Ageing‐associated increase in SGLT2 disrupts mitochondrial/sarcoplasmic reticulum Ca2+ homeostasis and promotes cardiac dysfunction

Abstract The prevalence of death from cardiovascular disease is significantly higher in elderly populations; the underlying factors that contribute to the age‐associated decline in cardiac performance are poorly understood. Herein, we identify the involvement of sodium/glucose co‐transporter gene (SGLT2) in disrupted cellular Ca2+‐homeostasis, and mitochondrial dysfunction in age‐associated cardiac dysfunction. In contrast to younger rats (6‐month of age), older rats (24‐month of age) exhibited severe cardiac ultrastructural defects, including deformed, fragmented mitochondria with high electron densities. Cardiomyocytes isolated from aged rats demonstrated increased reactive oxygen species (ROS), loss of mitochondrial membrane potential and altered mitochondrial dynamics, compared with younger controls. Moreover, mitochondrial defects were accompanied by mitochondrial and cytosolic Ca2+ ([Ca2+]i) overload, indicative of disrupted cellular Ca2+‐homeostasis. Interestingly, increased [Ca2+]i coincided with decreased phosphorylation of phospholamban (PLB) and contractility. Aged‐cardiomyocytes also displayed high Na+/Ca2+‐exchanger (NCX) activity and blood glucose levels compared with young‐controls. Interestingly, the protein level of SGLT2 was dramatically increased in the aged cardiomyocytes. Moreover, SGLT2 inhibition was sufficient to restore age‐associated defects in [Ca2+]i‐homeostasis, PLB phosphorylation, NCX activity and mitochondrial Ca2+‐loading. Hence, the present data suggest that deregulated SGLT2 during ageing disrupts mitochondrial function and cardiac contractility through a mechanism that impinges upon [Ca2+]i‐homeostasis. Our studies support the notion that interventions that modulate SGLT2‐activity can provide benefits in maintaining [Ca2+]i and cardiac function with advanced age.


| INTRODUC TI ON
Over the past several decades, the average lifespan of humans has increased worldwide. Notably, the prevalence of death from cardiovascular disease is significantly higher in elderly populations. 1 The epidemiological and demographic results show the development of new therapies and interventions to mitigate the ageing-related decline in cardiac performance. Growing evidence suggests that ageing triggers cellular defects at the molecular and genetic levels which can result in organ dysfunction. 2 In this regard, both sarcoplasmic reticulum (SR) and mitochondrial dysfunctions have been linked to ageing-associated disorders including cardiac dysfunction and contractile failure. 3 Mitochondrial function is essential for oxidative metabolism and ATP synthesis for maintaining tissue homeostasis and cell survival.
Also, the mitochondrion serves as a platform for calcium regulation and programmed cell death by apoptosis and necrosis. 4,5 Hence, the loss of normal mitochondrial function commonly associated with advanced age has been linked to abnormalities in sarcolemmal Ca 2+ transport and alteration in action potential duration. 3,6,7 However, the relationship between mitochondrial calcium handling defects and contractile dysfunction associated with ageing remains poorly understood.
Notably, we recently demonstrated that sodium/glucose co-transporter 2 (SGLT2), which is a member of the solute carrier family (SLC5A2) and is important in facilitating sodium-dependent glucose-transport and in the regulation of blood glucose levels, was significantly activated in the hearts of insulin-resistant animals with metabolic syndrome (MetS). Furthermore, we demonstrated that pharmacological inhibition of SGLT2 improved glucose tolerance and cardiac function. 8 SGLT2 inhibitors have also been shown to provide clinical benefits in normalizing glucose levels and cardiac function in diabetic and non-diabetics patients with insulin resistance. 9,10 While the underlying mechanisms by which SGLT2 inhibition confers cardioprotection are not well understood, several theories have been proposed including improved haemodynamics, more efficient metabolic function and cardiac/renal effects among others.
In particular, a recent study by Pasternak et al 10 demonstrated in the Scandinavian Cohort study that SGLT2 inhibitors significantly reduced the risk of heart failure and all-cause mortality, particularly in elderly patients. This finding raised awareness that SGLT2 inhibitors may play an important role in improving cardiac performance in the aged population. This view was substantiated by another recent study by Cintra et al, 11 who also showed beneficial effects of SGLT2 inhibition in elderly patients with heart diseases. At present, there are several theories that the underlying cardioprotection conferred by SGLT2 inhibition may converge on common signalling pathways that influence cardio-metabolic and bioenergetics involving Na + /H + exchange-1, (NHE-1). [12][13][14] Indeed, deregulated Na + and Ca 2+ levels are early drivers of mitochondrial dysfunction, ROS production and cardiovascular morbidity and mortality. 15 It has been shown that the increased intracellular Na + is an underlying cause of cardiac injury and metabolic dysfunction in diabetic cardiomyopathy, presumably from elevated SGLT2 activity. 14,16 In this regard, the studies have demonstrated that SGLT2 inhibition suppressed elevated NHE-1 activity and intracellular Na + levels in cardiomyocytes of diabetic animals. 16,17 Despite our previous published work and clinical studies linking SGLT2 in the pathogenesis of cardiac dysfunction and heart failure, the underlying mechanism by which SGLT2 inhibition improves cardiac function in the aged myocardium remains poorly understood. Therefore, in the present study, we explored the mechanism by which SGLT2 disrupts cardiac function associated with advanced age. Given that the mitochondria are a major nodal point in the regulation of cardiac metabolism and intracellular Ca 2+ ([Ca 2+ ] i ), we explored the impact of SGLT2 on mitochondrial Ca 2+ homeostasis in young and aged cardiac myocytes. Our data provide new important information that SGLT2 disrupts mitochondrial/SR Ca 2+homeostasis and thereby promotes cardiac dysfunction by impairing mitochondrial energy metabolism.

