Brain Research, 176 (1979) 91-100 91

© Elsevier/North-Holland Biomedical Press

ACUTE DENDROTOXIC CHANGES IN THE HIPPOCAMPUS OF

K A I N A T E T R E A T E D RATS

J. W. OLNEY, T. FULLER and T. DE GUBAREFF

Department of Psychiatry, Washington University School of Medicine, St. Louis, Mo. 63110 (U.S.A.)
(Accepted February 1st, 1979)

SUMMARY

Kainic acid (KA), a potent neuroexcitatory and neurotoxic analog of glutamate

(Glu), induces a widespread pattern of brain damage when administered subcutane-
ously to adult rats. The hippocampus is among the brain regions most consistently and
severely damaged. Here we describe acute swelling of certain spines and branchlets of
dendrites as the first detectable sign of KA neurotoxic changes in the hippocampus.
These swellings conform to a laminar pattern suggesting selective toxic interaction of
KA at specific levels of the dendritic trees of hippocampal pyramidal and dentate
granule neurons. The frequency and severity of involvement of each type of hippo-
campal neuron at each level of its dendritic tree was roughly estimated and neuronal
types were ranked from the most to least extensively involved (CA3 > CA4 > CA1 >
CA2 > dentate granules). The same rank order has been described for the vulnerabi-
lity of these neurons to acute destruction following intraventricular KA administra-
tion. Because the pattern of dendritic dilatations observed corresponds well with the
pattern of termination of putative glutamergic inputs to the hippocampus, we inter-
pret the findings as being consistent with the hypothesis that the toxic effects of KA
are mediated through glutamergic excitatory receptors. We propose that the sensitivity
of a given neuron to the neurodestructive action of KA may be determined by the
percentage of its dendritic surface occupied by Glu receptors. We suspect that most, if
not all, hippocampal neurons receive some glutamergic input and, therefore, are sensi-
tive to KA. That CA3 pyramids are substantially more sensitive than dentate granules
may signify that the former receive many more Glu terminals than the latter, an
assumption quite consistent with our observation that focal dendritic swellings were
both more densely and more widely distributed over the dendritic surfaces of the
former than the latter.
92

INTRODUCTION

Kainic acid (KA) is a rigid structural analog of glutamate (Glu) which far
exceeds the potency of Glu as either a neuroexcitanta, ~ or neurotoxin 11. It has been
postulated that an excitatory mechanism underlies the toxic action of KA and that the
action may be mediated by synaptic receptors specialized for glutamergic trans-
mission. If this is correct, neurons that are heavily innervated by glutamergic
terminals, and hence are abundantly endowed with excitatory Glu receptors, should be
particularly vulnerable to KA-induced degeneration and KA may be a useful tool for
mapping the location of such neurons.
Hypersensitivity of hippocampal neurons to KA was noted in 1974 by Olney et
al. 11, and has subsequently been explored by several authors3,6, 7,t4,19 Nadler et al. 6,7,
who examined the hippocampus of adult rats following slow intraventricular infusion
of small do~es of KA (0.1-3/tg), described an heirarchy of neuronal sensitivities to the
neurodestructive action of KA with CA3 and CA4 hippocampal pyramidal neurons
being the most vulnerable and dentate granule cells the least. Wuerthele et al. t9
recently described degeneration of CA3 and CA4 hippocampal pyramidal neurons as
an incidental finding in rats injected intrastriatally with KA. Hypersensitivity of CA3
and CA4 pyramids to KA is consistent with the hypothesis that Glu synaptic receptors
mediate KA toxicity as these neurons are heavily innervated by mossy fibers which are
thought to use Glu as transmitterL On the other hand, dentate granule cells also
receive fibers that are putatively glutamergic la as do many, if not all, hippocampal
neurons2,16,18. It was not determined in the above studies whether specific dendritic
fields upon which glutamergic fibers are thought to project are involved in the acute
toxic reaction to KA. Here we will address this question by describing the earliest
detectable ultrastructural changes in hippocampal neurons of KA-treated rats. In view
of the recent observation T M that the hippocampus is one among several brain regions
that is frequently damaged rather severely following subcutaneous or intraperitoneal
administration of sublethal doses (10-12 mg/kg) of KA to adult rats, we administered
KA systemically, thereby avoiding mechanical damage potentially associated with
intracranial delivery methods.

