Neuroscience Vol. 23, No. 3, pp. 95S968, 1987 0306-4522/87 $3.00 + 0.

00
Printedin Great Britain PergamonJournals Ltd
0 1987IBRO

THE FUNCTIONAL ANATOMY AND PAThOLOGY

OF LITHIUM-PILOCARPINE AND
HIGH-DOSE PILOCARPINE SEIZURES
D. B. CLIFFORD,*$ J. W. OLNEY,~A. MANIOTIS,R. C. COLLINS*and C. F. ZORUMSKI~~
*Department of Neurology and Neurological Surgery (Neurology), tDepartment of Psychiatry,
SMcDonnell Center for Studies of Higher Brain Function, Washington University School of Medicine,
St Louis, U.S.A.

Abstract-Subcutaneous treatment of rats with low doses of lithium and pilocarpine or a high dose of
pilocarpine results in a severe seizurebrain damage syndrome. Bats thus treated were studied with
multipledepth electrodes, quantitative [14CJ2-deoxyghtcose autoradiography, and light and electron
microscopy. Bats receiving lithium-pilocarpine did not differ from high-dose pilocarpine rats in
behavioral, electrographic, metabolic or histopathological findings, but lithium-pilocarpine reproduced
the syndrome more reliably and with a lower acute mortality rate.
Organized electrographic seizure activity developed just prior to the onset of behavioral forelimb clonus
and appeared to originate from ventral forebrain in the vicinity of the ventral pallidum and/or nucleus
accumbens. From these sites activity spread rapidly to involve other regions. Once initiated, electrographic
seizures persisted for hours. Increased glucose utilixation was found in most brain regions during the
period of continuous seizure activity. The greatest increases were found in the ventral pallidum, globus
pallidus, hippocampus, entorhinal cortex, amygdala, lateral septum, substantia nigra, ventrobasal and
mediodorsal tbalamus and frontal motor cortex. Animals sustaining seizures displayed a disseminated
pattern of neural degeneration not involving globus pallidus or ventral pallidum but otherwise coinciding
with the above pattern of enhanced glucose utilization. No consistent correlation was observed between
the pattern of brain damage and known regions of high muscarinic cholinergic receptor density.
Ultrastructurally, the cytopathological changes, like those associated with various other sustained seizure
syndromes, resemble the excitotoxic type of damage glutamate is known to cause.
This seizure-brain damage syndrome and that induced by systemic kainic acid appear to be similar in
behavioral but not in electrophysiological or metabolic manifestations. During kainic acid seizures,
electrographic changes are first recorded in the hippocampus while they are first detected in the ventral
forebrain repjon in pilocarpine seizures. Pilocarpine also induced metabolic activation of ventral forebrain
sites not activated by kainic acid. The cytopathology associated with the two syndromes is identical in
type but not in pattern, the cholinergic model being characterized by much greater neocortical and slightly
less hippocampal damage.
Further study of these choline@ models may provide new insights into the roles of the major excitatory
neurotransmitter systems (cholinergic and glutamergic) in limbic epilepsy.

It is postulated that choline@ mechanisms play an presses the spread of amygdaloid kindled seizures,3
important role in the onset and propagation of limbic while focal injections of cholinomimetic enhance
seizures. This belief is supported by early studiesioJ9hippocampal kindled discharges. These findings may
and by the more recent finding24 that focal, intra- be a result of the muscarinic action of acetylcholine,
mygdaloid injections of cholinomimetics producewhich produces long lasting neuronal depolarization
prolonged, recurring seizures and brain damage. In by blockade of a specific potassium conductance.
addition, rats administered repeated subconvulsive Recently Honchar et uf.* described a model of
amygdaloid injections of carbamylcholine develop a limbic seizures produced by systemic injection of
progressive behavioral seizure syndrome similar to pilocarpine (30 mg/kg) in rats pretreated with lithium
that seen with electrical kindling. Evidence support- chloride (3 mEq/kg). Behaviorally, these seizures re-
ing a role for the choline+ system in the propaga-semble other models of limbic seizures beginning with
tion of limbic seizure activity is provided by studies facial automatisms and head nodding and
demonstrating that lesions of the cholinergic cell progressing to forelimb clonus with rearing and
bodies in the substantia innominata inhibt the gener- falling. When seizures continued for several hours,
alization of electrically kindled amygdaloid seizures.i5
animals developed brain damage. A similar syndrome
Furthermore, atropine, a muscarinic antagonist, sup- of seizures and brain damage has been described by
Turski et aL3* in rats administered much higher
*Address for correspondence: Dr David B. Clifford, systemic doses of pilocarpine alone (4OOmg/kg).
Washington University School of Medicine, Box 8111, These pilocarpine seizure models provide an oppor-
Department of Neurology, 660 South Euclid Avenue,
St Louis, Missouri 63110, U.S.A. tunity to study the involvement of the central
Abbreviations: 2-DG, 2-deoxyglucose; EEG, electro- cholinergic system in the onset, propagation and
encephalogram; MlP, myo-inositol-l-phosphate. pathological consequences of limbic seizures. In this

953
954 D B. CLIFFORD cl al

paper we further characterize the functional anat- 2-Deoxyglucose autoradiography

