вход по аккаунту


Correlation between extracellular glucose and seizure susceptibility in adult rats.

код для вставкиСкачать
Correlation between Extracellular Glucose
and Seizure Susceptibility in Adult Rats
Evan M. Schwechter, BA,1,2 Jana Velı́s̆ková, MD, PhD,1,3,4 and Libor Velı́s̆ek, MD, PhD1,3,4
In adult diabetic patients, periods of hyperglycemia may be associated with exacerbation of focal seizures. Our objective
was to determine in the adult rats the correlation between seizure susceptibility and extracellular glucose concentration
in two models of seizures. Male rats were injected with two doses of streptozocin (40mg/kg IP) on 2 consecutive days to
induce diabetic hyperglycemia. Controls either received vehicle or were not injected. After 2 weeks, blood glucose concentration was measured, and the rats were subjected to flurothyl seizure test. Another group of rats received glucose
solution (20%, 5ml IP) 30 minutes before testing to induce nondiabetic hyperglycemia. Thresholds for flurothyl-induced
clonic and tonic-clonic seizures were determined. Finally, in vitro epileptiform activity was induced in the entorhinal
cortex-hippocampal slices from naive rats by perfusing with magnesium-free medium with various glucose concentrations. In additional slices, paired-pulse paradigm was determined in the perforant path. Susceptibility to clonic and
tonic-clonic flurothyl-induced seizures positively correlated with blood glucose concentrations as the increased glucose
concentration was associated with proconvulsant effects. Similarly, in the in vitro experiments, epileptiform activity was
promoted by increased and suppressed by decreased glucose concentrations. Data indicate that, in the adult rats, high
glucose concentrations are associated with proconvulsant effects.
Ann Neurol 2003;53:91–101
There are reports that in the adult diabetic patients hyperglycemia may precipitate seizures, especially those of
focal origin.1–3 This is in contrast with juvenile diabetic patients in whom the episodes of severe hypoglycemia frequently are associated with alterations of mental state, such as confusion or loss of orientation,4 and
eventually seizures and coma may occur.5–7 Hyperglycemia also is known to aggravate ischemic brain damage8 –10 (although there may be differential effects of
long-term and short-term hyperglycemia11–13), whereas
fasting-induced hypoglycemia has protective effects in
ischemia and neurotoxic neuronal damage.14 –17 Similarly, fasting-induced hypoglycemia is associated with
decreases in susceptibility to hyperbaric oxygen-induced
seizures.18 One might predict based on these data that
decreases in extracellular glucose levels decrease neuronal excitation leading to a decrease in seizure occurrence.
In vitro recordings in brain slices support the notion
that low extracellular glucose inhibits, whereas high extracellular glucose facilitates, synaptic transmission.
Perfusion with 1mM glucose (down from 5mM) reversibly suppresses extracellular excitatory postsynaptic
potentials in the cornu ammonis 1 region of hip-
pocampal slices induced by electrical stimulation of
Schaffer collaterals in slices from both euglycemic19
and diabetic20 rats. Similar data were collected using
intracellular recordings in the dorsolateral septal nucleus in euglycemic rats.21 The results may reflect a
glutamate decrease in synaptic terminals during hypoglycemia.22 Thus, the relationship between extracellular
glucose concentration, synaptic transmission, and seizure susceptibility is probably complex.
Here, we studied whether variations of extracellular
glucose concentration affect seizure susceptibility in
adult rats. Our hypothesis was that, in the adult brain,
increased glucose concentrations are associated with
proconvulsant effects, whereas decreased glucose concentrations are associated with anticonvulsant effects.
We chose streptozocin-induced diabetes23,24 as a model
of diabetic hyperglycemia. We initially elected to determine the effect of relatively short-term, 2-week diabetic hyperglycemia on seizure susceptibility in vivo.
We used short fasting for 24 hours and an injection of
glucose solution (as a model of hyperglycemia not confounded with streptozocin-induced metabolic and hormonal changes) as additional conditions. We also investigated in vivo whether diabetic hyperglycemia
From the 1Department of Neuroscience, Albert Einstein College
of Medicine, Bronx, NY; 2Brandeis University, Waltham, MA;
Department of Neurology, Albert Einstein College of Medicine;
and 4Einstein/Montefiore Epilepsy Management Center, Albert Einstein College of Medicine, Bronx, NY.
Address correspondence to Dr Velı́s̆ek, AECOM, K 314, 1410 Pelham Parkway South, Bronx, NY. E-mail:
Received Jun 10, 2002, and in revised form Sept 4. Accepted for
publication Sep 4, 2002.
© 2002 Wiley-Liss, Inc.
makes neurons in critical brain areas more prone to
seizure-induced damage. In the entorhinal cortexhippocampal slices from nontreated rats, we determined, in vitro, the effects of high glucose concentrations on epileptiform activity induced by low
extracellular [Mg2⫹]. Finally, we determined the effects
of high-glucose concentration on paired-pulse responses in the perforant path. Results clearly indicate
that in the adult male rats, high extracellular glucose
concentration is associated with proconvulsant effects,
whereas decreases in extracellular glucose may have anticonvulsant effects. Paired-pulse paradigm results suggest that the neuropeptide Y (NPY) neurotransmission
may be involved in these effects.
Materials and Methods
Experiments were conducted according to the Revised Guide
for the Care and Use of Laboratory Animals (NIH Guide,
Vol. 25, No. 28, August 16, 1996), and the procedures for
animal experimentation utilized were reviewed and approved
by the Institutional Animal Care and Use Committee. We
made every effort to use the minimum number of rats with
sufficient statistical power. Adult male Sprague-Dawley rats
from Taconic Farms (Germantown, NY) were used weighting 150 to 175gm.
