(1H-Imidazo[4,5-c]pyridin-2-yl)-1,2,5-oxadiazol-3-ylamine derivatives: Further optimisation as highly potent and selective MSK-1-inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.05.020

The novel imidazo[4,5-c]pyridine 1,2,5-oxadiazol-3-yl template affords an excellent start point for identification of inhibitors of a number of protein kinases. Here we report on its optimisation for mitogen and stress-activated protein kinase-1 (MSK-1) inhibitory activity, and selectivity over other kinases.

Full text of DOI:10.1016/j.bmcl.2005.05.020

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3407–3411
(1H-Imidazo[4,5-c]pyridin-2-yl)-1,2,5-oxadiazol-3-ylamine
derivatives: Further optimisation as highly potent
and selective MSK-1-inhibitors
Mark J. Bamford,* Nicholas Bailey, Susannah Davies, David K. Dean, Leann Francis,
Terence A. Panchal, Christopher A. Parr, Sanjeet Sehmi, Jon G. Steadman,
Andrew K. Takle, James T. Townsend and David M. Wilson
GlaxoSmithKline Pharmaceuticals, Neurology & GI Centre of Excellence for Drug Discovery, New Frontiers Science Park,
Third Avenue, Harlow, Essex CM19 5AW, UK
Received 12 January 2005; accepted 9 May 2005
Abstract—The novel imidazo[4,5-c]pyridine 1,2,5-oxadiazol-3-yl template affords an excellent start point for identification of inhib-
itors of a number of protein kinases. Here we report on its optimisation for mitogen and stress-activated protein kinase-1 (MSK-1)
inhibitory activity, and selectivity over other kinases.
Ó 2005 Published by Elsevier Ltd.
Mitogen and stress-activated protein kinases 1 (MSK-1)
and 2 are localised to the nucleus and are activated by
both mitogen- (ERK) and stress-activated (p38) protein
kinases. They phosphorylate a number of transcription
factors including CREB, ATF1, HMG-14 and histone
H3 in various cell types, and are involved in the tran-
scriptional activation of the NF-jB p65 subunit.1 Inhib-
itors may be of value as therapeutic agents since they are
implicated in the regulation of inflammatory mediators
and in excitotoxicity-induced neuronal death.2,3
This model proposes H-bonding interactions of the ami-
no-oxadiazole with the hinge region protein backbone,
and of N5 of the imidazopyridine with Lys81. These
are both common protein kinase features and, therefore,
2 represents an excellent start point for identification of
inhibitors of this class of target. Indeed, Table 1 shows
that 2 inhibits a number of representative kinases,
including CDK2, GSK3b, MAPKAP-K1a, p70S6K
and ROCK1. We now report on our studies to design
analogues of 2 which are selective for MSK-1.
We recently conducted a screening programme and
identified the 2-[3-amino-1,2,5-oxadiazol-5-yl] benzimid-
azole compound 1 (MSK-1 IC50 140 nM), which was
modified to afford the highly potent (1H-imidazo[4,5-
c]pyridin-2-yl)-1,2,5-oxadiazol-3-ylamine derivative
(MSK-1 IC50 3 nM).4
2
Using the docking program GOLD, we have suggested a
binding mode for 2 that explains its improved potency
relative to 1 (Fig. 1) and rationalises our initial SAR.4
We used GSK-3 and MAPKAP-K1 (also known as
RSK-1) to monitor selectivity as we explored the tem-
plate further. Figure 1 suggests the vector of the C4 po-
sition is in the direction of a lipophilic pocket beyond
the ‘‘gatekeeper’’ residue (Leu for MSK-1, GSK-3 and
RSK-1). Access to this pocket affords enhancements in
the activity and sometimes selectivity of inhibitors of
other kinases and was our first area of investigation.
Keywords: MSK-1; Mitogen and stress-activated protein kinase; Imi-
dazo[4,5-c]pyridine.
*
Corresponding author. Tel.: +44 1279 875078; fax: +44 1279
622260; e-mail: Mark.J.Bamford@gsk.com
0960-894X/$ - see front matter Ó 2005 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2005.05.020

