Anthony Renwick

PALB2, which encodes a BRCA2-interacting protein, is a breast cancer susceptibility gene

PALB2 interacts with BRCA2, and biallelic mutations in PALB2 (also known as FANCN), similar to biallelic BRCA2 mutations, cause Fanconi anemia. We identified monoallelic truncating PALB2 mutations in 10/923 individuals with familial breast cancer compared with 0/1,084 controls (P = 0.0004) and show that such mutations confer a 2.3-fold higher risk of breast cancer (95% confidence interval (c.i.) = 1.4–3.9, P = 0.0025). The results show that PALB2 is a breast cancer susceptibility gene and further demonstrate the close relationship of the Fanconi anemia–DNA repair pathway and breast cancer predisposition.

Truncating mutations in the Fanconi anemia J gene BRIP1 are low-penetrance breast cancer susceptibility alleles

We identified constitutional truncating mutations of the BRCA1-interacting helicase BRIP1 in 9/1,212 individuals with breast cancer from BRCA1/BRCA2 mutation–negative families but in only 2/2,081 controls (P = 0.0030), and we estimate that BRIP1 mutations confer a relative risk of breast cancer of 2.0 (95% confidence interval = 1.2–3.2, P = 0.012). Biallelic BRIP1 mutations were recently shown to cause Fanconi anemia complementation group J. Thus, inactivating truncating mutations of BRIP1, similar to those in BRCA2, cause Fanconi anemia in biallelic carriers and confer susceptibility to breast cancer in monoallelic carriers.

Genome-wide association study identifies five new breast cancer susceptibility loci

Breast cancer is the most common cancer in women in developed countries. To identify common breast cancer susceptibility alleles, we conducted a genome-wide association study in which 582,886 SNPs were genotyped in 3,659 cases with a family history of the disease and 4,897 controls. Promising associations were evaluated in a second stage, comprising 12,576 cases and 12,223 controls. We identified five new susceptibility loci, on chromosomes 9, 10 and 11 (P = 4.6 × 10−7 to P = 3.2 × 10−15). We also identified SNPs in the 6q25.1 (rs3757318, P = 2.9 × 10−6), 8q24 (rs1562430, P = 5.8 × 10−7) and LSP1 (rs909116, P = 7.3 × 10−7) regions that showed more significant association with risk than those reported previously. Previously identified breast cancer susceptibility loci were also found to show larger effect sizes in this study of familial breast cancer cases than in previous population-based studies, consistent with polygenic susceptibility to the disease.

ATM mutations that cause ataxia-telangiectasia are breast cancer susceptibility alleles

We screened individuals from 443 familial breast cancer pedigrees and 521 controls for ATM sequence variants and identified 12 mutations in affected individuals and two in controls (P = 0.0047). The results demonstrate that ATM mutations that cause ataxia-telangiectasia in biallelic carriers are breast cancer susceptibility alleles in monoallelic carriers, with an estimated relative risk of 2.37 (95% confidence interval (c.i.) = 1.51–3.78, P = 0.0003). There was no evidence that other classes of ATM variant confer a risk of breast cancer.

A common coding variant in CASP8 is associated with breast cancer risk

The Breast Cancer Association Consortium (BCAC) has been established to conduct combined case-control analyses with augmented statistical power to try to confirm putative genetic associations with breast cancer. We genotyped nine SNPs for which there was some prior evidence of an association with breast cancer: CASP8 D302H (rs1045485), IGFBP3 −202 C → A (rs2854744), SOD2 V16A (rs1799725), TGFB1 L10P (rs1982073), ATM S49C (rs1800054), ADH1B 3′ UTR A → G (rs1042026), CDKN1A S31R (rs1801270), ICAM5 V301I (rs1056538) and NUMA1 A794G (rs3750913). We included data from 9–15 studies, comprising 11,391–18,290 cases and 14,753–22,670 controls. We found evidence of an association with breast cancer for CASP8 D302H (with odds ratios (OR) of 0.89 (95% confidence interval (c.i.): 0.85–0.94) and 0.74 (95% c.i.: 0.62–0.87) for heterozygotes and rare homozygotes, respectively, compared with common homozygotes; Ptrend = 1.1 × 10−7) and weaker evidence for TGFB1 L10P (OR = 1.07 (95% c.i.: 1.02–1.13) and 1.16 (95% c.i.: 1.08–1.25), respectively; Ptrend = 2.8 × 10−5). These results demonstrate that common breast cancer susceptibility alleles with small effects on risk can be identified, given sufficiently powerful studies.NOTE: In the version of this article initially published, there was an error that affected the calculations of the odds ratios, confidence intervals, between-study heterogeneity, trend test and test for association for SNP ICAM5 V301I in Table 1 (ICAM5 V301I); genotype counts in Supplementary Table 2 (ICAM5; ICR_FBCS and Kuopio studies) and minor allele frequencies, trend test and odds ratios for heterozygotes and rare homozygotes in Supplementary Table 3 (ICAM5; ICR_FBCS and Kuopio studies). The errors in Table 1 have been corrected in the PDF version of the article. The errors in supplementary information have been corrected online.

