Received: 6 June 2019 | Accepted: 5 July 2019 DOI: 10.1002/mgg3.863 ORIGINAL ARTICLE Biallelic germline BRCA1 mutations in a patient with early onset breast cancer, mild Fanconi anemia‐like phenotype, and no chromosome fragility Katharina Keupp1 | Stephanie Hampp2 | Annette Hübbel1 | Monika Maringa1 | Sarah Kostezka 2 | Kerstin Rhiem 1 | Anke Waha 1 | Barbara Wappenschmidt 1 | Roser Pujol 3,4,5 | Jordi Surrallés 3,4,5 | Rita K. Schmutzler 1 | Lisa Wiesmüller 2 | Eric Hahnen1 1 Center for Familial Breast and Ovarian Cancer, Center for Integrated Oncology Abstract (CIO), University of Cologne, Faculty of Background: Biallelic BRCA1 mutations are regarded either embryonically lethal Medicine and University Hospital Cologne, or to cause Fanconi anemia (FA), a genomic instability syndrome characterized Cologne, Germany, University Hospital of Cologne, Cologne, Germany by bone marrow failure, developmental abnormalities, and cancer predisposition. 2 Department of Obstetrics and Gynecology, We report biallelic BRCA1 mutations c.181T > G (p.Cys61Gly) and c.5096G > A Ulm University, Ulm, Germany (p.Arg1699Gln) in a woman with breast cancer diagnosed at the age of 30 years. The 3 Department of Genetics and Biomedical common European founder mutation p.Cys61Gly confers high cancer risk, whereas Research Institute, Hospital de Sant Pau, Barcelona, Spain the deleterious p.Arg1699Gln is hypomorphic and was suggested to confer interme- 4 Department of Genetics and Microbiology, diate cancer risk. Universitat Autònoma de Barcelona, Methods and Results: Aside from significant toxicity from chemotherapy, the Barcelona, Spain patient showed mild FA‐like features (e.g., short stature, microcephaly, skin hy- 5 Center for Biomedical Network Research perpigmentation). Chromosome fragility, a hallmark of FA patient cells, was not on Rare Diseases, Barcelona, Spain present in patient‐derived peripheral blood lymphocytes. We demonstrated that the Correspondence p.Arg1699Gln mutation impairs DNA double‐strand break repair, elevates RAD51 Eric Hahnen, University Hospital Cologne, Kerpener Straße 34, 50931 Cologne, Germany foci levels at baseline, and compromises BRCA1 protein function in protecting from Email: eric.hahnen@uk-koeln.de replication stress. Although the p.Arg1699Gln mutation compromises BRCA1 func- Lisa Wiesmüller, Department of Obstetrics tion, the residual activity of the p.Arg1699Gln allele likely prevents from chromo- and Gynecology, Ulm University, some fragility and a more severe FA phenotype. Prittwitzstrasse 43, 89075 Ulm, Germany. Conclusion: Our data expand the clinical spectrum associated with biallelic BRCA1 Email: lisa.wiesmueller@uni-ulm.de mutations, ranging from embryonic lethality to a mild FA‐like phenotype and no Funding information chromosome fragility. Ministerio de Economía y Competitividad, Grant/Award Number: SAF2015-64152-R; KEYWORDS Deutsche Krebshilfe, Grant/Award Number: biallelic BRCA1, early onset breast cancer, Fanconi anemia, p.Arg1699Gln 110837, 70111850 and 70112504 Katharina Keupp and Stephanie Hampp contributed equally to this work. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2019 The Authors. Molecular Genetics & Genomic Medicine published by Wiley Periodicals, Inc. Mol Genet Genomic Med. 2019;7:e863.  wileyonlinelibrary.com/journal/mgg3 | 1 of 9 https://doi.org/10.1002/mgg3.863 2 of 9 |    KEUPP et al. 1 | IN TRO D U C T ION phenotype suggesting Fanconi anemia (FA). Medical re- cords revealed short stature, microcephaly, developmental Monoallelic germline mutations in the BRCA1 and BRCA2 delay, and significant toxicity from chemotherapy. The pa- confer high life‐time risks for breast cancer (BC) and ovarian tient carried a protein‐truncating variant (p.Asp821Ilefs*25) cancer (OC) (Kuchenbaecker et al., 2017) and were found in and a pathogenic missense variant (p.Val1736Ala) in trans approximately 24% of index patients who met the inclusion (Domchek et al., 2013). In 2015, Sawyer et al. reported bi- criteria of the German Consortium for Hereditary Breast and allelic BRCA1 mutations (p.Ser198Argfs*35 and p.Arg- Ovarian Cancer (GC‐HBOC) for germline testing (Kast et al., 1699Trp) in a woman with multiple congenital anomalies 2016). BRCA1/BRCA2‐related BC and OC susceptibility is consistent with a FA‐like disorder (phenotype subsequently inherited in an autosomal dominant manner with incomplete referred to as FANCS, MIM# 617883) and BC at the age of penetrance (Kuchenbaecker et al., 2017). 