Nanodisk-based glioma-targeted drug delivery enabled by a stable glycopeptide

Article abstract of DOI:10.1016/j.jconrel.2018.06.006

Heptapeptide ATWLPPR (A7R) binds specifically to vascular endothelial growth factor receptor 2 (VEGFR2) and neuropilin-1 (NRP-1) overexpressed in glioma cells, exhibiting high potential to achieve glioma targeted drug delivery. However, in vivo application of A7R peptide remains challenging due to the poor proteolytic stability and inaccessibility of A7R to the brain. To tackle these problems, we identified a glycosylated A7R derivative to enhance in vivo stability and brain transport efficacy. Our results showed that glycosylation of peptide could efficiently improve stability in serum, traverse the blood-brain barrier (BBB) and be uptaken by glioma cells. Furthermore, a novel glioma-targeted drug delivery system was constructed successfully employing glycopeptide as the targeting moiety and nanodisk as the carrier of paclitaxel (PTX). Physicochemical characterization showed that the nanodisk presented suitable size of 50 nm and adequate loading capacity of PTX. Compared to non-glycosylated nanodisk, glycopeptide modification could significantly enhance the uptake of disks by brain capillary endothelial cells through glucose transporter 1 (GLUT1). In vivo imaging and glioma fluorescence section results also indicated that nanodisks modified with glycopeptide showed a higher accumulation in glioma. The glycopeptide-enabled PTX delivery system exhibited superior anti-glioma efficacy in intracranial glioma xenograft model. These results suggested that glycosylation of peptides provided an efficient pathway to design multifunctional and stable brain targeting ligands.

Full text of DOI:10.1016/j.jconrel.2018.06.006

Accepted Manuscript
Nanodisk-based glioma-targeted drug delivery enabled by a stable
glycopeptide
Huan Wang, Xiaoyi Wang, Cao Xie, Mingfei Zhang, Huitong
Ruan, Songli Wang, Kuan Jiang, Fei Wang, Changyou Zhan,
Weiyue Lu, Hao Wang
PII:
DOI:
Reference:
S0168-3659(18)30331-6
doi:10.1016/j.jconrel.2018.06.006
COREL 9330
To appear in:
Journal of Controlled Release
Received date:
Revised date:
Accepted date:
19 February 2018
16 May 2018
4 June 2018
Please cite this article as: Huan Wang, Xiaoyi Wang, Cao Xie, Mingfei Zhang, Huitong
Ruan, Songli Wang, Kuan Jiang, Fei Wang, Changyou Zhan, Weiyue Lu, Hao Wang ,
Nanodisk-based glioma-targeted drug delivery enabled by a stable glycopeptide. Corel
(2017), doi:10.1016/j.jconrel.2018.06.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.

ACCEPTED MANUSCRIPT
Nanodisk-based glioma-targeted drug delivery enabled by a stable glycopeptide
Huan Wanga,b, Xiaoyi Wanga, Cao Xiea, Mingfei Zhanga, Huitong Ruana, Songli Wanga, Kuan
Jianga, Fei Wanga, Changyou Zhana,c, Weiyue Lua,d,**, Hao Wangb,*
aDepartment of Pharmaceutics, School of Pharmacy, and Key Laboratory of Smart Drug Delivery,
Fudan University, Shanghai 201203, & State Key Laboratory of Medical Neurobiology, and
Collaborative Innovation Center for Brain Science, Fudan University, Shanghai 200032, China.
bNational Pharmaceutical Engineering Research Center, China State Institute of Pharmaceutical
Industry, Shanghai 201203, China.
cDepartment of Pharmacology, School of Basic Medical Sciences, Fudan University, Shanghai
200032, & State Key Laboratory of Molecular Engineering of Polymers, Fudan University,
Shanghai, 200433, China.
dDepartment of Integrative Medicine, Huashan Hospital, Fudan University, & Institute of
Integrative Medicine of Fudan University, Shanghai 200041, and Minhang Branch, Zhongshan
Hospital and Institute of Fudan-Minghang Acadimic Health System, Minghang Hospital, Fudan
University, Shanghai 201199, China.
* Conrespondence to: Hao Wang. e-mail: wanghao10@sinopharm.com; tel.: +86-21-51320211;
fax: +86-21-51320728;
** Conrespondence to: Weiyue Lu. e-mail: wylu@shmu.edu.cn; tel.: +86-21-51980006; fax:
+86-21-51980090.

ACCEPTED MANUSCRIPT
Abstract
Heptapeptide ATWLPPR (A7R) binds specifically to vascular endothelial growth factor receptor 2
(VEGFR2) and neuropilin-1 (NRP-1) overexpressed in glioma cells, exhibiting high potential to
achieve glioma targeted drug delivery. However, in vivo application of A7R peptide remains
challenging due to the poor proteolytic stability and inaccessibility of A7R to the brain. To tackle
these problems, we identified a glycosylated A7R derivative to enhance in vivo stability and brain
transport efficacy. Our results showed that glycosylation of peptide could efficiently improve
stability in serum, traverse the blood-brain barrier (BBB) and be uptaken by glioma cells.
Furthermore, a novel glioma-targeted drug delivery system was constructed successfully
employing glycopeptide as the targeting moiety and nanodisk as the carrier of paclitaxel (PTX).
Physicochemical characterization showed that the nanodisk presented suitable size of 50 nm and
adequate loading capacity of PTX. Compared to non-glycosylated nanodisk, glycopeptide
modification could significantly enhance the uptake of disks by brain capillary endothelial cells
through glucose transporter 1 (GLUT1). In vivo imaging and glioma fluorescence section results
also indicated that nanodisks modified with glycopeptide showed a higher accumulation in glioma.
The glycopeptide-enabled PTX delivery system exhibited superior anti-glioma efficacy in
intracranial glioma xenograft model. These results suggested that glycosylation of peptides
provided an efficient pathway to design multifunctional and stable brain targeting ligands.
Keywords: Glycopeptide; Nanodisk; Glucose Transporter; blood-brain barrier; Glioma

ACCEPTED MANUSCRIPT
1. Introduction
Glioblastoma multiforme (GBM) is the most malignant primary brain tumor with high
mortality and poor prognosis. The median survival of GBM patients after surgery is 8 to 15
months and the overall 5-year survival rate is below 5%[1.2]. Given extremely infiltrative nature
of glioma, chemotherapy remains an effective adjunct treatment to surgical resection in clinic[3].
In recent years, actively targeted drug delivery systems have demonstrated the potential to
improve antitumor efficacy of chemotherapeutic agents by functionalizing targeting ligands on the
surface of nanocarriers. Peptide ligands have been extensively exploited for actively targeting
nanomedicines due to their ease of synthesis and screening[4-7].
A mutated phage library selected peptide A7R (ATWLPPR) exhibited specific binding to
vascular endothelial growth factor receptor 2 (VEGFR2) and neuropilin-1 (NRP-1), which are
overexpressed in glioma cells. In our previous study, A7R-functionalized drug delivery systems
have been developed for targeting subcutaneous glioma xenograft in nude mice, effectively
inhibiting tumor growth[8]. However, multiple obstacles restrict the precise delivery of
chemotherapeutic agents to the brain as well as the brain tumors. The A7R peptide is subject to
proteolysis in blood after systemic administration. The blood-brain barrier (BBB) would present
additional obstacle to brain tumor targeted delivery of A7R functionalized drug delivery systems.
Membrane transporters such as glucose transporter (GLUT) have been employed to facilitate
the brain transport of drug delivery systems. GLUT1, a representative member in the GLUT
family, is an endogenous transporter avidly presenting in brain microvessels[9]. D-glucose
moieties are the essential fuel for the large and uninterrupted energy demand of the brain[10]. Due
to this high capacity of glucose utilization and transport efficiency, glucose moieties are
considered good candidates to facilitate brain delivery[11-13].
Peptides have been stabilized by different methods, including the design of retro-all-D or
retro-inverso peptide analogues[14], head-to-tail cyclization[8], terminal amino acid modification
and introduction of non-natural amino acids[15]. Among them, glycosylation is among the
simplest and the most effective methods. Conjugation of sugar moieties to opioid peptides leads to
improved bioavailability in brain, as a result of prolonged metabolic and clearance half-live[16].
N-terminal glycosylation improved the resistance of glucagon-like peptide-1 to
dipeptidyl-peptidase-IV (DPP-IV) and exhibited substantially improved biological
activities[17.18].
Thus, we hypothesize that the introduction of glucose moieties on A7R would be
multifunctional. The stable glycopeptide may have brain and glioma dual-targeted effect. Firstly,
the glycosylated A7R has better biological stability and prolonged blood circulation time in vivo.
Then the glucose moiety enables brain transport of peptide via GLUT-mediated pathway. After
entry into brain, the glycosylated A7R peptide would further target the tumor region by
recognizing the overexpressed VEGFR2 and NRP-1.
Recently, polyethylene glycol (PEG)-stabilized nanodisk, formed in mixtures of
phospholipids, cholesterol and PEG lipids, has received increasing interest in drug delivery. The
nanodisk is a nano-sized flat circular lipid bilayer surrounded by a highly curved rim with
PEG-lipids readily distributing on the rim of lipid disks. A series of linear alpha-helical peptides,
such as melittin, showed high affinity for the rim of PEG-stabilized lipid disks. The
melittin-loaded lipid disks significantly reduced the hemolytic effect and efficiently inhibited
tumor growth in a tumor bearing mice model[19]. The PEG-stabilized disks also provided a larger

