Discovery of 4-(((4-(5-chloro-2-(((1s,4s)-4-((2-methoxyethyl)amino)cyclohexyl)amino)pyridin-4-yl)thiazol-2-yl)amino)methyl)tetrahydro-2H-pyran-4-carbonitrile (JSH-150) as a novel highly selective and potent CDK9 kinase inhibitor

Article abstract of DOI:10.1016/j.ejmech.2018.09.025

Through a structure-guided rational drug design approach, we have discovered a highly selective inhibitor compound 40 (JSH-150), which exhibited an IC₅₀ of 1 nM against CDK9 kinase in the biochemical assay and achieved around 300–10000-fold selectivity over other CDK kinase family members. In addition, it also displayed high selectivity over other 468 kinases/mutants (KINOMEscan S score(1) = 0.01). Compound 40 displayed potent antiproliferative effects against melanoma, neuroblastoma, hepatoma, colon cancer, lung cancer as well as leukemia cell lines. It could dose-dependently inhibit the phosphorylation of RNA Pol II, suppress the expression of MCL-1 and c-Myc, arrest the cell cycle and induce the apoptosis in the leukemia cells. In the MV4-11 cell-inoculated xenograft mouse model, 10 mg/kg dosage of 40 could almost completely suppress the tumor progression. The high selectivity and good in vivo PK/PD profile suggested that 40 would be a good pharmacological tool to study CDK9-mediated physiology and pathology as well as a potential drug candidate for leukemia and other cancers.

Full text of DOI:10.1016/j.ejmech.2018.09.025

Accepted Manuscript
Discovery of 4-(((4-(5-chloro-2-(((1s,4s)-4-((2-
methoxyethyl)amino)cyclohexyl)amino)pyridin-4-yl)thiazol-2-
yl)amino)methyl)tetrahydro-2H-pyran-4-carbonitrile (JSH-150) as a novel highly
selective and potent CDK9 kinase inhibitor
Beilei Wang, Jiaxin Wu, Yun Wu, Cheng Chen, Fengming Zou, Aoli Wang, Hong Wu,
Zhenquan Hu, Zongru Jiang, Qingwang Liu, Wei Wang, Yicong Zhang, Feiyang Liu,
Ming Zhao, Jie Hu, Tao Huang, Juan Ge, Li Wang, Tao Ren, Yuxin Wang, Jing Liu,
Qingsong Liu
PII:
S0223-5234(18)30797-9
10.1016/j.ejmech.2018.09.025
EJMECH 10736
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 30 March 2018
Revised Date: 6 September 2018
Accepted Date: 9 September 2018
Please cite this article as: B. Wang, J. Wu, Y. Wu, C. Chen, F. Zou, A. Wang, H. Wu, Z. Hu, Z. Jiang,
Q. Liu, W. Wang, Y. Zhang, F. Liu, M. Zhao, J. Hu, T. Huang, J. Ge, L. Wang, T. Ren, Y. Wang, J. Liu,
Q. Liu, Discovery of 4-(((4-(5-chloro-2-(((1s,4s)-4-((2-methoxyethyl)amino)cyclohexyl)amino)pyridin-4-
yl)thiazol-2-yl)amino)methyl)tetrahydro-2H-pyran-4-carbonitrile (JSH-150) as a novel highly selective
and potent CDK9 kinase inhibitor, European Journal of Medicinal Chemistry (2018), doi: https://
doi.org/10.1016/j.ejmech.2018.09.025.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Discovery
of
4
-(((4-(5-chloro-2-(((1s,4s)-4-((2-methoxyethyl)amino)cyclohexyl)amino)
pyridin-4-yl)thiazol-2-yl)amino)methyl)tetrahydro-2H-pyran-4-carbonit
rile (JSH-150) as a Novel Highly Selective and Potent CDK9 Kinase
Inhibitor
1
,2,3,7
1,2,3,7
1,3,7
1,2,3,7
1,3,7
Beilei Wang
, Jiaxin Wu
, Yun Wu , Cheng Chen
, Fengming Zou , Aoli
1
,3
1,3
1,3
1,2,3
4
1,3
Wang , Hong Wu , Zhenquan Hu , Zongru Jiang , Qingwang Liu , Wei Wang ,
1
,2,3
1,3
4,5
4,5
4,5
4,5
Yicong Zhang , Feiyang Liu , Ming Zhao , Jie Hu , Tao Huang , Juan Ge , Li
1
,2,3
4,5
5
1,3,4,5*
1,2,3,4,5,6*
Wang , Tao Ren , Yuxin Wang , Jing Liu
, Qingsong Liu
1
. High Magnetic Field Laboratory, Chinese Academy of Sciences, Mailbox 1110, 350
Shushanhu Road, Hefei, Anhui 230031, P. R. China
2
3
. University of Science and Technology of China, Hefei, Anhui 230036, P. R. China
. Key Laboratory of High Magnetic Field and Ion Beam Physical Biology, Hefei
Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui 230031,
P. R China
4
. Precision Targeted Therapy Discovery Center, Institute of Technology Innovation,
Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui
2
30088, P. R. China
5
6
. Precision Medicine Research Laboratory of Anhui Province, Hefei, Anhui 230088, P.
R. China
. Institute of Physical Science and Information Technology, Anhui University, Hefei,
Anhui 230601, P. R. China

