Dual Mechanisms of Cardiac Action Potential Prolongation by 4-Oxo-Nonenal Increasing the Risk of Arrhythmia; Late Na+ Current Induction and hERG K+ Channel Inhibition

4-Oxo-nonenal (4-ONE) is an endogenous lipid peroxidation product that is more reactive than 4-hydroxy-nonenal (4-HNE). We previously reported the arrhythmic potential of 4-HNE by suppression of cardiac human Ether-a-go-go Related Gene (hERG) K+ channels with prolonged action potential duration (APD) in cardiomyocytes. Here, we illustrate the higher arrhythmic risk of 4-ONE by modulating the cardiac hNaV1.5 channel currents (INaV). Although the peak amplitude of INaV was not significantly changed by 4-ONE up to 10 μM, the rate of INaV inactivation was slowed, and the late Na+ current (INaL) became larger by 10 μM 4-ONE. The chemical modification of specific residues in hNaV1.5 by 4-ONE was identified using MS-fingerprinting analysis. In addition to the changes in INaV, 4-ONE decreased the delayed rectifier K+ channel currents including the hERG current. The L-type Ca2+ channel current was decreased, whereas its inactivation was slowed by 4-ONE. The APD prolongation by 10 μM of 4-ONE was more prominent than that by 100 μM of 4-HNE. In the computational in silico cardiomyocyte simulation analysis, the changes of INaL by 4-ONE significantly exacerbated the risk of arrhythmia exhibited by the TdP marker, qNet. Our study suggests an arrhythmogenic effect of 4-ONE on cardiac ion channels, especially hNaV1.5.

We previously reported that 4-HNE has a potential arrhythmic effect on the heart by extending the action potential duration (APD), which was mediated by the inhibition of human Ether-a-go-go Related Gene (hERG) K + channel current (I Kr ) [13]. In addition to the voltage-gated K + channels, such as hERG, various functional disturbances of the human cardiac Na + channel (hNa V 1.5) are associated with an increased risk of arrhythmia [14]. The SCN5A gene encodes the hNa V 1.5 α-subunit, and mutations in SCN5A are associated with inherited susceptibility to ventricular arrhythmia, such as Brugada syndrome, long QT syndrome class 3 (LQT-3), or atrial fibrillation [15,16].
A gain-of-function mutation of SCN5A leads to increased Na + influx during systole, resulting in delayed action potential repolarization or early afterdepolarization (EAD) of the cardiac AP [16]. Specifically, the persistent or non-inactivating component of hNa V 1.5, called the late Na + current (I NaL ), could be responsible for the prolonged APD of LQT-3. However, the chemical modification of hNa V 1.5 and its arrhythmogenic effect, such as I NaL induction, has been rarely investigated. Interestingly, previous studies have shown that the oxidative condition of cardiac ischemia and heart failure enhanced I NaL [17][18][19]. The plausible changes of hNa V 1.5 current (I NaV ) and the putative induction of I NaL in ROS-mediated arrhythmia attracted us to investigate the modification of hNa V 1.5 activity by RCS.
In the present study, we highlighted the arrhythmic potentials of 4-ONE, which is formed from 4-hydroperoxy-2-nonenal, the same precursor as 4-HNE [1]. Structurally, 4-ONE differs at the C4 position with a ketone group instead of the hydroxyl group of 4-HNE, increasing the electrophilic reactivity of 4-ONE. Therefore, 4-ONE modifies various nucleophilic amino acids, such as cysteine (Cys), lysine (Lys), histidine (His), and arginine (Arg) [20][21][22]. However, a previous study on the effects of 4-ONE on ion channel activity was limited to TRPA1 and TRPV1 nonselective cation channels as the harmful sensory signals [23]. In addition to hNa V 1.5, we also examined the effects of 4-ONE on hERG (I Kr ), KCNQ1/KCNE1 (I Ks ), and L-type voltage-operated Ca 2+ channels (I Ca,L ). Finally, the relative contribution of the I NaV modulation to APD prolongation and arrhythmogenic risk was analyzed by a recently announced method of proarrhythmic risk analysis called Comprehensive in vitro Proarrhythmia Assay (CiPA), cooperatively using experimental data and in silico simulation [24,25].