ACCEPTED MANUSCRIPT
hydrophobic volume for chemotherapeutics compared to micelles and that their size could be
tailored for loading different drugs[20]. DOX-loaded disks showed enhanced bioavailability and
improved anticancer effect in comparison to free DOX[21].
In this work we synthesized a series of glycosylated analogues of A7R and compared their
serum stability, brain and glioma targeting ability. The optimal glycosylated peptide was
conjugated on the surface of nanodisk for dual-targeted drug delivery as illustrated in Scheme 1.
Cellular uptake efficiency, tumor spheroid penetration, cytotoxicity and biodistribution of
nanodisk formulations were investigated. The anti-glioma effects of paclitaxel-loaded nanodisk
formulations were evaluated in an intracranial glioma-bearing nude mice model.
Scheme 1. Schematic illustration of glycopetide-modified nanodisk as a dual-targeted drug
delivery system for anti-glioma therapy.
2. Materials and Methods
2.1 Materials and reagents
Fmoc-protected α-amino acids were supplied by GL Biochem Ltd. (Shanghai, China).
Fmoc-Rink Amide MBHA resin was supplied by Xi’an innovision bioscience Co. Ltd. (Xi’an,
China). Pentaacetate glucose was purchased from Aladdin (Shanghai, China).
O-benzotriazole-N,N,N′,N′-tetramethyl-uronium-hexafluorophosphate
(HBTU)
and
1-Hydroxybenzotriazole anhydrous (HOBt) were purchased from American Bioanalytical Co. Ltd.
(Massachusetts, USA). Diisopropylethylamine (DIEA) was supplied by Sigma-Aldrich (St. Louis,
MO, USA). Fluorescein-5-maleimide and Cyanine7-maleimide were obtained from Fanbo
Biochemicals (Beijing, China). Rat serum was supplied by Invitrogen (Grand Island, NY).
Matrigerl Basement Membrane Matrix was purchased from BD Biosciences (USA). POPC
(1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocho-line) and mPEG2000-DSPE were purchased from
Lipoid GmbH (Ludwigshafen, Germany). Cholesterol was purchased from Sinopharm Chemical
Reagent Co. Ltd (Shanghai, China). Mal-PEG3400-DSPE and NH2-PEG3400-DSPE were
purchased from Laysan Bio (Arab, USA). FITC was purchased from Sigma-Aldrich (St. Louis,
USA). Sephadex G50 was purchased from GE Healthcare (Uppsala, Sweden). DAPI
(4',6-diamidino-2-phenylindole) was supplied by Roche (Basel, Switzerland). DiR
(1,1'-dioctadecyl-3,3,3′,3′-tetramethyl
indotricarbocyanine
iodide),
DiD
(1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt dye) ,
and Paclitaxel were purchased from Meilun Biology Technology Co. Ltd. (Dalian, China).
Methanol and acetonitrile were purchased from Fisher Scientific (USA). N,N-dimethyformamide
(DMF), dichloromethane (DCM) and other chemical reagents were purchased from Sinopharm Co.
Ltd. (Shanghai, China). All chemicals were of analytic reagent grades.
Human glioblastoma cells (U87MG cells), Brain capillary endothelial cells (bEnd.3) and
Human umbilical vascular endothelial cells (HUVECs) were all obtained from Shanghai Institute
of Cell Biology and maintained in Dulbecco's Modified Eagle Medium (Gibco) supplemented
with 10% FBS (Gibco), 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C under a
humidified atmosphere containing 5% CO2.
Male ICR mice and BALB/c nude mice of 4-6 weeks age were purchased from the BK Lab
Animal Ltd. (Shanghai, China) and kept under SPF conditions. All animal experiments were
carried out in accordance with guidelines evaluated and approved by the ethics committee of
Fudan University.

ACCEPTED MANUSCRIPT
2.2 Synthesis and characterization of glycopeptides and functional materials
2.2.1 Chemical synthesis of building block of N-α-(9-fluorenylmethylcarbonyl)-3-O-(2,3,4,6-tetra-
O-acetyl-β-D-glucopyranosyl)-L-threonine
Scheme
2.
Synthesis
of
N-α-(9-fluorenylmethylcarbonyl)-3-O-(2,3,4,6-tetra-
O-acetyl-β-D-glucopyranosyl)-L-threonine by anomeric acetate activation method.
The lewis acid (BF3·Et2O, 10 mmol) was added to a solution of pentaacetate glucose
(2.3mmol) and Fmoc-threoine (2 mmol) in DCM. The mixture was stirred for 2 h and the solution
was then washed with aqueous HCl (1M), dried and concentrated (Scheme 2). The crude product
was purified by column chromatography and characterized by ESI-MS.
2.2.2. Synthesis and characterization of glycopeptide and fluorescein-labeled peptide
9
Scheme 3. Synthetic route to glycopeptide. The synthesis of G-A7R was taken as an example.
MBHA-functionalized Rink resin was used to provide the C-terminal amides upon cleavage after
classical Fmoc construction of the glycopeptide. Treatment with hydrazine hydrate (H2NNH2•H2O)
in methanol (CH3OH) was required to remove the acetates from the glycoside moiety prior to
cleavage from the Rink resin.
The peptides were synthesized by solid phase peptide synthesis using Rink-amide resin
(substitution = 0.54 mmol/g) according to Fmoc methodology (Scheme 3). All the solid phase
reactions were carried out manually in a sintered glasstube. Peptide condensation was facilitated
by using excess Fmoc amino acid (8.8 equiv.) activated with HBTU and HOBt in DMF. The
coupling reaction was allowed to proceed for 1 hour in DMF. For the glycoside-bearing residues,
1.5 equiv. of the glycosylated amino acid were used due to their expense, and the coupling
reaction was prolonged for 3 hours. The Fmoc deprotection was accomplished by treating the
peptide-bound resin with 20 % (vol/vol) piperidine/DMF for 15 min twice. Upon completion of
the synthesis, the acetate groups of the glycoside were removed with 80 % (vol/vol) hydrazine
hydrate in MeOH for 1 hour twice while on the solid phase. The resin bound peptide was washed
three times with DMF and three times with DCM and methanol and dried under vacuum. The
resin-bound peptides were deprotected and cleaved from the solid support with a cocktail mixture
(9.5 mL of TFA, 0.25 mL of Et3SiH and 0.25 mL of H2O, per 1.0 g of peptide resin) for 2 hour at
room temperature. After cleavage was completed, the resulting solution was concentrated to an oil
in vacuo and cold Et2O was poured over the peptide solutions to precipitate the samples.
Purification of cross-linked compounds was achieved by high performance liquid chromatography
(HPLC; Waters) on a reversed-phase C18 column (Waters) to yield the pure compounds. Chemical
composition of the pure products was confirmed by HPLC and ESI-MS. The analytical HPLC
chromatogram and mass spectral data of glycopeptide provided confirmation of purity and
structural integrity.
In order to investigate the effect of modification site and number of glucose moieties, a series
of threonine-based glycosylated analogues of A7R were designed. The peptides are:
7
(1) A7R: amino acid sequence is A6T5W4L3P2P1R; A N-terminus cysteine could be added
otherwise for conjugation of fluorescein and the peptide was denoted as A7R-Cys.
6
(2) G-A7R: The threonine of A7R was substituted by glycosylated threonine (Glu-Thr for short).
9
(3) G-A7R: Glu-Thr was coupled at the N-terminal of A7R-Cys;
(4) 9,10G-A7R: Two Glu-Thrs were coupled sequentially at the N-terminal of A7R-Cys;

