Air pollution is an environmental hazard that is associated with cardiovascular dysfunction. Phenanthrene is a three-ringed polyaromatic hydrocarbon that is a significant component of air pollution and crude oil and has been shown to cause cardiac dysfunction in marine fishes. We investigated the cardiotoxic effects of phenanthrene in zebrafish (Danio rerio), an animal model relevant to human cardiac electrophysiology, using whole-cell patch-clamp of ventricular cardiomyocytes. First, we show that phenanthrene significantly shortened action potential duration without altering resting membrane potential or upstroke velocity (dV/dt). L-type Ca2+ current was significantly decreased by phenanthrene, consistent with the decrease in action potential duration. Phenanthrene blocked the hERG orthologue (zfERG) native current, IKr, and accelerated IKr deactivation kinetics in a dose-dependent manner. Furthermore, we show that phenanthrene significantly inhibits the protective IKr current envelope, elicited by a paired ventricular AP-like command waveform protocol. Phenanthrene had no effect on other IK. These findings demonstrate that exposure to phenanthrene shortens action potential duration, which may reduce refractoriness and increase susceptibility to certain arrhythmia triggers, such as premature ventricular contractions. These data also reveal a previously unrecognized mechanism of polyaromatic hydrocarbon cardiotoxicity on zfERG by accelerating deactivation and decreasing IKr protective current.
The cardiac action potential (AP) is characterized by its long duration in contrast to nerve or skeletal muscle cells. Different voltage-gated ion channels shape the cardiac AP. Na+ channels are responsible for the rapid depolarization, the plateau phase is mediated by Ca2+ channels, and repolarization is due to multiple K+ channels. The shape depends on species and region of the heart, but the long duration is a hallmark of cardiac cells and safeguards against abnormal electrical activity, ensuring a long refractory period. Current elicited by L-type Ca2+ channels (ICaL) play critical roles in initiating the process of calcium-induced calcium release that drives cardiac excitation–contraction coupling and in providing depolarizing current that contributes to the plateau phase of cardiac APs (Bers and Despa, 2013). The rapid delayed rectifier K+ current (IKr) is a significant contributor to cardiomyocyte AP repolarization and is underpinned by the ether-à-go-go–related gene (ERG) encoded channel (Sanguinetti et al., 1995; Trudeau et al., 1995). Due to the unique kinetics of ERG channels, they generate rapid, transient outward currents in response to premature depolarization during AP repolarization, early in diastole, therefore providing some protection from premature ventricular excitation (Lu et al., 2001; Perry et al., 2015). Pharmacological inhibition of this channel leads to prolongation of the AP, and therefore, the QT interval of the electrocardiogram. The human ERG (hERG) channel is known to be susceptible to pharmacological inhibition by diverse drugs associated with the drug-induced form of acquired long QT syndrome and associated torsades de pointes arrhythmia (Sanguinetti and Tristani-Firouzi, 2006; Vandenberg et al., 2012).
Previous studies show that IKr and ICaL are inhibited by the low-molecular-weight tricyclic polyaromatic hydrocarbon (PAH) phenanthrene (see inset in Fig. 1 A for a structural diagram of the molecule) in cardiomyocytes from marine fish species (Brette et al., 2014; Brette et al., 2017) and freshwater trout (Ainerua et al., 2020; Vehniäinen et al., 2019). In vivo PAH exposure has also been shown to reduced cardiac output in large pelagic fish (Nelson et al., 2016) and cause heart block in zebrafish embryos (Incardona et al., 2004). PAHs are present in crude oil and are derived from the burning of fossil fuels. PAHs such as phenanthrene are a ubiquitous component of both air and water pollution, but their cardiotoxicity only came to light following the catastrophic 1989 Exxon Valdez oil spill and the 2010 Deepwater Horizon blowout, which had significant toxicological effects on aquatic ecosystems (Incardona et al., 2009; Brette et al., 2014; Brette et al., 2017). In addition, there is now clear evidence of a strong correlation between air pollution and human cardiovascular diseases such as cardiac arrhythmias and stroke (Shah et al., 2013; Franklin et al., 2015; Marris et al., 2020). However, the exact mechanism(s) of detrimental effects of PAH pollutants are yet to be understood, especially in a model relevant to human cardiac electrophysiology.
