The absence of insulin signaling in the heart induces changes in potassium channel expression and ventricular repolarization

Abstract

Diabetes mellitus increases the risk for cardiac dysfunction, heart failure, and sudden death. The wide array of neurohumoral changes associated with diabetes pose a challenge to understanding the roles of specific pathways that alter cardiac function. Here, we use a mouse model with cardiomyocyte-restricted deletion of insulin receptors (CIRKO, cardiac-specific insulin receptor knockout) to study the specific effects of impaired cardiac insulin signaling on ventricular repolarization, independent of the generalized metabolic derangements associated with diabetes. Impaired insulin action caused a reduction in mRNA and protein expression of several key K⁺ channels that dominate ventricular repolarization. Specifically, components of transient outward K⁺ current fast component (I_to,fast; Kv4.2 and KChiP2) were reduced, consistent with a reduction in the amplitude of I_to,fast in isolated left ventricular CIRKO myocytes, compared with littermate controls. The reduction in I_to,fast resulted in ventricular action potential prolongation and prolongation of the QT interval on the surface ECG. These results support the notion that the lack of insulin signaling in the heart is sufficient to cause the repolarization abnormalities described in other animal models of diabetes.

diabetic cardiomyopathy is associated with cardiac dysfunction, such as impaired diastolic and systolic function, in addition to structural and metabolic changes in the myocardium. Diabetes-associated cardiovascular dysfunction can lead to heart failure, independently of coronary artery disease and hypertension (3). Diabetes is associated with electrocardiogram (ECG) alterations that may increase the risk of arrhythmias. Specifically, prolongation of the QT interval has been reported as a consequence of an increase in ventricular action potential (AP) duration (APD) (6, 7). Electrophysiological studies have shown that a reduction in K⁺ repolarizing currents is responsible for the APD prolongation in cardiomyocytes isolated from diabetic animals (8, 15, 30). Specifically, attenuation of the rapidly activating and inactivating cardiac transient outward K⁺ current (I_to) underlies APD prolongation in type I and II diabetic models from a variety of species, including dog, rabbit, rat, and mouse (15, 33, 34). In concordance with these findings, gene and protein levels of the channel pore-forming subunits were decreased in diabetic hearts (20, 25).

Two I_to components with distinct recovery kinetics, referred to as fast and slow components of I_to (I_to,fast and I_to,slow, respectively), have been distinguished in most cardiac cells types (35). Both components of I_to are differentially expressed and contribute to regional heterogeneity in AP waveforms (5). In mouse myocytes, I_to is highly expressed and dominates the repolarization phase, causing an abbreviated AP, a typical feature of these cardiac cells (19, 21). Therefore, I_to has a prominent role in cardiac ventricular repolarization, and changes in this current can significantly affect APD in murine hearts (13, 21). I_to,fast channels are heteromultimers, composed of Kv4.2 and Kv4.3 α-subunits. In contrast, I_to,slow channels are encoded by the Kv1.4 α-subunit (5, 14). The cytoplasmic KChIP2 protein is an essential component of cardiac I_to,fast in that coexpression with KChIP subunits modulates the biophysical properties and increases the cell surface expression of Kv4 channels (26, 37). Other studies suggest that expression of Kv4 and KChIP2 proteins are reciprocally regulated (9, 10).

While various animal models have been used to characterize the functional consequences of diabetes in the heart (3, 28, 36), parsing the specific effects of insulin signaling in the heart from the confounding systemic effects of altered metabolism (such as hyperglycemia and hyperlipidemia) remains a challenge. In this work, we used a cardiac-specific insulin receptor knockout (CIRKO) mouse model that enabled us to examine the electrical consequences of absent insulin signaling in the heart at the molecular, cellular, and whole animal levels. The cardiac phenotype of CIRKO mice has been previously described and includes a reduction in heart size and changes in cardiac substrate utilization (2, 4, 27). There is a modest reduction in echocardiographic indexes of ventricular function (2, 4), although cardiac output remains normal (4). The modest reductions in cardiac contractility do not lead to heart failure.

METHODS

Mouse generation.

