[CANCER RESEARCH58, 149—158.
January 1, 19981
Spreading and Migration of Human Glioma and Rat C6 Cells on Central Nervous
System Myelin in Vitro Is Correlated with Tumor Malignancy and Involves a
Metalloproteolytic
Verena
R. Amberger,2
Activity1
Tamara
Hensel,
Nobuyoshi
Ogata,
and Martin
E. Schwab3
Brain Research Institute, University of Zurich, 8029 Zurich. Switzerland (V. R. A., T. H., M. E. SI, and Department of Neurosurgery, University Hospital of Zurich, 8091 Zurich,
Switzerland (N. 0.1
ABSTRACT
Malignant gliomas infiltrate the brain preferentially along myelinated
fiber tracts. Central nervous system (CNS) myelin, however, contains
inhibitory proteins that block axon regeneration, neurite outgrowth, and
cell spreading of astrocytes and fibroblasts. We tested 5 human brain
tumor cell lines, 1 rat brain tumor cell line, and 29 short-term cultured
specimens from human brain tumors for their abifity to spread and
migrate on a CNS myelin substrate. Low-grade and piocytic astrocytoma,
ependymoma,
medulloblastoma,
and meningioma cell lines as well as
primary cultures were strongly sensitive to the inhibitory proteins present
in the CNS
myelin.
In contrast,
glioblastomas,
anaplastlc
astrocytomas,
and oligodendrogliomas were able to spread and migrate on CNS myelin
coated culture dishes, demonstrating that within the gliomas, the abffity to
overcome
the inhibitory
effects of the CNS myelln is correlated
with the
grade ofmallgnancy ofthe original tumor. Cell spreading of glioblastomas
and anaplastic astrocytomas specifically on a CNS myelin substrate was
strongly inhibited by the metalloprotease blocker O-phenanthroline and
the peptide derivative carbobenzoxy-Phe-Ala-Phe-Tyr-amlde, whereas
blockers for serine, aspartyl, and cysteine proteases had no effect. Enzy
matic peptide degradation assays revealed the presence of a phosphor
amidon-sensitive and tlilorphan-lnsensitive metalloproteolytic activity in
the plasma membranes of hiaji-grade glioma cells. These results suggest a
crucial involvement
process of invasive
of a membrane-bound
metalloendoprotease
in the
migration of malignant gliomas along CNS white
matter fiber tracts.
INTRODUCTION
Malignant primary brain tumors, especially gliomas, are character
ized by extensive infiltration of the surrounding normal brain tissue
(1). It is therefore of primary importance to achieve a better under
standing of the mechanisms involved in brain tumor cell migration.
Single tumor cells migrate along blood vessels and seem to be arrested
by the glial limitans. Basement membrane proteins, which are present
around large blood vessels, have been shown to inhibit proliferation,
and some of them stimulate migration of glioma cell lines (2—7)and
primary glioma cells (8). The main route for glioma dissemination,
however, is along white matter fiber tracts and can lead to distant
spread, e.g., through the corpus callosum into the other hemisphere
(1, 9, 10).
One important component of the brain white matter is myelin. In
addition to structural proteins, myein contains specific membrane
proteins, the neurite growth inhibitors NI3S/250, which inhibit axon
regeneration, neurite outgrowth, and cell spreading of astrocytes and
fibroblasts (1 1—13).These inhibitory effects can be blocked with
monoclonal antibodies in vitro and in vivo (14—16).The fact that the
main route of malignant brain tumor infiltration is along white matter
fiber tracts suggests that gliomas use a specific mechanism to over
come the effect of these neurite growth-inhibitory proteins. Rat C6
glioma, a chemically induced glial brain tumor cell line (17), shows
rapid and diffuse infiltration of brain tissue in vivo (18, 19), very
similar to human glioblastomas. In vitro, C6 glioma cells use a
metalloproteolytic activity to spread and migrate on CNS4 tissue (20)
and could be blocked specifically by the metalloprotease blocker
O-phenanthroline or the newly designed tetrapeptide Cbz-Phe-Ala
Phe-Tyr-amide, whereas blockers of serine, aspartyl, or cysteine pro
teases had no effect (13, 20). We characterized the C6 enzyme as a
membrane-bound metalloendoprotease with a Mr between 60,000 and
90,000 that is highly sensitive to phosphoramidon but insensitive to
thiorphan and to blockers of matrix metalloproteases (13). A similar
metalloproteolytic activity combined with the capability to spread and
migrate on CNS myelin and white matter was found recently in CNS
glial precursor cells (21). These cells are known to be highly migra
tory cells in the developing and adult CNS (22—25).
In the present study, we tested five human cell lines and one rat
glioma cell line in three different in vitro assays: (a) a spreading assay
on CNS myelin extract; (b) a migration assay on CNS myelin extract;
and (c) a peptide degradation assay to evaluate metalloproteolytic
activity. Additionally, short-term cultures from human brain tumor
biopsies were tested for their ability to spread on purified CNS
myelin. We collected 24 specimens from glial tumors and 5 specimens
from meningiomas as a nonglial control tissue. We show evidence for
a direct correlation between the in vivo grade of malignancy of a given
glioma and the ability to spread and migrate on a CNS myelin
substrate. Cell spreading and migration of malignant glioma cells
were inhibited on CNS myelin by the general metalloprotease blocker
O-phenanthroline and by the tetrapeptide Cbz-Phe-Ala-Phe-Tyr
amide, suggesting the involvement of a metalloprotease similar to that
of C6 glioma cells and glial precursor cells.
MATERIALS
Reagents
1 Supported
by
the
Swiss
National
Science
Foundation
Grants
3 1-2998
1 .90
and
31-
45549.95, the Hartman-MUller Foundation (Zurich, Switzerland), and Regeneron Phar
maceuticals, Inc. (Tarrytown, NY). V. R. A. and T. H. contributed equally to this work.
2 Present
address:
Brain
Research
Laboratories,
London
Health
Science
Center,
and Materials.
Cell culture media were obtained
from Life
Technologies, Inc. (Basle, Switzerland); pepstatin A, leupeptin, and aprotinin
were obtained from Fluka (Buchs, Switzerland). The tetrapeptide Cbz-Phe
Ala-Phe-Tyr-amide
was synthesized
by Bachem
AG (BasIc,
Switzerland).
All
other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO).
The CSM (Apex Manufacturing, Inc.; Phoenix, AZ) used for the cell migration
assay was a kind gift of Dr. M. Berens (Neuro-Oncology Laboratory, Barrow
Neurological
Received 7/15/97; accepted 10/24/97.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
AND METHODS
Institute of St. Joseph's
Hospital,
Phoenix,
AZ).
Cell Lines. C6 glioma cells (passage 335) were a kind gift from Dr. D.
Monard (Friedrich Miescher Institute, Basle, Switzerland), U251 glioblastoma
cells (passage 580) were provided by Dr. R. Del Maestro (Brain Research
Laboratories,
London, Canada), and four permanent cell lines derived from
biopsy cultures
[IPSB-l8
(passage
35; Ref. 26), IPNT-H
(passage
10), IPPS-l
Vic
toria Campus, London, Ontario, N6A 405 Canada.
3 To
whom
requests
for
reprints
should
be
addressed,
at Brain
Research
University of Zurich, August-Forel-Strasse 1, 8029 Zurich, Switzerland.
I
(passage 15), and IPNN-8 (passage 15; Refs. 27—30)]
were provided by Dr. 0.
