ABT-333 (Dasabuvir) Increases Action Potential Duration and Provokes Early Afterdepolarizations in Canine Left Ventricular Cells via Inhibition of IKr

ABT-333 (dasabuvir) is an antiviral agent used in hepatitis C treatment. The molecule, similarly to some inhibitors of hERG channels, responsible for the delayed rectifier potassium current (IKr), contains the methanesulfonamide group. Reduced IKr current leads to long QT syndrome and early afterdepolarizations (EADs), therefore potentially causing life-threatening arrhythmias and sudden cardiac death. Our goal was to investigate the acute effects of ABT-333 in enzymatically isolated canine left ventricular myocardial cells. Action potentials (APs) and ion currents were recorded with a sharp microelectrode technique and whole-cell patch clamp, respectively. Application of 1 μM ABT-333 prolonged the AP in a reversible manner. The maximal rates of phases 0 and 1 were irreversibly decreased. Higher ABT-333 concentrations caused larger AP prolongation, elevation of the early plateau potential, and reduction of maximal rates of phases 0, 1, and 3. EADs occurred in some cells in 3–30 μM ABT-333 concentrations. The 10 μM ABT-333-sensitive current, recorded with AP voltage clamp, contained a late outward component corresponding to IKr and an early outward one corresponding to transient outward potassium current (Ito). ABT-333 reduced hERG-channel-mediated ion current in a concentration-dependent, partially reversible manner with a half-inhibitory concentration of 3.2 μM. As the therapeutic plasma concentration of ABT-333 is 1 nM, the arrhythmic risk of ABT-333 is very low, even in the case of drug overdose.


Introduction
The rapid component of delayed rectifier potassium current (I Kr ) is one of the most important potassium currents of the cardiac repolarization as it corresponds to the initiation of late repolarization (phase 3) [1,2]. I Kr is mediated by hERG channel proteins [3,4] and targeted by class III antiarrhythmics of the Vaughan Williams classification [5]. Early afterdepolarizations (EADs) can be generated on the basis of action potential (AP) prolongation induced by inhibition of I Kr [6,7]. The AP prolongation due to reduction of I Kr is known as long QT syndrome type 2 (LQTS2). This condition increases the short-term variability (SV)

1 µM ABT-333 Prolonged the Left Ventricular Action Potential
First, we perfused canine left ventricular cells with 1 μM ABT-333. During these experiments, ABT-333 perfusion lasted for 15 min, which was followed by a 20-minute-long washout. The analyzed parameters of the last ten action potentials recorded before ABT-333 perfusion were averaged to reduce the short-term variability of APs and are indicated as control: bicarbonate buffer containing Tyrode solution (BTY). Again, the average of the last ten APs before the washout was considered as ABT-333, while the average of other ten APs after the 20 min washout is indicated as Wo.
ABT-333, applied in 1 μM concentration, increased the AP duration of left ventricular cells from 258.3 ± 15.4 to 277.4 ± 15.3 ms (Figure 2A,B), resulting in a 7.84 ± 3.09% increase in the duration of the AP from the peak to 90% repolarization (APD90). This AP prolongation was not statistically significant (p = 0.08) for the duration of the AP from the peak to 50% repolarization (APD50), but the prolongation tendency was reversible upon washout, just as the ABT-333-induced APD90 increase. ABT-333 significantly reduced the maximal rates of phase 0 ( Figure 2C) and phase 1 (VPh1max, Figure 2D) to 82.8 ± 5.1 and 51.6 ± 11.3%, respectively, in a non-reversible manner. Other AP parameters, such as action potential amplitude (APA), the ratio of APD50 and APD90, overshoot potential (OSP), resting membrane potential (RMP), membrane potential differences between the RMP and the membrane potential value of at the 20 and 50% duration of APD90 (Plateau20 and Plateau50 amplitudes, respectively), and maximal rate of phase 3 (V−max) values did not change ABT-333, or dasabuvir, a non-nucleoside hepatitis C virus (HCV) polymerase inhibitor containing the methanesulfonamide group, is used to treat chronic HCV infection, often in combination with other compounds such as ombitasvir, paritaprevir, and ritonavir [15][16][17]. Chronic HCV infection is a silent worldwide epidemic, with 130-170 million individuals involved [18] and remains quiescent for decades before significant symptoms [19]. The disease contributes to high morbidity and mortality, as well as substantial costs in healthcare [20]. Cardiac side effects of ABT-333 containing anti-HCV medication were detected in some patients, and this was summarized by Li et al. [21]. These side effects included extreme bradycardia [22], cardiac arrest [23], and chest pain [24]. As ABT-333 is metabolized by hepatic cytochrome P450 2C8 enzyme (CYP2C8), strong inhibitors of CYP2C8 may increase ABT-333 plasma levels and carry increased risk of QT prolongation [25]. This can be the case with clopidogrel administration, as its glucuronidated metabolite inhibits CYP2C8 [26].
As ABT-333 contains a methanesulfonamide group and there is a risk of QT prolongation, we aimed to study its actions on the AP of canine ventricular isolated cardiomyocytes, a good model of human cardiomyocytes regarding cellular electrophysiology [27] and also on expressed hERG channels.

