Alk7 Depleted Mice Exhibit Prolonged Cardiac Repolarization and Are Predisposed to Ventricular Arrhythmia

We aimed to investigate the role of activin receptor-like kinase (ALK7) in regulating cardiac electrophysiology. Here, we showed that Alk7-/- mice exhibited prolonged QT intervals in telemetry ECG recordings. Furthermore, Langendorff-perfused Alk7-/- hearts had significantly longer action potential duration (APD) and greater incidence of ventricular arrhythmia (AV) induced by burst pacing. Using whole-cell patch clamp, we found that the densities of repolarizing K+ currents Ito and IK1 were profoundly reduced in Alk7-/- ventricular cardiomyocytes. Mechanistically, the expression of Kv4.2 (a major subunit of Ito carrying channel) and KCHIP2 (a key accessory subunit of Ito carrying channel), was markedly decreased in Alk7-/- hearts. These findings suggest that endogenous expression of ALK7 is necessary to maintain repolarizing K+ currents in ventricular cardiomyocytes, and finally prevent action potential prolongation and ventricular arrhythmia.


Introduction
Arrhythmia is one of the most serious types of cardiovascular disorders, and remains an important cause of mortality and morbidity. Dysregulation of repolarizing potassium current in cardiomyocytes leads to ventricular arrhythmia in humans and animal models [1][2][3][4][5][6], while the upstream signals that regulate repolarizing potassium currents are poorly understood.
The transforming growth factor beta (TGF-β) superfamily is a large family of structurally related regulatory proteins that function via binding to a distinct group of type I and type II serine/threonine kinase receptors [7]. Binding of TGF-β superfamily members to their type II receptors results in phosphorylation of type I receptors, which subsequently initiates an intracellular downstream signaling cascade in either Smad-dependent or -independent manner [8]. The activin receptor-like kinase 7 (ALK7) is one of the TGF-β type I receptors, and several tissue-specific ALK7 ligands have been reported, including growth differentiation factors (GDFs), nodal, activin B and activin AB [9][10][11][12].
The functions of ALK7 have been investigated in a variety of cell across different species. For example, ALK7 negatively regulates glucose-stimulated insulin release in pancreatic β-cell [13] and induces β-cells apoptosis [14]. ALK7 is also abundantly expressed in adipose tissue [15], and endogenous ALK7 is necessary for adipocyte differentiation [16]. ALK7 deleted mice develop age-dependent metabolic syndromes, including progressive hyperinsulinemia, reduced insulin sensitivity, liver steatosis, impaired glucose tolerance and pancreatic island enlargement [13]. Interestingly, Alk7 -/mice are partially resistant to diet-induced obesity and display reduced fat accumulation [10]. In the cardiovascular system, Alk7 gene polymorphism has been linked to cardiovascular remodeling in humans [17]. Studies from our group have recently demonstrated that endogenous ALK7 protects against pathological cardiac hypertrophy in mice, suggesting an essential role of ALK7 in cardiovascular remodeling [18].
Structural cardiac remodeling is commonly associated with the disruption of normal cardiac electrical signal and electrophysiology [19,20]. Given that ALK7 regulates cardiac structural remodeling in the animal model, we hypothesized that ALK7 might also play a role in the modulation of cardiac electrophysiology. In this study, we sought to investigate the role of ALK7 in cardiac electrophysiology remodeling and explore the underlying molecular mechanisms.

Experimental animals
All animal protocols were implemented under the NIH Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996). All the procedures were approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University. The experimental animals were 10-12 weeks old male Alk7 -/mice and littermate controls in C57BL/6 background. Mice were provided with food and water, and were maintained in a standard 12-hour light and dark cycle in temperature-and humidity-controlled housing. The physical condition of the animals was monitored once weekly by veterinarians at the institution. Animals were well in general, and no animals died prior to the experimental endpoint. Mice were anesthetized using pentobarbital sodium (60mg/kg, i.p.) to minimize potential suffering during the experiment. The genotyping of Alk7 -/vs. littermate control mice was assessed by PCR [18].