4 The
Institute,
abbreviations
used
are:
CNS,
central
nervous
system;
Cbz,
carbobenzoxy;
cell sedimentation manifold.
149
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
CSM,
@
@‘-‘@@.,@.@-.)I'/#
:
,
HUMAN AND RAT GLIOMAS ON CNS MYELIN
Pilkington (Institute of Psychiatry, London, United Kingdom). All cell lines
were grown as monolayer cultures in DMEM supplemented with 10% FCS,
penicillin (100 units/ml), and streptomycin (0.5 mg/ml) in the following called
culture medium under 5% CO2 at 37°C.Cells were trypsinized every fifth day
using 0.1% trypsin in Ca2@-free, Mg2@-free PBS and 0.025% EDTA.
anesulfonic acidIHCl buffer (pH 6.0), and stored at —70°C.
Plasma mem
branes were preincubated in the presence of a blocker mixture containing 10
p.M leupeptin, 10 p.M pepstatin A, and 1 mi@iphenylmethylsulfonyl
fluoride for
15 mm at 37°C.0-Phenanthroline and phosphoramidon were added to the
mixture when indicated. The reaction was started with 100 fmol of Cbz-Phe
Biopsy Tissue. Freshspecimens from primarybraintumorswere takenat
time of operation, and, in parallel, tissue samples were sent for neuropatho
logical diagnosis. The tissue was washed in ice-cold Ringer solution, and blood
Ala-Phe-'251-Tyr-amide,
and after
1 h of incubation
at 37°C, the components
in the incubation medium were separated by TLC, and iodinated peptides were
visualized by autoradiography. For quantification, the gel regions correspond
vessels and necrotic tissue were removed. After mincing the tissue in culture
ing to the bands on the autoradiogram
medium,
counter (LKB Wallace 1282 compu-gamma counter). Each experimental con
dition was performed in duplicate and was tested four times. The significance
was evaluated using the unpaired t test.
the cells were dissociated
size filter. After centrifugation,
by passing
them through
the cells were resuspended
a 70-sm
in culture
mesh
medium,
seeded on culture dishes (100 X 20 mm; Nunclon; Nunc, Roskilde, Denmark),
and kept under 9% CO2 at 37°Cfor 7—10days. The culture medium was
changed at day 2 and day 5 in culture. For the spreading assay, the cells were
trypsinized with 0.1% trypsin in Ca2@-free, Mg2@-free PBS and 0.025%
were excised
and counted
in a gamma
RESULTS
EDTA before plating.
Spreading of Human Glioma Cell Lines on Purified CNS My
Cell Spreading Assay and Quantification. CNS myelin was prepared elm. Five characterized human glioma cell lines generated from an
from rat spinal cord by sucrose density gradient centhfugation (12), extracted
ependymoma (IPPS-ll), a medulloblastoma (IPNN-8), a pilocytic
with buffer containing
20 mM Tris and 60 m@i 3-[(3-cholamidopropyl)dimeth
astrocytoma (IPNT-H; Refs. 27—30), an anaplastic astrocytoma
ylamino]-l-propanesulfonate,
and routinely tested for inhibitory
activity for
(IPSB-18; Ref. 26), and a glioblastoma (U25 1; Refs. 3 1 and 32) were
cell spreading with 3T3 fibroblasts (21). The myelin protein concentration
tested for their ability to attach and spread on culture dishes coated
resulting in 80% inhibited fibroblasts was used for the spreading and migration
assays. CNS myelin extract or extraction buffer alone (as control) was coated with a CNS myelin protein extract containing the neurite growth
on 4-well culture dishes (Greiner, Nurtingen, Germany) and incubated over inhibitory proteins. The tumor cells were plated either on control
substrate coated with extraction buffer or on culture dishes coated
night at 4°C. After washing the plates three times with HBSS, 10,000 cells/well
(I cm2) were plated on coated culture dishes, and as soon as the attached cells with CNS myelin proteins. After 1 h of incubation, most of the
on the control substrate were well spread, the assay was stopped by the IPNT-H cells (pilocytic astrocytoma; Figs. la and 2A), the IPSB-l8
addition of4% formaldehyde in 0.1 M phosphate buffer supplemented with 4%
cells (anaplastic astrocytoma; Figs. lc and 2A), and the U251 glio
sucrose. Protease blockers (3 @M
aprotinin, 10 p@ pepstatin A, and 10 @M blastoma cells (Fig. 2A) were firmly attached to the control substrate,
leupeptin) were added directly to the cell suspensions from lOX solutions 10
and the spreading process was advanced, as reflected by the flattening
rein before plating. Cell spreading was quantified by counting flat, spread cells
of the cells. On the CNS myelin-coated dishes, the pilocytic astrocy
the total well surface). Each experimental condition was performed in triplicate toma cells (IPNT-H) attached poorly, and spreading was strongly
inhibited (Figs. lb and 2A), whereas the anaplastic astrocytoma cells
(cell lines) or duplicate (primary cells). The statistical significance was eval
(IPSB- 18; Figs. ld and 14) as well as the U25 1 glioblastoma cells
uated using the unpaired t test.
Cell Migration Assay and Quantification. Teflon-printedmicroscope
versus
round
cells
inI2randomly
chosen
areas
ofeach
well
( about
12%
of
slides (Cell-Line
Associates,
Inc., Newfield,
NJ), which are subdivided
into 10
wells, were precoated with poly-L-lysine(5 pg/well) overnight. After blocking
the unspecific binding with 1%BSA in PBS for 30 mm, the slides were washed
control
.,
with Ca2tfree,
Mg2@-free PBS, dried, and stored at room temperature.
Some
slides were additionally
coated with rat CNS myelin extract (15 p@g/well) as
@
CNS myelin
[email protected].
.
. o••
described above. The migration assay was performed according to Berens et a!.
..‘
•
.
(3) with some modifications:coatedslides were covered with 70 ,,d of culture
medium each; the CSMs were placed on the slides; and 1 pi of cell suspension
(2,000 cells) was placed in the sedimentation cylinder. Protease blockers (1 p.M
aprotinin, 10 p.M leupeptin, 10 p.M pepstatsn A, 100 p.M 0-phenanthroline, and
40 p.M Cbz-Phe-Ala-Phe-Tyr-amide)
were added
directly
•@:@•
0•
L*@
-.
•
.
•
.
...
)O.
C
to the cell suspen
@?
0
.Q@--•0:@4
;‘0. ,@
sions from lOX solutions 10 mm before seeding. After incubating the cells for
2 h at 37°Cunder 5% CO2. the CSMs were removed, fresh culture medium
containing protease blockers was added, and the area covered with cells was
.,
@.
..I•b.0':
—C
recorded immediately and 6, 12, and 24 h after seeding with a black and white
video camera
(CCD;
Sony)
tive; Olympus,
IMT-2).
image analysis
system
attached
to an inverted
Cell migration
(Image
Access;
microscope
was measured
image-l),
(x 2.5 objec
and quantified
and the values obtained
Ii
with an
@6@)
at 2 h
after seeding were used as a baseline and subtracted from the values measured
at later time points. The numbers represent the mean measurement of three
wells, and each experiment was performed three times. The significance was
evaluated using the unpaired t test.
@
@
Nasma Membrane Preparation
and
Degradation Assay. Plasma membrane preparation, radioiodination
-@
@,bS,@‘.•0-,
-,0C
of Cbz
Phe-Ala-Phe-Tyr-amide, and the degradation assay were performed as de
scribed previously (13). Briefly, Cbz-Phe-Ala-Phe-Tyr-amide was iodinated
with Na'25!
