Brain Research, 495 (1989) 387-395 Elsevier
Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy N.C. de Lanerolle 1, J.H. Kim 2, R.J. R o b b i n s 3 and D . D . Spencer 1 Sections of INeurosurgery, 2Neuropathology and *Neuroendocrinology, Yale University School of Medicine, New Haven, CT 06510 (U.S.A.) (Accepted 2 May 1989) Key words: Hippocampus; Plasticity; Temporal lobe epilepsy; Neuronal loss
It has been hypothesized on the basis of animal models of epilepsy that abnormal neural activity in epilepsy may be related to reorganized neural circuits that facilitate epileptogenesis. Little evidence of this was available for human epilepsy. This paper provides the first evidence of such reorganization of a hippocampal seizure focus in human temporal lobe epilepsy (TLE). This reorganization involves the selective loss of somatostatin and neuropeptide Y immunoreactive interneurons, and axonal sprouting of other neuropeptide Y neurons and dynorphin-A immunoreactive granule cells. This set of changes is not exactly like those that are reported in animal models.
The association of localized brain damage with temporal lobe epilepsy (TLE) was recognized a hundred years ago by Hughlings Jackson ~7,18 and several studies since then have further documented patterns of hippocampal cell loss, referred to as hippocampal sclerosis, in this disease 4'6'23"35. The relationship of this cell loss to epileptogenicity has remained a puzzle. One possibility is that the hippocampal neuron loss results in the increased excitability of remaining hippocampal neurons. Since the excitability of neurons is governed by their participation in neural circuits involving excitatory and inhibitory neuroactive substances (neurotransmitters and neuromodulators), an elucidation of the distribution of neuroactive substances within the seizure focus could help define changes in the neurochemical circuits that may underlie the increased excitability of a seizure focus. Hippocampi removed from patients with TLE provide a unique source of tissue for the examination of chemically defined neural circuits in a seizure focus. We have studied the immunocytochemical localization of several neuroactive substances in such human hippo-
campi, and report here the first evidence of the selective vulnerability of specific subsets of somatostatin (SOM) and neuropeptide Y (NPY) interneurons, along with sprouting of other NPY interneurons and dynorphin (DYN) containing granule cells associated with a hippocampal seizure focus. Patients with medically intractable epilepsy referred to the Yale/West Haven V.A. epilepsy program undergo a comprehensive phased evaluation 36. Two groups of these patients are chosen as candidates for a unilateral en bloc resection of the anterior temporal neocortex and hippocampus. In one group intracranial masses (consisting primarily of low grade gliomas, hamartomas and vascular lesions) are identified in the anterior temporal lobe. In these patients medial temporal lobe structures (including hippocampus) are removed to provide pathologic assurance of negative margins for tumor cells and to maximize seizure control. This group of patients will be referred to as the tumor related temporal lobe epilepsy group and numbered 7 patients in this study. In the other group, no mass was present, but either non-invasive localization or depth electrode
Correspondence: N.C. de Lanerolle, Section of Neurosurgery, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06510, U.S.A.
Fig. 1. Photomicrographs of glutamate-like immunoreactivity in the human hippocampus. (A) Stained dentate granular neurons in TTLE patient showing a densely packed granule cell layer. (B) Dentate granule cell layer in a CTLE patient showing the loss of neurons and loosely packed granule cells. (C) and (F) CAI region from a TTLE and CTLE patient respectively showing immunoreactive pyramidal neurons. Note the extensive loss of pyramidal neurons in (F) and the atrophy of the region when compared to (C). (D) and (E) Area CA4 from TTLE and CTLE patients respectively. There is a significant loss of hilar neurons in (E). (G) CA3 m a TTLE patient. The region appears relatively normal with many immunoreactive pyramidal neurons. Bar = 100 Itm. The tissue for the above study was fixed by immersion in 5% acrolein in phosphate buffer, pH 7,4 for about 4-6 h and immunostaining carried out by the indirect antibody peroxidase-anti peroxidase method. The primary antibody was used at 1:10,000 dilution. Antibody specifity controls were carried out on adjacent sections.
'"~. . . . .
. . . . , . i."7 ; ? ."\ , ~
. . ., . . . . , . ..,
. . .i.
l ~ . .~,.,I, .. :
'~ ', ,',
"llll:. . . .
,~ . . . . . . . .
