Response Markers and the Molecular Mechanisms of Actionof Gleevec in Gastrointestinal Stromal Tumors1

Andrey Frolov, Santiago Chahwan, Michael Ochs,Juan Pablo Arnoletti, Zhong-Zong Pan,Olga Favorova, Jonathan Fletcher,Margaret von Mehren, Burton Eisenberg, andAndrew K. Godwin2

Departments of Medical Oncology [A. F., Z-Z. P., M. v. M., A. K. G.] andSurgical Oncology [S. C., J. P. A., B. E.], and Bioinformatics,Department of Information Science and Technology [M. O.], Fox ChaseCancer Center, Philadelphia, Pennsylvania 19111; Dana-Farber CancerInstitute, Boston, Massachusetts 02115 [J. F.]; and Graduate Programin the Department of Molecular Biology and Biotechnology, RussianState Medical University, 117869 Moscow, Russia [A. F., O. F.]

AbstractGastrointestinal stromal tumors (GISTs), defined by thepresence of constitutively activated KIT, are the mostcommon gastrointestinal mesenchymal malignancies.This observation has been successfully exploited inclinical trials of Gleevec (also known as imatinibmesylate, STI-571) for patients with unresectable and/or metastatic GISTs. The biological mechanisms ofGleevec as well as its downstream molecular effectsare generally unknown. We used a DNA microarray-based approach to identify gene expression patternsand signaling pathways that were altered in responseto Gleevec in GIST cells. We identified a total of 148genes or expressed sequence tags (of 10,367) thatwere differentially regulated; 7 known genes displayeda durable response after treatment. The significantlydown-regulated genes were SPRY4A, FZD8, PDE2A,RTP801, FLJ20898, and ARHGEF2. The only up-regulated gene was MAFbx. On a functional level, wedemonstrated that imatinib inhibited phosphorylation ofKIT, AKT, and extracellular signal-regulated kinase 1/2without affecting the total level of these proteins andthat differential expression of these response genesinvolved activation of mitogen-activated protein kinase-dependent and -independent pathways. In an attemptto correlate these in vitro findings to clinical data, weexamined GIST needle biopsy specimens taken frompatients before and after Gleevec administration

according to the CSTI571-B2222 Phase II trial anddemonstrated that expression levels of the two genetranscripts evaluated correlated well with clinicalresponse. This study emphasizes the potential value ofan in vitro cell model to investigate GIST response toimatinib in vivo, for the purpose of identifyingimportant genetic markers of clinical response,mechanisms of drug action, and possible therapeutictargets.

IntroductionGISTs3 are the most common mesenchymal malignancies ofthe gastrointestinal tract. GIST neoplastic cells appear tobelong to the same lineage as the interstitial cells of Cajal,which are the pacemaker cells of the gut (1). GISTs sharemany immunohistochemical, morphological, and ultrastruc-tural features with interstitial cells of Cajal, supporting thishypothesis (2). GISTs are characterized by the presence ofconstitutively activated KIT (CD117), the RTK encoded by thec-KIT proto-oncogene, also known as stem cell factor re-ceptor (3, 4). c-KIT is a member of the RTK subclass III familyand has structural homology to the receptors for FLT3, plate-let-derived growth factor, and macrophage colony-stimulat-ing factor (5). The proposed mechanism of constitutive KITphosphorylation in the majority of GISTs is a gain of functionmutation most commonly involving exon 11, which encodesa portion of the cytoplasmic juxtamembrane domain. Othermutations involve exon 9 (extracellular domain), exon 13 (firstpart of the split tyrosine kinase domain), or exon 17 (phos-photransferase domain; Ref. 6). In addition, a recent studyhas found activating mutations in the PDGFRA oncoproteinin GISTs lacking c-KIT mutations, indicating an alternativeoncogenic mechanism in GISTs (7).

GISTs are usually diagnosed in middle-aged to olderadults and occur anywhere in the gastrointestinal tract, mostcommonly in the stomach (8). Malignant GISTs often presentwith synchronous metastatic disease, and even those ame-nable to complete surgical resection recur locally, with a5-year actuarial survival rate of approximately 50% (9). Pa-tients with metastatic disease uniformly have a poor prog-nosis with a median survival of 6–8 months because GISTshave been historically resistant to conventional and investi-gational therapy.

Imatinib (imatinib mesylate, Gleevec; Novartis, Basel,Switzerland), formerly known as STI-571, is an oral 2-phenylaminopyrimidine derivative. In preclinical studies, ima-

Received 4/4/03; revised 5/30/03; accepted 6/4/03.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indi-cate this fact.1 Supported in part by the Eileen-Stein Jacoby Fund, GIST foundation (toM. v. M.), a grant from the NIH (Ovarian Cancer SPORE P50 CA83638), asupplement to 3 U10 CA21661-27 (to B. E.), Tobacco Formula Grant fromthe Pennsylvania Department of Health (to A. K. G.), and by an appropri-ation from the Commonwealth of Pennsylvania. Z-Z. P. was supported bya fellowship from the Department of Defense (Institutional Breast CancerTraining Program Grant DAMD17-00-1-0249).2 To whom requests for reprints should be addressed, at Fox ChaseCancer Center, 7701 Burholme Avenue, Philadelphia, PA 19111. Phone:(215) 728-2205; Fax: (215) 728-2741; E-mail: A_Godwin@fccc.edu.

3 The abbreviations used are: GIST, gastrointestinal stromal tumor; EST,expressed sequence tag; ERK, extracellular signal-regulated kinase;MAPK, mitogen-activated protein kinase; RTK, receptor tyrosine kinase;MEK, mitogen-activated protein/ERK kinase; PI3K, phosphatidylinositol3�-kinase; RT-PCR, reverse transcription-PCR; PDE, phosphodiesterase;PDGFRA, platelet-derived growth factor receptor alpha.

699Vol. 2, 699–709, August 2003 Molecular Cancer Therapeutics

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

tinib has been shown to inhibit the activity of c-ABL, BCR-ABL, and PDGFRA/B (10, 11). Several Phase I/II clinical trialshave demonstrated the efficacy of imatinib in the treatmentof chronic myelogenous leukemia patients with the BCR/ABLtranslocation (12–14). This drug has also been shown to be aspecific in vitro inhibitor of c-KIT phosphorylation in severaltumor cell lines (10, 15). A recent report indicated that ima-tinib rapidly inhibits c-KIT phosphorylation and tumor cellproliferation while inducing apoptosis in a human GIST cellline (16). The c-KIT target has been exploited in two com-pleted Phase I/II clinical trials of GISTs, with over 50% ofpatients with unresectable or metastatic GISTs demonstrat-ing a partial response, and only 13% of patients manifestingdisease progression while receiving imatinib (17, 18).

Despite its clinical success, there are limited preclinicaland clinical data on the molecular targets of imatinib down-stream from c-KIT. Therefore, our strategy was to charac-terize genetic markers of response to imatinib with a geneexpression profiling approach using an in vitro model ofGIST. Results of these studies were then used to furtherelucidate the molecular mechanisms of imatinib and to as-sess the panel of genetic markers in a limited number ofsequential biopsy specimens from GIST patients before andafter Gleevec treatment.

Materials and MethodsGIST Cell Culture. GIST882 is an immortal GIST cell linethat possesses a homozygous exon 13 missense mutation(i.e., K642E) in c-KIT (16). GIST882 cells were grown asreported previously (16).

Treatment of GIST Cells with Imatinib: Proliferation andApoptosis Assessment. For growth analysis, GIST882cells were seeded at 6.5 � 105 cells/60-mm dish. Imatinib(provided by Novartis Oncology and dissolved in water) wasadded directly to the media to achieve a final concentrationof 1 or 10 �M. Cells were refed with conditioned mediacontaining imatinib or with conditioned media without drugevery 12 h. Cells were harvested and stained for the cellnumber, cell viability, and induction of apoptosis usingGuava ViaCount and Guava Nexin reagents (Guava Technol-ogy Inc., Burlingame, CA). The cells were counted using aGuava Personal Cytometer, and the data were analyzedusing the Guava CytoSoft software package.

