High Glucose Regulates the Activity of Cardiac Sarcolemmal ATP-Sensitive K⁺ Channels via 1,3-Bisphosphoglycerate: A Novel Link Between Cardiac Membrane Excitability and Glucose Metabolism

Ventricular cardiomyocytes were dissociated from guinea pig hearts, as previously described (16). In brief, hearts were retrogradely perfused (at 37°C) with medium 199, followed by Ca²⁺-EGTA–buffered low-Ca²⁺ medium (pCa = 7) and finally low-Ca²⁺ medium containing pronase E (8 mg/100 ml), proteinase K (1.7 mg/100 ml), bovine albumin (0.1 g/100 ml, fraction V), and 200 μmol/l CaCl₂. Ventricles were cut into fragments in the low-Ca²⁺ medium enriched with 200 μmol/l CaCl₂. Single cells were isolated by stirring the tissue (at 37°C) in a solution containing pronase E and proteinase K supplemented with collagenase (5 mg/10 ml).

For conventional whole-cell electrophysiology, cells were superfused with Tyrode solution without glucose (in mmol/l: 136.5 NaCl, 5.4 KCl, 1.8 CaCl₂, 0.53 MgCl₂, and 5.5 HEPES-NaOH, pH 7.4). Pipettes (resistance 3–5 mol/lΩ) were filled with (in mmol/l) 140 KCl, 1 MgCl₂, 5 ATP, and 5 HEPES-KOH, pH 7.3. In some cells the whole-cell K⁺ current was measured using the perforated patch-clamp technique with essentially the same pipette solution as above, except ATP was omitted and amphotericin B (240 mg/ml; Sigma) was added. In the majority of experiments, we added (either in Tyrode or pipette solution, as stated in the text when appropriate) glucose or other substrates of glycolytic enzymes: glucose-6-phosphate, fructose-6-phosphate, fructose-1,6-bisphosphate, glyceraldehyde 3-phosphate, nicotinamide dinucleotide (NAD), 1,3-bisphosphoglycerate, and 3-phosphoglycerate (all from Sigma, Dorset, U.K.) and 1,3-bisphosphoglycerate, which was made as previously described by Graciet et al. (17). For all cells monitored, the membrane potential was normally held at −40 mV, and the currents evoked by a series of 400-ms depolarizing and hyperpolarizing current steps (−100 to +80 mV in 20-mV steps) were recorded directly to hard disk using an Axopatch-200B amplifier, Digidata-1321 interface, and pClamp8 software (Axon Instruments, Foster City, CA). The capacitance compensation was adjusted to null the additional whole-cell capacitative current. The slow capacitance component measured by this procedure was used as an approximation of the cell surface area and allowed normalization of current amplitude (i.e., current density). Currents were low-pass filtered at 2 kHz and sampled at 100-ms intervals. To monitor on-line behavior of single-channel molecules, the giagaohm seal patch-clamp technique was applied in the inside-out configuration. Cells were superfused with Ca²⁺-free and glucose-free Tyrode solution. Fire-polished pipettes, coated with Sylgard (resistance 5–7 mol/lΩ), were filled with (in mmol/l) 140 KCl, 1 CaCl₂, 1 MgCl₂, and 5 HEPES-KOH, pH 7.3. Recordings were made at room temperature (22°C) using a patch-clamp amplifier (Axopatch-200B). Single-channel activity was monitored on-line and stored on a personal computer. Channel activity, assayed by digitizing segments of current recordings and forming histograms of baseline and open level data points, were expressed as NP_o, where N is the number of channels in the patch and P_o is the probability of each channel to be open. Data were reproduced, low-pass filtered at 1 KHz (−3 dB), and sampled at the 100-μs rate for further analysis with pClamp8 software (18).

Data are presented as the means ± SE, with n representing the number of patched cells. Mean values were compared by the paired or unpaired Student’s t test or rank tests where appropriate. All statistical tests were performed using the SigmaStat program (Jandel Scientific). P < 0.05 was considered statistically significant.