The International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use S7B cardiac safety testing guidelines specify that model species used to investigate drug effects on cardiac electrophysiology should possess ventricular APs with a high plateau phase and human-relevant repolarization mechanisms (European Medicines Agency, 2005). Mice, for example, are poorly suited to this purpose as they possess abbreviated ventricular APs and do not rely on IKr or slow delayed rectifier potassium current (IKs) for repolarization (Wang et al., 1996; Nerbonne et al., 2001). By contrast, zebrafish (Danio rerio) exhibit a ventricular AP morphology that is similar to that in humans, with human-relevant repolarization mechanisms including IKr (Arnaout et al., 2007; Brette et al., 2008; Nemtsas et al., 2010; Abramochkin et al., 2018; Marris et al., 2020) and IKs (Abramochkin et al., 2018). The zebrafish ERG channel (zfERG) has a high level of sequence homology with hERG (Langheinrich et al., 2003; Marris et al., 2020), and zebrafish embryos have been shown to be sensitive to proarrhythmic drugs (Langheinrich et al., 2003). The zebrafish is therefore a translationally relevant model in which to study PAH cardiotoxicity.
The purpose of the present study was to characterize the effects of phenanthrene on the zebrafish ventricular AP and to study specific effects on calcium and potassium channel properties in order to determine the mechanism responsible for cardiac electrophysiological alteration in a model relevant to human cardiac physiology.
Material and methods
All experiments were conducted in accordance with The UK Animals (Scientific Procedures) Act, 1986, and European Union Directive EU/2010/63. Local ethical approval was obtained from the University of Manchester Animal Welfare and Ethical Review Board.
Following stunning, the brain of the zebrafish was destroyed via pithing using a 16-gauge needle. The isolated ventricle was cannulated with a blunt 35-gauge needle through the bulbus arteriosus, tied and perfused from a height of 5 cm with a Ca2+-free isolation solution (see Physiological solutions) for 5 min followed by enzymatic digestion (0.65 mg/ml collagenase type 1A, 0.4 mg/ml trypsin type III, and 0.7 mg/ml BSA in 10 ml isolation solution) for 4–5 min. The ventricle was removed, cut into small pieces, and gently triturated using a Pasteur pipette to further release of single cardiac myocytes. Isolated myocytes were held at room temperature in isolation solution and used within 6 h.
All solutions were prepared using ultrapure water supplied by a Milli-Q system (Millipore). Chemicals were reagent grade and purchased from Sigma-Aldrich, except for tetrodotoxin (TTX; Tocris). Purity of phenanthrene was at least 99%. Stocks of phenanthrene (25 mM) constituted in dimethyl sulfoxide (tissue culture grade; Sigma-Aldrich) were prepared fresh on the day of experiments. In all experiments, maximal DMSO concentration was <1:1,000 vol/vol.
Whole-cell patch clamp
Electrophysiological data were recorded as previously described (Ainerua et al., 2020). Briefly, isolated myocytes were allowed to settle for ∼10 min in the recording chamber at the beginning of each trial and then perfused with control physiological solution (see Physiological solutions). Data were recorded via a Digidata 1322A A/D converter (Axon Instruments) controlled by an Axopatch 200B (Axon Instruments) amplifier running pClamp 10.3 software (Axon Instruments). Signals were filtered at 2 kHz using an eight-pole Bessel low-pass filter before digitization at 10 kHz and storage. Patch pipette resistance was typically 2–3 MΩ when filled with intracellular solution. Series resistance ranged between 5 and 10 MΩ and was compensated up to 60% for ICaL voltage-clamp experiments (voltage error ≤ 2 mV).
APs were evoked using 1 ms supra-threshold current steps at a frequency of 0.5 Hz. All AP parameters were stable over the time of recording (<10 min) in controls.
ICaL was elicited using a double-pulse protocol (see Fig. 2 C, inset) that consisted of 2 s duration prepulses from −40 to +40 mV in 10-mV steps, a 20-ms inter-pulse interval at −40 mV, and a test pulse to 0 mV for 2 s. A prepulse to −40 mV from holding potential of −70 mV (500 ms) was used to inactivate Na+ and T-type Ca2+ current before the double-pulse protocol. I-V and steady-state activation curves were obtained from currents elicited during prepulse voltage steps. Availability (steady-state inactivation) curves were obtained from currents elicited during test pulse. The stimulus frequency was 0.1 Hz.
