7-[1H-Indol-2-yl]-2,3-dihydro-isoindol-1-ones as dual Aurora-A/VEGF-R2 kinase inhibitors: Design, synthesis, and biological activity

Article abstract of DOI:10.1016/j.bmcl.2008.07.090

A novel series of 7-[1H-indol-2-yl]-2,3-dihydro-isoindol-1-ones designed to be inhibitors of VEGF-R2 kinase was synthesized and found to potently inhibit VEGF-R2 and Aurora-A kinases. The structure-based design, synthesis, and initial SAR of the series are discussed.

Full text of DOI:10.1016/j.bmcl.2008.07.090

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5130–5133
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
7-[1H-Indol-2-yl]-2,3-dihydro-isoindol-1-ones as dual Aurora-A/VEGF-R2
kinase inhibitors: Design, synthesis, and biological activity
*
Terry V. Hughes , Stuart L. Emanuel, Harold R. O’Grady, Peter J. Connolly, Catherine Rugg,
Angel R. Fuentes-Pesquera, Prabha Karnachi, Richard Alexander, Steven A. Middleton
Johnson & Johnson Pharmaceutical Research and Development, L.L.C, PO Box 300, 1000 Route 202, Raritan, NJ 08869, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of 7-[1H-indol-2-yl]-2,3-dihydro-isoindol-1-ones designed to be inhibitors of VEGF-R2
kinase was synthesized and found to potently inhibit VEGF-R2 and Aurora-A kinases. The structure-based
design, synthesis, and initial SAR of the series are discussed.
Received 19 May 2008
Accepted 22 July 2008
Available online 26 July 2008
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
Cancer
Aurora-A
Aurora-B
Aurora-C
Kinase
Vascular endothelial
growth factor receptor 2
VEGF-R2
Kinase inhibition
Isoindolinone
2,3-Dihydro-isoindol-1-ones
H
O
N
Vascular endothelial growth factor (VEGF) is an important reg-
ulator of vascular growth and permeability. Overexpression of
VEGF in solid tumors has been shown to promote angiogenesis,
thereby facilitating tumor growth. VEGF acts on VEGF-R2, a recep-
tor tyrosine kinase present on the surface of endothelial cells. The
inhibition of VEGF-R2 by small molecules is a validated therapeutic
approach for treating cancer and has received much attention in
the literature.1 We have concentrated our efforts to find novel
inhibitors of the VEGF receptor kinases. By coupling the computer
modeling of virtual compounds with traditional medicinal chemis-
try principles, we designed a new kinase inhibitor scaffold 1 based
on known kinase inhibitors (Fig. 1).2 Initial modeling of our scaf-
fold using a homology structure of VEGF-R2 overlaid with known
VEGF-R2 inhibitors showed that key binding interactions were
maintained. Kinase screening of early examples of this series
revealed that the scaffold displayed inhibition of VEGF-R2 kinase
and unexpected inhibition of Aurora-A kinase.
H
N
H
N
O
H
N
H
N
Et2N
F
O
1
SU-11248 (sunitinib)
VEFG-R2 inhibitor
Figure 1. Pharmacophore 1 and known VEGF-R2 inhibitor SU-11248 (sunitinib).
The dashed lines depict intramolecular hydrogen bonds.
of the centrosome and Aurora-B regulates chromosomal
movement and cytokinesis. The biological function of Aurora-C is
still not understood. The inhibition of Aurora kinases by a small
molecule has been reported to induce apoptosis in a variety of can-
cer cell lines and suppress tumor growth in vivo.3,4 The dual inhi-
bition of VEGF-R2 and Aurora-A kinases is an attractive compound
profile and may provide for enhanced antitumor activity toward a
wide range of cancers.
In this letter, scaffold design, synthesis, and preliminary struc-
ture–activity relationships for the series are presented. Modeling
of our scaffold using a homology structure of Aurora-A, and an
X-ray structure of an analog in Aurora-A, have added to the under-
standing of the unexpected activity against Aurora-A, thereby
helping to direct our future synthetic efforts.
