Multiparameter Lead Optimization to Give an Oral Checkpoint Kinase 1 (CHK1) Inhibitor Clinical Candidate: (R)-5-((4-((Morpholin-2-ylmethyl)amino)-5-(trifluoromethyl)pyridin-2-yl)amino)pyrazine-2-carbonitrile (CCT245737)

Article abstract of DOI:10.1021/acs.jmedchem.5b01938

Multiparameter optimization of a series of 5-((4-aminopyridin-2-yl)amino)pyrazine-2-carbonitriles resulted in the identification of a potent and selective oral CHK1 preclinical development candidate with in vivo efficacy as a potentiator of deoxyribonucleic acid (DNA) damaging chemotherapy and as a single agent. Cellular mechanism of action assays were used to give an integrated assessment of compound selectivity during optimization resulting in a highly CHK1 selective adenosine triphosphate (ATP) competitive inhibitor. A single substituent vector directed away from the CHK1 kinase active site was unexpectedly found to drive the selective cellular efficacy of the compounds. Both CHK1 potency and off-target human ether-a-go-go-related gene (hERG) ion channel inhibition were dependent on lipophilicity and basicity in this series. Optimization of CHK1 cellular potency and in vivo pharmacokinetic-pharmacodynamic (PK-PD) properties gave a compound with low predicted doses and exposures in humans which mitigated the residual weak in vitro hERG inhibition.

Full text of DOI:10.1021/acs.jmedchem.5b01938

Subscriber access provided by University of Sussex Library
Article
Multi-parameter lead optimization to give an oral checkpoint kinase 1 (CHK1)
inhibitor clinical candidate: (R)-5-((4-((morpholin-2-ylmethyl)amino)-5-
(trifluoromethyl)pyridin-2-yl)amino)pyrazine-2-carbonitrile (CCT245737).
James D Osborne, Thomas Peter Matthews, Tatiana McHardy, Nicolas Proisy, Kwai-Ming Jack Cheung,
Michael Lainchbury, Nathan Brown, Michael I Walton, Paul D Eve, Katherine J. Boxall, Angela Hayes, Alan
T Henley, Melanie R Valenti, Alexis K De Haven Brandon, Gary Box, Yann Jamin, Simon P Robinson,
Isaac M Westwood, Rob Leon Marie van Montfort, Philip M Leonard, Marieke BAC Lamers, John C.
Reader, G Wynne Aherne, Florence I Raynaud, Suzanne A. Eccles, Michelle D Garrett, and Ian Collins
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01938 • Publication Date (Web): 11 May 2016
Downloaded from http://pubs.acs.org on May 17, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Cancer Therapeutics Unit
van Montfort, Rob; The Institute of Cancer Research, Cancer Research UK
Cancer Therapeutics Unit
Leonard, Philip; Sareum Ltd
Lamers, Marieke; Sareum Ltd
Reader, John C.; Sareum Ltd
Aherne, G; The Institute of Cancer Research, Cancer Research UK Cancer
Therapeutics Unit
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Raynaud, Florence; The Institute of Cancer Research, Cancer Research UK
Cancer Therapeutics Unit
Eccles, Suzanne; The Institute of Cancer Research, CRUK Cancer
Therapeutics Unit
Garrett, Michelle; The Institute of Cancer Research, Cancer Research UK
Cancer Therapeutics Unit
Collins, Ian; The Institute of Cancer Research, Cancer Research UK Cancer
Therapeutics Unit
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Multiꢀparameter lead optimization to give an oral
checkpoint kinase 1 (CHK1) inhibitor clinical
candidate: (R)ꢀ5ꢀ((4ꢀ((morpholinꢀ2ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ylmethyl)amino)ꢀ5ꢀ(trifluoromethyl)pyridinꢀ2ꢀ
yl)amino)pyrazineꢀ2ꢀcarbonitrile (CCT245737).
†
†
†
†
James D. Osborne, Thomas P. Matthews, Tatiana McHardy, Nicolas Proisy, Kwai-Ming J.
†
†
†
†
†
Cheung, Michael Lainchbury, Nathan Brown, Michael I. Walton, Paul D. Eve, Katherine J.
†
†
†
†
†
Boxall, Angela Hayes, Alan T. Henley, Melanie R. Valenti, Alexis K. De Haven Brandon,
†
§
§
†
†
Gary Box, Yann Jamin, Simon P. Robinson, Isaac M. Westwood, Rob L. M. van Montfort,
‡
‡
‡
†
Philip M. Leonard, Marieke B. A. C. Lamers, John C. Reader, G. Wynne Aherne, Florence I.
†
†
†
†,*
Raynaud, Suzanne A. Eccles, Michelle D. Garrett, Ian Collins,
†
§
Cancer Research UK Cancer Therapeutics Unit and Division of Radiotherapy and Imaging,
The Institute of Cancer Research, London, SM2 5NG, U.K.
‡
Sareum Ltd, Cambridge, CB22 3FX, U.K.
ACS Paragon Plus Environment
1

Page 3 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT
Multiꢀparameter optimization of a series of 5ꢀ((4ꢀaminopyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀ
carbonitriles resulted in the identification of a potent and selective, oral CHK1 preclinical
development candidate with in vivo efficacy as a potentiator of deoxyribonucleic acid (DNA)
damaging chemotherapy and as a single agent. Cellular mechanism of action assays were used to
give an integrated assessment of compound selectivity during optimization resulting in a highly
CHK1 selective, adenosine triphosphate (ATP) competitive inhibitor. A single substituent vector
directed away from the CHK1 kinase active site was unexpectedly found to drive the selective
cellular efficacy of the compounds. Both CHK1 potency and offꢀtarget human etherꢀaꢀgoꢀgoꢀ
related gene (hERG) ion channel inhibition were dependent on lipophilicity and basicity in this
series. Optimization of CHK1 cellular potency and in vivo pharmacokineticꢀpharmacodynamic
(PKꢀPD) properties gave a compound with low predicted doses and exposures in humans which
mitigated the residual weak in vitro hERG inhibition.
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
The serineꢀthreonine kinase CHK1 is a critical component of the cellular response to DNA
damage, especially the repair of DNA breaks arising during replication or caused by DNAꢀ
1
damaging chemotherapies and ionizing radiation. CHK1 is activated by the upstream kinases
ataxia telangiectasia and Rad3ꢀrelated (ATR), and to a lesser extent ataxia telangiectasia mutated
2
(
ATM). CHK1 is phosphorylated on residues S317 and S345 by ATR, and undergoes
3
autophosphorylation on S296. The activated enzyme phosphorylates a range of substrates,
including the cell division cycle 25 (CDC25) family of phosphatases that regulate cell cycle
ACS Paragon Plus Environment
2

