Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The functional roles of EAAs and their receptor subtypes in the NTS have been studied by many investigators. Microinjection of EAA agonists, such as glutamate, kainate, NMDA, quisqualate and L-aspartate, into NTS induces hypotension and/or bradycardia (Talman et al., 1980; Kubo and Kihara, 1988; Galloudec et al., 1989; Leone and Gordon, 1989; Ohta et al., 1993; and Tian and Hartle, 1994). Conversely, microinjection of EAA antagonists, such as GDEE, kynurenic acid, AP-5, MK-801 and CNQX, into NTS results in hypertension and blocks (or attenuates) the baroreflex (Talman et al., 1981; Guyenet et al., 1987; Talman, 1989; Leone and Gordon, 1989; Kubo and Kihara, 1988, 1991; Galloudec et al., 1989; Gordon and Leone, 1991; Vardhan et al., 1993; Tian and Hartle, 1994).

Although these microinjection studies clearly establish a relationship between EAAs and the baroreflex, many details of EAAs and their receptor subtypes at the level of single cells are still not understood because of limitations inherent in this approach. For example, the EAAs administered by microinjection nonselectively excite almost all neurons around the injection area, and some of these neurons affected by microinjecting EAAs are probably interneurons (such as GABAergic neurons). Thus indirect influences, such as the inhibitory influences mediated by EAA agonist-induced excitation of GABAergic neurons, are not excluded in microinjection studies. In addition, microinjection studies cannot discriminate which receptor subtypes are utilized in transmission from primary afferent to second-order neuron and in transmission between higher-order neurons.

In vitro electrophysiological studies have provided much information at the cellular level regarding EAAs and their receptor subtypes in neuronal transmission and integration in NTS. For example, perfusion of EAA agonists, such as glutamate, kainate, AMPA and trans-ACPD, in slice preparations or to acutely dissociated cells excites (or depolarizes) NTS neurons recorded extracellularly, intracellularly or by patch clamp (Drewe et al., 1990; Tell and Jean, 1990; Glaum and Miller, 1992; Glaum et al., 1993; Nabekura et al., 1994). Perfusion of EAA antagonists, such as GDEE, kynurenic acid, AP-5 and CNQX, into the brain slice preparation blocks the excitation of NTS neurons evoked by stimulation of the solitary tract (Miller and Felder, 1988; Andresen and Yang, 1990). Although in vitro studies have provided valuable information at the cellular level, these techniques also have limitations. First, one cannot identify the functional role(s) of cells in slice preparations, especially when the recording is performed in NTS, which receives a wide array of afferent inputs. A second limitation is the indirect influence of interneurons. As in microinjection studies, perfusion of slice preparations with EAA ligands affects a greater number of other neurons in the preparations in addition to the one being recorded. Indirect influences of EAAs on NTS neurons in some in vitro studies have been reported (Miller and Felder, 1988). The third limitation is that in vitro preparations change the environment around the recorded cells (for example, tonic afferent inputs, hormones and other central active substances are absent or reduced in the in vitro preparations), and the substantial dendritic trees of NTS neurons are severed during the preparation. These changes could influence the responses of the NTS cell to EAA ligands and thereby modify neurotransmission and integration of afferent inputs in NTS.

Experiments conducted recently in our laboratory were designed to bridge the gap between microinjection and in vitro studies by using in vivo electrophysiological single-unit recording techniques combined with microiontophoresis. Baroreflex-related NTS neurons in vivo were identified by their responses to stimulation of the ADN. The ADN in rat contains fibers that originate solely from baroreceptor endings (Krieger and Marseillan, 1963; Sapru and Krieger, 1977; Sapru et al., 1981; Numao et al., 1985). In addition, second-order (monosynaptic) or higher-order (polysynaptic) baroreflex neurons in NTS were identified by standard electrophysiological methods (Miles, 1986). To reduce indirect influences produced by drug administration, we used microiontophoresis, a commonly used techniques for local application of drugs in vivo. The tips of our electrodes for recording and microiontophoresis were maintained at a small size (2-2.5 µm), and the ejection currents were limited to 0 to 40 nA. It has been demonstrated that using this kind of electrode and ejection parameters, greatly reduces the indirect influences of drugs applied by microiontophoresis (Zhang et al., 1992, 1994). The experimental goals in the present study were threefold: 1) to test the responses of baroreflex-related NTS neurons to EAA agonists (including EAA agonists selective for certain receptor subtypes), 2) to compare the responses to EAA agonists of second-order and higher-order baroreflex-related NTS neurons and 3) to compare the responses to EAA agonists of baroreflex-related NTS neurons to those of non-ADN evoked NTS neurons.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Successful experiments were performed on 50 male Sprague-Dawley rats (350-500 g, Charles River Laboratories, Willington, MA). Rats were housed two per cage in a fully accredited (AAALAC and U.S.D.A.) laboratory animal room with free access to food and water. All experimental rats were given at least 1 week to acclimate before use. All experimental protocols were approved by the Institutional Animal Care and Use Committee.

