The purpose of the present study was to determine whether thoracic veins may act as ectopic pacemakers and whether nodelike cells and rich sympathetic innervation are present at the ectopic sites. We used a 1,792-electrode mapping system with 1-mm resolution to map ectopic atrial arrhythmias in eight normal dogs during in vivo right and left stellate ganglia (SG) stimulation before and after sinus node crushing. SG stimulation triggered significant elevations of transcardiac norepinephrine levels, sinus tachycardia in all dogs, and atrial tachycardia in two of eight dogs. Sinus node crushing resulted in a slow junctional rhythm (51 ± 6 beats/min). Subsequent SG stimulation induced 20 episodes of ectopic beats in seven dogs and seven episodes of pulmonary vein tachycardia in three dogs (cycle length 273 ± 35 ms, duration 16 ± 4 s). The ectopic beats arose from the pulmonary vein (n = 11), right atrium (n = 5), left atrium (n = 2), and the vein of Marshall (n = 2). There was no difference in arrhythmogenic effects of left vs. right SG stimulation (13/29 vs. 16/29 episodes, P = nonsignificant). There was a greater density of periodic acid Schiff-positive cells (P < 0.05) and sympathetic nerves (P < 0.05) at the ectopic sites compared with other nonectopic atrial sites. We conclude that, in the absence of a sinus node, thoracic veins may function as subsidiary pacemakers under heightened sympathetic tone, becoming the dominant sites of initiation of focal atrial arrhythmias that arise from sites with abundant sympathetic nerves and periodic acid Schiff-positive cells.
- atrial tachycardia
- autonomic nervous system
- sympathetic nerves
- computerized mapping
the mechanisms of rapid repetitive activities from thoracic veins remain unclear. In a rabbit sinocaval preparation, Ito et al. (10) demonstrated spontaneous diastolic depolarizations that could lead to automatic activity. Cheung (6) reported that isolated pulmonary veins (PVs) were capable of independent pacemaking activity. Masani (13) noted that clear nodelike cells distinct from normal myocardial cells are present in the myocardial layer of the rat PV. Light and electron microscope studies have subsequently confirmed the presence of these cells in human PVs (14). In addition to nodelike cells, thoracic veins share another common feature with nodal tissue in possessing rich sympathetic innervation (17). Whether or not increased densities of nodelike cells and rich sympathetic innervation are present in the thoracic vein sites that act as ectopic pacemakers is unclear. We possess a high-density computerized mapping system capable of mapping thoracic veins to a 1-mm resolution with 1,792 bipolar electrodes (18). Our laboratory also recently demonstrated that pacemaker cells could be identified at the sites of focal discharge in PVs by periodic acid Schiff (PAS) stain (7). The latter study was performed in isolated canine PVs, and the densities of the PAS-positive cells were not determined. In the present study, we sought to test the hypothesis that thoracic veins are common sites of origin of ectopic atrial arrhythmias in vivo during stellate ganglia (SG) stimulation, after the sinus node is crushed. We also hypothesize that these ectopic foci have higher densities of sympathetic nerves and PAS-positive specialized conduction cells than sites not serving as ectopic pacemakers.
The research protocol was approved by the Institutional Animal Care and Use Committees and conforms to the American Heart Association guidelines. We studied eight normal dogs (22–27 kg).
Under isoflurane anesthesia, the chest was opened via a median sternotomy. The left (LSG) and right SG (RSG) were visually identified. A pair of bipolar hook electrodes was attached to each SG to record SG nerve activity. Signal was filtered with a 30- to 500-Hz band pass and acquired at 979 samples/s using Prucka Labsys (Austin, TX) along with surface ECG.
High-density epicardial mapping of thoracic veins, including PVs and vein of Marshall (VOM) and adjoining left atrium (LA) and right atrium (RA), was performed using a 1,792-bipolar channel mapping system. Three or four flexible electrode plaques, each with 496 bipolar electrodes with a 1-mm resolution, were used to map the RA, LA, left PVs, and right PVs, according to methods published in a previous report (18).
