Circulation Journal ORIGINAL ARTICLE Circ J 2020; 84: 1931 – 1940 doi: 10.1253/circj.CJ-20-0288 /Electrophysiology

Binge Alcohol Exposure Triggers Atrial Through T-Type Ca2+ Channel Upregulation via Protein Kinase C (PKC) / Glycogen Synthesis Kinase 3β (GSK3β) / Nuclear Factor of Activated T-Cells (NFAT) Signaling ― An Experimental Account of Holiday Heart Syndrome ―

Yan Wang, MD, PhD; Masaki Morishima, PhD; Dan Li; Naohiko Takahashi, MD, PhD; Tetsunori Saikawa, MD, PhD; Stanley Nattel, MD, PhD; Katsushige Ono, MD, PhD

Background: The association between binge alcohol ingestion and (AF), often termed “holiday heart syndrome”, has long been recognized. However, the underlying cellular and molecular mechanisms are unknown.

Methods and Results: An experimental model of binge alcohol-induced AF was developed to elucidate the mechanisms linking acute ethanol exposure to changes in ion channel transcription and AF susceptibility. AF-susceptibility during transesophageal electrical stimulation was enhanced 8 h after, but not immediately or 24 h after, acute alcohol intake. T-type calcium channel (TCC) blockade and calcineurin inhibition diminished the AF-promoting effect of ethanol. Long-term (8–24 h) exposure to ethanol augmented TCC isoform-expression (Cav3.1 and Cav3.2) and currents in cardiomyocytes, accompanied by upregulation of the transcription factors, Csx/Nkx2.5 and nuclear factor of activated T-cells (NFAT), in the nucleus, and of phospho-glycogen synthesis kinase 3β (GSK3β) in the cytosol. Inhibition of protein kinase C (PKC) during the 7- to 8-h period following ethanol exposure attenuated susceptibility to AF, whereas acute exposure did not. GSK3β inhibition itself upregulated TCC expression and increased AF susceptibility.

Conclusions: The present study results suggest a crucial role for TCC upregulation in the AF substrate following binge alcohol- drinking, resulting from ethanol-induced PKC-activation that hyperphosphorylates GSK3β to cause enhanced calcineurin-NFAT-Csx/ Nkx2.5 signaling. These observations elucidate for the first time the potential mechanisms underlying the clinically well-recognized, but mechanistically enigmatic, “holiday heart syndrome”.

Key Words: Alcohol; Atrial fibrillation; Electrophysiology; Ion channels

trial fibrillation (AF) is the most commonly sus- tained cardiac arrhythmia. Although certain risk Editorial p 1909 factors, such as age, hypertension, serum metabolites A 1–3 6 and are well-established, the causes and within 24 h. Several potential mechanisms have been pos- mechanisms remain unknown in many patients. There is tulated by which an alcoholic binge could cause arrhyth- conflicting evidence regarding an association between mias. Alcohol-induced increases in plasma-free fatty acids long-term alcohol consumption and the risk of AF.4 In and catecholamines are thought to be arrhythmogenic,7 as contrast, an association between binge alcohol-drinking is the principal metabolite of alcohol, acetaldehyde.8 How- and AF, often termed “holiday heart syndrome”, has been ever, the precise electrophysiological mechanisms that connect widely recognized.5 Holiday heart syndrome refers to cardiac acute ethanol consumption to AF are largely unknown. , particularly AF, that occur after an alcoholic Pulmonary veins (PVs) are important foci for initiation binge in individuals showing no other evidence of heart of paroxysmal AF and are also associated with AF main- disease, which usually convert to normal sinus rhythm tenance.9 PVs contain cardiomyocytes with electrical activ-

Received March 31, 2020; revised manuscript received July 20, 2020; accepted August 7, 2020; J-STAGE Advance Publication released online October 7, 2020 Time for primary review: 20 days Department of Pathophysiology (Y.W., M.M., D.L., K.O.), Department of Clinical Examination and Diagnostics (Y.W., N.T., T.S., K.O.), Oita University School of Medicine, Yufu, Japan; Montreal Heart Institute Research Center, University of Montreal, Montreal (S.N.), Canada The first two authors contributed equally to this study (Y.W., M.M.). Mailing address: Katsushige Ono, MD, PhD, Department of Pathophysiology, Oita University School of Medicine, 1-1 Idaigaoka, Hasama, Yufu 879-5593, Japan. E-mail: [email protected] All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: [email protected] ISSN-1346-9843

Circulation Journal Vol.84, November 2020 1932 Wang Y et al.

Figure 1. Effect of ethanol on heart rate and atrial fibrillation (AF) inducibility. Changes in heart rate over time after (A) drinking water (20 mL) or of ethanol (20%, 20 mL), and (B) injection of saline (10 mL) or ethanol (20%, 10 mL) into the abdominal cavity, are shown. Examples of the electrocardiogram (ECG) record for vehicle (water or saline; ○) and ethanol (●) 8 h after the procedure are shown in the insets. Number of animals are indicated in parentheses. Examples of the ECG lead II deflec- tion from a rat 8 h after receiving (C) a saline injection, (D) an ethanol injection, (E) an ethanol injection with a short-term (1 h) application of kurtoxin (0.1 mg/kg), (F) an ethanol injection with a long-term (7 h) application of a PKC inhibitor, chelerythrine (5 mg/kg), or (G) an injection of a GSK3β inhibitor, BIO (0.1 μg/kg). The duration of AF was designated as the period from the end of electrical stimulation to the appearance of the first P wave (▼). Magnification of the ECGs for 1 s duration are shown in the insets (C,D). (H) Incidence of AF episodes in the different groups and (I) mean AF duration in rats treated with saline (10 mL) or with ethanol (20%, 10 mL) alone or cotreated with kurtoxin (0.1 mg/kg), mibefradil (5 mg/kg), cyclosporine (50 mg/kg), BIO (0.1 μg/kg), SB216763 (0.5 μg/kg), chelerythrine (5 mg/kg) or atenolol (5 mg/kg) for the durations indicated in parentheses. Each rat (n=10) received 10 sets of electrical stimulations with an interval of 10 min. *P<0.01 vs. vehicle, #P<0.01 vs. ethanol (8 h).