Journal of Medicinal Chemistry
Page 4 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
progression, with the result that the cell cycle is halted at checkpoints in Synthesis (S) or Gap 2
4
(
G2) phase. At the same time, CHK1 initiates DNA repair by homologous recombination
5
through signaling to the repair protein RAD51. The cell cycle checkpoints controlled by CHK1
therefore provide an opportunity for repair of the damaged DNA before the replicating cell enters
mitosis. If the damage is resolved the checkpoint is released and cell growth and division
recommence, while prolonged activation of the checkpoint leads to senescence or cell death.
The DNA damage response (DDR) mediated by ATRꢀCHK1 provides a mechanism for cancer
cells to resist DNAꢀdamaging antiꢀcancer therapies and is an attractive point of intervention for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
new treatments. A strong body of evidence shows that inhibition of CHK1 with small
molecules, or depletion by ribonucleic acid (RNA) interference, selectively sensitizes tumor cells
7
to genotoxic drugs and ionizing radiation. Normal cells have multiple DNAꢀdamage
checkpoints, including the important p53ꢀdependent Gap 1 (G1) phase checkpoint. However,
many cancer cells harbor defects in one or more of these checkpoints, for example due to
frequent deletion, mutation and loss of function of the tumor suppressor p53, and rely on the S
8
and G2 checkpoints mediated by CHK1 to effect repair of DNA damage.
The potential of CHK1 inhibition as a monoꢀtherapy has been demonstrated in tumor cells with
replication stress. These tumors rely on the CHK1ꢀmediated DDR to repair intrinsically high
levels of DNA strand breaks, for example arising from increased replication fork stalling in S
phase, and include cancers driven by vꢀmyc avian myelocytomatosis viral oncogene homolog
(MYC) and vꢀmyc avian myelocytomatosis viral oncogene neuroblastoma derived homolog
9
,10
11,12
(
MYCN) oncogenes , as well as some melanomas and tripleꢀnegative breast cancers.
There has been substantial activity in developing CHK1 inhibitors for clinical study in
7,13
combination with various DNAꢀdamaging agents.
While the first clinical studies have
ACS Paragon Plus Environment
3

Page 5 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
demonstrated the tolerability of CHK1 inhibition, further progression has often been challenging
and development of several agents has been halted after early phase clinical trials. Some of the
first generation inhibitors have low selectivity for CHK1, and were developed as intravenous
agents. Subsequent to the identification of these compounds, studies showed that maximum
potentiation of gemcitabine efficacy in preclinical models was associated with exposure to the
CHK1 inhibitor during a window beginning some 24 h after administration of the genotoxic
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
4,15
agent.
Consequently, an oral compound would offer advantages of ease of administration and
flexibility of scheduling. In addition, preclinical data on the emerging contexts for CHK1 monoꢀ
therapy suggests that daily dosing may be needed, for which an oral drug would again be
favored. The oral CHK1 inhibitors GDCꢀ0425 and GDCꢀ0575 have been reported in Phase I
1
6,17
clinical trials.
There remains a clear need for highly selective, oral CHK1 inhibitors from
different chemical series to advance to clinical studies.
We previously reported the discovery and optimization of the first published selective, oral
CHK1 inhibitor (R)ꢀ3ꢀ((1ꢀ(dimethylamino)propanꢀ2ꢀyl)oxy)ꢀ5ꢀ((4ꢀmethoxyꢀ5ꢀ(1ꢀmethylꢀ1Hꢀ
18
pyrazolꢀ4ꢀyl)pyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀcarbonitrile (1) through a scaffold hybridization
strategy starting from potent, selective but highly metabolized inhibitor (R)ꢀ5ꢀ((8ꢀ
chloroisoquinolinꢀ3ꢀyl)amino)ꢀ3ꢀ((1ꢀ(dimethylamino)propanꢀ2ꢀyl)oxy)pyrazineꢀ2ꢀcarbonitrile
(2) and the less potent but metabolically stable 5ꢀ((4ꢀaminopyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀ
1
9,20
carbonitrile 3 (Figure 1).
While effective in preclinical models, compound 1 was suboptimal
due to low metabolic stability in human hepatocytes, leading to predictions of poor human
pharmacokinetics (PK) and a requirement for high doses. In parallel with the discovery of 1, we
pursued the multiꢀparameter optimization of the 5ꢀ((4ꢀaminopyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀ
carbonitriles based on 3, described herein, resulting in the identification of the potent and highly
ACS Paragon Plus Environment
4

Journal of Medicinal Chemistry
Page 6 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
selective, oral CHK1 clinical candidate, (R)ꢀ5ꢀ((4ꢀ((morpholinꢀ2ꢀylmethyl)amino)ꢀ5ꢀ
2
1
(trifluoromethyl)pyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀcarbonitrile (4; CCT245737).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
Figure 1. Summary of the evolution of the oral inhibitor tool compound 1 and oral clinical
2
1
development candidate 4.
RESULTS AND DISCUSSION
Assay cascade and target candidate profile
The desired candidate profile was developed based on our experience with compounds 1 and 2.
Potent biochemical inhibition of CHK1 (IC50 <10 nM) was required. We measured CHK1
inhibition using a DissociationꢀEnhanced Lanthanide Fluorescent Immunoassay (DELFIA),
format and also developed a microfluidic assay for CHK1 substrate phosphorylation with a
larger dynamic range to accurately assess tightꢀbinding compounds. High selectivity for CHK1
over other kinases was considered essential. In particular, inhibition of the structurally distinct
DDR protein, checkpoint kinase 2 (CHK2), has been shown to be detrimental to the potentiation
2
2
of the cytotoxicity of DNA damaging agents and radiotherapy by CHK1 inhibition. In contrast
to CHEK1, encoding for the CHK1 protein, the CHEK2 gene is a tumor susceptibility gene, and
ACS Paragon Plus Environment
5

