Content uploaded by Joseph R. Testa
Author content
All content in this area was uploaded by Joseph R. Testa
Content may be subject to copyright.
Insulin-like growth factor 1 receptor is a potential
therapeutic target for gastrointestinal stromal tumors
Chi Tarn*
†
, Lori Rink*
†
, Erin Merkel*, Douglas Flieder
‡
, Harsh Pathak*, Daphne Koumbi
§
, Joseph R. Testa
§
,
Burton Eisenberg
¶
, Margaret von Mehren*, and Andrew K. Godwin*
储
Departments of *Medical Oncology and ‡Pathology and §Human Genetics Program, Fox Chase Cancer Center, Philadelphia, PA 19111; and ¶Norris Cotton
Cancer Center, Dartmouth–Hitchcock Medical Center, Lebanon, NH 03756
Communicated by Alfred G. Knudson, Jr., Institute for Cancer Research, Philadelphia, PA, April 8, 2008 (received for review January 25, 2008)
A subset of gastrointestinal stromal tumors (GISTs) lack gain-of-
function mutations in c-KIT and PDGFR
␣
. These so-called wild-type
(WT) GISTs tend to be less responsive to imatinib-based therapies
and have a poor prognosis. We identified amplification of IGF1R in
a SNP analysis of GIST and thus studied its potential as a thera-
peutic target in WT and mutant GIST. Expression of IGF1R and
downstream effectors in clinical GIST samples was examined by
using immunoblots and immunohistochemistry. The roles of IGF1R
signaling in GIST and viability were analyzed by using NVP-
AEW541, an inhibitor of IGF1R, alone and in combination with
imatinib, or via siRNA silencing of IGF1R. IGF1R was strongly
overexpressed, and IGF1R amplification was detected at a signif-
icantly higher frequency in WT GISTs, including a pediatric WT GIST,
compared with mutant GISTs (Pⴝ0.0173 and Pⴝ0.0163, respec-
tively). Inhibition of IGF1R activity in vitro with NVP-AEW541 or
down-regulation of expression with siIGF1R led to cytotoxicity and
induced apoptosis in GIST cell lines via AKT and MAPK signaling.
Combination of NVP-AEW541 and imatinib in GIST cell lines in-
duced a strong cytotoxicity response. Our results reveal that IGF1R
is amplified and the protein is overexpressed in WT and pediatric
GISTs. We also demonstrate that the aberrant expression of IGF1R
may be associated with oncogenesis in WT GISTs and suggest an
alternative and/or complementary therapeutic regimen in the
clinical management of all GISTs, especially in a subset of tumors
that respond less favorably to imatinib-based therapy.
pediatric GIST 兩tyrosine kinase inhibitors 兩imatinib mesylase 兩
adult wild-type GIST 兩NYP-AEW541
Gastrointestinal stromal tumors (GISTs) are the most com-
mon mesenchymal tumors of the digestive tract. Mazur and
Clark originally described these tumors in 1983, noting that they
contained smooth muscle and neural elements (1). Clinically,
diagnosis of GIST is typically confirmed by immunohistochem-
ical staining of a 145-kDa transmembrane glycoprotein, KIT,
referred to as CD117. Molecular genetic studies have shown that
the vast majority of primary GISTs (⬇70%) possess gain-of-
function mutations of c-KIT in exon 9, 11, 13, or 17, and that a
subset of GISTs (⬇10%) possess gain-of-function mutations of
PDGFR
␣
in exon 12, 14, or 18 (2–4).
Imatinib mesylate is an oral 2-phenylaminopyrimidine deriv-
ative that acts as a selective inhibitor against type III tyrosine
kinases such as KIT, PDGFR
␣
and BCR-ABL (the causative
chimeric fusion protein in chronic myelogenous leukemia) (5).
Since its approval by the FDA in 2002, it has been successfully
administered to treat patients with metastatic and/or unresect-
able GISTs (6, 7). Response to imatinib is correlated with the
type of c-KIT and PDGFR
␣
mutation present in a given tumor.
GIST patients with exon 11 c-KIT mutations have the best
response and disease-free survival, whereas GIST with non-exon
11 mutations or wild-type (WT) have a poorer disease-free
survival and overall survival (8, 9). The small but significant
portion of GIST patients (10–20%) whose tumors lack muta-
tions in either c-KIT or PDGFR
␣
or who possess ‘‘imatinib-
resistant’’ mutations (such as exon 17 mutations in KIT and exon
18 mutations in PDGFR
␣
) have lower response rates and shorter
disease-free progression compared with patients whose tumors
have exon 11 mutations.
A subset of GISTs arises in the pediatric age group. A typical
pediatric patient presents with anemia secondary to gastroin-
testinal bleeding, has a median age of 12, and is female (10). Only
15% of tumors will have evidence of a c-KIT or PDGFR
␣
mutation (11). In general, the growth of these tumors is more
indolent, and surgery is a mainstay of therapy; however, tumors
can metastasize. There is limited benefit of available drug
therapies, including imatinib in pediatric patients. Therefore,
identifying additional genetic factors that contribute to the
pathogenesis of GIST, independent of KIT and PDGFR
␣
, may
be helpful in developing additional, individually tailored
anti-GIST therapies.
Insulin-like growth factor (IGF) signaling plays a critical role
in the growth and development of many tissues and regulates
overall cell growth (12). The IGF system is composed of the IGF
ligands (IGF-1 and IGF-2), receptors (IGFR), the insulin re-
ceptor (IR) and six regulatory IGF-binding proteins (IGFBPs).
IGF1R is a transmembrane receptor that interacts with both
IGF-1 and IGF-2. IGF1R is activated via autophosphorylation
after ligand binding, thereby leading to activation of the mitogen-
activated protein kinase (MAPK) and phosphatidylinositol
3-kinase (PI3K) cascades (13). Because overexpression of
IGF1R has been identified in several tumor types (14–17) and
because of its role in metabolism, which potentially has relevance
to the survival of malignant cells, IGF1R has become a target for
anticancer therapy.
