| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 Department of Biology, Section of Physiology and Biophysics and Centre of Neuroscience, Ferrara University, 44100 Ferrara, Italy; 2 Centre of Neuroscience, Insubria University, 21100 Varese, Italy
Submitted 11 December 2002; accepted in final form 15 April 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
) that the
subthreshold behavior of the rat sympathetic neuron is actually determined by
a mix of at least three ionic components: a potassium conductance (most likely
the sum of 2 distinct conductance fractions) sensitive to TEA and Cs
treatments; a chloride component blocked by typical chloride channels blockers
and cancelled by substituting isethionate for chloride ions; and a leakage
component. The first two conductances exhibit clear-cut voltage-dependent
properties over the membrane potential range commonly considered as passive; a
stable passive behavior is actually obtained only with very high internal
negativity (about –120 mV) when the active conductances vanish and only
the voltage-independent leakage conductance survives. Moreover, the chloride
conductance generates a current that depends not only on the momentary
chloride conductance value but also on transmembrane chloride ion
distribution, which continuously varies as the membrane potential of the
neuron is artificially displaced. It was suggested that if voltage steps are
able to modify chloride conductance and internal chloride ion distribution,
the opposite might also hold true, namely that modifications in chloride
conductance will potentially influence membrane potential values and thus
neuronal excitability and responsiveness. The dynamic interplay, in the
subthreshold potential range, between voltage-dependent conductances and ion
distribution would provide the neuron with an intrinsic mechanism capable of
continuously controlling its resting membrane potential. Thus the latter would
no longer represent a predetermined electrical status, set by basal and
constant ionic and leakage conductances and their respective stationary
equilibrium potentials (as in common models of neuronal resting potential:
Forsythe and Redman 1988;
Jones 1989;
Lamas et al. 2002;
Xu and Adams 1992); the neuron
would instead be able to adjust its resting and active properties to the
continuously modifying conditions under which it is expected to operate.
In the present study, we demonstrate that this mechanism does exist in the
intact and mature rat sympathetic neuron, when the internal ionic medium is
not clamped but is free to readjust to the new voltage gradients. The chloride
conductance exhibits not only the previously demonstrated voltage dependence
over the potential range –40/–120 mV, but also a novel activity
dependence as it strongly increases after even moderate neuronal activity,
while the potassium conductance(s) is virtually unaffected. The cumulative
conductance profiles over voltage of the neuron at rest or after activity are,
therefore profoundly different, and these differences are sustained by
chloride conductance modifications. We are still unable to define the actual
trigger for this chloride conductance increase, but its involvement in
neuronal behavior is expected to be crucial as indicated by the unbalance in
the holding current under voltage-clamp conditions and by the ensuing
modifications in neuronal membrane potential when the latter is free to move.
On the other hand, sympathetic ganglia have represented a useful tool for the
analysis of posttetanic effects under current-clamp conditions, although
confusing and conflicting results were sometimes obtained; see, for example,
reviews by Katayama and Nishi
(1986) and Tokimasa and Akasu
(1995).
The present study represents the first attempt to quantitatively evaluate
and dissect the individual conductances underlying the "resting"
status of the rat sympathetic neuron, and the respective modifications after
physiological activity. It partially answers the starting observation of how
small changes in membrane potential can be amplified by the steepness of the
conductance-voltage relationship of the different currents activatable in the
neuron and how this mechanism might be of physiological interest. In fact,
they help appreciating the functional significance of previous recurrent
observations, highlighting the importance of the basal membrane potential
level as a crucial parameter in defining the mix of voltage-dependent and
synaptic channel types used by the neuron in its firing strategy
(Belluzzi and Sacchi 1988;
Sacchi et al.
1998,Sacchi et al.
1998).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
resistance). Recordings were obtained
under two-electrode current- or voltage-clamp conditions as described in
previous reports (Belluzzi and Sacchi 1985;
Sacchi et al.
1998,Sacchi et al.
1998). Preparations were never utilized for a period >2 h after
surgery. Measurements were performed while the ganglion was continuously
superfused with a standard medium [(in mM) 136 NaCl, 5.6 KCl, 5
CaCl2, 1.2 MgCl2, 1.2 NaH2PO4,
14.3 NaHCO3, 10–2 cholineCl, and 5.5
glucose] pre-gassed with 95% O2-5% CO2 to a final pH 7.3
at 37°C. Occasionally, when the stability of the long-lasting impalement
proved not to be compromised, the more physiological 2.2 mM
Ca2+ concentration was applied. The bath was grounded
through an agar-3 M KCl bridge. Synaptic stimulation of the neuron was obtained by applying single current pulses (0.3-ms duration) of variable strength to the cervical sympathetic trunk, which contains the entire preganglionic input to the neuronal population. Stimuli were also applied in trains of 15 Hz, 10-s duration. Direct stimulation under current-clamp conditions was obtained by applying current pulses of 3- to 5-ms duration and 2–7 nA intensity to the neuron through the current electrode, making minor intensity adjustments during tetanus to maintain them suprathreshold.
When TEACl (tetraethylammonium chloride, Sigma) was used, appropriate
amounts of NaCl were removed from the standard saline composition to maintain
isoosmolarity. Anthracene-9-carboxylic acid (9AC, Sigma) was dissolved in
ethanol at a final 0.8% concentration and bath applied by means of a
continuous and rapid superfusion system. When Cd2+ was
used, the initial bathing medium was switched to a phosphate- and
bicarbonate-free solution buffered with 15 mM Tris-HCl. The instantaneous cell
input conductance was evaluated by applying voltage steps of –40-mV
amplitude and 10-ms duration to the neuron held under voltage-clamp conditions
at membrane potential levels in the –40/–120 mV voltage range.
