Cognitive and socio-emotional deficits in platelet-derived growth factor receptor-β gene knockout mice.

Abstract

Platelet-derived growth factor (PDGF) is a potent mitogen. Extensive in vivo studies of PDGF and its receptor (PDGFR) genes have reported that PDGF plays an important role in embryogenesis and development of the central nervous system (CNS). Furthermore, PDGF and the β subunit of the PDGF receptor (PDGFR-β) have been reported to be associated with schizophrenia and autism. However, no study has reported on the effects of PDGF deletion on mice behavior. Here we generated novel mutant mice (PDGFR-β KO) in which PDGFR-β was conditionally deleted in CNS neurons using the Cre/loxP system. Mice without the Cre transgene but with floxed PDGFR-β were used as controls. Both groups of mice reached adulthood without any apparent anatomical defects. These mice were further examined by conducting several behavioral tests for spatial memory, social interaction, conditioning, prepulse inhibition, and forced swimming. The test results indicated that the PDGFR-β KO mice show deficits in all of these areas. Furthermore, an immunohistochemical study of the PDGFR-β KO mice brain indicated that the number of parvalbumin (calcium-binding protein)-positive (i.e., putatively γ-aminobutyric acid-ergic) neurons was low in the amygdala, hippocampus, and medial prefrontal cortex. Neurophysiological studies indicated that sensory-evoked gamma oscillation was low in the PDGFR-β KO mice, consistent with the observed reduction in the number of parvalbumin-positive neurons. These results suggest that PDGFR-β plays an important role in cognitive and socioemotional functions, and that deficits in this receptor may partly underlie the cognitive and socioemotional deficits observed in schizophrenic and autistic patients.