In the current study, we considered whether glucose has acute effect on whole-cell K⁺ currents. Therefore, we measured the whole-cell K⁺ current in cardiomyocytes, using perforated patch whole-cell recording. In perforated patch-clamp mode, intracellular milieu remains largely undisturbed because amphotericin B creates holes in the sarcolemma that are permeable to ions but impermeable to larger molecules (19). Under these conditions the steady-state voltage-current (I-V) relationship (Fig. 1A) was in an N shape presumably because of inward rectification of I_k1 channels and the absence of active K_ATP channels blocked by physiological levels of intracellular ATP (Fig. 1A). The whole-cell K⁺ current was largely steady during 10 min of recording because the current density at 80 mV was 3.6 ± 0.6 and 4.5 ± 0.5 pA/pF immediately after and 10 min after, respectively, the moment when whole-cell configuration was established (P = 0.26, n = 6) (Fig. 1A–C). When glucose (50 mmol/l) was added to the extracellular solution, after 2 min of lag period, it induced outward K⁺ currents that increased at potentials more positive than −70 mV (Fig. 1B–C). Specifically, the current density at 80 mV was 3.9 ± 0.6 pA/pF in the absence of glucose and 10.0 ± 0.8 pA/pF after a 2-min incubation with glucose (P = 0.0003, n = 6) (Fig. 1B–C). Extracellular glucose applied at a concentration of 30 mmol/l also induced a statistically significant increase in the whole-cell K⁺ current (the current density at 80 mV was 3.2 ± 0.6 pA/pF in the absence of glucose and 7.1.0 ± 0.9 pA/pF after a 2-min incubation with glucose, P = 0.03, n = 4). The inward rectification of the I-V relationship was much weaker in the presence of glucose, as reflected by the loss of N shape (Fig. 1B). The effect of glucose was reversible because 2 min after glucose removal, the N shape of I-V relationship was restored and the current density at 80 mV returned to 6.1 ± 0.6 pA/pF, which was significantly lower than that obtained just before glucose removal (at time point 6 min; 10.4 ± 0.9 pA/pF, n = 6, P = 0.0003) (Fig. 1).

Properties of glucose-induced outward K⁺ current resembled those observed when K⁺ current is conducted by the opening of sarcolemmal K_ATP channels in the heart (18). To further test the hypothesis that high glucose activates sarcolemmal K_ATP channels, we used HMR 1098, a selective antagonist of these channels (20). HMR 1098 (30 μmol/l) blocked the glucose-induced increase in outward K⁺ current (Fig. 2). Specifically, in this series of experiments, current density at 80 mV was 2.8 ± 0.2 pA/pF under control conditions. In the presence of glucose (50 mmol/l), current density at 80 mV rose to 5.3 ± 0.6 pA/pF (n = 5, P = 0.01 when compared with control) (Fig. 2), whereas in the presence of both glucose (50 mmol/l) and HMR 1098 (30 μmol/l), it was 3.3 ± 0.4 pA/pF (P = 0.02, n = 5) (Fig. 2).

It has been reported that the membrane-permeant guanosine 3′,5′-cyclic monophosphate (cGMP) analog 8-(4-chlorophenylthio)-cGMP (8-p-CPT-cGMP) causes a significant decrease in glucose transport in cardiomyocytes (21). Pretreatment of cardiomyocytes for 10 min with 200 μmol/l 8-p-CPT-cGMP abolished the glucose-induced K⁺ current (n = 4) (Fig. 3). In addition, 15-min pretreatment with 2-deoxyglucose (50 mmol/l), an inhibitor of glycolysis (22), also inhibited glucose-induced K⁺ current (n = 4) (Fig. 3).

