Discovery of N-substituted-3-phenyl-1,6-naphthyridinone derivatives bearing quinoline moiety as selective type II c-Met kinase inhibitors against VEGFR-2

Article abstract of DOI:10.1016/j.bmc.2020.115555

New N-substituted-3-phenyl-1,6-naphthyridinone derivatives are designed and synthesized, based on structural modification of our previously reported compound 3. Extensive enzyme-based SAR studies and PK evaluation led to the discovery of compound 4r, with

Full text of DOI:10.1016/j.bmc.2020.115555

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Bioorganic&MedicinalChemistry28(2020)115555
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Discovery of N-substituted-3-phenyl-1,6-naphthyridinone derivatives
bearing quinoline moiety as selective type II c-Met kinase inhibitors against
VEGFR-2
Hongchuang Xu1, Minshu Wang1, Fengxu Wu1, Linsheng Zhuo⁎, Wei Huang⁎, Nengfang She⁎
Key Laboratory of Pesticide & Chemical Biology of Ministry of Education, International Joint Research Center for Intelligent Biosensor Technology and Health, College of
Chemistry, Central China Normal University, Wuhan 430079, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
New N-substituted-3-phenyl-1,6-naphthyridinone derivatives are designed and synthesized, based on structural
modification of our previously reported compound 3. Extensive enzyme-based SAR studies and PK evaluation led
to the discovery of compound 4r, with comparable c-Met potency to that of Cabozantinib and high VEGFR-2
selectivity, while Cabozantinib displayed no VEGFR-2 selectivity. More importantly, at oral doses of 45 mg/kg
(Q.D.), compound 4r exhibits significant tumor growth inhibition (93%) in a U-87MG human gliobastoma xe-
nograft model. The promising selectivity against VEGFR-2 and excellent tumor growth inhibition of compound
4r suggest that it could be used as a new lead molecule for further discovery of selective type II c-Met inhibitors.
c-Met kinase inhibitor
1,6-Naphthyridone
VEGFR-2 selectivity
Antitumor efficacy
1. Introduction
conformation through interactions with DFG residues of the activation
loop to enter an allosteric region (adjacent to the ATP binding site)
The receptor tyrosine kinase c-Met, also known as hepatocyte
growth factor receptor (HGFR), which is encoded by c-Met proto-on-
cogene, has a high affinity for its naturally occurring ligand HGF (also
known as scatter factor).1–3 Once combined with HGF, the c-Met can be
activated and lead to the recruitment, phosphorylation, and activation
of multiple downstream signaling cascades, resulting in cell prolifera-
tion, survival, motility, induction of cell polarity, scattering, angio-
genesis, and invasion.3–5 Although the HGF/c-Met signaling pathways
are essential for embryogenesis and tissue injury repair, dysregulation
of the HGF/c-Met signaling pathway caused by c-Met protein over-ex-
pression, gene amplification or rearrangement, and transcriptional
regulation, as well as autocrine or paracrine ligand stimulation, has
been correlated with the promotion of multiple tumorigenic processes
and poor prognoses.6–10 Thus, the c-Met protein has been determined as
an important therapeutic target for cancer therapy.
which is only accessible in the inactive DFG-out conformation of the
kinase.3,14,15 Owing to their different binding modes, type I inhibitors
often have high selectivity against other kinases, while type II inhibitors
are often multi-target inhibitors. Despite the high selectivity of type I
inhibitors, drug resistance to certain mutations near the active site of c-
Met remains a significant issue in subsequent clinical treatment.
Hopefully, type II inhibitors may prove more effective against the
mutations, due to their capability of binding to the ATP-binding site, as
well as the allosteric site beyond the entrance of the active site of c-
Met.16,17 However, the potential off-target effects of type II inhibitors
on other protein kinases may cause significant clinical issues affecting
patient’s quality of life, such as VEGFR-2 related side effects.18 To date,
few selective type II c-Met inhibitors have been reported. Therefore,
developing novel type II c-Met inhibitors with increased selectivity over
other protein kinases is a challenging, but meaningful, subject.
