New drugs for epilepsy: A rapidly changing scene A. Richens University of Wales College of Medicine Heath Park
Introduction In all the time that Harry Meinardi has been an epileptologist, only three new antiepileptic drugs have been marketed in Holland, if benzodiazepine drugs are excluded. Valproate was given its first trials in the early 1960s and was marketed in France, where it was discovered, in 1967. Although it is now established worldwide as an effective antiepileptic drug in all seizure types, its mode of action has been elusive. The balance of evidence suggests that it increases activity in inhibitory pathways in the brain, but the way in which it achieves this at a molecular level is uncertain. More recently, vigabatrin with a known mode of action, and oxcarbazepine with a less certain mode of action, were introduced. It is undoubtedly valuable to have a new drug with which to treat our patients, but it is better still to have a compound which also teaches us something about synaptic function and its possible derangement in epilepsy. In the last 15 years various substances have been synthesised which modify excitatory and inhibitory amino-acid transmission and we are now seeing the benefits of this research in the clinic. I will deal first with the processes of synaptic inhibition and excitation and then go on to give a brief sketch of new drugs which are known to influence these systems, as well as several more whose mode of action is as yet uncertain.
Neurotransmitters Neurotransmitters in the central nervous system are divided into two types depending on the speed of the transduction process at the post-synaptic membrane. Fast neurotransmitters give rise to post-synaptic responses in
Address: A. Richens Department of Pharmacologyand Therapeutics Heath Park Cardiff CF4 4XN U.K.
the millisecond range, while slow neurotransmitters operate over seconds or minutes. These responses may be mediated either through ionic mechanisms (ionotropic) or through a second messenger system (rnetabotropic). Two major fast neurotransmitters in the brain are glutamate for excitatory responses and gamma-amino butyric acid (GABA) for inhibition. Signalling in most synapses in the central nervous system is by one or other of these transmitters. A small proportion of synapses, however, use slow neurotransmitters such as monoamines and neuropeptides which have a neuromodulatory role. While these may play some part in epilepsy, particularly accounting for the decrease in convulsive threshold following administration of psychoactive drugs, it seems likely that the major underlying cause is a defect in the transmission at those synapses using glutamate or GABA. It has to be said, however, that direct evidence from man for such a defect is still flimsy. Changes in receptor numbers and neurotransmitter concentrations in brain tissue taken from epileptic foci may be non-specific effects of brain damage or drug treatment rather than the underlying cause of the epilepsy.
GABA as an inhibitory neurotransmitter GABA has received much attention over the last 25 years and evidence has accumulated that some of the conventional antiepileptic drugs interact with inhibitory synaptic mechanisms using this substance as a transmitter. It is estimated that it is involved in 30 % of all synapses in the central nervous system. There are two types of inhibitory mechanism, presynaptic and postsynaptic. In the former, GABA acts on a presynaptic terminal of an excitatory neurone to prevent release of transmitter; this form of inhibition is found predominantly in the spinal cord. Postsynaptic inhibition is the main mechanism found in the brain, and it is at this site that many antiepileptic drugs have their action. When occupied by the endogenous ligand, GABA receptors invariably produce hyperpolarization of the postsynaptic neurone. There are two types of receptor, GABA, and GABA, (,). The former is blocked by a drug
Richens Activation of GABA, receptors opens a chloride channel in the postsynaptic membrane, allowing an influx of negatively charged chloride ions which leads to hyperpolarization of the neurone and consequently an inhibition of f i g . This is termed ionotropic transmission because it involves ionic fluxes. GABA is synthesised in presynaptic nerve terminals from glucose entering the tricarboxylic acid cycle (Figure 2). The transamination of -ketoglutamate generates Lglutamic acid (glutamate) the immediate precursor of GABA. This takes place in a closed loop called the ‘GABA shunt’ which is confined to the nerve terminal. GABA is formed by decarboxylation of L-glutamic acid by the enzyme glutamic acid decarboxylase (GAD), which is found only in neurones and is probably the rate-limiting step in the synthesis of GABA. The next reaction in the shunt is the transamination of GABA to succinic semialdehyde by GABA- transaminase (GABA-T), and at the same time a molecule of glutamate is synthesised for every molecule of GABA destroyed.
l GABA recognition
I
BenzodiazeDines
-
K . .
Chloride channnel
1
fig 1. An inhkhry synapse at which GABA is releasedto prcducepost-sppkinhibrbon at a M A Areceptor.
used in experimental work, bicucculine, and is the receptor which is involved in the treatment of epilepsy (Figure 1). It has recently been isolated by molecular cloning techniques (2).
