Insulin-like growth factor-1 content and pattern of expression correlates with histopathologic grade in diffusely infiltrating astrocytomas

Abstract

Studies of experimental tumorigenesis have strongly implicated signaling of the insulin-like growth factor 1 (IGF-1) as a key component in astrocytic neoplasia; however, its role in the growth of low-grade and malignant human tumors is not well understood. Correlative analyses of IGF-1, p53, and Ki-67 (MIB-1) immunohistochemistry and IGF-1 receptor (IGF-1R) mRNA expression were performed to examine the cellular pattern of IGF-1 signaling in 39 cases of astrocytoma (World Health Organization grades II-IV). Tumor cells expressing IGF-1 and IGF-1R were present in all tumor grades. The proportion of tumor cells that expressed IGF-1 correlated with both histopathologic grade and Ki-67 labeling indices, while expression of IGF-1R mRNA correlated with Ki-67 indices. In cases where stereotactic tissue sampling could be identified with a specific tumor area by neuroimaging features, the numbers of IGF-1 immunoreactive cells correlated with the tumor zones of highest cellularity and Ki-67 labeling. In glioblastomas, the localization of IGF-1 immunoreactivity was notable for several features: frequent accentuation in the perivascular tumor cells surrounding microvascular hyperplasia; increased levels in reactive astrocytes at the margins of tumor infiltration; and selective expression in microvascular cells exhibiting endothelial/pericytic hyperplasia. IGF-1R expression was particularly prominent in tumor cells adjacent to both microvascular hyperplasia and palisading necrosis. These data suggest that IGF-1 signaling occurs early in astroglial tumorigenesis in the setting of cell proliferation. The distinctive correlative patterns of IGF-1 and IGF-1R expression in glioblastomas also suggest that IGF-1 signaling has an association with the development of malignant phenotypes related to aberrant angiogenesis and invasive tumor interactions with reactive brain.

Full text

Neuro-Oncology APRIL 1 999 109
Insulin-like growth factor-1 content and
pattern of expression correlates with
histopathologic grade in diffusely
inltrating astrocytomas
Hirofumi Hirano, M. Beatriz S. Lopes, Edward R. Laws, Jr., Tetsuhiko Asakura,
Masamichi Goto, Joan E. Carpenter, Larry R. Karns, and Scott R. VandenBerg
1
Departments of Neurosurgery [H.H., T.A.] and Pathology [M.G.], Kagoshima University, Kagoshima, Japan
890-8520; and Department of Pathology, Division of Neuropathology [M.B.S.L., J.E.C., L.R.K., S.R.V.], and
Department of Neurological Surgery [E.R.L., Jr.], University of Virginia Health Sciences Center,
Charlottesville, VA 22908
Studies of experimental tumorigenesis have strongly
implicated signaling of the insulin-like growth factor 1
(IGF-1) as a key component in astrocytic neoplasia; how-
ever, its role in the growth of low-grade and malignant
human tumors is not well understood. Correlative analy-
ses of IGF-1, p53, and Ki-67 (MIB-1) immunohistochem-
istry and IGF-1 receptor (IGF-1R) mRNA expression
were performed to examine the cellular pattern of IGF-1
signaling in 39 cases of astrocytoma (World Health
Organization grades II–IV). Tumor cells expressing IGF-1
and IGF-1R were present in all tumor grades. The pro-
portion of tumor cells that expressed IGF-1 correlated
with both histopathologic grade and Ki-67 labeling
indices, while expression of IGF-1R mRNA correlated
with Ki-67 indices. In cases where stereotactic tissue sam-
pling could be identied with a specic tumor area by
neuroimaging features, the numbers of IGF-1 immunore-
active cells correlated with the tumor zones of highest cel-
lularity and Ki-67 labeling. In glioblastomas, the localiza-
tion of IGF-1 immunoreactivity was notable for several
features: frequent accentuation in the perivascular tumor
cells surrounding microvascular hyperplasia; increased
levels in reactive astrocytes at the margins of tumor inl-
tration; and selective expression in microvascular cells
exhibiting endothelial/pericytic hyperplasia. IGF-1R
expression was particularly prominent in tumor cells
adjacent to both microvascular hyperplasia and palisad-
ing necrosis. These data suggest that IGF-1 signaling
occurs early in astroglial tumorigenesis in the setting of
cell proliferation. The distinctive correlative patterns of
IGF-1 and IGF-1R expression in glioblastomas also sug-
gest that IGF-1 signaling has an association with the
development of malignant phenotypes related to aberrant
angiogenesis and invasive tumor interactions with reac-
tive brain. Neuro-Oncology 1, 109–119, 1999 (Posted
to Neuro-Oncology [serial online], Doc. 98-16, April 30,
1999. URL <neuro-oncology.mc.duke.edu>)
G
rowth factors have been implicated at multiple
stages of astroglial tumorigenesis (Louis, 1997).
In the development of low-grade astrocytomas,
these factors may play a role in stimulating cellular pro-
liferation in the absence of aberrant cell cycle regulation
(Cavenee et al., 1997). The possibility of aberrant
autocrine loops has been proposed for a number of
growth factor/receptor systems, and a causative role for
platelet-derived growth factor or its receptor as stimulat-
ing proliferation has been the most strongly implicated
(Westermark et al., 1995). Previous studies of astrocytic
tumors have demonstrated altered growth factor activity-
Received 24 August 1998, accepted 6 January 1999.
1
Address correspondence and reprint requests to Scott R. VandenBerg,
M.D., Ph.D., Department of Pathology, Neuropathology, University of
Virginia Health Science Center, Box 214, Charlottesville, VA 22908.
2
Abbreviations used are as follows: EGFR, epidermal growth factor
receptor; FF, form factor; IGF-1, insulin-like growth factor-1; IGF-1R,
insulin-like growth factor-1 receptor; LI, labeling indices; SSC, standard
sodium citrate (buffer); WHO, World Health Organization.
Neuro-Oncology
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

Neuro-Oncology AP R I L 1 99 9110
in addition to platelet-derived growth factor-associated
with either changes in ligand or receptor expression in
basic broblast growth factor, transforming growth fac-
tor-alpha or -beta, and IGF-12 (Ekstrand et al., 1991;
Hussaini et al., 1996; Jennings et al., 1991, Takahashi et
al., 1990; Van Meir, 1995; Zumkeller et al., 1993).
During anaplastic progression of astrocytomas or the
formation of de novo glioblastomas, aberrant up regula-
tion of EGF signaling pathway(s) via EGFR activation
(
EGFR
gene amplication) or via converging down-
stream signals (via PLC
g
, mitogen-activated protein
kinase, and phosphatidylinositol-3 kinase pathways)
(Yamada et al., 1997) from other growth factor receptors
may govern a variety of malignant phenotypic properties,
including invasiveness and angiogenesis. IGF-1 is one
such growth factor with partial convergent downstream
signaling that may have biologic effects both during early
astrocytic tumorigenesis and in malignant progression.
Although IGF-1 ligands and IGF-1Rs have been previ-
ously detected in human astrocytomas, the present stud-
ies address the previously uncharacterized spatial expres-
sion of IGF-1 and IGF-1R in both low-grade and malig-
nant astrocytomas in relation to the proliferative indices,
WHO grade, and p53 immunoreactivity of the tumors.
Alterations in p53 immunoreactive expression were
examined, since functional p53 may regulate IGF-1 sig-
naling by affecting either receptor or binding protein
expression (Buckbinder et al., 1995; Werner et al., 1996).
