Inhibition of cardiac late sodium current (late INa) is a strategy to suppress arrhythmias and sodium-dependent calcium overload associated with myocardial ischemia and heart failure. Current inhibitors of late INa are unselective and can be proarrhythmic. This study introduces GS967 (6-[4-(trifluoromethoxy)phenyl]-3-(trifluoromethyl)-[1,2,4]triazolo[4,3-a]pyridine), a potent and selective inhibitor of late INa, and demonstrates its effectiveness to suppress ventricular arrhythmias. The effects of GS967 on rabbit ventricular myocyte ion channel currents and action potentials were determined. Anti-arrhythmic actions of GS967 were characterized in ex vivo and in vivo rabbit models of reduced repolarization reserve and ischemia. GS967 inhibited Anemonia sulcata toxin II (ATX-II)–induced late INa in ventricular myocytes and isolated hearts with IC50 values of 0.13 and 0.21 µM, respectively. Reduction of peak INa by GS967 was minimal at a holding potential of −120 mV but increased at −80 mV. GS967 did not prolong action potential duration or the QRS interval. GS967 prevented and reversed proarrhythmic effects (afterdepolarizations and torsades de pointes) of the late INa enhancer ATX-II and the IKr inhibitor E-4031 in isolated ventricular myocytes and hearts. GS967 significantly attenuated the proarrhythmic effects of methoxamine+clofilium and suppressed ischemia-induced arrhythmias. GS967 was more potent and effective to reduce late INa and arrhythmias than either flecainide or ranolazine. Results of all studies and assays of binding and activity of GS967 at numerous receptors, transporters, and enzymes indicated that GS967 selectively inhibited late INa. In summary, GS967 selectively suppressed late INa and prevented and/or reduced the incidence of experimentally induced arrhythmias in rabbit myocytes and hearts.
Sodium (Na+) channel opening and influx of Na+ are responsible for the upstroke of the cardiac action potential (AP). When Na+ channels in myocytes fail to inactivate after opening, Na+ influx continues throughout the AP plateau. The resulting Na+ current (INa) is referred to as late INa to distinguish it from the larger and transient peak INa. Late INa in the normal heart is small, but its magnitude is increased in many pathologic conditions, such as in the failing and/or ischemic heart, in the heart exposed to oxidative stress, and in hearts of patients with congenital long QT3 syndromes (Ver Donck et al., 1993; Le Grand et al., 1995; Bennett et al., 1995; Wang et al., 1995; Ju et al., 1996; Maltsev et al., 1998; Maltsev and Undrovinas, 2006; Song et al., 2006; Sossalla et al., 2010).
Regardless of cause, an enhanced cardiac late INa is proarrhythmic (Boutjdir and El-Sherif, 1991; Sicouri et al., 1997; Undrovinas and Maltsev, 2008; Zaza et al., 2008). Late INa during the plateau of the AP reduces repolarization reserve (i.e., net outward current) and may prolong AP duration (APD). Prolongation of APD can result in early after-depolarizations as a result of L-type Ca2+ channel reopening (Ca2+ window current) (January and Riddle, 1989). Enhancement of late INa has been shown to elicit after-depolarizations, triggered arrhythmic activity, and torsades de pointes (TdP) tachycardia in studies of Purkinje fibers, isolated atrial and ventricular myocytes, isolated wedges of cardiac tissue, and intact hearts (Boutjdir and El-Sherif, 1991; Sicouri et al., 1997; Song et al., 2004, 2008). Conversely, drug-induced reduction of late INa has been associated with improvement of electrical function in myocytes isolated from failing hearts and in hearts made ischemic or that have been exposed to cardiac glycosides, hydrogen peroxide, enhancers of late INa, or drugs that block the rapidly activating delayed-rectifier K+ current (IKr) and reduce repolarization reserve (Ver Donck et al., 1993; Haigney et al., 1994; Le Grand et al., 1995; Sicouri et al., 1997; Song et al., 2004, 2006, 2008; Sossalla et al., 2010; Undrovinas and Maltsev 2008; Wu et al., 2011).
