RAT MODEL OF INFLUENZA-ASSOCIATED ENCEPHALOPATHY (IAE):

STUDIES OF ELECTROENCEPHALOGRAM (EEG) IN VIVO

Y. CISS,
a
S. WANG,
a
I. INOUE
b
AND H. KIDO
a
*
a
Division of Enzyme Chemistry, Institute for Enzyme Research, The
University of Tokushima, Tokushima 7708503, Japan
b
Division of Molecular Neurobiology, Institute for Enzyme Research,
The University of Tokushima, Tokushima 7708503, Japan
AbstractInuenza-associated encephalopathy (IAE) is ch-
aracterized by severe neurological complications during
high-grade fever with high morbidity and mortality in chil-
dren. The major neurological complications during high-
grade fever include convulsive seizures, loss of conscious-
ness, neuropsychiatric behavior (hallucination, meaningless
speech, disorientation, laughing alone); high voltage ampli-
tude slow waves and the occurrence of theta oscillation are
depicted on the electroencephalogram (EEG) in the IAE pa-
tients. At the early phase of the disease, the cytokines levels
increase in severe cases. To understand the neuronal prop-
erties in the CNS leading to these neurological complications
in IAE patients, we recorded EEG signals from the hippocam-
pus and cortex of rats infected with inuenza A/WSN/33 H1N1
virus (IAV) strain. Abnormal EEG activities were observed in
all infected rats under anesthesia, including high voltage
EEG burst amplitude and increased EEG spikes in the early
phase (8 hday 2) of infection, and these increases at the
early phase were in parallel with a signicant increase level of
interleukin-6 (IL-6) in the serum. When the infected rats were
heat-stressed by elevating the rat body core temperature to
3941 C, these abnormal EEG activities were enhanced, and
the oscillation pattern shifted in most of rats fromslow burst-
ing waves (<1 Hz) to theta oscillation (36 Hz). These results
indicate that the abnormal EEG activities in IAE patients
could be well reproduced in anesthetized IAV infected rats
under hyperthermia, hence this animal model will be useful
for further understandings the mechanism of neuronal com-
plications in IAE patient during high-grade fever. 2010
IBRO. Published by Elsevier Ltd. All rights reserved.
Key words: EEG, inuenza, hyperthermia, cytokines, theta
activity, hippocampus.
Inuenza-associated encephalopathy (IAE) is a complex
infectious disease which lead to serious neurological com-
plications (Kasai et al., 2000; Morishita et al., 2002) and
the levels of cytokines increase in the early phase of the
disease in some severe cases, particularly in pediatric
inuenza (Aiba et al., 2001; Fukumoto et al., 2007). The
rates of mortality and neurological complications are par-
ticularly high in children under 5 years (Cox and Subbarao,
1999; Morishita et al., 2002). The clinical features of the
neurological complications at high-grade fever include loss
of consciousness, coma, convulsive seizures, brain
edema, and multiple organs failure. In addition, some neu-
ropsychiatric behaviors (e.g., hallucination, meaningless
speech, periodic crying, and laughing alone) and at the
electroencephalogram (EEG) level, high voltage slow am-
plitude oscillation and the occurrence of theta activity are
often observed at the early phase of the disease in IAE
patients (Okumura et al., 2005; Fukumoto et al., 2007).
Although there have been many reports on IAE in the last
decade (Fujimoto et al., 1998; Mori et al., 1999; Togashi et
al., 2000; Chen et al., 2005), there is no experimental
electrophysiological model of IAE during hyperthermia.
Moreover, the EEG activities in animals infected with inu-
enza A/WSN/33 H1N1 virus (IAV) during hyperthermia
condition have not been reported so far.
To design an animal model of IAE, we examined the
effects of IAV infection on EEG activities in the hippocam-
pus and cortex of rats using EEG recording in vivo. The
hippocampus and the cortex were targeted for the present
model because these two structures are among the most
affected areas in the brain after IAV. For example, after
IAV infection in rats, the cellular proteases, such as pan-
creatic ectopic trypsin and matrix metalloproteinase 9
(MMP-9), are markedly upregulated in the cortex, but
mainly in the hippocampus (Le et al., 2006; Kido et al.,
2007, 2008) and theta oscillation, which is one of the major
oscillation patterns observed in IAE patients during high-
grade fever is most prominent in the hippocampus and it
occurs in the cortex as well (reviewed in Stan Leung,
1998).
In order to make the model compatible with IAE pa-
tients physiological behavior, the experiment required to
create hyperthermia condition. This is because rats did not
manifest spontaneous hyperthermia after IAV infection.
Hence rats were anesthetized and exposed to heat stress
by raising the core body temperature to a level similar to
human fever. Under this hyperthermia condition, the typi-
cal abnormal EEG activities similar to those observed in
IAE patients during high-grade fever were observed in the
infected rats. These abnormal EEG activities were in par-
allel with a signicant increases level of interleukin-6 (IL-6)
in the serum. It is known that infection of various types of
virus including IAV induces overproduction of proinam-
mation cytokines such as tumor necrosis factor alpha
(TNF-), IL-6 (Aiba et al., 2001; Fukumoto et al., 2007).
This is the rst electrophysiological analysis of brain
activity in animals infected with IAV demonstrating hyper-
*Corresponding author. Tel: 81-88-633-7425; fax: 81-88-633-7423.
E-mail address: kido@ier.tokushima-u.ac.jp (H. Kido).
Abbreviations: EEG, electroencephalogram; FFT, fast Fourier trans-
formation; IAE, Inuenza-associated encephalopathy; IAV, inuenza
A/WSN/33 H1N1 virus; IL-6, interleukin-6; MMP-9, matrix metallopro-
teinase 9; NMDA, N-methyl-D-aspartate; PCR, polymerase chain re-
action; TNF-, tumor necrosis factor alpha; 2D, two-dimension.
Neuroscience 165 (2010) 11271137
0306-4522/10 $ - see front matter 2010 IBRO. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.neuroscience.2009.10.062
1127
thermia-induced high voltage EEG burst amplitude slow
waves and theta oscillation, which are the major brain
activity in human during high-grade fever in IAE patients
(Okumura et al., 2005; Fukumoto et al., 2007).
EXPERIMENTAL PROCEDURES
Animals, virus infection and electrophysiology
This study was performed in accordance with the Guidelines for
Animal Care and Use approved by the animal care committee of
The University of Tokushima. We used male Wistar rats weighing
(80150 g) aged 34 weeks (P25P33) were used in the present
study. For virus infection, rats were anesthetized with ketamine
xylazine (62.612.4 mg/kg), then A/WSN/33 strain (110
5
plaque
forming unit in PBS, 60 L) was intranasally administrated. To
prevent air infection between the uninfected and infected rats, the
infected and uninfected rats were kept in two different rooms. EEG
activities in infected and uninfected rats were not recorded in the
same days and separate recording electrodes were used for
infected rats and uninfected ones. The rooms were designed in
accordance with the Guidelines for Animal Care and Use of the
University of Tokushima.
Anesthesia and hyperthermia
