Pazopanib, a Novel Multitargeted Kinase Inhibitor, Shows Potent in Vitro Antitumor Activity in Gastric Cancer Cell Lines with FGFR2 Ampliﬁcation

Published OnlineFirst September 23, 2014; DOI: 10.1158/1535-7163.MCT-14-0255

Molecular Cancer Small Molecule Therapeutics Therapeutics

Pazopanib, a Novel Multitargeted Kinase Inhibitor, Shows Potent In Vitro Antitumor Activity in Gastric Cancer Cell Lines with FGFR2 Ampliﬁcation

Seung Tae Kim1, Hye-Lim Jang1,2, Su Jin Lee1, Jeeyun Lee1, Yoon-La Choi3, Kyoung-Mee Kim3, Jeonghee Cho4, Se Hoon Park1, Young Suk Park1, Ho Yeong Lim1, Masakazu Yashiro5, Won Ki Kang1, and Joon Oh Park1

Abstract Pazopanib is an orally bioavailable, ATP-competitive, multitargeted tyrosine kinase inhibitor mainly targeting VEGFR2 and PDGFR tyrosine kinases, but the biologic sequences of pazopanib activities beyond antiangiogenesis are poorly deﬁned. We used a panel of 38 gastric cancer cell lines to test the efﬁcacy of pazopanib. In a growth inhibition assay, genomic changes indicated that pazopanib had differential effects on cell growth. Treatment of the KATO-III, OCUM-2M, SNU-16, and HSC-39 gastric cancer cell lines harboring

FGFR2 amplification with pazopanib resulted in marked decreases of cell survival with IC50 in ranges of 0.1 to 2.0 mmol/L, whereas the same treatment of those cell lines without FGFR2 amplification had no growth- inhibitory effects. In the ectopic FGFR2-expressing model, treatment with the indicated concentrations of pazopanib significantly inhibited cell growth and colony formation by FGFR2-expressing NIH 3T3 cells with wild-type (WT) FGFR2 and mutant FGFR2 (S252W). Pazopanib also selectively suppressed constitutive FGFR2 signaling and phosphorylation of downstream effectors. In cell-cycle analysis, FGFR2-amplified cells under-

went cell-cycle arrest at the G1–S phase after pazopanib treatment, whereas there were no signiﬁcant effects on cell-cycle progression in cells without FGFR2 ampliﬁcation treated with pazopanib. In addition, pazopanib

increased a substantial fraction of sub-G1 only in FGFR2-amplified cells. These findings show that the activation of FGFR2 signaling by amplification may be a critical mediator of cell proliferation in a small subset of gastric cancer patients and that pazopanib may provide genotype-correlated clinical benefits beyond the setting of highly vascular tumors. Mol Cancer Ther; 13(11); 2527–36. 2014 AACR.

Introduction gastric cancer accounts for 20.8% of all cancers in Korea Gastric cancer is the leading cause of cancer-related (3). Although overall survival (OS) rates of gastric cancer death worldwide with an incidence of 18.9 deaths/ have improved after the implementation of a national 100,000 people per year (1). There are approximately screening program in adults over 40 years in Korea, a 934,000 new cases of gastric cancer each year worldwide, large proportion of patients is still diagnosed at metastatic with 56% of the new cases occurring in East Asia (2). stages. Typically, ﬂuoropyrimidines and platinum com- According to the Central Tumor Registry data for 2002, pounds form the backbone of chemotherapy for patients with advanced gastric cancer (AGC) and have substan- tially improved outcomes compared with single-agent 1Division of Hematology-Oncology, Department of Medicine, Samsung chemotherapy or best supportive care (4). However, the Medical Center, Sungkyunkwan University School of Medicine, Seoul, 2 3 median survival time following cytotoxic chemotherapy Korea. Samsung Biomedical Research Institute, Seoul, Korea. Depart- ment of Pathology, Samsung Medical Center, Sungkyunkwan University is still less than 12 months, and thus metastatic gastric School of Medicine, Seoul, Korea. 4Samsung Genome Institute, Samsung cancer remains a therapeutic challenge for medical oncol- Medical Center, Seoul, Korea. 5Department of Surgical Oncology, Osaka City University Graduate School of Medicine, Osaka, Japan. ogists. The role of molecularly targeted therapy has not been adequately explored in gastric cancer when com- Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). pared with other common solid tumors such as non–small cell lung cancer, breast, or colorectal cancer. S.T. Kim and H.-L. Jang contributed equally to this article. There is increasing evidence that suggests that sensi- Corresponding Author: Joon Oh Park, Division of Hematology-Oncology, tivity to molecularly targeted agents among diverse Department of Medicine, Samsung Medical Center, Sungkyunkwan Uni- versity School of Medicine, 50 Irwon-dong Gangnam-gu, Seoul 135-710, human cancers is well correlated with underlying genetic Korea. Phone: 82-2-3410-3457; Fax: 82-2-3410-1754; E-mail: alterations of the tumor. Therefore, identifying critical [email protected] genes that play pivotal roles in controlling tumor growth doi: 10.1158/1535-7163.MCT-14-0255 and survival will establish the basis for developing ther- 2014 American Association for Cancer Research. apeutic targets. The ﬁbroblast growth factor receptor

www.aacrjournals.org 2527

Kim et al.

