Germline nonsense and canonical splice site variants identified in disease-causing genes are generally considered as loss-of-function (LoF) alleles and classified as pathogenic. However, a fraction of such variants could maintain function through their impact on RNA splicing. To test this hypothesis, we used the alternatively spliced BRCA2 exon 12 (E12) as a model system because its in-frame skipping leads to a potentially functional protein. All E12 variants corresponding to putative LoF variants or predicted to alter splicing (n = 40) were selected from human variation databases and characterized for their impact on splicing in minigene assays and, when available, in patient lymphoblastoid cell lines. Moreover, a selection of variants was analyzed in a mouse embryonic stem cell–based functional assay. Using these complementary approaches, we demonstrate that a subset of variants, including nonsense variants, induced in-frame E12 skipping through the modification of splice sites or regulatory elements and, consequently, led to an internally deleted but partially functional protein. These data provide evidence, for the first time in a cancer-predisposition gene, that certain presumed null variants can retain function due to their impact on splicing. Further studies are required to estimate cancer risk associated with these hypomorphic variants. More generally, our findings highlight the need to exercise caution in the interpretation of putative LoF variants susceptible to induce in-frame splicing modifications.

Significance:

This study presents evidence that certain presumed loss-of-function variants in a cancer predisposition gene can retain function due to their direct impact on RNA splicing.

Germline nonsense variants account for 20% of all Mendelian disease–causing single-nucleotide variants (SNV) reported within gene coding sequences (1). This type of variants, as well as frameshift deletions/insertions and SNVs located at the highly conserved intronic dinucleotides of splice sites (hereafter termed IVS±1/2), are presumed to lead to loss-of-function (LoF) because they are expected to introduce a premature termination codon (PTC). On the basis of these assumptions, these variants are generally regarded as null alleles and classified as pathogenic without further investigation. This is in particular the case for the iconic cancer predisposition genes, BRCA1 and BRCA2, where mono-allelic germline dominant LoF variants confer high risks of hereditary breast and ovarian cancer (HBOC; ref. 2).

Several consortia that provide guidelines for variant classification in the clinical setting, including the American College of Medical Genetics and Genomics (ACMG; ref. 3) and the Evidence-based Network for the Interpretation of Germline Mutant Alleles (ENIGMA; refs. 4, 5), recommend caution with the systematic pathogenic classification of such presumed null variants. Indeed, possible rescue mechanisms have been described for (partial) function preservation of certain presumed LoF variants in BRCA1 and BRCA2 genes. In particular, the polymorphic BRCA2 c.9976A>T (p.Lys3326*) nonsense variant results in a truncated protein with a minor loss in the C-terminal region that preserves functionality (6–8) and is associated with a modest increase in HBOC risks compared with typical LoF pathogenic BRCA2 variants (9). A second rescue mechanism involves specific naturally occurring alternative splicing events, as recently illustrated by the characterization of BRCA1 c.[594-2A>C; 641A>G] (5). This allele is not associated with high cancer risk despite its induced out-of-frame exon 10 skipping, potentially because the naturally occurring in-frame alternative transcript without exons 9 and 10 (ΔE9_E10) can encode for a functioning BRCA1 protein. Naturally occurring alternative splicing can also enable survival of patients homozygous for certain BRCA1 exon 11 nonsense variants, albeit with severe features of Fanconi anemia (10). Here, we raised the hypothesis that a subset of presumed LoF variants in HBOC susceptibility genes may potentially retain partial activity through their direct impact on splicing, leading eventually to questioning their clinical interpretation.

Virtually all variants may potentially induce a perturbation in RNA splicing, notably through the modification of either 3′ and 5′ splice sites (3′/5′ ss) or exonic splicing regulatory elements (ESR; refs. 11, 12). The most frequently observed splicing anomaly is exon skipping that often leads to a frameshift, causing LoF. However, some specific presumed LoF variants could induce in-frame skipping of a nonessential exon, resulting in a functional protein, at least partially. Here, we envisioned that this rescue mechanism may be triggered by specific variants located in the exon 12 (E12) of BRCA2. This exon does not encode a known functional domain involved in the BRCA2 tumor suppressor activity and its in-frame skipping leads to the internal deletion of a potentially dispensable 32-amino acid sequence (13). In a mouse embryonic stem cell (mESC)-based functional assay, the E12-deleted BRCA2 protein is able to complement the loss of endogenous wild-type Brca2 and exhibits the same functions with regard to (i) homologous recombination activity, (ii) radiation induced RAD51 foci formation, and (iii) sensitivity to DNA-damaging agents (13).

An additional argument supporting the tumor suppressor functionality of the BRCA2 E12-deleted protein came from the initial analysis of the clinical significance of a missense variant, c.6853A>G (p.Ile2285Val), that triggers E12 skipping. The family history likelihood ratio and the cooccurrence in trans with a pathogenic variant support the classification of this spliceogenic variant as neutral (13).

In this study, we selected, from human variation databases, a total of 40 variants in or near BRCA2 E12 that were either presumed LoF variants or predicted to alter E12 splicing, and characterized their impact on splicing in minigene assays and patient biological material. Moreover, a selection of variants was tested in a mESC-based functional assay to assess their combined effects on RNA splicing and protein activity. To evaluate their association with cancer, clinical and family data were collected for spliceogenic BRCA2 E12 variants. Altogether, these data provide evidence that presumed LoF variants, including nonsense SNVs, can induce splicing alterations that trigger in-frame deletion of a sequence not essential for BRCA2 function. The existence of this potential rescue mechanism instigates the need to reevaluate the interpretation of certain BRCA2 variations currently considered as LoF.

Variant selection

All variants located in BRCA2 E12 and its flanking intronic sequences were extracted from human variation databases (Supplementary Fig. S1). First, we selected the SNVs generally regarded as LoF variants, including: (i) IVS±1/2, (ii) nonsense, and (iii) frameshift variants. Second, we extended this selection to all variants predicted to affect E12 splicing based on two types of bioinformatics approaches. Predictions of the variant effects on 3′/5′ ss were obtained by using the MaxEntScan (MES) in silico tool (14), interrogated from the integrated software Alamut Visual 2.9 (Interactive Biosoftware). Predictions of variant impacts on ESR were generated by using a bioinformatics approach relying on ΔtESRseq values (total ESRseq score changes; refs. 15–17), based on exonic hexamer scores. Variants that were associated with either a decrease by at least 15% of MES score (ΔMES ≤ −15%) or a ΔtESRseq score inferior to −0.7 (ΔtESRseq < –0.7) were retained for splicing minigene analysis.

Splicing minigene assay

To evaluate the impact of the selected variants on E12 splicing, we performed an in cellulo splicing assay based on the use of the two-exon minigene pCAS2 vector (17, 18). Wild-type (WT) and mutant segments encompassing BRCA2 E12 and its intronic flanking sequences were amplified from patient genomic DNA (gDNA) by PCR using specific primers (Supplementary Table S1) and inserted into the intron of the pCAS2 minigene using BamHI and MluI restriction sites, yielding pCAS2-BRCA2-E12 minigenes (Fig. 1A). When patient gDNA was not available, the variant of interest was introduced by site-directed mutagenesis. WT and mutant pCAS2-BRCA2-E12 minigenes were transfected in parallel into HeLa cells using the FuGENE 6 Transfection Reagent (Roche Applied Science). Total RNA was extracted 24 hours posttransfection using the NucleoSpin RNA Kit (Macherey-Nagel). The splicing patterns were analyzed by semiquantitative fluorescence RT-PCR (25 cycles) by using the OneStep RT-PCR Kit (Qiagen), in a 25 μL reaction volume, with 200 ng total RNA and specific minigene primers (Fig. 1A; Supplementary Table S1). RT-PCR products were separated by electrophoresis on 2% agarose gels. All RT-PCR DNA bands were gel-purified and sequenced. Because of the formation of heteroduplexes between the different RT-PCR products, the images of the gels were not used for quantification. For semiquantitative analysis, fluorescence RT-PCR products were resolved under denaturing conditions by capillary electrophoresis on an automated 3500 Genetic Analyzer (Applied Biosystems). Electropherograms were then analyzed by using the GeneMapper v5.0 Software (Applied Biosystems) and peak areas were used to quantify the relative levels of each transcript.

Figure 1.

