Available via license: CC BY 4.0
Content may be subject to copyright.
Research Paper
Dendritic Homeostasis Disruption in a Novel Frontotemporal Dementia
Mouse Model Expressing Cytoplasmic Fused in Sarcoma
Gen Shiihashi
a
,DaisukeIto
a,
⁎, Itaru Arai
b
,YukiKobayashi
c
, Kanehiro Hayashi
d
, Shintaro Otsuka
b
,
Kazunori Nakajima
d
, Michisuke Yuzaki
b
, Shigeyoshi Itohara
c
,NorihiroSuzuki
a
a
Department of Neurology, Keio University School of Medicine, Shinjuku-ku, Tokyo 160-8582, Japan.
b
Department of Physiology, Keio University School of Medicine, Shinjuku-ku, Tokyo 160-8582, Japan
c
Laboratory for Behavioral Genetics, RIKEN Brain Science Institute, Wako, Saitama 351-0198, Japan
d
Department of Anatomy, Keio University School of Medicine, Shinjuku-ku, Tokyo 160-8582, Japan
abstractarticle info
Article history:
Received 4 June 2017
Received in revised form 17 August 2017
Accepted 5 September 2017
Available online 9 September 2017
Cytoplasmic aggregation of fused in sarcoma (FUS) is detected in brain regions affected by amyotrophic lateral
sclerosis (ALS) and frontotemporal dementia (FTD), which compose the disease spectrum, FUS proteinopathy.
To understand the pathomechanism of ALS-FTD-associated FUS, we examined the behavior and cellular proper-
ties of an ALS mouse model overexpressing FUS with nuclear localization signal deletion. Mutant FUS transgenic
mice showed hyperactivity, social interactional deficits, and impaired fear memory retrieval, all of which are
compatible with FTD phenotypes. Histological analyses showed decreased dendritic spine and synaptic density
in the frontal cortex before neuronal loss. Examination of cultured cells confirmed that mutant but not wild-
type FUS was associated with decreased dendritic growth, mRNA levels, and protein synthesis in dendrites.
These data suggest that cytoplasmic FUS aggregates impair dendritic mRNA trafficking and translation, in turn
leading to dendritic homeostasis disruption and the development of FTD phenotypes.
© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
1. Introduction
The identification of DNA- and RNA-binding protein TARDNA-bind-
ing protein (TDP-43)as a component of inclusions in amyotrophic later-
al sclerosis (ALS) and frontotemporal dementia (FTD) has led to an ALS
researchbreakthrough in 2006 (Arai et al., 2006; Neumannet al., 2006).
Since then,it has been proposed that FTD and ALSconstitute a novel dis-
ease spectrum sharing a common molecular basis (Ito and Suzuki,
2011). In 2009, mutations in the gene encoding the RNA-binding pro-
tein fused in sarcoma (FUS) (also known as translocated in liposarcoma
[TLS]) were identified as the cause of familial ALS (ALS6) (Kwiatkowski
et al., 2009; Vance et al., 2009). Subsequently, several investigators re-
ported observing FUS-positive inclusions in FTD that was classified as
basophilic inclusion body disease (BIBD) and neuronal intermediate fil-
ament inclusion disease (NIFID) (Munoz et al., 2009; Neumann et al.,
2009b). Thus, it has been proposed that mutant-FUS–linked ALS, atypi-
cal frontotemporal lobar degeneration with ubiquitinated inclusions,
BIBD, and NIFID also comprise an ALS-FTD disease spectrum: FUS
proteinopathy. Thereafter, a growing number of studies identified
many causative genes of ALS-FTD, such as C9orf72 G4C2 repeat
expansion and hnRNPA2B1, emphasizing the importance of this novel
disease spectrum (Kim et al., 2013; Majounie et al., 2012).
In FUS, which plays roles in DNA repair, transcription, alternative
splicing, translation, and RNA transport (Ito and Suzuki, 2011), ALS-
linked mutations are enriched at the nuclear localization signal at the
C-terminus. Cellular analysis reveals that the increased mislocalisation
of mutant protein into the cytoplasm is negatively correlated with the
age of disease onset (Dormann et al., 2010; Ito et al., 2010), which is
similar to polyglutamine diseases. The impairment of the nucleartrans-
port of FUS is therefore directly associated with neurodegeneration in
ALS and FTD. Furthermore, a mutation in a truncated protein lacking a
nuclear localization signal was identified to cause juvenile ALS with
rapid disease progression and cognitive impairment (Zou et al., 2013),
suggesting that impaired nucleocytoplasmic trafficking of FUS alone
can lead to the ALS phenotype (Waibel et al., 2010).
A major unresolved question is whether FUS-mediated neurodegen-
eration is caused by a toxic gain of function by cytoplasmic FUS aggre-
gates or a loss of normal FUS function in the nucleus. Recently, we
generated a transgenic mouse line overexpressing exogenous FUS
with nuclear localization signal deletion (ΔNLS-FUS), reflecting juvenile
ALS (Shiihashiet al., 2016). This FUS transgenic (tg) mouse, in which ex-
ogenous FUS protein is located strictly in cytoplasm at moderate level
(~80% of the endogenous FUS level), showed significant progressive
motor impairment from 20 weeks of age. Pathological analysis revealed
EBioMedicine 24 (2017) 102–115
⁎Corresponding a uthor at: Departm ent of Neurology, K eio University School of
Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan.
E-mail address: d-ito@jk9.so-net.ne.jp (D. Ito).
http://dx.doi.org/10.1016/j.ebiom.2017.09.005
2352-3964/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Contents lists available at ScienceDirect
EBioMedicine
journal homepage: www.ebiomedicine.com
ubiquitin- and p62-positive cytoplasmic ΔNLS-FUS aggregates and also
astrocytosis, microgliosis, and neuronal loss in the brains of FUS tg mice
at 1 year, indicating the recapitulation of some aspects of pathological
ALS. However, the expression level, nuclear distribution, and function
of RNA splicing of the endogenous FUS were not changed in FUS tg
mice. We therefore concluded that the toxicgain of function of cytoplas-
mic FUS aggregates was sufficient to lead to neurodegeneration
(Shiihashi et al., 2016).
To understand the pathomechanism of cytoplasmic FUS aggregates
in ALS-FTD, it is necessary to establish a novel FTD model of FUS
proteinopathy. Several groups have reported that transgenic mice
with mutant FUS showed ALS-like motor deficits (Qiuetal.,2014;
Scekic-Zahirovic et al., 2017; Scekic-Zahirovic et al., 2016; Sharma et
al., 2016); however, thus far none has evaluated FTD phenotypes. In
this study, we examined the behavior and cognitive function of ΔNLS-
FUS tg mice before the appearance of motor deficits. We also performed
anatomical and cellular examinations in order to understand the molec-
ular basis of neurodegeneration by cytoplasmic FUS aggregates. During
behavioral analysis, FUS tg mice showed hyperactivity, social interac-
tional deficits, and impaired fear memory retrieval, all of which are
compatible with FTD phenotypes. Regarding the pathological mecha-
nism of cytoplasmic aggregation of FUS, histological analysis revealed
that cytoplasmic FUS aggregates recruited robust mRNA and RNA trans-
porters such as Staufen and fragile X mental retardation protein
(FMRP). Golgi staining and electrophysiology showed decreased den-
dritic spine density in the frontal cortex and hippocampus of FUS tg
mice before neuronal loss. Moreover, examination of primary cortical
cultured neurons confirmed that mutant but not wild-type FUS de-
creased dendritic growth, mRNA levels, and protein synthesis in den-
drites. These data suggest that cytoplasmic FUS aggregates trap mRNA
and its transporters, impairing dendritic mRNA trafficking and transla-
tion, in turn leading to the disruption of dendritic homeostasis and the
development of FTD phenotypes.
2. Materials and Methods
2.1. Generation of ΔNLS-FUS Tg Mice
This procedure is described in full in the article of our previous study
(Shiihashi et al., 2016); briefly, a human FUS cDNA with deleted NLS
was inserted into the pTSC21k vector encoding the murine Thy1.2 ex-
pression cassette. Tg lines were bred and backcrossed more than five
generations with C57Bl6/J mice. Mice were housed in standardized ven-
tilated microisolation caging (four animals per cage). Genotyping from
tail DNA was performed using the following primer pairs: 5′-
AAGAAGACCTGGCCTCAAACG-3′and 5′-TATCCCTGGGGAGTTGACTG-
3′. Animal experiments were approved by the Committee on Animal
Care and Use, Keio University (09217- [3]) and conducted according
to the Animal Experimentation Guidelines of Keio University School of
Medicine.
