Sensitivity and Specificity of the In Vitro Guinea Pig Papillary Muscle Action Potential Duration for the Assessment of Drug-Induced Torsades De Pointes Liability in Humans.

The ICH S7B document, which provides guidance for the preclinical cardiovascular evaluation of pharmaceutical new chemical entities (NCE), is essentially focused on drug-induced QT lengthening, a biomarker for the proarrhythmic adverse drug reaction, torsades de pointes (TdP). In 2005, this guidance recommended the IKr assay and the in vivo QT telemetry study as mandatory assays for detecting potential torsades de pointes liability and relegated the cardiac action potential (AP) assay as a follow-up study. The IKr assay has become a mandatory screening tool in the early development and safety assessment process. Using only the IKr assay as a go/no go decision arbiter is regrettable since, due to the low specificity of the model (positives that are false for proarrhythmia liability, e.g. verapamil), promising, safe NCEs may be inadvertently discarded. Inclusion of additional medium throughput assays should be performed early to confirm or balance the putatively unfavourable IKr result with positive discovery model output (Pugsley et al., J Pharmacol Toxicol Methods 60:24-27, 2009). In the present chapter, the predictive value of in vitro guinea pig papillary muscle action potential assay will be discussed in terms of sensitivity and specificity and compared to currently available preclinical models such as IKr/hERG assay, dog Purkinje fibre action potential and in vivo QT measurements in dog and cynomolgus monkey.

I(Ks) blockade in border zone arrhythmias from guinea-pig ventricular myocardium submitted to simulated ischemia and reperfusion.

I(Ks) blockade might be a promising way to treat tachyarrhythmia because of the accumulation of activated potassium channels. However, I(Ks) blockade during ischemia/reperfusion has not been investigated. Thus, the electrophysiological effects of two I(Ks) blockers, chromanol 293B (10 µm) and HMR 1556 (1 µm), were assessed in an in vitro model of border zone between normal and ischemic/reperfused right ventricular myocardium from guinea-pigs, and classic electrophysiological parameters and the incidence of arrhythmias were studied. HMR 1556 and chromanol 293B exhibited slight conventional class III effects on action potential duration in the normal zone (NZ) (APD(90) : -2 ± 5%, not significant (NS); +6 ± 3%, NS; and +5 ± 1%, P < 0.05, respectively, in control, HMR 1556, and chromanol 293B groups) but failed to oppose its decrease after 30 min of simulated ischemic superfusion (APD(90) : -52 ± 5%, P < 0.01; -64 ± 5%, P < 0.01; and -61 ± 3%, P < 0.01, respectively, in control, HMR 1556, and chromanol 293B groups), leading to repolarization dispersion between normal and ischemic zones. Chromanol 293B and HMR 1556 prolonged APD(90) during reperfusion, respectively, by +11 ± 1%, P < 0.01 and +25 ± 4%, P < 0.01 in the NZ and by +13 ± 3%, NS and +31 ± 2%, P < 0.01 in the simulated ischemic zone. Both compounds exhibited neutral arrhythmogenic effects during ischemia or reperfusion. Thus, I(Ks) blockade was neutral on the occurrence of ventricular arrhythmias during ischemia and reperfusion in guinea-pig ventricular tissue.

I(Kr) vs. I(Ks) blockade and arrhythmogenicity in normoxic rabbit Purkinje fibers: does it really make a difference?

The electrophysiological (standard intracellular microelectrode technique) and pro-arrhythmic (occurrence of early after-depolarization) effects of five class III agents acting on delayed rectifier current (I(K)), rapid (I(Kr)), and/or slow (I(Ks)) components have been studied in rabbit Purkinje fibers taken near the septum and submitted in vitro to reduced stimulation rate (from 1 to 0.5 Hz) in the absence or presence of epinephrine (10 nm) during normoxic conditions. There were two I(Kr) blockers (d-sotalol and dofetilide), two I(Ks) blockers (chromanol 293B and HMR 1556), and a non-selective I(K) blocker (azimilide). d-sotalol, dofetilide, and azimilide lengthened APD(60) and APD(90) in a concentration-dependent manner. Both d-sotalol and dofetilide showed pro-arrhythmia at highest concentrations and in the presence of epinephrine and lower stimulation rate. Despite azimilide markedly lengthened APD(90), it was globally less pro-arrhythmic than dofetilide. Thus, in normoxic rabbit Purkinje fibers, I(Kr) blockade prolonged action potential duration (APD) and increased the incidence of early after-depolarizations, particularly so in the presence of adrenergic stimulation and bradycardia, I(Ks) blockade did neither, and non-selective I(K) blockade (by azimilide) behaved principally as I(Kr) blockade. It is concluded that in normoxic rabbit Purkinje fibers, I(Ks) blockade was neutral, whereas I(Kr) blockade was pro-arrhythmic, which may make a difference worth exploration in more complex models.

Dexrazoxane protects the heart from acute doxorubicin-induced QT prolongation: a key role for I(Ks).