(FGFR) tyrosine kinase family is composed of four kinases S252W were kindly provided Dr. Matthew Meyerson (FGFR 1, 2, 3, and 4) that differentially respond to 18 FGF (Dana-Farber Cancer Institute, Boston, MA) in 2010. YCC ligands and have long been implicated as causes of cancer cell lines and NIH 3T3 lines were maintained in DMEM (5). Recent research has shown that members of the FGFR (Gibco-BRL) and supplemented with 10% heat-inacti- family are promising targets that are deregulated by vated FBS, 100 U/mL penicillin, 100 U/mL streptomycin, amplification, point mutation, or translocation. Amplifi- and 2 mmol/L glutamine. The other cell lines were cul- cation or activation of FGFR1 has been reported in lung tured in RPMI-1640 medium (Gibco-BRL) and supple- cancer and translocations involving FGFR3, and activat- mented with 10% heat-inactivated FBS, 100 U/mL peni- ing in FGFR3, have been associated with multiple mye- cillin, 100 U/mL streptomycin, and 2 mmol/L glutamine. loma and bladder cancer. Missense mutations of FGFR2 All cells were incubated in a humidified atmosphere have also been identified in several types of human containing 5% CO2 at 37 C. All cell lines have been cancer, including lung squamous carcinoma, gastric can- authenticated using viability, morphology and growth cer, and endometrial cancer (6–8). These cancers are curve analysis on a regular basis, and tested negative for shown to be sensitive to FGFR2 kinase inhibition in the Mycoplasma. cell lines bearing such FGFR2 mutations, implicating The inhibition of growth was assessed by the 5-(3- FGFR2 as a novel therapeutic target (7). In gastric cancer, carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazoli- FGFR2 was first identified as the K-sam gene amplification um (MTS) assay according to previously established in the human gastric cancer cell line KATO-III, and is methods (16). Pazopanib (GW786034) was kindly provid- preferentially amplified or overexpressed in poorly dif- ed by GlaxoSmithKline (Brentford). Sorafenib and suni- ferentiated or undifferentiated types of gastric cancers tinib were purchased from LC Laboratories. Stock solu- (9, 10). Recent preclinical studies have demonstrated a tion of each compound was reconstituted in DMSO at a critical role for FGFR2 amplification in gastric cancer cell concentration of 10 mmol/L (final concentration of 0.1% proliferation and survival, and detected antitumor activ- DMSO for the MTS assay) and stored at 20C. For the ities of multitargeted tyrosine kinase inhibitors (MTKI) colony formation assay, FGFR2-expressing NIH 3T3 cells and pan-FGFR inhibitors in FGFR2-amplified gastric can- with WT FGFR2 and mutant FGFR2 (S252W) were used. cer cell lines, which suggests that FGFR2 amplification Two hundred cells per well were seeded in 6-well plates may be a promising therapeutic target in gastric cancer and treated with 0.1 to 1.0 mmol/L of pazopanib for 3 days. (11–15). Therefore, the present preclinical study primarily Triplicate cultures of each cell type were maintained at focused on the antitumor activities of pazopanib on 37 C for 14 days in an atmosphere of 5% CO2, with fresh FGFR2 signaling in gastric cancer. medium added after 7 days. Colonies, defined as groups of cells containing a minimum of 50 cells, were stained with 0.1% crystal violet and counted under an inverted Materials and Methods phase contrast microscope CKX31SF (Olympus Biosys- Cell culture, cell growth assay, and colony formation tems). Colonies were photographed after 2 weeks and the assay percentage of relative number of colonies was expressed Thirty-eight genetically defined human gastric cancer as (number of colonies from treated cells/number of cell lines representing broad genetic heterogeneity were colonies from controls) 100. The assay was repeated assembled to establish an in vitro platform that could be three times with duplicate samples. used to potentially identify genotype-correlated sensitivities and to eventually develop novel therapeutic strate- DNA sequencing for FGFR2 gies in gastric cancer. Human gastric cancer cell lines AGS, Analysis of DNA sequencing was performed for all KATO-III, MKN-1, MKN-28, MKN-45, MKN-74, N87, known mutation sites of FGFR2 exons. DNA was SNU-1, SNU-5, SNU-16, SNU216, SNU-484, SNU-520, extracted from five paraffin sections of 10-mm thickness SNU-601, SNU620, SNU-638, SNU-668, and SNU 719 cells containing a representative portion of each tumor block, were purchased from the Korean Cell Line Bank in 2008. using the QIAamp DNA Mini Kit (Qiagen). One hundred MKN-7, NUGC-3, and NUGC-4 were purchased from the nanograms of DNA was amplified in a 20-mL reaction Health Science Research Resources (Osaka, Japan) in 2008. solution containing 2 mLof10 buffer (Roche), 1.7 to 2.5 m YCC-1, YCC-2, YCC-3, YCC-6, YCC-7, YCC-10, YCC-11, mmol/L of MgCl2, 0.3 mol/L of each primer pairs, 250 YCC-16, and HS-741T were kindly provided by Dr. Sun mmol/L of deoxynucleotide triphosphate, and 2.5 U of Young Rha (Yonsei Cancer Center, Seoul, Korea) in 2008. DNA polymerase (Roche). Amplifications were per- HSC-39, HSC-44, HSC-57 HSC-58, HSC-60, and HSC-69 formed using a 5-minute initial denaturation at 94C; were kindly provided by Dr. Kazuyoshi Yanagihara followed by 30 cycles of 1 minute at 94C, 1 minute at (Yasuda Women’s University, Hiroshima, Japan) in 55C, and 1 minute at 72C, and a 10-minute final exten- 2009, TMK-1 by Dr. Wataru Yasui (Hiroshima University, sion at 72C. Polymerase chain reaction (PCR) products Hiroshima, Japan) in 2009, and OCUM-2M by Dr. Masa- were then 2% gel-purified with the QIAgen Gel Extraction kazu Yashiro (Osaka City University, Osaka, Japan) in Kit (Qiagen). DNA templates were processed for the DNA 2008. FGFR2-expressing NIH 3T3 cells with retroviruses sequencing reaction using the ABI-PRISM BigDye Termi- encoding isoform IIIb wild-type (WT) FGFR2 and FGFR2 nator version 3.1 (Applied Biosystems) with both forward

