Modulation of Ventricular Repolarization by a Premature Stimulus
Role of Epicardial Dispersion of Repolarization Kinetics Demonstrated by Optical Mapping of the Intact Guinea Pig Heart
Recent evidence suggests that ion channels governing the response of action potential duration (APD) to a premature stimulus (ie, APD restitution) are heterogeneously dispersed throughout the heart. However, because of limitations of conventional electrophysiological recording techniques, the effects of restitution in single cells on ventricular repolarization at the level of the intact heart are poorly understood. Using high-resolution optical mapping with voltage-sensitive dyes, we measured APD restitution kinetics at 128 simultaneous sites on the epicardial surface (1 cm2) of intact guinea pig hearts (n=15). During steady state baseline pacing, APD gradients that produced a spatial dispersion of repolarization were observed. Mean APD was shortened monotonically from 186±19 ms during baseline pacing (S1-S1 cycle length, 393±19 ms) to 120±4 ms as single premature stimuli were introduced at progressively shorter coupling intervals (shortest S1-S2, 190±15 ms). In contrast, premature stimuli caused biphasic modulation of APD dispersion (defined as the variance of APD measured throughout the mapping field). Over a broad range of increasingly premature coupling intervals, APD dispersion decreased from 70±29 ms2 to a minimum of 10±7 ms2 at a critical S1-S2 interval (216±18 ms), and then, at shorter premature coupling intervals, APD dispersion increased sharply to 66±25 ms2. Modulation of APD dispersion by premature stimuli was attributed to coupling interval–dependent changes in the magnitude and direction of ventricular APD gradients, which, in turn, were explained by systematic heterogeneities of APD restitution across the epicardial surface. There was a characteristic pattern in the spatial distribution of cellular restitution such that faster restitution kinetics were closely associated with longer baseline APD. This relationship explained the reversal of APD between single cells, inversion of APD gradients across the heart, and ECG T-wave inversion during closely coupled premature stimulation. Therefore, because of the heterogeneous distribution of cellular restitution kinetics across the epicardial surface, a single premature stimulus profoundly altered the pattern and synchronization of ventricular repolarization in the intact ventricle. This response has important mechanistic implications in the initiation of arrhythmias that are dependent on dispersion of repolarization.
It is well recognized from studies in single cells that cellular repolarization and APD are highly sensitive to the timing of a premature stimulus.1 2 APD shortens in an exponential fashion as the premature coupling interval is progressively shortened. The membrane ionic and intracellular processes that control the extent of APD shortening following a premature stimulus are referred to collectively as APD restitution.1 2 3 4 5 6 Because of the diversity of repolarizing currents, there are important differences in the kinetics of restitution throughout the heart. For example, restitution kinetics of Purkinje and ventricular muscle fibers differ.4 7 In addition, there is now compelling evidence that restitution properties may vary substantially between canine ventricular myocytes on the epicardial, midmyocardial, and endocardial surfaces,8 presumably on the basis of a heterogeneous distribution of IK9 or its components, IKr and IKs.10 However, it is not clear whether restitution kinetics of cells across the epicardial surface are heterogeneous, even though significant gradients of APD have been observed previously within ventricular epicardium.11 12 Moreover, because of limitations of conventional recording techniques, it is not known how regional differences in restitution at the level of the single cell affect the timing, gradients, and spatial synchronization (ie, dispersion) of repolarization at the level of the intact ventricle.
In the present study, high-resolution action potential mapping with voltage-sensitive dyes was used to establish how spatial heterogeneities of cellular restitution kinetics control the pattern of ventricular repolarization during a premature stimulus. An experimental system for recording high-fidelity transmembrane action potentials from multiple simultaneous sites permitted detailed and quantitative analysis of APD restitution in the intact heart. We found that the kinetics of APD restitution varied systematically across the epicardial surface and that such heterogeneities of restitution modulated ventricular repolarization during a premature stimulus. Since the spatial organization (ie, dispersion) of repolarization is important to the mechanism of reentrant arrhythmogenesis,13 14 modulation of ventricular repolarization by extrastimuli has important implications for the mechanism of arrhythmia initiation by premature stimuli.
Materials and Methods
Experiments were carried out in accordance with Public Health Service guidelines for the care and use of laboratory animals. Guinea pigs were anesthetized (30 mg/kg pentobarbital sodium [Nembutal] IP), and their hearts were rapidly excised and perfused as Langendorff preparations (perfusion pressure, 70 mm Hg) with oxygenated (95% O2/5% CO2) Tyrode's solution containing (mmol/L) NaCl 130, NaHCO3 25.0, MgSO4 1.2, KCl 4.75, dextrose 5.0, and CaCl2 1.25 (pH 7.40). The right atrium was excised to avoid competitive stimulation from the sinoatrial node. Hearts were stained with the voltage-sensitive dye di-4-ANEPPS (10 μmol/L) by direct coronary perfusion for 10 minutes. This staining procedure was not associated with any cardiac toxicity in our preparations. Restitution was measured in the following experimental groups: The primary analysis of restitution was made in group 1, where in seven experiments perfusion temperature was maintained at 32°C to reduce intrinsic automaticity permitting reliable stimulation of the ventricle throughout the restitution protocol. In group 2, the experiments performed in group 1 were repeated in three additional preparations using a perfusion temperature of 39°C to confirm reproducibility of our results at physiological temperatures. During perfusion at 39°C, the endocardial surface was eliminated using a cryoablation procedure described previously15 in order to prevent automaticity originating from the His-Purkinje system. This procedure produces a thin viable rim (≈800-μm depth) of epicardium having normal electrophysiological properties.15 In this group, the restitution measurements were repeated using a broad range of baseline pacing cycle lengths (1000, 350, 250, and 190 ms) to confirm that baseline cycle length did not significantly influence our results. In group 3, two additional experiments were performed during perfusion at 39°C but without endocardial cryoablation to test the effect of rapid baseline pacing cycle lengths (250 and 190 ms) at physiological temperatures with endocardium left intact (compared with group 2). In group 4, the experiments performed in group 1 were repeated in three additional preparations after endocardial cryoablation. These experiments were compared with experiments performed at the same temperature but without endocardial ablation (group 1) to determine whether endocardial tissue influences action potential restitution measured from cells on the epicardial surface.
