The effect of SGLT2 inhibitors on the endothelium and the microcirculation: from bench to bedside and beyond

Abstract

Aims

The beneficial cardiovascular effects of sodium-glucose cotransporter 2 (SGLT2) inhibitors irrespective of the presence of diabetes mellitus are nowadays well established and they already constitute a significant pillar for the management of heart failure, irrespective of the ejection fraction. The exact underlying mechanisms accountable for these effects, however, remain largely unknown. The direct effect on endothelial function and microcirculation is one of the most well studied. The broad range of studies presented in this review aims to link all available data from the bench to bedside and highlight the existing gaps as well as the future directions in the investigations concerning the effects of SGLT2 inhibitors on the endothelium and the microcirculation.

Methods and results

An extensive search has been conducted using the MEDLINE/PubMed database in order to identify the relevant studies. Preclinical data suggest that SGLT2 inhibitors directly affect endothelial function independently of glucose and specifically via several interplaying molecular pathways, resulting in improved vasodilation, increased NO production, enhanced mitochondrial homeostasis, endothelial cell viability, and angiogenesis as well as attenuation of oxidative stress and inflammation. Clinical data systematically confirm this beneficial effect on the endothelium, whereas the evidence concerning the effect on the microcirculation is conflicting.

Conclusion

Preclinical and clinical studies indicate that SGLT2 inhibitors attenuate endothelial and microvascular dysfunction via a combination of mechanisms, which play a role in their beneficial cardiovascular effect.

Graphical Abstract

Open in new tabDownload slide

Introduction

Sodium-glucose cotransporter 2 (SGLT2) inhibitors are a novel class of anti-diabetic drugs that lower serum glucose levels by blocking its SGLT2 mediated glucose reabsorption in the kidney, regardless of insulin secretion or action.¹ It was early recognized from the first clinical trials examining the effects of SGLT2 inhibition in diabetic patients, such as EMPA-REG OUTCOME,² that treatment with SGLT2 inhibitors reduces the risk of adverse cardiovascular outcomes and death from any cause in diabetic patients with cardiovascular disease. These findings were also consistent among future trials.^3,4 Further research examining the cardioprotective role of these drugs in cardiovascular disease, and specifically in heart failure^5–7 established that SGLT2 inhibition significantly reduces the risk of worsening heart failure or death from cardiovascular causes, regardless of the presence or absence of type 2 diabetes or HF phenotype. These results led to the recommendation of SGLT2 inhibitors for patients with HF in all phenotypes of the disease in the updated 2022 guidelines⁸ by the American Heart Association, the American College of Cardiology, and the Heart Failure Society of America.

Given the paramount significance of the positive results in HF, an increasing interest is being developed regarding the pathophysiologic pathways which SGLT2 inhibitors alter, resulting in cardiovascular protection. A handful of mechanisms have been proposed, including improved endothelial and microcirculatory function, reduction in myocardial fibrosis and arterial stiffness, enhancement of the systolic and diastolic cardiac function, weight loss and anti-hypertensive effect, diuresis-induced blood volume reduction, increased red blood cell mass, and improved myocardial bioenergetics.^1,9 However, the exact underlying mechanisms are yet to be clarified, while understanding them, especially in endothelial dysfunction, will lead to more effective strategies for treating microvascular dysfunction (MVD) in HF and a broad range of other disorders whose pathophysiology associates with endothelial or MVD, such as diabetes, brain small-vessel disease, and chronic kidney disease (CKD).^10,11

This review focuses on the effect of the SGLT2 inhibitors empagliflozin and dapagliflozin on the endothelium and the microcirculation, analyzing a broad range of data derived from research on endothelial cells and animal models, as well as from patient trials.

Pre-clinical data: assessing SGLT2 inhibition in the lab and animal studies

During the past few years, a large number of studies (Table 1) on murine and human endothelial cells have been conducted in order to evaluate the effect of SGLT2 inhibitors on the endothelium, proposing a variety of mechanisms by which they attenuate microvascular and endothelial dysfunction (Figure 1). These mechanisms, extensively analyzed below, mainly include nitric oxide production, mitochondrial function, oxidative stress, inflammation, cell viability, and angiogenesis.

The role of NO

Nitric oxide (NO) improves endothelial function through several intracellular pathways, ultimately leading to vasorelaxation, endothelial regeneration, inhibition of leukocyte chemotaxis, and platelet adhesion.¹ In specific, arginine is oxidized into citrulline and NO, via the activation of the endothelial nitric oxide synthase (eNOS). Protein kinase B, otherwise known as AKT, upregulates eNOS activity,¹² whereas asymmetric dimethylarginine (ADMA) is an endogenous inhibitor of NOs, blocking the formation of NO from arginine directly.¹

In vivo studies

Adingupu et al.¹ studied the effect of empagliflozin on coronary microcirculation and NO pathway, assessing the coronary flow velocity reserve (CFVR) and the L-arginine/ADMA ratio. After 10 weeks of treating prediabetic ob/ob^−/− mice with empagliflozin, the untreated group had lower CFVR compared with lean controls and the treated group. The L-arginine/ADMA ratio was also found to be higher in the treated mice, indicating increased NO bioavailability. The authors speculated that the increase in L-arginine could be attributed to increased kidney synthesis.