Selective inhibition of late INa is a therapeutic target for treatment of electrical and contractile dysfunction in cardiac ischemia and heart failure (Ver Donck et al., 1993; Undrovinas and Maltsev, 2008; Saint 2006). However, potent and selective inhibitors of late INa are currently available. The Na+ channel blocker tetrodotoxin inhibits late with greater potency than peak INa (Wu et al., 2009). Nevertheless, tetrodotoxin blocks neuronal and skeletal muscle isoforms of Na+ channels with much greater potency than it blocks the cardiac isoform NaV1.5 (Heinemann et al., 1992) and is therefore not suitable for in vivo block of cardiac late INa. Compounds R56865 and F15845 are potent inhibitors of cardiac late INa (Le Grand et al., 2008), but their efficacies for block of late versus peak INa and specificity for binding to Na+ channels relative to other proteins are unknown. Drugs such as lidocaine, mexiletine, and flecainide are prototypical Na+ channel blockers that inhibit both late and peak INa. The anti-ischemic, antianginal drug ranolazine is reported to inhibit late INa with greater potency (IC50 value of 6 µM) than it inhibits other ion currents (Antzelevitch et al., 2004). Ranolazine is more selective for inhibition of late INa relative to peak INa than either lidocaine or amiodarone (Undrovinas et al., 2006). However, similar to flecainide, ranolazine reduces IKr (IC50 value of 12–14 μM) (Antzelevitch et al., 2004). Ranolazine also blocks both α- and β-adrenergic receptors (Zhao et al., 2011), and in one (Parikh et al., 2012) but not in another (Galimberti and Knollmann, 2011) study, it was reported to regulate sarcoplasmic reticulum Ca2+ release. Thus, the interpretation of results of experimental studies with ranolazine and other nonselective inhibitors of late INa may not be straightforward, and the effect of selective inhibition of late INa on cardiac electrical function has yet to be clearly demonstrated.
Here, we describe a novel late INa inhibitor, GS-458967 (GS967), and use this agent to demonstrate that selective block of late INa has anti-arrhythmic actions. In some experiments, the sodium channel toxin Anemonia sulcata toxin II (ATX-II) was used to selectively enhance the Na+ channel late current. ATX-II alters Na+ channel gating to delay and increase recovery from Na+ channel inactivation (Chahine et al., 1996), thereby mimicking the effects of long-QT type-3 mutations (Bennett et al., 1995; Zimmer and Surber, 2008). For comparison with GS967, the drugs flecainide and ranolazine (nonselective inhibitors of both late INa and IKr), were used.
Materials and Methods
The use of rabbits (New Zealand White adult females, 2–4 kg; Western Oregon Rabbit Company, Philomath, OR) in this investigation conformed to the “Guide for the Care and Use of Laboratory Animals” (National Institutes of Health publication No. 85–23, revised 1996) and was approved by the Institutional Animal Care and Use Committees of Gilead Sciences and Zenas Technologies. The use of the female rabbit is based on previous reports that this is a sensitive model for detection of drug-induced proarrhythmia (Hondeghem et al., 2003) and for preclinical evaluation of new drugs believed to affect cardiac AP repolarization (Lengyel et al., 2001). Late INa is reported to be greater in myocytes isolated from female than from male mice (Lowe et al., 2012).
GS967 and ranolazine were synthesized at Gilead Sciences. GS967 (mol wt 347) is a triazolopyridine derivative, 6-(4-(trifluoromethoxy)phenyl)-3-(trifluoromethyl)-[1,2,4]triazolo[4,3-a]pyridine (Fig. 1). ATX-II, E-4031, and flecainide were purchased from Alomone Laboratories (Jerusalem, Israel), Tocris (Bristol, UK), and Sigma (St. Louis, MO), respectively. Stock solutions of GS967, flecainide, and ranolazine were prepared and stored in dimethylsulfoxide in glass vials.
Voltage/Current-Clamp Recording of Ion Currents in Rabbit-Isolated Ventricular Myocytes.
Ventricular myocytes were isolated from the septal portions of hearts of 2–4-kg New Zealand White female rabbits as previously described (Liu et al., 2012). For peak and late INa, myocytes were depolarized from a holding potential of −120 mV to a test potential of −20 mV for 20 or 220 ms, respectively, at a frequency of 0.1 Hz. ATX-II (10 nM) was used to increase late INa. For IKr measurement, myocytes were depolarized from −40 to +20 mV for 3 seconds, followed by a 2.5-second step to −40 mV to record tail current, at a frequency of 0.1 Hz. IKr was measured at 36±1°C. All other currents were measured at 22 ± 1°C (see Supplemental Data).
Rabbit Myocyte Intracellular Na+ and Ca2+ Concentrations.
Myocytes were enzymatically isolated from hearts of 2–3-kg adult New Zealand white female rabbits, and intracellular Na+ and Ca2+ concentrations were measured using confocal microscopy as previously described (Yao et al., 2011; see Supplemental Data).
Rabbit-Isolated Heart Experiments.
