Synthesis and SAR of aminopyrimidines as novel c-Jun N-terminal kinase (JNK) inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.03.078

The development of a series of novel aminopyrimidines as inhibitors of c-Jun N-terminal kinases is described. The synthesis, in vitro inhibitory values for JNK1, JNK2 and CDK2, and the in vitro inhibitory value for a c-Jun cellular assay are discussed.

Full text of DOI:10.1016/j.bmcl.2007.03.078

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3463–3467
Synthesis and SAR of aminopyrimidines as novel c-Jun
N-terminal kinase (JNK) inhibitors
Mahbub Alam, Rebekah E. Beevers, Tom Ceska, Richard J. Davenport,
Karen M. Dickson, Mara Fortunato, Lewis Gowers, Alan F. Haughan, Lynwen A. James,
Mark W. Jones, Natasha Kinsella,* Christopher Lowe, Johannes W. G. Meissner,
Anne-Lise Nicolas, Benjamin G. Perry, David J. Phillips, William R. Pitt, Adam Platt,
Andrew J. Ratcliffe, Andrew Sharpe and Laura J. Tait
UCB, Granta Park, Great Abington, Cambridge CB21 6GS, UK
Received 4 January 2007; revised 23 March 2007; accepted 24 March 2007
Available online 30 March 2007
Abstract—The development of a series of novel aminopyrimidines as inhibitors of c-Jun N-terminal kinases is described. The
synthesis, in vitro inhibitory values for JNK1, JNK2 and CDK2, and the in vitro inhibitory value for a c-Jun cellular assay are
discussed.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Inflammation is a key area of research for many
pharmaceutical companies. Patients suffering from
inflammatory disorders such as rheumatoid arthritis
(RA) require therapeutic agents that not only
demonstrate anti-inflammatory properties but also
protect against cartilage degradation.
In this communication, we report our efforts towards
a series of aminopyrimidines as novel ATP competitive
JNK1 and JNK2 inhibitors for the treatment of
inflammatory disorders.
SP600125 (1) is a reversible ATP competitive inhibitor
of all three JNK isoforms, identified from a high
throughput screen (HTS) of Celgene’s chemical library.6
Another JNK inhibitor (2) has subsequently been
published by Serono7 which has shown efficacy via oral
dosing in a model of RA. Further recent publications
from Astra-Zeneca8 and Abbott9 emphasise the growing
interest in JNK as a target (Fig. 1).
The c-Jun N-terminal protein kinases (JNKs) are a
family of serine/threonine protein kinases and members
of the mitogen-activated protein kinase (MAPK)
family. JNKs activate proteins such as c-Jun by
phosphorylation; these proteins then dimerise to form
transcription factors, including AP-1, which mediate
pro-inflammatory effects via induction of inflammatory
factors including cytokines (TNF-a and IL-2) and
matrix metalloproteinases (MMPs).1 Mammals have
three JNK genes Jnk1, Jnk2 and Jnk3.2 Gene knockout
studies of Jnk1 and 2 suggest a role for JNK in T cell
proliferation.3 Further studies also indicate that JNK
inhibition may afford protection against cartilage and
bone erosion.4 Current evidence suggests that selective
JNK1 and 2 inhibition would be beneficial for treating
patients with RA.3a,4,5
Screening of the UCB compound collection identified
several hits with varying potencies against JNK1, 2
and 3. One of the most promising hits was the
aminopyrimidine 5, which showed good JNK potency
and some selectivity against other protein kinases such
as p38 (Fig. 2).
In order to use computer-aided drug design an X-ray
crystal structure of an early lead, compound 6, bound
into the ATP binding site of JNK3 was obtained
(Fig. 3). JNK3 protein was used as a surrogate in the
absence of in-house JNK1 and JNK2 protein crystals
and all modelling work discussed here is from the
JNK3 protein. There is about 92% sequence identity
Keywords: Aminopyrimidines; JNK1; JNK2.
