Epilepsi, 4( Sup. 6):$90-S95, 2000 ippincet Wiliams & Wilkins, ine, Bakimore ‘Gfimematcnal League Again! Epilepsy What is GABAergic Inhibition? How Is it Modified in Epilepsy? C. Bernard, R. Cossart, J. C. Hirsch, M. Esclapez, and Y. Ben-Ari INSERM U29-INMED, Pare Scientifique de Luminy, Marseille Cédex, France Summary: A deficit of y-aminobutytic acid-ergic (GABAergic) inhibition is hypothesized to underlie most forms of epilepsy. Although apparently a straightforward and logical hypothesis to test, the search for a deficit of GABAergic inhibition in epileptic tissue has revealed itself to be as difficult as the quest fot the Holy Grail. The investigator faces many obstacles, in- cluding the multiplicity of GABAergic inhibitory pathways and the multiplicity of variables that characterize the potency of inhibition within each inhibitory pathway. Perhaps more importantly, there seems to be no consensual definition of GABAergic inhibition. The first goal of this review is to try to clarify the notion of GABAergic inhibition. The second goal is to summarize our current knowledge of the various alterations that occur in the GABAergic pathways in temporal lobe epi- lepsy. Two important features will emerge: (a) according to the variable used to measure GABAergic inhibition, it may appear ‘increased, decreased, or unchanged; and (b) these modifica- tions are brain area and inhibitory pathway-specific. The pos- sible functional consequences of these alterations are discussed. Key Words: Temporal lobe epilepsy—Hippocampus— GABA—Intemeurone, In the adult hippocampus, the output of principal cells (dentate granule cells, CA3 and CA1 pyramidal neurons) is tightly controlled by the activity of -y-aminobutyric acid-ergic (GABAergic) intermeurons. This control has at least three main components: (a) the activation of post- synaptic GABA receptors counteracts the membrane de- polarization induced by excitatory inputs via direct hy- perpolarizing or shunt effects; (b) GABA receptor mediated synaptic responses can directly block action potential firing (1,2); and (c) interneurons can synchro- nize the firing of principal neurons during oscillations (3-6). These three effects of GABAergic inhibition ulti- mately modify the probability of action potential firing in the postsynaptic neuron, In temporal lobe epilepsy (TLE), highly synchronous discharges occur simultaneously in large populations of neurons. Is such abnormal behavior related to a modifi- cation of the control of principal cells by GABAergic interneurons? Two nonexclusive hypotheses can be pro- posed: the ability of inhibition to counterbalance mem- brane depolarization and action potential firing is de- creased, and modifications occur within the interneuro- nal network to facilitate the synchronized firing of principal cells. Interneurons and principal cells fire syn- chronously during epileptiform discharges (7,8), but the ‘Address correspondence and reprint requests to Dz. Christophe ‘Bemard at Division of Neuroscience, S700, Baylor College of Medi- cine, 1 Baylor Plaza, Houston, TX 77030, U.S.A. E-mail: cbemnard@ Irpneuse.bem.ume.edu §90 involvement of interneurons in the synchronization of principal cells in chronic epilepsy remains to be investi- gated. Although a potentially important mechanism to explain epileptogenesis (9), this hypothesis will not be discussed further. The present review focuses on the ‘other hypothesis, ic., a deficit of inhibition in chronic epilepsy, particularly with regard to the “fate” of GABA, receptor-mediated inhibition, Because compounds that enhance GABA, receptor— mediated inhibition are used successfully in controlling some types of epilepsies in humans and because epilep- tiform activity can be triggered when GABA. receptor— ‘mediated inhibition is blocked in most brain structures (10), itis often taken for granted that GABAergic inhi- bition is decreased in epileptic tissue, although there is, some disagreement on this point (11,12). However, if A implies B, it does not necessarily mean that B implies A. Obviously, the hypothesis is a valid one, and it has been tested extensively, Before providing some of the most important results related to the fate of inhibition in TLE, it is necessary to ask a critical question: What is GABAergic inhibition? WHAT IS GABAERGIC INHIBITION? ‘The definition of GABAergic inhibition faces three major difficulties. (a) Because “inhibition” implies the ability to restrain, activation of postsynaptic GABA, receptors should de- crease the