A Novel Cardiotoxic Mechanism for a Pervasive Global Pollutant

The Deepwater Horizon disaster drew global attention to the toxicity of crude oil and the potential for adverse health effects amongst marine life and spill responders in the northern Gulf of Mexico. The blowout released complex mixtures of polycyclic aromatic hydrocarbons (PAHs) into critical pelagic spawning habitats for tunas, billfishes, and other ecologically important top predators. Crude oil disrupts cardiac function and has been associated with heart malformations in developing fish. However, the precise identity of cardiotoxic PAHs, and the mechanisms underlying contractile dysfunction are not known. Here we show that phenanthrene, a PAH with a benzene 3-ring structure, is the key moiety disrupting the physiology of heart muscle cells. Phenanthrene is a ubiquitous pollutant in water and air, and the cellular targets for this compound are highly conserved across vertebrates. Our findings therefore suggest that phenanthrene may be a major worldwide cause of vertebrate cardiac dysfunction.

Studies in zebrafish embryos have demonstrated cardiotoxic effects of many single PAH compounds and PAH mixtures, but also revealed a minimal binary division of mechanisms. PAHs that are strong carcinogens are potent agonists of the aryl hydrocarbon receptor (AHR), inducing their own metabolism and carcinogenic activation by cytochrome P4501A 8 . In general, these PAHs are also cardiotoxic to developing fish embryos at relatively high concentrations. Similar to dioxins and polychlorinated biphenyls 11,12 , these PAHs inappropriately activate the AHR in developing cardiomyocytes, leading to primary defects in cardiac morphogenesis (poor chamber looping and reduced cardiomyocyte proliferation) followed by secondary functional defects. This form of toxicity is entirely dependent on the AHR and is prevented by AHR gene knockdown. This has been demonstrated, for example, for 4-and 6-ring compounds such as benz(a)anthracene and benzo(a)pyrene [13][14][15][16] , and a C4-alkylated 3-ring compound, retene 17 . In contrast, exposure to either complex PAH mixtures derived from crude oil or single tricyclic compounds that dominate in these mixtures leads to cardiac function defects, followed by secondary morphological defects. Indeed, embryos of many fish species exposed to crude oil-derived PAHs show functional abnormalities that include bradycardia and arrhythmias characteristic of atrioventricular conduction block as well as reduced ventricular contractility [18][19][20][21][22] . Importantly, the cardiotoxicity of both crude oil and single non-alkylated tricyclic PAHs occurs without activation of the AHR in cardiomyocytes [18][19][20][21][22] , and is not prevented by AHR gene knockdown 11 . Moreover, the cardiotoxicity of crude oil is exacerbated by knockdown of cyp1a 23 , indicating that CYP1A-mediated metabolism of crude oil-derived PAHs is protective. Fractionation-based toxicity assays also linked the developmental cardiotoxicity of petroleum products to tricyclic PAH families 15 , as well as a limited number of studies in adult zebrafish [e.g., ref . 24]. Collectively, this work supports the existence of an AHR-independent mechanism by which crude oil-derived PAHs induce cardiac arrhythmia and reduce cardiomyocyte contractility.
We have recently demonstrated that complex PAH mixtures from crude oil affect excitation-contraction (EC) coupling in fish hearts 25 . EC coupling is the physiological process that links electrical excitation with contraction in a cardiomyocyte 26,27 . In fishes of the family Scombridae (e.g. mackerels, tunas 28,29 ), as in mammals 26 , EC coupling begins when an action potential (AP) depolarizes the surface membrane of the cardiomyocyte, and opens voltage-gated ion channels, allowing calcium (Ca 2+ ) entry into the cell through L-type Ca 2+ channels (I CaL ). This extracellular Ca 2+ influx triggers the release of additional Ca 2+ from the internal Ca 2+ stores of the sarcoplasmic reticulum (SR) via a process termed 'Ca 2+ -induced Ca 2+ -release' (CICR). The consequent systolic Ca 2+ transient, which activates the contractile machinery within the heart muscle cells, is the spatial and temporal sum of such local Ca 2+ releases 30,31 . Relaxation occurs when Ca 2+ is returned to resting levels by reuptake into the SR via the SR Ca 2+ ATPase (SERCA) and extrusion from the cell via the sodium calcium exchanger (NCX). Critical for AP repolarization are the opening and closing of voltage-gated sodium (Na + ), Ca 2+ and potassium (K + ) channels, which renew the EC coupling process at every heartbeat. Our earlier work 25 revealed crude oils disrupt these EC coupling pathways in scombrid fish cardiomyocytes, which explains the bradycardia and arrhythmia previously observed in the whole-heart.
