Background: TTN (Titin), the largest protein in humans, forms the molecular spring that spans half of the sarcomere to provide passive elasticity to the cardiomyocyte. Mutations that disrupt the TTN transcript are the most frequent cause of hereditary heart failure. We showed before that TTN produces a class of circular RNAs (circRNAs) that depend on RBM20 to be formed. In this study, we show that the back-splice junction formed by this class of circRNAs creates a unique motif that binds SRSF10 to enable it to regulate splicing. Furthermore, we show that one of these circRNAs (cTTN1) distorts both localization of and splicing by RBM20. Methods: We calculated genetic constraint of the identified motif in 125 748 exomes collected from the gnomAD database. Furthermore, we focused on the highest expressed RBM20-dependent circRNA in the human heart, which we named cTTN1. We used shRNAs directed to the back-splice junction to induce selective loss of cTTN1 in human induced pluripotent stem cell–derived cardiomyocytes. Results: Human genetics suggests reduced genetic tolerance of the generated motif, indicating that mutations in this motif might lead to disease. RNA immunoprecipitation confirmed binding of circRNAs with this motif to SRSF10. Selective loss of cTTN1 in human induced pluripotent stem cell–derived cardiomyocytes induced structural abnormalities, apoptosis, and reduced contractile force in engineered heart tissue. In line with its SRSF10 binding, loss of cTTN1 caused abnormal splicing of important cardiomyocyte SRSF10 targets such as MEF2A and CASQ2 . Strikingly, loss of cTTN1 also caused abnormal splicing of TTN itself. Mechanistically, we show that loss of cTTN1 distorts both localization of and splicing by RBM20. Conclusions: We demonstrate that circRNAs formed from the TTN transcript are essential for normal splicing of key muscle genes by enabling splice regulators RBM20 and SRSF10. This shows that the TTN transcript also has regulatory roles, besides its well-known signaling and structural function. In addition, we demonstrate that the specific sequence created by the back-splice junction of these circRNAs has important functions. This highlights the existence of functionally important sequences that cannot be recognized as such in the human genome but provides an as-yet unrecognized source for functional sequence variation.
Circular RNAs (circRNAs) are a class of non-coding RNA molecules that are gaining increasing attention for their roles in various pathophysiological processes. The RNA-binding protein quaking (QKI) has been identified as a regulator of circRNA formation. In this study, we investigate the role of QKI in the formation of circRNAs in the heart by performing RNA-sequencing on Qki-knockout mice. Loss of QKI resulted in the differential expression of 17% of the circRNAs in adult mouse hearts. Interestingly, the majority of the QKI-regulated circRNAs (58%) were derived from genes undergoing QKI-dependent splicing, indicating a relationship between back-splicing and linear splicing. We compared these QKI-dependent circRNAs with those regulated by RBM20, another cardiac splicing factor essential for circRNA formation. We found that QKI and RBM20 regulate the formation of a distinct, but partially overlapping set of circRNAs in the heart. Strikingly, many shared circRNAs were derived from the Ttn gene, and they were regulated in an opposite manner. Our findings indicate that QKI not only regulates alternative splicing in the heart but also the formation of circRNAs.
Long-QT syndrome type 1 (LQT1) is caused by mutations in KCNQ1. Patients heterozygous for such a mutation co-assemble both mutant and wild-type KCNQ1-encoded subunits into tetrameric Kv7.1 potassium channels. Here, we investigated whether allele-specific inhibition of mutant KCNQ1 by targeting a common variant can shift the balance towards increased incorporation of the wild-type allele to alleviate the disease in human-induced pluripotent stem-cell-derived cardiomyocytes (hiPSC-CMs). We identified the single nucleotide polymorphisms (SNP) rs1057128 (G/A) in KCNQ1, with a heterozygosity of 27% in the European population. Next, we determined allele-specificity of short-hairpin RNAs (shRNAs) targeting either allele of this SNP in hiPSC-CMs that carry an LQT1 mutation. Our shRNAs downregulated 60% of the A allele and 40% of the G allele without affecting the non-targeted allele. Suppression of the mutant KCNQ1 allele by 60% decreased the occurrence of arrhythmic events in hiPSC-CMs measured by a voltage-sensitive reporter, while suppression of the wild-type allele increased the occurrence of arrhythmic events. Furthermore, computer simulations based on another LQT1 mutation revealed that 60% suppression of the mutant KCNQ1 allele shortens the prolonged action potential in an adult cardiomyocyte model. We conclude that allele-specific inhibition of a mutant KCNQ1 allele by targeting a common variant may alleviate the disease. This novel approach avoids the need to design shRNAs to target every single mutation and opens up the exciting possibility of treating multiple LQT1-causing mutations with only two shRNAs.
