Identification of novel, selective and potent Chk2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.09.067

A series of isothiazole carboxamidine compounds were synthesized and discovered as novel and selective inhibitors for Chk2. They are not active against the related Chk1 kinase. The structure-activity relationship studies were performed on the scaffold, an

Full text of DOI:10.1016/j.bmcl.2006.09.067

Bioorganic & Medicinal Chemistry Letters 17 (2007) 172–175
Identification of novel, selective and potent Chk2 inhibitors
Gary Larson, Shunqi Yan, Huanming Chen, Frank Rong, Zhi Hong and Jim Zhen Wu*
Drug Discovery, Valeant Pharmaceutical Research and Development 3300 Hyland Avenue, Costa Mesa, CA 92626, USA
Received 28 August 2006; revised 21 September 2006; accepted 21 September 2006
Available online 10 October 2006
Abstract—A series of isothiazole carboxamidine compounds were synthesized and discovered as novel and selective inhibitors for
Chk2. They are not active against the related Chk1 kinase. The structure-activity relationship studies were performed on the scaf-
fold, and enzymatic kinetic analysis showed they are simple ATP competitive inhibitors with Ki values as low as 11 nM for Chk2.
Computer modeling studies were employed to comprehend the mechanism of action and SAR of these compounds.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Genetic material is exposed to many environmental
stresses such as radiation, ultraviolet light, and chemi-
cals that could damage DNA. Unrepaired DNA damag-
es in cells will accumulate erroneous genetic material
and eventually lead to tumorigenesis or apoptosis. With-
out the activation of DNA damage-signaling transduc-
tion pathways and cell cycle checkpoints, cells would
not have the chance to make the necessary repairs and
maintain genetic stability.
The crucial roles played by both Chk1 and Chk2 in
mediating cellular responses to DNA damages implicate
potential pharmacological value of their inhibitors in
cancer therapy. Addition of a Chk1 inhibitor to a
DNA damaging chemotherapy drug has been shown
to sensitize tumor cells to chemotherapy and enhance
drug’s cytotoxic effect.5 Conversely, selective inhibition
of Chk2 has been reported to protect normal cells from
DNA damage caused by ionizing radiation and enhance
5,6
the efficacy of genotoxic agents in tumor cells.
Since
Checkpoints are the fail-safes to ensure the cell cycle not
to progress to the next stage until the previous step is
completed.1 The protein kinases, checkpoint kinase 1
(Chk1) and 2 (Chk2), are the major effectors of the rep-
lication checkpoint.2 DNA damage can activate signal
transduction pathways that lead to cell cycle arrest or
apoptosis depending on the type and severity of the
damage. This signaling network is initiated by the cen-
tral sensors of DNA damage known as ATR and
ATM protein kinases.3 ATR generally responds to UV
damage and chemical-induced stall of DNA replication
forks, and subsequently promotes the degradation of
cdc25 by activating Chk1. ATM similarly initiates the
degradation of cdc25 but does so in response primarily
to the double strand breaks in the DNA.4 Activated
ATM phosphorylates both Chk1 and Chk2, which in
turn phosphorylate downstream cdc25A, cdc25C,
BRCA1, and p53, and lead to cell cycle arrest at G1/S
and G2/M checkpoints or apoptosis.
nearly half of the cancer patients undergo radiotherapy
and they all suffer from the side effects caused by radia-
tion damage to the sensitive tissues, a small molecular
Chk2 inhibitor may be beneficial in radiotherapy for
an improved safety window.
To date, a few Chk inhibitors have been discovered and
revealed in the literature. They include a broad spectrum
protein kinase inhibitor staurosporine and its analogs
such as UCN-01, Go6979, and isogranulatimide.7 Selec-
tive Chk inhibitors are rare but have been reported.
