Structure-based design of highly selective 2,4,5-trisubstituted pyrimidine CDK9 inhibitors as anti-cancer agents

Article abstract of DOI:10.1016/j.ejmech.2021.113244

Cyclin-dependent kinases (CDKs) are a family of Ser/Thr kinases involved in cell cycle and transcriptional regulation. CDK9 regulates transcriptional elongation and this unique property has made it a potential target for several diseases. Due to the conserved ATP binding site, designing selective CDK9 inhibitors has been challenging. Here we report our continued efforts in the optimization of 2,4,5-tri-substituted pyrimidine compounds as potent and selective CDK9 inhibitors. The most selective compound 30m was >100-fold selective for CDK9 over CDK1 and CDK2. These compounds showed broad anti-proliferative activities in various solid tumour cell lines and patient-derived chronic lymphocytic leukaemia (CLL) cells. Decreased phosphorylation of the carboxyl terminal domain (CTD) of RNAPII at Ser-2 and down-regulation of anti-apoptotic protein Mcl-1 were confirmed in both the ovarian cancer model A2780 and patient-derived CLL cells.

Full text of DOI:10.1016/j.ejmech.2021.113244

European Journal of Medicinal Chemistry 214 (2021) 113244
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Structure-based design of highly selective 2,4,5-trisubstituted
pyrimidine CDK9 inhibitors as anti-cancer agents
,
,
Hao Shao a 1, David W. Foley a 1, Shiliang Huang a, Abdullahi Y. Abbas a, Frankie Lam b,
Pavel Gershkovich a, Tracey D. Bradshaw a, Chris Pepper c, Peter M. Fischer a
,
,
**
Shudong Wang a
, b, *
a School of Pharmacy and Biodiscovery Institute, University of Nottingham, University Park, Nottingham, NG7 2RD, UK
b Drug Discovery and Development, Clinical and Health Sciences, University of South Australia, Adelaide, SA, 5001, Australia
c Brighton and Sussex Medical School, University of Sussex, Brighton, East Sussex, BN1 9PX, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Cyclin-dependent kinases (CDKs) are a family of Ser/Thr kinases involved in cell cycle and transcriptional
regulation. CDK9 regulates transcriptional elongation and this unique property has made it a potential
target for several diseases. Due to the conserved ATP binding site, designing selective CDK9 inhibitors has
been challenging. Here we report our continued efforts in the optimization of 2,4,5-tri-substituted py-
rimidine compounds as potent and selective CDK9 inhibitors. The most selective compound 30m was
>100-fold selective for CDK9 over CDK1 and CDK2. These compounds showed broad anti-proliferative
activities in various solid tumour cell lines and patient-derived chronic lymphocytic leukaemia (CLL)
cells. Decreased phosphorylation of the carboxyl terminal domain (CTD) of RNAPII at Ser-2 and down-
regulation of anti-apoptotic protein Mcl-1 were confirmed in both the ovarian cancer model A2780
and patient-derived CLL cells.
Received 12 December 2020
Received in revised form
25 January 2021
Accepted 25 January 2021
Available online 2 February 2021
Keywords:
RNA polymerase II
CDK9 inhibitor
Apoptosis
© 2021 Elsevier Masson SAS. All rights reserved.
Mcl-1
Chronic lymphocytic leukaemia
Anti-Cancer agents
1. Introduction
the transcription initiation complex. Subsequent phosphorylation
of Ser2, and occasionally Ser5, on the carboxyl terminal domain
Mammalian cells possess 21 cyclin-dependent kinases (CDKs)
that fulfil major roles in cell cycle and transcription regulation [1].
CDKs1, 2, 4 and 6 are required to facilitate cell cycle progression,
CDKs 8, 9, 12 and 19 are involved in the regulation of transcription
through phosphorylation of the C-terminus of RNA polymerase II.
While CDK7 is involved in both processes [2,3], CDK9 plays no
clearly defined role in cell-cycle regulation but promotes tran-
scription elongation. After initiation of RNA transcription, positive
transcription elongation factor b (P-TEFb), consisting of cyclinT-
CDK9, first phosphorylates the SpT5 (p160) subunit of DRB-
sensitivity-inducing factor (DSIF) and the RD subunit of negative
elongation factor (NELF) to remove promoter-proximal pausing on
(CTD) heptapeptide repeat of the largest subunit of RNAPII, pro-
motes productive elongation [4e8].
Constitutive expression of anti-apoptotic proteins is necessary
for cancer cell’s survival, which requires continuous activity of
RNAPII. Mcl-1, an anti-apoptotic protein belonging to the Bcl-2
family, is highly expressed in numerous tumour cells in compari-
son with normal cells [9,10]. It has been reported that the inhibition
of Mcl-1 is sufficient to promote apoptosis in acute myeloid
leukaemia (AML) and CLL cells [11,12] and to overcome the resis-
tance that cancer cells acquire upon treatment with Bcl-2 inhibitors
[13]. Inhibition of CDK9 can abrogate the production of short-lived
mRNA transcripts with anti-apoptotic functions and their corre-
sponding proteins, thus leading to apoptosis of cancer cells [4].
Selective CDK9 inhibitors could therefore be used as either a single
agent or in combination therapy. Indeed, multiple CDK9 inhibitors
have shown efficacy in both solid tumors and haematological
cancers [14,15]. The CDK9 inhibitors voruciclib and A-1592668 can
potentiate the anti-cancer efficacy of the Bcl-2 inhibitor ABT-199
[16,17]. We also reported that an orally bioavailable CDK9 inhibitor
* Corresponding author. Clinical and Health Sciences, University of South
Australia, Adelaide, SA, 5001, Australia.
** Corresponding author.
E-mail addresses: peter.fischer@nottingham.ac.uk (P.M. Fischer), shudong.
wang@unisa.edu.au (S. Wang).
1
co-first authors.
https://doi.org/10.1016/j.ejmech.2021.113244
0223-5234/© 2021 Elsevier Masson SAS. All rights reserved.

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
Abbreviations
KDR
Kinase Insert Domain Receptor
LDA
Lithium diisopropylamide;
myeloid cell leukaemia sequence 1
Acetonitrile;
Mixed-Lineage Kinase 1
N-bromosuccinimide;
ALL
acute lymphocytic leukaemia
Mcl-1
MeCN
MLK1
NBS
ATP
adenosine triphosphate
B-cell lymphoma 2
cyclin-dependent kinase
carboxyl terminal domain
dichloromethane
Bcl-2
CDK
CTD
DCM
DMF
NCS
N-chlorosuccinimide;
NELF
NEK2
PARP
P-TEFb
SAR
negative elongation factor
NIMA Related Kinase 2
poly(ADP-ribose) polymerase
positive transcription elongation factor b
structure and activity relationships
N,N-dimethylformamide;
DMF-DMA dimethylformamide dimethyl acetal
DMSO
DMAP
HRMS
dimethylsulfoxide;
dimethylaminopyridine;
high resolution mass spectra
CDKI-73 is active as a single agent in CLL [18], acute lymphoblastic
leukaemia (ALL), AML [19], and colorectal cancer models [20]. In
addition, CDKI-73 synergized with fludarabine in CLL [18], Bcl-2
inhibitor ABT-199 and BET inhibitor I-BET 151 in ALL and AML
[21,22], and PARP inhibitor olaparib in BRCA1 wide-type ovarian
cancer [23].
Due to the therapeutic potential of CDK9 inhibition in cancer,
tremendous efforts have been devoted to the CDK9 targeted drug
discoveries [15,24]. Flavopiridol was the first CDK inhibitor tested
in clinical trials and the mechanistic studies showed that it ach-
ieved its efficacy mainly through potent inhibition of CDK9 in CLL
[25]. More non-selective CDK9 inhibitors have since been discov-
ered and are currently in various preclinical and clinical stages of
development, summarized in a recent review [15]. In the past few
years, selective CDK9 inhibitors such as NVP-2 [26], JSH-150 [27],
AZD5473 [28] and compound 30i [29] have been reported. Our and
other studies showed that the selectivity for CDK9 can be achieved
through modification of the substituents located at the hydropho-
bic channel adjacent to solvent [27,30,31]. In addition to this po-
sition, we gained CDK9 selectivity through modifying the
substituents located at the hydrophobic pocket behind the gate-
keeper residue. In this manuscript, we report our continued efforts
in the design, synthesis and biological evaluation of 2,4,5-
trisubstituted pyrimidine compounds as potent and selective
CDK9 inhibitors.
Scheme 1. Reagents and conditions: (a) i. 1,1,1-trifluoropentane-2,4-dione, hydrox-
y(tosyloxy)iodobenzene, MeCN, , 1 h; ii. 1, , 4 h, 53%; (b) Boc2O, DMAP, CH2Cl2, RT,
2 h, 95%; (c) DMF-DMA, , overnight or microwave radiation, 30 min-1 h, 77e92%. (d)
DMF-DMA, CHCl3, RT, overnight, 98%; (e) CH3COCH2Cl, MeCN, , 4 h, 79%; (f) Br2,
, 4 h, 50e53%; (h) i. Acetaldehyde, LDA, THF,
D
D
D
D
MeOH, 0 ꢀC to rt 4 h, 76e77%; (g) 1,
D
ꢁ78 ꢀC, 2 h; ii. MnO2, CHCl3, RT, 4 h, 65e85%.
2. Results and discussions
2.1. Chemistry
Enaminones 11a-d were synthesized using the methods previ-
ously reported, which are shown in Scheme 1. Briefly, 1,1,1-
trifluoropentane-2,4-dione was treated with hydroxy-(tosyloxy)
iodobenzene, followed by reaction with 1-methylthiourea 1 to
obtain thiazole 2. Condensation of methyl-thiourea 1 with DMF-
DMA yielded 4 [32] followed by thiazole ring formation by cycli-
zation with chloroacetone to afford 5-acetyl-4-hydrogen thiazole 5
[33]. 5-Acetyl thiazoles 2 and 4 were Boc protected then converted
to the corresponding enaminones 11a-b using standard conditions.
Scheme 2. Reagents and conditions: (a) 3-chloro-2,4-pentadione, pyridine, EtOH, RT,
4 h, 98%; (b) DMF-DMA,
Boc2O, DMAP, CH2Cl2, room temp, 2 h; (e) propionaldehyde, LDA, THF, ꢁ78 ꢀC, 2 h; (f)
MnO2, CHCl3, RT, 4 h, 65e85%; (g) DMF-DMA, microwave radiation, 30 min-1 h.
D
, MeCN, overnight, 90%; (c) NCS, MeOH, 0 ꢀC, 2 h, 60%; (d)
corresponding enaminone 11e, which was converted to chlorine
substituted enaminone 11f using N-chlorosuccinimide in moderate
yields. 1-Methylthiourea 1 was treated with chloroacetone to afford
the thiazole 13. After Boc protection, compound 14 was deproto-
nated with LDA and reacted with propionaldehyde to form the
alcohol 15. Ketone 16 was formed by oxidation using manganese
dioxide. Enaminone 11g was synthesized using a microwave
method and used in the next step without purification.
a
-bromoketones 8 were formed from the corresponding ketones
using bromine and then condensed with 1-methylthiourea 1 to
afford the 5-H thiazoles 9. After Boc protection, 9 was deprotonated
with LDA and reacted with acetaldehyde to form the alcohols,
which were oxidized to ketones 10 using manganese dioxide [34].
Scheme 2 shows the synthesis of the enaminone 11f and 11g.
The synthesis of enaminone 11f started from 1-methylthiourea 1,
which was cyclized with 3-chloro-2,4-pentadione to yield 5-acetyl
thiazole 12 [35]. 12 was then treated with DMF-DMA to form the
The synthesis of enaminone 11h is outlined in Scheme 3.
2

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
and reacted with cyanamide as previously described [30]. The
amines employed were readily commercially available, with the
exception of the bicyclic amine 1,4-diazabicyclo[3.2.2]nonane 29,
which was synthesized via
a Beckmann rearrangement and
reduction of the commercially available quinuclidin-3-one 26.
Finally, the component parts of the final compounds were
combined in a condensation reaction using microwave conditions
as previously reported and summarized in Scheme 5 [30]. A Stille
coupling was used to introduce a vinyl substituent in place of a
chlorine at the C5-position of the pyrimidine, as shown in Scheme
5.
The introduction of a bromo substituent into the 5-position of
the pyrimidine differs from the previous synthetic method and is
summarized in Scheme 6. Briefly, the thiazole 13 was brominated
using N-bromosuccinimide in acetic acid to afford 31. After Boc
protection of the amine, compound 32 was treated with butyl
lithium for 30 min and then 5-bromo-2-chloropyrimidine. After
completion of the reaction, it was quenched with acetic acid and
DDQ was added to yield compound 33. The final compound 30q
was synthesized by treating compound 33 with 3-(4-acetyl-1-
piperazinyl)aniline.
Scheme 3. Reagents and conditions: (a) Br2, 48% HBr solution, 60 ꢀC, 3 h, 56%; (b)
Boc2O, DMAP, CH2Cl2, RT, 2 h; (c) trimethylphosphite, RT, 22 h; (d) FO2SCF2CO2CH3,
CuI/DMF, 80 ꢀC, overnight, 37% over 3 steps; (e) DMA-DMA,
D, 1 h, 64%.
Thiazole 12 was treated with two equivalents of a 48% solution of
bromine in HBr to obtain the dibromo compound 17 [36]. After Boc
protection, b-bromoenol phosphate 19 was obtained following a
Perkow reaction. Treatment of 19 with methyl 2,2-difluoro-2-(flu-
orosulfonyl)acetate in the presence of CuI led to the formation of
the desired trifluoromethylated intermediate 20. The enaminone
11h was then synthesized using the standard conditions.
The synthesis of the guanidine intermediates is summarized in
Scheme 4. Mono-substituted guanidines were synthesized as pre-
viously described via nucleophilic aromatic substitution (NAS) of
fluoro-nitrobenezene 21a [35]. For di-substituted guanidines the
corresponding NAS approach was not successful, so Buchwald
coupling of 22b-d was employed. The nitro group was then reduced
2.2. Ligand design, structure-activity relationships and molecular
docking
We have previously reported 2,4,5-trisubstituted pyrimidine
compounds as potent CDK9 inhibitors but selectivity for CDK9 over
CDK2 remained a challenge [30,31]. We elected to rationally design
inhibitors with improved selectivity for CDK9 by taking advantage
of the structural differences between CDK2 and CDK9. By
comparing the amino acid residues that form the active site of
CDK2 and CDK9, we hoped to identify non-conserved regions that
could be targeted with suitably designed compounds [37].
Although CDK2 and CDK9 share the same phenylalanine gate-
keeper residue, differences in the hydrophobic pocket behind the
gatekeeper residue have been successfully exploited to generate
selective CDK4 and CDK7 inhibitors [38,39]. In the hydrophobic
channel adjacent to solvent, residues are less conserved. For
example, CDK9 has a non-ionisable side chain in the Lys89 (posi-
tively charged) position of CDK2, and the electrostatic charge dif-
ferences of the hydrophobic channel can be used for improving the
CDK9 selectivity [40]. In addition, the region around Ala111-Gly112
in CDK9 is much more open than the corresponding Lys88-Lys89
region in CDK2 due to differences in size between alanine, glycine
and lysine. This pocket in CDK9 is therefore both larger and more
hydrophobic than CDK2, both features that could be exploited for
selectivity [40,41]. Residues around this region among CDKs are
less conserved and have been used for optimizing selectivity.
Several hypotheses were proposed to enable the design of
Scheme 4. Reagents and conditions: (a): H2SO4, NaNO2, EtOH,
D, 3 h, 69%e89%; (b):
Pd2(dba)3, (þ/ꢁ)-BINAP, Cs2CO3, Dioxane, 90 ꢀC, 5d, 16%e89%; (c): amines, DMSO,
100 ꢀC, 72 h, 53e77%; (d): RaNi, NH2NH2$H2O, 0 ꢀC, 30 min-1 h or H2, Pd/C, EtOH:
EtOAc 1:1, overnight, 90%e96%; (e): NH2CN, TMSCl, ACN,
D
,
6
h
or microwave,
D, 90 min, 65%;
100e140 ꢀC, 20e45 min, 35%e78%; (f): NH2OH$HCl, C2H3NaO2, H2O,
Scheme 5. Reagents and conditions. (a) 2-methoxyethanol, microwave, 100e140 ꢀC,
20e45 min; (b) (tributylstannyl)ethylene, 3% Pd[(t-Bu)3P]2, CsF, 1,4-dioxone, micro-
wave, 120 ꢀC, 2 h, 60%.
(g): H3PO4, 120e140 ꢀC, slow addition of 27 over 2 h, 38%; (h): LiAlH4, dioxane,
D, 6h,
83%.
3

