Synthesis and activity of quinolinyl-methylene-thiazolinones as potent and selective cyclin-dependent kinase 1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.01.081

A novel series of quinolinyl-methylene-thiazolinones has been identified as potent and selective cyclin-dependent kinase 1 (CDK1) inhibitors. Their synthesis and structure activity relationships (SAR) are described. Representative compounds from this class reversibly inhibit CDK1 activity in vitro, and block cell cycle progression in human tumor cell lines, suggesting a potential use as antitumor agents.

Full text of DOI:10.1016/j.bmcl.2007.01.081

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2134–2138
Synthesis and activity of quinolinyl-methylene-thiazolinones as
potent and selective cyclin-dependent kinase 1 inhibitors
Shaoqing Chen,* Li Chen, Nam T. Le, Chunlin Zhao, Achyutharao Sidduri, Jian Ping Lou,
Christophe Michoud, Louis Portland, Nicole Jackson, Jin-Jun Liu, Fred Konzelmann,
Feng Chi, Christian Tovar, Qing Xiang, Yingsi Chen, Yang Wen and Lyubomir T. Vassilev
Roche Research Center, Hoffmann-La Roche Inc., Nutley, NJ 07110, USA
Received 11 January 2007; accepted 29 January 2007
Available online 31 January 2007
Abstract—A novel series of quinolinyl-methylene-thiazolinones has been identified as potent and selective cyclin-dependent kinase 1
(CDK1) inhibitors. Their synthesis and structure activity relationships (SAR) are described. Representative compounds from this
class reversibly inhibit CDK1 activity in vitro, and block cell cycle progression in human tumor cell lines, suggesting a potential
use as antitumor agents.
Ó 2007 Elsevier Ltd. All rights reserved.
Cyclin-dependent kinases (CDKs) are a family of serine/
threonine kinases that play critical roles in controlling
cell cycle progression.1,2 After binding to a family of reg-
ulatory proteins cyclins that expressed at different times
during the cell cycle, CDKs are activated. Additional
control over CDK activity is exerted by phosphoryla-
tion on specific threonine residues. Abnormal CDK con-
trol of the cell cycle has been linked to the molecular
pathology of cancer, and thus inhibitors of cyclin-de-
pendent kinases are anticipated to have antiproliferative
activity and potential therapeutic utility. Therefore,
CDKs have become attractive therapeutic targets for
cancer therapy.3–5
RNA interference studies.6–10 Similar experiments with
CDK1 have shown that its kinase activity is critical
for cell cycle progression through mitosis. Therefore,
CDK1 inhibitors are expected to effectively arrest tumor
cell growth.
Many small-molecule CDK inhibitors have been identi-
fied11–13 and some of them have entered clinical evalua-
tion for the treatment of cancer.14–19 Although specific
CDK2 and CDK4 inhibitors have been reported, the
identification of CDK1-selective small-molecule inhibi-
tors has been challenging. Recently, we discovered a
quinolinyl-methylene-thiazolinone as potent and selec-
tive CDK1 inhibitor.20 Here, we report the synthesis,
structure activity relationships (SAR), and anti-prolifer-
ative activities of the quinolinyl-methylene-thiazolinone
class of CDK1 inhibitors.
It has been established that three CDKs (CDK1, CDK2,
and CDK4) and their activating cyclin partners (A, B,
D, and E) play key roles in mammalian cell cycle regu-
lation.1 CDK2/cyclin E and CDK4/cyclin D control pas-
sage through the G1-phase and G1- to S-phase
transition by phosphorylation of the retinoblastoma
phosphoprotein, pRB. CDK2/cyclin A regulates passage
through the S-phase. CDK1/cyclin B controls the G2
checkpoint and regulates entry into mitosis. However,
it was found that CDK2 and CDK4 may not be essen-
tial for cell cycle progression in recent genetic and
Most of the quinolinyl-methylene-thiazolinone analogs
described in this paper (see Table 1) were prepared by
the two general synthetic methods outlined in Scheme 1.
In method A, rhodanine was reacted with the appropriate
amine 1 in the presence of HgCl2 and DIEA to give the
intermediate 3. Knoevenagel condensation of 3 with the
appropriate aldehydes catalyzed by benzoic acid and
piperidine in toluene gave the desired products 4a. In
method B, condensation of rhodanine with the appropri-
ate ketone or aldehyde such as 6-quinoline-carboxalde-
hyde yielded intermediate 6. Although different
Knoevenagel condensation conditions can be used,
Keywords: Cyclin-dependent kinase; CDK1 inhibitor; Anticancer.
