Myofilament function critically determines diastolic cardiomyocyte stiffness, and myofilament stiffness is abnormally increased in HFpEF patients (Borbely et al., 2005). The giant elastic protein titin is known to influence passive stiffness via isoform shift (N2BA/N2B ratio) and posttranslational modifications such as oxidation and/or phosphorylation (Linke and Hamdani, 2014). In animal and human HFpEF myocardia, altered phosphorylation of titin and other small regulatory myofilament proteins have been shown to increase passive cardiomyocyte stiffness (Hamdani et al., 2013a,b, c; Linke and Hamdani, 2014). Our previous findings demonstrated that in human HFpEF myocardium, empagliflozin restores the pathologically altered phosphorylation of titin (Figure 1) and the small regulatory proteins troponin I and myosin binding protein C (Pabel et al., 2018). Consequently, empagliflozin treatment led to a reduction of pathological cardiomyocyte stiffness in human HFpEF myocardium (Pabel et al., 2018). Furthermore, we revealed that these observations were mediated by an improvement of cyclic guanosine monophosphate (cGMP)-dependent protein kinase or protein kinase G (PKG) signaling, which is typically diminished in HFpEF myocardium and is known to underlie diastolic stiffness in HFpEF (Paulus and Tschöpe, 2013). Accordingly, the improvement of cGMP pool, which regulates PKG activity, is thus considered as a potential therapeutic target in HFpEF (Greene et al., 2013). Interestingly, empagliflozin enhanced the NO-cGMP-PKG pathway after 8 weeks of treatment in diabetic mice (Xue et al., 2019). In human and rodent HFpEF myocardium we showed that nitric oxide bioavailability increased upon acute empagliflozin treatment resulting in elevated cGMP levels and increased PKG activity. As a consequence, PKG-dependent phosphorylation of myofilament proteins was restored (Kolijn et al., 2021). Recently, in pigs with myocardial infarction induced HFrEF 2 months of treatment with empagliflozin also resulted in an improved diastolic function in invasive and non-invasive analyses, which was associated with increased NO availability and PKG signaling (Santos-Gallego et al., 2021). As PKG is centrally involved in HFpEF pathophysiology, the impact of SGLT2i on PKG signaling and myofilament function could therefore be a key-stone effect in improving diastolic function in HFpEF hearts (Figure 2), thereby resulting in material change the disease trajectory (Pabel et al., 2020a).

Increased cardiac fibrosis adversely affects diastolic function and is a common feature in HFpEF patients (Zile et al., 2015). The etiology of fibrosis is heterogenous and the development of fibrotic tissue presumably takes place at later disease stages (Sweeney et al., 2020). Increased fibrosis reduces myocardial compliance thereby limiting diastolic filling (de Souza, 2002; Kasner et al., 2011; Sweeney et al., 2020). Limited evidence of the impact of SGLT2i on myocardial fibrosis has been reported. In a hypertensive HF model, 12 weeks of treatment with empagliflozin resulted in reduced cardiac remodeling with less atrial as well as ventricular fibrosis (Lee et al., 2019). In another rat model of myocardial infarction, dapagliflozin reduced myofibroblast and macrophages infiltration and thereby demonstrating antifibrotic properties (Lee et al., 2017). Reduced fibrotic content upon SGLT2i treatment was also observed in diabetic (Li et al., 2019) and afterload induced HFrEF mice (Shi et al., 2019). A possible mechanism has been provided in a diabetic mouse model, where dapagliflozin reduced myocardial fibrosis and proinflammatory markers, which was associated with regulation of AMPK (Ye et al., 2017). Accordingly, dapagliflozin increased AMPK phosphorylation in cardiac fibroblasts, which also resulted in a reduction of NHE1 mRNA expression (Ye et al., 2018).

Preventing the development of cardiac fibrosis using SGLT2i likely impedes the progression of myocardial stiffness and may therefore be advantageous for the causal treatment HFpEF patients. Nevertheless, evidence of these specific and direct mechanisms of SGLT2i on fibrosis is limited. Moreover, the effect of these agents on myocardial fibrosis may be confounded by their other cardiovascular effects which may deter disease progression via different mechanisms.

In patients presenting with HFpEF, diastolic dysfunction and cardiac hypertrophy are often concomitantly found. While the interaction of hypertrophy, diastolic function and HFpEF is complex, increased left ventricular hypertrophy impairs chamber geometry and may induce diastolic dysfunction per se (Heinzel et al., 2015). Along with the SGLT2i-induced improvement of diastolic function in different patient populations, an impact of SGLT2i on left ventricular hypertrophy has been demonstrated. Clinical data of a randomized placebo-controlled trial of 97 patients with preserved EF and T2DM as well as coronary artery disease demonstrated using cardiac MRI that empagliflozin reduced LV mass index after 6 months of treatment (Verma et al., 2019). A reduction of LV mass following SGLT2i therapy was also observed in echocardiographic measurements of patients with T2DM (Verma et al., 2016; Brown et al., 2020). However, in these clinical trials it is impossible to separate direct cardiac effects from secondary mechanisms such as changes of blood pressure or pre- and afterload. As of now, the mechanisms of the regression in hypertrophy warrants further investigation, in particular in HFpEF patients.

