Knowledge-based design of 7-azaindoles as selective B-Raf inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.07.019

The synthesis of a 7-azaindole series of novel, potent B-Raf kinase inhibitors using knowledge-based design was carried out. Compound 6h exhibits not only excellent potency in both the enzyme assay (IC₅₀ = 2.5 nM) and the cellular assay (IC

Full text of DOI:10.1016/j.bmcl.2008.07.019

Bioorganic & Medicinal Chemistry Letters 18 (2008) 4610–4614
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Knowledge-based design of 7-azaindoles as selective B-Raf inhibitors
b
b
b
a
a
Jun Tang a, , Toshihiro Hamajima , Masato Nakano , Hideyuki Sato , Scott H. Dickerson , Karen E. Lackey
*
a GlaxoSmithKline, Research Triangle Park, NC 27709, USA
b GlaxoSmithKline K. K. Tsukuba Research Laboratories, 43 Wadai Tsukuba, Ibaraki 300-4247, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The synthesis of a 7-azaindole series of novel, potent B-Raf kinase inhibitors using knowledge-based
design was carried out. Compound 6h exhibits not only excellent potency in both the enzyme assay
(IC50 = 2.5 nM) and the cellular assay (IC50 = 63 nM), but also has an outstanding selectivity profile against
other kinases.
Received 6 June 2008
Revised 3 July 2008
Accepted 8 July 2008
Available online 10 July 2008
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
B-Raf kinase
7-Azaindole
Knowledge-based design
As a target family, kinases have been widely accepted as one of
the largest druggable gene families for drug discovery.1 In order to
find a molecule for each druggable kinase target, High-Throughput
Screening can be used to obtain a primary starting point, which is
further subjected to multidimensional optimization. As more and
more crystal structures of protein kinases complexed with ATP or
an ATP-competitive inhibitor have been published in the Protein
Data Bank and accumulated in individual research groups, knowl-
edge-based design is emerging as the predominant design tool.2 In
particular, the refined pharmacophore map and molecular recogni-
tion of protein kinase binding pockets have recently been pub-
lished,3 which provide deep insight for medicinal chemists to
carry out relatively reliable rational design of the desired inhibi-
tors. Recent reviews describe the methodology very well.4 Herein,
we outline our strategy and illustrate a successful example of how
we developed a potent and selective B-Raf kinase inhibitor using
knowledge-based design.
Focusing B-Raf kinase as a drug intervention for cancer ema-
nates from the biological elucidation of the Raf-MEK-ERK signaling
pathway and the fact that Raf kinases play an important role in
activating MEK and promoting cell proliferation and survival.5 Fur-
thermore, the small molecule, Sorafenib (Nexavar, BAY-43-9006),
was approved as a Raf/VEGFR2 inhibitor for the treatment of renal
cell carcinoma in 2005.6 However, given the limited number of
selective B-Raf kinase inhibitors so far reported (Fig. 1), we wanted
to discover new selective B-Baf inhibitors, which could be used to
enhance our understanding of the problems due to lack of
selectivity.7
A 2-D pharmacophore map used in this study is described in
Figure 2. It depicts the seven critical binding regions of the ATP-
binding domain that can be accessed to generate ATP-competitive
kinase inhibitors. The 7-azaindole series was elaborated by incor-
porating functional groups that were designed to interact with
key features of these regions. The scaffold of an ATP-competitive
kinase inhibitor typically included a hinge binder as the starting
point for the design. Many heteroaromatic groups, such as pyri-
dine, pyrimidine, aminopyrazole, quinoline, and quinazoline, can
be incorporated into the molecule to access the hinge-binding re-
gion. We chose 7-azaindole as the hinge binder because this do-
nor-acceptor combination generally provided greater potency
than the others indicated above.8
H
CF3
O
Cl
N
HN
HN
O
N
N
N
H
O
O
N
Sorafenib
SB-590885
HN
O
N
* Corresponding author. Tel.: +1 919 483 4173; fax: +1 919 315 0430.
