Report

Biallelic Inactivation of BRCA2 in Fanconi Anemia

See allHide authors and affiliations

Science  26 Jul 2002:
Vol. 297, Issue 5581, pp. 606-609
DOI: 10.1126/science.1073834

Abstract

Fanconi anemia (FA) is a rare autosomal recessive cancer susceptibility disorder characterized by cellular hypersensitivity to mitomycin C (MMC). Six FA genes have been cloned, but the gene or genes corresponding to FA subtypes B and D1 remain unidentified. Here we show that cell lines derived from FA-B and FA-D1 patients have biallelic mutations in BRCA2 and express truncated BRCA2 proteins. Functional complementation of FA-D1 fibroblasts with wild-typeBRCA2 complementary DNA restores MMC resistance. Our results link the six cloned FA genes with BRCA1 and BRCA2in a common pathway. Germ-line mutation of genes in this pathway may result in cancer risks similar to those observed in families withBRCA1 or BRCA2 mutations.

Fanconi anemia (FA) is a rare autosomal recessive cancer susceptibility syndrome characterized by congenital abnormalities, progressive bone marrow failure, and cellular hypersensitivity to DNA cross-linking agents, such as MMC and cisplatin (1, 2). FA patients often develop acute myeloid leukemia (AML), but also develop squamous cell carcinomas, frequently of the head and neck or of the gynecologic system (3). Whether heterozygote carriers of FA gene mutations have an increased cancer risk remains unknown (4).

At least eight distinct complementation groups of FA (A, B, C, D1, D2, E, F, G) have been defined by somatic cell fusion studies (5–7), and six FA genes have been cloned (A, C, D2, E, F, G). The six known FA proteins interact in a common pathway (8). Five of the FA proteins (A, C, E, F, G) assemble in a multisubunit nuclear complex. Either in response to DNA damage (8) or during S phase of the cell cycle (9), this complex activates the monoubiquitination of the downstream D2 protein, thereby targeting D2 to BRCA1-containing nuclear foci. Biallelic mutation of an upstream FA gene disrupts the monoubiquitination of FANCD2, resulting in loss of FANCD2 foci and hypersensitivity to MMC.

Recent studies have suggested genetic interactions among the breast cancer susceptibility genes, BRCA1 andBRCA2, and the FA genes. First, disruption ofBRCA1 results in loss of DNA damage–inducible FANCD2 foci, suggesting that BRCA1 may act as an “organizer” of FA foci (8). Accordingly, the BRCA1 protein has a Ring Finger E3 ubiquitin ligase domain and may ubiquitinate FANCD2 in vivo (10). Second, BRCA1 –/– orBRCA2 –/– tumor cells exhibit MMC hypersensitivity and chromosome instability (11–13), similar to the defects observed in FA cells (fig. S1). Functional complementation ofBRCA2 –/– cells with murine wild-typeBrca2 restores MMC resistance (14). Third, targeted inactivation of the murine Brca2 gene, disrupting the COOH-terminus of the BRCA2 protein but sparing the NH2-terminus, results in viable mice with an FA-like phenotype (i.e., small size, skeletal defects, hypogonadism, cancer susceptibility, chromosome instability, and MMC hypersensitivity) (15, 16).

To investigate the relation between BRCA genes and FA, we sequenced BRCA1 and BRCA2 in cells derived from FA-B and FA-D1 patients (Table 1and Fig. 1A). Although noBRCA1 mutations were detected, biallelic mutations inBRCA2 were observed. A homozygous mutation (IVS19-1 G to A) was detected in a BRCA2 splice acceptor site in the FA-D1 reference line, HSC62, predicted to result in partial or complete loss of exon 20. In another FA-D1 line, EUFA423, two definitiveBRCA2 mutations were identified, 7691insAT (exon 15) and 9900insA (exon 27). These mutant alleles were not detected in a screen of 120 random genomic DNA samples from the general population. Both mutations create frameshifts and are predicted to encode COOH-terminal truncated BRCA2 proteins. The 9900insA mutant allele has previously been identified in a breast cancer kindred (Table 1) (17).

