Structural abnormalities in neurones - BMJ

J. clin. Path., 30, Suppl. (Roy. Coll. Path.), 11, 155-169

Structural abnormalities in neurones

A. W. BROWN

From the Medical Research Council Laboratories, Carshalton, Surrey

A wide variety of neuronal alterations ascribed to theeffect of hypoxia of different types has been des-cribed in man and in several species of experimentalanimals (Spielmeyer, 1922; Gildea and Cobb, 1930;Tureen, 1936; Hoff et at, 1945; Morrison, 1946;Grenell and Kabat, 1974; Krogh, 1952; Courville,1951; Meyer, 1963). In a comprehensive review of theliterature, Hoff et al (1945) concluded that thechanges were remarkably alike whatever the cause ofhypoxia. These changes, probably best seen with theNissl stain, range from non-specific forms where thecells appear very pale or are undergoing swelling andloss of stainable substance to the classical descrip-tions given by Spielmeyer (1922). These latterchanges, which include 'coagulation necrosis' or'ischaemic cell change', 'homogenizing cell change'and 'liquefaction necrosis', have been discussed interms of more modern views by Greenfield andMeyer (1963) and Brierley (1976). Nissl's 'acute celldisease' (Spielmeyer's 'acute swelling') and Nissl's'chronic cell degeneration' must be added to theabove neuronal alterations because they are so oftenmentioned as evidence of hypoxia and ischaemia.From this wide range of neuronal alterations

Spielmeyer's ischaemic cell change, the term origi-nally used to describe the alteration in nerve cellsafter general or local circulatory arrest, is nowrecognized as the degenerative response that is com-mon to all types of hypoxia. Thus it is also seen insubstrate deficiency (hypoglycaemia), anaemichypoxia represented by carbon monoxideintoxication, histotoxic hypoxia as exemplified bycyanide poisoning and the complex situation repre-sented by status epilepticus. However, it is importantto stress that the patterns of distribution of ischaemiccell change in the brain may differ widely amongthe categories of hypoxia.The great diversity ofthe morphological alterations

in neurones attributed to hypoxia is probably theoutcome of postmortem autolytic changes unavoid-able in human material, where there is a variabledelay between death and necropsy coupled with slowpenetration of the fixative, and to the use of immer-sion-fixation in experimental animals with theinevitable introduction of histological artefacts. ThusHicks (1968) described diminished staining of Nissl

bodies and other eosinophilic changes as the earliestneuronal alterations in anoxic necrosis in humanmaterial but considered them to be indistinguishablefrom postmortem changes.

There are two common cytological artefactsencountered in human and experimental animalbrains each of which has been interpreted as evidenceof hypoxic damage and reported as such. The hyper-chromatic neurone or 'dark cell' has been the subjectof numerous studies, including those of Scharrer(1938), Wolf and Cowaen (1949), Koenig and Koenig(1952), Cammermeyer (1960, 1961, 1962) and Cohenand Pappas (1969). This artefact is most frequentlyobserved after a fresh neurosurgical biopsy specimenis fixed in formaldehyde, when the brain of a normalanimal removed immediately after death is immersedin a fixative, or when a perfusion-fixed brain isremoved immediately from the skull. In paraffin andcelloidin sections the dark cells appear heavilystained and unevenly shrunken (fig 1). It is generallyagreed that dark cells result from postmortemtrauma incidental to removal of the brain. Thesecond type of artefact is known as 'hydropic cellchange' or 'water change' (Jakob, 1927; Cammer-meyer, 1960, 1961; Coimbra, 1964). In contrast tothe hyperchromatic neurone the hydropic cell isswollen and the cytoplasm stains less intensively thannormal (fig 2). It is sometimes seen in humanmaterial and particularly in infants. It occurs inimmersion-fixed animal brain and is a particularhazard in those that are incompletely or inadequatelyperfusion-fixed.Both the above artefacts and also artefactual peri-

neuronal and perivascular spaces can be excluded inthe experimental animal by the employment of care-fully controlled in-vivo intravascular perfusion-fixation and delayed removal of the brain. We haveemployed this method of fixation for our light andelectron microscopic studies of the evolution ofischaemic cell change in rodents and primates in thefollowing experimental models (1-6).

1 The preparation developed by Levine (1960) inthe rat which allowed ischaemic neuronal alterationsto be produced unilaterally when ligature of a com-mon carotid artery was followed by intermittentexposure to nitrogen was modified by Brown and

155

copyright. on January 17, 2022 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.s3-11.1.155 on 1 January 1977. Dow

nloaded from

http://jcp.bmj.com/

A. W. Brown

Fig 1 Immersion-fixed rat neocortex showing artefact ofhyperchromatic or dark neurones. One neurone hascorkscrew-like apical dendrite. Paraffin. Cresyl fastviolet. x 530.

Fig 2 Rhesus monkey. Inadequate perfusion-fixation ofcerebral cortex showing artefact of hydropic cell changewith swelling and peripheral vacuolation of thecytoplasm. Paraffin. Luxol fast blue and cresyl fastviolet. x 800.

Brierley (1966). The removal of a plastic 'carotidclasp' at the end of the exposure ensured therestoration of blood flow and subsequently the equalperfusion-fixation of both sides of the brain (Brownand Brierley, 1966, 1968, 1972, 1973).

2 Normal rats intermittently exposed to nitrogen(Brown and Brierley, 1966, 1968, 1972).

3 Squirrel monkeys (Saimiri sciureus) intermit-tently exposed to nitrogen with and without inter-ruption of blood flow in the right common carotidartery (Brown, 1973).

4 In the rhesus monkey (Macaca mulatta)ischaemic neuronal damage resulted from profoundarterial hypotension induced by a ganglion-blockingagent (Arfonad), head-up tilt and arterial bleeding(Brierley and Excell, 1966; Brierley et al, 1969). Afterrespiratory arrest mechanical ventilation ensurednormal arterial oxygenation. This model was used ina study restricted to mild cortical boundary zonelesions and the hippocampi (Brown, 1973).

5 Insulin-induced hypoglycaemia in M. mulattawithout hypoxaemia, hypotension or epilepsy(Brierley et al, 1971a and b).

6 Unilateral and bilateral interruption of bloodflow in the carotid arteries of gerbils (Brierley,Brown and Levy, in the press).For our light microscope studies we employed

formaldehyde, glacial acetic acid and absolutemethanol (FAM), 1:1:8 (David, 1955), and for com-bined light and electron microscopy 6 5 per centglutaraldehyde in cacodylate buffer, pH 7-4. Afterperfusion-fixation with FAM and glutaraldehyde thebrains were left in situ for two to three hours and onehour respectively.While it is recognized that the study of the very

earliest stages of ischaemic cell change in theinevitably immersion-fixed human brain is madedifficult by the presence of histological artefacts, thelater classical stages leading to cell loss and eventualgliomesodermal reaction are readily identified pro-vided the tissue is properly processed and stained.Thus our observations in the above experimentalmodels have been combined with those of DrBrierley in human brains in the following descriptionof the evolution of ischaemic cell change.