Surgical preparations. Rats were initially anesthetized with pentobarbital sodium (60 mg/kg i.p.) and were placed on a thermostatically controlled heating pad. Body temperature was maintained at 36-38°C throughout the experiment. After placement of a venous catheter (tail vein) and cannulation of the trachea, the animal was artificially ventilated with oxygenated room air, and subsequent anesthetic was given as an infusion of 10 to 20 mg/kg/hr (i.v.). Gallamine triethiodide (4 mg/kg/hr i.v.) was also given for paralysis. A femoral artery was cannulated for arterial blood pressure monitoring. Depth of anesthesia was assessed by monitoring the stability of arterial pressure and HR response to pinch of the hindpaw and was adjusted by appropriate changes in the infusion rate. ADNs were isolated bilaterally and marked with small pieces of black suture. After all surgery procedures were performed, the rat was placed in a stereotaxic head frame, and an occipital craniotomy was performed to expose the dorsal surface of medulla in the region of the obex. The ADN ipsilateral to the central recording site was mounted on bipolar stimulating electrodes.

Electrophysiological recordings. Extracellular action potential discharge was recorded with a five-barrel electrode. The recording barrel was filled with a solution of 0.5 M sodium acetate containing 2% Chicago sky blue (impedance 8-30 M). One side barrel of each five-barrel electrode was filled with a solution of 3 M NaCl and was used for current balancing. The remaining side barrels were filled with different drug solutions. All recordings were performed in a block of tissue 1.2 mm caudal and 0.5 mm rostral to the calamus scriptorius, 0 to 0.8 mm lateral to the midline and 0.2 to 1 mm below the surface. This region of the NTS was reported to be the primary site of termination of anterograde labeled aortic baroreceptor fibers (Mendelowitz et al., 1992). The electrode was lowered into the tissue in 2.0 to 2.5-µm steps by a stepdriver controller (Burleigh Instrument Inc., Fishers, NY). The ADN was stimulated with a single pulse (1 msec in duration, 0.5 Hz and 500 µA). Recorded action potentials were amplified by a DC amplifier (World Precision Instrument, New Haven, CT), passed through an AC filter and then sent to a digital oscilloscope (Nicolet Instrument Co., Madison, WI), an audiomonitor (Grass Instrument Co., Quincy, MA), a video tape recorder (A.R. Vetter, Co., Reberburg, PA) and a window discriminator (World Precision Instrument, Sarasota, FL). The window discriminator output was led to a Cambridge Electronic Design (CED1401, Cambridge, England) analog-to-digital converter interfaced with a PC. Spike2 or Sigavg data acquisition software was used for on-line and off-line analyses. After an ADN-evoked NTS neuron was found, MSNs or PSNs were characterized using previously described criteria (Scheuer et al., 1996). The main criterion used was the ability of the ADN-evoked responses of MSNs to follow two stimuli separated by 5 msec. PST histograms (duration 0.5 sec; bin width 1 msec; 40 sweeps) and rate meter histograms (duration 180 sec; bin width 1 sec) were collected to analyze evoked and spontaneous discharge, respectively. Arterial pressure was measured using a Cobe CDX transducer (Cobe Laboratories, Lakewood, CO). Mean arterial pressure and HR were determined from the pulsatile signal using a Coulburn blood pressure processor.

