RESEARCH ARTICLE

Glioma Cells in the Tumor Periphery Have a

Stem Cell Phenotype
Sune Munthe1,2,3*, Stine Asferg Petterson1, Rikke Hedegaard Dahlrot4, Frantz
Rom Poulsen2,3, Steinbjørn Hansen2,4, Bjarne Winther Kristensen1,2
1 Department of Pathology, Odense University Hospital, 5000 Odense C, Denmark, 2 Institute of Clinical
Research, University of Southern Denmark, 5000 Odense C, Denmark, 3 Department of Neurosurgery,
Odense University Hospital, 5000 Odense C, Denmark, 4 Department of Oncology, Odense University
Hospital, 5000 Odense C, Denmark

* [email protected]

Abstract
a11111 Gliomas are highly infiltrative tumors incurable with surgery. Although surgery removes the
bulk tumor, tumor cells in the periphery are left behind resulting in tumor relapses. The aim
of the present study was to characterize the phenotype of tumor cells in the periphery focus-
ing on tumor stemness, proliferation and chemo-resistance. This was investigated in situ in
patient glioma tissue as well as in orthotopic glioblastoma xenografts. We identified 26 glio-
mas having the R132 mutation in Isocitrate DeHydrogenase 1 (mIDH1). A double immuno-
OPEN ACCESS fluorescence approach identifying mIDH1 positive tumor cells and a panel of markers was
Citation: Munthe S, Petterson SA, Dahlrot RH, used. The panel comprised of six stem cell-related markers (CD133, Musashi-1, Bmi-1,
Poulsen FR, Hansen S, Kristensen BW (2016) Sox-2, Nestin and Glut-3), a proliferation marker (Ki-67) as well as a chemo-resistance
Glioma Cells in the Tumor Periphery Have a Stem
marker (MGMT). Computer-based automated classifiers were designed to measure the
Cell Phenotype. PLoS ONE 11(5): e0155106.
doi:10.1371/journal.pone.0155106 mIDH1 positive nucleus area-fraction of the chosen markers. Moreover, orthotopic glioblas-
toma xenografts from five different patient-derived spheroid cultures were obtained and the
Editor: Maria G Castro, University of Michigan
School of Medicine, UNITED STATES tumor cells identified by human specific immunohistochemical markers. The results showed
that tumor cells in the periphery of patient gliomas expressed stem cell markers, however
Received: November 16, 2015
for most markers at a significantly lower level than in the tumor core. The Ki-67 level was
Accepted: April 25, 2016
slightly reduced in the periphery, whereas the MGMT level was similar. In orthotopic glio-
Published: May 12, 2016 blastoma xenografts all markers showed similar levels in the core and periphery. In conclu-
Copyright: © 2016 Munthe et al. This is an open sion tumor cells in the periphery of patient gliomas have a stem cell phenotype, although it
access article distributed under the terms of the is less pronounced than in the tumor core. Novel therapies aiming at preventing recurrence
Creative Commons Attribution License, which permits
should therefore take tumor stemness into account. Migrating cells in orthotopic glioblas-
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are toma xenografts preserve expression and stem cell markers. The orthotopic model there-
credited. fore has a promising translational potential.
Data Availability Statement: Data are available from
Figshare (https://figshare.com/s/
939a79008964ca53bfc1).

Funding: The authors have no support or funding to

report. Introduction
Competing Interests: The authors have declared Treatment of gliomas is a major challenge. Despite treatment consisting of surgery, chemother-
that no competing interests exist. apy and radiation the mean survival of patients with the most common and malignant primary

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 1 / 16

 Glioma Tumor Periphery Display Stem Cell

brain tumor, the WHO grade IV Glioblastoma multiforme (GBM), is approximately 14.6
months [1]. A major challenge in treatment of gliomas is the high migratory potential of gli-
oma cells [2]. Migrating glioma cells are not eligible to surgery and therefore eradication by
radiation and chemotherapy is standard strategy.
The high resistance of gliomas against conventional radiation and chemotherapy has been
suggested to be due to the existence of immature cancer stem cells (CSC), which are tumor
cells with a stem cell-like phenotype sustaining glioma growth through asymmetric cell divi-
sion [3, 4]. These cells have been suggested to be resistant towards conventional treatment due
to enhanced DNA repair and enhanced expression of ATP-binding cassette drug transporters
[5]. The presence of CSCs in the periphery of gliomas has not previously been thoroughly
investigated. We therefore hypothesized that migrating glioma cells display a stem cell pheno-
type and express stem cell markers. Since proliferation and chemo-resistance are important
determinants for the effect of radiation and chemotherapy, the markers Ki-67 and MGMT
were included. The aim of this study was to characterize the expression of stem cell markers as
well as markers of proliferation and chemo-resistance in migrating glioma cells by using a dou-
ble immunofluorescence approach on patient glioma tissue and GBM xenografts.
In patient tumor tissue a mutated form of Isocitrate dehydrogenase 1 (mIDH1) [6] was
used as a tumor cell specific marker. The somatic point mutation that affects codon 132 is the
most frequent IDH1 mutation and a specific well described antibody recognizing mIDH1 R132
has recently been developed [7, 8]. This mutation is primarily associated with grade II and III
gliomas but is also found in secondary GBMs [9].
In glioma research the most preferred in vivo model is the orthotopic xenograft model. Here
GBM cells are implanted into the brain of immunosuppressed animals. The model is well
established in our laboratory and has been described by several groups [10, 11]. Using this
model together with a double immunofluorescence approach and human specific markers to
identify the tumor cells, the marker expression in migrating glioma cells was characterized. For
this characterization a panel of markers similar to that used for patient gliomas was used.
We have previously used fluorescence for quantification of biomarkers in gliomas [12]. In
this study we combined the tumor cell specific markers in a double immunofluorescence proto-
col with the following panel of stem cell markers: CD133 [4, 13–18], Nestin [18–24], Musashi-
1 [18, 25–28], Sox-2 [18, 29–32] and Bmi-1 [18, 29, 32, 33], to characterize the phenotype of
migrating glioma cells. We also investigated the glucose transporter type III (Glut-3) in the
core and periphery since it has been reported to be associated with stemness [34]. Furthermore,
we investigated the expression of the DNA repair enzyme O6-methylguanine-DNA methyl-
transferase (MGMT) in the core of the tumor and compared it to the periphery, since MGMT
is a strong prognostic and predictive marker for the effect of temozolomide in the upfront
GBM treatment [35]. The proliferation of migrating tumor cells has previously been investi-
gated by Sabit et al. [36] who investigated 11 patients. We wanted to futher explore this in our
cohort consisting of 26 patients as well as by using the orthotopic GBM xenograft model.

