Patru et al.

BMC Cancer 2010, 10:66

http://www.biomedcentral.com/1471-2407/10/66

RESEARCH ARTICLE Open Access

CD133, CD15/SSEA-1, CD34 or side populations

do not resume tumor-initiating properties of
long-term cultured cancer stem cells from human
malignant glio-neuronal tumors
Cristina Patru1,9, Luciana Romao1,2, Pascale Varlet1,3, Laure Coulombel1, Eric Raponi4, Josette Cadusseau5,
François Renault-Mihara1, Cécile Thirant1, Nadine Leonard1,3, Alain Berhneim6, Maria Mihalescu-Maingot1,
Jacques Haiech7, Ivan Bièche8, Vivaldo Moura-Neto2, Catherine Daumas-Duport1,3, Marie-Pierre Junier1,3,
Hervé Chneiweiss1*

Abstract
Background: Tumor initiating cells (TICs) provide a new paradigm for developing original therapeutic strategies.
Methods: We screened for TICs in 47 human adult brain malignant tumors. Cells forming floating spheres in
culture, and endowed with all of the features expected from tumor cells with stem-like properties were obtained
from glioblastomas, medulloblastoma but not oligodendrogliomas.
Results: A long-term self-renewal capacity was particularly observed for cells of malignant glio-neuronal tumors
(MGNTs). Cell sorting, karyotyping and proteomic analysis demonstrated cell stability throughout prolonged
passages. Xenografts of fewer than 500 cells in Nude mouse brains induced a progressively growing tumor. CD133,
CD15/LeX/Ssea-1, CD34 expressions, or exclusion of Hoechst dye occurred in subsets of cells forming spheres, but
was not predictive of their capacity to form secondary spheres or tumors, or to resist high doses of temozolomide.
Conclusions: Our results further highlight the specificity of a subset of high-grade gliomas, MGNT. TICs derived
from these tumors represent a new tool to screen for innovative therapies.

Background Historically, it has been proposed that gliomas (astro-

Tumor initiating cells (TICs) from various types of can- cytomas and oligodendrogliomas) originate respectively
cers have been isolated and characterized. The tumors from mature astrocytes or oligodendrocytes. The fact
of origin range from glioblastomas and medulloblasto- that these brain tumors frequently include a mixture of
mas [1-6] to epithelial tumors of the breast [7], lung [8], cells expressing neuronal and glial markers, has recently
colon [9], and prostate [10]. Gliomas represent the led to the alternative proposal that gliomas arise from
majority of primary tumors from the central nervous neural stem/progenitor cells. Support for this hypothesis
system (CNS) [11]. Difficulties in clinical management comes from mouse models in which changes in the
(e.g. treatment and prognosis) are related to the com- expression of oncogenes or tumor suppressors lead to
plex identity of gliomas, which lack reliable morphologi- experimental tumors [13]. Neural progenitor cells are,
cal and molecular signatures, precluding thus the for example, more sensitive than differentiated astro-
establishment of a clear cut classification discriminating cytes to the oncogenic effects of combined over-activa-
between different tumor subtypes [12]. tion of Ras and Akt signaling pathways [14]. It should
however be kept in mind that glioblastomas, the most
malignant form of gliomas, can be generated in mice by
retroviral transduction of oncogenes into mature glial
* Correspondence: [email protected] cells [14-16]. In good agreement, the conversion of
1
Glial Plasticity lab; Inserm UMR 894; University Paris Descartes; Paris, France

© 2010 Patru et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
Patru et al. BMC Cancer 2010, 10:66 Page 2 of 11
http://www.biomedcentral.com/1471-2407/10/66