2.2. Behavioral Tests
Male non-tg (n= 15) and FUS tg (n= 14) mice between 9 and
13 weeks of age were used for the following behavioral tests, which
were performed sequentially.
Home cage test: Each mouse (9 weeks old) was placed alone in a
testing cage (18.8 cm [W] × 28.8 cm [L] × 13.7 mm [H]) under a
12-h light-dark cycle (light on at 08:00) and with free access to
both food and water. Spontaneous activity in the cage was assessed
for 3 continuous days by counting the number of infrared beam
crossings (Scanet; Melquest).
Open-field test: Each mouse (11 weeks old) was placed in an open-
field arena (50 cm [W] × 50 cm [L] × 40 cm [H]; O'Hara & Co., Ltd.)
made of white polyvinyl chloride. The distance travelled in the
open field was recorded for 30 min using a video-imaging system
(Image OF9; O'Hara & Co., Ltd.). The central area was definedas
the central 18 cm × 18 cm region of the arena. The mice were tested
on two consecutive days.
Y-maze test: The Y-maze test was performedat 11 weeks of age after
the open-field test. The apparatus consisted of a plastic maze with
three V-shaped arms (40 cm [L] × 3 cm [W] at the bottom and
10 cm [W] at thetop opening× 12 cm [H], 120 degrees' separation).
Mice were allowed to move freely through the maze for 10 min. The
sequence and total number of arm entries were recorded by a video
camera. A “novel arm selection”was counted when a mouse succes-
sively entered three different arms. The recorded data were ana-
lyzed offline with EthoVision XT software (Noldus Information
Technology).
Three-chamber test: We evaluated social recognition and
response to social novelty using the three-chamber test at
12 weeks of age. The apparatus consisted of a white plastic box
(42 cm [W] × 61.5 cm [L] × 22 cm [H]) with partitions dividing the
box into three equal-sized chambers, with 10-cm openings between
chambers. The side chamber has a plastic triangular prism cage at
the distal corner for constraining a stranger mouse. Target subjects
(stranger one and stranger two) were age-matched males. This
task was carried out in three phases of 10 min each. In phase one,
the test mouse was placed in the middle chamber and left to explore
the area containing the empty cages for 10 min. In phase two, the
mouse was placed in the middle chamber, but an unfamiliar
mouse (stranger one) was placed into a cage in one of the side cham-
bers. In phase three, a novel stranger mouse (stranger two) was
placed in the previously empty cage and again the test mouse was
left to explore for 10 min. We measured the time the test mouse
spent in each chamber for each phase.
Trace fear conditioning test: This test was carried out in mice aged
13 weeks according to the procedure described previously, with
minor modifications (Suh et al., 2011). On day one, the mouse was
placed in a conditioning chamber. Each mouse was exposed to
tone-foot shock pairing (tone = 20 s, 65 dB white noise; footshock
= 2 s, 0.75 mA, 20 s after termination of tone). The tone-footshock
pairing was performed three times (beginning at 240, 400, and
560 s). Mice were returned to their home cages 98 s after the last
footshock. Fear memory of the tone was investigated on day two
by placing each mouse in a novel chamber. After 4 min of free explo-
ration, each mouse was exposed to three tones (60 s, 65 dB white
noise, separated by a 3-min interval). Animal images were captured
twice per second using a camera, and the area (in pixels) in which
the mouse moved was measured. Freezing was defined as a lack of
movement such that fewer than 20 pixels changed between succes-
sive video frames for at least 2 s. The electrical intensity required to
make the mouse jump was measured.
2.3. Immunoblot Analysis
Brain tissues were homogenized in cold lysis buffer containing
50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5% NP-40, 0.5% sodium
deoxycholate, 0.25% sodium dodecyl sulfate, 5 mM EDTA, and protease
inhibitor cocktail (Sigma). Total protein concentrations in the superna-
tants were determined using a Bio-Rad protein assay kit (Hercules). The
proteins were then analyzed by immunoblotting as follows. Protein
samples were separated by reducing sodium dodecyl sulfate polyacryl-
amide gel electrophoresis (SDS-PAGE) on 10% or 14% Tris-glycine
103G. Shiihashi et al. / EBioMedicine 24 ( 2017) 102–115
gradient gels and then transferred to polyvinylidene difluoride
membranes (MilliporeA). The membranes were incubated with prima-
ry antibodies and subsequently with horseradish peroxidase (HRP)-
conjugated secondary antibodies. Detection was performed using en-
hanced chemiluminescence reagents according to the manufacturer's
instructions (PerkinElmer Life Sciences, Norwalk, CT). The primary anti-
bodies used in this study were anti-c-Myc 9E10 (Santa Cruz, sc-40,
1:1000), anti-FUS (Thermo, PA5-23696, 1:1000; Bethul, A304-819A-
M, 1:1000), anti-GAPDH (Cell Signaling, 14C10, 1:1000), anti-Lamin
A/C (Cell Signaling, #2032, 1:1000), anti-puromycin (Millipore,
MABE343, 1:1000), anti-V5 (Invitrogen, 46-0705, 1:1000), and αtubu-
lin (Cell Signaling,#2125, 1:1000) as an internal loading control. Protein
levels were determined by densitometry using an Epson ES-2000 scan-
ner and ImageJ (National Institutes of Health).
2.4. Cytoplasmic and Nuclear Fractionation
Frozen brain was homogenized in 10 ml/g buffer containing 10 mM
HEPES,10mMNaCl,1mMKH
2
PO
4
, 5 mM NaHCO
3
, 5 mM EDTA, 1 mM
CaCl
2
, 0.5 mM MgCl
2
. After 10 min on ice, 0.5 ml/g 2.5 M sucrose was
added. Next, tissue was homogenized and centrifuged at 6300 ×gfor
10 min. The supernatant was collected as cytoplasmic fraction. The pel-
let was resuspended in RIPA buffer (50 mM Tris, 150 mM NaCl, 1% NP-
40, 5 mM EDTA, and 0.5% sodium deoxycholate) with 2% SDS as nuclear
fraction.
2.5. Immunohistochemical Staining
Mice were perfused intracardially with ice-cold PBS and then 4%
paraformaldehyde under deeply anesthesia. Brains were extracted,
postfixed in 4% paraformaldehyde overnight at 4 °C, transferred to a
30% sucrose solution for cryoprotection, frozen, and stored at −80 °C.
We sectioned brain samples with a cryostat (Leica). The sections were
then incubated with 0.2% triton-X for 15 min at room temperature in
preparation for immunohistochemistry. Immunohistochemistry was
performed using anti-NeuN (Chemicon, MAB377, 1:1000), anti-MAP2
(Merck, AB5622, 1:1000), anti-c-Myc 9E10 (1:1000), anti-Myc (Cell Sig-
naling, #06-549, 1:1000), anti-FMRP (Millipore, MAB2160, 1:1000),
anti-stau1 (Rockland, 600-401-EV4, 1:1000), anti-PSD 95 (NeuroMab,
75-028, 1:500), anti-vesicular glutamate transporter 1 (VGLUT1) (Syn-
aptic System, 135 303, 1:1000), and anti-puromycin (Millipore,
MABE343, 1:1000). Fluorescence intensity was determined by densi-
tometry on sections of the frontal cortex and hippocampus (n= 3 per
genotype) using ImageJ.
2.6. Golgi Staining
The Golgi-Cox impregnation stains were performed using the FD
Rapid GolgiStain Kit according to the manufacturer's instructions
(MTR Scientific Inc.). In brief, brains from 15-week-old FUS and non-tg
mice were immersed in impregnation solution for 2 weeks, transferred
to “Solution C”for 2 days, and cut into 100-μm sections using a cryostat
(CM3050S; Leica). Sections were mounted on APS-coated slides
(Matsunami) and allowed to dry for 2 days before staining with silver
nitrate solution (“Solution D and E”) and dehydrated through descend-
ing alcohol series before mounting with Permount (Falma). For dendrit-
ic spine analysis, z-stack images of the Golgi-stained pyramidal neurons
from Layer V in the frontal cortex and CA1 radiatum region in hippo-
campus were obtained using a 60× objective (BZ-9000; Keyence). We
analyzed the spines from 15 neurons (n = 3 mice per genotype). We
calculated the density by dividing the number of spines per 100 μmof
dendrite starting at a distance N30 μm away from the soma along the
dendrite over a distance 120 μm. The head width of dendritic spines
was measured using the line tool in ImageJ. The protrusion whose
width was b0.3 μm was labeled filopodia and distinguished from the
spine. A spine was labeled mushroom (mature spine) when its maxi-
mum head diameter was N0.6 μm and at least twice the neck diameter.