Beating and perfused hearts were placed in a custom-built Lexan chamber16 that was attached to a micromanipulator so that the mapping field (1 cm×1 cm) could be centered over the left anterior descending coronary artery ≈6 mm below its bifurcation with the diagonal coronary artery. To avoid surface cooling and the formation of intracardiac temperature gradients, the heart was immersed in the coronary effluent, which was maintained at a constant temperature (equal to the perfusion temperature) with a heat exchanger in the chamber. Gentle pressure was applied to the posterior surface of the heart with a movable piston to stabilize the anterior surface of the ventricle against an imaging window and to eliminate apparent motion artifact without altering action potential properties.15 17 This design permitted the heart to contract freely during action potential recordings, except for the 1-cm2 area within the mapping field, while avoiding electrophysiological side effects associated with pharmacological suppression of cardiac motion.17 18 19 Cardiac rhythm was monitored using three silver disk electrodes fixed to the chamber in positions roughly corresponding to ECG limb leads I, II, and III. The ECG signals were filtered (0.3 to 300 Hz), amplified (×1000), and displayed on a digital oscilloscope (DRO 1604, Gould Inc). To ensure the physiological stability of these preparations, action potentials were monitored from selected recording sites for any evidence of ischemia (ie, triangulation, decreased amplitude, or decreased duration). In addition to monitoring coronary flow (typically 10 to 15 mL/min), coronary resistance, and oxygen tension, at the conclusion of each experiment preparations were examined using 2,3,5-triphenyltetrazolium chloride staining and light microscopy to confirm tissue viability. In general, these preparations were stable and exhibited normal action potential characteristics for at least 3 hours of perfusion. The experimental protocol was typically completed within 1 hour.
Optical Mapping System
We have developed an optical action potential mapping system that is capable of recording high-fidelity action potentials simultaneously from 128 sites of the intact guinea pig heart.15 16 This system was designed to resolve membrane potential changes as small as 0.5 mV, thereby permitting quantitative analysis of action potential shape and duration.15 Fluorescence was excited from voltage-sensitive dye using a tungsten-filament lamp (250 W) and an interference filter (540±10 nm). Fluoresced light was long-wave pass–filtered (610 nm) and focused with a high–numerical aperture (effective numerical aperture, 0.36) photographic lens (Nikon 85 mm, F/1.4) onto a 12×12 element photodiode array (MD144-5T, Centronics Ltd). In the present study, an optical magnification of ×1.8 was used, which corresponded to a total mapping field of 10 mm×10 mm, 0.83-mm interpixel spatial resolution, and an estimated depth of field of 0.2 mm. Therefore, optical action potentials recorded from each photodiode element were derived from the summation of cellular transmembrane potentials within a 0.6-mm2 area and 0.2-mm depth (ie, 0.12-mm3 sample volume). We have previously found that in the region of the anterior left ventricle, cardiac fibers run essentially perpendicular to the left anterior descending coronary artery up to a depth of 1 mm.15 Therefore, optical action potentials were recorded from a uniform layer of myocardium having no significant rotational anisotropy. A 768×492 pixel CCD video camera (TM-7CN, PULNiX) was used to view and digitize (VideoPix Frame Grabber, Sun Microsystems) the position of the mapping array relative to anatomic features of the preparation (eg, coronary arteries) and to determine the precise location of the mapping array on the heart surface.
Photocurrent from 128 photodiodes was passed through low-noise current to voltage converters (250-MΩ feedback resistance) and then underwent postamplification (×50) with AC coupling (10-s time constant) and low-pass anti-aliasing filtering (500 Hz). Action potentials recorded from each photodiode and ECG signals were multiplexed and digitized with 12-bit precision at a sampling rate of 2000 Hz per channel to the hard disk of a UNIX workstation (Concurrent 5450S, Concurrent Computer Corp). A software interface designed to provide real-time measurements of action potential amplitude in each channel was used to modify excitation light intensity or amplifier gain as needed to fully use 12 bits of resolution. A burst sampling scheme was used to reduce the sampling interval (ie, skew) between channels to 2 μs.
The ventricular epicardial surface was stimulated (DCI-1114, Digital Cardiovascular Instruments Inc) at five times diastolic threshold current using a polytetrafluoroethylene-coated silver bipolar electrode (0.1-mm diameter; interelectrode spacing, 1.0 mm). To ensure steady state conditions, the preparation was paced at a cycle length of 400 ms until a constant QT interval (measured from ECG) was observed for at least 10 minutes. In addition, optical action potentials were periodically examined to verify the stability of repolarization over long time periods. In one experiment (group 1), the heart was paced at a cycle length of 350 ms to overdrive the intrinsic automaticity, which was slightly faster than 400 ms.
Once steady state was achieved, APD restitution was measured simultaneously from each of 128 ventricular recording sites by introducing a single premature stimulus (S2) after a 50-beat drive train (S1) at the baseline cycle length of 400 ms. The premature coupling interval (S1-S2) started at 400 ms and was progressively shortened by 5-ms decrements until ventricular refractoriness was reached. Action potentials were recorded from each mapping site during the last three beats of baseline pacing (S1) and during the prematurely stimulated beat (S2). This pacing protocol was repeated from the identical pacing site in two experiments to ensure that restitution measurements were reproducible over time. A similar pacing protocol was used when other baseline cycle lengths were tested (groups 2, 3, and 4).
Previous studies have shown that APD can be influenced by electrotonic loading between cells20 21 22 in addition to cellular membrane kinetics. Since the direction of wave-front propagation and tissue anisotropy can influence cell-to-cell loading during repolarization,22 it is possible that APD restitution measured in multicellular preparations can be influenced by the direction of propagation. To determine whether restitution kinetics measured at each ventricular site were an intrinsic property of the cells at each site (ie, independent of propagation direction), the restitution protocol was repeated from a second stimulation site located on the opposite side of the mapping field in five of the seven preparations in group 1.
Because of the large number of action potentials recorded during each experiment (≈20 000), automated algorithms were required to ensure objectivity and consistency in determining depolarization and repolarization times for each action potential. All computer-assigned depolarization and repolarization times were reviewed by the investigators.