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
significantly enhanced the CDK9/CDK2 selectivity to 44-fold. A
further improvement of CDK9/CDK2 selectivity was obtained by
replacing the morphine moiety with 4-acetyl-piperazine. Selec-
tivity was completely abolished when 30f was ring constrained into
the bridged analog 30g or when the C-5 methyl was replaced by
hydrogen (30v), which supports our hypotheses that inducing a
twist between the pyrimidine and thiazole rings could be exploited
to improve selectivity. In 30g, molecular modelling suggested that
the thiazole and pyrimidine rings were coplanar because of the
restricted rotation, whereas the clash of the methyl groups in 30f
resulted in a twist of the two rings, generating a non-planer
conformation (Fig. 2). It is possible that this conformational
change drives the selectivity between CDK9 and CDK2, and that
CDK9 can better accommodate the non-planer conformation due to
the greater flexibility of its binding pocket.
The kinase selectivity profile varied greatly in the C5-methyl
series of compounds when there were minor changes in the R3
substituents. The piperazine analog 30h gave the best selectivity
between CDK9 and CDK2 among the six membered-ring analogs,
~90-fold. Methylation of the piperazine resulted in 30i, which
decreased the Ki to 20 nM and the CDK9/CDK2 selectivity decreased
from 88-fold to 28-fold. The methylsulfonyl piperazine analog 30j
did not change the CDK9 potency, but the CDK9/CDK2 selectivity
was only ~12-fold. These analogs showed variable cellular potency,
Scheme 6. Reagents and conditions: (a) NBS, AcOH, RT, 2 h, 88% (crude); (b) Boc2O,
DMAP, CHCl3, RT, 2 h, 60%; (c) i. LDA, ꢁ78 ꢀC, 30 min; ii. 5-bromo-2-chloropyrimidine,
ꢁ78 ꢀC, 1 h; iii. AcOH, DDQ in THF; 77%; (d) 3-(4-acetyl-1-piperazinyl)aniline, 2-
methoxyethanol, microwave radiation, 140 ꢀC, 45 min, 28%.
highly selective CDK9 inhibitors based on the differences discussed
above: (i) Co-crystal structures of the lead compound 30a bound to
CDK2 revealed that the C-4 methyl of the thiazole was packed up
against the Phe80 gatekeeper residue (Fig. 1) [35]. Modification of
this R1 substituent may therefore dramatically change the kinase
inhibition profile; (ii) R2 could be extended into the hydrophobic
pocket to form favorable interactions with the gatekeeper residue
Phe103 of CDK9, evidenced by compound 30b (Fig. 1). In addition,
the introduction of substituents at this position induces a steric
clash with the R1 substituent on the thiazole ring. This causes a
twist in the conformation of the molecule when compared to the
observed co-planarity of the thiazole and pyrimidine rings in a co-
crystal structure of 30a with CDK2 [35]. Additionally, the steric
clash induced by substitution on both the thiazole and pyrimidine
rings could induce an even greater twist of these two ring systems
than that observed in the crystal structures. This would dramati-
cally alter the conformation of the molecule and could therefore
adjust the SAR. (iii) Various R3 substituents with different sizes and
electronic properties could be introduced to the molecule to opti-
mize the kinase potency and selectivity, as well as exploring
tolerance for future optimization of physicochemical properties.
Based on these rationales, we synthesized a series of com-
pounds and their structure activity relationships (SARs) are sum-
marized in Table 1. Compound 30c, containing chlorine at the C-5
position of the pyrimidine ring, was a low nanomolar potent CDK9
inhibitor and showed ~21-fold selectivity for CDK9 over CDK2.
When the methylpiperazine of R3 was changed to morpholine,
compound 30d showed comparable CDK9 potency with compound
30c and the CDK9/CDK2 selectivity dropped to 5-fold. Both of these
compounds were potent against HCT-116 and MCF-7 cell lines with
with GI50 values from 0.50 to 1.89 mM in the two cancer lines. The
CDK9/CDK2 selectivity was reduced when the NH was replaced
with an O (30e), acetylated (30f), methylated (30i) or converted to
the sulfonamide (30j), suggesting that a hydrogen bond donor or a
positively charged moiety in this position was critical for selectivity.
Introduction of the bulkier 7-membered 1,4-diazepane ring at
the R3 position led to the most selective compounds in this series,
exemplified by compound 30k, which showed ~156-fold selectivity
for CDK9 over CDK2. Yet again, disrupting the hydrogen bond donor
ability of the 1,4-diazepane by conversion to the lactam 30l greatly
reduced selectivity. Bridged diazepane analog 30m maintained the
selectivity for CDK9 over CDK2, ~120-fold. There was little change
in the cellular potencies of selective compounds compared to non-
selective compounds (cf 30m and 30d). This suggested that inhi-
bition of CDK9 also contributed to the observed anti-cancer activity
to a large extent.
To rationalize the selectivity of 30k and 30m, we docked the
molecules into active sites of the models generated from the crystal
structures 4BCG (CDK9) and 4BCP (CDK2) using a rigid protein
docking method as discussed previously. 30k was predicted to bind
to the hinge region of CDK9 and CDK2 in a similar manner to the
observed crystallographic binding mode of the inherent ligand
(Fig. 3A and B). The meta-anilino substituent of 30k was bound
differently in CDK9 and CDK2. In the CDK9/cyclinT/30k complex
(Fig. 3A), the 1,4-diazepane moiety was positioned towards the
thiazole ring, adjacent to the ribose position of ATP in the kinase
structures, which we termed as an ‘inward’ conformation. By
contrast, in the CDK2/cyclinA/30k complex, this moiety was posi-
tioned away from the thiazole and directed towards the hinge re-
gion (Fig. 3B), which was termed the ‘outward’ conformation.
However, when we attempted this docking with 30m, it became
clear that it could not be accommodated by the CDK9 model
generated from 4BCG. As the protein structure was rigid during the
docking process, the diazepane moiety was too bulky to fit in the
pocket. We therefore attempted to dock 30m into receptors
generated from alternative crystal structures of CDK9 [4BCF] and
CDK2 [4BCO]. In this case, compound 30m showed a similar
binding pose to compound 30k in CDK9 (Fig. 3C). A clash between
the methyl substituent of the thiazole ring with CDK2 was observed
(Fig. 3D). When the two sets of CDK9 crystal structures (4BCG and
4BCF) were overlaid, the different positions of glycine-rich loop
GI50 values of ~0.5
Keeping the morpholine moiety but replacing the C-5 chlorine
with methyl substituent, afforded compound 30e, which
mM.
a
Fig. 1. Binding poses of lead compounds 30a and 30b within CDK2 and CDK9: (a)
Crystal structure of 30a bound to CDK2 (PDB code 1PXN); (b) Crystal structure of 30b
bound to CDK9 (PDB code 4BCJ). Selected CDK2 and CDK9 residues are drawn in stick
representation and the carbons of CDK2/9 are green. The carbons of 30a and 30b are
grey. Hydrogen bonds are depicted by dotted red lines. The images were generated by
Pymol.
4

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
Table 1
Structures and biological activities..
Compd
Structure
R1
Ki (nM)a
SIc
GI50 (m
M)b
R2
R3
CDK9T1
CDK2A
HCT-116
MCF-7
30a
30b
30c
30d
30e
30f
30g
30h
30i
Me
Me
Me
Me
Me
Me
H
CN
Cl
p-OH
m-OH
m-methylpiperazine
m-morpholine
m-morpholine
e
70
12
e
e
e
11
5
4
10
12
34
8
20
7
1
21
6
0.03 0.005
0.66 0.06
0.43 0.08
0.36 0.11
0.55 0.01
0.06 0.02
0.83 0.04
1.29 0.16
0.50 0.05
1.37 0.70
2.12 0.58
1.07 0.04
4.80 0.14
4.05 0.21
>5
0.07 0.01
0.68 0.11
0.72 0.03
0.45 0.20
3.65 0.37
0.08 0.02
0.86 0.05
1.89 0.25
0.83 0.18
0.87 0.15
6.44 0.26
1.34 0.28
25.0 3.96
3.05 0.35
>5
107
22
444
794
87
711
586
87
1244
525
1248
1223
1934
503.5
125
14
Cl
Me
Me
e
44
66
2.5
89
29
12
156
28
125
13
40
8
16
2
7
3
8
m-4-acetylpiperazin-1-yl
e
e
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
CF3
H
cyclopropyl
isopropyl
Me
Me
Me
Me
Me
Me
Me
CF3
CF3
vinyl
Br
m-piperazine
m-methylpiperazine
m-4-(methylsulfonyl)piperazine
m-1,4-diazepane
m-1,4-diazepan-5-one
m-1,4-diazabicyclo[3.2.2]nonane
m-4-acetylpiperazin-1-yl
m-piperazine
m-morpholine
m-4-acetylpiperazin-1-yl
m-1,4-diazepane
30j
30k
30l
8
19
10
95
49
62
7.5
8
30m
30n
30o
30p
30q
30r
30s
30t
30u
30v
e
e
H
H
H
H
0.07 0.01
0.08 0.004
0.77 0.01
0.08 0.01
0.58 0.08
0.65 0.01
0.06 0.01
1.95 0.02
0.07 0.01
5.28 0.25
m-1,4-diazepane
3
22
m-CO-N-(1-methylpiperidin-4-yl)
m-4-acetylpiperazin-1-yl
m-4-acetylpiperazin-1-yl
40
91
32
128
680
281
H
9
a
The data given are mean values derived from two replicates of one experiment. Apparent inhibition constants (Ki) were calculated from IC50 values and the appropriate Km
(ATP) values for each kinase.
b
Anti-proliferative activity by MTT-48 h assay; the data given are mean values derived from at least three independent replicates, internal repeats n ¼ 3.
c
SI represents selectivity index (Ki, CDK2/Ki, CDK9).
Fig. 2. Overlay of compound 30f and 30g docked with CDK9 (PDB ID 4BCG). The kinase
is presented in green in cartoon representation. Carbon atoms of 30f and 30g are grey
and cyan respectively. Hydrogen bonds are denoted by red dashed lines. The images
were generated by Pymol.
Fig. 3. Docking poses for complexes of CDK9 and CDK2 with compound 30k and 30m.
(A): CDK9/30k, model generated from PDB 4BCG; (B): CDK2/30k, model generated
from PDB 4BCP; (C): CDK9/30m, model generated from 4BCF; (D): CDK2/30m, model
generated from 4BCO. The kinase is presented in green in cartoon representation and
selected residues are drawn in lines mode. Carbon atoms of 30k and 30m are grey and
hydrogen bonds are denoted by red dashed lines. The images were generated by
Pymol.
were observed and it is possible that the ligand binding pocket
became smaller because of the downward movement of the
glycine-rich loop in 4BCG compared to that in 4BCF. For proteins as
flexible as CDK9, modelling using the rigid docking platforms may
not be appropriate.
Methyl groups have been frequently replaced by trifluoromethyl
groups as bioisosteres in medicinal chemistry. The trifluoromethyl
5

