Discovery and optimization of a series of 3-substituted indazole derivatives as multi-target kinase inhibitors for the treatment of lung squamous cell carcinoma

European Journal of Medicinal Chemistry 163 (2019) 671e689

Contents lists available at ScienceDirect

European Journal of Medicinal Chemistry

journal homepage: h ttp://www.elsevier.com/locate/ejmech

Research paper

Discovery and optimization of a series of 3-substituted indazole

derivatives as multi-target kinase inhibitors for the treatment of lung

squamous cell carcinoma

,

c

,

Qi Wang ^{a 1}, Yang Dai ^{b 1}, Yinchun Ji ^{b 1}, Huanyu Shi ^{a c}, Zuhao Guo ^{a c}, Danqi Chen ^a,

Yuelei Chen ^a, Xia Peng ^b, Yinglei Gao ^b, Xin Wang ^a, Lin Chen ^a, Yuchen Jiang ^b

,

c

Meiyu Geng ^{b c}, Jingkang Shen ^a, Jing Ai ^{b **}, Bing Xiong ^a

,

c

,

, c, *

^aDepartment of Medicinal Chemistry, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 555 Zuchongzhi Road, Shanghai, 201203, China

^bDivision of Anti-tumor Pharmacology, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 555

Zuchongzhi Road, Shanghai, 201203, China

^cUniversity of Chinese Academy of Sciences, NO.19A Yuquan Road, Beijing, 100049, China

a r t i c l e i n f o

a b s t r a c t

Article history:

Although lung adenocarcinoma patients have beneﬁted from the development of targeted therapy, pa-

tients with lung squamous cell carcinoma (SqCC) have no effective treatment due to the complexity and

heterogeneity of the disease. Therefore, basing on the genetic analysis of mutations in lung squamous cell

carcinoma to design multi-target inhibitors represents a potential strategy for the medical treatment. In

this study, through screening an in-house focused library, we identiﬁed an interesting indazole scaffold.

And following with binding analysis, we elaborated the structure-activity relationship of this hit com-

pound by optimizing four parts guided by the DDR2 enzymatic assay, which resulted in a potent lead

compound 10a. We conducted further optimization of dual enzymatic inhibitions towards FGFR1 and

DDR2, two important kinases in lung squamous cell carcinoma. Finally, from the cellular antiproliferative

activity tests and in vivo pharmacokinetic test, 3-substituted indazole derivative 11k was found to be a

promising candidate and subjected to in vivo pharmacology study with the mouse xenograft models,

demonstrating profound anti-tumor efﬁcacy. Additional in vitro druglike assessment reinforced that

compound 11k could be valuable for SqCC drug development.

Received 25 September 2018

Received in revised form

4 December 2018

Accepted 6 December 2018

Available online 7 December 2018

Keywords:

Squamous cell carcinoma

Multi-target inhibitor

DDR

FGFR

1. Introduction

which poses a grand challenge for drug development [1]. The

predominant subtype of lung cancer is non-small cell lung cancer

Lung cancer is the leading cause of cancer-related death

worldwide, accounting for about one-quarter of all cancer deaths

and in USA alone resulting more than 150,000 people deaths

annually. Despite of the great advances in lung cancer treatment,

the ﬁve-year survival rate of lung cancer has not been improved

signiﬁcantly over the last three decades, and remains only about 5%,

(NSCLC), covering approximately 80%e85% of all lung cancers [2].

Adenocarcinoma (ADC, approximately 40e50% of cases) and

squamous cell carcinoma (SqCC, approximately 20e30%) are the

main histological subtypes of NSCLC [3]. Along with the advances of

human genomics and post-genomics technologies, comprehensive

target identiﬁcations have been launched for better understanding

the pathology of non-small cell lung cancer. Various oncogenic-

driven proteins in NSCLC were validated as drug targets, and

many small-molecule drugs targeting these onco-proteins have

been approved by FDA, such as geﬁtinib, afatinib and osimertinib

for EGFR mutants (15%), ceritinib and alectinib for ALK rearrange-

ment (5%), vemurafenib, dafrafenib and trametinib for BRAF-V600E

mutants (1e3%) [4]. However, these targeted drugs have largely

been conﬁned to patients with adenocarcinoma of NSCLC. While in

the case of SqCC, the drug development is lagged behind due to the

* Corresponding author. Department of Medicinal Chemistry, Shanghai Institute

of Materia Medica, Chinese Academy of Sciences, 555 Zuchongzhi Road, Shanghai,

201203, China.

** Corresponding author. Division of Anti-tumor Pharmacology, State Key Labo-

ratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of

Sciences, 555 Zuchongzhi Road, Shanghai, 201203, China.

E-mail addresses: jai@simm.ac.cn (J. Ai), bxiong@simm.ac.cn (B. Xiong).

These authors contributed equally.

1

https://doi.org/10.1016/j.ejmech.2018.12.015

672

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

poor understanding the pathology of this speciﬁc subtype of NSCLC,

and the standard ﬁrst-line therapy is still limited to the platinum-

based chemotherapy [5].

improving the potency to FGFR1 and DDR2, as well as the cellular

antiproliferation activity. Finally, on the basis of the bioactivities

and in vivo pharmacokinetic proﬁles, compound 11k was selected

for in vivo pharmacology evaluation, and the result showed it has

signiﬁcant anti-tumor activity in the xenograft mouse models,

demonstrating its potential utility in treatment of SqCC lung cancer.

SqCC is an intricate and heterogeneous disease and most genetic

alternations found in cell cycle regulation and squamous differen-

tiation are belong to receptor tyrosine kinases (RTK) related

signaling pathways [6]. Statistically, about 50e80% of SqCC have a

known oncogenic driver mutation [7], these mutations in receptors

or protein kinases can stimulate a complex cascade of signaling

pathways such as the RAS-RAF-MEK-ERK or MAPK, PI3K-AKT-

mTOR or JAK-STAT pathways. In the last decade, genetic aber-

rances identiﬁed and characterized in SqCC mainly include PIK3CA

2. Results and discussion

2.1. Hit identiﬁcation and binding mode analysis

To identify the starting molecules for development, an in-house

protein kinase inhibitor library containing 167 diversiﬁed mole-

cules was subjected to DDR2 inhibition assay, and compound 1

with indazole scaffold was found to be a moderate inhibitor,

showing enzymatic inhibition ratio about 76% at concentration of

mutation, FGFR1 ampliﬁcation, PDGFR-a ampliﬁcation/mutation

and DDR2 mutation [8]. However, the relationship between the

single genetic alteration and therapeutic response is not always

straightforward and other factors also play important roles in

modulating the response to a targeted therapy [9]. For example,

FGFR1 ampliﬁed tumors were dependent on FGF ligands and

increased ligand levels led to reduced sensitivity to FGFR1 inhibi-

tion; dasatinib treated lung squamous cells could bypassed acti-

vating a panel of RTKs and their signaling adaptor complexes [10];

Generally, these frequent occurring mutations in driver genes

implicated the complexity of personalized therapy in SqCC [11,12].