: ~',."'l." '" ' r ~"" t '"~ ";/'" ~;/" :" •, .,,., "-.::i-t;., ': ,'~. ,. /
~: \ '~ \ "_.. :' ' " l"
...-.... , ......
~' / "
• ?'/ ~;.';/
.: l ,7/ /
.--:" .'.,." ~,-~.~_~ ~.'~.~.~DD /
~ . . , "-. ', ".'.'-". '~ .-' 11 • "; -'.if
. <'~ "..~-~..~:.""
...... i:. ~'. "',
....... . .... I t_'i. " . ~ . , " :.. ""i~.~" ,: -"." .':'~'~, " .... "'.""-.:". ' '. \~. .i~,¢il . .."~.~il
" '.,. -
,," -", i,~, .' ,. ". ~ . \' % " .-.,."., . "" • :
'". . . . . .
".". ...... ;;'-~..-.. . . . . . .-~. "'~( "~i::':,,.:.;:-.-,}I ."~ • ,~>" //; .... ~, .... s,
I " " "\t .
\,, *,V \-,,.',,
I ,, L . " /
""' X":;',. ",,
\ ' ' \ , , , " , . " ~--~--" "_.,""-" , x~'_ ; , "" l I ...,,,:~..;............ ,.,
-..~,,, < , . . , , .~~.".,'.,F/
~','.",.-. . . . . ...'L>V "~.~..,.....1;,,~,v
~ / ' / !
,r t " v,.....,:;~ ~/',, ,...,......-....~..~-" %,,
E Fig. 2. Microscope digitizer plots of the exact distribution of immunoreactive somatostatin (A,C,E) and neuropeptide Y (B,D,F) neurons in representative sections from a TTLF~ (A and B), CTLF~ (C and D) and an atypical CTLE patient (E and F). The sections for somatostatin and neuropeptide Y are from the same patient. The immunoreactive neurons are indicated by dark symbols and the extent of the CA1 region is indicated by two arrows. Panels C and D show the paucity of somatostatin and neuropeptide Y neurons in the CA4 (hilus) region compared to the other groups.
390 and/or subdural electrode recording of electrical activity indicate that 80-100% of seizure activity originates from one hippocampus. This group will be referred to as the cryptogenic temporal lobe epilepsy group (or cryptogenic epilepsy) and consisted of 35 patients. The surgical removal of the appropriate temporal lobe tissue in both groups provides a high rate of seizure control 37. The hippocampal tissue removed by surgery from the above patients was immediately fixed by immersion in a solution consisting of 2% paraformaldehyde and 2% picric acid in phosphate buffer, pH 7.4 (1-2 h), followed by immersion in 5% acrolein in phosphate buffer, pH 7.4 (4-8 hr). Fifty micron coronal sections cut on a Vibratome were used for immunocytochemistry. Glutamyl-glutamate-like immunoreactivity was localized with a mouse monoclonal antibody 21 and the avidin-biotin technique ~6. SOM, NPY, D Y N and G A B A were localized with polyclonal antibodies generated in rabbit and the peroxidase-antiperoxidase (PAP) method 39. Antibody and tissue specificity controls were included. A comparison of neuron numbers in the dentate gyrus and A m m o n ' s Horn areas of the hippocampus in the two groups of epilepsy patients with those of non-epileptic autopsy cases (n = 11) revealed that the hippocampi of the tumor related group were much like that of non-epileptic persons. However, the hippocampi of the cryptogenic epilepsy group showed marked neuron loss in the dentate gyrus and all areas of A m m o n ' s Horn, the greatest loss being in area CA1 j4. This difference between groups is seen most dramatically in tissue stained with a monoclonal antibody directed against glutamyi-glutamate on the carboxy terminal end of tubulin 21 which was localized, almost exclusively in the granule cells, large hilar neurons and pyramidal cells (Fig. 1). Comparison of the 3 groups of hippocampi (tumor-related, cryptogenic and autopsy) revealed a
loss of granule cells, neurons in the subgranular polymorphic region and pyramidal neurons in Ammon's Horn in the cryptogenic epilepsy group (Fig. 1). There was no apparent cell loss in the subiculum or the entorhinal cortex of either the cryptogenic epilepsy or tumor-related epilepsy group. lmmunoreactivity for SOM localized to interneurons of the dentate hilus (CA4), the hippocampus proper, the subiculum and the entorhinal cortex in tumor-related epileptics. In the dentate hilus, the SOM neurons were found in a subgranular position (in the region described as the polymorphic layer in primate brain) as well as more deeply in the hilus. The axons and dendrites of the hilar SOM neurons form a network not only in this hilus, but also extend into the molecular layer of the dentate gyrus. This pattern of distribution is the same as that in non-epileptic controls ~. In the majority (29/35) of cryptogenic epilepsy patients, there were fewer SOM positive neurons in the hilus than in the tumor-related epilepsy group. The subgranular population of SOM neurons in particular was virtually always absent in the cryptogenic epilepsy group (Fig. 2). The pattern of distribution of SOM neurons in