Circulation Journal Vol.84, November 2020 Binge Alcohol-Induced AF Mechanism 1933

Figure 2. Changes in action potentials (APs) induced by ethanol applied in vivo and in vitro. (A) Representative APs recorded from pulmonary vein (PV) cardiomyocytes isolated from an adult rat 8 h after injection of saline (vehicle, 10 mL) or 20% ethanol (10 mL) into the abdominal cavity. (B) Representative APs recorded from neonatal cardiomyocytes cultured without ethanol (vehicle) or with ethanol (0.1%) for 24 h. (C–E) AP parameters of PV and neonatal cardiomyocytes (Neo) with or without ethanol procedures: AP cycle (C), V˙max at −50 mV (D), and the maximum diastolic potential (MDP) (E). (F) Group data of the beating rate of cardiomyocytes with (gray) or without (white, vehicle) ethanol exposure for 0, 1, 3, 6, 12, and 24 h. *P<0.01 vs. vehicle. #P<0.05 vs. vehicle. Number of cells (C–F) are indicated; each animal heart was assessed by using 10 different preparations.

ity, which have been suggested to function as subsidiary animal model of “holiday heart syndrome”; (2) to use the pacemakers and to induce atrial arrhythmias.10 In addi- model to clarify the molecular links between acute ethanol tion, PV cardiomyocytes have distinct electrophysiological consumption and subsequent transient AF; and (3) to characteristics and may possess arrhythmogenic activity establish the role, if any, of TCC dysregulation in experi- through several mechanisms.11 mental “holiday heart syndrome”. The T-type Ca2+ channel (TCC) current (ICa.T) is present in nodal cells, Purkinje fibers, and PV cardiomyocytes, but Methods is less important in normal atrial and ventricular myo- cytes.12 Cardiac pacemaker cells possess a greater density A detailed description of the experimental methods is pro- of ICa.T than non-pacemaker cells, and ICa.T contributes to vided in Supplementary File 1. the genesis of automaticity.12 In contrast, in pathological conditions like hypertrophy, , and Animal ECG Recordings and Electrophysiological heart failure, TCC is re-expressed in ventricular myo- Measurements cytes.13 TCC blockade may prevent the development of a The experimental protocol was approved in advance by the substrate for AF.14 Moreover, Chen et al demonstrated Ethics Review Committee for Animal Experimentation of that TCC blockade decreases spontaneous activity, sup- Oita University School of Medicine (Approval number presses delayed afterdepolarizations, and inhibits transient C004003). Detailed profiles and procedures for ECG, inward currents in PV cardiomyocytes, indicating a role development of an animal model for ethanol-induced AF, 15 for ICa.T in PV arrhythmogenesis. It has thus been sug- and patch clamp recordings were described in previous gested that TCC is correlated with the onset and/or the studies.17,18 Briefly, we applied 10 mL ethanol (20%) for maintenance of AF.16 Regarding the pathophysiological injections and 20 mL (20%) for the alcohol binge. Rats signals associated with the T-type Ca2+ channel, the role of voluntarily drank this amount of ethanol with ease after a the cardiac transcription factor, Csx/Nkx2.5, in positive 24- to 30-h water-fasting period. AF inducibility was assessed regulation of the Cav3.2-TCC has recently been pro- after a burst of transesophageal atrial electrical pacing posed.17 As a working hypothesis, we assumed that actions (70 V, 15 s, 33.3 Hz). For short-term ECG recording with a of TCCs may form part of the mechanism by which binge burst of transesophageal atrial pacing, a 400 series PowerLab alcohol consumption leads to AF. In addition, we consid- with Chart v4 software (ADInstruments, Bella Vista, NSW, ered that a sub-acute transient action of ethanol, rather Australia) connected with a BIO amp (ADInstruments) than an immediate or long-term effect, may create a sub- was used. strate favoring AF occurrence via modulation of TCC. The purpose of the present study was: (1) to create an

Circulation Journal Vol.84, November 2020 1934 Wang Y et al.

Statistical Analysis Data are expressed as mean±SEM. Between group and among group comparisons were conducted using one-way ANOVA with Scheffe’s test, respectively. P<0.05 was con- sidered significant.

