Introduction

Top Abstract Introduction Methods Results Discussion References

Several previous studies from this laboratory have focused on the characterization of the hypertensive response evoked through pharmacological stimulation of a spinal muscarinic receptor pressor pathway (Buccafusco and Magrí, 1990; Feldman et al., 1996; Feldman and Buccafusco, 1993a; Takahashi and Buccafusco, 1991a; 1992). In rats, administration of the direct muscarinic receptor agonist carbachol or the cholinesterase inhibitor neostigmine into the brain or spinal cord produces a significant vasopressor response that is reversed by administration of atropine (for review, see Buccafusco, 1996). The spinal segments involved in this cholinergic-sympathoexcitatory response were first identified through topical administration of neostigmine to the surface of the spinal cord. Neostigmine elicited the greatest pressor response (with relatively no change in HR) when the drug was applied between T7 and T11. These experiments were confirmed in subsequent studies in which carbachol was microinjected into discrete spinal sites at the T2 and T11 levels. Injection of carbachol directly into the IML at the T11 level failed to elicit a pressor response. However, increases in BP were observed after microinjection of the agonist into sites medial to the IML and throughout the ventral horn and lamina 7. The results of these experiments were consistent with those from anatomical studies by other investigators (Barber et al., 1984) in which the presence of relatively high levels of choline acetyltransferase were described in the areas we found to be sensitive to carbachol (Takahashi and Buccafusco, 1992). Thus, the pressor response obtained after spinal muscarinic receptor stimulation was not likely the result of direct activation of spinal preganglionic cell bodies located in the IML.

Cholinergic innervation within the rostral ventrolateral medulla provides tonic sympathoexcitatory output (Arneric et al., 1990) through interactions with a descending glutamatergic pathway (Bazil and Gordon, 1991; Gordon and McCann, 1988; Morrison et al., 1989). Although immunohistochemical evidence for a descending sympathetic glutamatergic sympathetic pathway is lacking, substantial pharmacological and electrophysiological evidence exist to support the concept (see Gordon, 1995). A component of the pressor response to spinal muscarinic receptor stimulation is mediated locally through facilitation of descending vasomotor tone (Takahashi and Buccafusco, 1992). We estimated that 40% to 50% of the pressor response to spinal stimulation of muscarinic cholinergic receptors is mediated within the spinal cord, whereas the remainder can be abolished by selective blockade of medullary muscarinic receptors (Feldman and Buccafusco, 1993a), a finding that is consistent with the possibility of an ascending neural link between spinal and medullary cholinergic vasomotor pathways. Preliminary studies from this laboratory have suggested that the pressor response to stimulation of spinal muscarinic receptors is mediated in part through the NMDA subtype of glutamate receptors (Feldman and Buccafusco, 1993b). In fact, stimulation of cholinergic receptors has been reported to enhance glutamatergic activity in the central nervous system (Andre et al., 1993; Markram and Segal, 1990).

The purpose of this study was to pharmacologically characterize the interactions between spinal cholinoceptive sites and amino acid neurotransmitters in the facilitation of descending spinal vasomotor and cardiac function. Activation of spinal muscarinic receptors was accomplished through direct i.t. administration of a standard dose of carbachol. Rats were pretreated by i.t. or i.c. injection with selective ligands for subtypes of glutamate or GABA receptors, and the cardiovascular response to subsequent injection of carbachol was measured. Experiments were performed in both conscious, freely moving and anesthetized rats. The former experiments were used to confirm the pharmacological responses without the potential for interference with the effects of general anesthesia. The latter experiments were performed in a preparation in which i.t. or i.c. drug injectates could be restricted in their distribution to specific spinal segments or to the lower medulla, respectively.

	Methods

Top Abstract Introduction Methods Results Discussion References

Experimental animals. Male Wistar rats (Harlan Sprague-Dawley, Indianapolis, IN), weighing 280 to 380 g were housed in an environmentally controlled room on a 12-hr/12-hr day/night cycle and were maintained on Wayne Rodent Blox and tap water. All animal protocols were previously approved by the Medical College of Georgia Committee on Animal Use for Research and Education.

Implantation of a chronic i.t. catheter. Rats were anesthetized with methohexital (65 mg/kg) and placed in a stereotaxic frame. Catheterization of the spinal subarachnoid space was performed under aseptic conditions by inserting a sterile saline-filled polyethylene (PE₁₀) catheter caudal to a microincision in the atlanto-occipital membrane. The catheter was advanced to the T12 level of the spinal cord (length was adjusted for each animal in accordance with weight and age). The distal end of the catheter was plugged with 30-gauge stainless steel wire and anchored to the skull with acrylic cement after being threaded through an occipital burr hole. Each rat was allowed a 2-day recovery period. Only normally moving, healthy animals were used in subsequent experiments. On the completion of an experiment, catheter placement was confirmed by dye injection and dorsal laminectomy.