firing probability of the inhibited neuron. However, the consequences of GABA, receptor activGABAERGIC INHIBITION S91 tion critically depend on the experimental conditions. ‘According to the concentration gradients of chloride and bicarbonate, the anions that permeate the GABA, iono- phore, the activation of GABA, receptors can lead to either a depolarization or a hyperpolarization of the postsynaptic membrane. The depolarizing action of GABA is not limited to neonates (13); it also occurs in the dendrites after intense activation of GABA, recep- tors (14,15). Even when GABA has a hyperpolarizing action, it can lead to delayed excitatory action via a di- rect interaction with intrinsic membrane conductances (4). For example, GABA, receptor-mediated hyper- polarization of the membrane could deinactivate Ca?* channels and secondarily boost excitatory inputs, facili- tate back-propagation of action potentials from the soma to the dendrites, and induce abnormal action potential firing in the dendrites (16). Therefore, the first dif- ficulty lies in the fact that “inhibition” does not seem appropriate to describe a context-dependent concept, ie, GABA, receptor function. For lack of a better term, we will still use GABAergic inhibition, and for the sake of simplicity in this review, we will consider only the GABA, receptor-mediated hyperpolarizing/ shunt effect. (b) What determines the potency of inhibition, i« the level of hyperpolarizing/shunt in the postsynaptic cell? The hyperpolarizing/shunt effect results from the opening of the GABA, ionophore after the fixation of synaptically released GABA on its receptor site. Therein lies the second difficulty in defining GABAergic inhibi- tion. The hyperpolarizing/shunt effect is controlled by a constellation of variables that range from the properties of the GABAergic neurons to those of the GABAergic synapse. Some of these are detailed below. At the end of the line lie the GABA, receptors, Their physiological properties (conductance, kinetics, inactiva- tion) depend on their subunit composition and on the action of intracellular (protein kinases, Ca™*) as well as extracellular (benzodiazepines, barbiturates, steroids, polyvalent cations, ethanol) factors (17). The activation of GABA, receptors usually results from the synaptic release of GABA. Two types of release have been identified; one requires the presence of an action potential in the presynaptic terminal, and the other occurs randomly in an action potential-independent ‘manner (e.g., miniature activity). In both cases, exocy- tosis of a GABA-containing vesicle is a probable event modulated by the properties of the terminal internal ma- chinery (including the various Ca** effectors), the past activity of the terminal, diffusible factors, polyvalent cat- ions, and a very large number of ionotropic and metabo- tropic receptors for neurotransmitters or neuromodula- tors located on the presynaptic terminal (18; see also 19,20 for additional references). During steady-state con- ditions in vitro, miniature activity in CA pyramidal cells represents as much as 60% of the GABAergic activity received by principal neuron somata (21,22). The rest (40%) is provided by an action potential-dependent pro- cess, e.g., the firing of GABAergic interneurons, Going farther upstream, the amount of activity- dependent GABAergic inhibition received by postsynap- tic neurons is a function of the number of presynaptic terminals made by a single GABAergic neuron, the number of presynaptic GABAergic neurons, the intrinsic membrane properties of the latter, the distribution/ properties of their various receptor channels, and the dis- tribution/properties of their presynaptic excitatory and inhibitory terminals. Thus, the number of variables that need to be assessed to describe GABAergic inhibition is truly mind boggling. (©) The third difficulty is attributable to the multiplic- ity of GABAergic pathways. Each pathway is defined by a specific class of inhibitory neuron with unique mor- phological, physiological, and functional features (1,20,23,24). Because of this heterogeneity, each subspe- cific inhibitory pathway should be investigated in iso- lation, ie., each of the variables cited in the previous paragraph should be assessed for each subspecific inhibi- tory pathway. ‘What, then, is GABAergic inhibition? GABAergic in- hibition can be defined as a set of variables distributed in multiple inhibitory pathways. Because the values of these variables are continuously changing and