Results Ethanol-Induced Increase of Heart Rate and Cellular Automaticity Possible actions of ethanol on heart rate change and arrhythmogenicity were examined over 24 h by using tele- metric ECG recordings from adult rats that drank ethanol (20%, 20 mL) or distilled water (20 mL), as well as from rats that were injected with ethanol (20%, 10 mL) or saline (10 mL) via the abdominal cavity (Figure 1). Heart rate increased after either binge drinking or injection of etha- nol, and reached its maximum 8 h later (Figure 1A,B). The increase in heart rate might not be a direct effect of ethanol or serum catecholamines, because the maximum ethanol and noradrenalin (NA) concentration in serum reached a peak 2 h after the ethanol binge (Supplementary Figure 1A). Importantly, ethanol concentration decreased significantly by 8 h (n=4–5), which did not correspond to the time Figure 3. Long-term actions of ethanol on L-type (LCC) and course of heart rate change (Supplementary Figure 1C). T-type Ca2+ channel (TCC) currents in PV cardiomyocytes. Cardiac autonomic nerve function was not directly related (A–C) Examples of current traces recorded in PV cardiomyo- to heart-rate change, but was based on heart rate variability cytes of adult rats at 8 h after a water or an ethanol binge. (A) in relation to serum concentrations of NA (Supplementary ICa obtained from a VHP of −100 mV; (ICa.L+ICa.T). (B) ICa Figure 1B,D,E). We then assessed the induction of AF by obtained from a VHP of −40 mV in the same patch; (ICa.L). (C) ICa obtained by subtraction of the traces in (B) from the traces transesophageal atrial burst pacing in rats with or without in (A), which represents ICa.T. (D,E) Current (I)-voltage (V) ethanol binging. After rapid transesophageal electrical relationship of the group data for ICa.L (D) and ICa.T (E) from 8 stimulation, ethanol consumption increased the probabil- cells out of 8 animals in each group. *P<0.01 vs. saline. ity of AF development by as much as 92±4%; AF was not substantially induced in animals that were injected with saline (Figure 1C–I). AF was not induced immediately after ethanol injection (1 h), but rather predominantly 8 h later; the probability of AF induction returned towards Cardiomyocyte Isolation and Culture baseline 24 h after ethanol injection. The incidence and the Neonatal cardiomyocytes were prepared from 2- to 3-day- duration of AF caused by ethanol was strongly reduced by old Wistar rat heart ventricles, as previously described.17 kurtoxin, a potent selective inhibitor of TCCs (Figure 1E), PV cardiomyocytes were prepared from adult Wistar rats suggesting that TCCs mediate AF promotion by binge (8 weeks old) using a procedure modified from that of ethanol. Pretreatment with the calcineurin (CN) inhibitor, Chen et al.15 cyclosporine, or the PKC inhibitor, chelerythrine, for 7 h significantly reduced the incidence and duration of AF fol- Measurement of Protein Kinase C (PKC) Activation lowing ethanol administration, although acute application Membranous PKC activity was measured using the (1 h) of cyclosporine or chelerythrine had no effect StressXpress PKC kinase activity assay kit (EKS-420A; (Figure 1H,I). Although the β-adrenergic blocking agent, Stressgen Bioreagents Corp., Victoria, British Columbia, atenolol (5 mg/kg), significantly reduced heart rate from Canada). 431±23 bpm to 367±20 bpm (P<0.01) 1 h after intraperito- neal application, it failed to suppress ethanol-related AF Western Blot Analysis (Figure 1H,I; Supplementary Figure 2C). Intriguingly, gly- Western blot analysis was performed according to a stan- cogen synthesis kinase 3β (GSK3β) inhibition by BIO or dard method.19 Total protein of the nucleus and the cytosol SB216763 greatly increased AF susceptibility and AF was extracted using the NE-PER Nuclear and Cytoplasmic duration (Figure 1G–I and Supplementary Figure 2E), sug- Extraction Reagents kit (Pierce, Rockford, IL, USA) and gesting PKC and GSK3β as potential signaling molecules was quantified using the BCA Protein Assay Kit (Pierce). associating binge alcohol drinking with AF. Increased excitability of PV cardiomyocytes due to etha- Transfection of siRNA and the Luciferase Activity Assay nol activity has been postulated as a possible trigger for siRNAs directed against rat Nkx2 (Nkx siRNA, pools of AF induction. We performed in vitro electrophysiological 3 target-specific 27 mer siRNAs) and the scrambled nega- experiments using spontaneously beating PV cardiomyo- tive control siRNA (Control siRNA) were purchased from cytes from adult rats. Examples of action potentials (APs) ORIGENE (SR512716; OriGene Technologies, MD, recorded from PV cardiomyocytes isolated from a rat 8 h USA). Transfection was performed in OPTI-MEM (Invi- after injection of saline or ethanol are shown in Figure 2A. trogen). PV automatic rate increased approximately 3-fold in PV cardiomyocytes obtained from ethanol-injected rats

Circulation Journal Vol.84, November 2020 Binge Alcohol-Induced AF Mechanism 1935

Figure 4. Long-term actions of ethanol on L-type (LCC) and T-type Ca2+ channel (TCC) expression in adult and neonatal cardio- myocytes. (A) The mRNA expression of Cav1.2 (A-a), Cav1.3 (A-b), Cav3.1 (A-c) and Cav3.2 (A-d) in atrial, ventricular and pulmo- nary vein (PV) cardiomyocytes of adult rats 8 h after drinking 20 mL of water (white bars) or 20 mL of 20% ethanol (gray bars). Data were normalized to Cav1.2 mRNA expression in atrial cardiomyocytes isolated from water-drinking rats, which was designated as 100. (B) Changes in Cav1.2 (B-a), Cav1.3 (B-b), Cav3.1 (B-c) and Cav3.2 (B-d) mRNA expression in neonatal cardiomyocytes treated with vehicle or ethanol (0.01%, 0.1% and 0.5%) in the culture medium for 24 h. (C) Time-dependent changes in Cav1.2 (C-a), Cav1.3 (C-b), Cav3.1 (C-c) and Cav3.2 (C-d) mRNA in neonatal myocytes treated with 0.1% ethanol in the culture medium for 0–24 h. Representative polymerase chain reaction (PCR) products are shown in the insets with the reference gene, GAPDH (below). *P<0.01 vs. vehicle (0% ethanol) (A,B) or specific time vs. 0 h (C).

