Inherited BRCA1 mutations confer elevated cancer risk. Recent studies have identified genes that encode proteins that interact with BRCA1 as modifiers of BRCA1-associated breast cancer. We evaluated a comprehensive set of genes that encode most known BRCA1 interactors to evaluate the role of these genes as modifiers of cancer risk. A cohort of 2,825 BRCA1 mutation carriers was used to evaluate the association of haplotypes at ATM, BRCC36, BRCC45 (BRE), BRIP1 (BACH1/FANCJ), CTIP, ABRA1 (FAM175A), MERIT40, MRE11A, NBS1, PALB2 (FANCN), RAD50, RAD51, RAP80, and TOPBP1, and was associated with time to breast and ovarian cancer diagnosis. Statistically significant false discovery rate (FDR) adjusted P values for overall association of haplotypes (PFDR) with breast cancer were identified at ATM (PFDR = 0.029), BRCC45 (PFDR = 0.019), BRIP1 (PFDR = 0.008), CTIP (PFDR = 0.017), MERIT40 (PFDR = 0.019), NBS1 (PFDR = 0.003), RAD50 (PFDR = 0.014), and TOPBP1 (PFDR = 0.011). Haplotypes at ABRA1 (PFDR = 0.007), BRCC45 (PFDR = 0.016 and PFDR = 0.005 in two haplotype blocks), and RAP80 (PFDR < 0.001) were associated with ovarian cancer risk. Overall, the data suggest that genomic variation at multiple loci that encode proteins that interact biologically with BRCA1 are associated with modified breast cancer and ovarian cancer risk in women who carry BRCA1 mutations. Cancer Res; 71(17); 5792–805. ©2011 AACR.
Inherited mutations in BRCA1 may be necessary to explain the Mendelian pattern of breast cancer in some families but are not sufficient to completely describe interindividual variability in age-specific cancer risk. A number of studies suggest that modifier loci influence breast cancer penetrance among BRCA1 mutation carriers. The most convincing evidence of such modifiers to date includes those genes that encode proteins that interact directly with BRCA1 or BRCA2. A recent genome-wide association study (GWAS) identified a locus on chromosome 19p13 as a modifier of breast cancer risk for BRCA1 mutation carriers (1). The most significant single-nucleotide polymorphisms (SNP) in this analysis were those that were located in a region containing 3 genes, among them C19ofr6, which encodes the BRCA1 interactor MERIT40, an interactor of BRCA1. However, fine mapping has yet to be done in this region, and other loci, including ANKLE1, were also highly significantly associated with breast cancer. A RAD51 SNP has been previously shown to modify breast cancer risk for BRCA2 mutation carriers in a number of independent studies (2–4). We recently reported that variation in genes that encode BRCA1- or BRCA2-interacting proteins was associated with ovarian cancer risk, including ATM, BARD1, BRIP1, MRE11A, and RAD51 (5). Many of these proteins have also been associated with risk of developing cancers and other diseases associated with DNA damage in the general population (6–10). BRCA1 influences genome integrity and tumorigenesis through the actions of a number of protein complexes (Supplementary Fig. S1 and Supplementary Table S1). Because the proteins encoded by these loci are part of multiprotein complexes that interact with BRCA1 to influence DNA damage and repair pathways, these are biologically plausible modifiers of cancer risk in BRCA1 mutation carriers. Based on this knowledge, the goal of this research is to more fully explore the observation that genes encoding BRCA1-interacting proteins modify cancer risk in BRCA1 mutation carriers. Because of limited statistical power in the current data set, future studies will also include interactors of BRCA2 when sample sizes become adequate for these analyses. In the present article, we evaluated whether variation at BRCA1 interactor–encoding loci modulates BRCA1-associated cancer risk by using a haplotype-tagged SNP (htSNP) approach to fully evaluate genomic variability at each of these loci.