interact- ing, the state of inhibition can only be approached when the system is frozen in a quasi-stable state. Only in this condition can inhibition be compared in two conditions, eg,, in epileptic and control neuronal networks. Unfor- tunately, this type of approach can only give a partial image of a multidimensional object. Many of these vari- ables have been assessed in human and experimental TLE, and it is now clear that multiple modifications oc- cur at different locations. We will now detail some of the modifications that occur in the various inhibitory pathways, from the source of GABAergic inhibition, the GABAergic intemeurons, to the targets, the postsynaptic GABA, receptors. LOSS OF INTERNEURONS IN TLE Numerous studies have consistently reported a loss of GABAergic interneurons in experimental (25-30) as well as in human (31-34) TLE. This loss of inhibitory interneurons should lead to a decreased number of in- hibitory synapses on the postsynaptie cells. Ultrastrue- tural studies indicate that the number of perisomatic GABAergic terminals on CAI pyramidal cells is not modified in the kainic acid and pilocarpine models of TLE before (35) and after spontaneous recurrent sei- zures have developed (36). Interestingly, in the kainic acid model, a loss of perisomatic parvalbumin- Epilepsia, Vol. 41, Suppl. 6, 2000S92 C. BERNARD ET AL. immunoreactive terminals was reported, whereas the number of these terminals around the initial segment of CAL pyramidal cells remained unchanged (28). These results suggest a transient deafferentation of pyramidal cell somata by GABAergic terminals followed by a re- afferentation, implying a reactive synaptogenesis. A transient loss of inhibitory terminals could explain the transient loss of paired pulse inhibition (37). The sprout- ing of the axon of GABAergic neurons has been reported in another preparation (38); in the hippocampus, it could occur in parallel with the well-documented recovery of excitatory terminals via the sprouting of excitatory axons in TLE (31,3947). GABAergic terminals along the dendrites of principal cells have not been quantified. A massive reorganization is expected along the dendrites because of the clearly identified loss of somatostatin-containing interneurons in TLE (48,49). In the CAI area, most somatostatinergic interneurons are located in the stratum oriens, where they represent the most prevalent population of GABAergic interneurons (23,50). The extensive axonal arbor of so- ‘matostatinergic interneurons forms symmetric synapses, preferentially on the distal dendrites of pyramidal cells (51), at the site of perforant path afferences in the lacu- nosum moleculare. The loss of somatostatinergic inter neurons may have important functional consequences, not only because somatostatin has anticonvulsant prop- erties (52,53) but also because of the resulting disinhi- bition of pyramidal cells. Somatostatinergic interneurons are reliably activated by excitatory inputs (54), they re- ceive a strong excitation from pyramidal cells (55-57), and, in the slice preparation, most of them (>80%) fire spontaneously, thus providing a strong, precise inhibi- tory barrage at the site of perforant path afferences (our unpublished observations). Because a decrease in the a tivity of these interneurons allows the direct excitation of CALL pyramidal cells by the temporoammonic pathway (57), their loss would have a direct disinhibitory effect in chronic epilepsy. This issue remains to be investigated In chronic epilepsy, the fate of other types of GABAergic interneurons containing other peptides or calcium-binding proteins (vasoactive intestinal polypep- tide [VIP], cholecystokinin (CCK), calretinin, etc.) has not been addressed. ‘The loss of GABAergic interneurons should logically result in the disinhibition of principal cells. This hypoth- esis is supported by the decrease in paired-pulse inhibi- tion (a quantity hypothesized to be related to the “strength” of inhibition) reported in vivo (37,58). How- ever, as reported above, paired-pulse inhibition recovers, 7 to 8 days after the initial lesion-induced status epilep- ticus (37), suggesting that the surviving inhibitory path- ways can still exert their function. This raises the ques- tion of the fate of the surviving GABAergic interneurons in TLE, Epilepsia, Vol. 41, Suppl. 