A deep learning platform to assess drug proarrhythmia risk.

Drug safety initiatives have endorsed human iPSC-derived cardiomyocytes (hiPSC-CMs) as an in vitro model for predicting drug-induced cardiac arrhythmia. However, the extent to which human-defined features of in vitro arrhythmia predict actual clinical risk has been much debated. Here, we trained a convolutional neural network classifier (CNN) to learn features of in vitro action potential recordings of hiPSC-CMs that are associated with lethal Torsade de Pointes arrhythmia. The CNN classifier accurately predicted the risk of drug-induced arrhythmia in people. The risk profile of the test drugs was similar across hiPSC-CMs derived from different healthy donors. In contrast, pathogenic mutations that cause arrhythmogenic cardiomyopathies in patients significantly increased the proarrhythmic propensity to certain intermediate and high-risk drugs in the hiPSC-CMs. Thus, deep learning can identify in vitro arrhythmic features that correlate with clinical arrhythmia and discern the influence of patient genetics on the risk of drug-induced arrhythmia.

Unfolded Protein Response as a Compensatory Mechanism and Potential Therapeutic Target in PLN R14del Cardiomyopathy.

BACKGROUND: Phospholamban (PLN) is a critical regulator of calcium cycling and contractility in the heart. The loss of arginine at position 14 in PLN (R14del) is associated with dilated cardiomyopathy with a high prevalence of ventricular arrhythmias. How the R14 deletion causes dilated cardiomyopathy is poorly understood, and there are no disease-specific therapies. METHODS: We used single-cell RNA sequencing to uncover PLN R14del disease mechanisms in human induced pluripotent stem cells (hiPSC-CMs). We used both 2-dimensional and 3-dimensional functional contractility assays to evaluate the impact of modulating disease-relevant pathways in PLN R14del hiPSC-CMs. RESULTS: Modeling of the PLN R14del cardiomyopathy with isogenic pairs of hiPSC-CMs recapitulated the contractile deficit associated with the disease in vitro. Single-cell RNA sequencing revealed the induction of the unfolded protein response (UPR) pathway in PLN R14del compared with isogenic control hiPSC-CMs. The activation of UPR was also evident in the hearts from PLN R14del patients. Silencing of each of the 3 main UPR signaling branches (IRE1, ATF6, or PERK) by siRNA exacerbated the contractile dysfunction of PLN R14del hiPSC-CMs. We explored the therapeutic potential of activating the UPR with a small molecule activator, BiP (binding immunoglobulin protein) inducer X. PLN R14del hiPSC-CMs treated with BiP protein inducer X showed a dose-dependent amelioration of the contractility deficit in both 2-dimensional cultures and 3-dimensional engineered heart tissues without affecting calcium homeostasis. CONCLUSIONS: Together, these findings suggest that the UPR exerts a protective effect in the setting of PLN R14del cardiomyopathy and that modulation of the UPR might be exploited therapeutically.

Metabolic Maturation Media Improve Physiological Function of Human iPSC-Derived Cardiomyocytes.

Induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) have enormous potential for the study of human cardiac disorders. However, their physiological immaturity severely limits their utility as a model system and their adoption for drug discovery. Here, we describe maturation media designed to provide oxidative substrates adapted to the metabolic needs of human iPSC (hiPSC)-CMs. Compared with conventionally cultured hiPSC-CMs, metabolically matured hiPSC-CMs contract with greater force and show an increased reliance on cardiac sodium (Na+) channels and sarcoplasmic reticulum calcium (Ca2+) cycling. The media enhance the function, long-term survival, and sarcomere structures in engineered heart tissues. Use of the maturation media made it possible to reliably model two genetic cardiac diseases: long QT syndrome type 3 due to a mutation in the cardiac Na+ channel SCN5A and dilated cardiomyopathy due to a mutation in the RNA splicing factor RBM20. The maturation media should increase the fidelity of hiPSC-CMs as disease models.