Hyperglycemia regulates cardiac K⁺ channels via O-GlcNAc-CaMKII and NOX2-ROS-PKC pathways

Abstract

Chronic hyperglycemia and diabetes lead to impaired cardiac repolarization, K⁺ channel remodeling and increased arrhythmia risk. However, the exact signaling mechanism by which diabetic hyperglycemia regulates cardiac K⁺ channels remains elusive. Here, we show that acute hyperglycemia increases inward rectifier K⁺ current (I_K1), but reduces the amplitude and inactivation recovery time of the transient outward K⁺ current (I_to) in mouse, rat, and rabbit myocytes. These changes were all critically dependent on intracellular O-GlcNAcylation. Additionally, I_K1 amplitude and I_to recovery effects (but not I_to amplitude) were prevented by the Ca²⁺/calmodulin-dependent kinase II (CaMKII) inhibitor autocamtide-2-related inhibitory peptide, CaMKIIδ-knockout, and O-GlcNAc-resistant CaMKIIδ-S280A knock-in. I_to reduction was prevented by inhibition of protein kinase C (PKC) and NADPH oxidase 2 (NOX2)-derived reactive oxygen species (ROS). In mouse models of chronic diabetes (streptozotocin, db/db, and high-fat diet), heart failure, and CaMKIIδ overexpression, both I_to and I_K1 were reduced in line with the downregulated K⁺ channel expression. However, I_K1 downregulation in diabetes was markedly attenuated in CaMKIIδ-S280A. We conclude that acute hyperglycemia enhances I_K1 and I_to recovery via CaMKIIδ-S280 O-GlcNAcylation, but reduces I_to amplitude via a NOX2-ROS-PKC pathway. Moreover, chronic hyperglycemia during diabetes and CaMKII activation downregulate K⁺ channel expression and function, which may further increase arrhythmia susceptibility.

Introduction

Diabetes mellitus (DM) and its hallmark metabolic abnormality, hyperglycemia, are closely associated with cardiovascular morbidity and mortality [9]. While the best characterized cardiac complications of DM include coronary artery disease and structural remodeling of the heart, DM is also associated with increased cardiac arrhythmia risk [18]. Among electrophysiological derangements that promote arrhythmias, altered K⁺ channel function leads to impaired action potential repolarization and is considered an important contributor to arrhythmogenesis [36, 40]. Critically, previous studies have reported downregulation of cardiac K⁺ channel expression and decreased K⁺ currents in animal models of both type 1 and type 2 DM (T1DM and T2DM, respectively) [22, 27, 29, 35, 39, 51, 54]. However, the signaling pathways by which diabetic hyperglycemia regulates K⁺ channels in the heart are unknown.

Calcium/calmodulin-dependent kinase II (CaMKII) is upregulated in diabetes and contributes to both cardiac remodeling and arrhythmias [11, 15, 21, 25]. Hyperglycemia has been shown to induce posttranslational modification of CaMKII by O-linked β-N-acetylglucosamine (O-GlcNAc) on Serine 280, inducing autonomous kinase activity [11]. In line with this, inhibition of either CaMKII or O-GlcNAcylation prevented spontaneous diastolic Ca²⁺ leak and arrhythmias in diabetic hyperglycemia [11]. CaMKII is also known to regulate K⁺ channels acutely by phosphorylation of the channel proteins and chronically by altering their expression [48]. However, the contribution of CaMKII to K⁺ channel remodeling during hyperglycemia and diabetes has not been investigated.

Oxidative stress is also frequently implicated in the pathophysiology of DM [13]. Multiple mechanisms can lead to increased generation of reactive oxygen species (ROS) in diabetic hyperglycemia, including enhanced NADPH oxidase 2 (NOX2) activity and a variety of derangements in mitochondrial metabolism [13, 24]. ROS regulates diverse cellular mechanisms, including excitation–contraction coupling, transcription, inflammation, and cell death [5]. ROS can also modulate ion channels either through direct protein modification or through oxidative stress-dependent activation of protein kinases, including protein kinase C (PKC) [12] and CaMKII [10]. However, the exact signaling mechanism of ROS-dependent K⁺ channel regulation are unclear. Therefore, mechanistic understanding of how diabetic hyperglycemia regulates K⁺ channels will be essential to understanding the mechanisms of arrhythmogenesis and designing more effective antiarrhythmic therapies in DM. We aimed to investigate the changes in major cardiac voltage-gated K⁺ currents during acute hyperglycemia and diabetes. We hypothesized that hyperglycemia regulates K⁺ channel biophysics and expression by activating CaMKII via O-GlcNAcylation and/or oxidation.

Methods

Animal models

Several types of adult male C57BL/6 J mice (10–12 weeks) were used, including wild-type (WT, Jackson Laboratory, Stock No. 000664), CaMKIIδ-knockout (KO) [23], CaMKIIδ_C (predominant cytosolic isoform)-overexpression (OE) [53], oxidation-resistant CaMKIIδ-MM281/282VV [25], and O-GlcNAc-resistant CaMKIIδ-S280A knock-in [24], and NOX2-knockout (Jackson Laboratory, Stock No. 002365). For rat and rabbit experiments, adult male Wistar rats (10–14 weeks) and New Zealand White rabbits (3–4 months) were used.

Isolation of left ventricular cardiomyocytes was performed as previously described [14]. Briefly, animals were injected with heparin (5000 U/kg) and anesthetized with isoflurane (5%). Hearts were excised and retrograde perfused on constant flow Langendorff apparatus (5 min, 37 °C) with Ca²⁺-free normal Tyrode’s solution, gassed with 100% O₂. Then, ventricular myocytes were digested using Liberase TM (0.225 mg/mL, Roche) for mice and rats, and collagenase type II (Worthington) and protease type XIV (Sigma-Aldrich) for rabbits. Ventricular myocytes were dispersed mechanically and filtered through a nylon mesh then allowed to sediment for 10 min. The sedimentation was repeated three times using increasing [Ca²⁺] from 0.125 to 0.25 then 0.5 mmol/L. Finally, ventricular myocytes were kept in Tyrode’s solution [0.5 mmol/L (Ca²⁺)] at room temperature until use.

