INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE ADENOSINE 3',5'-cyclic monophosphate (cAMP)-protein kinase A (PKA) pathway is a powerful modulator of neuronal excitability acting on both voltage-dependent and transmitter-regulated channels. Increases in intracellular cAMP result from stimulation of a variety of receptors; one of the most prevalent groups is serotonergic (5-HT) receptor subtypes such as 5-HT₄, 5-HT₆, and 5-HT₇ (see Ref. 37 for review). In hippocampal neurons recorded in vitro, stimulation of 5-HT₄ receptors increased cAMP, resulting in reduced calcium-activated potassium currents and enhanced neuronal excitability (22, 23, 32). It is also evident that either increasing intracellular cAMP or blocking calcium-dependent potassium channels can both result in changes at the systems level. For example, injection of an analog of cAMP into the caudal nucleus tractus solitarii (NTS) decreased respiratory frequency and tidal volume associated with transient hypotension and bradycardia (4). In addition, Lalley et al. (19) showed that intracellular injection of cAMP into expiratory neurons led to a depolarization and increased both firing intensity and burst duration in vivo. Recently, our laboratory (6) has demonstrated that blockade of calcium-dependent potassium channels in the NTS attenuated the cardiopulmonary reflex but potentiated the baroreceptor reflex.

In the present study we investigated a potential role for 5-HT₄ receptors in modulating cardiorespiratory reflex function at the level of the NTS, the central site of termination of visceral afferents (21). Specifically, we have considered a functional role for 5-HT₄ receptors in the NTS and whether their stimulation increases cAMP, which in turn modulates the cardiopulmonary reflex. This reflex can be evoked either naturally by pulmonary edema (25) or experimentally by a right atrial injection of phenyl biguanide (PBG, a 5-HT₃-receptor agonist). The evoked reflex is a potent defensive mechanism involving excitation of J receptors located in the lungs (25), as well as receptors in the heart, which both have vagal C-fiber afferents that project to and terminate in the caudal region of the NTS (5, 28, 33, 36). The resultant reflex involves pronounced decreases in both arterial pressure and heart rate, as well as various respiratory changes including rapid shallow breathing, which can be preceded by apnea (7). It has been reported that either non-N-methyl-D-aspartic acid (NMDA) receptors (36) or both non-NMDA and NMDA receptors (33) were implicated in vagal C-fiber neurotransmission within the NTS. Moreover, the cardiopulmonary reflex was modulated powerfully by 5-HT. Sévoz et al. (30) demonstrated that stimulation of 5-HT₃ receptors within the NTS depressed the cardiopulmonary reflex, whereas neurons receiving synaptic inputs from vagal C fibers were excited by 5-HT acting on 5-HT_1A or 5-HT₃ receptors (35).

Here we report that activation of 5-HT₄ receptors in the NTS depressed the cardiopulmonary reflex, the bradycardic component being mediated by a cAMP-dependent PKA pathway. Preliminary accounts of some of these observations have been published in abstract form (11, 12).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Adult male Sprague-Dawley rats (200-280 g, n = 37) were anesthetized with a mixture of -chloralose-urethan-pentobarbital sodium (69, 690, and 30 mg/kg ip, respectively) and supplemented with -chloralose (30 mg/ml) as required and as used previously by us (6). The level of anesthesia was assessed regularly by an absence of a withdrawal reflex to a toe pinch. The left femoral vein and artery were cannulated for subsequent administration of anesthetic and recording of arterial pressure, respectively. The right atrium was cannulated via the right jugular vein to allow administration of PBG. The trachea was cannulated below the larynx, and the animal was ventilated artificially with oxygen-enriched air. End-tidal CO₂ was sampled monitored and maintained at 4.5% by altering tidal volume.

