Brae Re.search. 32(~ ( 1985 ) 251 260 25 1 Elsevier
BRE lll-ltJt)
Axon Terminals of GABAergic Chandelier Cells Are Lost At Epileptic Foci
CHARLES E. RIBAK l)epartment ~{['Anatomy. University o,t CaliJornia. lrvine, ( A 92717 ( U. S.,4. ) (Accepted May 8th, 1984)
K<'r words': chandelier cells ecrct~ral cortex GABA ncurons -- epilcps 3
Axon terminals of chandelier cells were analvzed in monkcvs with cortical local epilepsy produced hy alumina gel to determine il this type of GABAergic terminal is lost at epileptic loci. These tcrminals form a dcnse plexus with the axon initial scgments of pyrami- dal neurons, especially those in layers II and II1. Axon initial segments of pyramidal neurons were traccd for at lea';l 4(i um in serial thin sections and hewed this point were observed to become mvelinated. In single sections, 10- 15 axon terminals were lound to form symmetric synapses throughout the entire length of the axon initial segments irom noncpileptic preparations and were observed to synapsc with only these structures and not adjacent dendritcs or spines. In epileptic cortex, the axon initial scgmenls el pyramidal neu- rons ~crc apposcd by glial profiles that contamcd cluslers of filaments typical of reactivc astrocytes. Only a few, small axon termimtls were observed to form symmetric synapses with these axon initial segments. Thus, thc chandelier cell axons appeared to degenerate in epileptic cortex. The highly strategic site of GABAergic inhibitory synapses on axon initial segments suggests that they cxcrt a strong influence on the output of pyramidal cells. The near absence of thcsc chandclicr cell axons in epileptic foci most likely conlributcs to the hyperexcitability of neurons.
INTRODt r('TI()N were classified as aspinous and sparsely spinous stel- late cells, subsequent Golgi studies have revealed It is now apparent that in some varieties of epilep- major subclasses of this category based on cell body sy, a severe deficit occurs in the cortical GABAergic shape, dendritic orientation and axonal distribu- system. Numerous studies of experimental models tion 13,2°,2<*,>. One cell type, the basket cell, was sug- that resemble post-traumatic epilepsy have shown a gested to be GABAergic because its axonal plexus loss of GABAergic terminals in epileptic foei e-' ca. A that forms pericellular nests around pyramidal cells quantitative assessment at the electron microscopic in laver V was GAD-positive in immunocytochemi- level has indicated that this loss is preferential for cal preparationsS,le,-'t,2L This identification was given GABAergic, symmetric synapses2e. Biochemical further support from recent studies that utilized col- studies also support this finding in that indices for chicine-treated immtinocytochemical preparations other neurotransmitters were not reduced as se- that displayed the cell bodies of this neuronal verely as those for GABA I,e,.<. These findings in ani- type li,iz. Since the terminals of this plexus form sym- mal models have received corroboration from bio- metric axosomatic synapses, they are readily quanti- chemical results of human epileptic loci >. Therefore, fied and were shown to be reduced by S()c~ al epilep- a loss of GABAergic terminals appears to be a hall- tic foci e-~. These data indicated that a basket cell defi- mark of epileptic loci. cit occurs at epileptic foci and as a result the laver V The initial inmmnocytochemical study on the mor- pyramidal cells at such sites arc probably more hy- phology of cortical GABAcrgic neurons revealed an percxcitable than normal. abundant variety of stained cell types distributed The present stud,,,' was undertaken to determine if throughout all layers of cortex -q. This study localized another cortical GABAergic cell lype was reduced in the synthesizing enzyme of GABA, glutamate decar- function at epileptic foci. The cell type chosen for this boxvlase (GAD). Although most of these neurons analysis was the chandelicr cell. Szcntfigothai and
('orrevl?omh'nce. Charles E. Ribak, Department of Anatomy, University of California, hvine, ('A 92717, U .S. ,,k.