Although aberrant expression of IGF1R is detected in many
different types of cancer, little is known regarding the molecular
mechanism behind it. It was first reported in 1994 that chromo-
some 15q26, where IGF1R is located, was amplified in ⬍10% of
breast cancers (18). Recently, others have reported IGF1R
amplification at low levels in pancreatic adenocarcinoma xeno-
grafts and in two gastric cancer cell lines and in a small
percentage of Wilms’ tumors (19, 20). In this work, we have
found that IGF1R is highly expressed in adult and pediatric WT
GISTs compared with GISTs with c-KIT or PDGFR
␣
mutations.
Using genomic real-time PCR and interphase fluorescence in
situ hybridization (FISH), we have determined that a significant
portion of WT GISTs and in a pediatric case have IGF1R gene
amplification. We also show that a tyrosine kinase inhibitor,
Author contributions: C.T., L.R., M.v.M., and A.K.G. designed research; C.T., L.R., E.M., H.P.,
and D.K. performed research; C.T., L.R., E.M., D.F., and J.R.T. analyzed data; C.T., L.R., B.E.,
M.v.M., and A.K.G. wrote the paper.
The authors declare no conflict of interest.
†C.T. and L.R. contributed equally to this work.
储To whom correspondence should be addressed at: Department of Medical Oncology,
Fox Chase Cancer Center, 333 Cottman Avenue, Philadelphia, PA 19111. E-mail:
andrew.godwin@fccc.edu.
This article contains supporting information online at www.pnas.org/cgi/content/full/
0803383105/DCSupplemental.
© 2008 by The National Academy of Sciences of the USA
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0803383105 PNAS
兩
June 17, 2008
兩
vol. 105
兩
no. 24
兩
8387–8392
MEDICAL SCIENCES
NVP-AEW541, which targets IGF1R (21), has significant inhib-
itory effects on IGF1R phosphorylation and on GIST cell
proliferation in vitro, regardless of the KIT mutational status and
IGF1R expression levels. Furthermore, knocking down IGF1R
expression alone by siRNA silencing could induce cytotoxicity,
even in the presence of activated KIT. Our findings support the
conclusion that IGF1R is driving GIST pathogenesis in tumors
lacking c-KIT and PDGFR
␣
activating mutations. Therefore,
IGF1R should be considered a therapeutic target in all GIST
patients, especially adult and pediatric GISTs that express high
levels of IGF1R.
Results
IGF1R Is Highly Expressed in WT GISTs. High-density single-
nucleotide polymorphism (SNP) arrays were used to evaluate
changes in gene copy number in GISTs. We subjected 24 GISTs
and two GIST cell lines, i.e., GIST 882 and GIST-T1 cells, to
copy number analysis by using the Affymetrix 50K Xba array. We
found frequent loss of chromosomal regions, such as 1p, and
monosomy for 9, 14, 15, and 22 that are common to most GISTs
(refs. 22 and 23 and data not shown). In addition, we uncovered
focal regions of amplification, including the IGF1R locus [sup-
porting information (SI) Table S1 and Y. Skorogabotko, M.
Belinsky, and A.K.G., unpublished data]. Based on these obser-
vations, immunoblotting was done on fresh-frozen GIST biop-
sies collected from Fox Chase Cancer Center for phospho-
IGF1R and total IGF1R expression. All tumors samples were
found to express KIT by standard immunohistochemical ap-
proaches. Of the 17 tumors examined, 14 possessed a c-KIT
mutation, 1 possessed two individual PDGFR
␣
missense muta-
tions within the same allele, and 2 lacked detectable mutations
in either gene. Mutation analysis was carried out as described in
ref. 4. Immunoblotting revealed that IGF1R was expressed and
activated in all GISTs but was markedly overexpressed (10- to
30-fold) in WT GISTs compared with mutant GISTs (Fig. 1A,
lanes with asterisk signs). However, phospho-IGF1R levels did
not correlate well with overall IGF1R levels. Because there is no
other known oncogenic driving force in WT GISTs, we initially
concluded that overexpression of IGF1R may be a key tumor-
igenic event for this subset of GISTs.
IGF1R
Mutational and Gene Amplification Analyses. We next sought
to determine whether IGF1R is mutated in WT GISTs. We were
able to isolate DNA from 10 fresh-frozen W T GISTs collected
by needle biopsy. We examined the tumor DNA for potential
gain-of-function mutations in IGF1R and performed mutational
analyses of the exons encoding the juxtamembrane domain and
the entire kinase domain of the receptor. No mutations in IGF1R
were found in the WT GISTs. We detected a polymorphism
(IGF1R-c.3129A ⬎G; p.E1013E) in exon 16 of IGF1R in 30%
of the WT GIST samples (3 of 10 samples) that was also found
in 40% of an age/race/gender-matched disease-free control
population (data not shown).
To validate the SNP array results and determine whether
enhanced expression of IGF1R might be associated with gene
amplification, we developed a genomic-based quantitative PCR
assay to evaluate IGF1R gene copy number in mutant and WT
GISTs. When tested on WT GISTs, we demonstrated that 7 of
the 10 WT GISTs possessed amplified IGF1R (copy number
range, 2.5–4 copies), compared with only 5 of 18 mutant GISTs
showing amplification (P⫽0.04) (Fig. S1). IGF1R gene ampli-
fication was also confirmed by FISH (Fig. S2 and Table S2).
These results confirm that enhanced expression of IGF1R in a
subset of GISTs is in part associated with gene amplification.
