Effects of anticonvulsants on excitability and GABA sensitivity of cat dorsal root ganglion cells

~28-3~8/81/0~~2?.07SO2.~/0 Pergamon Press Lid

Neuro~~rrnur~f~~~ Vol. 20. pp. 427 to 433. 1981 Printed in Great Britain

EFFECTS OF A~ICO~ULSA~S ON EXCITABILITY AND GABA SENSITIVITY OF CAT DORSAL ROOT GANGLION CELLS J. P. GALLAGHER,H. INOKUCHI*,J. NAKAMLJRA~ and PATRICIASHINNICK-GALLAGHER Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, TX 77550, U.S.A. (Accepred 4 December

1980)

the mammalian central nervous system, GABA is clearly an important inhibitory transmitter. As such, it has been implicated in the etiology of epilepsy. In this investigation an examination was made as to whether three clinically effective anticonvulsants, phenobarbi~l (PhB), phenytoin (DPH) and valproic acid (V) possessed a common action, namely whether they modified the interaction between GABA and the GABA receptor. The results demonstrated that PhB, at clinically effective concentrations, facilitated a GABA-induced depolarization recorded from cat dorsal root ganglion cells; i.e. enhanced its amplitude and prolonged its time course. In addition, PhB depressed direct action potentials. These effects occurred at concentrations below which PhB depolarized the membrane directly. On the other hand, neither DPH nor valproate facilitated a GABA depolarization. However, at clinically effective concentrations, DPH depressed direct action potentials; valproate did not decrease the excitability of these cells as did both PHB and DPH. These results suggest that the clinical anticonvulsant actions of PhB, DPH and valproate may occur through different mechanisms.

Summary-h

A maifunction in the GABAergic system has been implicated in the etiology of epilepsy (Meldrum, 1975; Tower, 1976). Three chemically dissimilar anticonvulsant drugs e.g. phenobarbital (PhB), phenytoin (DPH) and valproic acid (V) appear to interact with the GABA system. The ability of barbiturates to facilitate GABAmediated synaptic inhibition in mammals has been well documented (Nicoll, 1972; Ransom and Barker, 1975; Nicoll, Eccles, Oshima and Rubia, 1975; Nicoll, 1975; Ransom and Barker, 1976; Barker and Ransom, 1978). Although most of these studies deal primarily with the anesthetic barbiturate, pentobarbital, in some cases (Macdonald and Barker, 1978; Nicoll and Wojtowicz, 1980) a comparison has been made between phenobarbital (PhB) and other barbiturates. High concentrations of barbiturates have been suggested to act as GABAmimetics, at the receptor recognition site (Nicoll and Wojtowicz, 1980). On the other hand, there are reports that the depolarizing action obtained with the higher concentrations of barbiturates does not involve the GABA receptor recognition site but somehow affects the ionic channels associated with the GABA receptor complex or nonspecific neuroronal membrane sites (Barker and l Present address: Department of Physiology, Kurume University, School of Medicine, Kurume City, Japan 830. t Present address: Department of Nebropiychiatry, Kurume Universitv. School of Medicine. Kurume Citv. ,. Japan 830. _ Key words: GABA, phe,nobarbital, phenytoin, valproic acid, dorsal root ganglion, chloride channels.

N.P. 20/s--s

427

McBurney, 1979; Higashi and Nishi, 1979; Mathers and Barker, 1980). Another anticonvulsant, phenytoin (DPH), has been reported to have little effect on GABA receptors on frog motoneurons (Nicoll and Wojtowicz, 1980) or rat dorsal root ganglia (Connors, 1979). However, there are reports (Adams, Banks and Constanti, 1981; Ayala and Johnston, 1977; Deisz and Lux, 1977) that DPH enhances the GABA-mediated inhibitory postsynaptic potentials recorded from crayfish stretch receptors. The anticonvulsant valproic acid (V) was proposed to increase GABA brain levels by competitive inhibition of GABA transamina~, (Godin, Heiner, Mark and Mandel, 1969). On the other hand, others (Harvey, Bradford and Davison, 1975; Anzelark, Horton, Meldrum and Sawaya, 1975; Emson, 1976) showed that valproic acid inhibited mainly succinic acid seminaldehyde dehydrogenase. More recently, Frere, Young and Ma~onald (1980) demonstrate that valproic acid did not inhibit GABA transaminase or succinic acid semialdehyde dehydrogenase in primary dissociated cell cultures of mouse forebrain. In addition to possible metabolic effects, Macdonald and Bergey (1979) reported that valproic acid augmented postsynaptic GABA-mediated inhibition directly in cultured mouse spinal cord neurons. But, Geller (1980) has shown in cultures of rat hypothalamus neurons that valproic acid does not act acutely to potentiate GABA-mediated inhibition. It is apparent that the data currently available regarding the mechanisms by which these anti-