Ilil06-S993i85 $03.31i ~ IriS5 El~cxicr Science Publishers B.V 252
/\rbib :~ first described this neuron in Golgi prepara-. from the embedded specimens and stained with tions. This neuron is characterized by a small soma (I.05cf toluidine blue. Specimens were oriented in with aspinous dendrites in layers I1 and 111. The chan- the ultramicrotome to obtain SeCtions thai yielded delier cell's axon forms numerous vertical chains tha! the longest segments of apical dendrites. The ~dja- appear to be apposed to pyramidal cell apical den- cent thin sections from these specimens usually had drites in light microscopic preparations Is,~3. Howe~- the most number of identified axon initial segments cr, Somogyi > subsequently demonstrated in eleclron These sections were then stained with uranyl acetate microscopic preparations that these axons form swn- and lead citrate and examined ot~ ~ × 2 mm formvar, metric synapses with the axon initial segment,~ of p~- coated slot grids with the electron microscopc, ramidal cells in layers 11 and 111. Further studies have confirmed these electron microscopic obserw~tions in RESUI:I-S numerous species anti cortical regions +'`>.> including the hippocampus ~1. The morphology of these lermi- All observations were obtained from electron mi- nals and the localization of GAD-positive reaction croscopic preparations of monkey sensorimotor cor- product within these terminals s,>.et,~: have indicated tex. The analysis in the present study was limited to that they are GABAergic. These data suggest that the axon initial segments of pyramidal cells in layers chandelier cell axon terminals exert a strong inhibito- 11, IlI and V. Since the findings from the nonepiteptic ry influence on the output of pyramidal cells. A loss hemisphere were similar to those described in a pre- of such a plexus of axons might also contribute to ~hc vious study of normal primate sensorimotor cortex>, hyperexcitability of neurons in an epileptic focus. they will be presented first. Then, the data obtained from epileptic cortex will be compared with the data MATERIALS AND METHODS from the nonepileptic cortex.
The present study utilized specimens obtained Nonepileptic cortex from 3 of the 5 experimental monkeys from a prc- A previous study 22 from this laboratory described vious study-'~. All of the monkeys had received alumi- the features of nonepileptic monkey sensorimotor na gel applications to the left cerebral hemispheres to cortex in electron microscopic preparanons. De- produce seizure foci. Two of these monkeys (animals scnptions of layer V pyramidal neurons and adjacenl 3 and 5 from Ribak et al.>) received intracortical in- neuropil regions were provided. Briefly, layer V py- jections directly into both pre- and postcentral gyri. ramidal neurons have multiangular-shaped somata The remaining experimental monkey (animal 2 from formed by an array of basal dendrites, a single apical Ribak et al. ~3) had an injection of alumina gel limited dendrite and a single axon usually located between to the subarachnoid space in the area of the central the basal dendrites. Much of the soma is occupied by sulcus. Electrocorticography of all experimental ani- a large, rounded nucleus wh~se nucleoplasm con- mals verified epileptic foci > and subsequently, the tains a relatively homogeneous sprinkling of electron- monkeys were fixed by intracardiac perfusions of a opaque chromatin and a centrally located nucleo- mixture of two aldehydes. The fixative solution con- lus ]~.22. Pyramidal neurons contain numerous orga- tained 4% paraformaldehyde, 0.1% glutaraldehyde. nelles in the perikaryal cytoplasm including both free and 0.002% CaCI~ in a 0.12 M phosphate buffer. ribosomes and cisternae of granular endoptasmic re- Blocks of cortical tissue were obtained from the uculum. Terminals thai form axosomatic s}napses epileptic focus and the homologous area in the con- with these neurons make only ,~ymmetnc synapses. tralateral nonepileptic cortex. All blocks were sec- Pyramidal neurons in layers II and III are somewhat tioned on a Sorvall TC-2 