OUP user menu

NADPH oxidase activation by hyperglycaemia in cardiomyocytes is independent of glucose metabolism but requires SGLT1

Magali Balteau, Nicolas Tajeddine, Carole de Meester, Audrey Ginion, Christine Des Rosiers, Nathan R. Brady, Caroline Sommereyns, Sandrine Horman, Jean-Louis Vanoverschelde, Philippe Gailly, Louis Hue, Luc Bertrand, Christophe Beauloye
DOI: http://dx.doi.org/10.1093/cvr/cvr230 237-246 First published online: 22 August 2011


Aims Exposure to high glucose (HG) stimulates reactive oxygen species (ROS) production by NADPH oxidase in cardiomyocytes, but the underlying mechanism remains elusive. In this study, we have dissected the link between glucose transport and metabolism and NADPH oxidase activation under hyperglycaemic conditions.

Methods and results Primary cultures of adult rat cardiomyocytes were exposed to HG concentration (HG, 21 mM) and compared with the normal glucose level (LG, 5 mM). HG exposure activated Rac1GTP and induced p47phox translocation to the plasma membrane, resulting in NADPH oxidase (NOX2) activation, increased ROS production, insulin resistance, and eventually cell death. Comparison of the level of O-linked N-acetylglucosamine (O-GlcNAc) residues in LG- and HG-treated cells did not reveal any significant difference. Inhibition of the pentose phosphate pathway (PPP) by 6-aminonicotinamide counteracted ROS production in response to HG but did not prevent Rac-1 upregulation and p47phox translocation leading to NOX2 activation. Modulation of glucose uptake barely affected oxidative stress and toxicity induced by HG. More interestingly, non-metabolizable glucose analogues (i.e. 3-O-methyl-d-glucopyranoside and α-methyl-d-glucopyranoside) reproduced the toxic effect of HG. Inhibition of the sodium/glucose cotransporter SGLT1 by phlorizin counteracted HG-induced NOX2 activation and ROS production.

Conclusion Increased glucose metabolism by itself does not trigger NADPH oxidase activation, although PPP is required to provide NOX2 with NADPH and to produce ROS. NOX2 activation results from glucose transport through SGLT1, suggesting that an extracellular metabolic signal transduces into an intracellular ionic signal.

  • Glucotoxicity
  • Glucose metabolism
  • Oxidative stress
  • NADPH oxidase
  • SGLT

1. Introduction

The impact of high-glucose (HG) concentration on cardiac metabolism and function is controversial. Exposure to HG concentrations reduces ischaemic injury in isolated perfused hearts submitted to ischaemia and reperfusion.1 It stimulates glycolysis, delays ischaemic contracture, and improves recovery of left ventricular contractile function during reperfusion. As insulin also exerts a protective effect during reperfusion, glucose–insulin–potassium (GIK) solutions have been advocated to reduce ischaemic injuries and to metabolically support the ischaemic heart.2 The benefit of GIK solutions was reported in several studies conducted in patients with an ST-segment elevation myocardial infarction, indicating that recovery of the left ventricle function improved following GIK treatment.3 However, a more recent and large randomized trial questioned this conclusion and failed to demonstrate any beneficial effects of GIK on mortality.4 This failure could be related to the development of hyperglycaemia in patients receiving GIK, which could counteract the protective effect conferred by GIK.5 Several lines of evidence indicate that hyperglycaemia might result in toxic effects in several cell types including cardiomyocytes. Indeed, hyperglycaemia modifies myofibrillar structure and intercellular connections6 and chronic exposure of cardiomyocytes to HG concentrations induces cell death by an oxidative stress-dependent mechanism.79