Full text

Page 1

Cognitive and Socio-Emotional Deficits in Platelet-
Derived Growth Factor Receptor-b Gene Knockout Mice
Phuong Thi Hong Nguyen1, Tomoya Nakamura1, Etsuro Hori1, Susumu Urakawa2, Teruko Uwano3,
Juanjuan Zhao1, Ruixi Li1, Nguyen Duy Bac1, Takeru Hamashima4, Yoko Ishii4, Takako Matsushima4,
Taketoshi Ono2, Masakiyo Sasahara4, Hisao Nishijo1*
1System Emotional Science, Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama, Japan, 2Department of Judo Physiotherapy,
Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama, Japan, 3Integrative Neuroscience, Graduate School of Medicine and
Pharmaceutical Sciences, University of Toyama, Toyama, Japan, 4Department of Pathology, Graduate School of Medicine and Pharmaceutical Sciences, University of
Toyama, Toyama, Japan
Abstract
Platelet-derived growth factor (PDGF) is a potent mitogen. Extensive in vivo studies of PDGF and its receptor (PDGFR) genes
have reported that PDGF plays an important role in embryogenesis and development of the central nervous system (CNS).
Furthermore, PDGF and the b subunit of the PDGF receptor (PDGFR-b) have been reported to be associated with
schizophrenia and autism. However, no study has reported on the effects of PDGF deletion on mice behavior. Here we
generated novel mutant mice (PDGFR-b KO) in which PDGFR-b was conditionally deleted in CNS neurons using the Cre/loxP
system. Mice without the Cre transgene but with floxed PDGFR-b were used as controls. Both groups of mice reached
adulthood without any apparent anatomical defects. These mice were further examined by conducting several behavioral
tests for spatial memory, social interaction, conditioning, prepulse inhibition, and forced swimming. The test results
indicated that the PDGFR-b KO mice show deficits in all of these areas. Furthermore, an immunohistochemical study of the
PDGFR-b KO mice brain indicated that the number of parvalbumin (calcium-binding protein)-positive (i.e., putatively c-
aminobutyric acid-ergic) neurons was low in the amygdala, hippocampus, and medial prefrontal cortex. Neurophysiological
studies indicated that sensory-evoked gamma oscillation was low in the PDGFR-b KO mice, consistent with the observed
reduction in the number of parvalbumin-positive neurons. These results suggest that PDGFR-b plays an important role in
cognitive and socioemotional functions, and that deficits in this receptor may partly underlie the cognitive and
socioemotional deficits observed in schizophrenic and autistic patients.
Citation: Nguyen PTH, Nakamura T, Hori E, Urakawa S, Uwano T, et al. (2011) Cognitive and Socio-Emotional Deficits in Platelet-Derived Growth Factor Receptor-b
Gene Knockout Mice. PLoS ONE 6(3): e18004. doi:10.1371/journal.pone.0018004
Editor: Izumi Sugihara, Tokyo Medical and Dental University, Japan
Received January 12, 2011; Accepted February 17, 2011; Published March 18, 2011
Copyright: ? 2011 Nguyen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported partly by CREST (Core Research for Evolutional Science and Technology), JST (Japan Science and Technology Agency), Japan,
JSPS (Japan Society for the Promotion of Science) Asian Core Program, and the Ministry of Education, Science, Sports and Culture, Grant-in-Aid for Scientific
Research (A) (22240051). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: nishijo@med.u-toyama.ac.jp
Introduction
The family of platelet-derived growth factors (PDGF) comprises
4 members—PDGF-A, -B, -C, and -D—that are assembled from
disulfide-linked homo- or heterodimers of 2 distinct but related
chains (PDGF-AA, -AB, -BB, -CC, and -DD). Two receptor
subtypes of PDGF (PDGFR-a and -b) can form mature dimeric
receptor complexes that can bind to ligands with different affinities
[1]. PDGFR-a is largely expressed in oligodendroglial progenitors,
while PDGFR-b is predominantly expressed in neurons [2] and
upregulated in the neonatal rat brain [3]. PDGF-BB that
specifically binds to PDGFR-bb is abundantly expressed in
neurons and is upregulated in neonatal brains [4–6].
Altering PDGFR-b may result in abnormalities during the
development of the central nervous system (CNS) due to the loss of
paracrine and autocrine stimulation [7,8]. Consistent with this
hypothesis, evidence of the relationship between alteration of
PDGFR-b and -BB and human psychiatric disorders has been
reported. According to linkage analyses, PDGFR-b is located on
chromosome 5q31-q32 [9], which contains susceptibility genes for
schizophrenia [10–17]. Following the first report by Sherrington
et al. [18] that indicated the association between PDGFR-b and
schizophrenia, several researchers demonstrated that PDGFR-b
affects neurotransmitter systems related to schizophrenia and/or
autism. D2 dopamine receptor-mediated transactivation of
PDGFR-b alters excitatory neurotransmission mediated by the
N-methyl-D-aspartate subtype of glutamate receptors [19]. PDGF
exerts neurotrophic effects on both c-aminobutyric acid (GA-
BA)ergic and dopaminergic neurons [3,20,21] and has long-lasting
effectson N-methyl-D-aspartate
transmission in the hippocampus [22,23]. Furthermore, recent
studies reported that 3 single nucleotide polymorphisms and 2
haplotypes of PDGFR-b are associated with schizophrenia [24],
and that serum levels of PDGF-BB are high in autistic boys [25].
To investigate the role of PDGFR-b in CNS development in
vivo, conditional knockout (KO) mice with suppressed expression
of neuronal PDGFR-b in the Cre/loxP system (PDGFR-b KO
mice) were generated [26]. However, no apparent anatomical
defects were observed in these mice. The association of PDGFR-b
and PDGF-BB with schizophrenia and autism in humans (see
receptor-mediatedsynaptic
PLoS ONE | www.plosone.org1March 2011 | Volume 6 | Issue 3 | e18004