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
group is generally more metabolically stable than the methyl group
because of the strong CeF bond and has also been used for
improving oral bioavailability [42,43]. The trifluoromethyl group is
also a strong electron-withdrawing group and may help to improve
the hinge region binding between ligand and protein. Therefore,
the corresponding C-5 trifluoromethyl pyrimidine analogs were
synthesized. Compounds 30n and 30o were 10-40-fold selective
towards CDK9 over CDK2, but lost potency against both kinases and
cells compared with its C-5 methyl analogs 30f and 30h. These two
trifluoromethyl analogs had GI50 against two human cell lines
potency, although a similar selectivity profile to the corresponding
methyl analog 30v was observed. Interestingly, compound 30u was
much more potent against both cancer cells than 30v, indicating
potential off-target effects. In order to rationalize why compounds
with R1 as cyclopropyl and isopropyl substituents lost activities,
molecular docking simulations were carried out. Compound 30t
and 30u were docked into CDK9 (PDB code 4BCG) and the binding
poses of the top 5 scores were visualized. No inward conformation
was observed with compound 30t in the top 5 docking poses and
only one inward pose of compound 30u was observed among the
top 5 scores. More poses with the thiazole rings flipping about 180ꢀ
were observed, which may because the small hydrophobic pocket
around the gatekeeper residue Phe103 was too small to accom-
modate the isopropyl and cyclopropyl substituents.
Our above SAR clearly indicated a trend between increasing the
bulk of this 7-membered ring still further and improving CDK9
selectivity over CDK2. We therefore proposed that molecules con-
taining bulky groups at both meta positions of the aniline would
show preferential binding to CDK9, as these must place some bulk
in the inward conformation by virtue of their di-substitution
pattern. A series of such compounds was therefore synthesized
(30w-30ae, Table 2). In general, bulky disubstituted anilines were
more selective for CDK9 than CDK2, with compound 30ae being 30-
fold selective for CDK9 compared to 9-fold for 30v. However, dra-
matic improvements to selectivity were not observed and no
compound in this series approached the 100-fold selectivity
window.
We proposed three hypotheses to improve the selectivity for
CDK9 over other CDKs. Introducing substituents with various size
and electronic properties at R1 position did not change the selec-
tivity profile, but affected compounds’ potency against CDKs. Only
by introducing a substituent at R2 position, especially methyl group,
dramatically improved the selectivity for CDK9 over other CDKs.
This confirmed our hypothesis that a conformation change induced
by the clash of R1 and R2 substituents is important for selectivity.
Noticing that the pocket close to solvent exposed region was more
open in CDK9 than CDK2, we designed and synthesized bulky
mono- and di-substituted analogs. 7-membered 1,4-diazepane and
bridged diazepane substituted analogs 30k and 30m achieved the
ranging from 3.05 to 25.0 mM. A study of van der Waals volumes of
trifluoromethyl and various alkyl groups has suggested that a tri-
fluoromethyl group has a similar size to an ethyl group, but their
shapes are different [44,45]. Given that our previous SAR showed
that a C-5 ethyl pyrimidine analog lost both CDK9 and CDK2 po-
tency [30], we believe this steric effect explains why the tri-
fluoromethyl group was not tolerated. This is further supported by
the fact that a vinyl substituent in this position (30p) was also not
tolerated. The C-5 bromo pyrimidine analog 30q showed compa-
rable CDK9/CDK2 selectivity (16-fold) to its chlorine analog 30c.
From the above SAR, it can be concluded that both the steric and
electronic properties of substituents at the C-5 position of the py-
rimidine ring affect the potency and selectivity of kinase inhibition.
The C-5 chloro, bromo and methyl pyrimidine analogs have
different selectivity profiles presumably because of their different
electronic properties, considering their similar size. It was clear
from the above SAR that the introduction of a methyl group at the
C-5 of the pyrimidine ring was optimal for achieving selectivity for
CDK9.
After exploring the SAR of R2 position, we turned our attention
to the R1 position of the thiazole ring to further probe the gate-
keeper pocket. By keeping six- or seven-membered rings at the R3
position on the aniline ring, compounds 30r-30v with various R1
substituents were synthesized to interrogate SAR. The kinase po-
tency and cellular activity of compounds with small R1 groups (30r,
R1 ¼ CF3, and 30s, R1 ¼ H) was improved and selectivity for CDK9
over CDK2 was lost. When the bulk at R1 was increased to a
cyclopropyl unit, both potency and selectivity (compound 30t)
were decreased. An isopropyl analog 30u also showed a loss in
Table 2
Structures and biological activities..
Compd
Structure
R2
Ki (nM)a
CDK9T1
SIc
GI50 (m
M)b
R3
R4
CDK2A
HCT-116
MCF-7
30w
30x
30y
CN
CN
CN
CN
H
H
Me
Me
Me
azepane
azepane
piperazine
morpholine
morpholine
4-acetylpiperazin-1-yl
morpholine
4-acetylpiperazin-1-yl
4-acetylpiperazin-1-yl
6
25
6
44
52
45
16
831
2074
55
248
1369
7
2
8
3
18
11
6
12
32
1.54 0.23
4.05 0.49
0.08 0.01
0.49 0.07
0.34 0.08
0.20 0.07
0.02 0.001
0.70 0.04
0.78 0.15
4.13 0.81
7.04 0.64
0.61 0.01
0.95 0.01
0.38 0.04
0.42 0.07
0.09 0.001
2.40 0.57
2.00 0.35
30z
5
30aa
30 ab
30ac
30ad
30ae
morpholine
4-acetylpiperazin-1-yl
morpholine
4-acetylpiperazin-1-yl
4-acetylpiperazin-1-yl
47
187
9
21
43
H
CN
CN
Me
a
The data given are mean values derived from two replicates of one experiment. Apparent inhibition constants (Ki) were calculated from IC50 values and the appropriate Km
(ATP) values for each kinase.
b
Anti-proliferative activity by MTT-48 h assay; the data given are mean values derived from at least three independent replicates.
SI represents selectivity index (Ki, CDK2/Ki, CDK9).
c
6

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
Fig. 6. Activation of caspase-3 in A2780 cancer cell line after treatment with 30m for
24 h. Vertical bars represent the mean SD of two independent experiments, internal
repeats n ¼ 3. Values significantly (P ꢂ 0.05) different from DMSO diluent are marked
with an asterisk (*).
Fig. 4. Diversity kinase panel screening of 30m. The assay was carried out in duplicate.
Percentage of remaining kinase activity is shown.
best selectivity among this series. However, di-substituted analogs
30w-30ae did not further improve the selectivity.
2.3. Diversity kinase panel screening
As 30h, 30k and 30m showed good selectivity for CDK9 over
CDK2, they were selected to test against CDK1 and CDK7. All
showed approximately 100-fold selectivity for CDK9 over CDK1.
30h and 30m also showed a selectivity against CDK7 (>50-fold),
although only 10-fold selectivity was achieved by 30k. Compound
30m was further tested against a diversity panel of 60 kinases at a
concentration of 20 x CDK9 Ki (0.2 mM, Fig. 4). 30m showed a good
selectivity profile, only 5 other kinases (Aurora-A, MLK1, NEK2, KDR
and Fyn) being potently inhibited.
Fig. 7. Cell cycle effects of 30m. A2780 cells were treated with 30m for 24 h, followed
by propidium iodide (PI) staining and analyzed by flow cytometry. Representative of
three independent experiments is shown.
2.4. Biological evaluation CDK9 inhibitors in cancer cells
30m was further evaluated for its mechanism of action in
tumour cell lines and CLL patient-derived cells.
from 0.64 to 2.10 mM (Table 4). 30m was 10e20 fold less potent than
flavopiridol, likely due to its enhanced selectivity for CDK9
compared to flavopiridol.
2.4.1. 30m showed broad anti-proliferative activities
30m was found to be a potent anti-proliferative agent against a
panel of human cancer cell lines, including cancer of the colon,
breast, lung, ovary, cervix, and pancreas, with GI50 values ranging
2.4.2. 30m induced apoptosis of cancer cells
Induction of apoptosis in the A2780 ovarian cancer cell line 24 h
post treatment with 30m was confirmed by Annexin V/PI assay
(Fig. 5). 30m increased the percentage of apoptotic populations
following treatment with 1
mM 30m (18%) and reaching to 26% and
67% at 2
cells.
mM and 5 mM, respectively, when compared to untreated
Caspase-3 is the effector caspase responsible for proteolytic
cleavage of proteins involved in cell cycle regulation and DNA
repair. Cell death induced by anti-cancer agents can occur through
Table 3
Structures and biological activities.
Compd
Kinase inhibition Ki (nM)a
CDK9T1
CDK2A
CDK1B
CDK7H
30h
30k
30m
8
8
10
711
1244
1248
1120
785
1170
472
89
546
a
Fig. 5. Induction of apoptosis by 30m. A2780 cells were treated with 30m for 24 h at
the concentrations shown and analyzed by annexin V/PI staining. Representative of
three independent experiments is shown.
The data given are mean values derived from two replicates. Apparent inhibition
constants (Ki) were calculated from IC50 values and the appropriate Km (ATP) values
for each kinase.
7

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
Table 4
Anti-proliferative activity of 30m.
caspase-dependent or -independent mechanisms. To determine
whether the cytotoxicity induced by 30m is caspase-dependent, a
caspase-3 fluorometric assay was performed. As shown in Fig. 6,
caspase-3 was activated in a dose-dependent manner with the
Tissue Origin
Cell Line
GI50 (mM) S.D. (48h MTT)
Flavopiridol
30m
highest activity at 5 mM compared to untreated controls.
Colon
Breast
HCT-116
HT-29
MCF-7
MDA-MB-468
A549
A2780
0.056 0.013
0.131 0.013
0.092 0.005
0.096 0.006
0.145 0.006
0.064 0.001
0.043 0.001
0.078 0.018
1.07 0.04
2.10 0.45
1.34 0.28
1.46 0.58
1.54 0.57
0.64 0.06
0.80 0.15
1.07 0.13
2.4.3. 30m induced cell arrest in G2-M phase
The effect of 30m on cell cycle progression was investigated in
A2780 cells. Cells were treated with increasing concentrations of
30m for 24 h (Fig. 7). Accumulation of cells in the G2/M phase was
Lung
Ovary
Cervix
Pancreas
HeLa
Miacapa-2
detected at 1 and 2
in pre-G1 phase was observed, indicating cell death. The treatment
with 5 M of 30m resulted in 34% pre-G1 cells, which was
mM of 30m. A dose-dependent increase of cells
m
consistent with the annexin V/PI staining assay.
Table 5
Cytotoxicity of 30k and 30m in CLL patient derived cells.
Inhibition of CDK1B and CDK2A can arrest the G2/M cell cycle.
However, it is unlikely that the cell cycle effect is due to inhibition
of CDK1B and 2A, as 30m showed low activity against both kinases
(Ki value of 1,248 and 1,170 nM, respectively, Table 3). Inhibition of
CDK9 has been shown to block the G2/M cell cycle; for example,
flavopiridol caused a shift from G0/G1 to G2/M phase in multiple
cancer cells[46]. Another possible reason is the off-target effect,
since 30m inhibited aurora kinases A (Fig. 4), which has have a
known role in regulation of the G2 to mitosis.
aLD50
CLL B cells (48h, LD50, mM)
, mM
Compd
Comp
Patient A
Patient B
Patient C
Patient D
Patient E
30k
30m
1.4
6.2
0.83
2.3
18
82
0.81
1.5
0.94
2.9
a
Concentration required to kill 50% of CLL cells following treatment for 48 h.
tested (Table 4), regardless of its p53 status, so wildtype p53
function is not required for 30m to act. As in ovarian cancer cells,
30m demonstrated a similar mode of action in HCT-116 colorectal
cancer cells (data not shown). We also demonstrated a similar
mechanism of action to 30 m of 30k in the HCT-116 cell line
(Supplemental data).
2.4.4. 30m inhibited phosphorylation of RNAPII CTD and down-
regulated expression of anti-apoptotic proteins
CDK9 regulates mRNA transcription by phosphorylation of the
RNAPII at Ser-2 and Ser-5. Western blotting analysis of A2780
cancer cells 24 h-post treatment with 30m showed that the level of
phosphorylated Ser-2 and Ser-5 was reduced by 30m from ꢃ0.5
mM
in a concentration-dependent manner, confirming cellular CDK9
inhibition of 30m (Fig. 8).
2.4.5. Ex vivo antitumor activity in primary CLL cells
30m, as well as 30k, were further evaluated in patient-derived
CLL cells using an annexin V-FITC apoptosis assay and the LD50
values are shown in Table 5. The potency of the two compounds
seemed to be patient-dependent, since each patient had different
clinical characteristics (Supplemental Table 2). 30k was potent
against most of patient-derived CLL cells except patient C with LD50
Up-regulation of anti-apoptotic proteins, such as Mcl-1, Bcl-2
and XIAP, has been considered as the key mechanism of resistance
to cancer therapies. Transcriptional inhibition of expression of Mcl-
1 and XIAP is a general feature of CDK9 inhibitors. Consistently,
treatment of A2780 cancer cells with 30m for 6 or 12 h resulted in
the down-regulation of Mcl-1 at 5
Fig. 8). The extended treatment for 24 h reduced the level of Mcl-1
and XIAP starting from 0.5 M in a dose-dependent manner (right
mM concentration (left panel,
values ranging from 0.81 to 1.4
against the cells derived from patients A-B and D-E with
M. The cells derived from patient C were found to
mM. Similarly, 30m was potent
m
LD50 ¼ 1.5e6.2
m
panel, Fig. 8), which was accompanied by PARP cleavage, con-
firming apoptosis. 30m increased p53 levels and activated its
transcriptional target gene, as evidenced by p21 induction. How-
ever, there was little difference in the GI50 values in the cell lines
be resistant to 30k and 30m. Patient C was in the early stage of CLL,
and it is not clear why this disease was insensitive to both 30k and
30m. The cellular mode of action of 30k and 30m was confirmed in
patient B derived cells where reduction of CDK9-mediated Mcl-1
expression leading to apoptosis was observed (Supplemental
Fig. 5.).
3. Conclusion
We have described the rational design, synthesis and biological
evaluation of selective CDK9 inhibitors. Although the gatekeeper
residue is conserved among CDKs, we found that this region can be
exploited to improve the selectivity for CDK9 over other members
of CDK family. The selectivity for CDK9 over CDK2 and other kinases
was improved greatly compared with the lead compound, evi-
denced by 30k and 30m. Our study provides important strategies
for designing highly selective CDK9 inhibitors, which could also be
applied to other kinase inhibitors as well.
Non-selective CDK9 inhibitors showed more potent cellular
activity. As we improved the selectivity profile of the molecules,
potency against CDK9 remained relatively similar. This suggests
that our molecules have reduced off-target liabilities that are
contributing to the cellular effects on non-selective compounds.
Despite this, our lead molecules 30k and 30m retained potent anti-
Fig. 8. Cellular mode of action of 30m. A2780 cells were treated with 30m and
analyzed by Western blot. DMSO diluent was used as control in each experiment and
b
-actin antibody was used as an internal loading control.
8

