Insights into SGLT2 inhibitor treatment of diabetic cardiomyopathy: focus on the mechanisms

Diabetes may have been recognized more than 3000 years ago [1], and the incidence and prevalence of this ancient disease have increased with improvements in living and health standards. Currently, 1 in 11 people has diabetes, which is 4 times more frequent than 30 years ago [2]. In addition to eye, kidney, and foot disease, chronic hyperglycaemia can damage the myocardium, which manifests as diabetic cardiomyopathy. The pathology of diabetic cardiomyopathy includes abnormal carbohydrate and fatty acid metabolism, microvascular disease, endothelial dysfunction, mitochondrial dysfunction, myocardial diastolic dysfunction and fibrosis, excessive cardiomyocyte programmed cell death, increased oxidative stress, endoplasmic reticulum (ER) stress, and glycotoxicity [3‐5]. Adverse responses to abnormal blood glucose lead to structural remodelling and impaired function in the heart. Diabetic cardiomyopathy is known to significantly increase the risk of heart failure and can lead to a preserved or reduced ejection fraction [6].

Sodium-glucose cotransporter 2 inhibitors (SGLT2is) have long been used as drugs to treat diabetes and have recently been shown to improve cardiac function in diabetes. In a randomized controlled trial, dapagliflozin reduced heart failure by 27% in T2DM (type 2 diabetes mellitus) patients with cardiovascular disease. The effectiveness was 35% for empagliflozin and 33% for canagliflozin [7]. This review focuses on the benefits of SGLT2i for cardiomyopathy in diabetes, including the mechanism (Table 1). Future prospects are also discussed.

The first SGLT inhibitor was dihydrochalcone phlorizin, which is a nonselective SGLT inhibitor extracted from apple tree roots [8]. Dihydrochalcone phlorizin contains a glucose moiety and an aglycone in which two aromatic carbocycles are joined by an alkyl spacer. Later, the aromatic O-glycoside sergliflozin and the aromatic C-glycoside dapagliflozin officially opened the era of selective SGLT inhibitors [9]. The available SGLT2i are functionally similar but differ in selectivity, efficacy, and indication. For example, empagliflozin, dapagliflozin, and canagliflozin are 2600, 1200, and 150 times more selective for SGLT2 than for SGLT1 [10]. Ipragliflozin and dapagliflozin were approved to treat T1DM (type 1 diabetes mellitus) and T2DM, but most other SGLT2i were approved to treat only T2DM [10]. There are also some obvious differences among them; for example, compared with dapagliflozin and empagliflozin, canagliflozin has a stronger inhibitory effect on angiogenesis [11, 12].

The ways in which SGLT2 functions in the heart are not fully understood. SGLT1 is expressed by cardiomyocytes and may be a target of SGLT2i that improves heart failure [13]. Inhibiting SGLT1 with canagliflozin was shown to reduce nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activity in cardiomyocytes, thereby inhibiting the production of superoxide and decreasing inflammation [13]. SGLT2i may be useful as a first-line treatment for heart failure that is not only mediated by their target receptors. For example, dapagliflozin and canagliflozin were reported to directly inhibit Na⁺/H⁺ exchanger-1 (NHE1) and abrogate the increase in cytosolic Na⁺ in cardiomyocytes. Direct activation of AMP-activated protein kinase (AMPK) by dapagliflozin could reduce lipopolysaccharide (LPS)-induced myocardial fibrosis. That kind of intermolecular interaction is not dependent on SGLT2i-mediated inhibition of the sodium-glucose cotransporter [14]. Other studies have shown that the cardiac sodium channel Nav1.5 was a target of SGLT2i and that inhibiting sodium channels ameliorated dysfunctional sodium and calcium homeostasis, improved calcium overload, and decreased the incidence of malignant arrhythmias [15]. Ongoing study of SGLT2i is expected to reveal additional mechanisms of action.

