Hyperglycemia is a major manifestation of all forms of diabetes mellitus and is associated with increased risk of cardiovascular disease. It is well established that cardiac excitation-contraction (E-C) coupling is adversely affected in diabetic animals such that ventricular myocyte action potential duration is prolonged and intracellular Ca2+clearing and mechanical relaxation are slowed. We now report that ventricular myocytes incubated in a culture medium containing high extracellular glucose (25.5 mM) also exhibit these same changes in E-C coupling. These effects are not manifested for ∼24 h after exposure. Furthermore, in the presence of normal glucose (5.5 mM), relaxation is also prolonged by fructose (20 mM), yet is unaffected by equimolar concentrations of nonmetabolizable sugars such as l-glucose and mannitol, implying that the high glucose effects require glucose entry into the cell and metabolic processing. The prolonged relaxation can also be produced by 5 mM glucosamine (an intermediate of glycosylation) and is blocked by 0.5 μg/ml tunicamycin (an inhibitor of N-linked glycoprotein synthesis). Culturing myocytes with an inhibitor of glycation (10 mM aminoguanidine) does not prevent the high extracellular glucose concentration effects. Thus our data indicate that high extracellular glucose impairs cellular mechanisms contributing to myocardial relaxation and that this impairment may involve glycosylation of nascent proteins.
- high glucose
- calcium ion transient
diabetes mellitus alters intrinsic properties of ventricular myocytes that contribute to abnormal heart function. There is considerable evidence in both diabetic humans and animals that diastolic dysfunction occurs before systolic abnormalities and that these changes can occur independent of coronary artery disease (10, 11,42, 43). In ventricular myocytes isolated from diabetic animals, impaired relengthening is associated with prolonged action potentials (APs) and slowed cytosolic Ca2+clearing (19, 21, 22, 29, 32, 36). These myocyte abnormalities can develop as early as a few days after induction of experimental diabetes (32, 36). The factors associated with the pathogenesis of this cardiomyopathy are unknown.
We have recently taken advantage of an adult myocyte culture system and manipulated the culture conditions to mimic certain aspects associated with diabetes (e.g., hypoinsulinemia/hyperglycemia). Adult rat ventricular myocytes cultured in a chemically defined (i.e., serum-free) medium maintain an adult phenotype for days (9). By culturing normal ventricular myocytes for a few days in a “diabetic-like” medium, relaxation is impaired in a manner similar to diabetes (7). The diabeticlike medium contains five times less insulin and approximately five times more glucose than our normal medium (changes similar in magnitude to those that occur in rats after streptozotocin-induced diabetes).
We have preliminary evidence that high extracellular glucose, rather than low extracellular insulin, affects abnormal relaxation in our cell culture system (33). Hyperglycemia is linked to the etiology of vascular complications associated with diabetes (31, 37), but it is not known whether hyperglycemia alters the cellular mechanisms that contribute to prolonged cardiac relaxation. High extracellular glucose has been shown to elevate intracellular calcium concentration ([Ca2+]i) in both vascular smooth muscle cells (2) and cardiac myocytes (13), which in turn may alter protein function and gene expression. High glucose has been shown to alter intracellular signaling pathways [i.e., protein kinase C (PKC) and phospholipase A2] in vascular cells (20,38, 39). Diabetes also elevates cytosolic PKC activity and expression in both cardiac tissue (17, 40) and isolated myocytes (23).
The following experiments were designed to determine whether culturing ventricular myocytes in high extracellular glucose mediates changes in excitation-contraction (E-C) coupling and to begin to delineate the underlying cellular mechanisms associated with these changes. Cardiac myocyte E-C coupling was evaluated in cultured cells by use of perforated patch recording to measure APs, fluorescence spectroscopy to assess [Ca2+]itransients, and video edge detection to measure cell shortening and relengthening in response to electrical stimulation. By culturing adult rat ventricular myocytes in a serum-free medium for 1 day in the presence of high extracellular glucose (25.5 mM), we are able to show that elevated sugar changes cellular mechanisms that manifest in a prolonged AP, slower cytosolic Ca2+ clearing, and impaired relaxation.