Hearts were isolated and perfused by the method of Langendorff, as previously described (Wu et al., 2009). The atrioventricular nodal area was thermally ablated to produce heart block, and hearts were paced at a rate of 1 Hz. Monophasic APs (MAPs) from the left ventricular epicardium and pseudo 12-lead ECGs were recorded. After a 10–20-minute period of equilibration, hearts were exposed to vehicle (modified Krebs-Henseleit buffer), ATX-II, or E-4031 and then to increasing concentrations of either GS967 or ranolazine until a steady-state effect was reached. The duration of the MAP at the level at which repolarization is 90% complete (MAPD90) was measured. The QRS interval was determined from the ECG record.
Rabbit In Vivo Experiments.
The effect of GS967 on MAPD90 was determined as previously described (Wang et al., 2008). Hearts were atrial-paced beginning at a cycle length of 320 ms. Either vehicle (15% N-methyl-2-pyrrolidone, 10% solutol, and 75% water) or GS967 was then given intravenously as a bolus, followed by an infusion at each of 4 dose levels in ascending order: 30 µg/kg + 8 µg/kg/min, 30 µg/kg +16 µg/kg/min, 60 µg/kg + 32 µg/kg/min and 120 µg/kg + 64 µg/kg/min. Each dose was administered for 5 minutes. After recording the effects of the highest dose of GS967 or vehicle, the cardiac pacing cycle length was decreased in steps, from 320 to 200 ms, during continued infusion of the highest dose of GS967 or vehicle. Blood samples were taken at 1 and 5 minutes after the onset of infusion of each dose of GS967 to determine the plasma drug level.
To determine the effect of GS967 on the inducibility of TdP by clofilium in the presence of methoxamine, rabbits were first treated with either vehicle or GS967 (in randomized manner) given as a 60 µg/kg bolus, followed by a 16 µg/kg/min infusion that was maintained for the duration of an experiment. After 10 minutes, methoxamine was infused intravenously at 15 µg/kg/min, followed 10 minutes later by clofilium at 100 nmol/kg/min. The incidences of premature ventricular contractions (PVCs), ventricular tachycardia (VT; defined as three or more consecutive abnormal beats), and TdP were determined from the ECG recordings.
Rabbit Heart Ischemia–Induced Arrhythmia Model.
Rabbits were anesthetized using pentobarbital (30 mg/kg intravenous bolus, followed by infusion at a rate of 15 mg/kg/h). A coronary artery occluder was made by placing a 6-0 Prolene suture around the origin of the left circumflex artery and pulling both ends through 5 cm of PE-10 tubing. Animals were randomly assigned to vehicle, flecanide, ranolazine, and GS967 treatment groups. Vehicle or drugs were administered as an intravenous bolus injection followed by an infusion, as follows: flecanide, 1 mg/kg + 100 µg/kg/min; ranolazine, 0.75 mg/kg + 60 µg/kg/min; and GS967, 15 µg/kg + 4 µg/kg/min. After 30 min of drug infusion, the heart was subjected to 30 minutes of occlusion of the left circumflex artery. Electrical activity of the heart was monitored continuously for PVCs, VT, and ventricular fibrillation (VF). Blood samples were collected at 1, 5, 10, and 30 minutes after the onset of drug administration. At the end of an experiment, the heart was removed, and the ischemic area at risk was visualized by perfusion with 0.03% Evans Blue dye in saline after ligation of the left circumflex artery. Concentrations of drugs in blood plasma were determined after removal of protein by use of a high-performance liquid chromatograph–tandem mass spectrometric system.
Data are expressed as mean ± S.E.M. To determine the IC50 for a drug to inhibit an ion channel current, concentration-response data were fitted with the Hill equation. The statistical significance of differences in values before and after interventions in the same hearts was evaluated using repeated-measure one-way analysis of variance followed by a Student-Newman-Keuls test. An unpaired Student t test was used to determine the difference between values of two means obtained from different groups of cells or hearts. One-way analysis of variance with repeated measures, followed by Dunnett’s multiple comparison procedure, was used to determine the significance of drug and vehicle effects (at several rates of infusion in each animal) on values of MAPD90 recorded from hearts of anesthetized rabbits. Fisher’s exact test was used to analyze differences in categorical variables (i.e., the presence or absence of various arrhythmias and mortality). A difference with a P value <0.05 was considered to be statistically significant.
Effects of GS967, Flecainide, and Ranolazine on Ion Channel Currents.
