Optimization of triarylimidazoles for Tie2: Influence of conformation on potency

Article abstract of DOI:10.1016/j.bmcl.2007.08.052

In an effort to understand the effect of N-alkylation of triarylimidazoles on Tie2 inhibition, ortho-substituted C-2 aryl analogs were synthesized to investigate the effect of different torsion angles on potency. This exercise resulted in the identificati

Full text of DOI:10.1016/j.bmcl.2007.08.052

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5514–5517
Optimization of triarylimidazoles for Tie2: Influence
of conformation on potency
Neil W. Johnson,a,* Marcus Semones,a Jerry L. Adams,a
Michael Hansburyb and Jim Winklerc
aGlaxoSmithKline, 1250 South Collegeville Road, Collegeville, PA 19426, USA
bIncyte Pharm., Wilmington, DE, USA
cArray Biopharma, Boulder, CO, USA
Received 14 May 2007; revised 21 August 2007; accepted 23 August 2007
Available online 28 August 2007
Abstract—In an effort to understand the effect of N-alkylation of triarylimidazoles on Tie2 inhibition, ortho-substituted C-2 aryl
analogs were synthesized to investigate the effect of different torsion angles on potency. This exercise resulted in the identification
of a potent and selective tetrasubstituted imidazole that was efficacious in an animal model of angiogenesis.
Ó 2007 Elsevier Ltd. All rights reserved.
Several receptor tyrosine kinases (VEGF, FGF, PDGF)
are known to be involved in the process of angiogene-
sis—the formation of new capillaries from established
blood vessels. Similar to VEGFR2, the Tie family of
receptors, Tie1 and Tie2, are expressed predominately
on endothelial cells1 and are essential for vessel forma-
tion. Interference of the Tie2 pathway with an extracel-
lular receptor domain results in a significant inhibition
of tumor growth.2 Furthermore, targeted disruption
of the tie2 gene results in malformed vasculature
in embryonic mice.3 The recent approval of an anti-
VEGF monoclonal antibody (AvastinTM; Genentech,
and Roche)4 for the treatment of metastatic colorectal
cancers has validated the earlier hypothesis that disrup-
tion of angiogenesis could be an effective strategy for the
treatment of cancer. The involvement of Tie2 kinase
in angiogenesis makes it an attractive target for
investigation.
ole (Fig. 1, compound 1 vs compound 2) has been
difficult to rationalize.
To gain insight into the effect of N-alkylation of the
imidazole, small molecule X-ray crystal structures were
obtained for compounds 1 and 2.6 The torsion angles
of the aryl rings attached to the imidazole showed that
the C-2 aryl ring is slightly out of coplanarity with the
imidazole in compound 1, whereas in compound 2 it is
out of plane by almost 50°. When the methyl group is
23o
N
2o
N
H 16o
47o
O
N
N
O
S
S
N
N
39o
67o
O
O
1
2
Tie2 IC50 = 250 nM
60nM
Our group has been investigating a series of triarylimi-
dazoles as selective inhibitors of Tie2. As reported in
the accompanying manuscript,5 the incorporation of
an N-alkyl substituent on the triarylimidazole core was
found to improve Tie2 activity. The significant improve-
ment in Tie2 inhibition upon methylation of the imidaz-
30o
N
44o
O
N
N
S
37o
O
3
>20000nM
Keywords: Tie2 kinase; Angiogenesis; Cancer.
*
Figure 1. Representative triaryl imidazoles and their torsion angles
with Tie2 IC50’s.
Corresponding author. Tel.: +1 610 917 6292; fax: +1 610 917
6020; e-mail: neil.w.johnson@gsk.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.08.052

