gabaergic input to cholinergic nucleus basalis neurons

GABAERGIC INPUT TO CHOLINERGIC NUCLEUS BASALISNEURONS

A. KHATEB,* P. FORT,*§ S. WILLIAMS,*¶ M. SERAFIN,* M. MU}HLETHALER*and B. E. JONES‡†

*Departement de Physiologie, CMU, 1 Rue Michel-Servet, 1211 Geneve 4, Switzerland

†Department of Neurology and Neurosurgery, McGill University, Montreal Neurological Institute,3801 University Street, Montreal, Quebec, Canada, H3A 2B4

Abstract––The potential influence of GABAergic input to cholinergic basalis neurons was studied inguinea-pig basal forebrain slices. GABA and its agonists were applied to electrophysiologically-identifiedcholinergic neurons, of which some were labelled with biocytin and confirmed to be cholineacetyltransferase-immunoreactive. Immunohistochemistry for glutamate decarboxylase was also per-formed in some slices and revealed GABAergic varicosities in the vicinity of the biocytin-filled soma anddendrites of electrophysiologically-identified cholinergic cells. From rest (average "63 mV), the cholin-ergic cells were depolarized by GABA. The depolarization was associated with a decrease in membraneresistance and diminution in firing. The effect was mimicked by muscimol, the specific agonist for GABAA

receptors, and not by baclofen, the specific agonist for GABAB receptors, which had no discernible effect.The GABA- and muscimol-evoked depolarization and decrease in resistance were found to be post-synaptic since they persisted in the presence of solutions containing either high Mg2+/low Ca2+ ortetrodotoxin. They were confirmed as being mediated by a GABAA receptor, since they were antagonizedby bicuculline. The reversal potential for the muscimol effect was estimated to be 2"45 mV, which was"15 mV above the resting membrane potential. Finally, in some cholinergic cells, spontaneous subthresh-old depolarizing synaptic potentials (average 5 mV in amplitude), which were rarely associated with actionpotentials, were recorded and found to persist in the presence of glutamate receptor antagonists but to beeliminated by bicuculline.

These results suggest that GABAergic input may be depolarizing, yet predominantly inhibitory tocholinergic basalis neurons. ? 1998 IBRO. Published by Elsevier Science Ltd.

Key words: basal forebrain, glutamate decarboxylase, choline acetyltransferase, GABAA receptors,muscimol, sleep-wake states.

The cholinergic nucleus basalis neurons are known toplay an important role in cortical activation thatoccurs during the states of wakefulness and paradoxi-cal sleep (for review see Ref. 25). Emanating from thenerve terminals of these cells, acetylcholine releasein the cerebral cortex is known to be maximal dur-ing these states and minimal during slow-wavesleep.9,24,32,34,35,44 Of the potential afferent input tothe cholinergic neurons that could be responsible forsuch state-associated modulation, GABAergic inputwould be a likely candidate for the apparent decrease

in activity during slow-wave sleep. In fact, previousstudies had shown that local injections of GABAagonists into the basal forebrain resulted in adecrease of acetylcholine release from the cortex,indicating that GABA could provide an inhibitoryinput to these cells.8

A moderately dense plexus of GABAergic varicosi-ties is evident within the basal forebrain in the regionwhere the cholinergic neurons are distributed.18,37

Indeed, the majority of terminals in the basal fore-brain have been found to be GABAergic,10 andGABAergic terminals have been visualized on corti-cally projecting, presumed cholinergic cells in thisarea.23

In the present study, we investigated theGABAergic input to cholinergic nucleus basalisneurons by in vitro pharmacological study of GABA-mediated effects on electrophysiologically- andimmunohistochemically-identified cholinergic cells inguinea-pig basal forebrain slices.29 We also examinedby immunohistochemistry for glutamate decarboxy-lase (GAD), GABAergic varicosities in relation tothe biocytin-filled cholinergic neurons. The pharma-cological study of the role of GABA and its two

‡To whom correspondence should be addressed.§Present address: Departement de Medecine Experimentale,

INSERM U52, CNRS ERS 5645, Faculte de Medecine,Universite Claude Bernard, 69373 Lyon, France.

¶Present address: Department of Neuroscience, Universityof Calgary, Calgary, Alberta, Canada, TDN 4N1

Abbreviations: AMPA, á-amino-3-hydroxy-5-methyli-soxazole-4-propionate; BDHC, benzidine dihydrochlo-ride; ChAT, choline acetyltransferase; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; DAB, diaminobenzidine;D-AP5, (")-2-amino-5-phosphonopentanoic acid;GAD, glutamate decarboxylase; IPSP, inhibitory post-synaptic potential; NMDA, N-methyl--aspartate; RMP,resting membrane potential; TTX, tetrodotoxin.

Pergamon

Neuroscience Vol. 86, No. 3, pp. 937–947, 1998Copyright ? 1998 IBRO. Published by Elsevier Science Ltd

Printed in Great Britain. All rights reserved0306–4522/98 $19.00+0.00PII: S0306-4522(98)00094-3

937

major receptor subtypes, GABAA and GABAB, wasundertaken using current-clamp recording in theslice. Some of these results were reported previouslyin abstract form.41

EXPERIMENTAL PROCEDURES

Basal forebrain slices were obtained from young guinea-pigs (Cavia-Porcellus, OFA Strain, Arrare, Geneva) thatweighed ~100 to 300 g or by correspondence were approxi-mately four days to four weeks of age. Standard procedureswere followed for preparation of the slices.33 In brief,following anaesthesia with Nembutal (30 mg/kg, i.p.), theanimal was decapitated, the skull opened, and a tissue blockdissected containing the basal forebrain. The block wasfixed with cyanoacrylate to the stage of a vibrating micro-tome on which 400 µm-thick slices were cut. The slices werecollected and subsequently held in small vials containing asaline solution which was continuously oxygenated (95%O2/ 5% CO2) and held at room temperature. From there,slices were transferred to the recording chamber in whichthey were superfused at 32)C with the same oxygenatedsaline solution that contained 130 mM NaCl, 20 mMNaHCO3, 1.25 mM KH2PO4, 1.3 mM MgSO4, 5 mM KCl,2.4 mM CaCl2 and 10 mM glucose.