The arterial cannula was connected to a pressure transducer (Gould Statham), and the arterial pressure signal was amplified. Rectal temperature was monitored and maintained at 38°C using a thermostatically controlled heating blanket (Harvard, Bioscience). An electrocardiogram (ECG) was recorded via stainless steel pins placed subcutaneously into fore- and hindpaws; signals were amplified (Neurolog module 104), filtered (Neurolog module 125), and passed into a window discriminator. With use of the window discriminator, the R wave of the ECG was discriminated and used to generate transistor transistor logic (TTL) pulses, which were subsequently used to derive heart rate (see below). The head was placed in a stereotaxic frame (Kopf Instruments), and the dorsal surface of the medulla was exposed by removing the overlying nuchal muscles and the dura.

In the majority of experiments the right phrenic nerve was exposed and isolated via a lateral approach on the right side of the neck. Phrenic nerve activity was recorded via a bipolar silver wire hook electrode; signals were amplified, filtered (Neurolog module 104 and 125), rectified, and integrated (time constant 30-50 ms). The integrated phrenic nerve activity, TTL pulses derived from the R wave of the ECG, and arterial blood pressure signals were relayed to a Cambridge Electronics Design (CED) 1401 interface connected to a computer running Spike 2 software (CED). Heart rate was automatically computed from the inter-TTL pulse interval and displayed as an instantaneous rate in beats per minute.

Microinjection technique. Seven barreled glass microelectrodes (35- to 45-µm tip diameter) were constructed and placed around the caudal most extent of the area postrema (i.e., calamus scriptorius) under visual guidance using a binocular microscope and a stepper motor. Unilateral microinjections were made at ~900-1,100 µm caudal to the obex, 50-200 µm lateral to the midline, and between 400 and 450 µm ventral to the dorsal surface of the medulla. The volume microinjected (45 nl) was measured as the distance moved by the fluid meniscus as viewed through a binocular dissecting microscope fitted with an eyepiece graticule. Microinjected drugs were given slowly over 30-40 s. Microinjection sites were characterized histologically and physiologically using a microinjection of glutamate (100 mM). If the evoked cardiovascular response was qualitatively similar to that of a cardiopulmonary reflex (i.e., a decrease in both heart rate and arterial pressure, tachypnea or apnea) drugs were microinjected and the reflex was retested. Between one and four different drugs were used at a single microinjection site, depending on the exact experimental protocol (see below). Up to six microinjections were given at the same site (including a repetition of the same drug). Vehicle control microinjections were given in some experiments and did not change either baseline cardiorespiratory variables or the cardiopulmonary reflex.

Experimental protocol and analysis. The cardiopulmonary reflex was elicited by an injection of prewarmed PBG (1-µg bolus in 100 µl) into the right atrium. Ten-minute recovery periods were allowed between PBG injections to prevent tachyphylaxis. In all experiments three stable control reflexes were obtained; between the second and third control reflex tests a microinjection of glutamate was given into the NTS to physiologically identify the site. After a 10-min recovery period the final control cardiopulmonary reflex was tested. If this was quantitatively similar to the previous two reflexes, then modulatory drugs were microinjected into the NTS. If not, then the reflex was repeatedly tested every 10 min until three comparable reflex responses were obtained. In experiments in which the same drug was retested or a different drug was microinjected, three stable controls were again obtained before the microinjection. This was important because it took into consideration any change in reflex efficacy with time and also allowed us to assess preparation viability. After a microinjection into the NTS the cardiopulmonary reflex was restimulated 25-35 s later and then at 10-min intervals until recovery was obtained. Three further reflex control responses were obtained before a second microinjection of drug(s) into the NTS. The peak decreases in phrenic nerve cycle length (average of 2 respiratory cycles), heart rate, and mean arterial pressure were measured from the last (third) control cardiopulmonary reflex response and compared with the peak reflex response obtained after microinjection of a drug into the NTS. The peak depressor effect occurred 1-2 s after the peak fall in heart rate. Thus the fall in arterial pressure measured was presumed to be mediated by withdrawal of sympathetic activity rather than that associated with the powerful reflex bradycardia. In experiments in which combinations of drugs were given (i.e., an agonist in the presence of an antagonist) this was sequential from different barrels of the microelectrode and occurred within 10 s of one another. In most animals, one experiment was performed (i.e., agonist/antagonist). In a minority of rats (n = 5), two experiments were performed in which additional agonists were microinjected (e.g., isoprenaline and SKF-38393).

