Autophagy inhibition and antimalarials promote cell
death in gastrointestinal stromal tumor (GIST)
Anu Guptaa, Srirupa Royb, Alexander J. F. Lazarc,d, Wei-Lien Wangc,d, John C. McAuliffec,e,f, David Reynosoc,f,
James McMahong, Takahiro Taguchih, Giuseppe Florisi, Maria Debiec-Rychterj, Patrick Sch} offskii, Jonathan A. Trentc,f,
Jayanta Debnathb,1, and Brian P. Rubina,g,k,1
aDepartment of Molecular Genetics, Lerner Research Institute,gDepartment of Anatomic Pathology, andkTaussig Cancer Center, Cleveland Clinic, Cleveland,
OH 44195;bDepartment of Pathology and Diller Family Comprehensive Cancer Center, University of California, San Francisco, CA 94143;cSarcoma Research
Center and Departments ofdPathology andfSarcoma Medical Oncology, University of Texas M. D. Anderson Cancer Center, Houston, TX 77054;eMD/PhD
Program, University of Texas, Houston, TX 77054;hDivision of Human Health and Medical Science, Graduate School of Kuroshio Science, Kochi University,
Nankoku, Kochi 783-8505, Japan; and Departments ofiGeneral Medical Oncology andjHuman Genetics, Catholic University of Leuven, 3000 Leuven, Belgium
Edited* by Brian J. Druker, Oregon Health and Science University, Portland, OR, and approved June 22, 2010 (received for review January 8, 2010)
Although gastrointestinal stromal tumors (GISTs) harboring acti-
vating KIT or platelet-derived growth factor receptor A (PDGFRA)
mutations respond to treatment with targeted KIT/PDGFRA inhib-
Most often, a sizeable tumor cell subpopulation survives and re-
mains quiescent for years, eventually resulting in acquired resis-
tance and treatment failure. Here, we report that imatinib induces
autophagy as a survival pathway in quiescent GIST cells. Inhibiting
autophagy inhibition represents a potentially valuable strategy to
promote GIST cytotoxicity and to diminish both cellular quiescence
and acquired resistance in GIST patients.
tions resulting in ligand-independent, constitutive activation that
and an additional 7% have mutually exclusive platelet-derived
growth factor receptor A (PDGFRA) mutations (1). (1). As a re-
sult, imatinib mesylate, a small molecule tyrosine kinase inhibitor
that abrogates KIT and PDGFR activity, is highly effective as a
treatment for metastatic GIST (1). Before imatinib, recurrent or
metastatic GIST was uniformly fatal (2).
Although imatinib is quite effective in stabilizing disease pro-
gression in GIST, it is generally not curative. Less than 2% of pat-
ients experience complete radiographic regression (3). Even upon
prolonged imatinib treatment, most are left with a substantial and
stable tumor burden consisting of viable nonproliferating tumor
cells, indicating that significant numbers of GIST cells can survive
imatinib and remain quiescent. Moreover, imatinib withdrawal in
such individuals results in rapid disease progression, necessitating
lifelong imatinib therapy (4). The inability of imatinib to fully era-
dicate GIST cells also contributes to acquired imatinib resistance,
mostly due to intraallelic second-site KIT mutations that interfere
with imatinib binding (5). Hence, from a therapeutic standpoint, it
as single agents or in combination with imatinib.
As a result, we sought to dissect the contributions of macro-
autophagy (hereafter called autophagy), an evolutionarily con-
served lysosomal self-digestion process, to the survival of GIST
cells during imanitib-induced quiescence (6). Autophagy is a key
mechanism to recycle energy and nutrients during starvation or
stress (7). Although the precise role of autophagy in cell survival
versus death is highly context dependent (8, 9), growing evidence
indicates that autophagy can promote tumor cell survival in re-
sponse to both cytotoxic and targeted chemotherapies (10). Here,
we demonstrate that autophagy is robustly induced in GIST cells
astrointestinal stromal tumor (GIST) is the most common
mesenchymal neoplasm of the gastrointestinal tract (1). Im-
upon imatinib treatment; furthermore, inhibiting self-eating, ei-
ther alone or in combination with imatinib, promotes cell death.
Because GISTs are notable for their resistance to cell death, even
after targeted KIT/PDGFRA kinase inhibition, these results have
significant implications for GIST therapy.
