EPILEPSY OCTET New antiepileptic agents: strategies for drug development ROGER J.
After the clinical observations of Locock in the 1850s1 that documented the usefulness of bromides, and the equally astute conclusions of Hauptmann2 in 1912 that resulted in phenobarbitone, the experimental method has played a key part in the discovery of new antiepileptic drugs. Merritt and Putnam in 19373 were the first to examine systematically a series of compounds (they used an electroshock model in cats), and this research led to the discovery of phenytoin. Various manipulations of the cyclic ureide moiety from which phenytoin and phenobarbitone are derived (fig 1) over the next two decades provided additional barbiturates and hydantoins as well as diones and succinimides. Simple modifications of the molecule—eg, the rearrangement from phenytoin to ethosuximide-not only radically altered the efficacy for control of seizures but also dramatically changed the specificity of the clinical spectrum of the molecular
Fig 1-Chemical structures of phenobarbitone (A); phenytoin (B); trimethadione (C); and ethosuximide (D).4 1950s two rodent models of epilepsy were refined-the maximal electroshock model (useful for detecting drugs effective against partial and generalised tonic-clonic seizures in man) and the subcutaneous pentylenetetrazol (’Metrazol’) model (useful for detecting drugs effective against absence seizures).5 Although these models are still used today and may represent the final common pathway for many kinds of epileptic seizures, these tests tell us little about mechanisms of drug action and may not detect new and novel antiepileptic compounds.
During the early
not fared well in clinical studies in other countries. Vigabatrin (gamma-vinyl GABA), an irreversible inhibitor of the degradative enzyme GABA aminotransferase (fig 2) is potentially more successful. Nevertheless, this drug presents a special toxicological dilemma which I shall discuss later. Other approaches to the enhancement of GABA activity include GABA agonists and GABA uptake inhibitors.6
which, although marketed in France, has
Fig 2-Similarity of vigabatrin (left) and GABA (right). Vigabatrin acts as a "suicidal" substrate for the enzyme aminotransferase which degrades GABA.’2
More recently attention has turned to excitatory compounds, specifically aminoacids such as glutamate and aspartate. Whereas enhancement of GABA activity was the object of investigations of inhibitory processes, decrement of the abnormal activity of glutamate has been the focus of studies of excitatory processes. Initially, MK-801 was one of the most exciting compounds available for clinical study. It performed well in the maximal electroshock modeF and there was also a highly rational basis for its mode of action: it is a non-competitive antagonist of the N-methyl-Daspartate (NMDA) receptor. Unfortunately, the drug did not prove useful in clinical studies; it caused confusion and was not especially effective at tolerated doses. Yet another model of epilepsy-the kindling model-has been used to predict the ineffectiveness of MK-801. Kindling is a state of heightened seizure susceptibility produced by repeated electrical stimulation of a part (often the amygdala) of the animal brain. In studies of the kindled rat, McNamara and colleagues8 showed that MK-801 retarded the development of kindling but that it was not very useful in preventing seizures in fully kindled animals. These data suggest that the drug would not be useful in epileptic human beings, whose abnormal hyperexcitability has already developed. However, this study may have a profound influence on other approaches to searching for new drugs, in that the kindling model appears to distinguish drugs that affect the development of epilepsy-epileptogenesis-from those that prevent seizures in animals that are already epileptic.9 The best preclinical predictive tests for antiepileptic activity remain to be determined. Encouragingly, there are many more compounds available for clinical testing than there were only a decade ago; the Antiepileptic Drug Development Program of the Epilepsy Branch of the National Institute of Neurological Disorders and Strokes may have been partly responsible for this expansion.
Clinical Newer approaches to basic
In the past two decades several more rational approaches have been developed that may provide a higher yield of more effective and more targeted drugs for epilepsy. Recognising that the epileptic discharge is fundamentally one of neuronal hyperexcitability and that gamma-aminobutyric acid (GABA) is a prominent inhibitory neurotransmitter, investigators initially sought compounds that might enhance the activity of GABA and thereby decrease abnormal excitation and control seizures. Progabide was the earliest serious contender; it is essentially a GABA prodrug
powerful combination of complex pharmacological techniques, avant-garde statistical methods, and bias control has made the transition from pilot studies to definitive controlled clinical trials a natural investigative pathway. Epilepsy is not a disease but a heterogeneous symptom complex. to Thus the most important first step in the study of epilepsy is to recognise that the various seizure types and epilepsy syndromes are often strikingly different from one another, and that testing of any sort must begin with a patient population that has certain homogeneous characteristics. Generally, a study will be limited to patients with complex
partial seizures; as we become better at delineating the most important characteristics of these groups, our skill in defining populations for study will improve. To evaluate antiseizure drugs, one must have seizures to count-it is usually the decrement in seizure frequency caused by the drug that makes it valuable. Typically, therefore, patients with frequent seizures are studied so that valid statistical data can be obtained. Unfortunately, most such patients are already partly dependent on other drugs, which often cannot be discontinued for the study. These patients also have the most refractory epilepsies, and present a special, difficult hurdle for any new potential agent. Consequently, add-on studies are common and results are often difficult to interpret. Rigorous bias control is an absolute requirement for controlled trials in epilepsy. Since epileptic seizures are not under the patient’s control and since they appear to be easily seen and counted, one might suspect that placebo effects and investigator bias would not complicate adequate evaluation of new therapies-nothing could be further from the truth. As was well documented in a negative study of cinromide (now abandoned), the placebo effect may be extraordinary. 11 Most controlled clinical trials are now designed to take account of such difficulties, but this was not the case just a few years ago. Several promising compounds are under investigation. In addition to progabide and vigabatrin, potential drugs at various stages of clinical development include: clobazam,
denzimol, eterobarb, felbamate, flunarazine, gabapentin, lamotrigine, milacemide, oxcarbazepine, ralitoline, stiripentol, topiramate, and zonisamide.