Histological procedures. At the end of each experiment and before removal of the microelectrode from the NTS, Pontamine sky blue (45 nl) was microinjected into the NTS for subsequent histological verification of the location of the microinjection site. The brain stem was removed and fixed in a 2% paraformaldehyde and 20% sucrose solution over night. After fixation 100-µm transverse sections were cut on a freezing microtome. The sections were mounted on subbed slides, stained with neutral red, cleared, and coverslipped. Microinjection sites were documented on predrawn medullary sections (3).

Drugs. Microelectrodes were filled with a combination of the following drugs: glutamate (100 mM, Sigma); 8-bromoadenosine 3',5'-cyclic monophosphate (8-BrcAMP), a membrane-permeant analog of cAMP (200 mM, Sigma); Rp-adenosine 3',5'-cyclic monophosphorothioate (Rp-cAMPS), a membrane-permeant and selective cAMP-dependent PKA inhibitor (100 mM, Sigma; dose based on Ref. 1); forskolin, an activator of adenylate cyclase (100 µM, RBI); PBG, a 5-HT₃-receptor agonist (100 µM), isoprenaline, a nonselective ₁-adrenoceptor agonist [45 µM; dose as used previously (26)]; and 5-methoxytryptamine (5-MeOT), a nonselective 5-HT-receptor agonist (50 µM, Tocris; see Ref. 37), at a dose of 2.25 pmol. This dose was based on previous data describing the dose-dependent effect of 5-MeOT on cardiorespiratory reflexes (see Ref. 13) and produced a potent, but submaximal, attenuation of reflex responses without affecting baseline variables. Other drugs included: 1-[4-Amino-5-chloro-2-(3,5-dimethoxyphenyl)methyloxy-3-[1-[2-methylsulfonylamino]ethyl]piperidin-4-yl] propan-1-one (RS-39604 hydrochloride, selective 5-HT₄-receptor antagonist; 100 µM; Tocris); 3-(piperidin-1-yl)propyl 4-amino-5-chloro-2-methoxybenzoate RS-23597-190 hydrochloride (selective 5-HT₄-receptor antagonist; 100 µM; Tocris); N-(1-azabicylo[2.2.2]oct-3-yl)-6-chloro-4-methyl-3-oxo-3,4-dihydro-2H-1,4-benzoxazine-8-carboxamide (Y-25130 hydrochloride, 5-HT₃-receptor antagonist; 100 µM; Tocris); (±)-1-phenyl-2,3,4,5-tetrahydro-(1H)-3-benzazepine-7,8-diol (SKF-38393 hydrobromide, selective dopamine D₁-receptor agonist, 100 µM; Tocris; see Ref. 31); cAMP (200 mM, Sigma); saline (0.9%), and Pontamine sky blue (2%). All drugs were dissolved in 0.9% saline.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The resting levels of heart rate, arterial pressure, and phrenic nerve activity cycle length (i.e., time from the onset of a burst to the next) were 468 ± 9 beats/min, 102 ± 5 mmHg, and 0.87 ± 0.09 s (n = 19), respectively (means ± SE, n = 36). Right atrial injection of PBG (1 µg) resulted in reflex falls in both heart rate (337 ± 7 beats/min) and mean arterial pressure (31 ± 1 mmHg) and typically a decrease in phrenic nerve activity cycle length of 0.36 ± 0.02 s (n = 14 of 19 rats); the latter was the most consistent effect seen and is reported here. In 15 rats phrenic nerve activity was not recorded.