(500 p.Ci/p.l;
2000 Ci/mmol)
using chloramin
I and purified
d/,,@@4
by
sequentially running the sample on a C18-Sep-Pak column (Waters, Millipore,
Fig. 1. Anaplastic astrocytomas (IPSB-18) but not pilocytic astrocytomas (IPNT-H) are
Volketswil, Switzerland) and on a Cl8 reversed-phase high-performance liq
uid chromatography
activity
of the purified
column
probe
(Brownlee,
Applied
was 2000 cpm/fmol
Biosystem).
The
able to attach to and spread on CNS myelin. IPNT-H (a and b) and IPSB-18 (c and tO cells
were plated on control substrate (a and c) and on CNS myelin (b and d)-coated culture
dishes (10,000 cells/cm2). In contrast to pilocytic astrocytoma cells (IPNT-H), for which
specific
peptide.
CNS myelinis a nonpermissivesubstrate(b), anaplasticastrocytomacells (IPSB-18)
Cell membranes from the different tumor cell lines were purified on a
discontinuous
sucrose gradient, resuspended
attach to and spread on myelin (ti). All cells were fixed after 1 h, and pictures were taken
in 20 mM 2-(N-morpholino)eth
at a magnification of x 170.
150
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
HUMAN AND RAT GLIOMAS ON CNS MYEUN
A
100
0
80
@
control
CNS myelln
0
60
.@
4,
Q.
4,
40
a-
20
0
U251
IPSB-18
IPNT-H
IPPS-11
IPNN-8
B
100
0
80
IPSB-18 on control
IPSB-18 on CNS myelln
U251 on control
U251 on CNS myelln
a,
4,
C.,
60
‘a
a-
a,
40
20
0
blocker
cocktail
o-phenanthroline
cbz-phe-ala
phe-tyr-amide
Fig. 2. Pilocyticastrocytoma(IPWT-H),ependymoma(IPPS-11),and medulloblastoma(JPNN-8)cell spreadingon CNS myelin-coatedculturedishes is inhibited,whereas
glioblastoma (U251) and anaplastic astrocytoma (IPSB-18) spread equally well on CNS myelin and on control substrate and can be inhibited by metalloprotease blockers specifically
on CNS myein. A, cells were plated on control substrate (open bars) and on CNS myelin-coated culture dishes (hatched bars). After 1 (IPNT-H, IPSB-18, and U251), 5 (IPPS-l 1),
and 3 h (IPNN-8),the cellswerefixed,andthe percentageof spreadcellswasdetermined.B, beforeplating,U251and IPSB-18cellswerepreincubatedfor 15mmwitha blocker
mixture (10 @LM
pepstatin A, 10 gist leupeptin, and 3 psi aprotinin) and, where indicated, with 200 ,.u.i 0-phenanthroline or 20 @sstCbz-Phe-Ala-Phe-Tyr-amide.
After 1 h, the cells
were fixed and quantified. The values are the means of counted cells in 12 randomly chosen areas of 3 experiments. Bars, SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001.
(Fig. 2A) were attached and spread, demonstrating that the high-grade
glioma cells were not affected by the inhibitory proteins present in the
CNS myelin. Medulloblastoma cells (IPNN-8) and ependymoma cells
(IPPS-ll) slowly attached to and spread on the control substrate (3
and 5 h, respectively) and, at that time point, showed significantly
reduced cell spreading on CNS myelin (Fig. 2A; IPNN-8, P < 0.001;
IPPS-ll, P < 0.05). These results show that glioblastoma and ana
plastic astrocytoma cells have a mechanism that enables them to
spread on CNS myelin under conditions that are strongly inhibitory
for tumor cells derived from pilocytic astrocytomas, ependymomas,
and medulloblastomas.
Spreading of Glioblastoma and Anaplastic Astrocytoma Cells
on CNS Myelin Is Inhibited by Metalloprotease Blockers but not
by Blockers for Serine, Aspartyl, and Cysteine Proteases. IPSB-l8
astrocytoma and U25l glioblastoma cells were plated on CNS myelin
coated culture dishes in the presence of a mixture containing blockers
151
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
HUMAN AND RAT GLIOMAS ON CNS MYELIN
,
‘
f@'@
Z@e‘I':
\-@
0
•::@@..b--@‘.@-
@/0aA@
;i‘4
‘-@[email protected]‘
.@‘,p@.,,@.
b:
@
@/d@4
.@.
t!
eq
Fig. 3. Cell spreading of anaplastic astrocytoma cells on a CNS myelin substrate can
be inhibited by metalloprotease blockers, which do not affect cell spreading on the control
substrate. IPSB-l8 cells were preincubated for 15 mm with a blocker mixture containing
@
10 @LM
pepstatin A, 10 @.LM
leupeptin, 3 @aM
aprotinin (a), and 200 psi 0-phenanthroline (b)
or 20
cbz-Phe-Ala-Phe-Tyr-amide (c and d) and plated on CNS myelin-coated culture
dishes (a-c) and on control substrate (ti). After I h of incubation, the cells were fixed, and
pictures were taken at a magnification of X 170.
for serine (3 p.Maprotinin), aspartyl (10 p.Mpepstatin A), and cysteine
proteases (10 p.M leupeptin). After 1 h, the cells were attached and
spread equally well on both control and CNS myelin substrates (Figs.
2B and 3a), showing that this mixture of protease blockers did not
affect IPSB-l8 and U25l cell spreading on neither CNS myelin nor on
control substrate. In contrast, 200 p.M O-phenanthroline (Figs. 2B and
3b), a general metalloprotease blocker, or 20 p.M Cbz-Phe-Ala-Phe
Tyr-amide (Figs. 2B and 3c), a protease blocker inhibiting specifically
the metalloproteolytic activity involved in C6 glioma cell spreading
on CNS myelin (13), both strongly blocked cell spreading on CNS
myelin but had no effect on control substrate (Figs. 2B and 3d). The
strong blocking effects by O-phenanthroline and Cbz-Phe-Ala-Phe
Tyr-amide but not by the blocker mixture inhibiting serine, cysteine,
and aspartyl proteases suggest the involvement of a metalloproteolytic
activity that enables malignant astrocytoma cells to spread on CNS
myelin.
Spreading of Primary Human Brain Tumor Cells on CNS
Myelin. Short-term cultures from 29 specimens collected at the time
of operation, which were neuropathologically identified as 11 glio
blastoma multiforme, 2 anaplastic astrocytomas, 3 low-grade astro
cytomas, 3 pilocytic astrocytomas, 5 oligodendrogliomas, and 5 men
ingiomas, were tested for their ability to spread on CNS myelin-coated
culture dishes. Cell spreading of pilocytic and low-grade astrocytomas
was significantly inhibited by the CNS myelin (P < 0.001 and
P < 0.01, respectively) compared to anaplastic astrocytomas and
glioblastomas, which easily attached to and spread on the CNS myelin
substrate (Fig. 4A). These results reveal a correlation between the
ability to spread on CNS myelin and the in vivo malignancy of the
original astroglial tumors. Oligodendroglioma cells were not affected;
they attached and spread equally well on CNS myelin substrate and on
control substrate (Fig. 4A). Meningiomas originating from the arach
noid cells of the meninges were strongly impaired in cell spreading on
CNS myelin, and this inhibition was highly significant compared to
glioblastoma cell spreading (P < 0.001; Fig. 4A).
