Download NMDA Receptors Contribute to Primary Visceral Afferent

NMDA Receptors Contribute to Primary Visceral Afferent
Transmission in the Nucleus of the Solitary Tract
MARIA LUZ AYLWIN, 1 JOHN M. HOROWITZ, 2 AND ANN C. BONHAM 1
1
Department of Internal Medicine, Division of Cardiovascular Medicine and 2 Department of Neurobiology, Physiology
and Behavior, University of California, Davis, California 95616
INTRODUCTION
The nucleus of the solitary tract (NTS) is the first central
site where the reflex control of autonomic function is coordinated. The nucleus receives visceral afferent information
from sensory endings in the large blood vessels, heart, lungs,
and gastrointestinal organs; then, through primary reflex circuitries within the brain stem or through more elaborate
interconnections with higher brain regions, the sensory information is conditioned and ultimately transformed to regulate
autonomic output. Primary afferent fibers from the target
organs course in the solitary tract of the nucleus and then
exit to synapse onto second-order neurons within the nucleus. The fibers terminate, to a large extent viscerotopically;
cardiovascular, respiratory, and gastrointestinal afferent fibers terminate mainly in the intermediate and caudal NTS,
whereas gustatory afferent fibers terminate in the rostral NTS
(Loewy 1990).
Glutamate is the primary excitatory neurotransmitter released by the visceral afferent fibers in the intermediate and
caudal NTS (Talman et al. 1980). It is well established that
the non-N-methyl-D-aspartate [non-NMDA: a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and kainate] receptors, which mediate the fast component of glutamate signaling, are activated by visceral afferent transmission to NTS neurons (Andresen and Yang 1990; Brooks and
Spyer 1993; Brooks et al. 1992; Glaum and Miller 1993,
1995). Whether the NMDA receptors, which mediate a
slower-developing, longer-lasting component of glutamate
signaling, are also activated is still in question. If only the
non-NMDA receptors transmit primary sensory afferent signals to second-order neurons in the NTS, then the secondorder neurons most likely serve to simply relay information
from visceral sensory endings to higher-order neurons, as
has been classically described (Spyer 1981). If, on the other
hand, NMDA receptors are also activated by the primary
visceral afferent input, then the second-order neurons can
sustain a longer-lasting depolarization that allows for added
signal conditioning capabilities such as the temporal integration of multiple inputs, rhythmic firing, and synaptic plasticity (Cotman and Iversen 1987). Unlike the non-NMDA receptors, the activation of NMDA receptors requires not only
the presence of glutamate but also depolarization of the cell
because of the Mg 2/ block of the NMDA receptor channel
at resting membrane potentials (Nowak et al. 1984). Consequently, NMDA-receptor-mediated synaptic responses are
difficult to detect experimentally at resting membrane potentials, which may contribute to the uncertainty regarding the
participation of NMDA receptors to primary visceral afferent
transmission in the NTS.
Various experimental strategies have been used to investigate the role of non-NMDA and NMDA receptors in visceral
afferent transmission in the NTS. In whole animals with
0022-3077/97 $5.00 Copyright q 1997 The American Physiological Society
/ 9k11$$my22 J722-6
08-08-97 12:40:46
neupa
LP-Neurophys
2539
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
Aylwin, Maria Luz, John M. Horowitz, and Ann C. Bonham.
NMDA receptors contribute to primary visceral afferent transmission in the nucleus of the solitary tract. J. Neurophysiol. 77: 2539–
2548, 1997. The nucleus of the solitary tract (NTS) is a principal
site for coordinating the reflex control of autonomic function. The
nucleus receives and organizes primary visceral (sensory) afferent
inputs from the great vessels, heart, lung, and gastrointestinal organs. Glutamate, the excitatory neurotransmitter released by the
primary afferent fibers, activates non-N-methyl-D-aspartate (nonNMDA) receptors on second-order neurons in the NTS. Still in
question is whether NMDA receptors on the second-order neurons
are also activated. Accordingly, the purpose of this study was to
directly determine whether NMDA receptors contribute to synaptic
transmission of primary visceral afferent input to second-order
neurons in the NTS. Whole cell patch-clamp recordings were obtained from intermediate and caudal NTS neurons in rat coronal
medullary slices. Excitatory postsynaptic currents (EPSCs) were
evoked by stimulation of the solitary tract (1–25 V, 0.1 ms, 0.2
or 0.5 Hz) at membrane potentials ranging from 090 to /60 mV.
In 28 of 32 neurons in which current-voltage relationships were
obtained for solitary-tract-evoked EPSCs, the currents had short
onset latencies (3.42 { 1.03 ms, mean { SD), indicating that
they were the result of monosynaptic activation of second-order
neurons. Solitary-tract-evoked EPSCs had both a fast and a slow
component. The amplitude of the slow component was nonlinearly
related to voltage (being revealed only at membrane potentials
positive to 045 mV), blocked by the NMDA receptor antagonist
DL-2-amino-5-monophosphovaleric acid (APV, 50 mM; n Å 12;
P Å 0.0001), and enhanced in nominally Mg 2/ -free perfusate at
membrane potentials negative to 045 mV (n Å 5; P Å 0.016),
demonstrating that the slow component was mediated by NMDA
receptors. The amplitude of the fast component was linearly related
to voltage and blocked by the non-NMDA receptor antagonist 2,3dihydroxy-6-nitro-7-sulfamoylbenzo(F)quinoxaline (NBQX, 3
mM; n Å 9; P Å 0.0014), demonstrating that the fast component
was mediated by non-NMDA receptors. The slow component of
the EPSCs was not blocked by NBQX (n Å 6; P Å 0.134), nor
was the fast component blocked by APV (n Å 12; P Å 0.124).
These results show that both NMDA and non-NMDA receptors
coexist on the same second-order NTS neurons and mediate primary visceral afferent transmission in the NTS. The participation
of NMDA receptors suggests that second-order neurons in the NTS
may have previously unrecognized integrative capabilities in the
reflex control of autonomic function.