Numerous small-molecule c-Met kinase inhibitors have been de-
veloped. According to their binding mode in the c-Met kinase domain,
these inhibitors can be divided into two classes, known as type I and
type II, which are typified by Capmatinib and Cabozantinib, respec-
tively (see Fig. 1).11–13 Type I inhibitors are ATP-competitive inhibitors
which bind in a U-shaped conformation to the ATP-binding site at the
pocket entrance, while type II inhibitors adopt an extended
In light of our previous studies, the 2,7-naphthyridinone and 1,6-
naphthyridine moieties can serve as a key pharmacophoric group (block
C) to demonstrate c-Met inhibition, and the 2-fluoro-1,4-substituted
phenyl linker (block B) was generally conservative in the reported in-
hibitors. Moreover, fine-regulation of the functionalizable N(1) amine
group of the 1,6-naphthyridin-4(1H)-one moiety could lead to se-
lectivity improvement.19,20 Meanwhile, the quinoline moiety has been
⁎ Corresponding authors.
E-mail addresses: zhuolinsheng199010@126.com (L. Zhuo), weihuangwuhan@126.com (W. Huang), nfshe@mail.ccnu.edu.cn (N. She).
1 These authors contributed equally to this work.
https://doi.org/10.1016/j.bmc.2020.115555
Received 11 March 2020; Received in revised form 1 May 2020; Accepted 5 May 2020
Availableonline12May2020
0968-0896/©2020ElsevierLtd.Allrightsreserved.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
H. Xu, et al.
Bioorganic&MedicinalChemistry28(2020)115555
Fig. 1. Reported Type I and Type II c-Met inhibitors.
Fig. 2. Design of 1,6-naphthyridinone derivatives bearing a quinoline moiety in block A as new c-Met inhibitors.
proven to be a privileged scaffold in block A of type II c-Met inhibitors
naphthyridone building blocks were also synthesized from our reported
1,6-naphthyridone block through substitution reactions with different
halides or substituted oxiranes. Eventually, the target compounds were
prepared via an acid-catalyzed aromatic nucleophilic substitution of the
5-Cl group of 1,6-naphthyridone 10a–k with the aromatic amines 8a–d.
with favorable drug-like properties, such as Cabozantinib,13 TAS-115,21
Foretinib,22 RXDX-106,23 AMG-458,24 and Ningetinib.25 Optimization
of the substituents in the quinoline fragment could also improve po-
tency.19,21 Therefore, our reported 1,6-naphthyridone-based c-Met ki-
nase inhibitor 326 bearing a quinoline moiety in block A, which dis-
played moderate c-Met potency but no VEGFR-2 selectivity, was
selected as our starting point for this research, in order to improve its
potency and selectivity (Fig. 2). In this paper, we describe our efforts to
introduce structural modifications on the 1,6-naphthyridin-4(1H)-one
and quinolone moieties, as well as on the terminal phenyl ring. A series
of N-substituted-3-phenyl-1,6-naphthyridinone derivatives were pre-
pared and all synthesized compounds were evaluated for their enzyme-
based c-Met inhibition activity and VEGFR-2 selectivity. The most po-
tent compound 4r with high VEGFR-2 selectivity was also assessed for
tumor growth inhibition in the U-87 MG xenograft model.