Page 2

above) strongly suggests that the PDGFR-b KO mice show
cognitive and socioemotional deficits. Here the PDGFR-b KO
mice were examined using a battery of behavioral tests designed to
analyze such deficits. Furthermore, parvalbumin-positive neurons,
which are mostly GABAergic and closely involved in the
pathologies of schizophrenia and autism [27–29], were analyzed
in the amygdala, hippocampus, and medial prefrontal cortex.
Evoked gamma oscillation associated with GABAergic neurons,
which was reduced in schizophrenia [30–32], was also analyzed.
Methods
Ethics statement
All mice were housed in individual cages in a temperature-
controlled environment with a 12/12-h light/dark cycle (lights
were turned on and off at 08:00, and 20:00, respectively). Food
and water was supplied ad libitum. Mice (10- to 16-week-old) were
handled for 3 consecutive days before the start of the experiments.
All experimental protocols were performed in accordance with the
guidelines for care and use of laboratory animals approved by the
University of Toyama and the National Institutes of Health’s Guide
for the Care and Use of Laboratory Animals, and approved by the
Committee for Animal Experiments at the University of Toyama.
(License number: S-2009MED-9).
Generation of conditional PDGFR-b KO mice
The Cre/loxP system was used to develop conditional PDGFR-
b KO mutants. A previously established mutant mouse line was
used, in which exons 4–7 of PDGFR-b, which encode the
extracellular domain of the PDGFR-b protein, were flanked by
2 loxP sequences (floxed) positioned in introns 3 and 7 [33]. After
Cre-mediated recombination, deletion of the loxP-flanking region
and resulting frame shift mutation in the adjoining 39 region
occurred in PDGFR-b. To obtain conditional PDGFR-b KO, we
then crossed mutant mice harboring the PDGFR-b floxed allele
and those expressing Cre recombinase under the control of the
nestin promoter and enhancer (nestin-Cre+mouse, The Jackson
Laboratory, Bar Harbor, ME, USA) as previously described
[26,34]. Before this cross, both mutant mice harboring floxed
PDGFR-b and nestin-Cre+were outbred to the mice of C57BL/6J
(B6/J) strain for 14 generations to replace the genetic background
of our mutant mice with that of the B6/J strain. In the present
study, the following 2 types of 10- to 16-week-old male mice were
used: mice with the Cre transgene and floxed PDGFR-b (PDGFR-b
KO mice) and mice without the Cre transgene but with floxed
PDGFR-b (control mice).
Genotypes were confirmed by PCR of tail DNA, using
oligonucleotide primers pairs for floxed PDGFR-b and for the
Cre transgene as described previously [26]. The genotyping was
confirmed by Western blot of the total lysates of the adult mouse
brains to show that the PDGFR-b expression decreased to
undetectable levels in the PDGFR-b KO mice compared with
that in the control mice [26]. A total of 41 control male mice born
from 10 dams and 41 PDGFR-b KO male mice born from 21
dams were used in the present study. The following all behavioral,
histological, and neurophysiological testing was conducted by
experimenters blind to genotype.
Hot plate test
Sensitivity to thermal nociception [35] was evaluated using a
commercially available hot plate analgesia meter (Model DS37,
Ugo Basile, UK). The apparatus consisted of a metal plate
(24.5624.5 cm), which could be heated to a constant temperature,
on which a plastic cylinder (20 618 cm; diameter 6height) was
placed. Mice were brought to the testing room and allowed to
acclimate for 10 min prior to the test.
The latency to respond to the thermal stimulus (56.060.1uC)
was defined as the time between the moment the mouse was
placed inside the cylinder and when it licked or flicked its hind
paws or jolted or jumped off the hot plate. Each animal was tested
once per session.
Food search test
Mice were individually placed into the apparatus used for the
food search test (KUROBOX, Phenotype Analyzing Co., Ltd.,
Nagasaki, Japan) [36]. The apparatus consisted of 2 compart-
ments. The nest compartment (8061306210 mm; length6width
6 height) was separated from the observational compartment
(24062406210 mm) by a partition, and the mouse could freely
move through the partition into the square observation field.
Barycentric coordinates of each mouse were recorded using 64
infrared photosensors.
The nest compartment was covered to block light for 24 h. The
observation compartment was maintained in the 12/12-h light/
dark cycle. Food stations were set up at the 4 corners of the
observational compartment. Each food station consisted of a
polyvinylchloride wall unit (30645620 mm) and a food cup
containing powdered food. Although all 4 food cups were filled
with the powdered food, a mesh shutter covered 3 of the 4 food
cups. The mice did not have access to the food in the portions
covered by the mesh shutter; however, the same olfactory stimulus
emanated from all food cups. The rotary feeder moved the meshed
shutter in a counter-clockwise direction so that the food station
changed every 4 h. At any given time, the mice could only take
food through a single station that was not covered by the mesh
shutter. The supply of water was ad libitum in the observation
compartment.
The barycentric coordinates of each mouse were recorded every
second. Regions of interest (ROIs) were established at all 4 corners
of the observational field. Each ROI consisted of a 60660 mm2
area. When the barycenter of a mouse remained within the same
ROI for more than 4 s, this was counted as a visit to that food
station. The locomotion of each mouse over a period of 4 days,
number of visits to each ROI, and the order of these visits, which
were numbered until the mouse returned to the nest, were
recorded. Correct visit ratio was defined as the ratio of the number
of visits to the correct food station to the number of visits to all
stations.
Social interaction test
The testing cage consisted of a white plastic box (38.56
22.5620 cm). Individual mice were allowed to acclimate to the
testing cage for 20 min prior to the test. Pairs of mice from the
same group were placed in opposite corners of the box. Their
activities in the box were recorded using an overhead CCD
camera for 30 min. Frequencies and durations of social activities
(e.g., proximity behavior, approaching and leaving, following,
social sniffing, active and passive contact, and mounting) were
automatically analyzed using the SocialScan program (CleverSys
Inc., Reston, VA, USA). The plastic box was wiped with 70%
ethanol and air dried between the trials.
Social contact was defined as interbody distances between 2
mice of less than 20 mm. Furthermore, social contact was divided
into the following 2 subcategories: active and passive social
contact. When 1 mouse approached and actively contacted
another mouse, the mouse’s behavior was considered active and
the other mouse’s behavior was considered passive. These 2 types
of social contact were defined as follows: the active approaching
PDGFR-b Involvement in Socio-Emotional Development
PLoS ONE | www.plosone.org2March 2011 | Volume 6 | Issue 3 | e18004

