INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

After Reynolds (1969) reported that electrical stimulation in the rat midbrain periaqueductal gray (PAG) was analgesic, interest rapidly focused on study of endogenous pain inhibitory systems. The anatomy, physiology, and pharmacology of descending inhibition of nociceptive behaviors (e.g., Basbaum et al. 1977; Lewis and Gebhart 1977; Mayer et al. 1971; Oliveras et al. 1974) and spinal nociceptive neurons (e.g., Fields et al. 1977; Liebeskind et al. 1973; see Gebhart 1986 for review) were widely studied, and midbrain stimulation in humans was documented to be analgesic (Adams 1976; Hosobuchi et al. 1977; Richardson and Akil 1977a,b). Our current understanding is that the caudal, ventrolateral PAG is considered a nodal point in an endogenous, descending system of pain modulation that involves sites both rostral to the PAG (e.g., hypothalamus, Aimone et al. 1988; Lumb and Cervero 1989; Workman and Lumb 1997; cortex, Calejesan et al. 2000; Kuroda et al. 2001; Zhang et al. 1997; amygdala, Tershner and Helmstetter 2000) and caudal to the PAG, principally the rostroventral medial medulla (RVM) (for reviews, see Basbaum and Fields1984; Fields and Basbaum 1999; Gebhart and Randich 1990; Mason 2001; Sandkühler 1996).

Tonic descending inhibition had long been of interest to neurophysiologists, who focused principally on modulation of flexor reflexes (e.g., Fulton 1926; Sherrington and Sowton 1915), including those activated by stimulation of flexor reflex afferents (see Willis 1982; 1988 for reviews). The intensity of peripheral stimulation in these later studies likely activated nociceptors and thus anticipated more recent work on descending inhibition of nociceptive reflexes/neurons evoked by more natural stimuli (e.g., thermal stimulation) (see Gebhart 1986; Sandkühler 1996 for overviews). By far, the bulk of experimentation to date has employed noxious cutaneous stimuli or electrical activation of somatic nerve trunks.

Relatively few studies have examined descending modulatory influences on visceroceptive spinal neurons. Some studies have addressed tonic influences (e.g., Akeyson et al. 1990; Standish et al. 1992; Tattersall et al. 1986) and others stimulation-produced inhibition of spinal visceral transmission (Ammons et al. 1984, 1986; Chandler et al. 1989; Chapman et al. 1985; Gielser and Liebeskind 1976; Knuepfer and Holt 1991; Lumb 1990; Ness and Gebhart 1987; Okada et al. 1999). In previous work, we (Zhuo and Gebhart 1992, 1997) characterized descending influences from the RVM on noxious cutaneous stimuli, documenting the presence of facilitatory and inhibitory modulation of nociceptive transmission. The present study, undertaken to investigate whether spinal visceral nociceptive transmission was similarly modulated from the RVM, examined electrical- and glutamate-produced effects on spinal neuron responses to noxious colorectal distension in the rat.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surgical procedures

Twenty-nine adult, male Sprague-Dawley rats (Harlan, Indianapolis, IN) weighing 330-460 g were used. Rats were anesthetized with an intraperitoneal injection of pentobarbital sodium (45 mg/kg; Nembutal, Abbott Laboratories, Abbott Park, IL), and catheters were inserted into the trachea for mechanical ventilation, a femoral vein for administration of drugs, and a femoral artery for measurement of blood pressure and heart rate. All wound margins were covered with a local anesthetic ointment and closed with silk sutures. The Institutional Animal Care and Use Committee at The University of Iowa approved all experimental protocols.

The lumbosacral spinal cord was exposed by laminectomy, and rats were suspended by vertebral clamps rostral and caudal to the laminectomy. The skin was reflected laterally and tied to the frame to make a pool for agar (1.75% in saline) to minimize movement of the spinal cord. The head was fixed in a stereotaxic head holder. Arterial blood pressure and heart rate were monitored throughout experiments, and body temperature was maintained at 36 ± 0.5°C with a hot water heating pad and overhead lamps.

During the recording session, rats were paralyzed with pancuronium bromide (0.4 mg iv initially and 0.2 mg/h thereafter) and artificially ventilated. Anesthesia was maintained by inhalation of a nitrous oxide and oxygen (2:1) mixture and continuous intravenous infusion of pentobarbital sodium (5-10 mg · kg¹ · h¹). Tungsten microelectrodes (Micro Probe, Clarksburg, MD, 0.8-0.95 M) were used for extracellular recording of single units in the L₆-S₁ spinal segments, midline to 0.5 mm lateral, 0.2-1.4 mm from spinal cord dorsum.