Figure 1. Non-selective B-Raf inhibitor Sorafenib and selective B-Raf inhibitor
E-mail address: jun.tang@gsk.com (J. Tang).
SB-590885.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.07.019

J. Tang et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4610–4614
4611
Gate Keeper
induced fit
induced fit
NH
O
NH
inner
hydrophobic
inner
hydrophobic
Lys
NH
O
Glu
O
HO
N
H
Scaffold
N H
O
H
Phosphate
Binding
HN
O
OH
Asp
Sugar
Organizer
Pocket
Outer
Outer
Hydrophobic
Hydrophobic
Solvent Front
Solvent Front
Figure 2. 2-D pharmacophore map of the ATP kinase domain and the proposed scaffold interactions created by the inhibitor.
As designated by its name, the inner hydrophobic pocket prefers
an aromatic ring or one of its bioisosteres. On the other hand, the
induced fit pocket is a relatively big pocket featuring a flexible
phenylalanine (Phe594 in B-Raf kinase). With regard to potency
for B-Raf kinase inhibitors, either substituted or unsubstituted
hydrophobic aromatic rings are good options for binding in this
region. A reasonable linker to connect the two parts is a urea moi-
ety, which is found in the known B-Raf kinase inhibitor Sorafenib.
Additionally, the urea moiety should pick up some additional do-
nor-acceptor interactions with Glu500. If the molecule constructed
from the above fragments is not potent enough, pursuing further
interactions with a phenyl ring or other lipophilic ring systems in
the outer hydrophobic pocket might be beneficial. However, this
interaction may also deteriorate the selectivity over other kinases.
Functional groups in the solvent front area offer further extension
to the outside of ATP domain with water soluble residues.
Arranging these fragments together generated a molecule that
was unsuitable for fitting into the ATP domain. We required an
organizer motif which when substituted appropriately allowed
each fragment to fit precisely into the desired pockets. Considering
the tractability of the chemistry, we chose a 1,3,4-trisubstituted-
1H-pyrazole as the organizer motif.
Synthetic methods for building the designed compounds were
established and are summarized in Schemes 1–5.9 The two differ-
ent hinge binder pieces 2b and 3c used in this study were prepared
as boronic esters, exemplified in Schemes 1 and 2, and incorpo-
rated into the final molecule by a Suzuki coupling reaction. 4-Bro-
mo-1H-pyrrolo[2,3-b]pyridine 1 was prepared according to the
literature from commercially available reagents.10 The tosyl pro-
tecting group was incorporated using biphasic conditions to give
2a, which was transferred to the corresponding boronic ester 2b
in good yield.9 Ethyl 1H-pyrrolo[2,3-b]pyridine-2-carboxylate
7-oxide 3a was prepared as described in the literature,11 and was
readily brominated at the 4-position in good yield. Subsequent
reaction with bis(pinacolato)diboron under catalytic conditions
afforded the corresponding boronic ester 3c for further coupling
reaction.
The core consisting of the pyrazole organizer and a phenyl ring
at the 3-position was prepared as follows. Starting material 4a was
prepared according to the reported approach.12 The ethyl group
was easily installed by alkylation with ethyl iodide in the presence
of a base and phase transfer catalyst. The undesired isomer was ob-
tained as a minor byproduct and was readily removed from the
product 4b by silica gel chromatography. Reduction with Sn pow-
der gave the amine analogue 4c.
The final compounds were assembled using the chemistry indi-
cated below in Scheme 4. Suzuki coupling between compounds 4b
and 2b under microwave conditions afforded the adduct 5a. Subse-
quent Sn powder reduction gave 5b in moderate yield. The depro-
tection of 7-azaindole with aqueous NaOH at slightly elevated
temperature provided compound 5c, which reacted with isocya-
nates and isothiocyanates to give the final compounds 6a–l (Table
1).
Synthesis of the 3-bromo-7-azaindole derivative 5e required
deprotection of the tosyl group, followed by treatment with NBS
at room temperature as described in Scheme 5. Similarly, the final
compound 6m was obtained in the same fashion as described
above.