Oxaloacetate Acetyl
Clinical applications Enhancement of GABA mediated inhibition has an
,
anticonvulsant effect. The various ways in which this can be achieved are summarised in Table 1. GABA itself does not cross the blood-brain barrier and therefore in order to stimulate GABA receptors directly by systemic administration, it has to be linked to a lipid soluble moiety that facilitates penetration into the brain. Progabide was found to be effective in animals (3) but is less satisfactory in man (4). Alternatively, GABA agonist
Glut imate GAD
OC-Ketoglutarate Tr icarboxyIic acid cycle
GABA GABA-T
Citrate
Succinic semi-aldehyde
New Drugs for Epilepsy
Mechanism
Ep?limental
onrgsfor
CompOUnds
human use
GPB,agonism
MusdmlTHlP
Progab&
Bemodiarepine receptor agmism
Benzcd@nes
Direct action at C1-ionophore
EaMJrates
phemone
Reuptake inhibmon
Nipemtic add
liqabine
W - T inhibmon
EthmhmiW suiphate
Vgababln
Benzodiine
dws
molecules can be synthesised which mimic the actions of GABA. Muscimol and THIP are such compounds but unfortunately they do not seem to be active in human epilepsy. The reason for this is unclear. Another possible way to augment GABA inhibition is by using drugs which bind to various sites on the receptor complex and enhance the effect of synaptically released GABA. Theoretically, this type of pharmacological action should preserve the necessary spatial and temporal linkage to suppress abnormal neuronal output but have little effect on normal neuronal function. Benzodiazepines are known to act on specific benzodiazepine binding sites located on the GABA, receptor complex which facilitate GABA transmission and increase the frequency of the chloride channel opening. Barbiturates, on the other hand, act at a separate site associated with the channel, and prolong the open time. Inhibition of GABA reuptake, particularly into glial cells will enhance the inhibitory effect of the transmitter. Nipecotic acid does this following intracerebral injection into animals. Tiagabine is a compound comprising nipecotic acid linked to a lipophilic anchor, designed to allow its penetration into the brain following systemic administration. It is undergoing Phase II clinical testing (5). Inhibition of GABA-T, the enzyme that transamhates GABA, will cause an increase in cerebral levels of the transmitter, thereby producing an anticonvulsant effect. Vigabatrin is now an established drug in Europe used for treating partial and secondarily generalised seizures. It is a suicide inhibitor of the enzyme and has a long duration of action; recovery from its effects requires the synthesis of a new enzyme.
Glutamate as an excitatory neurotransmitter Besides being the immediate precursor for the synthesis of GABA, glutamate is the predominant excitatory amino-
acid neurotransmitter in the brain and spinal cord. Its postsynaptic action is always excitatory, mediated through either an ionotropic or metabotropic action. Stimulation of glutamate receptors causes depolarisation of the postsynaptic membrane with usually a fast onset and termination of effect. Four subtypes of post-synaptic receptor have been described, named after the agonists which preferentially bind to them: NMDA, kainate, quisqualate and AMPA (6). It appears that the NMDA (n-methyl-D-aspartate) receptor is the most likely candidate for a role in epilepsy because agonists for this receptor have been shown to be effective anticonvulsants in animals (7) (Figure 3). Like the GABA receptor complex, it has multiple binding sites which modulate the affinity of the endogenous ligand for its receptor. The Nh4DA complex surrounds a cation channel which allows the influx of sodium and calcium ions and the efflux of potassium ions. Within the channel there is a binding site for magnesium ions which at resting potential block the channel. Depolarisation of the postsynaptic neurone, by activation of quisqualate or AMPA receptors, is necessary before the magnesium block is removed, a so-called 'voltage dependent' block. The dissociative anaesthetics, ketamine and phencyclidine, have been shown to bind to another site
Presynaptic excitory nerve terminals containing glutamate in discrete vesicles
Glutamate recognition s
i
t
e
d+d N ~ +
Glycine recognition site l
Cell membrane
1 I
Mg++l
K+
'' I
Ion channel
Fg 3 An exdtatwy synapse at which glutamateacts on an NMDA receptor.
67
Richens within the channel to produce non-competitive blockade (8). This is termed the PCP site. Dizocilpine (MK801) is an experimental compound which acts on the site and shows potent anticonvulsant activity, but unfortunately has psychotomimetic adverse effects in man. Compounds have been developed that act as competitive antagonists at NMDA receptor sites and are anticonvulsant in animal models. The most potent are 2amino-5-phosphonoheptanoicacid (AP5) and 2-amino-7phosphonoheptanoic acid (AP7). The latter is effective against photically induced seizures in the baboon. The NMDA receptor has a glycine recognition site on it, in addition to the glutamate recognition site, and it facilitates excitatory transmission. In this way, it is similar to the benzodiazepine site on the GABA receptor complex. These are also polyamine and zinc recognition sites associated with the NMDA receptor. Compounds which influence these sites are of potential value as anticonvulsants.
Clinical applications Attenuation of glutamate mediated excitatory transmission has a powerful anticonvulsant effect in experimental animals. The various ways in which this can be achieved are summarised in Table 2. The synaptic release of glutamate can be reduced by stimulating auto receptors on the nerve terminal, in a similar manner to the autoregulation which occurs at many synapses. Lamotngine appears to act by reducing glutamate release, but whether its action is on autoreceptors or some other site on the presynaptic terminal is not known (9). As mentioned above decreasing the postsynaptic action of glutamate can be achieved either by competitive inhibition at the recognition site or by non-competitive inhibition at the PCP site in the ionophore. The competitive antagonists AP5 and AP7 enter the brain poorly when given systemically because they have a low lipid solubility. However, two experimental compounds, CGP37849 and CGP39551, have been synthesised and are orally active (10). They are undergoing Phase I studies in man. Remacemide is a novel acetamide compound which weakly blocks NMDA receptors in a non-competitive Table II.
manner, but its desglycine metabolite, to which it is transformed in man, is more potent in this respect. It is currently in Phase II studies in epileptic patients. So far, psychotomimetic effects do not seem to have occurred.
New antiepileptic drugs In Table 3 some of the more promising new drugs are !isted, together with their manufacturer and the stage of development reached. Three are already marketed in some European countries, but none in the USA. A brief review of several of the more advanced drugs is given below. Fully referenced reviews of these drugs have been published in Pisani et al. (1 1) Vigabatrin Vigabatrin (12, 13) is a synthetic derivative of the neurotransmitter, GABA. Chemically it is gamma-vinyl GABA, and it acts by irreversible (suicide) inhibition of GABA-T. Although it has a plasma elimination half life of 7-8 hours, its pharmacodynamic action is much longer because new enzyme has to be synthesized to replace that which is irreversibly bound to the drug. The pharmacodynamic half life has not been estimated, but is probably of the order of several days. This means that the drug can be given once or twice daily. From the pharmacokinetic point of view, vigabatrin is almost the ideal antiepileptic drug. Well absorbed, not protein bound, and renally excreted it has linear kinetics which vary little between patients with normal renal function. In the elderly or those with poor renal function an adjustment of the dose may be necessary, but in clinical use it seems to have a wide therapeutic index. Only one interaction has been reported, namely a slight reduction in plasma phenytoin levels, the reason for which is uncertain. In view of its mode of action, plasma level monitoring is unnecessary.