ACCEPTED MANUSCRIPT
7
. These authors contribute equally
AUTHOR INFORMATION
Corresponding Authors
*
E-mail address: jingliu@hmfl.ac.cn (J.Liu), E-mail address: qsliu97@hmfl.ac.cn (Q.
Liu).
ABSTRACT
Through a structure-guided rational drug design approach, we have discovered a highly
selective inhibitor compound 40 (JSH-150), which exhibited an IC50 of 1 nM against
CDK9 kinase in the biochemical assay and achieved around 300-10000-fold selectivity
over other CDK kinase family members. In addition, it also displayed high selectivity
over other 468 kinases/mutants (KINOMEscan S score(1)=0.01). Compound 40
displayed potent antiproliferative effects against melanoma, neuroblastoma, hepatoma,
colon cancer, lung cancer as well as leukemia cell lines. It could dose-dependently inhibit
the phosphorylation of RNA Pol II, suppress the expression of MCL-1 and C-Myc, arrest
the cell cycle and induce the apoptosis in the leukemia cells. In the MV4-11
cell-inoculated xenograft mouse model, 10 mg/kg dosage of 40 could almost completely
suppress the tumor progression. The high selectivity and good in vivo PK/PD profile
suggested that 40 would be a good pharmacological tool to study CDK9-mediated
physiology and pathology as well as a potential drug candidate for leukemia and other
cancers.
Keywords
CDK9, Kinase inhibitor, Structure-activity relationship, Leukemia

ACCEPTED MANUSCRIPT
Abbreviations used
AML, acute myeloid leukemia; CDK, cyclin-dependent kinase; CLL, chronic
lymphocytic leukemia; MCL, mantle cell lymphoma; SAR, structure-activity relationship;
DMSO, dimethyl sulfoxide; DMF, N, N-dimethylformamide; DCM, dimethyl chloride;
NBS, N-bromobutanimide; DMAP, 4-dimethylaminopyridine; LDA, lithium
diisopropylamide; DIPEA, N, N-diisopropylethylamine; DME, 1,2-dimethoxyethane;
DIAD, diisopropyl azodicarboxylate.
1
. Introduction
Cyclin-dependent kinases (CDKs) are a class of serine/threonine protein kinase which
play critical roles in the regulation of cell cycle or/and transcription [1,2]. To date, 21
different CDKs (1−11a, 11b−20) have been identified in the human genome and they can
be divided into two main groups based on their primary roles [3]. CDK1, CDK2, CDK4,
and CDK6 have been found to regulate the cell cycle progression upon binding to cyclin
proteins. CDK7, CDK8, CDK9 and CDK11 are key players in transcription regulation
[4,5]. These CDKs promote RNA synthesis for cell growth, differentiation and viral
pathogenesis [6,7] etc. Among them, CDK9 is a key regulator of transcription elongation
in eukaryotic cells and has been considered as a potential drug target for several diseases
including cardiac hypertrophy and certain cancers such as a variety of blood cancers as
well as solid tumors [8]. CDK9 phosphorylates Ser2 residues in the carboxy-terminal
domain (CTD) of RNA Pol II to initiate the transcription elongation [9]. Activation of
CDK9 kinase will promote the expression of antiapoptotic factor myeloid cell leukemia 1
(MCL-1) [10], which leads to malignant cell transformation. Inhibition of CDK9 activity
will result in the down-regulation of this short-lived antiapoptotic protein MCL-1, and

ACCEPTED MANUSCRIPT
consequently the induction of caspase-dependent apoptosis which is important for tumor
control [11-14].
Flavopiridol (alvocidib, 1), the first CDK inhibitor entered into clinical trials, exhibited
antileukemic activity in CLL patients via potent inhibition of the CDK9-meditated
down-regulation of antiapoptotic proteins transcription [15,16]. This has stimulated broad
research interests for CDK9 and a number of small molecule inhibitors have been
discovered and moved into clinical investigation (Figure 1). However, the high structural
homology of ATP binding pocket among CDKs made it a great challenge for developing
highly selective CDK9 kinase inhibitors. Most of the currently known CDK9 inhibitors
are non-selective inhibitors. For example, 1 exhibited high potency against CDK1, 2, 4, 5,
6
and 9; 2 (SNS-032) was potent to CDK 2, 7 and 9 [17,18]; compound 3 (RGB286638)
targeted CDK1, 2, 3, 4, 5 and 9. Compound 4 (LY2857785), which was evaluated in
phase I clinical trials for the treatment of solid tumors, was relatively more selective but
also simultaneously inhibited CDK7, 8, 9 [19,20]. Other multiple CDK inhibitors
including 5 [21] (AZD5438, CDK1/2/9), 6 [22] (AT7519, CDK1/2/4/6/9) and 7 [23-25]
(SCH727965, CDK1/2/5/9) have also been tested in clinical trials. The pyrimidine core
scaffold-based CDK9 inhibitors including 8 (CDKI-73) [13] and 9 also potently inhibited
both CDK7 and CDK9. However, given the fact that most of the CDK family members
play critical roles both in physiology and pathology conditions, e.g., CDK7 is directly
upstream of CDK9 and both of them control transcription, non-specific inhibition of
these kinases would increase the potential of the adverse events in the clinic and also
made it hard to dissect the specific roles in the cellular context. Therefore, a highly
selective CDK9 inhibitor is urgently desired for both of the physiological and

ACCEPTED MANUSCRIPT
pathological study. Here we reported our medicinal chemistry efforts starting from 9’s
core scaffold which led to the discovery of a highly selective and potent CDK9 inhibitor
4
0 (JSH-150). It is worth mentioning that during preparation of this manuscript, a
structurally similar inhibitor 10 (NVP-2), which also exhibited high selectivity and
potency against CDK9 kinase, was reported as a chemical biology research tool for
studying protein degradation [26].
Figure 1. Chemical structures of representative inhibitors that bear CDK9 kinase activity.
2
. Results and discussion
2
.1. Design rationale
Compound 9 is a potent inhibitor of CDK9 with an IC50 of 14 nM, but also displays
an IC value of 92 nM for CDK7 [27]. In order to compare the binding modes of 9 in
5
0
CDK7 and CDK9, we docked 9 into the CDK7 kinase (PDB ID: 1UA2) [7] (Figure

