Breast cancer is the most commonly diagnosed cancer in women in Europe today. A hereditary form of breast cancer is characterized by young age onset, increased risk of bilateral breast cancer, and its being frequently in association with ovarian cancer. The existence of an autosomal dominant pattern of inheritance accounting for 5–10% of the breast cancer cases has been demonstrated 1. Germline mutations in BRCA1 (OMIM#113705, Online Mendelian Inheritance in Man) and BRCA2 (OMIM#600185) genes are responsible for an important fraction of hereditary breast and ovarian cancers 2. A few hundred different mutations associated with inherited predisposition to breast and ovarian cancers have been identified in the BRCA1 and BRCA2 genes, as described at the Breast Cancer Information Core internet web site (BIC database) 3.

Most reported germline deleterious mutations are nonsense substitutions and small deletions/insertions causing truncations of BRCA1/2 proteins. In most populations tested, the observed frequencies of BRCA1 variations in high-risk breast and/or ovarian cancer families have been described as lower than predicted by linkage analysis. Pathogenic mutations in the coding region or in splice sites of the BRCA1 gene were found in approximately two-thirds of BRCA1-linked families 2. This finding suggests that methods generally used for mutation scanning fail to detect certain types of BRCA1 germline defects, such as large intragenic rearrangements. Most of the screening methods based on the PCR of genomic DNA are qualitative rather than quantitative 4. Partial or complete exon loss or amplification might be overlooked because of the presence of a wild-type allele that gives rise to a positive PCR signal and therefore a possible false-negative result. Several approaches have been used for detecting BRCA1 rearrangements: Southern blot 56789, long-range PCR 10, color bar coding of the BRCA1 gene on combed DNA 1112, semiquantitative-multiplex PCR 1314 or real-time PCR 15. Recently, multiplex ligation-dependent probe amplification (MLPA) has been widely used as a highly sensitive method for detecting the relative copy number of all BRCA1 exons in a high-throughput format 16.

Many different BRCA1 germline rearrangements with mapped breakpoints have been reported to date 17. These are scattered throughout the whole gene and most of them are deletions, but duplication, triplication or combined deletion/insertion events also have been described. A genetic structure of BRCA1 with extremely high density of intronic Alu repeats and the presence of a duplicated promoter region containing a BRCA1 pseudogene upstream of the BRCA1 could provide hotspots for unequal homologous recombination 1819. The proportion of genomic rearrangements in the BRCA1 mutation spectrum has been studied in several countries and often was found to vary from 8 to 15% 720212223. Higher values, probably due to a strong founder effect, have been presented by studies performed in the Netherlands and Italy 2425. The majority of known rearrangements create frame shifts that result in premature termination of translation. Therefore, the phenotype of patients carrying this type of mutations is not expected to be distinct from patients with other truncating mutations. Most mutations introducing a stop codon into BRCA1 have been described as leading to nonsense-mediated decay of mRNA, irrespective of their type 26. Large deletions of one or more exons maintaining the reading frame could cause loss of putative functional domains of the BRCA1 protein 8. However, no assay for the BRCA1 gene is currently available for testing the exact functional consequences of such mutations.

The presence of large rearrangements in the BRCA1 gene offers a promising outlook in clinical practice, and especially for probands with previously negative results of BRCA1/2 mutation screening. If causative mutation is determined, predictive testing can be performed to identify family members who may benefit from increased surveillance, chemoprevention or prophylactic surgery to reduce the risk of developing cancer 27. The aim of this study was to determine the frequency and type of BRCA1 intragenic rearrangements in Czech high-risk breast and/or ovarian cancer families where no deleterious mutations were previously found and to assess whether testing for such rearrangements should be included in standard mutation screening.

