The omega-3 (n-3) polyunsaturated fatty acids (omega-3 PUFAs) eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) from fish oil were recently reported to have an anti-atrial fibrillation effect in humans; however, the ionic mechanisms of this effect are not fully understood. The present study was designed to determine the effects of EPA and DHA on transient outward and ultra-rapid delayed rectifier potassium currents (Ito and IKur) and the voltage-gated sodium current (INa) in human atrial myocytes.
A whole-cell patch voltage clamp technique was employed to record Ito and IKur, and INa in human atrial myocytes. It was found that EPA and DHA inhibited Ito in a concentration-dependent manner (IC50: 6.2 µM for EPA; 4.1 µM for DHA) and positively shifted voltage-dependent activation of the current. In addition, IKur was suppressed by 1–50 µM EPA (IC50: 17.5 µM) and DHA (IC50: 4.3 µM). Moreover, EPA and DHA reduced INa in human atrial myocytes in a concentration-dependent manner (IC50: 10.8 µM for EPA; 41.2 µM for DHA) and negatively shifted the potential of INa availability. The INa block by EPA or DHA was use-independent.
The present study demonstrates for the first time that EPA and DHA inhibit human atrial Ito, IKur, and INa in a concentration-dependent manner; these effects may contribute, at least in part, to the anti-atrial fibrillation of omega-3 PUFAs in humans.
Atrial fibrillation (AF) is a common form of cardiac dysrhythmia, and the occurrence of AF increases with age: the prevalence rises from 0.5% of people in their 50s, to 5% of people over the age of 65 years, and to nearly 10% of the population over 80. AF is a major cause of morbidity and mortality as it increases the risk of death, congestive heart failure, and stroke in an ageing population.1,2 It is believed that AF is a lifetime risk in an ageing population, and therefore it is emerging as a major public-health concern.3,4 Anti-arrhythmic drug therapy remains the principal approach for suppressing AF and its recurrence.5
Recent experimental and clinical studies have shown that omega-3 polyunsaturated fatty acids (omega-3 PUFAs) from fish oil may be effective in preventing cardiac arrhythmias and sudden death.6–10 The omega-3 PUFAs were found to have a significant anti-arrhythmic action in rat atrial myocytes.11 Human consumption of fish has been associated with a lower incidence of AF in a follow-up study.6,12 Intake of omega-3 PUFAs is found to reduce the incidence of post-operative AF of coronary bypass surgery with a shorter hospital stay.10 A recent study demonstrated that omega-3 PUFAs suppressed congestive heart failure-induced electrical remodelling and AF induction in a dog model.13 Although omega-3 PUFAs have been found to block cardiac voltage-gated sodium current (INa), and L-type Ca2+ current (ICa.L) in cells from animal species,14 the ionic mechanism underlying the anti-AF action is not fully understood in humans, and a study with human atrium is therefore suggested.15 The transient outward and ultra-rapid delayed rectifier K+ currents (Ito and IKur) are found to play an important role in the repolarization of human atrium;16,17 thus both Ito and IKur are targets for the anti-AF study.5,18 The aim of this study was to investigate the effects of the omega-3 PUFAs eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) on Ito, IKur, and INa in adult human atrial myocytes using a whole-cell patch technique.
Atrial cells were isolated from specimens of human right atrial appendage obtained from patients undergoing coronary artery bypass grafting. The procedure for obtaining the tissues was approved by the Ethics Committee of the University of Hong Kong, based on the patients' consent. The investigation conforms with the principles outlined in the Declaration of Helsinki (see Cardiovasc Res 1997;35:2–4) for use of human tissue. All patients were free from supraventricular tachyarrythmias, and the atria were grossly normal at the time of surgery. Atrial myocytes were enzymatically dissociated as described previously19,20 and/or in Supplementary material online.
Data acquisition and analysis
The whole-cell patch-clamp technique was used for electrophysiological recording at room temperature (22–23°C) as described previously19,20 and in Supplementary material online. Nonlinear curve fitting was performed using Pulsefit (HEKA) and Sigmaplot (SPSS, Chicago, IL, USA). Paired and/or unpaired Student's t-test was used as appropriate to evaluate the statistical significance of differences between two group means, and analysis of variance was used for multiple groups. Values of P < 0.05 were considered to indicate statistical significance. Group data are expressed as mean ± SEM.