IGF-1 and IGF-1R are expressed in all tumor grades with
an increased percentage of IGF-1 immunoreactive cells in
tumors with higher proliferative indices and higher
histopathologic grades. Within glioblastomas, there
appears to be a more complex spatial expression of IGF-
1 and IGF-1R within distinct tumor zones as well as
within specic cell cycle subpopulations. This study sug-
gests that astroglial tumorigenesis may involve IGF-1 sig-
naling associated with cell proliferation in both low- and
high-grade tumors and that the distinctive patterns of
IGF-1 and IGF-1R expression in glioblastomas suggest
that IGF-1 signaling may have an additional role(s) in
malignant tumor progression and synergistic interactions
with adjacent, reactive brain.
Materials and Methods
Tumors
A total of 39 cases of primary astrocytic tumors were
analyzed for these studies. Five cases were sampled sepa-
rately by both stereotactic biopsy and open resection (2
WHO grade II, 2 WHO grade III, and 1 WHO grade IV),
23 cases by open resection only (4 WHO grade II, 3
WHO grade III, and 16 WHO grade IV), 11 cases by
stereotactic biopsy alone (7 WHO grade II, 4 WHO III).
Tissue samples from all cases were analyzed by ABC
immunohistochemistry for Ki-67 (MIB-1 epitope), p53
(mutant and wild-type epitopes), and IGF-1 peptide.
Thirty-four of the 39 cases were also analyzed for the
IGF-1R mRNA by in situ hybridization. When multiple
sections were available for study, only those sections that
demonstrated the diagnostic histopathologic features for
a given WHO grade were analyzed.
Seventy-four tissue specimens from the 16 stereotacti-
cally operated cases were classied according to the neu-
roimaging location: Area 1, peripheral to the tumor mar-
gin; Area 2, low density by CT and high intensity by T2-
weighted MRI; Area 3, interface zone between the con-
trast-enhanced area and the nonenhanced area (zone of
high intensity on T2-weighted MRIs; zone of slightly
higher density by CT); Area 4, contrast-enhanced area;
Area 5, nonenhanced center of tumor.
Immunohistochemistry
Four-micron parafn sections of formalin-xed tissues
were mounted on poly-L-lysine-coated glass slides. Tis-
sues were deparafnized, fully hydrated through graded
ethanols, and incubated in methanol with 0.5% H
2
O
2
for 30 min at 22°C to block endogenous peroxidase
activity. Specimens for Ki-67 and p53 were additionally
processed by heating with a 750-watt microwave 2 times
at 5 min each in 10 mM citrate buffer (pH 6). Primary
mouse monoclonal antibodies (Ki-67, MIB-1,
Immunotech, Westbrook, ME; p53, DO-1, Immunotech;
IGF-1, antihuman IGF-1, Upstate Biochemical, Lake
Placid, NY) were applied and allowed to react for 2 h at
22°C (Ki-67 and p53) or 18 h at 4°C (IGF-1). Secondary
biotinylated antisera were followed by the avidin-biotin-
complex (Vectastain, Vector Laboratories, Burlingame,
CA); the reaction was developed with 3-39-diaminoben-
zine-tetrahydrochloride and counterstained with Mayer’s
hematoxylin.
To calculate the nuclear LI for Ki-67 and p53, 200
nuclei for each stereotactic tissue section and 400–2000
nuclei for each tissue section from the open resection
specimens were analyzed. IGF-1 immunoreactivity was
graded according to either the relative percentage of
immunoreactive cells (IGF-1 stain ratio) or the relative
immunostaining intensity (IGF-1 stain intensity). For
IGF-1sr, the following indices were used: 0, for the
absence of any IGF-1 immunoreactivity; 1+, for <10%
labeled cells; 2+, for 10–50% labeled cells; or 3+, for
>50% labeled cells. For grading IGF-1 immunoreactive
intensity, relative levels of IGF-1si were graded from 0,
1+, 2+, or 3+.
In Situ Hybridization
A cDNA containing 291 bp of the 59 untranslated
region and 283 bp of the amino terminal coding region
of the
a
-peptide of the IGF-1R was excised from 4.4 kb
cDNA of the human IGF-1R (Kaleko et al., 1990). The
antisense and sense probe (
35
S labeled) were prepared
by in vitro transcription with RNA polymerase T7 and
polymerase T3, respectively. An antisense probe for
a
-actin was prepared from clone EST00003 (Genbank
M61955, Bethesda, MD) identied by TIGR (The
Institute for Genomic Research, Gaithersburg, MD)
(Adams et al., 1991) and purchased from the American
Type Culture Collection (Manassas, VA). A fragment
of approximately 500 bp (bases 125-653) representing
the 59 end was subcloned into the
Eco
R1 and
Eag
l
sites of pBluescript-SK. For in situ hybridization, the
plasmid was linearized with
Eco
R1, and the
35
S-labeled
H. Hirano et al.: IGF-1 in astrocytomas
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology APRIL 1 999 111
probe was prepared by in vitro transcription with
RNA polymerase T3.
Deparafnized, formalin-xed tissue sections were
prepared as described for immunohistochemistry. Sec-
tions were then immersed in 4% paraformaldehyde for
10 min, washed in 0.5 SSC buffer for 5 min, treated with
3 µg/ml proteinase K (Boehringer Mannheim, Indianapo-
lis, IN) in buffer (500 mM NaCl, 10 M Tris-HCl, pH
8.0) for 30 min, and washed in 0.5 SSC for 10 min. The
slides were then incubated for 1 h with 150 µl of prehy-
bridization buffer (50% formamide, 0.3 M NaCl, 20
mM Tris-HCl, 5 mM EDTA, 1 [Denhardt’s solution,
10% dextran sulfate, 10 mM dithiothreitol]). The
hybridization mixture (50 µl per slide) consisted of the
prehybridization buffer with the addition of 1 (10
6
cpm
of
35
S-labeled probe and 20 µg of yeast tRNA per slide).
Each slide was hybridized with 50 µl of hybridization
mixture and incubated overnight at 55°C in a humidied
chamber. After twice washing (10 min each with 2 SSC
with 10 mM
b
-mercaptoethanol and 1 mM EDTA),
slides were washed with 20 µg/ml of RNase A in the
buffer (500 mM NaCl, 10 M Tris-HCl, pH 8.0) for 30
min at 22°C. Continuous washing was performed under
the following conditions: 2 changes of 2 SSC with 10
mM
b
-mercaptoethanol and 1 mM EDTA at 22°C for 10
min, 2 changes of 0.1 SSC with 10 mM
b
-mercap-
toethanol and 1 mM EDTA at 55°C for 2 h, 2 changes of
0.5 SSC at room temperature for 10 min.
Following dehydration through graded ethanols with
0.3 M ammonium acetate, the slides were dried and
dipped in Kodak NTB2 Nuclear Emulsion diluted 1:1
with 0.6 M ammonium acetate. The slides were allowed
to dry at room temperature and transferred to a light-
proof, desiccation box at 4°C. After 3 weeks of exposure,
the slides were developed for 2 min in Kodak D-19,
washed in water, and xed for 2 min in Kodak Rapid
Fixer. The slides were washed and stained with hema-
toxylin and eosin. The expression of IGF-1R mRNA was
evaluated by the intensity of grain signal: 0, 1+, 2+, or 3+.
