Splicing factors in the heart: Uncovering shared and unique targets

Abstract

Alternative splicing generates specialized protein isoforms that allow the heart to adapt during development and disease. The recent discovery that mutations in the splicing factor RNA-binding protein 20 (RBM20) cause a severe form of familial dilated cardiomyopathy has sparked a great interest in alternative splicing in the field of cardiology. Since then, identification of splicing factors controlling alternative splicing in the heart has grown at a rapid pace. Despite the intriguing observation that a certain overlap exists between the targets of some splicing factors, an integrated and systematic analysis of their splicing networks is missing. Here, we compared the splicing networks of individual splicing factors by re-analyzing original RNA-sequencing data from eight previously published mouse models, in which a single splicing factor has been genetically deleted (i.e. HNRNPU, MBNL1/2, QKI, RBM20, RBM24, RBPMS, SRSF3, SRSF4). We show that key splicing events in Camk2d, Ryr2, Tpm1, Tpm2 and Pdlim5 require the combined action of the majority of these splicing factors. Additionally, we identified common targets and pathways among splicing factors, with the largest overlap between the splicing networks of MBNL, QKI and RBM24. We also re-analyzed a large-scale RNA-sequencing study on hearts of 128 heart failure patients. Here, we observed that MBNL1, QKI and RBM24 expression varied greatly. This variation in expression correlated with differential splicing of their downstream targets as found in mice, suggesting that aberrant splicing by MBNL1, QKI and RBM24 might contribute to the disease mechanism in heart failure.