Cardiomyocyte Ca²⁺ homeostasis mediates excitation-contraction coupling, thereby determining myocardial contraction and relaxation. During diastole cytosolic Ca²⁺ moves back into the sarcoplasmic reticulum via the SERCA2a transporter which is modulated by phospholamban. Ca²⁺ is also eliminated from the cardiomyocyte via the Na⁺/Ca²⁺ exchanger. As cytosolic Ca²⁺ levels decrease, the passive dissociation from Troponin C changes tropomyosin conformation resulting in myocardial relaxation (Bers, 2002). Thus, Ca²⁺ homeostasis plays a critical role for diastolic function. However, data on Ca²⁺ handling in HFpEF are scarce due to limited availability of human samples and limitations of HFpEF-like animal models. Reports from HFpEF-like animal models (i.e., via age, metabolic disorders or transverse aortic constriction) indicate that systolic Ca²⁺ release and cell shortening could be unaltered or increased, while Ca²⁺ reuptake and relaxation may be impaired along with elevated diastolic Ca²⁺ depending on the model studied (Peana and Domeier, 2017; Frisk et al., 2021). SGLT2i have been reported to influence cardiomyocyte Na⁺ homeostasis and thereby Ca²⁺ handling (firstly reported by Baartscheer et al., 2017), thus potentially modulating diastolic function. In failing ventricular murine and human cardiomyocytes treated in vitro with empagliflozin (24 h) Ca²⁺/calmodulin-dependent protein kinase IIδ (CaMKII) activity has been found to be diminished (Mustroph et al., 2018). Consequently, aberrant diastolic sarcoplasmic reticulum Ca²⁺ leak, which elevates cytosolic Ca²⁺ levels and thereby adversely increasing diastolic tension (Fischer et al., 2013) was reduced after exposure to empagliflozin (Mustroph et al., 2018). It has therefore been speculated that this mechanism could also be involved in HFpEF pathophysiology (Eisner et al., 2020). Interestingly, in obese diabetic mice, the improvement of diastolic function was associated with an increased phospholamban phosphorylation and thus SERCA2a activity (Hammoudi et al., 2017). However, the effects of SGLT2i on cardiomyocyte Ca²⁺ homeostasis are in part controversial and difficult to interpret as experimental protocols and (disease) models varied. In human cardiomyocytes from patients with HFrEF acute treatment with empagliflozin did not change systolic Ca²⁺ transient or diastolic cytosolic Ca²⁺ (Pabel et al., 2018). Likewise, we performed a blinded experimental long-term study of human induced pluripotent stem cell cardiomyocytes from healthy subjects as clinical effects occur after weeks or months which is in contrast to many experimental study designs where acute exposure to these drugs have been performed. In our respective study 2 months of treatment with empagliflozin showed no impact on Ca²⁺ homeostasis and EC-coupling proteins (Pabel et al., 2020b). In a study based on Dahl salt-sensitive rats with high-salt diet serving as a HFpEF model, the authors reported that dapagliflozin beneficially affects Ca²⁺ and Na⁺ overload after in vivo treatment but not after direct treatment of cardiomyocytes (Cappetta et al., 2020). Therefore, also non-cardiomyocyte targets might be involved in possible effects of SGLT2i on EC-coupling.

Besides Ca²⁺ cycling, Na⁺ homeostasis has shown to influence diastolic function. As an increased Na⁺ influx is counterbalanced via Na⁺/Ca²⁺ exchanger, mechanisms elevating cytosolic Na⁺ levels may also increase Ca²⁺ concentration. We have demonstrated that inhibition of the late Na⁺ current reduces increased cellular Na⁺ in HF and secondary diastolic Ca²⁺ which indeed led to a reduced diastolic dysfunction in human HF preparations (Sossalla et al., 2008). This mechanism may theoretically also reduce the arrhythmia potential and could explain the early mortality benefit seen (i.e., less sudden cardiac deaths) in this otherwise high risk population. Thus, distorted Na⁺/Ca²⁺ interplay detrimentally contributes to diastolic dysfunction in HFrEF (Sossalla et al., 2008).

Interestingly, SGLT2i has been recently reported to inhibit the late Na⁺ current in murine HF cardiomyocytes, which constitutes an abnormal Na⁺ influx throughout the action potential (Philippaert et al., 2021). In molecular docking simulations these effects were proposed to be driven by binding of empagliflozin on major cardiac Na⁺ channel isoform Na_V1.5 (Philippaert et al., 2021). While the role of SGLT2i for late Na⁺ current needs to be further investigated in HFpEF, the reduction of late Na⁺ current could favorably impact diastolic function (Sossalla et al., 2011). Interestingly, in patients with hypertrophic cardiomyopathy, typically characterized by normal systolic function but severely disturbed myocardial relaxation, the late Na⁺ current has been demonstrated to deteriorate cardiomyocyte Na⁺ and thus Ca²⁺ balance as a mechanism for impaired diastolic function (Coppini et al., 2013; Ferrantini et al., 2018). On the other hand, in patients with HFpEF and hypertensive heart disease an increased diastolic Ca²⁺ with impaired relaxation was found, which was, however, not caused by elevated Na⁺ levels in this model (Runte et al., 2017). Thus, further studies of cardiomyocyte Na⁺ homeostasis with respect to SGLT2i are needed, in particular in HFpEF myocardium.