2540
M. L. AYLWIN, J. M. HOROWITZ, AND A. C. BONHAM
/ 9k11$$my22 J722-6
not contribute to primary visceral afferent transmission to
second-order neurons in the NTS (Andresen and Yang
1990). However, because the recordings were made in neurons at their resting membrane potentials, the Mg 2/ block
of the NMDA receptor channel, maximal under these conditions, may have obscured NMDA-receptor-mediated responses. Indeed, NMDA receptor contribution to synaptic
transmission may be missed altogether unless the NMDA
currents are measured at membrane potentials positive to
045 mV. In the present study, to directly determine whether
NMDA receptors contribute to primary visceral afferent
transmission in the intermediate and caudal NTS, we used
voltage-clamp recording to examine excitatory postsynaptic
currents (EPSCs) monosynaptically evoked by solitary tract
stimulation at a range of membrane potentials from 090 to
/60 mV in the medullary slice.
METHODS
Experimental protocols followed in this work were reviewed
and approved by the Institutional Animal Care and Use Committee
in compliance with the Animal Welfare Act and in accordance
with Public Health Service Policy on Humane Care and Use of
Laboratory Animals.
SLICE PREPARATION. Male Sprague-Dawley rats 3–4 wk old
(60–120 g) were anesthetized with a combination of ketamine (35
mg/kg) and xylazine (2 mg/kg) given intramuscularly and were
decapitated with a guillotine. The brain was rapidly exposed and
submerged in ice-cold ( õ47C) high-sucrose artificial cerebrospinal
fluid that contained (in mM) 3 KCl, 2 MgCl2 , 1.25 NaH2PO4 , 26
NaHCO3 , 10 glucose, 220 sucrose, and 2 CaCl2 , pH 7.4, osmolality
300 m0sM, constantly bubbled with 95% O2-5% CO2 (Parkis et
al. 1995). After the brain stem was dissected, its caudal surface was
cemented to the stage of a Vibratome 1000 (Technical Products
International, St. Louis, MO) with cyanoacrylate glue, and its ventral surface was glued to an agar block. Coronal slices (300 mm
thick) were obtained between 300 mm rostral and 1,200 mm caudal
to the obex. Four slices were typically obtained from each brain
stem; one slice caudal to the area postrema, two slices containing
the area postrema, and one slice rostral to the area postrema. We
used longitudinal slices (250–300 mm) in a few experiments to
demonstrate that the findings in the coronal slices could be confirmed in this preparation, which has been used in other laboratories
(Andresen and Yang 1990). The slices were incubated for 45 min
at 377C in a holding chamber filled with the high-sucrose artificial
cerebrospinal fluid constantly bubbled with 95% O2-5% CO2 . The
slices were then transferred to a chamber filled with artificial cerebrospinal fluid (normal perfusate) containing (in mM) 125 NaCl,
2.5 KCl, 1 MgCl2 , 1.25 NaH2PO4 , 25 NaHCO3 , 10 glucose, and
2 CaCl2 , pH 7.4, osmolality 300 m0sM, continuously bubbled with
95% O2-5% CO2 and held at room temperature (227C). Slices were
incubated in the normal perfusate for ¢1 h before whole cell
recordings were begun, at which time a slice was transferred to
the recording chamber, held in place with a nylon mesh, and continually perfused with the normal perfusate at a rate of 3 ml/min at
room temperature. All the experiments were performed at 227C.
WHOLE CELL PATCH-CLAMP RECORDING. Borosilicate glass
electrodes were filled with a CsF solution containing (in mM)
145 CsF, 5 NaCl, 1 MgCl2 , 3 K-ATP, 0.2 sodium guanosine 5 *triphosphate (Na-GTP), 10 ethylene glycol-bis( b-aminoethyl
ether)-N,N,N *,N *-tetraacetic acid (EGTA), and 10 N-2-hydroxyethylpiperazine-N *-2-ethanesulfonic acid (HEPES), pH 7.4, 300
mosM, for voltage-clamp experiments, or with a KCl solution containing (in mM) 140 KCl, 5 NaCl, 1 MgCl2 , 3 K-ATP, 0.2 NaGTP, 10 EGTA, and 10 HEPES, pH 7.4, 300 m0sM, for current-
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
intact autonomic reflexes, microinjections of non-NMDA
and NMDA receptor agonists and antagonists in the NTS
have been utilized to mimic or block, respectively, the reflex
responses evoked by stimulation of the visceral afferent fibers. Such studies have variably implicated the participation
of either or both the non-NMDA and NMDA receptors on
NTS neurons in several autonomic reflex pathways: the baroreceptor (Gordon and Leone 1991; Kubo and Kihara 1988;
Ohta and Talman 1994), cardiopulmonary C fiber receptor
(Vardhan et al. 1993), Breuer-Hering (Bonham et al. 1993),
and superior laryngeal nerve inspiratory shortening reflex
pathways (Karius et al. 1994). These variable findings regarding the role of non-NMDA and NMDA receptors in
synaptic transmission in the NTS may be due to true differences in the glutamate receptor subtypes activated in these
different autonomic reflex pathways or to limitations of the
microinjection technique, which typically relies on relatively
large volumes (10–100 nl) of highly concentrated agents
that very likely affect many NTS neurons. Thus although
microinjection studies provide crucial data regarding the
functional significance of non-NMDA and NMDA receptors
on NTS neurons for full expression of various autonomic
reflex responses, they cannot determine with certainty the
extent to which the receptors are located on second-order,
higher-order, or other neurons not involved in afferent transmission. This limits the assessment of the relative contribution of the NMDA versus non-NMDA receptors to primary
visceral afferent transmission in the NTS.
Extracellular recordings of NTS unit activity in the whole
animal have more directly examined the role of non-NMDA
and NMDA receptors in synaptic transmission between functionally identified visceral afferent fibers and NTS neurons.
