Thalamic retrograde degeneration following cortical injury: An excitotoxic process?

:'v'w_m~rcie,~c~ Vol. 35. No. Printed m Great Britain

0306-4522/90 $3.00+0.00

3, pp. 525-550. 1990

Pergamon Press plc ? 1990IBRO

THALAMIC RETROGRADE DEGENERATION FOLLOWING CORTICAL INJURY: AN EXCITOTOXIC PROCESS? D. T. ROSS*? and F. F. EBNER~ *Head Injury Center, Division of Neurosurgery, University of Pennsylvania, 3400 Spruce St, Philadelphia, PA 19104, U.S.A. :Center for Neural Science. Box G, Brown University. Providence. RI 02912, U.S.A. Abstract-Traumatic or stroke-like injuries of the cerebral cortex result in the rapid retrograde degeneration of thalamic relay neurons that project to the damaged area. Although this phenomenon has been well documented, neither the basis for the relay neuron’s extreme sensitivity to axotomy nor the mechanisms involved in the degenerative process have been clearly identified. Physiological and biochemical studies of the thalamic response to cortical ablation indicate that pathological overexcitation might contribute to the degenerative process. The responses of thalamic projection neurons, protoplasmic astrocytes, and inhibitory thalamic reticular neurons in adult mice were examined from one to 120 days following ablation of the somatosensory cortex as part of an investigation of the role of excitotoxicity in thalamic retrograde degeneration. The responses of thalamic neurons to cortical ablation were compared with those produced by intracortical injection of the convulsant excitotoxin kainic acid, since the degeneration of neurons in connected brain structures distant to the site of kainic acid injection is also thought to occur via an excitotoxic mechanism. Within two days after either type of cortical injury, protoplasmic astrocytes in affected regions of the thalamic ventrobasal complex and the medial division of the posterior thalamic nuclei became reactive and expressed increased levels of immunohistochemically detectable glial fibrillary acidic protein. Within the affected regions of the ventrobasal complex an increased intensity of puncta positive for glutamate decarboxylase immunoreactivity. presumably due to an increase in its content within the terminals of the reciprocally interconnected thalamic reticular neurons, was also evident. These immunohistochemically detectable alterations in the milieu of the damaged thalamic neurons preceded the disappearance of the affected relay neurons by at least two days following cortical ablation and by seven to 10 days following intracortical kainic acid injection. Regions of the thalamus containing reactive astrocytes corresponded very closely to the regions undergoing retrograde degeneration. Protoplasmic astrocytes in these areas remained intensely reactive up to 60 days after cortical injury. Levels of glutamate decarboxylase were only transiently elevated in the degenerating regions of the ventrobasal complex following cortical ablation and returned to normal by I4 days. Increased glutamate decarboxylase immunoreactivity was transiently seen through the entire ventrobasal complex following intracortical kainic acid injection but was markedly more intense in degenerating regions. These patterns of labeling did not return to normal until 50 days after intracortical kainic acid injection, well after the death of the relay neurons. Cortical ablation and intracortical kainic acid injection produce similar alterations in thalamic neuronal and glial populations. The response of thalamic protoplasmic astrocytes to cortical injury may precipitate secondary pathological alterations that directly contribute to the overexcitation of affected relay neurons. Increased activity of inhibitory thalamic reticular neurons may contribute to the degenerative process by paradoxically increasing the excitability of relay neurons in degenerating thalamic regions. These results are consistent with the hypothesis that the thalamic degeneration induced by cortical ablation and intracortical kainic acid injection may share a common excitotoxic mechanism,

Some

express

neurons

in the CNS of adult mammals

a capacity

for sustained

clearly axonal regeneration

following injury.‘6~‘5~6x~‘06~‘“y~‘~5 All neurons may possess an intrinsic capacity for axonal regeneration but the expression of this capacity as sustained axonal regeneration is the exception and not the rule following axotomizing injury of adult mammalian CNS neurons. The severity of alterations that occur in both the internal and external milieu of many CNS neurons following axotomy may prevent the expression of their capacity for sustained axonal regeneration. The rapid retrograde degeneration of specific sensory relay neurons in the thalamus following cortical injury reflects an extreme sensitivity to axotomy that effectively precludes axonal regeneration.’ Accordingly, prevention of retrograde degeneration may be

?To whom

correspondence should be addressed. CL, centrolateral thalamic nucleus; GAD glutamate decarboxylase; GFAP, glial fibrillary acidic protein: HRP, horseradish peroxidase; KA, kainic acid; LGNd, lateral geniculate nucleus, dorsal division; LP, lateral posterior thalamic nucleus; MI, primary motor cortex; NMDA, N-methyl-D-aspartate; Porn, posterior thalamic nucleus, medial division: RT, thalamic reticular nucleus; SI. primary somatosensory cortex; VB. thalamic ventrobasal complex; VL, ventrolateral thalamic nucleus; VM, ventromedial thalamic nucleus; VPL, ventroposterior lateral thalamic nucleus; VPM, ventroposterior medial thalamic nucleus.

Ahhreciu/ions:

525

>‘(I

I>. T.

Ross and F. t-‘. EHNEK

suthcient to unmask the regenerative capacity of damaged thalamic relay neurons. Therefore. the crit~cal question in the study of thalamic regenerability is clearly “what causes the retrograde degeneration of axotomized thalamic relay neurons?” The phenomenon of thalamic retrograde degeneration following cortical injury has been effectively exploited in classical ncuroanatomical studies as ;I method for characterizing patterns of thalamocortical connections using Nissl stains.“” I”’ “‘.‘u The Wallerian (anterogradc) degeneration of corticothaiamic axons which occurs soon after cortical ablation has also been used as a technique for identifying corticothalamic projections in light microscopic studies employing reduced silver impregnationl”xK IO?and corticothalamic terminals in ultrastructural studies,‘“.“‘.X’.lJ’ Concurrent with the pcrikaryal degeneration of axotomized relay neurons the proximal segments of damaged thalamocortical axons exhibit what has been described as “indirect” Wallcrian degeneration, characterized by rapid structural breakdown of the entire axon from the initial segment on the perikaryon to the severed end.“*.“” The earliest events observed in ultrastructural analyses of thalamic retrograde degeneration include degeneration of corticothalamic terminals. swelling of thalamic astrocytes. and dissolution of rough cndoplasmic reticulum within the perikarya of affected ncUronS.:.h “I 1\1,IJIAs the degenerative process continues affected relay neurons show one of two reactions. tither an atrophied electron dense perikarya tilled with mitochondria, or an electron luccnt lysis 01 membrane bound organelles and the plasma mcmbranc.’ Despite the extensive description of this phenomenon. the cellular mechanisms that constitute the basis of the thalamic relay neuron’s extreme scnsitivity to axotomy remain poorly understood. Analyses of the metabolic and electrophysiological alterations which occur in the thalamus in response to cortical injury have not been as thoroughly documented as the morphological changes. Those physiological and biochemical alterations which have been observed may provide important clues about the nature of the degenerative process. There is now considerable evidence implicating an excitatory amino acid (glutamate. aspartatc or an analog) as the ncurotransmitter used by the terminals of both major afferent projections to the thalamic ventrobasal complex (VB), the corticothalamic projection from layer ‘4.WilK114 lii.l‘ii VI of the somatosensory cortex 229.41LJi and the ascending lemniscal projection from the gracilc, cuneatc and trigeminal nuclei.” ‘h.“‘X.“5 A marked increase in the rate of spontaneous activity and an enhanced response to sensory evoked stimulation can be recorded from axotomized neurons in the VB within a few hours after ablation of the primary somatosensory cortex (SI) in adult rats.“.“” Both the axotomized VB neurons’ elevated level of spontaneous activity and their enhanced responsiveness to sensory stimulation are completely and reversibly

blocked by the excitatory amino acid antagonist. kynurenic acid.“” Cortical ablation also results in a marked decrease in thalamic high-affinity uptake 01 glutamate and aspartate.‘.‘45 The failure of mechanisms which remove excitatory amino acids from the milieu of affected thalamic relay neurons is likely to precipitate their over-excitation, similar to that manifest during pharmacological blockade of thalamic glutamate uptake.” Is it possible that overexcitation is the driving force in the process of thalamic retrograde degeneration following cortical injury’! Application of convulsants to the surface of the somatosensory-motor cortex produces epileptiform tiring of thalamic relay neurons in the ventrolatcral thalamic nucleus (VL) and the ventroposterior lateral thalamic nucleus (VPL)~‘.‘~~h”~“.“”and results in the degeneration of the relay neurons” via an cxcitotoxic mechanism.‘” Intraparenchymal injection of kainic acid produces neuronal degeneration both at the injection site (“local cxcitotoxic degeneration”“.x”‘h ‘)‘) and in distant structures related to the injection site along seizure pathways (“distant excitotoxic degeneration”“.“7.“0.‘4h ). The present study was designed to compare the reaction of thalamic relay neurons. inhibitory neurons, and protoplasmic astrocytes following cortical ablation with the reaction ol the same elements following intracortical injection of the convulsant excitotoxin kainic acid. EXPERIMENTAL

PROCEDLRES

Adult female BALBIc

mace (two to three months old, with 0.4 ml of a 1.25%tribromoethanol solution and prepared for stereotaxic surgery. Parts of the left parietal. frontal. and temporal bones were removed to expose the ST. In one group of animals (II = 30) the left St was ablated by subpial aspiration. A second group (II = 28) received intracortical injections of the excitotoxin kainic acid (KA) into their left SI. The KA (0. I b(I ol a 0.5% solution in lactated Ringer) was injected with a 30.gauge needle inserted into the somatosensory cortex nearly parallel to the pial surface.

I9 24 g) were anesthetized

Following survival intervals from one to 120 days ammals were given a lethal dose of sodium pentobdrbitat (Nembutal) and perfused transcardialty with a 3.7”/0 formalin. 0.5”% lint salicylate solution (pH 6.3) followmg the protocol of Mugnaini and Dahl.” Corona) sections through the thalamus and damaged cortex were cut at 40pm on a cryostat and saved as three or four parallel sets of serial sections. One set w’as stained with the Nissl stain Cresyl Violet, and two sets were processed for immunohistochemistry. Sections reacted for glial fibrillary acidic protein (GFAP) immunohistochemistry were incubated in antiserum at a I : 750 dilution and processed according to the protocol of Leavitt and Rakic.” Sections reacted for glutamate decarboxylase (GAD) immunohistochemistry were incubated in antiserum at a I : 1500 dilution and processed according to the protocol of Mugnaini er ~1.~‘)For some cases a fourth set of sections was reacted for cytochrome oxidasc histochemistry using the method of Wong-Riley.“’

The neocortex. strtatum. hippocampus rpsilateral to the lesion and the neocortex

and thalamus contralateral to

Thalamic retrograde degeneration following cortical injury

Fig. I. A. Coronal section through the normal adult mouse sensorimotor cortex reacted for cytochrome oxidase histochemistry. The barrels in layer IV of the somatosensory cortex are darkly labeled. B. Corresponding, region of the contralateral cortex 14 days after ablation. The lateral-most two barrels spared by the lesion exhibit an intense cytochrome oxidase labeling. Scale bar for A and B is in B. C. Reactive astrocytes within the neocortex lateral to the lesion site, seven days after ablation, GFAP immunohistochemistry. Reactive astroghosis extends from the margin of the lesion cavity medially and laterally throughout ihe entire affected hemisphere. D. Reactive astrocytes within the neocortex ipsilaterai to the lesion site. 14 days after cortical ablation, GFAP immunohistochemistry. GFAP+-reactive astrocytes are confined to the margin of the lesion cavity and few are seen in adjacent undamaged cortical regions. Scale bar for C and D is in D.