Alternatively, voltage-ramp commands –50/+40 mV (relative to the
different holding potentials) and 200-ms duration were used with similar
results. It was previously reported that the membrane chord conductance in the
sympathetic neuron displays neither fast voltage sensitivity nor instantaneous
rectification at membrane potential values negative to –50 mV, and the
overall stability of the preparation under two-microelectrode impalement has
been proved (Sacchi et al.
1999).
Attention was systematically paid to avoid any spike discharge from each neuron tested before tetanus application (see following text). This was relevant in the case of synaptic stimulation, which is obligatorily applied to the preganglionic sympathetic trunk and thus involves activation of the whole neuron population. A practical consequence was that only one single neuron could be tested per ganglion.
Long-lasting recordings were filtered at 5 kHz and digitized continuously on tape (Biologic, DTR-1200; 0–10kHz). Data were analyzed on Pentium personal computers (AST) with pCLAMP (Axon Instruments) and MATLAB 386 (The MathWorks, Natick, MA) software packages.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
More than 100 rat sympathetic neurons were analyzed over periods of 
12
min, using the two-microelectrode voltage-clamp technique to measure their
cell input conductance. Figure
1 shows the effects of a conditioning tetanus (15 Hz, 10-s
duration) on neurons stimulated either directly through the current
microelectrode (A) or synaptically via the physiological
preganglionic fibers (B). The cell input conductance was thus
evaluated at –50-mV holding potential under voltage-clamp conditions
before and after (for 10 min) switching to the current-clamp condition for
the short period required to allow stimulation and spike discharge. For the
same neurons, the shifts in holding current level after stimulation are
reported (Fig. 1C). It
is evident that stimulation is followed by a relevant posttetanic increase in
the overall cell input conductance (a mean increase of +71.2% after 10 min in
the directly stimulated neurons, n = 8, vs. +115.8 in the case of
synaptic stimulation, n = 7), paralleled by the onset of an inward
current (see, for example, the initial part of
Fig. 5, A and
B). Despite the large variability from neuron to neuron,
we present the absolute conductance and current values without any attempt to
correct them for cell dimension. However, this general description
qualitatively applies to each single neuron tested here. The effects were long
lasting; they were followed systematically up to 10 min and occasionally even
longer. In isolated observations, the conductance increase vanished within 15
min after the tetanus, whereas in other neurons it was still evident after 30
min.
|
|
The mode of stimulation apparently was irrelevant in determining the time
course of the conductance increase and current modifications (current data
have been pooled). The magnitude of conductance change was slightly greater
after synaptic stimulation; in addition, an early and transient conductance
increase was observed in some neurons (prominent in 3/7 neurons) during the
first minute after synaptic— but not direct— stimulation. This
effect is most likely related to the short-lived activation of a postsynaptic
chloride conductance, exclusively mediated by stimulation of the neuronal
nicotinic receptors (gADPsin) (see
Sacchi et al. 2000).
In preliminary tests, direct stimulation at 15 Hz for 400 ms (7 spikes discharged) was able to evoke a slight increase in conductance in 2/4 neurons, while synaptic tetanus of the same intensity was ineffective in two other neurons. Ten spikes elicited by direct stimulation at 1 Hz were similarly sufficient to induce a detectable raise in conductance. The threshold for conductance increase was thus far below the standard intensity used in these experiments and the final effect on conductance obtained with increasing neuron activation was similarly graded. The standard 10-s tetanization was supramaximal to evoke peak conductance activation because any subsequent tetanization was unable to generate additional effects (see following text). Therefore we preferred to saturate the response to minimize any stimulation pattern-related variability.
Role of spikes and calcium entry on cell input conductance increase
In the preceding experiments, neurons discharged spikes throughout the entire stimulation time. This raises the question of whether spiking is a prerequisite for generating the observed input conductance increase. In a group of experiments, illustrated in Fig. 2A, the synaptic stimulation was applied under voltage-clamp conditions at a constant –50-mV holding potential. Preganglionic stimulation evoked in the neurons excitatory postsynaptic currents (EPSCs) of 20- to 32-nA initial current amplitude, which gradually declined during the train. The ensuing posttetanic effects on cell input conductance were absolutely similar to those observed when synaptic stimulation was performed under current-clamp conditions, both as concerns the short-lived initial transient increase and the final value at 10 min (a mean increase of +107.3%, n = 5). In one of these experiments, the +74.1% posttetanic conductance increase at 10 min vanished within another 5 min with a rebound to –19% of the pretetanic value at 25 min after stimulation.
|
Calcium ions enter the sympathetic neuron very effectively through
voltage-dependent channels during each spike and less efficiently through the
nicotinic synaptic channels during the synaptic current. From simulations in
the ideal rat sympathetic neuron, in fact, the amount of calcium charge
entering the neuron during a single spike arising from a –70-mV basal
membrane potential is 
11 pC, whereas the amount of calcium contaminating
a typical ganglionic EPSC, evoked under voltage-clamp conditions at the same
holding potential, is estimated to be 2–3 pC
(Sacchi et al. 2000). The
possible involvement of internal calcium increase in sustaining the rise in
posttetanic input conductance was examined in experiments in which the
transmembrane calcium movements during spikes evoked by standard direct
stimulation were blocked by 10 µM nifedipine + 0.5 mM CdCl2
(Fig. 2B). In further
experiments, the cell conductance increase, once induced in the neuron, proved
to be insensitive to subsequent modifications in external calcium
concentration in the 2- to 5-mM range (not shown). These results suggest that
the process is external calcium independent.