Effect of microinjecting 5-MeOT into NTS. 5-MeOT is a nonselective 5-HT-receptor agonist with a high affinity for 5-HT₄, 5-HT_2c, and 5-HT_1D receptors and a low affinity for 5-HT_1B, 5-HT₂, and especially 5-HT₃ subtypes (37). Microinjections of 5-MeOT (2.25 pmol) into the NTS did not induce significant changes in the baselines of the cardiorespiratory variables monitored. A subsequent injection of PBG evoked a reflex bradycardia that was attenuated relative to control (i.e., 329 ± 12 to 91 ± 20 beats/min; see Figs. 1-3 and Table 1, n = 13, P < 0.01). In addition, the reflex decrease in respiratory cycle length was also reduced from 0.41 ± 0.08 to 0.11 ± 0.07 s (Figs. 1-3 and Table 1; n = 6, P < 0.05). In contrast, the depressor response was not significantly different from control (Figs. 1-3 and Table 1; n = 13). This 5-MeOT-evoked attenuation recovered after 10 min (Figs. 1 and 2).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Microinjection of 5-methoxytryptamine (5-MeOT) into nucleus tractus solitarii (NTS) attenuates cardiopulmonary reflex, and this effect is antagonized by selective 5-HT4-receptor antagonist RS-39604. Control: a right atrial injection of phenyl biguanide (PBG, 1 µg) produced a decrease in phrenic nerve activity cycle length (PNA Cl), heart rate (HR), and arterial pressure (AP) response, which was attenuated by 5-MeOT (2.25 pmol). Reduction in depressor response was not significant. There was complete recovery in evoked reflex cardiopulmonary response within 10 min (Recovery). After a Control response microinjecting RS-39604 (4.5 pmol) and 5-MeOT (2.25 pmol) into the NTS blocked attenuating effect seen with 5-MeOT alone. Combination of drugs washed out and within 1 h of subsequent microinjection of 5-MeOT (2.25 pmol) alone attenuated cardiopulmonary reflex. bpm, Beats per minute.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Microinjection of 5-MeOT into NTS attenuates cardiopulmonary reflex, and this effect is blocked by cAMP-dependent protein kinase A (PKA) inhibitor Rp-adenosine 3',5'-cyclic monophosphorothioate (Rp-cAMPS). Control: right atrial injection of PBG (1 µg) produced a decrease in PNA Cl, bradycardia, and depressor response, which was attenuated by 5-MeOT (2.25 pmol). Reduction in depressor response was not significant. There was complete recovery in evoked reflex cardiopulmonary response within 10 min (Recovery). After Control response microinjecting Rp-cAMPS (4.5 pmol) and 5-MeOT (2.25 pmol) into NTS blocked attenuating effect seen with 5-MeOT alone. Combination of drugs washed out and within 1 h of subsequent microinjection of 5-MeOT (2.25 pmol) alone attenuated the cardiopulmonary reflex. Abbreviations as in Fig. 1.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Pooled data showing that microinjection of 5-MeOT attenuates cardiopulmonary reflex-induced bradycardia (A, n = 13) and decrease in PNA Cl (C, PNCl, n = 6) significantly. Open bars, control responses; solid bars, response after microinjection. There was no significant effect of microinjecting 5-MeOT into the NTS on the cardiopulmonary reflex depressor response (B). Microinjecting 5-MeOT with RS-39604 (n = 6) and 5-MeOT with Rp-cAMPS (n = 5) blocked attenuating effects of these drugs on cardiopulmonary reflex. These data indicate that 5-MeOT produced its attenuating effect by stimulating 5-HT4 receptors, which led to increase in intracellular cAMP-dependent PKA activity ** P < 0.01.