Imatinib-Sensitive GIST Cells Exhibit Reversible Quiescence. To de-
velop a cell-based model for imatinib-induced quiescence, we
evaluatedthe invitro response toandrecoveryfromimatinibtreat-
ment in three GIST cell lines, all of which contain constitutively
whereas GIST-T1-R is imatinib resistant. Imatinib blocked KIT
(Tyr721) phosphorylation in GIST-T1 and GIST882 (Fig. 1A). Af-
of 0.1 μM and higher in both sensitive cell lines (Fig. 1B). Flow
cytometry for DNA content corroborated a 3- to 5-fold decrease in
S1); furthermore, a complementary increase in the G0/G1fractions
maintained high level KIT phosphorylation in the presence of
imatinib and failed to show a proliferative block or G0/G1arrest
proliferation upon drug withdrawal after prolonged exposure to
imatinib, quiescent GIST-T1 and GIST882 cells were capable of
resuming proliferation (Fig. S1) and reentering the cell cycle (Fig.
1C) within 24 h of imatinib removal. Consistent with cell-cycle ar-
rest, the cell-cycle inhibitor p27 accumulated during imatinib treat-
ment and dropped significantly after drug withdrawal (Fig. 1D).
In both GIST-T1 and GIST882 cells, imatinib also induced sig-
nals associated with diminished protein translation, including in-
creased eIF2α Ser51 phosphorylation, a stress-induced suppressor
ribosomal protein, a downstream reporter of mTOR activation
(12) (Fig. 1D and Fig. S1). Again, imatinib withdrawal after 6 d of
treatment resulted in rapid reversal of these phosphorylation pat-
terns, correlating with cell cycle reentry (Fig. 1D). Finally, a 2-fold
Author contributions: A.G., J.D., and B.P.R. designed research; A.G., S.R., A.J.F.L., W.-L.W.,
J.C.M., D.R., J.M., G.F., M.D.-R., P.S., and J.A.T. performed research; T.T. contributed new
reagents/analytic tools; A.G., J.D., and B.P.R. analyzed data; and J.D. and B.P.R. wrote the
Conflict of interest statement: B.P.R. is a member of the Novarits Speakers Bureau, is
a consultant for Novartis and has designed educational material for Novartis. A.J.F.L. is
a member of the Novartis Speakers Bureau. J.A.T. is a member of the Novartis Speakers
Bureau and participates in Novartis-sponsored research and clinical trials. P.S. has received
a commercial research grant from Novartis. M.D.-R. has received honoraria from Novartis.
*This Direct Submission article had a prearranged editor.
1To whom correspondence may be addressed. E-mail: firstname.lastname@example.org or rubinb2@
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| August 10, 2010
| vol. 107
| no. 32
reduction in glucose uptake was observed in sensitive GIST lines
upon imatinib treatment that, again, reversed upon removal (Fig.
1E). Collectively, these data demonstrate that imatinib induces
importantly, these in vitro results recapitulate key biological fea-
tures observed in imatinib-treated GIST patients in vivo (4).
Imatinib Is Ineffective at Killing GIST Cells. After imatinib therapy in
vivo, most GISTs remain stable in size and contain viable cells
(13). Thus, we assessed the effects of imatinib on apoptosis in-
duction in vitro. Although GIST-T1 exhibited minimal caspase 3/7
activity after 1 μM imatinib for 72 h, GIST882 showed a 3-fold
increase (Fig. 2A). Terminal deoxynucleotidyl transferase dUTP
nick end labeling (TUNEL) analysis confirmed the lack of apo-
ptosis in GIST-T1, whereas ≈25% of GIST882 cells exhibited
positive TUNEL staining after treatment (Fig. 2B and Fig. S2).
To extend these results, we determined clonogenic recovery after
in an ≈60% decrease in clonogenic recovery, consistent with in-
creased caspase 3/7 activation (Fig. 2C). Notably, although a pro-
portion of GIST882 cells died after imatinib exposure, numerous
cells survived. Collectively, these results indicate that, even in
relatively death-sensitive GIST lines, a significant residual pop-
cell death observed in these sensitive cells strikingly resembles the
behavior of GIST treated with imatinib in vivo (3, 4).