527

I>. ‘T. Ross and 1.

for each case were microscope for indications of sections were examined for the

the Ica~on 111 all sets 01‘ parallel sections cwmmcd under the light dcpcncration Nissl-stained

prcsencc of degenerating neurons and gliosis. Sections proaged for GFAP immunohistochemistry were examined for lhc presence o!’GFAP-labeled (GFAP ) astrocytes around the Iebion sltc and m distant structures. Sections reacted for C;AD imlnunohistochemistry were esamined for changes in the number of GAD ’ neurons and in the intensity of GAD Ilnnrun~)labeling in the remaining neurons and punctae in lhc cortex and the thalamus. Sections I-acted for cytochrome oxldasc histochemistr) were cxamincd for signs of lncrcascd or dccrcased labehng. indicating increases or dccrcascs in (hc level> of oxidativc metabolic activity in region\ around the lesion sltr and In the diatanr structures.

In all cases the cortical ablation included the caudal and medial regions of the SI. In many cases the ablation extended into the SILprimary motor cortex (Ml) overlap LOX of the hindlimb cortex.” The most lateral and anterior part of the SI was often spared. leaving cytochrome oxidase-labeled barrel formations intact on the lateral border of the lesion cavity (Fig. I B). Damage to the cortical white matter including the optic radiations was present in some cases. but none of the cases included in this analysis showed evidence of direct damage to the underlying striatum or hippocampal formation. Within the cortical lesion site neuronal loss extended from the cortical surl‘ace to the white matter throughout the lesion. No neuronal loss was detected in cortical areas bordering the lesion at any time from two to I80 days after ablation. GFAP ’ -reactive astrocytes were seen throughout the entire neocortex ipsilateral to the lesion in cases examined from two to IO days after cortical ablation (Fig. IC). Cases examined 14 days or longer after ablation showed evidence of scar formation at the lesion margins, manifest in Nisslstained sections as an increased density of nonncuronal cells and in GFAP-reacted sections as an increase in reactive astrocytic processes (Fig. ID).

Cortical lesions produced by KA injection consisted of a needle track. a central area of complete neuronal loss, and a surrounding region of partial cell loss. The extent of the lesion varied according to the method of KA delivery. In cases where KA was delivered in three small increments of 0.03 111 with 5-min intervals between each injection. small focal lesions were produced (Fig. ZAPC). The central area of these lesions extended 100 -200 Itrn radially outward from the needle track. The surround region continued for 500-70O/tm from the needle track. Partial neuronal loss in surround regions was very striking in sections reacted for GAD immunohistochemistry because the density of GAD ’ neurons was markedly reduced in all cortical layers (Figs 2A and 3B. C). Most of the remaining GAD’ neurons in the surround region exhibited intensely dark immuno-

1. riHI1.R

labeling of their perikarya, proccsscs, and punctae (Fig. 39). Many of these perikarya were also markedly hypertrophied (Fig. 3C). No loss of GAD ’ cells was evident in cortical regions surrounding Icsions produced by ablation. Extensive lesions were produced when 0.1 ,uI KA was injected rapidly as ;I single bolus into the SI. These lesions were charactcri/cd hy a very large central area within the SI (Fig. 2E and F) and a surround region that cxtendcd throughout nearly all of the remaining SI. rostrally into the motor cortex, laterally into the SII. caudallq into the posterior part of the parietal cortex. and medially into the cingulate cortex. In some cases with extensive lesions the KA had apparently diffused rostrally and ventrally into the dorsal striatum. where ncuronal loss was also evident (Fig. 2E and F). GFAP ’ astrocytes were evident throughout almost the cntirc ipsilateral cortex from two to IO days aftcl either type of cortical injury (Fig. 2C and F). The exception was the center of extensive KA lesions, where cortical protoplasmic astrocytes did not cxpress GFAP at any time after KA injection. although GFAP ’ fibrous astrocytes were evident in the cortical white matter and corpus callosum (Fig. 2F). After I4 days most protoplasmic astrocytes in undamaged regions of the cortex lost their GFAP immunorcactivity and GFAP’ astrocytes remained only around the margins of extensive central lesions and in surroLmd regions of focal lesions (Fig. 3B). This reactive astrogliosis did not cxtcnd into the contralatcral cortex a( any time after either ablation or intracortical KA injection.

After cortical ablation the reciprocally interconnected regions of the VB and the posterior thalamic nucleus. medial division (Porn) showed the following sequence of changes in all cases. Between one and two days the relay neurons appeared slightly pale but were not overtly chromatolytic (Fig. 4B). By the fourth day many of the affected relay neurons appeared shrunken and stained darkly with Cresyl Violet and the packing density of small non-neuronal cells appeared to have increased slightly (Fig. 4D). By seven days after ablation no identifiable thalamic relay neurons remained. Between seven and I4 days after ablation there was a marked increase in the density of non-neuronal cells (Fig. 4H). Cases examined at survival times from 30 to I80 days after ablation did not show any evidence of recovery of the damaged relay neurons (Fig. 4J6N). Between 30 and 90 days the density of non-neuronal cells appeared to decrease (Fig. 45 and L). The VB ipsilatcral to the damaged cortex was clearly shrunken in casts cxamincd from 90 to 120 days after cortical ablation (Fig. 4N). In those cases where the cortical ablation had extended vcntrally into the white matter and damaged the optic radiations, neuronal degeneration was also seen in the dorsal lateral gcniculatc nucleus

.fhaldmic retrograde degeneration following cortical injury

Fig. 2. Reaction of cortical neurons and glia to injection of KA into the somatosensorimotor cortex. A. Pattern of GAD immunoreactivity rostra1 to the site of iniection. seven days after iniection of KA in the superficial layers. Intensively reactive GAD + punctae are-present in the lesion center region in layers I and II. Within the lesion surround region below, extending from Iavers III through VI and for about 500~1mlaterally and medially, there is a marked paucity of-GAD’ neurons compared with the normal distribution seen in the contralateral cortex (D). B. Section adjacent to A stained with Cresyl Violet showing the microglial aggregation in the lesion center and the partial loss of neurons in the surround region. C. Adjacent section reacted for GFAP immunohistochemistry. GFAP+-reactive astrocytes mark the extent of the surround region. Scale bar in D is for sections shown in A-D. E. Sensorimotor cortex 14 days after an extensive kainic acid injection, Cresyl Violet stain. A dense microglial reaction is evident along the needle track in the lesion center. F. Section adjacent to E reacted for GFAP immunohist~hemistry. An extensive lesion center rimmed by GFAP’ astrocytes delimits the area of complete neuronal loss in the SI and extends ventrally into the caudate putamen. Scale bar in F is for sections in E and F.

529

I). ‘I

Ross

~llld t‘. 1:. EHWX

Thaiamic retrograde degeneration following cortical injury

531

. .*

Fig. 4(A-II)

‘-,

Fig. 4, Thalamic rctragrade degeneration following cortical ablation. Pairs ai‘pllrttomicrographs showing aftcrations in the appeiilancc of Nissl staining in the VB and POm from one to 120 days after cortical ablation. Comparison of the VB ;nnd Porn at din‘erent stitges in the process of retragradc degeneratton following cortical ablation (rtght side. B. 0. f-‘. H, .I. L, N) with normal Ntssl staining of the corresponding region in the contr;rhiteral thalarnus an the lcfi side (A. C’. E, G. 1. K. M). A and B. One day ;tfter cartic;! ithlation r&y ncurans in the ipsilaterai ventropasterior medial thalamic nucleus (VPM) and POm (B) ;tt positions topagt-apbicafly r&tied to the lesion site in the SI exhibit :r decreased basophilnc compared with rclny ncurans at the corresponding position in the normal contralateral thalamus (A). C and D. Four di~ys uticr cortical ablation the density of cells within the VH and POm ipsilateral to the cortical lesion (D) :tppears to ho inoreascd. Within the degenerating region of the VPM and POm few large relay ncurans can be ili~ll~ified end many smaller ceils. either :ttrophied neurons or non-ncuron;tl ~41s are present. E ~rnd I-. Ssven days after ahltititrn or the forelimb and hindlimb region of the SI large relay ncurans, such 3s those seen III the contral~ttcml VPL (E). XC nat lhund within degenerating regions of the ipsilateral VPL (F). G :tnd il. Fourteen days :rl’ter cortical ablation it marked proliferation of’ non-neuronal cclb with it pale. f&my appeearing cytoplasm is evident in the degenerating portions r)f the VPM and POm :tnd no rchty neurons e:tn be tdentified within these rcgians (H). In spared regions of the VPM ventral tn the degcnerattng region large normal appearing rckty neurons arc present. I and J. Thirty days nltcr cortical ablation ;I cle;fr demarcntian 1s evident between the spined (lower left) nnd the degenerated (upper right) portions of the VPM (.I). Within the de~ener~ltcd regions few pale staining il(~n-Ileur~~n~~icells with the foamy cytoplasm :tre present I( itnd L. Sixty dayr ititer conical ablationthe VB ipsiiatcml to the lesion appears shrunken. Within the dcgcncrated region of the VPM and POm only intensely hasophilic. flnttcned glial ~~41s :tre present(I.1 in :t eandenstd puttcrn which suggests that the dcgencrated rceions of tbc th;tlamus hnd collapsed tn upon thcmscl\es. M and N. One hundred ;md twenty days after car&al ablation tlt~ V’B ipsilntcral to the cortical lcston ttppcars grassiy distartcd and it dense urc:t of’ Ilattsncd darkl> \t;rinirtg gluti cell\ II present :tt tire dors:tl c;~p tlf’ the VB (N). Wtthin these dcgcncratcd rcgtons no reia!

Thalamic

retrograde

.. . .

degeneration

.:

Fig. 5(A H)

following

cortical

injury

533

5 34

D. ‘1‘. Ross and f’. F. EWI;H

Fig. 5. Retrograde degeneration of neurons in the VB and POm following intracortical injection of KA. The progression of thalamic degeneration ipsilateral to the cortical KA lesion is shown in the series of photomicrographs from Nissl-stained sections on the right side (8, D. F, H, .I) and corresponding regions of the normal contralateral thalamus are shown on the left side (A, C, E, G, I) for comparison. A and B. Four days after intracortical injection of KA neurons in the topographically related regions of the VPM and POm appear slightly less hasophilic (B) than those in the corresponding region of the contralateral thalamus (A). C and D. Seven days after intracortical KA injection neurons in the VPM display a marked decrease in basophilia (D) but there is no evidence of either neuronal loss, atrophy, or proliferation of lion-neuronal cells when compared with the normal VB (C). E and F. Fourteen days following mtracorticel KA injection many neurons in the region of the VB and POm topographically related to the cortical KA lesion appear atrophied and the density of non-neuronal cells appears to be higher than normal (F). comparable to the reaction seen at four to seven days following cortical ablation (see f:tp. 4D and F). G and H. Thirty days after intracortical KA injection no relay neurons are evident in the degenerating regions of the VPM and POm (H). Relay neurons in the spared regions of the VPM appear less basophilic than those at corresponding positions on the control side (G). Within the degenerating regions n~rner~)us pale n~~n-ne~r(~nal cells with foamy cytoplasm are evident. I and J. Sixty days after a relatively large intracortical KA injection produced an extensive lesion in the Sf the degenerating and spared regions of the VB are very clearly demarcated (J). Within the degenerated regions no relay neurons arc present, few pale noI]-neuronal cells can be found, but a dense aggregation of darkly staming non-neuronal cells is evident. Scale bar for A I is in 1.