Ischaemic cell change

This is a destructive process with a definite timecourse that begins with an alteration in a single typeof organelle and ends with the disappearance of theneurone. Until 1966 the time course of this cellchange in the mammalian brain (including rat,monkey and man) remained unknown and it wasregarded as a relatively slow process. Thus the viewwas widely held that examination of human and

156copyright.

on January 17, 2022 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-11.1.155 on 1 January 1977. D

ownloaded from

http://jcp.bmj.com/


experimental animal brain material where the survivalperiod after a hypoxic stress was less than 12 to 24hours would be unrewarding with regard to demon-strating unequivocal neuronal alterations (Polaniand MacKeith, 1954; Richardson et al, 1959;Spector, 1961). In contrast, our investigations(Brown and Brierley, 1966, 1968, 1972, 1973) haveshown that ischaemic cell change is an extremelyrapid process. It is important to stress that its timecourse is not influenced by the severity of thehypoxic stress but only by the size of the cell, beingmore rapid in small neurones than in large. Differ-ences in severity of the hypoxic insults are reflectedonly in the number of nerve cells involved. It is wellestablished that ischaemic cell change is not randomlydistributed in the brain but occurs in regions whichexhibit 'selective vulnerability' to hypoxic stresses.Thus in all the types of hypoxia mentioned earlierthere is more or less involvement of some of thefollowing regions: (1) cerebral cortex, layers 3, 5 and6; (2) hippocampus, Sommer sector and endfolium;(3) amygdaloid nucleus, central and basolateralportions; (4) cerebellum, Purkinje and basket cells;(5) brain stem; (6) thalamus; (7) striatum; (8) pal-lidum and (9) subthalamic nucleus. The spinal cordis the most resistant.

In all our experimental models brain damage wasvariable with the greatest range of severity in models1, 2 and 6. In all models light microscopic examina-tion showed that classical ischaemic cell change wasalways preceded by a stage we have called 'micro-vacuolation' (Brown and Brierley, 1966). Micro-vacuolation has also been observed in the primatebrain after periods of status epilepticus (Meldrumand Brierley, 1973), and also in the brains of humansubjects dying one hour after cardiac arrest and threeto four hours after open-heart surgery (Brierley,1976).

In a neurone showing microvacuolation (fig 3) thenucleus is either normal or slightly shrunken, thenucleolus appears normal and shrinkage of the cellbody is appreciable only in the neocortex. Themicrovacuoles appear as apparently empty circularor oval spaces (0 16-2-5 ,um diameter) withincytoplasm of normal or slightly increased basophiliaand frequently extend into the proximal portions ofthe dendrites (fig 4). The electron microscope hasshown that the majority of microvacuoles are swol-len mitochondria (fig 5) which retain their doublemembranes despite progressive disruption of theirinternal structure (McGee-Russell et al, 1970; Brownand Brierley, 1972). Other microvacuoles correspondto dilated tubules and cisternae of the endoplasmicreticulum. There is an apparent accumulation ofabundant free ribosomes and ribosomal rosettescoupled with an increase in cytoplasmic matrix

density. Some of the neurones are more or less sur-rounded by the electron-lucent expanded processesof astrocytes (perineuronal spaces of light micro-scopy).

Microvacuolation has been seen in hippocampal(fig 6) and neocortical neurones of rat 'Levinepreparations' killed by perfusion-fixation immedi-ately after exposures to nitrogen of only five and 15minutes (Brown and Brierley, 1973). After fiveminutes a minority of the damaged neurones weresurrounded by swollen astrocytic processes whilethis was true of the majority after 15 minutes'exposure. Microvacuolation observed 16 minutesafter a period of hypotension and 30 minutes afterone of hypoglycaemia in the neocortex of rhesusmonkeys (Brierley et al, 1971a) and a similar altera-tion seen in the neocortex of a squirrel monkey killedimmediately after 30 minutes' intermittent exposureto nitrogen (Brown, 1973) represent the earliestevidence of neuronal damage after a hypoxic stress inthe primate. In all species examined microvacuola-tion can be seen earlier and persists for a shorter timein small neurones than in large in which it ceases tobe recognizable after about four hours. It has beenobserved in Purkinje cells up to two hours after aperiod of hypoxia in experimental primates and ina human subject surviving one hour after cardiacarrest (Brierley, 1976).

There is a gradual transition from the stage ofmicrovacuolation to that of classic ischaemic cellchange. The cell body is variably shrunken, and,together with its processes, stains more or lessintensively with aniline dyes (fig 7). The Nissl sub-stance is dispersed and finely granular and the cyto-plasm usually stains a vivid pink with eosin and blueor mauve with Luxol fast blue. The presence of a fewmicrovacuoles represents a transitional stage fromthat of microvacuolation. The nucleus is shrunken,often triangular and darkly staining with aniline dyesand dark blue with Luxol fast blue so that thenucleolus may be difficult to distinguish. In electronmicrographs vacuoles of various shapes, somederived from swollen mitochondria and others fromthe expanded cisternae and vesicles of both smoothand granular endoplasmic reticulum, contribute tothe fenestrated appearance (residual microvacuoles)of the dense cytoplasm (fig 8). Remnants of cristaecan be discerned in those vacuoles derived fromswollen mitochondria while the expanded smoothmembranes of the Golgi complex enclose very littlecondensed material. Occasionally undistended mem-branes of the Golgi complex participate in thesequestration of an area of cytoplasm reminiscent ofthe formation of an autophagic vacuole. Ribosomescan be identified on the expanded membranes of thegranular endoplasmic reticulum and abundant free

157copyright.


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

A. W. Brown

Fig 3 Rat Levine preparation, survival 30 minutes after40 minutes' intermittent exposure to nitrogen. Micro-vacuolation in ipsilateral hippocampus. The cell bodies areoccupied by small vacuoles separated by stronglybasophilic cytoplasm. The nuclei appear normal.Paraffin. Cresyl fast violet. x 800.

Fig 4 As figure 3, survival one hour. Ipsilateralhippocampus showing evenly distributed microvacuolesin pyramidal neurones also extending into dendrites.Perineuronal spaces are subdivided by trabeculae. 1 ,umEpon section. Toluidine blue. x 2000.

Fig 5 As figure 4. Electron micrograph ofa microvacuolated hippocampal pyramidal neurone. Swollen mitochondriawith disorganized cristae lie within cytoplasm of increased matrix density. The nucleus shows some distortion and anincrease in electron density. x 5200.

158



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


Fig 7

Fig 6 Rat Levine preparationperfusion-fixed immediately after fiveminutes' intermittent exposure tonitrogen. Electron micrographshowing swollen mitochondria innormal-looking cytoplasm ofapyramidal neurone of hippocampus.x 9500.

