Deletion of FoxO1 Leads to Shortening of QRS by Increasing Na⁺ Channel Activity through Enhanced Expression of both Cardiac Na_V1.5 and β3 Subunit

Abstract

Our in vitro studies revealed that a transcription factor, Forkhead box protein O1 (FoxO1), negatively regulates the expression of Na_V1.5, a main α subunit of the cardiac Na⁺ channel, by altering the promoter activity of SCN5a in HL-1 cardiomyocytes. The in vivo role of FoxO1 in the regulation of cardiac Na_V1.5 expression remains unknown. The present study aimed to define the role of FoxO1 in the regulation of Na_V1.5 expression and cardiac Na⁺ channel activity in mouse ventricular cardiomyocytes and assess the cardiac electrophysiological phenotype of mice with cardiac FoxO1 deletion. Tamoxifen-induced and cardiac-specific FoxO1 deletion was confirmed by polymerase chain reaction (PCR). Cardiac FoxO1 deletion failed to result in either cardiac functional changes or hypertrophy as assessed by echocardiography and individual ventricular cell capacitances, respectively. Western blotting showed that FoxO1 was significantly decreased while Na_V1.5 protein level was significantly increased in mouse hearts with FoxO1 deletion. Reverse transcription-PCR (RT-PCR) revealed that FoxO1 deletion led to an increase in Na_V1.5 and Na⁺ channel subunit β3 mRNA, but not β1, 2, 4, or connexin 43. Whole patch-clamp recordings demonstrated that cardiac Na⁺ currents were significantly augmented by FoxO1 deletion without affecting the steady-state activation and inactivation, leading to accelerated depolarization of action potentials in mouse ventricular cardiomyocytes. Electrocardiogram recordings showed that the QRS complex was significantly shortened and P wave amplitude was significantly increased in conscious and unrestrained mice with cardiac FoxO1 deletion. Na_V1.5 expression was decreased in the peri-infarct (border-zone) of mice with myocardial infarction and FoxO1 accumulated in the cardiomyocyte nuclei of chronic ischemic human hearts. Our findings indicate that FoxO1 plays an important role in the regulation of Na_V1.5 and β3 subunit expression as well as Na⁺ channel activity in the heart and that FoxO1 is involved in the modulation of Na_V1.5 expression in ischemic heart disease.

1. Introduction

The voltage-gated Na⁺ channel plays an important role in the membrane excitability of cardiomyocytes [1-3]. As a major Na⁺ channel isoform, the Na_V1.5 α subunit encoded by the SCN5a gene determines the cardiac rapid depolarization characteristic of ventricular myocytes and drives subsequent electrical propagation throughout the heart [1, 3]. Heterozygous deletion of SCN5a leads to prolongation of both the PR-interval and QRS complex as well as ventricular tachycardia in mice [4]. Numerous studies have shown that the dysfunction of Na⁺ channels contributes to the development of life-threatening ventricular arrhythmias in both inherited ion channelopathies and acquired cardiac diseases [1-3, 5-8]. Mutations of SCN5a gene alter the function of the Na⁺ channel and have been linked to Brugada syndrome, long QT syndromes (LQTs), cardiac conduction defects, and atrial fibrillation [1, 5, 9-10]. Dysregulation of Na_V1.5 expression has been widely reported in myocardial infarction, heart failure, and other heart diseases [2-3, 8]. Na⁺ channel activity is finely regulated by complex molecular mechanisms [11-12]. Na⁺ channel β subunits, ankyrin-G, fibroblast growth factor homologous factor 1B, caveolin-3, E3 ubiquitin-protein ligase Nedd4, glycerol-3-phosphphate dehydrogenase 1-like protein, plakophilin-2, MOG1, a multiprotein complex composed of syntrophins and dystrophin [13] interact with Na_V1.5 to regulate Na⁺ channel activity [14-21]. Na_V1.5 is directly phosphorylated by Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) and protein kinase A, resulting in changes of Na⁺ channel activity [11]. Na_V1.5 requires ankyrin-G-binding for normal physiological function [22] and its localization at the intercalated discs is required for β (IV)-spectrin-dependent targeting of CaMKII to a critical phosphorylation site, S571, in Na_V1.5 [23]. At the transcription level, NF-KappaB and TBX5 directly affect the SCN5a promoter activity and alter Na_V1.5 expression [12, 24]. NF-KappaB functions as suppressor [12] while TBX5 functions as an activator of SCN5a gene transcription in cardiac cells [24].

Accumulating evidence indicates that Forkhead box O (FoxO) transcription factors are critical in maintaining cardiac function and mediating oxidative stress [25]. Although FoxO1, 3, and 4 are all expressed in the heart, FoxO1 is the main isoform found in the adult heart [26]. The transcriptional activities of FoxOs largely depend upon their nuclear localization [27]. Deacetylation or phosphorylation of FoxO proteins typically determines their localization in either the nuclei or cytoplasm [27]. Generally, FoxO proteins directly bind to consensus DNA sequences and regulate the expression of target genes. The proximal regions of the mouse, rat, and human SCN5a promoters have FoxO-binding insulin response elements, indicating that FoxO1 may regulate Na_V1.5 expression [27]. Indeed, our previous studies have shown that FoxO1 reduces Na_V1.5 expression in vitro by binding to the SCN5a promoter and inhibiting its activity in HL-1 cardiomyocytes [27]. In our current study, we characterized the cardiac electrophysiological phenotype of mice with tamoxifen-induced and cardiac-specific deletion of FoxO1 and defined the in vivo role of FoxO1 in the regulation of cardiac Na⁺ channel activity.

