Activity of Dasatinib, a Dual SRC/ABL Kinase Inhibitor, and IPI-504, A

Cancer Therapy: Preclinical

Activity of Dasatinib, a Dual SRC/ABL Kinase Inhibitor, and IPI-504, aHeatShockProtein90Inhibitor,againstGastrointestinal Stromal Tumor ^ Associated PDGFRAD842V Mutation Barbara Dewaele,1BartoszWasag,1Jan Cools,1,4 Raf Sciot,2 Hans Prenen,3 PeterVandenberghe,1 AgnieszkaWozniak,3 Patrick Scho« ffski,3 Peter Marynen,1,4 and Maria Debiec-Rychter1

Abstract Purpose: Activating mutations in platelet-derived growthfactor receptor- a (PDGFRA) have been reported in f5% to 10% of patients withgastrointestinal stromal tumors (GIST). Imatinib efficiently inhibits the juxtamembrane PDGFRA mutations, whereas many tyrosine kinase domain activation loop PDGFRA mutations confer primary resistance to imatinib. In this study, we compared the efficacy of second-line tyrosine kinase inhibitors such as dasatinib, sorafenib, and nilotinib against two GIST-related PDGFRA mutants, PDGFRAD842V and PDGFRADDIM842-844. In addition, we sought to investigate the inhibitory effect of the heat shock protein 90 inhibitor, IPI-504, on these mutants. Experimental Design: Primary imatinib-resistant tumor cells and cell lines expressing imatinib- resistant PDGFRAD842V or imatinib-sensitive PDGFRADDIM842-844 mutants were treated with different concentrations of dasatinib, sorafenib, nilotinib, and IPI-504. The effect of treatment on proliferation, survival, and signaling was determined. Results: All inhibitors tested exhibited a high efficacy toward the PDGFRADDIM842-844 mutant. In contrast, ex vivo and in vitro assays revealed that only dasatinib potently inhibited the D842V PDGFRA isoform withan IC 50 value of 62 nmol/L. Sorafenib and nilotinib were significantly less efficacious against this mutation, inhibiting the PDGFRA kinase activity at >1, 0 0 0 a nd >5,000 nmol/L, and suppressing the proliferation of the cells expressing the PDGFRAD842V mutant withan IC 50 value of 239 and 1,310 nmol/L, respectively. IPI-504 treatment potently inhibited PDGFRA kinase activity by inducing the degradation of PDGFRAD842V and PDGFRADDIM842-844 at 256 and 182 nmol/L, respectively. Conclusions: Treatment with dasatinib or the heat shock protein 90 inhibitor IPI-504 may provide a therapeutic alternative for GIST patients whose tumors carry the imatinib-resistant PDGFRAD842V mutant isoform.

Gastrointestinal stromal tumors (GIST) comprise the largest tyrosine kinase homologous to KIT (5). Activating KIT and subset of mesenchymal tumors that develop along the PDGFRA mutations in GISTs differ in type and affect different gastrointestinal tract (1, 2). Constitutive activation of the KIT receptor domains (6). Therapeutic inhibition of KIT or tyrosine kinase through oncogenic mutations is an early and PDGFRA by the small-molecule inhibitor imatinib gives a probably initiating event in the majority of GISTs (3, 4). clinical response in the majority of advanced GISTs (3). Alternatively, a subset of GISTs express a constitutively activated However, early resistance has been reported in 10% to 20% platelet-derived growth factor receptor-a (PDGFRA) mutant, a of patients. It is well known that sensitivity to imatinib depends on the location of the KIT and PDGFRA gene mutation (3). Based on in vitro studies and subsequent clinical trials, it has been shown that a subset of GISTs harboring KIT and PDGFRA 1 2 3 Authors’Affiliations: Departments of Human Genetics, Pathology, and General mutations in the second tyrosine kinase domain (activation Medical Oncology, K.U. Leuven, and 4VIB Department of Molecular and Developmental Genetics,VIB, Leuven, Belgium loop domain) do not respond well to imatinib treatment Received 2/27/08; revised 4/28/08; accepted 5/22/08. (3, 7–10). The oncogenic PDGFRA mutations detected in Grant support: Life Raft Group, the Fonds voor Wetenschappelijk Onderzoek GISTs mainly involve exons 12, 14, or 18, and account for Vlaanderen (G.0286.05; M. Debiec-Rychter), and by a Concerted Action grant f5% to 10% of all GIST mutations. Wild-type PDGFRA is 2006/14 from the K.U. Leuven. sensitive to imatinib (11), but only a portion of the consti- The costs of publication of this article were defrayed in part by the payment of page advertisement tutively activated PDGFRA isoforms identified in GISTs are charges. This article must therefore be hereby marked in accordance D842V with18 U.S.C. Section 1734 solely to indicate this fact. potently inhibited by imatinib (3). The missense PDGFRA Requests for reprints: Maria Debiec-Rychter, Department of Human Genetics, mutation resulting in the substitution of aspartic acid to valine University of Leuven, O&N Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium. at codon 842 accounts for f60% of all PDGFRA mutations Phone: 32-16-347218; Fax: 32-16-346210; E-mail: Maria.Debiec-Rychter@ med.kuleuven.be. known in GISTs (5). It confers primary resistance to imatinib F 2008 American Association for Cancer Research. in vitro and in vivo (3, 5, 8, 12), resistance to sunitinib in vitro doi:10.1158/1078-0432.CCR-08-0533 (13), and resistance to nilotinib in vitro and in vivo (14).

www.aacrjournals.org 5749 Clin Cancer Res 2008;14(18) September 15, 2008 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Cancer Therapy: Preclinical

