Isoxazolo[3,4-b]quinoline-3,4(1H,9H)-diones as unique, potent and selective inhibitors for Pim-1 and Pim-2 kinases: Chemistry, biological activities, and molecular modeling

Article abstract of DOI:10.1016/j.bmcl.2008.08.079

A series of isoxazolo[3,4-b]quinoline-3,4(1H,9H)-diones were synthesized as potent inhibitors against Pim-1 and Pim-2 kinases. The structure-activity-relationship studies started from a high-throughput screening hit and was guided by molecular modeling of inhibitors in the active site of Pim-1 kinase. Installing a hydroxyl group on the benzene ring of the core has the potential to form a key hydrogen bond interaction to the hinge region of the binding pocket and thus resulted in the most potent inhibitor, 19, with K_i values at 2.5 and 43.5 nM against Pim-1 and Pim-2, respectively. Compound 19 also exhibited an activity profile with a high degree of kinase selectivity.

Full text of DOI:10.1016/j.bmcl.2008.08.079

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5206–5208
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Isoxazolo[3,4-b]quinoline-3,4(1H,9H)-diones as unique, potent and selective
inhibitors for Pim-1 and Pim-2 kinases: Chemistry, biological activities,
and molecular modeling
*
Yunsong Tong , Kent D. Stewart, Sheela Thomas, Magdalena Przytulinska, Eric F. Johnson, Vered Klinghofer,
Joel Leverson, Owen McCall, Niru B. Soni, Yan Luo, Nan-horng Lin, Thomas J. Sowin, Vincent L. Giranda,
Thomas D. Penning
Cancer Research, Global Pharmaceutical R&D, Abbott Laboratories, R47S, AP10, 100 Abbott Park Road, Abbott Park, IL 60064, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of isoxazolo[3,4-b]quinoline-3,4(1H,9H)-diones were synthesized as potent inhibitors against
Pim-1 and Pim-2 kinases. The structure–activity-relationship studies started from a high-throughput
screening hit and was guided by molecular modeling of inhibitors in the active site of Pim-1 kinase.
Installing a hydroxyl group on the benzene ring of the core has the potential to form a key hydrogen bond
interaction to the hinge region of the binding pocket and thus resulted in the most potent inhibitor, 19,
with Ki values at 2.5 and 43.5 nM against Pim-1 and Pim-2, respectively. Compound 19 also exhibited an
activity profile with a high degree of kinase selectivity.
Received 11 July 2008
Revised 20 August 2008
Accepted 22 August 2008
Available online 28 August 2008
Keywords:
Pim-1 kinase inhibitors
Pim-2 kinase inhibitors
Ó 2008 Elsevier Ltd. All rights reserved.
Pim-1 was identified as a Myc-cooperating oncogene linked to T
cell lymphomagenesis and is preferably activated by proviral inser-
tion in Moloney murine leukemia virus in lymphoblastic T cells.1
Related studies have led to the discovery of a small family of ser-
ine/threonine Pim kinases including Pim-1, Pim-2, and Pim-3 with
a high level of sequence homology among them.2 Pim-1 and Pim-2
kinases have been overexpressed in a variety of human hematopoi-
etic malignances such as myeloid or lymphoblastic leukemia and
different forms of lymphomas.3 In addition, the two kinases func-
tion in pathways distinct from/parallel to mTOR (a rapamycin tar-
get) in stimulated T cells, demonstrating an immunotherapeutical
implication.4 There is also evidence to show that Pim-1 may be in-
volved in prostate cancers thereby extending its role to solid tu-
mors.5 These findings suggest a potential role for small molecule
Pim-1 and/or Pim-2 inhibitors as anti-cancer agents. Recent re-
ports of X-ray co-crystal structures of an ATP analog and other
small molecule Pim-1 inhibitors bound to the active site of the
Pim-1 kinase6 may facilitate the drug discovery process.
3,4(1H,9H)-dione core, as
a
potent inhibitor for Pim-1
(Ki = 21.8 nM) and Pim-2 (Ki = 174 nM) kinases. This paper will dis-
close the general synthesis, preliminary structure–activity-rela-
tionships (SAR) studies, and the molecular modeling of this class
of compounds. A key finding was that, as postulated initially, a
suitable substitute at R1 could establish an important hydrogen
bond interaction with the hinge region of the active site motif of
the Pim-1 kinase and lead to higher potency.
