INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

CONDUCTION VELOCITY (CV) and safety of wave-front propagation depend on both the active membrane properties of each cell of the conducting pathway and of the cell-to-cell gap junctional conductance (coupling) (13, 24). The importance of junctional conductance in altering conduction and promoting conduction block and ventricular arrhythmias was emphasized in experimental (2, 4, 11, 13, 24) and simulation (12, 13, 22) studies. Recent studies (8) using transgenic mice with gap junctional connexin40 (Cx40) deficiency showed that burst atrial pacing in these mice can induce atrial conduction disturbances and atrial tachyarrhythmia. At present, the roles of selective and partial atrial gap junctional uncoupling in inducing atrial tachycardia (AT) and atrial fibrillation (AF) in larger animals remain poorly explored. It has been suggested that aging might be associated with increased cellular uncoupling (24), a phenomenon that might contribute to the increased incidence of AF in the elderly (3). We hypothesized that partial cellular uncoupling produced by heptanol at concentrations that do not alter the active membrane properties promotes wave break of the activation wave front during rapid pacing, leading to AF.

	METHODS AND MATERIALS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Isolated left atrial tissue preparation. The research protocol was approved by the Institutional Animal Care and Use Committee of Cedars-Sinai Medical Center and followed the guidelines of the American Heart Association. Twelve mongrel dogs of either sex weighing between 22 and 28 kg were anesthetized with intravenous pentobarbital sodium (35 mg/kg). The left atrium (LA) was isolated, cannulated, and perfused at a flow rate of 15 ml/min with 37°C oxygenated Tyrode solution as described previously (31). Two pairs of bipolar electrodes (one for pacing and one for recording) were placed 1.5-2 cm apart. In addition, two electrodes were placed at the edges of the tissue 4 cm apart from each other to record global atrial rhythm ("pseudoelectrocardiogram") (31).

Test of AF vulnerability: effects of heptanol. Development of conduction disturbances and changes in the vulnerability to AF were tested by rapid atrial pacing at 50 ms cycle length (CL) for 3 s by using 5-ms pulses of twice diastolic current threshold. Five attempts of AF induction were tested at baseline in each of the 12 isolated tissues (60 trials), and five attempts were tested after 2 µM heptanol infusion (60 trials). Because AF could be induced only with 2 µM heptanol (see RESULTS), the effects of 1, 5, 10, and 50 µM heptanol were tested in only five tissues. If the induced AF was sustained for >5 min, it was cardioverted with an external defibrillator (HVS-02, Ventritex). During rapid pacing at 50 ms CL, the mean shortest captured CL was determined before and after heptanol perfusion with a pair of recording bipolar electrodes located 1-2 cm away from the pacing electrode. Reversibility of the the effects of heptanol was evaluated after 30 min of drug-free Tyrode solution perfusion (n = 6).

Optical mapping. Wave-front dynamics during pacing and AF were analyzed using an optical mapping system. The LA was stained for 20 min with di-4-ANEPPS (10 µM). The patterns of activation were acquired with a charge-coupled device (CCD) camera with a sampling interval of 2.3 ms for a duration of 2.3 s in a recording field of ~28 by 28 mm with 128 by 128 pixel resolution. The optical mapping system is similar to the one described previously (18, 19). A solid laser with a wavelength of 532 nm was used for the light source, and the induced fluorescence was collected with 12-bit digital CCD camera. The fluorescent signals were baseline subtracted and inverted. The signals of each pixel were then spatially averaged with the signals of eight neighboring pixels to reduce noise. The computer first finds every adjacent pair of pixels in the frame that cross the average value of the data. If the intensity of the data on which the line coincides is increasing, that edge is identified as the wave front and is colored red. The color then progressively changes to yellow, green, light blue, and finally dark blue to represent increasing states of repolarization (Fig. 1). Isochronal activation maps were also constructed during pacing before and after heptanol exposure (26).

View larger version (49K): [in this window] [in a new window]   Fig. 1.   Snapshots and isochronal optical mapping before (Control) and after heptanol (2 µM) during 400- (A) and 130-ms (B) pacing cycle lengths (PCL) in an isolated canine left atrium. Heptanol causes conduction slowing at 400-ms PCL and conduction block across the fiber during 130-ms PCL (second paced beat). Bottom left and right show the color code for the snap shots (first 3 frames in each row) and isochronal maps (last frames in each row). Double-headed arrow shows the fiber orientation. *Pacing site. Number under each snapshot is time in milliseconds with time 0 denoting the onset of pacing. Note rate-dependent conduction slowing after heptanol on the snapshots and on the isochronal maps.

