We studied the origins of ectopic beats during low-flow reperfusion after acute regional ischemia in excised rat hearts. The left anterior descending coronary artery was cannulated. Perfusate was delivered to the cannula using an high-performance liquid chromatography pump. This provided not only precise control of flow rate but also avoided mechanical artifacts associated with vessel occlusion and deocclusion. Optical mapping of epicardial transmembrane potential served to identify activation wavefronts. Imaging of NADH fluorescence was used to quantify local ischemia. Our experiments suggest that low-flow reperfusion of ischemic myocardium leads to a highly heterogeneous ischemic substrate and that the degree of ischemia between adjacent patches of tissue changes in time. In contrast to transient ectopic activity observed during full-flow reperfusion, persistent ectopic arrhythmias were observed during low-flow reperfusion. The origins of ectopic beats were traceable to areas of high spatial gradients of changes in NADH fluorescence caused by low-flow reperfusion.
reperfusion arrhythmias in clinical settings are considered to be relatively transient and benign, particularly if the heart is structurally intact. This is because restoration of blood flow to reversibly damaged tissue quickly reinstates the ionic balance across the cell membrane and, consequently, normal conduction resumes. However, in some situations (i.e., a partially resolved blood clot, unsuccessful angioplasty, or the incomplete dilation of a coronary artery following a spasm), flow to the ischemic area is only partially restored. Indeed, an estimate based on Pouiselle's law of laminar flow predicts that a 50% reduction in a vessel's radius will result in a flow rate that is 6% of its original value. Low-flow reperfusion may also happen during cardiopulmonary resuscitation(29).
While providing tissue with minimal nutrients, low-flow reperfusion could be more arrhythmogenic than either ischemia or full-flow reperfusion. During low-flow reperfusion of an ischemic tissue bed, one can envision a continuously shifting pattern of flows between downstream vessels. This could be the result of autoregulation caused by ischemia-induced relaxation of smooth muscle cells, which increases the diameter of small coronary arteries, resulting in increased flow rates. Alternatively, it could be due to local hyperemia as adjacent tissue beds compete for the limited amount of flow. As a result of local shifts in flow, reperfused patches of myocardium might act as a substrate for persistent ectopic beats via calcium-overload mechanisms(2, 21). The effects of reperfusion-induced calcium overload can be further amplified by the release of norepinephrine from ischemic nerve endings(2, 15, 41).
The goal of our studies was to obtain direct experimental evidence for an increased incidence of ectopic beats during low-flow reperfusion. To avoid mechanical artifacts associated with vessel occlusion and deocclusion we remotely controlled flow to the left anterior descending coronary artery (LAD) by cannulating the vessel and delivering perfusate with an high-performance liquid chromatography (HPLC) pump. We imaged epicardial NADH fluorescence to measure local levels of ischemia. In parallel, transmembrane potentials were optically mapped to visualize activation wave fronts. A bipolar epicardial electrogram was recorded to continuously monitor heart rate. Data were analyzed to determine the incidence of arrhythmias and to reveal sources of ectopic beats.
In agreement with earlier studies (11, 14, 31), full-flow reperfusion of regionally ischemic myocardium was associated with brief arrhythmic episodes, which consisted of transient ectopic beats and occasional short-lived reentries. In contrast, we found that low-flow reperfusion of regionally ischemic myocardium led to persistent arrhythmias. These arrhythmias were fueled by multiple ectopic beats that were generated within areas having the highest spatial gradients of changes in NADH.
Hearts from male Sprague-Dawley rats (300–400 g) were excised and Langendorff-perfused with oxygenated Tyrode solution. The LAD was cannulated with polyimide tubing (127 μm ID), just distal to the coronary ostium (Fig. 1B). This is a technique we call “microcannulation.” With microcannulation, the flow of perfusate to a specific left ventricular myocardial volume was precisely controlled. The microcannula was connected to an HPLC pump (Series I, Scientific Systems) to deliver perfusate at 2.00 ml/min, the nominal flow rate. Perfusate was delivered to the right ventricle, septum, and any branches of the left coronary artery above the site of cannulation by perfusing the aorta at constant pressure (60 mmHg), as shown in Fig. 1C. A representative border between the tissue fed by the coronary ostia and the LAD cannula is shown in Fig. 1A. Intrinsic heart rate (∼30–90 beats/min) was stable throughout the experiment, unless arrhythmia was induced by reperfusion or other intervention. A detailed description of the microcannulation technique can be found in a separate technical report(37). All procedures involving animals have been approved by the Institutional Animal Care and Use Committee.
Hearts were stained with the potentiometric dye RH237 (10 μM solution, Molecular Probes), of which bolus injections (5 ml) were delivered to the aorta and the LAD. To prevent motion artifact Blebbistatin (10 μM) was added to the perfusate (13). After a stabilization period of 15–20 min, regional ischemia was induced by turning off the flow to the microcannula. The progression of ischemia was monitored at least every 2–3 min by imaging RH237 fluorescence for 12 s, followed by NADH imaging for 5 s. Flow to the LAD was resumed after 30 min of ischemia. A bipolar electrode was placed behind the heart to monitor basic electrical activity. Conclusions are based on a total of 15 successful microcannulation experiments, including 5 ischemia-full reperfusion experiments and 10 ischemia-low flow reperfusion studies. Typical electrograms, NADH, and transmembrane voltage data are shown.
