J Neurooncol (2017) 135:37–46

DOI 10.1007/s11060-017-2566-x

LABORATORY INVESTIGATION

CD133 positive U87 glioblastoma cells-derived exosomal

microRNAs in hypoxia- versus normoxia-microenviroment
Guobin Zhang1 · Yunsheng Zhang2 · Sen Cheng2 · Zhen Wu1 · Fusheng Liu2 ·
Junting Zhang1

Received: 1 February 2017 / Accepted: 13 July 2017 / Published online: 25 September 2017
© Springer Science+Business Media, LLC 2017

Abstract Hypoxia is a major regulator of glioma devel- Moreover, we analyze the microRNA content of exosomes
opment and aggressiveness. However, how CD133 positive produced in vitro by hypoxia and normoxia CD133+ U87
U87 glioblastoma cells adapt to hypoxia and communicate glioblastoma cells using Affymetrix microarrays. It appears
with their surrounding microenvironment during tumor that the exosome microRNA profiles are qualitatively quite
development remain important questions. Communica- similar. Nevertheless, their quantitative profiles are differ-
tion with host cells and stroma via exosomes represents ent and may be potentially taken as an opportunity to carry
one pathway by which tumors can modify their surround- out assays of diagnostic interest. We conclude that CD133+
ings to achieve a tumor-permissive environment. MicroR- U87 glioblastoma cells derived exosome-mediated miRNA
NAs are thought to be essential actors of tumorigenesis as transduction play an important role of mediating a proan-
they are able to control the expression of numerous genes. giogenic response and glioma cells proliferation, and that
Here, we show that exosomes derived from CD133+ U87 the exosomal pathway constitutes a potentially targetable
glioblastoma cells grown at hypoxic compared with nor- driver of hypoxia-dependent intercellular signaling during
moxic conditions are potent proliferation inducers of the tumor development.
tumor vasculature and glioma cells proliferation in vitro.
Keywords Exosome · Glioblastoma · Hypoxia ·
Normoxia · MicroRNA
Guobin Zhang and Yunsheng Zhang are the Joint first authors
and contributed equally to the work.

Electronic supplementary material The online version of this Introduction

article (doi:10.1007/s11060-017-2566-x) contains supplementary
material, which is available to authorized users.
Glioblastoma (GBM) is the most common and devastating
* Fusheng Liu primary brain tumor in adults, with a very aggressive course
liufushengs@hotmail.com and few therapeutic options. Much recent researches have
* Junting Zhang aimed to enhance our understanding of the mechanisms
zjt_ttyy@sina.com that drive GBM progression. Hypoxia is one of the diag-
1
nostic hallmarks of GBM. Since glioma stem cells (GSCs)
Department of Neurosurgery, China National Clinical
are enriched in hypoxic niches, it may be hypothesized that
Research Center for Neurological Diseases (NCRC‑ND),
Center of Brain Tumor, Beijing Institute for Brain they play an important role in mediating the hypoxia-driven
Disorders, Beijing Key Laboratory of Brain Tumor, effects on the tumor vasculature and thus contribute to the
Beijing Tiantan Hospital, Capital Medical University, generation of the perivascular niche [1, 2].
Tiantan Xili 6, Chongwen District, Beijing 100050,
The GSCs-niches interactions are in part mediated by the
People’s Republic of China
2
release of soluble factors such as growth factors, there is
Department of Neurosurgery, Brain Tumor Research Center,
growing interest in the role of small membrane-bound vesi-
Beijing Neurosurgical Institute, Beijing Tiantan Hospital,
Capital Medical University, Tiantan Xili 6, Chongwen cles in mediating such interactions. Recent studies point at
District, Beijing 100050, People’s Republic of China a hitherto unknown role of exosomes as important signaling