Noxious stimulation

The visceral stimulus was colorectal distension (CRD) with air of a flaccid latex balloon, 7-8 cm long, inserted intra-anally into the descending colon and rectum. Methods for balloon fabrication and controlled, constant pressure distension of the colon have been fully described (Gebhart and Sengupta 1996). Briefly, CRD was initiated by opening a solenoid gate to a constant pressure air reservoir; constant pressure stimuli at 20, 40, 60, 80, or 100 mmHg, 20 s duration were given at 4-min intervals. Brief (5 s) 80-mmHg distending stimuli were used to search for units, which were also tested for responses to nonnoxious stimuli (touch, brush, light pressure, and joint movement) and noxious stimuli (pinch with pair of forceps) to plot the convergent somatic receptive fields on the hind part of the body. Unit activity was continuously monitored by an analog delay circuit and storage oscilloscope, discriminated and counted on-line.

Brain stem stimulation and glutamate microinjection

Focal electrical brain stimulation, 5-100 µA, consisted of continuous 100-Hz constant-current cathodal pulses of 100 µs duration. Brain stimulation was started 5 s before and continued during the 20 s of colonic distension. This stimulation paradigm was the same as previously used in behavioral and electrophysiological studies (Zhuo and Gebhart 1990a, 1992, 1997). Monopolar stimulating electrodes (34-gauge, 0.15 mm OD), guided stereotaxically in the vertical plane (incisor bar at +3.3 mm) (Paxinos and Watson 1982), were inserted into the brain through a 26-gauge (0.45 mm OD) guide cannula. The electrodes were cut to extend 2 mm beyond the tip of the guide cannula. The stimulating electrode was lowered to a site in the RVM, and the effect of stimulation on spinal nociceptive transmission was tested at two to four sites in an experiment, 0.5 mm apart, in the same electrode track. Testing the effect of electrical stimulation always preceded testing of glutamate. The effects of electrical stimulation did not outlast the duration of stimulation by more than 3 min; the effects of glutamate were tested 15 min after electrical stimulation.

Monosodium-L-glutamate (5 or 50 nmoles, pH 6.7) was injected into the medulla in a volume of 0.5 µl via an injection cannula (33-gauge, 0.20 mm OD) inserted through the same 26-gauge guide cannula and extending 2 mm beyond its tip. Injection of glutamate was done by an electrically driven syringe pump at a speed of 0.5 µl/1.5 min. The progress of the microinjection was continuously monitored by following the movement of an air bubble in a length of calibrated tubing between the injection syringe and the injection cannula.

Histology

Anodal electrolytic lesions were made in the brain stem and spinal cord to mark the sites of stimulation and recording, respectively. At the end of experiments, rats were killed with an intravenous overdose of pentobarbital sodium. The brain and appropriate segments of spinal cord were removed and fixed in 10% Formalin, frozen and cut in 40-µm coronal sections, mounted on glass slides, and stained with hematoxylin-eosin for histological examination of lesion sites.

Data analysis and statistics

The resting activity of a neuron was counted for 10 s before the onset of CRD, and response to distension is defined as the increase in activity during distension above resting activity. Responses of neurons are presented as total number of impulses/20 s or normalized to a percentage of the response to 100 mmHg distension. Responses to three distensions (100 mmHg; 20 s), 4 min apart, were averaged to define the control response. Stimulus-response functions (SRF) were plotted for individual units and a least-squares regression was obtained from the linear part of the SRF. The regression line was extrapolated to the ordinate to estimate response threshold (in mmHg). Because electrical stimulation was started 5 s before the onset of CRD, the effects of stimulation on resting activity was assessed during these 5 s and compared with resting activity in the immediately preceding 5 s before the onset of electrical stimulation. Data are presented as means ± SE. Statistical comparisons were made using either one-way or two-way ANOVAs (Newman-Keuls test for post hoc comparisons). Student's t-test was applied for comparisons between paired groups. In all cases, P < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Unit sample

A total of 44 spinal units in the L₆ and S₁ spinal segments that responded to CRD (80 mmHg; 20 s) were studied. The neurons were divided into four groups based on their responses to CRD (Ness and Gebhart 1987): short-latency abrupt neurons (SLA; n = 22), which respond coincident with the onset of CRD, slowly adapt during CRD and cease responding within 1-2 s of termination of CRD; short-latency sustained neurons (SLS; n = 10), which also respond coincident with the onset of CRD but show no adaptation during CRD and exhibit a sustained discharge (180 s) after termination of CRD; long-latency neurons (LL; n = 3), which respond at a relatively long (6-11 s) latency after onset of CRD; and inhibited neurons (INHIB; n = 9), which are inhibited by CRD (Fig. 1). Forty-two units (95%) had convergent somatic receptive fields restricted to the caudal part of the body. Two (5%) SLA units responded only to CRD, and a convergent somatic receptive field could not be localized by noxious stimulation of hindlimbs, perianal skin, tail, or testis. Among the 42 units for which convergent receptive fields were found, the receptive fields of 30 units were localized to the caudal dorsal and ventral parts of the body, including hindlimbs, skin, and testis; the remaining 12 units responded to noxious pinch of the tail. The recording sites of the neurons were 0.42-1.31 mm (1.01 ± 0.03 mm, mean ± SE) below the cord dorsum. There were no differences in the depths of recording sites for the different groups of spinal units (i.e., SLS, SLA, LL, or INHIB; F(3,40) = 2.05. The recording sites of 22/44 neurons (SLA = 11; SLS = 6; LL = 1, and INHIB = 4) were histologically recovered (Fig. 1). The effects of electrical stimulation or L-glutamate microinjection into 83 sites in the RVM on spontaneous activity and responses to CRD (80 mmHg, 20 s) are described below. Typically, two sites of stimulation in the RVM were tested on one spinal neuron in each experiment. Electrical stimulation was tested before glutamate was tested; glutamate was tested at only one site in an experiment, typically the most ventral of the stimulation sites in an electrode penetration.

View larger version (16K): [in this window] [in a new window]   Fig. 1. Responses of spinal dorsal horn neurons to noxious colorectal distension (CRD). A: peristimulus time histograms (1-s binwidth) illustrating the 4 different types of responses to 80-mmHg (20 s) CRD. Short-latency abrupt (SLA) neurons respond at the onset of CRD and stop responding within 1 or 2 s after termination of CRD. Short latency sustained (SLS) neurons similarly respond at short latency with the onset of CRD, but continue responding after termination of CRD. Long-latency (LL) neurons respond at long latency following the onset of CRD and typically continue responding after termination of CRD. Inhibitory (INHIB) neurons generally have greater spontaneous activity than other types of neurons that respond to CRD and are inhibited during CRD. B: histologically confirmed sites of recording in the L6 and S1 spinal cord.

Modulation of spinal neuron responses by RVM electrical stimulation

A total of 83 sites in the brain stem were studied, and electrical stimulation (5-100 µA) produced biphasic, only inhibitory, or only facilitatory modulation of responses of spinal neurons to noxious CRD (Table 1 and Fig. 2).

View larger version (31K): [in this window] [in a new window]   Fig. 2. Summary of sites in the rostroventral medulla where electrical stimulation produced biphasic modulation (), facilitatory modulation (), inhibitory modulation (), or no effect () on spinal visceral nociceptive transmission. NGC, nucleus reticularis gigantocellularis; NGC, nucleus reticularis gigantocellularis pars alpha; NPGCl, nucleus reticularis paragigantocellularis lateralis; NRM, nucleus raphe magnus; NRO, nucleus raphe obscuris; NRP, nucleus raphe pallidus; DPGi, nucleus dorsal paragigantocellularis; Pyr, pyramidal tract; Sp5, spinal trigeminal tract; VII, facial nucleus (Paxinos and Watson 1986).

At 25 sites in the brain stem (, Fig. 2), electrical stimulation produced biphasic modulatory effects on neuron responses to CRD. At 38 sites in the brain stem (, Fig. 2), electrical stimulation produced only intensity-dependent inhibitory effects on responses of spinal units to noxious CRD. At 13 sites (, Fig. 2), electrical stimulation only facilitated, and at 7 sites (, Fig. 2) electrical stimulation did not affect responses of spinal units to noxious distension.

BIPHASIC MODULATION. Electrical stimulation in the nucleus reticularis gigantocellularis (NGC; n = 13), NGC (n = 7), nucleus raphe magnus (NRM; n = 4), or nucleus raphe pallidus (NRP; n = 1) produced biphasic effects. At lesser intensities of stimulation (5-25 µA), responses of SLA (n = 15), SLS (n = 4), and LL (n = 1) neurons to distension were significantly facilitated; at greater intensities of stimulation (50-100 µA), responses of the same neurons to the same intensity of distension were inhibited. Figure 3 shows an example of RVM stimulation-produced biphasic effects on responses of a SLA unit to CRD (80 mmHg; 20 s). In this example, the response to distension was increased to 112.4% of control at 10 µA RVM stimulation (from a control 307 to 345 total impulses/20 s) and decreased to 43.4% of control (133 total impulses/20 s) at a greater intensity of stimulation (100 µA). Because neurons excited by CRD (SLA, SLS, and LL) were affected similarly by stimulation, the data are grouped and presented together in Fig. 4A. At these 20 sites of stimulation, 10 µA stimulation produced a mean 120.8 ± 2.6% facilitation (P < 0.001) of unit responses to distension; inhibition of responses of the same units to 50% of control was produced at a mean 95.2 ± 15.2 µA of stimulation (Fig. 4A). The mean recruitment index for inhibition (% inhibition/20 µA increase in stimulation intensity) for these 20 sites was 30.0 ± 4.4. 

View larger version (23K): [in this window] [in a new window]   Fig. 3. Example of biphasic modulation of spinal visceral nociceptive transmission produced by electrical stimulation in nucleus gigantocellularis (NGC). A: peristimulus time histograms (1-s binwidth) and corresponding ocillographic records for a short-latency abrupt spinal unit. The control response to noxious colorectal distension (CRD; 80 mmHg, 20 s, illustrated by the filled horizontal bar) is shown at the top. The effects of different intensities of stimulation in NGC are shown below; the period of stimulation (25 s) is indicated by the arrows. B: graphic representation of the data in A. The point above zero represents the response to CRD in the absence of stimulation. C: stimulation site in NGC and recording site in the L6 spinal cord for the data presented in A.

View larger version (19K): [in this window] [in a new window]   Fig. 4. Summary of descending modulation from the rostral ventral medulla. Biphasic, inhibitory, and facilitatory modulation of responses to noxious colorectal distension (A) and spontaneous activity of the same spinal neurons (B).

INHIBITORY MODULATION. Thirty-three of the subtotal 38 sites where stimulation produced only inhibitory effects were located in the NGC (n = 16), NGC (n = 8), NRM (n = 2), NRP (n = 3), or the nucleus raphe obscuris (NRO; n = 4). The other five sites were located in the medial longitudinal fasciculus (n = 1), predorsal bundle (n = 2), the nucleus lateralis paragigantocellularis (n = 1), and dorsal paragigantocellularis (n = 1). Electrical stimulation at the intensities tested (10-100 µA) only inhibited responses to noxious colonic distension. The spinal neurons in these experiments were SLS (n = 8), SLA (n = 19), and LL (n = 3); inhibitory modulation of three INHIB neurons was not studied. Data from 24 sites of stimulation in the NGC (n = 16) and NGC (n = 8) are summarized in Fig. 4A. The estimated threshold of stimulation for inhibition was 4.7 ± 1.5 µA (n = 30), which is significantly less than the estimated threshold of stimulation for inhibition at "biphasic" sites in the RVM (25.3 ± 2.8 µA). The mean recruitment index for inhibition from these 30 inhibitory sites was 14.2 ± 2.0% inhibition/20 µA, which is significantly less than the mean recruitment index for inhibition from biphasic sites of stimulation (30.0 ± 4.4% inhibition/20 µA).

FACILITATORY MODULATION. Twelve of the subtotal 13 sites where only facilitation of responses to CRD were produced were located in the NGC (n = 6), NGC (n = 1), NRM (n = 2), or NRP (n = 3); one site was located in the nucleus dorsal paragigantocellularis (DPGi). In these 13 experiments, spinal units were SLA (n = 3), SLS (n = 4), LL (n = 1), and INHIB (n = 5). Electrical stimulation at 10 µA significantly facilitated responses of spinal units to distension to a mean 143.4 ± 15.7% of control. At a greater intensity of 100 µA, responses of spinal units were similarly facilitated to a mean 152.8 ± 18.8% of the control response to noxious distension (P < 0.01). The magnitude of the facilitatory effect was not dependent on the intensity of electrical stimulation [F(3,33) = 0.72]. The data from seven sites in the NGC and NGC are summarized in Fig. 4A.

SPONTANEOUS ACTIVITY. At the 20 sites in RVM where stimulation produced biphasic facilitatory and inhibitory effects, electrical stimulation did not affect spontaneous activity of spinal units [baseline 8 ± 3 impulses/s; F(4,100) = 0.11]. Similarly, at 33 inhibitory sites in the RVM, electrical stimulation produced intensity-dependent inhibition of responses of spinal units to distension without affecting spontaneous activity (mean 5 ± 1 impulses/s) of the same spinal units [F(4,124) = 1.00]. Stimulation at facilitatory sites (baseline 16 ± 5 impulses/s) also did not significantly affect spontaneous activity [F(3,33) = 0.72], although stimulation at 100 µA decreased activity to a mean 5 ± 5 impulses/s. Data are summarized in Fig. 4B.

INTENSITY CODING. SRFs of spinal neurons to graded intensities of CRD and their modulation by electrical stimulation in the RVM were also investigated. An example of stimulation-produced inhibition (at 100 µA) of responses of a SLA unit to graded distension is shown in Fig. 5. SRFs for six spinal units (SLA and SLS, n = 3 each) and their modulation by a facilitatory intensity of stimulation in the RVM are presented in Fig. 6B. Electrical stimulation at a mean intensity of 16.7 ± 3.1 µA significantly facilitated responses of these units to 100 mmHg distension to a mean 116.5 ± 4.6% of control and produced, at noxious intensities of distension (40-100 mmHg), a parallel shift of the SRF to the left without changing its slope (6 ± 2 impulses/mmHg vs. 7 ± 2 impulses/mmHg).

View larger version (23K): [in this window] [in a new window]   Fig. 5. Example of the effect of inhibitory stimulation in nucleus raphe magnus (NRM) on responses of a spinal neuron to graded colorectal distension (CRD). A: peristimulus time histograms (1-s binwidth) for responses to 60, 80, and 100 mmHg CRD in the absence (control, left column) and presence (right column) of stimulation (100 µA) in NRM. The period of CRD (20 s) is illustrated by the horizontal bars and the period of stimulation in NRM is indicated by the arrows. B: graphic representation of the data in A. C: stimulation site in NRM and recording site in the S1 spinal cord.

View larger version (30K): [in this window] [in a new window]   Fig. 6. Summary of the effects of stimulation in the rostroventral medulla on responses to graded colorectal distension (CRD). A and C: stimulus-response functions of individual spinal neurons in the absence of stimulation in the brain stem. B and D: summary of effects of stimulation in the rostroventral medulla on the stimulus response functions of the units illustrated in A and C, respectively. Data are presented as mean responses in the absence () and the presence () of stimulation. In B, a facilitatory intensity of stimulation (16.7 ± 3.1 µA) at sites illustrated in E shifted the stimulus-response function leftward. In D, an inhibitory intensity of stimulation (62.5 ± 9.5 µA) at sites illustrated in E shifted the stimulus-response function rightward.

GLUTAMATE-PRODUCED EFFECTS. Glutamate at a low concentration (5 nmoles, 0.5 µl) was administered into six sites in the RVM and significantly facilitated responses of spinal units to 80-mmHg distension. The example in Fig. 7 shows glutamate-produced facilitation of responses to CRD. Both electrical stimulation (25 µA) and glutamate microinjection (5 nmoles) facilitated unit responses to 30-mmHg distension. The response to 30-mmHg distension was increased 1 min after administration of glutamate to 275.7% of control (from a control 214 to 590 total impulses/20 s). Data from six experiments are summarized in Fig. 8A. The effects of glutamate were rapid in onset and short lasting (Fig. 8A). At 1 min after administration of glutamate, the responses of spinal units (n = 6) to 80-mmHg distension were significantly increased to a mean 125.1 ± 10.5% of control (407 ± 104 total impulses/ 20 s). Responses to 80-mmHg distension returned to preglutamate baseline (mean, 99.4 ± 3.8% of control) by 10 min after administration of glutamate. Spontaneous activity (mean 3 ± 3 impulses/s) of the same spinal units was not affected by glutamate administration at the same dose [5 nmoles; F(5,45) = 0.18]. In four of these six experiments, the facilitatory effect of glutamate was tested by a second administration of the same dose of glutamate (5 nmoles) into the same sites after the response of the unit returned to control. The percentage facilitation produced by this second administration of glutamate was to 130.2 ± 4.2% of control (P < 0.05), which was not significantly different from the magnitude of facilitation produced by the first microinjection of glutamate (Fig. 8B).

View larger version (21K): [in this window] [in a new window]   Fig. 7. Example of facilitation of spinal visceral transmission produced by electrical stimulation and glutamate in the nucleus raphe magnus (NRM). A: peristimulus time histograms (1-s binwidth) and corresponding ocillographic records in the absence (top histograms) and presence (bottom histograms) of electrical stimulation (25 µA) and glutamate (5 nmoles) given in the same site in NRM. The intensity and duration of colorectal distension is illustrated below; the period of electrical stimulation (25 s) is indicated by the arrows. B: summary of the data illustrated in A and time course of effect of glutamate given in NRM. The point above C represents the response to 30-mmHg colorectal distension; the point above stim represents the response to the same intensity of distension during stimulation in NRM. C: site of stimulation and injection of glutamate.

View larger version (30K): [in this window] [in a new window]   Fig. 8. Summary of glutamate-produced facilitation and inhibition of spinal visceral transmission. A: glutamate microinjection at a low dose (5 nmoles) facilitated responses to colorectal distension (CRD) and a greater dose (50 nmoles) inhibited responses to CRD. B: 15 min after the 1st injection of glutamate, a 2nd injection of either 5 or 50 nmoles produced the same magnitude of facilitation and inhibition, respectively, of responses to CRD. C: sites of microinjection of 5 nmoles glutamate ( and ) and 50 nmoles glutamate ( and ). Abbreviations as in Fig. 2.

TONIC DESCENDING FACILITATORY INFLUENCES. Lidocaine microinjection into seven sites in the RVM (NGC, n = 1; NGC, n = 2; NRM, n = 3 and NRP, n = 1; Fig. 9) significantly decreased spontaneous activity of spinal units to 51.9 ± 20.3% of control at a mean 5.7 ± 0.6 min (range 4-7 min) after lidocaine administration. These neurons included two INHIB, three SLS, and two SLA units. Spontaneous activity recovered to preinjection baseline (mean 109.2 ± 31.5% of control) 30 min later. In two other experiments, ibotenic acid (10 µg, 0.5 µl) was microinjected into the NGC; spontaneous activity of spinal dorsal horn neurons was also decreased (to 53.7 and 63.9% of control, respectively). The effects lasted throughout the period of observation (1 h).

View larger version (20K): [in this window] [in a new window]   Fig. 9. Effects of lidocaine and ibotenic acid injected into the rostroventral medulla (B) on spontaneous activity of spinal neurons that responded to colorectal distension (A).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study documents that electrical and/or chemical (glutamate) stimulation in the RVM produces intensity/dose-related modulation of spinal visceral nociceptive transmission. At the same sites in RVM, electrical stimulation in the RVM facilitated responses of spinal dorsal horn neurons to noxious CRD at lesser intensities of stimulation and inhibited responses at greater intensities of stimulation. At other sites in RVM, only inhibitory or only facilitatory effects were produced by electrical stimulation. Activation of glutamatergic receptors in the RVM replicated the biphasic modulatory effects produced by electrical stimulation. Stimulation- and glutamate-produced effects were rapid in onset, short lasting, and reproducible, as in previous reports of RVM modulation of spinal cutaneous nociceptive transmission (Zhuo and Gebhart 1992, 1997).

Descending inhibition of spinal visceral nociception

It has been well documented that spinal sensory transmission arising from somatic structures is modulated by descending inhibitory influences from the RVM (see INTRODUCTION). Spinal visceral sensory transmission has been shown to be subject to similar descending inhibitory modulation (Ammons et al. 1984; Chapman et al. 1985; Ness and Gebhart 1987; Tattersall et al. 1986). Previous studies of supraspinal origins of descending inhibitory influences on spinal visceral transmission and/or visceral nociceptive reflexes mainly focused on the PAG and NRM (Chandler et al. 1989; Ness and Gebhart 1987; Ammons et al. 1984; Chapman et al. 1985; Giesler and Liebeskind 1976). The results of the present study demonstrate that activation of cell bodies in the NGC and NGC as well as midline raphe nuclei by electrical stimulation or glutamate microinjection produce an inhibitory modulation of responses of spinal dorsal horn neurons to noxious colonic distension, complementing previous findings.

Descending facilitation: an important modulatory mechanism

Earlier studies suggested (Fields et al. 1977; Haber et al. 1980; McCreery et al. 1979) and subsequent studies documented that spinal somatic nociceptive transmission is subject to descending biphasic modulation from the RVM. Electrical stimulation in the RVM produced facilitation and/or inhibition of responses of spinal units to noxious mechanical or thermal stimulation of the skin, effects that were reproduced by microinjection of glutamate into the same sites in RVM (Zhuo and Gebhart 1992, 1997). Similarly, a spinal nociceptive reflex (tail-flick reflex) could also be biphasically modulated by activation of cell bodies in the RVM (Zhuo and Gebhart 1990a). Importantly, receptor-selective agonists (e.g., neurotensin, baclofen) given into RVM also have been documented to dose-dependently facilitate and inhibit somatic nociceptive transmission (Thomas et al. 1995; Urban and Gebhart 1997; Urban and Smith 1994). Descending excitatory or facilitatory influences from supraspinal structures on spinal visceral nociceptive transmission also have been noted in previous reports (Akeyson et al. 1990; Cervero 1983; Tattersall et al. 1986). Tattersall et al. (1986) reported that responses to visceral stimulation of 32 of a total 75 viscerosomatic neurons recorded in the lower thoracic spinal cord were either reduced or completely abolished by reversible spinal cold block. Others (Akeyson et al. 1990; Euchner et al. 1993; Hummel et al. 1997) also reported that viscerosomatic spinal neurons recorded in the thoracic spinal cord showed biphasic (increase or decrease) changes during reversible spinalization. These findings suggest that spinal visceral transmission is subject to descending excitatory influences from supraspinal structures, although the origin of this descending excitation is unknown.

In the present study, electrical and/or chemical (glutamate) stimulation in the RVM produced biphasic, only inhibitory, or only facilitatory effects on spinal visceral nociceptive transmission; facilitation of responses to noxious CRD was produced at biphasic sites in RVM at lower intensities/doses and inhibition of responses at greater intensities/doses. At sites in RVM where only facilitation was produced, all intensities of stimulation tested facilitated responses of spinal neurons to CRD. This is the first report to systematically characterize descending facilitation of spinal visceral nociceptive transmission. In previous work, we (Zhuo and Gebhart 2002) found that electrical and/or glutamate stimulation or chemical stimulation (N-methyl-D-aspartate) (Coutinho et al. 1998) in the RVM similarly modulates a visceromotor response to noxious CRD. These and other findings reveal the presence of descending influences from the brain stem that are capable of enhancing input from the viscera and potentially contributing to the discomfort and pain characteristic of functional bowel disorders.

Comparison with modulation of spinal cutaneous nociceptive transmission

In previous studies, we found that activation of cell bodies in the NGC/NGC and NRM produced biphasic modulatory effects on spinal neuron responses to noxious thermal stimulation of the hind paw (Urban and Gebhart 1997; Zhuo and Gebhart 1992, 1997). Electrical stimulation in the RVM, including NGC/NGC (n = 33 of a total 55 sites; 60%) (Zhuo and Gebhart 1992) and NRM (n = 12 of a total 30 sites; 40%) (Zhuo and Gebhart 1997), produced biphasic modulation of responses of spinal dorsal horn neurons to noxious heating of the skin. In the present study, electrical stimulation at 24 of a subtotal 63 sites (38%) in the NGC/NGC and NRM produced biphasic modulation of responses of spinal neurons to noxious CRD, confirming that descending biphasic modulatory system(s) from the RVM affect spinal cutaneous as well as visceral nociceptive transmission. However, in previous studies of cutaneous transmission, relatively few facilitatory sites (7/85 sites, 8.2%) in the NGC/NGC and NRM were found. A significantly greater number of sites (9/59 sites, 11.9%) in the RVM were found in the present study where electrical stimulation (5-100 µA) only facilitated responses of spinal dorsal horn neurons to noxious visceral stimulation. Future studies are clearly needed to determine whether the difference in the number of facilitatory sites may have biological significance, as suggested from results of experiments examining tonic descending influences. Interestingly, electrical stimulation-produced facilitation and inhibition from the NGC/NGC and NRM appears to be selective for evoked activity (also see Zhuo and Gebhart 1992, 1997). Electrical stimulation in the RVM in the present study did not affect the spontaneous activity of spinal dorsal horn neurons, whereas responses of the same spinal neurons to noxious visceral stimulation were facilitated or inhibited.

Chemical activation of the RVM

Glutamate microinjection into the RVM produced facilitation of responses to noxious CRD at a low concentration (5 nmol) and inhibition of responses at a greater concentration (50 nmol), consistent with previous studies of spinal cutaneous nociceptive transmission (Zhuo and Gebhart 1992, 1997). The effects of glutamate in these studies and effects of other chemicals (e.g., baclofen, Thomas et al. 1995; neurotensin, Urban and Gebhart 1997; Urban and Smith 1994) were dose dependent, time limited, and reversible. In related work (Coutinho et al. 1998), intra-RVM administration of N-methyl-D-aspartate or inhibitors of nitric oxide synthase dose-dependently and reversibly attenuated responses to noxious distension of the inflamed colon. Whereas electrical stimulation in the RVM certainly activated cells as well as axons, these results reveal that activation of cell bodies in the RVM is sufficient to engage descending facilitation of spinal visceral nociception. It is possible that the effects of glutamate reported here arose from a depolarization block produced by a high concentration of glutamate. This is unlikely for several reasons. Glutamate affected responses to CRD, but was without influence on resting activity of the same spinal neurons. Further, intra-RVM lidocaine blocked transmission and significantly affected spontaneous activity of spinal neurons. In addition, that dissimilar receptor agonists (e.g., baclofen, neurotensin) produced effects qualitatively indistinguishable from effects produced by intra-RVM glutamate argues against a glutamate-produced depolarization block as responsible for the effects produced.

Evoked versus spontaneous activity

It has been reported that background activity of spinal neurons can be affected by RVM stimulation. Chapman et al. (1985) reported that stimulation of the NRM at an average intensity of 300 µA inhibited spontaneous activity of spinal neurons (45 cells, 100%) recorded from in the thoracic spinal cord. Ness and Gebhart (1987) demonstrated that stimulation in the PAG or RVM produced inhibitory effects on spontaneous activity of 26 (RVM) and 5 (PAG) spinal neurons, while 5 other spinal neurons (RVM, 1; PAG, 4) were unaffected. Chandler et al. (1989) reported that electrical stimulation in the PAG (n = 35) and/or the midbrain reticular formation (n = 31) produced mixed effects (inhibitory, excitatory, or none) on spontaneous activity of spinal neurons recorded from in thoracic spinal cord. In the present study and in earlier studies (Zhuo and Gebhart 1992, 1997), electrical stimulation in the RVM produced biphasic, only inhibitory, or only facilitatory modulation of responses to noxious cutaneous or distending stimuli; the spontaneous activity of the same neurons was not significantly affected. The intensity of stimulation in the present study was less than used in other studies (Chandler et al. 1989; Chapman et al. 1985), which may explain the different effects on spontaneous activity observed.

Intensity coding

In spinal cutaneous nociceptive transmission, descending inhibitory and facilitatory influences from the brain stem on spinal nociceptive transmission was found to have parametrically different effects on the intensity coding of spinal neurons. Descending inhibitory influences produced by stimulation in the NGC or NGC (Zhuo and Gebhart 1992) significantly reduced the slope of SRFs (without changing the threshold for responses), while descending facilitatory influences produced a parallel, leftward shift of SRFs. These results suggest that different mechanism(s) are responsible for the descending facilitatory and inhibitory effects produced. In support, pharmacological studies found that descending inhibition and facilitation are mediated by different spinal neurotransmitter receptors (Zhuo and Gebhart 1990b, 1991), and previous studies reveal an anatomical separation in the spinal cord of descending inhibitory and facilitatory influences (Zhuo and Gebhart 1992, 1997).

In the present study, stimulation in the RVM at facilitatory intensities shifted the SRF to the left. This is consistent with previous reports of a parallel, leftward shift produced by facilitatory stimulation in RVM (Zhuo and Gebhart 1992, 1997), including a recent report that examined RVM modulation of a visceromotor response to CRD (Zhuo and Gebhart 2002). Stimulation in the RVM at inhibitory intensities in the present report significantly reduced responses to CRD. The slope of the SRF was reduced, consistent with previous reports (Zhuo and Gebhart 1992, 1997), but the reduction in slope was not statistically significant. Inspection of the data suggests a rightward parallel shift in the SRF at noxious intensities of CRD (40-100 mmHg) and an effect principally on the SRF slope at lower intensities of CRD. Accordingly, there is no change in response threshold and no statistically significant reduction in the mean slope of the SRF. A similar outcome was noted in a recent study of modulation of a visceromotor response to CRD (Zhuo and Gebhart 2002). In that study, the SRF was parallel shifted to the right between 40 and 100 mmHg CRD by inhibitory electrical stimulation in RVM. Differences between effects of facilitatory and inhibitory stimulation in RVM may represent pre- versus postsynaptic inhibition, the specific arrangement of bulbospinal neuron terminals on spinal neurons (Carstens et al. 1980), or different neurotransmitter and/or subtypes of receptors involved.

Tonic descending modulation

The presence of a tonic descending excitatory or facilitatory influence from supraspinal structures on spinal visceral nociceptive transmission has been reported (Akeyson et al. 1990; Euchner et al. 1993; Hummel et al. 1997; Tattersall et al. 1986). Reversible spinalization by cold reduced or abolished responses of spinal neurons to visceral stimuli. These results suggest that tonic descending excitatory influences from supraspinal sites modulates spinal nociceptive transmission. In the preset study, we found that local anesthetic injection into the NGC or NGC significantly decreased spontaneous activities of spinal units, suggesting that some descending excitatory influences originate from the RVM. This and other reports provide strong evidence that spinal visceral nociceptive transmission is subject to tonic excitatory modulation from the RVM.

In sum, the present results provide electrophysiological evidence that spinal visceral nociceptive transmission is modulated biphasically from the brain stem. Understanding mechanisms for descending facilitation and inhibition of spinal visceral nociceptive transmission will provide important insights for how visceral pain is transmitted and regulated in the CNS. Functional bowel disorders, for example, which exist in the absence of demonstrable organ pathology, may reflect prevailing facilitation of visceral spinal input (instead of the normally prepotent descending inhibition).