Main

Mechanical, electrical, and metabolic responses to changes in oxygen and glucose availability as well as the protective and antiarrhythmic properties of glucose during ischemia-reperfusion transitions have been extensively studied in a number of experimental models of adult heart (e.g. seeRefs. 1–3) but scarcely in the embryonic heart.

During fetal life, the heart is particularly vulnerable to hypoxemia(4), although it seems to be more dependent on glycolysis than on oxidative phosphorylation(5–7). It has also been shown that embryonic ventricular cells are more resistant to hypoxia than adult myocardial cells(8) and that glycolysis supports contractile activity during anoxia(9).

Throughout the 1st wk of development of the chick, the heart intensively catabolizes glucose via glycolysis and pentose phosphate pathway(10, 11), and its sensitivity to deprivation of oxygen and glucose increases(12, 13). The heart contains a large amount of glycogen(14), and its pulsatile activity is maintained for several hours either in normoxic glucose-free medium or in anoxic medium in the presence of glucose(15). Recently, we have found that the isolated heart of the 4-d-old chick embryo reacted rapidly and reproducibly to repeated anoxia-reoxygenation transitions, inducing reversible contractile disturbances(16, 17). Namely, reoxygenation induced a transient cardiac arrest followed by arrhythmias, which appeared to be the embryonic analog of the oxygen paradox in the adult heart. Under physiologic conditions, the embryonic myocardium is rather hypoxic(18) and utilizes preferentially glucose and lactate as energy substrates(19), contrary to the adult heart which oxidizes essentially FFA. Therefore, we hypothesized that changes in glucose availability should particularly affect the contractile activity of the embryonic chick heart subjected to anoxia-reoxygenation transitions and that fluctuations of glycolytic flux could induce arrhythmias.

METHODS

Fertilized eggs from Warren hens were incubated 96 h at 38°C and 90% relative humidity to obtain embryo at stage 24 according to Hamburger and Hamilton(20). Whole hearts (about 2 mm long) were dissected from the embryo and mounted in a thermostabilized (37.5°C) airtight chamber under controlled metabolic conditions, following the previously published procedure(21). Briefly, the spontaneously beating hearts were placed in the lower culture compartment (300μL) of the chamber which was separated from the upper gas compartment (300μL) by a transparent and thin 15-μm thick silicone membrane (RTV 141, Rhône Poulenc) allowing gas exchanges. As myocardial vascularization is absent at this stage and oxygen requirement is met exclusively by diffusion, the hearts were maintained in close contact with the silicone membrane to minimize diffusional barriers. The membrane slightly flattened the ventricle. The thickness of the ventricle wall being at most 200 μm, the Po2 at the tissue level could be controlled by flushing air (normoxia or reoxygenation, Po2 = 140 Torr) or pure nitrogen (anoxia, Po2 = 0 Torr) through the gas compartment.

The contractions were detected optically through a 10× objective using an adjustable phototransistor positioned over the projected image of the ventricular wall. This detector was connected to an Apple Macintosh LCII via an anolog-to-digital converter. The movement of the ventricular wall, partly intercepting the light, was defined as shortening (S). The amplitude of this movement between diastole and systole was referred to asshortening variation (SV) of the ventricle. The instantaneous contraction velocity (Cv) andrelaxation velocity (Rv) were obtained from the first derivative of the shortening (dS/dt) during each cardiac cycle. In this work, only the maximal values of Cv (mCv) andRv (mRv) were taken into consideration. The instantaneous HR was obtained from the reciprocal of the time interval between two successive systolic peaks. All parameters were recorded using a computerized microphoto-metric setup.

After a period of stabilization of about 40 min under normoxia, the hearts were rapidly (seconds) made anoxic by flushing N2 and reoxygenated by flushing air again in the gas compartment. Characteristic patterns of response to such anoxia-reoxygenation transitions were obtained for anoxia varying from 10 to 60 s. The hearts reproducibly and reversibly responded to variations of Po2 within 5-10 s and each measuring cycle, i.e. air-N2-air sequence, lasted 5 min. A delay of about 10 min was respected between two successive anoxic episodes to allow the cardiac activity to fully recover. The hearts were submitted to anoxia-reoxygenation transitions without renewal of the culture medium, which makes our preparation different from ischemic-reperfused isolated hearts. All hearts were treated according to the same experimental protocol and investigated for about 2 h. The chronotropic and inotropic effects observed in anoxia and at reoxygenation were expressed in percents of preanoxic values. The hearts were divided into three experimental groups according to the glucose concentration tested,i.e. 0, 8 (control group), and 20 mmol/L glucose. Glucose was either absent or present throughout the experiments.