CIRKO mice were generated by crossing mice that were homozygous for a floxed insulin receptor allele in which loxP sites flank exon 4 of the insulin receptor gene (IR^lox/lox) with IR^lox/lox transgenic mice with cardiac-specific expression of cre recombinase (driven by the α-myosin heavy-chain promoter). CIRKO mice have the genotype Cre-IR^lox/lox, and their littermate controls have the genotype IR^lox/lox [referred to hereafter as wild-type (WT) littermates], as in the studies of Abel and colleagues (1, 2). Previously, we demonstrated no differences in glucose tolerance, glucose uptake, concentration of insulin or metabolic substrates, body/heart weight, or contractile performance between lox/lox mice (loxP homozygous mice without cre expression) and WT mice (1).

Cell isolation.

Left ventricular (LV) myocytes were isolated from CIRKO (genotype Cre-IR^lox/lox) mice, and their control WT littermates (genotype IR^lox/lox) (22–25 wk old) by enzymatic digestion were previously described in detail (2, 28). Briefly, after heparin injection, mice were anesthetized with chloral hydrate (1 mg/g body wt) by intraperitoneal injection. The heart was removed, rapidly cannulated, and then retrogradely perfused on a modified Langendorff apparatus through the aorta with the following solution: (in mM) 126 NaCl; 4.4 KCl; 1 MgCl₂; 4 NaHCO₃; 10 HEPES; 30 2,3-butanedione monoxime; 5.5 glucose; 1.8 pyruvate; 0.025 CaCl₂ (pH 7.3); and 0.3 mg/ml collagenase type IV (Worthington Biochemical). Following collagenase type IV-containing solution, the LV free wall was dissected and gently triturated in the Kraftbruhe solution until cell dissociation. Isolated myocytes were stored in Kraftbruhe solution at 4°C until use. During the entire perfusion protocol, the solutions were maintained at 37°C. All animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Utah.

Patch-clamp recording.

Voltage-gated K⁺ currents were recorded in the whole cell configuration of the patch-clamp technique using an Axopatch 200b amplifier (Molecular Devices, Sunnyvale, CA). All experiments were performed at room temperature (21–23°C). Data acquisition and command potentials were controlled by pClamp 8.0 software (Molecular Devices). Patch pipettes were made from borosilicate capillary glass (World Precision Instruments, Sarasota, FL), with a resistance of 2 to 3 MΩ when filled with the following internal solution: (in mM) 135 KCl, 1 MgCl₂, 10 EGTA, 10 HEPES; and 5 glucose (pH 7.2). The bath solution contained (in mM) 136 NaCl; 4 KCl; 1 CaCl₂, 2 MgCl₂, 5 CoCl₂, 10 HEPES, and 10 glucose (pH 7.4). I_to,fast was recorded in the presence of tetraethylammonium (TEA; 25 mM). For current-clamp experiments, CoCl₂ was omitted from bath solution. Cell capacitance was measured by integrating the surface area of the capacitive transient evoked during ± 10-mV voltage steps. Compensation for series resistance (∼80%) and capacitance were performed electronically. For each test pulse, the amplitude of I_to was measured as the difference between the maximal outward current and the current level at the end of the depolarizing pulse.

In current-clamp recordings, myocyte APs were recorded with the Axoclamp-2A amplifier system (Axon Instruments) in bridge mode. APs were triggered by 2 ms square pulse of depolarizing intracellular current at 1 Hz and recorded after reaching the steady state.

Patch-clamp data were processed using Clampfit 10.2 (Molecular Devices) and analyzed in Origin 7.5 (OriginLab, Northampton, MA). Currents amplitudes were normalized for cell capacitance (in pApF⁻¹). The decay phases of the outward K⁺ currents were measured by fitting the sum of two exponential functions using the following equation:

$$y\left(t\right)=A_1\exp \left(-t/{\tau }_1\right)+A_2\exp \left(-t/{\tau }_2\right)+B$$

where τ is time, τ₁ and τ₂ are the time constants of decay of the inactivating K⁺ currents, A₁ and A₂ are the amplitudes of the inactivating current components, and B is the amplitude of the steady-state, noninactivating component of the total outward K⁺ currents. The voltage dependence of steady-state inactivation of I_to was determined by normalizing peak currents elicited by the test pulses as I_to/I_to,max and plotted as a function of conditioning V_m. The relationships were fitted to a Boltzman equation of the form:

$$I_{\text{to}}/I_{\text{to max}}=1/\left\{1+\exp \left[\left(V_{\text{m}}-V_{1/2}\right)/k\right]\right\}$$

where V_m is the conditioning voltage and V_1/2 and k are the half-maximal inactivation potential and the slope of the curve, respectively.