Electrophysiological Recording
Conventional whole-cell voltage and current-clamp were conducted for currents and AP recordings, respectively. For the I NaV recording, high giga-seal resistance (>2 GΩ), low series resistance (<10 MΩ), and the series resistance compensation (80%) were introduced to reduce voltage-clamp error. The extracellular bath solution for the I NaV and I NaL recordings in hNa V 1.5-HEK cells contained 130 mM NaCl, 10 mM HEPES (4-(2-hydroxyethyl)-1piperazineethanesulfonic acid), 4 mM CsCl, 1 mM MgCl 2 , 2 mM CaCl 2 , and 10 mM glucose adjusted to pH 7.4 with NaOH. The intracellular pipette solution for the I NaV and I NaL recordings contained 117 mM CsCl, 20 mM NaCl, 1 mM MgCl 2 , 5 mM HEPES, 5 mM EGTA, 5 mM MgATP, and 0.4 mM TrisGTP adjusted to pH 7.3 with CsOH. The extracellular bath solution for the I Kr and I Ks recordings contained 145 mM NaCl, 3.6 mM KCl, 10 mM HEPES, 1 mM MgCl 2 , 1.3 mM CaCl 2 , and 5 mM glucose adjusted to pH 7.4 with NaOH. The intracellular pipette solution for the I Kr and I Ks recordings contained 100 mM K-aspartate, 25 mM KCl, 5 mM NaCl, 10 mM HEPES, 1 mM MgCl 2 , 4 mM MgATP, and 10 mM BAPTA adjusted to pH 7.25 with KOH. The extracellular bath solution for I Ca,L contained 145 mM CsCl, 10 mM HEPES, 1 mM MgCl 2 , 1.8 mM CaCl 2 , and 5 mM glucose adjusted to pH 7.4 with CsOH. The intracellular pipette solution for I Ca,L contained 106 mM CsCl, 20 mM TEA-Cl, 5 mM NaCl, 10 mM HEPES, 5 mM MgATP, and 10 mM EGTA adjusted to pH 7.25 with CsOH. The compositions of the extracellular solutions used for the AP recording contained 145 mM NaCl, 5.4 mM KCl, 10 mM HEPES, 1 mM MgCl 2 , 1.8 mM CaCl 2 , and 5 mM glucose adjusted to pH 7.4 with NaOH. The intracellular solution contained 120 mM K-aspartate, 20 mM KCl, 5 mM NaCl, 2 mM CaCl 2 , 5 mM EGTA, 10 mM HEPES, and 5 mM MgATP adjusted to pH 7.25 with KOH.

In Silico Simulation
CiPAORdv1.0 (modified O'Hara-Rudy model) was used to simulate human ventricular AP and its changes due to the altered ionic currents (I Kr , I Ks , I Ca,L , I NaV , and I NaL ) by 4-ONE and 4-HNE. The levels of ionic current inhibition and the equations of inactivation time constant obtained from the experimental results are presented in Table 1.

Tandem Mass Spectrometry
The total lysates of the hNa V 1.5-HEK cells treated with 4-ONE (10 µM) were subjected to SDS-PAGE for mass spectrometry (MS). The hNa V 1.5 bands were cut from the SDS-PAGE gel and digested in gel with trypsin (Promega, Madison, WI, USA). The subsequent procedures were similar to the previous MS [13]. A fragment mass tolerance of 1.0 Da, peptide mass tolerance of 25 ppm, and maximum missed cleavage of 2 were set. The result filters were performed with charge states versus scores (XCorr by Sequest) where the minimal scores for the charge states were +1: 1.6, +2: 1.7, +3: 3.0, and >+4: 3.5. The carbamidomethylation (+57.021 Da) of cysteine (C) was set as a static modification, and the following variable modifications were allowed: Michael addition, +154 Da (C, H, K, R); Schiff base addition, +136 Da (C, H, K); and oxidation, +15.995 Da (M). The respective data for the post-translational modification (PTM) sites by 4-ONE were transformed and analyzed with Scaffold 4 program (Proteome Software, Portland, OR, USA).