ACCEPTED MANUSCRIPT
Fluorescein-labeled peptide was synthesized through sulfhydryl-maleimide covalently
conjugation. Briefly, 5 mg A7R-Cys was dissolved in phosphate buffer (0.1 M, pH = 7.2), 1 mg
fluorescein-5-maleimide in DMF was added and gently stirred for 2 h at room temperature in the
dark. The near-infrared fluorescent probe Cy7-labeled peptide was synthesized in a similar method.
Fluorescein-labeled Glu-Thr was synthesized similarly to the solid phase peptide synthesis. Briefly,
1.5 equiv. of N-α-(9-fluorenylmethylcarbonyl)-3-O-(2,3,4,6-tetra-O-acetyl-β-D-glucopyranosyl)-
L-threonin was coupled to Fmoc rink-amide resin. After Fmoc deprotection, 1.2 equiv. of FITC
was covalently conjugated in a mixture of pyridine, DMF and DCM (12:7:5) for 1 h. The product
was cleaved from resin after acetate groups deprotection and precipitate in cold n-hexane.
The target product was purified via preparative C18 reverse-phase HPLC. Pure product was
ascertained by HPLC and ESI-MS.
2.2.3 Synthesis and characterization of functional materials
9
The targeting material A7R-PEG3400-DSPE and G-A7R-PEG3400-DSPE were synthesized
through covalent conjugation between thiolated peptide and mal-PEG3400-DSPE.
9
Mal-PEG3400-DSPE dissolved in DMF was added to A7R-Cys or G-A7R dissolved in phosphate
buffer (0.1 M phosphate, pH 7.4) and the reaction was monitored by HPLC.
FITC-PEG3400-DSPE was synthesized through covalent conjugation by stirring FITC and
NH2-PEG3400-DSPE dissolved in DMF for 2 h. A ninhydrin test was adopted to confirm that no
free amino group existed, which indicated that all of the DSPE-PEG3400-NH2 was linked with
FITC.
The excessive peptide or FITC was removed via dialysis (MWCO 3.5 kDa) against distilled
water for 48 h. The pure functional materials were obtained after lyophilization and characterized
by 1H-NMR (Varian 400 MHz, Palo Alto, USA).
2.3 Preparation and characterization of nanodisks
The thin-film hydration method was used to prepare the nanodisks. A mixture of the desired
materials was dissolved in the chloroform, and the organic solvent was removed by rotary
evaporation to form a thin film. The lipid film was dried under vacuum overnight and hydrated in
phosphate buffered saline (10 mM phosphate, 150 mM NaCl, pH 7.4). Nanodisks were prepared
by sonication of the hydrated solution for 45 min in an ice-bath using a JY92-II sonicator (Scientz,
Ningbo, China). The resulting solution was filtered using a 0.22 μm filter to remove metal debris.
The PEG-stabilized nanodisk without modification (Disk) was composed of
POPC/cholesterol/ mPEG2000-DSPE (35:40:25, mol%). The A7R peptide or glycopeptide modified
nanodisks
(A7R-Disk
or
9G-A7R-Disk)
was
composed
mol%)
of
or
POPC/cholesterol/mPEG2000-DSPE/A7R-PEG3400-DSPE
POPC/cholesterol/mPEG2000-DSPE/9G-A7R-PEG3400-DSPE (35:40:23:2,mol%). For FITC-labeled
(35:40:23:2,
nanodisk,
the
components
and
ratios
were
POPC/cholesterol/mPEG2000-DSPE/FITC-PEG3400-DSPE (35:40:23:2, mol%) for Disk-FITC,
POPC/cholesterol/mPEG2000-DSPE/A7R-PEG3400-DSPE/FITC-PEG3400-DSPE
mol%) for A7R-Disk-FITC
(35:40:21:2:2,
and
POPC/cholesterol/mPEG2000-DSPE/9G-A7R-PEG3400-DSPE/FITC-PEG3400-DSPE (35:40:21:2:2,
mol%) for 9G-A7R-Disk-FITC.
For drug or fluorescein loading, PTX, Did or Dir was dissolved in the organic solvent

ACCEPTED MANUSCRIPT
together with the membrane materials, which was then hydrated with PBS and sonicated. The
unloaded PTX, Did or Dir was removed using a 0.1 μm filter membrane.
Cryogenic transmission electron microscope (Cryo-TEM) investigations were performed
using a FEI Tecnai G20 Transmission Electron Microscope (FEI, Hillsboro, USA). The
Cryo-TEM specimens were prepared by the following steps. A small drop of the Disk/PTX,
9
A7R-Disk/PTX or G-Disk/PTX was applied on Quantifoil grids, blotted for 4.5 s in a 100%
humidity chamber, and plugged into liquid ethane (cooled by liquid nitrogen) in an FEI Vitrobot
Mark IV vitrification robot. Samples were kept below -165 °C and protected from atmospheric
conditions during both transfer from the preparation chamber to the microscope and during
examination.
2.4 Serum stability of peptide and nanodisk
All the peptides were dissolved in phosphate buffered saline (PBS) (1 mg/mL). One hundred
microliter of each peptide was incubated with 0.9 mL 25% sterile rat serum. After 0.25, 0.5, 1, 2, 4,
8, 12 and 24 h of incubation at 37°C, 20 μL of trichloroacetic acid was added into 100 μL reaction
mixture. The mixture was stored at 4°C for 20 min, and then centrifuged at 12000 rpm for 10 min.
Twenty microliter supernatant was analyzed by RP-HPLC to monitor and quantify peptide
hydrolysis.
Nanodisks were suspended in 50% mouse serum and measurements were performed in
triplicate at room temperature for 2 days. Size (diameter, nm) of nanodisk were measured at
intervals by dynamic light scattering using a Zen 3600 Zetasizer (Malvern).
2.5 In vitro PTX release of nanodisks
Dialysis bags (MWCO 14 kDa, Millipore) with 200 μL of Disk/PTX, A7R-Disk/PTX or
9G-Disk/PTX were directly immersed into 10 mL of PBS (pH 7.4) plus with 50% mouse serum.
One milliliter aliquots were withdrawn from the solution periodically. The volume of solution was
held constant by adding up to 10 mL after each sampling. The amount of PTX released from
nanodisks was measured by absorbance at 227 nm using HPLC. The cumulative drug release was
calculated from the following relationship: cumulative drug release (%) =(Mt/M0) ×100%, where
Mt is the amount of drug released from lipid disks at time t, and M0 is the amount of drug loaded
into the lipid disks.
2.6 In vitro cellular uptake and subcellular localization
To evaluate the targeting effect in brain microvascular endothelial cells and glioma cells.
bEND.3 cells and U87 cells were incubated with 5 μM fluorescein-labeled peptide solution or
various FITC-labeled nanodisk formulations for 4 h at 37 °C. The cells were stained by DAPI and
visualized using a laser scanning confocal microscope (TCS SP5, Leica, Germany). To further
determine the in vitro targeting ability and stability in serum of the nanodisks, U87 cells were
incubated with FITC-labeled nanodisks (containing 5 μM of FITC) of different formulations
pre-incubated with 50% mouse serum. The fluorescein-positive cells were qualitatively measured
by a flow cytometer (FACS Aria; BD).
9
In order to investigate the cellular uptake mechanism of G-A7R, bEND.3 and U87 cells
were incubated with excess D-glucose (4.5 g/l) or A7R (100 μM) and 5 μM fluorescein-labeled
peptide solution at 37°C or 4 °C for 1 h. Then the cells were stained by DAPI for imaging or
gently suspended in PBS for quantitative analysis by flow cytometry.

