Artigo cientifico

1. Graphene quantum dots conjugated neuroprotective peptide improve
learning and memory capability
Songhua Xiao a, 1
, Daoyou Zhou b, 1
, Ping Luan c
, Beibei Gu d
, Longbao Feng e
,
Shengnuo Fan a
, Wang Liao a
, Wenli Fang a
, Lianhong Yang a
, Enxiang Tao a
, Rui Guo e, *
,
Jun Liu a, f, g, **
a
Department of Neurology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, 107 Yanjiang West Road, Guangzhou, 510120, China
b
Department of Neurology and Outpatient Department of Internal Medicine, Guangdong Provincial Hospital of Traditional Chinese Medicine, Guangzhou,
510120, China
c
School of Medicine, Shenzhen University, 3688 Nanhai Avenue, Shenzhen, 518060, China
d
Department of Anesthesiology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, 510120, China
e
Key Laboratory of Biomaterials of Guangdong Higher Education Institutes, Department of Biomedical Engineering, Jinan University, Guangzhou 510632,
China
f
Laboratory of RNA and Major Diseases of Brain and Heart, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, 510120, China
g
Guangdong Province Key Laboratory of Brain Function and Disease, Zhongshan School of Medicine, Sun Yat-sen University, Guangzhou 510080, China
a r t i c l e i n f o
Article history:
Received 13 June 2016
Received in revised form
8 August 2016
Accepted 14 August 2016
Available online 17 August 2016
Keywords:
Graphene quantum dots
Neuroprotective peptide
Amyloid-b (Ab)
Dendritic spine
Alzheimer disease (AD)
a b s t r a c t
Alzheimer disease (AD) is a neurodegenerative disorder and the most common form of dementia. His-
topathologically is characterized by the presence extracellular neuritic plaques and with a large number
of neurons lost. In this paper, we design a new nanomaterial, graphene quantum dots (GQDs) conjugated
neuroprotective peptide glycine-proline-glutamate (GQDG) and administer it to APP/PS1 transgenic
mice. The in vitro assays including ThT and CD proved that GQDs and GQDG could inhibit the aggregation
of Ab1-42 fibrils. Morris water maze was performed to exanimate learning and memory capacity of APP/
PS1 transgenic mice. The surface area of Ab plaque deposits reduced in the GQDG group compared to the
Tg Ctrl groups. Furthermore, newly generated neuronal precursor cell and neuron were test by immu-
nohistochemical. Besides, neurons were impregnated by DiI using gene gun to show dendritic spine.
Results indicated enhancement of learning and memory capacity and increased amounts of dendritic
spine were observed. Inflammation factors and amyloid-b (Ab) were tested with suspension array and
ELISA, respectively. Several pro-inflammatory cytokines (IL-1a, IL-1b, IL-6, IL-33, IL-17a, MIP-1b and TNF-
a) had decreased in GQDG group compared with Control group. Reversely, anti-inflammatory cytokines
(IL-4, IL-10) had increased in GQDG group compared with Control group. Thus, we demonstrate that the
GQDG is a promising drug in treatment of neurodegenerative diseases such as AD.
© 2016 Elsevier Ltd. All rights reserved.
1. Introduction
Alzheimer's disease (AD) is characterized by the progressive loss
of neurons [1]. One of the most widely accepted theory of AD is the
amyloid hypothesis, featuring aggregation and fibril formation of
amyloid-b (Ab) peptides [2]. Studies have shown that Ab is the
possible biomarkers for the AD diagnose [3]. Ab (38e43 residues)
comes from proteolysis of amyloid precursor protein (APP), and
Ab1-42 is the most toxic form presents in the brains of AD patients.
The aggregation of Ab in the brain is believed to be linked to the
neuron apoptosis and loss of cognitive function observed in pa-
tients with AD [4]. Some studies indicate that Ab can insert into the
membrane and form ion channels and cause calcium overload; or it
can alter the activity of NMDA receptors and modulating calcium
influx [5].
Neuroinflammation is now well recognized as a prominent
feature in Alzheimer's pathology and a potential target for therapy
and prevention of this disease [6]. Inflammatory components
* Corresponding author. Key Laboratory of Biomaterials of Guangdong Higher
Education Institutes, Department of Biomedical Engineering, Jinan University,
Guangzhou 510632, China.
** Corresponding author. Department of Neurology, Sun Yat-sen Memorial Hos-
pital, Sun Yat-sen University, 107 Yanjiang West Road, Guangzhou, 510120, China.
E-mail addresses: guorui@jnu.edu.cn (R. Guo), docliujun@hotmail.com (J. Liu).
1
These two authors contributed equally to this work.
Contents lists available at ScienceDirect
Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials
http://dx.doi.org/10.1016/j.biomaterials.2016.08.021
0142-9612/© 2016 Elsevier Ltd. All rights reserved.
Biomaterials 106 (2016) 98e110

2. include microglia, cytokines and chemokines [7,8]. Evidence shows
that accumulations of Ab can cause inflammatory response,
following the neurodegenerative condition [9]. Therefore, Ab is a
potential target for therapeutic intervention for AD.
Bioactive peptide which was defined as specific protein frag-
ments have a positive impact on body functions [10]. A endogenous
neuroprotective peptide glycine-proline-glutamate (Gly-Pro-Glu,
GPE), which is N-terminal tripeptide of insulin-like growth factor I
(IGF-1) [11], has been proved to be an effective peptide. The pro-
tective effects of GPE may be related with modulation of intracel-
lular calcium signaling, inhibition of glutamate binding to NMDA
receptors and protection against NMDA excitotoxicity [12]. GPE
could also rescue cell death induced by Ab and inhibit apoptotic
[13]. These suggests that GPE could be an important factor impli-
cated in the brain's protection [14].
Admittedly, the blood brain barrier (BBB) hampers the access of
systemically administered drugs to the brain [15]. So the BBB
permeability is one of the key factors for applications of bio-
materials. In the area of biomedicine, graphene nano-particles and
its derivatives have been highly anticipated to provide unique and
new opportunities for the developments of novel nanocarriers for
drug delivery [16]. Graphene quantum dots (GQDs), a new type of
carbon-based nanomaterial, is a promising new material for drug
carrier material owing to its small size, excellent solubility, large
specific surface area and low cost [17,18]. GQDs exhibits no
apparent toxicity in vitro and in vivo [19]. GQDs have been proved
that could across the Madin-Darby canine kidney (MDCK) cell
monolayer mainly through a lipid raft-mediated transcytosis [20].
Thus, GQDs may cross the BBB because of its small size. Moreover,
GQDs are demonstrated to inhibit inhibitor for aggregation of Ab1-
42 and rescue the cytotoxicity of Ab oligomers [21].
In our study, we conjugated the GQDs with neuroprotective
peptide GPE, we named the new nanomaterial as GQDG. We found
GQDG can inhibit the Ab1-42 aggregation in vitro. Then transmission
electron microscope, circular dichroism spectrum and Thioflavin-T
assay was used to confirm the GQDG could prevent the aggregation
of Ab1-42. Then GQDG was administrated to APP/PS1 transgenic
mice by intravenous injection to observe its therapeutic effect
through Morris water maze, immunohistochemical, ELISA, sus-
pension array and diolistic labeling in vivo.
2. Materials and methods
2.1. Preparation of materials
2.1.1. Preparation of graphene oxide
Natural graphite powder (<150 mm, Aldrich) was used for gra-
phene oxide (GO) synthesis by the modified Hummers method
[22]. An additional preoxidation procedure was performed before
the GO preparation. Solution contains H2SO4 (30 mL), K2S2O8 (10 g),
and P2O5 (10 g) was used to dissolve the graphite powder (8 g) and
the temperature was control at 80 C. The dark mixture was iso-
lated, cooled at room temperature overnight and then diluted with
ultrapure water to be neutral. After that the graphite was preoxi-
dized and then put into cold H2SO4 (180 mL) added with NaNO3
(4 g). We kept the temperature below 10 C and added KMnO4
(24 g) in to the mixture slowly. The mixture was then stirred at
around 40 C for 8 h. Then we added 400 mL water and heated the
mixture over 90 C for 15 min, large amount of water and 30% H2O2
solution was added to stop the reaction. The mixture was filtered
and washed with 1:8 HCl solution (800 mL). Lastly, after ultrapure
water washed for three times, GO product was collected.