6, 2000 MODIFICATIONS IN THE SURVIVING INHIBITORY PATHWAYS IN TLE ‘An ultrastructural study indicates that the number of GABAergic and non-GABAergic terminals is not modi- fied in various types of interneurons in TLE, although an increase in the number and length of GABAergic termi- nals in the perisomatic region of lacunosum moleculare interneurons was reported (59). A similar study per- formed in the dendrites of the various types of interneu- rons should give some insight into the reorganization of the distribution of synaptic terminals. These data are now available in control tissue and could be used for com- parison (60). ‘What are the properties of the excitatory and inhibi- tory synapses in the surviving interneurons in TLE? Nei- ther evoked excitatory nor inhibitory responses seem modified in interneurons, despite a transient decrease of the latter in acunosum moleculare interneurons (61,62). The functional properties of the presynaptic terminals, the distribution and properties of postsynaptic receptors, and their control by extracellular and intracellular factors, remain to be investigated. These issues are important to consider because these factors are modified in principal cells in TLE (36,63-65). Is the functioning mode of interneurons modified in ‘TLE? It is now clear that interneurons are not dormant in ‘TLE (66). All of the interneurons recorded in chronically epileptic animals fired bursts of action potentials after stimulation of their excitatory afferents (7,61). More im- portantly, synchronized epileptiform discharges occur si- ‘multaneously in interneurons and pyramidal cells during spontaneous seizures in chronic models of TLE (7) as well as in acute models (8). Epileptiform discharges in interneurons are dependent on the activation of N- methyl-D-aspartate receptors (7), as in pyramidal cells (67), suggesting a reorganization of the excitatory path- ways common to principal cells and interneurons. How- ever, many issues remain to be investigated, including the intrinsic membrane properties of the surviving inter- neurons and the distribution and properties of their ex- citatory and inhibitory synapses. ‘As mentioned above, principal cells receive a continu- ‘ous bombardment of inhibitory signals in vitro, part of which (40%) is provided by the firing of interneurons. In control tissue, one-third of the interneuronal population fires spontaneously (22,68~70) at an average frequency of 4.5 Hz (70). In TLE, four-fifths of the intemeuronal population fires spontaneously, and the average fre- quency is increased to 9 Hz (22,70). Therefore, the spon- taneous activity of interneurons is increased in TLE. This hyperactivity is dependent on the activation of glutama- tergic receptors (our unpublished observations) and seems to be attributable to the increased excitatory drive (700%) received by interneurons in TLE comparedGABAERGIC INHIBITION. $93 with controls (22,70). This glutamate receptor dependent hyperactivity is found in all morphologically identified intemeurons (22,70). It is possible that the in- creased excitatory drive reported in interneurons results from the sprouting of the CAI associational pathway (46,47). In CA1 pyramidal cells, the excitatory drive is increased by 500% in TLE, and it is directly related to the sprouting of the CA1-associated pathway (47). Thus, the modifications that occur within the interneu- ronal population are complex. On the one hand, the loss of inteeurons should logically result in a decreased inhibitory control of principal cells. On the other hand, the hyperexcitability and hyperactivity of the surviving interneurons should result in an increased inhibitory con- trol of principal cells. To determine which of these two phenomena is predominant, it is necessary to measure the amount of inhibition received by the principal cells. GABAERGIC INHIBITION IN PRINCIPAL CELLS IN TLE A large set of variables can describe the activity at a given GABAergic synapse, including the activity of its source (the emitting interneuron), the properties of the presynaptic GABAergic terminal and its control by neu- tomodulators, and the properties of the postsynaptic GABA, receptors and their control by extracellular and. intracellular factors. Because of the heterogeneity of in- terneuronal types (1,23,24), each set of variables should be assessed for each type of GABAergic synapse thus defined. So far, some of these variables have been char- acterized using mostly somatic recordings. Because it is not possible to clamp large dendritic structures (71-74), somatic recordings mostly provide information about perisomatic GABAergic synapses, ie., synapses origi- nating from either basket or chandelier cells (23). These interneurons appear to play a very important role because activation of the perisomatic inhibitory pathway can block action potential firing in principal cells (1) or syn- chronize their firing (4), ‘The postsynaptic perisomatic GABA, receptors un- dergo a profound reorganization in TLE. Modification of the GABA, receptor subunits expressed in dentate granule cells (65) results in larger evoked responses to GABA, a decreased sensitivity to zolpidem, and an in- creased inhibition by Zn?