MK-2461, a Novel Multitargeted Kinase Inhibitor, Preferentially Inhibits the Activated C-Met Receptor

Published OnlineFirst February 9, 2010; DOI: 10.1158/0008-5472.CAN-09-2541 Published Online First on February 9, 2010 as 10.1158/0008-5472.CAN-09-2541

Therapeutics, Targets, and Chemical Biology Cancer Research MK-2461, a Novel Multitargeted Kinase Inhibitor, Preferentially Inhibits the Activated c-Met Receptor

Bo-Sheng Pan, Grace K.Y. Chan, Melissa Chenard, An Chi, Lenora J. Davis, Sujal V. Deshmukh, Jackson B. Gibbs, Susana Gil, Gaozhen Hang, Harold Hatch, James P. Jewell, Ilona Kariv, Jason D. Katz, Kaiko Kunii, Wei Lu, Bart A. Lutterbach, Cloud P. Paweletz, Xianlu Qu, John F. Reilly, Alexander A. Szewczak, Qinwen Zeng, Nancy E. Kohl, and Christopher J. Dinsmore

Abstract The receptor tyrosine kinase c-Met is an attractive target for therapeutic blockade in cancer. Here, we describe MK-2461, a novel ATP-competitive multitargeted inhibitor of activated c-Met. MK-2461 inhibited in vitro phosphorylation of a peptide substrate recognized by wild-type or oncogenic c-Met kinases

(N1100Y, Y1230C, Y1230H, Y1235D, and M1250T) with IC50 values of 0.4 to 2.5 nmol/L. In contrast, MK- 2461 was several hundredfold less potent as an inhibitor of c-Met autophosphorylation at the kinase activation loop. In tumor cells, MK-2461 effectively suppressed constitutive or ligand-induced phosphorylation of the juxtamembrane domain and COOH-terminal docking site of c-Met, and its downstream signaling to the phos- phoinositide 3-kinase–AKT and Ras–extracellular signal-regulated kinase pathways, without inhibiting auto‐ phosphorylation of the c-Met activation loop. BIAcore studies indicated 6-fold tighter binding to c-Met when it was phosphorylated, suggesting that MK-2461 binds preferentially to activated c-Met. MK-2461 displayed significant inhibitory activities against fibroblast growth factor receptor (FGFR), platelet-derived growth factor receptor, and other receptor tyrosine kinases. In cell culture, MK-2461 inhibited hepatocyte growth factor/ c-Met–dependent mitogenesis, migration, cell scatter, and tubulogenesis. Seven of 10 MK-2461–sensitive tumor cell lines identified from a large panel harbored genomic amplification of MET or FGFR2. In a murine xenograft model of c-Met–dependent gastric cancer, a well-tolerated oral regimen of MK-2461 administered at 100 mg/kg twice daily effectively suppressed c-Met signaling and tumor growth. Similarly, MK-2461 inhibited the growth of tumors formed by s.c. injection of mouse NIH-3T3 cells expressing oncogenic c-Met mutants. Taken together, our findings support further preclinical development of MK-2461 for cancer therapy. Cancer Res; 70(4); 1524–33. ©2010 AACR.

Introduction phosphorylation of Y1349, Y1356, and Y1365 in the c-Met COOH terminus creates a docking site that interacts with scaf- c-Met, a member of a receptor tyrosine kinase subfamily fold and signaling molecules. The resulting signaling complex that also includes Ron, is a transmembrane protein com- activates pathways such as PI3K-PDK1-AKT-mTor, Ras-Rac- posed of an extracellular α chain disulfide-bonded to a mem- Pak, Ras-Raf-MEK-ERK, and PLC-γ (2–4). In addition to pro- brane-spanning β chain. The tyrosine kinase resides in the liferative and antiapoptotic activities, HGF elicits motogenic cytosolic portion of the β chain. Binding of hepatocyte and morphogenic cellular phenotypes (2, 5). HGF is also growth factor (HGF), the physiologic ligand, to c-Met induces known to be an angiogenic factor (6). c-Met dimerization and trans-phosphorylation of two tyro- c-Met was first identified as oncogenic fusion Tpr-Met (7). sine residues (Y1234 and Y1235) within the activation loop Hereditary human papillary renal carcinoma is causally relat- of the c-Met kinase domain, triggering conformational ed to germline mutations in the c-Met kinase domain (8, 9). changes that upregulate its catalytic activity (1, 2). Subsequent Aberrant c-Met signaling, resulting from MET genomic amplification, c-Met or HGF overexpression, or c-Met mutations, is found in a variety of human cancers and often correlated with poor clinical outcomes (2, 10). Recently, constitutive c-Met ac- Authors' Affiliation: Merck Research Laboratories, Boston, tivation due to MET amplification was found to be a driver of Massachusetts proliferation and survival of several gastric and lung cancer Note: Supplementary data for this article are available at Cancer cell lines (11, 12) and has been linked to acquired resistance Research Online (http://cancerres.aacrjournals.org/). of lung cancers to epidermal growth factor receptor (EGFR) Corresponding Author: Bo-Sheng Pan, Department of In Vitro Sciences, Merck Research Laboratories, BMB-11, 33 Avenue Louis inhibitors (13, 14). Pasteur, Boston, MA 02115. Phone: 617-992-2039; Fax: 617-992-2487; c-Met is an attractive cancer drug target. Small-molecule E-mail: [email protected]. c-Met kinase inhibitors and antibodies targeting c-Met or doi: 10.1158/0008-5472.CAN-09-2541 HGF have exhibited antitumor activities in preclinical mod- ©2010 American Association for Cancer Research. els (15–23). Some of these agents are undergoing clinical

