Download Do neurons have a reserve of sodium channels for the generation of

European Journal of Neuroscience, Vol. 12, pp. 1±7, 2000
ã European Neuroscience Association
Do neurons have a reserve of sodium channels for the
generation of action potentials? A study on acutely
isolated CA1 neurons from the guinea-pig hippocampus
Michael Madeja
Institute for Physiology, University of MuÈnster, Robert-Koch-Str. 27 A, D-48149 MuÈnster, Germany
Keywords: density of channels, frequency, recovery, repetitive ®ring, tetrodotoxin
Abstract
The density of voltage-gated sodium channels is high in several regions of the neuronal membrane. It is unclear if this density of
channels represents a reserve for the neuron, or if it ful®ls a special role in action potential ®ring. This problem was addressed by
studying sodium currents and action potentials in acutely isolated hippocampal CA1 neurons whose number of active sodium
channels was acutely changed by applying the sodium channel blocker tetrodotoxin (TTX) at different concentrations. The results
show that more than a third of the sodium channels can fail without affecting the single action potential. Thus, the neurons have a
remarkable surplus of sodium channels. The surplus, however, is necessary for repetitive action potential ®ring, as every decrease in
the fraction of sodium channels reduces the maximal frequency of action potentials that can be generated by the neuron.
Introduction
The generation of action potentials is one of the fundamental processes
in nervous systems. Although the general mechanisms of the action
potential and especially the role of the voltage-gated sodium channels
have been known for several decades (see Kandel, 1976), many
problems are still unsolved even now. A question of particular
importance concerns the number of sodium channels required for the
generation of action potentials. Thus, it is unknown if every reduction
of the number of sodium channels impairs action potential generation
or if neurons have a reserve that allows failure of a fraction of the
sodium channels without impairing the generation of action potentials.
An answer to this question is not only of interest for a further
understanding of the basic principles of neuronal activity, but might
have consequences for judging the relevance of sodium channel
blockade by pharmaceutical drugs and for a better understanding of
pathological conditions leading to a reduction of sodium channels.
In order to contribute to a solution of this problem, a neuronal cell
type was treated with tetrodotoxin (TTX) at different concentrations.
Because TTX is a speci®c blocker of the voltage-gated sodium
channels (Narahashi, 1974; Catterall, 1980; Ulbricht, 1981), the
momentary number of sodium channels available for the generation
of action potentials could be titrated by applying different
concentrations of TTX. Thus, action potential generation could be
studied in a neuronal preparation in which the number of active
sodium channels was set to different levels.
Materials and methods
Nerve cell isolation
The experiments were carried out on hippocampal neurons of adult
guinea-pigs (weight, 300±400 g). The brain was removed during ether
Correspondence: Dr M. Madeja, as above.
E-mail: [email protected]
Received 10 May 1999, revised 4 August 1999, accepted 2 September 1999
anaesthesia. The hippocampus was dissected and transverse slices
(thickness, 400±500 mm) were cut parallel to the alvear ®bres with a
McIlwain tissue chopper. The neurons were acutely isolated
according to the technique of Kay & Wong (1986), modi®ed as
described by Vreugdenhil & Wadman (1995). In short, the regions
containing CA1 neurons were dissected into sections of ~ 1 mm2.
These tissue pieces were incubated for 75 min at 32 °C in oxygenated
dissociation solution containing (in mmol/L): NaCl, 120; KCl, 5;
PIPES, 20; CaCl2, 1; MgCl2, 1; D-glucose, 25; 1 mg/mL trypsin
(40 U/mg, Merck, Darmstadt, Germany), pH 7.0. Thereafter, the
tissue was washed with enzyme-free solution and kept at room
temperature. Neurons were isolated by triturating the tissue pieces
through a series of Pasteur pipettes with decreasing tip diameter.
After settling onto the bottom of the recording chamber, neurons
which appeared bright and smooth under the microscope and had no
visible organelles were selected for recording.