Spreading of Glioblastoma Multiforme on CNS Myelin Is Im
paired by Metalloprotease Blockers. Seven short-term cultured hu
man glioblastoma multiforme specimens were tested for cell spread
ing on CNS myein and on control substrate in the presence of various
protease blockers. The presence of a mixture blocking serine, aspartyl,
and cysteine proteases (3 p.Maprotinin, 10 p.Mpepstatin A, and 10 ,.aM
leupeptin) did not affect cell spreading on either CNS myelin or
control substrate (data not shown). In contrast, the metalloprotease
blockers O-phenanthroline (150 p.M) or Cbz-Phe-Ala-Phe-Tyr-amide
(20 pM) inhibited cell spreading on CNS myein (Fig. 4B), and these
effects were highly significant (P < 0.001), whereas there was no
significant effect on cell spreading on control substrate. From these
results, we can conclude that primary glioblastoma cells use a metal
loproteolytic activity to spread on a CNS myelin substrate, reflecting
similar results obtained with glioma cell lines. Cell spreading of three
oligodendroglioma specimens in the presence of both metalloprotease
blockers was not significantly affected (data not shown), suggesting a
different mechanism for oligodendroglioma cell spreading on CNS
myelin.
Cell Migration of Five Human Cell Lines and One Rat Cell
Line on CNS Myelin. Five human (U251, IPSB-18, IPNT-H, IPPS
11, and IPNN-8) cell lines and one rat (C6) brain tumor cell line were
tested in a migration assay using slides that were coated with either
CNS myelin extract or poly-L-lysine as a control substrate. Areas
covered with migrating cells were measured after 2, 6, 12, and 24 h
(Fig. 5), and the value measured at 2 h was defined as the baseline
level, which was subtracted from the values measured at later time
points. At 24 h after plating on the control substrate U251 glioblas
toma, the anaplastic astrocytoma (IPSB-18) and the ependymoma
cells (IPPS-l 1) showed the highest migration activity (2.6 mm2),
followed by the C6 glioma (2.1 mm2) and the medulloblastoma cells
(IPNN-8; 1.7 mm2; Fig. 6). The pilocytic astrocytoma cells (IPNT-H)
showed very little migration (0.8 mm2; Fig. 6). After 24 h on the CNS
myelin substrate, pilocytic astrocytoma, medulloblastoma,
and
ependymoma cells showed hardly any migration activity (0.3, 0.4, and
0.5 mm2, respectively; Fig. 6), and migration of the anaplastic astro
cytoma cells was reduced by 50% compared to that of the control
substrate (1.2 mm2). In contrast, C6 and U251 glioblastoma cells were
not affected by CNS myelin. The cells migrated equally well on both
CNS myelin and control substrate (Fig. 6). These results demonstrate
that CNS myelin is a nonpermissive substrate for cell migration of
pilocytic astrocytomas, ependymomas, and medulloblastomas. The
sensitivity to the inhibitory effect of CNS myelin is correlated with the
malignancy of the original tumor, being lowest in the most malignant
glial tumors.
Cell Migration of Glioblastoma and Anaplastic Astrocytoma
Cells Is Inhibited in the Presence of Metalloprotease
Blockers.
C6, U251, and IPSB-18 cells were placed on CNS myelin-coated
slides and on control slides. The presence of a mixture blocking serine
(1 @LM
aprotinin), cysteine (10 @aM
leupeptin), and aspartyl proteases
(10 p.M pepstatin A) did not affect cell migration on both substrates
(Fig. 6). In contrast, the presence of 100 p.M O-phenanthroline
strongly inhibited cell migration of the anaplastic astrocytoma and the
two glioblastoma cell lines on CNS myelin-coated wells (Figs. 5 and
6). Most of the cells remained round, and actively migrating cells
were rare. The blocking effects of the tetrapeptide Cbz-Phe-Ala-Phe
Tyr-amide were much less pronounced. Cell migration of all three cell
lines was impaired within the first 12 h but was hardly affected after
24 h (Fig. 6). A similar short-term inhibition by Cbz-Phe-Ala-Phe
Tyr-amide was seen on cell spreading of C6 cells (13) and cell
spreading and migration of glial precursor cells (21) on CNS myelin.
Cbz-Phe-Ala-Phe-Tyr-amide
is a substrate for the C6 metalloprotease
and is degraded with a slow time course (13). The partial and transi
152
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
HUMAN AND RAT GLIOMAS ON CNS MYELIN
A
100
80
a,
@
60
C.,
@0
a,
0.
‘a
aa,
Fig. 4. Primary high-grade astrocytomas and oligo
dendrogliomas are able to spread on a CNS myelin
substrate, whereas pilocytic and low-grade astrocyto
‘5)55
and meningiomas
are strongly
inhibited.
40
20
A, short
term cultured cells from 3 piocytic, 3 low-grade, and
2 anaplastic astrocytomas; 11 glioblastomas; 5 oligo
dendrogliomas; and 5 meningiomas were plated on
control substrate (open bars) and on CNS myelin
0
coated culture dishes (hatched bars) and incubated for
glio
blastoma
90-135 missat 37°C,fixed, and quantified. B, seven
short-term cultured glioblastoma specimens were pre
incubated for 15 nun in the presence of a blocker
mixture (10 @Lsi
pepstatin A, 10 g@sileupeptin, and 3
@5M
aprotinin) and either 150 pat 0-phenanthroline or
20 ,.ssi cbz-Phe-Ala-Phe-Tyr-amide,
plated on control
astro
oligoden
cytoma I droglioma
menin
gloma
B
substrate (open bars) and on CNS myelin-coated cul
1 00
ture dishes (hatched bars), and incubated for 90 mm at
37°C.The values represent the means of counted cells
in 12 randomly chosen areas of 2 experiments. Bars,
@
astro
astro
cytoma Ill cytoma II
0 control
CNS myelln
SD; **, P < 0.01; ***, P < 0.001.
80
a,
4)
C.,
60
.@
4,
‘a
a0.
a,
40
20
0
o-phenanthroline
cbz-phe-ala
phe-tyr-amide
tory blocking effect of Cbz-Phe-Ala-Phe-Tyr-amide
on cell spreading
is therefore most probably due to peptide degradation. The strong
blocking effect of O-phenanthroline and the transient inhibition by the
tetrapeptide suggest a crucial involvement of a metalloproteolytic
activity in the migration process of malignant gliomas.
Malignant Gliomas Degrade Cbz-Phe-Ala-Phe-'@I-Tyr-amide
Using a Phosphoramidon-sensitive
and Thiorphan-insensitive
Metalloprotease.
To investigate the type of metalloproteolytic ac
tivity involved in cell migration and spreading of high-grade
gliomas, plasma membranes were prepared from U25 1, IPSB- 18,
IPNT-H, IPPS- 11, and IPNN-8 cells. The plasma membrane pro
teins were incubated with 100 fmol of Cbz-Phe-Ala-Phe-'25I-Tyr
amide for 1 h. To restrict peptide degradation to metallodependent
enzymes, a mixture containing blockers for serine, aspartyl, and
cysteine proteases (1 m@i phenylmethylsulfonyl
fluoride, 10 @M
leupeptin, and 10 p.M pepstatin A) was added 15 mm before the
incubation. U251 plasma membranes degraded 75 fmol peptide/
h/10 @gprotein (Fig. 7A). Membrane proteins from the anaplastic
astrocytoma (IPSB-18) and the pilocytic astrocytoma (IPNT-H; 60
and 55 fmol/h) as well as membranes from the ependymoma
(IPPS-1 1) and the medulloblastoma (IPNN-8; 48 and 56 fmol/h)
showed a significantly lower degradation rate compared to U25 1
membranes (Fig. 7A). The peptide derivative Cbz-Phe-Ala-Phe
Tyr-amide inhibits glioblastoma cell spreading in short-term as
153
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
@
@
:
-
•@
-@
HUMAN AND RAT GLIOMAS ON CN5 MYEUN
2h
6h
Fig. 5. U251 cells migrate on a CNS myelin
substrate, a process that is inhibited by the metal
loprotease
blocker 0-phenanthroline.