1524 Cancer Res; 70(4) February 15, 2010

MK-2461, a Novel Multitargeted Kinase Inhibitor

Figure 1. Differential effects of MK-2461 on phosphorylation of key tyrosine residues of c-Met. A, Western blots showing that MK-2461 was substantially more potent against in vitro ATP-induced autophosphorylation of the COOH-terminal docking site (Y1349 and Y1365) of c-Met than that of the c-Met activation loop (Y1230/Y1234/Y1235). B and C, Western blots showing that MK-2461 suppressed constitutive (B) and ligand-induced (C) phosphorylation of the juxtamembrane domain (Y1003) and the COOH-terminal docking site (Y1349/Y1365) of c-Met and phosphorylation of AKT (S473) and ERK1/2 (T202/Y204) without inhibiting c-Met activation loop (Y1234/35) phosphorylation in GTL-16 gastric cancer cells (B) and A549 lung cancer cells (C). D, the effects of MK-2461 (1 μmol/L) on c-Met phosphorylation in GTL-16 cells determined by SILAC mass spectrometry. Shown are the percent changes in abundance of the specified phospho-peptides in MK-2461 treated cells relative to vehicle-treated cells. development (22, 23). Here, we describe a novel multitar- School, Turin, Italy); OCUM-1 gastric cancer and EBC-1 geted kinase inhibitor that preferentially inhibits the acti- lung carcinoma cell lines were from Health Science vated c-Met. Research Resources Bank (Japan); Sum-52 breast cancer line was from Asterand, Inc.; NIH 3T3 cell lines harboring Materials and Methods c-Met mutations T3936C and T3997C were from L. Schmidt (National Cancer Institute, Bethesda, MD); and Madin- Recombinant kinases. The cytosolic domains of human Darby canine kidney (MDCK) type II cell line was from c-Met, c-Met mutants, Ron, TrkA, insulin-like growth factor Clontech. All other cell lines were from the American Type I receptor (IGF-IR), and Mer were expressed as NH2-terminal Culture Collection. glutathione S-transferase (GST) fusions using Baculogold Ba- Murine myeloid 32D cells harboring Tpr-Met fusions. culovirus Expression System (BD Bioscience). GST-tagged Tpr-Met and Tpr-Met (Y362) mutant were cloned into pLenti6/ KDR and EGFR kinase were from Keith Rickert of Merck Re- V5-DEST (Invitrogen) and transfected into 293FT packaging search Labs. His6-tagged Jak2, Flt1, Flt3, Flt4, fibroblast cells (Invitrogen) to produce lentiviruses. The lentiviruses growth factor receptor 1 (FGFR1), FGFR2, FGFR3, and TrkB were used to infect 32D cells, which were then passaged in in- kinases were from Upstate Biotechnology (Millipore). His6- terleukin 3 (IL-3)–free medium under blasticidin selection to tagged platelet-derived growth factor receptor β (PDGFRβ) achieve stable expression of the fusion proteins. and the c-Met cytosolic domain were from Invitrogen. Antibodies. The antibodies used included anti–phospho- Cell lines. The GTL-16 gastric cancer cell line was from c-Met (Y1365), anti–phospho-c-Met (Y1003), and anti–phospho- S. Giordano and P.M. Comoglio (University of Torino Medical c-Met (Y1230/Y1234/Y1235) from Biosource; anti–phospho-c-Met

www.aacrjournals.org Cancer Res; 70(4) February 15, 2010 1525

Pan et al.

(Y1234/Y1235), anti–phospho-c-Met (Y1349), anti–phospho- MK-2461 on the abundance of a phosphotyrosine residue(s) of FGFR (Y653/Y654), anti–phospho-PDGFR (Y857), anti–phospho- interest was gauged from ion-intensity ratio of the heavy (MK- AKT (S473), anti-phospho–extracellular signal-regulated kinase 2461–treated)– and light (vehicle-treated)–labeled tryptic pep- 1/2 (ERK1/2) (T202/Y204), anti–c-Met, anti-PDGFR, anti-AKT, tide containing the phosphotyrosine(s). The ratio was normal- and anti-ERK1/2 from Cell Signaling Technology; anti-FGFR2 ized using the heavy/light ratio of the self-to-self control. and anti-EGFR from R&D Systems; and anti–phospho-tyrosine Assessment of motogenic and morphogenic cellular phe- antibody 4G10 from Millipore. notypes. HGF-induced migration of HPAF II pancreatic can- In vitro kinase assays. c-Met–catalyzed phosphorylation cer cells was measured by a modified Boyden chamber method of N-biotinylated peptide (EQEDEPEGDYFEWLE-CONH2) using a FluoroBlock 96-Multiwell Insert System (BD Bio- was measured using a time-resolved fluorescence resonance sciences). The cells in DMEM plus 10% fetal calf serum were energy transfer assay adapted from Park and colleagues (24). plated into the upper chamber semisubmerged in the lower To evaluate kinase selectivity, a single concentration chamber containing the same medium. MK-2461 was added (1 μmol/L) of MK-2461 was tested using 216 kinases by Up- to both chambers and the plate was incubated for 2 h. HGF state Biotechnology, Inc. (Millipore), which also determined was then added to the lower chamber and the plate was incu- the MK-2461 IC50 for DRAK1, DYRK2, IRAK1, IRAK4, MELK, bated for an additional 20 h. Viable cells on the underside of and MLK1. The MK-2461 IC50 for Ron, Mer, Flt1, Flt3, Flt4, KDR, PDGFRβ, FGFR1, FGFR2, FGFR3, TrkA, and TrkB were determined using time-resolved fluorescence resonance energy transfer assays similar to the c-Met kinase assay. c-Met autophosphorylation assay. The c-Met cytosolic domain was preincubated with MK-2461. The autophosphorylation was initiated by addition of 50 μmol/L ATP and run for 30 min at 25°C. The reactions were stopped by a denaturing buffer and by boiling, and then the samples were subjected to Western blotting with antibodies targeting various phosphotyrosine residues of c-Met. Analysis of phosphorylation status of c-Met, FGFR2, PDGFR,EGFR,AKT,andERKincells.Tumor cells were treated for 2 h with MK-2461 or vehicle in RPMI 1640 sup- plemented with 10% fetal bovine serum and 10 mmol/L HEPES. When called for, the cells were stimulated with HGF or EGF during the last 10 min of the 2-h incubation. The cells were lysed with a denaturing or nondenaturing buffer containing phosphatase and protease inhibitors and subjected to Western blot or immunoprecipitation–Western blot analysis. Direct binding assay. The His6-tagged c-Met cytosolic domain (Invitrogen) was immobilized on two detection spots of a CM5 sensor chip in a BIAcore 51 instrument through an anti-His antibody and mild cross-linking. The c-Met on one spot was autophosphorylated by perfusion with MgATP. Binding kinetics of MK-2461 to the phosphorylated and unphosphorylated c-Met was then measured simultaneously using a protocol similar to that described by Nordin and colleagues (25). SILAC phosphoproteomic analysis of c-Met in GTL-16 cells. The cells were cultured for six passages in DMEM containing heavy arginine and heavy lysine (13C6-Arg and 13C6-Lys) or DMEM containing light (12C6) arginine and lysine. Heavy- Figure 2. Representative BIAcore data showing that MK-2461 binds labeledcellsweretreatedfor2hwith1μmol/L MK-2461, more strongly to phosphorylated c-Met than to unphosphorylated c-Met. The His6-tagged c-Met cytosolic domain was immobilized on two whereas the light-labeled cells were treated with the vehicle. detection spots of a sensor chip. One spot was perfused with ATP to c-Met was immunoprecipitated from a 1:1 mixture of the induce c-Met autophosphorylation. Both spots were then perfused with a heavy- and light-labeled cell lysates, gel-purified, tryptic di- MK-2461–containing buffer and then with a MK-2461–free buffer. gested, and subjected to liquid chromatography tandem mass A 2-fold dilution series of MK-2461 was tested. The surface plasmon spectrometry (LC-MS/MS) analysis as previously described resonance (SPR) signals recorded from the spot with unphosphorylated c-Met (A) and the spot with phosphorylated c-Met (B) are shown. The rate (26, 27). As a “self-to-self” control, a 1:1 mixture of lysates of constants kon and koff were obtained from fitting each data set globally vehicle-treated, heavy-labeled and vehicle-treated, light-la- with a 1:1 interaction model. The smooth lines represent the best fits. The K k k beled cells was subjected to the same procedures. The effect of equilibrium constant d was calculated as off/ on.