Electrophysiological techniques
Voltage-clamp recordings were performed in the whole-cell patchclamp con®guration (Hamill et al., 1981). Patch pipettes were pulled
from borosilicate glass and had resistances of 2±3 MW. The pipettes
were ®lled with an intracellular recording solution (in mmol/L): KF,
140; NaCl, 2; CaCl2, 1; MgCl2, 2; N-2-[hydroxyethyl]piperazineN¢-[2-ethanesulphonic acid] (HEPES), 10; ethylene glycol-bis(baminoethyl ether)-N,N,N¢,N¢-tetraacetic acid (EGTA), 10; MgATP,
0.5, pH 7.4. The acutely isolated neurons were superfused with an
extracellular solution of the following composition (in mmol/L):
NaCl, 130; KCl, 5.4; CaCl2, 1.8; MgCl2, 1; D-glucose, 25; HEPES,
10, pH 7.4. TTX (0.1 nmol/L up to 5 mmol/L; Sigma, Deisenhofen,
Germany) was added to the extracellular solution and was applied
with a perfusion pipette directed at the recorded neuron. The TTX
solutions were applied for at least 30 s before starting electrophysiological recordings. To avoid signi®cant rundown, each experimental
protocol was completed within 10 min. Thus, for the measurement of
2 M. Madeja
the recovery of action potentials and maximal discharge frequency,
only four different TTX concentrations were tested.
Sodium currents and action potentials were recorded with a EPC-7
ampli®er (List Electronic, Darmstadt, Germany). Deformations of
action potentials have been shown for recordings with patch-clamp
ampli®ers (Magistretti et al., 1996). The used EPC-7 ampli®er was
tested by applying voltage pulses with a resistance of 2 MW and a
capacitance of 3 pF in order to model the patch pipette. Rectangular
pulses showed a transient overshoot of 23% of the maximum voltage
amplitude which remained constant in the tested potential range from
10 mp to 100 mV. Thus, the absolute amplitude of the action potential
is not measured correctly, but the error can be assumed not to
contribute to the relative changes in action potential amplitude. The
bandwidth of the EPC-7 ampli®er (determined as the frequency at
which a sine wave was reduced to 50%) was 9.0 kHz. The maximum
rise time of the ampli®er was 4.64 mV/ms. Because the maximum rise
time of the action potentials was measured with ~ 0.9 mV/ms (see
Results), it can be assumed that the rise of the action potential is
suf®ciently well recorded by the ampli®er.
For the recording of sodium currents, the capacitive transients and
series resistances were compensated by > 80%. After seal formation
and membrane rupturing, the nerve cells were allowed to stabilize for
3 min before starting the pulse protocols. The holding potential was
±80 mV and voltage steps to ±20 mV were applied (duration 5 ms,
interpulse interval at least 4 s; for the measurement of recovery from
inactivation the interpulse interval was varied from 8 to 80 ms). For
the recording of action potentials, constant currents (up to 0.8 nA)
were applied to set the resting membrane potential to ±80 mV or more
negative. Short current pulses (duration 2 ms, interpulse intervals
10 ms up to 4 s) with an amplitude of 0.5 nA were applied to elicit
single action potentials or pairs of action potentials. Long current
pulses (duration 600 ms, interpulse interval 4 s) with increasing
amplitudes from 0.02 nA up to 0.2 nA were used to elicit repetitive
action potential ®ring. All experiments were carried out at room
temperature (22 6 1 °C).
Data acquisition and analysis
Currents were ®ltered with an eight-pole Bessel ®lter at a frequency
of 10 kHz and were transferred to a computer system (pCLAMP
program, Axon Instruments, Foster City, USA). Leakage currents
were subtracted online using the p/4 method (Bezanilla & Armstrong,
1977). The peak amplitudes of sodium currents were measured and
normalized to values under control conditions.
The amplitudes of action potentials were obtained after subtracting
the passive membrane potential changes caused by current injection.
The passive membrane potential changes were measured after
complete blockade of the sodium currents with 5 mmol/L TTX or
were calculated by exponential extrapolation of the passive
membrane potential changes measured in the sub-threshold potential
range. The maximum rate of rise was determined by differentiation of
the action potential (Schwarz et al., 1973). The action potential
duration was measured at the half-maximal amplitude.
Dose±response relations for the effects of TTX on the amplitudes
of the sodium currents and action potentials were determined by
®tting mean values to the Langmuir equation: y = (Km/c)n/[1 + (Km/
c)n], where y is the fraction of the control value, Km is the dissociation
constant, c is the concentration of TTX and n is the Hill coef®cient.
The maximal discharge frequency of action potentials was ®tted with
the modi®ed equation: y = a((Km/c)n/(1 + (Km/c)n)), where y is the
discharge frequency and a is the maximal value of the ®t. All other
curve ®ts were obtained by applying monoexponential functions.
Curve ®ttings and all mathematical procedures were obtained using
the program SigmaPlot (Jandel Scienti®c, Erkrath, Germany). The
data are given as mean 6 SEM.
