Screening for acute IKr block is insufficient to detect torsades de pointes liability: role of late sodium current

Abstract

Background: New drugs are routinely screened for IKr blocking properties thought to predict QT prolonging and arrhythmogenic liability. However, recent data suggest that chronic (hours) drug exposure to phosphoinositide 3-kinase inhibitors used in cancer can prolong QT by inhibiting potassium currents and increasing late sodium current (INa-L) in cardiomyocytes. We tested the extent to which IKr blockers with known QT liability generate arrhythmias through this pathway.

Methods and results: Acute exposure to dofetilide, an IKr blocker without other recognized electropharmacologic actions, produced no change in ion currents or action potentials in adult mouse cardiomyocytes, which lack IKr. By contrast, 2 to 48 hours of exposure to the drug generated arrhythmogenic afterdepolarizations and ≥15-fold increases in INa-L. Including phosphatidylinositol 3,4,5-trisphosphate, a downstream effector for the phosphoinositide 3-kinase pathway, in the pipette inhibited these effects. INa-L was also increased, and inhibitable by phosphatidylinositol 3,4,5-trisphosphate, with hours of dofetilide exposure in human-induced pluripotent stem cell-derived cardiomyocytes and in Chinese hamster ovary cells transfected with SCN5A, encoding sodium current. Cardiomyocytes from dofetilide-treated mice similarly demonstrated increased INa-L and afterdepolarizations. Other agents with variable IKr-blocking potencies and arrhythmia liability produced a range of effects on INa-L, from marked increases (E-4031, d-sotalol, thioridazine, and erythromycin) to little or no effect (haloperidol, moxifloxacin, and verapamil).

Conclusions: Some but not all drugs designated as arrhythmogenic IKr blockers can generate arrhythmias by augmenting INa-L through the phosphoinositide 3-kinase pathway. These data identify a potential mechanism for individual susceptibility to proarrhythmia and highlight the need for a new paradigm to screen drugs for QT prolonging and arrhythmogenic liability.

Keywords: KCNH2 potassium channel, human; Nav1.5 voltage-gated sodium channel; phosphatidylinositol 3-kinases; potassium channels.

Figure 1

Acute and chronic effects of dofetilide and 4-aminopyridine (4-AP) on action potentials and potassium current in mouse. A. After an acute (15 minute) exposure to dofetilide (dof), mouse cardiomyocyte action potentials were superimposable on those recorded in control. B. Acute exposure to 4-AP prolonged action potential duration. C. Effect of dofetilide and 4-AP on potassium currents. There was no change with dofetilide but a pronounced suppression with 4-AP, as expected from previous reports. Note that the signature of I_Kr, a deactivating tail current, is absent in these cells and that the major current is the transient outward potassium current (I_TO). D. Summary of the effects of dofetilide and 4-AP on I_TO (n=6–7). P-values are from t-tests assuming unequal variance, and are corrected for 3 statistical tests. These experiments were conducted in ventricular cardiomyocytes isolated from female C57 mice.

Figure 2

Effect of exposure to dofetilide for 5 hours in mouse cardiomyocytes. A. Chronic dofetilide prolonged action potentials at a stimulation rate of 1 Hz. An example is shown in the top panel, and summary data on APD at 50% and 90% repolarization (APD₅₀ and APD₉₀) for acute exposure to dofetilide and 4-AP (from Figure 1) and 5-hour exposure to dofetilide are shown below (n=6–8 each). P-values are from t-tests assuming unequal variance, and are corrected for 4 statistical tests. B. Example of action potentials recorded at a slow stimulation rate (0.1 Hz) 5 hours after isolation with no drug exposure. No abnormalities are seen. Traces recorded immediately after break-in (labeled 1^st min) and 5 minutes later (5^th min) are shown in this panel and in panels C and D. C. Example of action potentials recorded 5 hours after isolation with exposure to dofetilide. The insets show action potential prolongation and afterdepolarizations (small arrows). D. Example of action potentials recorded 5 hours after isolation with exposure to dofetilide, and with phosphatidylinositol 3,4,5-trisphosphate (PIP3) in the pipette. The insets show action potential prolongation at minute 1, and then action potential shortening by minute 5. There are no afterdepolarizations. E. Summary data showing frequency of afterdepolarizations in individual experiments. A t-test assuming unequal variance, and adjusted for 3 tests, was used for comparisons between groups.

