Lack of Direct Role for Calcium in Ischemic Diastolic Dysfunction in Isolated Hearts
Background—Ischemia is characterized by an increase in intracellular calcium and occurrence of diastolic dysfunction. We investigated whether the myocyte calcium level is an important direct determinant of ischemic diastolic dysfunction.
Methods and Results—We exposed isolated, perfused isovolumic (balloon in left ventricle) rat and rabbit hearts to low-flow ischemia and increased extracellular calcium (from 1.5 to 16 mmol/L) for brief periods. Intracellular calcium was measured by aequorin. Low-flow ischemia resulted in a 270% increase (P<0.05) in diastolic intracellular calcium, a 50% (P<0.05) calcium transient amplitude decrease, and a 52% (P<0.05) slowing of calcium transient decline. Diastolic pressure increased by 6±2 mm Hg (P<0.05), and rate of systolic pressure decay decreased by 65% (P<0.05). Experimentally increasing extracellular calcium doubled both intracellular diastolic calcium and calcium transient amplitude, concomitant with a developed pressure increase; however, there was no increase in ischemic diastolic pressure, slowing of the calcium transient decay, or further slowing of systolic pressure decay. Similarly, after 45 minutes of low-flow ischemia, after diastolic pressure had increased from 8.5±0.6 to 19.7±3.5 mm Hg (P<0.001), intracoronary high-molar calcium chloride infusion increased systolic pressure from 36±4 to 63±11 mm Hg (P<0.001), indicating an increase in intracellular calcium, but it decreased diastolic pressure from 19.7±3.5 to 17.5±3.7 mm Hg (P<0.01). Conversely, EGTA infusion decreased systolic pressure, indicating a decrease in intracellular calcium, but did not decrease diastolic pressure.
Conclusions—When calcium availability was experimentally altered during ischemia, there was no alteration in left ventricular diastolic pressure, suggesting that ischemic diastolic dysfunction is not directly mediated by a calcium activated tension.
Diastolic dysfunction occurs during ischemia; myocardial relaxation becomes impaired and diastolic chamber distensibility decreases progressively so that LV diastolic pressure is increased relative to diastolic volume. Such a reversible increase in diastolic chamber stiffness has been observed in patients with pacing- or exercise-induced angina and in several experimental models of myocardial ischemia and hypoxia.1
Postulated responsible mechanisms include an increase in intracellular free calcium and consequent elevated diastolic myofilament force production by persistent calcium-activated cross-bridge cycling in diastole. Alternatively, cross-bridges may lock in the rigor state as a result of a decrease in tissue ATP and creatine phosphate (CP) levels or increased ADP levels.2 3 4 The relative role of these 2 mechanisms (calcium-activated tension versus rigor bond formation) in causing ischemic diastolic dysfunction remains controversial. To help resolve this question, we created a state of ischemic diastolic dysfunction in isolated hearts and altered the intracellular calcium level with a brief perturbation, reasoning that ischemic diastolic chamber stiffness, if driven by abnormal calcium-activated tension, should be increased by an increase in [Ca2+]i or decreased by a decrease in [Ca2+]i.
The set of experiments of low-flow ischemia with concomitant measurement of [Ca2+]i by the aequorin technique (perfusion condition and protocol as below) was performed in the isolated, buffer-perfused rat heart. Male Wistar rats were anesthetized with diethylether, and hearts were quickly excised and retrogradely perfused with an oxygenated Krebs-Henseleit solution containing (mmol/L) NaCl 118, KCl 4.7, KH2PO4 1.2, NaHCO3 23, MgCl2 1.2, CaCl2 1.5, and dextrose 5.2.
The experiments to examine the effects of calcium perturbations during prolonged low-flow ischemia (perfusion condition and protocols as below) were performed in an isolated, isovolumic, red cell–perfused rabbit heart preparation as described previously in detail.5 Hearts of male New Zealand White rabbits were perfused with a red cell perfusate consisting of bovine red blood cells (at a final hematocrit of 40%) in Krebs-Henseleit buffer with 1.0 mmol/L lactate, 0.4 mmol/L palmitic acid, and 40 g/L BSA added. Final ionized calcium concentration was measured by a calcium electrode (NOVA 6, NOVA Biomedical Corp) and adjusted to 1.35 meq/L.