3408
M. J. Bamford et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3407–3411
ing hypothesis (Fig. 1) which predicts the butylamine
basic nitrogen to be in the region of Glu138. The corre-
sponding residues in GSK-3 and RSK-1 are Thr and
Asp, respectively.
Thus, further careful manipulation of the nature of the
N1 substituent and its interaction with the lipophilic
pocket in this region of the protein, together with careful
positioning of a basic centre may be expected to offer
both MSK-1 potency and enhanced selectivity. We next
set out to further explore this possibility.
There was no MSK1 potency advantage from straight
alkyl chain analogues 15 or 16 although the shorter
chain 15 did show some improved selectivity over
RSK-1. However, the more bulky 3-piperidinyl com-
pound 17 combined the improved MSK-1 potency and
modest RSK-1 selectivity of cyclohexyl derivative 11
with the modest GSK-3 selectivity of 4-piperidinyl ana-
logue 13. The selectivity over both GSK-3 and RSK-1
was further enhanced in the branched chain analogue
18. Further evidence that an appropriately positioned
basic centre is capable of affording selectivity is provided
by the excellent potency and significantly enhanced
GSK-3 selectivity of the 3-benzylamino derivative 20
when compared to the phenyl derivative 19. This is in
contrast to the 4-benzylamino derivative 21. Thus, we
were encouraged by these observations that access to
lipophilic interactions close to N1 with branched alkyl
or cycloalkyl functionality, together with a well posi-
tioned interaction with Glu138 may achieve the desired
combination of potency together with selectivity over
GSK-3 and RSK-1.
Figure 1. GOLD docking of compound 2 into an MSK-1 homology
model.
Whilst substitution at C4 with the small lipophilic Me
and Cl substituents in compounds 3 and 4, respectively,
was well tolerated (Table 2), increasing the size of the
substituent even moderately (compounds 5–7) produced
a significant reduction in activity. Larger substituents
(compounds 8–10) either abolished or greatly reduced
activity, presumably due to deleterious steric interac-
tions with gatekeeper residue Leu131 which prevent ac-
cess to the lipophilic pocket.
Interestingly, there was a moderate improvement in
selectivity of compound 3 compared with compound 2
for MSK-1 over GSK-3 and RSK-1. The reasons for
this are not clear since the gatekeeper residues for
GSK-3 and RSK-1 are the same as for MSK-1 (Leu)
and so other subtle requirements of the binding site
must be invoked.
Examination of the MSK-1 homology model (Fig. 1)
indicates that such interactions are also accessible
through modification of the C7 position of compound
2. We therefore next explored application of the above
findings at this position. The bromo substituent in com-
pound 22 (Table 4) retained good MSK-1 potency
although, as expected, lacked selectivity over either
GSK-3 or RSK-1. However, we were particularly
pleased to find that introduction of hetero-cycloalkyl
functionality (compounds 23–26) afforded greatly en-
hanced selectivity over both GSK-3 and RSK-1. Of par-
ticular note, 26 (SB-747651-A) shows a 10-fold increase
in MSK-1 potency and greatly enhanced selectivity over
23 suggesting an advantageous additional interaction
with MSK-1, possibly through Glu138 as predicted by
the MSK-1 homology model. Of interest to future stud-
ies will be the combination of this C7 substituent with
the C4-methyl group of compound 3.
We have previously shown4 that substitution at N1 with
lipophilic functionality such as cyclohexyl or cyclopro-
pyl affords potent MSK-1 inhibitory compounds 11
and 12, respectively (Table 3). Interestingly, whilst 11
had little selectivity over GSK-3, it showed a tendency
towards selectivity over RSK-1. Cyclopropyl derivative
12 showed improved GSK-3 selectivity. Furthermore,
introduction of substituents containing a basic nitrogen
such as the 4-piperidin-1-yl and the 4-butylamino deriv-
atives (13 and 14, respectively) retained MSK-1 activity
and showed a trend towards improved selectivity over
GSK-3 compared with that for 11 (Table 3). However,
selectivity over RSK-1 showed no improvement or was
reduced. These findings can be rationalised by our bind-
Table 1. Kinase selectivity profile of compound 2a
2
59
84
À6
15
94
90
13
À16
41
97
12
17
96
95
32
75
56
90
À7
18
6
67
a Values are %I at 10 lM using 100 lM ATP (see Ref. 5 for kinases used and assay details).

