INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE PHYSIOLOGICAL IMPORTANCE of autonomic neurons located on the heart, the intrinsic cardiac ganglia, is becoming progressively more apparent. One population of these ganglia in the dog resides within a fat pad located at the junction of the right pulmonary veins and the right atrium (2, 22). These neurons constitute what is known as the right atrial ganglionated plexus (RAGP). Localized electrical stimulation within this fat pad produces a prompt sinus bradycardia (4). Direct postganglionic nerve projections from the RAGP to the sinoatrial (SA) node have been identified by retrograde tracer techniques and electrophysiologically by using discrete bipolar electrical stimulation of intra-atrial nerve projections (4). In anesthetized dogs, localized neural blockade [hexamethonium (4) or tetrodotoxin (8) injected into the fat pad containing the RAGP] or surgical removal of the RAGP (3, 22) eliminated the bradycardia otherwise produced by activating the parasympathetic innervation of the SA node. In conscious dogs, heart rate (HR) was significantly increased (20, 21), and the high-frequency peak of the HR power spectrum was virtually eliminated (20) in dogs whose RAGP had been surgically removed. Taken together, these data indicate that the ganglia within the RAGP are innervated by parasympathetic preganglionic inputs and that the soma located within this ganglionated plexus project to, and thereby inhibit, the SA node pacemaker cells.

Another aggregate of intrinsic cardiac ganglia, the posterior atrial ganglionated plexus (PAGP), is anatomically close to the SA node; these neurons are located in fatty tissue on the rostral dorsal surface of the right atrium overlying the interatrial septum, immediately caudal to the right pulmonary artery and between the superior vena cava and ascending aorta (see Ref. 15 for anatomic figure). In dogs with -adrenergic blockade, electrical stimulation of the PAGP or stimulation in the immediately surrounding region produced no slowing in HR (15). Injections of fast blue into the SA node, with time allowed for retrograde transport, resulted in substantial staining of RAGP soma with little or no labeling of the PAGP neurons (15). Moreover, removal of the PAGP did not eliminate the bradycardia to vagal stimulation (19). The obvious question therefore arises as to the physiological role of these intrinsic cardiac ganglia in PAGP, especially because this region is contiguous with the principal extrinsic-to-intrapericardial entry point of right-sided sympathetic inputs into the SA node (2, 25). This is a particularly interesting question because evidence is accumulating that in many cases populations of intrinsic cardiac ganglia at different locations on the heart preferentially subserve specific functions in cardiac control. For example, the RAGP has been principally associated with direct vagal control of the SA node, whereas a group of intrinsic cardiac ganglia at the juncture of the inferior vena cava and inferior left atrium primarily subserves autonomic control of atrioventricular node function (2-4, 8, 14, 22, 24, 25). Moreover, recent data have indicated that the intrinsic cardiac ganglia contain multiple neuronal types including parasympathetic and sympathetic soma, sensory cells, and interneurons; together they are capable of mediating intracardiac reflexes and, potentially, of allowing for sympathetic-parasympathetic interactions at intracardiac sites separate from the end effectors (2, 6, 9, 14).

We have used a discriminative classical (i.e., Pavlovian) conditioning procedure to investigate the autonomic control of cardiac function (5, 16, 21). The procedure involves presenting a 30-s pulsed tone followed by shock; this tone, called a "conditional stimulus" (CS+), evokes a sudden, initial "phase 1" (P₁) tachycardia. HR tends to drop slightly toward baseline after the initial tachycardia before increasing again during "phase 2" (P₂). The rate of increase in HR (or slope) is greater for P₁ than P₂, although the total magnitude of the P₂ HR increase is larger than that occurring during P₁ (5). HR decreases toward its pretone value during the last portion of the tone (P₃, "phase 3"). Surgical removal of the RAGP essentially eliminates the P₁ tachycardia, which leads us to conclude that P₁ results from withdrawal of parasympathetic tone to the SA node (21). The magnitude and rate of HR increase during P₂ does not differ significantly before and after RAGP removal; conversely, this phase of the HR conditional response is attenuated 85% after -adrenergic blockade (21). These latter findings led us previously to conclude that the P₂ tachycardia is a reliable index of changes in SA node sympathetic activity (21). A steady 30-s tone, the CS, that is never followed by shock, serves as a behavioral discriminative stimulus: presentation of the CS evokes a dynamic and transient P₁ response that is almost identical to that for the CS+, but there is no P₂. Removal of the RAGP eliminates the short-latency (i.e., P₁) tachycardia, indicating that in the dog the brief HR response to the CS is attributable primarily to withdrawal of parasympathetic tone (13). In summary, the classical conditioning paradigm allows us to assess in the conscious animal the relative involvement of changes in cardiac sympathetic and parasympathetic nervous activity during a sudden, acute behavioral challenge, thereby correlating alterations in autonomic inputs with the resultant changes in chronotropism.

