Synthesis and evaluation of substituted benzoisoquinolinones as potent inhibitors of Chk1 kinase

Article abstract of DOI:10.1016/j.bmcl.2007.09.007

From HTS lead 1, a novel benzoisoquinolinone class of ATP-competitive Chk1 inhibitors was devised and synthesized via a photochemical route. Using X-ray crystallography as a guide, potency was rapidly enhanced through the installation of a tethered basic

Full text of DOI:10.1016/j.bmcl.2007.09.007

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 17 (2007) 6280–6285
Synthesis and evaluation of substituted benzoisoquinolinones
as potent inhibitors of Chk1 kinase
Robert M. Garbaccio,a,* Shaei Huang,a Edward S. Tasber,a Mark E. Fraley,a
Youwei Yan,b Sanjeev Munshi,b Mari Ikuta,b Lawrence Kuo,b Constanine Kreatsoulas,a
Steve Stirdivant,c Bob Drakas,c Keith Rickert,c Eileen S. Walsh,c Kelly A. Hamilton,c
Carolyn A. Buser,c James Hardwick,c Xianzhi Mao,c Stephen C. Beck,c Marc T. Abrams,c
Weikang Tao,c Rob Lobell,c Laura Sepp-Lorenzinoc and George D. Hartmana
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bDepartment of Structural Biology, Merck Research Laboratories, West Point, PA 19486, USA
cDepartment of Cancer Research, Merck Research Laboratories, West Point, PA 19486, USA
Received 1 August 2007; revised 28 August 2007; accepted 4 September 2007
Available online 7 September 2007
Abstract—From HTS lead 1, a novel benzoisoquinolinone class of ATP-competitive Chk1 inhibitors was devised and synthesized
via a photochemical route. Using X-ray crystallography as a guide, potency was rapidly enhanced through the installation of a teth-
ered basic amine designed to interact with an acidic residue (Glu91) in the enzyme pocket. Further SAR was explored at the solvent
front and near to the H1 pocket and resulted in the discovery of low MW, sub-nanomolar inhibitors of Chk1.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Despite their inherent toxicity, DNA damaging agents
continue to remain central in clinical cancer chemother-
apy. Therefore, strategies directed at improving their
therapeutic index are warranted. Following DNA dam-
age, normal cells arrest and attempt repair at the cell cy-
cle checkpoints G1 and S, via the tumor suppressor
protein p53, and at G2 and S via the checkpoint kinase
Chk1.1–3 Tumor cells, however, are often deficient in
p53 function (estimated 50–70% of all cancers) and thus,
must rely on Chk1 to induce arrest for survival. These
p53-deficient cancers should be more vulnerable to
Chk1 inhibition, leading to abrogation of DNA-dam-
age-induced arrest, premature progression into mitosis
and resulting in mitotic catastrophe and apoptosis. To
summarize, abrogation of the S and G2 checkpoints
should sensitize p53-deficient cancer cells to DNA dam-
aging agents without enhancing toxicity toward normal
proliferating cells. As such, Chk1 inhibitors4 have the
potential to widen the therapeutic window for clini-
cally-utilized DNA damaging agents in p53 deficient
tumors.
This effort initiated with the tetracyclic lead (1) that was
derived from an earlier Jak kinase effort at Merck.5 An
X-ray crystal structure was obtained for 1 and revealed
the expected interaction of the hydrogen bond donor/
acceptor pair of the pyridone with the kinase backbone
in the hinge region of the ATP active site. The 3-pyridyl
analog 3 demonstrated that some enhancement in po-
tency might be obtained by extension of a basic amine
in the region of the Glu91 and Glu134 residues in the
ATP pocket of Chk1.
One concern in this structure was the presence of the
imidazole group (calcd pKa for benzimidazole ꢀ5.8)
which contributed elevated polar surface area (PSA)
while likely not contributing to potency. Earlier studies
suggested that elevated PSA and multiple basic sites in
Chk1 inhibitors were detrimental to cell potency.6 In
keeping with these observations, a strategy emerged to
remove the imidazole from 1, therein producing a novel
benzoisoquinolinone core that had the potential to be
elaborated further to make key potency-enhancing con-
tacts (Fig. 1).
Keywords: Chk1 kinase; Chek1 kinase; Kinases; Photochemistry;
Benzoisoquinolinones; PSA; DNA damaging agents; Mitotic arrest;
Checkpoint escape; p53-Deficient cancer; Cancer; Apoptosis.
*
Corresponding author. Tel.: +1 215 652 5391; fax: +1 215 652
6345; e-mail: robert_garbaccio@merck.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.09.007