Chemicals
The compounds 4-ONE and 4-HNE were purchased from Cayman Chemical (Ann Arbor, MI, USA). The 4-ONE and 4-HNE were stored in 20 mM stocks in DMSO at −20 • C. Immediately prior to the application to the cells, 4-ONE and 4-HNE were freshly diluted with extracellular bath solution to the final target concentrations. Application of 4-ONE and 4-HNE was processed for at least 5 min to obtain stable electrophysiological responses. Other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Statistical Analysis
Data are expressed as mean ± S.E., and the statistical analyses were determined using paired or unpaired Student's t-tests. A p-value < 0.05 was considered as statistically significant.

Slowed hNa V 1.5 Inactivation and I NaL Induction by 4-ONE
The effect of 4-ONE on cardiac hNa V 1.5 was evaluated using stably overexpressing hNa V 1.5 in the HEK-293 cell line (hNa V 1.5-HEK cell). In the whole-cell voltage-clamp condition, the inward I NaV was recorded by applying −40 mV of depolarization pulse (300 ms) from −120 mV holding potential. After confirming the stable recording of I NaV , 4-ONE was applied to the bath perfusing solution, which reduced the peak amplitude of I NaV in a dose-dependent manner ( Figure 1A; remaining current after 4-ONE treatment: 94.25% for 1 µM, 88.22% for 10 µM, 72.53% for 30 µM, and 63.66% for 60 µM 4-ONE; n = 3, n = 6, n = 3, and n = 3;, respectively). The reduced I NaV was not restored by washing 4-ONE (data not shown), which was similar to the irreversible effect of 4-HNE on I hERG , as previously reported [13]. It has been reported that the in vivo concentration of 4-HNE under pathophysiological conditions ranges 1-100 µM [1]. In contrast, the in vivo concentration of 4-ONE has not been reported. However, an experiment on EA.hy 926 endothelial cells treated with ferrous sulfate suggested that the endogenous concentration of 4-ONE could be increased to 20 µM [26]. Therefore, we applied 10 µM 4-ONE in the subsequent experiments.
The current-voltage (I-V) relationship curves of the I NaV showed a minute decrease in the peak amplitude at 10 µM 4-ONE ( Figure 1C; peak inward current at −40 mV of −1071.8 ± 84.17 and −972.3 ± 85.52 pA/pF for control and 4-ONE, respectively, n = 6), whereas 4-HNE had no significant effect even at 100 µM ( Figure 1B). The I-V curves were converted to the conductance-voltage (G-V) curve for analyzing the voltage dependence of hNa V 1.5, which showed a slight left-shift, indicating that 4-ONE could reduce the threshold of activation ( Figure 1E; half-maximal voltage of activation of −53.7 and −56.3 mV for control and 4-ONE, respectively, n = 6). The steady-state inactivation property of hNa V 1.5 was analyzed by using the double pulse protocol ( Figure 1D, inset). The steady-state inactivation curve also showed a slight left-shift by 4-ONE ( Figure 1D; half-maximal voltage of inactivation of −74.5 and −77.5 mV for control and 4-ONE, respectively, n = 6).
Upon analysis of the inactivation speed of I NaV , the rate of inactivation that was slowed by 4-ONE (Figure 2A, left) was notable. When the normalized decaying components of I NaV were fit to a double exponential equation, both time constants for the fast and the slow components (τ fast and τ slow ) were increased by 4-ONE (Figure 2A, right). The delayed inactivation of I NaV suggested an increase in I NaL, which is the residual activity of hNa V 1.5 that was flowing after the large peak Na + current during AP. To analyze I NaL more specifically, we applied the AP-like voltage-clamp protocol, two-step depolarization followed by