|Home | About | Journals | Submit | Contact Us | Français|
The known breast cancer (BC) susceptibility polymorphisms in FGFR2, TNRC9/TOX3, MAP3K1,LSP1 and 2q35 confer increased risks of BC for BRCA1 or BRCA2 mutation carriers. We evaluated the associations of three additional SNPs, rs4973768 in SLC4A7/NEK10, rs6504950 in STXBP4/COX11 and rs10941679 at 5p12 and reanalyzed the previous associations using additional carriers in a sample of 12,525 BRCA1 and 7,409 BRCA2 carriers. Additionally, we investigated potential interactions between SNPs and assessed the implications for risk prediction. The minor alleles of rs4973768 and rs10941679 were associated with increased BC risk for BRCA2 carriers (per-allele Hazard Ratio (HR)=1.10, 95%CI:1.03-1.18, p=0.006 and HR=1.09, 95%CI:1.01-1.19, p=0.03, respectively). Neither SNP was associated with BC risk for BRCA1 carriers and rs6504950 was not associated with BC for either BRCA1 or BRCA2 carriers. Of the nine polymorphisms investigated, seven were associated with BC for BRCA2 carriers (FGFR2, TOX3, MAP3K1, LSP1, 2q35, SLC4A7, 5p12, p-values:7×10−11-0.03), but only TOX3 and 2q35 were associated with the risk for BRCA1 carriers (p=0.0049, 0.03 respectively). All risk associated polymorphisms appear to interact multiplicatively on BC risk for mutation carriers. Based on the joint genotype distribution of the seven risk associated SNPs in BRCA2 mutation carriers, the 5% of BRCA2 carriers at highest risk (i.e. between 95th and 100th percentiles) were predicted to have a probability between 80% and 96% of developing BC by age 80, compared with 42-50% for the 5% of carriers at lowest risk. Our findings indicated that these risk differences may be sufficient to influence the clinical management of mutation carriers.
Pathogenic mutations in BRCA1 and BRCA2 confer elevated risks of breast and ovarian cancer. Cancer risk estimates have been found to vary by the age at diagnosis or the cancer site of the proband that led to the family ascertainment (1-3) and studies have demonstrated significant variation in the breast cancer risks between families that segregate mutations in BRCA1 and BRCA2, according to the strength of family history (2, 4). Such evidence suggests that genetic or other factors that cluster in families may modify the cancer risks conferred by BRCA1 and BRCA2 mutations. Direct evidence of such modifiers of risk has been demonstrated through recent large scale association studies conducted by the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA(5)). These studies evaluated common genetic variants (single nucleotide polymorphisms, SNPs) , which have been shown to be associated with breast cancer risk in the general population through genome-wide association studies (GWAS) (6-9). The CIMBA results suggest that of the six variants investigated so far (rs2981582 in FGFR2, rs3803662 in TOX3/TNRC9, rs889312 in MAP3K1, rs3817198 in LSP1, rs13281615 on 8q24 and rs13387042 on 2q35) only the TOX3 and 2q35 polymorphisms were associated with breast cancer risk for BRCA1 mutation carriers. Five of the polymorphisms – all but the variant in the 8q24 region – were associated with breast cancer risk for BRCA2 mutation carriers. The estimated relative risk for the 8q24 SNP was consistent with that in the general population but was not statistically significant.
Since these investigations, eleven other breast cancer susceptibility variants have been identified through GWAS (10-14) including three SNPs rs4973768 in SLC4A7/NEK10, rs6504950 in STXBP4/COX11 and rs10941679 on 5p12. To evaluate whether these three polymorphisms are also associated with breast cancer risk for BRCA1 and BRCA2 mutation carriers we genotyped these polymorphisms in the CIMBA cohort. We also genotyped additional mutation carriers for the six polymorphisms previously investigated by CIMBA(6, 7). Here we present the updated results based on a larger number of female BRCA1 and BRCA2 mutation carriers. We also evaluated the evidence of interactions between the polymorphisms and the implications for risk prediction in BRCA1 and BRCA2 mutation carriers.
Female carriers of pathogenic mutations in BRCA1 and BRCA2 were recruited through the CIMBA initiative(5). Thirty-nine (39) studies contributed data for mutation carriers who were successfully genotyped for one or more of the nine SNPs investigated. The large majority of carriers were recruited through cancer genetics clinics offering genetic testing, and enrolled into national or regional studies. Some carriers were identified by population-based sampling of cases, and some by community recruitment (e.g. in Ashkenazi Jewish populations). Eligibility to participate in CIMBA is restricted to carriers of pathogenic BRCA1 or BRCA2 mutations who were 18 years old or over at recruitment. Information collected included the year of birth; mutation description, including nucleotide position and base change; age at last follow-up; ages at breast and ovarian cancer diagnoses; and age or date at bilateral prophylactic mastectomy. Information was also available on the country of residence, which was defined to be the country of the clinic at which the carrier family was recruited to the study. Related individuals were identified through a unique family identifier. Women were included in the analysis if they carried mutations that were pathogenic according to generally recognized criteria(15). Women who self-reported as “non-white” and those who carried pathogenic mutations in both BRCA1 and BRCA2 were excluded from the current analysis. All carriers participated in clinical or research studies at the host institutions under ethically approved protocols. Further details of the CIMBA initiative can be found elsewhere(5).