| Experimental animals
Our experimental protocols and all animal procedures are in coincidence with the guidelines of the Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes and are approved by Ankara University (reference # 2016-18-165).
Wistar male rats were grouped as 24-month-old (Aged group) vs. 6-month-old (Young group) and parameters such as blood glucose levels were analysed as described previously. 8 Hearts were rapidly excised following anaesthesia with sodium pentobarbital (30 mg/kg bodyweight).

| Cardiomyocytes isolation
Cardiomyocytes isolation was performed from the left ventricles of hearts, as described previously. 8 Briefly, hearts were cannulated in a Langendorff-perfusion system, followed by digestion with 1-mg/ mL collagenase (Type IV, Worthington, USA) for 30-35 minutes.
Cardiomyocytes were treated with either 100 nmol/L SGLT2 inhibitor (SGLT2i, dapagliflozin, D185360; for 4-5 hours at 37°C) or 0.1 µmol/L MitoTEMPO for further experiments. To validate the mitochondrial membrane integrity under control during measurements, we directly measured the mitochondria Ca 2+ level but not in the mitochondrial intermembrane space, all experiments were performed in cells under same experimental conditions (using Fluo-4AM and TMRM loaded permeabilized cells, by flow cytometry) through the determination of mitochondrial Ca 2+ level and mitochondrial membrane potential (MMP), simultaneously by using confocal microscopy (see Appendix S1 and Figure S1).

| Measurements of the resting levels of intracellular ions and global transient intracellular Ca 2+ -changes
The basal levels of intracellular free Ca 2+ , Na + and H + (or intracellular pH) levels ([Ca 2+ ] i , [Na + ] i , [H + ] i ) were measured in resting cardiomyocytes loaded with ion-specific fluorescence dyes (4-µmol/L Fura-2AM for Ca 2+ , 10-µmol/L SBFI for Na + and 5-µmol/L SNARF-1 for H + ). Fluorescence values were recorded by the use of a ratiometric micro-spectrofluorometer (PTI Ratio master and FELIX software; Photon Technology International, Inc, NJ USA) for Ca 2+ or a laser scanning microscope (confocal microscopy, Leica TCS SP5, Germany) for Na + , as described previously. 8 To determine the intracellular transient Ca 2+ -changes (Ca 2+transients) in cardiomyocytes (loaded with Fura-2AM, 4-µmol/L), an electric-field stimulation (with a 10-ms duration electrical pulses at a frequency of 0.2 Hz) was applied. We took into consideration the potential influences of a temperature-dependent effect on lowered amplitude of intracellular Ca 2+ -changes without changes in the time course of measurements, 21,22 as well as the potential effects of temperature on Ca 2+ -sensitive fluorescent probes, 23 we used room temperature for both cell-loading and recordings from loaded cells. Therefore, fluorescence intensity changes were measured at room temperature by using the micro-spectrofluorometer (PTI, Lawrenceville, NJ, USA) and FELIX software, as described, previously. 8
Cardiomyocytes were sampled and digitized at 5 kHz using an analog-to-digital converter and a software (Digidata 1200A and pCLAMP 10.0; Axon Instruments, USA). The amplifier has been used for either voltage-clamp mode (voltage-dependent K + -channel currents, I K ) or current-clamp mode (action potentials) to measure either ionic fluxes or membrane voltage changes, respectively, as described previously. 8 Action potential duration from the repolarization phase at 25, 50, 75, 90% (APD 25 , APD 50 , APD 75 , APD 90 ), the resting membrane potential, and the maximum amplitude of action potentials were calculated from original records (at least 15-20 records/cell).

