Available via license: CC BY 4.0
Content may be subject to copyright.
Pharmacological targeting of CSF1R inhibits
microglial proliferation and prevents the
progression of Alzheimer’s-like pathology
Adrian Olmos-Alonso,
1,
* Sjoerd T. T. Schetters,
1,
* Sarmi Sri,
1
Katharine Askew,
1
Renzo Mancuso,
1
Mariana Vargas-Caballero,
1,2
Christian Holscher,
3
V. Hugh Perry
1
and
Diego Gomez-Nicola
1
*These authors contributed equally to this work.
The proliferation and activation of microglial cells is a hallmark of several neurodegenerative conditions. This mechanism is
regulated by the activation of the colony-stimulating factor 1 receptor (CSF1R), thus providing a target that may prevent the
progression of conditions such as Alzheimer’s disease. However, the study of microglial proliferation in Alzheimer’s disease and
validation of the efficacy of CSF1R-inhibiting strategies have not yet been reported. In this study we found increased proliferation
of microglial cells in human Alzheimer’s disease, in line with an increased upregulation of the CSF1R-dependent pro-mitogenic
cascade, correlating with disease severity. Using a transgenic model of Alzheimer’s-like pathology (APPswe, PSEN1dE9; APP/PS1
mice) we define a CSF1R-dependent progressive increase in microglial proliferation, in the proximity of amyloid-bplaques.
Prolonged inhibition of CSF1R in APP/PS1 mice by an orally available tyrosine kinase inhibitor (GW2580) resulted in the blockade
of microglial proliferation and the shifting of the microglial inflammatory profile to an anti-inflammatory phenotype.
Pharmacological targeting of CSF1R in APP/PS1 mice resulted in an improved performance in memory and behavioural tasks
and a prevention of synaptic degeneration, although these changes were not correlated with a change in the number of amyloid-b
plaques. Our results provide the first proof of the efficacy of CSF1R inhibition in models of Alzheimer’s disease, and validate the
application of a therapeutic strategy aimed at modifying CSF1R activation as a promising approach to tackle microglial activation
and the progression of Alzheimer’s disease.
1 Centre for Biological Sciences, University of Southampton, Southampton, UK
2 Institute for Life Sciences, University of Southampton, Southampton, UK
3 Division of Biomedical and Life Sciences, Faculty of Health and Medicine, Lancaster University, Lancaster, LA1 4YQ, UK
Correspondence to: Diego Gomez-Nicola Ph.D.,
Centre for Biological Sciences, University of Southampton,
South Lab and Path Block, Mail Point 840 LD80C,
Southampton General Hospital,
Tremona Road, Southampton, SO16 6YD,
UK
E-mail: d.gomez-nicola@soton.ac.uk
Keywords: Alzheimer’s disease; microglia; gliosis; neurodegeneration; inflammation
Abbreviation: BrdU = bromodeoxyuridine
doi:10.1093/brain/awv379 BRAIN 2016: 139; 891–907 |891
Received July 8, 2015. Revised October 12, 2015. Accepted October 29, 2015. Advance Access publication January 8, 2016
ßThe Author (2016). Published by Oxford University Press on behalf of the Guarantors of Brain.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse,
distribution, and reproduction in any medium, provided the original work is properly cited.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Introduction
The neuropathology of Alzheimer’s disease shows a robust
innate immune response characterized by the presence of
activated microglia, with increased or de novo expression
of diverse macrophage antigens (Akiyama et al., 2000;
Edison et al., 2008), and production of inflammatory cyto-
kines (Dickson et al., 1993; Fernandez-Botran et al., 2011).
Evidence indicates that non-steroidal anti-inflammatory
drugs (NSAIDs) protect from the onset or progression of
Alzheimer’s disease (Hoozemans et al., 2011), suggestive of
the idea that inflammation is a causal component of the
disease rather than simply a consequence of the neurode-
generation. In fact, inflammation (Holmes et al., 2009),
together with tangle pathology (Nelson et al., 2012) or
neurodegeneration-related biomarkers (Wirth et al., 2013)
correlate better with cognitive decline than amyloid-baccu-
mulation, but the underlying mechanisms of the sequence
of events that contribute to the clinical symptoms are
poorly understood. The contribution of inflammation to
disease pathogenesis is supported by recent genome-wide
association studies, highlighting immune-related genes
such as CR1 (Jun et al., 2010), TREM2 (Guerreiro et al.,
2013; Jonsson et al., 2013) or HLA-DRB5–HLA-DRB1 in
association with Alzheimer’s disease (European Alzheimer’s
Disease et al., 2013). Additionally, a growing body of evi-
dence suggests that systemic inflammation may interact
with the innate immune response in the brain to act as a
‘driver’ of disease progression and exacerbate symptoms
(Holmes et al., 2009, 2011).
Microglial cells are the master regulators of the neuroin-
flammatory response associated with brain disease (Gomez-
Nicola and Perry, 2014a,b). Activated microglia have been
demonstrated in transgenic models of Alzheimer’s disease
(LaFerla and Oddo, 2005; Jucker, 2010) and have been
recently shown to dominate the gene expression landscape
of patients with Alzheimer’s disease (Zhang et al., 2013).
Recently, microglial activation through the transcription
factor PU.1 has been reported to be capital for the progres-
sion of Alzheimer’s disease, highlighting the role of micro-
glia in the disease-initiating steps (Gjoneska et al., 2015).
Results from our group, using a murine model of chronic
neurodegeneration (prion disease), show large numbers of
microglia with an activated phenotype (Perry et al., 2010)
and a cytokine profile similar to that of Alzheimer’s disease
(Cunningham et al., 2003). The expansion of the microglial
population during neurodegeneration is almost exclusively
dependent upon proliferation of resident cells (Gomez-
Nicola et al., 2013, 2014a;Liet al., 2013). An increased
microglial proliferative activity has also been described in a
mouse model of Alzheimer’s disease (Kamphuis et al.,
2012) and in post-mortem samples from patients with
Alzheimer’s disease (Gomez-Nicola et al., 2013, 2014b).
This proliferative activity is regulated by the activation of
the colony stimulating factor 1 receptor (CSF1R; Gomez-
Nicola et al., 2013). Pharmacological strategies inhibiting
the kinase activity of CSF1R provide beneficial effects on
the progression of chronic neurodegeneration, highlighting
the detrimental contribution of microglial proliferation
(Gomez-Nicola et al., 2013). The presence of a microglial
proliferative response with neurodegeneration is also sup-
ported by microarray analysis correlating clinical scores of
incipient Alzheimer’s disease with the expression of Cebpa
and Spi1 (PU.1), key transcription factors controlling
microglial lineage commitment and proliferation (Blalock
et al., 2004). Consistent with these data, Csf1r is upregu-
lated in mouse models of amyloidosis (Murphy et al.,
2000), as well as in human post-mortem samples from pa-
tients with Alzheimer’s disease (Akiyama et al., 1994).
Although these ideas would lead to the evaluation of the
efficacy of CSF1R inhibitors in Alzheimer’s disease, we
have little evidence regarding the level of microglial prolif-
eration in Alzheimer’s disease or the effects of CSF1R tar-
geting in animal models of Alzheimer’s disease-like
pathology. In this study, we set out to define the microglial
proliferative response in both human Alzheimer’s disease
and a mouse model of Alzheimer’s disease-like pathology,
as well as the activation of the CSF1R pathway. We pro-
vide evidence for a consistent and robust activation of a
microglial proliferative response, associated with the acti-
vation of CSF1R. We provide proof-of-target engagement
and efficacy of an orally available CSF1R inhibitor
(GW2580), which inhibits microglial proliferation and par-
tially prevents the pathological progression of Alzheimer’s
disease-like pathology, supporting the evaluation of
CSF1R-targeting approaches as a therapy for Alzheimer’s
disease.
Materials and methods
Study design
Sample size
The most suitable statistical test for our samples and experi-
ments, which reach the assumptions of normality and homo-
scedasticity is the one-way or two-way ANOVA. After
performing power calculations, to achieve a significant differ-
ence of P50.05, in light of a retrospective analysis of our
previous results (Gomez-Nicola et al., 2013, 2014a,b), we
needed a minimum of n=4 (immunohistochemistry,
RT-PCR) and n=6 (behaviour) to reach a power between
0.80–0.90, depending on the specific experimental conditions.
The calculations are the customary ones based on normal dis-
tributions and were performed following statistical advice from
the Research design and methodology Department of the
University of Southampton.
Data inclusion and exclusion
For RNA expression experiments performed in human post-
mortem samples, we excluded samples showing a low yield
and quality of recovered RNA, judged as having a difference
in the expression of the four selected housekeeping genes
higher than 5 Ct values and poor cross-correlation when
892 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
compared to the other samples. No samples were excluded
from the experiments using APP/PS1 mice.
Randomization and blinding
The experiments were designed in compliance with the
ARRIVE guidelines, including control groups for all experi-
ments, randomizing the procedures and applying double-
blinded analysis when possible.
Animals
APPswe/PSEN1dE9 mice (APP/PS1) on a C57BL/6 background
were originally obtained from The Jackson Laboratory
(McClean et al., 2011). Heterozygous males were bred with
wild-type female C57BL/6J (Harlan) at our local facilities.