Therefore, rational drug discovery for more effective multi-target

inhibitors may be necessary to treat SqCC patients. And among

them, we are particularly interested in two RTKs, Fibroblast Growth

Factor Receptor 1 (FGFR1) and Discoindin Domain Receptor 2

(DDR2), due to their important roles in SqCC.

20 mM. We performed the docking study to predict the binding

mode of this hit compound with DDR2. Currently, there is no solved

crystal structure of DDR2 deposited in PDB database, therefore

homology modeling protocol was adopted, and the crystal struc-

ture of DDR1 (PDB entry 5FDP) was selected as template to build

DDR2 model. Then Glide software was used for docking study, and

the predicted binding conformation was depicted in Fig. 1A. From

the docking prediction, compound 1 binds as a typical Type-II

(DFG-out) kinase inhibitor, using the nitrogen atoms in indazole

ring to form essential hydrogen bonds with hinge segment of DDR2.

The amide group of compound 1 strengths the interactions by

hydrogen bonding with the nearby residues Glu625 and

Asp728(Fig. 1A), while the methyl pyridinone was situated at the

hydrophobic sub-pocket near to the C-helix.

Type-II kinase inhibitors preferentially bind to the inactive

conformation of the protein kinases, and requisitely interact with

the hinge part that can contribute up to 40%e60% of the total

binding free energy of the inhibitor. Therefore, this part is very

essential for optimizing the activity of kinase inhibitors. Beside of

the obvious H-bond interaction between the backbone nitrogen of

Met657(Fig. 1A)and the N2 atom of indazole of compound 1, there

are two options of making further hydrogen bonding interaction

with carbonyl oxygen atom of Asp702 or Met657. By surveying the

reported type-II kinase inhibitors, we found that meta-CF₃

substituted benzene ring was usually employed to occupy the

allosteric sub-pocket near the C-helix in the inactive conformation

of kinase. Therefore, based on above-mentioned rationales, com-

pounds 2e5 were synthesized to verify the predicted binding mode

and to determine the scaffold for further optimization.

The bioactivities of compounds 2e5 were listed in Table 1. When

the group N-(3-substituted)-3-(triﬂuoromethyl) benzamide was

shifted from 5-position of indazole (3) to 6-position (2), activity

was dramatically enhanced to moderate inhibitory rate at 100 nM.

The bioactivity of compound with Methyl substituted on the ni-

trogen atom of indazole (4) was dramatically decreased, conﬁrming

that the NH of indazole played an important role in H-bond inter-

action with the hinge residues. Although an amine introduced to 3-

position may participate in forming an additional hydrogen

bonding with the hinge of kinase domain, however, as shown in

Table 1, the compounds (5 and 6) displayed signiﬁcantly reduced

bioactivity when the amine group was attached to the indazole.

FGFR1 is one of the most frequently ampliﬁed genes in human

cancer. Activation of FGFR1 led by endogenous ligand binding ini-

tializes three downstream pathways: the Ras-MAP kinase pathway,

the PI3K-Akt-mTOR pathway and PLCg

-Ca^2þpathway to control

cell proliferation, survival and differentiation [13]. Additionally,

FGFR1 ampliﬁcation occurs frequently in SqCC and can serve as a

negative prognostic marker in patients with early-stage SqCC [14].s.

The discoindin domain receptor (DDR2) is a non-integrin-type

receptor in collagen-mediated signaling with tyrosine kinase ac-

tivity [15]. DDR2 plays

a

critical role as regulator for

epithelialemesenchymal transition [16] and participates in cross-

talk with other cell surface receptors such as integrins and insulin-

like growth factor-1 receptor [17]. When interacts with the

collagen in the extracellular matrix, the DDR2 was activated, further

adhered the cell-cell connections, remodeled the extracellular ma-

trix, transduced the signals in the cell and ﬁnally increased the cell

proliferation [18]. Recent studies indicated that continuous activa-

tion of DDR2, whether through mutation or overexpression, was

frequently found in SqCC lung cancer cell lines. And inhibition of

DDR2 activity can suppress collagen-induced colony formation and

cell proliferation. Utilizing the sequencing method, Hammerman

et al. found that DDR2 mutations, similar to the incidence of EML-

ALK in lung adenocarcinomas, are present in 4% of lung SqCCs [12]

and most of these mutants are oncogenic and associated with gain-

of-function phenotype of cancer transformation, providing solid

evidence to select DDR2 as the drug target for SqCC lung cancer.

In the present study, we initially placed our efforts on opti-

mizing the DDR2 activity, because of its importance in SqCC and

limited reports of DDR2 inhibitors. To identify new chemotypes, we

screened an in-house focused kinase library against DDR2 in an

enzymatic assay. After obtaining the hit compound and probing the

interaction mode with ﬁrst round of structural modiﬁcation, we

regarded it binding to DDR2 as type II kinase inhibitor. Then, sys-

tematic SAR study was conducted to optimize the DDR2 inhibition

activity. After obtaining the lead compound 10a, we intended to

optimize the DDR2 and FGFR1 activities simultaneously and

revealed that N-substituted pyrazole played an important role in

2.2. Optimizing the groups B and C

After the identiﬁcation of proper hinge binder, the next step was

to identify a proper linker between group A and group D. Refering

to the reported type-II kinase inhibitors, Group C usually formed

two and more H-bonds with a glutamic acid residue in the C-helix

and an asparagic acid residue in the Asp-Phe-Gly (DFG) motif of

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

673

Fig. 1. A) Docking compound 1 into DDR2 kinase domain. And B) Optimizing four groups in SAR study based on the predicted binding mode. Group A, Hinge-binding Region. Group

B, Linker between A and C. Group C, H-binding Region with Glu625 and Asp728. Group D, Tail in the allosteric region.

active loop of kinase domain. Amide and urea were typical func-

tional groups to fulﬁll this H-bond capability. Therefore, amide and

reversed amide were synthesized (3 and 8, Table 1) and the activ-

ities from the DDR2 enzyme assay showed no apparent difference.

But when urea was applied to the molecule (7), a dramatic loss in

activity indicated that the length of linker is critical for optimal

interaction between the group D and the hydrophobic sub-pocket.

Substituent on the benzene of group B always plays a critical role in

the binding activity in Type II-kinase inhibitors. Therefore, methyl

group was introduced at different positions to scan the position of

“Flag Methyl”. As shown in Table 2, methyl at R¹position (9a)

dramatically enhanced the enzymatic activity of DDR2, which

indicated that this “Flag Methyl” was situated into a small hydro-

phobic cavity in the back pocket of DDR2 kinase. Further, different

substituents were also applied at the R¹position (9e-9g), and the

result indicated that larger or smaller substituents such as

methoxyl and ﬂuorine had detrimental impact on binding in-

teractions with this small hydrophobic cavity (see Table 3).

showed weak bioactivities in the FGFR1 enzymatic assay. To gain

deep insight into the molecular basis on inhibitory activity, we

analyzed the putative binding mode of compound 10a with mo-

lecular docking with the binding site model of FGFR1 protein pre-

pared from the PDB structure 4V04 (SI Fig. S3). From the docking

prediction, the residue Tyr563 at the hinge segment drew our

attention, as it might form a

p-p stacking interaction if we intro-

duce an aromatic ring at the 3-position of indazole, which will

enhance the binding energy with the FGFR1 protein. Therefore,

compound 11a was synthesized to verify our assumption. As shown

in Table 4, compound 11a displayed signiﬁcant potency not only in

the FGFR1 and DDR2 enzymatic assays, but also in the cellular

antiproliferative tests against the FGFR1-and DDR2-sensitive cell

lines KG-1 and NCI-H2286 respectively. Following this, analogs

(11b-11e) were prepared by adding different polar groups and

showed similar activities in enzymatic assays. We selected com-

pound 11a to perform the primary mouse PK assessment, which

resulting disappointed PK properties (Table 5). To solve this issue,

an approach of replacing the benzene ring with a more druglike

functional group pyrazole was adopted, producing compounds 11f-

11r. From the bioactivity tests, four compounds (11g, 11h, 11k and

11r) showed an obvious improvement on the activities.

