Proc. Natl. Acad. Sci. USAVol. 82, pp. 1558-1562, March 1985Neurobiology

GABAB-receptor-activated K+ current in voltage-clamped CA3pyramidal cells in hippocampal cultures

(badofen/inhlbition/caldum/barium)

BEAT H. GAHWILER* AND DAVID A. BROWNt*Preclinical Research, Sandoz Ltd., CH 4002 Basel, Switzerland; and fMedical Research Council Neuropharmacology Research Group, Department ofPharmacology, School of Pharmacy, University of London, 29/39 Brunswick Square, London, WC1N lAX, England

Communicated by Ewald R. Weibel, October 9, 1984

ABSTRACT GABAB receptors are a subclass of receptorsfor y-amino-n-butyric acid (GABA) that are also activated bythe antispastic drug fi-p-chlorophenyl-GABA (baclofen). Oneeffect of baclofen is to inhibit excitatory transmission fromCA3 to CA, hippocampal pyramidal cells. To identify the ionicmechanism of GABAB-receptor-mediated depression, we havestudied the effect of baclofen and GABA on ionic currents involtage-clamped CA3 pyramidal cell somata in rat hippocam-pal slice cultures. Baclofen (10 pM) induced an inwardly recti-fying outward current that reversed at -74 ± 4.3 mV (mean* SD). This appeared to be a K+ current since (') its reversalpotential showeo the expected shift when extracellular K+ con-centration was changed and (ii) it was blocked by externalBa2' or internal Cs'. The action of baclofen was closely imi-tated by GABA after the GABAA-mediated Cl- current hadbeen abolished with pitrazepin (10 #M); under these condi-tions, GABA (100 pM) also produced an inwardly rectifying,Ba2+-sensitive current with a reversal potential identical tothat of the badofen-induced current. When outward currentswere blocked with internal Cs', the residual inward voltage-dependent Ca21 current was not changed by baclofen. It isconcluded that the primary effect of GABAB-receptor activa-tion in these neurones is to increase K+ permeability ratherthan to reduce Ca21 permeability.

Two subclasses of receptors for the central inhibitory neuro-transmitter 4-amino-n-butyric acid (y-aminobutyric acid,GABA) have been defined and designated GABAA and GA-BAB receptors (1, 2). GABAA receptors are selectively acti-vated by the GABA analogue muscimol and mediate an in-creased Cl- conductance (3, 4). Effects of GABAA activa-tion are blocked by convulsants such as bicuculline andpicrotoxin. A selective agonist for the bicuculline-insensitiveGABAB receptor is the antispastic agent 8-p-chlorophenyl-GABA (baclofen) (1). The principal effect of baclofen is todepress transmitter release from certain excitatory path-ways, including intraspinal primary afferents (5-14) andSchaffer commissural fibers in the hippocampus (15-18).The ionic mechanism responsible for this effect of baclo-

fen action is not clear. In cell bodies of primary afferent fi-bers baclofen shortens the action potential (19, 20), fromwhich it has been suggested that it reduces voltage-depen-dent Ca2+ conductance (19). In hippocampal cell somata, onthe other hand, baclofen induces a membrane hyperpolariza-tion and an increased input conductance, suggesting an in-creased K+ permeability (21-23); if such an effect occurredat nerve terminals it could also affect action potential dura-tion and reduce Ca2' influx in the absence of a direct effecton Ca2+ currents.To obtain better resolution of the ionic permeability

changes induced by baclofen, we have studied its effect on

voltage-clamped hippocampal neurones under conditions inwhich actions on K+ and Ca2+ conductance can be discrimi-nated. We find that baclofen induces a K+ current withoutinhibiting directly the slow inward Ca2+ current. We alsoshow that the effect of baclofen can be replicated by GABAwhen GABAA receptors are blocked, so supporting the view(2) that GABA may be the endogenous ligand for baclofenreceptors.