ACCEPTED MANUSCRIPT
Intracellular distribution of nanodisks was observed by co-localization as determined by
9
confocal microscopy. bEND.3 cells and U87 cells were treated with A7R-Disk/FITC or G-A7R
Disk/FITC for 4 h at 37 °C. Cells were incubated with 50 nM LysoTracker Red (60 min) and 10
mM DAPI (10 min) to visualize lysosomes and nuclei, respectively. Then, the loading solution
was removed and the cell monolayers were washed three times with ice-cold PBS, and observed
with a laser scanning confocal microscope (TCS SP5, Leica, Germany).
2.7 Transport across BCECs monolayer
The BBB model was established as previously reported.[22] Rat primary brain capillary
endothelial cells were isolated and seeded onto rat tail collagen coated transwell chamber.
Transendothelial electrical resistance (TEER) was detected by an epithelial volt-Ωm
(Millicel-RES, Millipore, USA) to evaluate cell monolayer integrity. Monolayers with TEER over
250 Ω·cm2 were used for further experiments.
The culture medium in each apical chamber was replaced by 50 μM fluorescein-labeled
peptides in glucose-free DMEM with 10% FBS. For competition assay, cells were pre-incubated
for 30 min with excess glucose (4.5 g/l). After 0.5, 1, 1.5, 2, 3 and 4 h incubation at 37 °C,
fluorescence intensity of the solutions collected from lower compartment was detected by a
fluorescence detector (PowerWave XS, Bio-TEK, USA). At the end, the TEER was measured
again to verify the integrity of the BBB model.
2.8 Targeting ability on U87MG tumor spheroids and U87 tumor spheroid/BBB co-culture model
U87 tumor spheroids were cultured by seeding U87 cells onto 2% agarose-coated 48-well
plates at a density of 4 × 103 cells/400 μL per well. After 7 days, the tumor spheroids were
incubated with different nanodisks formulations containing 5 μM FITC in DMEM with 10% FBS
for 4 h. To evaluate the targeting ability in U87 tumor spheroid/BBB co-culture model, tumor
spheroids were moved to the lower chamber of BBB model, and lipid disks containing 50 μM
FITC were added in each apical chamber for 4 h incubation. Tumor spheroids were rinsed with
PBS and fixed by 4% paraformaldehyde for 30 min then imaged by a confocal laser microscopy.
2.9 In vivo brain and glioma targeting
To compare the brain targeting efficiency of different glycopeptide, male ICR mice 4~6
weeks of age were injected with 100 μL Cy7-labeled peptide phosphate buffer via tail vein. One
hour after injection, mice were sacrificed and the brains were dissected for ex vivo fluorescence
imaging using an in vivo imaging system (IVIS Spectrum, Caliper, USA).
9
To evaluate the relative tumor targeting efficiency of G-A7R, the subcutaneous xenograft
tumor models were established by inoculation of 4 × 106 U87MG cells (cells suspended in 100 μl
PBS) into the subcutaneous tissue of the right subaxillary of male Balb/c nude mice. After two
weeks when the tumor volume was about 200 mm3, the mice were injected with 100 μL of
Cy7-labeled peptide phosphate buffer. After 1 h of injection, mice were sacrificed and the tumors
were dissected for ex vivo fluorescence imaging.
9
To further investigate the intracranial glioma targeting efficiency of G-A7R, the U87MG
glioma mouse model was prepared by inoculation of 6 × 105 cells in 5 μl PBS into the right
striatum (1.8 mm lateral, 0.6 mm anterior to the bregma and 3 mm of depth) of male Balb/c nude
mice. After 7 days of implantation, the intracranial U87MG tumor bearing mice were injected with

ACCEPTED MANUSCRIPT
100 μL of Cy7-labeled peptide phosphate buffer. After 1 h post-injection, the fluorescent images
were detected using an in vivo image system, therefore tissues were harvested and imaged.
Glioma targeting ability of nanodisks of different formulations was also investigated in a
similar way, the model mice were respectively administrated with 100 μL of DiR encapsulated
nanodisks of different formulations via tail vein 7 days post-implantation. The fluorescence
images were acquired at predetermined time points and the brains were dissected for ex vivo
imaging 12 h after injection by the IVIS imaging system (IVIS Spectrum, Caliper, USA).
Immunofluorescence assay was performed by injecting FITC-labeled nanodisks into the model
mice 7 days post-implantation. After 4 h, the brains were harvested and frozen in Tissue Tek O.C.T.
compound after dehydration. Then, the brains were cut into 10 μm thick sections and stained with
300 nM DAPI and anti-CD31 antibody. The sections were examined under a confocal laser
scanning microscope.
2.10 Hemodynamics of the nanodisks
The kinetic properties of nanodisks with different formulations were analyzed in ICR mice.
Mice were injected with 100 μL DiD-loaded nanodisks via tail vein. At the time points of 1, 5, 10,
15, 30 min and 1, 2, 4, 8, 12, 24, 48, 72 h post injection, 50 μL blood was sampled from the
retro-orbital sinus. The blood samples were diluted 4 times for fluorescence quantification by a
microplate reader (Power Wave XS, Bio-TEK, USA). The pharmacokinetic parameters of the
nanodisks were analyzed by DAS2.0.
2.11Biodistribution of nanodisks in intracranial U87 xenograft bearing mice
Nude mice bearing intracranial U87 xenograft were injected with 100 μL DiD-loaded lipid
disks via tail vein. Twelve hours after injection, blood and homogenate of main organs and tumor
tissues (peeled off from the brain) were collected for fluorescence analysis by a microplate reader
(Power Wave XS, Bio-TEK, USA).
2.12 Cytotoxicity assay
The in vitro cytotoxicity of PTX-loaded nanodisks was determined using MTT assay. U87
cells were seeded into 96-well culture plates at a density of 4× 103 cells/100 μL per well. After 24
h cultivation at 37 °C, the cells were treated with different concentrations of PTX-loaded
nanodisks formulations. The cytotoxicity was determined at 72 h by MTT assay.
2.13Inhibition of tube formation in vitro
Matrigel basement was added into a 24-well plate and incubated at 37 °C for 30 min. Then
U87 cells or HUVEC resuspended with DMEM medium containing 100 ng/mL of PTX-loaded
nanodisks were seeded onto the pre-coated matrigel basement plate. The plate was incubated
overnight and then observed and photographed under an inverted phase contrast microscope
(DMI4000 B, Leica, Germany). The cells treated with blank DMEM medium were set as the
control group. The amount of tubes was counted within three randomly visions and collected for
statistically analysis.
2.14Anti-tumor effect in intracranial U87 xenograft -bearing nude mice
Nude mice bearing intracranial U87 xenograft were used for the anti-tumor studies. The mice
were randomly divided into five groups (n=8) and treated with saline, Taxol, Disk/PTX,
9
A7R-Disk/PTX and G-A7R-Disk/PTX via tail vein on the day 5, 7, 9, 11 and 13 after
implantation with a total PTX dose of 25 mg/kg. The survival times were recorded. One day after
the final treatment, mice were sacrificed and brains with tumor were obtained for

ACCEPTED MANUSCRIPT
immuno-histochemical analysis. The brains were fixed with 4% paraformaldehyde,
paraffin-embedded and sectioned. TUNEL assay was conducted to detect the apoptotic cells.
CD31/PAS dual staining was conducted to observe the microvessels. TUNEL-positive cells or
CD31-positive cells in three slices were counted.
2.15 Statistical analysis
Data were plotted as mean ± SD unless otherwise indicated. Comparison among the different
groups was measured by the one-way ANOVA with Bonferroni correction. p<0.05was considered
statistically significant.
3. Results
3.1 Characterization of glycopeptide
The threonine-based glycosylated analogues of A7R were synthesized via the Fmoc-protected
solid phase peptide synthesis method. The first step was the synthesis of building block, i.e.
glycosylated threonine residue (Glu-Thr). The crude peptide was purified and analyzed by HPLC
and the molecular mass were ascertained by ESI-MS (Fig. 1 and Fig. S1). The m/z[M+H]+ of
6
9
Glu-Thr, A7R, G-A7R, G-A7R and 9,10G-A7R were respective 672.6 Da, 943.4 Da, 1104.4 Da,
1205.4 Da and 1467.4 Da. The determined molecular mass was within experimental errors of the
expected values, indicating the successful synthesis of
a series of glycopeptides.
Fluorescein-labeled and Cy7-labeled peptides were also characterized by HPLC and ESI-MS (data
not shown).
Figure 1. Analysis and characterization of glycopeptide. HPLC and ESI-MS of A7R-Cys (A) and
9G-A7R (B). The conditions of HPLC analysis were as follows: temperature: 40 °C; C18 column:
YMC, 3.5 µm, 4.6×150 mm; running phase: 5-65% acetonitrile in water with 0.1% TFA; flow rate:
0.7 mL/min.
3.2 Comparison and screening of glycopeptide
A7R was shown to bind specifically to VEGFR2 and NRP-1 overexpressed on glioma.
Amino-acid deletion analysis revealed that four C-terminal residues (LPPR) and particularly the
C-terminal arginine with a free carboxyl group were essential for A7R activity[23]. The effect of
modification site and number of glucose moieties on A7R bioactivity were investigate. A series of
6
threonine-based glycosylated analogues of A7R were designed, including G-A7R with the
glucose moiety at the side chain, 9G-A7R with the glucose moiety at the N-termini and 9,10G-A7R
with the two glucose moieties at the N-termini. C-terminal modification was not conducted.
The stability of glycopeptides was studied by incubating with fresh rat serum and the intact
6
peptide after incubation was determined by RP-HPLC. As shown in Fig. 2A, A7R and G-A7R
9
degraded quickly in serum and almost vanished after 2 h incubation. In contrast, G-A7R and
9,10G-A7R were much more stable in serum than A7R. Fifty percent of glycopeptides remained
intact after 2 h incubation.
To investigate the effect of glycosylation on cellular uptake by glioma cells，fluorescently
labeled peptides were incubated with U87 cells and the uptake by U87 cells were quantitated by
flow cytometry. As shown in Fig. 2B, both A7R (77.7%) and glycosylated threonine residue
9
(Thr-Glu-Flu) (70%) showed high uptake efficiency by U87 cells. G-A7R and 9,10G-A7R
exhibited slightly enhanced or similar uptake efficiency of 82.2% and 75.2% compared with A7R,
respectively. However, 6G-A7R exhibited weaken uptake efficiency of 49.8%.