Figure 3

Chronic dofetilide exposure increases late sodium current in mouse ventricular myocytes. A. Family of potassium currents in mouse ventricular myocytes 5 hours after isolation in the absence of drug (control, left) or the presence of dofetilide (right). The summary current-voltage relations (n=6 each) are shown below. There was no effect of dofetilide on potassium current in these cells. B. Examples of late sodium current recorded 5 hours after isolation in the absence of drug (control), in the presence of dofetilide, or in the presence of dofetilide with PIP3 in the pipette. C. Individual and summary data. PIP3 almost completely inhibited the augmentation produced by dofetilide (data after 200 msec at −30 mV; n=6–10 each). P-values were computed using a t-test assuming unequal variances, and are adjusted for 3 statistical tests.

Figure 4

Augmented dofetilide-induced late sodium current (I_Na-L) in CHO cells. A. Time-dependence of dofetilide effect (n=6–8 each). Median and inter-quartile ranges are shown in the graph. Exposure to drug caused a time-dependent increase in I_Na-L that was statistically significant as early as 2 hours. The figure also shows that there was no effect on I_Na-L of including PIP3 in the pipette under control conditions, but that at 5 and 48 hours, significant inhibition of I_Na-L was seen. P-values were computed using a Mann-Whitney U test, and are adjusted for 10 statistical tests. B. Concentration dependence of the dofetilide effect, yielding an IC₅₀ of 103.4±5.4 nM (n=6–8 per data point). These data were generated with 48 hour drug exposures. There was a statistically significant increase in I_Na-L with even 10 nM (p<0.05). C. E-4031 and d-sotalol, previously considered to be specific I_Kr blockers, also increased I_Na-L as did erythromycin and thioridazine. By contrast, there was only a minor effect of haloperidol and there was no effect of either moxifloxacin or verapamil. Median and inter-quartile ranges are shown in the graph (Mann-Whitney U test, adjusted for 17 tests). D. Increased I_Na-L and peak I_Na induced by 48 hour exposure to dofetilide were both blocked by the selective sodium channel toxin blocker tetrodotoxin (TTX), and the effect was greater on I_Na-L (bottom right) than on peak current (bottom left). A paired t-test assuming unequal variance was used for data comparisons.

Figure 5

Action potential duration and late current are increased after dofetilide administration in vivo. A–C. Representative action potentials (A and B) and summary data (C) at 5 and 1 Hz in mouse ventricular myocytes recorded at day 5 after dofetilide (30 mg/Kg, intraperitoneally). T-tests (equal variance) were used for data comparisons. D–F. Representative action potentials (D and E) and summary data (F) at a slower stimulation rate (0.5 Hz). The data were collected from six female mice (3 for control and 3 for dofetilide). The numbers of cells in each data group are shown in panels C and F.

Figure 6

Five hour drug exposures to cardiomyocytes derived from human induced pluripotent stem cells (hiPSC-CMs). A. An example of a small I_Na-L in these cells. B. An example of current recorded after exposure to dofetilide. C. Effect of including PIP3 in the pipette solution in a dofetilide-exposed cell. The first and 20^th traces during stimuli at 0.2 Hz are shown. D. An example of a current recorded after exposure to moxifloxacin. E. Summary data. Median and inter-quartile ranges are shown in the graph. The Mann-Whitney U test, adjusted for 4 statistical tests, was used for data comparisons. I_Na-L was recorded after 200 msec at −30 mV (n=6–8 each). F. Examples of the effects of acute and chronic (5 hours) dofetilide on spontaneous ventricular-like action potentials in hiPSC-CMs.

Figure 7

Effects of 5 hours dofetilide exposure in CHO cells on the voltage-dependence of sodium current inactivation. Panels A and B show examples of tracings recorded using the protocol shown to determine the voltage-dependence of inactivation. In control, inactivation is near-complete by −70 mV, whereas after exposure to dofetilide, currents are not only larger (see also Supplemental Figure 3A–C) but are also incompletely inactivated at −70 mV. Panel C shows that dofetilide produced a striking positive shift in the voltage-dependence of channel inactivation, without a change in the voltage-dependence of activation. As a result, dofetilide increased the overlap between activation and inactivation (dotted box, panel C; shown at expanded scale in panel D). E. A slow voltage ramp protocol shows the peak voltage at which increased window current is observed coincides with the overlap shown in panels C and D.

Figure 8

Differing effects of dofetilide and moxifloxacin on Akt phosphorylation. A and B. Akt phosphorylation was significantly inhibited by dofetilide (Dof, at 3 and 10 µM, p<0.05), but not by moxifloxacin (Moxi, 100 µM), compared with the positive control PI3K inhibitor LY294002 (20 µM) in CHO cells transfected with SCN5A. Paired t-tests were used for data comparisons. A p<0.025 was considered statistically significant (adjusted for 2 tests). C and D. Chronic (5 hours) LY294002 (20 µM) enhanced late current in SCN5A-transfected CHO cells.