Rats were anesthetized by low-dose of ether, and tted to a
stereotaxic frame (model SN6N, Narishige, Tokyo, Japan). Then
the animals were gas-anesthetized with 1.51.7% isourane
mixed with 30% O
2
and 70% N
2
allowing spontaneous respiration.
In some rats (n5), ketamine (62.6 mg/kg) and xylazine (12.4
mg/kg) were used. The rat core body temperature was measured
before and after infection using digital rectal thermometer. During
the EEG recording, the body temperature was gradually increased
from 3741 C using a heating pad system (model 21051-00, Fine
Science Tools Inc., CA, USA), with a feed-back control probe
inserted rectally. The rats were covered with a wool sheet (15
cm20 cm) to reduce heat loss. The EEG recording for an ex-
periment lasted for a period of 2035 min. The slow and very
pronounced EEG waves were continuously monitored during the
whole experiment to ensure that the animal was well anesthetized
and painless.
EEG recordings
To record the EEG, enamel-coated tungsten wire electrodes with
uncoated diameter of 120 m (M.T. Giken Co, Tokyo, Japan)
were used. Craniotomy was performed without damaging the
underlying dura using a standard miniature drill equipped with a
0.5 mm diameter drill bit. The electrodes were inserted based on
George Paxinos and Charles Watson The rat brain stereotaxic
coordinates (Paxinos and Watson, 2005). To insert the EEG in the
hippocampal CA1CA3 area (left hemisphere), an electrode was
positioned 2.83.0 mm posterior to the bregma, 2.72.9 mm
lateral from the midline, 2.63.0 mm below the dura. To place the
electrode in the cortical area (right hemisphere), was positioned
2.03.0 mm anterior from the bregma, 2.03.5 mm lateral from
the midline and 0.50.8 mm below the dura. Signals were re-
corded using a dual microprobe system (WPI Instruments, New
Haven, CN, USA), and a home-made amplier (1000). The
baseline was adjusted to zero-level with a slow voltage clamp
system with a time constant of 2.2 s. The signal was low-pass
ltered at 0.53 kHz, sampled at 1 kHz and recorded using
Axopatch software (Axon Instruments, Palo Alto, CA, USA). To
verify the electrode position, the electrode tip was coated with a
lipophilic tracer dissolved in dimethylsulfoxide at a concentration
of 12.5 mg/mL, before insertion into the brain. After removal of
the electrodes, rats were anesthetized with ketaminexylazine
(62.612.4 mg/kg) and were transcardially perfused with saline,
followed by xation (4% paraformaldehyde). The brain was re-
moved and immediately put in sucrose and kept in a 4 C room.
Then 800 m thick sections were prepared. Red traces left by the
electrodes with the dye were observed under the microscope and
photographed.
Four rats were used for chronically implantation EEG record-
ings under non-anesthetized condition. After insertion of elec-
trodes into the brain, two screws were placed on the two hemi-
spheres and the skull was covered with dental cement and a bolt
in the cement to allow no painful xation of the rats head in the
stereotaxic frame. After surgery, the rats were kept individually for
23 days for recovery. Then, the EEG signals were recorded for
12 days before infection as control. Then, rats were infected and
EEG activities were recorded every day from 4 h to day 6 of post
infection time without heat-stressed.
Characterization of abnormal EEG signals
The EEG amplitudes, spikes, lower amplitude uctuation during
the burst suppressed periods, and the oscillation patterns were
used to characterise abnormal events in the control (uninfected)
and infected animals. The EEG burst amplitude was measured
from the baseline (l50 ms) to the peak of the burst as noted by
(ab) in Fig. 2A and the average (n20 EEG bursts) was used to
make an averaged single point. The spike was identied as a
sharp wave. It usually sprouts randomly within the burst and
during the burst suppressed periods. To distinguish the spike from
other waves, we carefully examined the wave patterns to nd the
spike has the characteristics of 1d15 ms and 1k35 ms (see
left, Fig. 2A).
The lower amplitude uctuation was characterised by mea-
suring the sweep amplitude (h) from the peak positive to the peak
negative and 15 sweeps were taken to obtain an averaged single
point (Fig. 3B, inset).
Real time-polymerase chain reaction (PCR) analysis
The total RNA was isolated from the hippocampus and cortex of
infected (n12) and uninfected (n3) rats using an RNeasy Mini
kit (Qiagen, Valencia, CA, USA) according to the protocol supplied
by the manufacturer and then reverse transcribed using Oligo
primers and SuperScript III RT (Gibro BRL, Rockville, MD, USA)
for cDNA synthesis. To determine the virus titre, forward (5=-
CAGCACTCTCGGTCTGGACAT-3=, nucleotide 167187), and
reverse primers (5=-TCCTTCAGAATCCGCTCCACTA-3=, nucleo-
tide 228238) were selected from the inuenza A virus NS1 gene
segment. PCR amplication consisted of 40 cycles of 15 s for
denaturation at 95 C and 1 min for extension at 58 C. To
normalize the quantities, mouse -actin was used for the control of
RNA amplication. Quantication of gene expression was per-
formed using Fast Start SYBR Green Master (Roche Diagnostics,
Tokyo, Japan) with an ABI Prism 7300 system (Applied Biosys-
tems, Foster City, CA, USA). The reactions were characterized by
the point at which the PCR product was rst detected (the thresh-
old cycle) instead of the amount of PCR product accumulating
after a xed number of cycles.
Measurement of serum levels of cytokines
The serum levels of IL-6 and TNF- were measured using 15 rats,
control (n3), and 8 h (n3), day 2 (n3), day 4 (n3), day 6
(n3) post infection. The EEG signals were rst recorded, then
blood was collected from the heart, and the plasma was separated
by centrifugation at 2000g for 15 min. The serum was collected
and stored. The concentration of serum IL-6 and TNF- were
measured according to the protocol of the immunoassay kit (In-
vitrogen Corporation, CA, USA).
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1128
Statistics
Numerical values are expressed as meanSD. P-value was ob-
tained by students paired t-test, and P0.05 was considered
statistically signicant.
RESULTS
EEG signals in chronically implanted rats
To mimic IAE patients EEGrecordings conditions, we used
chronically implanted non-anesthetize rats and tried to
create hyperthermia condition. Inuenza virus did not in-
duce spontaneous hyperthermia in the IAV infected rats.
To record EEG activities, the experiment requires stable
recordings under hyperthermia. When heat stress was
applied, the rats responded violently moving the body con-
tinuously during the heat application. Therefore the record-
ing could not be realized under head-stressed condition.
We could successfully record only without the heat stress
at the normal body temperature in three out of four rats
examined. In this condition, the recorded EEG activities
were similar to that observed in normal non-anesthetized
animal including small amplitude uctuation about 0.2 mV,
irregular activities and theta oscillation dominantly at 36
Hz. After IAV infection, the theta component increased in
30 20 10 0
EEG - Cortex
EEG - Hippocampus
1
2
1
2
0
5
Frequency (Hz)
Control
4 h
8 h
Day 2
Day 4
Day 6
P
o
s
t