IKr was activated by a test pulse to +20 mV (2.5 s) to fully activate K+ channels (as determined in preliminary experiments) and measured as the tail current at −40 mV (4.5 s) at a stimulation frequency of 0.1 Hz. A prepulse to −40 mV from a holding potential of −70 mV (500 ms) was used to inactivate any Na+ and T-type Ca2+ current. The instantaneous I-V relation was measured from currents recorded using an AP-like voltage-clamp protocol at a frequency of 0.5 Hz with an initial depolarizing step from −70 mV to +30 mV for 500 ms followed by repolarizing ramp (from +30 mV to −70 mV) of 150 ms. IKr rapid transient currents were generated using a paired AP-like voltage-clamp protocol as described above with depolarizing stimuli to 0 mV at increasing intervals (7.5 ms). IKr transient currents were obtained following each condition (control, phenanthrene, and E-4031) within the same cell. Peak transient currents obtained in presence of 2 µM E-4031 were subtracted from corresponding currents in both control and phenanthrene exposure. Peak IKr transients for each cell under control conditions were then normalized to the maximal current transient. Peak IKr transients obtained in presence of 3 µM phenanthrene are represented as percentage response of the corresponding control transient.
Other IK were activated by a depolarizing pulse to +40 mV (4 s) to fully activate K+ channels (as determined in preliminary experiments) at a frequency of 0.1 Hz, following IKr inhibition (see below). A prepulse to −40 mV from holding potential of −70 mV (500 ms) was used to inactivate any Na+ and T-type Ca2+ current (Abramochkin et al., 2018).
The isolation solution contained (in mM) 100 NaCl, 10 KCl, 1.2 KH2PO4, 4 MgSO4, 50 taurine, 20 glucose, and 10 HEPES, with pH set to 6.9 with KOH. The composition of the extracellular physiological solution (Ringer’s) used for electrophysiology contained (in mM) 150 NaCl, 3.5 KCl, 1.5 MgCl2, 2 CaCl2, 10 glucose, and 10 HEPES, with pH set to 7.7 with NaOH. To avoid overlapping ion currents during voltage-clamp recording of ICaL, IKr, and other IK, this solution was modified. For ICaL measurement, KCl was replaced by CsCl to inhibit K+ channels, and TTX (0.5 µM) was used to inhibit Na+. For barium current (IBaL) measurement, CaCl2 was replaced by equimolar BaCl2 in the above solution. For IKr recording, TTX (0.5 µM), nifedipine (10 µM), and glibenclamide (10 µM) were included in the Ringer’s solution to inhibit Na+, Ca2+, and ATP-sensitive K+ channels, respectively. The concentration of TTX used in this study to inhibit INa current, albeit low, is in accordance with previous studies (Haverinen et al., 2007) that showed greater sensitivity of TTX to Na+ channels in fish cardiomyocytes (in the nanomolar range) compared with mammalian cardiomyocytes (in the micromolar range). For recording net IK, 2 µM E-4031 was included in the external solution to inhibit the zfERG channel. The remaining K+ currents represent all outward IK except IKr. Solutions were locally superfused over the cell using a 6–1 manifold and VC-6 six-channel valve controller perfusion system (Warner Instruments) to switch between various solutions.
Pipette solutions were optimized for each electrophysiological recording condition. For AP measurement, the pipette solution contained (in mM) 10 NaCl, 140 KCl, 5 MgATP, 0.025 EGTA, 1 MgCl2, and 10 HEPES, with pH adjusted to 7.2 with KOH. For ICa measurement, the pipette solution contained (in mM) 130 CsCl, 15 TEA-Cl, 5 MgATP, 1 MgCl2, 5 Na2-phosphocreatine, 5 EGTA, 10 HEPES, and 0.03 Na2GTP, with pH adjusted to 7.2 with CsOH. CsCl and TEA-Cl were included to inhibit K+ currents. For IKr and other IK measurements, EGTA concentration was increased to 5 mM in the AP pipette solution.