(Figure 2C). In order to obtain an in vitro model, we stud- sistent with the results in PV cardiomyocytes, ethanol ied neonatal cardiomyocytes kept in primary culture with treatment (0.1%, 24 h) also increased ICa.T but not ICa.L, Ih, or without 0.1% ethanol for up to 24 h (Figure 2B). The or IKr in neonatal cardiomyocytes (Supplementary Figure 4). spontaneous beating rate was 88±13 bpm in the absence of To investigate the effects of ethanol on Ca2+ channel iso- ethanol (vehicle), and was 126±19 bpm in the presence of form expression, we used quantitative polymerase chain- ethanol (Figure 2C). Although the upstroke velocity at a reaction (qPCR) to analyze the mRNA expression of membrane potential of −50 mV (V˙ @−50 mV) was markedly Cav1.2 and Cav1.3, representing LCC isoforms, and Cav3.1 increased, the maximum diastolic potential (MDP) was and Cav3.2, representing TCC isoforms. The Cav1.2- and unchanged by ethanol both in adult PV cardiomyocytes Cav1.3-LCCs were expressed in PVs at lower levels than and in neonatal cardiomyocytes (Figure 2D,E). The beat- other cardiomyocytes (Figure 4A-a,b). Neither the mRNA ing rate of neonatal cardiomyocytes was significantly expression of Cav1.2 or Cav1.3 was changed in atrial, increased by 0.1% ethanol after incubation for 6–24 h ventricular or PV cardiomyocytes 8 h after an ethanol (Figure 2F). These results indicate that ethanol increases binge in adult rats (Figure 4A-a,b), nor in neonatal myo- beating rates and upstroke velocity of APs at potentials cytes cultured with ethanol (0.001–0.5%) for up to 24 h comparable to the pacemaker range several hours follow- (Figure 4B-a,b; Figure 4C-a,b). In contrast, Cav3.1- and ing ethanol drinking in vivo, and at comparable intervals Cav3.2-TCC isoforms were highly expressed in PV cardio- following ethanol treatment in vitro. myocytes, and were upregulated 8 h after an ethanol binge (Figure 4A-c,d). The upregulation of Cav3.1 and Cav3.2 by Ethanol Increases TCC Current and mRNA Expression ethanol in neonatal cardiomyocytes occurred at the con- As kurtoxin strongly decreased the effect of ethanol on AF centration of 0.1% or higher (Figure 4B-d), and it was in a induction (Figure 1E,H,I), actions of ethanol on TCC cur- time-dependent manner (Figure 4C-c,d). Unlike the tran- rents were suspected. The ICa.T was significantly increased sient upregulation of Cav3.1- and Cav3.2-TCC in rats after in PV myocytes 8 h after ethanol injection, although ICa.L an acute ethanol binge, the upregulation of these channels was unchanged (Figure 3). The increase in ICa.T nearly com- was sustained for more than 24 h when cardiomyocytes pletely reversed to baseline values in PV cardiomyocytes were exposed continuously to ethanol in vitro (Figure 4C-c,d). 24 h after an ethanol binge (Supplementary Figure 3). Con- As TCC upregulation occurred several hours after etha-

Circulation Journal Vol.84, November 2020 1936 Wang Y et al.

Figure 5. Effect of ethanol on transcription factors and signaling molecules. (A) Protein kinase C (PKC) activity in cardiomyocytes with or without 24 h treatment with culture medium containing 0.01%, 0.1% and 1% ethanol (n=8). (B) Changes in mRNA expres- sion of the transcription factors, Csx/Nkx2.5, GATA4, CREB, NFATc4 and NRSF in myocytes treated with or without 0.1% ethanol for 24 h. Expression of each mRNA without ethanol treatment was assigned as 100% (n=8). (C) Examples of Csx/Nkx2.5, NFAT, GSK3β, phosphorylated GSK3β (p-GSK3β) and GAPDH protein expression in the nucleus and cytosol of myocytes treated with or without 0.1% ethanol for 24 h. (D) Semi-quantitative assessment of protein expression levels of Csx/Nkx2.5, NFAT, GSK3β, and phosphorylated p-GSK3β determined based on the density of the blotted bands exemplified in panel (C) (n=5). The protein expres- sion of each sample was normalized to that in the nucleus of cells treated with vehicle. (E) A luciferase assay after transient transfection of cardiomyocytes with TOPFLASH and renilla; cells were treated with (gray bars) or without 0.1% ethanol for 24 h (open bars) in the presence or absence of PKC inhibitors (2 μmol/L chelerythrine, 5 nmol/L Gö 6976, 15–50 nmol 3-IYIAP), a Ca2+ chelator (50 μmol/L BAPTA-AM) or GSK3β inhibitors (20 nmol/L BIO, 10 μmol/L SB216763) (n=8). *P<0.01 vs. vehicle.