SG Stimulation and Electrophysiological Studies
We stimulated the SG using three protocols.
Bilateral SG stimulation and blood sampling for catecholamine.
We sought to demonstrate that SG stimulation increases cardiac sympathetic tone by comparing transcardiac (coronary sinus minus aorta) levels of catecholamines before and after bilateral SG stimulation. We sampled blood simultaneously from the coronary sinus, aorta, and cephalic vein. Both SG were then maximally simultaneously stimulated at a level at least thrice their thresholds (30 mA, 20 Hz, 2-ms pulse width) for a period of 30 s. The threshold is defined as current needed to produce a 20% or more rise in systolic blood pressure (SBP) or heart rate. Immediately thereafter, blood was resampled from the same sites. The samples were stored in ice and later processed for ELISA catecholamine assays.
Hemodynamic effects of unilateral SG stimulation.
To determine whether there is functional asymmetry between the hemodynamic effects of LSG and RSG activation, we performed graded stimulation of each SG (2–35 mA) and recorded their differential effects on heart rate, blood pressure, and SG nerve activity.
Arrhythmic effects of unilateral SG stimulation.
The aim of this protocol is to determine whether SG stimulation can induce arrhythmias and to determine whether there is functional asymmetry in the arrhythmic effects of left vs. right stellate stimulation. Each SG was stimulated separately at 20–40 mA. High-density mapping was performed 1) at baseline sinus rhythm without SG stimulation; 2) at baseline after SG stimulation; 3) after mechanical crushing of the sinus node but without SG stimulation; and 4) after sinus node crushing and after SG stimulation. When SG stimulation was performed, mapping was acquired immediately after the end of the stimulation train. We avoided atrial or PV electrical stimulation throughout the experiment, as these maneuvers are known to induce arrhythmias even without sympathetic stimulation.
During in vivo studies, the ectopic sites were identified on playback of high-density maps and marked on the back surface of mapping plaques. At the end of the experiment, the mapping plaques were removed from the tissue, and the sites were marked on the tissue using indelible Indian ink. Digital photos of the experimental setup with mapping plaques in situ aid in ensuring accurate electrical-to-anatomical correlation of the ectopic sites. The heart-PV preparation was fixed in a nonaqueous fixative, Carnoy's solution, to optimize staining for glycogen with PAS reagent. PAS staining is used to identify glycogen-rich specialized conduction cells, such as Purkinje cells (3). Multiple tissue blocks were sampled from the sites of ectopic focus, as well as other sites, including the LA and RA appendages, the PVs, VOM, and atrioventricular (AV) junction. The slides were stained with routine hematoxylin and eosin and Masson's trichrome for structural examination. The slides were also immunostained with antibodies to tyrosine hydroxylase (TH) for sympathetic nerves, according to methods described previously (4). Light microscopy was used to examine the slides. Nerve density and PAS staining were analyzed using computerized morphometry (Image Pro) (4). We quantified and compared the densities of sympathetic nerves and PAS staining at multiple locations. These included the ectopic atrial or venous sites, and other nonectopic atrial or venous sites that are at least 10 mm away from the focus. We defined the ectopic site histologically as any area within a 5-mm circle containing the Indian ink-marked spot. The Indian ink spot remains visible on the paraffin tissue block.
Serum epinephrine and norepinephrine concentrations were assayed by using a competitive enzyme immunoassay kit per manufacturer's protocol (Alpco Diagnostics, Salem, NH). Transcardiac catecholamine levels (pg/ml) were defined by coronary sinus minus aorta concentrations.
Electrophysiological Data Analyses
The absolute maximum time derivative of volume was selected as the time of local activation for bipolar electrograms. Dynamic display and isochronal maps were used to analyze the activation sequence. A focal activation was defined as an activation pattern with wavefronts spreading outward in all directions. An activation wavefront that originates from the edge of the mapped region was not considered focal, unless originating from the distal PV.
