Spatial Heterogeneity of Calcium Transient Alternans During the Early Phase of Myocardial Ischemia in the Blood-Perfused Rabbit Heart
Background Optical mapping of cytosolic calcium transients in intact mammalian hearts is now possible using long-wavelength [Ca2+]i indicators. We propose that beat-to-beat [Ca2+]i transient alternans during ischemia may lead to spatial and temporal heterogeneity of calcium-activated membrane currents.
Methods and Results To test this hypothesis, isolated rabbit hearts were loaded with the fluorescent [Ca2+]i indicator, rhod-2 AM, and imaged at 300 frames/sec during blood-perfused ischemic trials. High-quality [Ca2+]i transients were recorded in each of 8 hearts.[Ca2+]i transient alternans was never present in control records but occurred in each of the hearts during ischemia, with onset after 2 to 4 minutes. Alternans was confined to circumscribed regions of the heart surface 5 to 15 mm across. Multiple regions of alternans were found in most hearts, and regions that were out of phase with one another were found in 6 hearts. Quantitative maps of alternans were constructed by calculating an alternans ratio. This ratio behaved as a continuous variable that reached a maximum value in the center of the regions with alternans.
Conclusions These results demonstrate marked spatial heterogeneity of the [Ca2+]i transient during the early phase of ischemia, which could produce electrical instability and arrhythmias in large mammalian hearts.
Received May 25, 2001; revision received July 18, 2001; accepted July 19,2001.
Abnormalities in [Ca2+]i occur in the first few minutes of acute myocardial ischemia and may be an important cause of arrhythmias that affect humans with coronary disease. Most studies have shown an overall increase in [C2+]i in the first 5 minutes of ischemia.1–4 More pronounced, however, are fluctuations in the systolic and diastolic level of [Ca2+]i or calcium transient alternans.4,5 Calcium transient alternans during ischemia leads to fluctuations in contractile force and may also produce fluctuations in [Ca2+]i-dependent membrane currents that affect action potential duration (APD).6,7
When [Ca2+]i transient alternans is observed in whole hearts, it is unclear whether all of the cells contributing to the recording have [Ca2+]i transients that are fluctuating in the same way or to the same extent. Occurrence of [Ca2+]i transient alternans in circumscribed regions of the heart could be an important factor in the genesis of arrhythmias, because it could create regional variation of APD and refractoriness. Beat-to-beat changes in APD are believed to be responsible for T-wave alternans, which is an immediate harbinger of ventricular fibrillation (VF) in large mammalian hearts.
Better imaging techniques, including longer wavelength calcium indicators, now permit these issues to be studied in blood-perfused hearts, which show more alternans than saline-perfused hearts.5 One such indicator, rhod-2, has spectral properties that allow it to be used with optical mapping systems designed for potentiometric dyes.8,9 Rhod-2 has been used previously to study regional differences in the [Ca2+]i increase during simulated ischemia in brain slices10 and to study the relation between action potential and calcium transient alternans in rabbit hearts during rate jumps.9 The purpose of this study is to characterize regional heterogeneities of [C2+]i transient alternans in blood-perfused rabbit hearts during ischemia. Although stop-flow ischemia does not usually produce fibrillation in small hearts, the experimental conditions are otherwise similar to what occurs in large animals or humans.