mature astrocytes toward neural progenitors induced by approved by the institutional review board of Ste Anne
TGFa [17], a growth factor overexpressed early in the Hospital, Paris. Anatomopathological diagnosis classified
development of human gliomas [18]sensitizes them to tumors as glioblastoma multiformis (n = 6), Malignant
cancerous transformation [19]. The isolation from Glio-Neuronal Tumors (MGNTs, n = 15), medulloblas-
human glioblastoma biopsies of malignant cells that toma (n = 2), ganglioglioma (n = 1), oligodendroglioma
express markers of neural stem cells supports the exis- (n = 23). Immunolabeling of formalin-fixed, paraffin-
tence of tumor stem cells within gliomas [1-3,6]. Most embedded tissue sections was performed as previously
importantly, some of these cells exhibit the true proper- described [20].
ties of tumor initiating cells (TICs), including the ability
to give rise to a tumor identical to the one observed in Cell cultures
the patient upon orthotopic grafting in mouse brains Viable fragments were transferred to a beaker contain-
[1,3,6]. It remains, however, unknown whether these ing 0.25% trypsin in 0.1 mM EDTA (4:1), and slowly
TICs might help to discriminate between glioma sub- stirred at 37°C for 30-60 min. Dissociated cells were pla-
types. Moreover, the design of specific therapies awaits ted in 75 cm2 tissue culture flasks plated at 2500-5000
the identification of the molecular pathways presiding cells/cm2 in Dulbecco’s modified Eagle’s: F-12 medium
over the maintenance of the properties of these tumor (1:1) containing the N2, G5 (containing FGF and EGF)
stem cells. and B27 supplements (all from Invitrogen, France).
Here, we sought for tumor stem-like cells in 47 After 2 to 47 days in culture, spheres bloomed from
human adult malignant glial tumors. We identified a clusters of adherent cells. They were dissociated in sin-
subset of glial tumors that contain at high frequency of gle-cell suspension each week with a renewal of two
cells generating long-term self-renewing floating spheres third of their culture medium.
in vitro, and novel tumors in immunodeficient mice.
This subset corresponds to malignant glio-neuronal Characterization of sphere-forming cells
tumors (MGNTs) [20]. MGNTs are World Health Orga- Dissociated sphere-derived cells were plated in 18-mm
nization grade III and IV tumors that always present diameter wells in 1 ml volumes at a density of 200 000
numerous glial fibrillary acidic protein (GFAP)- and a cells/well. Cell proliferation/viability was evaluated using
few neurofilament protein (NFP)-positive tumor cells. the WST1 kit from Roche according to the manufac-
The other neuronal markers tested (NeuN, synaptophy- turer’s protocol. We verified that the results were the
sin, and chromogranin) are inconstantly expressed. Dis- same as those obtained by counting the absolute num-
tinction of MGNTs from other malignant gliomas is of ber of viable cells using trypan blue exclusion. Clonality
clinical importance since gross total surgical resection of was evaluated in 96-well plates seeded at a cell density
these tumors is the major prognostic factor predicting of 1-200 cells/well/0.1 ml. The number of wells contain-
long-term survival [20]. Flow cytometry and 2D-SDS- ing at least one secondary sphere was evaluated after
PAGE analyses showed stable and common proteomic 3-4 weeks of culture.
profiles of MGNT-derived tumor initiating cells growing
as floating spheres. These cells are highly resistant to Flow cytometry analysis of cell surface antigens
temozolomide and thus represent a novel tool to screen Single-cell suspensions were incubated with fluoro-
for more efficient therapies. chrome-conjugated monoclonal antibodies: fluorescein
(FITC)-coupled CD15, FITC-CD11b, FITC-CD45, phy-
Methods coerythrin (PE)-coupled CD34, PE-Thy-1 (CD90), PE-
Sample Classification CD133 and phycoerythrin cyanin 5 (PC5)-coupled
All of the samples were classified according to World CD34, PC5-CD56, and allophycocyanin (APC)-coupled
Health Organization guidelines (grade II, III or IV for CD133. CD133 antibodies were from Miltenyi (France),
gliomas), and the classification of Sainte Anne Hospital all others from Beckman-Coulter. Dead cells were
(low grade oligodendroglioma or type A, high grade oli- excluded from the analysis by 7-AAD. Acquisition was
godendroglioma or type B, glioblastomas and malignant performed on a Facscalibur (Becton Dickinson) and
glio-neuronal tumors, [20]. The biopsies were collected viable cells were analyzed with Cellquest software.
by a pathologist in the surgical room from July 2002 to
July 2005. All patients were 18 years old or older, had Assessment of the neural differentiation potential
signed a written agreement for participation to the Cells were plated at a density of 5.000 cells/cm 2 onto
research project after having being informed of the polyornithine-coated glass coverslips and onto Lab-
goals, potential interest of the research and methods. Tek™ II Chamber Slide™ System (Nalge Nunc Interna-
This biomedical research was conducted according to tional) http://www.nuncbrand.com/page.aspx?ID=234 in
the declaration of Helsinki, to the French laws, and was the presence of FCS or B27 (Invitrogen) for 2 to 14
Patru et al. BMC Cancer 2010, 10:66 Page 3 of 11
http://www.biomedcentral.com/1471-2407/10/66

days. Multiple immunofluorescence assays for neural 20°C with a maximum current setting of 50 mA/strip.
antigens were performed as previously described [17,20]. Immobiline DryStrip gels (IPG stripes, Amersham Bios-
ciences) with a pH gradient of 3-10 or 4-7 and a length
Real-time RT-PCR of 13 cm were rehydrated in 250 μl sample solution,
The theoretical and practical aspects of real-time quanti- containing 80 μg proteins. Prior to SDS-PAGE analysis,
tative RT-PCR using the ABI Prism 7900 Sequence IPG stripes were equilibrated in second-dimension equi-
Detection System (Perkin-Elmer Applied Biosystems) libration buffer (6 M urea, 2% SDS, 50 mM Tris-HCl,
have been described in detail elsewhere [21]. Briefly, pH 8.8, 30% glycerol) containing 1% DTT (Sigma) to
total RNA is reverse-transcribed before real-time PCR reduce the disulfide bonds of the proteins. The second
amplification. Quantitative values are obtained from the dimension was carried out using 12.5% gradient
threshold cycle (Ct) number at which the increase in polyacrylamide gels at constant 20 mA current per gel.
the signal associated with exponential growth of PCR Comparison of the protein maps was achieved using the
products begins to be detected using PE Biosystems ana- PD-Quest software according to the manufacturer
lysis software, according to the manufacturer’s manuals. instructions (BioRad).
We also quantified transcripts of TBP gene, which
encodes the TATA box-binding protein as the endogen- Results
ous RNA control, and each sample was normalized on Selection and expansion of cell-forming spheres from
the basis of its TBP content. adult human brain tumors
Results, expressed as N-fold differences in target gene In order to preserve tissue architecture and improve sam-
expression relative to the TBP gene, termed Ntarget, ple handling as well as the versatility of the tests that may
were determined by the formula: Ntarget = 2 ΔCt sample , be applied, we developed a two-step strategy for the 49
where ΔCt value of the sample was determined by sub- surgical samples of our survey. Tumor samples were first
tracting the Ct value of the target gene from the Ct sliced in the surgical room in fragments smaller than 1
value of the TBP gene. mm3, and placed on surgical sponges bathed with culture
The Ntarget values of the samples were subsequently medium. Tumor fragments were either dissociated within
normalized to a “basal mRNA level”, i.e. normalized to 2 hours after surgery (28 out of 49, 58%), or incubated in
the smallest amount of target gene mRNA detectable organotypic conditions for a maximal time of 2 weeks
and quantifiable by real-time quantitative RT-PCR prior dissociation. Tumor fragments cultured on surgical
assays (target gene Ct value = 35; Ntarget value = 1). sponges are viable and conserve the tumor architecture
The nucleotide sequences of primers for TBP and the for up to 6 weeks (additional file 1). In both cases, the
4 target genes are available on request. The thermal dissociated cells were plated at low density (<5 × 10 3
cycling conditions comprised an initial denaturation cells/cm2) in a serum-free medium containing EGF and
step at 95°C for 10 min and 50 cycles at 95°C for 15 s FGF2 (G5 supplement).
and 65°C for 1 min. Experiments were performed with In twelve of the fifteen cultures of MGNTs, floating
duplicates for each data point. cellular spheres were observed after 14 to 60 days in
culture, and could be amplified from 6 months up to 52
Chromosome Analysis of neurosphere-derived cells months. In contrast, none of the other 23 tumors classi-
Cells were treated with medium containing 10 μg/ml fied as World Health Organization grade II or III oligo-
Colchicine for 1 to 2 hours, and resuspended in hypo- dendrogliomas (n = 9 and 14, respectively) generated
tonic 1% sodium citrate at room temperature for 30 spheres up to 90 days, the time at which no viable cells
minutes. The cells were then washed in a methanol- remained in the cultures. Similarly, two benign lesions,
acetic acid (3:1, v/v) fixative solution for 30 minutes, one Taylor Dysplasia and one adult pilocytic astrocy-
and spread onto clean dry slides. Q-banding staining toma, did not yield any sphere. A third in vitro behavior
was then performed, and 10 metaphases were analyzed pattern was observed for cells derived from three of the
for each sample. six glioblastoma multiformis studied, two adult medullo-
blastomas and one adult ganglioglioma. In these cases,
Proteomic analysis spheres developed but stopped dividing after 2 to 5
Cells were washed three times with PBS for 10 min with months of culture, and disappeared suggesting limited
gentle shaking, prior to lyses in buffer containing 8 M auto-renewal capabilities (Table 1).
urea, 4% CHAPS and 40 mM Tris. The protein pellets These data extended previous reports of absence of
were dissolved in isoelectric focusing buffer and quanti- EGF/FGF2-responsive clonogenic precursors in
fied using the Micro BCA protein assay kit (Pierce). oligodendrogliomas [2-4,22]. They revealed the unique
First-dimension isoelectric focusing was performed with ability of MGNTs to yield cellular spheres that could be
the Ettan IPGphor system (Amersham Biosciences) at amplified for very long periods.
Patru et al. BMC Cancer 2010, 10:66 Page 4 of 11
http://www.biomedcentral.com/1471-2407/10/66

