Empagliflozin reverses late Na⁺ current enhancement and cardiomyocyte proarrhythmia in a translational murine model of heart failure with preserved ejection fraction

Empagliflozin, a sodium-glucose cotransporter-2 (SGLT2) inhibitor has been recently shown to improve cardiovascular outcomes in patients with heart failure and preserved ejection fraction (HFpEF), independent of diabetes status.¹ In addition to its advantageous renal and hemodynamic effects, empagliflozin can directly affect heart cells, and preclinical studies demonstrated empagliflozin effects in ex vivo perfused hearts and isolated cardiomyocytes.² However, the exact mechanism by which empagliflozin affects the heart remains elusive. In cardiomyocytes, SGLT2 expression is virtually absent, suggesting off-target beneficial effects. While controversial, the sodium-hydrogen exchanger-1 has been suggested as a potential empagliflozin target in cardiomyocytes, especially during ischemia. Recently, Philippaert et al.³ showed that empagliflozin directly inhibited late Na⁺ current (I_NaL) in a transverse aortic constriction-induced heart failure murine model and in recombinant human Na_V1.5 channels in HEK cells carrying long QT3 mutations or treated with H₂O₂. Moreover, empagliflozin reduces reactive oxygen species (ROS) production² and Ca²⁺/calmodulin-dependent kinase II (CaMKII) activity,⁴ important known regulators of I_NaL in heart diseases. Here, we tested for empagliflozin effects on I_NaL and consequent action potential (AP) changes in a translational two-hit murine model of HFpEF.⁵

HFpEF was induced in C57BL/6J wild-type mice (male, 10-wk-old, Jackson Laboratory) by combining high-fat diet (D12492, Research Diets) and inhibition of nitric oxide synthases with L-NG-nitroarginine methyl-ester (L-NAME, 0.5 g/L, Sigma-Aldrich) in drinking water, while control mice were fed normal chow (5k52, LabDiet) and water. After 15 weeks of HFpEF treatment, consistent with previous report,⁵ morphometric and echocardiographic evaluations (Figure [A]) revealed significant obesity, concentric cardiac hypertrophy, preserved ejection fraction, and significant diastolic dysfunction, as demonstrated by increased heart mass, left ventricular remodeling index, and elevated E/e’. Empagliflozin effects were studied in freshly isolated left ventricular myocytes using patch-clamp at 37°C. In voltage-clamp I_NaL measurements, pipette solution contained (mmol/L): 110 CsCl, 20 tetraethylammonium chloride, 5 MgATP, 10 HEPES, 5 phosphocreatine-Na₂, 0.0001 calmodulin, 10 EGTA, 4.1 CaCl₂ (free [Ca²⁺]=100 nmol/L), pH=7.20. Bath solution contained (mmol/L): 140 NaCl, 4 CsCl, 1.8 CaCl₂, 1 MgCl₂, 5 HEPES, 5 Na-HEPES, 5.5 glucose, 5 4-aminopyridine, 0.01 nifedipine, pH=7.40.

Figure. Empagliflozin reduces late Na⁺ current and action potential prolongation in a preclinical model of heart failure with preserved ejection fraction (HFpEF).

Wild-type mice (10-wk-old) on high-fat diet+L-NG-nitroarginine methyl ester (L-NAME) for 15 weeks were compared to those fed normal chow (N=10 per group). A, Body weight, heart weight to tibial length ratio (HW/TL), and membrane capacitance (C_m) of single ventricular cardiomyocytes. Left ventricular (LV) ejection fraction (EF), ratio between mitral E wave and e’ wave (E/e’), and LV remodeling index LVRI (LV mass/LV end-diastolic internal diameter, LVM/LVIDd). B, Representative late Na⁺ current (I_NaL) traces in control and HFpEF myocytes following acute empagliflozin (EMPA) and subsequent tetrodotoxin (TTX) applications. Insets shows voltage protocol, comparison of peak versus late I_Na, and the time course of EMPA and TTX effects on I_NaL. Peak I_Na was off-scale as shown. C, Four hour preincubation with empagliflozin reversed I_NaL upregulation in HFpEF. D, Action potential (AP) measurements in HFpEF myocytes paced at 1 Hz. Top left, Representative AP traces in control and HFpEF myocytes without or with preincubation with empagliflozin and autocamtide-2-related inhibitory peptide (AIP). Bottom left, Fifty consecutive AP durations at 90% repolarization (APD₉₀) for control and HFpEF as above. Right, Average data on APD₉₀ and short-term variability of APD₉₀ (STV). E, Concentration-dependent acute effect of empagliflozin on H₂O₂ (100 μmol/L, 5 min)-induced I_NaL in wild-type and CaMKII MM281/282VV (N=4 per group). IC₅₀, half-maximal inhibitory concentration. Data are presented as mean±SEM. Two-tailed Welch’s t-test was used for A. ANOVA followed by Tukey’s multiple-comparisons test was used for B, C, and D. All data, P values and n numbers are shown.