2.3. Exploration of the impact of group D on the binding

SAR study so far had led to compound 9a as a basic skeleton

because it showed a good potency towards DDR2 in the enzyme

assay. Base on compound 9a, exploration around group D was

conducted to assess its impact on the binding activity. Firstly, tri-

ﬂuoromethyl was replaced by some lager hydrophobic group such

as isopropyl and cyano substituted isopropyl (10b and 10c), which

showed similar bioactivities as 10a. When methylene was inserted

between the amide and benzene ring to change the chain length

(10d), the activity of DDR2 was slightly decreased. Piperazine as a

solubility-enhancing group was introduced into the molecules

(10e-10j), resulting in an interesting SAR: direct substitution of

triﬂuoromethyl with N-methyl piperazine (10e) sharply reduced

the potency on DDR2; shifting the N-methyl piperazine to para-

position (10f) showed similar inferior effects on bioactivity;

extending the N-methyl piperazine by adding methylene (10i) to

10e signiﬁcantly enhanced the enzymatic activity on DDR2, which

even ampliﬁed from the comparison of 10h and 10f; however,

adding the piperazine motif to the molecules with longer linker 10g

showed diminished activity with respect to both 10i. Based on the

SAR study described above, triﬂuoromethyl and piperazine motifs

were integrated into the molecules (10j-10m). As expected, these

four compounds showed high potency on DDR2 enzymatic assay.

2.5. In vivo PK study

Overall, considering the activity in enzymatic assays and cellular

antiproliferative assays, four compounds 11g, 11k, 11m and 11r

with different structural features were selected for in vivo phar-

macokinetic assessment (CD-1 mice) to identify the best one for

further evaluation. Among the selected compounds, 11k exhibited

excellent pharmacokinetic properties, with

a good exposure

(AUC_0ꢀ∞¼ 7636 ng*h/mL) and moderate half-life of 3.05 h. Incor-

porating previous optimization at Group D, we also prepared

compound 12 by adding N-methyl piperazine to the tail of 11k with

the hope to further improve the solubility as well as the pharma-

cokinetic properties. From Table 6, although compound 12 has

excellent inhibitory activities against FGFR1 and DDR2, it had much

inferior PK parameters with respect to 11k, especially the exposure

parameter. To ﬁnd the reason for this dramatic disparity, compound

11k and 12 were subjected to Caco-2 monolayer permeability assay

to assess the absorption issue. As shown in Table 6, although

compound 12 had higher solubility, it proved to be a substrate of P-

glycoprotein (P-gp) transporter with an efﬂux ratio more than 42

folds. Taking together, compound 11k was selected for further

study.

2.4. Improving the potency on FGFR1 and DDR2

Since FGFR1 also plays an important role in development of

SqCC, we selected DDR2-potent compounds 10a-10c and 10k-10n

to test the FGFR1 enzymatic inhibition. Generally, these compounds

2.6. Kinase selectivity proﬁle of 11k

The selectivity proﬁle of compound 11k was carried out over a

674

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

Table 1

SAR exploration of the hinge-binding region and group C^a.

Compd.

%@1000 nM

Inhibitory Rate (IR)

28.9%

%@100 nM

Inhibitory Rate (IR)

14%

2

3

90.8%

46.8%

4

5

29.8%

35.9%

7.6%

13.1%

6

7

24.8%

53.3%

22.8%

20.1%

8

84.7%

58.2%

Footnote.

a

Inhibition ratio are given as the mean from two separate experiments. And the positive control compound Dasatinib has the IC₅₀value 1.4 0.1 nM.

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

675

Table 2

Table 3

SAR exploration of group B.^a..

SAR Exploration of Tail part.^a..

Compd.

R

DDR2 IC₅₀(nM)

10a

1.5 0.1

1.2 0.2

2.7 0.3

Compd.

R1

R2

R3

R4

DDR2

IC50 (nM)

10b

10c

9a

9b

9c

9d

9e

9f

CH3

H

F

H

CH3

H

CH3

H

CH3

H

5.5 0.5

27.4%@ 100 nM^b

37.3%@ 100 nM

41.8% @ 100 nM

64.3 0.6

10d

10e

8.0 0.2

Cl

OCH3

H

9.5 1.3

67.9 2.4

9g

37.3%@100 nM^b

Footnote.

a

IC₅₀values or inhibition ratios are calculated from two separate experiments.

And the positive control compound Dasatinib has the IC₅₀value 1.4 0.1 nM.

10f

25.8%@100 nM

23.2%@100 nM

b

Means the enzymatic inhibition rate at ligand concentration 100 nM.

10g

panel of 30 kinases. As presented in Fig. 2, compound 11k was

clearly acting as a multi-target inhibitor, especially for the receptor

tyrosine kinase subfamily, and showed high inhibitory activities

towards VEGFR2 and EPH-B2. It also inhibited several kinases at

moderate level, such as c-Src, CSF-1R, FLT-3, Abl, CDK6 and PDGFR-

b

. For the rest of tested kinases,it only showed high inhibition rate

10h

10i

10j

6.7 0.5

75.9 7.6

4.7 0.8

at 1 M concentration. Although we only tested limited kinases, the

data implied that 11k has certain selectivity in the kinome.

m

2.7. Compound 11k blocks protein kinase phosphorylation and

downstream signaling in cells

To investigate the kinase inhibitory activity against FGFR and

DDR2 of compound 11k at cellular level, both FGFR2-addicted

cancer cell line SNU-16 cells and DDR2-dependent NCI-

H2286 cells were tested. Firstly, we analyzed its’ effects on the

phosphorylation of FGFR and its key downstream signaling mole-

cules Erk in SNU-16 cells. AZD4547,a selective FGFR1-3 inhibitor,

was used as the positive control. 11k showed signiﬁcant inhibition

of FGFR2 phosphorylation in a dose-dependent manner in SNU-

16 cells. The phosphorylation of ERK was also unanimously

inhibited (Fig. 3A). Thus, 11k potently inhibits FGFR signaling acti-

vation. Similar result could be observed in NCI-H2286, a DDR2-

dependent cell line with persistent activation of mutations in the

DDR2 (I638F) site. DDR2 activates important signaling components

such as ERK and PI3K/AKT [15,17]. We utilized multi-target inhib-

itor dasatinib as the positive control. 11k showed signiﬁcant inhi-

bition of DDR2 phosphorylation in a dose-dependent manner in

H2286 cells. The phosphorylation of AKT and ERK were also

unanimously inhibited (Fig. 3B). All these data indicated that 11k

could block FGFR and DDR2 signaling at the cellular level.