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
proliferative activity and a cellular mode of action consistent with a
CDK9-targeted mechanism.
430.1675.
Since pan-selective CDKs inhibitors have not achieved clinical
success, we suggest that this profile represents a significant
improvement and that highly selective CDK9 inhibitors have po-
tential to be developed as anti-cancer agents. This hypothesis must
be further tested in preclinical study and clinical development.
4.1.2. 5-(5-Chloro-2-(3-morpholinophenylamino)pyrimidin-4-yl)-
N,4-dimethylthiazol-2-amine (30d)
From 1-(3-morpholinophenyl)guanidine and 2-chloro-4-
methyl-1-(4-methyl-2-(methylamino)thiazol-5-yl)pent-2-en-1-
one. Yellow solid (37%), mp 230e232 ꢀC. Anal. RP-HPLC: tR
14.27 min (method A), 11.09 min (method D), purity 99%. 1H NMR
4. Experimental section
(400 MHz, DMSO‑d6)
d
: 2.33 (3H, s), 2.87 (3H, d, J ¼ 4.8 Hz), 3.08
(4H, apparent t, J ¼ 4.8 Hz), 3.74 (4H, apparent t, J ¼ 4.8 Hz), 6.58
(1H, dd, J ¼ 8.0 & 2.0 Hz), 7.13 (1H, apparent t, J ¼ 8.0 Hz), 7.21 (1H,
d, J ¼ 8.0 Hz), 7.38 (1H, apparent t, J ¼ 1.6 Hz), 7.97 (1H, q,
J ¼ 4.8 Hz), 8.47 (1H, s), 9.55 (1H, s). 13C NMR (100 MHz, DMSO‑d6):
4.1. Chemistry
Chemical reagents and solvents were obtained from commercial
sources. When necessary, solvents were dried and/or purified by
standard methods. 1H NMR and 13C NMR spectra were obtained
using a Bruker 400, 500 or 600 Ultrashield™ spectrometer at 400,
500, 600 MHz and 100, 125, 150 MHz respectively. These were
analyzed using the Bruker TOPSPIN 2.1 program. Chemical shifts
are reported in parts per million relative to the solvent chemical
shift. Coupling constants (J) are quoted to the nearest 0.1 Hz. The
following abbreviations (or combinations thereof) are used: s,
singlet; d, doublet; t, triplet; q, quartet; m, multiplet and br, broad.
High resolution mass spectra were obtained using a Waters 2795
single quadrupole mass spectrometer/micromass LCT platform.
Purity for final compounds was greater than 95% and was measured
using a Waters high performance liquid chromatography (Waters
d
19.3, 31.3, 49.1, 66.6, 106.7, 109.6, 111.1, 112.6, 116.3, 129.3, 141.3,
151.9, 153.5, 156.7, 158.1, 158.5, 170.7. HR-MS (ESIþ): m/z [M þ H]þ
calcd for C19H22ClN6OS, 417.1259; found 417.1177.
4.1.3. N,4-Dimethyl-5-(5-methyl-2-(3-morpholinophenylamino)
pyrimidin-4-yl)thiazol-2-amine (30e)
From 1-(3-morpholinophenyl)guanidine and 3-(dimethyla-
mino)-2-methyl-1-(4-methyl-2-(methylamino)thiazol-5-yl)prop-
2-en-1-one. Cream (32%), mp 214e216 ꢀC. Anal. RP-HPLC: tR
13.19 min (method A), 10.67 min (method D, purity 99%). 1H NMR
(400 MHz, DMSO‑d6):
d 2.15 (3H, s), 2.17 (3H, s), 2.85 (3H, d,
J ¼ 4.8 Hz), 3.08 (4H, appaent t, J ¼ 4.8 Hz), 3.74 (4H, apparent t,
J ¼ 4.8 Hz), 6.51 (1H, dd, J ¼ 8.0 & 2.0 Hz), 7.09 (1H, apparent t,
J ¼ 8.0 Hz), 7.19 (1H, dd, J ¼ 8.0 & 1.2 Hz), 7.40 (1H, apparent t,
2487 dual
l absorbance detector) with a Phenomenex Gemini-NX
5u C18 110A 250 ꢄ 4.60 mm column, UV detector at 254 nm, us-
ing system A: 10% MeOH containing 0.1% TFA for 4 min, followed by
linear gradient 10e100% MeOH over 10 min at a flow rate of 1 mL/
min; B: 10% MeOH containing 0.1% TFA for 4 min, followed by linear
gradient 10e100% MeOH over 6 min at a flow rate of 1 mL/min;
system C: 10% MeCN containing 0.1% TFA for 2 min, followed by
linear gradient 10e100% over 10 min at a flow rate of 1 mL/min; and
system D: 10% MeCN containing 0.1% TFA for 4 min, followed by
linear gradient 10e100% over 10 min at a flow rate of 1 mL/min.
Flash chromatography was performed using either glass column
packed with silica gel (200e400 mesh, Aldrich Chemical) or pre-
packed silica gel cartridges (FlashMaster systems, Biotage).
Melting points (m.p.) were determined with an Electrothermal
melting point apparatus.
J ¼ 2.0 Hz), 7.75 (1H, q, J ¼ 4.8 Hz, NH), 8.34 (1H,s), 9.28 (1H, s). 13
C
NMR (100 MHz, DMSO‑d6): d 16.4, 18.1, 31.3, 49.3, 66.6, 106.0, 108.7,
110.4,115.3,118.3,129.2,142.1, 149.4,151.9,158.5,158.6,160.4,169.4.
HRMS (ESIþ): m/z [M þ H]þ calcd for C20H25N6OS, 397.1805; found
397.1625.
4.1.4. 1-(4-(3-(5-Methyl-4-(2-(methylamino)thiazol-5-yl)
pyrimidin-2-ylamino)phenyl)piperazin-1-yl)ethanone (30f)
From 1-(4-acetylpiperazin-1-yl)phenyl)guanidine and 3-
(dimethylamino)-2-methyl-1-(4-methyl-2-(methylamino)thiazol-
5-yl)prop-2-en-1-one. Light yellow powder (20%), mp 149e151 ꢀC.
Anal. RP-HPLC: tR 13.27 min (method A), 10.65 min (method D),
purity 95%. 1H NMR (400 MHz, DMSO‑d6):
d 2.05 (3H, s), 2.16 (3H, s),
2.18 (3H, s), 2.86 (3H, d, J ¼ 4.8 Hz), 3.07 (2H, tm, J ¼ 4.8 Hz), 3.14
(2H, tm, J ¼ 4.8 Hz), 3.52e3.63 (4H, m), 6.53 (1H, dd, J ¼ 8.0 &
2.0 Hz), 7.10 (1H, apparent t, J ¼ 8.0 Hz), 7.22 (1H, d, J ¼ 8.0 Hz), 7.55
(1H, t, J ¼ 1.6 Hz), 7.82 (1H, q, J ¼ 4.8 Hz), 8.34 (1H, s), 9.30 (1H, s).
Detailed procedures for the synthesis of all intermediates are
included in the Supplementary Information.
General procedures for preparations of compounds 30c-30ad.
A
mixture of the appropriate enaminones and 1-
13C NMR (100 MHz, DMSO‑d6):
d 16.4, 18.0, 21.8, 31.3, 41.2, 45.9,
phenylguanidine (2 equivalent mmol) in 2-methoxyethanol
(3 mL) was heated at 100e140 ꢀC for 20e45 min in a Discovery
Microwave. Upon cooling, the residue was purified by flash chro-
matography using appropriate mixtures of EtOAc/PE or EtOAc/
MeOH as the eluant.
49.1, 49.5, 106.8, 109.5, 110.6, 115.3, 118.4, 129.3, 142.1, 149.1, 151.7,
158.45, 158.54, 160.5, 168.7, 169.4. HRMS (ESIþ): m/z [M þ H]þ calcd
for C22H28N7OS, 438.2071, found 438.2109.
4.1.5. 1-(4-(3-((2-Amino-4,5-dihydrothiazolo[4,5-h]quinazolin-8-
4.1.1. 5-(5-Chloro-2-((3-(4-methylpiperazin-1-yl)phenyl)amino)
yl)amino)phenyl)piperazin-1-yl)ethanone (30g)
pyrimidin-4-yl)-N,4-dimethyl thiazol-2-amine (30c)
From
N’-(6-((dimethylamino)methylene)-7-oxo-4,5,6,7-
dimethylformimidamide
From
2-chloro-3-(dimethylamino)-1-(4-methyl-2-(methyl-
and 1-(3-(4-
tetrahydrobenzo[d]thiazol-2-yl)-N,N-
amino)thiazol-5-yl)prop-2-en-1-one
and 1-(4-acetylpiperazin-1-yl)phenyl)guanidine. Brown solid (7%),
methylpiperazin-1-yl)phenyl)guanidine. Yellow solid (7%), mp
mp 291e293 ꢀC. Anal. RP-HPLC: tR 10.67 min (method D), purity
120e122 ꢀC. Anal. RP-HPLC: tR 13.17 min (method A), 10.82 min
95%. 1H NMR (400 MHz, DMSO‑d6):
d 2.06 (3H, s), 2.71e2.88 (4H,
(method D), purity 96%. 1H NMR (400 MHz, DMSO‑d6):
d
2.25 (3H,
m), 3.09 (2H, apparent t, J ¼ 4.8 Hz), 3.16 (2H, apparent t, J ¼ 4.8 Hz),
3.53e3.68 (4H, m), 6.51 (1H, dd, J ¼ 8.0 & 2.0 Hz), 7.08 (1H,
apparent t, J ¼ 8.0 Hz), 7.18 (1H, dd, J ¼ 8.0 & 0.8 Hz), 7.60 (1H, t,
J ¼ 2.0 Hz), 7.78 (2H, s), 8.07 (1H, s), 9.14 (1H, s). 13C NMR (125 MHz,
s), 2.34 (3H, s), 2.45e2.49 (4H, m), 2.86 (3H, d, J ¼ 4.8 Hz), 3.13 (4H,
apparent t, J ¼ 4.8 Hz), 6.57 (1H, dd, J ¼ 8.0 & 1.6 Hz), 7.11 (1H,
apparent t, J ¼ 8.0 Hz), 7.99 (1H, d, J ¼ 8.0 Hz), 7.36 (1H, t, J ¼ 2.0 Hz),
7.98 (1H, q, J ¼ 4.8 Hz), 8.47 (1H, s), 9.52 (1H, s). 13C NMR (100 MHz,
DMSO‑d6): d 21.7, 23.5, 25.8, 41.2, 46.0, 49.1, 49.5, 106.6, 109.2, 110.5,
DMSO‑d6):
d
19.3, 31.3, 46.0, 48.6, 54.9, 107.0, 110.0, 110.8, 112.6,
115.1, 115.4, 129.2, 142.3, 151.6, 153.6, 157.6, 159.4, 159.8, 168.8, 171.9.
HRMS (ESIþ): m/z [M þ H]þ calcd for C21H24N7OS, 422.1758, found
422.1685.
116.3, 129.3, 141.2, 151.8, 153.6, 156.7, 158.1, 158.5, 170.2. HR-MS
(ESIþ): m/z [M þ H]þ calcd for C20H25ClN7S, 430.1575; found
9