* (63-65). In parallel, the number of postsynaptic GABA, receptors is increased (75). In CA1 pyramidal cells, the reorganization is different. Although it is not known whether the ex- pression of GABA, receptor subunits is changed in these neurons, GABA, receptor-mediated responses are decreased (36,64) and the effect of clonazepam is re- duced (64). Despite the decrease of postsynaptic re- sponses, perisomatic GABA, receptor-mediated re- sponses can still block action potential firing in CAI pyramidal cells (7). A major conclusion from these ob- servations is that the alterations in the GABAergic path- ways are area-dependent. Presynaptic terminals also undergo modifications in experimental epilepsy. Perisomatic presynaptic boutons on dentate granule and CA1 pyramidal cells are enlarged (36,75), and the reserve pool of GABA-containing vesicles is depleted by 50% (36). The control of the presynaptic terminals is also altered. Zinc inhibits min- iature activity in dentate granule cells (63), whereas the inhibition of GABA release by the activation of GABAg autoreceptors disappears in CAI pyramidal cells in TLE (76). Finally, miniature activity, which provides 60% of the spontaneous inhibitory drive in controls (77), is de- creased by more than 50% in CAI pyramidal cells in TLE (36). Interestingly, miniature activity is increased in a model of febrile seizure (78) and in a cortical model of postlesional epilepsy (11). These observations indicate a profound modification in the function of presynaptic ter- minals that is area-dependent. However, the probability of release of GABA after an action potential in the pre- synaptic terminal, possibly the most important factor, has not been quantified. Taking into account the modifications that occur within the various inhibitory pathways (from the inter- neuron to the postsynaptic GABA. receptor), what is the amount of inhibition received by pyramidal cells? Two functional states should be investigated: steady state (outside epileptiform activity) and transients (during epi- leptiform activity). During steady state, the amount of inhibition received by CAI pyramidal cell somata is i creased by 50% because of the hyperactivity of periso- matic inhibitory interneurons (22,70), despite the deficit of GABA quantal release (36). In CAL pyramidal cell dendrites, the inhibitory drive is decreased by 35%, pos- sibly because of the loss of interneurons, in particular the somatostatinergic population (79). During transients, al- though CAI pyramidal cell somata still receive large GABAergic currents (7), these currents appear down- regulated by the large Ca”* influx that occurs at the same time (80,81), CONCLUSION The major conclusion to be drawn from these obser vations is a lesson in humility. Each modification of the factors that characterize inhibition is model-, time-, area-, and pathway-dependent. Within one brain region, for one model of epilepsy, at a given time, some modi- fications will appear as proepileptic and others will ap- peat as antiepileptic. In fact, the number of factors that, can be modified is so important that it is impossible to predict the resulting behavior of the neuronal network. An integrative framework is needed, and realistic com- puter models could prove useful to understanding the Epilepsta, Vol 41, Suppl 6, 2000sot outcome of these modifications. However, as mentioned above, some critical variables still need to be assessed in each subspecific inhibitory pathway. The latter point is particularly important to consider because there exist clear anatomical and functional organizing principles de- fined by the diverse types of GABAergic neurons and their synapses (20). REFERENCES 1. Miles R, Toth K, Gulyis AI, Hajos N, Freund TP. Differences ‘between somatic and dendritic inhibition in the hippocampus. New- ron 1996:16:815-23, 2, Dvorak-Carbone H, Schuman EM, Patterned activity in stratum lacunosum moleculare inhibits CAL pyramidal neuron firing. J Neurophysiol 1999:82:3213-22. 