This article is mentioned in:

Introduction

Inactivating mutations in the BRCA1 (MIM 113705) and BRCA2 (MIM 600185) genes confer a high risk of developing breast and ovarian cancer (1,2). Both genes have a contribution of approximately 16% to the risk of familial breast cancer (3). Genetic testing for BRCA1 and BRCA2 provides valuable information for determining the clinical management of patients with breast/ovarian cancer. However, the data provided are difficult to interpret due to the identification of many DNA variants of unknown pathological significance or unclassified variants (UVs) that hamper genetic counseling in hereditary breast and ovarian cancer (HBOC) (4).

UVs have the potential to alter protein function by altering the coding sequence of a transcript, or the level of the gene transcript by disrupting regulatory regions in promoters, untranslated regions, exons or introns (5). Such regulatory variants include those affecting the normal splicing of BRCA1 and BRCA2, many of which have been shown to be clinically significant using cDNA studies and multifactorial likelihood analysis methods that combine bioinformatics, as well as pathological and clinical information (6,7).

Assessing the impact of UVs on splicing is key to determining their pathogenicity. The accuracy of pre-mRNA splicing is determined by the recognition of well known 5′ and 3′ splice site consensus sequences. However, more discrete elements are also involved, such as exonic splicing enhancers (ESEs) that enhance pre-mRNA splicing when present in exons (8). As a result, each UV may potentially affect normal pre-mRNA splicing and be deleterious through the disruption of consensus sequences, or through the creation of de novo sequences or the alteration of splicing regulatory elements (9). Several BRCA1 isoforms have been identified in different tissues; however, their functional significance is not yet fully understood (10–15).

A recent study on the comprehensive annotation of BRCA1 splice junctions identified 63 independent alternative splicing events in RNA samples from healthy control individuals. Among these, 10 were predominant (Δ1Aq, Δ5, Δ5q, Δ8p, Δ9, Δ9–10, Δ9-10-11, Δ11q, Δ13p and Δ14p) and represented 5–30% of the full-length signal; 48 were minor and 5 were non-classifiable events (16).

Alternative splicing is a highly coordinated process. Mutations destroying the 5′ and 3′ splice site consensus sequences may alter the splicing patterns of one or more transcripts, interrupting the production or function of the encoded protein (17). A large number of pathological transcripts of the BRCA1 and BRCA2 genes, deriving from a mutation in the consensus regions, have been described in the literature (18).

The aim of this study was to identify aberrant transcript variants resulting from the alternative splicing of BRCA1 and BRCA2 genes in RNA extracted from blood lymphocytes from women with a family history of and/or early onset breast and/or ovarian cancer, in which genomic pathogenic alterations in BRCA1 and BRCA2 have not been detected by conventional analysis.

The analysis of all possible transcripts of the BRCA1/BRCA2 genes may allow us to uncover mRNA splicing defaults overlooked by conventional protocols and to confirm that the alternative splicing of the BRCA1 and BRCA2 genes plays an important role in BRCA1/2-driven tumorigenesis. This approach has the potential to complete the process of the characterization of mutations of BRCA1 and BRCA2 in HBOC.

Materials and methods

Sample acquisition

A total of 13 blood samples were collected from women with a family history of and/or early-onset breast and/or ovarian cancer and who tested negative for pathogenic mutations in the BRCA1 and BRCA2 genes. Sample data details and characteristics are presented in Table I.

Table I Summary of patient data and family history of cancer.

Table I

Summary of patient data and family history of cancer.

Patient	Personal history (age at onset, years)	Family history
		No. of breast cancer cases	No. of ovarian cancer cases	No. of cases of other types of cancer
P1	Br (57)	4	–	1
P2	Br (34)	1	–	1
P3	Br (43)	8	1	5
P4	Br (37)	1	1	4
P5	Br (55), Ov (61)	–	–	2
P6	Br (35)	3	–	1
P7	Br (55)	3	1	4
P8	Br bil (43)	4	1	3
P9	Br (56)	3	–	2
P10	Br bil (45 and 50)	4	–	9
P11	Br (25)	–	–	1
P12	Br (32)	–	–	1
P13	Br bil (38 and 48)	–	–	6

[i] Br, breast cancer; Ov, ovarian cancer; bil, bilateral.

