Discovery of Potent c-MET Inhibitors with New Scaffold Having Different Quinazoline, Pyridine and Tetrahydro-Pyridothienopyrimidine Headgroups

Cellular mesenchymal-epithelial transition factor (c-MET) is closely linked to human malignancies, which makes it an important target for treatment of cancer. In this study, a series of 3-methoxy-N-phenylbenzamide derivatives, N-(3-(tert-butyl)-1-phenyl-1H-pyrazol-5-yl) benzamide derivatives and N1-(3-fluoro-4-methoxyphenyl)-N3-(4-fluorophenyl) malonamide derivatives were designed and synthesized, some of them were identified as c-MET inhibitors. Among these compounds with new scaffolds having different quinazoline, pyridine and tetrahydro-pyridothienopyrimidine head groups, compound 11c, 11i, 13b, 13h exhibited both potent inhibitory activities against c-MET and high anticancer activity against tested cancer cell lines in vitro. In addition, kinase selectivity assay further demonstrated that both 13b and 13h are potent and selective c-MET inhibitors. Molecular docking supported that they bound well to c-MET and VEGFR2, which demonstrates that they are potential c-MET RTK inhibitors for cancer therapy.


Introduction
Tyrosine kinase is an enzyme that transfers a phosphate group from ATP to a protein, and it functions as an "on" or "off" switch in many cellular functions. They become potent oncogene that has the potential to cause cancer, when they are often mutated or expressed at high levels [1]. Several receptor tyrosine kinases (RTKs) inhibitors have been found to have effective anti-tumor activity and some of them have been approved or are in clinical trials. Recent FDA approved drugs Sorafenib (Nexavar) [2] is such example of multi-targeted agents ( Figure 1).

Introduction
Tyrosine kinase is an enzyme that transfers a phosphate group from ATP to a protein, and it functions as an "on" or "off" switch in many cellular functions. They become potent oncogene that has the potential to cause cancer, when they are often mutated or expressed at high levels [1]. Several receptor tyrosine kinases (RTKs) inhibitors have been found to have effective anti-tumor activity and some of them have been approved or are in clinical trials. Recent FDA approved drugs Sorafenib (Nexavar) [2] is such example of multi-targeted agents ( Figure 1).  Inspired by the structure of the lead compound 4, we have also designed a series of N-phenylbenzamide derivatives. The synthetic route is shown in Scheme 3. Compounds 10 were synthesized by a conventional peptide synthesis method with aniline and benzoic acid using ClCOCOCl at 0 °C. Then, a typical Williamson ether synthesis was carried on to afford compound 11a-i [29]. The chemistry described in Scheme 2 shows compound libraries could be made simply by using the reaction protocol with N-(3-(tert-butyl)-1-phenyl-1H-pyrazol-5-yl)-3-hydroxybenzamide (Compound 8) and different head groups. A cyclization reaction with nitrile and hydrazine was carried to afford 3-(tert-butyl)-1-phenyl-1H-pyrazol-5-amine. Compound 8 was synthesized by a conventional peptide synthesis method using ClCOCOCl at 0˝C [28]. The chemistry described in Scheme 2 shows compound libraries could be made simply by using the reaction protocol with N- (3-

(tert-butyl)-1-phenyl-1H-pyrazol-5-yl)-3-hydroxybenzamide (Compound 8)
and different head groups. A cyclization reaction with nitrile and hydrazine was carried to afford 3-(tert-butyl)-1-phenyl-1H-pyrazol-5-amine. Compound 8 was synthesized by a conventional peptide synthesis method using ClCOCOCl at 0 °C [28]. Inspired by the structure of the lead compound 4, we have also designed a series of N-phenylbenzamide derivatives. The synthetic route is shown in Scheme 3. Compounds 10 were synthesized by a conventional peptide synthesis method with aniline and benzoic acid using ClCOCOCl at 0 °C. Then, a typical Williamson ether synthesis was carried on to afford compound 11a-i [29]. We also combined different head groups to the side chain of lead compound 1, XL184. The synthetic route is shown in Scheme 4. After multi-step protection and de-protection reaction, a series ethyl hydrogen malonate derivative was obtained. Then, Compounds 12 were synthesized by a conventional peptide synthesis method using isobutyl chloroformate and 4-Methylmorpholine. After that, a typical Williamson ether synthesis was carried on to afford compound 13a-i [30].

Kinase Inhibitory Assay
All the synthesized compounds were assayed with the enzymatic activity against c-MET. The results were summarized in Tables 1-3. Also included was the representative c-MET inhibitor (XL184). Among these compounds, compound 11c, 11i, 13b, 13h showed potent inhibitory activity against

Molecular Docking and Molecular Dynamics Simulation Study
In order to better understand the interaction between compounds and kinases, molecular docking studies on the potent compound 11c and 11i were performed using the Discovery Studio 3.1/CDOCKER protocol [32].
In Figure  To further evaluate the binding affinity between compound 13b and c-MET/VEGFR-2. The molecular dynamics (MD) simulations were performed by using GROMACS package (version 4.5, University of Groningen, Groningen, The Netherlands), the root-mean-square deviation (RMSD) fluctuations is a principal criterion to evaluate the stability of the protein-ligand system.

