TIPS-April 1988 [Vol. 91
`
`~i::tm(fllJ~~+++rT ~T
`~E1£tii;_JhlfLttttt±t
`Is the anticonvulsant
`mechanism of valproate linked
`to its interaction with the
`cerebral y-hydroxybutyrate
`system?
`Philippe Vayer, Christopher D. Cash
`and Michel Maitre
`
`There is now evidence that y-hydroxybutyrate (GHB) may be a neuro(cid:173)
`modulator in the CNS. Administration of this compound to various mammals
`at sub-anaesthetic doses induces brain electrical activity resembling that of
`human absence epilepsy. This effect is antagonized by the anticonvulsant
`drugs valproate and ethosuximide, and by the opiate antagonist naloxone. In
`vitro valproate and ethosuximide reduce the depolarization-induced release of
`GHB from rat hippocampal slices, and in vivo valproate antagonizes the
`increase in hippocampal cGMP levels induced by prior GHB administration.
`Michel Maitre and colleagues therefori: propose that the anticonvulsant
`action of valproate may be linked to its interaction with the endogenous GHB
`system.
`
`127
`
`including
`anticonvulsant drugs
`valproate and also by ethosux(cid:173)
`imide and trimethadione11• These
`epileptic phenomena may be con(cid:173)
`sidered as the result of a hyper(cid:173)
`activity of an endogenous brain
`system and, taking into account
`its pharmacological and biochem(cid:173)
`ical properties, we would suggest
`that GHB plays a role in the
`resultant hyperactivity. Thus, we
`propose that the mechanism of
`action of some antiepileptic drugs
`is due to their interactions with
`the GHB system in the brain. We
`shall conc.eittrate our attention on
`the specific effects of valproate
`on cerebral GHB functions and
`propose that these biochemical
`actions indicate a heuristic mol(cid:173)
`ecular model for the mode of
`action of certain anticonvulsant
`drugs.
`
`Anticonvulsant mechanism of
`valproate: hypothesis
`In rodents, valproate induces an
`increase in the cerebral GABA
`Ievel12, coinciding with a protec(cid:173)
`tive ~ffect against various forms
`of experimental epilepsy. This
`GABA increase has been ascribed
`to the inhibitory effect of valpro(cid:173)
`ate on the GABA shunt enzymes,
`rat brain hippocampus4• More·
`GABA transaminase12 and succi(cid:173)
`nic semialdehyde dehydrogenase13
`over, its half-life in rat brain
`(see Fig. 1). The inhibitory effect is
`(about 28 min) is at least as short
`more pronounced on the latter
`as those of established neuro(cid:173)
`enzyme and is competitive with
`transmitters.
`Administration of GHB to ani(cid:173)
`respect to the substrate, succinic
`semialdehyde13
`a metabolite
`mals and humans induces various
`-
`which has only been detected in
`neuropharmacological and neuro(cid:173)
`minute quantities in brain tissue.
`physiological effects,
`the most
`Valproate has no effect on GABA
`salient of which are: (1) modula(cid:173)
`tion of dopaminergic activity5
`release
`from cortical synapto(cid:173)
`somes preloaded with JabelJed
`(especially in the striatum) and an
`GABA14•
`increase in 5-HT turnover'>, and (2)
`Concomitant with the GABA
`a pronounced induction of seda(cid:173)
`level increase induced by val·
`tion leading at higher GHB doses
`is a marked reduction
`loss of
`i::onsciousness and
`proate
`to
`level15•
`anaesthesia7
`in
`cerebral aspartate
`• This latter property
`However, no interaction of val(cid:173)
`has been exploited clinically.
`proate with possible excitatory
`However, our principal interest
`synapses which function with
`here lies in the electroencephalo(cid:173)
`aspartate has been demonstrated.
`graphic perturbations induced by
`The increase in the cerebellar
`GHB administration. These effects
`cGMP level frequently observed
`have been most studied in the cat8
`in numerous forms of experimen(cid:173)
`and in the rat9• In these animals,
`tal epilepsy is greatly diminished
`GHB induces authentic experi(cid:173)
`by prior administration of valpro(cid:173)
`mental epilepsies which are char(cid:173)
`ate, whereas cAMP
`levels are
`acterized by mono- and poly(cid:173)
`unchanged16• This
`increase
`in
`phasic discharges. These dis(cid:173)
`cGMP level is considered as play(cid:173)
`charges are analogous to those
`ing a role in the induction and
`in human petit mal
`observed
`continuation of
`epileptogenic
`epilepsy10• The abnormal elec·
`activity17•
`troencephalograms
`induced by
`Finally, valproate has no effect
`GHB are antagonized by various
`© 1988, lil••Vl•r Publlaitlons, Carnbrldga 0165 - 61471881$02.0ll
`
`is a
`y-Hydroxybutyrate (GHB)
`normal brain metabolite derived
`primarily from GABA. Recent bio(cid:173)
`chemical
`and pharmacological
`data (see Ref. 1) lend support to
`the hypothesis that this substance
`may play a neuromodulator role in
`brain.
