Hyperactivation of platelet‐derived growth factor signalling contributes to arrhythmogenesis in Brugada syndrome

We performed a comprehensive study to assess the pathogenicity of a novel transient receptor potential melastatin 4 (TRPM4) mutation and to pinpoint underlying molecular mechanisms using an induced pluripotent stem cell-derived cardiomyocyte (iPSC-CM) model. Patient-specific iPSC-CMs exhibited arrhythmic phenotype manifesting delayed afterdepolarizations (DADs) and paroxysmal cellular flutter (PCF), which were rescued by correction of the causal mutation. T262M conferred impaired TRPM4 channel function by enhanced ubiquitination for protein degradation via the lysosomal pathway, resulting in abnormal Ca2+ cycling and elevated diastolic intracellular Ca2+ ([Ca2+]i). Mechanistically, hyperactivation of platelet-derived growth factor receptor β (PDGFRB) signalling derived arrhythmic events in diseased iPSC-CMs. On the contrary, pharmacological and genetic inhibition of PDGFRB effectively restored diastolic [Ca2+]i and rescued the arrhythmic phenotype in T262M myocytes. Brugada syndrome (BrS) is an inherited arrhythmogenic disease featured by a high risk of sudden cardiac death.1 TRPM4 mutations have been reported to be associated with BrS,2-4 whereas the underlying mechanisms remain unknown. In this study, we recruited a 21-year-old male patient who is asymptomatic and no abnormality was found by echocardiography. However, his electrocardiogram showed a characteristic type 2 BrS pattern, with ST-segment morphology representing saddleback-type elevation in lead V2 (Figure S1A). In addition, we further monitored the electrocardiogram in lead V1 and V2 from one or two intercostal spaces higher than the standard position (Figure S1B). The genetic screening revealed a single missense mutation (c.785C > T, p.T262M) in TRPM4 (Figure S1C–E). The mutation, located at the N terminus of the TRPM4 channel, is highly conserved among spices and is graded as a variant of uncertain significance (Figure S1F,G). Our healthy control subject was a 21-year-old female. Skin fibroblasts were reprogrammed using nonintegrated Sendai virus and iPSCs were successfully generated and characterized (Figures S1H–L, S2 and S3). Genetic sequencing confirmed that T262M mutation was present in patient iPSCs but not in controls (Figure S1M). The iPSC-CMs were subsequently generated by a small molecule-based monolayer differentiation protocol (Figures S4 and S5). Single-cell patch clamp recordings revealed a uniform and rhythmic action potential (AP) profile in control iPSC-CMs (Figure 1A). However, arrhythmic waveforms were seen in a large proportion of patient iPSC-CMs, manifesting two distinct phenotypes: the more common phenotype was DAD; the other rare but more severe phenotype was PCF (Figure 1B,C,G). Moreover, we observed significantly increased peak interval variability and shortened AP duration (APD) in patient iPSC-CMs (Figure 1H,I and Table S1). We next generated isogenic control lines by CRISPR/Cas9-mediated genome editing technology (Figure S2 and S6). The gene-corrected (GC) iPSC-CMs showed a dramatic reduction of arrhythmic incidence and resembled the AP profile of controls (Figure 1D,G–I, Figures S4, S5 and Table S1). TRPM4 knockout (KO) iPSC lines were also generated by CRISPR/Cas9 (Figures S2, S7 and S8). KO iPSC-CMs recapitulated abnormal AP phenotype of patient iPSC-CMs (Figure 1E–I, Figures S4, S5 and Table S1). Taken together, these results demonstrate that TRPM4 T262M is a pathologic mutation that causes the arrhythmic phenotype. To assess if T262M gave rise to TRPM4 channel dysfunction, we next performed patch clamp on human embryonic kidney 293T cells transiently expressing TRPM4 (Figure 2A). The current density was significantly reduced in T262M channels as compared to wildtype (WT) (Figure 2B,C). Through molecular modelling, we observed no significant change of TRPM4 channel structure affected by T262M (Figure S9). Biotinylation assay revealed that total and surface