428

J. P.

GALLAGHER et al

convulsant drugs may affect GABAergic systems is conflictmg. Studies m this laboratory (Gallagher, Higashi and Nishi. 1978) have demonstrated that an irk citro cat dorsal root ganglion (DRG) preparation is a useful model system to investigate the interaction between drugs and a mammalian GABA receptor. The DRG is umque in that, unlike most III civo and in vitro preparations, it is devoid of synapses, and is not complicated by multiple sites of drug action other than the receptor and its associated ionic complex The purpose of this study with the isolated cat DRG was to determine whether anticonvulsant drugs (PhB. DPH and V) are acting: (1) on the GABAreceptor complex directly: (2) at known concentrations that may be related to effective anticonvulsant plasma levels: or (3) at other parts of the neuronal membrane or by mechanisms not involved with the GABA-receptor complex. METHODS

Dorsal root ganglia (L6-Sl) were removed from adult cats (3-5 kg) of either sex. Cats were anesthetlzed with a mixture of g-chloralose (60mg/kg. i.p.) and pentobarbitol (10 mg/kg, i.p.). After surgical anesthesia was achieved, a dorsal laminectomy was performed and the DRG dissected from their attachments to the spinal cord and peripheral nerves. The isolated DRG was maintained in a recording chamber by superfusion with a modified Krebs solution (Gallagher et al.. 1978) at 37°C and was gassed with 95”” 02-5”/, CO,. Individual DRG cells were irrnaled with either one or two microelectrodes (2( 0 MR impedance) filled with 2 M K citrate. In ad! ion, an extracellular electrode, filled with 0.5 M GABA (pH = 4.5) was used to iontophorese GABA onto the individual cells. A WPI electrometer (Model M-707) was employed for recording intracellular membrane responses. A second electrode was inserted into a cell to determine the GABA reversal potential (EGABA).With the second electrode a known constant current could be applied across the cell membrane to shift the membrane potential sufficiently in the depolarizing direction. Concentration-response curves were determined for GABA by applying various current intensities of equal duration to the iontophoretic GABA electrode. The amplitudes of the resultant GABA depolarizations were normalized so that the maximum response. represented by the plateau of the concentration response curve, was equal to I. The fact that the GABA responses plateaued with a maximum depolarization suggested that all the GABA receptors on the particular neuron were saturated. If a drug acted on the GABA-receptor complex directly, the amplitude and/‘or time course of the lontophoreticallyinduced GABA depolarization (GD) would be altered. If a drug affected a direct action potential (AP), generated by passing a cathodal current through the

recording electrode, the drug may have a non-specific action on the activated neuronal membrane, i.e. the ordinary, non-receptor area. The effect on the action potential was assessed by using an RC circuit to monitor the rate of rise (dV/dt) of the action potential, Finally, effects on the resting properties of the membrane (i.e. input resistance and resting potential) were contmuously monitored. Input resistance was measured by recording the amplitude of anelectrotonic potentials generated from repetitive hyperpolarizing current pulses applied to the cell through the recordmg electrode. The 3 drugs, phenobarbital (Sigma), 5.5-diphenylhydantoin Na (Sigma) and valprolc acid (Abbott Labs) were added to the superfuslon solution to yield their final bath concentrations. RESL’LTS