(LGNd) and the lateral posterior thalamic nucleus (LP). Similarly, in those cases where the ablation extended rostrally into the motor cortex degeneration was also seen in the VL. The process of retrograde degeneration in the LGNd. LP. and VL appeared to follow the same time-course as that seen in the VB nuclei and the Porn.

In all cases degeneration in regions of the VB and POm wcrc predictable. based upon the location of the KA injection into the SI. Neuronal loss in Nisslstained sections through the VB and POm was first evident 14 days after intracortical KA injection and the entire process of thalamic retrograde degeneration followed a more protracted time-course. VB and POm neurons examined 2, 4 and 7 days post lesion appeared pale and slightly hypertrophied (Fig. 5B, D). No dark chromatolytic neurons were seen and there did not appear to be an increase in the density of non-neuronal cells during this period. In cases examined 14 days after the lesion only a few large relay neurons could be identified and there was a very marked increase in the density of non-neuronal cells. By 17 days no large relay neurons could be identified in the VB and POm (Fig. 5F). Between 60 and 90 days after intracortical KA injection, the

density of small non-neuronal cells decreased and all relay neurons had completely disappeared (Fig. SH and J). In cases with only focal KA lesions very small regions of neuronal loss were seen in the VB, and the POm often showed no signs of degeneration in either Nissl-stained or GFAP-reacted sections.

The reaction of thalamic astrocytes was examined using GFAP immunohistochemistry. In contrast to the different time-course of the thalamic neuronal response following the two types of cortical injury, the temporal and spatial patterns of GFAP labeling in the thalamus were nearly identical following cortical ablation and intracortical KA injection. Protoplasmic astrocytes in the adult mouse VB and POm do not normally express detectable levels of GFAP (Figs 6A and 7A). GFAP’-reactive astrocytes were evident in the VB as early as two days following either cortical ablation (Fig. 6B) or intracortical KA injection (Fig. 7B). From two to 14 days after cortical injury GFAP ‘- processes of thalamic protoplasmic astrocytes appeared to increase in density but there was no indication of an increase in the number of astroglial perikarya in the degenerating regions of the VB and POm (Figs 6C, D and 7C. I)). Protoplasmic astrocytes in degenerated regions of the VB remained intensely reactive up to 30 days following cortical

Thalamic

retrograde

degeneration

following

cortical

injury

.

c

Fig. 6. Reaction of thalamic astrocytes to cortical ablation. Coronal sections through the thalamus of an adult mouse at the level of the VB and POm reacted for GFAP immunohistochemistry. A. GFAP+ fibrous astrocytes are present in the internal capsule and crus cerebri but thalamic protoplasmic astrocytes in the adult mouse VPM. VPL, and POm are not labeled. B. Two days after ablation of the ipsilateral SI protoplasmic astrocytes in regions of the VPL and POm topographically related to the cortical lesion are reactive and express GFAP. C. Fourteen days after ablation of the ipsilateral Sf intensely GFAP+-reactive astrocytes are present in regions of the VPM, VPL, and POm topographically corresponding to the damaged region of the SI. D. Thirty days after cortical ablation reactive protoplasmic astrocytes remain intensely GFAP+ in regions of the VPL and POm topographically corresponding to the site of the ablation in the SI. E. Sixty days cortical ablation the protoplasmic astrocytes in degenerating regions of the VB and POm are less reactive and fibrous astrocytes in the internal capsule and crus cerebri are more intensely GFAP. F. Within the VB 120 days after cortical ablation only a small dense patch of reactive astrocytes is present, corresponding to the region of neuronal loss at the dorsal cap of the distorted VB. Fibrous astrocytes in the internal capsule and crus cerebri at positions corresponding to the degenerated thalamocortical and corticofugal projections are intensely GFAP+. Scale bar for A F is in F.

535

A

/

PUm ‘_“.

I

/

/‘~

Thalamic retrograde degeneration following cortical injury injury (Figs 6E and 7D). Between 30 and 90 days the GFAP labeling in the VB and POm became noticeably paler (Figs 6D-F and 7D-F). Between 60 and 90 days after cortical injury a dense meshwork of intensety GFAP’ fibrous astrocytic processes became evident within the internal capsule and the cerebral peduncle at locations where fascicles of corticofugal and thalamocorticai axons had degenerated (Figs 6E, F and 7E). This reaction was most intense at 120 days (Fig. 6F). Regions of thalamic nuclei containing reactive astrocytes (Fig. 8C and E) corresponded very closely to regions in adjacent Nissl-stained sections where relay neurons were undergoing retrograde degeneration (Fig. 8B and E) in all 58 cases examined. Alterations in thalamic glutamate decarboxylase iabeling pattern

GAD labeling displayed a transient increase in degenerating regions of the VB following both types of cortical injury (Figs 9 and 10). Increased intensity of GAD+ puncta within degenerating regions of the VB was evident from two to seven days after cortical ablation (Fig. 9C-F), but returned to normal levels by 14 days (Fig. 9G and H). Within the VB only few GAD+ neurons were ever present (Fig. 9B, F and H). In regions of the VB that corresponded topographically to KA lesions in the SI, a slight increase in the intensity of GAD’ puncta labeling was seen two days after cortical injury (Fig. IOC and D). By four days after intracortical KA injection an increased intensity of GAD+ puncta was evident throughout the entire VB, but remained noticeably higher in the regions which topographically corresponded to the KA lesion in the cortex (Fig. IOE and F). This increase in GAD labeling was generally restricted to the VB and did not extend into the ventromedial thalamic nucleus (VM) or POm (Fig. 1OE-G). In cases examined between 14 and 17 days the GAD labeling within the degenerating regions of the VB was more intense, as was the intensity of perikaryal GAD labeling in the thalamic reticular nucleus (RT) (Fig. IOE and F). The intensity of GAD labeling in the VB and RT remained high 30 days after intracortical KA injection (Fig. IOG and H). Between 30 and 60 days the intensity of the GAD labeling in these structures decreased to the levels seen on the control side and remained low thereafter. A substantial population of GAD+ cells were the only neurons present in the LGNd in cases examined 14 days or longer after damage to the optic radiations or intracortical KA injection which spilled over into the visual cortex. DISCUSSION Retrograde degeneration .foUowing cortical injury

of t~a~ffrnic relay neurons

Neuronal loss, detectable at the light microscopic level, does not occur in either the VB or POm of adult

537

mice until four days after ablation of the SI. At earlier times (one to three days) the axotomized VB and Porn neurons appear noticeably pale in Nissl-stained sections. This decreased basophilia prior to neuronal death in the thaiamus has been extensively described in other species following cortical ablation.8~8i.83.‘4’ On the other hand, axotomized thalamic relay neurons exhibit few of the classic signs of chromatolysis, such as an intensely basophilic staining of Nissf granules in the peripheral cytopIasm.6~20*33~73~‘33 Eniarson4’ noted that neurons undergoing overexcitation during status epilepticus showed a marked decrease in basophilia when visualized in Nissl-stained preparations. At the same time, paroxysmal firing of neurons in epileptic cortex and thalamus decreases mRNA Ievels4’ and amino acid uptake2’ in these areas. Since axotomized thaiamic relay neurons exhibit abnormally high levels of firing when recorded extracellularly between 3 h and two days following cortical ablation,‘.“’ the decreased basophilia of Nissl bodies in affected neurons may reflect decreased levels of protein synthesis due to overexcitation and not as a result ‘of axotomy per se. The disappearance, and presumably death, of the relay neurons in the adult mouse VB occurs between four and 14 days following ablation. The time-course of neuronal loss in the degenerating thalamus of adult mice is characterized by both an earlier onset and a more rapid conclusion than the thalamic retrograde degeneration observed following similar cortical lesions in larger mammals. The first signs of neuronal loss following cortical ablation occur in the LGN and VI.3of rats at five to six days,9,83seven to 10 days for various thalamic nuclei in rabbits,25*8’IO days for the cat VB,‘” and 14 days for the monkey LGN.r4’ The first signs of the perikaryal retrograde response to axonal injury also occur faster in the thalamus of smafler species. These responses, which characterize degenerating relay neurons prior to their death, include decreased cytoplasmic basophilia in Nisslstained sections,22-25~83~‘4’ and profound disruption of granular endoplasmic reticulum with dispersion of ribosomes seen ultrastructurally.7~9~8’~n3~‘4’ Thalamocortical axons in species with smaller brains are propo~ionately shorter and, as a consequence, cortical ablation in smalIer animals axotomizes thalamic relay neurons closer to their perikarya. This crossspecies proximity effect appears to be analogous to the observation that peripheral nerve injuries closer to motoneuron somata initiate a chromatolytic response that is both more rapid in its onset and more severe in its manifestation than that of similar injuries at more distal positions.“7’.‘i”.‘4 In many early studies of thalamic retrograde degeneration the persistence of undegenerated neurons was observed within affected regions of the thalamus at long survival intervals following cortical ablation. 2’-.-‘5 ‘-~U.72.RI.X6.101.137.141 In cases examined several months following cortical ablation neurons within degenerated regions of many thalamic nuclei were

fig. 8. Correspondence of thalamic retrograde degeneration and reactrve astroghosis. A. Coronal section mouse at the levei of the VB. Porn, and RT. Crcsyl Violet stain. B. Corresponding region of the thalamus loss is evident in both the ventral portion of the VPL and the dorsomedial portion of the Porn. Scale bar reacted for GFAP immunohistochemistry. GFAP + protoplasmic astrocytes are present in regions of the VPL has occurred. D. Detail of normal neurons in the VPL region from box in A. E. AppGIranCe of corresponding of the hindlimb region of the Sf. detail of box From B. Within the degcnrrated area no large relay neurons are cells but in the spared dorsal portion of the VPL the relay neurons appear normal. F. Adjacent section reacted i’rwn c‘ Int~nscly GFhP-re:tcti\c pr[lt(~pi~Is~li~ astrocyte\ are pr-esent at the position where the retrograde

through the thalamus of a normal adult 14 days after cortical ablation. Neuronal for A and B is in l3. C. Adjacent section and POm where retrograde degeneration region of the VPL 14 days after ablation seen among the many small non-neuronal for GFAP immunohistochemistry. detail degeneration of the relay neurons had

Thalamic retrograde degeneration following cortical injury

f3

Fig. 9

’

539

./

*&l

540

I>. 7. Ross and f-‘. 1--.F.HNkK

Thalamic

retrograde

degeneration

totally unaffected or displayed only a very slight atrophy. In some species the thalamic relay nuclei contain significant numbers of small interneurons that do not send axonal projections to the cerebral cortex.‘05.“7 Most of these neurons contain immunohistochemically detectable levels of GAD and appear to be inhibitory local circuit neurons.6’.y5 The majority of neurons that persist within degenerated regions of thalamic relay nuclei appear to be interneurons which are not axotomized by cortical injury. In normal control cases and on the side contralateral to cortical injury in experimental cases very few GAD+ cells could be seen in any one section through the VB and such cells appeared to represent less than 1% of the total neuronal population of neurons in the VB, similar to the reported distribution of neurons in the rat VB immunoreactive for GABA itself.5Y Ipsilateral to the lesion these cells were intensely GAD+ during the early phases of the degenerative process but were difficult to identify within degenerated regions of the VB at times after the death of the relay neurons. The persistence of few, if any, neurons in affected regions of the VB following either

following

cortical

541

injury

cortical ablation is consistent with horseradish peroxidase (HRP) retrograde transport studies and ultrastructural studies82.84.“6.‘30which conclude that about 99% of the neurons in the rat and mouse VB are relay neurons. The e.ucitotoxic degeneration

hypothesis

of thalamic

retrograde

Several alterations within the thalamus following cortical injury suggest that overexcitation may be involved in the process of thalamic retrograde degeneration. First, axotomized thalamic relay neurons retain the large excitatory synaptic contacts on their proximal dendrites and perikarya. The synaptic contacts of lemniscal axons upon relay neurons in the VB and those of retinal axons upon relay neurons in the LGNd persist upon the axotomized neurons until they die,".~8.xl.l41 The stability of these synapses stands in sharp contrast to axotomized motoneurons which undergo a phase of synaptic stripping soon after peripheral nerve injury in which large excitatory contacts are removed from their perikarya and proximal dendrites.“.” Secondly, thalamic neurons

