Department of Pharmacology, Cardiovascular Research Centre and the Alberta Diabetes Institute, Li Ka Shing Research Centre, University of Alberta, Edmonton, AB, Canada
Peter E Light, Department of Pharmacology, Cardiovascular Research Centre and the Alberta Diabetes Institute, Li Ka Shing Research Centre, University of Alberta, i-OCS Liha Shing Research Centre, Edmonton, AB, Canada T6G 2E1. E-mail: email@example.com
Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with an increased risk for stroke, heart failure and cardiovascular-related mortality. Candidate targets for anti-AF drugs include a potassium channel Kv1.5, and the ionic currents IKACh and late INa, along with increased oxidative stress and activation of NFAT-mediated gene transcription. As pharmacological management of AF is currently suboptimal, we have designed and characterized a multifunctional small molecule, compound 1 (C1), to target these ion channels and pathways.
We made whole-cell patch-clamp recordings of recombinant ion channels, human atrial IKur, rat atrial IKACh, cellular recordings of contractility and calcium transient measurements in tsA201 cells, human atrial samples and rat myocytes. We also used a model of inducible AF in dogs.
C1 inhibited human peak and late Kv1.5 currents, frequency-dependently, with IC50 of 0.36 and 0.11 μmol·L−1 respectively. C1 inhibited IKACh (IC50 of 1.9 μmol·L−1) and the Nav1.5 sodium channel current (IC50s of 3 and 1 μmol·L−1 for peak and late components respectively). C1 (1 μmol·L−1) significantly delayed contractile and calcium dysfunction in rat ventricular myocytes treated with 3 nmol·L−1 sea anemone toxin (ATX-II). C1 weakly inhibited the hERG channel and maintained antioxidant and NFAT-inhibitory properties comparable to the parent molecule, resveratrol. In a model of inducible AF in conscious dogs, C1 (1 mg·kg−1) reduced the average and total AF duration.
Conclusion and Implications
C1 behaved as a promising multifunctional small molecule targeting a number of key pathways involved in AF.
Atrial fibrillation (AF) is the most common type of cardiac arrhythmia and the incidence of AF is increasing with an ageing population (Naccarelli et al., 2009; McManus et al., 2012). While AF does not generally induce sudden cardiac death, AF increases the risk of stroke fivefold (Wolf et al., 1991; Hart and Halperin, 2001; McManus et al., 2012), can cause tachycardia-induced cardiomyopathy (Nerheim et al., 2004) and adverse electrical and structural remodelling of the heart (Franz et al., 1997; Anter et al., 2009). Furthermore, ∼30% of patients undergoing coronary artery bypass graft (CABG) surgery develop transient post-operative AF (Andrews et al., 1991; Mathew et al., 2004). Current pharmacotherapy of AF is still suboptimal because some of the drugs used for rhythm control also display significant risk for acquired long QT syndrome and Torsades des Pointes arrhythmias (Taira et al., 2010) and may increase the incidence of vascular events (Connolly et al., 2011). Therefore, there is a significant need for the development of improved therapeutic agents for AF.
Over the past decade, studies on the cellular pathways in AF have revealed potential therapeutic targets to develop new anti-AF drugs. One important target is the ultra-rapid potassium channel (IKur/Kv1.5; channel and receptor nomenclature follows Alexander et al., 2013) preferentially expressed in the atria, but not the ventricle (Wang et al., 1993; Gaborit et al., 2007; Tamargo et al., 2009; Ravens and Wettwer, 2011). In addition to Kv1.5 channel, IKACh (Ehrlich and Nattel, 2009), voltage-gated sodium channels (Burashnikov et al., 2007; Burashnikov and Antzelevitch, 2012), oxidative stress (Carnes et al., 2001; Mihm et al., 2001) and activation of the transcription factor NFAT (Lin et al., 2004) have all been implicated in AF development. Given the complex aetiology of AF, it has been suggested that drugs targeting multiple pathways involved in AF development may be more effective (Dobrev and Nattel, 2010; Dobrev et al., 2012). To date, AF drugs have been exclusively targeted towards ion channels. However, non-ionic remodelling events also likely contribute to the initiation and maintenance of AF as well. It has been further suggested that targeting maladaptive remodelling events in AF may also be required for effective control of AF (Burashnikov and Antzelevitch, 2011). Therefore, based on these criteria, a multifunctional small-molecule drug for AF should ideally possess the following properties: (i) frequency-dependent Kv1.5 inhibition; (ii) IKACh inhibition; (iii) late sodium current inhibition; (iv) lack of hERG channel inhibition; (v) display atrial specificity, that is, no effect on repolarization and excitation–contraction (EC) coupling in ventricular tissue; (vi) antioxidant properties; (vii) NFAT inhibition; and (viii) functional efficacy in a large-animal model of inducible AF.