The BRCAPRO (http://www4.utsouthwestern.edu/breast-health/cagene/) and BOADICEA (https://pluto.srl.cam.ac.uk/cgi-bin/bd2/v2/bd.cgi) (19,20) programs were used, calculating a priori mutation carrier risk of >10% for each patient.

Ten blood samples from healthy women, aged 25–45 years, with no family history of breast and/or ovarian cancer, were used as the controls. All patients and healthy donors were recruited at the Interdepartmental Centre for Cancer Genetics, University Hospital of Santa Chiara in Pisa, Italy between 2004 and 2011. Prior to enrollment, informed consent for genetic analysis was obtained from all patients. All the experiments carried out complied with the current laws of the country in which they were performed (Italy).

RNA extraction and cDNA synthesis

All samples were subjected to total RNA extraction in 2 steps: peripheral blood mononuclear cells (PBMCs) were isolated from ethylenediaminetetraacetic acid (EDTA)-treated peripheral blood by standard density gradient centrifugation (Ficoll 15%; Bio-Rad Medical Diagnostics GmbH, Dreieich, Germany) and RNA extraction was obtained using a TRI Reagent kit (Molecular Research Center, Inc., Cincinnati, OH, USA) as recommended by the manufacturer. Isolated RNA was quantified using a NanoDrop spectrophotometer (NanoDrop Technologies, Inc., Wilmington, DE, USA) and examined for integrity on a 1.5% agarose/formaldehyde gel containing 0.5 μg/ml ethidium bromide (RNA Analysis Notebook; Promega Corp., Madison, WI, USA). First-strand cDNA was synthesized from at least 1,000 ng of total RNA using an oligo(dT) primer or random primer and SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions.

Reverse transcription-polymerase chain reaction (RT-PCR)

To perform the amplification of naturally occurring transcripts of the BRCA1 and BRCA2 genes, we used multiple combinations of forward and reverse primer pairs to amplify overlapping regions of the mRNA and to cover the entire open reading frame (Figs. 1 and 2). For each PCR reaction, 2 μl of cDNA were used. The PCR conditions were as follows: 5 min at 95°C followed by 4–45 cycles at 95°C for 1 min, melting temperature according to primer pair for 30 sec, 72°C for 1 min followed by 72°C for 1 min. To encompass multiple exons for large regions, Long Range PCR (Expand Long Template PCR system; Roche, Indianapolis, IN, USA) was used. The PCR products, eventually isolated on agarose gels, were sequenced on both strands using the BigDye Terminator v3.1 Cycle Sequencing kit and the 3130×l Genetic Analyzer (both from Life Technologies, Foster City, CA, USA). The electropherograms were analyzed using SeqScape Software v2.6 (Life Technologies).

Figure 1 Primers used for the scanning of exons in the BRCA1 gene. Primer pairs were used to amplify the cDNA covering the entire coding region of the BRCA1 gene overlapping the exonic region.

Figure 2 Primers used for the scanning of exons in the BRCA2 gene. Primer pairs were used to amplify the cDNA covering the entire coding region coding of the BRCA2 gene overlapping the exonic region.

Multiplex ligation-dependent probe amplification (MLPA)

To exclude large genomic deletions in the BRCA1 and BRCA2 genes, MLPA (MRC-Holland, Amsterdam, The Netherlands) was performed as recommended by the manufacturer’s instructions. The electropherograms were analyzed using GeneScan (Life Technologies) and Coffalyser software (MRC-Holland).