After demonstrating by Western blot analysis that IGF1R is
abundantly expressed in WT GISTs (Fig. 1 Aand data not
shown), we tested whether immunohistochemistry (IHC) could
be used to evaluate IGF1R levels in clinical samples rapidly. We
accessed 8 paraffin-embedded WT GISTs, a pediatric GIST, and
16 mutant GIST samples. Slides were stained for IGF1R and
KIT expression by IHC and scored according to the criteria
described in Materials and Methods. Fig. 1Bshows representative
examples of IGF1R expression for WT, mutant GISTs, and
pediatric GISTs. For the 16 mutant GISTs, the majority showed
low or no detectable levels of IGF1R, and none of these tumors
was found to express very high levels (overall score of ⬎2) (Table
S1). In comparison, all of the WT GISTs, including the pediatric
GIST, showed intense IGF1R staining throughout the tumor
(overall score of ⱖ2) (P⫽0.023, two-sided Fisher’s exact test).
These data further confirm the Western blotting results and
indicate that abundant IGF1R expression is much more prom-
inent in WT GISTs (both adult and pediatric) than in mutant
GISTs. We also evaluated expression of a number of signaling
molecules downstream of IGF1R and KIT, i.e., phospho-AKT
(pAKT), phospho-mTOR (pmTOR), and phospho-S6 ribosomal
protein (pS6), a protein target for activated mTOR. Interest-
ingly, we observed that the majority of mutant GISTs had higher
levels of pAKT and pmTOR compared with WT GISTs (Table
S3), suggesting that both WT and mutant GISTs are under
oncogenic stimulation albeit driven by different ‘‘oncogenic
forces.’’
Overall, we found that although IGF1R amplification is not
seen exclusively in WT GISTs, it is clearly more prevalent in this
subset of GISTs (70% in WT GISTs vs. 27.7% in mutant GISTs,
P⫽0.01632, two-sided Fisher’s exact test). Furthermore, al-
though only a single case was available, it is intriguing to
demonstrate that IGF1R is amplified and overexpressed in
pediatric GIST, which provides evidence that abnormal regula-
tion of IGF1R may be driving GIST pathogenesis in tumors
lacking c-KIT and PDGFR
␣
activating mutations. We concluded
that IGF1R could be considered a therapeutic target for WT
GISTs, including pediatric cases, if not for all GISTs, because the
protein is constitutively activated in the vast majority of tumors.
Fig. 1. IGF1R expression in GIST biopsies. (A) Immunoblot assays of 17
consecutive fresh-frozen GIST biopsies with anti-phospho-IGF1R and anti-
IGF1R antibodies. One hundred micrograms of WCE from each sample was
subjected to immunoblotting. The c-KIT or PDGFR
␣
genotype for each GIST is
listed below (⫹or ⫺). Asterisks indicate WT GISTs. (B) Immunohistochemistry
of WT, mutant, and pediatric GISTs for IGF1R expression. A score of 3 is
considered marked expression (all tumor cells express high levels of IGF1R).
8388
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0803383105 Tarn et al.
Inhibition of IGF1R Signaling with NVP-AEW541. To test the effi-
ciency of the small molecule inhibitor of IGF1R, NVP-AEW541
(Novartis) (24), we first examined MCF7 cells, which are known
to have high IGF1R levels (25), and compared the expression of
both pIGF1R and total IGF1R levels in MCF7 cells vs. two of
the WT GISTs (Fig. 2A). As indicated, no cell lines have yet to
be established from WT GISTs; therefore, we used MCF7 cells
as a surrogate to evaluate the activity of NVP-AEW541 initially.
As shown in Fig. 2A, MCF7 cells have ⬇2- to 10-fold higher
levels of IGF1R than the WT tumors, but they lack phosphor-
ylation in the absence of IGF-1 stimulation. Stimulation with
IGF-1 after serum starvation significantly enhanced pIGF1R
in MCF7 cells, which could be completely inhibited by
NVP-AEW541 (Fig. 2B).
Given that IGF1R is expressed and constitutively activated in
GISTs, we next examined the effect of targeting IGF1R by using
two GIST cell lines (26). Both GIST cell lines express IGF1R,
with higher constitutive levels detected in GIST 882 cells.
Importantly, IGF1R is autophosphorylated after serum starva-
tion, and stimulation with IGF-1 increased phosphor ylated levels
of IGF1R (Fig. 2Cand data not shown). As with MCF7 cells,
NVP-AEW541 could inhibit IGF1R phosphorylation after
ligand stimulation (Fig. 2C).
We next evaluated the consequence of inhibition of IGF1R by
using pAKT and pGSK3

, and pMAPK1/2 as molecular surro-
gates of downstream IGF1R signaling. We have evaluated GIST
882 and GIST-T1 cells for response to imatinib and have shown
them to be highly reliable models that represent the clinical
experience, in that T1 cells that possess an exon 11 c-KIT
mutation are more sensitive to imatinib than 882 cells that have
an exon 13 mutation (26, 27). Using standard cytotoxicity assays,
we determined the IC
50
of NVP-AEW541 for GIST 882 cells and
for GIST-T1 cells to be ⬇3.9 ⫾1.2
M and ⬇3.7 ⫾0.1
M,
respectively. We had shown that GIST 882 cells were signifi-
cantly more resistant to imatinib mesylate than to GIST-T1 cells
(26), but importantly, both were equally sensitive to NVP-
AEW541. Interestingly, the dose of NVP-AEW541 required to
achieve an IC
50
in MCF7 cells was ⬇3-fold lower (Fig. 2B) than
in the GIST-T1 and 882 cells, further suggesting that GIST with
high levels of IGF1R may be more sensitive to IGF1R-based
therapies. As expected, NVP-AEW541 did not have any effect on
pKIT levels. However, pAKT and pGSK3

were inhibited at 10
M NVP-AEW541 in both GIST cell lines (Fig. 2D). Interest-
ingly, the effect of IGF1R inhibition on MAPK1/2 signaling was
more cell type-dependent, perhaps associated with type of c-KIT
mutation, i.e., phosphorylation of MAPK1/2 was inhibited at 10
M NVP-AEW541 in GIST 882 cells, whereas only 1
M was
required in GIST-T1 cells (Fig. 2D).
NVP-AEW541 and Imatinib Have Additive Effects on GIST Cell Growth.
