New cardiac targets for empagliflozin: O-GlcNAcylation, CaMKII, and calcium handling

Empagliflozin, a sodium-glucose cotransporter-2 (SGLT2) inhibitor, reduces reabsorption of glucose in the renal proximal tubules and increases urinary excretion of glucose. This lowers blood glucose levels and provides important metabolic and hemodynamic benefits to patients with type 2 diabetes mellitus (T2DM). Diabetic cardiomyopathy is a severe consequence of T2DM associated with increased risk of coronary artery disease, heart failure (HF), and cardiac arrhythmias. Recent clinical trials showed that SGLT2 inhibitors reduce the composite outcome of cardiovascular mortality and hospitalization not only in patients with HF and T2DM but also in patients with HF but without T2DM, a benefit not observed with other glucose-lowering agents (1). In addition, mounting preclinical evidence from studies on ex vivo hearts and isolated cardiomyocytes demonstrated that empagliflozin directly affects heart cells. Since cardiomyocytes lack SGLT2 expression, off-target effects have been suggested, including the inhibition of Na⁺/H⁺ exchanger (NHE) and attenuation of reactive oxygen species (ROS) (1). However, the exact molecular target of empagliflozin in cardiomyocytes remains unclear.

In this issue of the American Journal of Physiology-Heart and Circulatory Physiology, Kadosaka et al. (2) reported that empagliflozin treatment (30–60 min, at a clinically relevant 1 µM concentration) significantly attenuated arrhythmia susceptibility and cardiomyocyte Ca²⁺handling impairments in the leptin receptor-deficient db/db (Lepr^db/db) mouse model of T2DM. They found that empagliflozin reduced cardiomyocyte glucose uptake, attenuated posttranslational modifications of intracellular proteins by O-linked-N-acetylglucosamine (O- GlcNAc), and prevented autonomous activation of Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) (inferred from reduced autophosphorylation) in db/db hearts (Fig. 1). Previous studies demonstrated that O-GlcNAcylation of CaMKII at serine-280 site is an important mediator of increased sarcoplasmic reticulum (SR) Ca²⁺ leak and arrhythmias in diabetic hyperglycemia (3, 4). In line with these data, Kadosaka et al. showed that db/db hearts exhibit increased number of premature ventricular complexes (PVCs) and increased phosphorylation of the ryanodine receptor (RyR) at serine 2814 [the key CaMKII phosphosite (5)], as well as higher frequency of Ca²⁺ sparks and Ca²⁺ waves in isolated cardiomyocytes. Second, the SR Ca²⁺ reuptake was slower in db/db, which, together with the marked SR Ca²⁺ leak, could contribute to reduced SR Ca²⁺ load in db/db. The underlying mechanism(s) for the reduced SR Ca²⁺ ATPase (SERCA) activity may include increased phospholamban (PLB) expression, frequently reported in T2DM, and O-GlcNAcylation of PLB, which may increase the inhibitory action of PLB on the SERCA (4). Importantly, Kadosaka et al. (2) demonstrated that empagliflozin reversed all these proarrhythmic changes in db/db. Although the effect of empagliflozin on diastolic function was not studied by Kadosaka et al., a similar study (6) reported that chronic empagliflozin treatment improved diastolic function in db/db mice, which was associated with reduced cardiac glucose concentrations and attenuated CaMKII activation with lower levels of RyR phosphorylation. In another study (7), empagliflozin treatment also attenuated diastolic dysfunction and interstitial fibrosis in female db/db mice; however, no significant fibrosis was found in male db/db mice by Kadosaka et al. (2), whose data may point to significant sex differences in diabetic cardiomyopathy.

Figure 1.

Schematic diagram of empagliflozin effects on cardiomyocyte Ca²⁺ cycling and signaling in diabetic cardiomyopathy. Diabetic cardiomyocytes exhibit higher glucose uptake, which stimulates the hexosamine biosynthetic pathway to provide a substrate for O-linked-N-acetylglucosamine (O-GlcNAc) modifications. O-GlcNAcylation of the Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) at serine-280 activates the kinase resulting in increased autophosphorylation, autonomous activity, and phosphorylation (P) of key molecular targets such as the ryanodine receptor (RyR) and the voltage-gated Na⁺ channel. CaMKII-dependent phosphorylation of the Na⁺ channel increases the late Na⁺ current (I_NaL) and prolongs the action potential (AP) duration (APD), which then increases Ca²⁺ entry through the L-type Ca²⁺ current (I_CaL). Phosphorylation of RyR at serine-2814 increases spontaneous Ca²⁺ leak from the sarcoplasmic reticulum (SR) to increase the frequency of Ca²⁺ sparks and Ca²⁺ waves. Ca²⁺ must then be removed from the cytosol, in part by the Na⁺/Ca²⁺ exchanger (NCX), which is electrogenic and may lead to delayed afterdepolarizations (DADs) and triggered APs. These may manifest as frequent premature ventricular complexes (PVCs) on surface electrocardiograms. At the same time, intracellular Ca²⁺ transient decay is generally prolonged and SR Ca²⁺ reuptake is slowed in diabetes, which might be explained by the increased expression and O-GlcNAcylation of phospholamban (PLB) that enhance the inhibitory action of PLB on the SR Ca²⁺ ATPase (SERCA). Importantly, Kadosaka et al. (2) demonstrated that the sodium-glucose cotransporter-2 (SGLT2) inhibitor empagliflozin reduced glucose uptake in diabetic ventricular myocytes despite the lack of SGLT2 expression in cardiomyocytes. Empagliflozin also reduced global O-GlcNAc levels and CaMKII autophosphorylation, and it attenuated the arrhythmogenic consequences of CaMKII activation. Empagliflozin may also inhibit both the Na⁺/H⁺ exchanger (NHE) and I_NaL, which could provide additional benefits to the diabetic hearts. Similar mechanisms may occur in the failing heart with or without diabetes.