The final compounds with a substituent at the 2-position of
7-azaindole were achieved via
a Suzuki reaction between
compounds 4c and 3c, followed by treatment with 4-trif-
luoromethylphenylisocyanate (Scheme 6). The corresponding
carboxylic acid analogue 6o was obtained by saponification of 6n
with aqueous NaOH at reflux.
Thus, a novel series of 7-azaindole analogues 6a–6o were pre-
pared (Table 1). Compounds were tested against B-Raf in a fluores-
cence anisotropy binding assay.9 The parent phenyl urea
compound 6b (IC50 = 17.4 nM) showed ꢀ20-fold improved inhibi-
tory activity compared to its thiourea analogue 6a. Direct connec-
tion of an aromatic moiety (R3) to the urea proved to be optimal
based on the reduced inhibitory activities observed when the R3
substituent was cyclohexyl 6c or benzyl 6d. A survey of substitu-
tions was performed to find the optimal group for the induced fit
pocket (6d–6l). Mono-substituted phenyl rings (6f, 6h, 6i, 6l) and
di-substituted phenyl rings (6g, 6j, 6k) were generally very potent
with IC50 values ranging from 2.1 to 25.7 nM, regardless of the
positions of the substituents. However, by the incorporation of
an oversized substituent, such as when R3 was benzyloxyphenyl
(6e), a dramatic drop in activity was observed. Among all of the
O
O
Br
N
B
N
Br
N
a
b
N
S
N
S
N
H
O
O
O
O
1
2a
2b
Scheme 1. Reagents and conditions: (a) TsCl, Bu4NHSO4 (cat), aqueous NaOH,
CH2Cl2, rt, 30 min, 70%; (b) Pd(dppf)Cl2, bis(pinacolato)diboron, KOAc, DMF, 90 °C,
48 h, 77%.

4612
J. Tang et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4610–4614
Br
O
O
a
b
B
N
COOEt
COOEt
N
H
N
O
N
H
COOEt
N
N
H
3a
3b
3c
Scheme 2. Reagents and conditions: (a) TBAB, (MeSO2)2O, DMF, rt, 6 h, 77%; (b) Pd(dppf)Cl2, bis(pinacolato)diboron, KOAc, DMF, 90 °C, 48 h, 68%.
O
O
NH2
N
N
O
O
a
b
Br
Br
Br
N
N
N
N
N
N
H
4a
4b
4c
Scheme 3. Reagents and conditions: (a) EtI, TBAB, aq NaOH, CH2Cl2, rt, 1 h, 81%; (b) Sn powder, aq HCl, EtOH, reflux, 4 h, 95%.
NO2
NO2
O
O
B
N
b
a
N
+
Br
O
S
O
N
N
N
N
S
N
N
O
O
4b
2b
5a
H
N
H
N
NH2
NH2
R3
c
d
A
N
N
N
HN
O
S
N
N
N
N
N
N
HN
N
O
5b
5c
6
Scheme 4. Reagents and conditions: (a) Pd(PPh3)4, aq Na2CO3, DME, MW heating, 120 °C, 1 h, 88%; (b) Sn powder, aq HCl, EtOH, reflux, 4 h, 78%; (c) aq NaOH, MeOH, 40 °C,
1 h, 99%; (d) Isocyanates or isothiocyanate, pyridine, rt, 1 h.
NO2
NO2
NO2
a
b
N
N
N
O
S
O
N
N
N
N
N
N
N
HN
HN
Br
5a
5d
5e
NH2
H
N
H
N
c
d
O
N
N
N
N
N
N
HN
HN
Br
Br
5f
6m
Scheme 5. Reagents and conditions: (a) aq NaOH, MeOH, rt, 12 h, 89%; (b) NBS, THF, rt, 12 h; (c) Sn powder, aq HCl, EtOH, reflux, 12 h; (d) Phenylisocyanate, pyridine, rt, 1 h.