Genotyping was performed using either the iPLEX or Taqman platforms. To ensure genotyping consistency, all genotyping centers were required to adhere to the CIMBA genotyping quality control criteria which are described in detail in Appendix 1 (Supplementary Material). After excluding samples that failed quality control, 19,934 unique mutation carriers (12,525 BRCA1, 7,409 BRCA2) from 39 studies had an observed genotype for one or more of the SNPs and were therefore included in the analysis (Supplementary Table 1).
The aim of the analysis was to evaluate the association between each genotype and breast cancer risk. The phenotype of each individual was therefore defined by her age at diagnosis of breast cancer or her age at last follow-up. For this purpose, individuals were censored at the age of the first breast cancer diagnosis, ovarian cancer diagnosis, or bilateral prophylactic mastectomy or the age at last observation. Mutation carriers censored at ovarian cancer diagnosis were considered unaffected. Since mutation carriers were not sampled randomly with respect to their disease status, standard methods of survival analysis (such as Cox regression) may lead to biased estimates of the hazard ratios (HR)(16). We therefore conducted the analysis by modelling the retrospective likelihood of the observed genotypes conditional on the disease phenotypes as previously described(15). The effect of each SNP was modeled either as a per-allele HR (multiplicative model) or as separate HRs for heterozygotes and homozygotes, and these were estimated on the log scale. Where there was evidence of deviation from the multiplicative model, dominant and recessive models were also fitted. The HRs were assumed to be independent of age (i.e. we used a Cox proportional-hazards model). The assumption of proportional hazards was tested by adding a “genotype × age” interaction term to the model in order to fit models in which the HR changed with age. Analyses were carried out with the pedigree-analysis software MENDEL(17). We examined between-study heterogeneity by comparing the models that allowed for study-specific log-hazard ratios against models in which the same log-hazard ratio was assumed to apply to all studies. All analyses were stratified by study group and country of residence and used calendar-year- and cohort-specific breast cancer incidence rates for BRCA1 and BRCA2 (4). Risk reducing salpingo-oophorectomy (RRSO) was not considered in the analysis as it is not expected to be associated with the underlying SNP genotype (i.e. it is not a confounder) and previous analyses of these SNPs suggested no marked effect in the associations after adjustment(6, 7). We used a robust variance-estimation approach to allow for the non-independence among related carriers(18).
To investigate whether our results were influenced by any of our assumptions we performed additional sensitivity analyses. If any of the SNPs were associated with disease survival, the inclusion of prevalent cases may influence the HR estimates. We therefore repeated our analysis by excluding mutation carriers diagnosed more than five years prior to the age at recruitment into the study.
We further investigated for interactions between the SNPs and estimated the absolute risk of developing breast cancer based on the joint distribution of all SNPs that were significantly associated with risk for either BRCA1 or BRCA2 mutation carriers. Details of these methods are described in appendix 2.
The proportions of the modifying variance explained by the set of associated SNPs were estimated by ln(c)/σ2, where c is the estimated coefficient of variation in incidences associated with SNP(19, 20) and σ2 is the estimated modifying variance (1.32 and 1.73 for BRCA1 and BRCA2 mutation carriers respectively(4)). We estimated the total proportion of the modifying variance due to all SNPs by adding the individual proportions, i.e. by assuming that the loci combined multiplicatively.
After the exclusions described in the methods section, a total of 12,525 BRCA1 and 7,409 BRCA2 mutation carries had an eligible genotype for at least one of the nine SNPs and were included in the analysis (total 19,934 mutation carriers, Supplementary Table 1). Of these 9,933 had an observed genotype at all nine SNPs. Subjects were followed until the first breast cancer diagnosis (10,546), ovarian cancer diagnosis (1,981) or bilateral prophylactic mastectomy (567). The remaining subjects were censored at the age they were last observed (6,840). Only individuals censored at a breast cancer diagnosis were assumed to be affected in the analysis. Table 1 summarizes the key characteristics of this CIMBA cohort.
The results for the three newly investigated polymorphisms in the SLC4A7/NEK10, 5p12, STXBP4/COX11 regions are shown in Table 2. rs4973768 in SLC4A7/NEK10 was associated with breast cancer risk for BRCA2 mutation carriers, where each copy of the minor allele was estimated to confer a HR of 1.10 (95% CI: 1.03-1.18, p-trend=0.006). There was no evidence that this SNP was associated with breast cancer risk for BRCA1 mutation carriers (HR 1.03, p-trend=0.26). There was no evidence of heterogeneity in the study HR estimates (p=0.08 and 0.66 for BRCA1 and BRCA2 respectively; Figures Figures11 and and2).2). Models which allowed for an age dependent HR did not fit better than the models with a constant HR (p=0.72 and 0.93 for BRCA1 and BRCA2 respectively).
The 5p12 SNP rs10941679 was also associated with breast cancer risk for BRCA2 mutation carriers (2df p=0.022 and p-trend=0.032). Although the HR estimate for the heterozygote carriers of the minor allele was greater than the risk for the homozygote carriers, there was no significant evidence that the heterogeneity model (separate HR parameter for heterozygote and homozygotes) fit better than the multiplicative model for the effect of the minor allele of this SNP (p=0.07). Under the multiplicative model, the per-allele HR was estimated to be 1.09 (95%CI: 1.01-1.19, p-trend=0.032). A model which assumed that the underlying model was dominant fitted equally well (HRdominant=1.15, 95%CI: 1.04-1.27, pdom=0.008). The 5p12 polymorphism was not associated with breast cancer for BRCA1 mutation carriers (HR 0.96 95%CI 0.90-1.02, p-trend=.16). There was no evidence that the HRs vary across studies (phet=0.33 and 0.77 for BRCA1 and BRCA2 respectively; Figures Figures11 and and2),2), or that the HRs vary with age for either BRCA1 or BRCA2 (p=0.45 and 0.37 respectively).