Statistical Analysis
The Kruskal-Wallis analysis was used as a nonparamet-
ric method, and Dunn’s method (Hollander and Wolfe,
1973) was used for multiple comparisons. For popula-
tions whose normal distribution and equality of variance
were conrmed, one-way analysis of variance (ANOVA)
was used and Fisher’s least signicant difference proce-
dure was used for multiple comparisons. Additionally,
Spearman’s rank correlation coefcient was used to esti-
mate the strength of correlations.
To further address the inuence of IGF-1 immunore-
activity, IGF-1R mRNA expression, or p53 LI on Ki-67
LI, we performed a multiple regression analysis. The
analysis with a stepwise method was performed using Ki-
67 LI as a dependent variable and p53 LI, IGF-1
immunoreactivity, and IGF-1R mRNA expression as
independent variables. Before the analysis, qualitative
scoring data were converted to a quantitative scale. For
example, IGF-1R expression (0, 1+, 2+, 3+) was coded
into 3 variables: IGF-1R1+ (0 or 1); IGF-1R2+ (0 or 1);
IGF-1R3+ (0 or 1). Other independent variables we used
were IGF-1sr1+, IGF-1sr2+, IGF-1sr3+, IGF-1si1+, IGF-
1si2+, IGF-1si3+, dark p53 LI, and light p53 LI.
Cell Counts and Morphometry in Stereotactic
Specimens
Four different areas per hematoxylin and eosin-stained
section were digitized as 200 µm
3
200 µm elds (1000
3
1000 pixels). The images were analyzed on a Macin-
tosh computer using “Image” software (National Insti-
tutes of Health, http://rsb.info.nih.gov/nih-image/). An FF
(form factor), as an index of nuclear pleomorphism, was
calculated by the formula FF = 4
3 p 3
A/L
2
, where A is
the nuclear area and L is the nuclear circumference for
each nucleus. Mean values for each section were calcu-
lated, based on nuclear counts. Low FF values corre-
sponded to increased nuclear pleomorphism. Although
the ratio of the longest and shortest nuclear diameter of
each digitized nucleus also correlated with the degree of
nuclear pleomorphism, only the FF was used for statisti-
cal analysis.
Results
Expression of IGF-1
IGF-1 immunoreactivity was detected in the cytoplasm
and cellular processes of a fraction of tumor cells in all
specimens. Gemistocytic cells were well stained in the
cytosol, whereas the smallest, more anaplastic cells
showed the lowest levels of IGF-1. Immunoreactive cells
were relatively scattered in a diffuse pattern in the low-
grade tumors (Fig. 1A). In contrast, the immunoreactiv-
ity was quite heterogeneous within the high-grade astro-
cytomas (Fig. 1B). In glioblastomas, both the number
and staining intensity of tumor cell processes were typi-
cally highest in perivascular areas (Fig. 1E). In addition,
the tumor microvessels exhibiting endothelial/pericytic
proliferation were intensely immunoreactive for IGF-1
(Fig. 2A), compared with nonproliferative vessels in
glioblastomas and lower-grade astrocytomas. IGF-1 was
also present in reactive astrocytosis at the brain-tumor
interface. The intensity of IGF-1 reaction was particu-
larly strong at the edge of glioblastomas (Fig. 2C). By
contrast, astrocytes in inltrating zones of low-grade
astrocytomas were weakly positive.
The highest levels of IGF-1 immunoreactivity, as
measured by both the percentage of immunoreactive cells
and relative staining intensity, were found in glioblas-
tomas. The relative percentage of IGF-1 immunoreactive
cells strongly correlated with the WHO grade of the
tumor (II–IV;
r
s
=0.489,
P
<0.0001), but the trend for
increased staining intensity did not show a similar corre-
lation (Table 1).
IGF-1R mRNA Expression
IGF-1R mRNA was expressed in all grades of astrocy-
tomas as determined by in situ hybridization (Figs. 1C
and 1D). Although IGF-1R expression tended to be
greatest in the higher-grade astrocytomas (Fig. 1D),
expression was also present in low-grade tumors (Fig. 1C)
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology AP R I L 1 99 9112
such that there was sufcient signal to distinguish them
from adjacent brain. Within glioblastomas, similar to
IGF-1 expression, the highest level of IGF-1R mRNA
was present in perivascular tumor cells (Fig. 1F). Dis-
tinct from the pattern of IGF-1 expression, the tumor
cells surrounding areas of necrosis, especially palisading
necrosis (Fig. 1G), showed marked expression of IGF-
1R without a similar difference in IGF-1 expression. In
addition, microvessels exhibiting endothelial/pericytic
proliferation had elevated IGF-1R mRNA similar to
IGF-1 expression (Fig. 2B). Statistical analysis demon-
strated that IGF-1R mRNA expression in tumor cells
had a slight correlation with WHO tumor grade
(
r
s
=0.292,
P
=0.0216; Table 1).
Fig. 1. Insulin-like growth factor 1 (IGF-1) and IGF-1 receptor (IGF-1R) expression in astrocytic tumors. A. IGF-1 immunoreactivity in astrocy-
toma WHO grade II. B. IGF-1 immunoreactivity in a glioblastoma. C. IGF-1R mRNA expression in an astrocytoma WHO grade II. D. IGF-1R
mRNA expression in a glioblastoma. E. Strong IGF-1 immunoreactivity in perivascular tumor cells in a glioblastoma. F. IGF-1R mRNA expres-
sion, which is strongest in perivascular tumor cells in a glioblastoma. G. Overexpression of IGF-1R mRNA in tumor cells surrounding palisading
necrosis in a glioblastoma.
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology APRIL 1 999 113
Ki-67 and p53 Immunohistochemistry
Staining intensity of nuclear immunoreactivity for Ki-67
was relatively homogeneous among the positive cells. Gen-
erally, the higher the grade, the higher the LI. The Ki-67 LI
correlated well with the WHO tumor grade (Table 2).
Nuclear immunoreactivity for p53 had two relative
levels of intensity, representing conspicuously dark and
light staining. Distinctly dark and light immunoreac-
tive nuclei were usually randomly mixed in any given
area (Fig. 2E). The differentiation between dark- and
light-stained nuclei was comparatively analyzed using
a case of glioblastoma with documented p53 gene
mutation as the dark standard. The percentage of
darkly stained p53 nuclei increased slightly with higher
WHO tumor grades (
r
s
=0.268,
P
=0.0364; Fig. 2D).
The light-stained p53 LI was relatively unchanged in
all WHO grades.
Correlation of IGF-1, IGF-1R, Ki-67, and p53 Data
The IGF-1 staining ratio had a correlation (
r
s
=0.504,
P
<0.0001; Fig. 3A left) with Ki-67 LI in the same sam-
Fig. 2. Insulin-like growth factor 1 (IGF-1) and IGF-1 receptor (IGF-1R expression and p53 protein reactivity in astrocytic tumors. A. IGF-1
immunoreactivity in microvessels exhibiting endothelial/pericytic proliferation in a glioblastoma. B. Similarly, high IGF-1R mRNA expression in
microvessels exhibiting endothelial/pericytic proliferation. C. IGF-1 immunoreactivity in reactive astrocytes at the periphery of a glioblastoma. D.
Dark p53 LI, which has a slight correlation with WHO grade (r
s
=0.268, P=0.0364). The r
s
of total p53 LI and light p53 LI with WHO grade was
0.414 (P=0.2721) and 0.035 (P=0.7873), respectively. E. p53 immunohistochemistry in a glioblastoma exhibiting darkly and lightly stained nuclei.
