Regional Differences in Electrophysiological Properties of Epicardium, Midmyocardium, and Endocardium
In Vitro and In Vivo Correlations
Background Microelectrode studies have described a population of cells within the midmyocardium (M cells) displaying a steep rate dependence of action potential duration (APD) and high V̇max compared with endocardial (Endo) and epicardial (Epi) cells.
Methods and Results We studied repolarization in different myocardial layers in vitro and in situ. In addition to confirming the results of earlier studies, we found that after abrupt lengthening of the cycle length (CL), APDs in M cells reached a new steady state faster than in Epi or Endo cells: the time to achieve 90% of the difference in APD (t90) was 13.3±0.7 minutes in Endo cells, 12.8±1.1 minutes in Epi cells, and 2.6±0.4 minutes in M cells (P<.05 compared with Epi or Endo) when CL changed from 400 to 1000 ms. In situ, we registered activation-recovery intervals (ARIs) in bipolar electrograms obtained from different myocardial layers in conditions of AV block and His-bundle pacing. At all CLs from 300 to 2000 ms, ARIs were equal in all myocardial layers from Epi to Endo cells. Steady-state ARIs coincided with APD of M cells registered in vitro in the physiological range of CL from 300 to 700 ms. When CL was changed from 300 to 1000 ms, the ARI followed the rapid time course typical of M cells (t90=2.6±0.5, 2.2±0.4, 2.5±0.4, 2.6±0.5, and 2.3±0.4 minutes for Epi; 3-, 5-, and 7-mm sub-Epi; and Endo cells, respectively).
Conclusions In contrast to in vitro results, there is no significant difference in repolarization among myocardial layers in the intact normal canine heart.
Solberg et al1 described different electrophysiological characteristics of ventricular surface cells and cells in deep myocardial layers. Action potentials recorded from deep subendocardial cells of sliced dog papillary muscles had longer durations and significantly higher maximum rates of rise (V̇max) than did surface cells in standard epicardial and endocardial preparations. APD-rate relationships in cells from the deep myocardial zone were as steep as those in Purkinje fibers, but unlike Purkinje fibers, they did not exhibit phase 4 depolarization.
A recent detailed analysis of the unique population of M cells in the ventricular midmyocardium2 3 4 5 6 revealed electrophysiological features intermediate between those of myocardial and specialized conducting (Purkinje) fibers. M cells are found in the deep subepicardial to midmyocardial layers of the canine ventricular free wall and in the deep subendocardial layers of the septum, papillary muscles, and trabeculae. M cells display a spike-and-dome action potential morphology typical of epicardium, but V̇max is considerably greater than that of either epicardium or endocardium. The APD-rate relationship of M cells is considerably steeper than that of epicardium and endocardium and more akin to that of Purkinje fibers. Cell subtypes resembling canine epicardial, endocardial, and M cells have also been identified in human heart.7 Differences in action potential characteristics, similar to those described in syncytial preparations, have also been found in single myocytes isolated from the epicardium, M region, and endocardium of the canine ventricular free wall.8 Liu and Antzelevitch9 have shown that the slowly activating component of the delayed rectifier potassium current is smaller in M than in epicardial or endocardial cells, which might explain the distinctive repolarization features of M cells. Finally, agents that produce a marked APD prolongation, after-depolarizations, and triggered activity in Purkinje fibers induce increases in APD, after-depolarizations, and triggered activity in M but not in endocardial or epicardial cells.3 10 11
Important roles for M cells as determinants of electrocardiographic T waves, U waves, and QT intervals and as factors in the development of cardiac arrhythmias have been suggested.4 5 7 11 12 However, regional differences in the electrophysiological properties of myocardial layers have not been clearly demonstrated in vivo. In fact, Janse13 was unable to detect systematic differences among refractory period durations across the canine left ventricular wall in situ. Those experiments were performed at relatively short stimulus CLs (≤600 ms) at which, according to in vitro data, the difference in APD among myocardial layers is comparatively small.2 In light of the above, the present study was designed to compare the electrophysiological characteristics of different myocardial layers in tissues in vitro and in hearts in situ over a wide range of CLs. The intent was to determine whether, in the normal heart, the function of M cells is expressed as a marked transmural gradient for ventricular repolarization. A preliminary report of this work has appeared elsewhere.14