The STXBP4/COX11 SNP rs6504950 was not associated with breast cancer risk for either BRCA1 (per-allele HR=1.02, 95% CI:0.96-1.08, p-trend=0.59) or BRCA2 mutation carriers (per-allele HR=1.03, 95%CI:0.95-1.11, p-trend=0.47). The HRs did not vary significantly with age for either BRCA1 (p=0.15) or BRCA2 (p=0.59). There was no evidence of heterogeneity in the HR estimates between studies (phet= 0.43 and 0.10 for BRCA1 and BRCA2 respectively, Figure Figure11 and and22).
To investigate whether our results may have been biased by the inclusion of prevalent cancers we repeated the analysis after excluding those who were diagnosed with breast or ovarian cancer more than 5 years prior to their recruitment into the study (i.e. long-term survivors). Individuals from studies in which the date/age at recruitment was not provided were also excluded from this analysis. The results for all three SNPs are summarised in Supplementary Table 2. The HR estimates were very similar to the analysis which included prevalent cancer patients. However, the p-values were larger and the 5p12 SNP was no longer significantly associated with breast cancer risk (p-trend=0.13, p-dominant=0.05) due to the smaller number of mutation carriers included in this analysis.
The updated results for SNPs rs2981582 in FGFR2, rs3803662 in TOX3/TNRC9, rs889312 in MAP3K1, rs3817198 in LSP1, rs13281615 in 8q24 and rs13387042 in 2q35, which include additional mutation carriers genotyped since they were originally published, are shown in table 3. The sample size increase varied from 1347 to1840 mutation carriers for the latest published SNPs in LSP1, 8q24 and 2q35 and from 3413 to 3854 mutation carriers for SNPs in FGFR2, TOX3/TNRC9 and MAP3K1. The pattern of associations of these SNPs with breast cancer risk for BRCA1 and BRCA2 mutation carriers were similar to that seen in the previously published CIMBA analyses, with the same SNPs significantly associated at the 5% level(6, 7). In the combined set of BRCA1 mutation carriers, only the TOX3/TNRC9 and 2q35 polymorphisms were associated with risk (p-trend=0.0049 and 2df p=0.01 respectively). In contrast, five of the six SNPs were associated with the risk of developing breast cancer in the combined set of BRCA2 mutation carriers. The most significant association was for the FGFR2 polymorphism (p-trend=6.8×10−11) in which each copy of the minor allele was estimated to confer a HR of 1.30 (95%CI:1.20-1.40), followed by TOX3/TNRC9 (per-allele HR=1.17, 95%CI: 1.07-1.27, p-trend=0.00029). These two SNPs had the largest increase in sample size since the previous analysis, and the significance of each association was correspondingly greater (p-trend=1.7×10−8 and 0.009 in the previous analysis for FGFR2 and TOX3/TNRC9 respectively). The significance of associations between the other SNPs (LSP1, MAP3K1, 2q35) and breast cancer risk for BRCA2 mutation carriers were similar to those reported previously (Table 3). The 8q24 SNP was not associated with breast cancer risk for BRCA2 mutation carriers (per-allele HR=1.06 95%CI 0.98-1.13, p-trend=0.13), but the number of additional BRCA2 mutation carriers included in this analysis was only 628, and the 95%CI still included the estimated relative risk in population-based studies. For all SNPs except TNRC9/TOX3, the inclusion of newly genotyped mutation carriers resulted in somewhat attenuated HR estimates, but narrower confidence intervals. The dominant model remained the most parsimonious model for the 2q35 SNP for both BRCA1 and BRCA2 carriers.
We evaluated all pairwise interactions between the SNPs that were associated with breast cancer risks for BRCA1 and BRCA2 separately (Supplementary Table 3). There was no evidence of any departure from a log-additive model for the TOX3/TNRC9 and 2q35 SNPs on the breast cancer risk for BRCA1 mutation carriers (p=0.22) or for any pairwise combination of the seven SNPs associated with BRCA2 breast cancer risk (p≥0.07).
Figure 3A shows the distribution of the combined HR across the 7 SNPs associated with breast cancer for BRCA2 mutation carriers, based on the estimates from the CIMBA sample and assuming that all SNPs interact multiplicatively. The HR varied from 1 for BRCA2 mutation carriers who were homozygous for the protective allele at all loci, to 5.75 for those who were homozygous for the risk allele at all loci. The median, 5th and 95th percentile HRs were 1.9, 1.3 and 3.0 respectively. Figure 3B translates the combined HRs into absolute risks of developing breast cancer by age 80. The estimated risk of developing breast cancer by 80 for BRCA2 mutation carriers varies from 42 to 96%. The median cumulative breast cancer risk is 64%, (5% and 95% percentile risk 50% and 80% respectively). Figure 4 shows the age-specific cumulative risks of developing breast cancer in BRCA2 mutation carriers by the combined genotype distribution at the seven associated SNPs. The risk of developing breast cancer by age 50 for the 5% of the mutation carriers at lowest risk is between 10-13%, compared with 29-47% for the 5% of the mutation carriers at highest risk. For comparison, we computed the cumulative risks using a risk score based on the published per-allele odds ratios for each SNP (all nine) in population-based studies (Supplementary Figure 1). The predicted combined HR and cumulative risks based on the median, the 5% and 95% percentiles of the genotype distribution were similar to those based on the CIMBA estimates.