Figure 1

FA-B and FA-D1 cells have biallelic BRCA2 mutations and express mutant BRCA2 proteins. (A) Schematic diagram of human BRCA2. The highly conserved BRC repeats, encoded by exon 11, mediate RAD51 interactions.BRCA2 mutations in EUFA423 (FA-D1), HSC62 (FA-D1), and HSC230 (FA-B) lymphoblasts are shown. Mutations were confirmed in primary cells. (B) Whole cell lysates were prepared from wild-type PD7 lymphoblasts, HeLa cells, CAPAN1, EUFA423 (FA-D1), HSC62 (FA-D1), and HSC230 (FA-B) lymphoblasts. CAPAN1 is a pancreatic carcinoma cell line that has lost one BRCA2allele and contains the 6174delT mutation in the remaining allele (19). Proteins were separated by electrophoresis and immunoblotted with a rabbit polyclonal antibody to BRCA2 (raised against amino acids 3245–3418 of BRCA2) (Ab-2, Oncogene Research) or (C) by a mouse monoclonal antibody to BRCA2 (raised against amino acids 1651–1821) (Ab-1, Oncogene Research). A protein in HeLa cell extracts (209 kD) was nonspecific. (D) Characterization of the BRCA2 protein in FA lymphoblasts from multiple complementation groups (subtypes A, C, D1, D2, E, F, and G). Proteins were immunoblotted with Ab-2. S, MMC sensitive; R, MMC resistant.

Table 1

FA patients with biallelic mutations inBRCA2. U/A, unassigned FA subtype. BIC, Breast Cancer Information Core. Seewww.nhgri.nih.gov/Intramural_research/Lab_transfer/Bic. Dashes indicate no recorded BIC entry.

View this table:

Paradoxically, the FA-B reference line, HSC230, also contained two abnormal BRCA2 alleles. One mutant allele contained a known 3033delAAAC frameshift mutation in exon 11, and the second allele contained the polymorphic stop codon (ter3326) in exon 27 (18). This latter allele has been detected in approximately 1% of normal controls in the U.S. population and is not associated with a strong cancer risk (18). That FA-D1 and FA-B cells had biallelic mutations in the same gene (BRCA2) suggests the possibility of intra-allelic complementation or phenotypic reversion to wild-type (6). Two additional cell lines from FA patients of unassigned subtype had biallelic mutations inBRCA2 (Table 1).

We next examined BRCA2 protein expression in the FA-D1 and FA-B cell lines (Fig. 1, B to D). An antibody to the COOH- terminus of BRCA2 (Ab-2) recognized full-length BRCA2 (380 kD) in normal control lymphoblasts, HeLa cells, and HSC62 (FA-D1) (Fig. 1B, lanes 1, 2, and 5). EUFA423 cells expressed a truncated BRCA2 protein (230 kD, lane 4), and no BRCA2 was detected in CAPAN1 (19) or HSC230 (FA-B) cells with this antibody (lanes 3 and 6). Probing again but with a different antibody (Ab-1, see epitope in Fig. 1A) revealed expression of BRCA2 in EUFA423 and HSC230 (Fig. 1C, lanes 4 and 6), suggesting that these BRCA2 polypeptides are truncated at the COOH-terminus (BRCA2C'Δ, 370 kD). Ab-1 also recognized the truncated BRCA2 protein (BRCA2-Trunc) in EUFA423 (lane 4), suggesting that this isoform has an internal deletion between the two antibody epitopes. Taken together, these results indicate that EUFA423 and HSC230 express BRCA2 polypeptides with small COOH-terminal truncations, consistent with the presence of mutations in exon 27 (Table 1). Cell lines from other FA subtypes displayed approximately equal levels of full-length BRCA2 (Fig. 1D).

Although HSC62 cells express BRCA2 protein of approximately normal size, the mutation (IVS19-1 G to A) predicts the presence of an abnormally spliced messenger RNA (mRNA). To test this, we performed reverse transcriptase–polymerase chain reaction (RT-PCR) of theBRCA2 mRNA, followed by direct cDNA sequencing (Fig. 2, A to C). As a result of this mutation, the BRCA2 mRNA lacks the first 12 bases of exon 20, corresponding to an in-frame deletion of four amino acids from BRCA2 (amino acids 2830 to 2833) (Fig. 2C, fig. S2). No normalBRCA2 mRNA was detected in HSC62 cells. The mutant protein may have partial activity because the HSC62 patient has a relatively mild clinical FA phenotype (table S1) and the HSC62 cells have only modest MMC sensitivity (20) (Table 2).

Figure 2

The FA-D1 reference line, HSC62, expresses a BRCA2 protein with an internal deletion of four amino acids. (A) Schematic representation of the RT-PCR reaction, resulting in specific amplification of a region of the BRCA2mRNA. PCR products from the indicated cell lines were analyzed on a 1% agarose gel. (B) PCR products were analyzed by direct DNA sequencing. (C) BRCA2 mRNA from HSC62 cells has an internal deletion of the first 12 bases from exon 20, resulting in an in-frame deletion of the indicated four amino acids. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; E, Glu; F, Phe; G, Gly; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; W, Trp; and Y, Tyr.

