Protease XIV abolishes NHE inhibition by empagliflozin in cardiac cells

Background: SGLT2i directly inhibit the cardiac sodium-hydrogen exchanger-1 (NHE1) in isolated ventricular cardiomyocytes (CMs). However, other studies with SGLT2i have yielded conflicting results. This may be explained by methodological factors including cell isolation techniques, cell types and ambient pH. In this study, we tested whether the use of protease XIV (PXIV) may abrogate inhibition of SGLT2i on cardiac NHE1 activity in isolated rabbit CMs or rat cardiomyoblast cells (H9c2), in a pH dependent manner. Methods: Rabbit ventricular CMs were enzymatically isolated from Langendorff-perfused hearts during a 30-min perfusion period followed by a 25-min after-dissociation period, using a collagenase mixture without or with a low dose PXIV (0.009 mg/mL) present for different periods. Empagliflozin (EMPA) inhibition on NHE activity was then assessed at pH of 7.0, 7.2 and 7.4. In addition, effects of 10 min PXIV treatment were also evaluated in H9c2 cells for EMPA and cariporide NHE inhibition. Results: EMPA reduced NHE activity in rabbit CMs that were not exposed to PXIV treatment or undergoing a 35-min PXIV treatment, independent of pH levels. However, when exposure time to PXIV was extended to 55 min, NHE inhibition by Empa was completely abolished at all three pH levels. In H9c2 cells, NHE inhibition by EMPA was evident in non-treated cells but lost after 10-min incubation with PXIV. NHE inhibition by cariporide was unaffected by PXIV. Conclusion: The use of protease XIV in cardiac cell isolation procedures obliterates the inhibitory effects of SGLT2i on NHE1 activity in isolated cardiac cells, independent of pH.

Direct Cardiac Actions of Sodium Glucose Cotransporter 2 Inhibitors Target Pathogenic Mechanisms Underlying Heart Failure in Diabetic Patients

Sodium glucose cotransporter 2 inhibitors (SGLT2i) are the first antidiabetic compounds that effectively reduce heart failure hospitalization and cardiovascular death in type 2 diabetics. Being explicitly designed to inhibit SGLT2 in the kidney, SGLT2i have lately been investigated for their off-target cardiac actions. Here, we review the direct effects of SGLT2i Empagliflozin (Empa), Dapagliflozin (Dapa), and Canagliflozin (Cana) on various cardiac cell types and cardiac function, and how these may contribute to the cardiovascular benefits observed in large clinical trials. SGLT2i impaired the Na⁺/H⁺ exchanger 1 (NHE-1), reduced cytosolic [Ca²⁺] and [Na⁺] and increased mitochondrial [Ca²⁺] in healthy cardiomyocytes. Empa, one of the best studied SGLT2i, maintained cell viability and ATP content following hypoxia/reoxygenation in cardiomyocytes and endothelial cells. SGLT2i recovered vasoreactivity of hyperglycemic and TNF-α-stimulated aortic rings and of hyperglycemic endothelial cells. Anti-inflammatory actions of Cana in IL-1β-treated HUVEC and of Dapa in LPS-treated cardiofibroblast were mediated by AMPK activation. In isolated mouse hearts, Empa and Cana, but not Dapa, induced vasodilation. In ischemia-reperfusion studies of the isolated heart, Empa delayed contracture development during ischemia and increased mitochondrial respiration post-ischemia. Direct cardiac effects of SGLT2i target well-known drivers of diabetes and heart failure (elevated cardiac cytosolic [Ca²⁺] and [Na⁺], activated NHE-1, elevated inflammation, impaired vasorelaxation, and reduced AMPK activity). These cardiac effects may contribute to the large beneficial clinical effects of these antidiabetic drugs.