MATERIALS AND METHODS

Twenty-four adult Sprague-Dawley male rats (Harlan Industries) weighing

275-325 g were used. Kainic acid (Sigma Chemical Co., St. Louis), dissolved in
distilled water with pH adjusted to 7.2 ± 0.2 with NaOH, was administered intraperi-
toneally in a dose of 12 mg/kg body wt. The animals were sacrificed by perfusion
fixation (cacodylate-buffered 1 ~ paraformaldehyde/1.5 ~ glutaraldehyde solution)
at various intervals from 1 to 24 h after treatment and brains were further processed by
methods described elsewhere s for light and electron microscopy. Untreated or saline
treated adult rats whose brains were prepared similarly for histological examination
served as controls,
 93

Fig. 1. a: the hippocampus from a saline treated control rat. b: the rat hippocampus 4 h following a 12
mg/kg i.p. dose of KA. Note the laminar pattern of edematous changes. The pathological reaction is
most conspicuous in regions containing either the cell bodies or dendritic processes of CA1, CA3 and
CA4 pyramidal neurons (x 50).

RESULTS

The 1-4 h post-treatment interval was found to be most favorable for analysis of
early KA-induced changes. A t this time, a conspicuous pattern o f swelling affecting
specific cellular elements was clearly discernable in the hippocampus (Figs. 1 and 2).
Quite prominent changes were observed in specific glia surrounding the cell bodies o f
pyramidal neurons (Fig. 2b and d). These cells underwent striking edematous swelling
94

similar to that described by Olney et al. 10 in the hypothalamus of the infant mouse
treated with DL-alpha aminoadipate, a non-excitatory structural analog of Glu that
does not damage neurons. Glial swelling has also been described as an accompaniment
of neuronal degeneration in the mouse hypothalamus following Glu treatment 8. While
the significance of the glial reaction to KA warrants further investigation, it will not be
 95

TABLE I
Summary of fields affected by KA treatment
A system of 1-3 plus was used to derive a combined rough estimate of the severity and frequency of
involvement of each dendritic field

Pyramids Dentate
granules
CA3a CA3b CA3c CA4 C.41 CA2

Apical
Proximal + ++ + + -4- + ÷+ - + +
Distal ++ + + + +
Basilar
Proximal
Distal + + + + + + + + + + + ++

given further a t t e n t i o n here as the p r i m a r y purpose of this report is to describe the

early n e u r o n a l changes associated with K A treatment.
The earliest changes in n e u r o n s were limited to dendritic elements a n d the
characteristic sign of toxicity was focal swelling confined to select dendritic branchlets
a n d spines distributed in a l a m i n a r p a t t e r n across several specific dendritic fields (Fig.
2). The n u m b e r s of dilated dendritic profiles in a given field a n d the overall reaction
p a t t e r n varied a m o n g animals b u t the latter was relatively consistent. Table I
summarizes the fields affected a n d gives a r o u g h estimate of the severity a n d frequency
o f i n v o l v e m e n t of each field. The distal basilar dendritic segments of C A 1 (Figs. 2b a n d
3) a n d sometimes CA2, 3 a n d 4 p y r a m i d a l n e u r o n s were affected, as were the distal apical
dendritic segments of CA1 (Fig, 2a), CA3a a n d CA3b pyramidal n e u r o n s (Fig. 4) a n d
dentate granule n e u r o n s (Fig. 2c). I n addition, we characteristically observed a wide
b a n d of rarified tissue exactly following the path of synaptic contacts made by mossy

Fig. 2. All scenes are from adult rat hippocampus 4 h following KA, 12 mg/kg i.p. except e which is
from a saline control, a: the band of vacuolated structures forming a layer between the arrows in the
upper field are dilated branchlets of distal apical dendrites of CAI neurons. The hippocampal fissure
(lower arrow) sharply divides the affected dendrites of CA 1 neurons from dendrites of dentate granule
cells which, in this scene, are unaffected. Swollen glia and dendrites of CA4 neurons at the bottom
right (immediately below the layer of dentate granule cells) give this zone a rarefied appearance, b: the
dark zone of myelinated axons at top is the alveus. Below it is a layer of clear vacuolated structures
which are dilated distal branchlets of basilar CA1 dendrites. The layer of CA1 pyramidal neurons
(lower field) are surrounded by an irregular meshwork of edematous glia, c: at the bottom of the field
are cell bodies of dentate granule neurons whose dendritic processes reach to the hippocampus fissure
(arrows). The layer of vacuoles immediately beneath the hippocampal fissure are dilated distal tips of
dentate granule cell dendrites. These distal segments are the only portion of the dendritic field of
dentate granule neurons which receives afferent input from the putatively glutamergic perforant path.
In d and e are shown CA3 pyramidal neurons from KA and saline treated animals respectively. The
sharply demarcated band of edematous tissue in d encompasses the row of pyramidal neurons and
their proximal dendritic shafts. Mid-dendritic levels of both apical and basilar fields (top and bottom)
are unaffected. The edematous structures immediately adjacent to the CA3 neuronal cell bodies are
swollen glial processes whereas those near the edges of the edematous zone are swollen dendritic
spines (see Fig. 5) upon which mossy fibers synapse (× 240).
96