The average risk of developing breast cancer for BRCA1 mutation carriers by age 80 was previously estimated to be approximately 66%(4). Based on the combined TOX3/TNRC9 – 2q35 genotype distribution, 13% of BRCA1 mutation carriers who were homozygous for the protective allele at both loci will have a risk of developing breast cancer of 61%, compared with 72% for the 2% of the BRCA1 mutation carriers who have the at-risk genotype at both loci.
We have investigated nine breast cancer susceptibility polymorphisms identified through genome wide association studies, for their associations with breast cancer risk for BRCA1 and BRCA2 mutation carriers. Of the three new polymorphisms investigated, the SLC4A7/NEK10 and 5p12 SNPs were associated with breast cancer risk for BRCA2 mutation carriers. In each case, the per-allele HR was similar to the published relative risks in population-based studies. For BRCA1 mutation carriers neither SNP showed an association with breast cancer risk, and in each case the 95%CI for the HR excluded the published point estimate for the general population. The STXBP4/COX11 SNP was not associated with breast cancer risk for either BRCA1 or BRCA2 mutation carriers. However, we cannot rule out that this SNP confers a HR for breast cancer in BRCA2 mutation carriers similar to the odds ratio estimated from population based studies as our confidence interval includes the 0.95 OR estimate(10). Given the magnitude of the effect in population-based studies, the current CIMBA sample of BRCA2 mutation carriers would have limited power to detect such an association (power of 31% at a 0.05 significance level). The estimated effects were not materially altered by inclusion of prevalent breast cancer patients in the analysis.
We have also incorporated newly-recruited mutation carriers in the analysis of the six SNPs that we previously investigated (FGFR2, TNRC9/TOX3, MAP3K1, LSP1, 8q24 and 2q35)(6, 7). The conclusions from these analyses were qualitatively similar to those previously reported, but there were some differences in the estimated HRs for the risk associated SNPs. With the exception of TOX3/TNRC9 in BRCA2, the HRs were somewhat attenuated perhaps reflecting a “winner’s curse” effect (i.e. HR overestimation) in the original investigation(21). The addition of new samples strengthened the associations for the FGFR2 and TOX3/TNRC9 SNPs which are the SNPs with largest estimated HRs, but the association p-values increased marginally for the other SNPs.
We focused on the associations of these SNPs with the risk of breast cancer for BRCA1 and BRCA2 mutation carriers. For this purpose, individuals who developed ovarian cancer first, were censored at the ovarian cancer diagnosis and were assumed to be unaffected in the analysis. If any of these polymorphisms were associated with ovarian cancer risk, this could potentially lead to biased estimates of the breast cancer HRs. However, previous analyses of these SNPs, that excluded mutation carriers who developed ovarian cancer, yielded similar HR estimates to the analysis that included these carriers (6). Moreover, there is no evidence from population based studies of ovarian cancer that any of these SNPs are associated with ovarian cancer risk in the general population (22, 23). A separate CIMBA study to estimate the effects of these polymorphisms on ovarian cancer risk for mutation carriers, assessed within a competing risks analysis framework is currently ongoing.
The associations between the nine SNPs and breast cancer risk differed substantially between BRCA1 and BRCA2 mutation carriers. Seven of the polymorphisms were associated with the risk of developing breast cancer for BRCA2 mutation carriers (FGFR2, TOX3/TNRC9, MAP3K1, LSP1, 2q35, SLC4A7/NEK10, 5p12). However, despite the larger sample size for BRCA1 carriers, only TOX3/TNRC9 and 2q35 were associated with the risk of breast cancer for BRCA1 mutation carriers. Significant differences in the HR between BRCA1 and BRCA2 were observed for FGFR2 (p = 3×10−6), MAP3K1 (p = 0.03) and 5p12 (p = 0.01). We have previously suggested that such differences could be explained by the differential effects of these SNPs by tumor subtype, specifically by ER status. Analyses by the Breast Cancer Association Consortium have indicated that many of the susceptibility loci confer higher relative risks for ER-positive disease, with weaker or absent association for ER-negative disease(24). Interestingly, the TOX3 and 2q35 SNPs, which exhibit associations for BRCA1 carriers, show the strongest evidence for association with ER-negative breast cancer risk in the general population, consistent with the observation that BRCA1 tumors are predominantly ER-negative (while BRCA2 tumors are predominantly ER-positive)(25). More specifically, these two SNPs were the only SNPs associated significantly with breast cancer expressing basal markers [Garcia-Closas, personal communication], the predominant subtype of breast cancer in BRCA1 carriers. The 5p12 and SLC4A7/NEK10 SNPs analyzed in the current study also conferred higher relative risks for ER-positive disease, consistent with this hypothesis(10, 11). Our results therefore provide further evidence for the distinct nature of the BRCA1related breast tumors. Overall, the seven SNPs associated with breast cancer risk for BRCA2 mutation carriers were estimated to account for approximately 4% of the genetic variability of breast cancer in BRCA2, while the TOX3/TNRC9 and 2q35 were estimated to account for 0.4% of the genetic variability in breast cancer risk in BRCA1. The estimated contribution to BRCA1 breast cancer risk variability is slightly lower than previously estimated(7), as a result of the attenuated HR estimates in the present analysis.