O
N
O
O
O
F
R1
Cl
O
O
N
N
Cl
N
H
R3
R2
1
The synthesis of compounds with R2 variations (1, 6–8) is shown in
Scheme 1. The commercially available bis-fluoro compound 2 was
treated with NaH and carbonic acid diethyl ester to give 3. The eno-
late of 3 underwent nucleophilic addition to an alkyl isothiocyanate
followed by methylsulfanylation led to intermediate 4. Upon de-
protonation, the amino group of 4 replaced the neighbouring fluoro
on the benzene ring to furnish the quinolinone. Subsequent sulfiny-
lation using mCPBA resulted in 5. The isoxazolone ring of the final
products was formed upon treatment of 5 with hydroxylamine.
The analogs with R3 variations (9-13, Scheme 2) were synthesized
conveniently from 5 (where R2 = ethyl) in the presence of an N-alkyl
Much of the interest in Pim kinases has been focused on biology
and structural biology aspects of the targets. Here, we report our
medicinal chemistry efforts leading to a novel class of potent and
selective inhibitors against both Pim-1 and Pim-2 kinases. High-
throughput-screening (HTS) of internal compound collections
identified 1, a molecule containing an isoxazolo[3,4-b]quinoline-
* Corresponding author. Tel.: +1 847 935 1252; fax: +1 847 935 5165.
E-mail address: yunsong.tong@abbott.com (Y. Tong).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.08.079

Y. Tong et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5206–5208
5207
O
ii
H3CO
Cl
Br
F
O
F
O
H3CO
i
F
i
F
ii, iii
OEt
Cl
F
15
14
16
Cl
F
Cl
2
3
O
O
H3CO
Cl
CO2Et
iii, iv, v, vi
vii
H3CO
Cl
O
O
O
F
CO2Et
iv, v
F
vi
OEt
N
S
F
F
HN
O
Cl
N
R2
S
Cl
SMe
R2
17
O
5
4
O
N
O
O
O
O
H3CO
Cl
HO
Cl
viii
O
O
O
R2=Et, 1
F
N
N
N
H
H
R2=i-Pr, 6
O
R2=n-Bu, 7
N
Cl
N
R2
18
19
H
R2=benzyl, 8
Scheme 3. Reagents and conditions: (i), Br2, H2SO4; (ii) a—PdCl2, PPh3,
CH2@CHOBu, Na2CO3; b—HCl, Ref. 8; (iii) NaH, EtNCS, DMF; (iv) MeI, 48% (2 steps);
(iv) NaH, THF; (vi) mCPBA, 63% (2 steps); (vii) NH2OHꢀHCl, NaOAc, THF, H2O, 41%;
(viii) AlCl3, toluene, 43%.
Scheme 1. Reagents and conditions: (i) NaH, (EtO)2CO, 50%; (ii) NaH, R2NCS, DMF;
(iii) MeI, 42–71% (2 steps); (iv) NaH, THF, 78–94%; (v) mCPBA, CH2Cl2, 69–87%; (vi)
NH2OHꢀHCl, NaOAc, THF, H2O, 8–77%.
hydroxylamine. It is noteworthy that N-methyl hydroxylamine gave
a 1:1 mixture of the N–O regio-isomers7 even though only the
structure of the desired isomer 9 is shown in Scheme 2. The mixture
was difficult to separate and thus was tested in the assay without
separation. Compound 13 was derived from de-benzylation of 12
using iodotrimethylsilane.
Scheme 3 outlines the synthesis of compounds 18 and 19. 2-
Chloro-4-fluoro-1-methoxy-benzene 14 was brominated to give
15, which was in turn acylated via a Heck coupling to afford 16 fol-
lowing literature procedures.8 Compound 16 was converted into
18 utilizing the same protocols as described in Scheme 1. De-meth-
ylation of 18 in the presence of AlCl3 provided the hydroxy-bearing
inhibitor 19.
