Pyridinylimidazole inhibitors of Tie2 kinase

Article abstract of DOI:10.1016/j.bmcl.2007.06.068

This communication details the evolution of the screening lead SB-203580, a known CSBP/p38 kinase inhibitor, into a potent and selective Tie2 tyrosine kinase inhibitor. The optimized compound 5 showed efficacy in an in vivo model of angiogenesis and a MOP

Full text of DOI:10.1016/j.bmcl.2007.06.068

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4756–4760
Pyridinylimidazole inhibitors of Tie2 kinase
Marcus Semones,a,* Yanhong Feng,a Neil Johnson,a Jerry L. Adams,a
Jim Winklerc and Michael Hansburyb
aGlaxoSmithKline, 1250 South Collegeville Road, Collegeville, PA 19426, USA
bIncyte Pharm., Wilmington, DE, USA
cArray Biopharma, Boulder, CO, USA
Received 24 April 2007; revised 18 June 2007; accepted 21 June 2007
Available online 27 June 2007
Abstract—This communication details the evolution of the screening lead SB-203580, a known CSBP/p38 kinase inhibitor, into a
potent and selective Tie2 tyrosine kinase inhibitor. The optimized compound 5 showed efficacy in an in vivo model of angiogenesis
and a MOPC-315 plasmacytoma xenograft model.
Ó 2007 Elsevier Ltd. All rights reserved.
Receptor tyrosine kinases (RTKs) are involved in the
process of angiogenesis, defined as the formation of
new capillaries from established blood vessels. In partic-
ular, angiogenesis is dependent on the vascular endothe-
lial growth factor (VEGFR2/KDR) and Tie2.1 Several
studies have shown that many tumors are inhibited by
the blockade of the VEGF/VEGF receptor pathway,
while others are unaffected, suggesting that alternative
pathways for vascular growth can drive tumor angio-
genesis.2,4,5 It has been demonstrated that blocking
Tie-2 activation with a recombinant Tie2 receptor
AdExTek inhibits tumor angiogenesis and tumor
growth in vivo.1,2 Therefore, there is an expectation that
small molecule inhibitors of Tie2 kinase would also be
attractive candidates as anti-angiogenic cancer chemo-
therapeutic agents.
of SB-203580 with p38 and mimicks many of the key
interactions demonstrated to be crucial for binding.
For example in the SB-203580 p38 co-crystal structure,
the 4-pyridyl nitrogen forms a hydrogen bond with the
backbone amide nitrogen of Met109. In analogy with
what has been demonstrated in other kinase crystal
structures, the 4-pyridyl nitrogen of compound 1 is
believed to mimic the interaction of the NÀ1 of adenine
in ATP with an NH of the amide backbone. Consistent
with this role of the 4-pyridyl group, the phenyl analog
of compound 1 was prepared and found to be inactive
(IC50 > 100 lM) for Tie2 kinase (data not shown).
Initially, we sought to investigate the SAR of the imid-
azole 4-position (R2 in Table 1) with the goals of
increasing Tie2 potency and enhancing selectivity
against p38 kinase. It was apparent from the initial
SAR (data not shown) that only naphthyl substitution
at the 4-position of the imidazole afforded Tie2 potency.
This aspect of the SAR can be rationalized by examina-
tion of the Tie2/compound 1 docking model (Fig. 1).
Thus, the naphthyl moiety is predicted to occupy the
aryl specificity pocket, lined by residues L876, I886,
L888, L900, and I902 in Tie2. The back pocket in Tie2
is considerably deeper than that of p38, and may there-
fore, favor increased potency for naphthyl-containing
inhibitors against Tie2 kinase and conversely disfavor
binding to p38. Further analysis of the docking model
in Tie2 suggested that introduction of functional groups
at the 6-position of the naphthyl ring might improve
potency for Tie2 and enhance selectivity against p38.
Screening efforts in our laboratories identified two tri-
substituted imidazoles, SB-203580 and 2-naphthyl
substituted compound 1 (Table 1), as Tie2 kinase inhib-
itors. SB-203580 had poor intrinsic potency and no cel-
lular activity, while compound 1 exhibited moderate
potency (Tie2 IC50 = 300 nM) and poor cellular activity
(cell IC50 = 30,000 nM).7
The binding model for compound 1 in Tie23,9 (Figs. 1
and 2) is based on the published co-crystal structures
Keywords: Tie2 kinase; Imidazole; Xenograft.
*
Corresponding author. Tel.: +1 610 917 5439; e-mail: marcus.a.
semones@gsk.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.06.068

