Structural Basis of AZD9291 Selectivity for EGFR T790M.

AZD9291 (Osimertinib) is highly effective in treating EGFR-mutated non-small-cell lung cancers (NSCLCs) with T790M-mediated drug resistance. Despite the remarkable success of AZD9291, its binding pose with EGFR T790M remains unclear. Here, we report unbiased, atomic-level molecular dynamics (MD) simulations in which spontaneous association of AZD9291 with EGFR kinases having WT and T790M mutant gatekeepers was observed. Simulation-generated structural models suggest that the binding pose of AZD9291 with T790M differs from its binding pose with the WT, and that AZD9291 interacts extensively with the gatekeeper residue (Met 790) in T790M but not with Thr 790 in the WT, which explains why AZD9291 binds T790M with higher affinity. The MD simulation-generated models were confirmed by experimentally determined EGFR/T790M complex crystal structures. This work may facilitate the rational design of drugs that can overcome resistance mutations to AZD9291, and more generally it suggests the potential of using unbiased MD simulation to elucidate small-molecule binding poses.

Structural insights into drug development strategy targeting EGFR T790M/C797S.

Treatment of non-small-cell lung cancers (NSCLCs) harboring primary EGFR oncogenic mutations such as L858R and exon 19 deletion delE746_A750 (Del-19) using gefitinib/erlotinib ultimately fails due to the emergence of T790M mutation. Though WZ4002/CO-1686/AZD9291 are effective in overcoming EGFR T790M by targeting Cys797 via covalent bonding, their efficacy is again limited due to the emergence of C797S mutation. New agents effectively inhibiting EGFR T790M without covalent linkage through Cys 797 may solve this problem. We presented here crystal structures of EGFR activating/drug-resistant mutants in complex with a panel of reversible inhibitors along with mutagenesis and enzyme kinetic data. These data revealed a previously un-described hydrophobic clamp structure in the EGFR kinase which may be exploited to facilitate development of next generation drugs targeting EGFR T790M with or without concomitant C797S. Interestingly, mutations in the hydrophobic clamp that hinder drug binding often also weaken ATP binding and/or abolish kinase activity, thus do not readily result in resistance to the drugs.

Discovery of 4-((N-(2-(dimethylamino)ethyl)acrylamido)methyl)-N-(4-methyl-3-((4-(pyridin-3-yl)pyrimidin-2-yl)amino)phenyl)benzamide (CHMFL-PDGFR-159) as a highly selective type II PDGFRα kinase inhibitor for PDGFRα driving chronic eosinophilic leukemia.

Through exploration of the non-highly conserved allosteric hydrophobic pocket generated by DFG-out shifting in the inactive conformation, we discovered a highly selective type II PDGFRα kinase inhibitor 15i (CHMFL-PDGFRα-159), which exhibited strong potency against purified PDGFRα (IC50: 132 nM) but not structurally similar PDGFRß, ABL, c-KIT and VEGFR2 kinases. In addition, it displayed a high selectivity profile (S score (10) = 0.02) at the concentration of 1 µM among 468 kinases/mutants in the KINOMEscan profiling. X-ray crystal structure of 15i in complex with PDGFRα revealed a distinct binding feature in the allosteric hydrophobic pocket which might help to expand the diversity of type II kinase inhibitors. Compound 15i potently inhibited the proliferation of PDGFRα driving Chronic Eosinophilic Leukemia (CEL) cell line EOL-1 through strong blockage of PDGFRα mediated signaling pathways, arresting cell cycle progression, and induction of apoptosis. Furthermore, compound 15i effectively suppressed the EOL-1 tumor progression in the xenograft model and increased the survival rate in the engraftment tumor model.

Structural basis of mutant-selectivity and drug-resistance related to CO-1686.

Non-small-cell lung cancers (NSCLCs) caused by activating mutations in the kinase domain of epidermal growth factor receptor (EGFR) initially respond to first-generation reversible drugs gefitinib and erlotinib. However, clinical efficacy is limited due to the development of drug-resistance that in more than half of the cases are driven by the secondary T790M mutation. CO-1686 is one of the third generation irreversible inhibitors that inhibits EGFR activating mutants, including those with concurrent T790M, while avoiding the off-target toxicity owing to inhibition of wild-type EGFR in treating EGFR mutation-positive NSCLCs. Despite the remarkable success, the experimentally determined structure of this agent in complex with EGFR T790M remains unknown. In this study, we determined crystal structures of EGFR T790M or L858R mutants covalently bound by CO-1686. Based on these structural data, we can explain why CO-1686 irreversibly inhibits EGFR and selectively prefers T790M, which may help improving this or similar compounds, and explain why EGFR L718Q and L844V mutations incur resistance to this agent.