The effects of GS967, flecainide, and ranolazine to reduce peak and late INa and IKr in rabbit isolated ventricular myocytes are shown in Fig. 2A–C and Table 1. The order of potency of the compounds to inhibit ATX-II–stimulated late INa was GS967 >flecainide > ranolazine (Table 1). All three compounds selectively inhibited late relative to peak INa. The ratios of values of IC50 for inhibition of peak and late INa were 78 (i.e., 1329/17 µM) and 25 (84/3.4 µM) for ranolazine and flecainide, respectively (Table 1). GS967 caused minimal inhibition (≤7.5%) of peak INa (Fig. 2, Table1) at concentrations up to 10 µM (the limit of solubility of the agent in aqueous buffer). The ratio of IC50 values for inhibition of peak and late INa by GS967 therefore exceeded 76-fold (>10/0.13 µM). To further clarify the effect of GS967 on peak INa, rabbit ventricular myocytes were paced at rates of 0.1, 1, and 3 Hz, and INa was elicited by depolarization to −20 mV from holding potentials of either −120 mV or −80 mV. GS967 (0.1, 1, and 5 µM) did not reduce peak INa in myocytes held at −120 mV and paced at rates of 0.1, 1, or 3 Hz. When myocytes were held at a diastolic potential of −80 mV, however, GS967 concentration dependently reduced peak INa at each pacing rate. GS967 at 0.1, 1, and 5 µM reduced peak INa by 17 ± 3%, 48 ± 7%, and 65 ± 4% at a pacing rate of 0.1 Hz; 18 ± 4%, 50 ± 7%, and 67 ± 4% at a rate of 1 Hz; and 25 ± 5%, 56 ± 8%, and 70 ± 4% at a rate of 3 Hz, respectively (n = 4 for all values). The results suggest that inhibition of peak INa by GS967 was significantly concentration- and voltage-dependent but minimally use-dependent. We have previously reported that block by ranolazine of NaV1.5 peak INa is use-dependent (Rajamani et al., 2009).
Ranolazine and flecainide were not selective for inhibition of late INa relative to IKr, whereas GS967 caused minimal inhibition of IKr (Fig. 2C and F). The ratios of values of IC50 for inhibition of IKr and late INa were ≤1-fold, <1-fold, and >76-fold (>10/0.13 µM) for ranolazine, flecainide, and GS967, respectively. Representative recordings of effects of GS967 on peak INa, late INa, and IKr are shown in Fig. 2D–F. GS967 (3 µM) had no significant effect on L- or T-type calcium channel currents or Na+-Ca2+ exchanger current (INCX), and 1 µM GS967 had minimal or no effect on the ATP-inhibited K+ current or on human cardiac ion channels expressed in human embryonic kidney 293 or Chinese hamster ovary cells (Supplemental Fig. 1). In addition, assays for the effects of GS967 on 162 receptors, ion channels, transporters, and enzymes, and on 442 kinases did not identify any target of the compound at concentrations that are likely to be used for inhibition of cardiac late INa (i.e., ≤ 1 µM) (Supplemental Tables 1–3).
Effects of GS967 on the Action Potential of Rabbit-Isolated Ventricular Myocytes.
GS967 (10, 100, 300 nM; Fig. 3) completely attenuated the effect of ATX-II (10 nM) to increase APD and APD variability in ventricular myocytes, with an apparent IC50 value of ∼10 nM (Fig. 3F) and decreased the beat-to-beat variability of APD (Fig. 3). The effects of 100 nM GS967 mimic the reported effects of 10 µM ranolazine (Fig. 5 in Song et al., 2004).
GS967 (0.1, 1, and 3 µM) had no significant effect on the resting (diastolic) membrane potential, APD50, or APD90 of ventricular myocytes stimulated at a rate of 1 Hz, although there was a trend for GS967 to decrease APD (Table 2). GS967 (3 µM) significantly decreased the upstroke velocity (Vmax) of the AP, whereas lower concentrations of GS967 did not (Table 2). The effect of GS967 (3 µM) to decrease Vmax of the AP upstroke was not significantly increased when the frequency of myocyte stimulation was increased from 1 to 2 and 3.3 Hz (data not shown). Flecainide was reported to reduce Vmax of the AP upstroke in rabbit ventricular myocytes by 52.5% at a concentration of 2.5 µM in a frequency (use)-dependent manner (Ikeda et al., 1985). Ranolazine was reported to reduce Vmax of the AP upstroke in dog Purkinje fibers at concentrations exceeding 10 µM (Antzelevitch et al., 2004).
GS967 Reduced Na+ and Ca2+ Overload in Rabbit-Isolated Ventricular Myocytes.
GS967 alone (0.1, 0.3, or 1 µM) had no significant effect on either systolic or diastolic Asante NaTRIUM Green or Fluo-4 fluorescence intensity (i.e., Na+ or Ca2+ concentration) in myocytes (n = 7–11) stimulated at a frequency of 1 Hz (Supplemental Fig. 2). ATX-II (5 nM) increased cytosolic concentrations of Na+ and Ca2+ during diastole, increased the duration of the Ca2+ transient, and caused loss of rhythmicity (Supplemental Fig. 3), consistent with previous findings (Yao et al., 2011). GS967 (0.1 µM) attenuated the ATX-II–induced increases of diastolic Na+ and Ca2+ concentrations by 85% and 82% (n = 6, P < 0.05), respectively, and restored normal rhythmic Ca2+ transients (Supplemental Fig. 3). Ranolazine similarly attenuated the effects of ATX-II to increase cytosolic Na+ and Ca2+ concentrations and arrhythmic activity (Yao et al., 2011).