2528 Mol Cancer Ther; 13(11) November 2014 Molecular Cancer Therapeutics

Pazopanib for FGFR2 Ampliﬁcation in Gastric Cancer

and reverse sequence-specific primers. Twenty nano- 10 nmol/L/L b-glycerophosphate, 1 mmol/L/L sodium grams of purified PCR products was used in a 20-mL vanadate, 0.5 mmol/L/L DTT, 4 mg/mL leupeptin, 4 mg/ sequencing reaction solution containing 8 mL of BigDye mL pepstatin, 4 mg/mL apoprotein, and 1 mmol/L/L Terminator v3.1 and 0.1 mmol/L of the same PCR primer. phenylmethylsulfonyl fluoride. Lysates were centrifuged Sequencing reactions were performed using 25 cycles of at 16,000 g for 5 minutes at 4C. The supernatant was 10 seconds at 96C, 5 seconds at 50C, and 4 minutes at used for subsequent procedures. For IPs, anti-FGFR2 60C. Sequence data were generated with the ABI PRISM antibody (Santa Cruz Biotechnology) was added to the 3100 DNA Analyzer (Applied Biosystems) and analyzed lysate and then incubated with protein A/G agarose by Sequencing Analyzing Software v5.3.1 (Applied Bio- beads (Santa Cruz Biotechnology). IPs were washed three systems) to compare variations. Primers used for FGFR2 times with ice-cold lysis buffer before boiling in 2 sequencing were as follows: exon 6, F: 50-TCTGGCAT- Lamelli sample buffer. Western blot analyses were con- GAGGTCACTG, R: 50-CACTCTCTGCTGGCTAGT; exon ducted after separation by SDS/PAGE electrophoresis 7, F: 50-ATGTGTAGGAGGTGAAGC, R: 50-ATCTCAC- and transfer to nitrocellulose membranes. Immunoblot- TGTGTGTGAAT; exon 8, F: 50-TAGTGACAGTGAC- ting was performed according to the antibody manufac- TACCT, R: 50-GCAACGCACCATGACTAG; exon 11, F: turer’s recommendations. Antibody binding was detected 50-TAACAGTAGCTGCCCATG, R: 50-CTTTGTGTTCA- using an enhanced chemiluminescence system (PerkinEl- TGGCTAGGA; exon 13, F: 50-AACTACTGACCTCA- mer, Waltham, MA, USA). The phospho-AKT (S473), AGTG, R: 50-CGGTTGCTGATTATTCTG. AKT, cleaved PARP, and cleaved caspase-3 antibodies were from Cell Signaling Technology (Beverly, MA, USA). FGFR2 amplification with quantitative real-time The phospho-FGFR and FGFR2 antibodies were from PCR and FISH R&D Systems (Minneapolis, MN, USA). The FGFR2 anti- Quantitative real-time PCR (qRT-PCR) with TaqMan body for IP was from Santa Cruz Biotechnology. The probes was performed to analyze FGFR2 copy number phosphotyrosine (4G10) antibody was from Upstate Bio- using an ABI PRISM 7000 Sequence Detection System technology (Lake Placid, NY, USA). (Applied Biosystems). The FGFR2 primers were F: At approximately 70% of cell confluence, protein CCCCCTCCACAATCATTCCT, R: ACCGGCGGCCTA- extracts of KATO-III, OCUM-2M, MKN-45 and N87 cells GAAAAC; the probe was labeled with the reporter dye 6- were also prepared for the Human Phospho-RTK Array carboxyfluorescein (6-FAM-TCGTCTAGCCTTTTCTTTT- (R&D Systems, Minneapolis, MN, USA). This array can MGBNFQ). The average FGFR2 copy number per cell was detect the phosphorylation level of 42 different receptor calculated from the differences in the threshold amplifica- tyrosine kinases (RTK) on the same nitrocellulose mem- tion cycles between EGFR and RNaseP. brane. Assays were performed in accordance with the Fluorescence in situ hybridization (FISH) was also per- manufacturer’s instructions. formed according to the established protocol using an FGFR2 probe at chromosome arm 10q26 and chromosome Fluorescence-activated cell sorting for cell-cycle 10 centromere probe (CEP10) purchased from Macrogen analysis (Supplementary Data). Four-micrometer tumor sections The fluorescence-activated cell sorting (FACS) analysis generated from TMA blocks were pretreated by depar- was performed using methods similar to those described affinizing in xylene and dehydrating in ethanol. The previously. Briefly, 1 to 1.5 106 cells were seeded into sections were immersed in Tris-base and EDTA (TE), 10-cm3 plates, and the drugs were added 24 hours later. washed in phosphate-buffered saline (PBS), and then After 24 hours of treatment, the cells were trypsinized and treated with Digest-All (Zymed). Sections were then fixed fixed overnight in ethanol at 4C. Fixed cells were then with formalin and dehydrated in ethanol. After codena- resuspended in 0.5% RNase A (Sigma), centrifuged, resus- turation of the tissue and the probe mixture (FGFR2 and pended in 5 mmol/L propidium iodine (Sigma) and 38 CEP10) at 70C for 3 minutes, the sections were hybridized mmol/L sodium citrate, incubated at room temperature at 37oC for 48 to 72 hours, washed with sodium citrate and for 30 minutes, and analyzed by the FACSCalibur flow Tween-20 containing buffers, and counterstained with cytometry (Becton Dickinson) with a cell-cycle test soft- DAPI. One hundred cells from each TMA core were ware, ModFit LT (Verity Software House Inc.). All experi- analyzed and the number of FGFR2 and CEP10 signals ments were repeated three times. was determined; tumors with FGFR2 to CEP10 2 or the presence of 10% gene cluster were defined as amplified. Patients and tissue specimens The outcomes of 544 patients with stage II–IV (M0) Immunoprecipitation, Western blot analysis, and gastric cancer who received adjuvant chemoradiation RTK array therapy after curative surgery were previously reported Cells grown under the previously specified conditions (17). Of these patients and an additional 23 stage IB were lysed in the following lysis buffer composition: 20 patients who were included in a previous study, forma- mmol/L/L Tris (pH 7.4), 150 mmol/L/L NaCl, 1% NP40, lin-fixed paraffin-embedded primary tumor tissues were 10% glycerol, 1 mmol/L/L EDTA, 1 mmol/L/L EGTA, 5 available from 482 patients. The postoperative adjuvant mmol/L/L sodium pyrophosphate, 50 mmol/L/L NaF, treatment that was adopted was the same as that used for

