- Email: [email protected]
Benzodiazepine antagonist flumazenil reduces bicuculline-induced enhancement of neuronal activity in the spinal cord
0028-3908/91 $3.00+0.00 Pergamon Press plc
Neuropharmacology Vo1.30, No.1, pp.107-111, 1991 Printed in Great Britain
BENZODIAZEPINE ANTAGONIST FLUMAZENIL REDUCES BICUCULLINE-INDUCED ENHANCEMENT OF NEURONAL ACTIVITY IN THE SPINAL CORD
P. Pole and I. DuEi6 Pharma Research Department, F. Hoffmann-La Roche Ltd, Grenzacherstrasse 124, CH-4002 Basel, Switzerland (Accepied 9 Novembeh. 19901
SUMMARY The interaction between the GABAA receptor antagonist bicuculline and the benzodiazepine receptor (BR) antagonist flumazenil was studied in the lumbosacral cord of spinal cats. Bicuculline (0.8 mg/kg i.v.) evoked epileptiform bursting of motoneurons in parallel with a depression of peripherally elicited dorsal root potentials (DRP) and dorsal root reflexes (DRR). A low dose of flumazenil (0.03 mg/kg i.v.) reduced the bursting activity of motoneurons as well as partially reversed the depression of DRP and DRR induced by bicuculline. It is suggested that an endogenous BR ligand down-regulating GABAA receptors is released during the bicuculline-evoked enhancement of neuronal activity and may contribute to the epileptiform discharge of motoneurons in the spinal cord. Key words: Cat spinal cord, GABAA receptor, endogenous ligand, epileptiform activity.
benzodiazepine
receptor
The spinal cord with its anatomically and physiologically defined afferent and efferent pathways as well as known epileptigenic properties of motoneurons has been proposed as a good model of experimental epilepsy (Schwindt and Crill, 1984). The deficiency of y-aminobutyric acid (GABA), a major inhibitory neurotransmittcr in the brain and spinal cord, is thought to be involved in some forms of epilepsy (Delgado-Escucta, Ward, Woodbury and Porter, 1986). Therefore, an enhancement of GABAergic inhibition may contribute to the action of anticpileptic drugs. The Binding that the specific GABAA receptor antagonist bicuculline consistently depressed primary afferent depolarization-related DRP and DRR gave an opportunity to study the effects of anticonvulsants on these mainly GABAmediated phenomena at the spinal level. Bcnzodiazcpines arc potent anticonvulsants drugs (Haefely, Pieri, Pole and Schaffncr, 1981) and enhance DRP after binding to and changing the conformation of high affinity binding sites located on the GABAA receptors ultimately leading to an increased chloride ion flux across the mcmbranc of primary affcrcnts (Haefcly and Pole, 1986). The exislencc of these sites. which mcdiatc many behaviorally rclcvant cffccts of bcnzodiazcpincs (bcnzodiazepinc rcceplor, BR), suggests the prcscnce of cndogcnous ligands for this receptor allosterically modulating the action of GABA in physiological and pathological states (Costa and Guidoui, 1987). In fact, intrinsic effects of the generally per SCinaclivc sclcclivc BR antagonist flumazcnil found under specific expcrimcntal conditions (File and Pcllow, 1986) should have given enough evidcncc for this proposal. However, relatively high doses of flumazenil used in thcsc cxpcriments prccludcd any reasonable interpretation, especially as at higher concentrations flumazcnil behaved like a partial BR agonist, i.e. slightly enhanced GABA-coupled depolarization and conductance in single cultured neurons (De Deyn and Macdonald, 1987; Chan and Farb, 1985). ‘Unexpcclcdly, flumazcnil was rcporlcd to reduce the frequency and severity of seizures in epileptic patients at doses which would have probably been below those inducing the partial agonism in animals (Scollo-Lavizzari, 1988). Being aware of problems when relating animal data to clinical observations in epilepsy, WChave tried ncvcrthclcss to explain the anticpileptic effect of flumazenil by using the spinal cord model. For this purpose, in the prcscnt study the interaction between bicuculline and flumazenil was 107
108
Preliminary
Notes
investigated under the condition of a marked enhancement of spinal neuronal activity induced by a convulsant dose of bicuculline. In addition to recording DRP and DRR, particular attention was paid to the observation of patterns of the motoneuronal discharge as a possible direct and sensitive measure of the convulsant action of bicucullinc and cvcmual anticonvulsant effects of flumazcnil.