Drugs and drug administration. After base-line discharge (spontaneous firing rate and/or evoked responses) was recorded, drugs were administered by application of microiontophoretic ejecting currents to the drug-containing barrels. For current-response curves, drugs were ejected for successive 10-sec periods separated by 10-sec intervals. The drug solutions for microiontophoresis were L-glutamic acid (monosodium salt, 100 mM, Sigma Chemical Co., St. Louis, MO), NMDA (N-methyl-D-aspartate) (100 mM, Sigma), AMPA [(RS)-alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid] (10 mM, Tocris Neuramin, Bristol, England), kainic acid (10 mM, Sigma) and trans-(1S,3R)-ACPD (trans-(1S, 3R)-1-amino-1,3-cyclopentanedicarboxylic acid) (100 mM, Research Biochemicals International, Natick, MA). All drugs were dissolved in 150 mM saline, and pH was adjusted to 8.0 to 8.5. All drugs were ejected as anions. Retaining currents of appropriate polarity were applied to the drug barrels to retrain the passive diffusion of the drug from the electrode tip during nonejection periods.

Data analysis. Data were analyzed with MANOVA (ANOVA) or ANCOVA with a repeated-measures design (StatSoft, Tulsa, OK). Newman-Keuls test was used for post-hoc comparisons. PSTs were used to analyze evoked discharge. Ratemeter records (spikes/sec) were used for statistical analysis of spontaneous discharge rate. To account for different levels of basal spontaneous discharge, the discharge rate increase (equal to the discharge rate during the drug application minus base-line discharge rate) was used for analysis. Significance was accepted at P < .05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Electrophysiological characterization of NTS neurons. Data were obtained from 169 NTS neurons: 51 that were evoked by stimulation of the ADN (ENs) and 118 that were not evoked by stimulation of the ADN (NENs). These numbers do not represent the relative percentages of ENs and NENs in NTS, because not every NEN encountered was studied. The vast majority of NTS neurons (80%-90%) were NENs. Of the 51 ENs, 25 were characterized as receiving a monosynaptic input from ADN afferents (MSNs), and 26 were characterized as receiving polysynaptic ADN inputs (PSNs). An example of an MSN and an example of a PSN are shown in figure 1. The average onset latency for MSNs (10.8 ± 1.9 msec, mean ± S.E.M., range 3-27 msec) was significantly (P < .001) shorter than that of PSNs (25.5 ± 1.7 msec, range 8-37 msec). Twenty (20/25) MSNs and 23 (23/26) PSNs were spontaneously active. The average firing rate for MSNs was 1.5 ± 0.4 Hz (range 0.1-29.1 Hz) and was not different (P = .20) from that for PSNs (2.2 ± 0.4 Hz, range 0.2-6.5 Hz). In addition, 101 of 118 NENs were also spontaneously active, with an average firing rate of 2.8 ± 0.3 Hz (range 0.1-19.2 Hz).

View larger version (26K): [in this window] [in a new window]   Fig. 1.   . Typical monosynaptic (panel A) and polysynaptic (panel B) NTS neurons receiving ADN inputs. The sweeps for each neuron are arranged similarly, with the top sweep illustrating action potentials. The arrows indicate a clear A-B break between an initial segment and a larger somato-dendritic spike. The next sweep downward illustrates paired-pulse. ADN stimulation (5-msec interval, stimulus artifacts indicated by *) In the sweep in panel A, two potentials were evoked (the second potential was only an initial segment spike), which indicates that this is a monosynaptic neuron. In panel B, one action potential was evoked, which indicates that this is a polysynaptic neuron. Immediately below, five responses evoked by ADN stimulation are displayed to illustrate the very consistent onset latencies between the stimulus artifacts and monosynaptic evoked responses (panel A) and the variable onset latencies for polysynaptic evoked responses (panel B). The bottom panels illustrate, using a peristimulus-time histogram of 40 evoked responses, the consistent onset latencies of a monosynaptic input (panel A) and the variables onset latencies for polysynaptic evoked responses (panel B).