nol application, a genomic mechanism was postulated. To out 0.1% ethanol. Actinomycin D in the culture medium investigate the potential signaling mechanisms and consid- prevented ethanol-induced increases in ICa.T and TCC- ering the suppressant effect of chelerythrine on ethanol- subunit expression; actinomycin D-induced downregula- related AF, we examined PKC activity and the involvement tion of the expression of Csx/Nkx2.5 mRNA by 40.2% was of transcription factors that regulate cardiac ion channel unaffected by ethanol addition (Supplementary Figure 5). expression. PKC activity in cardiomyocytes was signifi- To additionally test whether ethanol-mediated upregula- cantly enhanced when the cells were exposed to ethanol at tion of TCC is specifically due to inhibition of GSK-3β by a concentration of 0.1% or higher (Figure 5A). Under the PKC-mediated phosphorylation, ethanol-treated cardio- same conditions, the mRNA levels of the transcription myocytes were analyzed by TOPFLASH (a reporter plasmid factors, Csx/Nkx2.5, known to regulate Cav3.2 transcrip- containing multiple copies of wild-type Tcf-binding sites) tion, and GATA4 (but not those of CREB, NFAT or reporter assay with or without PKC inhibitors (chelery- NRSF), were increased by ethanol (Figure 5B). NFAT is thrine, Gö 6976, 3-IYIAP), and the results were compared an important target of GSK3β kinase activity in the heart. to the effects of the GSK-3β inhibitors, BIO and SB216763 Using a protein expression assay, we found that the levels (Figure 5E). Ethanol treatment enhanced TOPFLASH of Csx/Nkx2.5 and NFAT in the nuclear fraction were reporter activity, an action which depended upon intracel- more than doubled when cardiomyocytes were exposed to lular Ca2+ and PKC actions. ethanol. These increases were accompanied by an increase in phosphorylated (deactivated) GSK3β (pGSK), a sub- Effect of PKC-Dependent Signals on TCC Transcription strate of PKC, in the cytosol (Figure 5C,D). To confirm To further examine downstream target molecules that that ethanol modifies transcription of the TCC gene, we might mediate the effects of ethanol on TCC expression, cultured cardiomyocytes in the presence of actinomycin D we added the PKC inhibitor, chelerythrine, and/or the (0.01 μmol/L), an inhibitor of transcription, with or with- Ca2+ chelator, BAPTA-AM, to the culture medium for 24 h,

Circulation Journal Vol.84, November 2020 Binge Alcohol-Induced AF Mechanism 1937

Figure 6. Effects of protein kinase C (PKC), cellular Ca2+ and GSK3β on ICa.T, Cav3 and Csx/Nkx2.5 expression. (A) Examples of current traces of ICa.T recorded from saline-treated (white triangles) or ethanol-treated (filled triangles) neonatal cardiomyocytes cultured in normal medium for 24 h, with 2 μmol/L chelerythrine (Chele, a PKC inhibitor), with 50 μmol/L BAPTA-AM (BAPTA, a Ca2+ chelator), or with chelerythrine plus BAPTA-AM. (B,C) changes in Cav3.1 mRNA (B) and Cav3.2 mRNA (C) induced by treatment with 0.1% ethanol for 24 h, with or without chelerythrine or BAPTA-AM. Changes in the mRNA expression of (D) Cav3.1, (E) Cav3.2, and (F) Csx/Nkx2.5 induced by 0.1% ethanol treatment for 24 h with or without PKC inhibitors; the PKCα inhibitors Gö 6976 (5 nmol/L) or Ro-32-0432 (30 nmol/L), the PKCβI inhibitor, 3-IYIAP (15 nmol/L), or the PKCβI/II inhibitor, 3-IYIAP (50 nmol/L), or with or without the GSK3β inhibitor, BIO (20 nmol/L). *P<0.01 vs. control (as indicated). #P<0.01 vs. control (ethanol (–), PKC inhibitor (−) or BIO (−)). Representative polymerase chain reaction (PCR) products are shown in the insets with the reference gene, GAPDH (below).

with or without ethanol. The ethanol-induced increase in PKCβI and PKCβII, the upregulation of Cav3.2 by ethanol ICa.T and in the mRNA expression of TCCs was completely was completely abolished. Finally, to confirm the partici- eliminated in the presence of BAPTA (Figure 6A). Chel- pation of GSK3β in the ethanol signaling pathway, we erythrine also abolished the effect of ethanol on ICa.T as well examined the effects of ethanol in the absence or presence as on Cav3.1- and Cav3.2-mRNA expression (Figure 6A). of BIO (20 nmol/L), a GSK3β inhibitor. The expression of To determine which [Ca2+]i-dependent isoform of PKC Cav3.2-mRNA was strongly increased by BIO alone, and affects TCC transcription, we examined the effect of ethanol ethanol was unable to further increase the Cav3.2 level in the absence or presence of several PKC isoform-specific (Figure 6E,F). Ethanol regulation of Cav3.2 and Csx/ inhibitors: the PKCα inhibitors, Gö 6976 and Ro-32-0432, Nkx2.5 expression in the presence of specific PKC inhibitors and the PKCβ inhibitor, 3-IYIAP. The effect of ethanol on or a GSK3β inhibitor were strikingly similar (Figure 6E,F). Cav3.1 was abolished by Gö 6976 or Ro-32-0432, whereas Finally, to confirm our findings implicating Csx/Nkx2.5, its effect on Cav3.2 was unaffected (Figure 6D,E). When a short interfering RNAs for Nkx2.5 (Nkx siRNA) were low dose of 3-IYIAP was used, which inactivates PKCβII, applied along with ethanol treatment (Figure 7, the levels of both Cav3.1 and Cav3.2 were increased by Supplementary Figure 6). Nkx siRNA abolished the upreg- ethanol. In contrast, when a high concentration of ulation of Cav3.2 by ethanol as well as by PMA (Figure 7A,B). 3-IYIAP was used, which blocks the activity of both Nkx siRNA also reduced ethanol- and PMA-dependent

Circulation Journal Vol.84, November 2020 1938 Wang Y et al.