Effects oj‘phenobarbital

Phenobarbital (PhB) in concentrations from 1 x 10m5 M to 5 x 10e3 M was applied by superfusion to the DRG while iontophoretically-induced GDs and/or direct APs were recorded. Phenobarbital had multiple actions that were concentration-dependent (Table I ). No consistent effects were observed at 1 x 10e5 M PhB, although in 1 out of lOcells, the GD was enhanced in amplitude by lo?,,. At the lowest concentratlon that produced a significant effect, 5 x 10m5 M, PhB prolonged the t-decay time and enhanced the amplitude of the GD. Occasionally, PhB enhanced the amplitude of the GD without producing a significant prolongation of the f-decay time. However, the converse, 1.e. a prolonged f-decay without an increased amplitude. was not observed. At 5 x lo- ’ M PhB. neither the active nor passive properties of the neuronal membrane were changed. When the concentration was doubled to I x 10m4 M. in addition to the actions mentioned above (Fig. 1. top), PhB depressed the active membrane by slowing the rate of rise (dV/dt) of direct APs. The latter, general depressant action became more evident as the concentration was raised tenfold to 1 x 10d3 M. At 10e3 M and higher, direct APs were blocked (Fig. 1, bottom). At the 10e3 M concentration PhB had 3 different effects: (I) PhB enhanced the amplitude and prolonged the time course of the GD: (2) PhB blocked direct action potentials; and (3) PhB depolarized the membrane directly. The membrane depolarization was due to an increase in membrane conductance. Finally, at 5 x 10m3 M, the highest concentration examined, PhB depressed the GD without further changing its timecourse, blocked spikes and depolarized the membrane. The present authors believe the depression of the GABA response by 5 x 1O-3 M PhB is due to general membrane depressant actlon of PhB (see Discussion). It was interesting to examine the mechanism by which PhB facilitated the GD. For this reason, a concentration range of PhB (1 x 10-‘~5 x 10e4 M) was chosen as this produced the greatest facilitatory

1 = depressed;

T = Increased;

+

= prolonged;

19.5 & 5.0

212

5 x 1o-3 NC

NC = no change;

212

+45 -53

5116 216

7 33 8.3 1 19 + * 5.9

potential;

Depol.

= depolarized.

1

212

212

NC 1

16116 516

NC Depol. 3.6 + 1.7 mV Depol. 5.0 + 0 mV

16/16 516

NC

NC NC

A

717

Reststance

NC

lljll 13/13

properties Input lnctdence

717

membrane

NC NC

lljll 13113

RMP = resting

f 4.0 7.0

--+24 & 6.5

-+22 k 16 422 + 2.5

417

2111 3113

1 24 + 4.6

16.8 + 2.5 7 17 + 4.0

7116 516

717

2111 4113

membrane A

Passive RMP

effects of phenobarbital

Incidence

1. Concentration-dependent

GABA depolarization :-Decay Amplitude Inctdence ‘(,A + SD Incidence “/,A k SD

10-4 51 x 1o-3

2 x 1o-4

5 x 1o-5 1 x 1o-4

Cont. (Ml

Table

315 8/8

313 215

313 314 l/4 212

i Blocked Blocked

i

NC NC

action potential Amplitude Incidence A

Direct

J. P. GALLAGHER et al.

430 Control

PhB (lOQM) IlnA

PhB (109M)

Control

DIRECT Action Potential

Ftg. I. Acttons of phenobarbttal on GABA depolarizations and action potenttals. Top tracmgs: PhB (1 x IO-“ M) decreased the amphtude and prolonged the decay phase of an tontophoretically-induced GABA depolarization. At this concentratron there was httle effect upon the direct action potenttal. Bottom tracmgs: PhB (1 x 10m3M) depressed the rate of rtse and amplitude of direct action potenttals The membrane potential was also depolarized from the resting level (dashed line). effects on the GABA response,

while having no effect

upon passive membrane properties. Phenobarbital (1 x 10e4 M) shifted the concentration response curve for GABA in a parallel manner to the left and also increased the maximum response (Fig. 2). The GABA concentration response curve could be shifted in this manner if PhB enhanced the driving force by decreasing the GABA equilibrium potential (EGABA). TO examine this possibility, EGAaA was determined before, during, and after the addition of PhB (Fig. 3B). Although PhB did not alter ECARA(Fig. 3B). it is clear that the GD was enhanced and prolonged (Fig. 3A). In 5 paired experiments. the control EGABA was -20.2 k 6.9 mV and, after PhB (5 x 10e4 M). EoI\eA was -20.1 f 6.2 mV.