2.1.2. Preparation of GQDs
GQDs were synthesized using the method by published before
[23]. Briefly, GO was dissolved in N, N-Dimethylformamide (DMF)
with the concentrations of 30 mg/mL. Ultrasonication was used to
treat the complex for 30 min (120 W, 100 kHz). The complex was
added to a poly(tetrafluoroethylene) (Teflon)-lined autoclave
(40 mL) and heated at 200 C for 4 h. Then the container was cooled
to 25 C temperature by water. The product contained black sedi-
ments and brown transparent supernatant, and the black sedi-
ments (GQDs) were collected and washed by water. PBS was used to
conserve the GQDs.
2.1.3. GQDs and GPE coupling
The conjunction between GQDs and GPE was realized by the 1-
(3-(dimethylamino)propyl)-3-ethylcarbodiimide and N-hydrox-
ysuccinimide (EDC/NHS) coupling reaction [24]. Briefly, 8 mg of
EDC and 12 mg of NHS were added into GQDs solution (10 mL). We
added 10 mg of GPE (ChinaPeptide Co., Ltd, Shanghai, China) into
the solution and the reaction was conducted under stirring for 3 h
at room temperature. To remove residual EDC and NHS, the solu-
tion was dialyzed (MWCO, 500e1000) for 72 h (dialysate was
replaced every 12 h).
2.2. Animals
APPswe/PS1dE9 double transgenic mice (APP/PS1 mice) were
purchased from the Model Animal Research Center of Nanjing
University (Nanjing, China) (strain type B6C3-Tg [APPswe,
PSEN1dE9] 85Dbo/J, stock number 004462). In this study, 36 SPF 6-
month-old male APP/PS1 double transgenic (Tg) mice were used
(weight ¼ 27.70 g ± 3.47 g), 12 age-matched wild-type (Wt) lit-
termates were used as controls. All experimental procedures
involving animals were performed according to the regulations of
the Institutional Animal Care and Use Committee (IACUC) of Sun
Yat-sen University, Guangzhou, China. We kept all animals under
specific pathogenefree (SPF) conditions on a 12-h light, 12-h dark
cycle, and food and water were provided to the mice ad libitum.
The GQDG was resolved in the 0.01 M PBS with the final con-
centration of 200 mg/mL. We randomLy divided the mice into 4
groups: transgenic control (Tg Ctrl, n ¼ 12), transgenic GQDG
(GQDG, n ¼ 12), transgenic PBS (PBS, n ¼ 12), wild type control (Wt
Ctrl, n ¼ 12). Mice were administered GQDG or 0.01 M PBS every
day for 4 weeks. All of the mice were administered 50 mg/kg body
weight 5-bromo-2-deoxyuridine (BrdU, Sigma-Aldrich, USA) each
day for the first 5 days of administration.
2.3. In vitro assays
2.3.1. Thioflavin-T assay
The Thioflavin-T (ThT) fluorescence method was used [25], and
Ab1-42 (Sigma-Aldrich, USA) was dissolved in phosphate buffer (PB,
pH 7.4, 0.01 M) to give a 50 mM solution. GQDG was firstly dissolved
in PB at a concentration of 200 mg/mL. After incubating at 37 C for
48 h, ThT (5 mM in 50 mM glycine-NaOH buffer, pH 8.50) was added.
Fluorescence was measured at 450 nm and 485 nm. Each sample was
examined in triplicate. The fluorescence intensities were recorded,
and the percentage of inhibition on aggregation was calculated.
2.3.2. Transmission electronic microscopy (TEM)
Ab1-42 samples were prepared in PB (pH 7.4) at a concentration
of 50 mM. Then the Ab1-42 samples were incubated with or without
200 mg/mL GQDG for 48 h. To detect the structure of these Ab1-42
samples, 5 ml of samples to be imaged were spotted on 300-mesh
Formvar-carbon coated copper grid and stained with 1% uranyl
formate for 1 min. Afterwards, samples were air dried and observed
under the transmission electron microscope (FEI Inc., USA) with a
voltage of 80 kV.
S. Xiao et al. / Biomaterials 106 (2016) 98e110 99

3. 2.3.3. Hemolysis test
The method was described previously [26]. Briefly, normal sa-
line (NS, 0.9% sodium chloride) was used to dissolve GQDs, GQDG or
GEP. Whole blood was diluted with NS (4:5) and then added to
GQDs, GQDG or GEP solutions. After incubated at 37 C for 60 min,
the mixture was then centrifuged at 1500 rpm for10 min. The su-
pernatant was collected and transferred to a 96-well plate absor-
bance was measured at 545 nm using a Multiskan microplate
reader (Thermo Scientific, USA). Each sample was measured for
three times. Diluted blood with deionized water was used as pos-
itive control and diluted blood with NS was used as negative con-
trols Hemolysis degree was calculated as: Hemolysis rate
% ¼ (ODtest sampleÀODnegative control)/(ODpositive controlÀODnegative
control).
2.4. Morris water maze (MWM)
The MWM tests were conducted to evaluate the spatial learning
and memory ability of the APP/PS1 mice after GQDG administration
as described [27]. The MWM test was consists of an acquisition trial
for five days, a one-day probe test and a two-day reversal test [28].
There were four contiguous trails per day in the acquisition trail.
The timer was set to 60 s and automatically stopped once the
mouse reached the platform within the 60 s and remained on the
platform for 5 s. The mice were manually moved to the platform
and allowed to stay on the platform for 20 s if it could not find the
platform. The probe test was conducted 24 h after the end of the
acquisition trial, and the reversal test was conducted 24 h later. On
reversal test (days 7e8) the platform was placed at the quadrant
opposite the location from day1 to day5, and the mice were then
retrained in four sessions per day. After each trial, mice were dried
off in their housing facilities next to an electric heater for 30 min.
The trajectory and escape latency of the mice were recorded, and
the average escape latency was analyzed each day. In our experi-
ment, we recorded the platform-site crossovers, the path length of
each group in each quadrant and the duration of each group in each
quadrant in the extinction phase for 60 s.
2.5. Tissue preparation
All of the mice were sacrificed under deep anesthesia after the
behavioral tests. The venous bloods of the mice were collected with
the 1 mL syringes, followed by transcardial perfusions with ice-cold
PBS for 5 min, which included Protease Inhibitor Cocktail (Sigma-
Aldrich, USA). Then, the blood samples were allowed to precipitate
for 2 h at room temperature before being centrifuged for 15 min at
1000 Â g. The serum was collected and stored at À80 C. For
biochemical analysis, the brain used was removed and immediately
frozen at À80 C. For immunohistochemical (IHC) and hematoxylin
and eosin (HE) staining, the mice were perfused transcardially
and post-fixed in 4% paraformaldehyde (PFA) overnight, and
dehydrated in 30% sucrose in 0.1 M phosphate buffer for 48 h. For
the diolistic labeling, mice were perfused with 2% PFA and post-
fixed for 12 h; then, the coronal sections (200 mm) were cut on a
vibratome (Leica VT 1000 S, Germany) and one intact section was
stored in 0.01 M PBS for the diolistic labeling. The dentate gyrus and
adjacent cortex À2.2 mm to À2.4 mm relative to the bregma were
taken as the region of interest (ROI) in both analyses.
2.6. Immunohistochemical staining
The mice brain sections were blocked for 2 h with blocking
solution (PBS containing 10% normal FBS and 0.5% Triton X-100).