other regions of the hippocampus was similar in tumor-related and cryptogenic epilepsy, with an apparently normal distribution of immunoreactive neurons even in areas such as C A I where in the cryptogenic epilepsy group there is a great loss of pyramidal cells. Radioimmunoassay ( R I A ) measurements in the same population of patients were consistent with our immunohistochemistry in showing a significant decrease of SOM immunoreactivity in the dentate gyrus as well as CA4 but not in other regions in cryptogenic epileptics in comparison to tumor-related epileptics and non-epileptic controis 3°. NPY immunoreactive neurons have a distribution similar to that of S O S neurons in tumor-related -...}
Fig. 3. (A) and (B) Dark field photomicrographs of neuropeptide Y immunoreactive fibers in the molecular layer of the dentate gyrus in a TTLE (A) and CTLE (B) patient. The limits of the granule cell layer are depicted by two arrowheads. (C) and (D) are bright field photomicrographs of dynorphin-A like immunoreactivity in the granule cells and hilus of a TTLE and CTLE patient respectively. Note the presence of dark immunoreactivity in the inner part of the molecular layer in (D) but not in (C). The absence of immunoreactivity immediately under the granule cells layer in (D) compared to (C) reflects the loss of polymorphic layer neurons in the former. Such cells are normally innervated by dynorphin immunoreactive terminals as in (C). IM, inner third of molecular layer; OM, outer two thirds of molecular layer. Bar = 200/~m. Immunostaining was carried out on 50/~m Vibratome sections using the somatostatin antibodies mentioned in the text at a dilution of 1:1000 to 1:3000. Antibody specificity controls were carried out on adjacent sections with antisera preabsorbed with pure somatostatin and by omitting the antiserum in the staining procedure.
392 epilepsy and non-epileptic controls (Fig. 2). The axons of CA4 NPY neurons, in addition to forming a dense network in the hilus, also project through the granule cell layer into the molecular layer (Fig. 3). NPY axons are rather sparse in the inner parts of the molecular layer, but more numerous in the outer third. In the majority of cryptogenic epilepsy hippocampi 29"35, the hilar NPY neurons were missing, in particular the population of neurons in the subgranular polymorphic layer (Fig. 2). However, in these hippocampi, there is a very rich network of NPY immunoreactive fibers in the dentate molecular layer which ramify extensively in both the inner and outer parts of the molecular layer (Fig. 3). The appearence of this fiber network strongly suggests 'sprouting' of remaining NPY hilar neurons since these fibers can be seen to extend from the hilus through the granule cells into the molecular layer. A similar network of fibers was not observed for SOM. It is unlikely that a failure to detect such sprouting for SOM is due to technical limitations (fixation or antibody limitations) of immunocytochemistry, for we have failed to see it with 4 different antibodies (one monoclonal: courtesy Dr. A. Malcolm, and 3 polyclonal: Incstar, MN; $320 & $309, courtesy Dr. Benoit). Further, in the same specimens SOM fibers were well stained in the subiculum and neocortex. It thus appears that in cryptogenic epilepsy there is a specific loss of SOM and NPY immunoreactive neurons in the hilus, particularly in the subgranular zone, along with a sprouting of NPY fibers in the molecular layer. Since y-amino butyric acid (GABA) containing interneurons have been reported to be reduced in number in seizure foci of animal models 2~, we localized GABA-Iike immunoreactive neurons in our samples of hippocampal tissue ( G A B A antiserum, lncstar, Stillwater, MN). Microscope digitizer plots, similar to those shown for SOM and NPY, showed no obvious difference in the patterns of cell distribution. In both the timor-related epilepsy and cryptogenic epilepsy group immunoreactive neurons were found along the inner border of the granule cell layer, a few scattered neurons in the hilus and CA3, and many neurons in area CA1, the subiculum, along the perforant path within the hippocampus, and in the entorhinal cortex. In the cryptogenic epilepsy group there were many G A B A neurons in