Cell Cycle Analysis. Cells were trypsinized, centrifuged,and fixed in 70% ethanol at 4°C. Cell pellets were resus-pended in 50 �g/ml propidium iodide in PBS for 30 min at4°C. The stained cells were analyzed by flow cytometryperformed on a FACScan, and the data were analyzed withCell Quest software (Becton Dickinson).

Microarray Preparation. A total of 10,367 human cDNAfragments corresponding to known genes and ESTs from theintegrated molecular analysis of genomes and their expres-sion (I.M.A.G.E.) consortium library (Research Genetics) wereused to produce high-density cDNA microarrays in a mannerdescribed previously (19).

RNA Isolation and Microarray Hybridization. Imatinib(10 �M)-treated and untreated GIST882 cells were harvestedsimultaneously at 60–70% confluence. Total RNA was iso-lated using the guanidinium/isothiocyanate/phenol/chloro-

form method as described previously (20). Total RNA wasDNase treated using “DNA free” kit (Ambion, Austin, TX)according to the manufacturer’s specifications. Fifteen �g ofthis DNase-treated RNA were reverse transcribed, and ami-no-allyl-dUTP was incorporated in a reaction containing 500ng of oligo(dT) primers; 1� first strand buffer (Invitrogen,Carlsbad, CA); 0.01 M DTT; 500 �M each of dATP, dCTP,dGTP, and dTTP/amino-allyl-dUTP (2:3 ratio); 40 units ofrRNasin (Promega, Madison, WI); and 200 units of Super-Script II reverse transcriptase (Invitrogen). After brief dena-turation and annealing of the primers at 70°C for 8 min, thereaction was incubated at 42°C for 2 h, followed by alkalihydrolysis of RNA and cDNA purification using Microcon-30columns (Millipore, Bedford, MA) according to the manufac-turer’s instructions. The cDNA was then labeled with eitherCy3 or Cy5 dye by a coupling reaction using FluoroLinkmonofunctional dyes (Amersham Pharmacia Biotech, Pisca-taway, NJ) according to the manufacturer’s specifications.Probes were then purified using StrataPrep PCR PurificationKit (Stratagene, La Jolla, CA). Two of the samples (one la-beled with Cy3 and one labeled with Cy5) were combined,denatured, and preannealed in the presence of 10 �g ofCot-1 DNA (Invitrogen) and 10 �g of poly(dA) DNA. Hybrid-ization and washings were performed as described previ-ously (19). The above was repeated for each sample, exceptthat the dyes used to label the RNAs were reversed. Further-more, for each time point, the hybridizations were repeatedusing an independent source of RNA.

Intensity Extraction and Data Analysis. Intensity extrac-tion and spot quality characteristics were assessed as de-scribed previously (19). The data were then analyzed usingGeneSight 3 software (BioDiscovery, Inc., Marina Del Rey,CA) and with in-house software. The data were (a) correctedfor background by subtraction of the local group median, (b)normalized using a piecewise linear normalization with 5 binsand typically �1000 data points/bin, and (c) limited to havea minimum expression level equal to an estimate of theminimum background. The data were then converted to log2

values, and the mean and SD were determined for eachintensity ratio by combining “dye-flip” replicates. At least twodye-flip experiments were performed for each time point.The mean and the coefficient of variance were calculated forthese values and used for statistical analysis and clustering.Differential expression of individual genes was determinedby confidence analysis (21) and maximum likelihood analysis(22) to obtain a final list of candidates at a �95% confidencelevel. The data were also clustered into 11 clusters usinghierarchical clustering (23) with a Euclidean distance metric.

cDNA Microarray Clone Verification and SequenceAnalysis. cDNA fragments of interest were obtained fromoriginal stock plates used for microarray fabrication by theplasmid isolation from the bacteria. Clones were rese-quenced, and the correct annotation and homology wereidentified using the BLAST (National Center for Biotechnol-ogy Information) against the GenBank/EMBL database. DNAand amino acid sequence analyses were performed with theGCG/Wisconsin Package version 9.1 (Genetics ComputerGroup).

700 Markers of Response to Gleevec

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Northern Blotting of SPRY4A. Fifteen �g of total RNAwere used for standard Northern blot analysis, as describedpreviously (20). Blots were hybridized with 32P-labeled cDNAprobes corresponding to each of the genes.

Reverse Transcription of RNA. Two �g of DNase-treated RNA were reverse transcribed using Moloney murineleukemia virus reverse transcriptase (New England Biolabs,Beverly, MA) in a 20-�l reaction containing 1� reverse tran-scription buffer (NEB), 0.5 mM each dNTP, 4 �M oli-go(dT)16–18 primer, 10 units of RNase inhibitor (Promega),and 200 units of reverse transcriptase. Primers were prean-nealed for 10 min at 70°C, and the reaction was incubated for1 h at 42°C, followed by enzyme inactivation for 10 minat 90°C.

PCR. DNA fragments of SPRY4A, MAFbx, FZD8, PDE2A,RTP801, FLJ20898, ARHGEF2, and �-actin were amplifiedby PCR using cDNAs obtained by reverse transcription ofmRNA from GIST cells and tumor biopsies as a template andthe following sets of primers: Sprouty 4A (SPRY4A), 5�-CCGTTCTGTGGAGAGTCGATTTAC-3� and 5�-GTCCCTCA-GTGGCTCTCGACT-3�; Frizzled 8 (FZD8), 5�-ACAGTGTT-GATTGCTATTAGCATG-3� and 5�-GTGAAATCTGTGTATCT-GACTGC-3�; ARHGEF2, 5�-AAGGACGGAGAAAGGGAG-AA-3� and 5�-CAAGACAGCAGTGACCCTGA-3�; PDE2A, 5�-CCGCGATCTTTCTCGTAGTC-3� and 5�-CCCACTTCTGC-TACCTGCTC-3�; MAFbx, 5�-GTCCTGGGGTGAAAGTGAAA-3�and 5�-TCACAGCTCACATCCCTGAG-3�; FLJ20898, 5�-CCC-GAGTGACTCTGTTTTCC-3� and 5�-ACACCCAGTTGGAGGT-GAAG-3�; RTP801, 5�-AGACACGGCTTACCTGGATG-3� and5�-TTGATGACTCGGAAGCCAGT-3�; and �-actin, 5�-CTCAC-CATGGATGATGATATCGC-3� and 5�- CATGATGGAGTTGAA-GGTAGTTTCGT-3�. PCR was performed in a reaction volumeof 30 �l containing cDNA from 1 �l of the reverse transcriptionreaction described above as a template, 10 mM Tris-HCl (pH8.3), 50 mM KCl, 1.5 mM MgCl2, 0.001% gelatin, 0.5 �M of boththe forward and reverse primer, 60 �M of each deoxyribonu-cleotide, 5% DMSO, and 0.5 unit of platinum Taq DNA polym-erase (Invitrogen). After an initial denaturation step at 95°C for 5min, DNA was amplified through 19, 21, 22, 23, 24, 25, and 27cycles consisting of 5-s denaturing at 95°C, 60-s annealing at55°C, and 90-s extension at 72°C. The products were resolvedon a 1.5% agarose gel and visualized by UV light after stainingwith ethidium bromide. Images of the gels were obtained usingan Alpha Imager 2200 documentation and analysis system(Alpha Innotech, San Leandro, CA). Images were capturedwithin linear dynamic range and controlled for white color sat-uration. DNA bands were quantified using the Alpha Imagerversion 5.5 software package and the Fuji Image Gauge version3.11 software package (Fuji Photo Film Co., Ltd.).

