|Home | About | Journals | Submit | Contact Us | Français|
Genome-wide association studies have identified several novel risk alleles for breast cancer. We hypothesized that genetic variants that are associated with breast cancer, a hormone-related disease, would also be associated with endometrial cancer, another hormone-related disease. We conducted a case-control study nested within the Nurses’ Health Study and the Women’s Health Study to investigate the associations between these seven newly identified risk alleles for breast cancer and endometrial cancer risk using 692 invasive endometrial cancer cases and 1,723 matched controls. We used conditional logistic regression to calculate odds ratios (ORs) and 95% confidence intervals (CIs) to assess the risk of endometrial cancer. In contrast to the breast cancer findings, we did not observe an increased risk of endometrial cancer. We observed an inverse association among rs2981582 (FGFR2) variant carriers (OR= 0.75 (95% CI: 0.60, 0.95)). We also observed a non-significant inverse association with rs889312 (MAP3K1) variant carriers (OR = 0.85 (95% CI: 0.68, 1.05)) and rs1219648 (FGFR2) variant carriers (OR= 0.86 (95% CI: 0.69, 1.06) and endometrial cancer risk. We did not observe associations with the other single nucleotide polymorphisms (SNPs) and endometrial cancer risk. Replication studies investigating these polymorphisms and endometrial cancer risk are warranted. However, our findings do suggest the potential importance of biological differences between endometrial and breast cancer with respect to the genes identified in the scans. The molecular mechanisms underlying these differences remain to be defined.
Estrogen is well-recognized to play an important role in the development and progression of breast and endometrial cancers. Endometrial and breast cancers are considered to be ‘estrogen-dependent tumors,’ as they share common risk factors, such as early age at menarche, late age at menopause, nulliparity, late age at first birth, long-term postmenopausal hormone use, and increased body mass index (for postmenopausal breast cancer), that have been shown to influence the levels and duration of exposure to sex steroid hormones.
Genetic factors are also likely to play a major role in the development of endometrial cancer. Endometrial cancer is part of the hereditary cancer syndrome, hereditary nonpolyposis colorectal cancer (HNPCC), which is attributable to the inheritance of rare, highly penetrant mutated DNA repair genes 1, 2. Given the small percentage (2-5%) of endometrial cancers attributable to this hereditary cancer syndrome, other genetic models may play a role in the development of endometrial cancer. Women with a family history of endometrial cancer have their risk increased by 1.5 to 3-fold 3-5 suggesting that family history is an important risk factor in the population. Variation in genes involved in steroid hormone metabolism influence the levels of estrogen and progesterone and thus may be associated with an increased risk of endometrial cancer.
Several genome-wide association studies (GWAS) have identified novel risk alleles for breast cancer 6-9. Easton et al. 6 initially scanned 408 breast cancer cases with a strong family history of breast cancer and 400 controls for 266,722 SNPs followed by a second stage in which 12,711 SNPs were genotyped in 3,990 breast cancer cases and 3,916 controls. A confirmation phase followed in which 30 of the highest ranking SNPs that achieved a combined P < 0.00002 (based on either the Cochran-Armitage or a 2 degree freedom test) were genotyped in 21,860 cases and 22,578 controls to reveal five novel, independent loci with strong evidence of an association for breast cancer (P < 10-7). Four of these loci contained possible causative genes (FGFR2, TNRC9, MAP3K1, LSP1). The fifth polymorphism (rs13281615) lies on chromosome 8q24 and is correlated with SNPs in a 110kb linkage disequilibrium (LD) block that has no known genes6. Hunter et al.7 genotyped 528,173 SNPs in 1,145 breast cancer cases and 1,142 controls and identified four SNPs in intron 2 of the FGFR2 gene, which were significantly associated with sporadic postmenopausal breast cancer risk. This association was replicated in an additional 1,776 cases and 2,072 controls from breast cancer case-control studies nested within three prospective cohorts 7. Stacey et al.8 genotyped approximately 300,000 SNPs in 1,600 Icelandic individuals and 11,563 controls and then tested the ten highest ranking SNPs in five replication sets. Two SNPs (rs3803662 (TNRC9) and rs13387042) were consistently associated with estrogen receptor-positive breast cancer 8. The rs13387042 SNP is on chromosome 2q35, and the LD block containing this polymorphism has no known genes or human RNAs 8. A recent scan of 6,145 breast cancer cases and 33,016 controls observed a significant association with the previously identified FGFR2 SNP rs12196487 and estrogen receptor-positive breast cancer 9. Stacey et al. 8 and Easton et al. 6 both identified the rs3803662 (TNRC9) SNP, and Hunter et al.7 and Easton et al. 6 identified the same principal locus in FGFR2, as the two respective SNPs have an r2 of 1.0 (rs1219648 and rs2981582) in the HapMap CEU samples. Taken together, the genome-wide association studies have revealed, with some consistencies, a number of SNPs that are associated with breast cancer risk.