13
Vol.:(0123456789)
38 J Neurooncol (2017) 135:37–46

entities in the cross-talk between various cell types [3–6]. (EGF, Invitrogen, USA), and 20 ng/mL recombinant human
Interestingly, glioma-derived exosomes that contain angio- fibroblast growth factor-basic (FGF-b, Peprotech, USA).
genic proteins capable of stimulating endothelial cell growth CD133+ U87 glioblastoma cells at passages 2–6 were used
and proliferation in vitro [4]. Moreover, they also carry for experiments. Routine culture was done in a humidified
mRNA, including that derived from mutant oncogenes, as incubator maintained at 37 °C with 5% C­ O2 and 95% air. For
well as microRNA and even DNA, suggesting the scope and hypoxia experiments, the cells were cultured under 37 °C
mechanisms of their actions on recipient cells may extend with 5% C ­ O2 and 1% O
­ 2 (hypoxic) conditions, balanced with
beyond the promotion of local angiogenesis [4]. MicroR- ­N2 in a 3-gas incubator (Thermo, Waltham, MA, USA).
NAs (miRNAs) are small noncoding RNAs that are able to
regulate various physiological and pathological processes, CD133+ U87 glioblastoma cells identification,
including development, differentiation, invasion, prolifera- differentiation and immunofluorescent staining
tion and angiogenesis. They play a crucial role in controlling
the expression of many cellular proteins and therefore are CD133+ U87 glioblastoma cells spheres were cultivated in
involved in several oncogenic pathways [7]. Here, we have DMEM supplemented with 10% FBS for 7 days to induce
investigated the potential role of CD133+ U87 glioblas- neural lineage cell differentiation. The samples were incu-
toma cells-derived exosomal microRNAs in hypoxia- and bated with antibodies against CD133 (1:200, AC133, Milte-
normoxia-microenviroment. nyi, Bergisch Gladbach, Germany), glial fibrillary acidic
protein (GFAP) (1:100; Sigma, USA), neurofilament (NF)
(1:100; Boaosen, China) and Oligodendrocyte-myelin glyco-
Materials and methods protein (Omgp) (1:100; Boaosen, China) overnight at 4 °C.
Following washing, they were then incubated with second-
Culture of U87 glioblastoma cell line and human brain ary antibodies for an additional 1 h at room temperature. The
microvascular endothelial Cells (HBMECs) cell nuclei were stained with 4′,6-diamidino-2-phenylindole
(DAPI). The fluorescent signals were observed and photo-
U87 glioblastoma cells were purchased from the Chinese graphed under a fluorescence microscope (Leica, Solms,
Academy of Sciences Cell Bank, and cultured in DMEM Germany).
containing 10% fetal bovine serum (FBS) and 100 U/mL
penicillin–streptomycin at 37 °C in an atmosphere of 5% Exosome isolation
­CO2/95% air.
HBMECs (Sciencell, USA) were seeded into culture Exosomes were isolated by ultracentrifugation (UC)/sequen-
flasks coated with 1% gelatin (Amresco, USA) in M199 tial centrifugation. To this end, CD133+ U87 glioblastoma
medium supplemented with 20% FBS (Gibco, USA), 50 μg/ cells were cultured in neural basal medium at normoxic or
mL endothelial cell growth supplement (ECGS) (Macgene, hypoxic conditions. Conditioned media were collected from
China), and 100 μg/mL heparin (Macgene, China). 48 h cell cultures and centrifuged at 300g, 4 °C for 5 min to
eliminate cell debris. Supernatant fractions were further cen-
Cell sorting and culture of CD133+ U87 glioblastoma trifuged at 16,500g, 4 °C for 30 min. Afterwards, they were
cells filtered on a 0.2-μm filter, and finally ultracentrifuged for
2 h at 150,000g at 4 °C in a Beckman 70Ti rotor. Exosomes
CD133 + and CD133− U87 tumor cells were sorted by pre- were resuspended in PBS for functional assays or used for
paring single-cell suspensions, which were then incubated RNA isolation.
for 10 min at 4 °C with a CD133 antibody (AC133, Milte-
nyi, Bergisch Gladbach, Germany), or mouse IgG2b mono- Transmission electron microscopy
clonal antibodies as a control. Cell sorting was performed
using MoFlo™ XDP (Beckman Coulter, California, USA). Exosomes were characterised by transmission electron
The purity of the cells after sorting was detected using flow microscopy as previously described [8]. Briefly, isolated
cytometry. In addition, 7-aminoactinomycin D was added exosomes were washed in TBS twice, pelleted by ultra-
(1:10) to discriminate between viable and dead cells. Data centrifugation at 150,000g, 4 °C for 2 h, and fixed in 2%
acquisition and analysis were performed using a FACSCali- paraformaldehyde. Exosomes were adsorbed onto 200 mesh
bur (Becton–Dickinson). Sorted CD133+ cell populations Formvar-coated grids and stained with 0.75% uranyl formate
were cultured in neuralbasal medium composed of neu- (wt/vol). Samples were observed in a FEI Tecnai G2 Spirit
robasal (Invitrogen, USA) supplemented with 2% glutamate transmission electron microscope (North America Nano-
(Invitrogen, USA), 2% B27 supplement without Vitamin A Port) operated at 60 kV accelerating voltage. Images were
(Gibco, USA), 20 ng/mL human epidermal growth factor recorded with an Olympus SIS Veleta CCD camera.

13
J Neurooncol (2017) 135:37–46 39

Co‑cultures and cell viability assay Statistical analysis

Purified CD133+ U87 glioblastoma cells derived exosomes All statistical analyses were performed using SPSS 19.0 sta-
in hypoxia or normoxia condition were added to U87 or tistical software. All experiments were performed in trip-
HBMEC cultures at ~50% confluence at a ratio of 10 licate unless otherwise noted, and results were expressed
exosomes/U87 or HBMEC. U87 and HBMEC were co-cul- as mean ± standard deviation (SD). Comparisons between
tured with exosomes for 48 h and then U87 or HBMEC via- groups were made using Student’s t tests, and comparisons
bility was determined using a cell counting kit-8 (CCK-8). among multiple groups were made using one-way ANOVA.
All data are presented as the mean ± SD from three experi- A probability value of <0.05 in a Student’s t test was consid-
ments, and blank controls were also established. Every 12 h, ered statistically significant. The false discover rate (FDR)
10 μL CCK-8 (Dojindo, Japan) was added and incubated for was used to check the GeneChip microRNA Affymetrix
2 h. Plates containing were shaken for 10 min at 37 °C prior Microarrays correction.
to the measurement of OD values at 450 nm. The OD value
in each group was standardized by subtracting the average
value of the blank controls, and the percentage of viable Results
cells was calculated using standardized values divided by
the average value of the normal group. Isolation and characterization of CD133+ U87
glioblastoma cells
RNA samples
Several studies have shown that brain tumors originate
Before RNA purification, exosomes collected in the ultra- from CD133+ glioblastoma cells that express pluripotent
centrifugation pellets were resuspended in sterile 1X PBS stemness and multilineage differentiation potency [9–14].
and treated with RNaseA (Sigma, 0.1 μg/μL) for 15 min at We sorted CD133+ tumor cells from U87 glioblastoma cells
37 °C in order to remove extravesicular RNAs. RNase inhibi- using fluorescence-activated cell sorting. Approximately
tor (Applied Biosystem, 1 U/μL) was then added and the 0.16% of U87 cells were CD133+, and the purity of the
samples were resuspended in the extraction Lysis buffer of separated CD133+ cells was >80% (Fig. 1a). Immunofluo-
the mirVana miRNA isolation kit (Ambion). RNAs were rescent staining revealed the presence of stemness mark-
purified either from exosomes using the mirVana miRNA ers such as CD133 (Fig. 1b). CD133 + cells also exhibited
isolation protocol (Ambion). Low-molecular-weight (LMW) multilineage differentiation potential, demonstrated by high
RNAs (<200 bases) were purified and separated from high- expression levels of GFAP, NF and Ompg (Fig. 1c).
molecular-weight (HMW) RNAs (>200 bases). RNA quality
was checked by electrophoresis with the Bioanalyzer 2100 Isolation and characterization of CD133+ U87
(Agilent Technologies). glioblastoma cells derived exosomes