1 µM ABT-333 Prolonged the Left Ventricular Action Potential
First, we perfused canine left ventricular cells with 1 µM ABT-333. During these experiments, ABT-333 perfusion lasted for 15 min, which was followed by a 20-minute-long washout. The analyzed parameters of the last ten action potentials recorded before ABT-333 perfusion were averaged to reduce the short-term variability of APs and are indicated as control: bicarbonate buffer containing Tyrode solution (BTY). Again, the average of the last ten APs before the washout was considered as ABT-333, while the average of other ten APs after the 20 min washout is indicated as Wo.
ABT-333, applied in 1 µM concentration, increased the AP duration of left ventricular cells from 258.3 ± 15.4 to 277.4 ± 15.3 ms (Figure 2A,B), resulting in a 7.84 ± 3.09% increase in the duration of the AP from the peak to 90% repolarization (APD 90 ). This AP prolongation was not statistically significant (p = 0.08) for the duration of the AP from the peak to 50% repolarization (APD 50 ), but the prolongation tendency was reversible upon washout, just as the ABT-333-induced APD 90 increase. ABT-333 significantly reduced the maximal rates of phase 0 ( Figure 2C) and phase 1 (V Ph1 max, Figure 2D) to 82.8 ± 5.1 and 51.6 ± 11.3%, respectively, in a non-reversible manner. Other AP parameters, such as action potential amplitude (APA), the ratio of APD 50 and APD 90 , overshoot potential (OSP), resting membrane potential (RMP), membrane potential differences between the RMP and the membrane potential value of at the 20 and 50% duration of APD 90 (Plateau20 and Plateau50 amplitudes, respectively), and maximal rate of phase 3 (V − max) values did not change (Table 1).  Values of action potential duration measured at 90% of repolarization (APD90, (B)), maximal ra depolarization (V + max, (C)), and maximal rate of early repolarization (VPh1max, (D)) obtained examined cells isolated from 4 animals. Asterisks show statistically significant difference from trol (p < 0.05). Values of action potential duration measured at 90% of repolarization (APD 90 , (B)), maximal rate of depolarization (V + max, (C)), and maximal rate of early repolarization (V Ph1 max, (D)) obtained in 8 examined cells isolated from 4 animals. Asterisks show statistically significant difference from control (p < 0.05).

Development of Early Afterdepolarizations in the Presence of ABT-333
As mentioned previously, in the presence of ABT-333, we detected EADs in cer cells ( Figure 4). We summarized the EAD appearance in a way that EAD presence claimed in the given cell from the lowest concentration of ABT-333 in which the first E ABT-333 also reduced the maximal rate depolarization (phase 0; V + max), maximal rate of early repolarization (phase 1; V Ph1 max), and maximal rate of terminal repolarization (phase 3; V − max). Only 10 and 30 µM ABT-333 decreased V + max significantly to 84.4 ± 6.2 and 77.8 ± 10.9% of the control, respectively. Higher ABT-333 concentrations also reduced V − max values to 95.1 ± 0.9, 87.1 ± 4.2, and 84.4 ± 5.6% of the control in 3, 10, and 30 µM concentrations, respectively. ABT-333 reduced the maximal rates of phase 1 repolarization to 78.5 ± 6.3, 57.7 ± 6.8, and 40.0 ± 8.3% in the cases of 3, 10, and 30 µM, respectively.

Development of Early Afterdepolarizations in the Presence of ABT-333
As mentioned previously, in the presence of ABT-333, we detected EADs in certain cells ( Figure 4). We summarized the EAD appearance in a way that EAD presence was claimed in the given cell from the lowest concentration of ABT-333 in which the first EAD appeared and in all higher concentrations, even if the EADs did not persist. We did not observe any EADs in the control, and even in the presence of 1 µM ABT-333, but in higher concentrations of the drug, the percentage of cells developing EADs gradually increased ( Figure 4B). euticals 2023, 16,488

