INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THERE ARE MANY FACTORS that influence meal size and meal termination, but the process of satiation can be conceptualized as interplay between feed-forward or reinforcing signals and negative feedback or satiety signals (35). From an anatomical perspective, negative and positive feedback signals originate mainly from mechanosensors and chemosensors lining the alimentary canal and portal-hepatic axis and are carried to the brain via vagal and dorsal root primary afferents, whereas other signals can reach the brain in the form of hormones and other factors through the circulation.

Vagal sensory fibers have been recognized as one of the major mediators of negative feedback signals (15, 30, 36), and their ability to reduce meal size and possibly total food intake makes their neurochemical organization interesting as possible targets for new anorexic agents. Primary afferent vagal fibers terminate mainly in the nucleus of the solitary tract and other components of the dorsal vagal complex, and there seems to be a limited viscerotopography, in that gastric mechanoreceptor afferents can synaptically activate neurons predominantly in the lateral aspect of the medial subnucleus of the nucleus of the solitary tract (NTS; 16, 40) and small intestinal nutrient sensor afferents predominantly in the more medial and dorsal aspects of the medial subnucleus, as well as in the commissural and dorsomedial subnuclei and in the area postrema (29). However, the large areas of overlap suggest that there is either extensive collateralization at the periphery or convergence in the NTS, or both.

The identity of the neurotransmitter(s) released from primary afferent nerve terminals has been the subject of intense investigation. Most experiments have been directed toward cardiovascular vagal afferents or unspecified vagal afferents coursing through the cervical vagi. Based on morphological studies aimed at identifying glutamate in afferent vagal neurons (22) and nerve terminals (32) or glutamate receptors on second-order neurons in the NTS (Ref. 10 and unpublished observations), microdialysis studies (1, 37), and electrophysiological studies (3-5, 18), it becomes clear that glutamate and its N-methyl-D-aspartate (NMDA) and non-NMDA receptors play a prominent role at this first synapse of visceral sensory signaling pathways. It should be emphasized, however, that few other possible transmitter candidates have been studied in detail and that not all studies found supportive evidence (37).

Unlike cardiovascular afferents, there is not much direct experimental evidence that vagal afferents specifically innervating the gastrointestinal tract and carrying potential satiety signals use glutamate as central transmitter. There is some limited evidence from extracellular (24) and intracellular recording (44) experiments, showing that NTS neurons that respond to gastric distension can be inhibited by local application of kynurenic acid. Indirect evidence started with the observation that systemic administration of the noncompetitive NMDA receptor antagonist MK-801 facilitated ingestion of intraorally delivered sucrose, and food deprivation or MK-801 administration decreased glutamate levels in the NTS (7). Then, Burns and Ritter (12, 13) demonstrated that systemic MK-801-induced delay of satiation and increased meal size in rats trained to drink 15% sucrose solution after food deprivation depended on the integrity of vagal fibers. Finally, they found similar effects of the NMDA receptor blocker if directly infused into the fourth ventricle and NTS (39). Because the effect of the blocker was observed mainly during the later phases of a liquid meal and it was not seen with nonnutritive saccharin solutions, it appears to depend mainly on intestinal and portal-hepatic nutrient sensors, although an interaction with gastric mechanosensors cannot be excluded. To dissect the contribution of specific peripheral sensors in this MK-801-sensitive vagal sensory signaling pathway, we complemented the behavioral test paradigm with an anatomical approach that used neuronal Fos expression in the medulla after selective vagal afferent stimulation.

Thus the aims of the present study were to determine the ability of local, fourth ventricular infusion of the NMDA receptor blocker MK-801 to reduce gastric distension-induced neuronal activation in the dorsal vagal complex and to prolong food deprivation-induced sucrose drinking. We and others have previously shown that expression of the immediate-early gene product c-Fos can be used as an indicator of neuronal activation induced by gastric balloon distension in unanesthetized rats (16, 40). If second-order neurons in the medulla are synaptically activated by gastric mechanoreceptor stimulation via NMDA receptors, pretreatment with the selective NMDA receptor blocker MK-801 should prevent or reduce Fos expression.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. Adult male Sprague-Dawley rats (Harlan Industries, Indianapolis, IN), weighing 250-310 g at the time of surgery, were housed individually in hanging wire mesh cages in a climate-controlled room (22 ± 2°C) on a 12:12-h light-dark cycle with lights on at 0700. Food and water were available ad libitum except during the nights before surgery and testing. All testing was conducted in the light phase, between 0830 and 1300.