Treatment of GIST Cells with a MEK or a PI3K Inhibitor.The MEK1/2 inhibitor U0126 (Promega) was dissolved inDMSO. GIST cells were cultured to 60–70% confluence.MEK inhibitor was added directly to the media to achieve afinal concentration of 1, 10, and 30 �M. An equal amount ofDMSO was added to control untreated cells as a vehiclecontrol. The PI3K inhibitor LY294002 (CalBiochem, La Jolla,CA) was dissolved in DMSO and used at a concentration of30 �M. Cells were treated for 1, 3, 6, and 24 h before RNAisolation.

Cell Lysate Preparation, SDS-PAGE, and Western BlotAnalysis. Anti-�-actin monoclonal antibodies (Sigma, St.Louis, MO) were used at a dilution of 1:5000 in 5% dried milk.Anti-phospho-c-KIT (Tyr719) polyclonal antibodies (Cell Sig-naling, Beverly, MA) were used at a dilution of 1:500 in 5%BSA. Anti-c-KIT monoclonal antibodies (Santa Cruz Biotech-nology, Inc., Santa Cruz, CA) were used at a dilution of 1:100in 5% dried milk. Anti-phospho-ERK1/2 monoclonal antibod-ies (Cell Signaling), anti-ERK1 and anti-ERK2 polyclonalantibodies (Santa Cruz Biotechnology, Inc.), anti-AKT,anti-AKT/phospho-Thr308, and anti-AKT/phospho-Ser473

polyclonal antibodies (Cell Signaling) were each used at adilution of 1:1000 dilution in 5% BSA. Cell lysate preparationand Western blot analysis were performed as discribed inRef. 24.

Clinical Samples. With an institutional review board-approved study for human subjects and tissue collection andwith informed consent, tumor specimens were obtained fromseven patients enrolled in the CSTI571-B2222 clinical trialsponsored by Novartis Oncology. Eligibility criteria includedhistological confirmation of GIST with documentation of c-KIT expression, as well as evidence of unresectable recurrentand/or metastatic disease. Patients were randomized to re-ceive 400 or 600 mg p.o. once daily dose of imatinib (17).Ultrasound-guided 14G core biopsies of nonnecrotic tumorwere obtained before initiation of therapy as well as while ondrug or shortly after drug was discontinued. The biopsieswere typically obtained from patients after 7–31 days oftreatment with imatinib (Table 1). Standard histological eval-uation was used to assure adequacy of nonnecrotic tumortissue in these specimens. Specimens were flash frozen andkept in liquid N2. Patients were monitored for response toimatinib mesylate at 4–12-week intervals by one of the au-thors (M. v. M.). Patients who were subsequently found tohave disease recurrence underwent another biopsy or sur-gical resection if clinically appropriate.

Patients 1, 2, and 3 had GIST of either the stomach or thesmall bowel and had a partial response to imatinib treatmentdefined per standard criteria. Patient 7 had two measurableabdominal tumors, one of which responded to the therapy,whereas the other progressed. This patient, however, did nothave a pretreatment biopsy. Patient 4 had a partial responseto imatinib initially but developed disease progression 3months after initiation of the treatment. The patient wastaken off drug, and a core tumor biopsy was obtained. Pa-

Table 1 Duration of imatinib treatment and tumor response

Patient no. Duration on drug Tumor response

1 15 days �a

2 20 days �3 21 days �4 7 days �

13 months �5 60 days �6 31 days �7 18 months �

�

a �, responded to treatment; �, failed to respond to treatment.

701Molecular Cancer Therapeutics

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

tients 5 and 6, with large intestine and gastric GIST, respec-tively, had no response to imatinib treatment; one of them(patient 5) remained with stable disease, and the other one(patient 6) developed disease progression. Additional tumorwas obtained from patient 6 during palliative surgical resec-tion �31 days after drug therapy. One to 2 �g of total RNAis typically obtained from a single biopsy and was used toevaluate SPRY4 and MAFbx expression by RT-PCR analysis.

ResultsEffect of Imatinib Treatment on Growth and Proliferationof GIST Cells. We first assessed the response of GIST882cells to imatinib in vitro. We found that both 1 and 10 �M

imatinib dramatically inhibited cell growth within 24–48 hafter exposure (Fig. 1). However, GIST cells treated withimatinib show no significant increase (�3% of total cell num-ber) in either early or late apoptosis during 96-h treatmentperiod as determined by annexin-V and 7 amino-actinomycin

D staining. The fraction of viable cells was 93 � 5% for bothtreated and untreated cells during entire course of treatment.We next performed cell cycle analysis of drug-treated cellsand found a 57% (15.5% versus 6.6% of total cell amount)decrease in the amount of cells in S phase with addition of 1or 10 �M of imatinib (Fig. 1). The variation of the total numberof cells in S phase was �5% for independently repeatedexperiments.

Expression Profiling of GIST Cells Treated with Ima-tinib Using cDNA Microarrays. To identify potential ima-tinib-specific genetic targets, we treated GIST882 with 10 �M

imatinib for 6, 12, 24, and 48 h and isolated RNA. These RNAsamples were labeled with Cy3 or Cy5 dyes and hybridizedto cDNA microarrays containing known genes and ESTs. Thesignificance and confidence analysis produced a list of 148genes or ESTS that were differentially regulated in at leastone time point at 95% confidence. The complete list of thesegenes can be found at online.4 In the hierarchical clusteranalysis, one cluster showed significant down-regulation rel-ative to the control at all time points in the cell cultureexperiment and also contained a number of genes identifiedas confidently differentially regulated (Fig. 2). This clustercontained 62 of the 10,368 genes in the experiment as awhole, with 34 genes in the cluster being down-regulated at95% confidence in at least one condition (i.e., the intersec-tion of the results of the confidence analysis with the down-regulated cluster). The analysis revealed that this cluster isquite small and very close to the root of the tree, indicatingthat these genes cluster well away from the other genes inthe experiment. Because our interest was to identify potentialsurrogate genetic markers of response to imatinib, we furtherrefined this list to only those genes that showed a durableresponse for all time points (at least 2.4-fold). The reason forthis approach is that if genes fluctuate up and down aftertreatment, then they would be unlikely to be reliable markerswhen evaluating tumor samples from patients on clinicaltrials. Based on this criterion, we identified a total of sevenknown genes that displayed this pattern of expression aftertreatment. One gene was strongly up-regulated, and sixgenes were strongly down-regulated at all time points (Table2). To confirm the microarray results, we first revalidated theannotations of our gene list (as describe in “Materials andMethods”) and verified that the seven genes were SPRY4A,ARHGEF2, RTP801, FLJ20898, FZD8, PDE2A, and MAFbx.

To validate the array data, we isolated a panel of cDNAprobes and designed a set of PCR primers for each of thenamed genes, and performed Northern and RT-PCR analy-ses, respectively, using an aliquot of RNA reserved from thearray experiments (Fig. 3; data not shown). We controlled forsaturation of the PCR reaction by sampling each reactiontogether with a �-actin control reaction at different numbersof PCR cycles. All seven genes were differently expressed,as expected from our microarray screen (Fig. 3). SPRY4A,ARHGEF2, RTP801, FLJ20898, and FZD8 mRNA levels werediminished significantly within 6 h after addition of imatinib,and the level of PDE2A mRNA decreased measurably 12 h

4 http://bioinformatics.fccc.edu/supplement/Frolov_GIST_2003/.