In Vivo Experiments
Rats were randomly assigned to the following groups. (1)
Solvent controls (n ⫽ 12) were injected (intraperitoneally)
with two doses of 2ml/kg of 0.05N citrate buffer (pH 5.5)
on days 1 and 2 of the experiment (24 hours apart). (2)
Streptozocin rats (n ⫽ 22) were injected intraperitoneally
(IP) with two doses of 40mg/kg of streptozocin on days 1
and 2 of the experiment (2ml/kg; 24 hours apart; dissolved
in 0.05N citrate buffer) to produce diabetic hyperglycemia.
With sequential streptozocin administration in this dose
range, it is possible to obtain a relatively high survival rate
among the treated rats (⬎80%) with very predictable extracellular glucose concentrations.25 (3) Handled controls (n ⫽
10) were not injected but otherwise handled similar to solvent controls and streptozocin rats. (4) Fasting rats (n ⫽ 10)
were injected (IP) twice with 2ml/kg of 0.05N citrate buffer
on days 1 and 2 of the experiment (24 hours apart) and were
handled similar to solvent controls and streptozocin rats.
This group fasted for 24 hours before seizure testing and
glucose measurements with free access to tap water. (5) Glucose rats (n ⫽ 8) received 5ml of 20% glucose solution IP
30 minutes before seizure testing. Controls (n ⫽ 5) for this
group received 5ml of normal saline IP at the same pretreatment interval. In these two groups, glucose concentration
was determined before the injection (baseline) and after 30
minutes, just before seizure testing. The group of glucose injected rats provided nondiabetic hyperglycemia.
All rats were weighted five times per week. Seizure testing
was performed on day 15 of the experiments.
measurement of glucose concentrations was performed immediately before and after seizure testing (except for handled
Annals of Neurology
Vol 53
No 1
January 2003
controls). After snipping the tip of the rat’s tail, a small drop
of blood (⬃10␮l) was collected on a test stripe of a
Boehringer-Mannheim glucosemeter. The rat immediately
was placed into the flurothyl chamber and tested for flurothyl seizures. Glucose level reading (in mg per 100ml) was
available in 45 seconds and calculated in millimolars. As
soon as the rat displayed final seizure phase, that is, tonicclonic seizure, it was removed from the chamber, and another drop of blood was collected from the existing cut after
a gentle massage of the tail.
experiment was to determine correlation between cerebrospinal fluid (CSF) glucose concentration and glucose concentrations in the peripheral and central blood sample. Nine rats
were used after 24-hour fasting with free access to tap water.
Rats were introduced into deep urethane anesthesia and the
skull was opened using a 1mm-diameter precision drill.
Then, a needle of a Hamilton microsyringe (10␮l volume)
was stereotaxically inserted in the lateral ventricle (coordinates from bregma: anteroposterior, 0.8mm; lateral, 1.5mm;
depth, 3.8mm; incisor bar at ⫺3.5mm), and a 10␮l sample
of the CSF was withdrawn and immediately processed for
glucose measurement as described above. Once the measurement was finished, a drop of blood from the tip of the tail
was used to determine peripheral blood glucose concentration. The procedure was repeated (with the CSF withdrawn
from contralateral ventricle) and average value was used. We
obtained sufficient CSF samples in four rats. Finally, the rat
in deep anesthesia was decapitated, and the central blood
sample was analyzed for glucose.
Flurothyl is a gaseous convulsant,
which induces seizures by inhalation. In our experimental
setting, flurothyl was continuously (rate of 20␮l/min) delivered into an airtight chamber (9.34L) via a precise microinfusion pump (WPI, Sarasota, FL). The advantage of continuous flurothyl infusion is that seizures always occur. Two
seizure types develop after flurothyl exposure: clonic and
tonic-clonic seizures. Clonic seizures consist of clonic convulsions of head and forelimb muscles with preserved righting
reflex. Wild running followed by a loss of righting reflex is
indicative of tonic-clonic seizures26 consisting of a short,
tonic contraction of forelimb and hind limbs followed by
long, clonic convulsions of all four limbs. In adult rats,
clonic and tonic-clonic seizures are clearly separated, and,
usually, several episodes of clonic seizures precede the tonicclonic seizure. Because flurothyl was infused at a constant
rate, the latency to the onset of seizures allowed us to calculate the amount of infused flurothyl necessary to elicit seizures, that is, the flurothyl seizure threshold for clonic or
tonic-clonic seizures for our chamber size.27 The seizure
threshold therefore inversely reflects seizure susceptibility.
Thus, higher seizure threshold correlates with lower seizure
Two days after the seizures, rats under deep urethane anesthesia were perfused transcardially with 0.01M phosphatebuffered saline (pH 7.2; 4°C) followed by fixative (4% paraformaldehyde in 0.2 M phosphate buffer, pH 7.2; 4°C). The
brains were postfixed in the fixative for 48 hours at 4°C,
immersed in a cryoprotective solution of 20% sucrose, and
frozen at ⫺40°C in 2-methylbutane. Coronal sections
(40␮m) were cut using a cryostat. Neuronal damage in brain
areas susceptible to seizure-induced injury has been determined by silver stain and Fluoro-Jade B on adjacent sections.
We used a modified method of Gallyas and
colleagues.28 In brief, free floating sections were washed in
distilled water (3 ⫻ 5 minutes), pretreated with NaOH/
NH2N03 (1.2%), and incubated with silver impregnation solution (silver nitrate) for 10 minutes. After another washing,
sections were developed in a silicotungstate developer,
mounted on gelatine-coated slides, dehydrated in graded ethanols, and coverslipped.29
FLUORO-JADE B STAIN. Sections mounted on gelatinized
slides were immersed in alcohol with 1% NaOH, transferred
into 0.06% potassium permanganate, and then incubated
with Fluoro-Jade B (0.0004%), followed by washing in
0.1M phosphate-buffered saline, dried, soaked in xylene, and
coverslipped with Permount.30
MICROSCOPY. In the silver stain, we were searching for
“dark” neurons, which were evaluated according to the presence of silver deposits. We were searching in those brain areas, which have been demonstrated as sensitive to seizureinduced damage: thalamic nuclei, the hippocampus
including the dentate gyrus, the substantia nigra,31,32 and in
the olivary nuclei (which are sensitive to glycemia changes).33
In addition, we determined Fluoro-Jade B fluorescence using
an excitation filter at 488nm and an emission filter at 512nm
in these brain regions. This method may be more sensitive
than the silver stain.