2.2. In vitro biochemical c-Met/VEGFR-2 kinase inhibition activities
With the purpose to improve selectivity and potency, the primary
goal of our initial investigation efforts was to validate whether func-
tionalization of the N(1)–R1 group of the 1,6-naphthyridin-4(1H)-one
moiety could be tolerated for c-Met potency. Therefore, we initially
designed methyl (4a), ethyl (4b), isopropyl (4c), 2-methoxyethyl (4d),
and 2-methylpropan-2-ol (4e) substituents while keeping the OR4 and
R3 groups as methoxy and fluorine, respectively. As expected, com-
pounds 4a–e all displayed comparable c-Met inhibitory activities to that
of compound 3, suggesting that substitution onto the N(1)–R1 group
could be tolerated for c-Met potency; more importantly, a breakthrough
level in VEGFR-2 selectivity (IC50,VEGFR-2 > 3000 nM) was exhibited
by all of these compounds, indicating that substitution on the N(1)–R1
group could, indeed, improve VEGFR-2 selectivity. Encouraged by this,
the 7-OR4 group was optimized immediately, hypothesizing that opti-
mization of the 7-OR4 group could lead to the improvement of c-Met
potency while retaining high VEGFR-2 selectivity. We were gratified to
find that a consistent improvement of c-Met potency was observed by
replacing the methoxy group with methoxyethyl (4f), alicyclic amines
(4g), and 2-methylpropan-2-ol (4h) groups (see Table 1); where com-
pound 4h (IC50 = 12.7 nM, comparable to that of 2/Cabozantinib)
displayed a 3.3-fold increase in c-Met potency, compared to that of
compound 3, while maintaining high VEGFR-2 selectivity (IC50,VEGFR-
2. Result and discussion
2.1. Chemistry
The designed compounds 4a–r were synthesized following the
general synthetic route illustrated in Scheme 1. Based on our previous
study,20 chlorinated 1,6-naphthyridone and aromatic amine were de-
signed as two important synthetic building blocks. All compounds could
be synthesized, in a modular fashion, by using aromatic nucleophilic
substitution to join elaborated chlorinated 1,6-naphthyridone and
aromatic amine. The aromatic amine building blocks were obtained
using three steps, starting from our reported material 5.24,27 The de-
methylation of staring material 5 yielded 4-(2-fluoro-4-nitrophenoxy)
quinolin-7-ol 6. Then, a subsequent substitution reaction of inter-
mediate 6 with different halides or substituted oxiranes under a K2CO3/
DMF system generated the O-substituted intermediate 7. Finally, re-
duction of the intermediates 7 using iron powder produced the key
building blocks 8b–8d; the key building block 8a could be easily ob-
tained by the direct reduction of the starting material 5. The 1,6-
> 3000 nM), indicating that the 2-methylpropan-2-ol group was the
2
most suitable substituent for the 7-OR4 position. Due to the high c-Met
potency and great selectivity profile against VEGFR-2 of compound 4h,
a more detailed SAR study of the N(1)–R1 position was carried out, with
the 2-methylpropan-2-ol group serving as an optimized 7-OR4 sub-
stituent. It was found that compound 4i, possessing an ethyl substituent
2

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
H. Xu, et al.
Bioorganic&MedicinalChemistry28(2020)115555
Scheme 1. Reaction conditions and re-
agents: (i) BBr3, dry DCM, TEBA, 20 h, 68%;
(ii) R4Xor oxirane, K2CO3, DMF, 70 °C, 7 h;
(iii) Fe, NH4Cl, reflux, EtOH/H2O = 3/1,
1 h; (iv) K2CO3, DMF, R.T. to 50 °C; and (v)
P-toluenesulfonate of 4a-b, 4d-g, 4j, 4n-o,
4r: PTSA, propan-2-ol, 90 °C or vi) 4 h
Hydrochloride: HCl, propan-2-ol, 90 °C or
vii) DIPEA, EA, R.T..
However, the consistent introduction of a 2-methoxyethyl group (4m,
IC50 of 76.1 nM) resulted in a 3.4-fold loss of potency, compared with
compound 4k, and compound 4n (the demethylation product of 4m)
exhibited a slight increase in potency, compared with compound 4m
(40.0 nM vs 76.1 nM, 2.0-fold). Compounds 4o (IC50 of 41.0 nM) and
4p (IC50 of 40.3 nM), bearing two sterically bulky substituents, dis-
played comparable potency to that of compounds 4n and 3. The mor-
pholine-substituted derivative 4q (IC50 of 117.2 nM) showed a 9.2-fold
loss of potency, compared to that of compound 4h. Finally, compound
4r, which was generated by replacing the terminal 4-fluorophenyl ring
of 4h with a phenyl ring, displayed a comparable inhibitory activity
(IC50 of 10.6 nM) to that of compound 4h.