Identification in minigene assays of BRCA2 exon 12 spliceogenic variants among presumed LoF SNVs. A, Schematic representation of the pCAS2-BRCA2-E12 minigene used in the splicing reporter assay. Boxes indicate exons and lines in between represent introns. The pCAS2-BRCA2-E12 minigenes were generated by inserting a genomic fragment encompassing BRCA2 exon 12 into the intron of pCAS2 by using BamHI and MluI restriction sites. Arrows above the exons A and B indicate the position of primers used in RT-PCR analysis. CMV, CMV promoter. B, Distribution of 15 presumed LoF variants (IVS±1/2; nonsense and frameshift) in BRCA2 exon 12. The figure shows the nucleotide sequence of BRCA2 exon 12 and part of its flanking intronic regions, as well as the positions of the 15 selected variants, including four intronic variants located at position IVS±1/2 (below the sequence), five nonsense (above the sequence, in bold), and six frameshift (below the sequence) variants. C, Impact of the 15 presumed LoF variants on BRCA2 exon 12 splicing. WT and mutant pCAS2-BRCA2-E12 minigene constructs were transiently expressed in HeLa cells. The splicing patterns of the RNA produced from the different minigenes were analyzed by RT-PCR as described in Materials and Methods. Images show agarose gel electrophoresis of the RT-PCR products. The identities of the two major RT-PCR products, with (+E12) or without (ΔE12) exon 12, are indicated on the left of the gel. The other isoform identified (a) corresponds to the inclusion of exon 12 plus the retention of the first 40 nucleotides of intron 12 [+E12q(ins40)]. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex between the different RNA isoforms (noted by the star symbol on the gel images), as revealed by band sequencing. Instead, the relative quantifications of the different RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S2), as indicated in Materials and Methods. The data shown represent the average of at least three independent experiments. Error bars, SD values. nt, nucleotide.

Figure 1.

Identification in minigene assays of BRCA2 exon 12 spliceogenic variants among presumed LoF SNVs. A, Schematic representation of the pCAS2-BRCA2-E12 minigene used in the splicing reporter assay. Boxes indicate exons and lines in between represent introns. The pCAS2-BRCA2-E12 minigenes were generated by inserting a genomic fragment encompassing BRCA2 exon 12 into the intron of pCAS2 by using BamHI and MluI restriction sites. Arrows above the exons A and B indicate the position of primers used in RT-PCR analysis. CMV, CMV promoter. B, Distribution of 15 presumed LoF variants (IVS±1/2; nonsense and frameshift) in BRCA2 exon 12. The figure shows the nucleotide sequence of BRCA2 exon 12 and part of its flanking intronic regions, as well as the positions of the 15 selected variants, including four intronic variants located at position IVS±1/2 (below the sequence), five nonsense (above the sequence, in bold), and six frameshift (below the sequence) variants. C, Impact of the 15 presumed LoF variants on BRCA2 exon 12 splicing. WT and mutant pCAS2-BRCA2-E12 minigene constructs were transiently expressed in HeLa cells. The splicing patterns of the RNA produced from the different minigenes were analyzed by RT-PCR as described in Materials and Methods. Images show agarose gel electrophoresis of the RT-PCR products. The identities of the two major RT-PCR products, with (+E12) or without (ΔE12) exon 12, are indicated on the left of the gel. The other isoform identified (a) corresponds to the inclusion of exon 12 plus the retention of the first 40 nucleotides of intron 12 [+E12q(ins40)]. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex between the different RNA isoforms (noted by the star symbol on the gel images), as revealed by band sequencing. Instead, the relative quantifications of the different RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S2), as indicated in Materials and Methods. The data shown represent the average of at least three independent experiments. Error bars, SD values. nt, nucleotide.

Close modal

Analysis of E12 splicing patterns and allele-specific expression in lymphoblastoid cell lines

The EBV-immortalized lymphoblastoid cell lines (LCL) of healthy donors and patients were collected in collaboration with French HBOC diagnostic laboratories within the Group “Genetic and Cancer” (GGC, Unicancer; Supplementary Data). These LCLs were cultured and treated or mock-treated with puromycin as described previously (17). After extraction of total RNA, the E12 splicing patterns were analyzed by semiquantitative fluorescence RT-PCR (26 cycles), as described above, with primers mapping to BRCA2 exons 11 and 14 (Supplementary Table S1).

To measure allele-specific expression (ASE) of E12-containing transcripts (+E12), a SNaPshot quantitative primer extension assay was performed as described previously (17), with the same RT-PCR conditions as described above and using specific primers targeting the variants (Supplementary Table S1). The ASEVariant was then calculated after normalization of the cDNA ratio to the corresponding gDNA ratio.

To extend these ASE analyses to both transcripts with (+E12) and without (ΔE12) E12, we set up a SNaPshot protocol targeting the c.7242A>G (p.Ser2414 =) polymorphism located in BRCA2 exon 14 (SNP rs1799955, minor allele frequency in gnomAD = 22.5%). Transcripts +E12 and ΔE12 were selectively amplified in parallel RT-PCR reactions (32 cycles) using a forward primer overlapping either exons 11/12 or 11/13, respectively, combined with a reverse primer in exon 15 (Supplementary Table S1). The SNaPshot quantitative primer extension assay was then performed using a primer targeting the sequence immediately upstream the SNP. The ASESNP was then calculated as described above.

mESC–based assay

The mESC-based assay (8, 19) was used to assess the functional impact of BRCA2 E12 variants. The mESC expressing the human BRCA2 gene (hBRCA2) with specific variants were generated as described previously (20). For each transfection round a new ampule of the parental mES cell line was used and cell lines were checked for the presence of Mycoplasma monthly. After Cre-mediated deletion of the conditional mouse Brca2 allele, cell survival, BRCA2 protein expression, and homology-directed repair (HDR) activity were assessed as described previously (20). All experiments were performed in a 5 weeks timespan. For the drug sensitivity assays, mESC were seeded in quadruplicate at 10,000 cells per well on gelatin-coated 96-well plates to measure the sensitivity of BRCA2 variants to the DNA-damaging agent cisplatin (Accord Healthcare) and the PARP inhibitor, talazoparib (Axon Medchem). Cells were treated with the following concentrations of cisplatin: 0, 0.04, 0.08, 0.16, 0.31, and 0.63 μmol/L (in 1% PBS). Applied concentrations of talazoparib were 0, 0.08, 0.16, 0.31, 0.63, and 1.25 nmol/L (in 1% DMSO). Viable cells were counted using the Novocyte Flow Cytometer (ACEA Biosciences, Inc.) after 48 hours of treatment followed by 24 hours of additional culturing in normal culture medium without a DNA-damaging compound. The E12 splicing patterns were analyzed by fluorescence RT-PCR (23 cycles) as described above for LCLs.

Patient and family data

Clinical, cosegregation, and family data for patients carrying specific BRCA2 E12 variants were collected in collaboration with the French GGC network and within the framework of the COsegregation of VARiants in the BRCA1/2 and PALB2 genes (COVAR) clinical trial (https://clinicaltrials.gov/ct2/show/NCT01689584; Supplementary Data). These variants were previously detected in probands and in relatives undergoing genetic counseling. The criteria for diagnostic variant screening of the BRCA2 gene were applied according to the current French GGC recommendations and the French National Cancer Institute (INCa) guidelines. Informed written consent for genetic analysis was obtained from all patients.

Variant selection

From a total of 133 BRCA2 E12 variants collected from human variation databases, we selected all SNVs generally regarded as LoF variants (n = 15) encompassing four IVS±1/2, five nonsense, and six frameshift variants (Supplementary Fig. S1; Supplementary Tables S2 and S3). In addition, all other intronic, missense, or synonymous variants predicted to affect either 3′/5′ ss (ΔMES ≤ –15%, n = 10) or ESR (ΔtESRseq < –0.7, n = 14; Supplementary Fig. S1; Supplementary Tables S2 and S3) were retained for experimental analysis, as well as the previously described spliceogenic c.6853A>G variant (13). Altogether, this collection comprised 40 variants (Supplementary Fig. S1; Supplementary Tables S2 and S3) that were assessed for their impact on E12 splicing by using an in cellulo minigene assay.

Splicing minigene analysis of presumed LoF variants

In this assay, the WT pCAS2-BRCA2-E12 minigene only produced +E12 transcripts, whereas, as expected, the four IVS±1/2 variants had a drastic impact on splicing (Fig. 1B and C; Supplementary Fig. S2). Indeed, c.6842-1G>T, c.6937+1G>A, and c.6937+1G>T induced full E12 skipping (Fig. 1C; Supplementary Fig. S2), in agreement with the predicted destruction of the 3′ or 5′ ss (Supplementary Fig. S3). Surprisingly, despite bioinformatics predictions strictly identical to c.6937+1G>A and c.6937+1G>T, c.6937+2del generated, in addition to the transcript lacking E12 (ΔE12 = 51%), one RNA isoform with E12 plus retention of the 40 first nucleotides of intron 12 [+E12q(ins40) = 49%; Fig. 1C; Supplementary Fig. S2]. This latter isoform results from the activation of a cryptic intronic 5′ ss predicted for all three variants but only used in the context of c.6937+2del (Supplementary Fig. S3).

Remarkably, in minigene assays, four of the five selected nonsense variants affected splicing (Fig. 1B and C; Supplementary Fig. S2). More precisely, c.6898C>T and c.6920C>A induced a weak exon skipping (ΔE12 = 12% and 13%), whereas c.6844G>T and c.6901G>T led to a major exon skipping (ΔE12 = 73% and 72%; Fig. 1C; Supplementary Fig. S2). In contrast, none of the six frameshift variants affected E12 splicing (Fig. 1B and C; Supplementary Fig. S2).