DATA EVALUATION. Simple data (glucose concentrations)
were compared using one-way analysis of variance (ANOVA)
with post hoc Fisher protected least squares degree (PLSD)
test. Weight data before the beginning of the experiments,
and after 1 and 2 weeks, were compared using two-way
ANOVA for repeated measures (independent factor: groups;
within factor: days of experiment) with post hoc StudentNewman-Keuls test. Finally, flurothyl seizure thresholds were
correlated to glucose levels measured immediately before seizure testing by using logarithmic regression and to the
weight data using linear regression. Level of significance was
always preset to p value less than 0.05.
In Vitro Experiments
Euglycemic rats (n ⫽ 8) were killed under deep ether anesthesia, brains were removed, and horizontal entorhinal cortexhippocampal slices (400␮m thick) were cut at the level of
ventral hippocampus using a vibratome. Slices were placed in
an interface chamber. After 90 minutes of incubation in 33
to 34°C artificial CSF (ACSF; 2ml/min; composition in millimolars: NaCl 126; KCl 5; NaH2PO4 1.25; MgCl2 2;
CaCl2 2; NaHCO3 26; and glucose 10) oxygenated with
95% O2/5% CO2, pH 7.32 to 7.40, viability of slices was
tested in the cornu ammonis 1 area of the hippocampus.
Only slices responding to a stimulus with a single population
spike larger than 2mV were used. Note that, in these experiments, the baseline extracellular glucose concentration was
10mM, which is the common concentration used for in vitro
electrophysiology in brain slices.34 Osmolality of the solutions was determined using a microsmometer. After a stable
DC baseline recording was obtained in the medial entorhinal
cortex (MEC, layers IV–V), Mg2⫹-free ACSF was introduced. In the MEC, this treatment elicits long (1–7 minutes)
discharges with a negative DC shift (seizure-like events),
which eventually develop into a status of short recurrent discharges.34 At least 30 minutes after discrete seizure-like
events (see Fig 5a) developed into a status of short recurrent
discharges (see Fig 5b), the perfusion was switched to Mg2⫹free ACSF with a glucose concentration of 20mM. After 30
minutes, the perfusion was returned to Mg2⫹-free ACSF
with 10mM glucose, and the recording continued for at least
30 minutes. We evaluated the frequency of discharges as the
number of discharges per minute. The amplitude of discharges was expressed in microvolts. Baseline data were collected during the last minute before the perfusate was
switched from Mg2⫹-free ACSF with 10mM glucose to
Mg2⫹-free ACSF with altered glucose concentration. Treatment data were evaluated during the last minute before the
perfusate was switched Mg2⫹-free ACSF with 20mM glucose
concentration back to Mg2⫹-free ACSF with 10mM glucose.
Recovery data were recorded during a 1-minute period after
30 minutes of recovery in the Mg2⫹-free ACSF with 10mM
glucose. Data were further computed by repeated-measures
ANOVA (within factor: glucose concentration) with post
hoc Tukey test. Level of significance p value was preset to
PAIRED-PULSE PARADIGM. A additional nine rats were
used to determine the effects of 20mM glucose or a control
solution containing 10mM glucose and 10mM mannitol on
paired-pulse paradigm in the perforant path, which is the
output system of the entorhinal cortex. Brain slices were prepared as described above. After incubation, stimulation electrode was placed across the perforant path, and the recording
electrode was inserted in the layer of dentate gyrus granule
cells. Evoked population spikes were recorded in a response
to a pair of stimuli with interstimulus interval of 10 to 1,000
milliseconds.35 Always four responses were averaged for a
given interval. After the entire cycle was completed, perfusion was switched to either ACSF containing 20mM glucose
or 10mM glucose and 10mM mannitol (osmolarity control)
for at least 30 minutes, and the stimulation cycle was repeated. Thus, there were two groups of slices: high-glucose
and mannitol-treated, each with two cycles: control and after
the treatment. Responses to stimulation were expressed as
R2/R1*100. Differences between the groups were calculated
using Student’s t test. p value was always preset to 0.05.
In Vivo Experiments
Two weeks after the streptozocin injections, 14 of
22 rats displayed obvious diabetes with tail blood
glucose concentrations higher than 15.6mM. One
streptozocin-injected rat died during the latent period,
Schwechter et al: Extracellular Glucose and Seizures
and in one rat we were unable to establish a baseline
glucose concentration because the blood sample withdrawn before the seizure test was insufficient. In two
rats, streptozocin administration induced a prediabetic
state, with glucose concentrations of 7.22 and
8.11mM. These rats were excluded from further evaluation. In four rats, there was no diabetes after streptozocin injections (glucose concentrations in all four
rats were ⬍6.66mM; mean ⫾ SEM: 5.72 ⫾
0.37mM). Similarly, injections of solvent produced no
diabetes; glucose concentrations were 5.46 ⫾ 0.24mM.
These glucose concentrations were not significantly different from the handled, noninjected group (4.52 ⫾
0.22mM). Overall, there was no difference between
glucose concentrations in streptozocin-injected, nondiabetic rats (n ⫽ 4); solvent-injected rats (n ⫽ 12) and
handled controls (n ⫽ 10; Fig 1a). Conversely, 24hour fasting decreased the blood glucose concentration
to 2.81 ⫾ 0.21mM (n ⫽ 10). Therefore, glucose concentrations in both streptozocin-injected diabetic and
fasting groups significantly differed from all other
groups (ANOVA F[4,45] ⫽ 263.9 with post hoc
Fisher PLSD test; p ⬍ 0.05; see Fig 1a). Similar differences in glucose concentrations were measured immediately after flurothyl seizures. Streptozocin-injected
diabetic and fasting groups significantly differed from
all other groups (ANOVA F[4,44] ⫽ 200.3 with post
hoc Fisher PLSD test; p ⬍ 0.05; see Fig 1b). Finally,
when we determined weight gain as the difference in
body weight before drug administration (day 0 of experiment) and before seizure induction (day 14 of experiment; but before the onset of fasting), we found
that the streptozocin-injected diabetic group gained
significantly less weight than the other four groups
(ANOVA F[4,42] ⫽ 44.3 with post hoc Fisher PLSD
test; p ⬍ 0.05; see Fig 1c). Based on these results, for
further analysis we combined the three groups
(streptozocin-injected nondiabetic, solvent-treated, and
handled) with no baseline diabetes or hypoglycemia in
one group referred further to as nondiabetic controls.