Interestingly, Santos-Gallego et al.¹³ researched on an HF pig model and found that empagliflozin activates eNOS, increases NO availability, raises cGMP, enhances protein kinase G (PKG) and phosphorylates titin, hence, contributing to improvement in diastolic dysfunction and to benefits on HFpEF both on HF hospitalizations⁷ and on quality of life.¹⁴ The results of their research come in line with a paradigm proposed for HFpEF,¹⁵ where NO impairment causes diastolic dysfunction via the eNOS/NO/cGMP/PKG/hypophosphorylation of titin pathway.

Another research team,¹⁶ using a murine model of metabolic syndrome with HFpEF, showed that empagliflozin normalizes the NO-mediated endothelium-dependent relaxation and prevents the endothelium-dependent contractile response to acetylcholine (ACh), probably through averting the formation of cyclooxygenase-derived contractile prostanoids. Also, empagliflozin seems to maintain NO bioavailability through decreasing the levels of the senescence markers p53, p21, and p16.¹⁶

Furthermore, Steven et al.¹⁷ and Oelze et al.¹⁸ report a partial prevention of endothelial dysfunction, defined as impaired ACh-dependent relaxation, in diabetic rats after empagliflozin treatment. Empagliflozin reduced the upregulated cGMP-dependent kinase (cGK-I), which is known to critically modulate NO-dependent vasodilation, and increased the phosphorylation of VASP, a substrate protein of cGK-I. It also increased dihydrofolate reductase, an enzyme that converts BH₂ to BH₄.

Similarly, dapagliflozin has also been found to restore NO bioavailability. A study upon diabetic mice¹⁹ showed that treatment with dapagliflozin significantly improved the endothelium-dependent dilatation. This effect was observed using several doses of ACh with the total area under the curve being more than doubled.

Regarding eNOS signalling, some studies provide evidence supporting that SGLT2 inhibitors restore NO bioavailability by interfering with eNOS. In murine models^9,16,20,21empagliflozin restored the downregulated cellular levels of eNOS, as well as the upregulated ET-1 expression.²⁰ A study conducted upon a murine model of left ventricular pressure overload¹² specifically showed that empagliflozin activates the AKT/eNOS/NO pathway, which suppresses endothelial apoptosis, maintains capillarization and improves systolic dysfunction. Characteristically, empagliflozin increased citrulline levels in cardiac tissue and reduced levels of arginine, indicating enhanced metabolism from arginine to citrulline and NO. Their results also suggest that empagliflozin inhibits catecholamine-induced downregulation of the AKT/eNOS/NO pathway by increasing hydroxybutyrate, which could suppress catecholamine-induced increase of reactive oxygen species (ROS). Steven et al.¹⁷ also reports that the inhibitory phosphorylation of eNOS at Thr495 was diminished by treatment with empagliflozin in diabetic mice; however, the attenuation of endothelial dysfunction by empagliflozin is potentially a result of improved NO/cGMP signalling and prevention of oxidative damage in this pathway.

Nevertheless, not all studies support the alteration of eNOS signalling by SGLT2 inhibitors. Uthman et al.²² showed that empagliflozin and dapagliflozin restore NO bioavailability after TNF-α stimulation in human coronary artery endothelial cells (HCAECs) in a way that does not involve eNOS phosphorylation, expression or signaling. They found no involvement of eNOS_ser1177 or eNOS_Thr495 phosphorylation, as well as no changes in caveolin-1 (Cav-1) and Cav-1/eNOS correlation. Cav-1 is known to have an inhibitory effect on eNOS, contributes to the compartmentalization of the eNOS in caveolae and plays an important role in endothelial signaling pathways in diseases involving the vasculature.

In vitro studies

Some in vitro studies have also investigated the effect of SGLT2 inhibitors on the NO-dependent endothelial function and have displayed promising results. In HCAECs following ischemia/reperfusion (I/R) injury, empagliflozin²³and dapagliflozin²⁴ restored the downregulated cellular levels of eNOS, in addition to the upregulated ET-1 expression,²⁴ a finding consistent with the aforementioned in vivo studies.^9,16,20,21

Furthermore, regarding myocardial contractility, two studies conducted by Juni et al.,^25,26 demonstrated that human cardiac microvascular endothelial cells (CMECs) improve CM contraction and relaxation, with empagliflozin not only facilitating the endothelium to positively modulate CM function, but also maintaining it when CMECs were exposed to pro-inflammatory cytokines or uremic serum from patients with CKD. This beneficial effect of empagliflozin was found to be mediated by the endothelial-derived NO, whose bioavailability was increased due to reduced mitochondrial ROS generation. Interestingly, they also observed that the CMEC-derived NO was a more predominant contributor to CM contractile performance, rather than the endogenous NO. The cardiac microvascular endothelial enhancement of CM function that empagliflozin managed to restore implies a novel mechanism by which empagliflozin could restore CMEC function in a variety of heart failure types with MVD, including HFpEF.

The role of mitochondrial function

Mitochondria are involved in endothelial mobilization, senescence, growth, and proliferation as well as in censoring environmental stress and cellular adaptations. Their damage, therefore, contributes to endothelial dysfunction and cardiac microvascular I/R injury.^20,23 They are also a primary source of ROS, thus their dysfunction further contributes to oxidative stress.⁹ Moreover, the superoxides derived from mitochondria reduce NO production, while also promoting apoptosis by activating caspases.²³ SGLT2 inhibitors, therefore, could improve endothelial function through interfering with several of the above mechanisms.