`GHB is heterogenously distri(cid:173)
`buted in brain tissue, where it is
`synthesized by a specific enzyme
`located exclusively in neurones.
`The highest concentration of this
`substance is found in the synapto(cid:173)
`somal fraction2• GHB is released
`by depolarization of brain tissue
`a Ca2+ -dependent
`slices
`by
`mechanism and is transported by
`a high-affinity, energy· and Na+_
`dependent system. High-affinity
`specific GHB binding sites are
`primarily located on the rostral
`area of the brain, the richest
`region being the hippocampus3
`•
`In
`vivo, GHB
`administration
`induces accumulation of cGMP in
`
`Philippe Vayer and Christopher Cash are
`postdoctoral research scientists at the Centre
`de Neurochimie du CNRS and Unite INSERM
`U44, 5 rue Blaise Pascal, 67084 Strasbourg,
`France. Michel Ma/Ire is an Associate Profes·
`sor at tire Faculty of Medicine, Unlverslti
`Louis Pasteur, 67084 Strasbourg, France.
`
`Ranbaxy Ex. 1022
`IPR Petition - USP 9,050,302
`
`

`
`128
`
`Glu "'
`
`Fig. 1. The
`GABA shunt and
`Krebs cycle in the
`mitochondrion. Vatpro-
`ate
`(represented
`by
`sclsaorsJ blocks the oxidation
`of succinic semisldehyde (SSA) to
`succinate by succinic semialdehyde
`dehydrogenase (SSADH). The increased
`pool of succinic semiatdehyde may either
`inhibit GABA·transaminare (GABA·T) or be transaminated to GABA, leading in both
`cases to raised GABA levels.
`
`TIPS -April 1988 [Vol. 9]
`
`out of the mitochondria into the
`cytoplasm (Fig. 1).
`A more plausible mechanism for
`in
`valproate-induced
`increases
`GHB levels is the powerful inhibi(cid:173)
`tion of nonspedfic succinic semi(cid:173)
`aldehyde reductase by valproate21 •
`[This enzyme is identical to the
`previously described ALR1, also
`referred to as glucuronate reduc(cid:173)
`tase or SSR1 (Ref. 22).) It is reported
`to be responsible for the catabol(cid:173)
`ism of GHB to succinic semialde(cid:173)
`hyde both in vitro22 and in vivo23
`•
`Under physiological conditions in
`vitro, GHB is degraded to GABA
`via nonspecific succinic semi(cid:173)
`aldehyde reductase and GABA
`b·ansaminase22• The operation of
`this catabolic pathway explains
`why
`inhibitors of nonspecific
`semialdehyde reductase such as
`valproate or ethosuximide, or
`GABA
`transaminase
`inhibitors
`such as y-vinyl GABA, increase
`GHB levels2 (Fig 2).
`Valproate and ethosuximide
`also considerably reduce the Ca2+ -
`dependent depolarization-induced
`release of GHB from hippocamJ!al
`and striatal slices frcm rat brain24•
`The IC50s for these two drugs in
`this test are compatible with their
`brain levels after administration
`of anticonvulsant doses.
`The latter phenomenon explains
`why the epileptogenic effects of
`GHB administration are attenu(cid:173)
`ated by valproate or ethosuxi(cid:173)
`mide11, and also explains the para(cid:173)
`dox that drugs such as valproate
`increase GHB brain levels25 whilst
`antagonizing its effects.
`The increase in cGMP levels in
`certain brain regions (such as the
`hippocampus) commonly seen
`after administration of convulsant
`drugs is concomitant with the
`
`glu
`
`which reduces sur.dnic semialde(cid:173)
`hyde to GHB (specific succinic
`semialdehyde reductase) is un(cid:173)
`affected by valproate21• In fact the
`increase in GABA levels induced
`by valproate has been ascribed to
`succinic semialdeh~de dehydro(cid:173)
`inhibition 3 . However,
`genase
`succinic semialdehyde dehydro(cid:173)
`genase is a mitochondrial enzyme
`whereas specific succinic semi(cid:173)
`aldehyde reductase is cytoplasmic.