Resveratrol is a bioactive polyphenol found in red grape products that has been investigated extensively for its cardioprotective and anti-ageing properties (Brisdelli et al., 2009; Dolinsky and Dyck, 2011). We have previously shown that resveratrol (i) inhibits NFAT activation and hypertrophic remodelling in cardiac tissue (Chan et al., 2008; Dolinsky et al., 2009) and (ii) inhibits the late INa current preferentially over peak current (Wallace et al., 2006). Based on these multifunctional properties, the aim of the present study was to synthesize, using resveratrol as the parent molecule, a novel small molecule called compound 1 (C1) and to characterize its antioxidant, NFAT-inhibitory effects and electrophysiological profile with respect to Kv1.5, IKACh, late INa and hERG. We also determined its in vivo efficacy in a model of inducible AF in dogs.
Chemical synthesis of compounds
Details of the chemical synthesis used to generate four unique resveratrol hybrid bis-ortho-substituted biphenyl diamide derivatives are provided in Supporting Information Appendix S1.
The human Nav1.5 and hERG clones were generously provided by Dr Arthur Brown (Case Western Reserve University, USA) and Dr Henry Duff (University of Calgary, Canada) respectively. Expression vectors were co-transfected into the tsA201 cell line (Sigma Aldrich, St Louis, MO, USA) with a GFP expression vector using the calcium phosphate precipitation technique. TsA201 cells are a transformed HEK293 cell line expressing the temperature-sensitive antigen to increase expression levels of recombinant proteins. Transfected cells were then used within 48 h. All in vitro experiments were performed at room temperature (20–22°C).
Recordings were made from single tsA201 cells stably expressing the human heart Kv1.5 gene. Pipettes were pulled from borosilicate glass capillary tubing (Warner Instruments, Hamden, CT, USA) and tips were fire-polished, producing resistances of 1–3 MΩ. The pipette solution contained (in mmol·L−1): 140 KCl, 1 CaCl2, 1 MgCl2, 10 HEPES and 10 EGTA. The pH was adjusted to 7.3 with KOH. A concentration of 2 mmol·L−1 MgATP was added immediately before use. Cells were bathed in extracellular solution contained (in mmol·L−1): 135 NaCl, 5.4 KCl, 1 CaCl2, 1.2 MgCl2, 10 HEPES and 5 glucose (pH adjusted to 7.4 with NaOH). Solutions were applied to cells using a multi-input perfusion pipette (switch time <2 s). Once a giga-seal was formed, the patch was ruptured and the whole-cell voltage-clamp technique was used to record Kv1.5 currents. Currents were elicited by 120 ms duration test pulses to +40 mV from a holding potential of −100 mV with a cycle length of 0.3, 1 or 3 Hz respectively. Data were acquired using an Axopatch 200B patch-clamp amplifier and Clampex 8.2 software (Axon Instruments, Foster City, CA, USA).
Single tsA201 cells transfected with mammalian expression vectors encoding the human heart Nav1.5 channel were used for recording of whole-cell Nav1.5 currents, as described previously (Wallace et al., 2006). Voltage steps from a holding potential of −120–40 mV every 10 mV was used to generate the current–voltage (I –V) relationship for peak INa.
tsA201 cells were transfected with a mammalian expression vector encoding the hERG channel clone (Kv11.2) as described for Nav1.5 channels above and hERG currents were recorded as described previously (Kikuchi et al., 2005).
Human atrial IKur currents
For experiments on human atrial IKur (Kv1.5) atrial tissue was collected from 7 male patients aged 40-75 years undergoing CABG surgery at the University Hospital, University of Alberta, Canada. We obtained prior ethical approval from the University of Alberta human ethics board and informed patient consent. The atrial appendage samples were collected during surgery and enzymically dissociated into single atrial myocytes as described below. Human atrial appendage samples were cut into small pieces and washed 3 times for 10 min with 10ml of a calcium-free Tyrode's solution in a shaking water bath at 32oC. The sample was enzymically digested for 45 min in 10 ml of Tyrode's solution containing 20 mg Type II collagenase (Worthington, Lakewood NJ, USA) and 4 mg protease type XIV (Sigma Aldrich) in a shaking water bath at 32oC. After 45 min the supernatant is removed and 10 ml solution containing 20 mg Type II collagenase only was added and the sample further digested for 10 min at 32oC. Isolated cells are removed in the supernatant, centrifuged and resuspended in Tyrode's solution. IKur currents were then measured using a modification of the whole-cell recording protocol to inactivate Ito currents, as described previously (Cai et al., 2007).
All animal care and experimental procedures involving adult and neonatal rats complied with the guidelines of the Canadian Council for Animal Care and the project was approved by the University of Alberta research ethics office. All studies involving animals are reported in accordance with the ARRIVE guidelines for reporting experiments involving animals (Kilkenny et al., 2010; McGrath et al., 2010). A total of 15 adult rats, 69 neonatal rats and 5 dogs were used in the experiments described here.