Implantation of chronic gastric fistula. Overnight food-deprived animals were anesthetized with pentobarbital sodium (Nembutal, 60 mg/kg ip) and given atropine (1 mg/kg ip). When all reflexes were absent, the animals were shaved, and a midline laparotomy was performed, exposing the stomach. Stainless steel gastric drainage fistulas were implanted in accordance with a method described in detail earlier (41) with slight modifications. Briefly, the inner flange of the cannula was inserted into the lumen of the stomach through a small stab wound at the greater curvature near the border between fundus and corpus and secured in place by a purse-string suture. The outer flange was drawn through a hole in the peritoneal wall and positioned so that the stomach rested in its normal position within the peritoneal cavity.

Immediately after the gastric fistula implantation, the rat was placed in the stereotaxic frame for implantation of a fourth ventricular guide cannula. A small incision was made on the skull, and the exposed bone region was cleared of tissue. The flat-skull stereotaxic coordinates used for the placement of the 22-gauge stainless steel guide cannula were 3.0 mm caudal to lambda, 0.0 mm lateral, and 5.8 mm below the dura. The cannula was anchored to the skull with screws, and the assembly was embedded in dental cement. The guide cannula was occluded with a 26-gauge stylet. Postoperative care included application of a topical antibiotic (Furazolidone), a systemic antibiotic (Gentamycin, 1 mg/kg im; Schering-Plough, Kenilworth, NJ), and daily treatment of the area surrounding the gastric fistula with 3% hydrogen peroxide. For the behavioral tests only the fourth ventricular cannula was implanted. The correct position of the cannula was verified by examining the cannula track on frontal brain sections and required that the track completely penetrate through the cerebellum but not puncture the floor of the fourth ventricle. Six out of 37 rats did not meet this criterion and were excluded from the final analysis. Gastric fistulas and fourth ventricular cannulas were implanted at least 10 days before testing.

Adaptation and test procedure. After recovering from surgery, the rats were adapted to the experimental procedures to minimize any stress-induced Fos expression. For sucrose consumption, animals were put on a food deprivation schedule with normal lab chow ad libitum except for 15 h of overnight food deprivation every 3-4 days. On days after food deprivation, rats were trained to lick 15% sucrose from a drinking spout until 30 min sucrose consumption stabilized (typically after 5-6 trials). On two occasions the animals were adapted to inserting an injector attached to a length of PE-10 tubing into the guide cannula.

On experimental days, the stylet was removed and MK-801 (2 nmol in 3 µl sterile saline) or saline alone was infused over 30 s into the fourth ventricle by means of a 10-µl Hamilton syringe attached to the PE-10 tubing and injector. At the end of the injection, the injector was left in place for 2 min and then replaced by the stylet, and the rat was returned to its home cage. Fifteen minutes after the injection, the spout of a graduated drinking tube containing 15% sucrose solution was made available, and intake was measured every 5 min to the nearest 0.1 ml. Each rat was tested twice each with the drug and vehicle in random order.

In the gastric distension study, animals were habituated to the test cage for 1 h during each of the first 3-4 days. For the next 3-4 days, the animals were habituated to a fourth ventricular mock injection, to rinsing the stomach with warm saline, and to the insertion of a deflated balloon. The gastric balloon was 30 mm long and attached to a piece of polyethylene tubing with a 1-mm outside diameter. Fifteen hours before testing, all food was removed from the cage. On the test day, the stomach was briefly rinsed with prewarmed saline before the deflated balloon was inserted. After a 30-min resting period, either MK-801 or saline was infused into the fourth ventricle as described in the section above. Another 15 min later, the gastric balloon was filled with warm water through the extension tubing connected to a 20-ml plastic syringe mounted on a syringe pump. The balloon was filled at a rate of 1.4 ml/min over the first 5 min, then 0.6 ml/min for another 5 min (10 ml total), and then slowly deflated by 3 ml over the next 70 min, thus mimicking the dynamics occurring during voluntary intake of a liquid meal (16).

Seventy minutes after the end of the inflation procedure, the rats were killed with a lethal dose of pentobarbital sodium and transcardially perfused with 200 ml heparinized (20 U/ml) saline followed by 400 ml of cold, phosphate-buffered (0.1 M, pH 7.4), 4% paraformaldehyde containing 0.3% picric acid. The brain was removed and postfixed in the same fixative overnight at 4°C.