ACCEPTED MANUSCRIPT
To investigate the GLUT-meditated brain transport, uptake of peptides by bEND.3 cells was
analyzed by flow cytometry (Fig 2C). A7R could barely be taken up by bEND.3 cells, while
9
glycosylation at the N-termini of A7R, such as G-A7R and 9,10G-A7R, effectively facilitated
6
uptake by bEND.3 cells. G-A7R could not target brain capillary endothelial cells because the
glucose residue would hardly reach the deep binding pocket of GLUT1[9]. Cy7 was employed as
a fluorescent dye to further compare the brain distribution between A7R and glycopeptide in vivo.
One hour after injection, mice were sacrificed and the brains were dissected for imaging. The
6
fluorescence intensity of A7R-Cy7 and G-A7R in the brain was negligible, indicating that they
barely crossed the BBB in vivo (Fig. 2D). 9G-A7R and 9,10G-A7R could accumulate effectively in
the brain according to the semi-quantitative analysis of fluorescence intensity (Fig. 2E). However,
9
there was no significant difference between G-A7R and 9,10G-A7R, indicating more glucose
moieties did not significantly improve the brain transport of peptide. These findings were
consistent with the results of in vitro cellular uptake by bEND.3 cells.
9
Based on the aforementioned results, G-A7R and 9,10G-A7R showed advantages over A7R
6
9
and G-A7R. G-A7R was selected as for further studies for its ease of synthesis and high
targeting efficiency.
Figure 2. Performance of glycopeptides. (A) Stability of peptide in 25% rat serum at 37 °C
determined by HPLC. U87 (B) and bEND.3 cells (C) uptake of a series of glycopeptides analyzed
by flow cytometry. (D) Ex vivo near-infrared imaging of harvested brains 1 h post injection. (E)
Semi-quantitative analysis of the mean fluorescence intensity of the dissected brains (mean ± SD,
n = 3). ***p < 0.001, **p < 0.01.
3.3 Brain and glioma targeting ability of 9G-A7R in vitro and in vivo
3.3.1 Brain targeting ability of 9G-A7R on BBB model
The in vitro BCECs monolayer was established to evaluate the brain targeting ability of
9G-A7R peptide. The cell monolayer with qualified transendothelial electrical resistance (TEER >
250 Ω•cm2) was characterized for further experiments. The transport ratios for all groups
increased in a time-dependent manner (Fig. 3A). After 4 h, 1.16 ± 0.36 % of 9G-A7R traversed the
BBB, which was greater than that of A7R (0.61 ± 0.22%). A7R displayed lower efficiency at all
9
tested time points in comparison to G-A7R. For competition assay, excess D-glucose was
pre-incubated with BCECs monolayers for 30 min to saturate GLUT1 at the luminal side,
9
followed by incubation with fluorescein labeled 9G-A7R. Transport of G-A7R was dramatically
undermined with D-glucose pretreatment, confirming that the glucose transporter was involved in
the transport of 9G-A7R across the BBB.
3.3.2 Cellular uptake mechanism of 9G-A7R by bEND.3 cells and U87 cells
Glycopeptide is composed mainly of two parts: glucose moieties and A7R peptide. Fig. S2
illustrated cellular uptake and competition inhibition effect on the uptake by bNED.3 and U87
9
cells. The cellular uptake of G-A7R by bNED.3 cells was energy independent and could be
inhibited by glucose not A7R (Fig. S2A). However, uptake of 9G-A7R by U87 cells showed some
discrepancies. Fig. S2B showed that the cellular uptake by U87 cells was energy independent and
could be inhibited by A7R not glucose. In contrast, glucose might provide energy to promote
uptake of 9G-A7R by U87 cells. The data suggested that glycosylation indeed enabled A7R across
the BBB through the GLUT1.

ACCEPTED MANUSCRIPT
3.3.3 Tumor targeting ability in intracranial and subcutaneous U87 xenograft bearing nude mice
The glioma targeting ability of 9G-A7R with the existence of BBB were tested in intracranial
U87 xenograft bearing nude mice. Peptide was labled with near-infrared dye Cy7 for ex vivo
imaging. As shown in Fig. 3B and C, free A7R rarely distributed in the brain or brain tumor. The
fluorescence was higher in the tumor region after the treatment of 9G-A7R, suggesting of effective
and precise targeting for glioma.
To further verify the relative tumor targeting ability of 9G-A7R without the existence of BBB,
we also established subcutaneous U87 model. One hour after i.v. administration, main organs and
9
tumor tissues were dissected for ex vivo imaging. As shown in Fig. 3D and E, G-A7R displayed
higher accumulation in glioma in comparison to A7R. The average fluorescence intensity of
9G-A7R-Cy7 group was about 4.7 times higher than that of A7R-Cy7 group in tumor tissues.
9
Figure 3. Brain and tumor targeting ability of G-A7R in vitro and in vivo. (A) Transmembrane
9
transport efficiency of A7R and G-A7R in the in vitro BBB monolayer. Ex vivo near-infrared
imaging (B) and semi-quantitative analysis of the mean fluorescence intensity from the
Cy7-labeled peptide (C) of harvested brains 1 h post-injection in intracranial U87 xenograft
bearing nude mice. Ex vivo near-infrared imaging (D) and semi-quantitative analysis of the mean
fluorescence intensity from the Cy7-labeled peptide (E) of disserted organs 1 h post-injection in
subcutaneous U87 xenograft bearing nude mice. (Mean ± SD, n = 3). **p < 0.01, ***p < 0.001.
3.4 Characterization of PTX-loaded lipid-nano disk
9
The functional materials A7R-PEG3400-DSPE and G-A7R-PEG3400-DSPE were synthesized
using a sulfhydryl-maleimide coupling method. The purified materials were characterized by
1
1
HPLC and H-NMR analysis (Fig. S3A and B). In the H-NMR spectrum of mal-PEG3400-DSPE,
the maleimide group exhibited a peak at 6.7 ppm, while the peak completely disappeared in the
spectrum of A7R-PEG3400-DSPE and 9G-A7R-PEG3400-DSPE, indicating a successful synthesis of
the targeting materials.
Cryo-TEM enables direct visualization of the shape and particle size of nanodisks.
Cryo-TEM images revealed disk-shaped structures with a diameter of approximately 50 nm (Fig.
4). A7R or glycopeptide modification did not induce obvious morphological changes. All the
nanodisks exhibited good stabilities in mouse serum (Fig. S3F). All PTX-loaded nanodisk
formulations displayed similar PTX release profiles as shown in Fig. S3G. The zeta potentials of
9
Disk/PTX, A7R-Disk/PTX and G-A7R-Disk/PTX were respective -2.71 ± 1.05 mV, -1.58±
0.41mV and -1.49± 1.16 mV. The drug loading efficiency of Disk/PTX, A7R-Disk/PTX and
9G-A7R-Disk/PTX was respective 8.45 ± 0.058 %, 8.14 ± 0.12 % and 7.84 ± 0.067 % (Table S1).
Figure 4. Cryo-TEM images of Disk/PTX (C), A7R-Disk/PTX (D) and 9G-A7R-Disk/PTX (E).
3.5 Brain and tumor targeting ability of nanodisk in vitro and in vivo
3.5.1 In vitro cellular uptake by brain capillary endothelial cells and U87 cells
Cellular uptake of various FITC-labeled nano-disk formulations by bEND.3 and U87 cells
was conducted to evaluate the targeting property. As shown in Fig. 5A and C, 9G-A7R-Disk could
be significantly taken up by bEND.3 cells in comparison to Disk or A7R-Disk, indicating that
glycopeptide modification could enhance the uptake of disks by brain capillary endothelial cells
through glucose transporter 1 (GLUT1). Moreover, in the cellular uptake by U87 cells, the