Phenytoin (DPH). in concentrations of 1 x IO- ’ M2 x 10m4M. was applied by superfusion and GDs

.E 5

10

and direct APs were recorded in 34cells. Phenytoin did not facilitate the GD or prolong its time-course. In fact, DPH, at concentrations above 1 x 10e4M, depressed the amplitude of the GD (Fig. 4, top). In addition, DPH, at 5 x 10m5M, blocked direct action potentials (Fig. 4, bottom). The depressant actions of DPH on the GDs or APs were not associated with changes in the passive membrane properties. Efjpcts of valproic acid

Valproic acid (V) applied to 19 cells in concentrations from 1 x 10e4 to 1 x 10m3M did not affect significantly any of the parameters that were measured, namely: GD, passive membrane properties or APs. However, in 4 out of 13cells. at concentrations of 5 x 10m4M and greater, valproic acid reversibly depressed GDs by 5-15%; facilitation was not observed. At the higher concentrations no effects were observed on the passive membrane properties of cells or directly-induced APs.

0 Control l

PhB (10-4M 1

Log GABA

Ejection

Current

(x10-*A)

Fig 2. Effect of PhB on GABA concentration response curve; PhB (1 x 10m4M) was bath-applied. Circles represent experimental data from 1 cell; the solid lines are drawn by eye to fit the experimental data. All data have been normalized so that the control maximum GABA response ts equal to 1.0.

Anticonvulsants and GABA

(A)

’ -20

Control

PhB( 5~10-~M

l-

1

,- -

---z:

,--e

1 Set

0

Control

l

PhB

0 Rinse -6 -6

3<

-4 -2 -0

A 0

mV

// L4 Fig. 3. Effect of PhB on the GABA reversal potential. (A) Typical iontophoretically-induced GABA depolarizations recorded at their respective membrane potentials before (control) and during PhB (5 x 1O-4 M). (B) Plot of the GABA reversal potential from data in (A).

DISCUSSION

Phenobarbital, but not phenytoin or valproic acid, facilitated iontophoretically-induced GABA depolarizations. It was possible, on the basis of concentration, to separate a facilitatory action of PhB on the GABA response (i.e. enhanced amplitude and prolonged half decay time) from a depressant action on the direct action potential. There is a great deal of

Control

GABA Depolarization

431

evidence to support a facilitatory interaction between barbiturates and GABA. However, neither the precise mechanism nor the concentration of barbiturate responsible for this facilitation is known. At a GABA synapse, bartiturates could alter GABA synthesis, release, re-uptake or degradation. At the DRG, none of these possit$e mechanisms are operative (Desarmenien, Feltz and Headley, 1979). Alternatively, PhB could increase the input resistance of a cell and thereby enhance the response to each concentration of GABA. However, PhB did not increase resistance and, in fact, decreased the resistance at the higher concentrations (Table 1). Also, PhB could facilitate the GD by altering the driving force (ERMP- E,,,,) of the GD by either hyperpolarizing the membrane or shifting EGAeAto a more depolarized level; PhB did not hyperpolarize the DRG (Table 1) nor did PhB shift EGABA(Fig. 3). Barker and McBurney (1979) reported that when the GABAinduced current was enhanced in mouse spmal cord neurons grown in tissue culture, ECABA was not altered by PhB; however, the concentration of PhB employed in those experiments was not known because PhB was applied by iontophoresis. Another possible facilitatory mechanism could be that PhB was acting in low concentrations at the GABA receptor site to alter the affinity and/or efficacy of GABA without significantly changing the resting properties of the DRG. This mechanism would operate only when GABA was applied and would result in the shift of the GD concentration response curve to the left and upwards (Fig. 2). If this mechanism was effective then the apparent dissociation constant and/or the Hill’s slope of the GABA concentration response curve should have been altered. However, Higashi, Inokuchi and Nishi (1980) using voltage clamp techniques, demonstrated that pentobarbital (lo- ‘10e3 M) augmented the GABA-current of the cat DRG in a concentration-dependent manner without

OPH(2~10‘~hn)

, 1“A

-*v-r-,

1OmV ‘DPH

1

75nA

Control

DIRECT Actlon Potential

Fig. 4. Effect of phenytoin on GABA depolarization and action potentials. Top tracmg: DPH (2 x 10m4 M) slightly depressed the GABA depolarization wlthout altering Its decay phase of affectmg the membrane potential. Bottom tracing: DPH (5 x IO-’ M) depressed the rate of rise and amplitude of direct action

potentials.