Fig 7 Rhesus monkey, three-hoursurvival after hypotensive episode.Ischaemic cell change in largepyramidal neurones in parietalcortex. The cell bodies are shrunkenand the nuclei are triangular anddarkly stained. Celloidin. Luxolfast blue and cresyl fast violet.x 625.

Fig 8 As figure 7. Electronmicrograph showing ischaemic cellchange with residual vacuoles in aneurone in parietal cortex. Swollenmitochondria with remnants of cristaeand dilatations of the endoplasmicreticulum can be seen in th?extremely dense cytoplasm. Thenucleolus is visible in the densenucleoplasm. Swollen astrocyticprocesses surround the damagedneurone. x 4400.

Fig 6

159copyright.


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

A. W. Brown

ribosomes and ribosomal rosettes are scatteredthroughout the electron-dense cytoplasmic matrix.Breakdown of the parallel arrays of cisternae of thegranular endoplasmic reticulum accounts for thedissolution of discrete Nissl bodies as visualizedwith the light microscope. In some neurones theincrease in electron density of both nucleus and cyto-plasm makes it difficult to identify the nucleolus andnuclear membrane. The damaged neurones areusually surrounded by the swollen processes ofastrocytes containing few organelles. These processesfrequently impart a distorted contour to the damagedneuronal cell membrane.

This stage has been identified in two rat Levinepreparations and a single primate killed immediatelyafter exposures to nitrogen of only 15 and 30minutes respectively (Brown and Brierley, 1973;Brown, 1973). It evolves more rapidly in small than inlarge neurones where it can persist up to 24 hours. Inthe human brain this stage persists for at least sixhours (Brierley, 1976).

In the subsequent stage of ischaemic cell changewith incrustations further shrinkage of the neuronalcytoplasm and nucleus occurs. The cytoplasm and

Fig 9Fig 9 As figure 3. Survival 24 hours. Ischaemic cellchange with incrustations in hippocampal pyramidalneurones. Incrustations can be seen on the cell bodiesand their processes. Paraffin. Cresyl fast violet. x 600.

Fig 10 As figure 3. Survival two hours. Low-powerelectron micrograph showing ischaemic cell change withincrustations in a neocortical neurone. Vacuolatedelectron-dense portions of the perikaryon and dendritesaround the dense nucleus correlate with the incrustationson light microscopy. x 3900.

apical dendrite continue to stain pink with eosin andblue with Luxol fast blue. Typical heavily stained,small, spherical or irregular bodies (incrustations)lie on or close to the surface of the cell perikaryonand sometimes those of the dendrites (fig 9). At theultrastructural level both nucleus and cytoplasmshow a marked and comparable increase in electrondensity (fig 10). Within the electron-dense nucleo-plasm areas of extreme density and of variable sizeand shape are present as well as clusters of dense (41-51 nm) granules. The cytoplasm is invaginated byswollen astrocytic processes. Frequently only a nar-row rim of vacuolated dense cytoplasm remainsaround part of the shrunken nucleus (fig 10). Denseareas of cytoplasm occurring in the peripheralportions of the perikaryon and in the cell processescorrespond to the typical incrustations seen with thelight microscope. The many variable-sized vacuolesscattered throughout the dense cytoplasm probablyrepresent dilated cisternae and vesicles of bothsmooth and granular endoplasmic reticulum as wellas the remains of swollen mitochondria (figs 11, 12).Within the dense cytoplasm it is possible to identifyribosomes lying free and on the expanded mem-

Fig 10

160copyright.


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/


W,T O.

Fig 11 Fig 12Fig 11 As figure 10. Electron-opaque profiles of neuronal cytoplasm (arrows) within an astrocytic process contactingan incrusted neurone. The occasional absence of the plasmalemmas suggests phagocytosis. x 12 400.Fig 12 Norma rat intermittently exposed to nitrogen for 40 minutes, survival 45 minutes. Ischaemic cell changewith incrustations showing a mitochondrion in association with vacuolated granular material and surrounded by a singlemembrane (arrow). Axon terminals appear distorted by adjacent expanded astrocytic processes. x 14 300.

branes of the endoplasmic reticulum. Extremelydense mitochondria, multivesicular bodies, densebodies (small lysosomes), residual bodies (largelysosomes) and occasionally mitochondria in asso-ciation with granular endoplasmic reticulum orvacuolated granular aggregates and surrounded bysingle membranes (autophagic vacuoles) can also beseen (fig 12). Many of the vesicle-containing axonalprofiles in contact with the damaged neurones appearnormal with recognizable synaptic contacts whileothers appear distorted by adjacent expanded astro-cytic processes (fig 12). Some show depletion anddissolution of synaptic vesicles and other evidenceof degeneration. Electron-opaque profiles ofneuronalcytoplasm are occasionally seen within the expandedprocesses of astrocytes around the damaged cells (fig11). The occasional absence of their respective plas-malemmas suggests a process of phagocyto sis.

Ischaemic cell change with incrustations has beenidentified in the rat brain after a survival of 30minutes (Brown and Brierley, 1968) and after 90minutes in the rhesus monkey (Brierley et al, 1971b).It persists in large neurones of the rat up to about 24hours and in those of the monkey brain up to about

48 hours (Brown and Brierley, 1968; Brierley, 1972,1973).The progressive disappearance of incrustations

associated with increasing homogenization and lossof stainability of the cytoplasm represents the stageof homogenizing cell change (fig 13). The shrunkentriangular nucleus, which frequently shows frag-mentation after survivals exceeding 24 hours, issurrounded by remnants of uniformly eosinophiliccytoplasm (mauve staining with Luxol fast blue)lacking in Nissl substance. However, the appearanceof this stage in semi-thin (1 um) plastic sections withtheir superior cytological detail differs from that seenin the thicker paraffin sections in that the remnantsof cytoplasm around the shrunken nucleus stain lessuniformly and present a heterogeneous rather thanhomogeneous morphology. This is confirmed inultrastructural observations confined at present tothe rat (Brown, 1973). There is a generalized decreasein electron density of both nucleus and cytoplasm.Thus the nuclear membranes are more readilyidentified around the shrunken nucleus and thevariably sized clumps of dense nucleoplasm anddenser (41-51 nm) granules are more clearly visual-

161copyright.


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

A. W. Brown

Fig 13

Fig 13 As figure 9. Homogenizing cell change inneocortex. The shrunken triangular nucleus issurrounded by homogeneous cytoplasm. Paraffin. Luxolfast blue and cresyl fast violet. x 1260.

Fig 14 As figure 9. Homogenizing cell change in aneocortical neurone showing process of invagination(arrows) ofperipheral areas ofhomogeneous cytoplasmby astrocytic processes sometimes leaving only anarrow rim of degenerating cytoplasm around thenucleus. x 5000.