2. Methods

2.1. Polymerase chain reaction (PCR)

DNA was extracted from mouse-tails or hearts using Genomic DNA kit (QIAGEN) following the manufacturer’s instruction. Two pairs of primers (forward: 5’- CTA GGC CAC AGA ATT GAA AGA TCT -3’, reverse: 5’- GTA GGT GGA AAT TCT AGC ATC ATC C -3’; forward: 5’- GCG GTC TGG CAG TAA AAA CTA TC -3’, reverse: 5’- GTG AAA CAG CAT TGC TGT CAC TT -3’) were used to determine the presence of Cre, while two additional pairs of primers (forward: 5’- GCT TAG AGC AGA GAT GTT CTC AGA TT -3’, reverse 1: 5’- CCA GAG TCT TCG TAT CAG GCA AAT AA -3’, and reverse 2: 5’- CAA GTC CAT TAA TTC AGC ACA TTG A -3’) were employed to determine the presence of LoxP sites flanking FoxO1 exon 2 and confirm deletion of exon 2.

2.2. Reverse transcription PCR (RT-PCR)

Using RNeasy Fibrosis Tissue kit (QIAGEN) and following the manufacturer’s instruction, total RNA was extracted from adult mouse FoxO1^+/+ and FoxO1^-/- ventricles as well as the peri-infarct (border-zone) region defined as the 2-mm area encircling the area of pathologic infarction of the mouse hearts with anterior descending artery ligation or ventricular tissue from the hearts with a sham procedure. The RNA sample was then reversely transcribed to first-strand cDNA using a high capacity cDNA reverse transcription kit (AB Applied Biosystems). Semi-quantitative RT–PCR was performed to measure mRNA expression levels of Na_V1.5, Na⁺ channel β subunits, Connexin 43 (Cx43), and GAPDH in ventricular tissue. The primer sequences used to amplify the cDNA of these genes are provided in Supplemental Table 1.

2.3. Western blotting

Western blot analysis was performed on the proteins extracted from mouse ventricles with FoxO1^+/+ and FoxO1^-/-. Tissue from the peri-infarct (border-zone) region, defined as the 2-mm area encircling the area of pathologic infarction of the wild mouse hearts with anterior descending artery ligation, or ventricular tissue from the mice with the sham procedure were used as well. Ventricular tissues were homogenized in cell lysis buffer (Cell Signaling Technology, Inc). Fifteen to thirty micrograms of protein were equally loaded on 10% SDS polyacrylamide gels and transferred onto nitrocellulose membranes. The membranes were blocked with 5% non-fat milk for 1 hour, and then incubated with the indicated primary antibody overnight at 4°C. After one-hour incubation with the appropriate secondary antibody, the specific signals were revealed by enhanced chemiluminescence reagent (Pierce). Primary antibodies used in this study include rabbit anti-FoxO1 polyclonal (Cell Signaling, 1:1000 for Western blot and 1:200 for immunohistochemistry), rabbit anti-Na_V1.5 polyclonal (Alomone Labs, 1:500), rabbit anti-SCN3b polyclonal antibody (Abcam, 1:1000) and rabbit anti-GAPDH polyclonal (Millipore, 1:4000).

2.4. Whole-cell voltage clamp recording

Whole cell recording techniques were used to record Na⁺ currents in isolated ventricular myocytes from adult mice. The methods used to isolate cardiomyocytes were described previously [27]. Only Ca²⁺-tolerant ventricular cells with clear cross-striations and without spontaneous contractions or substantial granulation were selected for the experiments. All whole cell recordings were conducted at room temperature using a patch clamp amplifier (Axopatch 200B, Axon Instruments). The experiments were controlled by using pClampex 10.3 software (Axon Instruments Inc., Foster City, CA). The external solution for recording action potential (AP) contained (in mM): 136 NaCl, 4 KCl, 1 CaCl₂, 2 MgCl₂, 10 HEPES, and 10 glucose (pH 7.35 with NaOH); the pipette solution for recording AP contained (in mM): 135 KCl, 1 MgCl₂, 10 EGTA, 10 HEPES, and 5 glucose (pH 7.2 with KOH); the external solution for recording Na⁺ currents contained (in mM): 20 NaCl, 1 MgCl₂, 1 CaCl₂, 55 CsCl, 10 CsOH, 10 Glucose, 20 HEPES, 2 4-AP, 50 TEACl and 0.5 CdCl₂ (pH 7.35 with CsOH); the pipette solution contained (in mM) 5 NaCl, 135 CsF, 10 EGTA, 5 MgATP, 5 HEPES and 20 TEACl (pH 7.2 with CsOH). In all recordings, 80% of the series resistance was compensated. Data analysis was accomplished by using pClampfit 10.3 software (Axon Instruments Inc., Foster City, CA).