Fig. 1. Analysis of GIST882 cell growth in response to imatinib in vitro. A,growth curve of imatinib-treated and untreated control cells as deter-mined by Guava (see “Materials and Methods”). X axis represents hoursof treatment, and Y axis represents the total number of cells (�105).Untreated GIST882 cells, �; GIST882 cells treated with 1 �M imatinib, f;GIST882 cells treated with 10 �M imatinib, Œ. B, representative cell cycleprofiles of untreated and 10 �M imatinib-treated GIST882 cells at 96 h. C,numerical representation of the fraction of cells in G0/1, S, and G2-Mphases of the cell cycle for untreated and imatinib-treated (1 or 10 �M)GIST882 cells at the 72 and 96 h time points.

702 Markers of Response to Gleevec

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

after drug treatment. In contrast, MAFbx levels were dramat-ically induced within 6 h of imatinib exposure and remainedelevated throughout the treatment (Fig. 3).

Inhibition of c-KIT by Imatinib in GIST882 Cells. Wenext examined GIST882 cells for constitutively activated c-KIT, as well as downstream mediators ERK1/2 and AKT,using anti-phosphorylated antibodies. c-KIT, ERK1/2, andAKT were found to be constitutively activated in exponen-tially growing GIST882 cells (Fig. 4A). The inhibitory efficacyof the drug on the activity of c-KIT, ERK1/2, and AKT wasthen evaluated using either 1 or 10 �M imatinib for 0.5–6 h.Both drug concentrations resulted in loss of phosphorylatedc-KIT within 30 min and were specific only for the phospho-rylated protein because the total levels of KIT protein wereunaffected (Fig. 4A). Imatinib also potently inhibited the con-stitutive activation of AKT, as assessed using antibodiesspecific for phosphorylated threonine 308 or serine 473.There was a similar finding for phosphorylated ERK1/2. Aswith c-KIT, the total levels of AKT and ERK1/2 were unaf-fected by imatinib treatment (Fig. 4A). Furthermore, de-creased phosphorylation of AKT and ERK1/2 was rapid andparalleled inhibition of c-KIT (Fig. 4A). We next evaluated theeffects of imatinib on SPRY4A and MAFbx expression inconjunction with inhibition of c-KIT. As shown in Fig. 4B, theexpression of both genes was altered within 3 h of inhibition

of c-KIT. SPRY4A levels were noticeably decreased by 3 hand virtually undetectable by 6 h, and MAFbx mRNA levelswere dramatically increased by 3 h and maintained through-out the course of treatment. Overall, we were able to verifythat imatinib treatment resulted in decreased autophospho-rylation of a K642E mutant of c-KIT by inhibiting c-KIT kinaseactivity rather than by down-regulating expression of thec-KIT protein in GIST882 cells. Furthermore, this inhibitionleads to down-regulation of the activation of ERK1/2 andAKT.

SPRY4A, MAFbx, and ARHGEF2 Expression Is Regu-lated by an ERK1/2-dependent Pathway. Previous studieshave shown that the ERK pathway positively regulates theexpression of the Sprouty genes in mouse cells and that, ina limited number of tumor cell lines that exhibit constitutiveactivation of ERKs, SPRY1 and/or SPRY2 mRNA is elevated(25). To determine whether ERK1/2 pathways also regulatedSPRY4A in GIST cells, we treated the GIST882 line withU0126 (30 �M), a MEK1 inhibitor, for 30 min, 1, 3, and 6 h.Complete suppression of activated ERK1/2 was seen by 30min (Fig. 5A). Interestingly, U0126 inhibits phosphorylation ofAKT shortly after treatment, but this effect was transientbecause phospho-AKT returned to initial levels within 3 h oftreatment (Fig. 5A) and did not affect the levels of total AKT.We next sought to determine whether the panel of imatinib-

Table 2 Imatinib-responsive genes in GIST cells

Clone IDa Accession no. 6 h 12 h 24 h 48 h Mean ratio CVb

SPRY4 AF227516 �3.31 �1.91 �3.73 �2.38 �2.83 � 0.83 0.29FZD8 NM_031866 �2.22 �1.39 �2.54 �0.97 �1.78 � 0.72 0.41HIF-1 responsive gene (RTP801) AF335324 �2.11 �1.17 �1.67 �1.35 �1.58 � 0.41 0.26Hypothetical protein (FLJ20898) NM_024600 �1.63 �0.45 �2.5 �1.49 �1.52 � 0.84 0.55Rho/Rac guanine exchange factor

(ARHGEF2)NM_004723 �1.95 �0.8 �1.39 �1.49 �1.41 � 0.48 0.34

PDE2A NM_002599 �0.45 �0.72 �2.63 �1.25 �1.26 � 0.97 0.77MAFbx AY059629 1.81 1.13 1.61 0.72 1.32 � 0.49 0.37

a Listed are the genes that were differentially expressed at each time point upon treatment of GIST882 cells with imatinib and their accession numbers.Included are the log2 intensity ratios for each time point, and the overall mean log2 ratio for each gene with the standard deviation.b CV, coefficient of variance.

Fig. 2. Hierarchical clustering of themean values at the four data points. Thefull, normalized data were clustered inGeneSight using hierarchical clusteringwith a Euclidean metric into 11 clusters. Inthis image, red represents up-regulation,whereas green represents down-regula-tion. The colors shown overlaying the den-dogram roughly represent the locations ofthe 11 clusters, with the arrow showingthe point on the tree where the key clusterdiverges from the rest of the data.

703Molecular Cancer Therapeutics

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

responsive genes was dependent on ERK1/2 activation byMEK1. Expression patterns of SPRY4A, MAFbx, and ARH-GEF2 in U0126 cells were similar to those after imatinibtreatment (Fig. 5, B and C). PDE2A and RTP801 mRNA levelswere decreased transiently, and the levels of FLJ20898 andFZD8 did not appear to change noticeably (Fig. 5B). The levelof expression after addition of U0126 was comparable withthat seen with imatinib, except that the duration of suppres-sion with a single dose of imatinib was more prolonged (datanot shown). We also evaluated the effect of LY294002, aPI3K inhibitor (26), and we observed little or no effect onexpression levels of any of the imatinib-responsive genes inour restricted panel (Fig. 5D; data not shown). These resultssuggest that several of these genes are likely to be depend-ent on activation of the MAPK signaling pathway by c-KIT,whereas others may not be directly downstream of eitherMAPK or AKT signaling pathways.

Evaluation of Surrogate Markers of Imatinib Responsein Clinical GIST Samples. Ultrasound-guided needle corebiopsies obtained from GIST patients both before and afterinitiation of imatinib therapy were evaluated for expression ofa down-regulated transcript (SPRY4A) and an up-regulatedtranscript (MAFbx) by RT-PCR analysis. Each biopsy wasevaluated by a pathologist for evidence of viable GIST cellswith c-KIT expression by routine histology and immunohis-tochemistry, respectively, before molecular analysis. Fig. 6depicts the baseline and day 15 biopsy for patient 1 (aresponder). There is persistence of malignant tumor cellswith KIT expression in the biopsy taken while the patient wason imatinib. As evident in Fig. 7A, the expression levels of

SPRY4 were dramatically decreased after imatinib therapy intumors from patients (patients 1, 2, 3, and 7) who showedfavorable clinical responses to the drug. The two nonrespon-sive patients (patients 5 and 6) expressed high levels ofSPRY4A in both the pre- and posttreatment biopsies. In thepatient (patient 4) who initially responded to the drug treat-ment but subsequently relapsed, the SPRY4A levels in thetumor decreased dramatically in the biopsy taken duringclinical response but returned to the pretreatment levelsduring disease progression (Fig. 7A). We also assessed theexpression level of MAFbx in clinical samples (patients 1, 4,6, and 7). In patients 1 and 7, both of whom showed favor-able clinical responses, we detected high levels MAFbx tran-scripts in the treated tumor. In the tumor sample from patient4, the level of MAFbx increased during the response butreturned to pretreatment level during the disease progres-sion. In patient 6, who never responded to the drug, the levelof MAFbx was not changed (i.e., not detected) in the pairedsamples (Fig. 7B).

c-KIT, AKT, and ERK1/2 Expression in a Non-imatinib-responsive GIST. For patient 6, a larger biopsy sample (31days after imatinib treatment) was available, which allowedus to evaluate the c-KIT, AKT, and ERK1/2 protein levels afterimatinib exposure in vivo. A Western blot of this nonrespond-ing GIST indicated that c-KIT was constitutively activatedeven in the presence of imatinib therapy (Fig. 8A). For com-parison, we included the GIST882 cell line and a sample of amyxoid liposarcoma, a sarcoma histotype that is generallynot responsive to imatinib. As expected, the liposarcoma

Fig. 3. RT-PCR analysis of imatinib-responsive genes. RNA was isolatedfrom imatinib-treated (10 �M for 6, 12, 24, and 48 h) and untreatedGIST882 cells and reverse transcribed. Shown are ethidium bromide-stained agarose gels of the RT-PCR products for SPRY4A, MAFbx, FZD8,PDE2A, RTP801, FLJ20898, and ARHGEF2. �-Actin is shown as a control.