METHODS Dcmils of tbc experimental set-up were described previously (Pole, Mohler and Haefely, 1974; Pole, 1987). The spinal cord of eight male cats (2.8-3.3 kg) was transected under ether anaesthesia at the Cl level and the animals were vendlaled artificially, end-tidal pCO2 being kept at 3.5-4 vol. %. Both aa. carotis comm. were ligated and the caudal brain stem was destroyed by a blunt spatula LOavoid pain from the head. The blood pressure was monitored in the femoral artery; the femoral vein was cat.heterized for drug injections. The body temperature, as well as that of mineral oil pools covering the spinal cord and nerves in the left hind limb, was maintained at 36O C. The lumbosacral cord was exposed by lamincctomy. DRP and DRR were rccordcd by a pair of Ag-AgCl electrodes auachcd to a pcriphcmlly cut Icft dorsal roollct of LhcL6 dorsal root. Monosynaptic rcllexes were recorded from lhe cm left venlral root L7 by a bipolar platinum electrode. The cemral ends of me ipsilalerally transected hind limb nerve to the gastrocnemius muscle and the peroneus comm. nerve were mounted on a bipolar electrode for maximal stimulation of the group I and II muscle affercms (5-10 mA, 0.1 ms pulses at 0.5 Hz). In four cats, the prolonged inhibilion of monosynaptic rcflcxcs cvokcd by a stimulation of the gastrocncmius muscle affcrcnts was induced by consumt submaximal conditioning single shocks applied to the peroncus comm. ncrvc at intervals of 20, 40, 60, 100 and 140 ms prior to volleys in the gastrocnemius afferems. Filaments conmining Lhc axons of y-motoncurons (identified as background action potentials with amplitude ~100 WV) and a-motoncurons (no1 spontaneously active action potentials with amplitude >lOO pV, elicited by me peripheral ncrvc stimulation) were isolated from the ventral rmt S 1 and placed on fine platinum clccuodes. The potentials were amplified, displayed on an oscilloscope and recorded photographically. A microcomputer program was used for averaging eight consecutive DRP, DRR and monosynaptic reflexes induced by 0.5 Hz stimulation of peripheral nerves as well as for counling the spontaneously occurring spikes of motoneurons. The output of spike analyzers was fed into a chart recorder, and the integrated rate of pulses per second was continuously recorded. 4 h elapsed between the withdrawal of ether and the onset of experiments. Drug effecis were evaluated in the following way. During stable control periods of 20 min duration, records and compulcr measurcmcnls were taken 5 min after injection of 0.1 ml/kg placebo aqueous microsuspcnsion, followed by injection of 0.8 mg/kg bicuculline dissolved in 0.1 ml/kg of saline acidified Lo pH 4. The cffccls of bicucullinc were rccordcd 5 min after its administration. Flumazenil (0.03 mg/kg), dissolved in 0.1 ml/kg of placebo, was injected in four out of eight cals 15-20 min after bicuculline and its cffccts wcrc mcasurcd 5 min later. The continuous chart recording of tic discharge rale of motoncurons allowed to follow the time course of the action of bicuculline and flumazenil. In two preliminary expcrimcms only bicucullinc was administered and, since its effects started to decline 30-40 min aficr injcclion, flumazcnil was administered in four main expcrimcms during a “steady-state” of bicucullinc-cliched changes in ncuronal activity. In another two preliminary experiments, flumazenil was injected aflcr bicucullinc cithcr in a lower (0.01 mg/kg) or in a higher dose (0.1 m&g) in order to find out the dose-rcsponsc rclauonship. Bicucullinc was purchased from Fluka and flumazenil (Ro 151788; clhyl-8-fluoro-5,6-dihydro-5-mcthyl-6-oxo-4H-imid~o [1,5a] [1,4] benzodiazepine-3-carboxylate) was synihesizcd by Dr. Hunkclcr.
RESULTS The control periods were charactcrizcd by a more or less regular pattern of spontaneously active ‘ymoloncurons (Fig. 1) and DRP with superimposed DRR induced by a low frequency (0.5 Hz) stimulation of the pcroneus comm. ncrvc (Fig, 2). In four cats, the prolonged inhibition of monosynaptic rcflcxcs was apparent at 20 and 100 ms intervals bctwcen conditioning volleys applied to the pcroneus comm. nerve and test shocks in the gasuocnemius muscle afferems (Fig. 2). Bicucullinc (0.8 mg/kg) induced the following cffccts in eight cats studied. Within 2-4 min after ils administration, the spontaneous discharge ram of y-motoncurons was increased accompanied by the appcarancc of spikes of a-motoncurons, as shown in a rcprcscntativc cxpcrimcnt (Fig. 1). The patlcrn changed to cpilcpiiform bursls scparatcd by episodes of reduced activity at intervals of 0.5 set to 3 min. Concomitanl modcratc incrcascs of the blood prcssurc wcrc in no obvious relation lo the enhancement of moloncuronal aclivily (Fig. 1). Immcdiatcly thcreaficr rccordcd DRP wcrc smaller and DRR were virtually supprcsscd (Fig. 2). In three out of four cats, bicuculline abolished the inhibition of
Preliminary Notes
109
monosynaptic reflexes, as shown in a typical experiment (Fig. 2). The effects of bicuculline started to decline 30-40 min after its injection and lasted at a progressively reduced magnitude up to 60-70 min (two preliminary experiments, not shown).
control
bicuculline
flumazenil
0.8 mg/kg
0.03
mg/kg
I
I
a
b
100
P/S ’ !jmin’
C
!-h!
!
!
!
!
! ,!, !
!
!
!
!
,--c(-rl-l-+-+!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
! .!
!
!
! ,!
I_
Fig. 1. Reversal by flumazcnil of the bicuculline-induced epileptiform activity of motoneurons in a spinal cat. Injections of placebo (control) and drugs are at arrows. In the upper horizontal row (a), representative samples of oscilloscope traces are shown with spontaneously active y-motoneurons (control), bursts of y- and a-motoneurons (bicuculline) and background discharge of y-motoneurons (flumazenil). In (b) and (c), the chart record shows the continuously recorded firing rate of motoneurons and arterial blood pressure, respectively. Note the difference in time calibration between oscilloscope samples and chart records.