2.7. Mouse Primary Cultured Cortical Neurons and Cell Line Culture
The cerebral cortex obtained from E16.5 ICR mice (Japan SLC) or
E16.5 C57Bl6/J mice (Japan SLC) mated with FUS tg mice was excised,
incubated with 5 U/ml of papain (Nacalai) in BCG solution (0.02% bo-
vine albumin [Sigma], 0.02% L-cysteine [Sigma], 0.5% glucose [Wako])
at 37 °C, dissociated, and washed with neurobasal medium (Gibco).
The dissociated cells were cultured on the glass area of poly-L-lysine
coated glass-bottom dishes (Matsunami) in neurobasal medium sup-
plemented with B-27 (Life technologies). Neuro2a neuroblastoma
cells were cultured in αMEM (Gibco) supplemented with 10% fetal bo-
vine serum. Transfection was performed using Lipofectamine 2000 (Life
Technologies Corporation, Carlsbad, CA, USA) according to the
manufacturer's instructions at day 2 in vitro.
2.8. Reverse Transcriptase-polymerase Chain Reaction Analysis
The cDNA was synthesized from 1 μg total RNA using the
SuperscriptIII First-Strand Synthesis System (Life Technologies). Ex-
pression levels of glutamate receptor subunit were quantified using
the DyNAmo ColorFlash SYBR Green qPCR Kit (Thermo Fisher Scientific)
on a PikoReal 96 Real-Time PCR System (Thermo Fisher Scientific). Each
sample wasmeasured in triplicate and normalized to β-actin levels. The
primer sets are listed in Supplementary Table 1.
2.9. Electrophysiology
Coronal slices containing prefrontal cortex were prepared from 15-
week-old FUS-tag and non-tg mice (Gascon et al., 2014). Briefly, after
decapitation, the brain was rapidly removed and placed in ice-cold slic-
ing solution containing 87 mM NaCl, 25 mM NaHCO
3
, 2.5 mM KCl,
1.25 mM NaH
2
PO
4
,10mMD-glucose, 75 mM sucrose, 0.5 mM CaCl
2
,
and 7 mM MgCl
2
, (pH 7.4 in 95% O
2
/5% CO
2
, 325 mOsm). Coronal corti-
cal slices containing prefrontal cortex (300 μm thick) were cut using
DOSAKA linear slicer Pro 7. After a 20-min incubation at 34°C, the slices
were stored at room temperature. Experiments were performed at
room temperature. During experiments, slices were superfused with
an extracellular solution containing: 125 mM NaCl, 2.5 mM KCl,
25 mM NaHCO
3
,1.25mMNaH
2
PO
4
,25mMD-glucose, 2 mM CaCl
2
,
and 1 mM MgCl
2
(pH 7.4 in 95% O
2
/5% CO
2
, ~325 mOsm). Whole-cell
voltage-clamp recordings (holding potential of −70 mV) from visually
identified layer V pyramidal neurons in the prefrontal cortex were per-
formed to record mEPSC. Recording pipettes were fabricated from thin
borosilicate glass tubing (World Precision Instruments). Pipette resis-
tance was ~ 3 MΩwhen filled with intracellular solution containing
150 mM Cs-gluconate, 10 mM HEPES, 4 mM MgCl
2
,4mMNa
2
ATP,
1mMNa
2
GTP, 0.4 mM EGTA, and 5 mM QX-314 (pH adjusted to 7.3
with CsOH). A 5-mV hyperpolarizing test pulse was applied every 10 s
to monitor series resistance (R
s
). Data were discarded if R
s
changed
N20% or leak current was N250 pA. 1 μM TTX, 50 μM D-AP5, and 100
μM Picrotoxin were added to extracellular solution to pharmacological-
ly isolate the mEPSCs. In subsets of experiments, 10 μMNBQXwere
added at the end of recording to confirm that the recorded synaptic
events were mediated by the AMPA receptor. Total 9 neurons from 3
non-tg mice and 10 neurons from 4 FUS tg mice were analyzed. Data
were acquired with EPC-9 amplifier (HEKA), low-pass filtered at
2.9 kHz, and sampled at 20 kHz. Liquid junction potential was not
corrected. mEPSCs were detected using Igor pro 6.3.7 (Wavemetrics)
by template matching algorithm and visually verified (Chen et al.,
2017; Pernia-Andrade et al., 2012).
104 G. Shiihashi et al. / EBioMedicine 24 (2017) 102–115
2.10. Co-RNA FISH and Immunofluorescence Staining
The brain frozen sections or cultured neurons fixed in 4% parafor-
maldehyde for 15 min were incubated with pre-hybridization buffer
(50% formamide [Wako], 5× saline-sodium citrate buffer [SSC], 0.05%
heparin sodium [Wako], 0.02% ribonucleic acid from torula yeast
[Sigma])for 60 min at 65 °C and then hybridized with a fluorescently la-
beled locked nucleic acid (LNA) probes, PolyT(25)Vn 5′-fluorescein
(Exiqon 300510-04), Scramble-ISH 5′-fluorescein (Exiqon 300,514–
04), and 5′-FAM-AGGCCAGGTCTTCTTCAGAAATCA-3′(for exogenous
FUS mRNA) (Exiqon) diluted to final concentration of 80 nM, for 24 h
at 65 °C in a dark, humidified chamber. Next, sections were washed
with 2× SSC/formamide for 1 h at 65 °C and washed with TTBS (0.5 M
NaCl, 0.02 M Tris HCl [pH 8.0], 0.1% tween 20) three times. Then immu-
nohistochemistry was performed using anti-c-Myc 9E10 (1:1000), and
anti-V5 (Cell Signaling, #06-549, 1:1000). Fluorescence intensity was
determined by densitometry using ImageJ.
2.11. RNA Labeling With SYTO RNASelect
The brain frozen sections or cultured neurons fixed with −20 °C
methanol for 10 min were incubated with 500 nM SYTO RNASelect
(Life technologies) for 20 min at room temperature as per the
manufacturer's instruction. Then, immunohistochemistry was per-
formed using anti-c-Myc 9E10 (1:1000) and anti-V5 (Cell Signaling,
#06-549, 1:1000).
2.12. Puromycin Labeling
Cultured neurons and N2a cells were treated with 1.8 μMpuromycin
(Nacalai) for 15 min. Then immunohistochemistry or immune blotting
with anti-puromycin (Millipore, MABE343, 1:1000) was performed.
Fluorescence intensity was determined by densitometry using ImageJ.
2.13. Statistical Analysis
Results from non-tg and tg mice were compared using the Student's
t-test. The JMP 11 software package (SAS Institute, Inc.) was used for all
analyses.
3. Results
3.1. Expression and Distribution of Exogenous FUS (ΔNLS-FUS)
FUS proteinopathy defines the majority of tau- and TDP-43-negative
cases of frontotemporal dementia (Urwin et al., 2010). Several FUS mu-
tations have been reported to be linked to familial ALS with features of
frontotemporal dementia (Yan et al., 2010), and juvenile-onset ALS
with cognitive impairment (Hirayanagi et al., 2016). To explore the mo-
lecular mechanism underlying FUS proteinopathy in FTD, we examined
tg mice overexpressing Myc-tagged exogenous ΔNLS-FUS under Thy-1
promoter, which developed progressive motor weakness from
20 weeks of age and formed ubiquitin/p62-positive cytoplasmic aggre-
gates from 6 months of age (Shiihashi et al., 2016). Immunohistochem-
istry and western blot showed that the accumulation of exogenous FUS
protein was located prominently in the frontal cortex, hippocampus,
and entorhinal cortex, indicating that the accumulation of FUS was re-
stricted to specific cell types (Fig. 1a and b, Fig. S1a), consistent with
findings from the brains of patients with FTD (Neumann et al., 2009a;
Neumann et al., 2009b). Although the expressions of exogenous FUS
RNA at 15 weeks of age were significantly lower in the olfactory bulb
and striatum, these in the frontal cortex, hippocampus, occipital cortex,
and cerebellum were similar, indicating no correlation between the
levels of ΔNLS-FUS protein accumulation and exogenous FUS RNA ex-
pression (Fig. 1candd).Weconfirmed that the expression l evel of exog-
enous FUS was ~80% that of the expression level of endogenous FUS,
and exogenous FUS (ΔNLS-FUS) was strictly localized to the cytoplasm
in tg mice at 15 weeks of age; nuclear localization of endogenous mouse
FUS was not affected by ΔNLS-FUS expression (Shiihashi et al., 2016)
(Fig. S1b–d).