Vyhbin
Phase Ill-IV
hbgice
Phase Ill-IV
M e c h a n m by W h i i Glutamate&mrnisskm attenuated Phase Ill-IV
htrgine cGP37849
cGW1
Phase 111 Phase 111 -I1 Phase 11-111 Phase II phase II Phase II Phase I
New Drugs for Epilepsy Vigabatrin is effective in partial and secondary generalized seizures. In some types of primary generalized seizure it may also be effective but it appears to exacerbate myoclonus in some patients. It has been reported to be particularly beneficial in seizures associated with tuberous sclerosis. There is limited experience so far with vigabatrin monotherapy . Despite early fears about neurotoxicity (microvacuolation of the white matter occurred in the brains of experimental animals) no serious neurological adverse effects have so far been reported. However, psychosis can occasionally occur, both in patients with a past history of this disorder as well as those without. Animal teratological screening indicates that it is unlikely to be teratogenic in man, but insufficient human experience has so far accumulated to c o n f m this. Vigabatrin is licensed in most European countries at the time of writing but establishing its real place in the treatment of epilepsy will take several years more.
Lamotrigine This drug (14-16) is now marketed in the UK but not in Holland at the time of writing. It is a triazine derivative related to a series of antifolate compounds. It was identified as a potential antiepileptic drug in the belief that folate antagonism was a mechanism by which anticonvulsant effects can be achieved but, in fact, lamotrigine has weak antifolate effects. As described above, it is thought to inhibit the release of glutamate. Lamotrigine is well absorbed, has linear kinetics, and is metabolised to a glucuronide conjugate in the liver. This method of inactivation has a large capacity and is little affected by increasing age. However, its elimination is induced by phenytoin, phenobarbitone and carbamazepine, shortening its plasma elimination half life in adults from about 24 hours to an average of 15 hours when these drugs are being administered simultaneously. Sodium valproate has the opposite effect, markedly reducing the hepatic clearance of lamotrigine and lengthening its half life to about 60 hours on average. This latter interaction demands that a lower dose is administered to patients on valproate comedication. In Phase I1 and III studies it has been shown to be effective in partial and secondarily generalized seizures, about 30% of patients with intractable seizures obtaining a greater than 50% reduction in seizure frequency, possibly somewhat less than vigabatrin. However, preliminary observations suggest that it may have a more important role in primary generalized seizures, but the results of controlled studies are awaited. Patients with Lennox Gastaut syndrome have been reported to gain substantial benefit. Like all antiepileptic drugs, lamotrigine causes mild CNS adverse effects such as fatigue, sedation, diplopia and headache. Allergic skin rashes occur in about 2.5% of patients. They are usually mild and rapidly reversible on
stopping the drug, but occasionally more serious rashes have occurred resembling Stevens-Johnson syndrome. The frequency with which rashes are seen is dose related and is higher when lamotrigine is added to valproate, presumably because of a more rapid increase in the plasma level. Preclinical testing suggests that lamotrigine should be free of teratogenic effects.
Oxcarbazepine Oxcarbazepine ( I 7, 18) is an analogue of carbamazepine which is effectively a pro-drug. It is rapidly metabolized, much of it on frst pass through the liver, to IO,11-dihydro10-hydroxy carbazepine (DHC). Plasma levels of the parent drug are very low compared with those of DHC and it is the latter compound which is largely responsible for the therapeutic effect. It has a plasma half life of about 12 hours and can therefore be given twice daily. DHC is mainly conjugated to a glucuronide metabolite and its kinetics are linearly related to the dose of oxcarbazepine. No important interactions have been identified and oxcarbazepine is probably not an hepatic enzyme inducer. The efficacy of oxcarbazepine appears to be similar to that of carbamazepine; comparative studies have shown no difference but no placebo-controlled studies have been undertaken. The dose of oxcarbazepine is about 50% higher than the dose of carbamazepine. A major advantage of oxcarbazepine is that it less often causes allergic reactions than carbamazepine. Only about one half of patients who have had a skin rash when on carbamazepine experience a similar allergic reaction when they are switched to oxcarbazepine. On the other hand, hyponatraemia with high doses of oxcarbazepine appears to be more frequently encountered than with carbamazepine. Oxcarbazepine may therefore be used in patients who are unable to tolerate carbamazepine. It is possible that it may eventually replace the latter. Gabapentin Gabapentin (19, 20) is a structural analogue of GABA which was synthesized during a research programme aimed at developing a GAl3A-mimetic drug which could cross the blood-brain barrier. It demonstrated anticonvulsant activity in animal models but subsequent studies showed that this was not due to an agonist effect on GABA receptors; its mode of action is unclear. Gabapentin is excreted unchanged in the urine and has a plasma elimination half life of about 6 hours; three doses per day are necessary in view of the short half life. No drug interactions have been reported. In add-on studies in patients with resistant seizures, mainly partial with or without secondary generalization, gabapentin has been shown to reduce seizure frequency by 50% or more in 25-35% of patients. The drug appears to be well tolerated. Mild CNS adverse effects have been seen, 69
Richens particularly drowsiness and fatigue. Gabapentin was not licensed in any country at the time of writing.