expression levels of TRPM4 protein were significantly decreased in T262M channels, whereas the ratio of surface-to-total expression levels remained unchanged (Figure 2D–G). Consistently, the endogenous TRPM4 protein expression was markedly decreased in patient iPSC-CMs (Figure 2H,I). Given that the TRPM4 channel is activated by [Ca2+]i,5 we sought to investigate if T262M affected TRPM4 channel activation upon [Ca2+]i stimulation. Patch clamp recordings were performed in inside-out mode, allowing the cytosolic side of the patch perfused by escalated [Ca2+]i. We observed significantly reduced [Ca2+]i-activated TRPM4 current density and right-shifted current-[Ca2+]i curve in T262M channels, suggesting a weakened sensitivity to [Ca2+]i (Figure 2J–M). Moreover, we observed a markedly elevated level of ubiquitinated TRPM4 protein in T262M channels using the ubiquitin antibody P4D1 (Figure S10A,B). Inhibition of lysosome pathway by chloroquine effectively rescued T262M-induced down-regulation of TRPM4 protein and reduction of TRPM4 currents (Figure S10C–I). Collectively, these results indicate that T262M confers impaired TRPM4 channel function by enhanced ubiquitination for protein degradation via the lysosomal pathway. Alterations in Ca2+ cycling are a common trigger of cardiac arrhythmias.6 More attention has been attracted to the roles of Ca2+ signalling in arrhythmogenic mechanisms of BrS.7 To test whether T262M affects Ca2+ homeostasis, Ca2+ imaging was performed using fura-2 AM dye to ratiometrically record Ca2+ transients in iPSC-CMs (Figure S11A–C). Patient iPSC-CMs exhibited "arrhythmia-like" irregular transients, dysregulation of Ca2+ cycling and elevation of diastolic [Ca2+]i (Figure S11D–L and Table S2). Interestingly, L-type Ca2+ current density was markedly increased in T262M iPSC-CMs, and treatment of verapamil in T262M iPSC-CMs drastically ameliorated the incidence of irregular transients (Figure S11M–T). These results demonstrate that disrupted Ca2+ homeostasis is associated with arrhythmic phenotype caused by T262M. To understand the molecular mechanisms of how TRPM4 T262M causes arrhythmic phenotype, we performed genome-wide RNA sequencing by comparing control, KO, patient and GC iPSC-CMs (Figure 3A–C). Gene ontology analysis revealed that differentially expressed genes were enriched in ion channel-, Ca2+ signalling-, cardiac action potential and conduction, and PDGF signalling-related terms (Figure 3D). Interestingly, we found greatly enhanced PDGFRB expression in KO or patient iPSC-CMs (Figure 3D–H). Previous studies have reported that PDGFRB signalling regulates cardiomyocyte proliferation and myocardial regeneration, and over-activation of PDGFRB signalling is closely associated with atrial fibrillation and dilated cardiomyopathy.8-10 Notably, treatment of PDGFRB inhibitors or knockdown of PDGFRB expression in patient iPSC-CMs largely reduced PDGFRB protein expression, greatly attenuated proarrhythmic activities, and restored elevated diastolic [Ca2+]i (Figure 4). Altogether, these results suggest that hyperactivation of PDGFRB signalling contributes to arrhythmogenesis in TRPM4-related BrS. In conclusion, genome editing of iPSC-CMs can offer a precision medicine approach for identifying pathogenic mutation of BrS in a dish. More importantly, our findings reveal novel molecular mechanisms and potential therapeutic targets of TRPM4-related BrS (Figure S12). We would like to thank the core facility of Zhejiang University Institute of Translational Medicine for assistance with flow cytometry and confocal microscopy experiments. This work was supported by National Key R&D Program of China (No. 2017YFA0103700) (P.L.), National Natural Science Foundation of China (No. 81922006, 81870175, 31571528) (P.L.), and Natural Science Foundation of Zhejiang Province (No. LD21H020001, LR15H020001) (P.L.). P.L. would like to thank Natalie Liang and Michael Liang for their encouragement and consistent support. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.