Page 7 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
selective inhibition of CHK2 does not sensitize cancer cells to a range of DNAꢀdamaging
2
3,24
agents.
We also monitored compounds for inhibition of the cell cycle regulatory kinase,
cyclin dependent kinase 1 (CDK1). Continued cell cycle progression into mitosis is necessary to
expose cancer cells to the deleterious effects of enhanced DNA damage through checkpoint
abrogation. Inhibition of cell cycle regulatory kinases such as CDK1, itself an indirect
downstream target of activated CHK1, would lead to CHK1ꢀindependent cell cycle arrest that
would confound interpretation of the cellular activity of the compounds. In both cases, we
considered biochemical selectivity of >100ꢀfold a minimum requirement. As lipophilic bases
present a general risk of ion channel inhibition, we monitored hERG inhibition as a
representative and clinically relevant offꢀtarget during lead optimization.
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Potent and highly selective inhibition of CHK1ꢀdependent effects in human cancer cells was
considered essential and we specified an IC <100 nM for the abrogation of a CHK1ꢀdependent
5
0
2
0
etoposideꢀinduced G2/M checkpoint in p53ꢀdeficient HT29 human colon cancer cells. To
measure CHK1 selective effects in cells, the potency for the CHK1ꢀspecific checkpoint
abrogation was compared with the nonꢀspecific antiꢀproliferative effect of the inhibitors in the
2
0
same cell line, with a minimum requirement of fiveꢀfold cellular selectivity. We used this ratio
of cellular effects as an integrated surrogate for kinomeꢀwide or other large scale biochemical
profiling of multiple compounds during lead optimization. The growth of HT29 cells is not
highly sensitive to selective inhibition of CHK1, thus the ratio of the potency for this CHK1ꢀ
independent effect to that for the CHK1ꢀdependent checkpoint abrogation informs on the target
selectivity of the compounds. Compounds with sufficient cellular potency and selectivity were
assessed for their ability to potentiate the cytotoxicity of the DNAꢀdamaging agent gemcitabine
in SW620 human colon cancer cells. CHK1 inhibition has been shown to synergize maximally
ACS Paragon Plus Environment
6

Journal of Medicinal Chemistry
Page 8 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
with antiꢀmetabolites such as gemcitabine compared with other classes of DNAꢀdamaging
2
5
chemotherapies.
Cellular studies of the tool compound 1 had shown that sustained inhibition of CHK1 signaling
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
after administration of the DNAꢀdamaging agent was required for maximal potentiation of the
1
4
efficacy of the cytotoxic agent. Therefore; compounds with sufficient metabolic stability,
plasma concentration and pharmacodynamic (PD) duration of action to give 24 h inhibition of
the target following a single oral dose were sought. Inhibition of the autophosphorylation of
CHK1 on S296 in HT29 human tumor xenografts was used as a PD measure of target protein
14
engagement.
Optimization of in vitro and PK properties
Compound 3 (Table 1) presented an encouraging starting point for realization of the desired
clinical candidate profile. While requiring increases in cellular potency and selectivity, the
compound showed more than 400ꢀfold selectivity for in vitro inhibition of CHK1 over CHK2,
and was stable in mouse liver microsomes (MLM) (9% metabolized after 30 min; Table S1,
Supporting Information). To improve on 3 we chose to retain the 5ꢀaminopyrazineꢀ2ꢀcarbonitrile
group, as our extensive structural biology studies to develop the series had shown this group to
interact optimally with unique proteinꢀbound water molecules in CHK1, conferring high
1
9
selectivity over other kinases. Basic substitution at Cꢀ4 of the pyridine core was retained to
maximize affinity through potential interactions with polar residues in the ribose binding pocket
of the CHK1 active site, as we had previously shown for 5ꢀ((6ꢀ((piperidinꢀ4ꢀ
ylmethyl)amino)pyrimidinꢀ4ꢀyl)amino)pyrazineꢀ2ꢀcarbonitrile (PDB: 2YM7) and other
19
compounds prepared during the identification of 3.
ACS Paragon Plus Environment
7

Page 9 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1: In vitro properties of selected 4ꢀaminoꢀsubstituted methyl 6ꢀ((5ꢀcyanopyrazinꢀ2ꢀ
yl)amino)nicotinates
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
No
R
R
CHK1
CHK2
Checkpoint Cellular
hERG
a
b
IC (nM) IC (µM) abrogation
selectivity (%inhibition
5
0
50
c
d
e
IC (nM)
(fold)
9.6 (±2.4) 825 (±125) 3.8
11 (±0.1) 9.8 (±1.3) 440 (±130) 5.1
850 (±450) 4.3
@ 10 µM)
5
0
3
5
6
7
8
9
A
B
A
C
H
ꢀ
21 (±6)
39
55
82
69
58
91
60
42
24
Me 19 (±1)
ꢀ
f
ꢀ
2.9 (±0.1) 14
129 (±62)
4.4
6.2
7.5
6.8
3.5
3.5
f
f
(s)-D
H
1.0
11 (±5)
21
g
f,h
f
D
Me 0.5
2.2
12 (±6)
f
10 (s)-E
ꢀ
ꢀ
7.0 (±0.3)
ꢀ
2500
b,f
f
1
1
2
rac-F
56
ꢀ
1100
f
f
1
rac-G ꢀ
2.2 (±0.3) 5.0
310
a
b
Determined in Caliper microfluidic assay, mean (±SD) for n=2. Determined in DELFIA,
mean (±SD) for n=2. Abrogation of etoposideꢀinduced G2 checkpoint arrest in HT29 human
colon cancer cells, mean (±SD) for n=2. Ratio of antiproliferative GI (measured in duplicate
by sulforhodamine B (SRB) assay) to IC for CHK1ꢀmediated abrogation of etoposideꢀinduced
G2 checkpoint arrest in HT29 human cancer cells. Inhibition of hERG ion current in HEK cells
overexpressing hERG ion channel (PatchExpress, Millipore Inc.), single determination at 10 µM.
Single determination. Prepared as a mixture of (3s) and (3r) isomers. Below tightꢀbinding
c
d
5
0
50
e
f
g
h
limit of assay.
ACS Paragon Plus Environment
8

Journal of Medicinal Chemistry
Page 10 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Initially we examined a range of 4ꢀaminoꢀsubstituted methyl 6ꢀ((5ꢀcyanopyrazinꢀ2ꢀ
yl)amino)nicotinates to determine the scope for replacing the 4ꢀ(piperidineꢀ4ꢀmethyl) substituent
of 3, with selected examples shown in Table 1. Homologation (5) or truncation (7) of the linking
alkyl chain, and Nꢀmethylation (6 and 7) of the piperidine were tolerated, with increased CHK1
inhibition and cellular potency for 7. The tropane derivatives 8, 9 and 10 showed potent
inhibition of CHK1 activity (IC <10 nM), translating to cellular potency of <100 nM for 8 and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
9
and cellular selectivity exceeding the target threshold of fiveꢀfold. The racemic piperidinꢀ2ꢀ
methyl analogue 11 retained the profile of 3 while the racemic morpholinꢀ2ꢀylꢀmethyl derivative
2 had enhanced biochemical inhibition compared to the piperidines 3 and 11, with more potent
1
checkpoint abrogation in cells, but a greater differential between biochemical and cellular
activities.
While the increased lipophilicity of the tropane derivatives led to improved potency and
cellular selectivity, it was also associated with more potent hERG inhibition, particularly for the
Nꢀmethyl derivative 9. In contrast, the racemic morpholinꢀ2ꢀylꢀmethyl derivative 12 delivered
enhanced CHK1 affinity with minimal hERG inhibition. Reductions of amine lipophilicity or
2
6
basicity are documented as strategies for reducing hERG activity. hERG inhibition by a wider
group of 4ꢀaminoꢀsubstituted methyl 6ꢀ((5ꢀcyanopyrazinꢀ2ꢀyl)amino)nicotinates prepared during
this study, including those exemplified in Table 1, showed a correlation with calculated ALogP
(Figure 2). Less potent hERG inhibition was more frequently associated with secondary amines
with a total of 3 hydrogen bond donors (HBD) in the molecule, as shown by compounds 3, 11
and 12, compared to tertiary amines (2 HBD). This indicated that an increased risk of hERG
inhibition would be associated with compounds with higher calculated AlogP in this series, but
ACS Paragon Plus Environment
9