Implantation of an indwelling aortic catheter. After recovery from surgery, rats were again anesthetized with methohexital, and a midsagittal incision was made in the abdomen. The left iliac artery was then exposed, and a polyethylene catheter (PE₅₀) filled with heparinized saline (20 units/ml) was inserted into the abdominal aorta below the origin of the renal arteries. The distal end of the catheter was directed subcutaneously to emerge at the back of the neck. The catheter was passed through a spring support and connected to a watertight swivel mounted 300 mm above the cage floor. This surgical procedure allowed the chronically catheterized rat unrestricted movement to all areas of the cage for the duration of the experiment while receiving a constant infusion of heparinized saline (8 ml/day). Each rat was allowed a 2-day recovery period after the surgery. For BP measurements, the chronic indwelling catheter was connected to a pressure transducer coupled to a thermal array recorder, and the analog signals were amplified and digitized on a Buxco Electronics LS-14 Logging Analyzer. The analyzer provided 1-min average measurements of MAP and HR to a computer. Stable base-line MAP and HR were measured for at least 10 min before treatments to obtain base-line levels.

Femoral artery cannulation in anesthetized rats. Anesthesia was achieved with urethane (.86 g/kg intraperitoneal) that was supplemented with inhalational halothane. A PE₅₀ catheter filled with heparinized saline (2 units/ml) was inserted into the right femoral artery. BP was monitored continuously throughout the experiment. HR was obtained from the pressure pulses and displayed as an integrated signal by using a biotachometer. Intravenous access was accomplished by inserting a PE₅₀ catheter into the left femoral vein or left subclavian vein.

Placement of an i.t. catheter in anesthetized rats. In rats prepared with intra-arterial and intravenous lines as described above, halothane administration was terminated, and an additional i.v. injection of urethane (.30 g/kg) was administered to maintain anesthesia for the duration of the experiment. The i.t. cannulation procedure described below was developed to ensure that subsequent drug distribution was restricted to the thoracolumbar spinal cord (Feldman and Buccafusco, 1993a). Rats were placed in a stereotaxic frame, and the musculature between C7 and T3 was reflected. A dorsal laminectomy was performed at the level of T1-2, and the dura was longitudinally transected. The incision was extended to the lateral borders of the exposed spinal cord. A saline-filled catheter (PE₁₀) was then inserted 50 to 53 mm from the most caudal portion of the exposed cord to terminate in the T11-12 region. The length of the catheter was determined empirically based upon the weight of the animal. CSF was collected with microsponges (Alcon Surgical) to prevent redistribution of drug solution from the local spinal site to rostral brain regions. Artificial CSF was given to replace fluid loss. The i.t. injections were administered using a methodology similar to that described for the freely moving animals. Body temperature was maintained at 37°C throughout the experiment. The i.v. administration of 0.045% sterile saline plus 15 mEq KCl (5 ml/kg/hr) was used to maintain the animal's water balance and to help replace insensible losses. Tracheal intubation (PE₂₄₀ tubing) was performed to preserve normal respiratory dynamics. Arterial blood gases and blood glucose levels were routinely measured and were maintained. Animals were maintained with 2 l/min of room air supplemented with 100% O₂. Any break in the vascular integrity of the spinal cord or surrounding vessels terminated the experiment. Proper placement of the catheter was confirmed at necropsy.

Placement of an i.c. catheter in anesthetized rats. After the laminectomy, the dura was visualized between the occipital bone and C1. A 24-gauge needle fastened to PE₁₀ tubing was inserted past the dura into the subdural space. Paraffin wax was placed around the catheter and was melted with microelectric cautery. The catheter was stabilized by the addition of a thin coating of fibrin glue. Proper placement was ensured by visualizing CSF flow past the placement of the catheter. The i.c. injections were delivered directly onto the surface of the medulla. To ensure that drug distribution after i.c. injection did not go beyond the T1-2 level, microsponges were placed on and around the exposed spinal cord to absorb any leaking CSF or drug solution, and the lost volume was replaced by both i.c. and i.t. injection of artificial CSF. The i.c. injections were made in a manner similar to i.t. drug administration.