MATERIALS AND METHODSExperiments were performed on hippocampal slices pre-pared from 7-day-old Wistar rats and cultured organotypical-ly for 3-5 weeks in vitro as described previously (24, 25).The slice was superfused with Hanks' balanced salt solutionat 350C at a rate of 1 ml-min'. The composition of the solu-tion was (mM): Na+, 141.7; K+, 5.8; Cl-, 147.8; Ca2+, 2.8;Mg2+, 0.9; HCO3, 4.2; SO24-, 0.4; H2PO4, 0.4; D-glucose,5.6. Individual CA3 pyramidal cells were impaled with a sin-gle microelectrode containing either 3 M KCl or 2 M CsCl(tip resistance 25-35 MO or 15-25 MO, respectively) andvoltage-clamped with an amplifier (Axoclamp-2, Axon In-struments, Burlingame, CA) switching between current-passing and voltage-sampling modes at 2-6 kHz. A continu-ous voltage-output display was monitored to ensure that thevoltage drop across the electrode induced by the currentpulses had dissipated between voltage-sampling times, sothat the measured voltage referred solely to the cell mem-brane potential. Fast Na+ currents were eliminated by add-ing 1 jLM tetrodotoxin to the perfusion fluid. During the slowcurrents recorded in these experiments, measured voltageswere within 1-2 mV of the nominal command voltages. Allpotentials are corrected for electrode potentials registeredafter withdrawing the electrode from the cell. Baclofen wasobtained from CIBA-Geigy in (+), (-), and (±) forms; inmost experiments we used the (±) form.

RESULTSBaclofen-Induced Outward Currents. In cells impaled with

KCl-filled microelectrodes and voltage-clamped at or neartheir resting potential (-55 to -65 mV), (+)-baclofen (10pAM) always produced an outward (hyperpolarizing) mem-brane current (Fig. 1). During this current the amplitudes ofthe current deflections produced by superimposed short (1 s)voltage commands were greatly increased, showing that bac-lofen approximately doubled the cell's input conductance atthe holding potential. This effect was readily reversible onwashing out the baclofen and could be induced several timeson the same cell at intervals of 10 min or so without loss ofsensitivity. The action of (±)-baclofen was imitated by 10,uM (-)-baclofen but not by 10 ,uM (+)-baclofen.Outward Current Reversal Potential. "Steady-state" cur-

Abbreviation: GABA, -yaminobutyric acid.

1558

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. NatL Acad Sci. USA 82 (1985) 1559

10 4M baclofen

nIAW11~~I

I

1 min 1 s

FIG. 1. Effect of 10 ,uM baclofen on membrane currents in a voltage-clamped hippocampal CA3 pyramidal neurone. The cell was clamped atits initial resting potential (-61 mV) and subjected to 1-s voltage commands. The figure illustrates superimposed current responses (upperrecords) to three such commands, to -41, -81, and -101 mV (lower records). Currents evoked by the depolarizing command comprise (i) aninitial transient outward K+ current (IA; ref. 26), (it) an inward deflection (kCa; ref. 27), (iiM) a slow outward K+ current (IM + IC; refs. 28 and29), and (iv) an inward deactivating tail on repolarizing. Current responses to large hyperpolarizing commands show the inward IQ activation(28). During application of baclofen, 1-s -20-mV commands were applied every 5 s, while the recorder was slowed by a factor of 100.

rent-voltage (I/V) curves were constructed by measuringthe currents attained at the end of the 1-s commands andadding these to the holding current. The net current inducedby baclofen (Ibac) was calculated by subtracting the I/Vcurves obtained before and after adding baclofen from thecurve obtained during baclofen perfusion. In the cell illus-trated in Fig. 1, Ibac reversed from outward to inward at -82mV-i.e., 21 mV hyperpolarized to the resting potential of-61 mV (Fig. 2). (This is also apparent from the record inFig. 1, since the current level attained during the 20-mV hy-perpolarizing commands was the same in the absence andpresence of baclofen.) The mean (±SD) reversal potentialfor Ibac was -74 ± 4.3 mV in 11 cells.IbC Is a K+ Current. When [K+],,t was raised 3-fold, from

5.8 to 17.4 mM, the reversal potential shifted +26 mV, to-56 mV (Fig. 2); this is close to the +29-mV shift predictedfrom the Nernst equation for K+, suggesting that Ibac is sub-stantially or exclusively a K+ current.