Results
The investigated cells were acutely isolated CA1 neurons from the
guinea-pig hippocampus. Cells of similar size and shape (diameter of
soma ~ 20 mm, length of apical and basal dendritic stumps ~ 50 mm)
were chosen for electrophysiological recording. In the whole-cell
con®guration, most of the cells rounded up on the tip of the electrode,
allowing an estimation of their surface area. Assuming a spherical
model, the mean surface area was calculated to be 1420 6 110 mm2
(n = 10).
Because no protease inhibitor and only a low ATP concentration
were added to the intracellular solution, the calcium currents ran
down completely within 3 min after membrane rupturing (Madeja
et al., 1997). Thus, after blockade of the sodium currents with
1000 nmol/L TTX, no inward current was left indicating that the
calcium currents were abolished, and that in contrast to other cell
types (Scholz et al., 1998), TTX-resistant sodium currents were
absent (Fig. 1A). Although the potassium currents could not be
blocked due to their role in the repolarization of the action potential,
they can be assumed to interfere not signi®cantly with the
measurement of sodium currents as the peak of the sodium currents
appeared before the potassium currents developed (arrow in Fig. 1A).
The sodium current had a mean amplitude of 10.4 6 1.0 nA at a
potential of ±20 mV (n = 10).
Generation of single action potentials
In order to test how the reduction of sodium channels affects the size
and shape of single action potentials, sodium currents and action
potentials were elicited at various TTX concentrations. The
recordings revealed that the sodium currents were more sensitive to
TTX than the action potentials (Fig. 1A). Whereas 10 nmol/L TTX
reduced the sodium current to less than half of the control value, the
action potential amplitude was only slightly affected. Furthermore,
with the concentration of 100 nmol/L TTX, which reduced the
sodium current to < 6% of the control value, approximately half of the
action potential amplitude was left (Fig. 1A and B).
The different sensitivities are re¯ected in the dose±response curves
of sodium currents and action potential amplitudes (Fig. 1B).
Whereas the dose±response relation of the sodium currents has an
IC50 value of 6.4 nmol/L (Fig. 1B, ®lled circles) and a Hill coef®cient
of 0.91, the curve of the action potential amplitude is shifted to higher
concentrations and reveals an IC50 value of 104 nmol/L (Fig. 1B,
open circles). Furthermore, the curve has a steeper slope with a Hill
coef®cient of 1.23, indicating the threshold-dependent, snowballing
nature of the action potential.
Further parameters of action potential shape, the duration of the
action potential and (the most sensitive parameter) the maximum rate
of rise, were analysed (Fig. 1C). The maximum rate of rise under
control conditions was 0.92 6 0.09 mV/ms (n = 21). Whereas with
TTX the ®rst signi®cant decrease (i.e. a change of > 5% of control)
was obtained for the amplitude of the sodium current with 0.3 nmol/L
TTX, the ®rst signi®cant decrease of the action potential amplitude
appeared with 11.2 nmol/L TTX, the duration of the action potential
with 6.5 nmol/L TTX, and its maximum rate of rise with 3.1 nmol/L
TTX. At these concentrations, the sodium current was reduced to 36,
50 and 66% of the control value, respectively. Taking the most
sensitive parameter, it can be concluded that in these neurons more
than a third of the sodium current has to be blocked before generation
of single action potentials is affected.
Ó 2000 European Neuroscience Association, European Journal of Neuroscience, 12, 1±7
Do neurons have a reserve of sodium channels?
3
FIG. 1. Effects of TTX on sodium currents and single action potentials of acutely isolated CA1 neurons from the guinea-pig hippocampus. (A) Original recordings
of sodium currents (I, upper traces) and action potentials (MP, lower traces) under control conditions and at TTX concentrations from 0.1 up to 1000 nmol/L.
Currents were elicited from a holding potential of ±80 mV with voltage steps to ±20 mV (duration 5 ms, interpulse interval 4 s). The recordings are corrected for
leakage currents. Action potentials were elicited from a membrane potential of ±84 to ±91 mV by current injections of 0.5 nA (duration 2 ms, interpulse interval
4 s). The passive membrane potential changes caused by the current injection (obtained at complete blockade of the sodium currents in 5 mmol/L TTX) are
subtracted. The arrow in the recording with 1000 nmol/L TTX indicates the time of the peak sodium current under control conditions. (B) Dose±response curves for
the effects of TTX on the amplitudes of sodium currents (®lled circles) and action potentials (AP, open circles). The amplitudes are normalized to control values.