X-ray co-crystal structures of the ATP analog AMP-PNP (50aden-
ylyl-b, c
-imidodiphosphate)6a or small molecule Pim-1 inhibitors
bound to the active site of Pim-1 kinase have been reported re-
cently.6 When the HTS hit 1 was docked into the AMP-PNP co-crys-
tal structure (Fig. 1), it was notable that it lacked any strong
hydrogen bonding to the backbone carbonyl of Glu121 in the hinge
region. The interaction between the ligand and the protein was in-
stead anchored by a hydrogen bond between the carbonyl of the
isoxazolone and the charged amino group of Lys67 (distance was
about 3.0 Å). The oxygen on the quinolinone moiety of the inhibi-
tor also bound to Lys67 via a water molecule. This phenomenon
was observed with other Pim-1 inhibitors.6b,6e The region occupied
by R2 is exposed to the solvent front. Consequently, the modifica-
tions at R2 with bulkier and longer alkyl groups did not alter the
Ki value significantly (see Table 1, 1 vs 6, 7, and 8). The R3 group
projects toward the Gly-rich loop of the active site and is in close
proximity to the sidechain of Phe49 and the mainchain of Gly45.
These residues limit the steric size of the R3 group that can be tol-
Figure 1. Model of 1 (green carbon atoms) in Pim-1 kinase (orange carbon atoms).
The outline of the active site is shown as an orange surface. The 3-D structure of 1
was energy minimized (CFF force field, InsightII software, Accelrys, San Diego, CA)
and manually docked into the crystal structure of Pim-1 kinase (PDB code: 1YI3) so
as to optimize the Van der Waals contact with the hinge residues Glu 121 and Pro
123 and make H-bonds with Lys 67 and a conserved water molecule (Wat 21 in
1YI3, Ref. 6b; Wat 17 in 1YXV, Ref. 6c; Wat 64 in 2OBJ, Ref. 6e). Other orientations
were considered, but were observed to have sub-optimal H-bond or VdW
interactions. Hydrogen bond distances between heavy atoms are indicated in
Angstroms. The structure of AMP-PNP (1YXT) is shown in white for reference.
Table 1
Inhibitory activities10 against Pim-1 and Pim-2 kinases
Compound
Pim-1 Ki (nM)a
Pim-2 Ki (nM)a
O
1
6
7
8
21.8
12.3
15.0
36.5
50.1
1271
271
4894
1162
2698
2.5
174
62.8
165
558
>1424
>1424
>1424
>1424
>1424
>1424
43.5
O
O
F
F
CO2Et
i
O
N
R3
Cl
N
Cl
N
R2
S
9b
10
11
12
13
18
19
O
R3= Me, 9
5 (R2=ethyl)
R3= benzyl, 10
R3= CH2-(3-OH-Ph), 11
R3= (CH2)2OCH2Ph, 12
R3=(CH2)2OH, 13
ii
a
Scheme 2. Reagents and conditions: (i) R3-NHOH, NaOAc, THF, H2O, 7–26%; (ii)
Me3SiI, ClCH2CH2Cl, 32%.
Values are the mean of at least two measurements (except 12, n = 1).
1:1 mixture of two regio-isomers.
b

5208
Y. Tong et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5206–5208
erated. Replacing the proton at R3 of compound 1 with an alkyl
species larger than a methyl resulted in inhibitors with at least
12-fold less potency against Pim-1 (see Table 1, 9 to 13) and also
with less activity against Pim-2. The methylated inhibitor 9 was as-
sayed as a mixture (ꢁ1:1) of two regio-isomers.7 Since 9 was only
slightly less active than 1 (Pim-1 Ki: 50.1 nM vs 21.8 nM) it does
not appear that methylation had much impact electronically on
the H-bond between the carbonyl of the isoxazolone and Lys67.
In addition, neither isomer by itself would be significantly more
potent than 1. However, 9 did suffer a loss of potency against
Pim-2.
In order to take advantage of a potential hydrogen bond between
the ligand and the hinge region of Pim-1 kinase, the fluoro group on
the benzene ring of 1 was replaced with a hydroxyl. As shown in
Fig. 2, the hydroxyl group on inhibitor 19 can donate an H-bond to
Glu121 O with a desirable distance of 3.0 Å. The impact of this
designed hydrogen bond on affinity was uncertain because of the
complicated energetics of solvation/desolvation of ligand, protein,
and protein-ligand complex.9 Additionally, any lone-pair repulsion
between the fluoro group of 1 and Glu121 O would presumably be
eliminated in 19 and could also be a factor in any potency change.