Fig. 3. a: electron micrographic survey view of CA1 basilar dendritic field, distal portion, to show
dilated dendritic profiles (arrow heads) 4 h following KA treatment (12 mg/kg i.p.). Note normal
appearance of other neuropil components and of myelinated axons coursing through alveus at top. b :
a massively dilated dendritic branchlet (D) from the CA1 field in Fig. la. The synapse (arrowhead),
which is shown at higher magnification at lower right, identifies the swollen structure as a dendrite.
Note that the presynaptic axon terminal (A) appears normal despite marked edematous changes in tke
post-synaptic dendrite (D). a -- x 1200; b - x 7200; inset = x 45,000.

Fig. 4. a: electron micrographic survey view of CA3 a apical dendritic field, distal portion, following
KA treatment. Except for the acutely edematous dendritic profiles, the neuropil is normal, b: dilated
dendrite (D) from encircled region in Fig. 3a. Note that the swelling of this dendrite is confined to
portions distal to the point (arrowhead) where an axon terminal impinges synaptically upon it. a ~ ×
700; b = × 10,000.

fibers ( a x o n s o f d e n t a t e g r a n u l e cells) u p o n p r o x i m a l d e n d r i t i c s e g m e n t s o f C A 3 a n d
C A 4 p y r a m i d s (Figs. 1, 2d a n d 5).
By e l e c t r o n m i c r o s c o p i c e x a m i n a t i o n it was q u i t e c l e a r t h a t s t r u c t u r e s r e s p o n s -
ible f o r t h e rarified a p p e a r a n c e o f t h e m o s s y fiber p a t h w a y w e r e n o t t h e m o s s y fiber
t e r m i n a l s t h e m s e l v e s , b u t r a t h e r t h e d e n d r i t i c spines t h a t i n v a g i n a t e t h e s e t e r m i n a l s
 97

Fig. 5. A normal appearing mossy fiber (M) invaginated by two massively swollen dendritic spines (S)
in the CA3 ~ apical dendritic field, proximal portion from a KA treated rat. Note that the mossy fiber
is in synaptic contact (arrowhead) with one of the spines and that the postsynaptic density may be
manifesting pathological changes (x 36,000).

(Fig. 5). These spines - - and occasionally portions of the dendritic segments giving rise
to them, but not their cell bodies - - were markedly dilated despite absence of patho-
logical alterations in the mossy axon terminals embracing them. The edematous
neuronal elements, either in the mossy fiber zone or elsewhere, could be identified with
confidence as dendritic because of the axon terminals frequently found to be in pre-
synaptic contact with them (Figs. 3-5). These reactive dendritic elements had the
same appearance as has been described repeatedly in lesions induced by Glu or its
various excitatory analogsS,10A 1. Characteristically they were massively swollen and
appeared empty except for a variable amount of particulate debris and an occasional
degenerated mitochondrion. The presynaptic axon terminals in contact with such
structures typically exhibited little or no variation from normal; an increase in staining
density of the postsynaptic web and possible disarray of web components was
sometimes apparent (Fig. 4).
Changes affecting intracellular organelle systems in cell bodies of hippocampal
neurons were seen in the present study but were considered of secondary significance
as they occurred later than the above described dendritic swelling. Such changes
should probably be considered non-specific as they are observed in neurons under-
going acute necrosis from several kinds of insult, for example, anoxia or mechanical
trauma.
98