ACCEPTED MANUSCRIPT
2
A-B). The results showed that the aminopyrimidine formed two hydrogen bonds in the
hinge binding area with Met94 and Cys106 of CDK7 and CDK9 respectively. The
diazepane moiety of 9 formed a hydrogen bond with Leu143 and Thr29 of CDK7 and
CDK9 respectively. In addition, the NH of amino-thiazole moiety of 9 formed hydrogen
bonds with Lys41 and Thr29 of CDK7 and CDK9 respectively. After carefully analyzing
the structures of CDK7 and CDK9, we postulated that the P-loop of CDK9 was more
flexible and could accommodate bigger fragments at this position than CDK7. In
addition, there is a shallow hydrophobic pocket adjacent to the hinge binding area formed
by His108 and Gly112 of CDK9. While for CDK7, these two residues were replaced by
Thr96 and Val100, which indicated that a carefully selected small hydrophobic moiety
could be used here to explore the selectivity between CDK7 and CDK9 (Figure 2C-D).
Furthermore, the structural information showed that the gatekeeper residue in both CDK7
(Phe91) and CDK9 (Phe103) interacted with the CN group of 9, which might be replaced
with the Cl atom in order to achieve higher binding affinity due to the halogen bond
interaction [28]. Based on these analyses, we decided to start from 9’s core scaffold and
1
2
3
modify the R /R /R , X and Y moieties, which finally led to the discovery of the highly
potent and selective CDK9 kinase inhibitor 40 (JSH-150) (Figure 3).

ACCEPTED MANUSCRIPT
Figure 2. Binding mode of compound 9 with CDK7/9 kinase. (A) Docking of 9 in CDK7
kinase (PDB ID: 1UA2). (B) X-ray structure of 9 in complex with CDK9 kinase (PDB
ID: 4BCG). (C) Illustration of the P-loop area difference by overlaying CDK7 and CDK9.
(D) Illustration of the sequence difference of CDK7 and CDK9 adjacent to the hinge
binding area.
SH
O
SH
O
Cys106
N
H
Hinge binding
Cys106
N
H
HN
HN
H
R2
CN
Cl
N
N
X
H
S
N
N
N
N
S
NH
NH
NC
N
O
Y
N
R3
NH
9
40
HN
MeO
R1
Figure 3. Schematic illustration of discovery of compound 40 from compound 9.
2
.2. Chemistry and structure-activity relationship (SAR) investigation

ACCEPTED MANUSCRIPT
As shown in scheme 1, 5-bromothiazol-2-amine (46) was protected with Boc group
47) and intermediate 48 was prepared via Bromine dance reaction. Suzuki coupling
(
between 48 and pinacol boronic ester 45a or 45b offered intermediates 11 or 11a, which
then reacted with appropriate alcohols to afford intermediates 53g-m via Mitsunobu
reaction. Intermediates 56a-b were prepared from corresponding alcohols (54a-b) via
tosylation reaction (55a-b) and nucleophilic substitution, while 56c-e were synthesized
from tert-butyl ((1r,4r)-4-aminocyclohexyl)carbamate after nucleophilic substitution with
alkyl bromide and Boc deprotection. The target compounds 12-15, 21-26, 30, 32, 35,
3
7-38 and 43 were obtained by nucleophilic substitution of 53g-m with 56a-e.
Nucleophilic substitution of 53l with appropriate Boc-protected amine followed by
deprotection of the Boc group furnished the target compounds 16-17, and 18-20 were
obtained by acylation reaction of 16 with corresponding acyl chloride (Scheme 1).
a
Scheme 1. Synthesis of Compounds 11-26, 30, 32, 35, 37-38 and 43

ACCEPTED MANUSCRIPT
a
Reagents and conditions: (a) pinacol, toluene, reflux, overnight; (b) Boc
2
O, DMAP, THF,
rt, 2 days; (c) LDA, 0 °C, 2 h; (d) Pd(PPh
3
)
4
2 3 2
, Na CO , DME/H O/dioxane, 70 °C,
3
1
1
overnight; (e) R CH OH, PPh , DIAD, THF, N , 0 °C, overnight; (f) R NH or R OH,
2
3
2
2
DIEA, DMSO, 100-110 °C, 2 days; (g) NaH, TosCl, 0 °C-rt, overnight; (h)
6
trans-1,4-diaminocyclohexane, CH CN, 90 °C, overnight; (i) (i) R -Br, K CO , CH CN,
3
2
3
3
1
8
1
0 °C, 16 h; (ii) 4 N HCl, ethyl acetate, rt, 3 h; (j) (i) R NHBoc, DIEA, DMSO,
7
00-110 °C, 2 days; (ii) CF COOH, DCM, 0 °C-rt, 1.5 h; (k) R COCl, DCM, 0 °C, 1 h.
3

ACCEPTED MANUSCRIPT
The syntheses of compounds 28-29 were depicted in Scheme 2. Intermediate 60 was
synthesized from (tetrahydro-2H-pyran-4-yl)methanol (57) in three steps including
bromination (58), nucleophilic substitution (59) and reduction. Then a nucleophilic
substitution reaction between 57/60 with 2, 4-dibromothiazole provided intermediates
6
1-62. Suzuki coupling between 61/62 and 45a afforded 63-64, which was subsequently
subjected to nucleophilic substitution with 56c to yield the target compounds 28-29.
a
Scheme 2. Synthesis of Compounds 28-29
a
Reagents and conditions: (a) NBS, PPh , DCM, 0 °C, 1-2 h; (b) KSAc, DMF, 90 °C, 2 h;
3
(
c) LiAlH , THF, 0 °C-rt, overnight; (d) 2,4-dibromothiazole, NaH, THF, 0 °C-rt,
4
overnight; (e) 45a, Pd(PPh ) , Na CO , 1,4-dioxane/H O, 90 °C, overnight; (f) 56c,
3
4
2
3
2
K CO , DMSO, 100 °C, 2 days.
2
3
Starting from Mitsunobu reaction between 48 and a variety of alcohols, which
afforded intermediates 52b-f, compounds 31, 33-34, 36 and 39 were obtained from
nucleophilic substitution of 56c and 53b-f which were provided by Suzuki coupling
between 52b-f and 45a. For the syntheses of 40-42, intermediate 51 was obtained through
a two-step process including cyclization (50) and reduction. Then Mitsunobu reaction
between 51 and 48 offered intermediate 52a, which was subjected to a Suzuki coupling