omy, pharmacology and cytopathology of For metabolic experiments, the quantitative
seizure models. Preliminary accounts of this work [4C]2-deoxyglucose (2-DG) autoradiographic technique of
have been reported.6l34 Sokoloff et aLM was used to measure local cerebral glucose
utilization. Prior to experimentation, animals were fasted
overnight. On the day of an experiment animals were
EXPERIMENTAL PROCEDURES anesthetized with halothane and polyethylene catheters were
Male albino rats, weighing 25&35Og, were used in all inserted into the femoral artery and vein.
experiments. Drugs were purchased from Sigma Chemical The animals were then loosely restrained for purposes 01
Company (St Louis, MO) and were administered subcuta- restricting movement. Body temperature was monitored via
neously in a volume of O.llO.2 cm of normal saline. Two a rectal probe and kept at 3637C with a heat lamp.
treatment regimens were used for the main part of the study: Animals were allowed to recover from anesthesia for 4 h
(I) lithium chloride, 3 mEq/kg, followed after 24 h by before experimentation.
pilocarpine, 30 mg/kg; (2) pilocarpine, 400 mg/kg, without At the time of experiment animals were injected with
lithium pretreatment. Control rats comprised three groups: pilocarpine as described. After the first episode of forelimb
those that received only saline, those that received lithium, clonus, [4C]2-DG (60pCi/kg) was administered by intra-
3 mEq/kg, followed after 24 h by saline, and those that venous infusion over 30 s. Timed O.l-ml arterial blood
received saline followed after 24 h by pilocarpine, 30 mg/kg. samples were taken at i, $ lb 2, 3, 5, 7f 12, 20, 30 and
In addition, we conducted some pilot experiments in which 40min after 2-DG administration. The blood samples
the dose of pilocarpine was varied from 1 to 30 mg/kg in rats were centrifuged immediately and plasma radioactivity
pretreated with lithium, 3mEq/kg, or from 200 to and glucose concentrations were measured. Animals were
400 mg/kg in rats not pretreated with lithium. We also tested anesthetized with pentobarbitol45 min after 2-DO adminis-
the ability of a centrally active anticholinergic (a&opine) tration and perfused with phosphate bu&red 1% para-
and peripherally active anticholinergic (methscopolamine) formaldehyde fixative via an intracardiac cannula. The
or the anticonvulsant diazepam to modify the seizures and brains were rapidly removed, frozen and sectioned in a
neuropathological consequences of lithium-pilocarpine or cryostat.
high-dose pilocarpine treatment. Brain sections (20pm) were mounted with
[i4C]methyl-methacrylate standards (Amersham, Arlington
Electroencephalography Heights, IL) and exposed to Kodak X-AR film for 7 days.
For electrographic studies, animals were anesthetized Brain radioactivity and local glucose utilization were deter-
with halothane and concentric bipolar electrodes (Rhodes mined in regions of interest by quantitative autoradiography
Medical Instruments, Tujunga, CA) or bipolar stainless steel using a computerized densitometer. Separate measurements
twist electrodes were stereotactically placed in various re- of left- and right-sided structures were made and the means
gions according to the coordinates of Pellegrino et af.28 of these measurements were pooled for subsequent analysis.
Depth sites monitored in various animals included the Comparisons of experimental animals and controls were
hippocampus, entorhinal cortex, amygdala, lateral septal made using serial r-tests. Differences with P co.05 were
nucleus, caudate-putamen, nucleus accumbens, ventral considered significant.
globus pallidus, substantia nigra and ventrobsal thalamus.
Surface electroencenhalonram (EEG) activity was recorded
from Teflon-coated staiiess steel wires implanted in the RESULTS
inner table of bone above the motor cortex. Additional
stainless steel wires were positioned adjacent to the frontal
sinus and served as ground and indifferent references. All
electrodes were led to Amphenol pins in a connector board
The behavioral sequence of animals exhibiting
and the entire apparatus was affixed to the skull with lithium-pilocarpine seizures was identical to that seen
jewelers screws and dental acrylate. Each animal had 24 with high-dose pilocarpine seizures. For the purposes
sites monitored. After surgery animals were housed in ofthis study we have divided the syndrome into five
individual plexiglass cages and were allowed one weeks phases (Table 1). Within several mbmtes aftar ~@o-
recovery before experimental manipulation. Electrode
placements were confirmed histolgically at the termination carpine injection animals exhibited signs of per-
of the experiment. ipheral choline&c stimuiatkm. Pkereetion, sali-
At the time of study animals were connected to a Grass vation, tremor, chromodacryorrbea and dk&ea
Model 7D polygraph and EEG signals were amplised and were clearly evident 5 min after piloearpine @e&on.
displayed using 7 PS AC preamplifiers and driver amplifiers.
The cholinergic signs persisted for akmt 20 m&3&r
Animals were observed for 20-30 min prior to drug injection
and for up to 6 h after. injection but ten&d to wane as other~features d the
syndrome were manifested. Often coincident, but in
Histopathology
some animals lagging behind the cholinergic signs,
Animals were killed by intracardiac perfusion-tixation
under halothane anesthesia using a phosphate-buffered was a series of stereotyped behaviors. During this
solution of 1% narafomnthkhyde and 1.5% du- phase animals exhibited staring with oeea+!#oawll
taraldehyde. The b&ins were rant&& and cut into l-mm chewing, snif5ng and exploratory behavior. Wkhin
slabs which were additionally fhmd by immemion in 1% several minutes tbey showed a prettjtaction for mre-
osmium tetroxide for 24 h. These slabs were processed making in one comer of the cage with upward
through graded concentrations of ethanol, cleared in tolu-
ene and embedded in Arakk. For habt mieroscon~, I-rm extension of the nose and neck. Like the
sections were cut on an MT-2 Sonal hatome and a&ted signs these stereotyped movements~persisted f0r the
with Methylene Blue/Azure Il. Regkms of in- for first 20 tin after injection and eventualty gava Wrsyto
ultrastructural analysis were sediomd from the same behavioral signs 0f ictal activity. During t$e pet&d Of
blocks. Silver/gold sections were mounted on a formwar I&n
suspended over a 1 x 2 mm oval slot grid opening, stained stereotyped behavior tbe an&& c&d be distmmed
with lead citrate and uranyl acetate and viewed in a JEOL by tactile stimulation. Motor seizures @n \*itf a
10OB transmission electron microscope. noticeable increase in sniffing and chewing move-
 Lithium-pilocarpine and pilocarpine seizures 955

Table 1. Behavioral changes

Pilocarpine
Lithium-pilocarpine
Lithium (<200 (300-360 (370400
alone (10 mdk81 (30 mgjk.8) mg/k.@ m8/k8) m&k81
stagebehavior*
I Perioheral cholineraic
stimulation 0 0 100% 100% 100% 100% 100%
II Stereotyped movements 0 0 100% 100% 100% 100% 100%
III Motor seizures 0 0 0 97% 0 45% 81%
IV Status epilepticus 0 0 0 97% 0 45% 81%
V Braindamaged state 0 0 0 97% 0 45% 8l%t
N=5 5 8 34 5 11 43
lPercentage of animals displaying behavior.
tMany of these rats had to be killed in extremis l-3 h after seizure onset.