New Zealand rabbits of either gender (Kralek Farms, Turlock, Calif) weighing 2 to 3 kg were anesthetized with pentobarbital (50 mg/kg). The heart was quickly removed and was mounted on a Langendorff apparatus and perfused with oxygenated Tyrode’s solution4 containing probenecid5 at a rate of 20 mL/min. The perfusate was heated to maintain the temperature of the heart at 35±1.5°C. The heart was pushed from behind by a flat plunger with a set screw so that the surface of the left ventricle was stabilized against a heated glass plate. Heating of this plate limited the fall in temperature during ischemia (measured with a small thermistor) to 1.2°C at 2 minutes and 2.1°C at 4 minutes. The heart was paced from right ventricle. Rhod-2 AM 1 mg was dissolved in 1 mL of demethylsulfoxide containing Pluronic F-127 (20% wt/vol). This solution was diluted in 100 mL of Tyrode’s solution (final rhod-2 concentration=10 μmol/L) and infused over 4 minutes. After the infusion, 30 minutes were allowed for de-esterification of the dye. Ischemia was produced by cessation of coronary flow. All ischemic trials were preceded by 2 minutes of preperfusion with blood, which was diluted 50% with Tyrode’s solution and heparinized (50 U/mL).5
Light from a 2-W diode-pumped solid state laser (Spectra-Physics) with 532-nm single-line output was directed into a beam expander and onto the surface of four 6-mm fiber bundles, which were pointed at the heart surface. Fluorescence was collected by a Dalsa 12-bit digital camera fitted with a 600-nm long-pass filter and a 25-mm f 0.85 video lens (CA-D1-0256T, Dalsa Inc). Digital images of the heart surface were taken at a resolution of 96×96 pixels at 300 frames/sec for 2 seconds.11 The resulting image subtended a square region of heart surface. Images were stored to disk and processed with MATLAB software. A series of 600 images was acquired every 30 seconds during ischemia, and a shutter was used to minimize bleaching. Hearts were paced at a cycle length of either 200 ms or 330 ms during ischemic trials and control images. No images were obtained <2 minutes after a change in pacing rate. The duration of ischemia was 6 minutes, except in one heart, where it was 20 minutes. (That heart was paced at CL=400 ms for most of the ischemic period.) Image files were postprocessed by interrogation of specific pixels two at a time, with a mouse for display of fluorescence versus time. Pixels could also be interrogated by entering their X and Y coordinates (1 through 96) on a keyboard. Each pixel was approximately 250 μm across.
Motion artifact was studied in one heart by illumination with white light in the absence of rhod-2. The optical motion signal was ≈3% of total light, which is 3 to 8 times smaller than the rhod-2 fluorescence transients. The motion signal was reduced to 1% of total light after 2 minutes of ischemia and was abolished by 3 minutes. Some recordings were obtained with an 11-mm fiberoptic probe and a photomultiplier fitted with a 590±10-nm-band pass filter.
An example of [C2+]i transients obtained with rhod-2 AM is shown in Figure 1. The time course of the rhod-2 transients is not different than those obtained in intact hearts with indo-1 or Fura Red. Fluorescence transients involve a change in fluorescence (ΔF/Fpeak) of 26% (lower trace). Motion artifact obtained with this apparatus is only 5% of total light, and the motion signal does not begin until after the [Ca2+]i transient has reached its peak.4 The relative fluorescence change in Figure 1 is greater than the best results obtained with indo-11,3,4 and comparable with the best results with Fura Red.5
Figure 2 shows similar [Ca2+]i transients obtained from specific pixels with the optical imaging camera in a nonischemic heart (bottom panel). The location of the two pixels is shown in the accompanying map. Although the absolute intensities of signals at the two locations varies, the shape of the transients is identical. Moreover, there is no variation of [Ca2+]i transient amplitude. At constant pacing rates, alternans was never seen in the absence of ischemia. The 8 images to the left in Figure 2 show propagation of excitation. Each successive panel is 3.3 ms later, giving a total of 26 ms. Excitation first appears in a small region near the pacing electrode and spreads across the image. Because propagation of the action potential is responsible for the spread of excitation, a conduction velocity of 0.6 m/sec can be calculated, which is typical of ventricular myocardium. .