Data are expressed as means ± SD. Values between two groups were compared with Student's two-tailed t-tests, and multiple-group comparisons were made with ANOVA, followed by Newman-Keuls post hoc analysis. A 2 × 2 χ2 test was used to test the association between two categorical data. A P value of ≤0.05 was considered statistically significant.
SG Stimulation and Serum Catecholamine Levels
The SG was easily identified in all dogs (Fig. 1A). Histological section (Fig. 1B) showed the presence of TH-positive ganglion cells and nerve fibers, confirming the correct identification of this structure. Figure 1C shows the effects of bilateral SG stimulation on blood catecholamine levels sampled from the coronary sinus, aorta, and peripheral vein in all dogs studied. SG stimulation resulted in a rise of norepinephrine levels in the coronary sinus (P < 0.01), aorta (P < 0.01), and peripheral vein (P < 0.01), without a concomitant rise of epinephrine levels [P = nonsignificant (NS)]. Transcardiac (coronary sinus minus aorta) norepinephrine levels (P < 0.01), but not epinephrine levels (P = NS), were significantly elevated after SG stimulation, consistent with the fact that norepinephrine is the catecholamine released from nerve endings during SG stimulation.
Functional Asymmetry of LSG vs. RSG Stimulation
Figure 2 illustrates the results of graded stimulation of LSG vs. RSG. There was no difference in the threshold to induce a 20% rise in SBP and heart rate during LSG vs. RSG stimulation (6.8 ± 3.1 vs. 5.8 ± 3.3 mA, P = NS). At weak stimulation strengths (10–15 mA) (Fig. 2A), RSG stimulation induced a greater change in heart rate than LSG stimulation (P < 0.05 at both 10 and 15 mA), whereas LSG stimulation (Fig. 2B) produced a greater change in SBP than RSG stimulation (P < 0.01 at 15 mA). At stronger stimulation strengths (20–35 mA), there were no significant differences between the effects of LSG vs. RSG stimulation on either heart rate or SBP. Graded stimulation with increasing current strengths produced a progressive rise of heart rate and SBP until 30 mA. Beyond that, the rise in SBP after 35-mA SG stimulation was less compared with 30 mA, suggesting that fatigue had set in. Weak unilateral SG stimulation (≤15 mA) elicited neural afterdischarges from the ipsilateral ganglion only (Fig. 2B). On the other hand, stronger unilateral SG stimulation (20–40 mA) elicited bilateral SG afterdischarges (Fig. 2C). The threshold to elicit bilateral SG afterdischarges (the afterdischarge threshold) was 20.8 ± 8.6 mA for LSG and 17.0 ± 9.7 mA for RSG (P = NS).
Table 1 summarizes the results of high-density mapping.
Before sinus node crushing.
At baseline (before sinus node crushing), no arrhythmias were observed if SG was not stimulated. Weak SG stimulation (2–15 mA) did not induce any atrial arrhythmias. Stronger SG stimulation (20–40 mA) induced sinus tachycardia in all dogs, and atrial tachycardia in only two out of eight dogs (Table 1, column 1). The earliest activity was observed from the anterior RA [cycle length (CL) 293.1 ± 23.5 ms] in one dog, and left superior PV (LSPV) (CL 285.5 ± 12.2 ms) in the other dog. No atrial fibrillation was induced.
After sinus node crushing.
After sinus node crushing (but without SG stimulation), a slow junctional rhythm or ectopic atrial rhythm (CL 1,204 ± 152 ms) was observed in all dogs. Subsequent SG stimulation induced 20 episodes of ectopic beats in seven out of eight dogs. Among these 20 episodes, the earliest activity was observed to arise from the LA in 2 episodes, the RA in 5 episodes, the PVs in 11 episodes, and VOM in 2 episodes. SG stimulation after sinus node crushing also induced PV tachycardia (7 episodes) in three dogs (CL 273 ± 35 ms, duration 16 ± 4 s).