A number of mechanisms have been proposed to mediate hyperglycaemia-induced toxicity. They include protein kinase C (PKC) activation, increased flux through the hexosamine pathway, increased advanced glycation end-product formation, and increased production of reactive oxygen species (ROS).10 One source of ROS is the one-electron reduction in O2 to superoxide anion by NADPH oxidase. This enzyme was first described in macrophages and has later been identified in several other cell types, including cardiomyocytes. NADPH oxidases are multimeric enzymes, which contain a heterodimeric membrane-bound cytochrome b558 made of either NOX2 or NOX4 and p22phox subunits. NOX2 activation requires its association with four cytosolic subunits: p47phox, p67phox, p40phox, and Rac1, whereas the NOX4 complex is constitutively active and regulated by gene expression.11 The involvement of NADPH oxidase in hyperglycaemia-induced ROS production is well documented. Indeed, incubation of endothelial cells with HG concentrations induces expression of several NADPH oxidase subunits, translocation of p47phox and Rac1, and subsequent NADPH oxidase activation.12 Moreover, HG-dependent NADPH oxidase activation and ROS production lead to an increase in apoptosis.13 Superoxide generation by NADPH oxidase also depends on the availability of reducing equivalents. NADPH generated by the oxidative part of the pentose phosphate pathway (PPP) has been suggested to fuel NADPH oxidase and sustain ROS production in the heart.14,15 6-Aminonicotinamide (6-AN), a nicotinic analogue that inhibits PPP dehydrogenases (6-phosphogluconate dehydrogenase and to a lesser extent glucose-6-phosphate dehydrogenase),16 protects against oxidative stress during ischaemia–reperfusion in cardiomyocytes17 and in the heart of obese Zucker fa/fa rats.18

Although ROS production by NADPH oxidase plays a critical role in glucotoxicity in the heart, the mechanism linking HG concentrations to NADPH oxidase activation remains to be elucidated. Therefore, we investigated the connection between glucose transport, metabolism, NADPH oxidase activation, and subsequent ROS production under hyperglycaemia. In this paper, we describe for the first time that NADPH oxidase activation and ROS production in response to HG concentrations do not require glucose metabolism but rather its transport through a novel cardiac glucose transport system, the sodium-dependent glucose cotransporter, SGLT1.19

2. Methods

2.1 Animals and materials

Animal handlings were approved by local authorities (comité d'éthique facultaire pour l'expérimentation animale, UCL/MD/2007/049) and were performed in agreement with the guidelines on animal experimentation at our institution. Moreover, this study conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). All biochemicals were from Sigma, unless otherwise stated. The gp91ds-tat peptide and the corresponding scrambled peptide were kindly provided by V. Stroobant (Ludwig Institute, Brussels, Belgium).

2.2 Isolation and culture of adult rat cardiomyocytes

Adult male Wistar fed rats were anaesthetized with pentobarbital 50 mg/kg ip and sacrificed by cervical dislocation before removing the heart. Anaesthesia depth was monitored by limb withdrawal using toe pinching. The cardiomyocytes were isolated after heart perfusion with collagenase type II (1 mg/mL, Worthington), then purified and cultured as described previously.20 Briefly, cells were plated on laminin-coated dishes and incubated for 24 h in fresh medium 199 (Invitrogen, 5.5 mM d-glucose) containing 100 mg/L glutamine, 2 mM carnitine, 5 mM creatine, 5 mM taurine, 10−10 M T3, and 0.2% (wt/vol) BSA (fatty acid free) in the presence of 100 U/mL penicillin and 100 µg/mL streptomycin. Afterwards, primary cultures of cardiomyocytes were incubated with the indicated concentrations of d-glucose for the indicated periods of time as detailed in the figure legends. Other additions are as indicated in the figure legends.

