(CANCER RESEARCH 50. 6039-6044. September 15. I990|
Effect of Epidermal Growth Factor on Glioma Cell Growth, Migration, and
Invasion in Vitro1
Morten Lund-Johansen,2 Rolf Bjcrkvig, Peter A. Humphrey, Sandra H. Bigner, Darell D. Bigner, and
Ole-Didrik Laerum
The (jade Institute, Department of Pathology, L niversity of Bergen, llaukelanil Hospital, \-5U2l Bergen, \orway f.\l. I.-J., K. B., O-I). I../, and Department of Pathology
¡P.A. //., .V.//. B., n. D. R.I, antl the Preuss Laboratory for Brain Tumor Rest-arch fD. I). fÃ-.¡.
Duke I'nirersily Medical (enter, Durham, \orth ('anilina 277IO
ABSTRACT
Effects of epidermal growth factor (KGK) and an antibody (Ab-528)
reactive against the binding site for I ( .1 on human I ( ,1 receptors were
studied on multicellular tumor spheroids obtained from three human
glioma cell lines with high (D-37 MG), medium (D-247 MG), and low
(D-263 MG) levels of EG F receptor expression. The D-247 MG and D263 MG spheroids grew slowly or not at all in the absence of I (.I , while
in the presence of EGF they were growth stimulated. Tumor cell migra
tion, as measured by the spread of cells from spheroids on a plastic
substratum, was increased by the addition of KGF for all three cell lines.
Stimulation of migration could be blocked by a subsequent addition of
Ab-528 to the medium at a concentration of 50 MS/ml. Invasiveness of
glioma cell spheroids into fetal rat brain aggregates was related to FGF
receptor expression; the two lines with medium to high receptor expres
sion (D-247 MG and D-37 MG) were invasive, while the line with low
EGF receptor expression (D-263 MG) was noninvasive, as assessed by
an ¡nvitro coculture assay. In the D-247 MG cell line, morphometry
revealed EGF-enhanced invasiveness of the tumor cells. The addition of
the Ab-528 to EGF-treated cocultures reduced invasion in both Ü-247
MG and D-37 MG cell lines. Antibody Ab-528 alone did not affect
glioma cell growth or migration but did inhibit invasiveness. The present
study suggests that, in brain tumors with an increased number of normalsized M, 170,000 EGF receptors, EGF or an EGF-like ligand such as
transforming growth factor-a may selectively facilitate expansive tumor
growth and tumor cell invasion. This effect may in part be blocked or
retarded by specific antibodies to the EGF receptor.
INTRODUCTION
cells in vitro (18), but most EGF-rs in cultured glioma cells are
of apparently normal molecular weight (17).
EGF and its receptor have been implicated in bladder carci
noma invasiveness (19, 20). but to date the role of EGF and
EGF-r in glioma cell invasiveness has not been studied. Here
we report the effect of EGF on glioma growth, glioma cell
migration, and glioma invasiveness by using an in vitro spheroid
model system. Three different glioma cell lines expressing low,
medium, and high levels of normal-sized EGF-r were studied.
Medium or high levels of EGF-r were required for invasiveness
to occur. EGF stimulated the growth of the glioma spheroids
with the greatest stimulation observed at the lowest EGF-r
concentration and the least stimulation in the cell line with the
highest EGF-r levels. EGF stimulated cell migration in all 3
lines and enhanced invasiveness of 1 cell line.
An antibody reactive against the EGF-binding domain on the
EGF-r blocked these EGF-mediated effects; alone the antibody
did not affect growth or migration but did slow invasiveness.
These studies suggest that EGF, or a similar ligand such as
transforming growth factor-it, via binding to the normal-sized
A/r 170,000 EGF-r. plays a significant role in glioma cell
growth, migration, and invasiveness. The level of EGF-r expres
sion in the cells influenced these biological behaviors, but a
general effect was of EGF stimulation of glioma cell growth
and migration at all levels of EGF-r expression.
MATERIALS AND METHODS
EGF,' a 6045-Da polypeptide, has been demonstrated to
stimulate the in vitro growth of both nonneoplastic and neoplastic glial cells (1-5). Neoplastic glial cells also exhibited
migration in response to EGF (6). EGF transduces its proliferative signal by specific binding to the EGF-r, a M, 170,000
transmembrane glycoprotein with intrinsic tyrosine kinase ac
tivity. Glioma cells in vivo often overexpress the EGF-r due to
EGF-r gene amplification (7-9). Often associated with this
amplification is structural rearrangement of the gene (8-10).
which results in the expression of truncated EGF-rs (10-12).
In vitro, gliomas only rarely amplify the EGF-r gene (13), and
a wide range of EGF-r expression from IO4 to IO6 EGF-r
molecules per cell has been observed (14-17). High expression
of the EGF-r in culture in the absence of an amplified gene may
be related to increased gene dosage due to extra copies of
chromosome 7, to changes in transcriptional or translational
control, or to changes in EGF-r mRNA and/or protein turn
over. A few variant EGF-r proteins have been detected in glioma
Received 3/9/90; accepted 6/19/90.