In Vitro Experiments
Mongrel dogs weighing 10 to 20 kg were anesthetized with sodium pentobarbital (30 mg/kg IV). Hearts were removed through a left lateral thoracotomy and immersed in cold Tyrode's solution equilibrated with 95% O2/5% CO2 and containing (in mmol/L) NaCl 131, NaHCO3 18, KCl 4, CaCl2 2.7, MgCl2 0.5, NaH2PO4 1.8, and dextrose 5.5. Endocardial, epicardial, and transmural strips (≈1.5×1.0×0.1 cm) were filleted with surgical blades either parallel or perpendicular (transmural) to the surface of the anterobasal left ventricular free wall.2 4 The preparations were placed in a tissue bath, superfused with Tyrode's solution warmed to 37°C (pH 7.35±0.05), and allowed to equilibrate at a CL of 2000 ms. Solutions were pumped through the bath at a rate of 12 mL/min, with chamber content changed three times per minute. The bath was connected to ground with a 3-mol/L KCl/Ag/AgCl junction. Experiments were not started until preparations had fully recovered and displayed stable electrophysiological characteristics, which required 2 to 3 hours for endocardial and transmural and 4 to 6 hours for epicardial strips.
Transmembrane potentials were recorded with 3 mol/L KCl–filled glass capillary microelectrodes (tip resistances, 10 to 20 MΩ) coupled by an Ag/AgCl junction to an amplifier with high-input impedance and capacity neutralization (model KS-700, World Precision Instruments). Action potentials and V̇max were displayed on a digital storage oscilloscope (model 4074, Gould) and stored in digitized form in a personal computer for consequent analysis. V̇max was obtained by electronic differentiation with an operational amplifier, and the system was calibrated as previously described.15 For stimulation of preparations, standard techniques were used to deliver 1- to 2-ms square-wave pulses 2.0 times the threshold through bipolar Teflon-coated silver electrodes.15
In Situ Experiments
Experiments were done in 14 dogs anesthetized with sodium pentobarbital (30 mg/kg IV). Additional smaller doses were given as needed to maintain deep anesthesia. Under controlled ventilation, a thoracotomy at the level of the fourth right intercostal space was performed, and the lateral surface of the right atrium and basal portions of the right ventricle were exposed. To achieve a slow spontaneous heart rate, the heart was positioned in a pericardial cradle, and complete heart block was produced by injection of 0.1 to 0.3 mL of 40% formaldehyde into the AV node.16 The thorax was closed by suture and reopened through the fourth left intercostal space, and the heart was cradled in the open pericardium. Intramural plunge electrodes and bipolar epicardial and endocardial electrodes were implanted in the region of the left ventricular outflow tract. The epicardial surface electrodes (1-mm spacing) were anchored with superficial stitches. Plunge recordings from the endocardium were made with two Teflon-coated stainless steel wires (0.1-mm diameter).17 The wires were passed through 22-gauge needles and bent back at the bevel of the needle to form small hooks. After the electrodes were plunged through the left ventricular wall into the ventricular cavity, the needle was removed, thus allowing the hooks to touch the endocardium. The endocardial recording electrodes were positioned opposite to the epicardial electrodes. Multipolar plunge electrodes (0.5-mm diameter) for intramural electrogram recording, each with 10 thin insulated tungsten wires 50 μm in diameter,18 were introduced as close to the epicardial electrodes as possible. To decrease the injury induced by the movement of myocardial layers during contraction, the shaft of each needle was polished, and care was taken to avoid protrusion of wire terminals. The first pole was 2.5 mm from the epicardial surface, and all recording points were 1 mm apart. For bipolar recordings, two neighboring poles were used, permitting registration of five intramural bipolar electrograms from average depths of 3, 5, 7, 9, and 11 mm (measured from the epicardium). All steady-state recordings were made at least 30 minutes after placement of the electrodes. Electrical signals from seven amplifiers (0.5- to 1000-Hz bandwidth), together with the surface ECG (lead