An optical mapping system consisting of two CCD cameras (Andor IXON DV860s) fitted with a dual-port adapter (Andor CSU Adapter Dual Cam), a dichroic mirror (550 nm), and a single lens (Cosmicar 6 mm, F/1.0 with +27 closeup lenses) was used to image the epicardial fluorescence of RH237 and NADH from the same field of view. To record optical action potentials the epicardium was illuminated using two light emitting diodes (LumiLEDs, 530/35 nm). The resulting fluorescence of RH237 was long pass filtered at 680 nm and recorded at 250 fps. To record NADH the epicardium was illuminated with ultraviolet light (<360 nm) from a 100-W mercury lamp (Zeiss HBO100 W/2). The resulting fluorescence of NADH was band-pass filtered (475/50 nm) and recorded using the second CCD camera.
RH237 fluorescence was processed (7) to subtract background fluorescence from each image, and signals for each pixel were normalized. Noise was reduced using a digital low-pass Butterworth filter (passband and stopband frequencies at 30 and 40 Hz, respectively), and fluorescence signals were smoothed using a spatiotemporal conical filter with a radius of three pixels. The progression of ischemia was monitored as the average level of NADH fluorescence within a region of interest.
Processed datasets of RH237 fluorescence were analyzed to record the location of ectopic beats. They were identified as wave fronts having a localized source within the ischemic bed. The center point of the initial rise of RH237 fluorescence for each ectopic beat was used to pinpoint its source. Center points that had the same pixel location were automatically eliminated in subsequent spatial analyses of ectopic sources. Typically, between 50 and 150 spatially unique ectopic sources (center points) were identified during 15 min of low-flow reperfusion.
Local changes in NADH fluorescence [dNADH(i, j), Eq. 1] were determined by subtracting an NADH image acquired immediately before initiating low-flow reperfusion [NADHbefore(i, j)] from an NADH image acquired after low-flow reperfusion [NADHafter(i, j)]. The magnitude of the spatial gradient of dNADH [|∇dNADH(i, j)|] was then computed using Eq. 2. (1) (2)
Epicardial NADH during acute regional ischemia and full-flow reperfusion.
Epicardial NADH patterns before, during, and after acute regional ischemia for a typical experiment are shown in Fig. 2. A distinct area of elevated NADH fluorescence covered most of the left ventricle, as indicated by the white areas in Fig. 2A. Although the tissue bed perfused by the LAD was slightly different for each heart, the overall area and location was similar. The temporal progression of ischemia is shown in Fig. 2B. NADH levels steadily increased after the cessation of LAD flow, plateaued after 10.5 ± 0.9 min, and remained elevated until the reperfusion. The elevation of NADH was relatively uniform within the entire ischemic bed (Fig. 2A, top row). On the other hand, NADH during reperfusion was much more heterogeneous, with patches of tissue becoming less ischemic more rapidly than others (Fig. 2A, bottom row).
Ectopic beats during regional full-flow reperfusion.
Full-flow reperfusion of regionally ischemic myocardium was associated with a short period of tachyarrhythmias, shown in Fig. 2B as a transient peak in average heart rate. The two cases illustrate the similarities and variability between individual hearts. The transient tachyarrhythmic episodes consisted of ectopic beats and occasional short-lived reentries. A majority of the ectopic beats were traceable to the border area, in agreement with previous studies (31). An example of a reperfusion-induced propagation disturbance is shown in Fig. 2C. The top row shows a nominal beat and the bottom two rows illustrate the transition from ectopic activity to reentry.
Heart rate and spatial gradients of changes of NADH fluorescence caused by low-flow reperfusion.
Low-flow reperfusion (0.20 ml/min) of regionally ischemic myocardium caused sustained periods of tachyarrhythmias (Fig. 3A). The spatial heterogeneity of NADH fluorescence increased during low-flow reperfusion (Fig. 3B). Notably, NADH fluorescence decreased in some regions, whereas in other regions it increased or changed very little (Fig. 3C). The spatial gradient of changes in NADH fluorescence was computed to identify boundaries between tissue regions with different degrees of reflow as described in methods. An example of dNADH and the corresponding |∇dNADH| image as a pseudocolored surface is shown in Fig. 4A, right and left panels, respectively.
Local changes and spatial gradients of NADH at ectopic sources.