Ischemia is known to produce broadening of the [Ca2+]i transient, along with beat-to-beat variation in peak amplitude. Recordings with fluorescent [Ca2+]i indicators also show fluctuations in end-diastolic [Ca2+]i such that the smaller transients arise from a higher baseline. The degree of [Ca2+]i transient alternans produced by ischemia can be expressed as the alternans ratio,1-B/A, where B is the net amplitude of the smaller transients and A is the net amplitude of the larger transients.5 This measurement has been used to compare the degree of alternans in blood- versus saline-perfused hearts. Figure 3A shows an example of [Ca2+]i transient alternans recorded from one pixel in a blood-perfused ischemic heart. There is a marked degree of alternans, giving an alternans ratio of 81%. In a previous study of 9 blood-perfused hearts, the largest alternans ratio obtained was 48%, whereas the mean value was 27.9±3.7%.5 In that study, the area recorded from was 480 times larger. Thus, the degree of [Ca2+]i transient alternans can be much greater in small pixel regions. [Ca2+]i transient alternans occurred in each of 8 hearts, but control records taken before ischemia did not show alternans in 10 to 15 pixels sampled. Absence of alternans in control records is consistent with previous results.4,5 Alternans was never observed during the first minute of ischemia but was observed consistently at 2,3, or 4 minutes.
When [Ca2+]i transients from pairs of pixels are compared, 4 results are possible: (1) neither pixel shows alternans; (2) only one pixel shows alternans; (3) both pixels show alternans of the same phase; or (4) both pixels show alternans of opposite phase. Examples of (3) and (4) are shown in Figures 3B and 3C. The green trace shown in Figures 3B and 3C is from the same pixel and shows smaller transients for even-numbered beats. In Figure 3B, the white trace shows alternans that is in phase with the green pixel region (even beats smaller), whereas in Figure 3C, the white trace shows alternans that is out of phase (even beats larger). Another pair of pixels from the same image files showing alternans of opposite phase is seen in Figure 4. These pixels were found by entering specific X and Y coordinates to sample the image evenly. Of the 8 hearts studied, 6 showed [Ca2+]i transient alternans that was out of phase in at least some pixels, whereas two showed alternans that was of one phase only.
Qualitative and Quantitative Maps of [Ca2+]i Transient Alternans
Alternans was found to be confined to specific regions of heart surface. In general, pixels near one another were more likely to have a similar pattern than pixels far apart. Two types of alternans maps are possible, qualitative and quantitative. An example of a qualitative map is shown in Figure 5, which is from the same heart as Figures 3B, 3C, and 4⇑. Out of a total of 172 pixels that were interrogated, 79% could be classified as having uniform calcium transients (constant peak systolic and end-diastolic levels) or alternans. Fifty eight pixels, all located in the bottom three fifths of the map, showed alternans with the first and odd beats taller (•), whereas 58 pixels, all located in the top three fifths of the map, showed alternans with the second and even beats taller (X). Twenty pixels showed uniform [Ca2+]i transients (○). Each of these pixels is located at the transition between the even-beat tall and odd-beat tall regions, so that they tend to form a boundary between the two regions.
Another example of a qualitative alternans map is shown in Figure 7A. Here, 34 pixels, covering most of the illuminated region, show alternans that is of the same phase (first complete transient is short; ○). However, 7 pixels show alternans that is out of phase (first complete transient is tall; X). All of these points are clustered together near the top of the figure. Four pixels were found with uniform [Ca2+]i transients (○). These pixels were near the region having alternans that is out of phase. Original records from 8 of the 45 pixels in Figure 7A are shown in Figure 6. Their location is indicated by numerals in Figure 6. For 5 of these pixels, the first complete transient is shorter, and for two pixels, it is taller. One pixel shows no alternans. .
The second kind of alternans map that can be constructed is a quantitative map in which an alternans ratio is computed for each pixel analyzed. An example of such a map is shown in Figure 7B. Here only one area of the illuminated region, which is roughly 5 mm across, shows alternans. All pixels within this area show alternans of the same phase. Inspection of the values shows that pixels located near to one another have similar ratio values and that the ratio approaches zero near the boundary. This relationship also holds in Figure 7A (ratio values in legend) and in Figure 5 (values not shown). In each case, the alternans ratio behaves as a continuous variable, which is largest in the center of regions with alternans.