Splicing minigene analysis of variants with a predicted splicing effect

We extended our analysis to all SNVs located outside IVS±1/2 and predicted to alter the E12 3′/5′ ss (ΔMES ≤ –15%) (Fig. 2A; Supplementary Fig. S4). All 10 variants from this group had a splicing impact in the minigene assay, albeit to different extents (Fig. 2B; Supplementary Fig. S5A; Supplementary Table S3). Of note, additional transcripts deleted for either the first [+E12p(del3)] or the last [+E12q(del3)] three nucleotides of E12 were detected in the context of c.6842G>A and c.6935A>G, respectively (Fig. 2B; Supplementary S5A). Finally, the impact on E12 splicing of the variants predicted to modify ESR (ΔtESRseq < –0.7) was experimentally assessed. In this study, BRCA2 c.6853A>G was used as a positive spliceogenic control because this variant has been previously shown to promote E12 exclusion in patient LCLs, most likely through the alteration of ESR (13). In the minigene assay, we confirmed that this variant induces E12 skipping, albeit moderately (ΔE12 = 16%; Fig. 2C; Supplementary Fig. S5B; Supplementary Table S3). Of note, the ΔtESRseq score associated with this variant (ΔtESRseq = –0.3) did not correctly predict its splicing impact when taking into account a ΔtESRseq value inferior to –0.7 as a positive indicator. In parallel to this variant, we assessed in the minigene assay the splicing effect of 14 E12 variants predicted to affect ESR (ΔtESRseq < –0.7; Fig. 2A and C; Supplementary Fig. S5B; Supplementary Table S3). In comparison, among these selected variants, two variants induced a major or near-total exon skipping (ΔE12 = 58% and 95%), whereas seven variants exhibited a minor or moderate effect (3% ≤ ΔE12 ≤ 17%; Fig. 2C; Supplementary Fig. S5B; Supplementary Table S3). The remaining six selected variants did not cause splicing defect despite predictions of ESR alteration (ΔtESRseq < –0.7). Altogether, these data not only confirmed initial findings but, notably, uncovered new spliceogenic variants potentially altering ESR. They also showed that the approach relying on ΔtESRseq score is not fully reliable for predicting variant-induced E12 splicing effects.

Figure 2.

Identification in minigene assays of BRCA2 exon 12 spliceogenic variants among SNVs predicted to affect splicing. A, Distribution of 24 variants predicted to affect BRCA2 exon 12 splicing. The figure displays the nucleotide sequence of BRCA2 exon 12 and part of its flanking intronic regions. The relative positions of the 10 selected variants predicted to affect BRCA2 exon 12 splice sites (ΔMES ≤ –15%) are indicated above the sequence, whereas the 14 selected variants that may impact BRCA2 exon 12 splicing regulation based on ESR predictions (ΔtESRseq < –0.7) are shown under the sequence. The position of the control spliceogenic variant, c.6853A>G, is also indicated. B, Impact on BRCA2 exon 12 splicing of 10 variants predicted to alter splice sites (ΔMES ≤ –15%). WT and mutant pCAS2-BRCA2-E12 minigene constructs were transiently expressed in HeLa cells. The splicing patterns of the RNA produced from these different minigenes were analyzed by RT-PCR, as described in Materials and Methods. The image shows the agarose gel electrophoresis of the RT-PCR products. The identities of the two major RT-PCR products, with (+E12) or without (ΔE12) exon 12 are indicated on the left of the gel. The other isoforms identified, b and c, correspond to the inclusion of exon 12 deleted either for the first [+E12p(del3)] or the last [+E12q(del3)] three nucleotides. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex between the different RNA isoforms (noted by the star symbol on the gel images), as revealed by band sequencing. Instead, the relative quantifications of the different RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S5), as indicated in Materials and Methods. The data shown represent the average of at least three independent experiments. Error bars, SD values. C, Impact on BRCA2 exon 12 splicing of 14 variants predicted to modify regulatory elements (ΔtESRseq < −0.7). The effects on splicing of the variants selected on the basis of predictions of ESR alterations were assessed using the minigene reporter assay, as described in B. In parallel, we analyzed the previously described spliceogenic c.6853A>G variant (13). The image shows agarose gel electrophoresis of the RT-PCR products. For each minigene, the relative quantifications of the different RNA isoforms were determined as described in B. The data represent the average of at least three independent experiments. Error bars, SD values. nt, nucleotide.

Figure 2.

Identification in minigene assays of BRCA2 exon 12 spliceogenic variants among SNVs predicted to affect splicing. A, Distribution of 24 variants predicted to affect BRCA2 exon 12 splicing. The figure displays the nucleotide sequence of BRCA2 exon 12 and part of its flanking intronic regions. The relative positions of the 10 selected variants predicted to affect BRCA2 exon 12 splice sites (ΔMES ≤ –15%) are indicated above the sequence, whereas the 14 selected variants that may impact BRCA2 exon 12 splicing regulation based on ESR predictions (ΔtESRseq < –0.7) are shown under the sequence. The position of the control spliceogenic variant, c.6853A>G, is also indicated. B, Impact on BRCA2 exon 12 splicing of 10 variants predicted to alter splice sites (ΔMES ≤ –15%). WT and mutant pCAS2-BRCA2-E12 minigene constructs were transiently expressed in HeLa cells. The splicing patterns of the RNA produced from these different minigenes were analyzed by RT-PCR, as described in Materials and Methods. The image shows the agarose gel electrophoresis of the RT-PCR products. The identities of the two major RT-PCR products, with (+E12) or without (ΔE12) exon 12 are indicated on the left of the gel. The other isoforms identified, b and c, correspond to the inclusion of exon 12 deleted either for the first [+E12p(del3)] or the last [+E12q(del3)] three nucleotides. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex between the different RNA isoforms (noted by the star symbol on the gel images), as revealed by band sequencing. Instead, the relative quantifications of the different RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S5), as indicated in Materials and Methods. The data shown represent the average of at least three independent experiments. Error bars, SD values. C, Impact on BRCA2 exon 12 splicing of 14 variants predicted to modify regulatory elements (ΔtESRseq < −0.7). The effects on splicing of the variants selected on the basis of predictions of ESR alterations were assessed using the minigene reporter assay, as described in B. In parallel, we analyzed the previously described spliceogenic c.6853A>G variant (13). The image shows agarose gel electrophoresis of the RT-PCR products. For each minigene, the relative quantifications of the different RNA isoforms were determined as described in B. The data represent the average of at least three independent experiments. Error bars, SD values. nt, nucleotide.

Close modal

BRCA2 E12 splicing patterns in patient LCLs

In LCLs of three control individuals, we detected an alternative splicing pattern of BRCA2 E12 (average ΔE12 = 10.3%; Fig. 3A and B; Supplementary Fig. S6) in agreement with the published data (21, 22). In a LCL derived from a patient heterozygous for a genomic deletion of BRCA2 E12 (DelE12), a major increase in the relative levels of ΔE12 transcripts (ΔE12 = 64%) was observed (Fig. 3B; Supplementary Fig. S6), as expected. A comparable level of exon exclusion (ΔE12 = 61%) was measured for a patient carrying c.6937+1G>A (Fig. 3B; Supplementary Fig. S6), congruent with this variant being responsible for complete E12 skipping. Remarkably, in LCL of a patient carrying the nonsense c.6844G>T variant, the relative levels of ΔE12 transcripts were greatly increased (ΔE12 = 57%; Fig. 3B; Supplementary Fig. S6). Similar results were obtained for LCLs of 2 patients carrying the nonsense c.6901G>T variant (ΔE12 = 52% and 46%, respectively; Fig. 3B; Supplementary Fig. S6). Treatment of these LCLs with puromycin did not significantly affect transcript proportions (Supplementary Fig. S7), suggesting that the increase of ΔE12 levels does not result from nonsense-mediated mRNA decay of the PTC-containing +E12 RNA. In contrast, in LCLs with c.6935A>T or c.6853A>G, a weak or moderate increase in exon skipping was observed (Fig. 3B; Supplementary Fig. S6).

Figure 3.