*
Corresponding author. Tel.: +44 1223896638; e-mail: natasha.kinsella@
ucb-group.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.03.078

3464
M. Alam et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3463–3467
H
N
N
as our biological results were unaffected by JNK3 activ-
H
N
O
N
ity. The X-ray crystal structure of 6 (Fig. 3) shows the
aminopyrimidine nitrogens interacting with the hinge
region of the active site via Met 149, the indole sitting
in the pocket normally occupied by the ribose of the
natural ligand ATP and the 50-chloro group sterically
interacting with the sulfur of Met 146. Met 146 is also
known as the ‘gatekeeper’, a term used to describe the
residue of a kinase that blocks the classical hydrophobic
pocket. All protein kinases have a gatekeeper residue. A
feature so far found only in JNKs10 is the presence of a
moveable gatekeeper, which allows access to the classi-
cal hydrophobic pocket. All JNKs have a methionine
gatekeeper residue, p38 has a threonine residue and
CDK2 has a phenylalanine residue, which is not capable
of moving to expose the hydrophobic pocket. Modelling
suggested that this difference might be used to improve
selectivity. The aminopyrimidines described were
synthesized as shown in Scheme 1. Appropriately
substituted 2,4-dichloro pyrimidines (31) were reacted
via a Suzuki coupling with 1-phenylsulfonyl indole-3-
boronic acid, followed by a basic deprotected to give
intermediates (32). Intermediates (32) could be reacted
with substituted piperidines to give final compounds
via deprotection and further reaction with appropriate
capping groups. In order to synthesize compound 10,
indole was deprotonated using sodium hydride and
reacted with 2,4,5-trichloropyrimidine (33). The product
was then reacted as before with substituted piperidines,
followed by deprotection and reaction with ethylisocya-
nate. It can be noted that the chlorine of compound 10
can be removed via hydrogenation under basic condi-
tions to give compound 12. 3-Bromoimidazopyridine
(34) was reacted with trichloropyrimidine (33) via a
Negishi coupling. The intermediate dichloro compounds
were reacted with substituted piperidines as before to
give compound 9. Hydrogenation was used again to
synthesize compound 11. Finally the pyrazolopyridines
were synthesized via a coupling of dichloropyridine
(35) with a substituted acetylene followed by a
cyclisation reaction to give intermediates 36. Compound
5 was screened against a panel of protein kinases
including cyclin-dependent protein kinase 2 (CDK2)
which had been highlighted as a potential selectivity
issue. A further screen of other 50-cyano aminopyrimi-
dines showed that CDK2 activity was indeed an issue
that needed to be addressed.
N
N
S
N
CN
(2)
(1)
H
N
N
N
O
NH2
O
S
Cl
O
O
O
N
N
O
H
(3)
O
(4)
N
H
O
Figure 1. JNK inhibitors in the public domain. (1) SP600125, IC50
values: JNK1 110, JNK2 110, JNK3 150 nM. (2) AS601245, IC50
values: JNK1 150, JNK2 220, JNK3 70 nM.7 (3) Astra-Zeneca
compound, IC50 values: JNK1 384, JNK3 7 nM.9b (4) Abbott
compound, IC50 values: JNK1 36, JNK2 70 nM.9b
H
H
N
N
N
N
H
H
N
N
N
N
N
N
NC
Cl
O
O
(5)
N
(6)
N
H
H
F
Figure 2. Hit from HTS, compound 5, and early lead, compound 6.
IC50 values 5: JNK1 92, JNK2 67, JNK3 412, p38 4863 nM. IC50
values 6: JNK1 13, JNK2 25, JNK3 57, p38 5423 nM.