Effects of polyunsaturated fatty acids on Ito
Figure 1A shows the time course of peak Ito recorded in a human atrial myocyte with a 300 ms voltage step to +50 from −50 mV in the absence and presence of EPA. EPA at 10 µM gradually decreased Ito, and the sustained current (i.e. IKur). The inhibitory effect of the currents only slightly recovered on washout for 10 min. Bovine serum albumin (BSA) was reported to completely reverse the EPA-induced inhibitory effect on Ca2+ current in cardiac myocytes21 or INa in HEK 293 cell line.22 Therefore, a bath solution containing 0.1% BSA was applied to wash EPA. Figure 1B shows that the inhibition of Ito by EPA was rapidly reversed by 0.1% BSA.
Although the Ito measured was peak to the ‘quasi’-steady-state level, we suspected that the evaluation of the effect of EPA on Ito might not be accurate. We therefore used the selective Kv1.5 blocker diphenyl phosphine oxide-1 (DPO-1, Sigma-Aldrich)23 to separate Ito. DPO-1 (2 µM) induced a slight reduction of peak current amplitude and almost a full inhibition of IKur (Figure 1C). Ito was substantially suppressed by the combination of DPO-1 with 10 µM EPA, and the effect recovered on washout with 0.1% BSA.
Figure 1D illustrates the current–voltage (I–V) relationships of Ito density during control and in the presence of 1, 5, and 10 µM EPA. EPA inhibited Ito in a concentration-dependent manner. EPA at 1–10 µM suppressed Ito at test potentials of +10 to +60 mV (n = 7, P < 0.05 or P < 0.01 vs. control), and no significant voltage dependence was observed. The concentration–response relationship for the inhibition of Ito by EPA from 1 to 50 µM was evaluated at +40 mV (Figure 1D). The IC50 of EPA for inhibiting Ito was 6.2 µM, with a Hill coefficient of 0.9.
Figure 2A shows the time course of Ito recorded in the absence or presence of 10 µM DHA in a typical experiment without DPO-1 treatment. DHA slowly decreased Ito, and the effect reached a steady-state level with over 12 min superfusion. The effect was not reversible by 10 min washout; however, 0.1% BSA bath solution completely reversed DHA's inhibition of Ito in another representative cell (Figure 2B). DHA, as EPA, inhibited both Ito and IKur. Original traces are shown in the right of the panels. The inhibition of voltage-dependent Ito by 5 and 10 µM DHA in the cell pre-treated with 2 µM DPO-1 to suppress IKur was also fully reversed by BSA solution washout (Figure 2C). Figure 2D illustrates the I–V relationships of Ito during control and in the presence of 1, 5, and 10 µM DHA. DHA at 1–10 µM suppressed Ito at +10 to +60 mV (n = 6, P < 0.05 or P < 0.01). The IC50 of DHA for inhibiting Ito was 4.1 µM (Figure 2E), with a Hill coefficient of 0.9.
By fitting the inactivation process of Ito, we found that EPA and DHA (10 µM) had no significant effect on time-dependent inactivation (data not shown). The voltage dependence of activation conductance variable (g) of Ito was determined from I–V relationships for each cell (Figure 1C) as described previously.24 The normalized activation (g/gmax) value was fitted to the Boltzmann distribution to obtain half activation (V0.5) and slope factor of the current (Figure 3A). The V0.5 of Ito activation positively shifted 6.1 mV with 10 µM EPA (from 18.8 ± 0.8 mV in control to 24.9 ± 0.9 mV in EPA, n = 7, P < 0.01). With 10 µM DHA, the V0.5 positively shifted 6.6 mV from 18.1 ± 0.6 mV of control to 24.7 ± 1.1 mV (n = 6, P < 0.01). The slope factor was not significantly changed by EPA or DHA.
The variable (I/Imax) for the voltage-dependent inactivation (availability) of Ito was determined with the protocol as shown in the inset of Figure 3A and fitted to the Boltzmann distribution under control conditions and in the presence of 10 µM EPA. The V0.5 of Ito availability was not significantly changed by the application of EPA (−23.1 ± 0.8 mV for control, −26.5 ± 0.8 mV for EPA, n = 6, P = NS). Similarly, DHA at 10 µM had no effect on the V0.5 of Ito availability (−23.5 ± 0.6 mV in control, −26.7 ± 1.1 mV in DHA, n = 6, P = NS). The slope factor was not significantly altered by EPA or DHA.