Conversely, we have previously shown that PKC412 efficiently oncoprotein degradation, but the effect of HSP90 inhibitors on inhibited the PDGFRAD842V mutant at the concentration of GISTs bearing PDGFRA mutations has thus far not been tested. 1 Amol/L (12). In this report, we examined the efficacy of second-generation Because resistance of GIST patients to imatinib is a tyrosine kinase inhibitors nilotinib, dasatinib, and sorafenib for D continuous clinical challenge, multiple novel therapeutic the inhibition of PDGFRAD842V-expressing and PDGFRA DIM842-844– strategies are under development. Novel small molecule expressing cells. In addition, we tested the potency of the HSP90 compounds being tested in clinical trials include nilotinib, inhibitor, IPI-504, to induce PDGFRA oncoprotein degradation sorafenib, and dasatinib. as a therapeutic alternative for imatinib-resistant, PDGFRA mutant In addition, heat shock protein 90 (HSP90) inhibition, GISTs. causing degradation of wild-type and imatinib-resistant KIT mutant proteins, has recently been validated in preclinical Materials and Methods studies as an alternative therapeutic strategy to challenge the heterogeneity of KIT imatinib-resistant mutations in patients Inhibitors. The inhibitors, imatinib mesylate (Glivec/Gleevec, with GIST (15). The PDGFRA protein is also subject to HSP90- Novartis), sorafenib (BAY-43-9006, Bayer AG), nilotinib (AMN107, mediated protection from proteasomal degradation in cancer Novartis), dasatinib (BMS-354825, Bristol-Meyers-Squibb), and IPI-504 cells (16) and HSP90 inhibition may cause effective PDGFRA were provided by Infinity Pharmaceuticals, Inc. and MedImmune Inc.

Fig. 1. In vitro effect of imatinib, nilotinib, sorafenib, and dasatinib on PDGFRADDIM842-844 ^expressingandPDGFRAD842V-expressing Ba/F3 cells. The dose-response proliferation curves of PDGFRADDIM842-844 ^ transduced (A)andPDGFRAD842V-transduced (B) Ba/F3 cells, treated withimatinib, nilotinib, sorafenib, and dasatinib for 48 h. Points, the average results of two experiments done in triplicate plotted with the curve-fitting Origin software; bars, SD.The calculated IC50 for eachinhibitor is indicated. Representative results from apoptotic assays of PDGFRADDIM842-844 ^ transduced (C)andPDGFRAD842V-transduced (D) Ba/F3 cells. All compounds induced significant apoptosisbetween100and500nmol/LincellsexpressingPDGFRADDIM842-844. In contrast, only dasatinib showed significant induction of apoptosis at 500 nmol/L in cells expressing PDGFRAD842V.

Clin Cancer Res 2008;14(18) September 15, 2008 5750 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Therapeutic Alternative for PDGFRAD842V Mutant

Table 1. Comparison of in vitro apoptosis induction of imatinib-sensitive PDGFRADDIM842-844 and imatinib- resistant PDGFRAD842V mutants expressed in the Ba/F3 cells

PDGFRA#DIM842-844 PDGFRAD842V Normal (%) Necrotic (%) Apoptotic (%) Normal (%) Necrotic (%) Apoptotic (%) Vehicle 85.1 9 5.7 84.2 7.6 7.4 Imatinib 0.1 Amol/L 71.8 15.1 12.9 85.9 6.8 6.6 0.5 Amol/L 60.9 20.8 18.1 87.7 4.7 7.3 Nilotinib 0.1 Amol/L 85 7.8 7.1 87.7 6.9 4.5 0.5 Amol/L 73.8 13.4 12.5 88.4 5.7 5.2 Sorafenib 0.1 Amol/L 74.1 14.1 11.5 87.9 6.6 5 0.5 Amol/L 66.5 16.7 16.8 89.4 5.9 4.5 Dasatinib 0.1 Amol/L 46.1 26.6 27 77.9 13.3 8.3 0.5 Amol/L 39.3 30.3 28.9 59.4 24.1 15.4 IPI-504 0.1 Amol/L 73.1 14.6 12.1 80.3 11.9 6.5 0.5 Amol/L 48.4 28.6 22.1 37.8 37.8 22.8

