EP

EDJ*

BradrdttsPFsnaDaPvBaewfDcGodtG

TAf(spaapst

dsm

CLIN

ICA

L–A

LIMEN

TARY

TRA

CT

GASTROENTEROLOGY 2006;131:1734 –1742

ffects of PKC412, Nilotinib, and Imatinib Against GIST-AssociatedDGFRA Mutants With Differential Imatinib Sensitivity

LLEN WEISBERG,* RENEE D. WRIGHT,* JINGRUI JIANG,* ARGHYA RAY,* DAISY MORENO,* PAUL W. MANLEY,‡

ORIANO FABBRO,‡ ELIZABETH HALL–MEYERS,* LAURIE CATLEY,* KLAUS PODAR,* ANDREW L. KUNG,* andAMES D. GRIFFIN*

Dana Farber Cancer Institute, Boston, Massachusetts, and ‡Novartis Institutes of Biomedical Research, Basel, Switzerland

trsme

auaPpeFiPHtPt

olatwkpn(amPaaibiIir

Agt

ackground & Aims: Activating mutations in platelet-de-ived growth factor receptor � (PDGFRA) have been reported insubset of gastrointestinal stromal tumor (GIST) patients whoo not express the mutant stem cell factor receptor c-kit. Theesponsiveness of mutant PDGFRA-positive GIST to imatinibepends on the location of the PDGFRA mutation; for example,he V561D juxtamembrane domain mutation is more sensitiveo imatinib than the D842V kinase domain mutation. In thistudy, we compare the effects of 3 tyrosine kinase inhibitors,KC412 and nilotinib, and imatinib, on 2 GIST-related PDG-RA mutants, V561D and D842V, which possess differentialensitivity to imatinib. Methods: The effects of PKC412,ilotinib, and imatinib, alone and in combination, were evalu-ted via in vitro proliferation studies performed with V561D- or842V-PDGFRA mutants. The effects of nilotinib and PKC412,

lone and combined, were investigated in vivo. Results:KC412 potently inhibited the V561D-PDGFRA mutant initro and the D842V-PDGFRA mutant in vitro and in vivo.oth imatinib and nilotinib displayed potent activity in vitrogainst the V561D-PDGFRA mutant but were significantly lessfficacious against D842V-PDGFRA. However, when combinedith either imatinib or PKC412, nilotinib showed no evidence

or antagonism and acted in a cooperative fashion against842V-PDGFRA. Conclusions: Our findings support the

linical testing of PKC412 for treatment of mutant PDGFRA-IST. The data also support the use of nilotinib as a treatmentption for V561D-PDGFRA-associated GIST, although the re-uced sensitivity of D842V-PDGFRA probably limits the poten-ial of nilotinib monotherapy for D842V-PDGFRA-associatedIST.

he gastrointestinal stromal tumor (GIST) is the most com-mon mesenchymal tumor of the gastrointestinal (GI) tract.

pproximately 90% of GISTs have been shown to have gain-of-unction mutations in the stem cell factor receptor c-kit.1 ImatinibGleevec, STI571; Novartis Pharma AG, Basel, Switzerland) is amall molecule kinase inhibitor that selectively targets c-kit andlatelet-derived growth factor receptor (PDGFR) tyrosine kinases,s well as breakpoint cluster region-Abelson (Bcr-Abl).2 Imatinib isn effective therapy for GIST and is currently used to treat c-kit–ositive metastatic or unresectable tumors.1,3 However, despite theuccess of imatinib against this malignancy, acquired drug resis-ance can develop.

The responsiveness of GISTs to imatinib varies significantly,epending on which exon the c-kit mutation is located. For in-tance, whereas GIST patients with c-kit exon 11 mutations are the

ost prevalent and these patients respond well to imatinib, pa-

ients whose tumors express the c-kit exon 9 mutation are lessesponsive.3,4 These responses are in accordance with in vitrotudies that suggest that GISTs harboring regulatory-region c-kit

utations are more sensitive to imatinib than GISTs harboringnzymatic-region mutations.5

Of GISTs in which only wild-type c-kit can be detected,pproximately 16% harbor PDGFRA mutations, which result inp-regulation of the same downstream signaling componentss c-kit.4,6 As with c-kit–associated GISTs, the responsiveness ofDGFRA-associated GISTs to imatinib treatment varies de-ending on the exonic location of the PDGFRA mutation. Thexon 18, D842V-PDGFRA, mutation is the most common PDG-RA mutation and has been reported to be resistant to imatinib

n vitro.3,6,7 GIST patients possessing exon 18 mutations inDGFRA typically do not respond well to imatinib therapy.owever, the D842V-PDGFRA mutation does respond to other

yrosine kinase inhibitors, such as the staurosporine derivativeKC412.8 The rarer V561D mutation is located at the jux-amembrane domain of exon 12 in the PDGFRA gene.9

In light of the development of primary and secondary resistancef GIST toward imatinib emerging as a significant clinical chal-

enge in the treatment of this disease, there is a need to identifydditional compounds that can serve as alternative strategies forhe treatment of imatinib-resistant GIST patients. As such, weere interested in studying the in vitro and in vivo efficacy of othernown inhibitors of PDGFRA against some of the mutants seen inatients with GIST. Here, we report findings with PKC412, aonselective inhibitor of protein kinases,10,11 and nilotinib