II), were digitized with an analog-to-digital convertor (D-210, DATAQ Instruments Inc) and stored in a personal computer for further analysis. Then the signals were differentiated, and the activation-recovery intervals were measured from the steepest portion of the initial deflection of the QRS complex to the steepest portion of the terminal phase of the T wave. This would, of course, ignore any low-amplitude late events that might splay the end of the T wave. A strong correlation has been demonstrated between the activation-recovery intervals measured this way and the refractory periods19 or transmembrane APD.20 In several experiments, we also measured activation-recovery intervals from the beginning of the QRS complex to the end of the T wave. The resultant small systematic shift seen in all values did not significantly affect the essential conclusions of this study, which are based on the measurement of the activation-recovery intervals.
Hearts were stimulated by delivering 2-ms square-wave pulses 2.0 times the threshold through bipolar plunge electrodes (the same type used for endocardial electrogram recording) introduced into the middle part of the ventricular septum (His-bundle region) through the right ventricular wall. A femoral vein was cannulated for administration of supplemental doses of pentobarbital, and a femoral artery was cannulated for monitoring blood pressure by standard techniques.
All data are presented as mean±SEM. The statistical technique was ANOVA, and Bonferroni's test was used when the F value permitted. Significance was determined at P<.05.
In Vitro Experiments
The transmembrane potentials of cells in transmural slices of the left ventricle M cells were different from those of surface epicardial (or endocardial) cells. At a CL of 2000 ms, M cells generated action potentials with significantly longer plateaus and more rapid upstroke velocities than those observed in epicardium or endocardium (Fig 1A⇓ and the Table⇓). Also, midmyocardial (like epicardial but not endocardial) action potentials had a notch at the onset of the plateau: phase 1 amplitude (measured at the nadir of the notch) was significantly lower in midmyocardial compared with endocardial action potentials. Finally, the maximum diastolic potential in M cells was more negative than in epicardial or endocardial cells. Hence, the differences in the action potential parameters of endocardial, epicardial, and M cells observed in our experiments were practically identical to those described by Antzelevitch and coworkers.2 3 4
Steady-state rate dependence of the APD in M cells qualitatively resembled that in epicardial and endocardial cells; ie, the action potential became longer when the CL was increased from 400 to 8000 ms (Fig 1B⇑). However, similar to the report of Sicouri and Antzelevitch,6 the rate dependence of APD of M cells was considerably steeper than that of epicardial or endocardial cells: the midmyocardial action potentials remained significantly longer than epicardial or endocardial action potentials at all rates of stimulation, but this difference diminished when the CL was reduced to 400 ms. There were no significant differences in steady-state APD at all stimulation rates between endocardial and epicardial cell preparations.
In contrast to Antzelevitch and colleagues,2 3 4 who found cells with long APDs only in the deep subepicardial to midmyocardial layers of the canine left ventricular free wall, we registered action potentials typical of M cells in practically all loci of transmural slabs (Fig 2A⇓). We were not able to impale the cells on the epicardial and endocardial sides of transmural slabs; therefore, the electrical activity of epicardial and endocardial cells was recorded from standard slabs cut parallel to the ventricular surface. Distribution of V̇max across the left ventricular wall in our study was very similar to the distribution of APD90 (Fig 2B⇓) and essentially coincided with the results reported by Sicouri and Antzelevitch.6
That injury inflicted by cutting the tissue does not account for specific transmembrane potentials recorded in transmural slices is suggested by the results in Fig 3⇓. This is a typical recording obtained from endocardial (ie, not transmural) slabs with an obliquely cut edge. The impalement sites on the cut edges were at depths ≤0.2 mm, permitting us to compare subendocardial cells impaled here with those on the endocardial surface. Note that nearly identical action potentials were recorded from intact and cut surfaces at CLs of 400 to 4000 ms.