Data are reported as mean ± SEM. Due to a nonnormal distribution, paired data were analyzed with a Wilcoxon signed-rank test (percentage data) or a Mann–Whitney test with P values as indicated within the text and figure legends using GraphPad Prism 8 software. A two-way ANOVA with Holm–Šídák post hoc analysis was used for ICaL I-V plot. Samples sizes for animals are given as N and for myocytes as n in the text and figure legends.
Online supplemental material
Phenanthrene shortens APD in zebrafish ventricular cardiomyocytes
We recorded APs using a current clamp in the whole-cell configuration. Fig. 1 A shows shortening of the APD in response to 3 µM phenanthrene. Phenanthrene application at both 3 and 10 µM had no significant effect on resting membrane potential (Fig. 1 E) and upstroke velocity (dV/dtmax; Fig. 1 F), triangulation (measured as difference in APD at 30% and 90% of repolarization), or beat-to-beat variability rate from control values (data not shown). In contrast, APD50 and APD90 (measured at 50% and 90% of repolarization, respectively) were significantly decreased by acute superfusion of 3 µM phenanthrene (Fig. 1 A). This is in contrast to previous studies in bluefin tuna and brown trout in which AP prolongation was observed (Brette et al., 2017; Ainerua et al., 2020). Subsequent application of 3 µM phenanthrene with E-4031, a selective IKr inhibitor (at a supramaximal concentration of 2 µM), led to prolongation of APD90 (Fig. 1 A, green line). AP shortening was greater at 10 than 3 µM phenanthrene, leading to significant reduction of both APD50 and APD90 (Fig. 1 D). Application of E-4031 in the presence of 10 µM phenanthrene led to prolongation of APD90 (Fig. 1 B), suggesting that IKr was not fully blocked by phenanthrene at concentrations between 1 and 10 µM. Mean data indicate a significant reduction in APD50 and APD90 in 3 µM (n = 4; N = 2) and 10 µM (n = 9; N = 5) phenanthrene, respectively (Fig. 1, C and D; P < 0.05; Wilcoxon test). Partial recovery in both APD50 and APD90 was observed following 10 min washout (Fig. 1 B, dotted line). Taken together, these data suggest that IK1 (inwardly rectifying K+ current), the current responsible for resting membrane potential, and INa, the current responsible for the upstroke of AP, are not modified by phenanthrene, and that the likely ionic currents affected by phenanthrene are ICaL and K+ currents, resulting in an overall decrease in APD.
Phenanthrene inhibits L-type Ca2+ current
Fig. 2 A shows that phenanthrene significantly decreased ICaL in zebrafish ventricular myocytes, as described previously in other fish species (Brette et al., 2017). Partial recovery from inhibition was observed following 5 min wash (Fig. 2 A). Mean data at −10 mV showed a 39 ± 4% inhibition (Fig. 2 B; n = 10; N = 3; P < 0.05; Wilcoxon test). Construction of a concentration–response relation yielded an IC50 of 4.5 ± 0.8 µM (Fig. 2 C; nH = 1.3 ± 0.2; n = 5–10; N = 3). To examine this effect across a range of voltages, we used a double-pulse protocol (Fig. 2 E, inset), which showed significant voltage-dependent inhibition of ICaL currents within the range of −20 to +10 mV in presence of 10 µM phenanthrene (Fig. 2 D; n = 6; N = 2; P < 0.05; two-way ANOVA with Holm–Šídák post hoc test). Peak ICaL was elicited at −10 mV, and this showed no shift in the presence of phenanthrene. This is further confirmed by the lack of significant differences in steady-state activation and inactivation parameters, including window current, between control and phenanthrene conditions (Fig. 2 E). To better understand the mechanism of phenanthrene inhibition of ICaL, we repeated our studies with Ba2+ as the charge carrier (Fig. S1). Phenanthrene had less of an inhibitory effect on peak Ba2+ currents compared with peak Ca2+ currents (only 22 ± 6%, compared with 39 ± 4%; P < 0.05; Mann–Whitney test; Fig. S1 B; n = 5; N = 3). There was a trend for a reduction in the total charge (Q) transferred by Ba2+ compared with Ca2+ (Fig. S1, B–D). These data indicate a Ca2+-dependent inhibition by phenanthrene. However, further experiments are required to determine the exact mechanism. Together these results indicate that APD shortening in response to phenanthrene exposure may be accounted for by reduced ICa density. However, the AP plateau phase involves a delicate balance between the activity of L-type Ca2+ channels and K+ channels. Previous studies of phenanthrene on ventricular cardiomyocytes from several fish species showed significant AP prolongation linked to inhibition of IKr (Brette et al., 2014; Brette et al., 2017; Ainerua et al., 2020). Thus, we next assessed the effect of phenanthrene upon IKr.