AMN107), a novel inhibitor of Bcr-Abl, c-kit, and PDGFR.12 Inddition to inhibiting FLT3, which is mutated in 35% of acuteyeloblastic leukemia patients, PKC412 inhibits c-kit and

DGFR10,11; PKC412 is currently in clinical trials in patients withcute myeloblastic leukemia with mutations in FLT3. Nilotinib ispotent inhibitor of Bcr-Abl, designed to override resistance to

matinib caused by point mutations in Bcr-Abl that disrupt theinding of imatinib to its target12; nilotinib is currently undergo-

ng clinical evaluation in myeloproliferative diseases and in GIST.n this study, we investigated the antiproliferative and kinasenhibitory effects of PKC412, nilotinib, and imatinib against GIST-elated PDGFRA mutants showing varying degrees of imatinib

Abbreviations used in this paper: Bcr-Abl, breakpoint cluster region-belson; GIST, gastrointestinal stromal tumor; PDGFR, platelet-derivedrowth factor receptor; PDGFRA, platelet-derived growth factor recep-or �; PKC, protein kinase C.

© 2006 by the AGA Institute0016-5085/06/$32.00

doi:10.1053/j.gastro.2006.09.017

siaw

FtHCcsc5HHlwIw

vmSdBS1Bb(pd

sctrr

FccThDN A � 5

CLI

NIC

AL–

ALI

MEN

TARY

TRA

CT

December 2006 MUTANT PDGFRA: PKC412, NILOTINIB, IMATINIB 1735

ensitivity. We also explored the effects of combinations of the 3nhibitors against these mutants in vitro and investigated thentiproliferative activity of PKC412, alone and in combinationith nilotinib, against the D842V-PDGFRA mutant in vivo.

Materials and MethodsCell Lines and Cell CultureConstructs of D842V-, V561D-, and wild-type (wt) PDG-

RA complementary DNA (cDNA) cloned into pcDNA3.1 (ob-ained from M.C. Heinrich, Department of Pathology, Division ofematology and Oncology, Oregon Health & Science Universityancer Institute, Portland, OR) were stably transfected into Ba/F3

ells by electroporation, and cells were selected for neomycin re-istance and interleukin (IL)-3-independent growth. All cells wereultured in the presence of 5% CO2 at 37°C, at a concentration of� 105 cells/mL, in Cellgro RPMI 1640 medium (Mediatech, Inc.,erndon, VA), supplemented with 10% fetal calf serum (FCS;arlan Bioproducts, Indianapolis, IN), 1% glutamine, and penicil-

in/streptomycin. Parental Ba/F3 cells or wt-PDGFRA-Ba/F3 cellsere cultured with 15% WEHI-conditioned medium as a source of

L-3. All transfected cells were cultured in media supplemented

igure 1. Effects of PKC412, nilotinib, and imatinib on proliferationontrol (untreated) cells. Error bars represent the standard error of the mells (�IL-3) with PKC412 (n � 2). (B) Three-day treatment of D842V-hree-day treatment of D842V-PDGFRA-Ba/F3 cells (�IL-3) with imatinour with nilotinib, imatinib, and PKC412, respectively. Densitometry842V control No. 1 pTyr/PDGFRA � 9112/7680; AMN107 pTyr/PDGo. 2 pTyr/PDGFRA � 8431/6960; PKC412, 0.1 �mol/L pTyr/PDGFR

ith 1 mg/mL G418. c

Chemical Compounds, Antibodies, andImmunoblottingPKC412, nilotinib, and imatinib were synthesized at No-

artis Pharma AG, Basel, Switzerland, and were dissolved in di-ethyl sulfoxide (DMSO) to make 10 mmol/L stock solutions.

erial dilutions were then made, also in DMSO, to obtain finalilutions for cellular assays. Anti-p-Tyr (clone 4G10; Upstateiotechnology, Lake Placid, NY) and anti-p-Tyr (pY99, Sc-7020;anta Cruz Biotechnology, Santa Cruz, CA) were each used at:1000 for immunoblotting. PDGFRA antibody (C-20; Santa Cruziotechnology) was used at 1:200 for immunoblotting. The �-tu-ulin (clone DM1A) antibody was purchased from Sigma-AldrichSt. Louis, MO) and was used at a 1:2000 dilution. Protein lysisreparation and immunoblotting were carried out as previouslyescribed.11

Proliferation StudiesThe trypan blue exclusion assay has been previously de-

cribed11 and was used for all cell proliferation studies. For drugombination studies, PKC412 and nilotinib, PKC412 and ima-inib, or nilotinib and imatinib were added simultaneously at fixedatios to either D842V- or V561D-PDGFRA-Ba/F3 cells. Dose-esponse curves were generated, and combination indices were

842V-PDGFRA-Ba/F3 cells. Cell viability is reported as percentage ofor each data point. (A) Three-day treatment of D842V-PDGFRA-Ba/F3RA-Ba/F3 (�IL-3) and PDGFRA-Ba/F3 cells with nilotinib (n � 2). (C)� 2). (D) Immunoblot: Treatment of D842V-PDGFRA-Ba/F3 cells for 1is immunoblot yielded the following values, reported as area (pixels):� 8146/9632; imatinib pTyr/PDGFRA � 6859/10382; D842V control541/9609; PKC412, 1 �mol/L pTyr/PDGFRA � 5759/13461.

of Dean f

PDGFib (n

for thFRA

alculated as previously described by isobologram analysis.12

rpDMbvp

dtcblt

(6

FdDc .