ABT-333-Sensitive Current Profile with AP Voltage Clamp (APVC)
ABT-333-induced AP prolongation and reduction of V−max suggests the in of IKr as both of these changes can be a consequence of IKr reduction. ABT-333reduction of VPh1max suggests inhibition of Ito, as VPh1max is greatly influenced b plitude. To confirm this hypothesis, we recorded the effects of 10 μM ABT-333 du AP voltage clamp technique using a canonical AP as command potential (Figure analyzed the properties of the ABT-333-sensitive current. The ABT-333-sensitive was created by deducting the current trace in the presence of ABT-333 (blue trace ure 5B) from the one recorded before ABT-333 application (black trace on Fig Therefore, the ABT-333-sensitive current contained all those ion currents, which w ified by ABT-333 ( Figure 5C). On the graph, an outward (positive) current was see case of both the ABT-333-induced inhibition of an outward current as well as t 333-induced increase in an inward current.
The 10 μM ABT-333-sensitive current was outward throughout the AP. We o an early outward current peak with a density of 2.76 ± 0.64 pA/pF ( Figure 6B). Th rent traces peaked 2.90 ± 0.81 ms later than the time of the VPh1max of the comm On the basis of the shape of the own AP of the studied cells, there was a negative tion between the VPh1max values and the early outward peak current density valu ure 6C). The outward current at the half duration of the command AP was 0.8 pA/pF. The end of the sustained outward current (Iendsus, measured just before the returning to zero, indicated with d on Figure 6B) corresponding to terminal repola was 0.83 ± 0.16 pA/pF, and its position was 7.82 ± 1.64 ms before the time of the V the canonical AP. There was no correlation between the density of Iendsus and the values of the own AP of the measured cells ( Figure 6D).

ABT-333-Sensitive Current Profile with AP Voltage Clamp (APVC)
ABT-333-induced AP prolongation and reduction of V − max suggests the inhibition of I Kr as both of these changes can be a consequence of I Kr reduction. ABT-333-induced reduction of V Ph1 max suggests inhibition of I to , as V Ph1 max is greatly influenced by I to amplitude. To confirm this hypothesis, we recorded the effects of 10 µM ABT-333 during the AP voltage clamp technique using a canonical AP as command potential ( Figure 5A) and analyzed the properties of the ABT-333-sensitive current. The ABT-333-sensitive current was created by deducting the current trace in the presence of ABT-333 (blue trace on Figure 5B) from the one recorded before ABT-333 application (black trace on Figure 5B). Therefore, the ABT-333-sensitive current contained all those ion currents, which was modified by ABT-333 ( Figure 5C). On the graph, an outward (positive) current was seen in the case of both the ABT-333-induced inhibition of an outward current as well as the ABT-333-induced increase in an inward current.
The 10 µM ABT-333-sensitive current was outward throughout the AP. We observed an early outward current peak with a density of 2.76 ± 0.64 pA/pF ( Figure 6B). These current traces peaked 2.90 ± 0.81 ms later than the time of the V Ph1 max of the command AP. On the basis of the shape of the own AP of the studied cells, there was a negative correlation between the V Ph1 max values and the early outward peak current density values ( Figure 6C). The outward current at the half duration of the command AP was 0.89 ± 0.24 pA/pF. The end of the sustained outward current (I endsus , measured just before the start of returning to zero, indicated with d on Figure 6B) corresponding to terminal repolarization was 0.83 ± 0.16 pA/pF, and its position was 7.82 ± 1.64 ms before the time of the V − max of the canonical AP. There was no correlation between the density of I endsus and the V − max values of the own AP of the measured cells ( Figure 6D).

ABT-333 Blocked Expressed hERG Channels in a Time-and Concentration-Depen Manner
ABT-333 inhibited the hERG current in 30 μM concentration as observed ond-long depolarizing step to +20 mV followed by a repolarizing step to −40 m led to a characteristic fast recovery of inactivated hERG channels ( Figure 7A). T potential was −80 mV, and pulses were applied every 30 s. The application o caused a reduction of the hERG current both at +20 and −40 mV. The monotonic control current at +20 mV changed into a decaying current in the presence of which was fit with a single exponential function, yielding a time constant of 0 ( Figure 7B, n = 5). To test the concentration dependence of the inhibition, we ap 10, and 30 μM ABT-333. The remaining current fractions (RCFs) were 0.93 ± 0 0.05, 0.26 ± 0.04, and 0.15 ± 0.02, respectively (n ≥ 4 for each concentration). The tion-response curve yielded a half maximal inhibitory concentration (IC50) of 3. ure 7C). The effect of ABT-333 on the voltage sensing of hERG channels was te I-V protocol consisting of depolarizing pulses ranging from −50 mV to +50 mV increments. From the normalized tail current peaks, we created the G-V curves trol solution and the 30 μM ABT-333-superfused recordings. Fitting them with mann equation, the V1/2 value for the control solution was 6.50 ± 1.03 mV. For the V1/2 value was unable to be obtained this way, as its application reduced t to the extent where proper determination of the V1/2 value was not possible (Fi = 3). However, visually, a large voltage-shift was not apparent.

