Discovery of 8-Amino-Substituted 2-Phenyl-2,7-Naphthyridinone Derivatives as New c-Kit/VEGFR-2 Kinase Inhibitors

The 2,7-naphthyridone scaffold has been proposed as a novel lead structure of MET inhibitors by our group. To broaden the application of this new scaffold, a series of 8-amino-substituted 2-phenyl-2,7-naphthyridin-1(2H)-one derivatives were designed and synthesized. Preliminary biological screening resulted in the discovery of a new lead of c-Kit and VEGFR-2 kinase inhibitors. Compound 9k exhibited excellent c-Kit inhibitory activity, with an IC50 value of 8.5 nM, i.e., it is 38.8-fold more potent than compound 3 (IC50 of 329.6 nM). Moreover, the compounds 10l and 10r exhibited good VEGFR-2 inhibitory activity, with IC50 values of 56.5 and 31.7 nM, respectively, i.e., they are 5.0–8.8-fold more potent than compound 3 (IC50 of 279.9 nM). Molecular docking experiments provided further insight into the binding interactions of the new lead compounds with c-Kit and VEGFR-2 kinase. In this study, an 8-amino-substituted 2-phenyl-2,7-naphthyridin-1(2H)-one scaffold was identified as the new lead structure of c-Kit and VEGFR-2 kinase inhibitors.


Introduction
N-Heterocyclic scaffolds are ubiquitous building blocks that are used for pharmaceuticals and agrochemicals and in materials science. Novel ring structures may exhibit unique biological or physical properties [1]. Naphthyridines, which act as bioisosteres of quinolone, universally exist in nature and form an important class of nitrogen-containing heterocyclic compounds [2]. They possess a conjugated π-system and coplanar stiffness, and are therefore a highly valued building block. Since 1893, a number of naphthyridine derivatives have been synthesized and explored for their biological activities in the search for novel drugs. A broad spectrum of biological activities has been associated with the functional derivatives of naphthyridines [3]. Gemifloxacin (1, Figure 1), an antibacterial agent, is the most successful example of a naphthyridine-based drug [4,5]. In recent years, many naphthyridine derivatives have been reported to inhibit protein kinase activity (e.g., PI3Kδ [6], CK2 [7], Akt1/2 [8], Tpl2 [9,10], and MET [11][12][13]) for the treatment of many types of human diseases, including inflammation and cancer [14][15][16][17][18].
In our previous work, a novel 2,7-naphthyridone scaffold was designed to conformationally restrain the key pharmacophoric groups of class II MET inhibitors, resulting in the discovery of the potent preclinical candidate compound 3, which targets MET kinase with a favorable drug-likeness [11]. To further expand the application of the 2,7-naphthyridone scaffold, a series of 8-aminosubstituted 2-phenyl-2,7-naphthyridin-1(2H)-one derivatives were designed ( Figure 2). A small library of 2,7-naphthyridones with structural diversity in its 8-amino groups, substituted 2-phenyl groups, and 3-substituents was constructed to discover new kinase inhibitors.
In our previous work, a novel 2,7-naphthyridone scaffold was designed to conformationally restrain the key pharmacophoric groups of class II MET inhibitors, resulting in the discovery of the potent preclinical candidate compound 3, which targets MET kinase with a favorable drug-likeness [11]. To further expand the application of the 2,7-naphthyridone scaffold, a series of 8-aminosubstituted 2-phenyl-2,7-naphthyridin-1(2H)-one derivatives were designed ( Figure 2). A small library of 2,7-naphthyridones with structural diversity in its 8-amino groups, substituted 2-phenyl groups, and 3-substituents was constructed to discover new kinase inhibitors.
In our previous work, a novel 2,7-naphthyridone scaffold was designed to conformationally restrain the key pharmacophoric groups of class II MET inhibitors, resulting in the discovery of the potent preclinical candidate compound 3, which targets MET kinase with a favorable drug-likeness [11]. To further expand the application of the 2,7-naphthyridone scaffold, a series of 8-aminosubstituted 2-phenyl-2,7-naphthyridin-1(2H)-one derivatives were designed ( Figure 2). A small library of 2,7-naphthyridones with structural diversity in its 8-amino groups, substituted 2-phenyl groups, and 3-substituents was constructed to discover new kinase inhibitors.

Biological Evaluation
The ability of the synthesized compounds to inhibit MET, c-Kit, and VEGFR-2 activities was evaluated using an enzyme assay with a recombinant kinase domain [29]. Based on the structure activity relationship (SAR) of MET inhibitors [30], we proposed that the removal of the key diaryl ether fragments in compound 3 would result in a loss of MET activity. As shown in Table 1, compounds 9a-k exhibited no obvious MET inhibitory activity at 5000 nM, while our previously reported lead compound 3 exhibited excellent MET inhibitory activity (IC50 of 9.9 nM). Interestingly, compound 9g (n = 1, block A-6/4-pyridyl group) exhibited a moderate inhibitory activity against c-Kit (IC50 of 832.0 nM) that was only 2.5-fold less potent than that of compound 3 (IC50 of 329.6 nM). More importantly, 9k (n = 1, block A-9/4-quinolyl group) exhibited excellent c-Kit inhibitory activity (IC50 of 8.5 nM); 9k is 38.8-fold more potent than compound 3. Moreover, compounds 9c (n = 0, block A-3/2, 6-dichloro-phenyl group), 9g (block A-6), and 9k (block A-9) exhibited moderate VEGFR-2 inhibitory activity (IC50 values of 238.5-691.2 nM), which was comparable to compound 3 (IC50 of 279.9 nM). Table 1. Inhibitory activity of 9a-k against MET, c-Kit, and VEGFR-2.