The tissues were then stained overnight at 4 C with the following
combinations of primary antibodies: rabbit anti-Iba1 (ionized
calcium binding adapter molecule, 1:500; Wako Chemicals Inc.,
USA), goat anti-DCX (doublecortin, 1:400; Santa Cruz Biotech-
nology), mouse anti-NeuN (neuronal nuclear antigen, 1:800;
Sigma-Aldrich, USA)and mouse anti-Ab1-42 (1:1000, Sigma-Aldrich,
USA) and rat anti-BrdU (1:400; Oxford Biotechnology), labeled with
the secondary antibodies: Alexa Flour 555 goat anti-rabbit anti-
body, Alexa Flour 488 goat anti-mouse antibody, Alexa Fluor 488
goat anti-rabbit, Alexa Fluor 488 donkey anti-goat, Alexa Flour 594
donkey anti-rat, (1:400; all from Life Technology). DAPI (Roche)
was used to stain the nuclear. The surface area of the senile plaques
were measured and compared as percentage of the dentate gyrus
with Image-Pro Plus 6.0 (Media Cybernetics, Inc. Maryland, USA).
Meanwhile, a Zeiss LSM 710 laser confocal scanning microscope
(Zeiss, Germany) was used to photograph. The Ab immunoreaction
stained areas are expressed as a percentage of the total brain region
tissue area, as are the quantitative image analysis of microglial cell
immunoreactivity. For quantitative image analysis of neurogenesis
(BrdUþ, BrdUþ/DCXþ and BrdUþ/NeuNþ), cell proliferation was
assessed by unilateral counting in the dentate gyrus with an opti-
calfractionator stereology system stereo investigator (Micro-
BrightField, Williston, USA). The actual section thickness was
tested, and the appropriate guard zones at the top and the bottom
of the sections were defined to avoid oversampling. Measurements
were made in a systematic series of six 40-mm coronal sections,
240 mm apart.
2.7. Enzyme linked immunosorbent assay (ELISA)
We measured the concentration of Ab1-40, Ab1-42, brain derived
neurotrophic factor (BDNF) and nerve growth factor (NGF) in the
mice's brain with the ELISA method. The assays were performed
using the ELISA kits (Invitrogen for Ab1-40 and Ab1-42, Millipore for
BDNF and NGF) following the manufacturer's instructions. The
frozen brains were thawed and minced. Then, the cerebral tissues
of the brain were weighed; an aliquot of the tissue was homoge-
nized in a RIPA buffer containing 50 mM Tris (pH 7.4),150 mM NaCl,
1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, sodium orthova-
nadate, sodium fluoride, EDTA, leupeptin and a protease inhibitors
cocktail. The homogenates were centrifuged (100,000 Â g, 1 h,
4 C), and the supernatants were stored at À80 C for additional
analysis of soluble Ab, BDNF and NGF. Then, we sonicated the pellet
in 5 M guanidine Tris buffer; the samples were incubated for
30 min at room temperature, and then centrifuged (100,000 Â g,
1 h, 4 C), and the supernatants were stored for analysis of insoluble
Ab.
2.8. Bio-plex suspension array
Another aliquot of the brain tissue was prepared with the Bio-
Plex cell lysis kit developed specifically to prepare tissue lysate
samples for analysis with Bio-Plex total target assays. After
4500 Â g for 4 min, the supernatants were collected and used for
the suspension array. Multiplex Suspension Arrays analysis was
performed using the Bio-plex protein multi-array system, which
was equipped with Luminex-based technology. For screening and
exploring the significant unknown cytokines, the Bio-Plex Pro
Mouse Cytokine Group I, Group II (Bio-Rad, USA) was employed.
Samples, standards and blanks were prepared according to man-
ufacturer's instructions. Using biotinylated detection antibody and
streptavidin-phycoerythrin (SAPE), the signals were generated.
Raw median fluorescent intensity (MFI) data was captured by Bio-
Plex 200 System and analyzed using Bio-Plex Manager software 4.1
(Bio-Rad Laboratories, Hercules CA), under a high standard pho-
tomultiplier tube setting. The cytokines concentrations were
detected by computation from a standard curve that was based on
S. Xiao et al. / Biomaterials 106 (2016) 98e110100

4. the data from serial dilutions of an established reference sample
and its sensitivity level can reach at pg/mL. All the procedures
followed the recommendations of the manufacturer.
2.9. DiI-diolistic labelling and morphological analyses
The gene gun bullets were prepared as described [29,30]. We
mixed 4 mg of gold particles (1.6 mm diameter, Bio-Rad, USA) with
2.5 mg DiI (Sigma-Aldrich, USA) and dissolved in 250 ml methylene
chloride. After drying, the coated particles were collected in 1.5 mL
water and placed into a sonicator for 5 min. The solution was vor-
texed for 15 s and immediately transferred to a Tefzel tubing (Bio-
Rad, USA). The labeled sections were rinsed in 0.01 M PBS three
times and resuspended in PBS at 4 C overnight for the dye to
diffuse through the neuronal membranes. The images of DiI-
impregnated cells were taken using a Zeiss LSM 710 confocal mi-
croscope (Zeiss, Germany). The neuron was scanned at 1 mm in-
crements along the Z-axis and reconstructed to analyze the
dendrite segments. The density of the dendritic spines was
measured on 10 to 30 randomLy chosen dendrites from 3 to 6
neurons, calculated by quantifying the number of spines per unit
length of dendrite and normalized per 10 mm of dendrite length.
Three dimensional reconstructions of confocal images were per-
formed with an Imaris 6.4.2 (Bitplane, Zurich, Switzerland).
2.10. Statistical analysis
The data represented as the mean ± S.E.M. and p 0.05 was
considered to be significant. We use a two-way analysis of variance
(ANOVA) for repeated measures to analyze the MWM data. If the
data were significant, a Bonferroni post-hoc test was used to
perform the subsequent analysis. In the analyses of ELISA, and
diolistic labeling, we used a simple t-test to determine the differ-
ence between Wt Ctrl and Tg Ctrl, while the ANOVA was performed
to evaluate the effect of GQDG within the Tg Ctrl group. SPSS 13.0
software was used for all the statistical analyses.
3. Results and discussion
3.1. Synthesis and characterization of products
From the results, the GQDs possess excellent solubility in water
and physiological media such as PBS. Photographs of GQDs solution
taken under visible and UV lights, brown (visible lights) and blue
fluorescent color (UV lights) is observed by the naked eye (Fig. 1A
and B). From the UVeVis absorption of the GQDs (Fig. 1C), a typical
absorption peak at ca. 320 nm was observed, which was in agree-
ment with a previous report. The photoluminescence (PL) spectra
(Fig. 1D) show that when excited at wavelengths from 400 to
450 nm, the GQDs dispersion exhibits a strong peak at 515 nm
when excited at 420 nm, indicating the size uniformity of the GQDs.
The images of TEM show those GQDs and their size distributions
(Fig. 1E and F). The average diameters of GQDs are around 18 nm.
The Fourier transform infrared spectrum (FT-IR) shows that the
GQDs contain many chemical groups, including eOH, C]O, epoxy/
ether, and eCOeNH2 (Fig. 1G). The nitrogen elements originated
from decomposition of DMF. And the FT-IR of GPE is shown in the
supporting material (Fig. S1).
The presence of such a variety of functional groups can explain
why the GQDs have good solubility and excellent fluorescence
properties [31]. Furthermore, these functional groups clearly
facilitate the post-modification of the GQDs and the fabrication of a
multifunctional drug delivery system (DDS) using GQDs [17]. The
schematic diagram of GQDG is present in Fig. 1H.