i
n
f
e
c
t
i
o
n

t
i
m
e
Control
4 h
8 h
Day 2
Day 4
Day 6
P
o
s
t

i
n
f
e
c
t
i
o
n

t
i
m
e
2D - FFT
Cortex
Hippocampus
A
B C
Control p.i.: 4 h
p.i.: 8 h
p.i.: day 2
p.i.: day 4
p.i.: day 6
Cortex
Hippocampus
30 20 10 0
Frequency (Hz)
0
5
0
.
2

m
V
1 s
5 s
1

m
V
500 ms
1

m
V
200 ms
a
b
a
b
1

m
V
Fig. 1. Electroencephalogram (EEG) activity in chronically implanted non-anesthetize rats. (A) EEG activity in the cortex and hippocampus before
infection (control) and after infection at different post infection time (4 h 8 h, day 2, day 4, day 6). A section marked by the horizontal (bar and letter
ab) on the EEG traces are expanded (right). Note no striking difference between control and after infection. (B) The two-dimension fast fourier
transformation (2D-FFT) displayed the frequency distribution for a period of 20 s of recording. Note an activation of theta oscillation in the hippocampus
compared to the cortex at the post infection (4 h 8 h, day 2, day 4). (C) Paroxysmal activity after infection at day 4 in both the cortex and hippocampus.
Numbers (1 and 2) are expanded at the bottom. Note the presence of interictal activity and large burst in expanded traces. The 2D-FFT is calculated
for a period of 20 s EEG activity.
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1129
the hippocampus compared to before infection. The two-
dimension fast fourier transformation (2D-fast Fourier
transformation (FFT)) displayed the increased EEG activ-
ities at theta frequency (36 Hz) after infection in all rats
(Fig. 1A). Paroxysmal activity was also depicted at day 4 of
post infection in one out of the three infected rats (Fig. 1C).
Hence, some variability in individual infected rats may
exist. These ndings were different from those of IAE
patients under high-grade fever. Hence, the application
of the heat stress was necessary. To overcome this
technical limitation, the experiments required the use of
anesthesia in order to establish a stable non-moving
condition during the rats EEG recordings. Therefore, in
the following experiments, the EEG signals were re-
corded under anesthesia during normal and heat stress
conditions.
EEG signals in uninfected rats
For control experiments, we recorded EEG simultaneously
from the hippocampus and cortex of uninfected rats under
normal body temperature at 37 C and under heat stress
conditions at 3941 C. The EEG of rats at 37 C was
characterized by slow bursting waves (1 Hz) detected by
the FFT as shown in Fig. 2B and small amplitude oscilla-
tion (414 Hz) during the burst suppressed periods in the
ltered traces (Fig. 2A). Neuronal activities in animals dur-
ing sleep were described by Steriade et al. (2001) and
Timofeev et al. (2001). Slow EEG oscillation was de-
scribed in animals by Steriade et al. (1993a,b,c) and in
humans by Achermann and Borbly (1997); Watts and
Herrick (1999) and Rojas et al. (2006) demonstrated the
presence of slow oscillation under isourane anesthesia.
The low amplitude uctuation during the burst suppressed
l
a
3
b
d
k b
EEG burst Spike
F
i
l
t
e
r
e
d

(
4

-

1
4

H
z
)
Cortex
Hippocampus
A
41 37
E
E
G

b
u
r
s
t

a
m
p
l
i
t
u
d
e









(
m
V
)
N
u
m
b
e
r

o
f

s
p
i
k
e
s

/

m
i
n
Temperature ( C)
o
1.0
0.8
0.6
0.4
0.2
0.0
12
8
4
0
39
Hippocampus
Cortex
B
FFT
C
o
37 C 41 C
o
Cortex
Hippocampus
12 8 4 0

