Quantitative analysis of cryptic splicing associated with TDP-43 depletion

Abstract

Reliable exon recognition is key to the splicing of pre-mRNAs into mature mRNAs. TDP-43 is an RNA-binding protein whose nuclear loss and cytoplasmic aggregation are a hallmark pathology in amyotrophic lateral sclerosis and frontotemporal dementia (ALS/FTD). TDP-43 depletion causes the aberrant inclusion of cryptic exons into a range of transcripts, but their extent, relevance to disease pathogenesis and whether they are caused by other RNA-binding proteins implicated in ALS/FTD are unknown. We developed an analysis pipeline to discover and quantify cryptic exon inclusion and applied it to publicly available human and murine RNA-sequencing data. We detected widespread cryptic splicing in TDP-43 depletion datasets but almost none in another ALS/FTD-linked protein FUS. Sequence motif and iCLIP analysis of cryptic exons demonstrated that they are bound by TDP-43. Unlike the cryptic exons seen in hnRNP C depletion, those linked to TDP-43 do not originate from a particular family of transposable element. Cryptic exons are poorly conserved and inclusion overwhelmingly leads to nonsense-mediated decay of the host transcript, with reduced transcript levels observed in differential expression analysis. RNA-protein interaction data on 73 different RNA-binding proteins showed that, in addition to TDP-43, 7 specifically bind TDP-43 linked cryptic exons.This suggests that TDP-43 competes with other splicing factors for binding to cryptic exons and can repress cryptic exon inclusion. Our quantitative analysis pipeline confirms the presence of cryptic exons during TDP-43 depletion and will aid investigation of their relevance to disease.