Western blot and real-time quantitative PCR analysis.

Cell surface membrane preparations were extracted from LV tissue of WT and CIRKO hearts. Tissue samples used in the membrane preparations were rapidly frozen in liquid nitrogen after organ harvest. The tissues were then homogenized using a sucrose homogenization buffer containing a Halt protease and phosphatase inhibitor cocktail (Pierce, Rockford, IL) and centrifuged for 20 min at 4,000 rpm. The supernatant was transferred to ultracentrifuge tube and centrifuged for 50 min at 90,000 rpm. The resultant pellet membrane fraction was resuspended in radioimmunoprecipitation assay containing a Halt protease and phosphatase inhibitor cocktail and stored at −80°C until use. Western blot analysis was performed as previously described (31). Protein concentration of each of the solubilized samples was determined using the Micro BCA Protein Assay Kit (Pierce) with bovine serum albumin as the standard. For Western blot analysis, cell membrane-enriched protein was loaded and fractionated on a 10% SDS-polyacrylamide gel and electrotransferred to polyvinylidene difluoride membranes, blocked, and followed by overnight incubation at 4°C with one of the following monoclonal primary antibodies. The antibodies used were anti-Kv4.2, anti-Kv4.3, anti-Kv1.5, anti-KChIP2 [University of California (UC), Davis/National Institutes of Health (NIH) NeuroMab facility], at dilutions of 1:200. Coomassie blue (Bio-Rad, Hercules, CA) was used as loading control. Bound primary antibody was detected with a 1:10,000 dilution of goat anti-mouse Alexa Fluor 680 (Invitrogen) and goat anti-rabbit 800 (LI-COR Biosciences, Lincoln, NE). Detection was performed by measuring intensity of fluorescence from secondary antibodies using the Odyssey Infrared Imaging System, and quantification was done with their accompanying software (version 3.0, LI-COR Biosciences). We were unable to detect any signal in WT LV using two separate monoclonal antibodies against Kv2.1 (COOH-terminal and extracellular S1-S2 loop epitopes, respectively; UC Davis/NIH NeuroMab facility), despite detecting a strong signal in mouse brain (data not shown). Thus no further experiments were performed to measure Kv2.1 protein.

Real-time quantitative PCR analysis was achieved to determine the relative mRNA expression levels of genes encoding Kv4.2, Kv4.3, KChIP2, Kv1.5, Kv2.1, and mouse ether-a-go-go in LV of WT and CIRKO mouse hearts. RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA) and purified using the RNAeasy total RNA isolation kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. RNA (3 μg) was synthesized to cDNA using Superscript III Reverse Transcriptase (Invitrogen). Quantitative real-time PCR was performed with ABI Prism 7900 HT Real-Time PCR System (Applied Biosystems, Foster City, CA) using 384-well plates. SYBR green I fluorescent dye (Molecular Probes-Invitrogen) was used to quantify the relative mRNA levels. All reactions were performed in triplicate. Transcript levels for the constitutive housekeeping gene product cyclophilin were also quantitatively measured in each sample. PCR data were initially calculated as the number of transcripts per number of cyclophilin molecules, and the data were normalized to the mean WT value.

Surface ECG recording.

Standard surface ECG recordings were obtained from anesthetized WT and CIRKO (22–25 wk old) mice. Surface ECG were recorded with silver electrodes placed on the skin at optimized positions to obtain maximal amplitude recordings. Data were recorded and analyzed using LabChart7 Pro software (ADInstruments, Colorado Springs, CO), and QT, QTc, QRS, PR, and RR intervals were determined; average values for individual animals are reported.

Statistical analysis.

Results are presented as means ± SE. In all experiments, n represents the number of cells. Student's t-test or ANOVA was applied to compare individual data sets. A two-tailed probability value of <0.05 (P < 0.05) was considered statistically significant.

RESULTS

Ventricular Kv channels protein and mRNA levels were diminished in CIRKO mice.