a reverse-ramp voltage pulse ( Figure 2B, upper gray line). The resurgent inward current during the reverse-ramp period reflected the augmented I NaL by 4-ONE ( Figure 2B, b; current density of −2.79 ± 0.27 and −5.16 ± 0.53 pA/pF for control and 4-ONE, respectively, n = 13). The sustained current at 50 ms after the peak Na + influx was increased as well ( Figure 2B, a; current density of −3.44 ± 0.56 and −8.99 ± 0.96 pA/pF for control and 4-ONE, respectively, n = 13). The increased inward currents (a and b) in the presence of 4-ONE were reversed by additional application of 50 µM of ranolazine, a late Na + current inhibitor ( Figure 2B, −5.50 ± 1.04 and −3.14 ± 0.60 pA/pF, a and b, respectively, n = 6). In contrast to the significant induction of I NaL by 10 µM 4-ONE, the application of 100 µM 4-HNE induced neither slower inactivation nor I NaL ( Figure 2C,D).
The effects of 4-ONE on I NaV could be due to the PTM of hNa V 1.5, i.e., direct binding of 4-ONE with nucleophilic amino acids, such as Cys, His, Lys, and Arg [20,27] shows that His 445 , His 472 , and Lys 496 were commonly modified with Schiff base addition ( Figure 3A-C). Another spectrum shows the Michael addition of Arg 878 in the peptide 864 NYSELRDSDSGLLPr 878 ( Figure 3D). The 4-ONE-binding sites were visualized with the schematic topology of the hNa V 1.5 channel ( Figure 3E). The sites of the Schiff base addition were located at the intracellular linker between domain I (DI) and domain II (DII). The site of the Michael addition was located at the extracellular S5-S6 linker of DII. Upon analysis of the inactivation speed of INaV, the rate of inactivation that was slowed by 4-ONE (Figure 2A, left) was notable. When the normalized decaying components of INaV were fit to a double exponential equation, both time constants for the fast and the slow components (τfast and τslow) were increased by 4-ONE (Figure 2A, right). The de- The effects of 4-ONE on cardiac K + channels were evaluated using the hERG and KCNQ1/KCNE1 expressing HEK-293 cells. The acute treatment of 10 µM of 4-ONE reduced the peak amplitudes of I Kr (hERG K + current) and I Ks (KCNQ1/KCNE1 current) by 65% ( Figure 4A; peak current density of 73.91 ± 7.55 and 25.93 ± 5.32 pA/pF at 20 mV for control and 4-ONE, respectively, n = 8) and 29%, respectively ( Figure 4B; peak current density of 35.37 ± 8.95 and 25.66 ± 5.39 pA/pF at 40 mV for control and 4-ONE, respectively, n = 6). The cardiac I Ca,L was recorded from GPVMs. The peak amplitude of I Ca,L was decreased by 45% ( Figure 4C; peak inward current at 0 mV of −4.30 ± 0.29 and −2.36 ± 0.36 pA/pF for control and 4-ONE, respectively, n = 5). It was notable that 4-ONE also slowed the inactivation of I Ca,L ( Figure 4D). When the inactivation phase of I Ca,L was fit to double exponential function, the slow component of time constant (τ slow ) became larger by 4-ONE at 0 and 10 mV ( Figure 4D; τ slow at 0 mV of 122.8 ± 11.68 and 145.9 ± 10.94 ms and τ slow at 10 mV of 126.1 ± 9.92 and 144.3 ± 2.77 ms for control and 4-ONE, respectively, n = 7).
Antioxidants 2021, 10, x FOR PEER REVIEW  (B) I NaL through hNa V 1.5 channel was recorded by applying action potential-like repolarization pulse protocol. The I NaV was activated by short depolarization to −20 mV from −120 mV of hyperpolarized potential (a). The resurgent I NaL was then recorded during ramp pulse repolarization (b). (C,D) 4-HNE treatment induced neither the inactivation decay of I NaV nor the I NaL . All the data were analyzed using paired t-tests, where a p < 0.05 was considered statistically significant (*).