tissue sectioner at a thick- smaller than those in layer V. bul they share the same ness of 150 urn. Specimens that contained the entire ultrastructural characteristics, Since our prewous ul- cortical thickness were cut from these sections. trastructural studv 22 provided details on the somata These specimens were postfixed in 2% OsOa for I h. and dendrites of pyramidal neurons and the axon ter- dehydrated in ethanol and embedded in Epon. For minals thai formed synapses with these structures, light microscopy, semithin 1 um sections were cut the present study will involve primarily an analysis 253 of the synaptic contacts \`` itll aXOil initial Scglnenls of cells have somewhat fewer terminals that synapse pyranlidal netlrons. with their initial segments. These findings are consis- The axon initial segmeill {irises from the axon hill- tenl with a previous quantitative description of axon ock region of the neuronal cell hod\'. Peters ct ill. > initial segments of pyramidal ncclions ill prmlate seN- provide a COlllplctc description of the feattll-es of soIimotor cortex 2s. In lhat sttldv, Sloper and Powell > ~lxon initial segments. Such features will bl-icfh' bc reported that the total length of all axon initial seg- describcd lind attention will bc given to those slttlC- nlenl of a pyralllidal cell in ]avers I1 or I11 received ttlres and rehiiionships th{it appe:,lr differcnt in tilL' 13.4 synapses, on a\erage, while pyramidal ceils ill epileptic cortex. la\er V were contacted by 4.8 synapses per :.lX()ll ini- Axon initial segments usually appear al tile base of tial segment. It is iillportanl to note lllal these termi- file p_vranlidai cell bed) (Pig. 1). The botlnd{lr\ be- nals are not distributed evenl} along tile lengfll of the tweeI1 those two portions of tile lleuroll is :.lpp{ii+ellt lit axon initial segment. A gradienl exists in that fewer 1o+`` magnification because tile cisternao of the ~rantl- terminals arc present adjacent to the pl-oxinlal axon hlr endoplasmic rcticulum or Nissl bodies that reside and nlore {ire fotlnd associated with tile dislal portion within the soma do i1Ot enter the ~iX,Oll. The {iXOll hill- of the aXOll initial segment. This distrihution is simi- ock region is coniciil. \vhcl+eas the axon initial seg- [;.tr to file one described for {iXOllS of chandelier cells illel]l has ii const{inl cvlmdrical shape that is main- studied ill combined (hHgi-elociron microscopic tainud thl-oughc}ul its lcllgth. This ]atlcr lealul-c p re pa rat i o 11s" ,20.2u.31i stands in COlltrasl to dendrites which tistiallv Ila\C dif- fereill cross-section diameters over lheir entire f:t~ileptic corte.k length. lhe most disial part of lhe initial segillenl is Thin sections of cortical tissue lroln the epileptic cilarilcterizcd b\ the paranodal region of il in\olin focus disphiy similar structures to those observed in sheath (Figs. I alld _~ I. Thcsc major features t)f aXOll the nonepileptic cortex. Pyramidal ilCUl-Ons are read- initial SeglllelltS
255 in layers 11 and Ill of ten have no terminals that form cortical neuronal type, the chandclicr cell <>.>,>. synapses wth their axon initial segments. However, Although a similar study' of the chandelier cell has the larger pyramidal cells in layer V including two ex- not been made in the monkey scnsorimotor cortex, it amined Betz cells have the highest number of initial is likely that this cell exists in this region because So- segment terminals and this finding probably results mogyi et al. x<> have dcmonstratcd chandelier cells in from the much larger size of their axon initial scg- the motor cortex of cat alld in the monkevs visual mellts. cortex. In addition, the distribution of axon tcrrninals The second difference observed in these epileptic alongside the axon initial segments of pyramidal cells preparatives is the increased number of astrocvtic in laycrs 11 and III of the monkey sensorimotor cor- processes that lie adjacent and orient parallel to the tex es is very simihtr to the distribution of chandelier axon mitial segments. Most of these astrocytic proc- cell