APP/PS1/Macgreen (c-fms EGFP) mice maintained at our
local facilities, after breeding c-fms EGFP mice (Sasmono
et al., 2003) with the APP/PS1 line, allowing the Macgreen
transgene to be expressed in heterozygosis. Offspring were
ear punched and genotyped using PCR with primers specific
for the APP-sequence (forward: GAATTCCGACATGA
CTCAGG, reverse: GTTCTGCTGCATCTTGGACA). Mice
not expressing the transgene were used as wild-type controls.
Mice were housed in groups of 4 to 10, under a 12-h light/12-
h dark cycle at 21C, with food and water ad libitum.
For the evaluation of the effects of treatment with GW2580,
mice were 6 months of age when treatment began (n=6–8).
Mice were fed with a control diet (RM1) or a diet containing
GW2580 [Modified LabDietÕPicoLab EURodent Diet 14%,
5L0W (5LF2) with 0.1% (1000 ppm) GW2580 (LC
Laboratories); TestDiet] for 3 months, before behavioural
tasks were conducted. Alternatively, to test the effects of
increasing doses of GW2580 on microglial survival,
GW2580 (LC Laboratories) was suspended in 0.5% hydroxy-
propylmethylcellulose and 0.1% Tween 80 and was dosed
orally at 0.2 ml per mouse (75 mg/kg), daily for five consecu-
tive days to wild-type mice. Mice weight was monitored during
all experiments. Mice received one injection of intraperitoneal
bromodeoxyuridine (BrdU) (Sigma-Aldrich; 7.5 mg/ml, 0.1 ml/
10 g weight in sterile saline), 1 day before the end of the ex-
periment. All procedures were performed in accordance with
UK Home Office licensing.
Post-mortem samples of
Alzheimer’s disease
For immunohistochemical analysis, human brain autopsy
tissue samples (temporal cortex, paraffin-embedded, formalin-
fixed, 96% formic acid-treated, 6-mm sections) from the
National CJD Surveillance Unit Brain Bank (Edinburgh, UK)
were obtained from cases of Alzheimer’s disease (five females
and five males, age 58–76) or age-matched controls (four fe-
males and five males, age 58–79), in whom consent for use of
autopsy tissues for research had been obtained. All cases ful-
filled the criteria for the pathological diagnosis of Alzheimer’s
disease. Ethical permission for research on autopsy materials
stored in the National CJD Surveillance Unit was obtained
from Lothian Region Ethics Committee.
For mRNA analysis, human brain autopsy tissue samples
(temporal cortex, fresh-frozen tissue) were obtained from the
Human Tissue Authority licensed South West Dementia Brain
Bank, University of Bristol (UK). Samples were selected from
Alzheimer’s disease cases and age-matched controls
(Supplementary Table 1). Ethical permission for research on
autopsy materials stored in the South West Dementia Brain
Bank was obtained from Local Ethics Committee.
Behavioural tests
APP/PS1 or wild-type mice treated with control (RM1) or
GW2580 diet, as stated earlier, were tested on behavioural
tasks at 9 months of age (3 months into the diet): open-field
locomotor and exploratory activity and burrowing activity
(sickness behaviour). For the behavioural analysis, n=8–21
was used.
The open-field tests were carried out using activity monitor
software (Med Associated Inc.). The mice were placed in indi-
vidual cages of 27 27 0.3 cm for a period of 5 min, to
further analyse the total distance travelled (cm) and the
number of rears (vertical counts), using the average speed as
an internal control of the mouse motor abilities, during the test
period (5 min). For measuring anxiety-related behaviour, dif-
ferential exploratory activity (distance travelled or number of
entries) was analysed in a residual and an open zone, using
activity monitor software (Med Associated Inc.), as previously
described (Rothman et al., 2012).
For burrowing behaviour, plastic cylinders, 20-cm long and
6.8-cm in diameter were filled with 190 g of normal diet food
pellets and placed in individual mouse cages. Mice were placed
individually in the cages overnight, weighting the remaining
pellets at the end of each session, and calculating the
amount displaced (‘burrowed’). The mice were returned then
to their home cage. Body weights and diet consumption of all
mice were monitored during the course of the experiment.
Discrete trial spontaneous alternation in the T-maze was
performed as previously described (Deacon and Rawlins,
2006). The apparatus for this test consisted of a grey
T-shaped maze with 30 10 29-cm arms, with a central
partition extending 7 cm into the start arm from the back of
the maze, and two guillotine doors each having the potential
to block off the left or right goal arms. Mice were placed in
the start arm of the maze (facing the wall) and allowed to
make a choice to enter the left or right goal arm. Following
their choice, they were then enclosed in that arm for 30s to
facilitate habituation by sliding down the appropriate guil-
lotine door. Mice were then taken out of the maze, the cen-
tral partition removed and guillotine doors reopened. Mice
were once again placed in the start arm and allowed to
make another free choice of either goal arm. Whether or
not the mouse alternated was noted, with a score of 1
given if the mouse visited the other arm reflecting explora-
tory behaviour and memory of the first choice on its second
trial, and a score of 0 given if the mouse went to the same
arm on its second trial. The alternation ratio was obtained
as the mean of the 20 scores. If animals did not move in 90 s
or less these were considered failed trials and were not
scored; however, these did not account for 45% of all
trials and there was no effect of treatment or genotype for
missed trials (P= 0.97). The test was performed four times a
day for 5 days (a total of 20 trials) with an average spacing
of 2 h in between each trial.
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |893
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Immunohistochemistry
Coronal hippocampal sections were cut from paraformalde-
hyde-fixed, frozen or fresh brains. Mice perfusion, tissue
processing and immunohistochemical analysis was performed
as previously described (Gomez-Nicola et al., 2008, 2013),
using the following primary antibodies: rabbit anti-Iba1
(Wako), mouse anti-human Ki67 (Dako), mouse anti-BrdU
(DSHB), mouse anti-amyloid-b(6E10; Covance), mouse anti-
synaptophysin (SY38; Merck Millipore), rabbit anti-CSF1R
(Santa Cruz Biotechnologies) and rabbit anti-PU.1 (Cell
Signaling). Following primary antibody incubation, the sec-
tions were washed and incubated with the appropriate bioti-
nylated secondary antibody (Vector Labs), and/or with the
appropriate Alexa 405, 488 or 568 conjugated secondary anti-
body or streptavidin (Molecular Probes). For co-labelling of
Iba1 and Ki67 in human tissue, following primary antibody,
sections were incubated with an anti-rabbit biotinylated sec-
ondary antibody (for Iba1 detection) and the ImmPRESS-AP
Anti-Mouse (alkaline phosphatase) Polymer Detection Kit (for
Ki67 detection), as previously described (Gomez-Nicola and
Boche, 2015; Gomez-Nicola and Perry, 2016). For light mi-
croscopy, the sections were visualized using 3,3’-diaminoben-
zidine (DAB) precipitation or BCIP/NBT AP reaction, in a
Leica CTR 5000 microscope, coupled to a Leica DFC300FX
microscope camera. For PU.1 visualization, the DAB signal
was enhanced with 0.05% nickel ammonium sulphate, produ-
cing a black precipitate. After immunofluorescence labelling,
nuclei were visualized by DAPI staining and the sections
were mounted with Mowiol/DABCO (Sigma-Aldrich) mixture.
The sections were visualized on a Leica TCS-SP5 confocal
system, coupled to a Leica CTR6500 microscope.
The general immunohistochemistry protocol was modified
for the detection of BrdU, adding a DNA denaturation step
with 2 N HCl (30 min, 37C), as previously described (Gomez-
Nicola et al., 2011, 2013). For the detection of amyloid-b
plaques (6E10), sections were pretreated with 95% formic
acid (Sigma-Aldrich) for 10 min at room temperature. All
other histological stains (i.e. Congo red) were done according
to standard laboratory procedures.
The protocol used for immunohistochemistry on human sec-
tions was a modification of the general protocol
(DAB + alkaline phosphatase), with antigen unveiling in citrate
buffer being performed for 25 min, as previously described
(Gomez-Nicola et al., 2014b).
Golgi-Cox staining
A subgroup of APP/PS1 or wild-type mice treated with control
(RM1) or GW2580 diet (n=4), as stated before, were deeply
anaesthetized with sodium pentobarbital and then transcar-
dially perfused with artificial CSF. Brains were then rapidly
dissected and sliced with a vibrating microtome (200 mm;
Leica). The hippocampal slices were incubated in rapid
Golgi-Cox solutions, following manufacturer’s instructions
(FD Rapid Golgi Stain Kit, FD Neurotechnologies).The
slices were infused with a solution containing potassium di-
chromate, potassium chromate and mercuric chloride for 2
weeks, to be further developed into the Golgi-Cox staining
on free-floating plates. The slices were mounted onto gelati-
nized slides, dried, dehydrated, cleared with xylene and
mounted with DPX. Golgi-treated slices were analysed with
a Leica CTR 5000 microscope, coupled to a Leica
DFC300FX microscope camera. For the dendritic linear spine
density, Golgi-Cox-labelled apical dendritic processes of CA1
neurons were analysed.
Quantification and image analysis
The quantification of antigen positive cells (i.e. PU.1
+
) in the
cerebral cortex (n=4 fields/mouse, n=4–8 mice/group) was
performed after DAB immunohistochemistry. The number of
double positive cells (i.e. Iba1
+
BrdU
+
) in the specific area
(n=4 fields/mouse, n=4–8mice/group) was performed after
double immunofluorescence or double immunohistochemistry
with DAB/AP. Data were represented as number of positive
cells/mm
2
. The quantification of antigen-positive cells (i.e.
