Synthesis and evaluation of triazolones as checkpoint kinase 1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.01.043

Checkpoint kinase 1 (Chk1, CHEK1) is a Ser/Thr protein kinase that plays a key role in mediating the cellular response to DNA-damage. Synthesis and evaluation of a previously described class of Chk1 inhibitors, triazoloquinolones/triazolones (TZs) is further described herein. Our investigation of structure-activity relationships led to the identification of potent inhibitors 14c, 14h and 16e. Key challenges included modulation of physicochemical properties and pharmacokinetic (PK) parameters to enable compound testing in a Chk1 specific hollow fiber pharmacodynamic model. In this model, 16e was shown to abrogate topotecan-induced cell cycle arrest in a dose dependent manner. The demonstrated activity of TZs in this model in combination with a chemotherapeutic agent as well as radiotherapy validates this series of Chk1 inhibitors. X-ray crystal structures (PDB code: 2YEX and 2YER) for an initial lead and an optimized analog are also presented.

Full text of DOI:10.1016/j.bmcl.2012.01.043

Bioorganic & Medicinal Chemistry Letters 22 (2012) 2330–2337
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and evaluation of triazolones as checkpoint kinase 1 inhibitors
Vibha Oza a, , Susan Ashwell a, Patrick Brassil a, Jason Breed b, Jaychandran Ezhuthachan a, , Chun Deng a,
⇑
Michael Grondine a, Candice Horn a,à, DongFang Liu a,§, Paul Lyne a, Nicholas Newcombe b, Martin Pass b,
Jon Read b, Mei Su a, Dorin Toader a, Dingwei Yu a, Yan Yu a, Sonya Zabludoff a,–
a AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA
b AstraZeneca R&D Alderley Park, Cheshire, Macclesfield SK10 4TG, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Checkpoint kinase 1 (Chk1, CHEK1) is a Ser/Thr protein kinase that plays a key role in mediating the cel-
lular response to DNA-damage. Synthesis and evaluation of a previously described class of Chk1 inhibi-
tors, triazoloquinolones/triazolones (TZs) is further described herein. Our investigation of structure–
activity relationships led to the identification of potent inhibitors 14c, 14h and 16e. Key challenges
included modulation of physicochemical properties and pharmacokinetic (PK) parameters to enable com-
pound testing in a Chk1 specific hollow fiber pharmacodynamic model. In this model, 16e was shown to
abrogate topotecan-induced cell cycle arrest in a dose dependent manner. The demonstrated activity of
TZs in this model in combination with a chemotherapeutic agent as well as radiotherapy validates this
series of Chk1 inhibitors. X-ray crystal structures (PDB code: 2YEX and 2YER) for an initial lead and an
optimized analog are also presented.
Received 31 October 2011
Revised 12 January 2012
Accepted 13 January 2012
Available online 1 February 2012
Keywords:
Triazolones
CHK1
Checkpoint kinase 1
In vivo activity
Ó 2012 Elsevier Ltd. All rights reserved.
Overexpression of various protein kinases and rapid cell divi-
sion are recognized as two of the hallmarks of cancer.1 Thus, many
kinases that are involved in the regulation of the cell cycle are con-
sidered promising therapeutic targets for the treatment of cancer
including, the checkpoint kinases-1 and -2.2 Checkpoint kinase 1
(Chk1, CHEK1) is a Ser/Thr kinase that is involved in mediating
the cellular response to DNA-damage. In response to DNA-damage,
Chk1 is phosphorylated and hence activated by the ATM and ATR
kinases. Chk1 in turn phosphorylates a phosphatase, Cdc25C,
which then inactivates Cdc2 (Cdk1) resulting in cell cycle arrest
allowing the cell an opportunity to repair damaged DNA prior to
entering into mitosis. Abrogation of Chk1 mediated cell cycle
checkpoints results in the cellular population bearing damaged
DNA, entering mitosis prematurely and undergoing mitotic catas-
trophe and cell death.3 Thus when administered with a DNA dam-
aging agent such as standard cytotoxics or radiation therapy, a
Chk1 inhibitor has the ability to enhance its activity. Consequently,
Chk1 inhibitors have been highly sought after as chemo or radio-
potentiation agents.3 Triazolones (TZs) and thiophene carboxamide
ureas (TCUs) were identified from our HTS campaign against
Chk1.4 Using structure based drug design and a library synthesis
approach, these HTS hits were transformed into leads having
acceptable cellular potency, an activity that was lacking in the ori-
ginal hits, although both series were somewhat potent against the
Chk1 enzyme.4a,4b
Through structure based drug design, more potent analogs (4,
5a, 5b) of HTS hit 3 were identified. A key feature of the hit-to-lead
efforts in the TZ series included heteroaromatic substitution on the
triazolone ring which was found to be critical for cellular poten-
cy.4a Activities of TCU and TZ HTS hits and leads are summarized
in Figure 1.
In this publication, we wish to report additional structure–
activity relationships (SAR) generated around lead 5a which en-
abled us to identify compounds suitable for evaluation in vivo in
a hollow fiber pharmacodynamic (PD) model.4c Key SAR avenues
explored included replacement of the methyl group with alternate
functional groups and derivatization of the heteroaryl ring in order
to improve potency as well as to modulate physicochemical
properties.