Disorders of glucose and fatty acid metabolism are prominent features of diabetic cardiomyopathy. Under normal conditions, fatty acids are the first-choice energy source and account for 70–90% of the ATP produced by cardiomyocytes. Although each molecule of fatty acid produces more ATP than glucose, complete oxidation requires more oxygen. When the same amount of oxygen is consumed, fatty acids produce less ATP than glucose (2.33 vs. 2.58) [16]. However, compared with the absence of diabetes, oxidative metabolism in cardiomyocytes under diabetic conditions consumes a higher percentage of fatty acids and a lower percentage of glucose [17], and there is an increase in the use of ketone bodies [18]. The presence of diabetes increases oxygen consumption and decreases efficiency in the heart [17].

Aerobic respiration that occurs in the mitochondria of cardiomyocytes provides energy for cardiac contraction. Mitochondrial dysfunction occurs in diabetic cardiomyopathy as a consequence of high glucose, insulin resistance, and obesity. The consequences of these pathological changes include but are not limited to increased mitochondrial fission, autophagy, reduced oxidative phosphorylation and impaired ATP production, the accumulation of metabolic intermediates and reactive oxygen species (ROS), oxidative stress, apoptosis, and impaired mechanical function in the heart [38‐40].

Mitochondria frequently divide and fuse with each other [41]. When cells are subjected to mild stress, mitochondria form an extensive interconnected network. When cells are severely stressed, mitochondria undergo fission and are fragmented [42]. In response to hyperglycaemia, cardiomyocyte mitochondria undergo dynamin-related protein 1 (Drp1)-mediated fission, which, if excessive, can result in fragmentation, ROS production, increased oxidative stress and even cell death [43]. Tanajak et al. [44] studied the inhibitory effect of SGLT2i on excessive mitochondrial fission after cardiac ischaemia/reperfusion (I/R) injury in the hearts of obese rats with high-fat diet-induced insulin resistance. Dapagliflozin reversed the impairment in mitochondrial morphology and increased ROS levels. The total myocardial Drp1 protein level did not change significantly, but the myocardial mitochondrial Drp1 level decreased significantly [44]. The myocardial phosphorylation site Drp1 Ser-637 was activated by dapagliflozin, which may have accounted for the reduction in the mitochondrial translocation of Drp1 [45]. These data are consistent with those of Zhou et al. [46], who reported that the upstream signalling molecule of Drp1 was AMPK. Mitochondrial fusion and fission are also associated with mitofusins (MFNs) and optic atrophy 1 (OPA1). MFN1 and MFN2 mediate fusion of the mitochondrial outer membrane, OPA1 mediates inner membrane fusion [47], and the absence of any one of the three causes cells to produce different types of fragmented mitochondria [48, 49]. Therefore, MFNs and OPA1 are targets of SGLT2i therapy. Durak et al. [50] reported that dapagliflozin normalized the increase in MFN1, decrease in MFN2, and increase in OPA1 expression in the myocardium of rats with sucrose-induced metabolic syndrome. These results have also been reported for empagliflozin [51] (Fig. 1).

Mitochondrial biogenesis and the growth and division of preexisting mitochondria are regulated by AMPK, peroxisome-proliferator-activated receptor γ coactivator-1α (PGC-1α), and nuclear respiratory factors (NRF)-1 and -2 [52]. AMPK activation was shown to increase the transcription of the PGC-1α gene in rat cell nuclei, which then induced the expression of the transcription factor NRF-1 [53]. Mitochondrial transcription factor A (mtTFA) is a mitochondrial promoter that stimulates transcription. The proximal promoter of the human mtTFA gene is dependent on the activity of the recognition sites of the nuclear respiratory factors NRF-1 and NRF-2 [54]. Activation of PGC-1α, NRF-1, and mtTFA increases the transcription and replication of mitochondrial DNA. PGC-1α can also activate the expression of key enzymes in mitochondria, including cytochrome c oxidase [55]. This will allow the mitochondrial electron transport chain (ETC) to increase the efficiency of electronic transfer. These regulators of mitochondrial biogenesis are inhibited to varying degrees in diabetes. However, empagliflozin reversed the downregulation of PGC-1α, NRF-1, and mtTFA in a streptozotocin (STZ)-induced rat model of T2DM [51] (Fig. 1).