The purpose of the present experiment was to determine the role played by the PAGP in the control of the canine HR response to the classical conditioning test. Specifically, we examined 1) the effects of removal of the PAGP on resting HR, 2) the HR response to CS+ and CS before and after PAGPX, and 3) the role of -receptors in PAGP modulation of the conditional HR response. Although the conditional response is remarkably stable over time, we prepared a second group of sham-operated animals to test for possible training effects with time. We found an increase in the behaviorally induced tachycardia in the dogs before versus after ablation of the PAGP; the HR response in the sham-operated animals was virtually unchanged. We believe the augmented tachycardia after PAGP ablation is attributable to an "unleashing" of sympathetic excitation of the SA node. If so, neurons within the PAGP could selectively modulate sympathetic input to the SA node, thereby "tailoring" overall changes in sympathetic nervous activity to the specific needs of the cardiac pacemaker. Preliminary accounts of these findings have been published (17, 18).

	METHODS

Top Abstract Introduction Methods Results Discussion References

Subjects. Seventeen mongrel dogs (average weight = 23.5 kg) were assigned to one of two groups for use in this study. The PAGP was surgically ablated (PAGPX) in group 1 dogs (n = 9); the HR conditional response was examined in these animals before versus after PAGPX (i.e., a within-subjects design). Group 2 consisted of the remaining eight dogs that were sham operated; comparison of postsurgery responses in group 1 versus group 2 animals provided another test of the effects of PAGPX on HR control (i.e., a between-subjects design). All experiments were performed in accordance with the National Institutes of Health "Guide for the Care and Use of Laboratory Animals" [DHEW Publication No. (NIH) 85-23, Revised 1985, Office of Science and Health Reports, DRR/NIH, Bethesda, MD 20892] and were approved by the institutional animal care and use committee of the University of Kentucky.

Behavioral conditioning. The training was conducted in three stages: adaptation (1- to 2-wk duration), in which the animals were brought to the laboratory and placed in an isolation booth (1) for 1-2 h daily with no further manipulations; habituation (1 wk), in which the tones were presented but no shocks were given; and acquisition (4 wk), in which the dogs learned the association between one tone, the CS+, and shock. During habituation, a 30-s pulsed tone (that was eventually to become the CS+) and a nonpulsed tone of the same frequency (eventually to become the CS) were each presented five times daily. The habituation sessions continued daily until neither tone evoked any sustained HR response. The acquisition sessions were identical in procedure except that the 30-s CS+ tone was followed by a 0.5-s shock delivered across the animal's flank. The minimum amplitude of shock that caused the dog to flinch and its HR to increase was used; the current typically ranged from 3-5 mA and never exceeded 8 mA. The training sessions were complete when the animals showed consistent HR changes during the CS+ with no sustained changes during the CS.

Presurgical control data. Control observations (i.e., before the PAGP ablation or sham surgery) were collected once the animals had acquired a discriminative conditional HR response. Sessions were conducted in the unblocked state, after -adrenergic blockade (1 mg/kg iv DL-propranolol tested with an 0.5 µg/kg iv bolus of isoproterenol), after muscarinic receptor blockade (0.1 mg/kg iv atropine sulfate), and after combined -adrenergic and muscarinic blockade. Each condition was tested for a minimum of two CS+ and two CS trials on each of 3 days; a minimum of 1 day elapsed between drug tests.