Figure 7. Regulation of Cav3.2 by a tran- scription factor, Csx/Nkx2.5, and its modula- tion by siRNA-Csx/Nkx2.5 (Nkx siRNA) in cardiomyocytes. (A) Representative immu- nostaining of cardiomyocytes for Cav3.2 and DAPI incorporation with ethanol (EtOH) or a PKC activator, PMA (1 nmol/L) for 24 h. (B) Knockdown of Cav3.2-mRNA by Nkx siRNA. (C) Mean spontaneous beating rate of cardio- myocytes after application of EtOH (0.1%) or PMA (1 nmol/L) for 24 h. Significant reduction of the beating rate was observed in Nkx siRNA-transfected myocytes with or without EtOH/PMA. Data are expressed as mean ± SE. *P<0.05 as indicated. Nkx siRNA corresponds to Nkx (B) siRNA in Supplementary Figure 6.

increases in beating rates of cardiomyocytes (Figure 7C). These results confirm that activation of Csx/Nkx2.5 action by ethanol results in TCC current increases that enhance cardiomyocyte excitability.

Discussion Holiday Heart Syndrome and the Actions of Ethanol Clinicians have long recognized the temporal association of episodic alcohol use with the onset of AF,20 but the mechanisms have remained elusive. The present study demonstrated that acute ethanol exposure triggers tran- siently enhanced AF susceptibility in an animal model. Our detailed studies implicated TCC upregulation medi- ated by a PKC/GSK3β/NFAT/Csx pathway in the molec- ular pathophysiology of ethanol binge-related AF.

TCC Upregulation as a Possible Arrhythmogenic Substrate Although the pathophysiological actions of TCC in car- diac muscle are still controversial, recent investigations revealed that TCC current is larger in PVs than in other cardiac regions.15 In addition, the TCC blocker suppressed -induced atrial remodeling in experimental ani- mals.21 Previous studies indicate that PV cardiomyocytes are arrhythmogenic through the generation of triggered Figure 8. Schematic illustration of ethanol-protein kinase C activity.22 Furthermore, TCCs were highly expressed in PV (PKC) pathways that participate in T-type calcium channel cardiomyocytes and contributed to their pacemaker activ- (TCC) upregulation via GSB3β-calcineurin-NFAT-Csx signals ity and triggered activity.15,23 Ethanol acutely inhibits Ca2+ for atrial fibrillation (AF) promotion; a possible molecular 24 model of “holiday heart syndrome”. channels. However, here we noted that ethanol is a tran- scriptional modulator of TCCs in PVs and neonatal car- diomyocytes. Transient upregulation of TCCs following

Circulation Journal Vol.84, November 2020 Binge Alcohol-Induced AF Mechanism 1939 an experimental ethanol binge, along with suppression of tinct role of Csx/Nkx2.5 in the formation and identity of ethanol-induced AF promotion by a TCC blocker, suggest PV cardiomyocytes.34 The fact that perinatal loss of Csx/ that TCCs are involved in the initiation of ethanol-related Nkx2.5 results in a large reduction in cardiac TCC levels 35 AF in our model. (both Cav3.1 and Cav3.2) in the heart suggests that PV automaticity can be controlled by Csx/Nkx2.5 even in Downstream Signal Transduction adult animals. Ethanol has been shown to interact with membrane-asso- In fetal alcohol syndrome (FAS), the genomic action of ciated phospholipid substrates involved in signal transduc- ethanol is postulated to play a prominent role. FAS is tion pathways.25 Ethanol activates phospholipase C and characterized by organ defects, fetal growth restriction, triggers associated cellular signaling responses, including neurodevelopmental delays, and craniofacial malforma- the formation of inositol-1,4,5-triphosphate and diacylg- tions, with a high incidence of cardiac conduction defects, lycerol, leading to the stimulation of PKC.26 PKC is ventricular septal defects and valvular diseases.36 The car- involved in a variety of pathophysiological cell signaling diac anomalies in FAS are almost identical to those systems, including AF.27 Based on the structural and func- observed in mice with Csx/Nkx2.5 mutations.36 Our data tional properties of the kinase regulatory domain, enzymes raise the interesting possibility that ethanol-induced Csx/ of the PKC family are divided into 3 groups: conventional Nkx2.5 abnormalities might be involved in inducing FAS. (cPKC), novel (nPKC) and atypical (aPKC) isoforms.28 As the non-specific PKC inhibitor, chelerythrine, and the Study Limitations membrane-permeable