Phenanthrene inhibits IKr but not other IK currents in zebrafish cardiomyocytes
Zebrafish ventricular cardiomyocytes exhibited robust peak IKr tail currents (Fig. 3 A) at −40 mV following a voltage protocol incorporating activating commands to +20 mV (Fig. 3 A, inset), as previously described (Nemtsas et al., 2010), with a mean current density of 2.4 ± 0.3 pApF−1 (n = 9; N = 2). Phenanthrene at 3 µM, 10 µM, and 25 µM induced a significant 49 ± 3% (n = 9; N = 3), 84 ± 3% (n = 6; N = 3), and 94 ± 1% (n = 5; N = 2; P < 0.05; Wilcoxon test) decrease of peak IKr tail current, respectively. IKr tail amplitude was reduced in a concentration-dependent manner in response to phenanthrene exposure with an IC50 of 3.3 ± 0.2 µM and a Hill coefficient of 1.8 ± 0.2 (Fig. 3 B; n = 5–9; N = 3). A plot of the time-course curve showed a maximum of 69 ± 5% recovery from inhibition by 10 µM phenanthrene after 10 min of washout (Fig. 3 C; n = 6; N = 3). Phenanthrene also increased rate of decay of peak IKr tail currents in a concentration-dependent manner (Fig. 3 D). A significant reduction in time constant of fast-deactivating current component, τfast (Fig. 3 E; n = 8; N = 2) was associated with an increase in the contribution of fast-deactivating current component Af (Fig. 3 F; n = 8; N = 2).
To investigate whether other IK were modified by phenanthrene, we recorded remaining Ik during E-4031 superfusion before and after phenanthrene application. No significant effects at 3 and 25 µM were observed, strongly suggesting that IKr is the only IK modified by phenanthrene (n = 4–5; N = 2; Fig. S2, A and B).
To further investigate the effects of phenanthrene on IKr under more physiologically relevant conditions, we used an AP-like voltage clamp protocol (Fig. 4 A). IKr current showed a voltage dependence similar to that of the fully activated I-V relation (Fig. 4 B) due to rapid recovery from inactivation as AP repolarization proceeds. Therefore, zebrafish IKr during the AP-like protocol is very similar to that recorded from mammalian ventricular myocytes (Hancox et al., 1998; Zaza et al., 1998). The instantaneous I-V relation during the repolarizing ramp (from +30 to −70 mV) was bell-shaped with peak current at −46.4 ± 1.2 mV (n = 7; N = 2; Fig. 4 B). Interestingly, 3 µM phenanthrene had less (P < 0.05; Mann–Whitney test) of an inhibitory effect on the instantaneous peak current (24 ± 7% of peak current; Fig. 4 B) when compared with IKr peak-tail currents (49 ± 3% of peak current). In addition, phenanthrene induced a significant shift of −6 mV in mean peak I-V (from −46.4 ± 1.2 to −52.5 ± 1.5 mV; Fig. 4 B; P < 0.05; Wilcoxon test; n = 7; N = 2). This result raised the possibility that IKr may be less protective during the late repolarization phase following exposure to phenanthrene. To investigate this further, we tested the effect of phenanthrene upon rapid IKr transient outward currents elicited by a paired ventricular AP-like command waveform protocol.
Phenanthrene inhibits protective IKr transient currents
Fig. 5 A, inset, shows the double pulse protocol used to record IKr in response to premature depolarizing stimuli, introduced from APD90 −90 ms to APD90 +90 ms (Fig. 5 A). Under control conditions, IKr transient current peaked at 15 ms after APD90 (Fig. 5 E, b). This current is considered protective against cardiac arrhythmia (Perry et al., 2015), and 3 µM phenanthrene induced significant inhibition of this peak transient current over the entire range of depolarizing stimuli (Fig. 5 E). Notably, the magnitude of the peak IKr transient current inhibition (63 ± 6%; Fig. 5 E, b), is similar to that of IKr tail currents. Together, these results indicate that protective IKr transient currents were blunted by phenanthrene.