The Aurora kinases (Aurora-A, Aurora-B, and Aurora-C) are ser-
ine/threonine kinases that are essential for mitotic progression.
Aurora-A kinase plays a role in spindle formation and organization
* Corresponding author. Tel.: +1 610 270 6561; fax: +610 270 6609.
E-mail address: hughe007@hotmail.com (T.V. Hughes).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.07.090

T. V. Hughes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5130–5133
5131
We observed that the VEGF-R2 kinase inhibitor SU-11248 (sun-
H
N
H
N
O
1 O
itinib) as well as other potent kinase inhibitors were capable of
forming a pseudocycle via an intramolecular hydrogen bond (Fig.
1).2,5,6 We felt that the intramolecular hydrogen bonding in these
compounds held them in a flat, planar conformation that would
facilitate binding to the hinge region in the active site of protein
kinases. With this hypothesis in mind we proposed the novel
pharmacophore 1 as a scaffold for the inhibition of protein kinases.
Molecular modeling confirmed that in the minimum energy
conformation, the pyrrole NH of pharmacophore 1 is well situated
to participate in an intramolecular hydrogen bond with the
isoindolinone carbonyl oxygen (Fig. 1).
Boc
N
R
N
a
c
2
R
R
6 (R = H)
10 (R = H or OR2, R1 = Boc)
11 (R = H or OR2, R1 = H)
7 (R = OBn)
8 (R = OH)
b
d
Scheme 2. Reagents and conditions: (a) 1-Boc-indole-2-boronic acid or 5-benzyl-
oxy-1-Boc-indole-2-boronic acid, Pd(dppf)Cl2, 2 N Na2CO3, THF (70–80%); (b) H2
(balloon), 10% Pd/C, THF/MeOH, (100%); (c) Cs2CO3, DMF, 50 °C, R2Cl, (65–83%); (d)
TFA, DCM (95%) or 185 °C, neat (85–100%).
The synthesis of compound 1 is shown in Scheme 1. Suzuki cou-
pling of the known7 7-bromo-2,3-dihydro-isoindol-1-one (2) with
commercially available 1-(tert-butoxycarbonyl)pyrrole-2-boronic
acid afforded the Boc-protected core 3. Thermolytic cleavage of
the Boc-protecting group8 cleanly afforded the proposed pharma-
cophore 1. Additionally, the isoindolinone 3 was N-methylated
with methyl iodide to give compound 4. Subsequent thermal
deprotection of 4 afforded the N-methylated analog 5. Kinase data
for these analogs are shown in Table 1.
of the Boc group was achieved by either thermolytic cleavage or
under acidic conditions using TFA in DCM at 25 °C to give the ana-
logs 11. Various indolyl analogs (12c, 12d, 12f–12j) were prepared
using this general route (Scheme 2). Analogs 12a, 12b, and 12e
were synthesized from 12c via an oxidation/reductive-amination
sequence. The kinase data for representative analogs are summa-
rized in Table 2.
Changing the pyrrole ring of pharmacophore 1 to an indole ring
as in 12i provided 10-fold better VEGF-R2 potency (IC50 = 144 nM)
and a 3-fold increase in Aurora-A potency (IC50 = 100 nM). Molec-
ular modeling suggested that substitution at the 5-position of the
indole ring should be allowed as that position is solvent-exposed.
Extension of the series through alkylation of the phenol 12h affor-
ded very potent inhibitors (12a–12f) of VEGF-R2 and Aurora-A.
Generally, the VEGF-R2 potency tracked with the Aurora-A
potency, with the most potent analogs having a pendant amine
tethered to the indole ring by an alkyloxy linker. The length of
the linker, either 2-carbon (12d) or 3-carbon (12e), did not have
a significant effect on potency. All analogs of 12 were inactive
Gratifyingly, our proposed core 1 displayed promising activity
against VEGF-R2 kinase (IC50 = 1.10
lM), and good activity against
Aurora-A kinase (IC50 = 0.338 M). It was established that the NH
l
of the isoindolinone ring was important for kinase inhibition since
the N-methylated analog 5 displayed very weak kinase activity.