BRCA2 exon 12 splicing patterns and ASE in control and patient LCLs. A, Location of the primers used for RT-PCR analysis of BRCA2 exon 12 splicing patterns. Boxes indicate BRCA2 exons 11–14 and arrows above indicate the position of primers used in RT-PCR analysis. B, RT-PCR analysis of BRCA2 exon 12 splicing patterns. The BRCA2 exon 12 splicing profiles were assessed by RT-PCR in LCLs of three healthy donors (controls 1–3) and of patients carrying selected variants. For the nonsense c.6844G>T and c.6901G>T (bold), the missense c.6935A>T and c.6853A>G variants, the analyses were conducted in LCLs of one, two, or three unrelated patients, as indicated. The image shows the agarose gel electrophoresis of the RT-PCR products. The identities of the two RT-PCR products, with (+E12) or without (ΔE12) exon 12, are indicated on the left of the gel. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex between the different RNA isoforms (noted by the star symbol on the gel images), as revealed by band sequencing. Instead, for each LCL, the relative quantifications of these RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S6), as indicated in Materials and Methods. The data shown represent the average of at least three independent experiments. Error bars, SD values. C, ASE of transcripts including BRCA2 exon 12 assessed by SNaPshot analysis. To measure ASE of transcripts with exon 12 (+E12) in patient LCLs, a SNaPshot quantitative primer extension assay was performed, as described in Materials and Methods. Genomic segment (gDNA) and cDNA containing BRCA2 exon 12 (+E12) were amplified in parallel using specific primers (black arrows). The SNaPshot quantitative primer extension assay was then performed with a primer (white arrow) targeting the sequence immediately upstream the BRCA2 exon 12 variant (). Results from gDNA (top graphs) and cDNA (bottom graphs) are shown on the right. ASE measured using the variant (ASEVariant) was then calculated after normalization of the cDNA peak area ratio (Variant/WT) to the corresponding gDNA peak area ratio. On the basis of these data, a schematic representation of the ASE, expressed as percentage of the variant and WT alleles, is also presented. Results are representative of three independent experiments.

Figure 3.

BRCA2 exon 12 splicing patterns and ASE in control and patient LCLs. A, Location of the primers used for RT-PCR analysis of BRCA2 exon 12 splicing patterns. Boxes indicate BRCA2 exons 11–14 and arrows above indicate the position of primers used in RT-PCR analysis. B, RT-PCR analysis of BRCA2 exon 12 splicing patterns. The BRCA2 exon 12 splicing profiles were assessed by RT-PCR in LCLs of three healthy donors (controls 1–3) and of patients carrying selected variants. For the nonsense c.6844G>T and c.6901G>T (bold), the missense c.6935A>T and c.6853A>G variants, the analyses were conducted in LCLs of one, two, or three unrelated patients, as indicated. The image shows the agarose gel electrophoresis of the RT-PCR products. The identities of the two RT-PCR products, with (+E12) or without (ΔE12) exon 12, are indicated on the left of the gel. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex between the different RNA isoforms (noted by the star symbol on the gel images), as revealed by band sequencing. Instead, for each LCL, the relative quantifications of these RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S6), as indicated in Materials and Methods. The data shown represent the average of at least three independent experiments. Error bars, SD values. C, ASE of transcripts including BRCA2 exon 12 assessed by SNaPshot analysis. To measure ASE of transcripts with exon 12 (+E12) in patient LCLs, a SNaPshot quantitative primer extension assay was performed, as described in Materials and Methods. Genomic segment (gDNA) and cDNA containing BRCA2 exon 12 (+E12) were amplified in parallel using specific primers (black arrows). The SNaPshot quantitative primer extension assay was then performed with a primer (white arrow) targeting the sequence immediately upstream the BRCA2 exon 12 variant (). Results from gDNA (top graphs) and cDNA (bottom graphs) are shown on the right. ASE measured using the variant (ASEVariant) was then calculated after normalization of the cDNA peak area ratio (Variant/WT) to the corresponding gDNA peak area ratio. On the basis of these data, a schematic representation of the ASE, expressed as percentage of the variant and WT alleles, is also presented. Results are representative of three independent experiments.

Close modal

Altogether, the results obtained in patient LCLs are in agreement with minigene data. Still, it is important to note that the WT minigene did not reproduce the alternative splicing of BRCA2 E12 observed in control LCLs, most probably because of different cellular and/or genomic contexts. As a consequence, the minigene assay may underestimate the level of variant-induced exon skipping. It is also possible that some variants without an effect in the minigene assay exhibit a weak splicing impact in patient LCLs. Despite these limitations, the minigene assay represents a powerful approach for screening spliceogenic BRCA2 E12 variants, as all the variant-induced effects detected by this method were confirmed in available patient LCLs.

ASE of BRCA2 transcripts in patient LCLs

We took advantage of the quantitative SNaPshot assay to measure ASE in +E12 RNA by directly interrogating the presence of exonic variants. From the LCLs with the nonsense variants c.6844G>T or c.6901G>T, we observed an important allelic imbalance (AI) in +E12 transcripts in favor of the WT allele (Fig. 3C). In contrast, a moderate AI was detected in +E12 isoforms expressed in two LCLs with c.6935A>T and in three LCLs with c.6853A>G. Next, ASEs were measured in both +E12 and ΔE12 transcripts by using the c.7242A>G polymorphism located in BRCA2 E14 (Fig. 4AC). As expected, in control LCL, we observed, a nearly equal contribution of the two alleles in the expression of both RNA isoforms. In contrast, in LCL with c.6937+1G>A, a complete imbalance was observed in +E12 transcripts (Fig. 4B). Congruently, a strong AI was detected for ΔE12 transcripts, confirming that this isoform is predominantly produced from the minor allele, which, by deduction, also carries c.6937+1G>A (Fig. 4C).

Figure 4.

ASE of BRCA2 transcripts with and without exon 12 in control and patient LCLs. A, SNaPshot results obtained from genomic DNA. As indicated in the left panel, the genomic segment encompassing BRCA2 exon 14 was PCR amplified from gDNA of a control individual and patients carrying the heterozygous polymorphism located in BRCA2 exon 14, c.7242A>G (SNP, ) by using specific primers (black arrows). The SNaPshot quantitative primer extension assay was then performed with a primer (gray arrow) targeting the sequence immediately upstream the SNP. Results shown on the right panel are representative of three independent experiments. B, ASE of transcripts containing BRCA2 exon 12 (+E12). As shown in the left panel, +E12 cDNA was RT-PCR amplified from LCL RNA using specific primers (black arrows). The SNaPshot assay was next performed as described in A. Results are presented in the right panel. ASE measured in +E12 cDNA using the SNP (ASESNP) was calculated after normalization of the cDNA peak area ratio (minor allele/major allele) to the corresponding gDNA peak area ratio. Complete imbalance (c.i.) is indicated when only one allele was detected. A schematic representation of ASESNP, expressed as percentage of the minor and major alleles, is also presented. Results are representative of three independent experiments. C, ASE of transcripts lacking BRCA2 exon 12 (ΔE12). As indicated in the left panel, ΔE12 cDNA was RT-PCR amplified from LCL RNA using specific primers (black arrows). The SNaPshot quantitative primer extension assay was performed as described in A. Results are presented in right panel. ASE measured in ΔE12 cDNA using the SNP (ASESNP) was calculated after normalization of the cDNA peak area ratio (minor allele/major allele) to the corresponding gDNA peak area ratio. A schematic representation of ASESNP, expressed as percentage of the minor and major alleles, is also presented. Results are representative of three independent experiments.

Figure 4.

ASE of BRCA2 transcripts with and without exon 12 in control and patient LCLs. A, SNaPshot results obtained from genomic DNA. As indicated in the left panel, the genomic segment encompassing BRCA2 exon 14 was PCR amplified from gDNA of a control individual and patients carrying the heterozygous polymorphism located in BRCA2 exon 14, c.7242A>G (SNP, ) by using specific primers (black arrows). The SNaPshot quantitative primer extension assay was then performed with a primer (gray arrow) targeting the sequence immediately upstream the SNP. Results shown on the right panel are representative of three independent experiments. B, ASE of transcripts containing BRCA2 exon 12 (+E12). As shown in the left panel, +E12 cDNA was RT-PCR amplified from LCL RNA using specific primers (black arrows). The SNaPshot assay was next performed as described in A. Results are presented in the right panel. ASE measured in +E12 cDNA using the SNP (ASESNP) was calculated after normalization of the cDNA peak area ratio (minor allele/major allele) to the corresponding gDNA peak area ratio. Complete imbalance (c.i.) is indicated when only one allele was detected. A schematic representation of ASESNP, expressed as percentage of the minor and major alleles, is also presented. Results are representative of three independent experiments. C, ASE of transcripts lacking BRCA2 exon 12 (ΔE12). As indicated in the left panel, ΔE12 cDNA was RT-PCR amplified from LCL RNA using specific primers (black arrows). The SNaPshot quantitative primer extension assay was performed as described in A. Results are presented in right panel. ASE measured in ΔE12 cDNA using the SNP (ASESNP) was calculated after normalization of the cDNA peak area ratio (minor allele/major allele) to the corresponding gDNA peak area ratio. A schematic representation of ASESNP, expressed as percentage of the minor and major alleles, is also presented. Results are representative of three independent experiments.

Close modal

In LCLs of 2 patients heterozygous for both the c.6901G>T nonsense variant and the E14 SNP, we confirmed that +E12 transcripts are predominantly produced from the major/WT allele (Fig. 4B), whereas ΔE12 transcripts are predominantly produced from the minor/variant allele (Fig. 4C). These results strongly suggest that the c.6901G>T nonsense variant is responsible for major E12 skipping in patient samples, in good agreement with the minigene data.