Conduction velocity. CV was measured over the left epicardial appendage along the fiber during atrial pacing at different pacing CLs (PCLs, 600-200 ms) and during AF. The CV during pacing was calculated as the ratio of the distance traveled by the wave front to the time it took to travel between the pacing and the recording electrodes, which were 1.5-2 cm apart from each other (5). The CV during AF was determined by the ratio of the distance traveled continuously by the wave front to the time it took to travel the distance (i.e., 0.5 cm to 1.5 cm).

Transmembrane action potential recording. Transmembrane action potential (TMPs) were recorded with glass microelectrodes from the superficial cell layers (2-4 cells deep) of the epicardium of the LA free wall. The TMP data were continuously acquired using Axoscope software, a Digidata acquisition system (Axon Instruments) with a sampling interval of 200 µs and 14-bit resolution (14). The maximal voltage change over time (dV/dt_max) at PCLs of 400, 300, and 200 ms were measured from the digitized TMP data. Dynamic APD restitution curves (15) were then constructed during progressively shorter PCLs starting from 300 ms and decreasing by 10-ms increments (at PCL >190 ms) and by 2-ms increments (at PCL <80 ms) until block. The shortest 1:1 captured PCL was recorded before and after perfusion with 1-50 µM heptanol. The action potential duration (APD) repolarization to 90% (APD₉₀) against its previous diastolic interval was then plotted and fitted by a single exponential curve (Microcal, Origin 5.0).

Statistical analysis. All statistical analyses were performed using GB-STAT (7) Student's t-test for single comparison. ANOVA for multiple comparisons was used to compare the means. A Chi-square test was used to detect differences in the occurrence of sustained AF. The null hypothesis was rejected at a value of P  0.05. Results are expressed as means ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Effect of heptanol on atrial capture rate. At 1 µM, heptanol had no effect on the fastest 1:1 captured atrial PCL (111 ± 2.3 ms). However, when heptanol concentration was >1 µM, a concentration-dependent loss of regular 1:1 atrial capture developed at significantly (P < 0.01) longer PCL. At baseline, the shortest regular 1:1 atrial captured PCL was 111 ± 2 ms, which increased to 127 ± 27, 146 ± 34, 168 ± 40, and 182 ± 46 ms during perfusion with 2, 5, 10, and 50 µM heptanol, respectively (P < 0.01 for all comparisons). The effects of heptanol were reversible. Within 30 min of drug-free Tyrode perfusion, the atria could be captured at the same pre-heptanol shortest CLs (113 ± 2.8 vs. 111 ± 2.3 ms, P = not significant).

Effects of heptanol on CV. Heptanol at 2 µM but not at 1 µM caused significant (P < 0.001) CV slowing along the fiber at all three (600, 400, and 200 ms) PCLs tested. At baseline, the CV at these three PCLs were 64.2 ± 2.7, 53.8 ± 4.7, and 38.5 ± 4.5 cm/s, respectively, and decreased to 38.8 ± 0.9, 29.9 ± 3.3, 23.1 ± 2.4 cm/s, respectively, after heptanol (P < 0.05 for all comparisons). Figure 1 is a representative optical map showing conduction slowing after 2 µM heptanol perfusion at two different PCLs. At baseline, the wave front propagated with relative uniformity and without conduction slowing both at 400 and 130 ms PCL. However, after heptanol and during faster rates of pacing (CL < 200 ms), the CV became severely slowed and nonuniform (Fig. 1). When the CV was severely slowed, the propagation of the wave front became highly nonuniform, preventing its systematic and accurate measurement at pacing CLs <200 ms.

Induction of AF. At baseline, rapid pacing induced short runs of repetitive atrial activity that lasted 2.1 ± 1.5 s in 7 of 12 isolated atria (Fig. 2). No sustained AF (>30 s) could be induced at baseline in any of the 12 tissues studied. Perfusion with 1 µM heptanol did not change vulnerability to inducible AF. However, after 2 µM heptanol, sustained AF (>30 s) was induced in 9 of 12 tissues (P < 0.001). In 8 of 12 tissues, the induced AF lasted >5 min, requiring electrical shocks for termination. Shocks were applied through a pair of coil electrodes placed in the tissue bath at the edges of the isolated tissues. Figure 2 illustrates examples of sustained AF induced after 2 µM heptanol and no arrhythmia or nonsustained AF induced at baseline. In addition to sustained AF, nonsustained AF lasting on the average for 7.2 ± 4.5 s (P < 0.05 compared with baseline) could also be induced in all 12 tissues after 2 µM heptanol perfusion.