3.2. GQDG inhibit the aggregation of Ab1-42 fibrils and the
biocompatibility of GQDG in vitro
To investigate the inhibition effects of GQDs and GQDG in the
aggregation of Ab1-42, the ThT fluorescence assay was performed
(Fig. 2C). ThT associates rapidly with aggregated fibrils of the Ab1-42,
resulting in a new excitation maximum at 450 nm and enhanced
emission at 482 nm. This transformation depends on the aggre-
gated states of Ab1-42, as the monomeric or dimeric peptides would
not be reflected. As shown in Fig. 2C, GQDs and GQDG showed
better inhibitory effects on Ab1-42 aggregation at the concentration
of 200 mg/mL compared to the reference compound resveratrol,
which indicated that our compound was inhibitor for Ab1-42
aggregation.
And TEM was used to further verify the inhibitory effect of GQDs
and GQDG. We analyzed the morphology of the Ab1-42 with or
without GQDs and GQDG through the image of TEM. According to
the TEM images, the typical high-density long linear Ab1-42 fibrils
were detected in the image of untreated Ab1-42 (Fig. 2A). Those fi-
brils compacted in parallel bundles and intercrossed with each
other. In contrast, Ab1-42 samples incubated with GQDs and GQDG
contained only a few short linear fibrils (Fig. 2A) or few amount of
amorphous aggregates (Fig. 2A). The result was consistent with the
results showed in the ThT and CD assays, thereby proving that
GQDs and GQDG could inhibit the aggregation of Ab1-42 fibrils.
Hemocompatibility is the first-level evaluation before the GQDG
was administered via tail vein injection. As shown in Fig. 2B, he-
molysis degree of GQDG is relatively low and the value was 0.29%
(500 mg/mL), 0.18% (200 mg/mL) and 0.13% (50 mg/mL). These results
indicate that the GQDG is safe for intravenous administration.
According to the previous, the GQDs may specifically bind to the
central hydrophobic motif of Ab1-42 peptides. Moreover, the
negative charges of GQDs may interact with positively charged His
residues of Ab1-42 peptides [21]. It is widely accepted that hy-
drophobic interactions and electrostatic interactions are the two
most important factors of inhibiting the Ab1-42 peptide aggrega-
tion [32]. After modified by GPE, the surface area in contact of
GQDG become larger, it may enhance its inhibiting ability. We
concluded that hydrophobic interactions may play more dominant
roles in the process of inhibiting the aggregation. The GQDG may
also specifically bind to the central hydrophobic motif of Ab1-42. In
addition, the GQDG possessed good biocompatibility in vitro.
3.3. In vivo biocompatibility of GQDG
Then we want to further investigate the mechanisms underlying
the efficacy of GQDG in ameliorating the pathogenic processes of
AD in vivo, the APP/PS1 mice were intravenously (i.v.) injected with
20 mg/kg of GQDG (200 mg/mL) every other day for 14 days, no
obvious difference were observed between groups and all the mice
survived. No apparent histopathological abnormalities can be
observed in the kidney and liver from mice with GQDG (Fig. 3).
Our results felled in line with previous experiments. It is re-
ported that GQDs shows excellent biocompatibility, due to their
small size and thus fast clearance from kidneys [33]. Smaller
nanoparticles are more likely to be excreted in urine, thus enabling
fast elimination from the body and preventing excessive accumu-
lation in organs and tissues [34]. On the one hand, we guess that
GQDG might be eliminated through kidneys so it does not cause
obvious damage to mice. On the other hand, GQDs could alleviate
immune-mediated fulminant hepatitis by interfering with T cell
and macrophage activation and possibly by exerting a direct hep-
atoprotective effect [35]. The hepatoprotective effect of GQDs was
associated with the suppression of oxidative stress. This unique
property might contribute to the good biocompatibility of GQDG.
S. Xiao et al. / Biomaterials 106 (2016) 98e110 101

5. 3.4. GQDG attenuates the spatial learning impairment of the APP/
PS1 transgenic mice in the MWM
We conducted the MWM test to examine whether GQDG could
improve the cognitive abilities including reference memory and
working memory in the APP/PS1 transgenic mice (Fig. 4). In the first
day of the acquisition trial with a visible platform, no significant
difference was observed, indicating that there was no discriminable
visual distinction among those groups. During the 2nd day to the
5th day, all mice showed progressive decline in the escape la-
tencies. A two-way ANOVA and post-hoc tests showed that group
GQDG had decreased escape latency compared to the groups Tg Ctrl
and PBS (F [3,27] ¼ 105.27; **P 0.01) (Fig. 4A). In the probe test,
compared to groups Tg Ctrl and PBS, the mice administered with
GQDG showed more platform passing times (F [3,27] ¼ 12.93;
**P 0.01), increased path length (F [3,27] ¼ 10.24; **P 0.01) and
staying time (F [3,27] ¼ 9.48; **P 0.01) in the target quadrant
(Fig. 4BeD). The results illustrated that GQDG could attenuate the
reference memory deficit of the APP/PS1 mice. Moreover, the
decline of escape latency in the reversal test on the 7th and 8th days
displayed that GQDG might also improve the working memory of
the APP/PS1 mice (F [3,27] ¼ 95.51; **P 0.01) (Fig. 4A).
The APP/PS1 mice have an age-related increase in soluble and
insoluble Ab1-40 and Ab1-42, and develop Ab containing plaques
comparable to those observed in the post-mortem brains of human
AD patients at 5e6 months of age [36,37]. And this model is sug-
gested to be a relatively good model for AD progression. The
learning and memory deficit in the APP/PS1 mice were improved
by GQDG. It showed that GQDG treated group had better cognitive
ability than the control groups, which indicated that GQDG could
improve spatial and related forms of learning and memory.
3.5. GQDG decreases the amount of Ab in the brain and serum of
the APP/PS1 transgenic mice
To make a study on the potential inhibited effect of Ab, we
determined the level of Ab load in brain tissue and serum by ELISA
(Fig. 9A, B and C). We measured the quantity of soluble and
Fig. 1. Characterization of GQDs. GQDs dispersed in PBS and illuminated with visible (A) and UV (B) lights. (C) UVeVis absorption of the GQDs. (D) Photoluminescence (PL) spectrum
of GQDs when excited at wavelengths from 400 to 450 nm. (E) Transmission electron microscope image of GQDs. The scale bar is 100 nm. (F) Size distribution of GQDs. (G) Fourier
transform infrared spectrum (FT-IR)of GQDs. (H) Schematic diagram of GQDG.
S. Xiao et al. / Biomaterials 106 (2016) 98e110102

6. Fig. 2. GQDG prevents the aggregation of Ab1-42 in vitro. A, The effect of GQDG and GQDs on Ab1-42 using TEM is shown. A high density of typical linear Ab1-42 amyloid fibrils was
observed in the control group. In contrast, there is fewer fibrils and oligomers in the GQDG and GQDs groups. B, Hemolysis rate of GQDs, GQDG and GPE with different con-
centrations. C, ThT fluorescence assay when incubated with resveratrol, GQDG and GQDs. The scale bar is 100 nm and 200 nm, respectively.
S. Xiao et al. / Biomaterials 106 (2016) 98e110 103

7. insoluble Ab1-42 as well as soluble and insoluble Ab1-40 in the brain
homogenates of the APP/PS1 mice. Then the Ab1-42 and Ab1-40 were
also measured in the serum. Comparing to the Tg Ctrl group, the
quantity of soluble and insoluble Ab1-40 and Ab1-42 in the brain
homogenates (n ¼ 6, **p 0.01, ***p 0.001 in all four tests), and
soluble Ab1-42 in the serum (n ¼ 6, **p 0.01, ***p 0.001) in the
GQDG group decreased significantly. To ensure these results, we
then conducted the IHC staining on the brain slices of the mice
(Figs. 5 and 9D). The surface area of Ab plaque deposits reduced in
the GQDG group compared to the Tg Ctrl groups (n ¼ 6,
***p 0.001).
AD is characterized by Ab and can be observed neurons loss. We
believe that the broken central nervous system (CNS) homeostasis
is the beginning of a series of pathological changes including toxic
protein deposition. In accordance with results in vitro, we observed
a decreased level of the Ab in both the cortex and the hippocampus.