INTRODUCTION: Doxorubicin, an anthracycline widely used in the treatment of a broad range of tumours, causes acute QT prolongation. Dexrazoxane has been shown to prevent the QT prolongation induced by another anthracycline, epirubicin, but has not yet been reported to prevent that induced by doxorubicin. Thus, the present study was designed to test whether the acute QT effects induced by doxorubicin could be blocked by dexrazoxane and to explore the mechanism. Results were compared with those obtained with a reference human ether-a-go-go (hERG) channel blocker, moxifloxacin. METHODS: The effects of moxifloxacin (100 microM) and doxorubicin (30 microM), with or without dexrazoxane (from 3 to 30 microM), have been evaluated on the QTc interval in guinea-pig isolated hearts and on I(Kr) (rapid component of the delayed rectifier current) and I(Ks) (slow component of the delayed rectifier current) currents stably expressed in human embryonic kidney 293 cells. RESULTS: Moxifloxacin (100 microM), a potent hERG blocker, prolonged QTc by 22%, and this effect was not prevented by dexrazoxane. Doxorubicin (30 microM) also prolonged QTc by 13%, did not significantly block hERG channels and specifically inhibited I(Ks) (IC(50): 4.78 microM). Dexrazoxane significantly reduced the doxorubicin-induced QTc prolongation and prevented doxorubicin-induced inhibition of I(Ks). CONCLUSION AND IMPLICATIONS: Doxorubicin acutely prolonged the QT interval in guinea-pig heart by selective I(Ks) blockade. This effect was prevented by dexrazoxane. This result is important because it illustrates the danger of neglecting I(Ks) in favour of hERG screening alone, for early preclinical testing for possible induction of torsade de pointes.

Action potential experiments complete hERG assay and QT-interval measurements in cardiac preclinical studies.

INTRODUCTION: The ICHS7B guideline focused on hERG and QT assays, although other factors have also been linked with the induction of severe arrhythmias. Thus, the aim of the present study was to demonstrate that two in vitro action potential recordings constitute convincing models of predictive drug-induced Torsades de pointes (TdP) and re-entry arrhythmias. METHODS: The effects of D,L-sotalol, flecainide and quinidine were investigated on potassium (hERG) and sodium (Na(V)1.5) currents transfected in HEK-293 cells to determine the repercussion of the blockade of these currents on rabbit Purkinje fibre (PF) and atrial action potentials. Atrial conduction velocity was also investigated as a model of re-entry arrhythmias. RESULTS: hERG channels were blocked by D,L-sotalol, quinidine and flecainide (IC(50): 69, 0.33 and 0.74 micromol/L, respectively). D,L-sotalol (30 micromol/L) induced reverse-use dependent increases in action potential duration (APD(90): +31.7% and +81.2% at 1 and 0.2 Hz) and triangulation (APD(90-40): +34.7% and +73.6% at 1 and 0.2 Hz) in PF but not in atria. Quinidine (10 micromol/L) also increased APD(90) (+14.5% and +68.5% at 1 and 0.2 Hz) and APD(90-40) (+73.3% and +152.1% at 1 and 0.2 Hz) in PF. Flecainide (10 micromol/L) shortened APD(90) in PF (-26.0% and - 22.2% at 1 and 0.2 Hz). Quinidine and flecainide blocked Na(V)1.5 channels by 32.3% and 73.1%, respectively, and produced decreases in dV/dt(max) which were more marked in atria (-20.4% and -31.9%) compared to PF (-12.8% and 22.4%) at 1 Hz. Finally, quinidine and flecainide decreased atrial conduction speed by 14.6% and 30.8%, respectively. CONCLUSION: Results obtained with flecainide demonstrate that use of the hERG channel alone should not be considered as a useful single assay. Rabbit Purkinje fiber action potentials can be considered as a comparable model for detection of reverse-use dependent APD prolongation and triangulation whereas the rabbit atria can be considered as a useful model for detection of sodium channel blockade associated with decreases in dV/dt(max) and conduction velocity.

Class III effects of dofetilide and arrhythmias are modulated by [K+]o in an in vitro model of simulated-ischemia and reperfusion in guinea-pig ventricular myocardium.