Effects of glutamate on MSNs, PSNs and NENs. Microiontophoretic application of glutamate excited 5 of 9 MSNs (firing rate increase 2.3-11.6 Hz at 40-nA ejection current), 6 of 8 PSNs (firing rate increase 2.6-31.9 Hz at 40-nA ejection current) and 16 of 20 NENs (firing rate increase 1.1-69.0 Hz at 40-nA ejection current) (P < .01 for each group). Examples are shown in figures 2A, 3A and 4A. Six ENs (4 MSNs and 2 PSNs) and 4 NENs were not very sensitive to glutamate even at 40-nA ejecting currents and responded with a firing rate increase of less than 1 Hz. Examples are shown in figures 2B, 3B and 4B. The drug barrel of each electrode that recorded glutamate-insensitive NTS neurons was carefully checked to verify that this observation was not artifactual. First, during the recording and drug application, the iontophoretic circuitry indicated that neither drug nor balance barrel was blocked. Second, after the recording, we verified that a similar application of glutamate using the same barrel and the same ejecting currents could include excitation of a second NTS neuron recorded in the same electrode track. Two ENs (1 MSN and 1 PSN) and 1 NEN were supersensitive to glutamate, and these cells were driven into depolarization inactivation during 40 nA of glutamate ejection. Because more MSNs were apparently insensitive to glutamate, the dose-response curve of MSNs to glutamate was compared with those of PSNs and NENs (fig. 5A). There was no statistically significant difference among these three groups (P = .44 for between-groups comparison; P = .99 for group-dose interactions).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Glutamate-sensitive and -insensitive MSNs. A) A glutamate-sensitive MSN. Top panel: Five ADN stimuli evoked five responses with consistent latencies. The monosynaptic property of this neuron is indicated by double action potentials evoked by paired-pulse ADN stimuli. Middle panel: Glutamate (5-40 nA, two 40-nA ejections) and AMPA (5 nA) excited this neuron. Bottom panel: Arterial blood pressure. B) A glutamate-insensitive MSN. Top panel: Consistent evoked responses by ADN stimuli and monosynaptic property of this neuron identified by double-pulse ADN stimuli. Middle panel: This neuron was not sensitive to either glutamate (5-40 nA) or trans-ACPD (5-40 nA). Bottom panel: Arterial blood pressure.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Examples of different responses of ENs to EAA agonists. A) Different responses of a PSN to glutamate and trans-ACPD. Top panel: Cell identification indicates that this cell was a PSN. Middle panel: This neuron was very sensitive to glutamate (5-10 nA) but not to trans-ACPD (5-40 nA). Bottom panel: Arterial blood pressure. B) Different responses of a MSN to glutamate and kainate. Top panel: Cell identification. Middle Panel: This cell was not very sensitive to glutamate (5-40 nA) but was slightly sensitive to kainate. Bottom panel: Arterial blood pressure.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Responses of NENs to EAA agonists. A) NEN sensitive to glutamate and trans-ACPD. B) NEN insensitive to both glutamate and trans-ACPD. C) NEN sensitive to kainate but not to trans-ACPD. D) NEN sensitive to trans-ACPD but not to kainate.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Comparisons of dose-response relationships for different groups of NTS neurons to EAA agonists. Glutamate (A) and AMPA (B) excited all three groups of NTS neurons (P > .4 for each comparison). In contrast, NMDA was more potent on PSNs and NENs than on MSNs (panel C, P < .01 for each comparison), and trans-ACPD tended to be more potent on MSNs than on PSNs or NENs (panel D, P = .09 and .07).