Flumazcnil(O.03 mgkg) was injected 15-20 min after bicuculline, when the effects of the convulsant drug were pronounced and not yet in a declining phase. In all four cats, 1-2 min after the administration of flumazcnil the spikes of a-moloneurons disappeared and the firing rate of y-motoneurons reached the control levels with the regular pattcm obscrvcd before the injection of bicuculline (Fig. 1). Flumazenil antagonized the depressant effect of bicuculline on DRP and partially reversed the elimination by the convulsant of DRR (Fig. 2). In three out of four cats, the inhibition of monosynaptic reflexes was again visible at short conditioning-testing intervals as well as the 100 ms interval. The action of flumazenil lasted 15-25 min and as it disappeared, the effects of bicuculline, although to a smaller extent, reappeared (Fig. 1). In two preliminary experiments, a lower dose of flumazenil (0.01 mg/kg) was without effect whereas the higher dose (0.1 mg/kg) was not mom effective than the 0.03 my/kg dose in reducing the acl.ion of bicuculline (not shown). Thus, no dose-response relationship was found with doses of flumazenil used in this study.
DISCUSSION The present study shows that the bcnzodiazepinc antagonist flumazenil in a very low dose (0.03 mg/kg iv.) reversed disinhibitory efccts of the GABAA antagonist bicuculline in the cat spinal cord. If WCassume that bicucullinc disinhibitcd ‘I- as well as a-motoneurons by blocking GABAA receptors on spinal neurons, the reduction by flumazenil of the bicuculline-induced epileptiform activity might indicate an cnhanccment of GABAergic inhibition partially counteracting the blockade of GABAA receptors. By a similar action upon GABAA rcccptors located on primary afferent terminals, flumazenil may have reversed the depression of DRP and DRR as well as the reduction of prolonged inhibition of monosynaptic rcllcxes evoked by bicucullinc.
Preliminary Notes
110
control
0.5Hz
bicuculline
f lumazenil
IOSmV 50ms Fig. 2. Reversal by flumazenil of the bicuculline-evoked depression of DRP and DRR as well as of the reduced inhibition of monosynaptic reflexes in a spinal cat. Oscilloscope traces show (from left to right) averaged monosynaptic responses to stimulation of the gastrocnemius muscle afferents, elicited at various intervals (20, 40 60, 100 and 140 ms) after conditioning volleys in the peroneus comm. nerve. The last (sixth) spike in the sequence is the unconditioned monosynaptic reflex. It is followed by averaged DRP and superimposed DRR. evoked by the 1st conditioning volley and recorded immediately after the sequence of monosynaptic reflexes. From top to bottom: following the recording of potentials taken during a control period of 0.5 Hz stimulation (upper trace), records were obtained 5 min after injection of bicuculline (0.8 mgkg i.v.; intermediate trace) and 5 min after administration of flumazenil (0.03 mg/kg i.v.; lower trace), the latter drug being injected 15 min after bicuculline.
Basically, two mechanisms can underlie this action. 1) Flumazenil miaht not be a oure antagonist hraLiFpartialagonist at BR, Indeed, slight benzodiazcpine-like anticonvulsant effects of flumazenil were reported in some animal seizure models (File and Pcllow, 1986). However, in these studies doses of flumazcnil were higher than 1 mg/kg, suggesting that the partial agonism presumably responsible for these anticonvulsant cfects may not explain the anti-bicuculline action of flumazenil in the present investigation. At variance with a partial agonistic activity of flumazenil in our experiments is also the evidence that concentrations in the magnitude of 100 nM - 1 @JIof flumazenil were necessary to slightly enhance GABA-mediated effects in rat hippocampal slices (Kemp, Marshall, Wong and Woodruff, 1987) as well as in mouse and chick embryonic cultured neurons (De Deyn and Macdonald, 1987; Chan and Farb, 1985), whereas lower concentrations (10 nM - 40 nM) already antagonized the action of BR R agonists (DCDeyn and Macdonald, 1987). 2) * Fl m* nil m n lieand, which rcduccs the efficacy of GABA at GABAA receptors. The main candidate for such a negative allosteric modulator is diazcpam binding inhibitor (DBI) or one of its smaller fragments (Costa and Guidotti, 1987). DBI has been colocalized with GABA in distinct brain neurons, coreleased with GABA in brain slices and enriched in synaptosomes associated with synaptic vesicles, and its processing product octadecaneuropeptide (ODN) was found in the spinal cord (Ball, Burnett, Fountain, Ghatei and Bloom, 1986). DBI and/or ODN were also reported to induce flumazenil-reversible proconflict effects in rats and to reduce the GABA-coupled ionic current and depolarization in mouse embryonic spinal neurons (Costa and Guidotti, 1987). Regardless of the nature of endogenous ligand(s), its role in our experiments is suggested by the following considerations. The almost immediate reduction by flumazenil of the bicuculline-induced epilcptiform activity of motoneurons associated with a partial reversal of the convulsant’s action on DRP and DRR suggests a competition between flumazenil and an endogenous ligand down-regulating GABAA receptors rather than a complex interaction of flumazenil with seizureevoked receptor changes which follow a slower time course (HaefeFelyet al., 1981). Furthermore, if
