Abstract. As the majority of gliomas arise through malignant
transformation of astrocytes, we aimed at investigating the
interaction between malignant glioma cells and astrocytes in
a co-culture experimental model. For this purpose we analyzed
the expression of genes and proteins involved in tumor
promotion and invasion, such as glial fibrillary acidic protein
(GFAP), matrix metalloproteinase-2 (MMP-2), tissue inhibitor
of MMP-2 (TIMP-2), transforming growth factor-ß1 (TGF-ß1),
secreted protein acidic and rich in cysteine (SPARC), and
connexin 43 (CX43). Co-cultures of human neural stem
cell-derived astrocytes and U87 MG astrocytoma cells were
performed in a transwell system. Gene expression was
evaluated by real-time RT-PCR, and protein analysis was
performed by Western blotting, SDS-zymography, and
immunofluorescence. GFAP tended to be up-regulated in
astrocytes co-cultivated with U87, suggesting a reactive
response induced by glioma cells. CX43 mRNA tended to be
down- regulated in co-cultured astrocytes, as well as the non-
phosphorylated isoform at the protein level. MMP-2 mRNA
tended to be up-regulated, and MMP-2 protein levels were
significantly increased in astrocytes co-cultivated with U87.
TIMP-2 and SPARC mRNA decreased in astrocytes co-
cultivated with U87, showing lower expression in glioma
cells. By contrast, SPARC protein expression was strongly
induced in supernatants of co-cultured astrocytes. TGF-ß1 was
not modified. Our results suggest that U87 cells elicit
phenotype modifications in the neighbouring resident
astrocytes very likely mediated by soluble factors. Glioma/
astrocyte interaction could possibly trigger an astrocyte
phenotype modification consistent with a malignant
transformation, and favouring a more permissive environment
for glioma cells invasion.
Gliomas are the most common primary brain tumors,
accounting for >40% of all central nervous system neoplasms
(1). Malignant gliomas, deriving from neoplastic trans-
formation of astrocytes, are characterized by the aggressive
and widespread invasion of glioma cells into surrounding
brain tissue (2). The infiltrative potential of gliomas limits
the efficacy of surgical resection and targeted radiotherapy,
leading to an unfavorable prognosis even in response to multi-
disciplinary treatment strategies.
Gliomas are ‘intraparenchymally metastatic’ tumors (3),
invading the brain in a non-destructive manner that suggests
cooperation between invading glioma cells and their environ-
ment, possibly using resident astrocytes as substrate (4).
The mechanism by which glioma cells migrate and invade
adjacent normal brain tissue are not yet completely understood,
and the behaviour of resident normal astrocytes outside the
tumor has not been detailed.
Tumor growth is the result of an evolving cross-talk
between malignant and surrounding normal cells. The micro-
environment of tumor cells plays a key role in the growth of
the tumor, and its modification and remodeling allows tumor
invasion, in particular upon up-regulation of matrix metallo-
ONCOLOGY REPORTS 22: 1349-1356, 2009
Glioma-astrocyte interaction modifies the astrocyte
phenotype in a co-culture experimental model
NICOLETTA GAGLIANO1, FRANCESCO COSTA1, CHIARA COSSETTI2,3,
LETIZIA PETTINARI1, ROSARIA BASSI4, MAURIZIO CHIRIVA-INTERNATI5,
EVERARDO COBOS5, MAGDA GIOIA1and STEFANO PLUCHINO2
1Department of Human Morphology and Biomedical Sciences ‘Città Studi’, Extracellular Matrix Laboratory-EML,
School of Medicine, Università di Milano, via F.lli Cervi 93, 20090 Segrate; 2CNS Repair Unit, DIBIT2 and Institute of
Experimental Neurology (INSPE), Division of Neuroscience, San Raffaele Scientific Institute, via Olgettina 58, 20132
Segrate, Milan, Italy; 3GABBA - Graduate Program in Areas of Basic and Applied Biology, Instituto de Ciências
Biomedicas Abel Salazar (ICBAS), Universidade do Porto, 4099-002 Porto, Portugal; 4Department of Medical Chemistry,
Biochemistry and Biotechnology, L.I.T.A., Università di Milano, via F.lli Cervi 93, 20090 Segrate, Milan, Italy;
5Division of Hematology and Oncology, Texas Tech University Health Sciences Center and Southwest Cancer
Treatment and Research Center, 3601 4th St., MS 6591, Lubbock, TX 79430, USA
Received July 3, 2009; Accepted August 24, 2009
Correspondence to: Dr Nicoletta Gagliano, Department of Human
Morphology and Biomedical Sciences ‘Città Studi’, Extracellular
Matrix Laboratory-EML, School of Medicine, University of Milan,
Via Fratelli Cervi 93, 20090 Segrate, Milano, Italy
Key words: glioblastoma, astrocytes, glial fibrillary acidic protein,
connexin 43, matrix metalloproteinase-2, secreted protein acidic
and rich in cysteine
proteinase (MMP)-2, as observed in cell lines and resected
The previous observation of reactive astrocytes around
glioma cells in the brain (7) suggested the possibility that
these two cell types could be interacting. In particular, it was
hypothesized that this interaction results in exploitation of
the astrocyte environment by glioma cells, leading to a
remodeling of the surrounding matrix and increased tumor
Although the role of MMP in glioma tumorigenicity is
well established, it remains to be determined whether glioma-
astrocyte interactions affect other aspects of astrocytes.
Normal resident astrocytes and glioma cells interact via gap
junctions and growth factors, and the glioma cells could elicit
a phenotypic transformation of astrocytes in order to render
the brain parenchyma more permissive to glioma invasion.