Time-dependent recovery of Ito from inactivation was studied with the paired-pulse protocol as shown in the inset of Figure 3B. The recovery curves were fitted to a mono-exponential function in the absence or presence of 10 µM EPA. The recovery time constant of Ito was slightly slowed by EPA (72.3 ± 15.1 ms in control, 78.1 ± 17.3 ms in EPA, n = 7, P = NS). Similarly, DHA at 10 µM had no significant effect on the recovery time constant (73.6 ± 12.5 ms for control, 77.2 ± 14.4 ms after DHA, n = 6, P = NS) of Ito from inactivation.
Effects of eicosapentaenoic acid and docosahexaenoic acid on IKur
Figure 4A displays the time course of IKur in a representative experiment with 10 µM EPA superfusion. EPA remarkably reduced IKur, and the inhibition recovered on washout with 0.1% BSA. Voltage-dependent IKur traces recorded in a typical experiment in the absence or presence of EPA are illustrated in Figure 4B. EPA at 10 µM substantially inhibited both IKur and tail current. EPA inhibited IKur in a concentration-dependent manner (Figure 4C). The IC50 of EPA for inhibiting IKur was 17.5 µM, with a Hill coefficient of 1.4.
Figure 4D shows the time course of IKur in a typical experiment with 10 µM DHA superfusion. DHA, as EPA, remarkably reduced IKur, and the inhibition recovered on washout with 0.1% BSA solution. Figure 4E illustrates the effect of DHA on voltage-dependent IKur. DHA substantially decreased IKur in a concentration-dependent manner (Figure 4F). The inhibitory effect of DHA on IKur was stronger than that of EPA. The IC50 of DHA for suppressing IKur was 4.3 µM, with a Hill coefficient of 1.1.
Influence of eicosapentaenoic acid and docosahexaenoic acid on human atrial INa
Earlier studies reported that EPA and DHA inhibited cardiac INa in animal species and in HEK 293 cells expressing cloned Na+ channels.22,25,26 Here, we evaluated the effect of EPA and DHA on INa in freshly dissociated human atrial myocytes. Figure 5A shows the voltage-dependent INa traces recorded in a human atrial myocyte in the absence or presence of EPA. EPA at 10 µM remarkably reduced the amplitude of INa, and the inhibition was fully reversed by washout with 0.1% BSA bath solution. The I–V relationships of INa density shown in Figure 5B indicate that EPA significantly inhibits INa at −50 to −5 mV in a concentration-dependent manner. The IC50 of EPA for inhibiting INa (at −35 mV, Figure 5C) was 10.8 µM, with a Hill coefficient of 1.2.
DHA also exhibited an inhibitory effect on INa in human atrial myocytes; nevertheless, the inhibitory action was weaker than that of EPA (Figure 5D). The IC50 of DHA for inhibiting INa was 41.2 µM, with a Hill coefficient of 1.4.
Inactivation of INa traces was fitted to mono-exponential functions before and after 10 µM EPA or 20 µM DHA. The voltage-dependent inactivation time constants are illustrated in Figure 5E and F. EPA or DHA reduced the inactivation time constant; a significant effect was seen only at −50 to −40 or −30 mV (n = 7, P < 0.05 vs. control).
Use-dependent block of INa was examined with EPA or DHA using a train of 15 identical pulses (30-ms) from −120 to −35 mV. EPA (10 µM) or DHA (20 µM) inhibited INa at frequencies of 0.5, 1, 2, 5 Hz; however, the inhibition was use- or rate-independent (Figure 5G and H).
The activation conductance variable (g/gmax) of INa was determined from I–V relationships for each cell (Figure 5B) and was fitted to the Boltzmann distribution (Figure 6B). The voltage dependence of availability (I/Imax) of INa was determined as illustrated in Figure 6A, and also fitted to the Boltzmann distribution. Figure 6B shows that 10 µM EPA shifted the midpoint of INa availability towards negative potentials. The V0.5 of availability negatively shifted 9.7 mV, from −97.1 ± 1.5 mV in control to −106.9 ± 1.7 mV in EPA (n = 7, P < 0.01). There was no shift for the V0.5 of INa activation conductance with 10 µM EPA (−38.8 ± 1.2 mV in control, −39.4 ± 1.5 mV in EPA, n = 6, P = NS). DHA at 20 µM also negatively shifted V0.5 of INa availability by 6.5 mV (from −96.8 ± 1.3 mV in control to −103.3 ± 1.4 mV in DHA, n = 6, P < 0.01) and had no effect on the V0.5 of INa activation conductance.