In these voltage-clamp experiments, trains of negative voltage commands (–40 mV of 10-ms duration) were unable to generate any conductance modification and similarly ineffective were hyperpolarizing current pulses of the same intensity and frequency as those used to directly stimulate neurons under current-clamp conditions. These observations would rule out any effect on cell conductance due per se to artificial current application.
Effects of stimulation persist in the ganglion and do not summate
The two very different stimulation modalities, both resulting in cell input conductance increase, most likely converge on a common cellular mechanism because their effects do not summate when successively applied to the same neuron with supramaximal protocols. Different stimulation patterns were used in single neurons by applying direct or synaptic stimulation under either current- or voltage-clamp conditions in variable sequences (an example is illustrated in Fig. 3A); systematically, the rise in conductance, once fully developed, was completely insensitive to any form of additional stimulation.
|
In principle, the possibility exists that the observed conductance increase might be mediated by experimental handling of the neuron, which hosts two microelectrodes and is exposed to unphysiological current applications. A second point is relevant from a functional point of view, namely whether the synchronous massive synaptic stimulation of a group of neurons is actually able to induce widespread modifications in the membrane properties of all the involved neurons, sufficiently stable and long lasting to be ascertained at random in time after stimulation. This was verified by applying a preliminary standard tetanus to the sympathetic trunk (the unique input that feeds synaptically the whole ganglion neuronal population); thereafter, single neurons were tested by repeating the synaptic stimulation procedure within a few minutes. The results of these experiments are shown in Fig. 3, B and C, in which the time course of input conductance and holding current modifications due to the second tetanus is illustrated. It is evident that the second synaptic stimulation becomes ineffective even in neurons not exposed to preliminary manipulations; in these it was actually possible to ascertain the persistence of the effects of the first tetanus on conductance at various time intervals after its application. The time lapse between the first (neuron not impaled) and second stimulation (neuron impaled) was obviously crucial because the conductance increase is long lasting but reversible. In a fortunate example, entirely performed in normal 2 mM [Ca2+]e, the recovery from the first stimulation effects could be followed by analyzing the response to direct stimulation in three distinct neurons at increasing times from the conditioning tetanus. In the first neuron, the second stimulation was ineffective 18 min after the first tetanus; the second displayed a modest conductance increase (a maximum of +21%) at 30 min, while the third exhibited a strong recovery of the response (+67.1%) after 1 h.
Chloride conductance sustains the cell input conductance increase
Dissection of the conductances underlying the resting status of the
sympathetic neuron is based on pharmacological and ion substitution
experiments (Sacchi et al.
1999). Treatment with K channel blockers (20–50 mM TEACl + 5
mM CsCl) is expected to leave the chloride conductance (plus the constant
leakage conductance) unaffected, while treatment with chloride channel
blockers, or substitution of isethionate for chloride ions, isolates the
compound potassium component. This protocol was used to evaluate the relative
participation of the active conductances in the posttetanic effects.
Figure 4, A and
B, illustrates the effects of 0.5 mM 9AC, a chloride
channel blocker, on the posttetanic conductance and holding current response
in neurons stimulated either directly (n = 4) or synaptically
(n = 4). The results in the two subgroups were similar so that
conductance and current data were pooled. It is evident that the large
posttetanic conductance increase, typical of the normal neuron, and the
associated inward shift of the holding current, were replaced by a small
decrease in conductance and the onset of an outward current
component (compare with Fig. 1, A
and C). This suggests that the posttetanic effects are
sustained in the untreated neuron by a chloride conductance increase (but does
not still rule out the involvement of the potassium conductance; see, however,
the next section). Treatment with 9AC was able not only to prevent the
posttetanic conductance increase but also to reverse it once it had developed
in normal saline. This was systematically observed when cell conductance was
activated by neuron stimulation; an example is illustrated in
Fig. 5, A and
B, in which both the stimulation-related raise in cell
conductance and the holding current inward shifts are not only efficiently
cancelled by 9AC application but also reversed to decreased conductance
compared with the initial value in control saline and outward current.
|
The complementary demonstration that the posttetanic conductance increase is actually sustained by independent components and thus develops also when potassium channels are blocked (TEA + Cs) was complicated by the fact that under these conditions the neuronal spike duration is 30–120 ms and synaptic stimulation proves unpractical. The standard tetanization was therefore replaced by the discharge of 15–20 long-lasting, directly evoked spikes that proved to generate a conductance increase very similar in time course and relative magnitude (the potassium component is now cancelled) to that of normal neurons (Fig. 4C; n = 8).
Effect of stimulation on conductance voltage dependence
The conductance values discussed so far were measured at constant membrane
potential levels. As chloride distribution across the membrane depends on
membrane potential to assess the voltage-dependence of chloride conductance,
the latter must be measured at steady state after the cell has adapted its
chloride content to the new membrane potential level (see following text). We
previously demonstrated that the steady-state values of potassium and chloride
conductances, and the ratios between them, are continuously modified by
membrane potential as it moves in the subthreshold voltage range and only
vanish at membrane potential levels negative to –120 mV. A continuous
description of the active and leakage conductances over the
–40/–120 mV potential range was thus provided
(Sacchi et al. 1999). In those
experiments, there was no suspicion that any of these ganglionic conductances
might also be influenced by previous activity so that no care was taken to
avoid spike discharge in the neurons. They were actually stimulated at 1 Hz
throughout, and several neurons were successively tested in the same
stimulated ganglion. The data presented in the preceding sections urged
reconsideration of the problem: we therefore repeated the old protocol in new
experiments, testing only neurons that had not been stimulated before.