12 8 4 0
Frequency (Hz) Frequency (Hz)
Cortex
Hippocampus
37 C 41 C
o o
0
.
2

m
V
500 ms
2 s
1

m
V
Fig. 2. EEG recordings in control rat with or without heat stress. (A) EEG recorded simultaneously in the cortex and hippocampus. Left, top, EEG at
37 C displayed slow oscillation activity (1 Hz). A section of the top traces is ltered at 414 Hz. Right, top, EEG recorded at 41 C is characterized
by slow oscillation and increased burst frequency. EEG burst is indicated by the arrowarc and spike by arrowhead. Bottom, characterization of EEG
burst amplitude (left) and spike (right). (B) FFT plot shows EEG activity at 37 (left) and 41 C (right). Note a predominant slow oscillation peaks
indicated by the arrows. (C) Effect of hyperthermia on EEG burst amplitude and spikes. Note the constant EEG burst amplitude level.
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1130
periods recorded from the hippocampus was larger and
clear than that from the cortex (Fig. 2A). Hyperthermia had
no remarkable effects as depicted on the EEG signals (Fig.
2A). EEG burst amplitude or number of spikes in both the
cortex and hippocampus were almost the same as those at
normal body temperature (Fig. 2C).
Inuenza virus infection and hyperthermia induce
abnormal EEG signals
We then recorded the EEG signals from infected rats. We
analyzed the EEG burst amplitude and EEG spikes of
infected rats at different post infection time, 8 h (n5), day
1 (n4), day 2 (n6), day 3 (n5), day 4 (n4), day 5
(n3) and day 6 (n4). The abnormal activities were
observed from 8 h which were gradually decreased to day
6 of post infection.
Fig. 3A shows EEG activities simultaneously recorded
from the hippocampus and the cortex at 8 h and day 2
postinfection under hyperthermia (40 C), which displayed
an enhanced low amplitude uctuation (0.10.5 mV) at
414 Hz during the burst suppressed periods, high voltage
EEG burst amplitude and EEG spikes compared to unin-
fected rats. These abnormal activities were much pro-
nounced in the hippocampus compared to the cortex (Fig.
3A). The statistical analysis revealed that the low ampli-
tude uctuation was markedly enhanced in the hip-
pocampus of infected rats compared to the cortex of
infected rats at 8 h and day 2 of post infection (Fig. 3B).
The EEG burst amplitudes and the EEG spikes were
also markedly enhanced in the infected rats compared
to the uninfected. The enhancement in the hippocampus
(Fig. 3C, left) was slightly higher than in the cortex (Fig.
A
40 C
o 1
1
EEG - cortex
EEG - hippocampus p.i: 8 h p.i: day 2
0
.
2

m
V
200 ms
2 s
1

m
V
0.5
0.4
0.3
0.2
0.1
0.0
h
L
o
w

a
m
p
l
i
t
u
d
e

f
l
u
c
t
u
a
t
i
o
n

(
m
V
)
Uninfected 8 h
day 2
post infection time
Hippocampus
Cortex
Uninfected Infected (8 h) Infected (day 2)
37 39 40 41 37 39 40 41
Temperature ( C)
o
Temperature ( C)
o
E
E
G

b
u
r
s
t

a
m
p
l
i
t
u
d
e












(
m
V
)
N
u
m
b
e
r

o
f

s
p
i
k
e
s

/

m
i
n
E
E
G

b
u
r
s
t

a
m
p
l
i
t
u
d
e












(
m
V
)
N
u
m
b
e
r

o
f

s
p
i
k
e
s

/

m
i
n
Low amplitude fluctuation
B C
50
40
30
20
10
0
2.0
1.5
1.0
0.5
0
50
40
30
20
10
0
2.0
1.5
1.0
0.5
0
***
**
*** **
**
**
*** **
*
*
*
**
***
***
**
**
Hippocampus
Cortex
Fig. 3. Hyperthermia induces enhanced abnormal EEG activities at 8 h and day 2 of post infection in simultaneously recordings from hippocampus
and cortex in inuenza A/WSN/33 H1N1 virus (IAV) infected rats. (A) EEG recorded in the hippocampus and cortex at 8 h and day 2 after infection.
Left, superimposed hippocampal and cortical EEG traces at 40 C after infection at 8 h. Note the enhanced low amplitude uctuation at 414 Hz ltered
from a period marked by the horizontal (bar and 1) on the EEG traces. The spikes are indicated by arrowhead. Right, superimposed EEG traces of
the hippocampus and the cortex at 40 C after infection at day 2. (B) The enhancement of the low amplitude uctuation in IAV infected rats during
hyperthermia (40 C). Top (inset), the low amplitude uctuation (h) characterization. The low amplitude uctuation was signicantly enhanced in the
hippocampus of infected rats compared to the cortex of infected rats at 8 h and day 2 of post infection. (C) Left, A gradual enhancement a of abnormal
EEG burst amplitude (top) and increased EEG spikes (bottom) in both the hippocampus (left) and the cortex (right) of infected rats at 8 h and day 2
of post infections compared to uninfected ones. All data are meanSD (* P0.05, ** P0.01, *** P0.001; (n3: uninfected, post infection 8 h and
day 2)).
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1131
3C, right). Hence, to study the effects of inuenza virus
infection during hyperthermia, we analyzed mainly the
EEG recorded from the hippocampus in the subsequent
experiments.
Fig. 4A shows representative results of the effects of
heat stress on the EEG recorded from the hippocampus of
rats at 8 h postinfection at different body temperatures. At
37 C, the enhanced low amplitude uctuation at 414 Hz
was dominant, which is depicted in the expanded and
ltered trace (bar and 1). Under hyperthermia (41 C), the
pattern shifted to theta oscillation (36 Hz). The FFT plots
display this pattern shift with a peak at 3 Hz (Fig. 4B, Top).
The 2DFFT plots show the EEG frequency distribution at
different body temperatures (Fig. 4B, bottom). Theta oscil-
Infected
Control
3 Hz
0.8 Hz
0
5
37
38
39
40
41
T
e
m
p
e
r
a
t
u
r
e

