KIT Kinase Mutants Show Unique Mechanisms of Drug Resistance to Imatinib and Sunitinib in Gastrointestinal Stromal Tumor Patients
Total Page:16
File Type:pdf, Size:1020Kb
KIT kinase mutants show unique mechanisms of drug resistance to imatinib and sunitinib in gastrointestinal stromal tumor patients Ketan S. Gajiwalaa,1, Joe C. Wub,1, James Christensenc, Gayatri D. Deshmukhb, Wade Diehla, Jonathan P. DiNittob, Jessie M. Englishb,2, Michael J. Greiga, You-Ai Hea, Suzanne L. Jacquesb, Elizabeth A. Lunneya,3, Michele McTiguea, David Molinaa,4, Terri Quenzera, Peter A. Wellsd, Xiu Yua, Yan Zhangb, Aihua Zoud, Mark R. Emmette,f, Alan G. Marshalle,f, Hui-Min Zhange,g, and George D. Demetrih,3 aDepartments of Structural and Computational Biology, cCancer Biology, and dBiochemical Pharmacology, Pfizer Global Research and Development, 10777 Science Center Drive, La Jolla, CA 92121; bDepartment of Enzymology and Biochemistry, Pfizer Research Technology Center, 620 Memorial Drive, Cambridge, MA 02139; eNational High Magnetic Field Laboratory, Florida State University, 1800 East Paul Dirac Drive, Tallahassee, FL 32310; fDepartment of Chemistry and Biochemistry, Florida State University, Tallahassee, FL, 32306; gInstitute of Molecular Biophysics, Florida State University, Kasha Laboratory, MC4380, Tallahassee, FL 32306; and hLudwig Center at Dana–Farber/Harvard Cancer Center, 44 Binney Street, Boston, MA 02115 Communicated by Joseph Schlessinger, Yale University School of Medicine, New Haven, CT, December 8, 2008 (received for review September 24, 2008) Most gastrointestinal stromal tumors (GISTs) exhibit aberrant ac- T670I. However, certain imatinib-resistant mutants are also resis- tivation of the receptor tyrosine kinase (RTK) KIT. The efficacy of tant to sunitinib, including D816H/V (6), which are located in the the inhibitors imatinib mesylate and sunitinib malate in GIST activation loop (A-loop) of the KIT catalytic domain (Fig. 1A). patients has been linked to their inhibition of these mutant KIT Previous structural studies (7, 8) had shown that KIT can sample proteins. However, patients on imatinib can acquire secondary KIT diverse conformations (Fig. 1B): the unactivated, autoinhibited mutations that render the protein insensitive to the inhibitor. conformation [in which the Asp-810-Phe-811-Gly-812 (DFG) triad Sunitinib has shown efficacy against certain imatinib-resistant at the beginning of the A-loop is in the ‘‘DFG-out’’ orientation, with mutants, although a subset that resides in the activation loop, the DFG Phe oriented near the ATP-binding pocket, and the JM including D816H/V, remains resistant. Biochemical and structural domain bound in the pocket vacated by DFG Phe], the unactivated studies were undertaken to determine the molecular basis of DFG-out conformation with the JM domain oriented into the sunitinib resistance. Our results show that sunitinib targets the solvent, and the activated conformation (in which the canonical autoinhibited conformation of WT KIT and that the D816H mutant ‘‘DFG-in’’ conformation buries the Phe away from the ATP- undergoes a shift in conformational equilibrium toward the active binding pocket and the A-loop extends over the C terminus of the catalytic domain). The protein can be considered to be in equilib- state. These findings provide a structural and enzymologic expla- rium among these conformations, with a shift to the activated form nation for the resistance profile observed with the KIT inhibitors. upon phosphorylation. Prospectively, they have implications for understanding oncogenic The mechanism of KIT resistance to sunitinib could not be kinase mutants and for circumventing drug resistance. deduced in modeling studies using the published structures de- scribed above (7, 8). We therefore sought to investigate the mech- kinase inhibitor ͉ signal transduction ͉ targeted therapy ͉ anism using structural biological and functional enzymology studies resistance mechanism ͉ cancer of WT and mutated KIT proteins comprising the JM and kinase domains. cquired resistance to systemic therapy is a critical problem in Atreating metastatic cancers. Improved understanding of the Results molecular mechanisms underlying resistance should provide in- Sunitinib and Imatinib Prevent KIT Activation by Targeting the Unac- sights leading to development of alternative treatment strategies or tivated Conformation of KIT. Sunitinib and imatinib are more potent design of new therapeutic entities that could be used to circumvent inhibitors of unactivated WT KIT than fully activated, phosphor- desensitization. One example of acquired resistance is the second- ary mutants of KIT identified in GIST. Most GISTs have primary Author contributions: K.S.G., J.C.W., J.C., W.D., J.P.D., J.M.E., M.J.G., Y.-A.H., S.L.J., E.A.L., activating mutations in the genes encoding the closely related RTKs M.M., P.A.W., X.Y., M.R.E., and G. D. Demetri designed research; K.S.G., J.C.W., G. D. Desh- KIT (Ϸ85% of GIST patients) or platelet-derived growth factor mukh, W.D., J.P.D., Y.