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology AP R I L 1 99 9114
ples that were used for analysis of IGF-1R mRNA. This
correlation was also present in both the total specimen
group (153) from the 39 cases (
r
s
=0.592,
P
<0.0001) as
well the tissue limited to stereotactic biopsy (Table 4).
This result was independent of the selection of subgroup.
The Ki-67 LI and IGF-1R expression were correlated
(Fig. 3A right) and the Spearman’s rank correlation coef-
cient was 0.517 (
P
<0.0001). With respect to p53
immunohistochemistry, a slight correlation between dark
p53 LI and IGF-1R expression was present (
r
s
=0.299,
P
=0.0197; Fig. 3B right); however, there was no signi-
cant correlation between total p53 LI and IGF-1R
expression.
Stepwise multiple regression analysis identied three
effective independent variables: the highest percentage of
IGF-1 immunoreactive cells (IGF-1sr3+), the greatest lev-
els of IGF-1R expression (IGF-1R3+), and dark p53 LI.
Other variables were eliminated because their inuences
on the Ki-67 LI were not statistically signicant. The
equation nally obtained was as follows: tted Ki-67 LI
= 0.234(dark p53 LI) + 0.138(IGF-1R3+) + 0.103(IGF-
1sr3+), and the multiple correlation coefcient (
r
) was
0.828. Each
P
value for the regression coefcient was as
follows:
P
<0.0001 (dark p53 LI);
P
=0.0004 (IGF-1R3+);
P
=0.0002 (IGF-1sr3+). The distribution of real Ki-67 LI
and calculated values as tted Ki-67 LI obtained by the
multiple regression equation is shown in Fig. 4. This
analysis suggests that Ki-67 LI is strongly affected by
dark p53 LI (Fig. 3B left), a high expression of IGF-1R,
and high values for the percentage of IGF-1 immunore-
active cells. However, these data cannot exclude the con-
verse possibility that a highly proliferative state drives the
expression of IGF-1 and IGF-1R.
Stereotactic Biopsy Correlations
For the stereotactic cases, the mean cellularity, mor-
phometric nuclear FF, Ki-67 LI, and IGF-1 immuno-
histochemistry were analyzed according to the neu-
roimaging localization (areas 1–5) as previously
described. There was a signicant correlation between
Ki-67 LI and the centripetal location in tumors (Table 3)
with
P
=0.0004 and
r
s
=0.419. The mean cellularity also
signicantly correlated with the zonal location (Table 3).
The degree of nuclear pleomorphism, as quantied by
the mean FF, increased toward the centers of tumors
(Table 3). Both the percentage of IGF-1 immunoreac-
tive cells and the staining intensity correlated with the
zonal localization of the tumor cells (Table 3). The per-
centage of IGF-1 immunoreactive tumor cells also cor-
related with cellularity and the Ki-67 LI (Table 4), but
the IGF-1 staining intensity did not. Nuclear pleomor-
phism (FF) also correlated with levels of IGF-1
immunoreactivity (Table 4).
Discussion
Astrocytic tumorigenesis and anaplastic progression is a
complex multistep process that results in a sustained, het-
erogeneous cell proliferation; alterations in cell cycle reg-
ulation; increased brain invasiveness; and aberrant
angiogenesis. Although an increase in proliferative activ-
ity is currently postulated to underlie the initial develop-
ment of both low- and high-grade tumors, the roles of
specic growth factors for regulating this response in the
various tumor grades are not well understood. Func-
tional IGF-1R signaling appears to be essential for neo-
plastic transformation in multiple cell lineages, including
embryonal broblasts using SV40 large T-antigen trans-
fection (Sell et al., 1993). In the brain, IGF-1 functions as
a glial growth factor during development and in reactive
astrocytosis (Coppola et al., 1994; Gammeltoft et al.,
1988a; Sandberg et al., 1988). Experimental studies have
strongly suggested a key role for IGF-1 signaling in astro-
cytic tumorigenesis (Rubin and Baserga, 1995).
In cultured glioma cells, IGF-1 stimulates prolifera-
tion (Merrill and Edwards, 1990; Rubin and Baserga,
1995; Wang et al., 1997) and blocks apoptosis (Resnicoff
et al., 1995; Yang et al., 1996). Other glial mitogens (Van
Table 1. IGF-1 immunoreactivity and IGF-1R expression according to tumor grade by World Health Organization classication
Mean rank values
Grade Tissue section
a
IGF-1 staining ratio
b
IGF-1 staining intensity
c
IGF-1R expression
d
II 19 22.4
e
29.9 22.6
III 15 26.6 28.9 28.8
IV 29 41.1
f
35.1 37.2
IGF-1, insulin-like growth factor 1; IGF-1R, insulin-like growth factor-1 receptor.
a
The number of tissue sections examined.
b
Spearman’s rank correlation coefcient is 0.489 (P<0.0001).
c
Spearman’s rank correlation coefcient is 0.143 (P=0.2586).
d
Spearman’s rank correlation coefcient is 0.292 (P=0.0216).
e
versus
f
: P<0.005.
Table 2. Relationship between Ki-67 labeling indices (MIB-1) and
tumor grade by World Health Organization classication
Tumor grade Ki-67 LI% (mean) ±SEM
II 3.8 1.1
III 10.2 2.4
IV 19.0 3.0
LI, Labeling indices; SEM, standard error of the mean.
Ki-67 LI correlated with WHO grade (r
s
=0.517, P<0.0001). There was a difference among
three grades by Kruskal-Wallis (P=0.0002), and Dunn’s method was used for multiple
comparisons. Mean ranks of Ki-67 LI in each grade were II, 18.5; III, 32.0; IV, 40.8.
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology APRIL 1 999 115
der Ven et al., 1997), including EGF (Chernausek, 1993)
and raised levels of lactate, can increase IGF-1 secretion
by cultured glioma cells (Straus and Burke, 1995).
Reduced IGF-1 levels also accompany decreased cellular
proliferation after up regulation of gap junction forma-
tion (Bradshaw et al., 1993) or following treatment with
prostaglandin-A
2
(Bui et al., 1997) in cultured glioma
cells. Antisense knockout and mutant transfection stud-
ies with IGF-1R demonstrate that these tumorigenic
properties of IGF-1 are mediated via IGF-1R (Burgaud et
al., 1995; Resnicoff et al., 1995, 1996) and that consti-
tutive IGF-1R signaling is sufcient for autonomous
growth of glioma cells in serum-free cell culture
(Ambrose et al., 1994).
Previous studies have demonstrated that IGF-1 peptide,
IGF binding proteins, and IGF-1R are expressed in human
astrocytic tumors (Gammeltoft et al., 1988b; Glick et al.,
1989; McCusker et al., 1990; Merril and Edwards, 1990;
Sandberg et al., 1988; Zumkeller et al., 1993) at levels
higher than in fetal brain (Sandberg et al., 1988). The
present study is the rst to systematically examine the cel-
lular pattern of IGF-1 and IGF-1R expression in a large
series of diffuse-type astrocytomas (WHO grades II–IV)
with correlations to MIB-1 proliferative indices, p53
immunohistochemistry, and the previously well-character-
ized heterogeneous tumor zones dened by neuroimaging.