ACCEPTED MANUSCRIPT
with 45a to yield 53a. Finally, a base-mediated substitution reaction between 56a-c and
5
3a furnished the desired compounds 40-42 (Scheme 3).
a
Scheme 3. Synthesis of Compounds 31, 33-34, 36 and 39-42
a
3
Reagents and conditions: (a) R CH OH, PPh , DIAD, anhydrous THF, 0-40 °C, N , 1.5
2
3
2
h; (b) 45a, Pd(PPh ) , Na CO , DME/H O/dioxane, N , 70 °C, overnight; (c) 56c, DIEA,
3
4
2
3
2
2
DMSO, 100-110 °C, 2 days; (d) 1-bromo-2-(2-bromoethoxy)ethane, DBU, DMF, 85 °C,
3
h; (e) NaBH
, 1.5 h; (g) 56a-c, DIEA, DMSO, 100-110 °C, 2 days.
Intermediate thiazol boronic acid 66 was obtained from 65 which was synthesized
4 3
, DME/MeOH, 0 °C, 2 h; (f) 48, PPh , DIAD, anhydrous THF, 0-40 °C,
N
2
through Mitsunobu reaction between 48 and (tetrahydro-2H-pyran-4-yl)methanol.
Following Suzuki coupling and Buchwald coupling reactions provided intermediate 68.
After removal of the N-Boc protecting group, final compound 27 was afforded through a
nucleophilic substitution of 69 with 1-bromo-2-methoxyethane.
a
Scheme 4. Synthesis of Compound 27

ACCEPTED MANUSCRIPT
a
Reagents and conditions: (a) (tetrahydro-2H-pyran-4-yl)methanol, PPh , DIAD, THF,
3
-
10-50 °C, 3 h; (b) bis(pinacolato)diboron, Pd(dppf)Cl ·DCM, KOAc, dioxane, 100 °C, 3
2
h; (c) 4-bromo-2-chloro-5-methylpyridine, Pd(PPh
tert-butyl ((1r,4r)-4-aminocyclohexyl)carbamate, Pd
20 °C, overnight; (e) TFA, DCM, rt, 1 h; (f) BrCH
overnight.
3
)
4
, K
(dba)
CH OCH
3
PO
, Binap, tBuONa, toluene,
, K CO , DMF, 100 °C,
4
, dioxane, 100 °C, 4 h; (d)
2
3
1
2
2
3
2
3
In order to better explore the SAR of the inhibitors, besides the biochemical assay
with CDK9 protein (IC50), we also evaluated the inhibitor’s cellular antiproliferative
activity (GI50) using MV4-11 (acute myeloid leukemia, AML) cell and MEC-1 (chronic
lymphocytic leukemia, CLL) cell. In addition, we used normal Chinese hamster lung
(CHL) cell to monitor the inhibitor’s general cytotoxicity as well [29] (Table 1). We
2
first prepared the pyridinthiazole compound 11, which bears a Cl atom at R , a F atom at
Y and a slightly bigger Boc group instead of methyl group attached to NH at X position.
Unfortunately, it completely lost the activity against CDK9 kinase. Replacement of the
NHBoc at X with NH-methylenetetrahydropyrane moiety (12) started to gain back the
activity to CDK9 (IC50: 0.557
µM) but still did not show any cellular activity.
1
Substitution of the F atom with 2-methoxyethanamine (Y = NH, R = MeOCH CH )
2
2
afforded 13, which only showed weak activity against CDK9 and lacked the cellular

ACCEPTED MANUSCRIPT
1
activity. Installation of the phenyl group at R (14) slightly increased the activity to
1
CDK9 (IC : 0.238
µ
M) but also displayed no cellular activity. When R was the 4-fluoro
5
0
benzoyl group (15), it resulted in significant activity loss to CDK9 and started to show
general cytotoxicity. These results indicated that a more flexible group might be required
1
1
at the R position. We then tried to change the substituent group at R to 4-amino
cyclohexyl (16), which surprisingly showed good activity against CDK9 kinase (IC50
.022 M) and started to exhibit cellular activities in MV4-11 (GI : 0.14 M) and
MEC-1 (GI50: 0.27 M) cells. Meanwhile, it displayed good selectivity over normal CHL
cells (GI : 3.5 M). Addition of a methylene group between NH and 4-amino cyclohexyl
17) slightly decreased the activity to CDK9 (IC50: 0.084 M) and significantly lost the
activity in the cellular assays. A series of acyl substitutions of the 4-amino cyclohexyl
18-20) were also examined but all led to activity loss to CDK9 kinase. However,
aliphatic chain substitution at the same position such as methoxylethane (21) resulted in a
potent activity against CDK9 (IC50: 0.002 M), although it also displayed weak cellular
:
0
µ
µ
5
0
µ
µ
5
0
(
µ
(
µ
antiproliferative activity to cancer cell lines and cytotoxicity to CHL cells. Both of the
1
methoxypropane (22) and ethoxyethane (23) substitution of the 4-amino cyclohexyl at R
caused slight activity loss to CDK9 kinase as well as the cellular activities. Interestingly,
1
installation of a chiral center by introduction of (S)-1-methoxypropan-2-amine at R
position (24) completely lost the activity to CDK9 (IC50: >10
corresponding enantiomer (R)-1-methoxypropan-2-amine (25) started to gain back the
activity (IC : 0.733 M) although it was still much weaker than 21. Replacement of the
Cl in 21 with H atom (26) and methyl group (27) resulted in more than 50-fold and
26-fold activity loss to CDK9 indicating that the Cl atom might indeed form a halogen
µM), while the
µ
5
0
1