Rat atrial IKACh and recombinant IK1 currents
Rat atrial myocytes were isolated as described previously (Komukai et al., 2002) from 15 adult male Sprague-Dawley rats (150-200 g) obtained from the animal facility at Biological Sciences, University of Alberta. Whole-cell recordings of carbachol-induced IKACh currents were made using the same solutions used for Kv1.5 and a voltage step protocol from −110 to +50 mV with a holding potential of −80 mV. IK1 whole-cell currents were measured in tsA201 cells transiently transfected with the KIR2.1 clone using the same solutions and voltage protocols used for IKACh currents.
Cell shortening and calcium transient recordings
These were measured in rat ventricular myocytes prepared from adult Sprague-Dawley rats killed by an overdose of pentobarbital (150 mg kg−1, i.p.). The hearts were removed and right ventricular myocytes were then obtained by enzymic dissociation using standard protocols, which have been described previously (Bouchard et al. 1993; Light et al. 1998). Cell shortening and calcium transients were measured using standard procedures previously published by our laboratory (Baczko et al., 2005; Wallace et al., 2006).
NFAT reporter assay
Neonatal rat ventricular myocytes (NRVM) were isolated from the hearts of 1-3-day-old neonatal rat pups and cultured in DMEM/F-12, containing 5% fetal bovine serum, 10% horse serum, 50 μg ml−1 gentamicin, and 1% penicillin-streptomycin for 16h, as described previously (Kovacic et al. 2003). 24h prior to the initiation of experiments, the medium was changed to serum-free DMEM/F-12 containing 50 μg ml−1 gentamicin supplemented with 1× insulin, transferrin, sodium selenite (ITS)+3 liquid media supplement (Chan et al., 2008). All subsequent treatments of the myocytes were performed in this serum-free medium for either 1 or 24 h. NRVM were infected with Ad.GFP or Ad.NFAT-Luc-Promoter adenovirus (Seven Hills Bioreagents, Cincinnati, OH, USA) at an infection multiplicity of 10. Cells were treated 24 h post-infection with vehicle or varying concentrations of C1 or resveratrol and/or angiotensin II (1 μmol·L−1 for 24 h). Cells were harvested with the reporter lysis buffer supplied in the luciferase assay system kit (Promega, Madison, WI, USA) and luminescence according to the manufacturer's instructions.
A diphenylpicrylhydrazyl (DPPH) assay (Pyrzynska and Pekal, 2013) was used to determine the antioxidant capacity of C1 and resveratrol. A ethanolic solution of H2O2 (200 nmol·L−1) was mixed with DPPH alone, or DPPH with either resveratrol or C1 at various concentrations. The final concentration of ethanol in the experimental solution was 0.1%. The absorbance of these solutions was measured at 517 nm, with a lower absorbance representing an increase in antioxidant capacity.
Model of induced AF in vivo
All experiments were carried out in compliance with the Guide for the Care and Use of Laboratory Animals (USA NIH publication No 85–23, revised 1996) and approved by the Department of Animal Health and Food Control of the Ministry of Agriculture and Rural Development, Hungary (XIII/01031/000/2008). Adult male Beagle dogs (age: 24-26 months, weight: 14–16 kg; supplied by the University of Szeged Experimental Animal Facility, Hungary) were anaesthetized with ketamine (Richter Gedeon Ltd., Hungary; induction: 10 mg kg−1, maintenance: 2 mg kg−1, every 20 min) and xylazine (CP-Pharma Handelsges, Germany; induction: 1 mg kg−1, maintenance: 0.2 mg kg−1, every 20 min) and were mechanically ventilated. The implantation procedure was carried out under antibiotic coverage as follows: amoxicillin/clavulanic acid (1000 mg/200 mg i.v.; Richter Gedeon Ltd.) and gentamicin (40 mg i.v.; Sandoz GmbH, Kundl, Austria) given before the operation, amoxicillin/clavulanic acid (500 mg/125 mg orally, twice a day for 5 days; Augmentin 500mg/125mg®; GlaxoSmithKline Ltd., Hungary) given following the operation. For peri-operative analgesia metamizole sodium (1000 mg i.v., 1g/2ml; Sanofi-Aventis Hungary Ltd., Hungary) and tramadol (50 mg i.v., TEVA Ltd., Hungary, under licence from Grünenthal GmbH, Aachen, Germany) were administered.
The procedure involved the implantation of two bipolar pacemaker electrodes (Synox SX 53-JBP® and Synox SX 60/15-BP®, Biotronik Hungary Ltd., Hungary) into the right atrial appendage and apex of the right ventricle, the electrodes were connected to pacemakers (Logos DS® and Philos S®, Biotronik Hungary Ltd., Hungary) in subcutaneous pockets in the neck area, radiofrequency catheter ablation of the AV node was performed to avoid high atrial pacing rates propagating into the ventricles. The ventricular pacemaker was set between 80 to 90 beats/min, following the baseline heart rate of the dog before the operation.