Regression analysis showed that there is a significant
negative correlation between glucose concentration established before seizure testing and clonic flurothylinduced seizure thresholds (logarithmic regression
ANOVA F(1,47) ⫽ 13.066; p ⬍ 0.05; Fig 2a). Similarly, there was a significant negative correlation between glucose concentration and flurothyl-induced
tonic-clonic seizure threshold (logarithmic regression
ANOVA F(1,45) ⫽ 7.385; p ⬍ 0.05; see Fig 2b).
Thus, the higher the glucose concentration, the lower
the threshold for flurothyl-induced seizures, indicating
proconvulsant effects of high glucose concentrations in
flurothyl seizures. Note that if the data were grouped
(grouping was based on previously found difference in
glucose concentration: streptozocin-injected hyperglycemic, nondiabetic controls, and fasting groups) and
Annals of Neurology
Vol 53
No 1
January 2003
Fig 1. Blood glucose concentration measured immediately before flurothyl-induced seizures (a) and right after the seizures
(b). Body weight gain (c) during 15 days of experiments. Asterisks mark groups that are significantly different from all
other groups (ANOVA with post hoc Fisher PLSD test; p ⬍
0.05). Handled controls were not injected but otherwise handled similar to other groups. Solvent controls received two injections of vehicle. Streptozocin (STZ) nondiabetic rats were
injected with two doses of streptozocin; however, they did not
develop diabetes. Streptozocin diabetic rats developed full diabetes with blood glucose concentration greater than 15mM.
Fasting group was injected with vehicle and otherwise handled
similarly as other groups, however, was fasted for the last 24
hours before the seizure testing. Weights presented here for this
group were obtained before fasting has begun.
Fig 2. Correlation between blood glucose concentration and
seizure threshold for flurothyl-induced clonic (a) and tonicclonic seizures (b). For both clonic and tonic-clonic seizures
induced by flurothyl, there was a significant negative correlation between the two variables (independent: tail blood glucose
concentration; dependent: flurothyl seizure threshold) best described by logarithmic regression. All groups described in Figure 1 are included in this figure: handled controls, solvent
controls, STZ nondiabetic rats, STZ diabetic rats, and fasted
data were evaluated using ANOVA, there was a significant overall effect (ANOVA F[2,47] ⫽ 5.122, p ⬍
0.05; not illustrated) for clonic seizure threshold. Pairwise comparisons demonstrated that the threshold in
the streptozocin-injected hyperglycemic group was significantly lower than in the other two groups. For
tonic-clonic seizures, there was also an overall difference in seizure threshold (ANOVA POST 2,44) ⫽
5.486; p ⬍ 0.05; not illustrated). Pairwise comparisons
showed that in this case, fasting group had significantly
higher threshold than the two other groups. Because
streptozocin has additional metabolic and hormonal effects, which may result in altered seizure thresholds, we
used two additional groups of rats. One group was injected with 5ml of 20% glucose and the other group
with 5ml of normal saline 30 minutes before seizure
testing to determine effects of pure hyperglycemia on
seizure susceptibility. Measurements verified that these
two groups did not differ in blood glucose concentrations during baseline conditions (before injections; Fig
3a), but there was a significant hyperglycemia in glucose injected group 30 minutes after the injection, just
before seizure testing (see Fig 3a; Student’s t test; p ⬍
0.05). In flurothyl seizure test, there was a lower seizure threshold for clonic seizures in glucose-injected
rats (Fig 3b; Student’s t test; p ⬍ 0.05). However,
there was no difference in seizure threshold for
flurothyl-induced tonic-clonic seizures between the
groups (see Fig 3b).
Furthermore, we analyzed the difference in glucose
concentrations immediately before and after flurothylinduced seizures. Diabetic rats had a significantly lower
seizure-induced glucose concentration increase (1.06 ⫾
0.75mM) compared with nondiabetic controls (2.93 ⫾
0.27mM) and fasting rats (2.62 ⫾ 0.24mM; ANOVA
F[2,45] ⫽ 5.126, with post hoc Fisher PLSD test; p ⬍
0.05; Fig 4a). We analyzed the rats’ actual weights in
both groups on days 0, 8, and 14 of the experiment.
The data show that the rats randomly chosen for streptozocin injections, which then experienced hyperglycemia (diabetic-hyperglycemic group), were insignificantly heavier before the initial treatment but, during
the course of the experiment, gained significantly less
weight than nondiabetic controls and fasting rats (twoway repeated measures ANOVA; within factor: weight
of subjects; F[2,86] ⫽ 1270.69; p ⬍ 0.05; independent factor: groups; F[2,43] ⫽ 10.68; p ⬍ 0.05;
Student-Newman-Keuls post hoc test; see Fig 4b). For
illustration, Figure 4c demonstrates that the weight
gain determined as the difference between weight on
day 0 and day 15 (after fasting in the fasted group) of
experiments was smaller in diabetic rats (48.0 ⫾ 6.2g)
than in nondiabetic controls (106.3 ⫾ 2.4g) and fasting rats (82.1 ⫾ 3.1g; one-way ANOVA F(2,44) ⫽
65.624; with Fisher PLSD test; p ⬍ 0.05; see Fig 4c).