10k

3.9 0.7

10l

3.4 0.1

7.0 0.2

10m

Footnote.

a

IC₅₀values or inhibition ratios are calculated from two separate experiments.

And the positive control compound Dasatinib has the IC₅₀value 1.4 0.1 nM.

b

means the enzymatic inhibition rate at ligand concentration 100 nM.

cancer cell line NCI-H2286 (93.0 nM). Collectively, these results

highlighted the antiproliferative activities of 11k on FGFR-altered

and DDR2-driven cancer cell lines.

2.8. Cellular antiproliferation activity

To further evaluated the impact of 11k on FGFR and DDR2

overactivation-mediated cancer cell proliferation, ﬁve cancer cell

lines harboring frequently occurring oncogenic forms of different

FGFR members and DDR2 were chosen. As shown in Table 7, 11k

signiﬁcantly inhibited the proliferation of FGFR-driven cancer cell

lines with IC₅₀values of 108.4, 93.4, 31.8 and 306.6 against KG-1,

SNU-16, NCI-H716 and UMUC14 respectively. Moreover, 11k

demonstrated substantially activity against the DDR2-driven

2.9. In vivo antitumor efﬁcacy in mouse models in SqCC

Classical FGFR-dependent NCI-H1581 lung cancer cell line-

derived xenografts were ﬁrst exploited to evaluate the vivo efﬁ-

cacy of compound 11k. Nude mice bearing NCI-H1581 tumors were

randomized and treated with compound 11k at doses of 10 or

20 mg/kg daily consecutively. As shown in Fig. 4A, compound 11k

Table 4

SAR Exploration of hinge part.^a..

Compd.

R

FGFR1

KG-1

DDR2

NCI-H2286

IC₅₀(nM)

19.3 10.7

IC₅₀(nM)

136.4 5.0

IC₅₀(nM)

8.8 0.9

11a

89.5 5.6

11b

25.4 5.5

300.3 4.2

6.5 0.2

45%@100 nM^b

11c

11d

20.7 4.8

119.5 8.9

561 12.9

6.2 0.2

13%@100 nM

140.4 2.5

55.4 32.0

21.6 2.1

11e

9.0 2.6

176.7 4.8

8.7 0.01

56%@1000 nM

11f

>100

>1000

55.6 27.3

4.4 4.4

>1000

11g

6.4 0.4

65.7 12.9

24.3 2.5

11h

11i

6.7 1.0

257 31.1

358 38.7

2.6 1.7

3.2 0.8

19.8 0.01

12.9 2.6

41%@100 nM

11j

20.1 8.8

31.1 3.3

452 10.7

108.4 2.1

7.7 0.7

3.2 0.2

31%@100 nM

93.0 8.7

11k

11l

11.6 1.1

173.6 11.5

2.9 0.1

35%@100 nM

11m

11n

26.8 4.7

11.7 0.7

339 29.5

441 23.8

3.7 0.01

2.7 0.1

196.2 12.7

>1000

11o

11p

11.5 0.6

6.7 1.1

114. 16.4

303.6 1.8

2.6 0.3

6.2 0.1

49%100 nM

69.7 0.4

11q

11r

3.2 0.1

121.1 1.2

93.3 4.3

5.2 0.2

3.8 0.4

45%@100 nM

56.5 5.3

10.1 2.7

Footnote.

a

IC₅₀values or inhibition ratios are calculated from two separate experiments. And the positive control compound Dasatinib has the IC₅₀value 1.4 0.1 nM towards DDR2,

while the positive control compound AZD4547 has the IC₅₀value 1.1 0.4 nM towards FRFR1.

b

means the enzymatic inhibition rate at ligand concentration 100 nM.

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

677

could suppress tumor growth in a dose-dependent manner with

tumor growth inhibition rates (TGI) of 59.7% and 98.1% at doses of

SqCC. SCID mice bearing NCI-H2286 tumors were randomized and

treated with compound 11k at doses of 10 mg/kg for 10 consecutive

days. As shown in Fig. 4B, compound 11k could suppress tumor

growth with tumor growth inhibition rates (TGI) of 82.8%. Dasati-

nib in the positive control group suppressed tumor growth with

tumor growth inhibition rates (TGI) of 50.9% at doses of 50 mg/kg,

demonstrating that compound 11k elicited marked anti-tumor ef-

ﬁcacy in NCI-H2286 tumor model. Although compound 11k

showed less potent than AZD4547 and Dasatinib at blocking the

phosphorylation of FGFR2 and DDR2 in western blot experiment

(Fig. 3), 11k showed better tumor suppression effects in vivo, which

could be due to the fact that 11k acted as a multi-target inhibitor,

while AZD4547 only selectively targets the FGFRs and dasatinib has

very weak inhibitory activity to FGFRs. Another possible reason of

better efﬁcacy may stems from the good pharmacokinetics of

compound 11k, although we did not directed compare with the PK

proﬁles of two positive control compounds.

Table 5

Pharmacokinetics Proﬁles of compound 11a, 11g, 11k, 11 m,11r and 12.

Compd. dosage

mg/kg (po)

10

T_1/2

h

T_maxC_max

AUC_0-t

AUC_0-∞

MRT

h

(ng/ml) (h*ng/ml) (h*ng/ml)

11a

11g

11k

11m

11r

12

e

1.33 15.9

44.0

3475

7636

324

552

66.2

e

10

2.75 1.17 890

3.05 2.00 1125

0.75 0.25 202

2.75 1.33 170

3485

7680

334

563

610.10

3.26

5.71

1.37

3.96

3.2

1.4

1.67 111.2

Table 6

Compound 11k and 12 and the Caco-2 monolayer permeability assay of 11k and 12.

2.10. Primary assessment of drug-likeness of compound 11k

To further evaluate compound 11k as a potential candidate,

herein, we examined it in vitro drug metabolism by using a stan-

dard human and rat microsomal stability assay. The results indi-

cated compound 11k was very stable in the human and rat liver

microsomes (Table 8). As cytochrome P450s (CYPs) are the major

enzymes involved in the metabolism of various xenobiotics, we

selected ﬁve human hepatic CYP enzymes that play a dominant role

in the metabolism of drugs and other xenobiotics to investigate the

risk of drug-drug interaction. As shown in Table 8, compound 11k

had very weak inhibitory activities against the most relevant 3A4

and 2D6, but showed moderate inhibition towards 2C9 and 2C19

subtypes. Since 3A family is responsible for metabolism of about

50% of drugs and 2D6 is responsible for metabolism of about 30% of

drugs [21], we thought compound 11k poses lower risk in drug-

drug interactions. We also checked the time-dependent inhibition

(TDI) of 11k, which is to study the irreversible inhibition of the

compound. If TDI happens, it will take several days for the CYP

enzymes to be regenerated, which will pose serious issues for drug

elimination. Luckily, according to the result in Table 8, no TDI in-

hibition against the ﬁve CYP enzymes was found, even for 2C9 and

2C19. This further reinforced that compound 11k may have lower

risk in drug-drug interaction.