ments associated with eye blinks and head bobbing.
These symptoms rapidly progressed over 5-30 s to
include bilateral forelimb clonus with rearing and
falling. Subsequent to a motor seizure animals some-
times exhibited wet-dog shakes. The motor seizures
began by 24.6 f 1.8 (N = 26) min after pilocarpine
injection. Over the ensuing 10-15 min animals ex-
hibited 4.9 f 0.4 distinct motor seizures.
Following the period of recurring motor seizures,
by 34.6 f 2.4 min after pilocarpine injection, animals
developed continuous myoclonic jerks of the head
and forelimbs. During this period animals were not
distractible with even vigorous stimulation. This
status epilepticus behavior persisted for as long as
6 h, the longest period studied.
This behavioral sequence was typical of both
lithium-pilocarpine and high-dose pilocarpine seiz-
ures, although it showed more variability in time to
onset and specific manifestations in high-dose
pilocarpine-treated animals (see below). Control ani-
mals given lithium followed by saline exhibited none
of the behaviors described, whereas animals given
pilocarpine at doses less than 200 mg/kg exhibited
only the first two phases of the syndrome.
Pharmacology
In all studies involving lithium, a dose of 3 mEq/kg
was given. This dose has previously been reported to
give serum lithium levels of 0.10-0.20 mEq/l 24 h
after injection. * When the dose of pilocarpine was
less than 10 mg/kg, motor seizures were not produced
in lithium-pretreated rats, but doses ~20 mg/kg
reliably produced the entire syndrome (Table 1).
Neither lithium (3 mEq/kg) nor pilocarpine
(c 200 mg/kg) given alone produced motor seizures
or status epilepticus. At doses from 30&36Omg/kg,
pilocarpine produced seizures in less than 50% of
rats and at doses > 360 mg/kg the responses varied
erratically from no seizures to a severe syndrome
often terminating in death l-3 h after seizure onset.
In contrast, lithium pretreatment followed by low
dose pilocarpine (20 or 30 mg/kg) yielded predictable
results with nearly all rats developing the full seizure
brain damage syndrome and the mortality rate being
negligible.
For production of lithium-pilodarpine seizures we
routinely used a 2&24 h period between the lithium
and pilocarpine injections, although we have found
that seizures are reliably produced even if the time
interval is substantially reduced. Seizures did not
occur when the order of presentation was reversed
but co-administration of lithium (3 mEq/kg) and
pilocarpine (30 mg/kg) produced status epilepticus in
three out of five animals. When the interval between
lithium and pilocarpine injections was extended
to more than 48 h, seizures were not produced
(N = 2).
The entire behavioral syndrome described above
can be prevented by pretreatment with atropine
(1-5 mg/kg subcutaneously, 30 min before pilo-
carpine, N = 5). The peripherally acting antichohn-
ergic, methscopolamine (1 mg/kg) blocked peripheral
choline@ signs but did not prevent seizure activity.
Once motor seizures began, atropine had no effect.
Ictal behavior and electrographic status epilepticus
were aborted by intravenous injections of diazepam
P-10 mg/kg).
Electroencephalography
Ten different surface and depth sites were moni-
tored in 25 animals. In animals treated with lithium
chloride alone, pilocarpine (< 200 mg/kg) or lithium
chloride followed by pilocarpine ( I 10 mg/kg) no
spiking or organized electrographic seizures were
observed.
Figure 1 presents typical electrographic changes
recorded from an animal treated with lithium and
pilocarpine (30mg/kg). During behavioral stages 1
and 2 in which animals exhibited signs of cholinergic
stimulation and stereotyped movement, the EEG
showed minor changes. The background rhythm was
replaced by low voltage fast activity in several depth
sites, including the nucleus accumbens, ventral globus
pallidus and hippocampus, Late in the phase of
stereotyped movements, preceding the behavioral in-
crease in chewing, blinking and head nodding which
antedates motor seizures, the first spikes were seen.
of note was the fact that the stereotyped behaviors
seen in this syndrome were not associated with
organized electrographic seizure activity. At the time
956 D. B. CLIFFORD et ul

A NA
5-15-s bursts of ictal activity punctuated by 0.5.-I s
intervals of isoelectric activity.
In addition to the above electrographic-behavioral
VF
correlation we used multiple depth electrodes to map
the pattern of seizure onset. As depicted in Fig. 2,
when a single site of origin could be determined, the
earliest spikes and organized ictal discharges ema-
nated from ventral forebrain areas (ventral pallidum
and nucleus accumbens) and spread rapidly to in-
volve other sites. In other animals electrographic
seizures seemed to begin in multiple sites simulta-
B _I

neously. These findings contrast with the hippo-

campai onset of seizures seen with kainic acid (Fig.
2D). Based on 25 animals studied electro-
physiologically, it is our impression that lithium-
pilocarpine or high-dose pilocarpine seizures do not
preferentially originate in typical limbic sites such as
.---
the hippocampus or amygdala.

2-Deoxyglucose autoradiography
Autoradiograms depicting the pattern of glucose
utilization in experimental groups and contrds are
shown in Fig. 3 while metabolic rates of studied
structures are listed in Table 2. Visual inspection of
the autoradiograms reveals obvious differences be-
tween experimental and control brains in the pattern
of local glucose utilization. The overall pattern of
labeling was very similar in the two groups of experi-
L mental animals, although in many sites glucose util-
ization was higher in high-dose pilocarpine animals
compared to the lithium-pilocarpine group (Table 2).
Since glucose utPization was elevated in most
regions of the central nervous system during pilo-
carpine seizures, for the purposes of this report we
concentrated on those elevations in excess of four
times control. With such analysis, three major sys-
tems within the central nervous system show dra-
matic changes in glucose use: (1) fn ventral forebrain
Fig. 1. Electrographic recording obtaiaed from an animal regions, glucose utilization was markedfy increased in
fill&J the development of &hium-pM&rp& 4zures. the ventral pallidurn, ventromedial globus pallidus,
Sites monisond are nudeus 8ccumbea8 (top trace), ventral olfactory tubercie and pyriform cortex. Although the
FaMum(lni@ktalld (bottom trace). ventral pallidurn and globus pallidus contain cholin-
Caliition: 1.0 mV, 1 s. (A) r%cord&Uhafter ergic cell bodies, other ventral forebrain regions with
i@ection of litbilnn (3 &kg). T&tin* b not di&er
from pm-drug EEG (not shown). (B) EPeetmgM#&ic choline+ cell bodies did not show such marked
changes da the phase of early b&&ore1 (C) elevations. The medial septal nucleus and the vertical
and horizontal limbs of the diagona1 band nucleus
showed less marked increases of only two times
control. (2) Within the limbic system all hippocampal
subfields in both dorsal and ventral regions show
markedly increased glucose utilization as did the
entorhinal cortex, amygdala and lateral septal
nucleus. (3) Other sites with marked elevation of
of the first motor seizure, the isolated spikes or- glucose utilization included retions typically consid-
ganizadtojXoduce@rl sciaue. D&&g ered to be part of the motor system, such as the
the pariod of recur&g motor se&was there was substantia nigra and frontal motor cortex. Within the
cycli~oforga&cdi&tZ&ctitityw&hp&odsof thalamus the ventrobasal and medial dorsal nuclei
non-o&M elc&rographic aetitity. However, showed the greatest elevations. Regions of experi-
within S-10 min all wh and s&bee EEG s&es mental brains that displayed very littte change include
showed umtinuous orgaaiaed seizure activity. During superior colliculus, dorsolateral geniculate, and peri-
the period of status epilepticus all EEG sites exhibited aqueductal gray.
A HC C
vp

