Implications for selectivity of 3,4-diarylquinolinones as p38αMAP kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.12.073

In this study we report on the specificity profiling of the MAP kinase inhibitors 1, 2, and 3 in a panel of 78 protein kinases including the MAPK isoforms p38(α,β,γ,δ), JNK1/2/3, and ERK1/2/8 showing 3-(4-fluorophenyl)-4-pyridin-4-ylquinolin-2(1H)-one (1)

Full text of DOI:10.1016/j.bmcl.2007.12.073

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1431–1435
Implications for selectivity of 3,4-diarylquinolinones
as p38aMAP kinase inhibitors
Christian Peifer,a,* Robert Urich,a Verena Schattel,a Mohammed Abadleh,a Marc Ro¨ttig,b
Oliver Kohlbacherb and Stefan Laufera
aDepartment of Pharmacy, Eberhard-Karls-University, Auf derMorgenstelle 8B, D-72076 Tu¨bingen, Germany
bWilhelm-Schickard Institute of Bioinformatics, Eberhard-Karls-University, Sand 14, D-72076 Tu¨bingen, Germany
Received 28 November 2007; revised 27 December 2007; accepted 30 December 2007
Available online 5 January 2008
Abstract—In this study we report on the specificity profiling of the MAP kinase inhibitors 1, 2, and 3 in a panel of 78 protein kinases
including the MAPK isoforms p38(a,b,c,d), JNK1/2/3, and ERK1/2/8 showing 3-(4-fluorophenyl)-4-pyridin-4-ylquinolin-2(1H)-one
(1) to be highly selective for p38aMAPK with an IC50 of 1.8 lM. In contrast, besides p38a the isoxazoles 2 and 3 significantly inhib-
ited JNK2/3 and further kinases beyond the MAPK family such as PKA, PKD, Lck, and CK1. By using sequence alignment and
homology models of different members of the MAPK family the binding mode determining selectivity of 1 for the p38a isoform was
investigated. For lead optimization of 1 a straightforward tandem-Buchwald-aldol synthetic approach toward the flexible decora-
tion of the quinolin-2(1H)-one scaffold was employed. SAR for derivatives of 1 at the isolated p38aMAPK are presented.
ꢀ 2008 Elsevier Ltd. All rights reserved.
The well-characterized mitogen activated protein kinases
(MAPK) are a group of protein kinases (PK) that play
an essential role in signal transduction of diseases such
as in inflammation.1 In humans, at least 11 members of
the MAPK superfamily are known which can be divided
into 6 groups: extracellular signal-regulated protein ki-
nases (ERK1, ERK2); c-Jun N-terminal kinases (JNK1,
JNK2, JNK3); p38s (p38-alpha, p38-beta, p38-gamma,
p38-delta); ERK5; ERK3s (ERK3, p97-MAPK, ERK4);
and ERK7s (ERK7, ERK8). Among them especially
p38aMAPK (p38a) and JNK3 are considered to be prom-
isingtargetsfordrugdevelopment.2 Ingeneralthepotency
aswell as the selectivity ofPK inhibitors are important cri-
teria.3 However, in our former study compound 1 as an
inhibitor of p38a (IC50 = 1.8 lM) was found to be selec-
tive over the closely related JNK3 whereas the two iso-
details are available on SI). Herein, compound 1 signif-
icantly inhibited only the p38a isoform whereas 2 and 3
additionally blocked JNK2/3 (2 IC50 JNK2 = 1.28 lM).
Furthermore besides MAPK, 2 markedly inhibited
PKA7 (28% 2), PKD18 (60% 1), Lck9 (42% 0),
and CK1d10,11 (6% 2).
Compound 3 secondary blocked Aurora B (22% 2),
Aurora C12 (35% 3), and CK1d10 (55% 8). However,
since 1 (but not 2 and 3) was selective for p38a in this panel
we were particularly interested in differences of the bind-
ing modes in members of the MAPK family.
To address this question, rational binding modes of 1, 2,
and 3 were modeled in the ATP binding pocket of p38a
xazol-isomers
2
(p38a IC50 = 0.45 lM; JNK3
F
F
F
IC50 = 0.54 lM) and3 (p38a IC50 = 2.2 lM; JNK3
IC50 = 3.5 lM) inhibited both enzymes.4
N
N
N
In this communication we report on the profiling of 1, 2,
and 3 (Fig. 1) in a panel of 78 PKs5,6 including the
p38MAPK isoforms (a,b,c,d), JNK1/2/3 and ERK1/2/
8 (Table 1; the complete profile over 78 PK and assay
O
O
N
N
NH
O
H C
H C
3
3
CH
CH
3
3
1
2
3
Keywords: p38MAP kinase; Inhibitors; Selectivity.