Growing evidence suggests that monotherapies may not be
sufficient to eradicate tumor cells effectively and that targeting
multiple pathways may be beneficial (28). Therefore, we exam-
ined the consequences of single and combined inhibition of
IGF1R and KIT/PDGFR
␣
in GIST cells. As shown in Fig. 3,
NVP-AEW541 or imatinib alone (at 10 nM) had little effect on
cell viability. However, when GIST-T1 and GIST 882 cells were
treated with the combination at doses that induce ⬇40% cyto-
toxicity, the decrease in cell viability was similar to or even
greater than with NVP-AEW541 or imatinib at the highest doses
(10
M). However, inhibiting either KIT or IGF1R activity with
these high doses of imatinib or NVP-AEW541 alone appears to
be sufficient to induce cytotoxicity and growth inhibition in
GIST cells in vitro, indicating little synergistic effect of these two
drugs when various concentrations of imatinib are combined
Fig. 2. NVP-AEW541 treatment inhibits constitutive activation of effectors
downstream of IGF1R. (A) Immunoblot assays of MCF7 cells and two fresh-
frozen WT GIST biopsies with specific anti-phospho-IGF1R and anti-IGF1R
antibodies. Fifty micrograms of WCE from each sample was subjected to
immunoblotting. (B) MCF7 cells were serum-starved for 12 h followed by
either NVP-AEW541 treatment for 12 h and/or IGF-1 stimulation for 30 min.
Immunoblot assays of phospho- and total IGF1R were performed. IC50 for
NVP-AEW541 was quantitated based on four independent viability assays
listed. (C) Immunoblot assays of phospho- and total IGF1R. GIST 882 cells were
serum-starved for 12 h followed by either NVP-AEW541 treatment for 12 h
and/or IGF-1 stimulation for 30 min. (D) GIST-T1 cells and GIST 882 cells were
treated with NVP-AEW541 for6hattheindicated concentrations. Equal
amounts (40
g) of WCE from each sample were subjected to immunoblotting
with specific antibodies as indicated. In all cases, actin served as a loading
control. IC50 for NVP-AEW541 was quantitated based on four independent
viability assays for each cell line.
Fig. 3. NVP-AEW541 and imatinib have additive cytotoxicity effects on
mutant GIST cells. Cell viability assays of GIST-T1 and GIST 882 cells 72 h after
treatment with NVP-AEW541 alone or in combination with imatinib at indi-
cated concentrations are shown. *,Pvalues between 0.01 and 2.5 ⫻10⫺7.
Tarn et al. PNAS
兩
June 17, 2008
兩
vol. 105
兩
no. 24
兩
8389
MEDICAL SCIENCES
with 10 nM NVP-AEW541 and when increasing concentrations
of NVP-AEW541 are combined with 10 nM imatinib (Fig. S3).
Next, we determined the consequence of single and combined
inhibition on the molecular surrogates of downstream RTK
signaling. As expected, the 10 nM NVP-AEW541 alone or
combined with 10 nM imatinib did not inhibit KIT, IGF1R (in
the absence of IGF-1 stimulation) or their downstream effectors
in GIST cells (Fig. S4). However, when GIST cells were treated
with higher doses (10
M) of NVP-AEW541, or with imatinib
alone, or with a combination of NVP-AEW541 and imatinib at
the IC
40
concentrations, the phosphorylation of MAPK1/2,
AKT, and GSK3

was inhibited. These results strongly demon-
strate that imatinib and NVP-AEW541 show a potentially ad-
ditive effect on KIT and IGF1R signaling in GIST, which could
be clinically beneficial because imatinib alone is not always
curative, especially in metastatic GIST. Indeed, treatment of
GIST cells with combined therapy at the IC
40
doses of both drugs
achieved a greater inhibitory effect than did either single agent
at the highest dose. Importantly, the combined IC
40
doses or 10
M NVP-AEW541 alone induced caspase 3 activity as evi-
denced by poly(ADP-ribose)polymerase (PARP) cleavage in
GIST-T1 cells and to a lesser extent in GIST 882 cells (Fig. S4
and data not shown).
IGF1R Silencing Induces Cytotoxicity in c-
KIT
Mutant GIST Cells. We
further validated IGF1R as a therapeutic target in GIST by using
RNAi approaches. We used GIST-T1 cells for these experiments
because this cell line has much higher transfection efficiency
than GIST 882 cells. We observed that IGF1R levels were
depleted by ⬇70% after electroporation with small interfering
RNAs to IGF1R (siRNA-IGF1R) (Fig. S5A). Both immunoblots
and densitometry quantitations showed that pKIT and pAKT
levels remained similar in siRNA-control vs. siRNA-IGF1R-
transfected samples; however, there was ⬇50% decrease in the
pMAPK1/2 levels (Fig. S5A,Lower, sicontrol vs. siIGF1R with-
out imatinib treatment). These results suggest that in KIT
mutant GIST cells, blocking IGF1R has minimal effect on AKT
activation; however, IGF1R depletion does affect MAPK acti-
vation. As expected, treatment with imatinib at concentrations
as low as 60 nM inhibits KIT and AKT phosphorylation.
Importantly, siRNA against c-KIT resulted in KIT knockdown
and inactivation of pAKT and pMAPK1/2 (Fig. S5B). Viability
assays revealed that siRNA-IGF1R alone resulted in significant
cell death (⬇40%), whereas siRNA-KIT caused nearly 80% cell
death compared with siRNA-control (Fig. 4). These data confirm
that in these mutant GIST cells, KIT activation is the primary
oncogenic event but that IGF1R activation, even in the presence
of mutant KIT, may also be contributing to tumorigenesis. Our
results also suggest that NVP-AEW541 may possess some un-
known off-target activities or that the level of siRNA suppression
of IGF1R was not sufficient to elicit complete inhibition of the
IGF1R signaling pathway.