Despite evidence that the tricyclic PAH fraction causes the crude oil heart failure syndrome in developing fish 20 , a direct link between an individual tricyclic PAH and the disruption of EC coupling has not been established. The present work set out to define the molecular moiety(s) of cardiotoxic PAHs from crude oil in three scombrid fishes: the Pacific mackerel (Scomber japonicas), the yellowfin tuna (Thunnus albacares) and the Pacific bluefin tuna (Thunnis orientalis). We then reveal the mechanism for this disruption in atrial and ventricular myocytes from these pelagic predators.

Results
Disruption of cellular Ca 2+ dynamics by a single tricyclic PAH. Six PAHs found in crude oil (naphthalene, fluorene, dibenzothiophene, carbazole, phenanthrene and pyrene; Supplementary Fig. S1) were applied individually to ventricular cardiomyocytes isolated from the Pacific mackerel, and Ca 2+ dynamics were assessed using confocal microscopy. Figure 1A shows the effects of each PAH on the spatial and temporal properties of cellular Ca 2+ transients recorded using a calcium-sensitive dye, Fluo-4. Cellular Ca 2+ dynamics were unaffected by PAHs with two benzene rings fused to thiophene and pyrrole rings. By contrast, the 3-ringed PAH, phenanthrene, significantly decreased the Ca 2+ transient amplitude and slowed the decay ( Fig. 1B and C). Changes in Ca 2+ flux at the single myocyte level culminate in reduced strength and rate of contraction of the whole heart 32 and could underlie the known reduction in cardiac output in embryonic vertebrate hearts following exposure to crude oil and PAHs. Therefore, we next focused on elucidating the precise mechanism by which phenanthrene altered Ca 2+ cycling using a combination of confocal Ca 2+ imaging and electrophysiology in both atrial and ventricular cardiomyocytes.

Intra-and extracellular mechanisms for altered Ca 2+ transients in tuna myocytes. Similar to
Pacific mackerel, phenanthrene (5 μ M) impaired Ca 2+ transient amplitude and decay rate in both ventricular and atrial cardiomyocytes from Pacific bluefin tuna ( Fig. 2A and B) and yellowfin tuna ( Supplementary Fig. S2). Thus, in three ecologically and commercially important pelagic species, the toxic effects of a single tricyclic PAH on cardiac cellular dynamics mirror those observed in response to more chemically complex crude oil 25 . A reduction in Ca 2+ transient amplitude could be due to reduced extracellular Ca 2+ influx (I CaL ) through L-Type Ca 2+ channels and/or a smaller Ca 2+ release from SR internal Ca 2+ stores. To clarify this point, we first used whole-cell voltage-clamp to investigate extracellular Ca 2+ influx via I CaL . We found that phenanthrene decreased the amplitude of I CaL (Fig. 3A-C). I CaL is the major source of extracellular Ca 2+ entry across the sarcolemma during EC coupling in all vertebrate hearts, including tunas and mackerel 26,29,33 . I CaL contributes substantially to the amplitude of the Ca 2+ transient and is also the major extracellular Ca 2+ trigger for CICR from intracellular SR Ca 2+ stores 26  To examine these interactions, we next exposed atrial and ventricular cardiomyocytes to pharmacological inhibitors of SR Ca 2+ release (5 μ M ryanodine to inhibit the SR Ca 2+ release channel, ryanodine receptor) and SR Ca 2+ uptake (2 μ M thapsigargin to inhibit SERCA) for 30 minutes before exposing to phenanthrene. As anticipated from earlier work 28 , SR inhibition decreased the amplitude and the rate of decay of the bluefin tuna Ca 2+ transient (Fig. 4). This confirms that the hearts of these active predators utilize SR Ca 2+ stores during EC coupling, unlike many sedentary species of fish 29 . Pharmacological pre-blockade of SR Ca 2+ cycling eliminated the effects of phenanthrene on the amplitude and the decay of the cytosolic Ca 2+ transient ( Fig. 4A and B). To further investigate the role of SR internal Ca 2+ stores, we exposed atrial myocytes to a puff of caffeine (20 mM), which causes the ryanodine receptors to open and empty SR Ca 2+ into the cytosol. We found SR Ca 2+ content was significantly decreased in myocytes exposed to phenanthrene (Fig. 4C), indicating a diminished SR Ca 2+ load. Taken together, these results suggest that phenanthrene slows the decay of the transient by limiting the reuptake of Ca 2+ into the SR via SERCA. The smaller Ca 2+ transient amplitude in the presence of phenanthrene could be caused by a the reduction in the I Ca trigger for SR release, direct