Page 11 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
this would need to be balanced with the potential for improved CHK1 cellular potency and
selectivity with more lipophilic amines.
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Relationship of hERG inhibition to calculated AlogP and HBD count for 4ꢀaminoꢀ
substituted methyl 6ꢀ((5ꢀcyanopyrazinꢀ2ꢀyl)amino)nicotinates including examples specified in
Table 1.
During the optimization of the series, selected compounds were screened for metabolic
stability in MLM and typically showed <30% metabolism of the parent compound after 30 min
(see Table S1, Supporting Information). The mouse PK profiles of the esters 5 and 7 were
determined (Table 2). Compound 5 showed moderate oral bioavailability (44%), high volume of
distribution, and clearance greater than liver blood flow. In vivo clearance was substantially
increased for the Nꢀmethyl piperidine analogue 7 despite similar MLM in vitro stability (26%
and 28% metabolized after 30 min for 5 and 7, respectively; Table S1, Supporting Information).
A number of methyl esterꢀsubstituted analogues in the related 3ꢀalkoxyꢀ5ꢀ(pyridinꢀ2ꢀ
ACS Paragon Plus Environment
10

Journal of Medicinal Chemistry
Page 12 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
19
ylamino)pyrazineꢀ2ꢀcarbonitrile scaffold were assessed for stability in human plasma, but no
evidence for degradation was seen during a 60 min incubation (data not shown). To efficiently
discriminate compounds on the basis of PK performance, we used a limited sampling strategy
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
18
that compared plasma levels at 1 h and 6 h after i.v. and p.o. dosing to assess compounds in
vivo with promising in vitro potency and selectivity, or with structural features important to the
development of structureꢀactivity relationships. This strategy provided a useful ranking of
relative oral bioavailability and clearance. Selected compounds were subsequently progressed to
full PK determinations (Table 2).
Table 2: In vivo mouse PK properties of selected compounds following intravenous (i.v.) or oral
(p.o.) dosing.
a
a
a
c
No. i.v. Dose Cl
Vss
T
p.o. Dose F
(mg/kg)
1/2
b
b
(
mg/kg) (L/h/kg) (L/kg) (h)
(%)
44
10
0
5
7
1
5
5
5
8.45
54.7
7.85
60.5
20.2
7.25
11
5
5
5
0.36
1.6
5
25
30
36
38
40
41
4
10
1
6.75
0.50
8.25
6.35
6.70
2.75
2.65
26.8
1.15
10.4
14.8
43.9
10.1
9.25
4.3
2.0
3.8
2.2
8.1
3.4
2.9
10
5
53
16
10
68
68
84
105
10
5
10
5
10
10
10
10
10
10
a
b
c
Parameters from i.v. PK profile. Calculated based on a 20 g mouse body weight. Oral
bioavailability.
ACS Paragon Plus Environment
11

Page 13 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The crystal structure of 12 bound to CHK1 was determined (PDB: 5F4N, Figure 3). The 2ꢀ
aminopyridine core formed hydrogen bonds with Glu85 and Cys87 in the kinase hinge region, as
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
19
seen previously for compounds such as 2 (PDB: 2YM8). The cyanopyrazine was positioned
close to the sideꢀchain of Lys38. Importantly, both the nitrile and Nꢀ4 of the pyrazine ring were
positioned to interact with bound water at the entrance to the pocket beyond the gatekeeper
residue, one of a network of conserved water molecules resulting from the presence of the unique
polar residue Asn59 in this pocket in CHK1 instead of the more common lipophilic side chains.
Interactions with these bound waters are a CHK1 selectivity determinant for this and other series
2
7,28
of CHK1 inhibitors, including other pyrazines.
Consistent with this, compounds 3, 5, 6, 7, 9,
and 12 did not inhibit CDK1 (IC >10 µM).
5
0
Figure 3: Detail of the crystal structure (PDB: 5F4N) showing 12 (green) bound in the CHK1
ATP site with key residues (grey) and water molecules (red spheres) shown.
The crystallographic model indicated only the (R)ꢀenantiomer of the morpholine to be bound.
The protonated amine of the morpholine engaged in direct interactions with the sideꢀchain of
ACS Paragon Plus Environment
12

Journal of Medicinal Chemistry
Page 14 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Glu91 and the mainꢀchain carbonyl of Glu134. In addition, waterꢀmediated interactions to
Asn135 and Ser147 were observed, illustrating the rich set of potential contacts available to basic
substituents in this region of the kinase and consistent with the generally high CHK1 affinity of
the amine variants in Table 1. The morpholine projected towards the main chain of residues
Leu15, Gly16 and Glu17 in the Pꢀloop but did not form close contacts. At the entrance to the
ATPꢀpocket, the Cꢀ5 methyl ester substituent lay over the selectivity surface formed by Ser88,
Gly89 and Gly90. No direct polar interactions of the ester to this moderately hydrophobic
surface were observed, suggesting scope for optimization of contacts to the protein. Interestingly,
binding of a second molecule of the inhibitor was detected in the crystal structure at a separate
hydrophobic site in the Nꢀterminal domain defined by Trp9, Tyr71 and Phe83, however no
specific polar interactions of the ligand to the protein were evident and this is unlikely to
represent a biologically relevant site of action.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In parallel with variation of the Cꢀ4 substituent, investigation of substitution at the Cꢀ5 position
of 3 was undertaken. The Cꢀ5 ester substituent projects towards solvent and was anticipated to
provide a handle to optimize in vitro activity and allow for modulation of PK properties. A
diverse set of substituents from multiple scaffolds has been shown to be capable of binding in
2
7
this region of CHK1. Compounds were chosen to explore a range of physicochemical
properties and the resulting structureꢀactivity relationship is illustrated for specific examples of
different substituent types (Table 3).
ACS Paragon Plus Environment
13