Figure 3 shows an episode of PV tachycardia originating from the right superior PV (RSPV). Figure 3A shows baseline sinus rhythm. The earliest site (arrow) was from the sinus node direction. The blue color on the third panel shows the delayed activation of VOM. After sinus node crushing, RSG stimulation induced an RSPV tachycardia (Fig. 3B, beat b). The earliest activation site was within the RSPV, and the propagation was centrifugal from that early site (beat b, arrow). After its termination, an ectopic atrial rhythm resumed (Fig. 3B, beat c). Electrograms 1–4 are taken from sites adjacent to the site of initiation of tachycardia.
Figure 4 illustrates an example of a PV ectopic beat originating from the LSPV. Figure 4A shows three beats of RA rhythm (beat 1) after cessation of pacing (Fig. 4B). The next beat (beat 2) was premature. Its origin was within the LSPV (Fig. 4C). In addition to a single ectopic beat, we also mapped episodes of sustained (>10 s) ectopic rhythm. Figure 5 illustrates one example. Figure 5A shows baseline sinus rhythm without SG stimulation. After sinus node crushing, stellate stimulation produced an ectopic rhythm (Fig. 5B) with earliest activity in the left inferior PV.
In total, there were 20 episodes of ectopic beats induced by SG stimulation (11 PV, 2 VOM, 7 atria) and 9 episodes of atrial tachycardia (8 PV, 1 atrial). Among these, 13 of 29 episodes were induced by LSG stimulation compared with 16 of 29 episodes by RSG stimulation (P = NS). In dogs with multiple episodes of ectopic arrhythmias, the arrhythmias tended to arise from one or two sites only within a PV.
We performed histological analyses of a total of 38 tissue blocks from eight dogs. Table 1 also presents a summary of the histological examinations performed. Among the eight dogs, seven dogs developed ectopic foci after sinus node crushing and SG stimulation. Six of these seven dogs had thoracic vein ectopies, whereas one dog had only atrial ectopies (Table 1). For the six dogs with thoracic vein ectopies, five blocks were obtained from each dog (one each from LA, RA, AV junction, ectopic thoracic vein, and nonectopic vein). For the remaining two dogs without thoracic vein ectopies, four rather than five blocks per dog were obtained (one instead of two thoracic vein blocks per dog). Among the 38 blocks, 14 were from thoracic veins (6 ectopic veins and 8 nonectopic veins), 8 from LA, 8 from RA, and 8 from the AV junction. A total of 38 slides were stained for both PAS and TH (14 from PV, 8 each from LA, RA, and AV junction).
Purkinje-like cells in the PV.
Figure 6 shows the results of PAS staining. The PAS-positive cells, which stain entirely deep magenta, are cells with a greater amount of glycogen compared with regular myocardial cells (light purple). Figure 6A shows PAS staining of the LSPV. Arrows point to PAS-positive cells clustering on the endocardial surface as a thin layer up to three cells thick. Figure 6B shows NKX2.5 staining of the same PV, showing that PAS-positive cells also stain positive for NKX2.5. NKX2.5 is a myogenic transcription factor expressed in significantly higher levels by Purkinje fibers compared with normal myocardial cells (16). Its higher expression in PAS-positive cells than normal myocardial cells suggests that these cells are specialized conduction cells akin to Purkinje cells. We used canine AV junction (Fig. 6, C and D) and ventricular Purkinje cells (Fig. 6E) as internal positive controls. Figure 6C shows the bundle of His at the AV junction, which stained deep magenta. In contrast, the surrounding ventricular myocardium stained light purple. Figure 6D is a higher power view of the bundle of His, showing PAS-positive cells (arrows). Figure 6G shows Purkinje cells (arrows) on the subendocardial side of the ventricular myocardium in the same dog. Figure 6, E and F, show connexin 43 and 40 staining, respectively, in PV. We found that PV and atria stain positive for connexin 40 and 43, but negative for connexin 45. The subendocardial portion of the PV myocardium, where PAS-positive cells are clustered, have relatively less connexin staining compared with the midmyocardial portion. As opposed to canine sinus node, PV and atria were immunonegative for hyperpolarization-activated cation channels HCN2 and HCN4.