To access a variety of analogs, we designed an efficient synthetic
route utilizing the common intermediate, 9, namely ethyl 8-bro-
mo-1-oxo-1,2-dihydro[1,2,4]triazolo[4,3-a]quinoline-5-carboxyl-
ate as shown in Scheme 1.
⇑
Corresponding author. Tel.: +1 781 839 4448; fax: +1 781 839 4630.
E-mail addresses: Vibha.Oza@astrazeneca.com (V. Oza), jayan_e@hotmail.com (J.
Ezhuthachan), candice_horn@excite.com (C. Horn), dfliu@msn.com (DongFang Liu),
zabludoff@gmail.com (S. Zabludoff).
Reaction of commercially available 6-bromoisatin (6) with
malonic acid in a sodium acetate-acetic acid buffer system yielded
7. Fisher esterification of 7, as detailed by Buchman and co-work-
ers,5 generated 8 which was in turn reacted with POCl3 to generate
the corresponding chloroquinoline, 9. The reaction of 9 with ethyl
Present address: MA, USA.
Present address: Plainfiled, IN, USA.
Present address: Unionville, CT, USA.
Present address: San Diego, CA, USA.
à
§
–
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.01.043

V. Oza et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2330–2337
2331
H
N
Thiophene carboxamide ureas (TCUs)
HN
H
2 N
O
O
N
N
N
S
S
O
NH2
H
N
2
O
O
1, HTS Hit, IC50 0.15 µM
Triazolones (TZs)
2, TCU Lead, IC50 0.007 µM
EC50 0.02 µM
N
N
R
N
N
N
N
N
H
O
N
N
H
O
H
O
O
3, HTS Hit
IC50 0.8 µM
4, Potent Lead with no cell actvity
IC50 0.01 µM
Lead with cellular activity
H
IC50 0.0001 µM
N
5a, R =
EC50 0.51 µM
S
5b, R=
IC50 0.02 µM
EC50 2.0 µM
Figure 1. TCU and TZ Class of Chk1 inhibitors.
O
O
O
OH
O
b
a
O
Br
N
O
N
Br
N
O
Br
H
H
H
8
6
7
O
O
O
O
d
c
Br
N
H
O
Br
N
N
N
9
10
H
O
Scheme 1. Synthesis of ethyl 8-bromo-1-oxo-1,2-dihydro[1,2,4] triazolo [4,3-a]quinoline-5-carboxylate. Reagents and conditions: (a) malonic acid (1.1 equiv) acetic acid
(2.5 mL), sodium acetate (1.25 equiv), rt, 5 h then sodium acetate (1.25 equiv), 100 °C, overnight, 88%; (b) H2SO4, EtOH (abs, 1:1, 0.5 M), reflux, 45 min, 86%; (c) POCl3, P2O5,
reflux, 1.5 h, 76%; (d) ethyl carbazate (1.2 equiv), HCl/dioxane (4 M, cat.), EtOH (abs).
carbazate in the presence of absolute ethanol and a catalytic
amount of 4 M HCl in dioxane at 165 °C for 20 min under micro-
wave reaction conditions generated the key triazolone intermedi-
ate, 10, in high yield.
protecting groups were removed using either TFA in DCM or 4 M
HCl in 1,4-dioxane.
Analog design for TZs was based on the information generated
from a co-crystal structure of 5a with Chk1.4a As shown in Fig-
ure 2a, the structure of 5a with Chk1 protein revealed an altered
binding mode with respect to the HTS hit 3 which highlighted
the opportunity to explore SAR at the 4-methyl position and poten-
tially access additional interactions in the region of the sugar bind-
ing pocket or p-loop of Chk1 kinase.