2.5. Immunohistochemical staining

Formalin fixed and paraffin embedded tissues from the left ventricular free wall of human hearts, including 3 non-failing controls without a history of heart diseases (NFH) and 3 failing ones with ischemic heart disease (IHD) were used for immunohistochemical staining. The research protocol was approved by Institutional Review Boards at the University of Rochester. Four μm sections were cut from paraffin-embedded ventricular tissues and deparaffinized. Antigen retrieval was performed in sodium citrate buffer (pH 6.0) by heating to 99 °C for 20 minutes with PT Link system by Dako (Carpinteria, CA). After quenching endogenous peroxidase activity with 3% H₂O₂, the sections were blocked with 10% non-immune goat serum (Invitrogen, Carlsbad, CA) and then incubated overnight with primary anti-FoxO1 antibody (1:200) at 4 °C. The signals were amplified using the Histostain-SP Kit (Invitrogen) and detected with DAB substrate (Dako). Hematoxylin was used as a counterstain. Negative controls were incubated with serum instead of primary antibody under the same conditions.

2. 6. Permanent left anterior descending coronary artery ligation surgery

Wild type mice were anesthetized with 2.0% isoflurane inhalation and placed onto a heating pad (half inch plexiglass between the animal and the heating pad). Oral tracheal intubation was performed with a PE 90 tube (o.d. 1.27 mm). Mechanical ventilation (tidal volume of 250 μl at 130 breaths per minute) was started and 1.5% isoflurane was used for maintenance of anesthesia. After intubation, a left thoracotomy was performed in the fourth intercostal space and the pericardium was vertically opened. The left anterior coronary artery (LAD) was ligated with 9-0 polypropylene sutures. After observation of ventricle blanching and ST-segment elevation in the electrocardiograms (ECG), indicating successful occlusion of the artery, the chest was closed with 6-0 coated vicryl sutures. The skin was closed using 6-0 nylon, the anesthesia was stopped, and the mouse was allowed to recover for several minutes before the endotracheal tube was removed. Sham-operated animals served as surgical controls and were subjected to the same procedures as the experimental animals, except that the LAD was not ligated.

2.7. Electrocardiogram Recording

The method used to record electrocardiograms from conscious and unrestrained mice is as described in our previous study [28]. Standard lead II surface ECGs were recorded in adult (>25 weeks) mice. Two electrodes were placed on the right anterior and superior, and left anterior and inferior chest surfaces, respectively, and were connected to a Bioamplifier (AD instrument, CO) through an electrical swivel 6 channel (Promed-Tec Inc. Bellingham, MA). ECG recordings were obtained at a frequency response to 0.05 to 500 Hz. Signals were digitized at 2 kHz and recorded on a personal computer.

2.8. Echocardiography

Echocardiography was performed in conscious adult (>25 weeks) mice using a high-resolution mouse echo machine (Visualsonics Vevo 2100 ultrasound system with an 18-38 MHz MS-400 transducer). The following parameters were measured digitally in M-mode: LV dimensions at diastole (LVDd) and systole (LVDs), and anterior and posterior wall thickness at diastole and systole. All measurements were averaged over five consecutive cardiac cycles. The observer conducting the measurement was blind to the experiments and genotypes.

2.9. Statistics

Results are presented as means ± standard error; the statistical significance of differences between groups was assessed using a two-tailed Student t-test and the significance level was set at p<0.05.

3. Results

3.1. Cardiac-specific FoxO1 deletion mouse model

α-MHC-MerCreMer transgenic mice (mhy6-MCM^+/-) (Jackson Laboratory) were bred with mice containing loxP-flanked FoxO1 exon 2 (FoxO1^L/L) (which were generated from LoxP-flanked FoxO1,3,4 mice kindly provided by Dr. Ronald A DePinho in the Department of Cancer Biology, Division of Basic Science Research at the University of Texas MD Anderson Cancer Center, Houston, Texas) to produce mhy6-MCM^+/-/FoxO1^L. mhy6-MCM^+/-/FoxO1^L/+, mhy6-MCM^+/-/FoxO1^L/L, mhy6-MCM^-/-/FoxO1^L/+, and mhy6-MCM^+/-/ FoxO1^+/+ mice were generated from the further breeding, Tamoxifen was intraperitoneally injected in these mice for 5 consecutive days at a dose of 40 mg/kg/day. All mice treated with tamoxifen injection survived the process and remained healthy. PCR was performed on DNA extracted from the hearts of the tamoxifen treated mice one month after the injections. Tamoxifen-mediated cardiac-specific deletion of FoxO1 (FoxO1^-/-) induced by Cre was confirmed by the results presented in Supplemental Fig. 1. FoxO1^-/- mice were healthy and did not exhibit any abnormal behaviors.