Without a clear prediction from this complex interplay of molecular
forces, we were able to evaluate compound 19 and found that its hy-
droxy improved the Ki value (Pim-1) of 19 over 1 by about 8 fold
(2.5 nMvs21.8 nM, Table1). Compound19alsodisplayedhigherpo-
tency against Pim-2 as compared to 1 (43.5 nM vs 174 nM). The
importance of the hydroxyl group on 19 was further demonstrated
by inhibitor 18 where the hydroxyl was replaced by a methoxy,
showing much reduced activities against both Pim-1 and Pim-2
(2698 nM and >1424 nM, respectively, Table 1). Compound 19 was
assayed in a panel of 22 serine/threonine kinases (Table 2) and dis-
played an activity profile with a high degree of selectivity. As com-
pared to its potent inhibitory activity against Pim-1 (6.3 nM in this
assay format), this molecule was weak against all other kinases
Table 2
Kinase selectivity profile of 19
Kinase
Ki (
l
M)a
Kinase
Ki (lM)
Pim-1
Src
Zipk
0.0063
>1.94
0.0273
>8.91
>8.57
1.73
>8.21
>6.88
>6.67
>8.78
>9.52
CTAK1
Casein2
Gsk3a
MARK
PKA
PKCd
PKC c
PKCn
Plk1
Rock1
Rock2
>6.67
1.07
1.11
Cdc2
>8.44
>7.50
>8.89
>8.75
>8.33
>8.89
>9.11
>7.50
Cdk2
AMPK
Akt1
Aurora1
Aurora2
Chk1
Chk2
a
6-point screening format.11
pocket and thus resulted in a more potent inhibitor, 19, with Ki val-
ues at 2.5 and 43.5 nM against Pim-1 and Pim-2, respectively.
Compound 19 also exhibited a high degree of selectivity against
other serine/threonine kinases.
Acknowledgment
The authors thank Dr. Daniel Chu for earlier synthetic work.
References and notes
1. (a) Cuypers, H. T.; Selten, G.; Quint, W.; Zijlstra, M.; Robanus-Manndag, E.;
Boelens, W.; van Wezenbeek, P.; Melief, C.; Berns, A. Cell 1984, 37, 141; (b) van
Lohuizen, M.; Verbeek, S.; Krimpenfort, P.; Domen, J.; Saris, C.; Radaszkiewicz,
R.; Berns, A. Cell 1989, 56, 673.
2. (a) van der Lugt, N. M.; Domen, J.; Verhoeven, E.; Linders, K.; van der Gulden,
H.; Allen, J.; Berns, A. EMBO J. 1995, 14, 2536; (b) Mikkers, H.; Allen, J.;
Knipscheer, P.; Romeyn, L.; Hart, A.; Vink, E.; Berns, A. Nat. Genet. 2002, 32, 153.
3. review (a) Amaravadi, R.; Thompson, C. B. J. Clin. Invest. 2005, 115, 2618; (b)
Amson, R.; Sigaux, F.; Przedborski, S.; Flandrin, G.; Givol, D.; Telerman, A. Proc.
Natl. Acad. Sci. 1989, 86, 8857; c 1b.; (d) Cohen, A. M.; Grinblat, B.; Bessler, H.;
Kristt, D. A.; Kremer, A.; Shalom, S.; Schwartz, A.; Halperin, M.; Merkel, D.; Don,
J. Leuk. Lymphoma 2004, 45, 951.
4. Fox, C. J.; Hammerman, P. S.; Thompson, C. B. J. Exp. Med. 2005, 201, 259.
5. Valdman, A.; Fang, X.; Pang, S. T.; Ekman, P.; Egevad, L. Prostate 2004, 60, 367.
6. (a) Qian, K. C.; Wang, L.; Hickey, E. R.; Studts, J.; Barringer, K.; Peng, C.;
Kronkaitis, A.; Li, J.; White, A.; Mische, S.; Farmer, B. J. Biol. Chem. 2005, 280,
6130; (b) Jacobs, M. D.; Black, J.; Futer, O.; Swenson, L.; Hare, B.; Fleming, M.;
Saxena, K. J. Biol. Chem. 2005, 280, 13728; (c) Kumar, A.; Mandiyan, V.; Suzuki,
Y.; Zhang, C.; Rice, J.; Tsai, J.; Artis, D. R.; Ibrahim, P.; Bremer, R. J. Mol. Biol.