Table 1 Summary of the brain tumor cultures. See text for details.
Tumor type Number of cases Number of cultures containing Self-renewal length ≤ Self-renewal length >
floating cellular spheres 6 months 6 months
MGNT 15 12 0 12
Oligodendroglioma II 9 0 - -
Oligodendroglioma III 14 0 - -
Glioblastoma 6 3 3 0
Pilocytic astrocytoma 1 0 - -
Taylor Dysplasia 1 0 - -
Ganglioglioma 1 1 1 0
Medulloblastoma 2 2 2 0

MGNT contain cells endowed with long-term self-renewal immunoreactive for GFAP, an astroglial marker (Figure
and clonal properties 1B and 1I), whereas a small population was immunor-
MGNT cells were moderately proliferative but highly eactive for the neuronal marker ß3-tubulin (Figure 1D
clonogenic. The average proliferation rate, determined and 1I). Each individual sphere exhibited cellular
using cultures derived from MGNT 1, 6 and 7 (TG1, heterogeneity.
TG6 and TG7), was equal to 236% ± 27% within 48 When bFGF and EGF were removed and replaced by
hours. In two separate experiments, the percentage of serum, most but not all spheres underwent changes
cells initiating-spheres, determined by plating 500 disso- characteristic of differentiation: they adhered rapidly to
ciated cells in 6-well plates and assessing the number of the plastic substrate, the cells flattened and acquired a
secondary spheres, reached 25-30%. Similar results were fusiform shape. Cellular heterogeneity both between
obtained when cells were plated at a density of 200 cells spheres and within each sphere was maintained in these
per well in 96-well plates (TG1, 28.9% ± 8.0; TG6, conditions, as shown by immunocytochemical analysis
15.5% ± 7.0; TG7, 15.6% ± 5.0, mean ± sem, n = 3) (Figure 1E-G and 1I). When treated with 0.5% FCS for
(Table 2). Finally, 7 to 10% of the cells seeded at one 14 days, all cells exhibited nuclear Olig2-immunolabel-
cell/well generated at least one secondary sphere after ing (Figure 1G and 1I). In the presence of B27, a cock-
one month in culture (TGI, 9 ± 3%, n = 4, TG6, 7 ± tail designed to promote neuron survival and
1%, n = 3, TG7, 7 ± 2%, mean ± sem, n = 3). These development in cultures of normal nervous tissue, 30%
properties were reproducibly observed for TG1-cells of the cells within the spheres acquired the appearance
from week 40 to week 100, the latest time tested. of neuron-like cells with a rich neurite-like branching,
and ß3-tubulin expression (Figure 1H and 1I). Interest-
MGNT-derived cells forming spheres exhibit a neural ingly, a fraction of the differentiated cells co-expressed
progenitor phenotype in vitro neuronal and glial markers as in the original tumor (Fig-
Analysis was conducted with a panel of antibodies direc- ure 1F and 1I).
ted against antigens characteristic of stem/progenitor
cells of the CNS and other tissues, and of cells of the Cell surface phenotype and the functional state of
neuronal and glial lineages (Figure 1A-D). Numerous sphere-derived cells
cells expressed nestin and the bHLH-transcription factor The cell-surface phenotype of sphere-derived cells from
Olig2 (Figure 1A, 1C and 1I), normally expressed in different MGNT cultures (TG1, TG6, TG7) was repeat-
stem/progenitor cells. A majority of cells were edly characterized using flow cytometry and remained
stable over 6 months and up to 52 months for TG1, the
Table 2 Clonal properties of MGNT derived glioma stem
latest time tested (Figure 2). CD56/NCAM, CXCR4,
cells CD90/Thy-1, and VLA2, which are known to be express
by neural progenitors, were detected in over 90% of the
Culture name 200 cells/well 1 cell/well
cells. A second group of proteins was detected in less
TG1 28.9 ± 8 9±3
than 10% of the cells. They included CD15/Lex/ssea1,
TG6 15.5 ± 7.0 7±1
described on neural stem cells [23], and CD34 synthe-
TG7 15.6 ± 5.0 7±2 sized by stem/progenitor cells from several tissues, but
Example of clonal properties of cell-initiating spheres derived from MGNT. The which so far has not been reported on neural cells.
cellular spheres were dissociated, and plated at 200 or 1 cell/well in 96-well
plates. The percentage of wells containing spheres is presented as mean ±
CD34 is known to be also expressed by endothelial cells,
sem, n = 3. that may originate from neural stem cells in normal
Patru et al. BMC Cancer 2010, 10:66 Page 5 of 11
http://www.biomedcentral.com/1471-2407/10/66