Drugs. All drugs administered by i.t. or i.c. injection were dissolved in sterile-isotonic saline and were pH balanced. Carbachol chloride, atropine methyl sulfate and baclofen were purchased from Sigma Chemical (St. Louis, MO). D-AP7, MK801 (dizocilpine maleate) and CNQX were purchased from Research Biochemicals (Natick, MA).

Statistical analysis. The data depicted in the text and figures are presented as mean ± S.E.M. values. The existence of a significant difference between or among experimental groups was determined by analysis of variance with repeated measures using the raw data. A modified t test with Bonferroni's correction for multiple comparisons using the error mean-square term from the analysis of variance was used for analysis between groups of data. In some experiments, drug pretreatments altered the base-line values significantly from control. In these particular experiments, data were normalized before statistical analysis to ensure that the pretreatment base-line values would not confound any comparison between various treatment groups. The criterion for statistical significance was P < .05 for all comparisons. The AUC for time course data (including summation of both positive and negative areas, relative to base line) was determined by using the normalized (change from control) data obtained from 0.1 to 30 min after carbachol injection. AUC was determined by plotting data over a grid (grid block = 30 sec × 2 mmHg or 5 beats/min). Each AUC was then divided into incremental trapezoids. The areas of the trapezoids were determined individually and summed to determine the total AUC.

	Results

Top Abstract Introduction Methods Results Discussion References

Cardiovascular response to i.t. injection of carbachol in unanesthetized rats. In our previous studies, i.t. injection of carbachol was shown to evoke an atropine-reversible, dose-dependent increase in BP and HR (Buccafusco and Magrí, 1990; Magrí and Buccafusco, 1988: Marshall and Buccafusco, 1987). In rats pretreated with saline, subsequent i.t. administration of carbachol (5 µg, 27 nmol) elicited a reproducible peak increase in MAP of between 30 and 40 mm Hg and a duration of response of 30 to 45 min. The increase in MAP peaked at 4 min and was within 15 mm Hg of the predrug base-line level by 30 min after injection (Fig. 1). The pressor response to carbachol was accompanied by an immediate increase in HR that peaked at 2 min after injection (~75 beats/min). Immediately after peaking, HR values decreased slightly at the 5-min time point and then returned to the previous peak level (Fig. 2). Because the animals were unanesthetized and baroreceptor reflexes were intact, this biphasic nature to the HR response to i.t. carbachol most likely represents a transient reflex slowing of HR due to the pressor response. Thereafter, HR slowly returned to base-line levels within 45 to 60 min.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   The increase in MAP produced by i.t. injection of carbachol in conscious, freely moving rats. , Saline pretreatment. D-AP7 i.t. pretreatment with () 20, () 100 and () 200 nmol. D-AP7 significantly inhibited the pressor response to i.t. injection of carbachol in a dose-dependent manner. In a separate experimental group (), 5 rats were pretreated with the 200-nmol dose of D-AP7. At the peak of the reduced pressor response (~4 min after i.t. carbachol injection), each received an i.v. injection of 5 mg/kg tolazoline. Tolazoline treatment completely reversed the remainder of the pressor response to carbachol. Results are expressed as mean ± S.E.M. (n = 6 animals/group; P < .05 vs. control).

View larger version (31K): [in this window] [in a new window]   Fig. 2.   The change in HR produced by i.t. injection of carbachol in conscious, freely moving rats. , Saline pretreatment. D-AP7 i.t. pretreatment with () 20, () 100 and () 200 nmol. D-AP7 significantly inhibited the increase in HR to i.t. injection of carbachol in a dose-dependent manner. Results are expressed as mean ± S.E.M. (n = 6 animals/group; P < .05 vs. control).

Effects of pretreatment with D-AP7 on the cardiovascular response to i.t. injection of carbachol in unanesthetized rats. Unanesthetized, freely moving rats were randomly assigned to one of four pretreatment groups: saline (control), D-AP7 (20 nmol), D-AP7 (100 nmol) or D-AP7 (200 nmol). The doses of D-AP7 were selected based on the results of a preliminary dose-finding study (data not shown). After the initial experiment, rats were allowed to recover for a minimum of 4 days. They were then reassigned to receive a second pretreatment dose on a random basis. No rat was involved in more than two experiments. D-AP7 was administered by i.t. injection 20 min before the i.t. injection of carbachol. The i.t. pretreatment with 200 nmol of D-AP7 significantly lowered base-line MAP over the first 10 min after injection compared with corresponding saline controls. The two higher doses of D-AP7 significantly lowered resting MAP during the entire 20-min pretreatment period. Although there was no significant change in base-line HR after injection of either the 20- or 100-nmol doses of D-AP7, the 200-nmol dose significantly decreased resting HR during the pretreatment period (Table 1). However, these decreases in base-line MAP and HR to D-AP7 amounted to only 11% and 7% changes from predrug levels, respectively.

