Department of Physiology, University of Manitoba, Winnipeg, Manitoba R3E 0W3, Canada; and Department of Physiology, Emory University, Atlanta, Georgia 30322
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Garraway, Sandra M. and Shawn Hochman. Modulatory Actions of Serotonin, Norepinephrine, Dopamine, and Acetylcholine in Spinal Cord Deep Dorsal Horn Neurons. J. Neurophysiol. 86: 2183-2194, 2001. The deep dorsal horn represents a major site for the integration of spinal sensory information. The bulbospinal monoamine transmitters, released from serotonergic, noradrenergic, and dopaminergic systems, exert modulatory control over spinal sensory systems as does acetylcholine, an intrinsic spinal cord biogenic amine transmitter. Whole cell recordings of deep dorsal horn neurons in the rat spinal cord slice preparation were used to compare the cellular actions of serotonin, norepinephrine, dopamine, and acetylcholine on dorsal root stimulation-evoked afferent input and membrane cellular properties. In the majority of neurons, evoked excitatory postsynaptic potentials were depressed by the bulbospinal transmitters serotonin, norepinephrine, and dopamine. Although, the three descending transmitters could evoke common actions, in some neurons, individual transmitters evoked opposing actions. In comparison, acetylcholine generally facilitated the evoked responses, particularly the late, presumably N-methyl-D-aspartate receptor-mediated component. None of the transmitters modified neuronal passive membrane properties. In contrast, in response to depolarizing current steps, the biogenic amines significantly increased the number of spikes in 14/19 neurons that originally fired phasically (P < 0.01). Together, these results demonstrate that even though the deep dorsal horn contains many functionally distinct subpopulations of neurons, the bulbospinal monoamine transmitters can act at both synaptic and cellular sites to alter neuronal sensory integrative properties in a rather predictable manner, and clearly distinct from the actions of acetylcholine.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Neurons within the spinal
cord represent a primary site for the integration of somatosensory
input. Spinal sensory integration is a dynamic process regulated by
factors that include multisensory convergence and pathway selection
(Baldissera et al. 1981; Jankowska 1992
;
Lundberg 1979
), activity-dependent plasticity (see
Millan 1999
), and neuromodulation (see Randic
1996
). Neuromodulatory responses within the spinal cord include
actions mediated by monoaminergic systems that originate in the brain
stem. These bulbospinal monoaminergic nuclei can be divided into three
subtypes by their transmitter phenotype, serotonin (5-HT),
norepinephrine (NA), or dopamine (DA). Neurons within these nuclei are
characterized by their widespread projections throughout the spinal
cord (e.g., Clark and Proudfit 1991
, 1993
;
Holstege et al. 1996
; Marlier et al.
1991a
).
The monoaminergic modulation of two prominent spinal cord
functional systems has been examined in some detail. These are the control of motor output and nociception. Generally, the monoamines have
been reported to facilitate motor activity and inhibit sensory systems
(Basbaum and Fields 1984; Bell and Matsumiya
1981
; Jacobs and Fornal 1993
; Wallis
1994
; Willis and Coggeshall 1991
), consistent with a general hypothesis on 5-HT function in the CNS forwarded by
Jacobs and Fornal (1993)
. Because serotonergic,
noradrenergic, and dopaminergic systems have a similarly diffuse
distribution in the spinal cord (Clark and Proudfit 1991
,
1993
; Holstege et al. 1996
; Marlier et
al. 1991a
; Rajaofetra et al. 1989
, 1992
) and
their monoamine transmitters frequently exert similar actions (Belcher et al. 1978
; Bell and Matsumiya
1981
; Headley et al. 1978
; Weight and
Salmoiraghi 1966
), it is possible that these transmitter
systems act at similar spinal sites and by similar mechanisms. For
example, descending monoaminergic transmitters powerfully inhibit
nociceptive information in neurons by activation of serotonergic
5-HT1A, 5-HT1B,
2-adrenergic, and
D2-dopaminergic receptors (Kiritsy-Roy
1994
; Pertovaara 1993
; Zemlan
1994
) all of which are negatively coupled to adenylate cyclase
(reviewed in Barnes and Sharp 1999
; Bylund et al.
1994
; Vallone et al. 2000
). However, the
existence of many bulbospinal monoaminergic systems with heterogeneous
transmitter phenotypes (including co-transmitters) that act on a
variety of spinal metabotropic receptor subtypes (e.g., Huang
and Peroutka 1987
; Marlier et al. 1991b
;
Stone et al. 1998
; van Dijken et al.
1996
), suggest that neuromodulation in the spinal cord is a
highly differentiated process. Indeed, more recent findings indicate
that different noradrenergic or serotonergic nuclei can exert opposing
modulatory actions on spinal cord nociceptive function
(Calejesan et al. 1998
; Martin et al. 1999
). Further, the actions of 5-HT and NA on the
afferent-evoked recruitment of functionally identified spinal neurons
can differ considerably (Bras et al. 1989
;
Jankowska et al. 1997
, 2000
). For example, the
recruitment of ascending tract neurons following primary afferent
stimulation is commonly facilitated by 5-HT yet depressed by NA
(Jankowska et al. 1997
).