www.aacrjournals.org Mol Cancer Ther; 13(11) November 2014 2529

Kim et al.

m the INT-0116 (SWOG-9008) trial (18). All patients provid- IC50 in ranges of 0.1 to 2.0 mol/L, whereas similar ed written informed consent according to the institutional treatment of the cell lines without FGFR2 amplification guidelines and the study was approved by the Institu- had no effect (Fig. 2A). In addition, treatment with either tional Review Board at Samsung Medical Center (Seoul, sorafenib or sunitinib similarly showed significant growth Korea). inhibition in the gastric cancer lines harboring FGFR2 amplifications (Fig. 2B and C). To further test the efficacy Statistical analyses of pazopanib against ectopic expression of FGFR2, Disease-free survival (DFS) was defined as the time FGFR2-expressing NIH 3T3 cells with WT FGFR2 and from surgery to the first relapse of cancer, or death mutant FGFR2 (S252W) were used in a growth inhibition because of any cause. OS was calculated from the date and colony formation assay. Treatment with the indicated of surgery to the date of death. OS and DFS were calcu- concentrations of pazopanib significantly inhibited cell lated using the Kaplan–Meier method. Correlation anal- growth and colony formation by the FGFR2-expressing yses were performed using the two-sided x2 test or the NIH 3T3 cells with WT FGFR2 and mutant FGFR2 Fisher exact test. Differences in DFS and OS were com- (S252W), when compared with the NIH-3T3 cell with pared using the log-rank tests and the Cox proportional vector (Fig. 3A–C). hazards analysis. P values less than 0.05 were considered Cell-cycle analyses were performed to examine the statistically significant. mechanisms of growth inhibition of pazopanib in FGFR2-amplified cells. All cell lines were treated for 24 Results hours in the presence and absence of 1 mmol/L pazopanib. Correlation of sensitivity with MTKIs and FGFR2 The FGFR2-amplified cells underwent cell-cycle arrest at amplification in human gastric cancer cell lines the G1–S phase after pazopanib treatment, whereas there The efficacy of MTKI, including pazopanib, sorafenib, was no significant effect on cell-cycle progression in cells and sunitinib, whose clinical activity is not well recog- without FGFR2 amplification after treatment with pazo- nized in gastric cancer, was first tested using the in vitro panib. In addition, pazopanib caused an increase in the platform for modeling genotype-correlated drug sensitiv- substantial fraction of sub-G1 in FGFR2-amplified cells ity in a panel of 37 gastric cancer lines. Among the 37 only (Fig. 4). gastric cancer cell lines tested, most cells were not sensitive to MTKIs and some cells showed growth inhibition Pazopanib selectively inhibits FGFR2 based on the MTKI. However, four gastric cancer cells phosphorylation and downstream signaling (KATO-III, OCUM-2M, SNU-16, and HSC-39; 10.8%) molecules in FGFR2-amplified gastric cancer cell exhibited dramatic sensitivity, which could indicate they lines share the same genetic mechanism(s) for underlying drug The effects of drug treatment on FGFR2-dependent sensitivity to MTKIs (Fig. 1A). signaling were tested to address the mechanism by which To understand possible underlying genetic mechan- pazopanib triggers cell death in FGFR2-amplified cells. isms of MTKI sensitivity, genetic alterations were corre- Pazopanib selectively suppressed the constitutive phos- lated with their MTKI sensitivities. Of the four cell lines pho-FGFR2 in FGFR2-amplified KATO-III cells, whereas showing sensitivity to MTKIs, all were found to have high pazopanib did not suppress the phospho-MET and phos- FGFR2 gene copy numbers by qRT-PCR (Fig. 1B). In pho-HER2 in MKN-45 and N87 cells, respectively (Fig. addition, FISH analysis showed that the amplified gene 5A). Most importantly, treatment with this concentration copies were integrated within a chromosomal locus at of pazopanib also effectively abrogated the baseline phos- 10p26, consistent with the so-called homogeneously stain- phorylation of downstream effectors of growth factor ing regions (HSR) in these cell lines (Fig. 1C–H). The FISH receptors such as ERK-1/2 and AKT (Fig. 5B and C). results were consistent with the qRT-PCR results in iden- Sorafenib and sunitinib similarly downregulated FGFR2 tifying the subset of cell lines with FGFR2 amplification. signaling in KATO-III and OCUM-2M cell lines (Fig. 5A– However, any previously reported known oncogenic C). Thus, constitutive activation of these proliferative and FGFR2 mutations were not found in the current study of survival pathways in FGFR2-amplified cells appears to the gastric cancer cell lines (data not shown; refs. 7, 19). depend specifically on baseline FGFR2 signaling. In contrast, in cells without FGFR2 amplification where FGFR2 is Pazopanib blocks proliferation and survival of gastric not constitutively activated, pazopanib had no effect on cancer cell lines with FGFR2 amplification signaling. This indicates that these effectors are likely to Gastric cancer cell lines were first treated with pazo- be activated through alternative growth factor receptors panib to investigate whether inhibition of FGFR2 kinase (Fig. 5A). activity is effective against gastric cancer cell lines har- Induction of apoptosis has been linked to suppression boring FGFR2 amplification and to test the potential of essential growth factor–mediated survival pathways therapeutic relevance of these findings. Treatment of the and increased fraction of sub-G1 in cell-cycle analysis. KATO-III, OCUM-2M, SNU-16, and HSC-39 gastric can- PARP cleavage in Western blot analyses demonstrated cer cell lines harboring FGFR2 amplification with pazo- apoptosis in FGFR2-amplified cells treated with pazopa- panib resulted in marked decreases in cell survival with nib, but not in cells without FGFR2 amplification (Figs. 4

