55
Epilepsy Res., 10 (1991) 55-61 Elsevier EPIRES 10109
9. Excitotoxicity and epileptic brain damage
Brian Meldrum Department of Neurology, Instituteof Psychiatry,London (IJ. K.)
The two forms of epileptic brain damage, that found in patients with chronic epilepsy (post-mortem or in an anterior temporal lobectomy specimen) and that occurring acutely after status epilepticus, have much in common but are not identical. Hippocampal lesions ouxrring acutely after status epilepticus show a high degree of selectivity for hilar intemeurones, CA1 pyramidal neurones and CA3 pyramidal neurones. Hippocampal lesions in anterior temporal lobectomy specimens tend to involve the subfields less selectively with CA1 being only slightly more severely affected than dentate granule cells, CA3 and CA2 pyramidal neurones. The most severely damaged hippocampi may result from a combination of acute damage early in life (commonly from prolonged febrile convulsions) and cumulative damage associated with seizures. Less severe degrees of damage are probably a consequence of repeated seizures. The abnormal patterns of bring associated with epileptic activity are almost certainly responsible for cell death occurring acutely after status epilepticus; they may contribute to the progressive cell loss occmring in chronic epilepsy.
INTRODUCTION The concept of ‘epileptic brain damage’ is ambiguous. Traditionally it has been taken to refer to the selective pattern of damage (i.e., hippocampal sclerosis, cerebellar cortical atrophy, cerebral cortical atrophy) found in patients with chronic epilepsy dying in institutions”. Experimentalists, however, use the phrase to describe the selective pattern of damage encountered acutely after sustained seizures in animals and relate this to patterns of damage found in children or adults dying shortly after an episode of status epilepticus. These two pathologies may or may not be related. For example many authors have considered that the selective pathology in patients with chronic epilepsy arises from events early in life (in the
perinatal period or in infancy) and is a cause of epilepsy. These early events have been supposed by some authors to be of an anoxic/ischaemic nature and by others to be principally prolonged febrile convulsions. (Viral infections provide a third possibility.) The similarity in the pattern of damage was assumed to be attributable to the fact that cerebral anoxia/ischaemia contributed to the pathology after status epilepticus. In 1973, however, it became clear that damage following status epilepticus related to the excessive neuronal activity itself and not to systemic factor?. Progressively evidence has accumulated that so-called ‘excitotoxic’ mechanisms could account for the acute damage after status epilepticus. The possibility arises that this is not only an explanatory principle for acute epileptic brain damage but might provide
0920-1211/91/$03.50 0 1991 Elsevier Science Publishers B.V. All rights reserved
56
the required link between acute epileptic brain damage and the pathology found in patients with intractable seizures. The latter could resemble acute epileptic brain damage if it is excitotoxic regardless of whether it relates to (i) a febrile convulsion (ii) repeated brief seizures or (iii) events of an anoxitiischaemic kind or a viral infection early in life. We shall review evidence relating to the excitotoxic nature of ‘epileptic brain damage’ and consider how far ‘excitotoxicity’ can explain the traditional problems of epileptic brain damage. SELECTIVE DAMAGE AFTER STATUS EPILEPTICUS: IS EX~ITOT~~C~~ A FULL EXPLANATION? (a) Selectivepattern: regional The pattern of selective damage on a regional
basis shows an ~eq~vocal relationship with seizure activity. Thus for a wide range of seizure models damage is seen in those areas shown by electrophysiological or metabolic (CBF, oxygen or glucose consumption, and induction of immediate early genes) studies. For example in rodents experimental focal cortical seizures can induce thalamic damage7 and limbic seizures produce damage in a wide range of nuclei involved in limbic seizure activity (amygdala, medio-dorsal thalamus, olfactory bulb, globus pallidus and substantia nigra, pars reticulata). It is notable that in children cerebellar atrophy following prolonged hemiconvulsions (as in the HHE syndrome) involves the contralateral cerebellar hemisphere, i.e., depends on local electrical activity, not on vascular factors. This correlation is of course consistent with an ‘excitotoxic’ process but is hardly evidence for it in that metabolic disturbances and alterations in calcium homeostasis could be the key factors linking seizure activity and cell death. Focussing attention on the role of excitatory ne~o~~~~ers could be unhelpful to our understanding of the pathophysiology. (b) Selective pattern: cell types and glutamate receptors
Does the gumption that selective cell loss after status epilepticus is excitotoxic help to explain the pattern of selective cell loss in terms of the cell
types involved? One possibility would be a correlation between vulnerability and density and subtype of glutamate receptors. Most anatomical studies of glutamate receptor density and distribution have employed radioligands and autoradiography of frozen sections, This shows very dramatic differences in the distribution of NMDA and high affinity kainate receptors, but great similarity in the distribution of NMDA and AMPAfquisqualate receptors8926.Thus in the neocortex NMDA receptors are present in highest density in the outer two laminae, whereas high affinity kainate receptors are predomin~t in laminae 5 and 6. In the hip~mpus high affinity kainate receptors are extremely dense around the mossy fibre terminals (CA3 zona lucida) and are present in the inner third of the dentate molecular layer. In contrast NMDA receptors are absent in CA3 stratum lucidum but are very dense in the basal and apical dendritic fields of CA1 (strata oriens and radiatum). AMPA receptors like the NMDA receptors are in very high density over the CA1 dendritic layers, but are also dense over the CA1 pyramidal cell bodies. NMDA receptors show low density in the cerebellar molecular layer; physiological studies suggest that they are on the basket cells and are absent from Purkinje cells. Thus there are no specific correlations between the densities of glutamate receptor subtypes and selective neuronal loss (either following status epilepticus or following chronic epilepsy). However the selective pathology following kainate administration in experimental animals and domoate in people consuming blue mussels does show some relation to the ~s~bution of high a~nity kainate receptors. Thus in rodents following systemic or intracerebroventricular administration of kainate the CA3 region of the hippocampus is preferentially damaged (and administration of anticonvulsauts tends to protect remote regions-cortex and amygdala-but not the CA3 pyramidal neurons) . In patients dying after domoate intoxication neuronal loss occurs in the hippocampus (CA3 and CAl) and in the cortex laminae 5 and 6 (but not in 3 which is predominantly affected after status epilepticus) su~es~g a correlation with kainate receptor density 45. In the rat the globus pallidus and substantia nigra pars reticulata have exceptionally
57
low densities of NMDA receptor. Nevertheless these regions have a low threshold in the rodent for damage following either general&d seizures induced by flurothyl or mercaptopropio~c acid35937 or limbic seizures induced by pilocarpine. They are however resistant to damage in man, usually remaining intact even in patients dying with widespread damage in cortex, thalamus and even put~e~~uda~. (c) Excitotoxic cellular pathology The cytopathological changes occurring in vul-