Because weights of the rats varied and flurothyl, as
ether-like substance, may have affinity for adipose tissue and saturation of this tissue may change seizure
threshold, we tested whether there is a correlation between the weight of the rats and seizure thresholds. No
Schwechter et al: Extracellular Glucose and Seizures
diabetic rats for neuronal damage. We did not find any
silver deposits in either the diabetic rats or nondiabetic
controls in any of the areas investigated (data not
shown). Furthermore, there was no Fluoro-Jade B fluorescence in any of the areas of interest in either
diabetic-hyperglycemic or control rat (hippocampi
shown in Fig 6a and b, respectively) as compared with
the hippocampus of a positive control for staining
(kainic acid-induced neuronal damage) shown in Figure 6c.
most parallel measurements of glucose concentration in
the tail blood (peripheral), mixed arteriovenous blood
from carotid arteries and jugular veins (central), and
CSF sample showed that the lowest glucose concentration was found in the peripheral blood and the highest
in the mixed central blood. CSF glucose concentration
lied always in between these two values (Table). Average glucose concentration in the CSF after 24 hour
starving was 5.56mM (n ⫽ 4).
Mg2⫹-free ACSF (osmolality 308.3 ⫾ 4.6mOsm) regularly induced long discharges (seizure-like events; Fig
7a), which continually developed into short recurrent
discharges (see Fig 7b) persisting as long as the Mg2⫹
in the ACSF stayed low. An introduction of Mg2⫹-free
ACSF containing 20mM glucose (osmolality 329.0 ⫾
4.2mOsm) significantly and reversibly increased the
amplitude of epileptiform discharges (repeated measures ANOVA F[2,23] ⫽ 6.525; with post hoc Tukey
test; p ⬍ 0.05; see Fig 7c, e, f) bud did not affect the
frequency of discharges (repeated measures ANOVA
F[2,23] ⫽ 3.070; p ⬎ 0.05). Reperfusion with Mg2⫹free ACSF with 10mM glucose reversed the amplitude
of discharges back to values similar to those recorded
during baseline conditions, which were significantly
different from values recorded in 20mM glucose ACSF
(Fig 7d, e, and f). Thus, these in vitro data indicate
that extracellular glucose concentrations increased to
20mM range positively affect susceptibility to low
Mg2⫹-induced epileptiform activity in the entorhinal
cortex slices from euglycemic rats.
Fig 3. Effects of acute glucose injection associated with hyperglycemia on susceptibility to flurothyl-induced seizures. (a)
Glucose concentration before (baseline) and after the injection
of 5ml of 20% glucose in glucose-treated rats and 5ml of normal saline in controls. There was no difference in baseline
glucose concentration. Thirty minutes after injection, blood
glucose concentrations significantly differed (Student’s t test, p
⬍ 0.05). (b) The difference in flurothyl clonic (left) and
tonic-clonic (right) seizure threshold between the glucosetreated and control group. Glucose-treated rats had significantly lower threshold for clonic seizures (Student’s t test, p ⬍
such correlation was found for either clonic or tonicclonic flurothyl seizures (Fig 5a, b; p value always
Finally, we inspected the brains in diabetic and non-
Annals of Neurology
Vol 53
No 1
January 2003
in the control cycles (ACSF osmolality ⫽ 294.6 ⫾
1.6mOsm) of both groups did not differ significantly
( p always ⬎⬎0.20), those data were pooled for clarity.
Control cycle of paired pulses had a typical three-phase
pattern; an early depression (interstimulus intervals,
10 –20 milliseconds), intermediate facilitation (intervals, 50 –100 milliseconds), and a late depression (intervals, 125–1,000 milliseconds). After the control cycle, the slices were perfused with the ACSF containing
between these two treatments. High (20mM) glucose
concentration enhanced paired responses at 70, 250,
and 300 milliseconds (Fig 8; p ⬍ 0.05; Student’s t test)
compared with the solution containing 10mM glucose
plus 10mM mannitol with equal osmolarity.
Results of this study are consistent with those clinical
findings showing that increased glucose concentrations
in adult diabetic patients may provoke seizures.1,3,36 In
the adult rats, alloxan-induced diabetic hyperglycemia
was associated with proconvulsant effects in the electroshock seizure model,37 and streptozocin-induced diabetic hyperglycemia had proconvulsant effects in the
electroshock and bicuculline seizure models in the
adult mice.38 Our data confirm these studies and suggest that hyperglycemia by itself may have proconvulsant effects (as shown for clonic seizures in our glucose
injection experiment). On the other hand, the increased susceptibility to tonic-clonic seizures in
streptozocin-injected rats may have been caused by
other metabolic effects of streptozocin39 because similar proconvulsant effects were not seen after an acute
glucose injection. However, note here that the level of
hyperglycemia induced by glucose injection was lower
(⬃14mM) than the diabetic hyperglycemia after streptozocin (⬎15mM), and this may account for the differential effects on tonic-clonic seizure threshold. On
the other side, fasting led to slightly decreased glucose
concentrations associated with anticonvulsant effects.
Caloric restriction diet associated with decreased blood
glucose concentrations have anticonvulsant effects in
the epileptic El mice40 as well as neuroprotective effects
in rats after kainic acid.17 However, in all cases of caloric restriction or carbohydrate poor diet, the possible
anticonvulsant effects of ketogenesis should be considered.41 Our in vitro experiments support findings of
proconvulsant effects of high glucose concentrations in
our in vivo data. Another in vitro study demonstrated
that the hippocampal slices from streptozocin-induced
diabetic rats are more prone to produce epileptiform
response to a stimulation than control slices.42
In clinical studies, focal (multifocal) seizures were ex-
20mM glucose (osmolality, 317.25 ⫾ 2.45mOsm) or
10mM glucose plus 10mM mannitol (osmolality,
317 ⫾ 0.0mOsm). We observed significant differences
Fig 4. Seizure-associated blood glucose difference (a), weight
gain before (b) and after (c) fasting for a combined group of
nondiabetic rats, streptozocin (STZ) diabetic rats, and fasting
hypoglycemic rats. Group of nondiabetic rats includes handled
controls, solvent controls, and STZ nondiabetic rats as described in Materials and Methods and in the legend to Figure
1. Asterisks indicate significant difference compared with any
other group (ANOVA with post hoc Fisher PLSD test in panels a and c and two-way ANOVA with post hoc StudentNewman-Keuls test in panel b). In addition, pound sign in
panel b indicates a significant difference for the within factor
(repeated weight measurements).