The costs of publication of this article were defrayed ¡npart b\ the payment
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 I'.S.C. Section 1734 solely to indicate this fact.
1Supported by the Norwegian Cancer Society. Grant 88-47; the NIH. Cirants
CA 11898. NS 20023. and C A 32672: the American Cancer Society. Grant 89171; and by a grant from the McDonnell Foundation.
2 To whom requests for reprints should be addressed, at the Gade Institute of
Pathology. Haukeland Hospital. N-5021 Bergen. Norway.
•'
The abbreviations used arc: EGF. epidermal growth factor; EGF-r. epidermal
growth factor receptor; DMEM. Dulbccco's modified Eagle's medium.
Tissue Culture Conditions and Chemicals. The cells were cultured in
a standard tissue culture incubator (37°C,5% CO2, 95% air, 100%
humidity). All tissue culture plastic articles were from Nunc (Roskilde,
Denmark). The cells were cultured in DMEM supplemented with 10%
heat-inactivated newborn calf serum, 4 times the prescribed concentra
tion of nonessential amino acids, i.-glutamine. penicillin (100 lU/ml),
and streptomycin (100 ug/m\) (all from Flow Laboratories, Glasgow,
Scotland). The nonadhesive base coating of culture plastic for 3-dimensional cultures consisted of 0.75% Noble agar (Difco Laboratories,
Detroit. MI) dissolved in medium. Tissue culture grade EGF (Grand
Island Biological Co., Grand Island. NY) was reconstituted with Hanks'
balanced salt solution (Gibco, Paisley. Scotland) to 100 ng/ml. The
EGF-r antibody Ab-528 (Oncogene Science. Mineóla. NY) was dis
solved in 0.9% NaCl solution to 500 ug/m\ before use. All experiments
were done in triplicate.
Tumor Cells. Three different human glioblastoma multiforme cell
lines with a different density of EGF-rs were chosen for the experiments:
D-263 MG (2.9 x 10* EGF receptors/cell), D-247 MG (1.5 x IO5
receptors/cell), and D-37 MG (1.59 x IO6receptors/cell) were studied
(5. 21-23). The EGF-r in these 3 lines is normal sized at M, 170,000
and is fully functional in the high affinity binding of EGF and autophosphorylation (17). The quantitative expression of EGF-r in these
cell lines has been shown to be stable during several passages (5). The
cells were maintained as monolayer cultures in 80-cnr tissue culture
flasks in 20 ml DMEM and passed 1:5 (v/v) with 0.05% trypsin in
0.02% EDTA (Flow Laboratories) by confluence. The formation of
multicellular spheroids was obtained as described (24) by seeding 3 x
IO6single cells in 20 ml DMEM into 80-cm2 culture flasks base coated
with 10 ml medium agar. Within 2 to 3 days, spheroids were formed.
6039
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
GLIOMA CELLS AND EGF DURING INVASION
For the experiments, 6-day-old spheroids with a diameter of 300 ±50
(SD) fitn were chosen.
Invasion Assay. Normal brain aggregates were prepared according to
standard techniques described previously (25). Briefly, single cells were
prepared from fetal rat brain at the I8th day of gestation. The cells
were seeded into DM EM agar base-coated plastic dishes. During 20
days of culture, cellular aggregates were formed consisting of mature
neurons and glia cells with a prominent neuropil. These aggregates
(diameter, 300 ±50 ¿im)were placed in contact with tumor spheroids
in 96-well multiwell dishes base-coated with 100 ^1 agar. The culture
medium volume was 400 ¿il.The coculture time was 4 days; the
specimens were then fixed for light microscopic examination. There
were 4 experimental groups. The first group received 10 ng/ml EGF;
the second received 50 ^g/ml Ab-528; the third received 10 ng/ml EGF
and 50 Mg/ml Ab-528; and the fourth was the control group cultured
under standard conditions. The cocultures (n = 6 in each group) were
sized every day with an inverted phase contrast microscope with a
calibrated reticle in the eyepiece. Coculture volume was calculated by
the formula
r = 0.4 x a x b2
described by Attia and Weiss (26), where a is the largest and b is the
smallest diameter of the cocultures. Single spheroids and brain aggre
gates served as controls (n = 4 in each group).
