Alternative splicing in Alzheimer's disease: event-level evidence, risk-locus biology, and RNA-based therapeutic perspectives.

Genetic and transcriptomic studies have identified many Alzheimer's disease (AD) risk loci, but many remain difficult to interpret at the level of gene expression alone. Alternative splicing (AS) provides an event-level framework for linking genetic variation, transcript architecture, RNA surveillance, and protein-domain remodeling to AD biology. This Review evaluates AD-associated splice and isoform events according to event definition, human recurrence, genetic proximity, molecular consequence, perturbation evidence, and translational tractability. Current evidence is concentrated in several mechanistic areas, including amyloid processing and endolysosomal trafficking, tau biology and spliceosome dysfunction, TDP-43-associated cryptic splicing, immune and lipid-risk loci, and transport or endocytic risk loci. ApoER2 exon 19 correction provides AD-specific preclinical evidence for splice-event modulation, whereas BIIB080/MAPTRx illustrates the feasibility of CNS antisense oligonucleotide delivery and target-engagement monitoring in patients. By contrast, evidence for CD33, PSEN2, MAPT intron retention, TREM2, APOE/CLU, and MS4A-associated post-transcriptional regulation currently supports risk-locus interpretation, molecular stratification, or assay development rather than intervention-ready splice targets. CircRNAs, age-related loss of splicing fidelity, and sex-biased intron retention broaden the field, but their AD-specific functional relevance remains incompletely defined. Overall, AS should not be viewed as a single pathogenic mechanism in AD. Selected splice events may improve AD risk-locus interpretation, molecular stratification, biomarker development, and the design of RNA-based therapeutic strategies.