Genomic rearrangements in BRCA1 and BRCA2: A literature review

Women with mutations in the breast cancer genes BRCA1 or BRCA2 have an increased lifetime risk of developing breast, ovarian and other BRCA-associated cancers. However, the number of detected germline mutations in families with hereditary breast and ovarian cancer (HBOC) syndrome is lower than expected based upon genetic linkage data. Undetected deleterious mutations in the BRCA genes in some high-risk families are due to the presence of intragenic rearrangements such as deletions, duplications or insertions that span whole exons. This article reviews the molecular aspects of BRCA1 and BRCA2 rearrangements and their frequency among different populations. An overview of the techniques used to screen for large rearrangements in BRCA1 and BRCA2 is also presented. The detection of rearrangements in BRCA genes, especially BRCA1, offers a promising outlook for mutation screening in clinical practice, particularly in HBOC families that test negative for a germline mutation assessed by traditional methods.


Introduction
The precise identification of germline BRCA1 and BRCA2 mutations is a major concern for geneticists counseling families with a high risk of breast and ovarian cancers. The most frequent mutations encountered in these genes are deletions or insertions of a few bases or single-base substitutions that result in premature stop codons Narod and Foulkes, 2004). Such point mutations occur throughout the coding sequence of both genes and account for 10%-50% of the germline mutations encountered in hereditary breast and ovarian cancer (HBOC) families, depending on the inclusion criteria used (Agata et al., 2005;Vasickova et al., 2007).
The observed frequencies of BRCA1 mutations are lower than predicted by linkage analysis, with pathogenic variations in the coding region or splice sites of the gene being found in approximately two-thirds of BRCA1-linked families. This finding suggests that other dominant genes (Ford et al., 1998;Armour et al., 2002) and/or low penetrance alleles, such as the 1100delC mutation in CHEK2, may be associated with the HBOC phenotype (Puget et al., 1999;Nevanlinna and Barker, 2006). Indeed, breast and ovarian cancers have been associated with germline mutations in other genes that are involved in the maintenance of genomic integrity, such as TP53, PTEN, ATM, NBS1, RAD50, BRIP1 and PALB2. Inherited breast cancer is currently considered a highly heterogeneous genetic disease with respect to both the loci and alleles involved (Walsh et al., 2006;Walsh and King, 2007).
Large genomic rearrangements have recently been identified in HBOC families and account for a small but still significant proportion of cases in several populations. These mutations are usually pathogenic because deletions or insertions of large genomic sequences within a coding region result in out-of-frame translation and usually lead to a mutant peptide of abnormal structure and/or function (Preisler-Adams et al., 2006). These mutations may be overlooked by most of the available screening and diagnostic PCR-based methods that use qualitative rather than quantitative methods and do not detect partial or complete exon losses or gains (Armour et al., 2002). Large genomic rearrangements of BRCA1 may account for up to one-third of all disease-causing mutations in various populations, while large genomic rearrangements in BRCA2 are less frequently observed (Hansen et al., 2009).

Frequency of Large Rearrangements
As shown in Table 1, the frequency of large genomic rearrangements varies considerably among populations. Among HBOC families, the highest proportion of BRCA1 rearrangements has been observed in northern Italy, where large genomic deletions account for approximately onethird of the pathogenic BRCA1 mutations (Montagna et al., 2003) and the overall prevalence of rearrangements in the families studied is 23%. In the Netherlands, rearrangements also represent a high proportion of all deleterious mutations in BRCA1 (27%-36% of all germline mutations in the gene) and are attributable to founder mutations (Petrij-Bosch et al., 1997;Hogervorst et al., 2003). In contrast, western Danish families with HBOC have a BRCA1 rearrangement prevalence of 3.8% (Thomassen et al., 2006). Another study done in Finland failed to detect any rearrangements among 82 families with moderate or high risk for HBOC (Lahti-Domenici et al., 2001). The latter two studies indicate a lower frequency of genomic rearrangements in Nordic countries. Finally, a study in Canada found no evidence of BRCA1 or BRCA2 genomic rearrangements in high-risk French-Canadian breast/ovarian cancer families (Moisan et al., 2006). This wide range in the prevalence of rearrangements is most likely related to the different genetic backgrounds of the populations studied, although the heterogeneity of the clinical inclusion criteria used for HBOC in each study may also have influenced the results. Furthermore, the prevalence of rearrangements will be different in samples that include only BRCA mutation-negative individuals by sequencing compared to those that include previously untested individuals at risk for HBOC. More recent studies have encountered an intragenic rearrangement prevalence of 6% and 12%, respectively, in high-risk patients in families from the Czech Republic and the United States of America who were negative for BRCA1/2 point mutations by sequencing (Walsh et al., 2006;Vasickova et al., 2007). In Germany, the prevalence of BRCA1 rearrangements is lower, ranging from 1 in 59 (1.7%) to 1 in 17.5 (5.7%) among high-risk families who are mutation-negative by sequencing (Hofmann et al., 2003;Hartmann et al., 2004;Preisler-Adams et al., 2006).
Only a few studies have examined the prevalence of BRCA2 rearrangements in larger sets of high-risk patients. In a report from Australia, large genomic rearrangements in BRCA2 were identified in 2% of 149 high-risk families that tested negative for BRCA1 and BRCA2 point mutations (Woodward et al., 2005). Agata et al. (2005) found a similar frequency (2.5%) of BRCA2 rearrangements among 121 highly selected Italian families. In a recent study of Portuguese HBOC families, a single founder BRCA2 rearrangement (c.156_157insAlu) was identified in 8% of the families studied and is the most frequent BRCA2 rearrangement described to date (Machado et al., 2007).