Cell isolation and culture. Single ventricular myocytes were isolated enzymatically from the hearts of adult male Sprague-Dawley rats (200–250 g; Harlan Bioproducts, Indianapolis, IN) using the method described previously (32). Isolated myocytes were plated on glass coverslips precoated with laminin (10 μg/ml; Collaborative Biochemical Products, Bedford, MA) and maintained for 1–2 days in a defined medium consisting of medium 199 (Sigma) with Earle’s salts containing 25 mMN-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) and NaHCO3supplemented with 2 mg/ml albumin, 2 mM l-carnitine, 5 mM creatine, 5 mM taurine, 0.1 μM insulin, 0.1 nMd-triiodothyronine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 75 μg/ml gentamicin (9). The medium composition includes 5.5 mM d-glucose and will be referred to as “normal” medium.
Cell shortening/relengthening. Mechanical properties of cultured ventricular myocytes were assessed by a video-based edge detection system (Crescent Electronics, Salt Lake City, UT) as described previously (32). In brief, coverslips with cells attached were placed in a chamber mounted on the stage of an inverted microscope (Nikon Diaphot) with the temperature maintained at 37°C. The chamber was superfused (∼2 ml/min) with a buffer containing (in mM): 131 NaCl, 4 KCl, 1 CaCl2, 1 MgCl2, 10 glucose, and 10 HEPES, pH 7.4. The cells were field stimulated to contract at a frequency of 0.5 Hz. Shortening of rod-shaped myocytes was detected at both longitudinal edges at a video sweep speed of 120 Hz while sampling at 333 Hz. The indexes used to describe isotonic shortening were peak shortening amplitude normalized to cell length (PS), time to PS (starting at 10% above baseline and ending at 90% of maximal shortening), and area under the shortening phase normalized to PS. The indexes used to describe relaxation were time of relengthening (TR; starting at 90% of PS and ending at 10% above baseline) and area under the relengthening phase normalized to PS. All indexes (including those describing Ca2+ transients and APs) were analyzed off-line using Clampfit (Axon Instruments) and were determined after averaging 5–10 steady-state twitches for each myocyte.
Fluorescence measurement32). Myocytes were placed on an inverted microscope equipped with a heated (37°C) and light-tight chamber and were imaged through a ×40 oil objective. Field stimulation protocol and superfusing buffer were the same as those described above for mechanical measurements. Cells were exposed to light emitted by a 75-W lamp and were passed through either a 360- or a 380-nm filter (bandwidths were ±15 nm). Fluorescence emmisions were detected between 480 and 520 nm by a photomultiplier tube after first illuminating cells at 360 nm for 0.5 s and then at 380 nm for the duration of the recording protocol (333-Hz sampling rate). The 360 excitation scan was repeated at the end of the protocol, and an interpolated signal was calculated and used to calculate the ratio with the 380-nm emission (360/380 ratio). Diastolic Ca2+ was defined as resting 360/380 ratio, systolic Ca2+ was defined as peak 360/380 ratio, and rate of cytosolic clearing was determined by the time constant (τ) using a single exponential equation to describe the Ca2+transient decay from peak.
Perforated patch-clamp technique. Myocytes attached to coverslips were transferred to a temperature-controlled (37°C) bath chamber and were superfused with (in mM) 132 NaCl, 1.2 MgSO4, 20 HEPES, 11.1 glucose, 4 KCl, and 2 CaCl2, pH 7.4. Transmembrane APs were recorded through an Axopatch 200A amplifier (Axon Instruments), and myocytes were stimulated at 1 Hz with intracellular current pulses, typically 200–500 μs in duration. The pipette solution contained (in mM) 130 potassium aspartate, 15 KCl, 5 HEPES, 10 NaCl, and 0.5 CaCl2, pH 7.2. Amphotericin B was used (150 μg/ml, dissolved in dimethyl sulfoxide) as the perforating agent (12).
Statistical analyses. Data are presented as means ± SE. Statistical significance was estimated by analysis of variance (SYSTAT, SPSS, Chicago, IL). Follow-up tests for multiple comparisons were chosen depending on whether significance (P < 0.05) was identified in main effects and/or interaction terms.