The transfer of electrons in the ETC generates mitochondrial membrane potential. The reduced activity of ETC complexes in the hearts of diabetes patients is accompanied by a reduction in the mitochondrial respiration rate [56]. The decreases in mitochondrial state 3 respiration and mitochondrial membrane potential in rats with STZ-induced diabetes were reversed by high-dose empagliflozin [51]. Secker et al. [57] found that canagliflozin inhibited the activity of mitochondrial respiratory complex I and promoted complex II activity, and empagliflozin and dapagliflozin increased the activity of complex I. Improvements in mitochondrial function by SGLT2i may be related to a decrease in O-GlcNAcylation. The increase in O-GlcNAcylation secondary to hyperglycaemia leads to a decrease in the activity of ETC complexes I, III, and IV. This change is mainly due to the O-GlcNAcylated subunits that make up the respiratory chain complexes [58]. Dapagliflozin and other SGLT2i may improve the function of the mitochondrial respiratory chain by directly reducing O-GlcNAc transferase activity and O-GlcNAcylation [59].

Diabetes is known to increase the risk of cardiovascular events and cause myocardial microvascular complications [60, 61]. Damage to vascular endothelial cells caused by oxidative stress is followed by a series of pathological changes that include vascular inflammation, vasoconstriction, thrombosis and atherosclerosis [62]. Diabetes and glycotoxicity promote coronary atherosclerosis by exacerbating endothelial dysfunction and increasing oxidative stress, blood lipids, and autonomic dysfunction [63].

Myocardial fibrosis occurs in cardiomyopathy associated with diabetes, and it involves the accumulation of advanced glycation end products (AGEs) [94] and increased myocardial stiffness that interferes with ventricular diastole. Chronic hyperglycaemia is accompanied by the formation of AGEs that are produced by the nonenzymatic combination of glucose and proteins. Activation of the receptor for AGE (RAGE) by AGEs promotes the proliferation, function, and migration of cardiac fibroblasts, which exacerbates myocardial fibrosis and accelerates cardiac ageing [94‐96]. AGEs, RAGE and downstream protein kinase C (PKC)-ζ and mitogen-activated protein kinase (MAPK) promote tissue fibrosis. Activation of the MAPK subfamily ERK is the most significant event in this process. After being activated, ERK translocates to the nucleus, where it influences transcription factors such as cAMP-response element binding protein, ETS domain-containing protein-1 and Y-box binding protein-1 to regulate cell proliferation, differentiation, and extracellular matrix accumulation [94, 97, 98]. Empagliflozin can inhibit the AGE/RAGE axis in the kidney, and it is believed that evidence in the heart will follow [99].

Transforming growth factor-beta (TGF-β) also regulates tissue fibrosis via the SMAD protein, which is an intracellular effector downstream of the TGF-β receptor. Epithelial-myofibroblast transition and collagen expression are promoted by SMAD2 and SMAD3 and inhibited by SMAD7 [100, 101]. A recent study showed that myocardial fibrosis was improved by SGLT2i [102]. In a mouse model of type 2 diabetes, myocardial expression of collagen I and collagen III proteins and connective tissue fraction was increased compared with that in nondiabetic mice and was partially reversed by empagliflozin [102]. TGF-β/SMAD signalling was involved because, compared with that in untreated mice, the expression of TGF-β1, p-Smad2, and p-Smad3 in the heart tissue of mice treated with empagliflozin was significantly reduced (Fig. 4). Other studies have shown that blocking NOD-like receptor 3 (NLRP3) reduced myocardial fibrosis [103]. Interleukin (IL)-1β promotes TGF-β gene expression and promotes fibrosis through the TGF-β pathway. Other reasons involve nonpolymeric NLRP3 protein, but the mechanism is not very clear [104]. The TGF-β pathway is discussed in the pyroptosis section of this review (Fig. 5).