Schwechter et al: Extracellular Glucose and Seizures
Fig 5. Correlation between weight of the rats and seizure
threshold for flurothyl-induced clonic (a) and tonic-clonic seizures (b). For both clonic and tonic-clonic seizures induced by
flurothyl, there was no correlation (linear regression was used)
between the two variables (independent: weight of the rats;
dependent: flurothyl seizure threshold). All groups are included
in this figure, as described in Figure 1: handled controls, solvent controls, streptozocin (STZ) nondiabetic rats, STZ diabetic rats, and fasted rats.
Annals of Neurology
Vol 53
No 1
January 2003
Fig 6. Fluoro-Jade B staining for damaged neurons in the
streptozocin diabetic rat hippocampus (a), control rat hippocampus (b), and in the hippocampus of a positive control
(c). Cornu ammonis 1 (CA1) with long white arrows mark
the layer of pyramidal neurons in hippocampal area CA1;
dentate gyrus (DG) with long white arrows indicate the external blade of dentate gyrus granule cells. There were no
Fluoro-Jade B–positive cells in the hippocampus of
streptozocin-injected diabetic rats (a) or control rat hippocampus (b) after a single flurothyl-induced seizure. However, in
the positive control (c) produced by kainic acid–induced status
epilepticus in a euglycemic rat, there were many Fluoro-Jade
B–positive neurons (short white arrows) in the hilus of the
dentate gyrus, in the area of pyramidal neurons in the CA3c,
and in the bend of CA3. Scale bar ⫽ 200␮m.
Table. Correlation between Peripheral, Central, and CSF
Glucose Concentration
CSF Glucose
Peripheral blood sample was obtained from the tip of the tail, central blood sample was obtained after decapitation under deep anesthesia, and the CSF sample (10␮l) was withdrawn from the lateral
ventricle (attempts were successful in four rats). Samples were collected in euglycemic rats after 24-hour fasting with free access to tap
water. Note that, in all measured rats, CSF glucose concentration
was in between peripheral and central blood glucose concentration.
Fig 7. Model of low Mg2⫹-induced epileptiform activity in
the deep layers of the entorhinal cortex slices of the euglycemic
rat and the effects of 20mM extracellular glucose concentration. (a) Initial seizure-like discharges induced by perfusion
with Mg2⫹-free artificial cerebrospinal fluid (ACSF). (b) Development of short recurrent discharges in the Mg2⫹-free
ACSF (represents baseline condition, always with 10mM glucose). (c) Enhancement of short recurrent discharges by application of Mg2⫹-free ACSF containing 20mM glucose concentration (represents treatment condition). (d) Recovery of short
recurrent discharges after glucose in the ACSF was returned to
baseline 10mM concentration (represents recovery condition).
(e) Effect of 20mM glucose perfusion on the amplitude and
frequency (f) of the short recurrent discharges. Concentration of
20mM of glucose reversibly enhanced the amplitude (repeated
measures ANOVA with post hoc Dunn’s test). Asterisk indicates significant difference versus both baseline and recovery
conditions. Although the differences presented here appear
small, note that data were very consistent and were compared using a test for repeated measures; therefore, overall
and pairwise significance was found for these repeated measurements.
CSF ⫽ cerebrospinal fluid.
acerbated by increased glucose concentrations.1,3 Although flurothyl seizures used here are considered to be
primarily generalized,43 we cannot exclude a multifocal
origin in this model. Ictal phase of the low
magnesium-induced epileptiform activity is initiated in
the entorhinal cortex34 and thus has a focal origin. Previously mentioned studies in adult mice and rats37,38
investigating the effects of diabetic hyperglycemia used
an electroconvulsive shock, which is undoubtedly a
model of generalized seizures, and found similar effects
of glucose concentration on seizure susceptibility as we
did with flurothyl. Although we used two models of
epileptic activity, more studies using different models
of focal seizures are necessary.
In the clinical studies, hyperglycemia-exacerbated
seizures were of forebrain origin, and this corresponds
to our finding of the effects of hyperglycemia on
flurothyl-induced clonic seizure threshold. Clonic seizures are initiated in the forebrain.44 Similarly, in vitro
low magnesium-induced epileptiform activity used here
is initiated in the forebrain structures. However, tonicclonic seizures have their onset in the brainstem.44
Others found an effect of diabetic hyperglycemia on
this seizure type in electroshock and bicuculline models.37,38 Thus, additional metabolic effects in those
studies using diabetogenic drugs should be considered.
Our data further indicate that short-term diabetic hyperglycemia in combination with one seizure does not
induce neuronal damage in areas sensitive to seizureinduced impacts in the adult brain. Although we did not
perform neuronal counts, two methods for detecting
Schwechter et al: Extracellular Glucose and Seizures
Fig 8. Paired-pulse paradigm recorded in the layer of dentate
granule cells while stimulating perforant path. x axis: interstimulus interval in milliseconds. y axis: ratio of the second
(R2) to the first (R1) population spike in percentages. Thus,
values under 100% represent paired-pulse depression, whereas
the values above 100% represent paired-pulse potentiation.
Perfusion with 20mM glucose artificial cerebrospinal fluid
significantly increased responses at 70-, 250-, and 300millisecond interstimulus intervals compared with the osmolality control containing 10mM glucose and 10mM mannitol
(Student’s t test, p ⬍ 0.05).
neuronal injury (silver stain and Fluoro-Jade B28,30)
failed to show damaged neurons. Thus, there were no
morphological impacts of short-term streptozocininduced diabetic hyperglycemia and a single seizure on
This study also demonstrated that glucose concentrations used for in vitro electrophysiology are high compared with the CSF glucose concentration. Thus, electrophysiology experiments are commonly performed in
excess glucose. The difference between the average
(measured in vivo) CSF glucose concentration of 5.65
and 10mM commonly used for in vitro electrophysiology34 accounts for the difference in osmolarity of approximately 4mOsm (only 1.5% of total osmolarity).