N. W. Johnson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5514–5517
5515
condensation of benzaldehyde 5 with keto-oxime 65 gave
the triarylimidazole N-oxide that was then reduced to
the imidazole by heating with trimethylphosphite in
DMF. The resulting sulfide was oxidized with potassium
persulfate to give the ortho-methyl C-2 aryl analog 7.
This compound was shown to inhibit Tie2 with an
IC50 of 36 nM, an almost 2-fold improvement in po-
tency versus compound 2 and a 10-fold improvement
over compound 1. After separation of the sulfoxide
enantiomers via chiral HPLC, a small molecule X-ray
was determined for the S-enantiomer of compound 7.6
While the S-enantiomer was not as active as the R
(95 nM vs 22 nM, respectively) it still showed similar
activity to compound 2 and thus a 3-fold improvement
over compound 1. The torsion angle of the C-2 aryl
group was found to be 45°. The pyridine (30°) and naph-
thyl (48°) torsion angles were similar to those seen for
compound 1. Hence, the increase in potency can be par-
tially attributed to the C-2 aryl group being forced out
of plane with the imidazole.
Figure 2. Docking model of Compound 1 in Tie2.
adjacent to the naphthyl ring (compound 2) this forces
the ring out of plane by 67° allowing the pyridine to
be nearly coplanar with the imidazole. According to
modeling studies, alkylating adjacent to the pyridine
ring, as in compound 3, would force the pyridine ring
to be 30° out of plane and all other torsion angles to
be almost equivalent. In our proposed binding mode7
(Fig. 2), the pyridine hydrogen bonds Ala905 in the
hinge region of the ATP pocket with the naphthyl ring
occupying a lipophilic back pocket. As shown above,
methylating adjacent to the naphthyl ring is allowed
and preferred for Tie2 activity. The increase in the tor-
sion angle of the naphthyl group along with the pyridine
ring reaching coplanarity appears to be an optimal
conformation for binding to Tie2. What remains unclear
from this assessment is whether modifying the torsion
angle of the C-2 aryl ring has any impact on Tie2
activity. In order to independently evaluate this torsion
angle an ortho-methyl aryl analog of compound 1 was
synthesized.
Imidazole condensations were then performed with a
series of benzaldehydes and the pyridinyl 6-methoxy-
naphthyl keto-oxime 6 to investigate the SAR for ortho
substituted phenyl rings. Results are shown in Table 1.
The ortho-methyl phenyl analog 9 is significantly more
Table 1. Tie2 IC50’s for ortho-substituted C-2 phenyl triarylimidazoles
N
H
N
N
X
O
Compound
X
Tie2 IC50, nM
8
9
H
Me
60% at 10 lM
387
1600
157
82
10
11
12
13
14
15
16
F
Cl
The requisite ortho-substituted benzaldehyde was syn-
thesized from commercially available 4-bromo-2-methyl
benzonitrile 4, as shown in Scheme 1. The bromide was
displaced with sodium thiomethoxide followed by
reduction of the nitrile with DIBAL-H to afford the de-
sired 2-methyl-4-thiomethyl benzaldehyde 5. Using stan-
dard conditions for the formation of imidazoles,8
Br
CF3
2,6-DiCl
Ethyl
i-Propyl
260
78
346
836
Values are means of at least two experiments.
O
S
H
1) NaSMe/DMF
N
Br
2) Dibal-H/toluene
5
4
30o
N
N
1) NH4OH, 5
AcOH, 90 ºC
N
O
OH
O
S
N
N
2) P(OMe)3
45o
H
DMF, 90 ºC
3) K2S2O8
O
48o
O
6
7
36nM
Scheme 1. Synthesis of triaryl imidazole 7.

5516
N. W. Johnson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5514–5517
1) a) MeNH2 aq
AcOH
b) 6, 90 ºC
N-oxide. The remainder of the synthesis went forward as
performed previously.
N
X
X
O
H
O
S
N
N
S
2) P(OMe)3
The corresponding para-sulfide, sulfoxide and sulfone
analogs were tested along with the N-methylimidazole
analogs of the ortho-methylphenyl derivatives (Table 2).
For comparison, data for compounds 1, 2, and 7 are in-
cluded in the table. There was very little difference in po-
tency between the sulfones and sulfoxides (compound 7
vs 18 and 20 vs 21), except with compound 23 vs 24
where there was a nearly 10-fold difference. Compound
23, which possesses both the preferred ortho-methyl and
N-methyl imidazole, showed an additional 5-fold
increase in potency over the original N-methyl imidazole
2, and a 3-fold improvement in potency compared
to compound 7. Unlike the earlier result with mono-
substituted phenyls in Table 1, the ortho-chloro analogs
were not more potent than the ortho-methyl analogs
(i.e., compound 20 vs 7). The combined effect of increas-
ing the torsion angles of the naphthyl group and the
C-2 phenyl group is favorable for Tie2 inhibition.9
DMF, 90 ºC
3) K2S2O8
O
Scheme 2. Regioselective N-Methyl imidazoles.
potent than the unsubstituted phenyl analog 8 and still
has modest Tie2 inhibitory activity without the para-
methylsulfoxide. The ortho-fluoro analog 10 was much
less active, possibly due to its small size allowing a
conformation similar to the unsubstituted analog 8.
Potency was improved with ortho-chloro (compound
11) and even further with ortho-bromo and 2,6-di-
chloro (compounds 12 and 14) vs ortho-methyl (com-
pound 9). A modest improvement in Tie2 activity
was observed for the ortho-CF3 (compound 13) while
the ortho-ethyl was equivalent in activity and the
ortho-isopropyl had a negative impact on activity (com-
pounds 13 and 15, respectively). Given the increase in
size and lipophilicity of adding a bromo or 2,6-dichloro
substituent, the chloro and methyl substituents were
chosen as the best compromise between potency and
size.
Compound 23, our most potent Tie2 inhibitor, was
examined against a panel of kinases. Representative
data are given in Table 3. Very high selectivity is shown
against both tyrosine kinases and serine/threonine ki-
nases. Even for the most sensitive kinases in this panel,
c-fms and Alk5, selectivity remains good, being 25-fold
for Alk5 and 50-fold for c-fms.
We then wanted to combine the para-methylsulfoxide
with the optimized ortho-methyl and ortho-chloro C-2
phenyl rings as well as examine the effect of these
changes in the N-methylated triarylimidazoles. A regio-
selective synthesis of the N-methyl analogs was devel-
oped to simplify the synthesis of these compounds.
The appropriate benzaldehydes were synthesized using
a route similar to that used in Scheme 1. The benzalde-
hyde was stirred in acetic acid and aqueous methyl
amine (Scheme 2) to generate the corresponding methyl
imine. The keto-oxime 6 was then added and the reac-
tion mixture was heated at 90 °C. This condensation
method resulted in a single regioisomer of the imidazole
The in vivo efficacy of our optimized compound 23 was
tested in the matrigel model of angiogenesis in mice.10 In
this model, angiogenesis is stimulated with bFGF and
two doses of compound are given. The two control
groups are untreated mice and mice that are not given
bFGF. The mice are sacrificed after 6 days and the heme
content of the matrigel plug is measured. At two doses
of compound 23 (25 and 50 mg/kg p.o., b.i.d.) inhibition
of angiogenesis (35 and 80%, respectively) was observed
in this model.
In summary, we have successfully exploited a conforma-
tional hypothesis to optimize a series of triarylimidazoles
Table 2. Combination of ortho and para-substituents
N
X
Table 3. Kinase selectivity data for compound 23
N
S(O)n
Kinase
IC50, nM
N
R
IGF1-R
INS-R
FLT3
>30,000
>30,000
>30,000
>30,000
>30,000
>30,000
1500
O
KIT
Compound
R
X
n
Tie2 IC50, nM
PDGFR-a
PDGFR-b
VEGFR-2
FGF-R1
MET
1
2
H
H
2
1
0
1
2
0
1
2
0
1
2
250
60
Me
H
H
17
7
Me
Me
Me
Cl
1300
36
9700
H
>30,000
>30,000
>30,000
4100
18
19
20
21
22
23
24
H
40
ABL
PKC-b1
p38
H
1500
51
H
Cl
Cl
H
59
c-FMS
CDK2
GSK3
550
Me
Me
Me
Me
Me
Me
335
12
>30,000
>30,000
296
114
ALK5
AKT1
>30,000
Values are means of at least two experiments.