In vivo studies

A variety of studies provide evidence supporting that SGLT2 inhibitors exert their cardioprotective effects through interfering with mitochondrial homeostasis. More specifically, data indicate that SGLT2 inhibitors prevent mitochondrial fission,^20,21 a process with the potential to damage mitochondrial DNA during mitosis and induce endothelial apoptosis, which in turn causes microvessel swelling and vascular cell collapse.²⁰ Several pathways through which SGLT2 inhibitors prevent mitochondrial fission have been mentioned, such as reduction of Drp1 phosphorylation.²¹ The phosphorylation of Drp1 enables its shift onto mitochondria and is promoted by upstream signalling molecules, including Rho-associated coiled-coil containing protein kinase (ROCK1) and adenosine monophosphate-activated protein kinase (AMPK). Empagliflozin inhibits mitochondrial fission in an AMPK-dependent manner.²¹

Moreover, Cai et al.²⁰ found that empagliflozin suppresses mitochondrial fission by activating mitophagy on the CMECs of a murine I/R injury model, by specifically activating the FUN14 domain-containing 1 (FUNDC1)-dependent mitophagy through the AMPKα1/Unc-51-like autophagy activating kinase 1 (ULK1) pathway. Notably, mitophagy is a protective mechanism that repairs poorly structured mitochondria through normalization of mitochondrial fission and fusion, thus alleviating mitochondrial damage and preserving microvascular structure and function.²⁰

Aside from mitochondrial fission and fusion, empagliflozin has also been found to stabilize the mitochondrial membrane potential (MMP) and reduce the mitochondrial permeability transition pore opening rate in a murine model of I/R injury, ultimately reducing caspase-9-mediated mitochondrial apoptosis.²⁰ Moreover, SGLT2 inhibitors improve the energetics of the muscle and ATP generation by the mitochondria, as demonstrated with empagliflozin in pigs²⁷ and with dapagliflozin in humans.²⁸

In vitro studies

In line with the in vivo studies, the beneficial effects of SGLT2 inhibitors on mitochondrial function have also been shown in vitro. Zou et al.²³ demonstrated that empagliflozin inhibits mitochondrial fission through dephosphorylation of the mitochondrial fission 1 protein (Fis1) in I/R injured HCAECs. The SGLT2 inhibitor specifically protected the microvasculature by inhibiting the DNA-PKcs/Fis1/mitochondrial fission pathway, due to its anti-oxidative effects, resulting to a repressed mitochondrial fission and an improved mitochondrial metabolism. Dapagliflozin similarly restores the balance between mitochondrial fission and fusion. In HCAECs exposed to hypoxia-reoxygenation,²⁴dapagliflozin normalized mitochondrial morphology, by inhibiting the mitochondrial fission factors Drp1 and Mff, while upregulating the mitochondrial fusion regulators Opa1 and Mfn2.

SGLT2 inhibitors also have the ability to improve mitochondrial function in ways unaffiliated with their fission and fusion. In a murine model of hyperglycemia,²⁹empagliflozin and dapagliflozin managed to increase the basal oxygen consumption, promoting a more efficient mitochondrial metabolism, and decrease the proton leak, indicative of better mitochondrial integrity. Additionally, He et al.³⁰ established that dapagliflozin could increase a previously reduced MMP and prevent mitochondrial structural injury caused by palmitic acid (PA) in human umbilical vein endothelial cells (HUVECs) via the activation of the SIRT1/PGC-1a signaling pathway. Finally, in a recent study regarding HUVECs stimulated with high glucose levels, empagliflozin attenuated the mitochondrial Ca²⁺ overload and subsequently reduced the enhanced mitochondrial ROS production.³¹

The role of anti-oxidative effect

Moderate concentration of ROS play an important role in maintaining the proliferation and survival of endothelial cells, as well as in regulating vascular function and endothelial integrity.^32,33 Excessive levels, though, lead to deleterious vascular effects, such as increased endothelial permeability.^32,33 Mounting evidence reveals that empagliflozin displays anti-oxidative properties, thus protecting the endothelium, by inhibiting both intracellular and mitochondrial ROS (mtROS) production.^20,21,23 This anti-oxidative effect has been linked to elevated levels of anti-oxidative molecules such as glutathione (GSH), superoxide dismutase (SOD), and glutathione peroxidase (GPX).²⁰

In vivo studies

Empagliflozin treatment has been reported to reduce plasma levels of malondialdehyde (MDA), a marker of lipid peroxidation, as well as to up- or downregulate key proteins and pathways involved in ROS synthesis.⁹ Steven et al.¹⁷ and Oelze et al.¹⁸ also report a reduction in ROS formation after empagliflozin treatment in diabetic mice, as empagliflozin normalized the decreased level of the redox-sensitive enzyme mitochondrial aldehyde dehydrogenase (ALDH-2) in heart mitochondria, whereas it reduced the expression of 3-nitrotyrosine (3NT) or 4-hydroxynonenal (HNE)-positive proteins, as well as other parameters indicative of the oxidative stress extent.

Regarding the role of angiotensin II and SGLT2 inhibition, Bruckert et al.³⁴ evaluated the effect of empagliflozin on a murine model of angiotensin II-induced hypertension, reporting that empagliflozin preserves the diastolic function of the heart. This effect is mediated by averting endothelial cell activation in both the macro and coronary microcirculation and preventing the activation of the deleterious Ang II/NADPH oxidases/SGLT1- and 2 pro-oxidant pathways in endothelial cells. Empagliflozin specifically prevented the stimulatory effect of Ang II on the mRNA up-regulation of several deleterious NADPH oxidase subunits, such as p47phox, p22phox, and NOX1, as well as the downregulation of the vasoprotective NOX4.