As the majority of glioma arise through malignant trans-
formation of astrocytes, we aimed at investigating the effect of
the interaction between malignant glioma cells and astrocytes
in a co-culture experimental model. For this purpose, we
focused on astrocytes and we analyzed whether glioma cells
triggered any phenotypic change of astrocytes, by analyzing
the expression of genes and proteins typical of normal
astrocytes and of proteins involved in tumor promotion and
Materials and methods
Human neural stem/precursor cell isolation and expansion and
differentiation protocol. Human neural stem/precursor cells
(hNPCs) were derived from the telencephalon and dience-
phalon of a single 10.5 post-conception week human foetus,
as described (8). Briefly, cells were grown and expanded in a
chemically defined, serum-free medium in the presence of
basic fibroblast growth factor (FGF)-2 and epidermal growth
factor (EGF) (10 and 20 ng/ml, respectively) (growth medium).
To induce differentiation of hNPCs into astrocytes,
suspensions of single cells were plated on Matrigel (BD
Biosciences) coated 6-well plates (250,000 cells/well), in
growth medium w/o EGF for the first 72 h. The medium was
then replaced with fresh control medium (w/o growth factors)
plus 1% fetal calf serum (FCS). The cells were cultured for
further 5 days in vitro before processing for co-culture
Cell cultures. U87 MG cells were obtained from European
Cell Culture Collection and used at the 5th passage for co-
cultures experiments. Co-cultures of hNPC-derived astrocytes
and U87 astrocytoma cells were grown in a transwell system
with a 0.4 μm pore size. hNPC-derived astrocytes (AA) were
seeded in the lower compartment of a 6-well transwell
system (2.5x105cells); in the insert U87 (1.25x105or 2.5x105
cells) were cultured, or DMEM was placed. Cells were
cultured in DMEM supplemented with 10% heat-inactivated
fetal bovine serum (FBS), antibiotics (100 U/ml penicillin,
0.1 mg/ml streptomycin), and 0.0025 μg/ml ampotericin B.
For analysis cells were maintained for 48 h in serum-free
DMEM and then harvested. Each sample was cultured in
duplicate, and each co-culture experiment was repeated 3
Immunocytochemistry. AA and U87 were cultured on 12-mm
diameter round coverslips put into 24-well culture plates and
cultured for 48 h. Cells were washed in phosphate-buffered
saline (PBS), fixed in 4% paraformaldehyde in PBS containing
2% sucrose for 5 min at room temperature, and post-fixed in
70% ethanol and stored at -20˚C until use. The cells were
then washed in PBS 3 times and incubated overnight at
4˚C with monoclonal anti-GFAP primary antibody (1:800,
Chemicon, Temecula, CA). Secondary antibody conjugated
with rodhamine (1:500, Molecular Probes, Invitrogen) was
applied for 1 h at room temperature, followed by rinsing with
PBS. Negative controls were incubated omitting the primary
antibody. After the labelling procedure was completed, the
coverslips were mounted onto glass slides using a mounting
medium with DAPI. The cells were photographed by a digital
camera connected to the microscope.
Real-time RT-PCR. Total RNA was isolated by a modification
of the acid guanidinium thiocyanate-phenol-chloroform
method (Tri-Reagent, Sigma). Total RNA (1 μg) was reverse-
transcribed in 20 μl final volume of reaction mix (BioRad).
The gene expression of MMP-2, TIMP-2, CX43, TGF-ß1 and
SPARC was analyzed.
The primer sequences, designed with Beacon Designer 6.0
software (BioRad), were: GAPDH: sense CCCTTCATTG
ACCTCAACTACATG, antisense TGGGATTTCCATTGA
TGACAAGC; MMP-2 sense GCAGTGCAATACCTGA
ACACCTTC, antisense TCTGGTCAAGATCACCTGTC
TGG; TIMP-2 sense TGGAAACGACATTTATGGCAA
CCC, antisense CTCCAACGTCCAGCGAGACC; CX43
sense CTCTCGCCTATGTCTCCTCCTG, antisense TTTGC
TCACTTGCTTGCTTGTTG; TGF-ß1: sense GTGCGGCA
GTGGTTGAGC, antisense GGTAGTGAACCCGTTGA
TGTCC; TIMP-1: sense GGCTTCTGGCATCCTGTTGTTG,
antisense AAGGTGGTCTGGTTGACTTCTGG; SPARC:
sense GCGAGCTGGATGAGAACAACAC, antisense GTG
GAPDH was used as endogenous control to normalize for
differences in the amount of total RNA in each sample.
Amplification reactions were conducted in a 96-well plate in
a final volume of 20 μl per well containing 10 μl of 1X SYBR
Green Supermix (BioRad), 2 μl of template, 300 pmol of
each primer, and each sample was analyzed in triplicate. The
cycle threshold (Ct) was determined and gene expression
levels relative to that of GAPDH were calculated by the 2-ΔΔCt
SDS-zymography. MMP-2 protein levels and activity were
assessed by SDS-zymography in cell culture supernatants.
Culture media were mixed 3:1 with sample buffer (containing
10% SDS). Samples (5 μg total protein per sample) were run
under non-reducing conditions without heat denaturation onto a
10% polyacrylamide gel (SDS-PAGE) co-polymerized with
1 mg/ml of type I gelatin. The gels were run at 4˚C. After
SDS-PAGE, the gels were washed twice in 2.5% Triton X-100
for 30 min each and incubated overnight in a substrate buffer
at 37˚C (Tris-HCl 50 mM, CaCl25 mM, NaN30.02%, pH 7.5).
The matrix metalloproteinase (MMP) gelatinolytic activity
was detected after staining the gels with Coomassie brilliant
blue R250, as clear bands on a blue background (9).