Intracellular recordings were obtained with 3 M K+-acetate containing glass microelectrodes (resistance 80–150 MÙ) using a bridge amplifier (IR-183 or IR-283,Neurodata Instruments, New York, U.S.A.). Data wererecorded on videotapes using a Neurocorder (DR-384 orDR-484, Neurodata) and analysed off-line on a computerusing a home-made interface and dedicated software (withup to 80 kHz of sampling frequency). The recording siteswere determined by reference to a map of choline acetyl-transferase (ChAT)-immunoreactive neurons in the basalforebrain of the guinea-pig (Jones, unpublished obser-vation) and were focussed upon the substantia innominatabeneath the lateral wing of the anterior commissure.18

GABA and its agonists or antagonists were dissolved insaline and applied in the bath. The GABAA agonist, mus-cimol, and GABAB agonist, baclofen, were tested and theGABAA antagonist, bicuculline was employed. The agonistswere applied briefly for 30–60 s, whereas the antagonist wasapplied for longer periods to examine its effect alone and inthe presence of the agonists. For blocking of excitatoryamino acid effects, 6-cyano-7-nitroquinoxaline-2,3-dione(CNQX), and (")-2-amino-5-phosphonopentanoic acid(D-AP5) were also applied in the bath. Drugs were obtainedfrom Sigma (Buchs, Switzerland) or Tocris Cookson(Bristol, U.K.). To test the postsynaptic action of the drugs,they were administered in the presence of solutionscontaining high Mg2+ (6.3 mM), low Ca2+ (0.1 mM) ortetrodotoxin (TTX, 1 µM, Sigma) for synaptic uncoupling.

In a certain number of experiments, electrophysiologi-cally characterized cells were filled with 1–2% biocytin(Sigma) or neurobiotin (Vector, Burlingame, CA, U.S.A.)during recording, as previously published.29 Following fixa-tion and freezing, sections were cut on a freezing microtomeand processed for dual staining using peroxidase techniques.Of a previously published sample in which electrophysio-logically characterized, biocytin-filled cells were establishedas ChAT-immunoreactive,29 three had been tested withGABA and/or muscimol and were plotted onto a map ofChAT+ cells in the guinea-pig basal forebrain to showtheir approximate position in the present publication. Inaddition, 20 slices containing cells electrophysiologicallyidentified as cholinergic and filled with biocytin (or neuro-biotin) were sectioned (at 20 µm thickness) and processedfor dual sequential staining of biocytin (or neurobiotin) andGAD using peroxidase–antiperoxidase immunohistochem-istry with a rabbit anti-GAD antiserum (1:3000; Chemicon,Irvine, CA, U.S.A.).20 For this purpose, GAD was immu-

nostained using diaminobenzidine (DAB) as a chromogenin either the first or second position. Biocytin (or neuro-biotin) was stained using avidin–horseradish peroxidase(Jackson Immunoresearch Laboratories, West Grove, PA,U.S.A.) with nickel enhancement of DAB, as a blackchromogen applied in the first position, or benzidinedihydrochloride (BDHC), as a blue granular chromogenapplied in the second position.

The biocytin-filled neurons and GAD+ varicosities weredrawn at high power (with a 63# oil-immersion objective)using a drawing tube attached to a Leitz Dialux microscope.For publication, camera lucida drawings were scanned andfinalized using CorelDRAW computer software. Colourslide photographs of the sections were scanned with aNikon slide scanner and processed with Adobe Photoshop.

RESULTS

GABAergic varicosities near recorded cholinergic cells

The effects of GABA and its agonists were testedupon neurons located within the basal forebrain andcharacterized electrophysiologically as possessingthose properties previously shown to be characteristicof immunohistochemically-identified cholinergiccells of this region.3,29 Three cells upon whichGABA and/or muscimol were tested in the presentstudy were previously identified as ChAT-immunoreactive.29 These cells were situated in thesubstantia innominata beneath the anterior commis-sure and within the area where a large population ofChAT-immunoreactive neurons are concentrated(Fig. 1A) and delineated as the magnocellularpreoptic nucleus in the rat brain.18

The presence of GABAergic terminals was exam-ined by immunohistochemical staining for GAD insections containing biocytin-filled neurons, whichhad been electrophysiologically identified as cholin-ergic. GAD+ varicosities were seen in the immediatevicinity of the biocytin-labelled neurons as well asthrough the entire substantia innominata (Fig. 2A).GAD+ elements were drawn or plotted in the imme-diate surround of those cells which were adequatelylabelled (n=5, Fig. 1B). In all cases, GAD+ varicosi-ties were apparent in the vicinity of the soma anddendrites of the biocytin-labelled neurons. This wasthe case for distal as well as proximal dendriticprocesses (Fig. 2A, B). GAD-immunostained nervecell bodies were also seen in the same sections(Fig. 2B).