Moreover, a study conducted by Nikolaou et al.³⁵ in non-diabetic mice showed that chronic (6 weeks) administration of empagliflozin reduces myocardial infarct size and attenuates biomarkers of cardiac oxidative stress after I/R injury. It specifically lessened MDA and protein carbonyls (PCs) in the myocardium, while it upregulated the mRNA levels of superoxide dismutase 2 (Sod2), a regulator of oxidative stress. Regarding the underlying mechanism, chronic empagliflozin administration activated the STAT-3 pathway and improved the formation of Y(705)STAT-3 dimers, which regulate many anti-oxidant and anti-apoptotic genes. Notably, acute administration of empagliflozin did not activate STAT-3, while empagliflozin was also found to inhibit ROS production in the absence of STAT-3 activation.

Finally, a recent study by Santos-Gallego et al.⁵⁸ showed that empagliflozin reduces microvascular obstruction, decreases MI size, ameliorates ischemia/reperfusion injury, and reduces oxidative stress in a porcine model of percutaneous I/R. Therefore, this study further confirms the positive impact of empagliflozin in the reduction of oxidative stress, which could possibly be a result of the slight ketonemia induced by SGLT2 inhibitors and the alteration of the metabolic substrate for the ischemic myocardium.

In vitro studies

As aforementioned,^25,26empagliflozin reduces ROS production in human CMECs exposed to pro-inflammatory cytokines and uremic acid, contributing to an endothelial NO-mediated improvement of contraction and relaxation of cardiomyocytes. Moreover, in a human model of TNF-α-stimulated HCAECs,²²empagliflozin and dapagliflozin similarly reduced the increased ROS levels. Mone et al.,³¹ also studying human brain microvascular endothelial cells (hBMECs) treated with empagliflozin and then incubated with increasing concentrations of H₂O₂, showed a significant improvement of cell viability after 5 h, in the group pretreated with empagliflozin.

Recently, Li et al.³³ revealed that SGLT2 inhibitors prevent ROS production in a HCAEC model of enhanced cyclic stretch, thus displaying their ability to alleviate oxidative stress specifically caused by mechanical forces. The researchers used Na+/H + Exchanger 1 (NHE1) and NADPH oxidase inhibitors, which blocked ROS production induced by stretch. Notably, no further reduction in ROS production was noticed when empagliflozin was co-administered, suggesting that the anti-oxidative properties of empagliflozin are mediated by NHE1 and NOXs. NHE1 inhibition is also known to block the production of inflammatory molecules and protect the endothelium by lowering the accumulation of Na⁺ and Ca²⁺.³⁶ However, the role of NHE1 inhibition as a cardioprotective effect of SGLT2 inhibitors is still under investigation, as some contrasting data demonstrate an unrelated to NHE1 effect of empagliflozin.²⁶

The role of inflammation

Inflammatory cytokines are one of the main stimuli of ventricular remodelling, contributing to the diastolic left ventricular dysfunction in HFpEF.³⁷ SGLT2 inhibitors may exert their beneficial effects through suppressing microvascular inflammation.

In vivo studies

SGLT2 inhibitors have been well documented to attenuate endothelial dysfunction by suppressing inflammation. Ganbaatar et al.³⁸ demonstrated that 8-week empagliflozin treatment in diabetic apolipoprotein E-deficient mice significantly ameliorated diabetes-induced endothelial dysfunction, as determined by the vascular response to ACh, and reduced vasoconstrictive eicosanoids such as PGE2 and TXB2, as well as inflammatory and adhesion molecules in aortic endothelial cells and in perivascular adipose tissue. Steven et al.¹⁷ also reported a reduction in the expression of inflammatory molecules such as interferon-γ (IFN-γ), cyclooxygenase-2 (COX2), and inducible NO synthase (NOS2) after SGLT2 inhibition.

Dapagliflozin, in a similar manner, is also reported to reduce several circulating inflammatory markers. In diabetic mice, dapagliflozin decreased the levels of MCP-1, CCL5, IL-1β, IL-6, and IL-17,¹⁹ while in euglycemic mice on a high-salt diet, it attenuated the upregulated NF-kB levels,³⁶ in addition to myocardial galectin-3 (an NF-κB activator), MCP-1, IL-6, and E-selectin. This anti-inflammatory effect of dapagliflozin has also been reported in mice after cardiac I/R injury.²⁴

Moreover, SGLT2 inhibitors may prevent the expression of the adhesive proteins ICAM-1 and VCAM-1, as shown by several studies upon murine models.^{16–18,20,21,23,24,34,36,38,39} They also reduce the expression of E-selectin³⁶ and P-selectin.¹⁷ These observations bear significant clinical relevance, as the upregulation of adhesion molecule expression indicates endothelial dysfunction and predisposes to atherosclerosis development.³⁹

In vitro studies

Consistent with the results of the aforementioned in vivo studies in regard to circulating inflammatory markers, Gaspari et al.³⁹ reports a significant suppression of TNF-α-mediated NF-kB expression in HUVECs stimulated with TNF-α.

