Alleviating the Effects of Short QT Syndrome Type 3 by Allele-Specific Suppression of the KCNJ2 Mutant Allele

Figure 1. General topology of a Kir2.1 subunit with each of the four gain-of-function mutations that have been associated with the SQT3 syndrome. Each subunit consists of two membrane-spanning α-helical segments (M1 and M2) separated by the pore-forming domain P. The amino and carboxyl termini (NH2 and COOH, respectively) are located intracellularly. E1 and E2 indicate the two extracellular loops. The D172N mutation is located on the M2 inner helix. Four Kir2.1 α-subunits oligomerize to form a functional Kir2.1 potassium channel.

Figure 2. (A) Potential configurations of the tetrameric Kir2.1 channel in the case of a heterozygous mutation in KCNJ2. The rectangles indicate the five configurations with at most one mutant subunit (grey for exactly one mutant subunit; black for wild-type subunits only). (B) Binomial distribution of IK1 channels with 0–4 mutant subunits. Without suppression of the mutant allele, the distribution is symmetrical (cyan bars), and the fractions add up to 100%. With a 60% suppression of the mutant allele, the distribution is skewed (orange bars), and the fractions add up to 70% to account for the reduced channel density upon suppression and to allow a direct comparison of the amount of channels in each configuration.

Figure 3. (A,B) IK1 current–voltage relationships of the BPS2020 and ToR–ORd models. (C–F) Effects of each of the SQT3 mutations in KCNJ2 relative to the wild-type current. The maximum outward amplitude of the WT current is set to 100%. Note the difference in the ordinate scales of panels (C,D) on the one hand and panels (E,F) on the other hand.

Figure 4. Effects of the heterozygous D172 mutation in KCNJ2 (WT/D172N) on the action potential of a human ventricular cardiomyocyte, stimulated at 1 Hz, in the endocardial (‘endo’), midmyocardial (‘mid’), and epicardial (‘epi’) versions of the BPS2020 and ToR–ORd models, and effects of suppression of the mutant allele (dashed lines). (A–C) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the BPS2020 model. (D–F) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the ToR–ORd model. Vm denotes the membrane potential. Suppression settings “A” and “B” as set out in Section 2.1.

Figure 5. Effects of the heterozygous E299V mutation in KCNJ2 (WT/E299V) on the action potential of a human ventricular cardiomyocyte, stimulated at 1 Hz, in the endocardial (‘endo’), midmyocardial (‘mid’), and epicardial (‘epi’) versions of the BPS2020 and ToR–ORd models, and effects of the suppression of the mutant allele (dashed lines). (A–C) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the BPS2020 model. (D–F) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the ToR–ORd model. Suppression settings “A” and “B” as set out in Section 2.1.

Figure 6. Effects of the heterozygous M301K mutation in KCNJ2 (WT/M301K) on the action potential of a human ventricular cardiomyocyte, stimulated at 1 Hz, in the endocardial (‘endo’), midmyocardial (‘mid’), and epicardial (‘epi’) versions of the BPS2020 and ToR–ORd models, and effects of the suppression of the mutant allele (dashed lines). (A–C) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the BPS2020 model. (D–F) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the ToR–ORd model. Suppression settings “A” and “B” as set out in Section 2.1.

Figure 7. Effects of the heterozygous K346T mutation in KCNJ2 (WT/K346T) on the action potential of a human ventricular cardiomyocyte, stimulated at 1 Hz, in the endocardial (‘endo’), midmyocardial (‘mid’), and epicardial (‘epi’) versions of the BPS2020 and ToR–ORd models, and effects of the suppression of the mutant allele (dashed lines). (A–C) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the BPS2020 model. (D–F) Endo (left), mid (middle), and epi (right) action potentials (top) and associated IK1 (bottom) in the ToR–ORd model. Suppression settings “A” and “B” as set out in Section 2.1.

Figure 8. APD restitution curves in the case of the heterozygous D172N mutation obtained with an S1-S2 stimulus protocol. After a train of S1 stimuli at 1 Hz, an S2 stimulus was applied at various diastolic intervals, and the APD upon the S2 stimulus was determined. (A,B) APD restitution curves obtained with the (A) endo and (B) mid versions of the BPS2020 model. (C,D) APD restitution curves obtained with the (C) endo and (D) mid versions of the ToR–ORd model. To allow for a direct comparison, the APD range for the two endo models is identical, as is the APD range for the two mid models.

Figure 9. APD restitution curves in the case of the heterozygous E299V mutation obtained with an S1-S2 stimulus protocol. After a train of S1 stimuli at 1 Hz, an S2 stimulus was applied at various diastolic intervals, and the APD upon the S2 stimulus was determined. (A,B) APD restitution curves obtained with the (A) endo and (B) mid versions of the BPS2020 model. (C,D) APD restitution curves obtained with the (C) endo and (D) mid versions of the ToR–ORd model. To allow for a direct comparison, the APD range for the two endo models is identical, as is the APD range for the two mid models.

Figure 10. Effects of the heterozygous E299V mutation on the action potentials obtained during 1 Hz stimulation in simulations in which the IK1 densities of the BPS2020 and ToR–ORd models were matched by either scaling down the IK1 density of the ToR–ORd model by a factor of 3.2012 or scaling up the IK1 density of the BPS2020 model by this factor. (A) Membrane potential (Vm; top) and IK1 (bottom) of the default BPS2020 model. (B) Vm and IK1 of the ToR–ORd model with its IK1 density scaled down. (C) Vm and IK1 of the default ToR–ORd model. (D) Vm and IK1 of the BPS2020 model with its IK1 density scaled up. Note the difference in the IK1 ordinate scales between panels (A,B) on the one hand and panels (C,D) on the other hand.

Figure 11. Effect of replacing the original IK1 formulations of the BPS2020 and ToR–ORd models with Kir2.1-based formulations. (A) Current–voltage relationship of the original IK1 of the BPS2020 model and the Kir2.1-based one used in our simulations. (B) Vm and IK1 of the BPS2020 model during 1 Hz stimulation. (C) Current–voltage relationship of the original IK1 of the ToR–ORd model and the Kir2.1-based one used in our simulations. (D) Vm and IK1 of the ToR–ORd model during 1 Hz stimulation.