APD Prolongation and Increased Risk of Arrhythmia by 4-ONE
The effects of 4-ONE on the cardiac AP were analyzed in GPVM under the currentclamp condition and triggered at 1 Hz. The bath application of 10 μM of 4-ONE markedly prolonged the APD (Figure 5A, B; APD90, 309.2 ± 44.50 and 729.4 ± 67.07 ms for control and 4-ONE, respectively, n = 10), which was more prominent than the effect of 100 μM of

APD Prolongation and Increased Risk of Arrhythmia by 4-ONE
The effects of 4-ONE on the cardiac AP were analyzed in GPVM under the currentclamp condition and triggered at 1 Hz. The bath application of 10 µM of 4-ONE markedly prolonged the APD (Figure 5A, B; APD 90 , 309.2 ± 44.50 and 729.4 ± 67.07 ms for control and 4-ONE, respectively, n = 10), which was more prominent than the effect of 100 µM of 4-HNE, as reported previously [13]. The maximum depolarization speed and total amplitude of APs were not affected by 4-ONE. In addition, the resting membrane potential of GPVMs was not changed (Figure 5B, right).

Discussion
Our present study shows prominent cardiac APD prolongation by 4-ONE (10 μM) with multiple effects on the cardiac ion channels. We have previously reported that 100 μM of 4-HNE also induces APD prolongation with the inhibition of IKr [13]. In addition to the difference in the effective concentrations of the RCS, the APD prolongation and the risk of arrhythmia predicted by CiPA were commonly more prominent with 10 μM of 4-ONE than with 100 μM of 4-HNE ( Figure 5). The genetic dysfunction or pharmacological inhibition of IKr has been regarded as one of the main mechanisms of APD prolongation The contribution of APD prolongation simulated by I Ca,L and I Na (I NaV and I NaL ) added to I K input. (E) qNet (net charge carried by total ionic currents) was calculated under 4-HNE and 4-ONE inputs. (F) The contribution of qNet simulated by I Ca,L and I Na added to I K input. All the data were analyzed using paired t-tests, where a p < 0.05 was considered statistically significant (*).
The CiPA, different from the conventional cardiotoxicity analysis investigating I Kr only, covers the measurements of I Kr , I Ca,L , I NaV , and I NaL for the analysis using the in silico model (CiPAORdv1.0: modified O'Hara-Rudy ventricular myocyte model). Using CiPAORdv1.0, we simulated the AP reflecting the electrophysiological changes induced by 4-ONE treatment ( Figure 5C,D). For the calculation of the effects of 4-ONE, the relative conductance of I Kr and I Ks was decreased to 0.4 and 0.7, respectively. For I Ca,L , I NaV , and I NaL , in addition to the relative conductance, the changes of inactivation kinetics induced by 4-ONE were applied (Table 1). For 4-HNE simulation, the inputs with reduced I Kr and I Ks were applied according to our previous report [13]. The simulated APs revealed markedly prolonged APD by total input of 4-ONE. The decrease in I Kr was more effective than that of I Ks for the APD prolongation. However, it was notable that the modifications of both K + currents (I Kr and I Ks ) were insufficient to simulate the change by 4-ONE ( Figure 5C). The changes of inward currents (I NaV , I NaL , and I Ca,L ) were additionally introduced. While the changes of I Ca,L (slower inactivation and reduced conductance) had an insignificant effect, the increase in I NaL showed a significant additional prolongation of APD ( Figure 5D).
The risk of severe arrhythmia, such as Torsades de Pointes (TdP), is evaluated by a novel in silico biomarker, qNet (net charge carried by total ionic currents), proposed from CiPA [24,25]. The decrease in qNet by 10 µM of 4-ONE was more significant than that by 100 µM of 4-HNE ( Figure 5E), indicating a higher risk of 4-ONE for arrhythmia induction. In addition, the sufficient reduction of qNet was observed by combining the changes of I Kr , I Ks , and I Na , but not by the simulation using the changes of I Kr , I Ks , and I Ca,L ( Figure 5F), which were consistent with the results of the stepwise simulation of APD change induced by 4-ONE.