GAGLIANO et al: In vitro ANALYSIS OF GLIOMA-ASTROCYTE INTERACTION
Western blotting. Cells were lysed in Tris-HCl 50 mM pH 7.6,
150 mM NaCl, 1% Triton X-100, 5 mM EDTA, 150 mM
MgCl2, 1% SDS, protease inhibitors and 1 mM sodium
orthovanadate. Total proteins (20 μg) were diluted in SDS-
sample buffer, loaded on a 10% SDS-polyacrylamide gel,
separated under reducing and denaturing conditions at 80 V,
and transferred at 90 V to a nitrocellulose membrane in
0.025 M Tris, 192 mM glycine, 20% methanol, pH 8.3. After
electroblotting, the membranes were air-dried and blocked
for 1 h. After being washed in TBST (TBS/Tween-20, 0.05%),
membranes were incubated for 1 h at room temperature in
monoclonal antibody to GFAP (1:500 in TBST, Santa Cruz
Biotechnology, Santa Cruz, CA). After washing, membranes
were incubated in HRP-conjugated rabbit anti-mouse serum
(1:80,000 dilution in TBST, Sigma). To investigate CX43
protein levels, membranes were reacted with polyclonal
anti-CX43 (2 μg/ml in TBST, Zymed) and HRP-conjugated
anti-rabbit serum (1:20000 dilution, Sigma, Milan, Italy). To
confirm equal loading, membranes were reprobed by
monoclonal antibody to ·-tubulin (1:2000 dilution, Sigma).
Immunoreactive bands were revealed using the Amplified
Opti-4CN or the Opti-4CN substrate Bio Rad (BioRad).
SPARC protein levels were assessed in cell culture
medium. Total proteins (25 μg) were run on 10% SDS-PAGE
gels and electro-blotted as described above. Membranes were
incubated in anti-SPARC monoclonal antibody (1:100 in PBST,
Novocastra), followed by HRP-conjugated anti-mouse serum
(1:6000 dilution, Sigma), and revealed by Opty-4CN substrate.
The lysis and immunoreactive bands were analyzed by densito-
metric scanning (UVBand, Eppendorf).
Statistical analysis. All tests were run in duplicate. The
GraphPad Prism version 5.0 software package (GraphPad
Software) was used for the statistical analysis. Data are
expressed as mean ± standard error (SEM), and were
analyzed by one-way analysis of variance (ANOVA)
followed by the Student-Neumann-Keuls test. p<0.05 was
GFAP expression. The immunofluorescence analysis revealed
that AA co-cultivated with U87 (Fig. 1A, panel b) are strongly
positive for GFAP, compared to AA alone (Fig. 1A, panel a),
suggesting that glioma/astrocyte interaction induces
activation of AA. By contrast, U87 cells (Fig. 1A, panel c)
are weakly immunoreactive, confirming a down-regulation of
GFAP in malignant gliomas. This finding was confirmed by
Western blot analysis (Fig. 1B and C), showing a tendency in
GFAP up-regulation in co-cultured AA, compared to AA,
and a significant decrease of GFAP immunoreactivity in U87,
compared to AA and co-cultured AA (p<0.001 for U87 vs.
AA, AA+1.25x105U87, AA+2.5x105U87).
ONCOLOGY REPORTS 22: 1349-1356, 2009
Figure 1. (A) Immunofluorescence analysis of GFAP in AA (a), AA
co-cultured with U87 (b), and U87 (c). (B) Representative Western blot
analysis for GFAP. (C) Bar graphs showing GFAP protein levels in AA,
AA co-cultured with U87, and in U87 glioma cells. Changes in GFAP
expression are normalized on tubulin protein levels. Data are reported as
densitometric units after scanning of the immunoreactive bands. Values are
means ± ESM. *p<0.001 for U87 vs. AA, AA+1.25x105U87, AA+2.5x105
Figure 2. (A) Bar graphs showing MMP-2 mRNA levels in AA, AA
co-cultured with U87, and in U87 glioma cells. Changes in mRNA are
normalized on GAPDH gene expression. Values are means ± ESM. (B)
Representative gelatin zymogram of MMPs in serum-free conditioned
supernatants. The lytic bands correspond to proMMP-2. (C) Bar graphs
showing proMMP-2 activity in serum-free conditioned media after densito-
metric analysis of lytic bands following SDS-zymography. Data are expressed
as densitometric units and are means ± ESM. °p<0.001 for U87 vs.
AA+1.25x105U87, AA+2.5x105U87. *p<0.001 for AA vs. AA+1.25x105
U87, AA+2.5x105U87, U87. **p<0.001 for AA+1.25x105U87 vs.
AA+2.5x105U87. ***p<0.05 for U87 vs. AA+2.5x105U87.
MMP-2 and TIMP-2 expression. MMP-2 mRNA levels tended
to increase in AA co-cultivated with U87, and were highly
up-regulated in U87 (p<0.001 for U87 vs. AA+1.25x105U87,
AA+2.5x105U87) (Fig. 2A). Conversely, SDS-zymography
of cell supernatants revealed that the glioma/astrocyte
interaction strongly up-regulated MMP-2 at the protein level
(Fig. 2B and C). We observed a significant increase of
MMP-2 activity in all supernatants from co-cultured AA,
compared to AA alone (p<0.001 for AA vs. AA+1.25x105
U87, AA+2.5x105U87 and U87). Interestingly, there is a
significant increase of MMP-2 activity in supernatants from
co-cultured AA depending on the number of U87 cells
seeded in the insert of the transwell system (p<0.001 for
AA+2.5x105U87 vs. AA+1.25x105U87). The highest
activity was evident in U87 supernatants (p<0.05 for U87 vs.