GABA-induced depolarization in cholinergic cells

As depicted in Fig. 3, the intrinsic membraneproperties of nucleus basalis cholinergic neuronsrecorded in the present study were similar to thosedescribed previously.3,17,27–29 When such neuronswere depolarized with injection of current pulseswhile at rest, they responded by firing tonically (Fig.3A) at frequencies never exceeding 15 Hz. In con-trast, depolarizing the cells while being held at ahyperpolarized level with steady injection of current,yielded bursts of action potentials riding on low

938 A. Khateb et al.

Fig. 1. Location of GABA-responsive, biocytin-labelled, ChAT+ cells and of GABAergic varicosities inrelation to a biocytin-labelled cell. (A) Schematic drawing of a section through the basal forebrain slice ofthe guinea-pig, onto which have been mapped dual-stained biocytin-filled/ChAT+ cells29 that respondedto GABA and/or muscimol (large dots). They are shown in relation to single ChAT-immunostainedneurons located in this region (within a 25 µm-thick section, small dots). (B) Camera lucida drawing (fromone 20 µm-thick section) of GAD+ varicosities which are located in the immediate vicinity of abiocytin-filled cell that was electrophysiologically identified as cholinergic. ac, anterior commissure; CPu,

caudate–putamen; MS, medial septum; SI, substantia innominata.

GABAergic input to cholinergic neurons 939

threshold Ca2+ spikes (Fig. 3B). In addition, hyper-polarizing the cells by injection of current pulses wasfollowed by a delayed return to the baseline due tothe presence of an A-current (not shown). The data-base of the present study was composed of 69 cellsthat displayed these properties and had restingpotentials of at least "55 mV and action potentialamplitude of at least 55 mV. In 15 cells upon whichGABA (and/or muscimol) was tested, the averageresting membrane potential (RMP) was "62.9 mV(&1.0 mV, S.E.M.).

When applied in the bath, GABA (100 µM–10 mM) produced a membrane depolarization andinhibited firing in cells which were spontaneouslyactive (25% of the cells) (Fig. 3C). The depolarizingresponse occurred in almost every cholinergic neuronupon which GABA was tested (11/12), and a hyper-polarizing response was never observed. The effectwas accompanied by a decrease in membrane resist-

ance that was evidenced using manual voltage-clamp(Fig. 3C).

The receptor mediating the response to GABA wasinvestigated using the GABAA agonist, muscimol,and the GABAB agonist, baclofen. Bath applicationof muscimol (0.5–100 µM) produced a membranedepolarization and decrease in membrane resistance(40/47 cells), and it inhibited firing in those cellswhich were spontaneously active (Fig. 3D). A hyper-polarizing response was never observed in responseto muscimol. Bath application of baclofen (100–200 µM) produced no evident response either on themembrane potential or on the spontaneous discharge(10/10 cells, not shown).

The effect of muscimol was subsequently examinedin conditions of synaptic uncoupling to establishwhether it was postsynaptic. When applied in thepresence of solutions with high Mg2+/low Ca2+

(6.3 mM and 0.1 mM, respectively), muscimol

Fig. 2. Photomicrographs of GABAergic varicosities in relation to biocytin-labelled neurons, which wereelectrophysiologically identified as cholinergic. GAD is stained with a brown floccular chromogen (DAB)and biocytin with a blue granular chromogen (BDHC) in a sequential dual staining procedure. (A and A*)GAD+ varicosities in the area surrounding a biocytin-labelled cell within the substantia innominata (seeFig. 1B). (B and B*) GAD+ varicosities near the distal dendrite of another biocytin-labelled cell. A GAD+neuron is also found in this region. In A and B at both low and high (*) magnifications, note that onlyportions of DAB-stained GAD+ varicosities and BDHC-stained processes are in the plane of focus andthus visible within the 20 µm-thick sections. Examples of GAD+ boutons seen near dendrites and in the

plane of focus are shown by arrowheads. Scale bars=10 µm.


(100 µM, 6/6 cells, Fig. 3E) produced a depolariza-tion associated with a decrease in membrane resist-ance. Similarly in the presence of TTX (1 µM),muscimol (8/9 cells, Fig. 4, see below) produced amembrane depolarization and never evoked a mem-brane hyperpolarization. The depolarization variedfrom 2–7 mV in response to 10–20 µM muscimol(average 3.9&0.5 mV, n=8).

The effect of muscimol was antagonized with bicu-culline (100 µM), the GABAA antagonist, whenapplied prior to the application of muscimol (10–100 µM, 3/3 cells, Fig. 3F) or during the effect ofmuscimol (6/6 cells, not shown). This antagonismwas also observed in the presence of TTX.

Given its apparent postsynaptic action, the voltagedependence of muscimol’s effect was examined at

Fig. 3. Electrophysiological properties of cholinergic neurons and their response to GABA and muscimol.(A) From rest, a depolarizing current pulse evoked tonic discharge. (B) With d.c. hyperpolarization, thesame pulse evoked a rhythmic burst discharge in the same cell. Level of RMP, shown by arrowheads:"69 mV. (C) Chart paper record showing GABA effects on the membrane potential (upper trace) anddischarge rate (lower trace). Brief bath application of GABA (500 µM) produced a membrane depolari-zation that was accompanied by a decrease in membrane resistance, which is evidenced by manualvoltage-clamp allowing comparison of potentials produced by short hyperpolarizing current pulses duringthe maximum effect to those prior to application. These membrane effects were accompanied by a decreasein the spontaneous discharge of the cell. (D) In the same cell, bath application of muscimol (20 µM)produced similar effects. RMP: "58 mV. (E) In high Mg2+/low Ca2+ solution, application of muscimol(100 µM) still produced a membrane depolarization associated with a decrease in membrane resistance, asevident with manual voltage-clamp. RMP: "67mV. (F) The depolarization evoked by muscimol (100 µM)

was blocked by bicuculline (100 µM). RMP: "59 mV.


different levels of membrane potential in the presenceof TTX (Fig. 4). As compared to that at the restingmembrane potential ("56 mV), the membrane depo-larization increased in amplitude when the cell washyperpolarized to around "70 mV, and disappearedwhen the cell was depolarized to around "45 mV(Fig. 4). The average reversal potential for themuscimol effect was estimated at "45.7&1.2 mV(range: "44 to "48 mV), which was on average16.3 mV&3.18 above the resting membrane potential("62.0&4.2 mV) in these cells (n=3).