ACCEPTED MANUSCRIPT
intracellular fluorescence intensity in 9G-A7R-Disk treated cells was significant higher than that of
Disk or A7R-Disk (Fig. 5B), and it was attributed to the enhanced stability of glycopeptide[24]. As
confirmed in Fig. 5D, the pre-incubation with mouse serum would not obviously impair the
cellular uptake of 9G-A7R-Disk while the uptake of Disk or A7R-Disk was significantly reduced.
Lysosomes were labeled with Lyso-Tracker Red to track intracellular distribution of
9
nanodisks. It showed obvious co-localization of A7R-Disk or G-A7R-Disk after 4 h incubation
with bEND.3 or U87 cells (except that A7R-Disk could not internalized into bEND.3 cells),
suggesting lysosomes were involved in the internalization of the nanodisks (Fig. 5E and F). Thus
the stability of glioma targeting moiety during the BBB transcytosis process became particularly
important and the glycopeptide modified nanodisks would maintain targeting ability in the harsh
lysosomal environment.
Figure 5. Cellular uptake and subcellular localization of different nanodisks by bEND.3 (A, C and
E) and U87 cells (B, D and F). FITC-positive cells were calculated by a flow cytometer, and
intracellular fluorescence was captured by a confocal laser scanning microscope. Scale bars
represent 10 μm.
3.5.2 Targeting ability of nanodisks to U87 tumor spheroids and U87 tumor spheroid/BBB
co-culture model
U87MG tumor spheroids were established to assess the penetration efficiency of nanodisks.
As shown in Fig. 6A, both nanodisks modified with A7R or 9G-A7R showed efficient uptake, but
the fluorescent intensity of tumor spheroid treated with 9G-A7R-Disk was higher than that of plain
Disk or A7R-Disk. In order to better mimic the situation of brain tumor, a BBB/U87 tumor
spheroids co-culture model was established to assess the targeting ability of nanodisks modified
with A7R and 9G-A7R in vitro. As shown in Fig. 6B, A7R-Disk could not cross the BBB thus no
9
tumor spheroid targeting ability, whereas only the tumor spheroid treated with G-A7R-Disk
exhibited obviously uptake.
Figure 6. Uptake of FITC-labeled Disk, A7R-Disk and 9G-A7R-Disk by U87 tumor spheroids (A)
and U87 tumor spheroid/BBB co-culture model (B). Tumor spheroid penetration was examined by
a confocal microscope, with a 5 μm interval between consecutive slides.
3.5.3 In vivo brain glioma targeting effect
To assess the glioma targeting ability in vivo, DiR-loaded nanodisks were intravenously
injected into mice bearing intracranial U87MG tumor. The results of semi-quantitative ROI
9
analysis by the software of living image indicated that G-A7R-Disk led to the highest
accumulation in glioma, which reached peak value at 4 h post-injection (Fig. 7A and B). The
images of excised brains and its semi-quantitative analysis also indicated that nanodisks modified
with 9G-A7R showed the strongest fluorescence accumulation in glioma, which was around 4-fold
higher than that of A7R-Disk (Fig. 7C and D). Frozen brain glioma sections were obtained to
visualize the precise distribution of nanoparticles within the tumor site (Fig. 7E). The
9G-A7R-Disk were mostly distributed around blood vessels and taken up by glioma cells. In
contrast, only weak fluorescence signal from plain Disk or A7R-Disk was observed within tumor.
Figure 7. In vivo targeting ability of nanodisks in intracranial U87 glioma. In vivo fluorescence
imaging of intracranial U87MG tumor (A) and semi-quantitative ROI analysis (B) of in vivo
fluorescent images at different time points after intravenous injection of DiR-loaded nanodisks. Ex

ACCEPTED MANUSCRIPT
vivo near-infrared imaging of the harvested brains (C) and semi-quantitative ROI analysis of the
mean fluorescence intensity in tumors (D) from the DiR-loaded nanodisks 12 h post-injection.
(Mean ± SD, n =3). (E) The distribution of nanodisks in the glioma of tumor-bearing mice 7 days
post-implantation. Frozen sections were examined under a fluorescence microscope. Nuclei were
stained with DAPI (blue), blood vessels were labeled with anti-CD31 (red), while green represents
the FITC-labeled nanodisks (scale bar=50 μm).
3.5.4 Blood dynamics and biodistribution of nanodisks
To investigate the systemic circulation time of nanodisks, DiD-loaded disks were
administrated into each mouse via tail vein. Blood was sampled at different time points for
fluorescence measurements. Relative plasma concentration-time curves of nanodisks after i.v.
injection were presented in Fig. S4A and pharmacokinetic parameters were obtained by
noncompartmental analysis. The mean residence time (MRT) of Disk was as high as 21.76 hours,
9
however, modification with A7R or G-A7R reduced the MRT of Disk to some extent. The
biodistribution of nanodisks in intracranial U87 xenograft bearing mice was further studied.
DiD-loaded disks were administrated into each mouse via tail vein and blood and homogenate of
main organs were collected for fluorescence analysis 12 h post injection. As shown in Fig. S4B, all
the nanodisks had very high retention in blood compared to other organs. Modification with
glycopeptide of disk inevitably increased the uptake by the endothelial reticular system (RES),
which explained the reduced MRT of 9G-A7R-Disk. Meanwhile, the accumulation in tumor tissue
was significantly enhanced using glycopeptide to cross the BBB. The results indicated that
modification with glycopeptide could enable nanodisks traverse the BBB and significantly
elevated glioma accumulation.
3.6 In vitro and in vivo anti-tumor effect
3.6.1 Cytotoxicity of PTX-loaded nanodisks in vitro
In vitro cytotoxicity of PTX-loaded nanodisks were investigated in U87 by the MTT assay
(Fig. 8A). Inhibition of cell growth was studied after 72 h. The IC50 value of Taxol, Disk/PTX and
A7R-Disk/PTX were 43.66, 74.90 and 38.85 nM, respectively, while the IC50 value of
9G-A7R-Disk/PTX was 26.36 nM. The results indicated that the inhibitory effect to the
proliferation of U87 cells was markedly elevated after the incorporation of glycopeptide.
3.6.2 In vitro destroying ability against U87MG VM channels and HUVEC tubes
U87 vasculogenic mimicry (VM) and HUVEC tubes were used to evaluate the inhibitive
effects of PTX-loaded nanodisks (Fig. 8B-D). Cells treated with DMEM formed extensive and
enclosed tube networks. Networks of cells subjected to various PTX formulations exhibited a
9
different degree of fracture. G-A7R-Disk exhibited higher inhibitive effects in both U87 VM
tubes and HUVEC tubes than did A7R-Disk and plain Disk. These results were consistent with the
results of cytotoxicity, indicating the cytotoxic effect of nanodisks was significantly enhanced by
the glycopeptide modification.
Figure 8. Anti-tumor effect of PTX-loaded nanodisks in vitro. (A) Cytotoxicity of various
PTX-loaded nanodisks against U87 cells was assessed using MTT assay after 72 h incubation
(mean ± SD, n =3). The IC50 values were calculated using GraphPad Prism 7.0. (B) Effect of
various PTX-loaded nanodisks on U87 VM formation and HUVEC tube formation. The
quantitative data of the U87 VM formation (C) and HUVEC tube formation (D) treated by