3. Buzsdki G, Horvath Z, Urioste R, Hetke J, Wise K. High- frequency network oscillation in the hippocampus, Science 1992; 256:1025-1, 4. Cobb SR, Buhl EH, Halasy K, Paulsen ©, Somogyi P, Synchro- nization of neuronal activity in hippocampus by individual GABAergic interneurons, Nature 1995;378:75-8. 5. Whitington MA, Traub RD, Jefferys JG, Synchronized oscilla: tions in intemeuron networks driven by metabotropic glutamate receptor activation. Nature 1995;373:612-5, 6. Csicsvasi J, Hirase H, Caurko A, Mamiya A, Buzsaki G, Oscilla- tory coupling of hippocampal pyramidal cells and interneurons in the behaving rat. J Neurosci 1999:19:274-87, 7. Esclapez M, Hirsch JC, Khazipov R, Ben-Ari Y, Berard C, Op- erative GAB Acrgic inhibition in hippocampal CA\1 pyramidal neu- rons in experimental epilepsy. Proe Natl Acad Sci USA 199794: 1215-6, 8, Velazquez JL, Carlen PL. Synchronization of GABAcrgic inter- neuronal networks during seizure-like activity inthe rat horizontal hippocampal slice. Eur J Neurosci 1999;11:41 10-8. 9. Avoli M. GABA-mediated synchronous potentials and seizure generation. Epilepsia 1996;37:1035-42, 10, Prince DA. Neurophysiology of epilepsy. Annu Rev Neurosci 1978;1:395-415, 11. Prince DA, Jacobs KM, Salin PA, Hoffinan S, Parada I. Chronic focal neocortical epileptogenesis: does disinhibition play a role? Can J Physiol Pharmacol 1997375:500-. 12, Bush PC, Prince DA, Miller KD. Increased pyramidal excitability and NMDA conductance can explain posttraumatic epileptogen- esis without disinhibition: a model. J Neurophysiol 1999;82:1748— 58, 13, Cherubini, Gaiarsa JL, Ben-Ari Y. GABA: an excitatory trans- ‘mitter in early postnatal life, Trends Neurosci 1991;14515-9. 14, Andersen P, Dingledine R, Gjrstad L, Langmoen 1A, Laursen AM. Two different responses of hippocampal pyramidal cells to application of gamma-amino butyric acid. J Physiol 1980;305: 779-96. 15, Staley KI, Soldo BL, Proctor WR. Tonic mechanisms of neuronal excitation by inhibitory GABAA receptors. Science 1995;268: 977-81, 16, Magee J, Hoffman D, Colbert C, Johaston D. Electrical and cal- cium signaling in dendrites of hippocampal pyramidal neurons. Annu Rev Physiol 1998;60:327-46, 17. Hevers W, Luddens H. The diversity of GABAA receptors: phar- ‘macological and electrophysiological properties of GABAA chan- nel subtypes, Mo! Neurobiol 1998;18:35-86, 18, MacDermott AB, Role LW, Siegelbaum SA. Presynaptic iono- twopic receptors and the control of transmitter release, Annu Rey Neurosci 1999:22:443-88, 19, Cossart R, Esclapez M, Hirsch JC, Bemard C, Ben-Ari Y. GluRS kainate receptors activation in intereurons inereases tonic inhibi- tion of pyramidal cells. Nar Neurosci 1998;1:470-8. 20, Gupta A, Wang Y, Markram H. Organizing principles fora diver- sity of GABAergic intemeurons and synapses in the neocortex. Science 2000;281:273-8, Epilepsia, Vol. 41, Suppl. 6, 2000, 21 29, 30, 31 32. 33, 35, 38, 39, 4 2. 48, C. BERNARD ET AL. Soltesz I, Smetters DK, Mody I Tonic inhibition originates from synapses close to the soma. Neuron 1995;14:1273-83 Bemard C, Esclapez M, Hirsch JC, et al, Increased inhibitory GABAergic drive in the soma of CAI pyramidal colls in experi- mental epilepsy. Soc Newrosci Abstr 1999;29:340.10, Freund TP, Buzséki G. Interneurons of the hippocampus. Hippo- campus 1996;6:347-470, |. Parra P, Gulyas AL, Miles R. How many subtypes of inhibitory ceils in the hippocampus? Neuron 1998;20:983-3, Ben-Ari Y, Nitecka L, Woodson W, Robain ©, Gho M, Cherubini E. Vulnerabilite preferentielle des intemmeurones GABAergiques de hippocampe a Vacide kainique. L’encephale 1987; Sloviter RS. Decreased hippocampal inhibition and a selective loss of interneurons in experimental epilepsy, Science 1987;235:73-6. . Obenaus A, Esclapez M, Houser CR. Loss of glutamate decarbox- ylase mRNA-containing neurons inthe rat dentate gyrus following pilocarpine-induced seizures. J Neurosci 1993;13:4470-85, Best N, Mitchell J, Wheal HV, Ultrastructure of parvalbumin- ‘immunoreactive neurons in the CA1 area of the rat hippocampus following a kainic acid injection. Acta Neuropathol 1994:87:187- 95. Houser CR, Esclapez M. Vulnerability and plasticity of the GABA system inthe pilocarpine model of spontaneous recurrent seizures. Epilepsy Res 1996;26:207-18, ‘Morin F, Beaulien C, Lacalle IC. Selective loss of GABA neurons in area CAL of the rat hippocampus after intraventricular kainate Epilepsy Res 1998;32:363-9. DeLanerolle N, Kim J, Robbins R, Spencer D. Hippocampal in temeuron loss and plasticity in human temporal lobe