Spontaneous depolarizing potentials in cholinergic cells

As illustrated in Fig. 5, some cholinergic neuronsshowed spontaneous depolarizing potentials in theslice (n=5). The amplitude of these potentials variedfrom 3–9 mV (5.1&0.2 mV) and their frequency from1–7 Hz (Fig. 5A). In most cells (3/5), the potentialswere not associated with action potentials, and inthese (with RMP of 63.7 mV&1.2) the frequency ofthe depolarizing potentials was 1–2/s. In the remain-ing cells (2/5), the potentials were very occasionallyassociated with action potentials (not shown), and inthese (with RMPs of "56 and "57 mV), the averagefrequency of the depolarizing potentials was 7/sand of the action potentials was 0.1 and 0.3/s.Bath applications of the á-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA)-kainatereceptor antagonist, CNQX (50 µM, 3/3 cells), orN-methyl--aspartate (NMDA) antagonist, D-AP5

(50 µM, 3/3 cells), did not eliminate these spon-taneous depolarizing potentials (Fig. 5C, E1). Incontrast, bath application of bicuculline, at the con-centration (100 µM) previously shown to antagonizethe effect of muscimol, eliminated the spontaneousdepolarizing potentials and shifted the membranepotential to a slightly more hyperpolarized level (by<5 mV, 4/5 cells, Fig. 5B). A depolarizing shift inmembrane potential was never seen with bicuculline.The hyperpolarizing shift was associated with anincrease in resistance of 21 to 33% (27.0&3.46%,measured in three cells, Fig. 5C, D, E2).

DISCUSSION

The present study demonstrated that cholinergicnucleus basalis neurons (i) are surrounded byGABAergic varicosities, (ii) are depolarized yetinhibited by GABA and the GABAA agonist, musci-mol, and (iii) receive spontaneous subthreshold depo-larizing synaptic potentials which are eliminated bythe GABAA antagonist, bicuculline. These resultsindicate that the cholinergic cells receive a depolariz-ing GABAergic input which would tend to inhibitdischarge in these cells.

Immunohistochemical evidence for GABAergic inputto cholinergic basalis neurons

The present study allowed visualization of GAD-immunoreactive varicosities in relation to the soma

Fig. 4. Voltage dependence of postsynaptic muscimol effect. In the presence of TTX, muscimol produceda depolarization and decrease in resistance, as shown with manual voltage-clamp, with the cell at rest (B)and a greater depolarization with the cell hyperpolarized to "73 mV by d.c. current injection (C).Muscimol produced no change in membrane potential when the cell was depolarized to "45 mV (A).

RMP: "56 mV.


and dendrites of biocytin-filled neurons which wereelectrophysiologically identified as cholinergic in theguinea-pig basal forebrain slice. GAD+ varicositieswere apparent in the immediate vicinity of the cellbody and proximal to distal dendrites of the cholin-ergic neurons. In electron microscopic studies in therat brain, GAD-positive terminals have been ident-ified on cortically projecting, presumed cholinergicneurons in this region and found to representmore than 50% of the terminals on the soma andproximal dendrites of the cells.23 In the morerostral substantia innominata or ventral pallidum,amygdala-projecting, ChAT-immunoreactive cellswere also shown to be contacted on their soma andproximal dendrites by GAD-immunoreactive termi-nals.56 The GABAergic input to cholinergic basaliscells may derive as in other areas from localGABAergic neurons, since large numbers of small tomedium, in addition to larger, GAD+ cells areco-distributed with ChAT+ neurons through thebasal forebrain.18 GABAergic input could also arisefrom extrinsic sources, since GABAergic neurons arewidely distributed through the forebrain, includingthe ventral striatum, which projects to cholinergiccells in the anterior substantia innominata-ventralpallidum,55 the amygdala,40 the preoptic region andthe hypothalamus.19,37

Pharmacological evidence for depolarizing inhibitoryGABAergic action

To our knowledge, the present study representsthe first report on the effects of GABA and its

agonists on identified cholinergic basalis neurons.GABA produced a membrane depolarizationand increased conductance which were associatedwith an inhibition of tonic spike discharge inspontaneously active cholinergic cells. These effectswere mimicked by the GABAA agonist, muscimol,and not by the GABAB agonist, baclofen, suggestingthat they were mediated by a GABAA typereceptor. The effects of both GABA and muscimolpersisted in conditions of synaptic uncouplingindicating that they were postsynaptic, and wereblocked by bicuculline confirming that they weremediated by a GABAA receptor. A postsynapticdepolarizing inhibitory action of GABA, mediatedthrough the GABAA receptor has been describedon other cells in the brain, including the hippo-campal pyramidal cells,2,4,5 neocortical pyramidalcells,6,11,16 dentate gyrus granule cells46 andbrainstem neurons.22

In hippocampal pyramidal cells, the depolarizingresponse to GABA was shown to be prominent in theearly postnatal period (the first week) and to shift toa hyperpolarizing response with development (overthe second week) to adulthood (after the secondweek) in the rat.7 In the cholinergic cells studied herein the guinea-pig, which is more mature at birth thanthe rat, a hyperpolarizing response to GABA ormuscimol was not observed in animals of any weight-age (including animals of up to 300 g or approxi-mately four weeks of age), indicating that the lack ofa hyperpolarizing and presence of a depolarizingresponse to GABA cannot be attributed to immatu-rity of the animals. It is also significant that GABA