ACCEPTED MANUSCRIPT
different PTX-loaded nanodisks. Cells incubated with drug-free DMEM served as the control.
Data are presented as the percentages of the control group, which was set at 100%. *p<0.05,
**p<0.01, ***p<0.001. Bar = 200 μm.
3.6.3 Anti-tumor efficacy of in vivo
To evaluate the potential therapeutic efficacy of PTX-loaded nanodisks for the treatment of
brain tumors, intracranial U87 xenograft bearing nude mice were used and the survival results
were represented in a Kaplan-Meier plot in Fig. 9A. The median survival time of the mice treated
9
with G-A7R-Disk/PTX (33.5 days) was significantly longer than that of those treated with
A7R-Disk/PTX (28 days), Disk/PTX (28.5 days) and Taxol (26.5 days), thus increasing the
anti-glioma effect by 2.83, 2.43 and 5.67-fold, respectively. TUNEL and CD31/PAS
immunohistochemical staining were conducted (Fig. 9B-D). 9G-A7R-Disk/PTX treatment induced
significant apoptosis and remarkably inhibited angiogenesis in tumor site in comparison to other
treatments. These results indicated that nanodisks modified with glycopeptide exhibited a
significant improvement in anti-tumor activities, an effect attributed to its ability across the BBB.
Figure 9. Therapeutic efficacy of PTX-loaded nanodisks in vivo. (A) Kaplan–Meier survival
curves of nude mice bearing intracranial U87 glioma. Mice (n =8) were injected at 5, 7, 9, 11 and
13 days after glioma implantation with saline, Taxol, Disk/PTX, A7R-Disk/PTX and
9G-A7R-Disk/PTX. (B) CD31/PAS dual staining analysis and TUNEL immunohistochemical
staining of tumor sections after various treatments. Quantification of the percentage of TUNEL
positive apoptotic cells (C) and angiogenesis inhibition (D) in tumor treated with different disk
formulations in comparison to saline group. *p<0.05, **p<0.01. Scale bar =100 μm.
4. Discussion
Heptapeptide ATWLPPR (A7R) binds specifically to VEGFR2 and NRP-1 which holds
promise in tumor imaging[25.26] and active drug delivery[27-29]. Our previous studies have
demonstrated that A7R-modified liposomes exhibited excellent glioma-targeting ability in
subcutaneous tumor model[28.29]. Considering the BBB and enzymatic degradation of peptide in
treatments of glioma in vivo, strategy of glycosylation was applied to make A7R versatile.
Amino-acid deletion analysis revealed that four C-terminal residues (LPPR) and particularly
the C-terminal arginine with a free carboxyl group were essential for A7R activity[23]. Therefore
the effect of modification site and number of glucose moieties was investigated. We compared the
serum stability, the intrinsic targeting activity of glioma cells and uptake ability of brain
endothelial cells among the analogues of A7R. For serum stability, our results showed that A7R
6
and G-A7R (imitating natural O-linked glycopeptide with the glucose moieties on the side chain
of A7R) were degraded quickly in rat serum. By contrast, N-terminal glycopeptide (9G-A7R and
9
9,10G-A7R) exhibited enhanced stability (Fig. 2A). For the uptake of U87 cells, G-A7R and
9,10G-A7R had high uptake efficiency as A7R while 6G-A7R showed some decrease (Fig. 2B). We
speculate that the glucose moieties in 6G-A7R were so close to the C-terminal residuals to impede
the binding of A7R to the receptor. For the comparison of brain targeting ability, the N-terminal
6
glycopeptide (9G-A7R and 9,10G-A7R) showed a similar superior effect to A7R and G-A7R in
vitro and in vivo (Fig. 2C and D). Based on the experimental results above, 9G-A7R was selected
as the optimal one for further study considering its simplicity and validity.
It is noteworthy that tumor cells deviate from most normal tissues and exhibit a dysfunctional
and ravenous consumption of glucose, known as Warburg effect[30]. GLUT1 is also found to be

ACCEPTED MANUSCRIPT
overexpressed on brain tumor cells, thus can be utilized for tumor targeting[11.31]. However, the
9
glucose ligand did not increase the uptake of G-A7R by U87 cells through GLUT1 (Fig. S3B),
the increased accumulation in tumor tissues might result from enhanced serum stability. For brain
9
9
glioma targeting, G-A7R was proved to exhibit a two-stage targeting process. Firstly G-A7R
tranversed the BBB through GLUT-mediated transport and afterwards targeted the brain glioma
through receptor-mediated transcytosis.
A7R has been reported to decrease the tumor angiogenesis and growth at a relative high
9
dosage[32]. However, in our studies, neither high doses of A7R nor G-A7R could significantly
prolong the median survival time of glioma-bearing mice (Fig. S5). In contrast, the advantage of
9G-A7R is in line with its glioma targeting ability rather than therapeutic effect. Thus, it is of great
potential to use 9G-A7R as the targeting ligand in drug delivery system.
In order to enable A7R to cross the BBB, we performed glycosylation strategy, which would
inevitably increase the accumulation in reticuloendothelial system (RES) due to the introduction
of glucose molecule. However, biodistribution results in subcutaneous tumor mice suggest that
glycopeptide molecule has the most significant targeting effect at the tumor site, 4.7 fold more
potent than A7R. The relative targeting in the tumor site was much higher than that in other organs.
Modification of glycopeptides on the surface of nano-lipid nanodisks also significantly increased
the accumulation in brain tumor sites. It is noteworthy that, compared with other organs, the
nanodisk exhibited very high content in the blood, suggesting a long blood circulation capacity.
The hydrophobic core of the lipid disk provided a reservoir for the insoluble small molecule drug.
The PTX-loaded nanodisk has a good tumor killing effect and can significantly prolong the
survival of glioma-bearing mice.
5. Conclusions
9
In summary, a functional glycopeptide G-A7R was developed to overcome enzymatic and
9
BBB barriers by employing glycosylation strategy. G-A7R was resistant to serum enzymes and
could penetrate the BBB through GLUT-mediated transcytosis and afterwards target the brain
glioma through receptor-mediated endocytosis. A novel glioma-targeted drug delivery system was
constructed successfully employing 9G-A7R as the targeting moiety and nanodisk as the carrier of
paclitaxel. The delivery system could precisely target glioma in nude mouse xenograft models and
exhibit superior anti-glioma efficacy. The glycopetide modified nanodisk showed a promising
application for brain glioma therapy.
Acknowledgements
This work was supported by National Natural Science Foundation of China (81773657,
81690263
and
81473149),
Shanghai
Education
Commission
Major
Project
(2017-01-07-00-07-E00052) and Shanghai International Science and Technology Cooperation
Project (16430723800)