ACCEPTED MANUSCRIPT
bond interaction with Phe103 [28]. Change of the NH at X position of 21 to an O atom
(
28) or S atom (29) both led to activity loss (14 and 7-fold respectively) against CDK9
kinase.
Table 1. SAR Exploration Focused on the R /R /X/Y Moieties
1
2
a
1
2
CDK9
MV4-11
50
MEC-1
50
CHL
50
Compd.
R
R
X
Y
F
(
IC : ꢀM) (GI : ꢀM) (GI : ꢀM) (GI : ꢀM)
50
1
1
1
1
1
1
1
1
2
3
4
5
6
7
-
-
Cl
Cl
Cl
Cl
Cl
Cl
Cl
NHBoc
>10
>10
>10
>10
0
.557±
F
>10
>10
>10
>10
>10
>10
>10
>10
>10
0.012
1
.04±
NH
NH
NH
NH
NH
0.012
0
.238±
0.042
0
.55±0.00 3.6±0.023 0.93±0.02
2.3±
4
4
0.233
0
0
.022±
.007
0.14±
0.27±
3.5±
0
0.033
0.012
0.027
.084±
.008
1.9±
3.9±
>10
0
0.016
0.025

ACCEPTED MANUSCRIPT
0
.101±
.016
3.6±
1
1
2
2
2
2
2
2
2
2
2
2
8
9
0
1
2
3
4
5
6
7
8
9
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
>10
>10
>10
>10
>10
>10
0
0.042
0
.15±
>10
>10
0
.013
1
.89±
0
.808
0
0
0
.002±
0.17±
1.24±
0.534±
0
.0002
0.045
0.089
0.087
.012±
1.09±
2.61±
2.03±
0
.008
0.096
0.214
0.145
.051±
.005
0.497±
0.634±
2.82±
0
0.014
0.103
0.241
0
.44±
.021
>10
1.3±0.015 3.6±0.032
0
0
0
.733±
0.14±
.031
0.31±0.00
1.1±0.018
4
0
.503
0
.114±
.014
0.91±
.003
2.5±0.014 1.2±0.024
3.56±0.14 5.01±0.11
1.4±0.023 3.3±0.028
0.8±0.002 5.6±0.045
0
0
0
.253±0.0
1.62±
0.098
CH
Cl
Cl
3
2
4
0
.028±
.006
0.57±
0
0
.014
0
.014±
.002
0.34±
.025
0
0
a
All GI values were obtained by triplet testing.
5
0

ACCEPTED MANUSCRIPT
Because 21 exhibited potent inhibitory activity to CDK9 kinase in the biochemical
2
assay but still lacked cell antiproliferative efficacy, we then fixed the R group as Cl atom
1
3
and the X moiety as NH and further explored the R /Y/R moieties (Table 2).
3
Replacement of tetrahydropyran in 21 with H atom (30) at R decreased the activity to
CDK9 about 50-fold indicating that a relatively large size group here might be favorable
3
for the activity. A more hydrophobic CF group at R position (31) gained back the
3
activity against CDK9 to 0.017
µM (IC50), although it still lacked potent antiproliferative
3
activity in the cellular assays. Variation at the R site with different hydrophobic rings
(32-38) all retained reasonable activities to CDK9 kinase but none of them exhibited
3
potent cellular activity. When R was tetrahydrofuran group (39), it gained the activity to
CDK9 with an IC50 value of 0.09
µM and showed good selectivity window between
normal cells and cancer cells, although it was still not very potent in the cellular assays.
3
Interestingly, replacement of the R moiety with tetrahydropyran carbonitrile (40)
resulted in great activity against CDK9 kinase (IC : 0.002
µM). In addition, it also
5
0
exhibited potent antiproliferative efficacies in MV4-11 (GI : 0.012 M) and MEC-1
µ
5
0
(
(
(
GI50: 0.037
µ
M) cells, and meanwhile kept good selectivity over normal CHL cells
1
GI : 1.2
µ
M). Addition of a chiral center to the R moiety of 40 by installation of
5
0
R)-1-methoxypropan-2-amine (41) almost retained both of the biochemical and cellular
activities as well as the selectivity window. However, the corresponding S enantiomer
42) started to lose the cellular activities. Finally, changing NH at Y position of 21 to
Oxygen atom (43) led to significant activity loss to CDK9 kinase.
(
1
3
a
Table 2. SAR Exploration Focused on the R /Y/R Moieties

ACCEPTED MANUSCRIPT
1
3
CDK9
MV4-11
50
MEC-1
50
CHL
50
Compd.
R
R
Y
(
IC : ꢀM) (GI : ꢀM) (GI : ꢀM) (GI : ꢀM)
50
0
.101±0.0 0.53±0.01
3
3
3
3
3
3
3
3
3
3
0
1
2
3
4
5
6
7
8
9
H
NH
1.4±0.023 1.9±0.035
1
6
2
0
0
0
0
0
0
0
0
0
.017±0.0 0.752±0.0
21
0.877±0.0
CF3-
NH
NH
NH
NH
NH
NH
NH
NH
NH
1.1±0.032
12
0
7
.096±0.0 0.74±0.01 0.89±0.03
1.4±0.021
4
4
6
3
.107±0.0
4
2.26±0.10 1.29±0.08
>10
1
5
.080±0.0 2.16±0.11 2.94±0.08
>10
0
9
4
9
.027±0.0 0.82±0.03 0.85±0.02
8±0.214
0
3
4
3
.094±0.0 3.46±0.14 6.31±0.12 2.04±0.09
0
4
5
1
8
.014±0.0 0.33±0.00 0.47±0.02
1±0.017
0
2
6
5
.002±0.0 0.88±0.05 0.45±0.03
01
1.1±0.041
>10
0
3
2
.09±0.00 0.493±0.0 0.883±0.0
46 54
7