Fig. 4. Analysis of c-KIT in GIST cells treated with imatinib. A, analysis ofc-KIT, AKT, ERK1/2, and their phosphorylated counterparts in drug-treated and untreated GIST882 cells by immunoblotting. B, RT-PCR anal-ysis of SPRY4A, MAFbx, and �-actin levels in similarly treated cells.

704 Markers of Response to Gleevec

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

failed to express detectable levels of c-KIT (both phospho-rylated and unphosphorylated; Fig. 8A). We also evaluatedAKT and ERK1/2 in these samples. As one would predict,phosphorylated forms of AKT and ERK1/2 were present inthe nonresponding GIST (Fig. 8, B and C). Interestingly, in theliposarcoma specimen, both ERK1/2 and AKT were consti-tutively activated (Fig. 8B), and SPRY4A was expressed at alevel comparable with a typical GIST specimen (data notshown). These results confirm that c-KIT was not down-regulated in response to imatinib in this patient and that theexpression levels of both SPRY4A and MAFbx were predic-tive of this lack of response.

To assess further why this patient failed imatinib therapy,we evaluated c-KIT for a somatic mutation. RNA was isolatedfrom this tumor biopsy and reversed transcribed, and thecDNA was amplified by PCR. The entire transcript was se-quenced, and a mutation was found in exon 9 of the c-KITcDNA. The mutation identified was a 6-base insertion start-ing at nucleotide 1530 (1530ins6) that would result in theaddition of two extra amino acids (Fig. 8D). Multiple alleles

were sequenced, and it was determined that the tumor washeterozygous for the mutation (Fig. 8D; data not shown).These results suggest that this mutation in c-KIT may beassociated with the inability of imatinib to down-regulatec-KIT activity.

DiscussionThe CSTI571-B2222 clinical trial was one of the first to useimatinib for the treatment of a solid tumor (17). Although thisPhase II trial reported a high number of durable partial re-sponses, 13% of patients never responded, and some re-sponding patients have subsequently progressed. Questionsremain as to both the basis of this response and the basis ofprimary and acquired resistance. We hypothesized that byevaluating gene expression profiles from GIST cell lines andthen using these data to evaluate specimens from GISTpatients taken before and after imatinib therapy, we wouldidentify novel genetic biomarkers of this therapy and subse-quently define additional downstream mediators of re-

Fig. 5. MEK inhibition results in de-creased levels of activated ERK1/2levels and altered levels of imatinib-responsive genes. A, analysis ofphospho-ERK1/2/total ERK1/2 andphospho-AKT/total AKT levels inU0126-treated and untreated GISTcells by immunoblotting. RT-PCR anal-ysis of SPRY4A, MAFbx, FZD8,PDE2A, RTP801, FLJ20898, ARH-GEF2, and �-actin levels in U0126-treated (B), imatinib-treated (C), anduntreated GIST cells. D, RT-PCR anal-ysis of SPRY4A, MAFbx, and �-actinlevels in LY294002-treated and un-treated GIST cells.

705Molecular Cancer Therapeutics

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

sponse. Using cDNA microarray approaches, we identified apanel of genes that were responsive to imatinib inhibition ofc-KIT and its subsequent downstream signaling pathways.Our results showed that �1.4% (148 of 10,367) of the genesequences evaluated were significantly (P � 0.05) differen-tially expressed in a GIST cell line after KIT kinase inhibition.Hierarchical clustering identified a single cluster of 62 genesshowing down-regulation at all time points. Because hierar-chical clustering cannot assess significance, we further re-duced the focus from these 62 genes to only those 34 thatshowed statistically significant down-regulation. We nar-rowed this list of candidate genetic response markers to thegenes with the strongest fold changes across all treatmenttime points, seven of which were well annotated and showedsignificant changes in expression at each time point. We first

observed that GIST cells treated with imatinib undergo rapidgrowth arrest, without significant induction of apoptosis. Aprevious study of GIST882 cells found that a portion of thecells undergo programmed cell death in response to ima-tinib, albeit at later time points (16). Other studies haveshown that imatinib can lead to inhibition of cell growth in theabsence of measurable induction of apoptosis (27). Flowcytometry further confirmed that the imatinib-treated cellsaccumulate in G0–1 and G2, almost exhibiting a static growthstate. Furthermore, there was no detectable sub-G0 fraction,which would have been indicative of apoptotic cells. Thisparallels the clinical observation of the absence of initial rapiddecrease in tumor volume on computerized tomographyscan, whereas positron emission tomography scan indicatesa rapid abrogation of metabolic activity suggestive of tumor

Fig. 6. Histology and c-KIT immunohis-tochemistry of biopsies obtained from pa-tient 1. A depicts the biopsy of a livermetastasis before therapy, and B depictsthe biopsy of a liver metastasis after 15days of imatinib therapy. Tumor cells per-sist and express c-KIT as shown in thebottom panels. Magnification, �200.

Fig. 7. Analysis of SPRY4A and MAFbx expressionin pre- and post-imatinib GIST biopsies. A, RT-PCRanalysis of SPRY4A in GIST biopsies of patients whoresponded (R), failed to respond (NR), or respondedinitially and then progressed (R-P). Sample 7 is froman imatinib-treated patient in whom one tumor re-sponded, and one tumor did not. No pretreatmentbiopsy was available from patient 7. B, RT-PCRanalysis of MAFbx in four of patients with GIST. Allseven cases could not be tested for MAFbx due tolimited amounts of tissue in some of the biopsiesand/or loss of RNA. Bottom panel, �-actin as a con-trol.

706 Markers of Response to Gleevec

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

dormancy (28). The genocentric approach used in this studyidentified potential gene products and signaling pathwaysthat may be important in the therapeutic effects of imatinib.We found that the mRNA levels of Sprouty4A (SPRY4A) werethe most significantly down-regulated and that the mRNAlevels of MAFbx were dramatically up-regulated for the du-ration of drug treatment. We also observed and validateddown-regulation of several other known genes (i.e., FZD8,PDE2A, RTP801, FLJ20898, and ARHGEF2) in the cell cul-ture model. We then attempted to validate the gene tran-script most significantly down-regulated (SPRY4A) and the

only gene transcript up-regulated (MAFbx) in vitro by exam-ining sequential GIST biopsy specimens. SPRY4A expres-sion and MAFbx expression were found to be highly reliablepredictors of immediate response to drug, even in small coretumor biopsies (Fig. 7). The rationale behind evaluating thesetwo genes is that MAFbx has been reported to be expressedexclusively in heart and skeletal muscle (29). We have foundfor the first time that this gene is also expressed in smoothmuscle, albeit at much lower levels than seen after imatinibtreatment (data not shown). Therefore, by combining theexpression of both an up-regulated gene and a down-regu-lated gene, we believe that this lends further support thattumor cells are present in the biopsies.