nerable hippocampal or cortical neurones are closely similar to the changes induced by glutamate, kainate or other classical excitotoxic agents. Thus 30-90 min after 1-2 h of seizure activity induced by bicuculline, allylglycine or kainate pyramidal cells in the CA1 and CA3 hippocampal fields initially show focal dendritic swellings containing distended mitochondria and increased free calcium, nuclei with condensed chromatin and perikarya with dilated Golgi apparatus and endoplasmic reticulum and swollen mitochondria, progressing to a dense condensation of the cytoplasm with multiple va~olationt21i7. Similar appearances can be induced in hilar intemeurones and CA3 pyramidal neurons by sustained electrical stimulation of the perforant path4. These acute appearances are wholly consistent with an excitotoxic mech~ism being responsible for the selective loss of hilar and pyramidal neurons in the hippocampus following generalized or limbic status epilepticus. In the substantia nigra, pars reticulata, the acute cytopathological changes are not typical of excitotoxicity in that there is axonal swelling and a non-selective necrosis of neurones and glia, with the characteristics of an infarct35946. Nevertheless there is evidence that there is a sig nificant excitotoxic element in such damage. Thus unilateral lesioning of the frontal cortex (which destroys some of the ~ut~atergic afferents to the substantia nigra) protects against the nigral damage induced by sustained flurothyl-induced seizures20. (d) Release of gluta~~
or other ~citotoxi~
It is natural to assume that the dramatically enhanced electrical discharges that characterise sei-
zure activity are associated with a dramatic increase in synaptic release of glutamate. In practice however it is not easy to demonstrate this. The technique of in vivo microdialysis is well adapted for this task. However microdidlysis studies in Goteborg found minimal or no increases in extracellular glutamate concentration with either generalized status epilepticus or focal limbic seizuresZs. In collaboration with Drs. T. Obrenovitch and M. Millan we have been further investigating glutamate release using microdialysis probes in the hippocampus (CAl) with on-line analysis of glutamate in the dialysate. We fail to find sig~cant increases in extracelhtlar glutamate concentration when seizures are induced either with focal injections of bicuculline into the prepyriform cortex or with systemic injection of picrotoxin. It is possible that seizures are associated with both enhanced synaptic release of glutamate and enhanced reuptake of glutamate. The addition of a potent glutamate uptake blocker (dihydrokainate) to the perfusion release (see Fig. 1). Could there be a more significant release of other endogenous neurotr~smitte~ or excitotoxins? In experimental seizures Lehmann et al.25 found no change in extracellular aspartate. In three out of four patients with push-pull cannulae and recording electrodes inserted into a hippocampal focus increased aspartate levels were seen in relation to epileptic discharges”. Such damages were however transient (l-3 min duration). Increases in glutamate or in sulphur containing amino acids (cysteine sulphinic acid, homocysteine sulphinic acid, homocysteic acid, cysteic acid) were not seen. (e) Protection by excitatoryamino acid antagonists The most convincing evidence that excitotoxic
mechanisms are involved in ischaemic brain damage has been provided by experiments showing a cerebroprotective action of drugs that block the excitatory action of glutamatezs. Somewhat similar data are available for acute epileptic brain damage, NMDA antagonists can protect against the acute neuronal pathology. This protection is seen in thalamic neurons in rats with sustained focal motor seizures induced by pial application of bicuculline6. It is also seen for hippocampal dam-
DC 16
.@ .f&a
38.
6%.
Tine
98.
128
Cminute)
Fig. 1. ~nti~uo~ recording of DC potential, electroencephalogram, and extracellular glutamate concentration in the rat hippocampus, showing the effect of 100 mM KC (for 10 mm from arrow) and 500 @f ~by~ok~ate (continuous from DHKA arrow) applied via the perfusate and of 10 nmol of bicu~ne (injected focally into the prepays cortex). Glutamate is measured by a method employing ~u~mate dehy~ogenase and ~uo~rne~c detection of NADH. Whereas Kt produces a lo-fotd increase in extracellular ghttamate concentration, and uptake inhibition produces a 3-4-fold increase, sustained seixure activity (in the presence of uptake inhibition) does not significantly increase ghttamate concentration. (Unpublished observations ofM. Millan, T. Obrenovitch and B.S. Meldrum, see Millan et al., Epikpsy Res., 9 (1991) 86-91.)
age (in CA1 but not CA3) following seizures inIt might be argued that duced by k&ate 15*22,23. NMDA antagonists are powerful anticonvulsantss*X*31 and that reduction in the pathologi~l consequences of prolonged seizures is merely the ~on~quen~ of reducing the intensity of the local seizure activity, in much the same way as diazepam protects against remote pathology in limbic seizures induced by kainate3. In both the focal motor and the limbic seizure studies cited macroelectrode EEG recordings were presented to show that seizure activity was not reduced (indeed Fariello et all5 reported it to be enhan~d~. However studies in in vitro slice preparations show that NMDA antagonists commonly fail to reduce the frequency of burst firing but decrease the late components of a burst, decreasing in particular the elements comprising the Ca2* spikes. Thus the evi-
dence from NMDA antagonists can be inte~reted as supporting both the concept of excitotoxicity and also the direct interpretation of a link between epileptic activity and cell damage. rf Selective pattern: cdcium bkffering capacity We have seen that the concept of ‘excitoto~city’ may not offer any particular advantages as an explanatory framework for selective damage after prolonged seizures. It may be more appropriate to focus on failure of Ca2” homeostasis. In recent accounts of excitotoxieity an increase of intracellular [Ca2’] isusually considered a critical event in determining cell deathBe An increase in intracellular [Ca”‘] (consequent upon burst firing) has long been considered to be the determining link between seizure activity and selective cell 10s~‘~~~~. Selective eel1 death might thus be determined
59
EPILEPTIC BRAIN DAMAGE: EXCITOTOXICITY AND THE VARIOUS POTENTIAL PRIMARY MECHANISMS
have mesial temporal sclerosis as the primary pathology the proportion with a history of febrile convulsions is 50-75%“921. There is a strong correlation between a severe degree of cell loss in the dentate gyms, the endfolium and CA1 and prolonged or lateralised seizures in childhood3g. Milder degrees of cell loss are associated with a long history of seizures, sometimes with another primary pathology. Comparing hippocampal neuronal densities in 10 controls with 5 tumour associated epilepsies and 25 non-tumour epilepsies, Kim et a121found significant reductions in all hippocampal fields (including dentate granule cells and CA2 pyramidal cells) in the non-tumour epilepsies but not in the tumour epilepsies (the tumour epilepsies showed non-significant reductions in all fields relative to controls; the mean duration of epilepsy was 12.7 years in the tumour cases, 24.1 years in the non-tumour cases).