area CA1 despite almost a total loss of pyramidal cells. In tumor-related epilepsy patients DYN was localized principally in the hippocampal granule cells and their mossy fiber axons which innervate hilar and CA3 neurons. No DYN immunoreactivity was seen in the molecular layer of the dentate. In cryptogenic epilepsy patients, in addition to the pattern seen in the tumor-related epilepsy group, DYN immunoreactivity was also seen in the inner third of the dentate molecular layer (Fig. 3). This extension of DYN immunoreactivity into the molecular layer is consistent with a 'sprouting' of mossy fiber collaterals into the inner third of the molecular layer. Our observations thus reveal several new features in the pathology of human TLE. First, T L E is not invariably associated with hippocampal sclerosis. Sclerosis is clearly evident only in the cryptogenic epilepsy group even though the manifestation of the seizures in the two groups is the same. Variations in hippocampal pathology as assesed by cell numbers in TLE have been reported previously, with a small or no loss of neurons observed in cases with extrahippocampal temporal m a s s e s 4"~'13'23'35. However, the present study is the first to show that on several criteria tumor-related epilepsy hippocampi are more like those from non-epileptic autopsy cases. The similarity of the seizures in the two groups may be the result of the spread of the seizure through a common pathway rather than due to a common mechanism of seizure initiation. Second, these human tissue studies reveal that the patterns of neuron loss and reorganization seen in the cryptogenic epilepsy group may be related to the increased excitability of the tissue. It is notable that in 6/35 cryptogenic epilepsy patients, the SOM, NPY and DYN patterns resembled those in tumor-related epilepsy. In these 6 patients seizure control was not obtained following an anteromedial temporal lobectomy. This would suggest that the patterns of neural reorganization described, with perhaps others to be discovered in cryptogenic epilepsy, may play a significant role in the mechanisms of seizure generation. The primary seizure focus in 5 of these 6 atypical cryptogenic epilepsy patients awaits identification. In one patient of this group (IH63) no mass was detected before he underwent a left anterome-
393 dial temporal lobectomy. Following poor surgical results, the patient was subsequently found to have a small posterolateral tempero-occipital cortical mass upon MRI imaging. The removal of this mass produced seizure control. In the main cryptogenic epilepsy group 29'35 there is a selective loss of two specific populations of interneurons, those containing SOM and those containing NPY, located in the dentate hilus, particularly in the polymorphic zone. Many SOM and NPY neurons survive in other regions of the hippocampus, even in CA1 where there is a profound loss of pyramidal neurons. The loss of these interneurons in the hilus may be of particular significance. The role that hilar SOM and NPY neurons play depends upon the physiological effects of these peptides on the granular neurons, whether or not the NPY sprouts make functional synapses with the granular neurons, and the distribution of SOM and NPY receptors in the epileptic state. There are no studies that address these issues directly, and hence no definitive explanation is currently possible. However, the effects of SOM and NPY on CA1 pyramidal neurons has been studied s'11"15"26. A feature in common to both is their ability to inhibit the CA1 pyramidal neuron. If these peptides have similar effects on granule cells, and serve an inhibitory role, then the loss of these interneurons would presumably reduce the inhibition of granular neurons. Traditionally, neurons utilizing GABA are also thought to be involved in the inhibition of hippocampal neurons 2A9"29. However, in our human tissue no obvious loss of G A B A neurons was discernible in the cryptogenic epilepsy group compared to the tumor-related epilepsy group. Babb and coworkers 3 also report that they observed no loss of glutamic acid decarboxylase (the G A B A synthesizing enzyme) immunoreactive neurons in the hippocampi of their TLE population. This does not necessarily mean that a loss of GABA inhibition may not be involved in the increased excitability of the cryptogenic epilepsy hippocampus. At least a subpopulation of SOM and NPY hippocampal neurons are thought to also contain GABA 7'19, and thus may be lost along with the former. Masukawa et al. 22 have recently observed that upon perforant path stimulation there are abnormal discharges characterized by multiple population spikes in the dentate gyrus of