2.3 Measurement of cell death and of ROS

Dead cells were labelled by incubation with 50 µM propidium iodide. The percentage of dead cells was calculated from the number of dead (labelled) and living (unlabelled) cells observed under the microscope (1500 cells per condition and in triplicate for each experiment). Intracellular ROS production was measured by evaluating oxidation of the cell permeable fluorescent probe 2′,7′-dihydrodichlorofluorescein diacetate (H2DCF-DA, Invitrogen), which becomes fluorescent when oxidized by ROS. The fluorescence was detected by fluorescence microscopy (Zeiss, AxioVert200M, 485–530 nm). Cardiomyocytes were cultured in a Krebs–Henseleit medium containing 1.2 mM MgSO4, 1.2 mM KH2PO4, 4.7 mM KCl, 120 mM NaCl, 2.5 mM CaCl2, and 25 mM NaHCO3 in 5% CO2 at pH 7.4. Cells were first incubated with d-glucose, angiotensin II, and/or several inhibitors, and then loaded with 10 µM H2DCF-DA for 30 min at 37°C. The medium containing the probe was then removed and a fresh dye-free medium was added. ROS production was quantified, using AxioVision software. The threshold above which cells were considered to be positive, i.e. having a significant increase in ROS production, was fixed to the average fluorescence of the control cells plus two standard deviations. The results are expressed as the percentage of fluorescent cells.

2.4 Immunocytofluorescence analysis of p47phox membrane translocation

After treatment, cardiomyocytes were fixed with paraformaldehyde (4% v/v) in phosphate-buffered saline (PBS) for 30 min on ice and then permeabilized with Triton X-100 (0.2% v/v in PBS) and treated with PBS containing 5% (v/v) milk for 1 h at room temperature. Cells were incubated for 1 h with rabbit-anti-p47phox primary antibody (Santa Cruz) diluted 1/50 in PBS containing 1% (v/v) milk. After washing in PBS/milk 1% (v/v)/Triton (v/v) 0.2%, cells were incubated with 1/1000 Alexa-fluo-conjugated anti-rabbit secondary antibody (Alexa fluo 488, Molecular Probes) for 1 h at room temperature. After washing, a mounting medium for fluorescence with DAPI (Vectashield) was added. Cells were observed by fluorescence microscopy (Olympus IX71, 485–530 nm).

p47phox translocation to the plasma membrane was evaluated by measuring the disappearance of p47phox perinuclear staining.

2.5 Quantitative PCR

Total RNA was isolated from cardiomyocytes using a chloroform/isopropanol procedure (Tripur, Roche). Reverse transcription was performed for 1 h at 37°C with 1 µg RNA in the presence of primer poly(dT) (Roche) and Moloney-Murine Leukemia Virus Reverse Transcriptase (Invitrogen). Quantitative PCR was performed on IQ5 (Bio-Rad) using the QPCR Core kit for Sybergreen (Eurogentec). The mRNA level for each gene in each sample was normalized to β-actin mRNA. The nucleotide sequences of primers used were: NOX2: 5′-CCA-GTG-AAG-ATG-TGT-TCA-GCT-3′ (sense) and 5′-GCA-CAG-CCA-GTA-GAA-GTA-GAT-3′ (antisense); NOX4: 5′-GG A-TCA-CAG-AAG-GTC-CCT-AGC-AG-3′ (sense) and 5′-GCA-GC T-ACA-TGC-ACA-CCT-GAG-AA-3′ (antisense); p22phox: 5′-TGG-CC T-GAT-CCT-CAT-CAC-AG-3′ (sense) and 5′-AGG-CAC-GG A-CAG-CAG-TAA-GT-3′ (antisense); p47phox: 5′-TCA-CCG-AGA-TC T-ACG-AGT-TC-3′ (sense) and 5′-TCC-CAT-GAG-GCT-GTT-GA A-GT-3′ (antisense); and β-Actin: 5′-CGT-GCG-CT G-GTC-GTC-GA C-AAC-G-3′ (sense) and 5′-ATC-GTA-CTC-CTG-CT T-GCT-GAT-CC A-C-3′ (antisense).

2.6 Metabolic measurements

Glucose uptake was measured by the rate of detritiation of [2-3H]glucose, which occurs after glucose phosphorylation during the rapid isomerization of hexose-6-phosphates catalysed by phosphoglucose isomerize.21 Cardiomyocytes were treated as described in the figure legends. [2-3H]glucose (0.2 µCi/mL, Amersham) was added 30 min before the end of treatment. After stopping the reaction with HClO4 (5% v/v) and neutralization of the extracts, samples were taken to separate tritiated water from tritiated glucose by column chromatography. Radioactivity was measured by liquid scintillation counting (Liquid Scintillation Analyser Tri-Carb 2900, Perkin Elmer). Glucose-6-phosphate (G6P) levels were measured fluorimetrically in neutralized perchloric acid extracts by NADPH produced by glucose-6-phosphate dehydrogenase, which oxidizes G6P into 6-phosphogluconate.22 Glycogen was measured as glucose equivalent after enzymatic hydrolysis (amyloglucosidase) and enzymatic transformation into G6P and 6-phosphogluconate.23