A 10 mmol/L stock solution of the inhibitors, dissolved in DMSO or a citric salt buffer for IPI-504) and incubated for 2 h at +37jC. After a citric salt buffer (for IPI-504) was stored at -80jC. The inhibitors were wash in ice-cold PBS, cells were lysed. Lysates were separated by SDS- diluted in culture medium just before use. PAGE electrophoresis and immunoblotted using anti–phospho- In vitro assays using transduced Ba/F3 cells. The retroviral expres- PDGFRA(Tyr754), anti-PDGFRA, and anti-actin antibodies. D sion constructs pMSCV-PDGFRAD842V and pMSCV-PDGFRA DIM842-844 The 5-bromo-2¶-deoxyuridine cell proliferation (Roche Applied were described previously (12). Retrovirally transduced Ba/F3 cells, Science) and sulforhodamine B survival (Sigma-Aldrich) assays were stably expressing the above PDGFRA constructs, were maintained in used to evaluate the proliferation rate and the survival of primary GIST culture at 1 Â 106 cells/mL. For the Ba/F3 dose-response assays, cells, according to the manufacturer’s instructions (18, 19). Both tests retrovirally transduced Ba/F3 cells were grown in RPMI 1640 supple- were run simultaneously and in duplicate. In short, the adherent mented with 10% fetal bovine serum, with or without interleukin 3, primary GIST cells were seeded in 96-well plates (10,000 cells per well) in the presence of vehicle alone or with varying concentrations of the and incubated overnight at +37jC. On the second day, the medium was inhibitors. The number of viable cells was determined at the start and replaced and different inhibitors (or vehicle alone) were added for 72 h. after 48 h, using the CellTiter 96 AQueous One Solution proliferation Cell growth inhibition curves were plotted using curve-fitting Origin assay (Promega). software. Induction of apoptosis on transduced Ba/F3 cells was evaluated by flow cytometry using Annexin-V-FLUOS Staining Kit according to the manufacturer’s protocol (Roche Applied Science) in duplicate Results experiments. Ba/F3 cells at a density of 1 Â 106 were cultured for 48 h DDIM842-844 in 24-well plates with the presence of vehicle alone or with doses of The imatinib-sensitive PDGFRA mutant Ba/F3 cells inhibitors mentioned above. The data and results analysis was are highly sensitive to all second-generation tyrosine kinase conducted with a BD FACSCanto System, with application of BD inhibitors tested. As illustrated in Fig. 1, all second-line FACSDiVa software (BD Biosciences). inhibitors tested are potent inhibitors of the growth of the D In addition, the response of the transduced Ba/F3 cells was assessed PDGFRA DIM842-844 –expressing Ba/F3 cells in a dose-depen- by Western immunoassays. Cells were exposed to a range of inhibitors dent manner. Imatinib, nilotinib, sorafenib, and dasatinib D in different concentrations, or to vehicle alone (DMSO, or citric salt inhibited the cell growth of PDGFRA DIM842-844 Ba/F3 cells buffer for IPI-504) and incubated for 2 h at +37jC. After a wash in with IC s of 20, 56, 17, and 10 nmol/L, respectively. ice-cold PBS, cells were lysed. Lysates were separated by SDS-PAGE 50 electrophoresis and immunoblotted using anti– phospho-PDGFRA Furthermore, significant induction of apoptosis of these cells (Tyr754) and anti-PDGFRA (Santa Cruz Biotechnology), and anti-actin was recorded at 100 nmol/L of imatinib, sorafenib, and (Sigma-Aldrich) antibodies, as previously described (12). dasatinib, and at 500 nmol/L of nilotinib (Fig. 1; Table 1). Ex vivo assay using primary GIST cells. A surgical specimen of The Western immunoassays were in line with the dose- primary GIST from an imatinib-naBve patient was partially snap-frozen response experiments (Fig. 2). All four tyrosine kinase at -80jC after surgery and partially used for the primary cell cultures, as inhibitors tested significantly inhibited the phosphorylation D842V D previously described (12). The PDGFRA mutation was identified of the PDGFRA DIM842-844 Ba/F3 cells at low nanomolar by sequencing of the DNA isolated from the frozen specimen, according concentrations. to standard procedures (17). These data indicate that all four second-generation tyrosine For Western immunoblotting, primary GIST cells obtained from kinase inhibitors tested have a high activity towards the collagenase-disaggregated tumor specimens were seeded in duplicate at DDIM842-844 in vitro 80% confluence in 25 mm diameter cell culture dishes (Corning, Inc.) PDGFRA –expressing Ba/F3 cells . Notably, in vitro and grown in DMEM supplemented with 10% fetal bovine serum, dasatinib is even slightly more potent than imatinib . D842V 1.0 mmol/L of nonessential amino acids, and 1.0 mmol/L of sodium The imatinib-resistant PDGFRA -expressing Ba/F3 cells are D842V pyruvate for 24 h at +37jC. Next, the cells were exposed to a range of responsive to dasatinib in vitro. As expected, PDGFRA - inhibitors in different concentrations, or to vehicle alone (DMSO, or expressing Ba/F3 cells were resistant to imatinib and nilotinib,

www.aacrjournals.org 5751 Clin Cancer Res 2008;14(18) September 15, 2008 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Cancer Therapy: Preclinical

with IC50s of 642 and 1,310 nmol/L, respectively, based on (Fig. 1; Table 1). Moreover, dasatinib completely inhibited growth inhibition measurements (Fig. 1). The efficacy of the PDGFRAD842V phosphorylation at 500 nmol/L in Western sorafenib for PDGFRAD842V inhibition was better than imatinib immunoassay (Fig. 2). In line with proliferation and apoptosis (IC50 of 239 nmol/L), but dasatinib was the most potent experiments, imatinib, nilotinib, and sorafenib were much less D842V inhibitor of the PDGFRA -expressing Ba/F3, with an IC50 of efficient in reducing the tyrosine phosphorylation of the 62 nmol/L. An apoptosis assay confirmed the significant PDGFRAD842V protein in Ba/F3 because micromolar concen- induction of apoptosis in PDGFRAD842V Ba/F3 transformants trations were needed to inhibit PDGFRA kinase activity. The 48 h after treatment with 500 nmol/L of dasatinib effect on tyrosine phosphorylation of the PDGFRAD842V

Fig. 2. Western blot analyses of, respectively, PDGFRADDIM842-844 ^ transduced and PDGFRAD842V-transduced Ba/F3 cells after treatment withimatinib ( A and E), nilotinib (B and F), sorafenib (C and G), and dasatinib (D and H)for 90 min. Total cell lysates were analyzed with anti-PDGFRA or anti ^ phospho-PDGFRA(Tyr754) antibodies.

Clin Cancer Res 2008;14(18) September 15, 2008 5752 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Therapeutic Alternative for PDGFRAD842V Mutant

Fig. 3. Ex vivo effect of imatinib, sorafenib, anddasatinibonPDGFRAD842V mutant primary GISTcells. 5-Bromo-2¶-deoxyuridine proliferation and sulforhodamine B survival assay of PDGFRAD842V mutant primary GISTcells, treated withdasatinib and sorafenib (A). The proliferation rate and the survival of the cells were determined after 72 h. Points, average results of two experiments done in quadruplicate is plotted withthecurve-fitting Origin software; bars, SD. The calculated IC 50 for eachinhibitor is indicated.Western blot analysis of PDGFRAD842V mutant primary GISTcells after treatment withdasatinib ( B), sorafenib (C), or imatinib (D) for 2 h. Immunoblots were probed withanti-PDGFRA or anti ^ phospho-PDGFRA(Tyr754)antibodies.