Materials and Methods
Participants and data
Seventeen centers have contributed data and DNA from North America, Israel, Australia, and Europe: Baylor-Charles A. Sammons Cancer Center, Dallas, TX; Beth Israel Deaconess Medical Center, Boston, MA; Chaim Sheba Medical Center, Tel-Hashomer, Israel; Beckman Institute at the City of Hope National Medical Center, Duarte, CA; Creighton University, Omaha, NE; Dana Farber Cancer Institute, Boston, MA; Epidemiological Study of BRCA1 and BRCA2 Mutation Carriers (EMBRACE), University of Cambridge, UK; NorthShore University HealthSystem Center for Medical Genetics, Evanston, IL; Fox Chase Cancer Center, Philadelphia, PA; Georgetown University, Washington, DC; Jonsson Comprehensive Cancer Center at the University of California, Los Angeles, CA; Mayo Clinic College of Medicine, Rochester, MN; Ontario Cancer Genetics Network, Toronto, ON; University of Chicago, IL; University of California, Irvine, CA; Helsinki University Central Hospital, Helsinki, Finland; University of Pennsylvania Health System, Philadelphia, PA; University of Texas Health Sciences, San Antonio, TX; University of Vienna, Austria; and Women's College Hospital, Toronto, Ontario, Canada.
All participants were identified via high-risk programs for clinical and research purposes. Participants were referred by clinicians or self-referred because they were perceived to be at risk for hereditary breast and/or ovarian cancer. Genetic counseling and testing were done under clinical and/or research protocols specific to the Institutional Review Board (IRB) guidelines of each center. All centers identified women who had tested positive for BRCA1 mutations and provided eligibility information to the University of Pennsylvania, the coordinating center. The coordinating center determined eligible participants. Eligible participants included women older than 18 years, with documented disease-associated mutations in BRCA1, who had never been diagnosed with cancer at any site prior to center ascertainment. As only a small proportion of our cohort (<5%) comes from minority groups, we included only participants who were white, including Hispanic, non-Hispanic, and Jewish.
The BRCA1 mutation status of all subjects was confirmed by direct mutation testing, and subjects provided full informed consent for this study under protocols approved by the human subjects review boards at each institution. Some participants were simultaneously consented for both research and clinical BRCA1 testing, whereas others were consented separately for clinical testing and for research participation. Women with BRCA1 variants of unknown clinical significance were excluded. Mutations were included in the analysis if they were pathogenic according to generally recognized criteria, including (i) mutations generating a premature termination codon as a result of a nonsense substitution, a frame shift due to small deletion or insertion, aberrant splicing, or large genomic rearrangement; (ii) mutations resulting in loss of expression due to deletion of promoter and transcription start site; (iii) large in-frame deletions spanning 1 or more exons caused by aberrant splicing or large genomic rearrangement; and (iv) missense mutations classified as pathogenic using the algorithms of Goldgar and colleagues (11) and Chenevix-Trench and colleagues (12).
Data were obtained on all eligible participants by using medical records, telephone interviews, and/or self-administered questionnaires and included information on reproductive and exposure history, including hormone use and smoking. Vital status, cancer diagnoses, and prophylactic surgery data were verified by review of medical records, operative notes, and/or pathology reports if available. Time from ascertainment to cancer diagnosis or censoring was random with respect to risk-reducing salpingo-oophorectomy (RRSO), risk-reducing mastectomy (RRM), cancer occurrence, or death. In addition, because this was not a randomized clinical trial of RRSO or RRM, both the case and the control groups underwent a variety of cancer surveillance programs that were not controlled for in this study.
All participants included in this research provided informed consent for research participation under IRB-approved protocols at each of the participating center. The research presented here was also approved by the Committee on Research Involving Human Subjects at the University of Pennsylvania.
Genotype and haplotype data
Genomic DNA samples were extracted from peripheral blood at each center and shipped to the Penn Data Coordinating Center. Samples were genotyped using the SNPlex Genotyping System (Applied Biosystems) or the OpenArray Genotyping System (Applied Biosystems) following the standard protocols. Briefly for SNPlex, 40 ng of DNA extracted from peripheral blood was fragmented using heat. Samples then underwent the Oligo Ligation Assay (OLA) in which allele-specific oligos (ASO), each containing a unique identifying code (ZIPCode), were ligated to locus-specific oligos (LSO) to generate single-stranded products. The products were cleaned using exonuclease to remove all unligated products. Cleaned OLA product then underwent PCR. PCR products were then bound to a plate coated with streptavidin and underwent several washes in which reporters unique to each genotype (ZIPChutes) were hybridized to the products at the ZIPCode. ZIPChutes were then eluted and run on a 3130XL capillary sequencer. Genotypes were read using GeneMapper 4.0. OpenArray is a small-volume genotyping method based on TaqMan chemistry. Genotyping by the OpenArray was done on a 64-SNP array format. Following standard protocol, samples were mixed with the OpenArray Genotyping Mastermix and loaded onto the arrays with an Autoloader. Arrays underwent thermocycling and were imaged with the OpenArray NT Imager. Data were analyzed with the TaqMan Genotyper Software v. 1.0.