The composition of the culture medium was (in millimoles/L): 100 NaCl, 2 KCl, 0.75 CaCl2·2H2O, 0.78 MgCl2·6H2O, 10.7 Na2HPO4·2H2O, 2.7 KH2PO4. This solution had a pH of 7.4 and a buffering capacity of 6.9 μmol H+/mL·pH unit. The media containing 0, 8, and 20 mmol/L glucose were kept isosmotic (240 mosmol/kg H2O) by adding 20, 12, and 0 mmol/L sucrose, respectively. The addition of iodoacetate (100 μmol/L) in the medium was performed by using remote controlled push-pull syringes at a rate of 450-900 μL/min for 3-4 min.

Values are given as means ± SEM. Statistical analysis was performed using unpaired two-tailed t test or Mann-Whitney and Wilcoxon paired-sample tests.

The investigation conforms with the Guide for the care and use of laboratory animals published by the U.S. National Institutes of Health.

RESULTS

In total, 42 hearts were investigated. At room temperature (about 23°C), independently of the glucose concentration, the isolated hearts contracted spontaneously and regularly at 37 ± 1 beats/min (n= 6). After a period of stabilization of about 40 min, HR measured under normoxia at 37.5°C was 142 ± 5 (n = 16), 158 ± 5(n = 8) and 158 ± 5 (n = 18) beats/min for 0, 8 and 20 mmol/l of glucose, respectively. Whatever the concentration of glucose, the functional parameters such as HR, SV, mCv, and mRv remained stable for 2-3 h under normoxic steady-state conditions. Moreover, despite six successive anoxic episodes, the pattern of response to anoxia-rexygenation transitions was not significantly altered for at least 3 h of the experiment.

Sucrose at 12 or 20 mmol/L did not influence the functional changes observed during normoxia, anoxia, or reoxygenation in the presence or absence of glucose, showing that an increase of osmotic pressure corresponding to 20 mOsmol did not significantly affect the cardiac function. Under normoxic steady-state conditions, the actual wall motion of the contracting unloaded ventricle during one cardiac cycle was 37 ± 3 μm (n = 5) which corresponded to a shortening of the external perimeter of 2.1 ± 0.3%.

Anoxia

Figure 1 represents typical records of ventricular contractions during anoxia-reoxygenation cycle.

Figure 1
figure 1

Glucose-induced arrhythmias during anoxia and at reoxygenation. (A) In absence of glucose, heart beats stopped under anoxia (N2) and progressively recovered after reoxygenation(AIR). During early anoxia the amplitude of contraction decreased, whereas HR increased (lower trace). (B) In presence of glucose, the heart continued to beat but arrhythmias appeared during anoxia when HR reached a peak value (a1-a2). At reoxygenation, heart abruptly ceased to beat (embryonic oxygen paradox). HR was specially unstable during both anoxia and postanoxic recovery (lower trace). Upon resumption of mechanical activity potentiated contractions were observed.(C) Segments a1, a2, and b of the recordingB shown on expanded scale. The first pause in a1 was associated with relaxation and followed by an amplitude of contraction greater than normal (arrow); b shows types of arrhythmia induced by reoxygenation. The preanoxic shortening variation was about 30 μm.

Full size image

Resistance to anoxia. The spontaneous contractile activity under anoxia persisted during 50 ± 3, 80 ± 4, and 105 ± 15 s in the presence of 0, 8, and 20 mmol/L glucose, respectively (n= 6 for each condition; p < 0.05 versus 0 mmol/L glucose).