Iba1
+
BrdU
+
or Iba1
+
) in human brains was performed in
the white or grey matter of the temporal cortex after DAB/
AP immunohistochemistry (n=20 fields/brain, n=9–10
brains/group). The quantification of the distribution of micro-
glial cells (PU.1
+
) around amyloid-bplaques was performed
with an adapted version of the Sholl analysis, modified from
Frautschy et al. (1998). Briefly, we analysed all amyloid-b
plaques visible with Congo red staining (n=4–6 sections/
mouse, n=4–8 mice/group), tracing concentric circles starting
from the diameter of each individual plaque and setting radius
step size at 20 mm. The final radius was set when (i) an indi-
vidual plaque came into contact with a neighbouring plaque;
(ii) cell density reached wild-type levels; or (iii) a border of the
tissue was reached. PU.1
+
cells density (cells/mm
2
) contained
within each circle was quantified, considering that cells falling
at the interphase of two radii were counted as belonging to the
section containing 450% of the cell. The quantification of the
intensity of signal (i.e. synaptophysin) was performed after
immunofluorescence, and presented as per cent stained area.
The quantification of enhanced green fluorescent protein
(EGFP) intensity per cell in APP/PS1/Macgreen mice was per-
formed on confocal stacks, using a constant sampled area. All
quantifications were performed with the help of the ImageJ
image analysis software.
Analysis of gene expression by
reverse transcriptase PCR
APP/PS1 or wild-type mice treated with control (RM1) or
GW2580 diet (n=4–6/group) were processed to obtain samples
from the cortex by dissection under a microscope, after intra-
cardiac perfusion with heparinized 0.9% saline. RNA was ex-
tracted using TRIzolÕ(Life Technologies), quantified using
Nanodrop (Thermo Scientific), and reverse transcribed using
the iScript
TM
cDNA Synthesis Kit (Bio-Rad) following manufac-
turer’s instructions, after checking its integrity by electrophoresis
in a 2% agarose gel. cDNA libraries were analysed by quanti-
tative polymerase chain reaction (PCR) using the iTaq
TM
Universal SYBRÕGreen supermix (Bio-Rad) and the following
custom designed gene-specific primers (Sigma-Aldrich): csf1
(NM_007778.4; FW, agtattgccaaggaggtgtcag, RV, atctggcat-
gaagtctccattt), il34 (NM_001135100.1; FW, ctttgggaaacga-
gaatttggaga, RV, gcaatcctgtagttgatggggaag), csf1r
(NM_001037859.2; FW, gcagtaccaccatccacttgta, RV, gtgaga-
cactgtccttcagtgc), pu.1 (NM_011355.1; FW, cagaagggcaaccg-
caagaa, RV, gccgctgaactggtaggtga), c/ebpa (NM_007678.3;
894 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
FW, agcttacaacaggccaggtttc, RV, cggctggcgacatacagtac), runx1
(NM_001111021; FW, caggcaggacgaatcacact, RV, ctcgtgctg
gcatctctcat), irf8 (NM_008320; FW, cggggctgatctgggaaaat,
RV, cacagcgtaacctcgtcttc), il1b (NM_008361.3; FW, gaaatgc-
caccttttgacagtg, RV, tggatgctctcatcaggacag), tgfb (NM_011577;
FW, tgtacggcagtggctgaacc, RV, cgtttggggctgatcccgtt), il6
(NM_031168; FW, tagtccttcctaccccaatttcc, RV, ttggtccttagc-
cactccttc) or igf1 (NM_010512; FW, agatctgcctctgtgacttcttga,
RV, agcctgtgggcttgttgaagt). Quality of the primers and the
PCR reaction were evaluated by electrophoresis in a 1.5% agar-
ose gel, checking the PCR product size. Data were analysed
using the 2-Ct method with Primer Opticon 3 software,
using Gapdh (NM_008084.2; FW, tgaacgggaagctcactgg, RV,
tccaccaccctgttgctgta) as a housekeeping gene.
Frozen samples from Alzheimer’s disease cases or age-
matched controls (Supplementary Table 1) were processed
for RNA extraction and quantitative PCR analysis. RNA
was extracted using the RNAqueous-Micro Kit (Life
Technologies), quantified using Nanodrop (Thermo
Scientific), reverse transcribed using the iScript cDNA
Synthesis Kit (Bio-Rad), following the manufacturer’s instruc-
tions, after checking its integrity by electrophoresis on a 1.8%
agarose gel. Low quality and purity RNA samples were
excluded from consequent experimentation. cDNA libraries
were analysed by quantitative PCR using 96-wells custom-de-
signed TaqManÕarray plates with the 7500 Real-Time PCR
system (Applied Biosystems). Quality of the PCR reaction end
product was evaluated by electrophoresis in a 1.5% agarose
gel. Raw CT data were obtained from the SDS v.2.0.6 soft-
ware and normalized to the normalization factor (geometric
mean of four housekeeping genes; GAPDH,HPRT,18S and
GUSB) using the 2-CT method.
Antibody arrays
APP/PS1 or wild-type mice treated with control (RM1) or
GW2580 diet (n=4–6/group) were processed to obtain sam-
ples from the cortex by dissection under a microscope, after
intracardiac perfusion with heparinized 0.9% saline. Tissues
were homogenized in RIPA buffer and protein concentration
was quantified with a BCA protein assay kit (Pierce). The
concentration of 40 mouse cytokines and chemokines was
analysed with Mouse Quantibody Cytokine Arrays Q5
(QAM-CYT-5; Raybiotech) by the Raybiotech testing service,
according to manufacturer’s instructions. The concentration of
every cytokine was quantified using internal protein standards,
discarding data not reaching the detection threshold (LOD,
limit of detection). Data were represented as fold-change of
APP/PS1 + RM1 versus wild-type + RM1 and APP/
PS1 + GW2580 versus wild-type + GW2580.
Multiplex analysis of soluble
amyloid-b
38
, amyloid-b
40
and
amyloid-b
42
Protein samples obtained and quantified as above were ana-
lysed with an AbV-plex triple ultra-sensitive assay kit accord-
ing to the manufacturer’s instructions (Meso Scale Discovery).
Standards (amyloid-b
1–38
, amyloid-b
1–40
and amyloid-b
1–42
)
and samples (diluted 1:20) were added to the 96-well plates,
incubated, washed, and read in a Sector Imager plate reader
(Meso Scale Discovery) immediately after addition of the Meso
Scale Discovery read buffer. Amyloid-bconcentrations were
calculated with reference to the standard curves and expressed
as nanograms per millilitre.
Statistical analysis
Data were expressed as mean standard error of the mean
(SEM) and analysed with the GraphPad Prism 5 software
package (GraphPad Software). For all datasets, normality
and homoscedasticity assumptions were reached, validating
the application of the two-way ANOVA (two variables were
analysed in all cases), followed by the Tukey post hoc test
for multiple comparisons. Relative gene expression data
from human samples was analysed using a two-tailed
Fisher t-test. Correlation of relative gene expression in
human samples to Braak stage was tested using the
Kendall tau-b rank correlation test included in the SPSS ana-
lyticsoftware(v.17;IBM).Differenceswereconsidered
significant for P50.05.
Results
Microglial proliferation is increased in
Alzheimer’s disease and correlates
with disease severity
Although an increased microglial response has been docu-
mented in Alzheimer’s disease (Bondolfi et al., 2002;
Gomez-Nicola and Perry, 2014a,b), we identified the
need for a better understanding of microglial proliferation
and the expression of the potential pro-mitogenic compo-
nents in human tissue. The analysis of post-mortem cases of
Alzheimer’s disease (temporal cortex; Supplementary Table
1) evidenced an increase in the total number of microglial
cells in cortical grey matter (118% versus non-demented
controls, not significant) and white matter (120% versus
non-demented controls, P50.05) (Fig. 1A). We also
observed an increased microglial proliferation (Fig. 1B
and C), mostly evident in the grey matter (207% versus
non-demented controls, P50.05) than in the white
matter (155% versus non-demented controls, not signifi-
cant). At the gene expression level, we found an elevated
expression of the main components of the CSF1R pro-mito-
genic pathway (CSF1R,CSF1,SPI1,CEBPA,RUNX1; Fig.
1D). Several microglial markers are upregulated in
Alzheimer’s disease, while the majority of hematopoietic
stem cell or bone marrow-derived cell markers were
found to be unchanged when compared with non-demented
controls, with CD34,CD59 and CCL2 being upregulated
(Fig. 2A and B). We also evidenced an increased expression
of markers related to cell proliferation, such as PCNA,
(Figs 1D and 2C) and the activation of an inflammatory
response characterized by TGFB expression and low levels
of proinflammatory cytokines (IL1B,IL6), consistent with
previously reported findings (Gomez-Nicola and Perry,
2014a,b) (Fig. 2D and E). The analysis of the correlation
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |895
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
of gene expression with the Braak score evidenced a signifi-
cant association of disease severity with several markers of
microglial cells [ITGAM (CD11b), ITGAX (CD11c),
CD68,CX3CR1], cell proliferation (CEBPA,CSF1,
PCNA,SPI1) and inflammation (CCL2,CEBPB,
CX3CL1,TGFB or TREM2) (Supplementary Table 2).
