Correlation between human ether-a-go-go-related gene channel inhibition and action potential prolongation

Abstract

Background and purpose: Human ether-a-go-go-related gene (hERG; K_v 11.1) channel inhibition is a widely accepted predictor of cardiac arrhythmia. hERG channel inhibition alone is often insufficient to predict pro-arrhythmic drug effects. This study used a library of dofetilide derivatives to investigate the relationship between standard measures of hERG current block in an expression system and changes in action potential duration (APD) in human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs). The interference from accompanying block of Ca_v 1.2 and Na_v 1.5 channels was investigated along with an in silico AP model.

Experimental approach: Drug-induced changes in APD were assessed in hiPSC-CMs using voltage-sensitive dyes. The IC₅₀ values for dofetilide and 13 derivatives on hERG current were estimated in an HEK293 expression system. The relative potency of each drug on APD was estimated by calculating the dose (D₁₅₀ ) required to prolong the APD at 90% (APD₉₀ ) repolarization by 50%.

Key results: The D₁₅₀ in hiPSC-CMs was linearly correlated with IC₅₀ of hERG current. In silico simulations supported this finding. Three derivatives inhibited hERG without prolonging APD, and these compounds also inhibited Ca_v 1.2 and/or Na_v 1.5 in a channel state-dependent manner. Adding Ca_v 1.2 and Na_v 1.2 block to the in silico model recapitulated the direction but not the extent of the APD change.

Conclusions and implications: Potency of hERG current inhibition correlates linearly with an index of APD in hiPSC-CMs. The compounds that do not correlate have additional effects including concomitant block of Ca_v 1.2 and/or Na_v 1.5 channels. In silico simulations of hiPSC-CMs APs confirm the principle of the multiple ion channel effects.

Figure 1

Chemical structures of dofetilide and its derivatives.

Figure 2

Effects of dofetilide and its derivatives Dofe54 and Dofe33 on AP characteristics in hiPSC‐CMs. Representative AP recordings of hiPSC cardiomyocytes after incubating with dofetilide, n = 4–5 (A), the high affinity derivative Dofe54, n = 4 (B) and the low affinity derivative Dofe33, n = 4 (C) and plots of APD₉₀ as % of control versus concentrations of dofetilide, n = 4–5 (D), Dofe54, n = 4 (E) and Dofe33, n = 4 (F). The data points represent the mean ± SEM (see Table 1) and were fitted by a Hill equation for dofetilide and Dofe54. The data points for Dofe33 were connected by lines.

Figure 3

Effect of Dofe42, Dofe44 and Dofe45 on AP. (A–C) Representative AP traces of controls and in the presence of the indicated drugs. (D–F) Show dependence of APD₉₀ on the concentration of indicated derivatives (n = 4–6, see Table 1).

Figure 4

Effect of dofetilide and derivatives on potassium currents mediated through hERG channels expressed in HEK293 cells. (A) Representative current traces of control current (in the absence of drug) and in the presence of Dofe54 after steady state was reached at each concentration applied. The voltage protocol illustrated was applied every 3 s (A, upper panel). (B) Concentration‐inhibition curves for dofetilide (n = 5) and high affinity derivatives: Dofe54 (n = 5), Dofe81 (n = 6), Dofe60 (n = 5), Dofe78 (n = 7), Dofe35 (n = 8), Dofe45 (n = 5) and Dofe44 (n = 6). (C) Concentration‐inhibition curves for dofetilide and low affinity derivatives: Dofe33 (n = 6), Dofe31 (n = 7), Dofe30 (n = 7), Dofe41 (n = 8), Dofe42 (n = 8) and Dofe43 (n = 5). Peak tail current values (mean ± SEM, see Table 2) were fitted by the Hill equation.

Figure 5

Correlation between D₁₅₀ (concentration that prolongs AP in hiPSC‐CM by 50%) and IC₅₀ (half‐maximal concentration inhibiting hERG channels in HEK293 cells). A significant linear correlation (r = 0.94, P < 0.05) was observed for 12 data points (black circles) including dofetilide and 11 derivatives. Derivatives Dofe45, Dofe44 and Dofe42 (red circles) were not included in the correlation analysis. Dofe45 prolonged the AP in hiPSC‐CM by 50% only at 538 nM and Dofe42 and 44 at >600 nM. The red line represents a prediction of the mathematical simulation of the hiPSC‐CM's AP (see Figure 8).