(a) Prolonged febrile convulsions
(6) Repeated seizures and progressive cell loss
Children dying after prolonged febrile convulsions show essentially the same selective pattern of pathology as adults, involving CA1 pyramidal neurons and hilar intemeurons in the hippocampus, small neurons in the third cortical lamina and Purkinje cells in the cerebellum 47.It is reasonable to presume that the mechanisms involved are the same as in the adult. There may be a greater sensitivity to brain damage in the child compared with the adult, as the minimal duration for the occurrence of neurological sequelae appears to be 30 min in the child and 2 h in the adult’. This difference might be attributable to the exacerbating effect of hyperpyrexia, but it should be remembered that in adults with status epilepticus hyperpyrexia is common and is highly correlated with an unfa-W vourable neurological outcome’. The particular interest of this aetiological mechanism is the link provided with the ‘Ammon’s horn sclerosis’ found in the majority of anterior temporal lobes removed surgically from patients with intractable seizures originating in one temporal lobe. Ounsted et al.% and subsequently Falconerr4, drew attention to the high proportion of patients with complex partial seizures with a history of prolonged febrile convulsions in childhood. For patients undergoing anterior temporal lobectomy and found to
Experimental evidence that brief repeated seizures can cause cell loss in the hippocampus is less convincing than the evidence that status epilepticus causes hippocampal, cortical, thalamic and cerebellar damage. Nevertheless brief, closely repeated seizures in. baboons cause a restricted cell loss in CA1 and CA333.Kindled seizures in the rat may also be associated with a progressive loss of hilar neurons4. Quantitative studies of hippocampal cell loss in man are loss in man are most readily explained by a progressive non-selective loss of neurons associated with chronic seizures. Most studies show a correlation between duration of epilepsy and severity of hippocampal cell ~oss~*~,~‘~~~. Where the seizures are attributable to a glioma or hamartoma there is always a lesser degree of cell loss.
either by the capacity of the neuron to show sustained burst firing or by its calcium buffering capacity. Either or both these features may explain why in the hippocampus dentate granule cells are relatively invulnerable to acute cell death. Immunocytochemical studies4i show that calbindin is in high concentration in dentate granule cells but not in vulnerable hilar intemeurons or mossy cells. The possible importance of intrinsic calcium buffering in determining neuronal survival is sup ported by experiments in hippocampal slice#‘. Stimulation of the perforant path produced a rapid functional decline in mossy cells that was not seen when the intracellular recording electrodes were filled with the calcium chelator, BAPTA.
(c) Perinatalanoxialischaemia
In both rodents and man the distribution and density of the glutamate receptor subtypes is very different in the newborn and the adult (glutamate receptors in the newborn are principally of the NMDA subtype and are in highest density in the globus pallidus). The functional effect of activation of the receptors appears somewhat different in the infant. Birth injury, perinatal hypoxia/is-
60
chaemia or seizures in the neonatal period do not lead to a selective pattern of brain damage comparable to that seen after febrile convulsions in childhood or status epilepticus in adults. Infants coming to post mortem show haemorrhages (gross and micro) and diffuse damage involving the whole hemisphere, especially the cortex and basal ganglia. It is therefore likely that, in patients with complex partial seizures who show mesial temporal sclerosis and give a history of birth injury or perinatal asphyxia, cortical damage secondary to the birth injury facilitates seizures, which in turn induces the mesial temporal sclerosis.
REFERENCES 1 Aminoff, M.J. and Simon, R., Status epilepticus, causes, clinical features and consequences in 98 patients, Am. J. Med., 69 (1980) 657-666. 2 Babb, T.L., Research on the anatomy and pathology of epileptic tissue. IO: H. Luders (Ed.), Neurosurgery ofEpilepsy, Cleveland, OH, 1991, pp. 719-727. 3 Ben&i, Y., Tremblay, E., Ottersen, O.P. and Meldrum, B.S., The role of epileptic activity in hippocampal and ‘remote’ cerebral lesions induced by kainic acid, Brain Res., 191(1980) 79-97. 4 Cavazos, J.E. and Sutula, T.P., Progressive neuronal loss induced by kindling: a possible mechanism for mossy fiber synaptic reorganization and hippocampal sclerosis, Brain Res., 527 (1990) 1-6. 5 Chqamn, A.G. and Meldrum, B.S., Non-competitive Nmethyl+aspartate antagonists protect against sound-induced seizures in DBAl2 mice, Eur. J. Phamacol., 166 (1989) 201-211. 6 Clifford, D.B., Zorumski, C.F. and Olney, J.W., Ketamine and MK-801 prevent degeneration of thalamic neurons induced by focal cortical seizures, Exp. Neural., 105 (1989) 272-279. 7 Collins, R.C. and Olney, J.W., Focal distant thalamic lesions, Science, 218 (1982) 177-179. 8 Cotman, C.W., Monaghan, D.T., Ottersen, J. and StormMathisen, J., Anatomical organization of excitatory amino acid receptors and their pathways, Trends Neurosci., 10 (1987) 273-280. 9 Dam, A.M., Epilepsy and neuron loss in the hippocampus, Epilepsia, 21(1980) 617-629. 10 Do, K.Q., Klancnik, J., Gahwiler, B., Pershak, H., Wieser, H.G. and Cuenod, M., Release of EAA: animal studies and epileptic foci studies in human. In: B.S. Meldrum, F. Moroni, R.P. Simon and J.H. Woods (Eds.), Excitatory Amino Acids, Raven Press, New York, NY, 1991, pp. 677-685. 11 Duncan, J.S. and Sagar, H.J., Seizure characteristics, pa-
(d) Viral infections
Viral infections are usually associated with activation of tryptophan 2,3dioxygenase and enhanced synthesis of quinolinic acid. In AIDS dementia the CSF concentration of quinolinic acid can be increased 40-fold19. Viral infections can also be associated with the occurrence of selective hippocampal pathology. The question arises therefore whether quinolinic acid could, through an excitotoxic action, contribute to the hippocampal pathology occurring in viral encephalitis.