Diabetes increases the risk of myocardial infarction [60], and the infarct size is larger in patients with diabetes than in those without diabetes [105]. SGLT2i have been shown to improve ventricular remodelling after myocardial infarction and to alleviate cardiac fibrosis by modulating macrophage polarization. The mechanism is similar to that for atherosclerosis, as previously discussed. Studies have shown that in the postmortem myocardium, IL-10 can indirectly affect fibroblast activation by stimulating M2 macrophage polarization, thereby significantly reducing the level of collagen I, reducing the ratio of collagen I to collagen III in cardiac fibroblasts, and reducing collagen accumulation in the infarcted area of the mouse heart [106]. The increase in collagen I leads to increased fibril width and stiffness [107]. The effect of SGLT2i on this pathway was demonstrated by Lee et al. [108], who found that dapagliflozin induced macrophage polarization to the M2 phenotype and inhibited the M1 phenotype by stimulating signal transducer and activator of transcription 3 (STAT3) signalling in mice with myocardial infarction and attenuated the increase in collagen after infarction.

The effect of diabetes on ventricular diastolic dysfunction may be related to a reduction in cyclic guanosine monophosphate (cGMP) and a decrease in PKG activity because of increased nitrosative and oxidative stress and decreased bioavailability of NO [109, 110]. Oxidative and nitrosative stress caused by diabetes are discussed later in this review. PKG phosphorylates N2BA and N2B, the two main cardiac titin isoforms in the human left ventricle, resulting in a decrease in cardiomyofibrillar stiffness [111]. In human myocardial tissue from patients with heart failure with preserved ejection fraction (HFpEF) and ZDF obese rats, PKGIα was oxidized and was present as a dimer or polymer in the cardiomyocyte cell membrane. Empagliflozin decreased PKGIα oxidation and translocation of the reduced form into the cytoplasm of cardiomyocytes. Empagliflozin restored the damaged NO/soluble guanylate cyclase (sGC)/cGMP/PKG pathway in the HFpEF myocardium, significantly increased NO and cGMP concentrations, and increased sGC and PKGIα activity [112], which reduced cardiomyocyte stiffness. These changes were observed macroscopically. Echocardiography after intravenous injection of empagliflozin revealed significantly improved diastolic ventricular function in HFpEF rats and in humans [113].

Oxidative stress results from an imbalance in oxidative and antioxidative activity, and an excess of intermediate oxidative products damages cells. Hyperglycaemia and glucotoxicity result in oxidative stress in diabetes [114, 115], which is associated with lower antioxidant levels and higher oxidant levels than those in the absence of diabetes [116]. Therefore, a decrease in oxidative stress in the myocardium of diabetic patients would be key supporting evidence of their benefit in diabetic cardiomyopathy.

Diabetes promotes programmed cells in the myocardium [136, 137], which results in decreased cardiac contractility, heart failure, and other complications. In the classic apoptosis pathway, the activation of caspase-3, caspase-6 and caspase-7 causes membrane blebbing, cell shrinkage, the formation of apoptotic bodies, and chromosomal DNA fragmentation. The pyroptosis pathway is mediated by caspases-1, -4, -5, and -11, which can cleave the cytosolic protein gasdermin D, and the latter can form large oligomeric pores in the inner layer of the plasma membrane and intracellular organelles to kill the cell [138].

Autophagy digests long-lived proteins and cytoplasmic organelles to meet the metabolic needs of the cell and the renewal of certain organelles [154, 155]. Autophagy imbalance and disruption occur in diabetes. Acute induction of autophagy may be beneficial, but persistent autophagy induction may be harmful [156, 157].