Table 1
Activities of 4a–r against c-Met and VEGFR-2 kinasea and metabolic stability in
human liver microsomes (HLMS).
Compound
R1
R2(OR4)
R3
IC50, nMa
c-Met
HLMS
T1/2, min
VEGFR-2
4a
4b
4c
4d
4e
Me
Et
OMe
OMe
OMe
OMe
OMe
F
F
F
F
F
32.5
29.1
58.7
41.0
31.2
> 3000
> 3000
> 3000
> 3000
> 3000
351.7
–
–
–
–
i-Pr
2.3. Molecular docking study
Molecular docking experiments were further performed to explain
the selectivity of 4r between c-Met and VEGFR-2. As shown in Fig. 3A,
compound 4r was located deep in the c-Met pocket, and since five key
H-bond interactions were observed in the binding mode: the first was
established between the carbonyl group in block C and residue
Asp1222, the second was established between the amine group in block
C and residue Asp1222, the third formed between the N atom in block C
and Lys1110, the fourth formed between N atom in block A and
Met1160, the fifth formed between hydroxyl group in block A and
Tyr1159, 4r displayed comparable c-Met enzymatic activity to that of 2.
As shown in Fig. 3B, compound 4r was located deep in the VEGFR-2
pocket, only two key H-bond interactions were observed in the binding
mode: one was established between the N atom in block C and Lys868,
another formed between N atom in block A and Cys919. The deflection
of block B of 4r resulted in a disappearance of two H-bond interactions
between block C of 4r and the residue Asp1046. The VEGFR-2 allele
amino acid of Tyr1159 in c-Met was Phe918, which led to a dis-
appearance of the H-bond interaction between hydroxyl group in block
A and Tyr1159. And, the distance between methyl group in block C and
the residue Glu885 was just 2.8 Å, which led to a steric conflict. In
terms of van der Waals interaction, the hydrophobic interactions of 4r
with Leu1157, Phe1134, and Phe1200 in c-Met were stronger than the
allele residues Thr916, Ile892, and Cys1024 in VEGFR-2. Together,
these results suggest that the differences of key residues in active pocket
between c-Met and VEGFR-2 led to the selectivity of 4r. As shown in
Fig. 3D, compound 2 was located deep in the VEGFR-2 pocket, and
three key H-bond interactions were observed in the binding mode: the
first was established between the carbonyl group in block C and residue
Asp1046, the second was established between the carbonyl group in
block C and residue Lys868, the third formed between N atom in block
A and Cys919. In terms of van der Waals interaction, the hydrophobic
4f
Me
Me
F
35.6
> 3000
745.4
4g
F
22.0
> 3000
257.2
4h
4i
4j
4k
4l
4m
4n
4o
4p
4q
4r
Me
Et
i-Pr
F
F
F
F
F
F
F
F
F
F
H
12.7
9.5
> 3000
> 3000
> 3000
> 3000
> 3000
> 3000
> 3000
> 3000
> 3000
> 3000
> 3000
319.5
164.8
–
–
–
–
–
–
44.6
58.7
22.4
76.1
40.0
41.0
40.3
117.2
10.6
Bn
–
–
Me
1061.8
2 (Cabozantinib)
8.4
7.0
–
3
41.4
71.1
–
a
In vitro kinase assays were performed with the indicated purified re-
combinant c-Met or VEGFR-2 kinase domains (nM).
at the N(1)–R1 position, was as potent as compound 4h (IC50 of 9.5 nM,
1.3-fold to that of 4h). Next, the sterically bulky substituents isopropyl
(4j, IC50 of 44.6 nM) and cyclopropylmethyl (4k, IC50 of 58.7 nM) were
introduced, consistent losses of c-Met potency were observed (3.5-fold
and 4.6-fold compared with compound 4h, respectively), suggesting
that steric effects are the main cause for the slight loss of c-Met potency.