Selective antagonism of 5-HT₄ receptors prevents 5-MeOT attenuating effect on cardiopulmonary reflex. Microinjection of 5-HT₄-receptor antagonists (RS-23597 and RS-39604, both 4.5 pmol; n = 3 and 6, respectively) did not alter either resting baseline parameters or magnitude of the cardiopulmonary reflex (Table 1). However, microinjection of 5-HT₄ antagonists reversibly blocked the attenuating effect of 5-MeOT on the bradycardic (RS-39604 and RS-23597) and respiratory (RS-39604) components of the cardiopulmonary reflex (Figs. 1 and 3, Table 1). As seen in Table 1, there was a change in the control response in phrenic nerve cycle length during the experiment with 5-MeOT and RS-39604. The antagonizing effect of RS-39604 on the bradycardic component of the reflex was fully reversible as shown by a subsequent microinjection of 5-MeOT 1 h later, which again produced a significant attenuation of the cardiopulmonary reflex bradycardia from 310 ± 19 to 106 ± 20 beats/min (Fig. 1, Table 1, n = 6, P < 0.05).

Effect of 8-BrcAMP microinjected into NTS on cardiopulmonary reflex. Microinjection of 8-BrcAMP (9 nmol) did not cause any significant effect on basal heart rate, arterial pressure, or phrenic nerve activity. However, the cardiopulmonary reflex bradycardia was attenuated significantly from 352 ± 12 to 103 ± 20 beats/min (Fig. 4 and Table 2, n = 10, P < 0.01), and the decrease in phrenic nerve activity cycle length was significantly reduced from 0.31 ± 0.04 to 0.16 ± 0.07 s (Fig. 4 and Table 2, n = 5, P < 0.05). There was no significant reduction in the depressor response (Figs. 4 and 5 and Table 2, n = 10). The attenuating effect of 8-BrcAMP on both the reflex heart rate and phrenic nerve activity cycle length responses was reversible and recovered to control after 10 min (Fig. 6 and Table 2).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Microinjection of 8-bromoadenosine 3',5'-cyclic monophosphate (8-BrcAMP) into NTS attenuates cardiopulmonary reflex, and this effect is antagonized by Rp-cAMPS. Control: right atrial injection of PBG (1 µg) produced decrease in PNA Cl (i.e., increase in frequency), bradycardia, and depressor response that was attenuated by NTS microinjection of 8-BrcAMP (9 nmol). Reduction in depressor response was not significant. There was complete recovery in evoked reflex cardiopulmonary response within 10 min (Recovery). After Control response microinjecting Rp-cAMPS (4.5 nmol) and 8-BrcAMP (9 nmol) into the NTS blocked attenuating effect of 8-BrcAMP. Combination of drugs washed out and within 1 h a subsequent microinjection of 8-BrcAMP (9 nmol) alone attenuated cardiopulmonary reflex. Abbreviations as in Fig 1.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Pooled data showing percent fall in HR (A), mean arterial pressure (MAP, B), and PNA Cl (PNCl, C) under control conditions (open bars) and after microinjecting 8-BrcAMP (n = 10), forskolin (n = 5), Rp-cAMPS (n = 5), and Rp-cAMPS and 8-BrcAMP (n = 6) into the NTS (solid bars). 8-BrcAMP significantly reduced PBG-evoked reflex bradycardia and reduction in respiratory cycle length after stimulation of cardiopulmonary reflex. Forskolin significantly reduced reflex bradycardia but not depressor effect; it also attenuated reflex decrease in PNA Cl in 3 rats (data not shown). Although an NTS microinjection of Rp-cAMPS alone had no effect on cardiopulmonary reflex, it blocked attenuating effect of 8-BrcAMP on cardiopulmonary reflex bradycardia (** P < 0.01).

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Graph showing time course of attenuating effects after NTS microinjections of 8-BrcAMP, forskolin, 5-MeOT, and PBG on cardiopulmonary-induced reflex bradycardia. Control responses were at 10 s. Note greater duration of attenuating effect observed following NTS microinjection of PBG compared with other drugs.

