The Journal of Neuroscience, October 1989, 9(10): 35773583

GABA-Activated Cl- Channels in Astrocytes of Hippocampal Slices

B. A. MacVicar, F. W. Y. Tse, S. A. Crichton, and H. Kettenmann2
Department of Medical Physiology, University of Calgary, Calgary, Alberta, Canada T2N 4N1, and 2Department of
Neurobiology, University of Heidelberg, D-6900 Heidelberg 1, Federal Republic of Germany
We used kainic acid-lesioned hippocampal slices to ex-
amine glial responses to the inhibitory neurotransmitter GABA
in a neuron-free environment. Slices were prepared from rats
which received intracerebroventricular injections of kainic
acid 1 month prior to experiments. Astrocytes (membrane
potential averaged 81.4 f 5.5 mV; n = 48; mean -t SD) were
impaled in the CA3 region of the slice, which was completely
depleted of neurons. GABA (1 mM) application by bath per-
fusion depolarized membrane potential from 1 to 5 mV. The
GABA-induced depolarization was not affected by a tetro-
dotoxin (1 b.cM)/high-Mg*+/low-Ca2+ solution. Changing the Cl-
equilibrium potential by reducing extracellular Cl greatly in-
creased the GABA-induced depolarization. Muscimol mim-
icked the GABA response, while picrotoxin (0.1 mM), an an-
tagonist of the GABA-activated Cl- channel, resulted in a
60% blockade. The barbiturate, pentobarbital(O.1 mh!), and
the benzodiazepine agonist, flunitrazepam (1 mM), en-
hanced the depolarization by 60 and 40%, respectively. A
blocker of glial GABA uptake, j3-alanine (1 mM), did not affect
the GABA-induced membrane depolarization, indicating that
the depolarization is not caused by electrogenic uptake of
the amino acid. The pharmacological properties of the GABA
response of astrocytes from the hippocampal slice is similar
to that previously described for cultured astrocytes from rat
cerebral hemispheres. Our data suggest that GABA recep-
tors, which are coupled to Cl- channels, are also expressed
by astrocytes in an intact tissue.
When the major inhibitory neurotransmitter in the CNS, GABA,
is released, it binds to postsynaptic GABA, receptors and opens
Cl- selective channels (Krnjevic, 1974; Bormann, 1988). The
increase in the Cl- permeability ofthe neuronal membrane shunts
any current generated by excitation and is responsible for the
inhibition ofthe postsynaptic neuron. Depending on the relation
between resting membrane potential and Cl- equilibrium po-
tential, the neuronal membrane is either depolarized or hyper-
polarized. In pyramidal cells of the hippocampus, GABA hy-
perpolarizes the cell when applied at the soma and depolarizes
the cell when applied at the dendrites (Andersen et al., 1980).
Received Oct. 17, 1988; revised Mar. 1, 1989; accepted Apr. 18, 1989.
This work was supported by grants from Deutsche Forschungsgemeinschaft and
Bundesministerium fur Forschung und Technologie to H.K. and the Medical
Research Council (Canada) to B.A.M. F.W.Y.T. is an Alberta Heritage Foundation
for Medical Research (AHFMR) postdoctoral fellow, and B.A.M. is an AHFMR
Scholar, MRC Scientist, and Sloan Fellow.
Correspondence should be addressed to Dr. B. A. MacVicar, Neuroscience
Research Group, University of Calgary, 3330 Hospital Drive N.W., Calgary,
Alberta, Canada T2N 4Nl.
Copyright 0 1989 Society for Neuroscience 0270-6474/89/103577-07$02.00/O
Glial cells have been demonstrated to influence the size and
time course of GABA-induced conductance changes in neurons
by rapid removal of extracellular GABA from the synaptic re-
gion (e.g., Gallagher et al., 1983). Recent evidence from studies
on cultured astrocytes and oligodendrocytes indicates that glial
cells not only possess high-affinity uptake sites for GABA
(Schousboe, 198 1; Reynolds and Herschkovitz, 1984), but also
a GABA receptor-coupled Cl- channel (Kettenmann et al., 1984,
1987; Bormann and Kettenmann, 1989). Activation of GABA
receptor-coupled Cll channels always induced a depolarization
in glial cells because the intracellular Cl- in these cells is elevated
above passive distribution (Kettenmann et al., 1987; Hoppe
and Kettenmann, 1989). The pharmacological properties of the
glial GABA receptor were characterized in cultured astrocytes
from rat cortex, and it shared many similarities with the neu-
ronal GABA, receptor. Glial GABA responses were blocked by
the GABA, receptor antagonists picrotoxin and bicuculline, and
were mimicked by the agonist muscimol (Kettenmann and
Schachner, 1985). The glial GABA response was enhanced, as
in neurons, by pentobarbital and benzodiazepine agonists (Backus
et al., 1988); however, unlike the effect on neurons, inverse
benzodiazepine agonists increased GABA responses in astro-
cytes (Backus et al., 1988; Bormann and Kettenmann, 1989).