The tendency of pixels that are located near one another to have similar alternans ratios can be evaluated mathematically using a correlation coefficient. For the 9 pixels in Figure 7B that show alternans, the difference in alternans ratio for each of the 30 possible pairs was expressed as an absolute value, and the distance between the pixels was measured. A significant linear correlation was found between the pixel distance and the ratio difference (R=0.49; P<0.005). This is additional proof that some physical process acts to synchronize [Ca2+]i transient alternans in myocardial cells that are close together.
Although fluorescent [Ca2+]i indicators are now widely used in intact hearts, this is the first study in which [Ca2+]i transients have been recorded simultaneously from multiple points during ischemia. These recordings have yielded potentially important information about the spatial heterogeneity of [Ca2+]i transient alternans, which complements emerging knowledge about the spatial features of APD alternans during ischemia. These studies were made possible by advances in imaging techniques11 and by use of the newer [Ca2+]i indicator rhod-2 AM, which has favorable characteristics for imaging blood-perfused hearts. [Ca2+]i transients recorded with rhod-2 are identical in shape to those recorded with Fura Red and indo-1, including indo-1 transients obtained with the ratio method. This observation confirms other evidence that motion artifact has minimal effect on these transients when the heart is properly stabilized.
The present study did not involve calibration of rhod-2 fluorescence. Calibration is not necessary, because the alternans ratio is independent of absolute [Ca2+]i. There are reports that rhod-2 loads mitochondria preferentially under certain conditions, but these reports do not pertain to the conditions used here. Del Nido et al9 found that in rabbit hearts loaded with rhod-2 AM, <5% of total light, as determined by Mn2+ quenching, came from organelles. These authors obtained values of [Ca2+]i comparable to those obtained by microinjection of calcium indicators into the cytoplasm.
Studies with 3 cell permeant fluorescent indicators have now shown that [Ca2+]i transient alternans occurs during the first few minutes of ischemia in intact hearts.4,5 This study expands on previous work by showing that alternans of [Ca2+]i transients is confined to specific regions on the surface of the heart, which are a few millimeters to centimeters across.
Relation of [Ca2+]i Transient Alternans to Contraction Alternans
[Ca2+]i transient alternans should produce alternations in contraction strength, which would exhibit similar heterogeneity to that observed in the [Ca2+]i transients. This was first reported in experiments where intracavitary pressure in the perfused rabbit heart was compared with recordings from epicardial strain gauge transducers.4 After 3 minutes of ischemia, alternans was observed in both recordings, but was sometimes out of phase, so that the stronger contractions in the ventricle coincided with the weaker local contractions. This result implies that a small epicardial region could have [Ca2+]i transient alternans that was out of phase with the remainder of the heart, as shown here in Figure 7A. Epicardial strain gauge recordings during ischemia show alternations in diastolic force, which are especially marked. This also agrees with the effects of ischemia on [Ca2+]i.
Relation of [Ca2+]i Transient Alternans to the Action Potential and T Wave
Large fluctuations in [Ca2+]i transient amplitude, like those shown in Figure 3A, should also affect membrane current and membrane potential. Artificial release of calcium by 10 mmol/L caffeine when the membrane is clamped at the resting potential causes an inward current that is carried by electrogenic sodium/calcium exchange.6 An increase in systolic calcium may therefore produce broadening of the action potential through augmentation of this current. However, an increase in peak systolic calcium can sometimes shorten the action potential, probably through faster inactivation of the calcium current.7
Simultaneous recordings of [Ca2+]i transient alternans and action potentials have been obtained in ischemic rabbit hearts using the Franz monophasic action potential electrode.4 In these studies, [Ca2+]i transient alternans was accompanied by beat-to-beat fluctuation in action potential duration (phase III APD alternans). There was no variation in action potential amplitude, nor was there conduction block; and when the short and long action potentials were superimposed, the point of divergence occurred after the peak of the calcium transients. Experiments of this type may be useful in determining the causal relation between action potential and calcium transient alternans, provided that both signals can be obtained from the cells producing the transients.9,14
Besides producing alternans in APD and contraction strength, cardiac ischemia produces beat-to-beat alternation in the T wave of the ECG of large mammalian hearts, which is prominent just before the onset of VF.12 Carson et al13 performed multielectrode mapping of T-wave alternans in the regionally ischemic pig heart and found that alternans is localized to hot spots about 1 cm across. The present study suggests that regional localization of [Ca2+]i transient alternans could explain the T-wave alternans maps.