tyrosine kinase again correlated nicely with the dose-dependent sulforhodamine B assays (Fig. 3A). The proliferation of the inhibition of cell growth. These data show that dasatinib is a PDGFRAD842V mutant GIST primary tumor cells was almost potent inhibitor of PDGFRAD842V activity in Ba/F3 cells. completely inhibited at a 1,000 nmol/L concentration of D842V The imatinib-resistant PDGFRA mutant GIST cells are dasatinib, with an IC50 of 47 nmol/L. Moreover, f60% responsive to dasatinib ex vivo. As we found that dasatinib of PDGFRAD842V-expressing primary tumor cells died at D842V was a potent inhibitor of PDGFRA -expressing Ba/F3 1,000 nmol/L of dasatinib (IC50 of 513 nmol/L for cell cells in vitro, we investigated whether it had comparable survival). efficacy towards human GIST tumor cells carrying the same Sorafenib was almost 10 times less effective than dasatinib PDGFRAD842V mutation ex vivo, using cultures of primary in inhibiting the proliferation of PDGFRAD842V mutant GIST D842V tumor cells from an imatinib-naBve patient with PDGFRA primary tumor cells, with an IC50 of 425 nmol/L. After mutant GIST. We measured the effect of dasatinib and 72 hours of treatment, no more than 50% of the PDGFRAD842V- sorafenib on the proliferation rate and on the survival expressing primary cells underwent apoptosis by sorafenib in of primary GIST cells using 5-bromo-2¶-deoxyuridine and doses of up to 5 Amol/L.

www.aacrjournals.org 5753 Clin Cancer Res 2008;14(18) September 15, 2008 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Cancer Therapy: Preclinical

Western immunoblotting on PDGFRAD842V mutant primary activation of mitogen-activated protein kinase (MAPK) was GIST cells showed that 500 nmol/L of dasatinib efficiently clearly visible in both mutant forms. Interestingly, treatment inhibited the tyrosine kinase activity of PDGFRAD842V of PDGFRAD842V mutant Ba/F3 cells with imatinib led to (Fig. 3B). Sorafenib also inhibited the phosphorylation of the increased phosphorylation of MAPK. Conversely, a decrease of PDGFRAD842V protein in a dose-dependent manner, but at MAPKactivation was observed in PDGFRA deletion mutants 10-fold higher concentrations compared with dasatinib to after imatinib treatment, which is consistent with the imatinib obtain the same level of inhibition (Fig. 3C). In contrast, phenotype–sensitivity of this mutant. Similarly, treatment of D imatinib was not able to significantly reduce the ex vivo tyrosine PDGFRA DIM842-844 and PDGFRAD842V mutants with IPI-504 kinase activity of the PDGFRAD842V primary GIST cells at doses led to total and partial inhibition of MAPKphosphorylation of up to 5 Amol/L (Fig. 3D). levels at 0.5 and 1.0 Amol/L, respectively, being in line with the D PDGFRA DIM842-844 –expressing and PDGFRAD842V-expressing suppressive effect of the drug on the PDGFRA activation levels. Ba/F3 cells are sensitive to the HSP90 inhibitor IPI-504. As The dynamics of IPI-504–mediated PDGFRA inhibition illustrated by the proliferation curves, IPI-504 efficiently were determined in a time course experiment in which both D inhibited the growth of both PDGFRA DIM842-844 –expressing PDGFRA mutant Ba/F3 cell lines were treated with 500 nmol/L D842V and PDGFRA -expressing Ba/F3 cells with IC50 values of IPI-504 for different time intervals or with 1 Amol/L of of 182 and 256 nmol/L, respectively (Fig. 4A and B). Further- imatinib for 2 hours. Under IPI-504 inhibition of PDGFRAD842V, more, IPI-504 induced pronounced apoptosis of both phosphorylation was observed within 2 hours, followed by D PDGFRA DIM842-844 and PDGFRAD842V mutant Ba/F3 cells at protein degradation 6 hours or later (Fig. 6A). Similarly, D 500 nmol/L (Fig. 4C and D; Table 1). PDGFRA DIM842-844 was partially and completely inactivated Both PDGFRA mutant Ba/F3 cell lines were treated with after 2 and 6 hours of IPI-504 treatment, respectively (Fig. 6B). increasing concentrations of IPI-504 for 6 hours and analyzed As a reference, the PDGFRAD842V mutant in Ba/F3 cells was by Western blotting (Fig. 5). For comparison, both cell lines resistant to treatment with 1 Amol/L of imatinib after 2 hours D were also treated with 1 Amol/L of imatinib for 2 hours. For of treatment, whereas the PDGFRA DIM842-844 mutant was both mutants, the PDGFRA tyrosine kinase activity was clearly sensitive, confirming the differences in sensitivity to significantly inhibited by 500 nmol/L of IPI-504. In a similar imatinib of these two mutants. fashion, the total PDGFRA expression levels were significantly Imatinib-resistant PDGFRAD842V mutant GIST is responsive lowered at this concentration. Notably, PDGFRA inactivation to IPI-504 ex vivo. To test the sensitivity of primary GIST preceded total protein degradation, with the mature (glycosy- cells from a PDGFRAD842V-expressing GIST tumor to the lated) form of PDGFRA more substantially degraded in the inhibitor IPI-504, the cells were cultured and treated with D PDGFRA DIM842-844 deletion mutant. In addition, the signaling different concentrations of IPI-504. The proliferation rate of pathways of both mutants were investigated. There was a lack the PDGFRAD842V mutant GIST primary tumor cells was or only minor activation of the AKT signaling pathway in almost completely inhibited at 100 nmol/L of IPI-504, with DDIM842-844 D842V PDGFRA and PDGFRA , respectively, whereas an IC50 of 21 nmol/L (Fig. 7A). Moreover, after 72 hours,

Fig. 4. In vitro effect of IPI-504 on PDGFRADDIM842-844 êxpressingand PDGFRAD842V-expressing Ba/F3 cells. The proliferation curves of PDGFRADDIM842-844 êxpressing(A) and PDGFRAD842V-expressing (B) Ba/F3 cells treated withIPI-504 for 48 h.Points, average results of two experiments done in triplicate is plotted withthecurve-fitting Origin software; bars, SD.The calculated IC50 for eachPDGFRA mutant cell line is indicated. Induction of apoptosis in PDGFRADDIM842-844 êxpressing(C)and PDGFRAD842V-expressing (D) Ba/F3 cells treated with IPI-504 for 48 h, showing significant induction of apoptosis starting at 100 and at 500 nmol/L, respectively.