ACCEPTED MANUSCRIPT
0
0
0
1
.006±0.0 0.012±0.0 0.037±0.0
002 01
4
4
4
0
1
2
3
NH
NH
NH
O
1.2±0.067
1.1±0.017
0
1
.009±0.0 0.014±0.0 0.047±0.0
02 024 036
0
.01±0.00 0.121±0.0 0.537±0.0 1.25±0.05
2
41
25
7
.67±0.36
3
4
7.7±0.051
>10
>10
a
All GI values were obtained by triplet testing.
5
0
2
.3. Selectivity profiling of compound 40
Given the fact that 40 exhibited the best in vitro biochemical activity against CDK9
kinase, antiproliferative efficacy in cancer cell lines as well as the therapeutic window
between normal cells and cancer cells, we then chose this compound for further
selectivity evaluation. We first subjected 40 to Invitrogen’s SelectScreen assay to
examine its selectivity among structurally similar CDK kinase isoforms (Figure 4A-B).
Among 10 different CDK isoforms which are available from the Invitrogen’s platform,
4
0 only potently inhibited CDK9 kinase (IC : 0.001
ꢀM) and exhibited more than
5
0
1
000-fold selectivity over CDK1/2/3/5/7/8/11/14 and about 300-fold over CDK16. In
order to investigate the inhibitor’s selectivity among other protein kinases, we then tested
0 with DiscoverX’s KINOMEscan technology for the kinomewide selectivity profiling
30]. Among 468 kinases/mutants, 40 exhibited a S score(1) of 0.01 at the concentration
of 1 M. Besides CDK9 kinase, it only showed strong binding to DYRK1A and
DYRK1B kinases (Figure 4C-D, Supplemental table 1).
4
[
µ

ACCEPTED MANUSCRIPT
Figure 4. Selectivity profiling of compound 40. (A, B) Biochemical testing of 40 against
CDK isoforms with Invitrogen SelectScreen platform (Carlsbad, CA, USA). All data
were obtained by triplet testing. (C) Treespot demonstration of kinomewide selectivity of
4
0 with DiscoverX KINOMEscan technology. Measurements were performed at a
concentration of 1 ꢀM of the inhibitor. The affinity was defined with respect to a DMSO
control. (D) Targets that demonstrated strong binding to 40 in kinome profiling with a
percent control number less than 1.
In order to better understand 40’s binding mechanism, we next docked the inhibitor
into the X-ray structure of CDK9 kinase (PDB ID: 4BCG) (Figure 5A). The modeling

ACCEPTED MANUSCRIPT
results showed that the aminopyridine of 40 formed two hydrogen bonds with Cys106 in
1
the hinge binding region. In the R moiety, the oxygen atom formed a hydrogen bond
with Asp109 located adjacent to that area. In addition, the NH at X position formed a
hydrogen bond with Asp167 in the activation loop. Meanwhile, the CN group formed a
hydrogen bond with the Thr29 in the P-loop. In comparison, it was disfavored when this
binding mode was superimposed onto CDK7 kinase (PDB ID: 1UA2) (Figure 5B). First,
Val100 adjacent to the hinge binding region of CDK7 prevented the methoxy group of
1
the R moiety from directing into this position due to the steric hindrance, while in CDK9
kinase the residue Gly122 at the same position lacks the isopropyl side chain of Val100
and thus avoid this interaction. In addition, the Thr29 in the CDK9 P-loop was replaced
with Gly21 in the CDK7 P-loop, which lacks the capability to form hydrogen bond with
the CN group of 40 in CDK7.
Figure 5. Binding mode of compound 40 with CDK9 kinase and superposition mode of
CDK7 with CDK9 kinase. (A) Carton mode illustration of docking of 40 into CDK9
(PDB ID: 4BCG). (B) Illustration of the differences by overlaying CDK7 (PDB ID:
1
UA2) and CDK9.
2
.4. Cellular evaluation of compound 40

ACCEPTED MANUSCRIPT
We then examined the antiproliferative effect of 40 on a panel of established cancer
cell lines (Table 3). Not surprisingly, it exhibited potent antiproliferative activities in
solid tumor cell lines such as A375 (melanoma), A431 (squamous), BE(2)M17
(
neuroblastoma), GIST-T1 (GIST) and COLO205 (colon cancer) with GI values from
50
0
.002 to 0.044 µM. In the leukemia cell lines including AML, CLL and B cell lymphoma
cell lines, 40 also displayed strong growth inhibition efficacies with GI50s ranging from
single to double digit nM. In addition, 40 was much less sensitive in normal CHO cells
(
GI : 1.1
µ
M) compared with the cancer cell lines. Compound 7, which is a CDK1/2/5/9
multiple kinase inhibitor, also exhibited similar trends in these cancer cell lines except
that it also showed relatively potent antiproliferative effect on CHO cells (GI : 0.16 M)
5
0
µ
5
0
which might reflect its multiple-target property.
Table 3. Antiproliferation Effects of Compound 40 on a Panel of Cancer Cell Lines
a
Compd. 40
GI : M)
Compd. 7
(GI : M)
Cell line
Cell type
(
ꢀ
ꢀ
50
50
A375
A431
melanoma
squamous
neuroblastoma
GIST
0.044±0.0023
0.019±0.0022
0.034±0.0062
0.0021±0.0003
0.014±0.0032
0.013±0.0015
0.056±0.0024
0.012±0.0011
0.0011±0.0002
0.0066±0.0007
0.011±0.0014
0.011±0.0011
0.021±0.0013
BE(2)M17
GIST-T1
COLO205
Ramos
0.0088±0.0006
0.0068±0.0005
0.0086±0.0007
0.008±0.0003
0.0033±0.0004
0.0045±0.0004
0.013±0.0021
colon cancer
B cell lymphoma
AML
HL-60
MOLM13
MOLM14
OCI-AML-3
AML
AML
AML