For both buffer-perfused rat hearts and red cell–perfused rabbit heart experiments, coronary perfusion pressure (CPP) was monitored via a sidearm of the aortic cannula connected to a pressure transducer (Gould-Statham P23dB, Gould Inc) and recorded continuously on a multichannel physiological recorder (Gould Inc). A collapsed, thin-walled latex balloon was placed in the left ventricle (LV) via the left atrium. The balloon was connected to a pressure transducer to measure LV pressure and its first derivative.
Lactate concentrations in arterial and venous samples were measured with a specific enzymatic technique. At specified time points, rabbit hearts were freeze-clamped, and tissue levels of ATP, CP, and lactate were determined spectrophotometrically.5
All animal handling and procedures strictly adhered to the regulations of Boston University Animal Care and the National Society for Medical Research.
Protocol 1: Aequorin Experiments in Isolated Rat Hearts During Short Periods of Low-Flow Ischemia
Aequorin was macroinjected with a technique previously described for the ferret whole-heart preparation and slightly modified for the intact rat heart.6 7 Briefly, 3 to 5 μL of an aequorin-containing solution (1 μg/mL) was injected with a glass micropipette into the interstitium of the inferoapical region of the LV. The heart was positioned in an organ bath with the aequorin-loaded area of the LV directed toward the cathode of a photomultiplier (model 9635QA, Thorn-EMI, Gencom, Inc) and submerged in Krebs-Henseleit solution. The organ bath was enclosed in a light-occlusive photographic bellows designed for studies with aequorin-loaded muscles.6 The temperature was increased from 25°C to 37°C within 15 minutes, and the hearts were finally paced at 5 Hz.
Measurement of Intracellular Calcium
Aequorin light signals were recorded on a 4-channel recorder. At each time point of interest, 20 to 40 light transients were wave averaged. To compare light values of different experiments and to compensate for aequorin consumption, the method of fractional luminescence was used as described previously.7 8 Briefly, at the end of each experiment, the heart was perfused with a solution containing 20 mmol/L calcium and 5% Triton X-100 to lyse the aequorin-loaded cells and expose all the remaining aequorin to calcium. This resulted in an instantaneous burst of light, subsequently declining to baseline within 10 to 20 minutes. The area under the curve was integrated to obtain a value for the total amount of light (Lmax) emitted from the aequorin loaded into the myocytes as described by Kihara et al.6 To obtain Lmax(t) for an individual time point, the integral of the aequorin signal from the respective time point up to the end of the experiment was added to Lmax. Light values of a specific time point were expressed as L/Lmax(t).
Measurement of Time Constant of LV Pressure Decay and Constant of Decrease of Intracellular Calcium
In addition to the chart-strip recording, the LV pressure tracings and light signals were digitized by a 12-bit analog-to-digital converting board at a sampling rate of 1 kHz (DAP 800/3, Microstar). With the use of custom software, the time constants of exponential pressure decay (τp) and of light decrease (τca) were calculated through the variable asymptote method.9
After Aequorin loading and after steady-state conditions were reached, the intracardiac balloon volume of the rat hearts (n=12) was set to 50% of Volmax (balloon volume at peak developed pressure).10 This balloon volume was kept constant for the remainder of the experiment. Under this condition, an increase in LV isovolumic diastolic pressure signifies a decrease in diastolic chamber distensibility.5 After a stabilization period of 20 minutes, coronary flow was reduced to 5% of the initial value. After 15 minutes, when ischemic contracture, as defined by an increase in isovolumic diastolic pressure of >5 mm Hg, had occurred, the perfusate was switched to a buffer containing 16 mmol/L Ca2+ in form of CaCl2 fully equilibrated with identical measured pH. The coronary flow rate remained constant.