Kadosaka et al. (2) found that promotion of protein O-GlcNAcylation by inhibiting the O-GlcNAc removal enzyme (O-GlcNAcase) with thiamet-G opposed empagliflozin’s beneficial effects on intracellular Ca²⁺ handling in db/db. Moreover, KN-93, a cell-permeable CaMKII inhibitor also markedly reduced SR Ca²⁺ leak similar to empagliflozin but did not alter SR Ca²⁺ reuptake in db/db (2). These data pointed to the critical involvement of O-GlcNAcylation and CaMKII in mediating empagliflozin’s action on Ca²⁺ handling impairments in diabetic cardiomyocytes. Altogether, these data suggest that empagliflozin may have a molecular target upstream from O-GlcNAcylation and CaMKII. Such a target could potentially be an unidentified sarcolemmal glucose transporter. However, CaMKII can also regulate glucose uptake and cell metabolism (5); thus, the reduced glucose uptake might also be a consequence of CaMKII inhibition. Moreover, caution must be exercised when interpreting data with KN-93 and thiamet-G, because these small molecules may have significant off-target effects (5). Thus, CaMKII knockout animals in future studies are critically needed to confirm these findings and may further inform empagliflozin’s mechanism of action in diabetic cardiomyopathy. Kadosaka et al. (2) reported that NHE inhibition using cariporide showed some benefits to intracellular Ca²⁺ extrusion in diabetic cardiomyocytes; however, cariporide failed to ameliorate SR Ca²⁺ leak, SR Ca²⁺ content, and SERCA activity. These data suggest that empagliflozin’s effects go beyond NHE inhibition. Moreover, mitochondrial NHE inhibition and consequential changes in ROS production may also contribute to these effects, which require further study. It would also be interesting to know whether mitochondrial and nuclear Ca²⁺ signals can be affected by empagliflozin.

Diabetic cardiomyopathy and hyperglycemia are also associated with remodeling in multiple sarcolemmal ion channels and transporters that may further increase arrhythmia susceptibility (5). Although Kadosaka et al. (2) elegantly demonstrated empagliflozin’s advantageous effects on intracellular Ca²⁺ handling that may reduce arrhythmia trigger mechanisms, it remains to be determined how empagliflozin affects the function of key sarcolemmal ion channels in diabetes. In db/db mice with chronic aldosterone infusion (8) that induced HF with preserved ejection fraction (HFpEF), empagliflozin reversed the upregulation of the late Na⁺ current (I_NaL), another key target of CaMKII (5). Thus, empagliflozin may be antiarrhythmic by reversing a cardiomyocyte Na⁺ and Ca²⁺ handling vicious cycle (involving Ca²⁺-CaMKII-RyR-ROS-I_NaL) that has been shown to promote arrhythmias and contractile dysfunction in HF (9) and may also occur in T2DM and HFpEF. Further studies must determine whether, and how, glucose uptake or O-GlcNAcylation is also involved in this effect of empagliflozin on I_NaL in diabetes.

Importantly, SGLT2 inhibitors provide cardiovascular benefits not only in patients with T2DM but also in patients who are nondiabetic with HF. The proposed mechanism by Kadosaka et al. (2) might, at first, sound counterintuitive in this context; however, altered glucose uptake, increased O-GlcNAcylation, and CaMKII activation were all observed in the failing human heart without T2DM (3). Nonetheless, the contribution of O-GlcNAcylation to cardiac arrhythmias and contractile dysfunction in HF (with or without T2DM) and their potential reversal by empagliflozin merit further study. Another interesting and controversial topic is whether ROS or O-GlcNAcylation is the key mechanism for cardiac remodeling and arrhythmias in diabetes. Increases in ROS and O-GlcNAcylation have been reported in diabetes, and both ROS and O-GlcNAcylation can autonomously activate CaMKII (5), but the predominant mechanism may differ between diabetes types and stages of disease. ROS and O-GlcNAcylation may also differentially regulate cardiac ion channels in diabetic hyperglycemia (10). Importantly, empagliflozin reduces both ROS and O-GlcNAc levels in diabetic hearts (1, 2).

In conclusion, the study by Kadosaka et al. (2) highlights new potential targets of empagliflozin in diabetic cardiomyopathy, which include O-GlcNAcylation, CaMKII, and intracellular Ca²⁺ cycling. These mechanisms may also contribute to empagliflozin’s advantageous effect in other cell types, organs, and diseases, which are worthy of further investigations. Although these effects of empagliflozin reveal potential novel mechanism of action, the exact molecular target(s) of empagliflozin in cardiac myocytes is still to be determined.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants R01-HL142282 and P01-HL141084 (to D.M.B.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

B.H. and D.M.B. prepared the figure; drafted manuscript; edited and revised manuscript; and approved final version of manuscript.