The GQDG might bind to Ab monomer and thus prevent the
monomer assemble into aggregates.
3.6. GQDG reduces the microglial activation
In order to observe Ab associated microglial activation, Ab pla-
ques and activated microglia were double-labeled using IHC tech-
niques. The microglia in the APP/PS1 mice seemed to be activated
into amebocyte morphology from resting-state and tended to sur-
round Ab plaques (Fig. 6), which suggests that some form of
communication occurred between them. Representative immuno-
reactivity of Iba-1 positive cells (green) demonstrated microglial
activation was statistically decreased in the brain (Fig. 6) of GQDG-
treated mice. As observed in confocal micrograph, there seemed to
be less microglial activation surrounding unit Ab plaques in the
GQDG group than in the Tg Ctrl group. To further explore the
microglial activation associated with Ab plaques, the ratio of Ab
associated microglial cells was also calculated (Fig. 9E,
***p 0.001). The statistical analysis verified our prediction. The
microglia was significantly less activated in the GQDG-treated
group than in the Tg Ctrl group.
In the AD brain, Ab could trigger neuroinflammation by acti-
vated microglia that results in the release of pro-inflammatory
cytokine and inflammatory mediators [38]. Although early micro-
glial recruitment can promote Ab clearance and hinder the patho-
logic progression in AD, a persistent microglial accumulation can
also release cytotoxic molecules such as pro-inflammatory cyto-
kines [39]. Microglia are often found near Ab plaques in AD patients
[40] and microglia can facilitate Ab accumulation in return [41].
Sustained activation of microglia and successive release of inflam-
matory mediators resulted in chronic neuroinflammation [42],
which acts as a contributor to activate more microglia and en-
hances more Ab deposition leading to neuronal damage. GQDG
might decreased the microglia activation by reduce the Ab
aggregation.
3.7. Pro-inflammatory factors decreased and anti-inflammatory
increased in GQDG treated group
Inflammatory response is a serious pathology of AD, and the
precursors could be activated by microglia. In this study, inflam-
mation suspension array including cytokines and chemokine was
preformed to test pathological condition in APP/PS1 mice. Our re-
sults showed that several pro-inflammatory cytokines(IL-1a, IL-1b,
IL-6, IL-33, IL-17a, MIP-1b and TNF-a) (Fig. 7AeG, **p 0.01,
***p 0.001), which contribute to both the activation of microglia
and Ab deposition [43], had decreased in GQDG group compared
with Tg Ctrl group. Reversely, anti-inflammatory cytokines (IL-4, IL-
10) (Fig. 7HeI, ***p 0.001) had increased in GQDG group
compared with Tg Ctrl group.
IL-1a and IL-1b, the two distinct isoforms of IL-1, both decrease
greatly indicating that a gradually ameliorative neuroinflammation
in AD [44]. IL-lb stimulates the proliferation of astrocytes, induces
the release of IL-6 and regulates the synthesis of NGF. Activated
microglia may contribute to neuron fibrillary pathology in AD [45].
IL-1a positive microglia was observed in diffuse non-neuritic
Fig. 3. HE stained tissue slices (liver and kidney) of mice injected with GQDG and the Tg Ctrl group. No apparent histopathological abnormalities can be observed in the kidney
and liver from mice with GQDG. The scale bar is 100 nm.
S. Xiao et al. / Biomaterials 106 (2016) 98e110104

8. plaques and dense-core, imply that IL-1a positive microglia was
induced efficiently and may act as a core in forming amyloid pla-
ques formation [46]. The expression of was IL-6 demonstrated to be
induced by TNF-a and IL-1b in vitro [47,48]. As one of the most
significantly changed cytokine, IL-17A induces microglial activa-
tion, inflammation response and vascular pathology [49], which
can be regarded as a reflection to other cytokines such as TNF-a. IL-
33 is responsible for neuroinflammation and associated brain dis-
eases [50]. MIP-1b directly correlated with the age-related pro-
gression of Ab levels in APP/PS1 mice [51]. It plays a role in inducing
the release of IL-1 and TNF-a [52]. TNF-a is an inducible cytokine
with a broad range of pro-inflammatory and plays a cytotoxic role
in several CNS disorders [53]. Though TNF-a overproduction can
induce neuronal apoptosis.
On the contrary, IL-4 act as anti-inflammatory factors and their
production is induced. IL-4 has marked inhibitory effects on the
expression and release of the pro-inflammatory cytokines. It is able
to block the monocyte-derived cytokines, including IL-1, TNF-a, IL-
6, IL-8, IFN-g and MIP-1a. IL-10 is one of the main anti-
inflammatory cytokines and plays an important role in neuronal
homeostasis and cell survival. It exerts anti-inflammatory effect by
inhibiting monocyte/macrophage-derived TNF-a, IL-1, IL-6, IL-8, IL-
12, GMSF, MIP-1a, and MIP-3a [54].
Deposition of Ab is regarded as a beginning of the following
pathological changes in AD. Cytokines and chemokine are proved to
be associated with AD [55]. On the one hand, pro-inflammatory
usually show an aggravation in AD. IL-1a, IL-1b and IL-6 can in-
crease Ab production. Important pro-inflammatory cytokine
including IL-6 and IL-17 correlates closely with the degree of
inflammation [56,57]. It is reported that TNF-a is related to a
deleterious role induced by Ab on promote learning and memory
deficits and synaptic memory mechanisms in AD [58]. Other pro-
inflammatory cytokines such as MIP-1b are shown that they are
highly expressed in AD [59]. On the other hand, anti-inflammatory
factors can somehow block the inflammatory process. IL-4 has
marked inhibitory effects on the expression and release of the pro-
inflammatory cytokines and it is able to block or suppress the
monocyte-derived cytokines [54]. IL-10 has been suggested to play
an important role in neuronal homeostasis and cell survival. These
significant changed levels of cytokines reflected reduced inflam-
matory reaction and neurodegenerative disorder, implicating
treatment effect of GQDG.
3.8. Neuroprotective effect of GQDG
According to our result, NGF and BDNF levels increase signifi-
cantly in the APP/PS1 mice brain (Fig. 9G, H, **p 0.01) in the GQDG
group. NGF are important in the neuronal plasticity and survival of
forebrain cholinergic neurons, which are memory-related [60].
BDNF could regulate synaptic plasticity and neuronal differentia-
tion, so it is crucial to learning and memory [61,62]. These neuro-
trophic factors play an essential role in learning and memory due to
Fig. 4. GQDG improves the spatial learning abilities of APP/PS1 mice in MWM. A: comparison of latency time of each group in learning trails. B: the path length of each group in
each quadrant in the probe trial, with the first quadrant set as the target quadrant. Mice treated with GQDG have longer path length in the target quadrant than in any other
quadrants. C: the time of each group in each quadrant in the probe trial. D: the numbers passing the platform within 60s in the probe trial. Compared to the Tg ctrl group, mice
treated with GQDG were expected to spend more time, with a longer path length, in the target quadrant than in any other quadrant, and to pass the platform more. (n ¼ 12,
**P 0.01, ***P 0.001).
S. Xiao et al. / Biomaterials 106 (2016) 98e110 105

9. their effect on neuronal and synaptic plasticity [63]. We speculate
that because GQDG can reduce the Ab level, and then it might have
a positive effect on the expression of NGF and BDNF. Also, the
neuroprotective effect of GPE could contribute the increased level
of NGF and BDNF. These findings might provide new insights into
the neuroprotection mechanism of GQDG.
Fig. 5. GQDG decreases Ab plaques in the brains of APP/PS1 mice. The results were obtained after immunohistochemical staining of the hippocampus and cortex. Group Tg Ctrl and
group GQDG are presented. The results show a decrease in Ab1-42 aggregation in the GQDG-treated groups.