Page 3

mouse to move faster than 40 mm/s within 15 frames (corre-
sponding to 0.5 s) and the ratio of the ‘‘movement’’ of the active
mouse to that of the passive mouse to be greater than 2.0. Here,
‘‘movement’’ is defined as M(n)=1 – Intersect(A(n), A(n+1))/
Union(A(n), A(n+1)), where A(n) is denoted as the animal area at
previous frame n.
Approaching was defined by the direction, speed, and distance
covered. The direction criteria required the approaching mouse to
move towards the other mouse, and the angle between the
direction of the approaching mouse, which was defined as the
head direction detected by the mouse nose, and the line to the
other mouse to be less than 45u. The speed criteria required the
approaching mouse to move faster than 30 mm/s and the other
mouse to move slower than 100 mm/s within 15 frames (0.5 s).
The distance criteria required the approaching mouse to travel at
least 30 mm and the distance between the 2 mice to be less than
1000 mm. The duration of each approaching was defined as the
continuous duration of such action that lasted at least for 2.0 sec.
At the beginning of the approach, two animals should not be very
close, by default, inter distance more than 100 mm. At the ending
of the approach, two animals should be quite close.
Social sniffing was defined as the distance between the nose of
the sniffing mouse and the body of the other mouse being sniffed
of less than 30 mm. The duration of each social sniffing was
defined as the continuous duration of such action that lasted at
least for 0.27 sec.
Social leaving was also defined by the direction, speed, and
distance covered. The direction criteria required the leaving
mouse to leave the other mouse first, and the angle between the
direction of the leaving mouse and the line from the leaving mouse
to the other mouse to be greater than 100u. The speed criteria
required the speed of the leaving mouse to be greater than
30 mm/s within 15 frames (0.5 s). The distance criteria required
the distance between the 2 mice to be less than 1000 mm, and the
leaving mouse to travel at least 30 mm. Continuous actions
meeting these criteria and lasting for at least 0.5 s were considered
social leaving.
Social following was defined by the following criteria: (1) the
angle between the direction of the leaving mouse and the direction
of the following mouse to be less than 90u, (2) the angle between
the direction of the following mouse and the line from the
following mouse to the leaving mouse to be less than 30u, (3) the
angle between the direction of the leaving mouse and the line from
the leaving mouse to the following mouse to be greater than 100u,
(4) the distance between the 2 mice to be less than 300 mm, (5)
both the following and leaving mice to move at least 5 mm within
15 frames (0.5 s), and (6) the following mouse to travel at least
30 mm within 15 frames (0.5 s). Continuous actions meeting these
criteria and lasting for at least 0.5 s were considered social
following.
Social mounting was defined as a contact in which two-thirds of
the head or body of 1 mouse rode on top of the other mouse for at
least 20 frames (0.6 s). It is detected by the joint shape change of
two animals.
Contextual and cued fear conditioning
On day 0, each mouse was placed in a testing chamber
(11611612.5 cm; O’hara & Co, Ltd., Tokyo, Japan) inside a
sound-attenuated chamber with a brightness of 200 lux and
background white noise of 50 dB. Each mouse was allowed to
explore freely for 5 min. The chamber was equipped with a barred
metal floor that could deliver an electric shock. Initially, the
baseline level of freezing behavior was recorded for 5 min. Then, a
10 kHz, 65-dB tone, which served as the conditioned stimulus, was
presented for 20 s. This tone was followed by a footshock (0.4 mA