The ester amine acetylcholine (ACh) also modulates spinal sensory
processing in the dorsal horn (Myslinski and Randic
1977; Urban et al. 1989
). As it appears that
there are no descending cholinergic systems in the rat (refer to
Willis and Coggeshall 1991
, but see Bowker et al.
1983
), these actions probably arise from a population of
intrinsic cholinergic interneurons found in the dorsal horn
(Barber et al. 1984
; Todd 1991
).
Several studies have compared the actions of these biogenic amine
transmitters on the modulation of sensory input onto spinal neurons
(Belcher et al. 1978; Bras et al. 1989
;
Headley et al. 1978
; Jankowska et al.
1997
; Skoog and Noga 1995
; Todd and
Millar 1983
; Weight and Salmoiraghi 1966
;
Willcockson et al. 1984
). However, in these studies,
only modifications in extracellular spiking or field potentials were
recorded and transmitters were applied by iontophoresis (but see
Bras et al. 1989
). While it is apparent from these
studies that the transmitters have both common and distinct actions on
the modulation of spinal sensory input, the effects of monoamines on
intrinsic cellular properties and synaptic potentials in individual
neurons were not studied. Clearly, additional insight into monoamine
transmitter function may be achieved by a more direct examination of
their actions with intracellular recordings (e.g., Khasabov et
al. 1998
, 1999
; Lopez-Garcia 1998
; Lopez-Garcia and King 1996
).
Therefore in this study, we compared the effects of bath-applied 5-HT,
NA, DA, and ACh on cellular properties and primary afferent-evoked
synaptic responses in individual deep dorsal horn (DDH) neurons using
whole cell patch-clamp recordings (see Garraway et al.
1997; Hochman et al. 1997
). Neurons in the DDH
represent a functionally heterogeneous population, which include
several classes of ascending tract cells and spinal interneurons
(Willis and Coggeshall 1991
) and may thereby account for
the diversity in intrinsic and synaptic properties observed from
neurons in this region (e.g., Jiang et al. 1995
;
Morisset and Nagy 1998
, 1999
; also refer to
Hochman et al. 1997
). Parts of these results have been
presented in abstract form (Garraway and Hochman 1999
).
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Preparation of spinal cord slices
All experimental procedures complied with the Canadian Council of Animal Care guidelines. Neonatal rats (Sprague-Dawley postnatal days 10-14) were first anesthetized with 10% urethan (2 mg/kg ip body wt) and decapitated, and spinal segments L2-S1 were removed. The isolated spinal cord was embedded in Agar, 2.5% wt/vol (Type E, Sigma), and sliced on a vibrating blade microtome in 500-600 µm transverse sections (Leica VT1000S or Pelco 101) in cooled (<4°C) oxygenated high sucrose-containing artificial cerebrospinal fluid (ACSF) containing (in mM) 250 sucrose, 2.5 KCl, 1 CaCl2, 3 MgCl2, 25 glucose, 1.25 NaH2PO4, and 26 NaHCO3 at a pH of 7.4. Short dorsal rootlets remained attached to the spinal segments to allow for electrical stimulation of primary afferents.
Electrophysiology
Slices were incubated at 32°C for 1 h in normal ACSF
containing (in mM) 125 NaCl, 2.5 KCl, 2 CaCl2, 1 MgCl2, 25 glucose, 1.25 NaH2PO4, and 26 NaHCO3 at a pH of 7.4 and oxygenated with 95%
O2-5% CO2. For
experimentation, spinal cord slices were affixed to a recording chamber
using platinum U frames with a parallel array of nylon fibers glued
across (Edwards et al. 1989
). Patch electrodes were
prepared from 1.5-mm OD capillary tubes (Precision Instruments or
Warner) pulled in a two-stage process (Narishige PP83) producing
resistance values ranging from 4 to 7 M
with recording solution
containing (in mM) 140 K-gluconate, 0.2 EGTA, 10 HEPES, 4 Mg-ATP, and 1 GTP; pH 7.3. The recording chamber was continuously superfused with
oxygenated normal ACSF at a rate of ~2 ml/min. The whole cell
"blind" patch-clamp recording technique (Blanton et al.
1989
) was undertaken at room temperature (~20°C) using the
Axopatch 1D amplifier (Axon Instruments) filtered at 5 kHz (4-pole
low-pass Bessel). Voltage- and current-clamp data were acquired on
computer with the pCLAMP acquisition software (v 6.0; Axon Instruments).