DM was studied in three different mouse models: streptozotocin (STZ)-induced T1DM [21, 35], high-fat diet (HFD)-induced T2DM [32, 43] and db/db, a genetic model of T2DM (leptin-receptor deficiency) [2, 20]. In the first model, diabetes was induced in 6–8-week-old mice by intra-peritoneal injections with low-dose STZ (50 mg/kg body weight in 40 mmol/L sodium citrate, pH = 4.0) for 5 consecutive days and littermates received sodium citrate as vehicle control. Only mice exhibiting > 300 mg/dL blood glucose levels for 4 weeks following STZ injections (≈ 70% success rate) were included in the study. In the second model, mice were fed with HFD (60 kcal% fat, D12492, Research Diets) starting at 5 weeks of age for 16 weeks, whereas controls received low-fat diet (LFD, 10 kcal% fat, D12450J, Research Diets). In the third model, 10-week-old db/db mice (Jackson Laboratory, Stock No. 000642) were used and wild-type C57BLKS/J mice (Jackson Laboratory, Stock No. 000662) served as control. Blood glucose levels were assessed using OneTouch UltraMini (LifeScan) glucose monitoring system and test strips.

Transverse aortic constriction (TAC) with a 28G stenosis was performed in 8-week-old mice to induce pressure overload which led to heart failure (HF) in 8 weeks following the surgery. A sham procedure without banding the thoracic aorta was performed in control animals.

Electrophysiology

Isolated cardiomyocytes were transferred to a temperature-controlled chamber (Warner Instruments) mounted on a Leica DMI3000 B inverted microscope, and continuously perfused with Tyrode solution containing (in mmo/L): NaCl 140, KCl 4, CaCl₂ 1.8, MgCl₂ 1, HEPES 5, Na-HEPES 5, glucose 5.5 and mannitol 24.5; pH = 7.40 and osmolality = 320 ± 2 mOsm/L. High glucose effects were assessed 6-min following bath medium switch to a Tyrode solution containing 30 mmol/L glucose and 0 mannitol (osmolality and pH matched). Electrodes were fabricated from borosilicate glass (World Precision Instruments) having tip resistances of 2–2.5 MΩ when filled with internal solution containing (in mmol/L): K-Aspartate 100, KCl 20, NaCl 8, Mg-ATP 5, EGTA 10, CaCl₂ 4.1, HEPES 10, cAMP 0.002, phosphocreatine-K₂ 10, and calmodulin 0.0001, with pH = 7.2 free [Ca²⁺]_i = 100 nmol/L, calculated using the Web Maxc Extended version of MaxChelator software (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcE.htm). The electrodes were connected to the input of an Axopatch 200B amplifier (Axon Instruments). Outputs from the amplifier were digitized at 50 kHz using Digidata 1332A A/D card (Axon Instruments) under software control (pClamp 10). The series resistance was typically 3–5 MΩ and it was compensated by 90%. Experiments were discarded when the series resistance was high or increased by > 10%. Reported voltages are already corrected for liquid junction potentials.

Whole-cell K⁺ currents were recorded in the presence of Na⁺ and Ca²⁺ current inhibitors (10 μmol/L tetrodotoxin and 10 μmol/L nifedipine) in the bath solution. Different K⁺ current components were separated using appropriate voltage protocols and selective ionic channel inhibitors. The rapidly inactivating transient outward K⁺ current (I_to) and slowly inactivating K⁺ current (I_K,slow) in mouse were distinguished by biexponential fits (R² > 0.9 in each case) of the time-dependent decay of K⁺ current that is sensitive to 3 mmol/L 4-aminopyridine (4-AP) or by sensitivity of I_K,slow to 50 μmol/L 4-AP, as previously described [3, 50]. The non-inactivating steady-state K⁺ current (I_SS) in mouse was the 4-AP insensitive residual K⁺ current. The rapid delayed rectifier K⁺ current (I_Kr) and the inward rectifier K⁺ current (I_K1) were sensitive to 1 μmol/L E-4031 and 300 μmol/L Ba²⁺, respectively. Ionic currents were normalized to cell capacitance, determined in each cell using short (10 ms) hyperpolarizing pulses from – 10 mV to – 20 mV. All experiments were conducted at 37 ± 0.1 °C.

Detailed description of the voltage protocols and drug treatments are provided in the Supplemental Methods.

Transcript analysis using qRT-PCR

Total RNA from 8- to 10-week-old male CaMKIIδ knockout, overexpression, S280A and WT littermate (N = 3 each), and 16-week-old (8-week post-surgery) TAC and sham WT (N = 3 each) mouse hearts was extracted using RNeasy Mini Kit (Qiagen). Conversion to cDNA was performed using QuantiTect Reverse Transcription Kit (Qiagen). Transcript analysis of genes encoding CaMKIIδ (Camk2d), CaMKIIγ (Camk2g), markers of cardiomyocyte hypertrophy (β-myosin heavy chain, Myh7; atrial natriuretic factor, Anf), K⁺ channel pore forming and auxiliary subunits, as well as housekeeping control genes, including glyceraldehyde 3-phosphate dehydrogenase (Gapdh) and S18 ribosomal protein was carried out using quantitative real-time polymerase chain reaction (qRT-PCR) on an Applied Biosystems 7900HT Fast Real-Time PCR System and Primer–Probe detection. Three technical replicates were performed for each biological sample. The primers were from Eurofins Genomics and their sequences are provided in Suppl. Table 1. Transcript data were normalized to the arithmetic mean of Gapdh and S18, and analysed using the threshold cycle (C_T) relative quantification method, then linearized (2^−ΔCT) to make comparison.

Western blot analysis of K_ir2.1 protein expression

Hearts from WT, CaMKIIδ-KO, CaMKIIδ-OE and CaMKIIδ-S280A mice (N = 3 each) were collected and lysed in ice-cold buffer containing (in mmol/L): NaCl 150, HEPES (pH = 7.5) 10, NaF 50, sodium pyrophosphate 1, MgCl₂ 1, EGTA 1, EDTA 1, 1% Triton X-100, and protease and phosphatase inhibitors (EMD Millipore, set III and V, respectively). Protein concentration was determined by BCA assay (cat#23225, Thermo Fisher Scientific). Sample proteins were then separated by SDS-PAGE (4–20%) before transferring to a 0.2 μm nitrocellulose membrane. Immunoblots were blocked with 8% milk in 0.05% TBS-Triton (TBST). The blots were then incubated overnight at 4 °C with primary antibodies for K_ir2.1 (NeuroMab, Cat#: 75–210, Lot#: 455-6JD.13, 1:1000) and GAPDH (Abcam, Cat#: ab181602, Lot#: GR199633-12, 1:10,000). Following five TBST washes, specific secondary antibodies were applied for 2 h at room temperature. The blots were again washed sequentially in TBST, TBS and water before imaging on the Sapphire Biomolecular Imager (Azure Biosystem). Three blots were made for each protein sample.