Interestingly, SGLT2 inhibitors further exert their anti-inflammatory properties by protecting the endothelial integrity. VE-cadherins, claudin-5, and occludin are junctional proteins expressed on endothelial cells that play a vital role in endothelial integrity and vascular permeability. Studies on HCAECs after I/R injury or stretch-induced endothelial barrier dysfunction³³ showed that empagliflozin restores VE-cadherin expression by raising the sarcoma virus tyrosine kinase (Src) and the focal adhesion kinase (Fak), which regulate VE-cadherin activity.^20,23Dapagliflozin has also been found to revert stretch-induced VE-cadherin degradation.³³ In hBMECs stimulated with high glucose concentrations, empagliflozin restored the downregulated mRNA levels of claudin-5 and occludin.³¹

However, regarding adhesion molecule expression, the in vitro studies do not verify the results of the aforementioned in vivo models. Expression of ICAM-1 and VCAM-1 remained unchanged in human CMECs,²⁵ HCAECs,²² and HUVECs,²² all exposed to TNF-α, after treatment with an SGLT2 inhibitor. Furthermore, Cooper et al.⁴⁰ also demonstrated that empagliflozin does not significantly alter VCAM‑1 and ICAM‑1 expression on human abdominal aortic endothelial cells (HAAECs) under static and flow conditions. Collectively though, data suggest that SGLT2 inhibitors demonstrate a protective role towards the endothelial integrity and barrier function; an important effect, as it lessens the likelihood of microthrombus formation.^20,21,23

Another reported feature of empagliflozin is its ability to restore the integrity of the endothelial glycocalyx (GCX). The endothelial GCX is a dynamic structure that consists of proteoglycans, glycosaminoglycans, and glycolipids forming a functional layer lining the cell surface, which protects the microvascular environment. Loss of its integrity, caused for example by exposure to high levels of glucose or ROS, increases the atherogenic potential. Indeed, deterioration of endothelial GCX has been associated with endothelial dysfunction and lipid accumulation in the subendothelial space.⁴⁰ Cooper et al.⁴⁰ demonstrated that empagliflozin restores the level of heparan sulfate (HS) intensity, the most abundant component of the GCX, under static and flow culture conditions in heparinase III-treated HAAECs. Campeau et al.⁴¹ also showed the impact of GCX disruption on the induction of endoplasmic reticulum (ER) stress and the ability of empagliflozin to resolve the associated inflammatory response of endothelial cells through the transcriptional downregulation of key unfolded protein response genes. Their results indicated that empagliflozin attenuates EC inflammation independently of HS-mediated functions. Last but not least, of interest is the ERTU-SODIUM study, which is currently investigating the effects of SGLT2 inhibitors on glycocalyx, proteoglycans, and glycosaminoglycans in vivo in HFrEF patients.⁴²

The role of endothelial cell viability and angiogenesis

A variety of studies have investigated the impact of SGLT2 inhibitors on angiogenesis and EC's viability both in vitro and in vivo. It is well known that diabetes alters the dynamics of angiogenesis; therefore, treatment with SGLT2 inhibitors, on top of glycemic control, could attenuate the hyperglycemia-induced effects on the vasculature.³²

In vivo studies

Zhou et al.²¹ demonstrated upon a diabetic murine model that empagliflozin increases CMEC survival and delays vascular fibrosis, thus alleviating vascular degeneration, enriching microvascular density, and increasing cardiac perfusion. More specifically, empagliflozin preserved the angiogenic capacity of isolated CMECs through stabilization of F-actin, promotion of a well-arranged F-actin cytoskeleton and increased F-actin levels. Data are still conflicting; however, as Soares et al.⁹ report that empagliflozin decreased F-actin and P-cofilin content in the mesenteric arteries of a murine model. The reduction in F-actin and P-cofilin was associated with reduced arterial stiffness, which was also ameliorated by empagliflozin.

In vitro studies

In agreement with Zhou et al.,²¹ Ma et al.²⁴ similarly demonstrated the ability of dapagliflozin to sustain cytoskeletal integrity and preserve F-actin expression in vitro. Using HCAECs subjected to hypoxia-reoxygenation (H/R), they managed to demonstrate that dapagliflozin prevents H/R-mediated Ca2+/calmodulin-dependent protein kinase II (CaMKII) phosphorylation, which in turn leads to F-actin polymerization. This CaMKII inactivation is promoted by repression in calcium overload, which was found to be restricted by preventing XO-induced SERCA2 oxidation and inactivation. Interestingly, the inhibition of this pathway further alleviates oxidative stress, as the activation of XO in addition to cytoplasmic and mitochondrial calcium overload produce excess ROS, leading to reduced bioavailability of NO, microvascular contraction, and apoptosis.

Moreover, Nikolaou et al.³⁵ found that empagliflozin increases human microvascular endothelial cells viability in normoxia and hypoxia, with the protective effects being STAT-3 dependent under hypoxic conditions. They also demonstrated that empagliflozin upregulates VEGF levels, thus potentiating the survival and proliferation of endothelial cells. Indeed, STAT-3 and VEGF pathways have already been reported to have a mutual relationship. In support of these findings, dapagliflozin also restored VEGF expression in HCAECs after exposure to hypoxia-reoxygenation and counteracted the extent of apoptosis.²⁴ However, the mechanism by which SGLT2 inhibitors increase endothelial cell proliferation prior to the ischemia is still unclear and could also rely on other cells such as smooth muscle cells and fibroblasts.³⁵

Last but not least, dapagliflozin reversed the apoptosis in addition to the inhibition of angiogenesis caused by PA in HUVECs.³⁰ PA was used to mimic a high-fat environment. The expression of apoptosis-related proteins, including BAX, Bcl-2, and cleaved caspase 3, was also reduced by dapagliflozin.