hippocampal slices from cryptogenic epilepsy patients. They interpret this observation as consistent with a 10-20% decrease in GABAA mediated synaptic inhibition. Thus the role of G A B A in hippocampal excitability in human TLE remains to be further explored. The patterns of sprouting of NPY and DYN fibers, particularly into the inner third of the dentate molecular layer, permits a further inference. In the primate, the inner third of the molecular layer normally receives a terminal input from large mossy cells in the polymorphic layer which form the dentate gyrus associational and commissural pathway 31. The absence of nearly all cells in the polymorphic zone of the hilus in preparations stained for glutamylglutamate-like immunoreactivity and Nissl stained preparations, and the reduced number of neurons counted in the hilus 14 support the notion that the dentate gyrus associational pathway may be missing in cryptogenic epilepsy. The activation of the associational/commissural pathway in rats has been shown to inhibit the excitation of granule cells by perforant path stimulation 12. The loss of this pathway in hippocampi of patients with cryptogenic epilepsy may be another mechanism decreasing the inhibition of dentate granular neurons. The pattern of DYN immunoreactivity in cryptogenic epilepsy suggested that there may be recurrent collaterals from granule cells. Granule cells normally also contain the excitatory amino acids glutamate and aspartate 24. If the granule cell collateral sprouts form functional synapses with the granule cells5, this may be a mechanism for increasing the excitation of granule cells. It therefore appears that several of the cytochemical changes seen in the hippocampi of patients with cryptogenic epilepsy are consonant with probable mechanisms of epileptogenicity. Some of the changes seen in the cryptogenic epilepsy hippocampi have been reported in an experimental model of epilepsy in the rat. Sloviter33 induced epileptiform activity in the hippocampal granule cells of the rat by acute stimulation of the perforant path. Such a procedure produced the selective loss of S O M 33 and NPY neurons 34 in the dentate hilus but not neurons containing only GABA. The damage produced to the hilar neurons is thought to be by the release of excitatory amino acids in neurotoxic concentrations 32, and may be an
early event in the generation of seizure activity. A similar excitotoxic mechanism may be responsible
to the tumor-related epilepsy condition, where a
for the cell damage and resultant changes in inhibition in cryptogenic epilepsy patients. Axonal sprout-
particular focus - - chemical or electrical - - is the point of seizure initiation. These animal models may
ing was not observed in this model. However, Sutula and co-workers 39 have reported sprouting of granule
exhibit some of the changes seen in the cryptogenic epilepsy patients but n o n e mimic the spectrum of changes observed in the human. Genetic models of epilepsy, in which the hippocampus has been examined, such as the seizure sensitive gerbil m'25 do not
cell collaterals and synaptic reorganization throughout the dentate molecular layer as an early event in perforant path kindling in the rat, and Tauck and Nadler 4° have reported granule cell collaterals into the inner molecular layer with functional synapse formation in kainic acid lesion induced seizures in rat. The sprouting in the h u m a n seizure focus does not exactly parallel either of these models. Though granule cell sprouting in the h u m a n resembles the condition in the lesion induced seizure model, sprouting into the entire molecular layer in the h u m a n is from a different set of neurons to that seen in Sutula's model. Third, these studies on h u m a n tissue demonstrate
epilepsy that have been studied 2~'27 seem more akin
exactly parallel the situation in h u m a n T L E either. There is thus the need for the continued search for a better model for h u m a n T L E , and perhaps till then, careful studies on h u m a n epileptogenic tissue may be the best model for elucidating the mystery of h u m a n temporal lobe epilepsy.
We thank Ms. Ilona Kovacs, Dr. Kamal Thapar,
the need for the careful selection of animal models of h u m a n epilepsy and the need for cautious extrapolation of results from animal studies to the h u m a n condition. Many of the animal models of focal
Dr. Jonathan Partington and Mr. Sanjoy Sundaresan for expert assistance with immunohistochemistry; Drs. James Madl, A n d r e w Malcolm and Robert Benoit for generous donations of somatostatin antibody; and Dr. Robert S. Sloviter for valuable comments on this paper.