Immunohistochemistry. The brain stem block was placed in 25% sucrose overnight for cryoprotection. Coronal sections, 30 µm thick, were cut in a cryostat and collected as three series in 0.1 M PBS. Tissue sections from at least one rat of each group were processed at the same time. Sections were pretreated successively in 1) 1% sodium borohydride in PBS for 30 min to reduce background, 2) a solution of 1.5% hydrogen peroxide, 20% methanol, and 0.2% Triton X-100 and then rinsed four times in PBS until the bubbling stopped, to inactivate endogenous peroxidase activity, and 3) 5% normal goat serum (with 1% BSA, 0.5% Triton X-100) for 30 min to block nonspecific binding. Free-floating sections were then incubated with the c-Fos primary antibody (Ab-5; Oncogene Sciences, Cambridge, MA), at a dilution of 1:50,000, for 60 h at 8°C. All remaining steps were done at room temperature with gelatin-PBS (0.1% gelatin in PBS) for rinses. After four rinses, sections were incubated in biotinylated secondary antibody (goat anti-rabbit IgG, Jackson ImmunoResearch), diluted 1:1,000, for 90 min. After another three rinses, they were incubated in an avidin-biotin complex (1:500, Vectastain ABC Elite Kit, Vector Laboratories). After another three rinses, the antibody complex was then visualized by immersing the sections in a Ni/Co-enhanced DAB substrate (0.5% 3,3'-diaminobenzidine tetrahydrochloride, 1% cobalt chloride, and nickel chloride with stable hydrogen peroxide; Pierce Chemical, Rockford, IL) for 2-10 min, resulting in a dark brown-to-black nuclear stain. The sections were then rinsed three times in PBS and mounted in glycerol or, after dehydration, in Permount.

For negative controls, some sections were processed by leaving out the primary antibody or were incubated in primary antibody that had been preadsorbed by incubation with 10 µmol synthetic c-Fos protein (peptide-2, lot no. 954701, Oncogene Sciences). In all cases this resulted in total absence of staining.

Imaging and automated counting of Fos-positive neurons. Images of representative sections were obtained in a laser scanning microscope (Zeiss LSM 310) in the (nonconfocal) transmitted light mode using the red line of a helium/neon laser. Images were sharpened with internal filtering algorithms and printed in a digital dye-sublimation printer.

For automated counting, sections were viewed on a Zeiss Axiovert 135 inverted microscope with a ×5 objective and a black-and-white video camera. Images were captured by frame grabber, normalized to span the entire grayscale, and then analyzed with a Windows-based image analysis system (KS400; Kontron Electronics, Munich, Germany). The object-recognition parameters such as size, shape, and threshold gray value as applied to nuclear Fos staining were first optimized by manual counting of a representative area, and these parameters were then stored as a macro programming routine and applied to all sections and treatments.

For each brain stem, a total of nine sections was used for counting, covering a rostrocaudal area of 1,000 µm between the caudal end of the area postrema to a point rostral to the area postrema where the NTS no longer borders the fourth ventricle (see Fig. 3 for the exact location of each section). The dorsal vagal complex (DVC) was divided into NTS, area postrema, and dorsal motor nucleus (dmnX), and the NTS was further subdivided into subnuclei or other specified regions with use of the atlas of Paxinos and Watson (28) and the maps of Altschuler et al. (2). The number of Fos-positive neurons in each of these subareas was counted with the field count routine. Counts for some of these subareas, either from one section or across several sections, were then combined to give scores for entire subnuclei, nuclei, or sections. Corrections were made for double counting of labeled neurons falling on the lines dividing the subareas.

Statistical analysis. Data of sucrose intake were analyzed with a repeated-measure design, with intake as the response variable and time as the repeated factor. The covariance matrix was assumed to be unstructured. Similarly, the Fos distribution along the rostrocaudal axis (see Fig. 3) was analyzed with the same statistical model, with the number of Fos-positive neurons as response variable and location as repeated factor. For analysis of Fos-induction in the various areas and subnuclei (see Fig. 4, A and B), separate two-way ANOVAs were performed, with distension status and fourth ventricular infusion as the two factors, followed by Tukey-Kramer's adjustment. For all analyses the program PROC MIXED in SAS, version 6.12, was used.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of MK-801 on sucrose intake. Infusion of MK-801 into the fourth ventricle increased the amount of 15% sucrose drunk in 30 min compared with saline infusion (Fig. 1). The effect of MK-801 on sucrose consumption was manifest only during the later part of the liquid sucrose meal. Intake during the first 10 min was identical (12.4 ± 1.2 vs. 11.5 ± 0.9 ml), but during the final 20 min of the test, intake after MK-801 treatment was about threefold higher than after saline control treatment (8.9 ± 1.0 vs. 2.9 ± 0.8 ml, P < 0.01; Fig. 1).