The MLPA analysis of genomic DNA of 172 affected individuals from high-risk families with hereditary breast and/or ovarian cancer syndrome revealed six different deletions covering 1.9–36.9 kb of the BRCA1 genomic DNA (Figure 1, Table 3). The average age at the onset of the cancer diagnosis in probands with detected BRCA1 rearrangements was 40 years. The effect of decreasing age at onset of the illness in younger generations could be observed in some families. Except for two rearrangements including the exons 5–14 and 11–12, the deletions spanning the exons 1A/1B-2, 18–19, 20 and 21–22 had been previously described in the literature 68192539. However, the breakpoints of five out of six deletions detected in Czech patients were different from those characterized previously (Table 3).

The deletion of 1A/1B-2 exons was detected by MLPA in one family with the mother affected with ovarian cancer at age 43 and with two relatives affected with breast cancer at young age: her daughter at the age of 38 and her sister at age 39. The deletion of 1A/1B-2 exons was confirmed by long-range PCR with primers published by Puget et al. 19 (Figure 2A). The intensive PCR product was created only for the mutant allele and sequencing revealed the 36.9 kb deletion extending from intron 2 of the ψBRCA1 to intron 2 of the BRCA1 gene. The breakpoint region is consistent with that reported previously by Preisler-Adams et al. 40 and comprises 188 bp of perfect nucleotide homology between sequences located in the intron 2 of ψBRCA1 and intron 2 of the BRCA1 gene (Figure 2A). Based on the BAC clone sequence AC060780, the breakpoints occurred between nucleotides 71053–71240 in ψBRCA1 and nucleotides 34118–34305 in BRCA1 intron 2 40. The mutant allele harbors a chimeric gene consisting of the ψBRCA1 exons 1A, 1B and 2 fused to the BRCA1 exons 3–24 as a result of recombination between sequences located in the intron 2 of ψBRCA1 and intron 2 of the BRCA1 gene. The promoter was shown to be absent from this mutant allele and expression of the mutant allele was not observed 1925.

Deletion of the BRCA1 exons 5 to 14 was revealed by MLPA in four high-risk, severely affected families with young-age onset of cancer: two families with the hereditary breast-ovarian cancer syndrome and two families with only the breast cancer syndrome. Long-range PCR confirmed a deletion of 31.5 kb of genomic DNA spanning more than three-quarters of the BRCA1 gene coding sequence. This deletion interferes with the RING domain (N-terminal zinc finger domain) of the BRCA1 protein as well as with important interaction domains for multiple proteins, and it is suggestive of having a negative impact on the function of the BRCA1 protein. The exact breakpoints were characterized by sequencing as g.21716_53298del31583 (Figure 2B, Table 3). The breakpoint junctions determined in all four families were identical, thus supporting the likelihood of the founder effect. Moreover, three of the four individuals with confirmed deletion of exons 5–14 came from the same geographical region of the Czech Republic. Only haplotype analysis could reveal if repeated observation of these four deletions is due to the presence of founder mutation or only to the local instability in a region 722.

The deletion of the second part of the exon 11 and exon 12 was found by MLPA in a family with ovarian cancer phenotype only. Altogether, five women in two generations were affected with ovarian cancer at ages ranging from 34 to 56 years (Figure 3). Long-range PCR confirmed a deletion of 6.5 kb of genomic DNA spanning nearly one-half of the BRCA1 gene coding sequence. The exact breakpoints were characterized by sequencing as g.34845_41405del6561 (Figure 2C, Table 3). An aberrant splicing of mRNA that might further extend a defect on the BRCA1 protein is suspected in this case. This deletion lies within a central-risk region where mutations were associated with a significantly higher ovarian/breast cancer ratio 41.

The deletion of the exons 18 and 19 was revealed by MLPA in one woman diagnosed with both breast and ovarian cancers at the ages of 52 and 59, respectively, and with a family history of breast cancer in her second-degree relatives. Long-range PCR confirmed a deletion of nearly 2 kb of BRCA1 genomic DNA. The exact breakpoints were characterized by sequencing as g.63651_65590del1940 (Figure 4A, Table 3). The out-of-frame deletion of exons 18–19 is predicted to result in a truncation of the BRCA1 protein at the codon 1693. As displayed previously, most truncating mutations of the BRCA1 gene lead to nonsense-mediated mRNA decay and an allelic imbalance in the expression of the mutant versus wild-type allele 2642.