Each data point represents the mean 6 SEM of seven to nine experiments. The mean values were ®tted with Langmuir equations. The IC50 values and Hill
coef®cients were 6.4 nmol/L and 0.91 for the sodium currents, and 104 nmol/L and 1.23 for the action potentials, respectively (see dotted lines). (C) Change of
action potential (AP) amplitude (open circles), duration (open diamonds), maximum rate of rise (open squares) and sodium current amplitude (®lled circles) at
increasing TTX concentrations. All parameters are normalized to control conditions. Each data point represents the mean 6 SEM of ®ve to seven experiments. The
action potential parameters were ®tted with exponential functions; the ®t of the sodium current was obtained from B. A signi®cant change (> 5% of control) was
obtained for the amplitude of the sodium current with 0.3 nmol/L TTX. The respective values for the maximum rate of rise of the action potential were 3.1 nmol/L
TTX; for the duration of the action potential, 6.5 nmol/L TTX; and for the amplitude of the action potential, 11.2 nmol/L TTX (see dotted lines). At these
concentrations, the sodium currents were reduced by 34 to 64% of control (see dotted lines).
Generation of pairs of action potentials
To investigate if the failure of a small number of the sodium channels
impairs successive action potentials, pairs of action potentials were
elicited at those TTX concentrations at which the single action
potential was not affected. In a ®rst step, pairs of action potentials
were elicited with a ®xed interpulse interval in increasing TTX
concentrations. As can be seen from the original recordings in
Ó 2000 European Neuroscience Association, European Journal of Neuroscience, 12, 1±7
4 M. Madeja
FIG. 2. Effects of TTX on recovery after activation of the action potential. (A) Original recordings of pairs of action potentials under control conditions (CTRL),
and with 0.1, 1 and 10 nmol/L TTX. Current pulses (I, amplitude 0.5 nA, duration 2 ms) with an interpulse interval of 10 ms were applied at a membrane potential
(MP) of ±87 to ±94 mV. Due to the exponential decay of the stimulation artefact, the second current pulse of a pair was elicited at a MP of ±70 to ±71 mV. (B)
Recovery of the action potential at different interpulse intervals (10 to 64 ms) under CTRL (®lled circles) conditions, and with TTX at concentrations of 0.1 nmol/L
(open circles), 1 nmol/L (®lled squares) and 10 nmol/L (open squares). The amplitudes of the second action potential of a pair were normalized to the action
potential amplitude after complete recovery. Each data point represents the mean 6 SEM of six to eight experiments. The mean values were ®tted with exponential
equations. The inset shows superimposed original recordings of pairs of action potentials under control conditions and with 10 nmol/L TTX. The dotted line shows
the exponential ®t of the mean sodium currents' recovery from inactivation under control conditions of four experiments. (C) Time constants (t) for the recovery of
the action potential under CTRL and with TTX at concentrations up to 10 nmol/L. The time constants were obtained from the experiments and ®ts shown in B, and
have corresponding symbols. The dotted line indicates the time constant at the highest concentration of TTX found to have no effect on single action potentials. At
this concentration, the time constant of recovery was increased by 39% relative to the control.
Fig. 2A, the second action potential of a pair was reduced in
amplitude by TTX even at a concentration of 1 nmol/L (Fig. 2A,
arrow). At this concentration, the shape of single action potentials
remained unchanged (cf. Fig. 1C).
This impairment of the second action potential of a pair suggests an
effect on the recovery after activation (i.e. the recovery of the sodium
channel activation and the voltage-dependent availability of the
potassium channels). Pairs of action potentials were therefore elicited
at different interpulse intervals (Fig. 2B). Graphical evaluation of the
mean action potential amplitudes under control and TTX revealed a
slowing of the recovery with increasing TTX concentrations. The
time constants of the curves rose from 9.9 ms under control
conditions to 15.6 ms with 10 nmol/L TTX (Fig. 2C). Thus, at TTX
concentrations of 3.1 nmol/L, at which the single action potential was
Ó 2000 European Neuroscience Association, European Journal of Neuroscience, 12, 1±7
Do neurons have a reserve of sodium channels?
5
not affected, the time constant of the recovery of action potentials
was prolonged to 13.7 ms, corresponding to an increase of 39%
relative to the control. The recovery from inactivation of the sodium
currents had a longer time constant with a value of 21.7 ms (Fig. 2B,
dotted line, n = 4), further suggesting that only part of the sodium
channels is needed for the second action potential of a pair.