Each polymorphism was estimated to confer a modest HR. The largest per allele HR estimate was 1.30, for the FGFR2 association for BRCA2 mutation carriers. However, the combined effect of the susceptibility variants on risk can be much larger. Analysis of interactions between pairs of loci indicated that the combined effects were consistent with a multiplicative model. By defining a risk score based on this assumption, we estimated empirically that the highest 5% of the risk distribution had a HR of 2.64 (95%CI: 1.83-3.80, p=2.3×10−7) compared with the lowest 5%; this is very close to the predicted HR based on an assumed multiplicative model. We also conducted a similar analysis based on the estimated RRs from population studies, and the quantile-specific risk estimates were similar, indicating that the HRs were not exaggerated due to overfitting. Since we only considered pairwise interactions, it is possible that more complex interactions have been missed. However, given our results from the pairwise interactions and empirical score analysis, the multiplicative assumption seems plausible. A model with higher order interactions could lead to more powerful discrimination, but even with a study of this size there is insufficient power to fit higher order interactions reliably.
As BRCA2 mutations confer elevated risks of breast cancer, the combined HR estimates translate to large differences in the absolute risk of developing breast cancer between genotypes. Based on the combined effects of the seven SNPs we estimate that the 5% of BRCA2 mutation carriers at lowest risk will have a lifetime risk of developing breast cancer of 50% or lower whereas the 5% at highest risk will have a lifetime risk of 80% or higher. Such differences in risk could potentially be informative for genetic counselling purposes for classifying BRCA2 mutation carriers into different risk groups(26). A previous segregation analysis estimated that, based on the assumed distribution of modifiers of breast cancer risk, BRCA2 mutation carriers at the 5th percentile of risk distribution will have lifetime risk of developing the disease of 23% and those at the 95th percentile will have a lifetime risk of almost 100%(4). This analysis suggests that much greater improvements in risk profiling of carriers could be realised in the future if further modifiers of risk are identified. In contrast to BRCA2, only a limited number of risk modifying polymorphisms have been identified for BRCA1. This could reflect the fact that GWAS have so far focused on breast cancer patients unselected for tumor subtypes. Ongoing GWAS in BRCA1 mutation carriers and in ER-negative disease in the general population will be valuable in this respect.
In summary, our results indicate that the majority of the common breast cancer susceptibility variants identified through GWAS are associated with breast cancer risk for BRCA2 mutation carriers, to a similar relative extent as in the general population. Their combined effect results in substantial risk differences in absolute risk among SNP genotype categories. Such differences could inform genetic counselling and may lead to improved management of mutation carriers. Future studies in both the general population and mutation carriers that include GWAS, denser genotyping, exome and whole genome sequencing are likely to identify further variants associated with cancer risk for mutation carriers and will ultimately lead to more accurate risk prediction for these individuals.
Cancer Research - UK provided financial support for this work. ACA is a Senior Cancer Research UK Cancer Research Fellow. DFE is Cancer Research – UK Principal Research Fellow.
University of California Irvine (UCI): SLN and YCD were supported by NIH CA74415
MAGIC: The study is supported by NIH grants R01-CA083855 and R01-CA10277
Mayo Clinic Study (MAYO): The Mayo Clinic study was supported in part by the Breast Cancer Research Foundation (BCRF), a grant from Susan G. Komen for the Cure, the Mayo Clinic Breast Cancer SPORE (P50-CA116201) and NIH grants CA122340 and CA128978 to FJC.
CONsorzio Studi Italiani Tumori Ereditari Alla Mammella (CONSIT TEAM): The Italian study (CONsorzio Studi Italiani Tumori Ereditari Alla Mammella, CONSIT TEAM) is funded in part by grants from Fondazione Italiana per la Ricerca sul Cancro (Special Project “Hereditary tumors”), Associazione Italiana per la Ricerca sul Cancro (4017), Ministero della Salute (RFPS-2006-3-340203, Extraordinary National Cancer Program 2006 “Alleanza contro il Cancro”, and “Progetto Tumori Femminili”), Ministero dell’Universita’ e Ricerca (RBLAO3-BETH) and by funds from Italian citizens who allocated the 5×1000 share of their tax payment in support of the Fondazione IRCCS Istituto Nazionale Tumori, according to Italian laws (INT-Institutional strategic projects “5×1000”). CONSIT TEAM acknowledges Marco Pierotti, and Carla B. Ripamonti of the Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Bernardo Bonanni of the Istituto Europeo di Oncologia, Milan, Italy; Barbara Pasini of University of Turin, Turin, Italy; Laura Tizzoni and Loris Bernard of the Cogentech, Consortium for Genomic Technologies, Milan, Italy.