Compound

11k

12

DDR2

3.2 0.2 nM

5.7 1.4 nM

10 mM

12.0%

42.7

Minimal Soluble Concentration

Permeability (Mean Fabs)

Efﬂux Ratio

2

m

M

48.6%

1.1

10 and 20 mg/kg, respectively, indicating that compound 11k has

profound anti-tumor efﬁcacy in NCI-H1581 tumor model. Although

over 200 cell lines for NSCLC have been reported, only about 30 are

SqCC cell lines and even among these, some have been erroneously

categorized and are in fact poorly differentiated carcinomas [19].

Development of pre-clinical mouse models for SqCC has lagged far

behind due to the lack of validated driver mutations, unclear cell or

origin and difﬁculty replicating the complex genomic landscape of

the disease [20]. So we develop NCI-H2286 SCID mouse model for

In recent days, hERG inhibition brought the concern about heart

toxicity, as it may affect the cardiac action potential. This made

hERG inhibitory assay

a necessity for candidate evaluation.

Fig. 2. In vitro enzymatic screen assays of compound 11k.

678

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

Fig. 3. Compound 11k blocks tyrosine kinase phosphorylation and downstream signaling in cells. Cells including SNU-16 (A), NCI-H2286 (B) treated with 11k for 2 h at the indicated

concentrations were lysed and subjected to Western blot analysis.

Table 7

The anti-proliferative effects of 11k against related cell lines.

Cell lines

IC50 (nM)^a

11k

AZD4547

Dasatinib

KG-1

SNU-16

NCI-H716

UMUC14

NCI-H2286

FGFR1 fusion

108.4 2.1

93.4 0.9

31.8 8.8

306.6 48.3

93.0 8.7

5.8 0.6

19.6 0.5

5.9 2.1

4.5 0.1

/

FGFR2 ampliﬁcation

FGFR3 mutation

DDR2 mutation

33.3 6.3

Footnote.

a

Values are the mean SD of two independent assays.

Fig. 4. A) Compound 11k inhibits tumor growth in NCI-H1581 xenografts. Mice were randomly assigned to control vehicle and treatment groups (n ¼ 6 in treated group, n ¼ 12 in

vehicle group) when the tumor volume reached 100e150 mm³. The control groups were given vehicle alone, and the treatment groups received 11k and AZD4547 at the indicated

doses via oral administration once daily for 7 days. B) Compound 11k inhibits tumor growth in NCI-H2286 xenografts. Mice were randomly assigned to control vehicle and

treatment groups (n ¼ 5 in treated group, n ¼ 10 in vehicle group) when the tumor volume reached 100e150 mm³. The control groups were given vehicle alone, and the treatment

groups received 11k and Dasatinib at the indicated doses via oral administration once daily for 10 days. The results are expressed as the mean

SEM with **p < 0.01, and

***p < 0.001 vs control group determined using Student's t-test. The Tumor Volume was measured on the ﬁnal day of the study for the drug-treated mice compared with the control

mice.

Compound 11k did not show an evident inhibitory effect on hERG,

with an IC₅₀of more than 40 M from an automated patch clamp

electrophysiology measurement (Fig. S4 in SI.). Taken together

these druglike assessments, compound 11k showed promise as a

drug candidate for further development.

(triﬂuoromethyl) benzoic acid in DMF under the conditions of

HATU as a condensing agent to obtain the corresponding inter-

mediate 19a-g, which underwent another coupling to yield desired

compounds 9a-g. Compounds 10a-i in Scheme 3 were also ob-

tained by condensation reaction and Suzuki coupling reaction as

above. The active ester solution formed by 3-bromo-4-

methylbenzoic acid and HATU slowly added dropwise to a dilute

solution of 3-amino-5-(triﬂuoromethyl)phenyl) methanol under

low temperature, thereby obtaining intermediate 23, and then

benzyl bromide intermediate 24 is obtained by Appel reaction with

triphenylphosphine and NBS. N-substituted piperazine via

substituting with compound 24 to obtain intermediates 25j-m, and

by Suzuki coupling reaction to obtain ﬁnal product 10j-10m. The

intermediates 27a¡r could be obtained by the coupling of 6-

m

3. Chemistry

Starting from aniline (13) or carboxylic acid (16), the corre-

sponding boronate ester intermediate (14,15,17) was obtained by

condensation with carboxylic acid, isocyanate or amine,and then

coupled with the corresponding halogenated aromatic ring to

obtain compound 2e8 via Suzuki-Miyaura reaction (Scheme 1). In

Scheme 2, different substituted anilines were condensed with 3-

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

679

Table 8

Compound 11k In vitro liver microsome stability test.

HLM Stability

MLM Stability

Clint(ul/min/mg protein)

0

Reversible CYP Inhibition of 11k

(Direct Inhibition Mean,DI)^a

Irreversible CYP Inhibition of 11k (Time Dependent

Inhibition, TDI)^b

3A4-MDZ

3A4-Testo

2D6

2C9

1A2

16%

5%

10%

83%

20%

64%

3A4-MDZ

3A4-Testo

2D6

2C9

1A2

90

100

70

100

110

90

2C19

Footnote.

a

DI<20% No Inhibition; 20% < DI<50% Weak Inhibition; 50% < DI<70% Moderate Inhibition; DI >70% Strong Inhibition.

TDI>200 will be considered as irreversible CYP Inhibition.

b

Scheme 1. Synthesis route of compounds 2e8. ^aReagents and conditions: (i) 3-(triﬂuoromethyl) benzoic acid, HATU, DIPEA, DMF, 25 ^ꢁC; (ii) Aromatic halide, Pd(dppf)₂Cl₂$CH₂Cl₂,

K₂CO₃, 1,4-dioxane, H₂O, 80 ^ꢁC; (iii) 1-isocyanato-3-(triﬂuoromethyl) benzene, THF, 25 ^ꢁC; (iv) 6-bromo-1H-indazole, Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃,1,4-dioxane, H₂O, 80 ^ꢁC; (v) thionyl

chloride, 80 ^ꢁC, refﬂux; then 3-(triﬂuoromethyl) aniline, DIPEA, THF, 0 ^ꢁC (vi) 6-bromo-1H-indazole, Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃, 1,4-dioxane, H₂O, 80 ^ꢁC;

a

Scheme 2. The synthesis route of compounds 9a-9g. Reagents and conditions: (i) 3-(triﬂuoromethyl) benzoic acid, HATU, DIPEA, DMF, 25 ^ꢁC; (ii) 6-(4,4,5,5-tetramethyl-1,3,2-

dioxaborolan-2-yl)-1H-indazole Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃, 1,4-dioxane, H₂O,80 ^ꢁC;

bromo-3-iodo-1H-indazole with different phenyl boronates or

boronic acids at 1:1 equivalent. The following step Suzuki coupling

of 27a-r with intermediate 21a afforded desired compounds 11a-

11r (Scheme 3). Boronic ester 28 were provided from 25k via a

Miyaura borylation reaction. The coupling of 28 and 27k provided

desired compound 12 (see Scheme 4).

properties. Furthermore, in vivo pharmacology evaluations of

compound 11k showed signiﬁcant antitumor activity

(TGI ¼ 90.92%) in FGFR-driven NCI-H1581 xenograft model and

(TGI ¼ 89.00%) DDR2-driven NCI-H2286 xenograft model. These

imply that compound 11k deserves further research.