*
Corresponding author. Tel.: +1 9732354636; fax: +1 9732356084;
e-mail: shaoqing.chen@roche.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.081

S. Chen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2134–2138
2135
Table 1. Structure–activity relationship of quinolinyl-methylene-thiaz-
olin-ones against CDK1/cyclinBa
Table 1 (continued)
Compound
Rl
Ki (lM)
O
N
S
N
H
R1
53
54
0.089
S
*
>2.0
S
N
*
N
a See Ref. 20 for a description of the CDK1 assay.
Compound
Rl
H
Ki (lM)
12
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
0.080
>2.0
0.27
MeO
Me
n-Bu
NaOAc/HOAc gave the best result with an excellent yield
for 6a in the case of aldehydes (Z = H). In the case of
ketones (Z = Me), a variety of Knoevenagel condensation
conditions such as NaOAc/HOAc, NH4OAc/DMF, and
PhCOOH/Piperidine/PhMe were investigated but
produced low yields of the product 6b. However,
NH4OAc in toluene gave an almost quantitative yield of
6b. Methylation of 6 with MeI provided the thioether
intermediate 7. Subsequent introduction of the R1 group
was achieved by treatment of 7 with a variety of amines
(R1-NH2) to give thiazolinone product 4. In this step,
MeCN was the solvent of choice although other solvents
such as EtOH and DMF could also be used. Most of the
products precipitated at the end of the reaction
when MeCN was used as solvent and were isolated by
filtration.
0.31
0.27
cyc-Propyl
Ph
PhCH2
>2.0
0.15
PhCH2CH2
PhCH2CH2CH2
2-Thienyl-CH2
0.060
0.24
0.035
0.045
0.28
2-Thienyl-CH2CH2
2-Pyridyl-CH2CH2
2-Pyridyl-CH2
0.31
0.12
o-F-Ph-CH2CH2
m-F-Ph-CH2CH2
p-F-Ph-CH2CH2
p-Cl-Ph-CH2CH2
p-Br-Ph-CH2CH2
p-MeOPh-CH2CH2
m-Cl-Ph-CH2CH2
o-MeOPh-CH2CH2
o-EtO-Ph-CH2CH2
o-Cl-Ph-CH2CH2
o-Cl-Ph-CH2
0.060
0.062
0.41
0.64
0.39
0.13
0.063
0.087
0.046
0.052
0.038
0.086
1.04
Oxazolinone analogs were prepared as shown in Scheme 2
using a procedure similar to method B described above
starting from thioxo-oxazolidinone 8 prepared from
sodium cyanide, potassium thiocyanate, and formalde-
hyde. Knoevenagel condensation and methylation of 8
provided the intermediate 10, which was finally treated
with amine to give the product 11.
o-Br-Ph-CH2
o-MeO-Ph-CH2
2,6-Di-Cl-Ph-CH2
OH
*
Compounds 12, 13, 14, and 15 were simply prepared as
shown in Scheme 3 by Knoevenagel condensation con-
ducted using sodium acetate in acetic acid.
46
0.23
OH
*
To obtain the analog 19 with a single bond linker between
thiazolinone and quinoline rings for activity comparison,
direct reduction of 28 under reduction conditions such as
hydrogenation, reduction with LiBH4 or Hantzsch ester
was tried and proved to be unsuccessful. It was finally pre-
pared using the route shown in Scheme 4 starting from
aminoquinoline 17. Diazotization of 17 yielded diazoni-
um salt intermediate, which was reacted with acrylate in
the presence of catalytic CuBr to produce beta-bromo es-
ter 18. Compound 18 was reacted with a thiourea 16, pre-
pared from the corresponding amine, to form the
thiazolinone ring in the desired product 19.
47
48
2.0
OH
0.35
*
OH
49
50
51
0.58
>2.0
0.39
*
OMe
*
The quinolinyl-methylene-thiazolinones were initially
identified by their anti-proliferative activity in a cell-
based screen. A distinct block in G2 phase of the cell
cycle suggested that CDK1 inhibition might be involved.