Surgical preparation. The dogs were anesthetized [initially induced with thiopental sodium (15-25 mg/kg) and maintained on isoflurane] and prepared for sterile surgery. The left and right cervical vagi were isolated through a neck incision, and HR changes were recorded during electrical stimulation (30-s trains at 20 Hz, 2-ms duration, 2-6 V in 2-V increments). The animals were placed on positive pressure respiration, and their chests were opened through the right fourth intercostal space. The pericardium was incised to form a cradle. In group 1 dogs, the fatty tissue on the posterior surface of the right atrium and the inferior surface of the right pulmonary artery that contains the PAGP (15) was removed, after which these discrete surfaces were painted with phenol (88% carbolic acid). In group 2 dogs, the heart was exposed in an identical manner except that no tissue was removed. For all animals, bipolar electrodes were sutured onto the superior right atrium, the right atrial appendage, and the right ventricle for recording electrograms during the behavioral trials. Catheters were implanted in the right atrium or right femoral vein for eventual drug infusions. A catheter was also placed in the femoral artery, and a Konigsberg pressure transducer was inserted into the left ventricle for use in another experiment. The chest and leg incisions were closed, and negative intrathoracic pressure was reestablished. Finally, the stimulations of the cervical vagi were repeated, and the neck incision was closed. Prophylactic antibiotic therapy (e.g., Kefzol, 500 mg iv) was given immediately before surgery and was maintained postoperatively as indicated. Animals were given buprenorphine (0.2 mg im) at 8-h intervals for up to 24 h postoperatively or as needed for analgesia.

Postdenervation procedures. The conditioning sessions were reinstituted 1 wk or sooner after surgery, but no data were used in subsequent analyses until the dogs had recuperated from surgery for 2 wk. Data were collected with and without pharmacologic autonomic blockade, as described in Presurgical control data. Finally, at the end of the study, each animal was anesthetized [0.05 ml/kg Innovar-Vet (Pitman-Moore, Washington Crossing, NJ): 0.4 mg/ml fentanyl + 20 mg/ml droperidol (see Ref. 23) supplemented with 10 mg/kg pentobarbital sodium (n = 5) or with 30 mg/kg (n = 2) pentobarbital sodium], and the cervical vagi were restimulated for a final direct test of vagal slowing. The dog was then euthanized with an overdose of pentobarbital sodium.

Data acquisition and analysis. The right atrial and right ventricular electrograms and the left ventricular and arterial pressures were digitally sampled at 500 Hz using a Data Translation 2821 analog-to-digital converter in conjunction with an 80486 microprocessor. Data sampling started 30 s before the beginning of each CS+ (or CS) and ended 30 s after shock (or tone off). The microprocessor then computed HR on the basis of the interbeat interval and stored the results. The digital files from individual trials (6 minimum, but generally >15) were ensemble averaged to produce a "high-resolution analysis" of the conditional responses (5).

Statistical tests. The HR responses for selected aspects of the conditional response (i.e., resting HR, peak and average HR increases during CS+ vs. rest, and P₁ and P₂ amplitudes and slopes; see Ref. 5) were measured directly from the high-resolution files for each individual dog. A composite response to the CS+ and CS was computed by averaging the high-resolution files over all dogs. The statistical significance of the effects of the removal of the PAGP on these data was tested using a repeated-measures ANOVA with a term for surgery (control/PAGPX) and for -adrenergic blockade (unblocked/propranolol). A two-factor mixed ANOVA was also conducted with terms for group (PAGPX/sham operated between subjects) and for surgery (pre/post surgery repeated measures within subjects). Post hoc t-tests were conducted when allowed by the results of the ANOVA. Significance was accepted for P < 0.05. All data are given as means ± SD.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Figure 1 shows the group average HR in the absence of vagal stimulation (i.e., 0 V) and at the indicated intensities during the 30-s stimulation of the right (Fig. 1A) and left (Fig. 1B) cervical vagi. The data are for seven of the nine dogs; one animal (dog 5327) was eliminated from the experiment (see below), and one animal died before the final vagal stimulation was performed. The closed symbols in Fig. 1 show the HR during the stimulation of the vagus on the day of the thoracic surgery before PAGP ablation; the open symbols show the corresponding data for the same animals at the terminal experiment (average of 14 wk after PAGPX). There were no significant differences between the bradycardia induced by vagal stimulation before versus after PAGPX.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Average heart rate (HR) in anesthetized dogs (n = 7) determined from atrial electrogram for spontaneous rhythm (0 V) and during 30-s stimulation of right (A) and left (B) cervical vagus nerves before (, solid line) and 14 wk after (, dashed line) surgical ablation of posterior atrial ganglionated plexus (PAGPX). Error bars shown for 0 V only for clarity. Stimulus parameters were 2 ms in duration, at a frequency of 20 Hz, at 2, 4, and 6 V. There were no statistically significant differences in the responses across time, indicating that the direct vagal slowing of atrial rate remained intact after removal of PAGP. bpm, Beats/min.