Material and Methods

Patient selection and pathology
Adult residents in the Region of Southern Denmark diagnosed with primary brain cancer
between 1st of January 2005 and 31st of December 2009 were identified in the Danish Cancer
Register (DCR). Patients with the histopathological codes for gliomas (M 94003, M 94013, M
94403, M 94503, M 94513, M 93823, and M 93853) were considered for inclusion in the pres-
ent study, giving a total of 277 patients. The mIDH1 status was identified by immunohisto-
chemical staining; 47 patients were mIDH1 positive. Twenty six patients were identified with

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 2 / 16

 Glioma Tumor Periphery Display Stem Cell

Table 1. Summary of patients.

Histopathological diagnosis Number of mIDH1 patients Age (years) WHO grade

Median Range
diffuse astrocytoma 6 (23.1%) 41 30–62 II
oligodendroglioma 7 (26.9%) 48.9 31–81 II
oligoastrocytoma 5 (19.2%) 46.8 26–74 II
anaplastic astrocytoma 2 (7.7%) 46 34–58 III
anaplastic oligodendroglioma 1 (3.8%) 38 38 III
anaplastic oligoastrocytoma 2 (7.7%) 41 36–46 III
glioblastoma 3 (11.5 %) 60.3 43–75 IV
Total 26 46.7 26–81
doi:10.1371/journal.pone.0155106.t001

both a tumor core and a periphery zone present in the same tumor section. The different
mIDH1 positive gliomas are listed according to the WHO classification in Table 1. The histo-
pathological evaluation was carried out at the Department of Pathology at Odense University
Hospital. All tissue samples were evaluated by two pathologists and classified according to
WHO guidelines 2007 [37].

Xenograft model
For establishing the orthotopic xenograft model we used 5 different GBM spheroid cultures,
established in our laboratory from patient-derived GBM tumor tissue, collected at the Depart-
ment of Neurosurgery, Odense University Hospital. Tumor cells were grown and stem cell phe-
notypes validated as previously described in our laboratory [38]. The five GBM spheroid
cultures were named: T78, T86, T87, T111 and T113.
Female Balb c nu/nu mice 7–8 weeks of age were anesthetized subcutaneously with injection
of a mixture of Hypnorm and Dormicum (0,1ml/10g). The mice were placed in a stereotactic
frame (Model 900, David Kopf Instrument). A midline incision was made exposing bregma
and a burr hole was made 1 mm anterior and 2 mm lateral to bregma. A syringe (2 μl Hamilton
syringe) with a blunt needle containing 150.000 cells/μl was inserted 3 mm into the brain and
2 μl were injected slowly into the brain over several minutes. The needle was slowly removed to
prevent a vacuum causing the tumor cells to escape. The skin was sutured with resorbable
sutures. If the mice showed any signs of neurological deficit or weight loss more than 20%,
the mice were euthanized in a carbon dioxide chamber. The brains were immediately removed
and fixated in 4% formalin for 48 hours. Before paraffin embedding the brains were manually
divided by 1mm coronal sections. Histological sections were afterwards cut and immunohisto-
chemically stained with Vimentin, a human specific antibody. In mice implanted with the
GBM spheroid culture T86 CD56, another human specific antibody, was used for identification
of the tumor cells. The GBM spheroid cultures T86 and T113 were found to have unmethylated
MGMT promoter, whereas the remaining GBM spheroid cultures were methylated in the pro-
moter region. We therefore did not investigate the MGMT in the in vivo model. We defined
the core in our in vivo model as striatum and the periphery as corpus callosum in the contra
lateral side.

Immunohistochemical staining
Histological sections of three μm were cut on a microtome and placed on glass slides. Tissue
was deparaffinized, and antigen retrieval was carried out in a microwave oven using the Tris

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 3 / 16

 Glioma Tumor Periphery Display Stem Cell

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 4 / 16

 Glioma Tumor Periphery Display Stem Cell

Fig 1. Double immunofluorescence staining of core and periphery of patient glioblastomas. Histological sections were
stained with Dapi (blue), IDH1 (green) and CD133 (red) (AA-BE), Nestin (red) (CA-DE) and Musashi-1 (red) (EA to FE). The
software-based classifier is shown in the right column. The classifier illustrates tumor cells co-expressing markers of interest in
yellow and tumor cells not co-expressing markers of interest in green. The fluorescence stainings were quantified in both core
and periphery for CD133 (G,J), Nestin (H,K) and Musashi-1 (I,L). Statistical comparison was performed using student’s t-test
and ANOVA, ** p< 0.01. Scalebar: 200μm.
doi:10.1371/journal.pone.0155106.g001