axons. Taken together, these dala mdicate that esses are packed with nunlerous filaments (Fig. 7). most of the axon terminals that synapse with axon ini- Such processes are found adjacent to the somata of tial segments of pyramidal neurons arise from chan- these same pyramidal cells as described previously -~-'. delier cells. Often, the same astrocytic process will appose a por- Recent resuhs from immunocvtochen~ical studies tion of the cell bodx, the axon hillock and a l(I l;m which localize lhe GABA synthesizing enzyme, length of the axon inilial segment (Figs. 5 and ~3). Al- GAD, have indicated that chandelier cells tire though this orientation preference for the hmgitudi- GABAergic because most terminals that form sym- nal axis of the axon initial segment is observed most metric synapses with axon initial segments contain frequentl,v (Fig. 7), somc glial processes arc oriented GAD imn3unoreactivitv s.il,>. In fact. Freund et al. s transverse to this axis (double arrow in Fig. 4). This have provided direct cvideilce for this notion by com- apposition of glial processes is usually only one proc- bining Golgi impregnation with immuriocvtochemis- ess thick. However, inultiple layers of glial processes try m the same preparation. In one fortuitous case are found adjacent to the larger axons of Betz cells. from cat visual cortex, they demonstrated that 11 Golgi-imprcgnated bout(ms derived from a single DIS(t SSI()N chandelier cell contained OAD-posilive reaction product. The highly,' strategic site of these GABAer- The major finding of this electron microscopic gic synapses on axon initial segments suggests that study is that most of the axon terminals which form they exert a strong inhibitory effect on the pyramidal symmetric synapses with axon initial segments of py- neurons which they contact. The large loss of these ramidal netirons are lost at epileptic foci in monkeys. GABAergic synapses in epileptic loci indicates a sig- Although previous studies have demonstrated a loss nificant reduction of inhibition because (iABA has an of axosomatic synapses in epileptic loci 3-7,2-~,35. this inhibitory action tm cortical neuronsi4 i-. Thereforc, report is the first to document a loss of these initial a probable result of this loss is a hypercxcitability of segment symmetric synapses that are formed by termi- neurons at the epileptic focus. nals mainly derived from the chandelier cell (Fig. 8). Thcse findings are consistent with our previous re- Previous studies that utilized a combined Golgi- sults that dcmonstrated a severe loss of GAI)-posi- electron microscopic method have shown that axon tive axon terminals tit sites of alumintlnl gel-educed terminals forming synapses with laver II- I11 pyranli- epilepsy3L In addition, lhey coniplemeilt ~i previous dal neuron axon initial seglnents arise from a specific electron microscopic studv of ¢ixosonlalic and axe-
Fig. 1 [=lcctron micrograph el a laver Ill pyramidal soma and ~lxon initial scgillCtll lrom a nol'inal, nonepilepiic hemisphere, l'hc soma displap, ci round nucleus (iN) and several cislcrnao o[grariular cndoplasmic roticuhnn (E). Ihc axon hillock (H)is fori'ncd al the, base el lht.' SOll]{I:l[ItJ gives rise ~o the axon initial segment (nrrows) which Cili3[It' followed to the point whert., ii acquires ~i mvclill sheath. This ~lXOrl is coiltactcd b ~. ntllllc:l-ous Icrnlinals that form s,vmmctric synapscs (sce Fig. 2 I. ,x:500(I Fig. 2. Enlargcmcnt el the distal portion of lhc axon initial segmcnl shmvn m Fig. I. Numclous tcrmmals {|) arc located alongside this axon and manx term sx mmctric !