M. J. Bamford et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3407–3411
Table 2. MSK-1 activity and selectivity of C4 modified derivatives of compound 1
3409
Compound
R
MSK-1 IC50 (nM)
GSK-3/MSK-1a.
RSK-1/MSK-1b.
2
3
H
Me
3
8
11
104
13
11
—
—
1
8
21
19
4
Cl
Et
4
5
229
209
174
10,000
2240
813
1.5
6
OMe
OEt
1
6
1
3
1
7
8
OPh
N-Pyrrolidine
Ph
9
10
0.7
2
a Ratio of GSK-3 IC50 to MSK-1 IC50
b Ratio of RSK-1 IC50 to MSK-1 IC50
.
.
Table 3. MSK-1 activity and selectivity of N1-modified derivatives of compound 2
Compound
R1
MSK-1 IC50 (nM)
GSK-3/MSK-1a
RSK-1/MSK-1b
11
12
13
14
15
16
17
18
19
20
21
Cyclohexyl
Cyclopropyl
3
1.5
17
13
22
617
2
1.5
37
13
9
Piperidin-4-yl
(CH2)4NH2
(CH2)3NH2
12
41
17
2
23
33
15
33
83
12
8
(CH2)3NMe2
(R,S)-piperidin-3-yl
(R,S)-CH(Me)CH2NH2
Ph
6
17
9
69
10
6
3-(CH2NH2)Ph
4-(CH2NH2)Ph
2.5
4
634
13
11
a Ratio of GSK-3 IC50 to MSK-1 IC50
b Ratio of RSK-1 IC50 to MSK-1 IC50
.
.
Compound 26 (SB-747651-A) was further profiled5 (Ta-
ble 5) across those enzymes for which compound 2
showed pronounced activity (>70% inhibition at
10 lM, Table 1). These data confirm the significant
improvements in selectivity for 26 over these additional
kinases.
The C7 derivatives such as compounds 22 and 26 were
obtained as indicated in Scheme 2. Thus, ethyl-(3-
nitropyridine-4-yl)amine 284 was brominated in hot
acetic acid to afford the corresponding 5-bromo-ana-
logue 29 in moderate yield. In analogous fashion to
that employed for synthesis of compound 2,4 this
was then converted in three moderately yielding steps
to the key 7-bromo compound 22. This versatile inter-
mediate can be variously derivatised, including formy-
lated by treatment with n-BuLi followed by addition
of DMF to give the equally versatile 7-formyl interme-
diate 32. This can be reductively aminated with an
appropriate amine in the presence of polymer-bound
cyanoborohydride to give amines such as 33. Com-
pound 33 was deprotected with TFA to give desired
compound 26.
The N1-modified analogues were prepared6,7 as previ-
ously reported.4 The C4-modified analogues were gener-
ated as indicated in Scheme 1 for compound 3. Thus, an
acetic acid solution of compound 2 was treated with
30% aqueous hydrogen peroxide and heated at 90 °C
for 5 h to yield N-oxide 27 in moderate yield. This was
quantitatively transformed into chloro derivative 4 by
heating in phosphorus oxychloride for 6 h. Reaction
with an appropriate nucleophile then afforded the de-
sired C4-substituted products. For example, compound
4 in 1,4-dioxan was treated with a 2 M solution of tri-
methylaluminium in toluene and bis(triphenylphos-
phine)palladium (II) dichloride and then heated at
reflux for 4 h to give compound 3.
The (1H-imidazo[4,5-c]pyridin-2-yl)-1,2,5-oxadiazol-3-
ylamine template provides an excellent start-point for
identification of inhibitors of a number of kinases.
Potent and selective inhibitors of MSK-1 have been

3410
M. J. Bamford et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3407–3411
Table 4. MSK-1 activity and selectivity of selected C7-modified derivatives of compound 2
Compound
R
MSK-1 IC50 (nM)
GSK-3/MSK-1a
RSK-1/MSK-1b
22
23
Br
9
7
242
2
93
CH2-(1-piperidin-1-yl)
CH2-(1-piperazin-1-yl)
CH2-(4-NH2-piperidin-1-yl)
CH2NH(4-piperidin-4-yl)
7
24
25
1.4
3
257
635
81
69
26 (SB-747651-A)
0.5
3800
309
a Ratio of GSK-3 IC50 to MSK-1 IC50
b Ratio of RSK-1 IC50 to MSK-1 IC50
.
.
Table 5. Comparison of potency of compound 2 with C7-modified compound 26 for selected kinases
Compound
ROCK-1 IC50 (nM)6
CDK2 IC50 (nM)6
P70S6K IC50 (nM)
DYRK1a IC50 (nM)
2
26
13
100
169
>10000
70
270
110
4900
identified through optimisation using insights provided
by an MSK-1 protein homology model. In particular
this has led to the identification of exceptionally po-
tent MSK-1 inhibitor 26 (SB-747651-A) which has
an excellent kinase selectivity profile, including >300-
fold selectivity over RSK1 and >3000-fold selectivity
over GSK-3.
Acknowledgments
The authors acknowledge the assistance of Glaxo-
SmithKline colleagues from the Departments of Gene
Expression and Protein Biochemistry for the recombi-
nant human enzymes, Computational and Structural
Sciences for protein homology model generation, and
2
27
4
3
Scheme 1. Preparation of 4-substituted aza-benzimidazole derivatives 3 and 4. Reagents and conditions: (a) 30% H2O2, AcOH, 90 °C (53%); (b)
POCl3, 120 °C (100%); (c) 2 M Al(CH3)3 in toluene, 1,4-dioxan, Pd(PPh3)2Cl2, reflux (13%).
Scheme 2. Preparation of 7-substituted aza-benzimidazole derivatives 22 and 26. Reagents and conditions: (a) Br2, AcOH, 100 °C (43%); (b) sodium
dithionite, ethanol/water, 60 °C; (c) Ethyl cyanoacetate, 190 °C (35%, two steps); (d) (i) NaNO2, MeOH, HCl, rt; (ii) NH2OH, NaOH, reflux (45%, 2
steps); (e) n-BuLi, DMF, THF, À78 °C (41%); (f) 4-amino-piperidine-1-carboxylic acid tert-butyl ester, (polystyrylmethyl)trimethylammonium
cyanoborohydride, MeOH, AcOH, 21 °C (42%); (g) TFA, DCM, 21 °C (82%).