Intermediate 10 was used to generate a number of analogs as
summarized in Scheme 2. Saponification of 9 by LiOH generated
11. The latter was used to generate both secondary and tertiary
amides by carrying out HATU mediated coupling reactions. Acid
11 was also subjected to Curtius re-arrangement to generate the
corresponding Boc-protected 4-aminotriazolone which upon acidic
treatment gave the free amine. LiAlH4 reduction of the ester moiety
on 9 led to the generation of 12 which was easily converted to pro-
vide 1°, 2° or 3° amines by in situ generation of the mesylate fol-
lowed by reaction with the amine of choice at 0 °C. Analogs 13,
14, 15 and 16 or their Boc-protected precursors were generated
by microwave assisted Suzuki coupling reactions on the respective
penultimate bromo intermediates. Finally, where necessary, Boc
As evidenced by crystal structures of a series of TCUs,4b the
vicinity of the sugar binding pocket could be reached by attaching
an amide side chain bearing a basic amine. In the TCU series, we
had shown that with certain substitutions, it was possible to make
a salt bridge interaction with Asp-148 and charge-dipole interac-
tions with the side chain of Asn-135 and the Glu-134 main chain
carbonyl oxygen. An overlay of the crystal structures of 2 and 5a,
shown in Figure 2b, identified a hybridization design opportunity

2332
V. Oza et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2330–2337
OH
O
O
O
OH
a
b
Br
N
Br
N
N
N
Br
N
N
N
N
N
H
O
H
O
10
H
O
12
11
e, g*
d, e
f, g*
c
OH
NR1R2
O
NR1R2
NH2
N
R
N
R
N
R
N
N
N
N
H
N
N
H
R
N
N
N
H
O
16
O
15
O
H
O
13
14
Scheme 2. Synthesis of TZ analogs. Reagents and conditions: (a) LiOHÁH2O (1 equiv), THF–MeOH–H2O (1:1:1, 1.3 M), rt, 2 h., 64%; (b) LiAlH4 (1 M in THF, 1 equiv), THF
(0.23 M), rt, 30 min, 50%; (c) HATU (1.5 equiv), DMF (0.39 M), DIEA (5 equiv), amine substrate (1.5 equiv), rt, 1 h, then H2O, boronic acid (1.1 equiv), Cs2CO3 (4 equiv),
Pd(PPh3)4 (7 mol %), microwave, 165 °C, 10 min; (d) t-BuOH (0.2 M), (Ph)3P(O)N3 (1.09 equiv), reflux, 5 h, concentrate then TFA–DCM (5%), DCM (10 mL), 1 h., rt, 72.4%; (e)
RB(OH)2 (1.1 equiv), Cs2CO3 (4 equiv), Pd(PPh3)4 (7 mol %), dioxane–H2O (4:1 mL), microwave, 165 °C, 10 min; (f) TEA (2.2 equiv), NMP (0.4 M), MeSO2Cl (1.1 equiv), 0 °C,
30 min, warm to rt, 30 min, then amine substrate (1.1 equiv), 10 min, rt; (g) boronic acid (1.1 equiv), Cs2CO3 (4 equiv), TEA (4 equiv), Pd(PPh3)4 (10 mol %), microwave, 165 °C,
⁄
1200 s; (h) TFA–H2O (9:1), rt. Step g applied when BOC deprotection was required to yield the final compound).
at the 4-CH3 position of 5a. The hypothesis generated from this
overlay was that replacement of the 4-CH3 with a substituted
amide moiety like that depicted in TCU 2 would enable the substi-
tuted TZ 5a to pick up additional interactions with Asp-148,
Asn-135 side chains and Glu-134 backbone of Chk1. In order to test
this hypothesis, a set of 4-amides bearing basic side chains were
designed and synthesized using a library format.
modifying the 3-thienyl group resulted in the generation of 13n
and 13o. While such substitutions helped retain enzyme potency
and offered about a 10-fold increase in solubility, the compounds
were found to lose activity in the cellular assays (>30 lM). Reasons
for this were not investigated but may possibly be ascribed to the
lack of permeability due to the dibasic nature of these compounds.
The SAR trend evident from the data was that the presence of
the 2-aminoethyl moiety was key for activity against the Chk1 en-
zyme. Extension of the 2-aminoethyl group as in the case of 13h or
replacement of the terminal amine by a more neutral moiety such
as the 2-hydroxyethyl group as depicted in 13e and 13g resulted in
about a 10–20-fold drop in enzyme potency.
Most surprising from this set of analogs was the data obtained
for 13a and 13b. The hybridization design hypothesis generated
from the overlay of 5a with the TCU lead 2 shown in Figure 2b sug-
gested that replacement of the 4-CH3 position with the 2-aminopi-
peridine moiety would enhance interaction with the sugar binding
pocket region via H-bonding of the piperidine nitrogen with the
Asp-148 and/or Asn-135 side chains. While such SAR held true in
the case of the TCU series resulting in more potent compounds4b
(e.g. 2), in the case of triazolones we found that, 13a and 13b
exhibited a drop of several fold in the enzyme potency when com-
pared to 5a and 5b.
It is also noteworthy that while the amide moiety featured in
13a and 13b was instrumental in greatly improving cellular po-
tency for compounds in the TCU series, such a modification failed
to improve the cellular activity of inhibitors from the TZ series. This
snapshot of SAR did, however, highlight several promising com-
pounds, namely 13d, 13i and 13j, exhibiting a reasonable balance
of potency and solubility. We also explored a series of reversed
amides of these key compounds and found them to be equipotent
with their amide matched pairs. While not a general trend, we
were pleased to find that 14b shown in Table 2, and 13j, showed
about a five-fold increase in solubility while maintaining potency
in both the enzyme and cellular assays. Interestingly, during the
synthesis of 14b, routine screening identified the intermediate
14a as being active in the enzyme assay. This serendipitous discov-
ery led us to generate the additional 4-aminotriazolones 14c–14k
(Table 2). The heteroaryl groups incorporated in these analogs
were selected based on previously identified SAR trends.4a
Since the desired route of delivery of the Chk1 inhibitor in the
clinical setting was to be intravenous (iv), the presence of a basic
amino group was deemed advantageous in order to afford better
solubility. In the hit-to-lead exploration efforts, several key hetero-
cyclic groups were identified and these were retained in the gener-
ation of this library of compounds. Both cyclic and acyclic amides
containing either a 1°, 2° or 3° terminal basic amine or a neutral
alcohol side chain were generated. The compounds were profiled
in a Chk1 kinase assay.4c The cellular activity, represented as
EC50(CPT), was measured as a function of the ability of the com-
pound to abrogate cell cycle arrest at the G2/M checkpoint in
HT29 tumor cells. The checkpoint is induced by pre-treatment of
the cells with the DNA damaging agent, camptothecin (CPT) for
2 h prior to exposing them to the putative Chk1 inhibitor. The abil-
ity of a compound to abrogate this cell cycle arrest is quantified by
measuring the level of phosphohistone-H3 (a marker of mitosis-M
phase).4c Cellular activity data generated for cells with no
prior treatment with CPT was also recorded as EC50 (alone) to iden-
tify non-Chk1 mediated cell cycle affects arising from the
compound.