thology and outcome after temporal lobectomy, Neurology, 37 (1987) 405-409. 12 Evans, M., Griffiths, T. and Meldrum, B.S., Early changes in the rat hippocampus following seizures induced by bicuculline or L-aliyiglycine: a light and electron microscope study, Neuropathol. Appl. Neurobiol., 9 (1983) 39-52. 13 Evans, M.C., Griffiths, T. and Meldrum, B.S., Kainic acid seizures and the reversibility of calcium loading in tinerable neurons in the hippocampus, Neuropathol. Appl. Neurobiol., 9 (1984) 30-52. 14 Falconer, M.A., Genetic and related aetiological factors in temporal lobe epilepsy: a review, Epilepsia, 12 (1971) 13-31. 15 Fariello, R.G., Golden, G.T., Smith, G.G. and Reyes, P.F., Potentiation of kainic acid epileptogenicity and sparing from neuronal damage by an NMDA receptor antagonist, Epilepsy Res., 3 (1989) 206-213. 16 Gtiffiths, T., Evans, M.C. and Meldrum, B.S., Status epilepticus: the reversibility of calcium loading and acute neuronal pathological changes in the rat hippocampus, Neuroscience, 12 (1984) 557-567. 17 Griffiths, T., Evans, M.C. and Me&urn, B.S., Intracellular calcium accumulation in rat hippocampus during seizures induced by bicuculline or L-allylglycine, Neuroscience, 10 (1983) 385-395. 18 Heyes, M.P., Wyler, A.R., Devinsky, O., Yergey, J.A., Markey, S.P. and Nadi, N.S., Quinolinic acid concentrations in brain and cerebrospinal fluid of patients with intractable complex partial seizures, Epilepsia, 31 (1990) 172-177. 19 Heyes, M.P., Rubinow, D., Lane, C. and Markey, S.P., Cerebrospinal fluid quinolinic acid concentrations are increased in acquired immune deficiency syndrome, Ann. Neurol., 26 (1989) 275-277. 20 Ingvar, M., Seizure-induced damage in the substantia nigra pars reticulata: lesions in the frontal cortex prior to the seizure period mitigate the damage, Exp. Brain Res., 75 (1989) 369-374. 21 Kim, J.H., Guimaraes, P.G., Shen, M.Y., Masukawa, L.M. and Spencer, D.D., Hippocampal neuronal density in
61
22
23
24
25
26
27 28
29
30
31
32
33
34
temporal lobe epilepsy with and without ghomas, Acta Neuropathol., 80 (1990) 41-45. Labruyere, .I., Fuller, T.A., Olney, J.W., Price, M.T., Zorumsky, C. and Clifford, D., PCP and ketamine protect against KA-induced seizures and seizure related brain damage, Sot. Neurosci. Absr*., 12 (1986) 344-344. Lason, W., Simpson, J.N. and McGinty, J.F., Effects of p(-)-aminophosphonovalerate on behavioral and histological changes induced by systemic kainic acid, Neurosci. Len., 87 (1988) 23-28. Lehmann, A., Alterations in hippocampal extracelhdar amino acids and purine catabolites during Iimbic seizures induced by folate injections into the rabbit amygdala, Neuroscience, 22 (1987) 573-578. Lehmann, A., Hagberg, H., Jacobson, I. and Hamberger, A., Effect of status epilepticus on extracelhdar amino acids in the hippocampus, Brain Res., 359 (1985) 147-151. Maragos, W.F., Penney, J.B. and Young, A.B., Anatomic correlation of NMDA and 3H-TCP-labeled receptors in rat brain, J. Neurosci., 8 (1988) 493-501. Margerison, J.H. and Corsellis, J.A.N., Epilepsy and the temporal lobes, Brain, 89 (1966) 499-530. Meldrum, B.S., Protection against ischaemic neuronal damage by drugs acting on excitatory neurotransmission, Cereb. Brain Metabol. Rev., 2 (1990) 27-57. Meldrum, B.S. and Garthwaite, J., Excitatory amino acid and neurodegenerative disease, Trends Pharmacol. Sci., 11 (1990) 379-387. Meldrum, B.S., Chapman, A.G., Patel, S. and Swan, J.H., Competitive NMDA antagonists as drugs. In: J.C. Watkins and G.L. Collingridge (Eds.), The NMDA Receptor, IRL Press, Oxford, 1989, pp. 207-216. Meldrum, B.S., Croucher, M.J., Badman, G. and Collins, J.F., Antiepileptic action of excitatory amino acid antagonists in the photosensitive baboon, Papio papio, Neurosci. Len., 39 (1983) 101-104. Meldrum, B.S., Metabolic effects of prolonged epileptic seizures and the causation of epileptic brain damage. In: F.C. Rose (Ed,), Metabolic Disorders of the Nervous System, Pitman, London, 1981, pp. 175-187. Meldrum, B.S., Horton, R. W. and Brierley, J.B., Epileptic brain damage in adolescent baboons following seizures induced by ahylglycine, Bruin, 97 (1974) 417-428. Meldrum, B.S., Vigouroux, R.A. and Brierley, J.B., Systemic factors and epileptic brain damage. Prolonged seixures in paralysed artificially ventilated baboons, Arch. Neural., 29 (1973) 82-87.
35 Nevander, G., Ingvar, M., Auer, R. and Siesjo, B.K., Status epilepticus in well-oxygenated rats causes neuronal necrosis, Ann. Neural., 18 (1985) 281-290. 36 Norman, R.M., The neuropathology of status epilepticus, Med. Sci. Law, 4 (1964) 46-51. 37 O’Connell, B.K., Towfighi, J., Kofke, W.A. and Hawkins, R.A., Neuronai lesions in mercaptopropionic acid-induced status epilepticus, Actu Neuropathol., 77 (1988) 47-54. 38 Ounsted, C., Lindsay, J. and Norman, R., Biological Factors in Temporal Lobe Epilepsy, W. Heinemann, London, 1966, pp. l-135. 39 Sagar, H.J. and Oxbury, J.M., Hippocampal neuron loss in temporal lobe epilepsy: correlation with early childhood convulsions, Ann. Neural., 22 (1987) 334-340. 40 Scharfman, H.E. and Schwartxkroin, P.A., Protection of dentate hilar cells from prolonged stimulation by intracellular calcium chelation, Science, 246 (1989) 257-260. 41 Sloviter, R.S., Chemically defined hippocampal intemeurolls and their possible relationship to seizure mechanisms. In: Tome (Ed.), The Hippocampus - New Visms, Alan R. Liss, New York, NY, 1989, pp. 443-461. 42 Sloviter, R.S., Calcium-binding protein (CalbindinD28KO) and parvalbumin immunocytochemistry: locahxation in the rat hippocampus with specific reference to the selective vulnerability of hippocampal neurons to seixure activity, 1. Camp. Neural., 280 (1989)183-l%. 43 Sloviter, R.S. and Dempster, D.W., ‘Epileptic’ brain damage is replicated qualitatively in the rat hippocampus by central injection of glutamate or aspartate but not by GABA or acetylchohne, Brain Res. Bull., 15 (1985) 39-60. 44 Sloviter, R.S., ‘Epileptic’ brain damage in rats induced by sustained electrical stimulation of the perforant path. I. Acute electrophysiological and light microscopic studies, Brain Res. Bull., 10 (1983) 675-697. 45 Teitelbaum, J.S., Zatorre, R.J., Carpenter, S., Gendron, D., Evans, A.C., Gjedde, A. and Cashman, N.R., Neurologic sequelae of domoic acid intoxication due to the ingestion of contaminated mussels, New Engl. J. Med., 322 (1990) 1781-1787. 46 Towfighi, J., Kofke, W.A., O’Connell, B.K., Housman, C. and Graybeal, JM., Substantia nigra lesions in mercaptopropionic acid induced status epilepticus: a light and electron microscopic study, Acta Neuropathol,, 77 (1989) 612-620. 47 Zimmerman, H.M., The histopathology of convulsive disorders in children, J. Pediutr., 13 (1938) 859-890.