Determination of cell membrane properties
Immediately following rupture of the cell membrane (in voltage
clamp at 90 mV), the current-clamp recording configuration was used
to determine resting membrane potential. Series resistance was
subtracted in current-clamp mode (bridge balance), and junction potentials were measured and subtracted off-line. For the duration of
the experiment, leak conductance and bridge balance were monitored; if
their values were largely unaltered, the experiments were continued. Mean electrode series resistance was 33 ± 4 (SD) M
(n = 37). At an adjusted membrane potential of
70 mV,
a series of hyperpolarizing and depolarizing current steps were
undertaken to obtain estimates of membrane time constant, cell
resistance, rheobase, voltage threshold, action potential height, and
action potential duration at half-maximal amplitude (half-width). For
details on the estimation of these membrane properties, refer to
Hochman et al. (1997)
.
Primary afferent stimulation
Primary afferents were stimulated electrically with a constant
current stimulator (Eide 1972), using bipolar tungsten
electrodes. In the present comparative study, we used high stimulation
intensities to recruit the highest threshold unmyelinated afferents
and, hence, the majority of afferent fiber types, irrespective of age
(typically
500 µA, 500 µs) (see Thompson et al.
1990
). In the present sample, 29% of the neurons received
synaptic responses at intensities <500µA, 100µs; 49% received
synaptic input at 500µA, 100µs, while the remaining 22% of neurons
only received input at intensities
500 µA, 500 µs. Generally, the
evoked synaptic responses were first characterized as excitatory by
determining their reversal potential prior to collection of baseline
events. Neurons with short-latency inhibitory synaptic responses were
not included in this study. Excitatory postsynaptic potentials (EPSPs)
were evoked at low frequencies (once every 20-60 s) by stimulating dorsal rootlets for a baseline period of 10-15 min while maintaining the neuron at a holding potential of
90 mV. In all cases, membrane potential was carefully monitored, and any alterations in membrane potential were noted, then countered with intracellular current injection to maintain a holding potential of
90 mV.
Application of agonists
5-hydroxytryptamine HCl (5-HT), norepinephrine bitartrate (NA),
dopamine HCl (DA), and acetylcholine chloride (ACh) were obtained from
RBI/Sigma. The solutions were prepared on the day of the experiment
from 10 mM frozen stock solutions and bath applied at a final
concentration of 10 µM. Ascorbic acid (100 µM), an antioxidant, was
added to solutions containing 5-HT, NA, and DA to prevent their
oxidation (Krnjevic et al. 1978). A 10 µM
concentration was chosen based on previous studies involving bath
application of NA, 5-HT, and ACh (e.g., Baba et al.
2000
; Lopez-Garcia and King 1996
;
Miyazaki et al. 1998
; Urban et al. 1989
).
While this concentration may (e.g., Lopez-Garcia and King
1996
) or may not have a maximal physiological response (e.g.,
Baba et al. 2000
), it is likely to be below the
concentration range where nonspecific binding and actions have been
observed (Chesnoy-Marchais and Barthe 1996
; van
Wijngaarden et al. 1990
). All agonists were dissolved in normal
ACSF and bath applied from independent perfusion lines. Each agonist
was applied for a period of 7-15 min during which time, EPSPs were
continually recorded at the baseline parameters described in the
preceding section.
To compare the actions of more than one agonist on the primary afferent-evoked synaptic responses and cellular properties of the neurons, we allowed a washout/recovery period of 10-20 min before subsequent drug application. Due to the restrictions in recording duration with patch electrodes, we were often unable to observe the effects of all four agonists on a given neuron. However, in all cases, the actions of at least two drugs were compared. The following three combinations of drugs were compared in most cases: 5-HT and ACh, 5-HT, NA, and DA, and 5-HT, NA, DA, and ACh. These transmitters were applied in random order, and evoked EPSPs were always recorded at baseline parameters both during drug application and washout. In separate experiments, we compared the magnitude of modulatory actions evoked by independent application of 5-HT and NA to their co-application (5-HT/NA).
Analysis
Recordings were analyzed using pCLAMP (v 6.0, Axon Instruments).
Both the maximum amplitude of the synaptic response and the changes in
synaptic charge transfer calculated as the integral of the synaptic
response (area under the curve) of individual traces were measured.
Primary afferent-evoked synaptic responses in dorsal horn neurons are
generally glutamatergic consisting of both early and late components;
presumably (±)--amino-3-hydroxy-5-methylisoxasole-4-propionic acid
(AMPA)/kainate and N-methyl-D-aspartate (NMDA)
receptor-mediated respectively (e.g., Gerber and Randic
1989
). To determine whether, as a first approximation, the
drugs differentially modulated these components of the evoked
responses, we calculated area under the curve (AUC) at two time
intervals that approximately separate these events: early (<200 ms)
and late (
200
750 ms). These periods were chosen to
approximately separate AMPA/kainate from NMDA receptor-mediated actions
as described in an earlier study (Garraway and Hochman
2001a
) (Fig. 3A). Synaptic events generally returned
to baseline before 750 ms. Analysis of EPSPs within this time interval
excludes actions from the long-latency peptidergic component
(Urban and Randic 1984
).