Previous clinical studies on molecular targeted agents in SqCC

have unfortunately been negative, such as trials involving kinase

inhibitors sorafenib, motesanib, cediranib or iniparib. Squamous

cell lung cancer has long been considered a challenging tumor type

to treat, at least partly due to the lack of clear driver mutations.

Therefore, many questions remain to be solved, such as: Is the

DDR2 mutation concomitant with the appearance of FGFR1

ampliﬁcation, and/or PI3KCA mutations? Does the multi-target

strategy provide enough safety windows for treatment? Is there a

representative SqCC mouse model can be used for drug discovery?

We hope that compounds discovered in present study could

stimulate further study on SqCC drug development.

4. Conclusion

In summary, a novel series of 3-position substituted indazole

derivatives were designed, synthesized, and evaluated as multi-

target kinase inhibitors for the treatment of SqCC. The hit com-

pound was identiﬁed from an in-house kinase library. On the basis

of homology modeling and docking, we identiﬁed it to be a type II

inhibitor, and subjected for SAR explorationby deviding it into four

parts of Hinge binder, Spacer, Linker and Cap part. Among them,

compound 11k showed high potency against FGFR1 and DDR2 ki-

nase with good DMPK (Drug Metabolism and Pharmacokinetics)

680

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

Scheme 3. The synthesis route of compounds 10a-m. ^aReagents and conditions: (i) 4-methyl-3-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl) benzoic acid, HATU, DIPEA, DMF,

25 ^ꢁC; (ii) 6-bromo-1H-indazole, Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃, 1,4-dioxane, H₂O, 80 ^ꢁC; (iii) 3-bromo-4-methylbenzoic acid, HATU, DIPEA, DMF, 25 ^ꢁC; (iv) triphenylphosphine, N-

Bromosuccinimide, Toluene, 0e25 ^ꢁC(v) N-substituted piperazine, DIPEA, DCM, 25 ^ꢁC; (vi) 6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-indazole, Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃,

1,4-dioxane, H₂O, 80 ^ꢁC;

Scheme 4. The synthesis route of compounds 11a-11r and 12. ^aReagents and conditions: (i) Aromatic boric acid(esters), Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃, 1,4-dioxane, H₂O, 80 ^ꢁC; (ii)

21a, Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃, 1,4-dioxane, H₂O, 80 ^ꢁC; (iii) Bis(pinacolato)diboron, Pd(dppf)₂Cl₂$CH₂Cl₂, AcOK, DMSO, 80 ^ꢁC; (iv) 27k, Pd(dppf)₂Cl₂$CH₂Cl₂, K₂CO₃, 1,4-dioxane,

H₂O, 80 ^ꢁC;

5. Experimental section

a Finnigan MAT-95 mass spectrometer or an Agilent technologies

6224 TOF mass spectrometer. Purity of all compounds was deter-

mined by analytical Gilson high-performance liquid chromatog-

raphy (HPLC) using an YMC ODS3 column (50 mm ꢂ 4.6 mm,

5.1. Chemistry

5.1.1. General

5 mM). Conditions were as follows: CH₃CN/H₂O eluent at

¹H NMR (400 MHz) spectra were recorded with a Varian

Mercury-400 High Performance Digital FT-NMR spectrometer with

tetramethylsilane (TMS) as an internal standard. ¹³C NMR (100 MHz

or 125 MHz or 151 MHz) spectra were recorded by using a Varian

Mercury-400 High Performance Digital FT-NMR spectrometer or

Varian Mercury-500 High Performance Digital FT-NMR spectrom-

eter or Varian Mercury-600 High Performance Digital FT-NMR

spectrometer. Abbreviations for peak patterns in NMR spectra:

br ¼ broad, s ¼ singlet, d ¼ doublet, and m ¼ multiplet. Low-

resolution mass spectra were obtained with a Finnigan LCQ Deca

2.5 mL min^ꢀ1ﬂow [containing 0.1% triﬂuoroacetic acid (TFA)] at

35 ^ꢁC, 8min, gradient 5% CH₃CN to 95% CH₃CN, monitored by UV

absorption at 214 nm and 254 nm. TLC analysis was carried out

with glass precoated silica gel GF254 plates. TLC spots were visu-

alized under UV light. Flash column chromatography was per-

formed with a Teledyne ISCO CombiFlash Rf system. All solvents

and reagents were used directly as obtained commercially unless

otherwise noted. Anhydrous dimethylformamide was purchased

from Acros and was used without further drying. All air and

moisture sensitive reactions were carried out under an atmosphere

of dry Argon with heat-dried glassware and standard syringe

techniques.

XP mass spectrometer using

a

CAPCELL PAK C18

M) or an Agilent ZORBAX Eclipse XDB C18

M) in positive or negative electrospray mode.

(50 mm ꢂ 2.0 mm,5

m

(50 mm ꢂ 2.1 mm,5

Low-resolution mass spectra and high-resolution mass spectra

were recorded at an ionizing voltage of 70eV on a Finnigan/MAT95

spectrometer. High-resolution mass spectra were recorded by using

5.1.2. Synthetic procedures

5.1.2.1. N-(3-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)-3-

(triﬂuoromethyl)benzamide(14). 3-(triﬂuoromethyl) benzoic acid

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

681

(4.4g, 22.82 mmol) was stirred in a solvent of dimethyl formamide

for 396.1318, retention time 3.92 min, >99% pure.

(15 mL). The reaction solution was added with HATU (9.55 g,

25.10 mmol), DIPEA (7.54 mL, 45.64 mmol) and 3-(4,4,5,5-

tetramethyl-1,3,2-dioxaborolan-2-yl) aniline (5.00 g, 22.82 mmol,

13), followed by stirring for about 6 h at room temperature. The

reaction mixture was diluted with ethyl acetate and washed with a

saturated aqueous sodium bicarbonate solution and saline. The

organic layer thus obtained was dried over anhydrous sodium

sulfate, ﬁltered, and concentrated under reduced pressure. The

resulting mixture was concentrated to give the crude product,

which was puriﬁed by silica gel column chromatography (eluting

with 0e5% EtOAc in petroleum ether) to provide 14 as a white solid

5.1.4.3. N-(3-(3-amino-1H-indazol-5-yl)phenyl)-3-(triﬂuoromethyl)

benzamide (5). White solid, Yield: 73%. ¹H NMR (400 MHz, Meth-

anol-d₄)

d

8.31 (s, 1H), 8.24 (d, J ¼ 7.9 Hz, 1H), 8.07 (s, 1H), 7.99 (s,

1H), 7.94e7.87 (m, 1H), 7.74 (q, J ¼ 7.4 Hz, 1H), 7.7e7.62 (m, 2H),

7.51e7.43 (m, 2H), 7.39 (dd, J ¼ 8.6, 2.7 Hz, 1H). ¹³C NMR (126 MHz,

DMSO‑d₆)

d 164.50, 150.15, 142.19, 141.44, 139.76, 136.23, 132.31,

130.26, 130.23, 129.68 (d, J ¼ 32.0 Hz), 129.61, 128.63, 125.95, 124.69

(d, J ¼ 4.2 Hz),124.46 (d, J ¼ 272.6 Hz), 122.69, 119.23,119.00, 118.80,

115.20, 110.34. HRMS m/z (ESI) found 397.1285 (M

C

þ

H)^þ,

₂₁H₁₆F₃N₄O^þcalcd for 397.1271, retention time 3.10 min, >99%

pure.