Effects of pretreatment with baclofen on the cardiovascular response to i.t. injection of carbachol in unanesthetized rats. The GABA_B receptor agonist baclofen was administered by i.t. injection to freely moving animals to determine whether spinal GABA receptors play a role in the expression of the pressor response to spinal cholinergic stimulation. Baclofen was administered 15 min before i.t. injection of carbachol. Throughout the pretreatment period, baclofen did not significantly alter resting MAP or HR (Table 1), nor did the drug produce overt impairment of motor function. The i.t. pretreatment with baclofen resulted in a decrease in the magnitude of the pressor response elicited by i.t. injection of carbachol (Fig. 3). Although the time profiles of the carbachol-induced pressor responses for both control and baclofen-pretreated groups were similar, the baclofen-carbachol responses approached base-line values after only 15 min. The respective AUCs for the 0.2- and 1.0-nmol baclofen-pretreated groups were reduced by 69% (206.8 ± 27.5 mm Hg/30 min) and 77% (151.5 ± 36.2 mm Hg/30 min), respectively. The 0.2-nmol dose of baclofen abolished the tachycardia normally associated with the pressor response to carbachol. In rats pretreated with 1 nmol of baclofen, the inhibitory component of the biphasic HR response to carbachol (in control rats) was exaggerated (Fig. 3, inset).

View larger version (36K): [in this window] [in a new window]   Fig. 3.   The increase in MAP produced by i.t. injection of carbachol in conscious, freely moving rats. , Saline pretreatment. Baclofen i.t. pretreatment with () 0.2 and () 1.0 nmol. Each dose of baclofen significantly inhibited the pressor response to i.t. injection of carbachol. Inset, accompanying HR response. Baclofen significantly inhibited the increase in HR to i.t. injection of carbachol in a dose-dependent manner. Results are expressed as mean ± S.E.M. (n = 6 animals/group; P < .05 vs. control).

Cardiovascular response to i.t. injection of carbachol in anesthetized rats. After i.t. pretreatment with saline (20 min), carbachol (27 nmol) produced an increase in MAP that was qualitatively similar to that produced in unanesthetized rats (Figs. 1 and 4). Although the magnitude and duration of the pressor response were similar in both unanesthetized and anesthetized animals, the peak response was slightly delayed and prolonged (4 min after injection) in the anesthetized group. Unlike the response in unanesthetized rats, the pressor response to carbachol was not accompanied by a significant change in HR (Fig. 4, inset).

View larger version (37K): [in this window] [in a new window]   Fig. 4.   The increase in MAP produced by i.t. injection of carbachol in urethane-anesthetized rats. , Saline pretreatment. D-AP7 i.t. pretreatment with () 10, () 100 and () 1000 nmol. D-AP7 significantly inhibited the pressor response to i.t. injection of carbachol in a dose-dependent manner. Inset, accompanying HR response. The i.t. injection of carbachol evoked an increase in HR in rats pretreated with each of the three doses of D-AP7. Results are expressed as mean ± S.E.M. (n = 4 to 6 animals/group; P < .05 vs. control).

Effect of pretreatment regimens on base-line MAP and HR in anesthetized rats. Because the reproducibility of the cardiovascular response to i.t. injection of carbachol was consistent throughout all control experiments, a group of saline-pretreated rats were randomly interspersed among the three primary experimental groups. Therefore, 2 to 3 rats were used in each of the groups. This design served to both limit the numbers of control animals required for the study and provide confirmation that the typical control responses to i.t. injection of carbachol were obtained longitudinally throughout the study. Each animal was used only once.

Effect of pretreatment with D-AP7 on the cardiovascular response to i.t. injection of carbachol in anesthetized rats. Anesthetized rats were not used more than once in the following experiments. The i.t. pretreatment with 10 nmol of D-AP7 did not significantly alter the pressor response to subsequent i.t. injection of carbachol (P = .133). Both the 100- and 1000-nmol doses significantly inhibited the expression of the pressor response to i.t. carbachol (Fig. 4). After the two higher doses of D-AP7, carbachol evoked a transient depressor response (lasting ~2 min), followed by an attenuated pressor response that returned to base line within 20 min. Pretreatment with the 10-nmol dose of D-AP7 produced a 33% reduction in the AUC for the carbachol response (379.5 ± 46.7 mm Hg/30 min compared with the saline control mean of 575 ± 42.6 mm Hg/30 min). Pretreatment with the 100- or 1000-nmol doses resulted in an 80% (109.7 ± 40.1 mm Hg/30 min) and 85% (86.5 ± 23.6 mm Hg/30 min) reductions in the expression of the carbachol pressor response, respectively.