-Vll~lpsos arrow, s). This axorl ht.'C(/ll3CS t31velimltcd at the [~ttl[Olll o[ the i~hotomiciogrlq~h, x I ",Ylltl. :5, 257 dendritic SVlnlnctric synapses in the same nlonkev It is likely that the chandelier cells ha',e degener- preparations-'-L This latter study-'-" showed an 8()(~: atcd in these epileptic foci because most of their axon loss of axosonlatic symmetric synapses and a 5IV<; terminals are not present. Other cvMcnce to support loss of axodendritic synapses in cortical layer V of ep- this notion is derived from a report that dcmon- ileptic loci. Similar reductions in these two types of str:.tted neuronal degeneration in Fink -Hcimer prep- synapses were also obscrved in the superficial corti- aralions of alumina gel-treated epileptic i11Ollkevs'. cal layers where the present study was undertaken. It A common sequel to degeneration of lleur('Hls is the is interesting to note that the loss of symmetric syn- hypertrophy and proliferation of gila i". The results of apses with axon initial seglncnts was lnorc similar to thc prcsent study showed till illcreasc of glial profiles lhc larger reduction of axosomatic synapses than the adiacent to the axon initial segments of pyramidal smaller loss of axodendritic symlnetric synapses. neurons at epileptic loci. In fact, the s;tmc glial pro- Thercf(tre, tile most severe loss of GABAergic. sym- file was often found to apposc portions of the soma. metric synapses at epileptic foci suggests a degenera- axon hillock and axon initial segment (Fig. h ). Thcsc tion of two cortical (]ALIA cell types, basket and findings are consistent with previous results that chandelier cells. showed tip to a 5(Vi increase in glial proliferation at The terminals of basket and chandelier cells have epileptic foci and indicated degellCiution of net> SOIllC similarities. They send their axons to the soma i(/ns x.~'-111.22`N,~5, a likely cause of this prolifcration of and axon initial segment, respectively, of pyramidal glia alongside axon initial segments itl the alumina gel neurons, alld these two sites are most strategic for the model of epilepsy is the loss of GABAclgic axon ter- control of action potential generatiOl-l. Thus, both minals derived from chandelier cells. cell types can dramatically influence the activity and The results of this study add further support for output of the cortical proiection neurons. Second, the GABA hypothesis of epilepsy that suggests a re- these two types of terminals average more than onc duction of GABA-mcdiated inhitqtion ill the brain mitochondria per terminal. This fact has hecn doen- inav cause focal epilepsy e~. Biochemical sttidics of nlented in nlanv previous studies for chandelier cell cortical epileptic loci from hulnans and anilnal mod- terminals (sec also Fig. 3) and in a quantitative study els have shown a sclective loss of (iAI) activity, for the basket cell terminals-~-~. Thus, the chandelier GABA concentration and (IABA receptor bind- cell axonal plexus that forms initial segment Svlnnlet- in,,I-:.l".~a These data are consistent with thc results ric synapses lnay provide a tonically active inhibition from imlmUlOC\ tochemical studies that show a reduc- of cortical pnljcction neurons in a way proposed orig- tion of (;AD-conlaining terminals at epileptic imlllx for the pcriccllular basket cell axonal plcxus :-~. foci<-':L Mcldrunli7 and Krnjcxid 1" ha~,c sumnlarized Furthcr support for this similar function of thesc txvo in review chapters the physioh)gical and pharmaco- cell types arises fronl their similar degree of malfunc- logical data that sho,a' a (}ABA mechanism is in- tion in epileptic foci. Although a thor(/ugh quantita- wllved in epileptic activity. Therefore, tl selective re- tive assessment was not made in the present stud\, duction of GABA-lnediated inhibition appears to he one can conchlde from the electron microscopic data a hallmark (if cortical epileptic loci. A possible cause of initial segment synapses in epileptic foci that they for this loss of (}ABAcrgic neurons tit epileptic loci is are reduced m lnagnitudc to a similar amount as the hypoxia -~2,->'. Further studies nlust determine if the axOSOlllatic synapses as prc~ iously reportede-L
Fig. ~. ElllargCIllCilI ()l aoolhct axoll initial segment from a noncpilcptic hemisphere. The two characteristic feature', ol ~Lxon initial segments tire shown: a fasciculation of microtubules (M) and a dense undercoating that is continuous with the coaling o',cr a pinocytot- ic vesicle (V). Three terminals tire shown to form symmetric synapses (arrows). × 25,000.