The effects of SGLT2i on autophagy have been studied in diabetic cardiomyopathy. Empagliflozin has been reported to increase autophagy in the atrial tissue of ZDF rats. This treatment increased the microtubule-associated protein light chain 3 (LC3) II/I ratio and decreased the protein expression of p62 in the atrial tissues of ZDF rats [25]. Li et al. [158] reported that empagliflozin decreased the protein expression of CD36, which was associated with an increase in p-AMPK, which activated the AMPK/Unc-51-like kinase 1 (Ulk1)/Beclin1 pathway, increased Ulk1 and Beclin1 expression, and promoted autophagy. This process in liver cancer cell experiments does not seem to involve mammalian target of rapamycin (mTOR). However, dapagliflozin decreased p-mTOR/mTOR in rat colitis model cells, suggesting that the inhibition of mTOR by AMPK [159] is involved in AMPK-mediated enhancement of autophagy [160]. The difference might be related to nutritional status in the two models because AMPK activates Ulk1 to promote autophagy during glucose starvation. In the presence of adequate nutrients, high mTOR activity phosphorylates Ulk1 at Ser 757, which disrupts the interaction between Ulk1 and AMPK to prevent Ulk1 activation [161].

SGLT2i do not enhance autophagy unilaterally. Jiang et al. [162] reported that in T1DM or T2DM and myocardial infarction, cardiomyocyte survival benefitted from the inhibition of enhanced autophagy by empagliflozin, which depended on the downregulation of NHE1 and NHE1-related genes that induce autophagy, such as Beclin 1 and autophagy-related protein 5. In fact, the inhibitory effect of SGLT2i on NHE1 has been widely demonstrated [163], but it is not clear whether Beclin1 is a downstream target of NHE1. Deng et al. [164] found that Beclin1 but not NHE1 was targeted by empagliflozin. Empagliflozin was shown to inhibit the increase in autophagy caused by myocardial infarction with acute hyperglycaemia. Empagliflozin can also reverse the increase in the autophagy-related protein LC3II/I and decrease in P62 in cardiomyocytes induced by Tat-beclin1. This difference may be related to the difference in mouse models; the former model is a T2DM model with myocardial infarction, and the latter model is a nondiabetic myocardial infarction model with acute hyperglycaemia. However, these results show that the regulatory effect of empagliflozin on autophagy involves maintaining a balance, rather than unilaterally enhancing or weakening autophagy.

ER stress involves disturbances in Ca²⁺ or redox balance and the accumulation of misfolded or unfolded proteins that cannot be processed, which initiates an unfolded protein response (UPR) that leads to apoptosis [165]. The UPR involves three transducers: protein kinase RNA-like endoplasmic reticulum kinase (PERK), activated transcription factor 6 (ATF-6) and inositol-requiring protein-1α (IRE1α). The transducers bind to the ER chaperone glucose-regulated protein 78 (GRP78), and as unfolded proteins accumulate, GRP78 leaves the transducer and is involved in processing the accumulated proteins [166]. This leads to the activation of these three sensors and subsequent lethal effects [167]. Currently, the correlation between diabetes and ER stress has been well established [168].

PERK/eIF2α/ATF4/C/EBP-homologous protein (CHOP) are signalling pathways involved in ER stress. CHOP is a transcription factor that downregulates BCL2, BCL-XL, and MCL-1 expression and upregulates Bim, Bak and Bax expression. CHOP also upregulates the expression of the pseudokinase tribbles homologue 3 gene, which was shown to weaken the inhibition of Caspse-9 and Caspase-3 expression by AKT [169‐171].

SGLT2i ameliorate ER stress in animal models induced by heart-pressure overload or I/R injury [172, 173]. In these ER stress models induced by heart pressure overload or I/R injury, dapagliflozin and empagliflozin inhibited the increase in p-PERK and its downstream molecules associated with ER stress due to pressure overload by activating SIRT1 and preventing GRP78 detachment [172, 173]. Cell death was also significantly weakened. Dapagliflozin could significantly reduce GRP78, PERK, eIF-2α, ATF-4, and CHOP expression in the myocardium in an ER stress model induced by doxorubicin [174] (Fig. 6). SGLT2i have been shown to inhibit ER stress through similar downstream pathways in various organs. Ipraglifilozin inhibited the increase in GRP78 and PERK and their downstream signalling molecules associated with ectopic lipid deposition in mouse kidneys [175].

