Abstract
A potential role for neurotensin in the rostral ventromedial medulla (RVM) in modulation of visceral nociceptive transmission was examined in this study. Microinjection of neurotensin (3–3000 pmol) into the RVM of awake rats produced a dose-dependent inhibition of the visceromotor response (VMR) to noxious colorectal distension (CRD) that lasted 30 to 120 min. Additionally, intra-RVM injection of neurotensin (300 pmol) significantly reduced the slope of the stimulus-response function to graded CRD (20–80 mm Hg), whereas the greatest dose of neurotensin (3000 pmol) completely inhibited the VMR at all intensities of CRD. General motor function was unaffected after intra-RVM injection of neurotensin (3000 pmol). Intra-RVM injection of lesser doses of neurotensin (0.03–0.30 pmol) resulted an enhancement of the VMR to noxious CRD that had a short duration (18–30 min), and produced a leftward shift of the stimulus-response function to graded CRD without a change in the slope of the function. Additionally, intra-RVM injection of the neurotensin-receptor antagonist SR48692 (0.3–300 fmol) in naive animals produced dose-dependent inhibition of VMR to noxious CRD, whereas a lesser dose (0.03 fmol) enhanced the VMR. These data support a role for neurotensin in the RVM in biphasic modulation of visceral nociception. The results obtained with SR48692 suggest that endogenous neurotensin in the RVM modulates VMR to noxious CRD via a prominent interaction with neurotensin receptors that mediate facilitatory influences and a lesser interaction with neurotensin receptors that mediate masked inhibitory influences.
The tridecapeptide neurotensin is distributed throughout the central nervous system (CNS) and has been implicated in several physiological functions, including modulation of nociception. Initial studies examining the effect of neurotensin found the peptide to have an antinociceptive effect in animal models of both cutaneous and visceral pain after administration into the CNS or various specific brain loci (Clineschmidt and McGuffin, 1977; Kalivas et al., 1982; Fang et al., 1987; Behbehani, 1992). Several recent studies, however, have suggested a more complex role for neurotensin in the rostral ventromedial medulla (RVM) in descending modulation of spinal nociceptive transmission.
A role for neurotensin in descending pain modulation from the RVM has been the focus of several studies, based on the fact that high levels of neurotensin and neurotensin-binding sites have been identified in this site (Young and Kuhar, 1981; Jennes et al., 1982). The RVM has been identified as an important site involved in descending modulation of spinal nociceptive transmission (Gebhart and Randich, 1990; Fields et al., 1991; for review, see Fields and Basbaum, 1993). Electrical stimulation, or administration of various receptor-selective agonists into the RVM, has been shown to both inhibit and facilitate spinal behavioral and dorsal horn neuron responses to noxious stimulation, depending on the intensity/nature of the intra-RVM stimulus (Zhuo and Gebhart, 1992, 1997; Urban and Smith, 1993; Thomas et al., 1995; Urban and Gebhart, 1997). These dual influences from the RVM are believed to be mediated by anatomically distinct, independent systems that activate different receptors in the spinal cord (Zhuo and Gebhart, 1990–1992;Urban et al., 1996b).
Injection of neurotensin into RVM has a biphasic effect on behavioral nociceptive responses: greater doses inhibit, whereas lesser doses facilitate thermal nociceptive tail-flick and hot-plate responses (Urban and Smith, 1993). These results were supported in a subsequent study in which greater and lesser doses of neurotensin injected into the RVM were found to inhibit and facilitate, respectively, spinal dorsal horn neuron responses to noxious thermal stimulation (Urban and Gebhart, 1997). The dual actions of neurotensin on spinal nociceptive transmission are anatomically differentiated within the RVM, involve activation of anatomically distinct descending systems, and are mediated by different spinal receptors (Urban and Smith, 1994; Urban et al., 1996b; Urban and Gebhart, 1997). Additionally, the observation that low doses of the neurotensin-receptor antagonist SR48692 selectively blocks the descending inhibitory influence of neurotensin after sequential intra-RVM injection suggests that this receptor antagonist may be used to discriminate neurotensin-receptor subtypes in the RVM involved in the biphasic action of the peptide (Smith et al., 1997; Urban and Gebhart, 1997). Thus, inhibitory and facilitatory influences of neurotensin in the RVM on spinal nociceptive transmission are probably the result of an interaction with multiple neurotensin receptors in the RVM that activate independent descending systems.