Imatinib Induces Autophagy in GIST Cells. BecauseGISTcellspersist
through imatinib treatment, we hypothesized that autophagy may
serve as a survival pathway in GIST cells treated with imatinib. To
we created cells stably expressing GFP fused to microtubule-
associated protein light chain 3 (GFP-LC3). Untreated GIST-T1
was notable for a small number of GFP-LC3 puncta, indicative of
baselineautophagy. Punctate GFP-LC3rapidly increasedin num-
ber within 8 h of imatinib treatment, and subsequently decreased
at 16 and 24 h (Fig. S3), consistent with the formation and lyso-
somal turnover of autophagosomes (15, 16). Electron microscopy
confirmed increased autophagic vesicles in GIST-T1 cells after
treatment with imatinib (Fig. S3).
followed by a decrease at later timepoints, we hypothesized that
autophagosomes were rapidly maturing into autolysosomes, a hall-
prediction, we stably expressed a tandem mCherry-GFP-tagged
LC3 chimera in GIST-T1, which enabled simultaneous quantifica-
tion of autophagosome induction and autolysosome maturation.
Whereas the GFP signal is sensitive to the acidic and proteolytic
conditions of the lysosome, mCherry is stable (14, 16). After 8 h of
imatinib treatment, we observed marked induction of double pos-
itive (GFP+mCherry+) LC3 puncta, a marker of early autophago-
somes, which then decreased at 16 h and 24 h (Fig. 3 A and B). In
contrast, mCherry single positive puncta were progressively ele-
vated at later timepoints, indicating autophagosome maturation
GIST-T1-R cell lines. (A) Cells were grown in the absence or presence of imatinib for the indicated times, lysed, and immunobloted with antibodies to KIT, pKIT
(Y721), and β-actin. (B) Cell proliferation was measured after treatment with the indicated concentrations of imatinib for 72 h. (C) Cell-cycle analysis by flow
cytometry after treatment with 1 μM imatinib for 6 d and then 2 d after removal of imatinib. (D) GIST-T1 were grown in the absence (control) or presence of 1
μM imatinib for the indicated times, lysed, and immunoblotted with antibodies to KIT, pKIT (Y721), pEIF2α (Ser51), eIF2α, pS6 (Ser235/236), S6, p27, and
β-actin. (E) 2-NBDG uptake measured in GIST-T1 after incubation in imatinib at the indicated doses for 24 h. During washout, cells were incubated in 1 μM
imatinib for 24 h, followed by withdrawal of imatinib for 24 h.
Imatinib induces reversible quiescence in GIST cells. Experiments were performed in imatinib-sensitive GIST-T1 and GIST882 and imatinib-resistant
| www.pnas.org/cgi/doi/10.1073/pnas.1000248107 Gupta et al.
During autophagy, the lipidation of LC3 results in a faster mi-
be associated with either an increase or decrease in LC3-II, de-
pending on the rate of autophagosome turnover (17). In GIST-T1
during imatinib treatment (Fig. 3C). However, LC3-II remained
stable in the presence of the lysosomal cathepsin inhibitors E64d
(see “E/P+” lanes). In contrast, GIST-T1R showed high baseline
LC3-II that did not change after addition of imatinib (Fig. 3D). In
in GIST-T1, which did not change markedly after addition of
imatinib (Fig. 3E). This finding may reflect reduced autophagy in
S2). Nevertheless, both GIST-T1 and GIST882 treated with ima-
ubiquitin-binding scaffold protein selectively degraded by auto-
phagy (Fig. S4) (18). In contrast, p62 levels were not affected in
GIST-T1-R. Overall, these results indicate that bona fide auto-
phagic degradation is induced in sensitive GIST cells upon imati-
Punctate LC3-II Is Observed in Human GISTs Treated with Imatinib.
We next assessed whether imatinib treatment induced autopha-
gosome formation(punctateLC3)in humanGISTinvivo by using
tissue samples from GIST patients who were randomized to treat-
(19). KIT and PDGFRA mutation status was known in each case,
and in vitro sensitivities of each mutant to imatinib (5, 20). Signif-
icant viable tumor cell populations were observed in all samples
after treatment, indicating that imatinib is minimally cytotoxic in
GIST tumor cells at this early timepoint. Using immunohisto-
chemistry to detect the punctate distribution of LC3 within tumor
cells (Fig. 3D), cases were scored as negative (0% cells positive),
the imatinib-sensitive group, increasing amounts of punctate LC3
were seen in the 7-d treatment group compared with patients
treated for shorter times (Fig. 3E); in contrast, only negative or
focal staining was observed in the imatinib-resistant group (Fig.