Epilepsy Res., 10 (1991) 55-61 Elsevier EPIRES 10109
9. Excitotoxicity and epileptic brain damage
Brian Meldrum Department of Neurology, Instituteof Psychiatry,London (IJ. K.)
The two forms of epileptic brain damage, that found in patients with chronic epilepsy (post-mortem or in an anterior temporal lobectomy specimen) and that occurring acutely after status epilepticus, have much in common but are not identical. Hippocampal lesions ouxrring acutely after status epilepticus show a high degree of selectivity for hilar intemeurones, CA1 pyramidal neurones and CA3 pyramidal neurones. Hippocampal lesions in anterior temporal lobectomy specimens tend to involve the subfields less selectively with CA1 being only slightly more severely affected than dentate granule cells, CA3 and CA2 pyramidal neurones. The most severely damaged hippocampi may result from a combination of acute damage early in life (commonly from prolonged febrile convulsions) and cumulative damage associated with seizures. Less severe degrees of damage are probably a consequence of repeated seizures. The abnormal patterns of bring associated with epileptic activity are almost certainly responsible for cell death occurring acutely after status epilepticus; they may contribute to the progressive cell loss occmring in chronic epilepsy.
INTRODUCTION The concept of ‘epileptic brain damage’ is ambiguous. Traditionally it has been taken to refer to the selective pattern of damage (i.e., hippocampal sclerosis, cerebellar cortical atrophy, cerebral cortical atrophy) found in patients with chronic epilepsy dying in institutions”. Experimentalists, however, use the phrase to describe the selective pattern of damage encountered acutely after sustained seizures in animals and relate this to patterns of damage found in children or adults dying shortly after an episode of status epilepticus. These two pathologies may or may not be related. For example many authors have considered that the selective pathology in patients with chronic epilepsy arises from events early in life (in the
perinatal period or in infancy) and is a cause of epilepsy. These early events have been supposed by some authors to be of an anoxic/ischaemic nature and by others to be principally prolonged febrile convulsions. (Viral infections provide a third possibility.) The similarity in the pattern of damage was assumed to be attributable to the fact that cerebral anoxia/ischaemia contributed to the pathology after status epilepticus. In 1973, however, it became clear that damage following status epilepticus related to the excessive neuronal activity itself and not to systemic factor?. Progressively evidence has accumulated that so-called ‘excitotoxic’ mechanisms could account for the acute damage after status epilepticus. The possibility arises that this is not only an explanatory principle for acute epileptic brain damage but might provide
0920-1211/91/$03.50 0 1991 Elsevier Science Publishers B.V. All rights reserved
56
the required link between acute epileptic brain damage and the pathology found in patients with intractable seizures. The latter could resemble acute epileptic brain damage if it is excitotoxic regardless of whether it relates to (i) a febrile convulsion (ii) repeated brief seizures or (iii) events of an anoxitiischaemic kind or a viral infection early in life. We shall review evidence relating to the excitotoxic nature of ‘epileptic brain damage’ and consider how far ‘excitotoxicity’ can explain the traditional problems of epileptic brain damage. SELECTIVE DAMAGE AFTER STATUS EPILEPTICUS: IS EX~ITOT~~C~~ A FULL EXPLANATION? (a) Selectivepattern: regional The pattern of selective damage on a regional
basis shows an ~eq~vocal relationship with seizure activity. Thus for a wide range of seizure models damage is seen in those areas shown by electrophysiological or metabolic (CBF, oxygen or glucose consumption, and induction of immediate early genes) studies. For example in rodents experimental focal cortical seizures can induce thalamic damage7 and limbic seizures produce damage in a wide range of nuclei involved in limbic seizure activity (amygdala, medio-dorsal thalamus, olfactory bulb, globus pallidus and substantia nigra, pars reticulata). It is notable that in children cerebellar atrophy following prolonged hemiconvulsions (as in the HHE syndrome) involves the contralateral cerebellar hemisphere, i.e., depends on local electrical activity, not on vascular factors. This correlation is of course consistent with an ‘excitotoxic’ process but is hardly evidence for it in that metabolic disturbances and alterations in calcium homeostasis could be the key factors linking seizure activity and cell death. Focussing attention on the role of excitatory ne~o~~~~ers could be unhelpful to our understanding of the pathophysiology. (b) Selective pattern: cell types and glutamate receptors
Does the gumption that selective cell loss after status epilepticus is excitotoxic help to explain the pattern of selective cell loss in terms of the cell
types involved? One possibility would be a correlation between vulnerability and density and subtype of glutamate receptors. Most anatomical studies of glutamate receptor density and distribution have employed radioligands and autoradiography of frozen sections, This shows very dramatic differences in the distribution of NMDA and high affinity kainate receptors, but great similarity in the distribution of NMDA and AMPAfquisqualate receptors8926.Thus in the neocortex NMDA receptors are present in highest density in the outer two laminae, whereas high affinity kainate receptors are predomin~t in laminae 5 and 6. In the hip~mpus high affinity kainate receptors are extremely dense around the mossy fibre terminals (CA3 zona lucida) and are present in the inner third of the dentate molecular layer. In contrast NMDA receptors are absent in CA3 stratum lucidum but are very dense in the basal and apical dendritic fields of CA1 (strata oriens and radiatum). AMPA receptors like the NMDA receptors are in very high density over the CA1 dendritic layers, but are also dense over the CA1 pyramidal cell bodies. NMDA receptors show low density in the cerebellar molecular layer; physiological studies suggest that they are on the basket cells and are absent from Purkinje cells. Thus there are no specific correlations between the densities of glutamate receptor subtypes and selective neuronal loss (either following status epilepticus or following chronic epilepsy). However the selective pathology following kainate administration in experimental animals and domoate in people consuming blue mussels does show some relation to the ~s~bution of high a~nity kainate receptors. Thus in rodents following systemic or intracerebroventricular administration of kainate the CA3 region of the hippocampus is preferentially damaged (and administration of anticonvulsauts tends to protect remote regions-cortex and amygdala-but not the CA3 pyramidal neurons) . In patients dying after domoate intoxication neuronal loss occurs in the hippocampus (CA3 and CAl) and in the cortex laminae 5 and 6 (but not in 3 which is predominantly affected after status epilepticus) su~es~g a correlation with kainate receptor density 45. In the rat the globus pallidus and substantia nigra pars reticulata have exceptionally
57
low densities of NMDA receptor. Nevertheless these regions have a low threshold in the rodent for damage following either general&d seizures induced by flurothyl or mercaptopropio~c acid35937 or limbic seizures induced by pilocarpine. They are however resistant to damage in man, usually remaining intact even in patients dying with widespread damage in cortex, thalamus and even put~e~~uda~. (c) Excitotoxic cellular pathology The cytopathological changes occurring in vul-
