SGLT2-inhibitors improve the cardiovascular outcome and cardiac function in patients with heart failure. However, due to systemic effects, the mode of action of these drugs on the myocardium remains unclear. Therefore, we aimed to investigate direct effects of empagliflozin on contractility and autophagy in failing human and normal rabbit myocardial tissue in an in-vitro model. 

After written informed consent, we obtained LV tissue samples from 13 patients undergoing LVAD surgery or heart transplantation. A total of 47 living 300 µm thick myocardial slices were created with a vibratome and cultivated in biomimetic chambers for 6 (n=27) to 21 days (n=20) under continuous electrical stimulation with 0.5 Hz. M199 containing 5.5 mM glucose and insulin (1mg/ml) was used as culture medium. To investigate non-failing myocardium, we cultivated 65 LV tissue slices from 15 rabbit hearts for up to 6 days. After 1-2 days of culture, we added either 1µM empagliflozin (Empa) or vehicle (DMSO) to the culture medium. Contraction parameters and the force-frequency relationship (FFR) were assessed regularly. At the end of cultivation, we analyzed the ß-adrenergic response by adding increasing concentrations of isoprenaline. Furthermore, we assessed the autophagic flux by quantifying LC3B and p62 accumulation via Western blotting after 150 min of treatment with 100 nM bafilomycin.

In human myocardial slices, empagliflozin did not show significant effects on relative contraction force after 4 days of incubation compared to baseline amplitudes before incubation (1.28±0.4 Empa vs 1.23±0.28 DMSO, p=0.87) nor after 3 weeks of incubation (2.56±1.01 Empa vs 2.83±1.1 DMSO, p=0.62). Relative contraction duration compared to baseline values before incubation was not changed significantly either (0.86±0.08 Empa vs. 0.9±0.07 DMSO, p=0.59). The maximum frequency captured did not differ significantly between empagliflozin and control group (2.4±0.4Hz Empa vs. 2.64±0.3Hz DMSO, p=0.58 after 4 days, 2.94±0.12Hz Empa vs. 3.0±0.2Hz DMSO, p=0.74 after 3 weeks), although the relative contraction force at 2 Hz compared with the baseline stimulation frequency at 0.5 Hz showed a trend towards higher values in the empagliflozin group (1.13±0.18 Empa vs. 1.03±0.16 DMSO,p=0.07) after 4 days of incubation. The mean relative response to 1 µM isoprenaline was comparable in Empa (3.29±0.52) and DMSO (2.97±0.51, p=0.53). Similar results, i.e. no significant differences in the analyzed parameters, were obtained with non-failing rabbit myocardial slices. Autophagic flux assays in cultivated rabbit myocardium (n=6) revealed intact autophagy, as indicated by p62 and LC3BII accumulation after addition of bafilomycin, but no significant difference in autophagic flux between Empa and DMSO.

We conclude that with the chosen method and culture conditions no significant effects of empagliflozin can be detected on cultivated failing human and non-failing rabbit myocardium. Thus, the clinically observed effects of SGLT2 inhibitors might be due to mechanisms not considered in the in-vitro model used. However, if autophagic flux is altered by empagliflozin in human myocardium and if different culture conditions can reveal effects that are otherwise masked, is being investigated in this ongoing study.