Effect of blocking cAMP-dependent PKA pathway on cardiopulmonary reflex and on attenuating effect of 8-BrcAMP (n = 5). In all experiments, Rp-cAMPS (4.5 nmol) had no effect on baseline heart rate, arterial pressure, or phrenic nerve activity cycle length or the magnitude of the cardiopulmonary reflex (Fig. 5 and Table 2, n = 5). However, when a microinjection of 8-BrcAMP was given immediately after an injection of Rp-cAMPS there was no significant change in the magnitude of the cardiopulmonary reflex bradycardia and depressor response (n = 6) and respiratory response (n = 2; Figs. 4 and 5 and Table 2). After 1 h, a subsequent microinjection of 8-BrcAMP only into the NTS produced a significant attenuation of the reflex bradycardia response (Fig. 4, n = 6, P < 0.05) but no change in either the reflex depressor response or the decrease in respiratory cycle length.

Effects of forskolin on cardiopulmonary reflex. Microinjection of forskolin (4.5 pmol, n = 5) did not cause a significant change in basal arterial pressure, heart rate, or phrenic nerve activity cycle length. The reflex bradycardia elicited by cardiopulmonary stimulation was attenuated from 342 ± 8 to 97 ± 11 beats/min (Fig. 5 and Table 2, n = 5, P < 0.05), but the reflex change in arterial pressure was not affected significantly (Fig. 5 and Table 2; reflex changes in respiration were not measured). The attenuating effect on the reflex bradycardia was reversible after 10 min (Table 2). In addition, a microinjection of Rp-cAMPS, preceding the one of forskolin, blocked the attenuating effect of forskolin on the heart rate component of the cardiopulmonary reflex (Table 2, n = 4).

5-HT₄ receptors act via a cAMP-dependent PKA pathway in NTS. After recovery from an attenuation of the cardiopulmonary reflex with a microinjection of 5-MeOT, Rp-cAMPS was microinjected immediately before a microinjection of 5-MeOT. The attenuating effect of 5-MeOT on the bradycardic (Figs. 2 and 3, n = 6) component of the cardiopulmonary reflex was blocked consistently by Rp-cAMPS (Table 1). One hour after a microinjection of Rp-cAMPS into the NTS, a second microinjection of 5-MeOT alone attenuated the reflex bradycardia (n = 6, P < 0.05) but no significant changes were found in the depressor response (Table 1, n = 6). In three animals in which phrenic nerve activity was recorded, the reflex tachypnea was depressed by 5-MeOT and this was antagonized by Rp-cAMPS, but data did not reach significance.

Selectivity of putative 5-HT₄ receptor evoked attenuation of cardiopulmonary reflex. An NTS microinjection of PBG (4.5 pmol) attenuated the reflex bradycardia of the cardiopulmonary response (i.e., from 343 ± 4 to 89 ± 18 beats/min) without affecting the depressor or phrenic nerve activity cycle length response components. Neither Rp-cAMPS (n = 3) nor RS-39604 (n = 5) blocked the attenuating effect of PBG on the cardiac component of the cardiopulmonary reflex. Notably, after a microinjection of PBG into the NTS the cardiopulmonary reflex did not return to control for 50-60 min (n = 6). Conversely, a microinjection of Y-25130 (4.5 pmol), a 5-HT₃-receptor antagonist (16), did not alter the attenuating effects of a microinjection of 5-MeOT on the cardiopulmonary reflex (n = 5).

Microinjection site specificity. In control experiments either 8-BrcAMP (n = 3) or 5-MeOT (n = 3) was microinjected more rostrally (up to 1 mm) to the site where an attenuation of the cardiopulmonary reflex was obtained. This caused no significant change in any of the measured variables of the reflex.

Microinjection sites in NTS. Histological analysis of effective microinjection sites revealed Pontamine sky blue deposits located in the NTS. There were 13 sites successfully recovered and are depicted in Fig. 7. These were ~900-1,100 µm caudal to the obex, 50-200 µm lateral to the midline, and between 400 and 450 µm ventral to the dorsal surface of the medulla.