In addition periodic spot checks were carried out against Cdk1
as this activity would induce cell cycle arrest, and therefore mask
the cellular read out from abrogation of the G2/M checkpoint. Test-
ing of representative TZ analogs in the Cdk1 assay revealed that
none of the compounds possessed Cdk1 activity. A survey of the
key compounds summarized in Table 1 revealed that the strategy
of generating amides bearing a basic amino moiety enabled us to
maintain a high degree of enzymatic potency, and we were also
able to successfully identify compounds that exhibited good solu-
bility such as, 13d and its 3-thienyl matched pair, 13j, both bearing
N,N-dimethylammonium ethyl solvent tails. Unfortunately, cellu-
lar potency was not generally maintained, where limited enzy-
matic activity was seen. Additional efforts to functionalize 13j by

V. Oza et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2330–2337
2333
3-thienyl ring with basic groups resulted in compounds with in-
creased solubility. A particularly noteworthy example of this is
14e which has a solubility of over 50 lM and is also active in the
cellular assay. This is, again, an example of a modification that re-
tained cellular activity in the 4-amino subseries compared to the
previously published 4-methyl subseries.4a This data provided fur-
ther insight into how to improve cellular potency in the TZ class of
Chk1 inhibitors. Additionally, these findings helped us to identify
key compounds, namely, 14c and 14h with good cellular potency
and reasonable PK sufficient to allow evaluation in the PD model
even though these very potent compounds were poorly soluble
(less than 5 lM). In general, there appeared to be opposing trends
between solubility and target potency in this class of Chk1 inhibi-
tors, with modifications resulting in either potent compounds with
poor solubility or moderately potent compounds with good solubil-
ity and poor PK. Next, we investigated incorporating solubilizing
groups in place of the 4-amino group resulting in the generation
of the 4-aminomethyltriazolones summarized in Table 3.
Modification of 4-amino to 4-aminomethyl seemed a viable
strategy to take in order to increase solubility by increasing the ba-
sicity of the amine moiety. While there is a delicate balance be-
tween solubility and target potency for triazolones in general,
this seemed like a reasonable avenue to explore given the very po-
tent nature of the 4-aminotriazolones, particularly 14c and 14h.
We sought to synthesize the desired compounds in a library for-
mat where intermediate 11 was converted to the corresponding
amine in one pot via generation of the mesylate at 0 °C. Application
of Suzuki coupling conditions to the resultant bromo substituted
amine and an appropriate boronic acid followed by Boc deprotec-
tion under acidic conditions resulted in the subset of analogs
15a–15l listed in Table 3.
A key observation from this set of analogs is the increase in sol-
ubility (relative to the TZ analogs described in Tables 1 and 2) as
expected from the increased basicity. This also potentially explains
the concomitant drop in cellular activity, thought to be due to de-
creased cellular permeability as the compounds were shown to re-
tain enzymatic potency. Thus a recurring theme of a drop in
cellular potency for basic compounds with improved solubility
was observed. Given the challenges posed by the TZ series, we fo-
cused our efforts in order to identify compounds having PK proper-
ties suitable to allow in vivo profiling (PD assay) before further
attempting to improve solubility. To this end, we attempted mod-
ification of the triazolone core. This involved generation of a num-
ber of 4-hydroxymethyl matched analogs of some of the
compounds identified from the 4-aminotriazolones summarized
previously in Table 2.
Figure 2. Complexes of human Chk1 kinase in complex with inhibitors. The
structures are oriented with the N-terminal lobe towards the top left, the C-
terminal lobe towards the bottom right and the hinge region on the bottom left.
Main chain is shown as a cartoon. Hydrogen bonds are shown as dotted lines. The P-
loop has been removed for the purposes of clarity. (a) Superposition of X-ray
structure of Chk1 complexes of triazolone lead compound 5a (Pdb code: 2YER) with
triazolone hit compound 3 (Pdb code: 2X8D). Protein backbone cartoon and carbon
atoms of Chk1 complex with 5a are colored pale blue and those of 3 are colored
pink. (b) Superposition of X-ray structure of Chk1 complexes of triazolone lead
compound 5a with TCU lead compound 2 (Pdb code: 2YDK). Protein backbone
cartoon and carbon atoms of Chk1 complex with 5a are colored pale blue and those
of 2 are colored pink.
Additional efforts carried out in the 4-hydroxymethyl subseries
are summarized Table 4. These efforts resulted in the identification
of 16e as one of the most potent compounds in the 4-hydroxy-
methyl substituted subseries.