1526 Cancer Res; 70(4) February 15, 2010 Cancer Research

MK-2461, a Novel Multitargeted Kinase Inhibitor

the membrane were stained using Calcein AM (Invitrogen) and quantified using a fluorescence reader. HGF-induced scatter of HPAF II cells was measured as described previously (28). HGF-induced tubulogenesis of MDCK cells was moni- tored as described by Yamaguchi and colleagues (29). Proliferation and mitogenic assays. Proliferation and viability of tumor cells was measured using the ViaLight PLUS kit (Cambrex). HGF-dependent mitogenic activity of 4MBr monkey lung cells was measured by monitoring bro- modeoxyuridine (BrdUrd) incorporation into genomic DNA using the Roche Cell Proliferation ELISA, BrdUrd (chemilu- minescence) kit. Tumor xenograft models. GTL-16 cells or c-Met mutant- transformed NIH3T3 cells were inoculated s.c. into the flank of female nude CD-1 nu/nu mice. When mean tumor size reached a predetermined range, the mice were randomized and given vehicle or MK-2461 by p.o. gavage once or twice daily. Tumor volumes were determined using calipers. The percentage increase in the volume of a xenograft tumor on day n versus day 0 (the day when dosing of MK-2461 began) was calculated as (tumor volume on day n − tumor volume Figure 3. MK-2461 effectively inhibited phosphorylation of the activation on day 0) / tumor volume on day 0 × 100. The mean percentage loop of FGFR2 and PDGFR in cells. A, Western blots showing that of tumor growth inhibition in each MK-2461–treated group MK-2461 inhibited phosphorylation of the activation loop (Y653/Y654) relative to the vehicle-treated group was calculated as of FGFR2 in Kato III gastric cancer cells. B, Western blots showing (1 − mean percent increase of tumor volume in the MK- that MK-2461 inhibited phosphorylation of the activation loop (Y849) of – PDGFR in H1703 lung cancer cells. C, MK-2461 does not inhibit 2461 treated group/mean percent increase of the tumor EGF-induced phosphorylation of EGFR in A549 lung cancer cells. The volume in the vehicle-treated group) × 100. cells were incubated for 2 h with 2 μmol/L MK-2461, 2 μmol/L gefitinib Measurement of c-Met (Y1349) phosphorylation in or vehicle, stimulated with EGF for 10 min, and then lysed. EGFR xenograft tumors. Mice bearing GTL-16 tumors were eutha- was immunoprecipitated and subjected to Western blotting with an nized 1 h after p.o. administration of MK-2461. The tumors anti-phosphotyrosine antibody (4G10) and anti-total EGFR antibody. Note that 2 μmol/L gefitinib abrogated EGFR phosphorylation. were excised, snap-frozen, and dispersed using a Qiagen Tissue- Lyser in a nondenaturing lysis buffer containing protease and phosphatase inhibitors. The homogenate was lysed at 4°C the COOH-terminal docking site than the autophosphoryla- for 1 h, clarified by centrifugation, and then analyzed by quan- tion of the activation loop [IC50 was ∼900 nmol/L as deter- titative Western blotting for phospho–c-Met (Y1349) and total mined using an anti-pMet (Y1234/Y1235) antibody and was c-Met. The pMet (Y1349) signal of each c-Met band was nor- >2 μmol/L as determined using an anti-pMet (Y1230/1234/ malized with its total c-Met signal. To combine or compare 1235) antibody]. data from several gels, the pY1349/total Met ratio for each We next used GTL-16 gastric cancer cells, which harbor c-Met band was further normalized to the average pY1349/ MET genomic amplification and overexpress constitutively total c-Met ratio of the vehicle-treated tumor samples on activated c-Met protein, to assess the cellular activities of thesamegel. MK-2461 against phosphorylation of c-Met and its downstream signaling. MK-2461 potently inhibited the phosphorylation of three nonactivation loop tyrosine residues of c-Met, Results Y1003 in the juxtamembrane domain and Y1349 and Y1365 ≈ in the COOH-terminal docking site (IC50 50 nmol/L). This MK-2461 is a potent ATP-competitive kinase inhibitor was accompanied by a concentration-dependent inhibition that preferentially binds the activated c-Met. MK-2461 of phosphorylation of two key downstream signaling mole- (Supplementary Fig. S1) inhibited the kinase activity of cules, AKT (S473) and ERK1/2 (T202/Y204; Fig. 1B). In con- human c-Met with a mean IC50 of 2.5 nmol/L in the presence trast, MK-2461 had little effect on c-Met activation loop of 50 μmol/L ATP. Steady-state kinetics experiments (Sup- (Y1234/Y1235) phosphorylation, even at concentrations up plementary Fig. S2) showed that MK-2461 is competitive to 2.5 μmol/L (Fig. 1B). Very similar results were obtained from with ATP and noncompetitive with respect to the peptide A549 lung cancer cells, which were used to assess the effects of substrate. MK-2461 on HGF-induced phosphorylation of c-Met, AKT, and The effect of MK-2461 on ATP-induced autophosphoryla- ERK1/2 (Fig. 1C). tion of recombinant c-Met cytosolic domain was determined The effects of MK-2461 on c-Met phosphorylation in GTL- by Western blotting. As shown in Fig. 1A, MK-2461 was sub- 16 cells were further quantified using SILAC mass spectrom- stantially more potent against the autophosphorylation of etry. As shown in Fig. 1D, treatment of GTL-16 cells with μ Y1349 (IC50 ≈100 nmol/L) and Y1365 (IC50 ≈26 nmol/L) in 1 mol/L MK-2461 caused a 98% decrease in the abundance