3F). Because autophagy induction inversely correlates with cell
death in numerous experimental models, we next assessed apo-
caspase-3 (CC-3) (8, 9). Cases were scored as low (<20 positive
cells/5 high power fields) or high (≥20 positive cells/5 high power
fields). Importantly, those imatinib-sensitive cases that exhibited
moderatestaining for LC3 had low amounts of CC-3, whereas 9 of
15 imatinib-sensitive GIST with low or no staining for autophagy
showed high CC-3. These in vivo results are consistent with our in
vitro results, demonstrating that imatinib can induce autophago-
some formation in GIST; moreover, increased autophagosome
decreased levels of apoptosis.
ATG Depletion Promotes the Death of Imatinib-Treated GIST Cells.
We next asked whether autophagy promotes GIST survival during
imatinib treatment. Using RNAi against ATG7 and ATG12, two
critical autophagy regulators, we tested whether autophagy in-
hibition increased GIST cell death in combination with imatinib
(21). ATG knockdown significantly reduced GFP-LC3 puncta in
immunoblotting (Fig. S5). Furthermore, ATG7 or ATG12 deple-
was induced, we measured caspase 3/7 activation in imatinib-
in caspase 3/7 activity was observed in ATG knockdown cells after
treatment with 1 μM imatinib but not in nontargeting controls.
Statistical analysis confirmed that ATG knockdown synergized
withimatinib tokill GIST cells(Fig.4 A andB).Collectively,these
data support the hypothesis that autophagy contributes to the
survival of imatinib-treated GIST cells.
Cotreatment with Imatinib and Lysosomotrophic Antimalarial Agents
Promotes Cell Death and Abrogates Imatinib Resistance in GIST.
Emerging evidence indicates that cancer cells undergoing auto-
phagy are highly sensitive to treatment with lysosomotrophic
agents, such as the antimalarial chloroquine, which inhibits lyso-
somal acidification and blocks the terminal stages of autophagic
proteolysis (22, 23). We confirmed that chloroquine elicited an
increase in LC3-II in GIST-T1 cells, as did another antimalarial
agent, quinacrine (Fig. S6). Quinacrine also inhibited both base-
line and starvation-induced LC3-II turnover in MCF10A mam-
mary cells (Fig. S6), demonstrating that quinacrine functions sim-
ilarly to other lysosomotrophic agents with regard to its ability to
To assess how these drugs impact GIST cell survival, we treated
and absence of imatinib. Indeed, a synergistic decrease in clono-
genic survival was observed when GIST-T1 cells were cotreated
with imatinib and either chloroquine or quinacrine (Fig. 4C). Im-
portantly, GIST-T1 cells also exhibited significantly reduced via-
bility when treated with either chloroquine or quinacrine as single
agents; indeed, quinacrine was highly cytotoxic compared with
imatinib alone. Quinacrine also produced the greatest enhance-
different GIST cell lines. (A) Cells were grown for 72 h with or without 1 μM
imatinib and assayed for caspase 3/7 activity. (B) Percentage of TUNEL-
positive cells for GIST-T1 and GIST882. (C) Clonogenic replating efficiency of
cells after treatment with or without 1 μM imatinib for 72 h. Results are the
mean ± SD from three independent experiments.
Imatinib induces variable amounts of apoptosis and cell death in
Gupta et al. PNAS
| August 10, 2010
| vol. 107
| no. 32
ment in apoptosis, eliciting a 2-fold increase in caspase 3/7 activity
(Fig. 4D). In addition, a synergistic effectwas observed when com-
biningtheseagentswithimatinib, resultingin a 3.5-fold increasein
caspase activity compared with imatinib alone (Fig. 4D).
To morethoroughly scrutinize whetherimatinib andautophagy
inhibitors acted synergisitcally, several concentrations of imatinib
(0–5 μM) were combined with several concentrations of either
chloroquine (0–200 μM) or quinacrine (0–10 μM), upon which
caspase 3/7 activity was assayed. Chou–Talalay analysis demon-
strated that imatinib synergized with chloroquine over the entire
range of drug concentrations (Fig. S7), whereas imatinib syner-
gized with quinacrine at intermediate doses (Fig. S8) (24, 25).
