ATP1A3 Variants, Variably Penetrant Short QT Intervals, and Lethal Ventricular Arrhythmias

Abstract

Importance: Alternating hemiplegia of childhood (AHC) is a disorder that can result from pathogenic variants in ATP1A3-encoded sodium-potassium adenosine triphosphatase alpha 3 (ATP1A3). While AHC is primarily a neurologic disease, some individuals experience sudden unexplained death (SUD) potentially associated with cardiac arrhythmias.

Objective: To determine the impact of ATP1A3 variants on cardiac electrophysiology and whether lethal ventricular arrhythmias are associated with SUD in patients with AHC.

Design, setting, and participants: In this international, multicenter case-control study from 12 centers across 10 countries, patients with AHC were grouped by ATP1A3 variant status (positive vs negative) and into subgroups with the most common AHC variants (D801N, E815K, G947R, and other). A healthy control cohort was established for comparison. Blinded, manual measurements of QT intervals and corrected QT interval (QTc) were performed independently by 2 pediatric cardiac electrophysiologists. Induced pluripotent stem cell cardiomyocytes were derived from patients with AHC who were positive for the D801N variant of ATP1A3 (iPSC-CMD801N cells). Data analysis was performed from April to June 2022.

Exposure: Presence of ATP1A3 variant.

Main outcomes and measures: The primary outcome was QTc. Outcomes, including survival, were abstracted and variants were mapped on cryogenic electron microscopy structure maps. iPSC-CMD801N cells were used to validate ventricular repolarization and arrhythmic susceptibility in vitro.

Results: Among the 222 individuals included (148 with AHC and 74 control), the mean (SD) age at diagnostic electrocardiography was 11.0 (9.4) years and 119 (54%) were female. The cohort with AHC consisted of 148 largely unrelated probands (mean [SD] age at diagnostic electrocardiography, 11.5 [10.5] years). Of these, 123 individuals were ATP1A3 genotype positive, including 35 (28%) with the D801N variant, 21 (17%) with the E815K variant, 8 (7%) with the G947R variant, and 8 (7%) with a loss-of-function variant. Probands with the D801N variant had shorter mean (SD) QTcs (381.8 [36.6] milliseconds; 24 [69%] with QTc <370 milliseconds) compared with those who had the E815K variant (393.6 [43.1] milliseconds; P = .001; 4 [19%] with QTC <370 milliseconds), the G947R variant (388.4 [26.5] milliseconds; P = .02; 1 [13%] with QTc <370 milliseconds), a loss-of-function variant (403.0 [33.5] milliseconds; P < .001; 1 [13%] with QTc <370 milliseconds), all other variants (387.8 [37.1] milliseconds; P < .001; 44 [86%] with QTc <370 milliseconds), and healthy controls (415.4 [21.0] milliseconds; P < .001; 0 with QTc <370 milliseconds). Three D801N-positive individuals had a major cardiac event, compared with 0 major cardiac events in all other individuals (P = .02). The D801N variant and 4 rare variants (D805N, P323S, S772R, and C333F) found in individuals with the shortest QTcs localized to the potassium-binding domain of ATP1A3. IPSC-CMD801N lines demonstrated shortened action potential duration, higher mean diastolic potential, and delayed afterdepolarizations compared with controls.

Conclusions and relevance: Nearly 70% of individuals with D801N variants of ATP1A3 had short QTcs (<370 milliseconds), with an association between ventricular arrhythmias and cardiac arrest. This may underlie the SUD etiology in AHC.

Figure 1.. Design of Multicenter, International Study

A, Flow diagram showing the design of this study and the inclusion and exclusion criteria used to establish the patient cohort. B, Distribution of ATP1A3 genotype–positive vs ATP1A3 genotype–negative individuals within the study cohort. C, Variant breakdown of gene-positive individuals. AHC indicates alternating hemiplegia of childhood; IAHCRC, International Consortium for the Research on Alternating Hemiplegia of Childhood; LOF, loss of function; QTc, corrected QT interval.