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Effective NTS sites studied were physiologically and anatomically characterized (see METHODS). Representative transverse sections of dorsal medulla at levels 1.0-1.1 mm caudal to obex. Sites that were effective in attenuating cardiopulmonary reflex are indicated by asterisks. Microinjection of either 8-BrcAMP (n = 3) or 5-MeOT (n = 3) 1.0 mm rostral to these sites was ineffective in modulating cardiopulmonary reflex (sites not shown). Sections are adapted from Barraco et al. (3). C, central canal; Cu, cuneate nucleus; Gr, gracile nucleus; NTS, nucleus tractus solitarii; ts, solitary tract; 10, dorsal vagal motor nucleus; 12, hypoglossal motor nucleus.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study provides evidence that a microinjection of 5-MeOT within the NTS attenuates both the respiratory and bradycardic components of the cardiopulmonary reflex. This effect can be reversed by selective 5-HT₄-receptor antagonists. Our data also indicate that this attenuation, at least on the cardiac component of the reflex, can be blocked by antagonizing cAMP-dependent PKA activity with Rp-cAMPS.

Specificity of microinjection technique. Microinjection techniques provide a useful tool for understanding both the site of action and physiological effects of a neuromodulator at the systems level. We fully acknowledge the limitations of this technique, particularly in the precise location, specificity of drug action, and physiological significance of the action of an agonist on reflex function (20, 24). We carried out numerous controls. First, all effective sites recovered were found to be within the NTS at regions 0.9-1.1 mm caudal to obex as revealed by deposits of Pontamine sky blue dye (Fig. 7). Second, to test for site specificity we microinjected drugs into regions up to 1 mm rostral to those that attenuated the cardiopulmonary reflex. These sites failed to change the magnitude of the reflex, indicating that the effects observed were relatively localized to NTS regions caudal to the obex. Moreover, these effective regions coincided with NTS sites found essential for mediating (5, 34,) and modulating (6, 30) the cardiopulmonary reflex in the rat but are more caudal than those reported by others (33). However, they overlap with NTS regions containing neurons responding to cardiopulmonary vagal afferent stimulation in both rat (36) and mouse (28). Furthermore, they are analogous to those containing pulmonary vagal C-fiber terminals in the cat (18). Third, the dose of 5-MeOT used in the present study produced submaximal attenuating effects (40%) on cardiorespiratory reflex responses (unpublished data). Finally, it is most unlikely that the attenuating effects observed on the cardiopulmonary reflex resulted from a volume artifact because microinjection of both vehicle and other drugs (for example, cAMP and Y-25130) failed to affect the reflex response.

Cardiopulmonary reflex respiratory responses. In a recent study from our laboratory we described an increase in expiratory interval after PBG stimulation of the cardiopulmonary reflex because this pattern of response was recorded most consistently (6). This should not be interpreted as implying that we never saw a tachypnea comparable to that reported in the present study (see, e.g., Fig. 4 of Ref. 6). Indeed, in the present study we also observed increases in expiratory time, but it was the reflex increase in frequency of phrenic nerve activity that was most consistently observed. The reason for this difference is unclear. However, we are not alone in reporting inconsistent respiratory effects following cardiopulmonary receptor stimulation (7, 8, 25, 36). Certainly, differences in dose and sensitivity of an animal to PBG, as well as the depth of anesthesia, may all contribute to producing different patterns of reflex respiratory response both between and within studies. All told, both the increased expiratory interval and tachypnea responses are likely to be manifestations of the same reflex effect on the ventral respiratory network, i.e., an oscillation between early preinspiratory vs. postinspiratory neurons (27, 29) with the resultant effect dependent on the balance of activity between these oscillating neurons (27).