Effects of GS967, Flecainide, and Ranolazine on the Rabbit-Isolated Heart.
Three assays were used to assess block of IKr, late INa, and peak INa by GS967 in the intact heart: the effect of drug to (1) prolong the duration of the left ventricular epicardial MAPD90, (2) shorten the duration of the left ventricular epicardial MAPD90 in the presence of the late INa enhancer ATX-II, and (3) prolong the QRS interval in the ECG, respectively. GS967 at concentrations of 0.01–3 µM did not prolong but rather shortened MAPD90 by 4 ± 2% from 184 ± 4 to 174 ± 4 ms (n = 6, P <0.05, Fig. 4A). This result is consistent with the finding that GS967 blocked IKr in rabbit-isolated myocytes by only 17% at a concentration of 10 µM and tended to shorten the APD90 of the rabbit-isolated myocyte. In contrast, both flecainide and ranolazine prolonged MAPD90 in a concentration-dependent manner. Flecainide (0.01–30 µM) prolonged MAPD90 by 29 ± 4% from 185 ± 5 to 235 ± 7 ms (n = 7, P <0.01), whereas ranolazine (0.1–100 µM) similarly prolonged MAPD90 by 28±3% from 185.1 ± 5.7 to 236.9 ± 6.9 ms (n = 11, P <0.05, Fig. 4A). These results are consistent with previous reports that flecainide and ranolazine blocked IKr with IC50 values of 2.1 and 12–14 µM, respectively, and increased APD of ventricular myocytes (Follmer et al., 1992; Antzelevitch et al., 2004).
GS967 shortened the duration of the left ventricular epicardial MAPD90 in the presence of the late INa enhancer ATX-II (Fig. 4B). ATX-II (3 nM) prolonged MAPD90 by 68 ± 11%, from 181 ± 7 (control) to 306 ± 28 ms (n = 6, P <0.01) (not shown). GS967 (0.01–3 μM, n = 6) completely reversed the ATX-II–induced increase in MAPD90 with an IC50 value of 0.21 μM (Fig. 4B), which is consistent with the IC50 value of 0.13 µM for GS967 to block late INa in ventricular myocytes. Flecainide reversed the effect of ATX-II by up to 71 ± 4%, with an IC50 value of 1.48 µM (n = 5, Fig. 4B). Ranolazine (0.1–100 µM) also shortened MAPD90 in the presence of ATX-II by a maximum of 55 ± 9% at 100 µM (n = 6, Fig. 4B), with an IC50 value of 15 ± 4 µM (P <0.001, compared with GS967). Whereas the antagonism of an ATX-II–induced prolongation of MAPD90 by GS967 was complete, those by flecainide and ranolazine were only partial. The efficacies of flecainide and ranolazine to shorten the rabbit epicardial MAPD90 in the presence of ATX-II by blocking late INa may be limited (Fig. 4B) by their concomitant effects to block IKr (which causes an increase of MAPD90).
Neither GS967 (0.01–3 µM; n = 7) nor ranolazine (up to 100 µM; n = 9) caused a significant change in the QRS interval in hearts paced at a rate of 1 Hz (Fig. 4C). In contrast, flecainide (n = 7) at concentrations of 10 and 30 µM significantly prolonged the QRS interval (Fig. 4C), consistent with its effects to block peak INa in myocytes with an IC50 value of 84 µM (Table 1). An increase of the rate of pacing from 1 to 3 and 4 Hz increased the effect of flecainide (2 µM; n = 4) but not the effects of ranolazine (10 µM; n = 4) or GS967 (3 µM; n = 4–7) on the QRS interval (Fig. 4D).
GS967 Terminated TdP Induced by ATX-II or E-4031 in the Rabbit-Isolated Heart.
Enhancement of late INa with ATX-II or reduction of IKr with E-4031 prolonged APD and caused TdP-type VT in rabbit-isolated hearts (Fig. 5). GS967 (0.01–3 µM) concentration-dependently decreased MAPD90 in the presence of either 3 nM ATX-II (Fig. 5A, n = 11) or 60 nM E-4031 (Fig. 5B, n = 6) with IC50 values of 147 ± 22 and 460 ± 56 nM, respectively, and terminated VT (both panels). We have previously demonstrated that ranolazine (5, 10 and 30 µM) reduces MAPD90 and the incidence of TdP in the rabbit-isolated heart treated with 60 nM E-4031 (Wu et al., 2009).