M. J. Bamford et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3407–3411
3411
J.; Erhardt, J. A.; Ohlstein, E. H.; Hunter, A. J.; Harrison,
D. C.; Philpott, K.; Smith, B. R.; Adams, J. L.; Parsons, A.
A. Med. Res. Rev. 2001, 21, 129; (c) Gaiba, A.; Takle, A.
K.; Wilson, D. M. WO Patent 66540:2001; (d) Irving, E. A.;
Ray, A. M.; Staton, P. C.; Davis, R. P.; Hadingham, S. J.;
Campbell, C. A.; Roberts, J.; Reith, A. D.; Hunter, A. J.;
Parsons, A. A. J. Cerebr. Bl. Flow Metab. 2001, 21, S379.
4. Bamford, M. J.; Alberti, M. J.; Bailey, N.; Davies, S.; Dean,
D. K.; Gaiba, A.; Garland, S.; Harling, J. D.; Jung, D. K.;
Panchal, T. A.; Patt, C. A.; Steadman, J. G.; Takle, A. K.;
Townsend, J. T.; Wilson, D. M.; Witherington, J. Bioorg
Med. Chem. Lett. 2005, 15, 3402.
Assay Development & Compound Profiling for assay of
the compounds.
References and notes
1. (a) Deak, M.; Clifton, A. D.; Lucocq, L. M.; Alessi, D. R.
EMBO J. 1998, 17, 4426; (b) Thomson, S.; Clayton, A. L.;
Hazzalin, C. A.; Rose, S.; Barratt, M. J.; Mahadevan, L. C.
EMBO J. 1999, 18, 4779; (c) Wiggin, G. R.; Soloaga, A.;
Foster, J. M.; Murray-Tait, V.; Cohen, P.; Arthur, J. S. C.
Mol. Cell. Biol. 2002, 22, 2871; (d) Soloaga, A.; Thomson,
S.; Wiggin, G. R.; Rampersaud, N.; Dyson, M. H.;
Hazzalin, C. A.; Mahadevan, L. C.; Arthur, J. S. C.
EMBO J. 2003, 22, 2788; (e) Vermeulen, L.; De Wilde, G.;
Van Damme, P.; Vanden Berghe, W.; Haegeman, G.
EMBO J. 2003, 22, 1313.
2. Hughes, J. P.; Staton, P. C.; Wilkinson, M. G.; Strijbos, P.
J. L. M.; Skaper, S. D.; Arthur, J. S. C.; Reith, A. D.
J. Neurochem. 2003, 86, 25.
3. (a) Irving, E. A.; Bamford, M. J. Cerebr. Bl. Flow Metab.
2002, 22, 631; (b) Barone, F. C.; Irving, E. A.; Ray, A. M.;
Lee, J. C.; Kassis, S.; Kumar, S.; Badger, A. M.; Legos, J.
5. Davies, S. P.; Reddy, H.; Caivano, M.; Cohen, P. Biochem.
J. 2000, 351, 95.
6. MSK-1 was assayed as previously described.4 ROCK-1
was assayed as decribed in patent application WO2003/
080610. GSK-3 and CDK-2 were assayed by SPA-based
assays as described in, Tavares, F. X.; Boucheron, J. A.;
Dickerson, S. H.; Griffin, R. J.; Preugschat, F.; Thom-
son, S. A.; Wang, T. Y.; Zhou, H.-Q. J. Med. Chem.
2004, 47, 4716.
1
7. All novel compounds gave satisfactory H NMR and LC/
MS data in full agreement with their proposed structures.