www.aacrjournals.org Cancer Res; 70(4) February 15, 2010 1527

Pan et al.

of a c-Met COOH-terminal peptide dually phosphorylated increase (25%) in the abundance of the activation loop pep- at Y1349 and Y1356, and 83% decrease in the abundance of tide dually phosphorylated at Y1234 and Y1235. the same peptide singly phosphorylated at either Y1349 or The ability of MK-2461 to inhibit autophosphorylation of Y1356. This was accompanied by a 64% decrease in the abun- the juxtamembrane domain and COOH-terminal docking site dance of a peptide containing phosphorylated Y1003 of the of c-Met while remaining essentially inactive against dual ty- juxtamembrane domain and a 47% reduction in the abun- rosine (Y1234 and Y1235) phosphorylation of the c-Met acti- dance of an activation loop–derived peptide singly phosphor- vation loop is consistent with MK-2461 binding preferentially ylated at Y1234. In contrast MK-2461 caused a moderate to the activation loop–phosphorylated c-Met. To test this

Figure 4. MK-2461 inhibited HGF/ c-Met–dependent cellular phenotypes. A, MK-2461 inhibited HGF-induced scattering of HPAF II pancreatic cancer cells in a two-dimensional culture. The cells were stained with Hoechst 33342 and imaged on an INCell1000. B, MK-2461 inhibited HGF-induced tubulogenesis of MDCK cells in three-dimensional culture. Shown are the images of randomly selected MDCK cysts. C, the effects of MK-2461 on IL-3–independent proliferation of 32D cells transformed with Tpr-Met (○) or Tpr-Met (Y362C) mutant (▵), and on IL-3–depedent growth of parental 32D cells (▾). Each data point represents the average of duplicate determinations.

1528 Cancer Res; 70(4) February 15, 2010 Cancer Research

MK-2461, a Novel Multitargeted Kinase Inhibitor

model, the binding of MK-2461 to phosphorylated and unphosphorylated c-Met cytosolic domain was directly measured using BIAcore (Fig. 2). Two independent experiments showed that the average equilibrium dissociation con- K stant ( d) of MK-2461 for phosphorylated c-Met (4.0 nmol/L) is 6-fold lower than that for unphosphorylated c-Met (25.1 nmol/L). The higher affinity of MK-2461 for phosphorylated k c-Met resulted from faster association rate constant ( on) and slower dissociation rate constant (koff). These findings provided direct evidence that MK-2461 preferentially binds the activated conformations of c-Met. In vitro activity of MK-2461 against c-Met mutants. Several oncogenic c-Met mutations reportedly confer resistance to certain c-Met kinase inhibitors (15, 16, 30, 31). We thus compared in vitro potencies of MK-2461 against wild-type c-Met and five c-Met mutants, including three activation loop mutants (Y1230C, Y1230H, and Y1235D), a P + 1 loop mutant M1250T, and N1100Y, which is near the ATP binding site. In the presence of 50 μmol/L ATP, MK- 2461 is equally or more potent against the five mutants (mean IC50 ranging from 0.4 to 1.5 nmol/L) compared with the wild- type c-Met (mean IC50 = 2.5 nmol/L; Supplementary Table S1). Kinase selectivity of MK-2461. In vitro kinase assays revealed that Ron (IC50 = 7 nmol/L) and Flt1 (IC50 = 10 nmol/L) were inhibited by MK-2461 with similar potencies to c-Met (IC50 = 2.5 nmol/L), whereas nine other kinases, including FGFR1, FGFR2, FGFR3, PDGFRβ, KDR, Flt3, Flt4, TrkA, and TrkB, were found to be 8- to 30-fold less sensitive to MK-2461 than c-Met (Supplementary Table S1). We further evaluated the Figure 5. In vitro antiproliferative potencies of MK-2461 against a tumor cell line panel. The cells were treated in triplicate for 72 h with various effect of MK-2461 on autophosphorylation of FGFR2 and α concentrations of MK-2461. The numbers of viable cells were measured PDGFR- , respectively, in KATO III cells, which overexpress using Vialight Plus assay. The data were fitted with a four-parameter constitutively activated FGFR2 (32), and H1703 cells, which dose-response equation and the inflection point of the least-square fit overexpress constitutively activated PDGFR-α (33). MK-2461 curve was reported as IC50. Cells harboring MET gene amplification, potently inhibited phosphorylation of the activation loop of FGFR2 amplification, and constitutively activated PDGFR are labeled with a single (*), double (**), and triple asterisks (***), respectively. FGFR2 (Y653/Y654) and PDGFR-α (Y849) in the cells with IC50 <300 nmol/L (Fig. 3A and B). These observations contrast with the inability of MK-2461 to inhibit activation loop phos- To assess the cellular activity of MK-2461 against onco- phorylation in c-Met. Consistent with its lack of inhibitory genic c-Met activation loop mutants, we tested the effect of activity against recombinant EGFR kinase, MK-2461 did not MK-2461 on proliferation of 32D murine myeloid leukemia inhibit ligand-induced phosphorylation of EGFR in A549 lung cells transformed by Tpr-Met or Tpr-Met (Y362C) mutant, cancer cells (Fig. 3C). which is equivalent to the c-Met activation loop mutant Inhibition by MK-2461 of HGF/c-Met–dependent cellular Y1230C. The 32D/Tpr-Met and 32D/Tpr-Met (Y362C) cells phenotypes. MK-2461 inhibited HGF-induced mitogenesis acquired the ability to proliferate in the absence of IL-3, (measured as BrdUrd incorporation) of 4MBr-5 monkey lung which was needed to support proliferation and survival of epithelial cells (34) with an IC50 of 204 ± 44 nmol/L (mean ± SE, the parental 32D cells. As shown in Fig. 4C, MK-2461 potently n = 5). (The result of a representative experiment is shown inhibited IL-3–independent growth of 32D/Tpr-Met and ≈ in Supplementary Fig. S3.) The compound inhibited HGF- 32D/Tpr-Met (Y362C) cells (IC50 100 nmol/L). Importantly, induced scattering of HPAF II human pancreatic adenocar- MK-2461 was ∼100-fold less potent in inhibiting IL-3–dependent cinoma cells (35) in a dose-dependent manner (Fig. 4A), with proliferation of the parental 32D cells, indicating that the n a mean IC50 of 416 ± 26 nmol/L (SE, = 23) as determined by an potent antiproliferative effects of MK-2461 on the Tpr- automated imaging assay (28). Using a modified multiwell Met–expressing cells resulted from specific inhibition of Boyden chamber system, we found that MK-2461 inhibited Tpr-Met signaling. HGF-induced migration of HPAF II cells with a mean IC50 of Genomic amplification and constitutive activation of 404 ± 47 nmol/L (SE, n = 11). (A representative experiment is c-Met, FGFR2, or PDGFR are predictors of in vitro tumor shown in Supplementary Fig. S4.) MK-2461 inhibited, in a con- response to MK-2461. The potency of MK-2461 for inhib- centration-dependent manner, the HGF-induced branching iting in vitro proliferation and viability of tumor cells was tubulogenesis of MDCK cells in a three-dimensional collagen evaluated using a tumor cell line panel. The observed anti- gel matrix (Fig. 4B). proliferative potency of MK-2461 varied widely with mean