The expression of markers of perivascular macrophages
or bone marrow-derived cells (CD163,CCR2,CD34,
CD59) correlated with Alzheimer’s disease severity, while
we found no correlation with hematopoietic stem cell mar-
kers (KIT,MYB,ATXN1; Supplementary Table 2).
These findings highlight the relevance of microglial
proliferation and activation for the progression of
Alzheimer’s disease, and support the study of these mech-
anisms in animal models of Alzheimer’s disease-like
pathology.
Figure 1 Characterization of the microglial proliferative response in Alzheimer’s disease. (A–C) Immunohistochemical analysis and
quantification of the number of total microglial cells (Iba1
+
;A) or proliferating microglial cells (Iba1
+
Ki67
+
;B) in the grey (GM) and white matter
(WM) of the temporal cortex of Alzheimer’s disease cases (AD) and age-matched non-demented controls (NDC). (C) Representative pictures of
the localization of a marker of proliferation (Ki67, dark blue) in microglial cells (Iba1
+
, brown) in the grey matter of the temporal cortex of
non-demented controls or Alzheimer’s disease cases. (D) RT-PCR analysis of the mRNA expression of CSF1R,CSF1,IL34,SPI1 (PU.1), CEBPA,
RUNX1 and PCNA in the temporal cortex of Alzheimer’s disease cases and age-matched non-demented controls. Expression of mRNA repre-
sented as mean SEM and indicated as relative expression to the normalization factor (geometric mean of four housekeeping genes; GAPDH,
HPRT,18S and GUSB) using the 2-CT method. Statistical differences: *P50.05, **P50.01, ***P50.001. Data were analysed with a two-way
ANOVA and a post hoc Tukey test (Aand B) or with a two-tailed Fisher t-test (D). Scale bar in C=50mm.
896 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Figure 2 Gene expression analysis in human post-mortem Alzheimer’s disease cases and age-matched controls. Samples from Alzheimer’s disease (filled circles) cases or
age-matched controls (NDC, open circles) were analysed by quantitative PCR for the expression of microglia/macrophage markers (A), hematopoietic stem cell/bone marrow-derived cell (HSCs/
BMCs) markers (B), cell cycle activation markers (C), inflammation markers (D) or other (E). Samples were analysed using custom-designed TaqMan
Õ
array plates with the 7500 Real-Time PCR
system (Applied Biosystems). Expression of mRNA represented as mean SEM and indicated as relative expression to the normalization factor (geometric mean of four housekeeping genes;
GAPDH,HPRT,18S and GUSB) using the 2-CT method. Statistical differences: *P50.05, **P50.01, ***P50.001. Data were analysed with a two-tailed Fisher t-test.
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |897
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Increased microglial proliferation and
CSF1R activity are closely associated
with the progression of Alzheimer’s
disease-like pathology
We investigated the proliferative dynamics of microglial
cells and the expression of the components of the CSF1R
pathway in a relevant model of Alzheimer’s disease-like
pathology (APPswe, PSEN1dE9; APP/PS1). While micro-
glial accumulation around amyloid-bplaques has been
documented (Brawek et al., 2014), we took advantage of
the specificity of PU.1 expression in microglia to quantify
microglial numbers, as it allows an accurate identification
of microglial nuclei (Fig. 3A) (Gomez-Nicola et al., 2013,
2014a). Microglial numbers increase progressively in APP/
PS1 mice (from 9 to 14 months of age), when compared to
wild-type littermates, accumulating in the vicinity of amyl-
oid-bplaques (Fig. 3A). We observed that while the
number of plaques in APP/PS1 mice increases with age
(APP/PS1 9 months = 4.41 0.39 versus APP/PS1 14
months = 7.25 1.82 plaques/mm
2
), neither their size nor
the pattern of microglial distribution change (Fig. 3B).
Microglial proliferation was evidenced by the incorporation
of BrdU in Iba1
+
cells, analysed by confocal microscopy
(Fig. 3C). The number of Iba1
+
/BrdU
+
cells increases pro-
gressively in APP/PS1 mice (Fig. 3D), in close association
with amyloid-bplaques (Fig. 3C).
Further analysis of gene expression showed upregulation
of the main components of the CSF1R pathway in APP/PS1
brains, with some genes being upregulated from 9 months
onwards (Csf1,Il34,Csf1r,Cebpa) and others from 14
months [Spi1 (Pu.1), Runx1] (Fig. 3E). These changes co-
incide with the upregulation of key inflammatory genes
associated with microglial activation (Fig. 3E).
We next analysed the localization of CSF1R in APP/PS1/
Macgreen mice (c-fms EGFP), as an optimal reporter line
for CSF1R expression, as EGFP is driven by the Csf1r pro-
moter (c-fms) (Sasmono et al., 2003). APP/PS1/Macgreen
mice showed a differential increased expression levels of
this receptor in microglial cells associated to amyloid-b
plaques, when compared to microglia distal to plaques
(Fig. 3F), with EGFP levels and distance to plaque centre
showing inversed correlation (Fig. 3G). This correlation can
also be observed when using CSF1R immunostaining
(Supplementary Fig. 1) and it is suggestive of an association
of CSF1R levels with the observed proliferative response.
Pharmacological targeting of CSF1R
activation with an orally-available
inhibitor blocks microglial prolifera-
tion in APP/PS1 mice
As CSF1R is likely to drive microglial proliferation in APP/
PS1 mice, we targeted the activation of CSF1R with a
selective inhibitor (GW2580) from 6 to 9 months of age,
to evaluate target engagement and efficacy of a CSF1R-
inhibitory approach (Fig. 5). To determine if microglia
would be the only target of CSF1R inhibitors, we used
APP/PS1/Macgreen mice as reporters of CSF1R expression.
Using immunofluorescence and confocal microscopy we
found no evidence of expression of CSF1R in neurons
(NeuN
+
) or astrocytes (GFAP
+
) (Supplementary Fig. 2),
supporting that CSF1R expression is exclusive to microglia,
as previously described (Sasmono et al., 2003; Erblich
et al., 2011; Gomez-Nicola et al., 2013). While a dose of
GW2580 of 75 mg/kg has been used in the past without
causing significant changes in the survival of microglia
(Gomez-Nicola et al., 2013), recent evidence supports
that CSF1R inhibitors could cause the depletion of the
microglial population (Elmore et al., 2014). Therefore, we
assessed the effects of repeated increasing doses of
GW2580 (75, 150, 300 mg/kg) on the survival of the
microglial population (Supplementary Fig. 3). We found
that none of the doses of GW2580 caused any significant
change in the total number of microglial cells
(Supplementary Fig. 3A and B), supporting the continued
use of the 75 mg/kg to maintain consistency with our pre-
viously published data (Gomez-Nicola et al., 2013).
Prolonged treatment with GW2580 caused a significant
reduction in the number of microglial cells and a blockade
of microglial proliferation in APP/PS1, when compared
with APP/PS1 on control diet (RM1) (Fig. 4A–C). The
treatment with the CSF1R selective inhibitor GW2580 did
not cause depletion of the microglial population in either
wild-type or APP/PS1 mice (Fig. 4A and B), suggesting spe-
cific targeting of microglial proliferation and not microglial
survival. In line with these results, GW2580 downregulated
the mRNA expression of Csf1r,Csf1,Spi1 (PU.1), Cebpa
and Runx1 in APP/PS1 mice (Fig. 4D), all involved in con-
trolling the pro-mitogenic programme. Some differences
were observed in this dataset (Fig. 4D) compared to previ-
ous ones (Fig. 3E), suggesting significant variability in this
model.
To better understand the impact of the prevention of
microglial proliferation in Alzheimer’s disease-like path-
ology we screened the expression of several inflammatory
mediators at the mRNA level (Fig. 4D) and at the protein
level using quantitative antibody arrays (Fig. 4E). Blockade
of microglial proliferation induced a significant reduction in
the expression of several inflammatory mediators at the
mRNA level (Fig. 4D). CSF1R inhibition in APP/PS1 mice
returned the expression of pro-inflammatory mediators
such as IL1A, IL12, IL17, CD30L, CCL1, CCL2 or
CXCL13 to wild-type levels, while upregulating the levels
of anti-inflammatory cytokines such as IL4, IL5 or IL13
(Fig. 4E) or the chemokine CCL9 (Fig. 4E). Blockade of
CSF1R prevented overexpression of CSF1 in APP/PS1, as
well as the downregulation of ICAM1, leptin and TNFR1
(Fig. 4E), highlighting some potentially beneficial side-
effects of CSF1R inhibition in microglial cells. Other mol-
ecules were unaffected by CSF1R inhibition, including IL7,
TNFR2, CCL11 and CCL24 (Fig. 4E).