Another mechanism by which SGLT2i may influence myocardial Na⁺ homeostasis are inhibitory effects on the Na⁺/H⁺ exchanger 1 (NHE1), which were first reported for rabbit myocardium (Baartscheer et al., 2017), and later confirmed for murine cardiomyocytes (Uthman et al., 2018) and human atrial tissue (Trum et al., 2020). A study in healthy rabbit cardiomyocytes reported a consecutive acute reduction in cytosolic Ca²⁺ and Na⁺ by empagliflozin (Baartscheer et al., 2017). On the contrary the effects on Na⁺ homeostasis via NHE1 inhibition have been questioned in a study using healthy rat cardiomyocytes (Chung et al., 2020).

Finally, overall cytosolic Na⁺ levels were also decreased by empagliflozin after 30 min and 24 h treatment in murine wild-type mice (Mustroph et al., 2018). Therefore, SGLT2i-dependent changes in myocardial Na⁺ might constitute an important cardiac mechanism potentially also in HFpEF (Trum et al., 2021). In conclusion, studies in different experimental models reported an involvement of SGLT2i in cellular Ca²⁺ and Na⁺ homeostasis. Yet, the role of cellular Ca²⁺ and Na⁺ alterations with respect to treatment with SGLT2i in HFpEF is rather speculative and further studies in HFpEF myocardium are required to clarify this important question.

In HFpEF, inflammation and oxidative stress play a key role for the progression of structural and functional diastolic dysfunction and are associated with comorbidities typically found in HFpEF patients (Figure 3) such as chronic kidney disease or metabolic syndrome (Franssen et al., 2016; Zhazykbayeva et al., 2020). In particular in cardiorenal syndrome, kidney injury-associated chronic inflammation and oxidative activation may impair cardiac function as shown in different models of renal failure (Rangaswami et al., 2019). Thus, oxidative stress and inflammation are considered as central mechanisms linking the cardiac HFpEF phenotype with the multifaceted comorbidities in the HFpEF patient (Zhazykbayeva et al., 2020). Growing evidence demonstrates that SGLT2i attenuate inflammation and oxidative stress (Yaribeygi et al., 2019). In diabetic mice with myocardial infarction SGLT2i reduced oxidative stress and inflammatory markers (Ye et al., 2017; Yurista et al., 2019). Likewise, 4 weeks of treatment with ipragliflozin diminished oxidative stress and inflammation in diabetic mice (Tahara et al., 2013, 2014).

In HFpEF, also inflammation/oxidative stress-mediated endothelial dysfunction may impair cardiomyocyte function (Franssen et al., 2016). A recent work demonstrated that empagliflozin may reduce inflammation-dependent endothelial dysfunction resulting in improved cardiomyocyte contractility (Juni et al., 2019). In line with that, dapagliflozin reduced endothelial dysfunction and inflammation in a HFpEF rat model (Dahl salt-sensitive rats with high-salt diet) resulting in an improved diastolic function (Cappetta et al., 2020). Moreover, we showed that empagliflozin attenuated pathologically elevated levels of oxidative stress (H₂O₂, GSH, LPO) and inflammation (ICAM, VCAM, TNFα, and IL-6) in human HFpEF myocardium after in vitro treatment (Kolijn et al., 2021).

As empagliflozin reduced oxidative stress and inflammation in human HFpEF myocardium, NO bioavailability and PKG signaling were improved upon exposure to empagliflozin leading to lower myofilament stiffness and thereby improved diastolic function in human myocardium (Pabel et al., 2018; Kolijn et al., 2021). Thus, the attenuation of oxidative stress and inflammation due to SGLT2i treatment could be potentially helpful in HFpEF patients (Figure 3) at least via an improvement of contractility (Pabel et al., 2020a). Also, other potential secondary effects of SGLT2i driven by a reduction of oxidative stress and inflammation are conceivable. A potential oxidative CaMKII activation (Erickson et al., 2008) might be diminished as SGLT2i reduce oxidative stress, which could result in lower diastolic sarcoplasmic reticulum Ca²⁺ leak as well as reduced late Na⁺ current (Mustroph et al., 2018; Philippaert et al., 2021). Finally, hypertrophy and fibrosis are a common detrimental outcome of chronic inflammation and oxidative stress, and could thereby be ameliorated upon anti-oxidative and anti-inflammatory effects of SGLT2i (Zhazykbayeva et al., 2020). However, the molecular mechanisms need to be explored further.