Labeling and hybridization on GeneChip microRNA Exosomes purified from the media of the CD133+ U87
Affymetrix microarrays glioblastoma cells in hypoxia and normoxia conditions
were shown in Fig. 2, as visualized by transmission elec-
200 ng of small RNAs were labeled using the FlashTag tron microscopy (TEM). TEM revealed a relatively uniform
biotin-HSR RNA labeling kit (Genisphere). First, poly (A) population of membrane-bound vesicles ranging between 25
tailing was carried out at 37 °C for 15 min in a volume of and 65 nm in diameter, with an average diameter of 43 nm.
15 μL of reaction mix containing reaction buffer, MnCl2, Most exosomes exhibited a “cup” shaped morphology typi-
ATP, and poly (A) polymerase. Second, Genisphere biotin cal of exosomes under TEM. The count and diameter of
complex was ligated at room temperature for 30 min by add- CD133+ U87 glioblastoma cells derived exosomes between
ing FlashTag Ligation Mix Biotin and T4 DNA Ligase into hypoxia and normoxia conditions were similar.
the 15-μl reaction mix. Stop Solution was then added to stop
the reaction. Characterization of microRNA content of hypoxia
Subsequently, microRNA cocktails were hybridized and and normoxia CD133+ U87 glioblastoma cells exosomes
analyzed on microRNAs microarrays version 2 or version 3
(Affymetrix). Labeled RNAs were hybridized on GeneChip To characterize and compare the RNA content of derived
microarrays, washed, stained, and then scanned using the exosomes, we analyzed RNA samples purified both from
miRNA-2.0 library for microRNA microarrays version 2 and CD133+ U87 glioblastoma cells exosomes in hypoxia and
the miRNA-3.0 library for microRNA microarrays version normoxia conditions using Affymetrix microarrays. Small
3, according to Affymetrix’s specifications. RNA fractions were separated and purified following the

13
40 J Neurooncol (2017) 135:37–46

Fig. 1  Isolation and characterization of GSCs. a Isolation of stemness markers such as CD133 (magnification, ×200). c After cul-
CD133+ U87 glioblastoma cells using FACS. Approximately 0.16% ture in medium containing 10% FBS for 7 days, CD133+ cells exhib-
of U87 cells were CD133+. The purity of sorted CD133+ cells was ited multilineage differentiation potential, represented by high expres-
82.8%. b CD133+ cells grew into tumor spheres in neural basal sion levels of GFAP, and NF and Ompg. Nuclei were counterstained
medium. Immunofluorescent staining revealed the presence of several with DAPI (magnification, ×200)

mirVana protocol (Ambion) from exosome samples. To hypoxia and normoxia CD133+ U87 glioblastoma cells
assess the microRNA populations triplicate samples were exosomes miRNA profiles indicated that these two condi-
then independently analyzed on miRNA2.0 microarrays by tions of CD133+ U87 glioblastoma cells have in common
labeling identical amounts of small RNAs for each sample. a total of 152 miRNAs, included 91 miRNAs up-regulation
miRNAs were considered as detected in exosomes and kept and 61 miRNAs down-regulation (Supplementary file 1).
for further analysis when the averaged intensities of the cor- At the quantitative level, the microRNA profiles of the 20
responding array probe sets were found to be higher than miRNAs that were found in common in the two types of
two times the background in at least one exosome sample exosome samples were found to be very different in the
type (either hypoxia or normoxia CD133+ U87 glioblas- exosomes secreted by the hypoxia CD133+ U87 glioblas-
toma cells derived exosomes). Comparison between the toma cells versus the normoxia CD133+ U87 glioblastoma

13
J Neurooncol (2017) 135:37–46 41

Fig. 2  characterization of exosomes from CD133+ U87 in hypoxia 43 nm (scale bar 100 nm). b Representative transmission electron
and normoxia conditions. a Representative transmission electron micrograph of CD133+ U87 exosomes in normoxia condition shows
micrograph of CD133+ U87 exosomes in hypoxia condition show exosomes with a typical diameter of approximately 43 nm (scale bar
a relatively uniform population of membrane-bound vesicles rang- 50 nm)
ing between 25 and 65 nm in diameter, with an average diameter of

cells (Fig. 3). This was exemplified by the values of the miR-4428, miR-2681, miR-6847, miR-4283, miR-6876; with
ratios between the amounts of the individual microRNAs in the highest hypoxia-to-normoxia ratio up to 7.4) and, on the
hypoxia CD133+ U87 glioblastoma cells exosomes and nor- contrary, 6 microRNAs (miR-6871, miR-1273, miR-6807,
moxia CD133+ U87 glioblastoma cells exosomes (Fig. 4). miR-5585, miR-1292, miR-6891) were detected in signifi-
Significantly increased amounts were observed in hypoxia cantly larger amounts in normoxia CD133+ U87 glioblas-
CD133+ U87 glioblastoma cells exosomes compared toma cells exosomes compared with hypoxia CD133+ U87
with normoxia CD133+ U87 glioblastoma cells exosomes glioblastoma cells exosomes (with a hypoxia-to-normoxia
for 14 miRs (miR-3156, miR-4486, miR-6826, miR-373, level ratio lower than 0.5; Fig. 4). The results of FDR also
miR-7845, miR-4649, miR-4278, miR-1910, miR-6885, confirm our conclusion (Q-value, Supplement 1).