The membrane

potential

was not altered.

432

J.P.

GALLAGHER

affecting the dissoctation constant or the Hill’s slope of the GABA concentration response curve. The present results with PhB (Fig. 2) are in agreement with those of Higashi ef al. (1980). Finally, PhB may be affecting the chloride channel coupled to the GABA receptor; PhB could enhance the single channel conductance and/or prolong the life-time of a single channel. The present experiments could not distinguish between either of these possibilities, but the fact that the falling or decay phase of the GD was prolonged suggested that PhB may have prolonged the life-time of the channels. The fact that no observation was made of a prolongation of the decay phase in all experiments in which the amplitude of the GD was increased may be due to the voltage-dependence of the GD (Gallagher et al., unpublished observations, see also Onodera and Takeuchi, 1979). Barker and McBurney (1979) using voltage clamp techniques, concluded that PhB did not increase the single channel conductance but rather prolonged the life-time of a single channel activated by GABA. It has previously been demonstrated (Fig. 9 of Gallagher et al., 1978) that when bromide anion was substituted for chloride, a facilitatory action (enhanced amplitude and prolonged half decay time) was observed similar to that obtained with PhB (Fig. 1 top, Fig. 3A). All ions in a series of 17 anions prolonged the half decay of the GD. However, bromide was the only anion that produced a facilitation; all the rest depressed the amplitude of the GD. This facilitatory action may be related to the clinical effectiveness of bromide as an anticonvulsant. The mechanism for this facilitatory action for bromide may be similar to that of PhB, namely, an action on the anionic (chloride) channels. If the concentration of PhB was equal to or less than 5 x 10m4M, PhB did not directly depolarize the soma membrane or depress the GD. At high concentrations of 1 x 10m3M or greater, PhB depolarized the DRG, as well as depressed GD’s and blocked direct action potentials. Higashi and Nishi (1979) have reported that the depolarizing action observed with concentrations of pentobarbital greater than 10e4 M is not a GABAmimetic action (i.e. a specific increase in chloride conductance resulting from activation of the GABA receptor), but is rather a nonspecific increase in membrane conductance (but see Nicoll and Wojtowicz, 1980). Mathers and Barker (1980) have concluded that pentobarbital may open ion channels of similar conductance and 5 times the lifetime of those activated by GABA. It is important to note, as Mathers and Barker have suggested, that these ionic channels activated by pentobarbital (or perhaps the higher concentrations of PhB) may not be associated with the GABA-receptor complex, but are associated with some other part of the neuronal membrane. This non-specific action of PhB at these high concentrations is probably not related to its clinical effectiveness as an anticonvulsant. The effective plasma level for PhB as an anticonvul-