GS967 Decreased MAPD90 but Did Not Alter Cardiac Conduction Time in the Anesthetized Rabbit.
GS967 significantly decreased the left ventricular epicardial MAPD90 (relative to vehicle) in hearts paced at a cycle length of 320 ms (Fig. 6A). There were no significant changes in either PR or QRS intervals in the ECG at these concentrations of GS967, relative to vehicle (not shown). The PR intervals were 75 ± 5 ms before and during drug administration (n = 5 each), the QRS intervals varied from 34 ± 1 to 38 ± 2 ms in both vehicle and GS967-treated hearts (n = 4–5), and the duration of the QT interval was 169 ± 9 ms before drug administration and 159±5 ms during administration of 1.46 µM GS967 (n = 5; P >0.05). The administration of increasing doses of vehicle alone (n = 4 rabbits) caused no significant effect on MAPD90 (not shown). A decrease of the pacing cycle length from 320 to 200 ms in the presence of the highest tested dose of GS967 or vehicle (120 µg/kg bolus + 64 µg/kg/min infusion) was associated with a similar trend (P >0.05) toward a decrease of MAPD90 in both vehicle and GS967-treated (1.46 µM plasma concentration) animals (Fig. 6B), suggesting that the magnitude of the effect of GS967 on MAPD90 is not altered by a change in heart rate. We have previously observed that ranolazine (≤28 µM) did not alter the PR or QRS intervals but prolonged the QT interval in the anesthetized rabbit (Wang et al., 2008). Flecainide is known to increase QRS duration and slow conduction of electrical activity (Hellestrand et al., 1982).
GS967 Prevented the Induction of Arrhythmic Activity and TdP by Clofilium in Anesthetized Rabbits.
The induction of TdP in rabbits by administration of the IKr blocker clofilium in the presence of the α1-agonist methoxamine was inhibited by GS967 (Fig. 7A). Baseline values of mean arterial blood pressure, heart rate, and ECG intervals (PR, QRS, and QT) were 66 ± 1 mmHg, 172 ± 5 beats per minute, 76 ± 2 ms, 34 ± 2 ms, and 164 ± 4 ms (all n = 12), respectively, before drug administration. Either vehicle (n = 6 rabbits) or GS967 (n = 6 rabbits) was administered as a 60-µg/kg bolus, followed by a 16-µg/kg/min infusion. The plasma concentration of GS967 was 0.3 µmol/l at 5 minutes after onset of infusion. Neither vehicle nor GS967 alone had any significant effect on measured cardiovascular parameters. Methoxamine was then infused intravenously at a rate of 15 µg/kg/min. After the administration of methoxamine, no arrhythmias were observed in either vehicle- or GS967-treated animals, suggesting that GS967 does not have a proarrhythmic effect. The subsequent administration of clofilium in the continued presence of methoxamine prolonged the QT interval by 63 ± 4 ms from 170 ± 5 to 233 ± 6 ms in vehicle-pretreated rabbits. In the rabbits pretreated with GS967, clofilium-induced QT prolongation was significantly (P <0.01) reduced from 63 ± 4 to 46 ± 5 ms (i.e., from 158 ± 4 to 204 ± 6 ms). Clofilium and methoxamine caused PVCs, VT, and TdP in 5 (83%) of 6 rabbits pretreated with vehicle (Fig. 7A). Treatment of rabbits with GS967 before the administration of methoxamine and clofilium significantly reduced the incidences of PVCs, VT, and TdP to 2 of 6, 1 of 6, and 0 of 6 rabbits, respectively (Fig. 7A). Pretreatment of rabbits with ranolazine (14 and 28 µM) similarly decreased the incidence of TdP caused by methoxamine and clofilium in the anesthetized rabbit (Wang et al., 2008).
GS967 Decreased the Incidence of Ischemia-Induced Arrhythmias in Anesthetized Rabbits.