3.2. No cardiac hypertrophy identified in FoxO1^-/- heart

Global knockout of FoxO3 has been reported to cause mouse cardiac hypertrophy [29]. To determine if FoxO1 deletion was able to induce cardiac hypertrophy, mouse heart/weight ratios and individual ventricular cell capacitances were determined. There was no significant difference in heart/body weight ratio (mg/g) between FoxO1^+/+ and FoxO1^-/- mice (3.8 ± 0.1, n=7 versus 4.1 ± 0.1, n=6) (Fig. 1A). After establishing whole-cell recordings, the currents were recorded to determine ventricular cell capacitance by depolarizing to -60 mV and hyperpolarizing to -80 mV from a holding potential (HP) of -70 mV. The capacitances of ventricular myocytes isolated from FoxO1^-/- and FoxO1^+/+ mouse hearts were 117.1 ± 3.3 pF (n=26) and 133.2 ± 9.9 pF (n=18), respectively, and they were not statistically significantly different (Fig. 1B). These findings indicate that cardiac-specific FoxO1 deletion does not result in cardiac hypertrophy in mice.

Fig. 1. No cardiac hypertrophy and dysfunction were observed in FoxO1^-/- mouse hearts.

(A) Heart/body weight ratio was not significantly different between FoxO1^+/+ and FoxO1^-/- mice. (B) Capacitances of ventricular myocytes were not significantly changed by cardiac deletion of FoxO1. (C) Echocardiographic M-mode images of FoxO1^+/+ and FoxO1^-/- left ventricles. Cardiac function was not altered by FoxO1 deletion (see parameters in Supplemental table 2).

3.3. Heart function was not altered by FoxO1 deletion

To clarify the influence of FoxO1 deletion on mouse cardiac functions, echocardiography was performed. Representative M-mode images are shown in Fig. 1C. There were no significant differences in heart function parameters such as ejection fraction (EF) and fraction shortening (FS) between the FoxO1^-/-and FoxO1^+/+ mice (Supplemental table 2). Furthermore, other echocardiography parameters such as LVPWs, LVPWd, LVIDs, and LVIDd did not significantly differ between FoxO1^-/- and FoxO1^+/+ mice (Supplemental table 2). These findings indicate that FoxO1 does not significantly affect the cardiac functions of nor cause hypertrophy in adult mice.

3.4. Deletion of FoxO1 induced an increase in both cardiac Na_V1.5 and β3 subunit expression

FoxO1 negatively regulates Na_V1.5 expression in HL-1 cardiomyocytes[27]. To test the hypothesis that FoxO1 functions as a negative regulator of Na_V1.5 expression in mouse hearts, Western blot was performed on the protein extracted from FoxO1^+/+ and FoxO1^-/- mouse ventricles. As evidenced by Western blots (Fig. 2A) and quantitative analysis (Fig. 2B), Na_V1.5 and β3 protein levels were significantly (p<0.01, p<0.05) increased and FoxO1 was significantly (p<0.01) decreased in FoxO1^-/- hearts in comparison with FoxO1^+/+ hearts. Densitometry analyses of FoxO1, Na_V1.5, and β3 levels compared to GAPDH level showed that Na_V1.5 and β3 expression were increased 162% and 21% while FoxO1 expression was decreased 84% in FoxO1^-/-, compared to FoxO1^+/+ hearts (Fig. 2B). These findings indicate that FoxO1 is a suppressor of both Na_V1.5 and β3 expression in adult mouse hearts.

Fig. 2. FoxO1^-/- induced increases of cardiac Na_V1.5 protein and mRNA levels and β3 mRNA level.

(A) Western blot showed that Na_V1.5 protein was more expressed while FoxO1 was less expressed in FoxO1^-/- mouse ventricles as compared with FoxO1^+/+ control group. (B) Quantitative analyses of Na_V1.5, β3 and FoxO1 relative to GAPDH revealed that both Na_V1.5 and β3 were significantly increased (p<0.01 or p<0.05) increased and FoxO1 was significantly (p<0.01) decreased in FoxO1^-/- in comparison with FoxO1^+/+ ventricles (n=3 for each group). (C) RT-PCR was performed on the RNA extracts of mouse ventricles and the results showed that mRNA levels of Na_V1.5 and β3 subunits but not β1, β2 and β4, and Cx43 were increased in FoxO1^-/- ventricles compared with FoxO1^+/+ group. Quantitative analyses of Na_V1.5, β subunits, and Cx43 mRNA levels relative to GAPDH revealed that Na_V1.5 and β3 subunit were significantly increased (p<0.01).

3.5. mRNA levels of Na_V1.5 and β3 subunit were increased in FoxO1^-/- mouse hearts

The QRS complex represents ventricular depolarization. Na⁺ channel and Cx43 channel activities determine cardiac depolarization. Na⁺ channel activity is primarily determined by a main α subunit, Na_V1.5; however, multiple β subunits also play roles in the regulation of this channel activity. Increased Na⁺ channel activity due to elevated expression of Na_V1.5 protein was found in FoxO1^-/- mouse hearts and this contributed to the shortening of the QRS complex. Other components related to the electrical conduction between ventricular myocytes, such as β1, 2, 3, and 4 subunits as well as Cx43 were assessed in these cardiac FoxO1 deficient mice. RT-PCR analysis showed that the level of Na_V1.5 was significantly (p<0.01) increased (Fig. 2C). Interestingly, mRNA level of the β3 subunit was also significantly (p<0.01) increased, but β1, 2, 4, and Cx43 levels were not changed in mouse hearts with deletion of FoxO1 in comparison with the control group (Fig. 2C and D).