2.7 Immunoblot analysis and protein kinase B activity measurement

Cells were lysed in cold buffer containing 50 mM Hepes (pH 7.5), 50 mM KF, 1 mM KPi, 5 mM EDTA, 5 mM EGTA, 15 mM β-mercaptoethanol, 1 µg/mL leupeptin, 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 1 mM vanadate, and 0.2% v/v Triton X-100. Protein content was measured using the method of Bradford with bovine serum albumin as a standard. Immunoblots were performed with total extracts separated on 10% SDS–PAGE and transferred to polyvinylidene difluoride membranes. The membranes were blocked with milk or bovine serum albumin (5% v/v) and then blotted with antibodies against SGLT1 (Santa Cruz) or O-linked N-acetylglucosamine (O-GlcNAc) residues (Santa Cruz). After incubation with the appropriate secondary antibody, proteins were visualized using electrochemical luminescence (Pierce) and quantified using Image J. Protein kinase B (PKB) activity was measured in immunoprecipitates as described previously.24

2.8 Rac1-GTP pull down

Cardiomyocytes were lysed with a buffer containing 50 mM Tris–HCl (pH 8), 150 mM NaCl, 10 mM MgCl2, 1 mM EDTA, and 5% v/v Triton X-100. Rac-GTP was immunoprecipitated with protein-A-Sepharose (GE Healthcare) coupled to an anti-active Rac-GTP mouse monoclonal antibody (NewEast Biosciences). Extracts were separated on 15% SDS–PAGE, transferred to a polyvinylidene difluoride membrane, and incubated with an anti-Rac total antibody (Cell Signaling) and an appropriate secondary antibody. Proteins were visualized using electrochemical luminescence (Pierce) and quantified using Image J.

2.9 Statistical analysis

All the values are expressed as means ± SEM. Comparisons were performed using ANOVA followed by a Bonferroni post hoc test. A value of P< 0.05 was considered as statistically significant.

3. Results

3.1 Chronic exposure to HG induces cell death by increasing ROS production

Primary cultures of adult rat cardiomyocytes were incubated with 5–30 mM d-glucose for up to 24 h. Cell mortality, measured by propidium iodide staining, significantly increased when the cells were exposed to 21 mM or higher concentrations of d-glucose for at least 18 h (Figure 1A and B). Subsequently, cells were incubated for 24 h with either low (5 mM, LG) or high (21 mM, HG) d-glucose. Mortality was significantly higher in HG than in LG conditions (HG, 18.3 ± 0.3% vs. LG, 10.4 ± 0.5% of dead cells, P< 0.05). Addition of l-glucose, sucrose, or mannitol (up to 21 mM final) to the culture medium containing 5 mM glucose did not mimic the effect of HG (data not shown), indicating that glucotoxicity was not related to osmotic stress. The antioxidant agent, N-acetyl-l-cysteine or resveratrol, protected the cells against glucotoxicity (data not shown). In line with these observations, ROS production measured with the fluorescent dye H2DCF-DA was significantly increased in the HG group (HG, 28.9 ± 4.2% vs. 5.9 ± 3.6% of positive cells, P< 0.05). Ten micromolar allopurinol and 100 μM NG-nitro-l-arginine methyl ester, inhibitors of xanthine oxidase and NO synthase, respectively, failed to protect cardiomyocytes against cell death (data not shown), whereas the NADPH oxidase inhibitors, diphenylene iodonium (DPI) and apocynin, abrogated the HG-induced increase in cell mortality (Figure 1C and D). From these results, we conclude that NADPH oxidase activation and the resulting increased ROS production were responsible for the HG-induced cell death.