ModSqaud: C.I.S. is supported by the Mayo Rochester Early Career Development Award for Non-Clinician Scientists. We acknowledge the contributions of Petr Pohlreich and Zdenek Kleibl (Department of Biochemistry and Experimental Oncology, First Faculty of Medicine, Charles University, Prague, Czech Republic) and the support of the Grant Agency of the Czech Republic, project No. GP301/08/P103 (to M.Z.). We acknowledge the contribution of Kim De Leeneer and Anne De Paepe. This research was supported by grant 1.5.150.07 from the Fund for Scientific Research Flanders (FWO) to Kathleen Claes and by grant 12051203 from the Ghent university to Anne De Paepe. Bruce Poppe is Senior Clinical Investigator of the Fund for Scientific Research of Flanders (FWO - Vlaanderen). Kim De Leeneer is supported by the Vlaamse Liga tegen Kanker through a grant of the Foundation Emmanuel van der Schueren. L.F., Machackova Eva, and Lukesova Miroslava’s are supported through the Ministry of Health grant CR-MZ0 MOU 2005.
National Cancer Institute (NCI):: The research of Drs. PL Mai and MH Greene was supported by the Intramural Research Program of the US National Cancer Institute, and by support services contracts NO2-CP-11019-50 and N02-CP-65504 with Westat, Inc, Rockville, MD. Genoptyping of NCI DNA samples was performed by NCI’s Core Genotyping Facility, Gaithersburg, MD.
Ontario Cancer Genetics Network (OCGN): We wish to thank Mona Gill, Lucine Collins, Nalan Gokgoz, Teresa Selander, Nayana Weerasooriya and members of the Ontario Cancer Genetics Network for their contributions to the study.
Sheba Medical Center Study (SMC): The SMC study was supported in part by the Israel Cancer Association (ICA).
SWE-BRCA: SWE-BRCA collaborators: Per Karlsson, Margareta Nordling, Annika Bergman and Zakaria Einbeigi, Gothenburg, Sahlgrenska University Hospital; Sigrun Liedgren, Linkoping University Hospital; Niklas Loman, Håkan Olsson, Ulf Kristoffersson, Helena Jernström, Katja Harbst and Karin Henriksson, Lund University Hospital; Brita Arver, Anna von Wachenfeldt, Annelie Liljegren and Gisela Barbany-Bustinza Stockholm, Karolinska University Hospital; Henrik Grönberg, Eva-Lena Stattin, and Monica Emanuelsson, Umea University Hospital; Hans Ehrencrona, Richard Rosenquist Brandell, and Niklas Dahl, Uppsala University Hospital
University of Pennsylvania (UPENN): Breast Cancer Research Foundation (to KLN); Cancer Genetics Network (to SMD), Marjorie Cohen Foundation (to SMD)
International Hereditary Cancer Center (IHCC): IHCC was supported by Grant PBZ_KBN_122/P05/2004
Spanish National Cancer Center (CNIO): We thank R.M. Alonso, G Pita and R.M. Milne for their assistance. This study was partially supported by Fundación Mutua Madrileña, Asociación Española Contra el Cáncer and the Spanish Ministry of Science and Innovation (FIS PI08 1120). Funded in part by the Basque Foundation for Health Innovation and Research (BIOEF): BIO07/CA/006.
Deutsches Krebsforschungszentrum (DKFZ) study. The DKFZ study was supported by the DKFZ.
Epidemiological study of BRCA1 & BRCA2 mutation carriers (EMBRACE): Douglas F. Easton is the PI of the study. EMBRACE collaborators: North of Scotland Regional Genetics Service, Aberdeen: Helen Gregory, Zosia Miedzybrodzka. West Midlands Regional Clinical Genetics Service, Birmingham: Carole McKeown, Laura Boyes. South West Regional Genetics Service, Bristol: Alan Donaldson. Medical Genetics Services for Wales, Cardiff: Alexandra Murray, Mark Rogers, Emma McCann. St James’s Hospital, Dublin & National Center for Medical Genetics, Dublin: David Barton. Peninsula Clinical Genetics Service. Exeter: Carole Brewer, Emma Kivuva, Anne Searle, Selina Goodman. West of Scotland Regional Genetics Service, Glasgow: Victoria Murday, Nicola Bradshaw, Lesley Snadden, Mark Longmuir, Catherine Watt, Sarah Gibson. South East Thames Regional Genetics Service, Guys Hospital London: Louise Izatt, Chris Jacobs, Caroline Langman. Leicestershire Clinical Genetics Service, Leicester: Julian Barwell. Yorkshire Regional Genetics Service, Leeds: Carol Chu, Tim Bishop, Julie Miller. Merseyside & Cheshire Clinical Genetics Service. Liverpool: Ian Ellis. Manchester Regional Genetics Service, Manchester: Felicity Holt. North East Thames Regional Genetics Service, NE Thames: Alison Male, Lucy Side, Anne Robinson. Nottingham Center for Medical Genetics, Nottingham: Carol Gardiner. Northern Clinical Genetics Service, Newcastle: Fiona Douglas, Oonagh Claber. Oxford Regional Genetics Service, Oxford: Diane McLeod, Dorothy Halliday, Sarah Durrell, Barbara Stayner. The Institute of Cancer Research and Royal Marsden NHS Foundation Trust: Ros Eeles, Susan Shanley, Nazneen Rahman, Richard Houlston, Elizabeth Bancroft, Lucia D’Mello, Elizabeth Page, Audrey Ardern-Jones, Kelly Kohut, Jennifer Wiggins. Elena Castro, Lisa Robertson. North Trent Clinical Genetics Service, Sheffield: Oliver Quarrell, Cathryn Bardsley. South West Thames Regional Genetics Service, London: Sheila Goff, Glen Brice, Lizzie Winchester. Wessex Clinical Genetics Service. Princess Anne Hospital, Southampton: Diana Eccles, Anneke Lucassen, Gillian Crawford, Emma Tyler, Donna McBride. D.F.E., S.P., M.C., D.F. and C.O. are funded by Cancer Research-UK Grants C1287/A10118 and C1287/A8874. D.C. is supported by Cancer Research-UK Grant C8197/A10123.