The localization of the peptide and receptor are consistent
with the putative neoplastic paracrine/autocrine nature of
IGF-1 signaling in diffuse-type astrocytomas. The
increased expression of IGF-1 in reactive astrocytes, espe-
cially at the margin of glioblastomas, also raises the possi-
bility that IGF-1 may be a paracrine growth factor at the
A
B
Fig. 3. Relationship between insulin-like growth factor 1 (IGF-1), IGF-1 receptor (IGF-1R), Ki-67, and p53 in astrocytic tumors. A. Relationship
between Ki-67 LI and IGF-1 staining ratio and IGF-1R expression. There were differences among the mean ranks of Ki-67 LI by IGF-1 ratio
(P=0.0013 by Kruskal-Wallis) and among those by IGF-1R expression (P=0.0002 by Kruskal-Wallis ). The differences between two groups are
based on Dunn’s method of multiple comparison (#, P<0.05; ##, P<0.005; *, P<0.01; **, P<0.001). Spearman’s rank correlation coefcient
(r
s
=0.504, P<0.0001) indicated a correlation between IGF-1 staining ratio and Ki-67 LI. IGF-1R expression and Ki-67 LI also correlated well and
the r
s
was 0.517 (P<0.0001). However, no differences were seen among the mean ranks of Ki-67 by IGF-1 intensity (P=0.5237 by Kruskal-
Wallis ) and the grade of IGF-1 staining intensity (r
s
=0.079, P<0.533) (not shown). B. Relationship between dark p53 LI and Ki-67 LI and IGF-
1R expression. Ki-67 LI was affected by dark p53 LI (r
s
=0.669, P<0.0001). Although the statistical differences were not detected by Kruskal-
Wallis, Spearman’s rank correlation indicated a correlation between dark p53 LI and IGF-1R expression (r
s
=0.299, P=0.0197).
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology AP R I L 1 99 9116
inltrating tumor edge. The direct relationships between
IGF-1 content to the MIB-1 proliferative indices and
WHO grade in all tumor samples (open resection and
stereotactic biopsies) implicate IGF-1 as a mitogenic factor
in human astrocytomas. Within specic neuroimaging
zones in the stereotactic biopsies (WHO grades II–IV), the
strong association of IGF-1 expression with MIB-1 prolif-
erative indices, cellularity, and nuclear pleomorphism also
suggests that IGF-1 has a role in cellular proliferation in
both low- and high-grade astrocytic tumors. The more
robust correlation between the percentage of IGF-1
immunoreactive cells and proliferative indices or WHO
grade, compared with that for IGF-1 staining intensity,
suggests that these biologic and histopathologic parame-
ters are sensitive to threshold levels without additional
detectable effects at higher focal concentrations. Alterna-
tively, estimation of immunohistochemical staining inten-
sity is likely to have more technical limitations than sim-
ple determination of immunoreactive cells.
The geographic patterns of IGF-1 expression in
glioblastomas, as distinct from the random expression
pattern in low-grade tumors, suggest that IGF-1 signaling
may also have biologic activities related to the attenua-
tion of apoptosis and to aberrant microvascular hyper-
plasia. The up regulation of IGF-1R expression in malig-
nant astrocytes surrounding palisading necrosis without
an increase in IGF-1 expression would suggest that this
up regulation may be related to anti-apoptotic signaling
mediated by IGF-1 (Kulik et al., 1997). Recently, it was
reported that cells expressing the mutant receptor (1280-
1283) are fully responsive to IGF-1-mediated mitogene-
sis, but are not transformed (Li et al., 1996); whereas the
transforming domain of the IGF-1R appears to be local-
ized between residues 1245 and 1310 (Hongo et al.,
1996). Accordingly, there is a possibility that the down-
stream signaling via IGF-1R for proliferation and trans-
formation may be distinct. The abundant IGF-1R mRNA
expression that did not directly coincide with prolifera-
tive activity may be related to the role of IGF-1 signaling
in blocking apoptosis. This activity may be especially rel-
evant to necrotizing micro-environments, such as in
those cells palisading around geographic necrosis.
The tumor cells in areas immediately adjacent to the
aberrant hyperplastic microvasculature demonstrated
high levels of both IGF-1 and IGF-1R expression. Like-
wise, the microvascular cells associated with endothe-
lial/pericytic hyperplasia had notable up regulation of
both IGF-1 and IGF-1R expression. Such an up regula-
tion of IGF-1R expression may be, in part, stimulated by
increased IGF-1 levels (Roberts, 1996). Such a conspicu-
ous pattern suggests the possibility of both paracrine and
autocrine IGF-1 signaling in association with this
microvascular response. This may be related to the direct
role of IGF-1 signaling in angiogenesis (Bar et al., 1988;
Kluge et al., 1997; Nakao-Hayashi et al., 1992) or the
increased expression of vascular endothelial growth fac-
tor in response to IGF-1 signaling (Goad et al., 1996;
Table 3. Stereotatic biopsy correlations
Mean rank values
Tissue Cell Form Ki-67 LI
d
IGF-1 IGF-1
Location section
a
number
b
factor
c
% staining ratio
e
staining intensity
f
1 9 39.6
g
0.722
g
0.8
g
15.0
g
14.5
g
2 34 62.9 0.706
g
2.2 36.5
h
37.7
h
3 13 131.4
i
0.693 5.6
h
38.7
h
34.4
h
4 16 129.0
j
0.629
h
4.6
h
47.8
k
50.3
k
5 2 111.0 0.537
h
6.5
h
66.5
i
55.3
h
IGF-1, insulin-like growth factor 1; IGF-1R, insulin-like growth factor 1 receptor; LI, labeling indices.
a
The number of tissue sections examined.
b
Spearman’s rank correlation coefcient is 0.433 (P<0.0001).
c
Spearman’s rank correlation coefcient is –0.304 (P=0.0094).
d
Spearman’s rank correlation coefcient is 0.419 (P=0.0004).
e
Spearman’s rank correlation coefcient is 0.457 (P=0.0001).
f
Spearman’s rank correlation coefcient is 0.429 (P=0.0002).
g
versus
h
, P<0.05; versus
i
, P<0.01; versus
j
, P<0.005; versus
k
, P<0.001.
Fig. 4. Distribution of real Ki-67 LI and calculated Ki-67 LI. The dis-
tribution of real Ki-67 LI is shown on the y-axis and calculated Ki-67
LI on the x-axis. The equation was as follows: tted Ki-67
LI=0.234(dark p53)+0.138(IGF-1R3+)+0.103(IGF-1sr3+). Multiple
correlation coefcient (r) of the regression model was 0.828.
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology APRIL 1 999 117
Kim et al., 1998; Punglia et al., 1997). The secretion of
plasminogen-activating proteinases by the malignant
perivascular astrocytes (Landau et al., 1994; Yamamoto
et al., 1994) may also promote this IGF-1-induced
microangiogenesis (Sato et al., 1993). Despite these sug-
gestive data, the biologic mechanisms for the endothe-
lial/pericytic proliferation in glioblastomas, as an exam-
ple of a complex aberrant microvascular hyperplasia,
await further elucidation.
The increased numbers of reactive astrocytes express-
ing increased IGF-1 at the inltrating margin of glioblas-
tomas in comparison to low-grade astrocytomas and
normal brain raises the possibility of IGF-1 acting as a
paracrine growth factor at the inltrating tumor edge.