Previous findings reported that outward K⁺ currents were significantly decreased in CIRKO mice (28). Thus we focused our analysis on the expression levels of the genes encoding the principal murine ventricular Kv channels (Kv4.2, Kv4.3, KChIP2, Kv1.5, Kv2.1) in LV tissue from CIRKO and their control littermate hearts (WT) (22–25 wk old). As shown in Fig. 1, mRNA expression of Kv4.2, one of the components of I_to,fast, was reduced by ∼80% in CIRKO mice. In contrast, the other component of I_to,fast, Kv4.3, remained unchanged in CIRKO mice. KChIP2, the accessory subunit of I_to,fast, was reduced by ∼50%. Gene expression of the pore-forming α-subunits of the slowly inactivating K⁺ current (I_K,slow), Kv1.5 and Kv2.1, were both diminished by ∼70%.

Fig. 1.

Summary of the expression levels of genes encoding Kv4.2, Kv4.3, KChIP2, Kv1.5, and Kv2.1 in left ventricle (LV) from 6 wild-type (WT) and 6 cardiac-specific insulin receptor knockout (CIRKO) mice (22–25 wk old). mRNA from heart tissue was amplified by real-time PCR and normalized to cyclophilin. Expression levels of Kv4.2, KChIP2, Kv1.5, and Kv2.1 subunits mRNA were reduced in CIRKO mice (*P < 0.05) compared with WT.

We next investigated the effect of absent cardiac-specific insulin signaling on ion channel protein levels. We performed Western blot analysis of cell membrane-enriched protein in LVs from WT and CIRKO hearts (Fig. 2). Kv4.2 channel protein was reduced by ∼50% in CIRKO mice (Fig. 2). No difference was found for Kv4.3 membrane levels between WT and CIRKO hearts. Protein levels of KChIP2 were decreased by ∼40% in CIRKO hearts compared with WT. Additionally, Kv1.5 membrane band densities decreased by ∼50% in CIRKO hearts. These results parallel the decline in mRNA expression.

Fig. 2.

Cell surface membrane expression of Kv channels subunits in LV myocytes isolated from WT and CIRKO mice. A: representative examples of isolated cell surface membranes from LV of WT and CIRKO mouse hearts immunoblotted with the anti-Kv4.2, anti-Kv4.3, anti-KChIP2, and anti-Kv1.5 antibodies. Coomassie blue was used to normalize protein signals. KO, knockout. B: densitometric quantification of respective membrane protein expression levels of Kv4.2 (n = 6), Kv4.3 (n = 6), KChIP2 (n = 5), and Kv1.5 (n = 5). Kv4.2, KChIP2, and Kv1.5 abundance was significantly diminished in LVs of CIRKO mouse hearts (*P < 0.05). The expression level of Kv4.3 was unaffected in CIRKO mice.

Reduced outward K⁺ currents in CIRKO mice.

In mouse ventricular myocytes, different voltage-gated K⁺ currents have been identified, which are kinetically and pharmacologically distinct (35): I_to,fast, I_to,slow, I_K,slow, and steady-state outward current (I_ss). I_to,fast and I_to,slow are resistant to TEA exposure (25 mM), whereas both I_K,slow and I_ss are reduced by ∼60%. Thus we analyzed the current remaining after 25 mM of TEA (Fig. 3). Outward K⁺ currents were elicited by 4-s step depolarizations between −40 and +60 mV, from a −70 mV holding potential. Both peak outward K⁺ current densities (Fig. 3B) and the current at the end of the pulse (Fig. 3C) were decreased in CIRKO mice. The peak current density at +40 mV was significantly reduced in CIRKO (52 ± 2), compared with WT myocytes (16 ± 3, n = 5, *P < 0.05). As reported previously (35), the decay phases of the outward K⁺ currents evoked during long depolarization pulses are well described by the sum of two exponentials. Our data were described by the sum of two exponentials in the presence of 25 mM of TEA (means ± SE). inactivation time constants (τ_decay) in WT myocytes were 395 ± 14 and 36 ± 1 ms. These τ_decay values were not significantly different from CIRKO myocytes: 349 ± 11 and 37 ± 1 ms, respectively. Analysis of the decay phases are represented in Fig. 3D.

Fig. 3.

