ABSTRACT: Tyrosine kinase inhibitors (TKIs), as a class of small-molecule drugs that exert anti-tumor effects by inhibiting tyrosine kinase-catalyzed phosphorylation, have been used in the treatment of various cancers. Sorafenib, as a multi-targeted TKI drug, is the first-line treatment for advanced renal cell carcinoma and unresectable hepatocellular carcinoma. However, sorafenib has repeatedly been reported to cause cardiac events in patients without a history of heart diseases during clinical use, indicating that it has cardiotoxicity. Alternative splicing of cardiac contraction-related genes happens during heart development and cardiac diseases, and is critical for heart function. However, whether alternative splicing plays a role in drug-induced cardiotoxicity remains unexplored. RBM20 is an important cardiac-specific splicing factor, mutations of which cause dilated cardiomyopathy or other cardiac dysfunctions. Rbm20 also mediates alternative splicing of genes essential for heart contraction, which is often negatively affected in drug-induced cardiotoxicity. Existing studies do not fully explain the mechanism of sorafenib cardiotoxicity, and none of the relationship between cardiotoxicity of sorafenib and alternative splicing mediated by tissue-specific splicing factors, such as Rbm20, have been reported. In order to explore whether cardiac-specific alternative splicing plays a role in sorafenib-induced cardiotoxicity, we establish both cell and animal models of cardiotoxicity, and obtain the following results: (1) By constructing a rat animal model administered with sorafenib, we find that sorafenib causes abnormal cardiac function in rats, and the genes that undergo alternative splicing in rat hearts are related to cytoskeleton of actin; (2) Alternatively spliced genes induced by sorafenib in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) are enriched in sarcomere, actin filament, calcium transient regulation, mitochondria, all of which are critical for cardiac contraction. These genes are associated with dilated cardiomyopathy, hypertrophic cardiomyopathy and other cardiomyopathy; (3) Sorafenib induces a decrease in the expression of cardiac-specific splicing factor RBM20; (3) Many genes whose splicing are altered by sorafenib overlap with Rbm20 targets, indicating that sorafenib may affect alternative splicing through Rbm20; (4) Sorafenib induces pathogenic alternative splicing of FHOD3, which is a RBM20 target gene and participates in myocardial sarcomere formation. Sorafenib also affects alternative splicing of SLC25A3, which encodes a phosphate transporter on the mitochondrial inner membrane and regulates ATP synthesis; (5) Enhancing the expression of RBM20 rescues the cardiotoxicity of sorafenib by reducing apoptosis and increasing ATP levels, which is mediated by reversing the alternative splicing of FHOD3 and SLC25A3 induced by sorafenib. This paper uncovers that sorafenib reduces the expression of RBM20 to cause pathogenic alternative splicing of genes related to myocardial sarcomere and energy mechanism, resulting in abnormal myocardial function. Increasing the expression of RBM20 reverses the alternative splicing of FHOD3 and SLC25A3 associated with cardiac sarcomeres and mitochondria respectively, rescuing the cardiotoxicity of sorafenib.