Table 1 (row 1) shows the average ± SD resting HR for the group 1 dogs under various conditions. There were no differences before versus after PAGPX.

Figure 2 shows illustrative 90-s-long HR recordings from one dog starting 30 s before the beginning of a CS+ trial conducted before (Fig. 2A) and 14 days after PAGP ablation (Fig. 2B). The tone was presented between 30 and 60 s. Pretone resting HR was essentially unchanged in this animal before versus after PAGPX, and a pronounced respiratory sinus arrhythmia was present in both states. Selected portions of these same HR recordings are plotted on the same scale in Fig. 2C. The plots start 10 s before the beginning of the tone and extend through the peak of the response, and the control trial (thin line), PAGPX data (heavy line), and differences (shaded area) between the two trials are plotted. Both the peak HR (control = 160 beats/min, PAGPX = 188 beats/min) and absolute increase in HR relative to resting HR (control = +89 beats/min, PAGPX = +126 beats/min) were increased after PAGPX.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Illustrative HR responses to single conditional stimulus (CS+) trial before (A, control) and after (B, PAGPX) surgical removal of PAGP in dog 5840. Resting HR (HRr) was essentially unchanged before vs. after surgery, and respiratory sinus arrhythmia remained intact. C: plot of last 10 s of pretone rest and initial seconds of CS+ (horizontal bar) for control trial (thin line) and post-PAGPX trial (heavy line) on same axis to emphasize difference in amplitude (shaded area) of 2 conditional HR responses. Peak increase in HR (HRpk) during CS+ was larger (188  62 = 126 beats/min) after PAGPX than in control state (160  71 = 89 beats/min).

Figure 3 shows the composite HR response to CS+ and CS for the eight group 1 dogs for control trials conducted before PAGPX. Each plot was constructed by ensemble averaging the high-resolution files from the individual animals for the respective tones, and the duration of the tone is delineated. Both CS+ and CS evoked P₁ responses that were generally similar. Conversely, the CS+, but not the CS, evoked a large P₂. HR peaked at 42 s, after which it declined as part of P₃. The unconditional response due to shock delivery at the end of CS+ consisted of a sharp, but brief, tachycardia. Table 1 shows the average preganglionectomy (i.e., control) values for the components of the conditional response that are of primary interest.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Composite analysis (n = 8) of HR response to CS+ (solid line) and CS (dashed line) before removal of PAGP. Tones were delivered between 30 and 60 s, as indicated by horizontal bar on time axis. Response to CS+ consisted of an initial, rapid tachycardia (P1) followed by slower, but larger increase (P2). HR declined during last portion of CS+ (P3) but increased after shock delivery at 60 s [i.e., unconditional response (UR)]. Response to CS consisted of only initial P1.

Figure 4 is a composite analysis of the conditional HR response before and after removal of the PAGP and shows both data identical to the control CS+ data in Fig. 3 and the HR response for the same eight dogs after PAGPX. There are several important observations from this figure that are quantitatively confirmed in Table 1. First, the resting HR was unchanged by removal of the PAGP. Second, the peak and average amplitudes of the behaviorally induced tachycardia were decidedly elevated after the selective PAGP ganglionectomy [pre- vs. post-PAGPX post hoc paired t (degrees of freedom = 7) for peak = 4.72 (P < 0.002); for average increase, t = 4.20 (P < 0.004)]. Third, although the HR trajectory during the first 1-2 s of the conditional response was similar in the two situations, the overall amplitude of P₁ was larger after PAGPX (t = 2.86; P < 0.03). Finally, the amplitude (t = 2.90; P < 0.02) and slope (t = 2.83; P < 0.03) of P₂ were also larger after the PAGP ganglionectomy.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Composite analysis (n = 8) of HR response to CS+ before (thin line; identical to that of Fig. 3) and after (heavy line) removal of PAGP. Conditional increase in HR was much larger after PAGPX.