Furthermore, in a mouse GIST-T1 xenograft model, treatment
with imatinib in combination with either chloroquine or quina-
crine resulted in a statistically significant increase in apoptosis,
evidenced by increased CC-3 staining in tissue sections (Fig. 4E).
In addition, mice treated with the combination of imatinib and
quinacrine exhibited a statistically significant decrease in tumor
size in comparison with those treated with imatinib alone (Fig.
4F). The decrease in tumor size was especially remarkable be-
cause treatment lasted only 15 d. Overall, these results corrob-
orate that autophagy inhibition using antimalarials promotes
apoptosis and decreases GIST cell viability in vivo.
associated membrane glycoprotein-2 (LAMP2), which is required
for autophagosome–lysosome fusion (9). LAMP2 knockdown pro-
duced an increase in GFP-LC3 puncta and LC3-II and promoted
a synergistic increase in the death of imatinib-treated GIST cells as
Finally, we tested the ability of chloroquine and quinacrine to
attenuate the long-term outgrowth of imatinib-treated GIST cells
in vitro. GIST-T1 cells subject to imatinib-induced quiescence
exhibit robust cell cycle arrest and metabolic suppression in vitro,
consistent with stable disease; nonetheless, upon culture in ima-
tinib for extended periods, these cells do exhibit low-level pro-
liferation (Fig. 5). However, this outgrowth was synergistically
reduced when GIST-T1 cells were cultured in the presence of
imatinib combined with either quinacrine or chloroquine (Fig. 5).
single agent efficacy in preventing the outgrowth of GIST cells;
notably, quinacrine and imatinib demonstrated equivalent results
when used as solitary treatments. Overall, these results suggest
that chloroquine and quinacrine, two well-tolerated, inexpensive
(A) Representative images of GFP and mCherry fluores-
cent puncta in GIST-T1 cells expressing mCherry-GFP-LC3
grown in complete media with 1 μM imatinib for
the indicated times. (B) Quantification of GFP/mCherry
double-positive and mCherry single-positive puncta per
cell in control or cells treated with 1 μM imatinib for the
indicated times. GIST-T1 (C), GIST-T1-R (D), and GIST882
(E) were treated with 1 μM imatinib for the indicated
times, lysed, and immunoblotted with antibodies to LC3
a representative spindle cell gastric GIST from a patient
treated with neoadjuvant imatinib for 7 d. Note the
punctate staining pattern (blowup) consistent with the
induction of autophagosomes. (G) Table summarizing
results from imatinib-sensitive GISTs. (H) Table summa-
rizing results from imatinib-resistant GISTs. (I) Table
Imatinib induces autophagy in vitro and in vivo.
| www.pnas.org/cgi/doi/10.1073/pnas.1000248107Gupta et al.
drugs with a long history of use in humans, may be useful to at-
tenuate the expansion of both imatinib-sensitive and -resistant
Imatinib has revolutionized the treatment of GIST, resulting in
disease stabilization in ≈75% of patients (1); nonetheless, ima-
tinib is associated with two major clinical problems that must be
addressed to improve the long-term outcome. First, most, if not
all, patients with metastatic disease remain stable or progress;
total tumor regression is rare (3). Thus, lifelong therapy with im-
atinib is standard of care. Indeed, patients with inoperable GIST
have remained stable on imatinib for >8 y (3). Moreover, even
after extended therapy, GISTs progress rapidly after withdrawal
of imatinib (4). The second major problem is the emergence of
acquired resistance during therapy, typically the result of second-
site intraallelic KIT mutations that inhibit imatinib binding and
reactivate oncogenic KIT signaling (5). Because the cell culture
models described here recapitulate the key biological features
observed in imatinib-treated GIST in vivo, they create the mech-
anistic platform needed to address these therapeutic barriers.
Growing evidence indicates that autophagy contributes to che-
motherapeutic resistance (10). Here, we demonstrate that imati-
nib induces autophagy in imatinib-sensitive GIST in vitro and in
vivo. Imatinib has been shown to induce autophagy in a variety of
nonneoplastic and neoplastic cell lines (22, 26, 27). Although
imatinib may stimulate autophagy by inhibiting unidentified pro-
GIST cell line that has acquired an imatinib-resistant KIT muta-
tion, demonstrate decreased autophagy compared with imatinib-
kinase inhibition in GIST.