Pollutants disrupt the fine-tuned electrical activity of the heart and increase the risk of cardiac arrhythmias and sudden cardiac death (Shah et al., 2013; Franklin et al., 2015). In this study, we show that phenanthrene, a low-molecular-weight tricyclic PAH, decreases APD in zebrafish. We show that phenanthrene decreases L-type Ca2+ current and the zfERG-mediated IKr current while accelerating IKr deactivation. Lastly, we show that the protective role of IKr (in the form of rapid IKr transients) is diminished by this pervasive air and water pollutant.
Disruption of ionic current induces AP shortening by pollutant phenanthrene
Air pollution affects human cardiovascular function and can lead to death (Marris et al., 2020). Cardiac arrhythmias represent a significant proportion of these deaths (European Medicines Agency, 2017). Phenanthrene is present in the aquatic environment primarily due to crude oil spills and is also present in the atmosphere as a volatile substance and in association with particulate matter (e.g., diameter of 2.5 µm or less; Tsapakis and Stephanou, 2005; Incardona et al., 2014; Marris et al., 2020). The devastating oceanic oil spills focused studies on the impact of PAHs on fish cardiovascular function, whereas comparable studies on mammals associated with air pollution have lagged behind (Marris et al., 2020). In freshwater brown trout and various pelagic fishes, phenanthrene significantly prolonged the AP at both the cellular and whole heart levels (Brette et al., 2017; Ainerua et al., 2020). The prolongation of AP in these studies is attributed to inhibition of the IKr current. This differs from the present study in zebrafish showing a significant AP shortening (Fig. 1). A recent study found no effect of phenanthrene on the APD in rainbow trout ventricular cardiomyocytes at concentrations between 0.3 and 30 µM, but did find significant APD shortening after acute exposure to an alkylated phenanthrene (retene) at concentrations between 1 and 10 µM (Vehniäinen et al., 2019).
The relative APD shortening or lengthening in response to an agent depends largely on its inhibitory potency on inward (mainly Na+ and Ca2+) and outward (mainly K+) ionic conductances (Hancox et al., 2008). Thus, it is interesting to note that in both the rainbow trout study (Vehniäinen et al., 2019) and our present study, phenanthrene exerted significant inhibitory activity on IKr amplitude (Fig. 3, A–C), which is predicted to prolong APD (Brette et al., 2017; Ainerua et al., 2020). Our finding of APD shortening instead of prolongation despite the significant impact on IKr implicates significant inhibition of ICaL in response to phenanthrene exposure. Phenanthrene has been reported to inhibit L-type Ca2+ channel current and reduce the intracellular Ca2+ transient, thus affecting the contractility of ventricular myocytes in a number of marine fishes and in the freshwater brown trout (Brette et al., 2017; Ainerua et al., 2020). However, pharmacological potencies varied greatly in these studies, with ∼30% inhibition exhibited by 5 µM phenanthrene in bluefin tuna (Brette et al., 2017), whereas similar inhibition was observed only at 30 µM phenanthrene in brown trout (Ainerua et al., 2020). We show that phenanthrene exhibited a maximum inhibition of 60 ± 5% of the zebrafish L-type calcium current at 25 µM concentration (Fig. 2 C) with an IC50 value of 4.5 µM (Fig. 2 C), which may account for AP shortening following phenanthrene exposure. It is important to note that fish are more reliant on transsarcolemmal Ca2+ influx during excitation–contraction coupling than mammals (Shiels and Galli, 2014), which impacts the density of inward and outward currents during an AP across species. This is consistent with preliminary data reported in Marris et al. (2020) showing AP prolongation in sheep ventricular cardiomyocytes in the presence of phenanthrene. Thus, the AP shortening observed in our study could highlight important species-specific differences in the balance of Ca2+ influx to K+ efflux during the AP and the differential pharmacological potencies of phenanthrene toward Ca2+ and K+ channels. Such differences need to be kept in mind when extrapolating findings from zebrafish to humans.