In a recent report, Zhang and Mifflin (1996), using iontophoretic application of non-NMDA and NMDA receptor antagonists, concluded that non-NMDA receptors transmit primary input from baroreceptor afferent fibers to second-order
neurons, but that NMDA receptors transmit the information
to higher-order NTS neurons. We similarly concluded that
non-NMDA receptors largely mediate activation of NTS
neurons by cardiopulmonary C fiber afferent input (Wilson
et al. 1996). Together the data suggest that baroreceptor
and cardiopulmonary C fiber afferent input is transmitted to
second-order NTS neurons in the intermediate and caudal
NTS largely by non-NMDA receptors. However, because the
studies were performed on neurons at their resting membrane
potentials and because extracellular Mg 2/ could not be depleted to remove the voltage-dependent block of the NMDA
channel, the contribution of NMDA receptors to the synaptic
transmission may not have been detected.
In in vitro studies in the medullary slice, Andresen and
Yang (1990) reported that maximal selective blockade of
non-NMDA receptors decreased the amplitude of short-latency excitatory postsynaptic potentials (EPSPs) evoked by
stimulation of the solitary tract by ú85%, whereas NMDA
receptor blockade had relatively little effect, decreasing the
amplitude by õ20%. Brooks et al. (1992) similarly demonstrated that non-NMDA receptor antagonism abolished solitary-tract-evoked EPSPs, although whether the recordings
were made from second- or higher-order neurons was not
determined. These in vitro findings considered along with the
in vivo data have led to the proposal that NMDA receptors do
NMDA RECEPTORS IN PRIMARY VISCERAL AFFERENT TRANSMISSION
/ 9k11$$my22 J722-6
voltage. Currents obtained at voltages negative to 070 mV in
nominally Mg 2/ -free perfusate were fitted with two exponentials, with time constants and amplitudes determined with the
use of the pClamp6 simplex method. We used these same time
constants to fit the curves obtained in the normal-Mg 2/ perfusate
and then determined the amplitude of the two exponentials. Data
are expressed as means { SE unless otherwise indicated. The
EPSCs in the control conditions and after interventions were
compared by the use of the paired t-test. Differences were considered significant at P õ 0.05.
CHEMICALS. NaCl, KCl, MgCl2 , sucrose, and CdCl2 were purchased from Fisher; NaHCO3 , K-ATP, Na-GTP, EGTA, HEPES,
bicuculline, picrotoxin, and APV from Sigma; NaH2PO4 and CaCl2
from Mallinkdrot; glucose from Baker; CsF from Aldrich; and
NBQX from RBI.
RESULTS
NTS sites of whole cell patch-clamp recordings
Whole cell patch-clamp recordings were obtained from
177 NTS neurons, 126 of which received excitatory input
from the solitary tract. In all of the four cells tested, the
EPSCs evoked by solitary tract stimulation were blocked by
the calcium channel blocker CdCl2 (100 mM), indicating
that the currents evoked by solitary tract stimulation were
synaptically mediated. A photomicrograph of a medullary
slice in the recording chamber is shown in Fig. 1A, bottom.
The positions of the stimulating electrode in the solitary
tract and the recording electrode just medial to the tract are
indicated in Fig. 1A, inset. The locations of the recording
sites are shown in three schematic representations of coronal
slices in Fig. 1B. All recordings were made just medial to
the solitary tract and between 300 mm rostral and 1,200 mm
caudal to the obex. The majority of recordings was made in
the intermediate NTS at the level of the area postrema,
whereas fewer were made rostral or caudal to the area postrema.
Onset latencies and voltage dependence of solitary-tractevoked EPSCs
EPSCs evoked by single stimuli delivered to the solitary
tract were recorded in 32 neurons at a series of voltage steps
from 090 to /60 mV. In 28 of the 32 neurons, the solitarytract-evoked EPSCs had two components, a fast and a slow
component; in the remaining four neurons the EPSCs had
only a fast component. The fast component was maximal at
4.19 { 1.80 (SD) ms after the onset of the EPSC and was
negligible at 20 ms after the peak; the slow component developed more slowly, was clearly evident 20 ms after the peak,
and had a duration of 100 ms. For the cells with both components (Fig. 2A), the currents had a fast onset but decayed
slowly when the cell was depolarized to voltages positive
to 045 mV. For the cells with only a fast component, the
currents had a fast onset and decayed rapidly and similarly
at all voltages (Fig. 2B). For all 28 cells with both components, the ratio of the amplitude of the current measured at
20 ms after the peak (where the contribution of the fast
component was negligible) to the amplitude of the current
measured at the peak averaged 0.78 { 0.04 at voltages positive to /40 mV. For the remaining four cells, in which the
EPSCs had only a fast component, the ratio was 0.26 {
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
clamp experiments. Electrodes had a resistance of 3–5 MV. Whole
cell recordings in NTS cells were made with the Axoclamp 1D
patch-clamp amplifier (Axon Instruments, Foster City, CA) in voltage-clamp or current-clamp mode. Whole cell currents and voltages
were filtered at 2 kHz with the use of the Axoclamp amplifier fourpole Bessel filter, digitized at 10 kHz with the use of a DigiData
1200 Interface (Axon Instruments), and stored in a 386 DX computer. Data were analyzed off-line with the use of pClamp 6 software (Axon Instruments). EPSCs evoked by solitary tract stimulation were pharmacologically isolated by constant perfusion with the
competitive g-aminobutyric acid-A (GABAA ) receptor antagonist
picrotoxin (100 mM) or the noncompetitive GABAA receptor antagonist bicuculline (10 mM) or both. The synaptic currents were
no larger than 800 pA and the measured series resistance was
õ30 MV.
SOLITARY TRACT STIMULATION. Either a custom-made bipolar
electrode in which two 35-mm Pt/Ir wires were twisted and juxtaposed, or a commercially available bipolar tungsten stimulating
electrode with two 25-mm tips (Frederick Haer), was placed in
the solitary tract. Single stimuli (1–25 V, 100-ms pulses) were
delivered to the solitary tract ipsilateral to the recording site at 0.2
or 0.5 Hz. To determine that responses to solitary tract stimulation
were synaptically activated, we compared the solitary-tract-evoked
EPSCs before and in the presence of the calcium channel blocker
CdCl2 (100 mM).