Their CDK1 inhibitory activity was confirmed by testing
for direct inhibition of CDK1/cyclin B activity in vitro.20
The good correlation between CDK1 inhibitory activity
OH
*
52
0.023
*

2136
S. Chen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2134–2138
O
O
N
N
S
a
+
R1 NH2
S
R1
N
H
O
N
N
S
S
b
N
H
R1
1
3
2
z
O
O
N
N
z
O
N
S
S
S
e
O
N
z
d
S
c
z
4
S
+
4a: Z = H
S
4b: Z = Me
N
N
2
5
6
7
5a: Z = H
6a: Z = H
5b: Z = Me
6b: Z = Me
Scheme 1. General synthetic scheme for quinolinyl-methylene-thiazolin-ones. Reagents and conditions: (a) HgCl2, DIEA, MeCN, rt, 12 h;
(b) 6-quinolinecarboxaldehyde, PhCOOH/piperidine, toluene, 150 °C, micro-wave, 20 min (Z = H only); (c) NaOAc, HOAc, 130 °C, 24 h (Z = H); or
NH4OAc, PhMe, 130 °C, 24 h, (Z = Me); (d) MeI, DIEA, EtOH, rt, 12 h; (e) R1-NH2, DIEA, MeCN, microwave, 145 °C, 20 min.
O
N
O
O
O
N
O
N
O
O
S
NaCN
KSCN
+ HCHO
N
O
MeS
R1
N
H
b
a
c
d
S
+
8
N
N
N
9
11
10
Scheme 2. Synthesis of oxazolinone 11. Reagents and conditions: (a) concd HCl, H2O, 0 °C, 30 min, then rt, 12 h; (b) 6-quinolinecarbox-aldehyde,
NaOAc, HOAc, reflux, 12 h; (c) MeI, DIEA, EtOH, rt, 12 h; (d) R1-NH2, microwave, 100 °C, 10 min.
Y
H
O
N
X
amine resulted in the complete loss of the CDK1 activity
(54 vs. 28). Synthetic intermediate 7 was also inactive.
These results indicated that a free NH group, a hydro-
gen bond donor, is essential for CDK1 inhibitory activ-
ity in this position. The length of the spacer between the
aryl and amine groups also affected the activity dramat-
ically. Except for compound 24 which lacks a spacer,
several carbon spacers are tolerated, and 1–2 carbon
R
Y
N
X
a
R
+
N
N
12 (X = S, Y = O, R = NH2)
13 (X = S, Y = O, R = OH)
14 (X = N(Me), Y =O, R = NH2)
15 (X = S, Y = NH, R = NH2)
S
a, b
Scheme 3. Synthesis of compounds 12–15. Reagents and conditions:
(a) NaOAc, HOAc, 130 °C, 12 h.
S
N
H
NH2
S
NH2
16
Br
NH2
and anti-proliferative activity against tumor cells further
strengthens the notion that growth arrest is derived from
CDK1 inhibition (data not shown).
O
O
N
N
S
N
H
O
S
e
c, d
N
N
We next generated a series of analogs of structure 4a and
explored the SAR in their CDK1 activity. The data
(Table 1) indicated that R1 can tolerate a variety of
changes. The unsubstituted compound (12) has a Ki of
80 nM, while alkoxy substitution (20) resulted in the loss
of activity. Alkyl, cycloalkyl, aryl, and heteroaryl substi-
tution in this position is tolerated. Methylation of the
19
18
17
Scheme 4. Synthesis of compound 19. Reagents and conditions:
(a) PhCO-NCS, CHCl3, reflux, 2 h; (b) aq. Na2CO3, 70 °C, 1 h; (c)
HBr, NaNO2, Me2CO, H2O, 0 °C, 30 min; (d) Me acrylate, cat CuBr,
60 °C, 1 h; (e) 16, NaOAc, 2-methoxyethanol, 100 °C, 5 h.

S. Chen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2134–2138
2137
length gave the best result (24 vs. 25, 26, 27 or 28 vs. 29,
30 vs. 31). With right spacer, most aryl is tolerated, but
does not contribute significantly to potency.
In the case of phenyl (or heteroaromatic) alkyl (1–2 car-
bons) as an R1 substituent, we further investigated
changes at the C-a or C-b positions. Appropriate substi-
tution at these positions was tolerated. The hydroxyl-
methyl group was introduced to improve solubility but
resulted in a decrease in activity compared with the
unsubstituted analogs (46–51). Stereochemistry also af-
fects the potency. The S-isomer was more potent than
the R-isomer (46 vs. 47 and 48 vs. 49). The most potent
compound in this series with Ki of 23 nM was the more
rigid analog (52). Since R1 can tolerate many changes,
this is a good position to modify the molecule to im-
prove the physicochemical properties.