Fig. 9. Alterations of GAD immunolabeling in the VB and RT following cortical ablation. A. Coronal section through the RT at the level of the VB and POm reacted for GAD immunohistochemistry. GAD+ cells are present in the reticular nucleus as arrays of five to I2 cells which extend from the VPL laterally along fascicles of traversing axons. B. Detail of GAD labeling in VB from A. A few small GAD+ intrinsic neurons are labeled in the VPM but the labeling of punctae is barely detectable. C. Corresponding region of the thalamus two days after ablation of the SI. The intensity of GAD labeling in the RT is increased and punctae are more intensely labeled in the degenerating region of the VB. D. Detail of the intensely GAD+ punctae in the neuropil of the VB from C. E. Corresponding region of the thalamus four days after ablation of the SI. The degenerating region of the VB is clearly outlined by intensely GAD+ punctae and the intensity of GAD labeling in the RT is higher than normal. F. Detail of GAD labeling in the neuropil of the VB from E. The punctae are very intensely GAD+ as are the few GAD+ VB interneurons present in this section. G. Corresponding region of the thalamus I4 days after ablation of the SI. The intensity of GAD immunoreactivity in both the degenerating region of the VB and the associated regions of the RT is reduced. Scale bar for A. C, E and G is in G. H. Detail of GAD labeling in the neuropil of the VB from G. The punctae from the RT projection into the VB are barely detectable but the intrinsic GAD+ neurons of the VB and their processes remain intensely labeled. Scale bar for B, D, F, and H is in H.

Fig. IO. Alterations in patterns of GAD immunolabeling in the VB and RT following intracortical KA injection. A. Coronal section through the thalamus at the level of the VB, Porn, and RT of a normal adult mouse reacted for GAD immunohistochemistry. Arrays of five to I2 GAD+ RT neurons extend from the VPL to the internal capsule along fascicles of traversing axons. B. Detail of GAD labeling in the VB from A. Within the neuropil of the VB labeling of punctae from the RT projection is barely detectable. C. Two days after intracortical KA injection in the forelimb and hindlimb region of the SI the intensity of GAD labeling in the RT is increased and punctae from the RT projection into the VB are also detectable. D. Detail of GAD+ punctae labeling in the VPL from C. Punctae are visible only in regions of the VPL topographically associated with the site of the KA lesion in the SI. E. Seventeen days after injection of KA into the SI perikarya in the RT are intensely GAD+ and their punctae within the VB are also intensely labeled. Although GAD+ punctae show an increased intensity of labeling throughout the VB, the intensity is highest in degenerating regions of the VPM and VPL which topographically correspond to the center of the lesion in the SI. Note that the region of increased punctae labeling stops at the VPM/POm border. F. Detail of GAD labeling of punctae within the VB from E. The area of most intense GAD+ labeling corresponds to the lesion center in the SI but the punctae in the surround, non-degenerating regions of the VPM dorsal and ventral to this region also exhibit higher than normal levels of GAD labeling. G. Corresponding region of the thalamus 30 days after injection of KA in the SI. Increased intensity of GAD+ labeling of RT neurons and punctae within the degenerating VB persist. Scale bar for A, C, E, and G is in G. H. Detail of GAD+ punctae labeling in the neuropil of the degenerated VPM from G. For purposes of orientation, the asterisk in H corresponds to the asterisk in G. Scale bar for B, D, F. and H is in H.

542

D. T. Ross and F. F. EHNEK

become hyperexcitable following axotomy. as manifest by an increased rate of spontaneous firing and an enhanced response to sensory evoked stimulation.“.“” Motoneurons also exhibit hyperexcitability following axotomy. evidenced by the production of dendritic Ca’ ’ spikes in response to the stimulation of afferents that terminate upon the distal reaches of their dendritic arbors.” Whereas the stripping of large excitatory contacts from their perikarya and proximal dendrites may protect axotomized motoneurons from overexcitation. axotomized thalamic relay neurons do not appear to be afforded this protection. The cxcitotoxic hypothesis of thalamic retrograde degeneration holds that cortical injury renders thalamic relay neurons hyperexcitable and that a cascade of pathological alterations within the thalamus combine to excite the damaged relay neurons to death. Thalamic retrograde degeneration following both cortical ablation and intracortical kainic acid injection may share a common mechanism of progressive overexcitation. Alterations in thalamic protoplasmic astrocytes and inhibitory neurons in the RT which precede the retrograde degeneration of relay neurons may serkc both as indicators of overexcitation and as contributors to the cxcitotoxic process. Tiir contribution

of’c~oriicotllulunzic~ termid

lion to thubmic

retropak

degeneru -

degenrrution

Corticothalamic projections terminate upon the distal dendritic branches of thalamic relay neurons and may constitute up to 75% of the round vesicle asymmetrical (excitatory) contacts upon VB neurons.“” The Wallerian degeneration of corticothalamic projections therefore results in massive deafferentation of the relay neurons. Matthews” suggested that degeneration of corticothalamic terminals en masse may represent an anterograde transneuronal mechanism that accelerates the degenerative process of axotomized thalamic relay neurons by decreasing the neuronal activity of the axotomized thalamic relay neurons. Alternatively. it is possible that synchronous degeneration of large numbers of excitatory contacts may both increase relay neuron excitability and directly contribute to their overexcitation. Several lines of evidence support this interpretation. For example. one of the characteristic features of corticothalamic terminals undergoing Wallerian degeneration is a loss of identifiable synaptic vesicles.’ ” “’ Terminal degeneration may be accompanied by the release ol massive amounts of excitatory transmitter into the cxtracellular space around the degenerating thalamic neurons. The findings from numerous studies suggest that the excitotoxic amino acids glutamate or aspartate meet many of the necessary criteria for identification as the neurotransmitter at the corticothalamic synapse.‘” The perikarya of many layer V and VI pyramidal neurons contain immunohistochcmically detectable quantities of glutaminasc. an enzyme involved in glulamate and aspartate synthesis,‘“’ and also label seleclively with antibodies to glutamate

itself,?“,‘4.” Corticothalamic terminals have highaffinity mechanisms for glutamate uptake3,“’ “,I45 and cortical pyramidal neurons in layers V and VI retrogradely transport [‘H]aspartate.‘0.“4~‘3’ Synaptic activity at the corticothalamic synapse in the VB of adult rats can also be completely and reversibly antagonized by kynurenic acid, a broad spectrum excitatory amino acid antagonist.lu8 The degeneration of corticothalamic terminals following ablation of the somatosensory cortex in adult rats and cats decreases the specific uptake of glutamate from degenerating regions of the VPL.‘,‘“’ Iontophoresis of the glutamate uptake blocker glutamic acid dimethyl ester into the VB of normal cats results in an increase in the spontaneous activity of normal relay neurons.5’ Increases in the thalamic extracellular concentration of cndogenous excitatory amino acid neurotransmitters due to massive release from corticothalamic terminals’44 and the breakdown of mechanisms for their reuptake provide a potent combination of responses which could contribute to the overexcitation of relay neurons following cortical injury.

Expression of GFAP in thalamic protoplasmic astrocytes within affected regions of the thalamus consistently precedes relay neuron degeneration in these areas. Protoplasmic astrocytes within degenerating thalamic regions undergo a reactive astrogliosis that is characterized ultrastructurally by a hypertrophy of astroglial processes and an increase in the number and density of intermediate filaments within the processes.7,x’.83,‘4’ Protoplasmic astrocytes within the normal adult thalamus do not express levels 01 GFAP that are immunohistochemically detectable at the light microscopic level. Expression of GFAP by mature thalamic astrocytes has been demonstrated as a reaction to deep stab wounds in the cerebrum.‘5.‘“.x” Neither the topographic relationship between the site of cortical lesions and thalamic GFAP labeling patterns nor the close correspondence of GFAP reactivity and thalamic retrograde degeneration have been documented previously. The patterns of GFAP+-reactive astrogliosis in the thalamus corrcspond so well with the regions of neuronal degeneration that prior examination of adjacent sections reacted for GFAP immunohistochemistry significantly aids the rapid localization of zones of neuronal degeneration in Nissl-stained sections. Retrograde cell death is not likely to be the signal for the expression of GFAP by protoplasmic astrocytes in the affected areas of the thalamus since reactive astrogliosis precedes the loss of relay neurons in the VB and POm by at least two days following cortical ablation and by up to IO days following intracortical KA injection. Ablation of the Sl results in the anterograde degeneration of corticothalamic projections to the VM and centrolateral thalamic nucleus (CL) as well as the POm and VB but reactive

Thalamic

retrograde

degeneration

is never seen in either the CL or VM. The absence of consistent findings of intensely GFAP+reactive gliosis in other structures that normally receive a corticofugal projection (e.g. the contralateral cortex, striatum, thalamic reticular nucleus, superior colliculus, pons, and spinal cord) suggests that the rapid expression of massive quantities of GFAP by protoplasmic astrocytes in the VB and POm is not triggered directly by the anterograde (Wallerian) degeneration of corticothalamic terminals. Pathological alterations within the extracellular fluid compartment surrounding the neurons in the VB and POm as a result of abnormally high levels of neuronal activity, however, may be sufficient to trigger reactive astrogliosis. Studies of astroglial populations in vitru indicate that reactive astrogliosis may be triggered by high levels of K+ or CO> in the bathing medium. ‘J* During periods of high neuronal activity the extracellular pH decreases and extracellular K + levels are increased.53,y’.“h1?7 Abnormally high levels of thalamic neuronal activity, such as epileptiform discharges, raise extracellular K+ levels from about 3 mM up to levels approaching 12 mM.” The high-frequency injury discharge of thalamic relay neurons which immediately follows cortical ablation”” may transiently increase extracellular K+ and CO? enough to trigger the reaction of thalamic protoplasmic astrocytes and induce the expression of GFAP in ciao. Sustained levels of abnormally high relay neuron activity following axotomy4,“” or during distant excitotoxic degeneration following intracortical kainic acid degeneration may maintain the reactive state of the surrounding protoplasmic astrocytes. Reactive astroghosis may therefore serve as an indicator of overexcitation in degenerating thalamic nuclei. Other alterations involving thalamic protoplasmic astrocytes may contribute to the overexcitation of thalamic relay neurons following cortical injury. Swelling of protoplasmic astrocytes decreases the volume of extracellular space in the gray matter.27.s7.” This may effectively elevate the concentration of substances released into the extracellular space, including K+.‘* High K+ levels produce presynaptic depolarization, resulting in the pathologically induced local release of excitatory transmitter,“.4’ and inhibit high-affinity uptake mechanisms for aspartate and glutamate.i4 Increases in extracellular levels of endogenous excitatory amino acids, due to either increased presynaptic release, decreased uptake, or decreased extracellular volume which may accompany astroghal swelling, could contribute to the overexcitation of the affected relay neurons.