Structural pharmacological studies on EGFR T790M/C797S.

Drug-resistance is a major challenge in targeted therapy of EGFR mutated non-small cell lung cancers (NSCLCs). The third-generation irreversible inhibitors such as AZD9291, CO-1686 and WZ4002 can overcome EGFR T790M drug-resistance mutant through covalent binding through Cys 797, but ultimately lose their efficacy upon emergence of the new mutation C797S. To develop new reversible inhibitors not relying on covalent binding through Cys 797 is therefore urgently demanded. Gö6976 is a staurosporine-like reversible inhibitor targeting T790M while sparing the wild-type EGFR. In the present work, we reported the complex crystal structures of EGFR T790M/C797S + Gö6976 and T790M + Gö6976, along with enzyme kinetic data of EGFR wild-type, T790M and T790M/C797S. These data showed that the C797S mutation does not significantly alter the structure and function of the EGFR kinase, but increases the local hydrophilicity around residue 797. The complex crystal structures also elucidated the detailed binding mode of Gö6976 to EGFR and explained why this compound prefers binding to T790M mutant. These structural pharmacological data would facilitate future drug development studies.

Discovery and characterization of a novel irreversible EGFR mutants selective and potent kinase inhibitor CHMFL-EGFR-26 with a distinct binding mode.

EGFR T790M mutation accounts for about 40-55% drug resistance for the first generation EGFR kinase inhibitors in the NSCLC. Starting from ibrutinib, a highly potent irreversible BTK kinase inhibitor, which was also found to be moderately active to EGFR T790M mutant, we discovered a highly potent irreversible EGFR inhibitor CHMFL-EGFR-26, which is selectively potent against EGFR mutants including L858R, del19, and L858R/T790M. It displayed proper selectivity window between the EGFR mutants and the wide-type. CHMFL-EGFR-26 exhibited good selectivity profile among 468 kinases/mutants tested (S score (1)=0.02). In addition, X-ray crystallography revealed a distinct "DFG-in" and "cHelix-out" inactive binding mode between CHMFL-EGFR-26 and EGFR T790M protein. The compound showed highly potent anti-proliferative efficacy against EGFR mutant but not wide-type NSCLC cell lines through effective inhibition of the EGFR mediated signaling pathway, induction of apoptosis and arresting of cell cycle progression. CHMFL-EGFR-26 bore acceptable pharmacokinetic properties and demonstrated dose-dependent tumor growth suppression in the H1975 (EGFR L858R/T790M) and PC-9 (EGFR del19) inoculated xenograft mouse models. Currently CHMFL-EGFR-26 is undergoing extensive pre-clinical evaluation for the clinical trial purpose.

Discovery of (R)-1-(3-(4-Amino-3-(3-chloro-4-(pyridin-2-ylmethoxy)phenyl)-1H-pyrazolo[3,4-d]pyrimidin-1-yl)piperidin-1-yl)prop-2-en-1-one (CHMFL-EGFR-202) as a Novel Irreversible EGFR Mutant Kinase Inhibitor with a Distinct Binding Mode.

On the basis of Ibrutinib's core pharmacophore, which was moderately active to EGFR T790M mutant, we discovered novel epidermal growth factor receptor (EGFR) inhibitor compound 19 (CHMFL-EGFR-202), which potently inhibited EGFR primary mutants (L858R, del19) and drug-resistant mutant L858R/T790M. Compound 19 displayed a good selectivity profile among 468 kinases/mutants tested in the KINOMEscan assay (S score (1) = 0.02). In particular, it did not exhibit apparent activities against INSR and IGF1R kinases. The X-ray crystal structure revealed that this class of inhibitors formed a covalent bond with Cys797 in a distinct "DFG-in-C-helix-out" inactive EGFR conformation. Compound 19 displayed strong antiproliferative effects against EGFR mutant-driven nonsmall cell lung cancer (NSCLC) cell lines such as H1975, PC9, HCC827, and H3255 but not the wild-type EGFR expressing cells. In the H1975 and PC9 cell-inoculated xenograft mouse models, compound 19 exhibited dose-dependent tumor growth suppression efficacy without obvious toxicity. Compound 19 might be a potential drug candidate for EGFR mutant-driven NSCLC.

A structure-guided optimization of pyrido[2,3-d]pyrimidin-7-ones as selective inhibitors of EGFRL858R/T790M mutant with improved pharmacokinetic properties.