GABA-induced depolarizations in glial cells in situ have also
been observed (Krnjevic and Schwartz, 1967), but these de-
polarizations were speculatively attributed to the release of K+
from surrounding neurons (Constanti and Galvan, 1978).
The aim of the present study was to identify glial GABA
receptors in an intact tissue of the nervous system to verify that
these receptors are not only expressed in cultured astrocytes,
but that the expression of GABA receptor is a property of glial
cells in the CNS.
Materials and Methods
Hippocampal slices. Kainic acid (KA)-induced neuronal lesions were
used to produce neuron-free areas in hippocampal slices. KA (2-3 rg)
was injected bilaterally into the lateral ventricles of male Sprague-Daw-
ley rats (150-200 gm). After injections, animals experienced repetitive
seizures which gradually diminished in severity after a few hours. From
3 to 6 weeks after the injection, transverse hippocampal slices (500 pm)
were prepared from the KA-treated rats. As has been shown before
(Nadler et al., 1978), KA induced virtually complete neuronal degen-
eration in area CA3. In cresyl violet-stained tissue (Fig. l), the loss of
pyramidal cells and the proliferation of glial cells were evident.
Cell culture. Cultures of enriched astrocytes (> 90%) were obtained
from cerebral hemispheres of 0- to 2-d-old Sprague-Dawley rats ac-
cording to McCarthy and De Vellis (1980), with modifications described
by Keilhauer et al. (1985), and maintained in culture for 3-6 weeks. By
a combination of immunocytochemical and electrophysiological tech-
niques, it was previously shown that cells depolarized by GABA are
glial fibrillary acidic protein-positive astrocytes (Kettenmann et al., 1984).
3578 MacVicar et al. * GABAergic Responses in Astrocytes
Figure 1. GFAP labeled a LY-inject-
ed glial cell in a KA-lesioned hippo-
campal slice. a, Cresyl violet-stained
cryostat section of a hippocampal slice
from which the glial cell illustrated in
b, c was recorded. There are no neurons
visible in the CA3 region, which ex-
hibits intense gliosis. The granule cell
layer (CC) survived the KA lesion. A
recording was obtained from a glial cell
(in the area indicated by the arrow)
which depolarized in response to GABA
and was then injected with LY. The slice
was cryostat sectioned (1 O-l 5 pm), and
the section containing the LY-stained
cell body (b) was also immunohisto-
chemically stained for GFAP (c). The
adjacent section was stained with cresyl
violet (a). GFAP-positive fibers are
contained within the LY-stained cell
(arrows), indicating that this cell was an
astrocyte. The intensity of LY staining
was greatly diminished by the diami-
nobenzidine reaction to visualize HRP.
Scale bars: a, 200 pm; b and c, 20 pm.
Electrophysiologicul procedures. Hippocampal slices (500 hrn) or cell
cultures were maintained in a recording chamber at 35C (25-30C for
cell cultures) and were continuously perfused. Slices were submerged
and perfused on one side. Cells were penetrated by 2 M K-acetate or
KCl-filled electrodes (resistance 40-80 Ma). A similar electrode re-
corded extracellular potentials which were subtracted from the intra-
cellular recordings to determine the true transmembrane potential. For
intracellular injection of Lucifer yellow (LY, Stewart, 198 1), electrodes
were filled with 46% LY (Sigma, USA); 2-4 nA min of hyperpolarizing
current was sufficient to stain glial cells. Electrophysiological data were
amplified by conventional electrophysiological equipment, stored on
the floppy disk system of a digital oscilloscope (Nicolet, Madison, USA),
and plotted on an XY-plotter.
