Introduction

The breast cancer susceptibility gene BRCA2 was first identified by positional cloning in 1995.1 Subsequent family studies confirmed that BRCA2 fit into the tumour suppressor gene classical model,2,3 with a role in the maintenance of DNA integrity.4,5 When initially isolated, the human BRCA2 transcript was shown to consist of 27 exons, coding for 3418 amino acids.1 Naturally occurring variants of BRCA2 having altered regulatory activity have also been identified.6,7 The presence of these variants has complicated the interpretation of diagnostic tests that rely on the detection of protein species. Here, we report new BRCA2 isoforms first detected in leucocyte RNA from patients at high risk for a germ-line BRCA mutation.

Materials and methods

Cells, cell lines and tissues

Patient peripheral blood samples were obtained from individuals with early-onset breast cancer and a family history of breast cancer. Normal control blood and tissue samples were obtained from anonymous, unrelated healthy individuals. Epithelial ovarian tumours were flash-frozen, anonymous, archived specimens. Cell lines were obtained from the American Type Culture Collection.

RNA and RT/PCR

RNA was isolated using RNeasy (Qiagen, Valencia, CA, USA) and Trizol (Invitrogen Life Technologies, Carlsbad, CA, USA). Reverse transcription was performed with Superscript II Reverse Transcriptase (Invitrogen Life Technologies) using random hexamers, oligodT primers and 2 μg of DNAse I-treated, total RNA as template. The following primers were used: Exons 18–23 of BRCA2 forward: p3274, 5′-102593CTTGTTCTCTGTGTTTCTGAC-3′ or p3586: 5′-102000GAAAGGGTGCTTCTTCAAC-3′. (Note: A T7 promoter and translation/initiation sequence was added 5′ to the unique primer sequence for the purpose of protein truncation test (PTT)). Exons 18–23 of BRCA2 reverse: p3293, 5′-119209TTTTTGTCGCTGCTAACTGTA-3′ or p3472, 5′119366-TCCCGTGGCTGGTAAATCTGA-3′.

PCR conditions were as follows: 95°C 3 min, 5 × (95°C 60 s, 50°C 90 s, 72°C 90 s), 35 × (95°C 60 s, 47°C 90 s, 72°C 90 s), and one cycle of 72°C 10 min.

Seminested amplification of exon 20A-containing isoforms was performed on amplicons of exons 18–23 using reverse primer p3293, with an exon 20A-specific forward primer p850: 5′-114673CACTGAAGTCTTGAACCCCC-3′.

PCR conditions were as follows: 95°C 3 min, 35 × (95°C 60 s, 55°C 90 s, 72°C 90 s) and one cycle of 72°C 10 min.

Note: Nucleotide position of primers is based on PAC HS214K23 (gb Z74739).8

PTT analysis

A 3.75 μl PCR product aliquot was combined with [35S]methionine and the TNT/T7 coupled wheat germ system (Promega, Madison, WI, USA) according to the manufacturer's instructions. The protein products were separated on a 10% SDS-polyacrylamide minigel (PAGE) (BioRad, Hercules, CA, USA). The gels were fixed in 20% methanol/10% acetic acid, and dried at 80°C for 1 h followed by autoradiography.

Subcloning and sequencing of PCR products

DNA products from the amplification of exons 18–23 were cloned into PCR2.1 (Invitrogen Life Technologies) and end sequenced using universal primers. Alternatively, an M13-tailed reverse primer was used in the amplification of noncloned cDNA from the same region (exons 18–21) and sequenced using an M13 universal primer by automated sequence analysis (Applied Biosystems, Foster City, CA, USA; Visible Genetics, Toronto, ON, USA).

Results

Identification of new BRCA2 isoforms

A research project employing PTT to scan for BRCA2 mutations in sporadic ovarian tumours uncovered a consistent difference between ovarian tumour and leucocyte PTT patterns produced from BRCA2 exons 18–23. Ovarian tumour PTT gels had a prominent band in the expected size range for the cDNA amplicon, at approximately 46.9 kDa (Fragment A, Figure 1). However, leucocyte samples from patients at high risk for familial breast cancer displayed a second prominent PTT band of approximately 28.3 kDa (Fragment B, Figure 1). To test whether Fragment B was a product of leucocyte specific alternative splicing, the PCR products obtained from amplification of exons 18–23 from the leucocytes of a breast cancer patient were subcloned and sequenced. Several BRCA2 isoforms were identified, each containing a 64 bp in-frame insertion between exons 20 and 21. These were collectively called BRCA2_v20 isoforms.

Figure 1
figure 1

PTT analysis of exons 18–23 of BRCA2. SDS-PAGE gels showing PTT pattern obtained from leucocytes and ovarian tumour samples.

Full size image

A BLAST search revealed that the insertion was a cryptic exon derived from BRCA2 IVS 20 (nucleotides 114 658–114 721, PAC HS214K238). This exon, which we called exon 20A, encoded 21 amino acids ending with a nonsense codon at the junction with exon 21. The sequence of the insertion was 8632G_ TTACTTCCTCCACTGAAGTCTTGAACC CCCCAA AGTCATCCATGAGGGTTGGAATCAACTTCTG _A8633, translating to amino acid sequence: VTSSTEVLNPPKSSMRV GINFX.

In addition, two isoforms eliminated exon 22 by utilizing a cryptic splice site in exon 23 at nucleotide position 9232. All isoforms were predicted to produce the same truncated protein lacking the amino acids encoded by the last seven exons of the gene.