Figure 1

Hyperglycaemia-induced cell death. (A) Dose–response curve of glucose exposure on cell death. Cardiomyocytes were incubated for 24 h in the presence of the indicated glucose concentrations. (B) Time-dependent increase in cell death in response to HG (21 mM). Effects of (C) 100 nM DPI and (D) 10 μM apocynin (Apo) on HG-mediated cell. The data are means ± SEM, n= 3. *Values statistically different from the LG control. $Values statistically different from the corresponding HG sample, without inhibitors.

3.2 HG conditions activate NOX2 by Rac1 activation and p47phox translocation

NADPH oxidase activation can result from either an increase in expression of one or several subunits (NOX2 or NOX4) or by NOX2 activation resulting from Rac1 activation and p47phox translocation to the plasma membrane.25 HG treatment did not affect mRNA expression of NOX2, NOX4, p22phox, and p47phox (data not shown). In contrast, p47phox content of the perinuclear region was decreased in the HG, suggesting that p47 was translocated to the plasma membrane (Figure 2A). This disappearance corresponded to an increase in p47phox content in the plasma membrane fraction after cell fractionation (see Supplemental material online, Figure S1) and in the co-localization of p47phox with caveolin-3. p47phox-caveolin-3 co-localization doubled after HG exposure (data not shown). In addition, Rac1 was activated in the HG group (Figure 2B). The involvement of NOX2 in glucotoxicity was further confirmed by the use of the gp91ds-tat peptide, a specific inhibitor of NOX2 activation,26 which blocked the HG-induced ROS production and prevented glucotoxicity (Figure 2C and D).

Figure 2

Cell death and insulin resistance induced by NOX2 activation, under HG concentrations. (A) p47phox translocation and (B) Rac1-GTP were determined after 24 h of HG incubation and were compared with angiotensin II (AngII, 10−6 M) stimulation. Arrows indicate cells with positive perinuclear p47phox staining. (C and D) Effects of gp91ds-tat (2.5 µM) and the corresponding scrambled peptide (2.5 µM) on ROS production (C) and cell death (D) in cardiomyocytes after 6 h (ROS production) or 24 h (cell viability) incubation with HG. The data are means ± SEM, at least n= 3. *Values statistically different from the LG control. $Values statistically different from the corresponding HG sample incubated with the scrambled peptide. (E) PKB activity in cardiomyocytes incubated for 24 h with HG and then treated with 3.10−9 M insulin for 30 min. NADPH oxidase activation by HG was blocked by gp91ds-tat. The insulin response was compared in the presence of gp91ds-tat or of the corresponding scrambled peptide. The data are means ± SEM, n= 5. *Values statistically different from the LG control, with insulin. $Values statistically different from the corresponding sample incubated without insulin. #Values statistically different from the corresponding sample incubated in the presence of HG, insulin, and the corresponding scramble peptide.

3.3 HG conditions induce insulin resistance

Hyperglycaemia modified insulin signalling as measured by PKB phosphorylation and activity. It reduced insulin-induced PKB phosphorylation (data not shown) and activation (Figure 2E), thus reflecting insulin resistance. gp91ds-tat treatment restored normal insulin response under HG treatment (Figure 2E), suggesting that NADPH oxidase activation and ROS production were responsible for this insulin resistance.

3.4 NADPH production by the PPP is required for ROS production but not for NADPH oxidase activation

To determine the origin of NADPH required by NADPH oxidase for ROS production, we used 6-AN, an inhibitor of the oxidative component of PPP. 6-AN protected cardiomyocytes against HG-induced ROS production and cell death (Figure 3A and B). Similarly, the addition of methylene blue, which acts as an ‘electron sink’ for NADPH, lowered ROS generation and cell death in response to HG (Figure 3C and D). These data indicated that the oxidative component of the PPP was required for ROS production under HG conditions. Interestingly, inhibition by 6-AN of NADPH provision did not modify Rac1 activation and p47phox translocation to the plasma membrane (Figure 3E and F). We conclude that NOX2 activation and NADPH provision by the PPP are separate but necessary steps leading to HG-induced ROS production and cell toxicity.