We chose SNPs to tag haplotypes and putative functional SNPs in genes that interact with BRCA1 (Supplementary Tables S1 and 2). These include ATM, BRCC36, BRCC45, BRIP1 (BACH1/FANCJ), CTIP, ABRA1 (FAM175A), MERIT40, MRE11A, NBS1 (NBN), PALB2, RAD50, RAD51, RAP80, and TOPBP1. A total of 152 SNPs were identified from the publicly available HapMap data (release 36). htSNPs for each gene were selected using Tagging Wizard in SNPBrowser based on haplotype R2 > 0.95 limited to minor allele frequencies of 5% or greater. Thirty-eight putative functional SNPs were also included if they had been reported in HapMap data or in the literature. SNPs with genotyping failure rates of more than 20% or with low minor allele frequency (MAF: <1%) were excluded from subsequent analyses. In addition, SNPs that showed significant deviations from Hardy–Weinberg equilibrium in unrelated, unaffected Caucasian women were also excluded. After applying these quality control checks, we included a total of 108 SNPs in our analysis.
Analysis was undertaken using the weighted cohort approach of Antoniou and colleagues (13). The weighted approach was implemented to address the issue that study carriers may be ascertained nonrandomly with respect to their disease status. Mutation carriers in our study were ascertained by families that may have included multiple affected individuals who underwent BRCA1/2 mutation testing. Because the presence of disease may influence the likelihood of testing, affected carriers may be overrepresented in our cohort. Under this retrospective study design, standard methods of analysis such as Cox regression can lead to biased estimates of the risk ratios. To address this potential bias, we analyzed the data within a retrospective weighted cohort framework. This approach has been shown to provide risk ratio estimates that are nearly unbiased (13). Five-year interval weights were applied on the basis of published breast cancer incidences for BRCA1 mutation carriers (14).
For breast cancer analyses, women were followed from the initial time of their inclusion in the study to the earliest of the following censoring events: RRSO, RRM, death, or having reached the end of follow-up without a breast cancer or other censoring event. Time to event was computed from age at study inclusion to age at first breast cancer diagnosis or age at censoring. For ovarian cancer analyses, women were followed from the initial time of their inclusion in the study to the earliest of the following censoring events: RRSO, death, or having reached the end of follow-up without an ovarian cancer or other censoring event. For the analyses with ovarian cancer as the primary endpoint, breast cancer diagnoses were ignored. Time to event was computed from age at study inclusion to age at first ovarian cancer diagnosis or age at censoring. Cox proportional hazards models were used to analyze the data after the proportional hazards assumption underlying the Cox model was assessed using log(−log) plots and the Schoenfeld residuals test. To account for intracluster dependence due to multiple individuals from the same family, a robust sandwich variance estimate was used. Analyses were adjusted for ethnicity (Jewish, Hispanic, or non-Hispanic non-Jewish white) and birth cohort (decade of birth). Correction for multiple hypothesis testing was done using the Benjamini–Hochberg method to control the false discovery rate (FDR; ref. 15). All survival analyses were conducted in SAS version 9.1 (SAS Institute Inc.).