Ca2+-chelator, BAPTA-AM, com- Although the present study results provide compelling evi- pletely abolished the effects of ethanol on Cav3-mRNA dence for a contribution of TCC upregulation to ethanol- and ICa,T, we speculated that cPKC was involved in the induced PV automaticity and AF inducibility, we cannot regulation of Cav3 expression. Numerous studies indicate exclude the possibility that other changes may also con- that PKC can phosphorylate and inhibit GSK3β, which tribute. It is possible that the effects of ethanol-dependent regulates a wide variety of cardiac transcription factors.29 TCC upregulation on AF-susceptibility in rats are different For example, GSK3β inhibits endothelin-1 (ET1)-induced from those in humans, especially in diseased hearts. hypertrophy in neonatal cardiomyocytes, and prevents Accordingly, whether increased ethanol-PKC signaling NFAT transcriptional activation by retaining NFAT in and upregulation of TCC contribute to the development of the cytosol or by delaying ET1-induced nuclear importa- clinical AF remains to be determined. tion of NFAT.30 Because nuclear localization of NFAT is accelerated when it is dephosphorylated by CN, NFAT Conclusions phosphorylation by GSK3β should counteract dephos- phorylation by CN.31 In transgenic mice, overexpression of Here, we developed a novel animal model of binge alcohol- S9A (a mutant form of GSK3β, which cannot be phos- induced AF susceptibility. This model was used to assess phorylated at Ser 9 and is constitutively active), signifi- underlying cellular and molecular mechanisms. Ethanol cantly decreased the nuclear localization of NFAT.30 activated intracellular Ca2+-dependent PKC isoforms, One molecule that might function to connect NFAT and which inhibited GSK3β through enhanced phosphoryla- Cav3.2 is Csx/Nkx2.5. Chen and Cao recently demon- tion, thereby increasing nuclear NFAT translocation and strated that Csx/Nkx2.5 is a direct target of NFAT that Csx/Nks2.5 expression, leading to upregulation of TCCs. co-ordinates with other transcription factors such as Blockade of TCCs or suppression of TCC-upregulating GATA4 to regulate Csx/Nkx2.5 during cardiogenesis.32 pathways prevented AF promotion by binge alcohol. The Our previous research demonstrated that overexpression present study thus provides novel insights into the molecu- of Csx/Nkx2.5 by adenovirus-mediated gene transfer lar mechanisms underlying the previously enigmatic “holi- markedly increased the spontaneous beating rate, ICa.T and day heart syndrome”. 17 mRNA expression of Cav3.2. Our results therefore sug- gest that the effects of binge ethanol drinking to cause TCC Acknowledgments channel upregulation are mediated through cPKC/GSK3β/ The authors would like to thank Dr. F. Hamada and Miss Y. Akiyoshi NFAT and Csx/Nkx2.5 signaling pathways (Figure 8). for their technical support. This finding raises the question of whether other signals that phosphorylate GSK3β, such as protein kinase B Sources of Funding (Akt), regulate AF susceptibility. Because PI3-kinase and This work was supported, in part, by KAKEN grants #21590934 to Akt are known to play a role in the inactivation of GS3β K.O., #21.09356 to Y.W., and the Canadian Institutes of Health by phosphorylation, which occurs after activation of the Research and Canadian Heart Foundation grants to S.N. insulin receptor, transient hyperglycemia and/or during an insulin surge may account for AF epidemiology. Actually, Disclosures diabetes mellitus is an independent risk factor for AF. The N.T., S.N. are members of Circulation Journal’ Editorial Team. potential role of TCC changes in linking insulin dysregula- tion to AF substrates may be an interesting topic for future IRB Information exploration. Oita University granted an exemption for this study to require ethics approval. Physiological and Pathophysiological Transcription by Nkx/Csx2.5 References Csx/Nkx2.5 is an evolutionarily conserved transcription 1. Harada M, Melka J, Sobue Y, Nattel S. Metabolic consider- factor of the homeobox gene family, which plays a crucial ations in atrial fibrillation: Mechanistic insights and therapeutic role in cardiogenesis, and whose expression continues opportunities. Circ J 2017; 81: 1749 – 1757. throughout adulthood.33 A recent study identified a dis- 2. Ono K. How is uric acid related to atrial fibrillation? Circ J 2019;