Clin Cancer Res 2008;14(18) September 15, 2008 5754 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Therapeutic Alternative for PDGFRAD842V Mutant

Fig. 5. Western blot analysis of PDGFRADDIM842-844 ^expressing(A)and PDGFRAD842V-expressing (B) Ba/F3 cells after treatment withdifferent concentrations of IPI-504 for 6 hor with1 Amol/L of imatinib for 2 h(control). Membranes were probed with anti-PDGFRA, anti ^ phospho- PDGFRA(Tyr754), anti ^ phospho-AKT, anti-AKT, anti ^ phospho-MAPK, anti-MAPK, or anti-actin antibodies.

PDGFRAD842V-expressing primary tumor cells died dose- challenge. Therefore, it is important to validate novel thera- responsively with an IC50 of 21 nmol/L for cell survival, and peutic strategies for the specific KIT and PDGFRA genotypes. with f50% of PDGFRAD842V primary cells killed at 5 Amol/L Second-generation tyrosine kinase inhibitors, nilotinib and of IPI-504 (Fig. 7B). dasatinib, inhibit the kinase activity of the wild-type PDGFRA Western immunoblotting of PDGFRAD842V mutant primary (7, 14, 20, 21). The activity of sorafenib in this regard is GIST cells (Fig. 7C) showed that IPI-504 completely inhibited unknown. Given the lack of in vivo models for PDGFRA the tyrosine kinase activity of PDGFRAD842V at 0.5 Amol/L. mutant isoforms, in vitro experiments using transformed Ba/F3 The PDGFRAD842V protein (mature and immature forms) mutant cells as well as primary tumor GIST cells are the expression levels were both significantly lowered at IPI-504 approach of choice for efficacy testing. In this report, we concentrations of 1 Amol/L. Both, AKT and MAPK were compared the efficacy of nilotinib, sorafenib, and dasatinib activated in PDGFRAD842V mutant primary GIST cells, albeit toward two GIST-related PDGFRA mutants, PDGFRAD842V and D AKT to a lesser degree. Treatment of these cells with IPI-504 PDGFRA DIM842-844, which confer differential sensitivity to resulted in a clear inhibition of both signaling pathways at imatinib (12). 0.5 Amol/L, and parallel to PDGFRA tyrosine kinase activity In line with previous reports, we have found that D inhibition. These results indicate that IPI-504 is also a potent PDGFRA DIM842-844 –expressing Ba/F3 cells were efficiently D842V inhibitor of PDGFRA activity in human primary GIST inhibited by imatinib (with an IC50 of 20 nmol/L). Corre- cells ex vivo. spondingly, nilotinib, sorafenib, and dasatinib also effectively D inhibited the cell growth of PDGFRA DIM842-844 –transduced Discussion Ba/F3 cell lines, with IC50s of 56, 17, and 10 nmol/L, respectively. In contrast, the PDGFRAD842V mutant showed Although most patients with GIST initially respond to varying degrees of sensitivity to these compounds. Thus, therapy, 10% to 20% of patients still exhibit primary resistance nilotinib had a minor effect on the PDGFRAD842V-transduced to imatinib. Primary resistant tumors either carry no mutations Ba/F3 cells, displaying an IC50 of 1.31 Amol/L, and leading to in KIT or PDGFRA, or harbor specific activation loop mutations only a modest decrease in levels of PDGFRA phosphorylation in KIT or PDGFRA, including the PDGFRAD842V mutation at high 5.0 Amol/L concentrations. Similar findings were (3, 5, 8). The D842V mutation is the most common activating reported by Weisberg and coworkers (22), who showed similar D842V PDGFRA mutation in GIST, which has also been identified as IC50 values of PDGFRA for both nilotinib and imatinib a mechanism of secondary resistance to imatinib (12). in the 0.5 to 1.0 Amol/L range. The KIT and PDGFRA missense The development of primary and secondary resistance of mutations in the second kinase domain keep the activation GIST toward imatinib is emerging as a significant clinical loop of the tyrosine kinases in an ‘‘open’’ or active conformation

www.aacrjournals.org 5755 Clin Cancer Res 2008;14(18) September 15, 2008 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Cancer Therapy: Preclinical