Outward K⁺ currents are reduced in LV of CIRKO mice. A: representative whole cell voltage-gated outward K⁺ currents recorded from WT (top) and CIRKO (bottom) LV myocytes after exposure to 25 mM of tetraethylammonium (TEA). Inset: TEA-resistant currents were evoked using the voltage protocol. Mean current-voltage relationships are shown for peak current density (I_peak; B) and I_ss (C). Both, I_peak and I_ss (TEA resistant) current densities were significantly decreased in CIRKO myocytes (P < 0.05). D: analysis of the decay phases of the outward K⁺ currents revealed no significant differences in WT and CIRKO cells. Inactivation time constants were determined from the double exponential fits to the decay phases (τ_decay). V_m, conditioning voltage.

In the next experiments, we determined whether the absence of cardiac insulin signaling altered the biophysical properties of I_to,fast. Figure 4 compares the voltage dependence of steady-state inactivation of I_to,fast in LV myocytes isolated from WT and CIRKO mice. Currents were examined using a voltage steps protocol to +50 mV after 5-s conditioning prepulses to potentials between −100 and −10 mV. The amplitudes of I_to,fast in WT and CIRKO were determined from a single exponential fit to the decay phases of the currents evoked at +50 mV from each prepulse potential and then normalized to its respective maximal current amplitude evoked from −100 mV (in the same cell). Normalized I_to,fast amplitudes were plotted as a function of conditioning potential as illustrated in Fig. 4B and fitted to a single Boltzmann function. Neither V_1/2 nor k were significantly altered in CIRKO myocytes (V_1/2 = −31 ± 1 mV, and k = 3.7 ± 1 mV for WT; and V_1/2 = −34 ± 2 mV, and k = 4.3 ± 2 mV for CIRKO; n = 8). The analysis of the recovery of I_to,fast from steady-state inactivation is shown in Fig. 5. Cells were first depolarized to +50 mV for 5 s to inactivate I_to,fast, then hyperpolarized to −70 mV for varying times ranging from 0 to 1,300 ms, and finally stepped to +50 mV. The amplitude of I_to,fast evoked at +50 mV after the recovery period was determined from a single exponential fit to the decay phases of the total outward current and normalized to the current amplitude evoked after the recovery time (in the same cell). The normalized I_to,fast amplitudes plotted as a function of recovery time are presented in Fig. 5B. The recovery data for I_to,fast were well described by a single exponential with time constants of 19 ± 2 ms for WT and 16 ± 1 ms for CIRKO myocytes (n = 5). Our data show that I_to,fast current density was significantly diminished; however, neither the time nor voltage-dependent properties of I_to,fast were affected in CIRKO myocytes.

Fig. 4.

Comparison of voltage dependence of steady-state inactivation of transient outward K⁺ current fast component (I_to,fast) in LV myocytes isolated from WT and CIRKO. A: representative outward K⁺ current evoked by a 2-pulse protocol (inset). B: normalized I_to,fast amplitudes in WT and CIRKO were plotted as a function of conditioning potential. The amplitude of I_to,fast evoked at +50 mV from each test potential were determined from exponential fits to the decay phases and normalized to the current amplitudes evoked from −100 mV (in the same cell). Solid lines correspond to the Boltzmann fits with half-maximal inactivation potential and slope factors of −31 ± 1 mV (k = 3.7 ± 1 mV) for WT and −34 ± 2 mV (k = 4.3 ± 2 mV) for CIRKO myocytes. The voltage dependence of steady-state inactivation of I_to,fast was identical in both groups (n = 8).

Fig. 5.

Characterization of recovery from steady-state inactivation of I_to,fast in WT and CIRKO LV myocytes. A: typical outward K⁺ current showing the time course of recovery from inactivation of I_to,fast. Currents were evoked using the voltage protocol (inset). The recovery of the current is shown on an expanded time scale (dotted lines). B: the rates of recovery of I_to,fast from inactivation were well described by a single exponential. The solid lines represent the single exponential fits with time constants of 19 ± 2 ms for WT and 16 ± 1 ms for CIRKO myocytes (n = 5). There were no significant differences between both groups.

Previous studies of diabetic animals have shown that the inwardly rectifier K⁺ current (I_K1) is not affected by diabetes (8, 33). Here, we explored whether the absence of insulin receptor in the heart might affect I_K1. Figure 6 illustrates that I_K1 current density was not significantly changed in myocytes isolated from CIRKO mice.