Figure 3

Effects of inhibitors of NADPH supply on NADPH oxidase activation and ROS production in response to HG. Effect of (A and B) 200 µM 6-AN and of (C and D) 250 nM methylene blue on (A and C) ROS production and (B and D) cell death in cardiomyocytes incubated for 6 h (ROS production) or 24 h (cell viability) with HG. Effects of 6-AN on (E) p47phox translocation and on (F) Rac1-GTP in cardiomyocytes incubated for 24 h with HG. The data are means ± SEM, n= 4. *Values statistically different from the LG control. $Values statistically different from the corresponding HG sample incubated without 6-AN or methylene blue.

3.5 HG-induced mortality is not due to protein glycosylation

Utilization of glucose in excess by the hexosamine pathway might affect cell function by protein glycosylation, which has been invoked in the hyperglycaemia-induced cell damages.27 Therefore, we studied whether HG could increase protein glycosylation. Comparison of the level of O-GlcNAc residues in LG- and HG-treated cells did not reveal any significant difference (Figure 4A). In addition, O-2-acetamido-2-deoxy-d-glucopyranosylidene amino N-phenylcarbamate (PUGNAC), an inhibitor of O-GlcNAc-ase that increases O-GlcNAc residues28 (Figure 4A), did not reproduce the effect of HG on cell death (Figure 4B), ruling out that O-GlcNAc residue formation is involved in glucotoxicity in our model.

Figure 4

Effects of PUGNAC on O-GlcNAc residues and cell death. (A) Cardiomyocytes were incubated for 24 h with LG or HG in the presence of 200 µM PUGNAC. (B) Effects of 200 µM PUGNAC on HG-induced cell death in cardiomyocytes incubated for 24 h under HG conditions. The data are means ± SEM, n= 3. *Values statistically different from the LG control.

3.6 HG-induced mortality is not due to increased glucose uptake through facilitated-diffusion glucose transporters

Exposure to HG should increase glucose uptake and metabolism, as expected from the Km of the glucose transporters (GLUTs).29 Indeed, HG stimulated glucose uptake (Figure 5A) and increased G6P (LG: 0.34 ± 0.04 vs. HG: 0.67 ± 0.13 nmol/mg proteins, P= 0.05) and glycogen content (LG: 0.29 ± 0.04 vs. HG: 0.53 ± 0.13 µmol/mg proteins, P= 0.05). These changes in glucose metabolism correlated with an increase in ROS production and cell death. However, inhibition of glucose uptake by 40 µM phloretin, an inhibitor of GLUT1 and GLUT4,30 reduced glucose uptake by 80% but did not significantly affect ROS production (Figure 5B). Moreover, addition of insulin to HG further increased glucose uptake (Figure 5A) and G6P content (HG + Ins, 2.08 ± 0.37 vs. HG, 0.67 ± 0.18 nmol/mg proteins, P< 0.05) but did not influence cell death (Figure 5C) or ROS production (Figure 5D). Taken together, these data show that glucose metabolism is not directly linked to ROS production.

Figure 5

Relationship between glucose uptake and glucotoxicity. (A) Glucose uptake was measured in cells incubated 30 min with the indicated glucose concentrations (dose–response curve in the presence (filled circle) or in the absence (open circle) of 3.10−9 M insulin). (B) Effects of 40 µM phloretin on ROS production after 2 h incubation with HG. Forty micromolar phloretin decreased glucose transport by 80% (control, 0.064 vs. phloretin, 0.013 µmol glucose/h mg proteins). Effects of 3.10−9 M insulin on (C) cell death and (D) ROS production after 24 or 6 h HG incubation, respectively. The data are means ± SEM, n= 3. *Values statistically different from the LG control. $Values statistically different from the corresponding HG sample incubated without insulin (A).