Preliminary Notes
111
flumazenil only displaced an endogenous ligand from BR, the lack of dose-response relationship found in our study is precisely what one would expect in that case. This interpretation is supported by the finding that the higher frequency stimulation of primary afferents induced a depression of DRR and DRP. the prompt reversal of which by a low dose of flumazenil being interpreted as antagonism of an endogenous ligand (Pole and DuEi& “in press”). In this context, it is relevant that a neuronal coexistence of a ~classical neurotransmitler and a neuropeptide was found to have functional consequences depending on the (discharge rate in presynaptic neurons, the modulatory peptide being released upon higher frequency stimulation (Lundberg and Hbkfelt, 1983). By assuming a colocalization of GABA and a putative neuropeptide in GABAergic interneurons contacting primary afferents and motoneurons, then, according ito the concept of Lundberg and Hokfelt (1983). a high frequency neuronal discharge due to the bicuculline-evoked disinhibition might have released higher amounts of an endogenous BR ligand facilitating the epileptiform action of bicuculline. If a similar process occurs in epileptic foci where GABAergic inhibition seems to be reduced (Delgado-Escucta et al., 1986), the antiepileptic action of flumazenil may be explained by antagonism of a pathophysiologically released proconvulsant BR ligand contributing to epileptigenic phenomena by shifting the disturbed balance between excitatory and inhibitory neuronal activities even more towards disorganized excitation.
REFERENCES Elall, J.A., Burnett, P.W.J., Fountain, B.A., Ghatei, M.A. and Bloom. S.R. (1986). Octadecaneuropeptide, benzodiazepine ligand-like immunoreactivity in rat central nervous system, plasma and peripheral tissues. Neurosci. Lett. 72: 183-188. Chan,C.Y. and Farb, D.H. (1985). Modulation of neurouansmitler action: control of the y-aminobtyric acid response through the benzodiazcpine receptor. J. Neurosci. 5: 2365-2373. Costa, E. and Guidotti, A. (1987). Neuropeptides as cotmnsmitters: Modulatory effects at GABAergic synapses. In: Psychopharmacology: The Third Generation of Progress (Meltzer, H.Y., Ed.), pp. 425435, Raven Press, New York. De Deyn, P.P. and Macdonald, R.L. (1987). CGS 9896 and ZK 91296, but not CGS 8216 and Ro 151788, are pure benzodiazepine receptor antagonists on mouse neurons in culture. J. Pharmacol. Exp. Ther. 242: 48-55. Delgado-Escueta, A.V., Ward, A.A., Woodbury, D.M. and Porter, R.J. (1986). New wave of research in the epilepsies. In: Advances in Neurology, Vol. 44 (Delgado-Escueta, A.V.. Ward, A.A., Woodbury, D.M. and Porter, R.J., Eds.), pp. 3-55, Raven Press, New York. File, S. and Pellow, S. (1986). Intrinsic actions of the benzodiazepine receptor antagonist Ro 15-1788. Psychopharmacologia 88: 1- 11. Haefely, W. and Pole, P. (1986). Physiology of GABA enhancement by benzodiazepines and barbiturates In: Benzodiazepinc/GABA Receptors and Chloride Channels. Structural and Functional Properties (Olsen, R.W. and Venter, J.C., Eds.), pp. 97-133, Alan R. Liss, New York. Haefely, W., Pieri, L., Pole, P. and Schaffner, R. (1981). General pharmacology and neuropharmacology of benzodiazepinc derivatives. In: Handbook of Experimental Pharmacology, Vol. 55/B (Hoffmeister, F. and Stille, G., Eds.), pp. 1 l-262, Springer-Verlag, Berlin. Kemp, J.A., Marshall, G.R., Wong, E.H.F. and Woodruff, G.N. (1987). The affinities, polencies and efficacies of some bcnzodiazepinc receptor agonists, antagonists and invcrsc agonists al rat hippocampal GABAA receptors. Br. J. Pharmacol. 91: 601-608. Lundberg, J.M. and Hokfelt, T. (1983). Coexistence of peptides and classical neurotransmitters. TINS 6: 325-333. Pole, P. (1987). NMDA receptors mediate background and excessive activity of y-motoneurons in the spinal cord. Eur. J. Pharmacol. 144: 113-I 15. Pole, P. and DuEiC, I. (1990). Affcrcnt stimulation frequency modulates GABAergic phenomena in the spinal cord: rcvcrsal by bcnzodiazcpinc antagonists. Brain Rcs. (“in press”). Pole, P., Mohler H. and Haefcly, W. (1974). The effect of diazepam on spinal cord activities: possible sites and mechanisms of action. Naunyn Schmiedebergs Arch. Pharmacol. 284: 319-337. Schwindt, PC. and Crill, W.E. (1984). The spinal cord model of experimental epilepsy. In: Electrophysiology of Epilepsy (Schwarzkroin, P.A. and Wheal, H.V., Eds.), pp. 220-251, Academic Press, London. Scollo-Lavizzari, G. (1988). The clinical anticonvulsant effects of flumazenil, a benzodiazepine antagonist. Eur. J. Anaeslh., Suppl. 2: 129-138.