EGTA buffer. Slides were stained on the AutostainerPlus platform (DAKO, Glostrup, Den-
mark). For mIDH1 staining slides were incubated with antibody (mIDH1 R132H, Clone H09,
Dianova, 1:100) and the detection system ultraViewTM Universal DAB Detection Kit (Ven-
tana Medical Systems) was subsequently used. For each patient, tumor slide was stained by a
double immuno-fluorescence approach combining mIDH1 and 6 different stem cell-related
markers: CD133 (clone: W6B3C1, Miltenyi Biotec), Musashi-1 (clone:14H1, MBL Interna-
tional), Bmi-1 (clone: F6, Upstate Cell Signaling Solution), Sox-2 (clone: 245610, R&D Systems
Inc.), Nestin (clone: 196908, R&D Systems Inc.), Glut-3 (clone: HPA006539, Atlas Antibodies),
MGMT (clone: MT23.2, Invitrogen) and the proliferation marker Ki-67 (clone: MIB1, Beck-
man Coulter). The CSA II Biotin-free Tyramide signal Amplification System kit (DAKO) was
used for detection of mIDH1 (1+1600). Detection of the second antibody was performed using
a Tyramide Signal Amplification (TSA) Plus System with Cyanin 5 (Cy5); CD133 (1+40),
Musashi (1+200), Bmi-1 (1+200), Sox-2 (1+400), Nestin (1+200), GLUT-3 (1+100), MGMT
(1+100) and Ki67 (1+800). The nuclei were counterstained with 4´,6 diamidino-2-phemylin-
dole (Dapi) (VWR Internation ApS).
For each xenograft slide a double immuno-fluorescence staining with Vimentin and the same
six stem cell markers and Ki-67 was made. For xenograft slides with T86 tumors, CD56 was used as
a tumor marker since it was Vimentin negative. The Alexa Flour-488 donkey anti-rabbit (1+100)
was used to detect Vimentin (1+400, EP20, Epitomics) and CSA II Biotin-free Tyramide signal
Amplification System kit (DAKO) was used for detection of CD56 (1+3200, CO4-NCAM, Neo-
markers) in the T86 tissue samples. The nuclei were counterstained with 4´.6 diamidino-2-phemy-
lindole (Dapi) (VWR Internation ApS).

Automated Quantitative analysis

Super images of whole slides were taken at 1.25x magnification using a Leica DM6000 B micro-
scope with an Olympus DP72 camera with bright field settings. Subsequently the region of
interest (ROI), both the core and periphery, was manually outlined for each tumor section
using the Visiopharm Integrator System (VIS) version: 4.5.6.440 (Visiopharm, Hørsholm,
Denmark). Sampling was performed at 20x magnification with a minimum of 15 images in
each ROI. Images were reviewed to ensure that no artifacts or blurring were present. Then
images were analyzed using an algorithm developed in the Visiomorph software module. A
specific classifier was developed for each double-fluorescence staining according to the Visio-
pharm Manual. Classifiers were designed and trained so nuclear area corresponding to DAPI
staining of positive cells were measured. In each specific classifier we calculated the total area
of the nuclei of tumor cells of interest (labeled by tumor cell specific marker and marker of
interest) and the total area of remaining tumor cell nuclei (labeled by tumor cell specific marker
but not by marker of interest). From these areas, the nuclear area fraction of tumor cells labeled
by marker of interest was calculated in core and periphery. This fraction was determined based
on the whole mIDH1 positive cell population in each ROI. It was not determined on a single
cell basis, since core areas with very high cellularity did not allow complete separation of the
DAPI stained nuclei.

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 5 / 16

 Glioma Tumor Periphery Display Stem Cell

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 6 / 16

 Glioma Tumor Periphery Display Stem Cell

Fig 2. Double immunofluorescence staining of core and periphery of patient glioblastomas. Histological sections were
stained with Dapi (blue), IDH1 (green) and Sox-2 (red) (AA-BE), Bmi-1 (red) (CA-DE) and Glut-3 (red) (EA to FE). The software-
based classifier is shown in the right column. The classifier illustrates tumor cells co-expressing markers of interest in yellow and
tumor cells not co-expressing markers of interest in green. The fluorescence stainings were quantified in both core and periphery
for Sox-2 (G,J), Bmi-1 (H,K) and Glut-3 (I,L). Statistical comparison was performed using student’s t-test and ANOVA, ** p< 0.01.
Scalebar: 200μm.
doi:10.1371/journal.pone.0155106.g002

Statistics
Data was analyzed in GraphPad Prism version 5.01. The comparison of area fraction in the
core and periphery was performed with an unpaired t-test. Statistical significance was defined
as p < 0.05.

Ethics
The official Danish ethical review board named the Regional Scientific Ethical Committee of
the Region of Southern Demark approved the use of human glioma tissue (permission J. No. S-
2011 0022) in the current study. Written informant consent was obtained from all participants.
The use of animals in the present study was approved by The Animal Experiment Inspectorate
in Denmark (J. Nr. 2013/15-2934-00973).