Figure 1 Immunocytochemical characterization of cells derived from MGNT. MGNT-derived cells form spheres in culture and express neural
markers. Example of spheres from one tumor, TG1, cultured with FGF/EGF (A-D) or in the presence of 10% (E-F), or 0,5% FCS (G). A-D. In serum-
free medium, cells within spheres expressed nestin (green, A), GFAP (red, B), Olig2 (red, C) and low levels of ß3-tubulin (green, D). E-G. After 14
days in serum-containing medium, the numbers of cells expressing Nestin decreased (green, E), The differentiated cells retained nuclear Olig2
expression (red, G). Some cells exhibited co-immunolabelling for GFAP (red) and ß3-tubulin (green, F). H. ß3-tubulin-immunoreactive cells in
serum-free medium supplemented with B27. Nuclei were counterstained with DAPI or Topro. Scale bar = 40 μm. I. Quantification of cells
expressing nestin, GFAP, Olig2 and ß3-tubulin in serum-free medium containing EGF and bFGF, in medium supplemented with 0.5 and 10% FCS,
and in medium supplemented with B27.

brains [24]. However, no expression was detected for the the negative and positive fractions from TG1, and TG6
endothelial-specific markers CD31 and VE-cadherin by cultures (week 35 to 90), and determined their ability to
MGNT-derived cultured cells. Similarly, none of the generate spheres in culture. The ability to form spheres
hematopoietic markers CD45, CD3, CD11b, CD14, segregated neither with CD15 nor with CD34 expression
CD19, CD33, CD36, CD38 was expressed, thus exclud- (Table 3). A side population was identified using the
ing that CD34 expression reflected contamination by Hoechst 3342 dye. Cells excluding the dye amounted to
endothelial or hematopoietic progenitors. Likewise, we 2-10% of the cellular population (2 independent experi-
did not detect expression of the cytokine receptors c-kit ments). The capacity of the cells to form spheres (in
or Flt3, or CD44 and CD93. In addition, we looked for both primary and secondary assays) was independent of
CD133-immunoreactive cells that have been reported to their ability to exclude the dye, suggesting that this
represent up to 50% of dissociated tumor cells examined property depends from the cell activation state rather
by flow cytometry [1-6]. In agreement with these data, than its identity.
we found that 20 to 30% of cells isolated immediately RT-PCR analysis was additionally used to seek for
after tumor resection were positive for this antigen (n = transcripts usually observed in stem cells and considered
4). Surprisingly, only a few cultured cells expressed as involved in their self-renewal capacities. Oct 4,
CD133, suggesting a rapid down-regulation in vitro. Nanog, and hTERT transcripts were observed in all
Furthermore, CD133+ as well as CD133- cells were able MGNT-derived cells forming spheres analyzed (n = 3)
to form floating spheres as well as to initiate tumors (Table 4). Of note, CD133 transcripts were always close
upon brain grafting (see below). One may hypothesize to the limit of detection (Table 4).
that CD133 expression is not a characteristic of a sub- Molecular characterization of cells-forming spheres
population of cells endowed with stem-like properties at was further performed using 2D-SDS-PAGE. The stabi-
least in MGNTs, but rather depends on their lity of the proteomic profile of a given culture was
environment. determined using TG1 cells at two different passages
To determine if the sphere-forming capability segre- (week 40 and week 160). The proteomic profiles of cells
gated according to CD34 or CD15 expression, we sorted isolated from the spheres were also compared between
Patru et al. BMC Cancer 2010, 10:66 Page 6 of 11
http://www.biomedcentral.com/1471-2407/10/66

different MGNTs. Analysis after silver staining revealed

a very similar protein pattern for TG1 at both passages
examined, as well as between cells derived from distinct
patients (TG6, TG7) (Figure 3). Comparison of the pro-
tein maps using computer-assisted analysis (PD-Quest
software) showed a high coefficient of correlation (0,84
± 0.06). A differential subtractive approach comparing
protein patterns from cells collected from MGNT-
derived spheres and a frozen micro-dissected fragment
of the original tumor, allowed the MALDI-TOF identifi-
cation of a group of proteins enriched in cultured cells
(Table 5). The eight potential biomarkers identified
included the mannose-6-phosphate receptor binding
protein 1 reported on early stem cells [25], and two pro-
teins associated with multidrug resistance, Sorcin being
the most strikingly over-expressed. Sorcin mRNA and
protein levels have been reported in astrocytoma and
are positively correlated with the malignancy of the
tumor [26]. These data are consistent with the possibi-
lity that Sorcin expression signs for an enhanced occur-
rence of TICs in high-grade gliomas.
Altogether, these results demonstrate that cells form-
ing cells derived from MGNT exhibit a common and
stable surface antigen and proteomic profile, as well as
some features of stem/progenitor cells in long-term
cultures.