Our study also uncovered an activating mutation in c-KITthat was associated with a failure to respond to imatinib. Thisin-frame mutation (1530ins6) in exon 9 of c-KIT was the oneof the first reported outside the exon 11 juxtamembranedomain (30). Henrich et al. (31) have reported that GISTs withexon 9 mutations have a lower response rate than GISTs withexon 11 mutations. It is possible that this exon 9 mutationactivates the receptor via ligand-independent oligomeriza-tion. It was shown that imatinib binds to the Thr670 of theATP-binding site of the kinase (corresponds to exon 14).Importantly, imatinib can only bind to an inactive conforma-tion of the kinase, in which the NH2-terminal part of theactivation loop is folded into the ATP-binding site (32). There-fore, it is possible that the 1530ins6 mutation could lead toactivation of this site, thereby preventing imatinib from bind-ing and inhibiting the kinase activity. However, additionalstudies to evaluate a range of GIST-associated c-KIT muta-tions and their response to imatinib are needed.

The panel of genes uncovered in our study is also ofpotential interest from a functional aspect to further elucidatethe molecular mechanisms of imatinib. There are little dataregarding the therapeutic effect of imatinib on GISTs, otherthan that the drug inhibits activation of c-KIT and severaldownstream mediators. Our studies confirmed that both AKTand ERK1/2 signaling pathways are rapidly inhibited afterexposure to imatinib. Interestingly, our studies with inhibitorsof these pathways, LY294002, a PI3K inhibitor, and U0126, aMEK1 inhibitor, suggest that other signaling pathways mayalso be affected by imatinib treatment, and additional studiesshould elucidate this issue. Specifically, we found that mRNAlevels of SPRY4A and ARHGEF2 are down-regulated,whereas MAFbx levels are up-regulated after treatment withU0126 or imatinib. Interestingly, the kinetics and pattern bywhich the two drugs inhibit these markers are quite distinct.Even though we have demonstrated that all seven markersrespond to imatinib, albeit at different rates, inhibition of twomajor pathways downstream of c-KIT do not appear to ac-count for all of the differences we observe in expressionpatterns (compare Fig. 5, B, C, and D). These results suggestthat several of the markers identified are not directly down-stream of any of these signaling pathways and warrant fur-ther investigation regarding the action of imatinib.

We have also used hierarchical clustering to identify asubset of genes that includes SPRY4A, which showed thehighest degree of down-regulation. Expression levels of thetranscripts for SPRY4A and ARHGEF2, for example, are

Fig. 8. c-KIT, ERK1/2, and AKT expression in a non-imatinib-responsiveGIST (patient 6 posttreatment), a liposarcoma, and GIST cells by immu-noblotting. A, phospho-c-KIT and total c-KIT protein; B, phospho-AKTand total AKT levels; and C, phospho-ERK1/2 and total ERK1/2 levels inprotein extracts from a GIST biopsy 31 days after imatinib treatment, amyxoid liposarcoma biopsy, and untreated GIST882 cells. D, represent-ative sequence electropherograms of the mutant and wild-type c-KIT fromtumor RNA isolated from the post-imatinib biopsy of patient 6. Sequenc-ing of the mutant transcripts revealed a 6-base insertion at nucleotide1530 (i.e., 1530ins6) in exon 9. The sequence shown in the bottom panelrepresents the wild-type allele expressed in the tumor from patient 6.

707Molecular Cancer Therapeutics

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

consistently down-regulated by imatinib treatment in theGIST cell line. Interestingly the proteins encoded by thesetwo genes appeared to be involved in the control of feedbackloops for RAS signaling associated with coadapter proteinbinding and RAS activation (25). The coordinated down-regulation of these two genes and presumably their proteinssuggests a potential mechanism for immediate downstreamimatinib-induced inhibition of KIT signaling.

The down-regulation of the PDE2 transcript may directlyrelate to its role in the transduction of cell survival signals.c-GMP-stimulated PDE2 was shown to be up-regulated incertain cancers and functions to control the cAMP level in thecell (33). It was also shown that some RTKs may transductsignal not only through adapter proteins but also throughsecondary messengers. In this case, PDEs play a crucial rolein transmitting the signal from the receptor and the responseto various stimuli. It was recently shown that PDE3 plays avery important role in insulin-like growth factor I receptorsignaling in pancreatic � cells (34), and PDE2A was shown tobe down-regulated in exisulind-induced apoptosis (35). Itwas also shown that insulin-like growth factor I-dependentregulation of PDE3B might be linked to cell survival throughPI3K and not p42/p44 MAPK (36). Recent studies demon-strated that GISTs overexpress genes encoding G-protein-coupled receptor 20 and protein kinase C � (37). Both ofthese genes were not only highly expressed in GIST but alsoclustered very tight with it. PDEs were shown to be activatedby protein kinase C and enhance the degradation of cyclicmonophosphates (38), thus playing a critical role in signaltransduction. This evidence further supports the suggestionof possible horizontal interactions between RTK and G-protein-coupled receptor signaling.

The discovery of the up-regulation of the muscle atrophyF-box (MAFbx) is of particular interest. MAFbx, also knownas Atrogin-1, encodes a ubiquitin ligase, a protein that bindsand mediates ubiquitination of specific substrates (29, 39). Ithas been shown that overexpression of MAFbx in the myo-tubules of mice results in muscular atrophy when comparedwith mice deficient in this gene (29). The presence of theup-regulated MAFbx transcript in imatinib-sensitive GISTsuggests a potentially important role in promoting tumornecrosis, which is of interest, given that the GIST progenitorcells have myogenic differentiation properties. Furthermore,we found that treatment of GIST882 cells with LY294002resulted in a modest but detectable increase in MAFbxmRNA levels (Fig. 5D). This is interesting, given that severalproteins, including AKT in the PI3K/AKT cascade, have beenshown to have a positive effect on muscle hypertrophy (40).Therefore, down-regulation of AKT, as observed in GISTsafter treatment with imatinib, may have the opposite effect.Furthermore, it is intriguing to speculate that up-regulation ofproteins, such as MAFbx, may also contribute to the gener-alized muscle cramping that is a common side effect inpatients taking imatinib.

Clearly further work will need to be done in exploring thefunctional consequences of these genomic GIST profilescomparing molecular events before and after imatinib ad-ministration. This initial study with unique patient sampleshas yielded important observations regarding molecular pro-

filing of GIST patients on imatinib therapy. The correlation ofSPRY4A and MAFbx expression in relation to clinical re-sponse is quite uniform and compelling, particularly in pa-tient 4, with initial response (SPRY4A down-regulated andMAFbx up-regulated) followed by disease progression(SPRY4A up-regulated and MAFbx down-regulated).SPRY4A expression in these patients is interesting, particu-larly with reference to a recent publication supportive ofadditional evidence of SPRY4A involvement in GISTs (41).This report indicates that SPRY1, SPRY4, and c-KIT areamong the genes that distinguish GISTs from other sarco-mas. Therapeutic implications are speculative; however, pro-teins such as those encoded by SPRY4A and RTP801,whose expression is elevated in GISTs, provide potentialtargets for drug discovery purposes, especially in tumorsrefractory to imatinib.

In summary, we have used a genomic-based approach toidentify several surrogate response markers of imatinib ther-apy in a limited number of GIST cell lines and in GIST clinicalspecimens. Functional studies, which are under way, willhopefully resolve which of them are important in the thera-peutic effects of this small molecule targeted agent. Thisstudy provides a platform for continuing investigations intothe molecular basis for imatinib response in GISTs, whichmay result in the development of new therapeutic targets andreliable markers of clinical response.

