Design, Synthesis, and Biological Evaluation of 4-Phenoxyquinoline Derivatives as Potent c-Met Kinase Inhibitor
Abstract
A series of novel 4-phenoxyquinoline derivatives containing 3-oxo-3,4-dihydro-quinoxaline moiety were synthesized and evaluated for their antiproliferative activity against five human cancer cell lines (A549, H460, HT-29, MKN-45 and U87MG) in vitro. Most of the tested compounds exhibited more potent inhibitory activities than the positive control foretinib. Compound 1b, 1s and 1t were further examined for their inhibitory activity against c-Met kinase. The most promising compound 1s (with c-Met IC50 value of 1.42 nM) showed remarkable cytotoxicity against A549, H460, HT-29, MKN45 and U87MG cell lines with IC50 values of 0.39 μM, 0.18 μM,0.38 μM, 0.81 μM, respectively. Their preliminary structure-activity relationships (SARs) study indicated that the replacement of the aromatic ring with the cyclohexane improved their antiproliferative activity.The c-Met (Mesenchymal-epithelial transition factor) tyrosine kinase is a key member of receptor tyrosine kinases (RTKs) family 1-3, which can affect the occurrence, development, invasion and metastasis of tumor cells. The c-Met is a high-affinity receptor for hepatocyte growth factor (HGF) 4. Whereas, abnormal activation of HGF/c-Met signaling pathway is associated with cell processes such as tumor formation, migration and apoptosis. Furthermore, tumor angiogenesis, tumor invasion and metastasis and drug resistance are also closely related to this signaling pathway. Therefore, the HGF/c-Met signaling pathway has been emerging as an attractive target for cancer therapies 5.At present, most of the small molecule c-Met kinase inhibitors have been found to be competitive inhibitors of ATP, which inhibited c-Met kinase by blocking the phosphorylation of tyrosine. Small molecule c-Met kinase inhibitors were mainly divided into Type I and Type II, according to the way of combination. Recent research6 shows that Type II inhibitors 7,8, which have a wide variety of structural types and can be modified with more sites, target multiple links of the c-Met signaling pathway. Over the last decade, a large number of c-Met inhibitors containing 6,7-disubstituted-4-phenoxyquinolines have been reported as class II c-Met inhibitors, such as Kirin Brewery 9, Cabozantinib 10, Foretinib 11, AMG458 12, and AM7 13 (Figure 1)Fig. 1 Classical TypeⅡ c-Met inhibitors
These structures of 4-phenoxyquinolines as c-Met kinase inhibitors mainly consist of three parts: A (4-phenoxyquinoline core moiety), B (aryl hydrophobic moiety), and C (linker moiety). The modification of 4-phenoxyquinoline reported in the literature mainly included the following aspects: (1) Introduction of a hydrophilic group at the 7 position in moiety A; (2) Replacement the phenyl ring with various substituents (moiety B); (3) Transformation of moiety C.In our previous work, the C fragment was modified to obtain three series of compounds, containing 1,2,4-triazolone fragment, imidazole fragment and quinoxalinone fragment respectively 15-17 (Figure 2). Particularly, compounds bearing quinoxalinone fragment had been confirmed to exhibit comparable or superior anticancer activity to foretinib. Our latest study had focused on the effects of B ring substituents, which can affect the electrical properties of part B. In the above research, it was very fortunate for us to obtain the most potent compound 9, which showed 3-fold more potent activity against A549 and HT29 cell lines than foretinib. Meanwhile, its activity against H460, MKN-45 and U87MG cell lines is comparable to foretinib. However, according to the molecular simulation docking experiments of compound 9, using the co-crystal protein of foretinib and c-Met (PDB ID code: 3LQ8) as a docking model (Figure 3), we found that B ring does not fit into the binding site completely due to the rigidity of the B ring. Therefore, we hoped to change the flexibility of the B ring by adding a flexible carbon chain between the B ring and the C ring for better occupying the cavity. On this basis, we further explored the influence of carbon chain between the B and C parts, so compounds 1a-1p had been synthesized in order to prove our point of view. On the other hand, we are also interested to investigate the effect of the aromatic ring π electron of B on antitumor activity by replacing the aromatic ring with the non-planar aliphatic ring (compounds 1q-1x).
