Discovery of aminopyridine-containing spiro derivatives as EGFR mutations inhibitors

Abstract Neratinib is an oral pan HER inhibitor, that irreversibly inhibits EGFR and HER2 and was proven to be effective against multiple EGFR mutations. In previous study, we reported spiro [indoline-3, 4′-piperidine]-2-ones as anticancer agents. In this study, we designed aminopyridine-containing spiro [indoline-3,4′-piperidine] derivatives A1-A4 using Neratinib and spiro [indoline-3, 4′-piperidine]-2-one compound patented as lead structure, then replaced piperidine with cyclopropane to obtain B1-B7 and replaced indoline with benzmorpholine to get C1-C4 and D1-D2. We synthesized these compounds and evaluated their residual activities under 0.5 M drug concentration on EGFR and ERBB2. Most of compounds showed stronger inhibition on EGFR-wt and ERBB2, in which A1-A4 showed excellent inhibitory activity with inhibition percentage on EGFR-wt kinase of 7%, 6%, 19%, 27%, respectively and 9%, 5%, 12%, 34% on ERBB2 kinase compared with 2% and 6% of Neratinib.


Introduction
Overexpression of epidermal growth factor receptor (EGFR or HER1) and human epidermal growth factor receptor (HER2) is frequently found in different solid tumors. Coexpression of EGFR and HER2 has been reported in different tumors. They are also accompanied with postoperative adverse, radiotherapy, and chemotherapy resistance [1][2][3] . Therefore, it is more effective to dual target EGFR/HER2 rather than just EGFR inhibition 4 . Targeting of EGFR and HER2 is a proven anti-cancer strategy 5 . Both kinases have been an attractive therapeutic targets for cancer therapy. A variety of ATP-competitive EGFR/HER2 RTK dual inhibitors related to different scaffolds have been reported and many of them are currently in market or clinical trials for the treatment of cancer [6][7][8] .
Neratinib (HKI-272) is a highly selective inhibitor of HER2 and EGFR that leads to reduced phosphorylation and activation of downstream signaling pathways 9,10 . Neratinib was proven to be effective against multiple EGFR mutations including T790 mutation, L858R mutation and T790/L858. In complex with Neratinib, the EGFR kinase adopts an inactive conformation in which the regulatory C-helix is displaced from its active position. The enlarged hydrophobic pocket created by the outward rotation of the C-helix appears to be required to accommodate the bulky aniline substituent found in Neratinib, which contain additional aromatic groups appended to the 2-pyridinyl group and binds the inactive conformation of the kinase 11,12 . Neratinib is currently being tested in a number of clinical trials for its safety and efficacy in lung cancer, and colorectal, bladder, and breast cancers [13][14][15] .
In previous study, we reported spiro [indoline-3, 4 0 -piperidine]-2-ones as anticancer agents and several compounds showed stronger inhibition on tyrosine kinase 16 . In this study, we designed aminopyridine-containing spiro [indoline-3, 4 0 -piperidine] derivatives A1-A4 using neratinib and spiro [indoline-3, 4 0 -piperidine]-2one compound (a) patented as lead structure, then replaced piperidine with cyclopropane to obtained B1-B7 and replaced indoline with benzmorpholine to get C1-C4 ( Figure 1 and Tables  1-3). We synthesized these compounds and evaluated their residual activities under 0.5 M drug concentration on EGFR-wt and ERBB2. Then, we performed head to head comparative trial with neratinib to investigate inhibitory effects of compounds (A1-A2) with the best activity on EGFR-wt, HER2 and EGFR mutations. Molecular docking was adopted for all the synthesized compounds to confirm their mechanism of action.

Biological evaluation
2.2.1. EGFR-wt/HER2 protein kinase assay The inhibitory profile of 17 compounds was determined using two protein kinases EGFR-wt and HER2. All values are residual activities (% of control activity) were shown in Table 4. The testing of 17 compounds in singlicate at one concentration (5 Â 1 0 À7 M) with two protein kinases showed a differential inhibitory profile. Residual activity of compounds as followed: Neratinib and A1-A3 20%, A4 > 20% and 60%, C1,C3, B5 > 60% and 80%, the other compounds >80%. As can be seen from the results, compounds A1-A4 were lucky to show stronger inhibition on EGFR-wt and HER2 with inhibition percentage on EGFR-wt kinase of 7%, 6%, 19%, 27%, respectively and 9%, 5%, 12%, 34% on ERBB2 kinase, in which A1 and A2 was pretty much to neratinib with residual activities of 2% and 6%. Since A1 and A2 had superior activity on EGFR-wt, they will almost certainly be effect on various types of EGFR mutations, and was the strong possibility that A1 and A2 could therapy both HER2-positive breast cancer and lung cancer-resistance to the first and the second line EGFR inhibitors compared to neratinib. The compounds revealed that spiro[indoline-3, 4 0 -piperidine] could increase the activity. Further research works are currently under investigation and will be reported in due course.