Figure 6

Inhibition of Ca_v1.2 channel by dofetilide derivatives. (A) Superimposed barium currents through rabbit Ca_v1.2 in control (black) and in the presence of indicated concentrations of Dofe42 (left), Dofe44 (middle) and Dofe45 (right). Barium currents were recorded in response to 50 ms pulses (0.2 Hz) from the holding potential of −80 to +10 mV. (B) Concentration‐dependence of peak I_Ba inhibition by Dofe42 (IC₅₀ = 38 ± 9.3 μM, n = 5, left), Dofe44 (IC₅₀ > 200 μM, n = 5, middle) and Dofe45 (IC₅₀ = 192 ± 28 nM, n = 5, right). The IC₅₀ values were obtained by fitting the data by the Hill equation. (C) Barium currents through Ca_v1.2 during 1 Hz trains of 50 ms pulses from −80 to +10 mV under control conditions (absence of drug) and after 3 min incubation in the presence of the indicated concentrations of dofetilide derivatives. The first current in drug reflects the resting state inhibition. (D) Mean peak current amplitudes during 50 ms pulse trains in control and the presence of the indicated concentration of Dofe42, Dofe44 and Dofe45. The peak current decay after 20 pulses at 1 Hz in control indicates the development of inactivation. Peak current decay in the presence of Dofe42 (100 μM, 38 ± 2%, n = 5) and Dofe45 (100 nM, 42 ± 6%, n = 5) versus in control (12 ± 2%, n = 6) illustrates additional significant (P < 0.05) use‐dependent block.

Figure 7

Inhibition of Na_v1.5 by dofetilide derivatives. (A) Superimposed I_Na through human Na_v1.5 in control (black) and in the presence of indicated concentrations of Dofe42 (left), Dofe44 (middle) or Dofe45 (right). Sodium currents were recorded in response to 20 ms pulses (0.2 Hz) from a holding potential of −140 to −10 mV. (B) Concentration‐dependence of peak I_Na inhibition at a holding potential of −140 mV (squares) and −80 mV (circles) yielding IC₅₀ values for Dofe 42 of IC₅₀ = 77.9 ± 9.7 (at −140 mV, n = 6) and IC₅₀ = 13.8 ± 1.9 (at −80 mV, n = 5), Dofe 44 of IC₅₀ = 23.3 ± 1.9 (at −140 mV, n = 6) and IC₅₀ = 4.7 ± 2.0 (at −80 mV, n = 6) and Dofe 45 of IC₅₀ = 69.7 ± 1.0 (at −140 mV, n = 6) and IC₅₀ = 6.4 ± 1.0 μM (at −80 mV, n = 5).

Figure 8

Simulation of hiPSC‐CM AP at indicated levels of hERG, Ca_v1.2 and Na_v1.5 channel inhibition. (A) Simulation of hiPSC‐CM APs for different levels of selective hERG channel inhibition. (B) Dependence of the calculated APD₉₀ (as % of control) on the concentration of a selective hERG channel inhibitor. (C) Simulated APs at a Dofe45 concentration of 300 nM accounting for hERG inhibition (IC₅₀ = 40 nM) and simultaneous inhibition of Ca_v1.2 (IC₅₀ = 200 nM) and Na_v1.5 (IC₅₀ = 8.9 μM). (D) Comparison of simulated APs at different IC₅₀s of Ca_v1.2 and Na_v1.5 inhibition. Control AP is shown in dark blue and AP for selective hERG channel inhibition (IC₅₀ = 40 nM) in light blue. Red 100 nM (Ca_v1.2) and 1 μM (Na_v1.5), orange 100 nM (Ca_v1.2) and 10 μM (Na_v1.5), magenta 500 nM (Ca_v1.2) and 1 μM (Na_v1.5), green 500 nM (Ca_v1.2) and 10 μM (Na_v1.5). See also Supporting Information Table S1 comparing the values used in silico AP models of adult ventricular myocytes and hiPSC‐CM.