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
4.1.6. N,4-Dimethyl-5-(5-methyl-2-((3-(piperazin-1-yl)phenyl)
amino)pyrimidin-4-yl)thiazol-2-amine (30h)
1H). 13C NMR (100 MHz, DMSO‑d6):
d 16.4,18.0, 29.0, 31.3, 47.5, 47.8,
48.2, 51.8, 102.3, 105.2, 106.7, 115.2, 118.2, 129.4, 142.3, 149.0, 149.2,
A mixture of compound 30f 1-(4-(3-((5-methyl-4-(4-methyl-2-
(methylamino)thiazol-5-yl)pyrimidin-2-yl)amino)phenyl)piper-
azin-1-yl)ethanone in ethanol (5 mL) and 2 M HCl (4 mL) was
heated under reflux for 3 h. After completion of the reaction, the
mixture was neutralized by NaOH solution, concentrated and pu-
rified by column chromatography using CHCl3/MeOH as the eluant
to get the final product as cream (54%). Mp 131e133 ꢀC. Anal. RP-
HPLC: tR 10.90 min (method A), 8.44 min (method C), 10.47 min
158.6, 158.7, 160.3, 169.3. HRMS (ESIþ): m/z [M þ H]þ calcd for
C
21H28N7S, 410.2121, found 410.2049.
4.1.10. 1-(3-((5-Methyl-4-(4-methyl-2-(methylamino)thiazol-5-yl)
pyrimidin-2-yl)amino)phenyl)-1,4-diazepan-5-one (30l)
From
3-(dimethylamino)-2-methyl-1-(4-methyl-2-(methyl-
and 1-(3-(5-oxo-1,4-
amino)thiazol-5-yl)prop-2-en-1-one
diazepan-1-yl)phenyl)guanidine. Pale yellow solid (8 mg; 3%).
Melting point: 208e209 ꢀC. Anal. RP-HPLC: tR 10.87 min (method
B), purity 99%. Rf (9:1 CHCl3:MeOH): 0.39. nmax (KBr disk, cmꢁ1):
687.8, 752.0, 786.0, 847.5, 997.2, 1072.4, 1122.2, 1170.4, 1212.6,
1251.1, 1313.1, 1389.0, 1408.3, 1452.2, 1490.2, 1521.9, 1566.8, 1652.6,
(method D), purity 99%. 1H NMR (400 MHz, DMSO‑d6):
d 2.16 (3H,
s), 2.17 (3H, s), 2.85 (3H, d, J ¼ 4.8 Hz), 2.88 (4H, apparent t,
J ¼ 4.8 Hz), 3.06 (4H, apparent t, J ¼ 4.8 Hz), 6.49 (1H, dd, J ¼ 8.0 &
1.6 Hz), 7.07 (1H, apparent t, J ¼ 8.0 Hz), 7.18 (1H, d, J ¼ 8.0 Hz), 7.51
(1H, apparent t, J ¼ 1.6 Hz), 7.75 (1H, q, J ¼ 4.8 Hz), 8.33 (1H, s), 9.25
2918.7, 3206.4, 3297.1. 1H NMR (400 MHz, DMSO‑d6):
d 2.15 (3H, s),
(1H, s). 13C NMR (100 MHz, DMSO‑d6):
d
16.4, 18.1, 31.3, 45.7, 49.6,
2.17 (3H, s), 2.84 (3H, d, J ¼ 4.6 Hz), 3.12e3.23 (2H, m), 3.40e3.47
(2H, m), 3.47e3.53 (2H, m), 6.44 (1H, dd, J ¼ 8.2 & 2.0 Hz), 7.06 (1H,
apparent t, J ¼ 8.2 Hz), 7.12e7.20 (1H, m), 7.39 (1H, m), 7.59 (1H,
apparent t, J ¼ 5.3 Hz), 7.74 (1H, q, J ¼ 4.7 Hz), 8.32 (1H, s), 9.21 (1H,
br s).* dH (400 MHz; CDCl3): 2.22 (3H, s), 2.54 (3H, s), 2.73e2.80
(2H, m), 3.01 (3H, br s), 3.34e3.44 (2H, m), 3.55e3.60 (2H, m),
3.60e3.67 (2H, m), 5.45 (1H, br s), 5.98 (1H, br s), 6.50 (1H, dd,
J ¼ 8.2 & 2.0 Hz), 6.87 (1H, dd, J ¼ 8.0 & 1.4 Hz), 7.05 (1H, s), 7.19 (1H,
t, J ¼ 8.2 Hz), 7.54 (1H, t, J ¼ 2.1 Hz), 8.27 (1H, s). 13C NMR (100 MHz,
106.33, 109.0, 110.1, 115.3, 118.3, 129.2, 142.0, 149.4, 152.4, 158.5,
158.6, 160.4, 169.4. HRMS (ESIþ): m/z [M þ H]þ calcd for C20H26N7S,
396.1965, found 396.2151.
4.1.7. N,4-Dimethyl-5-(5-methyl-2-((3-(4-methylpiperazin-1-yl)
phenyl)amino)pyrimidin-4-yl) thiazol-2-amine (30i)
From
amino)
3-(dimethylamino)-2-methyl-1-(4-methyl-2-(methyl-
thiazol-5-yl)prop-2-en-1-one and 1-(3-(4-
methylpiperazin-1-yl)phenyl)guanidine. Yellow sold (20%), mp
DMSO‑d6): d 15.6,17.5, 30.7, 36.8, 40.4, 45.3, 52.3,105.2,108.0,108.5,
139e141 ꢀC. Anal. RP-HPLC: tR 10.52 min (method D), purity 100%.
114.6, 117.8, 129.0, 141.7, 148.8, 149.0, 157.99, 158.01, 159.8, 168.8,
175.5. HRMS (ESIþ): m/z [M þ H]þ calcd for C21H26N7OS 424.1914,
found 424.1963. HPLC RT (Method B, ACN): 10.80 (96%). *A CH2
signal is present under the water peak, as determined by 2D ex-
periments (COSY, HSQC).
1H NMR (400 MHz, DMSO‑d6):
d 2.16 (3H, s), 2.18 (3H, s), 2.23 (3H,
s), 2.46 (4H, apparent t, J ¼ 4.8 Hz), 2.85 (3H, d, J ¼ 4.8 Hz), 3.11 (4H,
apparent t, J ¼ 4.8 Hz), 6.50 (1H, dd, J ¼ 8.0 & 2.0 Hz), 7.07 (1H,
apparent t, J ¼ 8.0 Hz), 7.19 (1H, d, J ¼ 8.0 & 1.2 Hz), 7.49 (1H,
apparent t, J ¼ 2.0 Hz), 7.75 (1H, q, J ¼ 4.8 Hz), 8.33 (1H, s), 9.24 (1H,
s). 13C NMR (100 MHz, DMSO‑d6):
d
16.4, 18.1, 31.3, 46.2, 48.8, 55.1,
4.1.11. 5-(2-((3-(1,4-Diazabicyclo[3.2.2]nonan-4-yl)phenyl)amino)-
5-methylpyrimidin-4-yl)- N,4-dimethylthiazol-2-amine (30m)
A suspension of 3-(1,4-diazabicyclo[3.2.2]nonan-4-yl)aniline
(2.5 g; 11.5 mmol) in 20 mL acetonitrile had cyanamide (800 mg;
19.0 mmol) and TMSCl (2.9 mL; 23.0 mmol) added in sequence and
the resultant suspension was heated to reflux for 18 h, whereupon
MS showed complete reaction. The ACN was decanted leaving a
crude brown guanidine which was dried under vacuum (3.5 g; 92%
yield as di-hydrochloride salt). 2.4 g of this material was re-
dissolved in 20 mL 2-methoxyethanol and the pH adjusted with
excess K2CO3 to pH 9 (Suspension A). A solution of tert-butyl N-
methyl-N-(4-methyl-5-propanoyl-1,3-thiazol-2-yl)carbamate
(4.0 g; 14.1 mmol) and DMF-DMA (2.4 mL; 17.5 mmol) in 12 mL ACN
was heated under microwave conditions to 140 ꢀC for 45 min,
whereupon MS indicated near complete reaction. The reaction
mixture was concentrated to dryness, azeotrophed with toluene
and isopropanol, then re-dissolved in 20 mL 2-methoxyethanol.
This solution was added to Suspension A and conventionally heated
to reflux overnight. The reaction mixture was absorbed onto silica
and purified by column chromatography, using first 9:1
CHCl3:MeOH / 6:4 CHCl3:MeOH (to give 1.5 g of ~90% pure
product), which was further columned in 4:1 / 6:4 DCM:MeOH.
After concentrating to dryness residue was suspended in 200 mL
refluxing ether and filtered to give the final product as a light or-
ange solid (243 mg). A further 335 mg of product was isolated by
addition of PE to the filtrate (17% yield based on starting aniline).
Decomposes at 220 ꢀC. Anal. RP-HPLC: tR 10.92 min (method B),
10.73 min (method D), purity 97%. 1H NMR (400 MHz, MeOH-d4):
106.3,109.0,110.1,115.2,118.3,129.2,142.0,149.4,151.9,158.6,158.6,
160.3, 169.4. HRMS (ESIþ): m/z [M þ H]þ calcd for C21H28N7S,
410.2121; found 410.0873.
4.1.8. N,4-Dimethyl-5-(5-methyl-2-((3-(4-(methylsulfonyl)
piperazin-1-yl)phenyl)amino) pyrimidin-4-yl)thiazol-2-amine (30j)
From
3-(dimethylamino)-2-methyl-1-(4-methyl-2-(methyl-
amino)thiazol-5-yl)prop-2-en-1-one and 1-(3-(4-(methylsulfonyl)
piperazin-1-yl)phenyl)guanidine. Cream (12%), mp 266e268 ꢀC.
Anal. RP-HPLC: tR 13.57 min (method A), 11.22 min (method D),
purity 100%. 1H NMR (400 MHz, DMSO‑d6):
d 2.16 (3H, s), 2.17 (3H,
s), 2.85 (3H, d, J ¼ 4.8 Hz), 2.93 (3H, s), 3.18e3.28 (8H, m), 6.55 (1H,
dd, J ¼ 8.0 & 2.0 Hz), 7.11 (1H, apparent t, J ¼ 8.0 Hz), 7.24 (1H, dd,
J ¼ 8.0 & 1.2 Hz), 7.56 (1H, apparent t, J ¼ 2.0 Hz), 7.77 (1H, q,
J ¼ 4.8 Hz), 8.34 (1H, s), 9.30 (1H, s). 13C NMR (100 MHz, DMSO‑d6):
d
16.4, 18.1, 31.3, 34.3, 45.8, 48.9, 107.0, 109.6, 110.8, 115.2, 118.4,
129.3, 142.1, 149.4, 151.3, 158.5, 160.4, 169.4. HRMS (ESIþ): m/z [M þ
H]þ calcd for C21H28N7O2S2, 474.1740; found 474.1577.
4.1.9. 5-(2-((3-(1,4-Diazepan-1-yl)phenyl)amino)-5-
methylpyrimidin-4-yl)-N,4-dimethylthiazol-2-amine (30k)
From 1-(4-(3-((4-(4-methyl-2-(methylamino)thiazol-5-yl)pyr-
imidin-2-yl)amino)phenyl)-1,4-diazepan-1-yl)ethanone,
which
was synthesized from 3-(dimethylamino)-2-methyl-1-(4-methyl-
2-(methylamino)thiazol-5-yl)prop-2-en-1-one and 1-(3-(4-acetyl-
1,4-diazepan-1-yl)phenyl)guanidine. The crude product was used
in next step without further purification. Method as described for
30h. Cream (65%), mp 121e123 ꢀC. Anal. RP-HPLC: tR 10.87 min
d
1.54e1.73 (2H, m),1.93e2.07 (2H, m), 2.14 (3H, s), 2.15 (3H, s), 2.84
(method A), purity 99%. 1H NMR (400 MHz, MeOH-d4):
d
2.01e2.10
(3H, d, J ¼ 4.7 Hz), 2.85e3.00 (6H, m), 3.42e3.51 (2H, m), 3.88 (1H,
br s), 6.34 (1H, apparent dt, J ¼ 7.3 & 2.0 Hz), 7.34 (1H, br s), 7.72
(1H, q, J ¼ 7.5 Hz), 8.31 (1H, s), 9.15 (1H, s). 13C NMR (100 MHz,
(m, 2H), 2.22 (3H, s), 2.23 (3H, s), 2.86 (apparent t, 2H, J ¼ 5.6 Hz),
2.97 (s, 3H), 3.04 (apparent t, 2H, J ¼ 5.2 Hz), 3.59e3.66 (4H, m),
6.41 (dd, 1H, J ¼ 8.0 & 2.0 Hz), 6.85 (dd, 1H, J ¼ 8.0 & 2.0 Hz), 7.10
(apparent t, 1H, J ¼ 8.0 Hz), 7.40 (apparent t, 1H, J ¼ 2.4 Hz), 8.28 (s,
MeOH-d4): d 16.3, 17.2, 31.6, 43.3, 47.5, 52.5, 57.4, 106.6, 109.2, 110.3,
116.5, 120.2, 130.3, 142.8, 149.6, 151.0, 159.9, 160.0, 161.0, 172.1.
10