898 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Figure 3 Characterization of the microglial proliferative response in a mouse model of Alzheimer’s disease-like pathology
(APP/PS1). (A) Immunohistochemical analysis and quantification of the number of total microglial cells (PU.1
+
; black) in the cortex of APP/PS1
and wild-type mice at 9 and 14 months of age. Amyloid-bplaques are shown in red (Congo Red). Number of microglia represented as
mean SEM of PU.1
+
cells/mm
2
.(B) Analysis of the spatial distribution of microglial cells (PU.1
+
) around amyloid-bplaques in the cortex of
APP/PS1 mice at 9 and 14 months of age, using an adapted version of the Sholl analysis (see ‘Materials and methods’ section). Number of microglia
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |899
(continued)
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
CSF1R inhibition prevents the pro-
gression of Alzheimer’s disease-like
pathology
The prolonged inhibition of CSF1R by GW2580 prevented
the behavioural deficits observed in APP/PS1 mice (Fig. 5).
Blockade of microglial proliferation with GW2580 caused
a significant recovery of the deficits in short-term memory
observed in APP/PS1 mice, analysed as spontaneous alter-
nation in the T-maze (Fig. 5A). The increased activity in the
open-field task observed in APP/PS1 was prevented by the
treatment with GW2580-containing diet (Fig. 5B). These
effects were associated with the development of a hyper-
active behaviour in APP/PS1 mice, rather than a change in
anxiety-related behaviour, as evidenced by the lack of pref-
erential use of the different zones of the arena (Fig. 5B) and
the lack of differences in the number of entries in the open
zone (Fig. 5C). A task associated with sickness behaviour
(burrowing) was found to be unchanged in APP/PS1 mice
(Fig. 5D). The treatment of wild-type or APP/PS1 mice with
a GW2580-containing diet for 3 months did not cause any
significant effects on weight gain, when compared with
mice treated with a control diet (RM1) (Fig. 5E).
One of the main pathological hallmarks in APP/PS1 mice,
the deposition of amyloid-bplaques, was found unchanged
after the treatment with GW2580, as evidenced by multi-
plex analysis of soluble amyloid-b
38
, amyloid-b
40
and
amyloid-b
42
(Fig. 6A), 6E10 immunostaining (Figs 6B and
7C) and Congo Red staining (not shown), suggesting a
requirement for microglial proliferation/activation to asso-
ciate the amyloidogenic component with the behavioural
decline observed in Alzheimer’s disease-like pathology.
However, prolonged blockade of CSF1R prevented the syn-
aptic degeneration observed in the hippocampus of APP/
PS1 mice, as evidenced by a significant recovery of synaptic
density at the hippocampus (Fig. 7A and B). This was fur-
ther confirmed by analysing spine density in CA1 neurons
by Golgi-Cox staining (Fig. 7C and D). We observed that
treatment with GW2580 caused a significant prevention
of the loss of dendritic spines observed in APP/PS1 mice
(Fig. 7C and D).
The observed beneficial effects of prolonged and specific
targeting of CSF1R, with the orally available inhibitor
GW2580, provide a proof of target engagement and effi-
cacy in a model of Alzheimer’s disease-like pathology.
Discussion
The innate immune component has a clear influence over
the onset and progression of Alzheimer’s disease. The ana-
lysis of therapeutic approaches aimed at controlling neu-
roinflammation in Alzheimer’s disease is moving forward
at the preclinical and clinical level, with several clinical
trials aimed at modulating inflammatory components of
the disease. We have previously demonstrated that the pro-
liferation of microglial cells is a core component of the
neuroinflammatory response in a model of prion disease,
another chronic neurodegenerative disease, and is con-
trolled by the activation of CSF1R (Gomez-Nicola et al.,
2013). This aligns with recent reports pinpointing the
causative effect of the activation of the microglial prolifera-
tive response on the neurodegenerative events of human
and mouse Alzheimer’s disease, highlighting the activity
of the master regulator PU.1 (Gjoneska et al., 2015). Our
results provide a proof of efficacy of CSF1R inhibition for
the blockade of microglial proliferation in a model of
Alzheimer’s disease-like pathology. Treatment with the
orally available CSF1R kinase-inhibitor (GW2580) proves
to be an effective disease-modifying approach, partially im-
proving memory and behavioural performance, and pre-
venting synaptic degeneration. These results support the
previously reported link of the inflammatory response gen-
erated by microglia in models of Alzheimer’s disease with
the observed synaptic and behavioural deficits, regardless of
amyloid deposition (Jones and Lynch, 2014). Our findings
support the relevance of CSF1R signalling and microglial
proliferation in chronic neurodegeneration and validate the
evaluation of CSF1R inhibitors in clinical trials for
Alzheimer’s disease.
Our findings show that the inhibition of microglial pro-
liferation in a model of Alzheimer’s disease-like pathology
does not modify the burden of amyloid-bplaques, suggest-
ing an uncoupling of the amyloidogenic process from the
pathological progression of the disease. Although a recent
report correlated the accumulation of microglia with the
expansion of amyloid-bplaques, suggesting that microglial
Figure 3 Continued
represented as mean SEM of PU.1
+
cells/mm
2
.(Cand D) Analysis and quantification of microglial proliferation (Iba1
+
BrdU
+
, green and red,
respectively; C) by triple immunofluorescence analysed by confocal microscopy. Amyloid-bplaques are shown in blue (6E10). (D) Microglial
proliferation represented as mean SEM of Iba1
+
BrdU
+
cells/mm
2
.(E) RT-PCR analysis of the mRNA expression of Csf1r,Csf1,Il34,Spi1 (PU.1),
Cebpa,Runx1,Tgfb,Igf1,Il1b,Il6 and Irf8 in the cortex of APP/PS1 and wild-type (WT) mice at 9 and 14 months of age. Expression of mRNA
represented as mean SEM and indicated as relative expression compared to the housekeeping gene (Gapdh) using the 2-CT method. (F)
Immunofluorescent analysis of the expression of EGFP (green) driven by the Csf1r promoter in APP/PS1/Macgreen mice, around amyloid-b
plaques in the cortex of APP/PS1 mice at 14 months of age. Amyloid-b(6E10) is shown in red. Arrowheads indicate microglia with low CSF1R
expression. (G) Correlation analysis of the expression of EGFP in individual cells in APP/PS1/Macgreen mice with the distance to amyloid-b
plaques. Statistical differences: *P50.05, **P50.01, ***P50.001. Data were analysed with a two-way ANOVA and a post hoc Tukey test (A,D
and E). Scale bars: A= 100 mm, Cand F=50mm.
900 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Figure 4 Prolonged inhibition of CSF1R blocks microglial proliferation and rescues the inflammatory alterations of APP/PS1
mice. (A–C) Immunohistochemical analysis and quantification of the number of total microglial cells (PU.1
+
; black, Aand B) and proliferating
microglial cells (Iba1
+
BrdU
+
,C) in the cortex of APP/PS1 and wild-type mice at 9 months of age, after treatment for 3 months with a control diet
(RM1) or a diet containing GW2580. Amyloid-bplaques are shown in red (Congo Red). Number of microglia represented as mean SEM of
PU.1
+
or Iba1
+
BrdU
+
cells/mm
2
.(D) RT-PCR analysis of the mRNA expression of Csf1r,Csf1,Il34,Spi1 (PU.1), Cebpa,Runx1,Tgfb,Igf1,Il1b,Il6
and Irf8 in the cortex of APP/PS1 and wild-type mice at 9 months of age, after treatment for 3 months with a control diet (RM1) or a diet
containing GW2580. Expression of mRNA represented as mean SEM and indicated as relative expression compared to the housekeeping gene
(Gapdh) using the 2-CT method. (E) Quantification of protein concentration of 40 inflammatory mediators (grouped as cytokines, growth
factors and chemokines) by Mouse Quantibody Cytokine Arrays (see ‘Materials and methods’ section), in samples from the cortex of APP/PS1 and
wild-type mice at 9 months of age, after treatment for 3 months with a control diet (RM1) or a diet containing GW2580. Protein expression
represented as fold change ( + fold change = upregulation, fold change = downregulation) of the corresponding APP/PS1 group (with RM1 or
with GW2580) over its correspondent wild-type group. When protein concentration fell below the levels of detection of the assay for more than
half of the samples, data are shown as N/D (not detectable). Statistical differences: *P50.05, **P50.01. Data were analysed with a two-way
ANOVA and a post hoc Tukey test. Scale bar: A= 100 mm.
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |901
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
cells could act as a ‘barrier’ (Condello et al., 2015), our
current data provide evidence that removing microglia does
not alter the deposition of amyloid-b. This indicates a ne-
cessity of the executive role of microglia to linking amyloid
deposition with the progression of cognitive decline. In this
direction, recent reports using Trem2
+/
mice on an APP/
PS1 background evidenced an impaired amplification of the
microglia population in response to amyloid-b, suggesting
that deficient phagocytosis could link with impaired micro-
glial proliferation (Ulrich et al., 2014). Linking TREM2
activation with CSF1R-induced proliferation could be pos-
sible through the adaptor molecule DAP12, as suggested
Figure 5 CSF1R inhibition prevents behavioural deficits in APP/PS1 mice. (A) Spontaneous alternation in the T-maze of APP/PS1 and
wild-type mice at 9 months of age, after treatment for 3 months with a control diet (RM1) or a diet containing GW2580. Alternation was
measured as % election of the alternative arm in the second test (short-term memory). (Band C) Analysis of the behaviour in the open field,
measured as total distanced travelled (B) or number of entries in the open zone (C) of APP/PS1 and wild-type (WT) mice at 9 months of age, after
treatment for 3 months with a control diet (RM1) or a diet containing GW2580. Exploratory activity was measured as distance travelled (cm) in
the open field test, analysing the locomotor activity on an open zone versus residual zone as a correlate of anxiety. (D) Burrowing behaviour, a
measure of sickness behaviour, was measured as weight displaced (g) off the tube in 24 h. (E) Analysis of the effect of the influence of genotype
(wild-type versus APP/PS1) or diet (RM1 versus GW2580) on the average weekly weight change (relative to t = 0) mice at 9 months of age, after
treatment for 3 months with a control diet (RM1) or a diet containing GW2580. Statistical differences: *P50.05, **P50.01. Data were analysed
with a two-way ANOVA and a post hoc Tu k e y ( B–E) or Bonferroni (A) test.