(


C
)

0
37 C
o
38 C
o
41 C
o
Hippocampus
Frequency (Hz)
p.i: 8 h
1
1
A
2D - FFT
12 8 4 0
Frequency (Hz)
FFT
B C
Temperature ( C)
o
12 8 4 0
Frequency (Hz)
3 Hz
FFT
EEG - cortex
EEG - hippocampus
41 C
o
D
60
40
20
0
2.0
1.5
1.0
0.5
0.0
37 38 39 40 41
E
E
G

b
u
r
s
t

a
m
p
l
i
t
u
d
e









(
m
V
)
N
u
m
b
e
r

o
f

s
p
i
k
e
s

/

m
i
n
60
40
20
0
37 38 39 40
N
u
m
b
e
r

o
f

b
u
r
s
t

/

m
i
n
*
Control
Infected (p.i.: 8h)
2 s
0
.
5

m
V
1 s 500 ms
0
.
2

m
V
0
.
5

m
V
1 s
0
.
5

m
V
Fig. 4. Hyperthermia induces abnormal EEG activities in rat brain infected with inuenza A virus. (A) EEG in the hippocampus recorded at different
body temperatures (37, 38, and 41 C). At 37 C, the EEG pattern is characterized by slow oscillations. A segment of the activity showing enhanced
low amplitude oscillation is expanded and ltered at 414 Hz indicated by the horizontal bar and 1. At 38 C, the EEG activity is characterized by
increased burst frequency. A segment of the EEG is expanded to show the spikes indicated by arrowhead. At 41 C, the EEG pattern shifted to theta
oscillation (36 Hz). A segment of the activity is expanded to show the theta activity. (B) Top, FFT plot shows EEG recorded at 41 C in uninfected
and infected rats. Note that EEG activities peaked at 0.8 Hz, indicating the predominance of slow oscillation in uninfected rat, and at 3 Hz, indicating
the predominance of theta oscillation, in infected rat. Bottom, 2D-FFT plots. Note the irregular wide range of high frequency components in infected
rat. (C) Top, EEG activities displaying theta oscillation in the cortex and hippocampus of rat at post infection day 3. Bottom, FFT plots shows the
frequency distribution of the corresponding EEG (top) traces. Note the peak at 3 Hz. (D) Top, Effect of changes in body temperature on the number
of bursts in uninfected and infected rats; middle, the increased EEG burst amplitude; bottom, increase EEG spikes for low temperature at 37
compared to hyperthermia at 41 C. All data are meanSD (* P0.05, n3, post infection 8 h). The 2D-FFT is calculated for a period of a minute
of EEG activity.
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1132
lation appeared under hyperthermia, but the day of ap-
pearance varied between infected rats after infection; two
out of ve at 8 h, two of four at day 1, three of six at day 2,
three of ve at day 3, three of four at day 4, two of three at
day 5, and three of four at day 6. The theta oscillation also
appeared in the cortex (Fig. 4C), but none of the uninfected
rats (n10). Hyperthermia effect was also observed on the
EEG burst frequency, the EEG burst amplitude and the
EEG spikes. The number of EEG bursts increased signif-
icantly with the body temperature increase and remained
almost constant over 38 C (Fig. 4D, top). The EEG burst
amplitude and EEG spikes increased in parallel with the
rise in the rat body temperature (Fig. 4D, bottom). But at
37 C, the EEG burst amplitude was almost same as that
in the uninfected rats. In contrast, the number of EEG
spikes increased under the same temperature. This may
indicate some difference in the neuronal mechanism lead-
ing to EEG burst amplitude and EEG spikes increases.
These ndings including the high voltage EEG burst am-
plitude and theta oscillation induced under hyperthermia
are similar to IAE patients EEG signals recorded during
high-grade fever.
Fig. 5A shows the time-dependent changes in the EEG
burst amplitude and number of spikes at two body temper-
atures (37 and 40 C) after infection. Both parameters
reached peak levels during 8 h to day 2 postinfection, and
then gradually decreased to levels close to the control. The
rates of these changes under the heat stress were higher
than those under normal temperature.
Fig. 5B shows the time course of inuenza virus de-
tection in the brain and lung after infection. The inuenza
virus mRNA level in the lung was signicantly higher at day
2 but decreased to basal level at day 4. On the other hand,
the increase in mRNA level in the brain was delayed,
relative to the lung, with the peak at day 4 postinfection and
return to basal level at day 6. Similar nding has been
reported in mice infected with IAV (Wang et al., 2009, in
press). Correlation between EEG abnormalities and virus
detected in the brain showed that the former appeared
much earlier than the latter.
Effects of ketaminexylazine anesthesia
To determine whether the hyperthermia induced abnor-
mal neuronal responses in IAV infected rats were de-
pendent on the anesthesia, we used ketaminexylazine
anesthesia under heat-stressed condition. It is known
that ketaminexylazine anesthesia induces slow sleep
oscillation (1 Hz) without generating theta oscillation in
animal experiments (Steriade et al., 1993a; Timofeev et
al., 1996). However, as shown in Fig. 7A, the EEG
activities in the IAV infected rats recorded under ket-
aminexylazine anesthesia displayed theta oscillation
similar to that found under isourane anesthesia (Fig.
7A). In contrast, the EEG burst amplitude and the num-
ber of spike decreased compared with those observed
under isourane anesthesia (Fig. 7B). These ndings
suggest that the appearance of theta oscillation, high
voltage EEG burst amplitude and increased EEG spikes
Fig. 5. Hyperthermia increases EEG burst amplitude and number of
spikes in rats infected with inuenza A virus. (A) EEG burst amplitude
(top) and number of spikes (bottom) plotted against time after infec-
tion. Note the EEG burst amplitude and number of spikes reached
peak values at 8 hday 2 postinfection (* P0.05, n3). comparing
infected rats at 37 to at 40 C. Note the increased EEG spikes at 37 C.