Felbamate Felbamate (21, 22) is chemically related to meprobamate and has a wide spectrum of antiepileptic activity in experimental animals. In man it is metabolized in the liver and has a plasma elimination half life of 10-20 hours. When added to existing antiepileptic medication it causes an increase in phenytoin levels but, unusually, a fall in carbamazepine levels. Placebo-controlled, add-on trials in the USA have shown a reduction in seizure frequency with felbamate although in one study a significant effect was seen only when seizure frequency was corrected for a reduction in carbamazepine levels. European studies are scheduled to start soon.
References I. 2. 3.
4.
5. 6.
7.
8.
Tiagabine As mentioned earlier in this review, tiagabine (5) is a GABA reuptake inhibitor which showed potent anticonvulsant activity in preclinical testing. Initial placebo-controlled add-on studies are complete and the results were being analysed at the time of writing. The plasma elimination half life varies from 4.5 to 13 hours in healthy volunteers, but its metabolism is inducible and therefore the halflife is shorter in patients receiving enzyme inducing antiepileptic drugs.
9.
Remacemide
13.
Remacemide (23) is a diphenyl-ethyl-acetamidewhich, in man, is converted to a desglycine metabolite. The latter is a non-competitive blocker of NMDA receptor sites and is probably largely responsible for the drugs antiepileptic action. The parent drug has a short half Life, about 4 hours, but the active metabolite is more slowly eliminated with an apparent half life of 12-24 hours. The metabolism of 2 is inducible but the clinical relevance of this be determined in view of the fact that the -. metabolite is active. Initial placebo-controlled add-on studies have been completed; a preliminary analysis shows encouraging results (J Hutchinson, personal communication).
Conclusion A number of promising new drugs for epilepsy have been developed in the last few years, partly as a result of advances in our understanding of excitatory and inhibitory mechanisms in the brain, and partly by the traditional method of screening new compounds with CNS activity. Much work needs to be done to evaluate these drugs in human epilepsy but if some of them prove to be more efficacious, less toxic or both, the therapeutic management of seizures could improve considerably during the next decade.
10.
11. 12.
14.
15.
6.
7.
8.
19.
20. 21.
22.
23.
Matsumato RR. GABA receptors: are cellular differences reflected in function? Brain Res Reviews 1989: 1 4 203-25. Levitan ES, Schofield PR, Burt DR et al. Structural and functional basis for GABA, receptor heterogeneity. Nature 1988: 335: 76-9. Worms P, Depoortere H, Durand A et al. Aminobutyric acid (GABA) receptor stimulation. 1. Neuropharmacological profiles of progabide (SL 76002) and SL 75102 with emphasis on their anticonvulsant spectra. J Pharmacol Exp Therap 1982: 2 2 0 660. Dam M, Gram L, Philbert A et al. Progabide. A controlled trial in partial epilepsy. Epilepsia 1983: 24: 127. Pierce M W , Suzdak PD, Gustavson LE et al. Tiagabine. In: Pisani F et al. eds. Antiepileptic Drugs. Amsterdam: Elsevier 199I . Collingridge GL, Lester RAJ. Excitatory amino acid receptors in the vertebrate central nervous system. 1989 Pharmacological Reviews 40: 143-210. Meldrum BS, Chapman AG, Patel S, Swan J. Competitive NMDA antagonists as drugs. In: Watkins JC el al. eds. The NMDA Receptor. Oxford: Oxford University Press 1989: 207- 16. h a m JA, Martin D, Tomczyk M et al. Neocortical epileptogenesis in vitro: studies with N-methyl-D-aspartate, phencyclidine, sigma and dextromethorophan receptor ligands. J Pharmacol Exper Therapeutics 1989: 248: 320-28. Leach MJ, Marden CM, Miller AA. Pharmacological studies on lamotrigine, a novel potential antiepileptic drug. 11. Neurochemical studies and mechanism of action. Epilepsia 1986: 27: 490. . . Schmutz M, Portet CL, Jeker A et al. The competitive NMDA receptor antagonists CGP 37849 and CGP 39551 are potent, orallyactive anticonvulsants in rodents. Naunyn-Schmiedeberg's Arch Pharmacol 1990: 342: 61-6. Pisani F, Perucca E, Avanzini G, Richens A eds. New Antiepileptic Drugs. Epil Res 1991: S3. Richens A. Potential antiepileptic drugs. Vigabatrin. In: Levy R et al. eds. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 937-46. Mumford JF'. Lewis PJ. Vigabatrin. In: Pisani F et al. New Antiepileptic Drugs. Epil Res: 1991: Suppl 3: 161-8. Gram L. Potential antiepileptic drugs. Lamotrigine. In: Levy R et al. eds. Antiepileptic Drugs 3rd Edn. New York: Raven Press 1989: 947-53. . Richens A. The efficacy and safety of new antiepileptic drugs. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl3: 89-96. Yuen AWC. Lamotrigine. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 1991: 3: 115-24. Dam M, Klosterskov Jensen P. Potential antiepileptic drugs. Oxcarbazepine. In: Levy R et al. eds. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 913-24. Klosterskov Jensen P, Gram L, Schmutz M. Oxcarbazepine. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 199I : Suppl3: 135-40. Schmidt B. Potential antiepileptic drugs. Gabapentin. In: Levy R et al. eds. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 925-35. Foot M, Wallace J. Gabapentin. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl 3: 109-14. Leppik IE, Graves NM. Potential antiepileptic drugs. Felbamate. In: Levy R et al. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 983-90. Duane Sofia R. Kramer L, Perhach JL, Rosenberg A. Felbamate. In: Pisani Fetal. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl3: 103-8. Muir KT, Palmer GC. Remacemide. In: Pisani F. et al. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl 3: 147-52.