3.7 HG-induced ROS production does not require glucose metabolism but depends on the sodium/glucose cotransporter SGLT1

Incubation of cardiomyocytes with a non-metabolizable glucose analogue, that is, 3-O-methyl-d-glucopyranoside (3OMG) reproduced the effect of HG concentrations on ROS production (Figure 6A). This definitively demonstrated that the toxic effect of glucose is independent of its metabolism. A similar result was obtained with α-methyl-d-glucopyranoside (AMG), another glucose analogue, but not with 2-deoxyglucose (DOG) (Figure 6A). Accordingly, SGLT1 could sense increased glucose concentration and be responsible for NADPH oxidase activation. Indeed, SGLT1, which is expressed in cardiomyocytes (Figure 6B), transports 3OMG (Km 6 mM), AMG (Km 0.4 mM), and d-glucose (Km 0.5 mM) but not DOG (Km> 100 mM).31 However, little is known about the kinetic properties of heart SGLT1. Our hypothesis was confirmed by using the SGLT-1 inhibitor, phlorizin, which counteracted HG-induced ROS production (Figure 6C). More interestingly, it also blocked p47phox translocation (Figure 6D) and Rac1 activation (Figure 6E). Phlorizin also inhibited 3OMG-induced ROS production (data not shown).

Figure 6

Involvement of SGLT1 in HG-induced toxicity. (A) Effects of DOG, 3OMG, and AMG on ROS production, after 2 h incubation. These glucose analogues were added to the culture medium containing 5 mM d-glucose concentration. (B) Western blot representing SGLT1 in the small intestine, heart, cardiomyocytes, and liver. Effects of phlorizin on (C) ROS production, (D) p47phox translocation, and (E) Rac1 activation after 2 h incubation with HG. The data are means ± SEM, n= 3. *Values statistically different from the LG control. $Values statistically different from the corresponding HG sample, without phlorizin.

4. Discussion

The major finding of this study is that NOX2 activation (i.e. Rac1 activation and p47phox translocation to the plasma membrane) by HG is triggered by glucose transport but not by its metabolism, as non-metabolizable glucose analogues reproduce the effect of HG concentration on ROS production. The second original finding is that glucose transport through a sodium-dependent GLUT, SGLT1, is responsible for NADPH oxidase activation and subsequent increased ROS production in cardiomyocytes submitted to hyperglycaemic conditions. To our knowledge, this is the first description of a role of SGLT1 in the heart, which differs from the classical facilitated-diffusion GLUTs.

Concurring with earlier studies,13,32 we found that HG treatment increased cell death in cardiomyocytes by stimulating ROS production as a result of NADPH oxidase activation. It is also in agreement with data, showing that NADPH oxidase is one of the principal sources of ROS in cardiomyocytes33 and that myocardial NADPH oxidase activity is increased in response to HG levels.32 In addition, HG conditions induced insulin resistance, as evidenced by decreased insulin-induced PKB phosphorylation and activity, thus confirming the link between ROS production and insulin resistance that has been demonstrated in many tissues.34 This link was further demonstrated by the fact that inhibition of NADPH activation by gp91ds-tat restored the insulin-induced PKB phosphorylation and activation under HG concentrations.

HG-induced NADPH oxidase activation results from NOX2 activation via Rac1 activation and p47phox translocation to the plasma membrane. Such changes have originally been described in endothelial cells and more recently in cardiac tissue.12,35 Rac1 activation plays a crucial role in this phenomenon since Rac1 deletion decreased NADPH oxidase activation and ROS production in hyperglycaemic hearts.35

ROS production by NADPH oxidase requires a continuous supply of NADPH by the PPP as demonstrated by the protection conferred by both 6-AN and methylene blue. Interestingly, the inhibition of PPP did not affect NOX2 activation, as Rac1-GTP levels and p47phox translocation in response to HG exposure were not affected by 6-AN. NADPH oxidase activation by exposure to HG was thus the primary event and resulted in increased NADPH demand. This is in line with the generally accepted view36 that the flux trough the oxidative component of the PPP is driven by NADPH demand rather than the concentration of G6P, because NADPH is a strong inhibitor of G6P dehydrogenase. NADPH consumption by NADPH oxidase stimulates PPP flux, whereas increasing glucose metabolism is not expected to do the same independently of NADPH demand. In conclusion, NADPH oxidase activation precedes and triggers NADPH provision by PPP, which are separate events, but necessary steps for ROS production and cell toxicity.