The applied transmitters were considered to have a modulatory action if
they altered EPSP amplitude 10%. Because multiple drugs were added
in most experiments, and in several experiments evoked responses did
not return to naïve EPSP baseline values, the change in
synaptic amplitude was measured as a difference in the mean peak
amplitude or AUC during drug application compared with the mean values
just prior to drug application (control or washout/recovery).
Similarly, the estimated membrane properties and firing properties of
the neurons were compared before, during, and following washout of the
drugs tested. Following analysis, graphs were constructed using Sigma
Plot (SPSS) and imported into CorelDRAW (Corel) for final editing. Both
AUC and peak amplitude were measured, and as reported later, similar
results were obtained for the actions of 5-HT, NA, and DA. Thus unless
stated, all values are reported as means ± SD of peak changes
(maximum amplitude) of the synaptic response. Unless otherwise stated,
the effects of the monoamines are statistically compared with control
values using the Students's t-test. Multiple pairwise
comparisons on differences in values between the monoamine transmitters
are not reported.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Sample population
A total of 41 deep dorsal horn neurons (laminae III-VI) were
recorded. The approximated location of 36 of these neurons, determined using visual landmarks (King et al. 1988), is presented
in Fig. 1A. The membrane
properties of the neurons are summarized in Table 1. Neurons can also be grouped by their
firing properties in response to current injection. Phasically firing
neurons could not be induced to fire repetitively regardless of the
magnitude of current injection. Repetitive firing neurons always fired
repetitively in responses to larger magnitude current injections. In
the present sample, 15 neurons fired spikes repetitively, while the
remaining 26 neurons fired only phasically, 10 of which fired no more
than two spikes (1-2 spikes) (refer to Lopez-Garcia and King
1994
). Examples of neuronal firing patterns are illustrated in
Fig. 1B. The effects of the monoamines on neuronal membrane
properties will be considered later.
|
|
Effects of the monoamine transmitters on primary afferent-evoked EPSPs
The effects of bath-applied 5-HT, NA, DA, and ACh on the dorsal root stimulation-evoked EPSPs are summarized in Tables 2 and 3 and illustrated in Fig. 2. In the majority of neurons, 5-HT and NA depressed EPSP amplitude (P < 0.05), and a similar trend was observed for DA. In comparison ACh facilitated EPSP amplitude (P < 0.05; Table 3, Fig. 2A). When EPSP AUC values were compared, 5-HT, NA, and DA decreased while ACh increased AUC values (P < 0.05). These observations demonstrate that despite the heterogeneous population of neurons found in the DDH, the biogenic amine neurotransmitters altered afferent-evoked EPSPs rather predictably with the bulbospinal monoamines having similarly depressant actions. The relative incidence of the transmitters at modulating peak EPSP amplitude were 5-HT > ACh > NA > DA (Table 2). Following washout of NA and DA, partial recovery of EPSP amplitudes generally occurred (Fig. 2B). However, in some neurons, during the washout that followed 5-HT-evoked synaptic depression, a rebound potentiation of EPSP amplitude occurred, usually exceeding control values. In addition, EPSP amplitudes generally remained facilitated following washout of ACh.
|
|
|
We also compared the differential effects of the agonists on the early
versus the late occurring components of the evoked EPSPs by measuring
the AUC for early (<200 ms) and late (200
750 ms) components
of the EPSP (Table 4), periods that
approximately separate AMPA/kainate from NMDA receptor-mediated actions
(e.g., Fig. 3A) (see also
Garraway and Hochman 2001a
). There were no differences
in depression between early and late AUC values for 5-HT, NA, and DA
(Table 4), supporting a uniform depression of both AMPA/kainate and
NMDA receptor-mediated responses. However, AUC measures show that ACh
preferentially facilitated the late component of the EPSP supporting a
preferential facilitation of the NMDA receptor-evoked response (Table
4). Figure 3B illustrates a representative example of the
effects of 5-HT and ACh on evoked EPSPs in a given cell. Figure
3C compares the AUC increases for early and late components
of the EPSP following application of ACh in individual neurons. Note
that in many neurons, the later component of the EPSP is dramatically
facilitated compared with the early component of the EPSP.
|
|
The modulatory actions of ACh and 5-HT were compared in 12 neurons (Fig. 4). In six of the nine neurons where EPSPs were depressed by 5-HT, ACh facilitated the EPSPs. Thus 5-HT and ACh have different (P < 0.01; Wilcoxon signed-rank test) and predominantly opposite actions on spinal neurons in the deep dorsal horn.