M. Semones et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4756–4760
Table 1. Tie2 and p38 kinase activity of compounds 1–57
4757
Compound
Tie2 IC50 (nM)
p38 IC50 (nM)
N
H
N
O
S
N
R2
SB-203580
20,000
300
42
F
1
2
300
4571
339
Br
3
4
5
5500
219
16,500
741
CH3
O
250
50,000
Figure 1. Tie2 docking model for compound 1.9
Figure 2. Tie2 docking model for compound 5.9
To probe this hypothesis, compounds 2–5 containing
substituents at the 6-position of the naphthyl group
were synthesized according to the synthetic route out-
lined in Scheme 1. Conversion of the naphthoic acid
to the acid chloride followed by reaction with N,O-
dimethylhydroxylamine gave the desired Weinreb
amide 6. Nucleophilic addition of lithiated 4-picoline
to the latter provided ethanone 7 that was readily con-
verted to the keto-oxime 8 as a mixture of geometric
isomers. Cyclodehydration to the desired imidazol-1ol
was accomplished by in situ generation of the imine
from a suitable aldehyde [in this case, 4-(methylthio)
benzaldehyde] and ammonium acetate, followed by
addition of the keto-oxime in acetic acid. Finally,
reduction of the crude N-oxide 9 with trimethyl phos-
phate followed by oxidation of the thiomethyl group
with potassium persulfate afforded the requisite racemic
imidazoles.

4758
M. Semones et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4756–4760
O
O
SOCl2,CH2Cl2, Et3N
LDA, 4-picoline,
THF, 0 oC, 70%
OH
N
O
R
R
CH3ONHCH3 . HCl
90-100 %
6
N
N
1.p-CH3SC6H4CHO,
CH3CO2NH4
N
OH
NaNO2
O
O
1:1 3N HCl-water
5 oC 90%
CH3CO2H, reflux
R
R
8
7
R
R
N
2. (CH3O)3P, DMF,
reflux
N
3. K2S2O8, CH3CO2H,
H2O 50-60 %
N
N
HO
N
N
O
H
S
S
9
1-5
Scheme 1. The synthesis of compounds 1–5.
Compounds 4 and 5 with a methyl and a methoxy sub-
stituents at the 6-position of the naphthyl group, respec-
tively, showed only a slight improvement in Tie2
inhibition; however, a remarkable improvement in selec-
tivity over p38 versus compound 1 (13- and 200-fold in-
crease) was achieved. A substantial decrease in potency
against Tie-2 was observed with the 6-phenyl substituted
naphthyl group (3 vs 1).
of this observation will be addressed in the following
paper.
Compounds 5 and 10 were found to have moderate to
excellent cellular activities (cell IC50 = 232 nM and
24.4 nM, respectively). Compounds 5 and 10 displayed
similar pharmacokinetic parameters in rodents and were
advanced into an in vivo angiogenesis model.
Next, we investigated what effect an N-alkyl group
placed on the triarylimidazole core would have on
selectivity and potency of this series for Tie2 over
p38. Synthesis of the tetra-substituted imidazoles 10–
14 was accomplished by N-alkylation of imidazole 5
with an appropriate electrophile (Cs2CO3, THF °C
2 h). The resultant mixtures of regioisomers were sepa-
rated by preparative reverse phase HPLC and fully
We used the Matrigel model of angiogenesis in mice to
test the in vivo efficacy of our optimized compound.7
Angiogenesis is stimulated with bFGF and two doses
of compound are administered. The two control groups
are non-treated mice and mice that are not given bFGF.
The mice are sacrificed after 6 days and the Heme con-
tent of the Matrigel plug is measured. At doses of 25 and
50 mg/kg ip b.i.d, compound 5 showed a reduction of
41% and 70%, and compound 10 showed a reduction
of only 5% and 30%, respectively, of angiogenesis in this
model.
1
characterized by H NMR and small molecule X-ray
crystallography.6
The data in Table 2 indicate a clear preference for N-
methyl substitution on the imidazole ring adjacent to
the naphthyl group (compound 10 vs 11). To our de-
light, regioisomer 10 also displayed a fourfold in-
crease in potency over the parent tri-substituted
imidazole 5. Alkyl groups larger than methyl were
not as well tolerated. Our Tie2 homology model
places the N-methyl group in a space occupied by
the catalytic lysine relative to p38. The conserved ly-
sine is pushed further back in the pocket in Tie2,
allowing the methyl group to reside in that area with-
out hindering binding to the kinase. Further analysis
Based on the encouraging result from the angiogenesis
model, compound 57,8 was advanced into a MOPC-
315 plasmacytoma xenograft model. Figure 3 below
shows that compound 5 induced a modest dose depen-
dent delay in tumor growth.
In summary, we have optimized SB-203580, a known
CSBP/p38 kinase inhibitor, into a potent and selective
Tie2 tyrosine kinase inhibitor. The optimized compound
5 showed efficacy in an in vivo model of angiogenesis
and a MOPC-315 plasmacytoma xenograft model.