astrogliosis

Alterations in thalamic glutamate decarboxylase ing patterns

label-

One alteration that is consistently seen in the degenerating VB prior to the death of relay neurons is an increased level of GAD immunolabeling. Following cortical ablation a transient increase in the

following

cortical

injury

543

intensity of GAD labeling is seen in the perikarya of thalamic reticular neurons and their projections into the degenerating regions of the thalamus. The increase in the intensity of GAD labeling observed in degenerating thalamic regions does not appear to represent the sprouting of GABAergic processes from intact RT neurons in response to corticothalamic terminal degeneration because neither the number nor the distribution of flat symmetrical synaptic contacts upon the relay neurons change prior to their death.8~‘~‘x~n’~‘4’ The increase in immunohistochemically detectable GAD levels within the degenerating VB most likely represents an increase in neuronal and metabolic activity of the thalamic reticular neurons. Due to reciprocal connections between the VB and the RT,h5.““.“7 increased activity of thalamic relay neurons increases the activity of RT neurons.‘,’ Increased levels of GAD in RT neurons and their processes may reflect an increase in their metabolic activity due to increased synaptic driving by axotomized relay neurons, The decrease in the intensity of GAD labeling within the degenerating VB that occurs between seven and 14 days after ablation may reflect the end of overexcitation, occasioned by the death of the relay neurons. A transient increase in GAD labeling throughout the entire VB occurs following intracortical KA injection. The intensity of labeling is highest in the degenerating thalamic regions. Unlike the increased GAD labeling seen following cortical ablation, an increase is also detectable in non-degenerating regions of the VB. GAD levels in degenerating VB regions may be raised by increased levels of relay neuron activity via the mechanism described previously for GAD increases following cortical ablation. The increased GAD seen in the non-degenerating regions of the VB may reflect increased activity of spared corticothalamic neurons due to the selective loss of GABAergic inhibitory interneurons in the surround region of the cortical KA lesion. The loss of these cells may disinhibit the remaining corticothalamic neurons in the lesion surround zone, particularly if chandelier cells that make inhibitory contacts on pyramidal neuron initial segments43.64.‘00.“*.‘~y.‘~4 are lost. Release of layer V and VI corticothalamic pyramidal neurons from inhibition may increase the tonic driving of RT neurons that project to the VB surrounding the degenerating zone, increasing GAD levels. Alterations in the activity of RT neurons may also directly contribute to the degeneration of the axotomized relay neurons. Neurons in the RT receive excitatory input from the cerebral cortex”.65 via collaterals of the layer V and VI pyramidal cells that project to the thalamus.“‘5”7.‘43 They also receive excitatory inputs from thalamic relay neurons.65~‘04~‘oj~“7~‘43 The tonic inhibitory influence of cortical activity upon thalamic relay neurons mediated by the RT is suppressed during cortical spreading depression’ and is

lost tollowing cortical ablation.’ The loss of this tonic Inhibition following cortical ablation may directly increase the activity of the axotomized relay neurons. The influence of corticothalamic activity upon thalamic relay neurons has been characterized as excitatory.““.“” inhibitory,‘““J or ;I combination of both.’ “M”’ ‘I’ The balance between excitation. mediated by the d7rect corticothalamic projection to the VB. and inhibition. acting through the corticorctioulur thalamic loop, may serve as the neural substrate for sharpening response parameters 01 thalamic neurons.“’ but may also serve to hold the resting membrane potential of the relay neurons within the range optimal for high-fidelity sensory relay transmission.‘h.“2 “.“ The phasic inhibition of relay neurons. mediated through their reciprocal connections with the RT.’ I” persists following cortical ablation and is increased proportionally with relay neuron activity. It is possible that an increase in phasic inhibition actually increases the excitability of the aaotomized relay neurons because thalamic neurons possess lowthreshold voltage-sensitive Ca’ channels that are not normally opened unless the resting membrane potential is hyperpolarized prior to an excitatory stimuJus,‘li(‘i Opening of this channel produces a calcium spike upon which numerous fast sodium spikes arc imposed.“.“’ A dramatic alteration in the nature of the relay neuron’s rcsponsc to excitation occurs as a result. The normal “relay” mode of firing. characterized by high-fidelity following of synaptic sensory input with single Na’ spikes. is replaced by 21 hypermode.“’ similar to the paroexcitable “bursting” xysmal cpilcptiform bursting of cortical neurons. Paroxysmal bursting of relay neurons in the VPL. is produced when penicillin is applied to the cortic;ll surf~lce,~l.~?M1.‘11.1 1’) Regular afterhyperpolarizing potentials recorded between paroxysmal bursts”” may reflect the phasic activfity of RT neurons which continuously de-inactivate v,oltage-sensitive Ca’ channels and perpetuate the relay neuron’s bursting pattern of firing. A similar mechanism may operate following cortical ablation, though the pattern ol bursting of axotomized thalamic relay neurons appears to be less frequent and mom irregular. at least in anesthetized animals recorded in the tirst 6 h after ablation.‘“’ ‘I” Thus increased inhibition of axotomized thalamic relay neurons may paradoxically increase their excitability. If ovcrexcitation is involved 7n the process of thalamic retrograde dcpencration then the increased activity of inhibitory ncurons In the RT may contribute to the dcgcnerativc process through activation of the low-threshold Ca’ ’ channels.

The onset of thalamic neuronal loss following intracortical KA injection is actually seven to IO days later than that seen following cortical ablation. Why

is thalamic retrograde degeneration delayed following intracortical KA injection? One hypothesis previously put forward to account for thalamic retrograde degeneration following cortical injury, the neurotrophic hypothesis, holds that “maintenance”-type ncurotrophic factors necessary for the maintenance of normal thalamic function and relay neuron survival are produced by cortical elements and transported to the thalamus!’ As a result of neuronal lysis following intracortical KA injection, sufficient yuantities of such factors may be released into the cortex and remain available for retrograde transport by thalamocortical axons for several days. Thalamocortical axons, which are spared by the cortical KA lesion.“’ may continue to take up and transport these factors for some time after the cortical neurons have died. According to this hypothesis the more rapid degeneration of thalamic relay neurons following cortical ablation may occur because the source of cortically produced trophic factors has been removed. If retrograde degeneration of thalamic relay neurons occurs due to the cessation of trophic factor transport to the thalamus. then blockade of retrograde and antcrograde axonal transport from the somatosensory cortex for an extended period of time should result in the retrograde degeneration of relay neurons in the VB. Blockade of transport from the somatosensory cortex for I4 days by repeated intr;tcortical injection of colchicine was sufficient to prevent transport of HRP from the SI to the VB and POm but did not produce retrograde degeneration of thalamic relay neurons in any of the six cases examined (D. T. Ross. C. Weathcrly-White, and F. 1:. Ebner. unpublished observations). Although thala mic relay neurons may receive II continuous supply ol substances that arc transported retrogradcly from the neocortcx. interruption of this supply dots not appear sufficient to cause the thalamic rctrogradc dcgcneration following cortical injury. The application of convulsants to the cortical surface results in the degeneration of relay neurons m topographically related thalatnic nuclei” and intraccrcbral injection of KA produces distant dcgenerat7on of neurons in structures related to the site 01 in.jcction by seizure pathways.“,“7 ““JamNeuronal loss under both of these conditions is characterized by an excitotoxic component to the degenerative process.“’ Given these observations, the presence of an cxcitotoxic component to the degenerative process 01 thaiatnic relay neurons following intracortical KA in_jection is highly probable. If both cortical KA lesions and cortical ablation lead to the massive rclcasc of excitatory amino acids from corticothalamic terminals, antidromically depolarize VB relay neurons. and prccipitatc alterations in the thalamic microenvironment which contribute to overexcitation. why does it take seven to IO days longer fot relay neurons to die following intracortical KA injection? The explanation may lit in the obvious diffcrcncc that thalamic axons arc axotomizcd by cortical

Thalamic retrograde degeneration following cortical injury ablation but re.main intact within an abnormal extracellular environment following KA injection. KA injection into regions of dense axonal termination, unlike direct axotomy, does not produce an injury discharge. Ionic influxes that occur following axotomy may precipitate a metabolic crisis in the perikarya of damaged neurons which accelerates the degenerative process. Axotomy induces a massive influx of Nat and Cal+ down their concentration gradients into damaged axonal segments, resulting in an acute, profound antidromic depolarization which initiates the injury discharge. The perikaryal calcium influx consequent upon the injury discharge may bring the axotomized relay neurons very close to a state of metabolic distress where the cells’ internal mechanisms for sequestering or eliminating calcium can no longer maintain homeostasis. Abnormally high intracellular calcium levels activate pathological proteolysis”x and other degenerative processes which could ultimately kill the affected neurons.‘*’ The absence of an injury discharge related perturbation of the relay neuron’s intracellular environment, particularly the lack of an early build-up of intracellular calcium levels, may allow longer survival following intracortical KA injection. Thalamic relay neurons do eventually succumb to the pathological consequences of intracortical KA injection, but the integrity of intracellular mechanisms for maintaining ionic balance may allow them to hold out longer against the onslaught of local excitotoxic conditions in the degenerating thalamus. A progressitle excitotoxic retrograde degeneration?

mechanism

,for thulamic

Each aspect of the previous discussion is consistent with the interpretation that a cascade of pathological alterations occurs following cortical injury which may act to both increase the excitability of the damaged relay neurons and contribute to their overexcitation. Prolonged overexcitation may precipitate a condition of extreme metabolic distress that eventually proves fatal to the affected relay neurons. The progressive excitotoxic processes in thalamic retrograde degeneration may involve (I) depolarization of the thalamic relay neurons following cortical injury, (2) Wallerian degeneration of the corticothalamic projections, resulting in massive release of excitatory amino acids and the degradation of mechanisms for their uptake, (3) increased phasic activity of inhibitory thalamic reticular neurons reciprocally connected with affected relay neurons which “deinactivates” voltage-sensitive Cal+ channels and switches the relay neurons into a hyperexcitable “bursting” mode, and (4) alterations in astrocytic structure and function that may decrease the size of the extracellular fluid compartment and increase extracellular levels of potassium and glutamate in the degenerating thalamus. The cascade of pathological alterations in relay neuron excitation and excitability combined with their sustained synaptic driving by lemniscal