Fig 15 As figure 14. Expansion of the space between the two nuclear membranes has produced a blebcontaining finely granular material. At this point the inner of the two membranes has disappeared and theouter one has ruptured. To the left of the bleb the nuclear membrane has disappeared with consequentfusion of nucleoplasm and degraded cytoplasm. x 16 250.

162copyright.


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/


ized (figs 14, 15). Frequently the nuclear envelopedisappears with consequent fusion of chromaticmaterial with the surrounding degraded cytoplasm(fig 15). Occasional large intracytoplasmic blebs areformed by expansion of the space between the twomembranes of the nuclear envelope (fig 15). Thesecontain a homogeneous finely granular materialwhich appears to be released into the cytoplasm byrupture of the outer nuclear membrane. Densemitochondria are still recognizable organelles withinotherwise totally degenerating cytoplasm where allorganized structure has disappeared leaving a debrisof many small membrane-bound vacuoles, coateddense vesicles and small dense bodies amid a hetero-geneous granular matrix probably representingdegraded free ribosomes (fig 15). Other areas ofcytoplasm at the periphery of the degeneratingperikaryon have a more homogeneous charactercomparable to the contents of the nuclear blebs(fig 14). The whole of the neuronal cytoplasm isinvaginated by swollen astrocytic processes whichoccasionally appear to contain degenerating neuronalcytoplasm (fig 14). There is evidence (Brown, 1973)that such a process of invagination and phagocytosisby astrocytes and also the action of phagocyticmicroglia may be responsible for the subsequentstage of the 'ghost cell' where the cytoplasm iseither absent or reduced to a minimally stainingremnant around a shrunken dark-staining nucleusand eventual cell loss.Homogenizing cell change and cell loss have been

observed in neocortical boundary zone lesions aftersurvivals of only four hours in the rat (Brown andBrierley, 1968). In regions of selective neuronaldestruction in rat and gerbil brains homogenizingcell change and ghost cells are visible with the lightmicroscope up to 21 days and in an infarct only upto seven to 10 days (Brown and Brierley, unpublishedresults). They have been seen to persist in the brainsof primates for 10 days or more (Brierley, 1972,1973).In general the stage of the naked nucleus is reachedearlier in small than in large neurones.No increase in extracellular space was observed

at any stage in the evolution of ischaemic cell change.

Reversibility of the process of ischaemic cell change

This is a question of the greatest clinical importance.Ideally its assessment would require the study oflarge numbers of animals subjected to standardizedinsults and with minimal differences in susceptibilitybetween animals. The appearance of larger num-bers of damaged neurones at the shorter rather thanthe longer survival periods would then indicatereversibility. Our observations in a large number ofexperimental animals (rats, gerbils and monkeys)

never showed any appreciable difference between thenumber of nerve cells showing microvacuolationafter survivals of 0 to one hour and the numbershowing later stages after survivals of more than onehour. Thus the assessment of reversibility must betentatively based on ultrastructural criteria alone. Itseems probable that the greatly swollen mitochondriashowing loss of cristae are irreversibly damaged. Thisalteration, together with increasing density of theneuronal cytoplasm, a change that possibly repre-sents protein denaturation, and the disorganizationof other organelles present in the later stages ofmicrovacuolation suggest that reversibility might bepossible within at most the first half hour after thehypoxic-ischaemic stress.

However, there is evidence for a second type ofneuronal cell change which may be reversible.Previously this type of alteration in rat brain wasreferred to as 'severe cell change' (Brown andBrierley, 1968). This term was considered inappro-priate in view of the minor alterations observed inthe fine structure and 'scalloped cell' was substituted(Brown and Brierley, 1973). The slightly dense,mildly distorted, non-vacuolated neurones seen infour Levine preparations killed immediately aftershort exposures to nitrogen (Brown and Brierley,1973) probably represent the initial stages of this typeof neuronal alteration that can attain its final formafter a survival of 45 minutes. In paraffin sections thecytoplasm and nucleus show a progressive increasein staining intensity which is accompanied by dis-tortion of the contours of the cell by expansion of thesurrounding perineuronal spaces (fig 16). Its maxi-mum development is seen as a marked 'scalloping' ofthe perikaryon, a slightly distorted nucleus and amoderate increase in staining intensity of both. Thereis no vacuolation of the cytoplasm or pyknosis of thenucleus. In semi-thin plastic sections the superiorcytological detail confirms the relative normality ofthe nucleus and cytoplasm (beyond a progressiveincrease in staining intensity) even in the most heavilyscalloped neurones. The maximally distorted neu-rones occur most frequently in arterial boundaryzone lesions in the neocortex. Electron micrographsreveal a progressive increase in density of the cyto-plasmic matrix and nucleoplasm accompanying theexpansion of astrocytic processes indenting the peri-karyon (fig 17). Where this expansion produces mark-ed indentations, an end stage seen after survivals of24 hours as well as of 45 minutes, the nucleus may beslightly distorted. Breakup of polysomes frequentlyoccurs to give an even dispersion of ribosomesthroughout the cytoplasm and this is sometimesaccompanied by slight swelling of the granular endo-plasmic reticulum cisternae while the remainingcytoplasmic organelles appear to be normal. The

163copyright.


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

164

Fig 16 As figure 12. Neocortical neurones exhibitingvaryving degrees of scalloped cell change. Darklystained cytoplasm is indented by swollen astrocyticprocesses. Paraffin. Cresylfast violet. x 1000.

Fig 17 Asfigure 12. Electron micrograph showingscalloped cell change in a neocortical neurone. Theperikaryon is distorted by expanded astrocytic processes.Note the normality of the cytoplasiniL organelles and thetrabecalae bridging the perineuronal 'spaces'. x 3600.

A. W. Brown

expanded astrocytic processes indenting the neuronessometimes contain flocculent material of mediumelectron density but few organelles. The trabeculae,which can be seen by light microscopy traversing theperineuronal spaces, consist of apposed astrocyticmembranes and the extended profiles of other cel-lular processes including axon terminals (fig 17).There is no increase in the extracellular spaceassociated with this cell change.As these scalloped cells do not show the dis-

organization of mitochondria and other organellesthat is typical of all stages of ischaemic cell change, itis tempting to speculate that this alteration might bereversible if the indenting astrocytic processesreturned to their normal dimensions. This type ofcell change has been described in epileptic braindamage in adolescent baboons following seizuresinduced by allylglycine (Meldrum et al, 1974) and inthe focal and geographical brain lesions ascribedeither to air embolism or hypotension after open-heart surgery (Brierley, 1963). It is also seen inoedematous human cortex (Long et al, 1966). Thusthe application of therapeutic measures which wouldreverse astrocytic swelling might bring about clinicalimprovement.