Fig. 6.

Inward rectifier K⁺ currents are not affected in CIRKO mice. The current-voltage relationships of inwardly rectifier K⁺ current (I_K1) were not different between WT and CIRKO LV myocytes (n = 8).

CIRKO mice display APD and QT prolongation.

To determine the effects of absent cardiac insulin receptors at the cellular level, we recorded APs in LV myocytes isolated from WT and CIRKO mice. As shown in Fig. 7A, APs waveforms were prolonged in CIRKO myocytes. The quantitative analysis of APD at 50 and 90% of repolarization revealed significant prolongation in CIRKO (7.3 ± 0.9 and 78.3 ± 7 ms, n = 10) compared with WT myocytes (3 ± 0.2 and 25.8 ± 2.2 ms, n = 7). By contrast, resting membrane potentials, APD at 10%, APD at 25%, and AP amplitudes were unchanged (data are summarized in Table 1). Also, the maximal upstroke velocity was not different between the WT and CIRKO mice APs. Finally, we examined the effect of absent cardiac insulin receptors at the level of the whole organism by measuring ECG activity in anesthetized mice. Figure 8 illustrates the comparison of ECG recordings from WT and CIRKO mice. Analyses of the ECG parameters confirmed a significant prolongation in QRS, QT, and QTc intervals in CIRKO mice relative to control littermates (n = 6; P < 0.05). Both rate-corrected and -uncorrected QT intervals were significantly longer in CIRKO animals.

Fig. 7.

Action potential duration (APD) was prolonged in CIRKO LV cells. A: representative action potentials waveforms recorded from LV myocytes isolated from WT and CIRKO mice (22–25 wk old). Action potentials were significantly lengthened in LV cells from CIRKO mice. B: analyses of the active membrane properties in both groups revealed a significant difference in APD measured at 50 and 90% (APD₅₀ and APD₉₀) of repolarization in CIRKO cardiomyocytes.

Table 1.

Comparison of action potential parameters in left ventricular myocytes isolated form WT and CIRKO mice

	WT	CIRKO
RP, mV	−78.7 ± 0.6	−79.4 ± 0.6
APA, mV	128.7 ± 1.3	129.7 ± 1.2
APD10, ms	1.1 ± 0.1	1.5 ± 0.1
APD50, ms	3 ± 0.2	7.3 ± 0.9*
APD90, ms	25.8 ± 2.2	78.3 ± 7*
dV/dtmax, V/s	0.299 ± 18	0.292 ± 8

Values are means ± SE, showing action potential durations at 10 (APD₁₀), 50 (APD₅₀), and 90% (APD₉₀) of repolarization; maximal upstroke velocity (dV/dt_max); resting membrane potentials (RP); and action potential amplitudes (APA). WT, wild-type; CIRKO, cardiac-specific insulin receptor knockout.

^*P < 0.05, significantly different.

Fig. 8.

CIRKO mice display QT prolongation. A: analyses of ECG parameters revealed significant differences in QRS, QT and QTc intervals in WT and CIRKO mice (n = 6). BPM, beats/min. B: representative surface ECG recordings obtained from both groups studied. QT intervals are indicated below the records. C: summary representation of QT interval in WT and CIRKO mice. QT interval in CIRKO mice was significantly longer than in WT mice (*P < 0.05). D: observed QT intervals correlated with RR intervals and show no significant differences.

DISCUSSION

Diabetes mellitus affects more than 8% of the current United States population, and the worldwide epidemic of obesity and sedentary lifestyle is projected to result in over 300 million people with diabetes by 2025. Patients with diabetes are at a significant risk of developing cardiomyopathy, which is characterized by cardiac hypertrophy, ventricular dysfunction, and an increased risk of lethal arrhythmias. Cardiac insulin resistance is one factor that contributes to diabetic cardiomyopathy (17, 18). The wide array of neurohumoral changes associated with diabetes pose a challenge to understanding the roles of specific pathways that are altered in this complex disease. Here, we take advantage of CIRKO mice to study the specific effects of impaired cardiac insulin signaling on ventricular repolarization, independent of the generalized metabolic derangements associated with diabetes.

Reduced K⁺ channel gene and protein expression in CIRKO hearts.