Protocol 2: Effect of Short Periods of Calcium Overload and EGTA During Prolonged Ischemia Protocols in Rabbit Hearts
Before the protocols described below, all rabbit hearts were perfused at a constant CPP of 100 mm Hg for 30 minutes. LV balloon volume was adjusted so that LV isovolumic diastolic pressure was 10 mm Hg. Hearts were then exposed to 60 minutes of low-flow ischemia at a constant CPP of 8 mm Hg and reperfused at initial perfusion conditions, ie, at a constant CPP of 100 mm Hg for 30 minutes. After 5, 15, 30, and 45 minutes of low-flow ischemia, a control group (n=6) was subjected to a bolus of 0.08 mL of 0.9% NaCl, a calcium group (calcium constant CPP; n=6) to a bolus of 0.08 mL of the 260 mmol/L CaCl2 solution, and a EGTA group (n=6) to 0.08 mL of a 10−2 mol/L EGTA solution. Calcium infusion had a small initial vasodilatory effect, followed by a small vasoconstrictor effect after about 1 minute. Because this change in coronary blood flow (CBF), even though minimal, could have affected isovolumic diastolic pressure, an additional group of hearts (calcium constant CBF; n=7) was exposed to the bolus of the high-molar calcium chloride solution, with the system switched from a constant CPP mode to a constant CBF mode during the intervention and for 5 minutes afterward.
Data are reported as mean±SEM. Data acquired by repeated sequential measurements in individual hearts were tested by ANOVA for repeated measures. Comparison between ≥2 experimental groups was performed by 2-way ANOVA. Post hoc analysis was performed by the method of least significant differences. A difference of a single metabolic measurement between experimental groups within 1 protocol was tested by an unpaired t test. A value of P<0.05 was considered significant.
Protocol 1: Aequorin Experiments
During low-flow ischemia, LV developed pressure decreased to 9% of baseline (Table 1⇓ and Figure 1⇓). Isovolumic diastolic pressure increased from 4.0±0.7 to 10.1±1.9 mm Hg (P<0.05) after 15 minutes of ischemia. Typical tracings are shown in Figure 1A⇓ and 1B⇓. During ischemia, diastolic [Ca2+]i increased, as assessed by an increase of the aequorin light signal. As expected, increasing [Ca2+]o from 1.5 to 16 mmol/L increased diastolic [Ca2+]i, the calcium transient amplitude, and LV developed pressure. However, diastolic pressure did not increase; on the contrary, it decreased slightly, from 10.1±1.9 to 9.3±1.8 mm Hg (P<0.05). The time constant of light decrease, τca, increased during ischemia but did not increase significantly further during calcium infusion. The time constant of pressure decay, τp, increased during ischemia but decreased after calcium infusion.
Protocol 2: Effect of Calcium Perturbation During Prolonged Low-Flow Ischemia
Coronary flows were identical (1.2 to 1.3 mL · g−1 · min−1) during preischemia in the 4 groups that underwent the prolonged ischemia-reperfusion protocol. During ischemia, CBF fell comparably in all groups, initially to 15% to 23% and subsequently to 9% to 12% of baseline. During reperfusion, an initial hyperemic CBF subsequently returned to baseline values in all groups. Developed pressure (ie, systolic minus diastolic pressure) was similar in all groups at baseline, low-flow ischemia, and reperfusion (Figure 2⇓).
Effects of Calcium Overload
At 5 minutes of ischemia, diastolic pressure had not yet increased. Increasing [Ca2+]o, and thereby increasing [Ca2+]i, had no effect on diastolic pressure, although it dramatically increased systolic function (Figure 2⇑). At 45 minutes of ischemia, diastolic pressure had increased. A typical result of a calcium infusion at this time point is given in Figure 3⇓. There was a marked positive inotropic effect, indicating an increase in [Ca2+]i; however, similar to the aequorin experiments, LV diastolic pressure decreased slightly.
LV pressure decay, as assessed by maximal negative dP/dt, was not slowed but rather increased after experimental calcium overload at all times during ischemia (Figure 4⇓). This can be explained by the afterload dependency of maximal negative dP/dt.11 Therefore, we normalized maximal negative dP/dt per developed pressure to estimate the rate of pressure decay relative to the increase in afterload. At 45 minutes of ischemia, maximal negative dP/dt per developed pressure in the calcium constant CPP group was 12.1±1.0 seconds−1 and stayed at 12.4±1.4 seconds−1 (P=NS) after experimental calcium overload. In the calcium constant CBF group, this value increased from 7.4±1.1 to 11.5±1.3 seconds−1 (P<0.001) after the imposed calcium overload, indicating an increased rate of LV pressure decay, consistent with the observed increase in τp after the calcium bolus in the aequorin experiments.