Rectification of Ib,. The curves in Fig. 2 indicate that Ibacshows appreciable inward rectification-that is, the currentbecomes larger in the inward direction. A small time-depen-dent component of this inward current is visible in Fig. 1, butinward rectification was still apparent when currents weremeasured 20 ms after the imposed voltage step, immediatelyafter the capacity transients had settled.

External Ba2' Blocks Ib,. Some inwardly rectifying K+currents are blocked by external Ba2+ (30, 31). Fig. 3 showsthat Ibac is also effectively blocked by Ba2+. Ba2+ ions alsoblock the voltage-dependent K+ current IM in hippocampalneurones (28, 32). However, Ibac differs from IM in that Ibacbecame larger with hyperpolarization instead of smaller andwas not inhibited by 10 A.M acetylcholine, which suffices toinhibit IM in these cultured hippocampal cells (unpublishedobservations; cf. ref. 28).Ba2+ ions also inhibit the Ca2+-activated K+ current lc

previously described in hippocampal neurones (29). Unlike1c, however, Ibac was not inhibited by 100 ,M Cd2+ (see alsoref. 22), showing that Ibac did not depend on a prior influx ofCa2+ through Cd2+-sensitive Ca2+ channels.GABA Can Imitate Baclofen. Fig. 4 shows a comparison of

the effects of baclofen and GABA on a voltage-clamped hip-

pocampal neurone. In contrast to baclofen, application ofGABA (100 jM) at a clamp-holding potential of -62 mVproduced an inward current. This may be ascribed to a GA-BAA-mediated increase in Cl- conductance, the inward cur-rent resulting from the high intracellular Cl- concentrationattained with KCl electrodes (33-35). The mean (+SEM) re-versal potential for IGABA in six experiments was -46.7 ±4.2 mV-i.e., on average 26.0 ± 5.0 mV positive to the re-versal potential for Ibac in the same cells (Table 1). After 10

+0.5 -

lbac 'nA -0.5 -

-1.0-

-1.5

I .

)1P(5.8 mM K+

T/v ,/' 17.4 mM K+

U,

-120 -100 -80V, mV

-60 -40

FIG. 2. Voltage dependence of the baclofen-induced current(Ibac) recorded in 5.8 and 17.4 mM K+ solution (same cell as thatillustrated in Fig. 1). A series of 10-mV stepped voltage commandslike those shown in Fig. 1 to membrane potentials (V) between -101and -41 mV were delivered before, during, and after superfusionwith 10 ,uM baclofen, and the near-steady-state current levels at-tained at the end of the commands were measured with respect tothe initial holding current. Currents attained in the absence of baclo-fen were then subtracted from those in the presence of baclofen, toobtain the amount of extra current (+, outward; -, inward) inducedby baclofen (Ibac). Baclofen was applied twice in 5.8 mM K+ (0, A),then in 17.4 mM K+ (m), and finally in 5.8 mM K+ again (v).

Neurobiology: Gdhwiler and Brown

I I1

:,;T iiiiiiiii ..:TI.!,.1

!ill: .11:1.:.!ii .. ..

6-am

1560 Neurobiology: Gahwiler and Brown

10 ,uM baclofen

II

IIIii!r

1

nA

1 min

Table 1. Reversal potentials for the ionic currents induced byGABA and baclofen before and after addition of pitrazepin

Reversal potential, mVHoldingpotential, Control Pitrazepin

Cell mV GABA Baclofen GABA Baclofen

1 -61 -53 -76 -74 -762 -50 -46 -71 -71 -703 -46 -57 -73 -73 -764 -52 -52 -66 -66 -625 -61 -28 -76 -73 -716 -62 -44 -74 -73 -75

Mean -46.7 -72.7 -71.7 -71.7SEM ±4.2 ±1.5 ±1.2 ±2.2

Reversal potentials were determined from I/V curves as shown inFig. 2. Concentrations in the perfusion fluid were 100 ,uM GABA, 101AM baclofen, and 10 1LM pitrazepin.