The selective contribution of Cl– and K+
conductances were pharmacologically determined by applying appropriate
antagonists. The dashed lines in Fig.
6 reproduce the conductance values previously reported
(Sacchi et al. 1999) that
actually pertain to stimulated neurons and have been confirmed by new
observations in stimulated neurons (n = 3). New data obtained in
resting neurons are displayed in the same panels and clearly illustrate that
the overall conductance-voltage profile is profoundly different in the resting
and stimulated preparations (A; n = 9); after stimulation,
the increase in cell input conductance, described in the preceding text for
the unique –50-mV membrane potential level
(Fig. 1), has now been verified
over a wide membrane potential range. It is worth noting that the
stimulation-related increase in cell conductance at –50 mV (+66.7%) in
these independent experiments is comparable with the peak posttetanic increase
observed in the directly stimulated neurons of
Fig. 1A (+71.2%).
|
As regards cumulative potassium conductance (9AC treatment, n = 8; Fig. 6B), despite the great variability in neurons and experimental conditions, stimulated or resting neurons exhibited identical voltage dependence in the –40/–70 mV membrane potential range with minor, statistically not significant, differences (see also Table 1) for the –80 and –90 mV values. In two experiments, conductance measurements at rest were repeated after substituting 136 mM Na isethionate for an isoosmotic amount of NaCl (final [Cl–]e =18 mM), and similar conductance-voltage profiles were obtained.
|
On the contrary, large differences were revealed between the resting and stimulated chloride conductance profiles (TEA + Cs treatment, n = 7; Fig. 6C). The shapes of the curves are totally different: the resting chloride conductance exhibits a mild voltage dependence, whereas in stimulated neurons the conductance drastically increases in the –40/–70 mV range. This confirms and extends the results of Fig. 4C.
The virtually complete voltage independence of the leakage conductance component, and its constant low value, are confirmed by the new experiments in the unstimulated neurons (TEA + Cs + 9AC treatment, n = 9; Fig. 6D). In this regard, input conductances as low as 5–7 nS were occasionally measured with the present two-microelectrode technique at –120 mV. These figures favorably compare with the resistance values measured with patch-clamp techniques in mechanically and enzymatically treated neurons of the same origin. Thus the leakage conductance might well be affected by impalement damage, but the usually larger membrane conductance values observed in the intact neuron at less negative membrane potential levels might actually reflect the presence of active conductances, which become unavailable during manipulations required to isolate neurons.
The curves of Fig. 6 were
obtained from different neuron groups not readily comparable in dimensions and
properties. However, the individual ionic conductances, recovered as summated
mean resting gK (Fig.
6B) and gCl
(Fig. 6C) values,
individually corrected for the mean leakage conductance value (19.3 nS,
Fig. 6D), yield a good
recovery of the average total input conductance of the untreated resting
neurons illustrated in Fig.
6A (recovery of 104.2% at –50 mV; 112.4 at
–70 mV; 127.3 at –90 mV; 117.2 at –110 mV), suggesting that
the analysis is reasonably accurate. The systematic excess in the recovery of
the summated conductances might be due to an incomplete blockade of the
individual components, which moreover might not be equally affected at all
membrane potentials levels, and to the presence of minor conductances,
insensitive to the pharmacological treatments employed here. For example, the
A current is mildly inhibited by 10–50 mM TEACl and CsCl
(Belluzzi et al. 1985), whereas
9AC blockade of ClC-1 channels, the channel type presumably sustaining
gCl in this neuron, is strong (
95%) but incomplete
at the concentration used here (Astill et
al. 1996).
In the experiments illustrated in Fig.
6 a "staircase" voltage-clamp protocol was used, in
which the cell input conductance was measured at command voltages increasing
in successive 10-mV steps; the neuron remained at each potential level for a
period of 90 s to allow chloride ion redistribution and the progressive
slow closure of the chloride channels
(Sacchi et al. 1999). These
relatively slow voltage shifts are probably similar to those encountered by
the neuron during its physiological activity. However, two crucial issues
remained to be clarified, namely the sensitivity of posttetanic increase in
gCl to previous history of membrane potential and the
actual lack of posttetanic effects on the isolated gK,
even when tested with large instantaneous voltage steps. Two
–50/–90/–50 mV voltage sequences (the sojourn at –90
mV was now 150 s in duration), separated by the standard direct tetanus under
current-clamp conditions, were successively applied to the same neuron in
normal saline or after 0.5 mM 9AC treatment
(Table 1). The isolated
potassium conductance proved to be insensitive to stimulation, as expected,
and the relatively close numerical values at –50 and –90 mV are in
line with the results of Fig.