l Amaral, D.G., Insauti, R. and Campbell, M.J., Distribution of somatostatin immunoreactivity in the human dentate gyms, J. Neurosci., 8 (1988) 3306-3316. 2 Andersen, P., Holmquist, B. and Voorhoeve, EE., Entorhinal activation of dentate granule cells, Acta Physiol. Scand., 66 (1966) 448-460. 3 Babb, T.L., GABA mediated inhibition in the Ammon's Horn and pre-subiculum in human temporal lobe epilepsy: GAD immunocytochemistry. In G. Nistico, P.L. Morselli, K.G. Lloyd, R.G. Fariello and J. Engel (Eds.), Neurotransmitters, Seizures and Epilepsy 111, Raven Press, New York, 1986, 293-302. 4 Babb, T.L., Brown, W.J., Pretorius, J., Davenport, C., Lieb, J.P. and Crandall, P.H., Temporal lobe volumetric cell densities in temporal lobe epilepsy, Epilepsia, 25 (1984) 729-740. 5 Babb, T.L., Kupfer, W.R. and Pretorius, J.K., Recurrent excitatory circuits by 'sprouted' mossy fibers into the fascia dentata of human hippocampal epilepsy, Epilepsia, 29 (1988) 674. 6 Burton, C.J., The Neuropathology of Temporal Lobe Epilepsy, Oxford University Press, New York, 1988. 7 Chan-Palay, V., Somatostatin immunoreactive neurons in the human hippocampus and cortex shown by immunogold/ silver intensification on Vibratome sections: coexistence with neuropeptide Y neurons, and effects in Alzheimertype dementia, J. Comp. Neurol., 260 (1987) 201-223. 8 Colmers, W.F., Lukowiak, K. and Pittman, Q.J., Neuropeptide Y actions in the rat hippocampal slice: site and mechanism of presynaptic inhibition, J. Neurosci., 8 (t988)
3827-3837. 9 Curtis, D.R., Felix, D. and McLennan, H., GABA and hippocampal inhibition, Br. J. Pharmacol., 40 (1970) 881-883. 10 Dam, A.M., Bajoreck, J.C. and Lomax, P., Hippocampal neuron density and seizures in the mongolia gerbil, Epilepsia, 22 (1981) 667-674. 11 Dodd, J. and Kelly, J.S., Is somatostatin an excitatory transmitter in the hippocampus?, Nature (Lond.), 273 (1978) 674-675. 12 Douglas, R.M., McNaughton, B.L. and Goddard, G.V., Commissural inhibition and facilitation of granule cell discharge in fascia dentata, J. Comp. Neurol., 219 (1983) 285-294. 13 Falconer, M.A., Serafetinides, E.A. and Corsellis, J.A.N., Etiology and pathogenesis of temporal lobe epilepsy, Arch. Neurol., 10 (1964) 233-248. 14 Guimares, P.O., Kim, J.H., de Lanerolle, N.C., Robbins, R.J. and Spencer, D.D., A quantitative evaluation of hippocampal neuronal loss in temporal lobe epilepsy without primary glial tumor, J. Neuropathol. Exp. Neurol., 47 (1988) 372. 15 Hass, H.L., Hermann, A., Greene, R.W. and Chan-Palay, V., Action and localization of neuropeptide tyrosine (Y) on hippocampal neurons of the rat in slice preparations, J. Comp. Neurol., 257 (1987) 208-215. 16 Hsu, S.M., Raine, L. and Fanger, H., The use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures, J. Histochem. Cyto-
chem., 29 (1981) 577. 17 Jackson, H.H. and Beevor, C.E., Case of a tumor of the right temperosphenoidal lobe bearing on the location of the sense of smell and the interpretation of a particular variety of epilepsy, Brain, 12 (1889) 346-357. 18 Jackson, J.H. and Coleman, W.S., Case of epilepsy with tasting movements and 'dreamy state': very small patch softening in the left uncinate gyrus, Brain, 21 (1898) 580-590. 19 K6hler, C., Eriksson, L.G., Davies, S. and Chan-Palay, V., Co-localization of neuropeptide tyrosine and somatostatin immunoreactivity in neurons of individual subfields of the rat hippocampal region, Neurosci. Lett., 78 (1987) 1-6. 20 Louis, E.D., Williamson, P.D. and Darcey, T.M.. Experimental models of chronic focal epilepsy: a critical review, Yale J. Biol. Med., 60 (1987) 255-272. 