(B) Changes in viral RNA in the brain and lung after intranasal instil-
lation of the virus. Viral RNA was undetected at 8 h and reached a
peak at day 2 in the lung and at day 4 in the brain after intranasal
infection. C stands for the uninfected (control) rats. All data are
meanSD, points at 37 and 40 C for the same post infection time are
statistically compared (* P0.05, n3). For interpretation of the refer-
ences to color in this gure legend, the reader is referred to the Web
version of this article.
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1133
were heat-stress related events, but may also suggest
these abnormal EEG activities may be induced by dif-
ferent neuronal mechanism.
DISCUSSION
We have shown that in inuenza-infected rats, in the ab-
sence of heat stress, the EEG signals did not display
abnormal neuronal responses. However, under anesthesia
and during hyperthermia condition, EEGs displayed high
voltage EEG burst amplitudes, increased EEG spikes, the
EEG pattern shift from slow waves to theta oscillation.
These abnormalities reached the peak levels at 8 h to day
2 of postinfection and were enhanced under hyperthermia.
The major nding in the present study is that the abnormal
EEG activities (i.e high voltage EEG burst amplitude slow
waves and theta oscillation) in IAE patients could be well
reproduced in anesthetized IAV infected rats under hy-
perthermia.
The cytokines productions are usually associated with
neurological complications in IAE patients. We measured
the amounts of IL-6 and TNF- in the serum after mea-
surements of EEG activities under hyperthermia (n15).
The increases in the serum IL-6 levels were statistically
signicant at the early phase of post infection (8 hday 2)
(Fig. 6A). But the TNF- level was lower even undetect-
able in some rats. The IL-6 level correlate with the abnor-
mal EEG activities during hyperthermia in infected rats
(Fig. 6B). IL-6 is used in clinics as one of the diagnostic
parameter in IAE patients (Aiba et al., 2001; Fukumoto et
al., 2007). Its high concentration was associated with neu-
rological complications in IAE patients and once the serum
IL-6 level was increased to 15,000 pg/mL none of the
patients survived. High IL-6 levels have also been mea-
sured in mice infected with IAV (data not shown).
The hyperthermia condition was the most critical factor
to aggravate the abnormal EEG activities in IAV infected
animals. This was clear in the chronically implanted non-
anesthetized rats without heat stress (Fig. 1A). The EEG in
these rats at normal body temperature displayed patterns
commonly observed in rats during wakefulness, character-
ized by small amplitude, irregular activities and theta os-
cillation in the EEG of the cortex or hippocampus. But the
theta was dominant in the hippocampus after infection.
These characteristics are similar and common to other
mammals studied to date (Nita et al., 2008; Steriade and
Timofeev, 2003) including humans (Arnolds et al., 1980).
Two distinct rhythmic activities depicted on the EEG
were prominent and consistently present in the infected
rats during hyperthermia condition under the anesthesia.
These rhythmic activities were (1) theta (36 Hz) oscilla-
tion, and (2) low amplitude uctuation (414 Hz) during the
burst suppressed periods.
It is known that theta oscillation in the hippocampus is
most consistently present during exploration (Jouvet,
1969; Vanderwolf, 1969; Winson, 1972; Buzsaki, 2002)
and rapid eye movement (REM) sleep (Vanderwolf, 1969)
in non-anesthetized animals. In clinic, the EEG of IAE
patients with inuenza is characterized by high voltage
theta oscillation during high-grade fever (Okumura et al.,
2005; Fukumoto et al., 2007). Our nding is the rst ex-
perimental evidence to show theta oscillation during sleep
(slow wave) oscillation and under anesthesia with both
Fig. 6. Serum cytokines levels and EEG signals at different post
infection time. (A) The serum interleukin-6 (IL-6) level was high at the
early phase of post infection (8 hday 2). (B) EEG recording in the
hippocampus of the same rats at different post infection time. Note
high voltage EEG burst amplitude at post infection time (8 hday 2). All
data are meanSD, (* P0.05, n3).
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1134
isourane and ketaminexylazine in the infected rats dur-
ing hyperthermia.
The pattern shift to theta may explain in part the occa-
sional shift of neuronal activity in the brains of some pa-
tients infected with inuenza virus during high-grade fever
from a normal to a hyperactive brain state, similar to the
awake or REM state. Therefore, we assume this pattern
shift leading to an unbalance in the neuronal activity is the
most sensitive period for patients to undergo abnormal
behavior changes during sleep.
The low amplitude uctuation (414 Hz) during the
burst suppressed periods was spontaneous and present in
both uninfected and infected rats under anesthesia. This
rhythmic activity was slightly reected in the cortex (see
ltered traces, Figs. 23A). It was clear but lower ampli-
tude in the hippocampus of control animals, even more
pronounced in the hippocampus of infected animals (Fig.
3A, B) and was observed in all infected rats (8 hday 6).
This may suggests a possible involvement of other brain
structures within a large neuronal network in the genera-
tion of this rhythmic activity mainly present in the hip-
pocampus. It is known that hippocampus receives rhyth-
mic inputs from many brain areas including the medial
septum and diagonal band complex (MSDB) (Chandler
and Crutcher, 1983; Amaral and Kurz, 1985; Manseau et
al., 2008), the entorhinal cortex (perforant path axons)