Fig. 5. Spontaneous depolarizing synaptic potentials that are blocked by bicuculline. (A–B) Membranepotential recording illustrating the presence of spontaneous depolarizing synaptic potentials (A) whichwere eliminated by bath application of bicuculline (100 µM, B). The drug also produced a shift in themembrane potential to below resting (indicated by dashed line). RMP: "57 mV. (C–E) In another cell,spontaneous depolarizing potentials persisted in the presence of CNQX and D-AP5 (C, trace at arrowenlarged in E1, where time-scale equals 500 ms as in A) but were eliminated by application of bicuculline(100 µM, D). This effect of bicuculline was associated with an increase in membrane resistance, as evidentby the increased voltage deflection in response to hyperpolarizing current pulse injections (response incontrol, Ctr, indicated by dashed lines in C and D and superimposed upon that with bicuculline, Bic, in

E2, showing enlargements of responses near arrows in C and D). RMP: "62 mV.


hyperpolarizing responses have been observed hereunder the same conditions in the guinea-pig basalforebrain slice but on non-cholinergic basalisneurons.26

Depolarizing GABAA-mediated responses havenonetheless also been observed in mature animals(above), including rats and guinea-pigs, but in thesecases have been found to be (i) site dependent,being more prevalent on dendrites than on soma,(ii) concentration and time dependent, being moreprevalent with higher concentrations and prolongeddelivery or release of GABA, and (iii) ion depen-dent, being more prevalent in association with cur-rents carried by ions other than or in addition tochloride. In order to assess the possible role of thesefactors in producing the depolarizing responseobserved here in the cholinergic cells, we will con-sider each factor in sequence. (1) In hippocampaland neocortical pyramidal cells, depolarizing effectsof GABA applied by local iontophoretic applicationwere predominant on dendrites, whereas hyperpo-larizing effects were predominant on the soma ofthese cells.2,4,11,16 In the present study using bathapplications of agonists in the slice, we could notassess whether there was a differential GABAA

response on dendrites versus soma. However, inview of the fact that we did not observe anyhyperpolarizing response, we consider that thedepolarizing response could be present on either orboth dendrites and soma of the cholinergic neurons.(2) In the hippocampal pyramidal cells, the depolar-izing response to GABA applied by local pressureinjection was originally found to be present whenrelatively high concentrations of GABA wereapplied for a short time or lower concentrations fora long time.54 The responses were nonethelessbiphasic, with the depolarizing response always fol-lowing an initial transient hyperpolarization. Giventhe use of bath application here in the slice, it couldbe possible that the depolarizing response to GABAand muscimol by the cholinergic cells would be dueto relatively high concentrations of the agonistsapplied over relatively long periods. However, thedepolarization to GABA and muscimol was pro-duced at relatively low concentrations (100 and0.5 µM, respectively). Furthermore, no transienthyperpolarization occurred in response to GABA ormuscimol, suggesting the unique presence of aGABAA depolarizing response in these cells. (3)Whereas it is well established that GABAA hyper-polarizing current is carried by chloride (Cl") andreverses at the same potential as the equilibriumpotential for Cl" (ECl) in the soma of hippocampalneurons, it is doubted that the GABAA depolarizingresponse in the dendrites or soma of these cells iscarried solely by Cl" ions. Indeed, there is noconvincing evidence for ECl being at morepositive potentials in the dendrites or changingwith prolonged stimulation in the soma of thepyramidal cells in mature animals.21,30,42 Here, in

the cholinergic cells, the reversal potential for themuscimol-evoked response was &"45 mV whichwas 215 mV positive to the average RMP of thesecells and would thus appear to be positive to thecommon ECl measured in mature centralneurons.14,52 These results would suggest that as issuspected for pyramidal cell dendrites and soma21,30

and recently, brainstem auditory neurons,22 theGABAA-mediated depolarizing current in choliner-gic neurons may not be carried exclusively by Cl".More recently to explain site- and activity-dependent depolarizing responses to GABA in hip-pocampal pyramidal cells, a role was suggested forbicarbonate (HCO3

") ions, which permeate theGABAA ionophore and which have an equilibriumpotential positive to the resting membrane poten-tial.42,47 The possible involvement of HCO3

", inaddition to Cl", ions remains to be investigated infuture studies on the GABAA depolarizing responsein cholinergic basalis neurons.

Electrophysiological evidence for synaptic depolarizingGABAergic input

Spontaneous inhibitory postsynaptic potentials(IPSPs) and currents that are blocked by bicucullinehave been recorded in brain slices from multipleregions, including most particularly the hippocam-pus,1,39,45 and have indicated the presence of a tonicrelease of GABA due to spontaneous activity ofintact interneurons and/or transmitter release fromsectioned fibres. Spontaneous depolarizing poten-tials, which are blocked by bicuculline and not byexcitatory amino acid receptor antagonists, havealso been observed in the hippocampus underparticular conditions including, in the presence of4-aminopyridine,36,43 in the presence of zinc31 and inimmature rats.7 Here in the basal forebrain slice, werecorded in the cholinergic cells, spontaneous depo-larizing potentials in absence of any pharmacologicaltreatment in mature guinea-pigs. These potentialswere <10 mV in amplitude (5 mV on average) and inonly rare cases (<5% of potentials in 2/5 cells)reached spike threshold. They were not eliminated byAMPA and NMDA antagonists, but were eliminatedby the GABAA antagonist, bicuculline. In addition,the membrane potential was slightly hyperpolarized(<5 mV) by the application of bicuculline, in associ-ation with an increase in membrane resistance(230%). These results suggest that the spontaneousdepolarizing potentials and membrane depolariza-tion are synaptically mediated by GABA releasedfrom GABAergic terminals surrounding the cholin-ergic cells. They also suggest that spontaneousGABAergic input to the cholinergic neurons wouldbe associated with potentials that depolarize themembrane but do not reach threshold for sodiumspikes and that furthermore bring the membrane to alevel at which low-threshold calcium spikes areinactivated.