SN -

B HC

NA

AMY

Fig. 2. Pattern of electrographic seizure onset in two animals (A and C) treated with lithium-pilocarpine and one animal (B) treated with high-dose pilocarpine. For comparison
the pattern of electrographic seizure onset from an animal treated wtih kainic acid (12 mg/kg s.c.) is shown in (D). The traces display the first organized electrographic seizure in
each animal. HC, hippocampus; CTX, motor cortex; SN, substantia n&a; NA, nucleus accumbens; AMY, amygdala; VP, ventral pallidum; EC, entorhinal cortex. Calibration bars:
(A) HC = 1.5 mV, CTX = 0.5 mV, SN = 1.OmV, NA = 0.5 mV; (B) 0.5 mV for all; (C) VP = 1.OmV, EC = 0.5 mV, HC = 1.OmV, SN = 0.25 mV; (D) SN = 1.OmV, NA = 0.5 mV,
HC = 1.5 mV. Time = 1 s for all.
 Lithium-pilocarpine and pilocarpine seizures 959

Table 2. Autoradiography of rats with cholinergic seixures

Lithium/pilocarpine Control Pilocarpine
Mean S.E.M. Mean , S.E.M. Mean S.E.M.
Frontal cortex 295.67 25.50 73.77 5.41 323.97 39.57
CiWdotC 275.77 23.21 83.20 6.30 277.33 40.77
Medial accumbcns 174.57 16.75 70.10 7.03 189.70 41.38
Lateral accumbens 175.90 16.23 68.47 5.88 187.80 26.71
Pyriform cortex 321.13 21.41 73.00 6.66 312.73 28.10
Corpus collosum 113.13 15.59 17.00 2.47 117.93 16.44
Lateral septum 253.50 20.47 51.03 5.29 275.50 36.67
Mediil septum 156.00 14.70 68.60 5.98 173.53 29.61
Ventral diagonal band 138.97 18.49 66.40 3.43 163.57 31.13
Ventral pallidum 289.00 24.19 55.73 7.56 271.42 36.83
Olfactory tubercle 314.17 27.06 73.87 8.10 300.55 42.53
Globus palhdus 236.83 24.35 42.40 4.65 238.20 35.65
Horizontal diagonal band 144.00 27.25 65.10 4.10 156.30 32.67
Anterior cingulate cortex 240.63 30.30 89.60 2.77 295.17 55.13
Dorsal hippocampus CA1 172.83 13.41 35.03 8.24 249.70 37.43
Dorsal hippocampus perforant path 221.30 18.75 43.87 6.57 263.40 34.52
Dorsal hippocampus dor. dentate 213.27 18.49 47.83 8.71 278.43 37.98
Dorsal hippocampus ventral dentate 208.93 19.83 33.83 3.02 248.03 37.67
Dorsal hippocampus CA3 232.07 17.22 45.47 8.79 250.71 52.30
Amygdala 255.30 20.32 50.10 7.20 284.17 28.90
Ventrobasal thalamas 264.83 17.30 65.97 5.90 333.87 58.17
Medial dorsal thalamus 291.60 23.33 54.13 9.41 326.37 32.48
Sensory cortex 258.60 27.46 93.00 13.47 370.13 55.68
Hypothalamus 135.37 15.61 46.93 6.85 124.50 22.44
Hilum dentate. 193.50 22.22 39.33 6.68 229.37 32.07
Habenula 144.00 11.23 62.13 9.56 190.87 27.04
Retrosplenial cortex 203.87 25.64 72.13 7.31 284.80 46.59
Visual cortex 215.93 18.36 68.10 5.80 285.67 44.09
Auditory cortex 235.20 16.71 116.65 5.36 308.33 47.26
Entorhinal cortex 260.90 25.87 43.90 5.96 275.27 22.99
Subiculum 256.97 22.04 46.13 6.88 290.80 29.35
Ventral hippocampus CA1 230.532 14.56 34.30 4.49 263.80 29.86
Ventral hippocampus CA3 242.10 22.47 43.37 6.88 252.37 31.41
Ventral hippocampus dentate 274.43 33.40 52.10 7.11 281.00 39.51
Substantia nigra 271.23 11.83 48.60 4.97 301.57 37.91
Medial geniculate 231.43 23.15 99.47 5.63 317.57 47.13
Dorsal lat. geniculate 69.27 7.72 51.13 6.04 109.57 19.25
Medial posterior thalamus 65.10 2.07 45.03 6.03 77.33 18.47
Superior colliiulus 72.53 1.33 65.67 3.80 94.17 23.69
Periaquaductal gray 91.77 4.00 56.47 5.92 119.40 28.36