Shown schematically in Fig 1A⇓ is an action potential recorded from one site during the last beat of a drive train (S1), followed by a single premature stimulus (S2). For each prematurely stimulated action potential, DTp, RTp, APDp, and DI were measured from action potentials recorded from each mapping site using previously established techniques.11 15 DTp was defined as the point of maximum positive derivative in the action potential upstroke (dV/dtmax). RTp was defined as the maximum positive curvature (maximum positive second derivative) during repolarization and corresponds to ≈95% repolarization23 . DTp and RTp were determined by convolving the digitally sampled action potentials with a differentiating and smoothing digital filter.24 APDp was defined as the difference between DTp and RTp. For the last beat of baseline pacing, DTb, RTb, and APDb were determined using the same algorithms described above. DI was calculated by subtracting DTp from the repolarization time of the last drive-train beat (RTb).
Previously, restitution kinetics have been characterized by fitting the relationship between APDp and DI to single-exponential1 7 or double-exponential25 data models. However, since nonexponential behavior has been observed5 26 and since in a series of preliminary studies27 we have found that the characteristics of restitution curves vary when measured from hundreds of sites across the ventricular wall, we used an intuitive measure of restitution that does not assume a predefined mathematical relation between APDp and DI. Restitution kinetics were measured at each site by an empirical rate constant, RK, defined as follows (Fig 1B⇑):where ΔAPD is the extent to which APD was shortened over the range of DIs tested (ΔDI). Therefore, RK is an index of the “time course of restitution” and reflects overall responsiveness of APD to a change in DI measured during a single premature stimulus. We found that RK is a robust measure of the kinetics of the rapid phase of the restitution response, since the relative magnitude of RK is insensitive to the specific range of DIs used to define it, provided this range is consistently defined for each site and that the maximum DI is taken from the flat portion of the restitution curve.
The patterns of depolarization, APD, repolarization, and RK were depicted graphically using gray scale contour plots with spline interpolation. Four pixels in each corner of the mapping array (which were not used) and selected pixels containing inadequate signals were replaced by the weighted average of neighboring pixels.
To quantitatively analyze epicardial repolarization gradients, APD gradient vectors were calculated directly from action potential maps using a computer algorithm. First, a local APD gradient was determined for each pixel in the mapping array from the difference in APD between each pixel and its eight neighboring pixels. Second, local APD gradients were vectorally summed to determine a resultant APD gradient vector for the mapping surface. Each computer-determined APD gradient vector was inspected by the investigators to confirm that they had accurately represented APD contour maps.
Changes in mean APD and APD dispersion from baseline stimulation to premature stimulation were compared using a paired Student's t test. Bonferroni's correction was applied for multiple comparisons where appropriate. A value of P<.05 was considered statistically significant.
APD Restitution at Individual Recording Sites
An example of APD restitution measured from two representative ventricular sites is shown in Fig 2⇓. Optical action potentials (inset) recorded simultaneously from these two sites are also shown. The prematurely stimulated action potentials measured at the site at which APD during baseline pacing (APDb=bold action potentials) was long demonstrated a greater degree of APD shortening compared with the site at which APDb was short. The difference in APD shortening between these two sites is also evident in the restitution curves (Fig 2⇓). At long DI, the difference in APDp between these sites is equal to the difference in APDb. However, as DI decreases (ie, with greater S1-S2 prematurity), the restitution curves converge. This convergence is a result of greater APD shortening, ie, a faster time course of restitution (RK=0.47) at the site having longer APDb compared with relatively slow restitution (RK=0.29) at the site where APDb was short. Moreover, heterogeneities of restitution between epicardial cells were also evident when the extent of APD shortening was expressed as a percent change from baseline APD rather than the absolute change in APD. When comparing percent APD change in all experiments, APD was shortened by 41±6% in response to premature stimuli in cells having longest baseline APD (195±19 ms) compared with only 24±8% in cells with the shortest APD (172±19 ms) during baseline stimulation. The difference in APD response between these cell populations was highly significant (P<.0001).
Heterogeneity of APD Restitution Across the Epicardial Surface
RK was calculated at each of 128 ventricular sites to determine whether restitution kinetics were heterogeneously distributed across the epicardial surface. Fig 3⇓ shows the spatial distribution of RK measured from a representative experiment. In this example, pacing was performed from the left side of the mapping field (site A). Within this 1-cm2 area of epicardium, RK varied by as much as 500% (range, 0.24 to 0.04). Moreover, spatial heterogeneity of RK was not random; rather, there was an organized pattern of RK across the epicardial surface. This pattern closely paralleled the spatial distribution of APD measured during baseline pacing. By comparing the two contour plots in Fig 3⇓, it is evident that in regions where APDb (right) is longest, the kinetics of restitution (RK, left), are fastest. The graph in Fig 3⇓ (bottom) quantitatively demonstrates a close linear relationship between RK and APDb.
To confirm that heterogeneity of restitution kinetics measured in these experiments reflects heterogeneity of intrinsic membrane ionic currents and is not dependent on electrotonic loading during propagation (see “Stimulation Protocol”), RK was compared when pacing from opposite sides of the mapping field (sites A and B in Fig 3⇑). As shown in Fig 4⇓, RK was not significantly influenced by the direction of propagation. Similarly, APDb (Fig 4⇓) was independent of propagation, indicating that both RK and APDb are properties intrinsic to the cells at each recording site and also reaffirming the close quantitative relationship between RK and APDb.