Physiological significance of depolarizing GABAergicinput

That GABAA-mediated depolarizations can beinhibitory on mature neurons has been established inboth vertebrate and invertebrate nervous systems.The inhibitory effect of the depolarization has beenattributed to shunting of the cell membrane due toincreased conductance.4,15,46 In the cholinergic basa-lis neurons, the GABA- and muscimol-evokeddepolarization was associated with a decrease inspontaneous tonic spiking which could be explainedby the concomitant decrease in membrane resistance,according to the latter mechanism. On the otherhand, it is possible that the GABAA-dependent spon-taneous depolarizing synaptic potentials observedwould not necessarily be inhibitory and could also beexcitatory under certain circumstances, as has beenshown to be the case with pharmacologically-inducedGABAA-mediated spontaneous depolarizing poten-tials in the hippocampus.36,43 In the cholinergic cells,however, the GABAA-mediated depolarizationwould be associated with inactivation of the low-threshold calcium spikes and would thus preventphasic low-threshold bursting in these cells.29 Basedupon our parallel results of the muscimol-evokedchanges in discharge and membrane properties andof the bicuculline-sensitive spontaneous potentialsand changes in membrane properties, we thus con-sider that GABAergic input would tend to inhibittonic spiking and would most definitely prevent low-threshold bursting in the cholinergic basalis neurons.

Although it may be unwarranted to speculateconcerning state-dependent GABAergic input tocholinergic basalis cells, it is nonetheless interestingto consider the potential differences between theeffect of this input to basalocortical neurons with thatto thalamocortical neurons during slow-wave sleep.Thalamocortical relay cells are hyperpolarized byGABAergic input from nucleus reticularis neurons,that switches their mode of firing from tonic to phasicthrough activation of low-threshold calcium spikes inthe transition into and during slow-wave sleep (forreview see Refs 48, 49 and 50). The hyperpolarizationis mediated by fast GABAA IPSPs from which re-bound low-threshold calcium spikes appear in associ-ation with spindles.53 The reversal potential for theGABAA response in the relay cells was recentlymeasured as "82 mV.52 In addition, a slower hyper-polarization is mediated by GABAB receptors thatmay prime low-threshold bursting and favour sloweroscillations, which depend upon the reciprocal playof the low-threshold calcium current with thehyperpolarization-activated inward rectifying current

intrinsic to these neurons (for review see Ref. 12).Through these synaptic and intrinsic mechanisms,thalamocortical relay neurons discharge in rhythmiclow-threshold bursts during slow-wave sleep inassociation with spindle, delta and slow oscillations(for review see Refs 48 and 50). In contrast, ifbasalocortical neurons were submitted to GABAer-gic input during slow-wave sleep, they would bedepolarized through GABAA receptors. This depo-larization would be associated with increased con-ductance and potential shunting of the membrane,thus possibly inhibiting tonic discharge and mostparticularly, preventing the phasic, low-thresholdburst discharge. Such an action could explain thedecrease in acetylcholine release in the cortex thatoccurs during slow-wave sleep9,24 and would alsoconform to the demonstrated decrease in corticalacetylcholine release that occurs following injectionof muscimol into the basal forebrain.8 Althoughauthors have speculated upon the cholinergic identityof basal forebrain neurons recorded in vivo and thusupon changes in firing mode and frequency by thecholinergic cells across sleep-wake states,13,38,51 suchidentification and information awaits immunohisto-chemical staining of biocytin-filled neurons recordedin vivo as performed here for the cells recordedin vitro.

CONCLUSIONS

The results of the present study suggest that cholin-ergic basalis neurons receive a depolarizing inputfrom GABAergic terminals, which are locatedaround the soma and dendrites of these cells. Basedupon the observed parallel decrease in spontaneousdischarge and membrane resistance, the GABAA-mediated depolarization appears to be inhibitory. Asevidenced by bicuculline-sensitive, spontaneous sub-threshold depolarizing synaptic potentials and tonicdepolarization of the membrane, the GABAergicinput would not appear to reach threshold forsodium spikes and would moreover bring the mem-brane to a level at which low-threshold calciumspikes are inactivated.

Acknowledgements—We thank Lynda Mainville andDaniele Machard for their excellent technical assistance.This work was supported by grants from the Swiss FondsNational, the Sandoz and Roche Foundations to MM andthe Canadian Medical Research Council (MRC) to BEJ. PFwas supported by fellowships from Fondation Fyssen andFondation pour la Recherche Medicale (FRM); SW was therecipient of a post-doctoral fellowship from the CanadianMRC.

REFERENCES

1. Alger B. E. and Nicoll R. A. (1980) Spontaneous inhibitory postsynaptic potentials in hippocampus: mechanism fortonic inhibition. Brain Res. 20, 195–200.

2. Alger B. E. and Nicoll R. A. (1982) Pharmacological evidence for two kinds of GABA receptor on rat hippocampalpyramidal cells studied in vitro. J. Physiol., Lond. 328, 125–141.


3. Alonso A., Khateb A., Fort P., Jones B. E. and Muhlethaler M. (1996) Differential oscillatory properties of cholinergicand non-cholinergic nucleus basalis neurons in guinea pig brain slice. Eur. J. Neurosci. 8, 169–182.

4. Andersen P., Dingledine R., Gjerstad L., Langmoen I. A. and Mosfeldt Laursen A. (1980) Two different responses ofhippocampal pyramidal cells to application of gamma-aminobutyric acid. J. Physiol., Lond. 305, 279–296.