Discussion
IGF1R is activated and signals its downstream pathways after
binding of its ligands, IGF1 and IGF-II. IGF1R signaling has
been implicated in many different human cancers based on
experimental model systems and population studies (29). Drugs
inhibiting this pathway are actively being tested in phase I and II
clinical trials (30–33). Although it has been shown by others that
IGF1R is expressed in a number of hematological neoplasias and
solid tumors, the molecular mechanism by which overexpression
of IGF1R contributes to cancer is still not clear. In addition, the
role of IGFBP-3, a regulator of circulating IGF1 levels, has
recently been studied in GISTs. Interestingly, Trent et al. (34)
reported that IGFBP-3 expression is regulated by imatinib in
GISTs and may play a role in the antitumor activity of imatinib.
We have shown that IGF1R expression is enhanced greatly in
numerous GISTs and that this enhanced expression is found
primarily in GISTs that lack c-KIT and PDGFR
␣
mutations (Fig.
1). Importantly, our results also demonstrated that IGF1R is
overexpressed and that the gene may be amplified, in pediatric
GIST, a subset of GISTs that commonly lack known oncogenic
mechanisms and demonstrate few genomic alterations (11). The
finding of high expression of IGF1R in the pediatric GIST
sample is in accordance with a previous publication (35). Fur-
thermore, our studies suggest that IGF1R may be a relevant
target for both mutant and WT GISTs (including adult and
pediatric cases), although an in vitro WT GIST model does not
currently exist. Importantly, these studies provide insights into
the pathogenesis of this disease, namely that IGF1R overexpres-
sion may be an important oncogenic event in the pathogenesis
of GISTs that lack c-KIT and PDGFR
␣
mutations, including
pediatric GISTs.
It was demonstrated that imatinib at equivalent drug concen-
trations has equal efficacy in decreasing the phosphorylation of
both WT induced with steel factor and a mutated c-KIT (36).
This finding is in contrast with the clinical experience where the
WT GISTs are less likely to have responses to imatinib therapy
and have shorter event-free and overall survival compared with
GISTs containing exon 11 mutations (8, 9). WT GISTs have an
increased relative risk of disease progression and death from
GIST of 108% and 76%, respectively, compared with GISTs
containing an exon 11 mutation. Therefore, metastatic GIST
patients with WT tumors would benefit from an alternative
treatment to imatinib alone, and blockade of the IGF1R pathway
may be one such approach.
Our results strongly suggest, because IGF1R is commonly
overexpressed in most WT GISTs, that IHC analysis of IGF1R
could serve as a rapid means for prescreening GIST, before
c-KIT and PDGFR
␣
genotyping. Furthermore, we have shown
that GISTs lack mutations in the IGF1R juxtamembrane and
kinase domains. However, we have demonstrated by genomic
qPCR, and we confirmed by FISH, that the majority of WT
GISTs, including a pediatric GIST, possess IGF1R amplification,
a genomic abnormality that could contribute in part, but not
necessarily totally, to the high levels observed. Although IGF1R
amplification is not confined to WT GISTs, it is clearly more
prevalent compared with mutant GISTs (P⫽0.01632). We have
also determined that treatment of GIST-T1 and GIST 882 cell
lines, with a demethylation agent, 5-aza-2⬘-deoxycytidine, did
not result in increased levels of IGF1R (data not shown). This
result suggests that transcriptional factors, rather than DNA
methylation, may contribute to the regulation of IGF1R in
GISTs, although additional studies will be necessary to identify
the mechanism. Furthermore, downstream signaling effectors,
such as AKT and mTOR, remain active in both WT and mutant
tumors (Table S3), although this pathway seems to be more
prominently activated in mutant tumors. This result indicates
that both WT and mutant GISTs are under oncogenic stimula-
Fig. 4. The cytotoxic effect of KIT and IGF1R silencing in GIST. Cells were
treated with imatinib at the indicated doses 24 h before 48-h harvest time
followed by cytotoxicity assays. *,Pvalues between 0.003 and 5.1 ⫻10⫺5;**,
Pvalues between 3 ⫻10⫺4and 5.4 ⫻10⫺6.
8390
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0803383105 Tarn et al.
tion albeit driven by different oncogenic forces. This concept is
similar to those reported in other types of cancer in which
multiple receptor tyrosine kinases (RTKs) are coactivated and
thereby limit the efficacy of therapies targeting a single RTK
(37). Our results suggest that aberrant IGF1R expression may be
one of several oncogenic forces in this subset of GISTs and thus,
represents a therapeutic target to consider in combination with
clearly effective, but not always curative, therapies. Even in
mutant GISTs, concomitant activation of multiple RTKs, e.g.,
IGF1R, may serve to reduce dependence on only KIT for the
maintenance of critical downstream signaling, thus rendering
some tumors refractory to imatinib or related RTK inhibitors.
Furthermore, because we demonstrated that the vast majority of
GISTs have constitutive activation of IGF1R in the absence of
gain-of-function mutations, we examined an IGF1R ligand,
IGF-1, in GIST tissue samples and blood from GIST patients for
the potential existence of an IGF-1 autocrine loop. Overall,
we found comparable levels of IGF-1 in WT and mutant
GISTs (data not shown), suggesting similar levels of
autophosphorylation of IGF1R via an autocrine loop.
There are now multiple agents available for the inhibition of
IGF1R signaling in cancer cells (38, 39), including IGF-1- and
IGF-2-neutralizing antibodies as well as monoclonal antibodies
and small tyrosine kinase inhibitors of the receptor itself (29).
We examined the drug NVP-AEW541, which has specific activ-
ity against IGF1R, and showed inhibitory effect of cell signaling
and growth in mutant GIST cells, which was independent of KIT
signaling. Unlike published data that suggest that inhibition of
IGF1R enhances the therapeutic effect of or sensitizes tumor
cells to standard drug treatment (40, 41), our results do not show
a synergistic effect of NVP-AEW541 and imatinib on GIST cell
viability.