U25 I cells
were preincubated in the absence or presence of 80
laM0-phenanthroline for 15 mm and placed into the
CSMs. At 2 (immediately after removing the
CSMs), 6, 12,and 24 h, cells were fixed and stained
@
with H&E, and pictures were taken at a magnifica
tion of X35 (details, x350). Bar, 300 p.m.
12h @.
.@ ,
@
. . .:•..@
24h
@,,
..J,J
@
.
@
- -, ;@“:@@‘
,..-, ‘-.@-
,
says (Figs. 2B, 4B, and 6; Ref. 13) but is also a slowly degraded
metalloprotease substrate (13).
The chelator O-phenanthroline
(80 p.M) inhibited the degrada
tion of Cbz-Phe-Ala-Phe-'25I-Tyr-amide
by all tested cell mem
branes (U25 1, IPSB-18, IPNT-H, IPPS-1 1, and IPNN-8; Fig. 7B),
demonstrating a metallodependent peptide degradation. To further
characterize the metalloproteolytic activity, we tested two specific
metalloprotease blockers, phosphoramidon and thiorphan, that spe
cifically block endopeptidase 24.1 1 (enkephalinase; Refs. 33 and
34), whereas the metalloendoprotease
found in C6 glioma cell
membranes is sensitive to phosphoramidon but not to thiorphan
(13). Ten @.tM
phosphoramidon
strongly inhibited Cbz-Phe-Ala
Phe-'251-Tyr-amide degradation by C6 and U251 cell membranes
(Fig. 7B) and, to a lesser extent, inhibited degradation by anaplas
tic astrocytoma membranes. In contrast, a blocking effect of phos
phoramidon on the degrading activities in pilocytic astrocytoma,
ependymoma, and medulloblastoma
membranes was prominent
(Fig. 7B). In the presence of 10 p.Mthiorphan, Cbz-Phe-Ala-Phe
I25I-Tyr-amide
degradation
by all tested
cell
membranes
was
not
significantly affected (Fig. 7B). These results demonstrate a major
involvement of a phosphoramidon-sensitive
but thiorphan-insensi
tive metalloprotease in Cbz-Phe-Ala-Phe-'25I-Tyr-amide
degrada
tion by glioblastomas and high-grade astrocytomas, reflecting sim
ilar results obtained with C6 cell membranes (13). This enzyme
activity correlates with the grade of malignancy of the tested cell
lines. The degrading activity present in pilocytic astrocytoma,
ependymoma, and medulloblastoma
membranes was sensitive to
O-phenanthroline but not to phosphoramidon or to thiorphan.
DISCUSSION
In the present study, we tested five human cell lines, one rat brain
tumor cell line, and short-term cultured human brain tumor specimens
in their ability to spread and migrate on rat CNS myelin protein
extract that is strongly inhibitory for cell spreading of astrocytes and
fibroblasts (11—13).In our experiments, CNS myelin turned out to be
a nonpermissive substrate for cells from glial tumors, such as the
pilocytic and low-grade astrocytomas, ependymomas, and medullo
blastomas, and for cells from a mesenchymal tumor, such as the
memngioma, whereas anaplastic astrocytoma, glioblastoma, and rat
C6 glioma cells rapidly spread and migrated on this substrate. This
ability seems to correlate with the grade of malignancy of the glial
tumors and was strongly reduced by the metalloprotease blocker
O-phenanthroline and the peptide Cbz-Phe-Ala-Phe-Tyr-amide
but
not by blockers for serine, cysteine, and aspartyl proteases. Oligoden
droglioma cells spread rapidly on CNS myein-coated culture dishes
but were not affected by metalloprotease blockers or any other tested
protease blocker.
In the recent past, research on glioma invasion had been focused
mainly on the interaction between tumor cells and the extracellular
matrix and the production of matrix metalloproteases by the tumor
itself or neighboring cells (for reviews, see Refs. 35 and 36). Extra
cellular matrix proteins such as collagens, fibronectin, and laminin are
predominantly found in the basement membrane around blood yes
sels, at the leptomemnges, and in the glial limitans externa (37). In the
gray and white matter of the CNS, glycosaminoglycans, also called
lecticans (38), hyaluromc acid, and tenascin are expressed (35, 39, 40)
154
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
@
@
@1@
@;
1
HUMAN AND RAT GUOMAS ON CN5 MYELIN
4
4.
a
a
E
E
a
3
0
(5
C6
control
myelin
myellnlblocker cocktail
myelln/o-phenanthrollne
myelln/cbz-FAFY-amlde
a control
IPNT-H
a myelln
3 .
a
myelln/blocker
cocktaIl
a,
‘a
2
2
@0
a,
‘a
a0)
E
@
I
1
1@
0
0
_ @-@-@____________
4
@T-r
6
AlT
‘
ITT
12
24
12
24
U251
3
(‘4J
E
E
@
hL
‘5
a,
a-
2
‘a
a,
‘a
a-
1
E
:
@
@ir:
0:@L@
@
.@:..
6
12
24
6
4
4
IPSB-18
E
E
3
3
‘a
a,
a-
‘5
2
@0
a,
‘a
a-
a,
E
@
1
0 L@_L.
iL:
6
0
12
24
6
hours
12
24
hours
Fig.6. Timecourseofgliomaandmedulloblastoma
cellmigrationona controlandona CNSmyein substrate.C6andU25l glioblastomaandtheanaplasticastrocytoma(IPSB.18)
cells are able to migrate on a CNS myclin substrate, whereas piocytic astrocytoma (JPNT-h'), ependymoma (IPPS-11). and medulloblastoma (IPNN-8) cell migration was strongly
inhibited by CNS mydlin. Glioblastoma
(C6 and U251) and anaplastic astrocytoma (JPSB-18) cell migration is strongly inhibited by 0.phenanthroline
and is partially impaired by
Cbz.Phc-Ala-Pbe-Tyr-aniide.
Cellswereplacedon controlsubstrateandon CNSmyclin-coatedculturedishes,and the migrateddistancewas measuredat 6, 12,and 24 h. When
indicated. the cells were preincubated with a blocker mixture (10 p.si pepsiatin A, 10 p.si leupeptin, and 1 p.st aprotinin) and either 100 p.si 0-phenanthroline
or 40 p.si
Cbz.Phe-AIa-Phe-Tyr-amide.
The values represent the mean measurement of three slides and three independent experiments. Bars, SD; s. P < 0.05; **, P < 0.01; ***, P < 0.001.
155
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
HUMANAND RAT GLIOMASON CN5 MYELIN
C
A
a,
0
a-
0.
a,
80-
0
.C
a,
.D
**
60-
**
:L
***
0.
a,
0.