EGFR wt, HER2, and seven EGFR mutants protein kinase assay
Based on good inhibitions of A1-A2 on EGFR-wt and HER2, we performed head to head comparative trial with neratinib to investigate inhibition of A1-A2 on EGFR-wt, HER2 and seven EGFR mutations (EGF-R d747-749/A750P, EGF-R d752-759, EGF-R G719S, EGF-R L858R, EGF-R L861Q, EGF-R T790M,EGF-R T790M/L858R). The IC 50 values for three compounds in 9 protein kinase assays are shown in Table 5. The corresponding IC 50 curves of three compounds in 9 protein kinase assays are shown in Supporting Information. The results confirmed that A1 and A2 were dual EGFR/HER2 inhibitors, biochemical activities on ERBB2 were about the same to neratinib with the IC 50 values of 0.18 lM, 0.26 lM, and 0.31 lM, respectively, biochemical activities on EGFR wt were not more than that of neratinib but within five times. However, A1 and A2 had stronger inhibitory effect on various EGFR mutants, especially on T790M and L858R that were main mutants resistant to EGFR inhibitors in lung cancer. The inhibitory activities of A1 and A2 on EGFR mutants containing both T790M and L858R were 31 times stronger with IC 50 values of 0.09 lM and 0. 08 lM than neratinib of 2.5 lM. These results showed that A1 and A2could be developed as potential inhibitors to therapy breast cancer and drug-resistant lung cancer.

Docking study
The main structural difference between the active and inactive states is a structural change in the TKD activation loop and movement of the N-lobe helix, both located near the adenosine triphosphate (ATP) binding site. Unlike the active and inactive conformations present in most RTKs and EGFR, in HER2 the conformation as experimentally identified is in the middle of these active-inactive conformations, and has been named the "activelike conformation", due to the orientation of the helix-aC-ac, the DFG-in and the unformed secondary structure of the activation loop 17,18 .Structural analysis showed that the drugs impact differently the conformational space of active and inactive EGFR and energetic analysis pointed out that some ligands have better affinity for the inactive EGFR than the active EGFR state [19][20][21] .
The reports show that in complex with neratinib, the EGFR kinase adopts an inactive conformation in which the regulatory C-helix is displaced from its active position and the enlarged hydrophobic pocket created by the outward rotation of the Chelix appears to be required to accommodate the bulky aniline substituent found in neratinib, neratinib contain additional aromatic groups appended 2-pyridinyl group and binds the inactive conformation of the kinase, the quinoline core of neratinib forms a single hydrogen bond with the hinge region of the kinase. The 2-pyridinyl group of neratinib is surrounded by hydrophobic residues in the expanded pocket, including Met-766 in the C-helix, Phe-856, and Met-790, the mutant gatekeeper residue, the nitrile substituent of neratinib also approaches the gatekeeper residue 11,12 .
In this study, docking experiments were performed to elucidate the binding model of the compounds A1, A2 and neratinib with two mutant proteins 2jiv and 3w2q which was the crystal structure of EGFR kinase domain T790M mutation in compex with neratinib and EGFR kinase domain T790M/L858R mutant with neratinib, respectively. A1 and A2 could adopt the same combination mode as neratinib and bound to active sites of two proteins (shown in Figures 5 and 6). The binding modes of A1, A2 and neratinib with 2jiv showed that their binding energies (kcal/mol) is followed as Neratinib(-7.4)%A1(-8.1)%A2(-7.4), which is consistent with our further experimental activity with IC 50 values of 0.17 lM, 0.089 lM and 0.13 lM. the 2-pyridinyl group of compounds were surrounded by hydrophobic residues in the expanded pocket, including Met-766 in the C-helix, Phe-856, and Met-790, the mutant gatekeeper residue, the nitrile substituent of neratinib also approaches the gatekeeper residue, and the covalent bond was formed between Cys-797 at the edge of the active site cleft and the crotonamide Michael-acceptor group on the inhibitor, it can not match the active site owing to the tension of the spiral ring, the piperidine ring on the spiral ring mainly pointed to the hydrophilic surface, which is far from the hinge region. These effecte could explain that the EGFR kinase adopted an inactive conformation ( Figure 5). The binding modes of A1, A2 and neratinib with 3w2q showed that their binding

energies (kcal/mol) is followed as Neratinib
, which is consistent with our further experimental activity with IC 50 values of 2.5 lM, 0.09 lM and 0.08 lM, A1 and A2 could bind in the ATP-binding cleft with a covalent to Cys-797 in a fashion similar to neratinib and the Cys-797 side-chain orientation wasrelevant to the position and length of the acceptor substituent to accommodate covalent binding, a distinct conformation change for Met790 was observed in the vicinity of these compounds, the side-chain rearrangement of Met790 was necessary to accommodate binding of the cyano group for A1, A2 and neratinib, The moiesty of 3-chloro-4-(pyridin-2-ylmethoxy)benzenamine in A1, A2 and neratinib had similar binding mode with Val726, Leu718, Leu788, and it can not match the active site owing to the tension of the spiral ring, the piperidine ring on the spiral ring mainly pointed to the hydrophilic surface, which is far from the hinge region ( Figure 6).