Fig. 3  Exosomal miRNA profiles differ for both hypoxia and nor- tion intensities correspond to the measured values minus the thresh-
moxia CD133+ U87. Barplots represent the detection levels on old value (see “Materials and methods”). Errors bars mean ± SEM.
Affymetrix microarrays of the microRNAs in exosomes for hypoxia Y-axes are in a log10 scale
CD133+ U87 (grey) and normoxia CD133+ U87 (white). Detec-

13
42 J Neurooncol (2017) 135:37–46

Fig. 4  Hypoxia and normoxia CD133+ U87 derived exosomal RNA the intensities in normoxia CD133+ U87 exosomes. The ratios are
profiles are different. Barplot of the ratios of the measured intensities shown in a log2 scale
of microRNAs detected in hypoxia CD133+ U87 exosomes versus

Gene ontology analysis of the predicted targets of the glioma cells, cyto-activity of control cells was measured as
top 10 selectively packaged miRNAs revealed significant 1.423 ± 0.015 (n = 6) (Fig. 5b). Cell viability of cells treated
enrichment of several functional pathways (Supplementary with exosomes drawn from normoxia CD133+ U87 glioblas-
File 2); most highly enriched was the KEGG pathway of toma cells (1.642 ± 0.031, n = 6; Fig. 5b) showed no signifi-
Proteoglycans in cancer, a pathway that plays a critical role cant change compared with controls. While cells treated with
in tumour microenvironment and angiogenesis [15]. exosomes from hypoxia CD133+ U87 glioblastoma cells
exhibited significant increase in cell viability (1.966 ± 0.073,
CD133+ U87 glioblastoma cells derived exosomes n = 6, P < 0.05) compared to controls (Fig. 5b). These data
promote glioma cells and vascular endothelial cells suggest that exosomes released in hypoxia CD133+ U87
proliferation glioblastoma cells effectively increased cell viability of
glioma cells.
We exposed cultured human brain microvascular endothelial
cells (HBMEC) and U87 glioma cells to hypoxia and nor-
moxia conditions CD133+ U87 glioblastoma cells derived Discussion
exosomes respectively to determine whether the molecular
cargo of exosomes could affect the proliferation of HBMEC It has been recently demonstrated by experimental stud-
and U87 glioma cells. After 48 h of exposure the HBMEC ies that GSCs are enriched in specific niches around tumor
and U87 were washed and the cyto-activities were measured vessels and areas of necrosis, the latter associated with
by CCK-8. For HBMEC, cyto-activity of control cells was restricted oxygen levels. And accumulating opinions show
measured as 0.655 ± 0.013 (n = 6) (Fig. 5a). Cell viability of that hypoxia play pivotal roles in tumor neovascularization
cells treated with exosomes drawn from normoxia CD133+ [16]. Hence, GSCs display a symbiotic relationship with
U87 glioblastoma cells (0.765 ± 0.015, n = 6; Fig. 5a) proliferative and hypoxic niches [2, 17, 18]. Tumor cells
showed no significant change compared with controls. can reshape the microenvironment, which is done by the
While cells treated with exosomes from hypoxia CD133+ release of soluble factors, e.g., by creating a distinct pheno-
U87 glioblastoma cells exhibited significant increase in type of the tumor endothelium compared with that of normal
cell viability (1.034 ± 0.033, n = 6, P < 0.05) compared vasculature [19]. Progress in isolating and identifying these
to controls (Fig. 5a). These data suggest that exosomes factors contributed to a new era in cancer therapeutics devel-
released in hypoxia CD133+ U87 glioblastoma cells effec- opment. The complexity of the hypoxic signaling response
tively increased cell viability of endothelial cells. For U87 in the tumor microenvironment clearly needs to be better