for 1 s), which served as the unconditioned stimulus. Four more
conditioned stimulus/unconditioned stimulus pairings were pre-
sented with an interstimulus interval of 465–870 s. Freezing
behavior was analyzed for 20 s after the presentation of each
conditioned tone.
A contextual test was conducted 24 (day 1), 48 (day 2), and 72 h
(day 3) after conditioning in the same chamber without the
presentation of the tone or footshock. The mice were left in the
chamber for 11 min, and the freezing time during the initial 5 min
was analyzed. The animals were then returned to their home
cages. A cued test was conducted in a different environment 2 h
after the contextual test using a white plastic box with a brightness
of 50 lux, background white noise of 60 dB, and presence of the
smell of alcohol. The same conditioned tones were presented 5
times on each experimental day without the footshock after the
baseline level of freezing behavior was assessed for 5 min. Freezing
behavior was analyzed for 20 s after the presentation of each tone.
Freezing behavior was analyzed using video-captured images of
the mice. Images of the mice were captured at the rate of 1
frame/s. For each pair of successive frames, the distance (in pixels)
covered by the mouse was measured. When this distance was
below a certain threshold (20 pixels), the behavior was judged as
‘‘freezing.’’ This optimal threshold (number of pixels) to judge
freezing was determined by adjusting it to the amount of freezing
measured by human observation.
Prepulse inhibition (PPI)
The startle reflex measurement system (O’Hara & Co.) was used
to measure PPI. A test session began by placing a mouse in a
plastic cylinder in a sound-attenuated chamber and leaving it
undisturbed for 10 min. The background white noise level in the
chamber was 65 dB. The prepulse sound (0, 70, 72, 74, 78 and
82 dB) was presented for 120 ms before the startle stimulus
[120 dB white noise (40 ms), main pulse] was provided. Each test
session was composed of 36 trials. Six blocks of the 6 prepulse/
main pulse combination types were presented in a pseudorandom
order, such that each combination type was presented once within
a block. The average intertrial interval was 15 s (range, 10–20 s).
Startle responses were recorded for 140 ms (measuring the
response every 1 ms), starting with the onset of the prepulse
stimulus. These responses were corrected for the body weight of
each mouse. The PPI percentage was calculated using the
following formula: [(startle amplitude in trials without prepulse)
2 (startle amplitude in trials with prepulse)]/(startle amplitude in
trials without prepulse) 6100.
Forced swim test (FST)
FST used in this study was based on the original version used for
mice by Porsolt with modifications. Mice were placed in a cylinder
(15 cm622.5 cm; diameter 6 height) filled with water (15 cm
high). The mice were not able to touch the bottom. The water
temperature was set at 2561uC. The cylinder was placed in a box
with infrared cell sensors on the walls to detect swimming activity
(SCANET, Melquest Inc., Toyama, Japan). The software set up a
rectangle that circumscribed the body of an animal every 0.3 sec.
If the animal went out of the rectangle (i.e., a part of the animal
body was detected out side the rectangle) 0.3 sec after setting up
the rectangle, the software counted as ‘‘movement (swimming)’’ in
that period. ‘‘Immobility’’ was defined as such if the animal stayed
within the same rectangle 0.3 sec after setting up the rectangle.
On the first day, the mice were placed in water and forced to
swim in a single trial of 15 min. On the second day, the mice were
placed in water and forced to swim in a single trial of 5 min.
PDGFR-b Involvement in Socio-Emotional Development
PLoS ONE | www.plosone.org3 March 2011 | Volume 6 | Issue 3 | e18004