*
Figure 1. Structures of MAPK inhibitors 1, 2, and 3 possessing a
vicinal 4-fluorphenyl/pyridine pharmacophore.
Corresponding author. Tel.: +49 70 71 29 78797; fax: +49 70 71 29
5037; e-mail: Christian.Peifer@uni-tuebingen.de
0960-894X/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.12.073

1432
C. Peifer et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1431–1435
Table 1. Activity profile of 1, 2, and 3 at a concentration of 10 lM
over 10 PK of the MAPK superfamily represented as % residual
activity of PK (and SD, n = 3) compared to control
1
SD
2
SD
3
SD
ERK1
ERK2
JNK1
JNK2
JNK3
p38a
p38b
p38c
p38d
ERK8
84
93
86
89
2
3
92
86
67
13
28
20
95
84
86
84
8
1
87
88
98
59
63
50
98
94
95
88
7
1
0
6
2
4
7
9
3
6
6
1
6
1
106
53
97
89
91
73
4
0
4
6
7
13
0
3
9
10
6
6
(Fig. 2). According to the typical binding mode of com-
pounds bearing a vicinal 4-F-phenyl/pyridine pharmaco-
phore in p38a an H-bond from Met109 (hinge region) is
addressed to the pyridine nitrogen. The 4-F-phenyl moi-
ety is situated in the hydrophobic pocket II with the vic-
inal 4-F-phenyl/pyridine system clasping around
‘gatekeeper’13 residue Thr106 and thereby accounting
for p38(a)-selectivity.
Furthermore an H-bond from Lys53 toward the car-
bonyl-oxygen (1) and to the isoxazole ring-nitrogen (2)
or -oxygen (3) is formed, respectively. In light of the
compound’s differentiated inhibitory profile we com-
pared these key residues by a sequence alignment of
the highly conserved ATP binding pockets of MAPK
(shown in SI) and their homology models (Fig. 3). The
relevant residues Met and Lys are overall conserved
while the gatekeeper differs within these enzymes. The
role of more spacious Met146 in JNK3 compared to
Thr106 in p38a was found to determine selectivity of
six-membered core compound 1 for p38a but not for
the five-membered core compounds 2 and 3.4 Actually
p38a and p38b show a Thr residue as gatekeeper
whereas in line with the hypothesis a Met (p38c/d and
JNK1/2/3) or even bulkier residues (Gln ERK1/2, Phe
ERK8) are situated at this position (Fig. 3). Thus, with
the exception of p38b the loss of activity of 1 for these
MAPK can be sufficiently explained by the major gate-
keeper preventing access of 1 to these binding pockets.
Although sharing the same gatekeeper residue 106 the
only difference in the ATP pocket between the highly
Figure 3. ATP binding pockets of p38a (1A9U),14 JNK3 (1PMN)15
and homology models for p38b/c/d, JNK1/2, and ERK1/2/8. Key
residues are labeled (experimental details are given in supporting
information).
,
Figure 2. Modeled binding modes of 1, 2, and 3 in the ATP binding pocket of p38aMAPK (PDB 1AU9).14 Key residues and ligand–protein
interactions are shown.

C. Peifer et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1431–1435
1433
homologous p38a/b isoforms is to be seen at position
107 (p38a His, p38b Thr).
homologous binding pockets of p38a, JNK3 (and
JNK2) could account for the inhibition of these kinases
by 2 and 3. Interestingly, the compounds did not inhibit
p38b while comparable p38a inhibitors such as
SB203580 and SB202190 bearing the same prototypical
vicinal pyridine/4-F-phenyl pharmacophore also have
been shown to block p38b with similar potency.5,6 How-
ever, determining a X-ray structure of a complex of 2 in
p38a showing the decided protein conformation and li-
gand–protein interactions may answer this discrepancy.
To investigate if the different residues at position 107
could have an impact on the conformation of the neigh-
boring gatekeeper Thr106 we compared the conforma-
tions of these amino acids in p38a crystal structures
(Fig. 4).16,17
Herein, Thr106 is involved in water mediated H-bond
interactions thereby significantly affecting the orienta-
tion of this gatekeeper residue. Typically, the methylene
moiety of Thr106 is situated inside the pocket interact-
ing with the ligand whereas the Thr106-OH is fixed in
the backbone background to His107 (which is unique
to p38a) via two water molecules (Fig. 4, left). In con-
trast, p38a without a ligand is showing a flip of
Thr106 (and Met108) with the Thr-OH subsequently sit-
uated in the pocket and not involved in water mediated
interactions (Fig. 4, right). Hence a considerable impact
of His107 on the conformation of the gatekeeper Thr106
by ligand binding in p38a can be revealed. Concerning
Thr107 in p38b instead of His107/p38a may have a sig-
nificant consequence on the conformation of gatekeeper
Thr106 in p38b and subsequently on ligand binding,
too. However, this hypothesis is supported by the com-
plete loss of activity of 1, 2, and 3 for p38b (and p38c/d,
Table 1).
Furthermore, the strength of the H-bond from the lysine
residue of p38a and JNK2/3 addressing the isoxazole
ring nitrogen in 2 was reported to be higher than
the H-bond to the corresponding oxygen in 3.18 This
is reflected by the relatively better inhibition of JNK2/
3 and p38a of 2 compared to 3 (for IC50 values, see
Table 1).
For the efficient optimization of the lead structure 1
showing promising selectivity for p38a we used a general
synthetic approach toward the quinolin-2(1H)-one sys-
tem by combining a Buchwald amidation reaction and
an aldol condensation (Fig. 5).19,20
The quinolin-2(1H)-one-NH was not interacting to an
amino acid residue in the calculated binding mode of 1
in the ATP binding pocket of p38a (Fig. 2). Thus, in
order to reveal SAR data for this side a variety of
substituents were introduced to the quinolin-2(1H)-
one-nitrogen by SN reaction of deprotonated 1 and
alkylhalogenide.21 Due to the tautomeric situation in
the quinolinone nucleus22 during this reaction both
N-and O-alkylated products were obtained (Table 2).23
In terms of selectivity for p38a the situation for 2 and 3
seems to be more complex and further PK were inhib-
ited (see above). As previously described4 the highly
Furthermore, the 4-F-phenyl system of 1 was replaced
by an ortho-toluyl moiety (18) to investigate the restric-
tion of the conformation of this aromatic residue on the
biological activity. In the SN reaction of 18 to generate
19 the corresponding O-alkylated compound 20 (2-
[(2,2-dimethyl-1,3-dioxolan-4-yl)methoxy]-3-(2-methyl-
phenyl)-4-pyridin-4-ylquinoline) was generated. Com-
pound 23 (4-pyridin-4-ylquinolin-2(1H)-one) was
synthesized via an intramolecular SEreaction of 22 (see
SI) using PPA24 and compound 25 (3,4-diphenyl-quino-
lin-2(1H)-one) obtained via aldol condensation of 24
(see SI).4 The structures of compounds 9 (CCDC Nr
670663), 13 (CCDC Nr 670664), 14 (CCDC Nr
670665), and 19 (CCDC Nr 670666) have been proven
by X-ray analysis (further synthetic and analytical de-
tails for all compounds are available on SI).
Figure 4. Exemplified different H-bond networks in p38a crystal
structures involving water molecules (red dots) affecting the confor-
mation of gatekeeper Thr106 and residue His107 in the ATP binding
pocket (left, PDB 1W8416 with ligand L12 in cyan; right, PDB 1P38,17
no ligand bound).
O
O
Cs CO
Br
2
3
Pd /dba)
KOH
2
3
H N
2
NH
1
O
Xantphos
EtOH, RF
Toluol 100 ºC
5
+
6
O 7
F
F
N
N
Figure 5. Preparation of 1 by a tandem Buchwald-aldol-reaction approach: compounds 5 and 6 are reacted by a Buchwald amidation to generate
intermediate 7 which subsequently yields the quinolinone ring by an aldol condensation.

1434
C. Peifer et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1431–1435
Table 2. Substitution patterns and indication of compounds
R
F
F
CH
O
N
3
N
N
O
R
O
N
R
R
N
N
–H
–Methyl
1
8
9
—
—
10
12
14
16
—
18
—
—
—
—
19
21
–Ehyl
–Propinyl
11
13
15
17
–Phenylethyl
–[(2,2-Dimethyl-1,3-dioxolan-4-yl)methyl]
Butyl-1,2-diol
N
O
N
H
N
H
O
23
25
The compounds have been evaluated in vitro for their
potency against p38a25 (Table 3).
pounds 10, 12, 14, and 16 were inactive presumably
mainly due to clashes with Lys53 in the ATP binding
pocket (Fig. 2). Compound 18 was determined to have
an IC50 value of approximately 10 lM and the intro-
duction of the functionalized side chain (21) did not
significantly change the activity, as already seen for
the related derivatives 1/17. Furthermore, the vanished
activity of 23 on the one hand and 25 on the other
hand indicated the vicinal pyridine/4-F-phenyl system
in 1 to be favorable for p38a inhibition.
Within this series 1 was found again to inhibit the p38a
enzyme most potent with an IC50 value of 1.8 lM.4
The quinolinone-N-alkylated compounds 8, 9, 11, and
13 showed slightly less inhibition activity yet still in
the single digit lM range. In contrast to a promising
modeled binding mode of 17 in the ATP pocket of
p38a (not shown), the introduction of the two hydroxyl
moieties in the quinolinone-N-alkyl side chain (17)
slightly decreased the activity and hence this optimiza-
tion strategy for 1 was not successful. However, accu-
rate and actual virtual design of protein kinase
inhibitors remains challenging.26 As expected from
our modeling data, the quinolinone O-alkylated com-
Taken together, 1 is a promising lead compound for the
development of potent and selective p38a inhibitors.
Acknowledgments
The specificity testing of compounds 1, 2, and 3 over 78
PKs was performed by Jenny Bain in the group of Sir
Prof. P. Cohen, Screening Service at the Division of Sig-
nal Transduction Therapy, University of Dundee, Scot-
land and is gratefully acknowledged. The X-ray analysis
of compounds was performed by Dr. D. Schollmeyer,
Institute of Organic Chemistry, University of Mainz,
Germany and is gratefully acknowledged. The authors
thank the Merckle GmbH Blaubeuren and the DAAD
for financial support.
Table 3. Inhibition of p38a activity by test compounds reported as
IC50 values (lM) or
concentration of 10 lM
% residual activity of the enzyme at a
#
IC50 p38a (lM)
#
IC50 p38a (lM)
1
1.8
5.2
18
19
20
21
23
25
9.7
nd
8
9
2.2
66% at 10 lM
3.1
nd
10
11
12
13
14
15
16
17
80% at 10 lM
78% at 10 lM
64% at 10 lM
79% at 10 lM
8.2
—
55% at 10 lM
—
Supplementary data
57% at 10 lM
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2007.12.073.
— Indicates no significant inhibition at 10 lM compound concentra-
tion. nd, not determined.

C. Peifer et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1431–1435
1435
17. Wang, Z.; Harkins, P. C.; Ulevitch, R. J.; Han, J.; Cobb,
M. H.; Goldsmith, E. J. Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 2327.
18. Nobeli, I.; Price, S. L.; Lommerse, J. P. M.; Taylor, R.
J. Comput. Chem. 1997, 18, 2060.
19. Manley, P. J.; Bilodeau, M. T. Org. Lett. 2004, 6, 2433.
20. General method for the preparation of quinolin-2(1H)-
ones, exemplified by the synthesis of 1. The reaction was
References and notes
1. Richards, M. L. Curr. Top. Med. Chem. 2006, 6, 75.
2. Boldt, S.; Kolch, W. Curr. Pharm. Des. 2004, 10, 1885.
3. Scapin, G. Drug Discovery Today 2002, 7, 601.
4. Peifer, C.; Kinkel, K.; Abadleh, M.; Schollmeyer, D.;
Laufer, S. J. Med. Chem. 2007, 50, 1213.
5. (a) Bain, J.; McLauchlan, H.; Elliott, M.; Cohen, P.
Biochem. J. 2003, 371, 199; (b) Bain, J.; Plater, L.; Elliott,
M.; Shpiro, N.; Hastia, J.; McLauchlan, H.; Klevernic, I.;
Arthur, S.; Alessi, D.; Cohen, P. Biochem. J. 2007, 408, 297.
6. Davies, S. P.; Reddy, H.; Caivano, M.; Cohen, P.
Biochem. J. 2000, 351, 95.
7. Skaalhegg, B. S.; Funderud, A.; Henanger, H.; Hatledal,
H.; Tilahun, T.; Larsen, A. C.; Kvissel, A. K.; Eikvar, S.;
Oerstavik, S. Curr. Drug Targets 2005, 6, 655.
8. Rozengurt, E.; Rey, O.; Waldron, R. T. J. Biol. Chem.
2005, 280, 13205.
9. Kamens, J. S.; Ratnofsky, S. E.; Hirst, G. C. Curr. Opin.
Investig. Drugs 2001, 2, 1213.
10. Knippschild, U.; Wolff, S.; Giamas, G.; Brockschmidt, C.;
Wittau, M.; Wuerl, P. U.; Eismann, T.; Stoeter, M.
Onkologie 2005, 28, 508.
11. Godl, K.; Wissing, J.; Kurtenbach, A.; Habenberger, P.;
Blencke, S.; Gutbrod, H.; Salassidis, K.; Stein-Gerlach,
M.; Missio, A.; Cotten, M.; Daub, H. Proc. Natl. Acad.
Sci. U.S.A. 2003, 100, 15434.
12. Keen, N.; Taylor, S. Nat. Rev. Cancer 2004, 4, 927.
13. Prien, O. ChemMedChem 2006, 1, 1195.
14. Wang, Z.; Canagarajah, B. J.; Boehm, J. C.; Kassisa, S.;
Cobb, M. H.; Young, P. R.; Abdel-Meguid, S.; Adams, J.
L.; Goldsmith, E. J. Structure 1998, 6, 1117.
15. Scapin, G.; Patel, S. B.; Lisnock, J.; Becker, J. W.;
LoGrasso, P. V. Chem. Biol. 2003, 10, 705.
16. Gill, A. L.; Frederickson, M.; Cleasby, A.; Woodhead, S.
J.; Carr, M. G.; Woodhead, A. J.; Walker, M. T.;
Congreve, M. S.; Devine, L. A.; Tisi, D.; O’Reilly, M.;
Seavers, L. C. A.; Davis, D. J.; Curry, J.; Anthony, R.;
Padova, A.; Murray, C. W.; Carr, R. A. E.; Jhoti, H.
J. Med. Chem. 2005, 48, 414.
carried out under argon. Compound
3.43 mmol), (4-fluorophenyl) acetamide
5
(900 mg,
(630 mg,
4.11 mmol), Cs2CO3 (1.57 mg, 4.82 mmol), Pd2(dba)3
(32 mg, 0.03 mmol), and Xantphos (58 mg, 0.103 mmol)
were dissolved in 5 ml of dry toluol and refluxed for 18 h.
After cooling 20 ml of water was added and extracted by
2· 50 ml CH2Cl2. The organic phase was evaporated and
the crude mixture of products 7 and 1 was redissolved in
ethanol. After addition of 20 mg KOH the mixture was
refluxed for 30 min, then water added and extracted with
2· 50 ml CH2Cl2. The organic phases were dried over
Na2SO4, evaporated, and the product purified by column
chromatography to yield 451 mg (42%) of 1 as a white
solid.
21. General method for the SN of compound 1. In a dry 50 ml
three-necked round-bottomed flask 1 mmol of 1 and
2 mmol of a 60% NaH suspension were stirred for
30 min at rt in 5 ml dry DMSO. Through a septum
2 mmol of the corresponding halogenalkyl was added via
syringe and the mixture stirred for 60 min. Water (20 ml)
was added, the reaction mixture extracted by 3· 30 ml
ethyl acetate, the organic phase was separated, dried over
Na2SO4, and evaporated. The N- (and O-) alkylated
product(s) were purified by flash chromatography.
22. Gerega, A.; Lapinski, L.; Nowak, M. J.; Furmanchuk, A.;
Leszczynski, J. J. Phys. Chem. A 2007, 111, 4934.
23. Park, K. K.; Lee, J. J. Tetrahedron 2004, 60, 2993.
24. Larsen, R. D. In Science of Synthesis, 2004, Chapter 15.4.
25. Laufer, S.; Thuma, S.; Peifer, C.; Greim, C.; Herweh, Y.;
Albrecht, A.; Dehner, F. Anal. Biochem. 2005, 135.
26. Muegge, I.; Enyedy, I. J. Curr. Med. Chem. 2004, 11,
693.