Functional characterization of BRCA2 variants in mESC

For determination of the functional consequences of variant-induced enhanced production of the ΔE12 isoform, we employed the mESC-based BRCA2 functional assay (8, 19, 20). We tested one missense (c.6842G>T), one IVS±1/2 (c.6937+1G>A), and three nonsense (c.6844G>T, c.6859A>T, and c.6901G>T) variants selected on the basis of their splicing effects in minigene assays. First, we showed that all five selected variants were able to rescue the cell-lethal phenotype induced after Cre-mediated deletion of the conditional mouse Brca2 allele, as well as the two controls DelE12 and c.6853A>G (Fig. 5).

Figure 5.

Complementation phenotypes of BRCA2 variants expressed in mESC. Brca2 -/loxP; Pim1DR-GFP/WT cells without BRCA2 (−BRCA2), or expressing BRCA2 WT or variant were transfected with a Cre-GFP expression plasmid to induce loss of the conditional mouse Brca2 allele. Upon Cre-recombinase expression, cells become Brca2 deficient, which is lethal unless complemented by the expression of a (partially) functional BRCA2 variant. Thirteen days post Cre-GFP transfection, culture dishes were stained with methylene blue.

Figure 5.

Complementation phenotypes of BRCA2 variants expressed in mESC. Brca2 -/loxP; Pim1DR-GFP/WT cells without BRCA2 (−BRCA2), or expressing BRCA2 WT or variant were transfected with a Cre-GFP expression plasmid to induce loss of the conditional mouse Brca2 allele. Upon Cre-recombinase expression, cells become Brca2 deficient, which is lethal unless complemented by the expression of a (partially) functional BRCA2 variant. Thirteen days post Cre-GFP transfection, culture dishes were stained with methylene blue.

Close modal

As described previously (13), the WT BRCA2 gene expressed in this system reproduced E12 alternative splicing (ΔE12 = 18%), whereas the DelE12 variant produced solely ΔE12 transcripts (Fig. 6A; Supplementary Fig. S8). The analysis of mESC expressing c.6842G>T and c.6937+1G>A revealed almost complete or full E12 skipping, respectively (Fig. 6A; Supplementary Fig. S8). Importantly, the impact on E12 splicing of the two nonsense variants, c.6844G>T and c.6901G>T, were confirmed in mESC as they caused nearly full E12 skipping (Fig. 6A; Supplementary Fig. S8). In contrast, the third nonsense variant tested, c.6859A>T, exhibited a slight increase in the relative levels of ΔE12 transcripts, as compared with WT (Fig. 6A; Supplementary Fig. S8). For the known spliceogenic c.6853A>G variant (13), a more extensive increase of E12 skipping levels was observed.

Figure 6.

Functional analysis by using a mESC-based assay of BRCA2 variants that affect exon 12 splicing. A,BRCA2 exon 12 splicing patterns. The location of the primers used for RT-PCR analysis of BRCA2 exon 12 splicing patterns is presented on the top panel. Boxes indicate BRCA2 exons 11–14 and arrows represent primers used in RT-PCR analysis. The BRCA2 exon 12 splicing profiles were assessed by RT-PCR in mESC expressing either WT or mutant BRCA2 from a genomic copy of the human gene. The identities of the two RT-PCR products, with (+E12) or without (ΔE12) exon 12, are indicated on the left of the gel. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex (noted by the star symbol on the gel images) between the different RNA isoforms, as revealed by band sequencing. Instead, the relative quantifications of different RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S8). Data represent the average of at least three independent experiments. Error bars, SD values. B, BRCA2 protein expression. Expression of BRCA2 protein either full length (FL), or deleted for exon 12 (ΔE12) or truncated (TC) was detected by Western blot analysis. Vinculin was used as loading control. C, HDR activity. GFP signal was measured by flow cytometry in I-Sce1–expressing cells 2 days post transfection. HDR activity is expressed as the percentage observed in BRCA2 variant–expressing cells relative to BRCA2 WT–expressing cells. Brca2 represents the conditional Brca2−/loxPPim1DR-GFP/WT cell line expressing endogenous mouse Brca2. Error bars indicate the SD of six independent GFP measurements per variant as represented by the dots. Tukey multiple comparisons test was applied to calculate P values comparing DelE12 or c.6853A>G with exon 12 variants. *, P < 0.05; **, P < 0.01; ****, P < 0.0001.

Figure 6.

Functional analysis by using a mESC-based assay of BRCA2 variants that affect exon 12 splicing. A,BRCA2 exon 12 splicing patterns. The location of the primers used for RT-PCR analysis of BRCA2 exon 12 splicing patterns is presented on the top panel. Boxes indicate BRCA2 exons 11–14 and arrows represent primers used in RT-PCR analysis. The BRCA2 exon 12 splicing profiles were assessed by RT-PCR in mESC expressing either WT or mutant BRCA2 from a genomic copy of the human gene. The identities of the two RT-PCR products, with (+E12) or without (ΔE12) exon 12, are indicated on the left of the gel. Here, it is important to note that the gel images were not used for the quantification of the relative transcript levels because of the formation of heteroduplex (noted by the star symbol on the gel images) between the different RNA isoforms, as revealed by band sequencing. Instead, the relative quantifications of different RNA isoforms were determined by capillary electrophoresis of the fluorescence RT-PCR products (see Supplementary Fig. S8). Data represent the average of at least three independent experiments. Error bars, SD values. B, BRCA2 protein expression. Expression of BRCA2 protein either full length (FL), or deleted for exon 12 (ΔE12) or truncated (TC) was detected by Western blot analysis. Vinculin was used as loading control. C, HDR activity. GFP signal was measured by flow cytometry in I-Sce1–expressing cells 2 days post transfection. HDR activity is expressed as the percentage observed in BRCA2 variant–expressing cells relative to BRCA2 WT–expressing cells. Brca2 represents the conditional Brca2−/loxPPim1DR-GFP/WT cell line expressing endogenous mouse Brca2. Error bars indicate the SD of six independent GFP measurements per variant as represented by the dots. Tukey multiple comparisons test was applied to calculate P values comparing DelE12 or c.6853A>G with exon 12 variants. *, P < 0.05; **, P < 0.01; ****, P < 0.0001.

Close modal

Western blot analysis confirmed the expression of BRCA2 protein for all variants after Cre-mediated deletion of the conditional mBrca2 allele (Fig. 6B; Supplementary Fig. S9). It is important to note that the full-length and ΔE12 isoforms cannot be distinguished by Western blotting due to the small difference in size. A truncated BRCA2 protein at the expected size (∼260 kDa) was only detected for the nonsense variant c.6859A>T (p.Arg2287*) that expressed substantial amount of PTC-containing +E12 transcript. In contrast, no truncated protein isoforms were detected for the two other nonsense variants, c.6844G>T (p.Glu2282*) and c.6901G>T (p.Glu2301*), suggesting that these nonsense variants only produce BRCA2 ΔE12 protein, in good agreement with the RNA data (Supplementary Table S3).

For functional characterization of the variants, we assessed their impact on HDR activity, the most prominent BRCA2 tumor suppressor function (Fig. 6C). In a previous validation study, HDR ranges were defined to allow classification of variants in clinically relevant categories. BRCA2 variants with >50% HDR activity are considered to be (likely) benign, while variants with <30% HDR activity are (likely) pathogenic (20).

The DelE12 variant displayed a 50% reduction in HDR activity as compared with WT. A similar reduction in HDR activity was observed for the previously described spliceogenic c.6853A>G missense variant (13), as well as for c.6842G>T, c.6937+1G>A, and the two nonsense variants, c.6844G>T and c.6901G>T. In contrast, the c.6859A>T nonsense variant displayed a more severe functional impact with a reduction of 70% in HDR activity and is the only variant in this set that is associated with a statistically significant reduction in HDR activity as compared with the DelE12 variant (Tukey multiple comparisons test comparing DelE12 with c.6859A>T, P < 0.01). The extent of functional impairment is in-line with the observation that this nonsense variant displayed the lowest relative levels of the “functional” ΔE12 transcript.

Previous data from Li and colleagues (13) showed no difference in homologous recombination between ES cells expressing WT and the BRCA2 DelE12 and c.6853A>G transgenes, whereas here, we report 40%–50% reduction in HDR activity. The observed difference may be explained by differences in experimental set up both at the cell model level (single clones vs. polyclonal population of cells) and the assay used to determine HDR activity (targeted vector integration vs. HDR on a DR-GFP reporter assay). It is possible that the type or range of recombinational events assessed in these assays is not 100% identical. The DR-GFP reporter is an extensively used, well-established, and validated method to determine homologous recombination activity (20, 23). In addition, both our study and the data presented by Li and colleagues (13) show that there is no difference in cell viability of DelE12 and c.6853A>G variants as compared with WT.