Additionally, the NH of the pyrrole ring was also found to be
important for kinase activity since the N-Boc analog 3 was also
only weakly active. At this point we considered 1 to be a solid
hit for VEGF-R2 and Aurora-A kinase inhibition with surprisingly
good potency, low molecular weight (198), and a straightforward
synthesis. We then explored the SAR for the left hand pyrrole side
of the pharmacophore 1. The synthesis of a representative set of
analogs is shown in Scheme 2.
The Suzuki coupling of 2 with 1-Boc-indole-2-boronic acid
afforded 6 in good yield when Pd(dppf)Cl2 was used as the catalyst.
Similarly, coupling of 2 with 5-benzyloxy-1-Boc-indole-2-boronic
acid afforded 7. Removal of the benzyl group of 7 proceeded
smoothly under mild hydrogenolysis conditions to yield the phenol
8. Alkylation of the phenol with alkyl chlorides and Cs2CO3 in DMF
at 50 °C occurred to give protected intermediates 10. Final removal
against cyclin-dependent kinase—1 (CDK1, IC50 > 100 lM).
To continue variations of the left side of pharmacophore 1 we
synthesized the carboxylic acid intermediate 18 (Scheme 3). Reac-
tion of the commercially available 3-methyl phthalic anhydride
(13) with MeOH and H2SO4 afforded an isomeric mixture of phtha-
lic acid mono-esters. The isomeric mixture was converted to the
bis methyl ester 14 with TMS diazomethane in toluene/MeOH.
Treatment of 14 with NBS in CCl4 using AIBN as a catalyst afforded
the benzyl bromide 15. Direct conversion of 15 to the isoindoli-
Table 2
H
N
Me
N
VEGF-R2, Aurora-A, and CDK1 kinase activity for 12
H
N
O
O
O
H
R
N
R
N
N
O
a
b
H
N
Br
2
3 (R = Boc)
1 (R = H)
4 (R = Boc)
5 (R = H)
c
c
R
12
a
a
a
Compound R
VEGF-R2 IC50 Aurora-A IC50 CDK1 IC50
M) M) M)
Scheme 1. Reagents and conditions: (a) Pd(OAc)2, 2 N Na2CO3, P(o-tolyl)3, 1-(tert-
Butoxycarbonyl)pyrrole-2-boronic acid, DMF, (17%); (b) THF, NaH, MeI, (23%), (c)
N2, 185 °C, (73%).
(
l
(l
(l
12a
12b
O(CH2)34-ethyl-
piperazin-1-yl
O(CH2)34-hydroxy-
piperidin-1-yl
0.014
0.044
>100
0.018
0.061
>100
Table 1
VEGF-R2, Aurora-A, and CDK1 kinase activity for 1, 3–5
12c
12d
12e
12f
12g
12h
12i
O(CH2)3OH
O(CH2)2piperidin-1-yl
O(CH2)3piperidin-1-yl
O(CH2)2morpholin-4-yl 0.074
OMe
0.030
0.056
0.065
0.100
0.200
0.179
0.135
0.070
0.090
0.100
21.23
>100
>100
>100
>100
>100
13.3
a
a
a
Compound
VEGF-R2 IC50
(
lM)
Aurora-A IC50
(lM)
CDK1 IC50 (lM)
1
3
4
5
1.10
0.338
(25%)b
>100
(63%)b
(<10%)b
>100
(46%)b
>100
0.111
0.119
0.144
11.26
OH
H
(<10%)b
(37%)b
(24%)b
>100
>100
12j
OCH2C6H5
a
IC50 data are the average of at least two separate experiments. IC50 values listed
as >100 indicate no observed 50% inhibition at the highest dose tested, nor was an
a
IC50 data are the average of at least two separate experiments. IC50 values listed
inhibition maximum observed.
as >10 or >100 indicate no observed 50% inhibition at the highest dose tested, nor
was an inhibition maximum observed.
b
Values in parentheses indicate % inhibition at 100 lM.