Neuronal artefact

Light (Jakob, 1927; Scharrer, 1938; Cammermeyer,1960, 1961, 1962; Coimbra, 1964; Brown andBrierley, 1968) and combined light and electronmicroscope studies (Cohen and Pappas, 1969;Brown, 1973; Brown and Brierley, unpublishedobservations) have defined the considerable differ-ences between the appearances of the artefacts ofdark neurones and hydropic cell change and thoseof all the stages in ischaemicand scalloped cell change.In paraffin and plastic sections the contour of thehyperchromatic or dark cell is frequently irregularand indented, the apical dendrite being occasionallycorkscrew-like (fig 1). Both cytoplasm and nucleusare shrunken and stain more intensively than nor-mal. Nissl bodies cannot be identified in the cyto-plasm and 'clear' perinuclear spaces are sometimesseen (fig 18). The nucleolus appears to be of normalsize but is occasionally difficult to identify in thedark-staining nucleus. Perineuronal spaces aroundthese neurones are rarely seen except in the hippo-campus (Brown, 1973). Electron micrographs of thedark neurones show a great increase in electrondensity of both the nucleoplasm and cytoplasmicmatrix (fig 19). The cytoplasmic organelles appearnormal with the exception of the Golgi complexwhere the membranes are frequently expanded pro-ducing electron-lucent arrays which contrast with thedense cytoplasm (fig 19). These electron-lucent areas



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


Fig 18 Fig 19Fig 18 Rat brain removed immediately after perfusion-fixation with glutaraldehyde and subjected to deliberatetrauma. Fascia dentata. Hyperchromatic or dark neurones with 'clear' perinuclear cytoplasm. Epon. Toluidineblue. x 1000.Fig 19 As figure 18. Electron micrograph showing the great increase in electron density of the nucleoplasmand cytoplasm in the shrunken dark neurones of the fascia dentata. Note the electron-lucent perinuclear arraysof expanded Golgi membranes and occasional swollen perineuronal astrocytic processes. x 3300.

correlate with the clear perinuclear spaces observedwith the light microscope. Cohen and Pappas (1969)reported the additional dilatation of the granularendoplasmic reticulum. It is important to stress thatno expansion of the mitochondria occurs and thatswollen astrocytic processes in contact with thesedense neurones are rarely observed. In contrast tothe dark cell the hydropic cell is pale, swollen andvacuolated (fig 2). The periphery of the somafrequently has a 'fuzzy' outline and may contain oneor more vacuoles. Electron microscope observationsof this artefact in the cortical neurones of inade-quately perfused rhesus monkeys (Brown, 1973) andin rat neurones (Brown and Brierley, unpublishedobservations) show clumping of the nucleoplasm,swelling and rarefaction of the granular endoplasmicreticulum, and ballooning of the mitochondria,especially in the peripheral zones of the swollenperikaryon (fig 20).

It has been previously emphasized (Brierley et al,1973) that the frequency of both artefacts in the brainof a human subject or experimental animal dying after

hypoxia is not proportional to the intensity of thehypoxic stress. Their distribution does not coincidewith the selectively vulnerable regions. Thus both,but particularly dark cells, are more numerous overthe crests of cerebral gyri and cerebellar folia whereasneurones undergoing ischaemic cell change are morenumerous in the floors of sulci. Further, neither ofthese artefacts has a time course.Cammermeyer (1972) has suggested that the early

microvacuolated appearance of the neurone is anartefact caused by the superimposition of spongyneuropil upon the neuronal cytoplasm. He attributesthe sponginess to poor perfusion-fixation resultingfrom 'no-reflow', the phenomenon described byAmes et al (1968) who observed that after a period ofcirculatory arrest exceeding five minutes blood flowis not restored in particular regions of the brain. Wehave frequently observed microvacuolated neuronesin areas of tissue showing no spongy state (Brownand Brierley, 1966, 1968, 1972, 1973). Levy et al(1975b) described microvacuolation in 16 rats sub-jected to 10 or more minutes of a controlled hypoxic-

165copyright.


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

A. W. Brown

Fig 20 Rhesus monkey. A neurone exhibiting hydropiccell change in inadequately perfused occipital cortex.Note rarefaction of the granular endoplasmicreticulum, swelling of mitochondria and membranes ofthe endoplasmic reticulum, more pronounced at peripheryof the perikaryon, and clumping of the nucleoplasm.x 3000.

ischaemic stress. There was only one instance oflocalized spongy neuropil and one of impaired reper-fusion. Careful examination of thin paraffin and inparticular extra thin (1 ,um) plastic sections (fig 4) inour own material show that microvacuoles areunequivocally situated in the neuronal cytoplasm.Further, Cammermeyer's hypothesis, based uponlight microscopical observations only, totally dis-regards the evidence of the electron microscope.Cammermeyer (1973) also regards the classic

ischaemic cell change of Spielmeyer as an artefactand suggests it 'is the homologue of the dark neuronin intact tissue'. As emphasized above, both light andEM appearances permit the differentiation betweenthe dark neurone and ischaemic cell change. Theirfrequency in relation to the severity of the hypoxicstress and their distribution are in no way com-parable. Ischaemic cell change is a process withtransitional stages from the earliest microvacuolatedneurone through to the naked nucleus and glial-mesodermal reaction and is not a 'steady state' ascharacterized by the dark neurone.

Other studies

The histological and histochemical techniquesemployed by others on invariably immersion-fixedmaterial from Levine preparations failed to identifythe earliest neuronal alterations. Thus Becker andBarron (1961) described the earliest light micro-scopic evidence of ischaemic neuronal alterations as'focal neuronal damage' in the caudate nucleus afterthree hours, in the hippocampus and neocortex afterfour to six hours and Becker (1961) demonstratedmitochondrial swelling after 12 hours. They con-sidered that histochemical methods were necessary todetect the earliest ischaemic neuronal damage.Zeman (1963) reported 'loosening of tissue texturewith minimal shrinkage of nerve cell bodies' in theneocortex after one hour and Spector (1963) des-cribed 'alterations of staining reaction and shrinkageof nerve cell bodies' in the hippocampus after 10-18hours. This latter author concluded that a 'bio-chemical lesion', as defined by histochemical tech-niques, must occur before injury to the neurone canbe detected by light microscopy (Spector, 1963;MacDonald and Spector, 1963). Clendenon et al(1971) described histological alterations in the ipsi-lateral hemisphere after one hour which 'consisted ofpallor of the striatum and lateral cortex of Nisslstaining with evidence of swelling and loss of stain-ability of subcortical white matter'. In all thesereports typical ischaemic cell change was neither de-cribed nor illustrated.Some of the earlier studies of the fine structure of

hypoxic-ischaemic brain lesions have been limited bythe use of immersion or inadequate perfusion-fixation. Hills (1964), employing immersion-fixationin osmic acid in rat Levine preparations, describedswelling of astrocytic processes in the cerebral cortexafter 90 minutes and degeneration of neuronal andoligodendroglial cytoplasm after three hours. Thisauthor reported difficulty in interpreting some of theinitial changes because of the possibility of fixationartefact.