Effects of EGTA
A typical EGTA infusion at 45 minutes of ischemia is shown in Figure 3⇑. During ischemia, EGTA decreased systolic pressure, indicating a decrease in [Ca2+]i, but did not decrease LV isovolumic diastolic pressure (Figure 2⇑). EGTA did decrease maximal negative dP/dt. However, maximal negative dP/dt per developed pressure was unchanged, eg, 13.0±0.2 seconds−1 before EGTA and 13.0±1.2 seconds−1 (P=NS) after EGTA infusion at 45 minutes of ischemia.
Metabolic Effects of Ischemia-Reperfusion
During the preischemic period, all hearts used lactate as substrate and thus showed net lactate uptake or “negative” production. Low-flow ischemia resulted in a substantial lactate production in all groups in protocol 2. However, total lactate production, ie, combined lactate washout during ischemia and during reperfusion, was the smallest in the EGTA group (Table 2⇓).
In additional hearts undergoing the same low-flow ischemia protocol, ATP and CP tissue levels were measured after 35 minutes of ischemia. ATP and CP tissue levels in the control and calcium groups were significantly lower than after 30 minutes of normoxic perfusion (Table 2⇑). Therefore, at a time when diastolic chamber stiffness had slightly increased, high-energy phosphates were decreased 30% to 50%.
At the end of reperfusion in protocol 2, tissue levels of high-energy phosphates were similarly depleted by ≈50% compared with preischemic values in all 4 groups (Table 2⇑).
We investigated whether the myocyte calcium level is an important direct determinant of the diastolic chamber stiffness increase that occurs during ischemia. During low-flow ischemia, both isovolumic diastolic pressure and [Ca2+]i increased, but deliberate perturbation of [Ca2+]i revealed that it was not causally related to the increase in diastolic chamber stiffness. An imposed brief calcium load during ischemia further increased [Ca2+]i but did not increase diastolic pressure; conversely, decreasing [Ca2+]i with EGTA did not decrease diastolic chamber stiffness. These results were consistent in 2 species during brief periods of ischemia at a time when an increase in diastolic chamber stiffness was just beginning and after prolonged ischemia. They suggest that increased [Ca2+]i is not the direct cause of increased ischemic diastolic chamber stiffness, because the level of diastolic fiber tension was not altered by changes in [Ca2+]i, as it would be if it were calcium driven.
Similarly, during low-flow ischemia, the times required for calcium transient and LV pressure decay were prolonged; additional experimental calcium overload did not further prolong these decays, suggesting that their ischemic prolongation was not due to cytosolic calcium overload, because deliberate experimental worsening of the calcium overload did not lead to further prolongation of either pressure or calcium transient decay.
However, active tension, as reflected by LV developed pressure, was sensitive to the imposed alterations of [Ca2+]i during ischemia, reflecting its calcium-activated basis. Thus, during low-flow ischemia, systolic function changed in parallel with experimental perturbations of [Ca2+]i, but indexes of ischemic diastolic dysfunction were dissociated from changes in [Ca2+]i.
The development of rigor in a subendocardial population of myocytes would explain all the observations of our study. The myocytes in rigor would increase diastolic pressure in proportion to their number but would not be responsive to alteration of perfusate calcium levels; experiments in isolated myocytes have consistently demonstrated that myocytes in rigor are inexcitable and incapable of developing contractile force.12 13 The myocytes not in rigor are contracting actively and generating phasic LV pressures and are responsive to perturbations of calcium availability, and those in the epicardium are generating an observable aequorin calcium transient. Apparently, these ischemic but contracting myocytes do not contribute a calcium-activated tension to LV diastolic pressure, because there was no diastolic pressure increase in response to a calcium infusion or decrease in response to an EGTA infusion. The dissociation during low-flow ischemia between changes in the time constant of calcium decline and the time constant of the rate of pressure decline, as well as between the diastolic myocyte calcium levels and isovolumic LV end-diastolic pressure, reflects the finding that the epicardial myocytes from which calcium is being measured are not representative of the changes in calcium and function of all the regions of myocardium.