Nonneoplastic astrocytes up regulate IGF-1 expression in
response to injury (Garcia-Estrada et al., 1992), espe-
cially in subacute gemistocytic astrocytosis (Hussaini et
al., 1996). Although the receptor signaling pathways
mediating this response are not dened, EGFR activation
would be a putative mechanism (Ekstrand et al., 1995),
since increased EGF expression is a heterogeneous fea-
ture within a number of malignant astrocytic tumors
(Bello, 1996; Sanlippo et al., 1993). Focally increased
levels of IGF-1 and EGF would also produce a synergis-
tic effect in the invasive phenotype by increasing cellular
motility (Faber-Elman et al., 1996) via common down-
stream signaling pathways (Sasaoka et al., 1994;
Tournier et al., 1994). IGF-1 signaling, with respect to
specic biologic effects, may be more heterogeneous and
complex than what IGF-1/IGF-1R expression would sug-
gest. Activation of synergistic downstream pathways
with other growth factors and the variable tumoral
expression of a number of IGF binding proteins that also
modulate IGF-1 activity at the cell surface (Blum et al.,
1989; Busby et al., 1988; De Mellow and Baxter, 1988;
Ekstrand et al., 1995; Elgin et al., 1987; Jones and Clem-
mons, 1995; Mohan et al., 1989; Rao et al., 1995; Ross
et al., 1989; Shimasaki et al., 1990; Zheng et al., 1998)
would most likely confer additional levels of complexity.
Characterization of the specic mechanisms underlying
this putative complexity in situ are beyond the scope of
the present study, but clearly merit future work.
Wild p53 may down regulate IGF-1 signaling via sup-
pressing the IGF-1R promoter (Werner et al., 1996) or
decreasing the bioavailability of IGF-1 by inducing IGF
binding protein-3 expression (Buckbinder et al., 1995).
In this context, the distinction between dark and light
nuclear p53 immunoreactivity, as dened by a case of
glioblastoma with documented p53 gene mutation as
the dark standard, suggests an interesting possibility
with respect to IGF-1R expression. Variation in p53
immunohistochemical staining intensity has been previ-
ously described (Jaros et al., 1992; Rubio et al., 1993),
and it may be partly due to wild-type p53 protein accu-
mulation without mouse double-minute-2 (
MDM2
)
gene amplication (Fritsche et al., 1993; Rubio et al.,
1993). While not denitive, the population of light-
stained p53 cells may thus include those cells with high
levels of wild-type p53, but with no gene mutation,
while the dark p53 nuclear immunoreactivity may indi-
cate cells with aberrant p53 function. Dark p53
immunoreactivity correlated with increased IGF-1R
expression in this study, suggesting another putative
pathway affecting IGF-1 signaling in astrocytomas. The
multiple regression model in our study in which Ki-67 LI
has strong correlation with IGF-1, IGF-1R mRNA
expression and dark-stained p53 LI suggests the possi-
bility that IGF-1 signaling is up regulated in a complex
constellation of growth factors affecting astrocytic pro-
liferation in tumor development and growth. The char-
acteristic patterns of IGF-1 and IGF-1R expression in
glioblastomas also emphasize that multiple factors
affecting IGF-1 signaling may be a determinant of the
biological behavior in malignant astrocytomas.
Acknowledgments
We are grateful to Dr. Gerald R. Hankins for advice
regarding statistical analyses.
Adams, M.D., Kelley, J.M., Gocayne, J.D., Dubnick, M., Polymeropoulos,
M.H., Xiao, H., Merril, C.R., Wu, A., Olde, B., Moreno, R.F., Kerlavage,
A.R., McCombie, W.R., and Venter, J.C. (1991) Complementary DNA
sequencing: Expressed sequence tags and human genome project. Sci-
ence
252
, 1651–1666.
Ambrose, D., Resnicoff, M., Coppola, D., Sell, C., Miura, M., Jameson, S.,
Baserga, R., and Rubin, R. (1994) Growth regulation of human glioblas-
toma T98G cells by insulin-like growth factor-1 and its receptor. J. Cell.
Physiol.
159
, 92–100.
Bar, R.S., Boes, M., Dake, B.L., Booth. B.A., Henley, S.A., and Sandra, A.
Table 4. Relationship between IGF-1, Ki-67 labeling indices, form factor, and number of cells in stereotatic biopsy
Classication Number of cells Form factor Ki-67 LI (%)
IGF-1 ratio P=0.0314 (K-W) P=0.0361 (ANOVA) P=0.0008 (K-W)
(0, 1+, 2+, 3+) r
s
=0.296, P=0.0114 (S) r
s
=-0.313, P=0.0075 (S) r
s
=0.413, P=0.0004 (S)
IGF-1 intensity P=0.4621 (K-W) P=0.0071 (ANOVA) P=0.1575 (K-W)
(0, 1+, 2+, 3+) r
s
=0.215, P=0.1304 (S) r
s
=-0.366, P=0.0018 (S) r
s
=0.208, P=0.075 (S)
ANOVA, one-way ANOVA; IGF-1, insulin-like growth factor 1; IGF-1R, insulin-like growth factor 1 receptor; K-W, Kruskal-Wallis analysis; S, Spearman’s rank correlation (r
s
: Spearman’s
rank correlation coefcient).
References
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology AP R I L 1 99 9118
(1988) Insulin, insulin-like growth factors, and vascular endothelium. Am.
J. Med.
85
, 59–70.
Bello, L. (1996) Urinary epidermal growth factor (EGF) levels and epidermal
growth factor receptor (EGF-R) expression in human cerebral tumors.
Clinical considerations. J. Neurosurg. Sci.
40
, 167–182.
Blum, W.F., Jenne, E.W., Reppin, F., Kietzmann, K., Ranke, M.B., and Bierich,
J.R. (1989) Insulin-like growth factor I (IGF-I)-binding protein complex is
a better mitogen than free IGF-I. Endocrinology
125
, 766–772.
Bradshaw, S.L., Naus, C.C., Zhu, D., Kidder, G.M., D’Ercole, A.J., and Han,
V.K. (1993) Alterations in the synthesis of insulin-like growth factor bind-
ing proteins and insulin-like growth factors in rat C6 glioma cells trans-
fected with a gap junction connexin43 cDNA. Regul. Pept.
48
, 99–112.
Buckbinder, L., Talbott, R., Velasco-Miguel, S., Takenaka, I., Faha, B.,
Seizinger, B.R., and Kley, N. (1995) Induction of the growth inhibitor IGF-
binding protein 3 by p53. Nature
377
, 646–649.
Bui, T., Kuo, C., Rotwein, P., and Straus, D.S. (1997) Prostaglandin A2 specif-
ically represses insulin-like growth factor-I gene expression in C6 rat
glioma cells. Endocrinology
138
, 985–993.
Burgaud, J.L., Resnicoff, M., and Baserga, R. (1995) Mutant IGF-I receptors as
dominant negatives for growth and transformation. Biochem. Biophys.
Res. Commun.
214
, 475–481.
Busby, W.H., Jr., Klapper, D.G., and Clemmons, D.R. (1988) Purication of a
31,000-dalton insulin-like growth factor binding protein from human
amniotic uid. J. Biol. Chem.
263
, 14203–14210.
Cavenee, W.K., Bigner, D.D., Newcomb, E.W., Paulus, W., and Kleihues, P.
(1997) Diffuse astrocytomas. In: Kleihues, P., and Cavenee, W.K. (Eds.),
Pathology and Genetics of Tumours of the Nervous System. Lyon: Inter-
national Agency for Research on Cancer. pp. 2–9.
Chernausek, S.D. (1993) Insulin-like growth factor-1 (IGF-I) production by
astroglial cells regulation and importance for epidermal growth factor-
induced cell replication. J. Neurosci. Res.
34
, 189–197.
Coppola, D., Ferber, A., Miura, M., Sell, C., D’Ambrosio, C., Rubin, R., and
Baserga, R. (1994) A functional insulin-like growth factor I receptor is
required for the mitogenic and transforming activities of the epidermal
growth factor receptor. Mol. Cell. Biol.