The Hereditary Breast and Ovarian Cancer Research Group Netherlands (HEBON): HEBON Collaborating Centers: Coordinating center: Netherlands Cancer Institute, Amsterdam: Senno Verhoef, Martijn Verheus, Laura J. van ‘t Veer, Flora E. van Leeuwen; Erasmus Medical Center, Rotterdam: Margriet Collée, Ans M.W. van den Ouweland, Agnes Jager, Maartje J. Hooning, Madeleine M.A. Tilanus-Linthorst, Caroline Seynaeve; Leiden University Medical Center, Leiden: Juul T. Wijnen, Maaike P. Vreeswijk, Rob A. Tollenaar; Radboud University Nijmegen Medical Center, Nijmegen: Marjolijn J. Ligtenberg; University Medical Center Utrecht, Utrecht: Margreet G. Ausems; Amsterdam Medical Center: Theo A. van Os; VU University Medical Center, Amsterdam: Johan J.P. Gille, Quinten Waisfisz; University Hospital Maastricht, Maastricht: Encarna B. Gomez-Garcia, Cees E. van Roozendaal; University Medical Center Groningen University: Jan C. Oosterwijk, Annemarie H van der Hout, Marian J. Mourits; The Netherlands Foundation for the detection of hereditary tumours, Leiden, the Netherlands: Hans F. Vasen. The HEBON study is supported by the Dutch Cancer Society grants NKI 1998-1854, NKI 2004-3088 and NKI 2007-3756.
Fox Chase Cancer Center (FCCC): A.K.G. was funded by SPORE P-50CA83638, U01CA69631, 5U01CA113916, and the Eileen Stein Jacoby Fund.
Breast Cancer Family Registry (BCFR): This work was supported by the National Cancer Institute, National Institutes of Health under RFA-CA-06-503 and through cooperative agreements with members of the Breast Cancer Family Registry and Principal Investigators, including Cancer Care Ontario (U01 CA69467), Columbia University (U01 CA69398), Fox Chase Cancer Center (U01 CA69631), Huntsman Cancer Institute (U01 CA69446), Northern California Cancer Center (U01 CA69417), University of Melbourne (U01 CA69638), and Research Triangle Institute Informatics Support Center (RFP No. N02PC45022-46). Samples from the FCCC, HCI, and NCCC were processed and distributed by the Coriell institute for medical research. The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the BCFR, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government or the BCFR.
Genetic Modifiers of cancer risk in BRCA1/2 mutation carriers (GEMO): The GEMO study is supported by the Ligue National Contre le Cancer, Association for International Cancer Research Grant AICR-07-0454 and the Association “Le cancer du sein, parlons-en!” Award. We wish to thank all the GEMO collaborating groups for their contribution to this study. GEMO Collaborating Centers are: Coordinating Centres, Unité Mixte de Génétique Constitutionnelle des Cancers Fréquents, Centre Hospitalier Universitaire de Lyon / Centre Léon Bérard, & UMR5201 CNRS, Université de Lyon, Lyon: Laure Barjhoux, Sophie Giraud, Mélanie Léone, Sylvie; and INSERM U509, Service de Génétique Oncologique, Institut Curie, Paris: Marion Gauthier-Villars, Claude Houdayer, Virginie Moncoutier, Muriel Belotti. Institut Gustave Roussy, Villejuif: Brigitte Bressac-de-Paillerets, Audrey Remenieras, Véronique Byrde, Olivier Caron, Gilbert Lenoir. Centre Jean Perrin, Clermont–Ferrand: Yves-Jean Bignon, Nancy Uhrhammer. Institut Paoli Calmettes, Marseille: Violaine Bourdon, François Eisinger. Groupe Hospitalier Pitié-Salpétrière, Paris: Florence Coulet, Chrystelle Colas, Florent Soubrier. CHU de Arnaud-de-Villeneuve, Montpellier: Isabelle Coupier. Centre Oscar Lambret, Lille: Jean-Philippe Peyrat, Joëlle Fournier, Françoise Révillion, Philippe Vennin, Claude Adenis. Centre René Huguenin, St Cloud: Etienne Rouleau, Rosette Lidereau, Liliane Demange. Centre Paul Strauss, Strasbourg: Danièle Muller, Jean-Pierre Fricker. Institut Bergonié, Bordeaux: Michel Longy, Nicolas Sevenet. Institut Claudius Regaud, Toulouse: Christine Toulas, Rosine Guimbaud, Laurence Gladieff, Viviane Feillel. CHU de Grenoble: Christine Rebischung. CHU de Dijon: Cécile Cassini. CHU de St-Etienne: Fabienne Prieur. Hôtel Dieu Centre Hospitalier, Chambéry: Sandra Fert Ferrer.