We were concerned about the effect of the passage of time on the action potentials of the cells studied. Therefore, we determined the evolution of APD after abrupt changes in CL in all three types of tissue. For this purpose, preparations were stimulated for 5 minutes at a CL of 400 ms, and then the CL was abruptly increased to 1000, 2000, 4000, or 8000 ms for 30 minutes. As Fig 4A⇓ shows, the pattern of action potential changes in cells of transmural slices was qualitatively different from that in cells of epicardial or endocardial slabs; ie, when the CL increased abruptly from 400 to 8000 ms, epicardial and endocardial action potentials lengthened monotonically and attained a new steady state in 20 to 30 minutes. In contrast, the duration of midmyocardial action potentials prolonged and attained a maximum value in about 3 minutes, after which the duration shortened until a new steady state was reached at 20 to 30 minutes.
Fig 4B⇑ further illustrates two major differences between M and epicardial or endocardial cells with respect to the time course of APD changes induced by abrupt transitions from a CL of 400 ms to all CLs studied. First, at relatively long CLs (4000 and 8000 ms), APD changes were biphasic in M cells and monotonic in epicardial and endocardial cells. Second, at CLs of 1000 and 2000 ms, the changes in APD in M cells became monotonic but proceeded much more rapidly than in epicardial or endocardial cells. Fig 5⇓ presents averaged plots of changes in APD induced by abrupt changes in CL from 400 to 1000 ms (Fig 5A⇓) and to 8000 ms (Fig 5B⇓) in M, epicardial, and endocardial cells. At a CL of 1000 ms, the t90 between APDs at 400 and 1000 ms (Fig 5A⇓) was much longer in epicardium (12.8±1.1 minutes) and endocardium (13.3±0.7 minutes) than in midmyocardium (2.6±0.4 minutes; P<.05 compared with endocardium and epicardium). More marked differences were observed after abrupt lengthening of the CL to 8000 ms (Fig 5B⇓). Here, in contrast to a monotonic increase in APD always observed in epicardium and endocardium, the APD in M cells first reached a peak of 487±62 ms at 3.5±0.4 minutes and then slowly decayed to its steady-state level.
Thus, in addition to the distinctive properties of M cells described earlier in vitro (steep APD-rate relationship, high V̇max), we have found that the recovery of APD after an abrupt slowing of stimulation rate in M cells differs from that in epicardium and endocardium.
In Situ Experiments
Our goal was to determine whether the differences in electrophysiological properties between intramural and epicardial or endocardial cells registered in vitro are seen in situ. In vitro results show that these differences are more prominent at slower rates of stimulation. In 11 dogs with complete heart block, idioventricular CL was 1615±86 ms, and ventricular stimulation at a CL of 1000 ms could be performed. In 5 of these dogs, idioventricular rhythms were sufficiently slow to permit stimulation of the ventricles at a CL of 2000 ms, and in 1 dog we could stimulate the ventricles at a CL of 4000 ms without a dramatic fall in blood pressure.
Fig 6⇓ shows representative electrograms at CLs of 2000 and 4000 ms. No significant differences in activation-recovery intervals from different myocardial layers were seen; even at a CL of 4000 ms, the variation in activation-recovery intervals did not exceed 20 ms. Fig 7A⇓ shows the distribution of conduction times across the left ventricular wall. These increased significantly from endocardium to epicardium.