To investigate haplotype associations, the expectation–maximization (EM) algorithm (16, 17), which was implemented in R version 2.1.1 subroutine haplo.em (18), was used to estimate haplotype frequencies. To assess the association between haplotypes and survival outcome, we created a user-defined model matrix to estimate haplotype effects. First, haplo.em was used to estimate haplotype frequencies under the null hypothesis of no association (in the pool of all data). This approach enumerated all possible haplotype pairs per subject along with the posterior probabilities of each haplotype pair, conditional on the genotype data. The posterior probabilities were then used to average the rows of the model matrix per subject, and the resulting matrix was used in a Cox regression model. Global tests for association (to test the association between all haplotypes together with disease status) and haplotype-specific tests (to test the association between each haplotype and disease status) were conducted. To implement haplotype-specific survival analysis, we used the haplo.stat program, which is a set of R version 2.1.1 subroutines (18). For sparse data, haplo.stat computes simulation P values for all tests of association. The amount of phase ambiguity was quantified by estimating the percentage of uncertainty in the imputed diplotypes. Most haplotypes had a maximum posterior probability of more than 80%; hence, we felt comfortable proceeding with the haplotype association method outlined earlier rather than assigning the most likely haplotype pair to each subject. As a secondary analysis, we also undertook SNP associations by using Cox proportional hazards models as implemented in SAS version 9.1 (SAS Institute Inc.). All P values are based on 2-sided hypothesis tests.
Among the cohort of 2,825 BRCA1 female mutation carriers, 1,196 (42.3%) had a breast cancer diagnosis and 379 (13.4%) had an ovarian cancer diagnosis. As shown in Table 1, breast and ovarian cancer cases were significantly more likely to be older at the time of study inclusion and members of an older birth cohort than controls. Controls were more likely to have undergone RRSO or RRM than breast cancer cases.
Haplotype analysis was the primary approach used to identify associations. The sample sizes shown later and in the results tables vary due to sampling criteria that differ for breast and ovarian cancer analyses as well as missing genotype data for some SNPs and haplotypes. We considered an association to be important only if we observed a statistically significant overall difference in haplotype frequencies by disease (censoring) status, and there was at least 1 statistically significant haplotype associated with altered breast or ovarian cancer risk. If only one of these events occurred at a particular locus, we did not consider it to show a meaningful association. All inferences are based on 2-sided hypothesis tests with FDR-corrected P values to consider the number of tests conducted here.
An association was observed between breast cancer risk and haplotypes at C19orf62 (encoding MERIT40), in agreement with the prior GWAS finding at this locus (1). However, there was an overlap between the samples studied here and those used in the previous GWAS publication, so this result does not represent an independent replication of that finding. Also in support of the prior GWAS, we report the association of haplotypes at this locus and breast cancer (Table 2). Haplotypes B and E were significantly associated with an increase in breast cancer risk with HR = 1.15 and 1.22, respectively. The common SNP that is changed relative to the reference haplotype is an A>G change in SNP3 (rs3745185). One other haplotype (D) also contained this SNP and had a statistically nonsignificant increased HR effect, but another haplotype containing this SNP (A) did not show any increase in risk. SNP5 (rs8170) was associated with increased breast cancer risk (P = 0.010, Supplementary Table S3), in agreement with the prior GWAS publication (1). Haplotype E containing allele A of this SNP (Table 3) was also associated with increased breast cancer risk.
In addition to this previously reported association, we also identified a series of additional haplotype associations in genes encoding proteins involved in BRCA1 interactions. As shown in Table 2, we observed statistically significant increases in breast cancer risk with ATM haplotype A (HR = 1.18, 95% CI: 1.01–1.37); CTIP haplotype A (HR = 1.33, 95% CI: 1.05–1.67), and 2 rare haplotypes (HR = 1.84, 95% CI: 1.05–3.21); NBS1 haplotype E (HR = 1.66, 95% CI: 1.18–2.33); and TOPBP1 haplotypes B (HR = 1.64, 95% CI: 1.05–2.57) and C (HR = 1.30, 95% CI: 1.03–1.64). Statistically significant decreases in breast cancer risk were identified with BRIP1 haplotype G (HR = 0.52, 95% CI: 0.33–0.84) and RAD50 haplotype F (HR = 0.64, 95% CI: 0.43–0.97). The common change in both TOPBP1 haplotypes was for SNP 1 T>C (rs3732574), but this same SNP was found in another haplotype that was not associated with increased risk. Each of the haplotypes associated with altered risk differed from the reference haplotype by multiple SNPs, so it was not possible to infer which (if any) of the individual SNPs involved in these associations were primarily (or causally) responsible for the observed associations (Supplementary Table S3). No significant associations were observed for BRCC36, ABRA1, MRE11A, PALB2, RAD51, or RAP80.