Results

The aim of this study was to identify alternative transcripts of the BRCA1 and BRCA2 genes resulting from aberrant splicing events. The analysis was conducted on 13 total RNA samples extracted from PBMCs from women with a family history of and/or early-onset breast and/or ovarian cancer, more specifically, 5 cases of hereditary breast cancer (HBC), 4 cases of HBOC, 1 case of bilateral carcinoma, 1 case of breast/ovarian cancer and 2 cases of early-onset breast cancer (Table I.)

All women were affected by breast cancer, 1 women by breast and ovarian cancer, 3 by bilateral breast cancer and 2 by early-onset breast cancer. A total of 10 control RNA samples from healthy women, aged between 25 and 45 years with no family history of any form of cancer, were also included in this study.

The genomic DNA of each patient was analyzed by direct sequencing of the entire open reading frame, 5′ and 3′UTRs and exon/intron junctions of both genes. Approximately 100 bps from the 5′ and 3′ end of each intron were sequenced. All cases tested negative for the presence of germline mutations in the BRCA1 and BRCA2 genes. The presence of variants of unknown pathological significance was also excluded from our analysis. The presence of large genomic deletions or rearrangements was excluded by MLPA of DNA extracted from the peripheral blood lymphocytes of all patients and the controls.

The BRCA1 and BRCA2 mRNA in each patient was analyzed by dividing the cDNA into 9 amplicons for BRCA1 and 10 amplicons for BRCA2. The size of the full-length transcripts of the BRCA1 (5,592 bp) and BRCA2 (10,987 bp) genes prevents the analysis of the cDNA as a single amplicon.

Naturally occurring transcripts of BRCA1

The cDNA of the BRCA1 gene was amplified with primers localized in exonic sequences so that partially overlapping amplified products were obtained. The primers were selected in order to highlight all predominant naturally occurring alternative splicing isoforms and 28 of the minor transcripts, starting from exon 2, as previously reported by Colombo et al (16).

Under our experimental conditions, the following naturally occurring alternative transcripts were observed: the Δ9–10 transcript was detected in 3 samples, the Δ9 transcript was detected in 1 patient and Δ14p was detected in all the samples. Using the long range PCR approach, the following transcripts were detected: Δ11q (transcript lacking exon 11 except the initial 121 bp) in 10 samples and Δ9-10-11q (transcript lacking exons 9, 10 and 11 except the initial 121 bp) in 2 cases only.

Naturally occurring transcripts of BRCA2

The cDNA of the BRCA2 gene was amplified with primers localized in exonic sequences so that partially overlapping amplified products were obtained. A total of 5 predominant naturally occurring transcripts have been previously identified: Δ4, Δ4–7, Δ17–18, Δ18 and Δ20 (7,18,21). Under our experimental conditions, all 5 alternative transcripts were detected: More specifically, the Δ4 transcript was detected in 2 patients, the Δ4–7 transcript was detected in 2 patients, the Δ17–18 transcript was detected in 2 patients and the Δ18 transcript was detected in 7 patients. The Δ20 transcript was also detected in 1 case. All these BRCA1 and BRCA2 naturally occurring transcripts were detected in the controls.

Abberant transcripts

In addition to the above-mentioned predominant naturally occurring transcripts, we detected 3 aberrant transcripts in the BRCA1 gene in 2 patients (2 in patient P6 and 1 in patient P7). No aberrant transcripts were detected in the BRCA2 gene.

In patient P6, following the amplification of the cDNA of BRCA1 with primers localized in exons 16 and 21 in addition to the full-length transcript (600 bp), 2 additional transcripts of approximately 500 and 400 bp were obtained. These transcripts were not detected in the 10 healthy control samples. The aberrant transcripts were gel-purified and then sequenced as described in the Materials and methods. Sequence analysis allowed us to detect the presence of a transcript lacking exon 17 and another transcript lacking exons 17, 18 and 19 (Fig. 3). The transcript containing the deletion of exon 17 produced an abnormal stop signal at codon 1673 (HGVS codification: p.Val1665Serfs*8) and then a truncated protein lacking the last 192 amino acids. The transcript containing the deletion of exons 17, 18 and 19 did not produce an abnormal stop signal, lost 207 nucleotides and retained the open reading frame, producing a protein lacking 69 amino acids.