Echocardiography

Transthoracic echocardiography was performed in anesthetized (isoflurane, 0.5–2%) animals before and after STZ and TAC procedures. Short axis M-mode images of the left ventricles were acquired using a Vevo 2100 echocardiography system (FUJIFILM VisualSonics, Toronto, ON, Canada) equipped with a 40 MHz transducer. Body temperature was carefully monitored, and anesthesia was adjusted to achieve ≈ 450 beats/min heart rate in each animal.

Statistics

Data are presented as mean ± SEM. The number of cells/animals (n/N) in each experimental group was reported in the figures. Statistical significance was tested using paired, two-tailed Student’s t test or two-way ANOVA followed by a post-hoc Bonferroni or Tukey test when applicable. Graph-Pad Prism 8.0 and Origin 2016 software were used. A p value of less than 0.05 was considered significant.

Results

Acute hyperglycemia affects the function of K⁺ channels in mouse ventricular myocytes

First, we tested the effect of high glucose (30 mmol/L) corresponding to what is observed in severe diabetes versus osmotically-matched (glucose substituted with equimolar mannitol) low glucose (5.5 mmol/L) on major voltage-gated K⁺ channels in mouse ventricular myocytes at 37 °C (Fig. 1a, b). Acute hyperglycemia (6 min) significantly reduced I_to, whereas I_K,slow and I_SS were unchanged (Fig. 1c-e). The fast and slow inactivation kinetics of I_to were unaltered in hyperglycemia (Fig. 1f). However, the recovery from inactivation kinetics of I_to were significantly enhanced by 20–26% in hyperglycemia (Fig. 1g). Importantly, I_K1 was increased in hyperglycemia (Fig. 1h, paired data obtained in individual cells are shown in Suppl. Fig.1a). I_Kr was small in mice and unaltered by acute hyperglycemia (Suppl. Fig.1b).

Fig. 1.

Acute hyperglycemia regulates K⁺ channels in mouse ventricular cardiomyocytes. a Representative voltage-gated K⁺ current traces in normoglycemia and acute hyperglycemia. The transient outward K⁺ current (I_to) and slowly-inactivating K⁺ current (I_K,slow) were inhibited by 3 mmol/L 4-aminopyiridine. Inset shows enlarged I_to. The steady-state K⁺ current (I_SS) was insensitive to 4-aminopyridine. b Representative inward rectifier K⁺ current (I_K1) traces. c I_to was significantly reduced in acute hyperglycemia. d, e I_K,slow and I_SS were unaltered in acute hyperglycemia. f I_to fast and slow inactivation time constants (τ_fast and τ_slow) were unchanged in acute hyperglycemia. g I_to recovery from inactivation was faster in hyperglycemia. h I_K1 increased in acute hyperglycemia. Student’s paired t test, *p < 0.05, **p < 0.01, ***p < 0.001

Hyperglycemia effects on K⁺ channels are shared between species

K⁺ channel expression in the heart is significantly different between species [40]. Therefore, we also tested the effects of hyperglycemia in rat and rabbit ventricular myocytes (which more closely resemble human cardiomyocyte electrophysiology). Although basal I_to density was smaller (I_to: rabbit < rat < mouse) and no I_K,slow was observed in rat and rabbit myocytes, hyperglycemia induced similar I_to reduction in both species (Fig. 2a, b). However, I_to recovery from inactivation was faster in rat vs. mouse (Fig. 2c). In rabbit, roughly half of total I_to exhibited very slow recovery kinetics (often distinguished as I_to,slow vs. I_to,fast; Fig. 2d). Importantly, hyperglycemia enhanced I_to recovery in both species (Fig. 2c, d). In contrast to I_to, basal I_K1 density was larger in rat and rabbit myocytes compared to mouse (I_K1: mouse < rat < rabbit) and acute hyperglycemia further increased I_K1 in both species (Fig. 2e, f). These data demonstrate that the effects of acute hyperglycemia on K⁺ channels are similar across 3 species.

Fig. 2.

Acute hyperglycemia regulates K⁺ currents in rat and rabbit ventricular myocytes. a, b Representative I_to traces and averaged data in normoglycemia and hyperglycemia conditions in rat (a) and rabbit (b) ventricular myocytes. c, d Enhanced I_to recovery from inactivation kinetics in hyperglycemia in both rat (c) and rabbit (d) cardiomyocytes. e, f Representative I_K1 traces and averaged data in rat (e) and rabbit (f). Student’s paired t test; *p < 0.05, **p < 0.01, ***p < 0.001

O-GlcNAc modifications mediate the effects of hyperglycemia on K⁺ channels

Impairment of the hexosamine biosynthetic pathway (HBP) leading to enhanced O-GlcNAcylation has been implicated in diabetic arrhythmias [11]. Therefore, we examined the contribution of O-GlcNAc modifications to K⁺ channel regulation in hyperglycemia in mouse (Fig. 3a). HBP and its end-product UDP-GlcNAc are necessary for O-GlcNAc protein modifications. This pathway can be inhibited using the broad spectrum amidotransferase inhibitor 6-diazo-5-oxo-l-norleucine (DON) and the selective O-GlcNAc transferase (OGT) inhibitor OSMI-1. Strikingly, pretreatment with either DON or OSMI-1 prevented all acute hyperglycemia effects on K⁺ channels and had no effect in normoglycemia (Fig. 3b-f). Conversely, when O-GlcNAc removal by O-GlcNAcase (OGA) was inhibited using Thiamet-G (Thm-G, 6 min), the same K⁺ channel effects observed under hyperglycemia were also observed under normoglycemic conditions (Fig. 3b-f). These data suggest that O-GlcNAcylation is necessary and sufficient for the observed hyperglycemia-induced K⁺ channel regulation.

Fig. 3.