The key intermediates 6,7-disubstituted-4-phenoxyquinolines L1-L4 were synthesized by a convenient eight-step route starting from 1-(4-hydroxy-3-methoxyphenyl)ethanone, which was described in detail in our previous study18 (Scheme 1). Commercially available 1-(4-hydroxy-3-methoxyphenyl)ethanone reacted with 1-chloro-3-bromopropane under basic condition to obtain intermediate M1, which was then nitrified by fuming nitric acid at -20℃ for 6 h to give intermediate M2. Then, condensation of M2 with dimethyl formamide dimethyl acetal (DMF-DMA) in refluxing toluene provided a yellow solid M3, which was reduced and cyclized by glacial acetic acid and iron powder to afford hydroxyquinoline M4 with a high yield and purity. The condensation reaction of M4 with excessive secondary amines (piperidine, 4-methyl piperidine, N-methylpiperazine and morpholine) in acetonitrile at reflux provided intermediates M5, which were converted to 4-chloroquinoline M6 by refluxing in the mixed solvent of phosphorus oxychloride and acetonitrile in sequence. Next, the etherification reaction of M6 with 2-fluoro-4-nitrophenol afforded purified M7, which were reduced using iron powder and catalytic amounts of ammonium chloride in ethanol to obtain amides L1-L4.Scheme 1. Synthesis of intermediate L1-L4. Reagents and conditions: (i) Br(CH2)3Cl, acetone, 0 C, 30min, r.t., 12 h; (ii) 98% HNO3, CH2Cl2, -20 C, 4 h; (iii) DMF-DMA, toluene, 110 C, 10 h; (iv) Fe powder, AcOH, rt, 30 min, 80 C, 2 h; (v) secondary amines, CH3CN, 85 C, 10 h; (vi) POCl3, 85 C, 6 h; (vii) 2-fluoro-4-nitrophenol, PhCl, 140 C, 30 h; (viii) Fe powder, NH4Cl (sat.), EtOH/H2O, reflux, 5 h.
The synthetic route of the target compounds is shown in Scheme 2. The condensation of commercially available 1-fluoro-2-nitrobenzene with different substituted amines in the presence of sodium hydride in N,N-dimethylformamide afforded intermediates B1. Reduction of the nitro group of B1 with iron powder provided compounds B2. The cyclization of B2 with diethyl ketomalonate afforded 3-oxo-3,4-dihydro-quinoxaline-2-carboxylic acid esters B3, which were converted to acids B4 using lithium hydroxide monohydrate in THF/H2O at room temperature for 2 h. B4 were refluxed in toluene and SOCl2 for 5 h to afford acyl chlorides, which were condensed with intermediates L1-L4 in the presence of sodium carbonate in DCM at room temperature overnight to afford the target compounds 1a-1x.Scheme 2. The synthetic route of target compounds 1a-1x. Reagents and conditions: (a) amines, NaH, DMF, r.t., 16 h; (b) Fe powder, AcOH, CH3COOEt/H2O, reflux, 6 h; (c) diethyl ketomalonate, toluene, reflux, 12 h; (d) LiOH, THF/ H2O, r.t., 1.5 h, then 6 N HCl; (e) toluene, SOCl2, reflux, 5 h; (f) Na2CO3, CH2Cl2, 25 C, 3 h.In this paper, all the synthesized compounds (1a-1d, 1g, 1i-1x) were screened for their cytotoxic activity by the MTT-based assay using foretinib as a positive control. They were tested against five c-Met overexpressed human cancer cell lines, namely human colon cancer cell line (HT-29), human lung cancer cell lines(H460), human lung adenocarcinoma cell line (A549), human gastric cancer cell line (MKN-45) and human glioblastoma cell line (U87MG). The results were expressed as IC50 values and summarized in Table 1. The IC50 values were the average of at least three independent experiments.