Class effects of SGLT2 inhibitors in mouse cardiomyocytes and hearts: inhibition of Na+/H+ exchanger, lowering of cytosolic Na+ and vasodilation

AIMS/HYPOTHESIS

Sodium-glucose cotransporter 2 (SGLT2) inhibitors (SGLT2i) constitute a novel class of glucose-lowering (type 2) kidney-targeted agents. We recently reported that the SGLT2i empagliflozin (EMPA) reduced cardiac cytosolic Na⁺ ([Na⁺]_c) and cytosolic Ca²⁺ ([Ca²⁺]_c) concentrations through inhibition of Na⁺/H⁺ exchanger (NHE). Here, we examine (1) whether the SGLT2i dapagliflozin (DAPA) and canagliflozin (CANA) also inhibit NHE and reduce [Na⁺]_c; (2) a structural model for the interaction of SGLT2i to NHE; (3) to what extent SGLT2i affect the haemodynamic and metabolic performance of isolated hearts of healthy mice.

METHODS

Cardiac NHE activity and [Na⁺]_c in mouse cardiomyocytes were measured in the presence of clinically relevant concentrations of EMPA (1 μmol/l), DAPA (1 μmol/l), CANA (3 μmol/l) or vehicle. NHE docking simulation studies were applied to explore potential binding sites for SGTL2i. Constant-flow Langendorff-perfused mouse hearts were subjected to SGLT2i for 30 min, and cardiovascular function, O₂ consumption and energetics (phosphocreatine (PCr)/ATP) were determined.

RESULTS

EMPA, DAPA and CANA inhibited NHE activity (measured through low pH recovery after NH₄⁺ pulse: EMPA 6.69 ± 0.09, DAPA 6.77 ± 0.12 and CANA 6.80 ± 0.18 vs vehicle 7.09 ± 0.09; p < 0.001 for all three comparisons) and reduced [Na⁺]_c (in mmol/l: EMPA 10.0 ± 0.5, DAPA 10.7 ± 0.7 and CANA 11.0 ± 0.9 vs vehicle 12.7 ± 0.7; p < 0.001). Docking studies provided high binding affinity of all three SGLT2i with the extracellular Na⁺-binding site of NHE. EMPA and CANA, but not DAPA, induced coronary vasodilation of the intact heart. PCr/ATP remained unaffected.

CONCLUSIONS/INTERPRETATION

EMPA, DAPA and CANA directly inhibit cardiac NHE flux and reduce [Na⁺]_c, possibly by binding with the Na⁺-binding site of NHE-1. Furthermore, EMPA and CANA affect the healthy heart by inducing vasodilation. The [Na⁺]_c-lowering class effect of SGLT2i is a potential approach to combat elevated [Na⁺]_c that is known to occur in heart failure and diabetes.

Empagliflozin decreases myocardial cytoplasmic Na+ through inhibition of the cardiac Na+/H+ exchanger in rats and rabbits

AIMS/HYPOTHESIS

Empagliflozin (EMPA), an inhibitor of the renal sodium-glucose cotransporter (SGLT) 2, reduces the risk of cardiovascular death in patients with type 2 diabetes. The underlying mechanism of this effect is unknown. Elevated cardiac cytoplasmic Na⁺ ([Na⁺]_c) and Ca²⁺ ([Ca²⁺]_c) concentrations and decreased mitochondrial Ca²⁺ concentration ([Ca²⁺]_m) are drivers of heart failure and cardiac death. We therefore hypothesised that EMPA would directly modify [Na⁺]_c, [Ca²⁺]_c and [Ca²⁺]_m in cardiomyocytes.

METHODS

[Na⁺]_c, [Ca²⁺]_c, [Ca ²⁺]_m and Na⁺/H⁺ exchanger (NHE) activity were measured fluorometrically in isolated ventricular myocytes from rabbits and rats.

RESULTS

An increase in extracellular glucose, from 5.5 mmol/l to 11 mmol/l, resulted in increased [Na⁺]_c and [Ca²⁺]_c levels. EMPA treatment directly inhibited NHE flux, caused a reduction in [Na⁺]_c and [Ca²⁺]_c and increased [Ca²⁺]_m. After pretreatment with the NHE inhibitor, Cariporide, these effects of EMPA were strongly reduced. EMPA also affected [Na⁺]_c and NHE flux in the absence of extracellular glucose.