545

afferents may bring about retrograde degeneration via an excitotoxic process. From in vitro studies of glutamate neurotoxicity, two types of excitotoxic degeneration have been identified, differing in their ionic dependence, timecourse, and concentration of excitatory amino acids required to trigger the processes. The first type is dependent upon extracellular Na+ and Cl-, is marked by profound swelling of the affected neurons, and results in rapid catastrophic degeneration within several hours.‘4.“2 High extracellular levels of excitatory amino acids bring about steady depolarization characterized by massive Na+ influx and an attendant passive Cl- influx which precipitates osmotic neuronal lysis. Cell death occurs within several hours after the onset of excitotoxic conditions. This type of excitotoxic degeneration appears to occur following the direct application of these exogenous excitotoxins in vice 31.75.Y7.YR.I?0 or in tritro,?4,47.'03."* by anoxia in tissue culture,‘38 and possibly by sustained epileptiform activity in ciao.2x.50 The second type of excitotoxic degeneration can be initiated by much lower excitotoxin concentrations, is dependent upon extracellular calcium, and is much slower to develop.‘3~24~3’~“3Sustained overexcitation depolarizes the resting membrane potential of affected neurons and activates voltage sensitive calcium conductances associated with the N-methylD-aspartate (NMDA)-type excitatory amino acid receptor complex.‘3~77~“3 Elevated intracellular calcium levels appear to exert their neurotoxic effect via pathological stimulation of intracellular proteases and phospholipases which in turn may precipitate free radical lipid peroxidation, tearing cells up from the inside out.‘** The delayed (one- to fourday) degeneration of hippocampal CAI pyramidal cells following transient (IO-min) ischemia and reperfusion involves this type of Ca’+-dependent excitotoxicity.48.“4 Retrograde degeneration of thalamic relay neurons is not a rapid catastrophic event which takes place within a few hours after cortical injury. Thalamic degeneration following either cortical ablation or intracortical KA injection appears to be a progressive process. It is possible that thalamic extracellular glutamate levels progressively increase following cortical injury, and take several days to reach levels that precipitate rapid catastrophic excitotoxic degeneration. Alternatively, cortical injury may render thalamic neurons sufficiently hyperexcitable such that even normal levels of sustained afferent synaptic driving, reflected in little, if any, increase in extracellular glutamate levels, may succeed in overexciting them. As a consequence of sustained high levels of Ca?+ influx through LT or NMDA channels, intracellular mechanisms for removing or sequestering calcium may be pushed to the limit for a prolonged period of time before their capacity is finally exceeded and the fatal cascade of calcium toxicity and free radical lipid peroxidation take their toll. Descriptions of thalamic

s-10

D T. Ross and I-. I’. EHNlK

retrograde degeneration at the ultrastructural level leave open the possibility that both types of excitotoxic degeneration may be operating in the death 01 the relay neuron.” Why are thalamic relay neurons so sensitive to axotomy or application of excitotoxins in their terminal areas when other types of neurons in the adult CNS can survive these types of injuries and even. in some cases. regenerate their axons’?” K~~.“‘~.““‘.“~ The extreme sensitivity of thalamic neurons to these types of injuries may arise from a combination of the neurons’ own intrinsic membrane properties, the organization of their synaptic connections. and their failure to adapt to pathological alterations in their local microenvironment. Intrinsic properties of thalamic relay neurons. particularly the presence of lowthreshold Ca” channels and NMDA receptors with their associated voltage-sensitive Ca” channels. may place these neurons at greater risk. when pathological conditions develop within the local extraccllular microenvironment. than are other types of neurons which do not have thcsc Ca” channels. It is ironic that the presence of inhibitory synaptic contacts upon the pcrikarya of thalamic relay neurons might serve to increase rather than decrcasc their excitability once ovcrexcitation is begun. Other types of neurons share reciprocal connections with CiABAergic interneurons

but few of these also have low-threshold Ca’ ’ channels. The magnitude of corticothalamic terminal degeneration following cortical injury with its conscquent release and decreased capacity for reuptake of excitatory neurotransmitter within the thalamus may also put thalamic relay neurons at greater risk than other types of neurons which receive equivalent lcsions within their axon terminal areas. Although protective mechanisms which operate under normal physiological conditions to prevent overexcitation and counteract its adverse effects upon intracellular processes are clearly present within the thalamus. thcsc systems appear to be overwhelmed following cortical injury. The combination of a neuron’s intrinsic membrane properties, the nature and distribution of their synaptic inputs, and their degree of success in mounting a homeostatic response to the secondary pathological responses to injury may determine whether axotomy results in degeneration or regeneration. .-lc~krro~~‘k~c/~rnrrr~/., ----We thank Dr Larry Eng for the giti ot antisera to GFAP. Dr Donald Schmechel for his gift 01 antisera to GAD, and Michelle Dodge. Carl WeatherlyWhite. and Danny Cho for technical assistance in histology and immunohistochemistry. This work was partially supported by NIH postdoctoral fellowship NS074919 to DTR and NIH grant NS13031 to FFE.

REFERENCES

I. Albe-Fessard D., Condes-Lard M.. Kcsar S. and Sanderson P. (1983) Tonic controls acting on spontaneous and evoked thalamic activity. In Somuro,sm.vor~~ In&~:rutio,~ 111/hc T/~u/rtmu.s (eds Macchi G.. Rusitoni A. and Spreafico R.). pp. 273 286. Elsevier, New York. 2. Andersen P.. Junge K. and Sveen 0. (1972) Corticofugal facilitation of thalamic transmission. Bruin Behcrr-. Erolr~ 6, I70 184. 3 Angel A. (1983) The functional mterrelations between the somatosensory cortex and the thalamic reticular nucleus: their role in the control of information transfer across the specific somatosensory relay nucleus. In Somuto.ven.~~r~~ Inrrgra/ion in //IC T/danm (eds Macchi G.. Rusitoni A. and Spreafico R.), pp. 221 240. Elsevier, New York. of the forelimb in the ventrobasal thalamic complex 4. Angel A. and Clarke K. A. (1975) An analysis of the representation of the albino rat. J. Ph~~siol.. Lonrl. 249, 399 -413. 3 Auen E. L.. Bourke R. S.. Barron K. D.. San Fillip0 B. D. and Waldman J. B. (1979) Alterations in cat cerebrocortical capillary morphological parameters following potassium induced cerebrocortical swelling. ,4(,/u neuroprrth.. Bdir? 47,

175

I. 8. 9. IO.

I I. 12.

13. 14. IS

181.

K. D. (19X3)Comparative observations on the cytologic reactions of central and peripheral nerve cells to axotomy. In Spincll Coral Rac,onstrt/c,/k,n (eds Kao C. C.. Bunge R. P. and Reier P. J.). pp. 7 40. Raven Press, New York. Barron K. D. and Doolin P. F. (1968) Ultrastructural observations on retrograde atrophy of lateral geniculatc body II. The environs of the neuronal somata. J. Neuropatlz. c’_tp. Nrurol. 27, 401 420. Barron K. D., Doolin P. F. and Oldershaw J. B. (1967) Ultrastructural observations on retrograde atrophy of lateral geniculate body. I. Neuronal alterations. J. Nruroputh. c.~p. Neural. 26, 300 326. Barron K. D., Means E. D. and Larsen E. (1973) Ultrastructural evidence of retrograde degeneration in thalamus of rat. I. Neuronal soma and dendrites. J. Ncuropcr/h. c’_\-p.Nwrol. 32, 21% 244. Baughman R. W. and Gilbert C. D. (1981) Aspartate and glutamate as possible neurotransmitters of cells in layer 6 of the visual cortex. Nu/ure 287, X48-849. Ben-Ari Y., Tremblay E.. Ottersen 0. P. and Naquet R. (1979) Evidence suggesting secondary epileptogenic lesions after kainic acid: pretreatment with diarepam prevents distant but not local brain damage. Bruin Rc.r. 165, 362%36S. Benveniste H.. Drcjer J.. Schousoboe A. and Diemer N. H. (1984) Elevation of the extracellular concentrations of glutamate and aspartate in the rat hippocdmpus during transient cerebral ischemia monitored by intracerebral microdialysis. J. Ncurodwn. 43, 1369 -I 376. N-methylaspartate, and other Berdichevsky E.. Riveros N.. Sanchez-Armass S. and Orrcgo F. (1983) Kainate, excitatory amino acids increase calcium influx into rat brain cortex cells in c+/ro. Neurosci. Let/. 36, 75 80. Beresford W. A. A. (1965) A discussion on retrograde changes in nerve tibres. In Mechcrrhm of Nrurd Rc~,wwrtr/iotl (eds Singer M. and Schade J. P.). Pro,?. Bruin Rrs. 14, 33 56. Bignami A. and Dahl D. (1976) The astroglial response to stabbing: immunofluorescent studies with antibodies to astrocyte specific protein (GFA) in mammalian and submammalian vertebrates. Nc%!roparh. (I/$. Neurohiol. 2, 99 110.

6. Barron

Thalamic

retrograde

degeneration

following

cortical

injury

547

16. Bjiirkhmd A. and Stenevi U. (1971) Growth of central catecholamine neurons into smooth muscle grafts in the rat mesencephalon. Bruin Res. 31, I-20. 17. Blintzinger K. and Kreutzberg G. (1968) Displacement of synaptic terminals from regenerating motoneurons by microglial cells. Z. Zellforsch. mikrosk. Anaf. 85, 146157. 18. Bourke R. S.. Kimmelberg II. K., Daze M. A. and Popp A. J. (1979) Studies on the fo~ation of astroglial swelling and its inhibition by clinically useful agents. In Neural Trauma: Seminars in Neurological Surgery (eds Popp A. J., Bourke R. S., Nelson L. R. and Kimmelberg H. K.), pp. 955112. Raven Press, New York. 19. Burchfiel J. L. and Duffy F. H. (1974) Corticofugal influence upon cat thalamic ventrobasal complex. Brain Res. 70, 395-41 I. 20. Cammermeyer J. (1969) Species differences in acute retrograde neuronal reaction of facial and hypoglossal nuclei. J. Hirnsfirsch. 11, 13-29. 21. Carman J. B., Cowan W. M. and Powell T. P S. (1964) Cortical connexions of the thalamic reticular nucleus. J. Anat. 98, 5877598. 22. Carreras M., Lechi A., Zampollo A. and Juvarra 0. (1969) Residual neurons of the thalamic ventrobasat nucleus. Archs ital. &of. 107, 576-603. 23. Choi D. (1985) Glutamate neurotoxicity in cortical cell cultures is calcium dependant. Neurosci. Letf. 7, 293-297. 24. Choi D. (1987) Ionic dependence of glutamate neurotoxicity. J. Neurosci. 7, 369.--379. 25. Chow K. L. and Dewson J. H. III (1966) Numerical estimates of neurons and glia in lateral geniculate body during retrograde degeneration. J. camp. Neural. 128, 63-74. 26. Clifford D. B., Benz A., Olney J. W. and Zorumski C. F. (1987) MK 801 prevents thalamic damage induced by focal cortical seizures. Sot. Neurosci. Ahsfr. 13, 1077. 27. Collins R. C. and Nandi N. (1982) Focal seizures disrupt the incorporation of amino acids into proteins: an autoradiographic study using l-14C leucine. Brain Res. 248, 102-120. 28. Collins R. C. and Olney J. W. (1982) Focal cortical seizures cause distant thalamic lesions. Science 218, 1’77-179. 29. Conti F., Rustioni A.. Petrusz P. and Towle A. C. (1987) Glutamate-positive neurons in the somatic sensory cortex of rats and monkeys. J. Neurosci. 7, 188?--1901. 30. Coombs C. M. (1949) Fiber and cell degeneration in the albino rat brain after hemidecortication. J. camp. Neural. 90, 3733401. 31. Coyle J. T. (1983) Neurotoxic action of kainic acid. J. Neurochem. 41, l-1 I. 32. Coyie J. T. and Schwartz R. (1976) Lesion of striatat neurones with kainic acid provides a model for Huntington’s chorea. Nafure 263, 244246. 33. Cragg B. C. (1970) What is the signal for chromatolysis? Brain Res. 23, I-21. 34. Danbolt N. C. and Storm-Mathisen J. (1986) Inhibition by K+ of Na+ dependent D-aspartate uptake into brain membrane saccules. J. Neurochem. 47, 825-830. 35. David S. and Aguayo A. J. (1981) Axonal elongation into peripheral nervous system “bridges” after central nervous system injury. Science 214, 931-933. 36. Deschenes M., Paradis M., Roy J. P. and Steriade M. (1984) Electrophysiology of neurons of lateral thalamic nuclei in cat: resting properties and burst discharges. J. Neurophysiol. 51, 1196-1219. 37. Dietzel I., Heinemann G., Hofmeier G. and Lux H. D. (1980) Transient changes in the size of the extracellular space in the ~nsorimotor cortex of cats in relation to stimulus induced changes in potassium concentration. E_xpf Brain Re.y. 40, 432439. 38. Donoghue J. P. and Wells J. (1977) Synaptic rearrangement in the ventrobasal complex of the mouse following partial cortical deafferentation. Brain Res. 125, 351-355. 39. Donoghue J. P., Kerman K. L. and Ebner F. F. (1979) Evidence for two organizational plans in the somatic sensory-motor cortex of the rat. 1. camp. IVeur~~. 183, 647-664. 40. Donoghue J. P., Wenthotd R. J. and Altschuier R. A. (1986) Localization of glutaminase-like and aspartate aminotransferase-like immunoreactivity in neurons of the cerebral neocortex. J. Neurosci. 5, 2597-2608. 41. Drejer J., Beneviste H., Diemer N. H. and Schousboe A. (1985) Cellular origin of ischemia-induced glutamate release from brain tissue in oivo and in rirro. J. Neurochem. 45, 146-151. 42. Eniarson L. (1960) Nucleic acids as structural constituents of nerve cells. In Modern Scienf$c Aspeefs of Neurology (ed. Cumings J. N.), pp. I-68. Edward Arnold, London. 43. Farien A. and Valverde F. (1980) A specialized type of neuron in the visual cortex of the cat: a Golgi and electron microscopic study of chandalier cells. J. camp. Neural. 194, 761-779. 44. Fonnum F. (1984) Glutamate: a neurotransmitter in mammalian brain. J. Neurochem. 42, l-l I. 45. Fonnum F., Storm-Mathisen J. and Divac I. (1981) Biochemical evidence for glutamate as a neurotransmitter in corticostriatal and corticothalamic fibers in rat brain. Neuroscience 6, 863-873. 46. Gardner-Medwin A. R. (1981) Possible roles of vertebrate neuroglia in potassium dynamics, spreading depression and migraine. J. e.wp. Biol. 95, 111-127. 47. Garthwaite J., Garthwaite G. and Hajos F. (1986) Amino acid neurotoxicity: relationship to neuronal depolarization in rat cerebellar slices. f~e~ro~~~ence 18, 449460. 48. Gill R., Foster A. C. and Woodruff G. N. (1987) Systemic application of MK 801 protects against ischemic induced hippocampal neurodegeneration in the gerbil. J. Neurosci. 7, 3343-3349. 49. Giuffrida R., Livolsi G. and Panto M. R. (1983) Pyramidal influences upon ventral thalamic nuclei in the cat. Brain Res. 279, 254257. 50. Griffiths T.. Evans M. C. and MeIdrum B. S. (1983) Temporal lobe epilepsy: excitotoxins and the mechanism of selective neuronai loss. In ~x~~fofo.~i~~(eds Fuxe K., Roberts P. and Schwartz R.), pp. 331-342. Macmillan, London. 51. Gutnick M. J. and Prince D. A. (1972) Thalamocortical relay neurons: antidromic invasion of spikes from a cortical epileptogenic focus. Science 176, 424426. 52. Gutnick M. J. and Prince D. A. (1974) Effects of projected cortical epileptiform discharges on neuronal activities in cat VPL. 1. Interictal discharge. J. ~e~ro~h.y.~io/. 37, 1310-1327. 53. Gutnick M. J., Heinemann U. and Lux H. D. (1979) Stimulus induced and seizure related changes in extracellular potassium concentration in the cat thalamus (VPL). LYectroenceph. din. Neurophysiol. 47, 329-344.