Light Microscopy and Invasion Criteria. The specimens from the
invasion experiments were fixed in 2% glutaraldehyde in 0.1 M sucroseadjusted cacodylate buffer (pH 7.4, 300 ±50 mOsmol). Dehydration,
embedding in Epon 812. and polymerization were performed as de
scribed earlier (27). The specimens were positioned under a stereo
microscope and sectioned to obtain a plane parallel to the bispherical
shape of the cocultures. Semithin 1.S-^m sections were cut on a Reichert
2040 microtome (Reichert, Vienna, Austria), stained with toluidine
blue, and examined by light microscopy. Sections were taken from at
least 3 to 4 different areas of each coculture. representing serial crosssections 100 to 200 /urn apart from each other. Invasion was defined
according to criteria previously described (25, 27) as tumor cells pene
trating into the outer fibrous layer of the brain aggregate, thus occu
pying positions previously held by the brain cells. Furthermore, an illdefined border between the 2 tissues and a substantial loss of brain
tissue were required to fulfill the criteria.
Morphometry of Cocultures. Sections from the middle part of each
coculture representing the area of maximum invasion, were carefully
selected and examined at xlOO, using a light microscope connected via
a video camera to a Mitsubishi AT personal computer equipped with
the Videoplan morphometry software (Kontron. Eching. West Ger
many). Tumor cell invasion was quantified in the following way (Fig.
1). First, the tissue areas represented by the remaining brain (B) and
tumor cells (7") were measured. Thereafter, the percentage of brain
tissue which had been destroyed by the invading tumor was determined
by 2 separate methods, both taking the spherical shape of the brain
aggregate into consideration, (a) The invaded brain area (/) was encir
cled by drawing the cursor along the brain/tumor interface and then
through the tumor spheroid, the cursor line extrapolating the missing
part of the circular perimeter of the brain aggregate. The percentage of
invasion (% of/,) was thereby represented by the formula
%
Of
/,
=
/ + B
x 100.
(b) The maximum diameter of the brain aggregate (/>) was measured,
and since the brain aggregate was originally spherical, the area of the
uninvaded brain aggregate section (A) is:
A = 3.14 x
and the percentage of invasion (% of 72)was thereby calculated by the
formula
of h =
Percentage of invaded brain tissue was calculated from each value, and
the 2 different values were compared by Spearman correlation.
Immunohistochemistry of Cocultures. Specific immunohistochemical
staining of the tumor cell population in the confrontation cultures was
carried out to verify the morphometry data, which relied upon the
morphological differences between brain and tumor cells. Cocultures
between D-247 MG spheroids and brain aggregates (n = 10) were fixed
for 24 h in 4% buffered formaldehyde and then dehydrated and em
bedded in paraffin according to standard histological techniques. Sec
tions with a thickness of 5 ^m were cut on a Shandon Hypercut
microtome (Shandon, Cheshire, England) and incubated overnight at
50°Cto adhere the tissue to the object slides. Sections were deparaffinized in toluene twice for 5 min, and endogenous peroxidase activity
was blocked by incubation for 10 min in 3 ml 30% H2O2 in 200 ml
methanol. Rehydration was obtained by passing the slides through
increasing concentrations of water ¡nethanol. The sections were there
after rinsed 3 times for 10 min in 50 imi Tris-HCl buffer (pH 7.4, 0.5
M NaCl). The specimens were then incubated at 4°Cfor 24 h with 10
Mg/ml of the 3B4 pan-human species-reactive antibody from mouse
ascites fluid dissolved in the same buffer containing 10% rabbit serum.
3B4 is a pan-human glioma-reactive antibody that is nonreactive with
normal rat brain.J Thereafter, the slides were rinsed 3 times for 10 min
in Tris buffer, incubated for l h with peroxidase-conjugated rabbit antimouse IgG (Dako, Glostrup. Denmark), diluted 1:20 (v/v) in Tris
buffer, and then rinsed as above. The final staining was performed by
incubating the tissue sections for 9.5 min in a solution consisting of 75
mg 3,3'-diaminobenzidine tetrahydrochloride (Sigma) in 150 ml of 250
niM Tris-HCl buffer, and for 10 min in a similar mixture with 75 ß\
35% H2O2. The sections were then mounted and examined by light
microscopy.
Growth of Multicellular Spheroids. In each experiment, 24 spheroids
were transferred to 16-mm 24-well multiwell dishes base coated with
0.5 ml agar medium and filled with 1 ml of DMEM. EGF, at a final
concentration of 10 ng/ml, was added to 12 wells. The diameters of
each spheroid were measured regularly in the phase contrast microscope
over a 20-day period and the spheroid volume was thereafter deter
mined.
Tumor Cell Migration. Spheroids, placed in uncoated 24-well dishes
filled with 500 /il DMEM were divided into 4 groups with 3 spheroids
in each group. The 4 groups were treated with EGF and/or Ab-528, or
untreated as above. A colony was defined as the cells emanating from
one spheroid. Two orthogonal colony diameters were measured after 4
days of culture, and the mean area of the colony was calculated from
the average diameter by the formula
At Day 8 the cultures were fixed and stained with crystal violet, and
the final colony area was measured with the morphometry equipment.
Statistics. The significance of numerical value differences between
the different measurements was analyzed by using the Student's t test
on paired observations.