Circulation Journal Vol.84, November 2020 1940 Wang Y et al.

83: 705 – 706. 20. Greenspon AJ, Schaal SF. The “holiday heart”: Electrophysio- 3. Taufiq F, Maharani N, Li P, Kurata Y, Ikeda N, Kuwabara M, logic studies of alcohol effects in alcoholics. Ann Intern Med et al. Uric acid-induced enhancements of Kv1.5 protein expres- 1983; 98: 135 – 139. sion and channel activity via the Akt-HSF1-Hsp70 pathway in 21. Fareh S, Bénardeau A, Thibault B, Nattel S. The T-type Ca2+ HL-1 atrial myocytes. Circ J 2019; 83: 718 – 726. channel blocker mibefradil prevents the development of a sub- 4. Sano F, Ohira T, Kitamura A, Imano H, Cui R, Kiyama M, et strate for atrial fibrillation by tachycardia-induced atrial remod- al. Heavy alcohol consumption and risk of atrial fibrillation: The eling in dogs. Circulation 1999; 100: 2191 – 2197. Circulatory Risk in Communities Study (CIRCS). Circ J 2014; 22. Chen YJ, Chen SA, Chen YC, Yeh HI, Chan P, Chang MS, et 78: 955 – 961. al. Effects of rapid atrial pacing on the arrhythmogenic activity 5. Tonelo D, Providência R, Gonçalves L. Holiday heart syndrome of single cardiomyocytes from pulmonary veins: Implication revisited after 34 years. Arq Bras Cardiol 2013; 101: 183 – 189. in initiation of atrial fibrillation. Circulation 2001; 104: 2849 – 6. Schuckit MA. Alcohol and alcoholism. In: Fauci AS, Braunwald 2854. E, Kasper DL, Hauser SL, Longo DL, Jameson JL, et al, editors. 23. Malécot CO. Low voltage-activated channels in rat pulmonary Harrison’s principles of internal medicine (17th edn). New York, vein cardiomyocytes: Coexistence of a non-selective cationic USA: McGraw-Hill, 2008; 2724 – 2729. channel and of T-type Ca channels. Pflugers Arch 2020; 472: 7. Tansey MJ, Opie LH. Relation between plasma free fatty acids 1019 – 1029. and arrhythmias within the first twelve hours of acute myocar- 24. Habuchi Y, Furukawa T, Tanaka H, Lu LL, Morikawa J, dial infarction. Lancet 1983; 2: 419 – 422. Yoshimura M. Ethanol inhibition of Ca2+ and Na+ currents in 8. Condouris GA, Havelin DM. Acetaldehyde and cardiac arrhyth- the guinea-pig heart. Eur J Pharmacol 1995; 292: 143 – 149. mias. Arch Int Pharmacodyn Ther 1987; 285: 50 – 59. 25. Hoek JB, Rubin E. Alcohol and membrane-associated signal 9. Macle L, Jaïs P, Scavée C, Weerasooriya R, Shah DC, Hocini M, transduction. Alcohol Alcohol 1990; 25: 143 – 156. et al. Electrophysiologically guided pulmonary vein isolation 26. Kuroiwa M, Aoki H, Kobayashi S, Nishimura J, Kanaide H. during sustained atrial fibrillation. J Cardiovasc Electrophysiol Role of GTP-protein and endothelium in contraction induced by 2003; 14: 255 – 260. ethanol in pig coronary artery. J Physiol 1993; 470: 521 – 537. 10. Pappone C, Rosanio S, Oreto G, Tocchi M, Gugliotta F, 27. Christ T, Boknik P, Wöhrl S, Wettwer E, Graf EM, Bosch RF, Vicedomini G, et al. Circumferential radiofrequency ablation of et al. L-type Ca2+ current downregulation in chronic human pulmonary vein ostia: A new anatomic approach for curing atrial atrial fibrillation is associated with increased activity of protein fibrillation.Circulation 2000; 102: 2619 – 2628. phosphatases. Circulation 2004; 110: 2651 – 2657. 11. Zhou S, Chang CM, Wu TJ, Miyauchi Y, Okuyama Y, Park 28. McKinsey TA, Olson EN. Toward transcriptional therapies for AM, et al. Nonreentrant focal activations in pulmonary veins in the failing heart: Chemical screens to modulate genes. J Clin canine model of sustained atrial fibrillation. Am J Physiol 2002; Invest 2005; 115: 538 – 546. 283: H1244 – H1252. 29. Hardt SE, Sadoshima J. Glycogen synthase kinase-3β: A novel 12. Zhou Z, Lipsius SL. T-type calcium current in latent pacemaker regulator of cardiac hypertrophy and development. Circ Res cells isolated from cat right atrium. J Mol Cell Cardiol 1994; 26: 2002; 90: 1055 – 1063. 1211 – 1219. 30. Haq S, Choukroun G, Kang ZB, Ranu H, Matsui T, Rosenzweig 13. Rossier MF. Function and differential expression on T-type A, et al. Glycogen synthase kinase-3beta is a negative regulator calcium channels in various pathophysiological states. Recent of cardiomyocyte hypertrophy. J Cell Biol 2000; 151: 117 – 130. Res Dev Biochem 2003; 4: 13 – 29. 31. Fiedler B, Wollert KC. Interference of antihypertrophic mole- 14. Fareh S, Bénardeau A, Nattel S. Differential efficacy of L- and cules and signaling pathways with the Ca2+-calcineurin-NFAT T-type calcium channel blockers in preventing tachycardia- cascade in cardiac myocytes. Cardiovasc Res 2004; 63: 450 – 457. induced atrial remodeling in dogs. Cardiovasc Res 2001; 49: 32. Chen Y, Cao X. NFAT directly regulates Nkx2-5 transcription 762 – 770. during cardiac cell differentiation.Biol Cell 2009; 101: 335 – 349. 15. Chen YC, Chen SA, Chen YJ, Tai CT, Chan P, Lin CI. T-type 33. Komuro I, Izumo S. Csx: A murine homeobox-containing gene calcium current in electrical activity of cardiomyocytes isolated specifically expressed in the developing heart. Proc Natl Acad Sci from rabbit pulmonary vein. J Cardiovasc Electrophysiol 2004; USA 1993; 90: 8145 – 8149. 15: 567 – 571. 34. Mommersteeg MT, Brown NA, Prall OW, de Gier-de Vries C, 16. Garratt CJ, Fynn SP. Atrial electrical remodelling and atrial Harvey RP, Moorman AF, et al. Pitx2c and Nkx2-5 are required fibrillation.QJM 2000; 93: 563 – 565. for the formation and identity of the pulmonary myocardium. 17. Wang Y, Morishima M, Zheng M, Uchino T, Mannen K, Circ Res 2007; 101: 902 – 909. Takahashi A, et al. Transcription factors Csx/Nkx2.5 and 35. Briggs LE, Takeda M, Cuadra AE, Wakimoto H, Marks MH, GATA4 distinctly regulate expression of Ca2+ channels in neo- Walker AJ, et al. Perinatal loss of Nkx2-5 results in rapid con- natal rat heart. J Mol Cell Cardiol 2007; 42: 1045 – 1053. duction and contraction defects. Circ Res 2008; 103: 580 – 590. 18. Masuda K, Takanari H, Morishima M, Ma F, Wang Y, 36. Moushmoush B, Abi-Mansour P. Alcohol and the heart: The Takahashi N, et al. Testosterone-mediated upregulation of long-term effects of alcohol on the cardiovascular system. Arch delayed rectifier potassium channel in cardiomyocytes causes Intern Med 1991; 151: 36 – 42. abbreviation of QT intervals in rats. J Physiol Sci 2018; 68: 759 – 767. 19. Burger D, Lei M, Geoghegan-Morphet N, Lu X, Xenocostas A, Supplementary Files Feng Q. Erythropoietin protects cardiomyocytes from apoptosis via up-regulation of endothelial nitric oxide synthase. Cardiovasc Please find supplementary file(s); Res 2006; 72: 51 – 59. http://dx.doi.org/10.1253/circj.CJ-20-0288

Circulation Journal Vol.84, November 2020