5132
T. V. Hughes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5130–5133
HN
O
CH3
R
O
O
CO2H
NH
O
MeO2C
MeO2C
O
e
a,b
I
CO2H
O
I
b
O
O
a
I
RO2C
O
13
14 R = CH3
15 R = CH2Br
16 R = CH2N3
20
17 R = CH3
18 R = H
22
c
21
f
d
Scheme 4. Reagents and conditions: (a) Ac2O, reflux, (96%); (b) NH2CONH2, (85%).
Scheme 3. Reagents and conditions: (a) MeOH, H2SO4, reflux; (b) TMSCHN2, (95%);
(c) NBS, AIBN, CCl4, (80%); (d) NaN3, DMF, 25 °C, (88%); (e) PPh3, THF, H2O, 25 °C,
(52%); (f) LiOH, THF, MeOH, H2O, 25 °C, (83%).
phosphorylation on Serine 10 by Aurora B kinase is thought to be
crucial for proper chromosomal alignment and cytokinesis. The
inhibition of Aurora B kinase has been shown to reduce phosphor-
ylation of H3 on Serine 10. Compound 12d was shown to inhibit
none 17 with concentrated NH4OH in THF was problematic and
afforded the amide 19d due to unexpected reactivity of the aryl
methylester 17 to concd NH4OH.9 This was overcome by first con-
verting 15 into the benzyl azide 16 by treatment with sodium
azide in DMF. The benzyl azide 16 was then reduced in situ to
the benzyl amine, which cyclized to afford the isoindoline 17. Stan-
dard hydrolysis of the ester group of 17 using LiOH in THF/MeOH/
H2O afforded the desired carboxylic acid 18.
the phosphorylation of histone H3 at 2.5 lM (Fig. 2).
Polyploidy, the accumulation of cells with >4 N DNA content,
and multinucleation are an expected phenotypic result of Aurora
kinase inhibition.10 Treatment of HeLa cells with 2.5
lM of 12d
for 72 h induced a large number of cells to endoreduplicate result-
ing in a significant population of cells with a greater than 4 N DNA
content (Fig. 3).
Starting from carboxylic acid 18, series of analogs 19 were pre-
pared (kinase data are shown in Table 3). The reaction of carboxylic
acid 18 with 2-amino-6-methoxyaniline in refluxing 6N HCl affor-
ded the benzimidazole 19a. The kinase potencies of the 2-substi-
tuted benzimidazole analog 19a and the 2-substituted indole
analog 12g were similar. Peptide coupling of 18 with 4-methoxy-
aniline under standard coupling conditions afforded 19b. We were
surprised that 19b, an isostere of the potent indole 12g, displayed
To evaluate the ability of this class of compounds to effect
tumor suppression in vivo we performed tumor xenograft studies
using the A375 cell line in nude mice. When 12d was dosed
at 100 mg/kg ip, a 46% inhibition of tumor growth compared
to ip vehicle control (P = 0.08) was observed (Fig. 4). No statistically
significant inhibition was observed when 12d was dosed
100 mg/kg po.
lower potency against VEGF-R2 (IC50 = 5.94
ity against Aurora-A (45% inhibition at 100
l
M) and reduced activ-
lM). Changing the con-
p-H3 (Ser 10)
-actin
nection on the indole ring to the 3-position, analog 19c, resulted in
a significant reduction of VEGF-R2 and Aurora-A inhibitory activ-
ity. The 3-substituted indole analog 19c was synthesized from
the corresponding boronic acid as detailed in Scheme 2. The imide
analog 19g, prepared from 4-iodopthalimide (22, synthesized as
shown in Scheme 4) and 5-methoxy-1H-indol-2-boronic acid,
Nocodazole
+ DMSO
Nocodazole +
12d
was active at 0.1 lM but did not exhibit greater inhibition at high-
er concentrations, indicative of poor solubility under assay
conditions.
Figure 2. The level of p-H3 (Ser10) in lysates of HeLa cells synchronized in G2/M
with Nocodazole and then treated for an additional 8 h with either Nocodazole and
DMSO or Nocodazole and compound 12d.