6B. The strong voltage dependence of cell conductance
after activity was similarly confirmed in control neurons
(Table 1). The –50-mV
values of posttetanic cell input conductance in control solution, reported in
Table 1, are relatively low
because they were measured 90 s after the tetanus before the posttetanic
effects had fully developed. Data reported in
Table 1 raise an interesting
point: the pretetanus –90-mV values are significantly smaller than those
measured at the same potential during the staircase voltage sequences in
Fig. 6A (24.3 ±
3.4 vs. 54.8 ± 9.3 nS; P < 0.01, Student's
t-test). This observation was drawn from different neurons so that
the figures may not be comparable. However, the mean initial –50-mV
conductance values in the two groups were similar, so the differences might be
meaningful. This would suggest that the simple staircase protocol might be
sufficient to elicit changes in chloride conductance: thus
gCl might not only depend on voltage and previous
suprathreshold activity but also keep track of the mode in which the membrane
voltage potential varies. However, this intriguing point was not further
investigated.
In these experiments, a long-lasting command step of –40-mV amplitude
was instantaneously applied. This markedly displaces chloride from its
equilibrium and makes the chloride currents at the new membrane potential
level more evident. Membrane potential migrations, in fact, generate in the
sympathetic neuron measurable chloride currents, significant chloride
redistribution, and changes in intracellular chloride content and cell input
conductance. Hyperpolarizing steps produce inward chloride currents that
rapidly peak and decay thereafter with long-lasting time constant accompanied
by a parallel decrease in the overall cell input conductance. Examples of
these transient chloride currents during –50/–70/–90 mV
command cycles, recorded in different external media, are illustrated in
panels of Fig. 6. We previously
dissected the contribution of channel gating and chloride redistribution in
determining the relaxation currents, by measuring instantaneous cell input
conductance during the transients (by continuously applying voltage steps of
–40-mV amplitude and 10-ms duration to the neuron held under voltage
clamp) and showed that the two processes occur on similar time scales, but
channel closing in response to the hyperpolarizing step is faster than
chloride redistribution and therefore dominates the kinetics of the current
transients. We suggested that this behavior is sustained by specialized
chloride channels that are open at the moment in which the negative
voltage step is applied and then close during maintained
hyperpolarization according to a voltage-dependent rate constant
(Sacchi et al. 1999). These
processes, however, develop phenomenologically over a long time course and
become faster with increasing membrane polarization: the mean time constants
measured in previous experiments proved to be of 31–10 s in the
–50/–130 mV voltage range. This requires that the correct
conductance value pertinent to each membrane potential level has to be
measured only when the new steady-state is reached, usually 90–150 s
after imposing the new voltage command. The currents generated by the large
and long-lasting voltage steps used in the experiments of
Table 1 are particularly suited
to evaluate kinetic aspects of the readjustments occurring in the neuron. We
report here that gCl voltage dependence is low in the
unstimulated neuron (Fig.
6C) and markedly increases after stimulation. The
additional pertinent observation is that the decay time constant of
ICl is slow in the resting neurons (38.4 ± 4.0 s,
n = 8) and becomes significantly faster in the stimulated neurons
(22.9 ± 2.1 s; n = 8. P < 0.01). This indicates
that stimulation affects not only the final gCl magnitude,
but also its kinetic properties.
Posttetanic effects simulated in a neuron model
Conductances and ionic gradients are mutually linked in the neuron at rest.
All the conductances operating in the ganglion neuron have been previously
characterized experimentally by defining their magnitude and the voltage and
time dependence of their activation and inactivation properties
(Belluzzi and Sacchi 1991). The
resulting set of continuous equations, over a wide membrane potential range,
was included in a comprehensive mathematical model
(Sacchi et al.
1998,Sacchi et al.
1998). The voltage-dependent chloride conductance here addressed
was similarly characterized in neurons continuously stimulated at 1 Hz
(Sacchi et al. 1999), and its
contribution to neuron behavior was added to the model. Over the membrane
potential range here considered (–40/–120 mV), the main features
of the resulting model can be simplified as illustrated in the electrical
equivalent of Fig. 7B;
this scheme has been presented in a previous paper
(Sacchi et al. 1999) and is
reported here for clarity. Briefly, the new findings are represented by a
voltage-dependent compound potassium conductance fed by a constant potassium
battery and the novel voltage-dependent chloride conductance fed by a
voltage-dependent chloride battery, which continuously varies with membrane
potential migrations. This model does not take into account additional minor
current components insensitive to the blockers used and electrogenic pump
contributions, which have been considered to remain constant over time and
membrane potential level; it provides, however, a useful tool for the
understanding of the mode of interaction of the three conductances analyzed in
the present study in the subthreshold region. All the variables depicted have
been experimentally estimated (neuron stimulated) and mathematically
characterized by continuous equations over a wide membrane potential range;
they have now been complemented with new data concerning the truly
resting neuron (see legend to Fig.
7B). Simulations depicted in the initial part of
Fig. 7A predict the
values of resting chloride and potassium currents, and chloride equilibrium
potential, for resting potentials of either –50 or –70 mV. The
remaining part of the curves in Fig.
7A shows the effects of an instantaneous increase in
gCl elicited by ongoing 1-Hz spike discharge in an ideal,
initially silent sympathetic neuron (according to data in
Fig. 6C). This
simulation dissects the complicated mutual adjustments of the multiple
variables that contribute to accommodate the neuron to varying
voltage-dependent conductances and subthreshold membrane potential levels.
Some of the effects are not readily predictable: for example, the final
steady-state increase in cell input conductance at –70 mV is lower than
expected from the gCl amount initially imposed (100.3 vs.