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
HRMS (ESIþ): m/z [M þ H]þ calcd for C23H30N7S 436.2278, found
4.1.16. 5-(2-((3-(1,4-Diazepan-1-yl)phenyl)amino)pyrimidin-4-yl)-
N-methyl-4-(trifluoromethyl) thiazol-2-amine (30r)
436.2242.
From 1-(4-(3-((4-(2-(methylamino)-4-(trifluoromethyl)thiazol-
5-yl)pyrimidin-2-yl)amino)phenyl)-1,4-diazepan-1-yl)ethanone.
Method as described from 45. Yellow solid (64%), mp 194e196 ꢀC.
Anal. RP-HPLC: tR 12.10 min (method B), 11.75 (method D), purity
4.1.12. 1-(4-(3-((4-(4-Methyl-2-(methylamino)thiazol-5-yl)-5-
(trifluoromethyl)pyrimidin-2-yl) amino)phenyl)piperazin-1-yl)
ethanone (30n)
From tert-butyl (5-(3 (dimethylamino)-2-(trifluoromethyl)
100%. 1H NMR (400 MHz, DMSO- d6):
d 1.85-1.72 (2H, m), 2.62 (2H,
acryloyl)-4-methylthiazol-2-yl)(methyl)carbamate
and
1-(3-
apparent t, J ¼ 5.6 Hz), 2.84 (2H, apparent t, J ¼ 5.2 Hz), 2.89 (3H, s),
3.46 (2H, apparent t, J ¼ 5.2 Hz), 3.54 (2H, apparent t, J ¼ 6.0 Hz),
6.29e6.41 (1H, m), 6.92 (1H, d, J ¼ 5.2 Hz), 7.10-6.98 (3H, m), 8.46
(1H, br s), 8.48 (1H, d, J ¼ 5.2 Hz), 9.42 (1H, s). 13C NMR (100 MHz,
acetylpiperazin-1-yl)phenyl)guanidine. Pale yellow solid (35%),
mp 211e213 ꢀC. Anal. RP-HPLC: tR 14.50 min (method A), 11.37 min
(method D), purity 99%. 1H NMR (400 MHz, DMSO‑d6):
d 2.05 (3H,
s), 2.16 (3H, s), 2.85 (3H, d, J ¼ 4.8 Hz), 3.08 (2H, apparent t,
J ¼ 4.8 Hz), 3.14 (2H, apparent t, J ¼ 4.8 Hz), 3.54e3.64 (4H, m), 6.66
(1H, d, J ¼ 8.0 Hz), 7.17 (1H, apparent t, J ¼ 8.0 Hz), 7.22 (1H, d,
J ¼ 8.0 Hz), 7.45 (1H, s), 7.85 (1H, q, J ¼ 4.8 Hz), 8.78 (1H, s), 10.11
DMSO- d6): d 29.5, 31.4, 47.8, 47.9, 48.5, 49.1, 52.7, 102.7, 106.0, 107.0,
108.7, 121.3 (q, J ¼ 270 Hz), 126.0, 129.5, 137.3 (q, J ¼ 34 Hz), 141.5,
149.1, 156.0, 159.4, 160.2, 170.0. HRMS (ESIþ): m/z [M þ H]þ calcd for
C
20H23F3N7S, 450.1682, found 450.1458. 1-(4-(3-((4-(2-(Methyl-
(1H, s). 13C NMR (100 MHz, DMSO‑d6):
d
17.4, 21.7, 31.3, 41.1, 45.9,
amino)-4-(trifluoromethyl)thiazol-5-yl)pyrimidin-2-yl)amino)
phenyl)-1,4-diazepan-1-yl)ethanone. From tert-butyl methyl(5-(4-
methylpent-2-enoyl)-4-(trifluoromethyl)thiazol-2-yl)carbamate
and 1-(3-(4-acetyl-1,4-diazepan-1-yl)phenyl)guanidine. Yellow
48.9, 49.3, 108.2, 111.2, 111.9, 112.5 (q, J ¼ 25 Hz), 124.8 (q,
J ¼ 216 Hz), 129.5, 140.4, 150.6, 151.6, 158.1, 158.7, 160.9, 168.7, 169.7.
HRMS (ESIþ): m/z [M þ H]þ calcd for C22H25 F3N7OS, 492.1788;
found 492.1743.
solid (53%). 1H NMR (400 MHz, DMSO- d6):
d 1.92-1.77 (3.5H, m),
1.99 (1.5H, s), 2.89 (3H, d, J ¼ 3.6 Hz), 3.38-3.27 (2H, m), 3.60-3.46
(6H, m), 6.43-6.36 (1H, m), 6.92 (1H, d, J ¼ 4.8 Hz), 7.18-7.02 (3H, m),
8.46 (1H, br s), 8.49 (d, 1H, J ¼ 5.2 Hz), 9.45 (d, 1H, J ¼ 3.6 Hz). HRMS
(ESIþ): m/z [M þ H]þ calcd for C22H25F3N7OS, 492.1788, found
492.1731.
4.1.13. N,4-Dimethyl-5-(2-((3-(piperazin-1-yl)phenyl)amino)-5-
(trifluoromethyl)pyrimidin-4-yl)thiazol-2-amine (30o)
From compound 30n 1-(4-(3-((4-(4-methyl-2-(methylamino)
thiazol-5-yl)-5-(trifluoromethyl)pyrimidin-2-yl)amino)phenyl)
piperazin-1-yl)ethanone. Method as described from 30f. Yellow
solid (68%), mp 207e209 ꢀC. Anal. RP-HPLC: tR min (method A),
10.80 min (method D), purity 99%. 1H NMR (400 MHz, DMSO‑d6):
4.1.17. 5-(2-((3-(1,4-diazepan-1-yl)phenyl)amino)pyrimidin-4-yl)-
N-methylthiazol-2-amine (30s)
d
2.15 (3H,s), 2.85 (3H, d, J ¼ 4.8 Hz), 2.95 (4H, apparent t,
J ¼ 4.8 Hz), 3.12 (4H, apparent t, J ¼ 4.8 Hz), 6.66 (1H, d, J ¼ 8.0 Hz),
7.15 (1H, apparent t, J ¼ 8.0 Hz), 7.20 (1H, d, J ¼ 8.0 Hz), 7.42 (1H, s),
7.86 (1H, q, J ¼ 4.8 Hz), 8.77 (1H, s),10.09 (1H, s). 13C NMR (100 MHz,
From 1-(4-(3-((4-(2-(methylamino)thiazol-5-yl)pyrimidin-2-yl)
amino)phenyl)-1,4-diazepan-1-yl)ethanone (From tert-butyl (5-(3-
(dimethylamino)acryloyl)thiazol-2-yl)(methyl)carbamate and 1-
(3-(4-acetyl-1,4-diazepan-1-yl)phenyl)guanidine. After completion
of the reaction, the product was purified by EtOAc/MeOH to get the
crude product and used in the next reaction directly). Method as
described from 45. Grey solid (62%), mp 139e141 ꢀC. Anal. RP-
HPLC: tR 10.87 min (method B), 10.48 min (method D), purity
DMSO‑d6): d 14.55, 32.28, 44.98, 48.56, 107.95, 110.97, 111.71, 112.48
(q, J ¼ 25 Hz), 124.76 (q, J ¼ 216 Hz), 129.44, 140.40, 150.90, 151.94,
158.02, 158.83, 160.87, 169.74. HRMS (ESIþ): m/z [M þ H]þ calcd for
C
20H23 F3N7 S, 450.1688; found 450.1332.
4.1.14. N,4-Dimethyl-5-(2-((3-morpholinophenyl)amino)-5-
vinylpyrimidin-4-yl)thiazol-2-amine (30p)
100%. 1H NMR (400 MHz, DMSO- d6):
d 1.76e1.87 (2H, m), 2.63 (2H,
apparent t, J ¼ 5.2 Hz), 2.86 (2H, apparent t, J ¼ 5.2 Hz), 2.89 (3H, d,
J ¼ 4.8 Hz), 3.48 (2H, apparent t, J ¼ 5.2 Hz), 3.56 (2H, apparent t,
J ¼ 6.0 Hz), 6.26e6.35 (1H, m), 6.98-7.00-7.13 (2H, m), 7.11 (1H, d,
J ¼ 5.2 Hz), 7.22 (1H, s), 8.03 (1H, s), 8.18 (1H, q, J ¼ 4.8 Hz), 8.27 (1H,
From 5-(5-chloro-2-((3-morpholinophenyl)amino)pyrimidin-4-
yl)-N,4-dimethylthiazol-2-amine. Yellow solid (60%). 1H NMR
(500 MHz; DMSO‑d6):
d
2.09 (3H, s), 2.86 (3H, d, J ¼ 4.8 Hz), 3.11
(4H, apparent t, J ¼ 4.8 Hz), 3.75 (4H, apparent t, J ¼ 4.8 Hz), 5.23
(1H, d, J ¼ 12.0 Hz), 6.51 (1H, d, J ¼ 17.6 Hz), 6.45e6.60 (2H, m), 7.11
(1H, apparent t, J ¼ 8.0 Hz), 7.19 (1H, d, J ¼ 8.0 Hz), 7.60 (1H, s), 7.84
(1H, q, J ¼ 4.8 Hz), 8.76 (1H,s), 9.57 (1H, s). 13C NMR (125 MHz,
d, J ¼ 5.2 Hz), 9.16 (1H, s). 13C NMR (100 MHz, DMSO- d6):
d 29.4,
31.4, 47.7, 47.9, 48.2, 52.3, 102.4, 105.5, 105.6, 106.9, 124.6, 129.5,
142.1, 143.1, 149.1, 157.6, 158.9, 160.4, 172.7. HRMS (ESIþ): m/z [M þ
H]þ calcd for C19H24N7S, 382.1808, found 382.1975.
DMSO- d6):
d 18.4, 31.3, 49.3, 66.6, 106.2, 109.2, 110.7, 113.6, 115.3,
119.6, 129.3, 132.2, 141.6, 150.2, 152.0, 157.2, 157.3, 158.8, 169.9.
HRMS (ESIþ): m/z [M þ H]þ calcd for C21H25N6OS, 409.1805, found
409.1680.
4.1.18. 3-((4-(4-Cyclopropyl-2-(methylamino)thiazol-5-yl)
pyrimidin-2-yl)amino)-N-(1-methyl piperidin-4-yl)benzamide
(30t)
4.1.15. 1-(4-(3-((5-Bromo-4-(4-methyl-2-(methylamino)thiazol-5-
From tert-butyl (4-cyclopropyl-5-(3-(dimethylamino)acryloyl)
yl)pyrimidin-2-yl)amino)phenyl) piperazin-1-yl)ethanone (30q)
thiazol-2-yl)(methyl)carbamate
and
3-guanidino-N-(1-
From
tert-butyl
(5-(5-bromo-2-chloropyrimidin-4-yl)-4-
methylpiperidin-4-yl)benzamide. Cream (24%), mp 254e256 ꢀC.
Anal. RP-HPLC: tR min 11.30 (method A), 8.57 (method C), 10.65
methylthiazol-2-yl)(methyl)carbamate and 1-(4-(3-aminophenyl)
piperazin-1-yl)ethanone. Yellow solid (28%), mp 205e207 ꢀC. Anal.
RP-HPLC: tR 10.52 min (method D), purity 100%. 1H NMR (400 MHz,
(method D), purity 98%. 1H NMR (400 MHz, DMSO- d6):
d 0.90e1.05
(4H, m), 1.50e1.65 (2H, m), 1.72e1.82 (2H, m), 1.94 (2H, t,
J ¼ 11.6 Hz), 2.17 (3H, s), 2.58e2.68 (1H, m), 2.76 (2H, d, J ¼ 11.6 Hz),
2.82 (3H, d, J ¼ 4.8 Hz), 3.66e3.78 (1H, m), 7.01 (1H, d, J ¼ 5.6 Hz),
7.39e7.41 (2H, m), 7.91 (1H, d, J ¼ 8.0 Hz), 8.03 (1H, q, J ¼ 4.8 Hz),
8.09e8.18 (2H, m), 8.33 (1H, d, J ¼ 5.6 Hz), 9.52 (1H, s). 13C NMR
DMSO‑d6):
d
2.05 (3H, s), 2.30 (3H, s), 2.87 (3H, d, J ¼ 4.8 Hz), 3.08
(2H, apparent t, J ¼ 4.8 Hz), 3.14 (2H, apparent t, J ¼ 4.8 Hz),
3.50e3.53 (4H, m), 6.59 (1H, dd, J ¼ 8.0 & 1.6 Hz), 7.13 (1H, apparent
t, J ¼ 8.0 Hz), 7.20 (1H, br d, J ¼ 8.0 Hz), 7.42 (1H, br s), 7.41 (1H, q,
J ¼ 4.8 Hz), 8.57 (1H, s), 9.60 (1H, s). 13C NMR (100 MHz, DMSO‑d6):
(100 MHz, DMSO‑d6):
d 9.2, 13.1, 31.3, 31.9, 46.5, 47.0, 55.0, 108.1,
d
19.1, 21.7, 31.3, 41.1, 45.9, 49.4,106.3,107.5,110.3,111.2,114.0,129.4,
117.3, 119.0, 120.4, 121.9, 128.5, 136.1, 141.1, 158.1, 158.3, 159.3, 159.9,
166.7, 170.0. HRMS (ESIþ): m/z [M þ H]þ calcd for C24H30N7OS,
464.2227; found 464.2436.
141.2, 151.7, 152.4, 158.2, 158.5, 161.0, 168.7, 169.8. HRMS (ESIþ): m/z
[M þ H]þ calcd for C21H25 BrN7OS, 502.1019, found 502.1035.
11

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
4.1.19. 1-(4-(3-((4-(4-Isopropyl-2-(methylamino)thiazol-5-yl)
4.1.23. 4-(4-Methyl-2-(methylamino)thiazol-5-yl)-2-((3-methyl-5-
pyrimidin-2-yl)amino)phenyl) piperazin-1-yl)ethanone (30u)
(piperazin-1-yl)phenyl) amino)pyrimidine-5-carbonitrile (30y)
From
tert-butyl
(5-(3-(dimethylamino)acryloyl)-4-
and 1-(4-
From
3-(dimethylamino)-2-(4-methyl-2-(methylamino)thia-
isopropylthiazol-2-yl)(methyl)carbamate
zole-5-carbonyl)acrylonitrile and 1-(3-methyl-5-(piperazin-1-yl)
phenyl)guanidine. Brown solid (76%). Melting point: 119e121 ꢀC.
Anal. RP-HPLC: tR 11.42 min (method B), 10.97 min (method D),
acetylpiperazin-1-yl)phenyl)guanidine. Yellow solid (15%), mp
190e192 ꢀC. Anal. RP-HPLC: tR 11.19 min (method D), purity 100%.
1H NMR (400 MHz, DMSO- d6):
d
1.24 (6H, d, J ¼ 5.2 Hz), 2.05 (3H, s),
purity 95%. 1H NMR (400 MHz; DMSO‑d6):
d 2.23 (3H, s), 2.40 (3H,
2.85 (3H, d, J ¼ 4.8 Hz), 3.11 (2H, apparent t, J ¼ 4.8 Hz), 3.17 (2H,
apparent t, J ¼ 4.8 Hz), 3.48e3.51 (1H, m), 3.51e3.53 (4H, m), 6.58
(1H, dd, J ¼ 8.0 & 2.0 Hz), 6.83 (1H, d, J ¼ 5.6 Hz), 7.12 (1H, apparent
t, J ¼ 8.0 Hz), 7.22 (1H, dd, J ¼ 8.0 & 0.8 Hz), 7.44 (1H, t, J ¼ 2.0 Hz),
8.07 (1H, q, J ¼ 4.8 Hz), 8.33 (1H, d, J ¼ 5.6 Hz), 9.28 (1H, s). 13C NMR
br s), 2.78e2.85 (4H, m), 2.88 (3H, d, J ¼ 4.7 Hz), 2.98e3.06 (4H, m),
6.47 (1H, s), 7.02 (1H, s), 7.13 (1H, br s), 8.24 (1H, q, J ¼ 4.7 Hz), 8.75
(1H, s), 10.00 (1H, br s). 13C NMR (150 MHz; DMSO‑d6):
d 19.5, 21.7,
30.9, 45.6, 49.6, 93.3, 105.1, 111.5, 112.1, 117.9, 137.9, 139.3, 152.0,
154.9,159.0,161.0, 163.4,170.2. HRMS (ESIþ): m/z [M þ H]þ calcd for
(100 MHz, DMSO‑d6):
d
21.7, 22.5, 29.6, 31.5, 41.2, 45.98, 49.1, 49.6,
C21H25N8S 421.1917, found 421.1921.
107.2, 107.8, 110.0, 111.1, 117.3, 129.3, 141.8, 151.7, 158.4, 158.9, 160.1,
162.2, 168.7, 170.5. HRMS (ESIþ): m/z [M H]þ calcd for
23H30N7OS, 452.2227; found 452.2418.
þ
4.1.24. 4-(4-Methyl-2-(methylamino)thiazol-5-yl)-2-((3-methyl-5-
morpholinophenyl)amino) pyrimidine-5-carbonitrile (30z)
C
From
3-(dimethylamino)-2-(4-methyl-2-(methylamino)thia-
and 1-(3-methyl-5-
zole-5-carbonyl)acrylonitrile
4.1.20. 1-(4-(3-(4-(4-Methyl-2-(methylamino)thiazol-5-yl)
morpholinophenyl)guanidine. Yellow solid (55%). Melting point:
pyrimidin-2-ylamino)phenyl) piperazin-1-yl)ethanone (30v)
From 1-(4-acetylpiperazin-1-yl)phenyl)guanidine and 3-
(dimethylamino)-1-(4-methyl-2-(methylamino)thiazol-5-yl)prop-
2-en-1-one. Light yellow powder (78%), mp 217e219 ꢀC. Anal. RP-
HPLC: tR 13.42 min (method A), 10.62 min (method D), purity
243e245 ꢀC. Anal. RP-HPLC: tR 12.67 min (method B), 11.52 min
(method D), purity 96%. 1H NMR (400 MHz; DMSO‑d6):
d 2.24 (3H,
s), 2.39 (3H, br s), 2.87 (3H, s), 3.05e3.12 (4H, m), 3.68e3.82 (4H,
m), 6.50 (1H, s), 7.05 (1H, s), 7.15 (1H, s), 8.24 (1H, br s), 8.75 (1H, s),
10.03 (1H, br s). 13C NMR (100 MHz; DMSO‑d6):
d 19.5, 21.6, 30.8,
100%. 1H NMR (400 MHz, DMSO‑d6):
d
2.05 (3H, s), 2.47 (3H, s), 2.86
48.7, 66.1, 93.3, 104.8, 111.2, 112.6, 117.8, 138.1, 139.4, 151.4, 156.0,
(3H, d, J ¼ 4.8 Hz), 3.11 (2H, apparent t, J ¼ 5.2 Hz), 3.17 (2H,
apparent t, J ¼ 5.2 Hz), 3.54e3.64 (4H, m), 6.57 (1H, dd, J ¼ 8.0 &
1.6 Hz), 6.90 (1H, d, J ¼ 5.6 Hz), 7.12 (1H, apparent t, J ¼ 8.0 Hz), 7.22
(1H, d, J ¼ 8.0 Hz), 7.55 (1H, apparent t, J ¼ 1.6 Hz), 8.05 (1H, q,
J ¼ 4.8 Hz), 8.33 (1H, d, J ¼ 5.6 Hz), 9.28 (1H, s). 13C NMR (100 MHz,
161.0, 163.4, 170.2. HRMS (ESIþ): m/z [M þ H]þ Calculated for
C
23H28N7S 422.1758, found 422.1770.
4.1.25. 5-(2-((3,5-Dimorpholinophenyl)amino)pyrimidin-4-yl)-
N,4-dimethylthiazol-2-amine (30aa)
From 3-(dimethylamino)-1-(4-methyl-2-(methylamino)thiazol-
5-yl)prop-2-en-1-one and 1-(3,5-dimorpholinophenyl)guanidine.
Pale yellow solid (2%). Decomposes at 205 ꢀC. Anal. RP-HPLC: tR
11.30 min (method B), 10.63 min (method D), purity 97%. 1H NMR
DMSO‑d6):
d
19.1, 21.7, 31.3, 41.2, 46.0, 49.1, 49.6, 107.1, 107.1, 109.9,
111.0, 118.4, 129.3, 141.8, 151.7, 152.7, 158.1, 159.1, 160.1, 168.7, 169.8.
HRMS (ESIþ): m/z [M þ H]þ calcd for C21H26N7OS, 424.1914; found
424.1955.
(400 MHz; DMSO‑d6):
d
2.44 (3H, s), 2.84 (3H, d, J ¼ 4.8 Hz),
3.00e3.20 (8H, m), 3.64e3.81 (8H, m), 6.13 (1H, s), 6.88 (1H, d,
J ¼ 5.5 Hz), 6.94 (1H, d, J ¼ 1.9 Hz), 8.04 (1H, q, J ¼ 4.6 Hz), 8.31 (1H,
4.1.21. 2-((3-(Azepan-1-yl)phenyl)amino)-4-(4-methyl-2-
(methylamino)thiazol-5-yl)pyrimidine-5-carbonitrile (30w)
d, J ¼ 5.4 Hz), 9.10 (1H, s). 13C NMR (150 MHz; DMSO‑d6):
d 18.7,
From
3-(dimethylamino)-2-(4-methyl-2-(methylamino)thia-
30.7, 49.2, 66.2, 97.2, 98.4, 106.6, 117.9, 141.7, 152.2, 152.3, 157.6,
zole-5-carbonyl)acrylonitrile and 1-(3-(azepan-1-yl)phenyl)guani-
dine. Pale yellow solid (32%). Melting point: 249e252 ꢀC. Anal. RP-
HPLC: tR 13.15 min (method B), 11.75 min (method D), purity 99%.
158.5, 159.7, 169.5. HRMS (ESIþ): m/z [M
23H30N7O2S 468.2176, found 468.2203.
þ
H]þ calcd for
C
1H NMR (400 MHz, DMSO‑d6):
d
1.45 (4H, br s), 1.72 (4H, br s), 2.38
4.1.26. 1,1’-(4,4’-(5-((4-(4-Methyl-2-(methylamino)thiazol-5-yl)
pyrimidin-2-yl)amino)-1,3-phenylene)bis(piperazine-4,1-diyl))
diethanone (30 ab)
(3H, s), 2.87 (3H, s), 3.44 (4H, br t, J ¼ 6.0 Hz), 6.39 (1H, dd, J ¼ 8.0 &
1.6 Hz), 6.97e7.10 (3H, m), 8.21 (1H, br s), 8.74 (1H, s), 9.84 (1H, br
s). 13C NMR (100 MHz, DMSO‑d6):
d
19.3, 26.3, 26.8, 30.7, 48.7, 93.7,
From 3-(dimethylamino)-1-(4-methyl-2-(methylamino)thiazol-
103.3,106.4, 107.6,117.7,129.0,139.7, 148.6, 158.9,161.0,163.2,170.0.
HRMS (ESIþ): m/z [M þ H]þ calcd for C22H26N7S 420.1965, found
420.1969.
5-yl)prop-2-en-1-one
and
1-(3,5-bis(4-acetylpiperazin-1-yl)
phenyl)guanidine. Pale yellow solid (16%). Melting point:
200e202 ꢀC. Anal. RP-HPLC: tR 12.94 min (method B), 10.32 min
(method D), purity 95%. 1H NMR (400 MHz; DMSO‑d6; 1:1 mix of
rotamers):
d
2.04 (6H, s), 2.46 (3H, s), 2.85 (3H, d, J ¼ 4.8 Hz),
4.1.22. 2-((3-(Azepan-1-yl)-5-methylphenyl)amino)-4-(4-methyl-
2-(methylamino)thiazol-5-yl) pyrimidine-5-carbonitrile (30x)
3.05e3.11 (4H, m), 3.12e3.18 (4H, m), 3.51e3.61 (8H, m), 6.17 (1H,
s), 6.89 (1H, d, J ¼ 5.6 Hz), 6.96 (2H, d, J ¼ 1.9 Hz), 8.05 (1H, d,
J ¼ 4.8 Hz), 8.31 (1H, d, J ¼ 5.4 Hz), 9.13 (1H, s). 13C NMR (125 MHz;
From
3-(dimethylamino)-2-(4-methyl-2-(methylamino)thia-
and 1-(3-(azepan-1-yl)-5-
zole-5-carbonyl)acrylonitrile
DMSO‑d6; mix of rotamers):
d 18.7, 21.7, 31.0, 40.8, 45.6, 49.0, 49.4,
methylphenyl)guanidine. Yellow solid (18% over three steps).
Melting point: 217e219 ꢀC. Anal. RP-HPLC: tR 13.42 min (method B),
11.20 min (method D), purity 99%. 1H NMR (500 MHz; DMSO‑d6):
98.9, 99.3, 106.5, 117.9, 141.7, 151.9, 152.3, 157.6, 158.5, 159.6, 168.3,
169.5. HRMS (ESIþ): m/z [M þ H]þ calcd for C27H36N9O2S 550.2707,
found 550.2749.
d
1.35e1.52 (4H, m), 1.62e1.81 (4H, m), 2.21 (3H, s), 2.39 (3H, br s),
2.87 (3H, d, J ¼ 4.8 Hz), 3.42 (4H, t, J ¼ 6.0 Hz), 6.23 (1H, s), 6.85 (1H,
4.1.27. 2-((3,5-Dimorpholinophenyl)amino)-4-(4-methyl-2-
(methylamino)thiazol-5-yl)pyrimidine-5-carbonitrile (30ac)
s), 6.88 (1H, s), 8.34 (1H, q, J ¼ 4.6 Hz), 8.74 (1H, s), 9.96 (1H, br s).
13C NMR (150 MHz; DMSO‑d6):
d
19.9, 22.4, 26.7, 27.5, 31.3, 49.3,
From
3-(dimethylamino)-2-(4-methyl-2-(methylamino)thia-
93.7, 101.2, 107.7, 109.1, 118.4, 138.5, 140.1, 149.1, 155.0, 159.5, 161.5,
163.7, 170.6. HRMS (ESIþ): m/z [M þ H]þ calcd for C23H28N7S
434.2121, found 436.2106.
zole-5-carbonyl)acrylonitrile and 1-(3,5-dimorpholinophenyl)gua-
nidine. Pale yellow solid (2%). Melting point: 180e182 ꢀC. Anal. RP-
HPLC: tR 11.95 min (method B), 10.94 min (method D), purity 100%.
12