Modulation of APD Gradients (Dispersion) by a Premature Stimulus
To quantitatively assess the effect that heterogeneity of cellular restitution had on spatial dispersion of APD during a premature stimulus, contour maps of depolarization and APDp were analyzed for each S1-S2 premature coupling interval tested. Fig 5⇓ shows a representative example of the spatial patterns of epicardial depolarization (panels A, C, and E) and APD (panels B, D, and F) during baseline pacing (panels A and B), during a premature stimulus at an intermediate coupling interval (panels C and D), and during a premature stimulus at a short (within 15 ms of refractoriness at the pacing site) coupling interval (panels E and F). During baseline pacing (Fig 5A⇓), depolarization propagated uniformly from the site of pacing, and significant gradients of APD (APDb) were evident (Fig 5B⇓). A premature stimulus introduced at an intermediate coupling interval produced no significant change in the pattern of depolarization (Fig 5C⇓); however, APD became much more homogeneous (Fig 5D⇓), and the APD gradients that were evident during baseline pacing (Fig 5B⇓) were eliminated. Consequently, APD dispersion (defined as the variance of APD measured over all ventricular mapping sites) during the premature stimulus (7 ms2) was reduced substantially compared with dispersion during baseline stimulation (61 ms2). When a premature stimulus was introduced at a very short coupling interval, conduction slowed somewhat, as evidenced by slight crowding of isochrone lines (Fig 5E⇓) near the region of long APD (Fig 5B⇓), but the overall pattern of depolarization remained unchanged. In contrast, APD changed substantially (Fig 5F⇓). The dispersion in APD reappeared (72 ms2) and was comparable to dispersion measured during baseline pacing. However, the orientation of the APD gradient was inverted; ie, where APD was longest during baseline pacing became shortest during the premature beat, and vice versa (compare Fig 5F⇓ with 5B).
The spatial dispersion of APD measured during baseline stimulation and during premature stimulation at intermediate and short coupling intervals is summarized for all experiments in the Table⇓. Compared with mean APD during steady state baseline pacing (186±19 ms), mean APD decreased progressively at intermediate (162±10 ms) and short (120±4 ms) S1-S2 coupling intervals. In contrast, APD dispersion exhibited a biphasic dependence on premature coupling interval; ie, dispersion decreased to a minimum (from 70±29 to 10±7 ms2) at intermediate S1-S2 coupling intervals but then increased (to 66±25 ms2) as the coupling interval was progressively shortened beyond the intermediate value.
The APD gradient vector plots in Fig 6⇓ demonstrate the modulation of repolarization by premature stimuli and specifically highlight inversion in the orientation of APD gradients at short premature coupling intervals. The magnitude and orientation of APD gradients measured from the epicardial surface of each preparation are represented by the length and angle of each vector, respectively. During baseline stimulation (S1), APD gradients having similar magnitude (average, 32.3±10.1 ms/cm) were present in every experiment. Although some variability was present, APD gradients were generally oriented perpendicular to the left anterior descending coronary artery and parallel to cardiac fibers. Irrespective of the orientation of APD gradients during baseline stimulation, a closely coupled premature stimulus (S1-S2=190±15 ms; see the Table⇑) produced an APD gradient (average, 35.2±8.1 ms/cm) having comparable magnitude to the gradient present at baseline, but in each case the direction of these gradients was reversed nearly 180° from baseline (ie, “APD gradient inversion”).
A pattern of action potential responses identical to those shown in Figs 5 and 6⇑⇑ was observed in preparations in which endocardial and midmyocardial layers were eliminated by cryoablation, indicating that subepicardial tissues were not essential to this response. Similar results were also obtained in separate experiments performed at temperatures up to 39°C and at baseline pacing cycle lengths varying from 190 to 1000 ms, indicating that modulation of APD gradients by premature stimuli occurs over a broad range of physiological heart rates and temperatures.
Heterogeneity of APD Restitution and Its Effect on Repolarization
To examine the effect that spatial dispersion of restitution kinetics has on repolarization (RTp), we measured RTp and its components, depolarization time and APD (RTp=DTp+APDp). Fig 7⇓ compares depolarization (panels A and D), APD (panels B and E), and repolarization (panels C and F) during baseline pacing (S1) and after a premature stimulus introduced at a S1-S2 coupling interval near refractoriness. Compared with the pattern of depolarization during baseline stimulation (Fig 7A⇓), the pattern of depolarization during the premature stimulus is similar except for slight slowing of conduction (Fig 7D⇓, lower left) where APDb is relatively long (Fig 7B⇓, lower left). In contrast, the pattern of APD changed significantly (Fig 7B and 7E⇓⇓), as evidenced by the inversion of the APD gradient. Repolarization was also significantly changed by premature stimulation. During baseline pacing, earliest repolarization (Fig 7C⇓) began at a site in the mapping field opposite the pacing site. However, a premature stimulus increased the spatial dispersion of repolarization (crowding of repolarization isochrones, Fig 7F⇓) and shifted the earliest site of repolarization to the opposite side of the mapping field, toward the site of pacing. Inversion of repolarization during the premature stimulus is reflected in the ECG (bottom panel) by the change in polarity of the T wave, indicating that this phenomenon was not limited to cells within the mapping array. Since the pattern of depolarization was not substantially changed by the premature stimulus, changes in the pattern of repolarization were primarily due to modulation of APD gradients rather than conduction slowing.
In the present study, it was necessary to record high-fidelity action potential signals from multiple simultaneous sites to determine how restitution kinetics of individual cells and heterogeneity of restitution between cells control the pattern and synchronization of repolarization in the intact heart. Because of limitations associated with conventional single-site microelectrode recordings and nonsimultaneous refractory period mapping with extracellular electrodes, the effect of restitution properties of single cells on the functional organization of ventricular repolarization in the intact ventricle has not been well understood. Rather, investigations have focused on analyzing rate-dependent changes in total repolarization time or QT interval.1 2 28 We demonstrate that there are substantial gradients of cellular repolarization in guinea pig epicardium. The magnitude and orientation of these gradients are profoundly sensitive to the S1-S2 premature coupling interval and can be explained on the basis of heterogeneities of cellular restitution kinetics across the epicardial surface.
APD Gradients in Ventricular Myocardium
During steady state baseline stimulation, we consistently observed uniform gradients of APD across the ventricular epicardial surface (Fig 6⇑). In the absence of a premature stimulus, the magnitude and orientation of APD gradients remained essentially constant over a broad range of baseline stimulation rates and perfusion temperatures. Therefore, the premature stimulus, and not rapid pacing per se, was responsible for modulation of repolarization gradients. These findings are consistent with an earlier report from Rosenbaum et al,11 who found that when cycle length is abruptly shortened and maintained at a faster rate, APD gradients are quickly restored after steady state is achieved at the faster cycle length. The persistence of APD gradients we observed across the epicardial surface at rapid pacing rates is of interest, since transmural (ie, epicardium-to-endocardium) APD gradients in canine ventricle are not maintained at rapid heart rates.10 Moreover, we report significantly larger APD differences across the epicardial surface than APD differences measured previously between epicardium and endocardium.29 These findings suggest that APD differences across the epicardial surface, rather than APD differences across the ventricular wall, contribute more significantly to action potential heterogeneities, particularly during tachycardia.