Neuropharmacology Vo1.30, No.1, pp.107-111, 1991 Printed in Great Britain
BENZODIAZEPINE ANTAGONIST FLUMAZENIL REDUCES BICUCULLINE-INDUCED ENHANCEMENT OF NEURONAL ACTIVITY IN THE SPINAL CORD
P. Pole and I. DuEi6 Pharma Research Department, F. Hoffmann-La Roche Ltd, Grenzacherstrasse 124, CH-4002 Basel, Switzerland (Accepied 9 Novembeh. 19901
SUMMARY The interaction between the GABAA receptor antagonist bicuculline and the benzodiazepine receptor (BR) antagonist flumazenil was studied in the lumbosacral cord of spinal cats. Bicuculline (0.8 mg/kg i.v.) evoked epileptiform bursting of motoneurons in parallel with a depression of peripherally elicited dorsal root potentials (DRP) and dorsal root reflexes (DRR). A low dose of flumazenil (0.03 mg/kg i.v.) reduced the bursting activity of motoneurons as well as partially reversed the depression of DRP and DRR induced by bicuculline. It is suggested that an endogenous BR ligand down-regulating GABAA receptors is released during the bicuculline-evoked enhancement of neuronal activity and may contribute to the epileptiform discharge of motoneurons in the spinal cord. Key words: Cat spinal cord, GABAA receptor, endogenous ligand, epileptiform activity.
benzodiazepine
receptor
The spinal cord with its anatomically and physiologically defined afferent and efferent pathways as well as known epileptigenic properties of motoneurons has been proposed as a good model of experimental epilepsy (Schwindt and Crill, 1984). The deficiency of y-aminobutyric acid (GABA), a major inhibitory neurotransmittcr in the brain and spinal cord, is thought to be involved in some forms of epilepsy (Delgado-Escucta, Ward, Woodbury and Porter, 1986). Therefore, an enhancement of GABAergic inhibition may contribute to the action of anticpileptic drugs. The Binding that the specific GABAA receptor antagonist bicuculline consistently depressed primary afferent depolarization-related DRP and DRR gave an opportunity to study the effects of anticonvulsants on these mainly GABAmediated phenomena at the spinal level. Bcnzodiazcpines arc potent anticonvulsants drugs (Haefely, Pieri, Pole and Schaffncr, 1981) and enhance DRP after binding to and changing the conformation of high affinity binding sites located on the GABAA receptors ultimately leading to an increased chloride ion flux across the mcmbranc of primary affcrcnts (Haefcly and Pole, 1986). The exislencc of these sites. which mcdiatc many behaviorally rclcvant cffccts of bcnzodiazcpincs (bcnzodiazepinc rcceplor, BR), suggests the prcscnce of cndogcnous ligands for this receptor allosterically modulating the action of GABA in physiological and pathological states (Costa and Guidoui, 1987). In fact, intrinsic effects of the generally per SCinaclivc sclcclivc BR antagonist flumazcnil found under specific expcrimcntal conditions (File and Pcllow, 1986) should have given enough evidcncc for this proposal. However, relatively high doses of flumazenil used in thcsc cxpcriments prccludcd any reasonable interpretation, especially as at higher concentrations flumazcnil behaved like a partial BR agonist, i.e. slightly enhanced GABA-coupled depolarization and conductance in single cultured neurons (De Deyn and Macdonald, 1987; Chan and Farb, 1985). ‘Unexpcclcdly, flumazcnil was rcporlcd to reduce the frequency and severity of seizures in epileptic patients at doses which would have probably been below those inducing the partial agonism in animals (Scollo-Lavizzari, 1988). Being aware of problems when relating animal data to clinical observations in epilepsy, WChave tried ncvcrthclcss to explain the anticpileptic effect of flumazenil by using the spinal cord model. For this purpose, in the prcscnt study the interaction between bicuculline and flumazenil was 107
108
Preliminary
Notes
investigated under the condition of a marked enhancement of spinal neuronal activity induced by a convulsant dose of bicuculline. In addition to recording DRP and DRR, particular attention was paid to the observation of patterns of the motoneuronal discharge as a possible direct and sensitive measure of the convulsant action of bicucullinc and cvcmual anticonvulsant effects of flumazcnil.