Effects of AMPA and kainate on NTS neurons. AMPA and kainate were very potent excitants of ENs and NENs. AMPA stimulated 5 of 6 MSNs (2.8-44.3-Hz increases at 40-nA ejection current), all 7 PSNs (2.0-16.8-Hz increases at 40-nA ejection current) and 26 of 27 NENs (2.0-76.0-Hz at 40-nA ejection) currents. Examples are shown in figures 6A and 7B. Two ENs (2 PSNs) and 12 NENs were driven into depolarization inactivation by AMPA at 20-nA and 40-nA ejection currents. Microiontophoretic kainate excited 3 to 5 MSNs (4.9-12.1-Hz increases at 40 nA), all 3 PSNs (2.3-18.3-Hz increases at 40 nA) and 25 of 28 NENs (1.5-27.4-Hz at 40 nA). Examples are shown in figures 6B, 3B and 4C. Two ENs (1 MSN and 1 PSN) and 9 NENs were also very sensitive to kainate and were stimulated into depolarization inactivation at 20-nA and 40-nA current ejections. Some NTS neurons were resistant to AMPA or kainate, as illustrated in figures 7A and 4D. The drug barrels of each electrode that recorded insensitive NTS neurons were carefully checked as described previously. One MSN was omitted in the kainate group because this cell exhibited a huge firing rate increase (fig. 6B) that biased the statistical analysis. Statistical analysis did not reveal any significant difference in the potencies of AMPA or kainate on MSNs, PSNs and NENs (P > .05 for each comparison). The dose-response curves of these three groups of NTS neurons to AMPA are shown in figure 5B.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Examples of the responses of MSNs to EAA agonists. A) MSN with different responses to NMDA and AMPA. Top panel: The consistent evoked responses of this neuron to ADN stimuli. The monosynaptic property of this neuron was identified by double action potentials evoked by paired-pulse ADN stimuli. Middle panel: This neuron was not sensitive to NMDA (5-40 nA) but was very sensitive to AMPA (5-20 nA). Bottom panel: Arterial blood pressure. B) MSN sensitive to both trans-ACPD and kainate. Top panel: MSN was identified by its consistent responses to ADN stimuli and double action potentials evoked by paired-pulse ADN stimuli. Middle panel: This neuron was sensitive to trans-ACPD (5-40 nA) and kainate (5-10 nA). Bottom panel: Arterial blood pressure.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Examples of the responses of PSNs to EAA agonists. A) PSN sensitive to NMDA but insensitive to kainate. Top panel: Consistent evoked responses by ADN stimulation. The polysynaptic property was identified by only one evoked action potential in response to paired-pulse ADN stimuli. Middle panel: This neuron was very sensitive to NMDA (5-40 nA), but not to kainate (5-40 nA). Bottom panel: Arterial blood pressure. B) PSN sensitive to NMDA and AMPA. Top panel: Cell identification for polysynaptic NTS neuron. Middle panel: This cell responded to NMDA (5-40 nA) and AMPA (5-40 nA). Bottom panel: Arterial blood pressure.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Comparisons of the dose-responses of NTS neurons to different EAA agonists. A) Comparisons of the dose-responses of ENs to different EAA agonists. All EAA agonists tested excited ENs (P < .001 for each group) with similar potencies (P = .54 for between groups, P = .65 for group-dose interactions). B) Comparisons of dose-responses of NENs to different EAA agonists. All EAA agonists tested excited NENs (P < .01 for each group), and trans-ACPD was less potent than all other EAA agonists (P < .05 for each comparison).

Effects of NMDA on NTS neurons. In contrast to AMPA and kainate, NMDA weakly excited 4 of 9 MSNs (1.2-6.6-Hz increases at 40-nA ejection) and strongly excited 7 of 8 PSNs (6.2-26.7-Hz increases at 40-nA ejection) and 26 of 29 NENs (2.3-36.2-Hz increases at 40-nA ejection). Examples are shown in figures 6A, 7A and B. Only 4 NENs were driven into depolarization inactivation by NMDA at 40-nA ejection. Some NMDA-insensitive MSNs responded to other EAA agonists (fig. 6A), so the apparent lack of sensitivity of MSNs to NMDA was not due to a sampling problem. In contrast, some PSNs were apparently more sensitive to NMDA than to other EAA agonists (fig. 7A). Statistical analysis indicated that NMDA was more potent on PSNs and on NENs than on MSNs (P < .01 for each comparison). As shown in figure 5C, the dose-response curve of MSNs to NMDA was shifted downward compared to those of PSNs and NENs.

Effects of trans-ACPD on NTS neurons. Microiontophoretic trans-ACPD excited 4 of 5 MSNs (firing rate increases ranged from 1.5 Hz to 24.9 Hz at 40-nA ejection current) and weakly excited all 4 PSNs (1.2-8.4-Hz increases at 40-nA ejection current) and 23 of 26 NENs (2.6-16.7-Hz increases at 40-nA ejection current). Examples are shown in figures 2B, 3A, 4A-D and 6B. One EN (1 MSN) and 1 NEN exhibited depolarization inactivation during iontophoresis of trans-ACPD. Statistical analysis suggested that trans-ACPD was more potent on MSNs than on PSNs or NENs (p = .09 and 0.07 respectively). If one trans-ACPD-insensitive MSN (Note: this neuron was also not sensitive to glutamate; fig. 2B) and 3 insensitive NENs (fig. 4B and C) are omitted then trans-ACPD was significantly more potent on MSNs than on PSNs or NENs (P = .012 and P = .010, respectively). The dose-response curves of NTS neurons to trans-ACPD are shown in figure 5D. The dose-response curve of MSNs to trans-ACPD was clearly shifted upward compared with those of PSNs or NENs.