Seven EGFR mutants protein kinase assay
The compounds were serially diluted in semi-log steps with 100% DMSO in a 96 well microtiter plate. The concentration range of this serial dilution was 1 Â 10 À3 M to 3 Â 10 À8 M. Directly before use the test compounds were further diluted 1:10 with water, resulting in a concentration range from 1 Â 10 À4 M to 3 Â 1 0 À9 M in 10% DMSO. For the assays, 5 lL from each concentration were transferred into the assay. The final volume of the assay was 50 lL. The compounds were tested at 10 final assay concentrations in the range from 1 Â 10 À5 M to 3 Â 10 À10 M. The final DMSO concentration in the reaction cocktails was 1% in all cases. A radiometric protein kinase assay (33PanQinase V R Activity Assay) was used for measuring the kinase activity of the 9 protein kinases. All kinase assays were performed in 96-well FlashPlatesTM from Perkin Elmer (Boston, MA, USA) in a 50 lL reaction volume. The reaction cocktail was pipetted in 4 steps in the following order: 10 lL of non-radioactive ATP solution (in H 2 O), 25 lL of assay buffer/[c-33P]-ATP mixture, 5 lL of test sample in 10% DMSO, 10 lL of enzyme/substrate mixture. The assay for all protein kinases contained 70 mM HEPES-NaOH pH 7.5, 3 mM MgCl 2 , 3 mM MnCl 2 , 3 lM Na-orthovanadate, 1.2 mM DTT, ATP, [c-33P]-ATP, protein kinase, and substrate.The reaction cocktails were incubated at 30 C for 60 min. The reaction was stopped with 50 lL of 2% (v/v) H 3 PO 4 , plates were aspirated and washed two times with 200 lL 0.9% (w/v) NaCl. Incorporation of 33Pi (counting of "cpm") was determined with a microplate scintillation counter (Microbeta, Wallac). All assays were performed with a BeckmanCoulter Biomek 2000/SL robotic system.
For each kinase, the median value of the cpm of three wells with complete reaction cocktails, but without kinase, was defined as "low control" (n ¼ 3). This value reflects unspecific binding of radioactivity to the plate in the absence of protein kinase but in the presence of the substrate. Additionally, for each kinase the median value of the cpm of three other wells with the complete reaction cocktail, but without any compound, was taken as the "high control", i.e. full activity in the absence of any inhibitor (n ¼ 3). The difference between high and low control was taken as 100% activity for each kinase. As part of the data evaluation the low control value of each kinase was subtracted from the high control value as well as from their corresponding "compound values". The residual activity (in %) for each compound well was calculated by using the following formula: Res: Activity % ð Þ ¼ 100 Â cpm of compoundlow control ð Þ = Â high controllow control ð Þ Since 10 distinct concentrations of each compound were tested against each kinase, the evaluation of the raw data resulted in 10 values for residual activities per kinase. Based on each 10 corresponding residual activities, IC50 values were calculated using Prism 5.04 for Windows. The mathematical model used was "Sigmoidal response (variable slope)" with parameters "top" fixed at 100% and "bottom" at 0%. The fitting method used was a least-squares fit.

Molecular docking
The docking was performed with MOE (version 2019) 22 . The crystal structures of EGFR mutations (T790M/L858R and T790M) with neratinib (PDB code: 3W2Q and 2JIV) were downloaded from Protein Data Bank (PDB: http://www.rcsb.org/). Only A chain of EGFR and neratinib were kept, and then the complexes were prepared by using MOE QuickPrep module in MOE (Molecular Operating Environment, version 2019.01) with default parameters. The compounds A1 and A2 were optimized based on MMFF94X force field. The MOE covalent docking was employed to predict the covalent binding modes of EGFR protein with A1, A2 and neratinib. In our covalent docking study, this method relies on a reaction/transformation placement methodology to match the reactive groups (ethylene bond) on the A1, A2 and neratinib and the cysteine residue followed by the formation of the covalent bond between them. Covalent ligands were modeled and minimized in the prereaction form generated with the Builder panel module. The default postprocessing protocol using minimization by the Rigid Receptor refinement scheme and rescoring by the GBVI/WSA dG scoring function was used to estimate the binding free energy (Kcal/mol) of docked pose and EGFR protein. A total of 30 covalent docking poses for each compound were ranked by the estimated binding free energy.