Clinical trials: data on the effect of SGLT2 inhibitors in human studies

Even though the complex and interconnected pathophysiological mechanisms concerning the effect of SGLT2 inhibitors on the endothelium are not yet fully understood, evidence indicates their key role in the improvement of the endothelial function among a broad spectrum of patients. In this part of the review, we present trials assessing SGLT2 inhibition in endothelial cell function and microcirculation (Table 2). An attempt has been made to differentiate between the trials assessing endothelial function and microcirculation according to the used indexes. It should be noted, however, that endothelial and microvascular function are two concepts that overlap and share common underlying mechanisms, as microcirculation is regulated by both endothelial-dependent and independent mechanisms. Flow-mediated dilation (FMD) and reactive hyperaemia index (RHI) have been classified as endothelial function indexes, as the vasodilatory response to shear stress that they assess is thought to be NO-mediated. On the contrary, myocardial flow reserve (MFR) is assessed as a measure of microcirculatory function, mainly reflecting the myocyte tone of vessels, which is regulated in an endothelium-independent manner. Lastly, CFVR, is also considered a measure of microcirculatory function, as the stressors used to estimate it cause vasodilation through non-endothelial dependent mechanisms.

Trials regarding endothelial function

A variety of studies provide insight into the effect of SGLT2 inhibitors on endothelial function. Lunder et al.¹¹ studied 40 insulin-dependent male patients that were randomized into four groups: the empagliflozin, the metformin, the empagliflozin/metformin, and the placebo group. Endothelial function was assessed using brachial artery ultrasound-based FMD and RHI, measured at inclusion and after 12 weeks of treatment. Brachial artery FMD and RHI were significantly improved after treatment in all treatment groups and most prominently in the empagliflozin/metformin. In the placebo group, neither FMD nor RHI was found significantly changed. However, no comparison was made between the treatment groups and the control, thus the results of this study should be interpreted cautiously. The same research team also found a significant improvement in arterial stiffness after empagliflozin treatment compared to placebo, therefore, providing another mechanism through which SGLT2 inhibitors exert their cardioprotective properties.

Similarly, Irace et al.⁴³ investigated the effect of empagliflozin on endothelial function in patients with type 2 DM. The patients were treated either with empagliflozin (n = 20) or incretin (n = 15). Endothelial function was assessed by measuring brachial artery ultrasound-based FMD and shear stress at baseline and after 1 and 3 months. Patients treated with empagliflozin had a significant increase in both peak wall shear stress (61±20 vs. 68 ± 25 dynes/cm² after 3 months, P = 0.04) and FMD (4.8 ± 4.5% vs. 8.5 ± 5.6% after 3 months, P = 0.03), with no difference being observed in the incretin-based therapy group. The difference of peak wall shear stress and FMD was also significant when the empagliflozin group was compared to control (P < 0.05 and P = 0.04, respectively). Significant effects of empagliflozin were reported as early as after 1 month of treatment. The cohort was further divided into three FMD categories according to the time of maximal dilation such as early, late, and no dilators. Empagliflozin significantly increased the early dilators and decreased the late and no dilators, while no significant change was observed with the incretin-based therapy. This observation has important clinical relevance, as delayed dilation has been linked to underlying cardiovascular disease and cardiovascular risk factors.⁴³

Contrasting with the two previously described studies, the EMBLEM trial,⁴⁴ including a total of 117 adults with type 2 DM and cardiovascular disease, who were equally randomized to receive either empagliflozin or placebo for 24 weeks, showed no effect of empagliflozin treatment in endothelial function, as measured by RHI. It should be noted that the participants had relatively well-controlled concomitant diseases, such as blood pressure, BMI, and HbA1c.

Regarding dapagliflozin, a few studies have also investigated its effect on endothelial dysfunction. Sposito et al.⁴⁵ studied 98 diabetic patients with atherosclerotic disease, who were randomized 1:1 to 12 weeks of therapy with dapagliflozin or glibenclamide, in addition to metformin. Endothelial function was primarily assessed by measuring 1-min brachial artery ultrasound-based-FMD both at rest and after 15 min of ischemia followed by 15 min of reperfusion time (I/R). The post-treatment change in rest FMD at 1 min was significantly different between the two groups, with a median 3.3% increase being displayed in the dapagliflozin group, while the result remained statistically significant after adjustment for the office systolic and diastolic blood pressure change and pulse pressure (PP) change. Interestingly, dapagliflozin also improved the paradoxical arterial constriction that was observed in 30% of the patients after cuff deflation during the FMD protocol, which could be attributed to improved endothelial function. The authors additionally reported a significant difference between the two groups regarding the 1-min nitrite level, as well as the end diastolic velocity, peak systolic velocity, and resistive indices measured at 1 and 5 min. All aforementioned effects of dapagliflozin indicate an improvement in macro- and microvascular endothelial function in patients with type 2 diabetes mellitus and atherosclerotic burden.