AcknowledgmentsWe acknowledge Drs. Alfred Knudson and Phillip A. Godwin for insightfulcomments regarding the manuscript, Sharon Howard and the TissueCulture Facility for technical assistance, and the DNA microarray facility,the DNA sequencing facility, the bioinformatics facility, and the BiosampleRepository5 at Fox Chase Cancer Center for services and specimens. Wealso thank Dr. Rosaleen Parsons and Dr. Barton Milestone for expertassistance obtaining the core biopsies, Dr. Harry Cooper for pathologicassessment of the tumor biopsies, and Monica Davey for excellent care ofthe study patients and data management. We acknowledge the support ofDr. Sasa Dimitrijevic (Novartis, Basel, Switzerland).

References1. Graadt van Roggen, J. F., van Velthuysen, M. L., and Hogendoorn,P. C. The histopathological differential diagnosis of gastrointestinal stro-mal tumours. J. Clin. Pathol., 54: 96–102, 2001.2. Rumessen, J. J., Peters, S., and Thuneberg, L. Light- and electronmicroscopical studies of interstitial cells of Cajal and muscle cells at thesubmucosal border of human colon. Lab. Investig., 68: 481–495, 1993.3. Lasota, J., Jasinski, M., Sarlomo-Rikala, M., and Miettinen, M. Muta-tions in exon 11 of c-Kit occur preferentially in malignant versus benigngastrointestinal stromal tumors and do not occur in leiomyomas or leiomy-osarcomas. Am. J. Pathol., 154: 53–60, 1999.4. Rubin, B. P., Singer, S., Tsao, C., Duensing, A., Lux, M. L., Ruiz, R.,Hibbard, M. K., Chen, C. J., Xiao, S., Tuveson, D. A., Demetri, G. D.,Fletcher, C. D., and Fletcher, J. A. KIT activation is a ubiquitous feature ofgastrointestinal stromal tumors. Cancer Res., 61: 8118–8121, 2001.5. Qiu, F. H., Ray, P., Brown, K., Barker, P. E., Jhanwar, S., Ruddle, F. H., andBesmer, P. Primary structure of c-kit: relationship with the CSF-1/PDGFreceptor kinase family: oncogenic activation of v-kit involves deletion ofextracellular domain and C terminus. EMBO J., 7: 1003–1011, 1988.6. Fletcher, J. A., Fletcher, C. D., Rubin, B. P., Ashman, L. K., Corless,C. L., and Heinrich, M. C. KIT gene mutations in gastrointestinal stromal

5 http://www.fccc.edu/clinicalresearch/BRCF/.

708 Markers of Response to Gleevec

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

tumors: more complex than previously recognized? Am. J. Pathol., 161:737–738; author reply, 738–739, 2002.7. Heinrich, M. C., Corless, C. L., Duensing, A., McGreevey, L., Chen,C. J., Joseph, N., Singer, S., Griffith, D. J., Haley, A., Town, A., Demetri,G. D., Fletcher, C. D., and Fletcher, J. A. PDGFRA activating mutations ingastrointestinal stromal tumors. Science (Wash. DC), 299: 708–710, 2003.8. Miettinen, M., Sarlomo-Rikala, M., and Lasota, J. Gastrointestinal stro-mal tumors: recent advances in understanding of their biology. Hum.Pathol., 30: 1213–1220, 1999.9. DeMatteo, R. P., Lewis, J. J., Leung, D., Mudan, S. S., Woodruff, J. M.,and Brennan, M. F. Two hundred gastrointestinal stromal tumors: recurrencepatterns and prognostic factors for survival. Ann. Surg., 231: 51–58, 2000.10. Buchdunger, E., Cioffi, C. L., Law, N., Stover, D., Ohno-Jones, S.,Druker, B. J., and Lydon, N. B. Abl protein-tyrosine kinase inhibitor STI571inhibits in vitro signal transduction mediated by c-kit and platelet-derivedgrowth factor receptors. J. Pharmacol. Exp. Ther., 295: 139–145, 2000.11. Druker, B. J. STI571 (Gleevec) as a paradigm for cancer therapy.Trends Mol. Med., 8: S14�S18, 2002.12. Kantarjian, H., Sawyers, C., Hochhaus, A., Guilhot, F., Schiffer, C.,Gambacorti-Passerini, C., Niederwieser, D., Resta, D., Capdeville, R.,Zoellner, U., Talpaz, M., and Druker, B. Hematologic and cytogeneticresponses to imatinib mesylate in chronic myelogenous leukemia. N. Engl.J. Med., 346: 645–652, 2002.13. Talpaz, M., Silver, R. T., Druker, B. J., Goldman, J. M., Gambacorti-Passerini, C., Guilhot, F., Schiffer, C. A., Fischer, T., Deininger, M. W.,Lennard, A. L., et al. Imatinib induces durable hematologic and cytoge-netic responses in patients with accelerated phase chronic myeloid leu-kemia: results of a phase 2 study. Blood, 99: 1928–1937, 2002.14. Sawyers, C. L., Hochhaus, A., Feldman, E., Goldman, J. M., Miller,C. B., Ottmann, O. G., Schiffer, C. A., Talpaz, M., Guilhot, F., Deininger,M. W., Fischer, T., O’Brien, S. G., Stone, R. M., Gambacorti-Passerini,C. B., Russell, N. H., Reiffers, J. J., Shea, T. C., Chapuis, B., Coutre, S.,Tura, S., Morra, E., Larson, R. A., Saven, A., Peschel, C., Gratwohl, A.,Mandelli, F., Ben-Am, M., Gathmann, I., Capdeville, R., Paquette, R. L.,and Druker, B. J. Imatinib induces hematologic and cytogenetic re-sponses in patients with chronic myelogenous leukemia in myeloid blastcrisis: results of a Phase II study. Blood, 99: 3530–3539, 2002.15. Heinrich, M. C., Griffith, D. J., Druker, B. J., Wait, C. L., Ott, K. A., andZigler, A. J. Inhibition of c-kit receptor tyrosine kinase activity by STI 571,a selective tyrosine kinase inhibitor. Blood, 96: 925–932, 2000.16. Tuveson, D. A., Willis, N. A., Jacks, T., Griffin, J. D., Singer, S.,Fletcher, C. D., Fletcher, J. A., and Demetri, G. D. STI571 inactivation ofthe gastrointestinal stromal tumor c-KIT oncoprotein: biological and clin-ical implications. Oncogene, 20: 5054–5058, 2001.17. Demetri, G. D., von Mehren, M., Blanke, C. D., Van den Abbeele, A. D.,Eisenberg, B., Roberts, P. J., Heinrich, M. C., Tuveson, D. A., Singer, S.,Janicek, M., Fletcher, J. A., Silverman, S. G., Silberman, S. L., Capdeville, R.,Kiese, B., Peng, B., Dimitrijevic, S., Druker, B. J., Corless, C., Fletcher, C. D.,and Joensuu, H. Efficacy and safety of imatinib mesylate in advanced gas-trointestinal stromal tumors. N. Engl. J. Med., 347: 472–480, 2002.18. van Oosterom, A. T., Judson, I., Verweij, J., Stroobants, S., Donato diPaola, E., Dimitrijevic, S., Martens, M., Webb, A., Sciot, R., Van Glabbeke,M., Silberman, S., and Nielsen, O. S. Safety and efficacy of imatinib(STI571) in metastatic gastrointestinal stromal tumours: a Phase I study.Lancet, 358: 1421–1423, 2001.19. Frolov, A., Prowse, A. H., Vanderveer, L., Bove, B., Wu, H., and Godwin,A. K. DNA array-based method for detection of large rearrangements in theBRCA1 gene. Genes Chromosomes Cancer, 35: 232–241, 2002.20. Schultz, D. C., Vanderveer, L., Berman, D. B., Hamilton, T. C., Wong,A. J., and Godwin, A. K. Identification of two candidate tumor suppressorgenes on chromosome 17p13.3. Cancer Res., 56: 1997–2002, 1996.21. Kerr, M. K., Afshari, C. A., Bennett, L., Bushel, P., Martinez, J., Walker,N. J., and Churchill, G. A. Statistical analysis of a gene expression mi-croarray experiment with replication. Statistica Sinica, 12: 203–218, 2002.22. Ideker, T., Thorsson, V., Siegel, A. F., and Hood, L. E. Testing fordifferentially-expressed genes by maximum-likelihood analysis of mi-croarray data. J. Comput. Biol., 7: 805–817, 2000.23. Eisen, M. B., Spellman, P. T., Brown, P. O., and Botstein, D. Clusteranalysis and display of genome-wide expression patterns. Proc. Natl.Acad. Sci. USA, 95: 14863–14868, 1998.