Generation of repetitive action potentials
The slowing of the recovery after activation of action potentials in the
presence of TTX can be assumed to in¯uence repetitive action
potential ®ring. Thus, the effect of the blockade of different fractions
of sodium channels on the maximal number of action potentials
generated by the neuron was studied. The maximal discharge
frequency was obtained by applying long current stimuli with
increasing amplitudes (Fig. 3A; trace with the maximal number of
action potentials marked by an asterisk). With increasing concentrations of TTX, the maximal number of action potentials decreased
(Fig. 3B). The mean maximal discharge frequencies yielded a
decrease from 20.6 Hz under control conditions to 3.0 Hz with
10 nmol/L TTX (Fig. 3C). A ®t of the mean values revealed a
reduction by 56% of control at the TTX concentration of 3.1 nmol/L,
at which concentration the single action potential was not affected.
Furthermore, at a TTX concentration of 0.3 nmol/L, at which the ®rst
noticeable reduction of the sodium current appeared (cf. Fig. 1C), also
a decrease of the maximal discharge frequency was found (Fig. 3C;
reduction to 19.6 Hz, corresponding to a decrease by 5% of the
control).
Discussion
The results of this paper show that in these hippocampal neurons a
reduction of the sodium currents of 34% does not affect the size and
shape of single action potentials (amplitude, duration and maximum
rate of rise). With this reduction, however, the repetitive ®ring of
action potentials was strongly impaired. The time constant of
recovery after activation was prolonged by 39% and the maximal
discharge frequency was reduced by 56%. Furthermore, even smaller
reductions of the sodium current which did not affect single action
potentials decreased the maximal discharge frequency. Therefore, it
can be concluded that these neurons have a surplus of sodium
channels for the generation of single action potentials, but that this
surplus is needed for high-frequency action potential ®ring.
Pharmacological investigations also need to take the results of this
study into account, as the recording of unchanged single action
potentials does not necessarily exclude an impairment of the sodium
channels. If a pharmacological agent is suspected to have blocking
effects on sodium channels, this possibility cannot be ruled out
simply by measuring the size and shape of single action potentials
because, as shown in the neurons studied here, single action
potentials might remain unaffected whereas a great part of the
sodium channels is blocked. Because this `stability' of the action
potential to the blocking of sodium channels has also been described
in other neuronal preparations (myelinated nerve ®bres of frogs;
Schwarz et al., 1973), similar effects are likely to be present in other
nerve cells as well.
In order to estimate the sodium channel density in the investigated
neurons, the number of active sodium channels was calculated from
the amplitude of the sodium current. Under the assumption of an open
probability of ~ 90% at ±20 mV (calculations from the current±
voltage relationships measured by SteinhaÈuser et al., 1990 and Costa,
1996; see also single channel recordings from Magee & Johnston,
1995) and a single channel current at ±20 mV of 1.1 pA (Magee &
FIG. 3. Effects of TTX on the maximal discharge frequency of action
potentials. (A) Original recordings of repetitive action potentials during long
current pulses of increasing amplitude (amplitudes from 0.02 to 0.2 nA,
duration 600 ms). The trace with the maximal number of action potentials is
marked by an asterisk. (B) Original recordings at the maximal discharge
frequency under control conditions (CTRL), and with 0.1, 1 and 10 nmol/L
TTX. The action potentials are truncated. (C) Graphical evaluation of the
maximal discharge frequency at TTX concentrations up to 10 nmol/L. Each
data point represents the mean 6 SEM of six experiments. The mean values
were ®tted with a Langmuir equation. The dotted lines indicate the maximal
discharge frequency at the highest TTX concentration which did not affect
single action potentials, and at the concentration at which the ®rst signi®cant
reduction of the sodium current appeared. At these concentrations, the
maximal discharge frequency was decreased by 56% and 5% relative to the
control, respectively.
Johnston, 1995), the mean sodium current under control conditions of
10.4 nA can be roughly assumed to be carried by ~ 11 000 sodium
channels. With a mean surface area of 1420 mm2 of the neurons
Ó 2000 European Neuroscience Association, European Journal of Neuroscience, 12, 1±7
6 M. Madeja
FIG. 4. Relation between repetitive action potential generation and the density
of sodium channels in isolated hippocampal neurons. The data points (®lled
circles) and the points of the curve are depicted from the experiments on the
amplitude of sodium currents (shown in Fig. 1B) and on the maximal
discharge frequency with different concentrations of TTX (shown in Fig. 3C).