Page 15 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3: In vitro properties of selected substituted 5ꢀ((4ꢀ((piperidinꢀ4ꢀylmethyl)amino)pyridinꢀ2ꢀ
yl)amino)pyrazineꢀ2ꢀcarbonitriles
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
No. CHK1
IC (nM)
Checkpoint
abrogation
Cellular
hERG
a
selectivity (%inhibition@
5
0
b
c
d
IC (nM)
(fold)
10 µM)
5
0
13
14
15
16
17
18
398 (±48)
8.8 (±0.7)
ꢀ
ꢀ
ꢀ
e
4400 (±400)
500
5.6
7.6
ꢀ
76
26
ꢀ
f,g
55
54 (±4)
>10000
4000
f
4.7 (±0.8)
12 (±1)
4.8
4.8
5.9
3.0
7.6
52
34
59
ꢀ
3100
f,g
19
20
21
50
440
177 (±8)
92 (±7)
600
2500
64
a
b
Determined in Caliper microfluidic assay, mean (±SD) for n=2. Abrogation of etoposideꢀ
c
induced G2 checkpoint arrest in HT29 human colon cancer cells, single determination. Ratio of
antiproliferative GI50 (measured in duplicate by SRB assay) to IC50 for CHK1ꢀmediated
abrogation of etoposideꢀinduced G2 checkpoint arrest in HT29 human cancer cells. Inhibition
of hERG ion current in HEK cells overexpressing hERG ion channel (PatchExpress, Millipore
Inc.), single determination at 10 µM. Mean (±SD) for n=2. Determined in DELFIA assay.
Single determination.
d
e
f
g
ACS Paragon Plus Environment
14

Journal of Medicinal Chemistry
Page 16 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Replacement of the ester with tertiary alkyl amides led to decreased CHK1 inhibition, e.g. 13
Table 3). Secondary alkyl amides, e.g. 14, were better tolerated for CHK1 affinity but had poor
(
cellular activity, presumably as a result of the increased polarity and additional HBD compared
to 3. Secondary aryl amides such as 15 offered a better range of CHK1 potencies and retained
some cellular activity with increased cellular selectivity over the ester 3. Compound 15 also
showed low hERG inhibition despite the addition of the lipophilic aryl substituent. However, the
PK properties of 15 were discouraging, with high clearance and no detectable oral bioavailability
(Table 2). Substitution with alkyl, alkenyl and alkynyl groups at Cꢀ5 was tolerated for CHK1
inhibition, with a general preference for the unsaturated functionality at Cꢀ5 as represented by 17
and 18 compared to 16. Aryl and heteroaryl substitution also retained CHK1 inhibition, as for
example the phenyl 19 and oxadiazole 21 analogues. CHK1 inhibition was sensitive to
substitution on the Cꢀ5 aryl group, as exemplified by 3ꢀfluorophenyl analogue 20.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
As the inhibitors are ATPꢀcompetitive, binding to an active conformation of CHK1 kinase
(Figure 3), we expected a decrease in potency in transferring from the CHK1 biochemical assays
([ATP] = 30 µM) to cells with high endogenous [ATP]. Esters (see Table 1), aryl amides and
heteroaryl or aryl Cꢀ5 substituents generally gave the most favorable ratio of potency for G2/M
checkpoint abrogation to CHK1 biochemical inhibition (<100ꢀfold).
The SAR described above showed that identifying a clinical candidate CHK1 inhibitor would
require balancing the lipophilicity and basicity of the molecules to generate potent CHK1
biochemical inhibition that translated to effective cellular activity and favorable PK properties,
while minimizing the risk of ion channel inhibition. We focused on 5ꢀ((4ꢀ((morpholinꢀ2ꢀ
ylmethyl)amino)pyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀcarbonitriles related to morpholine 12. To help
prioritize the selection of Cꢀ5 substituents, three in silico classification models were generated.
ACS Paragon Plus Environment
15

Page 17 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
These considered: CHK1 biochemical potency, ratio of CHK1ꢀdependent cellular activity to
CHK1 biochemical potency, and inhibition of the hERG ion channel (see Supporting
Information). The CHK1 biochemical potency model classified compounds with IC < 25 nM as
5
0
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
active, while the model for the ratio of CHK1ꢀdependent cellular and enzyme activities classed
as desirable those compounds that had a less than 50ꢀfold ratio of cellular to enzyme potency.
The focus on the translation to cellular potency was driven by the anticipated need for highly
potent compounds to achieve a low projected dose in humans. As CHK1 inhibitors will be used
as potentiators of conventional cytotoxic chemotherapy, it is important to minimize nonꢀ
mechanism related toxicity to achieve tolerable combination schedules. Lastly, the hERG model
classified compounds as hERGꢀinactive if their percentage inhibition was less than 40% @ 10
µM. Virtual molecules were constructed from two chemotypes that varied in one substituent
position (Figure S1, Supporting Information). The first structure, relevant to this work, was
based on morpholine 12 with variation of the Cꢀ5 substituent. The other structure considered was
18
for the previously described series related to compound 1. Of the substituents selected to
replace the Cꢀ5 group in the virtual library, 135 were selected by medicinal chemists and 1514
were extracted from commercially available building blocks. The combinatorial library was
enumerated and, after the application of Lipinski RuleꢀofꢀFive heuristics, 1649 virtual structures
were taken forward for prediction in the three models. The resulting predictions for each of these
virtual molecules were used to aid selection of compounds for synthesis, using the normalized
probabilities of satisfying the objectives of each of the three models and visual inspection.
Representative
5ꢀ((4ꢀ((morpholinꢀ2ꢀylmethyl)amino)pyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀ
carbonitriles related to morpholine 12 are described in Table 4. Replacing the ester of 12 with
substituted phenyl groups led to the single enantiomers 22-29. In vitro inhibition of CHK1 was
ACS Paragon Plus Environment
16

Journal of Medicinal Chemistry
Page 18 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
higher for the morpholine (R)ꢀenantiomers (compare 22 vs. 23 and 25 vs. 26) consistent with the
observation of the (R)ꢀenantiomer of 12 bound to CHK1 (Figure 3). Compounds 22, 23, 25 and
26 were tested for inhibition of CHK2 and retained >100ꢀfold selectivity for CHK1. The cellular
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
profiles of these compounds were similar to 12, with no improvement in cellular potency except
for the 3ꢀfluorophenyl analogue 25. hERG inhibition was increased over 12 for these lipophilic
analogues. Replacing the Cꢀ5ꢀaryl groups by monocyclic heteroaromatics improved cellular
potency, e.g. thiazole 30 and Nꢀmethyl pyrazole 31. The (R)ꢀenantiomer of the morpholine was
again preferred (compare 31 vs. 32). Combining the polar pyrazole and lipophilic phenyl
functionalities in the Nꢀmethylindazoles 33 and 34 compromised the cellular potency and
selectivity compared to the corresponding pyrazole 31.
The ability of selected compounds to potentiate the efficacy of the cytotoxic drug gemcitabine
in p53ꢀdeficient SW620 human colon cancer cells was investigated (Table 4). The cell active Cꢀ
5ꢀaryl analogues 25 and 26 had moderate cellular selectivity, and both potentiated the efficacy of
gemcitabine. While the thiazole and pyrazole analogues 30ꢀ32 strongly potentiated gemcitabine
cytotoxicity in cells, the indazoles 33 and 34 surprisingly showed minimal activity. The cellular
selectivities of 33 and 34 were reduced compared with other analogues, suggesting that offꢀtarget
activities may be confounding the CHK1ꢀdependent effects.
ACS Paragon Plus Environment
17