Table 2

Chromosome breakage analysis of FA and control cell lines was performed as in (7). Groups of experiments for lymphoblasts are separated by a line space. I and II indicate experiment number for fibroblast cell lines. S, MMC-sensitive; R, MMC-resistant; L, EBV-transformed lymphoblasts; F, SV-40 transformed fibroblasts; ND, not determined; wt, wild type.

View this table:

We next determined, by means of genomic PCR with specific flanking primers and direct sequencing, whether the BRCA2 mutant alleles segregate in the EUFA423 kindred (Fig. 3). The paternal allele was 7691insAT and the maternal allele was 9900insA (Fig. 3A). The proband (EUFA423) was a compound heterozygote, whereas two of the three unaffected siblings were BRCA2 carriers. Lymphoblasts from all heterozygousBRCA2 carriers expressed full-length BRCA2 (Fig. 3B).

Figure 3

Segregation of BRCA2 mutant alleles in the EUFA423 pedigree. (A) The proband with FA subtype D1 is EUFA423. Genomic DNA was prepared from lymphoblasts from the indicated family members and was sequenced forBRCA2 mutations. (B) Expression of mutant BRCA2 polypeptides in lymphoblasts derived from EUFA423 kindred. Proteins were immunoblotted with Ab-2. (C) EUFA423F was transfected either with pcDNA3-empty vector or pcDNA3-HA-BRCA2 (28, 29), and stable G418-resistant cells were isolated. Cell lines were analyzed by immunoblot with Ab-2 and by the MMC chromosome breakage assay (Table 2).

Next, we stably transfected EUFA423 fibroblasts with cDNA encoding the full-length wild-type BRCA2 protein (Fig. 3C). G418-selected cells expressed full-length BRCA2 (Fig. 3C, lane 3) and exhibited a correction of their MMC sensitivity (Table 2). Similarly, transfection with human chromosome 13, which contains the wild-typeBRCA2 gene, corrected the MMC hypersensitivity (Table 2). Taken together, these results confirm that BRCA2 is a FA gene.

FA has an estimated incidence of less than 1 per 100,000 live births, and less than 5% of FA families are assigned to subtypes B and D1.BRCA2 mutations have a cumulative carrier frequency of approximately 1% of the U.S. population (17). ThisBRCA2 carrier frequency predicts a higher incidence ofBRCA2 homozygotes than the observed FA incidence. On the basis of our limited sample collection, FA patients have at least one mutation in the 3′ region of BRCA2. Thus, only a subset ofBRCA2 –/– individuals (namely, those expressing truncated BRCA2 proteins with partial activity) may manifest the FA phenotype. Homozygous disruption of the 5′ end of BRCA2, in contrast, may result in embryonic lethality, as it did in the mouse model (21–23).

Specific BRCA2 mutations may vary in cancer risk (17). The 6174delT mutation found in Ashkenazi Jews may confer a breast cancer risk as high as 70% by age 70. Other variant BRCA2 alleles, such as the polymorphic stop codon ter3326, appear to cause no increased cancer risk (18) but may cause FA in the compound heterozygous state. The smallest known cancer-associated deletion removes only 224 amino acids from the COOH-terminus of BRCA2 (24). Due to the unavailability of clinical records, we were unable to assess the cancer risk of the BRCA2 mutant alleles in these FA families (Table 1).

FA patients with biallelic BRCA2 mutations share clinical features with FA patients from other subtypes (i.e., congenital abnormalities, abnormal skin pigmentation, bone marrow failure, and cellular sensitivity to MMC) (25) (table S1). These similarities suggest that BRCA2 and other FA proteins cooperate in a common DNA damage response pathway, the FA/BRCA pathway (Model, fig. S3A). According to this model, DNA damage activates the monoubiquitination of FANCD2, thus targeting FANCD2 to DNA repair foci containing BRCA1 and BRCA2 (26). Previous studies have indicated that FA-B cells lack FANCD2 monoubiquitination, whereas FA-D1 cells express monoubiquitinated FANCD2 (8) (fig. S3B). BRCA2 may function upstream in the pathway, by promoting FA complex assembly and FANCD2 activation, and/or downstream in the pathway, by transducing signals from FA proteins to RAD51 and the homologous recombination machinery (27). The precise molecular function(s) of BRCA1 and BRCA2 in this pathway remain to be elucidated.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1073834/DC1

Materials and Methods

Figs. S1 to S3

Table S1

  • * To whom correspondence should be addressed. E-mail: alan_dandrea{at}dfci.harvard.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article