effects on SR Ca 2+ release through ryanodine receptors, or both. I Kr blockade and AP prolongation are the basis for cardiac arrhythmogenesis. EC coupling is initiated by the AP. To assess whether phenanthrene alters this essential electrical property of excitable cardiomyocytes, we used whole-cell current-clamp to record APs prior to and during application of ascending concentrations of phenanthrene. Phenanthrene caused a rapid (< 1 min) and significant prolongation of AP duration (APD) associated with a dose-dependent increase in triangulation (APD 90 -APD 30 ) (Fig. 5). Such proarrhythmic responses have been implicated in atrial and ventricular fibrillation and sudden cardiac death in a number of species 34,35 .
The delayed rectifier K + channel current (I Kr ), is the key ion current responsible for AP repolarization. To determine whether the observed AP prolongation was due to an inhibition of outward K + conductance, we evaluated the effects of phenanthrene on I Kr . As shown in Fig. 5D-F, I Kr was significantly reduced by ascending concentrations of phenanthrene. In addition to explaining the prolongation of the AP, this mechanism is consistent with the phenanthrene-and crude oil-induced arrhythmogenesis and atrioventricular conduction block previously reported in other fish species 18,20,22 . Phenanthrene did not affect the cardiomyocyte resting membrane potential, AP amplitude or the rate of rise of the AP (i.e., AP traces in Fig. 5A and Supplemental Fig. S3).

Discussion
Our data demonstrate that impairment of EC coupling in cardiomyocytes by phenanthrene is a key determinant of cardiotoxicity from crude oil. This disruption of excitable cell pathways now establishes a mechanism for crude oil cardiotoxicity that has proven elusive for decades. First, phenanthrene affects membrane excitability by prolonging AP duration by inhibiting K + efflux from the cardiomyocyte via I Kr . Second, phenanthrene reduces Ca 2+ influx into the cell by reducing I CaL . This reduces myofilament activation and thus myocyte contractility. Third, inhibition of I CaL has a knock-on effect as it reduces Ca 2+ release from the SR, which further impairs cardiac contractility. Combined, these disruptions to EC coupling in the myocyte can lead to contractile failure and abnormal contractile rhythm. Our findings are in agreement with previous studies showing that SERCA function is altered in phenanthrene-exposed zebrafish 36 and miR-133a, a microRNA known to regulate hypertrophy, is reduced in rats with phenanthrene-induced cardiac hypertrophy 37 . Although some of the larger 4-and 6-ring PAHs have also been shown to alter expression of genes involved in SR calcium handling (e.g., atp2a2/SERCA2 and ryr2/ ryanodine receptor), these effects must be downstream of AHR activation 38,39 .
By applying single PAHs directly to isolated cardiomyocytes, we have identified AHR-independent cellular mechanisms that likely underpin the whole-heart cardiotoxicity phenotypes previously observed in oil-exposed fish embryos. Because of the very small size of the embryonic fish heart (e.g., roughly 300 cardiomyocytes in zebrafish; ref. 11), its limited capacity for CYP1A-mediated detoxification, and its close proximity to PAH uptake across the embryonic epidermis, reducing potential first-pass metabolic protection by CYP1A in other tissues, isolated cardiomyocytes from relative mature fish are a good proxy for the intact hearts of developing fish embryos. Prior to liver formation and the associated capacity for metabolic detoxification, fish embryos bioconcentrate tricyclic PAHs to very high tissue concentrations -i.e., in the low parts-per-million or micromolar range (e.g. refs 18 and 40). For example, parent phenanthrene levels are higher than the alkylated homologs in Alaskan crude oil at the early stages of weathering. Pacific herring embryos exposed to this oil accumulated parent phenanthrene to tissue concentrations as high as 2.5 μ M (450 parts per billion) and showed pronounced and phenanthrene pre-exposed (Caf + Phe; n = 9, N = 2) are shown to the right. *Indicates significant difference, (P < 0.05, Student's t-test). cardiac arrhythmia 18 . Thus, the exposure concentrations used here for isolated cardiomyocytes correspond to the tissue levels that produce heart form and function defects in whole embryos. Additional contributions from AHR-dependent pathways 17,21 , if any, would increase the toxic potency of complex PAH mixtures.