Copenhagen Breast Cancer Study (CBCS): We wish to thank Bent Ejlertsen, Mette K. Andersen and Susanne Kjaergaard for clinical data. The work was supported by the Neye Foundation
Gynecologic Oncology Group (GOG): This study was supported by National Cancer Institute grants of the Gynecologic Oncology Group Administrative Office (CA 27469) and the Gynecologic Oncology Group Statistical and Data Center (CA 37517). GOG’s participation was supported through funding provided by both intramural (Clinical Genetics Branch, DCEG) and extramural (Community Oncology and Prevention Trials Program – COPTRG) NCI programs. Genotyping of GOG DNA samples was performed by NCI’s Core Genotyping Facility. The technical expertise of Tim Sheehy, Amy Hutchinson is gratefully acknowledged.
Ohio State University Clinical Cancer Genetics (OSU CCG): This work was funded by the OSU Comprehensive Cancer Center. We thank Kevin Sweet and Caroline Craven for patient accrual and data management, the Human Genetics Sample Bank for sample preparation and the OSU Nucleic Acids Shared Resource for plate reads.
Istituto Oncologico Veneto - Hereditary Breast Ovarian Cancer Study (IOVHBOCS): The study was supported by the Ministero dell’Università e della Ricerca, Ministero della Salute and Alleanza Contro il Cancro.
N.N. Petrov Institute of Oncology (NNPIO): The work is supported by the Russian Foundation for Basic Research (grants 08-04-00369-a, 09-04-90402 and 10-04-92110-a), the Commission of the European Communities (grant PITN-GA-2009-238132) and through a Royal Society International Joint grant (JP090615).
Baltic Familial Breast Ovarian Cancer Consortium (BFBOCC): We acknowledge the Genome Database of Latvian Population, Latvian Biomedical Research and Study Centre for providing data and DNA samples for BFBOCC (LV) and Ramunas Janavicius (Vilnius University Hospital Santariskiu Clinics, Lithuania) for data and DNA samples for BFBOCC (LT).
UK and Gilda Radner Familial Ovarian Cancer Registries (UKGRFOCR): UKFOCR was supported by a project grant from CRUK to Paul Pharoah. We thank Carole Pye and Patricia Harrington for family recruitment and technical support. We’d like to acknowledge the Roswell Park Alliance Foundation for their continued support of the Gilda Radner Ovarian Family Cancer Registry. GRFOCR qould like to acknowledge Kirsten Moysich (Department of Cancer Prevention and Control) and Kunle Odunsi (Departments Gynecologic Oncology and Immunology).
Women’s Cancer Research Institute (WCRI): This work was supported by the American Cancer Society Early Detection Professorship and Entertainment Industry Foundation.
The German Consortium of Hereditary Breast and Ovarian Cancer (GC-HBOC): GC-HBOC is supported by a grant of the German Cancer Aid (grant 107054). We thank Juliane Köhler for her excellent technical assistance, Ellen Kirsch, Isabell Eisenhauer, Hans-Jörg Plendl, Thomas Neumann, Ulrike Siebers-Rehnelt, Doris Steinemann, Britta Skawran, Patricia Steiner and the 12 centers of the GC-HBOC for providing samples and clinical data.
Helsinki Breast Cancer Study (HEBCS): HEBCS thanks Tuomas Heikkinen and Dr. Carl Blomqvist for their help with the patient data and samples. The HEBCS study has been financially supported by the Helsinki University Central Hospital Research Fund, Academy of Finland (132473), the Finnish Cancer Society, and the Sigrid Juselius Foundation.
Hospital Clinico San Carlos (HCSC): Trinidad Caldes and Miguel de la Hoya were supported by FIS 09/00859 and RD06/0020/0021 (RTICC; ISCIII) Spanish Ministry of Science and Innovation.
Interdisciplinary Health Research International Team Breast Cancer Susceptibility (INHERIT BRCAs): INHERIT collaborators: Francine Durocher, Rachel Laframboise, Marie Plante, Centre Hospitalier Universitaire de Quebec & Laval University, Quebec, Canada ; Peter Bridge, Molecular Diagnostic Laboratory, Alberta Children’s Hospital, Calgary, Canada; Jocelyne Chiquette, Hôpital du Saint-Sacrement, Quebec, Canada ; Bernard Lesperance, Hôpital du Sacré-Cœur de Montréal, Montréal, Canada. Jacques Simard- J.S. is Chairholder of the Canada Research Chair in Oncogenetics. This work was supported by the Canadian Institutes of Health Research for the “CIHR Team in Familial Risks of Breast Cancer” program and by the Canadian Breast Cancer Research Alliance-grant #019511.
Kathleen Cuningham Consortium for Research into Familial Breast Cancer (KCONFAB): We wish to thank Heather Thorne, Eveline Niedermayr, all the kConFab research nurses and staff, the heads and staff of the Family Cancer Clinics, and the Clinical Follow Up Study (funded by NHMRC grants 145684, 288704 and 454508) for their contributions to this resource, and the many families who contribute to kConFab. kConFab is supported by grants from the National Breast Cancer Foundation, the National Health and Medical Research Council (NHMRC) and by the Queensland Cancer Fund, the Cancer Councils of New South Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Western Australia , Amanda Spurdle is supported by an NHMRC Senior Research Fellowship, and Georgia Chenevix-Trench by an NHMRC Senior Principal Research Fellowship.
Grants: The Consortium of Investigators of Modifiers of BRCA1/2 is supported by grants from Cancer Research UK. All study-specific grants are listed in the acknowledgments section.
Conflicts of Interest: All authors declare no conflicts of interest.