AControl + Pitrazepin

BTFlofel FFI

d;l !i 1liisilil1iljWa

FIG. 3. Effect of 1 mM BaCl2 on the outward current generatedby 10 ,uM baclofen. Holding potential was -61 mV throughout.Thick downward deflections show currents induced by -20 mV hy-perpolarizing voltage commands. Gaps in records refer to intervalswhere variable voltage commands were applied in the presence ofbaclofen (not illustrated).

,uM pitrazepin, a GABAA antagonist (23), had been added tothe perfusion fluid, GABA produced an outward current,with a reversal potential of -71.7 ± 1.2 mV-identical withthat for Ibac (Table 1). Pitrazepin did not affect either themagnitude of Ibac or its reversal potential. In the presence ofpitrazepin, IGABA also showed the inward rectification previ-ously noted for Ibac; before pitrazepin IGABA did not showrectification. Like Ibac, IGABA in the presence of pitrazepinwas blocked by 1 mM Ba2+. These experiments suggest thatGABA and baclofen activate the same K+ conductance inhippocampal cells under conditions in which the increasedC1- conductance produced by GABA is suppressed.

Baclofen Does Not Inhibit Ca2+ Current. In Fig. 1 an in-ward current deflection may be noted after the initial tran-sient outward current during the depolarizing command.This is generated by the slow inward Ca2+ current Ice previ-ously described in hippocampal neurones (27, 36). ICa itselfis largely masked by subsequent outward K+ currents duringthe depolarizing step, but its continued presence throughoutthe command is suggested by the inward tail current whenthe holding potential is restored. This inward tail may reflectboth persistence ofICa (27) and a residual Ca2+-activated Cl-current (43). In baclofen solution, ICa is less apparent duringthe depolarizing command, but the repolarizing tail currentis still present; this suggests that the reduction in ICa may besecondary to the increased outward current. To test the ef-fect of baclofen on ICa itself, outward currents were inhibitedby injecting Cs+ into the neurone by using a CsCl-filled mi-croelectrode (36). This also inhibited the outward currentgenerated by baclofen. Under these conditions depolarizing

GABA > m 771|

0.5 nA1 min

B

I,nA

IIGABA

V, mV

FIG. 4. GABA imitates baclofen when GABAA-mediated Cl-conductance is suppressed. The records inA show currents inducedby 10 AM baclofen and 100 ,uM GABA applied before and after add-ing 10 u.M pitrazepin (23) to the perfusion fluid. Holding potentialwas -62 mV. Downward deflections are currents induced by 1-s 20-mV hyperpolarizing commands at 15-s intervals (cf. Fig. 1). Out-ward current upwards. Note that pitrazepin reversed the GABA-induced current from inward to outward but did not affect the baclo-fen-induced current. The graphs in B show the voltage dependenceof GABA- and baclofen-induced currents recorded in the absence(o, o) and presence (v, *) of pitrazepin. Graphs were obtained fromsubtracted I/V curves as described for Fig. 2.

Control

+ Ba2'

Proc. Nad Acad Sci. USA 82 (1985)

Proc. NatL Acad Sci USA 82 (1985) 1561

Control + Baclofen Wash

1 nA

7_... 120 mV

1is

FIG. 5. Lack of effect of baclofen (10 ,uM) on inward currentsrecorded in a cell impaled with a 2 M CsCl-filled microelectrodeduring 1-s step depolarization of +20 mV from a holding potential of40 mV. The inward currents show fast transient and slow sustainedresponses (cf. ref. 37), the former being incompletely clamped. Thelower records show corresponding "leak" currents induced by 20-mV hyperpolarizing commands.

commands generated a net inward Ca2+ current as shown inFig. 5. Two components of this current can be discerned, arapid transient current and a more sustained component, aspreviously reported in adult rat hippocampal cells (27, 37).These currents were inhibited in Ca2 -free solution or by ad-dition of 10-100 uM Cd2 . In contrast, baclofen did not in-hibit either component.

DISCUSSIONThe principal effect of baclofen on hippocampal neuronesdetected in these experiments is the generation of an out-ward membrane current Ibac. This we attribute to an in-creased K+ conductance for three reasons. (i) The reversalpotential for Ibac shows the expected dependence on externalK+ concentration; (it) Ibac is inhibited by external Ba2+; and(iii) it is also inhibited by internal Cs+. This current is pre-sumably responsible for the hyperpolarization of hippocam-pal cells previously reported (21-23). The inward rectifica-tion would explain why this hyperpolarization appeared todiminish when the cells were depolarized (22). A comparablecurrent probably underlies the hyperpolarizing effect of ba-clofen on Onchidium neurones (38).