40 -
@0
a,
.a
‘a
a.
a,
a,
20-
@0
Fig. 7. Plasma membrane protein extracts from
@
glioblastoma (U251), anaplastic astrocytoma (IPSB
@
18), pilocytic astrocytoma (1PNT-ff), ependymoma
(IPPS-11), and medulloblastoma (IPNN-8) degrade
the peptide
using a
0
E
U251
metalloproteolytic
activity.Tenp.gof plasmamem
10 p.s pepstatin
I
a
IPSB-18 IPNT-H IPPS-11 IPNN-8
o-phenanthrollne
phosphoramidon
0
0 thiorphan
B
bra.ne extracts prepared from different cell lines were
preincubated with a blocker mixture containing 10
I.SM leupeptin,
-I.-
0-
100'
A, and 1 inst phen
ylmethylsulfonyl fluoride (A) and with either 80 psi
O-phenanthroline.
10 psi phosphoramidon,
or 10 psi
II@
thiorphan (B) for 15 mm, followed by a 1-h incuba
tion with 100 fmol of Cbz-Phe-Ala-Phe-1@I-Tyr
amide at 37'C. The degradation products were sep
arated by TLC and quantified. The values represent
the mean of five experiments. Bars, SD; **,
P < 0.01; ***, P < 0.001. Significance in B is
compared to control values in A.
I
80@
I
4,
0.
4,
0.
60@
@0
@0
4,
*
‘a
a-
a,
4,
*@*
40@
*
*
*
*
*
*
*
*
20@
@
0@
I _
U251
IPSB-18
1* _1*
IPNT-H
IPPS-11
behavior could not be inhibited by antibodies against the neuronal cell
that interact with a multitude of membrane proteins of neuronal and
glial cells. Two membrane proteins of oligodendrocytes and myelin,
the neurite growth inhibitors N135/250, are strongly inhibitory mole
cules for neurons, fibroblasts, and astrocytes (11—13).These negative
signals in the CNS myelin have been found in many higher vertebrates
including rats, cattle, and humans (11, 12, 41).
In previous studies, we have shown that C6 glioma cells, a highly
invasive rat brain tumor cell line, rapidly spread and migrate in CNS
tissue and on a CNS myelin substrate, and that they express a
membrane-bound metalloprotease that neutralizes the inhibitory com
ponents present in the CNS myelin (13, 20), especially a highly
purified inhibitory protein from bovine bNI-220.5 Recently, Giese et
aL (42) could show that four different human glioma cell lines,
including U25l cells, and primary cell suspensions from two tumors
were able to attach to and migrate on CNS myelin. This cellular
adhesion
molecule
NCAM
and
was
only
slightly
inhibited
with
anti-@3l integrin antibodies, demonstrating that these membrane
constituents are not major players in these processes.
In our present experiments, human glioblastomas, anaplastic astro
cytomas, and rat C6 glioma rapidly spread and migrated on a CNS
myelin substrate under conditions that were inhibitory for cell spread
ing and migration of pilocytic and low-grade astrocytomas, ependy
momas, and medulloblastomas. Ependymoma cell spreading and mi
gration were strongly inhibited on a CNS myelin substrate, and the
pilocytic astrocytoma cells showed hardly any migration activity on
either control or CNS myelin substrate; both effects might be cx
planed by the noninvasive, benign behavior of these tumors in vivo
(1). In contrast, the low-grade astrocytomas and the medulloblastomas
are known as malignant, invasive tumors (1); however, in our exper
iments, their cell spreading and migration were significantly inhibited
on a CNS myelin substrate. Low-grade astrocytomas show a number
of differences in their invasive character compared to glioblastomas,
5 ‘F.
Hensel, V. R. Amberger, and M. E. Schwab. A metalloprotease, partially purified
from C6 glioma cells, inactivates the myelin-associated
neutralized by antibodies, submitted for publication.
*
growth inhibitors and can be
156
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
HUMAN AND RAT GLIOMAS ON CNS MYELIN
e.g., the invasion of the brain parenchyma is a much slower process
that lacks any evidence of tissue destruction (1). Furthermore, the
expression levels of the proteases cathepsin B and D, which are high
in glioblastomas, are low in grade II astrocytomas (43, 44), whereas
the tissue inhibitors for metalloproteinases 1 and 2 are highly cx
pressed in low-grade astrocytomas and are almost absent in glioblas
toma tissue (45). These findings, together with our results, suggest
that low-grade astrocytomas use a different mechanism to invade
surrounding brain tissue. Medulloblastomas are highly malignant
brain tumors in young children that mainly migrate along blood
vessels and invade the subarachnoid space; however, infiltration of the
brain parenchyma has been found as well (1). The reason for the
inhibitory effect of adult CNS myelin on medulloblastoma cell
spreading and migration might be found in the differences between
the pediatric and adult CNS, e.g., the lower content of myelin.
Myelination is a late developmental process; during the first year of
age, only about 20% of the nerve structures in the CNS are myeli
nated, during the second year, it reaches about 60%, and most parts of
the CNS are not myelinated before the age of 7 years (46).
These results demonstrate a correlation between the ability to
spread and migrate on CNS myelin and the grade of malignancy of a
glial tumor. Cell spreading and migration of glioblastomas and ana
plastic astrocytomas on CNS myelin were highly sensitive to the
metalloprotease blocker O-phenanthroline, whereas blockers for ser
me, aspartyl, and cysteine proteases had no effect. This blocker profile
suggests the involvement of a metalloproteolytic activity that enables
glioblastomas and anaplastic astrocytomas to spread and migrate on a
CNS myelin substrate. Rossi et aL (47) have investigated the influ
ence of sequential in vitro passaging on human brain tumor cell lines,
especially on IPSB-18, IPNN-8, and IPPS-1 1, which were used for the
present study. They described changes in the enzyme secretion of four
matrix metalloproteases, but without a defmed pattern. In our study,
we found that the results obtained with the short-term cultured brain
tumor specimens clearly corresponded to those of the cell lines,
showing the validity of these cell lines as a model. The only minor
difference was in the required inhibitory concentration of the metal
loprotease blocker O-phenanthroline (200 p.M for cell lines U251 and
IPSB-18 versus 150 pM for primary human glioblastoma cells). A
similar difference had been found earlier between a glial precursor
cell line (CG-4) and primary glial precursor cells (21).
In a previous study, we characterized a metalloendoprotease that is
expressed on plasma membranes of rat C6 glioma cells and that
enables these cells to spread and migrate on a CNS myelin substrate
(13). This metalloproteolytic activity degrades the peptide Cbz-Phe
Ala-Phe-'@I-Tyr-amide with a slow time course and is highly sensi
tive to phosphoramidon but is insensitive to thiorphan (both are very
potent blockers for endopeptidase 24.11, enkephalinase; Refs. 33 and
34) and to matrix metalloprotease blockers. This blocker profile
differentiates the C6 enzyme from most known metalloproteases (13).