AA). An opposite pattern was evident for TIMP-2 gene
expression (Fig. 3A), that slightly tended to decrease in AA
co-cultivated with U87, showing the lower expression in U87
(p<0.001 for U87 vs. AA; p<0.05 for U87 vs. AA+1.25x105
U87 and AA+2.5x105U87).
TGF-ß1 gene expression. TGF-ß1 gene expression was similar
in AA alone and in co-cultured AA, and tended to be slightly
up-regulated in U87 (Fig. 3B).
SPARC gene and protein expression. SPARC gene expression
(Fig. 4A) tended to decrease in AA co-cultivated with U87,
showing lower expression in glioma cells (p<0.001 for U87
vs. AA; p<0.05 for U87 vs. AA+1.25x105U87 and
AA+2.5x105U87). SPARC protein levels were similarly
expressed in cell lysates obtained from AA, co-cultured AA,
and U87 (Fig. 4B). By contrast, SPARC protein levels in
supernatants displayed an opposite pattern. SPARC was
almost undetectable in AA alone, but its expression was
up-regulated in co-cultured AA and in U87 (p<0.001 for AA
vs. U87; p<0.01 for AA vs. AA+2.5x105U87; p<0.05 for
AA+1.25x105U87 vs. U87) (Fig. 4C).
CX43 expression. CX43 expression was evaluated both at the
mRNA and at the protein level. Gene expression analysis
shows that CX43 tended to be down-regulated in AA co-
cultured with U87, reaching the lowest expression in U87
glioma cells (p<0.01 for U87 vs. AA) (Fig. 5A). Western blot
analysis demonstrates the pattern of CX43 protein expression
(Fig. 5B). Three isoforms of CX43 could be separated on
SDS-PAGE and visualized after immunoblotting as three
distinct bands: a non-phosphorylated isoform (NP), and the
GAGLIANO et al: In vitro ANALYSIS OF GLIOMA-ASTROCYTE INTERACTION
Figure 3. Bar graphs showing TIMP-2 (A) and TGF-ß1 (B) mRNA levels
in AA, AA co-cultured with U87, and in U87 glioma cells. Changes in
mRNA are normalized on GAPDH gene expression. Values are means ±
ESM. *p<0.01 for U87 vs. AA. **p<0.05 for U87 vs. AA+1.25x105U87,
Figure 4. (A) Bar graphs showing SPARC mRNA levels in AA, AA
co-cultured with U87, and in U87 glioma cells. Changes in mRNA are
normalized on GAPDH gene expression. Values are means ± ESM. (B) Bar
graphs showing SPARC protein levels in cell lysates. Data are reported as
densitometric units after scanning of the immunoreactive bands. Values
are means ± ESM. (C) Bar graphs showing SPARC protein levels in cell
supernatants. Data are reported as densitometric units after scanning of the
immunoreactive bands. Values are means ± ESM. *p<0.001 vs. AA.
**p<0.05 vs. AA+1.25x105U87, AA+2.5x105U87. °p<0.001 vs. U87.
°°p<0.01 vs. AA+2.5x105U87. °°°p<0.05 for U87 vs. AA+1.25x105U87.
two phosphorylated isoforms of the protein (P1 and P2).
CX43-NP isoform (open gap junction) immunoreactivity was
high in AA, tended to decrease in co-cultured AA, and was
completely undetectable in U87. In the same samples the
CX43-P isoform (closed gap junction) tended to increase in
co-cultured AA. Interestingly, glioma/astrocyte interaction
tended to decrease the CX43-NP/CX43-P ratio in co-cultured
AA, compared to AA alone (p<0.05 for U87 vs. AA,
AA+1.25x105U87, AA+2.5x105U87) (Fig. 5B and C),
suggesting that CX43 posttranscriptional phosphorylation
influences the open/close status of the gap junction.
Glioblastoma cells are characterized by cellular heterogeneity
(10) and, due to their exceptional migratory capacity, they
are able to diffusely infiltrate normal brain (11), making total
surgical removal impossible; therefore, patients have a poor
prognosis, even in response to multidisciplinary treatment
strategies including surgery, radiotherapy and chemotherapy
(12,13). Intra-parenchymal invasion of glioma occurs in a
non-destructive manner (3), suggesting that glioma cells may
establish relationships of cooperation with their environment
and in particular with resident brain astrocytes. We focused
our attention on the characterization of astrocyte phenotype
co-cultured with U87 glioma cells, in order to enlighten
how glioma and astrocytes interact, thus providing a possible
mechanism of glioma progression.
GFAP is the intermediate filament protein marker of
astroglial cells, whose assembly/disassembly status is
regulated by phosphorylation, thus affecting cytoskeletal
network, cell morphology, motility, differentiation, growth,
and mitosis (14). GFAP is also present in astroglial cell nuclei
as the major component of the nuclear lamina. With increasing
astrocytic malignancy, a progressive loss of GFAP production
was reported (15).
In our study, most glioblastoma cells exhibited GFAP
immunofluorescence with variations in intensity, and they
showed only a diffuse and weak fluorescent signal for GFAP.
This is consistent with the previously reported reduced GFAP-
positivity in high-grade astrocytomas (16), and a progressive
loss of GFAP production described with increasing astocytic
malignancy (17). A relationship between increased GFAP
protein expression and suppression of glial tumor growth
(18), inhibition of cell-cycle progression, and decreased
proliferation was previously reported (19).
By contrast, our data show that co-cultured astrocytes
tended to overexpress GFAP, compared to astrocytes cultured
alone. This evidence suggests that U87 may trigger astrocyte
activation and therefore a modification of their phenotype. A
strong GFAP immunopositivity was previously reported
in vitro (20) and in astrocytes at the periphery around the
tumor and, interestingly, these glial cells express primarily
MMP-2 and VEGF (21).