As mentioned in these ®gures, the data are normalized to the values under
control conditions.
tested, this corresponds to a mean density of ~ eight active sodium
channels per mm2 membrane area under control conditions. Thus, the
observed reductions of the sodium currents suggest that in the
isolated CA1 neurons the shape and size of the action potential is not
affected until the sodium channel density is decreased from eight to
~ ®ve active channels per mm2. For the estimation of the total number
of sodium channel density, it has to be considered that at the holding
potential of ±80 mV ~ 70% of the total number of sodium channels
are inactivated (Sah et al., 1988). Thus, the total sodium channel
density can be assumed to be ~ 25 channels per mm2.
The relation between the density of sodium channels and the
maximal discharge frequency (at room temperature) was examined.
Thus, the amplitudes of the sodium currents (see Fig. 1B) and the
maximal discharge frequencies of the action potential (see Fig. 3C)
obtained under control conditions and with different TTX concentrations were correlated (Fig. 4; the curve shows the respective ®ts from
both ®gures). The data points and the curve in Fig. 4 suggest that a
plateau maximal discharge frequency for action potentials is not
reached and that an increase of the sodium current (by increasing the
density of sodium channels to > 25 channels per mm2) would yield
higher maximal discharge frequencies. In accordance with these data,
the sodium channel density is higher in neuronal structures
specialized for the high-frequency conduction of action potentials
(> 50 channels per mm2 in nerve ®bres; see Hille, 1992).
Concerning the mechanism by which a partial block of sodium
channels modi®es repetitive action potential ®ring without affecting
the generation of the single action potential, the sodium channels'
recovery from inactivation appears to be mainly involved. From the
estimation shown above it can be assumed that a full-blown action
potential can be elicited when during recovery from inactivation
®ve channels per mm2 are available for activation again. Due to the
exponential time course of recovery from inactivation, this time
interval can be assumed to be short if the density of sodium channels
per mm2 is high. For a total of eight active channels per mm2, the time
interval was measured as ~ 23 ms (recovery of the sodium current to
66% of control; see Fig. 2B) suggesting a theoretical maximum
discharge frequency of 43 Hz. [For the measured, lower value of
21 Hz (see Fig. 3C) it has to be considered that the applied long
current pulses lead to a continuous depolarization inactivating part of
the sodium channels, and that in a series of action potentials the
generation of not full-blown action potentials leads to a reduction of
the repolarizing potassium currents. Both mechanisms impair the
ability of the neuron for repetitive ®ring and thus reduce the obtained
values of the measured maximum discharge frequencies.] An increase
of channel density to a total of e.g. 10 active channels per mm2 can be
assumed to reduce the time interval for the recovery of ®ve active
channels per mm2 to ~ 15 ms and to increase the theoretical maximum
discharge frequency to 67 Hz. However, if the channel density is
reduced to ®ve active channels per mm2, a full-blown, unchanged
action potential can be generated only after complete recovery from
inactivation, i.e. the maximum discharge frequency is low.
It was the aim of the present paper to study the effects of reduction
of sodium channel density in a neuronal cell membrane. Thus, the
isolated CA1 neuron was chosen as a simple and easily accessible
neuronal preparation. It is, however, hardly possible to infer from the
obtained results to the CA1 neuron in vivo due to an important
limitation: the obtained density of sodium channels in the isolated
neuron might not correspond to the somatic channel density of the
neuron in vivo. The isolated CA1 neurons do not possess axons and it
is unknown if the axon is lost completely or if part of the axon is
retracted into the cell body following isolation. In the latter case, part
of the recorded sodium channels was of axonal origin and did not
belong to the soma of the CA1 neuron. Because the density of sodium
channels is high in axons in comparison with cell bodies (Safronov
et al., 1999), the somatic sodium channel density in vivo might be
lower than estimated from the isolated neuron. On the contrary,
dendritic membrane areas are incorporated into the soma during
rounding up of the isolated CA1 neuron. Although sodium channels
are found in dendrites of CA1 neurons (Magee & Johnston, 1995), a
low dendritic density is likely. Thus, the mean sodium channel density
of the isolated and rounded neuron might be re-decreased. However,
although no information about distribution of channels in isolated CA1
neurons is available, TTX-binding studies on these neurons in brain
slices suggest a signi®cant amount of sodium channels within the cell
bodies (Mourre et al., 1988) which may be assumed to roughly
represent the situation in the acutely isolated neurons. However, for
more de®nite conclusions, binding studies quantifying the sodium
channel density in isolated CA1 neurons are needed.
This study was performed on a neuronal cell type. The loss of the
complexity of the neuron after isolation and the fact that in general
the same effect of a stronger effect of TTX on sodium currents than
on action potentials has been described for myelinated nerve ®bres of
frogs (Schwarz et al., 1973) may allow the discussion of the relevance
of the results for action potential generation in nerve ®bres as well.