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Effect of 4th ventricular infusion of N-methyl-D-aspartate (NMDA) receptor blocker MK-801 on ingestion of 15% sucrose solution in 15-h food-deprived rats. Infusions of MK-801 (2 nmol in 3 µl saline) or saline were made 15 min before access to sucrose. Curves show cumulative intake at 5-min intervals; inset shows early (0-10 min) and late (10-30 min) intake. Data are means ± SE of 6 rats. * P < 0.05, ** P < 0.01 (repeated-measures design, PROC MIX in SAS).

Effect of MK-801 on gastric distension-induced Fos expression. In animals subjected to the combined control treatments with sham gastric distension and fourth ventricular saline infusion, there were generally few Fos-positive neurons in the dorsal vagal complex at any rostrocaudal level (Figs. 2A and 3). Gastric distension with 10 ml induced Fos expression in a large number of neurons throughout the dorsal vagal complex, particularly in the medial subnucleus of the NTS (Fig. 2), as has been shown earlier (40). Infusion of MK-801 in the absence of distension also significantly increased the number of Fos-positive neurons, particularly at levels near the rostral end of the area postrema (Figs. 2 and 3). Finally, distension in the presence of the blocker produced Fos expression that was not different in general pattern and number from that with distension alone (Figs. 2 and 3). Analysis of the total Fos counts for the entire DVC at the nine rostrocaudal levels sampled revealed 1) a significant effect of rostrocaudal location for all treatments (F_8,21 = 3.5-30.5, P < 0.01) except the sham distension with saline infusion (F_8,21 = 0.74, P = 0.66); 2) a significant effect of distension at each location regardless of drug treatment (MK-801: F_1,21 = 5.1-23.9, P < 0.05; saline: F_1,21 = 18.8-81.9, P < 0.001); and 3) no drug effect on Fos induction in the presence of gastric distension at any location (F_1,21 = 0.01-0.79, P = 0.38-0.96).

View larger version (141K): [in this window] [in a new window]   Fig. 2.   Effects of 4th ventricular infusion of NMDA receptor blocker MK-801 on gastric distension-induced Fos expression in dorsal vagal complex. Four experimental conditions (n = 6 each) included sham gastric distension after saline infusion (A), gastric balloon distension (10 ml) after saline (B), sham gastric distension after MK-801 (2 nmol in 3 µl saline) infusion (C), and gastric distension after MK-801 infusion (D). Fourth ventricular infusions were made 15 min before start of gastric distension. Gastric distension was maintained for 1 h. ap, Area postrema; c, central canal; com, commissural nucleus of the solitary tract (NTS); dm, dorsomedial NTS; X, dorsal motor nucleus of vagus; med, medial NTS; t, solitary tract.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Quantitative analysis of Fos induction produced by four experimental conditions (see legend to Fig. 2). Total Fos counts for entire dorsal vagal complex at 9 section levels in rostrocaudal axis are indicated in inset. Values are means ± SE of Fos-positive neurons per 30-µm section (repeated-measures design, PROC MIX in SAS). There was no significant effect of NMDA receptor blocker MK-801 on distension-induced Fos expression at any level. The blocker itself increased Fos expression at each level. Dist, gastric distension.

Therefore, based on this broad, anatomically nonselective analysis, there is no evidence that the blocker decreased gastric distension-induced Fos expression at any rostrocaudal level investigated. A finer-grained analysis was subsequently carried out, taking into consideration the various subnuclei of the NTS, particularly those that had been identified as major recipients of central terminals of primary vagal afferent neurons from the stomach. The raw Fos counts for all subnuclei and the four treatment conditions are shown in Fig. 4A. As for the broader analysis above, gastric distension significantly increased the number of Fos-positive neurons in each subnucleus, and this was not affected by MK-801 treatment (F_1,21 = 13.1-139.0, P < 0.01). MK-801 alone also significantly increased Fos counts in the dorsomedial, commissural, and gelatinosus subnuclei of the NTS (P < 0.05), but not in the other subnuclei, the area postrema, and the dorsal motor nucleus. This analysis also demonstrated that, although gastric distension was able to significantly further increase the number of activated neurons in the presence of MK-801 in the dmnX and the medial subnucleus of the NTS, this was not the case in the dorsomedial, commissural, and gelatinosus subnuclei. This was indicated by a significant interaction between the main effects of drug treatment and distension status in the commissural (F_1,21 = 7.8, P < 0.05) and gelatinosus subnuclei (F_1,21 = 6.26, P < 0.05). In the dorsomedial subnucleus the interaction did not quite reach statistical significance (F_1,21 = 3.39, P = 0.08).