|
The generally depressant actions of the three descending transmitters 5-HT, NA, and DA were compared in 15 neurons (Fig. 5). With few exceptions, none of the transmitters had competing modulatory actions on the evoked EPSPs. In 6 of the 15 neurons, all three transmitters produced synaptic depression while in 1 neuron, all three drugs produced synaptic facilitation. In 8 of 15 neurons, common modulatory responses were not produced by all three transmitters, but only in 3 of these neurons were opposite actions >10% observed. An example of the common neuromodulatory actions of the brain stem monoamines in a single neuron is presented in Fig. 5B. Thus unlike ACh, which generally supported synaptic facilitation, the three descending monoamines 5-HT, NA, and DA commonly exerted similar functions on spinal cord sensory input to a given cell.
|
In 10 neurons, 5-HT and NA were applied individually as well as co-applied. Co-application of 5-HT and NA evoked synaptic depression of a greater magnitude than the sum of 5-HT and NA applied alone in 3 of the 10 neurons tested (Fig. 6A, asterisks). The effects of co-applied 5-HT and NA were not significantly different from the effects of 5-HT alone but significantly greater than effects of NA alone (P < 0.05; Tukey test), suggesting a prominent contribution from 5-HT on synaptic depression when transmitters are combined.
|
Effects of the agonists on cellular properties
As a population, none of the transmitters had significant effects on cell passive membrane properties, rheobase or voltage threshold (Table 1). The effects of the agonists on EPSP amplitude were plotted against cell input resistance (Rin) to identify corresponding actions that would support a postsynaptic site of action (Fig. 7A). However, with the exception of ACh, where a relatively weak relationship existed (r2 = 0.18), no relations were found.
|
The actions of the biogenic amines on firing properties were examined
in 32 neurons by counting spike numbers during 1.2-s depolarizing
current pulses at approximately twice rheobase current intensities.
Overall, the monoamines significantly increased firing in these neurons
(P < 0.05). However, when neurons were divided into
those that initially fired either phasically (n = 19)
or repetitively (n = 13) in response to current
injection, only the population of neurons that fired phasically had an
increased spike number (P < 0.01). In addition, when
the actions of the individual transmitters were separated and compared
in this population significant increases were produced only by 5-HT
(P < 0.01). In this population, 14 of the 19 cells had
increased number of spikes during current injection in the presence of
the transmitters (5/6 from neurons initially firing 1-2 spike
population; Fig. 7B, ). This increase in spike number was
largely attributable to the observation that the cells that originally
fired phasically were converted into neurons that fired repetitively
during monoamine transmitter application. An example is presented in
Fig. 7C. In contrast, in cells initially capable of
repetitive firing, no such trends were evident (Fig. 7B,
). No relationship existed between the type of neuronal firing observed in response to current injection (e.g., phasic vs. repetitive) and the effects of the agonists on EPSP amplitude.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Summary
In this study, we investigated the effect of each biogenic amine transmitter on primary afferent-evoked synaptic responses and membrane properties in deep dorsal horn neurons. First, we observed that 5-HT, NA, and DA generally had common actions on evoked EPSPs in individual neurons, with the dominant action being depression. In contrast, ACh generally increased EPSP amplitude, even in the same neurons where synaptic depression was evoked by the bulbospinal monoamines. The order of the incidence and magnitude of evoked modulatory actions on EPSP amplitude was 5-HT > ACh > NA > DA. Second, while 5-HT, NA, and DA tended to uniformly depress short- and long-latency components of the evoked EPSPs, the ACh-induced facilitatory response was significantly greater for the later, presumably NMDA receptor-mediated component of the EPSP. Third, we observed that co-application of 5-HT and NA could produce a much greater synaptic depression than either transmitter applied independently. Finally, while passive membrane and threshold properties of neurons were unaffected by the monoamines, membrane firing properties in the subpopulation of neurons initially expressing a phasic firing pattern were converted to repetitive.
Common actions of the monoamines on evoked EPSPs in individual neurons
Consistent with previous in vitro studies, we observed that 5-HT
generally produced synaptic depression (e.g., Khasabov et al.
1999; Lopez-Garcia 1998
; Lopez-Garcia and
King 1996
) although facilitation was observed in a few cells.
Like 5-HT, bath application of NA and DA also produced synaptic
depression in most cells, consistent with depressant actions observed
in previous studies that monitored alterations in firing frequency
(Fleetwood-Walker et al. 1988
; Headley et al.
1978
; Skoog and Noga 1995
; Willcockson et
al. 1984
). An important observation in this study is that 5-HT, NA, and DA commonly produced the same modulatory action on EPSPs when
applied independently to the same neuron. Thus it appears that despite
the diverse functional heterogeneity of the spinal cord dorsal horn
(Baldissera et al. 1981
; Jankowska 1992
;
Willis and Coggeshall 1991
), the bulbospinal monoamine
transmitters can produce a widespread depression of sensory synaptic
input onto deep dorsal horn neurons. It is, however, necessary to
mention, that in addition to the common actions generally evoked by the three transmitters arising from brain stem nuclei on individual neurons, these transmitters were also capable of inducing distinct modulatory actions on some neurons.