(23). Imatinib can only bind to the inactive conformation of systemic mastocytosis models (24). Furthermore, at 0.5 Amol/L KIT and PDGFRA (24), explaining the lack of activity against of dasatinib, numerous PDGFRAD842V –expressing BaF/3 cells these mutants. Nilotinib is a close relative of imatinib, binding entered programmed cell death. In addition, we have shown selectively to the inactive conformation of BCR-ABL, KIT, and ex vivo that 72 hours of treatment with 1 Amol/L dasatinib kills PDGFRs, but with 20-fold improved affinity for mutant and f70% of PDGFRAD842V-expressing primary GIST tumor cells wild-type BCR-ABL (22). Although the binding affinity of and significantly inhibits their proliferation. Pharmacokinetic nilotinib for PDGFR and KIT is very similar to that of imatinib studies done during a phase I dose-escalation study of dasatinib (21), in a phase I study, this drug has been reported to possess have shown that high nanomolar concentrations of the activity both as a single agent and in combination with compound can be safely achieved in humans (24). Given the imatinib in GIST patients resistant to prior tyrosine kinase fact that dasatinib is well tolerated, and that PDGFRAD842V is inhibitor treatment (25). inhibited by 500 nmol/L of dasatinib, treatment with dasatinib Dasatinib (BMS-354825) is an orally active, multitarget is a potential alternative for the patients with this mutation. inhibitor of BCR-ABL and SRC family kinases with a log2 Sorafenib is an orally active multikinase inhibitor which increased potency relative to imatinib (26). This drug was targets tumor cell growth and angiogenesis by potently developed for patients with chronic myelogenous leukemia, inhibiting, respectively, the RAF family of serine/threonine who were either intolerant to or who developed resistance to kinases and the tyrosine kinase receptors VEGFR-2 (KDR) and imatinib. Dasatinib is also a potent inhibitor of many other VEGFR-3 (FLT-4). Sorafenib also has selectivity for the tyrosine protein kinases, including KIT and PDGFRs. Studies of kinases PDGFRB, KIT, FLT-3, and RET (29, 30). Sorafenib is dasatinib in GIST and solid tumors are ongoing and preli- approved for the treatment of advanced renal cell carcinoma by minary data have suggested some antitumor activity (27). It was the Food and Drug Administration and European Medicines recently shown that dasatinib inhibits the kinase activity of the Agency (31). Recently, the efficacy of sorafenib in the treatment imatinib-resistant KIT D816V mutation with an IC50 between of patients with hepatocellular carcinoma was shown (25). In 50 and 100 nmol/L (24). The PDGFRAD842V mutation patients, a dose of 100 mg/d of sorafenib could result in steady- investigated in this study is analogous to the D816V mutation state serum drug concentrations of up to 4 Amol/L (32–34). in KIT. Therefore, we hypothesized that dasatinib might also In our study, although sorafenib was only slightly less potently inhibit the kinase activity of the PDGFRAD842V potent than dasatinib in inhibiting the kinase activity of D mutant. We have found that dasatinib completely inhibits the PDGFRA DIM842-844, it was markedly less potent than dasatinib kinase activity of the PDGFRAD842Vmutant isoform at nano- in inhibiting the imatinib-resistant PDGFRAD842V mutation. molar concentrations in vitro, albeit at somewhat higher Hence, a dose-dependent decrease of proliferation with an concentrations than needed for the inhibition of wild-type IC50 of 425 nmol/L, and incomplete inhibition of the PDGFRA, as evidenced by experiments done on A10 (rat PDGFRA tyrosine kinase activity at a concentration 1 Amol/L VSMC cells) and human AoSMC cell lines, in which PDGF- was observed. In line with this observation, it was recently stimulated wild-type PDGFR tyrosine phosphorylation was shown that sorafenib inhibits imatinib-resistant PDGFRB completely blocked by 50 nmol/L of BMS-354825 (28). Using gatekeeper mutant T681I (IC50 of 110 nmol/L), but is less D842V PDGFRA mutant primary GIST cells, we have shown that active (IC50 of 1.17 Amol/L) against the activation loop mutant dasatinib nearly totally inhibits the phosphorylation of of PDGFRBD850V, which is analogous with the PDGFRAD842V PDGFRAD842V at 0.5 Amol/L ex vivo. This is in the same range investigated in the present study (35). Moreover, the KITD816V D816V D842V as the IC50 found for the analogous KIT mutant in isoform, analogous with the PDGFRA isoform, is resistant

Fig. 6. Time course study of PDGFRAD842V-expressing (A)and PDGFRADDIM842-844 ^expressing(B) Ba/F3 cells treated with500 nmol/L of IPI-504 for different time intervals or with 1 Amol/L of imatinib for 2 h. Total cell lysates were analyzed withanti-PDGFRA, anti ^ phospho-PDGFRA(Tyr754), and anti-actin antibodies. Eachblot is representative of at least three independent experiments.

Clin Cancer Res 2008;14(18) September 15, 2008 5756 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Therapeutic Alternative for PDGFRAD842V Mutant

Fig. 7. Ex vivo effect of IPI-504 on PDGFRAD842V mutant primary GISTcells. 5-Bromo-2¶-deoxyuridine proliferation (A) and sulforhodamine B survival (B) assay of PDGFRAD842V mutant primary GISTcells, treated with IPI-504. The proliferation rate and the survival of the cells were determined after 72 h.Western blot analysis of PDGFRAD842V mutant primary GISTcells after treatment withdifferent concentrations of IPI-504 for 2 h( C). Blots were probed with anti-PDGFRA, anti ^ phospho-PDGFRA (Tyr754), anti ^ phospho-AKT, anti-AKT, anti ^ phospho- MAPK, anti-MAPK, or anti-actin antibodies.

to sorafenib (36). These differences may reflect a preference for degradation of HSP90-dependent PDGFRA proteins in trans- sorafenib to bind to the inactive form of the PDGFR as has been formed cells (16). We could expect that PDGFRA containing shown for binding of sorafenib to RAF kinase (37). activating mutations would even become more dependent on In the second part of this study, we examined the potency of HSP90 chaperoning. IPI-504, a HSP90 inhibitor, towards PDGFRA oncoproteins, as IPI-504 is a novel, highly soluble analogue of 17-AAG. It is an an alternative treatment option for imatinib-resistant PDGFRA active metabolite of 17-AAG, and it is slightly more potent than mutant GISTs. HSP90 is an emerging therapeutic target of 17-AAG (41). IPI-504 preferentially targets and accumulates interest for the treatment of cancer. HSP90 is a chaperone in tumor tissues. IPI-504 was recently evaluated in a murine molecule that maintains the structure and activity of certain model of chronic myelogenous leukemia and might be a new key signaling proteins within cancer cells. 17-Allylamino-18- therapeutic agent for the treatment of imatinib-resistant BCR- demethoxy-geldanamycin (17-AAG) is a geldanamycin deriva- ABL–induced leukemia (42). Multiple phase I and phase II tive that binds a conserved pocket in the HSP90 NH2-terminal clinical trials with IPI-504 are ongoing. Preliminary data from domain and prevents HSP90 from stabilizing clients (38). an ongoing dose escalation phase I clinical trial in patients with Concentrations of 1 Amol/L of 17-AAG can be achieved in vivo imatinib-resistant metastatic GISTs showed stable disease in in the blood as shown by preliminary data from phase I clinical 76% of patients (43). trials (39). A study by Fumo et al. (40) showed that 17-AAG led In the present study, we showed the potency of IPI-504 to to rapid degradation of the imatinib-resistant KITD816V mutant, inactivate and degrade imatinib-resistant PDGFRAD842V and D with a more profound effect at concentrations of >500 nmol/L. imatinib-sensitive PDGFRA DIM842-844 oncoproteins in Ba/F3 It was recently shown that 17-AAG can selectively promote the cell lines within 6 hours at a concentration of 500 nmol/L.

www.aacrjournals.org 5757 Clin Cancer Res 2008;14(18) September 15, 2008 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Cancer Therapy: Preclinical