N. W. Johnson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5514–5517
5517
application to CCDC, 12 Union Road, Cambridge CB2
1EZ, UK [fax: +44 (0)1223-336033 or e-mail:
deposit@ccdc.cam.ac.uk].
for Tie2 potency. The alkylation of the imidazole adja-
cent to the naphthyl ring and incorporation of a C-2
ortho-aryl substitution to increase its torsion angle
relative to the imidazole resulted in a potent and selec-
tive Tie2 inhibitor. The optimized compound 23 showed
efficacy in an in vivo model of angiogenesis.
7. 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 1–3 were manually docked into the
homology model based on the bound conformation
of SB-203850 in crystal structure 1a9u. Figure
depicts the manual docking model for compound 1 in
second-generation homology model of an active
2
References and notes
a
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 tem-
plates. Figures were generated using the program
pymol.
1. (a) Jones, N.; Iljin, K.; Dumont, D. J.; Alitalo, K. Nat.
Rev. Mol. Cell Biol. 2001, 2, 257; (b) Tsigkos, S.;
Koutsilieris, M.; Papapetropoulos, A. Expert Opin. Inves-
tig. Drugs 2003, 12, 935.
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.
3. Sato, T. N.; Tozawa, Y.; Deutsch, U.; Wolburg-Buchholz,
K.; Fujiwara, Y.; Gendron-Maguire, M.; Gridley, T.;
Wolburg, H.; Risau, W.; Ying, Q. Nature 1995, 376, 70.
4. Ferrara, N.; Hillan, K. J.; Gerber, H-P.; Novotny, W. Nat.
Rev. Drug Disc. 2004, 3, 391.
5. Semones, M.; Feng, Y.; Johnson, N. W.; L. Adams, J. L.;
Winkler J.; Hansbury, M. Biorg. Med. Chem. Lett. 2007,
17, 4756.
8. (a) Volkamer, K.; Zimmerman, H. Chem. Ber. 1969, 102,
4177; (b) Liverton, N. J.; Butcher, J. W.; Claiborne, C. F.;
Claremon, D. A.; Libbey, B. E.; Nguyen, K. T.; Pitzen-
berger, S. M.; Selnick, H. G.; Smith, G. R.; Tebben, A.;
Vacca, J. P.; Varga, S. L.; Agarwal, L.; Dancheck, K.;
Forsyth, A. J.; Fletcher, D. S.; Frantz, B.; Hanlon, W. A.;
Harper, C. F.; Hofsess, S. J.; Kostura, M.; Lin, J.; Luell,
S.; O’Neill, E. A.; Orevillo, C. J.; Pang, M.; Parsons, J.;
Rolando, A.; Sahly, Y.; Visco, D. M.; O’Keefe, S. J.
J. Med. Chem. 1999, 42, 2180.
9. Calculated dihedral angles for compound 23 using MOE
are; C-2 aryl to imidazole 65°, pyridine to imidazole 38°,
naphthyl to imidazole 66°.
10. (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.
6. Crystallographic data (excluding structure factors) for the
structures in this paper have been deposited with the
Cambridge Crystallographic Data Centre as supplemen-
tary publication numbers CCDC 646583 and 646584.
Copies of the data can be obtained, free of charge, on