Fig. 6. Confocal micrographs show that Ab plaques (red) and activated microglial cells (green) are doubly labeled immunohistochemically in the brain. The rigth are high-powered
magnification of individual plaques. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
S. Xiao et al. / Biomaterials 106 (2016) 98e110106

10. 3.9. GQDG could increase density of dendritic spines in the brain of
APP/PS1 mice
Diolistic labeling is a common method to observe the dendritic
spines of neuron in vivo, and the reduction of dendritic spines is
related to the decline of the synaptic plasticity and neuron
impairment. We suspect that GQDG could also protect the synapsis,
so we preformed the diolistic labeling in the brain slice. From the
result, we found that the number of dendritic spines increased in
the GQDG group compared to the Tg Ctrl group (Figs. 8 and 9F,
***p 0.001).
A significant decrease of dendritic spines is part of abnormal
changes of dendrites, which is associated with a decline in synaptic
plasticity in the cortex and hippocampus. Ab aggregation increased
oxidative stress level might lead to this lesion [64]. If Ab level could
be regulated early in AD, the cognitive decline might be prevented
[65]. We verified that GQDG could improve learning and memory
ability in APP/PS1 mice through MWM, it lead us to investigate the
effect of GQDG in the dendritic spines. So we examined the density
of dendritic spines of neurons in the hippocampus by diolistic la-
beling. Based on the result, we speculated that the synapsis pro-
tection effect might come from the reduction of Ab aggregation.
Also, we believed this effect could be the comprehensive results
improved microenvironment of whole brain, including the
increased neurotrophic factors level. However, further work should
be carried out to investigate this mechanism.
3.10. GQDG could promote neurogenesis in APP/PS1 mice
To further investigate neuroprotective mechanism of GQDG, we
examined the number of newborn NPCs and neurons in hippo-
campus. we found that the number of newborn NPCs (BrdUþ/
DCXþ) and neurons (BrdUþ/NeuNþ) in the GQDG group increased
compared with the Tg Ctrl group (Figs. 9I and 10) (n ¼ 6, **p 0.01,
***p 0.001). The significant increase of newborn NPCs and neu-
rons in GQDG group showed that GQDG has neurogenesis effects
and could attenuate the cognitive deficits in AD.
Fig. 7. The results of Bio-plex suspension array (IL-1a, IL-1b, IL-6, IL-33, IL-17a, MIP-1b, TNF-a, IL-4 and IL-10). (**p 0.01; ***P 0.001, n ¼ 6).
Fig. 8. GQDG increases the density of dendritic spines in the brains of APP/PS1 mice.
The GQDG-treated groups, especially Group 20 mg, have the densest dendritic spines.
S. Xiao et al. / Biomaterials 106 (2016) 98e110 107

11. Neurodegeneration and neuronal loss is the main pathological
feature in AD brains. There is a decline in hippocampal neuro-
genesis in the process of AD, it has been shown to link to the Ab
aggregation [66]. It could be a potential therapeutic strategy to
promote the neurogenesis in order to prevent cognitive decline in
AD. In this study, we used BrdU labeling to identify newborn NPCs
and neurons. Both the amount of two kinds of cells increased in the
GQDG group, which reflects the neurogenesis effect of GQDG. We
Fig. 9. A: insoluble Ab1-40/Ab1-42 in the brain. B: soluble Ab1-40/Ab1-42 in the brain. C: Ab1-40/Ab1-42 in the Serum. D: statistical analysis shows a remarkable reduction in Ab load in
the brain after the administration of GQDG compared to the Tg Ctrl group. E: The histogram shows a significantly smaller degree of microglial activation in GQDG groups compared
with Tg Ctrl group. F: statistical analysis of the spine density. The spine density of GQDG-treated groups significantly increase compared with the Tg Ctrl group. G: the ELISA result
for the BDNF quantity. The data show an obvious increase in BDNF levels in the GQDG group. H: the ELISA result for the NGF quantity. The data show an obvious increase in BDNF
levels in the GQDG group. I: statistical analysis shows that silibinin increases the number of neurons (BrdUþ/NeuNþ) and NPCs (BrdUþ/DCXþ). (**p 0.01; ***P 0.001, n ¼ 6).
Fig. 10. GQDG increases the number of newly generated cells in the hippocampus in APP/PS1 mice. The picture shows the increase in newly generated NPCs (BrdUþ/DCXþ) and
neurons (BrdUþ/NeuNþ).
S. Xiao et al. / Biomaterials 106 (2016) 98e110108

12. analyzed that GQDG might enhance the neurogenesis through the
following mechanisms. In the first place, GQDG could decrease Ab
aggregation and reduce the neuroinflammation caused by Ab. It
played an important in the improvement of the microenvironment
of whole brain and increase neurogenesis. Moreover, GPE can dose-
dependently protect hippocampus neurons from NMDA-induced
neuronal toxicity by reducing calcium overload in the brain of the
APP/PS1 mice. That might be necessary for neuron reproduction.
Thirdly, GPE might also be a direct stimulus for neural proliferation.
4. Conclusion
In conclusion, we proved that GQDs could act as a novel drug
carrier to transport neuroprotective peptide GPE to CNS, and an
in vitro and in vivo model has been described to investigate the
neuroprotective effect of GQDG. The inhibition effect on the ag-
gregation of Ab1-42 fibrils was verified by ThT and TEM. The learning
and memory capacity of APP/PS1 mice was improved in the Morris
water maze test. Also, the Ab plaque deposits reduced in the GQDG
group. More importantly, the number of newly generated neuronal
precursor cell and neuron increased. Finally, GQDG decreased some
pro-inflammatory cytokines including IL-1a, IL-1b, IL-6, IL-33, IL-
17a, MIP-1b and TNF-a and increased two anti-inflammatory cy-
tokines (IL-4, IL-10). These results can prevent the aggregation of Ab
and reduce the inflammatory response, thus protect the synapse
and promote the neurogenesis, ultimately improves the learning
and memory ability of APP/PS1 mice. It may change the microen-
vironment, accelerating the differentiation of newly proliferating
cells into neurons, thus contributing to improved behaviors. This
study may provide a novel point for AD therapy.
Acknowledgments
This study was supported by grants to Jun Liu from the National
Natural Science Foundation of China (No. 81372919), the Guang-
dong Natural Science Foundation (No. 2014A030313016), the grants
to Rui Guo from the National Natural Science Foundation of China
(No. 51303064), and the grants to Ping Luan from the Science and
Technology Planning Fundamental Research Project of Shenzhen
(No. JCYJ20150324140036853).
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.biomaterials.2016.08.021.
References
[1] F. Re, M. Gregori, M. Masserini, Nanotechnology for neurodegenerative dis-
orders, Nanomed- Nanotechnol. 8 (2012). S51eS8.
[2] J. Hardy, D.J. Selkoe, The amyloid hypothesis of Alzheimer's disease: progress
and problems on the road to therapeutics, Science 297 (2002) 353e356.
[3] T. Demeritte, B.P. Viraka Nellore, R. Kanchanapally, S.S. Sinha, A. Pramanik,
S.R. Chavva, et al., Hybrid graphene oxide based plasmonic-magnetic multi-
functional nanoplatform for selective separation and label-free identification
of Alzheimer's disease biomarkers, ACS Appl. Mater Inter 7 (2015)
13693e13700.
[4] E. Karran, M. Mercken, B. De Strooper, The amyloid cascade hypothesis for
Alzheimer's disease: an appraisal for the development of therapeutics, Nat.
Rev. Drug Discov. 10 (2011) 698e712.
[5] G.M. Shankar, B.L. Bloodgood, M. Townsend, D.M. Walsh, D.J. Selkoe,
B.L. Sabatini, Natural oligomers of the Alzheimer amyloid-b protein induce
reversible synapse loss by modulating an NMDA-type glutamate receptor-
dependent signaling pathway, J. Neurosci. 27 (2007) 2866e2875.