MGNT-derived spheres contain tumor-initiating cells

Karyotype analysis and in vivo behavior upon orthotopic
grafting were performed in order to determine whether
the cell-forming spheres derived from MGNTs were
indeed tumoral and behaved as tumor-initiating cells
(TICs).
Karyotype analysis was performed on cells collected
from spheres grown from 4 tumors (TG1, TG6, TG7,
TG10). Gains were observed for each cell examined on
chromosomes 1q, 5, 7, 9q, 12, and 14, whereas losses
were observed on chromosomes 9p and 18q (Fig 4).
To determine the presence of TICs, MGNT-derived
sphere cells from four different MGNTs were grafted
Figure 2 Expression of surface antigens by from freshly
dissociated MGNTs cells and MGNT-derived cells forming into the brain of nude mice (2 mice each, 200-1000 liv-
spheres in culture. A. Control IgGs. B. Fresh tumors were ing cells per graft). The recipients were sacrificed from
dissociated enzymatically with trypsin and immediately 12 to 24 weeks post-graft. Tissue analysis after hematox-
immunolabeled with either CD133 (left panel) or CD34 (right panel). ylin/eosin and immunolabeling with antibodies that
C. MGNT-derived cellular spheres were dissociated mechanically as
recognized only the human form of the glial marker
described (see text), and immunolabeled. Cells were analyzed by
flow cytometry and the positive gates indicated were defined based vimentin identified human cells in 7 out of 8 mice. In
on analysis done after double labeling with FITC- or PE-labeled six cases out of seven, the grafts had developed with cell
isotype-matched non-immune immunoglobulin and another proliferation (as evidenced by Mcm2 labeling), and
specific antibody. Results are expressed in a morphological gate migration over large distances from the injection site
including all viable nucleated cells. Data shown are from one
(Fig 5). Several migrating cells were observed along the
experiment out of three giving similar results.
neural fiber axis (Fig 5).
Patru et al. BMC Cancer 2010, 10:66 Page 7 of 11
http://www.biomedcentral.com/1471-2407/10/66

Table 3 Phenotypic characterization of cell-forming tumor cells that are clonogenic and are capable of form-
spheres. ing a tumor mass mimicking the original one. The pre-
% of wells containing spheres Number of sphere/well sent work reports on the characterization of tumor
CD15- 11/11 = 100% 15 ± 2.2 (day 21) initiating cells from 12 cases of a newly characterized
CD15+ 9/13 = 70% 5.9 ± 2.6 (day 21) form of human adult glioma. These cells exhibit long-
CD34- 14/14 = 100% 22 ± 4.1 (day 36) term stability in culture, and properties that support
their capacity to establish a novel brain tumor from a
CD34+ 12/15 = 80% 20 ± 3.6 (day 36)
few cells at least up to two years of continuous passages.
Non SP 10/10 = 100% 18.6 ± 0.5 (day 36)
Within the array of brain tumors studied in this work,
SP 10/10 = 100% 19.1 ± 0.2 (day 36)
only MGNTs [20] yielded at a high frequency cells
Cells cultured as spheres from either TG1, TG6 or TG7 were dissociated and forming spheres in culture (12 of 15 tumors). In con-
sorted according to their expression of CD15, CD34 or they capability to
exclude Hoechst 3342 dye. Sorted cells were then plated (100 cells/well) in trast, cells proliferating as spheres from glioblastomas,
round-bottomed 96-well plates and fed once a week. Wells were examined at adult medulloblastoma and ganglioglioma were rapidly
days 21 and 36, and the number of spheres (with more than 50 cells) per well
was recorded. Data are from 1 representative experiment out of 3.
lost upon serial passages after 4 to 6 months in accor-
dance with previous studies [3,4]. Most studies of glio-
Taken altogether these data demonstrate that long- blastomas have identified tumor stem cells in only half
term cultured MGNT-derived floating spheres contain or less of the glioblastomas studied [3,4]. Our results
tumor cells with neural stem/progenitor phenotype, thus raise the possibility that some of the glioblastomas
fulfilling the criteria of tumor initiating cells (TICs). assayed in these studies belong to the MGNT sub-class.
Review of 200 glioblastomas showing that 50% of them
TICs resist to high concentrations of temozolomide contain cells expressing NFP and GFAP supports such a
Temozolomide, an oral alkylating agent, is the main possibility (PV and CDD, unpublished observations).
drug used for high grade gliomas. The therapeutic con- The presence of a few tumor cells co-expressing neuro-
centration given to patients is around 60 μM. Cells cul- nal and glial antigens is one MGNT characteristic,
tured from tumors TG1 at three different passages, which can be easily looked for [20]. In addition, our
TG10 and TG16 were exposed for 48h to different con- large set of cultured oligodendrogliomas (WHO grade II
centrations of temozolomide. Cells from all cultures or III) never yielded any sphere and/or cells expressing
were highly resistant to temozolomide, 60% of the cells a stem-like phenotype. In good agreement with this, no
being still viable in presence of 1000 μM temozolomide long-term culture from human oligodendroglioma has
(Table 6). No difference in temozolomide resistance was yet been reported. Determination of the presence or
noted between long-term cultured spheres obtained absence of tumor stem cells is, however, insufficient in
from FACS-sorted CD133+ and CD133- cells. In addi- the present state of knowledge to conclude that
tion, removal of growth factors for 10 days or induction MGNTs, medulloblastomas and some glioblastomas
of differentiation for 7 days in the presence of 10% FCS result from the targeting of transforming mutations to
did not change TIC resistance to TMZ. an early progenitor/stem-like cell, whereas oligodendro-
gliomas result from mutations accumulating in a more
Discussion differentiated cell.
Tumor stem cells or tumor-initiating cells with stem- CD133 expression, because expressed also by normal
like properties (TICs) name a small subpopulation of neural stem cells, has been proposed to identify cells at
the top of the hierarchy formed by tumor stem cells and
their more differentiated progeny, and to be therefore
Table 4 Expression of stemness markers by MGNT- the paramount of glioma stem cells. Its pertinence as a
derived cells forming spheres. glioma stem cell marker is now highly controversial,
Samples OCT4/POU5F1 NANOG HTERT CD133/PROM1 several groups having demonstrated the tumor initiating
TG1 5937 a
1240 131 5 properties of CD133- cells [27]. In our hands, CD133
was rapidly down-regulated in cell spheres grown in
TG6 11857 1959 165 12
vitro, whereas expression of other stem cell markers
TG10 1686 169 517 3
such as CD15 or CD34 was maintained. CD133- cells
TG16 871 611 73 14
were also able to form floating cellular spheres with
OCT4/POU5F1, NANOG, HTERT and CD133/PROM1 genes expressions were properties undistinguishable from those of CD133+
evaluated using real-time quantitative PCR in 4 MGNT-derived cells forming
spheres. Tests were done with cultures at passage 30 ± 5. Data are from one cells. Similarly, long-term cultured sphere forming cells,
experiment out of three giving similar results. a N-fold differences in target whatever established from CD133+ or CD133- tumor
gene expression relative to the smallest amount of target gene mRNA
detectable and quantifiable by real-time quantitative RT-PCR assays (target
cells were equally resistant to temozolomide. CD133
gene Ct value = 35; Ntarget value = 1). expression most likely reflects the bioenergetics stress of
Patru et al. BMC Cancer 2010, 10:66 Page 8 of 11
http://www.biomedcentral.com/1471-2407/10/66