E-C coupling. We have recently shown that maintaining ventricular myocytes isolated from normal adult rats in a serum-free medium containing low insulin and high glucose impairs relaxation in a manner similar to diabetes (7, 32). In the present study, myocytes in low insulin/high glucose medium demonstrate normal contractile properties (only PS is shown in Fig.1) but prolonged relaxation, indicated by increase time for relaxation and area under the relaxation phase normalized to PS (Fig. 1). To distinguish the effects of insulin from those of glucose, we maintained cells in either low insulin or high glucose. High extracellular glucose concentration (high [glucose]), but not low insulin, prolonged relaxation in cultured myocytes (Fig. 1). It should be emphasized that each group was evaluated from multiple cultures and that all groups were evaluated on the same experimental day (e.g., Fig. 1). This procedure was used to control for any potential variability among cultures. The data are pooled values for each group where culture was not considered a factor (i.e., sample sizes represent the number of myocytes measured/group). The lower insulin concentration was chosen on the basis of changes in serum levels associated with streptozotocin-induced diabetes (∼ normal) and reflects a decrease relative to our normal medium insulin rather than physiologically relevant concentrations. When myocytes were cultured in our normal medium without insulin, relaxation was similar to that of myocytes in normal medium with insulin, and high glucose prolonged relaxation to the same extent in both groups [TR was longer by ∼35%, and the area under the relengthening phase normalized to PS (AreaR/PS) was larger by ∼22%]. TR tended to be slightly, but not statistically, longer in normal medium without insulin (114 ± 8 ms,n = 26) than in normal medium (99 ± 11 ms, n = 18), whereas AreaR/PS was the same (normal medium without insulin, 80 ± 5 μm ⋅ ms/μm,n = 26 and normal medium, 76 ± 5 μm ⋅ ms/μm, n = 18). These data strongly suggest that higher glucose and not lower insulin is the factor affecting abnormal relaxation in our culture system.
The membrane-permeable dye fura 2-AM was used to evaluate the time course of Ca2+ transients in myocytes cultured for 1 day in either normal or high [glucose] medium. The time course of the fluorescence signal decay was described by a single exponential equation, and the τ was used as a measure of the duration of free cytoplasmic Ca2+. In myocytes cultured in high [glucose], decay of the Ca2+ transients was significantly slower than in myocytes cultured in normal [glucose] (Fig.2). It is known that the rate of sarcoplasmic reticulum Ca2+ uptake is Ca2+ dependent and that the time course of the transient decay is influenced by peak cytosolic Ca2+ (4). We have recently shown that the relationship between peak Ca2+ and τ is altered in myocytes isolated from diabetic animals (32) and cells cultured in low insulin/high glucose medium (7). Here, we also analyzed the transient decay in a manner that takes into account differences in peak Ca2+ (Fig. 2). These data show that the relationship between the peak Ca2+ ratio and the rate of transient decay can be described by a linear equation in myocytes cultured in normal [glucose]; however, high [glucose] disrupts this relationship as indicated by the increase in data heterogeneity and an upward shift in the regression (suggesting slower decay across a range of peak [Ca2+]i). The longer τ (slower decay) is consistent with that seen in diabetic myocytes (21, 32) and likely contributes to impaired relaxation.
We recorded transmembrane potentials using perforated patch-clamp techniques to determine whether high [glucose] prolongs the AP in cultured myocytes in a manner similar to diabetes (19, 22, 36). After only 1 day in culture, the AP duration was longer in myocytes cultured in high [glucose] compared with cells cultured in normal [glucose] (Fig. 3). The resting membrane potential was significantly different between high [glucose] myocytes (−77 ± 1 mV) and normal cells (−81 ± 1 mV, P < 0.05), but whether this slight depolarization reflects a physiologically meaningful difference is uncertain.