902 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
for macrophages (Otero et al., 2009), although the func-
tional connection of amyloid-bphagocytosis and microglial
proliferation has not been reported to date. Our results are
in agreement with previously reported data using CD11b-
HSVTK mice under two different APP models, in which
microglial cells were depleted after treatment with gancic-
lovir (Grathwohl et al., 2009). The authors found no
change in the number of amyloid-bplaques despite the
fact that the brains of APP or APP/PS1 mice were virtually
devoid of microglia. Although the amyloid-baccumulation
is still considered a main driver of the pathogenic events in
Alzheimer’s disease, several studies support the notion that
additional factors are necessary for the development of the
cognitive decline. Amyloid-band tau accumulation are fre-
quently observed in the brains of non-demented people
(Neuropathology Group; Medical Research Council
Cognitive and Aging, 2001). Also, patients immunized
against amyloid-b, leading to the effective removal of
their plaques, continued to decline in cognitive function
(Boche et al., 2008). Evidence suggests that inflammation
(Holmes et al., 2009), tangle pathology (Nelson et al.,
2012) or neurodegeneration-related biomarkers (Wirth
et al., 2013) correlate better with cognitive decline than
amyloid-baccumulation alone. All this evidence, together
with our present data, support an uncoupling of some
pathological hallmarks of Alzheimer’s disease from amyl-
oid-b, and highlight a more indirect route through the in-
flammatory component. A better understanding of what
processes are amyloid-band/or tau dependent and which
ones are caused, or exacerbated by inflammation, would be
a way forward to dissect the pathophysiology of
Alzheimer’s disease.
Interestingly, the report by Grathwohl et al. (2009) indir-
ectly supports our present findings regarding a relatively
high rate of microglial proliferation in APP/PS1 mice. The
fast depletion of microglia in CD11b-HSVTK (2 or 4-weeks
with ganciclovir) would only occur if all microglia were
entering the cell cycle during that time, as thymidine
kinase can only kill dividing cells. Although this prolifer-
ation rate is high it does suggest that other factors, such as
the recently reported anti-proliferative actions of ganciclo-
vir alone (Ding et al., 2014), are contributing to the
observed decline in microglia in CD11b-HSVTK mice
(Grathwohl et al., 2009). Our data suggest that, on average
in 9- and 14-month-old APP/PS1 mice, the proliferation
index (PI = Iba1
+
BrdU
+
/ total Iba1
+
) is 1.9%, suggesting
that an estimated 53% of the microglia undergo prolifer-
ation during a 4-week period. This high rate is similar with
that found in human samples, where the proliferation index
(PI = Iba1
+
Ki67
+
/ total Iba1
+
) in Alzheimer’s disease
brains is 2.63% in grey matter and 1.52% in white
matter. The fact that the mild increase in total microglial
numbers in both APP/PS1 and human Alzheimer’s disease
is not justified by the reported high rates of proliferation
indicates the necessity of compensatory microglial death.
This aspect needs further exploration in future work, to
Figure 6 CSF1R inhibition does not alter the levels of amyloid-b.(A) Multiplexed analysis of the concentration of soluble amyloid-b
38
,
amyloid-b
40
and amyloid-b
42
in cortex homogenates of APP/PS1 and wild-type mice at 9 months of age, after treatment for 3 months with a
control diet (RM1) or a diet containing GW2580. Soluble amyloid-blevels represented as mean SEM of concentration (ng/ml). (Band C)
Immunohistochemical analysis and quantification of the number of amyloid-bplaques (6E10
+
; brown) in the cortex of APP/PS1 mice at 9 months
of age, after treatment for 3 months with a control diet (RM1) or a diet containing GW2580. Number of amyloid-bplaques represented as
mean SEM of 6E10
+
plaques/mm
2
. Statistical differences: ***P50.001. Data were analysed with a two-way ANOVA and a post hoc Tukey test.
Scale bar in C= 100 mm.
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |903
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
better understand the observed discrepancies in different
brain regions (i.e. human cortical grey versus white
matter). It is worth noting that the expansion of the micro-
glial population in the model used in the current study,
APP/PS1, courses without the contribution of circulating
monocytes, as evidenced by the use of CCR2 deficient
mice and bone-marrow transplant with head shielding
(Mildner et al., 2011).
While the mechanistic link of CSF1R with the observed
beneficial outcomes seems clear, there are variables to take
into account when evaluating potential candidate drugs tar-
geting CSF1R. The activation of CSF1R can be achieved by
two independent mitogens, CSF1 and IL34 (Hume and
MacDonald, 2012). Although both CSF1 and IL34 are ex-
pressed in many organs, IL34 appears particularly upregu-
lated in the developing and adult brain, suggesting specific
functions that do not overlap with those of CSF1 (Wei
et al., 2010). Recently, data arising from the generation
of IL34
LacZ/LacZ
mice showed that IL34 is a tissue-restricted
ligand, controlling the development of Langerhans cells in
the skin and microglia in the brain (Wang et al., 2012). In
contrast, targeted deletion of IL34 had little effect on other
myeloid-cell compartments and dendritic cell subsets were
largely unaffected (Greter et al., 2012; Wang et al., 2012).
In conclusion, these studies suggest that the maintenance of
the populations of microglia and Langerhans cells is de-
pendent on IL34–CSF1R signalling, supporting the particu-
lar ability of these cell populations to self-renew
throughout life (Merad et al., 2008; Ginhoux et al.,
2013). Therefore, and although our present data do not
show any significant change in IL34 levels, it would be
interesting to observe the impact of anti-IL34 therapeutic
approaches to provide a more targeted inhibition of the
functions controlled by CSF1R in the brain, without affect-
ing many populations in peripheral organs. This is a par-
ticularly relevant issue to be addressed, as a prolonged
systemic delivery of CSF1R inhibitors may cause an unba-
lanced response in CSF1R-dependent cells (Sauter et al.,
2014). The therapeutic potential of CSF1R inhibitors has
been suggested in inflammatory diseases, autoimmune dis-
orders, bone disease and cancer (Burns and Wilks, 2011;
Pyonteck et al., 2013). Monocyte-derived cell types are
most likely required for the long-term integrity of the
immune system. Therefore, the therapeutic benefit of influ-
encing their function in particular pathologies may result in
the alteration of the natural balance of the immune system,
Figure 7 CSF1R inhibition prevents synaptic degeneration in APP/PS1 mice. (Aand B) Immunohistochemical analysis and quanti-
fication of protein levels of synaptophysin in the hippocampus of APP/PS1 and wild-type (WT) mice at 9 months of age, after treatment for 3
months with a control diet (RM1) or a diet containing GW2580. Synaptophysin levels represented as mean SEM of %Synaptophysin
+
area (A).
Representative confocal images are shown in B.(Cand D) Analysis of spine density in the apical segment of hippocampal CA1 neurons of APP/
PS1 and wild-type mice at 9 months of age, after treatment for 3 months with a control diet (RM1) or a diet containing GW2580. Representative
images are shown in D. Statistical differences: *P50.05, **P50.01. Data were analysed with a two-way ANOVA and a post hoc Tukey test. Scale
bars: B=50mm, D=10mm.
904 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
the consequences of which are hard to predict. In this dir-
ection, the use of imatinib (a PDGFR inhibitor with potent
activity over CSF1R) in clinical trials has shown that these
fears are unfounded, not presenting immune-related side
effects (Burns and Wilks, 2011). Similarly, an orally avail-
able inhibitor, JNJ-28312141, has substantial activity
against CSF1R and the related receptor FLT3 (Manthey
et al., 2009). JNJ-28312141 was found to decrease
Kupffer cell numbers by 40%, and also to reduce macro-
phage numbers in transplanted tumours, constraining
tumour growth (Manthey et al., 2009). Therefore, a more
targeted analysis of the peripheral effects of CSF1R inhib-
ition in ongoing or planned clinical trials is necessary, to
better ponder the long-term side effects of potential thera-
pies for Alzheimer’s disease.