O-GlcNAc pathway mediates the K⁺ channel effects in hyperglycemia. a Schematic of O-GlcNAc modifications and targets for pharmacological interventions. b–f Inhibition of O-GlcNAc pathway using 6-diazo-5-oxo-l-norleucine (DON, 50 μmol/L) or the specific O-GlcNAc transferase (OGT) inhibitor OSMI-1 (50 μmol/L) prevented hyperglycemia-induced changes both in I_K1 density (b, c), I_to density (d), and I_to recovery kinetics (e, f) in mouse ventricular myocytes. Conversely, inhibiting the O-GlcNAc removal enzyme (O-Glc-NAcase, OGA) using Thiamet-G (Thm-G, 100 nmol/L) promoted glucose effects on K⁺ currents already in normoglycemia. Student’s paired t test; *p < 0.05, **p < 0.01, ***p < 0.001

Glucose dose–response effects and the impact of extracellular K⁺ concentration on I_K1

Next, we measured I_K1 in rabbit myocytes acutely exposed to increasing concentrations of glucose (osmolarity was matched using mannitol). I_K1 magnitudes in 5.5 mmol/L and 10 mmol/L glucose were indistinguishable, whereas 20 mmol/L glucose induced a nonsignificant trend towards increased I_K1, and 30 mmol/L glucose was sufficient to induce a significant increase in I_K1 (Fig. 4a, b). When we repeated the experiments in the presence of insulin (5 nmol/L) to increase glucose uptake in cardiomyocytes via GLUT4 [42], this only slightly promoted glucose effects on I_K1 (Fig. 4b). However, Thm-G markedly enhanced glucose effects on I_K1 (Fig. 4b). These data suggest that in healthy control myocytes, hyperglycemia can affect K⁺ channels in a graded manner which can be further shifted by enhanced glucose uptake or perturbation of the HBP (e.g., here by Thm-G). Indeed, the balance between addition and removal of O-GlcNAc is regulated through OGT and OGA, respectively, which may be impaired in DM [26].

Fig. 4.

Glucose dose-response on I_K1 and effect of changing extracellular K⁺ concentration. a Representative I_K1 traces measured using a ramp protocol (shown above) in the presence of increasing concentrations of glucose in rabbit ventricular myocytes. b Glucose dose–response curves on I_K1 at – 60 mV in control and following pretreatments with either insulin (5 nmol/L) or Thiamet-G (Thm-G, 100 nmol/L). c Representative I_K1 traces in hyperkalemia and hypokalemia. d Hyperkalemia and hypokalemia shifted the reversal potential of I_K1 (E_K) and the voltage, where outward I_K1 peaked. e Representative I_K1 traces in combined hyperkalemia and hyperglycemia. f Hyperglycemia markedly increased I_K1 in hyperkalemia and slightly increased in hypokalemia. Two-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001 vs. normoglycemia; ^†††p < 0.001 vs. control

I_K1 is also critically regulated by the level of extracellular K⁺, and it is known that both hypo- and hyperkalemia may frequently occur in diabetic patients [9]. Hypo- and hyperkalemia not only shifted the reversal potential (E_K) of I_K1, but also markedly altered I_K1 amplitude and the voltage, where outward I_K1 peaked (Fig. 4c, d). Hyperglycemia did not alter the measured E_K or V_m for peak outward I_K1 (Fig. 4d) but similarly increased I_K1 in both hyper- and hypokalemia (Fig. 4e, f).

CaMKII and PKC mediate the effects of hyperglycemia on K⁺ channels

We tested the contribution of protein kinases which may mediate the downstream effects of hyperglycemia and could be regulated by O-GlcNAc modifications (Fig. 5a). Cells were pretreated (for 15 min) with specific kinase inhibitors, including autocamtide-2-related inhibitory peptide (AIP), protein kinase inhibitor peptide (PKI), and bisindolylmaleimide II (BIM-II) to selectively inhibit CaMKII, protein kinase A (PKA), and PKC, respectively. AIP and PKI pre-incubations used myristoylated forms and the peptides were also included in pipette solution to optimize inhibition. At baseline, these inhibitors had no effect on K⁺ channels except for CaMKII inhibition using AIP, which slightly reduced I_K1 (Fig. 5b, c). However, AIP prevented the hyperglycemia effects on I_K1 and I_to recovery, but not on the reduction of I_to amplitude (Fig. 5b-f). PKI did not affect any of the measured K⁺ currents in hyperglycemia (Fig. 5b-f). Interestingly, the hyperglycemia-induced decrease in I_to amplitude was prevented only by the PKC inhibitor BIM-II (Fig. 5d). We further examined the involvement of PKC in regulating I_to using Go 6976 that selectively inhibits the Ca²⁺-dependent conventional PKC isoforms (PKCα and PKCβ; Fig. 5g). Cell pretreatment with Go 6976 did not change baseline I_to but prevented the hyperglycemia-induced reduction of I_to amplitude (Fig. 5h, i). Because PKC inhibitor staurosporine and derivates thereof may directly inhibit K⁺ channels (off-target effect) [45], we also used BIM-V, a structurally related but inactive analogue of these PKC inhibitors. BIM-V did not change I_to amplitude and did not prevent I_to reduction in hyperglycemia (Fig. 5h, i).

Fig. 5.

CaMKII and PKC mediate the hyperglycemia effects on K⁺ channels. a Schematic of potential protein kinase pathways that may mediate the hyperglycemia effects on K⁺ channels and their selective inhibitors. b–f Inhibition of CaMKII using the specific autocamp-tide-2 related inhibitory peptide (AIP, 1 μmol/L) prevented the hyperglycemia effects on I_K1 (b, c) and I_to recovery (e, f) in mouse ventricular myocytes. Selective inhibition of protein kinase C (PKC) using bisindolylmaleimide II (BIM-II, 100 nmol/L) prevented I_to reduction in hyperglycemia (d). Protein kinase A (PKA) inhibition using the selective protein kinase inhibitory peptide (PKI, 1 μmol/L) had no effect on K⁺ channels in hyperglycemia. g Schematic of PKC isoforms and inhibitors. h Go 6976 (100 nmol/L), a selective inhibitor of conventional, Ca²⁺-dependent PKC isoforms (PKCα and β) prevented I_to reduction in hyperglycemia. BIM-V (100 nmol/L), a structurally related but inactive analogue of PKC inhibitors had no effect on I_to in normal glucose and did not prevent I_to reduction in hyperglycemia. i I-V relationship of I_to following Go 6976 and BIM-V pretreatments. Two-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001 vs. normoglycemia; ^†p < 0.05 vs. control