T

DDDVV

N

CLIN

ICA

L–A

LIMEN

TARY

TRA

CT

1736 WEISBERG ET AL GASTROENTEROLOGY Vol. 131, No. 6

Mouse Studies and In Vivo ImagingD842V-PDGFRA-Ba/F3 cells were transduced with a ret-

ovirus encoding firefly luciferase (MSCV-Luc) and selected withuromycin at a concentration of 0.5–1 �g/mL to generate the842V-PDGFRA-Ba/F3-luciferase (luc�) cell line. Cells free ofycoplasma and viral contamination were resuspended in Hank’s

alanced salt solution (HBSS; Mediatech, Inc., VA) prior to intra-enous (IV) administration to mice. Solutions of nilotinib wererepared by dissolving 200 mg in 1.0 mL of 1-methyl-2-pyrroli-

igure 2. Proliferation studies showing combination effects of PKCay treatments of D842V-PDGFRA-Ba/F3 cells with nilotinib, imatinib842V-PDGFRA-Ba/F3 cells with PKC412, nilotinib, or a combination oells with PKC412, imatinib, or a combination of PKC412 plus imatinib

able 1. Combination Indices Calculated for Dose-Response

Cell lines (treatments) ED25

842V-Ba/F3 (nilotinib�Imatinib) 0.48812842V-Ba/F3 (nilotinib�PKC412) 1.00967842V-Ba/F3 (nilotinib�PKC412) 1.38372561D-Ba/F3 (nilotinib�Imatinib) 0.80100561D-Ba/F3 (nilotinib�PKC412) 1.12460

OTE. ED25–ED75, where ED50 is the median effective dose.

one to give a clear solution and diluted daily prior to adminis-ration with 9.0 mL PEG300. Six percent wt/wt PKC412 in Gelu-ire 44/14 (Gattefosse, France) was diluted with 1X phosphate-uffered saline (PBS) and warmed in a 42°C water bath until

iquid. Placebo mice received vehicles for both PKC412 and nilo-inib, administered 30–45 minutes apart.

Male NCr athymic nude outbred mice (CrTac:NCr-Foxn1�)5– 6 weeks of age; Taconic, NY) were administered a total of00,000 D842V-PDGFRA-Ba/F3-luc� cells by tail vein injec-

nilotinib, and imatinib against D842V-PDGFRA-Ba/F3 cells. (A) Three-a combination of nilotinib plus imatinib. (B) Three-day treatments of412 plus nilotinib. (C) Three-day treatments of D842V-PDGFRA-Ba/F3

es Shown in Figures 2 and 5

Combination indices

ED50 ED75 ED90

0.73221 1.09836 1.647620.80135 0.68007 0.618171.01669 0.77981 0.624621.00712 1.81556 5.230121.42547 1.81579 2.32421

412,, orf PKC

Curv

tacmaltmPwPwlrt

oh

aI�

gsmt

tspmtiPypFritm

wppPodE(idiea

tPBchacyoasbc

FtiaPntnuDap2p2

CLI

NIC

AL–

ALI

MEN

TARY

TRA

CT

December 2006 MUTANT PDGFRA: PKC412, NILOTINIB, IMATINIB 1737

ion. Mice were imaged, and total body luminescence quantifieds previously described.13 Baseline imaging 1 day after tumorell inoculation was used to establish treatment cohorts withatched tumor burden. Cohorts of mice were treated with oral

dministration of vehicle, 50 –100 mg/kg/day PKC412 (formu-ated as above; 4 – 6 days total treatment with 100 mg/kg, 6 daysotal treatment with 50 mg/kg), 150 mg/kg/day nilotinib (for-

ulated as above; 6 days total treatment), or a combination ofKC412 and nilotinib (6 days total treatment). Repeat imagingas performed at various intervals. At the planned end of theKC412 and nilotinib combination study (approximately 1eek following initial IV injection of D842V-PDGFRA-Ba/F3-

uc� cells), mice were killed, body and spleen weights wereecorded, and tissues were preserved in 10% formalin for his-opathologic analysis.

ResultsEffects of PKC412, Nilotinib, and Imatinibon Proliferation of D842V-PDGFRA-Ba/F3Cells and Tyrosine Phosphorylation inD842V-PDGFRA-Ba/F3 CellsD842V-PDGFRA-Ba/F3 cells were treated with a range

f concentrations of PKC412, nilotinib, or imatinib for 72

igure 3. Effects of combinations of PKC412, nilotinib, and imatinib onyrosine phosphorylation in D842V-Ba/F3 cells. PTYR immunoblot show-ng effects of nilotinib, imatinib, and PKC412—alone and combined—gainst D842V-Ba/F3 cells. Lane 1, control; lane 2, nilotinib alone; lane 3,KC412 alone; lane 4, imatinib alone; lane 5, nilotinib � PKC412; lane 6,ilotinib � imatinib; lane 7, imatinib � PKC412. D842V-Ba/F3 cells werereated for 4 hours with PKC412 (0.1 �mol/L), imatinib (0.5 �mol/L), orilotinib (0.5 �mol/L), alone or combined. The pTyr (pY99) antibody wassed for the pTyr immunoblot, and �-tubulin was used as a loading control.ensitometry for this immunoblot yielded the following values, reported asrea (pixels): D842V control pTyr/tubulin � 21983/19993; nilotinib aloneTyr/tubulin � 18897/20791; PKC412 alone pTyr/tubulin � 19030/4114; imatinib alone pTyr/tubulin � 19116/24576; nilotinib � PKC412Tyr/tubulin � 13103/25584; nilotinib � imatinib pTyr/tubulin � 14208/5865; imatinib � PKC412 pTyr/tubulin � 13571/27210.