-A.H., S.L.J., T.Q., P.A.W., Y.Z., A.Z., H.-M.Z., and G. D. Demetri receptor-␣ (PDGFRA; Ϸ5% of patients) (1). The majority of KIT performed research; W.D., D.M., and A.Z. contributed new reagents/analytic tools; K.S.G., mutations affect the juxtamembrane (JM) region of the protein J.C.W., J.P.D., J.M.E., S.L.J., E.A.L., M.M., T.Q., P.A.W., X.Y., Y.Z., A.Z., M.R.E., A.G.M., H.-M.Z., and G. D. Demetri analyzed data; and K.S.G., J.C.W., J.C., J.P.D., M.J.G., E.A.L., P.A.W., X.Y., encoded by exon 11 of the gene, such as V560D (Fig. 1A). Imatinib M.R.E., A.G.M., H.-M.Z., and G. D. Demetri wrote the paper. mesylate (Gleevec™; an inhibitor of KIT and PDGFRs, as well as Conflict of interest statement: K.S.G., J.C.W., J.C., G. D. Deshmukh, W.D., J.P.D., J.M.E., BCR-ABL) is currently first-line treatment for advanced GIST. M.J.G., Y-A.H., S.L.J., E.A.L., M.M., D.M., T.Q., P.A.W., X.Y., Y.Z., and A.Z. are Pfizer Clinical response and duration of clinical benefit with imatinib employees and stockholders or were at the time this work was carried out. M.R.E. holds correlate with KIT and PDGFRA genotype in GIST (2). Unfortu- Pfizer stock. G. D. Demetri has received consulting stipends and honoraria from Pfizer, Novartis, Bayer, and Infinity, and his institution has received research funding from Pfizer, nately, the majority of patients eventually show resistance to the Novartis, Bristol-Meyers Squibb, Infinity, and Exelixis. drug: Ϸ14% of patients are initially insensitive to imatinib, and 1K.S.G. and J.C.W. contributed equally to this work. Ϸ50% of patients develop resistance within 2 years (3, 4). The latter 2Present address: Department of Oncology, Merck Research Laboratories, 33 Avenue Louis resistance commonly occurs via secondary gene mutations in the Pasteur, Boston, MA 02115. KIT TK domains (Fig. 1A) (5). An effective second-line treatment 3To whom correspondence may be addressed. E-mail: [email protected] or is provided by sunitinib malate (Sutent™), which is approved beth.lunney@pfizer.com. multinationally for the treatment of advanced GIST after failure of 4Present address: Strain Development, Sapphire Energy, 3115 Merryfield Row, San Diego, imatinib due to resistance or intolerance. Sunitinib is an inhibitor CA 92121. of multiple RTKs, notably in this context, KIT and PDGFRA, and This article contains supporting information online at www.pnas.org/cgi/content/full/ has been shown to be effective against certain imatinib-resistant 0812413106/DCSupplemental. KIT mutants, such as the ATP-binding-pocket mutants V654A and © 2009 by The National Academy of Sciences of the USA 1542–1547 ͉ PNAS ͉ February 3, 2009 ͉ vol. 106 ͉ no. 5 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0812413106 Downloaded by guest on September 30, 2021 Fig. 1. Overview of the KIT gene and protein. (A) Schematic representation of KIT showing location of functional domains, primary (1°) and secondary (2°) mutations (mut.). Frequencies of primary KIT genotypes, specific secondary KIT mutations, and resistance (R) or sensitivity (S) to imatinib (IM) or sunitinib (SU) were those reported in a phase I/II trial of sunitinib in advanced GIST after imatinib failure (6). V560D, substitution of Asp for Val at residue 560. (B) The unactivated, autoinhibited and activated forms of WT KIT (7, 8). The JM domain (red), A-loop (green), and C ␣-helix (cyan) are oriented differently in the autoinhibited and activated states. *V560D generally occurs as a primary mutation. ylated WT KIT (Fig. 2). This was determined initially in biochem- containing primary and secondary mutations observed in the clinic ical experiments with activated WT KIT. In enzyme assays con- (V560D ϩ D816H, V560D ϩ D816V, and V560D ϩ V654A) in taining 1.2 mM ATP and activated WT KIT, sunitinib and imatinib sufficient quantity for structural experiments were unsuccessful. were only weakly active against the enzyme (IC50s: 7.0 and 3.3 M, However, crystallography-quality D816H protein was made in respectively). The unactivated enzyme state was then assayed: sufficient amounts for structural studies. This mutant, along with Unactivated WT KIT (40 nM) was incubated with sunitinib or the D816V and V654A single mutants, was studied biochemically, imatinib and 4 mM ATP for 16 h, followed by measurement of providing insight into the impact of point mutations. Additionally, kinase activity to quantify the amount of activated KIT produced. the clinically relevant mutants V560D and V560D ϩ T760I were Both inhibitors were very effective at inhibiting autoactivation even characterized and assayed. Biochemical studies showed that once at physiological ATP concentration (4 mM; Fig. 2) (IC50s: sunitinib, 42 nM; imatinib, 124 nM). Because imatinib has been shown to bind fully activated, all KIT mutants tested were virtually indistinguish- to KIT in the DFG-out, unactivated kinase conformation (8), it was able from WT (Table 1): Their catalytic efficiencies (kcat/Km) expected to be more effective against unactivated KIT than acti- were all within 3-fold of WT KIT.