Molecular Pathology of BRCA1 Rearrangements
Several BRCA1 germline rearrangements with well characterized breakpoints have been reported (Mazoyer, 2005). These rearrangements are scattered throughout the gene and although most of them are deletions, duplications, triplications or combined deletion/insertion events have also been described. The BRCA1 gene characteristically has an extremely high density of intronic Alu repeats and a duplicated promoter region containing a BRCA1 pseudogene that most likely account for the occurrence of "hot spots" that favor unequal homologous recombination events (Smith et al., 1996;Puget et al., 2002). Currently, 45 different large genomic rearrangements have been characterized worldwide, including deletions and duplications of one or more exons (Table 1).

Alu sequences
The human genome contains up to 1 million copies of interspersed Alu elements (approximately one Alu repeat for every 5 kb) that apparently mediate chromosomal rearrangements and homologous recombination events, resulting in translocations, duplications, inversions or deletions (Kolomietz et al., 2002;Tancredi et al., 2004). These sequences are named Alu because most of the members of this family of repeats are cleaved by the bacterial restriction endonuclease Alu I. Members of the Alu family show significant homology but do not have identical sequences. Around 500,000 members of the Alu family have been identified and it is estimated that together they comprise 3% of the human genome. Approximately 41.5% of the intronic sequences of BRCA1 consist of Alu elements (Figure 1) that range in size from 0.5 kb to 23.8 kb and are located throughout the entire gene (Montagna et al., 1999).
Alu sequences have often been regarded as genomic instability factors because they are responsible for recombinational "hot spots" in certain genes and are frequently involved in exon shuffling during meiosis as a result of non-homologous recombination. These sequences may also act as regulatory factors in transcription, with structural roles (as "physical separators" of protein-protein 438 Ewald et al.  interactions during chromosome condensation in cellular division) and functional roles (in alternative "splicing" or as a connection between transcription factors) being proposed.
The two most prevalent sub-classes of repetitive elements in the Alu family are LINEs (Long Interspersed Elements) and SINEs (Short Interspersed Elements). LINEs span 6-8 kb and represent~21% of the total human genome DNA, whereas SINEs, which are derived from RNA polymerase transcripts, are shorter (100-300 bp) and represent 13% of the human genome. LINEs and SINEs are mobile elements that move via reverse transcription (Gad et al., 2001).
The complete genomic sequence of BRCA1 was published by Smith et al. (1996), who identified 138 individual Alu elements within this gene. Rearrangements are less common in the BRCA2 gene, probably because of a lower frequency of Alu sequences (17%). In most of the well characterized rearrangements described in the literature, there is good evidence for the involvement of Alu repeat elements in the recombination event. For example, the BRCA1 exon 5-7 deletion described in German families results from a non-allelic homologous recombination between AluSx in intron 3 and AluSc in intron 7. Both Alu repeats share a ho-mologous region of 15 bp at the crossover site. (Preisler-Adams et al., 2006) Non-functional pseudogenes Another important cause of unequal recombination within the coding region of certain genes is the presence of non-functional pseudogenes with high sequence homology to at least parts of the functional gene. Pseudogenes are usually non-functional "relatives" of known genes that have lost their protein-coding ability or are no longer expressed in the cell (Vanin, 1985). Puget et al. (2002) were the first to report this mutational mechanism for the BRCA1 gene. In two families with HBOC, these authors showed that the first exons of the gene were replaced by those of the BRCA1 pseudogene, yBRCA1. This pseudogene had previously been shown to lie~30 kb upstream of BRCA1 (Barker et al., 1996;Brown et al., 1996). The presence of a duplication containing most of BRCA1 exons 1 and 2 and the identification of two different recombination events involving homologous regions located in the BRCA1 gene and yBRCA1, respectively, led the authors to postulate that these regions were strong "hot spots" for recombination. The mutant alleles identified in 440 Ewald et al. the study harbored a chimeric gene that consisted of yBRCA1 exons 1A, 1B, and 2 fused to BRCA1 exons 3-24. This chimeric gene lacked both the BRCA1 promoter and translation initiation codon and was therefore non-functional (Hofmann et al., 2003).