Fig, 4. Electron nlicrograph of ~il1 axon initial segment oI a laver Ill pyramidal nCUFOnfrom an epileptic focus, lhc axon hillock (1 l) is located at the top oI this micrograph and the initial scgmcnt (arrows) in included in its cntirct~ up to and hlcludm~2 lhe point (',ffrow- heads) at which it HCqLliIC5 {I myelin sheath Both proximal and distal to this point, the axon displays slight dilation~ lh:lt Icalur¢ 11111iiv disoriented mitoch(mdrm NtlmCIOUS large profiles of reactive ~ISIIOCVtCS ( A ) life 1(/1111¢] llcar fin', axon. ,~ >000 S~ 259
NONEPILEPTIC EPILEPTIC GABA hypothesis is responsible for other types of epilepsy;, such as those with genetic causes.
A('KN()WLEDGEMENTS
This work was supported bv NIH Grant NS-15669 and a Fellowship from the Klingenstein Foundation. The author gratefully acknowledges M. Brundage and Y, Jhurani for technical assistance, A. B. Harris IS for the monkey brain specimens, and N. Sepion for secretarial assistance.
Fig. 8. Summary diagram that shows the diftcrcnces observed for axon initial segments from nonepileptic and epileptic prepa- rations+ A reduction in the number of terminals that form sym+ metric synapses and an increase in gila arc the two major differ- ences. Some axon initial segments also displa} dilations of their distal portions. Such a loss of inhibitory, GABAergic terminals [] GLIA [] TERMINALS THAT FORM could contribute to the hyperexcitability of the ncunms at epi- [] SYMMETRIC SYNAPSES leptic loci (see Discussion ).
REFEREN([!S cerebral cortex. In F. O. Schmidt, F. (;. Worden, G. Adcl- man and S. G. Dennis (Eds.). The Organization of the Cere- 1 Bakav. R. A. E. and Harris, A. B., Neurotransmittcr. re- bral ('orlex, MI,T. Press, Cambridge, MA, 1981, pp. ceptor and biochemical changes in monkey cortical epilep- 325-345. tic loci, Brain Research, 20~~ ( 1981 ) 387-404. 6 Fair~Sn, A. and Valverde, F., A specialized type o[ neuron 2 Balcar, k..1., Pumain, R., Mark, J., Borg, J. and Mandel, in the visual cortex of cat. A Golgi and electron microscopic P, (iABA-mcdiated inhibition in the cpileptogenic focus, a study of chandelier ceils, J. uomp. "Veered., 194 (1980) process which may be involved in the mechanism of cobalt 761-779. induced epilepsy, Brain Re.search, 154 (1978) 182-185. 7 Fischer, J., Electron microscopic alterations m the vicinity 3 Brown, W. ,I.. Structural substrates of seizure foci in the hu- of epileptogenic cobalt-gelatine necrosis in the cerebral man temporal lobe. In M. A. B, Brazier (Ed.). Epilepsy. hs cortex of the rat, Acta Neuropathol., 14 (1969) 201 214. Phem?mem¢ z~l Man, Academic Press, New York, 1973, pp. 8 Freund, T. F., Martin, K. A. C., Smith, A. D. and Somo- 33~; 374. gyi, P., Glutamate decarboxylase-immunoreactive termi- 4 Brown. W. J. and Babb, T. L., Neuronal, dendritic and vas- nals of Golgi-impregnated axoaxonic cells and of presumed cular profiles of human temporal lobe epilepsy correlated basket cells in synaptic contact with pyramidal neurons of with cellular studies in vivo. In A. V. Delgado-Escueta, A. the cat's visual cortex, J. comp. Neurol., 221 (1983) A. Ward. Jr. and D. M. Woodbury (Eds.), Basic Mecha- 263-278. ;U,~tn.~ ~1 the k:pilepsies, Raven Press, New York, 1985, in 9 Harris, a. B., Degeneration in experimental epileptic fool, press. Arch. Neurol., 26 (1972) 434-44~,L 5 Emson, P. (" and Hunt, S. P., Anatomical chemistry of the 10 Harris, A. B.. Cortical neuroglia m experimental epilepsy,