Effect of pretreatment with MK801 on the cardiovascular response to i.t. injection of carbachol in anesthetized rats. In the next group of experiments, MK801 was used to further confirm the role of spinal NMDA receptors in the expression of the cardiovascular response to i.t muscarinic receptor stimulation. The i.t. pretreatment with MK801 reduced the pressor response to i.t. injection of carbachol in a dose-dependent manner (Fig. 5). In MK801-pretreated rats, the initial response to carbachol was a 10 to 20 mm Hg decrease in BP, which returned to base line within 2 to 3 min. MK801-pretreated animals also exhibited a prolonged time to peak for the carbachol response. Moreover, the carbachol responses elicited after 10- and 100-nmol doses of MK801 had returned to within 10 mm Hg of base-line MAP within 15 min. Pretreatment with 10 nmol of MK801 resulted in a 53% (268.1 ± 53.6 mm Hg/30 min) reduction in the pressor response to i.t. carbachol compared with control. For the higher doses, MK801 decreased the carbachol-induced hypertensive response by 63% and 73% (211.9 ± 45.1 and 153.6 ± 50.8 mm Hg/30 min). In contrast to D-AP7, carbachol evoked no significant changes in HR after MK801 pretreatment (Fig. 5, inset). The respective AUC values for HR were 32.5 ± 110.9 beats from base line/30 min for saline and 225 ± 33.9, 122 ± 37.1 and 126.3 ± 113.8 beats from base line/30 min each of the three doses of MK801, respectively (P = .10, .83 and .74, respectively). Two rats received MK801 (200 nmol) as an i.v. pretreatment. The pressor response to subsequent i.t. injection of carbachol was decreased by ~25%; resting HR was not affected.

View larger version (39K): [in this window] [in a new window]   Fig. 5.   The increase in MAP) produced by i.t. injection of carbachol in urethane-anesthetized rats. , Saline pretreatment. MK801 i.t. pretreatment with () 10, () 100 and () 1000 nmol. MK801 significantly inhibited the pressor response to i.t. carbachol in a dose-dependent manner (P < .05 vs. control). Inset, accompanying HR response. MK801 did not significantly alter the HR response to i.t. injection of carbachol. Results are expressed as mean ± S.E.M. (n = 4 to 6 animals/group).

Effect of pretreatment with CNQX on the cardiovascular response to i.t. injection of carbachol in anesthetized rats. In the next series of experiments, CNQX was used to produce blockade of spinal cord quisqualate/kainate glutamate receptors. Due to the hydrophobic nature of CNQX, a DMSO/saline combination (60:40) was used to solubilize the drug. Therefore, two sets of controls were used in this regimen to circumvent the possible action of DMSO alone. The i.t. pretreatment with either saline alone or with the DMSO/saline mixture did not significantly alter the pressor response to i.t. injection of (Fig. 6). The i.t. injection of CNQX (10 or 100 nmol) produced no significant change in resting MAP or HR (Tables 2 and 3). Moreover, neither dose of CNQX significantly altered the pressor response to i.t. injection of carbachol (459 ± 70.9 and 464 ± 35 mm Hg/30 min, control vs. CNQX pretreatment, respectively); P = .0821 and .2311). CNQX pretreatment resulted in carbachol-evoked dose-dependent changes in the HR (Fig. 6, inset). Pretreatment with the 10-nmol dose resulted in a profound bradycardia during the first 3 min after carbachol injection. Thereafter, HR quickly returned toward base line but never reverted completely during the observation period. In contrast, pretreatment with the 100-nmol dose of CNQX resulted in a mild tachycardia after carbachol administration, which was statistically greater than the response observed in rats pretreated with saline or DMSO/saline.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   The increase in MAP produced by i.t. injection of carbachol in urethane-anesthetized rats. , Saline pretreatment. CNQX i.t pretreatment with () 10 or () 100 nmol. , DMSO/saline control (60:40 DMSO/saline). CNQX did not significantly alter the pressor response to i.t. injection of carbachol. Results are expressed as mean ± S.E.M. (n = 4 to 6 animals/group; P > .05 vs. control). Inset, accompanying HR response. After pretreatment with CNQX, i.t. injection of carbachol produced a significant change in HR. The 10-nmol pretreatment dose of CNQX resulted in a carbachol-induced bradycardia, whereas the 100-nmol dose resulted in a carbachol-induced tachycardia. Results are expressed as mean ± S.E.M. (n = 4 to 6 animals/group; P < .05 vs. control).