RESULTS
Invasion and Immunohistochemistry. As observed by light
microscopy, there was a merging of the 2 cell populations in
the D-263 MG cocultures (n = 60), but there was a sharp,
demarcated border between the 2 tissues, and there was no
substantial destruction of brain tissue (data not shown). There
fore, according to the definition of in vitro invasion, this cell
line was noninvasive.
The D-247 MG and D-37 MG lines were invasive in the
assay according to the criteria. The tumor cells were morpho
logically different from the brain cells, and they invaded in a
broad, solid tissue manner without single cell invasion. This
x 100.
C. Wikstrand. unpublished results.
6040
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
GLIOMA CELLS AND EOF DURING INVASION
Fig. 1. D-247 MG cocultures showing, A.
strong invasion in EGF-treated coculture, and
C, inhibition of invasion in Ab-528 coculture
(Epon section, toluidine blue stain; A, x 160, C.
x 200). B and D, the corresponding morphometric images of the cocultures display the tu
mor population (7"), brain population (/?), and
the invading part of tumor (/). The diameter of
the brain aggregate is indicated (
). £,D247 MG coculture showing selective staining of
the tumor cell population obtained by the 3B4
pan-human species-reactive monoclonal anti
body (paraffin section, peroxidase stain, x 450).
D
*
was confirmed by examining the paraffin sections stained with
the 3B4 pan-human species-reactive antibody, which displayed
a specific staining of the human tumor cell population (Fig.
IE).
D-263 MG. In the EGF group, there was a volume increase
of the cocultures in the 4-day observation period; whereas in
the other groups, volume was maintained (Fig. 2). Since this
cell line was noninvasive, morphometry analysis was not per
formed.
D-247 MG. EGF caused a maintenance of the coculture
volume throughout the observation period; whereas in the other
experimental groups, the volume was decreased to approxi
mately 50% of the initial volume (Fig. 2). In the EGF-treated
group, invasion was increased, as shown by a greater loss of
brain tissue and a higher percentage of invasion than control
(Figs. 1 and 3Ä). Furthermore, there was an increase of the
area representing tumor tissue, suggesting a selective stimula
tion of tumor growth caused by EGF (Table 1). Cocultures
treated with Ab-528 in addition to EGF displayed tumor versus
brain areas and invasiveness smaller than that of control. The
addition of Ab-528 alone to the cocultures also reduced inva
sion, and there was more brain tissue preserved (Figs. 1C and
3B). However, the tumor cells were still capable of invasion
after the addition of Ab-528.
•—
0.031EE¿
4.
«*ti
'
0.03-D-247Mg
Coculture¿I^/A
0.02-•Eroa
Coculture1
°-°2-JT
^'^'~^è
/''\1AÕÕ-'"
I
^^Wk•
A;^.I
^^Â
N.TT¿
xV
—^A'
\?^"T*Control1
•i
0.01
-D~5oÜ
0.01
O
O:
""^^"0A
Ab-528A
— A:
Ab-528+EGF01
— A:
0.r
n.D-37Mg
23456Time
2345601
(in days)
Time (in days)
D-263Mg
Coculture
0123456
Time (in days)
Fig. 2. Volume of cocultures of glioma spheroids from 3 cell lines and fetal
rat brain aggregates. Bars. SEM.
D-37 MG. In the D-37 MG cell line, the coculture volume in
6041
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
GLIOMA CELLS AND EGF DURING INVASION
Table 1 Data obtained
hy morphometry
area
(mm2)0.014
groupD-247
Cell line and
MGControlEGFAb-528Ab-528
0.004"0.010
±
0.0030.01
±
±0.0030.01
5
±0.0020.015
6
EGFD-37 +
MGControlEGFAb-528Ab-528
±0.0100.016
0.0070.021
+
±0.0140.025
+ EGFBrain
±0.015Tumor
' Significant difference from control.
' Mean ±SD.
60-,
100
on inrasireness
of two human glioma cell lines
area
(mm2)0.007
I,19.5of
of/219.8
0.0040.020
±
0.0110.005
±
0.0040.004
±
0.0020.01
±
10.736.1
±
±11.715.2
±9.014.2
7.844.8
±
6.634.5
+
±11.812.9
±9.57.5
4.847.3
±
28.141.6+
+
±0.0060.016
2
12.423.8
0.0070.006
±
0.0030.008
±
19.625.2+
±
13.5%
±0.004Invasion%
29.840.0
±
14.023.3
±
1.023.0+
±2
15.5n201581710111711P"•CO.OOOI<0.0009<0.009None<O.O
Invasion
D-247VQ
•••
Ab-528
+EGF
Ab-528
D-37Mg
CD:
TO:
MU:
OS:
o -0-
=~
Control
EGF
Ab-528
Ab-528
+ EGF
Doy 1 Doy 4
Doy 8
20-
B
D-263Mg
Fig. 3. A, values of invasion. ci of /, and 'c of 12 plotted against each other.