Thus, in the node of Ranvier, very high sodium channel densities
have been found (up to 2000 channels per mm2; Conti et al., 1976;
Chiu, 1980; Waxman & Ritchie, 1985). It is tempting to speculate
that this clustering of sodium channels represents another advantage
of conduction in myelinated nerve ®bres. In comparison with
unmyelinated ®bres, the extreme sodium channel density at the
nodes would not only induce a faster generation and rise of the action
potential, thus further increasing the conduction velocity (Hille,
1992), but might also allow the generation of action potentials at
higher frequencies, and thus extend the operating range of the ®bre
for frequency-encoded information.
Ó 2000 European Neuroscience Association, European Journal of Neuroscience, 12, 1±7
Do neurons have a reserve of sodium channels?
Furthermore, especially under pathological conditions, changes of
action potential shape and sodium current amplitude (Figenschou
et al., 1996; Francke et al., 1996; Colino et al., 1998) might affect the
generation of action potentials at high frequencies. Thus, it can be
assumed that processes of demyelination which are known to reduce
the conduction velocity of nerve ®bres might cut off part of the
conducted information. In contrast to the nodes of Ranvier, the
internodal area contains only a low density of sodium channels
(< 25 sodium channels per mm2; Ritchie & Rogart, 1977; Black et al.,
1990; see Catterall, 1992). With demyelination, the saltatory
conduction of the ®bre is replaced by continuous conduction (see
Kocsis & Waxman, 1985) and action potentials have to be generated
in the former internodal areas, where the sodium channel density is
low. This might be supposed to reduce maximal discharge frequency
of action potentials in the impaired ®bre and thus to restrict the
operating range of the ®bre and block the information coded by
higher action potential frequencies (see Kocsis & Waxman, 1985).
Besides the reduced conduction velocity and the other pathogenic
mechanisms of demyelination, the loss of conducted information due
to action potential generation in areas of reduced sodium channel
density might contribute to the symptoms of demyelinating diseases.
Acknowledgements
I would like to thank Professor E.-J. Speckmann (MuÈnster, Germany) for
discussions of the results and Professor W. Ulbricht (Kiel, Germany) for
critical comments on the manuscript. I am grateful to Dr B.J. Corrette (Berlin,
Germany) for English corrections.
This article is dedicated to Professor V. Cioli (Florence, Italy).
Abbreviations
EGTA, ethylene glycol-bis(b-aminoethyl ether)-N,N,N¢,N¢-tetraacetic acid;
HEPES, N-2-[hydroxyethyl]piperazine-N¢-[2-ethanesulphonic acid]; TTX,
tetrodotoxin.
References
Bezanilla, F. & Armstrong, C.M. (1977) Inactivation of the sodium channel. I.
Sodium current experiments. J. Gen. Physiol., 70, 549±566.
Black, J.A., Kocsis, J.D. & Waxman, S.G. (1990) Ion channel organization of
the myelinated ®ber. Trends Neurosci., 13, 48±54.
Catterall, W.A. (1980) Neurotoxins that act on voltage-sensitive sodium
channels in excitable membranes. Annu. Rev. Pharmacol. Toxicol., 20, 15±
43.
Catterall, W.A. (1992) Cellular and molecular biology of voltage-gated
sodium channels. Physiol. Rev., 72, S15±S48.
Chiu, S.Y. (1980) Asymmetry currents in the mammalian myelinated nerve.
J. Physiol. (Lond.), 273, 499±519.
Colino, A., Garcia-Seoane, J.J. & Valentin, A. (1998) Action potential
broadening induced by lithium may cause a presynaptic enhancement of
excitatory synaptic transmission in neonatal rat hippocampus. Eur. J.
Neurosci., 10, 2433±2443.
Conti, F., Hille, B., Neumcke, B., Nonner, W. & StaÈmp¯i, R. (1976)
Measurement of the conductance of the sodium channel from current
¯uctuations at the node of Ranvier. J. Physiol. (Lond.), 262, 699±727.
Costa, P.F. (1996) The kinetic parameters of sodium currents in maturing
acutely isolated rat hippocampal CA1 neurones. Dev. Brain Res., 91, 29±40.
7
Figenschou, A., Guo-Yuan, H. & Storm, J.F. (1996) Cholinergic modulation of
the action potential in rat hippocampal neurons. Eur. J. Neurosci., 8, 211±
219.