Figure 3 Proteomic patterns of MGNT-derived cells forming spheres. Upper panel: two-dimensional gel electrophoresis analysis for TG1,
one of the cultures of MGNT-derived cells forming spheres, TG1, at 40 (left) and 160 (right) weeks of culture. Lower panel: the same area of the
gel as in the upper panels for protein extracts taken from two other cultures of MGNT-derived cells forming spheres, TG6 and TG7.

the cells rather than their stem-like properties [28]. Conclusions

CD15, also known as SSEA1 or Lewis X, has also been We previously reported that MGNTs represent a clini-
recently proposed as an enrichment marker of glioma cal entity with distinct clinical, anatomo-pathological
TICs [29]. Although we observed some differences in and radiological behaviors. We now show that they
sphere sizes and proliferation rates between CD15 posi- have also a distinct biological behavior among high-
tive and negative cells during the first week after sorting, grade glial tumors. The glioma initiating cells described
these variations disappeared with obtaining the second- here establish a novel model that may be used to routi-
ary spheres, which in both cases contained a mixture of nely establish adult human tumoral cell lines stable in
CD15+ and CD15- cells. long-term cultures in a define medium in a

Table 5 MALDI MS/MS characterization of proteins overexpressed in MGNT-derived cells forming spheres.
Entry SwissProt Masse Pi % cover Identification
RET1_HUMAN 15879.92 4,99 60 (P09455) Retinol-binding protein I
SORCN_HUMAN 21947.46 5,32 30 (P30626) Sorcin (22 kDa protein) (CP-22) (V19)
GSTP1_HUMAN 23438.07 5,44 27 (P09211) Glutathione S-transferase P (EC 2.5.1.18)
PRDX6_HUMAN 25002.19 6,02 25 (P30041) Peroxiredoxin 6 (EC 1.11.1.15)
ECHM_HUMAN 31823.30 8,34 29 (P30084) Enoyl-CoA hydratase
HSPB1_HUMAN 22825.51 5,98 66 (P04792) Heat-shock protein beta-1 (HspB1)
M6PBP_HUMAN 47189.00 5,30 53 (O60664) Mannose-6-phosphate receptor binding protein 1
Q96E67_HUMAN 40542.00 5,55 32 (Q96E67) ACTB protein (Fragment)
The proteins listed in the table were only detected in 2D-SDS silver stained gels obtained from MGNT-derived cells forming spheres and not observed in the
corresponding gel obtained with the same amount of proteins extracted from the original brain tumor. The proteins were identified by mass spectrometry
analysis.
Patru et al. BMC Cancer 2010, 10:66 Page 9 of 11
http://www.biomedcentral.com/1471-2407/10/66

Figure 4 Karyotype. Chromosomal analysis of TG1 cells by Multi FISH reveals a hypotriploidy with an over-representation of chromosomes 7
and 14, and one chromosome 11q+.

Figure 5 Intra-cerebral grafts. MGNT-derived cells forming spheres were grafted in the dorsal striatum of immunodeficient mice (inset in panel
B). Mice were sacrificed from 12 to 22 weeks post-graft and brain sections immunostained. A. Hematoxylin-eosin staining was performed to
identify tumor mass (inset: a cell undergoing mitosis within the tumor mass). B. Identification of human cells with an antibody specific of the
human vimentin (brown). C-D. Double immunolabeling for vimentin (brown) and minichromosome maintenance 2 (mcm2, red nuclear labeling)
identified cycling human cells (arrows). Cells were able to migrate at distance from the site of injection (D). Scale bar: 100 μm in B, 50 μm in A,
20 μm in C, D and inset in A. B-D: floxin counterstaining.
Patru et al. BMC Cancer 2010, 10:66 Page 10 of 11
http://www.biomedcentral.com/1471-2407/10/66