We found that impaired insulin action caused a reduction in message and protein expression of several key K⁺ channels that dominate ventricular repolarization; specifically, components of I_to,fast (Kv4.2 and KChiP2) and Kv1.5 were downregulated in CIRKO hearts. Similar to our study of absent insulin signaling in the heart, some studies in animal models of diabetes also reveal a reduction in the message and protein levels of the components of I_to,fast (25). By contrast, other studies using a rat model of type I diabetes reported that Kv4.2, but not Kv4.3, mRNA expression was downregulated (20). In our work, neither Kv4.3 message or protein expression levels were altered in CIRKO mice; rather, a reduction in Kv4.2 and KChiP2 message and protein provides the molecular basis for reduced I_to. A specific role for insulin in modulating K⁺ channel gene expression is supported by the observation that incubation of myocytes isolated from diabetic hearts with insulin restored I_to current (16, 29, 33). These findings demonstrate the modulatory effect of insulin on K⁺ channel gene and protein expression. In addition, a mechanistic link between insulin signaling and K⁺ channel gene expression comes from the observation that activation of insulin receptors by a neurotrophic factor induces mRNA and protein expression of Kv4.2 and increases the transient outward K⁺ current (I_A) densities in rat cerebellar granule neurons. Moreover, insulin receptor pathway blockade reduced neurotrophic factor-mediated induction of I_A current densities and Kv4.2 protein levels. In this context, we speculate that the cardiomyocyte deletion of insulin receptors affects insulin-mediated gene and protein regulation of specific K⁺ channels.

Heterologous coexpression studies demonstrated that I_to,fast channels are generated by the assembly of Kv4.2/Kv4.3 and KChIP2 subunits (12). In this work, we showed that Kv4.2 and KChIP2 gene and protein expression are downregulated in CIRKO hearts, and these results are consistent with the reduction of I_to,fast current densities. However, the alternative α-subunit component of I_to,fast, Kv4.3 is not affected. There is strong evidence that Kv4.3 channels are not required for the generation of functional I_to,fast channels in mice (21). Targeted deletion of Kv4.2 completely eliminates I_to,fast in adult mouse ventricles, whereas in mice lacking Kv4.3, the current densities and the time and voltage-dependent properties of I_to,fast were similar to WT mice. These results indicated that Kv4.2 subunits are crucial for the generation of ventricular I_to,fast channels in adult mouse ventricles (11, 21). Thus the changes in Kv4.2 and KChIP2 channel gene and protein surface membrane expression observed in our study significantly altered I_to,fast current density in CIRKO hearts. Likewise, Kv1.5 and Kv2.1 mRNA messages are diminished, suggesting I_K,slow current density reduction in mice with cardiomyocyte-selective deletion of the insulin receptor.

Outward K⁺ currents are reduced in CIRKO ventricle.

Consistent with the observation of decreased cell surface expression of Kv4.2 and KChiP2 protein, we observed a reduction in I_to current density in CIRKO ventricle. The voltage dependence and kinetics of I_to,fast were not affected by the absence of cardiac insulin signaling. The absence of voltage dependent/kinetic changes is consistent with the finding that both Kv4.2 and KChIP2 proteins were diminished. KChIP2 is known to increase the current density and modify the biophysical properties of Kv4 channels, altering the time- and voltage-dependent properties of currents (9). A reduction in KChIP2 expression alone, suppresses I_to,fast and slows the kinetics of the fast component of inactivation (23). However, we speculate that the combined reduction KChIP2 and Kv4.2 preserves the voltage-dependent properties of I_to,fast. We did not find any differences in the voltage dependence of inactivation or recovery from inactivation of I_to,fast in CIRKO mice compared with WT channels.

Significant APD and QT prolongation in CIRKO mice.

A reduction in the densities of the repolarizing Kv4 currents in adult mouse ventricles modifies AP waveforms and contributes to the abnormal dispersion of repolarization (13, 24). Our findings suggest that as a consequence of reduced expression of I_to,fast components, ventricular APD was prolonged in CIRKO LV myocytes, which in turn, produced prolongation of the QT interval on the surface ECG. Similarly, ventricular APD prolongation was reported in other studies of diabetes (8, 33). QT prolongation is a known risk factor for sudden cardiac death and QTc interval prolongation constitutes an independent predictive factor for future strokes in patients with type 2 diabetes (6). Our study highlights the specific contribution of impaired cardiomyocyte insulin signaling to K⁺ channel gene expression that leads to QT prolongation, independent of the systemic metabolic derangements that accompany diabetes. Further studies are warranted to delineate the molecular mechanisms whereby insulin controls ion channel gene expression in the heart.