Myocyte mechanics after culturing in glucose metabolites. To determine whether abnormal E-C coupling induced by high [glucose] in culture is dependent on metabolism of glucose, myocytes were maintained in medium containing 5.5 mM d-glucose and either 20 mM fructose (which can be converted to fructose 6-phosphate, an intermediate metabolite in glycolysis) or 20 mM l-glucose (a nonmetabolizable isomer of glucose). For these experiments, the culture medium always contained 5.5 mM d-glucose and 20 mM of the added hexose. High fructose concentration, like high [d-glucose], prolonged myocyte relaxation, whereas 20 mM l-glucose had no effect (Fig. 4). To determine whether these effects were due to changes in extracellular osmotic pressure, we also cultured myocytes in 20 mM mannitol (to which the cell is impermeable). Mannitol had no effect on relaxation (Fig. 4).
To investigate the possibility that nonenzymatic protein glycosylation (glycation) might be involved in prolonged relengthening, we cultured cells in high [glucose] with an inhibitor of glycation. Culturing myocytes in aminoguanidine (10 mM), which has been widely used to prevent glycation in many cell types in vitro and in vivo (5,28, 28), did not prevent abnormalities in myocyte relaxation induced by high [glucose]. Relaxation time in high [glucose] myocytes (116 ± 9 ms) and high [glucose] plus aminoguanidine (119 ± 9 ms) was significantly longer than in either normal myocytes (91 ± 4 ms) or normal plus aminoguanidine (91 ± 5 ms,n = 28–29/group). We also cultured myocytes in high [glucose] for a shorter time period. After culturing myocytes in either normal or high [glucose] for 4–6 h (rather than 24 h), we were unable to detect any differences in mechanical properties between normal and high [glucose] myocytes (e.g., relaxation time for normal myocytes = 91 ± 7 ms and for high [glucose] cells = 89 ± 7 ms, n = 24/group).
Glucose and fructose can be metabolized through glycolysis and oxidative phosphorylation, or converted to glucosamine 6-phosphate, which in turn is used for the glycosylation of proteins in the endoplasmic reticulum. A small portion of cellular fructose 6-phosphate is used for glycosylation (25). The addition of 5 mM glucosamine to the culture medium (in the presence of normal [glucose]) also prolonged myocyte relaxation (Fig.5 A), suggesting that glycosylation is involved in impaired relaxation. To further evaluate the role of glycosylation in our cell system, myocytes were cultured with high [glucose] and tunicamycin (0.5 μg/ml). Tunicamycin blocks the addition of UDP-N-acetylglucosamine to dolichol phosphate, the first step in the formation of oligosaccharides destined to become N-linked glycoproteins. Tunicamycin prevented abnormal relaxation induced by high [glucose] (Fig. 5 B). A key observation of these experiments is that the same concentration of tunicamycin had no effect on mechanical properties in cells cultured in normal [glucose] (Fig.5 B).
The principal findings of this investigation show that ventricular myocytes isolated from normal animals and maintained in a serum-free, high [glucose] medium exhibit prolonged AP, slower cytosolic Ca2+ clearing, and prolonged relaxation after only 1 day. The effects of high [glucose] medium on E-C coupling are the same as those seen in myocytes isolated from animals that have been diabetic for only a few days (32, 36). Pathway(s) by which high [glucose] alters E-C coupling mechanisms appear to involve changes in glucose metabolism, since they are dependent on the presence of biologically active sugars (e.g., d-glucose, fructose, and glucosamine).
Changes in Ca2+ transmembrane fluxes and/or intracellular Ca2+ binding may play a key role in affecting abnormal relaxation in high [glucose] myocytes. Prolonged relaxation measured in cells cultured in high [glucose] as well as myocytes from diabetic animals (21,32) may result, in part, from longer AP durations (Fig. 3). Sustained depolarizations could allow trigger Ca2+ to enter the cell for a longer period of time, thus contributing to longer cytosolic Ca2+ transients (slower decay; Fig. 2). The ionic mechanisms that produce the prolonged AP are presently being investigated. In myocytes isolated from diabetic animals, the transient outward K+current is significantly depressed (19, 22, 36). Changes in L-type Ca2+ current are more subtle and may or may not contribute to slower repolarizations (19, 36), at least in early stages of the disease and in our cell culture system.