In order to attain better CDK2 selectivity, while
retaining JNK potency, we investigated the SAR around
four distinct areas of the original hit (5), the 50-substitu-
ent, the bicycle, the heterocycle and the piperidine
substituent, referred to as the cap. The first two areas
of investigation were alternative 50-substituents and
bicycle replacements (Table 1, 5–13). From Table 1 it
can be seen that there is a subtle interplay between
50-substituents and the bicycles affecting aminopyrimi-
dine SAR. When 50 is chloro the 3-indole (6) has greater
potency against the JNKs than the imidazopyridine (9),
followed by the 1-indole (10), which shows only
moderate activity. Modelling suggests that the shorter
carbon–nitrogen of the 1-indoles results in a steric clash
with the 50-chlorine which in turn leads to a loss in
Figure 3. X-ray structure of compound 6 (PDB code 2P33).
between JNK1 and JNK3 and about 85% between
JNK2 and JNK3. JNK3 enzyme activity was routinely
checked to aid with our modelling and is shown in
Table 1. No further discussion will be given in this paper

M. Alam et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3463–3467
Table 1. Enzymatic and cellular activity of aminopyrimidine analogues
3465
H
H
H
O
N
N
cap
N
N
N
N
N
N
( )n
N
N
N
N
NH
Cl
Cl
X
cap
Bicycle
14-15
16-19
5-12, 20-28
N
N
H
H
Compound 50 X Bicycle
Heterocycle
Cap
JNK111
JNK211
JNK311
CDK212
IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM) IC50 (nM)
c-Jun13
5
6
CN 6-F-3-Indole
4-Piperidine
4-Piperidine
4-Piperidine
4-Piperidine
CONHEt
CONHEt
CONHEt
CONHEt
CONHEt
CONHEt
CONHEt
CONHEt
CONHEt
CONHEt
CONHEt
92
13
67
25
412
57
412
3700
704
Cl
Me
H
3-Indole
3-Indole
3-Indole
1517
IA
7
320
74
250
245
55
410
na
10,000
8091
7723
8
10,000
605
9
Cl
Cl
H
3-Imidazopyridine 4-Piperidine
1-Indole 4-Piperidine
3-Imidazopyridine 4-Piperidine
41
na
na
10
11
12
13
14
15
16
17
18
19
20
21
22
23
457
59
709
281
512
194
177
1551
15
4443
4219
10,000
8663
2483
na
>10,000
19,331
29,891
4000
708
na
H
1-Indole
Pyrazolopyridine
3-Indole
3-Indole
3-Indole
3-Indole
3-Indole
3-Indole
3-Indole
3-Indole
3-Indole
3-Indole
4-Piperidine
4-Piperidine
3-Pyrrolidine
Azetidine
71
69
H
na
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
360
1340
29
582
5000
32
6268
4272
3-(S)-Piperidine CONHEt
555
6995
3-(S)-Piperidine CH2CONHMe
3-(R)-Piperidine CONHEt
3-(R)-Piperidine CH2CONHMe
139
60
267
88
na
>10,000
1264
612
14,667
3883
2807
107
31
15
13
31
22
4-Piperidine
4-Piperidine
4-Piperidine
4-Piperidine
CH2CONHMe
COOEt
CONMe2
14
82
123
na
1769
37
15
49
37
na
na
na
4358
5281
741
2770
CO-(4Me
piperazine)
18
26
24
25
26
27
Cl
Cl
Cl
Cl
3-Indole
3-Indole
3-Indole
3-Indole
4-Piperidine
4-Piperidine
4-Piperidine
4-Piperidine
COCH2NHCOMe
COCH2NHMe
COCH2NMe2
CO-(4Me
28
67
57
15
46
85
45
14
na
551
1938
2028
2497
813
179
120
48
2672
1126
1895
piperidine)
CONH-(4-Me
piperidine)
28
Cl
3-Indole
4-Piperidine
47
62
na
6652
807
na, not available.
planarity and fit with the binding site. When 50 is hydro-
gen (8) (11) (12) there is an overall drop in JNK activity.
However in this case the 50-substituent has less effect,
with all bicycles having comparable potencies. This
interplay also has an effect on CDK2 activity. When
chloro, JNK1 and JNK2 activities are comparable, but
more potent than CDK2. When hydrogen, JNK1
activity is greater than JNK2, which in turn is more
potent than CDK2. 50-Chloro analogues have far
greater JNK potency, however selectivity will need to
be addressed as exemplified by high activity against
CDK2. Alternatively, 50-hydrogen analogues offer a
better JNK/CDK selectivity profile and so have
potential if JNK potency can be improved. Our next
area of interest was to look for heterocycles as replace-
ments for the 4-piperidine of the original hit (5)
(14–19). Changing the ring size from six (6) to five (14)
to four (15) gave a reduction in JNK activity and
CDK2 activity, suggesting a less optimal fit of the
smaller rings in the active site of both JNKs and
CDK2. The 4-piperidine ring seems to be occupying
space that is key to potency. To test this several 3-piper-
idine analogues (16) to (19) were designed. Overall, these
compounds were either equipotent or less active than the
4-piperidine compound (6), with the (S)-3-piperidine
compounds 16 and 17 having an undesirable effect on
CDK2 potency. Comparison of enantiomers suggests a
stereochemical preference that is dependent on the nat-
ure of the cap. Compounds 20–28 exemplify the final
area of diversification, which involved varying groups
attached to the nitrogen of the 4-piperidine. Many
groups were well tolerated and it can be seen from Table
1 that different groups can improve selectivity against
CDK2, whilst retaining JNK potencies. This area has
huge scope and is under investigation with the aim of
improving physical properties, such as solubility. As dis-
cussed earlier, modelling suggested that the gatekeeper
area could be used to induce selectivity for JNKs over
CDK2 by having R groups that move the gatekeeper
and access the classical hydrophobic pocket (Table 2,
Fig. 4). Compound 13 showed good JNK1 activity,
moderate JNK2 potency, with good selectivity over
CDK2, providing confidence that this would be a good
template to work from. Two further compounds were
made (29–30) and interestingly it seems that R can give
rise to potent and selective compounds.
When R is phenyl (30) JNK1 and JNK2 potency is 22
and 5 nM, respectively, with excellent selectivity over
CDK2. While this work needs further investigation the
hypothesis proposed by modelling has been supported
by experimental evidence.

3466
M. Alam et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3463–3467
H
N
N
N
Cl
N
Cl
a,b
c,d,e or f
N
N
O
N
N
X
X
X
N
Cl
N
N
(31)
H
H
(32)
(6)-(8), (14)-(28)
H
N
N
N
N
Cl
N
N
O
X
g, c, d, e
N
NH
Cl
Cl
(10) X=Cl
(12) X=H
h
(33)
H
N
N
N
N
O
Br
Cl
i, j
N
NH
N
N
N
(9) X=Cl
(11) X=H
h
(34)
H
N
N
Cl
N
Cl
R
N
k, l
j
N
N
N
N
O
Cl
NH
R
(35)
N N
N N
Figure 4. Docking of Compound 30 in JNK3 protein (GOLD, CCDC,
UK) into 1 pmn.10
(13), (29)-(30)
(36)
Scheme 1. Synthesis of aminopyrimidines. Reagents and conditions:
(a) 1-phenylsulfonyl indole-3-boronic acid, Pd(PPh3)4, Na2CO3,
MeCN, H2O, reflux, 90 min, 50–90%; (b) KOH, MeOH, reflux
30 min, 50–60%; (c) 1-bocaminopiperidine, DMF, 120 ꢁC, 6 h, 50–
70%; (d) 2 N HCl in ether, MeOH, rt, 40–60%; (e) EtNCO, Et3N,
Na2CO3, DCM, rt, 10–40%; (f) chloromethylacetamide, Na2CO3,
DMF, 60 ꢁC, 2 h, 10–40%; (g) indole, NaH, DMF, rt, 60–70%; (h)
Pd/C, H2, NaOH, MeOH, 80–85%; (i) n-BuLi, ZnBr, THF, then
2,4,5-trichloropyrimidine, Pd(PPh3)4, 32%; (j) 1-ethylurea-4-aminopi-
peridine, Et3N, DMF, 50%; (k) R-acetylene, PdCl2(PPh3)2, CuI,
Et3N, THF, rt, 18 h, 70–80%; (l) 1-aminopyridinium iodide, DBU,
MeCN, 50 ꢁC, 18 h, 20–50%.
submicromolar cellular potency (22, 27, 28). Further
work is underway to improve our understanding of
cellular potency; however at present other bicycles have
not demonstrated significant cellular activity.
A large number of compounds have been synthesized
and assayed against JNK1, JNK2 and CDK2 free
enzymes and tested in a c-Jun cellular assay. The results
presented highlight our success in synthesizing
compounds that led to improvements in JNK potency
and good selectivity over CDK2.
JNK inhibitors need to be cell permeable to reach their
intracellular targets; therefore a phospho c-Jun cell
based assay13 was utilized to test cellular potencies
(Table 1). Compound 4 gave moderate cellular activity
and seemed to be a good starting point for further
development, although only a few analogues showed
X-ray crystal structures and docking have been
employed to understand the complex SAR of this series
of molecules regarding their activity against JNK.
Predictions for improved selectivity, such as accessing
the classical hydrophobic pocket, have been tried
Table 2. Enzymatic and cellular activity of aminopyrimidine analogues
H
N
N
N
N
O
NH
R
N N
(13), (29)-(30)
Compound
50 X
Bicycle
R
JNK111
IC50 (nM)
JNK211
IC50 (nM)
JNK311
IC50 (nM)
CDK212
IC50 (nM)
c-Jun13
IC50 (nM)
6
13
29
30
Cl
H
H
H
3-Indole
H
13
69
25
194
698
5
57
na
na
5
1517
8663
704
4000
Pyrazolopyridine
Pyrazolopyridine
Pyrazolopyridine
H
iPr
Ph
520
22
>10,000
>10,000
21,160
3845
na, not available.

M. Alam et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3463–3467
3467
2006, 16, 2590; (b) Liu, G.; Zhao, H.; Liu, B.; Xin, Z.; Liu,
M.; Kosogof, C.; Szczepankiewicz, B. G.; Wang, S.;
Clampit, J.; Gum, R. J.; Haasch, D. L.; Trevillyan, J. M.;
Sham, H. L. Bioorg. Med. Chem. Lett. 2006, 16, 5723; (c)
Zhao, H.; Serby, M. D.; Xin, Z.; Szczepankiewicz, B. G.;
Liu, M.; Kosogof, C.; Liu, B.; Nelson, L. T. J.; Johnson,
E. F.; Wang, S.; Pederson, T.; Gum, R. J.; Clampit, J. E.;
Haasch, D. L.; Abad-Zapatero, C.; Fry, E. H.; Rondi-
none, C. M.; Trevillyan, J. M.; Sham, H. L.; Liu, G.
J. Med. Chem. 2006, 49, 4455; (d) Szczepankiewicz, B. G.;
Kosogof, C.; Nelson, L. T. J.; Liu, G.; Liu, B.; Zhao, H.;
Serby, M. D.; Xin, Z.; Liu, M.; Gum, R. J.; Haasch, D. L.;
Wang, S.; Clampit, J. E.; Johnson, E. F.; Lubben, T. H.;
Stashko, M. A.; Olejniczak, E. T.; Sun, C.; Dorwin, S. A.;
Haskins, K.; Abad-Zapatero, C.; Fry, E. H.; Hutchins, C.
W.; Sham, H. L.; Rondinone, C. M.; Trevillyan, J. M.
J. Med. Chem. 2006, 49, 3563.
experimentally and the evidence so far supports these
predictions. Further work on this series of compounds
and their selectivity profiles will be communicated in
due course. Work is underway to improve our
understanding of cellular potency; however at present
other bicycles have not demonstrated significant cellular
activity.
References and notes
1. (a) Kyriakis, J.; Avruch, J. Physiol. Rev. 2001, 81, 807; (b)
Pearson, G.; Robinson, R.; Gibson, T. B.; Xu, B. E.;
Karandikar, M.; Berman, K.; Cobb, M. H. Endocr. Rev.
2001, 22, 153.
2. Gupta, S.; Barrett, T.; Whitmarsh, A. J.; Cavanagh, J.; Sluss,
H. K.; Derijard, B.; Davis, R. J. EMBO J. 1996, 15, 2760.
3. (a) Bogoyevitch, M. A.; Boehm, I.; Oakley, A.; Ket-
terman, A. J.; Barr, R. K. Biochim. Biophys. Acta 2004,
1697, 89; (b) Sabapathy, K.; Hu, Y.; Kallunki, T.;
Schreiber, M.; David, J. P.; Jochum, W.; Wagner, E. F.;
Karin, M. Curr. Biol. 1999, 9, 116.
4. Bogoyevitch, M. A. Trends Mol. Med. 2005, 11, 231.
5. (a) Dong, C.; Yang, D. D.; Wysk, M.; Whitmarsh, A. J.;
Davis, R. J.; Flavell, R. A. Science 1998, 282, 2092; (b)
Sabapathy, K.; Hu, Y.; Kallunki, T.; Schreiber, M.;
David, J. P.; Jochum, W.; Wagner, E. F.; Karin, M. Curr.
Biol. 1999, 9, 116.
10. Scapin, G.; Patel, S. B.; Lisnock, J.; Becker, J. W.;
LoGrasso, P. V. Chem. Biol. 2003, 10, 705.
11. The JNK-1/2/3 in vitro enzyme assay determines the
effect of test compounds on phosphorylation of the
GST-c-Jun (1–89) substrate at the Ser73 site. GST-c-
Jun is pre-bound to glutathione plates. Plates are
washed prior to addition of MKK4 activated JNK-1/
2/3, test compound and ATP. Plates are washed and
the level of phosphorylation detected with an anti-
phospho Ser73 antibody.
12. The CDK2 in vitro enzyme assay determines the effect of
test compounds on the phosphorylation of a fluorescein
labelled Histone H1-derived peptide detected by IMAP
technology. FAM-peptide, ATP, CDK2/Cyclin A and test
compounds are added to a 384-well black plate. The
reaction mixture is incubated for 60 min at RT and
stopped by addition of 60 ll progressive binding buffer ‘A’
containing 1/400 IMAP beads.
6. Han, Z.; Chang, L.; Yamanishi, Y.; Karin, M.; Firestein,
G. S. Arthritis Rheum. 2002, 46, 818.
7. Gaillard, P.; Jeanclaude-Etter, I.; Ardissone, V.; Arkin-
stall, S.; Cambert, Y.; Camps, M.; Chabert, C.; Church,
D.; Cirillo, R.; Gretener, D.; Halazy, S.; Nichols, A.;
Szyndralewiez, C.; Vitte, P.; Gotteland, J. J. Med. Chem.
2005, 48, 4596.
8. (a) Stocks, M. J.; Barber, S.; Ford, R.; Leroux, F.; St-
Gallay, S.; Teague, S.; Xue, Y. Bioorg. Med. Chem. Lett.
2005, 15, 3459; (b) Teague, S. J.; Barber, S.; King, S.;
Stein, L. Tetrahedron Lett. 2005, 46, 4613.
13. A549 cells are plated out at 1.25 · 104/well in RPMI
media containing 10% FBS, penicillin/streptomycin and
glutamine and incubated overnight at 37 ꢁC/5% CO2.
The cells are stimulated with 1 ng/ml recombinant
human IL-1b and incubated for a further 30 min at
37 ꢁC/5% CO2. c-Jun translocation is then measured
9. (a) Liu, G.; Xin, Z.; Liu, M.; Wang, S.; Clampit, J. E.;
Gum, R. J.; Haasch, D. L.; Trevillyan, J. M.; Sham, H. L.;
Abad-Zapatero, C.; Fry, E. H. Bioorg. Med. Chem. Lett.
using
a c-Jun Activation Screening Hitkit product
(Cellomics).