2530 Mol Cancer Ther; 13(11) November 2014 Molecular Cancer Therapeutics

Pazopanib for FGFR2 Ampliﬁcation in Gastric Cancer

Figure 1. A, sensitivity of 38 gastric cancer cell lines to MTKIs. Sensitivity is measured by the MTS assay after treatment of multitargeted agents for 72 hours. IC50 value is calculated from at least three independent experiments. B, amplification of FGFR2 measured by qRT-PCR in 38 gastric cancer cell lines. Each column represents the average of three independent experiments. Error bars indicate standard deviations. C–H, FGFR2 amplification is also detected by dual- color FISH (FGFR2, red/CEP10, green) in gastric cancer cell lines; cells with FGFR2 amplification (C, KATO-III; D, SNU-16; E, OCUM-2M; F, HSC-39) and cells with no FGFR2 amplification (G, MKN-45; H, N87). and 5B). These findings suggest that FGFR2 signaling presumably reflects the effect of low-level copy-number plays a critical role in cell survival in these FGFR2- variability in the control locus used in qRT-PCR analysis, addicted cell lines. resulting in underestimation of the true extent of FGFR2 amplification. However, FGFR2 amplification was not Correlations between FGFR2 amplification and associated with clinicopathologic parameters such as age, clinical outcomes sex, histology, Bormann type, Lauren classification, tumor To assess FGFR2 amplification in clinical gastric cancer location, or tumor–node–metastasis (TNM) staging. In cases, tumor specimens from 482 patients with AGC were addition, FGFR2 amplification was not related to clinical analyzed by qRT-PCR using TaqMan probes and dual- outcomes (data not shown). color FISH (Fig. 6). Of the 482 patients, 24 (5%) of the patients had FGFR2 copy numbers greater than 4.0 copies. FISH and qRT-PCR analyses were also consistent in iden- Discussion tifying FGFR2 amplification in tissue specimens, with This study shows that FGFR2 is amplified in a subset of higher-fold amplification apparent by FISH (Fig. 6). This gastric cancer and is closely associated with sensitivity to

www.aacrjournals.org Mol Cancer Ther; 13(11) November 2014 2531

Kim et al.

Figure 2. A, effect of pazopanib on cell growth in gastric cancer cell lines with various genetic alterations. Pazopanib has differential effects on survival of gastric cancer cell lines according to FGFR2 amplification status. Sensitivity to pazopanib is well correlated with FGFR2 amplification status. Sensitivity is measured by the MTS assay after treatment of pazopanib for 72 hours and IC50 value is calculated from at least three independent experiments. B and C, effect of either sunitinib or sorafenib on IC50 value and cell growth in gastric cancer cell lines with various genetic alterations. Treatment with either sorafenib or sunitinib shows similar significant growth inhibition in the gastric cancer lines harboring FGFR2 amplifications, when compared with gastric cancer lines without FGFR2 amplifications. Sensitivity is measured by the MTS assay after treatment of either sunitinib or sorafenib for 72 hours and IC50 value is calculated from at least three independent experiments.

pazopanib. Pazopanib is an orally bioavailable, ATP- inhibitor and its consequent effects on angiogenesis (23). competitive, MTKI that mainly targets VEGFR2, PDGFR, Notably, RCC tumors are typically highly vascularized, and c-Kit tyrosine kinases in a low-nanomolar concentra- suggesting a potential critical requirement for angiogen- tion, resulting in selective inhibition of VEGF-induced esis in that disease setting. endothelial cell proliferation and the growth of a broad MTKIs are a relatively new class of therapeutics with a range of human tumor xenografts in mice (20, 21). On the propensity to inhibit multiple kinases, but the biologic basis of these preclinical results, pazopanib is under sequences of multiple kinase activities beyond antiangio- clinical development for the treatment of multiple tumor genesis are poorly defined (21). Moreover, although the types and has demonstrated significant clinical activity as ability of MTKIs to target additional kinases may contrib- monotherapy in patients with renal cell carcinoma (RCC) ute to clinical activities in other solid tumors, this possi- in recent phase III trial (22). Like other MTKIs, including bility has not been well investigated beyond PDGFRA sunitinib and sorafenib, the clinical success of pazopanib mutations that are largely confined to gastrointestinal in RCC may indicate the role of pazopanib as a VEGFR stromal tumor (GIST; ref. 24). In an in vitro kinase assay,