Although most studies investigating mechanisms of descending modulation have used models involving noxious cutaneous stimulation, spinal visceral nociceptive transmission has also been found to be subject to descending modulation (Cervero et al., 1985; Ness and Gebhart, 1987). A recent study expanded on this notion by demonstrating both inhibitory and facilitatory influences on visceromotor and spinal dorsal horn neuron responses to noxious colorectal distension after electrical stimulation or glutamate injection into the RVM (M. Zhuo and G.F.G., unpublished observations). Given the role of neurotensin in the RVM in biphasic modulation of cutaneous nociception, this study was designed to expand on this notion by examining an additional role for neurotensin in the RVM in the modulation of visceral nociceptive transmission.
Materials and Methods
Animal Care and Preparation.
Adult male Sprague-Dawley rats (400–450 g; Harlan, Indianapolis, IN) were used in all experiments and housed in the American Association for the Accreditation of Laboratory Animal Care-accredited facility in the Bowen Science Building, University of Iowa. The animals were housed individually, had free access to food and water, and were maintained on a standard light/dark cycle. All experimental protocols were approved by the Institutional Animal Care and Use Committee at the University of Iowa.
Animals were deeply anesthetized with pentobarbital sodium (45–50 mg/kg i.p.) and prepared with electrodes and intracerebral guide cannulas as previously described (Coutinho et al. 1998). Briefly, electrodes for electromyographic (EMG) recordings (Teflon-coated stainless steel wire; Cooner Sales, Chatworth, CA) were stitched into the external oblique musculature immediately superior to the inguinal ligament. The electrode leads were tunneled s.c. and exteriorized at the back of the neck, where they were secured with a suture. All surgical wounds were subsequently closed with silk sutures. Animals were then mounted in a stereotaxic apparatus (Kopf) and implanted with chronic intracerebral guide cannulas on the midline 3 mm dorsal to the nucleus raphe magnus (RMg, ventromedial RVM). The final coordinates for placement of cannulas relative to the interaural line were −2.0 mm (rostral-caudal), 0 mm (medial-lateral), and −6.5 mm (dorsal-ventral) according to the atlas of Paxinos and Watson (1986). The stainless steel guide cannulas (26-gauge needle shaft) had a length of 17.5 mm and were kept in place with acrylic dental cement secured by two stainless steel screws implanted in the skull. Each cannula was fitted with a 33-gauge stainless steel stylet to maintain patency. After these procedures, animals were individually housed and were given 5 to 7 days to recuperate before behavioral nociceptive testing.
Behavioral Nociceptive Testing.
Colorectal distension (CRD) was used as a noxious visceral stimulus in all experiments, and the visceromotor response (VMR), a contraction of the abdominal and hindlimb musculature, was measured as an index of visceral nociception (Ness and Gebhart, 1988). On the day of testing, a 6- to 8-cm-long lubricated, flexible latex balloon was inserted intra-anally into the descending colon and rectum of awake rats and secured by taping the balloon catheter (Tygon tubing) to the proximal end of the tail. Phasic CRD (20–80 mm Hg, 20 s) was produced by pressure-controlled air inflation of the balloon via opening of a solenoid gate to a constant-pressure air reservoir. The pressure in the reservoir was monitored and altered with a pressure control device (Bioengineering Department, University of Iowa, Iowa City, IA). The VMR to CRD was measured by recording EMG activity in the external oblique musculature. EMG activity was amplified, filtered, displayed on an oscilloscope, and quantified by recording activity that exceeded a preset voltage threshold (defined as the voltage that basal EMG activity did not exceed) as previously described (Coutinho et al., 1998). For each CRD trial, EMG activity was quantified 20 s before distension (baseline), during the 20-s distension period, and 20 s after distension. The increase in EMG activity during distension over the baseline was designated the visceromotor response. CRD (80 mm Hg) was given at 3-min intervals both before (control response) and after intra-RVM drug injection.
Intracerebral Drug Injection.