Limitations.

Insulin signaling plays an important role in regulating postnatal cardiac size, myosin isoform gene expression, and substrate utilization in the heart. Alterations in substrate metabolism and energy utilization in CIRKO mice likely result in secondary changes in expression of a variety of genes, some of which may influence the expression or function of cardiac ion channels to result in the observed changes in repolarization. As our focus was limited to specific cardiac ion channels, we cannot exclude the possibility that alterations in other genes also influence repolarization in CIRKO mice. Moreover, heart failure is known to induce changes in ion channel gene expression that can cause prolongation of the QT interval. While CIRKO mice exhibit modest changes in cardiac contractility (2, 4, 27), they do not develop heart failure and therefore the observed changes in ion channel expression in this study are not likely a consequence of heart failure per se.

There are significant differences in the repertoire of ion channel genes that modulate repolarization in the rodent and human ventricle. Whereas I_to dominates ventricular repolarization in rodents, the delayed rectifier K⁺ currents I_Kr and I_Ks are the principal repolarizing currents in human ventricle (19, 22, 23). While our results indicate that the absence of insulin signaling in the murine heart alters expression of repolarizing channel proteins, it remains unclear whether similar changes might happen in humans.

Several studies indicate the importance of the transmural gradient of I_to and regional differences in APD as a consequence of the ionic currents existing throughout the LV free wall (25, 32). We focused our investigation on myocytes isolated from LV as the mass of the LV primarily controls the QT interval on the surface ECG. However, transmural and regional differences in ion channel expression maybe be differentially regulated by insulin or the lack of insulin signaling. In addition, we do not have evidence that QT prolongation observed in CIRKO mice translates into ventricular arrhythmias, as we did not perform continuous telemetry.

While our project was focused primarily on the effects of cardiac insulin deficiency on ventricular repolarization, we also observed QRS prolongation in CIRKO mice. There was no difference in the maximal upstroke velocity between control and transgenic mice, suggesting that Na⁺ channel function was not appreciably different. It is possible that the absence of cardiac insulin deficiency or secondary consequences thereof may influence expression of connexin genes to prolong the QRS complex.

Conclusions.

In summary, impaired insulin action in the heart caused a reduction in message and protein expression of several key K⁺ channels that dominate ventricular repolarization. Kv4.2 and KChiP2 message and protein were reduced, leading to a reduction in the amplitude of I_to,fast in isolated LV CIRKO myocytes. As a consequence of reduced I_to,fast, ventricular APD and QT intervals were prolonged. These results support the concept that the lack of insulin signaling in the heart is sufficient to cause the repolarization abnormalities described in other animal models of diabetes.

GRANTS

This work was supported by the Nora Eccles Treadwell Foundation and by American Heart Association (AHA) Grants RO1DK092065 and UO1HL087947 (to E. D. Abel, an established AHA investigator). A. R. Wende was supported by Juvenile Diabetes Research Foundation Advanced Postdoctoral Fellowship 10-2009-672 and National Heart, Lung, and Blood Institute (NHLBI) Grant K99-HL-111322. R. O. Pereira was supported by AHA (Western Affiliates) postdoctoral fellowship from the and by NHLBI grant T32-HL-007576.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

A.L.-I., R.P., A.W., B.B.P., E.D.A., and M.T.-F. conception and design of research; A.L.-I., R.P., and A.W. performed experiments; A.L.-I., R.P., A.W., and M.T.-F. analyzed data; A.L.-I., R.P., A.W., B.B.P., E.D.A., and M.T.-F. interpreted results of experiments; A.L.-I. and A.W. prepared figures; A.L.-I. drafted manuscript; A.L.-I., R.P., A.W., B.B.P., E.D.A., and M.T.-F. edited and revised manuscript; A.L.-I., R.P., A.W., B.B.P., E.D.A., and M.T.-F. approved final version of manuscript.