ours. PKC412 exhibited an IC50 in the 0.05– 0.1 �mol/L range v

gainst D842V-PDGFRA-Ba/F3 cells (Figure 1A), whereas theC50 values for nilotinib and imatinib were in the 0.1– 0.5mol/L range (Figure 1B and C). Addition of IL-3, the normalrowth factor for Ba/F3 cells, resulted in near complete rescue,uggesting selective inhibition of D842V-PDGFRA with mini-

al nonspecific toxicity of each of these agents at the concen-rations tested.

The effects of PKC412, imatinib, and nilotinib on cellularyrosine kinase activity in D842V-PDGFRA-Ba/F3 cells weretudied. The robust signal observed for the pattern of phos-horylated bands in the D842V-PDGFRA-Ba/F3 cells wasodestly inhibited by 1-hour treatments with concentra-

ions of nilotinib and PKC412 that were effective in inhib-ting proliferation of cells (Figure 1D). Treatment of D842V-DGFRA-Ba/F3 cells with imatinib led to more pronounced,et still partial, inhibition of total cellular tyrosine phos-horylation levels with no observed decrease in D842V-PDG-RA expression (Figure 1D). The high level of cellular ty-osine phosphorylation and partial inhibition of this signaln the presence of the tyrosine kinase inhibitors are consis-ent with the aggressive, resilient phenotype of the D842V

utant.

Combination Effects of PKC412, Nilotinib,Imatinib: D842V-PDGFRA-Ba/F3Combinations of PKC412, nilotinib, and imatinib

ere tested against D842V-PDGFRA-Ba/F3 cells. Overall,ositive combination effects were observed between nilotiniblus imatinib, nilotinib plus PKC412, and imatinib plusKC412 (Figure 2). Calcusyn analysis of the combined effectsf nilotinib plus imatinib suggests synergistic to nearly ad-itive effects across a range of doses (ED25–ED75, whereD50 is the median effective dose), with antagonism at ED90

Table 1). The nilotinib plus PKC412 combination resultedn nearly additive to synergistic effects across a range ofoses (ED25–ED90) (Table 1), and the combined effects of

matinib plus PKC412 were nearly additive to synergisticffects across a range of doses (ED50 –ED90), with moderatentagonism observed at ED25 (Table 1).

Effects of Combinations of PKC412,Nilotinib, and Imatinib on TyrosinePhosphorylation in D842V-Ba/F3 CellsWe investigated the effects of combinations of nilo-

inib and imatinib, nilotinib and PKC412, and imatinib andKC412 on cellular tyrosine phosphorylation in D842V-a/F3 cells. We found that tyrosine phosphorylation, espe-ially of high-molecular-weight substrates, was modestly in-ibited by treatment for 4 hours with each drug as a singlegent at their approximate IC50 (as compared with vehicleontrol-treated cells) (Figure 3). Cellular tyrosine phosphor-lation was inhibited to a greater extent when combinationsf each of the 3 agents were used, as compared with singlegent- and vehicle control-treated cells (Figure 3). The resultsuggest that inhibition of cellular tyrosine phosphorylationy combined treatment of D842V-PDGFRA-expressing cellsorrelates with enhanced inhibition of cellular proliferation

ia combined treatment.

rPanwrPaT

Ptdptcamt

iB5aB

PufiFDfiwsw

4PbovPmvtt(snm

ih

CLIN

ICA

L–A

LIMEN

TARY

TRA

CT

1738 WEISBERG ET AL GASTROENTEROLOGY Vol. 131, No. 6

Effects of PKC412, Nilotinib, and Imatinibon Proliferation of V561D-PDGFRA-Ba/F3Cells and Tyrosine Phosphorylation inV561D-PDGFRA-Ba/F3 CellsV561D-PDGFRA-Ba/F3 cells were treated with a

ange of concentrations of PKC412, nilotinib, or imatinib.KC412 displayed an IC50 of approximately 0.025 �mol/Lgainst V561D-PDGFRA-Ba/F3 cells (Figure 4A). The IC50 ofilotinib or imatinib against V561D-PDGFRA-Ba/F3 cellsas approximately 0.01 �mol/L (Figure 4B and C). IL-3

escue was observed with all 3 inhibitors, and wild-typeDGFRA-expressing cells were unaffected by either inhibitort doses that inhibited mutant-PDGFRA-expressing cells.his suggests selective inhibition of mutant PDGFRA.

Treatment of V561D-PDGFRA-Ba/F3 cells for 1 hour withKC412, nilotinib, or imatinib led to a decrease in levels ofotal cellular tyrosine phosphorylation with no observedecrease in PDGFRA expression (Figure 4D). The increasedotency of the 3 inhibitors in decreasing levels of cellularyrosine phosphorylation in V561D-PDGFRA-Ba/F3 cells isonsistent with the increased potency of the inhibitorsgainst proliferation of Ba/F3 cells expressing the V561Dutant, as compared with cells expressing the D842V mu-

ant.