We found no significant differences among activation-recovery intervals measured at all depths during stimulation at all CLs (Fig 7B⇑). All measurements in Fig 7B⇑ were made 30 minutes after placement of the multipolar electrodes. In contrast, at earlier times we observed distortions of activation-recovery intervals that subsided by 15 to 20 minutes (Fig 8⇓). In different experiments, these distortions were observed in electrograms recorded from different myocardial layers. Fig 9⇓ summarizes the dependence of activation-recovery intervals on CL. All intervals reveal obvious rate dependence: they prolong with an increase in CL. Comparison with the in vitro data (dotted lines) demonstrates a close fit of the activation-recovery interval with the APD90 of M cells at CLs of 400 and 700 ms, whereas at longer CLs, the values of activation-recovery intervals in situ are intermediate between midmyocardial and epicardial (or endocardial) APDs.
To determine the temporal course of changes of activation-recovery intervals after abrupt slowing of stimulation rate, a protocol similar to that used in vitro was applied: the heart was stimulated at a CL of 300 ms for 5 minutes, and then the CL was abruptly increased to 1000 ms. In contrast to in vitro results, the time course of changes of activation-recovery intervals was practically the same on ECG and on surface and intramural electrograms (Fig 10⇓). The t90 values did not differ significantly in these recordings and were 2.3±0.4, 2.6±0.5, 2.3±0.4, 2.2±0.4, 2.5±0.4, and 2.6±0.5 minutes on ECG; surface epicardial; surface endocardial; and 3-, 5-, and 7-mm-deep intramural bipolar electrograms, respectively. These t90 values are very close to the t90 of M cells in vitro and are much shorter than the corresponding values for epicardium and endocardium found in vitro.
Thus, in contrast to the experiments with isolated ventricular preparations, we observed no differences among different myocardial layers with respect to either the values of activation-recovery intervals or the velocity of changes in these intervals after abrupt slowing of heart rate.
Our results from in vitro experiments are generally consistent with the results of Antzelevitch and coworkers2 3 5 6 with respect to the APD: the M cells have longer APDs than epicardial and endocardial cells, and this difference increases with prolongation of the CL. There are some quantitative differences between the results of Antzelevitch and colleagues and our results that can be explained as follows. In our experiments, steady-state data were collected 30 minutes after each change in CL, whereas in the experiments of Sicouri and Antzelevitch,6 data were recorded 60 to 90 seconds after each change. As our experiments show (see Fig 4⇑), this short time does not permit attainment of steady-state conditions.
There is another difference between our results and those of Antzelevitch et al2 6 : at low stimulation rates, they found cells with long APDs in canine left ventricle at depths from 1 mm to 5 to 7 mm from the epicardial surface. In our experiments, these cells could be demonstrated throughout the transmural slab (in vitro) or the ventricular wall (in vivo). In complete agreement with Sicouri and Antzelevitch,6 we found cells with significantly higher V̇max (the second major distinguishing feature of M cells) throughout the entire transmural slab. These results suggest that at the sites we studied, nearly all myocardial cells except for the superficial epicardial and endocardial cells have characteristics ascribed to M cells.
We noted a new feature distinguishing surface epicardial and endocardial cells from M cells in vitro, ie, different patterns of APD recovery after abrupt slowing of stimulation rate. Specifically, in the physiologically relevant range of stimulation rates (CLs of 400 ms to 1000 to 2000 ms), the transition of APD to a new steady state occurred significantly faster in M cells than in epicardial or endocardial cells. These different recovery rates may have significance in pathological settings for the following reasons: First, they suggest that after abrupt, marked slowing of rate, there can be even greater dispersion among midmyocardium, epicardium, and endocardium than might otherwise be expected (see Figs 4 and 5⇑⇑). Second, this dispersion is unstable, varying over several minutes. Moreover, the dispersion is greatest when the tissue sites studied are mechanically uncoupled from one another; ie, when epicardial and endocardial slabs are not in continuity with midmyocardium. This point is best illustrated by a comparison of the marked transmyocardial gradients in our isolated tissue studies with the minimal transmyocardial gradients in our in vivo work and that of others in normal hearts.13 21
A curious phenomenon was observed when the stimulation rate was abruptly made unphysiologically low (CL of 4000 or 8000 ms); unlike epicardial and endocardial cells, M cells showed a biphasic pattern of APD recovery with characteristic “humps” (see Figs 4 and 5B⇑⇑). Unpublished data kindly shared with us by Dr Charles Antzelevitch are consistent with our findings here and moreover indicate that the magnitude of these humps is Ca2+ dependent: the higher the [Ca2+]o, the greater the magnitude. This suggests that in settings of Ca2+ overload, the difference between midmyocardium and epicardium and endocardium may be further exaggerated. The experiment in Fig 8⇑ provides an independent indicator of the extent to which myocardial injury may in fact permit dispersion of repolarization to be manifested transmurally in vivo.