Following recovery from surgery (3 days), high frequency right atrial pacing was started at 400 beats/min, maintained for 6 to 7 weeks before the experiments to allow electrical remodeling of the atria (monitored by the measurement of the right atrial effective refractory period (AERP) every second day). The AERPs were measured at basic cycle lengths (BCL) of 150 and 300 ms with a train of 10 stimuli (S1) followed by an extrastimulus (S2), with the AERP defined as the longest S1S2 interval that did not produce a response. AERP shorter than 80 ms could not be measured in conscious animals (pacemaker measurement limit). No animals were lost between pacemaker, pacemaker electrode implantation and experimental use.
On the day of the experiment atrial pacing was stopped, continuous recording of the electrocardiogram commenced using precordial leads and the AERP was measured. A control set (25 times) of 10-second long rapid atrial bursts (800 beats/min, at twice threshold) were performed in order to induce atrial fibrillation in conscious dogs preceded by a bolus infusion of vehicle (20 mL of a mixture of DMSO + β-hydroxypropyl-cyclodextrin + saline (all from Sigma, St. Louis, USA), DMSO concentration less than 0.1%, infused in 15 min). During the control 25 bursts and subsequent AF episodes, a continuous infusion of vehicle was maintained (in a volume of 1.7 mL/kg/min). Following the measurement of AERP, Compound 1 (C1) was infused in a dose of 0.3 mg/kg (in 15 min bolus + maintenance) and AF was again induced 25 times. An identical procedure was repeated in every dog with 1 mg/kg dose of C1. The incidence of AF, the total duration of AF, the average duration of AF episodes were measured along with changes in atrial refractory period and QT interval. QT intervals were measured on dogs with pacemaker implantation before the 12th burst and were not corrected for heart rate since QT measurements were made at heart rate set to 80 beats/min by the ventricular pacemaker in each animal. Experiments were done in freely moving conscious dogs so that any interference from anesthetics could be ruled out.
HPLC assay of plasma levels of C1
Blood samples were collected from the vena cephalica from the right foreleg for later determination of the plasma levels of C1, before and after administration of C1 infusion at designated time points. For each time point (5 min before the administration of 0.3 mg/kg C1, 30 min after the administration of 0.3 and 1 mg/kg C1) 4 mL of blood was taken into tubes containing lithium heparin. The tubes were immediately centrifuged at 4 oC for 15 min at 3000g and the separated plasma was stored in a freezer at −20oC.
For each sample, 100 μL of plasma was combined with 100 μL of deionized water and the entire volume was loaded into individual wells of a 96-well Isolute SLE+ (supported liquid extraction) plate from Biotage (PN 820-00200-P01; Charlotte, NC, USA). The samples were eluted into a 96-well collection plate using 1 ml of tert-butyl methyl ether. The solvent was completely evaporated using a TurboVap 96 and samples were reconstituted in 200 μL of 0.1% formic acid in a 50/50 water/acetonitrile mixture. Analysis was performed using an Acquity UPLC-TQD (Waters) in multiple reaction monitoring mode (positive electrospray).
Data are expressed as mean ± SEM. In whole animal experiments, each animal served as its own control, after one-way anova, the groups were compared in pairs by means of Student's t-test. In cell-based experiments, one-way anovas or paired Student's t-tests were used as appropriate. A level of P < 0.05 was considered to be significant..
Angiotensin II, carbachol, DPPH, E-4031 and resveratrol were obtained from Sigma Aldrich. The anemone toxin ATX II, was obtained from Alomone Labs (Jerusalem, Israel).
Effects of C1 on Kv1.5 currents
The inhibitory effects on Kv1.5 currents of four novel resveratrol derivatives (Figure 1A) were compared with those of the parent molecule, resveratrol. Resveratrol is a weak inhibitor of Kv1.5 currents (IC50 = 66 μmol·L−1; Figure 1B). Of the four derivatives synthesized, C1 was found to be the most potent Kv1.5 inhibitor (IC50s = 0.36 μmol·L−1 and 0.11 μmol·L−1 for peak and late current inhibition, respectively; Figure 1B–E). Compounds 2–4 displayed intermediate Kv1.5 peak current inhibitory potencies of 8.3, 10.9 and 11.2 μmol·L−1 respectively (Figure 1B). Therefore, only C1 was selected for further characterization in this study.
Frequency-dependent effects of C1 on Kv1.5 currents
The frequency-dependent effects of C1 were tested on whole-cell Kv1.5 currents in the tsA201 cell line stably expressing Kv1.5 channels. At 1 Hz, 0.3 μmol·L−1 C1 displayed a time-dependent inhibition of Kv1.5 currents, reaching a steady-state after 15 s (Itest/Icontrol = 0.74 ± 0.02). In contrast, at a higher stimulation frequency of 3 Hz, the observed C1 inhibition was increased, reaching a steady-state level of 0.50 ± 0.03 when compared with control (n = 5 cells; Figure 2A,B).