METHODS Dcmils of tbc experimental set-up were described previously (Pole, Mohler and Haefely, 1974; Pole, 1987). The spinal cord of eight male cats (2.8-3.3 kg) was transected under ether anaesthesia at the Cl level and the animals were vendlaled artificially, end-tidal pCO2 being kept at 3.5-4 vol. %. Both aa. carotis comm. were ligated and the caudal brain stem was destroyed by a blunt spatula LOavoid pain from the head. The blood pressure was monitored in the femoral artery; the femoral vein was cat.heterized for drug injections. The body temperature, as well as that of mineral oil pools covering the spinal cord and nerves in the left hind limb, was maintained at 36O C. The lumbosacral cord was exposed by lamincctomy. DRP and DRR were rccordcd by a pair of Ag-AgCl electrodes auachcd to a pcriphcmlly cut Icft dorsal roollct of LhcL6 dorsal root. Monosynaptic rcllexes were recorded from lhe cm left venlral root L7 by a bipolar platinum electrode. The cemral ends of me ipsilalerally transected hind limb nerve to the gastrocnemius muscle and the peroneus comm. nerve were mounted on a bipolar electrode for maximal stimulation of the group I and II muscle affercms (5-10 mA, 0.1 ms pulses at 0.5 Hz). In four cats, the prolonged inhibilion of monosynaptic rcflcxcs cvokcd by a stimulation of the gastrocncmius muscle affcrcnts was induced by consumt submaximal conditioning single shocks applied to the peroncus comm. ncrvc at intervals of 20, 40, 60, 100 and 140 ms prior to volleys in the gastrocnemius afferems. Filaments conmining Lhc axons of y-motoncurons (identified as background action potentials with amplitude ~100 WV) and a-motoncurons (no1 spontaneously active action potentials with amplitude >lOO pV, elicited by me peripheral ncrvc stimulation) were isolated from the ventral rmt S 1 and placed on fine platinum clccuodes. The potentials were amplified, displayed on an oscilloscope and recorded photographically. A microcomputer program was used for averaging eight consecutive DRP, DRR and monosynaptic reflexes induced by 0.5 Hz stimulation of peripheral nerves as well as for counling the spontaneously occurring spikes of motoneurons. The output of spike analyzers was fed into a chart recorder, and the integrated rate of pulses per second was continuously recorded. 4 h elapsed between the withdrawal of ether and the onset of experiments. Drug effecis were evaluated in the following way. During stable control periods of 20 min duration, records and compulcr measurcmcnls were taken 5 min after injection of 0.1 ml/kg placebo aqueous microsuspcnsion, followed by injection of 0.8 mg/kg bicuculline dissolved in 0.1 ml/kg of saline acidified Lo pH 4. The cffccls of bicucullinc were rccordcd 5 min after its administration. Flumazenil (0.03 mg/kg), dissolved in 0.1 ml/kg of placebo, was injected in four out of eight cals 15-20 min after bicuculline and its cffccts wcrc mcasurcd 5 min later. The continuous chart recording of tic discharge rale of motoncurons allowed to follow the time course of the action of bicuculline and flumazenil. In two preliminary expcrimcms only bicucullinc was administered and, since its effects started to decline 30-40 min aficr injcclion, flumazcnil was administered in four main expcrimcms during a “steady-state” of bicucullinc-cliched changes in ncuronal activity. In another two preliminary experiments, flumazenil was injected aflcr bicucullinc cithcr in a lower (0.01 mg/kg) or in a higher dose (0.1 m&g) in order to find out the dose-rcsponsc rclauonship. Bicucullinc was purchased from Fluka and flumazenil (Ro 151788; clhyl-8-fluoro-5,6-dihydro-5-mcthyl-6-oxo-4H-imid~o [1,5a] [1,4] benzodiazepine-3-carboxylate) was synihesizcd by Dr. Hunkclcr.
RESULTS The control periods were charactcrizcd by a more or less regular pattern of spontaneously active ‘ymoloncurons (Fig. 1) and DRP with superimposed DRR induced by a low frequency (0.5 Hz) stimulation of the pcroneus comm. ncrvc (Fig, 2). In four cats, the prolonged inhibition of monosynaptic rcflcxcs was apparent at 20 and 100 ms intervals bctwcen conditioning volleys applied to the pcroneus comm. nerve and test shocks in the gasuocnemius muscle afferems (Fig. 2). Bicucullinc (0.8 mg/kg) induced the following cffccts in eight cats studied. Within 2-4 min after ils administration, the spontaneous discharge ram of y-motoncurons was increased accompanied by the appcarancc of spikes of a-motoncurons, as shown in a rcprcscntativc cxpcrimcnt (Fig. 1). The patlcrn changed to cpilcpiiform bursls scparatcd by episodes of reduced activity at intervals of 0.5 set to 3 min. Concomitanl modcratc incrcascs of the blood prcssurc wcrc in no obvious relation lo the enhancement of moloncuronal aclivily (Fig. 1). Immcdiatcly thcreaficr rccordcd DRP wcrc smaller and DRR were virtually supprcsscd (Fig. 2). In three out of four cats, bicuculline abolished the inhibition of
Preliminary Notes
109
monosynaptic reflexes, as shown in a typical experiment (Fig. 2). The effects of bicuculline started to decline 30-40 min after its injection and lasted at a progressively reduced magnitude up to 60-70 min (two preliminary experiments, not shown).
control
bicuculline
flumazenil
0.8 mg/kg
0.03
mg/kg
I
I
a
b
100
P/S ’ !jmin’
C
!-h!
!
!
!
!
! ,!, !
!
!
!
!
,--c(-rl-l-+-+!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
! .!
!
!
! ,!
I_
Fig. 1. Reversal by flumazcnil of the bicuculline-induced epileptiform activity of motoneurons in a spinal cat. Injections of placebo (control) and drugs are at arrows. In the upper horizontal row (a), representative samples of oscilloscope traces are shown with spontaneously active y-motoneurons (control), bursts of y- and a-motoneurons (bicuculline) and background discharge of y-motoneurons (flumazenil). In (b) and (c), the chart record shows the continuously recorded firing rate of motoneurons and arterial blood pressure, respectively. Note the difference in time calibration between oscilloscope samples and chart records.