14
, 4588–4595.
De Mellow, J.S.M., and Baxter, R.C. (1988) Growth hormone-dependent
insulin-like growth factor (IGF) binding protein both inhibits and potenti-
ates IGF-I stimulated DNA synthesis in human skin broblasts. Biochem.
Biophys. Res. Commun.
156
, 199–204.
Ekstrand, A.J., James, C.D., Cavenee, W.K., Seliger, B., Pettersson, R.F., and
Collins, V.P. (1991) Genes for epidermal growth factor receptor, trans-
forming growth factor alpha, and epidermal growth factor and their
expression in human gliomas in vivo. Cancer Res.
51
, 2164–2172.
Ekstrand, A.J., Liu, L., He, J., Hamid, M.L., Longo, N., Collins, V.P., and James,
C.D. (1995) Altered subcellular location of an activated and tumour-asso-
ciated epidermal growth factor receptor. Oncogene
10
, 1455–1460.
Elgin, R.G., Busby, W.H., Jr., and Clemmons, D.R. (1987) An insulin-like
growth factor (IGF) binding protein enhances the biologic response to
IGF-I. Proc. Natl. Acad. Sci. USA
84
, 3254–3258.
Faber-Elman, A., Solomon, A., Abraham, J.A., Marikovsky, M., and Schwartz,
M. (1996) Involvement of wound-associated factors in rat brain astrocyte
migratory response to axonal injury: In vitro simulation. J. Clin. Invest.
97
,
162–171.
Fritsche, M., Haessler, C., and Brandner, G. (1993) Induction of nuclear accu-
mulation of the tumor-suppressor protein p53 by DNA-damaging agents.
Oncogene
8
, 307–318.
Gammeltoft, S., Ballotti, R., Nielsen, F.C., Kowalski, A., and Van Obberghen,
E. (1988a) Receptors for insulin-like growth factors in the central nervous
system: sStructure and function. Horm. Metab. Res.
20
, 436–442.
Gammeltoft, S., Ballotti, R., Kowalski, A., Westermark, B., and Van
Obberghen, E. (1988b) Expression of two types of receptor for insulin-like
growth factors in human malignant glioma. Cancer Res.
48
, 1233–1237.
Garcia-Estrada, J., Garcia-Segura, L.M., and Torres-Aleman, I. (1992) Expres-
sion of insulin-like growth factor I by astrocytes in response to injury. Brain
Res.
592
, 343–347.
Glick, R.P., Gettleman, R., Patel, K., Lakshman, R., and Tsibris, J.C. (1989)
Insulin and insulin-like growth factor I in brain tumors: Binding and in vitro
effects. Neurosurgery
24
, 791–797.
Goad, D.L., Rubin, J., Wang, H., Tashjian, A.H., Jr., and Patterson, C. (1996)
Enhanced expression of vascular endothelial growth factor in human
SaOS-2 osteoblast-like cells and murine osteoblasts induced by insulin-like
growth factor I. Endocrinology
137
, 2262–2268.
Hollander, M., and Wolfe, D.A. (1973) Nonparametric Statistical Methods.
New York: John Wiley & Sons.
Hongo, A., D’Ambrosio, C., Miura, M., Morrione, A., and Baserga, R. (1996)
Mutational analysis of the mitogenic and transforming activities of the
insulin-like growth factor I receptor. Oncogene
12
, 1231–1238.
Hussaini, I.M., Lopes, M.B.S., Musante, D.B., Karns, L.R., and VandenBerg,
S.R. (1996) Expression of IGF-1 and IGF-1R in reactive astrocytes and
astrocytic tumors. J. Neuropathol. Exp. Neurol.
55
, 606. (Abstract)
Jaros, E., Perry, R.H., Adam, L., Kelly, P.J., Crawford, P.J., Kalbag, R.M. Mende-
low, A.D., Sengupta, R.P., and Pearson, A.D. (1992) Prognostic implica-
tions of p53 protein, epidermal growth factor receptor, and Ki-67 labelling
in brain tumours. Br. J. Cancer
66
, 373–385.
Jennings, M.T., Maciunas, R.J., Carver, R., Bascom, C.C., Juneau, P., Misulis,
K., and Moses, H.L. (1991) TGF beta 1 and TGF beta 2 are potential
growth regulators for low-grade and malignant gliomas in vitro: Evidence
in support of an autocrine hypothesis. Int. J. Cancer
49
, 129–139.
Jones, J.I., and Clemmons, D.R. (1995) Insulin-like growth factors and their
binding proteins: Biological actions. Endocr. Rev.
16
, 3–34.
Kaleko, M., Rutter, W.J., and Miller, A.D. (1990) Overexpression of the human
insulinlike growth factor I receptor promotes ligand-dependent neoplastic
transformation. Mol. Cell. Biol.
10
, 464–473.
Kim, K.W., Bae, S.K., Lee, O.H., Bac, M.H., Lee, M.J., and Park, B.C. (1998)
Insulin-like growth factor II induced by hypoxia may contribute to angio-
genesis of human hepatocellular carcinoma. Cancer Res.
58
, 348–351.
Kluge, A., Zimmermann, R., Weihrauch, D., Mohri, M., Sack, S., Schaper, J.,
and Schaper, W. (1997) Coordinate expression of the insulin-like growth
factor system after microembolisation in porcine heart. Cardiovasc. Res.
33
, 324–331.
Kulik, G., Klippel, A., and Weber, M.J. (1997) Antiapoptotic signalling by the
insulin-like growth factor I receptor, phosphatidylinositol 3-kinase, and
Akt. Mol. Cell. Biol.
17
, 1595–1606.
Landau, B.K., Kwaan, H.C., Verrusio, E.N., and Brem, S.S. (1994) Elevated lev-
els of urokinase-type plasminogen activator and plasminogen activator
inhibitor type-1 in malignant human brain tumors. Cancer Res.
54
,
1105–1108.
Li, S., Resnicoff M., and Baserga, R. (1996) Effect of mutations at serines
1280-1283 on the mitogenic and transforming activities of the insulin-like
growth factor I receptor. J. Biol. Chem.
271
, 12254–12260.
Louis, D.N. (1997) A molecular genetic model of astrocytoma histopathology.
Brain Pathol.
7
, 755–764.
McCusker, R.H., Camacho-Hubner, C., Bayne, M.L., Cascieri, M.A., and
Clemmons, D.R. (1990) Insulin-like growth factor (IGF) binding to human
broblast and glioblastoma cells: The modulating effect of cell released
IGF binding proteins (IGFBPs). J. Cell. Physiol.
144
, 244–253.
Merrill, M.J., and Edwards, N.A. (1990) Insulin-like growth factor-I receptors
in human glial tumors. J. Clin. Endocrinol. Metab.
71
, 199–209.
Mohan, S., Bautista, C.M., Wergedal, J., and Baylink, D.J. (1989) Isolation of
an inhibitory insulin-like growth factor (IGF) binding protein from bone
cell-conditioned medium a potential local regulator of IGF action. Proc.
Natl. Acad. Sci. USA
86
, 8338–8342.
Nakao-Hayashi, J., Ito, H., Kanayasu, T., Morita, I., and Murota, S. (1992)
Stimulatory effects of insulin and insulin-like growth factor I on migration
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from

H. Hirano et al.: IGF-1 in astrocytomas
Neuro-Oncology APRIL 1 999 119
and tube formation by vascular endothelial cells. Atherosclerosis
92
,
141–149.