Introduction In all the time that Harry Meinardi has been an epileptologist, only three new antiepileptic drugs have been marketed in Holland, if benzodiazepine drugs are excluded. Valproate was given its first trials in the early 1960s and was marketed in France, where it was discovered, in 1967. Although it is now established worldwide as an effective antiepileptic drug in all seizure types, its mode of action has been elusive. The balance of evidence suggests that it increases activity in inhibitory pathways in the brain, but the way in which it achieves this at a molecular level is uncertain. More recently, vigabatrin with a known mode of action, and oxcarbazepine with a less certain mode of action, were introduced. It is undoubtedly valuable to have a new drug with which to treat our patients, but it is better still to have a compound which also teaches us something about synaptic function and its possible derangement in epilepsy. In the last 15 years various substances have been synthesised which modify excitatory and inhibitory amino-acid transmission and we are now seeing the benefits of this research in the clinic. I will deal first with the processes of synaptic inhibition and excitation and then go on to give a brief sketch of new drugs which are known to influence these systems, as well as several more whose mode of action is as yet uncertain.
Neurotransmitters Neurotransmitters in the central nervous system are divided into two types depending on the speed of the transduction process at the post-synaptic membrane. Fast neurotransmitters give rise to post-synaptic responses in
Address: A. Richens Department of Pharmacologyand Therapeutics Heath Park Cardiff CF4 4XN U.K.
the millisecond range, while slow neurotransmitters operate over seconds or minutes. These responses may be mediated either through ionic mechanisms (ionotropic) or through a second messenger system (rnetabotropic). Two major fast neurotransmitters in the brain are glutamate for excitatory responses and gamma-amino butyric acid (GABA) for inhibition. Signalling in most synapses in the central nervous system is by one or other of these transmitters. A small proportion of synapses, however, use slow neurotransmitters such as monoamines and neuropeptides which have a neuromodulatory role. While these may play some part in epilepsy, particularly accounting for the decrease in convulsive threshold following administration of psychoactive drugs, it seems likely that the major underlying cause is a defect in the transmission at those synapses using glutamate or GABA. It has to be said, however, that direct evidence from man for such a defect is still flimsy. Changes in receptor numbers and neurotransmitter concentrations in brain tissue taken from epileptic foci may be non-specific effects of brain damage or drug treatment rather than the underlying cause of the epilepsy.
GABA as an inhibitory neurotransmitter GABA has received much attention over the last 25 years and evidence has accumulated that some of the conventional antiepileptic drugs interact with inhibitory synaptic mechanisms using this substance as a transmitter. It is estimated that it is involved in 30 % of all synapses in the central nervous system. There are two types of inhibitory mechanism, presynaptic and postsynaptic. In the former, GABA acts on a presynaptic terminal of an excitatory neurone to prevent release of transmitter; this form of inhibition is found predominantly in the spinal cord. Postsynaptic inhibition is the main mechanism found in the brain, and it is at this site that many antiepileptic drugs have their action. When occupied by the endogenous ligand, GABA receptors invariably produce hyperpolarization of the postsynaptic neurone. There are two types of receptor, GABA, and GABA, (,). The former is blocked by a drug
Richens Activation of GABA, receptors opens a chloride channel in the postsynaptic membrane, allowing an influx of negatively charged chloride ions which leads to hyperpolarization of the neurone and consequently an inhibition of f i g . This is termed ionotropic transmission because it involves ionic fluxes. GABA is synthesised in presynaptic nerve terminals from glucose entering the tricarboxylic acid cycle (Figure 2). The transamination of -ketoglutamate generates Lglutamic acid (glutamate) the immediate precursor of GABA. This takes place in a closed loop called the ‘GABA shunt’ which is confined to the nerve terminal. GABA is formed by decarboxylation of L-glutamic acid by the enzyme glutamic acid decarboxylase (GAD), which is found only in neurones and is probably the rate-limiting step in the synthesis of GABA. The next reaction in the shunt is the transamination of GABA to succinic semialdehyde by GABA- transaminase (GABA-T), and at the same time a molecule of glutamate is synthesised for every molecule of GABA destroyed.
l GABA recognition
I
BenzodiazeDines
-
K . .
Chloride channnel
1
fig 1. An inhkhry synapse at which GABA is releasedto prcducepost-sppkinhibrbon at a M A Areceptor.
used in experimental work, bicucculine, and is the receptor which is involved in the treatment of epilepsy (Figure 1). It has recently been isolated by molecular cloning techniques (2).
Oxaloacetate Acetyl
Clinical applications Enhancement of GABA mediated inhibition has an
,
anticonvulsant effect. The various ways in which this can be achieved are summarised in Table 1. GABA itself does not cross the blood-brain barrier and therefore in order to stimulate GABA receptors directly by systemic administration, it has to be linked to a lipid soluble moiety that facilitates penetration into the brain. Progabide was found to be effective in animals (3) but is less satisfactory in man (4). Alternatively, GABA agonist
Glut imate GAD
OC-Ketoglutarate Tr icarboxyIic acid cycle
GABA GABA-T
Citrate
Succinic semi-aldehyde
New Drugs for Epilepsy
Mechanism
Ep?limental
onrgsfor
CompOUnds
human use
GPB,agonism
MusdmlTHlP
Progab&
Bemodiarepine receptor agmism
Benzcd@nes
Direct action at C1-ionophore
EaMJrates
phemone
Reuptake inhibmon
Nipemtic add
liqabine
W - T inhibmon
EthmhmiW suiphate
Vgababln
Benzodiine
dws
molecules can be synthesised which mimic the actions of GABA. Muscimol and THIP are such compounds but unfortunately they do not seem to be active in human epilepsy. The reason for this is unclear. Another possible way to augment GABA inhibition is by using drugs which bind to various sites on the receptor complex and enhance the effect of synaptically released GABA. Theoretically, this type of pharmacological action should preserve the necessary spatial and temporal linkage to suppress abnormal neuronal output but have little effect on normal neuronal function. Benzodiazepines are known to act on specific benzodiazepine binding sites located on the GABA, receptor complex which facilitate GABA transmission and increase the frequency of the chloride channel opening. Barbiturates, on the other hand, act at a separate site associated with the channel, and prolong the open time. Inhibition of GABA reuptake, particularly into glial cells will enhance the inhibitory effect of the transmitter. Nipecotic acid does this following intracerebral injection into animals. Tiagabine is a compound comprising nipecotic acid linked to a lipophilic anchor, designed to allow its penetration into the brain following systemic administration. It is undergoing Phase II clinical testing (5). Inhibition of GABA-T, the enzyme that transamhates GABA, will cause an increase in cerebral levels of the transmitter, thereby producing an anticonvulsant effect. Vigabatrin is now an established drug in Europe used for treating partial and secondarily generalised seizures. It is a suicide inhibitor of the enzyme and has a long duration of action; recovery from its effects requires the synthesis of a new enzyme.