Combination Effects of Nilotinib WithImatinib and PKC412: V561DThe effects of combinations of nilotinib with either

matinib or PKC412 were evaluated against V561D-PDGFRA-a/F3 cells, and dose-response curves were generated (Figure). Generally, both combinations led to varying degrees ofntagonism across a range of doses in the V561D-PDGFRA-

a/F3 cell line (Table 1). w

In Vivo Analysis of Nilotinib and PKC412Against D842V-PDGFRA-Ba/F3 CellsTo assess directly the in vivo antitumor efficacy of

KC412 alone, as well as in combination with nilotinib, wetilized a mouse model in which tumor burden was quanti-ed by noninvasive imaging of luminescent D842V-PDG-RA-Ba/F3 cells (Figure 6). Murine Ba/F3 cells expressing the842V-PDGFRA mutant were engineered to stably expressrefly luciferase, and NCr nude mice were then inoculatedith these cells. Noninvasive imaging was used to assess

erially the tumor burden, and mice with established diseaseere divided into cohorts with similar tumor burden.

In 2 independent studies, mice were treated for a total ofor 6 days, respectively, with either vehicle (n � 3) or

KC412 at 100 mg/kg (n � 3). In these studies, tumorurden, as assessed by bioluminescence, was observed to beverall lower in mice treated with PKC412 as compared withehicle-treated mice (Figure 6A–D). In another study,KC412 (50 mg/kg) or nilotinib (150 mg/kg/day) was ad-inistered via oral gavage alone or in combination, as was

ehicle for a total of 6 days of treatment. At the end of thereatment period, tumor burden was lower in PKC412-reated mice but did not appear to be decreased by nilotinibFigure 6E). These results suggest efficacy of PKC412 as aingle agent against the D842V-PDGFRA mutant, whereasilotinib is a less effective inhibitor of the D842V-PDGFRAutant in vivo.

DiscussionThe vast majority of GISTs are characterized by activat-

ng mutations of c-kit or PDGFRA. The management of GISTsas been significantly changed by the development of imatinib,

Figure 4. Effects of PKC412,nilotinib, and imatinib on prolifera-tion of V561D-PDGFRA-Ba/F3cells. Cell viability is reported aspercentage of control (untreated)cells. Error bars represent thestandard error of the mean foreach data point. (A) Three-daytreatment of V561D-PDGFRA-Ba/F3 cells (�IL-3), PDGFRA-Ba/F3 cells, and parental Ba/F3cells with PKC412 (n � 2). (B)Three-day treatment of PDGFRA-V561D-Ba/F3 (�IL-3) and PDG-FRA-Ba/F3 cells with nilotinib (n �2). (C) Three-day treatmentof V561D-PDGFRA-Ba/F3 cells(�IL-3) and PDGFRA-Ba/F3 cellswith imatinib (n � 2). (D) Immuno-blot: Treatment of V561D-PDG-FRA-Ba/F3 cells for 1 hour withnilotinib, imatinib, and PKC412,respectively.

hich has proven to be an effective treatment for this malig-

noiRaPr

itfPoascpoptf

FPci

ctmaittD

ni2mccAaitPccng

Fct binat

CLI

NIC

AL–

ALI

MEN

TARY

TRA

CT

December 2006 MUTANT PDGFRA: PKC412, NILOTINIB, IMATINIB 1739

ancy by inducing clinical responses in patients and improvingverall survival. However, imatinib resistance in GIST is becom-

ng a significant challenge in the treatment of this disease.eports of resistance to imatinib after initial clinical responsere emerging, with secondary point mutations in either c-kit orDGFRA being identified as common mechanisms for acquiredesistance to imatinib.

The staurosporine derivative, PKC412, is a broad-spectrumnhibitor of Ser/Thr and Tyr protein kinases that has among itsargets protein kinase C (PKC), vascular endothelial growthactor receptor (VEGF-R2), FLT3, PDGFR, and c-kit.10,11

KC412 has previously been shown to be an effective inhibitorf the imatinib-resistant c-kit mutant D816V, which is associ-ted with systemic mast cell disease.14,15 PKC412 has also beenhown to be effective against the FIP1L1-PDGFRA kinase asso-iated with hypereosinophilic syndrome and chronic eosino-hilic leukemia16 and can override imatinib resistanceccurring in point-mutated FIP1L1-PDGFRA-induced myelo-roliferative disease.17 In addition, GIST patient samples posi-ive for the imatinib-resistant mutant c-kit or PDGFRA wereound to respond well to PKC412.8