Solutions. Slices or cultures were continuously perfused via pump or
gravity-feed systems during recording. For slices, the control solution
contained (in mM): NaCl, 124; KCl, 5; MgCl,, 1.3; NaHCO,, 26; CaCl,,
2; glucose, 10 (pH 7.35). The standard salt solution for cell culture
contained (in mM): KCl, 5.4; NaCl, 116.4; NaH,PO,, 1.0; MgSO,, 0.8;
CaCl,, 1.8; D-glucose, 5.6; NaHCO,, 26.2. The TTX/high-Mg2+/low-
Ca*+ solution was the control solution modified to include lp~ TTX,
10 mM Mg2+, and 0.2 mM CaZ+. NaCl (124 mM) was replaced with Na
gluconate (124 mM) in experiments to reduce extracellular Cl-. The
following substances were added to the salt solution: GABA (1, 10 mM),
muscimol(1 mh@, picrotoxin (0.1 mM), flunitrazepam (0. l-l mM), pen-
tobarbital (0.1 mM), fl-alanine (1 mM).
GFAP staining. After recordings were obtained from cells injected
with LY, slices were fixed in 4% paraformaldehyde in phosphate buffer
(0.1 M) for 24 hr. Following infiltration with 20% sucrose, sections (lO-
15 pm) were cut on a cryostat and observed on a fluorescence microscope
to determine which section(s) had the LY-stained cell. Immunohisto-
chemical staining for glial fibrillary acidic protein (GFAP) was accom-
plished using a Vectastain kit (Vector Lab.) and a monoclonal antibody
for GFAP (Labsystems, Helsinki). The final visualization of the HRP
was accomplished using a diaminobenzidine reaction. This, however,
The Journal of Neuroscience, October 1989, 9(10) 3579
GABA(1 mM)
A High K+
A
B
Muscimol
(1 mM)
30 set
I
2 mV
-so m &#Ll_____ -----___------_-- ---- -- ---- -___-----
1 min
Figure 2. GABA- and muscimol-induced membrane depolarization.
A, A stable intracellular recording was obtained from a glial cell in a
KA-lesioned slice. GABA (1 mM) was applied as indicated by bar and
reversibly depolarized glial membrane potential. The resting membrane
potential is indicated at the beginning of the trace. B, In another cell,
muscimol (1 mM), an agonist, was applied as indicated and also re-
versibly depolarized the glial cell.
greatly diminished the LY fluorescence. Adjacent sections were stained
with cresyl violet to examine the extent of the KA-induced lesion.
Results
Identification of astrocytes in the hippocampal slice
Hippocampal slices were obtained from rats which had received
an injection of KA 3-6 weeks prior to the experiment. Cells
were impaled in the CA3 area, which was free of neurons as
described previously (MacVicar et al., 1987). Neurons were rare-
ly impaled (e.g., only in 3 out of several hundred impalements;
in these cases, the slice was then no longer used because it
contained viable neurons). Stable penetration of glial cells re-
sulted in the recording of a very negative resting membrane
potential (8 1.4 * 5.5 mV; mean f SD, 12 = 46) with no apparent
spike or postsynaptic potential activity.
GABA-induced depolarizations in the slice
In all cells examined, application of GABA (1 or 10 mM) for 1
min depolarized the membrane potential reversibly (Fig. 2A).
The maximum depolarization induced by 1 or 10 mM GABA
was 1.6 + 0.7 mV (? SD, n = lo), or 3.5 & 0.9 (+ SD; n =
9), respectively. In 5 experiments, we verified that this GABA-
induced depolarization also occurred when synaptic and spike
activities in the slices were minimized by superfusing a
low-Ca*+TTX/high-Mg2+ solution (see Materials and Meth-
ods; data not shown); under these conditions the mean depo-
larization induced by 1mM GABA was 1.5 f 0.7 (k SD, n =
5).
In three out ofthree experiments we observed that the GABA-
induced depolarization could be reversed (Fig. 3) when the
membrane potential was depolarized to approximately - 15 mV
during a sustained application of 50 mM extracellular KC1 (Fig.
3; in TTX solution/high Mg2+llow Ca*). Furthermore, altering
the equilibrium potential for Cl- (E,,) by replacing extracellular
Cl- with gluconate, an impermeant anion, increased the GABA
GABA
-15 mV
------
I
3 mV
B Wash
GABA
25 set
-70 mV
__----
%
Figure 3. Reversal of GABA depolarizations in high K+. A stable
intracellular recording was obtained from a glial cell in 5 mM K+, then
K+ was increased to 50 mM by equimolar replacement of Na+ with K+.