(Bliss and Lomo, 1973) and the suppramammillary nuclei
(Vertes and Kocsis, 1997). The MSDB is believed to act
as a pacemaker for the hippocampal theta generation
(Goutagny et al., 2008). Thus, the interaction between
these inputs may play a role in the generation and the
spread of the rhythmic low amplitude uctuation (414 Hz)
in the hippocampus and the cortex of rats.
The factors that triggered the generation of the high
voltage EEG burst amplitude and the increased EEG spike
in IAV infected rats during hyperthermia and their physio-
logical meaning are not clear. Our experiments show that
these abnormal EEG activities under the ketaminexyla-
zine anesthesia were less than those under the isourane
anesthesia (Fig. 7). Ketamine is an antagonist of N-methyl-
D-aspartate (NMDA) receptor (Brooks et al., 1997; Cisse et
al., 2004; Hetman and Kharebava, 2006) and also known
to perform as useful analgesics. Ketamine attenuates neu-
ropathic pain in association with decreased EEGamplitude
and cortical somatosensory evoked potential amplitude
(Kochs et al., 1996; Oga et al., 2002). And also it exhibits
a neuroprotective effect in inammatory pain in neonatal
rodents (Anand et al., 2007). Based on these reports, IAV
infection that triggers abnormal biochemical signals may ex-
cessively stimulate NMDA receptors during hyperthermia.
In addition, we observed that intra-peritoneal adminis-
tration of picrotoxin (5 mg/kg) known as GABA receptors
blocker (Yoon et al., 1993) did not decrease the EEG burst
amplitude or EEG spikes in the infected rats (data not
shown). Also another factor to be considered is that isou-
rane blocks gap junctions (Peracchia, 1991), which form
contacts between inhibitory interneurons. This likely re-
duces synchrony between interneurons, and, therefore,
may affect excitatory interactions with overall result of
increased EEG burst amplitude. Hence, we assume that
the high voltage EEG burst amplitude and increased EEG
spikes are, at least in part, related to the activation of
NMDA receptors during hyperthermia. However, the
mechanism underlying the activation of the NMDA recep-
tors is unknown.
AcknowledgmentsThis work was supported in part by Grants-
in-Aid 20611013 and for The Special Coordination Funds for
Promoting Science and Technology of Ministry of Education, Cul-
ture, Sports, Science and Technology of Japan.
Fig. 7. A decrease of number of spikes and EEG burst amplitude in infected rats under ketaminexylazine anesthesia. (A) EEG recorded over the
hippocampus in rats at different body temperatures (37, 38, and 41 C). The EEG activity is characterized by slow oscillations at 3738 C and theta
oscillation at 41 C. Bottom, segment indicated by (1, 2, 3 and horizontal bar) on each EEG trace is expanded and ltered. (B) Effect of hyperthermia
and type of anesthesia on EEG burst amplitude and spikes. Note the reduced EEG burst amplitude and number of spikes in rats anesthetized with
ketaminexylazine compared with those with isourane anesthesia. Data are meanSD, (n3).
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1135
REFERENCES
Achermann P, Borbly AA(1997) Low-frequency (1Hz) oscillation in the
human sleep electroencephalogram. Neuroscience 81:213222.
Aiba H, Mochizuki M, Kimura M, Hojo H (2001) Predictive value of
serum interleukin-6 level in inuenza virus associated encephalop-
athy. Neurology 57:295299.
Amaral DC, Kurz J (1985) An analysis of the origins of the cholinergic
and noncholonergic septal projections to the hippocampal forma-
tion of the rat. J Comp Neurol 240:3759.
Anand KL, Garg S, Roynaghi CR, Narsinghani U, Bhutta AT, Hall, RW
(2007) Ketamine reduces cell death following inammatory pain in
newborn rat brain. Pediatr Res 62:283290.
Arnolds DE, Lopez da Silva FH, Aitink JW, Kamp A, Boeijinga P (1980)
The spectral properties of hippocampal EEG related to behaviour
in man. Electroencephalogr Clin Neurophysiol 50:324328.
Bliss TVP, Lomo T (1973) Long-lasting potentiation of synaptic trans-
mission in the dentate area of the aneasthetized rabbit following
stimulation of perforant path. J Physiol 232:331356.
Brooks WJ, Petit TL, LeBoutillier JC (1997) Effect of chronic adminis-
tration of NMDA antagonist on synaptic development. Synapse
26:104113.
Buzsaki G (2002) Theta oscillation in the hippocampus. Neuron
33:325340.
Chandler JP, Crutcher KA (1983) The septohippocampal projection in
the rat: an electron microscopic horseradish peroxidise study. Neu-
roscience 10:685696.
Chen Y, Mizuguchi H, Yao D, Ide M, Kuroda Y, Shigematsu Y, Yamaguchi
S, Yamaguchi M, Kinoshita M, Kido H(2005) Thermolabile phenotype of
carnitine palmitoyltransferase II variations as a predisposing factor for
inuenza-associated encephalopathy. FEBS Lett 579:20402044.
Cisse Y, Crochet S, Timofeev I, Steriade M (2004) Synaptic enhance-
ment induced through callosal pathways in cat associative cortex.
J Neurophysiol 92:32213232.
Cox NJ, Subbarao K (1999) Inuenza. Lancet 354:12771282.
Fukumoto Y, Okumura A, Hayakawa F, Suzuki M, Kato T, Watanabe
K, Morishima T (2007) Serum levels of cytokines and EEG nding
in children with inuenza associated with mild neurological com-
plications. Brain Dev 29:425430.
Fujimoto S, Kobayashi M, Uemura O, Iwasa M, Ando T, Katoh T,
Nakamura C, Maki N, Togari H, Wada Y (1998) PCR on cerebro-
spinal uid to show inuenza-associated acute encephalopathy or
encephalitis. Lancet 352:873875.
Goutagny R, Manseau F, Jackson J, Danik M, Williams S (2008) In
vitro activation of the medial septum-diagonal band complex gen-
erates atropine-sensitive and atropine-resistant hippocampal theta
rhythm: an investigation using a complete septohippocampal prep-