3.2. Hyperactivity of ΔNLS-FUS Tg Mice
To examine FTD phenotypes in our tg mice, we performed behavior-
al analyses from 9 to 13 weeks of age, before the appearance of motor
impairment (Shiihashi et al., 2016). First, we performed the home
cage test and open field test to assess locomotor activity from 9 to
11 weeks of age. In the home cage test, evaluating basal motor activity
in a familiar environment, there was no significant difference between
the total locomotor activity of FUS tg and non-tg mice over the 3 days
of observation (Fig. S2a). However, at the first adaptation period, FUS
tg mice showed significantly higher activity levels compared to non-tg
mice (Fig. 2a), indicating the hyperactivity of FUS tg mice in the novel
environment. In the open field test, FUS tg mice also showed greater
movement distances, durations, and speed compared to non-tg mice
(Fig. 2b). These results also indicate the hyperactivity of FUS tg mice.
However, there was no significant difference in the duration spent in
the centre region between FUS and non-tg mice (Fig. 2b), suggesting
that anxiety-related behavior was not altered in FUS tg mice.
Further, we performed the Y-maze test at 11 weeks of age. The num-
ber of novel arm selections did not differ between FUS tg and non-tg mice
(Fig. S2b), indicating that the spatial working memory of FUS tg mice was
not impaired. However, the total number of arm selections was higher in
FUS tg mice compared to non-tg mice, and FUS tg mice also travelled lon-
ger distances and at greater speeds than non-tg mice, indicating hyperac-
tivity (Fig. S2b). Taken together, these results indicate that ΔNLS-FUS
overexpression induces novelty-induced hyperactivity.
3.3. Social Interaction Deficit of FUS Tg Mice
To determinewhether the mice had disease-relevant behavioral def-
icits, we performed the three-chamber test when the mice were
12 weeks of age. In the second phase, both FUS tg and non-tg mice
spent more time in the chamber with the novel mouse than in the
chamber with the empty cage (Fig. 2c), indicating that FUS tg mice
could recognize its conspecific. However, in the third phase of the test,
while non-tg mice spent more time in the chamber with the novel
mouse than in that which contained the familiar mouse, FUS tg mice
spent almost the same amount of time in the chamber with the novel
mouse as with that of the familiar mouse (Fig. 2c). Although perfor-
mance in this test depends critically upon proper olfaction, the olfactory
bulb of FUS tg mice showed no exogenous FUS expression (Fig. 1b) and
no morphological changes (Fig. S2c), suggesting olfactory deficit was
unlikely in FUS tg mice. We concluded FUS tg mice failed to distinguish
novel and familiar mice, suggesting a deficit in social memory.
3.4. Memory Deficit of FUS Tg Mice
When the mice were 13 weeks of age, we conducted the trace fear-
conditioning test to evaluate long-term (24-h) memory formation. In
the paradigm, conditioned stimulus (CS) and unconditioned stimulus
(US) are separated by an interval period, which engages the hippocam-
pus and prefrontal cortex (Gilmartin and Helmstetter, 2010; Suh et al.,
2011). No significant difference in shock sensitivity was detected be-
tween FUS tg and non-tg mice (Fig. S2d). In the cue-retention condition,
FUS tg mice spent significantly less time freezing than non-tg controls
when they received the first and second tone (Fig. 2d). Notably, FUS tg
mice showed normal levels of freezing upon thethird tone (Fig. 2d), in-
dicating that fear memory was retained, and retrieved following repeti-
tion of the cue. These findings suggest that the impairment of 24-h
memory function is due toimpairment in the retrieval of stored memo-
ry rather than disrupted encoding or consolidation of the information
105G. Shiihashi et al. / EBioMedicine 24 ( 2017) 102–115
(Roy et al., 2016). In summary, in behavioral analyses, FUS tg mice
showed hyperactivity, social interactional deficits, and memory dys-
function, recapitulating most FTD phenotypes.
3.5. Decreased Spine Density in the Frontal Cortex and Hippocampus of FUS
Tg Mice
As shown in Fig. S3, and our previous study(Shiihashi et al., 2016),
there was no neuronal loss in the frontal cortex or hippocampus at
15 weeks of age even though behavioral dysfunction was observed. To
reveal the anatomical basis of the cognitive deficit, synaptic connections
were quantified by counting the total number of spines, and those of
mushroom-like morphology alone, in the dendritic region of the frontal
cortex and hippocampus. FUS tg mice showed decreased total spine
density in the frontal cortex and hippocampus relative to non-tg mice
(Fig. 3a). The number of mushroom-shaped mature spines was also
found to be decreased in both regions in FUS tg mice, but the ratio of
mature spines to all spines did not differ between FUS tg mice and
Fig. 1. Expression of exogenous FUS in transgenic mice at 15 weeksof age. (a)Immunohistochemistry with anti-Myc (red) antibody in each part of the brain. Sections counterstainedwith
4′,6-diamidino-2-phenylindole (DAPI; blue). Scale bar, 10 μm. (b) Western blot analysis of brain lysates using anti-Myc antibody. Expression of FUS with nuclear localization signal
deletion (ΔNLS-FUS) protein appears restricted to the frontal cortex and hippocampus. α-tubulin was used as an internal control. (c) Quantitative reverse transcriptase polymerase
chain reaction ( qRT-PCR) analysis for exogenous FUS mRNA in the brain of tg mice. (n= 3 per genotype; *Pb0.05 vs. frontal, occi pital cortex and cerebellum; #P b0.05 vs.
hippocampus by Student's t-test). (d) In situ hybridization showing the expression of theexogenous FUS mRNA (green) in the brain of tg mice. Scale bar, 10 μm.
106 G. Shiihashi et al. / EBioMedicine 24 (2017) 102–115
non-tg mice, indicating that spine formation, but not maturation, is im-
paired by ΔNLS-FUS expression. Next, immunohistochemical evaluation
of postsynaptic and presynaptic markers, PSD95 and VGLUT1, in the
frontal cortex and hippocampus revealed decreased synaptic density
in FUS tg mice relative to non-tg mice (Fig. 3b and c).
To explore thefunctional consequences of reduction of synaptic den-
sity, we performed whole-cell patch-clamp recording from layerV pyra-
midal neurons in brain slices acutely prepared from the frontal cortices
of non-tg and FUS tg mice at 15 weeks of age. Passive membrane prop-
erties were similar between non-tg [input resistance (R
input
), 453.1 ±
91.1 MΩ, membrane capacitance (C
m
), 149.7 ± 13.6 pF] and FUS tg
(R
input
, 318.2 ± 28.6 MΩ,C
m
, 221.4 ± 31.4 pF)] pyramidal neurons.
However, the frequency of miniature excitatory postsynaptic currents
(mEPSCs) was significantly reduced in FUS tg compared to non-tg pyra-
midal neurons (Fig. 3d-f). By contrast, the amplitude of mEPSCs did not
differ between non-tg and FUS tg pyramidal neurons (Fig. 3d–f). These
results indicate that although the function of each excitatory synaptic
input is spared, the total number of excitatory synaptic inputsis reduced
in FUS tg mice.
To further examine synaptic defects, we performed a quantitative
RT-PCR analysis of the expression levels of different glutamate receptor
subunits. Most N-methyl-D-aspartate (NMDA), α-amino-3-hydroxy-5-
Fig. 2. MutantFUS transgenic mice showed behavioral and cognitive impairment. Fourteen mutant FUSand fifteen non-tg mice wereevaluated on all batteries. (a) Home cage locomotor
activity during the adaptation period. (b) The total distance moved, total movement duration, average speed, and time spent in the center region in the open-field test. Transgenic mice
exhibited hyperactivity in their homecages and the open-field test. (c) Three-chamber testto assess sociability. The paradigm is brieflyillustrated. Graphsshow the time spent by the test
mouse in thechambers with and without a novel mouse (phase two: left panel), or the novel mouse and familiar mouse chambers (phasethree: right panel). (d) Trace fearconditioning
test to evaluate long-term memory function. The temporal relationshipsof conditioned stimulus (CS) and unconditioned stimulus (US) are schematically shown. Blue and red numbers
represent the onsets of CS andUS. Twenty-four hoursafter conditioning, freezing ratios were assessed along with theauditory cues. Asterisks (*) indicate significantdifferences vs. non-tg
mice (P b0.05 by Student's t-test). n.s., not significant. a.u., arbitrary units. All error bars represent standard deviation of the mean.