We also observed associations at BRCC45 in 1 of the 6 haplotype blocks considered (Table 3), namely, haplotype G of block 3 (HR = 1.20, 95% CI: 1.03–1.41). This common haplotype with a 21% frequency differed from the reference haplotype for 5 of the 6 SNPs that comprised the haplotype (SNPs 10, 11, 12, 13, and 15; Table 3). The association of haplotype B in block 4 suggested an association, but there was no overall significant test for differences among haplotypes in this block. None of the other blocks suggested an association of this locus with breast cancer.
We observed a statistically significant association of haplotype C at ABRA1 and ovarian cancer (HR = 1.42, 95% CI: 1.02–1.97; Table 4). This haplotype differed from the reference haplotype by only SNP1 (rs13125836). In SNP analysis (Supplementary Table S4), this C>T change was marginally significantly associated with risk. Carriage of any T allele was associated with HR = 1.44 (95% CI: 1.03–1.97).
Also associated with ovarian cancer risk was a single haplotype in RAP80. Carriage of haplotype G contained multiple differences from the reference haplotype and was associated with a significantly decreased risk of developing ovarian cancer compared with the reference group (HR = 0.69, 95% CI: 0.50–0.97; Table 4).
Finally, we also observed that multiple haplotype blocks in BRCC45 were associated with ovarian cancer risk (Table 5). Two of the 6 blocks (blocks 4 and 5) showed an overall difference in haplotype frequencies by case status (as judged by the FDR-corrected P values) and had significant individual haplotype associations. In block 4, the rare haplotype (frequency: 0.1%) was associated with an elevated risk of ovarian cancer compared with the reference haplotype in this block (HR = 13.58, 95% CI: 8.83–20.89). However, this result is based on a very small number of carriers of these rare haplotypes (N = 26 cases, N = 162 censored individuals). Caution must be used in interpreting this finding, and further validation will be required. In addition, haplotype D in block 5 was associated with an increase in relative risk of developing ovarian cancer (HR = 1.45, 95% CI: 1.01–2.08). This association represents a single difference in SNP27 (rs4666053) compared with the reference haplotype. SNP27 was not associated with ovarian cancer risk in SNP analysis (Supplementary Table S4).
We have identified a number of biologically plausible associations of breast or ovarian cancer with genes that encode proteins that are involved in DNA damage response and interaction with BRCA1. Each of these proteins interacts directly or indirectly with BRCA1 (Supplementary Fig. S1) and therefore may act in concert with a mutated BRCA1 to either confer further increased cancer risk or mitigate the effect of the BRCA1 mutation to lessen breast cancer risk. These results are therefore both biologically plausible and support earlier studies in which individual genes of interest in these pathways have been reported as cancer risk genotypes.
The list of potential modifier gene associations reported here does not include those that were previously reported by our group [e.g., BRCA1/2 interacting proteins in ovarian cancer (5) or others (e.g., RAD51 in BRCA2-associated breast cancer (4)]. However, in combination with our present results, the literature now provides strong evidence that interactors of BRCA1 and BRCA2 are breast and ovarian cancer risk modifiers in women who have inherited a disease-associated BRCA1 or BRCA2 mutation. As shown in Table 6, there is now evidence for genetic variability in ATM, BRCC45, BRIP1, CTIP, MERIT40, NBS1, RAD50, and TOPBP1 in BRCA1-associated breast cancer; BARD1 and RAD51 in BRCA2-associated breast cancer; ATM, BRCC45, and RAP80 in BRCA1-associated ovarian cancer; and ATM, BARD1, BRIP1, MRE11, and RAD51 in BRCA2-associated ovarian cancer. Therefore, there is strong evidence that genetic variation associated with proteins involved in BRCA1-associated multiprotein complexes contributes to breast and ovarian carcinogenesis.