21 Madl, J.E., Larsson, A.A. and Beitz, A.J., Monoclonal antibody specific for carbodiimide-fixed glutamate: immunocytochemical localization in the rat CNS, J. Histochem. Cytochem., 34 (1986) 317-326. 22 Masukawa. L.M., Higashima. M., Kim, J. and Spencer, D.D., Small decreases in synaptic inhibition can reproduce the abnormal discharges present in hippocampal brain slices from epileptic patients, Epilepsia, 29 (1988) 674. 23 Mauritzen Dam, A., Hippocampal neuron loss in epilepsy and after experimental seizures. Acta Neurol. Scand.. 66 (1982) 601-642. 24 Ottersen, O.P. and Storm-Mathisen, J., Differential neuronal localization of aspartate-like and glutamate-like immunoreactivities in the hippocampus of rat, guinea-pig, Senegalese baboon (Papio papio), with a note on the distribution of gamma-aminobutyrate, Neuroscience, 16 (1985) 589-606. 25 Peterson, G.M., Ribak, C.E. and Oertel. W.H., A regional increase in the number of hippocampal GABAergic neurons and terminals in the seizure-sensitive gerbil, Brain Research, 340 (1985) 384-389. 26 Pittman, Q.J. and Siggins, G.R., Somatostatin hyperpolarized hippocampal pyramidal cells in vitro, Brain Research, 221 (1981) 402-408. 27 Ribak, C.E., Contemporary methods in neurocytology and their application to the study of epilepsy. In A.V. DelgadoEscueta, A.A. Ward, D.M. Woodbury and R.J. Porter (Eds.). Basic Mechanisms of the Epilepsies: Molecular and Cellular Approaches, Advances in Neurology, 44, Raven Press, 1986, 739-764.
28 Ribak, C.E., Hunt, C.A., Bakay, R.A.E. and Oertel, W.H., A loss of small GABAergic somata is associated with the preferential loss of GABAergic terminals at epileptic foci, Brain Research, 363 (1986) 78-90. 29 Ribak, C.E., Vaughn, J.E. and Saito, K., Immunocytochemical localization of glutamic acid decarboxylase in neuronal somata following colchicine inhibition of axonal transport, Brain Research, 140 (1978) 315-332. 30 Robbins, R.J., Adrian, T., de Lanerolle, N.C., Kim, J.H., Halvonik, S. and Spencer, D.D., Hippocampal somatostatin, CCK and VIP levels in patients with temporal lobe epilepsy, Soc. Neurosc. Abstr., 13 (1987) 366. 31 Rosene, D.L. and Van Hoesen, G.W., The hippocampal formation of the primate brain. In E.G. Jones and A. Peters (Eds.), Cerebral Cortex, Volume 6, Plenum Press, New York, 1987, 345-456. 32 Sloviter, R.S., On the role of seizure activity and endogenous excitatory amino acids in mediating seizure-associated hippocampal damage. In R. Schwarcz and Y. Ben-Ari (Eds.), Excitatory Amino Acids and Epilepsy, Plenum, New York, 1986, 659-671. 33 Sloviter, R.S., Decreased hippocampal inhibition and a selective loss of interneurons in experimental epilepsy, Science, 235 (1987) 73-76. 34 Sloviter, R.S., Chemically defined hippocampal interneurons and their possible relationship to seizure mechanisms. In C. KOhler and V. Chan-Palay (Eds.), The Hippocampus: New Vistas, A.R. Liss, New York, in press. 35 Sommer, W., Erkrankung des Ammonshornes als aetiologisches Moment der Epilepsie, Arch. Psychiatr. Nervenkrankh., 10 (1880) 631-675. 36 Spencer, D.D., Spencer, S.S., Mattson, R.H., Williamson, P.D. and Novelly, R., Access to posterior medial temporal lobe structures in the surgical treatment of temporal lobe epilepsy, Neurosurgery, 15 (1984) 667-671. 37 Spencer, S.S., Schwarcz, S.S. and Spencer, D.D., The treatment of epilepsy with surgery, Merrit Putnam Quarterly, 5 (1988) 3-17. 38 Sternberger, L.A., Immunocytochemistry, 2nd edn., Wiley, New York, 1979. 39 Sutula, T., Xiao-Xian, H., Cavazos, J. and Scott, G., Synaptic reorganization in the hippocampus induced by abnormal functional activity, Science, 239 (1988) 11471150. 40 Tauck, D.L. and Nadler, J.V., Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid-treated rats, J. Neurosci., 5 (1985) 1016-1022.