Francke, M., Pannicke, T., Biedermann, B., Faude, F. & Reichelt, W. (1996)
Sodium current amplitude increases dramatically in human retinal glial cells
during diseases of the eye. Eur. J. Neurosci., 8, 2662±2670.
Hamill, O.P., Marty, A., Neher, E., Sakmann, B. & Sigworth, F.J. (1981)
Improved patch-clamp techniques for high-resolution current recording
from cells and cell-free membrane patches. P¯uÈgers Arch., 391, 85±100.
Hille, B. (1992) Counting channels. In Editors? (eds), Ionic Channels of
Excitable Membranes, 2nd ed., Chapter 12. Sinauer Associates, Sunderland,
MA, USA, pp. 315±336.
Kandel, E.R. (1976) Cellular mechanisms of neuronal function. In Kandel,
E.R. (ed.), Cellular Basis of Behavior, Chapter 6. Freeman, San Francisco,
pp. 137±210.
Kay, A.R. & Wong, R.K.S. (1986) Isolation of neurons suitable for patchclamping from adult mammalian central nervous systems. J. Neurosci.
Meth., 16, 227±238.
Kocsis, J.D. & Waxman, S.G. (1985) Demyelination: causes and mechanisms
of clinical abnormality and functional recovery. In Koetsier, J.C. (ed.),
Handbook of Clinical Neurology, Vol. 3, Demyelinating Diseases. Elsevier
Science, Amsterdam, pp. 29±47.
Madeja, M., Mubhoff, U., Binding, N., Witting, U. & Speckmann, E.-J. (1997)
Effects of Pb2+ on delayed-recti®er potassium channels in acutely isolated
hippocampal neurons. J. Neurophysiol., 78, 2649±2654.
Magee, J.C. & Johnston, D. (1995) Characterization of single voltage-gated
Na+ and Ca2+ channels in apical dendrites of rat CA1 pyramidal neurons.
J. Physiol. (Lond.), 487, 67±90.
Magistretti, J., Mantegazza, M., Guatteo, E. & Wanke, E. (1996) Action
potentials recorded with patch-clamp ampli®ers: are they genuine? Trends
Neurosci., 19, 530±534.
Mourre, C., Moll, C., Lombet, A. & Lazdunski, M. (1988) Distribution of
voltage-dependent Na+ channels identi®ed by high-af®nity receptors for
tetrodotoxin and saxitoxin in rat and human brains: quantitative
autoradiographic analysis. Brain Res., 448, 128±139.
Narahashi, T. (1974) Chemicals as tools in the study of excitable membranes.
Physiol. Rev., 54, 813±889.
Ritchie, J.M. & Rogart, R.B. (1977) The density of sodium channels in
mammalian myelinated nerve ®bers and the nature of the axonal membrane
under the myelin sheath. Proc. Natl Acad. Sci. USA, 74, 211±215.
Sah, P., Gibb, A.J. & Gage, P.W. (1988) The sodium current underlying action
potentials in guinea pig hippocampal CA1 neurons. J. Gen. Physiol., 91,
373±398.
Safronov, B.V., Wolff, M. & Vogel, W. (1999) Axonal expression of sodium
channels in rat spinal neurones during postnatal development. J. Physiol.
(Lond.), 514, 729±734.
Scholz, A., Appel, N. & Vogel, W. (1998) Two types of TTX-resistant and one
TTX-sensitive Na+ channel in rat dorsal root ganglion neurons and their
blockade by halothane. Eur. J. Neurosci., 10, 2547±2556.
Schwarz, J.R., Ulbricht, W. & Wagner, H.-H. (1973) The rate of action of
tetrodotoxin on myelinated nerve ®bres of Xenopus laevis and Rana
esculenta. J. Physiol. (Lond.), 233, 167±194.
SteinhaÈuser, C., Tennigkeit, M., Matthies, H. & GuÈndel, J. (1990) Properties of
the fast sodium channels in pyramidal neurones isolated from the CA1 and
CA3 areas of the hippocampus of postnatal rats. P¯uÈgers Arch., 415, 756±
761.
Ulbricht, W. (1981) Kinetics of drug action and equilibrium. Results at the
node of Ranvier. Physiol. Rev., 61, 785±828.
Vreugdenhil, M. & Wadman, W.J. (1995) Potassium currents in isolated CA1
neurons of the rat after kindling epileptogenesis. Neuroscience, 66, 805±
813.
Waxman, S.G. & Ritchie, J.M. (1985) Organization of ion channels in
myelinated nerve ®ber. Science, 228, 1502±1507.
Ó 2000 European Neuroscience Association, European Journal of Neuroscience, 12, 1±7