PROTOCOLS. The NTS and solitary tract were visualized in the
slice with a 110 objective. A bipolar stimulating electrode was
placed in the solitary tract, and whole cell recordings (Hamill et
al. 1981) were obtained with the use of the blind-patch technique.
The recording sites were drawn on a representative slice at the
time of recording and subsequently reconstructed on one of three
representative slices drawn from 40-mm histological sections with
the aid of a camera lucida drawing tube.
For each solitary-tract-evoked EPSC, we measured the amplitude of the current at the peak of the response, where the fast
component of the EPSC is predominant, and at 20 ms after the
peak, where the slow component is predominant (Hestrin et al.
1990). Solitary-tract-evoked EPSCs were determined to have both
a fast and a slow component if the ratio of the amplitude of the
current measured at 20 ms after the peak to the amplitude of the
current measured at the peak of the EPSC was ¢0.6 when measured
at voltages positive to /40 mV.
Cells were classified as second-order neurons if the onset latencies were short ( °4.5 ms) and varied by õ0.5 ms. To independently verify that these short, invariant onset latencies were consistent with monosynaptic activation, we assessed the ability of five
cells with latencies °4.5 ms to discharge action potentials to each
of two solitary tract stimuli delivered with an interpulse interval
of 5 ms. On the basis of previously established criteria (Miles
1986), each of the two stimuli were required to evoke an action
potential for the cell to be considered to receive monosynaptic
input.
For antagonist studies, solitary-tract-evoked EPSCs were obtained in the normal perfusate and then 2–6 min after application
of either the NMDA receptor antagonist DL-2-amino-5-monophosphovaleric acid (APV) or the non-NMDA receptor antagonist
2,3-dihydroxy-6-nitro-7-sulfamoylbenzo(F)quinoxaline (NBQX).
The responses recovered after 2–5 min in the presence of normal
perfusate. For the nominally Mg 2/ -free studies, EPSCs were first
obtained in the nominally Mg 2/ -free perfusate and then measured
4–6 min after the change to the normal perfusate (which contained
1 mM Mg 2/ ).
DATA ANALYSES. Currents shown in the traces and currentvoltage plots are averages of two to four traces. Each voltage
was corrected by subtracting the voltage error ( the product of
the series resistance and steady-state current ) from the command
2541
2542
M. L. AYLWIN, J. M. HOROWITZ, AND A. C. BONHAM
0.06, significantly less than that for the EPSCs with both
components (P õ 0.001, unpaired t-test).
The onset latencies of the 28 neurons with both the fast
and slow components averaged 3.42 { 1.03 (SD) ms and
ranged from 2.2. to 7.7 ms. The onset latencies for the remaining four neurons in which the EPSCs had only the
fast component were not significantly different (P Å 0.33;
unpaired t-test), averaging 4.00 { 1.79 (SD) ms and ranging
from 2.0 to 5.7 ms. The EPSCs with short ( °4.5 ms) and
invariant (varied by õ0.5 ms) onset latencies were considered to be evoked by monosynaptic activation. Independent
evidence that these onset latencies were the result of monosynaptic activation was obtained in five cells that had similar
onset latencies (3.02 { 0.94 ms) and consistently discharged
action potentials to each of two stimuli delivered at 5-ms
intervals (Miles 1986), thus confirming that the solitarytract-evoked EPSCs were measured in second-order neurons.
We also measured solitary-tract evoked EPSCs in NTS
cells in longitudinal slices in which the stimulating electrode
was placed in the solitary tract Ç1 mm away from the recording site. The current-voltage relationships for the peak
current and for the current at 20 ms after the peak were the
same as for those obtained in coronal slices. Three of six
neurons with onset latencies of 4.5 { 0.6 (SD) ms had both
/ 9k11$$my22 J722-6
the fast and slow components. The remaining three neurons
had only the fast component and had onset latencies of
4.1 { 0.8 (SD) ms.
Although the fast and slow components overlap in time
( Figs. 2 and 3 ) , the fast component of the EPSC is predominant at the peak of the response, with only a minor contribution from the slow component, and the slow component is
predominant at 20 ms after the peak, with a small contribution from the fast component. Thus the fast component of
the EPSC is represented in the peak current-voltage relationship, whereas the slow component is represented in the
current-voltage relationship for the current measured at 20
ms after the peak. The current-voltage relationships were
plotted for the solitary-tract-evoked EPSCs that met the
criteria for monosynaptic activation given above and had
both a fast and slow component. An example is shown in
Fig. 3. At hyperpolarized potentials, the EPSCs had a fast
onset and fast decay, corresponding to the predominance
of the fast component. At depolarized potentials, the EPSCs
had a fast onset and a much slower decay, reflecting the
presence of both components ( Fig. 3 A ) . The current-voltage relationships for the peak current and the current at 20
ms after the peak for this neuron are shown in Fig. 3 B.
The amplitude of the fast component was linearly related to
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
FIG . 1. Position of recording electrode, stimulating electrode, and recording sites. A: photomicrograph showing a medullary
slice in the recording chamber. A inset: showing stimulating electrode in the solitary tract and recording electrode located
medial to the tract. B: location of recording sites (hatched areas) superimposed on 3 schematic coronal sections. All recordings
were made medial to the solitary tract; most were made in the intermediate nucleus of the solitary tract (NTS) at the level
of the area postrema (AP; 0100 to 0800 mm caudal to the obex). ts, solitary tract; c, central canal.
NMDA RECEPTORS IN PRIMARY VISCERAL AFFERENT TRANSMISSION
2543
NMDA receptor currents are negligible, APV had no observable effect. The responses recovered after 2–5 min in the
presence of normal perfusate (data not shown). The currentvoltage relationships for the peak current and the current at
20 ms after the peak for this neuron are shown in Fig. 4B.
APV reduced the amplitude of the current measured at 20
ms after the peak at voltages positive to 045 mV and had
no effect on the peak current.