109.3 nS), due to a decrease in gK after membrane
depolarization from –70 to –61.4 mV (see
Fig. 6B).
|
Raised conductance and neuron depolarization are expected to engender potentially conflicting results because the first effect acts to stabilize the membrane, while the second moves the neuron toward its firing threshold. Simulations help to clarify this point. The resulting changes in neuronal excitability, expressed in terms of the threshold inward charge required to fire the neuron, are illustrated in the Fig. 7, bottom. In this example, membrane potential (depolarized by the increased gCl from a basal –70 mV level to –63 mV) is the major controller of neuron excitability, whereas input conductance produces appreciable but short-lived effects only in the moment in which it is abruptly raised. A different picture is seen when simulation is repeated at a more positive membrane potential. The increase in gCl is large (and cell input conductance) in the neuron stimulated at –50 mV and is stable over time, but it generates a membrane depolarization of 3.0 mV, which is only accompanied by a minor drop in threshold charge. The starting membrane potential level thus exerts a crucial role in conditioning not only resting excitability but also the final poststimulation effects, which vary in a complex, nonlinear manner and are difficult to test experimentally.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
). We describe here an additional crucial property,
exclusive of the chloride conductance, namely its dependence on previous
neuronal activity. Activation of the neuron results in a consistent, long-term
increase in cell input conductance, specifically sustained by the chloride
channel, that also acquires marked voltage dependence. It may be of interest
to observe that in the absence of previous stimulation chloride conductance
only varies by 40% over the –90- to –40-mV membrane potential
range, and this may help explaining why the existence of an active chloride
conductance in the subthreshold region of membrane potential had so far
escaped experimental detection. The mode of neuron activation appears
relatively unimportant as concerns the recording mode (current- or
voltage-clamp), the activation mode (direct or synaptic stimulation), and the
stimulation sequences or frequencies applied. In the present experiments, the
neuron was usually activated by high-frequency tetanization, but a maintained
preganglionic stimulation frequency as low as 1 Hz, well compatible with the
ongoing discharge frequencies operating physiologically in the autonomic
nervous system (reviewed by Jänig
1995), is suitable for its activation. We suspect that mere
membrane potential shifts in the subthreshold voltage region might be
sufficient to elicit a detectable response. The effect is long-lasting (it was
followed occasionally for up to 30 min and more), but transient and may revert
to decreased conductance, suggesting that the momentary chloride conductance
value represents a balance between activation and turn-off processes. The
neuron input conductance is expected to vary in the subthreshold range within
the physiological limits outlined by the curves in
Fig. 6A: between the
curve pertinent to the unstimulated neuron and that of the neuron synaptically
activated at 1 Hz. Such limits could possibly be even broader in the case of
strong synaptic tetanization, especially at membrane potential levels negative
to –70 mV, where 1-Hz stimulation appears less effective. The main question is what exactly does resting mean in this context. The absence of spike discharge during the experiment makes it implicit that the neuron has not been stimulated, but this gives no insight into its previous history. We have actually no information about the real memory of this process in the operating neuron, or of the modalities for cancelling any previous level of gCl activation. Experiments are being planned on denervated neurons in which synaptic stimulation and firing will be abolished for a controlled extent of time.
The cellular mechanism underlying this behavior is unidentified; we prove that neither membrane potential migrations associated with spike discharge, nor synaptic activity by itself, or external calcium entry into the cell, represent the required trigger for its induction and development. The molecular bases of the signaling pathway, and the reasons for the increased voltage dependence of the resting conductance and the acceleration of its kinetic properties following stimulation, however, remain to be elucidated.
It will be noted that a chloride-mediated posttetanic control may bear some
relevance to the neuronal functioning: two independent chloride-related
mechanisms actually operate in sequence in the sympathetic neuron. The first
is sustained by an early chloride conductance increase (of 20 nS)
mediated exclusively through activation of the nicotinic channels by native
acetylcholine. It generates a relatively long-lasting current (it decays with
a mean time constant of 370 ms), IADPSsyn, that
exhibits fast onset, is calcium-independent and insensitive to specific
chloride channel blockers (Sacchi et al.
2000). The second mechanism described here partially over-laps the
first (see points at 1 min after tetanization in Figs.
1B and
2A) but then prolongs
the effects initially brought about by pure nicotinic channel activation.
The presence of a persistent chloride current governing the resting status of the sympathetic neuron has been largely over-looked: it is of limited magnitude and mixed with other currents; it requires that its driving force be artificially increased by voltage steps before it becomes discernible; the internal chloride ion concentration must not be clamped but free to readjust according to the momentary membrane potential level; the two alternative gCl conditions, unstimulated-stimulated, have never been taken into account before. The importance of the chloride current arises from the equilibrium that persists in the resting neuron between out- and inward currents; one of the inward components is actually the chloride current, which is kept constantly inward by the internal chloride concentration, which is in turn maintained higher than its electrochemical equilibrium by active transport mechanisms. The potassium conductance is potentially voltage dependent, but will remain stable until an external event modifies its value; this trigger is not represented by any of the neuron activation procedures tested here. At steady state, any constant membrane potential level will thus generate a constant potassium current, being fed by a constant potassium battery. On the contrary, the chloride channel hosts an intrinsic mechanism, its activity dependence, which makes the chloride current a candidate for natural controller of the balance between the opposite resting currents, and thus of the final membrane potential level. It might represent an intrinsic mechanism to continuously adjust neuron basal status to its activation degree.
