Autonomic Nerve Stimulation Reverses Ventricular Repolarization Sequence in Rabbit Hearts

Heart Preparations

Fully innervated rabbit hearts were isolated using a modified method of that described by Ng et al.²⁴ Briefly, adult New Zealand White rabbits (3.5 to 4 kg, n=14) were premedicated with ketamine HCl (0.1 mg kg⁻¹, SC; Hospira Inc) and Meditomidine (1 mg, SC; Pfizer) followed by general anesthesia induced and maintained by Diprivan (1 to 2 mg kg⁻¹ IV; Astra Zeneca). Rabbits were ventilated after tracheotomy using a small-animal ventilator (60 breaths per minute; Harvard Apparatus Ltd, Edenbridge, Kent, UK) with an O₂/air mixture. The vagus nerves were isolated, and the blood vessels leading to and from the ribcage were ligated and dissected. The rabbit was euthanized with sodium pentobarbital and portions of the rib cage were removed and the descending aorta was cannulated. The pericardium was cut, and ice-cold Tyrode’s solution was applied to the surface of the heart to preserve myocardial function. The preparation extending from the neck to thorax was dissected from surrounding tissues as described previously.²⁴ An advantage of this nerve–heart preparation is that the peripheral vessels and baroreflexes are eliminated when the spinal cord is severed below the Medulla oblongata, and the heart and part of the rib cage are isolated (see Figure 1). All procedures were undertaken in accordance with the Animals (Scientific Procedures) Act 1986 and the current Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication No. 85-23, revised 1985).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Isolated innervated heart preparation. A, Side view of the heart–nerve preparation. The photograph shows the heart immobilized between a glass slide above and rubber slings (brown) below. This keeps the heart in position for optical mapping. B, Schematic of the isolated innervated heart preparation. The heart was placed in a custom-made chamber for optical mapping of APs. PA indicates pulmonary artery; LV, left ventricle; RV, right ventricle. The ascending aorta was perfused with a 5-F pulmonary wedge monitoring balloon catheter inserted into the right carotid artery. A quadripolar electrode was placed in the right ventricle via the pulmonary artery and used to pace the heart or measure a bipolar electrogram. A second quadripolar electrode was inserted in the spinal canal and used to stimulate the thoracic outflow of the sympathetic trunk at T2 level. C, Close-up view of the heart in position for optical mapping. Lowercase letters describe the following structures. a, Elastic rubber slings were placed posterior to the heart to immobilize the anterior surface against a glass slide. b, Thymus gland and associated fat pads. c, Glass slide with a transparent marker for exact positioning of the preparation. Note that the dark border has the dimensions of the optical field-of-view. d, Left anterior descending artery in the interventricular sulcus. e, Obtuse marginal artery. f, Left atrium. g, Right vagus nerve placed on stimulating electrodes. D, Schematic of the mapped region. The supine position of the heart held by rubber straps stabilizes an area of the left ventricle between the interventricular sulcus and the obtuse marginal artery that was mapped optically (area shown as a grid).

A new method was devised to improve dye and drug delivery to the heart and improve cardiac perfusion pressure and coronary flow rate. Two perfusion circuits were set up with 2 peristaltic pumps. One pump perfused the spinal cord from the descending aorta to the spinal artery, at low flow rates (10 to 15 mL/min). The other perfused the heart in a modified Langendorff apparatus at a higher constant flow rate (30 mL/min) by inserting a 5 French catheter (B. Braun Medical Inc, Bethlehem, Pa) up the ascending aorta above the aortic valve and inflating the balloon catheter to maintain perfusion pressure to the heart (65±5 mm Hg) and to separate the two perfusion circuits. This approach was critical to stain the heart with the voltage-sensitive dye, obtain reproducible drug actions at reasonable concentrations, and achieve physiological perfusion pressures to the heart and the spinal cord. The heart and nerves were perfused with Tyrode’s solution containing (in mmol/L): 130 NaCl, 24 NaHCO₃, 1.0 CaCl₂, 1.0 MgCl₂, 4.0 KCl, 1.2 NaH₂PO₄, 20 dextrose, at pH 7.4, and gassed with 95% O₂ plus 5% CO₂ at 37±1°C. Hearts were stained with the voltage sensitive dye, di-4 ANEPPS (40 μL of a 1 mg/mL stock solution in dimethyl sulfoxide). Two methods were used in combination to reduce motion artifacts. (1) The heart was placed horizontally in a custom designed chamber with an adjustable base and 2 straps to hold a glass slide against the region of epicardium to mechanical immobilize the region viewed by the photodiode array (Figure 1). (2) Cytochalasin D was used sparingly at 2 μmol/L in the perfusate for 10 minutes, followed by washout. When applied in this manner, cytochalasin D reduced contractions for 20 to 30 minutes without altering action potential (AP) characteristics, conduction velocity, or repolarization.