23 years (Sawyer et al., 2015). Freire et al. described a 2.5‐ Biallelic (i.e., homozygous/compound heterozygous) year‐old girl of consanguineous offspring with severe short mutations in the BRCA2 cause Fanconi anemia (FANCD1, stature, microcephaly, neurodevelopmental delay, congenital MIM# 605724), a recessive, congenital genomic instability heart disease and dysmorphic features, and the girl was a ho- syndrome generally characterized by bone marrow failure, mozygous carrier of a BRCA1 nonsense mutation in exon 11 developmental abnormalities in various organ systems and a (p.Cys903*) (Freire et al., 2018). In 2018, Seo et al. reported high predisposition for hematological or solid tumors. Despite multiple congenital anomalies and severe chromosomal fra- a similar mutation prevalence of the BRCA2 and BRCA1 in gility in children of consanguineous offspring, all homozy- the general population (Maxwell, Domchek, Nathanson, & gously carrying BRCA1 nonsense mutations affecting exon Robson, 2016), biallelic mutations affecting the BRCA1 are 11 (p.Trp372*, p.Leu431*)(Seo et al., 2018). Of note, mi- rarely described and have been suggested as lethal during em- tomycin C‐ or diepoxybutane (DEB)‐induced chromosome bryonic development (Goldgar et al., 2004), consistent with fragility assays performed by Seo et al. are routinely used to the embryonic lethality observed in Brca1 null mice (Evers & confirm clinical FA diagnosis (Castella et al., 2011). Jonkers, 2006). In 2013, however, Domchek et al. described In this study, we report a female index patient with two a 28‐year‐old woman with OC and an otherwise complex deleterious BRCA1 missense alterations in a compound F I G U R E 1 Pedigree and germline mutation status. (a) Cancer family history of the index patient (III‐1, marked with an arrow), including age at first diagnosis. BC = breast cancer; PC = prostate cancer; y = years. The index patient's mother (II‐2) had no personal history of cancer at the age of 60 years. The maternal grandfather of the index patient was diagnosed with prostate cancer at the age of 80 years. The index patient's father (II‐1) had no personal history of cancer at the age of 59 years. The paternal grandmother of the index patient was diagnosed with breast cancer at the age of 42 years and died at the age of 87 years. No other cancer cases were reported in the family. (b) Sanger sequencing electropherograms show BRCA1 genotypes (BRCA1 reference transcript NM_007294.3) of the index patient III‐1 and both parents (II‐1, II‐2). Red arrows mark the presence of a single nucleotide substitution KEUPP et al.    | 3 of 9 F I G U R E 2 (a) Analysis of DSB‐induced HR activities in individuals with wild‐type, mono‐ or biallelic mutations in the BRCA1. To measure the repair of DSBs by HR, we used the EGFP‐based reporter substrate HR‐EGFP/5'EGFP consisting of mutated EGFP genes with an I‐SceI endonuclease recognition sequence replacing 4 bp (HR‐EGFP) and a 3' truncation (5'EGFP), respectively. The DSB repair substrate for the determination of HR frequencies is schematically drawn on top (Akyuz et al., 2002). I‐SceI recognition sequence, red triangle; cross, truncating mutation; white bars, mutated EGFP genes; dark green bars, reconstituted EGFP; blue bar, spacer sequence; gray bar with kinked arrow, transcriptional promoter; scissors, I‐SceI endonuclease. HR measurements were performed 24 hr following transfection of PBLs (left diagram) or LCLs (right diagram) from individuals with BRCA1 (p.Arg1699Gln) or BRCA1 (p.Cys61Gly/p.Arg1699Gln) with HR‐EGFP/5´EGFP substrate plus I‐SceI expression plasmid for substrate cleavage. Measurements for PBLs with wild‐type BRCA1 were obtained from one (left diagram) and two (right diagram) independent control individuals. Percentages of EGFP‐positive live cells were normalized to the individually determined transfection efficiencies for HR frequency calculations. Mean values of wild‐type controls were set to 100% (absolute mean frequency: 0.2%). Mean values and standard errors of the mean (SEM) from 3 to 5 measurements are shown. (b) Immunofluorescence microscopic analysis of DNA damage and RAD51 nucleofilament formation. Nuclear foci of the DNA damage marker γH2AX (left diagram) and of RAD51 (right diagram), indicating nucleofilament formation, were analyzed in PBLs from individuals with wild‐type and mutated BRCA1 (p.Arg1699Gln, p.Cys61Gly/p. Arg1699Gln) 1 and 6 hr after treatment with 2 Gy ionizing radiation (IR) and in untreated samples. Immunolabeled foci were scored by automated quantification from 97 to 115 nuclei per sample. Mean foci/cell and SEM are graphically shown on the left side. Statistically significant differences were determined using nonparametric Mann–Whitney test for unpaired samples with the software GraphPad Prism version 7.03. *p < .05; **p < .01; ***p < .001; ****p < .0001. Representative images of green γH2AX and red RAD51 foci in DAPI‐stained nuclei (blue) 6 hr post‐IR are shown on the right side. The size bar indicates 5 µm. (c) Analysis of replication fork protection following stalling. Single DNA fiber analysis was performed as schematically outlined in the sketch on top. Representative fibers ∓ HU are shown for the patient carrying biallelic BRCA1 mutations (p.Cys61Gly/p.Arg1699Gln). The size bar indicates 5 µm. CldU track length distributions of CldU‐ and IdU‐positive DNA fibers from PBLs derived from individuals with wild‐type and mutated BRCA1 (p.Arg1699Gln, p.Cys61Gly/p.Arg1699Gln) were determined in the presence of HU, indicating replication stalling (right) or absence of HU for unperturbed replication. Mean values were calculated from fiber track lengths of 106–172 single fibers in each sample. Statistically significant differences between PBLs from individuals with wild‐type and mutated BRCA1 (p.Cys61Gly/p.Arg1699Gln) were calculated using Dunn's test. ***p < .001; ****p < .0001 4 of 9 |    KEUPP et al. heterozygous