Discussion
Our present study shows prominent cardiac APD prolongation by 4-ONE (10 µM) with multiple effects on the cardiac ion channels. We have previously reported that 100 µM of 4-HNE also induces APD prolongation with the inhibition of I Kr [13]. In addition to the difference in the effective concentrations of the RCS, the APD prolongation and the risk of arrhythmia predicted by CiPA were commonly more prominent with 10 µM of 4-ONE than with 100 µM of 4-HNE ( Figure 5). The genetic dysfunction or pharmacological inhibition of I Kr has been regarded as one of the main mechanisms of APD prolongation and EAD. While sharing the inhibitory effect on I Kr with 4-HNE, an additional intriguing finding was the augmentation of I NaL by 4-ONE ( Figure 2).

I NaL and Inactivation of Na V 1.5
The very rapid activation of hNa V 1.5 is responsible for the fast activation wave and synchronous initiation of cardiac contraction. The inactivation process is also rapid, which prevents wasteful Na + entry throughout the AP plateau in cardiomyocytes. However, cardiac I NaV also shows residual flow during the sustained depolarization. Although I NaL is relatively negligible to the fast component (0.1%-0.5% of peak I NaV ), the continuous activity in the AP plateau could contribute to determining the shape and duration of the cardiac AP. Congenital gain-of-function mutations in SCN5A coding hNa V 1.5 cause LQT-3. LQT-3 patients have a high risk not only for TdP but also for atrial fibrillation [14][15][16].
I NaL is generally thought to be a persistent opening of the channels modulated either to slow the inactivation or to reopen over the voltage ranges between steady-state activation and inactivation curves, called a "window" potential. An enlargement of the window potential could be induced by the shift of activation or inactivation curves and has been reported as a mechanism of LQT-3 [14,28,29]. In our results, 4-ONE slightly shifted the inactivation curve to the left, implying a narrowed window potential at the relatively positive ranges ( Figure 1D, right panel). Considering the voltage difference between the AP plateau (>0 mV, Figure 5) and the window potential under treatment with 4-ONE (below −40 mV, Figure 1E), it is unlikely that the current during the window period of AP could play a significant role for I NaL induction [30].
More importantly, we found that the speed of hNa V 1.5 inactivation was slowed by 4-ONE but not by 4-HNE ( Figure 2). Cardiac I NaV flows through a channel formed by the α-subunit encoded by SCN5A, which alone accounts for major features of I NaV including the fast inactivation. A previous study suggested a structure responsible for the fast inactivation of I NaV resides in IFM motif (isoleucine-phenylalanine-methionine) on the linker between the third and fourth repeat (DIII-DIV linker) as a "ball" or "lid" and on the bottom of the S4-S5 linker of each repeat ( Figure 3E) [31]. In addition to the classical domain for fast inactivation, the perturbation of many locations can destabilize the inactivation and cause pathological I NaL [31][32][33].
For the mechanisms of I NaL by physiological PTM of hNa V 1.5, CaMKII-dependent phosphorylation of Ser 571 [34] and PKC-dependent phosphorylation of Ser 1503 [35] have been reported. In addition, the nNOS (NOS1)-dependent S-nitrosylation was suggested, although the precise location of the candidate Cys has not been identified [36]. The noncongenital acquired increase in I NaL is often observed in cardiomyocytes isolated from ischemic hearts and may be due to oxidative stress with increased ROS [37][38][39]. However, no previous study has paid attention to the modification of hNa V 1.5 by 4-ONE that could be abundantly produced by ischemia/reperfusion conditions. In this regard, our present study might suggest a novel mechanism of I NaL induction by ischemia/reperfusion-induced oxidative stress of the heart.
Through the MS/MS analysis, we could identify the binding sites of 4-ONE to hNa V 1.5 (His 445 , His 472 , Lys 496 , and Arg 878 ). Since the electrophysiological changes by 4-ONE was not reversed by washout with control solution, we carefully suggest that PTM sites revealed by the MS/MS analysis might be the candidate for the slowed inactivation and the increase in I NaL (Figure 3). Although the modified residues are not equivalent to the reported mutations in the congenital LQT-3 patients [14,16,32], those sites are relatively close to the binding sites of a known I NaL activator, veratridine ( Figure 3E) [40,41]. The site-directed mutagenesis of hNa V 1.5 and the electrophysiological investigation are requested to identify the actual roles of the modified residues in the I NaL and the altered inactivation. Regretfully, we have not conducted the MS/MS analysis with hNa V 1.5-HEK cells treated with 4-HNE. Since the treatment with 4-HNE did not induce the functional changes in I NaV inactivation and I NaL , the comparative analysis might provide more specific information for the critical residue(s) of Na V 1.5 modified by 4-ONE.