Molecular Docking and Molecular Dynamics Simulation Study
In order to better understand the interaction between compounds and kinases, molecular docking studies on the potent compound 11c and 11i were performed using the Discovery Studio 3.1/CDOCKER protocol [32].
In Figure 2, we showed that compound 11c and 11i could bind to c-MET kinase (PDB: 4MXC) very well. In addition, they can form hydrophobic interaction in the ATP-binding sites of c-MET and VEGFR-2. Compounds formed hydrophobic interaction with residues ILE-1084, ALA-1108, LEU-1157, MET-1160 and ALA-1221 of c-MET.   To further evaluate the binding affinity between compound 13b and c-MET/VEGFR-2. The molecular dynamics (MD) simulations were performed by using GROMACS package (version 4.5, University of Groningen, Groningen, The Netherlands), the root-mean-square deviation (RMSD) fluctuations is a principal criterion to evaluate the stability of the protein-ligand system.
As shown in Figure 4, although the RMSD values for compound 13b fluctuated in a narrow range around 2700 ps, they remained stable for most of the simulation in c-MET complex. RMSD of compound 13b reach the equilibrium state after about 500 ps and kept stable among the rest of simulation in VEGFR-2 complex, indicating a stabilities of the dynamics equilibriums. In general, the maximum RMSD for each case was lower than 0.2 nm, suggesting that c-MET and VEGFR-2 complexes are reliable and the low RMSD fluctuations of system observed indicated stable binding models of compound 13b with c-MET and VEGFR-2, respectively. As shown in Figure 4, although the RMSD values for compound 13b fluctuated in a narrow range around 2700 ps, they remained stable for most of the simulation in c-MET complex. RMSD of compound 13b reach the equilibrium state after about 500 ps and kept stable among the rest of simulation in VEGFR-2 complex, indicating a stabilities of the dynamics equilibriums. In general, the maximum RMSD for each case was lower than 0.2 nm, suggesting that c-MET and VEGFR-2 complexes are reliable and the low RMSD fluctuations of system observed indicated stable binding models of compound 13b with c-MET and VEGFR-2, respectively. range around 2700 ps, they remained stable for most of the simulation in c-MET complex. RMSD of compound 13b reach the equilibrium state after about 500 ps and kept stable among the rest of simulation in VEGFR-2 complex, indicating a stabilities of the dynamics equilibriums. In general, the maximum RMSD for each case was lower than 0.2 nm, suggesting that c-MET and VEGFR-2 complexes are reliable and the low RMSD fluctuations of system observed indicated stable binding models of compound 13b with c-MET and VEGFR-2, respectively.

Chemical Synthesis
All reagents were purchased from commercial sources and used without further purification. Melting points are corrected. 1 H-NMR spectra were determined on a Bruker Avance III 400 MHz spectrometer (Bruker, Billerica, MA, USA) in CDCl 3 or DMSO-d 6 solution. J values were in Hz. Chemical shifts were expressed in ppm downfield from internal standard TMS. HRMS data were obtained using Bruker micro TOF-Q instrument (Bruker) or TOF-MS instrument (Bruker).

General Procedure for the Preparation of 4-Chloro-6,7-dimethoxyquinazoline (Compound 5)
2-amino-dimethyl aminobenzoic acid 2.02 g (10 mmol) and acetic acid formamidine 2.10 g (20 mmol) were added in 2-methoxy ethanol. The mixture was reflux in 125˝C. After evaporation of solvent, the residue was added in the 10% ammonia solution, stirring, then after filtering, the solid was washed by water to get the brown solid powder (yield 89%). 1.82 g (8.8 mmol) product was dissolved in 25 mL SOCl 2 , and 30 drops of DMF was added. After heating 7 h to reflux, the solvent was evaporated, then after filtering, the solid was washed by water to get the brown solid powder 4-chloro-6,7-dimethoxyquinazoline (yield 93%). The N-(tert-Butoxycarbonyl)-4-piperidone, NCCH 2 CO 2 Et and Et 3 N was mixed at room temperature then stirring for 16 h, and then heated to 120˝C for 16 h in DMF with the formamidine acetate. Then, the intermediate was reacted with POCl 3 and DIPEA in toluene to obtain the compound 7 (yield 89%).

General Procedure for the Preparation of 3-Methoxy-N-phenylbenzamide Derivatives
Compound 11a-i 138.6 mg (1 mmol) hydroxybenzoic acid in 5 mL THF was added three drops of DMF and reacted with oxalyl chloride 0.5 mL in 0˝C. After 1.5 h reaction, the reaction liquid was dissolved in THF and HCl, which was dissolved in THF solution of aniline derivatives. After 3 h, the reaction mixture was evaporated; the crude product was purified by column chromatography to obtain compounds 10. Compounds 10 were further reacted with compounds 5, 6, 7 in the presence of K 2 CO 3 to obtain the final products 11a-i.