3.6. Increase of I_Na in FoxO1^-/- mouse ventricular myocytes

To determine if FoxO1 deletion and increased Na_V1.5 and β3 subunit expression were capable of inhibiting cardiac Na⁺ channel activity, Na⁺ currents were recorded from FoxO1^+/+ and FoxO1^-/- left ventricular myocytes isolated from adult mice using the whole-cell voltage-clamp techniques. Typical traces of I_Na recorded from FoxO1^+/+ and FoxO1^-/- ventricular cells are displayed in Fig. 3A. Peak Na⁺ current densities were obtained from the peak currents divided by individual cardiomyocyte capacitances. The average current densities, from test potentials of -35 mV to -5 mV, were significantly larger (p<0.05) in FoxO1^-/- ventricular myocytes compared with FoxO1^+/+ ventricular cells (Fig. 3B). At the test potential of -30 mV, the peak current density was 26.9 ± 1.4 pA/pF (n=26) for FoxO1^-/- and 21.3 ± 1.8 pA/pF (n=18) for FoxO1^+/+. There was no significant shift of the test potential for the peak Na⁺ current density between the two groups (Fig. 3B).

Fig. 3. I_Na was increased in FoxO1^-/- mouse ventricular myocytes.

(A) Typical whole cell Na⁺ currents were recorded during 80 ms depolarizing voltage steps to potentials between -90 and +15 mV from a holding potential of −100 mV in left ventricular myocytes isolated from FoxO1^+/+ and FoxO1^-/- mouse hearts (n=18 and n=26, respectively). (B) The current-voltage relationship curve of Na⁺ currents showed that the current densities obtained from the peak currents divided by individual cell capacitances were significantly increased (p<0.05) from -35 mV to -5 mV in FoxO1^-/- as compared with FoxO1^+/+ ventricular myocytes. No voltage-current relationship shift was observed between these two groups.

3.7. Inactivation but not activation of Na⁺ channel was accelerated by FoxO1 deletion in ventricular myocytes

Time constants of activation (τ_activation) and inactivation (τ_inactivation) were analyzed by fitting the upstroke and decay traces of I_Na using a mono-exponential function, respectively. As shown in Fig. 4B, there was no statistical difference in the τ_activation of I_Na between FoxO1^-/- and FoxO1^+/+ mouse ventricular myocytes, with average τ_activation values of 0.50 ± 0.05 for FoxO1^-/- and 0.53 ± 0.04 for FoxO1^+/+ ventricular cells at -30 mV. (Fig. 4B). The τ_inactivation values, however, were significantly decreased (p<0.05) at both -30 mV and -20 mV in FoxO1^-/- ventricular myocytes compared to FoxO1^+/+ (Fig. 4C), indicating that deletion of FoxO1 accelerated the Na⁺ channel inactivation. The values of τ_inactivation for the Na⁺ channel, for example, at -30 mV in ventricular myocytes from FoxO1^-/- and FoxO1^+/+ mice were 1.91 ± 0.15 ms and 2.35 ± 0.16 ms, respectively.

Fig. 4. Faster inactivation of Na⁺ channel in FoxO1^-/- ventricular myocyets.

(A) Typical Na⁺ current traces recorded from FoxO1^+/+ and FoxO1^-/- ventricular myocytes at -30 mV from a holding potential of -100 mV. A monoexponential function was used to fit the activation and inactivation phases of I_Na from -40 mV to -10 mV and the activation and inactivation time constants were obtained. They were plotted with the different voltage levels. (B) τ_activation were not significantly different between FoxO1^+/+ and FoxO1^-/- ventricular myocytes from -40 mV to -10 mV. (C) τ_inactivation of decay traces of I_Na were significantly decreased (p<0.01 or p<0.05) in FoxO1^-/- in comparison with FoxO1^+/+ ventricular myocytes.

3.8. Recovery from the inactivation but not steady-state activation and inactivation of Na⁺ channel was affected by FoxO1 deletion in ventricular myocytes

In order to determine whether there were any differences in Na⁺ channel kinetics between FoxO1^+/+ and FoxO1^-/- ventricular myocytes, the steady-state activation, steady-state inactivation, and recovery from inactivation of I_Na were recorded and analyzed. The voltage-dependent steady-state activation of the Na⁺ channel was determined by a series of test potentials ranging from -90 to +60 mV at 0.1 Hz, from a holding potential of -100 mV (Fig. 5A). These results showed that the voltage-dependent activation of I_Na was unchanged between FoxO1^+/+ and FoxO1^-/- ventricular myocytes. The half-maximal activation voltage and slope factor were -45.03 ± 0.37 mV and 6.52 ± 0.34 in FoxO1^-/- ventricular myocytes (n=26), and -41.24 ± 0.27 mV and 5.32 ± 0.24 in control group (n=18), respectively, and they were not statistically significantly different. The voltage-dependent inactivation of I_Na was recorded by a two-pulse protocol with conditioning potentials from -130 mV to -10 mV, followed by a test potential of -30 mV (Fig. 5B). Similarly, there were no changes in the voltage-dependent inactivation curves of I_Na between the two groups. The half-maximal inactivation voltage and slope factor did not show a significant difference (-70.93 ± 0.37 mV and 7.20 ± 0.33, n=18 in FoxO1^-/- versus -71.60 ± 0.61 mV and 7.60 ± 0.53, n=16 in FoxO1^+/+ ventricular myocytes). Recovery of I_Na from inactivation was recorded by a paired-pulse with a variable inter-pulse duration (from 0 to 230 ms) at holding potential of -100 mV. The curves of recovery from inactivation were obtained by normalizing the peak current from a test pulse (-30 mV, 80 ms) to a conditioning pulse (-30 mV, 300 ms) and then fitting the data with the mono-exponential equation. These results (Fig. 5C) showed that the recovery of I_Na from inactivation was significantly different between FoxO1^-/- and FoxO1^+/+ ventricular myocytes (Tau=18.3 and 13.3; K slope factor = 0.075 ± 0.001and 0.055 ± 0.000 in FoxO1^-/- versus in FoxO1^+/+ mice; n=16 for each group; p<0.05). These findings indicate that the deletion of FoxO1 leads to accelerated Na⁺ channel recovery from inactivation.