Glutamate as an excitatory neurotransmitter Besides being the immediate precursor for the synthesis of GABA, glutamate is the predominant excitatory amino-
acid neurotransmitter in the brain and spinal cord. Its postsynaptic action is always excitatory, mediated through either an ionotropic or metabotropic action. Stimulation of glutamate receptors causes depolarisation of the postsynaptic membrane with usually a fast onset and termination of effect. Four subtypes of post-synaptic receptor have been described, named after the agonists which preferentially bind to them: NMDA, kainate, quisqualate and AMPA (6). It appears that the NMDA (n-methyl-D-aspartate) receptor is the most likely candidate for a role in epilepsy because agonists for this receptor have been shown to be effective anticonvulsants in animals (7) (Figure 3). Like the GABA receptor complex, it has multiple binding sites which modulate the affinity of the endogenous ligand for its receptor. The Nh4DA complex surrounds a cation channel which allows the influx of sodium and calcium ions and the efflux of potassium ions. Within the channel there is a binding site for magnesium ions which at resting potential block the channel. Depolarisation of the postsynaptic neurone, by activation of quisqualate or AMPA receptors, is necessary before the magnesium block is removed, a so-called 'voltage dependent' block. The dissociative anaesthetics, ketamine and phencyclidine, have been shown to bind to another site
Presynaptic excitory nerve terminals containing glutamate in discrete vesicles
Glutamate recognition s
i
t
e
d+d N ~ +
Glycine recognition site l
Cell membrane
1 I
Mg++l
K+
'' I
Ion channel
Fg 3 An exdtatwy synapse at which glutamateacts on an NMDA receptor.
67
Richens within the channel to produce non-competitive blockade (8). This is termed the PCP site. Dizocilpine (MK801) is an experimental compound which acts on the site and shows potent anticonvulsant activity, but unfortunately has psychotomimetic adverse effects in man. Compounds have been developed that act as competitive antagonists at NMDA receptor sites and are anticonvulsant in animal models. The most potent are 2amino-5-phosphonoheptanoicacid (AP5) and 2-amino-7phosphonoheptanoic acid (AP7). The latter is effective against photically induced seizures in the baboon. The NMDA receptor has a glycine recognition site on it, in addition to the glutamate recognition site, and it facilitates excitatory transmission. In this way, it is similar to the benzodiazepine site on the GABA receptor complex. These are also polyamine and zinc recognition sites associated with the NMDA receptor. Compounds which influence these sites are of potential value as anticonvulsants.
Clinical applications Attenuation of glutamate mediated excitatory transmission has a powerful anticonvulsant effect in experimental animals. The various ways in which this can be achieved are summarised in Table 2. The synaptic release of glutamate can be reduced by stimulating auto receptors on the nerve terminal, in a similar manner to the autoregulation which occurs at many synapses. Lamotngine appears to act by reducing glutamate release, but whether its action is on autoreceptors or some other site on the presynaptic terminal is not known (9). As mentioned above decreasing the postsynaptic action of glutamate can be achieved either by competitive inhibition at the recognition site or by non-competitive inhibition at the PCP site in the ionophore. The competitive antagonists AP5 and AP7 enter the brain poorly when given systemically because they have a low lipid solubility. However, two experimental compounds, CGP37849 and CGP39551, have been synthesised and are orally active (10). They are undergoing Phase I studies in man. Remacemide is a novel acetamide compound which weakly blocks NMDA receptors in a non-competitive Table II.
manner, but its desglycine metabolite, to which it is transformed in man, is more potent in this respect. It is currently in Phase II studies in epileptic patients. So far, psychotomimetic effects do not seem to have occurred.
New antiepileptic drugs In Table 3 some of the more promising new drugs are !isted, together with their manufacturer and the stage of development reached. Three are already marketed in some European countries, but none in the USA. A brief review of several of the more advanced drugs is given below. Fully referenced reviews of these drugs have been published in Pisani et al. (1 1) Vigabatrin Vigabatrin (12, 13) is a synthetic derivative of the neurotransmitter, GABA. Chemically it is gamma-vinyl GABA, and it acts by irreversible (suicide) inhibition of GABA-T. Although it has a plasma elimination half life of 7-8 hours, its pharmacodynamic action is much longer because new enzyme has to be synthesized to replace that which is irreversibly bound to the drug. The pharmacodynamic half life has not been estimated, but is probably of the order of several days. This means that the drug can be given once or twice daily. From the pharmacokinetic point of view, vigabatrin is almost the ideal antiepileptic drug. Well absorbed, not protein bound, and renally excreted it has linear kinetics which vary little between patients with normal renal function. In the elderly or those with poor renal function an adjustment of the dose may be necessary, but in clinical use it seems to have a wide therapeutic index. Only one interaction has been reported, namely a slight reduction in plasma phenytoin levels, the reason for which is uncertain. In view of its mode of action, plasma level monitoring is unnecessary.