(7.5 g, 84%).¹H NMR (400 MHz, Methanol-d4)

d 8.29 (s, 1H), 8.23 (d,

J ¼ 8.0 Hz, 1H), 8.07 (s, 1H), 7.91 (d, J ¼ 7.6 Hz, 1H), 7.86 (d, J ¼ 8.3 Hz,

1H), 7.75 (t, J ¼ 7.9 Hz, 1H), 7.57 (d, J ¼ 7.3 Hz, 1H), 7.41 (t, J ¼ 7.7 Hz,

1H), 1.38 (s, 12H). MS m/z: 392.1 ([MþH] ^þ).

5.1.4.4. N-(3-(3-amino-1H-indazol-6-yl)phenyl)-3-(triﬂuoromethyl)

benzamide (6). White solid, Yield: 73%. ¹H NMR (400 MHz, Meth-

anol-d₄)

d

8.32 (s, 1H), 8.26 (d, J ¼ 7.8 Hz, 1H), 8.10 (d, J ¼ 2.2 Hz, 1H),

5.1.3. General procedure for the preparation of compounds 2,3,4,5,6

5.1.3.1. N-(3-(1H-indazol-5-yl)phenyl)-3-(triﬂuoromethyl)benzamide

(2). To a resealable vial was added K₂CO₃(112 mg, 0.81 mmol), 14

(160 mg, 0.41 mmol), 5-bromo-1H-indazole (80 mg, 0.41 mmol).

The vial was sealed and evacuated and purged with Ar for 5 min

before addition of PdCl₂(dppf)-CH₂Cl₂Adduct (20 mg, 0.02 mmol),

dissolved in 1,4-dioxane/water (10 mL, 4:1, v/v) was then added to

this solution before the vial was heated to 80 ^ꢁC overnight. The

reaction was cooled to room temperature, which was then brought

to basic using aqueous sodium bicarbonate solution and extracted

with ethyl acetate. The organic layer was collected and distilled

under reduced pressure. The crude residue was puriﬁed via by silica

gel column chromatography (eluting with 0e20% EtOAc in petro-

leum ether) to afford 2 as a white solid (85 mg, 55%). ¹H NMR

7.93 (d, J ¼ 7.6 Hz, 1H), 7.81e7.70 (m, 3H), 7.55 (s, 1H), 7.52e7.48 (m,

2H), 7.33 (d, J ¼ 8.4,1H). ¹³C NMR (126 MHz, DMSO‑d₆)

d 164.55,

149.60, 142.58, 141.91, 139.79, 138.72, 136.22, 132.33, 130.23, 129.71,

129.67 (d, J ¼ 32.0 Hz), 128.66 (q, J ¼ 3.9 Hz), 124.70 (q, J ¼ 4.1 Hz),

124.46 (d, J ¼ 272.5 Hz), 123.16, 121.30, 119.75, 119.62, 117.49, 113.99,

107.50. HRMS m/z (ESI) found 397.1276 (M þ H)^þ, C₂₁H₁₆F₃N₄O^þ

calcd for 397.1271, retention time 2.94 min, >98% pure.

5.1.4.5. 1-(3-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)-3-

(3-(triﬂuoromethyl)phenyl)urea(15). The

3-(4,4,5,5-tetramethyl-

1,3,2-dioxaborolan-2-yl) aniline (200 mg, 0.91 mmol). in 50 ml of

dry tetrahydrofuran was added dropwise to the 1-isocyanato-3-

(triﬂuoromethyl) benzene (172 mg, 0.91 mmol) at 0 ^ꢁC with stir-

ring. The ice-bath was removed after the completion of addition of

the isocyanate and the mixture was stirred at room temperature for

4 h. The precipitate formed was isolated via ﬁltration. The product

as a white solid (350 mg, 94%) was dried in vacuo and carried for-

ward to the next step without further puriﬁcation. ¹H NMR

(400 MHz, Methanol-d₄)

d

8.32 (s, 1H), 8.27 (d, J ¼ 7.7 Hz, 1H), 8.14

(s, 1H), 8.09 (s, 1H), 8.06 (s, 1H), 7.93 (d, J ¼ 7.6 Hz, 1H), 7.80e7.73

(m, 2H), 7.72 (d, J ¼ 7.1 Hz, 1H), 7.65 (d, J ¼ 8.7 Hz, 1H), 7.54e7.50 (m,

1H), 7.48 (d, J ¼ 7.7 Hz, 1H). ¹³C NMR (126 MHz, DMSO‑d₆)

d 164.53,

141.80, 139.90, 139.80, 136.23, 134.49,133.06, 132.32, 130.24, 129.73,

129.67 (d, J ¼ 32.1 Hz), 128.66, 126.00, 124.70, 124.46 (d,

J ¼ 272.4 Hz), 124.01, 122.98, 119.50, 119.32, 118.70, 111.09. HRMS m/

z (ESI) found 382.116 (M þ H)^þ, C₂₁H₁₅F₃N₃O^þcalcd for 382.1162,

retention time 3.62 min, >99% pure.

(400 MHz, DMSO‑d₆) d 9.03 (s, 1H), 8.83 (s, 1H), 8.01 (s, 1H), 7.87 (d,

J ¼ 2.4 Hz, 1H), 7.57 (t, J ¼ 8.3, 1.5 Hz, 1H), 7.54e7.46 (m, 2H),

7.35e7.27 (m, 3H), 1.29 (s, 12H). MS m/z: 407.3 ([MþH]^þ).