Chronotropic effects. Whatever the glucose concentration, anoxia induced a transient tachycardia, reaching rapidly a peak value after 23± 2 s (n = 18), and the maximal value of HR was always significantly higher than the preanoxic value (Fig. 1 and Table 1). This initial tachycardia was followed by bradycardia or even cardiac arrest before reoxygenation was allowed (Fig. 1A). In the presence of 8 mmol/L glucose, the HR measured during anoxia was not significantly different from that measured in 20 mmol/L glucose, whereas in the absence of glucose it decreased abruptly after 40 s (Fig. 2A), and all hearts stopped within 60 s. After 20 and 30 s of anoxia, the variability of HR increased, which was apparently due to arrhythmias (see below).

Table 1 Anoxia-induced transient tachycardia
Full size table
Figure 2
figure 2

Effects of anoxia on HR (A), shortening variation (B), maximal velocities of contraction (C), and relaxation (D) of the ventricle wall determined every 10 s during 60 s of anoxia. The values reported are means ± SEM. Circles, absence of glucose; triangles, 8 mmol/L glucose; squares, 20 mmol/L glucose. n = 6 for each concentration. Note that, in absence of glucose, only 4/6 hearts still contracted at 50 s and all hearts stopped at 60 s of anoxia. * = Significantly different from 0 mmol/L glucose;# = significant difference between 8 and 20 mmol/L glucose (p < 0.05; t test).

Full size image

Inotropic effects. In glucose-free medium, SV, mCv, and mRv fell abruptly after 40 s of anoxia (Fig. 2,B-D). In presence of 8 or 20 mmol/L glucose they diminished progressively throughout the period of anoxia. Unexpectedly, after 20 s of anoxia, these parameters were significantly (p < 0.05) more depressed in the presence of 20 mmol/L than in the presence of 8 mmol/L.

During early anoxia, the diastolic relaxation became incomplete, and a contracture developed (Fig. 1). Whatever the glucose concentration, the maximal contracture represented 20 ± 2% (n= 33) of the preanoxic SV and was observed when the HR was maximal.

When HR was maximal, SV represented on average 80 ± 5%, 59± 4%, and 42 ± 7% of the preanoxic values for 0, 8, and 20 mmol/L glucose, respectively (n = 6 for each concentration). These differences were significant from each other (p < 0.002).

Reoxygenation

Embryonic oxygen paradox. All hearts submitted to anoxia for more than 20 s stopped beating after the onset of reoxygenation within less than 10 s (i.e. atrium, ventricle, and conotruncus ceased simultaneously to contract). The duration of such a cardioplegia at reoxygenation increased with the duration of preceding anoxia (Fig. 3). With respect to the control group, a glucose concentration of 20 mmol/L reduced significantly the duration of cardioplegia by 88, 50, and 33%, after 40, 50, and 60 s of anoxia, respectively. Interestingly, for anoxia lasting up to 40 s, the duration of cardiac arrest at reoxygenation in glucose free medium was not significantly different from that observed in 8 mmol/L glucose. However, after 50 s of anoxia, the heart activity became extremely unstable and stopped abruptly (Fig. 1A). This suggests that glucose can protect efficiently the embryonic chick heart against postanoxic cardioplegia but only at a concentration higher than 8 mmol/L.

Figure 3
figure 3

Oxygen paradox in the embryonic heart. The duration of cardiac arrest upon reoxygenation is related to the duration of the preceding anoxia and depends on glucose availability. The concentrations of glucose are indicated in the graph. n = 6 for each concentration. Under anoxia in absence of glucose, 2 of 6 hearts spontaneously stopped after 40 s, and all hearts stopped after 50 s. These hearts were not represented in this graph because their failure was due to anoxia and not to reoxygenation. Means± SEM are reported. * = Significantly different from 8 mM (0.02 <p < 0.05).

Full size image

Stunning and postanoxic recovery. Reoxygenation did not immediately restore the normal myocardial contractility but led to significant alterations of the chronotropic and inotropic parameters (Fig. 4). Such a prolonged and reversible contractile dysfunction (i.e. stunning) was dependent on the duration of preceding anoxia. After 1 min of reoxygenation preceded by 1 min of anoxia, the contractile function of the heart remained depressed by about 50% in the presence of 8 mmol/L glucose, whereas it was depressed by only 25% in presence of 20 mmol/L glucose (Fig. 4,B-D). Paradoxically, compared with 8 mmol/L glucose, deprivation of exogenous glucose, which was expected to worsen the stunning, did not delay the myocardial recovery.