Echocardiography and histological analysis
Echocardiography was performed by Sonos 5500 ultrasound (Philips) with a 15-MHz linear array ultrasound transducer. The LV was assessed in both parasternal long-axis and short-axis views at a frame rate of 120 Hz. End-systole or end-diastole was defined as the phase with the smallest or largest LV area, respectively [21,22]. LVEDD, LVESD, LVEF and FS were measured from the LV M-mode tracing with a sweep speed of 50 mm/s at the midpapillary muscle level [23]. Hearts were excised, washed with saline solution, fixed with 10% formalin and then cut transversely close to the apex to visualize the left and right ventricles. Serial heart sections (4-5 μm thick) were prepared and stained with Masson trichrome stain for collagen deposition and then visualized by light microscopy [24]. The percentage of light blue stain (collagen) was measured with an Image pro-plus 6.0 software.

Telemetry ECG recording
Mice were anesthetized using pentobarbital sodium (60mg/kg, i.p.), and were then positioned in an incubator to maintain body temperature at 37°C. Leads were tunneled subcutaneously to the right shoulder and left apex (Lead II). Recording began 24 hours after mice recovered from anesthesia. Lead II ECG characteristics were recorded from conscious, freely moving mice under basal conditions in each animal during 24 hours continuously. The ECG amplifier module consisted of high pass filters (set to 0.05Hz), low pass filters (set to 1 kHz) and a gain selection device (set to1000-fold). Signals were digitized at 1 kHz and continuously recorded using a data acquisition system (DSI, US), and analyzed by P3 software [25].

MAP recording and burst pacing
Monophasic action potential (MAP) was recorded from the left and right ventricular epicardium using a custom-made electrode, a string of two 0.25 mm Teflon-coated silver wires (99.99% in purity) galvanically coupled to chloride to eliminate DC offset. The paired platinum stimulating electrode was positioned on the basal surface of the right ventricle, and delivered regular pacing. Signals were amplified using an amplifier (AD Instruments, Australia) and band pass filtered between 0.3 Hz and 1 kHz. MAP waveforms were analyzed using Lab Chrat7.0 software. To induce arrhythmia, burst pacing (2 ms pulses at 50 Hz, 2 s burst duration) was applied for 3 minutes in both ventricular locations to determine susceptibility to ventricular arrhythmia [25].
K + currents recordings: For transient outward current (I to ), the current-voltage relationship was elicited by a series of 500ms test potentials (-60mV~+60mV, with 10mV increments) and a holding potential of -80mV at a frequency of 0.1 Hz. We applied inactivating prepulse (100ms, -40mV) before each test potential to avoid contamination from rapid sodium currents. 0.3mmol/L CdCl was added to the perfusion solution to avoid contamination from calcium currents. For inward rectified potassium currents (I K1 ), the current-voltage relationship was established using 350ms steady-state depolarizing pulses from -120 mV to -40 mV with 10mV increments and a holding potential of -80mV at a frequency of 0.1 Hz. I K1 current amplitude (pA) was calculated after subtracting the background currents in the presence of 100 μM BaCl 2 . Current densities (pA/pF) were obtained by normalizing the current amplitudes (pA) to membrane capacity C m (pF).
Steady-state inactivation for I to : The two step voltage-clamp protocol was used to assess steady-state inactivation of I to . Each inactivating prepulse (-110mV~-10mV, with 10mV increments) for 1000ms, was followed by a fixed test pulse (+30mV) for 1000ms. I to currents recorded at each test potential were normalized to the maximal current recorded (I/I max ). I/ I max was plotted for each inactivating prepulse voltage, and data were best fitted to the Boltzmann curve.
Recovery from inactivation for I to : an inactivating pulse (P1 = -80mV) with various time intervals (10ms-500ms) was followed by a 500ms test pulse (P2 = +30 mV). The ratio of the current amplitude at test pulse P2 to the current amplitude at inactivating pulse P1 was plotted as a function of time intervals. Time constant was calculated by data fitted to exponential functions [25].
Action potential and resting potential recording: resting potential and action potentials in isolated ventricular myocytes were measured using the whole-cell current-clamp mode [27].