With the initial SAR beginning to take shape we evaluated the
ability of some of the most promising analogs to inhibit aurora ki-
nase in cells. Histone H3 is a direct downstream target of the Aur-
ora kinases. It is involved in chromosomal condensation and
Table 3
VEGF-R2 and Aurora-A kinase activity for 19
H
N
R1
O
R
19
Compound
R
R1
VEGF-R2
IC50 (lM) IC50 (lM)
Aurora-A
a
a
19a
19b
19c
19d
19e
19f
5-methoxy-1H-benzoimidazol-2-yl CH2
0.142
5.94
0.205
CONH(4-methoxy-phenyl)
1H-Indol-3-yl
CONH2
CO2Me
CO2H
CH2
CH2
CH2
CH2
CH2
(45%)b
(60%)b
14.34
>100
(50%)b
>100
>100
>100
>100
19g
5-Methoxy-1H-indol-2-yl
C@O (35%)c
(30%)b
a
IC50 data are the average of at least two separate experiments. IC50 values listed
as >100 indicate no observed 50% inhibition at the highest dose tested, nor was an
inhibition maximum observed.
b
Values in parentheses indicate % inhibition at 100
l
M.
M.
Figure 3. Polyploidy of HeLa cells upon treatment with 2.5 lM of 12d for 72 h as
determined by FACS anaylsis.
c
Values in parentheses indicate % inhibition at 0.1
l

T. V. Hughes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5130–5133
5133
of ATP. The backbone carbonyl of Glu-211 and the NH of Ala-
213 in the hinge region form the characteristic dual H-bonds
with 12f. Additionally, the indole NH of compound 12f forms
an intramolecular H-bond with the C@O of the isoindolonone.
The morpholine group is solvent-exposed and this region may
provide a handle for modifying the physicochemical properties
of the series.12
In summary, we have described the design, synthesis, and initial
SAR development of 7-[1H-indol-2-yl]-2,3-dihydro-isoindol-1-
ones as potent dual VEGF-R2 and Aurora-A kinase inhibitors that
are orally available and display in vivo activity in a tumor xeno-
graft model. The design of the progenitor of this series, pharmaco-
phore 1, was based on the hypothesis that an intramolecular
hydrogen bond between the indole NH and the isoindolinone car-
bonyl O would hold the structure in a relatively flat conformation.
Molecular modeling and a crystal structure of analog 12f in Aur-
ora-A kinase supported this type of intramolecular hydrogen bond
and showed that 12f binds in the ATP binding site. The dual activ-
ity of this series of kinase inhibitors provides a unique profile that
may prove very potent for blocking the growth of cancer cells as it
attacks two important pathways that drive proliferation of almost
all cancers.
1200
900
600
300
0
Control
Cpd 12d (100 mg/kg IP)
Cpd 12d (100 mg/kg PO)
0
5
10
15
Day
Figure 4. Tumor inhibition of 12d in a A375 tumor xenograft model in nude mice.
-j- vehicle control, -.- 12d 100 mg/kg po, -ꢀ- 100 mg/kg ip.
References and notes
1. (a) Folkman, J. N. Engl. J. Med. 1995, 333, 1757; (b) Fong, T. A. T.; Shawver, L. K.;
Tang, C.; App, H.; Powell, T. J.; Kim, Y. H.; Schreck, R.; Wang, X.; Risau, W.;
Ullrich, A., et al Cancer Res. 1999, 59, 99; (c) Wood, J. M.; Bold, G.; Buchdunger,
E.; Cozens, R.; Ferrari, S.; Frei, J.; Hofmann, F.; Mestan, J.; Mett, H.; O’Reilly, T.,
et al Cancer Res. 2000, 60, 2178; (d) Wedge, S. R.; Ogilvie, D. J.; Dukes, M.;
Kendrew, J.; Chester, R.; Jackson, J. A.; Boffey, S. J.; Valentine, P. J.; Curwen, J. O.;
Musgrove, H. L., et al Cancer Res. 2002, 62, 4645; (e) Cardones, A. R.; Banez, L. L.