J. Tang et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4610–4614
4613
Table 1
B-Raf inhibitory activities of compounds 6a–6o
R3
A
N
NH
NH
N
HN
R1
N
R2
Compound
A
R1
R2
R3
B-Raf (FP, IC50) (nM)
6a
6b
6c
6d
6e
6f
6g
6h
6i
6j
6k
6l
6m
6n
6o
S
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Br
H
H
Phenyl
Phenyl
Cyclohexyl
Phenyl
4-Benzyloxyphenyl
2-Chlorophenyl
2,6-Difluorophenyl
4-Trifluoromethylphenyl
3-Chlorophenyl
2-Fluoro-5-trifluoromethylphenyl
4-Chloro-3-trifluoromethylphenyl
3-Methoxylphenyl
Phenyl
331
17.4
41.7
170
110
5.3
25.7
2.5
10.5
3.9
5.4
2.1
79.4
33.9
45.7
O
O
O
O
O
O
O
O
O
O
O
O
O
O
H
Ethyl formate
Formic acid
Phenyl
Phenyl
lipophilic terminal substituted phenyl rings examined, compound
6h13 with a trifluoromethyl group at the 40-position provided one
of the best inhibitory activities (IC50 = 2.5 nM) against B-Raf. We
chose 6h over 6l as the exemplar for this series due to a superior
p450 inhibition profile.14 Evaluation of this compound in a cellular
assay through the measurement of phosphorylation of MEK1 dem-
onstrated its excellent potency (IC50 = 63 nM).9 Substituent R2 may
be affected by steric interactions as evidenced by the drop in inhib-
itory activity for 6 m containing a bromide at the 3-position of
7-azaindole. A substituent at the 2-position of 7-azaindole such
as ethyl formate (6n) or formic acid (6o) was tolerated. Although
these substituents did not add potency value, they could poten-
tially be used to improve the physical and chemical properties.
Illustrated in Fig. 3 is a docking model of 6h complexed with an
inactive conformation of B-Raf.15 From this docking analysis, it is
noted that 7-azaindole binds to the kinase at the hinge region with
its nitrogen and hydrogen atoms forming H-bond interactions with
the backbone NH and CO of Cys531.16 The pyrazole ring resides at
the sugar pocket, and the connected phenyl at the 3-position is posi-
tioned into the inner hydrophobic pocket. The urea tethered to the
meta position of the central phenyl ring is positioned deep into
the induced fit pocket, surrounded by mostly lipophilic residues
including Leu504, Val503, Leu566, and His573. The urea moiety ap-
pears to form interactions with Glu500. Based on this docking anal-
ysis, substituents at the 2-position of the 7-azaindole scaffold, such
as ethyl formate, would reach the outer hydrophobic pocket and
would be expected to be well tolerated despite the dissimilarity of
the ethyl formate moiety to the preferred lipophilic ring systems.
The compounds described here were generally about 10-fold
more potent as inhibitors of B-Raf than other kinases evaluated.
NH2
NH2
O
O
a
B
N
N
+
N
HN
N
COOEt
Br
N
N
H
N
EtOOC
4c
3c
5g
H
N
H
N
H
N
H
N
O
O
b
c
CF3
CF3
N
N
N
N
HN
EtOOC
HN
HOOC
N
N
6n
6o
Scheme 6. Reagents and conditions: (a) Pd(PPh3)4, aq Na2CO3, DME, MW heating, 120 °C, 1 h, 17%; (b) 4-trifluoromethylphenylisocyanate, pyridine, rt, 1 h; (c) 6 N NaOH,
MeOH, reflux, 1 h.

4614
J. Tang et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4610–4614
5. Schreck, R.; Rapp, U. R. Intl. J. Cancer 2006, 119, 2261.
6. Li, N.; Batt, D.; Warmuth, M. Curr. Opin. Investig. Drugs 2007, 8, 452.
7. (a) King, A. J.; Patrick, D. R.; Batorsky, R. S.; Ho, M. L.; Do, H. T.; Zhang, S. Y.;
Kumar, R.; Rusnak, D. W.; Takle, A. K.; Wilson, D. M.; Hugger, E.; Wang, L.;
Karreth, F.; Lougheed, J. C.; Lee, J.; Chau, D.; Stout, T. J.; May, E. W.; Rominger, C.