Effects of forskolin and 8-BrcAMP in NTS on cardiopulmonary reflex. Stimulation of adenylate cyclase with forskolin or microinjection of 8-BrcAMP into the NTS both significantly attenuated the bradycardic component of the cardiopulmonary reflex; the reflex increase in phrenic nerve activity cycle length was also attenuated by 8-BrcAMP. Moreover, the cardiac effect was dependent on a cAMP-mediated increase in PKA activity as revealed by Rp-cAMPS. Although not quantitatively analyzed, a similar trend was found with the respiratory response.

Evidence for a role for 5-HT₄ receptors in NTS. Our data indicate that an NTS microinjection of 5-MeOT caused an attenuation of the cardiac and respiratory components of the cardiopulmonary reflex. The absence of a selective 5-HT₄-receptor agonist meant that we had to employ 5-MeOT. Because this compound has partial agonistic effects for a number of 5-HT receptors (see RESULTS and Ref. 37), it was important to demonstrate which 5-HT receptor subtype was involved. Our results indicate that the 5-MeOT-induced depression of the cardiopulmonary reflex was mediated by 5-HT₄ receptors because it was attenuated by two selective 5-HT₄-receptor antagonists (RS-23597 and RS-39604; see Refs. 14, 15, 17). In addition, the cardiopulmonary reflex was not affected by stimulation of other receptors also coupled to adenylate cyclase: neither -adrenoceptor nor D₁ receptor activation altered any measured component of the reflex response. It is possible that the doses of the agonists used were inadequate. However, the doses of isoprenaline we used were comparable to those previously employed both in the NTS (26) and other central nervous structures (22). Additionally, we used a concentration of SKF-38393 comparable to that used by others (31).

Plausible mechanism following 5-HT₄ receptor stimulation in NTS. It is known that 5-HT₄ receptors are coupled to adenylate cyclase causing intracellular levels of cAMP to rise (10, 32). If 5-MeOT causes activation of adenylate cyclase in the NTS to attenuate the bradycardic component of the cardiopulmonary reflex, one might expect this effect to be mimicked by both forskolin and 8-BrcAMP. Indeed, not only did forskolin and 8-BrcAMP produce a similar attenuation of the bradycardic component of the cardiopulmonary reflex, but the duration of effect was identical to that seen with 5-MeOT (see Fig. 6). We speculate that 5-MeOT, forskolin, and 8-BrcAMP may exert their effects via a common mechanism. Furthermore, the fact that Rp-cAMPS blocked the attenuating effect of 5-MeOT provides evidence that the 5-MeOT-mediated depression of the cardiac and respiratory components of the cardiopulmonary reflex results from activation of an intracellular cAMP-dependent PKA pathway within the NTS. However, our techniques at the systems level do not allow us to discern whether the NTS neurons with 5-HT₄ receptors are the same cells that contain the cAMP-PKA pathway.

Plausible physiological significance of 5-HT₄ receptors in NTS. The absence of effect of both Rp-cAMPS and 5-HT₄-receptor antagonists on the cardiopulmonary reflex indicates that there is no tonic cAMP-dependent PKA activity or 5-HT₄ receptor-mediated modulation, respectively, in the anesthetized rat. However, this does not rule out a physiological role for either cAMP or 5-HT₄ receptor modulation in the NTS. Indeed, anesthesia may depress this pathway and it might be expected that in a decerebrate or conscious animal there is tonic serotonergic "tone" acting on 5-HT₄ receptors to stimulate adenylate cyclase, thereby increasing intracellular cAMP-dependent PKA activity. Thus, under certain physiological or behavioral situations, 5-HT₄ modulation may be switched on to produce a short-term depression of this reflex. At present, it is not clear under what circumstances this might occur. Future experiments must identify plausible mechanisms that can activate 5-HT₄ receptors in the NTS; these may include both peripheral afferent and/or central projections. Once these mechanisms are identified, the physiological relevance of the 5-HT₄ receptor-mediated depression of the cardiopulmonary reflex may become more obvious.