Finally, we further examined the role of IGF1R in GIST cell
viability by using siRNA silencing. We show that, as with
NVP-AEW541 treatment at a high dose (10
M), IGF1R
knockdown induces cytotoxic effect in GIST cells, albeit to a
lesser degree compared with KIT knockdown (Fig. 4). This result
suggests that in GIST cell lines that harbor mutant c-KIT, the
gain-of-function of mutant KIT plays a dominant role in GIST
survival. However, unlike NVP-AEW541 treatment at a lethal
dose, IGF1R knockdown does not lead to AKT inactivation,
indicating that there may be some off-target effects of NVP-
AEW541. As with many newer molecularly targeted agents, few
have single target specificity. In fact, it is thought that ‘‘dirtier’’
drugs, such as sunitinib or AMG706, may be more effective in
the clinical setting as long as toxicities are manageable and
long-term administration is feasible (42, 43).
After we submitted this article, Braconi et al. (44) reported
that IGF1 and IGF2 expression correlated with relapse in GIST
patients, whereas IGF1R expression was strong in all cases.
Although different methods were used, taken together with our
current and previous observations it appears that abnormal
IGF1R signaling may indeed be involved in the pathogenesis of
GIST via KIT independent pathways (26).
In conclusion, we have shown that IGF1R is amplified and
overexpressed in the majority of GISTs that lack c-KIT or
PDGFR
␣
mutations. More importantly, we have shown that
imatinib-sensitive and -resistant GIST cells respond equally well
to a small molecular inhibitor of IGF1R, suggesting an alterna-
tive and/or complementary therapeutic regimen in the clinical
management of GIST, especially in tumors that respond less
favorably to imatinib-based therapy, including pediatric cases.
These findings are particularly exciting given the number of
agents targeting IGF1R that are currently being tested in clinical
trials. It is feasible in the near future to initiate clinical trials by
using IGF1R-targeted therapies for imatinib-refractor y GIST
patients, initially focusing on adult and pediatric GIST patients
lacking c-KIT or PDGFR
␣
mutations and expanding possibly to
all GIST patients.
Materials and Methods
Cell Cultures. GIST 882 cells that possess a homozygous mutation in KIT exon
13 and GIST-T1 cells possessing a heterozygous mutation in KIT exon 11 were
grown as described in ref. 26. MCF7 cells were cultured in DMEM with 10% FBS
and insulin (0.3 unit/ml). For drug treatment, drugs were added directly to the
cell medium at the indicated final concentration for the specified period.
Imatinib mesylate (Gleevec) was dissolved in sterile PBS and stored at ⫺20°C.
NVP-AEW541 was provided by Novartis and made into a 10 mM stock with
DMSO. All antibodies used in this work were purchased from Cell Signaling
Technologies, except

-actin, purchased from Sigma, and used according to
the manufacturer’s instructions. Lyophilized recombinant human IGF-1 was
purchased from R&D Systems, resuspended in PBS, and stored at ⫺20°C. The
secondary antibody and detection system used for immunohistochemical
staining was super-sensitive link-label (biotin-based) IHC detection systems
from Biogenex. For siRNA studies, a smart pool of double-stranded siRNA
against IGF1R (IGF1R-NM-000875) as well as nonspecific siRNA (D-001206-01)
were obtained from Dharmacon Tech. Cell line Nucleofector kit V for
electroporation was purchased from Amaxa Biosystems.
Preparation of Whole-Cell Extract (WCE) from Cells and Immunoblotting Assays.
The WCEs were prepared as described in ref. 26.
DNA Extraction and Mutational Analysis.
Genomic DNA isolation.
Frozen tumor
samples ⬇2 mm in diameter were homogenized in 200
l of lysis buffer [50
mM Tris䡠HCl (pH 7.5), 150 mM NaCl, 5 mM EDTA (pH 8.0), 0.5% Nonidet P-40].
After homogenization, genomic DNA was isolated by using Easy-DNA kit
(Invitrogen) following the manufacturer’s instructions. The purified DNA was
resuspended in 50
l of TE buffer.
Mutational analysis.
Primer pairs for IGF1R exons 15–20 are listed in Table S4.
Each PCR was performed in a 50-
l volume containing 10 ng of genomic DNA,
a15
M concentration of each primer, 0.2 mM dNTP, 5
lof10⫻reaction
buffer, 2.5 mM MgCl2, and 1 unit of AmpliTaq Gold DNA polymerase (Applied
Biosystems). The PCR conditions were: 30 s at 94°C, 30 s at 52°C, 1 min at 68°C
for 36 cycles followed by a 10-min extension at 68°C. PCR products were
analyzed in 2% agarose gel electrophoresis and purified by a PCR purification
kit (Qiagen). Direct sequencing was carried out from both directions by using
BigDye Terminator V3.1 cycle sequencing kit on an ABI PRISM 3100 genetic
analyzer (Applied Biosystems).
IHC Analysis. IHC staining was performed on 5-
m slides. After deparaffiniza-
tion and rehydration, sections were subjected to heat-induced epitope re-
trieval by immersion in a 0.01 M citrate buffer (pH 6.0). Endogenous peroxi-
dase activity was blocked for 15 min in 3% hydrogen peroxide in methanol.
Nonspecific binding was blocked by treatment with a blocking reagent (pro-
tein block serum-free; DAKO) for 30 min at room temperature. The slides were
then incubated overnight with primary antibody at 4°C in a humidified
chamber. All primary antibodies were diluted to a concentration of 1:50.
Immunodetection was performed by using the super-sensitive link-label
(biotin-based) IHC detection systems.
Histological scoring and analysis.
All IHC evaluation was performed in a blinded
manner by one author (D.F.). The following criteria were used to assess
distribution and intensity of positive tumor cell staining.
Distribution.
Absent tumor cell staining was scored as 0, ⬍10% of positive
tumor cells staining was scored 1, 10–50% of cells staining was scored as 2,
50 –90% of cells staining was scored as 3, and ⬎90 of cells staining was scored
as 4.
Intensity.