| Measurement of cellular ATP level
The cellular level of ATP in isolated left ventricular cardiomyocytes was measured using a colorimetric ATP assay kit (Abcam, ab83355) with some modification, as described previously. 8

| Histological analysis of cardiomyocytes
Electron microscopy was performed in cardiomyocytes, as described previously. 20 Ultra-thin sections were stained with uranyl acetate and lead citrate and cells were observed using an LEO 906 E TEM (80 kV, Oberkochen, Germany). A Sharpeye CCD and Image SP (Germany) digital imaging system were used to photograph.
To quantify the results, the mitochondrial aspect ratio as the ratio of length to width was calculated by using Image SP, the randomly selected five longitudinally arranged micrographs, as described previously. 26 For each cell, the lengths of 500 to 600 interfibrillar mitochondria were assessed, as described previously. 27

| Western-blot analysis
The protein samples were prepared from the cells and tissues as previously described. 20 Equal amount of cardiac cell lysates were run on SDS-polyacrylamide gels and incubated with an- Immunoreactive bands were detected by a chemiluminescent reaction (ECL kit, Amersham Pharmacia, USA). The densities of the bands are analysed using ImageJ software, and the results were presented as fold changes.

| QRT-PCR analysis
The mRNA levels were measured as described previously 1 . The fold changes in the genes were analysed based on the comparative (2 −ΔΔCt ) method. Primer sequence for SGLT2 is TCATTGCCGCGTATTTCCTG (forward) and AACACCACAAAGAGCGCATT (reverse). Histogram presents fold change in ATP levels (young vs aged group), measured by colorimetric assay. Data are presented as Mean ± SEM, statistical significance * P < 0.05 vs. young group interquartile range, we used a box and whisker graph. Comparisons between quantitative variables were assessed by the unpaired twosided Student's t test at the P < 0.05 significance level.

| Altered myocardial ultrastructure and cardiac function in aged hearts
To gain insight into the effects of ageing on cardiac dysfunction, we first determined whether the aged hearts (24 months) are hypertrophic and different from younger animals (6 months). As shown in

| Age-associated increase in reactive oxygen species (ROS) underlies mitochondrial defects
As we observed a decline in mitochondrial ATP in aged animals, suggesting that mitochondrial perturbations may be associated with contractile dysfunction. We used mitochondria targeting antioxidant (MitoTEMPO) that we have previously shown to reduce ROS level in the ventricular cardiomyocytes, 20 we assessed the effects of ageing on mitochondrial superoxide production in MitoSOX-loaded cardiomyocytes followed by flow cytometry analysis. As shown in Figure 2 (Figure 3 panel B). Besides, the time to peak amplitude and the half time for recovery of Ca 2+ transients in the aged group was significantly longer than those of the young group (Figure 3 panel C).

| Effects of ageing on intracellular Ca 2+ signalling in ventricular cardiomyocytes
To test whether the [Ca 2+ ] content of the sarcoplasmic reticulum (SR) could be altered with ageing, we performed further experiments using caffeine to assess SR Ca 2+ extrusion. The responses to acute caffeine (10-mmol/L) were significantly lower in the aged group compared with the corresponding young group (Figure 3 panel D). To characterize the underlying mechanisms for altered Ca 2+ transients, we monitored the protein expression of total and phosphorylated phospholamban (pPLB), which modulates the SR Ca 2+ -ATPase (SERCA) function. As shown in Figure 3 panels E,F, a marked decrease in phosphorylation of PLB (pPLB) was observed in aged cardiomyocytes.