ABT-333 Blocked Expressed hERG Channels in a Time-and Concentration-Dependent Manner
ABT-333 inhibited the hERG current in 30 µM concentration as observed by a 3-secondlong depolarizing step to +20 mV followed by a repolarizing step to −40 mV, which led to a characteristic fast recovery of inactivated hERG channels ( Figure 7A). The holding potential was −80 mV, and pulses were applied every 30 s. The application of ABT-333 caused a reduction of the hERG current both at +20 and −40 mV. The monotonic increasing control current at +20 mV changed into a decaying current in the presence of ABT-333, which was fit with a single exponential function, yielding a time constant of 0.87 ± 0.09 s ( Figure 7B, n = 5). To test the concentration dependence of the inhibition, we applied 1, 3, 10, and 30 µM ABT-333. The remaining current fractions (RCFs) were 0.93 ± 0.04, 0.63 ± 0.05, 0.26 ± 0.04, and 0.15 ± 0.02, respectively (n ≥ 4 for each concentration). The concentration-response curve yielded a half maximal inhibitory concentration (IC 50 ) of 3.2 µM ( Figure 7C). The effect of ABT-333 on the voltage sensing of hERG channels was tested by an I-V protocol consisting of depolarizing pulses ranging from −50 mV to +50 mV in +10 mV increments. From the normalized tail current peaks, we created the G-V curves of the control solution and the 30 µM ABT-333-superfused recordings. Fitting them with the Boltzmann equation, the V 1/2 value for the control solution was 6.50 ± 1.03 mV. For ABT-333, the V 1/2 value was unable to be obtained this way, as its application reduced the current to the extent where proper determination of the V 1/2 value was not possible ( Figure 7D, n = 3). However, visually, a large voltage-shift was not apparent.

Effects of 1 µM ABT-333
We first tested ABT-333 actions on cardiac AP in 1 µM, where APD 90 was significantly and reversibly increased, likely due to the expected I Kr blockade of the drug. I Kr is responsible for the initiation of phase 3 repolarization, but its maximal rate (the value of V − max) mainly, and it does not exclusively depend on the density of inward rectifier potassium current, I K1 [1]. Accordingly, ABT-333 did not reduce V − max in 1 µM concentration, suggesting that it did not inhibit I K1 . The APD 50 /APD 90 ratio is often used as a marker of AP triangulation. The smaller the value, the more triangular the shape of the cardiac APs, which is often seen during inhibition of I K1 . As the APD 50 /APD 90 ratio did not change significantly in the presence of ABT-333, it also confirms that the drug did not inhibit I K1 . The ABT-333-induced V Ph1 max and V + max reductions can be caused by the inhibition of transient outward potassium current (I to ) and sodium current (I Na ), respectively. These actions did not reverse the 20 min washout, raising the possibility of an irreversible channel blockade by ABT-333.

Actions of Higher Concentrations of ABT-333
Compared to 1 µM ABT-333, higher concentrations of the drug generated further changes of AP parameters ( Figure 3, Table 2). Both APD 50 and APD 90 values, as well as the amplitude of early plateau phase (Plateau20 amplitude), were increased in a concentrationdependent manner, suggesting again that ABT-333 reduced I Kr . V + max and V Ph1 max values were also smaller in the presence of ABT-333. These observations raise the possibility of I to and I Na inhibition apart from I Kr blockade.
At larger concentrations of ABT-333, EADs appeared, also suggesting the inhibition of I Kr , as AP prolongation often leads to EAD formation. In a computer simulation study, major (more than 90%) I Kr reduction was needed for EAD development [6]. The APD increase, however, is not always due to I Kr inhibition but can be due to the increase in sodium or calcium currents. The former one can probably be ruled out as the reduction of V + max seen with ABT-333 is not consistent with a larger sodium current. Calcium current increase by ABT-333 is a possibility, especially as it would lead to elevation of the early plateau potential, which was indeed detected with ABT-333 ( Figure 3B). Therefore, it is possible that ABT-333 greatly reduced I Kr or induced EADs by reduction of I Kr and simultaneous increase in the calcium current, especially in higher concentrations.