CONCLUSIONS/INTERPRETATION

The glucose lowering kidney-targeted agent, EMPA, demonstrates direct cardiac effects by lowering myocardial [Na⁺]_c and [Ca²⁺]_c and enhancing [Ca²⁺]_m, through impairment of myocardial NHE flux, independent of SGLT2 activity.

Atrial fibrosis and conduction slowing in the left atrial appendage of patients undergoing thoracoscopic surgical pulmonary vein isolation for atrial fibrillation

BACKGROUND

Atrial fibrosis is an important component of the arrhythmogenic substrate in patients with atrial fibrillation (AF). We studied the effect of interstitial fibrosis on conduction velocity (CV) in the left atrial appendage of patients with AF.

METHODS AND RESULTS

Thirty-five left atrial appendages were obtained during AF surgery. Preparations were superfused and stimulated at 100 beats per minute. Activation was recorded with optical mapping. Longitudinal CV (CVL), transverse CV (CVT), and activation times (> 2 mm distance) were measured. Interstitial collagen was quantified and graded qualitatively. The presence of fibroblasts and myofibroblasts was assessed immunohistochemically. Mean CVL was 0.55 ± 0.22 m/s, mean CVT was 0.25 ± 0.15 m/s, and the mean activation time was 9.31 ± 5.45 ms. The amount of fibrosis was unrelated to CV or patient characteristics. CVL was higher in left atrial appendages with thick compared with thin interstitial collagen strands (0.77 ± 0.22 versus 0.48 ± 0.19 m/s; P = 0.012), which were more frequently present in persistent patients with AF. CVT was not significantly different (P = 0.47), but activation time was 14.93 ± 4.12 versus 7.95 ± 4.12 ms in patients with thick versus thin interstitial collagen strands, respectively (P = 0.004). Fibroblasts were abundantly present and were associated with the presence of thick interstitial collagen strands (P = 0.008). Myofibroblasts were not detected in the left atrial appendage.

CONCLUSIONS

In patients with AF, thick interstitial collagen strands are associated with higher CVL and increased activation time. Our observations demonstrate that the severity and structure of local interstitial fibrosis is associated with atrial conduction abnormalities, presenting an arrhythmogenic substrate for atrial re-entry.

The Driving Force of the Na/Ca-Exchanger during Metabolic Inhibition

OBJECTIVE

Metabolic inhibition causes a decline in mechanical performance and, if prolonged, myocardial contracture and cell death. The decline in mechanical performance is mainly due to altered intracellular calcium handling, which is under control of the Na(+)/Ca(2+)-exchanger (NCX) The driving force of the NCX (ΔG(ncx)) determines the activity of NCX. The aim of this study was to describe the relation between ΔG(ncx) and calcium homeostasis during metabolic inhibition.

METHODS

In left ventricular rabbit myocytes, during metabolic inhibition (2 mmol/L sodium cyanide), sodium ([Na(+)](i)), calcium ([Ca(2+);](i)), and action potentials were determined with SBFI, indo-1, and the patch clamp technique. Changes of ΔG(ncx) were calculated.

RESULTS

During metabolic inhibition: The first 8 min [Na(+)](i) remained constant, systolic calcium decreased from 532 ± 28 to 82 ± 13 nM, diastolic calcium decreased from 121 ± 12 to 36 ± 10 nM and the sarcoplasmic reticulum (SR) calcium content was depleted for 85 ± 3%. After 8 min [Na(+);](i) and diastolic calcium started to increase to 30 ± 1.3 mmol/L and 500 ± 31 nM after 30 min respectively. The action potential duration shortened biphasically. In the first 5 min it shortened from 225 ± 12 to 153 ± 11 ms and remained almost constant until it shortened again after 10 min. After 14 min action potential and calcium transients disappeared due to unexcitability of the myocytes. This resulted in an increased of the time average of ΔG(ncx) from 6.2 ± 0.2 to 7.7 ± 0.3 kJ/mol during the first 3 min, where after it decreased and became negative after about 15 min.

CONCLUSION

Metabolic inhibition caused an early increase of ΔG(ncx) caused by shortening of the action potential. The increase of ΔG(ncx) contributed to decrease of diastolic calcium, calcium transient amplitude, SR calcium content, and contractility. The increase of diastolic calcium started after ΔG(ncx) became lower than under aerobic conditions.