NOX2-dependent ROS production mediates the effects of hyperglycemia on I_to

Increased production of ROS via NOX2 and mitochondrial mechanisms has been implicated in diabetic hyperglycemia [13]. Because ROS is known to regulate both CaMKII (by oxidizing 2 neighboring methionines, MM281/282) [10] and PKC [8], we tested the involvement of ROS in regulating K⁺ channels in acute hyperglycemia (Fig. 6a). Pretreatment of the cells with ROS scavengers l-glutathione (GSH) and N-acetylcysteine (NAC) slightly reduced baseline I_K1 but did not alter the hyperglycemia-induced I_K1 increase (Fig. 6b, c). This same lack of ROS scavenger effect was also true for the accelerated kinetics of I_to recovery (Fig. 6e, f), but ROS scavengers abolished the hyperglycemia-induced I_to amplitude reduction (Fig. 6d). NOX2-KO also abolished the effect of hyperglycemia on I_to amplitude, but not on I_K1 amplitude or I_to recovery kinetics (Fig. 6b-f). Oxidation-resistant CaMKIIδ-MM281/2VV knock-in showed normal baseline K⁺ channel phenotype and all acute hyperglycemia effects on K⁺ channels were present (Fig. 6b-s). These functional fingerprints in Figs. 5 and 6 indicate that acute hyperglycemia reduces I_to amplitude via a NOX2-ROS-PKC pathway, but that CaMKII mediates the I_K1 and I_to recovery effects, which are independent of ROS and CaMKII oxidation.

Fig. 6.

NOX2-dependent ROS production mediates I_to reduction in hyperglycemia. a Schematic signaling via reactive oxygen species (ROS) that may be involved in the hyperglycemia effects on K⁺ channels. b–f Pretreatment of mouse ventricular myocytes with ROS scavengers, reduced glutathione (GSH, 10 mmol/L) and N-acetylcysteine (NAC, 10 mmol/L), prevented I_to downregulation (d) but not I_K1 (b, c) and I_to recovery (e, f) effects in hyperglycemia. I_to in NADPH oxidase 2 (NOX2)-knockout was also resistant to acute hyperglycemia. Oxidation-resistant CaMKIIδ-MMVV knock-in was not protected against acute hyperglycemia effects on either I_K1 or I_to. Two-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001 vs. normoglycemia; ^†p < 0.05 vs. control

CaMKIIδ-S280 O-GlcNAcylation mediates hyperglycemia effects on I_to recovery and I_K1

We further investigated the effects of hyperglycemia in myocytes isolated from CaMKIIδ-KO, CaMKIIδ-OE, and O-GlcNAc-resistant CaMKIIδ-S280A mice. First, we tested the baseline K⁺ currents measured in normoglycemia in each CaMKII genotype to assess CaMKII-dependent chronic regulation (i.e., functional expression), then acute hyperglycemic effects therein to assess acute regulation (Fig. 7a).

Fig. 7.

CaMKII activation via S280 O-GlcNAcylation regulates I_K1 and I_to recovery in hyperglycemia. a Representative voltage-gated K⁺ currents and I_K1 traces in normoglycemia and hyperglycemia in CaMKIIδ-knockout (KO), overexpression (OE) and O-GlcNAc-resistant S280A knock-in mouse myocytes. b–j K⁺ current remodeling in CaMKIIδ-KO and OE in normoglycemia. CaMKIIδ-S280A had a normal baseline phenotype. Hyperglycemia effects on I_to recovery kinetics (f, g) and I_K1 (h, i) but not on I_to amplitude (b) were prevented in both CaMKIIδ-KO and S280A. Two-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001 vs. normoglycemia; ^†p < 0.05, ^††p < 0.01, ^†††p < 0.001 vs. WT control

CaMKIIδ-KO exhibited substantially larger amplitudes of I_to, I_K,slow and I_SS, faster I_to inactivation, and faster I_to recovery from inactivation vs. WT (Fig. 7b-g). Conversely, CaMKIIδ-OE reduced I_to and I_SS but enhanced the slow component of I_to recovery vs. WT, whereas I_K,slow and I_to inactivation were unchanged (Fig. 7b-g). I_K1 was slightly increased in CaMKIIδ-KO and markedly reduced in CaMKIIδ-OE (Fig. 7h, i). The small I_Kr in mouse was further reduced in CaMKIIδ-OE (Fig. 6J). Most of these baseline effects are consistent with the expected directions of effect for chronic inhibition or enhancement of CaMKII [48]. Importantly, all K⁺ currents were unchanged in CaMKIIδ-S280A at baseline (note dashed horizontal lines for WT control; Fig. 7a-j).

Acute hyperglycemia reduced I_to amplitude in all CaMKIIδ genotypes (Fig. 7b), in line with the data obtained using CaMKII inhibitor AIP (Fig. 5d), and that CaMKII is not involved in acute reduction of I_to amplitude. However, the acute hyperglycemia effects on I_to recovery (Fig. 7f, g) and I_K1 amplitude (Fig. 7h, i) were prevented in CaMKIIδ-KO, confirming the AIP data (Fig. 5b, c) and CaMKIIδ involvement. Importantly, these CaMKII-mediated hyperglycemia effects were also prevented in CaMKIIδ-S280A (Fig. 7f-i) indicating the critical role of CaMKIIδ-S280 O-GlcNAcylation in autonomously activating CaMKII in hyperglycemia. Moreover, the hyperglycemia-induced reduction in I_to amplitude that seemed CaMKII-independent above, was almost identical in the CaMKIIδ-S280A mouse to that observed in WT mice (Fig. 7b). Thus, the mediation of these hyperglycemia-induced changes in I_to and I_K1 seems very consistent with respect to molecular mechanism, involving either O-GlcNAcylation at CaMKIIδ-S280 and CaMKII-dependent phosphorylation (I_K1 and I_to recovery kinetics) and O-GlcNAc-dependent activation of NOX2 and PKC-dependent phosphorylation (I_to amplitude reduction).