Page 4

Swimming time was continuously recorded every 1 min of each
trial in each animal. After the test, the animals were dried with
towels and returned to their cages.
Immunohistochemistry
Under deep anesthesia with sodium pentobarbital (50 mg/kg
body weight, i.p.), the mice were transcardially perfused with
heparanized saline (0.9% w/v NaCl), followed by 4% parafor-
maldehyde dissolved in 0.1 M phosphate buffer (PB) (pH 7.4).
After perfusion, the brains were removed from the skull, coronarily
cut into small blocks, and postfixed in 4% paraformaldehyde
overnight. The fixed brain blocks were immersed in 30% sucrose
dissolved in 0.1 M PB until they sank down to the bottom. The
brain blocks were then frozen in dry ice and coronarily cut into
40 mm-thick sections. The sections were placed in 0.01 M
phosphate buffer saline (PBS) and then transferred into an
antifreeze solution (25% ethylene glycol, 25% glycerin, and 50%
0.1 M PB) and stored at 220uC until immunohistochemical
staining could be performed.
The sections were washed thrice in 0.01 M PBS for 15 min,
blocked with 3% normal horse serum in PBS for 30 min at room
temperature, and incubated overnight at 4uC with mouse
monoclonal anti-parvalbumin antibodies (1:10,000 dilution in
1% horse serum PBS, Sigma, St. Louis, MO, USA). These sections
were washed thrice with 0.01 M PBS for 10 min each time,
incubated with biotinylated horse anti-mouse IgG (1:200 dilution,
Vector, Burlingame, USA) for 50 min at room temperature, and
then, after washing, incubated in ABC reagent (Vector) for
50 min. Finally, the parvalbumin-immunoreactive elements were
visualized by reacting them with 25 mg 3,39-diaminobenzidine
and 30 ml 30% H2O2in 100 ml 0.01 M PBS (pH 7.4) for 5–
8 min. The sections were then rinsed several times in PBS,
dehydrated in graded concentrations of ethanol, cleared in xylene,
and cover-slipped with Entellan (MercK, Darmstadt, Germany).
Negative control sections were treated identically except for
omission of the primary antibody. No reaction product was
observed in any of the control sections.
To count the number of parvalbumin-positive neurons, 3 sections
of each level of the brain were selected from each animal, and digital
images of the stained sections were captured at +1.1, +1.3, and
+1.4 mm AP from bregma (medial prefrontal cortex, anterior
cingulate cortex, infralimbic and prelimbic areas), 20.5, 20.7,
21.1, 21.6, and 22.1 mm AP (amygdala), and 21.6 and 22.1 mm
AP (dorsal hippocampus). These digital images were analyzed using
the ImageJ software (NIH ImageJ; http://rsbweb.nih.gov/ij/).
Anatomical structures were delineated on the basis of the atlas of
the C57BL/6 mouse brain provided by Hof et al. [37]. Cell counts in
the sections above the AP level were averaged in each animal.
Neurophysiological recordings of auditory event-related
gamma oscillation
In this experiment, naive control (n=10) and PDGFR-b KO
(n=9) were used (the animals received no other behavioral
testings). Under anesthesia with avertin (187.5 mg/kg, i.p.), 2
screws, which later worked as EEG electrodes, were implanted
over the frontal cortex and cerebellum of the skulls of the mice. A
connector for the wires was connected to the screws and attached
to the skull using cranioplastic. After recovery from surgery (1
week), the animals were acclimated to all of the handling and
testing procedures described previously. On the days when
measurements were recorded in the dark phase, the animals were
put into a plastic recording box (23561856125 mm). Signals from
the electrodes were amplified and band-pass filtered at 1.5–
1000 Hz (3 dB corner, 6 dB octave/slope) using an amplifier
Figure 1. Social interactions of the PDGFR-b KO and control mice. Frequencies (A) and durations (B) of each social behavior are indicated.
Black columns, PDGFR-b KO mice; white columns, control mice; *, p,0.05; **, p,0.01; and ***, p,0.001 (Student’s t-test).
doi:10.1371/journal.pone.0018004.g001
PDGFR-b Involvement in Socio-Emotional Development
PLoS ONE | www.plosone.org4March 2011 | Volume 6 | Issue 3 | e18004