Findings that glucose transport and glycolysis is required for the activation of sarcolemmal K_ATP channels prompted us to hypothesize that some intermediate product of glycolysis may regulate the activity of K_ATP channels. To test this hypothesis we have measured membrane currents from a guinea pig ventricular myocyte using a conventional whole-cell configuration. In the whole-cell configuration of a patch-clamp technique, pipette solution replaces intracellular environment. which allowed us to fully control the composition of the intracellular environment. When physiological levels of ATP (5 mmol/l) were present in the patch pipette solution, the steady-state voltage-current (I-V) relationship (Fig. 4A) was in an N shape because of the strong inward rectification of I_k1 channels and the absence of active K_ATP channels blocked by high intracellular ATP (Fig. 4A). From the moment of establishing the whole-cell configuration, currents were measured every 2 min. A progressive decline in the magnitude of the whole-cell K⁺ current without changes in the shape of the I-V relationship was observed in every patch recorded under these conditions (n = 11) (Fig. 4A and C). The addition of 20 mmol/l glucose into the patch pipette did not significantly change the whole-cell K⁺ current in cardiomyocytes in terms of magnitude, shape, or time-dependent magnitude decline (n = 6) (Fig. 4B and C). Because these results suggested that glucose does not have direct effects on K⁺ currents, we hypothesized that glucose-induced activation of K_ATP channels might be mediated by some of the glycolytic products. Therefore, we measured the whole-cell K⁺ current in cells with intracellular environments containing only pipette solution, substrates of glycolytic enzymes, and, when ATP was not applied as a substrate, a physiological level of ATP (5 mmol/l). Specifically, we tested the effects of the following substrates on the whole-cell K⁺ current: glucose-6-phosphate/ATP (20 mmol/l each), fructose-6-phosphate/ATP (20 mmol/l each), or fructose-1,6-bisphosphate (20 mmol/l fructose 1,6-bisphosphate and 5 mmol/l ATP; this concentration of ATP was applied because fructose-1,6-bisphosphate is the sole substrate for aldolase, and ATP was applied not as a substrate but as a K_ATP channel ligand to assess whether the product would have a direct effect on the channels) (Fig. 5). None of these substrates had a significant effect on the whole-cell K⁺ current (n = 5 for each) (Fig. 5).

When substrates of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were concomitantly present (glyceraldehide-3-phosphate and NAD, 20 mmol/l each) in the pipette solution containing 5 mmol/l ATP, K_ATP channel–mediated current was observed starting from the 2-min time point, and recordings revealed that outward K⁺ currents associated with a much weaker inward rectification of the I-V relationship (n = 5) (Fig. 6). The current density at 80 mV rose from 3.1 ± 1.0 pA/pF at time 0 to 14.2 ± 3.0 pA/pF after a 2-min incubation with GAPDH substrates (P = 0.0003) (Fig. 6). This increase in K⁺ current was maintained by the end of the experiment (Fig. 6). Glyceraldehide-3-phosphate (20 mmol/l) and NAD (20 mmol/l) added into pipette solution (containing 5 mmol/l ATP) alone did not have any effect on the whole-cell K⁺ current in cardiomyocytes (n = 4 for each) (Fig. 6).

Because the opening of sarcolemmal K_ATP channels by GAPDH substrates was detected, it was possible that the GAPDH-catalyzed reaction product 1,3-bisphosphoglycerate activates K_ATP channels. This hypothesis was tested using both whole-cell and single-channel electrophysiology. We have added 1,3-bisphosphoglycerate (20 mmol/l) into the pipette solution containing ATP (5 mmol/l) and measured the whole-cell K⁺ current. Under these conditions, from the beginning of the experiment, whole-cell current had properties of the K⁺ current flowing through K_ATP channels, i.e., the N shape of I-V relationship—typical for the whole-cell K⁺ current in cardio- myocytes when K_ATP channels are closed—was lost (Fig. 7A). The addition of 1,3-bisphosphoglycerate (20 mmol/l) on the intracellular face of excised membrane patches induced vigorous opening of K_ATP channels in the presence of 5 mmol/l ATP (Fig. 7C).

In the current study, we have suggested that high glucose may activate sarcolemmal K_ATP channels in cardiomyocytes. The glucose-induced activation seems to be mediated by glycolysis and the production of 1,3-bisphos- phoglycerate, which directly targets and activates sarcolemmal K_ATP channels.