Optical Apparatus

The optical apparatus was previously described.¹⁸ Briefly, light from two 100-W tungsten–halogen lamps was collimated and passed through 520±30-nm interference filters. Fluorescence emitted from the stained heart was collected with a camera lens. Fluorescence emission from the anterior region of heart was passed through a cutoff filter (610 nm; Omega Optical, Brattleboro, Vt) and focused on a 16×16 elements photodiode array (C4675-103; Hamamatsu Corp, Bridgewater, NJ) (Figure 1). Outputs from the photodiode array were amplified, digitized (14-bit) at 2000 frames per second (Microstar Laboratories Inc, Bellevue, Wash) and were acquired for at least 40 seconds for subsequent data analysis.

Nerve Stimulation and Image Acquisition Protocols

Data Analysis

APDs were measured from the optical APs recorded at each site from the difference between the activation and repolarization time points, as previously described.¹⁵ The activation time point was calculated from the maximum first derivative of the AP upstroke (dF/dt)_max, which represented the time when most of the cells viewed by a diode fired an AP and the repolarization time point was calculated from the maximum second derivative of the AP downstroke, (d²F/dt²)_max.¹⁵ We previously showed that the inflection point, (d²F/dt²)_max, represents the refractory period of the AP based on extra stimuli applied at 1.5× threshold voltage. The second derivative fell at 97% and 90% to 96% recovery to baseline for APs recorded, respectively, with a microelectrode or optically.¹⁵ In hearts with a prominent plateau phase, such as the rabbit (but not the mouse or rat AP), the maximum second derivative of the AP downstroke provided a unique time point that is not distorted by motion artifacts. In addition, the second derivative algorithm offered a major advantage in that it was impervious to low levels of motion artifacts compared with APD70 or APD90.¹⁵ All parameters, APDs, DOR, activation intervals, and heart rate were expressed as the mean±SEM. Comparisons were made using Student’s t test on paired samples.^14,29,32 Comparisons were made between paired samples and t test was used to test the significance. A probability value of <0.05 was considered statistically significant. The isochronal lines (2-ms interval) were drawn using custom written software using IDL (Interactive Data Language).

Effects of Autonomic Nerve Stimulation and Perfusion With Neurotransmitters on Heart Rate

Figure 2 illustrates the time course of heart rate changes elicited by SNS (A), bilateral VNS (B), or perfusion with isoproterenol (C) or acetylcholine (D). All four interventions were applied until the maximum change in heart rate was reached then the intervention was stopped. As shown in Figure 2A and 2B, the interruption of autonomic nerve stimulation resulted in the recovery of the heart rate back to its original value in seconds. In contrast, the wash of neurotransmitters and the reversal of their effects on heart rate was considerably slower requiring several minutes. (Note time scale in Figure 2C and 2D.) Bipolar electrogram recordings were made continuously to verify that the hearts remained in sinus rhythm throughout the measurements. Electrogram recordings are illustrated for each intervention before (traces a), during (traces b), and after (traces c), the application of SNS (A), VNS (B), ISO (C) and ACH (D). In all the hearts analyzed, activation under sinus rhythm lasted 6±2 ms (n=30) and the patterns did not vary significantly (not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Changes in activation intervals (AI) and electrogram recordings as a function of time. Changes in activation intervals were plotted as a function of time during sinus rhythm before (a), during (b), and after (c) a physiological intervention. Below each plot, electrogram recordings are shown for three phases of these experiments before (a), during (b), and after (c) recovery from an intervention. The duration of each intervention is shown as a red line. Activation interval are in milliseconds. A, SNS. B, Bilateral VNS. C, Infusion of isoproterenol (ISO) (100 nmol/L). D, Infusion of acetylcholine (ACH) (1 μmol/L).