Moreover, IPI-504 induced a pronounced apoptosis of the the degradation of constitutively activated oncoproteins irre- tested cells after 48 hours of treatment. Secondly, we provided spective of the specific mutational activation mechanisms. ex vivo evidence of significant activity of IPI-504 against We can thus conclude that treatment with the second- PDGFRAD842V primary GIST cells in the high nanomolar range. generation tyrosine kinase inhibitor dasatinib or with the Finally, we have shown that 500 nmol/L of IPI-504 inhibited HSP90 inhibitor IPI-504 may provide a therapeutic alternative imatinib-sensitive and imatinib-resistant PDGFRA oncopro- option for patients with GIST whose tumors carry the imatinib- teins, with totally diminished PDGFRA phosphorylation and resistant PDGFRAD842V mutant isoform. Clinical use of these significantly reduced total PDGFRA expression after 2 and drugs as first-line treatment should be considered in patients 6 hours, respectively. Given the possible broad heterogeneity of with the primary imatinib-resistant PDGFRA mutant GISTs. secondary KIT or PDGFRA resistance mutations during imatinib treatment in a given patient, the HSP90 inhibitor IPI-504 Disclosure of Potential Conflicts of Interest might be more effective in the clinic compared with any other single second-line tyrosine kinase inhibitor because it induces M. Debiec-Rychter is a member of the speakers’ bureau.