Recovery of INa from inactivation was studied with a paired-pulse protocol. Superimposed currents and the time course of recovery are illustrated in Figure 6C and D. INa recovery was complete and well fitted to mono-exponential functions with time constants of 10.6 ± 2.1 ms in control and 18.1 ± 2.5 ms with 10 µM EPA (n = 7, P < 0.01). In another set of experiments, the recovery time constant of INa was 11.5 ± 2.4 ms in control and 16.9 ± 1.9 ms with 20 µM DHA (n = 6, P < 0.01).
The development of resting inactivation of INa was determined with a variable duration of conditioning pre-pulse from −120 mV to a subthreshold potential (−75 mV), followed by a 3 ms step back to −120 mV prior to a test pulse to −35 mV.27Figure 6E shows the voltage protocol (inset) and superimposed INa traces recorded during the test pulse in a typical experiment, showing that INa reduces as the conditioning pre-pulse duration increases. The normalized INa was plotted against the pre-pulse duration and fitted to mono-exponential functions (Figure 6F). The inactivation time constant reduced from 24.8 ± 3.2 ms in control to 12.3 ± 2.9 ms with 10 µM EPA (n = 7, P < 0.01). In the experiments with 20 µM DHA, the inactivation time constant decreased from 25.3 ± 2.5 ms in control to 16.3 ± 3.5 ms in DHA (n = 6, P < 0.01). These results demonstrate that EPA and DHA significantly increase the inactivation of INa from resting states.
It is well known that fish oil is rich in the omega-3 PUFAs EPA and DHA.25 A study of the prevention of ventricular fibrillation by dietary fish oil following coronary artery occlusion and reperfusion was reported three decades ago.28 Recent epidemiological studies indicate that fish oil consumption may reduce the risk of ventricular arrhythmias in humans, supporting an anti-arrhythmic effect of omega-3 PUFAs,15,29 although the contradictory result was also reported.30 In addition, Calo et al.10 recently demonstrated that the administration of omega-3 PUFAs (EPA and DHA) induced a reduction of the relative risk of post-operative AF in 54.4% patients who had undergone coronary artery bypass. However, the intake of fish oil supplement has been reported to have no effect or to be pro-arrhythmic in different populations.31,32 Therefore, further clinical observation and cellular investigation are suggested to define the possible anti-AF action of omega-3 PUFAs.15
Earlier cellular studies showed that EPA and DHA inhibited INa in neonatal rat ventricular myocytes26 and in HEK293 cells transfected with the human cardiac Na+ channel hH1,22 L-type Ca2+ current (ICa.L) in rat ventricular cells,33 and the potassium currents IK and Ito in ferret ventricular myocytes.14 It is believed that omega-3 PUFAs modify Na+ channels by directly binding to the channel proteins,34 although an decreased INa was not observed in a recent study in pigs with a fish oil diet for 8 weeks.35 The information obtained from animal species on the effects of omega-3 PUFAs on cardiac ion channels reveals important alterations in the ionic currents that may account for significant cardiac protection against fibrillation. However, we are not aware of PUFAs-mediated alteration of ion channel properties in human cardiac myocytes.