Sympathetic Nerve Stimulation

Stimulation of the spinal canal to induce SNS increased heart rate gradually from 143±6.3 (AI=427±18.3 ms) to a peak rate of 229±13.2 (AI=265±13.2 ms) bpm (Figure 2A). More remarkable, SNS changed the direction of repolarization from apex toward the base to base toward the apex (Table, n=7). Figure 3A illustrates a map of repolarization gradient and a series of 16 superimposed APs from apex to base at slow and fast sweep speeds, during sinus rhythm before SNS. Figure 3B shows the same measurement from the same heart after 50 seconds of SNS at 15 Hz. The reversal of the direction of repolarization during SNS was attributable to difference of its effects at the base and apex of the ventricles. As shown in the Table, SNS did not significantly change APDs at the apex (control: APD=159±4.2 ms; during SNS: APD=155±12 ms; P=NS), whereas SNS caused a marked shortening of APDs at the base (134±9.3 ms) compared with sinus rhythm (APD: 176±4.0 ms; P<0.001; n=7). DOR during sinus rhythm (APD base–APD apex) was 17±0.7 ms, and during SNS, DOR changed to −22±1.6 ms (P<0.001; n=7), meaning that the direction of repolarization was reversed because of heterogeneities of the effects of SNS being more pronounced at the base than at the apex.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Effects of sympathetic neuromodulation of ABDOR. Each map (left panels) represent the dispersion of repolarization for a single cardiac beat and is displayed as an isochronal map with lines 2-ms apart. Light to dark shades represent early to late repolarization time points. Arrow points to the direction of repolarization sequence. Traces designated as “a” illustrate the superposition of APs (aligned in time) from different locations on the heart and differences of APDs across the surface. Traces designated as “b” display the same traces at a faster sweep speed to visualize temporal differences in AP downstroke. CTRL indicates control; VNSR, right VNS; VNSL, left VNS; VNSBL, bilateral VNS.

Sinus Rhythm Versus Isoproterenol Infusion

Isoproterenol infusion (100 nmol/L) increased heart rate to 204±13.1 bpm (n=6) (P=0.0289) and DOR to 23±1.6 ms (P<0.01, n=5) (Figure 2C) but did not change the direction of repolarization, which remained from apex to base. Figure 4A and 4B compares maps of APDs recorded from the same heart first in sinus rhythm and 2 minutes after perfusion with 100 nmol/L isoproterenol.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Pharmacological modulation of (apex–base) DOR. DOR is mapped with isochronal lines (2 ms apart). Light to dark shades represent early to late repolarization time points. Arrow points to the direction of repolarization. Note that there is no change in the direction of the repolarization sequence caused by the adrenergic or muscarinic agonists. Isochronal lines are 2 ms apart. CTRL indicates control; ISO, isoproterenol; ACH, acetylcholine infusion.

Time Course of DOR Changes Caused by SNS

On SNS, DOR reversed from apex→base to base→apex by 15 seconds, and when SNS stimulation was stopped, heart rate returned to baseline and DOR returned to control values and to its original orientation (ie, apex to base), with a latency of 115±15 seconds. Figure 5 depicts the changes in DOR as a function of time for 4 hearts for which DOR was mapped before, during 50 seconds of SNS, and 150 seconds after the interruption of SNS. The slower recovery of DOR of ≈2 minutes implicates a cellular signaling process, but the possibility of a slow redistribution of neurotransmitter (ie, norepinephrine) or a slow dissociation constant from its receptor cannot be excluded.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Dynamic changes of DOR during SNS and recovery. Time course of changes in DOR before, during, and after sympathetic nerve stimulation (n=4 hearts). DOR was inverted within 15 seconds after SNS and remained invested in some experiments for as long as 15 minutes (not shown), whereas recovery had a longer latency of 115±15 seconds after the interruption of SNS. After the interruption of SNS at peak heart rate (≈50 seconds), different hearts reversed their DOR at different times. More precisely, the earliest reversal occurred at 115 seconds in 1 of 4 hearts; 3 hearts reversed at 165±5 seconds and 1 heart at 180 seconds.

Vagus Nerve Stimulation

Right VNS caused a gradual drop in heart rate to 95±4.5 bpm (n=6), and DOR did not change significantly in magnitude or direction (Table). Left VNS caused a significant a similar drop in heart rate as for right VNS (96±5.8 bpm), failed to produce a statistically significant change DOR, and did not reverse the orientation of repolarization (Table). In contrast, bilateral VNS reversed the direction of repolarization from apex→base to base→apex, with ΔAPD=−16±2 ms during bilateral VNS. Figure 3A and 3C through 3E compares maps of repolarization from the same during sinus rhythm (Figure 3A, control), right VNS (Figure 3C), left VNS (Figure 3D), and bilateral VNS (Figure 3E). Bilateral VNS reversed the repolarization sequence attributable to its different effects at the base and apex of the ventricles. APDs at the apex increased significantly during bilateral VNS compared with sinus rhythm, from 159±4.2 to 195±14 ms (n=6; p <=0.02), which may be attributed in large part to the decrease in heart rate from 143±6.3 to 91±6.6 bpm (Table). At the base, however, bilateral VNS did not cause a statistically significant APD prolongation from 176±4.0 ms in sinus rhythm to 178±13.8 ms with bilateral VNS (Table). When compared with the anticipated APD prolongation observed by pacing at long cycle lengths, bilateral VNS failed to produce the APD prolongation expected from the slower heart rate.

