Modeling and Genome-Editing Brugada Syndrome in a Dish

More than a quarter century ago, a report on sudden deaths among Thai construction workers, often in their sleep (1), mobilized an organized effort to understand the fatal disease. This inherited arrhythmia with heightened risk of ventricular fibrillation in apparently healthy, young or middle-aged adults was quickly recognized as a distinct clinical entity and named Brugada syndrome (BrS) (2). Fast forward some 20 years, and the recent advent of human induced pluripotent stem cell (hiPSC) technology has bulldozed its way into BrS with a prospect to model a human disease in a dish with a patient’s own cells. And the potent genome-editing technologies promise to heal a defective gene back to its wild-type sequence, perhaps in the same dish (3). In this issue of the Journal, Liang et al. (4) report exactly that.

The authors began by enlisting 4 subjects; 1 patient (BrSp1) with electrocardiographic characteristics of BrS, whose right precordial leads exhibited an ST-segment elevation and a T-wave inversion at rest; 1 patient (BrSp2) with a drug-induced manifestation of the Brugada electrocardiograms; and 2 control subjects with no known heart rhythm problems. The 2 patients exhibited revealing signs of severe ventricular arrhythmias, including recurrent syncope and family members with sudden death. Genetic screen identified mutations in the cardiac Na⁺ channel gene, SCN5A, which encodes the α-subunit of the cardiac Na⁺ channel, Nav1.5. BrSp1 carried 2 mutations (R620H/R811H) in SCN5A. The authors have previously reported that the R811H mutation makes Nav1.5 channels to recover from inactivation less efficiently, thereby decreasing the peak Na⁺ current density (5). BrSp2 carried a deletion mutation in Nav1.5, truncating the channel at the pore loop of domain III. The mutation deletes the entire domain IV, and thus is expected to render the Na⁺ channel nonfunctional.

At first glance, the hiPSC platform may seem unfit to model a disease such as BrS, which is viewed as a problem that originates at the tissue or organ level. Specifically, epicardial and transmural dispersion of repolarization or slowing of conduction at the right ventricular outflow tract are thought to provide arrhythmic substrates (6,7). Nevertheless, the authors proceeded to derive iPSC lines from the patients’ skin biopsies, and then differentiated the hiPSC lines to cardiac myocytes to investigate the electrophysio-logical phenotypes of the de novo cardiac myocytes. As expected, the Na⁺ current (I_Na) densities from BrSp1 and BrSp2 cardiomyocytes were dramatically reduced compared to those from control cardiomyocytes. This was accompanied by reduced Nav1.5 channel protein levels in BrSp1 and BrSp2 cardiomyocytes compared with control, perhaps reflecting the myocytes’ inherent mechanisms to degrade the defective proteins (8,9). The low I_Na density contributed to slowing of the action potential (AP) upstroke velocity in BrSp1 and BrSp2 cardiomyocytes. These were largely anticipated from the incredibly fast activation kinetics of Nav1.5 channels, most of which open within a couple of milliseconds upon initial depolarization (10).

The less intuitive but perhaps more interesting findings are that the cardiomyocytes derived from the BrS patients exhibited predisposition to arrhythmias at the single cell level. The AP from both BrSp1 and BrSp2 cardiomyocytes displayed a high variability in beat-to-beat intervals and triggered activities, which were largely absent in the control myocytes. The single-cell arrhythmias were accompanied with dysregulation of intracellular Ca²⁺ handling as well as decreased transcript levels of KCND3 and KCNJ2, which are the molecular correlates for transient outward K⁺ current (I_to) and inward rectifier K⁺ current (I_K1), respectively. In addition, the transcript level of SCN5A was lower in cardiomyocytes derived from the BrS patients compared with control. Although the SCN5A mutations in BrSp1 and BrSp2 cannot directly explain why the repolarization ion channel gene expression levels became lower, disorders of repolarization, including those of I_to and I_K1, are known substrates for ventricular arrhythmias (11). Indeed, deliberate suppression of I_K1 suffices to elicit spontaneous activity from quiescent adult ventricular myocytes (12). Upon demonstrating the strong genotype-phenotype correlation, the authors employed CRISPR/Cas9-mediated genome editing technology to correct the SCN5A deletion mutation in the BrSp2 iPSCs. The conversion to SCN5A wild-type lead to restoration of electrical properties, including normalization of beat-to-beat interval, a fast AP upstroke velocity, and Ca²⁺ transients that are as robust as those from the control myocytes. But, is this all we can learn from this single-cell study?

The first putative causal mutations for BrS were found in SCN5A in 1998 (13). Since then, more than 300 rare variants in SCN5A have been reported, although the contribution of other genes remains small (14,15). This seems to justify the framing of BrS as a monogenic disease with autosomal dominant mode of inheritance. However, hardly any of the large BrS family pedigrees is strictly dictated by Mendelian inheritance. To the contrary, most familial forms of the BrS show incomplete penetrance and remain genetically undiagnosed (16). Accordingly, the SCN5A mutations are unlikely to be the sole factor in determining the Brugada phenotype. In this regard, it would be motivating to know if the KCND3, KCNJ2, and SCN5A expression levels have returned back to their normal levels upon disease-correcting genome modification. That is, the gene-editing step affords us an opportunity to examine whether the collateral changes in those gene products are indeed compensatory mechanisms or unique effectors that are independent from the SCN5A mutations. This may lead to new correlative or predictive markers for BrS, a disease notorious for its difficulties in prognosis (17). As the authors acknowledge, the hiPSC model of BrS suffers from the usual list of drawbacks inherent to pluripotent stem cell–derived cardiomyocytes. Still, when combined with genome-editing tools, the hiPSC disease model is well positioned to move the field forward—even for BrS.