One issue that requires comment is the selectivity of
CSF1R inhibitor GW2580. CSF1R belongs to the family
of type III growth factor receptors, including PDGFR, c-
KIT, FLT3 and c-fms (CSF1R). The structural similarity of
these receptors favours CSF1R inhibitors having certain
degree of promiscuity, potentially influencing other targets
(Hume and MacDonald, 2012). An analysis of the activity
and selectivity of several CSF1R inhibitors confirmed the
high selectivity of GW2580 for CSF1R, as all the other
compounds studied were also inhibitors of PDGFRband
c-Kit (Uitdehaag et al., 2011). A novel inhibitor, PLX3397,
has been described to have potent activity over CSF1R,
inhibiting the survival of microglia in the healthy brain
and causing the rapid depletion of the population
(Elmore et al., 2014). However, PLX3397 also has potent
inhibitory activity over c-Kit, FTL3 and PDGFRb
(Patwardhan et al., 2014), confounding the reported effects
on the microglial population. Loss of PDGFbsignalling, for
example, would be expected to have an impact on survival
of the NG2 pericytes leading to damage of the blood–brain
barrier and neurodegeneration (Bell et al., 2010). The par-
ticular effects of these approaches on the microglial popu-
lation highlight an important aspect: a therapeutic
approach to control CSF1R activity should be effective in
inhibiting microglial proliferation, but not affecting the sur-
vival of the remaining microglial cells, compromising other-
wise the normal brain function (Gomez-Nicola and Perry,
2014a,b). Our evidence shows that GW2580 selectively
inhibits microglial proliferation, and not survival, as long-
term dosing does not cause a significant reduction in the
number of microglial cells in wild-type mice. CSF1R acti-
vation has pleiotropic effects, ranging from the control of
cell survival, proliferation or chemotaxis, based on the dif-
ferential binding of adapter proteins and phosphorylation
pattern of its intracellular domain (Pixley and Stanley,
2004). Evidence from conformational analysis supports
the idea that different CSF1R inhibitors can bind the
active or inactive forms of the receptor and also show dif-
ferent dissociation rates, suggesting that these biophysical
properties could underpin a differential functional activa-
tion of CSF1R (Uitdehaag et al., 2011). These ideas need to
be taken into account to fully understand the effects of a
sustained inhibition of CSF1R in the normal physiology of
the brain.
In summary, the present data provide strong evidence for
an increased proliferative response in microglia in
Alzheimer’s disease, as well as their dependence upon
CSF1R activation. Prolonged reduction of microglial acti-
vation and proliferation in Alzheimer’s disease mice using a
selective CSF1R inhibitor prevents cognitive decline, re-
gardless of amyloid plaque pathology. We provide support
for the efficacy of CSF1R inhibitory strategies in the treat-
ment of Alzheimer’s disease-like pathology to reduce micro-
glia numbers and reduce the potentially damaging
components of neuroinflammation, thus underpinning the
possible evaluation of CSF1R inhibitors in clinical trials for
Alzheimer’s disease.
Acknowledgements
We thank the National CJD Surveillance Unit Brain Bank
(Edinburgh, UK) and the South West Dementia Brain Bank
(SWDBB) for the provision of human brain samples. The
SWDBB is part of the Brains for Dementia Research pro-
gramme, jointly funded by Alzheimer’s Research UK and
Alzheimer’s Society and is supported by BRACE (Bristol
Research into Alzheimer’s and Care of the Elderly) and
the Medical Research Council. We thank Kerry Hunter
(University of Lancaster) for technical assistance.
Funding
The research was funded by the Medical Research Council
(MR/K022687/1) and by Alzheimer’s Research UK.
Supplementary material
Supplementary material is available at Brain online.
References
Akiyama H, Barger S, Barnum S, Bradt B, Bauer J, Cole GM, et al.
Inflammation and Alzheimer’s disease. Neurobiol Aging 2000;21:
383–421.
Akiyama H, Nishimura T, Kondo H, Ikeda K, Hayashi Y, McGeer PL.
Expression of the receptor for macrophage colony stimulating factor
by brain microglia and its upregulation in brains of patients with
Alzheimer’s disease and amyotrophic lateral sclerosis. Brain Res
1994;639:171–4.
Bell RD, Winkler EA, Sagare AP, Singh I, LaRue B, Deane R, et al.
Pericytes control key neurovascular functions and neuronal pheno-
type in the adult brain and during brain aging. Neuron 2010;68:
409–27.
Blalock EM, Geddes JW, Chen KC, Porter NM, Markesbery WR,
Landfield PW. Incipient Alzheimer’s disease: microarray correlation
analyses reveal major transcriptional and tumor suppressor re-
sponses. Proc Natl Acad Sci USA 2004;101:2173–8.
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |905
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Boche D, Zotova E, Weller RO, Love S, Neal JW, Pickering RM, et al.
Consequence of Abeta immunization on the vasculature of human
Alzheimer’s disease brain. Brain 2008;131(Pt 12):3299–310.
Bondolfi L, Calhoun M, Ermini F, Kuhn HG, Wiederhold KH, Walker
L, et al. Amyloid-associated neuron loss and gliogenesis in the neo-
cortex of amyloid precursor protein transgenic mice. J Neurosci
2002;22:515–22.
Brawek B, Schwendele B, Riester K, Kohsaka S, Lerdkrai C, Liang Y,
et al. Impairment of in vivo calcium signaling in amyloid plaque-
associated microglia. Acta Neuropathol 2014;127:495–505.
Burns CJ, Wilks AF. c-FMS inhibitors: a patent review. Expert Opin
Ther Patents 2011;21:147–65.
Condello C, Yuan P, Schain A, Grutzendler J. Microglia constitute a
barrier that prevents neurotoxic protofibrillar Abeta42 hotspots
around plaques. Nat Commun 2015;6:6176.
Cunningham C, Deacon R, Wells H, Boche D, Waters S, Diniz CP, et al.
Synaptic changes characterize early behavioural signs in the ME7
model of murine prion disease. Eur J Neurosci 2003;17:2147–55.
Deacon RM, Rawlins JN. T-maze alternation in the rodent. Nat
Protoc 2006;1:7–12.
Dickson DW, Lee SC, Mattiace LA, Yen SH, Brosnan C. Microglia
and cytokines in neurological disease, with special reference to AIDS
and Alzheimer’s disease. Glia 1993;7:75–83.
Ding Z, Mathur V, Ho PP, James ML, Lucin KM, Hoehne A, et al.
Antiviral drug ganciclovir is a potent inhibitor of microglial
proliferation and neuroinflammation. J Exp Med 2014;211:
189–98.
Edison P, Archer HA, Gerhard A, Hinz R, Pavese N, Turkheimer FE,
et al. Microglia, amyloid, and cognition in Alzheimer’s disease: an
[11C](R)PK11195-PET and [11C]PIB-PET study. Neurobiol Dis
2008;32:412–9.
Elmore MR, Najafi AR, Koike MA, Dagher NN, Spangenberg EE,
Rice RA, et al. Colony-stimulating factor 1 receptor signaling is
necessary for microglia viability, unmasking a microglia progenitor
cell in the adult brain. Neuron 2014;82:380–97.
Erblich B, Zhu L, Etgen AM, Dobrenis K, Pollard JW. Absence of
colony stimulation factor-1 receptor results in loss of microglia, dis-
rupted brain development and olfactory deficits. PloS One 2011;6:
e26317.
European Alzheimer’s Disease I, Genetic, Environmental Risk in
Alzheimer’s D, Alzheimer’s Disease Genetic C, Cohorts for H,
Aging Research in Genomic E, et al. Meta-analysis of 74,046 indi-
viduals identifies 11 new susceptibility loci for Alzheimer’s disease.
Nat Genet 2013;45:1452–8.
Fernandez-Botran R, Ahmed Z, Crespo FA, Gatenbee C, Gonzalez J,
Dickson DW, et al. Cytokine expression and microglial activation
in progressive supranuclear palsy. Parkinsonism Relat Disord
2011;17:683–8.
Frautschy SA, Yang F, Irrizarry M, Hyman B, Saido TC, Hsiao K,
et al. Microglial response to amyloid plaques in APPsw transgenic
mice. Am J Pathol 1998;152:307–17.
Ginhoux F, Lim S, Hoeffel G, Low D, Huber T. Origin and differen-
tiation of microglia. Front Cell Neurosci 2013;7:45.
Gjoneska E, Pfenning AR, Mathys H, Quon G, Kundaje A, Tsai LH,
et al. Conserved epigenomic signals in mice and humans reveal
immune basis of Alzheimer’s disease. Nature 2015;518:365–9.
Gomez-Nicola D, Boche D. Post-mortem analysis of neuroinflamma-
tory changes in human Alzheimer’s disease. Alzheimers Res Ther
2015;7:42.
Gomez-Nicola D, Fransen NL, Suzzi S, Perry VH. Regulation of micro-
glial proliferation during chronic neurodegeneration. J Neurosci
2013;33:2481–93.
Gomez-Nicola D, Perry VH. Microglial dynamics and role in the
healthy and diseased brain: a paradigm of functional plasticity.
Neuroscientist 2014a;21;169–84.
Gomez-Nicola D, Perry VH. Neurodegenerative diseases. In: Tremblay
M-E
`, Sierra A, editors. Microglia in health and disease: Springer:
New York; 2014b. p. 437–53.
Gomez-Nicola D, Perry VH. Analysis of microglial proliferation in
Alzheimer’s disease. Methods Mol Biol 2016;1303:185–93.
Gomez-Nicola D, Schetters ST, Hugh Perry V. Differential role of
CCR2 in the dynamics of microglia and perivascular macrophages
during prion disease. Glia 2014a;67:1041–52.
Gomez-Nicola D, Suzzi S, Vargas-Caballero M, Fransen NL, Al-
Malki H, Cebrian-Silla A, et al. Temporal dynamics of hippocam-
pal neurogenesis in chronic neurodegeneration. Brain 2014b;137:
2312–28.