A number of these associations have been reliably repeated in studies using a variety of approaches. We report here using a candidate haplotype association approach that MERIT40 is associated with BRCA1-associated breast cancer risk. In the prior GWAS, haplotype analysis included 6 SNPs (rs4808611, rs3745185, rs8170, rs4808075, rs100241, and rs2363956). Two of these SNPs (rs3745185 and rs8170, denoted SNP3 and SNP5 here, respectively; Supplementary Table S2) were also included in the present analysis. The inferred risk haplotypes from the GWAS were those that included the G allele at rs3745185. This is the same change found in the putative risk haplotype reported here (haplotypes B and E, Table 2). Therefore, both studies have reported that the same alleles (haplotypes) are associated with breast cancer risk in BRCA1 mutation carriers. MERIT40 is a component of the BRCA1 complex that also includes ABRA1 (FAM175A), BARD1, RAP80, BRCC36, and BRCC45 (Supplementary Table S1). In the present study, we also found BRCC45 to be an additional modifier of BRCA1-associated breast cancer risk. However, in a separate study, we found no association of BARD1 Cys557Ser or haplotypes and breast cancer risk (19). This complex is required for recruitment and retention of the BRCA1-BARD1 ubiquitin ligase and the BRCC36 deubiquitination enzyme at sites of DNA damage (20, 21). Thus, genetic variants that modify MERIT40 or BRCC45 function or expression might influence BRCA1-dependent DNA repair and checkpoint activity in BRCA1 mutation carriers sufficiently to increase the risk of developing breast cancer.
We also expanded this result by identifying additional associations with genes encoding proteins involved in other BRCA1-associated complexes (Supplementary Fig. S1 and Supplementary Table S1). We found associations with CTIP, NBS1, and RAD50, suggesting that mutations in genes responsible for regulation of CHK1 activation leading to homologous recombination also influence BRCA1-associated breast cancer risk. We also found associations with BRIP1 and TOPBP1, both of which interact with BRCA1 to regulate S-phase and G2 cell-cycle arrest and influence homologous recombination. Note that BRIP1 also interacts with BRCA2 in regulation of related pathways. We have previously reported that BRIP1 is also associated with ovarian cancer risk in BRCA2 mutation carriers.
In contrast, we found no associations with RAD51 or PALB2. RAD51 interacts with BRCA2. RAD51 genotypes and haplotypes have been associated with altered breast (2–4) and ovarian (5) cancer risk in BRCA2 mutation carriers but not in BRCA1 mutation carriers in most studies (Table 6). These results suggest that modifiers of BRCA1- or BRCA2-associated breast or ovarian cancer risk are more likely to involve those proteins that directly interact with either BRCA1 or BRCA2, rather than those that are indirectly associated with BRCA1 or BRCA2.
Strengths of this study include a relatively large cohort of BRCA1 mutation carriers and a focus on biologically plausible associations involving genes that encode proteins that interact with BRCA1. A limitation of this study is that the sample size remains insufficient to detect some small effects. For example, some analyses involve a relatively small number of ovarian cancer cases, and the effects found for rare haplotypes depend only on a small number of observations. As a result, we have in some cases detected relatively large HR effects that were not statistically significant. We did not have the power to study interactions or higher-order effects among genes or with exposures. Therefore, additional large-scale studies should be undertaken to confirm the results reported here. Finally, new BRCA1/2 interactors are being discovered and the evaluation of all of the genes that encode BRCA1/2 interactors remains incomplete (Table 6). Thus, additional research is required to completely assess these genes as risk modifiers.
Despite the biological plausibility of our results, we cannot make strong inferences about the mechanism of these associations. The SNPs selected here were chosen to optimally reflect haplotypes at these loci, but these haplotypes themselves may not be functionally relevant. We do not have information about the causative alleles that may be in linkage disequilibrium with the haplotype or SNP associations identified here. Significant SNP effects that did not correspond to significant haplotype effects (or vice versa) may in part be explained by limited sample size to detect haplotype effects or that the SNP effect was diluted across multiple (possibly rare) haplotypes. Thus, if the observed effect is greater in magnitude in haplotypes than in SNP analyses, then we may infer that the observed effect represents one or more variants that lie on that haplotype but may or may not have been measured here. If the observed effect is less for haplotypes than SNP results, but the risk estimate across all haplotypes that share that allele are similar or in the same direction, then the allele itself may be of interest. As an example of the limitations of our approach in making causative or mechanistic inferences, we observed statistically significant effects in opposite directions for BRIP1 variants in breast versus ovarian cancer among mutation carriers. Although the usual hypothesis to be made is that the effect of modifiers may be in the same direction for cancers at different sites, the effects of modifier genes may be tissue specific or have other functional consequences that vary by tumor site that the present analyses cannot address. These results again argue in favor of subsequent functional or mechanistic studies to elucidate the fundamental causes of our statistical associations. Finally, despite the relatively large sample sizes available to this study, we do not have sufficient statistical power to attempt analyses of interactions among loci. This research may provide important insights into the joint effects of multiple loci in this pathway in future analyses when even larger sample sets are available.