Figure 3 Alternative transcript isoforms in patient P6. Following the amplification of the cDNA of BRCA1 with primers localized in exons 16 21, the wild-type (wt) transcript (~600 bp) and 2 additional bands of (A) ~500 bp and (B) ~400 bp were obtained. Sequencing analysis allowed us to detect the presence of a transcript lacking (A) exon 17 and another transcript lacking (B) exons 17, 18 and 19.

In patient P7, following the amplification of the cDNA of the BRCA1 gene with primers localized in exon 13 and 16 in addition to the full-length transcript of approximately 420 bp, an additional transcript of approximately 220 bp was obtained (Fig. 4). Direct sequencing of this transcript allowed us to detect an exon 15 deletion. The transcript containing the deletion of exon 15 produced an abnormal stop signal at codon 1510 (HGVS codification: p.Ser1496Glyfs*14) and then a truncated protein lackcing the last 405 amino acids.

Figure 4 Alternative transcript isoform in patient P7. Following the amplification of the cDNA of BRCA1 with primers localized in exons 13 and 16 with primers 13F and 16R, an additional band of ~420 bp was obtained. The fragment had a deletion of exon 15 as detected by direct sequencing.

Phenotype-genotype correlation

Patient P6, at the age of 35 years, developed an infiltrating ductal carcinoma of the right breast, which was estrogen receptor-positive, progesterone receptor-negative and Her2/neu 3+, with loco-regional lymph node metastasis. The patient’s sister had been diagnosed with breast cancer of a similar phenotype at the age of 34. These were not the only cases of breast cancer in the family; the grandmother and maternal aunt (their mother’s twin sister) had also been afflicted by the disease. Segregation analysis of the aberrant transcript in this family was not possible as DNA samples were not available.

Patient P7 was diagnosed with breast cancer at the age of 55 years. The tumor was an infiltrating breast carcinoma of the right breast, and was estrogen receptor-positive, progesterone receptor-negative and Her2/neu-negative, without any lymph node metastasis. The patient’s mother had been affected by ovarian cancer at 77 years of age and 3 cases of breast cancer were reported in the family: 2 sisters of the mother and the mother’s cousin (36 years of age). On the mother’s side, an uncle had been affected by malignant melanoma, and an aunt by a brain tumor. These relatives were not available for co-segregation analysis, but this aberrant transcript was not detected in the 42-year-old healthy daughter.

Discussion

This study focused on the possibility that a proportion of patients with HBOC have mutations in the BRCA1 or BRCA2 genes that affect splicing, but are not detectable through sequencing of the gene exons or intronic sequences near the intron-exon boundaries. Several studies have demonstrated that, due to variations in splice sites, BRCA1/BRCA2 may generate truncated, non-functional proteins that may be associated with a predisposition to breast and ovarian cancer (22). However, to the best of our knowledge, no studies have evaluated the presence of BRCA1/BRCA2 pathological transcripts in patients without mutations identified in the canonical splice sites or regulatory sequences.

Recently, a systematic description of ‘naturally occurring’ alternative splicing at the BRCA1 locus was conducted by the Evidence-based Network for the Interpretation of Germline Mutant Alleles (ENIGMA) consortium (16). This led to the annotation of 63 splicing events, of which 35 were novel findings, even though most of them are rather minor, and it is likely that some do not qualify as ‘naturally occurring’ events, suggesting that the characterization of the full complexity of BRCA1 splicing requires further investigation (16).