Effect of pretreatment with i.c. injection of atropine or D-AP7 on the pressor response to i.t. injection of carbachol. In the last series of experiments, we sought to determine whether medullary muscarinic or NMDA receptors participated in the expression of the cardiovascular response to i.t. injection of carbachol. Rats were pretreated by the i.c. route of administration with either methylatropine or D-AP7. Although i.t. injection of D-AP7 (100 nmol) lowered base-line MAP (29.7 mm Hg), i.c. D-AP7 (100 nmol) actually increased base-line MAP by 12 mm Hg (Table 2). After i.c. pretreatment with D-AP7, MAP remained elevated compared with saline controls before the administration of carbachol. The resting level of HR remained unchanged (Table 3). The i.c. pretreatment with 26 nmol of methylatropine produced no significant change in resting MAP or HR. The i.c. pretreatment with either methylatropine or D-AP7 significantly inhibited the expression of the pressor response to i.t. injection of carbachol (Fig. 7). In methylatropine-pretreated rats, carbachol produced a peak pressor response of only 26 mm Hg just after injection that rapidly diminished thereafter. The pressor response to carbachol after D-AP7 pretreatment peaked (10 mm Hg) 3 to 5 min after injection and returned to base line within 15 min (Fig. 7). Carbachol AUC mean values for methylatropine- and D-AP7-pretreated rats were 341 ± 75.1 and 80.1 ± 49.5 mm Hg/30 min, respectively.

View larger version (29K): [in this window] [in a new window]   Fig. 7.   The increase in MAP produced by i.t. injection of carbachol in urethane-anesthetized rats. , Saline pretreatment. , Methylatropine (10 mg/10 µl) i.c. pretreatment. , D-AP7 (100 nmol) i.c. pretreatment. The i.c. pretreatment with atropine or D-AP7 significantly inhibited the pressor response to i.t. injection of carbachol. Results are expressed as mean ± S.E.M. (n = 4 to 6 animals/group; P < .05 vs. control). There were no changes in HR from base line produced by carbachol after either pretreatment (data not shown).

	Discussion

Top Abstract Introduction Methods Results Discussion References

The ability of the muscarinic receptor agonist carbachol to elicit an increase in MAP after i.t. injection into the lower thoracic level of the spinal cord has been pharmacologically characterized in our earlier studies (for a review, see Buccafusco, 1996). The ability of carbachol to elicit a concomitant tachycardic response in conscious rats suggests that stimulation of spinal muscarinic receptors results in enhanced cardiac sympathetic activity, withdrawal of vagal tone or both. This effect is sufficient to override the influence of vascular baroreceptor stimulation resulting from the accompanying pressor response. The influence of the baroreceptor reflex in the HR response to i.t. injection of carbachol is suggested by the observation that immediately after peaking, HR values decreased slightly and then returned to the previous peak level. This biphasic nature to the HR response to carbachol most likely represents a transient reflex slowing of HR caused by the pressor response. However, the independence of spinal cardiac pathway from the vasomotor pathway is suggested both by the difference in anatomical origin of each response (as discussed below) and by the observation that the HR response usually outlasted the pressor response.

Spinal sites responsive to the pressor-evoking effects of carbachol are not relegated to the site of origin of preganglionic neurons; rather, activation of spinal muscarinic receptors located in more medial regions of the lower thoracic spinal cord facilitates descending vasomotor tone (Takahashi and Buccafusco, 1992). However, ~50% of the expression of the pressor response observed after i.t. administration of carbachol is mediated through an ascending pathway, which, in turn, communicates with other cholinoceptive sites in the lower medulla (Buccafusco, 1990; Buccafusco and Magií, 1990; Feldman and Buccafusco, 1993a), possibly within the RVL. The experimental design of the present study was based on some of the neurotransmitter interactions within the RVL that are known to contribute to the maintenance of vasomotor tone. For example, citing primarily anatomical evidence, Arneric et al. (1990) suggested that cholinergic projection neurons terminating within the RVL inhibit the function of GABA interneurons that provide tonic inhibitory tone to bulbospinal, possibly glutamatergic vasomotor neurons. Accordingly, stimulation of muscarinic receptors has been reported to alter membrane potential and increase the firing rate of glutamatergic neurons long after muscarinic receptor activation has ceased (Agarwal and Calaresu, 1992; Andre et al., 1993; Downing and Kaneko, 1992; Gonzales et al., 1993). Vasomotor neurons originating within the RVL participate in generating tonic sympathetic discharge from the central nervous system (Bazil and Gordon, 1991; Gordon and McCann, 1988; Heary et al., 1993; Hong et al., 1994; Llewellyn-Smith et al., 1992; Morrison et al., 1989). That a population of these vasomotor neurons may be glutamatergic is supported by the results of anatomical studies that show a direct glutamatergic connection between the RVL and preganglionic sympathetic neurons in the IML (Llewellyn-Smith et al., 1992; Morrison et al., 1989).