107G. Shiihashi et al. / EBioMedicine 24 ( 2017) 102–115
methyl-4-isoxazolepropionic acid (AMPA), and kainate receptor sub-
units showed a tendency to decrease in FUS tg mice. mRNA encoding
AMPA receptor subunits Gria3 and Gria4 especially were significantly
decreased in FUS tg mice at 15 weeks of age (Fig. S4).
3.6. mRNA Transport was Impaired by the Cytoplasmic FUS Aggregate
We next explored the molecular basis of the dendritic spine deficits
observed in FUS tg mice. The RNA binding protein FUS contributes to
mRNA transport and regulates local translation in the dendrites (Fujii
et al., 2005; Yasuda et al., 2013). We asked whether cytoplasmic FUS ag-
gregates disturb mRNA transport in the dendrites. RNA distribution was
examined using SYTO RNASelect, which selectively stains RNA
(Knowles et al., 1996), in the frontal cortex and hippocampus of FUS
tg mice at 15 weeks of age. The SYTO signal was colocalized with the cy-
toplasmic FUS aggregates and increased with age (Fig. 4a). We further
performed fluorescence in situ hybridization (FISH) for the poly-A
tails of mRNA in the frontal cortex and hippocampus. The signal of
FISH in FUS aggregates also increased with age (Fig. 4b). These results
indicate that cytoplasmic FUS aggregates sequester RNA.
To evaluate mRNA transport in the dendrites, primary cultured crit-
ical neurons from the wild-type mouse embryos were transiently
transfected with mCherry and empty vector, wild-type FUS, or ΔNLS-
FUS at 2 days in vitro (DIV) before being stained by SYTO RNASelect at
Fig. 3. FUS transgenic mice showed synapticimpairment. (a) Representative figure of Golgi staining in the frontal cortex and hippocampus of FUS transgenic (n= 3) and non-tg (n= 3)
mice. Dendritic spine number was quantified along a 100-μmsegment from the origin of theprimary apical dendritic branches of thepyramidal neurons.Asterisks (*) indicate significant
differences vs. non-tg mice (Pb0.05 by Student's t-test). (b, c) Immunohistochemistry for quantification of postsynaptic density 95 (PSD95) protein (b) and vesicular glutamate
transporter 1 (VGLUT1) (c) in the frontal cortex and hippocampus of FUS transgenic and non-tg mice (n = 3 per genotype, P b0.05 by Student's t-test). Scale bar, 20 μm. (d) Whole-
cell patch-clamp recordings of mEPSCs in pyramidal neurons of the frontal cortex. The amplitude and frequency of these miniature events were compared between FUS transgenic and
non tg mice (P= 0.58 and 0.015 by Wilcoxon rank test for amplitude and frequency, respectively). (e, f) Cumulative distributions of mEPSC frequency and amplitude in the frontal
cortex of FUS transgenic and non tg mice (Pb0.01 and 0.16 by Kolmogorov–Smirnov test for frequency and amplitude, respectively).
108 G. Shiihashi et al. / EBioMedicine 24 (2017) 102–115
5 DIV. Nuclear localization of endogenous mouse FUS was not affected
by ΔNLS-FUS expression (Fig. S5). The intensity of SYTO decreased in
the dendrites of the cultured neurons expressing ΔNLS-FUS (Fig. 5a).
We also performed the FISH for poly-A tails in these cultured neurons.
The signal of FISH was clearly decreased in the dendrites of the cultured
neurons expressing ΔNLS-FUS compared to those expressing empty
Fig. 4. The signal of messenger ribonucleic acid (mRNA) was colocalized withthe cytoplasmic FUSaggregates. (a) Frontal cortexand hippocampus of mice at 15 weeks of age was stained
with anti-Myc (red) and SYTORNA select (green). (b) Double-fluorescence insitu hybridization staining for the poly-A tailof mRNA (green) and anti-Myc (red) inthe frontal cortex and
hippocampus of FUS transgenic and non-tg mice. The lower panel shows a scrambled probe staining serving as a negative control. Scale bar, 10 μm.
109G. Shiihashi et al. / EBioMedicine 24 ( 2017) 102–115
vector and wild-type FUS (Fig. 5b and c). We then validated these find-
ings in primary cultured neurons from FUS tg mice. We performed RNA
staining using SYTO RNASelect or FISH for poly-A tails in the cultured
neurons from FUS and non-tg mice embryos at 28 DIV, when exogenous
FUS under post mitotic neurons promoter Thy-1 was substantially
expressed as described below in detail (Fig. 7h). The RNA signal
was also decreased in the dendrites of the cultured neurons from
FUS tg mice (Fig. 5d–f). Collectively,these results indicate that cytoplas-
mic FUS aggregates sequester the mRNA and impair its dendritic
transport.
Fig. 5. The messenger ribonucleic acid(mRNA) signal was decreased in thedendrites of the cultured neuronsexpressing FUS withnuclear localization signal deletion (ΔNLS-FUS), but not
those expressing wild-type FUS. (a) Cultured neurons transfected with mCherry and empty vector, wild-type FUS, or ΔNLS-FUS tagged V5, were stained with SYTO RNASelect (green).
Exogenous FUS expression was confirmed using anti-V5 tag antibody (red; right panels). The SYTO RNASelect signal was decreased in the dendrites of neurons expressing ΔNLS-FUS.
Scale bar, 10 μm. (b) Fluorescence in situ hybridization (FISH) for the poly-A tail of mRNA (green) was performed for the cultured neurons transfected with mCherry and empty
vector, wild-type FUS, or ΔNLS-FUS. (d) Cultured neurons form mutant FUS and non-tg mice were stained with SYTO RNASelect (green). Scale bar, 10 μm. (e) FISH was performed on
the cultured neurons from transgenic mice. Scale bar, 10 μm. (c and f) Quantification of dendritic mRNA as achieved by FISH (c from b, f from e). The mRNA signal in each group was
normalized to the mean mCherry signal in each dendritic segment, intensities relative to the empty vector group were then calculated in (c). The mRNA signal was normalized to the
mean microtubul e associated protein 2 (MAP2) signal in eac h dendrite, and then to the mRNA signal of the non-tg group in (f). A total of 15 neurons from three independent
experiments were analyzed. *Pb0.05 by Student's t-test.
110 G. Shiihashi et al. / EBioMedicine 24 (2017) 102–115
3.7. Cytoplasmic FUS Aggregates Recruit RNA Transporters
FUS binds and retains other RNA-binding proteins via a prion-like
domain in vitro (Blokhuis et al., 2016; Kato et al., 2012; Murakami et
al., 2015). Inclusions in FTD-FUSalso co-accumulated with RNA-binding
proteinsin patients' brains (Neumann et al., 2012), and a previous study
demonstrates that ΔNLS-FUS inclusion sequesters the RNA granules
markers SMN and G3BP (Shiihashi et al., 2016). We therefore examined
Fig. 6. Ribonucleic acid transporters werecolocalized with the cytoplasmic FUS aggregates. (a, b) Frontal cortex and hippocampus fromFUS transgenic and non-tg mice at each age were
stainedwith anti-Myc (red)and FMRP (green;a) or Staufen (green;b). Sections werecounterstainedwith DAPI. Notethat cytoplasmicFUS aggregates recruit both FMRPand Staufen in an
age-dependent manner. Scale bar, 10 μm.
111G. Shiihashi et al. / EBioMedicine 24 ( 2017) 102–115
Fig. 7. Fused in sarcoma (FUS) with nuclear localization signal deletion (ΔNLS-FUS) but not wild-type FUS impaired protein synthesis. (a) Protein synthesis was detected by anti-
puromycin antibody (red) in the primary cultured neurons transfected with GFP and empty vector, wild-type FUS, or ΔNLS-FUS after pulse-treatment with puromycin. Dendrites were
identified by GFP expression. Scale bar, 20 μm. (b) Quantification of dendritic protein synthesis as assessed by puromycylation in Fig. 7a. The fluorescence intensities of each group
were normalized to the mean GFP signal in each dendritic segment. The graphs show the percentage of the mean puromycin signal (mean ± standard error) in each transfected cell to
the emptyvector group in eachdendritic segment.Fifteen transfected cellsfrom three independent experiments were analyzed. (c) Quantitativedata on the numberof dendritic branches
in the primary cultured neurons expressing wild-type FUS or ΔNLS-FUS. n= 15 per group. (d) Neuro2acells were transfected with empty vector, wild-type FUS, or ΔNLS-FUS, and then
treatedwith puromycin. Celllysates were analyzedby western blottingusing an anti-puromycin antibody. (e) Quantification of dendritic proteinsynthesis as assessedby puromycylation
in Fig. 7d. The graphs show the intensities of puromycylationin each cell relative to thatof the empty vector (mean ± standard error; n = 3). *P b0.05vs. empty vector by Student'st-test.