Previous isolated observations are in line with this view. GABA depolarizes
the rat sympathetic neuron (Adams and Brown
1975) by inducing an inward current when applied at membrane
potentials negative to the chloride equilibrium potential
(Sacchi et al. 1999). It has
been reported that isethionate substitution for chloride ions reduces cell
input conductance, as measured in current-clamp
(Adams and Brown 1975) or
voltage-clamp conditions (Sacchi et al.
1999), and hyperpolarizes the neuron
(Adams and Brown 1975); a
similar result has been obtained here by pharmacologically blocking the
chloride channels (not shown). Moreover, a distinct inward current
accompanying gCl increase has been recorded in the present
experiments under voltage-clamp conditions. Less recent observations, under
current-clamp conditions, have demonstrated that ganglionic tetanization was
followed by a sequela of membrane potentials shifts; one of these, the
"late slow excitatory postsynaptic potential" exhibited a slow
onset, an extremely prolonged time course, and a sufficient magnitude to
induce a long-lasting late afterdischarge in bullfrog neurons
(Nishi and Koketsu 1968). The
existence of an analogous response in sympathetic ganglia of the guinea pig
and rabbit has been demonstrated by Ashe and Libet
(1981). This latter description
is phenomenologically reminiscent of the present findings: the late
depolarization was resistant to nicotinic, muscarinic, or adrenergic
antagonists; it appeared with latencies in the range of seconds, rise times in
the minute time scale, duration of up to 20 min or more; even a 1-Hz
stimulation frequency applied in a sufficiently large number of pulses was
almost as effective as higher frequency tetanization. These results have not
actually been linked to chloride participation. The new findings presented in
this and the preceding paper (Sacchi et
al. 1999) would suggest that the envisaged mechanism associated
with gCl represents one of the physiological controllers
of the neuronal excitability machinery in which activity would control further
activity by modifying the cell input conductance and the basal membrane
potential level.
The active chloride movements, and their control, are a homeostatic process
not completely understood (Delpire
2000; Jentsch
1996; Misgeld et al.
1986; Rohrbough and Spitzer
1996; Russell
2000). Chloride conductance, on the other hand, plays important
roles in regulating excitability and the equilibrium between excitation and
inhibition. In muscle, the high resting chloride conductance and the passive
distribution of chloride ions suppress depolarizing inputs and stabilize the
membrane potential at its negative level
(Palade and Barchi 1977). In
neurons, active chloride distribution represents an important controller of
interneuronal communication mediated by GABA synapses: the developmental
shifts in the activity of chloride-related cotransporters control the
distribution of chloride ions; the ensuing equilibrium potential may vary
widely and lead to relevant changes in neuronal response to synaptic
activation. During development, responses to GABA switch in fact from
depolarization to hyperpolarization (reviewed by
Ben-Ari 2002). If an
activity-dependent gCl increase were a general, widespread
mechanism in other neurons as well, the posttetanic conductance increase would
generate similar effects in all of them by affecting their passive properties;
the final physiological result, however, would be conditioned by chloride
transmembrane distribution. In the sympathetic neuron, chloride ions are
constantly maintained at an internal concentration higher than predicted by
the nernstian equilibrium, by active mechanisms that move chloride ions into
the cell. The opposite holds true for motoneurons and CNS neurons in which
active chloride extrusion frequently occurs. Under these conditions,
gCl modifications will generate membrane potential shifts
in opposite directions, from the resting status, according to the momentary
chloride driving force. This principle is well documented for the short-lived
synaptic conductance modifications, occurring for example in motoneurons
(Coombs et al. 1955), in
hippocampal pyramidal neurons (Staley
1994), in spinal neurons
(Rohrbough and Spitzer 1996),
and in rat auditory neurons (Ehrlich et
al. 1999). It would also be applicable when modifications of the
mechanisms controlling chloride distribution develop with a long time course,
as in the present case, with widespread and long-term effects on neuronal
firing strategy. Interestingly, the presence of anion channels has been
demonstrated in rat brain synaptosomal membranes, and their possible role in
mediating transmitter release by controlling the polarization of the
presynaptic terminal membrane has been suggested
(Nomura and Sokabe 1991;
Yuto et al. 1997). These data
open a series of questions of general interest concerning the role of chloride
in synaptic signal processing, the mechanisms sustaining the chloride
gradients in different conditions, and the relationship between chloride
conductance and activity. They would stress, anyway, the relevance of the
processes governing chloride ion movements in distinct compartments and during
different aspects of neuron functioning; they would also explain apparently
conflicting posttetanic effects, which have been observed in the absence of a
detailed description of the actual chloride ion dynamics and knowledge of the
history of previous activity.
|
DISCLOSURES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests: O. Sacchi, Dept. of Biology, Section of Physiology and Biophysics, Via Borsari 46, I-44100 Ferrara, Italy (E-mail: sho{at}dns.unife.it).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
-aminobutyric acid on sympathetic ganglion
cells. J Physiol 250:
85–120, 1975.Alvarez-Leefmans FJ, Gamiño SM, Giraldez F, and Noguerön I. Intracellular chloride regulation in amphibian dorsal root ganglion neurons studied with ion-selective microelectrodes. J Physiol 406: 225–246, 1988.
Ashe JH and Libet B. Orthodromic production of non-cholinergic slow depolarizing response in the superior cervical ganglion of the rabbit. J Physiol 320: 333–346, 1981.