Role of ATP Depletion
A subpopulation of cells in rigor would also explain our ATP results. Studies in intact tissue have shown that rigor develops when ATP levels fall to ≈30% of control levels.14 Our observed decrease in the average myocardial ATP level of 30% to 50% during ischemia, concomitant with an increase in LV end-diastolic pressure, is consistent with rigor formation in the subendocardium, where the decrease in [ATP] would be greater than the average myocardial decrease.15 16 Metabolic factors might contribute to rigor development at such modest decreases in ATP. Metabolites of the creatine kinase reaction, such as a decrease in CP and an increase in ADP, can affect rigor tension development and stiffness.3 4 Additionally, moderate decreases in [ATP] can impair relaxation by mechanisms in addition to rigor. ATP in the millimolar range has a plasticizing effect to facilitate myofilament relaxation, and its loss can contribute to diastolic dysfunction.17
Although we did not directly measure transmural gradients of tension development, calcium transients, or high-energy phosphates in these experiments, a number of other studies from our laboratory and others support this explanation of the present results. Several studies have demonstrated that subendocardial ischemia is more severe than subepicardial ischemia for a given decrease in CPP.15 16 Similarly, recent 31P-NMR studies of isolated hearts subjected to global low-flow ischemia have identified 2 regions of different degrees of acidosis, whereby the extent of increase in LV end-diastolic pressure correlated closely with the size of the more severely acidotic (presumably subendocardial) region.18
Ultrastructural studies of isolated rabbit hearts subjected to global low-flow ischemia have shown a marked degree of heterogeneity, with some myocytes in rigor juxtaposed to adjacent cells with near-normal ultrastructure.19 Studies of isolated cardiomyocytes subjected to metabolic inhibition with cyanide have demonstrated a variable time to the onset of rigor, suggesting intrinsic cell-to-cell heterogeneity in sensitivity to ischemic contracture.20 Thus, all these studies support the concept of heterogeneity of ischemia in this isolated heart model.
Role of Calcium
An accumulation of calcium during ischemia or hypoxia has been demonstrated and is thought to result from a combination of increased calcium influx, a decrease in calcium efflux, and a decreased calcium reuptake by the sarcoplasmic reticulum during diastole. An increase in [Ca2+]i has been held responsible by many,6 including us,1 5 as the cause of the reversible increase in diastolic chamber stiffness during angina and in experimental ischemia. Increases in [Ca2+]i during hypoxia and zero-flow and low-flow ischemia in the whole heart have supported this hypothesis.14 21 22 Similarly, in isolated myocytes, a rise in [Ca2+]i preceded the onset of contracture during hypoxia or metabolic inhibition of oxidative phosphorylation.12 23 None of the techniques to measure [Ca2+]i directly measure [Ca2+]i bound to troponin (or available for binding to troponin), and it is the troponin-bound calcium that is responsible for calcium-activated tension. Therefore, none of these studies established a definitive cause-effect relationship between ischemia-induced changes in diastolic chamber stiffness and concomitant changes in [Ca2+]i, and several reports showed a clear dissociation between an increase in [Ca2+]i and the onset of hypoxic or ischemic contracture.24 25 26
Our aequorin experiments are consistent with previous observations showing an increase in [Ca2+]i during ischemia.6 22 However, after the onset of ischemic diastolic chamber stiffness, when we deliberately perturbed [Ca2+]i to be able to relate changes in diastolic chamber stiffness to changes in [Ca2+]i, we did not substantiate a cause-effect relationship between the increased [Ca2+]i and the increases in ischemic diastolic chamber stiffness.
A feature of ischemic diastolic dysfunction is prolonged LV pressure decay. In our experiments, the calcium transient decay was prolonged during ischemia in parallel with LV pressure decay, as previously observed.22 However, when we experimentally further increased [Ca2+]i τca was not further prolonged, and τp shortened. There are several explanations for this dissociation. First, as discussed above, the aequorin signal reports the calcium transient from epicardial myocytes, but the LV pressure wave is determined from the entire myocardium. Second, the decay of τca is a function of [Ca2+]i and calcium transient amplitude; a decrease in calcium transient amplitude during low-flow ischemia would be expected to result in an increase in τca and does not necessarily reflect impaired cytosolic calcium removal processes.27
This work was supported by grants (HL-38198 and HL-48715) from the National Heart, Lung, and Blood Institute (Dr Apstein) and a grant from the Deutsche Forschungsgemeinschaft [S.F.B.; 355 Pathophysiology of Heart Failure, project A3 (Dr Neubauer)]. Dr Eberli was the recipient of a grant for young investigators from the Swiss National Science Foundation.