ACCEPTED MANUSCRIPT
References:
[1] K. Anjum, B.I. Shagufta, S.Q. Abbas, S. Patel, I. Khan, S.A.A. Shah, N. Akhter, S.S.U.
Hassan, Current status and future therapeutic perspectives of glioblastoma multiforme (GBM)
therapy: A review, Biomed. Pharmacother. 92(2017) 681-689.
[2] M. Reni, E. Mazza, S. Zanon, G. Gatta, C.J. Vecht, Central nervous system gliomas, Critical
Reviews in Oncology/Hematology 113(2017) 213-234.
[3] K.A. Ghotme, G.E. Barreto, V. Echeverria, J. Gonzalez, R.H. Bustos, M. Sanchez, J. Leszek,
N.S. Yarla, R.M. Gomez, V.V. Tarasov, G.M. Ashraf, G. Aliev, Gliomas: New Perspectives
in Diagnosis, Treatment and Prognosis, Curr. Top. Med. Chem. 17(2017) 1438-1447.
[4] M.V. Wanjale, G. Kumar, Peptides as a therapeutic avenue for nanocarrier-aided targeting of
glioma, Expert Opin. Drug Deliv. 14(2017) 811-824.
[5] A.S. Wadajkar, J.G. Dancy, D.S. Hersh, P. Anastasiadis, N.L. Tran, G.F. Woodworth, J.A.
Winkles, A.J. Kim, Tumor-targeted nanotherapeutics: overcoming treatment barriers for
glioblastoma, Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 9(2017).
[6] L. Lu, H. Qi, J. Zhu, W.X. Sun, B. Zhang, C.Y. Tang, Q. Cheng, Vascular-homing peptides
for cancer therapy, Biomed. Pharmacother. 92(2017) 187-195.
[7] A. David, Peptide ligand-modified nanomedicines for targeting cells at the tumor
microenvironment, Adv. Drug Deliv. Rev. 119(2017) 120-142.
[8] M. Ying, Q. Shen, C. Zhan, X. Wei, J. Gao, C. Xie, B. Yao, W. Lu, A stabilized peptide
ligand for multifunctional glioma targeted drug delivery, J. Control. Release 243(2016)
86-98.
[9] D. Deng, C. Xu, P. Sun, J. Wu, C. Yan, M. Hu, N. Yan, Crystal structure of the human
glucose transporter GLUT1, Nature 510(2014) 121-125.
[10] X. Fu, G. Zhang, R. Liu, J. Wei, D. Zhang-Negrerie, X. Jian, Q. Gao, Mechanistic Study of
Human Glucose Transport Mediated by GLUT1, J. Chem. Inf. Model. 56(2016) 517-526.
[11] X. Jiang, H. Xin, Q. Ren, J. Gu, L. Zhu, F. Du, C. Feng, Y. Xie, X. Sha, X. Fang,
Nanoparticles
of
2-deoxy-d-glucose
functionalized
poly(ethylene
glycol)-co-poly(trimethylene carbonate) for dual-targeted drug delivery in glioma treatment,
Biomaterials 35(2014) 518-529.
[12] C. Zhang, W. Zhao, L. Liu, R. Ju, L. Mu, Y. Zhao, F. Zeng, H. Xie, Y. Yan, W. Lu, A
nanostructure of functional targeting epirubicin liposomes dually modified with aminophenyl
glucose and cyclic pentapeptide used for brain glioblastoma treatment, Oncotarget 6(2015)
32681.
[13] I. Harbi, B. Aljaeid, K.M. El-Say, A.S. Zidan, Glycosylated Sertraline-Loaded Liposomes for
Brain Targeting: QbD Study of Formulation Variabilities and Brain Transport, AAPS
Pharmscitech 17(2016) 1404-1420.
[14] X. Wei, C. Zhan, Q. Shen, W. Fu, C. Xie, J. Gao, C. Peng, P. Zheng, W. Lu, AD
-Peptide Ligand of Nicotine Acetylcholine Receptors for Brain-Targeted Drug Delivery,
Angew. Chem. Int. Ed. 54(2015) 3023-3027.
[15] B.D. Green, M.H. Mooney, V.A. Gault, N. Irwin, C.J. Bailey, P. Harriott, B. Greer, F.
O'Harte, P.R. Flatt, N-terminal His(7)-modification of glucagon-like peptide-1(7-36) amide
generates dipeptidyl peptidase IV-stable analogues with potent antihyperglycaemic activity, J.
Endocrinol. 180(2004) 379-388.
[16] N.O. Elmagbari, Antinociceptive Structure-Activity Studies with Enkephalin-Based Opioid

ACCEPTED MANUSCRIPT
Glycopeptides, J. Pharmacol. Exp. Ther. 311(2004) 290-297.
[17] B.D. Green, V.A. Gault, F.P. O'Harte, P.R. Flatt, A comparison of the cellular and biological
properties of DPP-IV-resistant N-glucitol analogues of glucagon-like peptide-1 and
glucose-dependent insulinotropic polypeptide, Diabetes Obes. Metab. 7(2005) 595-604.
[18] T. Ueda, K. Tomita, Y. Notsu, T. Ito, M. Fumoto, T. Takakura, H. Nagatome, A. Takimoto, S.
Mihara, H. Togame, K. Kawamoto, T. Iwasaki, K. Asakura, T. Oshima, K. Hanasaki, S.
Nishimura, H. Kondo, Chemoenzymatic synthesis of glycosylated glucagon-like peptide 1:
effect of glycosylation on proteolytic resistance and in vivo blood glucose-lowering activity,
J. Am. Chem. Soc. 131(2009) 6237-6245.
[19] J. Gao, C. Xie, M. Zhang, X. Wei, Z. Yan, Y. Ren, M. Ying, W. Lu, RGD-modified lipid
disks as drug carriers for tumor targeted drug delivery, Nanoscale 8(2016) 7209-7216.
[20] E. Johansson, C. Engvall, M. Arfvidsson, P. Lundahl, K. Edwards, Development and initial
evaluation of PEG-stabilized bilayer disks as novel model membranes, Biophys. Chem.
113(2005) 183-192.
[21] W. Zhang, J. Sun, Y. Liu, M. Tao, X. Ai, X. Su, C. Cai, Y. Tang, Z. Feng, X. Yan, G. Chen,
Z. He, PEG-Stabilized Bilayer Nanodisks As Carriers for Doxorubicin Delivery, Mol.
Pharmaceut. 11(2014) 3279-3290.
[22]
P.
Zhang,
L.
Hu,
Y.
Wang,
J.
Wang,
L.
Feng,
Y.
Li,
Poly(epsilon-caprolactone)-block-poly(ethyl ethylene phosphate) micelles for brain-targeting
drug delivery: in vitro and in vivo valuation, Pharm. Res. 27(2010) 2657-2669.
[23] P.L.R.V. Anna Starzec, C.H.D.P. Alda Navaza, Structure-function analysis of the
antiangiogenic ATWLPPR peptide inhibiting VEGF165 binding to neuropilin-1 and
molecular dynamics simulations of the ATWLPPR/neuropilin-1 complex, Peptides 28(2007)
2397-2402.
[24] M. Ying, Q. Shen, Y. Liu, Z. Yan, X. Wei, C. Zhan, J. Gao, C. Xie, B. Yao, W. Lu,
Stabilized Heptapeptide A7R for Enhanced Multifunctional Liposome-Based
Tumor-Targeted Drug Delivery, ACS Appl. Mater. Inter. 8(2016) 13232-13241.
[25] L. Mey, M. Hormann, N. Schleicher, P. Reuter, S. Donges, R. Kinscherf, M. Gassmann, T.
Gerriets, N. Al-Fakhri, Neuropilin-1 modulates vascular endothelial growth factor-induced
poly(ADP-ribose)-polymerase leading to reduced cerebrovascular apoptosis, Neurobiol. Dis.
59(2013) 111-125.
[26] H. Wu, H. Chen, D. Pan, Y. Ma, S. Liang, Y. Wan, Y. Fang, Imaging integrin alphavbeta 3
and NRP-1 positive gliomas with a novel fluorine-18 labeled RGD-ATWLPPR heterodimeric
peptide probe, Mol. Imaging Biol. 16(2014) 781-792.
[27] H. Slimani, E. Guenin, D. Briane, R. Coudert, N. Charnaux, A. Starzec, R. Vassy, M.
Lecouvey, Y.G. Perret, A. Cao, Lipopeptide-based liposomes for DNA delivery into cells
expressing neuropilin-1, J. Drug Target. 14(2006) 694-706.
[28] M. Ying, Q. Shen, Y. Liu, Z. Yan, X. Wei, C. Zhan, J. Gao, C. Xie, B. Yao, W. Lu,
Stabilized Heptapeptide A7R for Enhanced Multifunctional Liposome-Based
Tumor-Targeted Drug Delivery, ACS Appl. Mater. Inter. 8(2016) 13232-13241.
[29] M. Ying, C. Zhan, S. Wang, B. Yao, X. Hu, X. Song, M. Zhang, X. Wei, Y. Xiong, W. Lu,
Liposome-Based Systemic Glioma-Targeted Drug Delivery Enabled by All-d
Peptides, ACS Appl. Mater. Inter. 8(2016) 29977-29985.
[30] O. WARBURG, On the origin of cancer cells, Science 123(1956) 309-314.

ACCEPTED MANUSCRIPT
[31] Q.L. He, I. Minn, Q. Wang, P. Xu, S.A. Head, E. Datan, B. Yu, M.G. Pomper, J.O. Liu,
Targeted Delivery and Sustained Antitumor Activity of Triptolide through Glucose
Conjugation, Angew. Chem. Int. Ed. Engl. 55(2016) 12035-12039.
[32] A. Starzec, R. Vassy, A. Martin, M. Lecouvey, M. Di Benedetto, M. Crepin, G.Y. Perret,
Antiangiogenic and antitumor activities of peptide inhibiting the vascular endothelial growth
factor binding to neuropilin-1, Life Sci. 79(2006) 2370-2381.

Graphics Abstract

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9