Figure 2.. Comparison of Corrected QT Interval (QTc) Measurements by Genotype and ATP1A3 Variant Status

A, Representative electrocardiogram (ECG) tracings taken from lead V5 of representative probands with various ATP1A3 genotypes. B, QTc for ATP1A3 genotype–positive vs ATP1A3 genotype–negative individuals. Horizontal center lines indicate mean; error bars, SEM. C, QTc for each major variant in ATP1A3. Horizontal center lines indicate mean; error bars, SEM. D and E, Penetrance with percentage of individuals with QTc shorter than 370 milliseconds (D) and shorter than 350 milliseconds (E) for each major variant in ATP1A3. LOF indicates loss of function.

Figure 3.. Cardiac Event Frequency and Event-Free Survival

A, Percentages of cardiac events meeting the composite outcome of ventricular tachycardia or ventricular fibrillation arrest, sudden cardiac arrest, suspected cardiogenic syncope, or implantable cardioverter-defibrillator appropriately discharged and cardiac events as a function of total anesthesia exposures. Only individuals with the ATP1A3 D801N variant experienced cardiac outcomes in our cohort. The D801N group had an incidence of 0.6 events/100 life-years; the other groups had an incidence of 0 events/100 life-years. LOF indicates loss of function. B, Kaplan-Meier curve depicting freedom from the composite cardiac outcome showing that individuals with the ATP1A3 D801N variant were significantly more likely to experience one of the composite cardiac outcomes compared with individuals with all other ATP1A3 variants (P = .02). Shaded area shows the 95% CI for the D801N variant.

Figure 4.. Structural Mapping of ATP1A3 Amino Acid Positions Whose Missenses Are Associated With Variable Penetrance of Shortened Corrected QT Interval (QTc)

Structural mapping uses the crystal structure of human alpha 3 sodium-potassium adenosine triphosphatase (ATPase) in its potassium-occluded state. A, Crystal structure with the beta 1 subunit in blue, FXYD protein in purple, and the alpha 3 subunit in gray. The shortened QTc–associated amino acids D801, D805, P323, S772, and C333 are red. B, Closer view of the region of interest on the human alpha 3 sodium-potassium ATPase. C and D, Transmembrane region in anterior (C) and posterior (D) views showing where the variants resulting in shortened QTc with variable penetrance are located.

Figure 5.. Shorter Action Potential Duration and Predisposition to Delayed Afterdepolarizations (DADs) With ATP1A3 D801N Variant

A, Pedigrees of the proband donor of the induced pluripotent stem cell (iPSC) line with an ATP1A3 D801N missense variant (iPSC^D801N) and an unrelated proband donor of the iPSC^D801N-2 line with an ATP1A3 D801N variant. SQT indicates shortened QT interval; VF, ventricular fibrillation; arrowhead, proband; square, male; circle, female; open symbol, unaffected; shaded symbol, affected; slash, deceased; −, negative; +, positive. B, Representative action potential tracing of spontaneously depolarizing wild-type (WT) iPSC cardiomyocyte (iPSC-CM^WT) and iPSC-CM^D801N cells. cWT indicates clustered regularly interspaced short palindromic repeats (CRISPR)–corrected WT; arrowheads, DADs. C, Representative action potential tracing. D and E, Action potential duration (APD) at 50% repolarization (APD₅₀) (D) and 90% repolarization (APD₉₀) (E). F, Mean diastolic potential (MDP). D-F, Horizontal center lines indicate mean; error bars, SEM. G, Percentage of cells with DADs in iPSC-CM^cWT (0/5 cells [0%; 95% CI, 0.0%-43.4%]), iPSC-CM^D801N (2/9 cells [22.2%; 95% CI, 3.9%-54.7%]), iPSC-CM^WT-2 (0/24 cells [0%; 95% CI, 0.0%-10.7%]), and iPSC-CM^D801N-2 (3/11 cells [27.3%; 95% CI, 9.7%-56.5%]). CIs were calculated using the Wilson/Brown method. Mann-Whitney test was conducted to assess statistical significance. ^aP < 05. ^bP < .01.