Neuropathology quently affected were mediodorsal, laterodorsal, ven-

Cell damage produced by these seizures was identi-
cal whether seizures were produced by lithium-
pilocarpine or high-dose pilocarpine. Animals treated
with lithium alone or subconvulvsive doses of pilo-
carpine sustained neither seizures nor brain damage.
If seizures were blocked by pretreatment with
atropine or aborted by diaxepam, cell damage was
prevented.
Light microscopic studies revealed that in the
presence of ongoing seizure activity, many brain
regions sustained cell damage, including the anterior
olfactory nucleus, pyriform and entorhinal cortices,
thalamus, amygdala, hippocampus, subiculum, lat-
eral septum, bed nucleus of the stria terminalis,
claustrum, substantia nigra and neocortex (Figs 4-S).
In the most severely damaged brains most regions
of neocortex and several thalamic and amygdaloid
nuclei were involved. In neocortex, acute cyto-
pathological changes were most prominent in layers
2 and 3 but scattered, less conspicuous changes were
also detected in layers 46. Thalamic regions fre-
trobasal and pretectal nuclei. In the amygdala, the
cortical, medial and lateral (but not basolateral)
nuclei and the amygdalohippocampal area were most
consistently and severely damaged. In the hippo
campus, damage was consistently greater in the ven-
tral than dorsal regions although conspicuous neural
damage was consistently seen in both ventral and
dorsal ssubiculum. Layers CA3 and CA1 were both
usually involved, especially in the ventral hippo-
campus; in the dorsal hippocampus of a few animals,
acute changes were largely restricted to CA3. Edema-
tous changes affecting glia and intemeurons in the
dentate hilium were frequently present and tended
to vary in severity with the intensity of the CA3
reaction.
The distribution of acute cytopathological changes
following pilocarpine or lithium-pilocarpine seizures
was very similar to that observed as a consequence of
sustained seizures induced by systemic administration
of kainic acid. However, regional differences in de-
gree of damage could be discerned; for example, the
960 D. 8. CLBFORDet al

Fig. 4. All scenes are from the cortical nucleus of the rat amygdala 4 h after treatment with pilocarpine
(~rn~~~ subcutaneous. (a) Light micro~~pic overview depicting the major ~to~~hoi~~i changes
typically associated with Pilgrim-ind~d seizures, including numerous swolIen dendrites (smaI1
vacuous spaces conferring a Swiss cheese appearance) and neuronal cell bodies manifesting either
edematous swelling or dark ~11 changes. (b) Ekctron micrograph of a swollen neuron undergoing
edematous degeneration while in synaptic contact with a normal-appearing axon terminal (see boxed
region abstracted at higher magniEc&on). (c) Electron micrograph of a swollen dendrite contacted by
a normal-appearing presynaptic axon terminal (see boxed region and abstracted at higher ma~~cation).
(a) x 100; (b) x SGUO(inset x 30,WO); (c) x 4@30(inset x 29.000).
 Lit~~-~l~~e and pilocarpine seixures

Fig. 5. All scenes are from the hippocampus of a lithium-pilocarpine-treated rat killed after 3 h of status
seizure activity. (a) Light microscopic overview of the dentate hilus/CA3 region showing massive
edematous changes affecting both glial and neuronal elements at the inner margins of the dorsal and
ventral dentate. granule cell layers and forming a rarBed band engulfmg the cell bodies and proximal
dendrites of CA3 pymmidal neurons in a pattern caning to the mossy fiber innervation of these
neurons. (b) Electron micrograph of the CA1 pyramidal layer (lower portion), statum o&ens (middle) and
alveus (top) showing early dark cell changes in CA1 pyramidal cell bodies, swelling of glia surrounding
these neurons and massive dilatation of basilar dendritic processes in the outer portion of stratum oriens.
(c) A magnified view of a swollen dendrite from stratum oriens. Note the normal-appearing axon terminal
in synaptic contact (arrowhead) with this acutely degenerating dendrite. (a) x 80; (b) x 260, (c) x 14,000.
962 D. B. CLIFFORD er at?

Fig. 6. These scenes from a rat killed 3 h after highdose pifocarpine depict neurons in the lateral septal
nucleus (a) and mediodorsal nucleus of the thakunus (b) displaying massive edematous swelling in contrast
to the dark cell (vacuolar condensation) changes affecting a neuron in the dorsal lateral thalamic nucleus
of the same rat (c). Elsewhere we have described both the edematous and dark-cell type of neuro-
degenerative changes in association with various other seizure-related brain-damage syndromes or with
exogenous glutamate treatment. (a) and (b) x 160; (c) x6000.
Fig. 7. (b) and (c) Froth the piriform cortex and (e) from the cingulate cortex of a lithmm-pllocarpme
rat killed 2 h after onset of status seizure activity. (a) and (d) From the piriform and cingulate cortices,
reqectively, of control rats protected from lithium-pilocarpine-induced Azures and brain damage by
pretreatment with diazqam (a) or atropine (d). The massively swollen dendrite in (c) is from the basilar
dendritic field of the piriform cortical scene depicted in (b). Note that the abnormal dendrite in (c) is in
synaptic contact (arrow) with a normal-appearing axon terminal. (a) and (b) x 140, (c) x 9000; (d) and
(e) x50.
863
 D. B. CLIFFORLI et al.