Flumazcnil(O.03 mgkg) was injected 15-20 min after bicuculline, when the effects of the convulsant drug were pronounced and not yet in a declining phase. In all four cats, 1-2 min after the administration of flumazcnil the spikes of a-moloneurons disappeared and the firing rate of y-motoneurons reached the control levels with the regular pattcm obscrvcd before the injection of bicuculline (Fig. 1). Flumazenil antagonized the depressant effect of bicuculline on DRP and partially reversed the elimination by the convulsant of DRR (Fig. 2). In three out of four cats, the inhibition of monosynaptic reflexes was again visible at short conditioning-testing intervals as well as the 100 ms interval. The action of flumazenil lasted 15-25 min and as it disappeared, the effects of bicuculline, although to a smaller extent, reappeared (Fig. 1). In two preliminary experiments, a lower dose of flumazenil (0.01 mg/kg) was without effect whereas the higher dose (0.1 mg/kg) was not mom effective than the 0.03 my/kg dose in reducing the acl.ion of bicuculline (not shown). Thus, no dose-response relationship was found with doses of flumazenil used in this study.
DISCUSSION The present study shows that the bcnzodiazepinc antagonist flumazenil in a very low dose (0.03 mg/kg iv.) reversed disinhibitory efccts of the GABAA antagonist bicuculline in the cat spinal cord. If WCassume that bicucullinc disinhibitcd ‘I- as well as a-motoneurons by blocking GABAA receptors on spinal neurons, the reduction by flumazenil of the bicuculline-induced epileptiform activity might indicate an cnhanccment of GABAergic inhibition partially counteracting the blockade of GABAA receptors. By a similar action upon GABAA rcccptors located on primary afferent terminals, flumazenil may have reversed the depression of DRP and DRR as well as the reduction of prolonged inhibition of monosynaptic rcllcxes evoked by bicucullinc.
Preliminary Notes
110
control
0.5Hz
bicuculline
f lumazenil
IOSmV 50ms Fig. 2. Reversal by flumazenil of the bicuculline-evoked depression of DRP and DRR as well as of the reduced inhibition of monosynaptic reflexes in a spinal cat. Oscilloscope traces show (from left to right) averaged monosynaptic responses to stimulation of the gastrocnemius muscle afferents, elicited at various intervals (20, 40 60, 100 and 140 ms) after conditioning volleys in the peroneus comm. nerve. The last (sixth) spike in the sequence is the unconditioned monosynaptic reflex. It is followed by averaged DRP and superimposed DRR. evoked by the 1st conditioning volley and recorded immediately after the sequence of monosynaptic reflexes. From top to bottom: following the recording of potentials taken during a control period of 0.5 Hz stimulation (upper trace), records were obtained 5 min after injection of bicuculline (0.8 mgkg i.v.; intermediate trace) and 5 min after administration of flumazenil (0.03 mg/kg i.v.; lower trace), the latter drug being injected 15 min after bicuculline.
Basically, two mechanisms can underlie this action. 1) Flumazenil miaht not be a oure antagonist hraLiFpartialagonist at BR, Indeed, slight benzodiazcpine-like anticonvulsant effects of flumazenil were reported in some animal seizure models (File and Pcllow, 1986). However, in these studies doses of flumazcnil were higher than 1 mg/kg, suggesting that the partial agonism presumably responsible for these anticonvulsant cfects may not explain the anti-bicuculline action of flumazenil in the present investigation. At variance with a partial agonistic activity of flumazenil in our experiments is also the evidence that concentrations in the magnitude of 100 nM - 1 @JIof flumazenil were necessary to slightly enhance GABA-mediated effects in rat hippocampal slices (Kemp, Marshall, Wong and Woodruff, 1987) as well as in mouse and chick embryonic cultured neurons (De Deyn and Macdonald, 1987; Chan and Farb, 1985), whereas lower concentrations (10 nM - 40 nM) already antagonized the action of BR R agonists (DCDeyn and Macdonald, 1987). 2) * Fl m* nil m n lieand, which rcduccs the efficacy of GABA at GABAA receptors. The main candidate for such a negative allosteric modulator is diazcpam binding inhibitor (DBI) or one of its smaller fragments (Costa and Guidotti, 1987). DBI has been colocalized with GABA in distinct brain neurons, coreleased with GABA in brain slices and enriched in synaptosomes associated with synaptic vesicles, and its processing product octadecaneuropeptide (ODN) was found in the spinal cord (Ball, Burnett, Fountain, Ghatei and Bloom, 1986). DBI and/or ODN were also reported to induce flumazenil-reversible proconflict effects in rats and to reduce the GABA-coupled ionic current and depolarization in mouse embryonic spinal neurons (Costa and Guidotti, 1987). Regardless of the nature of endogenous ligand(s), its role in our experiments is suggested by the following considerations. The almost immediate reduction by flumazenil of the bicuculline-induced epilcptiform activity of motoneurons associated with a partial reversal of the convulsant’s action on DRP and DRR suggests a competition between flumazenil and an endogenous ligand down-regulating GABAA receptors rather than a complex interaction of flumazenil with seizureevoked receptor changes which follow a slower time course (HaefeFelyet al., 1981). Furthermore, if
Preliminary Notes
111
flumazenil only displaced an endogenous ligand from BR, the lack of dose-response relationship found in our study is precisely what one would expect in that case. This interpretation is supported by the finding that the higher frequency stimulation of primary afferents induced a depression of DRR and DRP. the prompt reversal of which by a low dose of flumazenil being interpreted as antagonism of an endogenous ligand (Pole and DuEi& “in press”). In this context, it is relevant that a neuronal coexistence of a ~classical neurotransmitler and a neuropeptide was found to have functional consequences depending on the (discharge rate in presynaptic neurons, the modulatory peptide being released upon higher frequency stimulation (Lundberg and Hbkfelt, 1983). By assuming a colocalization of GABA and a putative neuropeptide in GABAergic interneurons contacting primary afferents and motoneurons, then, according ito the concept of Lundberg and Hokfelt (1983). a high frequency neuronal discharge due to the bicuculline-evoked disinhibition might have released higher amounts of an endogenous BR ligand facilitating the epileptiform action of bicuculline. If a similar process occurs in epileptic foci where GABAergic inhibition seems to be reduced (Delgado-Escucta et al., 1986), the antiepileptic action of flumazenil may be explained by antagonism of a pathophysiologically released proconvulsant BR ligand contributing to epileptigenic phenomena by shifting the disturbed balance between excitatory and inhibitory neuronal activities even more towards disorganized excitation.