Page 19 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4: Properties of selected substituted 5ꢀ((4ꢀ((morpholinꢀ2ꢀylmethyl)amino)pyridinꢀ2ꢀ
yl)amino)pyrazineꢀ2ꢀcarbonitriles.
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
3
No.
R
R
R
CHK1
Checkpoint Cellular
Cellular hERG
a
IC (nM) abrogation
selectivity efficacy %inhibition
(fold)
5
0
b
c
d
e
IC (nM)
(fold)
(IC , µM)
5
0
50
22
23
24
25
26
(R)ꢀA
(S)ꢀA
(S)ꢀA
(R)ꢀA
(S)ꢀA
(R)ꢀA
(R)ꢀA
(R)ꢀA
(R)ꢀA
(R)ꢀA
(S)ꢀA
(R)ꢀA
(R)ꢀA
(R)ꢀA
B
B
B
B
B
B
B
B
C
4ꢀOMe 6.4 (±0.1) 900
4.3
4.5
6
ꢀ
86
f
4ꢀOMe 11
1500
ꢀ
73
4ꢀF
25 (±1)
800
ꢀ
64
3ꢀF
6.4 (±0.6) 120
5.8
10
2.2
ꢀ
7.9
4.8
ꢀ
71 (8.8)
71 (8.1)
63
3ꢀF
17 (±4)
5.2 (±1)
14 (±0.5)
12 (±2)
390
600
ꢀ
27
28
29
30
3ꢀCl
3,4ꢀF
ꢀ
ꢀ
2
3,5ꢀF
630
13
6.7
6.3
11
3.5
4.2
3.8
4.3
7.6
8.2
7.4
8
38
2
f,g
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
0.5
38 (56)
67 (8.6)
26 (22)
51 (16)
69
g
31
32
33
34
35
D
D
E
0.9 (±0.1) 60
3.2 (±0.8) 400
8.5 (±1)
1000
1.9
1.9
ꢀ
F
4.6 (±0.4) 240
f,g
G
0.5
280
39
ACS Paragon Plus Environment
18

Journal of Medicinal Chemistry
Page 20 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
3
3
3
4
4
6
7
8
9
0
(S)ꢀA
(S)ꢀA
(S)ꢀA
G
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
2.2 (±0.4) 180
2.1 (±0.1) 72
5.2 (±0.8) 100
9.2
12
29
24
9
8.6
5.6
22
ꢀ
28 (26)
47 (16)
68 (8.2)
8
H
ꢀC≡CH
racꢀ(A) ꢀC≡N
23 (±3)
1600
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(S)ꢀA
(R)ꢀA
(R)ꢀA
(R)ꢀA
ꢀcycꢀPr
7.4 (±1.4) 100
7.3
17
18
ꢀ
79 (6.6)
76 (6.2)
83 (6.1)
39
h
ꢀCF
ꢀCl
ꢀF
1.3 (±0.4) 30 (±11)
23
14
ꢀ
3
h
41
42
1.9 (±0.8) 157 (±37)
14 (±1)
ꢀ
a
b
Determined in Caliper microfluidic assay, mean (±SD) for n≥2. Abrogation of etoposideꢀ
induced G2 checkpoint arrest in HT29 human colon cancer cells, single determination. Ratio of
antiproliferative GI50 (measured in duplicate by SRB assay) to IC50 for CHK1ꢀmediated
c
d
abrogation of etoposideꢀinduced G2 checkpoint arrest in HT29 human cancer cells. Foldꢀ
potentiation of gemcitabine cytotoxicity in SW620 human colon cancer cells, determined as the
ratio of GI (test compound)/GI (test compound + fixed dose Gemcitabine) in an SRB assay
5
0
50
e
(
(
ref 19). Inhibition of hERG ion current in HEK cells overexpressing hERG ion channel
PatchExpress, Millipore Inc.), single determination at test concentration of 10 µM, IC
determination (n=1) in parentheses. Single determination in DELFIA assay. Below tightꢀ
binding limit of assay. Mean (±SD), n≥3.
5
0
f
g
h
Key Cꢀ5ꢀaryl and Cꢀ5ꢀheteroaryl analogues were screened for in vivo PK properties (Table 2).
The 3ꢀfluorophenyl derivative 25 showed a similar in vivo clearance and bioavailability to the
ester 5. The thiazole 30 had low in vivo clearance but poor oral bioavailability, presumed to be
due to poor absorption and consistent with the 20ꢀfold lower volume of distribution for this
compound. The potent pyrazole 31 was assessed in vivo by limited sampling PK, where a drop in
plasma levels at 6 h to <20% of those at 1 h indicated a high clearance relative to 5 and the
compound was not profiled further.
In tandem with (hetero)aryl groups, nonꢀaryl Cꢀ5ꢀsubstituents were prepared in the 5ꢀ((4ꢀ
((morpholinꢀ2ꢀylmethyl)amino)pyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀcarbonitrile series (Table 4). The
2
ꢀmethylbutꢀ3ꢀynꢀ2ꢀols 35 and 36 potently inhibited CHK1 with cellular activity and selectivity
ACS Paragon Plus Environment
19

Page 21 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
similar to the aryl substituted analogues. The 1ꢀethynylcyclopentanꢀ1ꢀol analogue 37 increased
potency in vitro and in cells. Both 36 and 37 potentiated the efficacy of gemcitabine in SW620
cells but the PK profile of 36 showed high clearance and low oral bioavailability (Table 2).
The terminal alkyne derivative 38 was found to retain cellular activity, and to have
exceptionally high cellular selectivity compared to other compounds in this series. The
potentiation of the antiꢀproliferative effects of gemcitabine by the alkyne 38 was also increased.
The nitrile analogue 39 gave very high cellular selectivity but with reduced biochemical and
cellular potency. Other small substituents at Cꢀ5 were investigated (Table 4) and the structureꢀ
activity relationship for the potentiation of gemcitabine cytoxicity in SW620 human colon cancer
cells revealed an inverse correlation to the size of the Cꢀ5 substituent, as measured by molecular
weight or heavy atom count (Figure 4). The cellular selectivity for CHK1ꢀmediated effects
followed a similar trend. The crystal structure of 12 bound to CHK1 (Figure 3) indicated that
substitution at the Cꢀ5ꢀposition would become directed into solvent with increasing size rather
than making contacts with the CHK1 protein. We therefore speculate that the structureꢀactivity
relationships for high cellular efficacy and selectivity reflected confounding offꢀtarget
interactions, possibly one or more kinases, that increased with increasing size of the Cꢀ5ꢀ
substituent and interfered with the ability of the CHK1 inhibitor to potentiate genotoxic efficacy.
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
20