4-(Aminoalkylamino)-3-benzimidazole quinolines are a
series of specific Chk1 inhibitors.8 A series of 2-aryl-
benzimidazole compounds were published recently as
selective and potent inhibitors for Chk2.9 By screening
against a protein kinase target with our chemical library,
we found a series of isothiazole carboxamidine com-
pounds that exhibited potent protein kinase inhibitory
activity.10 Some of them possess potent selective inhibi-
tory activity against Chk2 in the protein kinase specific-
ity studies. Here, we describe the SAR of this series of
compounds against Chk2, their enzymatic inhibition
kinetics, and computer model analysis of their mode
of action. A Chk2 enzymatic assay was developed to
determine IC50 values of these compounds. The assay
Keywords: Protein kinase inhibitor; Chk2 inhibitor; Chk; Chk2; DNA
damage.
*
Corresponding author. Tel.: +1 714 545 0100x3024; fax: +1 714 668
3142; e-mail: jwu@valeant.com
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.09.067

G. Larson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 172–175
CN
173
cophore in the scaffold. These simple SAR exercises sug-
gest that these compounds have tractable SAR, which
prompted us to perform a more systematic SAR study
of the scaffold. As summarized in Tables 1–3, three ser-
ies of compounds, A, B, and C, were synthesized with
modifications focused on the substitution R1 and R2,
the linker X between two phenyl rings, and the left phen-
yl ring replacement. Their activities against Chk2 were
determined.
H
N
NH2
NH2
NCS
(b)
(a)
S
O
R
R
R
(c)
R1
NH
HN
CN
H
H
(d)
N
OH
N
OH
N
N
S
S
R
R
As illustrated in Table 1, compounds in series A with a
NH bridge linker between two phenyls consistently dem-
onstrated higher inhibitory activity against Chk2 than
Scheme 1. Reagents and conditions: (a) CSCl2, aq K2CO3, CHCl3, rt,
2 h; (b) cyanoacetamide, KOH, DMF, rt, 16 h; (c) Br2, EtOAc, rt,
1–2 h; (d) R1–NH2, EtOH, 90 ꢁC, 16 h.
Table 2. The enzymatic IC50 values of series B compounds against
Chk2 (in lM)
conditions are described in the notes.14 The chemical
synthesis of these compounds was based on the pub-
lished protocol10 and is outlined in Scheme 1.
R1
NH
OH
HN
H
The original hit against Chk2 from the protein kinase
specificity analysis, 3-hydroxy-N-isopropyl-5-(4-phen-
oxy-phenylamino)isothiazole-4-carboximidamide (16),
has an IC50 of 2.3 lM (Table 2). Preliminary SAR stud-
ies by replacing the bridge oxygen between two phenyl
rings in 16 with an amine resulted in 7 with a 5-fold
improvement of IC50 (Table 1). Further addition of a
bromo to the para position of the left phenyl ring of 7
yielded 1 with an IC50 of 87 nM. Moreover the displace-
ment of the carboxamidine by a cyano group totally
abolished a compound’s activity, underscoring the
importance of carboxamidine as a part of key pharma-
N
S
N
H
X
B
Compound
R1
X
IC50
8.0
14
–CH2–
–N@N–
–O–
15
16
17
18
11
2.3
>40
37
–SO2–
–CH2–
Table 1. The enzymatic IC50 values of series A compounds against
Chk2 (in lM)
R1
19
–N@N–
11
NH
OH
HN
H
20
21
–S–
19
N
N
S
R2
–CO–
5.0
N
H
A
Compound R1
R2 IC50 Compound R2 IC50
Table 3. The enzymatic IC50 values of series C compounds against
Chk2 (in lM)
1
2
3
4
Br 0.087
Br 0.14
Br 0.22
Br 0.30
7
8
H
H
H
H
0.42
0.46
1.0
HO
HO
R1
NH
OH
HN
9
H
N
S
N
10
1.8
R2
C
HO
Compound
R1
R2
IC50
19
5
6
Br 0.31
Br 0.47
11
12
H
H
1.2
22
HO
HO
N
N
23
24
25
40
0.90
N
N
HO
HO
>40
17
13
H
0.46
HO

174
G. Larson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 172–175
those in series B. Most of IC50 values in series A are un-