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6280–6285
6281
N
H
N
H
O
O
N
X
N
X
K2CO3
hv, O2
F
C9
C4
MeOH
60 ºC
65%
t-BuOH
benzene
5 d, 30%
NH
R1
CN CHO
CN
6
N
C6
4
5
R1
Benzoisoquinolinones
Cmpd
Chk1 IC50
N
N
1. NaOH
EtOH, 80 ºC X
- reduced core polarity
- improve potency
- explore SAR
H
Ph
3-pyridine
1
2
3
610 nM
1080 nM
110 nM
X
LAH
2. HCl
Et2O, 0 ºC
MeOH, 80 ºC
82%
60%
CN
N
O
OMe
N
Figure 1. HTS lead 1 and preliminary SAR along with the proposed
benzoisoquinolinone core.
7
8
O
Further analysis of the X-ray structure of 1 suggested
the placement of a basic amine approximately 4–5 ang-
stroms from C6 of the benzoisoquinolinone core would
interact with the acidic residue Glu91 and would result
in greater potency. Accordingly, it was decided to devise
a general synthetic route that could access both an eth-
ylamine and propylamine substitution at C6 from a
common starting material (Schemes 1 and 2).
X
X
NC
P(OEt)2
DMP
>95%
DBU, LiCl
78%
O
10
OH
N
9
N
NaBH4
X
X
mCPBA
MeOH/pyr
85 ºC
85%
CH2Cl2
CN >95%
Stobbe-condensation of substituted benzonitriles 4 with
pyridine-4-carboxaldehyde 5 yielded the E-azastillbenes
6 in moderate yield.7 A key step in this synthesis was
the oxidative photocyclization of 6 to the azaphenanth-
rene 7. While slow and somewhat low-yielding (30%),
this reaction was run reproducibly on 80 g scale, used
O2(g) as an oxidant, and did not require high dilution.
In addition, upon concentration the pure product 7 pre-
cipitated from the reaction mixture and was easily iso-
lated. The presence of t-BuOH was found to accelerate
the reaction consistent with an earlier observation.8
11
12
CN
O-
H
N+
O
N
Ac2O
LAH
THF
25 ºC
62%
X
X
140 ºC
then
NaOH/MeOH
77%
CN
CN
13
14
H
N
H
N
O
O
HN
N
Br
ArB(OH)2
Elaboration of the C6–CN group into primary alcohol
(9) was accomplished in three high-yielding steps: (1)
hydrolysis to the acid, (2) Fisher esterification to the
ester and (3) reduction using LAH. The alcohol was
then oxidized to the aldehyde 10. Despite efforts to
shorten the sequence from 7 to 10, no simplified
route was identified. Importantly, however, no
chromatography was required for these first 6 steps.
Homologation of the aldehyde 10 to the acrylonitrile
11 and subsequent 1,4-reduction with NaBH4 gave the
C6-ethylnitrile 12. At this stage, N-oxidation with
m-CPBA and regioselective rearrangement to the pyri-
done provided benzoisoquinolinone 14. The regioselec-
tivity of this transformation is rationalized by
resonance arguments that maintain the aromaticity of
the naphthalene core. After considerable experimenta-
tion, it was found that LAH provided the optimal
conditions for reduction of the nitrile to the primary
amine 15.
Pd(PPh3)4
Na2CO3
DMF, 130ºC
63%
NH2
NH2
15
16
Scheme 1. Photochemical route to C6-propylamine benzoiso-
quinolinones.
Conversion to the primary mesylate 19 was followed
by direct displacement with NaCN to provide the nitrile
20. Here, LAH was found to lead to significant decom-
position whereas BH3–THF provided clean reduction to
the ethylamine 21. Again, for the C9–Br, metal-medi-
ated coupling reactions allowed for exploration at C9.
The final vector that was explored in this benzoisoquin-
olinone core emanated from the C4-position. Substitu-
tion at C4 would extend into the H1 pocket9 and
surrounding hydrophobic space that is left unused by
HTS lead 1. The reactivity of the pyridone moiety (even
in the presence of the pyrazole) in these benzoisoquinoli-
nones allowed for late-stage bromination at C4 and fur-
ther Suzuki-couplings to determine the effect of
extending substitution into that region (Scheme 3).
In the sequence for which X = Br, Suzuki couplings with
various boronic acids allowed for exploration of the C9
SAR (results reported below).
Alternatively, the intermediate alcohol 9 could be direc-
ted toward the formation of the C6-ethylamine benzo-
isoquinolinones 22 (Scheme 2). Again, regioselective
conversion to the pyridone 18 was strictly observed.
These benzoisoquinolinone inhibitors proved to be excel-
lent structural surrogates for the HTS lead 1 and through
this versatile synthetic route, superior enzymatic and