ACCEPTED MANUSCRIPT
SKM-1
U937
AML
0.012±0.0018
0.011±0.0032
0.01±0.0018
0.0036±0.0006
0.011±0.0029
0.16±0.013
AML
0.002±0.0003
0.011±0.0034
0.037±0.0028
1.1±0.021
MEC-1
MEC-2
CHO
CLL
CLL
Chinese hamster ovary cell
a
All GI values were obtained by triple testing.
5
0
Since CDK9 kinase inhibitors have been extensively evaluated for leukemia in the
clinical trials, we chose MV4-11 (AML), HL-60 (AML) and MEC-1(CLL) cell lines to
examine 40’s effects on the CDK9-mediated signaling pathways (Figure 6A). As
expected, upon 2 h drug treatment, phospho-CDK9 Thr186 was not affected however
phospho-RNA Pol II Ser2, which is the direct phosphorylation site in the CTD domain,
was dose-dependently inhibited (EC50: <100 nM) in all three cell lines. In addition,
phospho-RNA pol II Ser5, which is the phosphorylation site of CDK7 was not affected
even at 3 µM concentration of 40, which again reflected its selectivity over CDK7 kinase.
The expression level of antiapoptotic proteins such MCL-1 and c-MYC were
dose-dependently reduced while other antiapoptotic proteins such as XIAP and BCL-2
were not affected upon 2 h treatment of 40 [31]. Compound 7 exhibited a similar trend.
After 24 h drug treatment, by looking at the cleaved PARP and caspase-3 protein level,
dose-dependent apoptosis was observed in all three cell lines which were in accordance
with the observation of the antiapoptotic protein level reduction for MCL-1 and c-MYC
(
Figure 6B). Furthermore, the cell cycle was arrested at G0/G1 phase after 24 h treatment
of 40, indicating that the cells stopped growing after transcription was blocked (Figure
C).
6

ACCEPTED MANUSCRIPT
Figure 6. Cellular effects of compound 40 in AML cell lines MV4-11, HL-60 and CLL
cell line MEC-1. (A) Effects of 40 on CDK9-mediated signaling pathways. (B) Effects of
4
0 on apoptosis. (C) Effects of 40 on cell cycle progression.
.5. In vivo PK/PD evaluation of compound 40
2
The PK properties of 40 were evaluated in different species including mice,
Sprague-Dawley rats and beagle dogs through intravenous injection and oral
administration (Table 5). The data showed that 40 was absorbed rapidly in dogs and mice

ACCEPTED MANUSCRIPT
(T
max = 1.33 h and 2.00 h respectively) but slowly in rats (Tmax = 3.33 h). 40 also
displayed different half-lives in three different species via oral administration (T1/2 = 1.55
h in mice, 3.37 h in rats and 20.37 h in dogs), which indicated that it was metabolized
very slowly in dogs compared with mice and rats. The Vz values of 40 in mice, rats and
dogs were 18203.84, 44464.14 and 12783.38 mL/kg respectively, indicating that it
distributed widely in rats’ body while mainly in blood and extracellular fluid for mice and
dogs. In addition, 40 exhibited acceptable bioavailability in mice, rats and dogs (F =
4
5.01%, 45.10% and 39.15%, respectively). The PK properties indicated that 40 would
be suitable for oral administration.
Table 5. Pharmacokinetic Study of Compound 40 in Mice, Sprague Dawley Rats and
a
Beagle Dogs
Mice
n = 24)
Rats
(n = 3)
Dogs
(n = 3)
Species
(
iv
3 mg/kg)
po
iv
po
iv
(2 mg/kg)
po
4
0
(
(10 mg/kg)
(3 mg/kg)
(20 mg/kg)
(10 mg/kg)
T1/2
1
.72
1.55
2.00
6.80±0.47
0.02±0.00
3.37±0.45
13.14±2.63 20.37±17.09
(
h)
Tmax
0
.083
3.33±1.15
0.02±0.00
1.33±0.58
(
h)
C
max
1468.16±1
60.36
414.47±156 770.55±259.
.82 27
5
3
4
73.36
93.05
08.01
115.76
591.17
612.21
--
329.12±66.11
(
ng/mL)
AUC(0-t)
h*ng/mL)
624.78±38
.42
1780.21±
341.34
2014.82±47 3943.59±18
4.61 19.29
(
(
AUC(0-∞)
h*ng/mL)
662.30±47
.14
1989.38±
373.25
3121.69±12 16735.55±1
21.56
2152.92
Vz
mL/kg)
44464.14±
469.28
12783.38±2
464.08
18203.84
--
--
--
(
Cl
mL/h/kg)
4545.22±3
28.14
711.30±277
.78
7
352.79
--
--
(

ACCEPTED MANUSCRIPT
MRT(0-t)
9569.54±
2017.58
1
1
.25
.58
--
3.24
3.48
4.69±0.13
6.33±0.43
--
9.36±1.86
3.59±0.10
(
h)
MRT(0-∞)
h)
13311.12±23
12.20
20.36±7.04 29.25±24.28
(
F (%)
45.01
45.10±7.12
--
39.15±18.06
a
All data were obtained by triple testing (±SD) except mice.
Finally, we tested 40’s antitumor efficacy in MV4-11 cell-inoculated xenograft
mouse model via oral administration (Figure 7A-B). Treatment of 40 at all dosages, i.e.,
0, 20 and 30 mg/kg/day, could almost completely suppress the tumor progression in the
1
first two weeks and did not affect the animal’s weight indicating that there was no
general cytotoxicity at these doses. Interestingly, after stopped the treatment of 40, the
tumors of the animals treated with 10 mpk drug dosage started to grow again. However,
this tumor recurrence was not observed in the 20 and 30 mpk dosage groups during the
following week after administration of compound 40 was stopped and p values were
quantified on the 21st day, which were 0.042, 0.0035 and 0.0028, respectively.
Figure 7. Compound 40’s antitumor efficacy in MV4-11 xenograft mouse model. Female
nu/nu mice bearing established MV4-11 tumor xenografts were treated with 40 at 10, 20,
and 30 mg/kg/d dosage or vehicle. Daily oral administration was initiated when MV4-11