Figure 5 shows the composite HR response to CS+ after -adrenergic blockade before and after PAGPX. After propranolol pretreatment, there is no evidence for any difference in the HR response to CS+ before versus after PAGPX (see also Table 1).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Composite analysis (n = 8) of HR response to CS+ after -adrenergic blockade (propranolol, 1 mg/kg iv) before (thin line) and after PAGPX (heavy line). HR scale is same as in Fig. 3. There is no demonstrable difference in the 2 responses when -receptors are blocked.

Resting HR after muscarinic blockade (n = 7) was the same in the control state (159 ± 23 beats/min) and after PAGPX (160 ± 23 beats/min); likewise, there was no difference in the peak conditional HR increase before (+25 ± 20 beats/min) versus after (+33 ± 26 beats/min) atropine. Finally, the HR after combined muscarinic and -blockade (n = 7) was similar before (125 ± 13 beats/min) and after (122 ± 12 beats/min) PAGP ablation, as were the peak conditional increases (control, +11 ± 6 beats/min; PAGPX, +14 ± 5 beats/min).

Figure 6 shows the composite response to the CS tone before and after PAGPX and the control (i.e., presurgery) response to CS+ for reference. There were no significant differences in the amplitude of the P₁ HR response to the discriminative (CS) stimulus (23 ± 6 beats/min vs. 29.4 ± 14 beats/min, before vs. after PAGPX). Although the mixed-factor (i.e., between groups) ANOVA produced an F value (sham/PAGPX) that was (barely) significant (F_1,14 = 4.59; P  0.05) for P₁ slope, post hoc tests failed to detect a significant difference before (15.6 ± 6.1 beats · min¹ · s¹) versus after PAGPX (19.8 ± 4.4 beats · min¹ · s¹).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Composite analysis (n = 8) of HR response to CS before (thin line) and after (heavy line) PAGPX; dotted line shows predenervation response to CS+ for reference. Prior work shows that HR response to CS is due principally to withdrawal of vagal inhibition of sinoatrial (SA) node. Unlike response to CS+ (Fig. 3), there were no significant differences in HR change evoked by discriminative stimulus before and after PAGPX.

Table 2 and Fig. 7 summarize the resting HR and the HR response to the CS+ before and after sham surgery for the group 2 dogs. To facilitate comparison of the amplitudes and slopes of the two phases of the conditional response, we show the data as a change in HR (relative to the average pretone value) for the 30 s of the CS+ tone only. The response was stable despite the surgery and the time that separated the two data sets. Finally, the mixed-factor ANOVA detected a significant interaction (pre/post surgery × sham/PAGPX) for the peak and average HR increases and for the P₂ slope.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Average change () in HR in response to CS+ in 8 dogs before (thin line) and after (heavy line) sham operation. Changes were assessed against average pretone HR. Data are shown only for 30-s tone. There were no significant differences in response before vs. after surgery, demonstrating stability of conditional HR response.

The response of dog 5327 to the PAGPX differed substantially from the remaining eight dogs: vagal stimulation no longer slowed this animal's HR after surgery, its resting HR increased from 74 to 106 beats/min before versus after surgery, and the resting sinus arrhythmia was virtually eliminated. The HR conditioning results for this dog were consequently eliminated from the data set.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The intrinsic cardiac ganglia undoubtedly play an important role in the control of cardiac function. We had previously used Pavlovian conditioning protocols to produce controlled, time-locked changes in cardiac autonomic nervous activity and showed that removal of the RAGP elevated resting HR, abolished the respiratory sinus arrhythmia, and eliminated the rapid P₁ tachycardia. These results indicate that neurons within the RAGP provide direct vagal inhibition of the SA node and that the initial component of the conditional HR response (i.e., P₁) is caused primarily by withdrawal of parasympathetic inhibition of the SA node (21). In contrast, the major findings of the present study were that 1) the direct vagal control of HR remained after removal of the PAGP, 2) PAGP ablation potentiated the conditional HR response (P₁ and P₂ components), 3) there was no evidence for any difference in the behaviorally elicited (CS+) response before versus after PAGPX when the -adrenergic receptors were pharmacologically blocked, 4) there was no evidence for any difference in the behaviorally elicited (CS+) response before versus after PAGPX when the muscarinic receptors were pharmacologically blocked, 5) there was no evidence for an effect of PAGP ganglionectomy on the HR response to the discriminative stimulus (i.e., CS), and 6) there was no evidence for any effect of the sham operation on the conditional HR response. We believe that these findings provide new insight into the organization of the autonomic control of SA node function via the intrinsic cardiac ganglia: although the PAGP does not directly inhibit the SA node, it restrains the sympathoexcitatory effects directed to the pacemaker tissues.