We also assessed the effects of autophagy inhibition on GIST
cell fate and observed that ATG depletion enhanced GIST cell
death when combined with imatinib. Recently, the combination
of autophagy inhibition and p210BCR/ABLinhibition by imatinib
resulted in near complete eradication of CML stem cells, which
were otherwise resistant to imatinib alone (22). Little is known
about the progenitors that give rise to GIST (28). Whether these
precursor cells are susceptible to autophagy inhibitors, either
alone or in combination with imatinib, is an important topic for
phagy in response to targeted therapies exhibit sensitivity to lyso-
death in imatinib-treated GIST cells. (A) Clonogenic replating efficiency of
cells after 72 h of treatment with or without 1 μM imatinib in ATG7- or
ATG12-depleted cells. (B) Caspase 3/7 activity was measured in ATG7- or
ATG12-depleted cell lines with administration of 1 μM imatinib for 72 h. (C)
Clonogenic replating efficiency of GIST-T1 cells after 48 h of treatment with
1 μM imatinib, 50 μM chloroquine, or 3 μM quinacrine and each autophagy
inhibitor in combination with imatinib. (D) Caspase 3/7 activity was mea-
sured after treatment with imatinib, chloroquine, or quinacrine and each
autophagy inhibitor in combination with imatinib. Results are the mean ±
SD from three independent experiments. The percent increase in caspase 3/7
activity in cells treated with imatinib relative to untreated controls is shown.
NT, nontargeting siRNA pool. P values indicated with an asterisk are based
on statistical analysis of synergy between imatinib and inhibition of
autophagy (see details in SI Methods). (E) CC-3 positive cells ± SEM from
mouse GIST-T1 xenografts after treatment with imatinib alone or in com-
bination with either chloroquine or quinacrine. (F) Tumor weight ± SEM of
mouse GIST-T1 xenografts after treatment with imatinib alone or in com-
bination with either chloroquine or quinacrine. Note that untreated control
tumors underestimate tumor size (designated by *) because it is necessary to
euthanize all untreated control mice because of excessive tumor growth
before the end of the experiment.
ATG depletion and antimalarial lysosomal inhibitors promote cell
cells in vitro. (A) Representative images from GIST-T1 cells plated at 100
cells per well and treated for 14 d with 0.1 μM imatinib, 5 μM chloroquine, or
0.5 μM quinacrine either as single agents or in the designated combinations.
(B) Quantification of cells per colony under the various conditions; the data
are presented as mean ± SEM. P values indicated with an asterisk are based
on statistical analysis of synergy between imatinib and inhibition of auto-
phagy (see details in SI Methods).
Antimalarials attenuate the outgrowth of imatinib-resistant GIST
Gupta et al. PNAS
| August 10, 2010
| vol. 107
| no. 32
motrophic agents (22, 23). Accordingly, we now demonstrate that Download full-text
lysosomal inhibition sensitizes GIST to die when combined with
imatinib both in vitro and in vivo. Furthermore, the outgrowth of
GIST cells observed in the presence of imatinib alone is signifi-
with imatinib. Remarkably, the inability of imatinib to kill GIST
cells may facilitate the proliferation of resistant cells harboring
second-site KIT mutations; hence, combination therapy with ei-
that develops during extended imatinib treatment.
In comparison with chloroquine, our results indicate that quin-
acrineexhibitsmore robustcytotoxicity against GISTcells, both as
a single agent and in combination with imatinib. Like chloroquine,
quinacrine has a long history as an antimalarial, but to our knowl-
yeast vacuoles, the equivalent of the mammalian lysosome (29).
other established lysosomal inhibitors in terms of its ability to in-
hibit autophagic flux. Based on these results, we hypothesize that
the cytotoxic effects of quinacrine in GIST cells are due to auto-
phagy and lysosomal inhibition. Still, we recognize that the
cytotoxic effects of quinacrine and chloroquine may involve other
processes (30, 31). Because both agents are in clinical trials for
cancer, further studies are needed to dissect the precise con-
tributions of autophagy inhibition to their antineoplastic effects.