This indicates that osmolarity change should not interfere with the experiment. However, as others19 –21 and
we demonstrated, glucose concentration may significantly affect synaptic transmission, and the question remains whether the glucose concentration in the in vitro
experiments should more closely follow experimentally
determined CSF glucose concentrations.
In addition, we confirmed observations of others
that the development of streptozocin-induced diabetes
is associated with either slower weight gain or possibly
with a weight loss.39 Therefore, weight gain may be
used as an indicator of success in streptozocin-induced
diabetic rat models.
Finally, our paired-pulse paradigm data indicate that
20mM glucose concentration may significantly affect
the efficacy of synaptic transmission at 70-, 250-, and
300-millisecond intervals. Our data35 as well as the
data of others45 show that the changes in NPY trans-
Annals of Neurology
Vol 53
No 1
January 2003
mission may alter paired-pulse paradigm in the intervals 200 to 300 milliseconds, whereas the shorter, 70millisecond interval corresponds to changes in the
GABAB neurotransmission.35 NPY is associated with
regulation of food intake and, thus, glucose levels.46
NPY-containing neurons respond rapidly to increased
glucose levels by NPY production.47 There are NPYcontaining neurons in the hippocampal hilus, which
may affect the paired-pulse paradigm recorded in the
dentate granule cell layer.48,49 Therefore, it appears
possible that the regulatory function of glucose concentration on seizure susceptibility that we observed in this
study may be mediated via NPY neurotransmission, at
least in the hippocampus.
Supported in part by Heffer Family Medical Foundation grant, by
the grants from the NIH (NS-20253, NS-36238 and NS/HD41366), by the CURE Foundation, and the Albert Einstein College
of Medicine SURP.
1. Berkovic SF, Johns JA, Bladin PF. Focal seizures and systemic
metabolic disorders. Aust N Z J Med 1982;12:620 – 623.
2. Venna N, Sabin TD. Tonic focal seizures in nonketotic hyperglycemia of diabetes mellitus. Arch Neurol 1981;38:512–514.
3. Harden CL, Rosenbaum DH, Daras M. Hyperglycemia presenting with occipital seizures. Epilepsia 1991;32:215–220.
4. Davis EA, Keating B, Byrne GC, et al. Hypoglycemia: incidence and clinical predictors in a large population- based sample of children and adolescents with IDDM. Diabetes Care
5. Keene DL, Metrakos K, Belmonte M, et al. EEG abnormalities
and convulsions in juvenile diabetes mellitus. Can J Neurol Sci
1983;10:198 –199.
6. Shehadeh N, Kassem J, Tchaban I, et al. High incidence of
hypoglycemic episodes with neurologic manifestations in children with insulin dependent diabetes mellitus. J Pediatr Endocrinol Metab 1998;11(suppl 1):183–187.
7. Hypoglycemia in the Diabetes Control and Complications
Trial. The Diabetes Control and Complications Trial Research
Group. Diabetes 1997;46:271–286.
8. Li C, Li PA, He QP, et al. Effects of streptozotocin-induced
hyperglycemia on brain damage following transient ischemia.
Neurobiol Dis 1998;5:117–128.
9. Li PA, Shamloo M, Katsura K, et al. Critical values for plasma
glucose in aggravating ischaemic brain damage: correlation to
extracellular pH. Neurobiol Dis 1995;2:97–108.
10. Warner DS, Gionet TX, Todd MM, McAllister AM. Insulininduced normoglycemia improves ischemic outcome in hyperglycemic rats. Stroke 1992;23:1775–1780; discussion, 1781.
11. Li PA, He QP, Csiszar K, Siesjo BK. Does long-term glucose
infusion reduce brain damage after transient cerebral ischemia?
Brain Res 2001;912:203–205.
12. Schurr A. Glucose and the ischemic brain: a sour grape or a
sweet treat? Curr Opin Clin Nutr Metab Care 2001;4:
13. Schurr A, Payne RS, Miller JJ, Tseng MT. Preischemic
hyperglycemia-aggravated damage: evidence that lactate utilization is beneficial and glucose-induced corticosterone release is
detrimental. J Neurosci Res 2001;66:782–789.
14. Marie C, Bralet AM, Gueldry S, Bralet J. Fasting prior to transient cerebral ischemia reduces delayed neuronal necrosis.
Metab Brain Dis 1990;5:65–75.
15. Lin TN, Te J, Huang HC, et al. Prolongation and enhancement of postischemic c-fos expression after fasting. Stroke
1997;28:412– 418.
16. Go KG, Prenen GH, Korf J. Protective effect of fasting upon
cerebral hypoxic-ischemic injury. Metab Brain Dis 1988;3:
17. Bruce-Keller AJ, Umberger G, McFall R, Mattson MP. Food
restriction reduces brain damage and improves behavioral outcome following excitotoxic and metabolic insults. Ann Neurol
1999;45:8 –15.
18. Chavko M, Braisted JC, Harabin AL. Attenuation of brain hyperbaric oxygen toxicity by fasting is not related to ketosis. Undersea Hyperb Med 1999;26:99 –103.
19. Fan P, O’Regan PA, Szerb JC. Effect of low glucose concentration on synaptic transmission in the rat hippocampal slice.
Brain Res Bull 1988;21:741–747.
20. Tekkok S, Kriz J, Padjen AL, Krnjevic K. Higher sensitivity of
CA1 synapses to aglycemia in streptozotocin-diabetic rats is agedependent. Brain Res 1998;813:268 –273.
21. Shoji S. Glucose regulation of synaptic transmission in the dorsolateral septal nucleus of the rat. Synapse 1992;12:322–332.