Germline mutations in RAD51D confer susceptibility to ovarian cancer

Recently, RAD51C mutations were identified in families with breast and ovarian cancer. This observation prompted us to investigate the role of RAD51D in cancer susceptibility. We identified eight inactivating RAD51D mutations in unrelated individuals from 911 breast-ovarian cancer families compared with one inactivating mutation identified in 1,060 controls (P = 0.01). The association found here was principally with ovarian cancer, with three mutations identified in the 59 pedigrees with three or more individuals with ovarian cancer (P = 0.0005). The relative risk of ovarian cancer for RAD51D mutation carriers was estimated to be 6.30 (95% CI 2.86–13.85, P = 4.8 × 10−6). By contrast, we estimated the relative risk of breast cancer to be 1.32 (95% CI 0.59–2.96, P = 0.50). These data indicate that RAD51D mutation testing may have clinical utility in individuals with ovarian cancer and their families. Moreover, we show that cells deficient in RAD51D are sensitive to treatment with a PARP inhibitor, suggesting a possible therapeutic approach for cancers arising in RAD51D mutation carriers.

A genome-wide association study of testicular germ cell tumor

We conducted a genome-wide association study for testicular germ cell tumor (TGCT), genotyping 307,666 SNPs in 730 cases and 1,435 controls from the UK and replicating associations in a further 571 cases and 1,806 controls. We found strong evidence for susceptibility loci on chromosome 5 (per allele OR = 1.37 (95% CI = 1.19–1.58), P = 3 × 10−13), chromosome 6 (OR = 1.50 (95% CI = 1.28–1.75), P = 10−13) and chromosome 12 (OR = 2.55 (95% CI = 2.05–3.19), P = 10−31). KITLG, encoding the ligand for the receptor tyrosine kinase KIT, which has previously been implicated in the pathogenesis of TGCT and the biology of germ cells, may explain the association on chromosome 12.

Variants near DMRT1, TERT and ATF7IP are associated with testicular germ cell cancer.

We conducted a genome-wide association study for testicular germ cell tumor, genotyping 298,782 SNPs in 979 affected individuals and 4,947 controls from the UK and replicating associations in a further 664 cases and 3,456 controls. We identified three new susceptibility loci, two of which include genes that are involved in telomere regulation. We identified two independent signals within the TERT-CLPTM1L locus on chromosome 5, which has previously been associated with multiple other cancers (rs4635969, OR = 1.54, P = 1.14 × 10−23; rs2736100, OR = 1.33, P = 7.55 × 10−15). We also identified a locus on chromosome 12 (rs2900333, OR = 1.27, P = 6.16 × 10−10) that contains ATF7IP, a regulator of TERT expression. Finally, we identified a locus on chromosome 9 (rs755383, OR = 1.37, P = 1.12 × 10−23), containing the sex determination gene DMRT1, which has been linked to teratoma susceptibility in mice.

Mosaic PPM1D mutations are associated with predisposition to breast and ovarian cancer

Improved sequencing technologies offer unprecedented opportunities for investigating the role of rare genetic variation in common disease. However, there are considerable challenges with respect to study design, data analysis and replication. Using pooled next-generation sequencing of 507 genes implicated in the repair of DNA in 1,150 samples, an analytical strategy focused on protein-truncating variants (PTVs) and a large-scale sequencing case–control replication experiment in 13,642 individuals, here we show that rare PTVs in the p53-inducible protein phosphatase PPM1D are associated with predisposition to breast cancer and ovarian cancer. PPM1D PTV mutations were present in 25 out of 7,781 cases versus 1 out of 5,861 controls (P = 1.12 × 10−5), including 18 mutations in 6,912 individuals with breast cancer (P = 2.42 × 10−4) and 12 mutations in 1,121 individuals with ovarian cancer (P = 3.10 × 10−9). Notably, all of the identified PPM1D PTVs were mosaic in lymphocyte DNA and clustered within a 370-base-pair region in the final exon of the gene, carboxy-terminal to the phosphatase catalytic domain. Functional studies demonstrate that the mutations result in enhanced suppression of p53 in response to ionizing radiation exposure, suggesting that the mutant alleles encode hyperactive PPM1D isoforms. Thus, although the mutations cause premature protein truncation, they do not result in the simple loss-of-function effect typically associated with this class of variant, but instead probably have a gain-of-function effect. Our results have implications for the detection and management of breast and ovarian cancer risk. More generally, these data provide new insights into the role of rare and of mosaic genetic variants in common conditions, and the use of sequencing in their identification.