We then focused on the 1–2 carbon linker analogs. Most
substitutions including halogen and alkoxy groups at
the o- or m-positions of the phenyl ring improved or re-
tained the potency (32, 33, 38–44). p-Substitution (35–
37) decreased the activity except in the case of very small
groups such as fluorine (34). Di-substitution (45) result-
ed in reduced activity. The thiophenyl analog (28) with
heteroaryl substitution was most potent against CDK1
(Ki = 35 nM).
O
O
N
S
O
N
N
S
N
H
N
H
S
N
H
CH3
N
H
S
4b
54
Ki> 2 μM
(R1=2-Thienyl-CH2-)
19
Ki> 2 μM
N
N
Ki: > 2 μM
N
H
O
N
O
N
O
N
S
O
N
H
S
N
H
S
O
S
28
Ki: 0.04μM
11
13
Ki> 2 μM
Ki: 0.63 μM
N
N
N
O
N
N
NH
N
S
H2N
H2N
15
Ki: > 2 μM
14
Ki> 2 μM
N
N
Figure 1. Structure and their CDK1 activities.
Table 2. Selectivity profile against a panel of diverse kinases (Ki, lM)a
Compound
CDK1
CDK2
CDK4
PKAP1
AKT
PKCa
PKCd
ERK
GSKp1
FYN
EPHB3
12
28
52
0.15
0.035
0.02
2
>2
>2
>2
>2
>2
>2
>2
>2
>2
>2
>2
>2
1.7
>2
2
>2
0.5
>2
>2
>2
>2
>2
>2
>2
0.34
>2
0.32
0.48
0.22
a See Ref. 20 for a description of the kinase assays.
Table 3. Inhibitory activity on various tumor cells for representative compoundsa
Compound
Tumor cell (IC50, lM)
SW480 MDA435
1.28
HCT116
H460a
SJSA1
RKO
12
28
46
52
2.53
1.14
1.58
2.41
1.82
1.04
ND
2.49
3.10
2.20
3.07
6.58
2.84
4.55
3.24
1.85
1.86
2.00
1.39
2.34
1.14
2.03
3.34
a See Ref. 20 for a description of the cellular anti-proliferation assays.

2138
S. Chen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2134–2138
The possibility of replacement of the thiazolinone moie-
ty (Fig. 1, 11, 13–15) was also investigated. However,
any attempt to replace it with other heterocyclics was
unsuccessful. Only oxazolinone retained some potency.
Acknowledgments
We thank Drs. Jefferson Tilley and Adrian Cheung for
critically reading the manuscript, and Drs. Nader Fo-
touhi and John Roberts for the helpful discussions dur-
ing this work.
Next, we examined the linker between the thiazolinone
and quinoline rings and found that the double bond is
required for activity. Replacement of the double bond
with a single bond resulted in a total loss of activity
(19). A similar result was obtained with the methyl sub-
stitution of the hydrogen at the olefin position (4b).
These results suggested that the unsubstituted olefin
linker is important for maintaining the configuration
of the thiazolinone and quinoline rings, needed for
CDK1 binding.
References and notes
1. Morgan, D. O. Nature 1995, 374, 131.
2. Sherr, C. J. Science 1996, 274, 1672.
3. Mani, S.; Wang, C.; Wu, K.; Francis, R.; Pestell, R. Exp.
Opin. Invest. Drugs 2000, 9, 1849.
4. Deshpande, A.; Sicinski, P.; Hinds, P. W. Oncogene 2005,
24, 2909.
5. Fischer, P. M.; Gianella-Borradori, A. Exp. Opin. Invest.
Drugs 2005, 14, 457.
6. Malumbres, M.; Hunt, S. L.; Sotillo, R.; Martin, J.;
Odajima, J.; Martin, A.; Dubus, P.; Ortega, S.; Barbacid,
M. Adv. Exp. Med. Biol. 2003, 532, 1 (New Trends in
Cancer for the 21st Century).
7. Tetsu, O.; McCormick, F. Cancer Cell 2003, 3, 233.
8. Ortega, S.; Prieto, I.; Odajima, J.; Martin, A.; Dubus, P.;
Sotillo, R.; Barbero, J. L.; Malumbres, M.; Barbacid, M.
Nat. Genet. 2003, 35, 25.
9. Berthet, C.; Aleem, E.; Coppola, V.; Tessarollo, L.;
Kaldis, P. Curr. Biol. 2003, 13, 1775.