When membrane potential was approximately - 15 mV (after 10 min),
GABA (1 mM) was bath-applied and caused a membrane hyperpolar-
ization. Membrane potential was still slowly drifting and is indicated
by the dotted line. Extracellular K+ was then decreased to 5 mM, and
GABA now caused a depolarization as membrane potential was slowly
recovering to control levels.
(1 mM)-induced depolarization from 1.6 to 7 mV (n = 4; Fig.
4A). The above experiments suggest that GABA is activating a
Cll conductance and the normal E, in glial cells is less negative
than membrane potential. Six cells from these experiments were
stained by injecting -2 nA of LY for 2.0 min. As described
earlier (Gutnick et al., 1981; Connors et al., 1984), the dye
spread to adjacent cells indicating electrical coupling. Some cells
were subsequently identified as astrocytes by immunohisto-
chemically staining the slice for GFAP, a specific glial marker
(Fig. 1).
Pharmacology of GABA-induced depolarizations
We tested the effect of an agonist, an antagonist, and 2 modu-
lators of neuronal GABA, receptors to compare the GABA
responses in hippocampal astrocytes with those in neurons and
cultured glial cells. Muscimol, (1 mM), a GABA, receptor ag-
onist that is not taken up in brain slices (Kom and Dingledine,
1986), mimicked the action of GABA by inducing a maximum
depolarization of 2.7 * 1.8 mV (Fig. 2B; n = 5). Picrotoxin (0.1
mM), which has been shown to block the GABA-coupled Cl-
channel in neurons (Barker et al., 1983), reversibly reduced the
GABA-induced depolarization in the astrocytes by 40% (n = 4;
Fig. 4B) compared with a control obtained from the same cell.
Pentobarbital, a barbiturate, reversibly enhanced the GABA-
induced depolarization by 60% (n = 9; Fig. 5); a similar action
on the neuronal GABA, receptor has been reported (Study and
Barker, 198 1). Flunitrazepam (1 mM), a benzodiazepine recep-
tor agonist, enhanced the GABA response by 40% (n = 4; Fig.
5), showing further similarities to the neuronal GABA, receptor
(Bormann and Sakmann, 1984). A blocker of GABA uptake,
P-alanine (1 mM) (Shon and Kelly, 1974; Bowery et al., 1979;
Schousboe, 198 l), did not affect the size of the GABA-induced
depolarization (data not shown; n = 3). The pharmacological
properties of the GABA receptor in astrocytes from hippocam-
pus are thus similar to those of the GABA, receptor in neurons
and the GABA receptor in cultured astrocytes.
3580 MacVicar et al. * GABAergic Responses in Astrocytes
Figure 4. GABA-induced membrane
depolarizations are dependent on the
Cl- gradient and blocked by picrotoxin.
During an intracellular recording from
a glial cell in a KA-lesioned slice, GABA
(1 mM) was applied as indicated by bar.
Three responses to GABA were ob-
tained from the same cell and were su-
DerimDOSed. The individual aodica-
iions were separated by at least i 0 min
to avoid progressive desensitization.
First, as a control, GABA was applied.
For the next application, the experi-
mental solution (gluconatcAow Cl- or
picrotoxin) was applied for 2-10 min
and the effect of GABA was again tested
in the presence of the experimental so-
lution. Resting membrane potential was
not affected by either Cl- replacement
or picrotoxin. After a 20 min wash with
normal bathing solution, GABA was
again applied. A, Reducing extracellular
Cl- by replacing most Cl- with an im-
permeant anion, gluconate, greatly en-
hanced the depolarization from GABA
(1 mM) application. Note also that the
depolarization in gluconate quickly be-
gins to desensitize. This is also ob-
served during GABA applications on
cultured gliac cells. B, Picrotoxin (0.1
ins) reversibly decreased the GABA-
induced depolarization. Therefore, the
glial GABA response is altered by both
changing the [Cl-] and by pharmaco-
logically blocking the Cl- channel with
picrotoxin.