Data from 2 cell lines are shown («= 98). The linear regression line is drawn. B.
graphic presentation of invasion ('"<of/,) in 2 glioma cell lines confronted with
15-
rat brain aggregates for 96 h. Bars, SEM.
10
0.08-*|0.06-I.20.04-•o's»0.02-Q.Ul0
Doy I
Control•
O:
EGF^8:^-0
0.04-,•s.,
•B~Ç
=
•
—•
0 02
o-o^o •
— o-o-o--cD-37MgO
•O•
—
0
20
25
0
5
10
15
20
(in
days)0.080.06-D-263MgHT/iAy1
days)
Time (in
0
5
10
15
20
Doy 8
•
00
•0.08-D-247Mg0.06-x«
5 25Time
10
15
Doy 4
Fig. 5. Migration of cells from glioma spheroids derived from 3 different cell
lines, untreated, or treated with EGF. Ab-528. or both. Bars. SEM.
25
Time (in days)
Fig. 4. Volume growth curves of glioma spheroids from 3 cell lines cultured
in the presence and absence of EGF. Bars. SEM.
all groups dropped to approximately 50% of the initial volume
(Fig. 2). In the control group, this cell line was highly invasive
and destructive, in average causing deletion of nearly one-half
of the brain aggregate and occasionally causing extensive lysis
of the remaining brain tissue. Thus, it was significantly more
invasive than D-247 MG. EGF did not alter invasiveness of
this cell line, but treatment with Ab-528 reduced invasion to
50% of control, independent of whether EGF was added or not.
Furthermore, the addition of Ab-528 caused a reduction of
tumor size (Table 1).
The comparison of the 2 different methods (% of I, and %
of/:) for analyzing invasion (Fig. 3/1) showed a strong corre
lation, thus verifying the usefulness of the morphometry.
Spheroid Growth. In the D-263 MG spheroids, EGF caused
a strong growth stimulation; in the control group, volume was
constant. D-247 MG spheroids were significantly growth stim
ulated by EGF (P < 0.001), showing a volume doubling time of
20 days in the EGF group and a maintenance of a constant
volume in the untreated group. D-37 MG control spheroids
lost nearly 50% of their volume; whereas in the EGF group,
this volume reduction was on average 15% (Fig. 4).
Migration. After 24-h culturing of spheroids on plastic sub
stratum, cells were spreading from the colonies in all groups,
as shown by single cells surrounding the explanted spheroid. In
all cell lines, the EGF treatment significantly stimulated out
ward migration of the cells on the plastic surface, and Ab-528
significantly neutralized this stimulation, producing a colony
area similar to that of control (P < 0.05, all measurements)
(Fig. 5). The dose-response effect of EGF on migration of D37 MG cells was tested and revealed that maximal migration
was absorbed over the EGF concentration range of 5 to 50 ng/
ml (data not shown). Thus, the EGF concentration of 10 ng/
ml used in these migration experiments was capable of eliciting
a maximal response.
6042
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
GLIOMA CELLS AND ECF DURING INVASION
DISCUSSION
The data presented here demonstrate the significance of EGF
and the normal-sized M, 170,000 EGF-r in the in vitro growth,
migration, and invasiveness of glioma cells. The response of
the cell lines, which were selected due to low, medium, and
high levels of EGF-r expression, to EGF did vary according to
receptor level in growth stimulation and migration experiments,
but the overall effect was of stimulation at all levels of EGF-r
expression. Moreover, in this study on 3 glioma cell lines,
invasiveness appeared to be dependent on EGF-r level, as 1.5
x 10s EGF-rs/cell or greater were required for invasiveness to
occur. Study of larger numbers of glioma cell lines will be
required to establish more precisely the relationship between
the level of EGF-r expression and invasive capacity. In vivo, the
EGF-r is highly expressed due to gene amplification in about
40% of patients; these amplified genes are sometimes rear
ranged, leading to the expression of truncated EGF-rs. The in
vitro studies presented here suggest that normal-sized receptors
at all levels of expression are capable of transducing EGFmediated stimulatory events. Thus, EGF, transforming growth
factor-«,amphiregulin, or other ligands capable of binding and
activating the EGF-r may operate in vivo to stimulate glioma
cell growth, migration, and invasiveness, both in patients with
overexpressed normal-sized EGF-r products of amplified genes
and in patients with normal-sized EGF-r products of nonamplified genes. The effect of EGF and other ligands on glioma
cells expressing the truncated EGF receptors remains to be
studied.
The cell lines used in our study have been tested for respon
siveness to EGF in monolayers, showing increased monolayer
growth of D-37 and D-263 MG, but no stimulation of D-247
MG (5). The results obtained by Werner et al. (5) show some
inconsistencies with ours, in particular concerning the D-247
MG cell line, which was claimed to be nonresponsive to EGF.
However, the present study on 3-dimensional growth of the
cells is not necessarily directly comparable to their investiga
tions on monolayer culture.