et al

sant is 2Opg/ml (Woodbury and Fmgl, 1975) and this level roughly corresponds to the concentration (1 x 10m4M) of PhB that facilitated the GABA response but had minimal effects upon the direct action potential. A direct depressant effect of PhB was observed at twice the concentration that produced GABA-facilitating effects. which roughly corresponded to the plasma concentration that manifests untoward effects in patients. In comparison, phenytoin, did not facilitate the GABA response (Fig. 4). At concentrations (1 x 10m4M or greater) which blocked direct action potentials, DPH also depressed GABA depolarizations. The effective plasma level for DPH (lOpg/ml) corresponds to about 5 x 1O-5 M which was the concentration that began to block direct action potentials. At twice this concentration (1 x 10e4 M), DPH. in addition to blocking action potentials, also depressed GABA depolarizations. Finally, valproic acid did not share the actions of either PhB or DPH. It did not affect GABA depolarizations, or active or passive membrane properties. Based on the present results it appeared that if valproic acid was interacting with a GABA system, either the GABA synthetic, re-uptake, or metabolic processes,were involved but not the GABA-receptor complex. The results with DPH and valproic acid do not agree with data collected at the crayfish stretch receptor (Adams et al.. 1980: Ayala and Johnston, 1977; Deisz and Lux, 1977) and at mouse spinal cord neurons grown in tissue culture (Macdonald and Bergey, 1979) respectively (see Introduction). Possible explanations may be that there is a species difference in the action of these drugs at these GABA receptors or that the GABA receptors are different in various preparations. This latter possibility is considered in depth by Nistri and Constanti (1979). On the other hand, these anticonvulsant drugs may act on GABA release, re-uptake, synthesis, or metabolic mechanisms not active in the present system. For example, Hackman, Davidoff and Grayson, (1980) demonstrated that valproate is 99% effective in inhibiting re-uptake of GABA in the hemisected frog spinal cord preparation. This inhibition by valproate could explain both the absence of facilitation of the GABA depolarization at the cat dorsal root ganglion where re-uptake mechanisms may not be important (Desarmenien et al.. 1979) and the presence of facilitation on cultured mouse spinal cord neurons (Macdonald and Bergey. 1979) with an active GABA uptake system. In conclusion, PhB, at concentrations that are clinically relevant, may interact with the chloride channel associated with the GABA receptor, but not the GABA receptor itself, to facilitate a GABA depolarization Some of the anticonvulsant action of PhB may be due to this particular facilitatory action via enhancement of inhibition in the CNS. The anticonvulsant action of PhB could also be due to a direct depression of action potentials. A general decrease in

Anticonvulsants and GABA

433

membrane excitability observed with DPH may also contribute to its anticonvulsant action. On the other hand, since valproic acid did not appear to alter the GABA response under investigation or alter membrane excitability. its anticonvulsant effect may be due to actions on parameters not present in the present system, namely, release, re-uptake, or metabolism of GABA.

Godin, Y., Heiner, L., Mark, J. and Mandel, P (1969). Effects of di-n-propylacetate, an anti-convulsant compound, on GABA metabolism. J. Neurochem. 16: 869-873. Hackman, J. C., Davidoff. R. A. and Grayson, V. (1980). The actions of valproic acid on dorsal root fibers m the frog spinal cord. ioc. Neurosci. Absrr. 6: 793. Harvey, P. K. P., Bradford, H. F. and Davison. A. F. (1975). The Inhibitory effect of di-n-propylacetate, an anti-convulsant compound on GABA metabBhsm. iFEES

~~k~o~~e~~e~~~~fs-we thank Dr H. Higashi for a helpful review of the manuscript and Mrs Martha Myers for clerical assistance Supported by USPHS Grant NSl3727.

Higashi, H. and Nishi. S. (1979) The effect of barbiturates on the GABA receptor of cat primary afferent neurons. In: Integrative Control Functrons of the Brain (Ito, M . Kubota, K.. Tsukahara. N. and Yagl, K.. Eds), Vol. 2, pp. 18-19. Elsevier North-Holland Biomedical Press. Amsterdam. Higashi, H., Inokuchl, H. and Nislu, S. (1980). Barbiturates and GABA-induced currents of primary afferent neurons. J. Pension. Sot Japarr 42: 253 Macdonald. R. L. and Barker. J. L. (1978). Dtfferent actlons of anticonvulsant and anesthetic barbiturates revealed bv use of cultured mammalian neurons. Sclerzce 200: 775-777. Macdonald, R. L. and Bergey, G. K. (1979). Valprolc acid au ments GABA-mediated postsynaptic mhlbition m cuftured mammalian neurons. Brart~ Res. 170: 558-562. Mathers, D A. and Barker, J. L. (1980). (-)Pentobarbltal opens ion channels of long duratton m cultured mouse spinal neurons Scfer~e 209: 507-508. Meldrum. B. S. (1975). Epilepsy and y-ammobutyrlc acidmediated inhibition. Int. Rec. Neurobrol. 17: 1-36 Nicoll. R. A. (1972). The effects of anesthetlcs on synaptic exatatlon and inhibition m the olfactory bulb. J PhySJO!.223: 803-814. N~colf,R. A. (1975) Pentobarbital, action on frog motoneurons. &urn Rex 96: 119-123. Nicoll, R. A and WoJtowicz. J. M. (1980). The effects of pentobarbltal and related compounds on frog motoneurons. Bruiu Res. 191: 225--237 N~coll, R. A., Eccles. J. C., Oshima, T.. and Rubla. E (l975). Prolongation of hippocampal mhlbitory postsynaptic potentials by barbiturates. Nature 258: 625-627 Ntstrt, A. and Constant%. A. (1979) Pharmacologtcal characterlzatlon of dtfferent types of GABA and glutamate receptors in vertebrate and Invertebrates. Prog. Neurobrol. 13: 117-135. Onodera, K. and Takeucl-n. A. (1979). An analysis of the mhlbltory postsynaptic current in the voltage-clamped crayfish muscle. J. Pby~iol., Land. 2%: 265-282. Ransom. B. R. and Barker, .I. I_. (1975). Pentobarbltal modulates transmitter effects on mouse spinal neurones grown in tissue culture Nuture 254: 703- 705. Ransom, B R. and Barker, J L. (1976). Pentobarbltal selectively enhances GABA-mediated postsynaptlc inhIbItIon in tissue cultured mouse spinal neurons. &am Rrs. 114: 530-535. Tower. D B (1976) GABA and seizures: Clintcal correlates in man. In: GAB.4 8t1 Nervous System FWKGOII (Roberts, E.. Chase, T. N. and Tower. D. B., Eds), pp. 461478. Raven Press, New York. Woodbury. D M. and Fmgl. E (1975) Drugs effective in the therapy of epllepsles In: Tlte Pharmacologtcal &MS of’ Therapeutw (Goodman, L. S. and GIlman Eds), 5th Edition. pp. 220-221. Macmrllan, New York