nerable hippocampal or cortical neurones are closely similar to the changes induced by glutamate, kainate or other classical excitotoxic agents. Thus 30-90 min after 1-2 h of seizure activity induced by bicuculline, allylglycine or kainate pyramidal cells in the CA1 and CA3 hippocampal fields initially show focal dendritic swellings containing distended mitochondria and increased free calcium, nuclei with condensed chromatin and perikarya with dilated Golgi apparatus and endoplasmic reticulum and swollen mitochondria, progressing to a dense condensation of the cytoplasm with multiple va~olationt21i7. Similar appearances can be induced in hilar intemeurones and CA3 pyramidal neurons by sustained electrical stimulation of the perforant path4. These acute appearances are wholly consistent with an excitotoxic mech~ism being responsible for the selective loss of hilar and pyramidal neurons in the hippocampus following generalized or limbic status epilepticus. In the substantia nigra, pars reticulata, the acute cytopathological changes are not typical of excitotoxicity in that there is axonal swelling and a non-selective necrosis of neurones and glia, with the characteristics of an infarct35946. Nevertheless there is evidence that there is a sig nificant excitotoxic element in such damage. Thus unilateral lesioning of the frontal cortex (which destroys some of the ~ut~atergic afferents to the substantia nigra) protects against the nigral damage induced by sustained flurothyl-induced seizures20. (d) Release of gluta~~
or other ~citotoxi~
It is natural to assume that the dramatically enhanced electrical discharges that characterise sei-
zure activity are associated with a dramatic increase in synaptic release of glutamate. In practice however it is not easy to demonstrate this. The technique of in vivo microdialysis is well adapted for this task. However microdidlysis studies in Goteborg found minimal or no increases in extracellular glutamate concentration with either generalized status epilepticus or focal limbic seizuresZs. In collaboration with Drs. T. Obrenovitch and M. Millan we have been further investigating glutamate release using microdialysis probes in the hippocampus (CAl) with on-line analysis of glutamate in the dialysate. We fail to find sig~cant increases in extracelhtlar glutamate concentration when seizures are induced either with focal injections of bicuculline into the prepyriform cortex or with systemic injection of picrotoxin. It is possible that seizures are associated with both enhanced synaptic release of glutamate and enhanced reuptake of glutamate. The addition of a potent glutamate uptake blocker (dihydrokainate) to the perfusion release (see Fig. 1). Could there be a more significant release of other endogenous neurotr~smitte~ or excitotoxins? In experimental seizures Lehmann et al.25 found no change in extracellular aspartate. In three out of four patients with push-pull cannulae and recording electrodes inserted into a hippocampal focus increased aspartate levels were seen in relation to epileptic discharges”. Such damages were however transient (l-3 min duration). Increases in glutamate or in sulphur containing amino acids (cysteine sulphinic acid, homocysteine sulphinic acid, homocysteic acid, cysteic acid) were not seen. (e) Protection by excitatoryamino acid antagonists The most convincing evidence that excitotoxic
mechanisms are involved in ischaemic brain damage has been provided by experiments showing a cerebroprotective action of drugs that block the excitatory action of glutamatezs. Somewhat similar data are available for acute epileptic brain damage, NMDA antagonists can protect against the acute neuronal pathology. This protection is seen in thalamic neurons in rats with sustained focal motor seizures induced by pial application of bicuculline6. It is also seen for hippocampal dam-
DC 16
.@ .f&a
38.
6%.
Tine
98.
128
Cminute)
Fig. 1. ~nti~uo~ recording of DC potential, electroencephalogram, and extracellular glutamate concentration in the rat hippocampus, showing the effect of 100 mM KC (for 10 mm from arrow) and 500 @f ~by~ok~ate (continuous from DHKA arrow) applied via the perfusate and of 10 nmol of bicu~ne (injected focally into the prepays cortex). Glutamate is measured by a method employing ~u~mate dehy~ogenase and ~uo~rne~c detection of NADH. Whereas Kt produces a lo-fotd increase in extracellular ghttamate concentration, and uptake inhibition produces a 3-4-fold increase, sustained seixure activity (in the presence of uptake inhibition) does not significantly increase ghttamate concentration. (Unpublished observations ofM. Millan, T. Obrenovitch and B.S. Meldrum, see Millan et al., Epikpsy Res., 9 (1991) 86-91.)
age (in CA1 but not CA3) following seizures inIt might be argued that duced by k&ate 15*22,23. NMDA antagonists are powerful anticonvulsantss*X*31 and that reduction in the pathologi~l consequences of prolonged seizures is merely the ~on~quen~ of reducing the intensity of the local seizure activity, in much the same way as diazepam protects against remote pathology in limbic seizures induced by kainate3. In both the focal motor and the limbic seizure studies cited macroelectrode EEG recordings were presented to show that seizure activity was not reduced (indeed Fariello et all5 reported it to be enhan~d~. However studies in in vitro slice preparations show that NMDA antagonists commonly fail to reduce the frequency of burst firing but decrease the late components of a burst, decreasing in particular the elements comprising the Ca2* spikes. Thus the evi-
dence from NMDA antagonists can be inte~reted as supporting both the concept of excitotoxicity and also the direct interpretation of a link between epileptic activity and cell damage. rf Selective pattern: cdcium bkffering capacity We have seen that the concept of ‘excitoto~city’ may not offer any particular advantages as an explanatory framework for selective damage after prolonged seizures. It may be more appropriate to focus on failure of Ca2” homeostasis. In recent accounts of excitotoxieity an increase of intracellular [Ca2’] isusually considered a critical event in determining cell deathBe An increase in intracellular [Ca”‘] (consequent upon burst firing) has long been considered to be the determining link between seizure activity and selective cell 10s~‘~~~~. Selective eel1 death might thus be determined
59
EPILEPTIC BRAIN DAMAGE: EXCITOTOXICITY AND THE VARIOUS POTENTIAL PRIMARY MECHANISMS
have mesial temporal sclerosis as the primary pathology the proportion with a history of febrile convulsions is 50-75%“921. There is a strong correlation between a severe degree of cell loss in the dentate gyms, the endfolium and CA1 and prolonged or lateralised seizures in childhood3g. Milder degrees of cell loss are associated with a long history of seizures, sometimes with another primary pathology. Comparing hippocampal neuronal densities in 10 controls with 5 tumour associated epilepsies and 25 non-tumour epilepsies, Kim et a121found significant reductions in all hippocampal fields (including dentate granule cells and CA2 pyramidal cells) in the non-tumour epilepsies but not in the tumour epilepsies (the tumour epilepsies showed non-significant reductions in all fields relative to controls; the mean duration of epilepsy was 12.7 years in the tumour cases, 24.1 years in the non-tumour cases).