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Effect of 4 experimental conditions (see legend to Fig. 2) on Fos expression across components of dorsal vagal complex and individual subnuclei of NTS. Values in A are means ± SE of Fos-positive neurons from 1 or more 30-µm section(s). Values in B are difference scores (as indicated) obtained from mean scores in A, showing net effect of distension under conditions of saline (open bars) or MK-801 infusion (hatched bars) into 4th ventricle and assuming that distension and MK-801 induce Fos in 2 separate populations of NTS neurons. In A, means that do not share same letter are significantly different (P < 0.05, based on separate 2-way ANOVAs followed by Tukey-Kramer's adjustments). In B, percent relative suppression of distension to induce Fos expression by MK-801 pretreatment is indicated (* P < 0.05). dmnX, dorsal motor nucleus of vagus; gel, gelatinosus; medm and medl, medial and lateral parts of medial subnucleus of NTS; rest, sum of all other subnuclei (central, lateral, interstitial, and intermediate) of NTS.

If this capacity of gastric distension to activate neurons in the absence or presence of MK-801 is graphically depicted by the subtraction scores (Fig. 4B), it is clear that the dorsomedial, commissural, and gelatinosus subnuclei are differently affected by the drug than the rest of the nuclei. If the Fos counts of these three areas are pooled (dorsomedial + commissural + gelatinosus), a significant relative suppression (75%, P < 0.05) of distension-induced neuronal activation by MK-801 becomes evident. In the other sampled areas there was a smaller relative suppression that did not reach statistical significance. However, this analysis of relative suppression is only useful and relevant under the assumption that gastric distension and MK-801 induce Fos in two completely separate populations of NTS neurons.

In addition to the DVC, fourth ventricular infusion of MK-801 induced Fos expression in other areas of the medulla, including a prominent cluster in the inferior olive, as has been reported before (25).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

This study used both a behavioral and an anatomical approach to test the hypothesis that glutamate is the transmitter released from central processes of vagal primary afferent neurons activated by satiety signals from the gastrointestinal tract. Although our behavioral results are consistent with this hypothesis, the results of the Fos activation study are less clear and can only be interpreted as supporting the hypothesis if certain anatomical assumptions are made.

Effect of MK-801 on satiety. Burns and co-workers (11-13) have convincingly shown that systemic administration of MK-801 postpones satiety in rats and that the effect depends on the integrity of the subdiaphragmatic vagus and, more specifically, visceral afferent fibers carrying nutritional information. Although these findings suggested that this effect might be due to blockade of glutamatergic transmission from primary vagal afferents to second-order neurons in the brain stem, it could have been due to such blockade at any site in the brain where vagal afferent signals are propagated. We have shown in the present study that infusion of the blocker into the fourth ventricle has exactly the same effect in postponing the onset of satiety, confirming the most recent observation by Treece et al. (39). In addition to infusing MK-801 into the fourth ventricle, these authors found essentially the same effect with very localized injections directly into the NTS. Taken together, these studies strongly support the hypothesis that glutamate is one of the transmitters used by vagal afferent fibers to signal satiety to the brain stem. However, even local injection into the NTS does not distinguish whether MK-801 blocks the action of glutamate released from primary afferents or from other sources, such as interneurons or projections from outside the injected area. Anatomical methods as discussed below will be necessary to address this problem.

The effect of glutamatergic blockade was only seen after ~10 min, when a substantial amount of sucrose had been ingested. Because at this time some of the ingested sucrose is in the small intestine and absorption is in full swing, not only gastric mechanoreceptors but also intestinal and perhaps portal-hepatic glucose and other nutrient sensors could generate the vagal afferent signals that lead to glutamate release in the medulla. On the basis of the fact that deprivation-induced saccharin drinking was not augmented by systemically administered MK-801 (11), it could be concluded that feedback from nutrient sensors but not gastric mechanosensors may use glutamate as a transmitter. However, because gastric mechanosensor activity is modulated by small intestinal nutrient sensor activity via the release of CCK (34) and because the volume of saccharin solution drunk was relatively modest, a role for glutamate receptors in the processing of gastric distension signals cannot be ruled out. In an attempt to determine the involvement of brain stem glutamatergic transmission in specific gastrointestinal feedback signals, we used the gastric distension-induced Fos expression paradigm (16, 40).