Effects of phenanthrene on IKr and other IK
Inhibition of IKr by various molecules has previously been shown to induce 2:1 atrio-ventricular block and arrhythmias in zebrafish (Langheinrich et al., 2003; Milan et al., 2003; Incardona et al., 2004; Mittelstadt et al., 2008). Atrio-ventricular block and bradycardia have also been reported following phenanthrene exposure in in vivo studies using zebrafish embryos (Incardona et al., 2004; Incardona et al., 2005; Zhang et al., 2013). Interestingly, in the present study, phenanthrene exhibited differential inhibition of the potassium currents. Notably, only 24 ± 7% inhibition in the presence of 3 µM phenanthrene (Fig. 4 B) was observed during the AP-like ramp protocol (Fig. 4 A), compared with greater (50%) inhibition observed for peak-tail currents (Fig. 3 B). A significant (−6 mV) shift in peak repolarization current voltage (Fig. 4 B) during the AP-like ramp protocol was also observed. Contrary to this, the antimalarial drug halofantrine (which shares structural similarity with phenanthrene) exhibited similar inhibition of hERG channel under both AP-clamp and tail-current voltage protocol (Mbai et al., 2002). Halofantrine is a potent (21.6 nM IC50) hERG channel inhibitor in mammals that has been proposed to require channel gating to occur (Mbai et al., 2002). Under ventricular AP-clamp, halofantrine’s inhibitory action was greatest during phases 2 and 3 of the AP (Mbai et al., 2002). The lower potency (3.3 µM IC50) of IKr inhibition observed here is consistent with the smaller size of phenanthrene than halofantrine, as this would be expected to result in fewer simultaneous binding contacts with residues on the ERG/IKr channel. Alanine mutagenesis has revealed key interactions between halofantrine and aromatic (Y652, F656) residues of the hERG canonical drug binding site (Sánchez-Chapula et al., 2004).
In parallel with this study on zebrafish, structure–functional experiments on phenanthrene interactions with the human zfERG counterpart, hERG, have been undertaken to determine the nature of the phenanthrene interaction site(s) on the channel. These have demonstrated voltage and time dependence of block and a dependence of inhibition on intact inactivation gating (Al-Moubarak et al., 2020; Al-Moubarak et al. 2019. Proc Physiol Soc 43, PC016). These features are compatible with the actions of a range of known gating-dependent, pore-blocking inhibitors (Sanguinetti and Mitcheson, 2005; Hancox et al., 2008). Interestingly, phenanthrene inhibition of hERG appears to be sensitive to mutation of an S5 aromatic residue (Al-Moubarak et al., 2020; Al-Moubarak et al. 2019. Proc Physiol Soc 43, PC016).
Recently, IKs has been recorded in isolated zebrafish cardiomyocytes, although it differs from that in ventricular myocytes from mammalian species, with faster activation, most likely due to a reduced contribution of KCNE1 (Abramochkin et al., 2018). Similar to IKr, inhibition or loss-of-function mutations of IKs are associated with long QT phenotypes in mammals (Veerman et al., 2013). Indeed, IKs is thought to provide “repolarization reserve” and is thus cardio-protective against arrhythmias but via a mechanism that is distinct from the protection from premature stimulation observed here for IKr (Varró and Baczkó, 2011). The lack of effect of phenanthrene on zebrafish E4031 resistant–Ik reported here indicates that this mechanism remains intact during exposure (Fig. S2), which may contribute to the absence of APD prolongation observed in this study. Moreover, this finding supports the notion that the APD shortening effect of phenanthrene was attributable to decreased calcium currents without concomitant activation of outward potassium currents.