Interestingly, the introduction of an oxiran-2-ylmethyl group (4l, IC50
of 22.4 nM) exhibited a 2.6-fold increase of c-Met potency, compared
with compound 4k; it is hypothesized that the introduction of the
oxygen atom formed a new interaction with c-Met protein, positively
protecting against the inhibitory activity as a result of steric effects.
3

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
H. Xu, et al.
Bioorganic&MedicinalChemistry28(2020)115555
Fig. 3. (A) The proposed binding mode of 4r with c-Met; (B) The proposed binding mode of 4r with VEGFR-2; (C) The binding mode overlay of 4r with c-Met (green)
and 4r with VEGFR-2 (light purple); (D) The proposed binding mode of 2 with VEGFR-2; (E) The proposed binding mode of 3 with VEGFR-2; (F) The proposed
binding mode of 4a with VEGFR-2. Each dashed red line represents hydrogen bonds between residues with 2, 3, 4a or 4r.
interactions of 2 with Val848, Leu840, Ile892 and Cys1024 were ob-
served. Better binding efficiency of 2 displayed better VEGFR-2 enzy-
matic activity to that of 4r. As shown in Fig. 3E and 3F, compounds 3
and 4a were located deep in the VEGFR-2 pocket, two same key H-bond
interactions were observed in the binding mode: one was established
between the N atom in block C and Lys868, another formed between N
atom in block A and Cys919. However, the distance between methyl
group in block C of 4a and the residue Glu885 was just 2.8 Å and was
smaller than the distance between the N atom in block C of 3 and
Glu885 4.8 Å, which led to a steric conflict. Therefore, these results
suggest that the steric conflict of between methyl group and the residue
Glu885 in active pocket of VEGFR-2 led to the selectivity of 4a.
2.5. In vitro antiproliferative activity
Antiproliferative activities of the selected compounds against three
c-Met dependent cancer cell lines including U-87 MG (human glio-
blastoma), NIH-H460 (human lung cancer), and MKN-45 (human gas-
tric cancer) were also evaluated in vitro using compound
2
(Cabozantinib) as a positive control. As shown in Table 2, compound 4h
showed only moderate to weak antiproliferative activity. Compound 4r
displayed good and broad-spectrum antiproliferative activity (IC50 of
3.4–15.9 μM) against the tested three cancer cell lines, which is com-
parable to or slightly less active than compound 2 (Cabozantinib, IC50
of 4.3–6.7 μM) (see Table 2).
2.6. In vitro and in vivo PK study
2.4. In vitro kinase activity profiling
In order to evaluate the in vivo pharmacokinetic (PK) profiles of the
highly potent compounds 4h and 4r quickly and efficiently, the PK
properties of 4h were first determined in a rat model. To our delight,
compound 4h displayed favorable overall PK profiles after oral dose (see
Table 3), with a long half-life (8.1 h), moderate maximal plasma con-
The inhibitory activity of compound 4r against a panel of other forty
five kinases was also assayed. As shown in Fig. 4, compound 4r only
displayed high inhibitory effects (inhibition rate > 80% in 1 μM)
against AXL, Mer and TYRO3 kinases. These kinase profiling data
suggested that compound 4r could be used a novel scaffold for further
kinase selectivity enhancement.
centration (Cmax = 0.8 μg/ml), high plasma exposure (AUC0-∞
=
10.1 h*µg/mL), favorable total clearance (CL = 0.5 l/h/kg), and
Fig. 4. Preliminary results of kinase activity profiling of 4r (Inhibitory rate in 1 μM).
4