We reasoned that genetic variants that are associated with breast cancer, a hormone-related disease, would also be associated with endometrial cancer, another hormone-related disease. Therefore, we conducted a nested case-control study within the Nurses’ Health Study and the Women’s Health Study to investigate the associations between these seven newly identified breast cancer risk alleles and endometrial cancer risk, using 692 invasive endometrial cancer cases and 1,723 matched controls.
The Nurses’ Health Study (NHS) is a prospective cohort study of 121,700 women enrolled in 1976. During 1989-90, blood samples were collected from 32,826 women. In 2000-2002, buccal cell samples were collected from 32,883 NHS women who had not provided a blood sample. Cases were women with pathologically confirmed invasive endometrial cancer that had been diagnosed after cohort inception and prior to June 1, 2004. Controls were randomly selected from participants who had not had a hysterectomy and were free of diagnosed cancer. Controls were matched to cases according to age, menopausal status and postmenopausal hormone use at blood draw, and type of biospecimen. Completion of the self-administered questionnaire was considered to imply informed consent, and written informed consent was received for the biospecimen samples.
The Women’s Health Study (WHS) is a completed randomized, double-blind, placebo-controlled trial investigating the benefits and risks of aspirin and vitamin E in the primary prevention of cancer and cardiovascular disease among 39,876 female health professionals, with enrollment beginning in April 1993. Blood samples have been collected from 28,345 women. WHS cases consisted of women with pathologically confirmed invasive endometrial cancer who had been diagnosed after blood collection and prior to June 1, 2002. Controls were randomly selected participants who had given a blood sample, had not had a hysterectomy, and were free of diagnosed cancer. Controls were matched to cases according to age, menopausal status and postmenopausal hormone use at time of blood draw. Written informed consent was obtained from all women before their entry into the trial. Each study protocol was approved by the Committee on Use of Human Subjects of the Brigham and Women’s Hospital (BWH), Boston, MA.
Genotyping was performed at the Dana Farber/Harvard Cancer Center High Throughput Genotyping Core. DNA was extracted from buffy coat/buccal cell fractions with a QIAGEN QIAmp Blood Kit (QIAGEN, Chatsworth, CA). Polymorphisms were genotyped by the 5′ nuclease assay (Taqman) with the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA). Laboratory personnel were blinded to case-control status, and 5% blinded quality control samples were inserted to validate genotyping procedures; concordance for blinded samples was 100%. The percentage of missing genotyping data was <5%.
Conditional logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) to assess the risk of endometrial cancer. In addition to the matching factors, analyses were adjusted for endometrial risk factors: body mass index (BMI) at diagnosis (kg/m2); age at menarche; parity/age at first birth; smoking status; menopausal status at diagnosis; age at menopause; postmenopausal hormone use; and first-degree family history of colorectal cancer. We created indicator variables for the different genotypes using individuals homozygous for the most common allele as the reference category. Gene dosage effects were modeled by assigning the values of 0, 1, and 2 to a genotype trend variable according to the subject’s number of variant alleles (zero, one, and two variant alleles, respectively). After cohort-specific analyses were completed, tests of heterogeneity were conducted to determine the appropriateness of pooling the two datasets for analysis. We used the DerSimonian and Laird random effects model to combine results from the two cohorts after testing for heterogeneity. Because of sample size considerations, we combined heterozygotes and homozygote variants in the interaction analyses. As an exploratory analysis, we tested pairwise statistical interactions between the different genotypes in conditional logistic regression models. We used a likelihood-ratio test (LRT) to compare nested models that included terms for all combinations of the genotypes to the models with indicator variables for the main effects only (nominal LRT). All p-values were two-sided. All analyses were restricted to whites, and numbers may vary for the different analyses due to missing genotype data. We used the SAS Version 9.1 software (SAS Institute, Cary, NC).
In summary, we have 555 NHS endometrial cancer cases and 1,312 matched controls (n = 247 cases and 494 controls from the NHS buccal cell cohort and 308 cases and 818 controls from NHS blood cohort), and 137 WHS endometrial cancer cases and 411 matched controls for a total sample size of 692 cases and 1,723 matched controls. Population characteristics of the two study populations stratified by case-control status and restricted to whites only are displayed in Table 1.