When biological response modifiers like EGF are tested in
vitro to evaluate effects on tumor cells, care should be taken to
mimic the in vivo situation as much as possible. Multicellular
spheroids are regarded as more representative of an in vivo
tumor than monolayers because there are increased intercellular
surface contacts and generation of extracellular matrix. In
addition, there exists a hypoxic cell population in the center of
large spheroids (for a review see Ref. 28). The actual induction
of spheroid growth caused by EGF in 2 of the cell lines suggests
that EGF may be essential to the 3-dimensional growth of these
tumor cells. This is consistent with the results of Velu et al.
(29), showing EGF-dependent tumorigenicity of an EGF-rpositive cell line. Westermark et al. (6) have shown that EGF
increases migration of human glioma cells in tissue culture
plastic dishes. They, therefore, suggested that EGF might affect
invasion of human gliomas (6). Our results confirm their data
on migration and, furthermore, suggest a possible relationship
between EGF-r and glioma cell invasiveness.
Internal penetration and expansion of EGF-treated tumor
cells may in consequence widen the total diameter of the coculture with an outer layer of remaining brain tissue. This could
increase the calculated value of the degree of invasion, since the
formula used to calculate invasion relies on this diameter or the
perimeter, which would also be widened. However, the aggre
gate diameter was similar in all groups (data not shown).
Our study indicates that EGF may be an important factor
but is not the sole factor in glioma invasion since invasion took
place in the presence of EGF-r antibody. The propensity of
gliomas to invade brain tissue may suggest involvement of
membrane-bound cell recognition molecules specific for the
central nervous system. These molecules might include glycosphingolipids such as gangliosides (30) and cell adhesion mol
ecules (31, 32). The EGF-r is not a central nervous systemspecific cell recognition molecule, since it is expressed in a
variety of tumors and normal tissues (33). However, the
EGF-r complex stimulates several generally important func
tions in the tumor cell, leading to an increased proliferation (5,
33), activation of cell membranes, increased cell motility (6),
and in some cell lines increased extracellular proteolysis (34),
all of which theoretically favor tumor invasion.
Although EGF stimulated all 3 cell lines, there were quanti
tative differences, and responses varied in the applied test
systems. For example, D-37 MG, which showed little growth
stimulation with EGF, was strongly stimulated in the migration
assay. Several authors have shown that monolayers of tumor
cells with a very high number of EGF-rs show little growth
response to EGF, or the growth may even be inhibited (see for
example Ref. 35). Our study suggests that this may also be the
case in 3-dimensional growth (e.g., D-37 MG, high receptor
number, no growth stimulation; D-263 MG, low number of
receptors, growth stimulation). This relationship may, however,
be an artifact of the in vitro growth environment, as in vivo
EGF stimulates the growth of tumors expressing high levels of
the EGF-r (36, 37).
The variations in EGF versus antibody effects on the cell
lines may be due to autocrine stimulation of the receptor (38).
Specifically, glioma cells may synthesize transforming growth
factor-«,as several studies have demonstrated increased trans
forming growth factor-« ¡mmunoreactivity in both cultured
glioma cells and in glioma biopsies compared to nonneoplastic
glial cells (39, 40). A population of EGF-rs may then already
be in a complexed form with ligand; cells bearing such com
plexes may be less responsive to exogenous ligand. Further
more, coexpression of the receptor and the growth factor may
lead to cytoplasmic receptor activation without involvement of
the cell surface membrane. These processes, the extent of which
are unknown in the investigated cell lines, would not necessarily
be influenced by exogenous EGF or Ab-528.
Our results give further evidence to the significance of EGF
(or an EGF-like ligand) and the EGF receptor in glioma growth
control. EGF in general stimulated the malignant behavior of
the tumor cells, and these effects were neutralized by the anti
body to the EGF receptor. The cell lines showed variations in
their responsiveness, and this is consistent with the well-known
heterogeneity of malignant human gliomas.
REFERENCES
1. Brunk. II., Schellcns. J.. and Westermark. B. Influence of epidermal growth
factor (EGF) on ruffling activity, pinocytosis and proliferation of cultured
human glial cells. Exp. Cell Res.. 103: 295-302. 1976.
2. Simpson. D. L.. Morrison. R.. de Vellis. J.. and Herschman. H. R. Epidermal
growth factor binding and mitogenic activity on purified populations of cells
from the central nervous system. J. Ncurosci. Res.. 8: 453-462, 1982.
3. Wang, S. L., Shiverick. K. T.. Ogilvie. S.. Dunn. W. A., and Raizada. M. K.
Gharacteri/ation of epidermal growth factor receptors in astrocytic glial and
neuronal cells in primary culture. Endocrinology. 124: 240-247, 1989.
4. Westermark. B. Density dependent proliferation of human glia cells stimu
lated by epidermal growth factor. Biochem. Biophys. Res. Commun., 69:
304-310. 1976.