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
6282
R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6280–6285
O-
N+
Table 1. Ethyl- and propylamine benzoisoquinolinones
N
H
O
N
Br
Br
Br
Br
Br
Br
Ac2O
mCPBA
CH2Cl2
>95%
X
140 o
then
NaOH/MeOH
80%
C
OH
OH
C6
9
17
a
Chek1 IC50 (nM)
H
N
H
N
Compound
X
C6
O
O
MsCl
Et3N
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Cl
Cl
Br
Cl
Br
Cl
Br
Cl
Br
Cl
Br
Br
Br
Br
Br
H
580
1060
400
360
300
510
420
12
NaCN
CN
CN
THF
DMF/H2O
(10:1)
72% two steps
CO2Me
CO2Me
OH
OMs
18
19
(CH2)2CN
(CH2)2CN
(CH2)3NH2
(CH2)3NH2
(CH2)2NH2
(CH2)2NH2
CH2OH
H
N
H
N
O
O
BH3-THF
80 oC
ArB(OH)2
17
11
Pd(PPh3)4
Na2CO3
3
NH2
440
780
44
DMF
CN
130 o
C
20
21
(CH2)2OH
(CH2)2NHMe
(CH2)2NMe2
60% two steps
H
O
N
HN
N
130
a Run at Km for ATP (0.1 mm).
NH2
protonated primary amine and Glu91 (detailed analysis
of an X-ray structure presented below).
22
Scheme 2. Conversion of intermediate
benzoisoquinolinones.
9
into C6-ethylamine
With either a propyl- or ethylamine at C6, exploration
of C9 commenced (Table 2). Substitutions at C9 are di-
rected toward the solvent front of the Chk1 ATP-bind-
ing pocket and those substitutions that would form
beneficial interactions in this environment while mini-
H
H
O
N
HN
N
O
N
HN
N
2
˚
Br2
mizing molecular PSA <100 A were sought out to max-
Br
imize cellular potency. These inhibitors were tested
against the enzyme and also in a cell-based checkpoint
escape assay that measures the ability of a Chk1 inhibi-
tor to release H1299 tumor cells from camptothecin-in-
duced cell cycle arrest.11,12 The C9–Br derivative 32 or
Cl derivative 33 gave approximately a 7-fold improve-
ment in potency relative to the unsubstituted derivative
40. Substitution to the nitrile 41 or the corresponding
amide 42 resulted in minimal intrinsic potency change,
but diminished cellular potency (note PSA increase).13
Conversion to the 3-pyrazole or 4-pyrazole (16, 22,
43–44) gave a further 3- to 4-fold potency increase and
provided compounds with excellent cell potency. Supris-
ingly, the logP value for 16 was measured to be 0.0 and
yet it appears to have sufficient cell permeability. C9-
Pyrroles 45 and 46 were sub-nanomolar inhibitors in
this series with exceptional cell potency. The hydro-
gen-bond donor ability of these pyrazoles and pyrroles
may contribute to potency as the methylated pyrazole
47 was 3-fold less potent. Inhibitors 48–50 were designed
to interact with solvent, but their additional basic amine
and increased PSA appeared to have a negative impact
on cell potency.6 Due to potential stability and meta-
bolic activation issues seen with pyrroles, the C6-pyra-
zole (as in 16 and 22) was selected as the preferred
substitution.
AcOH
>95%
+
NH3
NH2
16
23
-OAc
H
N
O
HN
N
ArB(OH)2
Ar
Pd(PPh3)4
Na2CO3
DMF
130 oC
65%
NH2
24
Scheme 3. C4 bromination and Suzuki couplings.
cellular-potency were achieved in structures with low
molecular weight (MW <300). In Table 1, a comparison
of C6 substitutions establishes that the benzoisoquinoli-
none core is potent and that both the ethyl- and propyl-
linked amines provide excellent potency enhancement.
The core itself with a C9–Cl group (25) gave an
IC50 = 580 nM against Chk1 confirming that the imidaz-
ole was not needed for potency.10 Various neutral substi-
tutions (26–31) did not greatly alter potency, however,
both the propyl amine (32 and 33) and the ethyl amine
(34 and 35) provided approximately 30-fold improve-
ments in potency relative to 25. This potency enhance-
ment is lost completely by replacement of the amine
with an alcohol (36), and is even diminished with mono-
(38) and bismethylation (39) of the amine. These observa-
tions are consistent with an ionic interaction between a
Benzoisoquinolinone 16 was brominated selectively at
C4 and a library of Suzuki reactions was conducted to