Scholz and Hager (1959), Hager et al (1960) andHager (1963) used perfusion-fixation with osmic acidin their studies of the neocortex of the Syrian hamsterkilled immediately after repeated asphyxiation withnitrogen for 12 to 15 minutes. The nature of theirfindings and their frequency (discussed by Brown andBrierley 1972, 1973), together with their reports ofincomplete perfusion-fixation, suggest that poorpreservation accounts for the greater part of theirobservations. Their fig 2 (Hager et al, 1960) appearsto be the ultrastructural equivalent of the artefact ofhydropic cell change and is similar to the electronmicrograph of the same artefact illustrated by

166



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


Brierley et al (1973) (their fig 2) and fig 20 in the pres-ent communication.

In a combined light and EM study, Olsson andHossmann (1971) subjected cats to transient cerebralischaemia with or without rinsing of the intra-cerebral vessels during the ischaemia and reportedchanges in the neurones when ischaemia (withoutrinsing) lasted more than eight minutes and thesurvival period was longer than 30 minutes. Althoughtypical microvacuolation was illustrated it wasneither described nor recognized as such. It is alsounfortunate that the artefact of the dark neuroneseems to have been accepted as evidence of ischaemiccell change. The overall interpretation of their find-ings is difficult as the evidence was based uponcortical biopsies fixed by immersion and brainsremoved immediately after perfusion-fixation.More recently Shay and Gonatas (1973) produced

circulatory arrest for 45 minutes in the lumbar spinalcord of the cat by occlusion of the descending aortaand concurrent arterial hypotension. The spinal cordwas fixed by perfusion after survivals of 75 minutesand its removal was delayed. In the anterior hornneurones they observed swelling of the mito-chondria and Golgi apparatus, dispersal of ribo-somes and an increase in electron density of thecytoplasm. The lysosomes were unchanged inappearance and size.Means et al (1976), using perfusion-fixation and

delayed removal, described microvacuolation inanterior horn cells of the cat as early as seven min-utes after compression of the spinal cord. The micro-vacuoles corresponded to swollen mitochondria. Theyalso reported ischaemic nerve cell change and incrus-tations in 15-minute, 30-minute, one-hour, four-hourand eight-hour survivals.The neuronal alterations in developing cortical

infarction in squirrel monkeys, and, in particular therole of lysosomes in the production of ischaemic cellchange, were studied by Little et al (1974a and b).Cortical ischaemia, ranging in duration from 45minutes to 24 hours, was produced by occluding theright middle cerebral artery with a clip which wasremoved at the commencement of perfusion. Theydetected no changes with the light microscope at 45minutes but observed slight swelling of the roughendoplasmic reticulum at the EM level. After 90minutes swelling of mitochondria was also seen. Theydescribed swelling and fragmentation of largelysosomes at 12 hours and concluded that 'lysosomalenzyme release is not related to the early cellularchanges' but probably contributes to the final de-struction of the ischaemic neurone. Theirfindings thatlysosomes appear to be less susceptible to ischaemiathan other neuronal organelles correspond with ourown observations and those of Clendenon et al (1971)

in the rat.Microvacuolation and its later stages, depending

on the survival times employed, have also beenobserved with the light microscope in rat Levinepreparations with physiological control (Salfordet al, 1973; Levy et al, 1975b), in status epilepticus inbaboons induced by bicuculline (Meldrum andHorton, 1973; Meldrum and Brierley, 1973;Meldrum et al, 1973) and in gerbils subjected to onehour of unilateral carotid artery occlusion (Levy et al,1975a). The later stages only (shortest survival 18hours) were seen after subatmospheric decompres-sion (hypoxic hypoxia) in baboons and rhesusmonkeys (Brierley and Nicholson, 1969; Nicholsonetal, 1970).

General conclusions

It has been possible by the employment of perfusion-fixation and standard neuropathological techniquesin experimental animals to define the nature and timecourse of ischaemic cell change and to endorse thelong established conclusion from human neuro-pathological studies that this cell change is theneuronal alteration common to all types of hypoxia.

Microvacuolation of the neuronal cytoplasm is theearliest cytological response to a hypoxic stress inrodent, monkey and man and the majority of micro-vacuoles correspond to swollen mitochondria.Swelling of the mitochondria, together with a slightincrease in density of the cytoplasmic matrix andnucleoplasm (fig 6), were the only ultrastructuralalterations in the damaged neurones of rats killedimmediately after intermittent exposure to nitrogenof five minutes (Brown and Brierley, 1973). Sincesimilar changes were accompanied by other ultra-structural alterations such as dilatation of the endo-plasmic reticulum and progressive condensation ofthe cytoplasm and nucleoplasm (described above)when longer exposures to nitrogen and longersurvivals were employed (McGee-Russell et al, 1970;Brown and Brierley, 1972, 1973), it must be con-cluded that the mitochondrion is probably the mostsensitive organelle to a hypoxic stress. This is to beexpected as the enzyme complexes responsible forrespiration and oxidative phosphorylation arelocalized within these organelles.The ultrastructural finding that microvacuolation

of the neuronal cytoplasm precedes any alteration inthe astrocytic cytoplasm (Brown and Brierley, 1973)is in keeping with the light microscope reports ofdamaged neurones but undamaged astrocytes andblood vessels in physiologically controlled rat Levinepreparations (Salford et al, 1973; Levy et al, 1975b)and underlines the vulnerability of the neurone toeven brief hypoxic insults. Both these latter groups of

167



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

168 A. W. Brown

investigators, using a physiologically controlledLevine preparation which excluded the systemicvariables of apnoea, hypotension, bradycardia andacidosis, convincingly demonstrated that impairedvascular reperfusion was not a factor in the genesisof the neuronal alterations.The report by Salford et al (1973) that existing bio-

chemical methods lacked the sensitivity to detect thepresence of small numbers of neurones undergoingischaemic cell change after brief periods of hypoxia,while neuropathological techniques demonstratedsuch changes in the brains of similarly treatedanimals, endorses the view that these latter techni-ques, when applied to optimally perfused tissue,remain the most sensitive and reliable indicators ofearly hypoxic brain damage.

The author wishes to thank Miss J. Harrow, Mrs M.Kublik and Mr P. Georgiades for skilled technicalassistance.

References

Ames, A., III, Wright, R. L., Kowada, M., Thurston, J. M.,and Majno, G. (1968). Cerebral ischemia. II. The no-reflowphenomenon. American Journal of Pathology, 52, 437-453.

Becker, N. H. (1961). The cytochemistry ofanoxic and anoxic-ischemic encephalopathy in rats. II. Alterations in neuronalmitochondria identified by diphosphopyridine and tri-phosphopyridine nucleotide diaphorases. American Journalof Pathology, 38, 587-597.