On the basis of our in vitro observations, we evaluated APD-rate relationships and the time course of the APD change after abrupt slowing of the stimulation rate in the intact heart experiments. However, two major technical questions arose in the intact heart experiment. The first was how to measure the APDs of surface and intramural cells in situ. It has been shown that the maximum first derivative of the T wave of the local electrogram corresponds to the time of the maximum rate of change of voltage during phase 3 of the action potential.20 Moreover, theoretical and simulation20 22 as well as experimental23 24 studies have demonstrated that the activation-recovery interval measured as the time between the first peak of the derivative of the QRS and maximum derivative of the T wave in local electrograms is a good estimate of APD independent of T-wave waveform, drive site, or the presence of a localized steep repolarization gradient. The second technical question was how to control the ventricular rate. Because our experiments in vitro showed that the longer the CL, the more the M cells differed from epicardial and endocardial cells with respect to APD, we had to slow ventricular rate maximally. This could not be done by selective suppression of the sinus node because of the high rate of the canine AV junctional rhythm (88±3 bpm).25 Only complete AV block could slow heart rate sufficiently to permit sustained stimulation of the ventricles at long (1000 or 2000 ms) CLs.26
In our in situ experiments, we found no differences in duration of activation-recovery intervals registered from different myocardial layers at all stimulus CLs. This led us to conclude that in the intact heart, there is little heterogeneity in APD distribution across the free wall of the canine left ventricle, near the base. This observation is surprising because according to Sicouri and Antzelevitch2 6 and our own experiments in vitro, the difference in APDs between epicardial or endocardial and M cells should have been 90 to 100 ms at a CL of 2000 ms and 110 to 150 ms at a CL of 4000 ms. In contrast to our results, abstracts by Wang et al27 and Hariman et al28 showed a difference in activation-recovery time between epicardial and intramural (1 to 4 mm deep to the epicardium) layers of canine left ventricle. We suggest that in these studies, injury inflicted by specially constructed electrodes with protruding poles (used to register monophasic action potentials) resulted in the APD variance registered at different depths under the epicardium. In fact, in only one of these reports27 was a relatively long APD found, at a depth of 9 mm. Moreover, the second report28 demonstrated a systematic shift of about 50 ms in all APD values at the same CL of ventricular pacing as used in the first report. In our study using smooth-surface electrodes, we did note transient distortions in electrogram durations for several minutes after insertion, as shown in Fig 8⇑. These probably reflect an injury effect that required time for resolution.
In contrast to our in vitro experiments, we found no differences across the wall of the left ventricle in vivo in the time course of activation-recovery intervals. In all myocardial layers, the time to attain a new steady state after abrupt slowing of heart rate coincided with the time registered for M cells in vitro. Thus, our results show that the behavior of myocardial activation-recovery intervals in the intact heart manifests a minimal gradient. The characteristics of the component cell layers correspond to those of M cells in vitro rather than to epicardial or endocardial cells in that (1) in the physiological range of heart rates, absolute values of activation-recovery intervals in all myocardial layers are closer to the APD of M cells in vitro; (2) the transition rates for all activation-recovery intervals after abrupt increases in CL coincide with the values typical for M cells in vitro; and (3) the flat distribution of APD in transmural slabs (from epicardium to endocardium) corresponds to the flat distribution of activation-recovery intervals across the left ventricular wall in situ.