Quantitative real-time RT-PCR and Western blotting
Total RNA was extracted from frozen, pulverized mouse cardiac tissue using TRIZol (Invitrogen), and cDNA was synthesized using oligo (dT) primers with the Advantage RT-for-PCR kit (BD Biosciences) [28]. The relative amount of cDNA was measured by real-time PCR using SYBR Green PCR Master Mix (Applied Biosystems), and then normalized to that of the housekeeping gene Gapdh. Cardiac tissue was lysed in RIPA lysis buffer. Fifty micrograms of cell lysate were used for SDS/PAGE [29], and proteins were then transferred to an Immobilon-P membrane (Millipore). Membranes were then blotted with indicated primary antibodies (all purchased from Sigma-Aldrich) followed by fluorescent labeled secondary antibody (purchased from Li-Cor) [30]. The blots were scanned with the infrared Li-Cor scanner, and quantification of Western blots was performed using an Odyssey infrared imaging system (Li-Cor Biosciences) [23].

Statistical analysis
Results are presented as mean±SEM [31]. Comparisons between two groups were performed by unpaired Student t test or nonparametric Mann-Whitney U test using SPSS 16.0. Cross Tabulation (Chi-Square Test) with Fischer's exact test was used for categorical variables. P<0.05 was considered statistically significant. Patch clamp data were analyzed using Origin 9.0 for nonlinear curve fitting.