In previous studies, the effect of high glucose on cardiac membrane excitability has been rarely investigated under nonstress conditions. In genetically and streptozocin-induced diabetic rats, transient outward K⁺ current is greatly reduced (3,23), and acute hyperglycemia inhibits the opening of HERG K⁺ channels, which are ion channels controlling the rapid component of the delayed rectifier K⁺ current in the heart, which is the major repolarizing current in the plateau voltage range of cardiac action potential (6). It has been established since the 1970s that elevating extracellular glucose dramatically attenuates action potential duration shortening during hypoxia in intact cardiac tissue, and that action potential duration shortening is even more rapid in the absence of glucose (24–26). However, it should be said that, as opposed to hearts exposed to ischemia/hypoxia, in hearts under physiological conditions, glycolysis is only a minor source of intracellular ATP (27). Considering that the bioenergetic status of the heart is linked with its membrane excitability via K_ATP channels, the effect that hyperglycemia would have in normoxic and hypoxic hearts may be quite different. More specifically, intracellular ATP inhibits K_ATP channels (15), and it is believed that ischemia/hypoxia-induced decreases in cytosolic ATP concentration activate these channels, leading to efflux of K⁺ and shortening of the action membrane potential. During hypoxia, glycolysis takes over fatty acids in producing ATP, which in turn may lead to glycolysis-mediated inhibition of sarcolemmal K_ATP channels (indirectly via ATP production) and attenuation of action potential duration shortening during hypoxia. In contrast, in normoxic hearts K_ATP channels are closed anyway, and in any case, glucose is not an important source of ATP, so hyperglycemia may not have such dramatic effects on action potential duration. Because hyperglycemia per se inhibits HERG K⁺ channels (6), it should significantly increase the action potential duration in normoxic hearts. However, it has been recently reported that hyperglycemia has no significant effects on action potential duration in nonstressed hearts (28), which may be explained by the activation of K⁺ currents (such as K_ATP channel–mediated currents), which would counteract the effect of hyperglycemia on HERG K⁺ channels. Thus, although on first sight our results seem to contradict the established dogma that hyperglycemia increases action potential duration, they actually fit very well into the most recent findings showing that although hyperglycemia blocks HERG K⁺ channels, it still does not affect action potential duration in the heart.

The K_ATP channel–opening effect of glucose was not observed when conventional whole-cell instead of perforated-patch electrophysiology was used. This would suggest that an intact nondialyzed intracellular environment was essential for glucose-induced activation of K_ATP channels. 8-p-CPT-cGMP and 2-deoxyglucose, inhibitors of glucose transport and glycolysis, respectively (21,22), inhibited the effect of glucose, suggesting that the transport of glucose into cells and glycolysis are required for the opening of sarcolemmal K_ATP channels. Because pipette solution replaces cytosol during whole-cell patch-clamp recordings, enzymes that are not attached to the cellular cytoskeleton and other structural proteins are being removed from their physiological sites of action. This, in turn, may explain why glucose was without any effect when K⁺ current was measured using conventional whole-cell electrophysiology.