| Ageing up-regulates SGLT2 and disrupts the intracellular ionic balance of cardiomyocytes
As action potential durations were prolonged in aged cardiomyocytes, we assessed whether intracellular ionic levels were altered in aged cardiomyocytes. To test this possibility, we assessed changes in the Na + -influx (via voltage-dependent Na + channel current, I Na ) and Ca 2+ -influx (via voltage-dependent L-type Ca 2+channel currents, I CaL ) on the AP characteristics. However, there was no significant change in any parameter of the channel currents between young and aged groups ( Figure S1 and Figure S2, panels A,B).
Next, we tested the resting levels of cytosolic Ca 2+ ([Ca 2+ ] i ), measured by fluorescence microscopy by Fura-2 AM. As shown in Figure 4 panel A, cytosolic Ca 2 was found to be significantly Actin serves as a housekeeping gene. All data are presented, as Mean ± SEM. Cardiomyocytes were isolated from 5-6 rats/groups. Statistical significance P < 0.05 vs. young group, # P < 0.05 vs. aged group, statistical significance analysed by unpaired two-sided Student's t test higher in aged cardiomyocytes compared with the younger cardiomyocytes. Moreover, we observed an increase in basal levels of cytosolic free Na + ([Na + ] i ) and H + ([H + ] i ), measured as SNARF-1 intensity changes, in the aged cardiomyocytes, Figure 4 panels We previously have shown in a rat model of insulin-resistance (MetS) that ionic balance was disrupted in an SGLT2 dependent manner. 8 Therefore, we assessed whether age-related abnormalities in ionic balance were related to the deregulated expression of SGLT2. We assessed SGLT2 mRNA and protein expression in young and aged cardiomyocytes. As shown in Figure 4 panels D,E, the mRNA and protein levels of SGLT2 were significantly higher in aged cardiomyocytes.
To validate the possible correlation with SGLT2 activation and insulin resistance in left ventricular aged cardiomyocytes, we performed an oral glucose tolerance test (OGTT) in aged and young rats. As shown in Figure 4 panel F, the peak blood glucose levels measured at 30, 60 and 120 minutes were significantly higher in the aged rats compared with corresponding young rats. The HOMO-IR index was also high in the aged rats compared with the young rats (data not shown).

| Inhibition of SGLT2 suppress NaK/NCX activity and restores [Ca 2+ ] i -homeostasis in aged cardiomyocytes
As significantly higher [Ca 2+ ] i and [Na + ] i were observed in aged cardiomyocytes with increased SGLT2 activity, we examined whether SGLT2 influences Na + /K + -pump (NaK) or NCX activity. Interestingly, we observed a twofold increase in NaK current (I NaK ) in the aged cardiomyocytes compared with corresponding young hearts, Figure 6 panel A. Furthermore, we observed a similar increase in the NCXcurrents in both forward and reversed directions (I NCXin and I NCXou , respectively) in the aged cardiac myocytes, Figure 6, panel B.
Notably, the SGLT2 inhibitor dapagliflozin (100 nmol/L) restored I NaK and I NCX currents, respectively, in the aged cardiomyocytes highlighting the importance of SGLT2 in the regulation of intracellular ion-exchange currents, Figure 6 panels A,B.
Given these findings, we examined the effects of inhibition of SGLT2 on action potential (AP) prolongation in aged cardiac myocytes. As shown in Figure 6 panel C, SGLT2 inhibition significantly improved AP repolarization in the aged cardiomyocytes.

