Transcript fusion between CDKN1A and RAB44 caused by exon skipping like mechanism due to disruption of a splice site

Splicing contributes to gene regulation and protein diversity, while abnormal splicing underlies both hereditary diseases and cancers. Various mutations that disrupt splicing factors, exonic or intronic splicing enhancers or silencers, as well as splice sites, could be responsible for abnormal splicing. Characterization of abnormal splicing events is not only helpful for understanding the molecular processes linking mutations to disease phenotypes, but also provides promising targets for targeted therapies. In addition, CRISPR/Cas9 editing could be benefited once more attention is given to potential abnormal splicing outcomes other than off-target effects at the DNA level. Although large-scale multiplexed genome editing has been demonstrated in yeast, and has also been attempted for particular exons or genes in other eukaryotic cells to achieve saturation, in practice it is much more difficult to measure splicing consequences with genome-wide saturation editing in human cells. Instead, massive somatic mutations accumulated in cancer cohorts provide invaluable opportunities to study somatic mutation-associated splicing events. Abnormal splicing is not necessarily limited to single genes. Transcript fusion is a special form of abnormal splicing that connects two or more genes due to splicing on a transcriptional level (rather than chromosomal translocations such as BCR-ABL in chronic myeloid leukemia). It could arise from conventional splicing on read-through transcripts when the two genes are next to each other and on the same strand, or from trans-splicing when two genes are on different chromosomes, strands or far away - a few cases had been reported. However, it was found that these fusions not only occurred in tumors but also in normal tissues; there was limited investigation regarding how the fusion could happen, whether it be due to mutations or not, and what the downstream perturbations were. Here, in an effort to characterize somatic mutation-associated abnormal splicing (especially in its simplest form, exon skipping events), we identified over one hundred such events in various tumors, including those in MET, PTEN and TP53. Surprisingly, we detected a recurrent, but previously undescribed, tumor-specific transcript fusion event between the cyclin-dependent kinase inhibitor CDKN1A and the RAS oncogene family gene RAB44. By creating genome-edited cell lines, we demonstrate a causal relationship between splice-site mutations in CDKN1A and the fusion to the RAB44 transcript. We further provide evidence that the fusion arises from a readthrough transcript that escapes exosome-mediated degradation when the splice-site mutation occurred, and we show that the presence of the fusion transcript correlates with TP53 inactivation and CDK activation. The strong tissue specificity of RAB44 and the relatively high prevalence of this transcript fusion in multiple types of cancers warrants further study which could inform subclassifications of these cancers and the development of targeted therapies.