Table 6 TICs-derived from MGNTs are resistant to following grants: ARC 3952 (HC), Ligue National contre le Cancer (La Ligue
Temozolomide 2007, HC), Cancéropole (INCa/Région Ile de France) (CT), Cancer Stem Cell
Network Ile de France (HC, MPJ).
TMZ 62,5 TMZ 250 TMZ 500 TMZ 1000
μM μM μM μM Author details
1
TG1 P35 95 ± 7 94 ± 4 95 ± 5 67 ± 4 Glial Plasticity lab; Inserm UMR 894; University Paris Descartes; Paris, France.
2
Department of Anatomy, UFRJ, Rio, Brazil. 3Department of Neuropathology,
TG1 P70 95 ± 5 87 ± 5 83 ± 4 66 ± 6 Sainte-Anne Hospital, Paris, France. 4Grenoble - Institut de Neurosciences
TG1 P90 96 ± 5 87 ± 5 78 ± 7 58 ± 7 INSERM U83638706 La Tronche cedex, France. 5Inserm U955, Team 10,
University of Paris 12, Créteil, France. 6Laboratoire de génomique cellulaire
des cancers (FRE 2939 CNRS), Institut Gustave Roussy, Villejuif, France. 7CNRS
TG10 98 ± 6 97 ± 5 80 ± 6 59 ± 4 UMR 7175; Institut Gilbert Laustriat; University Strasbourg I; Illkirch, France.
8
Génétique et Biothérapies des Maladies Dégénératives et Prolifératives du
TG16 90 ± 5 80 ± 7 75 ± 5 62 ± 6 Système Nerveux INSERM U745; IFR71; Université Paris Descartes; Paris,
France. 9Present address: Department of Neurosurgery, Annecy, France.
TG16 CD133+ 92 ± 5 81 ± 7 74 ± 5 67 ± 8 Authors’ contributions
TG16 CD133- 91 ± 6 75 ± 6 71 ± 5 60 ± 6 CP and LR carried out the cell cultures, ICC and IHC, cell sorting, proteomic
studies, and drafted the manuscript. PV, NL and CDD carried out the
neuropathological analysis of tumor samples and xenografts. ER and JC
TG1 without GF 92 ± 6 85 ± 6 78 ± 6 62 ± 7 carried out the xenografts. AB carried out chromosomal analysis. LC
participated in the FACS analysis. IB, CT, MM and HC participated in Q-PCR
TG1 10% SVF 90 ± 5 80 ± 5 82 ± 6 79 ± 5
analysis. FRM, CT and MMM carried out the pharmacological experiments.
Cells cultured from tumors TG1, TG10 and TG16 were exposed for 48h to MPJ, VMN and JH participated in the design of the study, and MPJ to
different concentrations of temozolomide (TMZ). The analysis was performed redaction of he manuscript. HC conceived of the study, and participated in
on three passages of TG1 (weeks 35, 70 and 90 after clonal selection), and on its design and coordination and drafted the manuscript. All authors read
TG10 and TG16 at passage 40. It was additionally performed on cultures of and approved the final manuscript.
TG16 initiated from CD133+ (TG16 CD133+) or CD133- (TG16 CD133-) sphere-
forming cells. The influence of growth factors (GF) or serum (10%SVF) on the Competing interests
cells response to temozolomide was evaluated using cells cultured for 10 days The authors declare that they have no competing interests.
without FGF/EGF or for 7 days in the presence of 10% fetal calf serum, prior
temozolomide addition. Cell viability was determined using WST1 test. Results Received: 29 July 2009
are presented as % of control. Mean ± sem of three independent
Accepted: 24 February 2010 Published: 24 February 2010
experiments, each experiment performed in triplicates.

References
1. Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S, Fiocco R,
reproducible fashion. This may open new ways to iden- Foroni C, Dimeco F, Vescovi A: Isolation and characterization of
tify novel tumor cell markers and surface receptor pro- tumorigenic, stem-like neural precursors from human glioblastoma.
Cancer Res 2004, 64(19):7011-7021.
files for therapeutic and diagnostic purposes and to 2. Hemmati HD, Nakano I, Lazareff JA, Masterman-Smith M, Geschwind DH,
develop patient-tailored pharmacologic approaches for Bronner-Fraser M, Kornblum HI: Cancerous stem cells can arise from
the cure of gliomas. pediatric brain tumors. Proc Natl Acad Sci USA 2003, 100(25):15178-15183.
3. Ignatova TN, Kukekov VG, Laywell ED, Suslov ON, Vrionis FD, Steindler DA:
Human cortical glial tumors contain neural stem-like cells expressing
List of abbreviations used astroglial and neuronal markers in vitro. Glia 2002, 39(3):193-206.
TIC: tumor initiating cells; MGNT: Malignant glio-neu- 4. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB:
Identification of a cancer stem cell in human brain tumors. Cancer Res
ronal tumors 2003, 63(18):5821-5828.
5. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM,
Additional file 1: Maintenance of tumoral architecture and cell Cusimano MD, Dirks PB: Identification of human brain tumour initiating
viability in organotypic cultures of human brain tumors. cells. Nature 2004, 432(7015):396-401.
Immunohistochemical staining of an original MGNT tumor (upper panels) 6. Yuan X, Curtin J, Xiong Y, Liu G, Waschsmann-Hogiu S, Farkas DL, Black KL,
and after 4 weeks in organotypic culture (lower panels). The tumor Yu JS: Isolation of cancer stem cells from adult glioblastoma multiforme.
samples were collected by an anatomopathologist in the surgical room, Oncogene 2004, 23(58):9392-9400.
sliced into small fragments (less than 1 mm3), and layered on surgical 7. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF:
sponge fragments floating over a culture medium composed of RPMI Prospective identification of tumorigenic breast cancer cells. Proc Natl
with 6% FCS. These organotypic cultures were maintained up to 6 weeks Acad Sci USA 2003, 100(7):3983-3988.
without visible alteration of the tumoral architecture, as compared to the 8. Kim CF, Jackson EL, Woolfenden AE, Lawrence S, Babar I, Vogel S,
original histological examination. Crowley D, Bronson RT, Jacks T: Identification of bronchioalveolar stem
Click here for file cells in normal lung and lung cancer. Cell 2005, 121(6):823-835.
[ http://www.biomedcentral.com/content/supplementary/1471-2407-10- 9. O’Brien CA, Pollett A, Gallinger S, Dick JE: A human colon cancer cell
66-S1.PDF ] capable of initiating tumour growth in immunodeficient mice. Nature
2007, 445(7123):106-110.
10. Patrawala L, Calhoun T, Schneider-Broussard R, Li H, Bhatia B, Tang S,
Reilly JG, Chandra D, Zhou J, Claypool K, et al: Highly purified CD44+
prostate cancer cells from xenograft human tumors are enriched in
Acknowledgements tumorigenic and metastatic progenitor cells. Oncogene 2006,
We thank Dr. Alain Klause for his expert technical assistance for karyotype 25(12):1696-1708.
analysis, Jocelyne Cordier for help to maintain cell cultures, Jean-Philippe 11. Daumas-Duport C, Beuvon F, Varlet P, Fallet-Bianco C: [Gliomas: WHO and
Deloulme for help in grafting, Edith Demettre and Philippe Marin for their Sainte-Anne Hospital classifications]. Ann Pathol 2000, 20(5):413-428.
help in proteomic analysis. This work was supported by Inserm and the
Patru et al. BMC Cancer 2010, 10:66 Page 11 of 11
http://www.biomedcentral.com/1471-2407/10/66