Of the various heterocyclic substituted 4-aminotriazolones syn-
thesized and tested, 14c and 14h are particularly noteworthy for
their high cellular potency. However, the solubility of these com-
pounds (2.6 and 2.8
l
M, respectively) required further improve-
From the SAR efforts described so far, 14c and 14h stood out for
their cellular potencies. While 16e was not as potent as 14c and
14h, it did offer an overall balance of potency and PK thus proving
to be a compound of choice for further assessment (Fig. 3).
Additionally, 16e represented an example which demonstrates
that reasonable PK was indeed achievable in the 4-hydrox-
ymethyltriazolone subseries. Profiling of 16e against a panel of
33 kinases revealed it to be selective against most kinases with
the exception of KDR (319 nM) which was still about 15-fold less
potent than its activity for Chk1. With the exception of the four
kinases listed in Figure 3, the activity of 16e against other kinases
ment. Towards this goal, derivatization of 14b gave 14e and 14f.
However, even though this modification improved solubility as
anticipated (by around 10-fold), cellular potencies were decreased.
Furthermore, poor PK of 14e and 14f led us to abandon such a
derivatization strategy. Similar to the SAR trend observed previ-
ously amongst the five-membered heterocycles, 2- as well as 3-thi-
eno and 2-pyrrolo groups, as exemplified by 14c, 14d and 14h,
respectively, were found to be the most potent analogs in terms
of both enzyme and cellular potency. Additionally, 4-pyridyl
substituted analog 14k was somewhat active in the Chk1 cellular
assay while the corresponding 4-methyl triazolone was not, sug-
gesting a key contribution of the 4-amino moiety in this scaffold
to the observed cellular potency.4a
in this panel ranged from ꢀ2 to >100
lM against the other kinases
from the panel. In addition, selectivity was also observed in the cell
assay where the ratio of camptothecin untreated cells to treated
cells maintains a 10-fold differential. Compound 16e was also
found to lack activity against the MDS panel of targets with% inhi-
Another interesting observation is the decrease in the potency of
14g compared to its 4-methyl analog.4a Derviatization of the 2- and

2334
V. Oza et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2330–2337
Table 1
7-Substituted-4-aminotriazolones
R2
1
R
N
N
N
H
O
Compound
R1
R2
IC50
(lM)
EC50 (CPT) (
>13
lM)
Solubility (
lM)
H
N
O
*
13a
0.16
59
NH
NH
N
H
*
*
O
O
S
13b
0.42
>13
N.A.
N
H
*
H
N
H
N
*
13c
13d
13e
13f
0.02
0.02
0.11
0.44
3.3
1
60
N
H
*
*
*
*
*
H
N
O
O
O
O
*
*
*
*
N
47
N
H
H
N
OH
>13
>13
N.A.
N.A.
N
H
H
N
N
H
OH
H
N
N
13g
0.44
>13
N.A.
OH
H
N
O
*
13h
13i
0.56
0.04
>13
N.A.
94
N
*
*
*
N
*
*
*
O
O
S
S
S
H
N
0.85
N
H
N
N
13j
0.03
0.85
0.75
>13
14.8
N.A.
N
H
O
O
13k
N
H
*
*
*
*
O
S
S
13l
0.02
0.6
0.58
>13
>13
>13
52
N
N
O
O
O
13m
13n
N.A.
59
N
H
OH
*
*
*
*
S
H
N
N
N
0.009
0.02
N
H
*
S
N
13o
66
N
N
H
N.A., not available.
IC50 and EC50 values are mean values of three determinations.
Equilibrium solubility values at pH 7.4 reported.

V. Oza et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2330–2337
2335
Table 2
Table 3
7-Substituted-4-aminotriazolones
2
7-Substituted 4-aminomethyltriazolones
1
HN
R
R
H
N
1
R
N
O
N
N
N
N
H
N
H
O
Compound R1
R2
H
IC50
M)
EC50
(CPT)
EC50
(alone)
(lM)
Compound R1
IC50
M)
EC50 (CPT)
M)
Solubility
(lM)
(l
(l
(l
(l
M)
H
14a
14b
Br
0.3
31
N.A.
N
H
*
*
15a
15b
0.03
0.1
2
31
57
*
O
S
N
0.04
1.3
N.A.
0.97
6.6
N
*
H
S
*
N
15c
15d
0.3
11
14c
14d
H
0.01
0.08
13
N
H
*
H
*
S
H
H
0.006 N.A.
0.005 0.63
4.7
N
0.04
3.3
N.A.
N
H
*
*
*
S
14e
6.3
H
N
N
H
N
15e
15f
15g
0.98
0.03
0.09
3.5
N.A.
54
*
S
OH
OH
N
N
H
14f
H
H
0.004 0.38
3.1
0.96
3.5
*
*
*
N
N
H
*
N.A.
HN
14g
0.02
0.03
2.1
N.A.
N
N
N
H
N
15h
15i
0.03
0.24
0.9
4.9
40
38
*
*
14h
H
0.02
1.7
OH
N
*
O
14i
14j
H
H
0.04
0.01
0.21
0.24
13
H
O
*
N
1.3
15j
0.09
3.3
N.A.
*
N
H
*
*
14k
H
0.13
1.1
100
N
N
15k
15l
0.21
0.4
6
54.5
N.A.