The impact of intestinal flora imbalance on extraintestinal organs has received increasing attention in recent years, and the heart is no exception. The damaged intestinal wall can lead to the entry of intestinal bacteria into the circulation, causing inflammatory responses in multiple organs [176]. After entering the circulation, LPS from gram-negative bacteria is recognized by TLRs on the surface of immune cells and induces the release of proinflammatory cytokines [177]. Continuous infusion of low-dose LPS to mimic metabolic endotoxaemia leads to obesity, insulin resistance, T2DM, and atherosclerosis [178]. The metabolism of cholesterol and lipids by the gut microbiota affects the development of atherosclerotic plaques [179], and bacterial metabolites, such as trimethylamine N-oxide (TMAO), were shown to lead to myocardial fibrosis, abnormal metabolism, impaired endothelial function, and heart failure following the activation of various signalling pathways [180]. These pathological changes are consistent with the development of diabetic cardiomyopathy.

Dapagliflozin has been associated with changes in the gut microbiota, the function of extraintestinal blood vessels, and improvements in cardiovascular dysfunction caused by diabetes. Eight weeks of dapagliflozin treatment significantly increased Akkermansia muciniphila in the gut microbiota and improved generalized vascular dysfunction in mice with T2DM [181]. An increase in the abundance of A. muciniphila and improvements in glucose tolerance and blood glucose levels have been confirmed and correlated with induction of Foxp3-positive regulatory T cells [182]. An increase in A. muciniphila was also shown to reduce endotoxaemia and inflammation and prevent atherosclerosis by inducing the expression of tight junction proteins in the gut [183]. In addition, short-chain fatty acids (SCFAs) have been shown to thicken the mucin layer and strengthen the intestinal barrier by stimulating the release of IL-22 from lymphocytes, which prevents endotoxins and LPS from entering the body [184, 185]. Luseogliflozin increased the abundance of Syntrophothermus lipocalidus, family Syntrophomonadaceae, Parabacteroidesdistasonis distasonis, and genus Anaerotignum, which produce SCFAs [186]. Empagliflozin has been associated with an increase in the population of SCFA-producing bacteria and improvements in diabetes and cardiovascular function [187]. Roseburia, Eubacterium, and Faecalibacterium species were increased by empagliflozin treatment, and harmful bacteria, including Escherichia and Shigella, were decreased.

SGLT1 is the predominant receptor in the gut compared with SGLT2 [188]. Studies of the dual SGLT1/2 inhibitor canagliflozin and the SGLT1 inhibitor SGL5213 showed that these agents reversed the expansion of Firmicutes and contraction of Bacteroidetes in the gut microbiota of mice with adenine-induced renal failure [189, 190]. Increases in SCFAs such as acetate, butyrate, and propionate in the caecum and decreases in plasma TMAO levels were also found in a renal failure mouse model after treatment. These findings suggest that inhibition of SGLT1 reduces glucose uptake, while unabsorbed glucose temporarily reaches the lower small intestine [191], resulting in an altered glucose load and a possibly altered gut microbiota composition.

The goal of basic research is to support clinical use. We have reviewed animal and clinical trials that evaluated the therapeutic effect of SGLT2i. The results of some animal trials support the performance of clinical trials. Here, we show some representative clinical evidence in response to the results of theoretical research (Table 2).

First, SGLT2i improves myocardial metabolism in diabetic patients. SGLT2i (dapagliflozin, empagliflozin, canagliflozin) reduce the expression of PPAR-γ in the hearts of patients. Similar to the mechanism described above, this effect can reduce the accumulation of fatty acids in myocardial cells [192], which is exactly what happened.