The peptide Cbz-Phe-Ala-Phe-Tyr-amide,
which blocks the spread of
C6 cells and partially blocks their cell migration on a CNS myelin
substrate, acts as a competitive substrate for a metalloprotease and is
itself slowly degraded (13), which could explain the transient blocking
effect on cell migration found in this study; 20 @aM
Cbz-Phe-Ala-Phe
Tyr-amide strongly blocked cell spreading of malignant gliomas
within 1 h of incubation, but even a higher concentration of the
tetrapeptide (40 @.LM)
only slightly affected cell migration within the
first 12 h and had almost no effect at 24 h, suggesting that the peptide
was degraded by C6, U251, and anaplastic astrocytoma cells (IPSB
18). The results obtained in the peptide degradation assay confirmed
this assumption; plasma membranes of rat C6 and human U25 1 cells
degraded the iodinated peptide Cbz-Phe-Ala-Phe-125I-Tyr-amide,
whereas membranes from the anaplastic astrocytoma (IPSB-l8) and
the pilocytic astrocytoma (IPNT-H) degraded significantly less pep
tide per hour, showing a correlation between the peptide-degrading
activity and the grade of malignancy of the original tumor.
The peptide-degrading activities of all tested cell lines were inhib
ited by O-phenanthroline, identifying them as metallodependent ac
tivities. Further characterization of the peptide-degrading enzymes
revealed a high sensitivity to phosphoramidon in U25l glioblastoma
cells, very similar to results found with C6 cell membranes (13),
whereas in pilocytic astrocytoma (IPNT-H), ependymoma (IPPS-ll),
and medulloblastoma (IPNN-8), phosphoramidon had no major effect
on peptide degradation. Thiorphan did not affect peptide degradation
in all tested cell lines, differentiating the involved enzymes from
endopeptidase 24. 11 (enkephalinase).
In conclusion, spreading and migration of human primary glioma
cells and human and rat glioma cell lines on a CNS myelin substrate
are strongly dependent on the action of a membrane-bound metalloen
doprotease. This enzyme degrades the peptide Cbz-Phe-Ala-Phe-Tyr
amide and is phosphoramidon sensitive and thiorphan insensitive,
showing similarities to the meta.llodependent enzymes found in C6
glioma (13) and in glial precursor cells (21). The activity of the
metalloprotease seems to correlate with the grade of malignancy of
the original tumor and is highest in the most malignant glioma. Future
high-affinity blockers for this type of protease may allow us to inhibit
or significantly slow down the infiltration of white matter by malig
nant gliomas.
ACKNOWLEDGMENTS
We thank G. J. Pilkingion (Department of Neuropathology, Institute of
Psychiatry, London, United Kingdom) for providing four human tumor cell
lines and Prof. Y. Yonekawa, Dr. Könti,Dr. R. Bernais, and Dr. K. von
Ammon (Department of Neurosurgery, University Hospital, Zurich, Switzer
land) for providing tumor specimens. Special thanks are due to Dr. Dr. W.
Born (Calcium Laboratory, Orthopaedic Hospital Balgrist, University of Zur
ich, Switzerland) for iodinating the tetrapeptide and to R. Schoeb and R. Foster
for preparing the figures.
REFERENCES
1. Russel, D. C., and Rubinstein, L. Tumours of central neuroepithelial origin. in: E.
Arnold (ed), Pathology of Tumours of the Nervous System, Ed. 5., pp. 83—289.
London: Hodder & Stoughton, 1989.
2. Giese, A., Rief, M. D., Leo, M. A., and Berens, M. E. Determinants
of human
astrocytoma migration. Cancer Res., 54: 3897—3904,1994.
3. Berens, M. E., Rief, M. D., Lao, M. A., and Giese, A. The role ofextracellular matrix
in human astrocytoma migration and proliferation studied in a microliter scale assay.
Clin. Exp. Metastasis, 12: 405—415, 1994.
4. Koochekpour, S., Merzak, A., and Pilkington, G. I. Extracellular matrix proteins
inhibit proliferation, up-regulate migration and induce morphological changes in
human glioma cell lines. Eur. i. Cancer, 3/A: 375—380,1995.
5. Giese, A., Loo, M. A., Rief, M. D., Iran, N., and Berens, M. E. Substrates for
astrocytoma invasion. Neurosurgery (Baltimore), 37: 294—301,1995.
6. Giese, A., Lao, M. A., Iran, N., Hasken, D., Coons, S. W., and Berens, M. E.
Dichotomy of astrocytoma migration and proliferation. mt. I. Cancer, 67: 275—282,
1996.
7. Tysnes, B. B., Larsen, L. F., Ness, G. 0., Mahesparan, R., Edvardsen, K., Garcia
Cabrera, I,, and Bjerkvig, R. Stimulation of glioma cell migration by laminin and
inhibition by anti-a3 and anti-el integrin antibodies. tnt. J. Cancer, 67: 777—784,
1996.
8. Friedlander, D. R., Zagzag, D., Shiff, B., Cohen, H., Allen, i. C., Kelly, P. 1., and
Grumet, M. Migration of brain tumor cells on extracellular matrix proteins in vitro
correlates with tumor type and grade and involves av and @lintegrmns.Cancer Res.,
56: 1939—1947,1996.
9. Burger, P. C., Heinz, E. R., Shibata, T., and Kleihues, P. Topographic anatomy and
er correlations
in theuntreated
glioblastoma
multiforme.
J. Neurosurg.,
68: 698—
704. 1988.
10. Pedersen, P-H., Edvardsen, K., Garcia-Cabrera, I., Mahesparan, R., Thorsen, i.,
Mathisen, B., Rosenblum, M., and Bjerkvig, R. Migratory patterns of Iac-z trans
fected human glioma cells in the rat brain. mt. J. Cancer, 62: 767—771,1995.
11. Schwab, M. E., and Caroni, P. Oligodendrocytes and CNS myelin are nonpermissive
substratesfor neuritegrowthand fibroblastspreadingin vitro. I. Neurosci.,8:
2381—2393,
1988.
12. Caroni, P., and Schwab, M. E. Two membrane protein fractions from rat central
157
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
HUMAN AND RAT GLIOMAS ON CNS MYELIN
myelin with inhibitory properties for neurite growth and fibroblast spreading. i. Cell
30. Pilkingion, G. J., Akinwunmi, i., Ognjenovic, N., and Rogers, J. P. Differential
binding of anti-CD44 on human gliomas in vitro. Neuroreport, 4: 259—262,1993.
31. Bigner, D. D., Bigner, S. H., Ponten, J., Westermark, B., Mahaley, M. S., Ruoslahti,
E., Herschman, H., Eng, L F., and WiICStrand,C. J. Heterogeneity of genotypic and
Biol.,106: 1281—1288,
1988a.
13. Amberger, V. R., Paganeni, P. A., Seulberger, H., Eldering, J. A., and Schwab, M. E.
Characterization
of a membrane-bound
metalloendoprotease
of rat C6 glioblastoma
cells. Cancer Res., 54: 4017—4025,1994.
14. Caroni, P., and Schwab, M. E. Antibody against myelin-associated inhibitor of neurite
phenotypic characteristics
of fifteen permanent cell lines derived from human glio
mas. J. Neuropathol. Exp. Neurol., 40: 201—229,1981.
32. Ponten, J., and Macintyre, E. H. Long term culture of normal and neoplastic human
growth neutralizes non-permissive substrate properties of CNS white matter. Neuron,
1: 85—96,1988b.
glia. Acts Pathol. Microbiol. Scand., 74: 465-486, 1968.
33, Almenoff, J., and Orlowski, M. Membrane-bound kidney neutral metalloendopepti
15. Schnell, L., and Schwab, M. E. Axonal regeneration in the spinal cord produced by
an antibody against myelin-associated neurite growth inhibitors. Nature (Lend.), 343:
dase: interaction with synthetic substrates, natural peptides, and inhibitors. Biochem
isU)f, 22: 590—599, 1983.