Our first finding was that stimulation of the cervical vagi still evoked a profound bradycardia after removal of the PAGP. This is in marked contrast to what we observed after removal of the RAGP (compare Fig. 1 with Fig. 8 of Ref. 21). Moreover, resting HR was not significantly different before and after PAGPX. Conversely, surgical removal of the RAGP (RAGPX) increased resting HR from 85 to 114 beats/min (21). Finally, the respiratory sinus arrhythmia, an index of direct parasympathetic efferent input to the SA node, was preserved in the present experiment but not after removal of the RAGP (20). Each of these findings indicates that in contradistinction to RAGPX, the ability of the parasympathetic nervous system to control HR directly remained intact after PAGPX.

Our second finding is that the excision of the PAGP resulted in a significant increase in the size of the conditional tachycardia. In particular, PAGPX increased the amplitude and slope of the more slowly developed (i.e., longer latency) P₂ tachycardia. We have previously reported that P₂ is due primarily to increased sympathetic activity rather than withdrawal of parasympathetic activity (21). The most parsimonious interpretation of our second finding therefore is that the PAGPX augmented the HR increase by potentiating the action of sympathetic nervous activity at the SA node. However, the PAGP ganglionectomy also significantly increased the amplitude of P₁. This surprised us because we had previously attributed P₁ almost exclusively to withdrawal of parasympathetic nervous activity. Although we still believe that this is the primary force behind P₁, it is entirely possible that the earliest effects of increased sympathetic activity are actually expressed during P₁. The action of PAGPX to unleash the sympathetic cardioacceleration simply allowed us to detect these early effects, or perhaps even the effects of any resting sympathetic tone.

There was no tangible difference between the HR responses to CS+ before and after PAGPX in the presence of -adrenergic blockade. That is, the potentiation of the conditional HR response by the PAGPX depended on the presence of functioning -receptors. In fact, Table 1 indicates that virtually all aspects of the conditional HR response were nearly identical in the group 1 dogs before and after the PAGPX when the -adrenergic receptors were no longer responsive. Whatever the mechanism for the effects of PAGPX on HR control, it depends on an intact -adrenergic receptor for its expression.

In marked contrast to the effects of PAGPX on the tachycardia evoked by CS+, the ganglionectomy had no demonstrable effect on the HR response to CS. The latter consists of a brief tachycardia that is almost identical to the P₁ response to CS+. The CS evokes a response presumably because it takes the animal a moment to determine whether the tone is steady (i.e., the CS) or whether it is pulsed (i.e., the CS+) and thus reinforced by shock. The sudden, short-lived P₁ HR response to CS, which constitutes an "alpha response" akin to an orienting response, is eliminated by removal of the RAGP (13). We attributed it therefore primarily to withdrawal of parasympathetic nervous activity. There is no demonstrable sustained tachycardia resultant from increased sympathetic drive. It is as though the animal is able to discriminate between the tones so rapidly that it doesn't "bother" to activate the cardiac sympathetic nerves in response to the neutral CS. Therefore, our failure to demonstrate an effect of PAGPX on the HR response to CS accords with our postulated organization of the neural mediation of this response.

In this experiment we tested whether the increase in the conditional HR response after the PAGP ganglionectomy might be due to a training effect: the amplitude of the response simply increased with an increased duration of the animal's exposure to the conditioning paradigm. The stability of the conditional HR response in the sham-operated animals argues strongly against this possibility.

Dog 3375 in our previous study of the effects of RAGPX (21) showed a similar response as seen in the present group 1 animals. Conversely, the response to PAGPX in dog 5327 in the current study was almost identical to what we had observed earlier with removal of the RAGP (unpublished observation). One possibility is that our surgical procedure, or perhaps the phenol paint, removed other neural structures in these two dogs besides those within the ganglionated plexus. It is also possible, of course, that there is some variability across animals in the interactions that occur within and between the RAGP and PAGP.