Empagliflozin is encouraging in terms of improving ventricular remodelling in diabetic patients. For example, in a clinical randomized controlled trial, the SGLT2i empagliflozin significantly reduced the left ventricular mass relative to body surface area [193]. Similarly, a randomized controlled trial of dapagliflozin demonstrated a significant reduction in left ventricular mass (LVM) in patients with T2DM and left ventricular hypertrophy (LVH) [194].

In diabetes patients with atherosclerotic disease, endothelial changes occur in the microvasculature and macrovasculature, and 12 weeks of dapagliflozin treatment resulted in a significant increase in FMD [195]. Another study demonstrated that 4 weeks of dapagliflozin increased myocardial flow reserve (MFR) in patients with stable coronary artery disease and T2DM [196]. In addition, clinical studies have confirmed that dapagliflozin can prevent microcirculation remodelling in diabetes patients and can decrease the stiffness of large blood vessels [197]. T2DM patients with acute myocardial infarction (AMI) who had received long-term treatment with SGLT2i before admission had smaller infarct sizes and lower inflammatory markers than those receiving other oral hypoglycaemic agents [198].

The effectiveness of SGLT2i for improving hyperuricaemia has been clinically verified. Dapagliflozin, empagliflozin, and canagliflozin have been shown to promote uric acid excretion in patients with or without diabetes [199‐201]. This evidence shows that SGLT2i maintain cardiovascular function and reduce the risk of cardiovascular events.

The regulation of intestinal flora by SGLT2i is less clear. In a 12 week double-blind randomized trial, significant changes in gut microbiota diversity or composition were not observed in T2DM patients after dapagliflozin treatment [202]. The results may have been influenced by the administration of other drugs, including previous metformin monotherapy. In addition, the doses of dapagliflozin that are appropriate for treating diabetes may not be appropriate for regulating the gut microbiota. Additional studies of the impact of SGLT2i on the human intestinal microbiota are needed.

Several meta-analyses have provided more extensive and persuasive evidence that SGLT2i improve diabetes-induced cardiac insufficiency. T2DM increases the risk of atrial fibrillation (AF) and atrial flutter (AFL), and dapagliflozin significantly reduces AF and AFL [203]. Treatment of patients with AHF with SGLT2 inhibitors reduced the risk of rehospitalization due to heart failure and improved Kansas City Cardiomyopathy Questionnaire (KCCQ) scores [204]. SGLT2i improved diabetes-associated cardiac structure and function. A meta-analysis showed that SGLT2i partially reduced plasma NT-proBNP concentrations and improved cardiac diastolic function. However, SGLT2i improve left ventricular ejection fraction (LVEF) only in heart failure with reduced ejection fraction (HFrEF) in stage C heart failure [205]. On the other hand, in patients with HFpEF, especially those with stage A-B heart failure, the effect of SGLT2i treatment is not significant.