Sinus Rhythm Versus Acetylcholine Infusion

Perfusion with acetylcholine (1 μmol/L) to mimic parasympathetic nerve activity decreased heart rate to 105±10.2 bpm with significant increase in (DOR 32±3.0 ms) but without any change in the direction of repolarization sequence (Table). Figure 4C illustrates a repolarization map and the superposition of APs (apex→base) that were recorded during perfusion with acetylcholine (1 μmol/L). As shown in Figure 4C, repolarization advanced from apex to base, similar to repolarization sequence recorded in the absence of autonomic nerve activity.

Effects of Autonomic Nerve Stimulation on the Repolarization Sequence

Figure 6 summarizes the changes in DOR measured on the left ventricle of rabbit hearts during SNS, right, and left and bilateral VNS compared with isoproterenol or acetylcholine perfusion. The averaged data reveal a statistically significant reversal of the repolarization sequence, being oriented from apex→base in controls to base→apex during SNS or bilateral VNS. Changes in the magnitude of DOR can be attributed to rate and cellular signaling effects caused by norepinephrine and acetylcholine, but the reversals of the repolarization sequence implicate spatial heterogeneities of autonomic nerves along the apex–base axis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. Summary of autonomic modulation of DOR. A, Summary of DOR changes in sinus rhythm compared with various autonomic nerve stimulation protocols. n indicates the number of rabbit hearts tested; CTRL, control; VNSR, right VNS; VNSL, left VNS; VNSBL, bilateral VNS; Iso, isoproterenol (100 nmol/L); Ach, acetylcholine (1 μmol/L). **P<0.05 by paired t test vs control.

Ventricular Repolarization

Ventricular activation sequences have been extensively characterized in intact hearts by mapping the time delays of QRS spikes measured from closely spaced extracellular bipolar electrodes.³³ In contrast, the repolarization sequence or recovery to baseline has been poorly defined, in large part, because of technical difficulties and the sensitivity of that process to numerous variables. Changes in heart rate, temperature depth of anesthesia, sympathetic tone, and electrolyte balance all influence the APD locally and, thus, ventricular repolarization. APs can be recorded from the epicardial surface with microelectrodes or suction electrodes but only from a limited number of sites, leaving surface electrograms as the most practical method to track activation and repolarization sequences.³³ However, unlike activation, there is no electrogram deflection that accurately identifies the onset or termination of repolarization. Nevertheless, local activation recovery intervals (ARIs) measured with surface electrodes (ie, activation time point (minimum dV/dt of the QRS complex) minus the recovery (maximum dV/dt of the T-wave) time point) were used as a surrogate for APDs measured with an intracellular electrode. In early studies on dog hearts, ventricular repolarization determined from ARIs produced opposite findings; with repolarization sequence proceeding from base to apex³⁴ or apex to base.³⁵ Although there was a good correlation between ARIs and refractory periods and APDs, most ARIs were shorter than refractory periods³⁶ and differed from APDs by as much as 24 ms.³⁷ Alternatively, repolarization sequences determined from the local refractory periods indicated that repolarization and APD gradients proceeds from base to apex (short-to-long APDs) in situ in dog hearts.^38,39 A more recent study on canine and human hearts reported the opposite result that APDs are shorter in ventricular myocytes from the apex than the base caused by higher densities of K⁺ currents involved in ventricular repolarization.¹¹

The advent of voltage sensitive dyes and optical mapping techniques has provided new tools to characterize ventricular repolarization.⁴⁰ In isolated perfused hearts, ventricular repolarization was found to proceed from apex to base in a wide range of species: (1) mouse,^12,41 (2) rat,⁴² (3) guinea pig,¹⁴ (4) rabbit,³² and perhaps (5) dog and human hearts. In all of these species, repolarization is driven by a different set of ion channels; yet their spatial distribution preserves short to long APDs in going from apex to base, so as to guide the sequence of repolarization and relaxation.