Figure 7, A–C, compares the density of PAS staining at ectopic sites vs. nonectopic sites. Figure 7A shows PAS-positive cells at the origin of ectopic focus. Figure 7B shows the staining results for a randomly selected nonectopic site. Figure 7C shows quantitative analyses of all tissues studied, including both thoracic vein and atrial sites. The densities of PAS-positive cells were significantly higher at the ectopic sites than randomly selected nonectopic sites within the atria or PV, but significantly lower than that at the AV junction. The data from ectopic site in Figure 7C includes six thoracic veins and three atrial sites from seven dogs. There was no significant difference between thoracic vein ectopic sites vs. atrial ectopic sites. Hence the data were combined. The density of PAS-positive cells at ectopic sites was 102,626 ± 31,681 μm2/mm2 compared with 24,998 ± 20,647 μm2/mm2 10 mm away from ectopic sites (P < 0.01), 37,705 ± 35,085 μm2/mm2 in nonectopic veins (P < 0.05), 32,826 ± 12,996 μm2/mm2 in the RA (P < 0.01), 43,546 ± 17,601 μm2/mm2 in LA (P < 0.05), and 190,376 ± 9,141 μm2/mm2 in the AV junction (P < 0.01).
Density of sympathetic nerves.
Figure 7, D and E, shows that the density of sympathetic nerves at ectopic sites compared with nonectopic sites. Figure 7D shows an example of nerve structures stained positive for TH (sympathetic nerves) at an ectopic site. Brown twigs (arrows) are sympathetic nerves. Figure 7E shows a positive stained nerve (arrow) at a randomly selected nonectopic site within the PV. Figure 7F shows the density of sympathetic nerves was higher at ectopic sites compared with a site 10 mm away from the ectopic site within the same vein and randomly selected nonectopic RA and LA sites. However, the nerve density at the ectopic site was not statistically different than random sites in a nonectopic PV or from the AV junction. The data from ectopic site in Figure 7F includes six thoracic veins and three atrial sites. The density of sympathetic nerves at ectopic sites was 20,377 ± 13,336 μm2/mm2 compared with 2,771 ± 901 μm2/mm2 10 mm away from ectopic sites (P < 0.05), 9,581 ± 8,558 μm2/mm2 in nonectopic veins (P = NS), 1,067 ± 618 μm2/mm2 in RA (P < 0.05), and 1,228 ± 828 μm2/mm2 in the LA (P < 0.05) and 12,838 ± 8,443 μm2/mm2in the AV junction (P = NS).
We demonstrated that thoracic veins become the commonest sites of initiation of ectopic atrial rhythm with sympathetic nerve stimulation following the elimination of a dominant sinus pacemaker influence. The sites of ectopy initiation are not randomly distributed. Rather, they are distinguishable by higher densities of PAS-positive cells and higher densities of sympathetic innervation than other randomly selected nonectopic atrial site.
Ectopic Pacemakers After Sinus Node Crushing
Sympathetic nerve activity regulates intracellular calcium dynamics (2). Because intracellular calcium release plays important roles in atrial pacemaker activity (7, 12), sympathetic stimulation can induce focal discharges from the VOM and the PVs in dogs (1, 9). In this study of normal dogs, we demonstrated that, after sinus node crushing, sympathetic nerve stimulation can induce focal ectopic atrial beats and tachycardia arising most commonly from thoracic veins. However, the arrhythmias may also arise from the LA and RA. These findings suggest that latent ectopic pacemakers are constitutively present in normal dogs and are unmasked by the elimination of the dominant sinus node influence and with sufficient adrenergic stimulation. It is possible that, in patients with sick sinus node, these latent ectopic pacemakers may take over and serve as either the subsidiary pacemakers or sources of rapid activations that help trigger or sustain atrial arrhythmia.