Intra-RVM microinjections were performed in awake animals by lowering a 33-gauge injection needle that extended 3 mm beyond the tip of the guide cannula and delivering 0.5 μl of drug over 30 s into the medial RVM (RMg). The injection needle was connected to a Hamilton Co., 10-μl syringe by polyethylene tubing (PE10), and an air bubble was maintained in the tubing to monitor the flow of the drug solution. Each animal received a single injection of one dose of neurotensin or SR48692. At the conclusion of the experiment, correct placement of intracerebral guide cannulas was verified for each animal by injecting methylene blue, removing the brain, and placing the brain in 10% formaldehyde solution (Formalin) for 24 h followed by sucrose for an additional 24 to 48 h. Cryostat sections (40 μm) were subsequently examined for location of injection sites.
Motor Function.
In a separate group of animals, effects of intra-RVM neurotensin injection on motor function were determined with an inclined plane test (Murphy et al., 1995). Motor function was assessed before and after intra-RVM neurotensin injection (at the time of maximal effect on the VMR to CRD) by placing the animal on the upper edge of a rectangular Plexiglas plane (60 × 100 cm) inclined at a fixed angle (25–45°) approximately 10 cm from the top. The ability to remain near the top without freely sliding within 5 s was scored a success, and animals were tested for four trials at each angle.
Data Analysis and Statistics.
All experimental groups consisted of five to eight animals. For each animal, the VMR to noxious CRD (80 mm Hg) was measured, and the mean of three stable responses was designated as the control (predrug) response. The response was considered stable if the intratrial variation was less than 10%. After intra-RVM drug injection, VMR was expressed as the percentage of control VMR (% control). For dose-response functions, the overall effect was represented as the area under the curve (AUC). The AUC was obtained from time-response functions by calculating the change in postdrug VMR from control VMR plotted against time with the trapezoidal rule. In some cases, stimulus-response functions were generated for graded CRD (20–80 mm Hg). Effects of intra-RVM neurotensin on stimulus-response functions were determined by generating a stimulus-response function during the period the neurotensin effect was determined to be stable and maximal. These data were normalized to a percentage of the control (predrug) response to the greatest intensity of CRD (80 mm Hg).
Statistical analysis of dose-dependent drug effects on the VMR to CRD was performed with one-way ANOVA with Fisher’s test for post hoc comparisons. Comparisons between two groups were performed with at test. The slopes of stimulus-response functions were determined by linear regression with the use of a curve-fitting computer program (Graphpad Prism). P < .05 was considered statistically significant in all tests.
Drugs.
The drugs used in these experiments were neurotensin (Sigma Chemical Co., St. Louis, MO) and SR48692 (a gift from Sanofi Recherche, Toulouse, France). Neurotensin and SR48692 were dissolved in sterile saline (0.9%) and dimethyl sulfoxide (100%), respectively.
Results
Microinjection Sites for Neurotensin and SR48692.
All doses of neurotensin and SR48692 were microinjected into the ventromedial RVM (RMg). Microinjection of neurotensin or SR48692 into sites located beyond the border of the RVM were without effect, and these data were excluded from the experimental groups (Fig.1). Although not intentionally studied, the absence of drug effect when injected outside of the RVM served to establish specificity of site of action.
Effects of Neurotensin Microinjected into RVM on VMR to Noxious CRD.
Similar to effects on the nociceptive tail-flick reflex (Urban and Smith, 1993), intra-RVM injection of neurotensin into awake rats produced a dose-dependent, biphasic effect on the VMR to noxious, phasic CRD (80 mm Hg, 20 s). Intra-RVM injection of neurotensin at lesser doses (0.03–0.30 pmol) produced an enhancement of the VMR to noxious CRD that was apparent and maximal 6 min after injection and had a relatively short duration of 18 to 27 min (Fig.2, A and C). In contrast, intra-RVM injection of greater doses (30–3000 pmol) of neurotensin inhibited the VMR to noxious CRD. This inhibitory effect of neurotensin was apparent 3 min after injection, was maximal after 15 min, and had a relatively long duration of 30 to 120 min (Fig. 2, A and C). These facilitatory and inhibitory effects of intra-RVM neurotensin resulted in a biphasic dose-response function (Fig. 2B). No observable change in animal behavior was noted after intra-RVM injection of either facilitatory or inhibitory doses of neurotensin. Additionally, intra-RVM neurotensin injection had no effect on baseline EMG activity before noxious CRD.
Effects of Neurotensin Microinjected into RVM on VMR to Graded CRD.