Overall, we demonstrate that a significant proportion of GIST
cells survive imatinib therapy by entering a state of reversible
quiescence and activating an autophagy-dependent survival
mechanism. Autophagy inhibition, using either ATG knockdown
or pharmacological lysosomal inhibition with antimalarials,
potentiates imatinib cytotoxicity against GIST cells, both in vitro
and in vivo. Based on these results, we predict that combination
inexpensive antimalarial agents with well-known toxicity profiles,
has the potential to greatly improve clinical outcome in
Autophagy Analysis. Multiple complementary assays were used to measure
autophagosome formation and autophagic flux, including: quantification of
punctate LC3 using immunohistochemistry (in tissues) or fluorescence mi-
croscopy (in GIST cells stablyexpressing LC3fluorescentproteinchimeras), the
formation and lysosomal turnover of lipidated LC3 (LC3-II), and degradation
of the autophagy substrate, p62. All techniques were performed in accor-
dance with established guidelines (14). Details are found in SI Methods.
Cell Colony Outgrowth Assay. Cells were plated at 100 cells per well in six-well
tissueculture plates invariousconditions.Colonies weregrown out for14din
the presence of 0.1 μM imatinib, 5 μM chloroquine, or 0.5 μM quinacrine
alone or in combination with imatinib, fixed with methanol, and stained
with 0.2% crystal violet. The number of cells per colony was enumerated in
50 randomly oriented colonies. Each experiment was repeated three times.
Details for cell culture, antibodies and chemicals, generation of stable
lines, cell-cycle analysis, cell proliferation and apoptosis analyses, glucose
uptake analysis, analysis of punctate GFP-LC3 and GFP-mCherry-LC3, analysis
of LC3 by immunohistochemistry, clonogenic replating assay, RNA in-
terference, immunoblot analysis, electron microscopy, treatment of GIST-T1
xenografts, human subjects, and statistics are given in SI Methods.
ACKNOWLEDGMENTS. We thank James Bena, PhD (Department of Quan-
titative Health Sciences, Cleveland Clinic), for assistance with statistical
evaluations and Ms. Paula Carver for excellent technical assistance. GIST
882 was a gift of Dr. Jonathan Fletcher (Brigham and Women’s Hospital,
Boston). A grant from the Life Raft Group (to B.P.R. and M.D.-R.), National
Cancer Institute Grant R01 CA126792 (to J.D.), a Culpeper Medical Scholar
Award (Partnership For Cures) (to J.D.), an AACR-Genentech BioOncology
Award (to J.D.), and a Howard Hughes Medical Institute Physician-Scientist
Early Career Award (to J.D.) supported this work.
1. Rubin BP, Heinrich MC, Corless CL (2007) Gastrointestinal stromal tumour. Lancet 369:
2. Dematteo RP, Heinrich MC, El-Rifai WM, Demetri G (2002) Clinical management of
gastrointestinal stromal tumors: Before and after STI-571. Hum Pathol 33:466–477.
3. Blanke CD, et al. (2008) Long-term results from a randomized phase II trial of
standard- versus higher-dose imatinib mesylate for patients with unresectable or
metastatic gastrointestinal stromal tumors expressing KIT. J Clin Oncol 26:620–625.
4. Blay JY, et al. (2007) Prospective multicentric randomized phase III study of imatinib in
patients with advanced gastrointestinal stromal tumors comparing interruption
versus continuation of treatment beyond 1 year: The French Sarcoma Group. J Clin
5. Heinrich MC, et al. (2006) Molecular correlates of imatinib resistance in gastrointestinal
stromal tumors. J Clin Oncol 24:4764–4774.
6. Levine B, Kroemer G (2008) Autophagy in the pathogenesis of disease. Cell 132:27–42.
7. Klionsky DJ (2007) Autophagy: from phenomenology to molecular understanding in
less than a decade. Nat Rev Mol Cell Biol 8:931–937.
8. Debnath J, Baehrecke EH, Kroemer G (2005) Does autophagy contribute to cell death?
9. Thorburn A (2008) Apoptosis and autophagy: Regulatory connections between two
supposedly different processes. Apoptosis 13:1–9.
10. Kondo Y, Kanzawa T, Sawaya R, Kondo S (2005) The role of autophagy in cancer
development and response to therapy. Nat Rev Cancer 5:726–734.