Fig. >. Another clcctron micrograph froin epileptic cortex that shows a layer II1 pyramidal soma, proximal apical dendritc (AP) and axon initial segment (arrows). The soma contains a large nucleus (N) and a perikaryal cytoplasm that contains the typical organclles as well as :m abundance ol lipofuscm granules. The initial segment is apposed to a gliaI profile that also contacts the axon hillock and soma. × 13.000
Fig. 6. Enlargement o[ the initial segment (IS) and adjacent reactive glial process (A) shown m Fig. 5. ×23.1)()0. Fig. 7. En]argcment of a distal portion of another axon initial segment (IS) from a layer It[ pyramidal neuron from an epileptic prepa- ration. 1his axon displays fasciculation of microtubules (M) and a dense coating beneath the axolemma (arrov,). Profiles of reactive astroQtes (A) that contain filaments lie adjacent to this axon initial segment. Axon terminals that form symmetric synapses are not present. × 2g,()O0. 261l
Exp. Neurol.. 49 (1975) 691-715. sites of focal epilepsy. Science 205 ( 19791 211-214. 11 Hendry, S. H. C,, Houser, C. R., Jones, E. G. and 24 Ribak. C E. and Reiffenstein. R, .t Selective inhibitor~ Vaughn, J. E., Synaptic organization of immunocytochemi- synapse loss in chronic cortical slabs: ,~ morphological basis cally identified GABA neurons in the monkey sensory-mo- for epileptic susceptibility. Canad. J l'hvsiol PharmacoL. tor cortex, J. Neuroo, tol,. 12 ( 19831 639-660. 60 ( 19821 864-871/ 12 Houser, C. R.. ttendry, S. H. C., Jones, E. G. and 25 Roberts. E.. Epilepsy and amiepitepnc clrugs: a speculauve Vaughn, J. E., Morphological diversity of immunocytoehc- synthesis. In G. H Glaser. J K Penr~ and D. M Wood- mically identified GABA neurons in the monkey sensory- bury (Eds.). Antiepileptie Drug.s Mechanisms ," A¢~lon. motor cortex, J. Neuroeytol., 12 (1983) 617-638. Raven Press, New York. 1980. pp. 6~7 -713. 13 Jones, E. G., Anatomy of cerebral cortex: columnar input- 26 Sloper. J. J . Johnson. P. and Powcll, T. P. S.. Selective de- output organization. In F. O. Schmidt. F. G. Worden, G. generation of mterneurons in the motor cortex ol infanl Adelman and S. G. Dennis (Eds.), The Organization of the monkeys following controlled hypoxia: a possible cause of Cerebral Cortex, M.I.T. Press, Cambridge, MA, 198I, pp. epilepsy, Brain Research. 198 (19803 204-.209. 199-235. 27 Sloper, J. J. and Powell. T P. S.. Observations on the axon 14 KrnjeviC K., Chemical nature of synaptic transmission in initial segment and other structures m the neocortex using vertebrates, Physiol. Rev., 54 ( 19741 418-540. conventional staining and ethanolic phosphotungstic acid. 15 Krnjevi~, K., GABA-mcdiated inhibitory mechanisms in Brain Research. 