For all 12 cells, APV significantly decreased the amplitude
of the slow component measured at voltages positive to /40
mV from 87 { 12 pA to 21 { 6 pA (P Å 0.0001; paired ttest) without significantly altering the fast component, which
was 117 { 14 pA before and 95 { 13 pA in the presence
of APV (P Å 0.124).
Effect of absence of Mg 2/ on solitary-tract-evoked EPSCs
Effect of APV on solitary-tract-evoked EPSCs
Effect of NBQX on solitary-tract-evoked EPSCs
To determine whether the slow component of solitarytract-evoked EPSCs obtained in second-order neurons was
mediated by NMDA receptors, we used the selective NMDA
receptor antagonist APV in cells receiving monosynaptic
input. The solitary-tract-evoked EPSCs were examined before and in the presence of 50 mM APV (n Å 12). An
example is shown in Fig. 4A. At /25 mV, where NMDA
receptor currents are present, APV markedly reduced the
amplitude of the slow component while having no effect on
the fast component. At 026 mV, where the NMDA receptor
currents are not yet blocked by Mg 2/ , APV diminished the
amplitude of the slow component. At 069 mV, where
We compared the EPSCs of nine neurons before and after
application of the non-NMDA receptor antagonist NBQX
(3 mM; n Å 9). Six of the nine neurons tested with NBQX
exhibited both fast and slow components; the remaining
three neurons tested with NBQX had only the fast component. An example of the effect of NBQX on the fast and
slow components of solitary-tract-evoked EPSCs at three
voltages before and in the presence of NBQX is shown in
Fig. 6. At all three voltages NBQX nearly abolished the fast
component of the response and had no effect on the slow
component.
For the six neurons exhibiting both fast and slow compo-
FIG . 2. Representative solitary-tract-evoked excitatory postsynaptic currents (EPSCs) measured at 088, 029, 0, /29, and /44 mV. A: cell showing
prolonged EPSC with both fast and slow component. Presence of slow
component is indicated by persistent outward current seen at 20 ms after
the peak (– – –) at positive voltages. B: cell with only a fast component
with a peak current that corresponds to peak of the EPSC (– – – labeled
peak). Absence of slow component is indicated by negligible outward
current seen at 20 ms after the peak.
/ 9k11$$my22 J722-6
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
voltage. By contrast, the slow component was nonlinearly
related to voltage, with currents that were only evident at
voltages positive to 045 mV. The current-voltage relationship for the slow component of the solitary-tract-evoked
EPSCs is characteristic of an NMDA-receptor-mediated
current ( Nowak et al. 1984 ) .
NMDA receptor currents are relatively small at membrane
potentials negative to 045 mV because of Mg 2/ block of
the channel (Nowak et al. 1984). To further demonstrate
that the slow component of the solitary-tract-evoked EPSCs
is mediated by NMDA receptors, we examined the EPSCs
in the absence of Mg 2/ (nominally Mg 2/ -free perfusate)
and then in the presence of Mg 2/ (normal perfusate)
(n Å 5). An example is shown in Fig. 5A. At 082 mV, the
amplitude of the slow component measured in the absence
of Mg 2/ was greater than in the presence of Mg 2/ . At /1
and /8 mV, where the Mg 2/ block of the NMDA channel
is negligible, the amplitude of the slow component was similar in the absence or presence of Mg 2/ . The current-voltage
relationship of the slow component for this same neuron in
the absence of Mg 2/ shows a linear behavior of the current
over a wider voltage range than in the presence of Mg 2/ ,
reflecting the voltage-dependent Mg 2/ block of the NMDA
receptor channel (Fig. 5B).
For all five neurons tested, the NMDA current at 082 mV
was significantly greater in the absence ( 0177 { 56 pA)
than in the presence ( 052 { 21 pA) of Mg 2/ (P Å 0.03;
paired t-test). Moreover, when the decay of the currents was
fitted with two exponentials at 090 mV, the ratio of the
amplitude of the slow to the fast exponential was significantly greater in the absence (2.37 { 0.49) than in the
presence (0.613 { 0.37) of Mg 2/ (P Å 0.007). The amplitude of the slow component at voltages positive to /45 mV
was not different in the absence (127 { 16 pA) or in the
presence (139 { 24 pA) of Mg 2/ (P Å 0.707).
2544
M. L. AYLWIN, J. M. HOROWITZ, AND A. C. BONHAM
nents of solitary-tract-evoked EPSCs, NBQX significantly
decreased the amplitude of the fast component, measured at
voltages positive to /40 mV, from 102 { 25 pA to 35 {
12 pA (P Å 0.014; paired t-test) while having no effect on
the slow component, which was 85 { 18 pA before and
61 { 18 pA in the presence of NBQX (P Å 0.134; paired
t-test). The responses recovered after 2–5 min in the presence of normal perfusate.
DISCUSSION
This is the first study to demonstrate that dual non-NMDA
and NMDA receptor mechanisms operate in synaptic transmission between primary visceral (sensory) afferent fibers
and second-order neurons in the intermediate and caudal
NTS. Two issues critical to the conclusions drawn from
this study are whether the solitary-tract-evoked EPSCs were
obtained in second-order neurons and whether the solitarytract-evoked EPSCs obtained in these second-order neurons
had an NMDA-receptor-mediated component.
Two related observations indicate that the recordings in
the present study were obtained from second-order neurons
that received monosynaptic input from the solitary tract: 1)
the onset latencies of the synaptically evoked EPSCs were
short and invariant (Figs. 2–5), and 2) the neurons reliably
discharged an action potential after each of two pulses separated by 5 ms was delivered to the solitary tract (Miles
1986). The onset latencies in our study are the same as those
observed by others for solitary-tract-evoked EPSCs (Glaum
and Miller 1993) and EPSPs (Andresen and Yang 1990;
Champagnat et al. 1986) in second-order neurons in this
region of the NTS.