Pathophysiological Implication of 4-ONE and I NaL
4-ONE-mediated I NaL induction might have a pathophysiological significance. Increased I NaL in the heart can lead to arrhythmia by prolonging APD in a direct manner and by causing Ca 2+ overload in an indirect manner. As for the former mechanism, the resurgent I NaL at the repolarization phase of AP interferes with rapid repolarization and can cause EAD-associated arrhythmia. For the latter mechanism, the prolonged APD leads to Ca 2+ overload by I Ca,L and Na + -Ca 2+ exchanger, triggering pathological Ca 2+ release from intracellular Ca 2+ storing organelles. The Ca 2+ overload also causes diastolic dysfunction, increased wall stress, and ischemic risk [42]. In this regard, I NaL has been suggested as an attractive therapeutic target to treat arrhythmia, heart failure, and angina. Ranolazine, the most selective clinical I NaL inhibitor, has been used to suppress both arrhythmia events and angina [42,43]. In our result, 4-ONE-mediacted I NaL was effectively reduced by 50 µM ranolazine ( Figure 2B), further implying the pathophysiological role of 4-ONE in terms of the cardiac ischemia-associated arrhythmia.

Application of CiPA in Silico Model
To assess the arrhythmogenic risk of 4-ONE, we applied the CiPA in silico model. The inhibition of I Kr alone could suggest a pathophysiological implication of 4-ONE. Interestingly, the qNet analysis and AP simulations revealed a higher risk of 4-ONE than of 4-HNE, which is due to the I NaL induction. Such insight could not be obtained from the conventional cardiotoxicity test of the I Kr analysis alone, which reflects the strength of CiPA that includes the integrative simulation of the multiple types of cardiac ion channels.
Another interesting feature of the present study was the slowed inactivation and the reduced peak amplitude of I Ca,L by 4-ONE treatment, which was not observed in the previous study of 4-HNE [13]. However, according to the CiPA analysis, the enhanced persistent Ca 2+ current modulated by the slowed I Ca,L inactivation did not induce significant changes of the qNet and the simulated APD ( Figure 5D,F), which appears to be due to the compensation by the decrease in peak current activation ( Figure 4C).

Conclusions
Using electrophysiological investigation of cardiac ion channel currents, for the first time, we discovered the multichannel effects of 4-ONE, among which the inhibition of I Kr and the induction of I NaL were noteworthy, as confirmed by the qNet reduction indicating arrhythmogenic risk.