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Effects of heptanol on atrial rhythm during rapid atrial pacing. A: at baseline (control) no arrhythmia is induced. However, after 2 µM of heptanol, atrial fibrillation (AF) is induced. B: conversion of nonsustained AF (2.2 s) induced at baseline (control) to sustained AF after 2 µM heptanol. Stim, stimulus artifact; ECG, pseudoelectrocardiogram; BEg, bipolar atrial electrogram.

Wave-front dynamics during induced AF. There were significant differences between the characteristics of AF induced at baseline and AF induced during heptanol perfusion. The CL of the induced AF was significantly shorter at baseline than after 2 µM heptanol perfusion (114 ± 8 vs. 128 ± 12 ms for AF CL, P < 0.01). The AF, at baseline and during heptanol perfusion, was characterized by the presence of multiple independent wave fronts with a characteristic fibrillation-like electrogram both on the pseudoelectrocardiogram and the local atrial electrogram (Fig. 2). The AF was maintained by continuous wave-front breakups, providing a source of multiple independent wave fronts. The average number of independent wave fronts (i.e., number of wave fronts propagating in different directions per second in the optical field of view) during AF was significantly (P < 0.01) smaller at baseline (1.6 ± 0.3) than after heptanol (2.2 ± 0.4). In 26 episodes of heptanol-induced sustained AF, 69 events of spontaneous wavebreaks were captured and analyzed. Of these 69 events, 46 episodes of wave breaks developed across and 23 along the fiber. Block across the fiber developed at significantly longer coupling intervals (i.e., the interval between the previous activation and the activation causing wave break) than those along the fiber orientation (102 ± 24 vs. 74 ± 15 ms; P < 0.001). During the induced AF, the incidence of wave breaks in the optical field of view was significantly higher in the presence of heptanol than at baseline (3.8 ± 1.6 vs. 6.3 ± 2.1 s¹, respectively, P < 0.001). During the induced AF, the average CV of the wavelets was significantly (P < 0.01) faster at baseline than during 2 µM heptanol perfusion (27.3 ± 8.6 vs. 19.3 ± 10.2 cm/s, respectively).

Mechanism of heptanol-induced AF: conduction block and reentry. Figure 3 illustrates activation pattern at baseline during 2:1 atrial capture at a stimulating CL of 50 ms CL (i.e., captured CL = 100 ms). At baseline, the wave front propagated with a relatively uniform pattern without block during the entire pacing sequence (only the first two beats are shown). However, after 2 µM heptanol, at 50 ms CL of stimulation, the first captured wave front propagated slowly and in a nonuniform fashion. During the second captured beat at a CL of 100 ms, the wave front underwent conduction block across the fiber (Fig. 3). Block was then followed by a clockwise rotation of the wave front at a CL of 75 ms. The block was functional because propagation proceeded without block at baseline (Fig. 3A). During the second rotation, another wave front entered the mapped region from the bottom, signaling the onset of AF (Fig. 3). Induction of AF with heptanol was highly concentration dependent. Raising heptanol concentration to 5 µM and higher (up to 50 µM) caused loss of rapid atrial capture and inability to induce AF.

View larger version (64K): [in this window] [in a new window]   Fig. 3.   Induction of reentry during rapid atrial pacing in a canine left atrium exposed to 2 µM heptanol. A: baseline (control) showing 2:1 capture during pacing at 50-ms PCL with regular and relatively uniform propagation after each paced beat. However, after 2 µM heptanol (B), the first captured beat propagates slowly and nonuniformly (arrows), whereas the second captured beat undergoes functional conduction block (double lines). Block then leads to the formation of a clockwise reentrant activation (arrows). After the completion of the first rotation at a cycle length of 72 ms (snapshots 133.1 ms and 202.4 ms), another wave front enters the mapped region from the lower right part of the mapped field, signaling the onset of AF. Abbreviations are as in Fig. 1. The last frame shows the pacing site (*) and the fiber orientation (double-headed arrow).