N
N.A., not available.
IC50 and EC50 values are mean values of at least three determinations.
OH
N
*
13
N
N.A., not availableIC50 and EC50 values are mean values of at least three
determinations.
Equilibrium solubility values at pH 7.4 reported.
bition <25% at 10
Figure 3.
lM. The overall profile of 16e is outlined in
Thus 16e provided us an opportunity to investigate in vivo
activity in a Chk1 specific PD model. A relatively high throughput
hollow fiber PD assay was developed in the Ncr strain nude mice
(Charles River Labs, Taconic farms) to screen compounds for their
ability to modulate Chk1 activity in vivo.4c In this model, PD
activity is assessed by calculation of the odds ratio. The odds ratio
reflects the ratio of the odd in the treatment arm over odd in the
control arm. The odd itself is defined as the % of cells in the G2 over
G1 phases of the cell cycle.4c Hence an odds ratio greater than 1 is
representative of the ability of the Chk1 inhibitor to abrogate the
cell cycle arrest induced by chemotherapy or radiation treatment.
The higher the odds ratio, the greater the ability of the inhibitor
to abrogate chemo or radiotherapy-induced arrest. HCT116 cells
in the hollow fiber PD model were treated with topotecan (50
nM) followed by 16e which was delivered at varying doses ranging
from 2.5 to 50 mpk (iv). As shown in Figure 4a, the maximum effect
was achieved at 25 mpk of 16e which then leveled off as depicted
in Figure 4a. In addition to chemopotentiation, radiopotentiation is
another expected clinical outcome based on the mode of action of
Chk1. In vitro and in vivo radio sensitization by the TCU class of
Chk1 inhibitors has previously been demonstrated.6 Based on these
previous findings, we sought to recapitulate the radio sensitization
effects observed with TCU inhibitors with inhibitors from the TZ
class as shown in Figure 4b.
Ncr nude mice implanted with hollow fibers bearing HCT116
cells were treated first with radiation (7.5 Gy) then with varying
doses of 16e. While the results were not statistically significant,
they were trending towards demonstration of dose dependence
which was encouraging for the series. Thus inhibitors in the TZ
class as exemplified by 16e showed modulation of Chk1 in the
PD assay as measured by abrogation of the DNA-damage induced
cell cycle check points.
The crystal structure of 16e bound to Chk1 kinase is shown in
Figure 5.7,13
The compound binds in an almost identical manner to that of 5a
as shown in Figure 2. Additionally, the hydroxymethyl group of
16e makes a hydrogen bonding interaction with Glu-91 of Chk1
and displaces a bound water molecule when compared with the
structure of 5a. Despite this, there is a reduced affinity of 16e for
Chk1 (IC50 20 nM) with respect to 5a (IC50 0.1 nM). This indicates
that the desolvation penalty paid upon binding of the hydroxyl
group to the protein is not compensated for by the hydrogen bond
created. It also highlighted additional SAR opportunities such as

2336
V. Oza et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2330–2337
Table 4
a
Chemopotentiation.
Cell cycle abrogation following topotecan treatment
2
1.5
1
7-Substituted 4-hydroxymethyltriazolones
2
R
O
1
R
N
N
0.5
0
N
H
O
Compound R1
R2
IC50
M)
EC50 (CPT)
M)
EC50 (alone)
(lM)
16e (2.5mpk)
16e (25mpk)
16e (50mpk)
(
l
(l
Topotecan 50nM
Topotecan 50nM
S
*
16a
H
H
0.01
0.01
2.2
5
Treatment
*
S
16b
16c
1.1
1.1
2.5
4.8
b
Radiopotentiation.
*
S
CH3 0.003
Cell cycle abrogation following irradation
2
1.5
1
*
HN
N
H
N
16d
16e
H
H
0.02
0.02
13
3.8
2.7
0.24
*
O
0.5
0
*
*
16f
16g
16h
H
H
H
0.02
0.03
0.05
2
5.3
5.3
3.7
O
4.5
13
16e (25mpk)
16e (50mpk)
Radiation 7.5Gy
*
Radiation 7.5Gy
Radiation 7.5Gy
N
Treatment
N.A., not available.
IC50 and EC50 values are mean values of at least three determinations.
p values statistically not significant with overall trend
towards dose dependence
Figure 4. (a) Dose dependent chemopotentiation of topotecan by 16e. (b) Dose
dependent radiopotentiaion by 16e.
the potential for interactions involving Asp-148 of the DFG loop
and another with Asn-134.
In summary, a novel series of Chk1 inhibitors, TZs, were identi-
fied by HTS. Using structure based drug design, we were able to
identify inhibitors that access both the sugar binding pocket region
as well as the solvent channel within the Chk1 kinase domain
resulting in a dramatic improvement in enzymatic potency. Cellu-
lar potency in this scaffold was achieved by replacing aryl groups
with heteroaryl groups. Additional SAR efforts were carried out
to generate analogs of acceptable ADME properties. This led to
the identification of compounds 14c, 14h and 16e. While further
improvement in the physicochemical properties in these com-
pounds as well as better cellular potency (especially for 16e) is still
desirable, these inhibitors enabled validation of the TZ class of
Chk1 inhibitors in a hollow fiber PD model. In addition abrogation
of topotecan-induced cell cycle arrest by 16e, we observed a simi-
lar trend in abrogation of the arrest following radiation treatment.