Results
CD133 expression was significantly reduced in migrating tumor cells in the tumor periphery com-
pared to tumor cells in the core region when comparing levels in all tumor samples. Similar but
not significant results were obtained in the different grades and glioma subtypes (Fig 1G and 1J).
Nestin expression was significantly reduced in the periphery for tumor grade II and IV and
for all gliomas together (Fig 1H). For the different glioma subtypes only Oligo-Astrocytoma
(OA) and GBM revealed significantly reduced expression in the periphery although the same
trend was found in the other subtypes (Fig 1K).
The expression of Musashi-1 was significantly reduced in the periphery for grade II and IV and
in all gliomas together (Fig 1I). For the different subtypes of gliomas only GBM had a reduced
expression in the periphery, although the same trend was found in the other subtypes (Fig 1L).
The expression of Sox-2 was reduced in the periphery for all tumor grades and subtypes (S1
Fig) except for the Anaplastic Oligodendroglioma (AO), but without reaching significance (Fig
2G and 2J).
The expression of Bmi-1 was significantly reduced in the periphery of grade II tumors and
all gliomas together (Fig 2H). For subtypes, only the Diffuse Astrocytoma (DA) had a signifi-
cantly reduced expression in the periphery. The expression of Bmi-1 was reduced in the
periphery of all subtypes except the Anaplastic Oligodendroglioma (AO), but without reaching
significance (Fig 2K).
Glut-3 expression was apparently expressed at similar levels in core and periphery (Fig 2I
and 2L).
The expression of Ki-67 was significantly reduced in the periphery for all gliomas together
(Fig 3E). Similar but no significantly reduced expression of Ki-67 in periphery compared to
core was obtained in the different grades and glioma subtypes (Fig 3E and 3G).
MGMT was expressed at similar levels in core and periphery (Fig 3F and 3H).
Comparing expression of the different markers between grades, we did observe a significant
increase in the Ki-67 expression in the core area from grade II to grade IV (Fig 3E). There was
no significant increase of expression of other markers with grade.

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 7 / 16

 Glioma Tumor Periphery Display Stem Cell

Fig 3. Double immunofluorescence staining of core and periphery of patient gliomas. Two glioblastomas (AA-BE and CA-DE) were stained with Dapi
(blue), IDH1 (green) and Ki-67 (red) and MGMT (red). The software-based classifier is shown in the right column. The classifier illustrates tumor cells co-
expressing markers of interest in yellow and tumor cells not co-expressing markers of interest in green. The fluorescence stainings were quantified in both
core and periphery for Ki-67 (E,G) and MGMT (F,H). Statistical comparison was performed using student’s t-test and ANOVA, ** p< 0.01. Scalebar: 200μm.
doi:10.1371/journal.pone.0155106.g003

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 8 / 16

 Glioma Tumor Periphery Display Stem Cell

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 9 / 16

 Glioma Tumor Periphery Display Stem Cell

Fig 4. Double immunofluorescence staining of core and periphery in orthotopic model with five different patient-derived GBM spheroid
cultures. Tissues were stained with Dapi (blue), Vimentin/CD56 (green) and CD133 (red), Nestin (red) and Musashi-1 (red). The software-based
classifier is shown in the right column. The classifier illustrates tumor cells co-expressing markers of interest in yellow and tumor cells not co-
expressing markers of interest in green. The fluorescence stainings were quantified in both core and periphery for CD133 (G), Nestin (H) and Musashi-
1 (I). Statistical comparison was performed using student’s t-test. Scalebar: 200μm.
doi:10.1371/journal.pone.0155106.g004

In the GBM xenografts the different markers were expressed at similar levels in core and
periphery (Figs 4, 5 and 6).

Discussion
Our results suggest that tumor cells in the periphery express CSC-markers both in patient glio-
mas and in orthotopic xenografts. The CSC hypothesis states that tumor growth is driven by a
subpopulation of tumorigenic CSCs [39]. Huang et al. and Bao et al. showed that CSCs in
GBMs were extremely resistant to conventional radiation and chemotherapies [3, 40]. Together
these results suggest that the CSC hypothesis also extend to tumor cells left in the periphery
after surgery, although the vast majority of studies have been performed on removed central
tumor material. The presence of CSCs in the periphery in all glioma grades and sub-types indi-
cates that these CSCs could be the reason for treatment failure and recurrence in glioma
patients. Extending the CSC hypothesis to tumor cells left in the periphery after surgery for all
gliomas is fully in line with the well know observation that macro radical resection of the
tumors increases survival compared to biopsies and suboptimal resection [41, 42].
Expression of CD133 was found at lower level compared to the other stem cell markers.
This observation corresponds to what has been shown in other studies [43]. We have previ-
ously shown that it is of great importance which CD133 clone is used; e.g. whether the CD133
clone targets the non-glycosylated epitope (i.e. the C24B9 clone) or the glycosylated extracellu-
lar epitope (i.e. the AC133 or W6B3C1)[44]. In this study we used the W6B3C1 clone, which
we previously have identified as an antibody clone labeling both membrane and cytoplasm of
tumor cells [44–46]. Most importantly this clone has been shown to be associated with asym-
metric tumor cell division, which is a hallmark of cancer stem cells [44, 47].
For Ki-67, the expression was slightly reduced in the periphery for all gliomas together.
Comparing the expression between grades, we did however observe a significant increase in the
Ki-67 expression in the core area from grade II to grade IV (Fig 3E). There was no significant
increase in expression of other markers together with tumor grade. This is in line with the
results from Sabit et al. [36] who used a similar mIDH1 double fluorescence approach (11
patients). However Sabit et al. did not use a computer-based algorithm, but two independent
neuropathologists to evaluate the tumor samples. The slightly reduced proliferation in the
periphery suggest that both temozolomide and radiation therapy is less efficient in these areas
compared to central tumor areas. This may explain why Ki-67 labeling indices found and used
in prognostic studies, where tumor bulk is removed, not have shown to have a prognostic value
predicted [48, 49]. In most patients, it is in fact the tumor periphery left in the patients, which
receives temozolomide and radiation therapy.
The Ki-67 labeling index correlated with survival should ideally be the index obtained from
the periphery.
For MGMT a similar distribution in the core and periphery was found for all tumor grades
and subtypes. This means that resistance to temozolomide was preserved in migrating tumor
cells. Strategies aiming at reducing chemoresistance by compromising the MGMT enzyme are
therefore highly relevant for migrating glioma cells and new drugs added to temozolomide
should also reach these tumor cells to be efficient.

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 10 / 16

 Glioma Tumor Periphery Display Stem Cell

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 11 / 16

 Glioma Tumor Periphery Display Stem Cell

Fig 5. Double immunofluorescence staining of core and periphery in orthotopic model with five different patient derived GBM spheroid
cultures. Tissues were stained with Dapi (blue), Vimentin (green) and Sox-2 (red), Bmi-1 (red) and Glut-3 (red). The software-based classifier is
shown in the right column. The classifier illustrates tumor cells co-expressing markers of interest in yellow and tumor cells not co-expressing markers
of interest in green. The fluorescence stainings were quantified in both central part and periphery for Sox-2 (G), Bmi-1 (H) and Glut-3 (I). Statistical
comparison was performed using student’s t-test. Scalebar: 200μm.
doi:10.1371/journal.pone.0155106.g005

In the GBM xenograft mouse model the results showed a similar expression of stem cell
markers in core and periphery for all stem cell markers. Before implantation in mice brains,
the spheroid cultures were grown in neural stem cell medium, which favor a more stem cell
like phenotype. This might explain the preservation of the stem cell phenotype in migrating
xenograft tumor cells compared to the reduced expression of stem cell markers found in
migrating glioma cells in patients. However, to a large extend the overall levels in tumor core in
patients and mice were similar, except for musashi-1 and sox-2 which were expressed at much
higher or higher levels in both core and periphery in patients GBMs compared to GBM xeno-
grafts. Reasons for this may be differences e.g. in tumor microenvironment. Hypoxia is pro-
nounced in patient high grade gliomas and known to increase expression of stem cell markers
[50–52], but necrosis is not found in our GBM xenografts suggesting that this micro-environ-
mental stimulus is less pronounced patients. Another explanation of differences between
results obtained in patient tumors versus GBM xenografts is IDH1 status. The patient tumors
included in this study were all IDH1 mutated whereas all GBM xenografts were obtained from
IDH1 wild type GBMs. Accordingly, CD133 has been found to be expressed at higher levels in
IDH1 wild type GBMs compared to IDH1 mutated GBMs [53]. However, another marker of

Fig 6. Double immunofluorescence staining of core and periphery in orthotopic model with five different patient derived GBM spheroid cultures.
Tissues were stained with Dapi (blue), Vimentin (green) and Ki-67 (red). The software-based classifier is shown in the right column. The classifier illustrates
tumor cells co-expressing markers of interest in yellow and tumor cells not co-expressing markers of interest in green. The fluorescence stainings were
quantified in both core and periphery for Ki-67 (C). Statistical comparison was performed using student’s t-test. Scalebar: 200μm.
doi:10.1371/journal.pone.0155106.g006

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 12 / 16

 Glioma Tumor Periphery Display Stem Cell

tumor aggressiveness–Ki-67 appeared to be expressed at similar levels in patient GBMs and

GBM xenografts. The expression of Ki-67 in the GBM xenograft model showed a slightly
reduced but not significant reduction in the periphery (n = 5), whereas the slightly reduced
level was significant in patient GBMs (n = 26). In general, our results indicate that the orthoto-
pic model to some degree mimics the migration scenario seen in patients Studies focusing on
characteristics of migrating glioma cells and how they should be targeted could therefore to
some extend be done in the orthotopic model.
In conclusion tumor cells in the periphery of patient gliomas have a stem cell phenotype,
although it is less pronounced than in the tumor core. Novel therapies aiming at preventing
recurrence should therefore take tumor stemness into account. In addition, known resistance
factors like MGMT also seem to be preserved in the periphery of patient gliomas. Migrating
cells in orthotopic glioblastoma xenografts preserve expression and stem cell markers. The
orthotopic model therefore has a promising translational potential.

Supporting Information
S1 Fig. Double immunofluorescence staining of core and periphery of patient gliomas. Dif-
fuse astrocytoma (AA-BE), oligo-astrocytoma (CA-DE), oligodendroglioma (EA-FE), anaplas-
tic astrocytoma (GA-HE) and anaplastic oligo-astrocytoma (IA-JE) were stained with Dapi
(blue), IDH1 (green) and Sox-2 (red). The software-based classifier is shown in the right col-
umn. The classifier illustrates tumor cells co-expressing Sox-2 in yellow and tumor cells not
co-expressing Sox-2 in green. Scalebar: 200μm.
(TIF)

Acknowledgments
We would like to thank Helle Wohlleben and Tanja Dreehsen Højgaard for assistance with the
immunohistochemical and fluorescence staining.

Author Contributions
Conceived and designed the experiments: SM FRP BWK. Performed the experiments: SM
SAP. Analyzed the data: SM RHD BWK. Contributed reagents/materials/analysis tools: SM
FRP BWK. Wrote the paper: SM SAP RHD FRP SH BWK.

References
1. Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ, Janzer RC, et al. Effects of radiother-
apy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma
in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. The lancet oncology. 2009;
10(5):459–66. Epub 2009/03/10. doi: 10.1016/s1470-2045(09)70025-7 PMID: 19269895.
2. Giese A, Westphal M. Glioma invasion in the central nervous system. Neurosurgery. 1996; 39(2):235–
50; discussion 50–2. Epub 1996/08/01. PMID: 8832660.
3. Huang Z, Cheng L, Guryanova OA, Wu Q, Bao S. Cancer stem cells in glioblastoma—molecular signal-
ing and therapeutic targeting. Protein & cell. 2010; 1(7):638–55. Epub 2011/01/05. doi: 10.1007/
s13238-010-0078-y PMID: 21203936.
4. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, et al. Identification of human brain tumour
initiating cells. Nature. 2004; 432(7015):396–401. Epub 2004/11/19. doi: 10.1038/nature03128 PMID:
15549107.
5. Dean M, Fojo T, Bates S. Tumour stem cells and drug resistance. Nature reviews Cancer. 2005; 5
(4):275–84. Epub 2005/04/02. doi: 10.1038/nrc1590 PMID: 15803154.
6. Capper D, Zentgraf H, Balss J, Hartmann C, von Deimling A. Monoclonal antibody specific for IDH1
R132H mutation. Acta neuropathologica. 2009; 118(5):599–601. Epub 2009/10/03. doi: 10.1007/
s00401-009-0595-z PMID: 19798509.

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 13 / 16

 Glioma Tumor Periphery Display Stem Cell

7. Capper D, Sahm F, Hartmann C, Meyermann R, von Deimling A, Schittenhelm J. Application of mutant

IDH1 antibody to differentiate diffuse glioma from nonneoplastic central nervous system lesions and
therapy-induced changes. The American journal of surgical pathology. 2010; 34(8):1199–204. Epub
2010/07/28. doi: 10.1097/PAS.0b013e3181e7740d PMID: 20661018.
8. von Deimling A, Korshunov A, Hartmann C. The next generation of glioma biomarkers: MGMT methyla-
tion, BRAF fusions and IDH1 mutations. Brain pathology (Zurich, Switzerland). 2011; 21(1):74–87.
Epub 2010/12/07. doi: 10.1111/j.1750-3639.2010.00454.x PMID: 21129061.
9. Zheng PP, van der Weiden M, van der Spek PJ, Vincent AJ, Kros JM. Isocitrate dehydrogenase
1R132H mutation in microglia/macrophages in gliomas: Indication of a significant role of microglia/mac-
rophages in glial tumorigenesis. Cancer Biology & Therapy. 2012; 13(10):836–9. doi: 10.4161/cbt.
20836 PMID: 22785212; PubMed Central PMCID: PMCPmc3414408.
10. Joe YA, Hong YK, Chung DS, Yang YJ, Kang JK, Lee YS, et al. Inhibition of human malignant glioma
growth in vivo by human recombinant plasminogen kringles 1–3. International journal of cancer Journal
international du cancer. 1999; 82(5):694–9. Epub 1999/07/27. PMID: 10417767.
11. Huszthy PC, Daphu I, Niclou SP, Stieber D, Nigro JM, Sakariassen PO, et al. In vivo models of primary
brain tumors: pitfalls and perspectives. Neuro-oncology. 2012; 14(8):979–93. Epub 2012/06/09. doi:
10.1093/neuonc/nos135 PMID: 22679124; PubMed Central PMCID: PMCPmc3408261.
12. Dahlrot RH, Hansen S, Herrstedt J, Schroder HD, Hjelmborg J, Kristensen BW. Prognostic value of
Musashi-1 in gliomas. Journal of neuro-oncology. 2013; 115(3):453–61. Epub 2013/09/24. doi: 10.
1007/s11060-013-1246-8 PMID: 24057325.
13. Hermansen SK, Kristensen BW. MicroRNA biomarkers in glioblastoma. Journal of neuro-oncology.
2013; 114(1):13–23. Epub 2013/05/24. doi: 10.1007/s11060-013-1155-x PMID: 23700324.
14. Corti S, Nizzardo M, Nardini M, Donadoni C, Locatelli F, Papadimitriou D, et al. Isolation and characteri-
zation of murine neural stem/progenitor cells based on Prominin-1 expression. Experimental neurology.
2007; 205(2):547–62. Epub 2007/05/01. doi: 10.1016/j.expneurol.2007.03.021 PMID: 17466977.
15. Lee A, Kessler JD, Read TA, Kaiser C, Corbeil D, Huttner WB, et al. Isolation of neural stem cells from
the postnatal cerebellum. Nature neuroscience. 2005; 8(6):723–9. Epub 2005/05/24. doi: 10.1038/
nn1473 PMID: 15908947; PubMed Central PMCID: PMCPmc2377345.
16. Uchida N, Buck DW, He D, Reitsma MJ, Masek M, Phan TV, et al. Direct isolation of human central ner-
vous system stem cells. Proceedings of the National Academy of Sciences of the United States of
America. 2000; 97(26):14720–5. Epub 2000/12/20. doi: 10.1073/pnas.97.26.14720 PMID: 11121071;
PubMed Central PMCID: PMCPmc18985.
17. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, et al. Identification of a cancer stem
cell in human brain tumors. Cancer research. 2003; 63(18):5821–8. Epub 2003/10/03. PMID:
14522905.
18. Dahlrot RH, Hermansen SK, Hansen S, Kristensen BW. What is the clinical value of cancer stem cell
markers in gliomas? International journal of clinical and experimental pathology. 2013; 6(3):334–48.
Epub 2013/02/16. PMID: 23412423; PubMed Central PMCID: PMCPmc3563206.
19. Arai H, Ikota H, Sugawara K, Nobusawa S, Hirato J, Nakazato Y. Nestin expression in brain tumors: its
utility for pathological diagnosis and correlation with the prognosis of high-grade gliomas. Brain tumor
pathology. 2012; 29(3):160–7. Epub 2012/02/22. doi: 10.1007/s10014-012-0081-5 PMID: 22350668.
20. Dahlstrand J, Collins VP, Lendahl U. Expression of the class VI intermediate filament nestin in human
central nervous system tumors. Cancer research. 1992; 52(19):5334–41. Epub 1992/10/01. PMID:
1382841.
21. Ehrmann J, Kolar Z, Mokry J. Nestin as a diagnostic and prognostic marker: immunohistochemical
analysis of its expression in different tumours. Journal of clinical pathology. 2005; 58(2):222–3. Epub
2005/01/29. doi: 10.1136/jcp.2004.021238 PMID: 15677549; PubMed Central PMCID:
PMCPmc1770570.
22. Kanamori M, Kumabe T, Sonoda Y, Nishino Y, Watanabe M, Tominaga T. Predictive factors for overall
and progression-free survival, and dissemination in oligodendroglial tumors. Journal of neuro-oncology.
2009; 93(2):219–28. Epub 2008/12/23. doi: 10.1007/s11060-008-9762-7 PMID: 19099201.
23. Maderna E, Salmaggi A, Calatozzolo C, Limido L, Pollo B. Nestin, PDGFRbeta, CXCL12 and VEGF in
glioma patients: different profiles of (pro-angiogenic) molecule expression are related with tumor grade
and may provide prognostic information. Cancer Biology & Therapy. 2007; 6(7):1018–24. Epub 2007/
07/06. PMID: 17611402.
24. Kim KJ, Lee KH, Kim HS, Moon KS, Jung TY, Jung S, et al. The presence of stem cell marker-express-
ing cells is not prognostically significant in glioblastomas. Neuropathology: official journal of the Japa-
nese Society of Neuropathology. 2011; 31(5):494–502. Epub 2011/01/29. doi: 10.1111/j.1440-1789.
2010.01194.x PMID: 21269333.

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 14 / 16

 Glioma Tumor Periphery Display Stem Cell

25. Kanemura Y, Mori K, Sakakibara S, Fujikawa H, Hayashi H, Nakano A, et al. Musashi1, an evolution-
arily conserved neural RNA-binding protein, is a versatile marker of human glioma cells in determining
their cellular origin, malignancy, and proliferative activity. Differentiation; research in biological diversity.
2001; 68(2–3):141–52. Epub 2001/11/01. PMID: 11686236.
26. Toda M, Iizuka Y, Yu W, Imai T, Ikeda E, Yoshida K, et al. Expression of the neural RNA-binding protein
Musashi1 in human gliomas. Glia. 2001; 34(1):1–7. Epub 2001/04/03. PMID: 11284014.
27. Ma YH, Mentlein R, Knerlich F, Kruse ML, Mehdorn HM, Held-Feindt J. Expression of stem cell markers
in human astrocytomas of different WHO grades. Journal of neuro-oncology. 2008; 86(1):31–45. Epub
2007/07/06. doi: 10.1007/s11060-007-9439-7 PMID: 17611714.
28. Strojnik T, Rosland GV, Sakariassen PO, Kavalar R, Lah T. Neural stem cell markers, nestin and musa-
shi proteins, in the progression of human glioma: correlation of nestin with prognosis of patient survival.
Surgical neurology. 2007; 68(2):133–43; discussion 43–4. Epub 2007/06/01. doi: 10.1016/j.surneu.
2006.10.050 PMID: 17537489.
29. Phi JH, Park SH, Kim SK, Paek SH, Kim JH, Lee YJ, et al. Sox2 expression in brain tumors: a reflection
of the neuroglial differentiation pathway. The American journal of surgical pathology. 2008; 32(1):103–
12. Epub 2007/12/29. doi: 10.1097/PAS.0b013e31812f6ba6 PMID: 18162777.
30. Berezovsky AD, Poisson LM, Cherba D, Webb CP, Transou AD, Lemke NW, et al. Sox2 promotes
malignancy in glioblastoma by regulating plasticity and astrocytic differentiation. Neoplasia. 2014; 16
(3):193–206,.e19-25. Epub 2014/04/15. doi: 10.1016/j.neo.2014.03.006 PMID: 24726753; PubMed
Central PMCID: PMCPmc4094829.
31. Tam WL, Ng HH. Sox2: masterminding the root of cancer. Cancer cell. 2014; 26(1):3–5. Epub 2014/07/
16. doi: 10.1016/j.ccr.2014.06.024 PMID: 25026204.
32. Yan GN, Yang L, Lv YF, Shi Y, Shen LL, Yao XH, et al. Endothelial cells promote stem-like phenotype
of glioma cells through activating the Hedgehog pathway. The Journal of pathology. 2014; 234(1):11–
22. Epub 2014/03/08. doi: 10.1002/path.4349 PMID: 24604164; PubMed Central PMCID:
PMCPmc4260128.
33. Dong Y, Han Q, Zou Y, Deng Z, Lu X, Wang X, et al. Long-term exposure to imatinib reduced cancer
stem cell ability through induction of cell differentiation via activation of MAPK signaling in glioblastoma
cells. Molecular and cellular biochemistry. 2012; 370(1–2):89–102. Epub 2012/07/26. doi: 10.1007/
s11010-012-1401-0 PMID: 22829019.
34. Flavahan WA, Wu Q, Hitomi M, Rahim N, Kim Y, Sloan AE, et al. Brain tumor initiating cells adapt to
restricted nutrition through preferential glucose uptake. Nature neuroscience. 2013; 16(10):1373–82.
Epub 2013/09/03. doi: 10.1038/nn.3510 PMID: 23995067; PubMed Central PMCID:
PMCPmc3930177.
35. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al. Radiotherapy plus con-
comitant and adjuvant temozolomide for glioblastoma. The New England journal of medicine. 2005;
352(10):987–96. Epub 2005/03/11. doi: 10.1056/NEJMoa043330 PMID: 15758009.
36. Sabit H, Nakada M, Furuta T, Watanabe T, Hayashi Y, Sato H, et al. Characterizing invading glioma
cells based on IDH1-R132H and Ki-67 immunofluorescence. Brain tumor pathology. 2014. Epub 2014/
01/05. doi: 10.1007/s10014-013-0172-y PMID: 24384677.
37. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, et al. The 2007 WHO Classifica-
tion of Tumours of the Central Nervous System. Acta neuropathologica. 2007; 114(2):97–109. doi: 10.
1007/s00401-007-0243-4 PMID: 17618441; PubMed Central PMCID: PMCPmc1929165.
38. Jensen SS, Aaberg-Jessen C, Andersen C, Schroder HD, Kristensen BW. Glioma spheroids obtained
via ultrasonic aspiration are viable and express stem cell markers: a new tissue resource for glioma
research. Neurosurgery. 2013; 73(5):868–86; discussion 86. Epub 2013/07/28. doi: 10.1227/neu.
0000000000000118 PMID: 23887192.
39. Cabrera MC, Hollingsworth RE, Hurt EM. Cancer stem cell plasticity and tumor hierarchy. World journal
of stem cells. 2015; 7(1):27–36. doi: 10.4252/wjsc.v7.i1.27 PMID: 25621103; PubMed Central PMCID:
PMCPMC4300934.
40. Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, et al. Glioma stem cells promote radioresis-
tance by preferential activation of the DNA damage response. Nature. 2006; 444(7120):756–60. Epub
2006/10/20. doi: 10.1038/nature05236 PMID: 17051156.
41. Eyupoglu IY, Buchfelder M, Savaskan NE. Surgical resection of malignant gliomas-role in optimizing
patient outcome. Nature reviews Neurology. 2013; 9(3):141–51. Epub 2013/01/30. doi: 10.1038/
nrneurol.2012.279 PMID: 23358480.
42. McGirt MJ, Chaichana KL, Attenello FJ, Weingart JD, Than K, Burger PC, et al. Extent of surgical resec-
tion is independently associated with survival in patients with hemispheric infiltrating low-grade gliomas.
Neurosurgery. 2008; 63(4):700–7; author reply 7–8. doi: 10.1227/01.NEU.0000325729.41085.73
PMID: 18981880.

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 15 / 16

 Glioma Tumor Periphery Display Stem Cell

43. Zeppernick F, Ahmadi R, Campos B, Dictus C, Helmke BM, Becker N, et al. Stem cell marker CD133
affects clinical outcome in glioma patients. Clinical cancer research: an official journal of the American
Association for Cancer Research. 2008; 14(1):123–9. doi: 10.1158/1078-0432.CCR-07-0932 PMID:
18172261.
44. Hermansen SK, Christensen KG, Jensen SS, Kristensen BW. Inconsistent immunohistochemical
expression patterns of four different CD133 antibody clones in glioblastoma. The journal of histochem-
istry and cytochemistry: official journal of the Histochemistry Society. 2011; 59(4):391–407. Epub 2011/
03/18. doi: 10.1369/0022155411400867 PMID: 21411810; PubMed Central PMCID:
PMCPmc3201141.
45. Dahlrot RH, Hansen S, Jensen SS, Schroder HD, Hjelmborg J, Kristensen BW. Clinical value of CD133
and nestin in patients with glioma: a population-based study. International journal of clinical and experi-
mental pathology. 2014; 7(7):3739–51. Epub 2014/08/15. PMID: 25120750; PubMed Central PMCID:
PMCPmc4128985.
46. Christensen K, Schroder HD, Kristensen BW. CD133 identifies perivascular niches in grade II-IV astro-
cytomas. Journal of neuro-oncology. 2008; 90(2):157–70. Epub 2008/07/10. doi: 10.1007/s11060-008-
9648-8 PMID: 18612800.
47. Lathia JD, Hitomi M, Gallagher J, Gadani SP, Adkins J, Vasanji A, et al. Distribution of CD133 reveals
glioma stem cells self-renew through symmetric and asymmetric cell divisions. Cell death & disease.
2011; 2:e200. doi: 10.1038/cddis.2011.80 PMID: 21881602; PubMed Central PMCID:
PMCPMC3186899.
48. Lind-Landstrom T, Varughese RK, Sundstrom S, Torp SH. Expression and clinical significance of the
proliferation marker minichromosome maintenance protein 2 (Mcm2) in diffuse astrocytomas WHO
grade II. Diagn Pathol. 2013; 8:67. doi: 10.1186/1746-1596-8-67 PMID: 23618321; PubMed Central
PMCID: PMCPMC3648352.
49. Ambroise MM, Khosla C, Ghosh M, Mallikarjuna VS, Annapurneswari S. Practical value of MIB-1 index
in predicting behavior of astrocytomas. Indian journal of pathology & microbiology. 2011; 54(3):520–5.
doi: 10.4103/0377-4929.85085 PMID: 21934213.
50. Kolenda J, Jensen SS, Aaberg-Jessen C, Christensen K, Andersen C, Brunner N, et al. Effects of hyp-
oxia on expression of a panel of stem cell and chemoresistance markers in glioblastoma-derived spher-
oids. Journal of neuro-oncology. 2011; 103(1):43–58. Epub 2010/09/14. doi: 10.1007/s11060-010-
0357-8 PMID: 20835751.
51. Heddleston JM, Wu Q, Rivera M, Minhas S, Lathia JD, Sloan AE, et al. Hypoxia-induced mixed-lineage
leukemia 1 regulates glioma stem cell tumorigenic potential. Cell Death Differ. 2012; 19(3):428–39. doi:
10.1038/cdd.2011.109 PMID: 21836617; PubMed Central PMCID: PMCPMC3229666.
52. Li Z, Bao S, Wu Q, Wang H, Eyler C, Sathornsumetee S, et al. Hypoxia-inducible factors regulate
tumorigenic capacity of glioma stem cells. Cancer cell. 2009; 15(6):501–13. doi: 10.1016/j.ccr.2009.03.
018 PMID: 19477429; PubMed Central PMCID: PMCPMC2693960.
53. Shin JH, Lee YS, Hong YK, Kang CS. Correlation between the prognostic value and the expression of
the stem cell marker CD133 and isocitrate dehydrogenase1 in glioblastomas. Journal of neuro-oncol-
ogy. 2013; 115(3):333–41. doi: 10.1007/s11060-013-1234-z PMID: 24129546.

PLOS ONE | DOI:10.1371/journal.pone.0155106 May 12, 2016 16 / 16