M.; Schaber, M. D.; Luo, L.; Lakdawala, A. S.; Adams, J. L.; Contractor, R. G.;
Smalley, K. S. M.; Herlyn, M.; Morrissey, M. M.; Tuveson, D. A.; Huang, P. S.
Cancer Res. 2006, 66, 11100; (b) Takle, A. K.; Brown,M.-J. B. ; Davies, S.; Dean, D.
K.; Francis, G.; Gaiba, A.; Hird, A. W.; King, F. D.; Lovell, P. J.; Naylor, A.; Reith, A.
D.; Steadman, J. G.; Wilson, D. M. Bioorg. Med. Chem. 2006, 16, 378.
8. Four similarly substituted compounds with different hinge binders are shown
below. The 7-azaindole hinge binder provides the most potent inhibitory
activities against 4 kinases. Potency is represented by pIC50 value.
O
S
O
S
R
R
N
H
N
H
Figure 3. Homology model of compound 6h bound to inactive form of B-Raf.
Inhibitor atoms colored as follows: C, gray; N, blue; O, red.
N
N
N
AKT1 FL
Aurora A
PKA
5.8
6.8
6.6
8.4
AKT1 FL < 5.5
Aurora A
PKA
6.1
5.1
6.5
ROCK1
ROCK1
O
S
O
S
R
R
N
N
H
H
N
N
NH2
N
N
H
Figure 4. Kinase selectivity profile of compound 6h. Potency is represented by
pIC50 on the Y axis. ꢁ denotes no activity at maximum concentration (30
lM) tested.
AKT1 FL
Aurora A
5.6
6.7
AKT1 FL
Aurora A
7.3
8.0
PKA
7.3
PKA
8.0
ROCK1
8.6
ROCK1
8.8
For example, compound 6h bearing a trifluoromethyl group at the
40-position, which was proposed to fit the induced fit pocket very
well, showed more than 1000-fold selectivity against a variety of
other kinases screened, and almost 100-fold selectivity against
VEGFR2 and c-Met (Fig. 4).
In conclusion, we discovered a series of novel, potent B-Raf
inhibitors using knowledge-based design. As exemplified by com-
pound 6h, we achieved not only excellent potency in both enzyme
and cellular assays, but also an outstanding selectivity profile
against other kinases.
9. Tang, J.; Nakano, M.; Hamajima, T. PCT Int. Application, WO2007090141.
Enzyme and cellular assay description can be found in this patent. Cellular
assays: B-Raf mediated phosphorylation of MEK1 was measured in the cellular
assay. Expression constructs for B-Raf and FLAG-tagged MEK1 were co-
transfected in 3T3 cells and gene expression was induced using the
GeneSwitch (TM) system for inducible mammalian expression (Invitrogen).
Four hours following the induction of expression of B-Raf and MEK1, cells were
exposed to the test compounds for 2 h. The cells were lysed, and then an
immunoassay was performed using anti-phospho-MEK1/2 (Cell Signaling
Technologies) to detect the percent inhibition of MEK1 phosphorylation.
10. Thibault, C.; L’Heureux, A.; Bhide, R. S.; Ruel, R. Organic Lett. 2003, 5, 5023.
11. Adams, D. R.; Bentley, J. M.; Davidson, J.; Duncton, M. A. J.; Porter, R. H. P. PCT
Int. Application, WO2000044753.
Acknowledgments
12. Yoshida, I.; Yoneda, N.; Ohashi, Y.; Suzuki, S.; Miyamoto, M.; Miyazaki, F.;
Seshimo, H.; Kamata, J.; Takase, Y.; Shirato, M.; Shimokubo, D.; Sakuma, Y.;
Yokohama, H. PCT Int. Application, WO2002088107.