der 1 lM, and they tolerate the different R1 substitu-
tions. For example, the addition of an extra hydroxyl
group to R1 slightly changes IC50 of 1 from 0.087 to
0.14 lM compared to 2. Removal of one methyl group
of R1 from 2 only reduces 1.5-fold of potency in 3. A
longer and more flexible hydroxyl (5) or diol R1 (6) low-
ers the potency by a few folds. Meanwhile, it seems that
a more hydrophobic R1 such as in 4 is tolerable but not
favorable. On the other hand, IC50 values appear to be
more sensitive to the substitution R2 on the left side of
the molecules. An unsubstituted R2 (7) is 5-fold less po-
tent than a bromo substituted 1. Within the same series
of compounds with R2 = H (7–12), their inhibitory
activities range from 0.42 to 1.8 lM, and they are con-
sistently 3- to 5-fold less active than their bromo substi-
tuted counterparts. Nevertheless, both R1 and R2
substitutions are relatively unimportant in determining
their inhibitory activity against Chk2. The most potent
compound from this series is 1, which has an IC50 of
87 nM.
200
150
100
50
0
0
5
10
15
20
1/ATP (10-3)
Figure 1. Lineweaver–Burk analysis of inhibition kinetics of 2 against
Chk2. The study was performed at various concentrations of 2 (0 ( ),
•
25 (s), 50 (j), 100 (h), and 200 (m) nM) with ATP concentrations
varied from 0 to 200 lM. The double reciprocal plots were fitted best
to a simple competitive inhibition model.
exhibit some weak activity against two other protein
kinases, the IC50s are more than 20-fold higher than that
for Chk2.
The effect of the linker atom X between the two phenyl
rings was investigated in series B. It was found that any
change from amine linker is detrimental to the activity
(Table 2). A methylene linker in 14 and 18 resulted in
a 20-fold loss of IC50 in comparison to the similar
NH-linked 7 and 10. Other linkers such as oxygen in
16, thio in 20, or –N@N– in 15 and 19 also yielded com-
pounds with less potency. A sulfone substitution in 17
completely abolishes activity against Chk2.
To further comprehend the molecular mechanism of
how these compounds inhibit Chk2 and why their activ-
ities differ so modestly for various R1 and R2 substitu-
ents but so radically for any change on the linker X,
we performed computer modeling studies. As the previ-
ous kinetic analysis indicated, they are ATP competitive
inhibitors and they likely bind directly to the ATP bind-
ing pocket of Chk2. Using a recently solved Chk2
co-crystal structure with a debromohymenialdisin inhib-
itor as a template,13 compound 7 was docked into the
ATP binding site of Chk2 (PDB code: 2CN8) using the
Glide docking program.12 All water molecules around
the ATP site were removed and hydrogen atoms were
added using Pprep of Glide. Grids for Glide docking were
calculated using the bound ligand as the reference of
binding site in the Chk2 protein. The Glide docking
was performed using extra-precision mode with up to 10
poses saved for analysis. The top-scored pose is depicted
in Figure 2.
We further assessed the importance of the left phenyl
ring and the linker effect in the series C of compounds.
As summarized in Table 3, a direct attachment of R2
to the middle phenyl ring gave rise to either very weak
or totally inactive compounds regardless of the nature
of R1 and R2. For example, compounds 22 and 25 with
a cyclohexyl linked directly to phenyl have IC50 values in
double digit lM range. Compound 23 with R2 of 1H-
pyrazole is only marginally active. A combination of
R1 of 2-isopropyl-1-ol- and R2 of 1H -pyrazole in 24
completely eliminates any activity against Chk2.