ABT-333-Sensitive Current Profile with APVC
ABT-333-induced reduction of V Ph1 max suggested inhibition of I to on top of the I Kr inhibition. We tried to confirm this with the APVC measurements. The properties of the early outward peak of the average ABT-333-sensitive current were very similar to that shown earlier with the I to inhibitor 4-AP [1,28]. It was found that 1 mM 4-AP inhibits approximately 70% of I to and resulted in approximately 3 pA/pF peak current density (see Figure 1 in Banyasz et al. [1]). In the current study, the ABT-333-sensitive current early peak density was slightly smaller: 2.76 ± 0.64 pA/pF, probably due to the higher presence of cells with small I to current in the current study. Moreover, the time of the ABT-333-sensitive current peak was always after the time of V Ph1 max but before the deepest point of the incisura of the command potential. Further proof that ABT-333 reduces I to is the good correlation between early current peak density values and V Ph1 max values of own APs ( Figure 6C). Moreover, the early outward current decayed very rapidly, similarly to the 4-AP-sensitive current [29]. An amount of 100 µM chromanol 293B can also be used to study I to , wherein the decay of the current was less rapid [30]. Of note, the ABT-333-sensitive current did not decay to nearly zero, as seen with the 4-AP-sensitive one [1]. This might have been due to the effect of ABT-333 on other ion channels conducting during this early plateau phase of the AP. Nevertheless, on the basis of the results, it seems that 10 µM ABT-333 inhibits I to substantially.
We also confirmed the expected I Kr inhibition of ABT-333 with APVC. The position of I endsus was 7.82 ± 1.64 ms before the time of the V − max of command potential, which is in good agreement with earlier results [1,29]. The peak value of the 1 µM E4031-sensitive current was approximately 0.6 pA/pF, slightly smaller than the ABT-333-sensitive current of the current study (approx. 0.8 pA/pF). This could have been due to larger I Kr currents in cells of the current study than those in our previous one (see Figure 2 in Banyasz et al. [1]) as 1 µM E4031 inhibits approximately 80% of I Kr [31], similarly to the large reduction of hERG by 10 µM ABT-333 ( Figure 7C). In addition, it is possible that ABT-333-sensitive current might contain an activated calcium current component besides I to and I Kr , as suggested previously. This activated calcium current, if still active during the late phase of plateau, can add to the late peak value of the ABT-333-sensitive current on top of the I Kr , resulting in a larger last part than the E4031 sensitive current. This possibility is evident if one observes the fingerprints of I to and I Kr , as I to decays to zero at the latest after 50 ms of the plateau phase, while I Kr only starts to activate about 70-80 ms after the AP peak [1]. However, I to can decay more slowly [30] and therefore might contribute to the maintained outward component of the ABT-333-sensitive current during the early plateau phase ( Figure 6B). The ABT-333-induced augmentation of late sodium or calcium current can be also responsible, as that will appear as an outward current. The contribution of the late sodium current is, however, unlikely, as ABT-333 reduced the V + max value of the AP, which argues against sodium current activation.

ABT-333-Induced Reduction of hERG Channel Currents
ABT-333-induced AP parameter changes, namely, the increase in APD 50 and APD 90 , mainly without the alteration of their ratio; the reduction of V − max suggested the inhibition of I Kr . This was expected due to the presence of the methanesulfonamide group in the structure of ABT-333. Moreover, the ABT-333-sensitive current was also outward during the late plateau phase, suggesting I Kr inhibition as well. To confirm this, we carried out experiments on HEK cells stably expressing hERG channels. The hERG-mediated current was indeed reduced by ABT-333 in a concentration-dependent manner with an IC 50 of 3.2 µM. Similarly to some AP parameters where ABT-333-induced changes were (partly) irreversible upon washout, the reduction of hERG current by ABT-333 also was unable to be fully reversed. Moreover, applying ABT-333 in 30 µM concentration caused the time-dependent inhibition during the depolarizing pulse with a time constant of 0.87 ± 0.09 s ( Figure 7B). This led us to investigate the possible effects of the ABT-333 on the hERG channel gating transitions ( Figure S1). On this basis, 3 µM ABT-333 significantly decreased the deactivation time constant ratio (τ ABT-333 /τ control = 0.46 ± 0.04, p = 0.0001), and the superfusion of 30 µM ABT-333 decreased the inactivation time constant ratio to 0.78 ± 0.15 (mean ± SEM). These results suggest that ABT-333 blocks the channel pore mainly during the open state of the channel.