(f) Primary cultured neuronsfrom FUS and non-tg mice were treated with puromycin and immunostaining by anti-MAP2 (green) andpuromycin (red) antibodies was performed. Scale
bar, 20 μm. (g) Quantificationof dendritic proteinsynthesis as assessed by puromycylation in Fig.7f. The fluorescence intensitiesof each group were normalized to themean MAP2 signal.
Fifteen neurons from three independent experiments were analyzed. (h) Cell lysates of primary neurons from the FUS and non-tg mouse embryos were treated with puromycin and
assessed by western blotting. (i) Intensiti es of puromycylation i n the neurons at 28 DIV from transg enic FUS mice relative to those of non-tg mice. *P b0.05 vs. non-tg mi ce by
Student's t-test.
112 G. Shiihashi et al. / EBioMedicine 24 (2017) 102–115
the distribution of other RNA-binding proteins contributing to mRNA
transport (Bardoni et al., 2006; Zhang et al., 1995). We performed im-
munohistochemical tests for FMRP and Staufen (St Johnston et al.,
1991) in the frontal cortex and hippocampus. Both RNA transporters
were colocalized with the cytoplasmic FUS aggregates, and the size of
these aggregates increased with age (Fig. 6a and b). Collectively, cyto-
plasmic FUS aggregates trap RNA transporters, possibly leading pertur-
bation of dendritic transport of their mRNA targets.
3.8. Protein Synthesis was Impaired by ΔNLS-FUS
We next asked whether ΔNLS-FUS expression affected translation in
the dendrites. In primary cultured neurons transiently transfected with
green fluorescence protein (GFP) and empty vector, wild-type FUS, or
ΔNLS-FUS, and pulse-treated bypuromycin, which labels newly synthe-
sized proteins (Schmidt et al., 2009). The intensity of puromycin was
decreased in the dendrites of the cultured neurons expressing ΔNLS-
FUS, but not those expressing wild-type FUS (Fig. 7a and b). Moreover,
the number of longdendrites (N60 μm) was decreased in the case of cul-
tured neurons expressing ΔNLS-FUS (Fig. 7c).
Further, mouse neuroblastoma Neuro2a cells were transiently
transfected with empty vector, wild-type FUS, or ΔNLS-FUS, and the in-
tensity of puromycin was assessed by western blot analysis. Total new
protein synthesis was significantly reduced in cells expressing ΔNLS-
FUS, but rather increased in cells expressing wild-type FUS for some un-
known reason, indicating that ΔNLS-FUS expression impaired protein
synthesis (Fig. 7d and e). Furthermore, we validated these findings in
primary cultured neurons from FUS tg mice. Because Thy-1 expression
is initiated in post mitotic neurons in the perinatal period, substantial
expression of exogenous FUS was detected at 28 DIV, but not at 14
DIV (Fig. 7h). As shown in Fig. 7f–i, neurons from ΔNLS-FUS tg mice
showed decreased protein synthesis at 28 DIV. Taken together, these
data indicate that cytoplasmic FUS impairs protein synthesis in den-
drites, leading to the perturbation of dendritic homeostasis (Fig. 8).
4. Discussion
The present study provides several novel insights into the molecular
basis of FUS-linked ALS-FTD's pathology. First, ΔNLS-FUS tg mice,
previously shown to exhibit the toxic gain of function ALS phenotype
(Shiihashi et al., 2016), also recapitulate several aspects of FTD pheno-
types, hyperactivity, social interactional deficits, and impaired fear
memory retrieval before the appearance of progressive motor deficits
(Fig. 2), establishing them as a novel ALS-FTD mouse model. Our previ-
ous study showed that endogenous FUS expression level, nuclear local-
ization, and splicing activity were not altered in this FUS tg mouse line
(Shiihashi et al., 2016). Therefore, the present study indicates that cyto-
plasmic FUS is also sufficient for an FTD phenotype via a dominant toxic
effect without loss of FUS function. Histological and electrophysiological
analysis revealed decreased dendritic spine and synaptic density in the
frontal cortex and hippocampus of FUS tg mice at 15 weeks (Fig. 3), in
which neuronal loss was not observed (Shiihashi et al., 2016) (Fig.
S3). Glutamate receptor expression analysis reveals that the expression
of AMPA receptor subunits Gria3 and Gria4 was significantly decreased
in FUS tg mice (Fig. S4), consistent with previousstudies in which alter-
ations in AMPA receptors contribute to behavioral deficits mimicking
FTD (Adamczyk et al., 2012; Gascon et al., 2014; Udagawa et al.,
2015). Taken together, the findings of the present study indicate that
synaptic dysfunction is primary and critical for the development of
FTD phenotypes.
Recently, a single-copy mouse model of mutant cytoplasmic FUS has
been reported to display mild motor neuron phenotypes, but no ubiqui-
tin/p62-positive FUS inclusions (Scekic-Zahirovic et al., 2017). In con-
trast, our mice form ubiquitin- and p62-positive cytoplasmic ΔNLS-
FUS aggregates in affected neuron, indicating the recapitulation of as-
pect of pathological ALS (Shiihashi et al., 2016). Another finding of the
present study is that age-dependent progressive cytoplasmic FUS aggre-
gates recruit robust mRNA and RNA transporters such as Staufen and
FMRP (Figs. 4 and 6). Examination of cultured cells confirms that mu-
tant but not wild-type FUS decreases dendritic growth, mRNA levels,
and protein synthesis in dendrites (Figs. 5 and 7). Collectively, RNA dys-
regulation by cytoplasmic FUS aggregates leads to dendritic
dyshomeostasis, resulting in ALS-FTD-linked neuronal dysfunctions.
Notably, FUS aggregates capture various RNA-binding proteins, includ-
ing RNA transporters and stress granule components, and trap a broad
selection of mRNA molecules. Therefore, we propose that the correction
of specific RNAs regulated by FUS is insufficient as a therapeutic mech-
anism. Rather, the correction of widespread RNA dysregulation is
Fig. 8. The possible pathological cascadeof fused in sarcoma(FUS) proteinopathy.ALS-linkedmutants and/or certainpathologicalstresses lead the mislocalization of FUS to the cytoplasm.
ExcessivecytoplasmicFUS assembles as RNAgranules, and thelocal concentration of proteinincreases, resulting in theformation of toxicand tight aggregates/inclusion bodies.Aggregates
trap robust messenger RNA and RNA transporters, and impair dendritic messenger RNA transporters. Disruption of dendritic homeostasis results in the development of neurological
phenotypes, finally causing neuronalloss.
113G. Shiihashi et al. / EBioMedicine 24 ( 2017) 102–115
requiredto ameliorate FUS proteinopathy. A morereasonable molecular
approach for the development of therapeutics should be the clearance
and/or inhibition of the aggregation of FUS itself.
Several limitations and alternative interpretations should be consid-
ered alongside the findings of this study. First, although the frontal cor-
tex and hippocampus showed a dense accumulation of mutant FUS
(Fig. 1), spatial working memory measured by the Y maze, which de-
pends upon hippocampal-prefrontal-cortical circuits, was not impaired
in FUS tg mice until at least 15 weeks of age (Fig. S2b). Behavioral stud-
ies of other neurodegenerative mouse models overexpressing amyloid
precursor protein and Tau, were reported spatial working memory
was preserved in younger mice(Chen et al., 2000; Morgan et al., 2000;
Pennanen and Gotz, 2005). It is possible that the Y maze is not suffi-
ciently sensitive to detect the subtle spatial working memory deficits
of younger mice. Future studies ought to address whether these spatial
working memory impairments develop with age and/or whether other
preserved neuronal circuits compensate for the defective hippocampal-
frontal circuits of FUS tg mice.
In our trace fear conditioning test (Fig. 2d), the third toneelicited the
retrieval of stored memory information, suggesting a retrieval problem,
rather than a storage impairment, in the mutant mice. We speculate
that a reduction in dendritic spine density in the hippocampus and pre-
frontal cortex leads to the impairment of memory retrieval, while mem-
ory engramcells survive until 15 weeks of age. Similarly, it was recently
reported that optogenetic activation of hippocampal memory engram
cells results in memory retrieval in an Alzheimer's disease mouse
model, which also showed a progressive reduction in dendritic spine
density in the hippocampus (Roy et al., 2016). Thus, selective rescue
of dendritic spine density on surviving cells may represent an effective
strategy for treating cognitive dysfunction, and possibly alsomotor def-
icits, in the early stages of ALS-FTD.
