Minigene-based characterization and classification of splice-associated variants in succinate dehydrogenase B

Abstract

Abstract Pathogenic germline variants in SDHB are associated with an increased risk for tumors such as pheochromocytoma and paraganglioma. However, limited functional evidence for rare variants poses a major challenge for clinical decision making. To systematically investigate potentially splice-associated variants of unknown significance (VUS) in SDHB, we created a minigene spanning exons 2-5 and assessed minigene-derived SDHB transcripts in HEK293T cells using targeted NGS analysis. We characterized the transcriptional impact of 48 variants prioritized by SpliceAI (Δ ≥ 0.42), along with two negative controls as well as endogenous SDHB, and compared effects with primary cancer data (n=2). While 19 variants (38%) showed ≥90% wildtype expression, 17 variants (34%) exhibited ≥90% aberrant splicing. Across all variants, an average of 2.3 transcripts per variant was detected, yielding a total of 113 transcripts. Applying a customized PVS1/BP7 decision tree, weighted transcript strengths could be assigned to 104 transcripts (92%). For a total of 26 classified variants, this yielded a PVS1_Strong (RNA) code for 10 variants (38%), including eight canonical splice-site variants, one missense variant and one stop-gain variant, a PVS1_Moderate (RNA) code for two non-canonical intronic variants (8%), and a BP7_Strong (RNA) code for 14 intronic or synonymous variants (54%). Integration of minigene RNA data resulted in an average ACMG point change (Δ) of 2.7 (median Δ = 3.5; range: 1–4), with an increase for three variants (11.5%) and a decrease for 23 variants (88.5%). Reclassification occurred in 13 variants (50%), with 12 variants downgraded from VUS to likely benign, and one variant (c.402T>A) downgraded from likely pathogenic to VUS. We conclude that targeted RNA sequencing of minigene derived transcripts represents a precise and scalable approach for assessing splice-associated variants for precision oncology. Nevertheless, RNA-based evidence should be interpreted in the context of complementary functional and clinical data to ensure robust variant classification.