Our findings will help refine natural resource injury assessments for future oil spills in fish spawning habitats. As spilled oil weathers it becomes relatively enriched in phenanthrenes, thereby explaining why weathered oil (by mass) is proportionally more toxic to the fish heart. Our results also help explain why certain geologically distinct and phenanthrene-enriched crude oils have more significant cardiotoxic impacts, although we acknowledge that many PAHs and mixtures of PAHs exert their toxicity via multiple different mechanisms. This new insight may simplify the interpretation of water samples collected during oil spills for natural resource injury assessments. Future assessments should give particular weight to measured and modeled levels of phenanthrenes, in addition to complex PAH mixtures that dynamically shift in space and time throughout regions impacted by spilled oil.
Lastly, given that phenanthrenes are a near-ubiquitous component of complex environmental PAH mixtures in the oceans and the atmosphere, our findings have implications for humans that extend well beyond the Deepwater Horizon spill. A major area of public health research over the last decade has focused on the acute cardiac impacts of urban air pollution, but the precise etiology of these effects remains elusive 41 . The pro-arrhythmic actions of phenanthrene may be particularly relevant in this regard. In humans, I Kr is generated by the voltage-gated potassium channels hERG1 or hERG2 42 . Blockade of the hERG channel can lead to life-threatening arrhythmias, making this channel an important therapeutic target. Genetic and chemical studies in zebrafish indicate that the function and pharmacology of these channels are nearly identical across vertebrates [43][44][45] . Consequently, we suggest that atmospheric phenanthrene should be a concern for human cardiology, particularly due to its high abundance in urban air 46 and its rapid absorption into the bloodstream after inhalation 47 . This study should raise global interests in this important environmental pollutant given the prevalence of petroleum and PAHs in our environment.

Conclusion
We have identified a cardio-toxicant compound prominent in crude oil, and shown how it alters cardiac force and cardiac rhythm in pelagic fish. This provides a new framework for evaluating the effects of PAHs on cardiac function that we believe can be extended across a wide range of vertebrates, including humans. Reducing future releases of this pro-arrhythmic chemical should substantively benefit human and ecological health. Methods Fish origin and care. Mackerel (fish mass = 0.69 ± 0.14 kg, heart mass = 1.3 ± 0.3 g, mean ± SEM, N = 11), bluefin tuna (fish mass = 14.1 ± 1.0 kg, heart mass = 52.9 ± 4.8 g, mean ± SEM, N = 16) and yellowfin tuna (fish mass = 12.2 ± 2.0 kg, heart mass = 30.9 ± 37 g, mean ± SEM, N = 6) were captured off San Diego, CA, held aboard the F/V Shogun in seawater wells, and then transported by truck to the Tuna Research and Conservation Center (Pacific Grove, CA). Mackerel and tunas were held in a 30-m 3 and 109-m 3 tank respectively at 20 ± 1 °C and fed a diet of squid, sardines, and enriched gelatin, as previously described 48,49 . Fish were acclimated to 20 °C for at least 4 weeks before experimentation. Individual cardiomyocytes were isolated using protocols previously described in detail 28 . All procedures were in accordance with Stanford University Institutional Animal Care and Use Committee protocols. All experimental protocols were approved by the Stanford University Animal Care and Use Committee.
Chemicals. All solutions were prepared using ultrapure water supplied by a Milli-Q system (Millipore, USA).