GABA
A
lb
gluconate/low chloride
3 mV
-83 mV
1 min
GABA
B -
-78 mv
I
2 mV
--------____----__-----------
1 min
Comparison of glial GABA responses in situ and in vitro
The GABA-induced depolarizations in astrocytes from the rat
hippocampal slice were much smaller in size than those de-
scribed for cultured astrocytes from rat cortex (Kettenmann et
al., 1984). This difference can be attributed to 2 aspects: first,
the diffusion of applied GABA to astrocytes is slower in the
slice than in culture because of the thickness of the slice; second,
there may be a difference in E,,. The contribution from the
former was directly tested. To obtain an estimate of the velocity
of GABA diffusion in the slice, we increased the extracellular
K+ concentration from 5 to 50 mM and recorded the resulting
membrane depolarization of an astrocyte (not shown). The
membrane potential depolarized by 60 mV and reached a pla-
teau within 4.5 min; 90% of the depolarization was reached
within 3.5 min. We then compared GABA- and K+-induced
depolarization in cultured astrocytes and varied the velocity of
application by varying the speed ofbath perfusion. The exchange
rate of the bath was monitored by recording the time course of
the depolarization induced by an increase in the extracellular
K+ concentration. An exchange rate ofat least 90% ofthe bathing
solutions within 15 set was required to elicit GABA(lm@-
induced depolarizations, which were larger than 10 mV. A re-
duction of the perfusion rate from 90% exchange/l0 set to 90%
exchange/60 set in culture to mimic the slow diffusion of GABA
into the slice reduced the GABA-induced depolarization from
>20 mV to <5 mV (Fig. 6). We conclude that slow application
of GABA results only in small depolarizations even at optimal
GABA concentrations.
Discussion
Presence of glial GABA receptors in situ
Previous studies have already shown that glial cells in the intact
tissue were depolarized by GABA (Kmjevic and Schwartz, 1967;
Constanti and Galvan, 1978). It was, however, suggested that
the depolarization is caused by either electrogenic uptake of
GABA (e.g., Krnjevic, 1984) or by the increase of extracellular
K+ concentration released from adjacent neurons (Constanti and
Galvan, 1978). In these experiments, however, the CA3 neurons
have been lesioned by a KA injection. The GABA-induced de-
polarization was not blocked by inhibiting any possible residual
neuronal activity or synaptic transmission with a combination
of TTX, raised extracellular MgZ+, and reduced Ca*+. For cul-
tured astrocytes, it has been unequivocally demonstrated that
the GABA-induced depolarization is caused by the opening of
Cl- channels activated by GABA (Kettenmann et al., 1987;
Bormann and Kettenmann, 1989). GABA responses in glial cells
in the slice could be enhanced by the barbiturate pentobarbital
and blocked by the GABA, receptor antagonist picrotoxin. In
addition, the blocker of glial GABA uptake, p-alanine (Shon
and Kelly, 1974; Bowery et al., 1979), did not affect the GABA-
induced depolarization, indicating that electrogenic uptake did
not cause the depolarization. Also, muscimol, a GABA, agonist
which is not taken up in brain slices (Johnston et al., 1978;
White and Snodgrass, 1979; Korn and Dingledine, 1986) de-
polarized glial cells in this study. Therefore, the pharmacological
properties of glial GABA responses in the KA-lesioned slice are
in line with those described for cultured astrocytes and are thus
The Journal of Neuroscience, October 1989, 9(10) 3581
A
GABA
1 min
B GABA
coJtr(l mV
1 min
Figure 5. Effect of pentobarbital and flunitrazepam on the GABA-
induced depolarization. GABA alone was applied for the control re-
sponse; the experimental solutions (pentobarbital or flunitrazepam) were
applied for 2 min and the response to GABA was then tested again. A,
In the presence of 0.1 mM pentobarbital (PB), the GABA-induced de-
polarization was enhanced compared with the control. B, Similar to the
response in A, the addition of flunitrazepam (1 mM) enhanced the am-
plitude of the GABA-induced depolarization. Effects of both drugs were
reversible upon wash.
similar to the properties of the neuronal GABA, receptor. This
suggests that the expression of GABA, receptors in astrocytes
is not induced by the culture conditions, but is an intrinsic
property of these cells. We cannot exclude, however, that the
lack of neurons and the reactive gliosis might affect the expres-
sion of channels in astrocytes.