24. Pan, Z. Z., Bruening, W., Giasson, B. I., Lee, V. M., and Godwin, A. K.�-Synuclein promotes cancer cell survival and inhibits stress- and chem-otherapy drug-induced apoptosis by modulating MAPK pathways. J. Biol.Chem., 277: 35050–35060, 2002.25. Ozaki, K., Kadomoto, R., Asato, K., Tanimura, S., Itoh, N., and Kohno,M. ERK pathway positively regulates the expression of Sprouty genes.Biochem. Biophys. Res. Commun., 285: 1084–1088, 2001.26. Vlahos, C. J., Matter, W. F., Hui, K. Y., and Brown, R. F. A specificinhibitor of phosphatidylinositol 3-kinase, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002). J. Biol. Chem., 269: 5241–5248, 1994.27. Druker, B. J., Tamura, S., Buchdunger, E., Ohno, S., Segal, G. M.,Fanning, S., Zimmermann, J., and Lydon, N. B. Effects of a selectiveinhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells.Nat. Med., 2: 561–566, 1996.28. von Mehren, M., Blanke, C. D., Joensuu, H., Heinrich, M. C., Roberts,P., Eisenberg, B., Sliberman, S., Dimitrijevic, S., Kies, B., Fletcher, J., et al.High incidence of durable responses induced by imatinib mesylate(Gleevec) in patients with unresectable and metastatic gastrointestinalstromal tumors (GISTs). Proc. Am. Soc. Clin. Oncol., 21: 1608a, 2002.29. Bodine, S. C., Latres, E., Baumhueter, S., Lai, V. K., Nunez, L., Clarke,B. A., Poueymirou, W. T., Panaro, F. J., Na, E., Dharmarajan, K., Pan,Z. Q., Valenzuela, D. M., DeChiara, T. M., Stitt, T. N., Yancopoulos, G. D.,and Glass, D. J. Identification of ubiquitin ligases required for skeletalmuscle atrophy. Science (Wash. DC), 294: 1704–1708, 2001.30. Lux, M. L., Rubin, B. P., Biase, T. L., Chen, C. J., Maclure, T., Demetri,G., Xiao, S., Singer, S., Fletcher, C. D., and Fletcher, J. A. KIT extracellularand kinase domain mutations in gastrointestinal stromal tumors. Am. J.Pathol., 156: 791–795, 2000.31. Henrich, M. C., Croless, C. L., Banke, C., Demetri, G. D., Joensuu, H.,von Mehren, M., McGreevey, L. S., Wait, C. L., Griffin, J. D., Chen, C. J.,Haley, A., Kiese, B., Druker, B., Roberts, P., Eisenberg, B., Singer, S.,Silberman, S., Dimitrijevic, S., Fletcher, C. D. M., and Fletcher, J. A. KITmutational status predicts clinical response to STI571 in patients withmetastatic gastrointestinal stromal tumors (GISTs). Proc. Am. Soc. Clin.Oncol., 21: 2a, 2002.32. Manley, P. W., Cowan-Jacob, S. W., Buchdunger, E., Fabbro, D., Fen-drich, G., Furet, P., Meyer, T., and Zimmermann, J. Imatinib: a selectivetyrosine kinase inhibitor. Eur. J. Cancer, 38 (Suppl. 5): S19�S27, 2002.33. Solntseva, T. I., and Belousova, A. K. Some features of cyclic aden-osine monophosphate metabolism in mouse liver and hepatoma 22 (inRussian). Biokhimia, 42: 1331–1337, 1977.34. Harndahl, L., Jing, X. J., Ivarsson, R., Degerman, E., Ahren, B., Manga-niello, V. C., Renstrom, E., and Holst, L. S. Important role of phosphodies-terase 3B for the stimulatory action of cAMP on pancreatic �-cell exocytosisand release of insulin. J. Biol. Chem., 277: 37446–37455, 2002.35. Thompson, W. J., Piazza, G. A., Li, H., Liu, L., Fetter, J., Zhu, B., Sperl,G., Ahnen, D., and Pamukcu, R. Exisulind induction of apoptosis involvesguanosine 3�,5�-cyclic monophosphate phosphodiesterase inhibition,protein kinase G activation, and attenuated �-catenin. Cancer Res., 60:3338–3342, 2000.36. Ahmad, M., Flatt, P. R., Furman, B. L., and Pyne, N. J. The role of thecyclic GMP-inhibited cyclic AMP-specific phosphodiesterase (PDE3) inregulating clonal BRIN-BD11 insulin secreting cell survival. Cell Signalling,12: 541–548, 2000.37. Allander, S. V., Nupponen, N. N., Ringner, M., Hostetter, G., Maher,G. W., Goldberger, N., Chen, Y., Carpten, J., Elkahloun, A. G., and Meltzer,P. S. Gastrointestinal stromal tumors with KIT mutations exhibit a remarkablyhomogeneous gene expression profile. Cancer Res., 61: 8624–8628, 2001.38. Bian, J. S., Zhang, W. M., Pei, J. M., and Wong, T. M. The role ofphosphodiesterase in mediating the effect of protein kinase C on cyclicAMP accumulation upon �-opioid receptor stimulation in the rat heart.J. Pharmacol. Exp. Ther., 292: 1065–1070, 2000.39. Gomes, M. D., Lecker, S. H., Jagoe, R. T., Navon, A., and Goldberg,A. L. Atrogin-1, a muscle-specific F-box protein highly expressed duringmuscle atrophy. Proc. Natl. Acad. Sci. USA, 98: 14440–14445, 2001.40. Glass, D. J. Signalling pathways that mediate skeletal muscle hyper-trophy and atrophy. Nat. Cell Biol., 5: 87–90, 2003.41. Nielsen, T. O., West, R. B., Linn, S. C., Alter, O., Knowling, M. A.,O’Connell, J. X., Zhu, S., Fero, M., Sherlock, G., Pollack, J. R., Brown,P. O., Botstein, D., and van de Rijn, M. Molecular characterisation of softtissue tumours: a gene expression study. Lancet, 359: 1301–1307, 2002.

709Molecular Cancer Therapeutics

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

2003;2:699-709. Mol Cancer Ther   Andrey Frolov, Santiago Chahwan, Michael Ochs, et al.  

1of Gleevec in Gastrointestinal Stromal TumorsResponse Markers and the Molecular Mechanisms of Action

  Updated version

  http://mct.aacrjournals.org/content/2/8/699

Access the most recent version of this article at:

   

   

  Cited articles

  http://mct.aacrjournals.org/content/2/8/699.full#ref-list-1

This article cites 40 articles, 17 of which you can access for free at:

  Citing articles

  http://mct.aacrjournals.org/content/2/8/699.full#related-urls

This article has been cited by 15 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  SubscriptionsReprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://mct.aacrjournals.org/content/2/8/699To request permission to re-use all or part of this article, use this link

on May 10, 2020. © 2003 American Association for Cancer Research. mct.aacrjournals.org Downloaded from