As previously reported (Coutinho et al., 1996), graded CRD (20–80 mm Hg, 20 s) resulted in reliable, reproducible, stimulus-response functions for the VMR (Fig.3, predrug). The effects of inhibitory doses (300–3000 pmol) of neurotensin injected into the RVM on stimulus-response functions were determined during the time the neurotensin effect was maximal and stable (12–24 min). Intra-RVM injection of the greatest dose (3000 pmol) of neurotensin resulted in nearly complete inhibition of the VMR to CRD at all stimulation intensities and significantly reduced the slope of the stimulus-response function to graded CRD (from 1.6 ± 0.2 to 0.07 ± 0.04%/mm Hg; Fig. 3). Intra-RVM injection of a lesser dose (300 pmol) of neurotensin resulted in partial inhibition of the VMR to CRD at all stimulation intensities and significantly reduced the slope of the stimulus-response function to graded CRD (from 1.6 ± 0.2 to 0.79 ± 0.11%/mm Hg; Fig. 3).
The effect of intra-RVM injection of a facilitatory dose of neurotensin (0.03 pmol) on the stimulus-response function to graded CRD was also determined while this effect was maximal and stable (6–15 min). Intra-RVM injection of this facilitatory dose of neurotensin produced an enhanced VMR response to CRD at all stimulation intensities, which resulted in a 2.5-fold leftward shift of the stimulus-response function to graded CRD but did not affect the slope (1.6 ± 0.2 and 1.7 ± 0.24%/mm Hg for control and neurotensin, respectively; Fig. 3).
Effects of SR48692 Microinjected into RVM on VMR to Noxious CRD.
To confirm that the effects of neurotensin on the VMR to CRD are neurotensin-receptor mediated, the effects of a nonpeptide neurotensin-receptor antagonist, SR48692, on neurotensin responses were to be determined after sequential intra-RVM injection. This was not possible to resolve, however, because intra-RVM injection of SR48692 alone was found to have a biphasic effect on the VMR to noxious CRD (80 mm Hg). Intra-RVM injection of SR48692 at greater doses (0.3–300 fmol) produced a dose-dependent inhibition of the VMR to CRD that was maximal 3 min after injection and had a relatively short duration of 15 min (Fig. 4A). In contrast, injection of a lesser dose (0.03 fmol) of SR48692 produced short-lived enhancement of the VMR to CRD. Thus, similar to neurotensin, intra-RVM injection of SR48692 resulted in a biphasic dose-response function for the VMR to noxious CRD (Fig. 4B). Intra-RVM SR48692 injection had no effect on baseline EMG activity before noxious CRD.
Effects of Neurotensin Microinjected into RVM on Motor Function.
To support the notion that intra-RVM neurotensin is modulating nociceptive visceral input and not affecting general motor function, the effect of intra-RVM injection of the greatest inhibitory dose (3000 pmol) of neurotensin on motor function was assessed in a separate group of animals with an inclined plane test. The effects of intra-RVM neurotensin injection on inclined plane performance were determined at the time of maximal inhibition of the VMR to CRD (12–24 min). Intra-RVM injection of neurotensin (3000 pmol) produced no observable change in general motor function and did not affect the animals performance on an inclined plane of varying angles (25–45°) compared with the predrug performance (Fig.5).
Discussion
The results from this study demonstrate biphasic modulation of visceral nociception by neurotensin in the RVM. RVM has been identified as an important area that modulates spinal nociceptive transmission via projections that descend through the spinal funiculi and terminate in both the medullary and spinal cord dorsal horn (Fields et al., 1991; for review, see Fields and Basbaum, 1993). Electrical stimulation or injection of various receptor-selective agonists into the RVM has been shown to both inhibit and facilitate spinal behavioral and dorsal horn neuron responses to noxious stimulation (Haber et al., 1980; Zhuo and Gebhart, 1992, 1997; Thomas et al., 1995). Activation of inhibitory and facilitatory influences from the RVM is intensity/dose-dependent, and these dual influences appear to be mediated by anatomically distinct, independent descending systems that involve different spinal components (Zhuo and Gebhart, 1990–1992, 1997).