We are grateful to the enzymologists for contributions to the
screening work and Hiroshi Sootome for the evaluation of cellular
potency.
13. Analytic data of potent inhibitors 6h: 1H NMR (400 MHz, DMSO-d6) ppm 1.49
(t, 3H, J = 7.3 Hz), 4.26 (q, 2H, J = 7.3 Hz), 6.26 (dd, 1H, J = 2.0, 3.3 Hz), 6.81 (d,
1H, J = 5.1 Hz), 6.90–6.95 (m, 1H), 7.18 (dd, 1H, J = 7.8, 8.0 Hz), 7.39 (dd, 1H,
J = 2.5, 3.3 Hz), 7.47 (ddd, 1H, J = 1.0, 2.0, 8.1 Hz), 7.51–7.54 (m, 1H), 7.59–7.65
(m, 4H), 8.09 (d, 1H, J = 5.1 Hz), 8.18 (s, 1H), 8.80 (s, 1H), 9.00 (s, 1H), 11.62 (brs,
1H). LC/MS: m/z 491 (M+1)+.
References
1. (a) Hopkins, A. L.; Groom, C. R. Nat. Rev. 2002, 1, 727; (b) Melnikova, I.; Golden,
J. Nat. Rev. Drug Discov. 2004, 3, 993.
14. Compound 6h: 2C9 (pIC50 = 5.9); 2D6 (pIC50 = 5.3). Compound 6l: 2C9
(pIC50 = 6.5); 2D6 (pIC50 = 6.7).
15. Model built from X-ray crystal structure PDB ID 1UWH with MacroModel,
Schrödinger LLC.
16. This was numbered as Cys532 in some recent papers, for example: Tsai,
J.; Lee, J. T.; Wang, W.; Zhang, J.; Cho, H.; Mamo, S.; Bremer, R.; Gillette,
S.; Kong, J.; Haass, N. K.; Sproesser, K.; Li, L.; Smalley, K. S. M.; Fong, D.;
Zhu, Y.-L.; Marimuthu, A.; Nguyen, H.; Lam, B.; Liu, J.; Cheung, I.; Rice, J.;
Suzuki, Y.; Luu, C.; Settachatgul, C.; Shellooe, R.; Cantwell, J.; Kim, S.-H.;
Schlessinger, J.; Zhang, K. Y. J.; West, B. L.; Powell, B.; Habets, G.; Zhang,
C.; Ibrahim, P. N.; Hirth, P.; Artis, D. R.; Herlyn, M.; Bollag, G. Proc. Natl.
Acad. Sci. USA 2008, 105, 3041.
2. (a) Nie, Z.; Perretta, C.; Erickson, P.; Margosiak, S.; Almassy, R.; Lu, J.; Averill, A.;
Yager, K. M.; Chu, S. Bioorg. Med. Chem. Lett. 2007, 17, 4191; (b) Tao, Z.-F.; Wang,
L.; Stewart, K. D.; Chen, Z.; Gu, W.; Bui, M.-H.; Merta, P.; Zhang, H.; Kovar, P.;
Johnson, E.; Park, C.; Judge, R.; Rosenberg, S.; Sowin, T.; Lin, N.-H. J. Med. Chem.
2007, 50, 1514; (c) Collins, I.; Caldwell, J.; Fonseca, T.; Donald, A.; Bavetsias, V.;
Hunter, L.-J. K.; Garrett, M. D.; Rowlands, M. G.; Aherne, G. W.; Davies, T. G.;
Berdini, V.; Woodhead, S. J.; Davis, D.; Seavers, L. C. A.; Wyatt, P. G.; Workman,
P.; McDonald, E. Bioorg. Med. Chem. 2006, 14, 1255.
3. (a) McGregor, M. J. J. Chem. Inf. Model 2007, 47, 2374; (b). J. Med. Chem. 2007, 50,
409.
4. (a) Liu, Y.; Gray, N. S. Nat. Chem. Biol. 2006, 2, 358; (b) Southall, N. T.; Ajay J.
Med. Chem. 2006, 49, 2103.