www.aacrjournals.org Cancer Res; 70(4) February 15, 2010 1529

Pan et al.

Figure 6. Orally administered MK-2461 inhibited c-Met phosphorylation and xenograft tumor growth in nude mice. A, inhibition by MK-2461 of c-Met (Y1349) phosphorylation in GTL-16 tumors. The tumor-bearing mice were given a single p.o. dose of 3, 10, 30, or 100 mg/kg of MK-2461 and euthanized 1 h after dosing. The extent of c-Met (Y1349) phosphorylation in the tumors was determined by quantitative Western blotting. Shown are representative images of pMet (Y1349) and total c-Met blots. Also shown is the relationship between the phospho–c-Met (Y1349) levels of all individual tumors and the corresponding host plasma total MK-2461 concentrations. The effects of MK-2461 on the growth of GTL-16 tumor (B) or NIH3T3 tumor harboring single-nucleotide c-Met mutation T3936C (C) or T3997C (D). Asterisks indicate statistically significant differences from vehicle-treated group based on two-way repeated measures ANOVA (*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001).

IC50 values ranging from 0.1 to >30 μmol/L (Fig. 5). The cell genomic amplification and constitutive activation of PDGFR lines were classified into three groups, described as being are predictive biomarkers of in vitro tumor response to ≤ μ μ sensitive (IC50 1 mol/L), moderately sensitive (1 mol/L < MK-2461. IC50 ≤ 5 μmol/L), and insensitive (IC50 >5 μmol/L) to MK- MK-2461 inhibited c-Met signaling and tumor growth in 2461. Among the 10 MK-2461–sensitive cell lines, GTL-16 tumor xenograft models in mice. To assess the in vivo po- and SNU5 gastric cancer cell lines and EBC-1 and H1993 tency of MK-2461 for inhibiting c-Met, the extent of c-Met non–small cell lung carcinoma cell lines were previously Y1349 phosphorylation in GTL-16 tumors was measured reported to overexpress constitutively activated c-Met pro- 1 hour after the tumor-bearing mice were given an oral bolus tein due to MET genomic amplification (11, 12, 36). In of MK-2461 at several dose levels. Plasma samples were addition, KATO III, SNU16 gastric cancer cell lines, and analyzed for total concentration of MK-2461. Dose-dependent SUM-52 breast cancer cell line were previously shown to over- inhibition of phosphorylation of tumor c-Met (Y1349) was express constitutively activated FGFR2 protein due to FGFR2 observed and found to be correlated with the concentration genomic amplification (32, 37). The H1703 non–small cell lung of MK-2461 in the plasma. From the relationship between the carcinoma cell line was confirmed, by Western blotting anal- plasma MK-2461 concentration and tumor phospho–c-Met in vivo ysis (Fig. 3B), to harbor a high level of constitutively activated (Y1349), the IC50 of MK-2461 for c-Met inhibition PDGFR as previously reported (33). Neither overexpression was estimated to be 1.0 μmol/L (Fig. 6A). nor constitutive activation of MET, FGFR, or PDGFR was de- Two independent efficacy studies were conducted to as- tected (data not shown) in the two remaining MK-2461–senstive sess the in vivo antitumor activity of MK-2461 in nude mice cell lines, KM12 colon cancer cell line and H1048 small-cell lung bearing established subcutaneous GTL-16 tumors. In the first carcinoma cell line. With the exception of the MKN-45 gastric experiment (Fig. 6B), mice were treated with vehicle or one of cancer line (IC50 =1.2μmol/L), known to harbor MET four 21-day p.o. regimens of MK-2461 [10, 50, and 100 mg/kg genomic amplification (36), all tumor cell lines classified twice daily (bid) and 200 mg/kg once daily], which inhibited as moderately sensitive or insensitive to MK-2461 did not tumor growth by 62%, 77%, 75%, and 90%, respectively. The show MET or FGFR2 amplification and express only low mean terminal tumor weights of the four MK-2461 cohorts levels of functional c-Met protein, which was not phosphor- were 427 ± 91 mg (SE, n = 10), 358 ± 81 mg (n = 12), 340 ± ylated under the condition of the proliferation assay (data 46 mg (n = 13), and 232 ± 52 mg (n = 13), compared with the not shown). These results show that c-Met or FGFR2 mean tumor weight of 1,129 ± 231 mg (n = 9) for the vehicle