Glucose must be metabolized to impair relaxation, and it appears that some form of posttranslational protein modification plays a role in these high [glucose]-induced effects. In the presence of normal d-glucose, neither mannitol norl-glucose affect relaxation (Fig. 4); thus increases in osmotic pressure cannot account for this effect. Furthermore, myocytes exposed to high [glucose] for only 4–6 h (rather than 1 day) demonstrate normal mechanics, suggesting that some type of processing must take place rather than the acute activation of a regulatory mechanism that directly affects relaxation. For example, elevated [glucose] has been shown to be detrimental to cell function when sugars react nonenzymatically with free amino groups of proteins to form a reversible Schiff base and then undergo further rearrangement and cross-linking to form irreversible advanced glycation end products (5, 6). Glycation is generally viewed as a slow process, usually taking weeks to months to manifest changes in the amount of glycated proteins. However, to determine whether these changes occur in our cell culture system, we maintained myocytes in high [glucose] and aminoguanidine (an inhibitor of glycation). Aminoguanidine did not prevent abnormal relaxation in myocytes cultured in high [glucose]. Because of the apparent time course of the high [glucose] effect, our data suggest that intracellular processing (i.e., protein synthesis or degradation) is involved in prolonging relaxation as opposed to covalent or allosteric modification of proteins.
Marshall et al. (24) demonstrated that enzymatic glycosylation increases in cells with prolonged exposure to high [glucose]. With increased availability of glucose, metabolizable sugars may be preferably shunted through glycosylation pathway(s). Nishio et al. (27) have recently shown that a gene believed to express a regulatory enzyme involved inO-linked glycosylation is increased in heart by either diabetes or by culturing myocytes in high extracellular glucose. Both glucose and fructose can also be used inN-linked glycosylation of proteins. Fructose 6-phosphate can be converted to glucosamine 6-phosphate then UDP-N-acetylglucosamine, which in turn is coupled to a dolichol phosphate carrier in the endoplasmic reticulum. Tunicamycin competes with UDP-N-acetylglucosamine and inhibitsN-linked glycosylation. Tunicamycin blocks the adverse effects of high [glucose], suggesting that elevated glycosylation plays a role in abnormal E-C coupling (Fig.5). Tunicamycin may attenuate glucose uptake and prevent the abnormal effects of high [glucose] by altering the glycosylation of protein(s) in the glucose transporter GLUT-4 (14). In the presence of normal [glucose], tunicamycin had no detrimental effects on myocyte mechanics, thus inhibition of protein synthesis through this pathway can not account for the adverse effects of high [glucose]. Figure 5 presents data further supporting the view that abnormal glycosylation plays a role in our high [glucose] effects. Maintaining myocytes in normal [glucose] and glucosamine (which can enter cells and be phosphorylated) produces the same prolonged relaxation as that seen in myocytes cultured in high [glucose].
It is important to emphasize that our cell culture system targets the pathogenesis of diabetic cardiomyopathy, but it is clearly inadequate to assess the pathophysiology of long-term diabetes. We recognize the limitations of our model system when relating the effects of high [glucose] in vitro to those in vivo. Our cultured myocytes are quiescent and thus do not undergo the dynamic nature of a beating heart (e.g., Ca2+ fluctuations on a beat-to-beat basis) and are isolated from other cells and endogenous factors. This may, in part, contribute to the variability that we see among cell cultures. For example, indexes for relengthening (TR andA R) were ∼25% faster in Fig. 5 than in Figs. 1 and 4. We were unable to detect any differences in other mechanical indexes [e.g., cell length, PS, time to PS, maximum rates of shortening and relengthening (some data not shown)] that would indicate that these normal cells are different from other normal cells. We therefore conclude that there is interculture variability. We would like to reiterate that each study represents data from at least three separate cultures and always includes normal and high [glucose] groups in parallel with other test groups. Regardless of the TR in normal cells (i.e., 100 or 75 ms), high [glucose] increased TR by ∼59%. Consistent changes were also seen in the indexA R, with an average increase of ∼37% with high [glucose].