An interesting question arises from these results: Why is the difference in electrophysiological properties between surface and transmural cells of canine left ventricle observed in vitro but not in the intact heart? We can suggest two answers. The first derives from the explanation of the electrophysiological differences between surface and M cells in vitro proposed by Solberg et al.1 They speculated that these differences could result from differences in local environmental conditions. A relatively large amount of connective tissue on the endocardial and epicardial surfaces may diminish perfusion and impair elimination of metabolic waste products from subendocardial and subepicardial cells. According to this explanation, the electrophysiological differences between surface and M cells observed in vitro are an artifact resulting from underperfusion of subendocardial and subepicardial cells. If this is true, transmural slices might be preferable for in vitro experiments requiring multicellular preparations. The second possible explanation of the differences between in vitro and in situ results is that the electrophysiological properties of cells in thin epicardial and endocardial layers really do differ from those of M cells. However, in situ, the cells in these thin surfaces are extensively coupled electrically with M cells and therefore are electrotonically influenced by M cells to the extent that the differences in APD among all cell layers are minimized. In contrast, in vitro the cells in standard epicardial and endocardial slabs are disconnected from midmyocardium and thus express their intrinsic properties. Certainly, results obtained with myocytes isolated from the epicardium, midmyocardial region, and endocardium would support the latter speculation in that the APD-rate relationship is steeper in cells from the midmyocardial region.8 9 However, these data are not completely conclusive because of the overlap of APD recorded from cells isolated from different myocardial layers.9
Finally, although we found no heterogeneity in repolarization across the canine left ventricular wall in situ, our data are referable only to the normal heart. They certainly do not exclude the possibility that heterogeneity may occur in settings such as uncoupling, injury, or drug administration because cells from different myocardial locations may have different sensitivities to these interventions.29 30 31 32
Selected Abbreviations and Acronyms
|APD||=||action potential duration|
|APD90||=||action potential duration at 90% repolarization|
|t90||=||time to achieve 90% of the difference in APDs|
These studies were supported in part by USPHS-NHLBI grant HL-28958; in part by Helopharm, AG; and in part by the Wild Wings Foundation. We express our gratitude to Drs Charles Antzelevitch and Vladislav Nesterenko for their review of and constructive discussion concerning the manuscript. We also thank Dr Natalia Egorova for assisting in the performance of the experiments and Eileen Franey for her careful attention to the preparation of the manuscript.
- Received January 11, 1996.
- Revision received April 22, 1996.
- Accepted May 6, 1996.
- Copyright © 1996 by American Heart Association
Solberg LE, Singer DH, Ten Eick RE, Duffin EG Jr. Glass microelectrode studies on intramural papillary muscle cells. Circ Res. 1974;34:783-796.
Sicouri S, Antzelevitch C. A subpopulation of cells with unique electrophysiological properties in the deep subepicardium of the canine ventricle. Circ Res. 1991;68:1729-1741.
Antzelevitch C, Sicouri S, Litovsky S, Lukas A, Krishnan SC, DiDiego JM, Gintant GA, Liu D-W. Heterogeneity within the ventricular wall. Circ Res. 1991;69:1427-1448.
Antzelevitch C, Sicouri S, Lukas A, Nesterenko VV, Liu D-W, DiDiego JM. Regional differences in the electrophysiology of ventricular cells: physiological and clinical implications. In: Zipes DP, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. 2nd ed. New York, NY; WB Saunders Co; 1994:228-245.
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.
Liu D-W, Antzelevitch C. Characteristics of the delayed rectifier current (Ikr and I ks) in canine ventricular epicardial, midmyocardial, and endocardial myocytes. Circ Res. 1995;76:351-365.