Because most GISTs express gain-of-function c-kit or PDG-RA, it was of interest to investigate the effectiveness ofKC412 as a single agent in vitro against mutant PDGFRAonferring differential sensitivity toward imatinib. We were also

igure 5. Proliferation studies showing combined effects of nilotiniells. (A) Three-day treatments of V561D-PDGFRA-Ba/F3 cells with nreatments of V561D-PDGFRA-Ba/F3 with nilotinib, PKC412, or a com

nterested in investigating the in vitro effects of PKC412 in b

ombination with the Bcr-Abl inhibitors imatinib and nilotinib,he latter of which has not yet been investigated with respect to

utant PDGFRA-associated GIST. Additionally, we studied thectivity of nilotinib alone, and nilotinib in combination withmatinib, against differentially imatinib-sensitive PDGFRA mu-ants. Finally, we present here a novel in vivo imaging approacho studying the effects of tyrosine kinase inhibitors against the842V-PDGFRA mutant.Cooperative effects have been observed in vitro with combi-

ations of nilotinib and imatinib in c-kit and Bcr-Abl-express-ng cells, with differences in the cellular influx and efflux of the

agents being implicated as contributing to the underlyingechanism of the additive/synergistic effects.18 –20 It has re-

ently been shown that nilotinib and imatinib can interact withell transporters, such as the multidrug efflux transporterBCG2, which has been linked to chemoresistance,21 or Oct-1,n organic cation transporter found to be important for thenflux of imatinib but not for nilotinib uptake.20 Alternatively,he synergy observed between nilotinib plus imatinib orKC412, or imatinib plus PKC412, against D842V-Ba/F3 cellsould be attributed to the presence of signaling components/ellular proteins unique to D842V-PDGFRA expression (andot associated with V561D-PDGFRA expression) that are tar-eted by one or the other of the agents whose combination has

h imatinib and PKC412, respectively, against V561D-PDGFRA-Ba/F3ib, imatinib, or a combination of nilotinib plus imatinib. (B) Three-dayion of nilotinib plus PKC412.

b witilotin

een investigated in this study.

pcDcvsaVbPsaFlsF

titVlt�arsVpatc

FNB(foVV 3), n

CLIN

ICA

L–A

LIMEN

TARY

TRA

CT

1740 WEISBERG ET AL GASTROENTEROLOGY Vol. 131, No. 6

The potency of imatinib against D842V-PDGFRA-Ba/F3roliferation and cellular tyrosine phosphorylation was in ac-ordance with previous observations.8,22 The proliferation of842V-PDGFRA-Ba/F3 cells was inhibited by nilotinib at con-

entrations similar to those required for imatinib. Generally, initro D842V-PDGFRA-Ba/F3 cells were significantly less re-ponsive than V561D-PDGFRA-Ba/F3 cells toward imatinibnd nilotinib; the differential sensitivity of the D842V and561D mutants to these inhibitors is consistent with what haseen reported in the literature.7 However, the potency ofKC412 in inhibiting proliferation of the D842V mutant wasimilar to the potency of PKC412 against the V561D mutant. Inddition, treatment of mice injected with D842V-PDGFRA-Ba/3-luc� cells with a well-tolerated dose of PKC412 reduced

ower tumor burden in comparison with vehicle-treated mice,uggesting that PKC412 is efficacious against the D842V-PDG-

igure 6. In vivo analysis of PKC412 alone and PKC412 combined wCr nude mice treated for 4 days with vehicle vs PKC412 (100 mg/kg), or C) and number (0, 1, 3, 30) of each mouse. (B) Bioluminescence plo

shown as percentage baseline luminescence). Vehicle (n � 3) and PKCor 6 days with vehicle vs PKC412 (100 mg/kg). Nomenclature (C1, D0,f each mouse. (D) Bioluminescence plotted for (C) as a measure of tumoehicle (n � 3) and PKC412 (100 mg/kg) (n � 3). (E) Bioluminescence asehicle (n � 3), PKC412 (50 mg/kg) (n � 3), nilotinib (150 mg/kg) (n �

RA mutant in vivo. s

V561D-PDGFRA-Ba/F3 cells responded well to PKC412reatment in vitro, with high potency observed both in terms ofnhibition of cellular proliferation and tyrosine phosphoryla-ion. Nilotinib also displayed significant potency against the561D-PDGFRA mutant in terms of inhibition of cellular pro-

iferation and total cellular tyrosine phosphorylation. The ac-ivity of nilotinib as a single agent (IC50 approximately 0.00125mol/L) was similar to that of imatinib as a single agent (IC50

pproximately 0.00125 �mol/L); imatinib has previously beeneported to be a potent inhibitor of the V561D mutant.7 De-pite the impressive activity of each single agent against the561D-PDGFRA mutant, the combination of either nilotiniblus imatinib, or nilotinib plus PKC412, did not result indditive to synergistic effects across a range of doses. In con-rast, varying degrees of antagonism were observed between theompounds against the V561D-PDGFRA mutant. These results

ilotinib against D842V-PDGFRA-Ba/F3 cells. (A) Imaging data showingenclature (A0, B0, C30, C0, A3, and B1) corresponds to the cage (A,

for (A) as a measure of tumor burden in vehicle or PKC412-treated mice(100 mg/kg) (n � 3). (C) Imaging data showing NCr nude mice treatedD1, C0, C3) corresponds to the cage (C or D) and number (0, 1, 3, 33)

den in vehicle or PKC412-treated mice (shown as percentage baseline).asure of tumor burden in vehicle- or PKC412- or nilotinib-treated mice.

ilotinib � PKC412 (n � 3).

ith n. Nomtted412D33,r bura me

uggest that nilotinib, imatinib, or PKC412 could potentially be

uPcP

ticwiemtttDrdonaci

tptOaaniwtTpmtf

1

1

1

1

1

1

1

1

1

CLI

NIC

AL–

ALI

MEN

TARY

TRA

CT

December 2006 MUTANT PDGFRA: PKC412, NILOTINIB, IMATINIB 1741

sed as single agents against GISTs harboring the V561D-DGFRA mutant. However, the results do not support theombined use of nilotinib plus imatinib, or nilotinib plusKC412, against this mutation.