Reprint requests to Franz R. Eberli, MD, Swiss Cardiovascular Center Bern, University Hospital, 3010 Bern, Switzerland.
- Received November 19, 1999.
- Revision received June 22, 2000.
- Accepted June 22, 2000.
- Copyright © 2000 by American Heart Association
Ventura-Clapier R, Vksler V. Myocardial ischemic contracture: metabolites affect rigor tension development and stiffness. Circ Res. 1994;74:920–929.
Eberli FR, Weinberg EO, Grice WN, et al. Protective effect of increased glycolytic substrate against systolic and diastolic dysfunction and increased coronary resistance from prolonged global underperfusion and reperfusion in isolated rabbit hearts perfused with erythrocyte suspensions. Circ Res. 1991;68:466–481.
Kihara Y, Grossman W, Morgan JP. Direct measurement of changes in intracellular calcium transients during hypoxia, ischemia, and reperfusion of the intact mammalian heart. Circ Res. 1989;65:1029–1044.
Meissner A, Morgan JP. Contractile dysfunction and abnormal Ca2+ modulation during postischemic reperfusion in rat heart. Am J Physiol. 1995;268:H100–H111.
Raff GL, Glantz SA. Volume loading slows left ventricular isovolumic relaxation rate: evidence of load-dependent relaxation in the intact dog heart. Circ Res. 1981;48:813–824.
Stromer H, Cittadini A, Szymanska G, et al. Validation of different methods to compare isovolumic cardiac function in isolated hearts of varying sizes. Am J Physiol. 1997;272:H501–H510.
Ikenouchi H, Zhao L, Barry WH. Effect of 2,3-butanedione monoxime on myocyte resting force during prolonged metabolic inhibition. Am J Physiol. 1994;267:H419–H430.
Steenbergen C, Murphy E, Watts JA, et al. Correlation between cytosolic free calcium, contracture, ATP, and irreversible ischemic injury in perfused rat heart. Circ Res. 1990;66:135–146.
Momomura S, Ingwall JS, Parker JA, et al. The relationships of high energy phosphates, tissue pH, and regional blood flow to diastolic distensibility in the ischemic dog myocardium. Circ Res. 1985;57:822–835.
Zhang J, Ishibashi Y, Zhang Y, et al. Myocardial bioenergetics during acute hibernation. Am J Physiol. 1997;273:H1452–H1463.
Katz A. Physiology of the Heart. New York, NY: Raven Press; 1992.
Saupe KW, Lim C, Ingwall JS, et al. Comparison of hearts with two types of pressure-overload left ventricular hypertrophy. Hypertension. 2000;35:1167–1172.
Apstein CS, Gravino FN, Haudenschild CC. Determinants of a protective effect of glucose and insulin on the ischemic myocardium. Circ Res. 1983;52:515–526.
Kondo RP, Apstein CS, Eberli FR, et al. Increased calcium loading and inotropy without greater cell death in hypoxic rat cardiomyocytes. Am J Physiol. 1998;275:H2272–H2282.
Mohabir R, Lee HC, Kurz RW, et al. Effects of ischemia and hypercarbic acidosis on myocyte calcium transients, contraction, and pHi in perfused rabbit hearts. Circ Res. 1991;69:1525–1537.
Camacho SA, Brandes R, Figueredo VM, et al. Ca2+ transient decline and myocardial relaxation are slowed during low flow ischemia in rat hearts. J Clin Invest. 1994;93:951–957.
Barry WH, Peeters GA, Rasmussen CA Jr, et al. Role of changes in [Ca2+]i in energy deprivation contracture. Circ Res. 1987;61:726–734.
Allshire A, Piper HM, Cuthbertson KS, et al. Cytosolic free Ca2+ in single rat heart cells during anoxia and reoxygenation. Biochem J. 1987;244:381–385.
Bers DM, Berlin JR. Kinetics of [Ca]i decline in cardiac myocytes depend on peak [Ca]i. Am J Physiol. 1995;268:C271–C277.