1530 Cancer Res; 70(4) February 15, 2010 Cancer Research

MK-2461, a Novel Multitargeted Kinase Inhibitor

cohort. Statistical analyses showed that each of the four tors preferentially binding to the unactivated conformation, MK-2461 regimens had significant growth-inhibitory effect inhibitors preferentially binding to the active conformation compared with the vehicle. In the second efficacy study, face stronger competition from cellular ATP and are thus which evaluated five 21-day p.o. MK-2461 regimens (1, 3, likely to show a larger potency reduction in cell-based assays 10, 30, and 100 mg/kg bid), the 100 mg/kg bid regimen caused relative to cell-free assays. Imatinib, which preferentially essentially complete (99%) tumor growth inhibition, whereas binds an unactivated conformation of BCR-Abl, is known the other four MK-2461 regimens did not result in statistically to bind an active, DFG-in conformation of spleen tyrosine ki- significant tumor growth inhibition (Supplemental Fig. S5). nase, providing a precedent that a tyrosine kinase inhibitor All MK-2461 regimens tested in the two studies were well adopts different binding modes in different kinase contexts (44). tolerated, causing neither statistically significant loss in mean In a majority of receptor tyrosine kinases, including c-Met, body weight nor treatment-related death. the phosphotyrosine residues in the juxtamembrane domain We also tested a MK-2461 p.o. regimen (134 mg/kg bid) in and in the region COOH-terminal to the conserved kinase two mouse xenograft models derived from NIH3T3 cells domain seem to be primarily responsible for recruiting adap- transformed by c-Met single nucleotide mutants T3936C tor proteins and signaling molecules. However, the phos- and T3997C, respectively. The mutations are equivalent to phorylated activation loop of insulin receptor tyrosine Y1230H and M1250T mutations of human c-Met. Based on kinase has been shown to bind and signal through adaptor the tumor growth curves from day 0 to day 16, MK-2461 inhibited proteins APS and Grb10/14 (45, 46). MK-2461 does not block growth of T3936C and T3997C tumors by 78% (Fig. 6C) and activation loop phosphorylation of c-Met and would be inef- 62% (Fig. 6D), respectively, and the growth-inhibitory ef- fective in inhibiting signaling events that emanate from the fects of MK-2461 in both models were statistically significant. signaling or adaptor proteins that interact with the activa- These results showed the capacity of MK-2461 to inhibit c-Met tion loop of c-Met. Comparative studies using MK-2461 mutation–driven tumors in vivo. and c-Met inhibitors capable of inhibiting c-Met activation loop phosphorylation may allow identification of signaling molecules and pathways that are differentially modulated Discussion by c-Met inhibitors of different modes of action. MK-2461 was equally or slightly more potent against five Here, we have identified MK-2461, a novel ATP-competitive naturally occurring oncogenic human c-Met mutants multitargeted kinase inhibitor. Several lines of findings (N1100Y, Y1230C, Y1230H, Y1235D, and M1250T) compared strongly suggest that MK-2461 preferentially binds to the with wild-type c-Met and inhibited the growth of NIH3T3 activated c-Met with dual-phosphorylated activation loop tumors harboring murine equivalents of human c-Met and inhibits c-Met signaling by suppressing phosphorylation Y1230H and M1250T mutants. To date, about 15 c-Met ki- of tyrosine residues in the juxtamembrane domain and nase domain missense mutations have been detected in pap- COOH-terminal docking site of c-Met. These properties illary renal carcinomas, with the tyrosine residues Y1230 and distinguish MK-2461 from other known ATP-competitive Y1235 in the c-Met activation loop being “hotspots” for tyrosine kinase inhibitors, which either bind to the unacti- mutations (2). The missense mutations of Y1230 and vated and active kinase with similar affinity or preferentially Y1235 confer absolute or partial resistance to several bind the unactivated kinase (38, 39). Additional studies will be known c-Met inhibitors, including PF-04217903, a c-Met in- needed to decipher the molecular mechanism that allows the hibitor currently undergoing clinical development (15, 16, activation loop in MK-2461–bound c-Met molecules to re- 30, 31). A novel class of c-Met inhibitors represented by main fully phosphorylated in the face of a cytosolic environ- AM7 (16) exhibits uncompromised inhibitory activities ment rich in protein phosphatases. against these c-Met mutants. MK-2461, which is structur- Significantly, we found that MK-2461 effectively inhibited ally distinct from AM7, represents a new pharmacophore autophosphorylation of the activation loop of FGFR2 and with this desirable biological property. The c-Met inhibi- PDGFR in cells (Fig. 3). These observations suggest that tors potent against a broad spectrum of activating c-Met MK-2461 is capable of preventing activation of FGFR2 and mutants may hold promise for treatment of papillary renal PDGFR, presumably by binding to the unactivated conforma- cell carcinoma. tion of these kinases. The differences between the binding Genomic amplification of c-Met or FGFR2 occurs at higher modes of MK-2461 to c-Met and to FGFR2 may contribute frequencies in gastric cancers (47, 48) and is believed to be a to the observation that MK-2461 exhibits more potent anti- major driver of malignancy (11, 12, 32). A dual c-Met/FGFR2 proliferative activity in FGFR2-driven cells than in c-Met– inhibitor such as MK-2461 is, potentially, a promising thera- driven cells, despite in vitro kinase assays showing that peutic for the subset of gastric cancers that display either MK-2461 is more potent against c-Met than against FGFR2. Met or FGFR2 amplification. Structural studies of several unactivated tyrosine kinases re- In summary, MK-2461 is a novel orally available multi- vealed that parts of the activation loop and/or other struc- targeted kinase inhibitor, which inhibits c-Met signaling tural elements assume configurations that border or block and HGF/c-Met–dependent cellular phenotypes by selec- the ATP binding site (40–43). Therefore, the unactivated con- tively targeting the active conformation of c-Met. It also formation may have significantly lower affinity for ATP than exhibits significant activities toward several other oncology does the active conformation (38, 39). Compared with inhibi- kinase targets, including FGFR and PDGFR. It has signifi-

www.aacrjournals.org Cancer Res; 70(4) February 15, 2010 1531

Pan et al.