Although this series did not possess the physicochemical proper-
ties necessary for iv dosing, we were able to achieve potent Chk1
inhibitors that demonstrated activity in vivo in a hollow fiber PD
assay, thus demonstrating this scaffold to be a useful building
block on which to design novel kinase inhibitors.
OH
IC50: 20 nM
EC50: 240 nM
Pk (nude mouse): AUC 5.47 (uM.h)
Cl 33 mL/min/kg
H
N
T1/2 1.8 h
Vss 3.4 L/kg
Phys. Props: Solubility: 2.7 µM
N
N
N
Log D:
2.19
H
O
PPB (% Free): <5.2% (human)
<1.6% (mouse)
16e
MDS pharma profile: Clean
Kinase selectivity: No significant activity in a panel of 33 kinases
KDR
319 nM
650 nM
385 nM
JAK2
FGFR
MAP3K7 650 nM
Figure 3. Compound profile for 16e.

V. Oza et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2330–2337
2337
5133; (b) Janetka, James W.; Almeida, Lynsie; Ashwell, Susan; Brassil, Patrick J.;
Daly, Kevin; Deng, Chun; Gero, Thomas; Glynn, Roberta E.; Horn, Candice L.;
Ioannidis, Stephanos; Lyne, Paul; Newcombe, Nicholas J.; Oza, Vibha B.; Pass,
Martin; Springer, Stephanie K.; Su, Mei; Toader, Dorin; Vasbinder, Melissa M.;
Yu, Dingwei; Yu, Yan; Zabludoff, Sonya D. Bioorg. Med. Chem. Lett. 2008, 18,
4242; (c) Zabludoff, Sonya D.; Deng, Chun; Grondine, Michael R.; Sheehy, Adam
M.; Ashwell, Susan; Caleb, Benjamin L.; Green, Stephen; Haye, Heather R.;
Horn, Candice L.; Janetka, James W.; Liu, Dongfang; Mouchet, Elizabeth; Ready,
Shannon; Rosenthal, Judith L.; Queva, Christophe; Schwartz, Gary K.; Taylor,
Karen J.; Tse, Archie N.; Walker, Graeme E.; White, Anne M. Mol. Cancer Ther.
2008, 7, 2955.
5. Buchman, E. R.; Sargent, H.; Myers, T. C.; Seneker, J. A. J. Am. Chem. Soc. 1946, 68,
2692.
6. Mitchell, James B.; Choudhuri, Rajani; Fabre, Kristin; Sowers, Anastasia L.;
Citrin, Deborah; Zabludoff, Sonya D.; Cook, John A. Clin. Cancer Res. 2010, 16,
2076.
7. For cloning, sequencing, overexpression, purification and crystallization of Chk1
kinase: Protein and co-crystals with compounds 5a and 16e were obtained as
described previously.4a Crystals were cryo-protected with ethylene glycol (20%
v/v) in mother liquor and flash cooled in a cryo-stream at 100 K. Diffraction
data for complex 5a were collected at the ESRF beamline ID29 equipped with a
ADSC Quantum 4 X-ray detector, using a Si111 mono-chromated wavelength of
1.265 Å at 100 K. Diffraction data for complex 16e were collected on a Rigaku
Fre X-ray generator using CuK
a X-ray radiation, nickel filters and Osmic
Figure 5. X-ray structure of Lead TZ Inhibitor 16e in complex with Chk1 kinase
(PDB code: 2YER). The structures is oriented in the with the N-terminal lobe
towards the top left, the C-terminal lobe towards the bottom right and the hinge
region on the bottom left. Protein backbone cartoon and carbon atoms of the
protein are colored wheat and protein carbon atoms of the inhibitor are colored
green. Hydrogen bonds and are shown as dotted lines. Electron density (2Fo À Fc) is
VariMax HF mirrors equipped with a Rigaku Saturn 944+ CCD X-ray detector.
Data for all complexes were processed using MOSFLM8 and SCALA and were
reduced using CCP4 software.9 Data collection, processing and refinement
statistics for crystal structures of compounds 5a and 16e bound in human
CHK1 kinase are described in Ref. 17. The structures were solved by molecular
replacement using coordinates of the Chk1 kinase domain (pdb code 2x8d) as a
trial model using CCP4 software. Protein and inhibitor were modeled into the
electron density using either Quanta,12 for the complex of Chk1 with 5a or
COOT,11 and Flynn15 for the complex of Chk1 with 16e. The model of complex
5a was refined using CNX10 and the CCP4 program Refmac514 and the model of
complex 16e was refined solely with Refmac5. Atomic coordinates and
structure factors for the Chk complexes with 5a and 16e have been
deposited in the Protein Data Bank (2YEX, and 2YER, respectively) along
with structure factors and detailed experimental conditions. The
stereochemical quality of Chk1 protein was checked using PROCHECK.16
contoured at 1r is shown as a wire mesh.