Punglia, R.S., Lu, M., Hsu, J., Kuroki, M., Tolentino, M.J., Leough, K., Levy,
A.P., Levy, N.S., Goldberg, M.A., D’Amato, R.J., and Adamis, A.P. (1997)
Regulation of vascular endothelial growth factor expression by insulin-like
growth factor I. Diabetes
46
, 1619–1626.
Rao, G.N., Delafontaine, P., and Runge, M.S. (1995) Thrombin stimulates
phosphorylation of insulin-like growth factor-1 receptor, insulin receptor
substrate-1, and phospholipase C-gamma 1 in rat aortic smooth muscle
cells. J. Biol. Chem.
270
, 27871–27875.
Resnicoff, M., Burgaud, J.L., Rotman, H.L., Abraham, D., and Baserga, R.
(1995) Correlation between apoptosis, tumorigenesis, and levels of
insulin-like growth factor I receptors. Cancer Res.
55
, 3739–3741.
Resnicoff, M., Li, W., Basak, S., Herlyn, D., Baserga, R., and Rubin, R. (1996)
Inhibition of rat C6 glioblastoma tumor growth by expression of insulin-
like growth factor I receptor antisense mRNA. Cancer Immunol.
Immunother.
42
, 64–68.
Roberts, C.T., Jr. (1996) Control of insulin-like growth factor (IGF) action by
regulation of IGF-I receptor expression. Endocrine J.
43
(Suppl.), S49–55.
Ross, M., Francis, G.L., Szabo, L., Wallace, J.C., and Ballard, F.J. (1989) Insulin-
like growth factor (IGF)-binding proteins inhibit the biological activities of
IGF-I and IGF-2 but not des-(1-3)-IGF-1. Biochem. J.
258
, 267–272.
Rubin, R., and Baserga, R. (1995) Insulin-like growth factor-I receptor. Its role
in cell proliferation, apoptosis, and tumorigenicity. Lab. Invest.
73
,
311–331.
Rubio, M.P., von Deimling, A., Yandell, D.W., Wiestler, O.D., Gusella, J.F., and
Louis, D.N. (1993) Accumulation of wild-type p53 protein in human
astrocytomas. Cancer Res.
53
, 3465–3467.
Sandberg, A.C., Engberg, C., Lake, M., von Holst, H., and Sara, V.R. (1988)
The expression of insulin-like growth factor I and insulin-like growth fac-
tor II genes in the human fetal and adult brain and in glioma. Neurosci.
Lett.
93
, 114–119.
Sanlippo, J.S., Rao, C.V., Guarnaschelli, J.J., Woost, P.G., Byrd, V.M., Jones,
E., and Schultz, G.S. (1993) Detection of epidermal growth factor and
transforming growth factor alpha protein in meningiomas and other
tumors of the central nervous system in human beings. Surg. Gynecol.
Obstet.
177
, 488–496.
Sasaoka, T., Rose, D.W., Jhun, B.H., Saltiel, A.R., Draznin, B., and Olefsky, J.M.
(1994) Evidence for a functional role of Shc proteins in mitogenic signal-
ing induced by insulin, insulin-like growth factor-1 and epidermal growth
factor. J. Biol. Chem.
269
, 13689–13694.
Sato, Y., Okamura, K., Morimoto, A., Hamanaka, R., Hamaguchi, K., Shimada,
T., Ono, M., Kohno, K., Sakata, T., and Kuwano, M. (1993) Indispensable
role of tissue-type plasminogen activator in growth factor-dependent tube
formation of human microvascular endothelial cells in vitro. Exp. Cell Res.
204
, 223–229.
Sell, C., Rubini, M., Rubin, R., Liu, J.P., Efstratiadis, A., and Baserga, R. (1993)
Simian virus 40 large tumor antigen is unable to transform mouse embry-
onic broblasts lacking type 1 insulin-like growth factor receptor. Proc.
Natl. Acad. Sci. USA
90
, 11217–11221.
Shimasaki, S., Uchiyama, F., Shimonaka, M., and Ling, N. (1990) Molecular
cloning of the cDNAs encoding a novel insulin-like growth factor-binding
protein from rat and human. Mol. Endocrinol.
4
, 1451–1458.
Straus, D.S., and Burke, E.J. (1995) Glucose stimulates IGF-I gene expression
in C6 glioma cells. Endocrinology
136
, 365–368.
Takahashi, J.A., Mori, H., Fukumoto, M., Igarashi, K., Jaye, M., Oda, Y.,
Kikuchi, H., and Hatanaka, M. (1990) Gene expression of broblast
growth factors in human gliomas and meningiomas: Demonstration of
cellular source of basic broblastic growth factor mRNA and peptide in
tumor tissues. Proc. Natl. Acad. Sci. USA
87
, 5710–5714.
Tournier, C., Pomerance, M., Gavaret, J.M., and Pierre, M. (1994) MAP kinase
cascade in astrocytes. Glia
10
, 81–88.
Van der Ven, L.T., Van Buul-Offers, S.C., Gloudemans, T., Roholl, P.J., Sussen-
bach, J.S., and Den Otter, W. (1997) Histamine-stimulated expression of
insulin-like growth factors in human glioma cells. Br. J. Cancer
75
,
1091–1097.
Van Meir, E.G. (1995) Cytokines and tumors of the central nervous system.
Glia
15
, 264–288.
Wang, Z.H., Ma, J., Zeng, B.J., Catanese, V.M., Samuels, S., Gama Sosa, M.A.,
and Kolodny, E.H. (1997) Correlation of glioma cell regression with inhi-
bition of insulin-like growth factor 1 and insulin-like growth factor-bind-
ing protein-2 expression. Neuroendocrinology
66
, 203–211.
Werner, H., Karnieli, E., Rauscher, F.J., and LeRoith, D. (1996) Wild-type and
mutant p53 differentially regulate transcription of the insulin-like growth
factor I receptor gene. Proc. Natl. Acad. Sci. USA
93
, 8318–8323.
Westermark, B., Heldin, C.H., and Nister, M. (1995) Platelet-derived growth
factor in human glioma. Glia
15
, 257–263.
Yamada, M., Ikeuchi, T., and Hatanaka, H. (1997) The neurotrophic action
and signalling of epidermal growth factor. Prog. Neurobiol.
51
, 19–37.
Yamamoto, M., Sawaya, R., Mohanam, S., Rao, V.H., Bruner, J.M., Nicolson,
G.L., Ohshima, K., and Rao, J.S. (1994) Activities, localizations, and roles
of serine proteases and their inhibitors in human brain tumor progression.
J. Neurooncol.
22
, 139–151.
Yang, B.C., Chang, H.M., Wang, Y.S., Chen, R.F., and Lin, S.J. (1996) Tran-
sient induction of apoptosis in serum-starved glioma cells by insulin and
IGF-1. Biochim. Biophys. Acta
1314
, 83–92.
Zheng, B., Clarke, J.B., Busby, W.H., Duan, C., and Clemmons, D.R. (1998)
Insulin-like growth factor-binding protein-5 is cleaved by physiological
concentrations of thrombin. Endocrinology
139
, 1708–1714.
Zumkeller, W., Saaf, M., and Rahn, T. (1993) Insulin-like growth factor (IGF)-
I, -II and IGF-binding proteins in the cyst uid of a patient with astrocy-
toma. Childs Nerv. Sys.
9
, 100–103.
by guest on July 14, 2011neuro-oncology.oxfordjournals.orgDownloaded from