34. Roques, B. P., Fournié-Zaluski, M-C., Soroca, E., Lecomte, J. M., Maifroy, B.,
269—272, 1990.
16. Schnell, L., Schneider, R., Kolbeck, R., Barde, Y-A., and Schwab, M. E. Neurotro
phin-3 enhances sprouting of corticospinal tract during development and after spinal
Llorens, C., and Schwartz, i-C. The enkephalinase inhibitor thiorphan shows antino
ciceptive activity in mice. Nature (Land.). 288: 286—288,1980.
cord lesion. Nature (Lend.), 367: 170—173,1994.
17. Bends, P., Lighthody, J., Sate, G., Levine, L, and Sweet, W. Differentiated rat glial
cell strain in tissue culture. Science (Washington DC), 161: 370—371,1968.
35. DeClerck,Y.A.,Shimada,H.,Gonzalez-Gomez,
L,andRaffel,C.Tumoralinvasion
18. San-Galli,F., Vrignaud,P., Robert,J. M.,Coindre,J. M., andCohadon,F. Assess
ment of the experimental model of transplanted C6 glioblastoma. 3. Neuro-oncol., 7:
36. Uhm, 1. H., Dooley, N. P., Villemure, i-G., and Yong, V. W. Mechanisms of glioma
invasion: role of matrix-metalloproteinases. Can. 1. Neurol. Sci.. 24: 3—15,1997.
299—304, 1989.
37. Rutka, I. 1., Apodaca, 0., Stem, R., and Rosenblum, M. The extracellular matrix of
in the central nervous system. 1. Neuro-oncol.,
thecentralandperipheral
nervoussystem:structure
andfunction.1. Neurosurg.,
69:
19. Chicoine, M. R., and Silbergeld, D. L. Assessment ofbrain tumorcell motility in vivo
and in vitro. 3. Neurosurg., 82: 615—622, 1995.
155—170,1988.
20. Paganeni, P. A., Carnal, P., and Schwab, M. E. Glioblastoma infiltration into central
nervous system tissue in vitro: involvement of a metalloprotease. I. Cell Biol., 107:
38. Ruoslahti, E. Brain extracellular matrix. Glycobiology, 6: 489—492,1996.
39. Bignami,A.,andAsher,R. Someobservationson the localizationof hyaluronicacid
in adult, newborn and embryonal rat brain. list. 1. Dev. Neurosci., 10: 45—Si,1992.
2281—2291,
1988.
21. Amberger,V. R., Avellana-Adalid,
V., Hensel,1., and Baron-VanEvercooren,A.
Oligodendrocyte-type
18: 111—121,1994.
2 astrocyte progenitors
use a metalloendoprotease
to spread
and migrate on CNS myelin. Eur. J. Neurosci., 9: 151—162,
1997.
40. Zisch, A. H., Alessandri, L, Rantsch, B., Falchetto, R., Winterhalter, K. H., and
Vaughan, L. Neuronal cell adhesion molecule contactin/Fll binds to tenascin via its
immunoglobulin-like domains. 1. Cell Biol., 119: 203—213.1992.
22. Small,R. K., Riddle,P., andNoble,M. Evidencefor migrationof oligodendrocyte 41. Spillmann, A. A., Amberger, V. R., and Schwab, M. E. High molecular weight
protein of human central nervous system myelin inhibits neurite outgrowth: an effect
which can be neutralized by the monoclonal antibody IN-l. Eur. 1. Neurosci., 9:
type-2 astrocyte progenitor cells into the developing rat optic nerve. Nature (Lend.),
328:
155—157, 1987.
23. Waif, B. C., Fok-Seang,J., and Miller,R. H. Evidencefor the ventraloriginof
oligodendrocyte
precursorsin therat spinalcord.J. Neurosci.,11:2477—2488,
1991.
549—555, 1997.
42.
Giese,
A., Kiuwe,
L,
Laube,
B., Meisner,
H., Berens,
M. E., and Westphal,
M.
Migrationof humangliomacellson myelin.Neurosurgery,38: 755-764,1996.
24. Noll, E., and Miller, R. H. Oligodendrocyte precursors originate at the ventral
ventricular zone dorsal to the ventral midline in the embryonic rat spinal cord.
Development (Camb.), 118: 563—573,1993.
43. Sivaparvathi, M., Sawaya, R., Wang, S. W., RayfOrd, A., Yamamoto, M., Lioda,
L A., Nicolson, G. L, and Ran, 1. S. Overexpression and localization ofcathepsin B
25. Fok-Seang, J., and Miller, R. H. Distribution and differentiation of A2B5+ glial
precursors in the developing rat spinal cord. i. Neurosci. Res., 37: 219—235, 1994.
44, Sivaparvathi, M., Sawaya, R., Chintala, S. K., Go, Y., Gokaslan, Z. L, and Rao, 1. S.
26. Knott, 3. C. A., Edwards, A. J., Gullan, R. W., Clarke, T. M., and Pilkington, G. J. A
human glioma cellline retaining expression ofGFAP and gangliosides, recognized by
A2B5 and LB1 antibodies, after prolonged passage. Neuropathol. AppI. Neurobiol.,
16: 489—500, 1990.
27. Merzak, A., Koochekpour,
S., McCrea, S., Roxanis, Y., and Pilkington, G. i. Gan
gliosides modulate proliferation, migration, and invasiveness of human brain tumor
cells in vitro. Mol. Chem. Neuropathol., 24: 121—135,
1995.
duringthe progressionof humangliomas.Clin.Exp.Metastasis,13: 49—56,
1995.
Expression ofcathepsin D during the progression ofhuman gliomas. Neurosci. Left.,
208: 171—174,1996.
45. Mohanam, S., Wang, S. W., Rayford, A.. Yamamoto, M., Sawaya, R., Nakajima, M.,
Liotta, L A.. Nicolson, G. L, Stetler-Stevenson, W. G., and Ran, J. S. Expression of
tissue inhibitors of metalloproteinases: negative regulators of human glioblastoma
invasion in vivo. Clin. Exp. Metastasis, 13: 57—62,1995.
46. Yakoclev, P. I., and Lecours, A. R. Regional development of the brain in early life.
in: A. Minkowski (ed.), The Myelogenic Cycles ofRegional Maturation ofthe Brain,
glioma cell growth, migration and invasion in vitro by transforming growth factor @l.
pp. 24-38. Oxford, United Kingdom: Blackwell Scientific Publications, Ltd., 1967.
28. Merzak, A., McCrea, S., Koochekpour,
S., and Pilkington, G. J. Control of human
Br. i. Cancer, 70: 199-203, 1994.
29. Merzak, A., Koochekpour, S., and Pilkington, G. J. CD44 mediates human glioma
cell adhesion and invasion in vitro. Cancer Res., 54: 3988—3992,1994.
47. Rossi, M.. Roopral, H. K., Maidment, S. L, Rucklidge, G. 1., and Pilkington, G. i.
The influence of sequential, in vitro passage on secretion of matrix metalloproteinases
by human brain tumour cells. Anticancer Res., 16: 121—128,
1996.
158
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.
Spreading and Migration of Human Glioma and Rat C6 Cells on
Central Nervous System Myelin in Vitro Is Correlated with
Tumor Malignancy and Involves a Metalloproteolytic Activity
Verena R. Amberger, Tamara Hensel, Nobuyoshi Ogata, et al.
Cancer Res 1998;58:149-158.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/58/1/149
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1998 American Association for Cancer Research.