It is evident from Fig 6⇑ that APD gradients measured during steady state pacing were typically oriented perpendicular to the left anterior descending coronary artery and parallel to cardiac muscle fibers of this preparation.15 APD always increased progressively from apical to more basal regions of both ventricles (Figs 5 and 6⇑⇑). Recent studies have also reported longer APDs near the base of the guinea pig ventricle compared with the apex12 23 29 and have suggested a possible relationship between repolarization and fiber structure.29 However, some earlier reports have suggested that APD is longer at the apex rather than the base of the ventricle.11 30 It is not likely that our results differed from these earlier studies because of differences in pacing cycle length and perfusion temperatures, since we found that the magnitude and orientation of APD gradients are not substantially altered by steady state cycle length or a wide range of temperatures. However, there were important differences in the experimental preparations, as one of these studies was performed in working Langendorff hearts.30 It is also possible that the location of the mapping field used in earlier investigations may have differed from the present study. We incorporate a CCD video camera into our present mapping system design that uses a common optical path with the photodiode array to directly measure the precise position of the mapping array with respect to the heart's surface (see Fig 3⇑ for actual video image). In earlier studies, the location of the mapping field was estimated visually and was therefore not as well controlled.
Control of Ventricular Repolarization During a Premature Stimulus
A principal aim of the present study was to investigate the mechanisms responsible for modulation of cardiac repolarization and, specifically, to establish the role played by heterogeneities of cellular restitution kinetics across the epicardial surface. One can examine how regional differences in cellular restitution kinetics affect dispersion of APD by comparing the time course of restitution at two ventricular sites. The dashed lines connecting the restitution curves in Fig 2⇑ correspond to APD measured simultaneously from the two ventricular sites at each S1-S2 coupling interval. During baseline stimulation, or when premature coupling intervals are relatively long (ie, long DI in Fig 2⇑), there is a dispersion of APD between these sites that is equal to the vertical distance between the points connected by dashed lines. However, as the S1-S2 coupling interval is shortened to an intermediate value, APD dispersion is eliminated (ie, the dashed lines become horizontal). With further S1-S2 shortening to very premature coupling intervals, the dashed lines develop a steep positive slope rather than the negative slope that was present at long coupling intervals, indicating a “reversal” of APD; ie, sites with longest baseline APD developed the shortest APD and vice versa. APD reversal causes a restoration of APD gradients (ie, dispersion) similar in magnitude but opposite in orientation to the gradients present during baseline stimulation (Fig 6⇑). Such modulation of APD depicted from the two sites shown in Fig 2⇑ is consistent with changes in APD dispersion that we observed throughout the epicardial surface (Fig 5⇑).
There were two conditions required for modulation of dispersion by premature stimuli: (1) dispersion of cellular restitution kinetics (ie, regional differences in the restitution response itself) and (2) dispersion of DIs between cells (ie, differences in the input to restitution). The former condition is dependent on the kinetics of membrane repolarization, whereas the latter is dependent on membrane repolarization and conduction. Since restitution kinetics (RK) is faster at sites having longer baseline APD (Fig 3⇑), APD shortens more rapidly at these sites compared with sites with shorter APDs, resulting in narrowing of APD dispersion (ie, convergence of restitution curves) as the S1-S2 interval is shortened (Fig 2⇑). In addition to heterogeneities of restitution, propagation introduces heterogeneities of DI. A site having shorter DI compared with a neighboring site is operating on a steeper portion of its restitution curve, introducing greater dispersion of APD.
The importance of heterogeneities of restitution properties across the epicardial surface to the modulation of APD dispersion can be derived from the restitution curve shown in Fig 8A⇓. In that figure, we simulate a hypothetical situation in which spatial heterogeneities of restitution are absent (ie, there is a single restitution curve); hence, APD dispersion between two recording sites (filled and open circles in Fig 8A⇓) is determined exclusively by differences in DI between the sites. Under such circumstances of homogeneous restitution, APD dispersion would increase monotonically with S1-S2 prematurity (Fig 8B⇓), which is not consistent with our experimental results. We measured biphasic modulation of APD dispersion in our experiments (Fig 8B⇓); ie, APD dispersion decreased and then increased as the S1-S2 coupling interval was progressively shortened. Therefore, spatial heterogeneities of restitution contributed importantly to the modulation of APD dispersion by premature stimuli, and regional diversity of membrane responses was an essential mechanism for reducing dispersion.
Since both RK and DI are influenced by baseline APD, they are not entirely independent of one another. Changes in APD dispersion provoked by a premature stimulus are therefore determined by the interdependence of two highly nonlinear processes, illustrating the complexity of applying the principles of cellular restitution in multicellular tissues. For example, a site with long APD will have a faster time course of restitution (Fig 3⇑) and at the same time a shorter DI (ie, stronger input to restitution). Therefore, a small increase in baseline APD dispersion is expected to produce marked modulation of APD dispersion by a premature stimulus. One would therefore predict that pathological conditions that either enhance baseline dispersion of refractoriness14 31 or slow conduction32 will result in even greater modulation of repolarization during a premature stimulus.
Role of Passive Membrane Properties
We found that restitution properties inherent to the cells at each recording site were the primary determinant of epicardial repolarization. It is important to consider that passive electrotonic coupling between cells may additionally influence patterns of repolarization.20 22 Toyoshima and Burgess21 have suggested that electrotonic loading may cause APD to shorten in the direction of propagation during steady state ventricular stimulation. However, we found that during steady state stimulation APD was independent of propagation direction (Fig 4⇑) and that during a premature stimulus APD gradients were determined by restitution rather than propagation direction (Fig 7⇑). Therefore, it is likely that passive electrical properties played a relatively minor role in the modulation of repolarization during a premature stimulus.