Degradation of the ECM in the cell microenvironment
is thought to be critical for tumor migration and invasion.
MMP-2 is one of the major gelatinolytic MMPs produced in
the glioblastomas (22). MMP-2 is the gelatinase mainly
involved in extracellular matrix remodeling during tumor
invasion, allowing tumor cells to break down basement
membranes and diffuse into the surrounding tissues.
Its expression is related to the glioma malignancy grade
(23-25), and it was suggested that the modulatory effects of
MMP-2 on tumor cell migration could be attributed to both
matrix degradation and alterations of adhesion receptors on
the cell surface (26), as well as to its involvement in actin
mediated motility (27).
The role of TIMP-2, the main inhibitor of MMP-2, is
complex. A decrease in overall TIMP-2 expression is
reported with increasing grade in gliomas, particularly when
measured relative to increasing MMP-2 (28,29).
We observed a significant MMP-2 protein up-regulation
in astrocytes co-cultivated with U87, and the concomitant
tendency to down-regulation of its inhibitor, TIMP-2,
suggesting that U87 could induce a modification of the
phenotype of neighboring astrocytes; this phenotype is con-
sistent with a higher ability in remodeling the extracellular
matrix of the tumor microenvironment and thus favoring and
allowing tumor invasion. At the same time, however, since
MMP-2 is reported to be increased in relation to glioma
ONCOLOGY REPORTS 22: 1349-1356, 2009
Figure 5. (A) Bar graphs showing CX43 mRNA levels in AA, AA co-cultured
with U87, and in U87 glioma cells. Changes in mRNA are normalized on
GAPDH gene expression. Values are means ± ESM. (B) Representative
Western blot analysis for CX43. The antibody identifies a lower molecular
weight non-phosphorylated isoform (NP), and the two phosphorylated
isoforms of the protein (P1 and P2). (C) Bar graphs showing CX43-NP/
CX43-P ratio in AA, AA co-cultured with U87, and in U87 glioma cells.
Changes in GFAP expression are normalized on tubulin protein levels. Data
are reported as densitometric units after scanning of the immunoreactive
bands. Values are means ± ESM. *p<0.01 vs. AA+1.25x105U87, AA+2.5x105
malignancy grade, we hypothesize that MMP-2 over-
expression induced by U87 in co-cultured astrocytes may be
related to a modification of astrocyte phenothype, and that
U87 may trigger a malignant transformation of the resident
surrounding normal astrocytes.
SPARC is an important multifunctional glycoprotein
that influences several biological processes including cell-
extracellular matrix interaction, and cell differentiation,
migration and proliferation. SPARC is highly expressed in
human brain tumors (30,31) and in gliomas, regardless of
grade, SPARC was highly expressed in the peripheral regions
of tumors as well as adjacent to the normal brain tissue (30).
Increased SPARC expression correlates with glioma invasion
in vitro (32,33) and in vivo (33-35). This is accomplished
by up-regulation and activation of MMP-2, and a possible
mechanism involved in increased SPARC-dependent ECM
degradation via MMP may involve galectin-3 (36). However,
the effects of SPARC are complex, and the ability of SPARC
to promote invasion depends on the level of its secretion and
to the local tumor environment (35). Both low and higher
levels of SPARC correlated with more proliferation and
promoted a level of ECM adherence conductive for migration
in vivo, whilst intermediated levels would promote stronger
adherence and less migration. In addition, it was hypothesized
that increased glioma invasion would be accompanied by
delayed cell growth (35,37). SPARC has a suppressive effect
on glioma cell proliferation, delaying cell growth in vitro
(37), and delaying tumor growth in rat brains in vivo (35).
Intracellular SPARC down-regulation, possibly as a con-
sequence of promoter methylation, was previously reported
in human ovarian cancer (38), inducing a delay of cell cycle
progression; so the higher level of SPARC is related to the
greater percentage of cells in G0/G1 or G2/M phase (37).
High levels of SPARC are therefore consistent with a reduced
ability to enter the cell cycle and, as a consequence, with a
reduced cell proliferation. If cell cycle progression is inhibited
by SPARC, it is likely that SPARC functions also in cell
differentiation. In fact, SPARC has exhibited prominent
expression during the terminal differentiation of cultured
human keratinocytes (39), and is believed to regulate terminal
differentiation of lens epithelial cells (40).
It was suggested that SPARC may elicit different effects
extra- and intracellularly. It is known that SPARC exerts its
de-adhesive effects extracellularly through an antagonistic
action involving integrin signalling (41). As previously
reported (42), SPARC may be translocated in the nucleus
after endocytosis, where the protein may exert its effects on
the cell cycle.
Our data show that astrocytes express higher levels of
SPARC mRNA whilst astrocytes co-cultured with U87
tended to down-regulate SPARC gene expression, suggesting
that U87 could interact with astrocytes inducing their pro-
liferation and, possibly, their dedifferentiation, and therefore
triggering their malignant transformation. To confirm this
hypothesis we also analyzed SPARC protein levels in cell
lysates. Our results show that AA, co-cultured AA and U87
display similar intracellular SPARC protein levels, suggesting
that the proliferative potential of AA is not affected by U87.
By contrast, U87 induced SPARC protein level up-regulation
in supernatants of co-cultured astrocytes, suggesting that U87
affect the phenotype of normal resident astrocytes rendering
them more able to remodel the tumor micro-environment and
more invasive and, very likely, malignant. This hypothesis is
consistent with the results of a recent study showing that
intracerebral injection of glioma cells transfected with
SPARC siRNA in nude mice resulted in the formation of a
non-invasive tumor (33).