M. Semones et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4756–4760
Table 2. Tie2 and p38 kinase activity of alkylated compounds
4759
Compound
Tetra-substituted imidazole
Tie2 IC50 (nM)
p38 IC50 (nM)
N
O
N
10
60
16,000
S
N
O
N
N
O
S
11
12
13
14
20,000
5370
3890
437
16,000
N
O
N
O
S
N
N
>16,000
>16,000
>16,000
O
N
O
N
N
S
O
N
O
N
N
S
O
2. (a) Lin, P.; Buxton, J.; Acheson, A.; Radziejewski, C.;
Maisonpierre, P.; Yancopoulos, G.; Channon, K.; Hale, L.;
Dewhirst, M.; George, E.; Peters, K. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 8829; (b) Siemeister, G.; Schirner, M.;
Weindel, K.; Reusch, P.; Menrad, A.; Marme, D.; Martiny-
Baron, G. Cancer Res. 1999, 59, 3185.
Effect of Early Treatment with Compound 5
Tumor Growth of sc MOPC-315
104
1000
100
10
50 mg
25 mg
12.5 mg
6.25 mg
3.12 mg
3. (a) Sato, T. N.; Tozawa, Y.; Deutsch, U.; Wolburg-Buch-
holz, K.; Fujiwara, Y.; Gendron-Maguire, M.; Gridley, T.;
Wolburg, H.; Risau, W.; Qin, Y. Structure 1998, 6, 1117,
London; (b) 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,
London; (c) Shewchuk, L. M.; Hassell, A. M.; Ellis, B.;
Holmes, W. D.; Davis, R.; Horne, E. L.; Kadwell, S. H.;
McKee, D. D.; Moore, J. T. Structure 2000, 8, 1105, London.
4. (a) Volkamer, K.; Zimmerman, H. Chem. Ber. 1969, 102,
4177; (b) Liverton, N. et al. J. Med. Chem. 1999, 42, 2180.
5. (a) Biancone, L.; De Martino, A.; Orlandi, V.; Conaldi, P.
G.; Toniolo, A.; Camussi, G. J. Exp. Med. 1997, 186, 147;
(b) Drabkin, D. L.; Austin, J. H. J. Biol. Chem. 1935, 112,
51.
Vehicle Control
1
10
14
18
22
26
30
Days Postimplantation
Compound 5: ip, mg/kg, BID
Vehicle: ip, BID
5% EtOH + 5% Cremophor + 90% H2O
Figure 3. MOPC-315 xenograft model for compound 5.7
References and notes
6. An ortep view of compound 10. Crystallographic data
(excluding structure factors) for the structures in this paper
have been deposited with the Cambridge Crystallographic
1. (a) Jones, N.; Iljin, K.; Dumont, D. J.; Alitalo, K. Nat. Rev.
Mol. Cell Biol. 2001, 2, 257; (b) Tsigkos, S. et al. Expert
Opin. Investig. Drugs 2003, 12, 935.

4760
M. Semones et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4756–4760
Data Centre as supplementary publication numbers CCDC
assay. Compounds 5 and 10 were tested in our Tie2
autophosphorylation assay. For assay conditions see
Semones, M. A. PCT Int. Appl. (2002), WO 2002060382
and Adams, J. L.; Kasparec, J.; Silva, D.; Yuan, C. C. K.
PCT Int. Appl. (2003), WO 2003029209.
644904. Copies of the data can be obtained, free of charge,
on application to CCDC, 12 Union Road, Cambridge CB2
1EZ, UK [fax: +44 (0)1223 336033 or e-mail: deposit@ccdc.
cam.ac.uk].
8. Compound 5 has a >10-fold selectivity against VEGFR2,
VEGFR3, and PDGFR1b.
9. The homology model of Tie2 was generated in the program
MOE using the crystal structure FGFR-2 kinase (PDB
code 1oec) as the structural template; compounds 1 and 5
were manually docked into the homology model based on
the bound conformation of SB-203850 in crystal structure
1a9u. Figure 2 depicts the manual docking model for
compound 5 in a second-generation homology model of an
active form of Tie2; this second-generation homology
model was generated in the program MOE using the
crystal structures of inactive Tie2 (PDB code 1fvr) and
active FGFR-2 kinase (PDB code 1oec) as structural
templates. Figures were generated using the program
pymol.
7. Compound 1 was tested in our Tie2 receptor signal
transduction assay and our Tie2 autophosphorylation