Molecular Modeling
Molecular docking experiments were further performed to determine the SAR [31][32][33][34][35][36][37]. As shown in Figure 3A-C and Table 3, the entire molecules of 9g, 9k, and 10r were favorably located in the c-Kit binding pocket. The main weak interactions between 9g and c-Kit included: (1) the H-bond interactions with residues Asp810 and Cys673; (2) the ion-π interaction with Lys623; and (3) the hydrophobic interaction with Leu799. The stronger H-bond interactions with Cys673 and additional hydrophobic interactions with residues Leu595 and Tyr672 of compound 10r rendered 10r five-times more potent than compound 9g. The further enhancement of key H-bond interactions with residues Asp810 and Cys673 resulted in the significantly improved c-Kit inhibitory activity of 9k.

Molecular Modeling
Molecular docking experiments were further performed to determine the SAR [31][32][33][34][35][36][37]. As shown in Figure 3A-C and Table 3, the entire molecules of 9g, 9k, and 10r were favorably located in the c-Kit binding pocket. The main weak interactions between 9g and c-Kit included: (1) the H-bond interactions with residues Asp810 and Cys673; (2) the ion-π interaction with Lys623; and (3) the hydrophobic interaction with Leu799. The stronger H-bond interactions with Cys673 and additional hydrophobic interactions with residues Leu595 and Tyr672 of compound 10r rendered 10r five-times more potent than compound 9g. The further enhancement of key H-bond interactions with residues Asp810 and Cys673 resulted in the significantly improved c-Kit inhibitory activity of 9k.  As shown in Figure 3D-F and Table 3, compounds 9g, 9k, and 10r were entirely located in the VEGFR-2 binding pocket. The main weak interactions between compound 9g and VEGFR-2 included: (1) the H-bond interactions with residues Asp1046 and Cys919; (2) ion-π interaction with Lys868; (3) the hydrophobic interaction with Leu1035. The stronger H-bond interactions with residues Cys919 and additional hydrophobic interactions with residues Leu840 and Tyr918 rendered compound 9k more potent than 9g. The further enhancement of key H-bond interactions with residues Asp1046 and additional hydrophobic interaction with residues Ile892 and Leu1019 resulted in the significantly improved VEGFR-2 inhibitory activity of 10r.
Taking the molecular docking results into account, we hypothesize that the hydrogen bond acceptor (N containing heterocycle) and hydrophobic effects (fused aromatic ring) of 8-amino substituents are crucial to improve the inhibitory activity for c-Kit and VEGFR-2 kinase.

Conclusions
In summary, we described the design and synthesis of a series of 8-amino-substituted 2-phenyl-2,7-naphthyridin-1(2H)-one derivatives to broaden the application of this new 2,7-naphthyridone scaffold. Preliminary biological screening resulted in the discovery of new lead compounds 9k, 10l, and 10r, which exhibit more potent c-Kit and VEGFR-2 kinase inhibitory activity than the previously reported lead compound, 3. Molecular docking results provided further insight into the binding interactions of the new lead compounds with c-Kit and VEGFR-2 kinase. We identified 8-aminosubstituted 2-phenyl-2,7-naphthyridin-1(2H)-one as a new lead scaffold of c-Kit and VEGFR-2 kinase inhibitors.

Biochemical Kinase Assays
The ability of compounds to inhibit the activity of three kinases (MET, c-Kit, and VEGFR-2) was tested in vitro [30]. Enzyme assays were run in homogeneous time-resolved fluorescence (HTRF) format in 384-well microtiter plates using purified kinases purchased from Invitrogen (Carlsbad, CA, US). The HTRF KinEASE TK kit (contains substrate-biotin, antibody-cryptate, streptavidin-XL665, As shown in Figure 3D-F and Table 3, compounds 9g, 9k, and 10r were entirely located in the VEGFR-2 binding pocket. The main weak interactions between compound 9g and VEGFR-2 included: (1) the H-bond interactions with residues Asp1046 and Cys919; (2) ion-π interaction with Lys868; (3) the hydrophobic interaction with Leu1035. The stronger H-bond interactions with residues Cys919 and additional hydrophobic interactions with residues Leu840 and Tyr918 rendered compound 9k more potent than 9g. The further enhancement of key H-bond interactions with residues Asp1046 and additional hydrophobic interaction with residues Ile892 and Leu1019 resulted in the significantly improved VEGFR-2 inhibitory activity of 10r.
Taking the molecular docking results into account, we hypothesize that the hydrogen bond acceptor (N containing heterocycle) and hydrophobic effects (fused aromatic ring) of 8-amino substituents are crucial to improve the inhibitory activity for c-Kit and VEGFR-2 kinase.