Structural optimization of pyrido[2,3-d]pyrimidin-7-ones was conducted to yield a series of new selective EGFRT790M inhibitors with improved pharmacokinetic properties. One of the most promising compound 9s potently suppressed EGFRL858R/T790M kinase and inhibited the proliferation of H1975 cells with IC50 values of 2.0 nM and 40 nM, respectively. The compound dose-dependently induced reduction of the phosphorylation of EGFR and downstream activation of ERK in NCIH1975 cells. It also exhibited moderate plasma exposure after oral administration and an oral bioavailability value of 16%. Compound 9s may serve as a promising lead compound for further drug discovery overcoming the acquired resistance of non-small cell lung cancer (NSCLC) patients.

Ibrutinib targets mutant-EGFR kinase with a distinct binding conformation.

Ibrutinib, a clinically approved irreversible BTK kinase inhibitor for Mantle Cell Lymphoma (MCL) and Chronic Lymphocytic Leukemia (CLL) etc, has been reported to be potent against EGFR mutant kinase and currently being evaluated in clinic for Non Small Cell Lung Cancer (NSCLC). Through EGFR wt/mutant engineered isogenic BaF3 cell lines we confirmed the irreversible binding mode of Ibrutinib with EGFR wt/mutant kinase via Cys797. However, comparing to typical irreversible EGFR inhibitor, such as WZ4002, the washing-out experiments revealed a much less efficient covalent binding for Ibrutinib. The biochemical binding affinity examination in the EGFR L858R/T790M kinase revealed that, comparing to more efficient irreversible inhibitor WZ4002 (Kd: 0.074 µM), Ibrutinib exhibited less efficient binding (Kd: 0.18 µM). An X-ray crystal structure of EGFR (T790M) in complex with Ibrutinib exhibited a unique DFG-in/c-Helix-out inactive binding conformation, which partially explained the less efficiency of covalent binding and provided insight for further development of highly efficient irreversible binding inhibitor for the EGFR mutant kinase. These results also imply that, unlike the canonical irreversible inhibitor, sustained effective concentration might be required for Ibrutinib in order to achieve the maximal efficacy in the clinic application against EGFR driven NSCLC.

Structural and biochemical studies of the PDGFRA kinase domain.

Platelet-derived growth factor receptor α (PDGFRA) is a Type III receptor tyrosine kinase, and this kinase is a target for treatment of gastrointestinal stromal tumors (GIST) as it is frequently mutated in these cancers. Most of the mutations that cause constitutive activation of PDGFRA occur in either the activation loop (A-loop) or in the juxtamembrane (JM) domain, such as the mutations D842V or V561D respectively. Treatment of PDGFRA-mutated GIST with imatinib is successful in some cases, but the D842V mutation is imatinib-resistant. To better understand the mechanism of PDGFRA drug-resistance, we have determined the crystal structure of the PDGFRA kinase domain in the auto-inhibited form, and studied the kinetics of the D842V mutation. Auto-inhibited PDGFRA is stabilized by the JM domain, which inserts into the active site of the kinase. The conserved residue Asp842 makes extensive contacts with several A-loop residues to maintain PDGFRA in the "DFG out" conformation, which stabilizes the kinase in the inactive state and facilitates the binding of imatinib. The D842V mutation would therefore be expected to activate the kinase and hinder the binding of drug through destabilizing the "DFG out" conformation. Furthermore, our kinetic data show that drug resistance in the D842V mutation may also in part result from its increased affinity for ATP. The PDGFRA kinase domain structure we report in this study has potential to facilitate development of new agents which can inhibit this kinase, including both its activating and drug-resistant mutations.

Ibrutinib selectively and irreversibly targets EGFR (L858R, Del19) mutant but is moderately resistant to EGFR (T790M) mutant NSCLC Cells.

Through comprehensive comparison study, we found that ibrutinib, a clinically approved covalent BTK kinase inhibitor, was highly active against EGFR (L858R, del19) mutant driven NSCLC cells, but moderately active to the T790M 'gatekeeper' mutant cells and not active to wild-type EGFR NSCLC cells. Ibrutinib strongly affected EGFR mediated signaling pathways and induced apoptosis and cell cycle arrest (G0/G1) in mutant EGFR but not wt EGFR cells. However, ibrutinib only slowed down tumor progression in PC-9 and H1975 xenograft models. MEK kinase inhibitor, GSK1120212, could potentiate ibrutinib's effect against the EGFR (L858R/T790M) mutation in vitro but not in vivo. These results suggest that special drug administration might be required to achieve best clinical response in the ongoing phase I/II clinical trial with ibrutinib for NSCLC.