Fig. 8. (a) and (b) From the somatosensory cortex of a rat treated with lithium-pilocarpine (a) or diazepa
plus lithium-pilocarpine (b). (c) An acute lesion affecting the substantia nigra, pars reticulata of
lithium-pilocarpine-treated rat. Note the remarkably discrete *punched out appearance of this lesia
(a) and (b) x 180; (c) x 60.
 Lithium-pilocarpine and pilocarpine seizures
hippocampus and neocortex are usually damaged to
some extent in either the kainic acid or cholinergic
seizure model but the predilection is somewhat
greater for hippocampal damage in the kainic acid
syndrome and much greater for neocortical damage
in the cholinergic syndrome.
At the electron microscopic level the cell damage
associated with either pilocarpine or lithium-pilo-
cat-pine seizures consisted of massive swelling of
dendrites, swelling or vacuolar condensation of neu-
ronal cell bodies and marked dilatation of astroglial
elements with relative sparing of axonal components
(Figs 4-6,7). This is the same type of cytopathology
that has been described in association with prolonged
seizures induced by any of several means.26 Elsewhere
we recently compared this type of cytopathology
with that induced by direct exposure of brain tissue
to exogenous glutamate and concluded that the
two forms of cytopathology are indistinguishable-
all features of the excitotoxic reaction of brain tissue
to exogenous glutamate are reproduced in these
seizure-related cytopathological syndromes and all
features of the cytopathology associated with pro-
longed seizures have previously been described as a
consequence of glutamate neurotoxicity.*
DISCUSSION
Turski and colleagues32 recently described electro-
graphic and light microscopic histopathological
changes associated with high-dose pilocarpine
(4OOmg/kg) seizures in the rat. The pattern of brain
damage they found corresponds very closely with the
pattern we describe here. An important difference
between our findings and theirs, however, is that they
identified the hippocampus rather than ventral fore-
brain as the site of origin of seizure activity. This
apparent contradiction is readily explained by the
fact that they did not study ventral forebrain record-
ing sites.
Our observations suggest that the syndromes
produced by lithium-pilocarpine and high-dose
pilocarpine are behaviorally, metabolically, electro-
graphically and neuropathologically indistinguish-
able. However, the fully developed syndrome is
produced more reliably and with less mortality by the
combination of lithium and pilocarpine than by a
high dose of pilocarpine alone. Although our study
does not clarify the exact mechanisms underlying
seizure induction and propagation, muscarinic recep-
tor activation appears to play an important role. This
presumption is supported by evidence that the syn-
drome is induced by an injection of pilocarpine, a
muscarinic agonist, and can be blocked by systemic
administration of atropine, a muscarinic antagonist.
Once seizures are initiated, however, other mech-
anisms presumably become operative and muscarinic
receptors may play a less prominent role.
965
Comparing this cholinergic syndrome with another
extensively studied seizurebrain damage syndrome,
that induced by kainic acid, reveals interesting simi-
larities but signiticant differences. Pilocarpine seizure
activity was first detected in the ventral forebrain,
whereas seizures induced by kainic acid appear to
originate in the hippocampus. The exceedingly high
density of kainic acid receptors in the hippocampus3
presumably explains the origin of kainic acid seizures
in this structure, and by analogy, one might argue
that cholinoceptive neurons, either located in the
ventral forebrain or feeding directly into it, may be
the principal generator of the pilocarpine seizure
syndrome. According to our findings, the ventral
pallidum or nucleus accumbens might be the site of
origin of pilocarpine-induced seizure activity. It
seems more likely that nucleus accumbens is the
primary site since it has been shown to have the
highest density of muscarinic receptors in rat
brain.6*23The ventral pallidum, on the other hand, is
rich in cholinergic cell bodies but not in cholinocep-
tive neurons. To explain the early electrical activity
and strong increase in metabolic activity in the
ventral pallidum one might postulate a monosynaptic
relay of activity from nucleus accumbens to ventral
pallidum in the early phase of seizure propagation.
Activation of ventral pallidal neurons would, in turn,
explain the diffuse activation of neocortex since ven-
tral pallidal neurons project diffusely to the neo-
cortex, although primary activation of neocortical
cholinoceptive neurons by pilocarpine may also be a
contributing factor. Conversely, relative containment
of the kainic acid syndrome within limbic structures
with minimal involvement of neocortex is under-
standable, in terms of a hippocampal origin of this
syndrome and lack of hippocampal-neocortical
projections.
In contrast to kainic acid or hippocampal kindled
seizures, which entail a gradual build up and
progressive spread of electrographic activity through
the limbic system, pilocarpine-induced seizures typi-
cally become generalized much more rapidly. It is
possible that ventral forebrain sites with direct
projections to neocortex, amygdala and olfactory
cortical regions play an important role in facilitating
rapid seizure generalization. Stimulation of striatal
cholinoceptive neurons with spread of activity
through the striatonigral tract to substantia nigra
may also contribute in view of the important role that
has recently been attributed to substantia nigra in the
development and maintenance of status epilepticus
syndromes.
Since the neocortex was diffusely damaged by
pilocarpine and the neocortex is diffusely supplied
with cholinergic receptors, the question arises
whether cholinergic receptor stimulation might be
directly responsible for the cytopathology associated
with pilocarpine seizures. It is difficult to rule out this
possibility on the basis of neocortex alone. However,
several subcortical regions that have high muscarinic
966 D. B. CLIFFORD et a/.
receptor concentrations, including nucleus accum-
bens and caudate nucleus, were not damaged. In the
hippocampus the highest cholinergic receptor densi-
ties are in CA1 and dentate gyrus but the region most
consistently and severely damaged was CA3. Simi-
larly, in the amygdala the highest cholinergic receptor
density is in the basolateral nucleus while the most
severely damaged regions were the cortical, medial
and lateral nuclei.
The pattern of increased metabolic activity associ-
ated with pilocarpine seizures does not correlate well
with muscarinic receptor binding patterns.6*2 This is
not surprising since the spread of seizure activity
beyond the initial focus would entail activation of
numerous non-cholinergic pathways and involvement
of various other transmitter systems. The same need
not be true for the cytopathology since it is possible
that the receptors for a specific transmitter system
might mediate the neurotoxic process. Indeed, a
leading hypothesis to explain seizure-related brain
damage centers on the role that glutamate receptors
may play. The type of cytopathology seen with
pilocarpine seizures is identical to that associated
with other sustained seizure syndromes which, in
turn, is identical to the excitotoxic changes glutamate
is known to cause: massive swelling of dendrites and
astroglial processes and edematous swelling or vacu-
olar dark cell degeneration of neuronal cell bodies,
with sparing of axons. Moreover, it has been shown
that continuous electrical or pharmacological stimu-
lation of glutamergic tracts results in glutamate-like
cytopathology in neurons innervated by such
tracts.*** The correspondence between regions of
high glutamate receptor density and pilocarpine
seizure-related brain damage was relatively close but
not perfect. Areas of relatively high glutamate recep-
tor density which were damaged include: neocortex
(superficial layers especially), hippocampus (CA1 and
CA3), pyriform cortex, lateral septum, anterior olfac-
tory nucleus and several thalamic and amygdaloid
nuclei. To support the glutamate hypothesis, the
brain damage would not have to occur in every
region harboring a high density of glutamate recep-
tors, but rather should occur in those regions that are
both rich in glutamate receptors and are intensely
activated. Given this qualification, our findings
strongly support the glutamate hypothesis with one
notable exception-the substantia nigra was severely
damaged (Fig. 8c) in approximately 25% of these
animals and it reportedly has a very low concen-
tration of glutamate receptors.22
The intensity of glucose utilization in these seizures
deserves comment. In both seizure groups, glucose
utilization increased as much as 7-fold in some
structures, reaching levels of glucose metabolism in
excess of 3OO~mol/1OOg/min (see Table 2), levels
matched only by lateral septum, CA3 of the hippo-
campus and late measurements of subicular activity
during kainic acid-induced seizures reported by
Lothman and Collins. Areas in which we observed
at least quintupling of glucose utilization over
controls include several hippocampal subfields, the
dentate gyrus, globus pallidus, ventral pallidum,
amygdala, substantia nigra and medial dorsal thala-
mus (see Table 2). Areas slightly less intensely acti-
vated with increases of 3-4-fold over control include
lateral septum, pyriform cortex, olfactory tubercle,
the ventrobasal thalamus, frontal and visual cortex
and the caudate. A surprising aspect of the pattern of
increased metabolism is the marked increase in activ-
ity in the corpus collosum. While white matter has
substantially lower baseline values than other regions
of the brain, the change in callosal metabohc activity
is among the most marked. The explanation for this
finding is unclear, but does illustrate the hazard of
non-quantitative deoxyglucose studies in which the
corpus callosum is used as an internal reference.
Similar elevations of glucose utilization have been
observed in the white matter mammillothalamic tract
during pentylenetetrazole seizures* indicating that
this phenomenon may occur in a variety of seizure
models.
The pattern of enhanced glucose utilization re-
ported here is similar but not identical to that re-
ported in seizures produced with the cholinesterase-
inhibiting toxin, soman. Soman-induced seizures led
to 2-5-fold increases in glucose utilization in the
hippocampal body, dentate gyrus, lateral septum,
frontoparietal cortex, ventral thalamus, caudate,
medial geniculate, interpeduncular nucleus and sub-
stantia nigra reticularis. zB.r)Differences between the
soman and pilocarpine syndromes might be explained
in part by the fact that choline&erase inhibition
would not be limited to muscarinic stimulation but
rather would entail activation of all chohnergic recep
tors in the CNS; alternatively, it is possible that
soman has actions other than pure cholinesterase
inhibition.
A major question regarding the lithium-pilocar-
pine seizure model is the mechanism by which lithium
alters the response to pilocarpine. It is clear from our
studies that pretreatment with lithium results in
nearly a 20-fold shift in the pilocarpine dose-response
curve for producing seizures. There is some evidence
to suggest that lithium is a miid proeonvulsant but
none that would predict the order of magnitude
change in epileptic properties which we demonstrate
in these experiments. Lithium is known to have
several actions. It has been shown to decrease nor-
epinephrine and dopamine release at stimulated nerve
terminals.4 The decreased norepinephrine release
might predispose to increased seizure susceptibility
since other investigators have reported that bfockage
of noradrenergic function increases kindled seizure
activity.3 Lithium also appears to directly influence
cholinergic function in several ways. Lithium is re-
ported to increase release of acetylcholine.~i4 Other
investigators have reported changes in muscarinic
binding due to lithium. One intriguing way that
lithium interacts with the muscarinic system involves
 Li~~il~ine and pilocarpine seixures 967