Figure 4
figure 4

Stunning of the embryonic heart observed after 1 min of reoxygenation preceded by anoxia of variable duration. Stunning, which affects both chronotropic (A) and inotropic (B-D) parameters, is related to the duration of the preceding anoxia and is attenuated by 20 mmol/L of glucose. The changes are expressed as percent of the preanoxic values. The values reported are means ± SEM. n = 6 for each concentration. Same symbols as in Figure 2. * = Significantly different from 0 mmol/L glucose; # = significant difference between 8 and 20 mmol/L glucose (p < 0.05; t test).

Full size image

When the ventricle ceased to contract either under anoxia (Fig. 1A) or at the onset of reoxygenation (Fig. 1B), it showed progressive relaxation due to a slight compression exerted by the silicone membrane. The maximal relaxation observed after 60 s of anoxia represented 11 ± 2% (n = 9), 30± 4% (n = 8), and 19 ± 3% (n = 16) of the preanoxic shortening for 0, 8, and 20 mmol/L glucose, respectively. This relaxation was significantly more important in 8 mmol/L than in 0 or 20 mmol/L glucose. The contractile activity, as well as the degree of relaxation, recovered progressively throughout reoxygenation.

Arrhythmias

Glucose-induced arrhythmias in anoxia and at reoxygenation. An irregular myocardial activity appeared during anoxia as well as during reoxygenation, but exclusively in the presence of glucose (Fig. 1). Whatever the glucose concentration, no ectopic ventricular beats were observed, and arrhythmias always originated from atrium. The incidence of arrhythmia during anoxia and after onset of reoxygenation depended on the duration of anoxia and on the glucose concentration (Fig. 5). The incidence observed in 8 mmol/L glucose was higher than that in 20 mmol/L glucose.

Figure 5
figure 5

Incidences of arrhythmias during anoxia (A) or reoxygenation (B) depended on both the duration of anoxia and glucose concentration. Each point represents the mean of 16, 7, and 7 experiments at 0 (circles), 8 (triangles), and 20(squares) mmol/L glucose, respectively. In glucose-free medium, arrhythmias were never observed, whatever the duration of anoxia tested. Note that the incidences were generally higher at 8 than at 20 mmol/L.

Full size image

The arrhythmias induced by anoxia appeared when HR reached its maximal value (i.e. after about 20 s) and consisted systematically in irregular contractions as those shown in Fig. 1,B and C.

The arrhythmias induced upon reoxygenation were of two types: 1) irregular activity punctuated with pauses every two or three contractions characteristic for the initial phase of the postanoxic recovery (Fig. 1C); 2) bursts of contractile activity occurring at a frequency of 5-6/min and separated by quiescent periods of 3-4 s (Figs. 6A and 7). These bursts presented an initial potentiated contraction (reminiscent of rest potentiation) followed by a positive staircase phenomenon (Fig. 6B). Moreover, immediately after each potentiated contraction, HR was elevated and then progressively decreased (Fig. 6A, lower trace). These types of bursts could be induced after periods of anoxia longer than 60 s.

Figure 6
figure 6

Pattern of arrhythmia observed upon reoxygenation in presence of glucose. (A) Reoxygenation-induced bursts of contractions and corresponding oscillations of instantaneous HR (lower trace). (B) The region framed in A is shown on an expanded scale and is representative of the bursts observed at reoxygenation. The overshoot (arrows) was followed by an ascending staircase phenomenon associated with a rapid decrease of HR.

Full size image
Figure 7
figure 7

Block of glycolysis suppressed arrhythmias. Alterations of shortening during normoxia-anoxia-normoxia transitions were recorded successively on the same heart. Addition of iodoacetate (100 μmol/L) reduced resistance to anoxia but abolished bursts of contractions during reoxygenation within about 30 min.

Full size image

Simultaneous recordings of contractions along the heart tube and analysis of video recordings have shown that the changes in HR as well as the arrhythmias occurring during anoxia or at reoxygenation originated in the atrium. Interestingly, whatever the concentration of glucose, the ventricle stopped beating under anoxia before the atrium, whereas atrial activity resumed first during reoxygenation(16). These observations indicate that normoxia-anoxia-reoxygenation transitions could induce conduction disturbances and/or that a differential sensitivity to oxygen availability could exist along the heart tube (this is being currently investigated in our laboratory).