Previous patch clamping and intracellular studies in vitro
/ 9k11$$my22 J722-6
have provided persuasive evidence that non-NMDA receptors mediate primary visceral afferent inputs from the solitary tract to second-order neurons within the intermediate
and caudal NTS (Andresen and Yang 1990; Brooks and
Spyer 1993; Brooks et al. 1992; Drewe et al. 1990). Our
finding that NBQX blocked the fast component of solitarytract-evoked EPSCs is consistent with these previous observations, and confirms that non-NMDA receptors mediate the
fast component of glutamate signaling in primary visceral
afferent transmission.
It is well established that functional NMDA receptors are
present on NTS neurons, as evidenced by in vivo and in vitro
findings that direct application of NMDA evokes excitatory
responses in NTS cells receiving visceral afferent input
(Brooks et al. 1992; Drewe et al. 1990; Wilson et al. 1996).
However, whether NMDA receptors contribute to visceral
afferent synaptic transmission to NTS cells has been more
controversial. Recording in a medullary slice, Brooks et al.
(1992) determined that non-NMDA receptor blockade abolished solitary-tract-evoked EPSPs. Although it was not determined whether the EPSPs were recorded from second- or
higher-order neurons, the data suggest that NMDA receptors
play no significant role in visceral afferent transmission in
the NTS. Andresen and Yang (1990) similarly found a predominant contribution of non-NMDA receptors to short-latency, solitary-tract-evoked EPSPs; this has led to the suggestion that, in contrast to the non-NMDA receptors, the
NMDA receptors probably do not mediate synaptic transmission between primary visceral afferent fibers and secondorder neurons, but rather may mediate transmission to
higher-order NTS neurons, presumably via polysynaptic
pathways (Andresen and Kunze 1994). This proposal that
NMDA receptors mediate visceral afferent transmission to
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
FIG . 3. Voltage dependence of solitary-tract-evoked EPSCs. A: EPSCs were recorded at 088, 073, 059, 043, 029,
014, 0, /13, /25, /38, and /50 mV. B: current-voltage plot for same cell at peak current shows a linear fast component
( s ). Current-voltage plot for current at 20 ms after peak shows a nonlinear slow component ( ● ).
NMDA RECEPTORS IN PRIMARY VISCERAL AFFERENT TRANSMISSION
2545
higher-order neurons is supported by findings by Brooks and
Spyer (1993) that, in the presence of non-NMDA receptor
blockade, stimulation in the area of the solitary tract occasionally evoked small NMDA-receptor-mediated EPSPs,
suggesting that NMDA receptors may contribute to synaptic
responses in higher-order neurons evoked by excitation of
neurons outside the solitary tract.
However, the previous observations were made in experiments in which EPSPs were recorded in NTS neurons at
their resting membrane potentials, conditions in which the
voltage-dependent block of NMDA channels may have obscured the contribution of NMDA-receptor-mediated currents. In the present study, in which we examined excitatory
currents in NTS neurons over a range of membrane potentials, we were able to observe, in addition to the fast nonNMDA-receptor-mediated component, a slow component of
solitary-tract-evoked EPSCs in second-order neurons. The
following lines of evidence suggest that the slow component
was mediated by NMDA receptors: 1) the current-voltage
relationship for the current measured at 20 ms after the peak
of the EPSCs was characteristic of NMDA receptor currents
(Figs. 2A and 3), exhibiting appreciable outward currents at
positive potentials but only small inward currents at negative
potentials (Hestrin et al. 1990); 2) the slow component was
/ 9k11$$my22 J722-6
significantly attenuated by NMDA receptor antagonism with
APV (Fig. 4), but not by non-NMDA receptor blockade
with NBQX; and 3) in the absence of Mg 2/ from the perfusate, the slow component was significantly enhanced at
hyperpolarized membrane potentials (Fig. 5), at which
Mg 2/ block of the NMDA receptor channel prevents conductance (Nowak et al. 1984). This enhanced amplitude of
the slow component in the absence of Mg 2/ was corroborated by an increased contribution of the exponential with a
slow time constant. Taken together, the findings indicate
that NMDA receptors participate in synaptic transmission
between visceral afferent fibers and second-order neurons in
the NTS.
On the other hand, certain limitations of this study must
be considered. First, we demonstrated that NMDA receptors
contributed to visceral afferent transmission in the NTS in
3- to 4-wk-old rats. Although NMDA receptor characteristics
have not been specifically studied during postnatal development in the rat NTS, data from other CNS regions suggest
that NMDA receptor density is maximal in 2- to 3-wk-old
rats and slowly declines up to the age of 1 yr (Erdo and
Wolff 1990), and that NMDA receptors are less sensitive
to Mg 2/ blockade in 3- to 4-wk-old rats compared with adult
rats 10–12 wk old (Morrisett et al. 1990). These observa-
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
FIG . 4. Effect of DL-2-amino-5-monophosphovaleric acid (APV) on solitary-tract-evoked EPSCs. A: EPSCs shown before
and in the presence of APV at the 3 voltages. B: current-voltage plots of the peak current before ( s ) and in the presence
of ( h ) APV, and current-voltage plots of the current at 20 ms after the peak before ( ● ) and in the presence of ( j ) APV.
2546
M. L. AYLWIN, J. M. HOROWITZ, AND A. C. BONHAM
tions raise the possibility that NMDA receptor contribution
to synaptic transmission may become less important in adult
rats. Second, we cannot unequivocally rule out the possibility
that during solitary tract stimulation the voltage spread outside the solitary tract and excited intervening neurons that
then synaptically activated the neurons from which the recordings were made. This possibility seems unlikely because
stimulating voltages (which averaged 4.4 { 3.3 V, mean {
SD, and ranged from 2 to 20 V) used to evoke EPSCs having
both NMDA- and non-NMDA-receptor-mediated components were no greater than the voltages (which averaged
7.2 { 2.2 V, mean { SD, and ranged from 5 to 10 V)
required to evoke EPSCs having only the non-NMDA component. In addition, EPSCs with short onset latencies having
both non-NMDA- and NMDA-receptor-mediated components were evoked not only in coronal but also in longitudinal slices, where the recording electrode was separated from
the stimulating electrode by up to 1 mm.