Incorporated sarcolemmal fish oil fatty acids shorten pig ventricular action potentials

BACKGROUND

Omega-3 polyunsaturated fatty acids (omega3-PUFAs) from fish oil reduce the risk of sudden death presumably by preventing life-threatening arrhythmias. Acutely administered omega3-PUFAs modulate the activity of several cardiac ion channels, but the chronic effects of a diet enriched with fish oil leading to omega3-PUFA-incorporation into the sarcolemma on membrane currents are unknown.

METHODS

Pigs received a diet either rich in omega3-PUFAs or in omega9-fatty acids for 8 weeks. Ventricular myocytes (VMs) were isolated and used for patch-clamp studies.

RESULTS

omega3-VMs contained higher amounts of omega3-PUFAs and had a shorter action potential (AP) with a more negative plateau than control VM. In omega3 VMs, L-type Ca(2+) current (I(Ca,L)) and Na(+)-Ca(2+) exchange current (I(NCX)) were reduced by approximately 20% and 60%, respectively, and inward rectifier K(+) current (I(K1)) and slow delayed rectifier K(+) current (I(Ks)) were increased by approximately 50% and 70%, respectively, compared to control. Densities of rapid delayed rectifier K(+) current, Ca(2+)-activated Cl(-) current, and Na(+) current (I(Na)) were unchanged, although voltage-dependence of I(Na) inactivation was more negative in omega3 VMs.

CONCLUSIONS

A fish oil diet increases omega3-PUFA content in the ventricular sarcolemma, decreases I(Ca,L) and I(NCX), and increases I(K1) and I(Ks), resulting in AP shortening. Incorporation of omega3-PUFAs in the sarcolemma may have consequences for arrhythmias independent of circulating omega3-PUFAs.

Chronic inhibition of Na/H-exchanger attenuates cardiac hypertrophy and prevents cellular remodeling in heart failure

OBJECTIVE

In patients with heart disease, the transition from compensatory hypertrophy to heart failure (HF) is associated with altered calcium handling. Up-regulated Na(+)/H(+)-exchanger (NHE-1) activity underlies increased [Na(+)](i) and disturbance of cellular calcium handling in HF. We hypothesize that chronic inhibition of NHE-1 activity prevents the hypertrophic response, cellular remodeling, and development of HF.

METHODS

Rabbits received a control or cariporide (inhibitor of NHE-1) diet for 3 months, starting after induction of combined volume and pressure overload. Age-matched animals served as control. Development of HF was examined echocardiographically and electrocardiographically after 3 months. [Na(+)](i), [Ca(2+)](i), pH(i), and action potentials were measured in left ventricular midmural myocytes with SBFI, indo-1, SNARF, and di-4-anepps. Sarcoplasmic reticulum calcium content was calculated from the response of [Ca(2+)](i) to rapid cooling. Calcium after-transients were elicited by cessation of rapid stimulation (3 Hz) in the presence of 100 nmol/l noradrenalin.

RESULTS

Chronic treatment of rabbits with the specific Na(+)/H(+)-exchanger activity inhibitor cariporide greatly attenuated development of hypertrophy and entirely abolished development of HF; the heart/body weight ratio increased only little, no change in lung weight occurred, left ventricular dimensions and fractional shortening changed mildly, ascites was not present, QT duration did not increase, and sudden death did not occur. Chronic cariporide treatment also prevented cellular electrical and ionic remodeling. Myocyte dimensions were unaltered, action potentials were not prolonged, cytoplasmic sodium and NHE-1 activity did not increase, cytoplasmic and SR calcium handling remained undisturbed, and no increase of the incidence of calcium after-transient dependent delayed after depolarizations (DADs) occurred.

CONCLUSION

We conclude that enhanced activity of NHE-1 underlies cardiac cellular electrical and ionic remodeling in experimental heart failure, and that chronic dietary treatment with cariporide attenuates hypertrophy, development of HF, and cellular remodeling.