Chronic CaMKII activation downregulates K⁺ channel expression

Because chronic CaMKII activation led to hypertrophic cardiac remodeling and changes the functional expression of K⁺ channels (Fig. 7), we examined which genes were responsible for the K⁺ current remodeling. Using qRT-PCR we found markedly increased expression of hypertrophic genes (Anf, Myh7) in CaMKIIδ-OE (having ~ 50-fold higher Camk2d transcript level, Fig. 8a and Suppl. Fig.2). In CaMKIIδ-OE, several K⁺ channel genes were downregulated, including Kcnd2, Kcnd3, Kcnip2, Kcna5, Kcnk3, Kcnj2, Kcnh2a, Kcnh2b, and Kcnq1, whereas Kcne1 was upregulated (Fig. 8b). Interestingly, some of these K⁺ channel genes were reciprocally regulated in CaMKIIδ-KO, including upregulation of Kcnd3, Kcnb1 and Kcnk3, and downregulation of Kcne1 (Fig. 8b). However, the expression of Kcnd2, Kcnip2, Kcna5, Kcnh2a, Kcnh2b and Kcnq1 was unchanged in CaMKIIδ-KO (Fig. 8b). No change in the expression of any genes was found in CaMKIIδ-S280A (Fig. 8b), in line with its normal baseline phenotype (Fig. 8a). The protein expression profile of K_ir2.1 (predominant I_K1 channel isoform) closely followed its gene (Kcnj2) expression level with marked reduction in CaMKIIδ-OE, a nonsignificant increasing trend in CaMKIIδ-KO, and no change in CaMKIIδ-S280A (Fig. 8c). These expression data are in line with the K⁺ current densities in CaMKIIδ mutants. Moreover, these data provide insights on which K⁺ channels are responsible for the observed differences in K⁺ current densities in CaMKIIδ-KO and CaMKIIδ-OE (Fig. 7).

Fig. 8.

CaMKII regulates cardiac hypertrophy and K⁺ channel expression. a qRT-PCR data on expression of Camk2d, Camk2g, and hypertrophic genes (Anf, Myh7) and morphometric parameters in CaMKIIδ-knockout (KO), overexpression (OE) and O-GlcNAc-resistant S280A knock-in mouse hearts. b mRNA expression of K⁺ channel genes normalized to WT control. c Western blot analysis of K_ir2.1 expression normalized to WT control. ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001

K⁺ channel functional expression and regulation in chronic diabetes

Next, we measured K⁺ currents in three diabetic animal models (Fig. 9a) including STZ-induced T1DM vs. vehicle-injected controls [21, 35], HFD-induced T2DM vs. LFD [32, 43], and db/db vs. WT C57BLKS/J [2, 20], see “Animal models” in “Methods” for details. Diabetic animals exhibited increased blood glucose levels and moderate cardiomyocyte hypertrophy (Suppl. Fig. 3a, b). The K⁺ channel remodeling was studied in an early phase of the disease (e.g., 4-week post-STZ) with preserved cardiac systolic contractile function (no change in fractional shortening and ventricular diameters) as measured by echocardiogram (Suppl. Table 2).

Fig. 9.

Diabetes induces remodeling in K⁺ currents in mouse ventricular myocytes. a Representative voltage-gated K⁺ currents and I_K1 traces in normoglycemia and hyperglycemia in streptozotocin (STZ)-induced T1DM, high-fat diet (HFD)-induced T2DM, and db/db mouse myocytes. b–g Decreases of K⁺ currents in DM vs. corresponding controls: vehicle-injection for STZ, low-fat diet (LFD) for HFD and C57BLKS/J for db/db. Hyperglycemia further reduced I_to and increased I_K1 in STZ (and all controls), but not in HFD and db/db. h Representative I_K1 traces measured at – 150 mV in STZ-treated WT, O-GlcNAc-resistant CaMKIIδ-S280A knock-in, and vehicle-treated WT control. i The decrease in I_K1 density was markedly attenuated in STZ-treated CaMKIIδ-S280A. j The cell capacitance correlated with the decrease in I_K1 density in STZ-treated WT, whereas no correlation was found in WT control and STZ-treated CaMKIIδ-S280A. Lines indicate linear regression. Two-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001 vs. normoglycemia; ^†p < 0.05, ^††p < 0.01, ^†††p < 0.001 vs. control

Importantly, all K⁺ currents, including I_to, I_K,slow, I_SS, I_K1 and I_Kr, were already reduced in all three diabetic models in this early, cardiac-compensated phase of DM. The largest reduction in K⁺ currents was found in the STZ model (Fig. 9b-g), similar to that observed in CaMKIIδ-OE (Fig. 7). STZ mice showed no change in I_to inactivation kinetics and only a slight enhancement of I_to slow recovery (Suppl. Fig. 3c-e). However, in STZ myocytes acute hyperglycemia (as occur in vivo in these diabetic animals) induced a further reduction in I_to (Fig. 9b), but increased I_K1 (Fig. 9e, f) and further enhanced I_to recovery (Suppl. Fig. 3d, e). This recapitulates the situation that there is still dynamic hyperglycemic modulation of channel properties that rides on top of changes in expression in the STZ model. Surprisingly, cardiomyocytes from HFD and db/db were resistant to the effects of acute hyperglycemia on both I_to amplitude and recovery, and I_K1 magnitude (Fig. 9b, e, f, and Suppl. Fig. 3d, e).

Next, we tested whether CaMKIIδ-S280A is also protective against chronic hyperglycemia effects. Due to limited number of animals available and parallel uses of cells in different studies, we focused on the decreased I_K1 density, because it was the most pronounced alteration in DM. Importantly, I_K1 reduction was markedly attenuated in STZ-treated CaMKIIδ-S280A (Fig. 9h, i). In individual cells, the membrane capacitance (an indirect measure of cell size) was correlated with decreasing I_K1 density in STZ-treated WT, but this relationship was absent in WT control and STZ-treated CaMKIIδ-S280A (Fig. 9j). This data suggests that CaMKIIδ-S280 O-GlcNAcylation in DM is responsible for mediating both hypertrophic and I_K1 functional effects at a single cell level.

K⁺ channel functional expression and hyperglycemic response in heart failure

HF is a frequent comorbidity in diabetic patients, and diabetes increases the risk of newly developing HF [9]. CaMKII is also upregulated in HF and contributes to maladaptive cardiac remodeling and arrhythmias [23]. We, therefore, tested the effect of hyperglycemia in a mouse model of HF. We used the well-established TAC model in mice to induce cardiac hypertrophy and HF. 8-week post-surgery mice exhibited 45% decrease in fractional shortening, 30% increase in systolic ventricular diameter, 70% increase in heart weight/body weight ratio, and pulmonary edema (Suppl. Table 3). The voltage-gated K⁺ currents were all significantly decreased in TAC vs. sham (Fig. 10a-c), similar to that seen in CaMKIIδ-OE (Fig. 7) and DM (Fig. 9). The expression of the K⁺ channels was also downregulated in TAC-induced HF (Fig. 10d) in accordance with the changes in K⁺ currents. The K⁺ channel expression profile in TAC (in which Camk2d was twofold upregulated) largely overlapped with that of CaMKIIδ-OE. However, a few differences existed, including Kcna4 and Kcne1 downregulation in TAC mice (Fig. 10d) vs. their upregulation in CaMKIIδ-OE mice (Fig. 8c). This suggests additional CaMKIIδ-independent K⁺ channel regulatory mechanisms in HF. Moreover, acute hyperglycemia further reduced I_to, but enhanced its recovery and increased the markedly decreased I_K1 in TAC (Fig. 10a-c).