12. Noble M, Dietrich J: The complex identity of brain tumors: emerging

concerns regarding origin, diversity and plasticity. Trends Neurosci 2004,
27(3):148-154.
13. McConville P, Hambardzumyan D, Moody JB, Leopold WR, Kreger AR,
Woolliscroft MJ, Rehemtulla A, Ross BD, Holland EC: Magnetic resonance
imaging determination of tumor grade and early response to
temozolomide in a genetically engineered mouse model of glioma. Clin
Cancer Res 2007, 13(10):2897-2904.
14. Holland EC: Gliomagenesis: genetic alterations and mouse models. Nat
Rev Genet 2001, 2(2):120-129.
15. Ding H, Guha A: Mouse astrocytoma models: embryonic stem cell
mediated transgenesis. J Neurooncol 2001, 53(3):289-296.
16. Weiss WA, Burns MJ, Hackett C, Aldape K, Hill JR, Kuriyama H, Kuriyama N,
Milshteyn N, Roberts T, Wendland MF, et al: Genetic determinants of
malignancy in a mouse model for oligodendroglioma. Cancer Res 2003,
63(7):1589-1595.
17. Sharif A, Legendre P, Prevot V, Allet C, Romao L, Studler JM, Chneiweiss H,
Junier MP: Transforming growth factor alpha promotes sequential
conversion of mature astrocytes into neural progenitors and stem cells.
Oncogene 2007, 26(19):2695-2706.
18. Junier MP: What role(s) for TGFalpha in the central nervous system?. Prog
Neurobiol 2000, 62(5):443-473.
19. Dufour C, Cadusseau J, Varlet P, Surena AL, de Faria GP, Dias-Morais A,
Auger N, Leonard N, Daudigeos E, Dantas-Barbosa C, et al: Astrocytes
Reverted to a Neural Progenitor-like State with Transforming Growth
Factor Alpha Are Sensitized to Cancerous Transformation. Stem Cells
2009, 27(10):2373-2382.
20. Varlet P, Soni D, Miquel C, Roux FX, Meder JF, Chneiweiss H, Daumas-
Duport C: New variants of malignant glioneuronal tumors: a
clinicopathological study of 40 cases. Neurosurgery 2004, 55(6):1377-1391,
discussion 1391-1372.
21. Bieche I, Parfait B, Le Doussal V, Olivi M, Rio MC, Lidereau R, Vidaud M:
Identification of CGA as a novel estrogen receptor-responsive gene in
breast cancer: an outstanding candidate marker to predict the response
to endocrine therapy. Cancer Res 2001, 61(4):1652-1658.
22. Piccirillo SG, Reynolds BA, Zanetti N, Lamorte G, Binda E, Broggi G, Brem H,
Olivi A, Dimeco F, Vescovi AL: Bone morphogenetic proteins inhibit the
tumorigenic potential of human brain tumour-initiating cells. Nature
2006, 444(7120):761-765.
23. Capela A, Temple S: LeX/ssea-1 is expressed by adult mouse CNS stem
cells, identifying them as nonependymal. Neuron 2002, 35(5):865-875.
24. Wurmser AE, Nakashima K, Summers RG, Toni N, D’Amour KA, Lie DC,
Gage FH: Cell fusion-independent differentiation of neural stem cells to
the endothelial lineage. Nature 2004, 430(6997):350-356.
25. Assou S, Le Carrour T, Tondeur S, Strom S, Gabelle A, Marty S, Nadal L,
Pantesco V, Reme T, Hugnot JP, et al: A meta-analysis of human
embryonic stem cells transcriptome integrated into a web-based
expression atlas. Stem Cells 2007, 25(4):961-973.
26. Yokota T, Kouno J, Adachi K, Takahashi H, Teramoto A, Matsumoto K,
Sugisaki Y, Onda M, Tsunoda T: Identification of histological markers for
malignant glioma by genome-wide expression analysis: dynein, alpha-
PIX and sorcin. Acta Neuropathol 2006, 111(1):29-38.
27. Beier D, Hau P, Proescholdt M, Lohmeier A, Wischhusen J, Oefner PJ,
Aigner L, Brawanski A, Bogdahn U, Beier CP: CD133(+) and CD133(-)
glioblastoma-derived cancer stem cells show differential growth
characteristics and molecular profiles. Cancer Res 2007, 67(9):4010-4015.
28. Griguer CE, Oliva CR, Gobin E, Marcorelles P, Benos DJ, Lancaster JR Jr,
Gillespie GY: CD133 is a marker of bioenergetic stress in human glioma.
PLoS ONE 2008, 3(11):e3655. Submit your next manuscript to BioMed Central
29. Son MJ, Woolard K, Nam DH, Lee J, Fine HA: SSEA-1 is an enrichment and take full advantage of:
marker for tumor-initiating cells in human glioblastoma. Cell Stem Cell
2009, 4(5):440-452.
• Convenient online submission
Pre-publication history • Thorough peer review
The pre-publication history for this paper can be accessed here:http://www. • No space constraints or color figure charges
biomedcentral.com/1471-2407/10/66/prepub
• Immediate publication on acceptance
doi:10.1186/1471-2407-10-66 • Inclusion in PubMed, CAS, Scopus and Google Scholar
Cite this article as: Patru et al.: CD133, CD15/SSEA-1, CD34 or side
populations do not resume tumor-initiating properties of long-term • Research which is freely available for redistribution
cultured cancer stem cells from human malignant glio-neuronal tumors.
BMC Cancer 2010 10:66. Submit your manuscript at
www.biomedcentral.com/submit