Mapping intronic mutational hotspots by in silico mutagenesis enables single antisense oligonucleotide correction of multiple variants

Abstract

Deep intronic variants remain an understudied class of pathogenic variation, primarily due to their absence from standard exome and gene panel datasets and the complexity of non-coding genome interpretation. We hypothesized that pathogenic deep intronic variants are not randomly distributed but instead cluster within intronic "hotspots" inherently prone to pseudoexon activation, and that mapping such regions could improve molecular diagnosis. Importantly, such intronic hotspots nominate targets for antisense oligonucleotide (ASO) therapy. Using X-linked Alport syndrome as a proof-of-concept model and based on our previous work, we screened unsolved patients across multiple European diagnostic centers for variants within a defined region of COL4A5 intron 6. We identified eight independent variants in more than 35 affected individuals from ten unrelated families, which led to two pseudoexon inclusion events, both using the same strong cryptic splice donor site. In all, RNA sequencing and/or minigene assays confirmed aberrant splicing, even when prediction tools were discordant or fell below clinical thresholds. A single ASO targeting the shared donor site restored normal COL4A5 mRNA and 5(IV) collagen protein expression in patient-derived cells regardless of the causative variant. Extending this analysis gene-wide using the AlphaGenome sequence-to-function model, we confirmed intron 6 as one of the most critical COL4A5 splicing hotspot and identified additional potential hotspots harboring novel predicted spliceogenic variants. This gene-agnostic framework establishes a systematic strategy for identifying intronic mutational hotspots and matching patients to scalable, mutation-agnostic ASO-based precision therapies. Significance StatementDeep intronic variants are an underrecognized cause of genetic disease, undetectable to standard exome and gene-panel testing and difficult to interpret. Using X-linked Alport syndrome as an exemplar, we identified a recurrent deep intronic mutational hotspot in which independent pathogenic variants from unrelated families converge on the same cryptic splice donor site to activate pseudoexon inclusion. A single antisense oligonucleotide therapy corrected this shared splicing defect and restored normal collagen IV protein production in patient cells, regardless of the underlying causative variant. Extending this analysis gene-wide using an AlphaGenome sequence-to-function splicing model, we uncovered additional deep intronic hotspots across the gene, establishing a gene-agnostic strategy for systematically identifying hidden pathogenic variants and matching patients to mutation-agnostic, RNA-based precision therapies in Mendelian disorders.