Results with inhibitors of glucose transport and glycolysis allowed a supposition that one of the intermediary products of the glycolytic pathway might possess previously unrecognized K_ATP channel–opening properties. To examine this hypothesis, we have measured the whole-cell K⁺ current in cells with intracellular environments containing only pipette solution, substrates of glycolytic enzymes, and physiological levels of ATP (5 mmol/l). Under these conditions, glucose-6-phosphate (a substrate of phosphoglucose isomerase generating fructose-6-phosphate), fructose-6-phosphate/ATP (substrates of phosphofructokinase generating fructose-1,6-bisphosphate), and fructose-1,6-bisphosphate (a substrate of aldolase generating glyceraldehide-3-phosphate) did not have any effect on the whole-cell K⁺ current. This would stand against the possibility that some of these substrates, either directly or via generated product(s), regulate the activity of sarcolemmal K_ATP channels. However, substrates of GAPDH (glyceraldehyde-3-phosphate and NAD) activated K_ATP channels. Because this was observed only when glyceraldehyde-3-phosphate and NAD were applied together, it was likely that the opening of K_ATP channels was associated with the GAPDH catalytic activity, and that the product of GAPDH-catalyzed reaction, 1,3-bisphosphoglycerate, may be responsible for the activation of sarcolemmal K_ATP channels. Whether 1,3-bisphosphoglycerate is an opener of sarcolemmal K_ATP channels was tested using both whole-cell and single-channel electrophysiology. When present in pipette solution, 1,3-bisphosphoglycerate induced whole-cell outward K⁺ currents, despite the concomitant presence of millimolar ATP, suggesting that this compound opens sarcolemmal K_ATP channels. Direct application of 1,3-bisphosphoglycerate on the intracellular side of excised membrane patches activated K_ATP channels, and this further supports the notion that this intermediary product of glycolysis targets and activates K_ATP channels.

The consequences of opening the sarcolemmal K_ATP channels for cardiac function are complex, and they may depend on the extent of channel activation and probably other as yet unknown factors. On one hand, the opening of the channels increases the resistance of cardiomyocytes against ischemia/hypoxia and protects the heart against myocardial infarction (29,30). On the other hand, the activation of K_ATP channels may be associated with the appearance of arrhythmias. It has been suggested that the activation of sarcolemmal K_ATP channels could be both pro- and antiarrhythmogenic (31,32). Increased K⁺ conductance shifts resting membrane potential toward K⁺ equilibrium, which should reduce ectopic pacemaker activity. However, the activation of K_ATP channels accelerates the depolarization of the action potential, reducing the refractory period of the cell, which may increase the probability of reentrant arrhythmias (33). It has been shown that action potential duration is much longer in diabetic that nondiabetic hearts (28). It has been suggested that this effect of diabetes is a consequence of downregulation of ion channels controlling transient outward K⁺ currents (3,23). However, although this diabetes-induced change of action potential duration in the heart is chronic, it can be acutely corrected with insulin (28). Because insulin does not directly activate K⁺-selective conductance in the heart (34,35), our data would explain very well how insulin, by stimulating glucose transport, glycolysis, and 1,3-bisphosphoglycerate production, shortens/corrects action potential duration in diabetic hearts.

Previously, it was shown that lipids/fatty acids, which are the major source of ATP in the heart, are directly linked with heart membrane excitability via phosphatidylinositol phosphates and long-chain acyl-CoA esters (36,37). Analogously to the role that phosphatidylinositol phosphates and long-chain acyl-CoA esters play in linking fatty acids with K_ATP channels, 1,3-bisphosphoglycerate seems to play a role in linking glucose metabolism with cardiac electrophysiology. The existence of direct, ATP-independent signaling pathways in regulating the K_ATP channels of metabolites that produce ATP probably allows subtlety in the regulation of these channels. As an example, because glucose is not the only source of ATP in the heart, ATP-sensing by K_ATP channels would be reflective of the overall metabolic status of cardiac cells but not of the status of glucose metabolism. As opposed to ATP, levels of 1,3-bisphosphoglycerate depend solely on glucose levels and metabolism. This means that glucose metabolism, independently of ATP production or any other factors, is transduced into changes in membrane excitability that make the heart responsive to changes in glucose levels, irrespective of the conditions of other metabolic pathways and/or total energy status. This type of link between K_ATP channel activity and glucose metabolism may be important not only for cardiac cells but also for other cell types with functions regulated by K_ATP channels, including insulin-secreting β-cells.

This research was supported by grants from the British Heart Foundation, the BBSRC (Biotechnology and Biological Sciences Research Council), the MRC (Medical Research Council), Wellcome Trust, Anonymous Trust, and Tenovus-Scotland. We thank Aventis Pharma (Frankfurt, Germany) for HMR 1098.

A majority of the cardiomyocytes used in this study were isolated by G.R. Budas.