Anatomical Distribution of Autonomic Nerves to the Heart

Sympathetic outflow to the heart emerges from the thoracic intermediolateral horns of the spinal cord. Preganglionic neurons in the spinal cord project to the heart via axons to postganglionic neurons to innervate the heart via paravertebral ganglia and the various thoracic and cardiac ganglia.^43,44 Studies on dogs reported that different branches of the sympathetic nerves innervated distinct regions of the ventricles but there is a controversy regarding such specific anatomical nerve distribution. In our experiment, spinal canal stimulation at the level of T2 vertebrae allowed us to broadly activate bilaterally the sympathetic nerves thus, circumventing potential controversies, inaccuracies, and animal-to-animal variability of cardiac nerve distribution.^45,46

Effects and Mechanism of Sympathetic Activity on DOR

Changes in APD during SNS were heterogeneous and striking. During the SNS, DOR not only increased across the ventricle but its direction was inverted. Under control conditions, APDs were longer at the base than the apex, whereas during SNS, APDs at the base were significantly shorter than at the apex. Changes of DOR on SNS occurred rapidly in 5 to 15 seconds, and DOR reversed back to control orientation but with a slower time course of 130 to 140 seconds. These findings are consistent with current concepts in the literature. The autonomic nerves and of delayed rectifying K⁺ channels involved in the repolarization of the heart are heterogeneously distributed along the apex to base axis of the heart. In human hearts, autonomic nerves (both sympathetic and parasympathetic) are considerably more abundant at the base than the apex of left ventricles.⁴⁷ The delayed rectifying K⁺ currents are nonuniform on the surface of the heart, with the slow component, I_Ks, being more abundant at the base and the fast component, I_Kr, being more abundant at the apex of the rabbit left ventricle.^32,48 Moreover, I_Ks tends to be a weak current, difficult to measure until activated by a β-adrenergic agonist such as isoproterenol, which decreases the refractory period by a protein kinase A–dependent phosphorylation of I_Ks but has no effect on I_Kr.^49,50 Therefore in the absence of sympathetic activity, I_Kr dominates and repolarization advances from apex to base. In contrast, during SNS, activation of I_Ks by norepinephrine is more abundant at the base, resulting in greater APD shortening at the base than at the apex and a reversal of the repolarization sequence. The preferential shortening of APDs at the base under sympathetic influence occurred in ≈15 seconds (ie, time to phosphorylate I_Ks) and was reversible on interruption of SNS with a latency of 180 seconds consistent with the time course of I_Ks dephosphorylation.⁴⁹

Interestingly, perfusion of the heart with isoproterenol had a tendency to enhance DOR (Table) by shortening APDs slightly more at the apex than the base, but the direction of repolarization was not reversed. These data highlight the complexity of the system because SNS releases norepinephrine, a β- and α-adrenergic agonist that stimulates nearly all ionic currents and their kinetics. In addition, SNS may activate afferent as well efferent nerves and release other neurotransmitters such as neuropeptide Y, ATP, and substance P. Nevertheless, the results from the isoproterenol perfusion allows us to reasonably conclude that regional (apex–base) heterogeneities of sympathetic innervation (efferent and/or afferent nerves) is the dominant factor responsible for the reversal of DOR during SNS rather than I_Kr, I_Ks, or β₁ adrenergic receptor distribution across the ventricle.²⁸

Effects and Mechanism of Parasympathetic Activity on DOR

Right or left VNS did not change the DOR significantly, but bilateral VNS reversed the direction of DOR, this time by prolonging APD more at the apex than the base. All 3 methods of VNS decreased heart rate to the same extent, so that differences in APD caused by the 3 conditions cannot be attributed to heart rate alone (Table and Figure 5) but to enhanced recruitment of parasympathetic and acetylcholine release during bilateral VNS. Acetylcholine released by vagus nerves may shorten ventricular APDs by increasing the K⁺ conductance via the muscarinic receptors in the ventricle.⁵¹ Immunohistochemistry⁴⁷ of parasympathetic nerve distribution to the ventricle suggests a greater density of nerves at the base than the apex, which may explain the lack of APD prolongation at the base, despite the slow heart rate during bilateral VNS. Thus, as for SNS, the effects of VNS on DOR are most likely caused by heterogeneities in nerve distribution resulting in the modulation of the repolarization sequence.

Neuronal modulation of ventricular repolarization is unique and distinct from pharmacological modulation. These findings reveal that under physiological conditions autonomic activity has a marked impact on cardiac repolarization and broaden our views on why and how “autonomic imbalance” may be an important factor contributing to arrhythmia vulnerability during physical or emotional stress. Further studies are needed to elucidate the effects of autonomic nerve activity on transmural repolarization, on hearts with long QT or ischemia to gain deeper insights into the contribution of neural activity to arrhythmia vulnerability.