the inositol phospholipid second messenger system.
Allison et al.* have reported that lithium decreases
free inositol and increases myo-inositol-l-phosphate
(MlP) in rat cerebral cortex due to inhibition of
my~ino~toI- 1-phosphatase. This alteration appears
to be related to cholinergic function since pre-
treatment of animals with atropine abolishes the
response. It is interesting that pilocarpine alone
causes a Cfold increase in MlP levels whereas the
combination of lithium and pilocarpine used in our
model causes a @-fold increase.* This l&fold change
of ratio roughly mirrors the ~plification of pilo-
carpine effect on seizures we have observed. Thus, it
seems possible that alterations in second messenger
systems or their targets play a part in this interaction.
behavioral, electrographic, metabolic and neuro-
pathological changes associated with these two
cholinergic models suggests that the initiation and
spread of seizure and neurotoxic activity in the two
models may occur by the same m~han~(s). Since
the full seizure/brain damage syndrome is more
dependably reproduced by lithium-pilocarpine than
by high-dose pilocarpine, the former regimen may
prove to be the more useful for studying the role
of choline&c mechanisms in epilepsy. Moreover,
research focusing on the ~thi~~ii~~ine syn-
drome offers the additional advantage that it may
eventually help to elucidate the mechanism by which
lithium augments the epileptogenic and neurotoxic
properties of choline@ agents.

Acknowledgements-We wish to thank Arm Benz for excel-

CONCLUSION
We have studied the response of rats to two
cholinergic treatment regimens-low-dose pilocar-
pine following lithium pretreatment or pilocarpine
alone at a much higher dose. The similarity in the
lent technical support and Nancy Miller for preparation of
the manuscript. This study was supported by The Institute
of Medical Education and Research @B.C.), McDonnell
Center for Higher Brain Function (C.F.Z.), NIH ES07066
(C.F.Z.), RSA-MN 38894 (J.W.O.), DAMD 17-86-C-6010
(J.W.O.).