Medical Relevance
According to our hypothesis, ABT-333 indeed blocked the I Kr current. This was supported by (i) prolongation of canine ventricular AP; (ii) reduction of maximal rate of repolarization (V − max decrease); (iii) detecting I Kr -like profile of the late phase of the ABT-333-sensitive current with APVC; and (iv) inhibition the current of expressed hERG channels, the pore forming protein responsible for I Kr . In this study, we used ABT-333 between 1-30 µM. The IC 50 value of ABT-333 against recombinant HCV polymerase (NS5B) was in the range of 2-10 nM [32]. Similar values (2-8 nM) were observed for half effective ABT-333 concentrations using cell culture assays for inhibition of replication of HCV subgenomic replicons for two genotypes (1a (H77) and 1b (Con1)), but the values were greatly (approximately 12-13-fold) increased in the presence of 40% human plasma to 20-100 nM [32]. Even more important is to compare the concentrations of ABT-333 in the current study with plasma concentrations of the drug in patients taking the compound during anti-HCV treatment. The maximal plasma concentration (Cmax) in HCV and HIV co-infected patients taking ABT-333 in combination was approximately 600 ng/mL according to King et al. [33], which corresponds to approximately 1.2 nM. In healthy volunteers taking a 200 mg ABT-333-containing tablet alone twice daily resulted in maximal concentration of 500 ng/mL, approximately 1 nM [34]. Increasing the dose resulted in linearly higher Cmax values (1.8, 2.9, and 4.2 nM for 400, 600, and 1000 mg containing tablets twice daily, respectively). There was no major daily fluctuation in plasma levels, and also in accumulation ratio, except for the 1000 mg dose where 1.6 times higher concentrations were seen by the 10th day [34]. Moreover, renal and hepatic impairment (unless severe in the case of the latter one) did not alter the metabolism of ABT-333, hardy influencing the maximal plasma concentration [34]. The plasma concentration of ABT-333 can be increased in patients taking clopidogrel [25] and also gemfibrozil [35]. Taken together, it is rather unlikely that ABT-333 can reach higher than 1 nM concentration in blood plasma, given the fact that the recommended dose is 250 mg twice daily. Therefore, even in the case of ABT-333 overdose, the arrhythmic risk of ABT-333 is very low.
A compound can cause dual action on certain ion channels, as in the case of 2-ABP with TRPM7 channels [36] or channels responsible for Ca2+-release-activated Ca2+ current [37]. Moreover, several I Kr activators such as RPR260243 [38], NS1643 [39], NS3623 [40], ICA-105574 [41], and A-935142 [42] were shown to reduce I Kr in large concentrations. Given the large difference between the therapeutic dose of ABT-333 and the concentrations we studied above, we also tested the therapeutic concentration of ABT-333 (1 nM). The amount of 1 nM ABT-333 did not influence either the AP of left ventricular cells or expressed hERG currents ( Figure S1, Table S1).

Summary and Conclusions
ABT-333, on the basis of its molecular structure, is expected to block hERG channels and the I Kr current. This was confirmed as ABT-333 application led to prolongation of the cardiac AP and generation of EADs. ABT-333 most likely inhibits I to current too and might activate calcium current. Despite these factors, ABT-333 can be considered safe as its plasma level is unlikely to reach a high enough concentration for the abovementioned actions, even during overdose, and also due to lack of action in the therapeutic concentration.

Isolation of Canine Ventricular Myocytes
Cell isolation was carried out with the segment perfusion technique by enzymatic digestion, as described previously [2]. Intramuscular application of 10 mg/kg ketamine hydrochloride (Calypsol, Richter Gedeon, Budapest, Hungary) and 1 mg/kg xylazine hydrochloride (Sedaxylan, Eurovet Animal Health BV, Bladel, The Netherlands) was used to achieve complete narcosis in adult mongrel dogs of either sex according to protocols approved by the local ethical committee (license no. gassed with a mixture of 95% O 2 and 5% CO 2 and supplemented with 2.5 g/L taurine, 200 mg/L NaH 2 PO 4 , 1.4 g/L NaHCO 3 , 175 mg/L pyruvic acid, 13.5 mg/L allopurinol, and 750 mg/L D-ribose, pH 6.8 at 37 • C. Then, atria were cut off, and a wedge-shaped section of the left ventricular wall supplied by the left anterior descending coronary artery was perfused. Following further 5 min of perfusion to completely remove blood from the tissue, 0.9 g/L collagenase (type II, 245 U/mg; Worthington Biochemical Co., Lakewood, NJ, USA), 2 g/L bovine serum albumin (Fraction V.), and 50 µM CaCl 2 were added to the JMM solution. During the 30-40 min-long enzymatic digestion, the solutions were kept at 37 • C and gassed with a mixture of 95% O 2 and 5% CO 2 . Cells were sedimented and filtered four times to remove large chunks. During this procedure, the Ca 2+ concentration of the JMM solution was gradually restored to the final 1.8 mmol/L. After this, cells were stored in MEM solution (Minimum Essential Medium Eagle, product no. M0643) supplemented with the following: 2.5 g/L taurine, 200 mg/L NaH 2 PO 4 , 2.2 g/L NaHCO 3 , 175 mg/L pyruvic acid, 13.5 mg/L allopurinol, and 750 mg/L D-ribose (pH = 7.3, equilibrated with a mixture 95% O 2 and 5% CO 2 ) at 15 • C until further use within 36 h after isolation. The percentage of living cells (having clear cytoplasm, sharp edges, and clear striations) were usually 30-60%, and only these were used for experiments.