2005, 348, 183; (d) Pierce, A. C.; Jacobs, M.; Stuver-Moody, C. J. Med. Chem.
2008, 51, 1972; (e) Cheney, I. W.; Yan, S.; Appleby, T.; Walker, H.; Vo, T.; Yao,
N.; Hamatake, R.; Hong, Z.; Wu, J. Z. Bioorg. Med. Chem. Lett. 2007, 17, 1679.
7. 9 is a mixture (ꢁ1:1) of the following two compounds:
tested (Ki values at least over 1 lM) except Zipk (Ki = 27.3 nM).
In conclusion, guided by molecular modeling and known X-ray
co-crystal structures, we have carried out a concise SAR study fol-
lowing an HTS hit (1) and demonstrated that compounds with a
novel isoxazolo[3,4-b]quinoline-3,4(1H,9H)-dione core can serve
as potent inhibitors for Pim-1 and Pim-2 kinases. The R2 position
of the tricyclic core can tolerate various alkyl groups due to the
availability of space in the kinase active site. At R3 only a small
group, such as a proton or a methyl, can maintain the inhibitory
activity of the molecules. Installing a hydroxyl group on the ben-
zene ring (replacing F on 1) of the core has the potential to form
an optimal H-bond interaction to the hinge region of the binding
O
O
O
O
F
F
+
O
N
N
O
Cl
N
Cl
N
.
8. Takaishi, H.; Kishida, M.; Miura, Y. Japan Patent 4-356438, 1992.
9. (a) Foloppe, N.; Fisher, L. M.; Howes, R.; Kierstan, P.; Potter, A.; Robertson, A. G.
S.; Surgenor, A. E. J. Med. Chem. 2005, 48, 4332; (b) Hunenberger, P. H.; Helms,
V.; Narayana, N.; Taylor, S. S.; McCammon, J. A. Biochemistry 1999, 38, 2358.
10. Pim kinase assays: Assays were conducted as follows with final concentrations
as listed. In 384-well v-bottom polypropylene plates, 10 ml compound (2%
DMSO), was mixed with 20 ml of Pim-1 (50pM), or Pim-2 (500pM) and peptide
substrate (biotin-C6linker-VRRLRRLTAREAA) (2 mM), followed by immediate
initiation with 20 ml k-[33 P]-ATP (5 mM, 2 mCi/mmol) using a reaction buffer
comprising 25 mM HEPES, pH 7.5, 0.5 mM DTT, 10 mM MgCl2, 100 mM
Na3VO4, 0.075 mg/ml Triton X-100. Reactions were quenched after 1hr by
the addition of 50 ml stop buffer (50 mM EDTA, 2 M NaCl). 80 mL of the
stopped reactions were transferred to 384-well streptavidin-coated plates
(FlashPlate Plus, Perkin-Elmer), incubated 30 min at rt and washed 3 times
with 0.05% Tween-20/PBS using an ELX-405 automated plate washer (BioTek),
and counted on a TopCount Scintillation Plate Reader (Packard).
11. In vitro kinase selectivity assays: Assays were conducted exactly as described
for the Pim in vitro assay (Ref. 10) except for the enzyme concentration and
substrate used, which were custom to each kinase. Substrates were either
chosen from those described in literature/vendor protocols or identified
through screening of a 720 peptide Jerini Kinase Substrate Set (Jerinin AG).
Kinases specific reagents (phosphotidyl serine, diacyl glycerol, calcium
chloride, calmodulin, cGMP) were employed only where required.
Figure 2. Model of 19 in Pim-1 kinase with similar protein interactions as 1 (Fig. 1)
with the designed hydrogen bond to the backbone carbonyl of Glu 123.