Acute empagliflozin treatment (1 μmol/L, 3 min), in contrast to a previous report,³ did not change I_NaL in control and HFpEF, whereas tetrodotoxin (TTX, 10 μmol/L) readily inhibited I_NaL in both cases (Figure [B]). Importantly, I_NaL density (current amplitude normalized to cell capacitance) was significantly increased in HFpEF. However, empagliflozin preincubation (4 hr) versus vehicle (0.01% dimethyl sulfoxide) reversed the TTX-sensitive I_NaL upregulation in HFpEF to the control level but had no effects in control myocytes (Figure [C]).

We also measured APs to assess antiarrhythmic properties of empagliflozin. For physiological AP measurements, Cs⁺ was replaced with K⁺ in pipette and bath solutions without K⁺ and Ca²⁺ current inhibitors. Pipette EGTA was reduced to 10 μmol/L and CaCl₂ was omitted. In line with upregulated I_NaL, myocyte AP duration (APD) was significantly prolonged, and the short-term APD-variability (STV) was markedly increased in HFpEF. These proarrhythmic APD changes were reversed in cells preincubated with empagliflozin (Figure [D]). Because CaMKII is an important regulator of I_NaL in heart diseases, we also tested the effect of cell-pretreatment with the selective CaMKII inhibitor autocamtide-2-related inhibitory peptide (AIP, myristoylated, 1 μmol/L). Importantly, CaMKII inhibition also reversed APD prolongation and temporal variability in HFpEF (Figure [D]).

Finally, we measured concentration-response effects of empagliflozin on ROS-promoted I_NaL. Empagliflozin acutely inhibited the H₂O₂-induced I_NaL, but significantly only above 3 µmol/L with a half-maximal inhibition (IC₅₀) at 5.68 μmol/L (Figure [E]). Critically, myocytes with oxidation-resistant mutations in CaMKII (MM281/282VV) prevented both the H₂O₂-induced increase in I_NaL and the empagliflozin effect (Figure [E]).

Our data show that I_NaL is elevated in a preclinical murine model of HFpEF and that empagliflozin at a clinically-relevant 1 μmol/L concentration reverts I_NaL upregulation and arrhythmogenic AP changes. However, this empagliflozin effect on the Na⁺ channel in HFpEF is not acute inhibition (unlike the direct acute selective I_NaL blocker TTX) but requires drug preincubation, and this effect is mimicked by CaMKII inhibition and absent when CaMKII lacks two critical oxidation sites. Moreover, empagliflozin acutely inhibits H₂O₂-induced I_NaL with a 7 times higher IC₅₀ in cardiomyocytes than in HEK cells³ (5.68 μmol/L versus 0.79 μmol/L). Thus, we suggest that empagliflozin likely regulates I_NaL in HFpEF by an indirect mechanism, which may involve CaMKII. That agrees with results that empagliflozin treatment for 24 hr (but not 30 min) suppresses CaMKII activity and sarcoplasmic reticulum Ca²⁺ leak, and enhanced Ca²⁺ transients.⁴ Mechanisms could include reduced ROS and its ability to activate CaMKII. While the key molecular targets of empagliflozin in cardiomyocytes remain unclear, I_NaL and CaMKII suppression may be important mediators of its beneficial effects that merit further investigation. Our study is limited in that we studied empagliflozin effects in isolated murine myocytes and focused on cell electrophysiology. Further studies are required to confirm these findings in human cardiomyocytes and on other phenotypical aspects of HFpEF.

All procedures involving animals were approved by institutional and national authorities. All supporting data are available within the article.