Background: Cardiomyocyte progenitor cells (CMPCs) are a promising cell source for regenerative cell therapy to improve cardiac function after myocardial infarction. However, it is unknown whether undifferentiated CMPCs have arrhythmogenic risks. We investigate whether undifferentiated, regionally applied, human fetal CMPCs form a pro-arrhythmic substrate in co-culture with neonatal rat ventricular myocytes (NRVMs). Method: Unipolar extracellular electrograms, derived from micro-electrode arrays (8x8 electrodes) containing monolayers of NRVMs (control), or co-cultures of NRVMs and locally seeded CMPCs were used to determine conduction velocity and the incidence of tachy-arrhythmias. Micro-electrodes were used to record action potentials. Conditioned medium (Cme) of CMPCs was used to distinguish between coupling or paracrine effects. Results: Co-cultures demonstrated conduction slowing (5.6±0.3 cm/s, n=50) compared to control monolayers (13.4±0.4 cm/s, n=26) and monolayers subjected to Cme (13.7±0.6 cm/s, n=11, all p<0.001). Furthermore, co-cultures had a more depolarized resting membrane than control monolayers (-47.3±17.4 vs. - 64.8±7.7mV, p<0.001) and monolayers subjected to Cme (-64.4±8.1 mV, p<0.001). Upstroke velocity was significantly decreased in co-cultures and action potential duration was prolonged. The CMPC region was characterized by local ST-elevation in the recorded electrograms. The spontaneous rhythm was faster and tachy-arrhythmias occurred more often in co-cultured monolayers than in control monolayers (42.0 vs. 5.4%, p<0.001). Conclusion: CMPCs form a pro-arrhythmic substrate when co-cultured with neonatal cardiomyocytes. Electrical coupling between both cell types leads to current flow between a, slowly conducting, depolarized and the normal region leading to local ST-elevations and the occurrence of tachy-arrhythmias originating from the non-depolarized zone.
Dilated cardiomyopathy (DCM) is a major cause of heart failure (HF) defined by ventricular dilatation and systolic dysfunction. Although microRNAs (miRNAs) are known to affect HF development, little is known about the contribution of genetic variants in miRNAs or their precursors to the susceptibility or pathogenesis of DCM. We screened 1640 DCM cases for variants in cardiac miR-208a and miR-208b and their precursors. We identified four variants in the miR-208a pre-miRNA, which are present at very low frequencies in the general population. Two of these variants (+42G > T and +68G > T) alter a highly conserved nucleotide and the predicted pre-miRNA secondary structure. Both variants result in reduced mature miR-208a levels in overexpression experiments. The variant +42G > T also increased pre-miR-208a levels in these experiments, which indicates a maturation deficiency. Co-transfection of the overexpression constructs with a luciferase construct containing six miRNA binding sites revealed that both variants also impair repression of luciferase expression by miR-208a, indicative of also a loss of miR208a function. Together this indicates that these DCM-associated variants impair formation of mature miR208a. Combined with the role of miR-208a in cardiac contractility this suggests that variants +42G > T and +68G > T in pre-miR-208a may contribute to the DCM phenotype observed in these patients.
Dilated cardiomyopathy (DCM) caused by rare pathogenic variants in LMNA is a severe hereditary heart disease characterized by lethal arrythmias and progressive heart failure. Here, we describe a human induced pluripotent stem cell (hiPSC) line reprogrammed by non-integrative Sendai virus from dermal fibroblasts of a female DCM patient carrying the heterozygous LMNA p.Q493X rare pathogenic variant. The obtained hiPSCs have a normal karyotype, show high expression of pluripotency markers at the mRNA and protein level, and are able to differentiate into derivatives of all three germ layers. Therefore, this newly-generated hiPSC line enables modeling of DCM caused by rare pathogenic variants in LMNA.