state, namely, c.181T > G (p.Cys61Gly) and ethics committee of Ulm University), from which aliquots c.5096G > A (p.Arg1699Gln) (Bouwman et al., 2013; Lindor of the same blood draw had previously served as references et al., 2012). The c.181T > G (p.Cys61Gly) alteration is one in a recent case–control study (Deniz, 2017). LCLs from of the most common pathogenic BRCA1 founder mutations three reference individuals age 49–56 were used as con- in Europe (Rebbeck et al., 2018) and a well‐established high‐ trols during measurements of microhomology‐mediated risk variant (Lindor et al., 2012), whereas the deleterious end joining (Figure S1a). Thawed PBLs were cultivated p.Arg1699Gln mutation is hypomorphic and was suggested in PB‐MAXTM Karyotyping Medium (Gibco/Invitrogen, to confer intermediate cancer risk (Bouwman et al., 2013; Carlsbad, CA, USA) including 2% phytohemagglutinin Moghadasi et al., 2018; Spurdle et al., 2012). (PAA, Pasching, Germany) for 72 hr. LCLs were culti- vated as previously described (Gatz et al., 2016) and used for HR measurements at passage 2–6. 2 | M AT E R IA L S A N D ME T HODS 2.1 | Testing for germline mutations in 2.3 | EGFP‐based reporter for DSB repair cancer predisposition genes measurements The index patient III‐1 and her parents (Figure 1a) were coun- To analyze DSB repair via HR or microhomology‐mediated seled at the Center for Hereditary Breast and Ovarian Cancer, end joining, we used our enhanced green fluorescent protein University Hospital of Cologne, Germany. Physicians quali- (EGFP)‐based test system, as described earlier (Akyuz et al., fied in genetic counseling recorded personal and family can- 2002; Deniz, 2017; Obermeier et al., 2016). Briefly, the ex- cer history, information regarding age at first diagnosis, tumor pression plasmid for the endonuclease I‐SceI (pCMV‐I‐SceI) receptor status, and treatment of the index patient. Written in- together with HR or microhomology‐mediated end joining formed consent was obtained from all individuals, and ethical substrate (Figure 2a and Supp. Figure S1a) and filler plas- approval was granted by the ethics committee of the University mid (pBlueScriptII KS; Stratagene, Heidelberg, Germany) of Cologne (07‐048). Genomic DNA was isolated from venous or wild‐type EGFP reporter plasmid (for determination of blood samples. Next generation sequencing technology was transfection efficiencies) were transfected into PBLs or applied using an Illumina MiSeq sequencing device (Illumina, LCLs using amaxa reagents (Lonza, Cologne, Germany). San Diego, USA) and the customized diagnostic TruRisk® After 24 hr of cultivation, reconstitution of wild‐type EGFP multigene panel for target enrichment (Agilent, Santa Clara, was determined by FACS analysis of the fraction of green USA) covering BRCA1, BRCA2, and additional breast/ovarian fluorescent cells among living cells identified by gating in cancer predisposition genes (ATM, BRIP1, CDH1, CHEK2, the FL1/FL2 and side scatter/forward scatter dot plots fol- PALB2, RAD51C, RAD51D, TP53). Variant classification was lowing laser excitation at 488 nm (FACS Calibur FACScan, performed in accordance with the regulations of the interna- Becton Dickinson, Heidelberg, Germany). Each quantifica- tional ENIGMA consortium (Evidence‐based Network for the tion of green fluorescent cells monitoring DSB repair was Interpretation of Germline Mutant Alleles; https​://enigm​acons​ normalized using the individually determined transfection ortium.org; version 1.1:26 March 2015). efficiency (20%–60%) to calculate the DSB repair frequency. 2.2 | Blood‐derived cell cultures for 2.4 | Nuclear foci analysis by functional analyses immunofluorescence microscopy The DEB‐induced chromosome fragility assay was per- For in situ analysis of DSB repair by HR, we performed im- formed as described previously (Castella et al., 2011). munofluorescence microscopic analysis of the DNA damage PBLs were partially isolated from heparinized blood sam- marker γH2AX and the recombinase RAD51 in PBLs after ples by Ficoll‐Paque‐PLUS, (LSM 1077 Lymphocyte, GE exposure to 2 Gy of ionizing radiation (IR) as previously de- Healthcare, Germany) gradient centrifugation followed by scribed (Gatz et al., 2016; Obermeier et al., 2016). Briefly, several washing steps with PBS to remove thrombocytes, PBLs were harvested by cytospinning at the indicated time and the samples were then gently frozen as described points postirradiation, fixed with 3.7% formaldehyde fol- (Deniz, 2017). The remaining fractions of the blood sam- lowed by permeabilization with 0.5% TritionX‐100. Primary ples from the index patient (III‐1, Figure 1a) and her antibodies were directed against γH2AX (Ser139, clone mother (II‐2, Figure 1a) were subjected to the establish- JBW301, Millipore, Billerica, MA, USA) and RAD51 (H‐92, ment of LCLs by EBV‐immortalization of peripheral blood Santa Cruz Biotechnology, Heidelberg, Germany), and the mononuclear cells. Aliquots from two reference blood secondary antibodies were AlexaFluor488 and 555‐labeled samples were obtained from healthy female donors aged 28 (Invitrogen, Karlsruhe, Germany). Immunostained cells and 26 (approval #157/2010 and amendment 2016 by the were mounted with VectaShield mounting media containing KEUPP et al.    | 5 of 9 DAPI (Vector laboratories, Burlingame, CA, USA). Focal 3 | RESULTS accumulations of 53BP1 and RAD51 in ≥97 DAPI‐stained nuclei from two independent slides were analyzed using a 3.1 | Clinical findings Keyence BZ‐9000 microscope equipped with Keyence BZ‐II The female index patient presented at the age of 30 years Analyzer software (Keyence, Neu‐Isenburg, Germany). with an invasive‐ductal carcinoma of her left breast (G3; Ki67 30%). The 35‐mm tumor expressed estrogen receptors 2.5 | Cellular sensitivity to PARP but tested negative for progesterone receptor and epidermal inhibition and carboplatin growth factor receptor 2 expression/overexpression. At the time of the BC diagnosis, the patient's blood count was incon- To assess the response to the PARP inhibitor olaparib spicuous. During her childhood, she obtained diagnoses of (Biozol, Eching, Germany), PBLs were treated with vari- hearing loss (right side), celiac disease and congenital hip dis- ous olaparib concentrations (0.5–512 μmol/L) or carboplatin location (left side) and had a hip operation at 11 years. A dys- (0.125–2048 µmol/L) for 24 hr and subsequently re‐cultivated morphology exam revealed a short height (150 cm, <5%ile in fresh medium for 24 hr. Cell viabilities after the different for adult height) (de Onis et al., 2007), microcephaly (52 cm, treatments were determined in duplicate using the 3‐(4,5‐di- <5%ile for adult head circumference), triangular face with methylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide (MTT) low placed ears, and skin hyperpigmentation (several café au assay as described (Deniz, 2017). lait macules). No further FA‐like features were observed (no limb defects; no ophthalmic, endocrine, genitourinary tract, 2.6 | DNA fiber assay for nascent spine or neck anomalies; no congenital heart defect; no gas- DNA synthesis trointestinal malformations; no central nervous system anom- Cells were labeled with 20 µmol/L 5‐chloro‐2‐deoxyu- alities). Her parent's heights were unconspicuous (mother: ridine (CldU, Sigma‐Aldrich, Steinheim, Germany) for 179 cm, father: 185 cm). Following BC diagnosis, she re- 30 min and centrifuged for 2 min to discard CldU‐contain- ceived neoadjuvant chemotherapy starting with four cycles ing medium before treating the cells with 5 mmol/L hy- epirubicin/cyclophosphamide (90/600 mg/m2 body surface). droxyurea (HU, Sigma‐Aldrich, Steinheim, Germany) for Doses were reduced during the fourth cycle due to hemato- 30 min or mock‐treated. To discard HU‐containing me- toxicity (75/400 mg/m2 body surface). She subsequently was dium, the cells were centrifuged for 2 min, then, the cells treated with paclitaxel (80 mg/m2 body surface), which was were incubated with 200 µmol/L 5‐Iodo‐2 deoxyuridine combined with carboplatin (AUC 5) starting from the sec- (IdU, Sigma‐Aldrich, Steinheim, Germany) for 30 min. For ond cycle. Due to recurring thrombopenia, neutropenia, and cell harvesting, the cells were centrifuged for 2 min and anemia, she received reduced doses of paclitaxel (70 mg/m2 washed once with 1 × DPBS before the cell pellet was re‐ body surface) and carboplatin (AUC 1.5). As a result of the suspended in 1 × DPBS. Cell‐spotting, cell lysis, fixation prolonged thrombopenia, neutropenia, and anemia, chemo- and immunofluorescence staining were performed as pre- therapy was discontinued after four cycles, followed by mas- viously described (Hampp et al., 2016). Briefly, the cells tectomy, and radiotherapy and antihormone therapy. were spotted on a slide, and the DNA spreading was per- formed via gravity before fixation in methanol/acetic acid. 3.2 | Cancer family history and germline For denaturation/depurination, glass slides were incubated mutation status in HCl before immunofluorescence staining. The pedigree of the index patient III‐1 is shown in Figure 1a. The index patient met the inclusion criteria of the 2.7 | Statistical analyses German Consortium for Hereditary Breast and Ovarian Statistically significant differences between mean values Cancer (GC‐HBOC) for germline testing, which was per- of DSB repair frequencies or nuclear foci/cell numbers formed by next generation sequencing (NGS) using the were calculated using the Mann–Whitney test for unpaired TruRisk® gene panel. Two heterozygous missense mu- samples with the software GraphPad Prism version 7.03. tations, c.5096G > A (p.Arg1699Gln) and c.181T > G GraphPadPrism 7.03 was also used for calculating IC50 val- (p.Cys61Gly), in the BRCA1 were detected, which both ues and to test for significant differences between olapa- were confirmed by Sanger sequencing (Figure 1b). Genetic rib‐response curves using the extra sum‐of‐squares F‐test. testing of the patient's parents via Sanger sequencing re- Kruskal–Wallis based Dunn's multiple comparison test was vealed that the c.5096G > A (p.Arg1699Gln) mutation used for the calculation of statistically significant differences was inherited from the patient's mother (II‐2), whereas the in nascent DNA synthesis. All p‐values are two‐sided with c.181T > G (p.Cys61Gly) mutation was inherited from the p < .05 considered statistically significant. patient's father (II‐1). 