The central administration of NMDA receptor agonists and antagonists have been reported to increase and decrease BP, respectively (Bazil and Gordon, 1991; Gilbey and Spyer, 1993; Gordon, 1995; Kao et al., 1991). Our results indicate that the receptor antagonists D-AP7 and MK801 each decreased base-line BP and HR in both conscious and anesthetized rats. In fact, this is the first study to report that blockade of NMDA receptors decreases both resting BP and HR in conscious, freely moving rats. However, the magnitude of the decreases were not dramatic and certainly not consistent with an ability to completely prevent the transmission of central sympathetic tone to peripheral pressor systems. The i.t. pretreatment with D-AP7 in conscious rats inhibited the expression of the BP and HR responses evoked by i.t. injection of carbachol in a dose-dependent manner. In no case was the cardiovascular response to carbachol completely abolished. These findings support the hypothesis that the vasomotor responses normally observed after i.t. administration of a muscarinic receptor agonist are mediated, at least in part, by excitatory amino acid-induced stimulation of the NMDA subtype of glutamate receptors. However, it is likely that other descending or local neurotransmitters play a role in helping to maintain spinal sympathetic tone (see Gilbey and Spyer, 1993). This possibility is supported by our observation that the remainder of the D-AP7 inhibited carbachol pressor response was mediated by sympathetic alpha adrenergic nervous tone.

The i.t. administration of carbachol in anesthetized rats allowed us to localize the drug solution to the lower thoracic spinal cord. Similar profiles (magnitude and duration) of carbachol-induced pressor responses were produced under control (i.t. saline pretreatment) conditions in both conscious and anesthetized animals. The inability of carbachol to elicit a significant tachycardic response in anesthetized rats was expected because muscarinic cholinoceptive sites are somatotopically organized so that tachycardic responses to carbachol are elicited predominantly at cervical and upper thoracic levels (Marshall and Buccafusco, 1987; Takahashi and Buccafusco, 1991a, 1992). The selectivity of response observed to carbachol in anesthetized rats further supports the contention that drug localization to lower thoracic regions did indeed occur in anesthetized rats. The fact that blockade of lower thoracic spinal NMDA receptors in anesthetized rats produced a decrease in HR as well as in BP suggests that the influence of these receptors may not be somatotopically organized as are the responses to muscarinic receptor stimulation.

In keeping with its effects in conscious rats, i.t. pretreatment with D-AP7 inhibited the expression of the pressor response to carbachol in anesthetized rats. A higher dose of D-AP7 was required to attenuate the pressor response to carbachol in anesthetized rats compared with that in conscious rats. The reason for this difference in glutamate receptor antagonist sensitivity is unclear, however, it may reside in the neural effects of the urethane anesthesia. The ability of i.t. pretreatment with the non-competitive NMDA receptor antagonist MK801 to block the pressor response to i.t. injection of carbachol in anesthetized rats supported the results with D-AP7. As with D-AP7 pretreatment, carbachol did not significantly alter resting HR after MK801 pretreatment. HR did increase after carbachol injection when animals were pretreated with the 100 nmol dose of D-AP7, although the significance of the results for this one dose of D-AP7 remain unclear. Intravenous administration of MK801 appeared to inhibit the pressor response to i.t. injection of carbachol, but the magnitude of the response was less than half of the response observed after i.t. injection of a equimolar dose of MK801. Therefore, it is unlikely that an appreciable component of the depressor response to i.t. MK801 could be attributed to redistribution of the drug to the systemic circulation. Rather, the small inhibitory action produced after the i.v. administration of MK801 may be accounted for entry of the drug into the central nervous system and/or its weak ganglionic nicotinic inhibitory activity (Amador and Dani, 1991; Lewis et al., 1989).