cant in vitro and in vivo antitumor activities against Acknowledgments c-Met– and FGFR2-driven tumors. The results described in this report provided justification for advancing MK-2461 to We thank Drs. Lex Van der Ploeg, Giulio Draetta, and Mark Goulet for their preclinical development. support. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in Disclosure of Potential Conflicts of Interest accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Received 7/15/09; revised 10/27/09; accepted 11/17/09; published No potential conflicts of interest were disclosed. OnlineFirst 2/9/10.

References 1. Longati P, Bardelli A, Ponzetto C, Naldini L, Comoglio PM. Tyro- [3,2-b]pyridin-7-yloxy)phenyl)-2-oxo-3-phenylimidazolidine-1-car- sines1234-1235 are critical for activation of the tyrosine kinase boxamides: a novel series of dual c-Met/VEGFR2 receptor tyrosine encoded by the MET proto-oncogene (HGF receptor). Oncogene kinase inhibitors. Bioorg Med Chem Lett 2009;19:1323–8. 1994;9:49–57. 19. Schroeder GM, An Y, Cai Z-W, et al. Discovery of N-(4-(2-Amino-3- 2. Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF. Met, me- chloropyridin-4-yloxy)-3-fluorophenyl)-4-ethoxy-1-(4-fluorophenyl)- tastasis, motility and more. Nat Rev Mol Cell Biol 2003;4:915–25. 2-oxo-1,2-dihydropyridine-3-carboxamide (BMS-777607), a selective 3. Benvenuti S, Comoglio PM. The MET receptor tyrosine kinase in in- and orally efficacious inhibitor of the Met kinase superfamily. J Med vasion and metastasis. J Cell Physiol 2007;213:316–25. Chem 2009;52:1251–4. 4. Ponzetto C, Bardelli A, Zhen Z, et al. A multifunctional docking site 20. Burgess T, Coxon A, Meyer S, et al. Fully human monoclonal anti- mediates signaling and transformation by the hepatocyte growth factor/ bodies to hepatocyte growth factor with therapeutic potential scatter factor receptor family. Cell 1994;77:261–71. against hepatocyte growth factor/c-Met-dependent human tumors. 5. Brinkmann V, Foroutan H, Sachs M, Weidner KM, Birchmeier W. Cancer Res 2006;66:1721–9. Hepatocyte growth factor/scatter factor induces a variety of tissue- 21. Martens T, Schmidt NO, Eckerich C, et al. A novel one-armed specific morphogenic programs in epithelial cells. J Cell Biol 1995; anti-c-Met antibody inhibits glioblastoma growth in vivo.Clin 131:1573–86. Cancer Res 2006;12:6144–52. 6. Bussolino F, Di Renzo MF, Ziche M, et al. Hepatocyte growth factor 22. Liu X, Yao W, Newton RC, Scherle PA. Targeting the c-MET signaling path- is a potent angiogenic factor which stimulates endothelial cell motil- way for cancer therapy. Expert Opin Investig Drugs 2008;17:997–1011. ity and growth. J Cell Biol 1992;119:629–41. 23. Migliore C, Giordano S. Molecular cancer therapy: can our expecta- 7. Gonzatti-Haces M, Seth A, Park M, Copeland T, Oroszlan S, Vande tion be MET? Eur J Cancer 2008;44:641–51. Woude GF. Characterization of the TPR-MET oncogene p65 and the 24. Park Y-W, Cummings RT, Wu L, et al. Homogeneous proximity tyro- MET protooncogene p140 protein-tyrosine kinases. Proc Natl Acad sine kinase assays: scintillation proximity assay versus homoge- Sci U S A 1988;85:21–5. neous time-resolved fluorescence. Anal Biochem 1999;269:94–104. 8. Schmidt L, Duh FM, Chen F, et al. Germline and somatic mutations in 25. Nordin H, Jungnelius M, Karlsson R, Karlsson OP. Kinetic studies of the tyrosine kinase domain of the MET proto-oncogene in papillary small molecule interactions with protein kinases using biosensor renal carcinomas. Nat Genet 1997;16:68–73. technology. Anal Biochem 2005;340:359–68. 9. Jeffers M, Fiscella M, Webb CP, Anver M, Koochekpour S, Vande 26. Meng FY, Wiener MC, Sachs JR, et al. Quantitative analysis of com- Woude GF. The mutationally activated Met receptor mediates motil- plex peptide mixtures using FTMS and differential mass spectrome- ity and metastasis. Proc Natl Acad Sci U S A 1998;95:14417–22. try. J Am Soc Mass Spectrom 2007;18:226–33. 10. Christensen JG, Burrows J, Salgia R. c-Met as a target for human 27. Yates JR III, Morgan SF, Gatlin CL, Griffin PR, Eng JK. Method to com- cancer and characterization of inhibitors for therapeutic intervention. pare collision-induced dissociation spectra of peptides: potential for li- Cancer Lett 2005;225:1–26. brary searching and subtractive analysis. Anal Chem 1998;70:3557–65. 11. Smolen GA, Sordella R, Muir B, et al. Amplification of MET may iden- 28. Chan GK, Lutterbach BA, Pan BS, Kariv I, Szewczak AA. High- tify a subset of cancers with extreme sensitivity to the selective ty- throughput analysis of HGF-stimulated cell scattering. J Biomol rosine kinase inhibitor PHA-665752. Proc Natl Acad Sci U S A 2006; Screen 2008;13:847–54. 103:2316–21. 29. Yamaguchi H, Miki H, Takenawa T. Neural Wiskott-Aldrich syndrome 12. Lutterbach B, Zeng Q, Davis LJ, et al. Lung cancer cell lines harbor- protein is involved in hepatocyte growth factor-induced migration, ing MET gene amplification are dependent on Met for growth and invasion, and tubulogenesis of epithelial cells. Cancer Res 2002;62: survival. Cancer Res 2007;67:2081–8. 2503–9. 13. Engelman JA, Zejnullahu K, Mitsudomi T, et al. MET amplification 30. Berthou S, Aebersold DM, Schmidt LS, et al. The Met kinase inhibitor leads to gefitinib resistance in lung cancer by activating ERBB3 sig- SU11274 exhibits a selective inhibition pattern toward different re- naling. Science 2007;316:1039–43. ceptor mutated variants. Oncogene 2004;23:5387–93. 14. Bean J, Brennan C, Shih JY, et al. MET amplification occurs with or 31. Timofeevski SL, Ryan K, Cui J, et al. Enzymatic characterization of c-Met without T790M mutations in EGFR mutant lung tumors with acquired receptor tyrosine kinase oncogenic mutants and kinetic studies with amino- resistance to gefitinib or erlotinib. Proc Natl Acad Sci U S A 2007; pyridine and triazolopyrazine inhibitors. Biochemistry 2009;48:5339–49. 104:20932–7. 32. Kunii K, Davis L, Gorenstein J, et al. FGFR2-amplified gastric cancer 15. Zou HY, Li QH, Lee JH, et al. An orally available small-molecule incell lines require FGFR2 and Erbb3 signaling for growth and survival. hibitor of c-met, PF-2341066, exhibits cytoreductive antitumor effi- Cancer Res 2008;68:2340–8. cacy through antiproliferative and antiangiogenic mechanisms. 33. Rikova K, Guo A, Zeng Q, et al. Global survey of phosphotyrosine signal- Cancer Res 2007;67:4408–17. ing identifies oncogenic kinases in lung cancer. Cell 2007;131:1190–203. 16. Bellon SF, Kaplan-Lefko P, Yang Y, et al. c-Met inhibitors with novel 34. Rafferty B, Maile P, Rigsby P, et al. International standards for hepa- binding mode show activity against several hereditary papillary renal tocyte growth factor/scatter factor: initial assessment of candidate cell carcinoma-related mutations. J Biol Chem 2008;283:2675–83. materials and their evaluation by multicentre collaborative study. 17. Zhang YH, Kaplan-Lefko PJ, Rex K, et al. Identification of a novel J Immunol Methods 2001;258:1–11. recepteur d'origine nantais/c-Met small-molecule kinase inhibitor 35. Di Renzo MF, Poulsom R, Olivero M, et al. Expression of the Met/ with antitumor activity in vivo. Cancer Res 2008;68:6680–7. hepatocyte growth factor receptor in human pancreatic cancer. Can- 18. Raeppel S, Claridge S, Saavedra O, et al. N-(3-fluoro-4-(2-arylthieno cer Res 1995;55:1129–38.

