Background: The identification of pretherapeutic somatic BRCA variants can have considerable clinical impact given that they affect response to the new poly (ADP-ribose) polymerase (PARP)-targeted therapy. One major issue with this type of testing is the identification of splicing variants of uncertain significance (VUS) on degraded somatic messenger RNA. It is therefore important to be able to quickly characterize these splice variants.
Objective: As part of PARP inhibitor targeted therapy, we have investigated a method for the direct confirmation of potential pathogenic somatic splice variants of BRCA1 found in fixed tumor samples. Previously these VUS have commonly only been tested by in silico analysis.
Methods: Five BRCA1 variants affecting splicing were characterized from formalin-fixed, paraffin-embedded (FFPE) ovarian carcinoma tissues by next-generation sequencing (NGS). Three patient samples had already been functionally characterized and were used as controls. Total somatic RNA from samples was extracted, reverse-transcribed, and amplified with several primer pairs encompassing the target exon. The polymerase chain reaction (PCR) products were analyzed by capillary gel electrophoresis to assess possible changes in size due to splicing alterations. Finally, we confirmed our results by cloning, followed by Sanger sequencing, and analyzed the expression of the aberrant forms.
Results: Our molecular approach made it possible to visualize the splicing outcomes of three variants (c.5194-2A>G, c.5434C>G, and c.547+1G>A) already identified and present in databases and/or identified with prediction tools (ClinVar, UMD, ARUP Utah database, and Human Splice Finder splices sites prediction) and to confirm their exon skipping consequences, their expression in tumors, and thus their pathogenicity. The c.4484+5G>A variant was not found in databases and was predicted to have no impact on splicing, but was found to display altered processing in tumor tissue. This variant also had a major detrimental impact on transcriptional expression.
Conclusion: In a break from purely in silico approaches, we propose a simple and rapid pretherapeutic functional analysis of somatic BRCA1 variants potentially involved in splicing alterations. This approach will allow more ovarian cancer patients to benefit from new therapies targeting PARP.