The deletion of the exon 20 was found by MLPA in a woman affected with breast and colorectal cancers at the ages of 37 and 35, respectively. Her mother was affected with bilateral breast cancer at the ages of 39 and 46. Long-range PCR confirmed a deletion of about 7 kb of BRCA1 genomic DNA. The exact breakpoints were characterized by sequencing as g.68764_75792del7029 (Figure 4B, Table 3). The genomic deletion of exon 20 causes at least in-frame deletion of exon 20 in mRNA and results in the removing of 28 amino acids of the BRCA1 protein in position 1732–1759. This deletion affects the highly conserved area of the BRCA1 protein corresponding to the linker between two BRCT (BRCA1 C-terminal) domains, and it has been described to be involved in DNA repair and transcription activation 43.

The next in-frame deletion of the exons 21 and 22 was detected by MLPA in two high-risk families with hereditary breast cancer phenotype only (Figure 5). Long-range PCR confirmed a deletion of about 3.5 kb of BRCA1 genomic DNA. Sequencing of the breakpoint's region revealed the deletion/insertion event characterized as g.77128_80906del3779ins236 (Figure 6, Table 3). However, the in-frame deletion of exons 21–22 removes at least 43 amino acids of the BRCA1 protein in position 1760–1802, corresponding to a part of the C-terminal BRCT domain.

The loss of a part of the conservative domain of BRCA1 protein might have an effect on protein function and is suspected to be causative of cancer susceptibility. Even missense mutations located in this region (for example, P1749R and M1775R) were described to segregate with the disease and to have a destabilizing effect on the BRCT domain 44. However, the exact changes in the function of the BRCA1 protein cannot be determined because a functional assay is lacking. Therefore, the structural model of deletions in the BRCT domain was constructed to help with the interpretation of an effect on the stability of the BRCT domain (Figure 7), which might be indirectly related to the disease risk 44.

Inspection of the crystal structures of the BRCT repeat region from the BRCA1 revealed that the mutant with genomic deletion of exon 20 encodes the protein with the missing linker region between the N-terminal and the C-terminal BRCT repeats of BRCA1, while the in-frame deletion of exons 21 and 22 results in the protein with the missing C-terminal repeat of BRCT (Figure 7). The two BRCT domains interact in a head-to-tail fashion, burying about 1600 A² of solvent-accessible surface area in the interface 45. The truncation of the linker or the exclusion of any of the two BRCT domains will result in the exposure of the interface and the protein's unfolding (Figure 7). It is obvious that the protein with a disrupted carboxyl-terminal BRCT repeat region cannot fulfill its tumor suppressor function. This structural interpretation is in accordance with the study of Williams et al. 4446, who used a protease-based assay to assess the sensitivity of the folding of the BRCT domain to an extensive set of truncation and single amino acid substitutions derived from breast cancer screening programs. The protein can tolerate truncations of up to eight amino acids, but further deletion leads to BRCT folding defects.

It would have been interesting to determine whether detected deletions segregate with disease in affected families, but no informative data are available. In families with the deletion of exons 11+12 and the deletion of exons 21+22, all affected patients tested (3 and 2) were found to be carriers of the deletion (Figs. 3 and 5). In the cases of the remaining eight families, as there were no other affected individuals alive, only healthy individuals from mainly younger generations could be tested.