Becker, N. H., and Barron, K. D. (1961). The cytochemistryof anoxic and anoxic-ischemic encephalopathy in rats. I.Alterations in neuronal lysosomes identified by acidphosphatase activity. American Journal of Pathology, 38,161-175.

Brierley, J. B. (1963). Neuropathological findings in patientsdying after open-heart surgery. Thorax, 18, 291-304.

Brierley, J. B. (1972). The neuropathology of brain hypoxia.In Scientific Foundations of Neurology, edited by M.Critchley, J. L. O'Leary, and B. Jennett, pp. 243-252.Heinemann, London.

Brierley, J. B. (1973). Pathology of cerebral ischemia. InCerebral Vascular Diseases (Eighth Conference), pp. 59-75.Grune and Stratton, New York.

Brierley, J. B. (1976). Cerebral hypoxia. In Greenfield'sNeuropathology, 3rd edn, edited by W. Blackwood andJ. A. N. Corsellis, pp. 43-85. Arnold, London.

Brierley, J. B., Brown, A. W., Excell, B. J., and Meldrum,B. S. (1969). Brain damage in the rhesus monkey resultingfrom profound arterial hypotension. I. Its nature, distribu-tion and general physiological correlates. Brain Research,13, 68-100.

Brierley, J. B., Brown, A. W., and Levy, D. E. (1976).Relationship between ischemic neuronal damage and edemain primate and rodent brain. In Workshop on CerebralEdema, May 1976. McGill University, Montreal. (In press).

Brierley, J. B., Brown, A. W., and Meldrum, B. S. (1971a).The neuropathology of insulin-induced hypoglycaemia in aprimate (M. mulatta). Topography and cellular nature. InBrain Hypoxia (Clinics in Developmental Medicine 39/40),edited by J. B. Brierley and B. S. Meldrum, pp. 225-229.Heinemann, London.

Brierley, J. B., Brown, A. W., and Meldrum, B. S. (1971b).The nature and time course of the neuronal alterations

resulting from oligaemia and hypoglycaemia in the brain ofMacaca mulatta. Brain Research, 25, 483-499.

Brierley, J. B., and Excell, B. J. (1966). The effects ofprofoundsystemic hypotension upon the brain of M. rhesus. Physio-logical and pathological observations. Brain, 89, 269-298.

Brierley, J. B., Meldrum, B. S., and Brown, A. W. (1973). Thethreshold and neuropathology of cerebral 'anoxic-ischemic'cell change. Archives of Neurology, 29, 367-374.

Brierley, J. B., and Nicholson, A. N. (1969). Neuropathologi-cal correlates of neurological impairment following pro-longed decompression. Aerospace Medicine, 40, 148-152.

Brown, A. W. (1973). Morphological characteristics andevolution of anoxic-ischaemic brain damage in differentspecies. MPhil Thesis, Council for National AcademicAwards.

Brown, A. W., and Brierley, J. B. (1966). Evidence for earlyanoxic-ischaemic cell damage in the rat brain. Experientia(Basel), 22, 546-547.

Brown, A. W., and Brierley, J. B. (1968). The nature,distribution and earliest stages of anoxic-ischaemic nervecell damage in the rat brain as defined by the opticalmicroscope. British Journal of Experimental Pathology, 49,87-106.

Brown, A. W., and Brierley, J. B. (1972). Anoxic-ischaemiccell change in rat brain. Light microscopic and fine-structural observations. Journal of the NeurologicalSciences, 16, 59-84.

Brown, A. W., and Brierley, J. B. (1973). The earliestalterations in rat neurones and astrocytes after anoxia-ischaemia. Acta Neuropathologica, 23, 9-22.

Cammermeyer, J. (1960). The post-mortem origin andmechanism of neuronal hyperchromatosis and nuclearpyknosis. Experimental Neurology, 2, 379-405.

Cammermeyer, J. (1961). The importance of avoiding 'dark'neurons in experimental neuropathology. Acta Neuro-pathologica, 1, 245-270.

Cammermeyer, J. (1962). An evaluation of the significance ofthe 'dark' neuron. Ergebnisse der Anatomie und Entwick-lungsgeschichte, 36, 1-61.

Cammermeyer, J. (1972). Nonspecific changes of the centralnervous system in normal and experimental material. InThe Structure and Function of Nervous Tissue, edited byG. H. Bourne, vol. 6, pp. 131-251. Academic Press, NewYork.

Cammermeyer, J. (1973). 'Ischemic neuronal disease' ofSpielmeyer: a reevaluation. Archives of Neurology, 29, 391-393.

Clendenon, N. R., Allen, N., Komatsu, T., Liss, L., Gordon,W. A., and Heimberger, K. (1971). Biochemical alterationsin the anoxic-ischemic lesion of rat brain. Archives ofNeurology, 25,432-448.

Cohen, E. B., and Pappas, G. D. (1969). Dark profiles in theapparently-normal central nervous system: a problem inthe electron microscopic identification of early anterogradeaxonal degeneration. Journal of Comparative Neurology,136, 375-395.

Coimbra, A. (1964). Nerve cell changes in the experimentalocclusion of the middle cerebral artery. Histological andhistochemical study. Acta Neuropathologica, 3, 547-557.

Courville, C. B. (1953). Contributions to the Study of Anoxia.San Lucas Press, Los Angeles.

David, G. B. (1955). The effect of eliminating shrinkageartifacts on degenerative changes seen in CNS material.Excerpta Medica, Section VIII, Neurology and Psychiatry,8,777-778.

Gildea, E. F., and Cobb, S. (1930). The effects of anemia onthe cerebral cortex of the cat. Archives of Neurology andPsychiatry, 23, 876-903.

Greenfield, J. G., and Meyer, A. (1963). General pathologyof the nerve cell and neuroglia. In Greenfield's Neuropatho-



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Structural abnormalities in neurones 169

logy, 2nd edn, edited by W. Blackwood, et. al, pp. 1-70.Arnold, London.

Grenell, R. G., and Kabat, H. (1947). Central nervous systemresistance. n1. Lack of correlation between vascularity andresistance to circulatory arrest in hypothalamic nuclei.Journal of Neuropathology and Experimnental Neurology, 6,35-43.

Hager, H. (1963). Electron microscopical observations on theearly changes in neurons caused by hypoxidosis and on theultrastructural aspects of neuronal necrosis in the cerebralcortex of mammals. In Selective Vulnerability of the Brainin Hypoxaemia, edited by J. P. Schad6 and W. H.McMenemey, pp. 125-136. Blackwell, Oxford.

Hager, H., Hirschberger, W., and Scholz, W. (1960). Electronmicroscopic changes in brain tissue of Syrian hamstersfollowing acute hypoxia. Aerospace Medicine, 31, 379-387.