13
J Neurooncol (2017) 135:37–46 43

exosome-associated miRNAs in target cells. These results

support the notion that hypoxic exosomes mediate proan-
giogenic response and glioma cells proliferation through the
concerted action of hundreds or even thousands of effector
molecules. Nevertheless, there are some insufficiency still
exits in our research, e.g. the effect of the subpopulations
of exosomes did not explore, and cell free DNA as well as
the supernatant after exosomes isolation should also be ana-
lyzed. In the future, whether the subpopulations of exosomes
have (or have not) different effects will be study in depth.
Previous studies have suggested that hypoxic compared
with normoxic cells produce higher amounts of exosomes
[27], and that more than half of the secreted proteome from
hypoxic carcinoma cells may be associated with exosomes
[28], providing quantitative data to support a complex role
of exosomes in the hypoxic response. Strategies aimed at
targeting this system should thus focus on general mecha-
nisms of exosome-dependent intercellular communication,
e.g., exosome uptake by recipient cells, exosome formation,
and exosome stability in the extracellular milieu, rather than
on single exosome constituents. In addition, some scholars
have explained the role of exosomes in vivo. For example,
Fig. 5  Exosomes derived from CD133+ U87 enhanced viability of HERNÁN et al. have performed a Matrigel plug assay to find
in vitro HBMECs and U87 glioma cells. a Effects of exosomes were
that Jagged 1 incorporated in exosomes induces angiogen-
tested on the viability of HBMECs. Cell viability of HBMECs treated
with hypoxia or normoxia CD133+ U87 derived exosomes was esis in a HIF-1α-dependent manner in athymic nude mice
shown after 48 h incubation. b Effects of exosomes were tested on [29]. Others like Tomohiro et al. established hypoxia-resist-
the viability of U87 glioma cells. Cell viability of U87 treated with ant cells that can mimic in vivo conditions of hypoxic bone
hypoxia or normoxia CD133+ U87 derived exosomes was shown
marrow [30]. These in vivo results allow us to understand
after 48 h incubation. The viability was measured by CCK-8. Data
were represented as the mean ± SD from three independent experi- the effect of exosomes better.
ments (*P < 0.05) MiRNAs is a novel class of regulators, which are the
key elements in exosomes. Sequence analysis showed that
miRNAs were the most abundant in a diverse collection
defined to develop more rational therapies [20–22]. In the of the exosomal RNA species, making up over 76% of all
previous published studies, Paulina et al. showed exosomes mappable reads [31]. Most previous studies on the RNA
reflected the hypoxic status of glioma cells and mediated content of exosomes have focused on miRNAs [4, 5, 32,
hypoxia-dependent activation of vascular cells during tumor 33]; these small noncoding RNAs that base-pair with
development [23]. Bethany et al. reviewed some literatures the 3′ untranslated regions (UTRs) of protein-encoding
and showed that exosome-mediated intercommunication mRNAs, resulting in mRNA destabilization or transla-
among cancer stem cells, adult stem cells, cancer cells, tional inhibition [34]. In this study, we have found that
and stromal cells within the tumor microenvironment [24]. exosomes released from CD133+ U87 glioblastoma cells
Here, we provide evidence that exosome vesicles constitute are predominantly of exosomal origin and carry a complex
a potent mediator of hypoxia-dependent intercellular com- cargo of miRNA species, many of which are novel and
munication between CD133+ U87 glioblastoma cells and of unknown function. Taking an unbiased approach with
vascular cells, suggesting an important role of exosomes in miRNA deep sequencing, we have found that the repertoire
proliferation of the tumor vasculature. Another significant of miRNAs in exosomes are strikingly distinct between
finding of our work is that exosome vesicles from hypoxic the hypoxia and normoxia microenvironment, for exam-
conditions promote the glioma cells proliferation. Exosomes ple, miR-3156, miR-373, miR-6826, miR-494 attracted
may participate in intercellular signaling at several levels, our attention, especially miR-373. miR-373 secreted from
e.g., during the cell-to-cell communication, exosomes exosome vesicles expression in the larger differences, this
released by tumor may have a function of spreading their result have shown in supplement1. Chen et al. showed that
oncogenic activity among cells, which may include the miR-373 stimulates the epithelial-to-mesenchymal transi-
propagation of tumor-associated miRNAs [4, 5, 25, 26]. tion (EMT), migration and invasion through thioredoxin-
Our data show that hypoxic regulation of a multitude of interacting protein (TXNIP) dependent reactive oxygen