(a) Prolonged febrile convulsions
(6) Repeated seizures and progressive cell loss
Children dying after prolonged febrile convulsions show essentially the same selective pattern of pathology as adults, involving CA1 pyramidal neurons and hilar intemeurons in the hippocampus, small neurons in the third cortical lamina and Purkinje cells in the cerebellum 47.It is reasonable to presume that the mechanisms involved are the same as in the adult. There may be a greater sensitivity to brain damage in the child compared with the adult, as the minimal duration for the occurrence of neurological sequelae appears to be 30 min in the child and 2 h in the adult’. This difference might be attributable to the exacerbating effect of hyperpyrexia, but it should be remembered that in adults with status epilepticus hyperpyrexia is common and is highly correlated with an unfa-W vourable neurological outcome’. The particular interest of this aetiological mechanism is the link provided with the ‘Ammon’s horn sclerosis’ found in the majority of anterior temporal lobes removed surgically from patients with intractable seizures originating in one temporal lobe. Ounsted et al.% and subsequently Falconerr4, drew attention to the high proportion of patients with complex partial seizures with a history of prolonged febrile convulsions in childhood. For patients undergoing anterior temporal lobectomy and found to
Experimental evidence that brief repeated seizures can cause cell loss in the hippocampus is less convincing than the evidence that status epilepticus causes hippocampal, cortical, thalamic and cerebellar damage. Nevertheless brief, closely repeated seizures in. baboons cause a restricted cell loss in CA1 and CA333.Kindled seizures in the rat may also be associated with a progressive loss of hilar neurons4. Quantitative studies of hippocampal cell loss in man are loss in man are most readily explained by a progressive non-selective loss of neurons associated with chronic seizures. Most studies show a correlation between duration of epilepsy and severity of hippocampal cell ~oss~*~,~‘~~~. Where the seizures are attributable to a glioma or hamartoma there is always a lesser degree of cell loss.
either by the capacity of the neuron to show sustained burst firing or by its calcium buffering capacity. Either or both these features may explain why in the hippocampus dentate granule cells are relatively invulnerable to acute cell death. Immunocytochemical studies4i show that calbindin is in high concentration in dentate granule cells but not in vulnerable hilar intemeurons or mossy cells. The possible importance of intrinsic calcium buffering in determining neuronal survival is sup ported by experiments in hippocampal slice#‘. Stimulation of the perforant path produced a rapid functional decline in mossy cells that was not seen when the intracellular recording electrodes were filled with the calcium chelator, BAPTA.
(c) Perinatalanoxialischaemia
In both rodents and man the distribution and density of the glutamate receptor subtypes is very different in the newborn and the adult (glutamate receptors in the newborn are principally of the NMDA subtype and are in highest density in the globus pallidus). The functional effect of activation of the receptors appears somewhat different in the infant. Birth injury, perinatal hypoxia/is-
60
chaemia or seizures in the neonatal period do not lead to a selective pattern of brain damage comparable to that seen after febrile convulsions in childhood or status epilepticus in adults. Infants coming to post mortem show haemorrhages (gross and micro) and diffuse damage involving the whole hemisphere, especially the cortex and basal ganglia. It is therefore likely that, in patients with complex partial seizures who show mesial temporal sclerosis and give a history of birth injury or perinatal asphyxia, cortical damage secondary to the birth injury facilitates seizures, which in turn induces the mesial temporal sclerosis.
REFERENCES 1 Aminoff, M.J. and Simon, R., Status epilepticus, causes, clinical features and consequences in 98 patients, Am. J. Med., 69 (1980) 657-666. 2 Babb, T.L., Research on the anatomy and pathology of epileptic tissue. IO: H. Luders (Ed.), Neurosurgery ofEpilepsy, Cleveland, OH, 1991, pp. 719-727. 3 Ben&i, Y., Tremblay, E., Ottersen, O.P. and Meldrum, B.S., The role of epileptic activity in hippocampal and ‘remote’ cerebral lesions induced by kainic acid, Brain Res., 191(1980) 79-97. 4 Cavazos, J.E. and Sutula, T.P., Progressive neuronal loss induced by kindling: a possible mechanism for mossy fiber synaptic reorganization and hippocampal sclerosis, Brain Res., 527 (1990) 1-6. 5 Chqamn, A.G. and Meldrum, B.S., Non-competitive Nmethyl+aspartate antagonists protect against sound-induced seizures in DBAl2 mice, Eur. J. Phamacol., 166 (1989) 201-211. 6 Clifford, D.B., Zorumski, C.F. and Olney, J.W., Ketamine and MK-801 prevent degeneration of thalamic neurons induced by focal cortical seizures, Exp. Neural., 105 (1989) 272-279. 7 Collins, R.C. and Olney, J.W., Focal distant thalamic lesions, Science, 218 (1982) 177-179. 8 Cotman, C.W., Monaghan, D.T., Ottersen, J. and StormMathisen, J., Anatomical organization of excitatory amino acid receptors and their pathways, Trends Neurosci., 10 (1987) 273-280. 9 Dam, A.M., Epilepsy and neuron loss in the hippocampus, Epilepsia, 21(1980) 617-629. 10 Do, K.Q., Klancnik, J., Gahwiler, B., Pershak, H., Wieser, H.G. and Cuenod, M., Release of EAA: animal studies and epileptic foci studies in human. In: B.S. Meldrum, F. Moroni, R.P. Simon and J.H. Woods (Eds.), Excitatory Amino Acids, Raven Press, New York, NY, 1991, pp. 677-685. 11 Duncan, J.S. and Sagar, H.J., Seizure characteristics, pa-
(d) Viral infections
Viral infections are usually associated with activation of tryptophan 2,3dioxygenase and enhanced synthesis of quinolinic acid. In AIDS dementia the CSF concentration of quinolinic acid can be increased 40-fold19. Viral infections can also be associated with the occurrence of selective hippocampal pathology. The question arises therefore whether quinolinic acid could, through an excitotoxic action, contribute to the hippocampal pathology occurring in viral encephalitis.