Blockade of glycolysis suppressed arrhythmias. The inhibition of glycolysis with iodoacetate (100 μmol/L) in presence of glucose suppressed the burts of myocardial activity at reoxygenation, whereas it reduced the resistance to anoxia in the same way as in glucose-free medium (Fig. 7).

DISCUSSION

Method and protocol. Our preparation appears to be an adequate experimental model for studying the reactivity of embryonic heart to alterations of oxygen and substrate availability. Indeed, 1) myocardial oxygenation and the composition of the medium are strictly controlled, 2) spontaneous contractile activity remains stable during at least 3 h despite repeated anoxic episodes, 3) spontaneous HR is comparable to that measured in situ, and 4) responses to Po2 changes are rapid, reproducible, and reversible. Moreover, the renewal of the medium only every six successive measuring cycles did not alter the pattern of response to anoxia-reoxygenation transitions throughout the experiment. The fact that the response of the heart was reproducible despite repeated episodes of anoxia-reoxygenation indicate that this preparation is not a model of preconditionning. Taking into account the metabolic activity of the myocardium(21) and the buffering capacity of the medium, the change of external pH after 2 h under stop-flow conditions would be 0.01 at most and should not interfere with cardiac activity.

The instantaneous profile of Po2 within the embryonic myocardium during anoxia-reoxygenation transitions has been evaluated using a finite element computation of the oxygen concentration according to the Fick's equation of diffusion(21). The results were that the deepest cell layers would be totally deoxygenated within 14 s of anoxia and start to be reoxygenated 6 s after readmission of air in the chamber. These time lags, although slightly overestimated by the model, are of the same order as the delays of response to anoxia and reoxygenation observed experimentally.

It should be mentioned that reoxygenation of the embryonic heart was independent of reperfusion and was not accompanied by a washout of metabolites accumulated during anoxia. Therefore, the reoxygenation-induced dysfunction observed in the embryonic heart should be considered as an oxygen-dependent phenomenon rather than a flow-dependent phenomenon.

The reason to perform experiments in the presence of 8 and 20 mmol/L glucose was that these concentrations correspond to those found in whole blood and egg yolk and albumen, respectively [estimated from Romanoff(22)].

Responses to anoxia and reoxygenation. Three major observations concerning glucose emerge from this study: 1) glucose delays the anoxic contractile failure, shortens the reoxygenation-induced cardiac arrest, and improves the recovery of contractile activity; 2) glucose attenuates stunning at 20 mmol/L, but worsens it at 8 mmol/L; and 3) paradoxically, glucose (especially at 8 mmol/L) is arrhythmogenic during anoxia and reoxygenation.

Anoxia. Contractile dysfunction during the first minutes of anoxia in the presence of glucose has been attributed to accumulation of metabolites rather than ATP depletion(23–25). In the embryonic chick heart, the fact that the chronotropic and inotropic parameters were significantly altered, within the first 10-20 s of anoxia, indicates that involvement of ATP depletion is unlikely. Anaerobic glycolysis produce protons which are known to exert a negative inotropic effect. The fact that the contractile depression was more marked at 20 mmol/L than at 8 mmol/L glucose suggests that a high glucose concentration could stimulate glycolytic flux, resulting in a further intracellular acidification. The latter could activate Na+/H+ exchange and, secondarily, Na+/Ca2+ exchange, resulting in an arrhythmogenic calcium overload. It should be stressed that the excitation-contraction coupling in the embryonic myocardium seems to depend on extracellular and cytosolic calcium rather than calcium release from the relatively immature sarcoplasmic reticulum.

Inversely, removal of glucose suppressed arrhythmias but resulted in contractile failure within 50 s of anoxia, most likely due to a rapid ATP depletion resulting from the lack of substrate.

Stunning and recovery during reoxygenation. It is generally accepted that the dysfunction of the ischemic-reperfused adult heart is associated with disturbances of ionic homeostasis(26, 27), pH regulation(28, 29), energy metabolism(30), and production of oxygen free radicals(31, 32).