Because of the emergence of primary and secondary resis-ance of GIST to imatinib, there is a growing need for thedentification and development of additional inhibitors thatan be used therapeutically for imatinib-resistant GIST orhich can be safely and effectively used in combination with

matinib against GIST. In general, PKC412 appears to be anffective inhibitor of both the D842V- and V561D-PDGFRAutants as a single agent, whereas each mutation is differen-

ially sensitive to imatinib and nilotinib, respectively. In addi-ion, the combination of PKC412 with either nilotinib or ima-inib, respectively, in vitro resulted in greater inhibition of842V-PDGFRA-Ba/F3 cellular proliferation and cellular ty-

osine phosphorylation than any one agent alone, with varyingegrees of additive-synergistic effects between the compoundsbserved across a wide range of doses. The combinations ofilotinib, imatinib, and PKC412 could result in a cooperativentiproliferative effect from effects on multiple targets, espe-ially in the case of PKC412, which is a broad-spectrum inhib-tor.

The strategy of administering 2 different inhibitors could inheory lead to improved patient responsiveness, as well as sup-ression of the emergence of drug-resistant mutations charac-eristic of development of some forms of acquired resistance.ur data suggest that high doses of nilotinib would need to be

dministered to patients harboring the D842V mutant tochieve clinical responsiveness. Therefore, the benefits of usingilotinib in combination with another tyrosine kinase inhibitor

n D842V-PDGFRA-associated GIST patients would have to beeighed against the increased risk of adverse effects because of

he high dose of nilotinib required to cause tumor regression.he demonstrated efficacy of PKC412 as a single agent, and theositive enhancement of effects of PKC412 against the D842Vutant in the presence of Abl inhibitors, lends conviction to

he potential use of PKC412 alone or in combination therapyor GIST.

References

1. Hirota S, Isozaki K. Pathology of gastrointestinal stromal tumors.Pathol Int 2006;56:1–2.

2. Druker BJ, Tamura S, Buchdunger E, Ohno S, Segal GM, FanningS, Zimmermann J, Lydon NB. Effects of a selective inhibitor of theAbl tyrosine kinase on the growth of Bcr-Abl positive cells. NatMed 1996;2:561–566.

3. Heinrich MC, Corless CL, Demetri GD, Blanke CD, von Mehren M,Joensuu H, McGreevey LS, Chen CJ, Van den Abbeele AD, DrukerBJ, Kiese B, Eisenberg B, Roberts PJ, Singer S, Fletcher CD,Silberman S, Dimitrijevic S, Fletcher JA. Kinase mutations andimatinib response in patients with metastatic gastrointestinalstromal tumor. J Clin Oncol 2003;21:4342–4349.

4. Debiec-Rychter M, Sciot R, Le Cesne A, Schlemmer M, Hohen-berger P, van Oosterom AT, Blay J-Y, Leyvraz S, Stul M, Casali PG,Zalcberg J, Veerweij J, Van Glabbeke M, Hagemeijer A, Judson I.c-kit mutations and dose selection for imatinib in patients withadvanced gastrointestinal stromal tumors. Eur J Cancer 2006;42:1093–1103.

5. Heinrich MC, Rubin BP, Longley BJ, Fletcher JA. Biology andgenetic aspects of gastrointestinal stromal tumors: c-kit activa-

tion and cytogenetic alterations. Hum Pathol 2002;33:484–495.

6. Heinrich MC, Corless CL, Duensing A, McGreevey L, Chen CJ,Joseph N, Singer S, Griffith DJ, Haley A, Town A, Demetri GD,Fletcher CD, Fletcher JA. PDGFRA activating mutations in gastro-intestinal stromal tumors. Science 2003;299:708–710.

7. Corless CL, Schroeder A, Griffith D, Town A, McGreevey L, HarrellP, Shiraga S, Bainbridge T, Morich J, Heinrich MC. PDGFRA mu-tations in gastrointestinal stromal tumors: frequency, spectrumand in vitro sensitivity to imatinib. J Clin Oncol 2005;23:5357–5364.

8. Debiec-Rychter M, Cools J, Dumez H, Sciot R, Stul M, Mentens N,Vranckx H, Wasag B, Prenen H, Roesel J, Hagemeijer A, VanOosterom A, Marynen P. Mechanisms of resistance to imatinibmesylate in gastrointestinal stromal tumors and activity of thePKC412 inhibitor against imatinib-resistant mutants. Gastroen-terology 2005;128:270–279.

9. Yi ES, Strong CR, Piao Z, Perucho M, Weidner N. Epithelioidgastrointestinal stromal tumor with PDGFRA activating mutationand immunoreactivity. Appl Immunohistochem Molecul Morphol2005;13:157–161.

0. Fabbro D, Ruetz S, Bodis S, Pruschy M, Csermak K, Man A,Campochiaro P, Wood J, O’Reilly T, Meyer T. PKC412—a proteinkinase inhibitor with a broad therapeutic potential. AnticancerDrug Des 2000;15:17–28.