Results
Normal baseline echocardiogram and cardiac histology in Alk7 -/mice At baseline, Alk7 -/and control mice had comparable heart rates and cardiac functions assessed by echocardiograpy (S1A Fig, S1 Table). The heart-to-body weight ratio was also similar between the two groups, indicating no gross cardiac hypertrophy in Alk7 -/mice (P>0.05) (S1B Fig). In addition, Alk7 -/hearts did not show any cellular morphologic abnormality or enhanced fibrosis (S1C Fig). These observations are consistent with our previous findings that Alk7 -/mice exhibited normal heart function and did not develop spontaneous hypertrophic cardiomyopathy at baseline [18].
Prolonged QT interval in telemetry ECG recording in Alk7 -/mice To explore the function of ALK7 in cardiac electrophysiology, we utilized implantable telemetry ECG to record spontaneous arrhythmia in Alk7 -/and control mice in vivo. During the 24-hour telemetry recording, we captured spontaneous sinus arrest, atrioventricular block and escape rhythm in two out of eight (25%) Alk7 -/mice, whereas none of the control animals displayed spontaneous arrhythmia (Fig 1). One out of eight (12.5%) Alk7 -/mice showed sustained premature ventricular contractions (PVC >15 events/min). Notably, QT interval and corrected QT interval (QTc) were significantly prolonged in Alk7 -/mice by 50% (P<0.05), while the RR, PR or QRS duration were comparable between the two groups (Table 1). Taken together, Alk7 -/mice exhibited significantly prolonged QT interval in telemetry ECG recording in vivo.
Prolonged action potential duration (APD) and increased susceptibility to arrhythmia in Langendorff-perfused Alk7 -/hearts We further utilized the Langendorff-perfused heart system to characterize action potential duration (APD) and ventricular repolarization in two groups of mice. Langendorff-perfused hearts were paced at 125 ms cycle length, and action potential duration (APD) at different percentages of repolarization (APD30, 50, 70 and 90) were quantified in Table 2. We observed that the late-phase repolarization duration (APD70 and APD90) were markedly prolonged in the left ventricles of Alk7 -/hearts (Fig 2A and Table 2). Because prolonged APD enhanced susceptibility to arrhythmia, we predicted that the incidence of arrhythmia induced by burst pacing would be increased in Langendorff-perfused Alk7 -/hearts. Indeed, the overall incidence of ventricular arrhythmia (VA) was markedly increased in Alk7 -/hearts (52.0% in Alk7 -/vs. 19.2% in controls) ( Fig 2B and Table 2.). Remarkably, burst pacing-triggered sustained VA (duration > 30s) was recorded in five out of twenty-five (20%) Alk7 -/hearts but not in any control hearts (Table 3).
Inwardly rectified K + current (I K1 ) contributes significant repolarizing current during the terminal phase of repolarization. Reduction of I K1 is sufficient to prolong the cardiac action potential and QT interval in humans [1,35]. We observed that the density of inwardly I K1 current was significantly attenuated at voltages ranging from -70 mV to -120 mV (n = 8/group; P<0.01) (Fig 3F). I K1 also serves as the primary conductance controlling the diastolic resting membrane potential (Em) in ventricular myocytes [36]. We therefore measured resting Em in two groups of isolated ventricular myocytes and found that the resting Em were comparable (83±9.1 mV for control and 81±7.3 mV for Alk7 -/-, n = 5/group, P > 0.05) (S2 Fig). Collectively, Alk7 -/cardiomyocytes exhibited a significant reduction in repolarizing I to and I K1 , which could be responsible for the prolonged APD in Alk7 -/hearts.  Altered expression of voltage-dependent K + channel subunits in Alk7-/hearts We further explored the molecular mechanisms by profiling ion channel gene expression in Alk7 -/and control hearts. mRNA levels of Nav1.5, Cav 1.2 and Cav1.3, the subunits of sodium and calcium channels, were similar in Alk7 -/vs. control hearts (Fig 4A). However, mRNA levels of Kv4.2 and Kv4.3, subunits of potassium channel that carried I to current, were markedly reduced in Alk7 -/hearts by 44.3% and 37.3%, respectively ( Fig 4B). Notably, Kv4.2 protein level was also decreased by 30.5% in Alk7 -/hearts, whereas Kv4.3 protein level was not altered (Fig 4C). K + channel-interacting protein 2 (KCHIP2) is the accessory subunit of Kv4.2 and plays a key role in regulating Kv4.2 functionality and I to current. We found that KCHIP2 protein was reduced by 32.5% but KCHIP2 mRNA was not decreased in Alk7 -/hearts (Fig 4B  right panel and 4C). Thus, the diminished I to current in Alk7 -/cardiomyocytes could be attributable to the reduction of K + channel subunit Kv4.2 and its accessory subunit KCHIP2. Rapid delayed rectifier currents (I Kr ) and slow delayed rectifier currents (I Ks ) are two major rectifier outward currents during ventricular repolarization [37]. mRNA levels of KCNH2 (Erg) and KCNQ1 (Kv7.1), the pore-forming subunit for I Kr and I Ks respectively, were moderately increased by 20% in Alk7 -/mice relative to controls (Fig 4B). Interestingly, mRNA of KCNE, the accessory subunit responsible for both I Ks , was significantly reduced by 33.8% in Alk7 -/hearts (Fig 4C right panel). Kir2.1 is the main subunit of I K1 carrying channel. Although the current of I K1 was significantly reduced in Alk7 cardiomyocytes (Fig 3F), we did not observe any downregulation of Kir2.1 in Alk7 -/hearts (Fig 4B left panel and 4C). Taken together, depletion of ALK7 selectively suppressed the expression of Kv4.2, KCHIP2 and KCNE in cardiomyocytes.