Lett. 52: 25 i-254.

REFERENCES

Adams, P. R., Banks F. W. and Constanti, A. (1981). Voltage clamp analysis of inhibitory synaptic action in crayfish stretch receptor neurons. Fedn Proc. Fedn Am. Sacs exp. Btol. In press. Anzelark, G., Horton, R. W.. Meldrum, B. S. and Sawaya, M. C. B. (1975). Anticonvulsant action of ethanolamineo-sulfate and di-n-propylacetate and the metabohsm of y-aminobutyric acid (GABA) m mice with audiogenic seizures. Biochem. Pharmac. 25: 4 13-417. Ayala, G. F and Johnston. D. (1977) The influence of phenytom on the fundamental electrical propertEes of simple neural systems. Epilepsra 18: 299-307. Barker, J. L. and McBurney. R. N. (1979). Phenobarbltone modulation of postsynaptIc GABA receptor function on cultured mammalian neurons. Proc R. Sot. Lond. E 206: 3 19-327.

Barker J. L. and Ransom, B. R. (1978). Pentobarbltone pharmacology of mammalIan central neurones grown in tissue culture. .!. Phvsrol., Land. 280: 355-372. Connors. B. W. (1979). Pentobarbital and diphenylhy dantoin effects on the excltabihty and GABA sensitivity of rat dorsal root ganghon cells. Sot. Neuroscl. Abstr. 5: 587.

Deisz, R. A. and Lux, H. D. (1977). Diphenylhydantoin prolongs postsynaptic inhibition and iontophoretic GABA action m the crayfish stretch receptor. Neurosct. Left. 5: 199-203.

Desarmemen, M., Feltz, P. and Headley, P. M. (I 979). The depolarizmg responses to GABA m rat sensory gangha in cico and m utro. A study of the role of glial uptake. J Physlol, Paris 75: 661-665. Emson, P. C (1976). Effects of chronic treatment with amino-oxy-acetic acid on sodium f~-dlpropylacetate on brain GABA levels and the develoiTment and regression of cobalt epileptic foci in rats.’ J. Neuroch&. 27: 1489- 1494.

Frere, R. C.. Young, A. B. and Macdonald. R. L. (1980). Effects of valproic acid and y-acetylemc GABA on metabohc enzymes in primary dissociated cell culture of mouse forebrain. Sot. Neuroscr. Absrr 6: 56 Gallagher, J. P., Higashi, H and Nishi. S. (1978). Characterization and tonic basis of GABA-Induced depolanzations recorded rn crrro from cat primary afferent neurones. J. Physlol., Land. 275: 263-282. Geller, H. M. (1980). Valproate interactlons with GABA ml-ubition in hypothalamic tissue cultures. Sot. Neurosci. Abstr. 6: 538.