Taken together, these results provide strong evidence that genes encoding proteins that interact directly with BRCA1 are modifiers of BRCA1-associated breast cancer risk. These associations may be valuable in a number of ways. First, these associations may be used to identify a set of modifiers to assist in risk assessment of BRCA1/2-associated cancers. Second, these associations may provide additional information about the biological relationships of the multiprotein complexes associated with BRCA1 and/or BRCA2 and expand the list of therapeutic targets.
Disclosure of Potential Conflicts of Interest
Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the State of Nebraska or the Nebraska Department of Health and Human Services. No potential conflicts of interest were disclosed.
This publication was supported in part by revenue from Nebraska cigarette taxes awarded to Creighton University by the Nebraska Department of Health and Human Services. Support was also received from NIH grants 5UO1 CA86389 (to H.T. Lynch) and R01-CA083855, R01-CA74415 (to S.L. Neuhausen), R01-CA102776, and R01-CA083855 (to T.R. Rebbeck); the Breast Cancer Research Foundation (K.L. Nathanson); and the NIH grants P30-CA051008 (to C. Isaacs). The HEBCS study is supported by the Helsinki University Central Hospital Research Fund, The Academy of Finland (132473), The Finnish Cancer Society, and The Sigrid Juselius Foundation. The OCGN is supported by Cancer Care Ontario. kConFab is supported by grants from the National Breast Cancer Foundation, the National Health and Medical Research Council (NHMRC), the Queensland Cancer Fund, the Cancer Councils of New South Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Western Australia. EMBRACE is supported by Cancer Research UK grants C1287/A10118 and C1287/A11990. D. Gareth Evans and Fiona Lalloo are supported by an NIHR grant to the Biomedical Research Centre, Manchester. The investigators at The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust are supported by an NIHR grant to the Biomedical Research Centre at The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust. Ros Eeles, E. Bancroft, and Lucia d'Mello are also supported by Cancer Research UK grant C5047/A8385. Clinical Follow Up Study was funded by NHMRC grants 145684, 288704, and 454508.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The MAGIC Consortium includes the following centers and individuals: Baylor-Charles A. Sammons Cancer Center (J.L. Blum, MD, PhD, Estelle Brothers, RN, Gaby Ethington), Baylor College of Medicine (Claire Noll, Sharon Plon, MD, PhD), Beth Israel Deaconess Medical Center (N. Tung, MD), City of Hope National Medical Center (Veronica Lagos, J.N. Weitzel, MD), Creighton University (C.L. Snyder, BA, H.T. Lynch, MD, Patrice Watson, PhD), Dana Farber Cancer Institute (Kathryn Stoeckert, J.E. Garber, MD, MPH), Duke University (Sydnee Crankshaw, Joellen Schildkraut, PhD), NorthShore University HealthSystem Center for Medical Genetics (Christina Selkirk, Scott M. Weissman, W.S. Rubinstein, MD, PhD), Fox Chase Cancer Center (M.B. Daly, MD, PhD, Andrew Godwin, PhD), Queensland Institute of Medical Research (Georgia Chenevix-Trench), Georgetown University (C. Isaacs, MD), Jonsson Comprehensive Cancer Center at the University of California, Los Angeles (Joyce Seldon, P.A. Ganz, MD), Mayo Clinic College of Medicine (Linda Wadum, F.J. Couch, PhD), University of Chicago (Shelly Cummings, O.I. Olopade, MD), University of California, Irvine (S.L. Neuhausen, PhD, Linda Steele), University of Pennsylvania Health System (S.M. Domchek, MD, K.L. Nathanson, MD, T.M. Friebel, MPH., T.R. Rebbeck, PhD), University