5.1.4.6. 1-(3-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)-3-

(3-(triﬂuoromethyl)phenyl)urea (7). To a resealable vial was added

K₂CO₃(112 mg, 0.81 mmol), 15 (181 mg, 0.45 mmol), 6-bromo-1H-

indazole (80 mg, 0.41 mmol). The vial was sealed and evacuated

and purged with Ar for 5 min before addition of PdCl₂(dppf)-CH₂Cl₂

Adduct (20 mg, 0.02 mmol), dissolved in 1,4-dioxane/water (10 mL,

4:1, v/v) was then added to this solution before the vial was heated

to 80 ^ꢁC overnight. The reaction was cooled to room temperature,

which was then brought to basic using aqueous sodium bicarbon-

ate solution and extracted with ethyl acetate. The organic layer was

collected and distilled under reduced pressure. The crude residue

was puriﬁed via by silica gel column chromatography (eluting with

0e30% EtOAc in petroleum ether) to afford the product as a white

5.1.4. Compound 3,4,5,6 was prepared with a similar procedure as

that used for 2

5.1.4.1. N-(3-(1H-indazol-6-yl)phenyl)-3-(triﬂuoromethyl)benzamide

(3). White solid, Yield: 63%. ¹H NMR (400 MHz, Methanol-d₄)

d

8.32 (s,1H), 8.26 (d, J ¼ 8.0 Hz,1H), 8.14 (s,1H), 8.09 (s,1H), 7.93 (d,

J ¼ 7.8 Hz, 1H), 7.87 (dd, J ¼ 8.5, 0.7 Hz, 1H), 7.79 (s, 1H), 7.78e7.72

(m, 2H), 7.54e7.52 (m, 1H), 7.50 (s, 1H), 7.48 (d, J ¼ 1.4 Hz, 1H). ¹³

C

NMR (126 MHz, DMSO‑d₆)

d 164.58, 141.59, 141.00, 139.86, 138.59,

136.21, 133.90, 132.34, 130.24, 129.81, 129.68 (d, J ¼ 32.1 Hz), 128.67,

124.72, 124.46 (d, J ¼ 272.7 Hz), 123.27, 122.78, 121.53, 120.43,

119.91, 119.72, 108.05.HRMS m/z (ESI) found 382.1164 (M þ H)^þ,

C

₂₁H₁₅F₃N₃O^þcalcd for 382.1162, retention time 3.45 min, >99%

solid (92 mg, 57%). ¹H NMR (400 MHz, Methanol-d₄)

d 8.08 (s, 1H),

pure.

7.96 (s, 1H), 7.90 (s, 1H), 7.86 (d, J ¼ 8.4 Hz, 1H), 7.78 (s, 1H), 7.64 (d,

J ¼ 8.2 Hz, 1H), 7.58e7.45 (m, 2H), 7.42 (m, 3H), 7.32 (d, J ¼ 7.8 Hz,

1H). HRMS m/z (ESI) found 397.1273 (M þ H)^þ, C₂₁H₁₆F₃N₄O^þcalcd

for 397.1271, >98% pure.

5.1.4.2. N-(3-(1-methyl-1H-indazol-6-yl)phenyl)-3-(triﬂuoromethyl)

benzamide (4). White solid, Yield: 47%. ¹H NMR (400 MHz, Meth-

anol-d₄)

d

8.33 (s, 1H), 8.27 (d, J ¼ 7.9 Hz, 1H), 8.16 (s, 1H), 8.04 (s,

1H), 7.93 (d, J ¼ 8.4 Hz, 1H), 7.85 (d, J ¼ 8.4 Hz, 1H), 7.81 (s, 1H),

7.73e7.80 (m, 2H), 7.58 (d, J ¼ 7.9 Hz, 1H), 7.49e7.54 (m, 2H), 4.14 (s,

5.1.4.7. 3-(4,4,5,5-Tetramethyl-1,3,2-dioxaborolan-2-yl)-N-(3-(tri-

ﬂuoromethyl)phenyl)benzamide (17).

3H). ¹³C NMR (126 MHz, Methanol-d4)

d

165.71, 141.95, 140.51,

3-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzoic

acid

139.70, 138.77, 135.84, 132.21, 130.94, 130.62 (d, J ¼ 32.6 Hz), 129.25,

128.99, 128.98 (q, J ¼ 3.5 Hz), 124.17 (q, J ¼ 4.1 Hz), 123.97 (d,

J ¼ 271.5 Hz), 123.58, 123.14, 121.02, 120.53, 120.02, 119.95, 106.94,

34.16.HRMS m/z (ESI) found 396.1311 (M þ H)^þ, C₂₂H₁₇F₃N₃O^þcalcd

(300 mg, 1.21 mmol, 16) was dissolved in thionyl chloride (3 mL)

and a catalytic amount of DMF were added, and the mixture was

stirred at 78 ^ꢁC for 3 h. The reaction mixture was concentrated

under reduced pressure to give acid chloride. The acid chloride was

682

Q. Wang et al. / European Journal of Medicinal Chemistry 163 (2019) 671e689

dissolved in THF (5 mL), 3-(triﬂuoromethyl) aniline (195 mg,

1.21 mmol) and DIPEA (0.400 mL, 2.42 mmol) were added, and the

mixture was stirred at room temperature for 5 h. The reaction

mixture was diluted with ethyl acetate and washed with brine. The

organic layer was evaporated under reduced pressure and the ob-

tained crude product was puriﬁed via by silica gel column chro-

matography (eluting with 0e10% EtOAc in petroleum ether) to

afford the product as a white solid (350 mg, 74%). ¹H NMR

this solution before the vial was heated to 80 ^ꢁC overnight. The

reaction was cooled to room temperature, which was then brought

to basic using aqueous sodium bicarbonate solution and extracted

with ethyl acetate. The resulting mixture was concentrated to give

the crude product, which was puriﬁed by silica gel column chro-

matography. Pale yellow solid. Yield: 64%. 1H NMR (400 MHz,

Methanol-d4)

d

8.31 (s, 1H), 8.25 (d, J ¼ 7.9 Hz, 1H), 8.08 (s, 1H), 7.92

(dt, J ¼ 5.1, 2.3 Hz, 2H), 7.85 (d, J ¼ 8.5 Hz, 1H), 7.80e7.72 (m, 2H),

(400 MHz, DMSO‑d₆)

d

10.66 (s, 1H), 8.26 (d, J ¼ 13.3, 1.7 Hz, 2H),

7.58 (s, 1H), 7.47 (dd, J ¼ 8.5, 1.3 Hz, 1H), 7.36 (s, 1H), 2.47 (s, 3H). ¹³

C

8.10 (ddd, J ¼ 8.4, 4.9, 3.3 Hz, 2H), 7.90 (dt, J ¼ 7.3, 1.2 Hz, 1H), 7.59

(dt, J ¼ 11.1, 7.8 Hz, 2H), 7.51e7.41 (m, 1H), 1.33 (s, 12H). MS m/z:

392.3 ([MþH]^þ).

NMR (126 MHz, DMSO‑d₆) d 164.31, 142.22, 140.45, 139.54, 137.11,

136.23, 133.89, 132.26, 131.02, 130.98, 130.19, 129.64 (d, J ¼ 32.0 Hz),

128.56 (d, J ¼ 4.0 Hz), 124.63 (d, J ¼ 4.2 Hz), 124.45 (d, J ¼ 272.5 Hz),

122.48, 122.33, 122.27, 120.66, 119.88, 110.24, 20.16.MS m/z: 396.2

([MþH] ^þ), HRMS m/z (ESI) found 396.1313(M þ H)^þ, C₂₂H₁₇F₃N₃O^þ

calcd for 396.1318, retention time 3.58 min, 100% pure.

5.1.4.8. 3-(1H-indazol-6-yl)-N-(3-(triﬂuoromethyl)phenyl)benza-

mide (8). To

a resealable vial was added K₂CO₃(112 mg,

0.81 mmol), 17 (175 mg, 0.45 mmol), 6-bromo-1H-indazole

(80.0 mg, 0.41 mmol). The vial was sealed and evacuated and

purged with Ar for 5 min before addition of PdCl₂(dppf)-CH₂Cl₂