The majority (88%) of the cells had NMDA receptor
currents in addition to non-NMDA currents, indicating that
in the intermediate and caudal NTS, NMDA receptors con-
/ 9k11$$my22 J722-6
tribute to primary visceral afferent transmission to a high
percentage of second-order neurons in the NTS. Although
the modalities of the sensory endings and thus the function
of the individual NTS cells cannot be determined in the
slice, the majority of the cells had both NMDA and nonNMDA receptor currents. Thus it seems likely that the dual
receptors are a common mechanism for primary afferent
transmission in the NTS, where neurons receive primary
afferent input largely from cardiovascular and respiratory
sensory endings (Jordan and Spyer 1986; Loewy 1990;
Spyer 1981). This dual participation of NMDA and nonNMDA receptors in synaptic transmission has also been observed in the hippocampus (Hestrin et al. 1990) and visual
cortex (Artola and Singer 1987).
The significance of NMDA receptors in primary visceral
afferent transmission in the NTS may reside in three characteristics of the receptors that allow for added signal processing capacity beyond that provided by non-NMDA receptors. First, NMDA receptor channel conductance is maximal
both when glutamate is present and the cell is depolarized.
This suggests that NMDA receptor contribution may be en-
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
2/
FIG . 5. Effect of Mg
on solitary-tract-evoked EPSCs. A: EPSCs recorded initially in nominally Mg 2/ -free medium
( 0Mg 2/ ) and in the presence of 1 mM Mg 2/ (normal perfusate). B: current-voltage plots measured at 20 ms after the peak
of the EPSC in the absence of Mg 2/ ( m ) and in the presence of 1 mM Mg 2/ ( ● ).
NMDA RECEPTORS IN PRIMARY VISCERAL AFFERENT TRANSMISSION
2547
neuron, in the coordination of reflex control of autonomic
function.
The authors acknowledge Dr. Pedro Maldonado for the stimulating electrode, J. Stewart for technical assistance; Dr. Pam Pappone, Dr. Jesse Joad,
and A. Gupta for reviewing the manuscript; and E. Walker for clerical
support.
This work was supported by National Institutes of Health Grants HL52165, HL-48584, and DK-32907.
Address for reprint requests: A. C. Bonham, University of California,
Davis, Cardiovascular Medicine, TB 172 Bioletti Way, Davis, CA 95616.
Received 10 September 1996; accepted in final form 9 January 1997.
REFERENCES
hanced during high-frequency sensory input from visceral
afferent fibers or under conditions in which excitatory inputs
from other sources are integrated. In that regard, bursting
patterns of second-order neurons in the NTS have been observed following high-frequency stimulation of the solitary
tract (20–50 Hz for 100–600 ms) (Tell and Jean 1991) and
from direct application of NMDA (Tell and Jean 1993).
It may be that these second-order neurons having NMDA
receptors can transduce sensory input, which has no apparent
pattern, into these bursting patterns for signal transmission at
later synapses in the reflex circuitry. Second, the prolonged
depolarization due to NMDA receptor currents lengthens
the time during which otherwise ineffective inputs can be
integrated with the primary visceral afferent input. Third,
calcium entry through the channel activates second-messenger cascades, which are important in modification of synaptic
efficacy such as occurs in long-term potentiation (Artola
and Singer 1987; Collingridge et al. 1983) and in sustained
increases in discharge (windup) of dorsal horn neurons
(Mendell 1966; Woolf and Swett 1984).
In conclusion, this study offers the first direct evidence
for the contribution of NMDA receptors to synaptic transmission between primary visceral afferent fibers and secondorder neurons in the NTS. The second-order neurons may
play a more complex role, beyond that of a simple relay
/ 9k11$$my22 J722-6
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
FIG . 6. Effect of 2,3-dihydroxy-6-nitro-7-sulfamoylbenzo(F)quinoxaline (NBQX) on solitary-tract-evoked EPSCs. Currents shown before and
in the presence of NBQX at 3 different voltages.
ANDRESEN, M. C. AND KUNZE, D. L. Nucleus tractus solitarius—gateway
to neural circulatory control. Annu. Rev. Physiol. 56: 93–116, 1994.
ANDRESEN, M. C. AND YANG, M. Non-NMDA receptors mediate sensory
afferent synaptic transmission in medial nucleus tractus solitarius. Am.
J. Physiol. 259 (Heart Circ. Physiol. 28): H1307–H1311, 1990.
ARTOLA, A. AND SINGER, W. Long-term potentiation and NMDA receptors
in visual cortex. Nature Lond. 330: 649–652, 1987.
BONHAM, A. C., COLES, S. K., AND MCCRIMMON, D. R. Pulmonary stretch
receptor afferents activate excitatory amino acid receptors in the nucleus
tractus solitarius in rats. J. Physiol. Lond. 464: 725–745, 1993.
BROOKS, P. A., GLAUM, S. R., MILLER, R. J., AND SPYER, K. M. The actions
of baclofen on neurones and synaptic transmission in the nucleus tractus
solitarii of the rat in vitro. J. Physiol. Lond. 457: 115–129, 1992.
BROOKS, P. A. AND SPYER, K. M. Evidence for NMDA receptor-mediated
synaptic events in the rat nucleus tractus solitarii in vitro (Abstract). J.
Physiol. Lond. 467: 21, 1993.
CHAMPAGNAT, J., DENAVIT-SAUBIÉ, M., GRANT, K., AND SHEN, K. F. Organization of synaptic transmission in the mammalian solitary complex,
studied in vitro. J. Physiol. Lond. 381: 551–573, 1986.
COLLINGRIDGE, G. L., KEHL, S. J., AND MC LENNAN, H. Excitatory amino
acids in synaptic transmission in the Schaeffer collateral–commissural
pathway of the rat hippocampus. J. Physiol. Lond. 334: 33–46, 1983.
COTMAN, C. W. AND IVERSEN, L. L. Excitatory amino acids in the brain—
focus on NMDA receptors. Trends Neurosci. 10: 263–272, 1987.