Effects of C1 on human atrial IKur currents
Single atrial myocytes were isolated from human atrial appendage samples obtained during CABG surgery and whole-cell IKur currents measured. Application of C1 (0.3 μmol·L−1) reversibly inhibited IKur to 31.0 ± 1.8% of control values (n = 3 cells; Figure 2C,D), a value similar to that observed on recombinant Kv1.5 (Figures 1D and 2A,B).
Effects of C1 on rat atrial IKACh and recombinant IK1 currents
Single atrial cells were isolated from rat hearts and IKACh currents were elicited by application of carbachol. C1 significantly inhibited carbachol-induced rat atrial IKACh currents with an IC50 of 1.9 μmol·L−1 (Figure 3). In contrast, 3 μmol·L−1 C1 had no significant inhibitory effect on recombinant IK1 (Kir2.1) whole-cell currents transiently expressed in tsA201 cells (see Supporting Information Appendix S1 for results).
Effects of C1 on human Nav1.5 currents
The effects of C1 were tested on peak and late recombinant Nav1.5 whole-cell currents. Application of 3 μmol·L−1 C1 resulted in a 50% inhibition of peak current (Figure 4A,B). To establish the concentration–response curves for peak and late Nav1.5 currents, cells were treated with the sea anemone toxin (ATX-II) (3 nmol·L−1) to induce the late-current component and to test the inhibitory effects of C1 at differing concentrations. C1 preferentially inhibits the late current when compared with peak current (IC50s = 1.1 μmol·L−1 vs. 3.2 μmol·L−1, respectively; Figure 4C,D).
Effects of C1 on hERG currents
Inhibition of the hERG (IKr, Kv11.2) is highly undesirable and may result in arrhythmogenic QT prolongation (Sanguinetti et al., 1996; Sanguinetti and Tristani-Firouzi, 2006; Taira et al., 2010). Therefore, we tested the effects of C1 on recombinant whole-cell hERG channel currents. hERG currents were positively identified by their characteristic inactivation at positive holding potentials, large tail currents and inhibition by 1 μmol·L−1 E-4031 (Figure 5A). Construction of concentration–inhibition curves revealed that C1 is a weak inhibitor of peak and tail hERG currents (Figure 5B,C), with IC50s of 30 and 25 μmol·L−1, respectively, values that are 100-fold higher than those inhibiting peak and late Kv1.5 currents.
C1 antioxidant capacity and NFAT activation
Antioxidant therapy has been shown to reduce the incidence of post-operative AF in humans (Carnes et al., 2001), and reactive oxygen species (ROS) also directly activate Kv1.5 channels (Caouette et al., 2003). As resveratrol is a known antioxidant, we compared the antioxidant properties of C1 with resveratrol. At 10 μmol·L−1, resveratrol and C1 displayed significant antioxidant capacities (0.59 ± 0.04 vs. 0.77 ± 0.02 of maximal DPPH 517 nm absorbance signal; Figure 6A). At 100 μmol·L−1, these values were 0.09 ± 0.02 and 0.19 ± 0.02 for resveratrol and C1 respectively.
We have previously shown that resveratrol inhibits NFAT activation induced by phenylephrine contributing to a reduction in maladaptive hypertrophy in NRVMs (Chan et al., 2008). Accordingly, we tested the effects of C1 on NFAT activation in NRVMs. Treatment of NRVMs with 0.01–25 μmol·L−1 of C1 resulted in a significant reduction in NFAT activity when compared to no treatment (Figure 6B). These experiments revealed that C1 significantly reduced NFAT activity at concentrations of 0.1 μmol·L−1 and higher, with 1 μmol·L−1 of C1 exhibiting half-maximal inhibition (Figure 6B). As 1 μmol·L−1 of C1 inhibited NFAT activation by 50%, we used this concentration to determine whether C1 was able to inhibit NFAT activation in response to a pro-hypertrophic stimuli. NRVMs were infected as above and treated with vehicle, AngII (1 μmol·L−1), C1 (1 μmol·L−1) or AngII + C1 (1 μmol·L−1) for 24 h. As expected, treatment of NRVM with 1 μmol·L−1 of AngII significantly increased NFAT activation by 504 ± 73%, whereas 1 μmol·L−1 of C1 inhibited NFAT activation by approximately 50%. C1 also significantly reduced AngII-induced activation of NFAT to 160 ± 23% of control values (Figure 6C.).
Effects of C1 on ventricular myocyte excitation–contraction coupling
A desirable property of potential compounds for AF is a minimal effect on EC coupling in ventricular myocytes. Single rat ventricular myocytes were field-stimulated at 1 Hz; contractility and calcium transients were then measured by edge detection and calcium imaging. Application of 3 μmol·L−1 C1 had no significant effect on (i) diastolic resting cell length or calcium levels or (ii) peak systolic contractility or calcium transient amplitude (Figure 7A,B).