5. Werner. M. H.. Humphrey. P., Bigner. D. D., and Bigner. S. H. Growth
effects of EGF and a monoclonal antibody against the EGF receptor on four
glioma cell lines. Acta Neuropathol.. 77: 196-201. 1988.
6. Westermark. B.. Magnusson. A., and Heldin. C-H. Effect of epidermal
6043
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
GLIOMA CELLS AND EGF DURING INVASION
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
growth Tactor in membrane motility and cell locomotion in cultures of human
glioma cells. J. Neurosci. Res., 8: 491-507, 1982.
Libermann, T. A., Razón.N., Banal. A. D., Varden. Y., Schlessinger, J., and
Soreq, H. Expression of epidermal growth factor receptors in human brain
tumors. Cancer Res., 44: 753-760, 1984.
Libermann. T. A.. Nusbaum. H. R.. Razón,N.. Kris. R.. Lax, I., Soreq, M.,
Whittle. N.. Waterfield, M. D.. Ullrich, A., and Schlessinger. J. Amplifica
tion, enhanced expression and possible rearrangement of EGF receptor gene
in primary human brain tumors of glial origin. Nature (Lond.), 313: 144147, 1985.
Wong. A. J.. Bigner. S. H., Bigner, D. D.. Kinzler. K. W.. Hamilton, S. R.,
and Vogelstein, B. Increased expression of the EGF receptor gene in malig
nant gliomas is invariably associated with gene amplification. Proc. Nati.
Acad. Sci. USA, 84: 6899-6903. 1987.
Humphrey, P. A., Wong, A. J., Friedman. H. S.. Werner. M. H.. Bigner, D.
D., and Bigner. S. H. Amplification and expression of the epidermal growth
factor receptor gene in human glioma xenografts. Cancer Res., 48: 22312238. 1988.
Maiden. L. T., Novak. U., Kaye, A. H.. and Burgess. A. W. Selective
amplification of the cytoplasmic domain of the epidermal growth factor
receptor gene in glioblastoma multiforme. Cancer Res., 48:2711 -2714,1988.
Yamazaki, H.. Fukui. Y., Ueyama, Y.. Tamaoki. N., Kawamoto. T., Taniguchi. S.. and Shibuya. M. Amplification of the structurally and functionally
altered epidermal growth factor receptor gene (c-erbB) in human brain
tumors. Mol. Cell. Biol., 8: 1816-1820. 1988.
Filmus, J., Pollak, M. N.. Cairncross. J. G., and Buick. R. N. Amplified,
overexpressed and rearranged epidermal growth factor receptor gene in a
human astrocytoma cell line. Biochem. Biophys. Res. Commun., I3I: 207215, 1985.
Bell, D.. Harsh. G., IV, Rosenblum, M., Meltzer. P., and Trent. J. Numeric
and structural alterations of chromosome 7 in human brain tumors: correla
tion with expression of epidermal growth factor receptors. Proc. Am. Assoc.
Cancer Res., 27: 37, 1986.
Steck. P. A.. Gallick. G. E.. Maxwell, S. A.. Kloetzer. W. S.. Arlinghaus, R.
B., Moser. R. P.. Gutlterman. J. U.. and Yung. W. K. A. Expression of the
epidermal growth factor receptor and associated glycoprotein on cultured
human brain tumor cells. J. Cell. Biochem.. 32: 1-10, 1986.
Yung. W. K. A.. Gallick, G. E., Waterfield, M. D., Moser. R. P., and Steck.
P. A. Expression of epidermal growth factor receptor in cultured human
brain tumor cells. Mol. Cell Biol.. 8: 1816-1820. 1988.
Humphrey, P. A., Wong, A. J., Vogelstein. B.. Friedman. H. S.. Bigner, D.
D.. and Bigner. S. H. Loss of epidermal growth factor receptor gene ampli
fication as malignant human gliomas establish in culture. Proc. Am. Assoc.
Cancer Res., 29: 78. 1988.
Steck. P. A.. Lee. P., Hung, M-C.. and Yung, W. K. A. Expression of an
altered epidermal growth factor receptor by human glioblastoma cells. Cancer
Res.. 48: 5433-5439, 1988.
Neal. D. E., Marsh. C.. Bennett. M. K.. Abel. P. D.. Hall. R. R.. Sainsbury.
J. R. C.. and Harris. A. L. Epidermal growth factor receptors in human
bladder cancer: comparison of invasive and superficial tumors. Lancet. /:
366-368. 1985.
Smith. K.. Fennelly, J. A., Neal. D. E., Hall, R. R.. and Harris. A. L.
Characterization and quantitation of epidermal growth factor receptor in
invasive and superficial bladder tumors. Cancer Res.. 49: 5810-5815. 1989.
Bigner. D. D.. Bigner, S. H.. Pontén,J., Westermark. B.. Mahaley, M. S..