The bulbospinal monoamine transmitter systems project widely throughout
the spinal cord (Holstege et al. 1996; Marlier et al. 1991a
; Rajaofetra et al. 1992
), and there
are many serotonergic, noradrenergic, and dopaminergic receptors in the
dorsal horn (e.g., Huang and Peroutka 1987
;
Marlier et al. 1991b
; van Dijken et al. 1996
). Hence, the effects of the bath-applied bulbospinal
monoamine transmitters observed here probably reflect the actions of
these transmitters on their respective families of receptors. However, it is also possible that part of the observed depression of the longer-latency portion of the EPSP is due to the transmitter having a
direct voltage-dependent block of the NMDA receptor ionophore (Chesnoy-Marchais and Barthe 1996
). That 5-HT, NA, and
DA had common actions in most cells suggests that receptors for all
three transmitters are co-localized on many neurons and/or primary
afferent terminals. In another study, we have demonstrated that the
synaptic depression produced by 5-HT in similarly located neurons is
only partly mediated by 5-HT1A receptors while
5-HT7 receptors are predominantly responsible for
producing synaptic facilitation (Garraway and Hochman
2001b
).
In contrast to the depressant actions of the monoamines observed here,
Jankowska and colleagues (1997, 2000
) demonstrated that
nonnociceptive afferent input from different afferents to different
groups of spinal interneurons or ascending tract cells is modulated by
NA and 5-HT in a highly differentiated manner. Depending on the neuron
and afferent fiber type, they observed that NA and 5-HT could have
common facilitatory, inhibitory, or opposing modulatory actions on
synaptic input strength as measured using peristimulus time histogram
measures of extracellular spike latency and frequency. An explanation
for the observed differences between their results and ours is our
stimulation at high intensities to also recruit high-threshold C and
A
fibers, which comprise the largest fraction of primary afferent
fibers (Snider and McMahon 1998
; Willis and
Coggeshall 1991
). Therefore the strong depressant actions of
the monoamines in our study may result from a dominating activation of
high-threshold afferents that mask more subtle differential modulatory
actions on the low-threshold afferents studied by Jankowska and
colleagues. Another explanation for our observed differences may relate
to our finding that EPSP amplitude can decrease concomitant with
increases in firing postsynaptically. Together, these studies support a
complex and functionally differentiated modulation of sensory-evoked
firing properties in spinal cord neurons.
Despite the similar actions of the three transmitters arising from the
brain stem, the magnitudes of synaptic depression differed in order of
5-HT > NA > DA. In addition, the depression evoked by 5-HT
was more widespread as a greater proportion of cells underwent synaptic
depression by 5-HT. The differences in depression may reflect the
relative effectiveness of different bulbospinal systems in mediating
modulatory actions. Alternatively, the relative magnitude of effects
may be due to differences in potency, as the transmitters may not be
working at the same relative point on their dose-response curves (e.g.,
Elliot and Wallace 1992).
The actions of 5-HT and NA on antinociception have been extensively
studied (Basbaum and Fields 1984; Fitzgerald
1986
; Jones 1991
; Millan 1995
).
Previous studies have demonstrated that noxious input leads to the
release of both NA and 5-HT in the spinal cord (Satoh and Omote
1996
; Tyce and Yaksh 1981
; Yaksh and Tyce
1981
), and these transmitters can also produce antinociception
following release after stimulation of specific supraspinal sites
(e.g., Cui et al. 1999
; Sorkin et al.
1993
). In this study, we observed that co-application of 5-HT
and NA could produce synaptic depression of a greater magnitude than
5-HT or NA applied alone. It is possible that both transmitters are
co-released physiologically under conditions where a maximal sensory
depression is sought. While nociceptive input does not appear to evoke
release of dopamine (Satoh and Omote 1996
), both
stimulation of the A11 dopaminergic cell group and exogenous
application of DA can elicit antinociception (Fleetwood-Walker et al. 1988
). Overall, the role of DA in mediating
antinociception remains poorly studied, and it is possible that DA may
play a different role in the modulation of primary afferent input than 5-HT and NA.
Facilitatory actions of ACh
In contrast to the bulbospinal transmitters, ACh generally
facilitated primary afferent-evoked responses. Although some studies have reported inhibitory or antinociceptive effects of cholinergic agonists (e.g., Bleazard and Morris 1993), facilitatory
or excitatory actions, consistent with our observations, have also been
reported. For instance, Urban et al. (1989)
reported an
increase in excitability of spinal cord dorsal horn neurons by ACh,
while Baba et al. (1998)
reported a muscarinic-induced
facilitation of GABA release in substantia gelatinosa neurons of the
rat. We also observed that ACh caused significantly greater
facilitation of the late, largely NMDA receptor-mediated, component of
the evoked EPSP. This is consistent with previous studies demonstrating
the facilitatory effects of cholinergic agonists on NMDA
receptor-mediated events in various CNS regions including hippocampus
(Marino et al. 1998
) and striatum (Calabresi et
al. 1998
) via M1-like receptor activation. M1 receptor activation leads to an increase in protein
kinase C (PKC), and PKC enhances NMDA receptor activity (e.g.,
Chen and Huang 1992
; Xiong et al. 1998
).