Page 5

(MEG-5200G, Nihon Kohden, Tokyo, Japan). The amplified
analog signals were digitized at 2 kHz and stored using the Micro
1401mkII hardware and Spike2 (ver. 6) software (Cambridge
Electronic Design, Cambridge, UK). Auditory stimuli (5 kHz pure
tone, 500 ms duration) were generated using a sound generator
(DPS-725T, DIA Medical System, Tokyo, Japan) and delivered
using a speaker after amplification to 85 dB at an interval of 3 s
with a background noise of 50 dB.
Recent studies have reported a reduction in early phase-locked
gamma oscillation (evoked gamma responses; ERPs) as a
characteristic of schizophrenia [30–32]. In the present study,
ERPs in response to auditory stimulation were analyzed [38].
ERPs recorded in individual trials were digitally band-pass filtered
between 30–80 Hz using infinite impulse response filters and then
averaged across the trial for each animal. Phase synchronized
gamma oscillation with respect to stimulus onset would survive the
averaging process and can be seen in the averaged ERPs. The
averaged gamma data were rectified by squaring to produce a
positive value for gamma activity at each point. This enabled the
measurement of the area under the curve (AUC) for the time
window between 0–100 ms from stimulus onset. The mean AUC
across the animals was compared between the 2 groups of mice.
Statistical data analysis
Quantitative data were expressed as the mean 6 SEM. The
data were analyzed by two-way mixed repeated measures
ANOVA based on general linear model followed by Bonferroni
test, or by Student’s t-tests using SPSS 19.0 (SPSS Inc., Chicago,
IL). The statistical significance level was set at p,0.05.
Results
Social interaction
Figure 1 shows the frequencies (A) and durations (B) of each
social behavior studied in both groups. In comparison with the
control mice, the PDGFR-b KO mice showed significantly lower
frequencies and durations of proximity behaviors (Student’s t-test,
p,0.05; control, n=9; KO, n=9). Consistent with these results,
the PDGFR-b KO mice showed decreases in the other categories
of social behaviors, including social sniffing, active social contact,
passive social contact, and mounting (Student’s t-test, p,0.001;
control, n=9; KO, n=9), except for duration of social sniffing
(Student’s t-test, p.0.05). On the other hand, the PDGFR-b KO
mice showed significantly higher frequencies and durations for
approaching and leaving (Student’s t-test, p,0.001). Videotapes
were also analyzed manually afterwards to check aggressive
behaviors. However, no aggressive behaviors were observed in
both the groups.
PPI
PPI is a paradigm for sensorimotor gating that is most widely
used in animal models of schizophrenia and autism. PPI can be
easily measured in rodents in a manner almost identical to
procedures used in humans [39–41]. Statistical comparisons by
two-way mixed repeated measures ANOVA indicated that the
PDGFR-b KO mice have significant deficits in sensorimotor
gating (Figure 2A); there was a significant main effect of group in
PPI [F(1, 30)=5.902; p,0.05; control, n=16; KO, n=16],
although there was no significant interaction between group and
prepulse intensity [F(4, 120)=0.166; p.0.05; control, n=16; KO,
n=16]. Furthermore, no significant differences were observed in
the amplitudes of the startle responses to the main pulse without a
prepulse between the control and PDGFR-b KO mice (Student’s
t-test, p.0.05; control, n=16; KO, n=16) (Figure 2B). These
results indicate that sensorimotor gating was disturbed in the
PDGFR-b KO mice.
Sensitivity to nociception and fear conditioning
First, sensitivity to nociception was analyzed in the PDGFR-b
KO mice. Figure 3A shows the mean escape latencies measured in
the hot plate test. Statistical comparisons indicated no significant
difference between the control and PDGFR-b KO mice (Student’s
t-test, p.0.05; control, n=10; KO, n=9). This finding suggests
that sensitivity to nociception in the PDGFR-b KO mice did not
differ from that in the control mice.
To test the associative learning capacity of the PDGFR-b KO
mice, animals were tested for contextual and cued fear
conditioning (Figure 3B–D). In conditioning (Figure 3B), compar-
ison by two-way mixed repeated measures ANOVA (group 6
conditioning trial No.) indicated that there were marginally
significant main effect of group [F(1, 14)=3.411, p=0.086;
control, n=9; KO, n=7], and significant interaction between
group and conditioning trial No. [F(4, 56)=4.173, p=0.005;
control, n=9; KO, n=7]. Furthermore, when the data in the last
Figure 2. Prepulse inhibition (PPI) of acoustic startle responses.
A: PPI (%) at 5 different prepulse intensities (70, 72, 74, 78, and 82 dB).
The PDGFR-b KO mice displayed significantly less PPI than the control
mice [F(1, 30)=5.902; p,0.05; control, n=16; KO, n=16]. B: Acoustic
startle amplitudes measured in trials without a prepulse. No significant
differences were observed in the acoustic startle amplitudes of the 2
groups of mice. Values indicate the mean 6 SE. White columns, control
mice, n=16; black columns, PDGFR-b KO mice, n=16.
doi:10.1371/journal.pone.0018004.g002
PDGFR-b Involvement in Socio-Emotional Development
PLoS ONE | www.plosone.org5March 2011 | Volume 6 | Issue 3 | e18004