DREWE, J. A., MILES, R., AND KUNZE, D. L. Excitatory amino acid receptors
of guinea pig medial nucleus tractus solitarius neurons. Am. J. Physiol.
259 (Heart Circ. Physiol. 28): H1389–H1395, 1990.
ERDO, S. L. AND WOLFF, J. R. Postnatal development of the excitatory
amino acid system in visual cortex of the rat. Changes in ligand binidng
to NMDA, quisqualate and kainate receptors. Int. J. Dev. Neurosci. 8:
199–204, 1990.
GLAUM, S. R. AND MILLER, R. J. Activation of metabotropic glutamate receptors produces reciprocal regulation of ionotropic glutamate and GABA
responses in the nucleus of the tractus solitarius of the rat. J. Neurosci.
13: 1636–1641, 1993.
GLAUM, S. R. AND MILLER, R. J. Presynaptic metabotropic glutamate receptors modulate v-conotoxin-GVIA-insensitive calcium channels in the rat
medulla. Neuropharmacology 34: 953–964, 1995.
GORDON, F. J. AND LEONE, C. Non-NMDA receptors in the nucleus of the
tractus solitarius play the predominant role in mediating aortic baroreceptor reflexes. Brain Res. 568: 319–322, 1991.
HAMILL, O. P., MARTY, A., NEHER, E., SAKMANN, B., AND SIGWORTH, F. J.
Improved patch-clamp techniques for high-resolution current recording
from cells and cell-free membrane patches. Pfluegers Arch. 391: 85–
100, 1981.
HESTRIN, S., NICOLL, R. A., PERKEL, D. J., AND SAH, P. Analysis of excitatory synaptic action in pyramidal cells using whole-cell recording from
rat hippocampal slices. J. Physiol. Lond. 422: 203–225, 1990.
JORDAN, D. AND SPYER, K. M. Brainstem integration of cardiovascular and
pulmonary afferent activity. Prog. Brain Res. 67: 295–314, 1986.
KARIUS, D. R., LIMING, L., AND SPECK, F. S. Nucleus tractus solitarius and
excitatory amino acids in afferent-evoked inspiratory termination. J.
Appl. Physiol. 76: 1293–1301, 1994.
KUBO, T. AND KIHARA, M. Evidence of N-methyl-D-aspartate receptormediated modulation of the aortic baroreceptor reflex in the rat nucleus
tractus solitarii. Neurosci. Lett. 87: 69–74, 1988.
LOEWY, A. D. Central autonomic pathways. In: Central Regulation of Auto-
2548
M. L. AYLWIN, J. M. HOROWITZ, AND A. C. BONHAM
/ 9k11$$my22 J722-6
as the neurotransmitter of primary baroreceptor afferent nerve fibers.
Science Wash. DC 290: 813–815, 1980.
TELL, F. AND JEAN, A. Bursting discharges evoked in vitro, by solitary tract
stimulation or application of N-methyl-D-aspartate, in neurons of the rat
nucleus tractus solitarii. Neurosci. Lett. 124: 221–224, 1991.
TELL, F. AND JEAN, A. Ionic basis for endogenous rhythmic patterns induced
by activation of N-methyl-D-aspartate receptors in neurons of the rat
nucleus solitarii. J. Neurophysiol. 70: 2379–2390, 1993.
VARDHAN, A., KACHROO, A., AND SAPRU, H. N. Excitatory amino acid
receptors in the nucleus tractus solitarius mediate the response to the
stimulation of cardio-pulmonary vagal afferent C fiber endings. Brain
Res. 618: 23–31, 1993.
WILSON, C. G., ZHANG, Z., AND BONHAM, A. C. Non-NMDA receptors
transmit cardiopulmonary C-fibre input in nucleus tractus solitarii (NTS)
in rats. J. Physiol. Lond. 496: 773–785, 1996.
WOOLF, C. J. AND SWETT, J. E. The cutaneous contribution of the hamstring
flexion in the rat, an electrophysiological and anatomical study. Brain
Res. 303: 299–312, 1984.
ZHANG, J. AND MIFFLIN, S. Differential role for NMDA and non-NMDA
receptor subtypes in baroreceptor afferent integration in rat NTS (Abstract). FASEB J. 10: 338, 1996.
08-08-97 12:40:46
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.5 on April 30, 2017
nomic Functions, edited by A. D. Loewy and K. M. Spyer. New York:
Oxford, 1990, p. 88–103.
MENDELL, L. M. Physiological properties of unmyelinated fiber projection
to the spinal cord. Exp. Neurol. 16: 316–332, 1966.
MILES, R. Frequency dependence of synaptic transmission in nucleus of
the solitary tract in vitro. J. Neurophysiol. 55: 1076–1090, 1986.
MORRISETT, R. A., MOTT, D. D., LEWIS, D. V., WILSON, W. A., AND
SWARTZWELDER, H. S. Reduced sensitivity of the N-methyl-D-aspartate
component of synaptic transmission to magnesium in hippocampal slices
from immature rats. Dev. Brain Res. 56: 257–262, 1990.
NOWAK, L., BREGESTOVSKI, P., ASCHER, P., HERBET, A., AND PROCHIANTZ,
A. Magnesium gates glutamate-activated channels in mouse central neurones. Nature Lond. 307: 462–465, 1984.
OHTA, H. AND TALMAN, W. T. Both NMDA and non-NMDA receptors in
the NTS participate in the baroreceptor reflex in rats. Am. J. Physiol. 267
(Regulatory Integrative Comp. Physiol. 36): R1065–R1070, 1994.
PARKIS, M. A., BAYLISS, D. A., AND BERGER, A. J. Actions of norepinephrine on rat hypoglossal motoneurons. J. Neurophysiol. 74: 1911–1919,
1995.
SPYER, K. M. Neural organization of the barorecepter reflex. Rev. Physiol.
Biochem. Pharmacol. 88: 23–124, 1981.
TALMAN, W. T., PERRONE, M. H., AND REIS, D. J. Evidence for L-glutamate