It has also been postulated that epicardial repolarization is determined by “propagation” of repolarization along cardiac fibers from the endocardial surface.29 However, we demonstrated that modulation of repolarization by a premature stimulus was not dependent on subepicardial or endocardial tissue, as epicardial repolarization was not affected by elimination of subepicardial muscle layers after endocardial cryoablation. In addition, we found that APD is highly dependent on the premature coupling interval and not fiber orientation. Although propagation direction and fiber structure were constant, the pattern of repolarization was completely reversed by an appropriately timed premature stimulus (Figs 5 and 7⇑⇑), proving that intrinsic restitution was a far more important determinant of repolarization than underlying fiber structure.
Ionic Mechanisms Responsible for Spatial Dispersion of Restitution Kinetics
In the present study, we demonstrate that steady state APD and restitution kinetics vary considerably between cells across the epicardial surface of the heart. Since recordings were made from a uniform, viable, and well-coupled layer of epicardial tissue, where electrotonic interaction between cells may otherwise minimize regional heterogeneities of APD,20 our findings suggest the presence of considerable heterogeneity of membrane ionic processes across the epicardial surface. Moreover, we found that the pattern of heterogeneity of cellular repolarization was not random but varied systematically across the epicardial surface (Fig 3⇑), forming uniform gradients of APD parallel to cardiac fibers. Therefore, regional APD differences were greatest in a direction where cell-to-cell coupling is ordinarily expected to minimize such differences. This finding further supports the existence of substantial heterogeneities in membrane ion kinetics across the epicardial surface. In contrast to known differences in the distribution of ion channels between the epicardial, midmyocardial, and endocardial ventricular walls,8 the extent to which the channels responsible for repolarization vary across the epicardial surface is less well understood. It is not likely that regional variation of IK1 or Ito8 33 can account for the heterogeneities of restitution we observed, since the activity of IK1 is uniformly distributed throughout the ventricular muscle34 and Ito is absent in guinea pig ventricle. On the other hand, regional heterogeneity of IK9 and its components, IKr and IKs,10 has been well documented in ventricular myocardium.
One possible mechanism for heterogeneity of restitution across the epicardial surface is spatial heterogeneities in the relative ratio of IKr to IKs channels. Restitution in regions having decreased IKr-to-IKs ratios can be predicted from the effects of selective pharmacological blockade of IKr,35 36 which produces significant APD prolongation at slow (baseline) rates but not at faster rates, at which IKr has relatively little influence on repolarization compared with IKs (ie, reverse use dependence37 ). Therefore, regions having decreased IKr-to-IKs ratios are expected to exhibit prolonged APDs at baseline in association with more rapid APD shortening during a premature stimulus (ie, a faster time course of restitution). In the present study, we also report a close quantitative relationship (Fig 3⇑) between APDs during baseline pacing (APDb) and the rate of restitution (RK), which is consistent with the hypothesis that epicardial heterogeneities of restitution are determined by regional variation of the IKr-to-IKs ratio. However, individual ionic currents were not measured and require further study to clarify the mechanism responsible for restitution heterogeneities on the epicardial surface of the heart. Furthermore, these mechanisms may not apply to other species in which Ito may play a more significant role in repolarization38 or in which only a single type of IK is present.39
Role of Conduction Versus APD in Dispersion of Repolarization
Our data indicate that the timing of repolarization at any ventricular site is determined by the time required for the cardiac impulse to propagate to and depolarize the site, plus the time course of local cellular repolarization (ie, as determined by APD restitution). Accordingly, spatial dispersion of repolarization is determined by spatial heterogeneity of propagation (ie, the pattern and velocity of conduction) and regional heterogeneity of APD. Using extracellular recording techniques, Han et al40 previously demonstrated marked increases in dispersion of recovery near the point of a prematurely stimulated impulse. Increased dispersion was attributed to conduction slowing near the site of pacing and was associated with decreased ventricular fibrillation threshold. Avitall et al41 have reasoned that conduction slowing–induced dispersion of repolarization may be responsible for latency of local electrograms recorded during closely coupled premature stimuli that initiate ventricular fibrillation. In contrast, by using high-resolution multisite optical action potential recordings, we discovered that the pattern of repolarization during a premature stimulus was influenced by changes in primary repolarization (ie, APD) much more so than conduction (Figs 7 and 8⇑⇑). For example, in Fig 5⇑ APD gradients are substantially altered (panel D) before any change in conduction occurs (panel C). Conventional extracellular recording techniques used previously could not distinguish the contributions of conduction and APD to dispersion of repolarization. In addition, by restricting their analysis to a single premature coupling interval, Han et al40 and others14 could not appreciate the extent to which repolarization gradients are dependent on the specific premature coupling interval tested (Figs 5 and 8⇑⇑).
In the present study, we limited our analysis to a uniform layer of epicardium on the anterior left ventricular surface. We made no attempt to measure heterogeneities of cellular repolarization elsewhere on the epicardial surface or in midmyocardial or endocardial muscle layers.8 Since we observed substantial changes in the morphology and polarity of the T wave of the ECG when APD gradient inversions were present within the mapping array (Fig 7⇑), APD gradients were probably modulated in many regions of the heart in a manner similar to those measured on the 1×1-cm surface mapped in these studies. In other words, it is highly unlikely that modulated repolarization is a phenomenon that is restricted to the region covered by our mapping array.
It is important to consider APD gradients between cells located on the epicardial and endocardial surfaces, which were recently demonstrated in guinea pig29 and dog.8 However, gradients of APD across the epicardial surface probably underwent the most significant modulation, since in the guinea pig, epicardial-to-endocardial APD differences are relatively small12 29 compared with the intraepicardial APD differences reported in the present study. Also, APD is randomly distributed across the endocardial surface and lacks the consistent gradients that we observed across the epicardium.29
The empirical rate constant, RK, used in the present study cannot distinguish between different shapes of restitution curves. However, the fact that RK is insensitive to subtle variations in the shape of restitution curves measured from different ventricular sites was a distinct advantage, since similar variations in the shape of restitution would substantially affect the exponential time constant of conventional single- or double-exponential models. It was critical for us to develop a restitution parameter that consistently reflected the time course of restitution measured from different ventricular sites independent of the precise shape of the restitution response. Therefore, although there are limitations in using RK to quantitatively characterize restitution in single cells, it proved extremely useful for measuring the distribution of restitution in a population of cells.