Journal of Medicinal Chemistry
Page 22 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Relationship of efficacy and selectivity in SW620 colon cancer cells to molecular
weight for 5ꢀ((4ꢀ((morpholinꢀ2ꢀylmethyl)amino)pyridinꢀ2ꢀyl)amino)pyrazineꢀ2ꢀcarbonitriles
with Cꢀ5 substituents of various sizes, exemplified in Table 4. Points are colorꢀcoded by CHK1
cellular selectivity (red – low, green – high) and labelled with compound identifiers.
Among the smaller substituents investigated at Cꢀ5, the cyclopropyl analogue 40 maintained
potency but with decreased cellular selectivity and efficacy. In contrast, both the trifluoromethyl-
4
and chloroꢀderivative 41 had excellent potency and cellular selectivity, and were strong
potentiators of gemcitabine efficacy in cancer cells. Consistent with the cellular profile, these
compounds were more than 1000ꢀfold selective for CHK1 over CHK2 and CDK1. The
ACS Paragon Plus Environment
21

Page 23 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
importance of the presence of a suitable Cꢀ5 substituent for CHK1 affinity was shown by the
fluoro analogue 42 which was tenꢀfold less potent than 4. Although more potent and efficacious
in cells, the smaller and more lipophilic Cꢀ5 substituents were associated with micromolar hERG
inhibition. This was balanced by an improvement in PK properties. Thus, the terminal alkyne 38
and cyclopropane 40 had improved oral bioavailability compared to 5, while both the
trifluoromethyl 4 and chloro analogue 41 had substantially reduced in vivo clearance and
increased oral bioavailability (Table 2). Inhibition of CYP450 enzymes was determined for
selected analogues, and compounds 36, 38, 40, 41 and 4 all showed IC >10 µM against the 6
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0
isoforms tested (1A2, 2A6, 2C9, 2C19, 2D6, 3A4).
Optimization of in vivo target engagement and efficacy
In parallel with the optimization of in vitro and PK properties, the PD properties of this series
of CHK1 inhibitors were assessed in vivo. Following DNA damage, CHK1 is activated through
phosphorylation on Ser317 and Ser345 by the upstream kinase ATR, and undergoes
autophosphorylation on Ser296. We therefore chose inhibition of the gemcitabineꢀinduced
CHK1 Ser296 autophosphorylation as a direct PD biomarker of CHK1 inhibition in vivo.
Selected compounds with moderate to good oral bioavailability, high cellular selectivity and
gemcitabineꢀpotentiating efficacy were assessed for tolerability in mice when administered in
combination with gemcitabine (60 mg/kg i.v.). The compounds were then tested at the maximum
tolerated combination dose for their ability to inhibit gemcitabineꢀinduced pSer296 CHK1 in
SW620 human colon cancer xenografts. It has been shown that maximum potentiation of
genotoxic efficacy in cells and in vivo requires sustained inhibition of CHK1 signaling to prevent
effective tumor cell DNA repair, and inhibition of CHK1 for at least 24 – 48 h following
ACS Paragon Plus Environment
22

Journal of Medicinal Chemistry
Page 24 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
14,15
administration may be desirable.
We therefore sought compounds showing robust pSer296
inhibition up to 24 h following a single oral dose.
The dose of compound 25 (80 mg/kg) was restricted by toxicity, although a partial inhibition
of the CHK1 pSer296 signal was seen at 6 h (data not shown). In contrast, single doses of
compounds 38 (300 mg/kg), 41 (300 mg/kg), and 4 (300 mg/kg) were very well tolerated.
Compound 38 inhibited the gemcitabineꢀinduced autophosphorylation of CHK1 at 6 h (not
shown), but CHK1 kinase activity had recovered by 24 h (Figure 5A). In contrast, the two
compounds 41 and 4 inhibited gemcitabineꢀinduced pS296 CHK1 for up to 24 h (Figure 5B, C).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
23

Page 25 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5: Inhibition of gemcitabine (60 mg/kg i.v.)ꢀinduced pSer296 CHK1 autophosphorylation
at 24 h by 300 mg/kg p.o. of A) 38, B) 41 and C) 4 in SW620 human colon cancer xenografts.
Dashed line indicates where western blot image has been cut. D) Total inhibitor concentrations
of 38, 41 and 4 in plasma and SW620 tumor xenografts at 24 h after a single 300 mg/kg p.o. dose
(mean±SD, n=3).
Inhibitor concentrations at 24 h were measured in plasma and the tumor xenografts (Figure
5D). Plasma levels of between 4 – 8 µM were seen for 38, 41 and 4. Similar levels of 38 were
observed despite the higher clearance of this compound compared with 41 and 4 (Table 2).
Distribution to tumor tissue was seen for all compounds, with 4 achieving the highest tumor
concentration.
The antitumor efficacies of 41 and 4 were investigated in combination with gemcitabine in
HT29 human colon cancer xenografts. The dosing schedule was selected based on our previous
1
4
studies of the oral CHK1 tool compound 1 and involved the administration of two doses of the
CHK1 inhibitors at 24 and 48 h after gemcitabine (100 mg/kg i.v.) in each of three, weekly
treatment cycles. Compounds 41 and 4 were dosed at their maximum tolerated dose (MTD) (150
ACS Paragon Plus Environment
24

Journal of Medicinal Chemistry
Page 26 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mg.kg p.o.) for this schedule in the xenograft tumor model (Figure 6). Both compounds were
well tolerated in combination with no significant decrease in body weight. A strong potentiation
of the antiꢀtumor effect of gemcitabine was seen, consistent with the cellꢀbased studies (Table 4),
with minimal effects of the CHK1 inhibitors alone in this model. At these doses, tumor
regressions were observed in 83% (41) and 100% (4) of the combinationꢀtreated cohorts at day
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
17 immediately following completion of the final treatment cycle.
Figure 6: Antiꢀtumor efficacies of 41 (150 mg/kg p.o.) and 4 (150 mg/kg p.o.) in combination
with gemcitabine (100 mg/kg i.v.) in HT29 colon cancer xenografts. The CHK1 inhibitors were
administered as suspensions in DMSOꢀTween80ꢀPEG400ꢀwater (0.05:0.2:0.65:0.1) 24 and 48 h
after gemcitabine in each of three weekly treatment cycles. Inset: body weights of cohorts as
described in main figure. N=4ꢀ6 per cohort.
ACS Paragon Plus Environment
25