Electrophysiology
Cells were placed in a plexiglass chamber with a volume of 1 mL and continuously superfused with bicarbonate buffer containing Tyrode solution (BTY) containing (in mmol/L): NaCl 121; KCl 4; MgCl 2 1; CaCl 2 1.3; HEPES 10; glucose 10; NaHCO 3 25 (pH = 7.3; adjusted with NaOH) supplied by a gravity-driven system at a speed of 2 mL/min. During experiments, the bath temperature was set to 37 • C by a temperature controller (Cell Micro Controls, Norfolk, VA, USA). Cells were visualized by inverted microscopes placed in a Faraday cage on an anti-vibration table (Newport, Rochester, NY, USA). Electrical signals were recorded with intracellular amplifiers (MultiClamp 700A or 700B, Molecular Devices, San Jose, CA, USA) after analogue-digital conversion (Digidata 1440A or 1332, Molecular Devices, San Jose, CA, USA) and recorded with pClamp 10 software (Molecular Devices, San Jose, CA, USA). Cells were perfused with 1 µM ABT-333 (Cayman Chemical, Ann Arbor, MI, USA, product no. 18482) for 15 min, followed by a 20-minute-long period of washout. When higher ABT-333 concentrations were used (1-3-10-30 µM), each concentration was applied for 5 min in a cumulative manner. ABT-333 was dissolved in DMSO so that the maximal DMSO concentration (in case of 30 µM ABT-333) was 1.5%, which did not affect any of the AP parameters studied (Table S2.).

Recording of Action Potentials
Action potentials (APs) were measured with 3 mol/L KCl containing borosilicate microelectrodes with a tip resistance of 20-50 MΩ. A 1 s cycle length steady-state pacing was achieved for supra-threshold current pulses (2 ms long, 120-130% of threshold) produced by an electronic stimulator (DS-R3; Főnixcomp Ltd., Debrecen, Hungary). APs were digitized at 50 kHz, and upon the off-line analysis of APs, the following parameters were determined in ten consecutive APs then averaged: APD 50 and APD 90 values (duration of the AP from the peak to 50 and 90% of repolarization, respectively); maximal rate of phases 0, 1, and 3 (V + max, V Ph1 max, and V − max, respectively); resting membrane potential (RMP); overshoot potential (OSP); APA (action potential amplitude, determined as the difference between OSP and RMP); and membrane potential difference between the RMP and the membrane potential value of at the 20% and 50% durations of APD 90 (Plateau20 and Plateau50 amplitudes, respectively).

Action Potential Voltage Clamp (APVC) Studies
Action potential voltage clamp experiments were conducted according to the method described previously [2], and signals were digitized at 50 kHz. In the experiments, a previously recorded typical AP (recorded with 700 ms long cycle length pacing on a midmyocardial cell, termed as "canonic" AP) was applied to the voltage clamped cells as command signal. In action potential voltage clamp experiments, ABT-333-sensitive current was obtained by pharmacological subtraction calculated by deducting the current signals recorded in the presence of 10 µM ABT-333 from those measured in the control condition (BTY) [2]. On this ABT-333-sensitive current, an outward (positive) current will be seen in cases of both the ABT-333-induced inhibition of an outward current as well as the ABT-333-induced increase in an inward current.
Membrane currents were recorded using the patch-clamp technique [43] in wholecell configuration. The cells were superfused with bicarbonate buffer containing Tyrode solution (see above for composition) at 37 • C. Borosilicate glass micropipettes had tip resistance of 2-3 MΩ after filling with pipette solution containing (in mmol/L) K-aspartate 120, KCl 30, MgATP 3, HEPES 10, Na 2 -phosphocreatine 3, EGTA 0.01, cAMP 0.002, and KOH 10 at pH = 7.3 with an osmolarity of 285 mOsm measured with a vapor pressure osmometer (Vapro 5520, Wescor Inc., Logan, UT, USA). After establishing high (1-10 GΩ)resistance seal by gentle suction, the cell membrane beneath the tip of the electrode was disrupted by further suction and/or by applying 1.5 V electrical pulses for 1 ms. The series resistance was typically 4-6 MΩ. Experiments were discarded when the series resistance was high or substantially increased during the measurement. Ion currents were normalized to cell capacitance using short hyperpolarizing pulses to −10 mV for 45 ms from 20 ms long depolarization to 0 mV from the holding voltage of −80 mV applied at 10 Hz. The average value of cell capacitance was 154.2 ± 6.4 pF in the 7 myocytes studied.