Acknowledgments
The authors wish to thank Anne White and Graham Walker for
their respective work on the development and implementation of
the high-throughput Chk1 kinase screening assay. The authors also
thank Elizabeth Mouchet for the development of cellular abroga-
tion assay. Authors would also like to thank Heather Haye for pro-
tein production, Hannah Pollard for the earlier structures and
Richard Pauptit for proof-reading this manuscript.
8. Mosflm: Leslie, A.G.W., Joint CCP4
Crystallography 1992, 26.
9. CCP4, Acta Crystallogr., Sect. D 1994, 50, 760.
10. CNX: Version 2000.1, Accelrys.
11. Coot: Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D 2004, 60, 2126.
12. Quanta: QUANTA2000, Accelrys.
+ ESF-EAMCB Newsletter on Protein
13. Pymol: DeLano, W. L. The PyMOL Molecular Graphics System; DeLano
Scientific: San Carlos, CA, USA, 2002. http://www.pymol.org.
14. Refmac: Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystallogr., Sect. D
1997, 53, 240.
15. Flynn: version 1.1.1, OpenEye Scientific Software, Inc., Santa Fe, NM, USA, 2010.
www.eyesopen.com.
16. Procheck: Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J. M. J.
Appl. Crystallogr. 1993, 26, 283.
17. Crystallographic statistics for the CHK1-compound 5a complex are as follows:
Numbers in parenthesis characterize the higher resolution shell. Space group
P212121, unit cell 41.4, 70.8, 104.3 Å, resolution 24.9–1.3 Å (1.37–1.30), 62,083
reflections with an overall redundancy of 2.8 (1.7) give 83.6% (34.3)
References and notes
1. Hanahan, Douglas; Weinberg, Robert Cell 2011, 144, 646.
2. (a) Ashwell, Susan; Zabludoff, Sonya Clin. Cancer Res. 2008, 14, 4032; (b)
Bucher, N.; Britten, C. D. Br. J. Cancer 2008, 98, 523; (c) Kong, Norman; Fotouhi,
Nader; Wovkulich, Peter M.; Roberts, John Drugs Future 2003, 28, 881.
3. (a) Garrett, Michelle D.; Collins, Ian Trends Pharmacol. Sci. 2011, 32, 308; (b)
Lainchbury, Michael; Collins, Ian Exp. Opin. Ther. Patents 2011, 21, 1191; (c)
Dent, Paul; Tang, Yong; Yacoub, Adly; Dai, Yun; Fisher, Paul B.; Grant, Steven
Mol. Interv. 2011, 11, 133; (d) Walton, Michael I.; Eve, Paul D.; Hayes, Angela;
Valenti, Melanie; De Haven Brandon, Alexis; Box, Gary; Boxall, Kathy J.; Aherne,
G. Wynne; Eccles, Suzanne A.; Raynaud, Florence I.; Williams, David H.; Reader,
John C.; Collins, Ian; Garrett, Michelle D. Mol. Cancer Ther. 2010, 9, 89; (e) Du,
Juan; Xi, Lili; Lei, Beilei; Lu, Jing; Li, Jiazhong; Liu, Huanxiang; Yao, Xiaojun J.
Comp. Chem. 2010, 31, 2783; (f) Janetka, James W.; Ashwell, Susan Exp. Opin.
Ther. Patents 2009, 19, 165; (g) Jackson, Jeffrey R.; Gilmartin, Aidan; Imburgia,
Christina; Winkler, James D.; Marshall, Lisa A.; Roshak, Amy Cancer Res. 2000,
60, 566; (h) Tao, Zhi-Fu; Lin, Nan-Horng Anti-Cancer Agents Med. Chem. 2006, 6,
377.
completeness with Rmerge of 5.6% (51.8%) and mean I/r(I) of 13.3 (1.7). The
final model containing 2191 protein, 449 solvent, and 20 compound atoms has
an R-factor of 13.6% (Rfree using 5% of the data 16.4%). Mean temperature
factors for the protein and the ligand are 10.4 and 8.5 Å2, respectively.
Crystallographic statistics for the CHK1-compound 16e complex are as follows:
Space group P212121, unit cell 41.2, 69.9, 103.5 Å, resolution 103.5 1.83 Å
(1.93–1.83), 22,780 reflections with an overall redundancy of 5.6 (1.7) give
83.9% (25.0) completeness with Rmerge of 3.3% (4.8%) and mean I/r(I) of
36.5(11.9). The final model containing 2158 protein, 407 solvent, and 21
compound atoms has an R-factor of 14.7% (Rfree using 5% of the data 17.5%).
Mean temperature factors for the protein and the ligand are 10.6 and 10.4 Å2,
respectively.
4. (a) Oza, Vibha; Ashwell, Susan; Brassil, Patrick; Breed, Jason; Deng, Chun;
Ezhuthachan, Jay; Haye, Heather; Horn, Candice; Janetka, James; Lyne, Paul;
Newcombe, Nicholas; Otterbien, Ludo; Pass, Martin; Read, Jon; Roswell, Sian;
Su, Mei; Toader, Dorin; Yu, Dingwei; Yu, Yan; Valentine, Anna; Webborn, Peter;
White, Ann; Zabludoff, Sonya; Zheng, Xiaolan Bioorg. Med. Chem. Lett. 2010, 20,