Effects of heptanol on TMP and APD restitution. In three dogs, we recorded TMP before and after 2 µM heptanol. There was no significant difference in the dV/dt_max at PCL of 400, 300, and 200 ms. At baseline, the dV/dt_max were 40.8 ± 4.1, 33.4 ± 2.8, and 31.5 ± 2.3 V/s, respectively. After 2 µM heptanol perfusion the dV/dt_max at these three PCLs were 37.9 ± 3.6, 33.4 ± 2.8, and 29.6 ± 2.1 V/s, respectively (P = not significant for all comparisons) (Fig. 4). Heptanol (2 µM) had no significant effect on the APD at all pacing CLs studied (400-200 ms) and no maximum slope of the dynamic APD restitution curve in all three tissues (0.85 ± 0.08 vs. 0.80 ± 0.14; P = not significant) (Fig. 4).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Dynamic transmembrane action potential duration (APD) restitution curves to 90% repolarization (APD90) before (A) and after (B) 2 µM heptanol in three atrial tissues (cases 1-3). Note that the maximum (max) slope of the APD90 remains <1 after heptanol. C: atrial transmembrane action potential recordings during 400- and 200-ms PCL before (control) and during 2 µM heptanol perfusion. Heptanol produces no detectable effect on either APD restitution curves in all three tissues or transmembrane action potential properties. DI is diastolic interval.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Major findings. The ability of partial cellular uncoupling with heptanol to increase the incidence and duration of inducible AF in normal canine atria without changing the active cellular properties constitutes the major finding of this study. Critical uncoupling (2 µM) seemed to be necessary for increased vulnerability to inducible AF. Heptanol at 1 µM, 5 µM, and higher failed to promote inducible AF.

Mechanisms of heptanol-induced AF. Heptanol-induced AF was preceded by nonuniform atrial conduction slowing that eventually underwent block with subsequent formation of reentry leading to AF. The ability of heptanol to cause nonuniform atrial conduction slowing is consistent with previous experimental (28) and simulation (16, 27) studies. For example, Jalife et al. (11) have shown that low concentrations of heptanol cause significant ventricular conduction slowing with no or only minimal changes in the fast sodium current and action potential properties. These findings in the ventricular myocardium are consistent with the present data in atrial tissue. Simulation studies have shown that decreased cellular coupling (i.e., increased gap junctional resistance) unmasks dispersion of APD, a phenomenon that leads to increased spatial dispersion of excitability (16, 27). Because heptanol increases gap junctional resistance (25), it is possible that the nonuniformity of wave-front propagation seen in the present study could have resulted from increased spatial dispersion of atrial excitability (6).

Conduction slowing and block. Heptanol-induced conduction slowing, conduction block, and reentry appear to be independent of the potential influence of heptanol on the active cellular properties of the atrium (20, 25). This suggestion is supported by the following two findings. First, the concentration of heptanol (2 µM) that promoted inducible AF was not associated with any significant effect on the dV/dt_max of the action potential upstroke, indicating that the observed wave break was independent of the potential of heptanol to block sodium ion channels (20). Second, heptanol at 2 µM did not change the slope of the dynamic APD restitution curve or atrial APD, indicating that the mechanism of heptanol for increasing AF vulnerability is not mediated to the APD restitution hypothesis (29). In line with the cellular uncoupling hypothesis of atrial vulnerability to tachyarrhythmias is the demonstration of atrial conduction disturbances and rapid atrial pacing-induced atrial tachyarrhythmias in transgenic mice with Cx40 deficiency (8).

Heptanol and the mechanism of AF maintenance. During AF, frequent wave breaks developed both across and along the fiber. Wave breaks developed more frequently across than along the fiber (67% vs. 33%). The recovery period at the site of the wave break along and across the fiber was significantly shorter than the mean AF CL. Whereas a temporal excitable gap is known to exist during AF (10, 30), early premature wave fronts may block as they encroach the refractory tail of the previous waves (i.e., outside the excitable gap). In the present study it is possible that the wave fronts that blocked along the fiber at coupling intervals of ~58% of the mean AF CL may have resulted from the residual refractoriness left behind by the previous wave front. In contrast, block across the fiber developing at coupling intervals longer or equal to 80% of the AF CL suggests, but does not prove, that block across the fiber can result from a heptanol-induced increase in side-to-side gap junctional resistance (24). Wave breaks were often followed by the formation of a reentrant wave front of excitation that could be sustained for a short period (atrial flutter), or underwent breakup, leading to multiple and irregular wavelets characteristic of AF (9, 31).

Limitation of the study. It may be argued that two-dimensional mapping of a three-dimensionable structure may miss important insights. However, our ability to demonstrate heptanol-induced wave break during rapid pacing leading to reentry and subsequent breakups, causing sustained AF, allowed us to effectively test the partial cellular uncoupling hypothesis of increased vulnerability to inducible AF in the canine atria.

Clinical implications. Atrial remodeling causing enhanced AF vulnerability engages either transmembrane ionic changes (electrical remodeling) (6, 32) or atrial structural changes, including interstitial atrial fibrosis (anatomic remodeling) (1, 17). Remodeling of the gap junctions that may result from cardiac adaptive processes (23) or from cardiac diseases (21) may also lead to enhanced AF vulnerability.