11. Ron D, Walter P (2007) Signal integration in the endoplasmic reticulum unfolded
protein response. Nat Rev Mol Cell Biol 8:519–529.
12. Guertin DA, Sabatini DM (2007) Defining the role of mTOR in cancer. Cancer Cell 12:
13. Agaram NP, et al. (2007) Pathologic and molecular heterogeneity in imatinib-stable
or imatinib-responsive gastrointestinal stromal tumors. Clin Cancer Res 13:170–181.
14. Klionsky DJ, et al. (2008) Guidelines for the use and interpretation of assays for
monitoring autophagy in higher eukaryotes. Autophagy 4:151–175.
15. Kabeya Y, et al. (2000) LC3, a mammalian homologue of yeast Apg8p, is localized in
autophagosome membranes after processing. EMBO J 19:5720–5728.
16. Kimura S, Noda T, Yoshimori T (2007) Dissection of the autophagosome maturation
process by a novel reporter protein, tandem fluorescent-tagged LC3. Autophagy 3:
17. Tanida I, Minematsu-Ikeguchi N, Ueno T, Kominami E (2005) Lysosomal turnover, but
not a cellular level, of endogenous LC3 is a marker for autophagy. Autophagy 1:
18. Bjørkøy G, et al. (2005) p62/SQSTM1 forms protein aggregates degraded by autophagy
and has a protective effect on huntingtin-induced cell death. J Cell Biol 171:603–614.
19. McAuliffe JC, et al. (2009) A randomized, phase II study of preoperative plus posto-
perative imatinib in GIST: Evidence of rapid radiographic response and temporal in-
duction of tumor cell apoptosis. Ann Surg Oncol 16:910–919.
20. Heinrich MC, et al. (2008) Correlation of kinase genotype and clinical outcome in
the North American Intergroup Phase III Trial of imatinib mesylate for treatment of
advanced gastrointestinal stromal tumor: CALGB 150105 Study by Cancer and
Leukemia Group B and Southwest Oncology Group. J Clin Oncol 26:5360–5367.
21. Maiuri MC, Zalckvar E, Kimchi A, Kroemer G (2007) Self-eating and self-killing: crosstalk
between autophagy and apoptosis. Nat Rev Mol Cell Biol 8:741–752.
22. Bellodi C, et al. (2009) Targeting autophagy potentiates tyrosine kinase inhibitor-
induced cell death in Philadelphia chromosome-positive cells, including primary CML
stem cells. J Clin Invest 119:1109–1123.
23. Degtyarev M, et al. (2008) Akt inhibition promotes autophagy and sensitizes PTEN-
null tumors to lysosomotropic agents. J Cell Biol 183:101–116.
24. Chou T (1991) The median-effect principle and the combination index for quantitation of
synergism and antagonism. Synergism and antagonism in chemotherapy, eds Chou TR,
Rideout DC (Academic, San Diego), pp 61–102.
25. Chou TC, Talalay P (1984) Quantitative analysis of dose-effect relationships: the
combined effects of multiple drugs or enzyme inhibitors. Adv Enzyme Regul 22:
26. Ertmer A, et al. (2007) The anticancer drug imatinib induces cellular autophagy.
27. Shingu T, et al. (2009) Inhibition of autophagy at a late stage enhances imatinib-
induced cytotoxicity in human malignant glioma cells. Int J Cancer 124:1060–1071.
28. Lorincz A, et al. (2008) Progenitors of interstitial cells of cajal in the postnatal murine
stomach. Gastroenterology 134:1083–1093.
29. Roberts CJ, Raymond CK, Yamashiro CT, Stevens TH (1991) Methods for studying the
yeast vacuole. Methods Enzymol 194:644–661.
30. Guo C, et al. (2009) 9-Aminoacridine-based anticancer drugs target the PI3K/AKT/
mTOR, NF-kappaB and p53 pathways. Oncogene 28:1151–1161.
31. Gurova KV, et al. (2005) Small molecules that reactivate p53 in renal cell carcinoma
reveal a NF-kappaB-dependent mechanism of p53 suppression in tumors. Proc Natl
Acad Sci USA 102:17448–17453.
| www.pnas.org/cgi/doi/10.1073/pnas.1000248107Gupta et al.