Attenuation of the HDR pathway sensitizes cells to cisplatin and PARP inhibitors (20). Congruent with their substantial residual capacity to perform HDR (i.e., 43% ≤ HDR ≤ 61%), the major spliceogenic E12 variants displayed a relatively mild sensitivity to cisplatin and PARP inhibitor (talazoparib), whereas the c.6859A>T nonsense variant with a minor splicing effect and a more severe functional impact (i.e., HDR = 26%) is associated with the highest drug sensitivity (Supplementary Fig. S10).

Clinical and family data of patients carrying E12 spliceogenic variants

We focused on families of 10 probands (P) that carry either variants inducing complete E12 skipping [DelE12 (P1), c.6842G>T (P2), and c.6937+1G>A (P3 and P4)] or nonsense variants that trigger major E12 skipping [c.6844G>T (P5 and P6) and c.6901G>T (P7–P10); Fig. 7; Supplementary Table S4]. When taking into account women over 45 years of age, we observed a partial cosegregation of the variants with breast cancer. Indeed, besides probands, we identified: three carriers with breast cancer, five cancer-free carriers, and three noncarriers with breast cancer. Of note, cooccurrence of c.6842G>T with a pathogenic BRCA2 exon 24 variant, c.9195_9196delinsAT (p.Phe3065_Gln3066delinsLeu*), was observed in P2. Yet, the chromosome phasing of these two variants remains unknown.

Figure 7.

Pedigrees of patients carrying BRCA2 exon 12 spliceogenic variants. Here, the most informative pedigrees are presented. Probands referred to as P2, P3, P5, and P10, respectively, carriers of BRCA2 c.6842G>T (p.Gly2281Val), c.6937+1G>A, c.6844G>T (p.Glu2282*), and c.6901G>T (p.Glu2301*), are indicated by an arrow. Note that in patient P2, both BRCA2 exon 12 variant (c.6842G>T) and BRCA2 exon 24 pathogenic frameshift variant (c.9195_9196delinsAT) were identified. Filled symbols represent patients with cancer (BC, breast cancer; Bilat., bilateral; OC, ovary cancer; CC, colon cancer). Ages at cancer diagnosis are indicated. For some individuals, prophylactic salpingo-oophorectomy (PSO) was performed. Heterozygote carriers for the variant are indicated by the symbol [+], obligate carriers by the symbol (+), and noncarriers by the symbol [–]. CF, cancer free.

Figure 7.

Pedigrees of patients carrying BRCA2 exon 12 spliceogenic variants. Here, the most informative pedigrees are presented. Probands referred to as P2, P3, P5, and P10, respectively, carriers of BRCA2 c.6842G>T (p.Gly2281Val), c.6937+1G>A, c.6844G>T (p.Glu2282*), and c.6901G>T (p.Glu2301*), are indicated by an arrow. Note that in patient P2, both BRCA2 exon 12 variant (c.6842G>T) and BRCA2 exon 24 pathogenic frameshift variant (c.9195_9196delinsAT) were identified. Filled symbols represent patients with cancer (BC, breast cancer; Bilat., bilateral; OC, ovary cancer; CC, colon cancer). Ages at cancer diagnosis are indicated. For some individuals, prophylactic salpingo-oophorectomy (PSO) was performed. Heterozygote carriers for the variant are indicated by the symbol [+], obligate carriers by the symbol (+), and noncarriers by the symbol [–]. CF, cancer free.

Close modal

It is now well established that mis-splicing triggered by specific SNVs is a common cause of Mendelian diseases (24). As a shift of paradigm, this study provides evidence for the first time in a cancer susceptibility gene, as far as we know, that variant-induced perturbation of splicing can rescue complete LoF of presumed pathogenic variants, at least partially.

For this demonstration, we used BRCA2 E12 as a model system and took advantage of complementary experimental approaches to demonstrate that (i) a subset of presumed LoF IVS±1/2 variants as well as nonsense SNVs enhance in-frame skipping of BRCA2 E12 and, as a consequence, (ii) lead to the production of an internally deleted but partially functional BRCA2 protein. Altogether, these data call into question the clinical interpretation of specific spliceogenic hypomorphic IVS±1/2 and nonsense variants.

Whereas the severe impact of IVS±1/2 variants on BRCA2 E12 splicing was expected given the disruption of the highly conserved 3′ or 5′ canonical ss, the effects of the five nonsense variants on splicing, albeit to different extents, were not anticipated and are most probably due to the modification of ESR, via either the destruction of exonic splicing enhancers (ESE) and/or the creation of exonic splicing silencers (ESS; refs. 11, 25). Interestingly, the two major spliceogenic nonsense variants described here, BRCA2 c.6844G>T (p.Glu2282*) and c.6901G>T (p.Glu2301*), lie within purine-rich 5′-AGAA-3′ motifs resembling those described as binding sites for the splicing activator protein Tra2-β1 (26). Most of the selected nonsense variants (4/5) promoted E12 skipping, albeit to different degrees, whereas none of the six frameshift variants tested had a splicing effect in minigene assays. Considering the relative limited number of variants in these two groups, the apparent specific association between nonsense mutations and E12 skipping might be coincidental as, in theory, small insertions or deletions within an exon could also fortuitously disrupt critical ESEs and/or create putative ESSs, both mechanisms potentially promoting exon skipping. Still, the association between nonsense mutations and E12 skipping could also reflect the fact that putative stop codon trinucleotides are enriched in ESS hexamers (27–29).

Nonsense-associated altered splicing has been previously described in cancer susceptibly genes (11). In particular, it has been shown that BRCA1 c.5080G>T (p.Glu1694*) causes in-frame skipping of exon 18, resulting in a nonfunctional protein, given the disruption of its first BRCT domain (30, 31). In contrast, for BRCA2 E12 spliceogenic nonsense variants, the in-frame skipping of the PTC-containing exon not only circumvents protein truncation, but also increases the production of an internally deleted partially functional isoform, as revealed in the mESC assay.

Recently, it has been shown that the mESC assay is a reliable method to predict pathogenicity of missense BRCA2 variants (20). The complementation phenotype, the intermediate HDR activity, and the mild effect on cisplatin and PARP inhibitor sensitivity measured for the two major spliceogenic BRCA2 E12 nonsense variants (c.6844G>T and c.6901G>T) suggest that these initially presumed null variants do not lead to complete LoF. The hypomorphic activity observed for these variants might confer moderate or low risks of breast cancer, as shown for previously characterized hypomorphic missense BRCA1 and BRCA2 variants (32–34). In contrast, the nonsense variant with the lowest level of E12 skipping (c.6859A>T) displays a significant decrease in HDR activity comparable with a previously reported pathogenic missense variant (20). Consistent with the impact on HDR, this variant shows hypersensitivity to cisplatin-induced DNA damage and to PARP inhibition.

Besides the 15 presumed LoF variants, we extended the screening of BRCA2 E12 spliceogenic variants to 25 additional SNVs (intronic, missense, and synonymous variants). Up to this date, only three missense variants have been reported to increase the level of alternative splicing of BRCA2 E12 (13, 35, 36). This study not only confirmed those initial findings but, notably, uncovered 17 new spliceogenic variants affecting either 3′/5′ ss or ESR.

In this study, we confirmed that the variant c.6853A>G promotes E12 skipping, as previously shown (13). However, this effect is partial: the mutant allele also significantly contributes to the production of the full-length +E12 transcript encoding the BRCA2 p.Ile2285Val protein. This observation does not call into question the nonpathogenic status of this variant. However, it prevents (i) from considering the neutral classification of this variant as an additional argument in favor of the functionality of the BRCA2 E12-deleted protein, as initially reported (13) and (ii) from extending this classification to all other BRCA2 E12 spliceogenic variants.

To apprehend the clinical relevance of BRCA2 E12 spliceogenic variants with respect to cancer risk, we analyzed patient clinical data and family history available within the French diagnostic network. Several observations (i.e., identification of both cancer-free carriers and noncarriers with breast cancer and cooccurrence with a pathogenic BRCA2 variant) could be in-line with the hypomorphic activity associated with these variants. Still, the interpretation of these spliceogenic variants cannot be inferred from these limited patient and family information and further larger-scale cosegregation and case–control studies will be needed to accurately evaluate cancer risks conferred by BRCA2 E12 spliceogenic variants.

To our knowledge, the existence of a rescue mechanism for nonsense variants through their direct impact on in-frame splicing has never been described for cancer susceptibility genes. Yet, it has been documented in genes involved in other Mendelian diseases. For instance, in the DMD gene, several truncating SNVs induce in-frame skipping of the PTC-containing exon, resulting in an internally deleted dystrophin, which exhibits partial function and consequently leads to a milder phenotype known as Becker muscular dystrophy (37–39). It is quite possible that presumed LoF BRCA2 E12 variants would confer severe, moderate, or minimal reduction in function, depending of their impact on splicing (i.e., levels of ΔE12 transcript), and would hence be associated with variable degrees of cancer susceptibility.