DISCUSSION

The mechanism of KA neurotoxicity remains an unsettled issue. Several years

ago Olney et a1.11,12 postulated that KA, and other excitatory Glu analogs which
mimic the neurotoxic activity of Glu, do so by an excitatory mechanism. They
suggested that each analog interacts directly with the excitatory receptors through
which iontophoretically applied Glu or aspartate depolarize neurons. McGeer et al. 5,
noting that surgical ablation of the putatively glutamergic corticostriatal tract renders
KA less effective as a striatal neurotoxin, reasoned that KA may act primarily by
enhancing the toxic potential of Glu at its own synaptic receptors. Campochiaro and
Coyle 1 have argued similarly, based on evidence that striatal neurons are insensitive to
KA toxicity during early periods of development when the corticostriatal tract has not
yet established striatal synapses. Nadler et al. 7 have suggested that KA may induce
repetitive hippocampal seizure activity analogous to an episode of status epilepticus
and that a particularly low seizure threshold of CA3 pyramidal neurons may underly
their extreme vulnerability to degeneration from KA exposure. The pathological
process leading to cell death, according to their view, would be repetitive depolariza-
tion rather than sustained depolarization as proposed by Olney et al.ll,~L Nadler et
al. 7, citing several lines of evidence, particularly the low sensitivity of dentate granule
cells to KA toxicity, tend to reject the hypothesis that Glu synaptic receptors mediate
KA hippocampal toxicity.
In the present study we were impressed by the precise localization of acute toxic
effects of KA to select dendritic branchlets and spinous processes, since these are the
specific portions of the neuron that receive the bulk of the synaptic terminals through
which excitatory signals are transmitted to hippocampal neurons. It is further to be
noted that the neuronal populations which Nadler et al. found most sensitive to the
neurodestructive effects of KA, namely CA3a and CA3b pyramids, were the only
neurons which we found to have acute dendritic dilations at three separate levels on
their dendritic trees (both distal and proximal portions of the apical dendritic field and
distal portions of the basilar field). Fiber tracts innervating these dendritic fields are
entorhinal perforant fibers, mossy fibers and commissural fibers respectively ~3, all
three of which are putative glutamergic pathways2A6,18. Three additional neuronal
populations, CA3e, CA4 and CA1 pyramids, exhibited acute pathological changes at
two levels on their dendritic trees; these neuronal groups were judged by Nadler et al.
to be highly sensitive, although not quite as sensitive as CA3 a and CA3b neurons to
KA toxicity. Again, the pathways innervating the specific dendritic fields involved are
believed to be glutamergic; i.e., the involved fields of all three neuronal populations
receive commissural fibers, the involved fields of CA3e and CA4 pyramids receive
mossy fibers and those of CAI pyramids receive entorhinal perforant fibers 13. Finally,
CA2 pyramids and dentate granule cells were considered relatively insensitive by
Nadler et al. and we detected involvement of their dendritic trees only at one level.
Thus, there are important similarities between our descriptive findings and those of
Nadler et al. 7. We are not inclined, however, to interpret our findings as they
interpreted theirs. Since the earliest toxic changes we detected were localized to
 99

portions of hippocampal neurons that house excitatory synaptic receptors and the
particular dendritic fields involved were specific fields thought to receive glutamergic
inputs, we are hesitant to reject the hypothesis that Glu synaptic receptors mediate KA
toxicity.
Accepting the above hypothesis for discussion purposes, the differential sensiti-
vities of hippocampal neurons to KA toxicity might be explained as follows. First, any
neuron endowed with at least a few synaptic receptors specialized for glutamergic
transmission may be sensitive to KA in the sense that dendritic segments having such
receptors may be focally affected. Secondly, the percentage of a neuron's surface that is
occupied by Glu receptors may be a very important determinant of differential vulner-
ability to cell death. If focal dendritic swellings following KA treatment do mark the
loci of Glu synaptic receptors, our findings suggest that a large percentage of the
dendritic surfaces of CA3a and CA3b pyramids may be occupied by Glu receptors
whereas a relatively smaller percentage of the dendritic surfaces of dendate granules
are thus occupied. Low tissue concentrations of KA would be expected to kill the
former but not the latter neurons if, as we postulate, the cytotoxic threshold varies
inversely with the surface area occupied by Glu synaptic receptors. Thirdly, the level
on a neuron's dendritic tree at which Glu receptors are located may be a further
determinant of its vulnerability to the neurodestructive action of KA. All neurons
innervated by mossy fibers are highly vulnerable. Mossy fibers innervate hippocampal
pyramids on spinous excrescences given off by their proximal dendritic segments.
Stimulation by KA of these elements may lead efficiently to cell death because the
locus of stimulation is so close to the cell body. If, as Olney et al. have proposed 12,
sustained depolarization-induced changes in plasma membrane permeability are the
mechanisms triggering KA-induced neuronal necrosis, interaction of KA at multiple
loci immediately adjacent to the cell body may be more effective in causing lethal
permeability changes than if impingement were distributed over the distal tips of the
dendritic arborization.
Finally, if the relative insensitivity of CA2 pyramids and dendate granule cells
is not adequately explained by their surfaces being only sparsely occupied by
glutamergic terminals, an alternate explanation might be considered. Recent evidence 17
suggests that there may be more than one type of receptor with which Glu interacts to
depolarize central neurons; this raises the distinct possibility that KA, being a rigid
analog, may be able to interact readily with only a specific sub-class of Glu receptor so
that the pattern of dendritic dilatations observed in the present study might represent
the distribution of a sub-class of Glu receptor rather than the full inventory of Glu
receptors present on the surfaces of hippocampal neurons. If this were the case, KA
might be viewed as a potential tool for mapping a specific sub-class of Glu receptor.
To explore this possibility further, the hippocampus - - a highly organized, laminar
structure whose afferent connections are fairly well understood - - may be an ideal
brain region to focus upon and systemic administration of KA to adult rats with
analysis of toxic changes only a few hours following treatment is, in our estimation, a
fruitful approach, provided techniques employed are optimally suited for revealing
focal pathological changes in dendritic processes.
100