Hicks, S. P. (1968). Vascular pathophysiology and acute andchronic oxygen deprivation. In Pathology of the NervousSystem, edited by J. Minckler, vol. 1, pp. 341-350. McGraw-Hill, New York.

Hills, C. P. (1964). The ultrastructure of anoxic-ischaemiclesions in the cerebral cortex of the adult rat brain. Guy'sHospital Reports, 113, 333-348.

Hoff, E. C., Grenell, R. G., and Fulton, J. F. (1945). Histo-pathology of the central nervous system after exposures tohigh altitudes, hypoglycemia and other conditions associa-ted with central anoxia. Medicine (Baltimore), 24, 161-217.

Jakob, A. (1927). Normale and pathologische Anatomie undHistologie des Grosshirns. Band I, Normnale Anatomie undHistologie und allgemeine Histopathologie des Grosshirns.Deuticke, Leipzig.

Koenig, R. S., and Koenig, H. (1952). An experimental studyof post mortem alterations in neurons of the central nervoussystem. Journal of Neuropathology, and ExperimentalNeurology, 11, 69-78.

Krogh, E. (1952). (Discussion). Proceedings of the Ist Inter-national Congress of Neuropathology, Rome Vol. 1, pp. 453-457.

Levine, S. (1960). Anoxic-ischemic encephalopathy in rats.American Journal of Pathology, 36, 1-17.

Levy, D. E., Brierley, J. B., and Plum, F. (1975a). Ischaemicbrain damage in the gerbil in the absence of 'no-reflow'.Journal of Neurology, Neurosurgery and Psychiatry, 38,1197-1205.

Levy, D. E., Brierley, J. B., Silverman, D. G., and Plum, F.(1975b). Brief hypoxia-ischemia initially damages cerebralneurons. Archives of Neurology, 32, 450-456.

Little, J. R., Kerr, F. W. L., and Sundt, T. M.,Jr. (1974a). Therole of lysosomes in production of ischemic nerve cellchanges. Archives of Neurology, 30, 448-455.

Little, J. R., Sundt, T. M., Jr., and Kerr, F. W. L. (1974b).Neuronal alterations in developing cortical infarction.Journal of Neurosurgery, 40, 186-198.

Long, D. M., Hartmann, J. F., and French, L. A. (1966). Theultrastructure of human cerebral edema. Journal ofNeuropathology and Experimental Neurology, 25, 373-395.

MacDonald, M., and Spector, R. G. (1963). The influence ofanoxia on respiratory enzymes in rat brain. British Journalof Experimental Pathology, 44, 11-15.

McGee-Russell, S. M., Brown, A. W., and Brierley, J. B.(1970). A combined light and electron microscope study ofearly anoxic-ischaemic cell change in rat brain. BrainResearch, 20, 193-200.

Means, E. D., Anderson, D. K., and Gutierrez, C. (1976).Light and electron microscopy of gray matter followingexperimental spinal cord compression injury. (Abstract).Journal of Neuropathology and Experimental Neurology,

35, 348.Meldrum, B. S., and Brierley, J. B. (1973). Prolonged

epileptic seizures in primates: ischemic cell change and itsrelation to ictal physiological events. Archives ofNe;arology,28, 10-17.

Meldrum, B. S., and Horton, R. W. (1973). Physiology ofstatus epilepticus in primates. Archives of Neurology, 28,1-9.

Meldrum, B. S., Horton, R. W., and Brierley, J. B. (1974).Epileptic brain damage in adolescent baboons followingseizures induced by allylglycine. Brain, 97, 407-418.

Meldrum, B. S., Vigouroux, R. A., and Brierley, J B. (1973).Systemic factors and epileptic braindamage: prolongedseizures in paralyzed, artificially ventilated baboons.Archives of Neurology, 29, 82-87.

Meyer, A. (1963). Intoxications. In Greenfield's Neuro-pathology, 2nd edn, edited by W. Blackwood, et al, pp.235-287. Arnold, London.

Morrison, L. R. (1946). Histopathologic effect of anoxia onthe central nervous system. Archives of Neurology andPsychiatry, 55,1-34.

Nicholson, A. N., Freeland, S. A., and Brierley, J. B. (1970).A behavioural and neuropathological study of the sequelaeof profound hypoxia. Brain Research, 22, 327-345.

Olsson, Y., and Hossmann, K. A. (1971). The effect of intra-vascular saline perfusion on the sequelae of transientcerebral ischemia. Light and electron microscopicalobservations. Acta Neuropathologica, 17, 68-79.

Polani, P. E., and MacKeith, R. (1954). The sequelae ofanoxia. Guy's Hospital Reports, 103, 54-58.

Richardson, J. C., Chambers, R. A., and Heywood, P. M.(1959). Encephalopathies of anoxia and hypoglycemia.Archives of Neurology, 1, 178-190.

Salford, L. G., Plum, F., and Brierley, J. B. (1973). Gradedhypoxia-oligemia in rat brain: II. Neuropathologicalalterations and their implications. Archives of Neurology,29, 234-238.

Scharrer, E. (1938). On dark and light cells in the brain and inthe liver. Anatomical Record, 72, 53-65.

Scholz, W., and Hager, H. (1959). Toxicity changes in thecentral nervous system. Oxygen deficiency and its influenceon the central nervous system. U.S.A.F. Tech. (Final) Rep.,Contract No. AF61(514)-945, Part II, pp. 1-24.

Shay, J., and Gonatas, N. K. (1973). Electron microscopy ofcat spinal cord subject to circulatory arrest and deep localhypothermia (15C). American Joiurnal of Pathology, 72,369-396.

Spector, R. G. (1961). Water content of the brain in anoxic-ischaemic encephalopathy in adult rats. British Journal ofExperimental Pathology, 42, 623-630.

Spector, R. G. (1963). Selective changes in dehydrogenaseenzymes and pyridine nucleotides in rat brain in anoxic-ischaemic encephalopathy. British Journal of ExperimentalPathology, 44, 312-316.

Spielmeyer, W. (1922). Histopathologie des Nervensystems,pp. 74-79. Springer, Berlin.

Tureen, L. L. (1936). Effect of experimental temporaryvascular occlusion on the spinal cord. I. Correlationbetween structural and functional changes. Archives ofNeurology and Psychiatry, 35, 789-807.

Wolf, A., and Cowaen, D. (1949). Pathology. In SelectivePartial Ablation of the Frontal Cortex, edited by F. A.Mettler, p. 469. Hoeber, New York.

Zeman, W. (1963). Histochemical and metabolic changes inbrain tissue after hypoxaemia. In Selective Vulnerability ofthe Brain in Hypoxaemia, edited by J. P. Schade andW. H. McMenemey, pp. 327-348. Blackwell, Oxford.



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Documents

Structural abnormalities in neurones - BMJ