22. Madl JE, Royer SM. Glutamate in synaptic terminals is reduced
by lack of glucose but not hypoxia in rat hippocampal slices.
Neuroscience 1999;94:417– 430.
23. Courteix C, Bourget P, Caussade F, et al. Is the reduced efficacy of morphine in diabetic rats caused by alterations of opiate
receptors or of morphine pharmacokinetics? J Pharmacol Exp
Ther 1998;285:63–70.
24. Vallon V, Albinus M, Blach D. Effects of KAPT channel blocker
U37883A on renal function in experimental diabetes mellitus
in rats. J Pharmacol Exp Ther 1998;286:1215–1221.
25. Havel PJ, Uriu-Hare JY, Liu T, et al. Marked and rapid decreases of circulating leptin in streptozotocin diabetic rats: reversal by insulin. Am J Physiol 1998;274:R1482–R1491.
26. Velı́s̆ková J, Velı́s̆ek L, Nunes M, Moshé S. Developmental regulation of regional functionality of substantia nigra GABAA receptors involved in seizures. Eur J Pharmacol 1996;309:
27. Lánský P, Velı́s̆ková J, Velı́s̆ek L. An indirect method for absorption rate estimation: flurothyl-induced seizures. Bull Math
Biol 1997;59:569 –579.
28. Gallyas F, Wolff JR, Bottcher H, Zaborszky L. A reliable and
sensitive method to localize terminal degeneration and lysosomes in the central nervous system. Stain Technol 1980;55:
299 –306.
29. Velı́s̆ková J, Velı́s̆ek L, Galanopoulou AS, Sperber EF. Neuroprotective effects of estrogens on hippocampal cells in adult female rats after status epilepticus. Epilepsia 2000;41:S30S35.
30. Schmued LC, Hopkins KJ. Fluoro-Jade B: a high affinity fluorescent marker for the localization of neuronal degeneration.
Brain Res 2000;874:123–130.
31. Sutula T, Cavazos J, Golarai G. Alteration of long-lasting structural and functional effects of kainic acid in the hippocampus
by brief treatment with phenobarbital. J Neurosci 1992;12:
4173– 4187.
32. Okada R, Nagishi N, Nagaya H. The role of the nigrotegmental GABAergic pathway in the propagation of pentylenetetrazol
induced seizures. Brain Res 1989;480:383–387.
33. Kadekaro M, Savaki H, Sokoloff L. Metabolic mapping of neural pathways involved in gastrosecretory response to insulin hypoglycaemia in the rat. J Physiol 1980;300:393– 407.
34. Dreier JP, Heinemann U. Regional and time dependent variations of low Mg2⫹ induced epileptiform activity in rat temporal cortex slices. Exp Brain Res 1991;87:581–596.
35. Velı́s̆ek L, Velı́s̆ková J. Estrogen treatment protects GABA(B)
inhibition in the dentate gyrus of female rats after kainic acidinduced status epilepticus. Epilepsia 2002;43(suppl 5):
146 –151.
36. Bush GH, Steward DJ. Can persistent cerebral damage be
caused by hyperglycaemia? Paediatr Anaesth 1995;5:385–387.
37. Koltai M, Minker E. Changes of electro-shock seizure threshold
in alloxan diabetic rats. Experientia 1975;31:1369.
38. Tutka P, Sawiniec J, Kleinrok Z. Experimental diabetes sensitizes mice to electrical- and bicuculline-induced convulsions.
Pol J Pharmacol 1998;50:92–93.
39. Shafrir E. Diabetes in animals: contribution to the understanding of diabetes by study of its etiopathology in animal models.
In: Porte DJ, Sherwin RS, eds. Ellenberg and Rifkin’s Diabetes
Mellitus; Theory and Practice. New York: Appleton & Lange,
40. Greene AE, Todorova MT, McGowan R, Seyfried TN. Caloric
restriction inhibits seizure susceptibility in epileptic EL mice by
reducing blood glucose. Epilepsia 2001;42:1371–1378.
41. Freeman JM, Vining EP. Seizures decrease rapidly after fasting:
preliminary studies of the ketogenic diet. Arch Pediatr Adolesc
Med 1999;153:946 –949.
42. Margineanu DG, Niespodziany I, Wulfert E. Hippocampal
slices from long-term streptozotocin-injected rats are prone to
epileptiform responses. Neurosci Lett 1998;252:183–186.
43. Truitt EB, Ebesberg EM, Ling ASG. Measurement of brain excitability by use of hexaflurodiethyl ether (Indoclon). J Pharm
Exp Ther 1960;129:445– 453.
44. Browning RA. Role of the brain-stem reticular formation in
tonic-clonic seizures: lesion and pharmacological studies. Fed
Proc 1985;44:2425–2431.
45. Penix LP, Wasterlain CG. Selective protection of neuropeptide
containing dentate hilar interneurons by non-NMDA receptor
blockade in an animal model of status epilepticus. Brain Res
1994;644:19 –24.
46. Levin BE. Glucosensing neurons do more than just sense glucose.
Int J Obes Relat Metab Disord 2001;25(suppl 5):S68 –S72.
47. Wang J, Dourmashkin JT, Yun R, Leibowitz SF. Rapid
changes in hypothalamic neuropeptide Y produced by
carbohydrate-rich meals that enhance corticosterone and glucose levels. Brain Res 1999;848:124 –136.
48. Sloviter RS, Ali-Akbarian L, Horvath KD, Menkens KA. Substance P receptor expression by inhibitory interneurons of the
rat hippocampus: enhanced detection using improved immunocytochemical methods for the preservation and colocalization of
GABA and other neuronal markers. J Comp Neurol 2001;430:
49. Freund TF, Buzsaki G. Interneurons of the hippocampus. Hippocampus 1996;6:347– 470.
Schwechter et al: Extracellular Glucose and Seizures
Без категории
Размер файла
258 Кб
adults, correlation, extracellular, seizure, rats, susceptibility, glucose
Пожаловаться на содержимое документа