1532 Cancer Res; 70(4) February 15, 2010 Cancer Research

MK-2461, a Novel Multitargeted Kinase Inhibitor

36. Rege-Cambrin G, Scaravaglio P, Carozzi F, et al. Karyotypic analysis 43. Wang W, Marimuthu A, Tsai J, et al. Structural characterization of of gastric carcinoma cell lines carrying an amplified c-met oncogene. autoinhibited c-Met kinase produced by coexpression in bacteria Cancer Genet Cytogenet 1992;64:170–3. with phosphatase. Proc Natl Acad Sci U S A 2006;103:3563–8. 37. Moffa AB, Ethier SP. Differential signal transduction of alternatively 44. Atwell S, Adams JM, Badger J, et al. A novel mode of Gleevec bind- spliced FGFR2 variants expressed in human mammary epithelial ing is revealed by the structure of spleen tyrosine kinase. J Biol cells. J Cell Physiol 2007;210:720–31. Chem 2004;279:55827–32. 38. Liu Y, Gray NS. Rational design of inhibitors that bind to inactive ki- 45. Hubbard SR, Miller WT. Receptor tyrosine kinases: mechanisms of nase conformations. Nat Chem Biol 2006;2:358–64. activation and signaling. Curr Opin Cell Biol 2007;19:117–23. 39. Knight ZA, Shokat KM. Features of selective kinase inhibitors. Chem 46. Nouaille S, Blanquart C, Zilberfarb V, et al. Interaction with Grb14 Biol 2005;12:621–37. results in site-specific regulation of tyrosine phosphorylation of the 40. Blume-Jensen P, Hunter T. Oncogenic kinase signaling. Nature insulin receptor. EMBO Rep 2006;7:512–8. 2001;411:355–65. 47. Nakajima M, Sawada H, Yamada Y, et al. The prognostic significance 41. Hubbard SR. Protein tyrosine kinases: autoregulation and small- of amplification and overexpression of c-met and c-erb B-2 in human molecule inhibition. Curr Opin Struct Biol 2002;12:735–41. gastric carcinomas. Cancer 1999;85:1894–902. 42. Hubbard SR. Juxtamembrane autoinhibition in receptor tyrosine ki- 48. Hattori Y, Itoh H, Uchino S, et al. Immunohistochemical detection of nases. Nat Rev Mol Cell Biol 2004;5:464–71. K-sam protein in stomach cancer. Clin Cancer Res 1996;2:1373–81.

www.aacrjournals.org Cancer Res; 70(4) February 15, 2010 1533

MK-2461, a Novel Multitargeted Kinase Inhibitor, Preferentially Inhibits the Activated c-Met Receptor

Bo-Sheng Pan, Grace K.Y. Chan, Melissa Chenard, et al.

Cancer Res Published OnlineFirst February 9, 2010.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-09-2541

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2010/02/08/0008-5472.CAN-09-2541.DC1

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://cancerres.aacrjournals.org/content/early/2010/02/09/0008-5472.CAN-09-2541. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.