Since experiments were performed in nonworking Langendorff-perfused hearts, our results may not account for possible effects of mechanoelectrical feedback on APD.42 Therefore, the pattern and distribution of APD gradients reported in the present study may, in theory, change during mechanical loading of the left ventricle. However, Kanai and Salama29 recently reported a detailed analysis of action potential gradients in guinea pig ventricle under various states of mechanical stretch and found that stretch altered APD but had no significant effect on APD gradients across the epicardial surface. Therefore, mechanoelectrical feedback is not expected to substantially influence modulation of APD dispersion by premature stimuli as reported in the present study.
These data establish the presence of considerable heterogeneity in the kinetics of APD restitution across the epicardial surface of the heart. Moreover, there was a characteristic pattern in the spatial distribution of restitution properties such that faster restitution kinetics were closely associated with longer APDs during steady state baseline stimulation. This relationship accounted for the essential features associated with modulation of repolarization during a premature stimulus, including reversal of APD between cells located at different epicardial sites, inversion of APD gradients across the epicardial surface, and ECG T-wave inversions.
Wiggers and Wegria43 and Han and colleagues13 40 originally hypothesized that vulnerability to ventricular fibrillation is directly related to regional gradients (ie, dispersion) of repolarization. Therefore, we have demonstrated a response to premature stimulation that is expected to have important implications regarding the state of electrical instability in the heart. Repolarization gradients were significantly altered by a premature stimulus, and the precise pattern of APD gradients (ie, the major determinant of repolarization) changed systematically and predictably as a function of the premature coupling interval (Fig 8B⇑, solid line). Therefore, in addition to serving as a potential “trigger” for reentry, a premature stimulus may directly modulate the underlying electrophysiological substrate for reentry and ventricular fibrillation.
Based on the concept of “peeling back” refractoriness, vulnerability to reentry is thought to increase monotonically with greater degrees of extrastimulus prematurity.40 However, this notion only considers changes in mean refractoriness rather than changes in dispersion of refractoriness during a premature stimulus. Although refractoriness (ie, APD) did indeed shorten monotonically with the S1-S2 coupling interval, APD dispersion did not (Table⇑). Instead, we observed a biphasic response in which a premature stimulus attenuates dispersion over a broad range of premature coupling intervals and then dispersion rises sharply as the coupling interval is shortened beyond a critical value (Fig 8⇑). It is possible that the attenuation of dispersion by a premature stimulus may serve as a protective response against ventricular fibrillation in electrophysiologically normal myocardium. On the other hand, the rapid increase in dispersion at very short coupling intervals may explain why the initiation of ventricular fibrillation in normal hearts typically requires multiple closely coupled premature stimuli.
In addition to changes in the magnitude of dispersion, our results demonstrate significant changes in the orientation of repolarization gradients during a premature stimulus compared with baseline pacing (Figs 5, 6, and 7⇑⇑⇑). Kuo et al14 and Spach et al44 have previously emphasized the importance of repolarization gradient orientation as a potential determinant of vulnerability to reentry. We found that with tightly coupled premature stimuli, steep repolarization gradients form that are oriented away from the site of stimulation (Fig 7⇑). This pattern of dispersion has been shown to greatly enhance susceptibility to fibrillation.14 45 Further studies are required to confirm that modulation of the electrophysiological substrate by a premature stimulus influences the state of electrical vulnerability in the heart and therefore plays a role in the mechanism of cardiac arrhythmias.
Selected Abbreviations and Acronyms
|APD||=||action potential duration|
|APDb||=||APD of the baseline beat|
|APDp||=||APD of the premature beat|
|DTb||=||depolarization time of the baseline beat|
|DTp||=||depolarization time of the premature beat|
|IK||=||delayed rectifier K+ current|
|IK1||=||inward rectifier K+ current|
|IKr, IKs||=||rapidly and slowly activating components of IK|
|Ito||=||transient outward current|
|RK||=||rate constant of restitution|
|RTb||=||repolarization time of the baseline beat|
|RTp||=||repolarization time of the premature beat|
This study was supported by National Institutes of Health grant HL-54807, the Medical Research Service of the Department of Veterans Affairs, The Whitaker Foundation, and the American Heart Association, Northeast Ohio Affiliate, Inc. We are extremely grateful to Drs Yoram Rudy and Albert Waldo for their helpful advice.
Presented in part at the 15th Annual Scientific Sessions of the North American Society of Pacing and Electrophysiology, Nashville, Tenn, 1994.
- Received July 21, 1995.
- Accepted May 24, 1996.
Elharrar V, Surawicz B. Cycle length effect on restitution of action potential duration in dog cardiac fibers. Am J Physiol. 1983;244:H782-H792.
Franz MR, Swerdlow CD, Liem LB, Schaefer J. Cycle length dependence of human action potential duration in vivo. J Clin Invest. 1988;82:972-979.
Bass BG. Restitution of the action potential in cat papillary muscle. Am J Physiol. 1975;228:1717-1724.
Saitoh H, Bailey J, Surawicz B. Action potential duration alternans in dog Purkinje and ventricular muscle fibers. Circulation. 1989;80:1421-1431.
Antzelevitch C, Sicouri S, Litovsky SH, Lukas A, Krishnan SC, Di Diego JM, Gintant GA, Liu D-W. Heterogeneity within the ventricular wall: electrophysiology and pharmacology of epicardial, endocardial, and M cells. Circ Res. 1991;69:1427-1449.
Furukawa T, Kimura S, Furukawa N, Bassett AL, Myerburg RJ. Potassium rectifier currents differ in myocytes of endocardial and epicardial origin. Circ Res. 1992;70:91-103.
Liu D-W, Antzelevitch C. Characteristics of the delayed rectifier current (IKr and IKs) in canine ventricular epicardial, midmyocardial, and endocardial myocytes: a weaker IKs contributes to the longer action potential of the M cell. Circ Res. 1995;76:351-365.