of Texas, Southwestern (G.E. Tomlinson, MD), University of Vienna (C.F. Singer, MD), and Women's College Hospital (S.A. Narod, MD). The authors 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 for their contributions to this resource and the many families who contribute to kConFab. The HEBCS Study thanks Dr. Kristiina Aittomäki and Tuomas Heikkinen for their help with the patient data and samples. The Ontario Cancer Genetics Network (OCGN) thanks Gord Glendon, Hilmi Ozcelik, Teresa Selander, Nayana Weerasooriya, and members of the Ontario Cancer Genetics Network for their contributions to the study. D.F. Easton is the principal investigator of the Epidemiological Study of BRCA1 and BRCA2 Mutation Carriers (EMBRACE). EMBRACE Collaborating Centers are as follows: Coordinating Centre, Cambridge: S. Peock, Margaret Cook, and Debra Frost. North of Scotland Regional Genetics Service, Aberdeen: Zosia Miedzybrodzka and Helen Gregory. Northern Ireland Regional Genetics Service, Belfast: Patrick Morrison and Lisa Jeffers. West Midlands Regional Clinical Genetics Service, Birmingham: Trevor Cole, Kai-ren Ong, and Jonathan Hoffman. South West Regional Genetics Service, Bristol: A. Donaldson and Margaret James. East Anglian Regional Genetics Service, Cambridge: J. Paterson, Sarah Downing, and Amy Taylor. Medical Genetics Services for Wales, Cardiff: Alexandra Murray, Mark T. Rogers, and Emma McCann. St James's Hospital, Dublin & National Centre for Medical Genetics, Dublin: M.J. Kennedy and David Barton. South East of Scotland Regional Genetics Service, Edinburgh: Mary Porteous, Sarah Drummond. Peninsula Clinical Genetics Service, Exeter: Carole Brewer, Emma Kivuva, Anne Searle, Selina Goodman, and Kathryn Hill. West of Scotland Regional Genetics Service, Glasgow: Rosemarie Davidson and Victoria Murday, Nicola Bradshaw, Lesley Snadden, Mark Longmuir, Catherine Watt, Sarah Gibson, Eshika Haque, Ed Tobias, and Alexis Duncan. South East Thames Regional Genetics Service, Guy's Hospital London: Louise Izatt, Chris Jacobs, Caroline Langman, and Anna Whaite. North West Thames Regional Genetics Service, Harrow: H. Dorkins. Leicestershire Clinical Genetics Service, Leicester: Julian Barwell. Yorkshire Regional Genetics Service, Leeds: Julian Adlard, Carol Chu, and Julie Miller. Merseyside & Cheshire Clinical Genetics Service, Liverpool: Ian Ellis and Catherine Houghton. Manchester Regional Genetics Service, Manchester: D. Gareth Evans, Fiona Lalloo, and Jane Taylor. North East Thames Regional Genetics Service, NE Thames, London: Lucy Side, Alison Male, and Cheryl Berlin. Nottingham Centre for Medical Genetics, Nottingham: Jacqueline Eason and Rebecca Collier. Northern Clinical Genetics Service, Newcastle: Fiona Douglas, Oonagh Claber, and Irene Jobson. Oxford Regional Genetics Service, Oxford: Lisa Walker, Diane McLeod, Dorothy Halliday, Sarah Durell, and 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, Anita Mitra, and Lisa Robertson. North Trent Clinical Genetics Service, Sheffield: Jackie Cook, Oliver Quarrell, and Cathryn Bardsley. South West Thames Regional Genetics Service, London: Shirley Hodgson, Sheila Goff, Glen Brice, Lizzie Winchester, Charlotte Eddy, Vishakha Tripathi, and Virginia Attard. Wessex Clinical Genetics Service, Princess Anne Hospital, Southampton: Diana Eccles, Anneke Lucassen, Gillian Crawford, Donna McBride, and Sarah Smalley.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received March 8, 2011.
- Revision received May 16, 2011.
- Accepted June 8, 2011.
- ©2011 American Association for Cancer Research.