511 [ 1973~ 163 - 16~ relation to epileptic discharges. In H. H. Jasper and N. M. 28 Sloper, J. J. and Powell. T. P S . ~ study of the axon mitia van Gelder (Eds.), Basic Mechani.sms of Neuronal Hyper- segment and proximal axon ot n~ urons m the primate mo- excitability, Alan R, Liss, New York, 1983, pp. 249-28(t. tor and somatic sensory cornces. Phil Trans roy. %e 16 Lloyd, K. G., Munari, C., Bossi, L., Bancaud, J., Talai- 1,ond.. 285 ( 19751 173-197 rach, J. and Morselli, P. L., Biochemical evidence for the 29 Somogyi. P., A specific "axo-axonal'mteHleuron in the visu- alterations of GABA-mediated synaptic transmission in hu- al cortex of the rat, Brain Research. !3(~ 19771 345-350 man epileptic loci. In P. L. Morselti, K. G. Lloyd, W. 30 Somogyi, P.. Freund, T. F. and Cowe~~. A.. The axo-axomc L6scher, B. Meldrum and E. H. Reynolds (Eds.), Neuro- interneuron m the cerebral cortex o[ the rat. cat and mon- transmitters, Seizures and Epilepsy, Raven Press, New key, Neuros~ i.. - ( 19821 2577-2608. York, 1981, pp. 331-354. 31 Somogyi, P., Nunzi. M. G.. Gorio. A. and Smith. A. D.. A 17 Meldrum, B. S., Epilepsy and 7-aminobutyric acid-me- new type ot specific interneuron n the monkey hippocan> diated inhibition, Int. Rev. Neurobiol., 17 (1975) 1-3f~. pus forming synapses exclusively with the axon initial seg- 18 M611er-Paschinger, I.-B., T6mb6t, T. and Petsche, tt.. ments of pyramidal cells Brain l,~es'earch 259 ~t9831 Chandelier neurons within the rabbits' cerebral cortex. A 137- 142 Golgi study, Anat. Embrvol.. 166 (1983) 149-154. 32 Somogyi, P.. Smith. A D.. Nunzi. M G.. Gono A.. TaM> 19 Peters, A., Palay, S. L. and Webster, H. de F., The Fine gi, H and Wu. J .-Y.. Glutamate decarboxvlase immunore- Structure of the Nervous System: The Neurons and Support- activity in the hippocampus of the cat. Distribution of lm- ing Cells, W. B. Saunders, Philadelphia, 1976. munoreacuve terminals with special reference to the axon 20 Peters, A., Proskauer, C. C. and Ribak, C. E., Chandelier initial segment of pyramidal neurons. J. Neurosci.. 3 (I983, cells in rat visual cortex, J. coral?. Neurol., 2(16 (19821 1450-1468. 397-416. 33 Szent~,gothai. J. and Arbib. M.. Conceptual models of neu- 2t Ribak, C. E., Aspinous and sparsely-spinous stellatc neu- ral orgamzation, Neurosci. Res. Prc,~r. Bull,. 12 (1974~ rons in the visual cortex of rats contain glutamic acid decar- 307-5l(t boxylase, J. Neuroo, tol., 7 (1978) 461-478. 34 Van Gelder. N. M. and Courtols_ A ('lose correlatmn be- 22 Ribak, C. E., Bradburne, R. M. and Harris, A. B., A pref- tween changing content of specific anmro acids in epilepto- erential loss of GABAergic, symmetric synapses in epilep- genie cortex ot cats. and severity ~! epilepsy, Brain Re- tic foci: a quantitative ultrastructural analysis of monkey search. 43 119721 477-484 neocortex, J. Neurosci., 2 (19821 1725-1735. 35 Williams. R. S.. Lott, I. 1.. Ferrantc. R. J. and Caviness, 23 Ribak, C. E., Harris, A. B., Vaughn, J. E. and Roberts, V. S.. It. The cellular pathology o! neuronal ceroid lipo- E., Inhibitory, GABAer~ic nerve terminals decrease at fuscinosis, Arch Neurol.. 34119771 298-305