The comprehension of the very complex mechanisms
regulating the different extra- and intracellular SPARC
protein levels should be useful in understanding the role of
SPARC as a possible target for glioma treatment.
CX43. Gap junctions are ubiquitously expressed by many
types of cells, and they are involved in diverse processes
related to the regulation of major cell functions including
proliferation, differentiation, and cell homeostasis (43-46).
CX43 is the most abundant gap junction connexin expressed
in astrocytes and glioma cells, and a reduced expression of
CX43 has been demonstrated in numerous cancer cells
(47,48), including glioma cells. Loss of gap junction
coupling has been implicated in malignant transformation:
CX43 expression decreases as the grade of glioma/
astrocytoma worsens and an inverse correlation between
CX43 expression and tumor malignancy grade exists (44,49-
51). Furthermore, forced expression of CX43 has been
shown to inhibit proliferation of C6 glioma cells both in vitro
and in vivo, suggesting that the CX43 gene can act as a tumor
suppressor (52,53). Since gap junctions are closely associated
with the control of cell growth (54), growing evidence
suggests that CX43 may function as a tumor suppressor gene
CX43 was previously analyzed in direct co-culture
experimental models (4,55) and in brain tumor tissues (51),
and resulted strongly up-regulated at the margin of tumor
invasion as a result of a direct astrocytes-glioma interaction.
In our experimental model, we found decreased CX43 gene
expression and down-regulation of the non-phosphorylated/
phosphorylated CX43 ratio in co-cultured astrocytes,
possibly exerted by soluble factors produced by U87.
Interestingly, CX phosphorylation represents a further
mechanism involved in gap junction intercellular com-
munication, since it may favour its internalization and
degradation (56). On the basis of our results, we hypothesize
that glioma/astrocyte interaction modulates CX43 gene and
protein expression, in particular acting on the post-
translational regulation of CX43 by phosphorylation. We
observed a decrease of the NP isoform (open gap junction)
of CX43. This result suggests that glioma cells influence
neighbouring astrocytes very likely modifying their phenotype
towards a malignant-like phenotype in which CX43 is down-
The tumor microenvironment is increasingly recognized
as shaping tumor phenotypes, including invasiveness, and
therefore the assessment of the relative roles of glioma and
astrocytes in glioma progression is useful. Considered as a
whole, our results suggest that U87 may elicit phenotype
modifications of the surrounding resident astrocytes very
likely mediated by soluble factors. Our experimental model
suggests that glioma/astrocytes interaction could induce
astrocyte activation and their phenotype modification.
GAGLIANO et al: In vitro ANALYSIS OF GLIOMA-ASTROCYTE INTERACTION
On the basis of our results, we can hypothesize that these
phenotypic changes may be consistent with a malignant
transformation of astrocytes triggered by glioma cells, and/or
may render the brain parenchyma permissive to glioma
Our study contributes to the understanding of the mech-
anisms that regulate glioma histogenesis and tumorigenesis,
and may be useful in the development of new therapeutic
strategies to improve the prognosis of patients with incurable
C.C. is recipient of a fellowship (SFRH/BD/15899/2005)
from the Fundação para a Ciência e a Tecnologia (FCT). This
work was supported via the Italian Multiple Sclerosis
Foundation (FISM, grants 2004/R/15 to S.P.), the National
Multiple Sclerosis Society (NMSS, partial grants RG-4001-
A1 to S.P.) and Banca Agricola Popolare di Ragusa (BAPR,
unrestricted grant to S.P.).
1. Kleihues P, Soylemezoglu F, Schauble B, Scheithauer BW and
Burger PC: Histopathology, classification, and grading of
gliomas. Glia 15: 211-221, 1995.
2. Schiffer D, Cavalla P, Dutto A and Borsotti L: Cell proliferation
and invasion in malignant gliomas. Anticancer Res 17: 61-69,
3. Bernstein JJ: Local invasion and intraparenchymal metastasis
of astrocytomas. Neuropathol Appl Neurobiol 22: 421-424,
4. Oliveira R, Christov C, Guillamo JS, de Boüard S, Palfi S,
Venance L, Tardy M and Peschanski M: Contribution of gap
junctional communication between tumor cells and astroglia to
the invasion of the brain parenchyma by human glioblastomas.
BMC Cell Biol 6: 7, 2005.
5. Forsyth PA, Laing TD, Gibson AW, Rewcastle NB, Brasher P,
Sutherland G, Johnston RN and Edwards DR: High levels of
gelatinase-B and active gelatinase-A in metastatic glioblastoma.
J Neurooncol 36: 21-29, 1998.
6. Saxena A, Shriml LM, Dean M and Ali IU: Comparative
molecular genetic profiles of anaplastic astrocytomas/
glioblastomas multiforme and their subsequent recurrences.
Oncogene 18: 1385-1390, 1999.
7. Le DM, Besson A, Fogg DK, Choi KS, Waisman DM,
Goodyer CG, Rewcastle B and Yong VW: Exploitation of
astrocytes by glioma cells to facilitate invasiveness: a mechanism
involving matrix metalloproteinase-2 and the urokinase-type
plasminogen activator-plasmin cascade. J Neurosci 23: 4034-4043,
8. Vescovi AL, Parati EA, Gritti A Poulin P, Ferrario M, Wanke E,
Frölichsthal-Schoeller P, Cova L, Arcellana-Panlilio M,
Colombo A and Galli R: Isolation and cloning of multipotential
stem cells from the embryonic human CNS and establishment
of transplantable human neural stem cell lines by epigenetic
stimulation. Exp Neurol 156: 71-83, 1999.