aration. Hippocampus 18:531535.
Hetman M, Kharebava G (2006) Survival signaling pathways activated
by NMDA receptors. Curr Top Med Chem 6:787799.
Jouvet M (1969) Biogenic amines and the state of sleep. Science
163:3241.
Kasai T, Togashi T, Morishima T (2000) Encephalopathy associated
with inuenza epidemics. Lancet 355:15581559.
Kido H, Okumura Y, Yamada H, Le TQ, Yano M (2007) Protease
essential for human inuenza virus entry into cells and their inhib-
itors as potential therapeutic agents. Curr Pharm Des 13:403412.
Kido H, Okumura Y, Takahashi E, Pan HY, Wang S, Chida J, Le TQ,
Yano M (2008) Host envelope glycoprotein processing proteases
are indispensable for entry into human cells by seasonal and highly
pathogenic avian inuenza virus. J Mol Genet Med 3:167175.
Kochs E, Scharein E, Molenberg O, Bromm B, Sculte J (1996) Anal-
gesic efcacy of low-dose ketamine. Anesthesiology 85:304314.
Le QT, Kawachi M, Yamada H, Shiota M, Okumura Y, Kido H (2006)
Identication of trypsin I as a candidate for inuenza A virus and
Sendai virus envelope glycoprotein processing protease in rat
brain. Biol Chem 387:467475.
Manseau F, Goutagny R, Danik M, Williams S (2008) The hippocampo-
septal pathway generates thythmic ring of GABAergic neurons in the
medial septumand diagonal bands: an investigation using a complete
septohippocampal preparation in vitro. J Neurosci 28:40964107.
Mori SI, Nagashima M, Sasaki Y, Mori K, Tabei Y, Yoshida Y,
Yamazaki K, Hirata I, Sekine H, Ito T, Suzuki S (1999) A novel
amino acid substitution at the receptor-biding site on the hemag-
glutinin of H3N3 inuenza A viruses isolated from 6 cases with
acute encephalopathy during the 19971998 season in Tokyo.
Arch Virol 144:147155.
Morishita T, Togashi T, Yokota S, Okuno Y, Miyazaki C, Tashiro M,
Okabe N (2002) Encephalitis and encephalopathy associated with
an inuenza epidemic in Japan. Clin Infect Dis 35:512517.
Nita DA, Cisse Y, Timofeev I (2008) State-dependent slow outlast-
ing activities following neocortical kindling in cats. Exp Neurol
211:456468.
Oga K, Kojima T, Matsuura M, Nagashima M, Kato J, Saeki S, Ogawa
S (2002) Effect of low-dose ketamine on neuropathic pain: an
electroencephalogram-electrooculogram/behavioural study. Psy-
chiatry Clin Neurosci 56:355363.
Okumura A, Takashi N, Fukumoto Y, Higuchi K, Kamiya H, Watanabe
K, Morishima T (2005) Delirious behavior in children with inuenza:
its clinical features and EEG ndings. Brain Dev 27:271274.
Paxinos G, Watson C (2005) The rat brain in stereotaxic coordinates.
5th ed, pp 5558. MA: Elsevier Academic Press.
Peracchia C (1991) Effects of anesthetics heptanol, halothane and
isourane on gap junction conductance in craysh septate axons:
a calcium- and hydrogen-independent phenomenon potentiated by
caffeine and theophylline, and inhibited by 4-aminophyridine. J
Membr Biol 121:6778.
Rojas MJ, Navas JA, Rector DM (2006) Evoked response potential
markers for anesthetic and behavioural states. Am J Physiol Regul
Integr Comp Physiol 291:R189R196.
Stan Leung L (1998) Generation of theta and gamma rhythms in the
hippocampus. Neurosci Biobehav Rev 22:275290.
Steriade M, Nuez A, Amzica F (1993a) A novel slow (1Hz) oscilla-
tion of neocortical neurons in vivo: depolarizing and hyperpolariz-
ing components. J Neurosci 13:32523285.
Steriade M, Contreras R, Curr Dossi R, Nuez A (1993b) The slow
(1Hz) oscillation in reticular thalamic and thalamocortical neu-
rons: scenario of sleep rhythm generation in interacting thalamic
and neocortical networks. J Neurosci 13:32843299.
Steriade M, Nuez A, Amzica F (1993c) Intracellular analysis of relation
between the slow (1 Hz) neocorical oscillation and other sleep
rhythms of electroencephalogram. J Neurosci 13:32663283.
Steriade M, Timofeev I, Grenier F (2001) Natural waking and sleep
states: a view from inside neocortical neurons. J Neurophysiol
85:19691985.
Steriade M, Timofeev I (2003) Neuronal plasticity in thalamocortical
networks during sleep and waking oscillation. Neuron 37:
563576.
Timofeev I, Contreras D, Steriade M (1996) Synaptic responsiveness
of cortical and thalamic neurones during various phases of slow
sleep oscillation in cat. J Physiol 494:265278.
Timofeev I, Grenier F, Steriade M (2001) Disfacilitation and active
inhibition in the neocortex during the natural sleep-wake cycle: an
intracellular study. Proc Natl Acad Sci U S A 98:19241929.
Togashi T, Matsuzono Y, Narita M (2000) Epidemiology of inuenza-
associated encephalitis-encephalopathy in Hokkaido, the north-
ernmost island of Japan. Pediatr Int 42:192196.
Vanderwolf CH (1969) Hippocampal electrical activity and voluntary
movement in the rat. Electroencephalogr Clin Neurophysiol 26:
407418.
Vertes RP, Kocsis B (1997) Brainstem-diencephalo-septohippocam-
pal systems controlling the theta rhythm of the hippocampus. Neu-
roscience 81:893926.
Wang S, Le TQ, Chida J, Cisse Y, Yano M, Kido H (2009) Mechanisms
of matrix metalloproteinase-9 upregulation and tissue destruction
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1136
in various organs in inuenza A virus infection. J Med Invest, in
press.
Watts ADJ, Herrick IA (1999) The effect of sevourane and isourane
anesthesia on interictal spike activity among patients with refrac-
tory epilepsy. Anesth Analg 89:12751281.
Winson J (1972) Interspecies differences in the occurrence of theta.
Behav Biol 7:479487.
Yoon KW, Convey FD, Rothman SM (1993) Multiple mechanism of
picrotoxin block of GABA-induced current in rat hippocampal neu-
rons. J Physiol 464:423439.
(Accepted 29 October 2009)
(Available online 3 November 2009)
Y. Ciss et al. / Neuroscience 165 (2010) 11271137 1137