The lack of AP prolongation or increased beat-to-beat variability seen here with phenanthrene is not consistent with increased proarrhythmic risk linked to repolarization delay. To our knowledge, there are at present no published studies on the impact of phenanthrene on ventricular APs and underlying ionic conductances in a human experimental system. Thus, the translational relevance of these observations to the human heart remains to be established. Regardless, the suppression of the response of IKr to premature stimulation (Fig. 5 E) is consistent with increased vulnerability to premature electrical excitation. Phenanthrene significantly increased the rate of deactivation of IKr in a concentration-dependent manner (Fig. 3, D–F), and this acceleration of deactivation together with the channel-blocking effect of the drug may account for the reduced ability of IKr to generate rapid transient currents compared with control (Fig. 5 E). Under normal conditions, this transient current would be expected to counter premature depolarization (Lu et al., 2001; Perry et al., 2015), and such a response would be expected to be impaired in the presence of phenanthrene. Zebrafish APD shortening combined with increased IKr deactivation could lead to abbreviation of post-repolarization refractoriness, increasing susceptibility to arrhythmias. Post-repolarization refractoriness acts as a protective mechanism in the heart by preventing multiple compound APs from occurring (i.e., it limits the frequency of depolarization and therefore excitation rate; Grunnet et al., 2008). Consistent with this notion, accelerated IKr/hERG deactivation (as a response to extracellular acidosis) has previously been shown to reduce the protective role of IKr early in diastole and decrease AP excitation threshold in a human ventricular AP model (Du et al., 2010). This is also pharmacologically exemplified by a recent study wherein excessive AP abbreviation by type 2 hERG activator, ICA-105574, was found to be proarrhythmic (Perry et al., 2020). Phenanthrene might be anticipated to produce a similar effect, which could be further augmented under conditions of reduced IKr contribution to repolarization and protective transient currents, as would occur with loss-of-function hERG mutations (long QT syndrome 2; Perry et al., 2016). Further evidence of a key role for IKr deactivation in influencing arrhythmia susceptibility comes from a recent study with the small molecule RPR260243 that only slows hERG channel deactivation kinetics (Shi et al., 2020). This compound exhibited anti-arrhythmic properties by increasing post-repolarization refractoriness and attenuated the irregular actional potential firing by dofetilide (IKr inhibitor). Future experiments are now warranted to determine the overall effect of phenanthrene on arrhythmia susceptibility at the intact zebrafish heart level, using an in vivo or ex vivo heart model.
Plasma and urine concentrations of PAHs in humans are in the nanomolar range, with a recent study reporting phenanthrene concentrations of ∼200 nM in human blood (Huang et al., 2016). However, it is important to note that phenanthrene, like most tricyclic PAHs, is highly lipophilic (logP [octanol/water] coefficient of ∼4.4), explaining the higher tissue than plasma concentrations observed in exposure studies (Carls et al., 1999; Heintz et al., 1999; West et al., 2014). The low micromolar potency of phenanthrene on both IKr (3.3 µM; Fig. 3 C) and ICaL (4.5 µM; Fig. 2 C) reported here is therefore concerning, as concentrations as high as 3 µM have been reported in animal tissues (Camacho et al., 2012; Dhananjayan and Muralidharan, 2013).
The present study demonstrates the cardiotoxic effect of the air and water pollutant phenanthrene in an animal model with human cardiac relevance. This work demonstrates inhibition of zebrafish L-type Ca2+ channel current (IC50 = 4.5 µM) and IKr (IC50 = 3.3 µM) and also identifies acceleration of IKr deactivation as a significant action of phenanthrene, accompanied by a reduction in IKr protective transient currents at a concentration (3 µM) previously reported in human tissues (Jacob and Seidel, 2002). Interestingly, a comprehensive study looking at hERG inhibition of ∼4,800 tricyclic aromatic compounds revealed an average IC50 value of ∼2 µM (Ritchie et al., 2011). Therefore, phenanthrene and other PAHs in petroleum-based pollution could have detrimental effects on human cardiac health, warranting future studies to evaluate the impact of phenanthrene exposure on large animal models.
Jeanne M. Nerbonne served as editor.
This work is supported by British Heart Foundation grant PG/17/77/33125 (to H.A. Shiels, J.C. Hancox, and F. Brette) and in part, supported by the French National Funding Agency for Research (ANR; ANR-10-IAHU04-LIRYC and ANR-20-CE17-0010-01, to F. Brette).
The authors declare no competing financial interests.
Author contributions: S.N. Kompella was involved in conceptualization, performed experiments, analyzed data, prepared the figures, and drafted the manuscript. H.A. Shiels, J.C. Hancox, and F. Brette were involved in conceptualization; funding acquisition; supervision; and writing, reviewing, and editing the manuscript. All authors approved the final version of the manuscript.