Several reports suggest that glucotoxicity could be mediated by the hexosamine pathway resulting in protein glycosylation.37 However, no change in O-GlcNAc residues was detected in cardiomyocytes incubated for 24 h under HG conditions. Incubation with PUGNAC protected cardiomyocytes against glucotoxicity, although it increased protein glycosylation. This obviously ruled out a role of protein glycosylation in our experimental model of glucotoxicity. The role of protein glycosylation in cardiac survival remained controversial. Indeed, increased O-GlcNAc levels also exert cardioprotective effects against ischaemic injuries38 and are involved in the ischaemic preconditioning39 and in the protection conferred by glucosamine.40

Incubation with HG led to an increase in ROS production that correlated with enhanced glucose uptake, and G6P and glycogen content, suggesting that increased glucose metabolism was related to glucose toxicity. To test this relationship, we inhibited facilitated glucose uptake with phloretin, or stimulated it with insulin, and evaluated the consequences on ROS production. The results clearly showed that opposite changes in facilitated glucose uptake barely affected ROS production and that increased glucose uptake by facilitated-diffusion transporters per se did not lead to NOX2 activation and ROS production. This is in agreement with previous data, showing that cardiac overexpression of GLUT1, which increased glucose uptake, did not affect basal heart morphology and function.41 In contrast, incubation of cardiomyocytes with non-metabolizable glucose analogues mimicked the deleterious effect of HG. This was not the case for DOG. This difference corresponds to the specificity of the GLUT SGLT1 and suggested that this transporter, and not the classical GLUTs, was mediating the deleterious effects of HG. This hypothesis was further confirmed by phlorizin, which inhibits SGLT1.

SGLT1 is expressed in cardiomyocytes.19 Glucose transport into the cell is driven by the sodium gradient across the plasma membrane. The contribution of SGLT transporters to glucose transport and metabolism is marginal. Indeed, phlorizin barely affected overall glucose transport and phosphorylation at doses (10−4 M) that blocked HG-induced ROS production (see Supplemental material online, Figure S2). SGLT should increase intracellular sodium concentration, which in turn could act as a signal leading to NOX2 activation, thus transducing a metabolic signal into an ionic signal. It could act as a glucose sensor as it was already described in hypothalamic neurons.42 PKC might be involved in glucotoxicity because our preliminary results suggest that a general PKC inhibitor reduced glucotoxicity. Interestingly, SGLT1 expression is increased in type 2 diabetes and myocardial ischaemia.19 It could contribute to cellular damage and is thus a potential therapeutic target. Supposedly, the detrimental effect of HG due to this glucose sensor will be reinforced in the longer term by the process of non-enzymatic glycation because glycated proteins are known to stimulate NADPH oxidase and ROS production.43

5. Conclusion

Exposure of adult rat cardiomyocytes to HG activated signalling pathways leading to NOX2 activation by a mechanism that is not dependent on GLUTs but that requires SGLT1. Once activated, NADPH oxidase consumes NADPH provided by the PPP to produce ROS, which mediates insulin resistance and glucotoxicity.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Conflict of interest: none declared.


This work was supported by grants from the Fonds National de la Recherche Scientifique et Médicale (FNRS), Belgium, and from the Actions de Recherche Concertées, Belgium. M.B. and C.M. were supported by the Fund for Scientific Research in Industry and Agriculture, Belgium, and by a ‘Bourse du Patrimoine’ from Université catholique de Louvain. S.H. and L.B. are Research Associate and C.B. is MD Post-doctoral fellow of the FNRS, Belgium.


We thank V. Stroobant (Ludwig Institute, Brussels, Belgium) for kindly providing us with gp91ds-tat and the corresponding scrambled peptides, O. Feron and J.-L. Balligand for their comments and interest, and O. Van Caenegem for technical help.


View Abstract