GS967 (0.1–0.2 µM, the approximate IC50 value for inhibition of late INa) suppressed cardiac arrhythmias caused by acute ischemia in the anesthetized rabbit (Fig. 7B). Ligation of the left circumflex coronary artery for 30 minutes resulted in occurrences of VT, VF, and mortality in 5 of 10, 6 of 10, and 4 of 10 vehicle-treated rabbits, respectively (Fig. 7B). Treatment of rabbits with flecainide at a therapeutically relevant plasma concentration of 0.5 µM before and during ischemia caused VT and VF in 7 of 7 and mortality in 6 of 7 rabbits (Fig. 7B). In rabbits treated with a therapeutically relevant concentration of ranolazine (4 µM), the incidence of VT (4 of 9), VF (6 of 9), and mortality (4 of 9), was similar to that in vehicle-treated rabbits (Fig. 7B). In contrast, in rabbits treated with GS967, the incidences of ischemia-induced VT, VF, and mortality were 2 of 8, 2 of 8, and 2 of 8 rabbits, respectively (Fig. 7B). The mean time to occurrence of fatal arrhythmia after ligation of the circumflex coronary artery in vehicle, flecainide, ranolazine, and GS967 treatment groups was 16 ± 3 (n = 4 deaths), 7 ± 1 (n = 6), 11 ± 2 (n = 4), and 8 (n = 2) min, respectively. The mean size of the ischemic area at risk (30%–34% of the left ventricular mass) was not significantly different among treatment groups.
The results of this study demonstrate that GS967 is a potent and highly selective blocker of late INa in female rabbit hearts and isolated ventricular myocytes. Moreover, GS967 was not proarrhythmic, but rather reduced the induction of arrhythmias by an enhancer of late INa, inhibitors of IKr, or myocardial ischemia in rabbit hearts ex vivo and in vivo (Figs. 5, 7). The results strongly implicate late INa as a cause of arrhythmogenesis. Because of its selectivity, GS967 should be a useful tool for determination of the roles of late INa in physiologic and pathologic processes in cardiac and other excitable cells, both in vitro and in vivo.
The potency of GS967 to inhibit late INa induced by 10 nM ATX-II in isolated ventricular myocytes was more than 10-fold greater than that of either flecainide or ranolazine: 0.13 ± 0.01 versus 3.4 ± 0.4 and 17.1 ± 1.2 µM, respectively. In intact rabbits, GS967 plasma concentrations of 0.1–0.3 µM were effective to suppress arrhythmic activity induced by clofilium-methoxamine or ischemia (Fig. 7). In comparison, ranolazine plasma concentrations of 14–28 µM are required to reduce the incidence of TdP in clofilium-methoxamine–treated rabbits (Wang et al., 2008). GS967 (10 µM) inhibited IKr and peak INa by only 17% and 7%, respectively. Consistent with the absence of an effect to block IKr or peak INa in myocytes, GS967 did not prolong APD or increase the duration of the QRS interval in either isolated hearts or anesthetized rabbits. In contrast, ranolazine and flecainide increased APD in isolated perfused hearts (Fig. 4), and flecainide increased the QRS interval (Fig. 4) and mortality during acute myocardial ischemia (Fig. 7). Thus, GS967 is a more selective inhibitor of late INa than either flecainide or the current standard ranolazine, which is reported to inhibit IKr and blocks α- and β-adrenergic receptors at concentrations only two-fold higher than the 2–8 μM used to inhibit late INa (Antzelevitch et al., 2004; Zhao et al., 2011).
Physiologic Importance of Late INa.
The magnitude of endogenous late INa is small (≤60 pA) in the normal heart, but its contribution as an inward current during the AP plateau is proarrhythmic in the presence of blockers of IKr (Wu et al., 2009, 2011) (Fig. 5B). Because of the high input resistance during the AP plateau, even a small inward current, such as late INa, can contribute significantly to the AP waveform and duration. However, GS967 caused small but consistent decreases in APD90, MAPD90, and the QT interval in isolated myocytes (Table 2) and hearts (Fig. 4A) and anesthetized rabbits (Fig. 6). This result suggests that endogenous late INa contributes to action potential duration. In hearts treated with either E-4031 (Fig. 5) or clofilium-methoxamine (Fig. 7) to reduce IKr, GS967 shortened MAPD90 and prevented (Fig. 7) or terminated (Fig. 5) the induction of arrhythmic activity. This finding confirms earlier data demonstrating the proarrhythmic role of endogenous late INa when repolarization reserve is reduced (Wu et al., 2009, 2011). GS967 was more potent and efficacious in reversing the effects of ATX-II than that of E-4031 (Fig. 5). An explanation for this finding is that, in the presence of ATX-II, GS967 inhibits both the endogenous and enhanced late INa, whereas in the presence of E-4031, GS967 reduces only the endogenous late INa. Taken together, these findings provide evidence that reduction of endogenous late INa is beneficial in preventing ventricular arrhythmias when repolarization reserve is reduced.
Pathologic Role for late INa.