Comparison of different EAA agonists on ENs and NENs. Ignoring differences between MSNs and PSNs, ENs were excited by all EAAs (P < .001 for each group) and showed similar responses to all agonists (P = .54 for between-groups and P = .65 for group-dose interactions) (fig. 8A). In contrast, NENs were apparently less sensitive to trans-ACPD than to other EAA agonists (P < .05 for each comparison) (fig. 8B). When two or more drugs were applied to the same NTS neuron, some NTS neurons showed great individual differences in responding to different EAA agonists. Many neurons exhibited a greater response to one particular EAA agonist than to another, whereas other neurons appeared to respond equally to different agonists. These different patterns of responses are shown in figures 3, 4, 6 and 7.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The responses of NTS neurons to various EAA agonists were tested in the anesthetized rat. Some of these NTS neurons were identified as baroreflex-related NTS neurons by their responses to the ADN stimulation. We also determined whether these baroreflex-related NTS neurons received monosynaptic or polysynaptic inputs from the ADN by using criteria initially described in an in vitro study (Miles, 1986) and subsequently used in in vivo studies (Rogers et al., 1993; Scheuer et al., 1996). Some of the NTS neurons that did not respond to ADN stimulation (NENs) could also be baroreflex-related neurons. For example, they may receive baroreceptor afferents from nerves other than ADN, such as the carotid sinus or vagus nerves. The functional role of these "unidentified neurons," however, is not clear in the present study.

All EAA agonists tested in the present study excited some NTS neurons that received baroreceptor afferent inputs. EAA-induced excitation was not limited to baroreflex-related NTS neurons. All three groups of NTS neuronsMSNs, PSNs and NENsresponded similarly to the non-NMDA receptor agonists AMPA and kainate. These results indicate that EAAs are involved in the integration of other reflexes in NTS and that AMPA and/or kainate receptors may play a ubiquitous role in neurotransmission and integration in NTS, as suggested by Zhang and Mifflin (1995).

The iontophoretic approach was presented as a means to bridge the gap between the limitations inherent in in vivo microinjection studies and in vitro cellular studies. However, iontophoretic studies also have limitations. Our results make it possible to compare the effects of the same EAA agonist on different types of NTS neurons. For example, one can compare the effects of AMPA on MSNs and PSNs because the drug concentration, ejecting currents and parameters of the recording electrode are always the same. However, it is not appropriate to make a quantitative comparison of the potency of different EAA agonists (such as AMPA vs. NMDA) on the same type of NTS neuron. This is because the concentrations of the drugs used in the present study were different. The transport numbers of glutamate, NMDA, AMPA and kainate are similar (Stone, 1985). Several indirect observations, however, suggest that AMPA and kainate may be more potent on NTS neurons than other EAA agonists. First, the concentrations used for AMPA and kainate (10 mM) to generate similar excitatory influences on NTS neurons in the present study were an order of magnitude lower than those used for other EAA agonists (100 mM). Second, inactivation of depolarization was observed in more NTS neurons during AMPA or kainate application than during the application of other EAA agonists at even lower concentrations. Third, when two or more EAAs were applied to the same NTS neuron, AMPA or kainate excited some NTS neurons that were not sensitive to glutamate (fig. 3B). Similar findings have also been reported in microinjection studies (Talman et al., 1980; Leone and Gordon, 1989). These investigators found that lower doses of kainate than of glutamate or NMDA were needed to induce cardiovascular changes when microinjected into NTS. Because information about the binding profiles for EAA receptor subtypes in NTS is currently lacking [there is only one report about NMDA receptor binding in NTS (Monaghan and Cotman, 1985)], the basis of these apparently different potencies of EAA agonists on NTS neurons is currently not known.

A surprising finding in the present study was that some NTS neurons, including some baroreflex-related NTS neurons, were insensitive to EAA agonists, even the non-NMDA agonists. This was not a technical problem arising from the microiontophoresis technique, because electrodes that recorded insensitive NTS neurons were carefully checked. We frequently found that in the same rat, even when the same electrode and the same ejecting currents were used, some NTS neurons were very sensitive to EAA agonists and some were not. In most other brain regions (for example, in substantia nigra), microiontophoretic application of EAA agonists excites all cells (Zhang et al., 1992, 1994).