D. T. Ross and l-. F. EHK.EK

%lx

54. Haldeman S. and McLennan H. (1972) The antagonistic action of glutamic acid dicthyl Euler towards amino acid induced and synaptic excitations of central neurons. Brtrin RCA. 45, 393-400. 55. Haldeman S. and McLennan H. (1973) The actions of two inhibitors of glutamate uptake upon amino acid induced and synaptic excitations of thalamic neurons. Bruir~ Re.v. 63, I23 129. 5h. Ilaldeman S., Huffman R. D. and McLennan H. (1972) the antagonism of the glutamate induced and synaptic excitations of thalamic neurons. Brrri~~ Rex. 39, 419 425. 57. Hansen A. .I. and Olsen C. E. (1980) Brain extracellular space during spreading depression and ischemia. Acrcr p/r\,.tio/. \c,trrlt/. 108, 355 365. SX. Hansen A. J. and Zeuthcn T. (1081) Extracellular ion concentration during spreading depresston and ischemia in the rat brain cortex. Ac/tr $~.rrol. .rccr&. 113, 437 445. 5Y. Harris R. M. and Hendrickson A. E. ( 1985) Local circuit neurons in the rat ventrobasal thalamus. Sot,. Ncuro.sc.r. ilh.vrr. Il. 561 60. Heinemann U. and Gutnick M. J. (1979) Relation between extracellular potassium concentration and neuronal activities in cat thalamus (VPL) during projection of cortical epileptiform discharge. EIcc~~roc~nc,q~/~. c/in. Nmrophv.\iol. 41, 345 351. 61. Houser C. R.. Vaughn J. E., Barber R. A. and Roberts E. (19X0) GABA neurons are the major cell type in the thalamic reticular nucleus. Bruin Rrs. 200, 341 354. 62. Jahnsen H. and Llinis R. (19X4) Electrophysiological properties of guinea-pig thalamic neurons: an in c.itro study. .J. Pl~~.\iol.. Lord. 349, 205 226. 63. Jahnscn H. and Llmas R. (1984) Ionic basis for the electroresponsiveness and oscillatory properties of guinea-pig thalamic neurons BI ri/ro. ./. Phy.viol., Land. 349, 237 247. 64. Jones E. G. (1975) Varieties and distribution of nonpyramidal cells in the somatic sensory cortex of the squirrel monkq J. cony. Nrurd. 160, 205 267. 65. Jones E. G. (1975) Some aspects of the organization of the thalamic reticular complex. J. (‘onq. N~,uro/. 162, 2X5 30X. 66. Jones E. G. (1985) Tl~r Thalamu.c. Plenum Press, New York. 67. Kearns J. M. and Hinsman E. .I. (1973) Neuroglial responses to sciatic neurectomy II. Electron microscopy. J. ~nrp. ‘vwtY~/. 151, 255~ 279. 6X. Kromer L. F.. Bjiirklund A. and Stenevi U. (1981) Innervation of embryonic hippocampal implants by regenerating axons of cholincrgic septal neurons in the adult rat. Bruirl Rrs. 210, I53 171. 69. Kuno M. and Llinas R. (1970) Enhancement in synaptic transmission of dendritic potentials in chromatolyscd motoneurons of the cat. J. P/~~wo/.. Lad 210, 807 821. 70. Latov N.. Nilaver G.. Zimmerman E. A.. Johnson W. G.. Silverman A.-J.. Defcndini R. and Cote L. (1979) Fibrillar> astrocytes proliferate in response to brain injury. Deal Viol. 72, 381 384. 71. Lcavitt P. and Rakic P. (1980) lmmunoperoxidase localization of glial fibrillary acid protein m radial glial cells and astroglia of the developing rhesus monkey brain. J. c’on~p. Newel. 193, 815 X40. 72. LeGros Clark W. E. and Russel W. R. (1940) Atrophy of the thalamus in a case of acquired hemiplegia associated with diffuse porencephaly and sclerosis of the left cerebral hemisphere. J. Nrurol. P.cychiut. 3, I23 140. 73. Lieberman A. R. (1971) The axon reaction: a review of the principal features of the perikaryal response to axotom!. /)I/. Rw. ~Vczmhiol. 14, 49 11-4. 74. Llinas R. and Jahnsen F-l. (1982) Electrophysiology of mammalian thalamlc neurons 1,~ ri/ro. ~Yoruw 297, 406 40x. 75. Lucas D. R. and Newhousc J. P. (1957) Toxic effect of sodium L-glutamate on the inner layers of the retma .,lr&. Op//?trl!llo/. 58, I93 20 I 76. Lund-Karlsen R. and Fonnum F. (197X) Evidence for glutamate as a neurotransmitter in the corticofugal tibers to the dorsal lateral geniculate body and the superior coll%ulus in rats. Bruin Res. 151, 457-467. 77. MacDermott A. B.. Maver M. L.. Westbrook G. C.. Smith S. J. and Barker J. Z. (1986) NMDA receutor activation increases cytoplasmic c&urn concentration in cultured spinal cord neurones. Nuture j21, 519 522. ’ 7X. Madarasr M.. Somogyi J.. Silakob V. L. and Hamori J. (19X3) Residual neurons in the lateral geniculate nucleus ol adult cats following chronic disconnection from the cortex. E.rp/ Bruin RCT. 52, 363 374. 70. Mad1 J. E.. Larson A. E. and Beitz A. J. (1986) Monoclonal antibodies specific for carbodimide fixed glutamateImmunocytochemicdl localization in the rat CNS. J. His/odwm. Cy/ochen~. 34, 317 326. X0. Mathewson A. J. and Berry M. (1985) Observations on the astrocytic response to a cerebral stab wound m adult rats. UI.N;II Rcs. 327, 61 6’). Xl. Matthews M. A. (1973) Death ofthe central neuron: an electron microscopic study of thalamic retrograde degencratlon folloumg cortical ablation. J. N~woc~~~tol. 2, 265~-288. X2. Matthews M. A. and Faciene C. C. (1977) Electron microscopy of the development of synapttc pattern> 111the ccntrobasal complex of the rat. Brcrlfl Rcs. 135, 197 215. X3. Matthews M. A.. Nardyanan C. H.. Narayanan Y. and St. Onge M. F. (1977) Neuronal maturation and synaptogencsis in the rat ventrobasal complex: alignment with developmental changes in the rate and severity of the axon reaction. J. con?,,. Neural. 173, 745 771 X4, McAllister J. P. and Wells J. (1981) The structural organization of the ventroposterolateral nucleus of the rat. .I. c’f,nl,~. !Vc,rrro/. 197, 271 301. X5. McGeer E. G. and McGeer P. L. (1976) Duplication of biochemical changes of Huntington’s chorea by intrastriatal Injections of glutamic and kainic acids. Ntirurc 263, 517 519. 86. Minkowski M. (1913) Experimentalle Untersuchengen uber die besondes zun Corpus geniculatum externum. .-lrh. Himmur. X7. Molinari

Imr.

Lurk/?

%en/ren

7, 255~ 362.

M., Minciacchi D.. Bentivoglio M. and Macchl G. (19X5) Efferent fibers from the motor cortex terminate bilaterally in the thalamus of rats and cats. Espl Bruin Rcs. 57, 305 312. XX. Mnntcro V. M. and Guillery R. W. (1968) Degeneration of the dorsal lateral geniculate nucleus of the rat following interruption of the retinal or cortical connections. J. c’0nl,,. ,Yruro/. 134, 21 I 241. 89. Mugnaini E. and Dahl A. L. (1983) Zinc-aldehyde fixation for light-microscopic immunocytochemlstry of nervou\ tlssuc. .I. Hi.c/ohnr. C’~~/odrern. 31, 1435 1438. 90. Mugnalni E.. Oertcl W H.. Schmechel D. E.. Tappar M. L. and Kopin I. J. (19X2) The nnmunocytochemlcal

Thalamic

91. 92.

93. 94. 95.

96. 97. 98. 99. 100. 101. 102. 103. 104.

105. 106. 107. 108. 109.