Vyhbin
Phase Ill-IV
hbgice
Phase Ill-IV
M e c h a n m by W h i i Glutamate&mrnisskm attenuated Phase Ill-IV
htrgine cGP37849
cGW1
Phase 111 Phase 111 -I1 Phase 11-111 Phase II phase II Phase II Phase I
New Drugs for Epilepsy Vigabatrin is effective in partial and secondary generalized seizures. In some types of primary generalized seizure it may also be effective but it appears to exacerbate myoclonus in some patients. It has been reported to be particularly beneficial in seizures associated with tuberous sclerosis. There is limited experience so far with vigabatrin monotherapy . Despite early fears about neurotoxicity (microvacuolation of the white matter occurred in the brains of experimental animals) no serious neurological adverse effects have so far been reported. However, psychosis can occasionally occur, both in patients with a past history of this disorder as well as those without. Animal teratological screening indicates that it is unlikely to be teratogenic in man, but insufficient human experience has so far accumulated to c o n f m this. Vigabatrin is licensed in most European countries at the time of writing but establishing its real place in the treatment of epilepsy will take several years more.
Lamotrigine This drug (14-16) is now marketed in the UK but not in Holland at the time of writing. It is a triazine derivative related to a series of antifolate compounds. It was identified as a potential antiepileptic drug in the belief that folate antagonism was a mechanism by which anticonvulsant effects can be achieved but, in fact, lamotrigine has weak antifolate effects. As described above, it is thought to inhibit the release of glutamate. Lamotrigine is well absorbed, has linear kinetics, and is metabolised to a glucuronide conjugate in the liver. This method of inactivation has a large capacity and is little affected by increasing age. However, its elimination is induced by phenytoin, phenobarbitone and carbamazepine, shortening its plasma elimination half life in adults from about 24 hours to an average of 15 hours when these drugs are being administered simultaneously. Sodium valproate has the opposite effect, markedly reducing the hepatic clearance of lamotrigine and lengthening its half life to about 60 hours on average. This latter interaction demands that a lower dose is administered to patients on valproate comedication. In Phase I1 and III studies it has been shown to be effective in partial and secondarily generalized seizures, about 30% of patients with intractable seizures obtaining a greater than 50% reduction in seizure frequency, possibly somewhat less than vigabatrin. However, preliminary observations suggest that it may have a more important role in primary generalized seizures, but the results of controlled studies are awaited. Patients with Lennox Gastaut syndrome have been reported to gain substantial benefit. Like all antiepileptic drugs, lamotrigine causes mild CNS adverse effects such as fatigue, sedation, diplopia and headache. Allergic skin rashes occur in about 2.5% of patients. They are usually mild and rapidly reversible on
stopping the drug, but occasionally more serious rashes have occurred resembling Stevens-Johnson syndrome. The frequency with which rashes are seen is dose related and is higher when lamotrigine is added to valproate, presumably because of a more rapid increase in the plasma level. Preclinical testing suggests that lamotrigine should be free of teratogenic effects.
Oxcarbazepine Oxcarbazepine ( I 7, 18) is an analogue of carbamazepine which is effectively a pro-drug. It is rapidly metabolized, much of it on frst pass through the liver, to IO,11-dihydro10-hydroxy carbazepine (DHC). Plasma levels of the parent drug are very low compared with those of DHC and it is the latter compound which is largely responsible for the therapeutic effect. It has a plasma half life of about 12 hours and can therefore be given twice daily. DHC is mainly conjugated to a glucuronide metabolite and its kinetics are linearly related to the dose of oxcarbazepine. No important interactions have been identified and oxcarbazepine is probably not an hepatic enzyme inducer. The efficacy of oxcarbazepine appears to be similar to that of carbamazepine; comparative studies have shown no difference but no placebo-controlled studies have been undertaken. The dose of oxcarbazepine is about 50% higher than the dose of carbamazepine. A major advantage of oxcarbazepine is that it less often causes allergic reactions than carbamazepine. Only about one half of patients who have had a skin rash when on carbamazepine experience a similar allergic reaction when they are switched to oxcarbazepine. On the other hand, hyponatraemia with high doses of oxcarbazepine appears to be more frequently encountered than with carbamazepine. Oxcarbazepine may therefore be used in patients who are unable to tolerate carbamazepine. It is possible that it may eventually replace the latter. Gabapentin Gabapentin (19, 20) is a structural analogue of GABA which was synthesized during a research programme aimed at developing a GAl3A-mimetic drug which could cross the blood-brain barrier. It demonstrated anticonvulsant activity in animal models but subsequent studies showed that this was not due to an agonist effect on GABA receptors; its mode of action is unclear. Gabapentin is excreted unchanged in the urine and has a plasma elimination half life of about 6 hours; three doses per day are necessary in view of the short half life. No drug interactions have been reported. In add-on studies in patients with resistant seizures, mainly partial with or without secondary generalization, gabapentin has been shown to reduce seizure frequency by 50% or more in 25-35% of patients. The drug appears to be well tolerated. Mild CNS adverse effects have been seen, 69
Richens particularly drowsiness and fatigue. Gabapentin was not licensed in any country at the time of writing.