Histological Findings of the Ectopic Pacemakers
We found high densities of PAS-positive cells at the sites serving as ectopic pacemakers. The high NKX2.5 expression further suggests that these cells are specialized conduction cells. Because of the large differences of densities between the ectopic pacemaker site and other sites, these PAS-positive cells most likely played a role in the generation of the ectopic activity. This conclusion is further strengthened by the finding that, within the same dog, the pacemaker sites tended to be fixed, rather than randomly distributed throughout the thoracic veins and atria. We also found that there is significantly increased sympathetic innervation at the ectopic sites. We hypothesize that the higher density of sympathetic innervation at the ectopic sites translates to exposure of the pacemaker cells at those sites to more intense adrenergic stimuli during each burst of sympathetic nerve activity. This, in turn, promotes both automaticity and triggered activity at those sites. Kwong et al. (11) observed that connexin 43 staining clustered around the edges of the canine sinus node. The authors hypothesized that this distribution allows connexins to act as specialized conduction bundles directing propagation out of the sinus node. In our study, we observed relatively less connexin 43 staining on the subendocardial layer of PV, where PAS-positive cells are clustered.
Functional Asymmetry of LSG and RSG stimulation
Crampton et al. (8) reported that excessive or unopposed activity of the LSG, or subnormal activity of the RSG, account for the pathophysiological manifestations of the long QT syndrome. Schwartz et al. (15) reported that LSG ablation was an effective treatment for ventricular arrhythmias in patients with long QT. These data suggest that there is functional asymmetry between the effects of LSG and RSG stimulation in terms of ventricular arrhythmogenesis. However, less is known about the effects of LSG vs. RSG activation on atrial electrophysiology. One aim of the present study was to determine whether there is functional asymmetry of RSG and LSG stimulation on atrial arrhythmias. LSG stimulation had a greater effect on blood pressure than heart rate, whereas the opposite was true for RSG at lower stimulation strengths. However, there was no significant functional asymmetry in the hemodynamic effects of right and left stellate stimulation at higher stimulation strengths. There was also no difference in arrhythmogenic potential of left- vs. right-sided stellate stimulation. One potential explanation for this observation is that, at higher stimulation levels, stimulation of one SG activates both SG.
This study did not perform intracellular recordings directly from the ectopic PAS-positive cells. Therefore, we cannot determine the exact origin of the ectopic rhythms. Other investigators have performed intracellular recordings from isolated PV cell preparations and demonstrated the presence of nodal action potentials (5). Their data and ours strongly support the hypothesis that PVs contain specialized conduction cells, and that these cells play an arrhythmogenic role under autonomic stimuli.
We conclude that thoracic veins can serve as subsidiary pacemakers and dominant sources of ectopic atrial tachycardias under hyperadrenergic stimuli, once the sinus node is eliminated, because of the presence of abundant sympathetic innervation and PAS-positive specialized conduction cells.
This study was supported by the National Heart, Lung, and Blood Institute Grants P01 HL78931, R01 HL78932, 58533, and 71140; American Heart Association Scientist Development Grant 0435135N; Heart Rhythm Society Fellowship in Pacing and Electrophysiology; Pauline and Harold Price Endowments; a Chun Hwang Fellowship for Cardiac Arrhythmia Honoring Dr. Asher Kimchi; and the Cardiac Arrhythmia Research Enhancement Support Group Inc., Los Angeles, CA.
Medtronic, St. Jude, and Cryocath donated research equipment to our laboratory. P.S. Chen is a consultant to Medtronic, Inc.
We thank Dr. C. Thomas Peter for support and Elaine Lebowitz, Avile McCullen, and Lei Lin for assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 by the American Physiological Society