RNA splicing is a fundamental biological process that generates protein diversity from a finite set of genes. Recurrent somatic mutations of splicing factor genes are rare in Acute Myeloid Leukemia (AML, < 20%). We examined whether RNA splicing differences might exist in AML even in the absence of splicing factor mutations. Analysing RNA-seq data from two independent cohorts of AML patients, we have identified recurrent differential alternative splicing between AML patients with poor or favourable prognosis. The alternately spliced genes are enriched for specific molecular pathways, implicating aberrant RNA splicing as a driver of leukemia.
Developing informatics tools to predict the functional impact of alternative splicing on the translated protein, we have discovered that ~45% of the splicing events directly affect highly-conserved protein domains. A number of splicing factors are themselves misspliced, resulting in specific splicing alterations of their target transcripts that we have observed. By studying differential gene expression within the same patients, we have identified that alternative splicing results in the induction of a stress response and upregulation of inflammation-related genes in AML patients with adverse outcomes. Lastly, using machine learning techniques, we have identified a set of genes whose alternative splicing can be used to refine the accuracy of existing risk prognosis schemes in AML.
Taken together, we report the presence of extensive deregulation of RNA splicing in AML patients, even in the absence of any splicing factor mutations. Many of these events were shared in patients with adverse outcomes and their impact on the AML transcriptome points towards vulnerabilities that could be targeted.