Of note, for certain SNVs, the BRCA2 ΔE12 RNA levels were somewhat variable in the three complementary approaches used in this study (i.e., minigene, patient LCL, and mESC assays), most probably because of different cellular and genomic contexts. Ideally, E12 splicing patterns should also be quantified in the patient relevant tissues (normal and tumoral breast tissues). Intriguingly, it has been previously reported that BRCA2 E12 alternative splicing is increased in some breast tumors as compared with their matched normal tissues (40) and upregulated in several mammary and ovarian cancer cell lines (41). Some of these cell lines could carry specific spliceogenic variants and provide attractive cellular model systems for future investigations on BRCA2 E12-deleted protein function. Importantly, during the preparation of this article, it has been shown that the highly conserved region encoded by E12 may be involved in the reproductive rather than the tumor-suppressing function of BRCA2 (42, 43). Indeed, the E12-encoded region (amino acid 2281–2313) is part of a binding domain for MEILB2/HSF2BP. This testis-expressed protein recruits BRCA2 to DNA double-strand breaks to ensure meiotic homologous recombination during spermatogenesis. Altogether, these data strongly suggest that BRCA2 E12 is essential to the successful completion of meiotic homologous recombination but not for the role of BRCA2 in mitotic homologous recombination, as a cancer suppressor.

In conclusion, this study provides for the first time experimental arguments to reevaluate the clinical interpretation of a subset of BRCA2 IVS±1/2 and nonsense spliceogenic variants that enhance in-frame skipping of a potentially nonessential exon. On the basis of these findings and following the ACMG (3) and ENIGMA (4) recommendations, we propose to reclassify these SNVs as variants of unknown significance until further genetic analyses establish the cancer risks associated with these variants. This study emphasizes the requirement of complementary RNA and protein analyses to ensure comprehensive variant interpretation, with particular caution taken for presumed LoF variants located in in-frame dispensable exons. In the next-generation sequencing era, the main challenges for medical geneticists are not only the classification of the variants detected in patients but also the quantification of the associated risk, especially for variants retaining partial function. Indeed, our results highlight the need to move forward to a more accurate interpretation of variants in terms of both pathogenicity and penetrance.

H. Tubeuf is a PhD student (paid consultant) at Interactive Biosoftware for the time period October 2015-September 2018 in the context of a public-private PhD project partnership between INSERM and Interactive Biosoftware (CIFRE fellowship #2015/0335). Astra Zeneca contributed financially to the COVAR study (https://clinicaltrials.gov/ct2/show/NCT01689584) coordinated by S. Caputo and D. Stoppa-Lyonnet. No potential conflicts of interest were disclosed by the other authors.

Members of the French GGC network and of the group COVAR (COsegregation of VARiants in the BRCA1/2 and PALB2 genes) are listed in the Supplementary Data.

Conception and design: P. Gaildrat

Development of methodology: H. Vrieling, M.P.G. Vreeswijk

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Meulemans, R.L.S. Mesman, S.M. Caputo, S. Krieger, M. Guillaud-Bataille, N. Boutry-Kryza, J. Sokolowska, F. Révillion, C. Delnatte, H. Tubeuf, O. Soukarieh, F. Bonnet-Dorion, M. Bronner, V. Bourdon, S. Lizard, P. Vilquin, M. Privat, C. Grout, F.M.G.R. Calléja, L. Golmard, D. Stoppa-Lyonnet, C. Houdayer, T. Frebourg, P. Gaildrat

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L. Meulemans, R.L.S. Mesman, S.M. Caputo, S. Krieger, H. Tubeuf, T. Frebourg, M.P.G. Vreeswijk, A. Martins, P. Gaildrat

Writing, review, and/or revision of the manuscript: L. Meulemans, R.L.S. Mesman, S.M. Caputo, L. Golmard, H. Vrieling, C. Houdayer, T. Frebourg, M.P.G. Vreeswijk, A. Martins, P. Gaildrat

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.M. Caputo, V. Caux-Moncoutier, A. Drouet, F.M.G.R. Calléja, C. Houdayer