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6280–6285
6283
Table 2. C9-substituted benzoisoquinolinones
Table 3. C4-substituted benzoisoquinolinones
H
H
O
N
O
N
N
HN
C9
C4
(CH2)nNH2
NH2
Compound
n
C9
Chek1
a
IC50 (nM) (nM)
Cell EC50 PSA
2
(A )
Compound
C4
Chek1
a
IC50 (nM)
Cell
EC50 (nM)
PSA
2
(A )
˚
˚
40
32
33
41
42
3
3
3
3
3
H
98
12
17
25
11
n.t.
230
62
61
16
23
51
52
53
54
55
56
57
58
59
60
61
62
H
Br
3.5
2.3
3.6
2.0
2.0
1.0
0.3
2.6
3.7
3.5
0.8
11.5
1.2
1.2
64
140
210
140
230
67
96
95
CL
Br
310
2200
>10,000
61
88
CH@CH2
Ph
97
97
CN
CONH2
107
2-(OH)–Ph
3-(OH)–Ph
4-(OH)–Ph
2-(F)–Ph
3-(F)–Ph
4-(F)–Ph
2-(Cl)–Ph
3-(Cl)–Ph
4-(Cl)–Ph
2,4-(Cl)2–Ph
120
120
120
96
N
HN
22
16
2
3
3.0
3.5
79
64
96
96
47
370
300
370
130
1050
320
270
96
96
N
NH
43
44
2
3
2.5
4.3
100
150
95
93
97
97
97
96
NH
45
46
2
3
0.5
2.4
14
44
75
75
a Run at Km for ATP (0.1 mM).
N
MeN
47
3
9.2
957
82
caspase activation assay—a measure of apoptosis.14
HT29 cells were treated sequentially with camptothecin
followed by compound 22, then assayed for caspase 3
activity. While camptothecin by itself induces some level
of caspase activity, addition of increasing concentrations
of compound 22 following camptothecin exposure in-
creased the activity over camptothecin alone (Fig. 2).
Compound 22 by itself had no effect on caspase activity.
The results are consistent with the hypothesis that re-
lease of cells from DNA damage checkpoint arrest re-
sults in mitotic catastrophe and cell death.
H2N
48
3
6.5
1085
90
N
O
49
50
2
3
0.6
2.0
130
79
95
O
Me2N
>10,000
N
H
a Run at Km for ATP (0.1 mM).
An X-ray crystal structure of the inhibitor-bound com-
plex with Chk1 for compound 54 was obtained and ana-
lyzed in order to understand the multiple interactions
with the enzyme (Fig. 3).15 Several key hydrogen bonds
are made by this molecule that rationalize the increase in
potency afforded by the pyrazole, propylamine, and
phenol moeties and help to explain the SAR. First, the
complete our SAR assessment in the H1 pocket. While
some potency improvements were seen in terms of enzy-
matic activity, these improvements typically did not
translate into elevated cell potency. At the risk of
increasing molecular weight, it was found that consider-
able steric bulk could be accommodated at this position.
The bromide 23 was more potent against Chk1 than 16
but less potent in the checkpoint escape cellular assay.
Among the numerous groups employed, the phenols
54–56 and the chlorides 59 and 61 did increase enzy-
matic potency, but the expected increase in cell potency
was not realized relative to 16. In the case of the phenols
53–55, PSA is increased, and it is possible that this cor-
relates with some decreased permeability (relative to 16).
For the aryl halides 56–62 it is not clear why cell potency
decreases relative to 16. Lastly, the potency with the
combination of the ortho- and para-Cl substitution
was not additive in 62. The larger molecular weight of
the compounds in Table 3 (>400) may contribute to
the lack of improvement in cellular potency for these
C4-substituted inhibitors.
15000
10000
5000
0
Caspase Induction - HT29 Cells
The ability of a Chek1 inhibitor to sensitize a tumor cell
line to apoptosis and cell death was examined in a
Figure 2. Caspase-induction assay with camptothecin and 22.