1. Weisberg E, Boulton C, Kelly LM, Manley P, Fabbro D, Meyer T,Gilliland DG, Griffin JD. Inhibition of mutant FLT3 receptors inleukemia cells by the small molecule tyrosine kinase inhibitorsPKC412. Cancer Cell 2002;1:433–443.

2. Weisberg E, Manley PW, Breitenstein W, Bruggen J, Cowan-JacobSW, Ray A, Huntly B, Fabbro D, Fendrich G, Hall-Meyers E, KungAL, Mestan J, Daley GQ, Callahan L, Catley L, Cavazza C, Azam M,Neuberg D, Wright RD, Gilliland DG, Griffin JD. Characterization ofAMN107, a selective inhibitor of native and mutant Bcr-Abl.Cancer Cell 2005;7:129–141.

3. Armstrong SA, Kung AL, Mabon ME, Silverman LB, Stam RW, DenBoer ML, Pieters R, Kersey JH, Sallan SE, Fletcher JA, Golub TR,Griffin JD, Korsmeyer SJ. Inhibition of FLT3 in MLL. Validation ofa therapeutic target identified by gene expression based classi-fication. Cancer Cell 2003;3:173–183.

4. Growney JD, Clark JJ, Adelsperger J, Stone R, Fabbro D, Griffin JD,Gilliland DG. Activation mutations of human c-KIT resistant toimatinib mesylate are sensitive to the tyrosine kinase inhibitorPKC412. Blood 2005;106:721–724.

5. Gleixner KV, Mayerhofer M, Aichberger KJ, Derdak S, Sonneck K,Bohm A, Gruze A, Samorapoompichit P, Manley PW, Fabbro D,Pickl WF, Sillaber C, Valent P. PKC412 inhibits in vitro growth ofneoplastic human mast cells expressing the D816V-mutatedvariant of c-kit: comparison with AMN107, imatinib, and cladrib-ine (2CdA) and evaluation of cooperative drug effects. Blood2006;107:752–759.

6. Cools J, Quentmeier H, Huntly BJ, Marynen P, Griffin JD, DrexlerHG, Gilliland DG. The EOL-1 cell line as an in vitro model for thestudy of FIP1L1-PDGFRA-positive chronic eosinophilic leukemia.Blood 2004;103:2802–2805.

7. Cools J, Stover EH, Boulton CL, Gotlib J, Legare RD, Amaral SM,Curley DP, Duclos N, Rowan R, Kutok JL, Lee BH, Williams IR,Coutre SE, Stone RM, DeAngelo DJ, Marynen P, Manley PW,Meyer T, Fabbro D, Neuberg D, Weisberg E, Griffin JD, GillilandDG. PKC412 overcomes resistance to imatinib in a murine modelof FIP1L1-PDGFR�-induced myeloproliferative disease. CancerCell 2003;3:459–469.

8. LeCoutre P, Baskaynak G, Tamm I, Westermann J, Duyster JV,Bonin M, Pursche S, Mayer P, Manley P, Mestan J, Schleyer E,Dorken B. Activity and induction of apoptosis of the specifictyrosine kinase inhibitor AMN107 in Bcr-Abl� cell lines and inimatinib resistant primary cells from CML patients (abstract No.

762). Blood 2004;104:218.

1

2

2

2

Mt(

S

S

CLIN

ICA

L–A

LIMEN

TARY

TRA

CT

1742 WEISBERG ET AL GASTROENTEROLOGY Vol. 131, No. 6

9. Prenen H, Guetens G, de Boeck G, Debiec-Rychter M, Manley P,Schoffski P, van Oosterom AT, de Bruijn E. Cellular uptake of thetyrosine kinase inhibitors imatinib and AMN107 in gastrointestinalstromal tumor cell lines. Pharmacology 2006;77:11–16.

0. White DL, Saunders VA, Dang P, Engler J, Zannettino AC, Cam-bareri AC, Quinn SR, Manley PW, Hughes TP. Oct-1 mediatedinflux is a key determinant of the intracellular uptake of imatinibbut not nilotinib (AMN107); reduced OCT-1 activity is the cause oflow in vitro sensitivity to imatinib. Blood 2006;108:697–704.

1. Jorgensen H, Allan E, Jordanides N, Hamilton A, Mountford T.AMN107 appears equipotent and may synergise with imatinib atthe CML stem cell level through interaction with ABCG2. Blood2005;106:1080.

2. Ohashi A, Kinoshita K, Isozaki K, Nishida T, Shinomura Y, Kitamura

Y, Hirota S. Different inhibitor effect of imatinib on phosphorylation p

of mitogen-activated protein kinase and Akt and on proliferation incells expressing different types of mutant platelet-derived growthfactor receptor-�. Int J Cancer 2004;111:317–321.

Received January 24, 2006. Accepted August 31, 2006.Address requests for reprints to: Ellen Weisberg, PhD, Department ofedical Oncology, Dana Farber Cancer Institute, 44 Binney Street, Bos-

on, Massachusetts 02115. e-mail: Ellen_Weisberg@dfci.harvard.edu; fax:617) 632-4388.

P.W.M. and D.F. are employees of Novartis Pharma AG, Basel,witzerland.J.D.G. has a financial interest with Novartis Pharma AG.Supported by NIH grants CA66996, CA36167, and DK50654 and a

pecialized Center of Research Award from the Leukemia and Lym-

homa Society (all to J.D.G.).