As illustrated in Table 1, most of targeted compounds showed great activity against five cancer cell lines in comparison with foretinib in vitro. In our design, two different schemes have been carried out, such as prolonging the carbon chain between the B and C parts and replacing the aromatic ring with the aliphatic ring. In the study of prolonging the carbon chain, compounds 1a, 1b, and 1i bearing 4-fluorobenzyl or 4-methoxyphenethyl was obtained and their activity against the five cell lines were significantly better than the positive control foretinib. In addition, in the exploration of the replacement of the aromatic ring by the aliphatic ring, compounds 1q and 1r with cyclohexyl group exhibited similar activity levels in vitro. Notably, compounds 1s and 1t showed approximately 5-fold more potent against A549, H460 and HT-29 cell lines than foretinib, which showed much more potent than compound 9 as we mentioned above.Based on these activity results in vitro, the SAR revealed that the series of compounds obtained by extending the carbon chain between the B moiety and the C moiety exhibits comparable or superior activity to the positive control foretinib, but do not improve compared to compound 9. It’s frustrating that the removal of the 4-position substituent led to a significant decrease in activity. As shown in table 1, compound 1g showed a 2- to 10-fold activity loss compared to compound 1c, in which R1 substituent is morpholinyl. Therefore, we only synthesize one for this type of compounds.
It could be known that the substituent on the B ring is very important. The potency will suffer a heavy loss when the substituent is absent. When there were two carbon atoms between the B ring and the C moiety, an obviously decrease on anticancer potency can be observed, which seemly indicated that the narrow space of the cavity occupied by the B ring does not allow further extension of the carbon chain. Additionally, in the study of the replacement of the aromatic ring by an aliphatic ring, when the incorporation of cyclohexyl group, a significant enhancement was observed in cell activity, which indicated that the π-electron of the aromatic ring is not necessary. Instead, the cyclohexyl group can exert a stronger hydrophobic interaction with the protein cavity, thereby increasing the biological activity of the compound. However, when R2 is cyclopentyl group, the activity was drastically reduced compared to cyclohexyl, which indicated that shrinking the fat ring will cause a negative effect on activity.On the basis of cellular assay results, three most potent compounds (1b, 1s, 1t) were selected for c-Met enzymatic activity assay using homogeneous time-resolved fluorescence (HTRF). The results were outlined in table 2. All the three compounds showed excellent c-Met enzymatic potency with IC50 values ranging from 1.42 to 2.56 nM while the IC50 values of positive control was 1.63 nM. Apparently, the data of enzymatic potency was consistent with the cell cytotoxic activity, which elucidated that c-Met inhibition may lead to their antitumor effect.
Moreover, 1s exhibited the most potent activity with an IC50 values of 1.42 nM , which was preferable to foretinib.To further elucidate the binding mode of the target compounds, the docking analysis of compound 1s was performed. Similarly, the co-crystal protein of foretinib with c-Met (PDB ID code: 3LQ8) were used as a docking mode. The docking simulation was conducted using Glide XP (Schrödinger 2014), since Glide uses a hierarchical series of filters to search for possible locations of the ligand in the active-site region of the receptor. Then the image files were generated using Accelrys DS visualizer 4.0 system. As shown in Figure 4, all these interactions with Met1160, Asp1222 and Lys1110 that played an important role in stabilizing the conformation of ligand-protein complex were retained. In addition ,the cyclohexyl group formed three hydrophobic interactions with His1202, Phe1134 and Val1139, which suggested that terminal cyclohexyl group could inspire a new optimization idea of c-Met inhibitor.Fig. 5 (A) Binding poses of compound 1s with c-Met ; (B) 2D interactions of the docking model of 1s to c-Met
In summary, a series of 4-phenoxyquinoline derivatives 3-oxo-3,4-dihydro-quinoxaline moiety as novel c-Met receptor tyrosine kinase inhibitors were designed and synthesized. Most of the tested compounds exhibited more potent inhibitory activities than the positive control foretinib. Compounds 1b, 1s and 1t were proved to be the best compounds of the series, showing anti-proliferative activity in the low microgram range on the five human cancer cell lines (A549, H460, HT-29, MKN-45 and U87MG), and remarkable potency on c-Met kinase(IC50 values of 2.56 nM, 1.42 nM and 2.34 nM ). Additionally, compound 1s exhibited the most potent activity with an IC50 values of 1.42 nM, which was preferable to foretinib. Further exploration of the structure-activity relationship has been carried out by the docking simulation analysis of compound 1s with c-Met. It can be observed that cyclohexyl group formed three hydrophobic interactions with His1202, Phe1134 and Val1139, which suggested that terminal cyclohexyl group might produce new structural optimization inspiration of c-Met inhibitor. Further optimization of the terminal structure to enhance inhibitory activity against c-Met is in progress.