Most of the previously characterized rearrangements in the BRCA1 gene result from an unequal homologous recombination of Alu repeats 17. The presence of Alu elements revealed by the Repeat Masker program in breakpoint junctions in three out of five novel deletions characterized in this study supports this hypothesis. Contrary to the Alu-mediated deletions of exons 18–19, 20 and 20–21, no Alu repeats were found in the breakpoints of the deletions including exons 11–12 and 5–14 (Table 3). In these two cases, L1 repetitive sequences present near one side of the breakpoint do not correlate with any recognizable repeat motifs opposite, suggesting nonhomologous events or other mechanisms 1039. The role of the L1 repetitive elements in relation to the BRCA1 intragenic rearrangements has not yet been mentioned.

The intragenic deletions were detected in almost 6% of all high-risk families previously considered negative for the mutation in BRCA1 or BRCA2 genes. It represents 10.4% of all detected BRCA1 mutations and 7.8% of all mutations detected in BRCA1 and BRCA2 genes (in 290 high-risk families tested for mutations during 1999–2006, Table 1). This ratio is dependent on the selection of patients and the mutation detection rate. As can be seen from Table 1, BRCA1 rearrangements were found either in breast plus ovarian cancer families (5 cases) or in breast cancer families with at least one individual with breast cancer diagnosed under 50 years of age (5 cases). The deletions were identified in all three families, each with four individuals affected by breast cancer under age 50. By contrast, no rearrangement was detected in families with late-onset multiple breast cancer cases. In affected families, ovarian cancer or at least one case of breast cancer under age 50 seems to suggest the presence of BRCA1 gross rearrangement.

The proportion of intragenic BRCA1 rearrangements could be overestimated because of a higher objective amount of pathogenic mutations in the Czech population. BRCA1/2 mutations were excluded in this study using a combination of heteroduplex analysis and protein truncation tests. Heteroduplex analysis is aimed to detect small insertions and deletions and not at detecting single base changes. The protein truncation test is a convenient method for rapidly scanning relatively large fragments for protein-terminating variants, but it is incapable of identifying some potentially risky missense variants or small in-frame deletions located in exon 11 of BRCA1 or exon 10 and 11 of BRCA2.

Our results are in good concordance with those obtained by studies performed in other countries: the proportion of the BRCA1 intragenic rearrangements is slightly higher than are those estimated in France, Germany, Spain and in the United States 7202123 but lower than in the populations of Australia or New Zealand 22. Higher proportions of BRCA1 rearrangements have been observed in the Netherlands, due to the founder mutations representing 23% of all BRCA1 mutations found 24, and in a small population in Northern Italy 25. On the other hand, a study performed in Finland failed to detect any rearrangements in the BRCA1 gene 47. The most likely explanation for varying prevalencies of large rearrangements is in the differing genetic backgrounds of the populations studied. The study size and selection bias may be relevant, too. Methodology is not supposed to be a major factor, because the same commercially available MLPA kit is widely used. Our results indicate that MLPA is a rapid, reliable and sensitive technique allowing high-throughput screening for the BRCA1 rearrangements.

Using MLPA technique, intragenic rearrangements were detected in approximately 6% of the Czech high-risk families previously designated as BRCA1/2 mutation negative. Six different intragenic deletions represent more than 10% of all detected BRCA1 mutations. Our results prove the usefulness of testing for large BRCA1 rearrangements in the Czech population. These results are important for counseling purposes and clinical management of patients as well as for the possibility of predictive testing of relatives. MLPA testing of the BRCA2 rearrangements is now under examination and might further improve the sensitivity of testing.

The author(s) declare that they have no competing interests.

PV and EM contributed equally to this work.

PV, EM and LF participated in design of this study.

PV, EM and ML selected patients and their family members appropriate for this study, carried out molecular analyses and interpreted the results.

LF and MN participated in genetic counseling and selection of patients.

PV drafted the paper and finalized the manuscript with help of EM and LF.

OH carried out capillary electrophoresis on an ABI PRISM 310 genetic analyzer.

JD performed structural analysis of BRCT domain and its interpretation.

HP, VK and JK participated in mutation screening.

All authors read and approved the final manuscript.