Tandemly arranged short sequence repeats
Gross chromosomal deletions and/or insertions may also be mediated by tandemly arranged short sequence repeats. Highly repetitive nonconding human DNA often occurs in arrays (or blocks) of tandem repeats of sequences which may be simple (1-10 nucleotides) or moderately complex (tens to hundreds of nucleotides). Individual arrays can occur at a few or many different chromosomal locations. Satellite DNA, which constitutes most of the heterochromatic regions of the genome and is particularly noticeable in the vicinity of centromeres, consists of very large arrays of tandemly repeated DNA. Short repeats may cause slipped mispairing during replication, resulting in deletions or duplications of varying sizes. Recombination involving tandemly arranged short sequence repeats underlies the 244 bp deletion in BRCA1 exon 5 described in German HBOC families (Preisler-Adams et al., 2006).

BRCA2 Rearrangements
Only a few studies have investigated the presence and frequencies of deleterious BRCA2 rearrangements, and most of these were either done on a relatively small number of families or used cumbersome mutation detection methods of variable sensitivity (Agata et al., 2005).
Until recently, only two genomic rearrangements had been identified in six studies that analyzed hereditary breast cancer patients or primary breast tumors among diverse European populations (Peelen et al., 2000;Lahti-Domenici et al., 2001;Chin et al., 2001;Wang et al., 2001;Gad et al., 2002;Bunyan et al., 2004). The greatly reduced incidence of large genomic alterations that affect BRCA2 compared to BRCA1 most likely reflects differences in the density of Alu repeat sequences at the two loci, and these initial studies were not very supportive of the inclusion of this type of analysis in routine mutation testing of HBOC families (Preisler-Adams et al., 2006).
To date, 16 BRCA2 germline rearrangements have been reported. More recent studies have reported the frequent occurrence of large genomic BRCA2 rearrangements in male breast cancer families. Woodward et al. (2005) reported three BRCA2 rearrangements in 25 families with at least one male breast cancer, but no BRCA2 rearrangements in 114 families without male breast cancer, and Tournier et al. (2004) described three BRCA2 rearrangements in 39 French families with at least one case of male cancer. These findings indicate that large genomic rearrangements in BRCA2 are more frequent in families with male breast cancer.

Methods for Detecting Rearrangements
Classic methods for mutation detection (such as sequencing) are usually unable to identify large genomic rearrangements. Consequently, several alternative methods have been developed for the analysis of structural genomic abnormalities. These methods, which are designed to target either one or a few specific loci, or to scan the whole genome, include Southern blotting, long-range PCR, fluorescent in situ hybridization (FISH), quantitative multiplex PCR of short fluorescent fragments (QMPSF), protein truncation test (PTT), comparative genomic hybridization (CGH), real-time or quantitative PCR (RT-PCR or qPCR) and multiplex ligation-dependent probe amplification (MLPA). Although each of these methods has potential advantages and limitations, there have been very few largescale comparative analyses of these techniques. A brief summary of the most common detection methods is provided below.