Neurotoxicity and drug development One of the most vexing problems in the development of new drugs for central nervous system diseases arises from the fundamental and self-evident need for all such agents to enter the brain and to be "neuroactive." Because our experience with neuroactive drugs is limited, data on neurotoxicity are incomplete. As we cleverly devise new neuroactive drugs, we therefore encounter novel forms of neurotoxicity, some of which will undoubtedly prove to be benign neuropathological curiosities whereas others may be genuinely troublesome. The real difficulty is trying to decide which new finding is toxic and which is not. The fate of many important and potentially efficacious drugs hangs in the balance. Nowhere is this thesis more evident than with vigabatrin, which has some effectiveness in both the maximal electroshock model and newer models. Its proposed mechanism of action is to increase brain GABA at synaptic sites; it also increases GABA levels in brain and spinal fluid. Early pilot clinical studies were conducted in the USA and in Europe; results were encouraging. Unfortunately, small lesions were discovered in the brains of animals undergoing chronic toxicity testing. These lesions, which were vacuolar and appeared to be oedema of the myelin, were reversible but not species specific. Moreover, they appeared at doses not extraordinarily distant from those proposed for human beings with epilepsy. Are these events benign and of little consequence, and in any case unlikely to occur in man, or are
they of long-term importance
with potentially devastating those who may be sentenced to a lifetime of ingestion of antiepileptic drugs? Perhaps the truth lies somewhere in between? Curiously, the diverse regulatory bodies throughout the
in the process of permitting an answer to these questions. In July, 1983, the US Food and Drug Administration stopped all testing of vigabatrin in the USA, with the exception that refractory patients who seemed to respond and were already taking the drug could continue. As a result the number of patients in the USA has dwindled to 20-30. However, in many European countries the alarm was not sounded. The manufacturers were allowed to continue with clinical testing but they were admonished to continue animal studies and to attempt to provide a mechanism for toxicity detection in man. Unfortunately, no really adequate measure of potential human toxicity has been developed; despite various claims made in studies of evoked potentials, it is not certain that these techniques detect the vacuolar lesions. So, should patients with severe epilepsy be placed at risk-level of risk unknown-in the hope that control of their seizures might be improved with the new drug? There might be long-term toxic effects of vigabatrin, which could even remain undetectable until irreparable brain damage had occurred. Conversely, severe epilepsy itself is a devastating and occasionally fatal disorder. More than 2000 vigabatrin-treated patients, mostly in Europe, have been studied and there is no evidence of toxicity from myelinic oedema. Moreover, in the limited necropsy series provided by fate and by temporal lobectomy (for surgical treatment of epilepsy), no lesions have been found. Thus there is mounting evidence that this particular novel toxic finding may well be of little importance. Those who are conservative in their recommendations for study of vigabatrin would simply say that we have been lucky; the drug turned out to be safe. Those who are more permissive, in the laudable name of helping people with epilepsy, would say that the risks were not so great. What seems abundantly clear is that this sequence of events will often be repeated as we develop new neuroactive drugs with new and novel neurotoxicities. world
paper by E. H. Sieveking: analysis of 52 cases of epilepsy observed by the author. Lancet 1857; i: 527. 2. Hauptmann A. Luminal bei Epilepsie. Munch Med Wochenschr 1912; 59: 1. Locock C. Discussion of
1907-09. 3. Merritt HH, Putnam TJ. A new series of anticonvulsant drugs tested by experiments on animals. Arch Neurol Psychiatry 1938; 39: 1003-15. 4. Porter RJ. New antiepileptic drugs. In: Pedley TA, Meldrum BS, eds. Recent advances in epilepsy 4. Edinburgh: Churchill Livingstone, 1988: 161-79. 5. Porter RJ, Cereghino JJ, Gladding GD, et al. Antiepileptic drug development program. Cleveland Clin Quart 1984; 51: 293-305. 6. Meldrum BS. Pharmacological approaches to the treatment of epilepsy. In: Meldrum BS, Porter RJ, eds. New anticonvulsant drugs. London: Libbey, 1986: 17-30. 7. Clineschmidt BV, Martin GE, Bunting PR. Anticonvulsant activity of
(+)- 5-Methyl-10, 11-dihydro-5H-dibenzo[a,d] cyclohepten-5, 10imine (MK-801), a substance with potent anticonvulsant, central
anxiolytic properties. Drug Dev Res
1982; 2: 123-34. 8. McNamara JO, Russell RD, Rigsbee L, Bonhaus DW. Anticonvulsant 9.
and antiepileptic actions of MK-801 in the kindling and electroshock models. Neuropharmacology 1988; 27: 563-68. McNamara JO. Development of new pharmacologic agents for epilepsy: Lessons from the kindling model. Epilepsia 1989; 30 (suppl 1): S13-S18. Porter RJ. Epilepsy: 100 elementary principles. 2nd ed. London: WB Saunders, 1989. The Group for the Evaluation of Cinromide in the Lennox-Gastaut Syndrome. Double-blind, placebo-controlled evaluation of cinromide in patients with the Lennox-Gastaut syndrome. Epilepsia 1989; 30: 422-29. Shechter PJ. Vigabatrin. In: Meldrum BS, Porter RJ, eds. New anticonvulsant drugs. London: Libbey, 1986: 265-75.