Neurotensin has been shown to have an antinociceptive action in models of cutaneous nociception after injection into the CNS or various brain sites (Kalivas et al., 1982; Behbehani and Pert, 1984; Fang et al., 1987). Additionally, neurotensin injection into the RVM has been shown to dose-dependently facilitate and inhibit nociceptive tail-flick and hot-plate responses (Urban and Smith, 1993). Biphasic modulation of cutaneous spinal nociceptive transmission was supported in a subsequent study demonstrating that intra-RVM injection of neurotensin facilitates and inhibits spinal dorsal horn neuron responses to noxious thermal stimulation (Urban and Gebhart, 1997). The data from this study expand on these results by demonstrating a similar biphasic effect of intra-RVM neurotensin on visceral nociceptive transmission. Although most studies investigating effects of neurotensin on nociception have focused on cutaneous nociception, an antinociceptive action of neurotensin on visceral nociception has been reported after intracisternal injection and measurement of acetic acid-induced writhing (Clineschmidt and McGuffin, 1977). In our study, neurotensin injection into RVM was found to dose-dependently facilitate and inhibit the VMR to noxious CRD at lesser and greater doses, respectively. Modulation of the VMR by neurotensin is probably a specific effect on visceral nociceptive input, because intra-RVM neurotensin produced no observable change in general motor function and did not affect performance on an inclined plane. Additionally, these biphasic effects of neurotensin are probably localized in the ventromedial RVM (RMg), because previous studies with dyes and radiolabeled substances have reported intracerebral microinjections of 0.5 μl to diffuse approximately 0.5 mm from the injection site (Myers, 1966; Myers and Hoch, 1978). The localization of neurotensin effects is further supported in this study by the observation that microinjections extending 1 to 2 mm beyond the border of the RVM were without effect.
Interestingly, the dose range and time course of effects of intra-RVM neurotensin on biphasic modulation of visceral nociception appears to be similar to that found for modulation of cutaneous nociception. Lesser doses (fmol) of neurotensin produced a relatively short-lived facilitation of the VMR to CRD in the current study (18–27 min), as well as a similar short-lived facilitation of tail-flick and dorsal horn nociceptive neuron responses (Urban and Smith, 1993; Urban and Gebhart, 1997). Similarly, greater doses of neurotensin (pmol) produced a longer lasting inhibition of the VMR to CRD and longer lasting inhibition of tail-flick and dorsal horn nociceptive neuron responses. Descending modulation of spinal visceral nociceptive transmission has been reported after stimulation of various supraspinal sites, including the RVM (Giesler and Liebeskind, 1976; Cervero et al., 1985; Ness and Gebhart, 1987). The results from this study suggest that, similar to cutaneous nociception, intra-RVM neurotensin activates descending facilitatory and inhibitory influences on visceral nociceptive transmisson.
The biphasic effects of intra-RVM neurotensin on visceral nociceptive transmission are probably the result of activation of distinct, independent descending pain-modulatory systems. For example, descending facilitation produced by intra-RVM neurotensin has been shown to be localized within a large area of the RVM, involves descending projections in the ventrolateral funiculi, and is mediated by spinal CCKB receptors. In contrast, descending inhibitory influences of neurotensin are localized within the medial RVM (RMg), involve descending projections in the dorsolateral funiculi, and are probably mediated by a spinal noradrenergic component (Behbehani, 1992; Urban and Smith, 1994; Urban et al., 1996b; Urban and Gebhart, 1997). Although the aforementioned studies examined neurotensin effects on cutaneous nociception, the similarities between effects on cutaneous and visceral nociception suggest a similar activation of independent, descending facilitatory and inhibitory systems.
In this study, facilitatory and inhibitory effects of intra-RVM neurotensin were additionally discriminated by differential effects on intensity coding to graded CRD. Whereas inhibitory doses of intra-RVM neurotensin produced a dose-dependent reduction in the slope of the stimulus-response function to graded CRD, a facilitatory dose did not affect the slope and produced a 2.5-fold leftward shift of the function. These results demonstrate significant differences in the modulation of intensity coding of noxious visceral stimulation by the facilitatory and inhibitory actions of intra-RVM neurotensin and support the notion that independent systems mediate these effects.
Activation of independent descending facilitatory and inhibitory systems by neurotensin is probably the result of an interaction of multiple neurotensin-receptor subtypes in the RVM. Neurotensin-binding sites have been identified throughout the CNS, including the RVM, and both high- and low-affinity (levocabastine-sensitive) neurotensin receptors have been cloned (Tanaka et al., 1990; Chalon et al., 1996). A nonpeptide neurotensin-receptor antagonist, SR48692 (Gully et al., 1993), has also been used to discriminate neurotensin-receptor subtypes in various behavioral and electrophysiological studies (Dubuc et al., 1994; Steinberg et al., 1994). Based on the dose-dependent nature of neurotensin to activate descending facilitatory and inhibitory systems from the RVM, Smith et al. (1997) used SR48692 to discriminate neurotensin-receptor subtypes in the RVM involved in mediating these biphasic effects. The results from that study demonstrated that 1) descending facilitation involves activation of a high-affinity neurotensin receptor that has relatively low affinity for SR48692 (pmol), 2) descending inhibition involves activation of a low-affinity neurotensin receptor that has relatively high affinity for SR48692 (fmol), and 3) the low-affinity neurotensin receptor is not the levocabastine-sensitive, cloned neurotensin receptor.
In this study, intra-RVM injection of SR48692 produced a biphasic effect on the VMR to noxious CRD, suggesting a role for endogenous neurotensin in the RVM in modulating visceral nociception. Greater doses of SR48692 (low pmol) resulted in dose-dependent inhibition of the VMR to noxious CRD, whereas a lesser dose (low fmol) enhanced the response. These data suggest that endogenous neurotensin in the RVM modulates the VMR to CRD via concurrent activation of a dominant descending facilitatory influence and masked inhibitory influence revealed by greater and lesser doses of SR48692, respectively. This is consistent with a more prominent interaction of endogenous neurotensin in the RVM with the high-affinity neurotensin receptor involved in descending facilitatory influences and a relatively reduced but apparent interaction of endogenous neurotensin with the low-affinity receptor-mediating inhibition. The doses of SR48692 that revealed the facilitatory and inhibitory effects of endogenous neurotensin are consistent with the dose range of SR48692 previously found to discriminate neurotensin-receptor subtypes in the RVM involved in these biphasic effects (Smith et al., 1997; see previous paragraph). Although the facilitatory effect of neurotensin after intra-RVM injection appears to be less prominent than the inhibitory effect in this study, as well as others (Urban and Smith, 1993; Urban and Gebhart, 1997), the results obtained with SR48692 are consistent with previous reports demonstrating that facilitation of nociception is the principal physiologically relevant action of endogenous neurotensin in the RVM (Urban and Smith, 1993; Urban et al., 1996a; Smith et al., 1997). The notion that noxious, phasic CRD may concurrently activate prominent descending facilitatory and masked inhibitory influences is supported by several studies that have demonstrated concurrent activation of descending dual influences after acute or persistent nociceptive stimulation (Morgan et al., 1994; Wiertelak et al., 1994; Coutinho et al., 1998). The results from this study support a similar role for endogenous neurotensin in RVM in the modulation of VMR after acute, noxious visceral stimulation.
In summary, the results from this study demonstrate that lesser and greater doses of neurotensin injected into RVM facilitate and inhibit the VMR to noxious CRD. This biphasic effect of intra-RVM neurotensin on visceral nociception is consistent with a previously established role for this peptide in biphasic modulation of spinal cutaneous nociception. These biphasic effects probably result from an interaction with neurotensin-receptor subtypes in the RVM that activate independent descending inhibitory and facilitatory systems. Additionally, endogenous neurotensin in the RVM appears to modulate the VMR to noxious CRD via activation of prominent facilitatory and masked inhibitory influences. These results support a role for this neuropeptide in the modulation of both cutaneous and visceral nociception.
Acknowledgments
We thank Michael Burcham for preparation of the graphics.
Footnotes
-
Send reprint requests to: Mark O. Urban, Ph.D., Department of Pharmacology, Bowen Science Building, University of Iowa, Iowa City, IA 52242. E-mail: murban{at}blue.weeg.uiowa.edu
-
↵1 The study was supported by National Institutes of Health Grants DA-11431 (M.O.U.) and DA-02879 (G.F.G.).
-
↵2 Present address: UCLA/CURE Neuroenteric Disease Program, WLA VA Medical Center, Building 115, Room 223, 11301 Wilshire Boulevard, Los Angeles, CA 90073.
- Abbreviations:
- CNS
- central nervous system
- CRD
- colorectal distension
- RMg
- nucleus raphe magnus
- RVM
- rostral ventromedial medulla
- VMR
- visceromotor response
- Received December 28, 1998.
- Accepted March 2, 1999.
- The American Society for Pharmacology and Experimental Therapeutics