The results of experiments with the non-NMDA antagonist CNQX did not support a primary role for kainate/quisqualate receptors in the maintenance of resting BP because the drug did not produce any significant effect on base-line BP or HR. The pressor response to i.t. injection of carbachol was not significantly different from that for either saline or DMSO/saline controls after pretreatment with 10 or 100 nmol of CNQX. However, carbachol evoked a significant fall in HR in the group of rats pretreated with the lower (10 nmol) dose of CNQX. The explanation for this effect of carbachol is not apparent because as discussed above, carbachol-evoked HR changes occur at higher thoracic levels than those being affected in the anesthetized rat preparation. One possibility is that the DMSO vehicle may have altered the distribution pattern of carbachol. This possibility does not obviate the results of the BP component of the response because (as with the conscious preparation) the pressor response to carbachol occurs at both high and low thoracic levels.

From our earlier studies, we accumulated evidence to suggest that the pressor response to spinal muscarinic receptor activation is dependent to a large extent on an independent population of muscarinic receptors located in the lower medulla (Buccafusco, 1996). Some of the more pertinent findings include the observations that (1) the pressor response to spinal muscarinic receptor stimulation was not affected by midcollicular transection but was almost abolished by prior transection of the cervical spinal cord (Marshall and Buccafusco, 1987; Takahashi and Buccafusco, 1991a); (2) stimulation of spinal muscarinic receptors produces an atropine-sensitive facilitation of descending vasomotor activity evoked through electrical stimulation of the lateral funiculus (Takahashi and Buccafusco, 1992); (3) the pressor response to carbachol and neostigmine can be elicited (albeit at different levels of activation related to muscarinic receptor density) from cervical through upper lumbar levels (Takahashi and Buccafusco, 1992) and (4) ~50% of the pressor response to spinal cholinergic receptor stimulation was shown to be dependent on higher centers because selective blockade of medullary muscarinic receptors with atropine significantly reduced the magnitude of the pressor response to i.t. injection of carbachol or neostigmine (Feldman and Buccafusco, 1993a, 1993b). In further support of the possibility that the pressor response to spinal muscarinic receptor stimulation is mediated through an interaction with higher centers, we found in the present study that i.c. pretreatment with either atropine or D-AP7 partially inhibited the response to i.t. injection of carbachol. Also, a major descending cholinergic projection to the spinal cord has not been described (Sherriff et al., 1991; Woolf, 1991).

Although there have been no systematic anatomical (tracer) studies of the cholinergic vasomotor neurons that we propose to exist in the spinal cord, one candidate for this possibility is the afferent vasomotor pathway that is activated by electrical stimulation of the sciatic nerve. Stornetta et al. (1989) described elegant physiological and histochemical experiments with results that indicate that the somatic pressor response is mediated over a pathway that ascends via spinoreticular afferents traveling in the anterior or lateral funiculi of the spinal cord and that the C1 area of the RVL is critical region for the integration of the pressor reflex. Accordingly, we reported that the pressor response to sciatic nerve stimulation was modified by i.t. administration of neostigmine (Takahashi and Buccafusco, 1991b). Further studies will be required to determine whether somatic afferent reflex pathways provide the substrate for the spinal cholinergic pressor system that we have described. Alternatively, the anatomical substrate mediating the cardiovascular response to cholinergic agonists could be the central canal or partition cells originally described by Barber et al. (1984), whose intraspinal circuitry has been shown to ascend up to six spinal segments (Sherriff and Henderson, 1994). These choline acetyltranserase-containing cells are located within the regions that we found to be sensitive to intraspinal injection of carbachol (Takahashi and Buccafusco, 1992).

Although a number of explanations exist that may explain our previous and present findings, the data have provided a working hypothesis. Based on the known neural interactions within the RVL (discussed above) that involve cholinergic, GABAergic and glutamatergic pathways, it is possible that this relationship in the medulla exists as the rostral component of a continuum of similar interactions descending in sympathetic pathways. The model we propose as the basis for future experiments holds that in the spinal cord (as in the RVL), stimulation of cholinergic muscarinic receptors inhibits the release of GABA, which is inhibitory to putative glutamatergic cells at the level of spinal preganglionic cells. Disinhibition of glutamate neurons through stimulation of muscarinic receptors results in enhanced sympathetic tone and a pressor response. Although the physiological stimulus for the release of spinal acetylcholine in this pressor pathway is unknown, the possibility exists that this pathway may serve to modify the vasomotor response to a painful stimulus (as discussed above). Two other possibilities include a role for spinal cholinergic pressor neurons in the maintenance of experimental (genetic) hypertension (Buccafusco and Magrí, 1990) or in mediating the vasopressor response elicited during morphine withdrawal (Buccafusco, 1991; Holland et al., 1993; Marshall and Buccafusco, 1987).