5. Avoli M. (1992) Synaptic activation of GABAA receptors causes a depolarizing potential under physiologicalconditions in rat hippocampal pyramidal cells. Eur. J. Neurosci. 4, 16–26.

6. Avoli M. and Olivier A. (1989) Electrophysiological properties and synaptic responses in the deep layers of the humanepileptogenic neocortex in vitro. J. Neurophysiol. 61, 589–606.

7. Ben-Ari Y., Cherubini E., Corradetti R. and Gaiarsa J.-L. (1989) Giant synaptic potentials in immature rat CA3hippocampal neurones. J. Physiol., Lond. 416, 303–325.

8. Casamenti F., Deffenu G., Abbamondi A. L. and Pepeu G. (1986) Changes in cortical acetylcholine output inducedby modulation of the nucleus basalis. Brain Res. Bull. 16, 689–695.

9. Celesia G. G. and Jasper H. H. (1966) Acetylcholine released from cerebral cortex in relation to state of activation.Neurology 16, 1053–1064.

10. Chang H. T., Tian Q. and Herron P. (1995) GABAergic axons in the ventral forebrain of the rat: an electronmicroscopic study. Neuroscience 68, 207–220.

11. Connors B. W., Malenka R. C. and Silva L. R. (1988) Two inhibitory postsynaptic potentials, and GABAA andGABAB receptor-mediated responses in neocortex of rat and cat. J. Physiol., Lond. 406, 443–468.

12. Crunelli V. and Leresche N. (1991) A role for GABAB receptors in excitation and inhibition of thalamocortical cells.Trends Neurosci. 14, 16–21.

13. Detari L. and Vanderwolf C. H. (1987) Activity of identified cortically projecting and other basal forebrain neuronesduring large slow waves and cortical activation. Brain Res. 437, 1–8.

14. Ebihara S., Shirato K., Harata N. and Akaike N. (1995) Gramicidin-perforated patch recording: GABA response inmammalian neurones with intact intracellular chloride. J. Physiol. 484, 77–86.

15. Edwards D. H. (1990) Mechanisms of depolarizing inhibition at the crayfish giant motor synapse. I. Electrophysiol-ogy. J. Neurophysiol. 64, 532–540.

16. El-Beheiry H. and Puil E. (1990) Unusual features of GABA responses in layers IV–V neurons of neocortex. Neurosci.Lett. 119, 83–85.

17. Fort P., Khateb A., Pegna A., Muhlethaler M. and Jones B. E. (1995) Noradrenergic modulation of cholinergicnucleus basalis neurons demonstrated by in vitro pharmacological and immunohistochemical evidence in the guineapig brain. Eur. J. Neurosci. 7, 1502–1511.

18. Gritti I., Mainville L. and Jones B. E. (1993) Codistribution of GABA- with acetylcholine-synthesizing neurons in thebasal forebrain of the rat. J. comp. Neurol. 329, 438–457.

19. Gritti I., Mainville L. and Jones B. E. (1994) Projections of GABAergic and cholinergic basal forebrain andGABAergic preoptic-anterior hypothalamic neurons to the posterior lateral hypothalamus of the rat. J. comp. Neurol.339, 251–268.

20. Gritti I., Mainville L., Mancia M. and Jones B. E. (1997) GABAergic and other non-cholinergic basal forebrainneurons project together with cholinergic neurons to meso- and iso-cortex in the rat. J. comp. Neurol. 383, 163–177.

21. Grover L. M., Lambert N. A., Schwartzkroin P. A. and Teyler T. J. (1993) Role of HCO3" ions in depolarizing

GABAA receptor-mediated responses in pyramidal cells of rat hippocampus. J. Neurophysiol. 69, 1541–1555.22. Hyson R. L., Reyes A. D. and Rubel E. W. (1995) A depolarizing inhibitory response to GABA in brainstem auditory

neurons of the chick. Brain Res. 677, 117–126.23. Ingham C. A., Bolam J. P. and Smith A. D. (1988) GABA-immunoreactive synaptic boutons in the rat basal

forebrain: comparison of neurons that project to the neocortex with pallidosubthalamic neurons. J. comp. Neurol. 273,263–282.

24. Jasper H. H. and Tessier J. (1971) Acetylcholine liberation from cerebral cortex during paradoxical (REM) sleep.Science 172, 601–602.

25. Jones B. E. (1993) The organization of central cholinergic systems and their functional importance in sleep-wakingstates. In Cholinergic Function and Dysfunction, (ed. Cuello A. C.), pp. 61–71. Progress in Brain Research. Vol. 98,Elsevier, Amsterdam.

26. Khateb A. (1998) Physiologie et Neuromodulation d’un Systeme Relais de l’Activation Corticale: les NeuronesCholinergiques du Noyau Basal. Universite de Geneve, Geneva.

27. Khateb A., Fort P., Alonso A., Jones B. E. and Muhlethaler M. (1993) Pharmacological and immunohistochemicalevidence for a serotonergic input to cholinergic nucleus basalis neurons. Eur. J. Neurosci. 5, 541–547.

28. Khateb A., Fort P., Serafin M., Jones B. E. and Muhlethaler M. (1995) Rhythmical bursts induced by NMDA incholinergic nucleus basalis neurones in vitro. J. Physiol., Lond. 487.3, 623–638.

29. Khateb A., Muhlethaler M., Alonso A., Serafin M., Mainville L. and Jones B. E. (1992) Cholinergic nucleus basalisneurons display the capacity for rhythmic bursting activity mediated by low threshold calcium spikes. Neuroscience 51,489–494.