6 of 9 |    KEUPP et al. fivefold elevated microhomology‐mediated end joining com- 3.3 | Diepoxybutane (DEB)‐induced pared with three wild‐type LCLs (p = .0182–.0091; Figure chromosome fragility assay S1). These results indicated a defect in the HR machinery Chromosome fragility is usually reported as “percentage of of the mother and the index patient, in particular, causing a aberrant cells” and “breaks/cell.” The number of “breaks/ derepression of error‐prone DSB repair. Notably, error‐prone cell” is more than 10 times higher in the FA population than DSB repair, such as microhomology‐mediated end joining, in the non‐FA population, while the “percentage of aber- associates with BC/OC risk as previously shown for cells rant cells” is increased 60 times in FA patients (Castella et from BRCA1, BRCA2, and PALB2 mutation carriers as well al., 2011). To determine whether a FA phenotype may be as in a case–control format independently of the individual present on the cellular level using peripheral blood lym- risk genotype (Keimling et al., 2012, 2011; Obermeier et al., phocytes (PBLs) derived from the index patient (III‐1) and 2016). her unaffected mother (II‐2), we used an established DEB‐ induced chromosome fragility assay described previously (Castella et al., 2011). Both the index patient (6% of cells 3.5 | Nuclear foci analyses (γH2AX and with breaks, 0.12 breaks/cell) and her unaffected mother RAD51) by immunofluorescence microscopy (10% of cells with breaks, 0.17 breaks/cell) showed DEB‐ To elucidate the molecular defect observed, we performed induced chromosome fragility within the control range (4% immunofluorescence microscopic analysis of focal nuclear of cells with breaks, range 0%–22%; 0.05 breaks/cell, range structures indicative of the accumulation and/or removal of 0–0.5). These inconspicuous results prompted us to inves- DSBs using antibodies against γH2AX and of the assem- tigate the hypomorphic p.Arg1699Gln allele, which was bly of the HR machinery using antibodies against RAD51. suggested to confer intermediate cancer risk (Bouwman et Prior to treatment, we observed 1.4‐fold elevated basal al., 2013; Moghadasi et al., 2018; Spurdle et al., 2012), γH2AX foci levels in patient PBLs compared to PBLs from in more detail. The BRCA1 BRCT domain p.Arg1699Gln the mother and 1.8‐fold compared to the wild‐type control mutation in the index patient III‐1 and her mother II‐2 (Figure 2b). Moreover, basal RAD51 foci numbers were (Figure 1a) was described to weaken BACH1 phospho‐ threefold higher in PBLs from the patient and mother than peptide interactions (Clapperton et al., 2004), which are those in the control (Figure 2b). Because γH2AX foci were required for efficient error‐free homologous recombination reported to mark not only DSBs but also other DNA lesions (HR) repair (Xie et al., 2010). (Gagou, Zuazua‐Villar, & Meuth, 2010), and in particular, stalled replication forks triggering recombinative lesion by- pass, our observation was consistent with pronounced repli- 3.4 | EGFP‐based reporter for DSB repair cation stress marked by both γH2AX and RAD51 in patient measurements cells. One hour after DSB induction by treatment of PBLs Given the close association between HR deficiency and the with 2 Gy ionizing radiation, γH2AX foci accumulated two‐ pathogenicity of BRCA1 mutations (Bouwman et al., 2013), to threefold in all cell types (p < .001), thereby maintaining we applied an EGFP‐based DNA double‐strand break (DSB) a relative twofold increment of γH2AX foci in patient cells repair assay on fresh PBLs derived from the index patient, (Figure 2b). At the same time point, RAD51 foci numbers her unaffected mother and two female healthy donors func- still showed a similar pattern as before irradiation, because tionally representing wild‐type controls (Deniz, 2017; Gatz canonical nonhomologous end joining dominates during et al., 2016; Keimling et al., 2012) (Figure 2a). HR analysis this early period of DSB repair (Deniz, 2017) (Figure 2b). indicated threefold reduced HR frequencies in PBLs from However, 6 hr after irradiation, RAD51 foci were 3.4‐fold the patient as well as from the mother compared to the wild‐ and 1.9‐fold elevated in cells from the control and mother, type reference, although not reaching statistical significance respectively (p < .001), when compared with 1 hr postirra- (p = .1000), most likely due to the limited number of meas- diation (Figure 2b). No increase was observed in the patient urements. This observation was supported in a second experi- cells suggesting a problem in RAD51 filament assembly ment using early passage lymphoblastoid cell lines (LCLs) during the late phase of DSB repair involving end process- derived from the same blood draws of the patient and her ing‐mediated pathways such as HR (Deniz, 2017). Since mother. Compared to the mean values for wild‐type BRCA1 γH2AX‐labeled DSB scores were similar in the different status (obtained with two wild‐type PBL controls), a statisti- PBL cultures at this late time point (Figure 2b), DSB repair cally significant twofold decrease in HR was scored for the in patient cells may have shifted towards mutagenic path- LCLs from the mother carrying the BRCA1 p.Arg1699Gln ways. Microhomology‐mediated end joining was reported mutation and sixfold for the LCLs from the index patient to become unleashed in HR defective cells (Keimling et al., carrying both mutations, p.Cys61Gly and p.Arg1699Gln 2011) as demonstrated for the patient cells investigated here (p = .0357). Concomitantly, both LCLs showed four‐ to (Supp. Figure S1). Altogether, these data suggested that the KEUPP et al.    | 7 of 9 acquisition of the BRCA1 variant causing the p.Cys61Gly p.Cys61Gly and the p.Arg1699Gln mutations are compro- mutation on top of p.Arg1699Gln exchange severely al- mised in protecting nascent DNA from nucleolytic degrada- tered HR in the patient, resulting in accumulation of basal tion during replication stress, whereas defective RAD51 foci DNA damage and failure to assemble functional RAD51 formation postirradiation and olaparib hypersensitivity was nucleofilaments. observed with patient cells only. 3.6 | Cellular sensitivity to Poly(ADP‐ ribose) polymerase (PARP) inhibition and 4 | DISCUSSION carboplatin We show the co‐occurrence of two deleterious BRCA1 al- PARP inhibition causes synthetic lethality in combina- terations (p.Cys61Gly, p.Arg1699Gln) in trans with no cy- tion with severe HR deficiency such as of tumor cells from togenetic features of FA observed, indicating that at least one BRCA1 or BRCA2 mutation carriers upon inactivation of mutation is hypomorphic. The inconspicuous cytogenetic and the wild‐type allele (Bryant et al., 2005). Having identified comparatively mild FA‐like clinical phenotype presented here increasing defectiveness in HR in the biallelically versus is most likely due to incomplete impairment of BRCA1 func- monoallelically BRCA1 mutated and wild‐type cells, we tion by the p.Arg1699Gln risk allele, which was described to examined cell survival after exposure to increasing concen- intermediately increase BC/OC risks by the age of 70 years trations of the PARP inhibitor olaparib. Patient cells were to 20% and 6%, respectively (Moghadasi et al., 2018; Spurdle highly sensitive to olaparib treatment as compared to the con- et al., 2012). These calculated risks were considerably lower trol and to cells from the mother, for which we determined than those described for pathogenic mutations overall, which an IC50 value showing only a trend towards a significant were associated with BC/OC risks by the age of 70 years of decrease compared with the control (p < .1) (Supp. Table approximately 60% and 40%, respectively (Kuchenbaecker S1). Conversely, after treatment with the DNA cross‐link- et al., 2017). Using PBLs/LCLs derived from the patient's ing agent carboplatin, we found no statistically significant unaffected mother, we demonstrated that the p.Arg1699Gln difference between the IC50 values for the cells from the pa- mutation impairs DSB repair causing a derepression of error‐ tient, mother or control (Table S1); although in parallel, we prone DSB repair (Figure 2a and Figure S1), elevates RAD51 measured sensitivity changes in PBLs from a cohort of high foci levels at baseline (Figure 2b), and compromises BRCA1 BC/OC risk individuals (Deniz, 2017). These discrepant re- protein function in protecting from replication stress (Figure sults could be because of DNA interstrand cross‐links, as 2c). Although the p.Arg1699Gln mutation impairs DNA dou- discovered recently (Semlow, Zhang, Budzowska, Drohat, ble‐strand break repair and compromises BRCA1 function in & Walter, 2016), and can be repaired by both HR‐dependent protection from replication stress, the residual function of and ‐independent pathways. the p.Arg1699Gln allele likely prevents from chromosome fragility and a more severe FA phenotype. In addition, we cannot exclude a residual function of the BRCA1 protein car- 3.7 | DNA fiber assay for nascent rying the p.Cys61Gly mutation. DNA synthesis In concert with published findings (Domchek et al., 2013; Schlacher, Wu, & Jasin (2012) demonstrated that BRCA2 Freire et al., 2018; Sawyer et al., 2015; Seo et al., 2018), exerts functions in stabilizing DNA replication forks that can our data suggest that the phenotypic variability of biallelic be separated from its role in promoting HR. Livingston and BRCA1 mutations ranges from embryonic lethality to short co‐workers (Pathania et al., 2014) observed that BRCA1 is body height, microcephaly, early onset BC and toxicity from haploinsufficient for the suppression of replication stress but chemotherapy, whereas residual BRCA1 protein function de- not for RAD51 foci formation or PARP inhibitor sensitivity. termines the disease severity. For FA patients homozygously To investigate the response of patient cells in the presence carrying nonsense mutations in exon 11 of the BRCA1 (Freire of replication stress, we analyzed the fate of nascent DNA et al., 2018; Seo et al., 2018), it has been demonstrated that replication tracks after treatment with hydroxyurea (HU), a naturally occurring alternative splicing isoform can enable which causes the depletion of the deoxyribonucleotide pool survival, albeit with severe consequences (Seo et al., 2018). (Figure 2c). In the absence of HU, replication track lengths The index patient