ACCEPTED MANUSCRIPT
3
tumors had reached a size of 200−400 mm . Each group contained five animals. Data,
−
1
−1
mean ± SEM mpk, mg kg day . (A) Body weight, and (B) Tumor size measurements
from MV4-11 xenograft mice after 40 administration. Initial body weight and tumor size
were set as 100%. ‘*’, p<0.05. ‘**’, p<0.01.
3
. Conclusions
Starting from the core scaffold of CDK7/9 inhibitor 9, a structure-guided drug design
approach led to the discovery of a novel highly potent and selective CDK9 kinase
inhibitor 40. It exhibited single digit nM potency against CDK9 kinase in the biochemical
assay. Compared to most of the currently known CDK kinase inhibitors, 40 achieved
superior selectivity among CDK subfamily, especially the high selectivity between
CDK9 and CDK7 kinase (over 1700-fold), both of which are critical transcription
regulators. Therefore 40 would provide an ideal pharmacological tool for dissecting the
CDK7- and CDK9-related physiology and pathology. In addition, 40 exhibited great
antiproliferative activities in a range of established cancer cell lines including solid tumor
and leukemia cells through inducing the apoptotic cell death. Meanwhile it showed a
reasonable therapeutic window over normal cells, indicating that selective CDK9 kinase
inhibitor might have potential antitumor efficacy in a wide range of the cancers. The
impressive antitumor efficacy of 40 in the in vivo experiment especially that the tumor
remained un-relapsed after stopping the drug treatment indicated a possible
treatment-holiday-treatment drug application approach, which would provide a better
therapeutic window and a more economical way for clinical usage. The high potency and
selectivity, acceptable PK profile as well as good in vivo antitumor efficacy of 40 would
make it a potential drug candidate for a range of cancers especially leukemias.

ACCEPTED MANUSCRIPT
4
. Experimental section
4
.1. Chemistry. All reagents and solvents were purchased from commercial sources and
1
used as obtained. H NMR spectra were recorded with a Bruker 400 MHz NMR
3
spectrometer and referenced to deuterated methanol (CD OD), deuterium dimethyl
sulfoxide (DMSO-d ) or deuterium chloroform (CDCl ). Chemical shifts are expressed in
6
3
ppm. In the NMR tabulation, s indicates singlet; d, doublet; t, triplet; q, quartet; m,
multiplet; and br, broad peak. LC/MS experiments were performed on an Agilent 6224
TOF using an ESI source coupled to an Agilent 1260 Infinity HPLC system operating in
reverse mode with an Agilent Eclipse Plus C18 1.8 ꢀm, 3.0 mm
× 50 mm column. Flash
column chromatography was conducted using silica gel (Silicycle 40−64 ꢀm). The
purities of all final compounds were determined to be above 95% by HPLC.
4
.1.1. tert-Butyl (4-(5-chloro-2-fluoropyridin-4-yl)thiazol-2-yl)carbamate (11) A mixture
of 48 (3.61 g, 12.9 mmol) and Na CO (3.4 g, 32.3 mmol) in DME/H O/dioxane
2
3
2
(240/48/48
mL)
was
exchanged
with
N
2
twice,
and
5
2
-chloro-2-fluoro-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)pyridine (45a) (6.6 g,
5.9 mmol) and Pd(PPh (1.45 g, 1.25 mmol) were added to the mixture. The reaction
3 4
)
mixture was heated to 100 °C and stirred for 6 h under argon atmosphere. The solid was
removed by centrifugation at 3000 rpm, 25 °C for 20 min. The supernatant was
concentrated and the crude product was purified by silica gel flash chromatography
(eluting with ethyl acetate in petroleum ether 2-5%) to give 11 as a white solid (1.57 g,
1
yield = 37%). H NMR (400 MHz, DMSO-d ) δ 11.79 (s, 1H), 8.44 (s, 1H), 8.11 (s, 1H),
6
+
7
.61 (s, 1H), 1.50 (s, 9H); LC/MS (ESI, m/z) 274.05 [M+H] .

ACCEPTED MANUSCRIPT
4
.1.2.
-(5-Chloro-2-fluoropyridin-4-yl)-N-((tetrahydro-2H-pyran-4-yl)methyl)thiazol-2-amine
and
-(5-chloro-2-((2-methoxyethyl)amino)pyridin-4-yl)-N-((tetrahydro-2H-pyran-4-yl)methy
4
(12)
4
l)thiazol-2-amine (13) 53l (0.3 g, 0.71 mmol), 2-methoxyethan-1-amine (0.16 g, 2.1
mmol) and DIEA (0.36 g, 2.8 mmol) were dissolved in DMSO (10 mL) under N2
o
protection and the reaction mixture was stirred at 100-110 C for two days. The resulting
mixture was poured into cold water and extracted with ethyl acetate (3 × 20 mL) and
washed with water (3 × 20 mL) and brine (20 mL). The combined organic layers were
dried over anhydrous Na SO and evaporated to dryness. The crude product was purified
2
4
by silica gel flash chromatography (eluting with ethyl acetate in petroleum ether 2-5%) to
give 12 (53 mg, yield = 23%) and 13 (0.33 g, yield = 43%) as white solid. Compound 12,
1
H NMR (400 MHz, DMSO-d ) δ 8.39 (s, 1H), 7.95 (s, 1H), 7.63 (s, 2H), 3.85 (s, 2H),
6
+
3
.22 (s, 4H), 1.65-1.84 (m, 3H), 1.22 (s, 2H); LC/MS (ESI, m/z) 328.07 [M+H] ;
1
Compound 13, H NMR (400 MHz, DMSO-d
6
) δ 7.94 (s, 1H), 7.76 (s, 1H), 7.28 (s, 1H),
7
.09 (s, 1H), 6.88 (s, 1H), 3.86 (s, 2H), 3.38-3.46 (m, 5H), 3.08-3.14 (m, 3H), 1.57-1.87
+
(
m, 4H), 1.24 (s, 3H); LC/MS (ESI, m/z) 383.13 [M+H] .
.1.3 Compounds 14-15 were prepared following the synthetic procedure of 12/13 only
4
．
the base DIEA was replaced with NaH. Compounds 21-26, 30-43 were prepared
following the synthetic procedure of 12/13. Compounds 28 and 29 were prepared
following the same procedure only the base DIEA was replaced with K CO .
2
3
4
.1.3.1.
-(5-Chloro-2-(phenylamino)pyridin-4-yl)-N-((tetrahydro-2H-pyran-4-yl)methyl)thiazol-
4