The results presented in Figure 3 indicate that C1 inhibits the late component of the cardiac voltage-gated sodium channel current compared with peak current. Therefore, the effects of C1 were tested in a single-ventricular myocyte model of late sodium current-induced disturbances in calcium homeostasis by pretreatment of myocytes with 3 nmol·L−1 ATX-II. ATX-II pretreatment results in markedly elevated diastolic calcium levels and the loss of uniform calcium transients (Figure 7C,E,F). In contrast, application of C1 (1 μmol·L−1) significantly reduced the magnitude of ATX-II-induced diastolic calcium elevations (C1 = 5.9 ± 2.9% vs. control = 57.8 ± 5.9% of maximal transient amplitude, P < 0.01, n = 5 cells) and significantly delayed the onset of calcium transient dysfunction (C1 = 5.3 ± 0.7 min vs 2.4 ± 0.4 min for control, P < 0.01, n = 10 cells, Figure 7D, E, F).
Effects of C1 in a model of inducible AF in dogs
Right atrial effective refractory period (AERP) measurements before the commencement of rapid atrial pacing at 400 beats·min−1 yielded values of 118 ± 3.7 and 130 ± 3.2 ms in conscious dogs (n = 5; at basic cycle lengths of 150 and 300 ms respectively). Rapid right atrial pacing for 6–7 weeks resulted in a significant decrease of right AERP, as AERP decreased below 80 ms in all five animals, and were defined as 79 ms for the following reason: the lower adjustable limit for S1–S2 intervals in the pacemakers available for this study was 80 ms, therefore the exact AERP after 7 weeks of rapid atrial pacing and immediately before C1 administration could not be determined (measured AERP values were less than 80 ms in all dogs at both basic cycle lengths (BCLs) as an S1–S2 interval of 80 ms still evoked a P wave), although statistical comparison of AERP values was not possible. AERP measurements following C1 administration yielded 87.5 ± 2.50 ms at 150 ms BCL in four animals (in one animal AERP was less than 80 ms) and 90 ± 3.16 ms at 300 ms BCL, respectively, following 0.3 mg·kg−1 C1; and 90.0 ± 3.16 ms at 150 ms BCL, 100 ± 3.16 ms at 300 ms BCL, respectively, following 1 mg·kg−1 C1 (n = 5; with the exception of AERP measurement at 150 ms BCL following 0.3 mg·kg−1 C1).
The incidence of AF was not influenced by administration of C1 (86.4 ± 6.4% in control vs. 71.2 ± 15.7% and 66.4 ± 15.7% following 0.3 and 1 mg·kg−1 C1, respectively, both P > 0.05). However, as typical ECG recordings and grouped data show, the total duration of AF was significantly reduced by 1 mg·kg−1 C1 administration (Figure 8B–D) and the average duration of AF episodes was also significantly decreased by 1 mg·kg−1 C1 (Figure 8A–D). These results clearly demonstrate in vivo efficacy of C1 against AF in conscious dogs. HPLC analysis of blood plasma collected within 5 min of i.v. bolus injection showed that we obtained concentration ranges of 0.32–0.79 and 0.7–3.0 μmol·L−1 for the 0.3 and 1 mg·kg−1 doses respectively.
None of the investigated doses exhibited any QT interval prolonging effect in five conscious dogs, yielding 261.7 ± 9.37 ms in control vs. 260.9 ± 8.77 ms (P = 0.15) following 0.3 mg·kg−1 C1 and 257.3 ± 9.05 ms (P = 0.55) following 1 mg·kg−1 C1.
Ionic effects of C1
As decreased atrial action potential refractory period contributes to the development of AF, the design of atrial-specific repolarization-delaying drugs has received much recent attention (Tamargo et al., 2009; Dobrev and Nattel, 2010). Specifically, inhibition of the IKur/Kv1.5 repolarizing potassium channel that is predominantly expressed in atria rather than ventricle (Wang et al., 1993; Gaborit et al., 2007; Tamargo et al., 2009) has emerged as an attractive therapeutic target (Tamargo et al., 2009). Results generated in this study on recombinant Kv1.5 and human atrial myocyte IKur currents indicate that C1 is an effective inhibitor of this atrial-specific ion channel. The calculated IC50 values for Kv1.5 channels are in the 0.11–0.36 μmol·L−1 range for late and peak Kv1.5 current inhibition, which are 180- to 600-fold lower than observed for the parent molecule, resveratrol (66 μmol·L−1). Furthermore, these IC50 values are 10- to 30-fold lower than those reported for the recently developed AF drug vernakalant that is in clinical use in Europe (Camm et al., 2012). Further analysis revealed that C1 is a more effective Kv1.5 channel inhibitor at a higher stimulatory frequency of 3 Hz compared with 1 Hz, indicating that C1 may display increased inhibitory efficacy in fibrillatory tissue. Therefore, our results indicate that C1 has a favourable pharmacological and frequency-dependent inhibitory profile for Kv1.5 currents that are both desirable properties for novel AF therapeutic agents.