110.

I I 1. I 12. 113. 114.

I IS. 116. 117. 118. 119. 120. 121. 122. 123.

retrograde

degeneration

following

cortical

injury

549

demonstration of gamma-amino butyric acidergic neurons-methods and application. In Cytochemicul Methods in Neurounufomy, pp~ 2977329. Alan R. Liss, New York. Nicholson C., ten Bruggencate G.. Stockle H. and Steinberg R. (1978) Calcium and potassium changes in the extracellular microenvironment of rat cerebellar cortex. J. Neurophysiol. 41, 102GIO39. Nicholson C., Phillips J. M., Tobiasz C. and Kraig R. P. (1981) Extracellular potassium and calcium and volume protiles during spreading depression. In Ion Sensitive Microelectrodes and Their Use in Excitable Tissue (eds Syskova V., Hrik P. and Vylicky L.). Plenum, New York. Nobels F. L. and Prince D. A. (1978) Excitability changes in thalamocorti~dl relay neurons during synchronous discharges in cat neocortex. J. Neuroph,ysiol. 41, 1282-1296. Ogden T. E. (1960) Cortical control of thalamic somatosensory relay nuclei. Elecrroenceph. clin. Neurophysiol. 12, 62 I-634. Ohara P. T., Lieberman A. R., Hunt S. P. and Wu J.-Y. (1983) Neural elements containing giutamic acid decarboxyiase (GAD) in the dorsal lateral geniculate nucleus of the rat: immunohistochemical studies by light and electron microscopy. Neuroscience 8, 189-2 12. Olney J. W. (1978) Neurotoxicity of excitatory amino acids. In Kuinic Acid as a Tool in Neurobiology (eds McCeer E. G., Olney J. W. and McGeer P. L.), pp. 95-121. Raven Press, New York. Olney J. W. (1984) Excitotoxins: an overview. In E~~i,itoto.~in.~ feds Fuxe K., Roberts P. and Schwartz R.), pp. 82-96. Macmillan. London. Olney J. W., Rhee V. and Ho 0. L. (1974) Kainic acid: a powerful neurotoxic analog of glutamate. Brain Res. 77, 507-512. Ottersen 0. P. and Storm-Mathisen J. (1984) Glutamate and GABA containing neurons in the mouse and rat brain as demonstrated with a new immunohistochemical technique. J. camp. Neural. 229, 374-391. Peters A., Proskauer C. C. and Ribak C. E. (1982) Chandalier cells in the rat visual cortex. J. romp. Neural. 206, 397-416. Powell T. P. S. (1952) Residual neurons in the human thalamus following hemidecortication. Bruin 75, 571.-583. Powell T. P. S. and Cowan W. M. (1964) A note on retrograde fibre degeneration. J. Anaf. 98, 579-585. Price M. T., Olney J. W., Samson L. and Labruyere T. (1985) Calcium influx accompanies but does not cause excitotoxin induced neural necrosis in the retina. Brain Res. Bull. 14, 369--376. Ralston H. J. III (1983) The synaptic organization of the ventrobasal thalamus in the rat, cat, and monkey. in Somato.~en~~[~r~ integrations in thr Thalomu.~ feds Macchi G., Rustioni A. and Spreaftco R.), pp. 241-250. Elsevier, Amsterdam. Ramon y Cajal S. (191 I) Histologic du Sjxteme Nerveu.\-de I’Homme et des Vertebres, Vol. II. Trans. by L. Azoulay. Maloine. Paris. Ramon y Cajal S. (1928) Degeneration und Regenerufiun in the Nervous Swtem, Vols. I-II. Trans. by R. May. Hafner, New York. Rose J. E. and Woolsey C. N. (1949) Organization of the mammalian thalamus and its relationship to the cerebral cortex. Electroenceph. clin. Neurophysiol. 1, 391401. Ross D. T. (1987) Antagonism of normal and abnormal activity of relay neurons in the rat’s ventrobasal thalamus using excitatory amino acid antagonists. Sot. ~e~~os~j. Abstr. 13, 762. Ross D. T. and Das G. D. (1985) Regenerative growth of retinofugal axons into and through neocortical transplants following transection of the optic tract in adult rats. In Neural Tissue Transplantation and Regeneration (eds Wallace R. B. and Das G. D.), pp. 181-227. Springer, New York. Ross D. T. and Ebner F. F. (1986) Increased excitability of thalamic relay neurons in the ventrobasal complex following ablation of the SI cortex in the adult rat. Sot. Neurosci. Abstr. 12, 983. Rothman S. M. (I 984) Synaptic release of excitatory amino acid neurotransmitters mediates anoxic neuronal death. J. Neurosci. 4, 18841891. Rothman S. M. (1985) The neurotoxicity of excitatory amino acids is produced by passive chloride ingux. J. Neurosri. 5, 1483-1489. Rothman S. M., Thurston J. H. and Hauhart R. E. (1987) Delayed neurotoxicity of excitatory amino acids in rifro. Neuroscience 22, 47 1480. Rustioni A., Schmechel D. E.. Spreafico R., Cheema S. and Cuneod M. (1983) Excitatory and inhibitory amino acid putative neurotransmitters in the ventralis posterior complex: an autoradiographic and immunocytochemical study in rats and cats. In Somatosensor~ Integrution in the Tlzol~~nzu.~ (eds Macchi G.. Rustioni A. and Spreafico R.), pp. 365-383. Elsevier, Amsterdam. Salt T. E. (1986) Mediation of thalamic sensory input by both NMDA and non-NMDA receptors. Nature 322, 263-265. Saporta S. and Kruger L. (1977) The organization of thalam~ortical relay neurons in the rat ventrobasdl complex studied by the retrograde transport of horseradish peroxidase. J. camp. Neuroi. 174, 187-208. Schiebel M. E. and Schiebel A. B. (1966) The organization of the nucleus reticularis thalami: a Golgi study. Bruin Res. 1, 4342. Schlaepfer W. W., Zimmerman U.-J. P. and Micko S. (1981) Neurofilament proteolysis in rat peripheral nerve: homologies with calcium activated proteolysis of other tissues. Ce# Co&urn 2, 235-250. Schwartzkroin P. A.. Mutani R. and Prince D. A. (1975) Orthodromic and antidromic effects of a cortical epileptogenic focus on the ventrolateral nucleus of the cat. J. Neurophysiol. 38, 795-81 I. Schwab J. E., Fuller T., Price M. L. and Olney J. W. (1980) Widespread patterns of neuronal damage following systemic or intracerebral injections of kainic acid: an histolo~~al study. neuroscience 5, 991llOI4. Shimazu N.. Yanagisawa N. and Garoutte B. (1965) Cortico-pyramidal influences on thalamic somatosensory transmission in the cat. Jup. J. Physiol. 15, 101-124. Siesjo B. K. and Wieloch T. (1985) Brain injury: neurochemical aspects. In Cenrral Nervous System Trauma Stutus Report. NIH-NINCDS, pp. ,513.--532. Simon R. P.. Griffiths T., Evans M. C., Swan J. H. and Meldrum 8. S. (1984) Calcium overload in selectively vulnerable neurons of the hippocampus during and after ischemia: an electron microscopic study in the rat. J. cerebr. Blood Flow Me&b. 4, 350-361.

‘50

11. T. Ross and F. F. LHXl’.K

174. Simon R. P., Swan J. H.. Grifhths J. H. and Meldrum B. S. (1984) Blockade of N-methyl-aspartate receptors may protect against ischemic damage in the brain. Science 226, X50 X52. 125. Singer W. (1977) Control of thalamic transmission by corticofugal and ascending reticular pathways in the visual system. Phr.riol. Rw 57, 386420. 126. Singer W. and Lux H. D. (1975) Presynaptic depolarization and extracellular potassium in the cat lateral geniculate nucleus. Broirt Re.c. 64, I7 33. 137. Somjen G. G. (1979) Extracellular potassium in the mammalian CNS. A. Rec. Physiol. 41, 159~ 177. 128. Somogyi P. (1977) A specific axo-axonal interneuron in the visual cortex of the rat. Brain Res. 136, 3455350. 129. Somogyi P.. Hodgson A. J. and Smith A. D. (1979) An approach to tracing neuron networks in the cerebral cortex and basal ganglia: combination of Golgi staining, retrograde transport of horseradish peroxidase and anterograde degeneration of synapttc boutons in the same material. Neuro.vcirttcr, 4, 1805 1852. 130. Spacek J. and Lieberman A. R. (1974) Ultrastructure and three dimensional organization of synaptic glomeruli in rat somatosensory thalamus. /. Anur. 117, 487 -516. I3 I. Streit P. (I 980) Selective retrograde labelling indicating the transmitter of neuronal perikarya. J. c’omp. Neural. 191, 329 463. 132. Sugitani M. (1979) Electrophysiological and sensory properties of the thalamic reticular neurones related to somatic sensation in rats. b. Ph~siol. 290, 79 95. 133. Sunderland S. (1968) ~Vert,cz.vuntf Nrrnc fn&rir.s. Churchill Livingstone. London. 134. Szentagothai J. and Arbib M. A. (1974) Conceptual models of neural organization. Neuro.sc,i. RH. Progr. Bull. 12, 307 5 IO. ITS. Tello F. (I91 I) La inhuencia del neurotropismo en la regeneration de 10s centros nerviosos. Truh. Inst. Cajul hrt,.r/. hid. 5, I73 159. 136. Van Gelder N. M. (1983) Metabolic interaction between neurons and astroglia: glutamine synthetase, carbonic anhydrase and water balance. In Basic Mrchunims of Neuronul H~perrxc~i/ohi/i/~~ (eds Jasper H. H. and van Gelder N. M.). pp, 5- 29. Alan R. Liss, New York. 137. Walker A. E. (1938) The Primure Thulumus. University of Chicago Press, Chicago, IL. 138. Walker A. E. (1966) Internal structure and afferent relationships of the thalamus. In The Thu/umus (eds Purpura D. and Yahr M. D.), pp. I 12. Columbia University Press, New York. 139. Wailer H. J. and Feldman S. M. (1967) Somatosensory thalamic neurons: effects of cortical depression. Srienc~ 157, 1074 1077. 140. Watson W. E. (1968) Observattons on the nucleolar and total cell body nucleic acid of injured nerve cells J. Phr.vio/.. I>otrtl. 196, 655 -672. 141. Wang-Riley M. (1972) Changes in the dorsal lateral geniculate nucleus of the squirrel monkey after unilateral ablation of the visual cortex. J. camp. Neural. 146, 519-548. 142. Wong-Riley M. (1979) Changes in the visual system of monocularly sutured or enucledted cats demonstrated with cytochrome oxidase histochemistry. Bruin Rex. 171, I I 2X. 133. Yen C. T. and Jones E. G. (1983) Intracellular staining of physiologically identified neurons and axons in the somatoscnsory thalamus of the cat. Bruin Res. 280, 148 154. 144. Yoon M. and Ross D. T. (1989) Reaction of protoplasmic astrocytes in the adult rat thalamus to ablation of the somatosensory cortex. And. Rw. 223, 12SA. 145. Young A. B.. Bromberg M. B. and Penney J. B. (1981) Decreased glutamate uptake in subcortical areas deafferented by scnsorimotor cortical ablation in the cat. J. /Veuro.sci. I, 241 ~249. 146. Zacrek R.. Simonton S. and Coyle J. T. (1980) Local and distant neuronal degeneration following intrastriatal inJectton of kainic acid. J. ,Ycuropurh. c’.\-p. Neural. 39, 245 264. (/l
19X9)