Does MK-801 block the ability of gastric distension to activate neurons in the DVC? In some areas (medial NTS, lateral NTS, area postrema, and dmnX) MK-801 did not significantly reduce gastric distension-induced Fos expression, and in the absence of distension it did not in itself increase Fos expression. For these areas of the dorsal medulla, NMDA receptor-mediated glutamatergic transmission clearly does not seem to play a major role in the propagation of gastric mechanoreceptor signals. However, at least two questions arise from these observations. First, gastric distension produced the largest number of activated neurons in the medial subnucleus of the NTS, but based on tracing of primary afferent neurons from the stomach, the major termination site for such afferents seems to be the area of the gelatinosus, commissural, and dorsomedial subnuclei (2, 26). This apparent discrepancy could have several explanations. Most likely, second-order neurons may send dendrites into the gelatinosus subnucleus, but their cell bodies may be located outside, such as the medial subnucleus. Extensive dendritic trees of NTS neurons have been shown by several investigators (31, 43). Alternatively, neurons in the medial NTS that show gastric distension-induced Fos expression may be third- or higher-order neurons, with local interneurons interspersed. If this were the case, the results would still not suggest a role for NMDA receptor-mediated glutamatergic transmission at any of the synapses involved, because MK-801 would have interrupted the entire pathway and induction of Fos.

For other areas, such as the dorsomedial, commissural, and gelatinosus subnuclei of the NTS, the results are complicated by the fact that MK-801 alone, in the absence of gastric distension, induced substantial and significant Fos expression. For these subnuclei, two basic assumptions can be made, each resulting in a different interpretation of our findings. The first assumption is that the blocker and distension both activate the same population of neurons (converging activation). If this were the case, the finding that there was no reduction of distension-induced neuronal activation by MK-801 in these subnuclei would again not support a role for NMDA receptor-mediated glutamatergic transmission in gastric signal processing.

The second assumption is that the blocker and distension each activate a separate population of neurons (parallel or additive activation). If this were the case, the finding that the ability of gastric distension to increase Fos expression is significantly and substantially (75%) reduced by MK-801 would support a role for glutamatergic transmission in these subnuclei.

The literature does not provide enough information to clearly favor one of these assumptions. Of course, the central question we may ask is: what is the mechanism by which a receptor antagonist that blocks excitatory synaptic transmission induces Fos expression? This phenomenon has been reported by others (21, 33), and the simplest explanation has been to invoke a mechanism of disinhibition. The presence of GABAergic inhibitory interneurons and GABA_A receptors in the NTS is well documented (23). For MK-801 to activate a population of neurons via disinhibition, there has to be an NMDA receptor-dependent tonically active excitation of GABA interneurons, which is removed by the blocker. Any neurons normally inhibited by these GABA interneurons could thus become active, provided they receive some other excitatory input. Some models of neural processing by the NTS (9) suggest that second-order sensory neurons are among the neurons tonically inhibited by GABA interneurons, and by this model MK-801 could induce Fos in these same second-order neurons that can be activated by gastric distension. Only further experiments using additional anatomical and neurochemical identification procedures will be able to verify these possibilities.

Taken together, in contrast to the behavioral experiment, our Fos study provides little evidence that the NMDA receptor plays a major role in mediating the vagal gastric mechanoreceptor signal at the level of the NTS. However, as discussed above, the effect in the behavioral experiment may have been mainly due to postgastric sensory signals, and the Fos study does not rule out mediation of intestinal or portal-hepatic chemosensor signals by the NMDA receptor. Experiments with selective stimulation of these sensors by duodenal and portal nutrient infusion are underway.

Perspectives

In addition to glutamate, other transmitters may be released from primary afferent terminals in the DVC. Based on immunohistochemical analysis of nodosal neurons, these candidates include CCK (19), serotonin (27), calcitonin gene-related peptide (20), and substance P (6). Specific receptors for all of these substances are present in the NTS (38). In a related sensory system, selective blockade of substance P and neurokinin receptors has been shown to significantly reduce corneal stimulation-induced Fos expression in the trigeminal sensory nucleus of the rat (8). The potential involvement of these transmission systems in gastrointestinal sensory processing remains to be evaluated.