Absent staining in tumor cells was scored as 0, equivocal was scored
as 1, clearly positive was scored as 2, and strong positive staining was scored
as 3.
Scoring.
The results for intensity and distribution were summed, and a ‘‘score’’
assigned as follows: sum of 0, no staining (score 0); sum of 1–3, slight staining
(score 1); sum of 4–5, moderate staining (score 2); and sum of 6 –7, marked
staining (score 3).
Cell Proliferation/Viability Assay. Proliferation and viability were assessed with
an assay based on the cleavage of the tetrazolium salt WST-1 to formazan by
cellular mitochondrial dehydrogenases (Roche). GIST 882 and GIST-T1 cells
were seeded in 96-well plates at a density of 1.5 ⫻105cells per well. Twenty-
four hours later, cells were treated with varying doses of NVP-AEW541 and/or
imatinib mesylate. Cell proliferation and viability were measured 72 h after
Tarn et al. PNAS
兩
June 17, 2008
兩
vol. 105
兩
no. 24
兩
8391
MEDICAL SCIENCES
treatment with WST-1 reagent. The metabolic activity of viable cells was
quantified 4 h later with an EnVision microplate reader (PerkinElmer).
siRNA Transfection. GIST-T1 cells were trypsin-treated and resuspended in Amaxa
Nucleofector solution V at a concentration of 1 ⫻106cells per 100
l. Nucleofec-
tion was done by using the T20 program on the Nucleofector II device. Five
microliters of each 20
M siRNA was used for electroporation. After electropo-
ration, cells were washed with 3 ml of growth medium, counted, and seeded
according to experimental designs.
Statistical Analysis. All reported values are the means ⫾SEM. Statistical com-
parisons were determined with either one-sided (comparing tumor with normal
control) or two-sided (comparing WT GISTs with mutant GISTs) Fisher’s exact test.
Results were considered statistically significant if the Pvalue was ⬍0.05.
ACKNOWLEDGMENTS. We acknowledge the valuable input of Ms. Yuliya
Skorogabotko and Drs. Chong Xu, Samuel Litwin, and Alfred Knudson in this
work. We acknowledge the support of Ms. Tania Stutman and the GIST Cancer
Research Fund. This work was supported in part by National Institutes of
Health Grant CA106588 (to A.K.G.), FCCC Translational Research Committee
Grant 5P30CA06927-44 (to C.T, M.v.M., and A.K.G.), and by National Institutes
of Health Training Institutional National Research Service Award
CA009035-31 (to L.R.). J.R.T. was supported by National Institutes of Health
Grants CA77429, and J.R.T., M.V.M., and A.K.G. are supported by P30
CA06927. L.R. was supported by a GIST Cancer Research Fund fellowship.
1. Mazur MT, Clark HB (1983) Gastric stromal tumors: Reappraisal of histogenesis. Am J
Surg Pathol 7:507–519.
2. Heinrich MC, Corless CL (2005) Gastric GI stromal tumors (GISTs): The role of surgery in
the era of targeted therapy. J Surg Oncol 90:195–207.
3. Subramanian S, et al. (2004) Gastrointestinal stromal tumors (GISTs) with KIT and
PDGFRA mutations have distinct gene expression profiles. Oncogene 23:7780–7790.
4. Tarn C, et al. (2005) Analysis of KIT mutations in sporadic and familial gastrointestinal
stromal tumors: Therapeutic implications through protein modeling. Clin Cancer Res
11:3668–3677.
5. BuchdungerE, et al. (2000) Abl protein-tyrosine kinase inhibitor STI571 inhibits in vitro
signal transduction mediated by c-kit and platelet-derived growth factor receptors.
J Pharmacol Exp Ther 295:139–145.
6. Demetri GD, et al. (2002) Efficacy and safety of imatinib mesylate in advanced gastro-
intestinal stromal tumors. N Engl J Med 347:472–480.
7. Verweij J, et al. (2004) Progression-free survival in gastrointestinal stromal tumours
with high-dose imatinib: Randomised trial. Lancet 364:1127–1134.
8. Debiec-RychterM, et al. (2004) Use of c-KIT/PDGFRA mutational analysis to predict the
clinical response to imatinib in patients with advanced gastrointestinal stromal tu-
mours entered on phase I and II studies of the EORTC Soft Tissue and Bone Sarcoma
Group. Eur J Cancer 40:689– 695.
9. Heinrich MC, et al. (2003) Kinase mutations and imatinib response in patients with
metastatic gastrointestinal stromal tumor. J Clin Oncol 21:4342–4349.
10. Price VE, Zielenska M, Chilton-MacNeill S, Smith CR, Pappo AS (2005) Clinical and
molecular characteristics of pediatric gastrointestinal stromal tumors (GISTs). Pediatr
Blood Cancer 45:20–24.
11. Janeway KA, et al. (2007) Pediatric KIT wild-type and platelet-derived growth factor
receptor
␣
-wild-type gastrointestinal stromal tumors share KIT activation but not
mechanisms of genetic progression with adult gastrointestinal stromal tumors. Cancer
Res 67:9084–9088.
12. LeRoith D, Roberts CT, Jr (2003) The insulin-like growth factor system and cancer.
Cancer Lett 195:127–137.
13. Garrouste F, et al. (2002) Prevention of cytokine-induced apoptosis by insulin-like
growth factor-I is independent of cell adhesion molecules in HT29–D4 colon carcinoma
cells: Evidence for a NF-
B-dependent survival mechanism. Cell Death Differ 9:768–
779.
14. Gotlieb WH, et al. (2006) Insulin-like growth factor receptor I targeting in epithelial
ovarian cancer. Gynecol Oncol 100:389–396.
15. Sarfstein R, Maor S, Reizner N, Abramovitch S, Werner H (2006) Transcriptional regu-
lation of the insulin-like growth factor-I receptor gene in breast cancer. Mol Cell
Endocrinol 252:241–246.