9. Kleiner DE and Stetler-Stevenson WG: Quantitative zymography:
detection of picogram quantities of gelatinases. Anal Biochem
10. Misra A, Chattopadhyay P, Dinda AK, Sarkar C, Mahapatra AK,
Hasnain SE and Sinha S: Extensive intratumor heterogeneity in
primary glial tumors as a result of locus non-specific genomic
alterations. J Neurooncol 48: 1-12, 2000.
11. Lipinski CA, Tran NL, Bay C, Kloss J, McDonough WS,
Beaudry C, Berens ME and Loftus JC: Differential role of
proline-rich tyrosine kinase 2 and focal adhesion kinase in
determining glioblastoma migration and proliferation. Mol
Cancer Res 1: 323-332, 2003.
12. Daumas-Duport C, Scheithauer B, O'Fallon J and Kelly P:
Grading of astrocytomas. A simple and reproducible method.
Cancer 62: 2152-2165, 1988.
13. Shapiro WR: Current therapy for brain tumors: Arch Neurol 56:
14. Inagaki M, Nakamura Y, Takeda M, Nishimura T and Inagaki N:
Glial fibrillary acidic protein: dynamic property and regulation
by phosphorylation. Brain Pathol 4: 239-243, 1994.
15. Duffy PE, Huang YY and Rapport MM: The relationship of
GFAP to cell shape, motility and differentiation of human
astrocytoma cells. Exp Cell Res 139: 145-157, 1982.
16. Shiras A, Bhosale A, Shepal V, Shukla R, Baburao VS,
Prabhakara K and Shastry P: A unique model system for tunor
progression in GBM comprising two developed human neuro-
epithelial cell lines with differential transforming potential and
coexpressing neuronal and glial markers. Neoplasia 5: 520-532,
17. Rutka JT, Murakami M, Dirks PB, Hubbard SL, Becker LE,
Fukuyama K, Jung S, Tsugu A and Matsuzawa K: Role of
glial filaments in cells and tumors of glial origin: a review. J
Neurosurg 87: 420-430, 1997.
18. Toda M, Miura M, Asou, Sugiyama I, Kawase T and Uyemura K:
Suppression of glial tumor growth by expression of glial
fibrillary acidic protein. Neurochem Res 24: 339-343, 1999.
19. Asklunnd T, Appelskog IB, Ammerpohl O, Ekstrom TJ and
Almqvist PM: Histone deacetylase inhibitor 4-phenilbutyrate
modulates glial fibrillary acidic protein and connexin 43
expression, and enhances gap-junction communication, in
human glioblastoma cells. Eur J Cancer 40: 1073-1081, 2004.
20. Lal PG, Ghirnikar RS and Eng LF: Astrocyte-astrocytoma cell
line interactions in culture. J Neurosci Res 44: 216-222, 1996.
21. Nagashima G, Suzuki R, Asai J and Fujimoto T: Immuno-
histochemical analysis of reactive astrocytes around glioblastoma:
an immunohistochemical study of postmortem glioblastoma
cases. Clin Neurol Neurosurg 104: 125-131, 2002.
22. Rao JS: Molecular mechanisms of glioma invasiveness: the role
of proteases. Nat Rev Cancer 3: 489-501, 2003.
23. Nakada M, Nakamura H, Ikeda E, Fujimoto N, Yamashita J,
Sato H, Seiki M and Okada Y: Expression and tissue localization
of membrane-type 1, 2, and 3 matrix metalloproteinases in
human astrocytic tumors. Am J Pathol 154: 417-428, 1999.
24. Nakagawa T, Kubota T, Kabuto M, Sato K, Kawano H,
Hayakawa T and Okada Y: Production of matrix metallopro-
teinases and tissue inhibitor of metalloproteinases-1 by human
brain tumors. J Neurosurg 81: 69-77,1994.
25. Rao JS, Steck PA, Mohanam S, Stetler-Stevenson WG, Liotta LA
and Sawaya R: Elevated levels of M(r) 92,000 type IV collagenase
in human brain tumors. Cancer Res 53: 2208-2211, 1993.
26. Deryugina EI, Bourdon MA, Luo GX, Reisfeld RA and
Strongin A: Matrix metalloproteinase-2 activation modulates
glioma cell migration. J Cell Sci 110: 2473-2482, 1997.
27. Ogier C, Bernard A, Chollet AM, LE Diguardher T, Hanessian S,
Charton G, Khrestchatisky M and Rivera S: Matrix metallo-
proteinase-2 (MMP-2) regulates astrocyte motility in connection
with the actin cytoskeleton and integrins. Glia 54: 272-284, 2006.
28. Kachra Z, Beaulieu E, Delbecchi L, Mousseau N, Berthelet F,
Moumdjian R, Del Maestro R and Béliveau R: Expression of
matrix metalloproteinases and their inhibitors in human brain
tumors. Clin Exp Metastasis 17: 555-566, 1999.
29. Merzak A, Parker C, Koochekpour S, Sherbet GV and
Pilkington GJ: Overexpression of the 18A2/mts1 gene and down-
regulation of the TIMP-2 gene in invasive human glioma cell
lines in vitro. Neuropathol Appl Neurobiol 20: 614-619, 1994.
30. Rempel SA, Golembieski WA, Ge S, Lemke N, Elisevich K,
Mikkelsen T and Gutiérrez JA: SPARC: a signal of astrocytic
neoplastic transformation and reactive response in human
primary and xenograft gliomas. J Neuropathol Exp Neurol 57:
31. Rempel SA, Ge S and Gutiérrez JA: SPARC: a potential
diagnostic marker of invasive meningiomas. Clin Cancer Res 5:
32. Golembieski WA, Ge S, Nelson K, Mikkelsen T and Rempel SA:
Increased SPARC expression promotes U87 glioblastoma
invasion in vitro. Int J Dev Neurosci 17: 463-472, 1999.