To evaluate class III effects of clinically relevant concentrations of dofetilide (5 and 10 nmol/l) and the effects of extracellular potassium [K+]o modulation of arrhythmias onset at the level of the "border zone," we used a previously reported in vitro model whereby normoxic and ischemic/reperfused zones were studied. Guinea-pig right ventricular strips (driven at 1 Hz at 36.5+/-0.5 degrees C) were superfused with Tyrode's solution in oxygenated (HCO3- 25 mmol/l, K+ 4 mmol/l, pH 7.35+/-0.05, glucose 5.5 mmol/l: normal zone) and ischemia-simulating conditions (HCO3- 9 mmol/l, pH 6.90+/-0.05, no oxygen and no glucose: altered zone) having either [K+]o 4 (n=20), 8 (n=20) or 12 (n=20) mmol/l. Action potentials in normal and altered zones were recorded simultaneously during 30 min of simulated-ischemia and after 30 min of reperfusion with oxygenated Tyrode's solution. Each preparation served as control for successive phases of dofetilide studies (at 5 and 10 nmol/l) and action potential values were normalized to those present at the beginning of the experiment. During simulated-ischemia, the higher the [K+]o the worse were action potential changes, although full recovery was seen upon 30 min of reperfusion in all [K+]o groups. A high incidence of ischemia/reperfusion arrhythmias was observed in 4 and 12 mmol/l [K+]o groups as opposed to a low incidence of arrhythmias in 8 mmol/l [K+]o group. Dofetilide at 5 and 10 nmol/l with all [K+]o explored: (i) exhibited class III effects, (ii) was effective (or neutral) against ventricular arrhythmias during both simulated-ischemia and reperfusion, and (iii) did not globally increase the dispersion of action potential durations between normal and altered zones. Different arrhythmogenic mechanisms are involved in this model at different [K+]o with 8 mmol/l providing relative protection. Class III effects of dofetilide are evident in the normal zone when in the ischemic-like zone [K+]o ranges from 4 to 12 mmol/l. Thus dofetilide did not increase dispersion of repolarization and had either an antiarrhythmic or a neutral effect during ischemia/reperfusion.

Electrophysiological effects of azimilide in an in vitro model of simulated-ischemia and reperfusion in guinea-pig ventricular myocardium.

There are few investigations on azimilide effects during ischemia/reperfusion. We have therefore investigated low concentrations of azimilide (0.1 and 0.5 micromol/l) versus Controls on action potential parameters and occurrence of repetitive responses during simulated ischemia and reperfusion. An in vitro model of "border zone" in guinea-pig ventricular myocardium (n=30) was used. Azimilide 0.5 micromol/l lengthened action potential duration in normoxic but not in ischemic-like conditions. Therefore an increased dispersion of action potential duration at 90% of repolarization during simulated ischemia in presence of azimilide was seen. Upon reperfusion, both normal and reperfused myocardium showed azimilide-induced action potential duration increase. There was a neutral effect on the occurrence of arrhythmias during simulated ischemia; however azimilide showed significant (P=0.033) antiarrhythmic properties following reperfusion. To mimic I(Kr) and I(Ks) blocking properties of azimilide we further used dofetilide 10 nmol/l with HMR 1556 1 nmol/l (N=9), which was accompanied by less severe shortening (P<0.05) of action potential duration at 90% of repolarization at 30 min of ischemic-like conditions (-43+/-9%), as compared with azimilide 0.5 micromol/l (-64+/-5%) but similar to what seen with azimilide 0.1 micromol/l (-53+/-5%) and Controls (-52+/-6%). During reperfusion, 2/9 (22%) preparations had sustained activities, which was less than what observed in Controls (5/10, 50%) and with azimilide 0.5 micromol/l (0/10, 0%), although not statistically different (respectively, P=0.35 and P=0.21). Lack versus homogenous class III effects of azimilide in respectively simulated ischemia and reperfusion may explain its different efficacy on arrhythmias, although prevention of reperfusion arrhythmias calls for other than just its I(Kr) and I(Ks) blocking properties.

Additive effects of ziprasidone and D,L-sotalol on the action potential in rabbit Purkinje fibres and on the hERG potassium current.

INTRODUCTION: Ziprasidone, an atypical antipsychotic has been shown to be devoid of cardiac adverse effects in spite of its propensity to prolong the QT-interval via a hERG current inhibition. However, the effects of ziprasidone on the action potential (AP) parameters have not been published yet. Moreover, very little information is available concerning pharmacodynamic interactions between ziprasidone and other hERG channel blockers. Thus, we investigated the putative interaction between ziprasidone and D,L-sotalol on the hERG channels at therapeutic concentrations and their consequences on the action potential prolongation. METHODS: AP were recorded at 1 and 0.2 Hz. Increasing concentrations of ziprasidone (0.01-10 micromol/L) were successively superfused for 30 min alone or in D,L-sotalol 10 micromol/L pre-treated fibres. Moreover, the effects of ziprasidone, alone or in association with d,l-sotalol, were investigated on the hERG current. RESULTS: Ziprasidone (1-10 microM) induced a concentration and reverse frequency-dependent increase in APD(90) (APD(90): +27% and +36%, respectively at 1 Hz and +50% and +70%, respectively at 0.2 Hz) due to a hERG current blockade (IC50: 0.24 micromol/L). A pre-treatment with D,L-sotalol 10 micromol/L led to an increase in APD(90) of +23% at 1 Hz, stable at 66+/-4 min. In these pre-treated fibres, ziprasidone (1 and 10 micromol/L) induced an additional AP prolongation (APD(90): +16% and +18%, respectively at 1 Hz) as compared to D,L-sotalol pre-treatment. Moreover, D,L-sotalol did not interact with the pharmacological profile of ziprasidone on the hERG channel. CONCLUSION: The present study demonstrates that ziprasidone induces an AP prolongation due to its propensity to block the hERG channel. Moreover, ziprasidone and d,l-sotalol, superfused concomitantly exhibit additive effects on the AP duration since they do not interact as competitors for the hERG channel.