Fig. 10.

K⁺ channel remodeling and hyperglycemia effects in heart failure. a Representative voltage-gated K⁺ currents and I_K1 traces in normoglycemia and hyperglycemia in transverse aortic constriction (TAC)-induced heart failure (8-week post-TAC) vs. sham mouse myocytes. b K⁺ current remodeling in TAC. Hyperglycemia further reduced I_to in TAC. c Hyperglycemia slightly enhanced I_to recovery kinetics and increased the markedly downregulated I_K1 in TAC. d Normalized qRT-PCR data on expression of Camk2d and Camk2g, hypertrophic markers (Anf, Myh7) and K⁺ channel genes. ANOVA; *p < 0.05. **p < 0.01, ***p < 0.001 vs. normoglycemia; ^†p < 0.05, ^††p < 0.01, ^†††p < 0.001 vs. sham

Discussion

The increased arrhythmia risk and the lack of effective antiarrhythmic therapies in DM highlight the need for a better understanding of the molecular underpinnings of arrhythmogenesis in this disease. The pathophysiology of DM is undisputedly complex, but K⁺ channel remodeling is strongly implicated as a main driver of arrhythmogenicity [15]. K⁺ channel downregulation contributes to action potential and QT prolongation frequently observed in patients with DM [4], enhancing arrhythmogenic substrate in the heart. Moreover, due to the increased membrane resistance when I_K1 is reduced, a given inward current can cause larger depolarizations, promoting triggered activities [34]. Previous reports in line with our data here (Fig. 9) showed decreases in several K⁺ currents in DM, including I_to, I_K,slow, I_K1, I_Kr, and I_Ks in accordance with the reduced expression of K_V4.2 (Kcnd2), K_V4.3 (Kcnd3), KChIP2 (Kcnip2), K_V1.5 (Kcna5), K_ir2.1 (Kcnj2), hERG/K_V11.1 (Kcnh2), and K_V7.1/minK (Kcnq1/Kcne1), respectively [22, 27, 29, 35, 39, 51, 54].

Acute hyperglycemia may also increase arrhythmia risk [15]. In healthy human volunteers, hyperglycemic-clamp to 15 mmol/L plasma glucose for 2 h slightly prolonged QTc interval (from 413 to 442 ms) and markedly increased QTc dispersion (from 32 to 55 ms) independent of insulin action [28]. Clinical data in diabetic patients also demonstrates that postprandial hyperglycemia and glucose-variability are independent risk factors of cardiovascular disease [6]. Moreover, patients with T2DM had a slightly higher risk of ventricular premature beats during their hyperglycemic vs. normoglycemic periods [7]. Furthermore, hyperglycemia is also strongly associated with increased risk of early-onset ventricular tachycardia following myocardial infarction both in diabetic and non-diabetic patients [46]. However, the underlying molecular mechanisms for promoting arrhythmias in either chronic or acute hyperglycemia are still incompletely understood.

Acute hyperglycemia regulates K⁺ currents via O-GlcNAcylation and CaMKII and PKC

Existing work on the mechanisms of acute hyperglycemic alteration of K⁺ channels is sparse and focused only on ATP-sensitive K⁺ channels in cardiomyocytes [19] and large-conductance Ca²⁺-activated K⁺ channels in arterial smooth muscle cells [32]. Here, in three species, we found that acute hyperglycemia significantly altered I_to and I_K1 in ventricular cardiomyocytes, which are known to critically influence AP morphology and arrhythmias. However, the acute reduction in I_to amplitude observed may be partially compensated by enhanced I_to recovery kinetics at higher heart rates. Moreover, the increased I_K1 observed may provide some protection against hyperglycemia-induced diastolic Ca²⁺ leak and afterdepolarizations, another characteristic consequence of CaMKII activation [11, 15]. However, changes in [K⁺]_o (Fig. 4) and chronic remodeling in K⁺ channels in DM (Fig. 9) and HF (Fig. 10) may shift the relative K⁺ channel balance (mostly downregulation) to promote arrhythmias.

Signaling mechanisms contributing to K⁺ channel regulation in acute and chronic hyperglycemia are poorly characterized, but crucial for potential therapeutic targeting in DM. We found that the hyperglycemic effects on both I_to and I_K1 were mediated by the HBP and in particular via O-GlcNAcylation (Figs. 1, 2, 3). Furthermore, two separate pathways mediated these effects: I_to downregulation was dependent on the NOX2-ROS-PKC signaling axis (Figs. 5, 6), while the enhanced I_to recovery and I_K1 amplitude were dependent entirely on activation of CaMKIIδ by S280 O-GlcNAcylation (Fig. 7). Moreover, the acute CaMKII-dependent enhancement of I_K1 amplitude and I_to recovery were maintained in CaMKIIδ-OE, T1DM and HF, and in altered [K⁺]_o (for I_K1) and mimicked simply by inhibition of O-GlcNAcase. However, these acute hyperglycemic effects were abolished by inhibition of CaMKII or O-GlcNAc transferase and in CaMKIIδ-KO, -S280A mutant and T2DM (both HFD and db/db, Fig. 9). The underlying mechanism for this loss of acute hyperglycemic I_K1 enhancement in T2DM is unclear, but may include altered glucose uptake, insulin signaling and cell metabolism, or differential baseline protein O-GlcNAcylation [26].

Environmental factors such as hyperglycemia and oxidative stress can lead to posttranslational modifications and play an important role in the pathophysiology of DM [13, 26]. O-GlcNAcylation was demonstrated previously to activate CaMKIIδ [11] and O-GlcNAc modifications may also regulate PKCα, yet the level of O-GlcNAc-modified PKCα correlated negatively with its activity [37]. Nevertheless, oxidative stress activates both CaMKII [10] and PKC [12]. Moreover, both CaMKII [24, 31] and PKC [17] may further induce NOX2 and increase ROS production, thus forming a vicious positive feedback cycle in DM-induced progression of cardiac dysfunction.