2532 Mol Cancer Ther; 13(11) November 2014 Molecular Cancer Therapeutics

Pazopanib for FGFR2 Ampliﬁcation in Gastric Cancer

Figure 3. A, ectopic expression of FGFR2 WT and FGFR2 S252 mutation in NIH 3T3 cells is correlated with sensitivity to PD173074 and pazopanib. Treatment of the FGFR2-expressing NIH 3T3 cells with 1 mmol/L of PD173074 or pazopanib significantly inhibits cell growth as compared with the control. B and C, treatment with the indicated concentrations of PD173074 or pazopanib significantly inhibits colony formation by the FGFR2-expressing NIH 3T3 cells, when compared with the control cell. All experiments were repeated three times. pazopanib had modest activity against FGFR2 kinase with In this preclinical study, pazopanib sensitivity was nanomolar concentrations expecting antitumor activity in restricted to a small number of gastric cancer cell lines tumors having dysregulated FGFR2 signaling (21). harboring FGFR2 amplification. This study also showed

Figure 4. Pazopanib induces cell- cycle arrest and apoptosis in the FGFR2-amplified cells in cell-cycle analysis. KATO-III and OCUM-2M, representative of the FGFR2- amplified cells, undergo G1–S and G2–M arrest in response to pazopanib and show a sub-G1 peak following pazopanib exposure. However, there is no significant effect on cell-cycle progression and sub-G1 peak in cells without FGFR2 amplification after treatment of pazopanib. All experiments were repeated three times.

www.aacrjournals.org Mol Cancer Ther; 13(11) November 2014 2533

Kim et al.

Figure 5. A, pazopanib effectively inhibits FGFR2 signaling and induces apoptosis in the FGFR2- amplified gastric cancer cell lines. A phospho-RTK array reveals that phospho-FGFR2 is selectively suppressed in the presence of pazopanib in FGFR2-amplified KATO-III cells, while phospho-MET or phospho-HER2 was not inhibited in MET-amplified or HER2-amplified cells. Each cell line was treated with 1 mmol/L pazopanib, 2 mmol/L sorafenib, or 2 mmol/L sunitinib for 6 hours. In the array, each RTK is spotted in duplicate and hybridization signals at the corners serve as controls. B, pazopanib effectively abrogates FGFR2 signaling and Akt phosphorylation in OCUM-2M cells. C, there is a marked increase in the cleaved PARP (89 kDa) product in OCUM-2M after both pazopanib and sorafenib treatment. The cells were exposed to the indicated concentrations of either pazopanib, sorafenib, or sunitinib for 6 hours. For PARP, the cells were treated with each drug for 24 hours. The cells were lysed and the proteins were separated by Western blotting, transferred to a nitrocellulose membrane, and probed with the indicated antibodies. Cell extracts were also immunoprecipitated with an antibody to FGFR2. The precipitated proteins were determined by immunoblotting with the indicated antibodies.

that FGFR2 activation is coupled to critical downstream We found that FGFR2 amplification was also observed effectors, such as Erk and Akt, and that disrupting these in a relatively small subset (5%) of gastric cancer clinical pathways seems to mediate the inhibitory effects of pazo- samplesinthisstudy,whichisconsistentwithrecent panib on proliferation and cell survival in the FGFR2- Japanese and Chinese study on genomic alterations (25, addicted cells in vitro. Hence, these findings suggest that 26). Although the frequency of FGFR2 amplification has activated FGFR2 signaling by amplification may be a not been well investigated in a large cohort, relatively critical target, and that pazopanib might be useful as an low incidence of FGFR2 amplification in gastric cancer FGFR2 inhibitor in the clinical management of the subset specimens might be a possible impediment in develop- of tumors that exhibit FGFR2 activation. Interestingly, ing FGFR2-targeted therapeutics in gastric cancer. there was no evidence of FGFR2 mutations in 38 gastric Moreover, because a large-scale analysis of FGFR2 sta- cancer cell lines, which argues in favor of amplification tus has not been performed in gastric cancer tissues or rather than mutation as the preferred mechanism of DNA samples, the cutoff value for FGFR2 amplification FGFR2 activation in a subset of gastric cancer. In contrast, or FGFR2 expression has not yet been determined. FGFR2 activation is mainly associated with oncogenic Therefore, companion diagnostics to reliably detect mutations in endometrial and breast cancers (7). FGFR2 activation measured by FGFR2 amplification or

2534 Mol Cancer Ther; 13(11) November 2014 Molecular Cancer Therapeutics

Pazopanib for FGFR2 Ampliﬁcation in Gastric Cancer

Figure 6. A–D, FGFR2 amplification is also detected by dual-color FISH in the specimens from patients with gastric cancer; no FGFR2 amplification (A) and FGFR2 amplification (B–D). Tumors with FGFR2 to CEP10 2 or presence of 10% gene cluster were defined as amplified.