thology and outcome after temporal lobectomy, Neurology, 37 (1987) 405-409. 12 Evans, M., Griffiths, T. and Meldrum, B.S., Early changes in the rat hippocampus following seizures induced by bicuculline or L-aliyiglycine: a light and electron microscope study, Neuropathol. Appl. Neurobiol., 9 (1983) 39-52. 13 Evans, M.C., Griffiths, T. and Meldrum, B.S., Kainic acid seizures and the reversibility of calcium loading in tinerable neurons in the hippocampus, Neuropathol. Appl. Neurobiol., 9 (1984) 30-52. 14 Falconer, M.A., Genetic and related aetiological factors in temporal lobe epilepsy: a review, Epilepsia, 12 (1971) 13-31. 15 Fariello, R.G., Golden, G.T., Smith, G.G. and Reyes, P.F., Potentiation of kainic acid epileptogenicity and sparing from neuronal damage by an NMDA receptor antagonist, Epilepsy Res., 3 (1989) 206-213. 16 Gtiffiths, T., Evans, M.C. and Meldrum, B.S., Status epilepticus: the reversibility of calcium loading and acute neuronal pathological changes in the rat hippocampus, Neuroscience, 12 (1984) 557-567. 17 Griffiths, T., Evans, M.C. and Me&urn, B.S., Intracellular calcium accumulation in rat hippocampus during seizures induced by bicuculline or L-allylglycine, Neuroscience, 10 (1983) 385-395. 18 Heyes, M.P., Wyler, A.R., Devinsky, O., Yergey, J.A., Markey, S.P. and Nadi, N.S., Quinolinic acid concentrations in brain and cerebrospinal fluid of patients with intractable complex partial seizures, Epilepsia, 31 (1990) 172-177. 19 Heyes, M.P., Rubinow, D., Lane, C. and Markey, S.P., Cerebrospinal fluid quinolinic acid concentrations are increased in acquired immune deficiency syndrome, Ann. Neurol., 26 (1989) 275-277. 20 Ingvar, M., Seizure-induced damage in the substantia nigra pars reticulata: lesions in the frontal cortex prior to the seizure period mitigate the damage, Exp. Brain Res., 75 (1989) 369-374. 21 Kim, J.H., Guimaraes, P.G., Shen, M.Y., Masukawa, L.M. and Spencer, D.D., Hippocampal neuronal density in
61
22
23
24
25
26
27 28
29
30
31
32
33
34
temporal lobe epilepsy with and without ghomas, Acta Neuropathol., 80 (1990) 41-45. Labruyere, .I., Fuller, T.A., Olney, J.W., Price, M.T., Zorumsky, C. and Clifford, D., PCP and ketamine protect against KA-induced seizures and seizure related brain damage, Sot. Neurosci. Absr*., 12 (1986) 344-344. Lason, W., Simpson, J.N. and McGinty, J.F., Effects of p(-)-aminophosphonovalerate on behavioral and histological changes induced by systemic kainic acid, Neurosci. Len., 87 (1988) 23-28. Lehmann, A., Alterations in hippocampal extracelhdar amino acids and purine catabolites during Iimbic seizures induced by folate injections into the rabbit amygdala, Neuroscience, 22 (1987) 573-578. Lehmann, A., Hagberg, H., Jacobson, I. and Hamberger, A., Effect of status epilepticus on extracelhdar amino acids in the hippocampus, Brain Res., 359 (1985) 147-151. Maragos, W.F., Penney, J.B. and Young, A.B., Anatomic correlation of NMDA and 3H-TCP-labeled receptors in rat brain, J. Neurosci., 8 (1988) 493-501. Margerison, J.H. and Corsellis, J.A.N., Epilepsy and the temporal lobes, Brain, 89 (1966) 499-530. Meldrum, B.S., Protection against ischaemic neuronal damage by drugs acting on excitatory neurotransmission, Cereb. Brain Metabol. Rev., 2 (1990) 27-57. Meldrum, B.S. and Garthwaite, J., Excitatory amino acid and neurodegenerative disease, Trends Pharmacol. Sci., 11 (1990) 379-387. Meldrum, B.S., Chapman, A.G., Patel, S. and Swan, J.H., Competitive NMDA antagonists as drugs. In: J.C. Watkins and G.L. Collingridge (Eds.), The NMDA Receptor, IRL Press, Oxford, 1989, pp. 207-216. Meldrum, B.S., Croucher, M.J., Badman, G. and Collins, J.F., Antiepileptic action of excitatory amino acid antagonists in the photosensitive baboon, Papio papio, Neurosci. Len., 39 (1983) 101-104. Meldrum, B.S., Metabolic effects of prolonged epileptic seizures and the causation of epileptic brain damage. In: F.C. Rose (Ed,), Metabolic Disorders of the Nervous System, Pitman, London, 1981, pp. 175-187. Meldrum, B.S., Horton, R. W. and Brierley, J.B., Epileptic brain damage in adolescent baboons following seizures induced by ahylglycine, Bruin, 97 (1974) 417-428. Meldrum, B.S., Vigouroux, R.A. and Brierley, J.B., Systemic factors and epileptic brain damage. Prolonged seixures in paralysed artificially ventilated baboons, Arch. Neural., 29 (1973) 82-87.
35 Nevander, G., Ingvar, M., Auer, R. and Siesjo, B.K., Status epilepticus in well-oxygenated rats causes neuronal necrosis, Ann. Neural., 18 (1985) 281-290. 36 Norman, R.M., The neuropathology of status epilepticus, Med. Sci. Law, 4 (1964) 46-51. 37 O’Connell, B.K., Towfighi, J., Kofke, W.A. and Hawkins, R.A., Neuronai lesions in mercaptopropionic acid-induced status epilepticus, Actu Neuropathol., 77 (1988) 47-54. 38 Ounsted, C., Lindsay, J. and Norman, R., Biological Factors in Temporal Lobe Epilepsy, W. Heinemann, London, 1966, pp. l-135. 39 Sagar, H.J. and Oxbury, J.M., Hippocampal neuron loss in temporal lobe epilepsy: correlation with early childhood convulsions, Ann. Neural., 22 (1987) 334-340. 40 Scharfman, H.E. and Schwartxkroin, P.A., Protection of dentate hilar cells from prolonged stimulation by intracellular calcium chelation, Science, 246 (1989) 257-260. 41 Sloviter, R.S., Chemically defined hippocampal intemeurolls and their possible relationship to seizure mechanisms. In: Tome (Ed.), The Hippocampus - New Visms, Alan R. Liss, New York, NY, 1989, pp. 443-461. 42 Sloviter, R.S., Calcium-binding protein (CalbindinD28KO) and parvalbumin immunocytochemistry: locahxation in the rat hippocampus with specific reference to the selective vulnerability of hippocampal neurons to seixure activity, 1. Camp. Neural., 280 (1989)183-l%. 43 Sloviter, R.S. and Dempster, D.W., ‘Epileptic’ brain damage is replicated qualitatively in the rat hippocampus by central injection of glutamate or aspartate but not by GABA or acetylchohne, Brain Res. Bull., 15 (1985) 39-60. 44 Sloviter, R.S., ‘Epileptic’ brain damage in rats induced by sustained electrical stimulation of the perforant path. I. Acute electrophysiological and light microscopic studies, Brain Res. Bull., 10 (1983) 675-697. 45 Teitelbaum, J.S., Zatorre, R.J., Carpenter, S., Gendron, D., Evans, A.C., Gjedde, A. and Cashman, N.R., Neurologic sequelae of domoic acid intoxication due to the ingestion of contaminated mussels, New Engl. J. Med., 322 (1990) 1781-1787. 46 Towfighi, J., Kofke, W.A., O’Connell, B.K., Housman, C. and Graybeal, JM., Substantia nigra lesions in mercaptopropionic acid induced status epilepticus: a light and electron microscopic study, Acta Neuropathol,, 77 (1989) 612-620. 47 Zimmerman, H.M., The histopathology of convulsive disorders in children, J. Pediutr., 13 (1938) 859-890.