Finally,FUS knockout and knockdown did not leadto the ALS pheno-
types, but instead resulted in FTD-like behavioral phenotypes, such as
hyperactivity and reduced anxiety-related behavior (Kino et al., 2015;
Udagawa et al., 2015), indicating that the loss of FUS function also con-
tributes to the development of FTD phenotypes. Although in our previ-
ous study we could not detect evidence of loss of FUS function in ΔNLS-
FUS tg mice (Shiihashi et al., 2016), we cannot rule out the possibility
that the function of cytoplasmic FUS, such as in mRNA trafficking, was
partially lost as a result of its aggregation in restricted cell types, and
that this is associated with the observed cognitive dysfunction.
In summary, our data demonstrate that cytoplasmic FUS aggregates
trap robust mRNA and RNA transporters and lead to synaptic and den-
dritic spinedysfunction resulting in ALS-FTD phenotypes before the ap-
pearance of neuronal loss. Thus, we propose that strategies targeting
dendritic homeostasis might support future approaches to treating the
motor/cognitive deficits of ALS-FTD. The mouse model here reported
on will contribute to a more detailed understanding of ALS-FTD patho-
genesis and to the development of new therapeutic strategies.
Author Contributions
G.S. and D.I. designed the research. G.S. performed experiments and
analyzed the data. I.A., Y.K., K.H.and S.O. performed electrophysiological
experiments, behavior test, primary culture, and Golgi staining, respec-
tively. G.S., D.I., K.N., M. Y., S. I. and N. S. wrote the manuscript.
Conflicts of Interest
The authors declare no conflict of interest.
Funding Sources
This work was supported by Eisai Co. Ltd., KANAE Foundation for the
Promotion of Medical Science, Nakabayashi Trust For ALS Research,
Japan Society for the Promotion of Science (No. 16 J05812) and Life
Science Foundation of Japan and the Ministry of Education, Culture,
Sports, Science and Technology of Japan (No. 15 K09323).
Acknowledgement
We are grateful to Dr. Herman van der Putten (Novartis Institutes for
Biomedical Research, Basel, Switzerland) for providing pTSC21k and to
Naomi Kogo, Chie Sano, Atsuko Ohba (Laboratory for Behavioral Genet-
ics, RIKEN Brain Science Institute), and Tamao Kitamura (Department of
Neurology, Keio University School of Medicine) for technical assistance.
We thank the Collaborative Research Resources, Keio University School
of Medicine, for technical assistance.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.ebiom.2017.09.005.
References
Adamczyk, A., Mejias, R., Takamiya, K., Yocum, J., Krasnova, I.N., Calderon, J., Cadet, J.L.,
Huganir, R.L., Pletnikov, M.V., Wang, T., 2012. GluA3-deficiency in mice is associated
with increased social and aggressive behavior and elevated dopamine in striatum.
Behav. Brain Res. 229, 265–272.
Arai, T., Hasegawa, M., Akiyama, H., Ikeda, K., Nonaka, T., Mori, H., Mann, D., Tsuchiya, K.,
Yoshida, M., Hashizume, Y., et al., 2006. TDP-43 is a component of ubiquitin-positive
tau-negative inclusions in frontotemporal lobar degeneration and amyotrophic later-
al sclerosis. Biochem. Biophys. Res. Commun. 351, 602–611.
Bardoni, B., Davidovic, L., Bensaid, M., Khandjian, E.W., 2006. The fragile X syndrome: ex-
ploring its molecular basis and seeking a treatment. Expert Rev. Mol. Med. 8, 1–16.
Blokhuis, A.M., Koppers, M., Groen, E.J., van den Heuvel, D.M., Dini Modigliani, S., Anink,
J.J., Fumoto, K., van Digg elen, F., Snelting, A., Sodaar, P., et al., 2016. Com parative
interactomics analysis of different ALS-associated proteins identifies converging mo-
lecular pathways. Acta Neuropathol. 132, 175–196.
Chen, G., Chen, K.S., Knox, J., Inglis, J., Bernard, A., Martin, S.J., Justice, A., McConlogue, L.,
Games, D., Freedman, S.B., et al., 2000. Alearningdeficit related to age and beta-am-
yloid plaques in a mouse model of Alzheimer's disease. Nature 408, 975–979.
Chen, C., Arai, I., Satterfield, R., Young Jr., S.M., Jonas, P., 2017. Synaptotagmin 2 is the fast
Ca2+ sensor at a central inhibitory synapse. Cell Rep. 18, 723–736.
Dormann, D., Rodde, R., Edbauer, D., Bentmann, E., Fis cher, I., Hruscha, A., Than, M.E.,
Mackenzie, I.R., Capell, A., Schmid, B., et al., 2010. ALS-associated fused in sarcoma
(FUS) mutations d isrupt Transpor tin-mediated nuclear import. EMBO J. 29,
2841–2857.
Fujii, R., Okabe, S., Urushido, T., Inoue, K., Yoshimura, A., Tachibana, T., Nishikawa, T.,
Hicks, G.G., Takumi, T., 2005. The RNA bindingprotein TLS is translocated to dendritic
spines by mGluR5 activation and regulates spinemorphology. Curr.Biol. 15, 587–593.
Gascon, E., Lynch, K., Ruan, H., Almeida, S., Verheyden, J.M., Seeley, W.W., Dickson, D.W.,
Petrucelli, L., Sun, D., Jiao, J., et al., 2014. Alterations in microRNA-124 and AMPA re-
ceptors contribute to social behavioral deficits in frontotemporal dementia. Nat. Med.
20, 1444–1451.
Gilmartin, M.R., Helmstetter, F.J., 2010. Trace and contextual fear conditioning require
neural activity and NMDA receptor-dependent transmission in the medial prefrontal
cortex. Learn. Mem. 17, 289–296.
Hirayanagi, K., Sato, M., Furuta, N., Makioka, K., Ikeda, Y., 2016. Juvenile-onset sporadic
amyotrophic lateral sclerosis with a frameshift FUS gene mutation presenting unique
neuroradiological findings and cognitive impairment. Intern. Med. 55, 689–693.
Ito, D., Suzuki, N., 2011. Conjoint pathologic cascades mediated by ALS/FTLD-U linked
RNA-binding proteins TDP-43 and FUS. Neurology 77, 1636–1643.
Ito, D., Seki, M., Tsunoda, Y., Uchiyama, H., Suzuki,N., 2010. Nuclear transport impairment
of ALS-linked mutations in FUS/TLS. Ann. Neurol. 69, 152–162.
Kato, M., Han, T.W., Xie,S., Shi, K., Du, X., Wu,L.C., Mirzaei, H.,Goldsmith, E.J., Longgood, J.,
Pei, J., et al., 2012. Cell-free formation of RNA granules: low complexity sequence do-
mains form dynamic fibers within hydrogels. Cell 149, 753–767.
Kim, H.J., Kim, N.C ., Wang, Y.D., Scarborough, E.A., Moore, J., Diaz, Z., Ma cLea, K.S.,
Freibaum, B., Li , S., Molliex, A., et al., 2013. Mutations in pr ion-like domains in
hnRNPA2B1 and hnRNPA1 cause multisystem proteinopathy and ALS. Nature 495,
467–473.
Kino, Y.,Washizu, C., Kurosawa,M., Yamada, M., Miyazaki, H., Akagi, T., Hashikawa, T., Doi,
H., Takumi,T., Hicks, G.G., et al.,2015. FUS/TLS deficiency causes behavioral and path-
ological abnormalities distinct from amyotrophic lateral sclerosis. Acta Neuropathol.
Commun. 3, 24.
Knowles, R.B., Sabry, J.H., Martone, M.E., Deerinck, T.J., Ellisman, M.H., Bassell, G.J., Kosik,
K.S., 1996. Translocation of RNA granul es in living neurons. J. Neurosci. 16,
7812–7820.
Kwiatkowski Jr., T.J., Bosco, D.A., Leclerc, A.L., Tamrazian, E., Vanderburg, C.R., Russ, C.,
Davis, A., Gilchrist, J., Kasarskis, E.J., Munsat, T., et al., 2009. Mutations in the FUS/
TLS gene on chromosome 16 cause familial amyotrophic lateral sclerosis . Science
323, 1205–1208.