Southern blotting
Southern blotting is the transfer of DNA fragments from an electrophoretic gel to a membrane support that results in immobilization of the fragments on the membrane and in a semipermanent reproduction of the banding pattern of the gel. This technique can be used to detect changes in copy number (deletions and duplications) when samples are run in parallel (concomitantly) with an internal standard. In addition, large rearrangements may also be detected by a size shift in the blotted DNA fragments. Although frequently used in the past, this method has lost popularity as a routine diagnostic procedure since it is laborious, time consuming, requires large amounts of highmolecular weight DNA and its interpretation may be hampered by false-negative results (Unger et al., 2000;Brown, 2001;De Lellis et al., 2007).

Long-range PCR
Long-range PCR uses a mixture of two thermostable DNA polymerases (proofreading and non-proofreading), thereby increasing the product size to 35 kb. The method has been useful for identifying specific large aberrations, including intragenic deletions, insertions, duplications and chromosomal breakpoints in several disorders. Long-range PCR was originally designed to detect changes in gene copy number rather than translocations or inversions, requires small amounts of DNA and is excellent for locus-specific identification of known rearrangements. These features make it ideal for diagnostic purposes. However, this technique is limited by its low throughput and is unable to provide a genome-wide view of rearrangements, which therefore restricts its usefulness to the analysis of a specific genomic region delimited by the primers that are used (Vasickova et al., 2007;Morozova and Marra, 2008).

Fluorescent in situ hybridization (FISH)
FISH is based on the hybridization of fluorescent probes to metaphase or interphase nuclei followed by analysis with a fluorescence microscope. FISH can detect variations in copy number (deletions and duplications), translocations and inversions. Copy number is assessed by microscopic visualization. The most commonly used conventional in situ hybridization protocol in cancer research is dual-color FISH. This method involves labeling centromeres and the DNA region of interest with different colors and estimating the probe copy number from the ratio of the centromeric to noncentromeric signal. Dual-color FISH is used to detect chromosomal gains or losses (aneuploidy), intrachromosomal insertions, deletions, inversions, amplifications and chromosomal translocations. The advantages of FISH include the ability to analyze single cells, applicability to a wide range of substrates, including fixed samples (such as paraffin-embedded tissue), and relative simplicity of use. The method cannot provide a genome-wide assessment of DNA rearrangements, with the exception of gross chromosomal aberrations detected by multifluor-based techniques, and is thus of limited value for genome-wide identification of smaller-scale chromosomal aberrations (De Lellis et al., 2007;Morozova and Marra, 2008).

Quantitative multiplex PCR of short fluorescent fragments (QMPSF)
QMPSF is a sensitive method for the detection of genomic deletions or duplications based on the simultaneous amplification of short genomic fragments using dye-labelled primers under quantitative conditions. The PCR products are analyzed on a sequencing platform used in the fragment analysis mode and the peak height and area are proportional to the quantity of template present for each target sequence. In this setting, the height or area of peaks corresponding to the loss of one allele will be half that of normal samples, whereas a gain of one allele will result in a 50% increase. This method is rapid and sensitive and has been used to screen for BRCA1 rearrangements (Casilli et al., 2002;Bastard et al., 2007;Weitzel et al., 2007). However, it is not easily implemented in a routine mutation analysis laboratory and requires a fair amount of previous experience.

Protein truncation test (PTT)
The PTT method is a straightforward approach to screen for biologically relevant gene mutations. The method is based on the size analysis of products resulting from transcription and translation in vitro. Proteins of lower mass than the expected full-length protein represent translation products derived from truncating frameshift or non-sense mutations in the analyzed gene. Mutation detection may be limited by the size and location of the rearrangement in relation to the primers used in the assay. In addition, because of the low sensitivity of conventional PTT, mutations can be detected only in samples that harbor a relatively high number of mutated gene copies (Peelen et al., 2000;Hauss and Müller, 2007).

Comparative genomic hybridization (CGH)
CGH (also known as chromosomal microarray analysis or CMA) is a molecular-cytogenetic method that has been used to analyze variations in copy number (gains or losses) of DNA from patients and/or tumor cells. The method is based on the hybridization of fluorescently labeled tumor DNA and normal DNA to normal human metaphase preparations. Using epifluorescence microscopy and quantitative image analysis, regional differences in the fluorescence ratio of gains/losses vs. control DNA can be detected and used to identify abnormal regions in the genome. CGH does not identify structural chromosomal aberrations such as balanced reciprocal translocations or inversions since they do not change the copy number. Although CGH is a complex technique that requires significant previous experience in cytogenetics and a specific set-up in terms of infra-structure, it is an efficient method for genome-wide screening of rearrangements (Rouleau et al., 2007).