10. Sherr, C. J.; Roberts, J. M. Genes Dev. 2004, 18, 2699.
11. Toogood, P. L. Med. Res. Rev. 2001, 21, 487.
12. Hirai, H.; Kawanishi, N.; Iwasawa, Y. Curr. Top. Med.
Chem. 2005, 5, 167.
13. Pevarello, P.; Villa, M. Exp. Opin. Ther. Patents 2005, 15,
675.
The selectivity of the most potent CDK1 inhibitors
was examined against a panel of kinases including
CDK2 and CDK4 (Table 2). CDK1 and CDK2 pro-
teins share a 66% identity and 84% similarity.21 Mod-
eling studies have shown that the ATP-binding sites of
these two kinases are very similar.22 There are only
two differences in their amino acid composition, at
positions 84 and 85, and the side chains at these
positions project outside of the ATP binding pocket.
Nevertheless, compounds 12, 28, and 52 show at least
10-fold selectivity for CDK1. The lack of CDK1 crys-
tal structure makes it difficult to explain the structural
basis for this selectivity. These compounds also
showed at least 10-fold selectivity against all other
tested human kinases (Table 2).
Four of the most potent compounds 12, 28, 46, and 52
were selected for evaluation of their in vitro cellular
activity against a panel of human cancer cell lines: colon
cancer (HCT116, SW480, RKO), lung (H460a), breast
(MDA435), and osteosarcoma (SJSA1). They showed
antiproliferative activity in all cell lines in the range of
1–7 lM IC50 (Table 3). We reported previously that
the antiproliferative activity of 28 is due to a reversible
block of cell cycle progression at the G2/M phase bor-
der.20 However, prolonged exposure to the compound
leads to induction of apoptosis and suggests that selec-
tive CDK1 inhibitors may have utility as anticancer
agents.20
14. Zhai, S.; Senderowicz, A. M.; Sausville, E. A.; Figg, W. D.
Ann. Pharmacother. 2002, 36, 905.
15. Senderowicz, A. M. Cancer Biol. Ther. 2003, 2(4 suppl. 1),
s84.
16. Fischer, P. M.; Gianella-Borradori, A. Exp. Opin. Invest.
Drugs 2003, 12, 955.
17. Misra, R. N. Drugs Future 2006, 31, 43.
18. Chu, X.-J.; DePinto, W.; Bartkovitz, D.; So, S.-S.; Vu, B.
T.; Packman, K.; Lukacs, C.; Ding, Q.; Jiang, N.; Wang,
K.; Goelzer, P.; Yin, X.; Smith, M. A.; Higgins, B. X.;
Chen, Y.; Xiang, Q.; Moliterni, J.; Kaplan, G.; Graves, B.;
Lovey, A.; Fotouhi, N. J. Med. Chem. 2006, 49, 6549.
19. DePinto, W.; Chu, X.-J.; Yin, X.; Smith, M.; Packman,
K.; Goelzer, P.; Lovey, A.; Chen, Y.; Qian, H.; Hamid, R.;
Xiang, Q.; Tovar, C.; Blain, R.; Nevins, T.; Higgins, B.;
Luistro, L.; Kolinsky, K.; Felix, B.; Hussain, S.; Heim-
brook, D. J. Mol. Cancer Ther. 2006, 5, 2644.
20. Vassilev, L. T.; Tovar, C.; Chen, S.; Knezevic, D.; Zhao,
X.; Sun, H.; Heimbrook, D. C.; Chen, L. Proc. Natl. Acad.
Sci. U.S.A. 2006, 103, 10660.
In conclusion, we have identified the quinolinyl-meth-
ylene-thiazolinone class as potent and selective CDK1
inhibitors. SAR studies indicated that the quinoline
and thiazoline moieties are important for binding,
while the amine position can tolerate more modifica-
tions. The antiproliferative and pro-apoptotic activity
of these compounds suggests that members of this
class could be developed as agents for cancer
therapy.
21. McGrath, C. F.; Pattabiraman, N.; Kellogg, G. E.;
Lemcke, T.; Kunick, C.; Sausville, E. A.; Zaharevitz, D.
W.; Gussio, R. J. Biomol. Struct. Dyn. 2005, 22, 493.
22. Furet, P.; Zimmermann, J.; Capraro, H.-G.; Meyer, T.;
Imbach, P. J. Comput. Aided Mol. Des. 2000, 14, 403.