[6] Z. Cai, M.D. Hussain, L.-J. Yan, Microglia, neuroinflammation, and beta-amyloid
protein in Alzheimer's disease, Int. J. Neurosci. 124 (2014) 307e321.
[7] S.J. Hopkins, N.J. Rothwell, Cytokines and the nervous system I: expression
and recognition, Trends Neurosci. 18 (1995) 83e88.
[8] N. Ikeda, H. Suganuma, N. Ohkawa, S. Nagata, H. Shoji, T. Shimizu, Measure-
ment of cytokine levels in cerebrospinal fluid over time in
neonatalEnterococcal meningitis, Pediatr. Int. 56 (2014) e45ee7.
[9] C.S. Raine, Inflammation in Alzheimer's disease: a view from the periphery,
Neurobiol. Aging 21 (2000) 437e440 discussion 51e3.
[10] D.D. Kitts, K. Weiler, Bioactive proteins and peptides from food sources. Ap-
plications of bioprocesses used in isolation and recovery, Curr. Pharm. Des. 9
(2003) 1309e1323.
[11] J. Guan, G. Thomas, H. Lin, S. Mathai, D. Bachelor, S. George, et al., Neuro-
protective effects of the N-terminal tripeptide of insulin-like growth factor-1,
glycine-proline-glutamate (GPE) following intravenous infusion in hypo-
xiceischemic adult rats, Neuropharmacology 47 (2004) 892e903.
[12] J. Guan, P. Harris, M. Brimble, Y. Lei, J. Lu, Y. Yang, et al., The role for IGF-1-
derived small neuropeptides as a therapeutic target for neurological disor-
ders, Expert Opin. Ther. Tar 19 (2015) 785e793.
[13] E. Burgos-Ramos, G.A. Martos-Moreno, M.G. Lopez, R. Herranz, D. Aguado-
Llera, J. Egea, et al., The N-terminal tripeptide of insulin-like growth factor-I
protects against b-amyloid-induced somatostatin depletion by calcium and
glycogen synthase kinase 3b modulation, J. Neurochem. 109 (2009) 360e370.
[14] D. Aguado-Llera, M. Martín-Martínez, M.T. García-Lopez, E. Arilla-Ferreiro,
V. Barrios, Gly-Pro-Glu protects b-amyloid-induced somatostatin depletion in
the rat cortex, Neuroreport 15 (2004) 1979e1982.
[15] G. Tosi, L. Costantino, F. Rivasi, B. Ruozi, E. Leo, A. Vergoni, et al., Targeting the
central nervous system: in vivo experiments with peptide-derivatized
nanoparticles loaded with Loperamide and Rhodamine-123, J. Control
Release 122 (2007) 1e9.
[16] Y. Zhu, S. Murali, W. Cai, X. Li, J.W. Suk, J.R. Potts, et al., Graphene and gra-
phene oxide: synthesis, properties, and applications, Adv. Mater 22 (2010)
3906e3924.
[17] X. Sun, Z. Liu, K. Welsher, J.T. Robinson, A. Goodwin, S. Zaric, et al., Nano-
graphene oxide for cellular imaging and drug delivery, Nano Res. 1 (2008)
203e212.
[18] L. Li, G. Wu, G. Yang, J. Peng, J. Zhao, J.-J. Zhu, Focusing on luminescent gra-
phene quantum dots: current status and future perspectives, Nanoscale 5
(2013) 4015e4039.
[19] M. Nurunnabi, Z. Khatun, K.M. Huh, S.Y. Park, D.Y. Lee, K.J. Cho, et al., In vivo
biodistribution and toxicology of carboxylated graphene quantum dots, ACS
Nano 7 (2013) 6858e6867.
[20] Wang X-y, R. Lei, Huang H-d, N. Wang, L. Yuan, R.-Y. Xiao, et al., The
permeability and transport mechanism of graphene quantum dots (GQDs)
across the biological barrier, Nanoscale 7 (2015) 2034e2041.
[21] Y. Liu, L.-P. Xu, W. Dai, H. Dong, Y. Wen, X. Zhang, Graphene quantum dots for
the inhibition of b amyloid aggregation, Nanoscale 7 (2015) 19060e19065.
[22] W.S. Hummers Jr., R.E. Offeman, Preparation of graphitic oxide, J. Am. Chem.
Soc. 80 (1958) 1339.
[23] S. Zhu, J. Zhang, C. Qiao, S. Tang, Y. Li, W. Yuan, et al., Strongly green-
photoluminescent graphene quantum dots for bioimaging applications,
Chem. Commun. 47 (2011) 6858e6860.
[24] J. Qiu, R. Zhang, J. Li, Y. Sang, W. Tang, P.R. Gil, et al., Fluorescent graphene
quantum dots as traceable, ph-sensitive drug delivery systems, Int. J.
Nanomed 10 (2015) 6709.
[25] H. Levine, Thioflavine T interaction with synthetic Alzheimer's disease b-
amyloid peptides: detection of amyloid aggregation in solution, Protein Sci. 2
(1993) 404e410.
[26] J. Tan, N. Meng, Y. Fan, Y. Su, M. Zhang, Y. Xiao, et al., Hydroxypropyl-b-
cyclodextrinegraphene oxide conjugates: carriers for anti-cancer drugs,
Mater Sci. Eng. C 61 (2016) 681e687.
[27] C.V. Vorhees, M.T. Williams, Morris water maze: procedures for assessing
spatial and related forms of learning and memory, Nat. Protoc. 1 (2006)
848e858.
[28] Y. Ziv, N. Ron, O. Butovsky, G. Landa, E. Sudai, N. Greenberg, et al., Immune
cells contribute to the maintenance of neurogenesis and spatial learning
abilities in adulthood, Nat. Neurosci. 9 (2006) 268e275.
[29] J.A. O'Brien, S.C. Lummis, Diolistic labeling of neuronal cultures and intact
tissue using a hand-held gene gun, Nat. Protoc. 1 (2006) 1517e1521.
[30] G.K. Seabold, J.B. Daunais, A. Rau, K.A. Grant, V.A. Alvarez, DiOLISTIC labeling
of neurons from rodent and non-human primate brain slices, J. Vis. Exp. 41
(2010).
[31] X. Zhou, S. Guo, J. Zhang, Solution-processable graphene quantum dots,
ChemPhysChem. 14 (2013) 2627e2640.
[32] M. Mahmoudi, O. Akhavan, M. Ghavami, F. Rezaee, S.M.A. Ghiasi, Graphene
oxide strongly inhibits amyloid beta fibrillation, Nanoscale 4 (2012)
7322e7325.
[33] Y. Chong, Y. Ma, H. Shen, X. Tu, X. Zhou, J. Xu, et al., The in vitro and in vivo
toxicity of graphene quantum dots, Biomaterials 35 (2014) 5041e5048.
[34] M. Nurunnabi, Z. Khatun, M. Nafiujjaman, D-g Lee, Y-k Lee, Surface coating of
graphene quantum dots using mussel-inspired polydopamine for biomedical
optical imaging, ACS Appl. Mater Inter 5 (2013) 8246e8253.
[35] V. Volarevic, V. Paunovic, Z. Markovic, B. Simovic Markovic, M. Misirkic-
Marjanovic, B. Todorovic-Markovic, et al., Large graphene quantum dots
alleviate immune-mediated liver damage, ACS Nano 8 (2014) 12098e12109.
[36] R. Minkeviciene, J. Ihalainen, T. Malm, O. Matilainen, V. Keksa-Goldsteine,
G. Goldsteins, et al., Age-related decrease in stimulated glutamate release and
vesicular glutamate transporters in APP/PS1 transgenic and wild-type mice,
J. Neurochem. 105 (2008) 584e594.
[37] B. Laursen, A. Mørk, N. Plath, U. Kristiansen, J.F. Bastlund, Cholinergic
degeneration is associated with increased plaque deposition and cognitive
S. Xiao et al. / Biomaterials 106 (2016) 98e110 109

13. impairment in APPswe/PS1dE9 mice, Behav. Brain Res. 240 (2013) 146e152.