Prior works on PKC and CaMKII-dependent regulation of K⁺ channels are in line with our data (Fig. 5). PKC activation was found to reduce I_to in adult rat cardiomyocytes and both K_V4.2 and K_V4.3 currents expressed in Xenopus oocytes without significant changes in channel biophysics [30]. PKC was further shown to phosphorylate the C-terminus (T503) of the long isoform of K_V4.3 (but not the short splice variant), and reduce I_to in human cardiomyocytes [33]. We found that this PKC-mediated I_to reduction in diabetic hyperglycemia was critically dependent on oxidative stress (Fig. 6) in agreement with an earlier report [51]. CaMKII, on the other hand, was previously demonstrated to enhance I_to recovery [48] by phosphorylating K_V4.3 at S550 [41] and K_V1.4 at S123 [38], suggesting a mechanistic basis for hyperglycemia-induced enhancement of both the fast (K_V4.2/4.3) and slow (K_V1.4) components of I_to recovery seen in Figs. 1, 2. Acute I_K1 enhancement by CaMKII activation (Figs. 1, 2, 3) was also previously reported [16, 48]; however, the exact phosphorylation site of CaMKII on K_ir2.1 has not yet been identified. CaMKII may also acutely increase the slow delayed rectifier K⁺ current (I_Ks) [16], the ATP-sensitive K⁺ current [52] and the Ca²⁺-activated K⁺ current [44]. The regulation of these channels in hyperglycemia requires further studies.

Chronic K⁺ channel expression and current reduction with CaMKIIδ, DM and HF

K⁺ channel remodeling during T1DM and T2DM development in mouse models (Fig. 9) resembled that seen in CaMKIIδ-OE (Figs. 7, 8) and HF (Fig. 10), which is consistent with CaMKII involvement in the regulation of K⁺ channel expression in DM at the level of gene expression/transcription. We show direct evidence for this mechanism in that the decrease in I_K1 following STZ-induced T1DM was markedly attenuated in CaMKIIδ-S280A (Fig. 9). We also demonstrated a correlation between I_K1 reduction and increasing cell capacitance in individual cells following STZ treatment (Fig. 9). This relationship was abolished in STZ-treated CaMKIIδ-S280A mice, suggesting that CaMKII O-GlcNAcylation initiates myocyte transcriptional remodeling and expression of a hypertrophic gene program [53]. This hypothesis should be explored in future studies.

Upon chronic activation, both PKC [47] and CaMKII [49] phosphorylate class II histone deacetylases (HDACs) which lead to their nuclear export and altered gene transcription. Accordingly, CaMKIIδ-OE was shown here (Fig. 8) and previously [48] to decrease I_K1 and I_to, in line with the downregulation of mRNA and protein levels of K_V4.2 (Kcnd2), K_V4.3 (Kcnd3), KChIP2 (Kcnip2) and K_ir2.1 (Kcnj2). CaMKIIδ-OE also downregulated I_K,slow, I_SS and I_Kr in line with the decreased expression of Kcna5, Kcnk3 and Kcnh2a/2b. Expression of I_Ks channel mRNAs were also altered in CaMKIIδ-OE with reduction of the pore-forming Kcnq1 subunit and an increase in the β subunit Kcne1 (Fig. 8). Interestingly, reduced Kcne1 expression mediated by the Ca²⁺/calcineurin/NFAT axis has been described following sustained β-adrenergic stimulation [1]. In line with this finding, in TAC-induced HF we observed a tendency for reduced Kcne1 expression (Fig. 10).

The present study has advanced fundamental understanding of the mechanisms of both acute and chronic K⁺ channel alterations in hyperglycemia and diabetes. The broad downregulation of K⁺ channels in both DM and HF and the role of CaMKII paints a picture of reduced repolarization reserve that may contribute similarly to increased arrhythmogenic propensity in both pathologies. Due to altered myocyte glucose metabolism in HF, enhanced CaMKII O-GlcNAcylation has also been observed in the hearts of non-diabetic HF patients, and even higher levels of O-GlcNAcylated CaMKII has been observed in HF patients with DM [11]. Moreover, CaMKII O-GlcNAcylation has also been shown in brain samples of diabetic patients [11] and may occur in other organs as well. Overall, these data highlight the need for future studies to characterize the role of CaMKII O-GlcNAcylation in the pathogenesis of both cardiac and other systemic complications of DM.

Conclusions

We found that O-GlcNAcylation of CaMKII at S280 enhanced I_K1 and I_to recovery from inactivation during acute hyperglycemia. However, chronic CaMKII activation in diabetes and HF significantly downregulated the functional expression of most K⁺ channels. Additionally, a NOX2-ROS-PKC pathway significantly decreased I_to amplitude in hyperglycemia. These data suggest that while acute hyperglycemia effects on K⁺ channels may be compensated for in a healthy heart, chronic hyperglycemia and CaMKII activation can cause significant arrhythmogenic electrophysiological remodeling. Thus, a better understanding of mechanisms of K⁺ channel regulation in diabetic hyperglycemia may elucidate specific downstream targets amenable to precision medicine interventions, enabling design of better antiarrhythmic therapeutics in DM.

Data availability

All data and materials are available on reasonable request to the corresponding author.

Abbreviations

4-AP

4-Aminopyridine

AIP

Autocamtide-2-related inhibitory peptide

ANF

Atrial natriuretic factor

BIM

Bisindolylmaleimide

CaMKII

Ca²⁺/calmodulin-dependent kinase II

DM

Diabetes mellitus

DON

6-Diazo-5-oxo-l-norleucine

HBP

Hexosamine biosynthetic pathway

HF

Heart failure

HFD

High-fat diet

I_Kr

Rapid delayed rectifier K⁺ current

I_Ks

Slow delayed rectifier K⁺ current

I_Kslow

Slowly inactivating K⁺ current

I_K1

Inward rectifier K⁺ current

I_SS

Steady-state K⁺ current

I_to

Transient outward K⁺ current

LFD

Low-fat diet

NOX2

NADPH oxidase 2

Myh7

β-Myosin heavy chain

OE

Overexpression

O-GlcNAc

O-Linked β-N-acetylglucosamine

OGA

O-GlcNAcase

OGT

O-GlcNAc transferase

PKI

Protein kinase inhibitor

PKC

Protein kinase C

ROS

Reactive oxygen species

STZ

Streptozotocin

TAC

Transverse aortic constriction

Thm-G

Thiamet-G

T1DM

Type 1 diabetes mellitus

T2DM

Type 2 diabetes mellitus