Intracellular [Ca 2+
] measurements. Confocal Ca 2+ imaging was performed using a laser-scanning unit attached to an Olympus inverted microscope. Control and PAH-exposed myocytes were loaded with 4 μ M Fluo-4 AM (Molecular Probes) for 20 min, washed via dilution to de-esterify and then perfused with standard Ringer solution. The dye was excited at 488 nm and fluorescence measured at > 500 nm. Transverse line scans were acquired at 5 ms intervals. Cells were electrically stimulated at 0.5 Hz via extracellular electrodes. Batches of myocytes were incubated with 5 μ M of PAH for at least 1 hour. Control experiments were performed on time-matched untreated cells. Some cells from both control and PAH-exposure groups were incubated with SR inhibitors (5 μ M ryanodine and 2 μ M thapsigargin) for at least 30 minutes prior to imaging. In some experiments, a caffeine pulse (20 mM) was applied via a home-built rapid solution system. All line scan images are presented as original raw fluorescence (F). Background fluorescence (F 0 ) was measured in each cell in a region that did not have localized or transient fluorescent elevation.
Electrophysiological recordings. Electrophysiological data was recorded as previously described 51 .
Briefly, at the beginning of each trial, myocytes were allowed to settle for ~10 minutes in the recording chamber, and then perfused with control external solution. Membrane potentials and currents were recorded from each myocyte in whole-cell mode under baseline (control) conditions, and again at increasing concentrations of phenanthrene in the extracellular solution. Stimuli used to elicit ion currents and APs are provided in the figure legends. Briefly, the L-type Ca 2+ channel current (I CaL ) was elicited by a pulse to 0 mV (the approximate peak of the current-voltage relationship in cardiomyocytes) after a pre-pulse to − 40 mV to inactivate Na + current. I CaL was measured as the difference between peak and the end of pulse current. Trains of depolarizing pulses were applied at 0.2 Hz. Action potentials (APs) were evoked using 10 ms sub-threshold current steps at a frequency of 0.5 Hz. The delayed rectifier K + current (I Kr ) was measured using an established protocol adapted from previous studies on Pacific bluefin tuna 52 and rainbow trout 53 . I Kr was activated by a pre-pulse to + 40 mV (to fully activate K + channels) and measured as the tail current at − 20 mV, the maximum tail current in tuna myocytes 53 . To separate rapid K + current (I Kr ), tail I K amplitude was measured as the current sensitive to a specific I Kr inhibitor (2 μ M E-4031). Trains of depolarizing pulses were applied 0.2 Hz every 20 seconds. Data was recorded via a Digidata 1322 A A/D converter (Axon Instruments, CA) controlled by an Axopatch 200B (Axon Instruments, CA) amplifier running pClamp software (Axon Instruments, CA). Signals were filtered at 1-10 kHz using an 8-pole Bessel low pass filter before digitization at 10-20 kHz and storage. Patch pipette resistance was typically 1.5-3 MΩ when filled with intracellular solution (below). Cell membrane capacitance was measured using the "membrane test module" in Clampex (fitting the decay of the capacitance current recorded during a 10 mV depolarizing pulse from a holding potential of − 80 mV). Data analysis. Ca 2+ transients from confocal experiments were analyzed using Image J, Clampfit (Axon Instruments, CA), and Origin (OriginLab Corporation, MA). At least three traces at steady state were analyzed and averaged. The decay of the Ca 2+ transient was fitted with a single exponential to calculate Tau of decay (i.e., time to decrease to 37% of the peak amplitude). The amplitude of the Ca 2+ transient is defined as peak increase Scientific RepoRts | 7:41476 | DOI: 10.1038/srep41476 and basal Ca 2+ (F/F 0 ) 54 . Electrophysiological data were analyzed using Clampfit and Origin software. All AP parameters were stable over the time of recording (< 10 min) in controls. Currents are expressed as current density (pA/pF). I CaL was measured as the difference between the peak inward current and the current at the end of the depolarizing pulse. I Kr amplitude was measured as the current sensitive to E-4031.
Statistics. Data are presented as mean ± SEM Statistical analysis was performed using SigmaStat software (Systat Software, CA). Electrophysiological data with a confirmed normal distribution and equal variance were analyzed using One Way Repeated Measures Analysis of Variance or Friedman Repeated Measures Analysis of Variance on Ranks followed by a Student-Newman-Keuls test for post hoc analysis. For confocal data, unpaired Student's t-test were used within the same species of tuna, one-way ANOVAs were used to test for the effect of PAHS in mackerel, and two-way ANOVAs were used to determine interactions between species, SR inhibition and PAHs. P < 0.05 was considered significant.