Germline RAD51C mutations confer susceptibility to ovarian cancer

To the Editor: In 2010, Meindl and colleagues proposed that germline RAD51C mutations confer high risk for breast and ovarian cancer, comparable to BRCA1 and BRCA2 mutations1,2. However, multiple follow-up studies have provided no supportive evidence that RAD51C mutations predispose to breast cancer3–12. Following the original report, we began investigating the role of other RAD51 paralogs in breast and ovarian cancer susceptibility. This led to our recent discovery that germline RAD51D mutations predispose to ovarian cancer13. We identified truncating RAD51D mutations in 8 of 911 familial breast-ovarian cancer pedigrees and 1 of 1,060 population controls. Our analysis of simultaneous association with both breast and ovarian cancer risk showed that RAD51D mutations confer a sixfold increased risk of ovarian cancer (relative risk (RR) = 6.30, 95% confidence interval (CI) = 2.86–13.85; P = 4.8 °— 10-6) but do not affect or cause only a small increase in breast cancer risk (RR = 1.32, 95% CI = 0.59– 2.96; P = 0.50). This result was supported by our analysis of 737 familial breast cancer pedigrees with no ovarian cancer, in which we detected no RAD51D mutations. These findings prompted us to reevaluate the role of RAD51C in cancer susceptibility. We sequenced the full coding region and intronexon boundaries of RAD51C in 1,132 probands from families with a history of ovarian cancer occurring with or without breast cancer, 272 individuals with ovarian cancer from a hospital-based unselected case series and 1,156 population-based controls (Supplementary Tables 1 and 2 and Supplementary Methods). We identified 12 mutations that result in premature protein truncation in cases compared to 1 such mutation in controls (P = 0.009) (Table 1 and Supplementary Fig. 1). Nine mutations were identified among the 1,132 familial cases, and there was a higher prevalence of mutations in families with multiple ovarian cancer cases: 4 mutations were detected in 311 families with 2 or more cases of ovarian cancer, and 2 mutations were detected in the 67 families with 3 or more cases of ovarian cancer. Three mutations were identified among the 272 individuals with ovarian cancer unselected for family history, suggesting that ~1% of ovarian cancer cases harbor germline RAD51C mutations. We also identified a total of 12 nonsynonymous RAD51C variants (Supplementary Table 3). Four variants were identified in cases and controls; only one, c.790G>A, encoding a p.Gly264Ser amino-acid change, showed any evidence of association with cancer (P = 0.02), consistent with other studies2,9,12. Of note, this variant is predicted to be benign by in silico analyses and has limited impact on RAD51C function2. The remaining eight nonsynonymous variants were each identified in a single individual; there was no significant difference in the overall frequency (P = 0.36), position or predicted functional effects of these variants between cases and controls (Supplementary Table 3). These data exemplify the inherent complexities of evaluating the clinical consequences of missense variants (outside simple Mendelian disorders) and underscore why non-truncating and truncating variants should be considered separately. Analyzing controls for specific rare variants detected in cases and concluding that their absence in controls is evidence of pathogenicity can result in over-interpretation of the data. Such findings confirm that the specific variant is rare but can seldom provide conclusive evidence of disease association. Full sequencing of the gene in both cases and controls is a more appropriate analysis, as it allows the spectrum of variants in cases and controls to be directly compared. Functional and conservation data can be useful in the evaluation of variants, but in vitro functional effects do not necessarily imply that the variant has clinical sequelae. Moreover, as we and others have shown (for example, in studies of the breast cancer susceptibility genes BRIP1 and ATM), such an assumption can result in incorrect attribution of pathogenicity14,15. Better information is provided when mutational and functional analyses are equally ascertained in both cases and controls. To estimate the risk associated with RAD51C mutations, we undertook modified segregation analysis, in which we simultaneously modeled the risks of ovarian and breast cancer and incorporated control data and information from the full pedigrees of mutation-positive and mutation-negative families (Supplementary Methods). The relative risk of ovarian cancer for RAD51C mutation carriers was estimated to be 5.88 (95% CI = 2.91–11.88; P = 7.65 × 10-7), which constitutes a >9% cumulative risk by age 80. In contrast, there was no evidence of an association with breast cancer (RR = 0.91, 95% CI = 0.45–1.86; P = 0.8). Thus, the cancer risk estimates for RAD51C mutations were similar to those estimated for RAD51D mutations13. These data are fully consistent with the results presented by Meindl et al. and provide a likely explanation for why Meindl et al. identified RAD51C mutations only in breast cancer cases that had relatives with ovarian cancer and not in 620 familial breast cancer pedigrees without ovarian cancer. As RAD51C