Recent research has suggested that inhibition of peak and late atrial sodium currents may also be an attractive strategy to suppress AF (Burashnikov et al., 2007; Burashnikov and Antzelevitch, 2012). Inhibition of peak sodium current in a frequency-dependent manner may reduce the occurrence of premature action potentials and reduce the incidence of AF. Whereas the induction of the Nav1.5 late current may not only increase the risk of early after depolarization (EAD)-induced arrhythmias but also lead to excessive sodium loading within cells that may contribute to chronic calcium loading that is considered to be a primary trigger for EADs and calcineurin-mediated activation of the NFAT gene transcriptional pathway that contributes to maladaptive hypertrophic remodelling observed in chronic AF (Lin et al., 2004). Previous research from our group reported that the C1 parent molecule resveratrol preferentially inhibited late over peak Nav1.5 currents (Wallace et al., 2006; 26 μmol·L−1 vs. 77 μmol·L−1 respectively). Our results indicate that C1 inhibits the peak and late components of recombinant human heart Nav1.5 currents more potently than resveratrol with IC50s of 3.2 and 1.1 μmol·L−1 respectively. While we did not directly compare atrial and ventricular sodium currents, these findings warrant further investigation with respect to atrial-specific sodium channel inhibition by C1. However, results from ventricular myocytes indicate that C1 does not affect EC coupling at concentrations in the low micromolar range. In contrast, C1 was effective at reducing calcium-transient abnormalities caused by late sodium current induction with ATX-II. These findings suggest C1 will not interfere with ventricular EC coupling at concentrations effective at inhibiting atrial Kv1.5 and peak/late sodium channel currents.
Inhibition of the hERG (Kv11.2) potassium channel is a highly undesirable property of any new drug as significant hERG inhibition can lead to excessive action potential prolongation and life-threatening Torsades des Pointes arrhythmias (Sanguinetti and Tristani-Firouzi, 2006). Accordingly, we tested the effects of C1 on recombinant hERG channel whole-cell currents and observed that C1 is a weak inhibitor of peak and tail hERG currents with IC50 values of 30 and 25 μmol·L−1 respectively. These results indicate that C1 is unlikely to exhibit any QT prolongation via hERG channel inhibition at concentrations shown to be effective at inhibiting Kv1.5, Nav1.5 and IKACh channels.
With respect to the overall inhibitory profile on ion channels tested in this study, C1 displays marked selectivity in the following order of IC50 values Kv1.5 (late) < Kv1.5 (peak) < Nav1.5 (late) < IKACh < Nav1.5 (peak) << hERG (tail) < hERG (peak; Figure 9). These results are summarized in Figure 9 and indicate that C1 displays a favourable ion channel inhibitory profile that merits further characterization as a potential treatment for AF.
Non-ionic effects of C1
The generation of ROS in the fibrillating atria is likely to contribute to the disturbances in the ionic and non-ionic milieu observed in AF (Carnes et al., 2001; Mihm et al., 2001). Furthermore, Kv1.5 activity is increased in the presence of the ROS H2O2 (Caouette et al., 2003). Resveratrol exhibits a widely described antioxidant capacity that contributes to its well-documented biological effects (Brisdelli et al., 2009; Dolinsky and Dyck, 2011; Wu and Hsieh, 2011; Chung et al., 2012). Our results show that C1 possesses an antioxidant capacity similar to that of resveratrol. While the concentrations of C1 required to yield significant quenching of H2O2 are higher than those exerting its Kv1.5 and Nav1.5 inhibitory effects, C1 levels may accumulate in the plasma-membrane as resveratrol is known to partition into lipid membranes leading to significant tissue accumulation (Sale et al., 2004). Furthermore, binding of C1 to Kv1.5 channels may additionally contribute to inhibition of Kv1.5 currents in the presence of ROS by decreasing the ROS-induced activation (Caouette et al., 2003).
Activation of the NFAT transcription pathway is an important signalling pathway that contributes to the initiation of maladaptive hypertrophy and chronic AF (Cha et al., 2004; Wilkins et al., 2004) and activation and nuclear translocation of NFAT is increased in AF (Lin et al., 2004). Our previous research indicated that resveratrol inhibited NFAT nuclear translocation and reduced phenylephrine-induced hypertrophy in cardiac myocytes (Chan et al., 2008). Interestingly, our present results also indicated that C1 inhibited basal and angiotensin-II-induced NFAT activation at similar concentrations (IC50 ∼ 1 μmol·L−1) to resveratrol. While this effect may not play a major role in the reduction of AF in the acute setting, it may become more important in the prevention of transient AF conversion into chronic AF as the atria becomes remodelled.
In vivo efficacy
To test the in vivo efficacy of C1, we used a model of AF in chronically instrumented, conscious dogs, where electrophysiological changes favouring the development of AF were elicited by chronic rapid right atrial pacing (Figure 8A). Atrial tachypacing in dogs is an established AF model leading to electrical and structural remodelling (Morillo et al., 1995; Gaspo et al., 1997). Atrial tachycardia and flutter have also been shown to lead to similar atrial remodelling in humans (Franz et al., 1997). Therefore, the significantly decreased duration of AF (Figure 8B–D) and possibly the increased AERP following C1 administration provides in vivo evidence that C1 may be beneficial in the treatment of certain forms of AF in patients.