33. Seno T, Harada H, Kohno S, Teraoka M, Inoue A and Ohnishi T:
Downregulation of SPARC expression inhibits cell migration and
invasion in malignant gliomas. Int J Oncol 34: 707-715, 2009.
34. Rich JN, Shi Q, Hjelmeland M, Cummings TJ, Kuan CT,
Bigner DD, Counter CM and Wang XF: Bone-related genes
expressed in advanced malignancies induce invasion and
metastasis in a genetically defined human cancer model. J Biol
Chem 278: 15951-15957, 2003.
ONCOLOGY REPORTS 22: 1349-1356, 2009
35. Schultz C, Lemke N, Ge S, Golembieski WA and Rempel SA:
Secreted protein acidic and rich in cysteine promotes glioma
invasion and delays tumor growth in vivo. Cancer Res 62:
36. McClung HM, Thomas SL, Osenkowski P, Toth M, Menon P,
Raz A, Fridman R and Rempel SA: SPARC upregulates MT1-
MMP expression, MMP-2 activation, and the secretion and
cleavage of galectin-3 in U87MG glioma cells. Neurosci Lett
419: 172-177, 2007.
37. Rempel SA, Golembieski WA, Fisher JL, Maile M and Nakeff A:
SPARC modulates cell growth, attachment and migration of
U87 glioma cells on brain extracellular matrix proteins. J
Neurooncol 53: 149-160, 2001.
38. Said N and Motamed K: Absence of host-secreted protein acidic
and rich in cysteine (SPARC) augments peritoneal ovarian
carcinomatosis. Am J Pathol 167: 1739-1752, 2005.
39. Ford R, Wang G, Jannati P, Adler D, Racanelli P, Higgins PJ
and Staiano-Coico L: Modulation of SPARC expression during
butyrate-induced terminal differentiation of cultured human
keratinocytes: regulation via a TGF-beta-dependent pathway.
Exp Cell Res 206: 261-275, 1993.
40. Bassuk JA, Birkebak T, Rothmier JD, Clark JM, Bradshaw A,
Muchowski PJ, Howe CC, Clark JI and Sage EH: Disruption of
the Sparc locus in mice alters the differentiation of lenticular
epithelial cells and leads to cataract formation. Exp Eye Res 68:
41. Motamed K and Sage EH: Regulation of vascular morphogenesis
by the matricellular protein SPARC. Kidney Int 51: 1383-1387,
42. Gooden MD, Vernon RB, Bassuk JA and Sage EH: Cell cycle-
dependent nuclear location of the matricellular protein SPARC:
association with the nuclear matrix. J Cell Biochem 74: 152-167,
43. Goldberg GS, Moreno AP and Lampe PD: Gap junctions between
cells expressing connexin 43 or 32 show inverse permselectivity
to adenosine and ATP. J Biol Chem 277: 36725-36730, 2002.
44. Huang RP, Hossain MZ, Sehgal A and Boynton AL: Reduced
connexin 43 expression in high-grade human brain glioma cells.
J Surg Oncol 70: 21-24, 1999.
45. Huang R, Lin Y, Wang CC, Gano J, Lin B, Shi Q, Boynton A,
Burke J and Huang RP: Connexin 43 suppresses human
glioblastoma cell growth by down-regulation of monocyte
chemotactic protein 1, as discovered using protein array
technology. Cancer Res 62: 2806-2812, 2002.
46. Huang RP, Fan Y, Hossain MZ, Peng A, Zeng ZL and
Boynton AL: Reversion of the neoplastic phenotype of human
glioblastoma cells by connexin 43 (cx43). Cancer Res 58:
47. Dermietzel R and Spray DC: Gap junctions in the brain: where,
what type, how many and why? Trends Neurosci 16: 186-192,
48. Yamasaki H and Naus CC: Role of connexin genes in growth
control. Carcinogenesis 17: 1199-1213, 1996.
49. Pu P, Xia Z, Yu S and Huang Q: Altered expression of Cx43 in
astrocytic tumors. Clin Neurol Neurosurg 107: 49-54, 2004.
50. Shinoura N, Chen L, Wani MA, Kim YG, Larson JJ, Warnick RE,
Simon M, Menon AG, Bi WL and Stambrook PJ: Protein and
messenger RNA expression of connexin43 in astrocytomas:
implications in brain tumor gene therapy. J Neurosurg 84:
51. Soroceanu L, Manning TJ and Sontheimer H: Reduced expression
of connexin-43 and functional gap junction coupling in human
gliomas. Glia 33: 107-117, 2001.
52. Omori Y and Yamasaki H: Mutated connexin43 proteins inhibit
rat glioma cell growth suppression mediated by wild-type
connexin43 in a dominant-negative manner. Int J Cancer 78:
53. Zhu D, Kidder GM, Caveney S and Naus CC: Growth retardation
in glioma cells cocultured with cells overexpressing a gap
junction protein. Proc Natl Acad Sci USA 89: 10218-10221,
54. Goodenough DA, Goliger JA and Paul DL: Connexins,
connexons, and intercellular communication. Ann Rev Biochem
65: 475-502, 1996.
55. Zhang W, Couldwell WT, Simard MF, Song H, Lin JH and
Nedergaard M: Direct gap junction communication between
malignant glioma cells and astrocytes. Cancer Res 59: 1994-2003,
56. Laird DW: Connexin phosphorylation as a regulatory event
linked to gap junction internalization and degradation. Biochim
Biophys Acta 1711: 172-182, 2005.
GAGLIANO et al: In vitro ANALYSIS OF GLIOMA-ASTROCYTE INTERACTION