| D ISCUSS I ON
Ageing is a major factor that predisposes cardiac myocytes to metabolic and contractile abnormalities that leads to heart failure. Lipid peroxidation, oxidative stress and mitochondrial injury are central features of insulin resistance and obesity associated with advanced age. [31][32][33] Notably, mitochondrial perturbations, coupled with abnormalities in SR Ca 2+ dysfunction have been identified as important contributors to the metabolic derangements with the ageing process in muscle cells. 34 Inhibition of SGLT2 rescues phosphorylation of PLB and reduces mitochondrial Ca 2+ load in the aged cardiomyocytes. Panel A, Western blot analysis of the lysate derived from young, aged and aged + SGLT inhibited (SGLT2i) cardiomyocytes for protein expression of sarcoplasmic reticulum Ca 2+ -ATPase (SERCA2a), sodium/calcium-exchanger (NCX), GAPDH; total phospholamban (PLB) and phosphorylated (pPLB) and Actin. Panels B-C, Histogram presents fold change in SERCA2, NCX and pPLB levels of the younger group. Panel D, Histogram depicts the level of mitochondrial Ca 2+ ([Ca 2+ ] Mit ) in Fluo4-AM loaded (4-µmol/L for 45-min) young, aged and aged + SGLT inhibited (SGLT2i) cardiomyocytes. All data are presented as Mean ± SEM, cardiomyocytes were isolated from 5-6 rats/groups. Statistical significance *P < 0.05 vs. young group, # P < 0.05 vs. aged group, analysed by the unpaired two-sided Student's t test cardiac myocytes. Our present data are in line with recent studies highlighting a role for SGLT2 inhibitors, on cardiometabolic, bioenergetics as well as the cardiac structure and NHE-1 activity. [12][13][14] Previous, studies have shown that an increase in myocardial Na + level and Ca 2+ level are a potential early drivers of death from cardiovascular diseases including heart failure, via induction of increased oxidative stress. 15 Furthermore, it has also been shown that increased intracellular Na + can trigger mitochondrial ROS production in the diabetic heart through activation of SGLT2. 14,16 Moreover, our present data demonstrating elevated levels of Na + , Ca 2+ and H + in aged cardiomyocytes is concordant with previous studies which demonstrated normalization of the activated NHE-1 and increased intracellular Na + in cardiomyocytes under pathological conditions. 12,13,16,17 Our data highlight that de-regulated SGLT2 activity in cardiomyocytes promotes defective Ca 2+ -handling that impairs contractile function with advanced age. This view is supported by recent clinical studies, demonstrating the use of metformin along with SGLT2 inhibition, in elderly patients. 41,42 Further, several clinical trials have demonstrated the beneficial effects of SGLT2 inhibitors on reducing cardiac morbidity risk of heart failure in the aged population. [10][11][12]43 Hence, our data suggest that deregulated SGLT2 during ageing disrupts mitochondrial function and cardiac contractility. Those effects seem to be associated with a mechanism that impinges upon deregulated [Ca 2+ ] i homeostasis. Our studies support the notion that interventions that modulate SGLT2 activity may provide benefits in maintaining [Ca 2+ ] i levels and cardiac function with advanced age.
More importantly, our data demonstrate for the first time that ageing increases SGLT2 expression which negatively impacts mitochondrial/SR Ca 2+ -homeostasis and cardiac function. In aggregate, our data provide novel evidence that SGLT2 is an important regulator of cardiac dysfunction with advanced age. Our studies further supported the notion that interventions that modulate SGLT2 activity may provide benefits in maintaining [Ca 2+ ] i levels and cardiac function with advanced age. Hence, targeted therapy for SGL2 inhibition may prove beneficial in protection against age-associated contractile dysfunction.

F I G U R E 6
Inhibition of SGLT2 suppress deregulated electrical activities of NaK and NCX and restore [Ca 2+ ] i -homeostasis in aged left ventricular cardiomyocytes. Panel A, Na + /K + -pump currents (I NaK ) traces (left panel) in young, aged and aged + SGLT2 inhibited (SGLT2i) cardiomyocytes (treated with dapagliflozin; 100-nmol/L D185360; for 4-5 h), Right panel: Histogram presents Mean ± SEM values for I NaK current. Panel B, Left panel, shows Na + /Ca 2+ -exchanges currents (I NCX ) (inward; I NCXin and outward; I NCXou ) and the maximum amplitudes of the currents, Right panel: Histogram presenting Mean ± SEM values for NCX activity. Panel C, The effect of SGLT2 inhibition (SGLT2i) on prolonged AP repolarization phases in aged cardiomyocytes (measured at AP 25 , AP 50 , AP 75 and AP 90 ). The histogram presents Mean ± SEM values, statistical significance *P < 0.05 vs. aged group. Panel D, SGLT2 inhibition provided important protection in the depressed Ca 2+release and re-uptake (amplitude and time course of [Ca 2+ ] i changes such as TP: the time to the maximum amplitude of [Ca 2+ ] i and RT 50 : the half time re-uptake of [Ca 2+ ] i ; left and right panels respectively) under electrical-field stimulation in the aged cardiomyocytes. The bar graphs present a per cent (%) change from the aged group. The statistical significance level was * P < 0.05 vs. aged group, analysed by unpaired twosided Student's t test

ACK N OWLED G EM ENTS
The authors would like to thank Dr CV Bitirim (Stem cell Institute, Ankara University) for her help in flow cytometry.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon reasonable request.