The present study was performed in accordance with a standard assay design and detection methods formulated by ENIGMA Consortium members (23) in order to detect the predominant naturally occurring alternative splicing isoforms of BRCA1, as reported by Orban et al (13,14), which were the only data available at the moment of the study design. There is no standard assay design for the detection of the predominant naturally occurring alternative splicing isoforms of BRCA2; thus we referred to the transcripts described in Ensembl (http://www.ensembl.org/index.html).

Using this strategy, we detected the Δ9, Δ9–10, Δ11q, Δ9-10-11q and Δ14p BRCA1 isoforms and the Δ4, Δ4–7, Δ17–18, Δ18 and Δ20 BRCA2 isoforms. In addition these predominant BRCA1 transcripts, we detected the Δ17 and the Δ17–19 isoforms in patient P6 and the Δ15 isoform in patient P7. No minor transcripts were detected for BRCA2.

The Δ17 transcript lost the open reading frame, leading to the formation of a truncated protein. The Δ17 transcript was not detected in any of the 10 healthy controls in our study; however, it has previously been found in 28 out of 103 (27%) independent control samples or technical replicas in the ENIGMA Consortium study (16). The Δ17 transcript, due to the genomic deletion of exon 17, has frequently been detected in Italian families (24) possibly as a result of a non-homologous rearrangement among ‘Alu repeats’. The Δ17 transcript is not derived from a genomic rearrangement as demonstrated by MLPA of the patient’s DNA. The Δ17 transcript, even if not derived from a genomic rearrangement or from aberrant mRNA maturation due to canonical splice site mutations, it has to be considered as pathogenetic as it encodes for a truncated protein.

The Δ17–19 transcript gave rise to a protein with an in-frame deletion of 69 amino acids. The Δ17–19 isoform was not detected in either our 10 healthy control samples or the 223 independent control samples and technical replicas in the ENIGMA Consortium study (16). Although the Δ17–19 transcript had an in-frame deletion, a recent study demonstrated that the expression of the human BRCA1 alternative splicing variant, Δ17–19, in the MCF-7 cells resulted in an impaired assembly of DNA repair complexes and aberrant DNA damage response (25), thus contributing to an increased risk of developing cancer in the carrier subjects.

In patient P7, the BRCA1 alternative transcript Δ15 was detected. The alternative Δ15 transcript lost the open reading frame and produced an abnormal stop signal to position p.Ser1510Glyfs*14, leading to a truncated protein lacking the last 405 amino acids. To evaluate the presence of Δ15 at the genomic level, we performed MLPA, which did not reveal any deletion. This isoform was not detected in our 10 healthy PBMC control samples or in the 8 PBMC samples analyzed by Colombo et al (16). They found this transcript in only 4 of 10 replicas derived from samples of leukocytes, PHA-stimulated peripheral blood leukocytes and lymphoblastoid cell lines (they did not specify if these were independent samples or replicas of the same sample).

In conclusion, the present study on 13 patients with a family history of breast and/or ovarian cancer detected 3 alternative transcripts, probably pathogenic and not predictable by genomic screening, in 2 patients. A number of studies have described these 3 transcripts as consequences of large genomic rearrangements or mutations in canonical splice sites (24–28). These aberrant alternative transcripts may undergo nonsense-mediated decay (NMD) as they give rise to truncated proteins (29); therefore, the accurate quantification of these alternative transcripts is crucial to discriminating what is considered pathological and what is neutral for clinical relevance. The relative abundance of the minor transcripts detected in this study may prove to be useful in evaluating their possible role in cancer predisposition and risk modification.

Acknowledgments

The present study was funded by the Istituto Toscano Tumori (ITT) grant ‘BRCA1 role in repair of DNA damage: from the yeast Saccharomyces cerevisiae genetics to human breast carcinogenesis’ (2009–2012) and the Associazione Italiana per la Ricerca sul Cancro (AIRC) grant [no. IG2013 N.14477 (2013–2015)].

References