This effect of baclofen probably results from an action onGABAB receptors since it could be replicated by GABA whenthe GABAA-mediated increase in Cl- conductance wasblocked with pitrazepin. Newberry and Nicoll (39) have alsodetected a shift in the reversal potential for the effect ofGABA on hippocampal CA1 neurones to more hyperpolar

ized levels in the presence of the GABAA antagonist bicucul-line. In the absence of a selective GABAB antagonist we can-not be certain that this reflects an effect ofGABA and baclo-fen on the same receptor, as distinct from independentactivation of the same conductance mechanism via differentreceptors, although ligand-binding studies (2) support a com-mon receptor interaction.At rest potentials between -50 and -70 mV the increased

K+ conductance produced by GABA via GABAB receptorswas much less than the increased Cl- conductance associat-ed with GABAA-receptor activation, so that in the absenceof GABAA antagonists the latter dominates the overall re-sponse to exogenous GABA. Thus, inhibition of the K+component with Ba2' did not noticeably change the reversalpotential for GABA in the absence of pitrazepin. The in-creased K+ conductance becomes greater at membrane po-tentials beyond -80 mV, but it is difficult to imagine its sig-nificance during simultaneous activation of GABAA recep-tors. Activation of this K+ current might be of more im-portance to the effects of synaptically released GABA if, assuggested by Newberry and Nicoll (39), there is a temporaldissociation between the activation of GABAA and GABABreceptors.We could detect no effect of baclofen on ICa in hippocam-

pal neurones under conditions in which the K+ conductancewas blocked. We therefore suggest that the inhibition oftransmitter release by baclofen postulated to occur in the ter-minals of CA3 cell axons (13-15) is more likely to result fromthe indirect effects of the increased K+ conductance on Ca2+influx than from a direct effect on Ca2+ conductance per se.This may also be true for the effect of baclofen on mammali-an primary afferent fibers, since baclofen failed to reducethe "Ca2+ component" of the action potential in sensory spi-nal ganglion cells impaled with Cs'-filled microelectrodes(20) in which lbac would be suppressed.

In subsequent experiments we have observed that the in-ward currents produced by baclofen and GABA (in the pres-ence of pitrazepin) at membrane potentials negative to theirreversal potential were also blocked in a voltage-dependentmanner by 1 mM external Cs' (unpublished observations).Since other inwardly rectifying K+ currents show voltage-dependent block by Cs' (40-42) this supports our view thatboth compounds activate the same species of inward-rectify-ing K+ conductance.

This work was partly supported by a grant from the U.K. MedicalResearch Council to D.A.B.

1. Bowery, N. G., Hill, D. R., Hudson, A. L., Doble, A., Mid-dlemiss, D. N., Shaw, J. & Turnbull, M. (1980) Nature (Lon-don) 283, 92-94.

2. Hill, D. R. & Bowery, N. G. (1981) Nature (London) 290, 149-152.

3. Krnjevic, K. (1974) Physiol. Rev. 54, 418-540.4. Nistri, A. & Constanti, A. (1979) Prog. Neurobiol. 13, 117-

235.5. Davidoff, R. A. & Sears, E. S. (1974) Neurology 24, 957-963.6. Pierau, F.-K., Mattheson, G. K. & Wurster, R. D. (1975) Exp.

Neurol. 48, 343-351.7. Nistri, A. (1975) Experientia 31, 1066-1068.8. Fukuda, H., Kudo, Y. & Ono, H. (1977) Eur. J. Pharmacol.

44, 17-24.9. Fox, S., Kmjevic, K., Morris, M. E., Puil, E. & Werman, R.

(1978) Neuroscience 3, 495-515.10. Ono, H., Fukuda, H. & Kudo, Y. (1979) Neuropharmacology

18, 647-653.11. Capek, R. & Esplin, B. (1981) Can. J. Physiol. Pharmacol. 60,

160-166.12. Johnston, G. A. R., Hailstone, M. H. & Freeman, C. G.