Temporal changes and the spatial gradients of NADH fluorescence in the vicinity of each ectopic source were analyzed to reveal whether the location of ectopic sources was determined by the change in the NADH in time or time and space (Fig. 4A). NADH levels changed over several minutes (Fig. 3C), whereas ectopic beats from individual sources appeared much more rapidly (within seconds). Because of this, sites of multiple ectopic sources could be superimposed with one image of dNADH (Fig. 4A). A visual analysis of the sites of ectopic sources combined with NADH data (red dots plotted with dNADH and black dots plotted with |∇dNADH|, Fig. 4A) suggested a strong correlation between ectopic locations and the spatial gradient of dNADH. A detailed statistical analysis was used to confirm this conclusion. First, dNADH and |∇dNADH| maps were spatially smoothed by averaging values within a 2.5-pixel radius at each pixel. Next, two sets of dNADH and |∇dNADH| values were analyzed: one set contained those values at sites of ectopic sources (set A), and another set contained those values for a uniformly distributed set of random sites (set B). Set B equaled in number to set A. Nonpaired t-tests and Kolmogorov-Smirnov nonparametric tests were used to examine the null hypothesis that there was no difference between dNADH and |∇dNADH| in sets A and B. The means of dNADH, |∇dNADH|, and their histograms are shown in Fig. 4, B and C, for sets A (ectopic sites), B (random sites), and C (all sites) from two representative studies. We found that dNADH was not significantly different between ectopic sites and random sites. However, |∇dNADH| was highly significantly different between ectopic sites, random sites, and all sites (P < 1.0e−5).
To the best of our knowledge this is the first study that has employed high-resolution optical mapping of NADH and transmembrane potential to correlate sites of ectopic beats with the underlying spatial distribution and degree of acute ischemia. This was accomplished by controlling flow to the LAD to mimic low-flow reperfusion. The technique allowed mechanical artifacts associated with vessel occlusion and deocclusion to be avoided, which was essential for mapping the exact locations of ectopic sources. As such, our studies provide the first intact-heart optical mapping recordings of ectopic beats originating from regionally ischemic and reperfused tissue. Zaitsev et al. (40) have described wave-break formation during ventricular fibrillation in regionally ischemic pig hearts, but they did not study reperfusion arrhythmias. Other studies have only used optical mapping to describe activity during global ischemia and/or global reperfusion (18, 24, 25, 38) or in heart preparations with healed infarcts (28).
Many mechanistic insights of reperfusion arrhythmias have been provided by earlier studies using multielectrode recordings, biochemical assessment, or both (5, 11, 14, 20, 22, 39). Acute ischemia results in acute biochemical changes, including acidification, accelerated efflux of intracellular potassium, accumulation of acyl-carnitines, and lysophospholipids (9, 12, 19, 26). The ensuing depolarization of resting membrane potentials, slowing of conduction, and increased dispersion of refractoriness all contribute to reentrant arrhythmias, which are prevalent during the first phase (1a) of an ischemic episode (19, 20, 27). This phase is followed by a short period when arrhythmia occurrence diminishes, likely due to the diffusion of K+ from the occluded area and the release of endogenous catecholamines, which temporarily improve tissue conductivity and excitability (9, 30). This is followed by phase 1b (20), during which a second rise of K+ occurs, gap junctional coupling decreases, and catecholamine-mediated triggered activity via delayed afterdepolarizations leads to an increased incidence of ectopic arrhythmias (10, 17, 23, 27, 30, 36). If the occlusion persists, irreversible cell injury ensues and is associated with progressive cell uncoupling, calcium overload, and cell death (8, 9, 23). Reperfusion of ischemic myocardium leads to a high incidence of arrhythmias, mostly via nonreentrant mechanisms (32, 34) with intracellular calcium overload as the major culprit (3, 9, 21). Electrode mapping studies in open-chest canines have shown that during acute ischemia 25% and 75% of arrhythmias occurring during ischemia and reperfusion, respectively, can be traced to ectopic sources–with a majority located near ischemic border zones (10, 33).
The data we have presented are in agreement with the findings described above, specifically with the transient nature of reperfusion arrhythmias (6), the localization of ectopic sources to ischemic boundaries (10, 33), and heterogeneity of reflow during reperfusion(16). However, our main emphasis is not on additional evidence that confirms previously established ectopic and/or reentry-based ischemia-reperfusion events. Instead, we call attention to the marked difference in electrical activity between full-flow and low-flow reperfusion. Low-flow reperfusion and the resulting patterns of flow create, in effect, a continuous boundary between partially reperfused and ischemic tissue within an ischemic tissue bed. This fuels a persistent generation of ectopic beats during low-flow reperfusion, creating a highly arrhythmogenic substrate. These findings match our previous in vitro studies where we hypothesized that ectopic beats generated during ischemia could be caused by microreperfusion of the ischemic border zone (1, 2, 35). Such transient ectopic waves can later transition into longer-lived reentrant sources upon changes in cell-cell coupling, tissue excitability, or position of the boundary (4, 35).
Studies were performed using unloaded blood-free excised rat hearts that were perfused with Tyrode solution. Therefore, the direct applicability of our results to clinical cases is limited and additional studies are needed to fully translate the observed phenomena to human patients.
Financial support by the American Heart Association, Mid-Atlantic Affiliate (BGIA 0665377U to M. W. Kay) and the National Heart, Lung, Blood Institute (HL-076722 to N. Sarvazyan) is gratefully acknowledged.
We thank Vishal Khatri for excellent technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 by the American Physiological Society