described by Sawyer et al. carried a pro- were similar for PBLs from the mother or patient compared tein‐truncating BRCA1 mutation, p.Ser198Argfs*35, along with the control (Figure 2c). However, tracks of PBLs from with the p.Arg1699Trp high‐risk variant which was shown the mother and the patient were 30% and 40%, respectively, to elevate the BC risk to 58% by the age of 70 years (Spurdle shorter after HU treatment compared to the control after HU et al., 2012). Analyses of patient‐derived cells revealed de- treatment. Taken together, cells from both the mother carry- fective RAD51 foci formation and olaparib hypersensitivity, ing the p.Arg1699Gln mutation and the patient with both the similar to the patient characterized in our study. In contrast, 8 of 9 |    KEUPP et al. however, the FANCS patient described by Sawyer et al. rather than non‐homologous end‐joining predicts hereditary ovarian showed DEB‐caused cross‐linker–induced chromosomal ab- carcinoma. Oncotarget, 8(58), 98660–98676. errations, a hallmark of FA patient cells, which was not pres- Domchek, S. M., Tang, J., Stopfer, J., Lilli, D. R., Hamel, N., Tischkowitz, M., … Greenberg, R. A. (2013). Biallelic deleterious ent in our patient. BRCA1 mutations in a woman with early‐onset ovarian cancer. Cancer Discovery, 3(4), 399–405. https​ ://doi.org/10.1158/2159- ACKNOWLEDGMENTS 8290.CD-12-0421 Evers, B., & Jonkers, J. (2006). Mouse models of BRCA1 and BRCA2 We thank Mrs. Marlies Just, Ulm University, for EBV‐medi- deficiency: Past lessons, current understanding and future pros- ated LCL generation. This work was supported by the German pects. Oncogene, 25(43), 5885–5897. https​ ://doi.org/10.1038/ Cancer Aid to LW (Deutsche Krebshilfe #70112504) and to sj.onc.1209871 RKS (Deutsche Krebshilfe #110837 and #70111850). JS is Freire, B. L., Homma, T. K., Funari, M. F. A., Lerario, A. M., Leal, A. M., Velloso, E. D. R. P., … Jorge, A. A. L. (2018). Homozygous funded by MINECO grant SAF2015‐64152‐R, European loss of function BRCA1 variant causing a Fanconi‐anemia‐like Commission (EUROFANCOLEN), ICREA Academia and phenotype, a clinical report and review of previous patients. Fondos FEDER. European Journal of Medical Genetics, 61(3), 130–133. https​://doi. org/10.1016/j.ejmg.2017.11.003 Gagou, M. E., Zuazua‐Villar, P., & Meuth, M. (2010). Enhanced H2AX CONFLICT OF INTEREST phosphorylation, DNA replication fork arrest, and cell death in the The authors declare that they have no competing interests. absence of Chk1. Molecular Biology of the Cell, 21(5), 739–752. https​://doi.org/10.1091/mbc.E09-07-0618 Gatz, S. A., Salles, D., Jacobsen, E. M., Dork, T., Rausch, T., Aydin, ORCID S., … Wiesmuller, L. (2016). MCM3AP and POMP mutations cause a DNA‐repair and DNA‐damage‐signaling defect in an im- Eric Hahnen https://orcid.org/0000-0002-1152-8367 munodeficient child. Human Mutation, 37(3), 257–268. https​://doi. org/10.1002/humu.22939​ Goldgar, D. E., Easton, D. F., Deffenbaugh, A. M., Monteiro, A. N., R E F E R E NC E S Tavtigian, S. V., Couch, F. J., … C., (2004). Integrated evalua- tion of DNA sequence variants of unknown clinical significance: Akyuz, N., Boehden, G. S., Susse, S., Rimek, A., Preuss, U., Application to BRCA1 and BRCA2. American Journal of Human Scheidtmann, K. H., & Wiesmuller, L. (2002). DNA substrate de- Genetics, 75(4), 535–544. https​://doi.org/10.1086/424388 pendence of p53‐mediated regulation of double‐strand break repair. Hampp, S., Kiessling, T., Buechle, K., Mansilla, S. F., Thomale, J., Rall, Molecular and Cellular Biology, 22(17), 6306–6317. https​://doi. M., … Wiesmuller, L. (2016). DNA damage tolerance pathway in- org/10.1128/MCB.22.17.6306-6317.2002 volving DNA polymerase iota and the tumor suppressor p53 reg- Bouwman, P., van der Gulden, H., van der Heijden, I., Drost, R., Klijn, ulates DNA replication fork progression. Proc Natl Acad Sci U S C. N., Prasetyanti, P., … Jonkers, J. (2013). A high‐throughput A, 113(30), E4311–4319. https​://doi.org/10.1073/pnas.16058​28113​ functional complementation assay for classification of BRCA1 mis- Kast, K., Rhiem, K., Wappenschmidt, B., Hahnen, E., Hauke, J., sense variants. Cancer Discovery, 3(10), 1142–1155. https​://doi. Bluemcke, B., … Ovarian, C. (2016). Prevalence of BRCA1/2 org/10.1158/2159-8290.CD-13-0094 germline mutations in 21 401 families with breast and ovarian Bryant, H. E., Schultz, N., Thomas, H. D., Parker, K. M., Flower, D., cancer. Journal of Medical Genetics, 53(7), 465–471. https​://doi. Lopez, E., … Helleday, T. (2005). Specific killing of BRCA2‐de- org/10.1136/jmedg​enet-2015-103672 ficient tumours with inhibitors of poly(ADP‐ribose) polymerase. Keimling, M., Deniz, M., Varga, D., Stahl, A., Schrezenmeier, H., Nature, 434(7035), 913–917. https​://doi.org/10.1038/natur​e03443 Kreienberg, R., … Wiesmüller, L. (2012). The power of DNA Castella, M., Pujol, R., Callen, E., Ramirez, M. J., Casado, J. A., double‐strand break (DSB) repair testing to predict breast cancer Talavera, M., … Surralles, J. (2011). Chromosome fragility in pa- susceptibility. The FASEB Journal, 26(5), 2094–2104. https​://doi. tients with Fanconi anaemia: Diagnostic implications and clinical org/10.1096/fj.11-200790 impact. Journal