`Thus the hypothesis that valproate
`increases GHB levels via an in(cid:173)
`creased synthesis due to higher
`precursor availability would de(cid:173)
`pend on an as yet unknown
`mechanism whereby excess suc(cid:173)
`cinic semialdehyde is transported
`
`1
`SSA -----~GHB
`
`on the GABAA receptor nor on
`benzodiazepine receptors, ruthough
`it displaces the convulsant dihydro(cid:173)
`picrotoxin from its binding site18•
`In general, it is thought that
`valproate potentiates GABAergic
`neurotransmission. However, it
`decreases its turnover rate and it
`does not modify its release. A
`possible direct GABA-like agonist
`action of valproate on a post(cid:173)
`synaptic site has yet
`to be
`demonstrated18, although a direct
`action of sodium valproate on
`action potentials of cultured neur(cid:173)
`ons has been reported19•
`
`Interactions of valproate with
`cerebral GHB system
`Several workers have reported
`that valproate administration to
`rodents brings about an increase
`in the brain GHB level20• This
`and dose(cid:173)
`increase
`is
`time-
`dependent, and appears to be due
`factors: direct
`two related.
`to
`inhibition of catabolic enzymes
`and a
`reduction
`in synaptic
`release.
`As the principal precursor of
`GHB is GABA, the mechanism of
`this GHB increase could be due to
`inhibition of the mitochondrial
`enzyme succinic semiald.ehyde de(cid:173)
`hydrogenase (ECl.2.1.24), which
`causes the pool of succinic semi(cid:173)
`aldehyde (the direct precursor of
`GHB) to be elevated. The enzyme
`
`valproata
`athoaux:lm:lda
`Fig. 2. Inhibition of GHB catabollsm by valproate, ethosuxlmlde and y-vinyt GABA
`(GVG). (1), Specific succinlc semlaldehyde reductase; (2), nonspecific succ/nlc
`semlaldehyde reductase; (3), GABA·transamlflase. Valproate and ethoauxlmlde
`directly inhibit nonspecific succinic semlaldehyde reductase, whereas GVG Inhibits
`GABA-transaminase leading to an Increase In succlnlc semlaldehyde (SSA), which Is a
`product ir.lllbltor of the dehydrogenase.
`
`2-ketcglu
`
`

`
`TIPS-April 1988 {Vol. 9]
`
`Fig. 3. Proposed chain-breaking action of
`valproate and ethoauxlmlde on the GHB
`system: GHB release is inhibited, and
`release of endorphin (endo) from GHB·
`activated presynaptic opioid-peptide term·
`inals is not stimulated. Naloxone (NLX)
`antagonizes the action of peptictergic synap·
`ses directly and thus indirectly antagonizes
`the epileptic phenomena induced by GHB.
`
`cGMP increase
`+epileptic disorders
`
`commencement and generaliza(cid:173)
`tion of massive depolarization
`phenomena17
`• The increase we
`have observed in hippocampal
`cGMP after GHB administration4
`could be considered as either a
`side-effect or a precursor to the
`evolution of the epilepsy induced
`by GHB, the latter directly or
`indirectly provoking a perturba(cid:173)
`tion in the membrane polarization
`of the hippocampus, a region rich
`in high-affinity GHB binding
`sites3• In various experimental
`epilepsies, therapeutic doses of
`\'!<lproate antagonize increases in
`cGMP levels; valproate also antag(cid:173)
`onizes, both in vivo and in vitro,
`the increase in cGMP levels seen
`after GHB administration4• Etho(cid:173)
`suximide also antagonizes this
`increase4 and thus it can be infer(cid:173)
`red
`that . their anticonvulsant
`effects are mediated by the inhibi(cid:173)
`tion of Ca2+ ~dependent release of
`GHB.
`In addition, the opiate receptor
`antagonist naloxone inhibits the
`GHB-induced cGMP increase4 and
`attenuate!'\
`the abnormal EEG
`activity2l. Taking into account the
`fact
`that GHB administration
`cause& an incnasc in dyno~hin
`level
`in the hippocampus , a
`region rich ;n µ-receptors, and
`that administration to the rat of
`certain opiates induces epileptic
`seizures which are antagonized
`by both valproate and etho(cid:173)
`suximide28, it could be suggested
`that these anticonvulsants act via a
`GHB-ergic mechanism which is
`linked to the endogenous opioid
`
`system (Fig. 3).
`GHB administration to animals
`most often brings about epileptic
`spiking activity. These pheno(cid:173)
`mena may represent functional
`overload of
`synapses which
`release GHB. There is much sup(cid:173)
`port for the existence of such
`synapses in the CNS. Valproate
`and ethosuximide modify
`the
`characteristics and functions of
`this group of synapses by inhibit(cid:173)
`ing the degradation of GHB and
`by blocking its Ca2+ adependent
`depolarization-induced
`release.