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
1H NMR (400 MHz; DMSO‑d6):
d
2.36 (3H, br s), 2.88 (3H, d,
4.3.2. Cell culture
J ¼ 4.1 Hz), 3.07 (8H, dd, J ¼ 6.9 & 4.7 Hz), 3.73 (8H, dd, J ¼ 6.3 &
Cancer cell lines were obtained from the American Type Tissue
Collection (ATTC) global bioresource centre and cultured in RPMI-
1640 with 10% FBS.
4.8 Hz), 6.23 (1H, s), 6.91 (2H, s), 8.34 (1H, br s), 8.76 (1H, s), 10.00
(1H, br s). 13C NMR (100 MHz; DMSO‑d6):
d 19.3, 31.0, 49.0, 66.1,
93.6, 98.7, 99.2, 100.0, 117.8, 140.0, 152.0, 159.0, 163.5. HRMS (ESIþ):
m/z [M þ H]þ calcd for C24H29N8O2S 493.2129, found 493.2134.
4.3.3. Proliferation assays
MTT
(3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide, Sigma) assays were performed as reported previously
[35]. Compound concentrations required to inhibit 50% of cell
growth (GI50) were calculated using non-linear regression analysis.
4.1.28. 2-((3,5-Bis(4-acetylpiperazin-1-yl)phenyl)amino)-4-(4-
methyl-2-(methylamino)thiazol-5-yl)pyrimidine-5-carbonitrile
(30ad)
From
3-(dimethylamino)-2-(4-methyl-2-(methylamino)thia-
4.3.4. Caspase-3 assay
zole-5-carbonyl)acrylonitrile and 1-(3,5-bis(4-acetylpiperazin-1-
yl)phenyl)guanidine. Decomposes at 167 ꢀC. Anal. RP-HPLC: tR
13.47 min (method B), 10.72 min (method D), purity 100%. 1H NMR
Activity of caspase 3 was measured using the fuorimetric cas-
pase 3 assay kit (Sigma-Aldrich, cat: CASP3F-1 KT). Method as
described from the kit.
(400 MHz; DMSO‑d6; 1:1 mix of rotamers): d 2.04 (6H, s), 2.36 (3H,
br s), 2.87 (3H, d, J ¼ 4.7 Hz), 3.02e3.11 (4H, m), 3.12e3.17 (4H, m),
3.46e3.66 (8H, m), 6.27 (1H, d, J ¼ 2.1 Hz), 6.91 (2H, d, J ¼ 2.0 Hz),
8.33 (1H, br s), 8.76 (1H, s), 10.0 (1H, br s). 13C NMR (125 MHz;
4.3.5. Cell cycle analysis and detection of apoptosis
Cells (4 ꢄ 105) were cultured for 24 h in medium alone or with
varying concentrations of inhibitor. Cells were harvested and pre-
pared in hypotonic fluorochrome solution (0.1% sodium citrate;
DMSO‑d6; mix of rotamers):
d 19.5, 21.2, 30.9, 40.7, 45.5, 48.7, 49.1,
93.5, 100.0, 100.6, 140.0, 151.9, 159.0, 161.1, 136.2, 168.3, 170.2. HRMS
(ESIþ): m/z [M þ H]þ calcd for C28H35N10O2S 575.2660, found
575.2664.
0.1% Triton X-100; 50 mg/ml propidium iodide; 100ug/ml RNase A in
dH2O). Cell cycle status was analyzed using a Beckman Coulter
EPICS-XL MCL™ flow cytometer and data were analyzed using
EXPO32™ software.
Apoptotic populations were examined using FITC annexinV/PI
(Annexin V-FITC Apoptosis Detection Kit I, BD Biosciences, cat:
556547) staining after cells were cultured in medium only or with
varying concentrations of inhibitors according to the protocols (BD
Bioscience). The annexin V/PI-positive apoptotic cells were
4.1.29. 1,1’-(4,4’-(5-((5-Methyl-4-(4-methyl-2-(methylamino)
thiazol-5-yl)pyrimidin-2-yl)amino)-1,3-phenylene)bis(piperazine-
4,1-diyl))diethanone (30ae)
From
3-(dimethylamino)-2-methyl-1-(4-methyl-2-(methyl-
and 1-(3,5-bis(4-
amino)thiazol-5-yl)prop-2-en-1-one
acetylpiperazin-1-yl)phenyl)guanidine. Pale yellow solid (10%).
Melting point: 121e123 ꢀC. Anal. RP-HPLC: tR 10.34 min (method
A),10.34 min (method D), purity 95%. 1H NMR (400 MHz; CDCl3; 1:1
enumerated using
cytometer.
a Beckman Coulter EPICS-XL MCL™ flow
mix of rotamers): d 2.13 (6H, s), 2.22 (3H, s), 2.23 (3H, s), 3.02 (3H, d,
4.3.6. Detection of apoptosis in primary CLL cells
J ¼ 3.8 Hz), 3.14e3.22 (8H, m), 3.55e3.65 (4H, m), 3.71e3.80 (4H,
m), 5.63 (1H, br s), 6.17 (1H, apparent t, J ¼ 2.0 Hz), 6.90 (2H, d,
J ¼ 2.0 Hz), 7.04 (1H, s), 8.27 (1H, s). 13C NMR (100 MHz; CDCl3; mix
Freshly isolated primary CLL cells and normal B- and T-cells
were cultured in RPMI with 10% foetal calf serum and L-glutamine,
penicillin, and streptomycin. Cells were maintained at 37 ꢀC in an
atmosphere containing 95% air and 5% CO2 (vol/vol). CLL cells (106/
mL) were treated with inhibitor for 48 h. Subsequently, cells were
of rotamers):
d 16.2, 17.7, 21.3, 32.0, 41.4, 46.3, 49.7, 50.1, 100.2,
100.5, 116.3, 119.1, 141.6, 149.2, 152.5, 158.9, 160.0, 169.0, 170.8.
HRMS (ESIþ): m/z [M þ H]þ calcd for C28H38N9O2S 564.2864, found
564.2852.
labelled with CD19-APC (Caltag) and then re-suspended in 200
binding buffer containing 4
mL
mL annexin VeFITC (Bender Medsys-
tems, Vienna, Austria). Apoptosis was quantified in the CD19þ CLL
cells, CD19þ normal B-cells and CD3þ normal T-cells using an
Accuri C6 flow cytometer and FlowJo software (TreeStar). LD50
values were calculated from line-of-best-fit analysis of the
sigmoidal dose response curves.
4.2. Molecular modelling
Modelling was performed using OpenEye OEDocking software.
Up to 2000 conformations of the individual compounds to be
docked were generated using OpenEye OMEGA program and
docked using OpenEye FRED into receptors generated by the
OpenEye MAKE RECEPTOR program using default settings from
reported crystal structures for CDK9 ([4BCF] and [4BCG]) and CDK2
([4BCP] and [4BCO]). FRED Chemgauss 4 was used as the scoring
function. The number of alternate poses retuned was set to 20 and
all other settings were on default. Results were visualized using
OpenEye VIDA software and images generated using DeLano Sci-
entific LLC PyMOL freeware.
4.3.7. Western blots
Western blotting was performed as described [25]. Antibodies
used were as follows: total RNAP-II (8WG16), phosphorylated
RNAP-II Ser-2, phosphorylated RNAP-II Ser-5 (Covance), p53
(Dako), b-actin (Sigma-Aldrich), Mcl-1, Cleaved PARP (Cell Signal-
ling Technologies), p21 (Santa Cruz). Both anti-mouse and anti-
rabbit immunoglobulin G (IgG) horse-radish peroxidase-conju-
gated antibodies were obtained from Dako.
4.3. Biological assays
4.3.8. Statistical analysis
All experiments were performed in triplicate and repeated at
least twice; representative experiments were selected for figures.
Statistical significance was determined using one-way analysis of
variance (ANOVA), with a minimal level of significance at p < 0.01.
4.3.1. Kinase assay
Inhibition of CDKs and other kinases was measured by radio-
metric assay using the Millipore KinaseProfiler services using
Km(app) concentration of ATP of each kinase. Half-maximal inhibi-
tion (IC50) values were calculated from 10-point dose-response
curves and apparent inhibition constants (Ki) were calculated
from the IC50 values and appropriate Km (ATP) values for the ki-
nases in question.
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
13