In a similar manner, Zainordin et al.⁴⁶ investigated the effects of dapagliflozin on endothelial dysfunction in type 2 diabetics with established ischemic heart disease, but no significant difference in FMD was found neither between the dapagliflozin and the placebo group nor within each group. However, the placebo group showed a worsening of ΔFMD, whereas the dapagliflozin group did not display any change. The researchers also measured the nitroglycerin-mediated dilation (NMD) and found that after 12 weeks of therapy, it was significantly higher in the dapagliflozin group compared to placebo. The endothelial function was additionally evaluated by surrogate markers, one of which (ICAM-1) was significantly reduced after treatment with dapagliflozin (−83.9 ± 205.9 ng/mL, P < 0.02), in contrast to the placebo group, where it remained unchanged.

Shigiyama et al.⁴⁷ examined a total of 80 patients with type 2 DM who were randomized to receive either metformin or metformin on top of dapagliflozin for 16 weeks. FMD displayed a tendency towards improvement in the dapagliflozin group, but ΔFMD was comparable between the two groups. However, in the subgroup analysis of patients with HbA1c≥7.0%, FMD showed a significant improvement in the dapagliflozin group compared with the metformin group (ΔFMD = 1.05 ± 2.59 and −0.94 ± 2.39, respectively, P = 0.041). Considering the comparable glycemic control in the two groups, the authors speculate that dapagliflozin's action is mediated through mechanisms other than its glucose lowering effect.

Moreover, Solini et al.,⁴⁸ in a pilot study, showed a significant increase in FMD independent of blood pressure when 16 patients with type 2 DM were treated with dapagliflozin for 2 days (2.8 ± 2.2 to 4.0 ± 2.1%, P < 0.05), compared to 10 patients treated with hydrochlorothiazide, whose FMD did not present any significant change. In a subsequent study,⁴⁹ 40 patients with type 2 DM and hypertension were randomized to receive for 4 weeks either dapagliflozin or hydrochlorothiazide, but neither group presented any significant change in FMD, while miRNAs linked to endothelial dysfunction (miR-200b and miR-27b) were found equally increased in both groups.

Sugiyama et al.⁵⁰ studied 54 patients with uncontrolled type 2 DM, who were non-randomly allocated to either dapagliflozin or non-SGLT2 inhibitor medication for 6 months in order to assess the peripheral microvascular endothelial function, using reactive hyperemia-peripheral arterial tonometry (RT-PAT) and calculating the natural logarithmic transformed value of the RH-PAT index (LnRHI). Treatment with dapagliflozin resulted in a significant intra-group improvement of the endothelial function (LnRHI: from 0.450 ± 0.135 to 0.662 ± 0.230, P < 0.01). The absolute and percentage changes in the LnRHI were also significantly greater when the dapagliflozin group was compared to control (P = 0.028). The beneficial role of dapagliflozin is further highlighted by the fact that 74.1% of the patients in this group presented with an increase in LnRHI exceeding 15%, as opposed to the control group, where only 40.7% had a similar increase (P = 0.01).

In contrast to the results of the aforementioned studies, a double-blind, crossover, placebo-controlled trial⁵¹ demonstrated a worsening of microvascular endothelial function in 19 type 2 diabetic HFrEF patients after a 2-week treatment with dapagliflozin. The patients were specifically found with a reduced peripheral capillary-derived RHI [1.40 (1.23, 1.84) vs. 1.29 (1.21, 1.56), P < 0.05]. However, the results should be interpreted cautiously, as the number of participants was relatively small.

Trials on microcirculation

Several studies have investigated the effects of SGLT2 inhibitors on coronary microcirculation. The SIMPLE trial⁵² evaluated 90 patients with type 2 DM, who were randomly assigned to either empagliflozin or placebo for 13 weeks. MFR was quantified noninvasively by Rubidium-82 positron emission tomography/computed tomography (PET/CT) at inclusion and at week 13. There was no significant change in MFR between baseline and week 13 in either the empagliflozin or placebo groups. The result remained insignificant when adjusted for the rate–pressure product to account for baseline workload. The researchers also evaluated myocardial blood flow (MBF) during rest and stress, as well as reversible cardiac ischemia; however, no difference was observed between the groups. Some possible explanations behind the failure of the study to display a statistically significant result, as speculated by the authors, are the high prevalence of concomitant treatment that could have blunted the impact of empagliflozin on MFR, in addition to the relatively high baseline MFR of this cohort.

Lauritsen et al.⁵³ also studied the effect of empagliflozin on MFR and MBF in patients with type 2 DM. A total of 13 patients were randomized to either empagliflozin treatment or placebo and, after 4 weeks, ¹⁵O-H₂O PET/CT was used to evaluate MFR and MBF. Interestingly, compared to placebo, empagliflozin significantly reduced resting MBF by 13% (empagliflozin: 0.74 ± 0.10 mL/g/min; placebo: 0.85 ± 0.10 mL/g/min; P < 0.01), a result that remained significant even after adjustment for cardiac workload. On the other hand, empagliflozin treatment did not significantly affect stress MBF or MFR, but it should be noted that the latter displayed a tendency towards improvement (P = 0.09).

Moreover, Suhrs et al.⁵⁴ analysed the effect of empagliflozin on CFVR in type 2 diabetics. A total of 26 patients were randomized 1:1 to each treatment sequence, treated with empagliflozin and placebo for 12 weeks with a 2-week wash-out period. CFVR was measured with transthoracic doppler echocardiography. No significant effect was observed on CFVR, neither after empagliflozin treatment nor placebo, thus not supporting the hypothesis that the beneficial effect of SGLT2 inhibitors is due to improvement of the non-endothelial dependent coronary microvascular function. When empagliflozin treatment was compared with placebo, a significant difference (P = 0.044) was shown in hyperemia CFV, but it was small and regarded as a change finding.