Collectively, these data support the view that high extracellular glucose impairs myocyte relaxation, in part, by prolonging systolic Ca2+ in a glycosylation-dependent manner. The signaling pathway for glucose-induced changes in intracellular Ca2+ regulation in ventricular myocytes is unknown. It has been shown in vascular smooth muscle cells that exposure to high [glucose] in vitro alters the activity and gene expression of a number of proteins (e.g., Na-K-ATPase, PKC, phospholipase A2, transforming growth factor-α; see Refs. 18, 26, 38, 39). In red blood cells, high [glucose] inhibits sarcolemmal Ca2+-ATPase, apparently through a glycosylation-dependent mechanism, and is independent of nonenzymatic glycation (8). Although sarcolemmal Ca2+-ATPase is relatively unimportant in regulating [Ca2+]iduring the rat cardiac contractile cycle, it may play an important role in regulating resting [Ca2+]iin ventricular myocytes (3). Another mechanism by which culturing myocytes in high glucose may alter relaxation is through modification of intra- or extracellular matrices (e.g., cytoskeletal filaments). It has been shown that high [glucose] can modify structural molecules (e.g., F-actin and collagen) through glycation (16, 44), but we are unaware of evidence showing that glycosylation plays a role in cytoskeletal modification in muscle. At this time, we cannot rule out any target molecules; however, the effects of glycation do not appear to be directly involved.
In the context of diabetes, there are clearly multifactorial changes that occur in cardiac tissue. For instance, myocardial membrane preparations and cardiac tissue homogenates from diabetic animals demonstrate impaired sarcoplasmic reticulum Ca2+-ATPase, Na-K-ATPase myosin ATPase activity, depressed Na/Ca exchange activity, and possibly sarcoplasmic reticulum Ca2+release (1, 30, 34). Consistent with the data from multicellular preparations, ventricular myocytes isolated from diabetic animals demonstrate prolonged AP and Ca2+transients, attenuated rapid cooling contractures, and diminished myofilament Ca2+ sensitivity (7,15, 36, 41). Diabetes has been shown to elevate translocation of PKC, increase membrane PKC activity (18, 40), and increase expression of the PKC epsilon isoform (23). A number of the cellular changes in E-C coupling occur only days after induction of diabetes (32, 36). Our data show that some of the same abnormalities in E-C coupling manifest in myocytes after a short exposure to high [glucose] (in vitro).
The signaling pathways that lead to changes in cardiomyocyte E-C coupling in diabetes have not been defined. However, the quick onset (24 h) of the high [glucose] effects on myocyte E-C coupling may indicate that these early events represent changes in modification of either target proteins (i.e., ion channels or pumps) or modulatory proteins (i.e., PKC or calmodulin kinase II). Many regulatory proteins are known to directly alter ion channel and pump function (e.g., PKC depresses the transient outward K+ current and alters sarcoplasmic reticulum Ca2+ release channel activity). Depressed transient outward K+ current (in rats) would lead to a prolonged AP, which could account for our observations of slowed cytosolic Ca2+ clearing and relaxation. Changes in regulatory factors may also affect gene expression.
We have begun to delineate a novel regulatory process that should impact our understanding of physiological as well as pathophysiological events in heart. Our data demonstrate that high [glucose] (in culture) alters E-C coupling in normal myocytes by prolonging AP duration, Ca2+ transients, and relaxation. These effects may, in part, be mediated by posttranslational mechanisms (supported by the glucosamine and tunicamycin experiments). Changes in function and/or expression of E-C coupling proteins may be due to abnormal modification (i.e., ion channels or pumps are not inserted into membranes properly or do not function properly; see Ref. 35) or through modifications in regulatory proteins (e.g., PKC). Our data indicate that high extracellular glucose modulates cellular mechanisms associated with E-C coupling in the heart.
We thank Nidas Undrovinas for expert technical assistance and Drs. James Marsh and Mary Schwanke for critically reviewing this manuscript.
Address for reprint requests: A. J. Davidoff, Dept. of Pharmacology, College of Osteopathic Medicine, University of New England, 11 Hills Beach Rd., Biddeford, ME 04005.
This research was partially supported by American Heart Association-Michigan Grant 72GB967 (to A. J. Davidoff), National Heart, Lung, and Blood Institute Grant HL-49918 (to G. A. Gintant), and by a Louis M. and Mollie Elliman Foundation Postdoctoral Fellowship (to J. Ren).
- Copyright © 1997 the American Physiological Society