Intrinsic heterogeneity in repolarization is increased in isolated failing rabbit cardiomyocytes during simulated ischemia

OBJECTIVE

Myocardial ischemia and ventricular arrhythmias often complicate congestive heart failure. Ischemia-induced dispersion in repolarization is an important arrhythmogenic factor that might be caused by intrinsic cellular differences in response to simulated ischemia (SI) or by changed coupling of myocytes. We hypothesized that intrinsic heterogeneity in action potential duration (APD) or the occurrence of rigor is larger in failing than in normal rabbit myocytes during SI.

METHODS

Heart failure (HF) was induced with volume and pressure overload. Left ventricular myocytes from apex, free wall and base were enzymatically isolated and exposed to SI with NaCN.

RESULTS

There were no baseline differences in APD before SI. During SI no differences in time to inexcitability occurred but the range in APD increased more in HF than in normal cells. Rigor occurred after 16.8+/-3.5 and 23.0+/-7.5 min (P<0.05) in normal and HF myocytes, with no differences between apical, free wall or base cells. Variance in time to rigor was larger in HF than in normal cells (55.7 versus 12.4 min(2)). Blockade of anaerobic reserve decreased variance in time to rigor, also when normalized to mean, in HF and normal myocytes. In coupled normal and HF cell pairs, no delay in action potential propagation or differences in APD occurred during SI, and time to rigor was synchronized (P<0.05 vs. single cells).

CONCLUSIONS

Intercellular differences in APD and in time of rigor arise in normal and HF myocytes subjected to SI, and are inhibited by blockade of anaerobic glycolysis. Dispersion in APD and tolerance to SI is increased in HF cells. APD and time to rigor are completely synchronized in coupled cell pairs.

SR calcium handling and calcium after-transients in a rabbit model of heart failure

OBJECTIVE

After-depolarization associated arrhythmias are frequently observed in heart failure and associated with spontaneous calcium release from sarcoplasmic reticulum (SR), calcium after-transients. We hypothesize that disturbed SR calcium handling underlies calcium after-transients in heart failure (HF).

METHODS

We measured the stimulation rate dependence (0.2-3 Hz) of diastolic calcium, calcium transient amplitude and SR calcium content in left ventricular myocytes isolated from hearts of rabbits with pressure and volume overload-induced HF and age-matched control animals. Cytosolic calcium was measured with indo-1. In some experiments, delayed after-depolarizations (DADs) were monitored with the voltage sensitive dye di-4-Annepps. SR calcium content was estimated from the response to rapid cooling (RC). After-transients were elicited in the presence of norepinephrine (100 nmol/l) after cessation of burst pacing.

RESULTS

With increasing stimulation rate (0.2-3.0 Hz): (1) steady state diastolic [Ca](i) increased from 102 to 174 nmol/l in HF and from 44 to 103 nmol/l in control, (2) calcium transient amplitudes decreased from 310 to 254 nmol/l in HF and increased from 186 to 429 nmol/l in control, (3) SR calcium content decreased from 1.25 to 1.09 mmol/l in HF and increased from 1.51 to 2.48 mmol/l in control, (4) in HF and control, the end diastolic SR membrane calcium gradient decreased by about 30%; at any stimulation rate, the magnitude of gradient in HF was one-third of control, (5) systolic depletion of SR was 85% in HF and 60% in control. In HF, noradrenaline (100 nmol/l) increased SR calcium content and SR membrane gradient by 40% versus about 7% in control. Calcium after-transients were observed in 14 out of 18 HF rabbits, and none in eight control animals and were associated with DADs. Calcium after-transients were associated with a 35% decrease in SR calcium content. The frequency of occurrence of calcium after-transients was related to diastolic calcium.

CONCLUSIONS

in HF, diastolic calcium is increased and both SR calcium content and SR membrane calcium gradient are decreased in a stimulation rate-dependent manner. In HF, beta-adrenergic stimulation can partly restore the SR calcium content and SR membrane gradient at higher stimulation rates in a meta-stable condition; upon transition to low stimulation rates, the SR membrane can no longer maintain this high unbalanced SR calcium load at increased diastolic calcium, the magnitude of which is causally related to the occurrence of calcium after-transients.