FGFR2 protein expression needs to be standardized. As data of this trial will be available in the late 2014, and the the growth curves for HSC-39 and SNU-16 only reach a significance of FGFR2 amplification and other biomar- lower limit of 40% survival in this preclinical study (Fig. kers on clinical outcomes in this context of the trial will 2A), it is presumed that the relatively lower level of be also investigated. As several FGFR kinase inhibitors amplification in SNU-16 and HSC-39 (Fig. 1B) may be are now in clinical trials, including brivanib, dovitinib, related to somewhat lower response to pazopanib. BIBF 1120, and SU-6668, it may be useful to test these Therefore, a predictive role of FGFR2 amplification for inhibitors on patients with gastric cancer harboring treatment response according the cutoff value should be focal FGFR2 amplifications (28–31). further validated as a novel therapeutic target in pro- In conclusion, the findings of this study show that spective clinical studies. In addition, given that pazo- activation of FGFR2 signaling by amplification may be panib is less potent against the FGFRs versus the a critical mediator of cell proliferation in a small subset of VEGFRs, it is important whether an efficacious concen- gastric cancer patients and may sensitize these cancer cells tration of pazopanib can be achieved in patients with to pazopanib. These findings suggest that pazopanib as gastric cancer. Although no pharmacokinetic (PK) data well as other MTKIs may provide genotype-associated on pazopanib in patients with gastric cancer are avail- clinical benefits beyond the setting of highly vascular able, steady-state exposure plateaued at doses 800 tumors such as RCCs. Although molecular targeted ther- mg/day and was 15 mg/mL (34 mmol/L) in 93% of apy has been studied less extensively in AGC than in other patients receiving a dose of 800 mg daily from PK data solid tumor types, future research should focus on novel in a phase I study (27), which plasma concentration can therapeutic targets in clinical trials. be clinically achieved might be enough to block FGFRs. This Ctrough plasma pazopanib concentration of 15 Disclosure of Potential Conflicts of Interest mg/mL also appeared to correlate with clinical activity J.O. Park received a commercial research grant from GlaxoSmithKline. in patients with RCC, which could be effective for No potential conflicts of interest were disclosed by the other authors. patients with gastric cancer with FGFR2 amplification as well. A phase II trial is currently being conducted Authors' Contributions involving pazopanib combined with capecitabine and Conception and design: H.-L. Jang, S.H. Park, Y.S. Park, H.Y. Lim, J.O. Park oxaliplatin (CapeOX) in AGC (ClinicalTrials.gov Iden- Development of methodology: K.-M. Kim, J.O. Park tifier: NCT01130805). Of 66 patients enrolled in this Acquisition of data (provided animals, acquired and managed patients, investigator-sponsored trial as of July 2014, 3 cases provided facilities, etc.): S.T. Kim, H.-L. Jang, S.J. Lee, J. Lee, Y.-L. Choi, K.-M. Kim, S.H. Park, H.Y. Lim, W.K. Kang, J.O. Park (7.1%) of 42 patients tested were found to have FGFR2 Analysis and interpretation of data (e.g., statistical analysis, biostatis- amplification by dual-color FISH. The complete clinical tics, computational analysis): S.T. Kim, H.-L. Jang, J. Lee, J.O. Park

www.aacrjournals.org Mol Cancer Ther; 13(11) November 2014 2535

Kim et al.

Writing, review, and/or revision of the manuscript: S.T. Kim, K.-M. Kim, Program through the National Research Foundation of Korea NRF-2009- J. Cho, S.H. Park, Y.S. Park, W.K. Kang, J.O. Park 0074138 (to J.O. Park). Administrative, technical, or material support (i.e., reporting or orga- The costs of publication of this article were defrayed in part by the nizing data, constructing databases): Y.S. Park, M. Yashiro, J.O. Park payment of page charges. This article must therefore be hereby marked Study supervision: Y.S. Park, M. Yashiro, J.O. Park advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Grant Support This work was supported by the Samsung Biomedical Research Insti- Received March 26, 2014; revised August 14, 2014; accepted September 6, tute Grant CA-8-213 (to J.O. Park), and by the Basic Science Research 2014; published OnlineFirst September 23, 2014.