One explanation for the insensitivity of these NTS neurons to EAAs could be that the recorded action potentials in NTS were from axons (there are no EAA receptors on the axons) instead of from the somatodendritic area of the neurons. For the following four reasons, we do not feel that this is a reasonable explanation. 1) The action potential waveform of the recorded NTS neurons had an A-B break (fig. 1). which is characteristic of somatic recordings (Lipski, 1981). 2) The tips of the multi-barrel electrodes used for recording and microiontophoresis were larger (2-2.5 µM) than those used in our single-barrel electrode (< 1 µM) experiments. The amplitude of electrical signals collected from axons using electrodes with such large tips would expected to be very small. 3) In other experiments using similar methods, microiontophoretic application of GABA agonists, such as GABA and muscimol, inhibited all NTS neurons (including baroreflex-related NTS neurons) (Mifflin and Zhang, 1996). When both GABA and glutamate (or other EAAs) were applied onto the same NTS neuron, all neurons were sensitive to GABA, but only some responded to glutamate (Zhang, J. and Mifflin, S.W. unpublished data). 3) In vitro studies (Drewe et al., 1990; Drewe and Kunze, 1994) have reported that only about 80% of NTS neurons were sensitive to EAA agonists when these EAAs were applied by perfusion to acutely dissociated neurons.

A second possible explanation for the insensitivity of NTS neurons to EAAs could be that EAA receptors are mainly located on remote dendrites, not on the somatodendritic area of NTS neurons, and our iontophoretically applied EAAs did not reach these remote areas. Both the soma and the proximal dendrites of NTS neurons have been reported to receive synaptic inputs from ADN afferent fiber terminals (Mendelowitz et al., 1992). However, though it has not been reported, an uneven distribution of EAA receptors on the soma and dendrites of NTS neurons may exist. Finally, the lack of response to EAA agonists could be due to the fact that these NTS neurons do not possess EAA receptors in the somatodendritic area. The insensitivity of these NTS neurons to EAAs suggests that in these neurons, ADN-evoked responses are mediated by other neurotransmitters.

The third finding in present study is that the different NTS neurons responded to selective EAA agonists differently. Two of five EAA agonists tested in the present study were found to have different potencies on different classes of NTS neurons. NMDA was more potent on PSNs and NENs than on MSNs. Some MSNs were totally insensitive to NMDA even though they were very sensitive to other EAA agonists (fig. 6A). These results agree with previous in vitro findings that NMDA receptors are probably not involved in the integration of monosynaptic tractus inputs to NTS neurons (Andresen and Yang, 1990; Drewe and Kunze, 1994). The endogenous EAAs found in NTS are considered to be glutamate and L-aspartate (Dietrich et al., 1982; Granata et al., 1984; Kubo and Kihara, 1988), and neither is selective for any EAA receptor subtype. Therefore, the insensitivity of MSNs to NMDA receptor agonists could explain the observation that evoked responses to some MSNs were blocked by non-NMDA antagonists and not by NMDA antagonists (Andresen and Yang, 1990; Zhang and Mifflin, 1996).

In contrast to ionotropic EAA receptors, metabotropic receptors are G protein-coupled receptors and indirectly regulate electrical signaling and activate various second messenger systems (for reviews, see Schoepp, 1994; Riedel, 1996). Activation of metabotropic receptors can produce postsynaptic and/or presynaptic influences on neurons. In addition, metabotropic receptors can modulate the EAA-induced excitation (via other EAA receptor subtypes) of some neurons. Perfusion of brain slices with the metabotropic receptor agonist trans-ACPD depolarized NTS neurons, reduced their membrane conductance, blocked excitatory or inhibitory synaptic activity induced by tractus stimulation and potentiated AMPA-induced inward currents (Glaum and Miller, 1992, 1994; Glaum et al., 1993). Microinjection studies suggest that metabotropic receptors may play an important role in glutamate-induced cardiovascular changes (Talman, 1989; Leone and Gordon, 1989). In the present study, we found that trans-ACPD predominantly influenced MSNs, which suggests that the primary role of metabotropic receptors in the NTS might be to modulate the integration of baroreceptor inputs during their initial processing by MSNs.

Summary and conclusions. Our results indicate the following: 1) Not every NTS neuron receiving an ADN input is sensitive to EAA agonists, which suggests that EAAs may not be the only transmitter used in baroreceptor afferent integration. 2) EAAs are probably not limited to the neuronal transmission and integration of baroreceptor afferent inputs in NTS. AMPA and/or kainate receptors may play a general role in neurotransmission and integration in NTS, as originally suggested by Zhang and Mifflin (1995). 3) NMDA receptors are likely to play a role in baroreceptor afferent integration at the level of PSNs, not MSNs. 4) Metabotropic receptor agonists appear to be more potent on MSNs than on PSNs or NENs, which suggests that metabotropic receptors may play a special role in the neuronal transmission and integration of the baroreflex.