REFERENCES Elall, J.A., Burnett, P.W.J., Fountain, B.A., Ghatei, M.A. and Bloom. S.R. (1986). Octadecaneuropeptide, benzodiazepine ligand-like immunoreactivity in rat central nervous system, plasma and peripheral tissues. Neurosci. Lett. 72: 183-188. Chan,C.Y. and Farb, D.H. (1985). Modulation of neurouansmitler action: control of the y-aminobtyric acid response through the benzodiazcpine receptor. J. Neurosci. 5: 2365-2373. Costa, E. and Guidotti, A. (1987). Neuropeptides as cotmnsmitters: Modulatory effects at GABAergic synapses. In: Psychopharmacology: The Third Generation of Progress (Meltzer, H.Y., Ed.), pp. 425435, Raven Press, New York. De Deyn, P.P. and Macdonald, R.L. (1987). CGS 9896 and ZK 91296, but not CGS 8216 and Ro 151788, are pure benzodiazepine receptor antagonists on mouse neurons in culture. J. Pharmacol. Exp. Ther. 242: 48-55. Delgado-Escueta, A.V., Ward, A.A., Woodbury, D.M. and Porter, R.J. (1986). New wave of research in the epilepsies. In: Advances in Neurology, Vol. 44 (Delgado-Escueta, A.V.. Ward, A.A., Woodbury, D.M. and Porter, R.J., Eds.), pp. 3-55, Raven Press, New York. File, S. and Pellow, S. (1986). Intrinsic actions of the benzodiazepine receptor antagonist Ro 15-1788. Psychopharmacologia 88: 1- 11. Haefely, W. and Pole, P. (1986). Physiology of GABA enhancement by benzodiazepines and barbiturates In: Benzodiazepinc/GABA Receptors and Chloride Channels. Structural and Functional Properties (Olsen, R.W. and Venter, J.C., Eds.), pp. 97-133, Alan R. Liss, New York. Haefely, W., Pieri, L., Pole, P. and Schaffner, R. (1981). General pharmacology and neuropharmacology of benzodiazepinc derivatives. In: Handbook of Experimental Pharmacology, Vol. 55/B (Hoffmeister, F. and Stille, G., Eds.), pp. 1 l-262, Springer-Verlag, Berlin. Kemp, J.A., Marshall, G.R., Wong, E.H.F. and Woodruff, G.N. (1987). The affinities, polencies and efficacies of some bcnzodiazepinc receptor agonists, antagonists and invcrsc agonists al rat hippocampal GABAA receptors. Br. J. Pharmacol. 91: 601-608. Lundberg, J.M. and Hokfelt, T. (1983). Coexistence of peptides and classical neurotransmitters. TINS 6: 325-333. Pole, P. (1987). NMDA receptors mediate background and excessive activity of y-motoneurons in the spinal cord. Eur. J. Pharmacol. 144: 113-I 15. Pole, P. and DuEiC, I. (1990). Affcrcnt stimulation frequency modulates GABAergic phenomena in the spinal cord: rcvcrsal by bcnzodiazcpinc antagonists. Brain Rcs. (“in press”). Pole, P., Mohler H. and Haefcly, W. (1974). The effect of diazepam on spinal cord activities: possible sites and mechanisms of action. Naunyn Schmiedebergs Arch. Pharmacol. 284: 319-337. Schwindt, PC. and Crill, W.E. (1984). The spinal cord model of experimental epilepsy. In: Electrophysiology of Epilepsy (Schwarzkroin, P.A. and Wheal, H.V., Eds.), pp. 220-251, Academic Press, London. Scollo-Lavizzari, G. (1988). The clinical anticonvulsant effects of flumazenil, a benzodiazepine antagonist. Eur. J. Anaeslh., Suppl. 2: 129-138.