It appears that such mechanisms are implicated also in the anoxic-reoxygenated embryonic chick heart. Indeed, alteration of the calcium homeostasis results in periodic or chaotic electrical and mechanical activities(33–35) resembling that observed at reoxygenation. External acidosis (pH 6.5) depresses contractile function, prolonges reoxygenation-induced cardioplegia, and provokes arrhythmias(36), whereas calcium antagonists (e.g. Verapamil) and antioxidants (superoxide dismutase/catalase, allopurinol) are protective(17). However, the main differences with respect to the adult heart are the following: 1) glucose is arrhythmogenic specially when present at the physiologic concentration of 8 mmol/L and 2) reoxygenation of the embryonic heart suppresses transiently the mechanical activity and does not result in a contracture, irrespective of the glucose concentration.

A high glucose concentration (20 mmol/L) improved significantly the recovery of the contractile function and decreased the incidence of arrhythmias with respect to control group. These beneficial effects could result from increased rates of 1) glucose transport, 2) glucose oxidation supplying ATP, and 3) production of reducing equivalents required for antioxidant systems and derived from the pentose phosphate pathway.

Arrhythmogenesis. The arrhythmias observed under our conditions originated exclusively in the atrial region and were related to tachycardia, and their incidences were dependent on the duration of anoxia. Interestingly, the removal of glucose or blockade of glycolysis not only decreased the anoxia-induced acceleration of the heartbeat, but also suppressed totally arrhythmias during anoxia and reoxygenation. This is in contrast with observations in the adult heart where an increase of glucose concentration protects automaticity and conduction during hypoxia(1) and reduces incidences of arrhythmias(37, 38) during reperfusion.

As far as the relationship between glucose availability and rhythm disturbances is concerned, the role of fluctuations of metabolic activity in the cardiac dysfunction has to be taken into consideration especially when glucose is the preferential substrate. Indeed, spontaneous glycolytic oscillations have been shown to be potentially arrhythmogenic and triggered exclusively within a critical range of glucose concentration and glycolytic rate(39). Interestingly, in the embryonic myocardium, incidences of arrhythmias during anoxia and reoxygenation were the highest at a glucose concentration of 8 mmol/L, whereas they decreased at 20 mmol/L and were zero in glucose-free medium. These results support the hypothesis that arrhythmias in the early developing heart appear also within a critical range of glucose concentration. The activity of ion pumps (e.g. Na+,K+- and Ca2+-ATPases) and some ion channels(e.g. ATP-sensitive K+ channels) is thought to be regulated by ATP-generated preferentially by glycolysis(40) and influenced by metabolites such as lactate and protons. Therefore, variations of glycolytic flux induced by deprivation and readmission of oxygen should result in potentially arrhythmogenic fluctuations of membrane potential.

The fact that arrhythmias in the embryonic chick heart were suppressed by removal of glucose or addition of iodoacetate strengthens the hypothesis that arrhythmogenesis might depend on accumulation of glycolytic metabolites. This is also supported by the observation that the nonrenewal of the medium during anoxia-reoxygenation cycles repeated at short intervals worsened cardiac dysfunction in the presence of glucose, whereas it did not induce arrhythmias in glucose-free medium.

In glucose-free medium, the production of glucose from the large stores of glycogen in the embryonic chick heart(14) was not sufficient to sustain activity during anoxia and to induce arrythmias during anoxia-reoxygenation transitions. This findings suggests that glycogenolysis provides glucose at a slower rate than transport of exogenous glucose. In the fetal heart as well, it has been observed that anaerobic glycolysis is primarily dependent on exogenous glucose rather than on glycogen(7). Moreover, in cardiomyocytes, glycogenolysis and glycolysis seem to be compartmentalized(41), and ion homeostasis appears to depend greatly on glycolytic flux from exogenous glucose. Thus, alteration of glucose availability could particularly affect membrane excitability and provoke arrhythmias.

Conclusion. The effects of glucose as the sole exogenous substrate on the activity of the anoxic-reoxygenated embryonic chick heart were determined for the first time under strictly controlled conditions. The results show similarities but also differences with respect to the adult heart. Indeed, glucose prolonged contractile function under anoxia and improved recovery upon reoxygenation but, paradoxically, induced rhythm disturbances. The latter phenomenon seems to be characteristic of the young embryonic heart. Our findings underscore the important role that fluctuations of glycolytic activity may play in the reactivity of the embryonic myocardium to anoxia-reoxygenation transitions.