Modulation of neuronal firing properties
The values of membrane and threshold properties obtained here are
comparable to those obtained from an earlier study on the properties of
similarly located neurons (Hochman et al. 1997). The
monoamine transmitters did not significantly alter the passive membrane
or threshold properties of DDH neurons. However, in neurons that
originally fired phasically, the monoamines could reversibly transform
firing behavior from phasic to repetitive. These neurons did not
display accelerated discharges in response to injection of depolarizing
current steps or continued membrane depolarization and firing after
current step termination. Thus the increased duration of firing is
unlikely to be due to the activation of a plateau potential (see
Morisset and Nagy 1998
). Interestingly, Lopez-Garcia and King (1994)
showed that firing patterns
in response to current injection are functionally correlated to the
source of primary afferent input. For example, wide dynamic range (WDR) neurons receive convergent input from both low- and high-threshold afferents, and these cells generally fire repetitively. If we applied
this classification to the neuronal firing properties observed
following the application of monoamines, the monoamines could convert
the functional properties of neurons from those that previously had
restricted sensory convergence to a WDR profile. Interestingly, the
majority of neurons in the dorsal horn of awake sheep, obviously having
normal bulbospinal activity, are WDR (Herrero and Headley
1995
). It is also possible that the monoamines do not alter
convergent properties to DDH neurons but rather, the classification
scheme developed by Lopez-Garcia and King (1994)
does
not apply to the behavior of neurons in the presence of monoamine transmitters.
It may seem inconsistent that the changes in neuronal firing properties
were unrelated to the observed changes in EPSP amplitude. However,
since different receptor subtypes may be found at pre- and postsynaptic
sites, it is not surprising that the monoamine transmitters exert
different actions on the synaptic and firing properties of spinal
neurons. For instance, while one class of receptor may depress sensory
input (e.g., Khasabov et al. 1999), another class may
increase the excitability of these neurons (e.g., Wallis et al.
1991
). Opposing pre- and postsynaptic actions may be a widely
employed strategy to alter the network properties of spinal neurons.
For example, a depressed sensory input with increased neuronal
responsiveness could support a transfer of control from peripheral to
descending command systems.
Possible sites of action
Although the bulbospinal monoamines are capable of exerting direct
postsynaptic actions, several observations suggest that the depression
they produced on EPSPs is mediated via presynaptic mechanisms. First,
none of these transmitters had effects on the passive membrane
properties that would support a reduction in EPSP amplitudes by
postsynaptic mechanisms (i.e., a decreased m
or Rin). In addition, the identical
percent depression on early, presumable AMPA/kainate and late,
presumably NMDA receptor-mediated, components can be explained by a
reduction in glutamate transmitter release. Consistent with a
presynaptic site of action, all three transmitters, 5-HT, NA, and DA,
have been shown to depress synaptic responses presynaptically as a
result of their ability to increase potassium conductance (e.g.,
North and Yoshimura 1984
, see also Barnes and
Sharp 1999
) or by inhibiting both N- and L-type calcium channels (e.g., Wikström et al. 1999
). Other
studies suggested that 5-HT and NA may mediate presynaptic inhibition
of glutamate release from primary afferents in the guinea pig
(Travagli and Williams 1996
), while presynaptic
D2 dopamine receptor mediates depression of
spinal reflexes (Gajendiran et al. 1996
). In addition, many monoaminergic receptors are present on primary afferent terminals (e.g., Daval et al. 1987
; Hamon et al.
1989
; Kidd et al. 1993
; Ridet et al.
1994
; Stone et al. 1998
), and application of
5-HT has been shown to generate a primary afferent depolarization (PAD) that corresponds to presynaptic inhibition of primary afferents (Khasabov et al. 1998
, 1999
; Lopez-Garcia and
King 1996
). Interestingly, Khasabov et al.
(1998)
demonstrated that capsaicin treatment, which selectively
destroys unmyelinated primary afferent fibers, significantly reduced
5-HT-induced PAD in the neonatal rat. Thus 5-HT may mediate spinal
inhibitory actions by reducing transmitter release from nociceptors.
The ability of the monoamine transmitters to reduce nociceptive
transmitter release at the first CNS synaptic input site supports a
critical role in antinociceptive function.