Astill DStJ, Rychkov G, Clarke JD, Hughes BP, Roberts ML, and Bretag AH. Characteristics of skeletal muscle chloride channel ClC-1 and point mutant R304E expressed in Sf-9 insect cells. Biochim Biophys Acta 1280: 178–186, 1996.
Ballanyi K and
Grafe P. An intracellular analysis of -aminobutyric acid-associated
ion movements in rat sympathetic neurones. J Physiol
365: 41–58,
1985.
Belluzzi O, Sacchi O, and Wanke E. A fast transient outward current in the rat sympathetic neuron studied under voltage-clamp conditions. J Physiol 358: 91–108, 1985.
Belluzzi O and Sacchi O. The interactions between potassium and sodium currents in generating action potentials in the rat sympathetic neuron. J Physiol 397: 127–147, 1988.
Belluzzi O and Sacchi O. A five-conductance model of the action potential in the rat sympathetic neuron. Prog Biophys Mol Biol 55: 1–30, 1991.
Ben-Ari Y. Excitatory actions of GABA during development: the nature of the nurture. Nat Rev Neurosci 3: 728–739, 2002.
Coombs JS, Eccles JC, and Fatt P. The specific ionic conductances and the ionic movements across the motoneuronal membrane that produce the inhibitory post-synaptic potential. J Physiol 130: 326–373, 1955.
Delpire E. Cation-chloride cotransporters in neuronal communication. News Physiol Sci 15: 309–312, 2000.
Ehrlich I, Löhrke S, and Friauf E. Shift from depolarizing to hyperpolarizing glycine action in rat auditory neurones is due to age-dependent Cl– regulation. J Physiol 520: 121–137, 1999.
Forsythe ID and Redman SJ. The dependence of motoneuron membrane potential on extracellular ion concentrations studied in isolated rat spinal cord. J Physiol 404: 83–99, 1988.
Galvan M, Dörge A, Beck F, and Rick R. Intracellular electrolyte concentrations in rat sympathetic neurones measured with an electron microprobe. Pfluegers 400: 274–279, 1984.
Jänig WM. Ganglionic transmission in vivo. In: Autonomic ganglia, edited by McLachlan EM. Luxembourgh: Harwood, 1995, p. 349–395.
Jentsch TJ. Chloride channels: a molecular perspective. Curr Opin Neurobiol 6: 303–310, 1996.
Jones SW. On the resting potential of isolated frog sympathetic neurons. Neuron 3: 153–161, 1989.
Katayama Y and Nishi S. Peptidergic transmission. In: Autonomic and Enteric Ganglia, edited by Karczmar AG, Koketsu K, and Nishi S. New York: Plenum, 1986, p. 181–200.
Lamas JA, Reboreda A, and Codesido V. Ionic basis of the resting membrane potential in cultured rat sympathetic neurons. Neuroreport 13: 585–591, 2002.
Misgeld U, Deisz RA, Dodt HU, and Lux HD. The role of chloride transport in postsynaptic inhibition of hippocampal neurons. Science 232: 1413–1415, 1986.
Nishi S and Koketsu K. Early and late after-discharges of amphibian sympathetic ganglion cells. J Neurophysiol 31: 109–121, 1968.
Nomura K and Sokabe M. Anion channels from rat brain synaptosomal membranes incorporated into planar bilayers. J Membr Biol 124: 53–62, 1991.
Palade PT and Barchi RL. Characteristics of the chloride conductance in muscle fibers of the rat diaphragm. J Gen Physiol 69: 325–342, 1977.
Rohrbough J and Spitzer NC. Regulation of intracellular Cl– by Na+-dependent Cl– cotransport distinguishes depolarizing from hyperpolarizing GABAA receptor-mediated responses in spinal neurons. J Neurosci 16: 82–91, 1996.
Russell JM. Sodium-potassium-chloride cotransport. Physiol Rev 80: 211–276, 2000.
Sacchi O, Belluzzi O, Canella R, and Fesce R. A model of signal processing at a mammalian sympathetic neuron. J Neurosci Methods 80: 171–180, 1998.
Sacchi O, Rossi ML, and Canella R. Synaptic stimulation of nicotinic receptors in rat sympathetic ganglia is followed by slow activation of postsynaptic potassium or chloride conductances. Eur J Neurosci 12: 2651–2661, 2000.
Sacchi O, Rossi ML, Canella R, and Fesce R. Synaptic current at the rat ganglionic synapse and its interactions with the neuronal voltage-dependent currents. J Neurophysiol 79: 727–742, 1998.
Sacchi O, Rossi ML, Canella R, and Fesce R. Participation of a chloride conductance in the subthreshold behavior of the rat sympathetic neuron. J Neurophysiol 82: 1662–1675, 1999.
Staley K. The role of an inwardly rectifying chloride conductance in postsynaptic inhibition. J Neurophysiol 72: 273–284, 1994.
Tokimasa T and Akasu T. Biochemical gating for voltage-gated channels: mechanisms for slow synaptic potentials in autonomic ganglia. In: Autonomic Ganglia, edited by McLachlan EM. Luxembourgh: Harwood, 1995, p. 259–295.
Xu Z-J and Adams DJ. Resting membrane potential and potassium currents in cultured parasympathetic neurons from rat intracardiac ganglia. J Physiol 456: 405–424, 1992.
Yuto J, Ide T, and Kasai M. ATP-sensistive anion channel from rat brain synaptosomal membranes incorporated into planar lipid bilayers. Biophys J 72: 720–727, 1997.
| ||||||||||||||||||||||||||||||||||||||||||||||||