epilepsy. Brain Res 1989:495:387-95. Robbins RJ, Brines JL, Kim JH, etal. A selective loss of somato- statin in the hippocampus of patients with temporal lobe epilepsy. Ann Neurol 1991;29:325-32. Mather GW, Babb TL, Pretorius JK, Leite JP. Reactive synapto- {genesis and neuron densities for neuropeptide Y, somatostatin, and alutamate decarboxylase immunoreactivity in'the epileptogenic hhuman fascia dentata. J Neurosci 1995;15:3990-4008. Marco P, Sola R, Alijarde M, Sanches. A, Ramon y Cajl S, De Felipe J. Inhibitory neurons in the human epileptogenic temporal eocortex: an immunoeytochemical study. Brain 1996;119:1327— 41, ‘Chapman CA, Lacaille JC. Intrinsic theta-frequeney membrane potential oscillations in hippocampal CAL intemeurons of stratum lacunosum-moleculare. J Neurophysiol 1999;81:1296-307. Hirsch JC, Agassandian C, Merchén-Pérez A, et al, Deficit of {quantal release of GABA in experimental temporal lobe epilepsy. Nat Neurosci 1999:2:499-500, Hellier JL, Patrylo PR, Don P, Nett M, Rose GM, Dudek FE, ‘Assessment of inhibition and epileptiform activity in the, septal dentate gyrus of freely behaving rats during the first week after kainate treatment. J Neurosci 1999;19:10033-64, ‘Sil FI, Drake-Baumann R, Leiman AM, Heradon RM, Tiekotter KL. Morphological correlates of altered neuronal activity in onga- notypic cerebellar cultures chronically exposed to anti-GABA agents. Dev Brain Res 1994;77:123-32, Nadler J, Perry B, Cotman C. Selective reinnervation of hippo- campal area CAL and fascia dentate after destruction of CA3-CA4 afferents with kainic acid. Brain Res 1980%182:1-9. [Nadler J, Perry B, Gentry C, Cotman C. Loss and reacquisition of hippocampal synapses after selective destruction of CA3-CA4 af- ferents with kainic acid. Brain Res 1980;191:387-403, ‘Tanck D, Nadler J. Evidence for functional mossy fiber sprouting in hippocampal formation of kainic acid treated rats. J Neurosci 1985:5:1016-22, Represa A, Tremblay E, Ben-Ari ¥. Kainate binding sites in the hippocampal mossy fibers: localization and plasticity, Neurosei- ence 1987;20:739-48, Phelps S, Mitchell J, Wheal H. Changes to synaptic ultrastructure in field CA1 of the rat hippocampus following intracerebroven- ticular injection of kainic acid. Neuroscience 1991;40:687-99, Cronin J, Obenaus A, Houser CR, Dudek FE, Eleetrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers. Brain Res 1992;573:305-10,45, 41, 48. 49, 30, 31 53. 54, 55. 56, 37, 58 6. Cy a, GABAERGIC INHIBITION Reptesa A, Ben-Ari Y. Kindling is associated withthe formation of novel mossy fibre synapses in the CA3 region. Exp Brain Res 1992:99:69-78 Perez Y, Morin F, Sutras I, Beaulieu C, Lacaille JC. CAL pyra- ‘midal cells in hyperexcitable slices of kainate-reated rats show sprouting of local axon collaterals. Eur J Neurosci 1996;8:736-48, Esclapez M, Hirsch JC, Ben-Ari Y, Bernard C, Newly formed excitatory pathways provide a substrate for hyperexcitability in experimental temporal lobe epilepsy. J Comp Neurol 1999408: 449-60. Morin F, Beaulieu C, Lacalle JC, Selective loss of GABA neurons in area CAL of the rat hippocampus after intraventricular kainate, Epilepsy Res 1998;32:363-9. Buckmaster PS, Jongen-Relo AL. Highly specific neuron loss pre- serves lateral inhibitory cicuits in the dentate gyrus of kainate- induced epileptic rats. J Neurosci 1999;19:9519-29, Kosaka T, Wu J-¥, Benoit R, GABAergic neurons containing Somatostain-ike immunoreactivity in the rat hippocampus and dentate gyrus. Exp Brain Res 1988:71:388-98, Katona I, Acsady L, Freund TF. Postsynaptic targets of somatostatin-immunoreactive interneurons in the rat hippocampus. Neuroscience 1999;88:37-55. Tallent MK, Siggins GR. Somatostatin acts in CAI and CA3 to reduce hippocampal epileptiform activity. J Neurophysiol 1999; 81:1626-35. Vezzani A, Hoyer D. Brain somatostatin: a candidate inhibitory role in seizures and epileptogenesis. ur J Neurosci 1999311; Martina M, Vida I, Jonas P. Distal initiation and active propagation of action potentials in interneuron dendrites. Science 2000:287: 295-300. Lacaille JC, Mueller A, Kunkel D, Schwartzkroin PA. Local circuit interactions between oriens/alveus intemeurons and CAL pyrami- dal cells in hippocampal slices: electrophysiology and morphol ogy. J Newrosci 198737:1979-93, Biasco-Ibanez JM, Freund TF. Synaptic input of horizontal inter- ‘neurons in stratum oriens ofthe hippocampal CAL subfield; strac- tural bass of feed-back activation. Eur J Neurosci 1995;7:2170- 80. Maccaferri G, MeBain CJ. Passive propagation