Discussion
Using a genetically modified Alk7 -/mouse model, we demonstrated for the first time that endogenous ALK7 expression in ventricular cardiomyocyte was necessary to prevent QT interval prolongation, maintain normal APD and protect against ventricular arrhythmia. The arrhythmogenic consequence of ALK7 deletion could be directly mediated by ALK7 signaling, or it could probably be related to adaptive processes after genetic alteration. Thus, to test whether the acute inhibition of ALK7 signaling reproduces the phenotypes of Alk7 -/mice be helpful to distinguish the two possibilities. Interestingly, Alk7 -/mice do not have alterations in cardiac histology and cardiac function at baseline, suggesting that ALK7 is either dispensable in maintaining normal cardiac histology/functionality or completely compensated by other pathways.
Prolonged ventricular APD could trigger ventricular tachycardia in humans and in murine models. Alk7 -/mice exhibited increased QT intervals in telemetry ECG recordings. Furthermore, they had prolonged APD and elevated incidence of ventricular tachycardia in Langendorff-perfused hearts ex vivo. Surprisingly, Alk7 -/mice did not develop spontaneous ventricular in vivo. Such discrepancy could probably be explained by the fact that rodents are more resistant to ventricular tachycardia than humans due to the immensely speedy repolarization, small heart sizes and different activation pattern of ventricular myocardium in murine hearts [38]. Because Alk7 gene polymorphism has been linked to cardiovascular remodeling in humans [17], our findings in murine cardiomyocytes indicate that ALK7 may serve as a potential therapeutic target for ventricular tachycardia in humans. Interestingly, two out of eight (25%) Alk7 -/mice developed spontaneous sinus arrest, atrioventricular block and escape rhythm, which could lead to bradycardia. This unexpected in-vivo finding imply the potential role of ALK7 in pacemaker cells and cardiac electrical conduction system.
Mechanistically, Alk7 deletion in cardiomyocytes resulted in~40% reduction of I to and a left shift of its steady-state inactivation curve. The decreased I to density could be due to the reduction of Kv4.2 proteins in Alk7 -/hearts. The left shift of steady-state inactivation could probably be explained by the decreased level of KCHIP2, the accessary subunit responsible for I to inactivation [39,40]. Notably, the reduced KCHIP2 protein was not associated with a parallel change in KCHIP2 mRNA, suggesting certain posttranscriptional mechanism could modulate KCHIP2 translation or protein stability. Kv4 proteins are critical in maintaining the stability of KCHIP during channel complex trafficking to the membrane and vice versa [41]. Therefore, we speculated that the~40% reduction in KCHIP2 protein was due to the downregulation of Kv4.2 and subsequent enhancement of KCHIP2 protein instability in Alk7 -/cardiomyocytes.
Another interesting observation in this study was that the I K1 current densities were attenuated in Alk7 -/cardiomyocytes (Fig 3F), while Kir2.1 mRNA or protein levels were not significantly reduced (Fig 4B left panel and 4C). These results suggest that the decrease of I K1 current in Alk7 -/mice may be due to the post-translational modification of channel properties, and the underlying mechanisms require further investigation.
This study has raised an important question: what are the molecular mechanisms linking ALK7 to potassium channel remodeling and the transcription of Kv4.2 in cardiomyocytes? Several candidate molecular signals have been implicated to regulate the transcription of Kv4.2. For example, activation of AKT and mTOR pathway can enhance Kv4.2 transcription in cerebellar granule neurons and Hela cells [42]. GATA4 in cardiomyocytes also upregulates Kv4.2 transcription [43]. Conversely, calcineurin and NFATc3 signaling suppresses Kv4 expression in the mouse left ventricle during chronic beta adrenergic stimulation [44]. Thus, the expression and function of the abovementioned candidate transcription factors needs to be examined in Alk7 -/cardiomyocytes. An unbiased transcriptome analysis in control vs. Alk7 -/cardiomyocytes will provide a more comprehensive survey of differentially expressed genes and further reveal novel functions of ALK7 in cardiomyocytes.

Conclusion
In the present study, we showed that ALK7 depletion reduced repolarization K + currents, prolonged QT intervals and enhanced the susceptibility to ventricular arrhythmia in mice. The alteration of cardiac electrophysiology in Alk7 -/mice is not associated with cardiac fibrosis or impaired cardiac function. These findings uncover an unexpected anti-arrhythmogenic role of endogenous ALK7 beyond its role in preventing cardiac fibrosis and structural remodeling. The long QT syndrome may be either congenital or acquired [45]. Given that Alk7 -/mice exhibit similar phenotypes to that of acquired long-QT syndrome without structural remodeling or heart functions impairment, Alk7 -/mice are the potentially useful experimental model to study the mechanisms of acquired long-QT syndrome and dysregulated ventricular repolarization.  Table. Echocardiography parameters in control and Alk7-/-mice. (PDF)