To understand the action mechanism of these inhibitors,
we picked a representative compound 2 for Chk2 inhibi-
tion kinetic analysis. Using a radioactive-based filter
14
binding assay
with the assay conditions highlighted
in Figure 1, we determined this compound’s inhibition
kinetics as simple competitive with ATP (Fig. 1). The
inhibition constant Ki was 11 nM. This result suggests
that these compounds likely act by directly binding to
the ATP site of Chk2. Because they are ATP competi-
tive, the IC50 values reported here are highly influenced
by the ATP concentration in the assay. In our assay, a
relatively high concentration of ATP at 100 lM was
used, in which the IC50 for 2 was 140 nM. Decreased
ATP concentration of 10 lM resulted in a much lower
IC50 of 43 nM. In a similar Chk1 enzymatic assay, the
IC50 of 2 was more than 40 lM, suggesting that 2 is a
selective Chk2 inhibitor. Moreover, we chose compound
8 for protein kinase inhibition specificity studies using
Upstate’s KinaseProfilerTM assays. Although it did
Figure 2. The predicted binding mode of 7 in the ATP site of Chk2
(PDB code: 2CN8).

G. Larson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 172–175
175
The predicted pose of 7 in the ATP binding site reveals a
number of key hydrogen bonds and hydrophobic inter-
action of Chk2 with the inhibitor. Both R1 and R2 are
located in the vicinity of solvent-exposed area. Specifi-
cally, the NH atoms of the protonated amidine group
interact extensively with side chains of Chk2 by hydro-
gen bonds. For example, one of its NH is engaged in
the hydrogen bonding with OH of Thr367, while the
ATP competitive inhibitors with a tractable SAR. They
possess cellular activity to regulate the Chk2 mediated
cell cycle arrest and apoptosis. They may be useful as
a radiation protection agent in anticancer radiotherapy.
Acknowledgments
À
other one forms a hydrogen bond with CO2 of
Glu308 (Fig. 2). The same CO2 group also makes a
The authors want to acknowledge Dr. D. Delia of Insti-
tuto Nazionale Tumori, Italy, for helping biological
characterization and Dr. D. Smith for initially evaluat-
ing compounds.
À
hydrogen bond with NH located between A and C rings
(Fig. 2). The NHR1 in amidine group forms a hydrogen
bond with C@O of side chain of Asn352, and the OH
group on the isothiazole ring C hydrogen bonds with
+
NH3 of Lys249. These hydrogen bonds are similar to
References and notes
the contacts made by the ribose and phosphate groups
of bound ADP in a typical protein kinase-ADP complex
structure, and are almost identical to those hydrogen
bonds between Chk2 and debromohymenialdisin inhib-
itor reported in the literature.13 More importantly, the
NH linker between A and B rings interacts with the
backbone –C@O of Met304 by a hydrogen bond, which
is the only interaction between the inhibitor and the
hinge area of Chk2 (Fig. 2). Additionally, the phenyl
rings in the isothiazole series of inhibitors seem to make
extensive hydrophobic contacts with the side chains of
Leu226 and Leu354.
1. Kastan, M. B.; Bartek, J. Nature 2004, 432, 316.
2. Bartek, J.; Lukas, J. Cancer Cell 2003, 3, 421.
3. Bakkenist, C. J.; Kastan, M. B. Nature 2003, 421, 499.
4. Buscemi, G.; Perego, P.; Carenini, N.; Nakanishi, M.;
Chessa, L.; Chen, J.; Khanna, K.; Delia, D. Oncogene
2004, 23, 7691.
5. Zhou, B. B.; Bartek, J. Nat. Rev. Cancer 2004, 4, 216.
6. Pommier, Y.; Sordet, O.; Rao, V. A.; Zhang, H.; Kohn,
K. W. Curr. Pharm. Des. 2005, 11, 2855.
7. Collins, I.; Garrett, M. D. Curr. Opin. Pharmacol. 2005, 5,
366.
8. Ni, Z. J.; Barsanti, P.; Brammeier, N.; Diebes, A.; Poon,
D. J.; Ng, S.; Pecchi, S.; Pfister, K.; Renhowe, P. A.;
Ramurthy, S.; Wagman, A. S.; Bussiere, D. E.; Le, V.;
Zhou, Y.; Jansen, J. M.; Ma, S.; Gesner, T. G. Bioorg.