Fig. 5. FoxO1^-/- did not affect the voltage-dependent and steady-state activation and inactivation of Na⁺ channel but accelerated this channel recovery from inactivation.

(A, B) Voltage-dependent and steady-state activation and inactivation of Na⁺ channel were not significantly different between FoxO1^+/+ and FoxO1^-/- ventricular myocytes. (C) Representative I_Na recovery traces were recorded in left ventricular myocytes by using the stimulation protocol shown in the inset from FoxO1^+/+ and FoxO1^-/- ventricular myocytes, respectively. Mean ± SEM normalized recovery data for peak I_Na were plotted and they are well described by a single exponential.

3.9. Deletion of FoxO1 accelerated the depolarization of action potentials in ventricular myocytes

We further investigated whether FoxO1 deletion influenced the ventricular myocyte action potential. Action potentials were recorded at 1 Hz from FoxO1^-/- and FoxO1^+/+ ventricular myocytes. Typical action potential traces, as shown in Fig. 6A, exhibited that the depolarization of action potentials was faster in FoxO1^-/- than FoxO1^+/+ ventricular myocytes (Fig. 6A). The values of maximal upstroke velocity (V_max) of the action potentials recorded from FoxO1^-/- and FoxO1^+/+ ventricular myocytes (Fig. 6B) were 240 ± 13 mV/ms (n=18) and 147 ± 22 mV/ms (n=12), respectively, and were significantly different (p<0.05). The amplitude of action potentials was larger in FoxO1^-/- ventricular myocytes, but it was not significantly different from FoxO1^+/+ (145 ± 4 mV, n=18 versus 127 ± 8 mV, n=12), and the resting membrane potential was also not statistically significant between FoxO1^-/- (-86.8 ± 3.1 mV, n = 18) and FoxO1^+/+ (-89.0 ± 3.8 mV, n=12). Both action potential duration (APD)₅₀ and APD₉₀ were not significantly different (Fig. 6B) between FoxO1^+/+ and FoxO1^-/- ventricular myocytes (APD₅₀: 3.7 ± 0.6 ms and 5.3 ± 0.9 ms; APD₉₀: 13.8 ± 1.6 ms and 16.0 ± 1.9 ms; n=18 and 12 in FoxO1^-/- and FoxO1^+/+ ventricular myocytes, respectively). These findings suggest that FoxO1 deletion induces an increase in Na_V1.5 expression and Na⁺ channel activity, which leads to faster depolarization of action potentials.

Fig. 6. Increased depolarization in FoxO1^-/- ventricular myocytes.

(A) Representative action potentials were recorded from FoxO1^+/+ and FoxO1^-/- left ventricular myocytes, respectively and the rising phase was faster in FoxO1^-/- cardiomyocytes as compared with FoxO1^+/+ cells. (B) The amplitude of action potentials, resting membrane potential and duration of action potentials (APD50 and APD 90) were not significantly different but the maximum upstroke velocity of the action potential (V_max) was significantly increased (p<0.05) in FoxO1^-/- than in FoxO1^+/+ ventricular myocytes.

3.10. QRS shortening in mice with cardiac deletion of FoxO1

Electrocardiogram (ECG) recording was performed on conscious and unrestrained mice (Fig. 7A). No arrhythmias were observed in the mice with cardiac deletion of FoxO1 or controls without this deletion. Although the PR-, QTc-, and RR-intervals were not prolonged in the mice with cardiac deletion of FoxO1, the QRS complex and P wave were markedly different from the FoxO1^+/+ mice. The QRS complex was significantly shortened (p<0.05) to 9.2 ± 0.3 ms (n=5) in mice with cardiac deletion of FoxO1 from 10.2 ± 0.3 ms (n=5) in the control group (Fig. 7B); the P wave amplitude was significantly (p<0.05) increased to 0.17 ± 0.01 mV (n=5) in mice with cardiac deletion of FoxO1 from 0.13 ± 0.01 mV (n=5) in the control group.

Fig. 7. FoxO1^-/- led to shortening of QRS complex and increasing of P wave amplitude.