Recording of hERG Currents
Whole-cell currents of voltage-clamped HEK (human embryonic kidney) cells stably expressing hERG channels were recorded by manual patch-clamp according to standard protocols using Axopatch 200B amplifiers connected to computers via Digidata 1550B digitizers (Molecular Devices, San Jose, CA, USA). Pipettes were pulled from GC 150F-15 borosilicate glass capillaries (Harvard Apparatus, Hollister, MA, USA) in three stages with 3-5 MΩ resistance. Immediately before the measurement, the cells were maintained in the recording Petri dish in bath solution (i.e., control solution, composition in mmol/L: choline chloride 140, KCl 5, MgCl 2 2, CaCl 2 2, glucose 20, HEPES 10, and CdCl 2 0.1; pH = 7.35 (adjusted with NaOH)). For the recordings, the composition of the solution used in patch pipette (internal solution) was made of in mmol/L: KCl 140, HEPES 10, MgCl 2 2, and EGTA 10, pH = 7.3 (adjusted with KOH). To test if the charge carrier was K + , a high concentration potassium solution (in mmol/L: KCl 150, HEPES 10, glucose 5.5, MgCl 2 1, CaCl 2 1; pH = 7.35, adjusted with KOH) was used as a positive control. ABT-333 was diluted to 30, 10, 3, and 1 µM concentrations. As ABT-333 was dissolved in DMSO, DMSO was also added to the control solution in 1.5% V/V. Solution exchange was achieved by using a gravity-flow system with continuous excess fluid removal. To avoid the changing of junction potentials during solution changes, the reference electrode, placed in a dish containing internal solution, was connected to the bath solution with an agar bridge. For the voltage-clamp measurements, a holding potential of -80 mV was used. Patch-clamp data were acquired with pClamp10 (Molecular Devices, San Jose, CA, USA). In general, currents were low pass filtered using the built-in analog four-pole Bessel filters of the amplifiers and sampled at 20 kHz. Before analysis, whole-cell current traces were digitally filtered (five-point boxcar smoothing). Experiments were performed at room temperature, ranging between 20 and 24 • C. Clampfit 10.7 (Molecular Devices, San Jose, CA, USA) and Graphpad Prism 7 (Graphpad, San Diego, CA, USA) were used for data display and analysis. The G-V curves were assembled by the normalization of the tail currents and fit with the Boltzmann equation (G = 1/(1 + e ((V 1/2 −V)/k) )), where k is the slope factor, V is the set membrane potential, and V 1/2 is the midpoint voltage. The concentration-response curve was fit with the Hill equation (RCF = 1/(1 + ([c] nH /IC 50 nH ))), where RCF is the remaining current fraction, nH is the Hill coefficient, [c] is the concentration of the ABT-333, and the IC 50 is the concentration where half of the channels are inhibited. The RCF values were calculated by the following equation: RCF = I ABT /I control , where I ABT is the tail current measured in the presence of ABT-333, and I control is the tail current measured in control solution. The time dependence was fit with a one-phase exponential decay equation Y = (Y 0 − C) × e −t/τ + C , where Y 0 is the Y value when time (t) is zero, C is the plateau where Y saturates over time, and τ is the time constant.

Statistical Analysis
All values are presented as arithmetic means ± standard error of the mean (SEM). Given the biological variability among cells, each cell was treated as independent in the statistical tests, although more cells could be obtained from the same animal. The statistical significance of differences was evaluated using paired Student's t-tests. Differences were considered significant when p was less than 0.05, and this is indicated with asterisks on graphs.
Supplementary Materials: The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/ph16040488/s1, Figure S1: The effect of the ABT-333 on the gating transitions of the hERG channel; Table S1: AP parameters obtained with 1 nM ABT-333; Table S2: AP parameters obtained with 1.5% DMSO.

Conflicts of Interest:
The authors declare no conflict of interest.  another name of the dasabuvir molecule AP action potential APA action potential amplitude APD action potential duration APD 50 action potential duration at 50% of repolarization APD 90 action potential duration at 90% of repolarization