Felbamate Felbamate (21, 22) is chemically related to meprobamate and has a wide spectrum of antiepileptic activity in experimental animals. In man it is metabolized in the liver and has a plasma elimination half life of 10-20 hours. When added to existing antiepileptic medication it causes an increase in phenytoin levels but, unusually, a fall in carbamazepine levels. Placebo-controlled, add-on trials in the USA have shown a reduction in seizure frequency with felbamate although in one study a significant effect was seen only when seizure frequency was corrected for a reduction in carbamazepine levels. European studies are scheduled to start soon.
References I. 2. 3.
4.
5. 6.
7.
8.
Tiagabine As mentioned earlier in this review, tiagabine (5) is a GABA reuptake inhibitor which showed potent anticonvulsant activity in preclinical testing. Initial placebo-controlled add-on studies are complete and the results were being analysed at the time of writing. The plasma elimination half life varies from 4.5 to 13 hours in healthy volunteers, but its metabolism is inducible and therefore the halflife is shorter in patients receiving enzyme inducing antiepileptic drugs.
9.
Remacemide
13.
Remacemide (23) is a diphenyl-ethyl-acetamidewhich, in man, is converted to a desglycine metabolite. The latter is a non-competitive blocker of NMDA receptor sites and is probably largely responsible for the drugs antiepileptic action. The parent drug has a short half Life, about 4 hours, but the active metabolite is more slowly eliminated with an apparent half life of 12-24 hours. The metabolism of 2 is inducible but the clinical relevance of this be determined in view of the fact that the -. metabolite is active. Initial placebo-controlled add-on studies have been completed; a preliminary analysis shows encouraging results (J Hutchinson, personal communication).
Conclusion A number of promising new drugs for epilepsy have been developed in the last few years, partly as a result of advances in our understanding of excitatory and inhibitory mechanisms in the brain, and partly by the traditional method of screening new compounds with CNS activity. Much work needs to be done to evaluate these drugs in human epilepsy but if some of them prove to be more efficacious, less toxic or both, the therapeutic management of seizures could improve considerably during the next decade.
10.
11. 12.
14.
15.
6.
7.
8.
19.
20. 21.
22.
23.
Matsumato RR. GABA receptors: are cellular differences reflected in function? Brain Res Reviews 1989: 1 4 203-25. Levitan ES, Schofield PR, Burt DR et al. Structural and functional basis for GABA, receptor heterogeneity. Nature 1988: 335: 76-9. Worms P, Depoortere H, Durand A et al. Aminobutyric acid (GABA) receptor stimulation. 1. Neuropharmacological profiles of progabide (SL 76002) and SL 75102 with emphasis on their anticonvulsant spectra. J Pharmacol Exp Therap 1982: 2 2 0 660. Dam M, Gram L, Philbert A et al. Progabide. A controlled trial in partial epilepsy. Epilepsia 1983: 24: 127. Pierce M W , Suzdak PD, Gustavson LE et al. Tiagabine. In: Pisani F et al. eds. Antiepileptic Drugs. Amsterdam: Elsevier 199I . Collingridge GL, Lester RAJ. Excitatory amino acid receptors in the vertebrate central nervous system. 1989 Pharmacological Reviews 40: 143-210. Meldrum BS, Chapman AG, Patel S, Swan J. Competitive NMDA antagonists as drugs. In: Watkins JC el al. eds. The NMDA Receptor. Oxford: Oxford University Press 1989: 207- 16. h a m JA, Martin D, Tomczyk M et al. Neocortical epileptogenesis in vitro: studies with N-methyl-D-aspartate, phencyclidine, sigma and dextromethorophan receptor ligands. J Pharmacol Exper Therapeutics 1989: 248: 320-28. Leach MJ, Marden CM, Miller AA. Pharmacological studies on lamotrigine, a novel potential antiepileptic drug. 11. Neurochemical studies and mechanism of action. Epilepsia 1986: 27: 490. . . Schmutz M, Portet CL, Jeker A et al. The competitive NMDA receptor antagonists CGP 37849 and CGP 39551 are potent, orallyactive anticonvulsants in rodents. Naunyn-Schmiedeberg's Arch Pharmacol 1990: 342: 61-6. Pisani F, Perucca E, Avanzini G, Richens A eds. New Antiepileptic Drugs. Epil Res 1991: S3. Richens A. Potential antiepileptic drugs. Vigabatrin. In: Levy R et al. eds. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 937-46. Mumford JF'. Lewis PJ. Vigabatrin. In: Pisani F et al. New Antiepileptic Drugs. Epil Res: 1991: Suppl 3: 161-8. Gram L. Potential antiepileptic drugs. Lamotrigine. In: Levy R et al. eds. Antiepileptic Drugs 3rd Edn. New York: Raven Press 1989: 947-53. . Richens A. The efficacy and safety of new antiepileptic drugs. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl3: 89-96. Yuen AWC. Lamotrigine. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 1991: 3: 115-24. Dam M, Klosterskov Jensen P. Potential antiepileptic drugs. Oxcarbazepine. In: Levy R et al. eds. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 913-24. Klosterskov Jensen P, Gram L, Schmutz M. Oxcarbazepine. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 199I : Suppl3: 135-40. Schmidt B. Potential antiepileptic drugs. Gabapentin. In: Levy R et al. eds. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 925-35. Foot M, Wallace J. Gabapentin. In: Pisani F et al. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl 3: 109-14. Leppik IE, Graves NM. Potential antiepileptic drugs. Felbamate. In: Levy R et al. Antiepileptic Drugs. 3rd Edn. New York: Raven Press 1989: 983-90. Duane Sofia R. Kramer L, Perhach JL, Rosenberg A. Felbamate. In: Pisani Fetal. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl3: 103-8. Muir KT, Palmer GC. Remacemide. In: Pisani F. et al. eds. New Antiepileptic Drugs. Epil Res 1991: Suppl 3: 147-52.