Majounie,E., Renton, A.E., Mok,K., Dopper, E.G., Waite, A., Rollinson, S., Chio, A., Restagno,
G., Nicolaou, N. , Simon-Sanchez, J., et al., 2012. Frequency of the C9orf72
114 G. Shiihashi et al. / EBioMedicine 24 (2017) 102–115
hexanucleotide repeat expansion in patients with amyotrophic lateral sclerosis and
frontotemporal dementia: a cross-sectional study. Lancet Neurol. 11, 323–330.
Morgan, D., Diamo nd, D.M., Gottschall, P.E., Ugen, K. E., Dickey, C., Hardy, J., Duff, K.,
Jantzen, P., DiCarlo, G., Wilcock, D., et al., 2000. A beta peptide vaccination prevents
memory loss in an animal model of Alzheimer's disease. Nature 408, 982–985.
Munoz, D.G., Neumann, M. , Kusaka, H., Yokota, O., Ishihara, K., Terada, S., Ku roda, S.,
Mackenzie, I.R. , 2009. FUS pathology in basophilic inclusion body dis ease. Acta
Neuropathol. 118, 617–627.
Murakami, T., Qamar, S., Lin, J.Q., Schierle, G.S., Rees, E., Miyashita, A., Costa, A.R., Dodd,
R.B., Chan, F.T., Michel, C.H., et al., 2015. ALS/FTD mutation-induced phase transition
of FUS liquid droplets and reversible hydrogels into irreversible hydrogels impairs
RNP granule function. Neuron 88, 678–690.
Neumann, M., Sampathu, D.M., Kwong, L.K., Truax, A.C., Micsenyi, M.C., Chou, T.T., Bruce,
J., Schuck, T., G rossman, M., Clark, C.M., et al., 2006. Ubiquitinated TDP-43 in
frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science 314,
130–133.
Neumann, M., Rademakers, R., Roeber, S., Baker, M., Kretzschmar, H.A., Mackenzie, I.R.,
2009a. A new subtype of frontotemporal lobar degenerat ion with FUS pathology.
Brain 132, 2922–2931.
Neumann, M., Roeber, S., Kretzschmar, H.A., Rademakers, R., Baker, M., Mackenzie, I.R.,
2009b. Abundant FUS-immunoreactive pathology in neuronal intermediate filament
inclusion disease. Acta Neuropathol. 118, 605–616.
Neumann, M., Valori, C.F., Ansorge, O., Kretzschmar, H.A., Munoz, D.G.,Kusaka, H., Yokota,
O., Ishihara, K., Ang, L.C., Bilbao,J.M., et al., 2012. Transportin 1 accumulates specifical-
ly with FET proteins but no other transportin cargos in FTLD-FUS and is absent in FUS
inclusions in ALS with FUS mutations. Acta Neuropathol. 124, 705–716.
Pennanen,L., Gotz, J., 2005. Different tau epitopes define Abeta42-mediatedtau insolubil-
ity. Biochem. Biophys. Res. Commun. 337, 1097–1101.
Pernia-Andrade, A.J., Goswami, S.P., Stickler, Y., Frobe, U., Schlogl, A., Jonas, P., 2012. A
deconvolution-based method with high sensitivity and temporal resolution for de-
tection of sponta neous synaptic currents in vitro and in vivo. Biophys. J. 103,
1429–1439.
Qiu, H., Lee, S., Shang, Y., Wang, W.Y., Au, K.F., Kamiya, S., Barmada, S.J., Finkbeiner, S., Lui,
H., Carlton, C.E., et al.,2014. ALS-associated mutationFUS-R521C causes DNA damage
and RNA splicing defects. J. Clin. Invest. 124, 981–999.
Roy, D.S., Arons, A., Mitchell, T.I.,Pignatelli, M., Ryan, T.J., Tonegawa, S., 2016. Memory re-
trieval by activating engram cells in mouse models of early Alzheimer's disease. Na-
ture 531, 508–512.
Scekic-Zahirovic, J., Sendscheid, O., El Oussini, H., Jambeau, M., Sun, Y., Mersmann, S.,
Wagner, M., Dieterle, S., Sinniger, J., Dirrig-Grosch, S., et al., 2016. Toxic gain of func-
tion from mutant FUS proteinis crucial to trigger cell autonomous motorneuron loss.
EMBO J. 35, 1077–1097.
Scekic-Zahiro vic, J., Oussini, H.E., Mersmann, S., Drenner, K., Wagner, M., Sun, Y. ,
Allmeroth, K., Dieterle, S., Sinniger, J., Dirrig-Grosch, S., et al., 2017. Motor neuron
intrinsic and extrinsic mechanisms contribute to the pathogenesis of FUS-associated
amyotrophic lateral sclerosis. Acta Neuropathol. 133, 887–906.
Schmidt, E.K., Clavarino, G., Ceppi, M., Pierre, P., 2009. SUnSET, a nonradioactive method
to monitor protein synthesis. Nat. Methods 6, 275–277.
Sharma, A., Lyas hchenko, A.K., Lu, L., Nasr abady, S.E., Elmaleh, M., Mendelsohn, M.,
Nemes, A., Tapia, J.C., Mentis, G.Z., Shneider, N.A., 2016. ALS-associated mutant FUS
induces selective motor neuron degeneration through toxic gain of function. Nat.
Commun. 7, 10465.
Shiihashi, G., Ito, D., Yagi, T., Nihei, Y., Ebine, T., Suzuki, N., 2016. Mislocated FUS is suffi-
cient for gain-of-toxic-function amyotrophic lateral sclerosis phenotypes in mice.
Brain 139, 238 0–2394.
St Johnston, D., Beuchle,D., Nusslein-Volhard, C., 1991. Staufen,a gene required to localize
maternal RNAs in the Drosophila egg. Cell 66, 51–63.
Suh, J., Rivest, A.J., Nakashiba, T., Tominaga, T., Tonegawa, S., 2011. Entorhinal cortexlayer
III input to the hippocampus is crucialfor temporal association memory.Science 334,
1415–1420.
Udagawa, T., Fujioka,Y., Tanaka, M., Honda, D., Yokoi, S., Riku, Y., Ibi, D., Nagai, T., Yamada,
K.,Watanabe,H.,etal.,2015.FUS regulates AMPA receptor function andFTLD/ALS-as-
sociated behaviour via GluA1 mRNA stabilization. Nat. Commun. 6, 7098.
Urwin, H., Josephs, K.A., Rohrer, J.D., Mackenzie, I.R., Neumann, M., Authier, A., Seelaar, H.,
Van Swieten, J.C., Brown, J.M., Johannsen, P., et al., 2010. FUS pathology defines the
majority of tau- and TDP-43-negati ve frontotempor al lobar degeneration. Acta
Neuropathol. 120, 33–41.
Vance, C., Rogelj, B., Hortobagyi, T., De Vos, K.J., Nishimura, A.L., Sreedharan, J., Hu, X.,
Smith, B., Ruddy, D., Wright, P., et al., 2009. Mutations in FUS, an RNA processing pro-
tein, cause familial amyotrophic lateral sclerosis type 6. Science 323, 1208–1211.
Waibel, S., Neumann, M., Rabe, M., Meyer, T., Ludolph, A.C., 2010. Novel missense and
truncating mutations in FUS/TLS in familial ALS. Neurology 75, 815–817.
Yan, J., Deng, H.X., Siddique, N., Fecto,F., Chen, W., Yang, Y., Liu, E., Donkervoort, S., Zheng,
J.G., Shi, Y., et al., 2010. Frameshift and novel mutations in FUS in familial amyotro-
phic lateral sclerosis and ALS/dementia. Neurology 75, 807–814.
Yasuda, K.,Zhang, H., Loiselle, D., Haystead, T., Macara,I.G., Mili, S., 2013.The RNA-binding
protein Fus directs translation of localized mRNAs in APC-RNP granules. J. Cell Biol.
203, 737–746.
Zhang, Y., O'Connor, J.P., Siomi, M.C., Srinivasan, S., Dutra, A., Nussbaum, R.L., Dreyfuss,G.,
1995. The fragile X mental retardation syndrome protein interacts with novel homo-
logs FXR1 and FXR2. EMBO J. 14, 5358–5366.
Zou, Z.Y., Cui, L.Y., Sun, Q., Li, X.G., Liu, M.S., Xu, Y., Zhou, Y., Yang, X.Z., 2013. De novo FUS
gene mutations are associated with juvenile-onset sporadic amyotrophic lateral scle-
rosis in China. Neurobiol. Aging 34 (1312), e1311–1318.
115G. Shiihashi et al. / EBioMedicine 24 ( 2017) 102–115