Page 27 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The in vitro and in vivo PK properties of 41 and 4 were investigated further (Table 5). Both
compounds had low oral bioavailability in rats. Although both were high clearance compounds,
4
had a lower clearance than 41. Pleasingly, determination of the intrinsic clearance (CLint) in
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
hepatocytes showed both compounds were substantially more stable in human hepatocytes than
mouse, and 4 was more stable in human than rat and dog hepatocytes. In particular, 4 had a
favorable low human hepatocyte CLint. Predictions of human clearance were made using
2
9
physiologicalꢀbased PK (PBPK) modelling and suggested that 4 would have a clearance in
humans of 6% liver blood flow.
Table 5: Additional in vivo and in vitro PK data for 41 and 4.
41
4
Rat PK
CL (L/h/kg) 5.17
3.50
Vss (L/kg)
F (%)
7.45
4.60
22
38
b
Hepatocyte CLint
Mouse
22.1 (±2.6)
25.7 (±3.4)
43.4 (±9.96)
55.0 (±2.1)
2.72 (±0.61)
6
a
c
(
µL/min/10 cells)
Rat
n.d
c
Dog
n.d
d
Human
4.82 (±0.91)
a
b
Determined at Cyprotex. Incubations at 3 µM. Mean (±SE) Incubations in male CDꢀ1
c
d
mouse hepatocytes. not determined. Incubations in mixed gender human hepatocytes.
In vitro and in vivo profile of compound 4
Compound 4 was further profiled in vitro and in vivo. The solubility of 4 in pure water was
ꢀ1
ꢀ1
0
.09 mg mL , rising to 1.5 mg mL in pH 6.5 phosphate buffer, consistent with the moderate
3
0
30
basicity (pKa = 7.87 ±0.07) and lipophilicity (LogD = 1.47 ±0.04) and indicating good
7
.4
ACS Paragon Plus Environment
26

Journal of Medicinal Chemistry
Page 28 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
prospects for formulation as a salt. Compound 4 showed passive permeability across a CaCo2
ꢀ6
ꢀ1
cell monolayer (P = 25 x10 cms ). The plasma protein binding of 4 was moderate (mouse Fu =
e
0.19; human Fu = 0.21, measured using dialysis with pH control). No change in the metabolic
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
stability of 4 was observed in human liver microsomes in the presence and absence of
glutathione, indicating that 4 does not undergo metabolic activation. While some electronꢀ
deficient heterocyclic nitriles have been reported to be chemically reactive under physiological
31
conditions, we speculate that the balanced electronic donorꢀacceptor substitution pattern of the
5ꢀaminopyrazineꢀ2ꢀcarbonitrile prevents this reactivity. Compound 4 was not metabolized by
human aldehyde oxidase.
The kinome selectivity of 4 was assessed against a panel of 124 representative kinases in
3
2
substrate phosphorylation assays ([ATP] = Km,ATP). Compound 4 showed <50% inhibition at 10
µM for 85/121 kinases, representing a selectivity for CHK1 inhibition of >5000ꢀfold over these
enzymes, while 12/124 (including CHK1) showed >80% inhibition at 10 µM (Table S2,
Supporting Information). The activity of these enzymes, along with the potential offꢀtargets
3
3
CDK1/2 and 5'ꢀAMPꢀactivated protein kinase (AMPK) were titrated (Table S3, Supporting
Information). Subꢀmicromolar activity was identified for seven kinases in this panel, with 4
showing 100ꢀfold selectivity for CHK1 over the most potent offꢀtarget kinase activity identified
(mitogenꢀactivated protein kinase 15, ERK8) and more than 200ꢀfold selectivity for CHK1 over
other kinases tested. These data confirmed the expectation of high CHK1 selectivity from the
cellular selectivity of 4. While mediumꢀthroughput wholeꢀkinome profiling during lead
optimization is feasible, interpretation of the functional relevance of offꢀtarget binding assays
34
remains difficult. Our experience with this inhibitor series shows that appropriate cellꢀbased
ACS Paragon Plus Environment
27

Page 29 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mechanistic assays can drive compounds towards high kinome selectivity based on an integrative
assessment of onꢀ vs. offꢀtarget outcomes in cancer cell lines.
The ability of compound 4 to potentiate the antiꢀcancer efficacy of DNAꢀdamaging agents was
10
11
12
13
14
15
16
17
18
19
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
21
determined in a wider range of cancer cell lines (Supporting Information, Table S4).
Potentiation of the efficacy of the antiꢀmetabolite gemcitabine (8 – 23ꢀfold) was significantly
greater than that for SN38 (1.8 – 3ꢀfold), the active metabolite of the clinical topoisomerase 1
inhibitor irinotecan. Importantly, in addition to efficacy in colon cancer cell lines, compound 4
strongly potentiated gemcitabine efficacy in nonꢀsmall cell lung cancer (NSCLC) and pancreatic
cancer cell lines. These represent clinical indications where gemcitabine is part of current
chemotherapy regimens, and where tumors have a high frequency of p53 mutation or lossꢀofꢀ
3
5,36
function.
Compromise of the p53ꢀdependent DDR checkpoint may enhance the effectiveness
7
of CHK1 inhibition in combination with genotoxic agents. The inhibition of CHK1 and antiꢀ
tumor efficacy of 4 was confirmed in NSCLC and pancreatic cancer tumor xenografts. In
particular, 4 potentiated the efficacy of a gemcitabine and carboplatin combination in RAS
21
mutant Calu6 human NSCLC xenografts. In vivo efficacy in combination with gemcitabine was
further demonstrated in the KPC1 genetically engineered mouse model (GEMM) of p53ꢀ and Vꢀ
Kiꢀras2 Kirsten rat sarcoma viral oncogene homolog (KRAS)ꢀdeficient pancreatic ductal
3
7
adenocarcinoma. In cells derived from this model, 4 potentiated the antiꢀproliferative effect of
gemcitabine 18ꢀfold in vitro. Allografts from established KPC1 mouse tumors were grown
subcutaneously in female NCr athymic mice, and treated with gemcitabine and three dose levels
of 4 using the schedule developed during lead optimization. Some single agent efficacy of the
highest dose of CHK1 inhibitor was observed in this model in addition to doseꢀdependent
potentiation of gemcitabine efficacy (Figure 7).
ACS Paragon Plus Environment
28