Ruoslahti. E.. Herschman. H.. Eng. L. F., and Wikstrand. C. J. Heterogeneity
of genotypic and phenotypic characteristics of fifteen permanent cell lines
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
derived from human gliomas. J. Neuropathol. Exp. Neurol.. 15: 201-227.
1981.
Bigner, S. H., Bullard, D. E.. Pegram, C. N., Wikstrand. C. J.. and Bigner,
D. D. Relationship of in vitro morphologic and growth characteristics of
established human glioma-derived cell lines to their tumorigenicity in athymic
nude mice. J. Neuropathol. Exp. Neurol.. 40: 390-409, 1981.
Wikstrand, C. J.. Grahmann. F. C., McComb. R. D., and Bigner, D. D.
Antigenic heterogeneity of human anaplastic gliomas and glioma-derived cell
lines defined by monoclonal antibodies. J. Neuropathol. Exp. Neurol.. 44:
229-241, 1985!
Lund-Johansen, M., Bjerkvig, R., and Andersen, K-J. Multicellular tumor
spheroids in serum-free culture. Anticancer Res., 9: 413-420,1989.
Bjerkvig. R.. Laerum. O. D.. and Mella. O. Glioma cell interaction with fetal
rat brain aggregates in vitro and with brain tissue in vivo. Cancer Res., 46:
4071-4079, 1986.
Attia, A., and Weiss, D. W. Immunology of spontaneous mammary carci
nomas in mice. V. Acquired tumor resistance and enhancement in strain A
mice infected with mammary tumor virus. Cancer Res., 26: 1787-1800.
1966.
Lund-Johansen. M. Interactions between human glioma spheroids and fetal
rat brain aggregates studied in a chemically defined medium. Invasion Me
tastasis. 10: 113-128, 1990.
Sutherland. R. M. Cell and environment interactions in tumor microregions:
the multicellular spheroid model. Science (Washington DC), 240: 177-184,
1988.
Velu, T. J., Beguinot. L., Vass. W. C.. Willingham, M. C, Merlino, G. T.,
Pastan. !.. and Lowy, D. R. Epidermal growth factor-dependent transfor
mation by a human EGF receptor proto-oncogene. Science (Washington
DC), 238: 1408-1410. 1987.
Hakomori, S. Glycosphingolipids in cellular interaction, differentiation and
oncogenesis. Annu. Rev. Biochem.. 50: 733-764. 1981.
Kuppner. M. C.. Hamon, M. F., Sawamura. Y., and de Tribolet, N. Cytokine
regulation of intercellular adhesion molecule ICAM-I expression on human
glioma cells. J. Neuro-Oncol., 7(Suppl.): SI7, 1989.
Brackenbury. R. Molecular mechanisms of cell adhesion in normal and
transformed cells. Cancer Metastasis Rev., 4: 41-58, 1985.
Carpenter, G. Epidermal growth factor. Annu. Rev. Biochem., 48: 193-216,
1979.
Lee, L-S.. and Weinstein, I. B. Epidermal growth factor, like phorbol esters,
induces plasminogen activator in HeLa cells. Nature (Lond.). 274: 696-697,
1978.
Filmus, J.. Pollak, M. N.. Cailleau, R., and Buick. R. N. MDA-468, a human
breast cancer cell line with a high number of epidermal growth factor (EGF)
receptors, has an amplified EGF receptor gene and is growth inhibited by
EGF. Biochem. Biophys. Res. Commun., 128: 898-905, 1985.
Ginsburg, E., and Vonderhaar. B. K. Epidermal growth factor stimulates the
growth of A431 tumors in athymic mice. Cancer Lett., 28: 143-150. 1985.
Ozawa, S., Ueda, M., Ando, W., Abe, O., Hirai, M., and Shimizu. N.
Stimulation by EGF of the growth of EGF receptor-hyperproducing tumor
cells in athymic mice. Int. J. Cancer., 40: 706-710, 1987.
Westermark. B., Nister, M., and Heldin, C. H. Growth factors and oncogenes
in human malignant glioma. Neurol. Clin., 3: 785-799. 1985.
Cerosa, M. A.. Talarico. D., Fognani. C.. Raimondi, E.. Colombatti, M.,
Tridente, G.. DeCarli, L., and Della Valle. G. Overexpression of H-ras
oncogene and epidermal growth factor receptor gene in human glioblastomas.
J. Nati. Cancer Inst.. 81: 63-67. 1989.
Samuels. V., Barrett, J. M., Bockman. S., Pantazis. C. G., and Allen, M. B.,
Jr. Immunocytochemical study of transforming growth factor expression in
benign and malignant gliomas. Am. J. Pathol., 134: 895-902, 1989.
6044
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
Effect of Epidermal Growth Factor on Glioma Cell Growth,
Migration, and Invasion in Vitro
Morten Lund-Johansen, Rolf Bjerkvig, Peter A. Humphrey, et al.
Cancer Res 1990;50:6039-6044.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/50/18/6039
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 17, 2017. © 1990 American Association for Cancer Research.