Sensitivity of the glial GABA receptor
In this study we have used higher concentrations of GABA than
similar studies performed on neurons or on cultured astrocytes
and have detected only small changes in the resting membrane
potential of astrocytes in the slice. While the single-channel
properties of the GABA receptors in neurons and cultured as-
trocytes are similar (Bormann and Kettenmann, 1989), several
properties of glial cells in an intact tissue make it difficult to
detect membrane depolarizations upon GABA receptor acti-
vation. First, glial cells are strongly electrically coupled, thus
forming a large syncytium (Gutnick et al., 1981; Kettenmann
and Ransom, 1988). When GABA is applied, it first activates
receptors located on the surface of the slice. However, because
of the extensive coupling, the depolarization is strongly atten-
uated by the glial cells located deeper in the slice. Subsequently,
when GABA has penetrated further into the slice, receptors on
the surface are inactivated and thus attenuate the depolarization
of the glial cells deeper in the slice. The extensive coupling also
results in small membrane potential changes when substances
are applied iontophoretically, since only a relatively small mem-
brane area of the whole syncytium can be activated as it has
been done in cortical glial cells (Krnjevic and Schwartz, 1967).
Second, the input resistance of glial cells is much lower than
;/mmvl A
-20
-40
-60
r-J-4
-80
GABA
I I
-20
-40
-60
-80
B
1 min
E
I
GABA
D
I
SOmMK*
Figure 6. Influence of the bath perfusion speed on the size of the
GABA-induced depolarization in cultured astrocytes. Membrane po-
tential (VJ was recorded from a cultured astrocyte from rat cortex. In
A, GABA (1 mM) was applied as indicated by bar, and in B, with the
same speed of bath perfusion, [K+] was increased from 5.4 to 50 mM.
In C and D, the velocity of bath perfusion was reduced and GABA (c)
resulted in a smaller depolarization. The slowed perfusion speed is
reflected in the slower depolarization induced by the increase in [K+]
CD).
that of neurons. This results in a smaller membrane depolari-
zation when comparable numbers of receptors are activated.
Third, the normal E,, in glial cells in the slice preparations may
be closer to resting membrane potential. Shifting E,, to a more
depolarized potential by reducing extracellular Cl- greatly en-
hanced the amplitude of the depolarization. These 3 factors may
explain why GABA responses can be easily detected in neurons,
but not in glial cells of a slice preparation. Compared with
cultured astrocytes, responses are smaller because of the slow
diffusion of the transmitter in the slice when applied by bath
exchange. We obtained an estimate of the time course for dif-
fusion in the slice during a solution change by increasing extra-
cellular K+ and recording the membrane depolarization of an
astrocyte. Diffusion into the slice greatly reduces the rate at
which agents can be introduced onto cells. Thus, activating re-
ceptors by slowly increasing the concentration in the slice led
to a small depolarization of the glial cell, in contrast to a step-
like solution change causing a large depolarization in culture
(see Fig. 6). Considering these less-than-optimal conditions for
detection of GABA responses, we conclude that glial cells could
express GABA receptors in significant amounts and that they
could be functionally relevant. Receptor activation leads only
to small depolarization, but could locally induce large Cl- cur-
rents (Kettenmann et al., 1988). The presence of GABA, re-
ceptors and GABA-activated Cl- currents in glial cells may have
widespread implications for GABA-binding studies and GABA-
activated Cl- flux studies. In most biochemical studies in the
CNS, it has been assumed that GABA-activated Cl- channels
3582 MacVicar et al. - GABAergic Responses in Astrocytes
ASTROCYTE
TERMINAL
Figure 7. Hypothetical function ofglial GABA receptors. The diagram
displays the presumptive spatial arrangement ofglial and neuronal GABA
receptors necessary to explain the role of astrocytes in the Cll homeo-
stasis. The close vicinity of an astrocytic process at a synapse to a neuron
is assumed. At lower right is a synaptic region marked by the synaptic
vesicles. This part illustrates that GABA is released from the presynaptic
terminal and activates GABA receptors at the neuronal postsynaptic
membrane. The GABA receptor Cl- channel opens and triggers a Cl-
flux from the extracellular space into the neuron. GABA also activates
astrocytic GABA receptors, which leads to an efflux of Cl- from the
glial cell into the extracellular space. Movements of other ions, partic-
ularly K+, may also occur, and extracellular [K+] could also increase
due to a K+ efflux from GABA-depolarized astrocytes.
are expressed only on neurons. This work indicates that there
should be more caution attributing a purely neuronal site for
some neurotransmitters. There is also evidence for glutamate
(kainate type; MacVicar et al., 1988) and muscarinic (S. A.
Crichton and B. A. MacVicar, unpublished observations) re-
ceptors on glial cells in the KA-lesioned hippocampal slice.
Function of &al GABA receptors
The possibility that GABA receptors are expressed by glial cells
in an intact tissue leads to the question of their functional role.
These receptors could serve as a signal from an inhibitory neu-
ron to an adjacent glial cell. Furthermore, we present, in the
following, a hypothesis by which glial cells could be enabled to
regulate the extracellular Cl- concentration in the vicinity of a
GABAergic synapse (Fig. 7).
We assume that glial GABA receptors are concentrated at
processes or endfeet facing the synaptic cleft of a GABAergic
synapse. When neuronal GABA receptors are activated, it leads
to an influx of Cl- into those neurons which hyperpolarize upon
application of GABA. The Cl- influx leads, as a consequence,
to a decrease in extracellular [Cl-] as described by Miiller et al.
(1988). In contrast, activation of glial GABA receptors leads to
an efflux of Cl-. Glial cells could thereby be enabled to regulate
[Cl-]. This mechanism ofextracellular Cl- regulation could serve
to stabilize [Cl-] in the synaptic cleft at times of high inhibitory
activity and supports the concept of glial cells being regulatory
units of extracellular ion concentrations. Since the efflux of Cl-
is also accompanied by an efflux of K+ (Kettenmann et al., 1988),
a rise in extracellular [K+] should be observed during GABA
application. An efflux of K+ has been observed during a KA-
induced depolarization of glial cells (MacVicar et al., 1988).
This could lead to a K+-induced depolarization of neurons, and
indeed, hyperpolarizing IPSPs are followed by a depolarization
(Diesz and Prince, 1989). GABA-evoked depolarizations of neu-
rons could reflect a glial-generated [K+] increase.
References
Andersen, P., R. Dingledine, L. Gjerstad, I. A. Langmoen, and A. M.
Laursen (1980) Two different responses of hippocampal pyramidal
cells to auolication of gamma-aminobutvric acid. J. Phvsiol. (Land.)
305: 2791i96. -
Backus, K. H., H. Kettenmann, and M. Schachner (1988) Effect of
benzodiazepines and pentobarbital on the GABA-induced depolari-
zation in cultured astrocytes. Glia I: 132-140.
Barker, J. L., R. N. McBurney, and D. A. Mathers (1983) Convulsant-
induced depression of amino acid responses in cultured mouse spinal
neurones studied under voltage clamp. Br. J. Pharmacol. 80: 619-
629.
Bormann, J. (1988) Electrophysiology of GABA, and GABA, receptor
subtypes. Trends Neurosci. I I: 112-l 16.
Bormann, J., and H. Kettenmann (1989) Patch clamp study ofGABA
receptor Cll channels in cultured astrocytes. Proc. Natl. Acad. Sci.
USA (in press).
Bormann, J., .and B. Sakmann (1984) Properties of inhibitory Cl-
channels and the action of diazepam. IUPHAR 9th Int. Cong. Phar-
macol., London, Abstr. S 13-S 14.
Bowery, N. G., D. A. Brown, R. D. White, and G. Yamini (1979) [Hl
Gamma-amino butyric acid uptake into neuroglial cells ofrat superior
cervical svmuathetic ganglia. J. Phvsiol. (Land.) 293: 5 l-74.
Connors, B.-W.; L. S. Benardo, and D.A. Prince (1984) Carbon dioxide
sensitivity of dye coupling among glia and neurons of the neocortex.
J. Neurosci. 4: 1324-1330.
Constanti, A., and M. Galvan (1978) Amino-acid-evoked depolari-
zation of electrically inexcitable neuroglial cells in the guinea pig
olfactory cortex slice. Brain Res. 153: 183-l 87.
Deisz, R. A., and D. A. Prince (1989) Frequency dependent depression
of inhibition in the guinea pig neocortex in vitro by GABA, receptor
feedback on GABA release. J. Physiol. (Lond.) (in press).
Gallagher, J. P., J. Nakamura, and P. Shinnick-Gallagher (1983) Ef-
fects of glial uptake and desensitization on the activity of -y-amino-
butyric acid (GABA) and its analogs at the cat dorsal root ganglion.
J. Pharmacol. Exp. Ther. 26: 876-884.
Gutnick, M. J., B. W. Conners, and B. R. Ransom (198 1) Dye-coupling
between glial cells in the guinea pig neocortical slice. Brain Res. 213:
486-492.
Hoppe, D., and H. Kettenmann (1989) GABA triggers a Cl- efflux
from cultured mouse oligodendrocytes. Neurosci. Lett. (in press).