[38] A.M. Szczepanik, G.E. Ringheim, IL-10 and glucocorticoids inhibit Ab (1e42)-
and lipopolysaccharide-induced pro-inflammatory cytokine and chemokine
induction in the central nervous system, J. Alzheimers Dis. 5 (2003) 105e117.
[39] S.E. Hickman, E.K. Allison, J. El Khoury, Microglial dysfunction and defective b-
amyloid clearance pathways in aging Alzheimer's disease mice, J. Neurosci. 28
(2008) 8354e8360.
[40] J. Rozemuller, P. Eikelenboom, Activated microglia and cerebral amyloid de-
posits in Alzheimer's disease, Res. Immunol. 143 (1992) 646e649.
[41] S.A. Grathwohl, R.E. K€alin, T. Bolmont, S. Prokop, G. Winkelmann, S.A. Kaeser,
et al., Formation and maintenance of Alzheimer's disease b-amyloid plaques
in the absence of microglia, Nat. Neurosci. 12 (2009) 1361.
[42] W.J. Streit, R.E. Mrak, W.S.T. Griffin, Microglia and neuroinflammation: a
pathological perspective, J. Neuroinflamm 1 (2004) 14.
[43] S. Zaheer, R. Thangavel, Y. Wu, M.M. Khan, D. Kempuraj, A. Zaheer, Enhanced
expression of glia maturation factor correlates with glial activation in the
brain of triple transgenic Alzheimer's disease mice, Neurochem. Res. 38
(2013) 218e225.
[44] S Shaftel Solomon, WSTGMKOB, The role of interleukin-1 in neuro-
inflammation and Alzheimer, 2008.
[45] L.L. Yuekui Li, W. Steven, W. Barger, Sue T. Griffin, Interleukin-1 mediates
pathological effects of microglia on tau, 2003.
[46] W.S. Griffin, J.G. Sheng, G.W. Roberts, R.E. Mrak, Interleukin-1 expression in
different plaque types in Alzheimer's disease: significance in plaque evolution,
J. Neuropath Exp. Neur 54 (1995) 276e281.
[47] G.E. Ringheim, K.L. Burgher, J.A. Heroux, Interleukin-6 mRNA expression by
cortical neurons in culture: evidence for neuronal sources of interleukin-6
production in the brain, J. Neuroimmunol. 63 (1995) 113e123.
[48] R.A. Gadient, U.H. Otten, Interleukin-6 (IL-6)ea molecule with both beneficial
and destructive potentials, Prog. Neurobiol. 52 (1997) 379e390.
[49] J. Zimmermann, M. Krauthausen, M.J. Hofer, M.T. Heneka, I.L. Campbell,
M. Muller, CNS-targeted production of IL-17A induces glial activation,
microvascular pathology and enhances the neuroinflammatory response to
systemic endotoxemia, PLoS One 8 (2013) e57307.
[50] G. Singhal, E.J. Jaehne, F. Corrigan, C. Toben, B.T. Baune, Inflammasomes in
neuroinflammation and changes in brain function: a focused review, Front.
Neurosci. 8 (2014) 315.
[51] M. Zhu, J.S. Allard, Y. Zhang, E. Perez, E.L. Spangler, K.G. Becker, et al., Age-
related brain expression and regulation of the chemokine CCL4/MIP-1beta in
APP/PS1 double-transgenic mice, J. Neuropath Exp. Neur 73 (2014) 362e374.
[52] A. Nath, S Fau Chattopadhya, U. Chattopadhyay, U Fau Chattopadhyay,
N.K. Sharma, N.K. Sharma, Macrophage inflammatory protein (MIP)1alpha
and MIP1beta differentially regulate release of inflammatory cytokines and
generation of tumoricidal monocytes in malignancy, 2006.
[53] V. Alvarez, I.F. Mata, P. Gonzalez, C.H. Lahoz, C. Martinez, J. Pena, et al.,
Association between the TNFalpha-308 A/G polymorphism and the onset-age
of Alzheimer disease, Am. J. Med. Genet. 114 (2002) 574e577.
[54] J.M. Rubio-Perez, J.M. Morillas-Ruiz, A review: inflammatory process in Alz-
heimer's disease, role of cytokines, Sci. World J. 2012 (2012) 756357.
[55] K.K. Modi, A. Roy, S. Brahmachari, S.B. Rangasamy, K. Pahan, Cinnamon and its
metabolite sodium benzoate attenuate the activation of p21rac and protect
memory and learning in an animal model of Alzheimer's disease, PLoS One 10
(2015) e0130398.
[56] R.A. Quintanilla, D.I. Orellana, C. Gon_zalez-Billault, R.B. Maccioni, Interleukin-6
induces Alzheimer-type phosphorylation of tau protein by deregulating the
cdk5/p35 pathway, Exper Cell Res. 295 (2004) 245e257.
[57] P. Chakrabarty, K. Jansen-West, A. Beccard, C. Ceballos-Diaz, Y. Levites,
C. Verbeeck, et al., Massive gliosis induced by interleukin-6 suppresses A?
deposition in vivo: evidence against inflammation as a driving force for am-
yloid deposition, FASEB J. 24 (2010) 548e559.
[58] F.E. McAlpine, J.K. Lee, A.S. Harms, K.A. Ruhn, M. Blurton-Jones, J. Hong, et al.,
Inhibition of soluble TNF signaling in a mouse model of Alzheimer's disease
prevents pre-plaque amyloid-associated neuropathology, Neurobiol. Dis. 34
(2009) 163e177.
[59] P. Grammas, R. Ovase, Inflammatory factors are elevated in brain microvessels
in Alzheimer's disease? Neurobio Aging 22 (2001) 837e842.
[60] S. Korsching, G. Auburger, R. Heumann, J. Scott, H. Thoenen, Levels of nerve
growth factor and its mRNA in the central nervous system of the rat correlate
with cholinergic innervation, EMBO J. 4 (1985) 1389e1393.
[61] W.J. Tyler, M. Alonso, C.R. Bramham, L.D. Pozzo-Miller, From acquisition to
consolidation: on the role of brain-derived neurotrophic factor signaling in
hippocampal-dependent learning, Learn Mem. 9 (2002) 224e237.
[62] M.G. Murer, Q. Yan, R. Raisman-Vozari, Brain-derived neurotrophic factor in
the control human brain, and in Alzheimer's disease and Parkinson's disease,
Prog. Neurobio 63 (2001) 71e124.
[63] F. Angelucci, F. Gelfo, M. Fiore, N. Croce, A.A. Mathe, S. Bernardini, et al., The
effect of neuropeptide Y on cell survival and neurotrophin expression in in-
vitro models of Alzheimer's disease, Can. J. Physiol. Pharmacol. 92 (2014)
621e630.
[64] D.L. Moolman, O.V. Vitolo, J.-P.G. Vonsattel, M.L. Shelanski, Dendrite and
dendritic spine alterations in Alzheimer models, J. Neurocyto 33 (2004)
377e387.
[65] J.J. Palop, L. Mucke, Amyloid-[beta]-induced neuronal dysfunction in Alz-
heimer's disease: from synapses toward neural networks, Nat. Neurosci. 13
(2010) 812e818.
[66] A. Faure, L. Verret, B. Bozon, N. El Tannir El Tayara, M. Ly, F. Kober, et al.,
Impaired neurogenesis, neuronal loss, and brain functional deficits in the
APPxPS1-Ki mouse model of Alzheimer's disease, Neurobio Aging 32 (2011)
407e418.
S. Xiao et al. / Biomaterials 106 (2016) 98e110110