Figure 2 shows the BRCA2_v20 variants. RNA species ‘A’ conformed to the published cDNA sequence for exons 18–23 of BRCA2. Isoforms v20B–v20E all contained exon 20A, but showed alternative splicing of exons 22 and 23. The PTT fragment obtained from primer set p3274/p3293 was predicted to produce a protein product of 345 amino acids (species ‘A’). The theoretical size of the product produced by the alternative isoforms was 213 amino acids, resulting in a size ratio of 0.62:1 for the alternative isoforms relative to species ‘A’. This was in agreement with the observed 0.60:1 size ratio of PTT Fragment B relative to PTT Fragment A (Figure 1).

Figure 2
figure 2

Diagram showing isoform structure. The numbers above the boxes indicate the individual exons and the shaded boxes represent regions eliminated via alternative splicing.

Full size image

Expression of BRCA2 isoforms

In order to determine if the isoforms were expressed in leucocytes from normal individuals, seminested PCR using an internal forward primer (p850) designed to amplify exon 20A-containing isoforms only was performed on exon 18–23 cDNA amplicons from 19 healthy individuals. First round PCR produced a blurred band in the 1100 bp range, as expected, since the various BRCA2 amplicons range in size between 889 and 1139 bp (not shown). Seminested PCR produced several electrophoretic bands in all lanes, corresponding in size to each of the BRCA2_v20 isoforms (Figure 3). The most prevalent band corresponded in size to the v20D isoform. Similar analyses using total RNA from normal breast and ovarian tissues, ovarian tumours and tumour cell lines (16 examined in total) demonstrated a single sharp band, corresponding in size to species A in the first round (data not shown). This agreed with the PTT results obtained from ovarian tissues that showed a single prominent band (Figure 1). Second round PCR using the exon 20A-specific primer set, which excluded species A, revealed a single major band consistent in size with isoform v20B (514 bp) in all lanes (Figure 3).

Figure 3
figure 3

Seminested RT-PCR amplification of exons 18–23. Left: Examples of second round PCR amplification of leucocytes from normal healthy controls using primers specific for the v20 isoforms: v20B: 514 bp; v20C, 463 bp; v20D, 315 bp (predominant); v20E, 264 bp. Right: Examples of second round PCR amplification of tissue from breast and ovarian samples expressing v20B only. M – 123 bp DNA ladder; D – cloned v20D (315 bp).

Full size image

Detection of overexpressed exon 20A in breast cancer patients

In two independent, collaborating clinical laboratories, mutational analysis of BRCA2 was performed routinely by PTT. The PTT pattern for the region containing exons 18–23 showed a secondary band consistent with the band seen in the leucocyte samples used in the ovarian tumour experiments. However, several unrelated patients appeared to have a stronger secondary band in the same position, interpreted to be a possible truncation. Electrophoresis of the PCR products revealed a second prominent PCR product of lower molecular weight, suggesting a splicing defect. Sequencing of the cDNA from these patients revealed a 64 bp insertion between exons 20 and 21 identical to that cloned in the research laboratory. In these cases, genomic sequencing failed to reveal any mutations in the region.

A review of the clinical leucocyte specimens revealed that PTT Fragment B was strongest in blood samples that had been stored at room temperature over a period of at least 24 h. Fresh samples were obtained from the patients for repeat analysis by PTT. PTT experiments performed using fresh RNA resulted in a diminished Fragment B, confirming the suspicion that the previous samples had been affected by conditional alternative splicing.

Discussion

While nonsense-mediated mRNA decay may cause false-negative results in PTT experiments,9 false-positive results due to illegitimate splicing have not been previously described as a possible pitfall of PTT. In our series of PTT analyses, the BRCA2 splice variants were observed in both patient and control tissues. Also, no genomic mutations that might explain the alternative splicing were found in any of the samples exhibiting increased isoform expression. Since overexpression of the v20 variants was most evident in aged blood samples, it likely represented a response to environmental stress. At least three distinct illegitimate splice forms arising from aged blood samples have also been found in the 3′ end of BRCA1 (data not shown). Therefore, the possibility of false-positive PTT results for the 3′ region of BRCA2 as well as BRCA1 must be considered when interpreting test results.

The prominence of frameshift and premature stop codons in BRCA2 mutation families suggests that BRCA2 truncating mutations are important in breast cancer susceptibility. Also, the involvement of the 3′ terminus of BRCA2 in DNA repair pathways and protein localization has been proven experimentally.10,11,12 Differences in the range of expression of individual v20 splice variants in breast- and ovarian-derived tissues as compared to leucocytes suggested physiological tissue specificity and not just illegitimate splicing alone. It was therefore important to evaluate the expression patterns of the newly discovered alternative BRCA2 isoforms to determine if they were of aetiologic significance to high-risk breast cancer patients. Heterogeneous expression of BRCA2, as described here, appears to be similar to the temperature-dependent expression and tissue-based variability seen in the NF1 message.13,14,15 Tumour suppressor genes TP53 and APC also appear to undergo similar, unexplained, truncating alternative splicing.16,17 Acquired defects in pre-mRNA processing have been proposed as a factor in causing and predisposing individuals to disease.18 Since tumour suppressor genes are acutely susceptible to inactivation through truncation events, it is possible that epigenetic or environmental factors leading to increased expression of truncating isoforms in sensitive tissues impact on disease progression.

When characterizing the expression of tumour suppressor genes in vitro or when evaluating patient samples for expression abnormalities, it is important to rule out conditional alternative splicing as a possible mechanism of abnormal expression. An increased understanding of the conditions leading to alternative splicing in clinical samples is also extremely important so that compromised interpretation of RNA-based analyses is avoided.