SGLT2i are safe, well-tolerated drugs. They can improve all glycaemic parameters and have some additional benefits, such as weight and BP reductions, low risk of hypoglycaemia, improvements in β-cell function and insulin sensitivity, and reductions in macrovascular and microvascular events [206]. However, there are still many challenges to be solved in clinical practice. These challenges include but are not limited to decreased blood pressure, urinary and genital tract infections, amputation, ketoacidosis, kidney injury and fracture. Reduced blood pressure in SGLT2i users is understandable, which is caused by insufficient circulating blood volume due to osmotic diuresis. Meta-analyses showed that SGLT2i induced an average reduction in systolic/diastolic BP of 3.62/1.70 mmHg in 24 h ambulatory blood pressure [207]. However, the use of SGLT2i alone was very unlikely to cause postural hypotension. The use of SGLT2i has been reported to increase the risk of urinary and genital tract infections by more than 3 times [208]. The main reason may be related to the increase in urine sugar concentration. As previously mentioned, canagliflozin caused twice the risk of lower limb amputation compared to the control, possibly due to the inhibition of angiogenesis [209]. An increased incidence of ketoacidosis has been reported in SGLT2i users [210]. The prevailing belief is that this outcome is associated with glucose loss, increased hyperglucagonemia, constant or decreased insulin levels, mild infections, and decreased blood volume. This mechanism may induce euglycaemic diabetic ketoacidosis. The evidence that SGLT2i may cause acute renal injury is available [211], and the associated factors include circulatory failure, increased uric acid in urine, and multiple drug use [212, 213]. However, this treatment is not recognized as an inducer of acute renal injury, and several clinical trials have confirmed that SGLT2i have positive effects on kidney function [214]. In assessing the increased risk of fracture associated with SGLT2i, canagliflozin was directly associated with bone mineral density loss, especially in the hip. Dapagliflozin was associated with fractures unrelated to osteoporosis, mainly due to an increased risk of falls caused by fluctuations in blood pressure and hypoglycaemia caused by the combined use of hypoglycaemic drugs. However, empagliflozin was considered safe [215]. In addition, there are still some unresolved questions. The limited clinic evidence showed that SGLT2i improves the concentration of circulating proteins in plasma, serum or urine that are known to have beneficial effects on the heart [216]. These proteins include insulin-like growth factor-binding protein 1, transferrin receptor protein 1, erythropoietin and so on. But the exact target and signalling pathway of SGLT2i in clinic treatment are still not fully understood. The optimal dose of SGLT2i for diabetic cardiomyopathy and whether this dose is consistent with the dose for diabetic treatment still need experimental verification [217, 218].

In light of the current challenges associated with SGLT2i, we reviewed some of the guidelines and provided some insight into future strategies for the use of SGLT2i [218]. The use of SGLT2i is strongly recommended in patients with HFrEF and with or without T2DM, patients with T2DM and coronary heart disease, patients with T2DM were over 50 years old who had risk factors associated with coronary heart disease, and patients with albuminuric renal disease with or without T2DM. The benefits of SGLT2i on these patients have been well documented and reduce the risk of HF hospitalization. Moreover, we may avoid some possible complications by individualized medications and appropriate periodic examinations. For patients with suspected ketoacidosis symptoms such as nausea, vomiting, abdominal pain, and dyspnoea during SGLT2i use, blood and urine ketone body analysis should be performed in time. Patients who use SGLT2i should be advised to pay attention to perineal hygiene. For patients with urinary tract infection during medication use, SGLT2i administration should be paused, and anti-infective therapy should be given. SGLT2i should be discontinued in patients with recurrent urinary tract infections. Patients taking SGLT2i and diuretics together need to monitor their blood pressure. Canagliflozin is contraindicated in patients with lower extremity vascular stenosis or with osteoporosis. For renal insufficiency, patients with glomerular filtration rates less than 30 mL/min/1.73 m² should be treated with low-dose SGLT2i or not treated with SGLT2i.

In this review, we mainly focused on basic studies to elucidate the pharmacological mechanism of SGLT2i and related signalling pathways or targets for the treatment of diabetic cardiomyopathy. Studies have shown that SGLT2i can improve diabetic cardiomyopathy by improving myocardial metabolism, restoring mitochondrial function, improving microcirculatory disorders, reducing myocardial fibrosis, reducing oxidative stress, inhibiting programmed death, regulating autophagy, inhibiting ER stress and regulating the intestinal flora. In addition, we present representative clinical trials demonstrating the beneficial effects of SGLT2i in clinical use. Finally, we briefly discuss the current challenges and possible future strategies for the utilization of SGLT2i.

Our aim was to arouse more attention from readers through this review to further explore the potential mechanisms and targets of SGLT2i in the treatment of diabetic cardiomyopathy in the clinic and lay a solid foundation for the reasonable use of SGLT2i; SGLT2i can save more patients with diabetic cardiomyopathy. In addition, these findings could help in the development of new drugs for diabetic cardiomyopathy based on relevant signals or therapeutic targets.