ACKNOWLEDGEMENTS

S u p p o r t e d by N . I . H . G r a n t s NS-09156, D A - 0 0 2 5 9 , ES-07066 a n d R . S . D . A w a r d
MH-38894 (J.W.O.)

REFERENCES

1 Campochiaro, P. and Coyle, J. T., Ontogenetic development of kainate neurotoxicity: Correlates

with glutamatergic innervation, Neurobiology, 75 (1978) 2025-2029.
2 Crawford, I. L. and Connor, J. D., Localization and release of glutamic acid in relation to the
hippocampal massy fibre pathway, Nature (Lond.), 244 (1973) 442~-43.
3 Fuller, T. and Olney, J. W., Naloxone and the effects of systemically administered kainic acid,
Neurosci. Abstr., 4 (1978) 423.
4 Johnston, G. A. R., Curtis, D. R., Davies, J. and McCulloch,R. M., Spinal interneurone excitation
by conformationally restricted analogues of L-glutamic acid, Nature (Lond.), 248 (1974) 804-805.
5 McGeer, E. G., McGeer, P. L. and Singh, K., Kainate-induced degeneration of neostriatal
neurons: dependency upon corticostriatal tract, Brain Research, 139 (1978) 381-383.
6 Nadler, J. V., Perry, B. W. and Cotman, C. W., Selectivity in the destruction of hippocampal
neurons by kainic acid, Neurosci. Abstr., 3 (1977) 202.
7 Nadler, J. V., Perry, B. W. and Cotman, C. W., Preferential vulnerability of hippocampus to
intraventricular kainic acid. In E. McGeer, J. Olney and P. McGeer (Eds.), Kainic ,4cidin Neuro-
biology, Raven Press, New York, 1978.
8 0 l n e y , J. W., Glutamate-induced neuronal necrosis in the infant mouse hypothalamus. An
electron microscopic study, J. Neuropath. exp. NeuroL, 30 (1971) 75-90.
9 0 l n e y , J. W., Neurotoxicity of excitatory amino acids. In E. McGeer, J. Olney and P. McGeer
(Eds.), Kainic Acid in Neurobiology, Raven Press, New York, 1978.
10 Olney, J. W., Ho, O. L. and Rhee, V., Cytotoxic effects of acidic and sulphur-containing amino
acids on the infant mouse central nervous system, Exp. Brain Res., 14 (1971) 61-76.
11 Olney, J. W., Rhee, V. and Ho, O. L., Kainic acid : A powerful neurotoxic analogue of glutamate,
Brain Research, 77 (1974) 507-512.
12 Olney, J. W., Sharpe, L. G. and de Gubareff, T., Excitotoxic amino acids, Neurosci. Abstr., I
(1975) 371.
13 Raisman, G., Cowan, W. M. and Powell, T. P. S., The extrinsic afferent, commissural and asso-
ciation fibres of the hippocampus, Brain, 88 (1965) 963-998.
14 Schwob, J. E., Fuller, T. and Price, J. L., The distribution of cellular degeneration following
systemic or intracerebral injection of kainic acid, Neurosci. Abstr., 4 (1978) 227.
15 Shinozaki, H. and Konishi, S., Actions of several anthelmintics and insecticides on rat cortical
neurones, Brain Research, 24 (1970) 368-371.
16 Storm-Mathison, J., Localization of transmitter candidates in the brain: The hippocampal for-
mation as a model, Progr. NeurobioL, 8 (1977) 119-181.
17 Watkins, J. C., Excitatory amino acids. In E. McGeer, J. Olney and P. McGeer (Eds.), Kainic Acid
in Neurobiology, Raven Press, New York, 1978.
18 White, W. F., Nadler, J. V., Hamberger, A. and Cotman, C. W., Glutamate as transmitter of
hippocampal perforant path, Nature (Lond.), 270 (1977) 356-357.
19 Wuerthele, S. M., Lovell, K. L., Jones, M. Z. and Moore, K. E., A histological study of kainic
acid-induced lesions in the rat brain, Brain Research, 149 (1978) 489-497.