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
synergistically enhances cell death induced by the Bcl-2 selective inhibitor
venetoclax in preclinical models of acute myeloid leukemia, Signal Transduct.
Target. Ther. 5 (2020) 17.
appeared to influence the work reported in this paper.
Acknowledgement
[18] E. Walsby, G. Pratt, H. Shao, A.Y. Abbas, P.M. Fischer, T.D. Bradshaw,
P. Brennan, C. Fegan, S. Wang, C. Pepper, A novel Cdk9 inhibitor preferentially
targets tumor cells and synergizes with fludarabine, Oncotarget 5 (2014)
375e385.
[19] M.H. Rahaman, Y. Yu, L. Zhong, J. Adams, F. Lam, P. Li, B. Noll, R. Milne, J. Peng,
S. Wang, CDKI-73: an orally bioavailable and highly efficacious CDK9 inhibitor
against acute myeloid leukemia, Invest. N. Drugs 37 (2019) 625e635.
[20] M.H. Rahaman, F. Lam, L. Zhong, T. Teo, J. Adams, M. Yu, R.W. Milne, C. Pepper,
N.A. Lokman, C. Ricciardelli, M.K. Oehler, S. Wang, Targeting CDK9 for treat-
ment of colorectal cancer, Mol. Oncol. 13 (2019) 2178e2193.
This study was supported by Cancer Research UK grants C21568/
A8988 and, C21568/A12474. OpenEye Scientific Software Inc. is
acknowledged for the software using for docking studies.
InstantJChem (5.3.1) and JChem for Excel (5.3.1), ChemAxon
(https://www.chemaxon.com). were used for displaying, charac-
terizing and capturing chemical structures and assoicated physi-
cochemical and biological properties.
[21] S. Xie, H. Jiang, X.W. Zhai, F. Wei, S.D. Wang, J. Ding, Y. Chen, Antitumor action
of CDK inhibitor LS-007 as a single agent and in combination with ABT-199
against human acute leukemia cells, Acta Pharmacol. Sin. 37 (2016)
1481e1489.
Appendix A. Supplementary data
[22] H. McCalmont, K.L. Li, L. Jones, J. Toubia, S.C. Bray, D.A. Casolari, C. Mayoh,
S.E. Samaraweera, I.D. Lewis, R.K. Prinjha, N. Smithers, S. Wang, R.B. Lock,
R.J. D’Andrea, Efficacy of combined CDK9/BET inhibition in preclinical models
of MLL-rearranged acute leukemia, Blood Adv 4 (2020) 296e300.
[23] J. Li, X. Zhi, S. Chen, X. Shen, C. Chen, L. Yuan, J. Guo, D. Meng, M. Chen, L. Yao,
CDK9 inhibitor CDKI-73 is synergetic lethal with PARP inhibitor olaparib in
BRCA1 wide-type ovarian cancer, Am. J. Cancer Res. 10 (2020) 1140e1155.
[24] Y.A. Sonawane, M.A. Taylor, J.V. Napoleon, S. Rana, J.I. Contreras, A. Natarajan,
Cyclin dependent kinase 9 inhibitors for cancer therapy, J. Med. Chem. 59
(2016) 8667e8684.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ejmech.2021.113244.
Author contributions
The manuscript was written through contributions of all au-
thors. All authors have given approval to the final version of the
manuscript.
[25] R. Chen, M.J. Keating, V. Gandhi, W. Plunkett, Transcription inhibition by
flavopiridol: mechanism of chronic lymphocytic leukemia cell death, Blood
106 (2005) 2513e2519.
[26] C.M. Olson, B. Jiang, M.A. Erb, Y. Liang, Z.M. Doctor, Z. Zhang, T. Zhang,
N. Kwiatkowski, M. Boukhali, J.L. Green, W. Haas, T. Nomanbhoy, E.S. Fischer,
R.A. Young, J.E. Bradner, G.E. Winter, N.S. Gray, Pharmacological perturbation
of CDK9 using selective CDK9 inhibition or degradation, Nat. Chem. Biol. 14
(2018) 163e170.
References
[1] M. Malumbres, M. Barbacid, Mammalian cyclin-dependent kinases, Trends
Biochem. Sci. 30 (2005) 630e641.
[27] 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-
[2] J. Chou, D.A. Quigley, T.M. Robinson, F.Y. Feng, A. Ashworth, Transcription-
associated cyclin-dependent kinases as targets and biomarkers for cancer
therapy, Canc. Discov. 10 (2020) 351e370.
[3] P.K. Parua, R.P. Fisher, Dissecting the Pol II transcription cycle and derailing
cancer with CDK inhibitors, Nat. Chem. Biol. 16 (2020) 716e724.
methoxyethyl)amino)cyclohexyl)amino)pyridin-4-
yl)thiazol-2-yl)amino)
methyl)tetrahydro-2H-pyran-4-carbonitrile (JSH-150) as a novel highly se-
lective and potent CDK9 kinase inhibitor, Eur. J. Med. Chem. 158 (2018)
896e916.
[4] S. Wang, P.M. Fischer, Cyclin-dependent kinase 9:
a key transcriptional
regulator and potential drug target in oncology, virology and cardiology,
Trends Pharmacol. Sci. 29 (2008) 302e313.
[5] G. Romano, A. Giordano, Role of the cyclin-dependent kinase 9-related
pathway in mammalian gene expression and human diseases, Cell Cycle 7
(2008) 3664e3668.
[6] B. Palancade, O. Bensaude, Investigating RNA polymerase II carboxyl-terminal
domain (CTD) phosphorylation, Eur. J. Biochem. 270 (2003) 3859e3870.
[7] J. Garriga, X. Grana, Cellular control of gene expression by T-type cyclin/CDK9
complexes, Gene 337 (2004) 15e23.
[28] J. Cidado, S. Boiko, T. Proia, D. Ferguson, S.W. Criscione, M. San Martin, P. Pop-
Damkov, N. Su, V.N. Roamio Franklin, C. Sekhar Reddy Chilamakuri,
C.S. D’Santos, W. Shao, J.C. Saeh, R. Koch, D.M. Weinstock, M. Zinda, S.E. Fawell,
L. Drew, AZD4573 is a highly selective CDK9 inhibitor that suppresses MCL-1
and induces apoptosis in hematologic cancer cells, Clin. Canc. Res. 26 (2020)
922e934.
[29] J. Xu, H. Li, X. Wang, J. Huang, S. Li, C. Liu, R. Dong, G. Zhu, C. Duan, F. Jiang,
Y. Zhang, Y. Zhu, T. Zhang, Y. Chen, W. Tang, T. Lu, Discovery of coumarin
derivatives as potent and selective cyclin-dependent kinase 9 (CDK9) in-
hibitors with high antitumour activity, Eur. J. Med. Chem. 200 (2020) 112424.
[30] H. Shao, S. Shi, S. Huang, A.J. Hole, A.Y. Abbas, S. Baumli, X. Liu, F. Lam,
D.W. Foley, P.M. Fischer, M. Noble, J.A. Endicott, C. Pepper, S. Wang,
Substituted 4-(thiazol-5-yl)-2-(phenylamino)pyrimidines are highly active
CDK9 inhibitors: synthesis, X-ray crystal structures, structure-activity rela-
tionship, and anticancer activities, J. Med. Chem. 56 (2013) 640e659.
[31] H. Shao, S. Shi, D.W. Foley, F. Lam, A.Y. Abbas, X. Liu, S. Huang, X. Jiang,
N. Baharin, P.M. Fischer, S. Wang, Synthesis, structure-activity relationship
and biological evaluation of 2,4,5-trisubstituted pyrimidine CDK inhibitors as
potential anti-tumour agents, Eur. J. Med. Chem. 70 (2013) 447e455.
[32] S. Grosjean, S. Triki, J.-C. Meslin, K. Julienne, D. Deniaud, Synthesis of nitrogen
bicyclic scaffolds: pyrimido[1,2-a]pyrimidine-2,6-diones, Tetrahedron 66
(2010) 9912e9924.
[33] A. Noack, H. Hartmann, Synthesis and characterisation of N,N-disubstituted 2-
amino-5-acylthiophenes and 2-amino-5-acylthiazoles, Tetrahedron 58 (2002)
2137e2146.
[34] R.J. Highet, W.C. Wildman, Solid manganese dioxide as an oxidizing agent,
J. Am. Chem. Soc. 77 (1955) 4399e4401.
[35] S. Wang, C. Meades, G. Wood, A. Osnowski, S. Anderson, R. Yuill, M. Thomas,
M. Mezna, W. Jackson, C. Midgley, G. Griffiths, I. Fleming, S. Green, I. McNae,
S.Y. Wu, C. McInnes, D. Zheleva, M.D. Walkinshaw, P.M. Fischer, 2-Anilino-4-
(thiazol-5-yl)pyrimidine CDK inhibitors: synthesis, SAR analysis, X-ray crys-
tallography, and biological activity, J. Med. Chem. 47 (2004) 1662e1675.
[8] C.W. Bacon, I. D’Orso, CDK9:
Transcription 10 (2019) 57e75.
a signaling hub for transcriptional control,
[9] A.M. Mandelin 2nd, R.M. Pope, Myeloid cell leukemia-1 as a therapeutic
target, Expert Opin. Ther. Targets 11 (2007) 363e373.
[10] M.R. Warr, G.C. Shore, Unique biology of Mcl-1: therapeutic opportunities in
cancer, Curr. Mol. Med. 8 (2008) 138e147.
[11] S. Caenepeel, S.P. Brown, B. Belmontes, G. Moody, K.S. Keegan, D. Chui,
D.A. Whittington, X. Huang, L. Poppe, A.C. Cheng, M. Cardozo, J. Houze, Y. Li,
B. Lucas, N.A. Paras, X. Wang, J.P. Taygerly, M. Vimolratana, M. Zancanella,
L. Zhu, E. Cajulis, T. Osgood, J. Sun, L. Damon, R.K. Egan, P. Greninger,
J.D. McClanaghan, J. Gong, D. Moujalled, G. Pomilio, P. Beltran, C.H. Benes,
A.W. Roberts, D.C. Huang, A. Wei, J. Canon, A. Coxon, P.E. Hughes, AMG 176, a
selective MCL1 inhibitor, is effective in hematologic cancer models alone and
in combination with established therapies, Canc. Discov. 8 (2018) 1582e1597.
[12] X. Yi, A. Sarkar, G. Kismali, B. Aslan, M. Ayres, L.R. Iles, M.J. Keating,
W.G. Wierda, J.P. Long, M.T.S. Bertilaccio, V. Gandhi, AMG-176, an mcl-1
antagonist, shows preclinical efficacy in chronic lymphocytic leukemia, Clin.
Canc. Res. 26 (2020) 3856e3867.
[13] A. Nencioni, F. Hua, C.P. Dillon, R. Yokoo, C. Scheiermann, M.H. Cardone,
E. Barbieri, I. Rocco, A. Garuti, S. Wesselborg, C. Belka, P. Brossart, F. Patrone,
A. Ballestrero, Evidence for a protective role of Mcl-1 in proteasome inhibitor-
induced apoptosis, Blood 105 (2005) 3255e3262.
[14] F. Morales, A. Giordano, Overview of CDK9 as a target in cancer research, Cell
Cycle 15 (2016) 519e527.
[36] X. Huang, Y. He, Y. Ding, The novel syntheses of
from -bromoenol phosphates, phosphorus, sulfur, Silicon Relat. Elem. 174
(2006) 201e207.
a-trifluoromethylated ketones
[15] T. Wu, Z. Qin, Y. Tian, J. Wang, C. Xu, Z. Li, J. Bian, Recent developments in the
biology and medicinal chemistry of CDK9 inhibitors: an update, J. Med. Chem.
63 (2020) 13228e13257.
b
[37] V. Krystof, I. Chamrad, R. Jorda, J. Kohoutek, Pharmacological targeting of
CDK9 in cardiac hypertrophy, Med. Res. Rev. 30 (2010) 646e666.
[38] T. Shimamura, J. Shibata, H. Kurihara, T. Mita, S. Otsuki, T. Sagara, H. Hirai,
Y. Iwasawa, Identification of potent 5-pyrimidinyl-2-aminothiazole CDK4, 6
inhibitors with significant selectivity over CDK1, 2, 5, 7, and 9, Bioorg. Med.
Chem. Lett 16 (2006) 3751e3754.
[16] D.C. Phillips, S. Jin, G.P. Gregory, Q. Zhang, J. Xue, X. Zhao, J. Chen, Y. Tong,
H. Zhang, M. Smith, S.K. Tahir, R.F. Clark, T.D. Penning, J.R. Devlin, J. Shortt,
E.D. Hsi, D.H. Albert, M. Konopleva, R.W. Johnstone, J.D. Leverson, A.J. Souers,
A novel CDK9 inhibitor increases the efficacy of venetoclax (ABT-199) in
multiple models of hematologic malignancies, Leukemia 34 (2020)
1646e1657.
[39] S. Ali, D.A. Heathcote, S.H. Kroll, A.S. Jogalekar, B. Scheiper, H. Patel, J. Brackow,
A. Siwicka, M.J. Fuchter, M. Periyasamy, R.S. Tolhurst, S.K. Kanneganti,
[17] D.A. Luedtke, Y. Su, J. Ma, X. Li, S.A. Buck, H. Edwards, L. Polin, J. Kushner,
S.H. Dzinic, K. White, H. Lin, J.W. Taub, Y. Ge, Inhibition of CDK9 by voruciclib
14

H. Shao, D.W. Foley, S. Huang et al.
European Journal of Medicinal Chemistry 214 (2021) 113244
J.P. Snyder, D.C. Liotta, E.O. Aboagye, A.G. Barrett, R.C. Coombes, The devel-
opment of a selective cyclin-dependent kinase inhibitor that shows antitumor
activity, Canc. Res. 69 (2009) 6208e6215.
trifluoromethyl substitution provides a global protective effect against hepatic
metabolism, J. Med. Chem. 38 (1995) 1355e1371.
[43] R. Filler, R. Saha, Fluorine in medicinal chemistry: a century of progress and a
60-year retrospective of selected highlights, Future Med. Chem. 1 (2009)
777e791.
[44] F. Leroux, Atropisomerism, biphenyls, and fluorine: a comparison of rotational
barriers and twist angles, Chembiochem 5 (2004) 644e649.
[45] K. Muller, C. Faeh, F. Diederich, Fluorine in pharmaceuticals: looking beyond
intuition, Science 317 (2007) 1881e1886.
[40] S. Wang, G. Griffiths, C.A. Midgley, A.L. Barnett, M. Cooper, J. Grabarek,
L. Ingram, W. Jackson, G. Kontopidis, S.J. McClue, C. McInnes, J. McLachlan,
C. Meades, M. Mezna, I. Stuart, M.P. Thomas, D.I. Zheleva, D.P. Lane,
R.C. Jackson, D.M. Glover, D.G. Blake, P.M. Fischer, Discovery and character-
ization of 2-anilino-4- (thiazol-5-yl)pyrimidine transcriptional CDK inhibitors
as anticancer agents, Chem. Biol. 17 (2010) 1111e1121.
[41] C. McInnes, S. Wang, S. Anderson, J. O’Boyle, W. Jackson, G. Kontopidis,
C. Meades, M. Mezna, M. Thomas, G. Wood, D.P. Lane, P.M. Fischer, Structural
determinants of CDK4 inhibition and design of selective ATP competitive in-
hibitors, Chem. Biol. 11 (2004) 525e534.
[42] G.D. Diana, P. Rudewicz, D.C. Pevear, T.J. Nitz, S.C. Aldous, D.J. Aldous,
D.T. Robinson, T. Draper, F.J. Dutko, C. Aldi, et al., Picornavirus inhibitors:
[46] V. Caracciolo, G. Laurenti, G. Romano, V. Carnevale, A.M. Cimini, C. Crozier-
Fitzgerald, E. Gentile Warschauer, G. Russo, A. Giordano, Flavopiridol induces
phosphorylation of AKT in a human glioblastoma cell line, in contrast to
siRNA-mediated silencing of Cdk9: implications for drug design and devel-
opment, Cell Cycle 11 (2012) 1202e1216.
15