Interestingly, treatment with C1 (0.3 and 1 mg·kg−1, i.v.) did not reduce the incidence of burst-induced AF in this study; however, it significantly reduced the duration of AF episodes. Selective and non-selective IKur blocking compounds have been shown to reduce both the incidence and the duration of AF (Ford and Milnes, 2008). However, one must consider the electrical and structural remodelling of the atria following chronic AF in patients and thtat following chronic rapid atrial pacing in animal models (Nattel et al., 2007). Remodelling can alter the efficacy and effective doses of compounds developed for the treatment of AF. Although no change in IKur density was previously observed in a tachypaced dog AF model (Yue et al., 1997), the down-regulation of IKur has been demonstrated in humans with chronic AF (Van Wagoner et al., 1997; Bosch et al., 1999; Caballero et al., 2010). Such down-regulation of Kv1.5 currents in this model, if present, might have influenced the in vivo effect of C1, making the IKur current-blocking effect less important in this particular model. Other compounds blocking IKur that we recently investigated in our laboratory (unpublished data) in the same model (with similar IC50 values for IKur) significantly decreased the incidence of burst-induced AF in i.v. doses of 3 mg·kg−1 or higher. A recent review summarizing data on the cardiac electrophysiological effects of novel IKur blockers also found that they were effective against AF induction in i.v. doses of 1–3 mg·kg−1 and higher (Ford and Milnes, 2008). In the present study, due to limits in the synthesized amount of C1 (required for several parallel in vitro studies), it was not feasible to investigate C1 in vivo at a dose of 3 mg·kg−1.
Importantly, C1 did not prolong the QT interval in conscious dogs in the present study, suggesting that C1 is unlikely to provoke ventricular arrhythmias based on repolarization disturbances. These latter findings support our results on weak C1 hERG channel inhibition and a lack of effect of C1 on EC coupling in single ventricular myocytes. As this animal model represents AF in the setting of previously remodelled atria leading to an increased induction of AF, the pronounced acute effects observed with C1 are likely to be mediated by its ion channel inhibitory and antioxidant properties rather than inhibition of NFAT activation. Whether chronic treatment with C1 reduces maladaptive remodelling remains to be determined by long-term studies.
Limitations of the study
While the effects of C1 on human atrial cell IKur were measured, the effects of C1 on human atrial cell action potentials were not studied due to the limited supply of atrial appendage tissue and the limited number of viable cells following enzymic digestion. For practical reasons, all ionic currents were measured at room temperature (20–22°C) rather than at 37°C. Moreover, we acknowledge that the cardiac rhythm status of the donor could influence the electrophysiological response to C1 in human atrial cells; however, no information was available on the cardiac rhythm status of the patients whose atrial tissue samples were used in this study. In the conscious dog model of AF, the AERP could not be measured below 80 ms as this is the shortest S1–S2 duration that can be set in the human pacemakers used in this study. The conscious animal experimental set-up did not allow us to directly measure AERP as that would have required anaesthesia and thoracotomy. Therefore, the exact statistical evaluation of the possible AERP-prolonging effect of C1 was not performed in this study.
In conclusion, the aetiology of AF is complex, and a number of ionic and non-ionic pathways contribute to the initiation and maintenance of AF (Franz et al., 1997; Lin et al., 2004). Therefore, inhibition of more than one of these pathways is likely to provide greater anti-fibrillatory efficacy than one pathway alone (Ehrlich and Nattel, 2009; Dobrev and Nattel, 2010; Dobrev et al., 2012). Our results indicate that the resveratrol derivative C1 is an effective inhibitor of several potential targets involved in AF development and confirmed that C1 is also effective at reducing the duration of AF episodes in a model of inducible AF in dogs.
We would like to acknowledge the expert technical assistance of Beth Hunter and Lynn Jones for the isolation of human and rat cardiac myocytes. We would also like to thank Irina Manisali and the Centre for Drug Research and Development (Vancouver, Canada) for their help in obtaining the C1 HPLC data from dog plasma.
Funding was provided to P. E. L and J. R. B. D. by the Canadian Institutes of Health Research and Technology Entrepreneurs and Companies (TEC) Edmonton. P. E. L. is holder of the Charles A. Allard Chair in Diabetes Research. J. R. B. D is a Senior Scholar of Alberta Innovates Health Solutions. This work was also funded by grants from the Hungarian National Office for Research and Technology, Ányos Jedlik Programs (NKFP_07_01-RYT07_AF), the Hungarian National Development Agency co-financed by the European Regional Fund (TÁMOP-4.2.2/B-10/1-2010-0012), the Hungarian Academy of Sciences and the Hungarian Scientific Research Fund (OTKA NK-104331).
Conflict of interest
The authors declare no conflicts of interest. It should be noted that C1 and derivatives thereof are the subject of a US Provisional Patent application filed in May 2013.