30. Lambert N. A., Borroni A. M., Grover L. M. and Teyler T. J. (1991) Hyperpolarizing and depolarizing GABAA

receptor-mediated dendritic inhibition in area CA1 of the rat hippocampus. J. Neurophysiol. 66, 1538–1548.31. Lambert N. A., Levitin M. and Harrison N. L. (1992) Induction of giant depolarizing potentials by zinc in area CA1

of the rat hippocampus does not result from block of GABAB receptors. Neurosci. Lett. 135, 215–218.32. Lehmann J., Nagy J. I., Atmadja S. and Fibiger H. C. (1980) The nucleus basalis magnocellularis: the origin of a

cholinergic projection to the neocortex of the rat. Neuroscience 5, 1161–1174.33. Llinas R. and Sugimori M. (1980) Electrophysiological properties of in vitro Purkinje cell somata in mammalian

cerebellar slices. J. Physiol., Lond. 305, 171–195.34. Lo Conte G., Casamenti F., Bigi V., Milaneschi E. and Pepeu G. (1982) Effect of magnocellular forebrain nuclei

lesions on acetylcholine output from the cerebral cortex, electrocorticogram and behaviour. Archs ital. Biol. 120,176–188.

35. Marrosu F., Portas C., Mascia S., Casu M. A., Fa M., Giagheddu M., Imperato A. and Gessa G. L. (1995)Microdialysis measurement of cortical and hippocampal acetylcholine release during sleep-wake cycle in freely movingcats. Brain Res. 671, 329–332.


36. Michelson H. B. and Wong R. K. S. (1991) Excitatory synaptic responses mediated by GABAA receptors in thehippocampus. Science 253, 1420–1423.

37. Mugnaini E. and Oertel W. H. (1985) An atlas of the distribution of GABAergic neurons and terminals. In Handbookof Chemical Neuroanatomy. GABA and Neuropeptides in the CNS, Part I (eds Bjorklund A. and Hokfelt T.), Vol. 4,pp. 436–608. Elsevier Science Publishers, Amsterdam.

38. Nunez A. (1996) Unit activity of rat basal forebrain neurons: relationship to cortical activity. Neuroscience 72,757–766.

39. Otis T. S., Staley K. J. and Mody I. (1991) Perpetual inhibitory activity in brain slices generated by spontaneousGABA release. Brain Res. 545, 142–150.

40. Pare D. and Smith Y. (1994) GABAergic projection from the intercalated cell masses of the amygdala to the basalforebrain in cats. J. comp. Neurol. 344, 33–49.

41. Pegna A., Khateb A., Fort P., Jones B. E. and Muhlethaler M. (1993) GABAergic action on identified cholinergicneurones of the nucleus basalis in guinea-pig brain slices. Soc. Neurosci. Abstr. 19, 1137.

42. Perkins K. L. and Wong R. K. S. (1996) Ionic basis of the postsynaptic depolarizing GABA response in hippocampalpyramidal cells. J. Neurophysiol. 76, 3886–3894.

43. Perrault P. and Avoli M. (1991) Effects of low concentrations of 4-aminopyridine on CA1 pyramidal cells of thehippocampus. J. Neurophysiol. 61, 953–970.

44. Rasmusson D. D., Clow K. and Szerb J. C. (1994) Modification of neocortical acetylcholine release and electroen-cephalogram desynchronization due to brainstem stimulation by drugs applied to the basal forebrain. Neuroscience 60,665–677.

45. Staley K. J. and Mody I. (1991) Integrity of perforant path fibres and the frequency of action potential independentexcitatory and inhibitory synaptic events in dentate gyrus granule cells. Synapse 9, 219–224.

46. Staley K. J. and Mody I. (1992) Shunting of excitatory input to dentate gyrus granule cells by a depolarizing GABAA

receptor-mediated postsynaptic conductance. J. Neurophysiol. 68, 197–212.47. Staley K. J., Soldo B. L. and Proctor W. R. (1995) Ionic mechanisms of neuronal excitation by inhibitory GABAA

receptors. Science 269, 977–981.48. Steriade M., Contreras D. and Amzica F. (1994) Synchronized sleep oscillations and their paroxysmal developments.

Trends Neurosci. 17, 199–208.49. Steriade M. and Deschenes M. (1984) The thalamus as a neuronal oscillator. Brain Res. Rev. 8, 1–63.50. Steriade M. and Llinas R. R. (1988) The functional states of the thalamus and the associated neuronal interplay.

Physiol. Rev. 68, 649–742.51. Szymusiak R. and McGinty D. (1989) Sleep-waking discharge of basal forebrain projection neurons in cats. Brain Res.

Bull. 22, 423–430.52. Ulrich D. and Huguenard J. R. (1997) Nucleus-specific chloride homeostasis in rat thalamus. J. Neurosci. 17,

2348–2354.53. von Krosigk M., Bal T. and McCormick D. A. (1993) Cellular mechanisms of a synchronized oscillation in the

thalamus. Science 261, 361–364.54. Wong R. K. S. and Watkins D. B. (1982) Cellular factors influencing GABA response in hippocampal pyramidal cells.

J. Neurophysiol. 48, 938–951.55. Zaborszky L. and Cullinan W. E. (1992) Projections from the nucleus accumbens to cholinergic neurons of the ventral

pallidum: a correlated light and electron microscopic double-immunolabeling study in rat. Brain Res. 570, 92–101.56. Zaborszky L., Heimer L., Eckenstein F. and Leranth C. (1986) GABAergic input to cholinergic forebrain neurons: an

ultrastructural study using retrograde tracing of HRP and double immunolabeling. J. comp. Neurol. 250, 282–295.

(Accepted 9 February 1998)


gabaergic input to cholinergic nucleus basalis neurons

Documents