Therapeutic targeting of alternative splicing caused by a lethal noncoding structural variant in X-linked dystonia parkinsonism.

X-linked Dystonia-Parkinsonism (XDP) is a lethal adult-onset neurodegenerative disorder that exhibits features of dystonia and parkinsonism and is exclusively associated with a causal founder haplotype that is indigenous to the Philippines and affects Filipino males. Using patient-specific fibroblasts, neural stem cells (NSC), and other neuronal models, we discovered that cryptic alternative splicing caused by a novel SINE-VNTR-Alu (SVA) mobile element insertion into intron 32 of TAF1 is a mechanistic hallmark of XDP. We leveraged postmortem brain samples from an XDP-specific brain bank to demonstrate that the molecular hallmarks of XDP observed in neural stem cells (NSCs) mirror abnormalities observed in brain tissues from affected patients. Based on these findings that patient-specific NSCs reproduce mechanistic signatures found in the brain, we sought to develop a bespoke precision therapeutic for XDP and evaluate its relative efficacy in ameliorating transcriptomic signatures in neuronal models. We first used CRISPR-based excision of the SVA and demonstrated ablation of all aberrant splicing and dysregulation of TAF1 expression in NSCs across 30 independent clones. CRISPR-based correction of the XDP haplotype also restored the expression of 424 of 1,490 (30%) differentially expressed genes (DEGs) that were altered in XDP patient lines and greatly exceeded what would be expected by chance (p-value = 9.89e-87). While in vivo delivery of a gold standard CRISPR therapy is currently not feasible for XDP, we evaluated a tractable approach for Filipino patients by exploring the potential to modulate alternative splicing in XDP patients using antisense oligonucleotides (ASOs). To accomplish this, we developed a large-scale and well controlled functional genomics platform that screened eighty ASOs targeting intron 32 of XDP patients, followed by prioritization of lead ASOs based on attenuation of the alternative splicing signature. In transcriptome analyses across 1,550 libraries, we found that 8 of the 12 lead ASOs ameliorated the targeted XDP aberrant splicing. Moreover, we found that the two lead ASOs exhibited 38% and 43% rescue of XDP-specific DEGs that were also rescued by CRISPR excision of the SVA (enrichment p-values = 2.06e-13 and 2.27e-05, respectively). These rescues represented restoration of key molecular functions previously implicated in XDP, such as synaptic function, DNA-binding transcription factor activity, and gliogenesis. This study highlights a path to a potential targeted therapeutic for XDP and the capacity to exploit functional genomic signatures in patient-derived neural models to develop a scalable precision therapeutic platform for rare genetic disorders.