References 1. Cunningham D, Jost LM, Purkalne G, Oliveira J. ESMO minimum resection with D2 nodal dissection for adenocarcinoma of the stom- clinical recommendations for diagnosis, treatment and follow-up of ach. Int J Radiat Oncol Biol Phys 2005;63:1279–85. gastric cancer. Ann Oncol 2005;16:i22–3. 18. Macdonald JS, Smalley SR, Benedetti J, Hundahl SA, Estes NC, 2. Inoue M, Tsugane S. Epidemiology of gastric cancer in Japan. Post- Stemmermann GN, et al. Chemoradiotherapy after surgery compared grad Med J 2005;81:419–24. with surgery alone for adenocarcinoma of the stomach or gastro- 3. Bae JM WY, Jung KW, Park JG. Annual report of the Korea central esophageal junction. N Engl J Med 2001;345:725–30. cancer registry program 2000: based on registered data from 131 19. Katoh M. Cancer genomics and genetics of FGFR2 (review). Int J Oncol hospitals. Cancer Res Treat 2002;34:77–83. 2008;33:233–7. 4. Wagner AD, Grothe W, Haerting J, Kleber G, Grothey A, Fleig WE. 20. Kumar R, Knick VB, Rudolph SK, Johnson JH, Crosby RM, Crouthamel Chemotherapy in advanced gastric cancer: a systematic review and M-C, et al. Pharmacokinetic-pharmacodynamic correlation from meta-analysis based on aggregate data. J Clin Oncol 2006;24:2903–9. mouse to human with pazopanib, a multikinase angiogenesis inhibitor 5. Turner N, Grose R. Fibroblast growth factor signalling: from develop- with potent antitumor and antiangiogenic activity. Mol Cancer Ther ment to cancer. Nat Rev Cancer 2010;10:116–29. 2007;6:2012–21. 6. Davies H, Hunter C, Smith R, Stephens P, Greenman C, Bignell G, et al. 21. Karaman MW, Herrgard S, Treiber DK, Gallant P, Atteridge CE, Camp- Somatic mutations of the protein kinase gene family in human lung bell BT, et al. A quantitative analysis of kinase inhibitor selectivity. Nat cancer. Cancer Res 2005;65:7591–5. Biotechnol 2008;26:127–32. 7. Dutt A, Salvesen HB, Chen T-H, Ramos AH, Onofrio RC, Hatton C, et al. 22. Sternberg CN, Davis ID, Mardiak J, Szczylik C, Lee E, Wagstaff J, et al. Drug-sensitive FGFR2 mutations in endometrial carcinoma. Proc Natl Pazopanib in locally advanced or metastatic renal cell carcinoma: Acad Sci U S A 2008;105:8713–7. results of a randomized phase III trial. J Clin Oncol 2010;28:1061–8. 8. Jang J-H, Shin K-H, Park J-G. Mutations in fibroblast growth factor 23. Xu C-F, Bing NX, Ball HA, Rajagopalan D, Sternberg CN, Hutson TE, receptor2 and fibroblast growth factorreceptor3 genes associated with et al. Pazopanib efficacy in renal cell carcinoma: evidence for predic- human gastric and colorectal cancers. Cancer Res 2001;61:3541–3. tive genetic markers in angiogenesis-related and exposure-related 9. Hattori Y, Odagiri H, Nakatani H, Miyagawa K, Naito K, Sakamoto H, genes. J Clin Oncol 2011;29:2557–64. et al. K-sam, an amplified gene in stomach cancer, is a member of the 24. Heinrich MC, Corless CL, Duensing A, McGreevey L, Chen C-J, heparin-binding growth factor receptor genes. Proc Natl Acad Sci U S Joseph N, et al. PDGFRA activating mutations in gastrointestinal A 1990;87:5983–7. stromal tumors. Science 2003;299:708–10. 10. Hattori Y, Itoh H, Uchino S, Hosokawa K, Ochiai A, Ino Y, et al. 25. Matsumoto K, Arao T, Hamaguchi T, Shimada Y, Kato K, Oda I, et al. Immunohistochemical detection of K-sam protein in stomach cancer. FGFR2 gene amplification and clinicopathological features in gastric Clin Cancer Res 1996;2:1373–81. cancer. Br J Cancer 2012;106:727–32. 11. Takeda M, Arao T, Yokote H, Komatsu T, Yanagihara K, Sasaki H, et al. 26. Liu YJ, Shen D, Yin X, Gavine P, Zhang T, Su X, et al. HER2, MET and AZD2171 shows potent antitumor activity against gastric cancer over- FGFR2 oncogenic driver alterations define distinct molecular seg- expressing fibroblast growth factor receptor 2/keratinocyte growth ments for targeted therapies in gastric carcinoma. Br J Cancer 2014; factor receptor. Clin Cancer Res 2007;13:3051–7. 110:1169–78. 12. Yashiro M, Shinto O, Nakamura K, Tendo M, Matsuoka T, Matsuzaki T, 27. Hamberg P, Verweij J, Sleijfer S. (Pre-)Clinical pharmacology and et al. Synergistic antitumor effects of FGFR2 inhibitor with 5-fluoro- activity of pazopanib, a novel multikinase angiogenesis inhibitor. uracil on scirrhous gastric carcinoma. Int J Cancer 2010;126:1004–16. Oncologist 2010;15:539–47. 13. Kunii K, Davis L, Gorenstein J, Hatch H, Yashiro M, Di Bacco A, et al. 28. Chen J, Lee BH, Williams IR, Kutok JL, Mitsiades CS, Duclos N, et al. FGFR2-amplified gastric cancer cell lines require FGFR2 and Erbb3 FGFR3 as a therapeutic target of the small molecule inhibitor PKC412 signaling for growth and survival. Cancer Res 2008;68:2340–8. in hematopoietic malignancies. Oncogene 2005;24:8259–67. 14. Gozgit JM, Wong MJ, Moran L, Wardwell S, Mohemmad QK, Nar- 29. Sarker D, Molife R, Evans TRJ, Hardie M, Marriott C, Butzberger- asimhan NI, et al. Ponatinib (AP24534), a multitargeted pan-FGFR Zimmerli P, et al. A phase I pharmacokinetic and pharmacodynamic inhibitor with activity in multiple FGFR-amplified or mutated cancer study of TKI258, an oral, multitargeted receptor tyrosine kinase inhib- models. Mol Cancer Ther 2012;11:690–9. itor in patients with advanced solid tumors. Clin Cancer Res 2008; 15. Xie L, Su X, Zhang L, Yin X, Tang L, Zhang X, et al. FGFR2 gene 14:2075–81. amplification in gastric cancer predicts sensitivity to the selective 30. Hilberg F, Roth GJ, Krssak M, Kautschitsch S, Sommergruber W, FGFR inhibitor AZD4547. Clin Cancer Res 2013;19:2572–83. Tontsch-Grunt U, et al. BIBF 1120: triple angiokinase inhibitor with 16. Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO, sustained receptor blockade and good antitumor efficacy. Cancer Res et al. MET amplification leads to gefitinib resistance in lung cancer by 2008;68:4774–82. activating ERBB3 signaling. Science 2007;316:1039–43. 31. Machida S, Saga Y, Takei Y, Mizuno I, Takayama T, Kohno T, et al. 17. Kim S, Lim DH, Lee J, Kang WK, MacDonald JS, Park CH, et al. An Inhibition of peritoneal dissemination of ovarian cancer by tyrosine observational study suggesting clinical benefit for adjuvant postop- kinase receptor inhibitor SU6668 (TSU-68). Int J Cancer 2005;114: erative chemoradiation in a population of over 500 cases after gastric 224–9.

2536 Mol Cancer Ther; 13(11) November 2014 Molecular Cancer Therapeutics

Pazopanib, a Novel Multitargeted Kinase Inhibitor, Shows Potent In Vitro Antitumor Activity in Gastric Cancer Cell Lines with FGFR2 Amplification

Seung Tae Kim, Hye-Lim Jang, Su Jin Lee, et al.

Mol Cancer Ther 2014;13:2527-2536. Published OnlineFirst September 23, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-14-0255

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2014/09/24/1535-7163.MCT-14-0255.DC1

Cited articles This article cites 31 articles, 19 of which you can access for free at: http://mct.aacrjournals.org/content/13/11/2527.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/13/11/2527.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 Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/13/11/2527. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.