Finally, the DAPAHEART trial⁵⁵ demonstrated for the first time in humans a significant increase in MFR after treatment with a SGLT2 inhibitor. A total of 16 patients with type 2 DM and stable coronary artery disease were equally randomized to receive for 4 weeks either dapagliflozin or placebo. MFR was assessed using ¹³N-ammonia PET/CT. Patients in the dapagliflozin group showed a significant improvement in MFR (2.56 ± 0.26 vs. 3.59 ± 0.35, P = 0.006) compared with the placebo group. The result remained significant even after correction for resting rate–pressure product (2.22 ± 0.25 vs. 3.23 ± 0.4, P = 0.008). The dapagliflozin arm also displayed a significantly lower resting MBF and a trend towards an increased stress MBF, compared with the placebo group.

Just as importantly, SGLT2 inhibitors have also been found to improve arterial stiffness and peripheral vascular resistance. A double-blind, placebo-controlled randomized clinical trial⁵⁶ investigated the effects of empagliflozin in nondiabetic patients with HFrEF and found that pulse wave velocity (PWV) was significantly reduced in the treatment group compared to controls (−0.58 cm/s vs. 0.60 cm/s, P < 0.01). Similarly, in a post hoc analysis⁵⁷ of data from phase III trials including T2DM patients, empagliflozin was shown to significantly (P < 0.001) reduce PP, mean arterial pressure, and double product (DP or RPP). There was also a trend towards reduction in the ambulatory arterial stiffness index with empagliflozin in one of the cohorts (P = 0.059).

Clinical perspectives and future directions

Currently SGLT2 inhibitors constitute the only drug class that has shown positive results in HFpEF and is suggested by heart failure guidelines^8,59 for treating patients with HF regardless of their EF. Endothelial and microvascular dysfunction have a key role in the pathogenesis of HF, with myocardial and endothelial inflammation, oxidative stress and impaired vasodilation being some of the microcirculation-related pathogenetic mechanisms. Notably, specifically in HFpEF, MVD is highly prevalent, reaching up to 70–80% of patients.^60,61 In addition, the association of common comorbidities among HFpEF patients (hypertension, metabolic syndrome, and dyslipidemia) with the development of endothelial dysfunction are suggestive of the key role MVD has in the pathogenesis of HFpEF.¹⁵ Intriguingly, other agents acting on the endothelium and microcirculation, such as ACE and PDE-5 inhibitors, have failed until now to show any clinical benefit in HFpEF.

As thoroughly described thus far, plenty preclinical and clinical data demonstrate the endothelial and microcirculatory function improvement induced by empagliflozin and dapagliflozin. Yet still, these studies have limitations. In this area of research with small animal studies, it is likely that there is a publication bias with some neutral results not being published. Furthermore, the in vivo studies have almost exclusively used mice and rats as their models, thus deterring us from extrapolating the results to humans. The in vitro studies using human cells overcome this limitation; however, these studies are limited. Moreover, the cells used are of high diversity in origin and developmental stage, in addition to being cultured mainly under static conditions. Lastly, the evaluation of methodological quality in animal studies—as opposed to human clinical trials—is challenging due to lack of standardized assessment scales.

Regarding clinical trials, most of them verify the beneficial effects displayed in the preclinical models. However, an important aspect that should be clarified is that some of the effects of SGLT2 inhibitors on the endothelium and microcirculation could also be attributed to the accompanying weight loss. On top of that, a few trials have failed to show significant beneficial SGLT2 effects and they should not be disregarded. The conflicting results might be explained by the differences seen in patient population, treatment protocols, indexes assessed, and no thorough pre-protocol assessment of the microcirculation status. This highlights the need for more intensified research on this particular topic. Additionally, there are still no trials assessing endothelial or microvascular functional indexes after treatment with SGLT2 inhibitors specifically in patients with HFpEF. Therefore, the focus from now onwards should be redirected on examining the impact of SGLT2 inhibition on the microcirculation of HFpEF patients. Finally, MVD characterizes several other cardiovascular diseases with limited therapeutic options, such as Myocardial Infarction with No Obstructed Coronary Arteries, Refractory Angina, and Takotsubo Cardiomyopathy. Given the pleiotropic effects of SGLT2 inhibitors on microcirculation and endothelial function, it is possible that this drug class could be of therapeutic benefit in these patients. Thus, future trials should assess whether SGLT2 inhibition could provide clinical benefit in other pathologies with microcirculation disorders, besides HFpEF.

Conclusion

SGLT2 inhibitors have well-documented cardioprotective effects and constitute a new therapeutic option for heart failure patients regardless of the presence of diabetes. Interestingly, preclinical studies establish several molecular mechanisms whereby empagliflozin and dapagliflozin improve endothelial and microcirculatory function, while the clinical trials verify these positive effects. However, there is still much debate about the exact underlying mechanisms for the impact of SGLT2 on the cardiovascular system, as conflicting data still exist. Therefore, larger scale studies are needed to verify the existing evidence and shed light on the complex responsible pathways for the rather unique beneficial actions of SGLT2 in humans.

Acknowledgement

None.

Funding

None.

Conflict of interest: None.

Data availability statement

The data for this review are available under request from the authors.

References