Gomez-Nicola D, Valle-Argos B, Pallas-Bazarra N, Nieto-Sampedro
M. Interleukin-15 regulates proliferation and self-renewal of adult
neural stem cells. Mol Biol Cell 2011;22:1960–70.
Gomez-Nicola D, Valle-Argos B, Pita-Thomas DW, Nieto-Sampedro M.
Interleukin 15 expression in the CNS: Blockade of its activity prevents
glial activation after an inflammatory injury. Glia 2008;56:494–505.
Grathwohl SA, Kalin RE, Bolmont T, Prokop S, Winkelmann G,
Kaeser SA, et al. Formation and maintenance of Alzheimer’s disease
beta-amyloid plaques in the absence of microglia. Nat Neurosci
2009;12:1361–3.
Greter M, Lelios I, Pelczar P, Hoeffel G, Price J, Leboeuf M, et al.
Stroma-derived interleukin-34 controls the development and main-
tenance of langerhans cells and the maintenance of microglia.
Immunity 2012;37:1050–60.
Guerreiro R, Wojtas A, Bras J, Carrasquillo M, Rogaeva E, Majounie
E, et al. TREM2 variants in Alzheimer’s disease. N Engl J Med
2013;368:117–27.
Holmes C, Cunningham C, Zotova E, Culliford D, Perry VH.
Proinflammatory cytokines, sickness behavior, and Alzheimer dis-
ease. Neurology 2011;77:212–8.
Holmes C, Cunningham C, Zotova E, Woolford J, Dean C, Kerr S,
et al. Systemic inflammation and disease progression in Alzheimer
disease. Neurology 2009;73:768–74.
Hoozemans JJ, Veerhuis R, Rozemuller JM, Eikelenboom P. Soothing
the inflamed brain: effect of non-steroidal anti-inflammatory drugs
on Alzheimer’s disease pathology. CNS Neurol Disord Drug Targets
2011;10:57–67.
Hume DA, MacDonald KP. Therapeutic applications of macrophage
colony-stimulating factor-1 (CSF-1) and antagonists of CSF-1 recep-
tor (CSF-1R) signaling. Blood 2012;119:1810–20.
Jones RS, Lynch MA. How dependent is synaptic plasticity on micro-
glial phenotype? Neuropharmacology 2014;96:3–10.
Jonsson T, Stefansson H, Steinberg S, Jonsdottir I, Jonsson PV,
Snaedal J, et al. Variant of TREM2 associated with the risk of
Alzheimer’s disease. N Engl J Med 2013;368:107–16.
Jucker M. The benefits and limitations of animal models for transla-
tional research in neurodegenerative diseases. Nat Med 2010;16:
1210–4.
Jun G, Naj AC, Beecham GW, Wang LS, Buros J, Gallins PJ, et al.
Meta-analysis confirms CR1, CLU, and PICALM as alzheimer dis-
ease risk loci and reveals interactions with APOE genotypes. Arch
Neurol 2010;67:1473–84.
Kamphuis W, Orre M, Kooijman L, Dahmen M, Hol EM. Differential
cell proliferation in the cortex of the APPswePS1dE9 Alzheimer’s
disease mouse model. Glia 2012;60:615–29.
LaFerla FM, Oddo S. Alzheimer’s disease: Abeta, tau and synaptic
dysfunction. Trends Mol Med 2005;11:170–6.
Li T, Pang S, Yu Y, Wu X, Guo J, Zhang S. Proliferation of paren-
chymal microglia is the main source of microgliosis after ischaemic
stroke. Brain 2013;136(Pt 12):3578–88.
Manthey CL, Johnson DL, Illig CR, Tuman RW, Zhou Z, Baker JF,
et al. JNJ-28312141, a novel orally active colony-stimulating factor-
1 receptor/FMS-related receptor tyrosine kinase-3 receptor tyrosine
kinase inhibitor with potential utility in solid tumors, bone metas-
tases, and acute myeloid leukemia. Mol Cancer Ther 2009;8:
3151–61.
McClean PL, Parthsarathy V, Faivre E, Holscher C. The diabetes drug
liraglutide prevents degenerative processes in a mouse model of
Alzheimer’s disease. J Neurosci 2011;31:6587–94.
906 |BRAIN 2016: 139; 891–907 A. Olmos-Alonso et al.
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
Merad M, Ginhoux F, Collin M. Origin, homeostasis and function of
Langerhans cells and other langerin-expressing dendritic cells. Nat
Rev Immunol 2008;8:935–47.
Mildner A, Schlevogt B, Kierdorf K, Bottcher C, Erny D, Kummer MP,
et al. Distinct and non-redundant roles of microglia and myeloid
subsets in mouse models of Alzheimer’s disease. J Neurosci
2011;31:11159–71.
Murphy GM Jr, Zhao F, Yang L, Cordell B. Expression of macrophage
colony-stimulating factor receptor is increased in the AbetaPP(V717F)
transgenic mouse model of Alzheimer’s disease. Am J Pathol
2000;157:895–904.
Nelson PT, Alafuzoff I, Bigio EH, Bouras C, Braak H, Cairns NJ,
et al. Correlation of Alzheimer disease neuropathologic changes
with cognitive status: a review of the literature. J Neuropathol
Exp Neurol 2012;71:362–81.
Neuropathology Group. Medical Research Council Cognitive F, Aging
S. Pathological correlates of late-onset dementia in a multicentre, com-
munity-based population in England and Wales. Neuropathology
Group of the Medical Research Council Cognitive Function and
Ageing Study (MRC CFAS). Lancet 2001;357:169–75.
Otero K, Turnbull IR, Poliani PL, Vermi W, Cerutti E, Aoshi T, et al.
Macrophage colony-stimulating factor induces the proliferation and
survival of macrophages via a pathway involving DAP12 and beta-
catenin. Nat Immunol 2009;10:734–43.
Patwardhan PP, Surriga O, Beckman MJ, De Stanchina E, Dematteo
R, Tap WD, et al. Sustained inhibition of receptor tyrosine kinases
and macrophage depletion by PLX3397 and rapamycin as a poten-
tial new approach for the treatment of MPNSTs. Clin Cancer Res
2014;20:3146–58.
Perry VH, Nicoll JA, Holmes C. Microglia in neurodegenerative dis-
ease. Nat Rev Neurol 2010;6:193–201.
Pixley FJ, Stanley ER. CSF-1 regulation of the wandering macrophage:
complexity in action. Trends Cell Biol 2004;14:628–38.
Pyonteck SM, Akkari L, Schuhmacher AJ, Bowman RL, Sevenich L,
Quail DF, et al. CSF-1R inhibition alters macrophage polarization
and blocks glioma progression. Nat Med 2013;19:1264–72.
Rothman SM, Herdener N, Camandola S, Texel SJ, Mughal MR,
Cong WN, et al. 3xTgAD mice exhibit altered behavior and elevated
Abeta after chronic mild social stress. Neurobiol Aging 2012;33:830
e1–12.
Sasmono RT, Oceandy D, Pollard JW, Tong W, Pavli P,
Wainwright BJ, et al. A macrophage colony-stimulating factor
receptor-green fluorescent protein transgene is expressed through-
out the mononuclear phagocyte system of the mouse. Blood
2003;101:1155–63.
Sauter KA, Pridans C, Sehgal A, Tsai YT, Bradford BM, Raza S, et al.
Pleiotropic effects of extended blockade of CSF1R signaling in adult
mice. J Leukoc Biol 2014;96:265–74.
Uitdehaag JC, Sunnen CM, van Doornmalen AM, de Rouw N, Oubrie
A, Azevedo R, et al. Multidimensional profiling of CSF1R screening
hits and inhibitors: assessing cellular activity, target residence time,
and selectivity in a higher throughput way. J Biomol Screen
2011;16:1007–17.
Ulrich JD, Finn MB, Wang Y, Shen A, Mahan TE, Jiang H, et al.
Altered microglial response to Abeta plaques in APPPS1-21 mice
heterozygous for TREM2. Mol Neurodegener 2014;9:20.
Wang Y, Szretter KJ, Vermi W, Gilfillan S, Rossini C, Cella M, et al.
IL-34 is a tissue-restricted ligand of CSF1R required for the devel-
opment of Langerhans cells and microglia. Nat Immunol
2012;13:753–60.
Wei S, Nandi S, Chitu V, Yeung YG, Yu W, Huang M, et al.
Functional overlap but differential expression of CSF-1 and IL-34
in their CSF-1 receptor-mediated regulation of myeloid cells. J
Leukoc Biol 2010;88:495–505.
Wirth M, Madison CM, Rabinovici GD, Oh H, Landau SM, Jagust
WJ. Alzheimer’s disease neurodegenerative biomarkers are asso-
ciated with decreased cognitive function but not beta-amyloid in
cognitively normal older individuals. J Neurosci 2013;33:5553–
63.
Zhang B, Gaiteri C, Bodea LG, Wang Z, McElwee J, Podtelezhnikov
AA, et al. Integrated systems approach identifies genetic nodes and
networks in late-onset Alzheimer’s disease. Cell 2013;153:707–20.
CSF1R inhibition ameliorates Alzheimer’s disease-like pathology BRAIN 2016: 139; 891–907 |907
Downloaded from https://academic.oup.com/brain/article/139/3/891/2468754 by guest on 21 March 2023