Med. Chem. Lett. 2006, 16, 3121.
9. Arienti, K. L.; Brunmark, A.; Axe, F. U.; McClure, K.;
Lee, A.; Blevitt, J.; Neff, D. K.; Huang, L.; Crawford, S.;
Pandit, C. R.; Karlsson, L.; Breitenbucher, J. G. J. Med.
Chem. 2005, 48, 1873.
10. Varaprasad, C. V.; Barawkar, D.; El Abdellaoui, H.;
Chakravarty, S.; Allan, M.; Chen, H.; Zhang, W.; Wu, J.
Z.; Tam, R.; Hamatake, R.; Lang, S.; Hong, Z. Bioorg.
Med. Chem. Lett. 2006, 16, 3975.
11. Carlessi, L.; Buscemi, G.; Larson, G.; Hong, Z.; Wu, J.Z.;
Delia, D. Mol. Cancer Ther., submitted for publication.
12. www.schrodinger.com; Schrodinger, New York.
13. Oliver, A. W.; Paul, A.; Boxall, K. J.; Barrie, S. E.;
Aherne, G. W.; Garrett, M. D.; Mittnacht, S.; Pearl, L. H.
EMBO J. 2006, 25, 3179.
These modeling results support our SAR findings.
Docking results predict that both R1 and R2 are near
solvent accessible area, suggesting the moderate effect
of both R1 and R2 substitution on the inhibitory
activity. This is in good agreement with experimental
results summarized in the three tables. More impor-
tantly when the linker X = NH is altered, the sole
hydrogen bond between the inhibitor and the hinge
area of Chk2 is disrupted. Such a hydrogen bond
interaction is believed to be crucial to achieve substan-
tial affinity between a protein kinase and an inhibitor.
This insight agrees well with our observations that
compounds with various other linkers tend to be
weaker or totally inactive due to the loss of NH linker
at the position.
14. Chk1 and Chk2 kinase activities were assayed using
recombinant human GST-CHK1 and GST-CHK2 pro-
teins from Upstate (Lake Placid, NY). Briefly, 10 nM of
CHK1 or CHK2 was used to phosphorylate 25 lM myelin
basic protein (MBP) (Invitrogen, Carlsbad, CA) in a
buffer containing 8 mM MOPS, pH 7.2, 10 mM b-glycerol
phosphate, 1.5 mM EGTA, 0.4 mM EDTA, 0.4 mM
sodium ortho-vanadate, 100 lM ATP, 1 lCi [c-33P]ATP,
15 mM MgCl2, 0.4 mM DTT, 0.006% Brij-35, 1% glycerol,
and 0.2 mg/ml BSA in a final volume of 25 ll. The reaction
ran for 30 min at 24 ꢁC and was quenched by adding
100 ll of 1% trichloroacetic acid. The quenched solution
was subsequently transferred to a 96-well white GF/B filter
plate (Perkin-Elmer, Wellesley, MA) using a Perkin-Elmer
Filtermate Universal Harvester. The radioactivity that was
incorporated into MBP was trapped on the filter plate and
counted using a Perkin-Elmer TopCount.
Because of the potency and selectivity our Chk2 inhibi-
tors had achieved, we chose one representative com-
pound
2
for further biochemical and cellular
characterization. This compound was shown to suppress
the ionizing radiation-induced activation of Chk2 in
cells and prevented the radiation-induced Chk2-depen-
dent degradation of HDMX protein, a negative regula-
tor of p53. More importantly, the compound was
capable of attenuating radiation induced cell apoptosis,
suggesting that it does possess the radiation protective
effect as we hoped. Detailed biological studies of this
compound will be published elsewhere.11
In summary, a series of novel and selective Chk2 inhib-
itors were synthesized and discovered. They are simple