`Increases in cGMP brought about
`by GHB are blocked by valproate,
`ethosuximide and also by nalox(cid:173)
`one. These three drugs antagonize
`the epileptogenic activity of GHB.
`The effect of naloxone may indi(cid:173)
`cate that endorphins participate in
`the aetiology of this epileptic
`phenomenon. This biochemical
`mechanism could constitute a
`model for anticonvulsant agents.
`It would thus be of interest to
`synthesize new molecules which
`either reduce synaptic release of
`GHB or are antagonists at its
`receptor sites.
`
`Acknowledgement
`Supported by grant from DRET
`(no. 85/1200).
`
`References
`1 Vayer, P., Mandel, P. and Maitre, M.
`(1987) Life Sci. 41, 1547-1557
`2 Snead, O. C. (1987) J. Neurochein. 48,
`196-201
`3 Hechler, v .. WeiBsman, D., Mach, E.,
`Pujol, J. F. and Maitre, M.
`(1987)
`/. Neurochem. 49, 1025-1032
`
`129
`
`4 Vayer, P., Gobaille, S., Mandel, P. and
`Maitre, M. {1987) Life Sci. 41, 605-610
`5 Roth, R. H., Doherty, J.P. and Walters,
`J. R. {1980) Brain Res. 189, 556--560
`6 Hedner, Th. and Lundbor~, P. {1983)
`J. Neural Transm. 57, 39--48
`7 Laborit, H. (1964) Int. J. Neuropharma(cid:173)
`co/. 3, 433-452
`8 Winters, W. D. and Spooner, C. E.
`(1965) Int. /. Neuropharmacol. 4, 197-200
`9 Marcus, R. J., Winters, W. D., Mori, K.
`and Spooner, C. E. (1967) Int. /. Pharma(cid:173)
`col. 6, 175-185
`10 Godschalk, M., Dzoljic, M. R. and
`Bonta, I. L. {1977) Eur. /. Pharmacol. 44,
`105-111
`11 Godschalk, M., Dzoljic, M. R. and
`Bonta, I. L. (1976) Neurosci. Lett. 3, 145-
`150
`12 Simler, S., Ciesielski, L., Maitre, M.,
`Randrianarosa, H. and Mandel, P.
`(1973) Biochem. Pharmacol. 22, 1701-
`1708
`13 Van der Laan, J. W. and De Boer, Th.
`(1979) /. Neurochem. 32, 1769-1780
`14 Abdul-Ghani, A. S., Coutinho-Netto, J.,
`Druce, D. and Bradford, H. F. (1981)
`Biochem. Pharmacol. 30, 363-368
`15 Schechter, P. J., Tranier, Y. and Grove,}.
`(1978) ]. Neurochem. 31, 1325-1327
`16 McCandless, D. W., Feussner, G. K.,
`Lust, W. D. and Passoneau, J. V. (1979)
`J. Neurochem. 32, 755-760
`17 Gross, R. A. and Ferrendelli, J. A. (1979)
`Ann. Neural. 6, 296-310
`18 Chapman, A., Keane, P. E., Meldram,
`B. S., Simiand, J. and Vemieres, J.C.
`(1982) Prag. Neurobiol. 19, 315-359
`19 McLean, M. and McDonald, L. R. (1986)
`J. Pharmacol. Exp. Ther. 237, 1001-1011
`20 Snead, 0. C., Bearden, L. J. and Pegram,
`V. (1980) Neuropharmacology 19, 47-52
`21 Rumigny, J. F., Maitre, M., Cash, C. and
`Mandel, P. (1980) FEBS Lett. 117, 111-
`116
`22 Vayer, P., Schmitt, M., Bourguignon,
`J. J., Mandel, P. end Maitre, M. (1985)
`FEBS Lett. 190, 55-60
`23 Kaufman, E. end Nelson, Th. (1987)
`/. Neurochem. 48, 1935-1941
`24 Vayer, P., Charlier, B., Mandel, P. and
`(1987) J. Neurochem. 49,
`Maitre, M.
`1022-1024
`25 Snead, O. C., Bearden, L. J and Pegram,
`V. (1980) Neuropharmacology 19, 47-52
`26 Snead, O. C. and Bearden, L. J. (1980)
`Neurology 30, 832-838
`27 Lason, W., Przewlocka, B. and Przew·
`locka, R. (1983) Life Sci. 33, 599-602
`28 Snead, 0. C. and Bearden, L. J. (1982)
`Neuropharmacology 21, 1137-1144
`
`Transmembrane
`signalling
`
`Single copies of this centrefold
`can be purchased from our
`Gambridge office.
`
`See page 126.