16. WuJD, et al. (2006) Interaction of IGF signaling and the androgen receptor in prostate
cancer progression J Cell Biochem 99:392–401.
17. YanochkoGM, Eckhart W (2006) Type I insulin-like growth factor receptor overexpres-
sion induces proliferation and antiapoptotic signaling in a three-dimensional culture
model of breast epithelial cells. Breast Cancer Res 8:R18.
18. Almeida A, Muleris M, Dutrillaux B, Malfoy B (1994) The insulin-like growth factor I
receptor gene is the target for the 15q26 amplicon in breast cancer. Genes Chromo-
somes Cancer 11:63–65.
19. Armengol G, et al. (2000) DNA copy number changes and evaluation of MYC, IGF1R,
and FES amplification in xenografts of pancreatic adenocarcinoma. Cancer Genet
Cytogenet 116:133–141.
20. Natrajan R, et al. (2006) Blastemal expression of type I insulin-like growth factor
receptor in Wilms’ tumors is driven by increased copy number and correlates with
relapse. Cancer Res 66:11148–11155.
21. Hopfner M, et al. (2006) Blockade of IGF-1 receptor tyrosine kinase has antineoplastic
effects in hepatocellular carcinoma cells. Biochem Pharmacol 71:1435–1448.
22. El-Rifai W, Sarlomo-Rikala M, Andersson LC, Knuutila S, Miettinen M (2000) DNA
sequence copy number changes in gastrointestinal stromal tumors: Tumor progression
and prognostic significance. Cancer Res 60:3899–3903.
23. Meza-Zepeda LA, et al. (2006) Array comparative genomic hybridization reveals
distinct DNA copy number differences between gastrointestinal stromal tumors and
leiomyosarcomas. Cancer Res 66:8984– 8993.
24. Garcia-Echeverria C, et al. (2004) In vivo antitumor activity of NVP-AEW541: A novel,
potent, and selective inhibitor of the IGF-IR kinase. Cancer Cell 5:231–239.
25. HaileyJ, et al. (2002) Neutralizing anti-insulin-like growth factor receptor 1 antibodies
inhibit receptor function and induce receptor degradation in tumor cells. Mol Cancer
Ther 1:1349–1353.
26. Tarn C, et al. (2006) Therapeutic effect of imatinib in gastrointestinal stromal
tumors: AKT signaling dependent and independent mechanisms. Cancer Res
66:5477–5486.
27. Frolov A, et al. (2003) Response markers and the molecular mechanisms of action of
Gleevec in gastrointestinal stromal tumors. Mol Cancer Ther 2:699–709.
28. Yanada M, Naoe T (2006) Imatinib combined chemotherapy for Philadelphia
chromosome-positive acute lymphoblastic leukemia: Major challenges in current
practice. Leuk Lymphoma 47:1747–1753.
29. MillerBS, Yee D (2005) Type I insulin-like growth factor receptor as a therapeutic target
in cancer. Cancer Res 65:10123–10127.
30. Haluska P, et al. (2007) in ASCO Annual Meeting (American Society of Clinical Oncol-
ogy, Chicago, IL), p 3586.
31. KarpDD, et al. (2007) in ASCO Annual Meeting (American Society of Clinical Oncology,
Chicago, IL), p 7506.
32. RodonJ, et al. (2007) in ASCO Annual Meeting (American Society of Clinical Oncology,
Chicago, IL), p 3590.
33. Tolcher AW, et al. (2007) in ASCO Annual Meeting (American Society of Clinical
Oncology, Chicago, IL), p 3002.
34. TrentJC, et al. (2006) Early effects of imatinib mesylate on the expression of insulin-like
growth factor binding protein-3 and positron emission tomography in patients with
gastrointestinal stromal tumor. Cancer 107:1898–1908.
35. Prakash S, et al. (2005) Gastrointestinal stromal tumors in children and young adults:
A clinicopathologic, molecular, and genomic study of 15 cases and review of the
literature. J Pediatr Hematol Oncol 27:179–187.
36. HeinrichMC, et al. (2000) Inhibition of c-kit receptor tyrosine kinase activity by STI 571,
a selective tyrosine kinase inhibitor. Blood 96:925–932.
37. StommelJM, et al. (2007) Coactivation of receptor tyrosine kinases affects the response
of tumor cells to targeted therapies. Science 318:287–290.
38. Hopfner M, Baradari V, Huether A, Schofl C, Scherubl H (2006) The insulin-like growth
factor receptor 1 is a promising target for novel treatment approaches in neuroendo-
crine gastrointestinal tumours. Endocr Relat Cancer 13:135–149.
39. Sachdev D, Singh R, Fujita-Yamaguchi Y, Yee D (2006) Down-regulation of insulin
receptor by antibodies against the type I insulin-like growth factor receptor: Implica-
tions for anti-insulin-like growth factor therapy in breast cancer. Cancer Res 66:2391–
2402.
40. O’Reilly KE, et al. (2006) mTOR inhibition induces upstream receptor tyrosine kinase
signaling and activates Akt. Cancer Res 66:1500–1508.
41. Warshamana-Greene GS, et al. (2005) The insulin-like growth factor-I receptor kinase
inhibitor, NVP-ADW742, sensitizes small cell lung cancer cell lines to the effects of
chemotherapy. Clin Cancer Res 11:1563–1571.
42. Faivre S, et al. (2006) Safety, pharmacokinetic, and antitumor activity of SU11248, a
novel oral multitarget tyrosine kinase inhibitor, in patients with cancer. J Clin Oncol
24:25–35.
43. Morabito A, De Maio E, Di Maio M, Normanno N, Perrone F (2006) Tyrosine kinase
inhibitors of vascular endothelial growth factor receptors in clinical trials: Current
status and future directions. Oncologist 11:753–764.
44. Braconi C, et al. (2008) Insulin-like growth factor (IGF) 1 and 2 help to predict disease
outcome in GIST patients. Ann Oncol 10.1093/annonc/mdn040.
8392
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0803383105 Tarn et al.