of Medical Genetics, 48(4), 242–250. https​://doi. Keimling, M., Volcic, M., Csernok, A., Wieland, B., Dork, T., & org/10.1136/jmg.2010.084210 Wiesmuller, L. (2011). Functional characterization connects indi- Clapperton, J. A., Manke, I. A., Lowery, D. M., Ho, T., Haire, L. F., vidual patient mutations in ataxia telangiectasia mutated (ATM) Yaffe, M. B., & Smerdon, S. J. (2004). Structure and mechanism of with dysfunction of specific DNA double‐strand break‐repair sig- BRCA1 BRCT domain recognition of phosphorylated BACH1 with naling pathways. The FASEB Journal, 25(11), 3849–3860. https​:// implications for cancer. Nature Structural & Molecular Biology, doi.org/10.1096/fj.11-185546 11(6), 512–518. https​://doi.org/10.1038/nsmb775 Kuchenbaecker, K. B., Hopper, J. L., Barnes, D. R., Phillips, K. A., de Onis, M., Onyango, A. W., Borghi, E., Siyam, A., Nishida, C., & Mooij, T. M., Roos‐Blom, M. J., … Olsson, H. (2017). Risks of Siekmann, J. (2007). Development of a WHO growth reference breast, ovarian, and contralateral breast cancer for BRCA1 and for school‐aged children and adolescents. Bulletin of the World BRCA2 mutation carriers. JAMA, 317(23), 2402–2416. https​://doi. Health Organization, 85(9), 660–667. https​ ://doi.org/10.2471/ org/10.1001/jama.2017.7112 BLT.07.043497 Lindor, N. M., Guidugli, L., Wang, X., Vallée, M. P., Monteiro, A. N. Deniz, M. R. T., Kostezka, S., Faul, A., Gundelach, T., & Moreno‐ A., Tavtigian, S., … Couch, F. J. (2012). A review of a multifactorial Villanueva, M. et al. (2017). Increased single‐strand annealing KEUPP et al.    | 9 of 9 probability‐based model for classification of BRCA1 and BRCA2 Semlow, D. R., Zhang, J., Budzowska, M., Drohat, A. C., & Walter, variants of uncertain significance (VUS). Human Mutation, 33(1), J. C. (2016). Replication‐dependent unhooking of DNA interstrand 8–21. https​://doi.org/10.1002/humu.21627​ cross‐links by the NEIL3 glycosylase. Cell, 167(2), 498–511, e414. Maxwell, K. N., Domchek, S. M., Nathanson, K. L., & Robson, M. https​://doi.org/10.1016/j.cell.2016.09.008 E. (2016). Population frequency of germline BRCA1/2 mutations. Seo, A., Steinberg‐Shemer, O., Unal, S., Casadei, S., Walsh, T., Gumruk, Journal of Clinical Oncology, 34(34), 4183–4185. https​ ://doi. F., … King, M. C. (2018). Mechanism for survival of homozygous org/10.1200/JCO.2016.67.0554 nonsense mutations in the tumor suppressor gene BRCA1. Proc Moghadasi, S., Meeks, H. D., Vreeswijk, M. P., Janssen, L. A., Borg, A., Natl Acad Sci U S A, 115(20), 5241–5246. https​://doi.org/10.1073/ Ehrencrona, H., … Garcia, E. G. (2018). The BRCA1 c. 5096G>A pnas.18017​96115​ p.Arg1699Gln (R1699Q) intermediate risk variant: breast and ovar- Spurdle, A. B., Whiley, P. J., Thompson, B., Feng, B., Healey, S., ian cancer risk estimation and recommendations for clinical manage- Brown, M. A. … Goldgar, D. E. (2012). BRCA1 R1699Q variant ment from the ENIGMA consortium. Journal of Medical Genetics, displaying ambiguous functional abrogation confers intermediate 55(1), 15–20. https​://doi.org/10.1136/jmedg​enet-2017-104560 breast and ovarian cancer risk. Journal of Medical Genetics, 49(8), Obermeier, K., Sachsenweger, J., Friedl, T. W., Pospiech, H., Winqvist, 525–532. https​://doi.org/10.1136/jmedg​enet-2012-101037 R., & Wiesmuller, L. (2016). Heterozygous PALB2 c.1592delT mu- Xie, J., Litman, R., Wang, S., Peng, M., Guillemette, S., Rooney, T., tation channels DNA double‐strand break repair into error‐prone & Cantor, S. B. (2010). Targeting the FANCJ‐BRCA1 interaction pathways in breast cancer patients. Oncogene, 35(29), 3796–3806. promotes a switch from recombination to poleta‐dependent bypass. https​://doi.org/10.1038/onc.2015.448 Oncogene, 29(17), 2499–2508. https​://doi.org/10.1038/onc.2010.18 Pathania, S., Bade, S., Le Guillou, M., Burke, K., Reed, R., Bowman‐ Colin, C., … Livingston, D. M. (2014). BRCA1 haploinsuffi- ciency for replication stress suppression in primary cells. Nature Communications, 5, 5496. https​://doi.org/10.1038/ncomm​s6496​ SUPPORTING INFORMATION Rebbeck, T. R., Friebel, T. M., Friedman, E., Hamann, U., Huo, D., Additional supporting information may be found online in Kwong, A., … Nathanson, K. L. (2018). Mutational spectrum in a worldwide study of 29,700 families with BRCA1 or BRCA2 muta- the Supporting Information section at the end of the article. tions. Human Mutation, https​://doi.org/10.1002/humu.23406​ Sawyer, S. L., Tian, L., Kahkonen, M., Schwartzentruber, J., Kircher, M., … Greenberg, R. A.; University of Washington Center for How to cite this article: Keupp K, Hampp S, Hübbel Mendelian Genomics (2015). Biallelic mutations in BRCA1 cause A, et al. Biallelic germline BRCA1 mutations in a patient a new Fanconi anemia subtype. Cancer Discovery, 5(2), 135–142. with early onset breast cancer, mild Fanconi anemia‐like https​://doi.org/10.1158/2159-8290.CD-14-1156 phenotype, and no chromosome fragility. Mol Genet Schlacher, K., Wu, H., & Jasin, M. (2012). A distinct replication fork Genomic Med. 2019;7:e863. https​://doi.org/10.1002/ protection pathway connects Fanconi anemia tumor suppressors mgg3.863 to RAD51‐BRCA1/2. Cancer Cell, 22(1), 106–116. https​ ://doi. org/10.1016/j.ccr.2012.05.015