13
44 J Neurooncol (2017) 135:37–46

species (ROS) reduction, and also correlation with HIF-1α noninvasive biomarker to assess the oxygenation status and
[35]. Eichelser et al. showed that serum levels of exosomal aggressiveness of malignant tumors.
miR-373 are linked to triple negative and more aggres- The complexity of the hypoxic signaling response in the
sive breast carcinomas [36]. These findings confirm that tumor microenvironment clearly needs to be better defined
miR-373 plays an important role in breast cancer, how- to develop more rational therapies [20, 23, 47]. Here, we
ever, limited reported have shown miR-373 function in provide evidence that exosomes constitute a potent mediator
glioma, especially in the exosomes. Cheryl et al. showed of hypoxia-dependent intercellular communication between
that exposure of HBMECs to glioma exosome resulted in CD133+ U87 glioblastoma cells and vascular cells, sug-
significant changes in endothelial cell gene expression, gesting an important role of CD133+ U87 glioblastoma
which reflects the capacity of these particles to mediate cells derived exosomes in maintenance tumor microenvi-
intercellular communication [37]. Their data suggest that ronment, mediating a proangiogenic response and glioma
gliomas can modify the transcriptional landscape of their cells proliferation.
local environment through a variety of exosome-delivered
RNA-based pathways. Another studies have also found
that after irradiated cells, the protein and miRNA can also Conclusions
change, e.g. Ramesh et al. reported that radiation-induced
changes occurred in the protein and miRNA cargo of Exosomes are novel and unique effectors carrying out many
ELVs isolated from plasma, such as miR-204-5p, miR- biological messages. Aside from their conventional basal
92a-3p and miR-31-5p [38], and Arscott et al. showed intracellular communications, exosomes can also actively
that radiation influenced exosome abundance, specifically participate in triggering signal pathways and exclusively
can alters their molecular composition, and on uptake, transfer nucleic acids as miRNA clusters. Due to its charac-
promotes a migratory phenotype in glioblastoma [39]. teristics, CD133+ U87 glioblastoma cells derived exosome-
In our data, we indicate CD133+ U87 glioblastoma cells mediated miRNA transduction is one of the autogenous
derived exosomes may deliver the hypoxia related miR- mediators to exchange specific and endogenous tumor-
NAs through the regulation of hypoxia microenvironment. related genetic signals among multiple types of cells, which
A previous study on glioma exosomes used antibody has attracted more attention on CD133+ glioblastoma cells-
arrays shows that angiogenic proteins (such as angiogenin, glioma cell and CD133+ glioblastoma cells-endothelial cell
IL-6 and IL-8) were capable of stimulating endothelial cell interaction in the tumor microenvironment. Furthermore,
growth and tubule formation in vitro [4]. Cheryl et al. found our experimental evidence more or less strongly supports a
that many transcripts down-regulated in endothelial cells hypotheses which hypoxia play a critical role in gliomas by
exposed to glioma exosome were predicted targets of miR- creating a microenvironment. Therapy by exosome mediated
NAs that were either abundant or specifically enriched in the miRNA transduction may provide new therapeutic strategies
exosome [37]. KEGG pathway analysis of these predicted for glioma.
miRNA targets identified GAG synthesis, C ­ a2+ reabsorp-
tion, and ECM-receptor interactions as highly significant, Acknowledgements This study was supported by grants from
over-represented functions. These pathways were relevant the Beijing Tiantan Hospital “Nursery Project” Foundation (No.
2014MP09), the National Natural Science Foundation of China (Nos.
to glioma biology [40–42]. In our study, the KEGG path- 81372354, 81302186), the Beijing Municipal Natural Science Founda-
way analysis of the predicted miRNA targets identified tion (No. 7151002), the Beijing Health System High-level Personnel
Proteoglycans in cancer, Glutamatergic synapse, Signaling Building Foundation (No. 2013-3-018), the Beijing Laboratory of Bio-
pathways regulating pluripotency of stem cells, PI3K-Akt medical Materials Foundation (PXM2014_014226_000005) and the
Beijing Municipal Administration of Hospitals’ Youth Program (No.
signaling pathway, mTOR signaling pathway, TGF-beta QML20150505). We would like to thank Prof. Yilin Sun and Cuiping
signaling pathway, HIF-1 signaling pathway, Ras signaling Zhang (Pathological Department, Beijing Neurosurgical Institute) for
pathway, Hippo signaling pathway, and Wnt signaling path- observing exosomes by transmission electron microscopy.
way as the highly significant functions. All of these path-
Compliance with ethical standards
ways were relevant to glioma microenvironment, growth and
angiogenesis, e.g. Wnt signaling pathway was a powerful
Conflict of interest No potential conflicts of interest were disclosed.
modulator of tumor growth, invasion and angiogenesis [43],
PI3K-Akt, mTOR and Hippo signaling pathway modulated
cell death and proliferation [44, 45], and TGF-beta signaling
pathway were critical for tumor infiltration [46]. In addi- References
tion, our data suggest that exosomes constitute a potentially
1. Sundar SJ, Hsieh JK, Manjila S, Lathia JD, Sloan A (2014)
targetable mediator of hypoxia-driven tumor development, The role of cancer stem cells in glioblastoma. Neurosurg Focus
and that the exosomal molecular signature may serve as a 37(6):E6. doi:10.3171/2014.9.FOCUS14494

13
J Neurooncol (2017) 135:37–46 45

2. Calabrese C, Poppleton H, Kocak M, Hogg TL, Fuller C, Ham- evidence for tumor propagation by glioblastoma cancer stem cells.
ner B, Oh EY, Gaber MW, Finklestein D, Allen M, Frank A, PLoS ONE 6(9):e24807. doi:10.1371/journal.pone.0024807
Bayazitov IT, Zakharenko SS, Gajjar A, Davidoff A, Gilbertson 18. Li Z, Bao S, Wu Q, Wang H, Eyler C, Sathornsumetee S, Shi
RJ (2007) A perivascular niche for brain tumor stem cells. Can- Q, Cao Y, Lathia J, McLendon RE, Hjelmeland AB, Rich JN
cer Cell 11(1):69–82. doi:10.1016/j.ccr.2006.11.020 (2009) Hypoxia-inducible factors regulate tumorigenic capacity
3. Kosaka N (2016) Decoding the secret of cancer by means of of glioma stem cells. Cancer Cell 15(6):501–513. doi:10.1016/j.
extracellular vesicles. J Clin Med. doi:10.3390/jcm5020022 ccr.2009.03.018
4. Skog J, Wurdinger T, van Rijn S, Meijer DH, Gainche L, Sena- 19. Hardee ME, Zagzag D (2012) Mechanisms of glioma-asso-
Esteves M, Curry WT Jr, Carter BS, Krichevsky AM, Breake- ciated neovascularization. Am J Pathol 181(4):1126–1141.
field XO (2008) Glioblastoma microvesicles transport RNA and doi:10.1016/j.ajpath.2012.06.030
proteins that promote tumour growth and provide diagnostic 20. Wilson WR, Hay MP (2011) Targeting hypoxia in cancer therapy.
biomarkers. Nat Cell Biol 10(12):1470–1476. doi:10.1038/ Nat Rev Cancer 11(6):393–410. doi:10.1038/nrc3064
ncb1800 21. Liao D, Johnson RS (2007) Hypoxia: a key regulator of angi-
5. Valadi H, Ekstrom K, Bossios A, Sjostrand M, Lee JJ, Lotvall JO ogenesis in cancer. Cancer Metastasis Rev 26(2):281–290.
(2007) Exosome-mediated transfer of mRNAs and microRNAs is doi:10.1007/s10555-007-9066-y
a novel mechanism of genetic exchange between cells. Nat Cell 22. Pouyssegur J, Dayan F, Mazure NM (2006) Hypoxia signalling
Biol 9(6):654–659. doi:10.1038/ncb1596 in cancer and approaches to enforce tumour regression. Nature
6. Ratajczak J, Wysoczynski M, Hayek F, Janowska-Wieczorek A, 441(7092):437–443. doi:10.1038/nature04871
Ratajczak MZ (2006) Membrane-derived microvesicles: impor- 23. Kucharzewska P, Christianson HC, Welch JE, Svensson KJ, Fred-
tant and underappreciated mediators of cell-to-cell communica- lund E, Ringner M, Morgelin M, Bourseau-Guilmain E, Beng-
tion. Leukemia 20(9):1487–1495. doi:10.1038/sj.leu.2404296 zon J, Belting M (2013) Exosomes reflect the hypoxic status of
7. Lages E, Ipas H, Guttin A, Nesr H, Berger F, Issartel JP (2012) glioma cells and mediate hypoxia-dependent activation of vas-
MicroRNAs: molecular features and role in cancer. Front Biosci cular cells during tumor development. Proc Natl Acad Sci USA
17:2508–2540 110(18):7312–7317. doi:10.1073/pnas.1220998110
8. Zhou W, Fong MY, Min Y, Somlo G, Liu L, Palomares MR, Yu 24. Hannafon BN, Ding WQ (2015) Cancer stem cells and exo-
Y, Chow A, O’Connor ST, Chin AR, Yen Y, Wang Y, Marcus- some signaling. Stem Cell Investig 2:11. doi:10.3978/j.
son EG, Chu P, Wu J, Wu X, Li AX, Li Z, Gao H, Ren X, Boldin issn.2306-9759.2015.05.02
MP, Lin PC, Wang SE (2014) Cancer-secreted miR-105 destroys 25. Schwarzenbach H (2015) The clinical relevance of circulating,
vascular endothelial barriers to promote metastasis. Cancer Cell exosomal miRNAs as biomarkers for cancer. Expert Rev Mol
25(4):501–515. doi:10.1016/j.ccr.2014.03.007 Diagn 15(9):1159–1169. doi:10.1586/14737159.2015.1069183
9. Zhang G, Jin G, Nie X, Mi R, Zhu G, Jia W, Liu F (2014) 26. Al-Nedawi K, Meehan B, Rak J (2009) Microvesicles: messengers
Enhanced antitumor efficacy of an oncolytic herpes simplex and mediators of tumor progression. Cell Cycle 8(13):2014–2018.
virus expressing an endostatin-angiostatin fusion gene in human doi:10.4161/cc.8.13.8988
glioblastoma stem cell xenografts. PLoS ONE 9(4):e95872. 27. King HW, Michael MZ, Gleadle JM (2012) Hypoxic enhancement
doi:10.1371/journal.pone.0095872 of exosome release by breast cancer cells. BMC Cancer 12:421.
10. Zhu G, Su W, Jin G, Xu F, Hao S, Guan F, Jia W, Liu F doi:10.1186/1471-2407-12-421
(2011) Glioma stem cells targeted by oncolytic virus carrying 28. Park JE, Tan HS, Datta A, Lai RC, Zhang H, Meng W, Lim SK,
endostatin-angiostatin fusion gene and the expression of its Sze SK (2010) Hypoxic tumor cell modulates its microenviron-
exogenous gene in vitro. Brain Res 1390:59–69. doi:10.1016/j. ment to enhance angiogenic and metastatic potential by secretion
brainres.2011.03.050 of proteins and exosomes. Mol Cell Proteomics 9(6):1085–1099.
11. Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, doi:10.1074/mcp.M900381-MCP200
Dewhirst MW, Bigner DD, Rich JN (2006) Glioma stem cells 29. Gonzalez-King H, Garcia NA, Ontoria-Oviedo I, Ciria M,
promote radioresistance by preferential activation of the DNA Montero JA, Sepulveda P (2017) Hypoxia inducible factor-1alpha
damage response. Nature 444(7120):756–760. doi:10.1038/ potentiates jagged 1-mediated angiogenesis by mesenchymal stem
nature05236 cell-derived exosomes. Stem Cells. doi:10.1002/stem.2618
12. Bao S, Wu Q, Sathornsumetee S, Hao Y, Li Z, Hjelmeland AB, 30. Umezu T, Tadokoro H, Azuma K, Yoshizawa S, Ohyashiki K,
Shi Q, McLendon RE, Bigner DD, Rich JN (2006) Stem cell- Ohyashiki JH (2014) Exosomal miR-135b shed from hypoxic
like glioma cells promote tumor angiogenesis through vascu- multiple myeloma cells enhances angiogenesis by targeting fac-
lar endothelial growth factor. Cancer Res 66(16):7843–7848. tor-inhibiting HIF-1. Blood 124(25):3748–3757. doi:10.1182/
doi:10.1158/0008-5472.CAN-06-1010 blood-2014-05-576116
13. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire 31. Huang X, Yuan T, Tschannen M, Sun Z, Jacob H, Du M, Liang
J, Dirks PB (2003) Identification of a cancer stem cell in human M, Dittmar RL, Liu Y, Liang M, Kohli M, Thibodeau SN, Board-
brain tumors. Cancer Res 63(18):5821–5828 man L, Wang L (2013) Characterization of human plasma-derived
14. Yuan X, Curtin J, Xiong Y, Liu G, Waschsmann-Hogiu S, Farkas exosomal RNAs by deep sequencing. BMC Genomics 14:319.
DL, Black KL, Yu JS (2004) Isolation of cancer stem cells from doi:10.1186/1471-2164-14-319
adult glioblastoma multiforme. Oncogene 23(58):9392–9400. 32. Jaiswal R, Luk F, Gong J, Mathys JM, Grau GE, Bebawy M
doi:10.1038/sj.onc.1208311 (2012) Microparticle conferred microRNA profiles–implications
15. Iozzo RV, Sanderson RD (2011) Proteoglycans in cancer biol- in the transfer and dominance of cancer traits. Mol Cancer 11:37.
ogy, tumour microenvironment and angiogenesis. J Cell Mol Med doi:10.1186/1476-4598-11-37
15(5):1013–1031. doi:10.1111/j.1582-4934.2010.01236.x 33. Pegtel DM, Cosmopoulos K, Thorley-Lawson DA, van Eijndhoven
16. Zagzag D, Zhong H, Scalzitti JM, Laughner E, Simons JW, MA, Hopmans ES, Lindenberg JL, de Gruijl TD, Wurdinger T,
Semenza GL (2000) Expression of hypoxia-inducible factor Middeldorp JM (2010) Functional delivery of viral miRNAs
1alpha in brain tumors: association with angiogenesis, invasion, via exosomes. Proc Natl Acad Sci USA 107(14):6328–6333.
and progression. Cancer 88(11):2606–2618 doi:10.1073/pnas.0914843107
17. Lathia JD, Gallagher J, Myers JT, Li M, Vasanji A, McLendon 34. Bartel DP (2009) MicroRNAs: target recognition and regulatory
RE, Hjelmeland AB, Huang AY, Rich JN (2011) Direct in vivo functions. Cell 136(2):215–233. doi:10.1016/j.cell.2009.01.002

13
46 J Neurooncol (2017) 135:37–46

35. Chen D, Dang BL, Huang JZ, Chen M, Wu D, Xu ML, Li R, Yan 42. Sasisekharan R, Shriver Z, Venkataraman G, Narayanasami U
GR (2015) MiR-373 drives the epithelial-to-mesenchymal transi- (2002) Roles of heparan-sulphate glycosaminoglycans in cancer.
tion and metastasis via the miR-373-TXNIP-HIF1alpha-TWIST Nat Rev Cancer 2(7):521–528. doi:10.1038/nrc842
signaling axis in breast cancer. Oncotarget 6(32):32701–32712. 43. Hu B, Wang Q, Wang YA, Hua S, Sauve CG, Ong D, Lan ZD,
doi:10.18632/oncotarget.4702 Chang Q, Ho YW, Monasterio MM, Lu X, Zhong Y, Zhang J,
36. Eichelser C, Stuckrath I, Muller V, Milde-Langosch K, Wikman Deng P, Tan Z, Wang G, Liao WT, Corley LJ, Yan H, Zhang
H, Pantel K, Schwarzenbach H (2014) Increased serum levels of J, You Y, Liu N, Cai L, Finocchiaro G, Phillips JJ, Berger MS,
circulating exosomal microRNA-373 in receptor-negative breast Spring DJ, Hu J, Sulman EP, Fuller GN, Chin L, Verhaak RG,
cancer patients. Oncotarget 5(20):9650–9663. doi:10.18632/ DePinho RA (2016) Epigenetic activation of WNT5A drives
oncotarget.2520 glioblastoma stem cell differentiation and invasive growth. Cell
37. Li CC, Eaton SA, Young PE, Lee M, Shuttleworth R, Humphreys 167(5):1281–1295 e1218. doi:10.1016/j.cell.2016.10.039
DT, Grau GE, Combes V, Bebawy M, Gong J, Brammah S, Buck- 44. Chao Y, Wang Y, Liu X, Ma P, Shi Y, Gao J, Shi Q, Hu J, Yu R,
land ME, Suter CM (2013) Glioma microvesicles carry selectively Zhou X (2015) Mst1 regulates glioma cell proliferation via the
packaged coding and non-coding RNAs which alter gene expres- AKT/mTOR signaling pathway. J Neurooncol 121(2):279–288.
sion in recipient cells. RNA Biol 10(8):1333–1344. doi:10.4161/ doi:10.1007/s11060-014-1654-4
rna.25281 45. Artinian N, Cloninger C, Holmes B, Benavides-Serrato A, Bashir
38. Yentrapalli R, Merl-Pham J, Azimzadeh O, Mutschelknaus L, T, Gera J (2015) Phosphorylation of the hippo pathway compo-
Peters C, Hauck SM, Atkinson MJ, Tapio S, Moertl S (2017) nent AMOTL2 by the mTORC2 kinase promotes YAP signaling,
Quantitative changes in the protein and miRNA cargo of plasma resulting in enhanced glioblastoma growth and invasiveness. J
exosome-like vesicles after exposure to ionizing radiation. Int J Biol Chem 290(32):19387–19401. doi:10.1074/jbc.M115.656587
Radiat Biol. doi:10.1080/09553002.2017.1294772 46. Roy LO, Poirier MB, Fortin D (2015) Transforming growth fac-
39. Arscott WT, Tandle AT, Zhao S, Shabason JE, Gordon IK, Schlaff tor-beta and its implication in the malignancy of gliomas. Target
CD, Zhang G, Tofilon PJ, Camphausen KA (2013) Ionizing radia- Oncol 10(1):1–14. doi:10.1007/s11523-014-0308-y
tion and glioblastoma exosomes: implications in tumor biology 47. Semenza GL (2010) Defining the role of hypoxia-inducible factor
and cell migration. Transl Oncol 6(6):638–648 1 in cancer biology and therapeutics. Oncogene 29(5):625–634.
40. Lu P, Weaver VM, Werb Z (2012) The extracellular matrix: a doi:10.1038/onc.2009.441
dynamic niche in cancer progression. J Cell Biol 196(4):395–406.
doi:10.1083/jcb.201102147
41. Monteith GR, McAndrew D, Faddy HM, Roberts-Thomson SJ
(2007) Calcium and cancer: targeting Ca2+ transport. Nat Rev
Cancer 7(7):519–530. doi:10.1038/nrc2171

13