Real time polymerase chain reaction (qPCR)
Real time PCR, also known as quantitative real time polymerase chain reaction (qPCR), is a polymerase chain reaction-based technique used to amplify and simultaneously quantify a target DNA molecule. qPCR allows the detection and quantification (as absolute number of copies or relative amount when normalized to DNA input or additional normalizing genes) of a specific sequence in a DNA sample. The procedure follows the general principle of PCR, the key difference being that the amplified DNA is quantified as it accumulates in the reaction in real time after each amplification cycle. Two common methods of quantification are the use of fluorescent dyes that intercalate with double-stranded DNA, and modified DNA oligonucleotide probes that fluoresce when hybridized with a complementary DNA. Although this method is rapid and does not require a large amount of starting material, it has a limited throughput. It is not suitable for the detection of translocations or inversions or for genome-wide screening of rearrangements Morozova and Marra, 2008).

Multiplex ligation-dependent probe amplification (MLPA)
MLPA is a multiplex PCR method developed to detect abnormal copy numbers of different genomic DNA sequences. Each MLPA probe consists of two oligonu-cleotides that can be ligated to each other when hybridized to a target sequence. All ligated probes have identical sequences at their 5' and 3' ends, permitting simultaneous amplification in a PCR containing only one primer pair. One of the two oligonucleotides of each MLPA probe has a common sequence used for PCR amplification at the 5' end and a target-specific sequence at the 3' end. The 5' region of the second oligonucleotide of each probe is designed to hybridise to the target sequence immediately adjacent to the first oligonucleotide and its 3' region has a common sequence used for PCR amplification and a "stuffer" sequence with different a specific length. Each probe gives rise to an amplification product of unique size, due to the variation in the stuffer sequence length. Because only ligated probes will be exponentially amplified during the subsequent PCR reactions the number of probe ligation products is a measure for the number of target sequences in the sample. The amplification products of different sizes are separated using capillary electrophoresis (Schouten et al., 2002). Nevertheless, MLPA has certain drawbacks, including false-negative scores when probes are designed outside the region of interest, i.e., outside the region involved in the rearrangement. MLPA is primarily used as a screening tool to identify rearrangements, and the precise location of the deletion or duplication breakpoints in the usually very large intronic or affected flanking regions must be refined by sequencing (Staaf et al., 2008). In addi-tion, in rare cases, MLPA may give a false-positive result for a deletion due to occurrence of a point mutation within the sequence of MLPA probe hibridisation (Gomez et al., 2009). However, compared to most other techniques, MLPA is an inexpensive, sensitive, relatively simple, and high-throughput method (Hogervorst et al., 2003;Dunnen and White, 2006;Ratajska et al., 2008). The use of MLPA has facilitated the screening of genomic rearrangements in BRCA1 (Montagna et al., 2003;Hartmann et al., 2004) and BRCA2 (Woodward et al., 2005).

Conclusion
Point mutations in the BRCA genes are the most common deleterious mutations encountered in HBOC families, and full gene sequencing and other PCR-based methods remain the gold standard for initial mutation identification. However, rearrangements in these genes have been described in a significant proportion of HBOC families, and are responsible for up to one-third of the identifiable BRCA mutations in certain populations. Consequently, in HBOC families that test negative for BRCA point mutations by conventional approaches, screening for large gene rearrangements in BRCA1 and probably also BRCA2 should be strongly considered. A suggested flowchart for investigation in these families is presented in Figure 2. The availability of relatively inexpensive and technically straightfor-Rearrangements in BRCA genes 443 ward screening methods has greatly simplified this process, but often more than one method must be used to fully characterize a deletion or duplication in a given patient. Several studies in different populations have proven the usefulness of screening for BRCA1 rearrangements, however the prevalence of such mutations in a given population should be known before definitive recommendations are made regarding the routine testing for rearrangements. In populations where there are highly prevalent founder rearrangements, preliminary screening for pathogenic BRCA gene mutations may be a cost-effective initial strategy.

Acknowledgments
This study was partly supported by grants from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, grant no. 477990/2006-1), Fundação de Apoio à Pesquisa do Hospital de Clínicas de Porto Alegre (FIPE, grant no. 04-081) and Susan G Komen for the Cure (POP 0403033). IPE was supported by a grant from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), EIP and SLC were supported by grants from CNPq and PILR was supported by a grant from Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS).