ISOLATION OF A NEW CELL POPULATION IN THE GLIOBLASTOMA MICROENVIRONMENT
Anne Clavreul1,2,3, Amandine Etcheverry4,5,6, Agnès Chassevent7, Véronique Quillien5,8, Tony Avril8, Marie-Lise
Jourdan9, Sophie Michalak10, Patrick François11, Jean-Luc Carré12, Jean Mosser4,5,6, Philippe Menei1,2,3
and the Grand Ouest Glioma Project network13
1
LUNAM Université, Ingénierie de la Vectorisation Particulaire, F-49933 Angers, France
2
INSERM, U646, Unit 646, F-49933 Angers, France
3
Département de Neurochirurgie, Centre Hospitalier Universitaire d'Angers, 4 rue Larrey, 49100 Angers, France
4
Plateforme Puces à ADN, Biogenouest®, 35043 Rennes, France
5
UMR 6061 CNRS, Université de Rennes 1, IFR 140, 35043, France
6
Service de Génétique Moléculaire et Génomique, Hôpital Pontchaillou, CHU, 35043 Rennes, France
7
Centre Régional de Lutte Contre le Cancer Paul Papin, 49100 Angers, France
8
Centre Régional de Lutte Contre le Cancer Eugène Marquis, 35042 Rennes, France
9
Laboratoire de Cancérologie, INSERM U921, CHU, 37044 Tours, France
10
Laboratoire Pathologie Cellulaire et Tissulaire, CHU, 49933 Angers, France
11
Service de Neurochirurgie, CHU, 37044 Tours, France
12
Groupe Gliome Brestois, Service de Biochimie et Biologie Moléculaire, UFR Médecine, CHU, 29238 Brest,
France (J.-L.C.)
13
Cancéropôle Grand Ouest, 44093 Nantes, France (G.O.G.P.)
Financial support: This work was supported by the French National Cancer Institute (INCa) and the Ligue
Départementale de lutte contre le cancer.
Corresponding author: Philippe Menei, Département de Neurochirurgie, CHU, 49933 Angers, France. Phone: +33
241 354822; Fax: +33 241 354508; E-mail: PhMenei@chu-angers.fr
1
Abstract
Glioblastoma (GB) is a highly infiltrative tumor recurring in 90% of cases within a few centimeters of the resection
cavity, even in cases of complete tumor resection and adjuvant chemo/radiotherapy. This observation highlights the
importance of understanding this special zone of brain tissue surrounding the tumor. It is becoming clear that the non
neoplastic stromal compartment of most solid cancers plays an active role in tumor proliferation, invasion and
metastasis. Very little information, other than that concerning angiogenesis and immune cells, has been collected for
stromal cells from GB. As part of a translational research program, we have isolated a new stromal cell population
surrounding GB by computer-guided stereotaxic biopsies and primary culture. We named these cells GB-associated
stromal cells (GASCs). GASCs are diploid, do not display the genomic alterations typical of GB cells and have
phenotypic and functional properties in common with the cancer-associated fibroblasts (CAFs) described in the
stroma of carcinomas. In particular, GASCs express markers associated with CAFs such as fibroblast surface protein
(FSP), alpha-smooth muscle actin (-SMA) and platelet derived growth factor receptor-beta (PDGFR).
Furthermore, GASCs have a molecular expression profile different from that of control stromal cells derived from
non-GB peripheral brain tissues. GASCs were also found to have tumor-promoting effects on glioma cells in vitro
and in vivo. The isolation of GASCs in a tumor of neuroepithelial origin was unexpected and further studies are
required to determine their potential as a target for anti-glioma treatment.
Running title: GASCs in the GB microenvironment
Keywords: Translational study, glioblastoma, microenvironment, cancer-associated fibroblasts, primary cultures.
Abbreviations: CAFs, cancer-associated fibroblasts; DMEM, Dulbecco’s modified Eagles’ medium; FCA,
absolute fold-change; GASCs, GB-associated stromal cells; GB, glioblastoma; HABS, human AB serum; hMSCs,
human mesenchymal stem cells; IZ, interface zone; NZ, necrotic zone; PBZ, peripheral brain zone PDGFR,
platelet derived growth factor receptor-beta; Q-PCR, real-time quantitative PCR; -SMA, alpha-smooth muscle
actin; TZ, tumor zone
2
Introduction
Glioblastoma (GB) is the most common and deadliest malignant primary brain tumor. GB accounted for about 20%
of the estimated 44,500 new primary brain tumors diagnosed in the United States in 2005 (CBTRUS 2005). On MRI
scans, GB often appears as a ring-enhancing lesion, with a central necrotic area, in an apparently well delimited
florid contrast-enhanced tumor surrounded by edematous brain parenchyma. The standard treatment is surgical
resection followed by fractionated radiotherapy, with concomitant and adjuvant chemotherapy with temozolomide
[1]. Despite this treatment, median survival is around 16 months and less than 10% of patients survive more than
five years [1]. GB generally recurs at the resection margin highlighting the importance of understanding this
peripheral tumor brain area [2-4].
Several recent studies have shown that the progression of malignant tumors is not solely dependent on the cellautonomous properties of the cancer cells, and may be influenced by the tumor stroma. The concept that cancer cells
may be controlled by their microenvironment has thus become increasingly important in cancer biology. Recent
studies have suggested an important role for macrophages, pericytes, endothelial cells and cancer-associated
fibroblasts (CAFs) in the growth and progression of tumors [5,6]. CAFs, which are also known as reactive
fibroblasts or myofibroblasts, are the most important cell type in the stroma of carcinomas [7-9]. Solid evidence has
been presented for synergy between CAFs and malignant epithelial cells in tumor initiation, proliferation and
invasion [7-9]. We now have some knowledge of the cancer stem cell biology of GB [10,11] but, other than for
angiogenesis and immune cells, very little is known about the stromal cells of GB [12-14].
A translational research program (the Grand-Ouest glioma project) has been initiated by the French National Cancer
Institute (INCa), with the aim of improving our understanding of the heterogeneity of GB cells. A comprehensive
network of university hospitals, research laboratories and platforms has participated in this program. This project
focused on the analysis, by imaging, cell biology, genomic and proteomic techniques, of the four areas of the GB,
from the core to the periphery: the necrotic zone (NZ), the tumor zone (TZ), the interface zone between the tumor
and the parenchyma (IZ) and the peripheral brain zone (PBZ).
A key finding of this project was the isolation of a new cell population in the periphery of GB. These cells, called
GB-associated stromal cells (GASCs), were found to have phenotype and functional properties in common with
CAFs. Their isolation in a tumor of neuroepithelial origin was unexpected and raises questions about the origin of
these cells.
3
Materials and Methods
Patients
The entire project was approved by the local ethical research committee (CPP Ouest II). All patients undergoing
surgery for presumed GB were pre-included and signed an informed consent form for participation in this study.
Patients were definitively included in the study when the diagnosis of GB (WHO 2007 classification [15]) was
confirmed by a central committee of neuropathologists.
Biopsies
The areas to be biopsied in the four zones (NZ, TZ, IZ and PBZ) were defined on preoperative T1 gadoliniumenhanced 3D MRI (Figure 1a). Stereotaxic biopsies were performed in the operating theater with computer-assisted
neurosurgery (BrainLAB®) (Figure 1b). Each biopsy specimen was separated in four sections for mirror histologic
analysis, culture, genomic and proteomic analysis and for inclusion in a tumor bank. For histologic analysis,
formalin-fixed, paraffin-embedded sections of the biopsy specimens were stained with hematoxylin-phloxin-saffron
(Figure 1c). We present here only the data for histologic analysis and for the culture of TZ and PBZ specimens.
Control samples were isolated from brain tissues obtained during other types of surgery, for meningioma (control
cells 1), epilepsy (control cells 2) and metastasis (control cells 3).
Cell culture
Before cell isolation, pia mater was resected and blood vessels visible to the naked eye were removed from TZ
biopsies. For PBZ biopsies, these large vessels and cortical zones with pia mater were absent. Biopsies were then
minced and mechanically dissociated in Dulbecco’s modified Eagles’ medium (DMEM) (Lonza, Verviers, Belgium).
Aliquots of the resulting suspension corresponding to samples before culture were stored at -80°C for further
analysis by real-time quantitative PCR (Q-PCR) and flow cytometry analysis of DNA. The remainder of the
suspension was used to seed T80 flasks (Nunc, Dominique Dutsher, Brumath, France) containing DMEM
supplemented with 10% human AB serum (HABS) (EFS, Lyon, France) and 1% antibiotics (Sigma-Aldrich, Saint
Quentin Fallavier, France). Cells were grown at 37°C in a humidified incubator, under an atmosphere containing 5%
CO2, and the medium was changed twice weekly. Primary cultures were used under passage 2 and 10 for TZ and
PBZ specimens respectively.
The A172 and U87MG GB cell lines were provided by ATCC (LGC Promochem, Molsheim, France). Cells were
4
maintained in the same medium as cell suspensions from GB biopsies, at 37°C, under an atmosphere containing 5%
CO2, until they reached 80 to 90% confluence.
Flow cytometry analysis of DNA
Vindelov’s protocol was used to stain DNA [16] and flow cytometry was performed with a FACScan flow cytometer
(BD Biosciences, Le Pont de Claix, France). DNA index (DI) and the percentage of cell populations identified on
the basis of different DNA contents were calculated with Modfit version 5.2 software (Verity Software House,
Topsham, Maine). The DNA diploid peak was located on DNA histograms, according to an external standardization
procedure using normal human lymphocytes and an internal standardization procedure using trout red blood cells.
DI was calculated as the ratio of the mean DNA content of tumor cells to the mean DNA content of diploid cells.
Cases with a DI value of 1 were classified as DNA-diploid and cases with DI values < or > 1 were classified as
aneuploid.
Phenotypic analysis by flow cytometry
The following primary antibodies were used for flow cytometry: CD34 (clone 563), CD45 (clone HI30), CD31
(clone WM59), CD227 (clone HMPV), CD14 (clone MOP9), CD90 (clone 5E10), CD29 (clone MAR4), CD44
(clone L178), CD73 (clone AD2) and platelet derived growth factor receptor-beta (PDGFR) (CD140b, clone 28D4)
were from BD Biosciences - CDH6 (clone 427909), CD105 (clone 166707) and alpha-smooth muscle actin (SMA, clone 1A4) were from R/D Systems (Lille, France) - glial fibrillary acidic protein (GFAP, clone 2.2B10), NG2
(clone 7.1) and fibroblast surface protein (FSP, clone 1B10) were from Zymed (AbCys, Paris, France), Chemicon
(Millipore, St-Quentin-en-Ynes, France) and Sigma-Aldrich respectively. Briefly, GASCs and control stromal cells
resuspended in PBS supplemented with 5% FCS and 0.02% NaN 3 were incubated on ice for one hour with test
antibodies or isotype control antibodies. For GFAP and -SMA, cells were permeabilized with the BD
cytofix/cytopermTM fixation/permeabilization kit (BD Biosciences) before the addition of the antibody. Following
incubation, the cells were washed and further stained with fluorescein isothiocyanate (FITC)-conjugated goat
F(ab’)2 anti-mouse immunoglobulin or anti-rat immunoglobulin (Dako, Trappes, France) for 30 minutes on ice. The
cells were then washed and fixed in 2% formaldehyde. The stained cells were analyzed in a FACScan flow
cytometer with CellQuest Software (BD Biosciences).
In vitro differentiation
Evaluation of the in vitro differentiation ability of GASCs into classical mesenchymal lineages was carried out as
5
previously described for human mesenchymal stem cells (hMSCs) [17]. Briefly, GASCs were treated with either
adipogenic differentiation medium supplemented with IBMX (0.5 mM), dexamethasone (1 µM), insulin (10 µg/ml),
and indomethacin (100 µM) (Sigma-Aldrich), or with osteogenic differentiation medium supplemented with
ascorbic acid (100 µM), -glycerophosphate (10 mM) (Sigma-Aldrich), and dexamethasone (10 nM). The
diffentiation medium was changed twice weekly until day 21. Adipogenic differentiation was assessed by the
cellular accumulation of neutral lipid vacuoles after cells were fixed with 4% paraformaldehyde (PFA, SigmaAldrich) pH 7.4 and stained with Oil Red O. For evaluation of osteogenic differentiation, cells were fixed with 70%
EtOH and stained with Alizarin Red to detect calcium deposits. HMSCs, kindly provided by Dr C. Montero-Menei
(INSERM U646, Angers, France), were used as positive controls.
DNA and RNA extraction
DNA and total RNA were isolated with the NucleoSpin Tissue and NucleoSpin RNAII Kits, respectively
(Macherey-Nagel, Hoerdt, France). DNA quality was assessed by electrophoresis in a 1% agarose gel. RNA integrity
(RNA Integrity Number ≥ 8) was confirmed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara,
California).
Array CGH profiling
Array CGH profiling was performed with the Human CGH 8  60K Microarray Kit (Agilent Technologies). DNA
from TZ primary cultures, GASCs and control stromal cells were labeled with cyanin 5 and two reference DNAs
(from single male or female subjects) were labeled with cyanin 3. Sample and reference DNAs were pooled and
hybridized on the arrays. Genomic profiles were computed and compared, with Nexus software (Biodiscovery).
DNA copy number alterations were identified with the Gain and Loss Analysis of DNA (GLAD) algorithm [18].
Thresholds for low-copy number gain and hemizygous deletion were set at +0.4 and -0.4, respectively. Thresholds
for large-amplitude events were set at +0.7 for amplification and -0.7 for homozygous deletion.
Gene expression profiling
Gene expression profiling was performed with the Agilent Whole Human Genome 4  44K Microarray Kit (Agilent
Technologies). RNAs extracted from GASCs and control stromal cells were labeled with cyanin 3 and hybridized
with the arrays. Data were log2-transformed, normalized (quantile normalization and baseline transformation) and
selected with GeneSpring GX software (Agilent Technologies). Probes were selected on the basis of their expression
level (intensity greater than 100 in at least one of the two conditions).
6
Q-PCR
Q-PCR was performed with the 7900HT Fast Real-Time PCR System (Applied Biosystems, Courtaboeuf, France),
according to manufacturer's protocol. Triplicate reactions were carried out for each data point. The relative
expression levels of genes were determined with SYBRTMGreen PCR Master Mix (CDH6, APOL6, RAB20, NRGN
and MYH11) and Taqman FAM-labeled probes (IL-13R2). GAPDH, B2M and HPRT were chosen as housekeeping
genes. All gene transcripts were quantified by the ∆∆Ct method as described by the manufacturer.
Functional analysis
In vitro proliferation assay
We assessed the effect of GASCs on glioma cell proliferation in vitro, using the non tumorigenic, aneuploid glioma
cell line A172, to distinguish tumor cells from diploid GASCs or control stromal cells. GASCs or control stromal
cells were seeded at 96  103 per well in 6-well plates and cultured in DMEM supplemented with 10% HABS and
1% antibiotics. After 24 hours, 96  103 A172 cells were added into the wells. The cells were cocultured for 3 days
and then harvested. Flow cytometry analysis of DNA was performed as described above. Cell cycle analysis was
carried out on A172 cells and the proliferation index (PI) of these cells was evaluated as S + G2M%. This
experiment was performed in triplicate, and two samples of GASCs and control stromal cells were analyzed.
In vivo implantation
We carried out implantation experiments to evaluate the contribution of GASCs to the growth of glioma cells in
vivo. Tumorigenic U87MG cells and non tumorigenic A172 cells were used. Swiss nude female mice (8-10 weeks
old) were obtained from Charles River Laboratories (L’Arbresle, France). The experiments were conducted
according to French Ministry of Agriculture regulations and the European Communities Council Directive of 24
November 1986 (86/609/EEC). The mice were housed in a controlled-temperature and controlled-humidity room,
with 12-hour light-dark cycles, under aseptic conditions. Mice were assigned to groups receiving subcutaneous
injections into the right flank of (a) U87MG cells alone (0.5  106 cells); (b) A172 cells alone (0.5  106 cells) (c)
GASCs alone (0.5  106 cells) (d) U87MG cells plus GASCs or control stromal cells (0.5  106 cells, ratio 1:1) (e)
A172 cells plus GASCs or control stromal cells (0.5  106 cells, ratio 1:1). Tumor volume within the flank was
determined by direct measurement with a digital caliper, twice weekly, and was calculated as follows: 0.5  (large
7
diameter)  (small diameter)2. The results are expressed as medians  SEM.
Statistical analysis
Results are expressed as means  SEM and medians  SEM for in vivo experiments. The significance of differences
was assessed with Student’s t tests. Differences were considered significant if P < .05. For gene expression profiling,
probes were considered to display significant differential expression if P < .005 and an absolute fold-change (FCA)
> 2.
Results
Primary cultures of GB biopsy specimens
Primary cultures of TZ and PBZ biopsy specimens were performed for 10 patients with de novo GB. The histology
of the biopsy specimens are presented in Table 1. All TZ biopsy specimens had the histologic features associated
with the tumor. Their corresponding PBZ biopsy specimens corresponded to apparently normal brain parenchyma
except for three cases where an infiltration of tumor cells was identified.
As described in our previous study [19], culture of TZ specimens in HABS-containing medium preserved the tumor
cell population of interest present in the initial biopsy during the two first passages as shown by the presence of
aneuploid groups and IL-13R2 mRNA expression and the decrease of cells with a DI =1 of which the normal cells
belong to (Figure 2a). Cells derived from TZ cultures were spindle-shaped or clumped into irregular spheroids, often
several layers thick (Figure 2c).
Culture of PBZ specimens led, in majority of cases (7/10), to a diploid cell population after two passages which did
not express IL-13R2 mRNA (Figure 2b). These diploid cells, which we called GASCs, were star-shaped and
unable to assembly into multiple-layered structures and could be maintained in culture under passage 10 (Figure 2d).
In the remaining PBZ cultures (3/10), in which tumor cells were present to a significant level in the biopsy specimen
before culture, these cells proliferated (data not shown). For comparison, peripheral brain tissues isolated during
other types of surgery (epilepsy, meningioma and metastasis) were cultured and gave rise to morphologically similar
diploid cells, which we referred to as control stromal cells (Figure 2e).
Characterization of GASCs
Array CGH analysis
8
In TZ cultures, the most frequent imbalances were: gain of chromosome 7 (up to 80%), loss of chromosome 9 (up to
40%) and chromosome 10 (up to 80%). No large genomic aberrations were observed in GASCs and control stromal
cells. Some copy number variation (CNV) imbalances, as listed in the Database of Genomic Variants, were
considered to be copy number polymorphisms and were not studied further (Figure 3a).
Gene expression profiling
The gene expression profiles for GASCs and control stromal cells were obtained with the Agilent Whole Human
Genome 4  44K Microarray Kit. We found that 24 of the 18,198 selected probes were differentially expressed
between GASCs and control stromal cells: 13 probes (11 genes) were overexpressed and 11 probes (11 genes) were
underexpressed in GASCs (Table 2). Hierarchical cluster analysis of the 18,198 selected probes (data not shown) or
the 24 differentially expressed probes clustered the samples into two distinct groups: one consisting of three GASC
samples and the other comprising the control stromal cells and one GASC sample (Figure 3b). APOL6
underexpression and RAB20, CDH6, MYH11 and NRGN overexpression were validated by Q-PCR (Figure 3c).
CDH6 overexpression was validated by flow cytometry (Figure 3d).
Phenotypic analysis
For definition of the phenotype of GASCs, a panel of markers associated to different cell types (CAFs, pericytes,
MSCs, epithelial cells, astrocytes, hematopoietic stem cells, macrophages and endothelial cells) was analyzed by
flow cytometry (Figure 4a). GASCs displayed expression for CAF (-SMA, PDGFR, FSP) and MSC (CD105,
CD44, CD73, CD29, CD90) associated markers. By contrast, GASCs were negative for endothelial cell (CD31),
macrophage (CD14), astrocyte (GFAP) and hematopietic stem cell (CD34, CD45) markers and expressed pericyte
(NG2) and epithelial cell (CD227) markers only weakly (Figure 4a). Control stromal cells had a similar antigenic
profile but significantly stronger expression of the CD29, CD44 and PDGFR markers was observed in GASCs
(Figure 4a).
The ability of GASCs to differentiate into the classical mesenchymal lineages was examined to determine if these
cells shared the properties of MSCs. GASCs were capable of differentiating into osteoblasts but not into adipocytes
(Figure 4b).
Functional analysis
Functional analysis was performed with two human cell lines (the aneuploid non tumorigenic A172 cell line and the
diploid tumorigenic U87MG cell line) and with two samples of GASCs (GASCs 2 and 3) and control stromal cells
9
(control cells 1 and 3) presented in Figure 3b. In vivo, the subcutaneous injection into nude mice of GASCs together
with the U87MG cell line resulted in significantly stronger tumor volume induction than the injection of U87MG
alone (Figure 5a). The injection of control stromal cells together with U87MG cells also increased tumor volume,
but this difference was not significantly greater than that observed with U87MG cells alone (Figure 5a). By contrast,
the subcutaneous injection into nude mice of GASCs alone, A172 cells alone and A172 cells together with GASCs
or control stromal cells did not result in the formation of tumors after five months. However, in vitro, the addition of
GASCs to the non tumorigenic A172 cells induced an increase in tumor cell proliferation which was significantly
stronger than observed after the addition of control stromal cells (Figure 5b). A172 cells did not promote growth of
GASCs or control stromal cells (data not shown).
Discussion
GB frequently recurs after therapy at the site of its initial treatment [2-4]. Recurrence is thought to result from the
proliferation of tumor cells infiltrating the brain tissue surrounding the tumor. Such cells have been isolated, by the
culture in a serum-free medium, of biopsy specimens taken from the brain tissue surrounding the resection cavity
[20]. These infiltrating tumor cells display the alterations typical of GB, but can be distinguished from tumor cells
isolated from the tumor core on the basis of their distinctive molecular marker profiles and responses to drug and
irradiation challenges in vitro [20].
Using a different type of culture medium, we have isolated, from the brain tissue surrounding GB, another
unexpected population of cells from the stroma which we have named GASCs.
Phenotypic analysis of GASCs
GASCs are diploid and do not display the genomic alterations typical of GB cells. Such diploid cells have been
described before, following the culture of normal or neoplastic brain tissues from human adults. However, little
attention was paid to these cells, which were considered to be glial-like cells or contaminating mesenchymal cells
from the meninges or vascular connective tissue [21-23]. We demonstrate here, that GASCs, very closely resemble
the CAFs described in carcinomas.
Alpha-SMA, FSP and PDGFR are the most commonly used markers for CAF identification, although there is not
marker absolutely specific to these cells. GASCs expressed these three markers and, like CAFs, they were negative
for the macrophage marker CD14 and the endothelial marker CD31. A similar cell population was isolated from
10
non-GB peripheral brain tissues. Comparison of the gene expression profiles of GASCs and of these cells showed
that only 24 of the 18,198 selected probes were differentially expressed between these two cell types. This small
difference may be due to the control stromal cells used being derived from cultures of pathological rather than
healthy brain tissues, which would have been difficult to obtain. Nevertheless, GASCs displayed significantly
stronger expression of the PDGFR marker than control stromal cells. Furthermore, the MYH11 gene, which
encodes a smooth-muscle myosin heavy chain, was significantly upregulated in GASCs. The gene encoding another
smooth muscle-specific protein, desmin, also seemed to be more strongly expressed in GASCs (data not shown).
These markers are used to distinguish between CAFs and normal fibroblasts or smooth muscle cells [9,24].
Functional analysis of GASCs
Although there is no CAF specific markers, it is well recognized that CAFs contribute to tumorigenesis [7-9]. In this
study, we observed that among the markers and genes overexpressed in GASCs, several are involved in tumor
progression, invasion, treatment resistance and/or brain angiogenesis, such as PDGFR [25], Rab20 [26], AKAP12
[27], DHFR [28], TLE3 [29], CDH6 [30,31], CD29 and CD44 [32]. Consistent with this observation, we showed
that GASCs had tumor-promoting effects in vitro on the non tumorigenic cell line A172, and that these effects were
significantly stronger than those of control stromal cells. However, unlike CAFs, which can induce the
transformation of non tumorigenic prostatic epithelial cell lines [33], GASCs did not trigger the malignant
transformation of the A172 cell line in vivo. Nevertheless, the injection of GASCs together with tumorigenic
U87MG cells into nude mice resulted in the induction of tumors larger than those induced by injection of the glioma
cell line alone or together with control stromal cells. Thus, GASCs have functional attributes similar to those of
CAFs.
Origin of the tumor-promoting phenotype of GASCs
As reported for CAFs [34], the tumor-promoting phenotype of GASCs may persist even if they are passaged in vitro
without continuous exposure to cancer cells. GASCs have a normal CGH array profile. Several studies have
suggested that the tumor-promoting phenotype of CAFs may be due to genetic alterations (mutations, loss of
heterozygosity - LOH). However, conflicting findings have arisen and counterarguments have been raised [35].
Recent findings indicate that epigenetic changes constitute a potential mechanism by which CAFs may acquire their
novel phenotype [35]. These mechanisms may also apply to GASCs.
11
Heterogeneity of GASCs
One of the GASC samples clustered with control stromal cells. In this case, the GASCs were derived from biopsies
in a 33-year-old patient. The median age at diagnosis in primary GB is 64 years [36] and the median age of the 24
patients studied was 68 years. GB in young adults under the age of 40 years is unusual and probably represents a
particular subtype of GB, although no specific cytogenetic or molecular markers for this subtype have yet been
isolated. CAFs are known to differ within and between histologic types of tumor and this CAF heterogeneity is of
prognostic value and may predict response to treatment [37]. This may also be the case for GASCs and the
identification of biologically relevant GASC subsets in GB is therefore warranted.
Origins of GASCs
Local fibroblasts were originally considered to be the main source of CAFs. However, recent studies have identified
other sources, such as endothelial cells, epithelial cells and bone marrow MSCs [35].
The brain parenchyma contains a very low proportion of fibroblasts and epithelial cells. GASCs may result from the
transition of GB tumor cells to mesenchymal cells, according to the epithelial-mesenchymal transition (EMT)
described for carcinomas. EMT and MET (mesenchymal-epithelial transition) are key regulators of cellular
plasticity in carcinomas and play a key role in treatment resistance, tumor recurrence and metastatic progression
[38]. Recent studies have shown that GBs have mesenchymal properties [39-44]. However, the normal CGH array
profile of GASCs, contrasting with the abnormal profile of GB tumor cells, is not consistent with this hypothesis.
Alternatively, GASCs may be derived from an endothelial-mesenchymal transformation at the invasive edge of the
tumor, from pericytes, vascular smooth muscle cells or from MSCs. This latter hypothesis is supported by many
studies showing the tropism of MSCs for glioma [17,45,46]. Furthermore, cells resembling bone marrow-MSCs
were recently isolated in normal mouse brain and in fresh glioma specimens obtained at surgery [47]. Moreover,
Mishra et al. [48] have shown that conditioned medium from U87MG cells can induce the transformation of MSCs
to a CAF phenotype. We showed that GASCs expressed several markers characteristic of MSCs, such as CD90,
CD105 and CD73. We also showed that GASCs had the potential to differentiate into osteoblasts but not in
adipocytes in vitro, suggesting that these cells have several properties in common with MSCs.
Conclusion
In this translational study, we have isolated, from brain tissue surrounding the GB, a new cell population with
12
similar properties to CAFs: GASCs. Like CAFs, GASCs express mesenchymal markers and have tumor-promoting
effects. However, these cells differ from CAFs. In particular, whereas CAFs constitute 50 to 70% of the tumor mass
in pancreatic, gastric and breast cancers, GASCs make up a smaller proportion of the cells in the GB
microenvironment. Immunohistochemical staining for -SMA, one of the markers of GASCs, in the brain tissue
surrounding GB showed that -SMA-positive cells were located exclusively within blood vessels suggesting that
GASCs may be a component of the vascular niche (data not shown).
Further studies are required to determine the origin of GASCs, the molecular mechanisms involved in their tumorpromoting functions and their potential as a target for anti-glioma treatment.
Acknowledgments: We thank neurosurgeons (Eric Lioret, CHU, Tours), Phong Dam Hieu (CHU, Brest),
Abderrahmane Hamlat (CHU, Rennes) for supplying tumor samples and the central committee of neuropathologists
(Delphine Loussouarn, Stephan Saikali, Sophie Michalak and Isabelle Quintin). We also thank Emilie Munnier
(Service d’Oncologie Médicale, CHRU, Tours), Jérôme Cayon (INSERM U646), Anne Bordron (Laboratoire de
Thérapie Cellulaire, CHU Morvan, Brest), Lucie Karayan-Tapon (Université Poitiers EA3805, CHU, Poitiers) and
Lisa Oliver (INSERM U601, Nantes) for their help in the preparation of cell samples. We thank Pierre Legras and
Jérôme Roux (Service Commun d’Animalerie Hospitalo-Universitaire, Angers), Catherine Guillet (Service Commun
de Cytométrie et d’Analyse Nucléotidique, Angers), Marina Denyset and the members of the Cancéropole Grand
Ouest for allowing us to use their facilities.
References
1.
Stupp R, Hegi ME, Mason WP et al (2009) Effects of radiotherapy with concomitant and adjuvant
temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5year analysis of the EORTC-NCIC trial. Lancet Oncol 10: 459-466
2.
Giese A, Kucinski T, Knopp U et al (2004) Pattern of recurrence following local chemotherapy with
biodegradable carmustine (BCNU) implants in patients with glioblastoma. J Neurooncol 66: 351-360
3.
Lefranc F, Brotchi J, Kiss R (2005) Possible future issues in the treatment of glioblastomas: special
emphasis on cell migration and the resistance of migrating glioblastoma cells to apoptosis. J Clin Oncol 23:
2411-2422
13
4.
Liang BC, Thornton AF, Jr., Sandler HM et al (1991) Malignant astrocytomas: focal tumor recurrence after
focal external beam radiation therapy. J Neurosurg 75: 559-563
5.
Tlsty TD, Coussens LM (2006) Tumor stroma and regulation of cancer development. Annu Rev Pathol 1:
119-150
6.
Ostman A, Augsten M (2009) Cancer-associated fibroblasts and tumor growth--bystanders turning into key
players. Curr Opin Genet Dev 19: 67-73
7.
Bhowmick NA, Neilson EG, Moses HL (2004) Stromal fibroblasts in cancer initiation and progression.
Nature 432: 332-337
8.
Shimoda M, Mellody KT, Orimo A (2010) Carcinoma-associated fibroblasts are a rate-limiting determinant
for tumour progression. Semin Cell Dev Biol 21: 19-25
9.
Franco OE, Shaw AK, Strand DW et al (2010) Cancer associated fibroblasts in cancer pathogenesis. Semin
Cell Dev Biol 21: 33-39
10.
Xie Z (2009) Brain tumor stem cells. Neurochem Res 34: 2055-2066
11.
Chen R, Nishimura MC, Bumbaca SM et al (2010) A hierarchy of self-renewing tumor-initiating cell types
in glioblastoma. Cancer Cell 17: 362-375
12.
Jain RK, di Tomaso E, Duda DG et al (2007) Angiogenesis in brain tumours. Nat Rev Neurosci 8: 610-622
13.
Yang I, Han SJ, Kaur G et al (2010) The role of microglia in central nervous system immunity and glioma
immunology. J Clin Neurosci 17: 6-10
14.
Sonabend AM, Rolle CE, Lesniak MS (2008) The role of regulatory T cells in malignant glioma.
Anticancer Res 28: 1143-1150
15.
Louis DN, Ohgaki H, Wiestler OD et al (2007) The 2007 WHO classification of tumours of the central
nervous system. Acta Neuropathol 114: 97-109
16.
Vindelov LL, Christensen IJ, Nissen NI (1983) A detergent-trypsin method for the preparation of nuclei for
flow cytometric DNA analysis. Cytometry 3: 323-327
17.
Roger M, Clavreul A, Venier-Julienne MC et al (2010) Mesenchymal stem cells as cellular vehicles for
delivery of nanoparticles to brain tumors. Biomaterials 31: 8393-8401
18.
Hupe P, Stransky N, Thiery JP et al (2004) Analysis of array CGH data: from signal ratio to gain and loss of
DNA regions. Bioinformatics 20: 3413-3422
14
19.
Clavreul A, Jean I, Preisser L et al (2009) Human glioma cell culture: two FCS-free media could be
recommended for clinical use in immunotherapy. In Vitro Cell Dev Biol Anim 45: 500-511
20.
Glas M, Rath BH, Simon M et al (2010) Residual tumor cells are unique cellular targets in glioblastoma.
Ann Neurol 68: 264-269
21.
Westermark B, Ponten J, Hugosson R (1973) Determinants for the establishment of permanent tissue
culture lines from human gliomas. Acta Pathol Microbiol Scand A 81: 791-805
22.
Rutka JT, Giblin JR, Dougherty DY et al (1987) Establishment and characterization of five cell lines
derived from human malignant gliomas. Acta Neuropathol 75: 92-103
23.
Gibbons HM, Hughes SM, Van Roon-Mom W et al (2007) Cellular composition of human glial cultures
from adult biopsy brain tissue. J Neurosci Methods 166: 89-98
24.
Orimo A, Weinberg RA (2007) Heterogeneity of stromal fibroblasts in tumors. Cancer Biol Ther 6: 618-619
25.
Raica M, Cimpean, A.M. (2010) Platelet-derived growth factor (PDGF)/PDGF receptors (PDGFR) axis as
target for antitumor and antiangiogenic therapy. Pharmaceuticals 3: 572-599
26.
Amillet JM, Ferbus D, Real FX et al (2006) Characterization of human Rab20 overexpressed in exocrine
pancreatic carcinoma. Hum Pathol 37: 256-263
27.
Lee HS, Han J, Bai HJ et al (2009) Brain angiogenesis in developmental and pathological processes:
regulation, molecular and cellular communication at the neurovascular interface. Febs J 276: 4622-4635
28.
Idbaih A, Carvalho Silva R, Criniere E et al (2008) Genomic changes in progression of low-grade gliomas.
J Neurooncol 90: 133-140
29.
van Agthoven T, Sieuwerts AM, Meijer-van Gelder ME et al (2009) Relevance of breast cancer
antiestrogen resistance genes in human breast cancer progression and tamoxifen resistance. J Clin Oncol
27: 542-549
30.
Blaschuk OW, Devemy E (2009) Cadherins as novel targets for anti-cancer therapy. Eur J Pharmacol 625:
195-198
31.
Krishna K, Redies C (2009) Expression of cadherin superfamily genes in brain vascular development. J
Cereb Blood Flow Metab 29: 224-229
32.
Teodorczyk M, Martin-Villalba A (2009) Sensing invasion: cell surface receptors driving spreading of
glioblastoma. J Cell Physiol 222: 1-10
15
33.
Hayward SW, Wang Y, Cao M et al (2001) Malignant transformation in a nontumorigenic human prostatic
epithelial cell line. Cancer Res 61: 8135-8142
34.
Orimo A, Gupta PB, Sgroi DC et al (2005) Stromal fibroblasts present in invasive human breast carcinomas
promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121: 335-348
35.
Gonda TA, Varro A, Wang TC et al (2010) Molecular biology of cancer-associated fibroblasts: can these
cells be targeted in anti-cancer therapy? Semin Cell Dev Biol 21: 2-10
36.
Bauchet L, Mathieu-Daude H, Fabbro-Peray P et al (2010) Oncological patterns of care and outcome for
952 patients with newly diagnosed glioblastoma in 2004. Neuro Oncol
37.
Pietras K, Ostman A (2010) Hallmarks of cancer: interactions with the tumor stroma. Exp Cell Res 316:
1324-1331
38.
Polyak K, Weinberg RA (2009) Transitions between epithelial and mesenchymal states: acquisition of
malignant and stem cell traits. Nat Rev Cancer 9: 265-273
39.
Phillips HS, Kharbanda S, Chen R et al (2006) Molecular subclasses of high-grade glioma predict
prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 9:
157-173
40.
Tso CL, Shintaku P, Chen J et al (2006) Primary glioblastomas express mesenchymal stem-like properties.
Mol Cancer Res 4: 607-619
41.
Ricci-Vitiani L, Pallini R, Larocca LM et al (2008) Mesenchymal differentiation of glioblastoma stem cells.
Cell Death Differ 15: 1491-1498
42.
Gunther HS, Schmidt NO, Phillips HS et al (2008) Glioblastoma-derived stem cell-enriched cultures form
distinct subgroups according to molecular and phenotypic criteria. Oncogene 27: 2897-2909
43.
Rieske P, Golanska E, Zakrzewska M et al (2009) Arrested neural and advanced mesenchymal
differentiation of glioblastoma cells-comparative study with neural progenitors. BMC Cancer 9: 54
44.
Carro MS, Lim WK, Alvarez MJ et al (2010) The transcriptional network for mesenchymal transformation
of brain tumours. Nature 463: 318-325
45.
Nakamizo A, Marini F, Amano T et al (2005) Human bone marrow-derived mesenchymal stem cells in the
treatment of gliomas. Cancer Res 65: 3307-3318
46.
Nakamura K, Ito Y, Kawano Y et al (2004) Antitumor effect of genetically engineered mesenchymal stem
16
cells in a rat glioma model. Gene Ther 11: 1155-1164
47.
Kang SG, Shinojima N, Hossain A et al (2010) Isolation and perivascular localization of mesenchymal stem
cells from mouse brain. Neurosurgery 67: 711-720
48.
Mishra PJ, Humeniuk R, Medina DJ et al (2008) Carcinoma-associated fibroblast-like differentiation of
human mesenchymal stem cells. Cancer Res 68: 4331-4339
17
Figure legends
Figure 1: (a) Computer screen of programmed surgery, showing the four zones of the GB (NZ, TZ, IZ and PBZ) in
a 3D approach. (b) Stereotaxic computer-assisted biopsy in the operating theater. (c) Mirror histologic control of the
four biopsy specimens (original magnification 100).
Figure 2: Examples of changes in the proportions of the cell contingents identified on the basis of their DI and IL13R2 mRNA expression during the culture of TZ (a) and PBZ biopsy specimens (b). Phase-contrast microscopy of
a TZ primary culture (P2) (c), GASCs (P3) (d) and control stromal cells (P3) (e). The bar indicates 50 µm.
Figure 3: (a) DNA copy number alterations in control stromal cells (n = 3), GASCs (n = 4), and TZ primary cultures
(n = 5). For each cell type, the y-axis indicates the percentage of the samples having an aberration at a specific point
in the genome. Amplifications are shown in green. Deletions are shown in red. Dark green and dark red indicate the
gain of two or more copies and homozygous loss, respectively. (b) Hierarchical clustering of the 24 probes
differentially expressed in GASCs and control stromal cells. The expression values presented are normalized, logtransformed intensities. (c) Validation of the differential expression of APOL-6, CDH6, RAB20, NRGN and MYH11
between GASCs (n = 4) and control stromal cells (n = 3) by Q-PCR. Results are expressed as the -fold change in
mRNA levels with respect to control stromal cells. (d) Representative flow cytometry analysis of CDH6 expression
in control stromal cells and GASCs. White and gray histograms represent isotype control and CDH6 labeling,
respectively.
Figure 4: (a) Flow cytometry immunologic analysis of GASCs (n = 5) and control stromal cells (n = 4). Results are
expressed as the mean percentage of positive cells  SEM (*, P < .05 versus control stromal cells). (b)
Differentiation of hMSCs and GASCs into classical mesenchymal lineages. Adipogenic and osteogenic
differentiation were detected by Oil Red O staining and Alizarin Red staining respectively. Four samples of GASCs
were analyzed and a representative sample is shown. GASCs were negative for adipogenesis while they were
positive for osteogenesis. HMSCs were used as positive controls for the differentiation into adipocytes and
osteoblasts. The bar indicates 100 µm.
Figure 5: (a) U87MG cells were injected alone (n = 10) or co-injected with GASCs (n = 4) or control stromal cells
(n = 5) subcutaneously into nude mice. Tumor volume was ploted in indicated days as medians  SEM (*, P < .05,
versus U87MG alone). (b) Direct coculture assay of GASCs or control stromal cells with the aneuploid glioma cell
18
line A172. Flow cytometry histograms show the DNA content of GASCs (DI = 1), control stromal cells (DI = 1) and
A172 cells (DI = 1.75). In each panel, the value of the proliferation index (PI) of A172 cells (means  SEM from
triplicate wells) is indicated (*, P < .05 versus A172 cells alone and A172 cells plus control stromal cells). For these
in vivo and in vitro experiments, two samples of GASCs (GASCs 2 and 3) and control stromal cells (control cells 1
and 3) were analyzed and a representative sample of each cell type is shown.
19