Curr. Pharm. Des. 2006, 12, 387.
2. Sun, Li.; Liang, C.; Shirazian, S.; Zhou, Y.; Miller, T.; Cui, J.; Fukuda, J. Y.; Chu, J.;
Nematalla, A.; Wang, X.; Chen, H.; Sistla, A.; Luu, T. C.; Tang, F.; Wei, J.; Tang, C.
J. Med. Chem. 2003, 1116.
3. Harrington, E. A.; Bebbington, D.; Moore, J.; Rasmussen, R. K.; Ajose-Adeogun,
A. O.; Nakayama, T.; Graham, J. A.; Demure, C.; Hercend, T.; Diu-Hercend, A.; Su,
M.; Golec, J. M. C.; Miller, K. M. Nat. Med. 2004, 262.
4. Keen, N.; Taylor, S. Nat. Rev. 2004, 927.
5. Sepp-Lorenzino, L.; Rands, E.; Mao, X.; Connolly, B.; Shipman, J.; Antanavage, J.;
Hill, S.; Davis, L.; Beck, S.; Rickert, K.; Coll, K.; Ciecko, P.; Fraley, M.; Hoffman,
W.; Hartman, G.; Heimbrook, D.; Gibbs, J.; Kohl, N.; Thomas, K. Cancer Res.
2004, 751.
6. Liu, J.; Dermatakis, A.; Lukacs, C.; Konzelmann, F.; Chen, Y.; Kammlott, U.;
Depinto, W.; Yang, H.; Yin, X.; Chen, Y.; Schutt, A.; Simcox, M. E.; Luk, K. Bioorg.
Med. Chem. Lett. 2003, 2465.
Figure 5. Model of compound 12f docked into the ATP binding site of human
Aurora-A kinase. The protein is depicted as a ribbon structure with backbone atoms
of key residues 210–214 shown explicitly. Amino acids 138, 139, and 147–150 are
hidden for clarity. Potential hydrogen bonds are shown in green.
7. Rupert, K. C.; Dodd, J. H.; Henry, J. R. Heterocycles 1997, 2217.
8. Rawal, V. H.; Cava, M. P. Tetrahedron Lett. 1985, 6141.
9. Complete conversion of 17 to 19d was observed when 17 was treated with
NH4OH/THF for 30 min at 25 °C.
10. (a) Ditchfield, C.; Johnson, V. L.; Tighe, A.; Ellston, R.; Haworth, C.; Johnson, T.;
Mortlock, A.; Keen, N.; Taylor, S. S. J. Cell Biol. 2003, 267; (b) Harrington, E. A.;
Bebbington, D.; Moore, J.; Rasmussen, R. K.; Ajose-Adeogun, A. O.; Nakayama,
T.; Graham, J. A.; Demur, C.; Hercend, T.; Diu-Hercend, A.; Su, M.; Golec, J. M. C.;
Miller, K. M. Nat. Med. 2004, 262.
Compound 12d has a short half-life but is orally available in rat
(po dose 10 mg/kg, F = 20%, Cmax = 391 ng/mL, t1/2 = 0.9 h). When
12d was dosed in nude mice at 10 mg/kg drug plasma levels of
154, 160, 41, and 12 ng/mL were observed for 0.5, 1, 2, and 4 h
timepoints, respectively.
To investigate potential binding modes of compound 12f, an
energy-minimized structure was docked into the ATP-binding
site of a published crystal structure of human Aurora-A kinase
(Fig. 5).11 In this model, the isoindolinone portion of 12f occu-
pies the same space in the active site as the amino pyrimidine
11. Fancelli, D.; Berta, D.; Bindi, S.; Cameron, A.; Cappella, P.; Carpinelli, P.; et al. J.
Med. Chem. 2005, 48, 3080; RCSB Protein Data Bank ID: 2BMC, http://
www.pdb.org/.
12.
A low-resolution (3.1 Å) crystal structure of compound 12f with murine
Aurora-A kinase was obtained, confirming the modeling results (data not
shown).