Johnston, G. A. R., S. M. E. Kennedy, and D. Lodge (1978) Muscimol
uptake, release and binding in rat brain slices. J. Neurochem. 34:
1519-1523.
Keilhauer, G., D. H. Meier, S. Kuhlmann, J. Nieke, and M. Schachner
(1985) Astrocytes support incomplete differentiation of an oligoden-
drocvte precursor cell. EMBO J. 4: 2499-2504.
Kettenmann, H., and B. R. Ransom (1988) Electrical coupling between
astrocytes and between oligodendrocytes in mammalian cell cultures.
Glia 1: 64-73.
Kettenmann, H., and M. Schachner (1985) Pharmacological properties
of GABA, glutamate and aspartate induced depolarizations in cul-
tured astrocvtes. J. Neurosci. 5: 3295-3301.
Kettenmann, H., K. H. Backus, and M. Schachner (1984) Aspartate,
glutamate and gamma-aminobutyric acid depolarize cultured astro-
cytes. Neurosci. Lett. 52: 25-29.
Kettenmann, H., K. H. Backus, and M. Schachner (1987) GABA opens
Cl- channels in cultured astrocytes. Brain Res. 404: l-9.
Kettenmann, H., K. H. Backus, and M. Schachner (1988) GABA
receptors on cultured astrocytes. In Gliul Cell Receutors. H. K. Ki-
melberg, ed., pp. 95-106, Raven Press, New York. -
Korn. S. J.. and R. Dinaledine (1986) Inhibition of GABA untake in
therat hippocampal Gice. Brain Res. 368: 247-255. *
Krnjevic, K. (1974) Chemical nature of synaptic transmission in ver-
tebrates. Physiol. Rev. 2: 4 19-540.
Kmjevic, K. (1984) Some functional consequences of Na+-linked
GABA uptake. Neurosci. Lett. 47: 283-287.
The Journal of Neuroscience, October 1989, 9(10) 3533
Krnjevic, K., and S. Schwartz (1967) Some properties of unresponsive
cells in the cerebral cortex. Exp. Brain Res. 3: 306-3 19.
MacVicar, B. A., S. A. Crichton, D. M. Bumard, and F. W. Y. Tse
(1987) Phorbol ester induced endogenous membrane conductance
oscillations in astrocytes. Nature 329: 242-243.
MacVicar, B. A., K. Baker, and S. A. Crichton (1988) Kainic acid
evokes a potassium efflux from astrocytes. Neuroscience 25: 721-
725.
McCarthy, D. D., and J. De Vellis (1980) Preparation of separate
astroglial and oligodendroglial cell-cultures from rat cerebral tissue.
J. Cell. Biol. 85: 890-902.
Miiller, W., U. Misgeld, and H. D. Lux (1988) GABA mediated ion
movements in the guinea pig hippocampal slice. Pfltlgers Arch. (Suppl.)
411: R149.
Nadler, J. V., B. W. Perry, and C. W. Cotman (1978) Preferential
vulnerability of hippocampus to intraventricular kainic acid. In Kai-
nit Acid as a Tool in Neurobiology, E. G. McGeer, J. W. Olney, and
P. L. McGeer, eds., pp. 219-237, Raven, New York.
Reynolds, R., and N. Herschkowitz (1984) Uptake of [jH] GABA by
oligodendrocytes in dissociated brain cell culture: A combined au-
toradiographic and immunocytochemical study. Brain Res. 322: 17-
31.
Schousboe, A. (198 1) Transport and metabolism of glutamate and
GABA in neurons and glial cells. Int. Rev. Neurobiol. 22: l-45.
Shon, F., and Kelly, J. S. (1974) The characterization of [H3] GABA
uptake into satellite glial cells of rat sensory ganglia. Brain Res. 66:
289-300.
Stewart, W. W. (198 1) Lucifer dyes-highly fluorescent dyes for bi-
ological tracing. Nature 292: 17-2 1.
Study, R. E., and J. L. Barker (198 1) Diazepam and (-)pentobarbital
fluxuation analysis reveals different mechanisms for potentiation of
gamma-aminobutyric acid responses in cultured central neurons. Proc.
Natl. Acad. Sci. USA 78: 7 180-7 184.
White, W. F., and S. R. Snodgrass (1979) Isoguavacine binding uptake
and release: Relation to the GABA system. J. Neurochem. 33: 18 l-
189.