Cellular Basis of [Ca2+]i Transient Alternans
Existing evidence suggests that effects of ischemia on the SR are the primary cause of [Ca2+]i transient alternans. Development of alternans is preceded by broadening of the [Ca2+]i transients. In nearly all recordings, the smaller transients arise from a higher end-diastolic level. This suggests that the SR takes up less calcium before the short transients and therefore releases less. [Ca2+]i transient alternans during ischemia may not be caused by variations in APD.4 This view is supported by the work of Chudin et al,14 who used voltage clamp to impose action potentials of fixed duration on rabbit ventricular myocytes that had shown APD and [Ca2+]i transient alternans when stimulated rapidly without voltage clamp. Their recordings show the same degree of [Ca2+]i transient alternans when APD was fixed as when it varied.
If [Ca2+]i transient amplitude is not controlled by APD, then some other mechanism must be responsible for synchronizing [Ca2+]i transient alternans in adjoining cells. Because each pixel samples from hundreds of myocytes, alternans would not be observed by optical mapping if each myocyte developed alternans independently of its neighbors. (Fluorescence of long-wavelength [Ca2+]i indicators arises primarily from cells within 1 mm of the heart surface.5 If cardiac myocytes are cylinders 30×120 μm, then each pixel contains about 360 myocytes). One mechanism that could explain the synchronization is the ability of calcium ions to cross gap junctions between cardiac cells. In the absence of action potentials, waves of calcium-induced calcium release can propagate between coupled cells and travel hundreds of microns in intact trabeculae.15,16 It follows that an unusually large calcium transient in one cell would accentuate calcium release in all of the adjoining cells to which it was coupled.
Another factor that favors the development of [Ca2+]i transient alternans during ischemia is perfusion of the heart with blood. [Ca2+]i transient alternans occurs nearly twice as often in blood-perfused hearts as in saline-perfused hearts.5 Long-wavelength indicators may be used to record [Ca2+]i transients in vivo from open-chest dog or pig hearts, which reliably fibrillate during the first few minutes of ischemia. Such experiments would permit direct proof of a relationship between [Ca2+]i transient alternans and VF.
This work was supported in part by NIH Grant HL58533. Animal care and use was as specified in the NIH Guide for Care and Use of Laboratory Animals (1996). We thank Dr Hon-Chi Lee for helpful comments and Rose Province of St Jude Medical for assistance.
Mohabir R, Lee H-C, 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.
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.
Figueredo VM, Brandes R, Weiner MW, et al. Endocardial versus epicardial differences of intracellular free calcium under normal and ischemic conditions. Circ Res. 1993; 72: 1082–1090.
Lee H,-C, Mohabir R, Smith N, et al. Effect of ischemia on calcium-dependent fluorescence transients in rabbit hearts containing indo 1. Circulation. 1988; 78: 1047–1059.
Wu Y-M, Clusin WT. Calcium transient alternans in blood-perfused ischemic hearts: observations with fluorescent indicator Fura Red. Am J Physiol. 1997; 273: H2161–H2169.
Del Nido PJ, Glynn P, Buenaventura P, et al. Fluorescence measurement of calcium transients in perfused rabbit heart using rhod-2. Am J Physiol. 1998; 274: H728–H741.
Konta T, Kozue I, Yamaki M. Significance of discordant ST alternans in ventricular fibrillation. Circulation. 1990; 82: 2185–3189.
Lamont C, Luther PW, Balke CW, et al. Intracellular Ca2+ waves in rat heart muscle. J Physiol. 1998; 51: 512–513.
Miura M, Boyden PA, Ter Keurs HEDJ. [Ca2+]i waves during triggered propagated contractions in intact trabeculae. Am J Physiol. 1998; 274: H266–H276.