In human heart, the transient outward K+ current Ito, the ultra-rapid delayed rectifier K+ current IKur, and the rapid and slow components IKr and IKs of delayed rectifier K+ current are important repolarization currents.16,17,36IKur is found to be functionally expressed in human atrium, but not in the ventricle.16 Therefore, the drugs that specifically inhibit the unique IKur may provide a means of preventing AF without the risk of ventricular pro-arrhythmia.5 The inhibition of Ito and/or IKur was reported to prolong the action potential duration in human atrium.37,38
The present study showed that EPA and DHA significantly inhibited Ito and IKur in human atrial myocytes (Figures 1–4). The relatively irreversible effects of EPA and DHA on Ito and IKur are likely related to their directly binding to the channel proteins.34 The suppression of Ito and IKur is useful for anti-AF.5 Therefore, the inhibitory effects of EPA and DHA on IKur and Ito would be most likely one of ionic mechanisms of anti-AF observed in humans.6,10,39 However, a reduced Ito was consistently found in atrial myocytes from patients with AF,40,41 while increased, unchanged, or decreased IKur was reported in different observations.40–42 This may account for the observation of unsuccessful AF prevention.43
As previously reported in neonatal rat ventricular myocytes and HEK 293 cells expressing human cardiac Na+ sodium channels (hH1),22,26 EPA and DHA inhibited INa in human atrial myocytes in a use-independent way (Figure 5), which is different from that of the anti-arrhythmic drug lidocaine.44 EPA and DHA negatively shifted the potential of INa availability and increased the development of INa inactivation at resting states (Figure 6). These effects would be cardioprotection against cardiac fibrillation.14,25 EPA had a stronger suppression of INa (Figure 5), while DHA showed a stronger inhibition of Ito and IKur in human atrial myocytes (Figures 1–4). Therefore, fish oil containing both EPA and DHA would be more effective in anti-AF in humans. However, on the other hand, the INa block may be related to the increase of ventricular arrhythmias observed in some people taking fish oil supplement.30,45
The inhibitory effect of EPA (IC50: 6.2 µM) or DHA (IC50: 4.1 µM) on Ito was slightly stronger in human atrial myocytes than that (IC50: ∼7.5 µM) in ferret ventricular myocytes46 However, the inhibition of INa by EPA (IC50: 10.8 µM), and especially DHA (IC50: 41.2 µM), was weaker in human atrial myocytes than that (IC50: <5 µM) in neonatal rat ventricular cells,26 whereas the inhibitory effect of EPA (IC50: 17.5 µM), and especially DHA (IC50: 4.3 µM), on IKur was stronger in human atrial myocytes than that (IC50: 20–30 µM for DHA) observed in Kv1.5 cell line.47 These differences are likely related to various species and/or types of cells. Nevertheless, the ‘acute’ omega-3 PUFAs effects on cardiac ion channel function may not be completely applicable to the chronic effects of fish oil supplements in the ‘real world’ setting, which is a limitation of an ‘acute’ in vitro study vs. chronic in vivo effect. The information for the concentration levels in cardiac tissue and/or blood plasma remains to be established. It is important to note the superfusion of BSA rapidly reversed the omega-3 PUFAs effect (Figures 1, 2, 4, and 5), indicating that omega-3 PUFAs have a high affinity to blood plasma albumin. Therefore, the effects of PUFAs on ion channels observed in the present study would most likely only be reproducible when the cardiac tissue concentration levels and/or free blood plasma concentration levels reached >4–6 µM.
Another limitation of the present study was that the specific IKur blocker DPO-1 used to separate Ito is a partial open channel blocker and shows less block of Kv1.5 at initial opening of the channel.23 The incomplete inhibition of IKur at initial opening of the channel would underestimate the Ito inhibition by omege-3 PUFAs.
Although the study with human atrial specimens from the patients with dietary fish oil is indicated,15 an ethical issue arises for such experiments. Here, we only focused the study on the ‘acute’ effects of omega-3 PUFAs on Ito, IKur, and INa in human atrial cells with the limited human atrial specimens. The effects of omega-3 PUFAs on other ion channel currents (e.g. ICa.L, IKr, and IKs) and action potentials were not verified in human atrial myocytes due to the shortage of human atrial specimens, which occurred because the cardiac coronary artery bypass surgical procedure has been recently improved to avoid cutting atrial tissue. It was reported that DHA inhibited IK in ferret ventricular myocytes46 and hERG channel expressed in Chinese hamster ovary cells,48 which may also contribute to anti-arrhythmias.
In summary, the present study demonstrates the novel information that the omega-3 PUFAs EPA and DHA from fish oil inhibited Ito, IKur, and INa in human atrial myocytes. These effects likely contribute at least in part to the anti-AF action observed in humans.
The work was supported by Sun Chieh Yeh Heart Foundation of Hong Kong.
The authors thank Heather J. Ballard for her critical reading of the manuscript.
Conflict of interest: none declared.