Conclusions
In summary, we described the design and synthesis of a series of 8-amino-substituted 2-phenyl-2,7naphthyridin-1(2H)-one derivatives to broaden the application of this new 2,7-naphthyridone scaffold. Preliminary biological screening resulted in the discovery of new lead compounds 9k, 10l, and 10r, which exhibit more potent c-Kit and VEGFR-2 kinase inhibitory activity than the previously reported lead compound, 3. Molecular docking results provided further insight into the binding interactions of the new lead compounds with c-Kit and VEGFR-2 kinase. We identified 8-amino-substituted 2-phenyl-2,7-naphthyridin-1(2H)-one as a new lead scaffold of c-Kit and VEGFR-2 kinase inhibitors.

Biochemical Kinase Assays
The ability of compounds to inhibit the activity of three kinases (MET, c-Kit, and VEGFR-2) was tested in vitro [30]. Enzyme assays were run in homogeneous time-resolved fluorescence (HTRF) format in 384-well microtiter plates using purified kinases purchased from Invitrogen (Carlsbad, CA, US). The HTRF KinEASE TK kit (contains substrate-biotin, antibody-cryptate, streptavidin-XL665, 5×enzymatic buffer, and detection buffer) was purchased from Cisbio (Codolet, France), and the kinase assays were performed according to the manufacturer's instructions. After the kinases and the compounds incubated at 25~30 • C for 5 min, the reactions were initiated by the addition of 2 µL of mixed substrate solution (mixed solution of ATP (Sigma, Shanghai, China) and substrate-biotin). The final concentrations of kinases were at EC 80 and the total reaction volume was 8 µL. Plates were incubated at 30 • C for 30~60 min, then the reactions were quenched by the addition 8 µL mixed detection solution (mixed solution of antibody-cryptate and streptavidin-XL665 in detection buffer). The fluorescence excitation wavelength was 320 nM. The fluorescence at 665 nm (acceptor emission wavelength) and 620 nm (donor emission wavelength) was measured with a PHERAstar FS plate reader (BMG, LABTECH, Ortenberg, Germany) using a time delay of 50 µs. All kinase assays were conducted using ATP concentrations below the enzyme K mapp and kinase-specific biotinylated substrate peptides.
The data for dose responses were plotted as percentage of inhibition calculated with the data reduction formula 100 × [1 − (U 1 − C 2 )/(C 1 − C 2 )] versus concentration of compound, where U is the emission ratio of 665 nm and 620 nm of test sample, C 1 is the average value obtained for solvent control (2% DMSO), and C 2 is the average value obtained for no reaction control (no kinase sample). Inhibition curves were generated by plotting percentage of control activity versus log 10 of the concentration of each kinase. The IC 50 values were calculated by nonlinear regression with Graphpad Prism 5 (GraphPad Software, San Diego, CA).

Molecular Modeling
The three-dimensional structures of the small molecules were constructed and primarily optimized by Sybyl 2.0 software. Steepest descent and conjugate gradient methods were used in the optimization process. Autodock Tools was used to assign Gasteiger charges for both the receptors and inhibitors.

General Information
Unless otherwise noted, all chemical reagents were commercially available and treated with standard methods. Silica gel column chromatography (CC). Silica gel (200-400 Mesh; Qingdao Makall Group Co., Ltd.; Qingdao; China). Solvents were dried in a routine way and redistilled. All reactions involving air-or moisture-sensitive reagents were performed under a nitrogen or argon atmosphere. 1 HNMR spectra (400 MHz) and 13 CNMR (100 MHz) spectra were recorded on a Bruker BioSpin AG (Ultrashield Plus AV 400, Fällanden, Switzerland) spectrometer as deuterochloroform (CDCl 3 ) or dimethyl sulfoxide-d 6 (DMSO-d 6 ) solutions using tetramethylsilane (TMS) as an internal standard (δ = 0) unless noted otherwise. MS spectra were obtained on an Agilent technologies 6120 quadrupole LC/MS (ESI). All reactions were monitored using thin-layer chromatography (TLC) on silica gel plates. Yields were of purified compounds and were not optimized.

General Procedure for the Preparation of Intermediates 7a-f
The intermediates 7a-f were prepared according to our previous report [11].

General Procedure for the Preparation of Targets 9a-k and 10a-s
An oven-dried Schlenk tube was charged with 7 (0.4 mmol), Pd 2 (dba) 3 (0.02 mmol), xantphos (0.04 mmol), t-BuONa (0.8 mmol), and amine (0.48 mmol), and then purged with argon three times. Ultra-dry dioxane (15 mL) was added to the Schlenk tube with a syringe at argon atmosphere. The mixture was stirred at 110 • C overnight. After cooling to room temperature, the mixture was concentrated in vacuo and the residue purified by flash chromatography on silica gel using DCM/MeOH (100:1) as eluent to obtain 9a-k and 10a-s.