Study supervision: H. Vrieling, M.P.G. Vreeswijk, A. Martins, P. Gaildrat

This work was financially supported by the Fédération Hospitalo-Universitaire (FHU) Normandy Centre for Genomic and Personalized Medicine (NGP) and the Gefluc. The Dutch Cancer Society KWF (UL2012-5649) supported R.L.S. Mesman and F.M.G.R. Calléja. L. Meulemans was funded by a FHU-NGP PhD fellowship and H. Tubeuf by a CIFRE PhD fellowship (ANRT public-private partnership Inserm/Interactive Biosoftware#2015/0335). The COVAR study is supported by GENETICANCER association. The authors thank Françoise Charbonnier, Anne Rovelet-Lecrux, and Anne-Claire Richard (Inserm U1245, Rouen, France), Philippe Lafitte (Institut Curie, France), and Florence Rousselet (CHRU Nancy, France) for technical support; Jos Jonkers and Peter Bouwman (Cancer Institute, Amsterdam, the Netherlands) for the I-SceI-mCherry plasmid; Shyam Sharan (NCI, Frederick, MD) for the Pl2F7 conditional Brca2 knockout mES cell line; Maria Jasin (Memorial Sloan-Kettering Cancer Center, New York, NY) for the DR-GFP reporter plasmid, and Angela Rosania Solano (Faculty of Medicine, University of Buenos Aires, Buenos Aires, Argentina Argentina) for providing the patient samples.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Mort
M
,
Ivanov
D
,
Cooper
DN
,
Chuzhanova
NA
. 
A meta-analysis of nonsense mutations causing human genetic disease
.
Hum Mutat
2008
;
29
:
1037
47
.
2.
Kuchenbaecker
KB
,
Hopper
JL
,
Barnes
DR
,
Phillips
K-A
,
Mooij
TM
,
Roos-Blom
M-J
, et al
Risks of breast, ovarian, and contralateral breast cancer for BRCA1 and BRCA2 mutation carriers
.
JAMA
2017
;
317
:
2402
16
.
3.
Richards
S
,
Aziz
N
,
Bale
S
,
Bick
D
,
Das
S
,
Gastier-Foster
J
, et al
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology
.
Genet Med
2015
;
17
:
405
23
.
4.
Spurdle
AB
,
Healey
S
,
Devereau
A
,
Hogervorst
FBL
,
Monteiro
ANA
,
Nathanson
KL
, et al
ENIGMA-evidence-based network for the interpretation of germline mutant alleles: an international initiative to evaluate risk and clinical significance associated with sequence variation in BRCA1 and BRCA2 genes
.
Hum Mutat
2012
;
33
:
2
7
.
5.
de la Hoya
M
,
Soukarieh
O
,
López-Perolio
I
,
Vega
A
,
Walker
LC
,
van Ierland
Y
, et al
Combined genetic and splicing analysis of BRCA1 c.[594-2A>C; 641A>G] highlights the relevance of naturally occurring in-frame transcripts for developing disease gene variant classification algorithms
.
Hum Mol Genet
2016
;
25
:
2256
68
.
6.
Mazoyer
S
,
Dunning
AM
,
Serova
O
,
Dearden
J
,
Puget
N
,
Healey
CS
, et al
A polymorphic stop codon in BRCA2
.
Nat Genet
1996
;
14
:
253
4
.
7.
Wu
K
,
Hinson
SR
,
Ohashi
A
,
Farrugia
D
,
Wendt
P
,
Tavtigian
SV
, et al
Functional evaluation and cancer risk assessment of BRCA2 unclassified variants
.
Cancer Res
2005
;
65
:
417
26
.
8.
Kuznetsov
SG
,
Liu
P
,
Sharan
SK
. 
Mouse embryonic stem cell–based functional assay to evaluate mutations in BRCA2
.
Nat Med
2008
;
14
:
875
81
.
9.
Meeks
HD
,
Song
H
,
Michailidou
K
,
Bolla
MK
,
Dennis
J
,
Wang
Q
, et al
BRCA2 polymorphic stop codon K3326X and the risk of breast, prostate, and ovarian cancers
.
J Natl Cancer Inst
2016
;
108
:
djv315
.
10.
Seo
A
,
Steinberg-Shemer
O
,
Unal
S
,
Casadei
S
,
Walsh
T
,
Gumruk
F
, et al
Mechanism for survival of homozygous nonsense mutations in the tumor suppressor gene BRCA1
.
Proc Natl Acad Sci U S A
2018
;
115
:
5241
6
.
11.
Cartegni
L
,
Chew
SL
,
Krainer
AR
. 
Listening to silence and understanding nonsense: exonic mutations that affect splicing
.
Nat Rev Genet
2002
;
3
:
285
98
.
12.
Scotti
MM
,
Swanson
MS
. 
RNA mis-splicing in disease
.
Nat Rev Genet
2016
;
17
:
19
32
.
13.
Li
L
,
Biswas
K
,
Habib
LA
,
Kuznetsov
SG
,
Hamel
N
,
Kirchhoff
T
, et al
Functional redundancy of exon 12 of BRCA2 revealed by a comprehensive analysis of the c.6853A>G (p.I2285V) variant
.
Hum Mutat
2009
;
30
:
1543
50
.
14.
Yeo
G
,
Burge
CB
. 
Maximum entropy modeling of short sequence motifs with applications to RNA splicing signals
.
J Comput Biol
2004
;
11
:
377
94
.
15.
Ke
S
,
Shang
S
,
Kalachikov
SM
,
Morozova
I
,
Yu
L
,
Russo
JJ
, et al
Quantitative evaluation of all hexamers as exonic splicing elements
.
Genome Res
2011
;
21
:
1360
74
.
16.
Di Giacomo
D
,
Gaildrat
P
,
Abuli
A
,
Abdat
J
,
Frébourg
T
,
Tosi
M
, et al
Functional analysis of a large set of BRCA2 exon 7 variants highlights the predictive value of hexamer scores in detecting alterations of exonic splicing regulatory elements
.
Hum Mutat
2013
;
34
:
1547
57
.
17.
Soukarieh
O
,
Gaildrat
P
,
Hamieh
M
,
Drouet
A
,
Baert-Desurmont
S
,
Frébourg
T
, et al
Exonic splicing mutations are more prevalent than currently estimated and can be predicted by using in silico tools
.
PLoS Genet
2016
;
12
:
e1005756
.
18.
Gaildrat
P
,
Killian
A
,
Martins
A
,
Tournier
I
,
Frébourg
T
,
Tosi
M
. 
Use of splicing reporter minigene assay to evaluate the effect on splicing of unclassified genetic variants
.
Methods Mol Biol
2010
;
653
:
249
57
.
19.
Hendriks
G
,
Morolli
B
,
Calléja
FMGR
,
Plomp
A
,
Mesman
RLS
,
Meijers
M
, et al
An efficient pipeline for the generation and functional analysis of human BRCA2 variants of uncertain significance
.
Hum Mutat
2014
;
35
:
1382
91
.
20.
Mesman
RLS
,
Calléja
FMGR
,
Hendriks
G
,
Morolli
B
,
Misovic
B
,
Devilee
P
, et al
The functional impact of variants of uncertain significance in BRCA2
.
Genet Med
2019
;
21
:
293
302
.
21.
Fackenthal
JD
,
Yoshimatsu
T
,
Zhang
B
,
de Garibay
GR
,
Colombo
M
,
De Vecchi
G
, et al
Naturally occurring BRCA2 alternative mRNA splicing events in clinically relevant samples
.
J Med Genet
2016
;
53
:
548
58
.
22.
Davy
G
,
Rousselin
A
,
Goardon
N
,
Castéra
L
,
Harter
V
,
Legros
A
, et al
Detecting splicing patterns in genes involved in hereditary breast and ovarian cancer
.
Eur J Hum Genet
2017
;
25
:
1147
54
.
23.
Guidugli
L
,
Shimelis
H
,
Masica
DL
,
Pankratz
VS
,
Lipton
GB
,
Singh
N
, et al
Assessment of the clinical relevance of BRCA2 missense variants by functional and computational approaches
.
Am J Hum Genet
2018
;
102
:
233
48
.
24.
Soemedi
R
,
Cygan
KJ
,
Rhine
CL
,
Wang
J
,
Bulacan
C
,
Yang
J
, et al
Pathogenic variants that alter protein code often disrupt splicing
.
Nat Genet
2017
;
49
:
848
55
.
25.
Wang
Z
,
Burge
CB
. 
Splicing regulation: from a parts list of regulatory elements to an integrated splicing code
.
RNA
2008
;
14
:
802
13
.
26.
Cléry
A
,
Jayne
S
,
Benderska
N
,
Dominguez
C
,
Stamm
S
,
Allain
FH-T
. 
Molecular basis of purine-rich RNA recognition by the human SR-like protein Tra2-β1
.
Nat Struct Mol Biol
2011
;
18
:
443
50
.
27.
Wang
Z
,
Rolish
ME
,
Yeo
G
,
Tung
V
,
Mawson
M
,
Burge
CB
. 
Systematic identification and analysis of exonic splicing silencers
.
Cell
2004
;
119
:
831
45
.
28.
Sterne-Weiler
T
,
Howard
J
,
Mort
M
,
Cooper
DN
,
Sanford
JR
. 
Loss of exon identity is a common mechanism of human inherited disease
.
Genome Res
2011
;
21
:
1563
71
.
29.
Ke
S
,
Anquetil
V
,
Zamalloa
JR
,
Maity
A
,
Yang
A
,
Arias
MA
, et al
Saturation mutagenesis reveals manifold determinants of exon definition
.
Genome Res
2018
;
28
:
11
24
.
30.
Mazoyer
S
,
Puget
N
,
Perrin-Vidoz
L
,
Lynch
HT
,
Serova-Sinilnikova
OM
,
Lenoir
GM
. 
A BRCA1 nonsense mutation causes exon skipping
.
Am J Hum Genet
1998
;
62
:
713
5
.
31.
Liu
H-X
,
Cartegni
L
,
Zhang
MQ
,
Krainer
AR
. 
A mechanism for exon skipping caused by nonsense or missense mutations in BRCA1 and other genes
.
Nat Genet
2001
;
27
:
55
8
.
32.
Spurdle
AB
,
Whiley
PJ
,
Thompson
B
,
Feng
B
,
Healey
S
,
Brown
MA
, et al
BRCA1 R1699Q variant displaying ambiguous functional abrogation confers intermediate breast and ovarian cancer risk
.
J Med Genet
2012
;
49
:
525
32
.
33.
Shimelis
H
,
Mesman
RLS
,
Von Nicolai
C
,
Ehlen
A
,
Guidugli
L
,
Martin
C
, et al
BRCA2 hypomorphic missense variants confer moderate risks of breast cancer
.
Cancer Res
2017
;
77
:
2789
99
.
34.
Moghadasi
S
,
Meeks
HD
,
Vreeswijk
MP
,
Janssen
LA
,
Borg
Å
,
Ehrencrona
H
, et al
The BRCA1 c. 5096G>A p.Arg1699Gln (R1699Q) intermediate risk variant: breast and ovarian cancer risk estimation and recommendations for clinical management from the ENIGMA consortium
.
J Med Genet
2018
;
55
:
15
20
.
35.
Brandão
RD
,
van Roozendaal
K
,
Tserpelis
D
,
García
EG
,
Blok
MJ
. 
Characterisation of unclassified variants in the BRCA1/2 genes with a putative effect on splicing
.
Breast Cancer Res Treat
2011
;
129
:
971
82
.
36.
Grodecká
L
,
Lockerová
P
,
Ravčuková
B
,
Buratti
E
,
Baralle
FE
,
Dušek
L
, et al
Exon first nucleotide mutations in splicing: evaluation of in silico prediction tools
.
PLoS One
2014
;
9
:
e89570
.
37.
Shiga
N
,
Takeshima
Y
,
Sakamoto
H
,
Inoue
K
,
Yokota
Y
,
Yokoyama
M
, et al
Disruption of the splicing enhancer sequence within exon 27 of the dystrophin gene by a nonsense mutation induces partial skipping of the exon and is responsible for Becker muscular dystrophy
.
J Clin Invest
1997
;
100
:
2204
10
.
38.
Ginjaar
IB
,
Kneppers
AL
,
v d Meulen
J-DM
,
Anderson
LV
,
Bremmer-Bout
M
,
van Deutekom
JC
, et al
Dystrophin nonsense mutation induces different levels of exon 29 skipping and leads to variable phenotypes within one BMD family
.
Eur J Hum Genet
2000
;
8
:
793
6
.
39.
Tuffery-Giraud
S
,
Saquet
C
,
Thorel
D
,
Disset
A
,
Rivier
F
,
Malcolm
S
, et al
Mutation spectrum leading to an attenuated phenotype in dystrophinopathies
.
Eur J Hum Genet
2005
;
13
:
1254
60
.
40.
Bièche
I
,
Lidereau
R
. 
Increased level of exon 12 alternatively spliced BRCA2 transcripts in tumor breast tissue compared with normal tissue
.
Cancer Res
1999
;
59
:
2546
50
.
41.
Rauh-Adelmann
C
,
Lau
K-M
,
Sabeti
N
,
Long
JP
,
Mok
SC
,
Ho
S-M
. 
Altered expression of BRCA1, BRCA2, and a newly identified BRCA2 exon 12 deletion variant in malignant human ovarian, prostate, and breast cancer cell lines
.
Mol Carcinog
2000
;
28
:
236
46
.
42.
Zhang
J
,
Fujiwara
Y
,
Yamamoto
S
,
Shibuya
H
. 
A meiosis-specific BRCA2 binding protein recruits recombinases to DNA double-strand breaks to ensure homologous recombination
.
Nat Commun
2019
;
10
:
722
.
43.
Brandsma
I
,
Sato
K
,
van Rossum-Fikkert
SE
,
van Vliet
N
,
Sleddens
E
,
Reuter
M
, et al
HSF2BP interacts with a conserved domain of BRCA2 and is required for mouse spermatogenesis
.
Cell Rep
2019
;
27
:
3790
8
.

Supplementary data