An enhanced late INa is pathologic, because it destabilizes AP repolarization, causes Na+-dependent Ca2+ overload and intracellular acidosis, and promotes phosphorylation and activation of CaMKII (Yao et al., 2011). The late INa enhancer ATX-II caused Na+ and Ca2+ loading of rabbit isolated myocytes (Supplemental Fig. 3), markedly prolonged the AP (Fig. 3), increased the variability of APD (Fig. 3), and caused TdP-like VT (Fig. 5A). GS967 (0.1 µmol/l) reduced ATX-II–induced late INa (Figs. 2 and 3) and reversed the Na+-dependent Ca2+ overload (Supplemental Fig. 3) and the proarrhythmic effects of ATX-II. The present results are consistent with the increasingly recognized roles of an enhanced late INa to prolong repolarization, to cause Na+ and Ca2+ loading of myocardial cells and, consequently, electrical dysfunction and arrhythmias (Ver Donck et al., 1993; Haigney et al., 1994; Le Grand et al., 1995; Sicouri et al., 1997; Saint, 2006; Undrovinas and Maltsev, 2008; Wu et al., 2009).
Blockers of Late INa as Therapeutic Agents: Opportunities and Challenges.
The effects of GS967 on rabbit ventricular myocytes, rabbit isolated hearts, and anesthetized rabbits show that the compound reduces the proarrhythmic consequences of either enhancing late INa or inhibiting IKr. The finding suggests that selective inhibition of late INa will be effective to increase repolarization reserve and prevent the induction of arrhythmias in the face of various interventions that delay and/or increase the instability of ventricular repolarization in the heart. In failing hearts, late INa is enhanced, and nonselective inhibitors of the current, albeit not as potent, selective, and efficacious as GS967, shorten the APD, improve diastolic function, and suppress arrhythmic activity (Undrovinas et al., 2006; Undrovinas and Maltsev, 2008). Inhibition of late INa has not been shown to have adverse consequences in either normal or diseased hearts. The adverse effects of nonselective inhibitors of late INa, such as flecainide and ranolazine, can be attributed to their effects to cause use-dependent reduction of peak INa and/or reduction of IKr, respectively. The challenges are therefore to develop a selective inhibitor of late INa with good pharmaceutical properties and to confirm that inhibition of late INa is indeed safe and beneficial therapeutically. The potent effect of GS967 to block late INa, reduction by the compound of peak INa in a voltage- but not a use-dependent manner, and its effect to decrease arrhythmic activity in the ischemic rabbit heart suggest that the compound may be beneficial in the context of myocardial ischemia.
The effects of GS967, ranolazine, and flecainide to reduce arrhythmic activity induced by ischemia in the anesthetized rabbit were determined using single doses of each agent. These doses produced therapeutically relevant plasma concentrations of ranolazine and flecainide and an IC50 concentration of GS967 to reduce late INa. However, it is possible that improved efficacy could be demonstrated using different dosing regimens. Although ranolazine (4 µM) did not reduce the arrhythmic activity induced by ischemia in the present study, it was reported to do so in a study of anesthetized rats (Dhalla et al., 2009). The different results may reflect either a species or a methodological (i.e., 5 versus 30 minutes duration of ischemia) difference between the studies. Additional studies are also needed to confirm the high potency of GS967 that is reported in this study of rabbit heart tissues. Although we have tested for effects of GS967 on hundreds of potential targets, further testing is nonetheless needed (e.g., the effect of GS967 on ryanodine receptors). This testing should include preparations of remodeled and normal cardiac tissues.
In conclusion, the results demonstrate that GS967 is a selective inhibitor of late INa. GS967 shortened the duration of the ventricular action potential and reduced the proarryhythmic effects of ATX-II, ischemia, and IKr block. GS967 was not proarrhythmic. The results provide compelling evidence of the major contribution of late INa to repolarization reserve and genesis of cardiac arrhythmias and indicate a potential important contribution of this current to Na+-dependent intracellular Ca2+ homeostasis.
The authors thank Dr. Wayne Giles for a critical review of the manuscript.
Participated in research design: Belardinelli, Shryock, Rajamani, Dhalla, Wu, Smith-Maxwell, Crumb.
Conducted experiments: Liu, Wang, El-Bizri, Hirakawa, Karpinski, C. H. Li, Hu, X.-J. Li, Crumb.
Contributed new reagents or analytic tools: Koltun, Zablocki.
Performed data analysis: Crumb, El-Bizri, Hirakawa, Wang, Rajamani, Wu, Dhalla.
Wrote or contributed to the writing of the manuscript: Shryock, Belardinelli, Rajamani, Dhalla, Smith-Maxwell, Crumb, Yao.
- action potential
- action potential duration
- Anemonia sulcata toxin
- rapidly activated delayed rectifier potassium current
- sodium current
- monophasic action potential
- monophasic action potential duration
- premature ventricular contraction
- torsades de pointes
- ventricular fibrillation
- ventricular tachycardia
- Received July 29, 2012.
- Accepted September 19, 2012.
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics