INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE MIDLINE CAUDAL medullary raphe (MR) area contains neurons that influence sympathetic vasomotor outflow (1, 15, 22, 31, 33). However, the effects of MR stimulation reported by various investigators are frequently contradictory (8, 16, 32, 33). Dampney (7) has suggested that these variable effects may be due to stimulation of mixed populations of intermingled sympathoexcitatory and sympathoinhibitory cells (7). A depressor region in the caudal midline medulla of the rabbit has been characterized recently (4), and an analogous region has been identified in the rat (12). These studies used chemical excitants to selectively stimulate neurons within the MR region and demonstrated a profound influence of caudal midline medullary neurons on sympathetic vasomotor outflow.

Lesions of the midline medullary area did not alter resting arterial blood pressure but elevated the discharge of the inferior cardiac nerve (17). Furthermore, Zhong and colleagues (35) reported that chemical inactivation of neurons within the cat MR uniformly reduced the 10-Hz rhythmic discharges of the left and right inferior cardiac nerves (35). These observations suggest that neurons within the midline medulla are part of a neuronal network associated with sympathetic rhythm generation.

Studies that have used intermittent electrical stimulation paradigms have identified short- and long-latency sympathoexcitatory responses associated with MR stimulation (18, 36). The contrasting effects reported in these studies may depend largely on the stimulation paradigm used. Thus the use of chemical excitants avoids activation of fibers of passage, but the observed effect may be the net result of stimulation of mixed populations of neurons. On the other hand, intermittent electrical stimulation techniques have the disadvantage of indiscriminately stimulating all neuronal elements, including axons of passage, but allow temporal resolution of the various components of an evoked response.

In this study, we have focused on the mechanism of the sympathoinhibitory vasomotor responses evoked by stimulation of the MR by examining the effects on lumbar sympathetic vasomotor discharge and on the discharge of premotor sympathoexcitatory neurons of the rostral ventrolateral medulla (RVLM). By comparing the effects of electrical and chemical stimulation of these sites, we have emphasized the heterogeneity of caudal midline medullary neurons with regard to sympathetic vasomotor function. In addition, we demonstrate that in the rat, depressor and sympathoinhibitory responses elicited by MR stimulation are mediated by inhibition of the discharge of sympathoexcitatory neurons of the RVLM (11, 14, 19).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

General procedures. All experiments were performed on 27 male Sprague-Dawley rats (290-380 g), were approved by the ethical review committee of the Austin and Repatriation Medical Centre, and complied with the guidelines of the National Health and Medical Research Council of Australia. Animals were anesthetized initially by placement into a container saturated with enflurane vapor (Ethrane; Abbott Australasia, Kurnell, Australia), and when a surgical level of anesthesia was attained, the trachea was cannulated. All animals were ventilated artificially with 100% O₂ (1 ml/100 g body wt, 60-70 breaths/min) containing ~1.4% halothane (Fluothane; Zeneca, Macclesfield, UK). This concentration of halothane produced a deep surgical level of anesthesia and was maintained throughout the entire surgical procedure. The depth of anesthesia was verified by noting the absence of responses to firm paw pinch and corneal probing. Core temperature was maintained at 36.5-37.5°C with a servocontrolled heating pad. Arterial blood pressure was measured from the right carotid artery, and drugs were injected intravenously via the right jugular vein.

Single-unit recordings. Single-unit recordings of RVLM premotor sympathoexcitatory neurons were made as described previously (2, 10, 29). In the unit recording experiments, either a monopolar or a bipolar stimulating electrode was placed into the dorsolateral quadrant of the spinal cord ipsilateral to the recording electrode after a laminectomy at the T₃-T₄ level for antidromic activation of medullospinal units. An indifferent electrode was connected to a silver wire implanted in a muscle adjacent to the stimulation site (monopolar electrode only). The antidromic nature of spikes elicited by spinal stimulation (0.5 Hz, 0.5-ms duration, and 0.3-2.5-mA intensity) was established by demonstration of invariant latency and the collision test. Barosensitivity of RVLM neurons was tested by inflation of a subdiaphragmatic abdominal aortic snare to increase systemic arterial blood pressure. These neurons were inhibited by elevation of arterial blood pressure in a reproducible and proportional manner.

Electrical stimulation of caudal MR region. Before insertion of a stimulating electrode into the MR, the location of the caudal pole of the facial motor nucleus was determined using antidromic field potential recording. This procedure localized both the caudal tip and the ventral boundary of the facial motor nucleus. Thereafter, the glass recording microelectrode was replaced with a monopolar stimulating electrode (2-3 M impedance at 1 kHz, Frederick Haer) that was lowered stereotaxically through the cerebellum to the midline medulla oblongata to a depth ~2 mm dorsal to the ventral boundary of the facial motor nucleus. Stimuli were applied to the electrode as intermittent single pulses (0.4-0.5 Hz, 0.5-ms pulse duration, 25-200 µA). The selection of these stimulation parameters was based on the results of preliminary experiments conducted by the authors as well as those reported earlier by other investigators (13, 36). The electrode was lowered through the medulla incrementally (1,000-µm steps), and electrical stimulation (25-200 µA, 0.5 Hz, 0.5-ms pulses) was applied at each site. When electrical stimulation of the midline medulla was combined with extracellular single-unit recording of RVLM premotor sympathoexcitatory neurons, the RVLM was initially located as described above, and then the stimulation electrode was lowered into the brain using an additional micropositioner.

Chemical stimulation of caudal MR region. Chemical stimulation of the MR area was achieved by microinjection of sodium glutamate (Glu; 0.2 M, pH 7.4, 30-40 nl, in artificial cerebrospinal fluid) through a glass micropipette (tip diameter 20-40 µm) cemented to the tip of a 1-µl glass microsyringe fixed to a micropositioner. Rhodamine-tagged microbeads (LumaFluor, New York, NY, or Molecular Probes, Eugene, OR) were incorporated into the solutions of Glu to facilitate histological localization of the injection sites using fluorescence microscopy.

Chemical stimulation of caudal MR region and single-unit recording in RVLM. Combination of microinjection with single-unit recording posed a special problem associated with the size of the glass capillaries and the proximity of the microinjection and recording sites. This problem was overcome by introducing a 10-15° angle in the shank of the micropipette. The caudal contour of the facial motor nucleus on the right side of the medulla was mapped electrophysiologically as described above (see Single-unit recordings). The recording microelectrode was then replaced with the angled microinjector and moved to the midline. The facial motor nucleus was then mapped on the left side of the medulla, and RVLM barosensitive bulbospinal neurons were located as described above. The depressor area of the midline medulla was located by microinjecting Glu as described above.

Sympathetic nerve recording. Sympathetic vasomotor discharge was recorded from the lumbar sympathetic nerve trunk (L₃-L₅) as described previously (24, 26, 27). At the end of each experiment, clonidine (200 µg/kg iv) was administered. The signal remaining after clonidine administration was regarded as electrical noise, systematically subtracted from the rectified signal on computer analysis, and considered as the zero level of the sympathetic discharge. The resting level of sympathetic nerve discharge measured at the beginning of the experiment was used as the 100% level. Thus the sympathetic discharge was quantified as arbitrary "units" of activity (24, 26, 27).

Histological analysis of stimulation, injection, and recording sites. Electrical stimulation sites within the MR area were marked by electrolytic deposition of iron that was identified using the Prussian blue reaction as described previously (26, 27). Only the most ventral site of the most caudal microstimulation electrode penetration was marked. All other stimulation sites were determined using the location of the marked site and the stereotaxic coordinates as references. Recording sites within the RVLM were marked by iontophoretic deposition of Pontamine sky blue from the recording electrode. Stimulation and recording sites were then identified using light microscopy and mapped onto standard maps of the rat brain with reference to a rat brain atlas (21). Injection sites within the MR area were located by visualizing the rhodamine-tagged microbeads using fluorescence microscopy.

Data analysis and statistics. Arterial blood pressure, sympathetic nerve discharge, stimulation pulses, and single-unit activity were recorded on a videocassette recorder in pulse code modulation format (Vetter Instruments). A microcomputer-based data acquisition system was then used to construct peristimulus time histograms and histograms of evoked sympathetic activity as described previously (28, 29). The magnitude of the inhibition of RVLM unit firing produced by electrical stimulation of the MR area was calculated by dividing the total number of spikes observed during the period of inhibition by the number of spikes observed during an equivalent period of prestimulus firing. This figure was subtracted from unity to yield an index of inhibition. The mean of the differences between the number of spikes observed in the prestimulus period and the poststimulus period was subjected to a Student's paired t-test in which the null hypothesis was that the mean of the differences does not differ from zero. The level of significance was taken as 0.05. All data are expressed as means ± SE. Changes in neuronal firing rate induced by chemical stimulation of the MR were calculated by expressing the difference between the resting firing rate before Glu microinjection and that observed after Glu as a percentage of the resting firing rate.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Electrical stimulation of caudal MR region and lumbar sympathetic nerve discharge. Intermittent electrical stimulation (0.5 Hz, single 0.5-ms rectangular pulses, 100-200 µA) of the MR area produced an evoked response that consisted of sympathoexcitatory and sympathoinhibitory components. In most cases (n = 5 experiments), this pattern consisted of a long-latency sympathoexcitatory response superimposed on a sympathoinhibitory phase. The onset latency of the long-latency sympathoexcitatory response was 146 ± 7 ms, and the peak latency was 176 ± 51 ms. Occasionally (2 of 5 experiments), a short-latency sympathoexcitatory response was also observed (onset latency = 40 ms). The inhibitory phase (onset latency = 59 ± 10 ms; duration = 268 ± 39 ms) always preceded the long-latency sympathoexcitatory response. Figure 1 depicts a representative series of responses observed on stimulation at various depths and rostrocaudal locations (Fig. 2) within the midline medullary region in a single experiment. The computer-averaged evoked sympathetic responses consisted of a prominent inhibitory trough whose nadir occurred ~250 ms after the onset of stimulation (Fig. 1, F, H, and I). The trough is most prominent at ventral and caudal locations within the region of the medulla investigated. A large sympathoexcitatory peak is also apparent at ~175 ms (Fig. 1, C, F, and I). It should be noted that, due to the brevity of the evoked sympathetic response, no obvious changes in arterial blood pressure were observed. Consequently, the response was not influenced by secondary effects due to activation of the arterial baroreflex.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Effect of intermittent electrical stimulation (0.5 Hz, 0.5-ms pulse duration, 200 µA) of midline medulla on averaged evoked sympathetic response recorded from lumbar sympathetic nerve in a typical experiment. Each histogram is an average of 150 repetitions of response to stimulus presented at t = 0. Sympathetic discharge is expressed as units, where resting level of rectified and filtered signal is equivalent to 100 units and signal remaining after clonidine (200 µg/kg iv) is regarded as zero. Sharp peak at t = 0 appearing in some histograms is a stimulation artifact. A-I represent responses obtained by stimulation at sites identified in Fig. 2.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Locations of electrical stimulation sites within midline medullary area. A-I refer to responses obtained to intermittent electrical stimulation depicted in Fig. 1. Numbers at left of each section represent distance (in mm) from bregma. VII, facial motor nucleus; Amb, ambiguus nucleus, compact formation; NTS, solitary tract nucleus; py, pyramid; RMg, raphe magnus nucleus; ROb, raphe obscurus nucleus; Sp5I, spinal trigeminal nucleus, interpolar region.

Chemical stimulation of caudal MR region and lumbar sympathetic nerve discharge. Microinjection of Glu typically elicited depressor and sympathoinhibitory responses whose magnitude was dependent on the location within the midline medullary area. Figure 3 shows blood pressure and sympathetic nerve responses to Glu microinjection obtained from two medullary sites depicted in the inset. In this example, the depressor and sympathoinhibitory responses were preceded by a transient (3-4 s) pressor and sympathoexcitatory response. This pattern of response was observed in three of seven cases. Pressor and sympathoexcitatory responses unaccompanied by depressor and sympathoinhibitory responses were never observed in this study. The lumbar sympathetic nerve responses always preceded the arterial blood pressure changes. Heart rate was either unchanged during the depressor responses, or bradycardic (10-30 beats/min) responses were observed (data not shown). Depressor responses to Glu ranged from 10 to 60 mmHg and lasted up to 30 s. These were always accompanied by prominent inhibition of lumbar sympathetic nerve discharge that returned to baseline levels just before return of blood pressure to preinjection levels. Results from several separate experiments (n = 7 rats) are depicted in Fig. 4. These data indicate that Glu-induced depressor responses may be elicited from a number of locations within the midline medulla oblongata and that the most effective sites are found within the ventral medulla at a level rostral to the area postrema and at the middle third of the inferior olive.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   A and B: effects of chemical stimulation [glutamate (Glu), 6 nmol/30 nl, injected at ] of midline medulla on mean arterial blood pressure (MAP) and integrated lumbar sympathetic nerve discharge (LSND). Note transient pressor and sympathoexcitatory response that precedes depressor and sympathoinhibitory responses. Microinjection sites are depicted in inset. Bar represents 60 s. XII, hypoglossal nucleus; LRt, lateral reticular nucleus.

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Summary of Glu microinjection experiments. Small circles, depressor responses <29 mmHg; medium circles, depressor responses >30 mmHg and <49 mmHg; large circles, depressor responses >50 mmHg. Numbers at right of each section represent distance (in mm) from bregma. RPa, raphe pallidus nucleus; RVLM, rostral ventrolateral medulla.

Caudal MR stimulation and RVLM premotor sympathoexcitatory neuronal discharge. Electrical stimulation (25-200 µA) of the MR region inhibited the discharge of 12 of 15 (80%) neurons classified as RVLM premotor sympathoexcitatory cells (2, 9, 26). All of these neurons were found within 300 µm of the caudal pole of the facial motor nucleus and were barosensitive, their discharge was pulse modulated, and they had a mean basal discharge rate of 13 ± 3 spikes/s (n = 15, range = 1-38 spikes/s). Approximately 53% (8 of 15) of these neurons were activated antidromically by spinal cord stimulation (mean antidromic latency = 30 ± 10 ms, range = 12-82 ms), indicating a spinally projecting axon. The onset latency of the inhibition induced by midline medullary stimulation was 20 ± 2 ms, and the duration of the inhibition was 42 ± 4 ms. Figure 5, A-C, depicts peristimulus time histograms of the discharge of an RVLM neuron produced by intermittent electrical stimulation (25-100 µA) of the midline medulla. At a low current intensity (25 µA), inhibition of this unit was still detectable. An oscilloscope view of the inhibition of RVLM unit discharge (50-µA stimulation intensity) induced by MR stimulation is also depicted (Fig. 5D). The mean inhibition index for these neurons was 0.85 ± 0.06 (0.40-1.00, n = 12, P < 0.05). Two neurons displayed evidence of excitation that preceded the inhibition produced by midline medullary stimulation (onset = 10 ms) at the highest current tested only (200 µA). In the remaining 10 cells, no evidence of an excitatory response was detected. RVLM premotor sympathoexcitatory neurons were not antidromically activated by MR stimulation.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Peristimulus-time histograms of an RVLM sympathoexcitatory vasomotor neuron obtained during stimulation of vasodepressor sites within midline medulla oblongata (0.5 Hz, 0.5-ms pulses, 150 sweeps). A: 25 µA, B: 50 µA, C: 100 µA. Electrical stimulation was presented at t = 0 (note stimulation artifact at t = 0). D: oscilloscope record of RVLM unit showing inhibition of firing produced by midline medullary stimulation (50 µA, 250 sweeps).

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Inhibition of an RVLM sympathoexcitatory neuron by activation of midline medullary depressor area with Glu (40 nl, 0.2 M). Elevation of MAP by gradual aortic occlusion (AOc) reproducibly inhibited discharge of neuron. Reduction of MAP with sodium nitroprusside (SNP; 5 µg/kg iv) resulted in baroreceptor unloading and an increase in neuronal discharge rate. Glu-induced depressor response is accompanied by a reduction in discharge rate of cell.

Localization of stimulation and recording sites. The locations of the stimulation electrode tips and recording sites in the single-unit recording experiments are shown in Fig. 7. These were based on the histological identification of the Prussian blue-stained spots in the midline and the locations of the Pontamine sky blue deposits made within the RVLM. All RVLM units were recorded within 300 µm of these sites. Locations of the Glu injection sites and RVLM recording sites are depicted in Fig. 8.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Locations of midline medullary electrical stimulation sites and RVLM recording sites. Large filled circles represent locations of Prussian blue spots that identify location of tip of stimulation electrode. Small filled circles represent locations of Pontamine sky blue deposited iontophoretically from recording microelectrode in unit recording experiments. All neurons were recorded within 200-300 µm of these marks of the Pontamine sky blue marks.

View larger version (36K): [in this window] [in a new window]   Fig. 8.   Locations of midline medullary chemical stimulation sites and RVLM recording sites. Small filled circles represent locations of Pontamine sky blue spots that identify location of tip of recording microelectrode. All neurons were recorded within 200-300 µm of these marks. Large filled circles represent locations of deposits of fluorescent microbeads incorporated into solutions of Glu used for microinjection. Numbers at right of each section represent distance (in mm) from bregma.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

This study has provided evidence that the rat MR region contains neurons that exert a powerful inhibitory influence on sympathetic vasomotor discharge via an inhibitory action on RVLM premotor sympathoexcitatory neurons. A number of other studies have reported similar findings in studies in which more rostral midline structures were stimulated (16, 30). We have extended these investigations by studying the effects of electrical and chemical stimulation of the midline depressor area on sympathetic nerve discharge and RVLM premotor sympathoexcitatory neuronal discharge. The powerful influence of the MR on sympathetic vasomotor activity was evident after both chemical and intermittent electrical stimulation of the MR.

The findings of the present study suggest that the MR area contains mixed populations of neurons capable of influencing sympathetic vasomotor function. Intermittent electrical stimulation of the MR allowed discrimination of sympathoexcitatory and sympathoinhibitory components largely because these occurred at different latencies. On the other hand, responses to chemical excitants reflect the net effect of stimulation of both excitatory and inhibitory elements that may not be discriminated unless these have differing thresholds or are separated spatially. Despite this, chemical stimulation of the midline medullary region occasionally produced transient pressor and sympathoexcitatory responses that preceded the profound depressor and sympathoinhibitory responses. Although these observations support the notion of intermingled sympathoinhibitory and sympathoexcitatory elements within the midline medulla, it is also possible that the sympathoexcitatory effects were mediated in part by activation of passing fibers.

Short-latency sympathoexcitatory responses have been recorded from the splanchnic nerve in anesthetized rats on stimulation of rostral raphe obscurus (36). In contrast, Huangfu and colleagues (13) reported only a long-latency sympathoexcitatory response that was predominantly, but not exclusively, mediated by stimulation of spinal serotonin receptors. These investigators also reported the presence of a distinct sympathoinhibitory phase on stimulation of the midline medulla, but this was not pursued further. In the present study, long-latency sympathoexcitatory responses predominated, but occasionally indications of a short-latency component were present. These discrepancies may be associated with differences in the specific sympathetic outflow studied in each investigation (e.g., lumbar sympathetic nerve trunk and splanchnic sympathetic nerve).

Because RVLM premotor sympathoexcitatory neurons were inhibited by MR stimulation, it is most probable that the depressor and sympathoinhibitory response is mediated by a propriomedullary pathway, which terminates onto RVLM neurons. In addition, the sympathoexcitatory response may be mediated by spinally projecting serotonergic neurons of the raphe obscurus/pallidus (13, 36). The net result of stimulation of these diverse elements may depend on experimental conditions (e.g., type of anesthetic and depth of anesthesia), location within the medulla, and also the nature of the excitatory amino acid receptor activated (6).

The most effective location for Glu-induced depressor responses coincided approximately with the location of the sites at which intermittent electrical stimulation produced an evoked sympathetic response containing a profound sympathoinhibitory component. This area, which was within the rostrocaudal level of the middle third of the inferior olive, is similar to that described by Henderson and colleagues (12) in the rat but more rostral to that identified in the rabbit (4).

A major question arising from these studies is whether the input from the midline medullary depressor area to RVLM premotor sympathoexcitatory neurons is mono- or polysynaptic. Evidence from experiments performed in the cat would suggest that the latter alternative is most probable (16). This study found that neurons with the characteristics of premotor sympathoexcitatory vasomotor neurons were inhibited by rostral midline medullary stimulation. However, rostral midline medullary neurons were not antidromically activated by stimulation within the RVLM pressor area. In addition, some cells were excited by depressor area stimulation at a shorter latency than those cells inhibited by microstimulation. The inhibition of RVLM neuronal discharge induced by rostral midline medullary stimulation was blocked by iontophoretic application of the GABA_A receptor antagonist bicuculline. Taken together, these observations suggest that midline medullary stimulation activates GABAergic interneurons that synapse onto RVLM premotor sympathoexcitatory neurons. Recent investigations performed in the rabbit have also supported the notion of an inhibitory GABAergic link (5). The location of this GABAergic pathway is presently unknown.

Chemical stimulation of the MR also inhibited the discharge of the majority of the RVLM medullospinal sympathoexcitatory neurons tested. Under pentobarbital sodium anesthesia, these neurons are under the influence of a powerful baroreceptor-mediated inhibition. This was reflected in the large increase in neuronal discharge observed on unloading the baroreceptors with sodium nitroprusside. Indeed, this effect is likely to result in underestimation of the magnitude of the inhibitory effect of MR stimulation with Glu, because the resultant depressor response would serve to unload the baroreceptors in a manner similar to the actions of the vasodilator agent. Although this confounding factor could have been avoided by using baroreceptor-denervated animals, it would also make positive identification of RVLM barosensitive cells difficult. Neurons that were not affected by Glu microinjection were nevertheless activated by reduction of arterial blood pressure. This suggests that these cells were also inhibited by Glu microinjection, but less powerfully. Importantly, the results of the chemical stimulation experiments support those obtained on intermittent electrical stimulation of the MR.

Evidence obtained from anterograde tracing experiments suggests that the midline medullary region projects directly to the RVLM premotor sympathoexcitatory neurons (20, 34). On the other hand, studies using retrograde tracers injected into the RVLM have not identified a major projection from the MR (3, 25). Anterograde tracing experiments using biotin-dextran (BD) conducted in our laboratory (unpublished observations) have indicated that direct contacts between BD-labeled efferents from the MR and tyrosine hydroxylase (TH)-labeled cells of the RVLM were infrequent, although BD-labeled varicose fibers were intermingled with the TH-positive neurons. The discrepancies between these studies may be explained by the fact that the midline medullary injection sites were not characterized functionally. Thus it is possible that direct projections to RVLM TH-positive neurons from the midline depressor sites may be much less prominent than from midline regions more rostral to the depressor area.

Perspectives