Representative surface ECGs were recorded from conscious and unrestraint mice with or without cardiac-specific deletion of FoxO1 in Lead II. (A) QRS complex was shortened and P wave amplitude was increased in FoxO1^-/- mice. (B) Surface ECG parameters, heart rate, QRS duration, P wave amplitude, QTc-, PR-, and RR-intervals were analyzed. The data showed that QRS duration was significantly (p<0.05) decreased and P wave amplitude was significantly (p<0.05) increased in FoxO1^-/- mice, compared with those in FoxO1^+/+ mice but heart rate, QTc-, PR-, and RR- intervals were not changed.

3.11. Inverse relationship of nuclear FoxO1 to Na_V1.5 expression in ischemic hearts confirmed that Na_V1.5 is in vivo FoxO1 target during oxidative stress

In order to investigate the potential role of FoxO1 in the regulation of Na_V1.5 expression in the IHD, immunohistochemical staining was performed on ventricular tissue from human hearts with IHD to show more nuclear FoxO1 in cardiomyocytes from human IHD (n=3) than that from NFH (n=3) (Fig. 8A). Immunohistochemical staining was performed on sections taken from mouse hearts and no FoxO1 signaling was detected in our hands. RT-PCR was performed on the RNA extracts, and Western blot was done on the protein extracts from the peri-infarct (border zone) region of the mouse hearts (which was defined as the 2-mm area encircling the area of pathologic infarction) and ventricular tissue from the sham-control group. Both Na_V1.5 protein and mRNA levels were significantly (p<0.05 or p<0.01) decreased in the peri-infarct region (n=3), compared to those in the ventricular tissue from the control group (n=3) (Fig. 8B).

Fig. 8. Inverse relationship of nuclear FoxO1 to Na_V1.5 expression in ischemic hearts.

(A) Immunohistochemical staining revealed that FoxO1 accumulated in the cardiomyocyte nuclei in the human failing hearts with ischemic heart disease (IHD) (n=3) in comparison with non-failing controls without a history of heart diseases (NFH) (n=3). (B) Na_V1.5 expression was decreased in the peri-infarct (border-zone) of mouse hearts with myocardial infarction. (a) RT-PCR analysis showed that mRNA level of Na_V1.5 was significantly (p<0.01) decreased; (b) Western blot exhibited that protein level of Na_V1.5 was significantly (p<0.05) downregulated as well.

4. Discussion

This study uncovered that cardiac-specific deletion of FoxO1 in adult mice enhances voltage-gated Na⁺ channel activity by increasing Na_V1.5 and β3 subunit expression, increases the speed of Na⁺ channel inactivation, and increases the recovery of Na⁺ channel from inactivation. This increase of Na⁺ channel activity leads to an acceleration of cardiac depolarization as evidenced by the larger V_max of action potentials recorded from FoxO1^-/- ventricular myocytes as well as the shortening of the QRS complex and the increase in P wave amplitude in surface ECGs recorded from conscious and unrestrained mice with cardiac deletion of FoxO1. Cardiac function was not changed by FoxO1 deletion. These in vivo findings support the conclusions from our previous in vitro studies that FoxO1 negatively regulates Na⁺ channel activity by altering Na_V1.5 expression in HL-1 cardiomyocytes[27]. Furthermore, these studies showed that FoxO1 binds to the insulin response elements in the promoter of SCN5a and suppressed the promoter’s activity, leading to the inhibition of Na_V1.5 expression [27]. In addition, our studies showed that accumulation of FoxO1 in the cardiomyocyte nuclei of human IHD and decrease of Na_V1.5 expression in the peri-infarct (border-zone) of mouse hearts with myocardial infarction. This further indicates that Na_V1.5 is in vivo FoxO1 target during oxidative stress.

It has been reported that cardiac-specific overexpression of the SCN5a gene did not significantly increase the surface density of the Na⁺ channels in ventricular myocytes, indicating that Na_V1.5 localization to the membrane and formation of functional channels are limited [30]. Na⁺ currents were increased in FoxO1^-/- ventricular myocytes, but our finding that FoxO1 deletion led to a significant increase in Na_V1.5 expression may not be able to explain the enhancement of the Na⁺ channel activity. It has been shown that the β3 subunit can increase Na_V1.5 membrane expression and alter the channel kinetics in oocytes [31]. The subsequent promotion of Na⁺ channel β3 subunit expression following FoxO1 deletion contributes to an increase in the Na_V1.5 surface expression and densities of the functional cardiac Na⁺ channel. The findings are most consistent with the decrease in I_Na observed in cardiomyocytes of the β3 knockout mouse [32]. The acceleration of inactivation and the recovery from inactivation of the Na⁺ channel is likely due to increased β3 subunit expression, although steady-state activation and inactivation of this channel were not changed in FoxO1^-/-[31]. β3 regulating Na_V1.5 may have similar mechanisms by which β3 increases both expression of the core-glycosylated form of Na_V1.7 and Na⁺ channel activity [33].