The frequency of the genotypes were similar to previously reported frequencies in white populations (Table 2) 6-9, and all genotypes were in accordance with Hardy-Weinberg equilibrium. All p-values for the tests of heterogeneity comparing the NHS and WHS results were >0.05. We observed an inverse association with rs2981582 (FGFR2) and endometrial cancer risk (OR = 0.75 (95% CI: 0.60, 0.95)) (Table 3). We also observed a suggestive inverse association with rs889312 (MAP3K1) (OR = 0.85 (95% CI: 0.68, 1.05)) and rs1219648 (FGFR2) (OR = 0.86 (95% CI: 0.69, 1.06)). The result with rs1219648 is not surprising as rs1219648 and rs2981582 have an r2 of 0.80 in our data. We did not observe associations with the additional four SNPs and endometrial cancer risk (Table 3). We did not observe any statistically significant gene-gene interactions in our exploratory analyses, although we had limited power to detect minor to modest associations.
We observed inverse associations with the rs2981582 (FGFR2) (ORtrend = 0.86 (95% CI: 0.73, 1.01)) and rs1219648 (FGFR2) (ORtrend = 0.87 (95% CI: 0.75, 1.01)) SNPs and endometrial cancer risk. In the breast cancer genome-wide association scans, Easton et al. 6 observed an ORtrend of 1.26 (95% CI: 1.23-1.30) for the rs2981582 polymorphism, and Hunter et al. 7 observed an ORtrend of 1.32 (95% CI: 1.17-1.49) for the rs1219648 polymorphism. Recently, Garcia-Closas et al. 10 observed positive associations with the rs2981582 and rs13281615 SNPs and estrogen receptor-positive breast cancer in a case-control study of 23,039 breast cancer cases and 26,273 controls from 20 studies. Fibroblast growth factor receptor 2, FGFR2, encodes a receptor tyrosine kinase and has been observed to be amplified and overexpressed in 5-10% of breast tumors 11-13. FGFR2 is expressed in the human endometrium 14, 15, and its expression in blood vessels in the secretory endometrium suggests involvement in the regulation of angiogenesis and blood vessel function in the human endometrium 15. Recent findings have revealed 11 different gain-of-function FGFR2 mutations in 30% of endometrial cancer cell lines and 10% of primary uterine tumors (primarily of the endometrioid histologic subtype), suggesting that the activation of FGFR2 is involved in endometrial tumorigenesis 16. The majority of somatic mutations that were identified resulted in receptor activation 16. Recently, Meyer et al. 17 observed that the rs2981582 and rs1219648 SNPs alter the binding of two transcription factors, Oct-1/Runx2 and C/EBPβ resulting in an increase in FGFR2 gene expression.
We also observed a non-significant inverse association with variant carriers of SNP rs889312 (OR= 0.85 (95% CI: 0.68, 1.05)) and endometrial cancer risk. SNP rs889312 is located in a LD block that contains the MAP3K1 (also known as MEKK) gene that encodes a serine/threonine kinase protein. This kinase has been shown to have a critical role in a network of phosphorylating enzymes integrating cellular responses to a number of mitogenic and metabolic stimuli, including insulin and many growth factors 18.
We did not perform gene-environment interactions as there is no prior biological evidence to suggest that the SNPs may interact differently with varying levels of environmental exposures. As an exploratory analysis, we investigated gene-gene interactions between the seven polymorphisms, and did not observe any significant associations.
We focused our efforts on the highly significant associations observed from the genome-wide association breast cancer studies and genotyped the seven polymorphisms in our endometrial cancer nested case-control study. The use of the NHS and WHS allowed us the opportunity to replicate the association between these polymorphisms and endometrial cancer risk simultaneously in two large, well-defined cohorts. Our study is the first prospective case-control study to investigate these SNPs in relation to endometrial cancer risk. However, we may have limited power to detect minor to modest genotype-disease associations; therefore, such associations cannot be ruled out. To assess the power of our study, we conducted power calculations by using Quanto19. We estimated that we had >80% power to identify an odds ratio of 1.23 per allele or 0.83 per allele in the pooled analysis (assuming a log additive model, a matched case-control study design, a two-sided test, minor allele frequencies ranging from 30 to 50%, and α = 0.05). In contrast to the breast cancer findings6-9, we did not observe an increased risk of endometrial cancer associated with the polymorphisms. Our findings do suggest the importance of biological differences between endometrial and breast cancer with respect to the genes identified in the scans yet replication studies investigating these polymorphisms and endometrial cancer risk are warranted.
We thank H. Ranu, P. Soule, and M. Chown for assistance, and we thank the participants in the Nurses’ Health Study and the Women’s Health Study for their dedication and commitment. This work is supported by National Institute of Health Grants: CA87969, CA49449, CA82838, NICHD K12 HD051959-01, a grant from the American Cancer Society: RSG-00-061-04-CCE.