REFRRENCES

1. Adams P. R., Brown D. A. and Constanti A. (1982) M-currents and other potassium currents in bullfrog sympathetic
neurons. J. Physioi., Lmd. 330, 557-572.
2. Allison J. H., Boshans R. L., Hallcher L. M., Packman P. M. and Sherman W. R. (1980) The e&c& of lithium on
myo-inositol levels of frontal cortex, in cerebellum, and in corpus collisum of the rat. 1. Neurockm. 34, 456-458.
3. Arnold P. S., Racine R. J. and Wise R. A. (1973) Effects of atropine, reserlrine, ~hydroxydo~ne, and handling
on seixure development in the rat. Expt Nearol. 4fh457-470.
4. BaMessarini R. J. (1985) Drugs and the treatment of psychiatric disorders. In The PItarmacologic~ Basic of
Therapeutics, 7th edn (eds Gihnan A. G., Goodman L. S., Rall T. W. and Murad F. P.), p. 427. Macmillan, New York.
5. Bumhfiel J. L., Duchowny M. S. and DulTy F. N. (1979) Neuronal supersensitivity to acetylcholine induced by kindling
in the rat hippocampus. Science, N.Y. 2@4, 1096-1098.
6. Clifford D. B., Olney J. W., Colins R. C., Maniotis A. and Zorumski C. F. (1986) Lithium-pilocarpine and pilocatpine
induced se&urea; metabolism and neuropathology. Sot. Newosci. A&r. 12, 6%.
7. CliBord D. B, Podolsky A. and Zorumski C. F. (1985) Acute effects of lithium on hippocampal kindled seizures.
Epikpssia 26, 689692.
8. Collins R. C. and Olney J. W. (1982) Focal cortical seizures cause distant thalamic lesions. Science, N. Y. 218,177-179.
9. Collins R. C., Teame R. G. and Lothman E. W. (1983) Functional anatomy of limbic seizures: Focal discharges from
medial entorhinal cortex in rat. Bruin Res. 280,2-.
10. Grossman S. P. (1%3) Chemically induced epileptiform seixures in the cat. S&ace, N. Y. 142, 409411.
11. Haas H. L. and Ryall R. W. (1977) An excitatory action of iontophoretically administered lithium on mammalian
central neurons. Br. J. Pharmac. 60, 185-195.
12. Honchar M. P., Olney J. W. and Sherman W. R. (1983) Systemic cholinergic agents induce seixures and brain damage
in lithium-treated rats. Science, N.Y. 228, 323-325.
13. Hruska R. E.. Ludmer L.M.. Pert A. and Bunnev W. E. (1984) Effects of lithium on 3H-ouinuclidinvl benzilate bindina
to rat brain muscarinic cholinergic receptors. J: Neurosh. R&. 11, 171-177.
14. Jope R. S. (1979) Effects of lithium treatment in vi00 and in r&o on acetylcholine metabolism in rat brain. J. Neurochem.
a487-495.
15. Kimura H., Kmeko Y. and Wada J. A. (1981) Catecholamlnergic and cholinergic systems and amygdaloid kindling.
In K&&i&g 2 (ad. Wada J. A.), pp. 265-287. Raven Press, New York.
16. Kobayashi R. hi., Palkovits M., Hruska R. E., Rot~~ld R. and Y~~um H. I. (1978) Regional dist~bution of
muscarinic cholinergic receptors in rat brain. Brain Res. 154, 13-23.
17. L&man E. W. and Collins R. C. (1981) Kainic acid induced limbic seizures: Metabolic, behavioral, electro-
13KCphalO~mphiC and neuropathological correlates. Brain Res. 218, 299-318.
18. Lothman E. W., Hatleid J. M. and Zorumski C. F. (1985) Functional mapping of limbic seizures originating in the
hippocampus: a combined 2deoxyglucose and electrophysiologic study. Brain Res. 360, 92-100.
19. MacLean P. D. and Delgado I. M. R. (1953) Electrical and chemical stimulation of frontotemporal portion of limbic
system in the waking animal. E~~r~nce~. clin. N~o~ys~~. !5,91-100.
20. MeDonough J. H., Hackley B. II., Cross R., Samson F. and Nelson S. (1983) Brain regional glucose use during
Soman-induced seixwea. Newotoxicology 4, 203-2 10.
21. Mirski M. A. and Ferrendelli J. A. (1986) Selective metabolic-activation of the mamillary bodies and their connections
during ethosuximide-induced suppression of pentylenetetraxole seixures. Epilepsia 27, M-203.
22. Monaghan D. T. and Cotman C. W. (1985) Distribution of N-methyl-waspartate-sensitive t.-[HI-glutamate binding
sites in rat brain. J. Neurosci. 5, 2909-2919.
968 D. B. CLIFFORDet ~1.

23. Nonaka R. and Moroii T. (1984) Ouantitative autoradioarauhv

_ . I of muscarinic cholinereic
- recenters
. in rat brain. Bruin
Res. 296, 295-303. _ _
24 Olney J. W., de Gubareff T. and Labruyere J. (1983) Seizure-related brain damage induced by cholinergic agents. Nature
301, 52&522.
25. Olney J. W., deGubareff T. and Sloviter R. S. (1983) Epileptic brain damage in rats induced by sustained electrical
stimulation of the perforant path. II. Ultrastructural analysis of acute hippocampal pathology. Brain Res. Eul/. 10,
6999712.
26. Olney J. W., Collins R. C. and Sloviter R. S. (1986) Excitotoxic mechanisms of epileptic brain damage. In Aduunces
in Neurology, Vol. 44, Basic Mechanisms of the Epilepsies: Molecular and Cellular Approach (eds Delgado-Escueta
A. V., Ward A. A., Woodbury D. M. and Porter R. J.), pp. 857-878. Raven Press, New York.
27. Olney J. W. (In press) Inciting excitotoxic cytocide among central neurons. In Excitatory Amino Acids and Seizure
Disorders (eds Ben-Ari Y. and Schwartz R.), Plenum Press, New York.
28. Pellegrino L. J., Pellegrino A. S. and Cushman A. J. (1979) A Sfereotaxic Atlas of the Rat Brain, 2nd edn. Plenum
Press, New York.
29. Samson F. E., Pazdemik T. L., Cross R. S., Giesler M. P., Mewes K., Nelson S. R. and McDonough J. H. (1984)
Soman induced changes in brain regional glucose use. Fundam. appl. Toxic. 4, 5173-5183.
30. Sokoloff L., Reivich M., Kennedy C., Des Rosiers M. H., Patlak C. S., Pettigrew K. D., Sakurada 0. and Shinohara
M. (1977)The [Qieoxyglucose method for the measurement of local cerebral glucose utilization: Theory, procedures
and normal values in the conscious and anesthetized albino rat. J. Neurochem. 28, 897-916.
31. Unnerstall J. R. and Wamsby J. K. (1983) Autoradiographic localization of high-affinity [Hlkainic acid binding sites
in the rat forebrain. Eur. J. Pharmac. 86, 361-371.
32. Turski W. A., Cavalhiero E. A., Schwarz M., Czuczwar S. J., Kleinrok Z. and Turski L. (1983) Limbic seizures
produced by pilocarpine in rats: behavioral, electroencephalographic and neuropathological study. Behau. Brain Res.
9, 315-335.
33. Wasterlain C. G. and Jonec V. (1983) Chemical kindling by muscarinic amygdaloid stimulation in the rat. Brain Res.
271, 31 l-323.
34. Zorumski C. F., Collins R. C., Olney J. W. and Clifford D. B. (1984) Lithium-pilocarpine seizures: In uivo studies.
Sot. Neurosci. Abstr. 10, 552.

(Accepted 1 April 1987)