of LTD to stratum otiens-alveus inhibitory neurons modulates the temporoammonic ‘npot to the hippocampal CAI region. Newron 1995;15:137—45. Sloviter RS. Permanently altered hippocampal structure, excitabil- ty, and inhibition after experimental status epilepticus in the rat the “dormant basket cell” hypothesis and its relevance to temporal lobe epilepsy. Hippocampus 1991;1:41-6, Morin F, Beaulieu C, Lacaille JC. Alterations of perisomatic GABA synapses on hippocampal CA\ inhibitory intemeurons and Pyramidal cells in the kainate model of epilepsy. Neuroscience 1999;93:457-67, Gulyas AI, Megias M, Emi Z, Freund TR, Total number and ratio of excitatory and inhibitory synapses converging onto single in- temeurons of different types in the CA1 area ofthe rat hippocam- pus. J Neurosci 1999;19:10082-97. ‘Rempe DA, Bertram EH, Williamson JM, Lothman EW, Intereu- rons in area CAI stratum radiatum and stratum oriens remain functionally connected to excitatory synaptic input ia chronically epileptic animals. J Neurophysiol 1997;78:1504-18. ‘Morin F, Beaulieu C, Lacaille J. Cell-specific alterations in syn- aptic properties of hippocampal CA interneuroas after kainate tweatment, J Neurophysiol 1998;80:2836-47. Buhl E, Otis T, Mody I. Zinc-induced collapse of augmented in- 64, 65, o, 10, 71 n, 2B. m 15. 16. 78. 0, 80, al. $95 hibition by GABA in a temporal lobe epilepsy model. Science 1996;271:369-73. Gibbs TW, Shumate MD, Coulter DA. Differential epilepsy- associated alterations in postsynaptic GABAA receptor function in dentate granule cells and CA neurons. J Neurophysiol 1997577; 1924-38, ‘Brooks-Kayal AR, Shumate MD, Jin H, Rikhter TY, Coulter DA. ‘Selective changes in single cell GABA(A) receptor subunit expres- sion and function in temporal lobe epilepsy. Nar Med 1998; 1166-72. ‘Bemard C, Esclapez M, Hirsch JC, Ben-Ari Y, Intereurones are not so dormant in temporal lobe epilepsy: a critical reappraisal of the dormant basket cell hypothesis. Epilepsy Res 1998;32:93-103, ‘Tumer D, Wheal H. Exctatory synaptic potential in kainic acid- denervated rat CA1 pyramidal neurons. J Neurosci 1991;11:2786— 94, Fricker D, Verheugen JA, Miles R, Cell-ttached measurements of the firing threshold of rat hippocampal neurons. J Physio! 1999; 517:791-804, Vetheugen JA, Fricker D, Miles R. Noninvasive measurements of the membrane potential and GABAergic action in hippocampal intemeurons. J Neurosci 1999;19:2546-55, Hirsch IC, CossartR, Esciapez M, Ben-Ari Y, Berard C. Reversal ‘of the balance between excitation and inhibition in hippocampal CAL intemeurones in experimental epilepsy. Soc Neurosci Abstr 1999:20:340.9, Tumer D. Waveform and amplitude characteristics of evoked re- sponses to dendritic stimulation of CAL guines-pig pyramidal cells. J Physiol 1988:395 419-29, Andersen P, Storm J, Wheal HT. Thresholds of action potentials evoked by synapses On the dendrites of pyramidal cells in the rat hippocampus in vitro. J Physiol 1987;383:509-26. ‘Tumer DA. Conductance transients onto dendritic spines in a seg- mental cable model of hispocampal neurons. Biophys J 1984:46: 85-96, Spruston N, Jaffe D, Willams S, Johnston D. Voltage- and space- ‘lamp errors associated with the measurement of electronically remote synaptic events. J Neurophysiol 1993;70:781-802. [Nusser Z, Hajos N, Somogyi P, Mody 1 Increased number of synaptic GABAA) recepiors underlies potentiation at hippocem- pal inhibitory synapses. Nature 1998;395:172-7, ‘Mangan PS, Lothman EW. Profound disturbances of pre- and post- synaptic GABAB-recepto-mediated processes in region CAL in & chronic model of temporal lobe epilepsy. J Neurophysiol 1996:76: 1282-96, tis TS, Staley KJ, Mody I. Perpetual inhibitory activity in mam- ‘malian brain slices generated by spontaneous GABA release. Brain Res 1991;545:142-50. Chen K, Baram TZ, Soltesz I. Febrile seizures in the developing brain result in persistent modification of neuronal excitability in limbic circuits. Nar Med 1999;5:888-94, Bemard C, Esclapez M, Agid F, Ben-Ari Y, Hirsch JC. Selective loss of GABAergic inhibition in the apical dendrites of CAL py- ramidal neurons in temporal lobe epilepsy. Soc Neurosci Abstr 1997;23:838.10. Beau FL, Alger BE, Trarsient suppression of GABAA-teceptor- mediated IPSPs aller epileptiform burst discharges in CA1 pyra- ‘midal cells. J Neurophysiol 1998;79:659-69, Isokawa M. Modulation cf GABAA receptor-mediated inhibition bby postsynaptic calcium in epileptic hippocampal neurons. Brain Res 1998;810:241-50, Epilepsia, Vol. 41, Suppl. 6, 2000