In contrast to the bulbospinal monoaminergic transmitters, ACh probably
mediates its facilitatory actions predominantly postsynaptically. ACh
had a preferential facilitatory action on the late, presumably NMDA
receptor-mediated, component of the EPSP, consistent with observed
modulatory actions of M1 muscarinic receptor activation on
NMDA receptor activity (Calabresi et al. 1998;
Marino et al. 1998
). If facilitation in synaptic
strength involved only presynaptic mechanisms that increase glutamate
release, one would expect to see a uniform facilitation of both
AMPA/kainate and NMDA receptor-mediated responses. Interestingly, the
presumed postsynaptic action of ACh in this study is consistent with
the dominant postsynaptic localization of cholinergic receptors in the
spinal cord dorsal horn (Gillberg and Askmark 1991
),
particularly the muscarinic receptors which are expressed about two to
three times higher than the nicotinic receptors (Gillberg et al.
1988
).
Experimental issues
There are several potential limitations to our experimental
approach. 1) Bath-applied transmitters will activate all
their respective receptors simultaneously, whereas physiologically, it
is possible that there is some degree of preferential activation of
receptor subtypes by separate descending serotonergic systems (Wei et al. 1999). Thus physiological conditions may
exist where synaptic facilitation dominates over the depression
observed in this study. Nonetheless, for 5-HT at least, the dorsal horn
contains many 5-HT varicosities that end blindly and do not form
conventional synapses (Poulat et al. 1992
; Ridet
et al. 1994
) and in this regard have been suggested to modulate
spinal activity via volume transmission (Zoli and Agnati
1996
). Thus if 5-HT and the other descending monoamine
transmitters act via volume transmission, their actions would
nonspecifically activate all receptor subtypes. Interestingly, in this
regard, the net effect in most neurons would be a significant synaptic
depression. 2) Our use of high-intensity electrical
stimulation to recruit the majority of afferents (Thompson et
al. 1990
) prevents a comparison of modulatory actions between
specific afferent fiber populations. For example, Jankowska and
co-workers (Jankowska et al. 1995
, 1997
, 2000
;
Noga et al. 1992
; Riddell et al. 1993
) demonstrated that 5-HT and raphe-spinal stimulation can exert a
differential control over primary afferents of different modalities. 3) The DDH spinal region is functionally heterogeneous, yet
we treated all dorsal horn neurons as a single population.
Jankowska and co-workers (2000)
have demonstrated that
the actions of 5-HT may depend on the functional identity of the spinal
neuron studied. 4) The animals were recorded from immature
rats (P10-14), and hence the actions may be different in adult rats.
However, this is the preferred age range to study sensory integrative
mechanisms in vitro due to its near-intact circuitry, and near mature
developmental status (Fitzgerald 1985
; Fitzgerald
and Koltzenburg 1986
; Lopez-Garcia and King
1994
; Thompson et al. 1990
; Woolf and
King 1987
). Furthermore, the modulatory actions observed here
are unlikely to differ from those seen in the adult. If we take 5-HT as
an example, 5-HT immunoreactivity is found in the gray matter at all
spinal cord levels at birth, and staining density increases, peaking at
P7 in cervical and P14 in lumbar cord (Bregman 1987
). In
addition, Wallis et al. (1993)
showed that the strong
descending inhibition of the monosynaptic reflex in the P1 rat is
mediated by serotonin. Thus at least for 5-HT, descending modulatory
systems are present at birth (but see Fitzgerald and Koltzenburg
1986
).
The aforementioned limitations notwithstanding, the present observations demonstrate that the descending monoamine transmitters 5-HT, NA, and DA have broadly similar depressant actions in the spinal cord and on the same neurons, and these actions are opposite to the facilitatory actions of ACh, which is normally released from intrinsic spinal neurons. That these effects were similar in the majority of neurons sampled suggests that descending monoaminergic systems are capable of modulating spinal sensory integration in a diffuse and general matter.
Significance
In conclusion, we provide the first comparative analysis of the actions of the biogenic amine transmitters on synaptic and cellular properties of DDH spinal neurons. 5-HT, NA, and DA are involved in the descending control of spinal sensory integration, and the present observations suggest that the separate brain stem monoaminergic systems can affect spinal sensory integration in a remarkably similar manner. With respect to the control of sensory input, these studies increase our understanding of how these transmitters act, perhaps to maximize antinociceptive actions.
![]() |
ACKNOWLEDGMENTS |
---|
We thank C. Gibbs for expert technical assistance.
This work was supported by the Christopher Reeve Paralysis Foundation. S. M. Garraway obtained studentship salary support from the Rick Hansen Institute.
![]() |
FOOTNOTES |
---|
Address for reprint requests: S. Hochman, Rm. 362, Physiology Building, Emory University School of Medicine, 1648 Pierce Dr., Atlanta, GA 30322 (E-mail: shochman{at}physio.emory.edu).
Received 20 December 2000; accepted in final form 25 July 2001.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|