Janse MJ. The Effect of Changes in Heart Rate on the Refractory Period of the Heart. Amsterdam, Netherlands: Mondeel-Offset-Drukkerij; 1971:41-42. Thesis.
Sosunov EA, Anyukhovsky EP, Rosen MR. Comparison of repolarization of cells from different layers of myocardium in vitro and in vivo. Biophys J. 1996;70(pt 2):A276. Abstract.
Rosen MR, Merker C, Gelbland H, Hoffman BF. Effects of procainamide on the electrophysiological properties of the canine ventricular conducting system. J Pharmacol Exp Ther. 1973;185:438-446.
Scherlag BJ, Kosowsky BD, Damato AN. A technique for ventricular pacing from the His bundle of the intact heart. J Appl Physiol. 1967;22:584-587.
Scherlag BJ, El-Sherif N, Hope R, Lazzara R. Characterization and localization of ventricular arrhythmias resulting from myocardial ischemia and infarction. Circ Res. 1974;35:372-383.
Spach MS, Barr RC. Ventricular intramural and epicardial potential distributions during ventricular activation and repolarization in the intact dog. Circ Res. 1975;37:243-257.
Millar CK, Kralios FA, Lux RL. Correlation between refractory periods and activation-recovery intervals from electrograms: effects of rate and adrenergic interventions. Circulation. 1985;72:1372-1379.
Haws CW, Lux RL. Correlation between in vivo transmembrane action potential durations and activation-recovery intervals from electrograms. Circulation. 1990;81:281-288.
Freigang KD, Becker R, Bauer A, Voss F, Senges J, Brachman J, Schoels W. Electrophysiological properties of individual muscle layers in the in vivo canine heart. J Am Coll Cardiol. 1996;27(suppl A):124A. Abstract.
Steinhaus BM. Estimating cardiac transmembrane activation and recovery times from unipolar and bipolar extracellular electrograms: a simulation study. Circ Res. 1989;64:449-462.
Wyatt RF. Comparison of estimates of activation and recovery times from bipolar and unipolar electrograms to in vivo transmembrane action potential duration. Proc IEEE/Eng Med Biol Soc. 1980;2:22-25.
Wyatt RF, Burgess MJ, Evans AK, Lux RL, Abildskov JA. Estimation of ventricular transmembrane action potential durations and repolarization times from unipolar electrograms. Am J Cardiol. 1981;47:488. Abstract.
Wang Y, Hariman RJ, Gintant GA, Louie EK, Hwang MH, Loeb HS. A method for recording of intramyocardial monophasic action potential in intact dogs: in vivo evidence of M cells. PACE Pacing Clin Electrophysiol. 1992;15:559. Abstract.
Hariman RJ, Wang Y, Sharif MA, Louie EK, Hwang MH, Loeb HS, Scanlon PJ. Early afterdepolarizations arising from M cells in intact dogs. Circulation. 1992;86(suppl I):I-300. Abstract.
Weissenburger J, Nesterenko VV, Antzelevitch C. Intramural monophasic action potentials (MAP) display steeper APD-rate relations and higher sensitivity to class III agents than epicardial and endocardial MAPs: characteristics of the M cell in vivo. Circulation. 1995;92(suppl I):I-300. Abstract.
Lukas A, Antzelevitch C. Differences in the electrophysiological response of canine ventricular epicardium and endocardium to ischemia: role of the transient outward current. Circulation. 1993;88:2903-2915.
Yan G-X, Antzelevitch C. Contribution of M cells to the electrocardiographic U wave: direct evidence from arterially perfused canine left ventricle. PACE. 1995;18(suppl II):II-830. Abstract.
Burashnikov A, Antzelevitch C. Alpha-agonists produce opposite effect on action potential duration in Purkinje and M cells isolated from the canine left ventricle. PACE. 1995;18(suppl II):II-3. Abstract.