Rosenbaum DS, Kaplan DT, Kanai A, Jackson L, Garan H, Cohen RJ, Salama G. Repolarization inhomogeneities in ventricular myocardium change dynamically with abrupt cycle length shortening. Circulation. 1991;84:1333-1345.
Watanabe T, Rautaharju P, McDonald T. Ventricular action potentials, ventricular extracellular potentials, and the ECG of guinea pig. Circ Res. 1985;57:362-373.
Han J, Moe G. Nonuniform recovery of excitability in ventricular muscle. Circ Res. 1964;14:44-60.
Kuo C, Munakata K, Reddy CP, Surawicz B. Characteristics and possible mechanisms of ventricular arrhythmia dependent on the dispersion of action potential durations. Circulation. 1983;67:1356-1357.
Girouard SD, Pastore JM, Laurita KR, Gregory KW, Rosenbaum DS. Optical mapping in a new guinea pig model of ventricular tachycardia reveals mechanisms for multiple wavelengths in a single reentrant circuit. Circulation. 1996;93:603-613.
Girouard SD, Laurita KR, Rosenbaum DS. Unique properties of cardiac action potentials recorded with voltage-sensitive dyes. J Cardiovasc Electrophysiol. In press.
Girouard S, Laurita K, Rosenbaum DS. Optical mapping can resolve propagation and recovery in the intact beating heart. Pacing Clin Electrophysiol. 1993;16:II-104. Abstract.
Tovar O, Amanna A, Milne K, Krauthamer V, Jones J. Effects of diacetyl monoxime on cardiac action potential duration and fibrillation cycle length. Pacing Clin Electrophysiol. 1994;17(pt II):824. Abstract.
Liu Y, Cabo C, Salomonsz R, Delmar M, Davidenko J, Jalife J. Effects of diacetyl monoxime on the electrical properties of sheep and guinea pig ventricular muscle. Cardiovasc Res. 1993;27:1991-1997.
Lesh MD, Pring M, Spear JF. Cellular uncoupling can unmask dispersion of action potential duration in ventricular myocardium: a computer modeling study. Circ Res. 1989;65:1426-1440.
Toyoshima H, Burgess MJ. Electrotonic interaction during canine ventricular repolarization. Circ Res. 1978;43:348-356.
Osaka T, Kodama I, Tsuboi N, Toyama J, Yamada K. Effects of activation sequence and anisotropic cellular geometry on repolarization phase of action potential of dog ventricular muscles. Circulation. 1987;76:226-236.
Efimov IR, Huang DT, Rendt JM, Salama G. Optical mapping of repolarization and refractoriness from intact hearts. Circulation. 1994;90:1469-1480.
Salama G, Rosenbaum DS, Kanai T, Cohen RJ, Kaplan DK. Measuring spatial inhomogeneities in action potential duration and repolarization using optical transmembrane potentials. IEEE Eng Med Biol Soc. 1989;1:222-223.
Kobayashi Y, Peters W, Khan SS, Mandel WJ, Karagueuzian HS. Cellular mechanisms of differential action potential duration restitution in canine ventricular muscle cells during single versus double premature stimuli. Circulation. 1992;86:955-967.
Laurita KR, Girouard SD, Rudy Y, Rosenbaum DS. Influence of electrotonus on repolarization during premature stimulation. Circulation. 1994;90(suppl I):I-412. Abstract.
Bazett HC. An analysis of the time relationship of electrocardiograms. Heart J. 1920;7:353-370.
Kanai A, Salama G. Optical mapping reveals that repolarization spreads anisotropically and is guided by fiber orientation in guinea pig hearts. Circ Res. 1995;77:784-802.
Salama G, Lombardi R, Elson J. Maps of optical action potentials and NADH fluorescence in intact working hearts. Am J Physiol. 1987;252:H384-H394.
Vassallo J, Cassidy D, Kindwall E, Marchlinski F, Josephson M. Nonuniform recovery of excitability in the left ventricle. Circulation. 1988;78:1365-1372.
Liu D-W, Gintant GA, Antzelevitch C. Ionic bases for electrophysiological distinctions among epicardial, midmyocardial, and endocardial myocytes from the free wall of the canine left ventricle. Circ Res. 1993;72:671-687.
Zeng J, Laurita KR, Rosenbaum DS, Rudy Y. Two components of the delayed rectifier K+ current in ventricular myocytes of the guinea pig type: theoretical formulation and their role in repolarization. Circ Res. 1995;77:140-152.
Jurkiewicz NK, Sanguinetti MC. Rate-dependent prolongation of cardiac action potentials by a methanesulfonanilide class III antiarrhythmic agent: specific block of rapidly activating delayed rectifier K+ current by dofetilide. Circ Res. 1993;72:75-83.
Carmeliet E. Voltage- and time-dependent block of delayed K+ current in cardiac myocytes by dofetilide. J Pharmacol Exp Ther. 1993;262:809-817.
Han J, de Jalon G, Moe GK. Fibrillation threshold of premature ventricular responses. Circ Res. 1966;18:18-25.
Avitall BJ, Mckinnie J, Jazayeri M, Akhtar M, Anderson AJ, Tchou P. Induction of ventricular fibrillation versus monomorphic ventricular tachycardia during programmed stimulation: role of premature beat conduction delay. Circulation. 1992;85:1271-1278.
Lab MJ. Contraction-excitation feedback in myocardium. Circ Res. 1982;50:757-766.
Wiggers CJ, Wegria R. Ventricular fibrillation due to single localized induction and condenser shocks applied during the vulnerable phase of ventricular systole. Am J Physiol. 1940;128:500-505.
Spach MS, Dolber PC, Heidlage JF. Interaction of inhomogeneities of repolarization with anisotropic propagation in dog atria: a mechanism for both preventing and initiating reentry. Circ Res. 1989;65:1612-1631.
Kuo C, Atarashi H, Reddy P, Surawicz B. Dispersion of ventricular repolarization and arrhythmia: study of two consecutive premature complexes. Circulation. 1985;72:370-376.