false: an approach for molecular epidemiology studies. J Natl Cancer Inst 96: 434–442.
- 24. Washington K (2010) 7th edition of the AJCC cancer staging manual: stomach. Ann Surg Oncol 17: 3077–3079.
- 25. Reed E (1998) Platinum-DNA adduct, nucleotide excision repair and platinum based anti-cancer chemotherapy. Cancer Treat Rev 24: 331–344.
- 26. O’Donovan A, Davies AA, Moggs JG, West SC, Wood RD (1994) XPG endonuclease makes the 3′ incision in human DNA nucleotide excision repair. Nature 371: 432–435.
- 27. Friedberg EC (2003) DNA damage and repair. Nature 421: 436–440.
- 28. Isla D, Sarries C, Rosell R, Alonso G, Domine M, et al. (2004) Single nucleotide polymorphisms and outcome in docetaxel-cisplatin-treated advanced non-small-cell lung cancer. Ann Oncol 15: 1194–1203.
- 29. Deng Q, Sheng L, Su D, Zhang L, Liu P, et al. (2011) Genetic polymorphisms in ATM, ERCC1, APE1 and iASPP genes and lung cancer risk in a population of southeast China. Med Oncol 28: 667–672.
- 30. Neumann AS, Sturgis EM, Wei Q (2005) Nucleotide excision repair as a marker for susceptibility to tobacco-related cancers: a review of molecular epidemiological studies. Mol Carcinog 42: 65–92.
- 31. Sasco AJ, Vainio H (1999) From in utero and childhood exposure to parental smoking to childhood cancer: a possible link and the need for action. Hum Exp Toxicol 18: 192–201.
- 32. Thun MJ, Hannan LM, Adams-Campbell LL, Boffetta P, Buring JE, et al. (2008) Lung cancer occurrence in never-smokers: an analysis of 13 cohorts and 22 cancer registry studies. PLoS Med 5: e185.
- 33. Shi Y, Hu Z, Wu C, Dai J, Li H, et al. (2011) A genome-wide association study identifies new susceptibility loci for non-cardia gastric cancer at 3q13.31 and 5p13.1. Nat Genet 43: 1215–1218.
- 34. Rodriguez G, Bilbao C, Ramirez R, Falcon O, Leon L, et al. (2006) Alleles with short CAG and GGN repeats in the androgen receptor gene are associated with benign endometrial cancer. Int J Cancer 118: 1420–1425.