Na_V1.5, encoded by the SCN5a gene, is the main α subunit of the cardiac Na⁺ channel and its activity determines cardiac excitability and electrical conduction. The importance of the cardiac Na⁺ channel for normal cardiac electrical activity is emphasized by the occurrence of potentially lethal arrhythmias in the setting of inherited and acquired functional defects. Other extensive studies have been more focused on a large number of protein relationships, including β subunits, ankyrin-G, and MOGl interactions with Na_V1.5 [34] or concerned with post-translational modifications, such as the effects of phosphorylation, glycosylation, S-nitrosylation, ubiquitination, and methylation on the function of the Na_V1.5 channel [35]. Knowledge of the mechanisms behind the regulation of Na_V1.5 expression at the transcriptional level is largely limited. Currently, it has been shown that four transcriptional NF-KappaB and TBX5 directly affect the SCN5a promoter activity and alter Na_V1.5 expression [12, 24]. In vivo studies showed that TBX5 is only expressed in the mouse cardiac conducting cells and that specific deletion of TBX5 in these cells led to a severe reduction in cardiac conduction [24]. Accumulating evidence suggests that FoxO1 is the main isoform among FoxOs in the heart [26] and that it plays an important role in the regulation of cardiac function under different stresses [36]. It typically functions as a transcription activator and binds to the DNA promoter consensus sequences of target genes to upregulate their expression in response to cardiac stresses [36]. Both our in vitro [27] and in vivo studies are the first to show that FoxO1 negatively regulates Na_V1.5 expression. These findings strongly suggest that FoxO1 functions as a suppressor of SCN5a gene transcription. The mechanism of β3 subunit regulation by FoxO1 is unclear. FoxO proteins target a conserved DNA binding sequence, 5′-GTAAA(C/T)A-3′, or insulin response element (IRE), 5′-CAAAA(C/T)A-3′ in the promoters of their target genes to affect these gene expressions. Analysis of the β3 promoter revealed that there are IREs, CAAAA(C/T)A, in mice and humans. It is possible that FoxO1 regulates β3 expression by affecting the promoter activity. Additional studies including promoter activity and chromatin immunoprecipitation assays will be pursued to reach a definite conclusion in the future.

FoxO1 is a transcriptional factor and its nuclear localization is important for its transcriptional activity. The nuclear localization is determined by the status of FoxO1 phosphorylation [36] and acetylation [37]. Any physiological and pathological event may alter the FoxO1 nuclear localization, subsequently regulating Na_V1.5 expression. Previous studies showed that FoxO proteins were phosphorylated in thoracic aortic banding, induced cardiac hypertrophy, and by several hypertrophic ligands, and their transcriptional activity was changed through a mechanism involving the PI3K/Akt pathway [29]. Consistent with these findings, we previously reported that increased Na⁺ current density was associated with higher Na_V1.5 protein levels and increased FoxO1 phosphorylation in mouse left ventricular hypertrophy and failure induced by transverse aortic constriction. Akt has been reported to phosphorylate FoxO1 at threonine 24, serine 256 and serine 319 and makes FoxO1 less localized in the nuclei [38]. This process potentially involves less FoxO1’s suppressive effects on Na_V1.5 expression in mouse cardiac hypertrophy and failure [39].

Epidemiologic data indicate that ischemic heart disease (IHD), a leading cause of death worldwide, is responsible for 80% of fatal arrhythmias and includes conditions such as ventricular tachycardia and ventricular fibrillation [40-43]. Current therapies for cardiac arrhythmias include pharmacologic intervention, A-V node ablation, and electronic devices, which target ion channels with varying degrees of success and sometimes with severe side effects. It is crucial to identify the signaling pathways involved in cardiac ion channel regulations for IHD. Our studies revealed that FoxO1 is significantly increased in the nuclei of human cardiomyocytes during chronic ischemic cardiomyopathy and that both mRNA and protein levels of Na_V1.5 are significantly decreased in the peri-infarct (border-zone) of mouse ventricles. Elevated levels of reactive oxygen species (ROS) in IHD [44-46] induced the nuclear localization of FoxO1 in both HL-1 [27] and rat neonatal cardiomyocytes [26] potentially through acetylation [37] and phosphorylation at serine 212 by mammalian sterile 20-like kinase 1 [47] and suppresses Na_V1.5 expression through FoxO1 mediation of the SCN5a promoter activity [27]. Our findings indicate that FoxO1 plays an important role in the regulation of Na_V1.5 expression in IHD.

Research highlights.

• ➡Cardiac deletion of FoxO1 increased Na_V1.5 and β3 subunit expression.
• ➡Cardiac deletion of FoxO1 enhanced Na⁺ channel activity.
• ➡Cardiac deletion of FoxO1 accelerated the cardiac depolarization.
• ➡Cardiac deletion of FoxO1 led to the shortening of QRS wave and increasing of P wave amplitude.
• ➡FoxO1 plays an important role in the regulation of Na⁺ channel expression in both physiological and pathological conditions.

Abbreviation

CaMKII

Ca²⁺/calmodulin-dependent protein kinase II

LQTs

Long QT syndromes

LVPWs

Systolic left ventricular posterior wall thickness

LVPWd

Diastolic left ventricular posterior wall thickness

LVIDs

Systolic left ventricular internal diameter

LVIDd

Diastolic left ventricular internal diameter

EF

Ejection fraction

FS

Fraction shortening

APD

Action potential duration

IHD

Ischemic heart disease

ROS

Reactive oxygen species