Perspectives

The inhibition of sympathetic tachycardia observed during stimulation of the vagus has been called "accentuated antagonism" (10-12). Figure 8A represents a classic view of the organization of the parasympathetic and sympathetic pathways innervating the SA node. Postjunctional and prejunctional mechanisms whereby vagal stimulation can antagonize sympathetically induced chronotropic effects are shown. Postjunctionally at the SA node the interaction occurs at the level of adenylyl cyclase, where parasympathetic stimulation attenuates the sympathetically induced rise in intracellular cAMP levels via separate and antagonistic G protein-coupled receptor mechanisms (11, 12). Also shown are the presynaptic interactions likely to occur at the end-effector site. Note that in this model a single population of the parasympathetic intrinsic ganglion cells is shown giving rise to both the direct vagal inhibition of the SA node and the presynaptic inhibition of sympathetic neurotransmitter release.

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Figurative illustrations of potential peripheral autonomic neural interactions between parasympathetic and sympathetic efferent fibers innervating SA node. , Pre- and postganglionic parasympathetic projections; , pre- and postganglionic sympathetic projections. In the classic representation (A), a cardiac intrinsic parasympathetic neuron provides both direct inhibition of the SA node, thereby slowing HR, and presynaptic inhibition of neurotransmitter release from sympathetic nerves. B: same labeling system used to propose an updated model for neural interactions occurring within and between intrinsic cardiac ganglionated plexus and at SA node pacemaker tissues. Within intrinsic cardiac ganglia (G), note that some efferent sympathetic projections are depicted as fibers of passage whereas others terminate on cell bodies. Disruption of right atrial ganglionated plexus (RAGP) eliminates direct parasympathetic efferent projections to SA node but leaves intact a major parasympathetic-mediated suppression of sympathetic projections to the SA node. It is proposed that this residual parasympathetic-sympathetic interaction occurs within the posterior atrial cardiac ganglionated plexus, thereby suppressing sympathetic efferent inputs prejunctional to the SA node. Removal of PAGP would be expected to potentiate sympathoexcitation of the SA node, as found in current experiment.

Our current experiments, in conjunction with other recent work (14, 19), indicate that sympathetic-parasympathetic interactions occur at intrapericardial sites in addition to those that occur at the end effector. Figure 8B is a model that incorporates these additional interactions that we believe occur between sympathetic and parasympathetic neurons contained within the intrinsic cardiac ganglionated plexus (14, 19). They may also include axo-axonal interactions between parasympathetic neurons contained within the intrinsic cardiac ganglionated plexus and axons of passage of the sympathetic postganglionic projections to the SA node (14, 25). Further studies of the anatomy of the various ganglionated plexuses, of course, may lead us to modify the model. Nonetheless, although some sympathetic fibers to the SA node course into and through the ventral RAGP, the major sympathetic projection to the SA node, at least from the right side, courses between the superior vena cava and ascending aorta (25). This same area contains the PAGP (15). Therefore, the PAGP neurons are ideally located to exert modulating effects on adjacent sympathetic efferent projections.

The new model helps explain the previously puzzling observation (21) that HR increased from 114 ± 17 beats/min to 159 ± 28 beats/min after muscarinic blockade in resting dogs that had been subject to "selective SA node parasympathectomy" (i.e., ablation of the RAGP). The new hypothetical model (Fig. 8B) clearly shows that the parasympathetic inhibition of sympathetic neurotransmitter release would have remained intact in our previous study so that the atropine could have caused a tachycardia by disinhibiting sympathetic neuronal release of catecholamines. In fact, we reexamined these previous data, selecting only those trials in which -blockade (propranolol, 1 mg/kg) was instituted first and found that subsequent administration of atropine (0.1 mg/kg) increased HR to only 117 ± 10 beats/min (n = 4; Ref. 21).

One of the most obvious implications of the new model is that the nervous system would potentially retain independent control of direct parasympathetic inhibition of the SA node and parasympathetic inhibition of sympathetic cardioacceleration. This would allow for greater versatility of neural control. For example, parasympathetic ganglia within the PAGP could selectively modify sympathetic input to the SA node independently of direct vagal inhibition of pacemaker activity (via the RAGP), thereby tailoring a more generalized level of sympathetic nervous activity to meet the organism's specific needs for HR control.