(1980) J. Pharm. Pharmacol. 32, 230-231.13. Curtis, D. R., Lodge, D., Bornstein, J. C. & Peet, M. J. (1981)

Exp. Brain Res. 42, 158-170.14. Henrv. J. L. (1982) Neuropharmacology 21, 1085-1093.

Neurobiology: Gdhwiler and Brown

r -L

1562 Neurobiology: Gahwiler and Brown

15. Lanthorn, T. H. & Cotman, C. W. (1981) Brain Res. 225, 171-178.

16. Olpe, H.-R., Baudry, M., Fagni, L. & Lynch, G. (1982) J.Neurosci. 2, 698-703.

17. Ault, B. & Nadler, J. V. (1982) J. Pharmacol. Exp. Ther. 223,291-297.

18. Ault, B. & Nadler, J. V. (1983) Br. J. Pharmacol. 80, 211-219.19. Dunlap, K. (1981) Br. J. Pharmacol. 74, 579-585.20. Desarmenien, M., Feltz, P., Occhipinti, G., Santangelo, F. &

Schlichter, R. (1984) Br. J. Pharmacol. 81, 327-334.21. Misgeld, V., Klee, M. R. & Zeise, M. L. (1982) in Physiology

and Pharmacology of Epileptogenic Phenomena, eds. Klee,M. R., Lux, H. D. & Speckmann, E.-J. (Raven, New York),pp. 131-139.

22. Newberry, N. R. & Nicoll, R. A. (1984) Nature (London) 308,450-452.

23. Gahwiler, B. H., Maurer, R. & Wuethrich, H. J. (1984) Neur-osci. Lett. 45, 311-316.

24. Gahwiler, B. H. (1981) J. Neurosci. Methods 4, 329-342.25. Gihwiler, B. H. (1984) Experientia 40, 235-243.26. Gustafsson, B., Galvan, M., Grafte, P. & Wigstrom, H. (1982)

Nature (London) 299, 252-254.27. Brown, D. A. & Griffith, W. H. (1983) J. Physiol. (London)

337, 303-320.28. Halliwell, J. V. & Adams, P. R. (1982) Brain Res. 250, 71-92.29. Brown, D. A. & Griffith, W. H. (1983) J. Physiol. (London)

337, 287-301.

30. Hagiwara, S., Miyazaki, S., Moody, W. & Patlak, J. (1978) J.Physiol. (London) 279, 167-185.

31. Standen, N. B. & Stanfield, P. R. (1978) J. Physiol. (London)280, 169-191.

32. Constanti, A., Adams, P. R. & Brown, D. A. (1981) Brain Res.206, 244-250.

33. Alger, B. E. & Nicoll, R. A. (1979) Nature (London) 281, 315-317.

34. Anderson, P., Dingledine, R., Gjerstadt, L., Langmoen, I. A.& Laursen, A. M. (1980) J. Physiol. (London) 305, 279-2%.

35. Ben-Ari, Y., Krnjevic, K., Reiffenstein, R. J. & Reinhardt, W.(1981) Neuroscience 6, 2445-2463.

36. Johnston, D., Hablitz, J. J. & Wilson, W. A. (1980) Nature(London) 286, 391-393.

37. Halliwell, J. V. (1983) J. Physiol. (London) 341, 1OP-l1P.38. Shimuzu, N., Akaike, N., Oomura, Y., Maruhashi, J. & Klee,

M. (1983) Jpn. J. Physiol. 33, 459-467.39. Newberry, N. R. & Nicoll, R. A. (1984) J. Physiol. (London)

353, 99P.40. Hagiwara, S., Miyazaki, S. & Rosenthal, N. P. (1976) J. Gen.

Physiol. 67, 621-638.41. Gay, L. A. & Stanfield, P. R. (1977) Nature (London) 267,

169-170.42. Constanti, A. & Galvan, M. (1983) J. Physiol. (London) 335,

153-178.43. Owen, D. G., Segal, M. & Barker, J. L. (1984) Nature (Lon-

don) 311, 567-570.

Proc. NatL Acad Sci. USA 82 (1985)