References 1. Fletcher CD, Berman JJ, Corless C, et al. Diagnosis of 16. Matei D, Satpathy M, Cao L, Lai YC, Nakshatri H, exhibits broad spectrum oral antitumor activity and gastrointestinal stromal tumors: A consensus ap- Donner DB. The platelet-derived growth factor recep- targets the RAF/MEK/ERK pathway and receptor ty- proach. Hum Pathol 2002;33:459 ^ 65. tor a is destabilized by geldanamycins in cancer cells. rosine kinases involved in tumor progression and an- 2. KindblomLG,RemottiHE,AldenborgF,Meis- J Biol Chem 2007;282:445 ^ 53. giogenesis. Cancer Res 2004;64:7099 ^ 109. Kindblom JM. Gastrointestinal pacemaker cell tumor 17. Debiec-Rychter M,Wasag B, Stul M, et al. Gastroin- 31. Anonymous.Treatment for advanced kidney cancer. (GIPACT): gastrointestinal stromal tumors show phe- testinal stromal tumours (GISTs) negative for KIT FDA Consum 2006;40:3. notypic characteristics of the interstitial cells of Cajal. (CD117 antigen) immunoreactivity. J Pathol 2004; 32. Strumberg D. Preclinical and clinical development of Am J Pathol 1998;152:1259 ^ 69. 202:430 ^ 8. the oral multikinase inhibitor sorafenib in cancer treat- 3. HeinrichMC, Corless CL, Demetri GD, et al. Kinase 18. Skehan P, Storeng R, Scudiero D, et al. New colori- ment. DrugsToday (Barc) 2005;41:773 ^ 84. mutations and imatinib response in patients with metric cytotoxicity assay for anticancer-drug screen- 33. Strumberg D, Clark JW, Awada A, et al. Safety, metastatic gastrointestinal stromal tumor. J Clin Oncol ing. J Natl Cancer Inst 1990;82:1107 ^ 12. pharmacokinetics, and preliminary antitumor activity 2003;21:4342^9. 19. Monks A, Scudiero D, Skehan P, et al. Feasibility of of sorafenib: a review of four phase I trials in patients 4. Hirota S, Isozaki K, Moriyama Y, et al. Gain-of- a high-flux anticancer drug screen using a diverse withadvanced refractory solid tumors. Oncologist function mutations of c-kit in human gastrointestinal panel of cultured human tumor cell lines. JNatl Cancer 2007;12:426 ^ 37. stromal tumors. Science 1998;279:577^ 80. Inst 1991;83:757 ^ 66. 34. Awada A, Hendlisz A, Gil T, et al. Phase I safety 5. HeinrichMC,CorlessCL,DuensingA,etal.PDGFRA 20. Lombardo LJ, Lee, FY, Chen P, et al. Discovery of and pharmacokinetics of BAY 43-9006 adminis- activating mutations in gastrointestinal stromal N-(2-chloro-6-methyl-phenyl)-2-(6-(4-(2-hydrox- tered for 21 days on/7 days off in patients withad- tumors. Science 2003;299:708 ^ 10. yethyl)-piperazin-1-yl)-2-methylpyrimidin-4-ylami- vanced, refractory solid tumours. Br J Cancer 2005; 6. Corless CL, Fletcher JA, Heinrich MC. Biology of no)thiazole-5-carboxamide (BMS-354825), a dual 92:1855 ^ 61. gastrointestinal stromal tumors. J Clin Oncol 2004; Src/Abl kinase inhibitor with potent antitumor activity 35. Guida T, Anaganti S, Provitera L, et al. Sorafenib 22:3813^ 25. in preclinical assays. JMed Chem 2004;47:6658 ^ 61. inhibits imatinib-resistant KIT and platelet-derived 7. Chen H, Isozaki K, Kinoshita K, et al. Imatinib inhibits 21. O’Hare T, Walters DK, Deininger MW, Druker BJ. growthfactor receptor h gatekeeper mutants. Clin various types of activating mutant kit found in gastro- AMN107: tightening the grip of imatinib. Cancer Cell Cancer Res 2007;13:3363 ^ 9. intestinal stromal tumors. Int J Cancer 2003;105: 2005;7:117^ 9. 36. Lierman E, Lahortiga I, Van MH, et al. The ability 13 0 ^ 5. 22. Weisberg E, Manley P, Mestan J, Cowan-Jacob S, of sorafenib to inhibit oncogenic PDGFRh and 8. Corless CL, Schroeder A, Griffith D, et al. PDGFRA Ray A, Griffin JD. AMN107 (nilotinib): a novel and se- FLT3 mutants and overcome resistance to other mutations in gastrointestinal stromal tumors: frequen- lective inhibitor of BCR-ABL. Br J Cancer 2006;94: small molecule inhibitors. Haematologica 2007;92: cy, spectrum and in vitro sensitivity to imatinib. J Clin 17 6 5 ^ 9 . 27 ^ 34. Oncol 2005;23:5357 ^ 64. 23. Foster R, GriffithR, Ferrao P, Ashman L. Molecular 37. Wan PT, Garnett MJ, Roe SM, et al. Mechanism 9. HirotaS,OhashiA,NishidaT,etal.Gain-of-function basis of the constitutive activity and STI571resistance of activation of the RAF-ERK signaling pathway by mutations of platelet-derived growthfactor receptor a of Asp816Val mutant KIT receptor tyrosine kinase. oncogenic mutations of B-RAF. Cell 2004;116: gene in gastrointestinal stromal tumors. Gastroenter- J Mol GraphModel 2004;23:139 ^ 52. 855^67. ology 2003;125:660 ^ 7. 24. Shah NP, Lee FY,Luo R, JiangY,Donker M, Akin C. 38. Neckers L, Lee YS. Cancer: the rules of attraction. 10. Longley BJ, Reguera MJ, MaY.Classes of c-KIT ac- Dasatinib (BMS-354825) inhibits KITD816V, an ima- Nature 2003;425:357 ^ 9. tivating mutations: proposed mechanisms of action tinib-resistant activating mutation that triggers neo- 39. Solit DB, Ivy SP, Kopil C, et al. Phase I trial of 17- and implications for disease classification and therapy. plastic growthin most patients withsystemic allylamino-17-demethoxygeldanamycin in patients Leuk Res 2001;25:571 ^ 6. mastocytosis. Blood 2006;108:286 ^ 91. withadvanced cancer. Clin Cancer Res 2007;13: 11. Buchdunger E, Cioffi CL, Law N, et al. Abl protein- 25. Llovet J, Ricci S, Mazzaferro V, et al. Sorafenib 1775 ^ 82. tyrosine kinase inhibitor STI571inhibits in vitro signal improves survival in advanced hepatocellular carcino- 40. Fumo G, Akin C, Metcalfe DD, Neckers L. 17- transduction mediated by c-kit and platelet-derived ma (HCC): results of a phase III randomized placebo- Allylamino-17-demethoxygeldanamycin (17-AAG) is growthfactor receptors. J Pharmacol ExpTher2000; controlled trial (SHARP trial) [abstract]. J Clin Oncol effective in down-regulating mutated, constitutively 295:139 ^ 45. 2007;25:LBA1. activated KIT protein in human mast cells. Blood 12. Debiec-Rychter M, Cools J, Dumez H, et al. Mecha- 26. Shah NP,Tran C, Lee FY, Chen P, Norris D, Sawyers 2004;103:1078 ^ 84. nisms of resistance to imatinib mesylate in gastroin- CL. Overriding imatinib resistance witha novel ABL ki- 41. SydorJR, Normant E, Pien CS, et al. Development of testinal stromal tumors and activity of the PKC412 nase inhibitor. Science 2004;305:399 ^ 401. 17-allylamino-17-demethoxygeldanamycin hydroqui- inhibitor against imatinib-resistant mutants. Gastroen- 27. Evans TR, Morgan JA, van den Abbeele AD, et al. none hydrochloride (IPI-504), an anti-cancer agent terology 2005;128:270 ^ 9. Phase I dose-escalation study of the SRC and multi- directed against Hsp90. Proc Natl Acad Sci U S A 13. Prenen H, Cools J, Mentens N, et al. Efficacy of the kinase inhibitor BMS-354825 in patients (pts) with 2006;103:17408 ^ 13. kinase inhibitor SU11248 against gastrointestinal stro- GIST and other solid tumors [abstract]. J Clin Oncol 42. Peng C, Brain J, Hu Y, et al. Inhibition of heat shock mal tumor mutants refractory to imatinib mesylate. 2005;23:3034. protein 90 prolongs survival of mice withBCR-ABL- Clin Cancer Res 2006;12:2622 ^ 7. 28. Chen Z, Lee FY, Bhalla KN,Wu J. Potent inhibition of T315I-induced leukemia and suppresses leukemic 14. Weisberg E, Wright RD, Jiang J, et al. Effects of platelet-derived growthfactor-induced responses in stem cells. Blood 2007;110:678 ^ 85. PKC412, nilotinib, and imatinib against GIST-associat- vascular smoothmuscle cells by BMS-354825 (dasa- 43. Demetri GD, George S, Morgan JA, et al. Inhibition ed PDGFRA mutants withdifferential imatinib sensi- tinib).MolPharmacol2006;69:1527^33. of the Heat Shock Protein 90 (Hsp90) chaperone tivity. Gastroenterology 2006;131:1734 ^ 42. 29. Carlomagno F, Anaganti S, GuidaT, et al. BAY 43 ^ withthenovel agent IPI-504 to overcome resistance 15. Bauer S,YuLK, Demetri GD, FletcherJA. Heat shock 9006 inhibition of oncogenic RET mutants. J Natl to tyrosine kinase inhibitors (TKIs) in metastatic GIST: protein 90 inhibition in imatinib-resistant gastrointesti- Cancer Inst 2006;98:326^ 34. updated results of a phase I trial [abstract]. J Clin nal stromal tumour. Cancer Res 2006;66:9153 ^ 61. 30.Wilhelm SM, Carter C,Tang L, et al. BAY 43 ^ 9006 Oncol 2007;25:10024.

Clin Cancer Res 2008;14(18) September 15, 2008 5758 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Activity of Dasatinib, a Dual SRC/ABL Kinase Inhibitor, and IPI-504, a Heat Shock Protein 90 Inhibitor, against Gastrointestinal Stromal Tumor−Associated PDGFRAD842V Mutation

Barbara Dewaele, Bartosz Wasag, Jan Cools, et al.

Clin Cancer Res 2008;14:5749-5758.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/14/18/5749

Cited articles This article cites 43 articles, 20 of which you can access for free at: http://clincancerres.aacrjournals.org/content/14/18/5749.full#ref-list-1

Citing articles This article has been cited by 18 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/14/18/5749.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/14/18/5749. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.