Discovery of protein-protein binding disruptors using multi-component condensations small molecules

Article abstract of DOI:10.1016/j.bmcl.2006.05.014

A series of small molecule compounds interfering with the binding process of VEGF and NRP1 has been identified and further optimized. Full synthetic details as well as SAR are reported which demonstrate that expeditious MCC-based syntheses may lead to val

Full text of DOI:10.1016/j.bmcl.2006.05.014

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3998–4001
Discovery of protein–protein binding disruptors using
multi-component condensations small molecules
*
´
´
Karim Bedjeguelal, Hugues Bienayme, Antoine Dumoulin, Stephane Poigny,
Philippe Schmitt and Eric Tam
Pierre Fabre Urologie, 11 Av. Albert Einstein, 69100 Villeurbanne, France
Received 28 March 2006; revised 21 April 2006; accepted 4 May 2006
Available online 9 June 2006
Abstract—A series of small molecule compounds interfering with the binding process of VEGF and NRP1 has been identified and
further optimized. Full synthetic details as well as SAR are reported which demonstrate that expeditious MCC-based syntheses may
lead to valuable molecules addressing challenging targets such as protein–protein interactions. Preliminary functional assay data
confirm that these compounds may be further developed toward drug candidates.
Ó 2006 Elsevier Ltd. All rights reserved.
Vascular endothelial growth factor (VEGF) has been
described as a cytokine triggering one of the major path-
ways controlling angiogenesis. This growth factor,
which is expressed as two major isoforms, VEGF165
and VEGF121, binds several tyrosine kinase receptors
like VEGFR2/KDR or FLT1. Neuropilin-1 (NRP1), a
non-tyrosine kinase receptor, has been recently identi-
fied as a VEGF165 co-receptor. Coexpression of
NRP1 and VEGFR2 in endothelial cells (EC) promotes
VEGF165 binding to VEGFR2 leading to EC migration
and proliferation, compared with EC expressing
VEGFR2 alone.1 NRP1 has also been associated with
neuronal guidance2 and epithelial cell migration. In a
prior study,3 NRP1 has been validated as a target of
interest while addressing prostate cancer, showing that
this co-receptor is involved in migration and prolifera-
tion of tumor cell lines. Prevention of VEGF–NRP1
hetero-dimerization by a small molecule appeared there-
fore to us as both a promising and a challenging task of
relevance in the pursuit of novel prostate cancer
treatments.4
example have been proposed to inhibit the p53–
HDM2 interaction which has also proved to be of phar-
maceutical relevance while developing new cancer
cures.5 In a similar approach, the Bcl-XL hetero-dimer
has also been reported as a target for small molecule
disruptors.6,7
Multi-component condensations (MCCs) are among the
most useful chemical reactions known in the context of
drug discovery. Because they combine two important
attributes of synthetic efficiency, convergence and atom
economy, they were naturally recognized and used in
combinatorial chemistry.8 One aspect of MCCs that
have been overlooked in the past is their capacity to cre-
ate structurally complex molecules in a single chemical
step.9 This property is of obvious interest in the context
of protein–protein interactions, where structurally elab-
orate natural products have a demonstrated track-
record.
In this report are described for the first time to our
knowledge the use of MCCs for the discovery of small
molecule antagonists of a protein–protein interaction
between VEGF and NRP1, and the identification of
potent and functional antagonists.
Several examples of protein–protein binding disruption
by small molecules have already been reported proving
that this approach constitutes a viable therapeutic
strategy. Nutlins and 1,4-benzodiazepine-2,5-diones for
To assess the potential of organics to disrupt the interac-
tion between VEGF and NRP1, an original ELISA-
based binding assay was developed. This assay is using
a luminescent mode of detection supported by a VEGF
specific antibody–enzyme (HRP) conjugate and recom-
binant NRP1.10 The screening was performed with a
Keywords: VEGF-NRP1; Protein–protein interaction; Multi-compo-
nent condensations; 3-Amino-1-cyano-indolizines.
*
Corresponding author. Fax: +33437472381; e-mail: stephane.poigny @
pierre-fabre-urologie.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.05.014

K. Bedjeguelal et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3998–4001
3999
CN
CN
CN
NH
Ph
N
CN
Ph
NH2
N
b
a
Ph
NH
Cl
Cl
HN
N
N
OMe
Cl
1
Cl
O
R
Figure 1. 5-Chloro-3-cyclohexylamino-2-phenyl-indolizine-1-carbo-
nitrile.
OMe
11
9
12: R=Ph
13: R=Me
Scheme 2. Reagent and conditions: (a) TFA, DCM, 1 h; (b) BzCl for
12, AcCl for 13, TEA, DCM, 16 h.
collection of approximately 20,000 small molecules
selected for their structural relevance and diversity:
11,800 compounds purchased from dedicated suppliers
and 8160 compounds prepared in-house using 22 differ-
ent, often proprietary, MCCs.
tyl chloride, giving the N-acylated compounds 12 and
13, respectively.
While none of the commercially available compounds
inhibited significantly the VEGF–NRP1 interaction,
several compounds clustered on three different MCC-
based chemotypes showed some activity. Among those
chemotypes, 3-amino-1-cyano-indolizines were the
most potent antagonists. A prototypical indolizine, 5-
chloro-3-cyclohexylamino-2-phenyl-indolizine-1-carbo-
nitrile (1) (Fig. 1), showed an IC50 of 29 lM in our
ELISA assay.
An optimization program was therefore designed from 1
in which the three substitutable positions, R1 to R3, were
sequentially and orthogonally varied. Despite several
variations at the R1 position, no improvement over the
chloro group was found and this substituent was there-
fore considered to be optimum and conserved for the
next round of optimization (Table 1). Whereas chloro
derivatives seem to be associated with the strongest
inhibitory activities, it should be noted that significant
activities remain by changing it for other substituents,
ruling out consequently that activity is mediated by
the alkylation of chloro-indolizine through a SNAr
mechanism.
3-Amino-1-cyano-indolizines were easily prepared using
a recent three-component condensation discovered in
our laboratory (Scheme 1, step a).11 Isocyanides under-
went condensation with aldehydes and cyano-2-methy-
lene-pyridines, in the presence of a catalytic amount of
base, to give 3-amino-1-cyano-indolizines in fair to good
yields. The mechanism is thought to proceed through
the cheletropic addition of the isocyanide moiety to
the Knoevenagel-type hetero-diene intermediate.12
R2 optimization was conducted by running MCCs using
an extended set of isocyanides of either commercial
origin or custom-made according to published proce-
dures (Table 2).13 Great variations among different sub-
stituents were observed for this position. While aromatic
moieties seemed to be better tolerated, a two-carbon ali-
phatic linker was preferred to maintain the higher inhib-
itory activities (compounds 23–25). However, acylation
of the amino function did not show any biological activ-
ity (compounds 12 and 13).
A great benefit of MCC is that each substituent is
addressable in the course of the reaction, greatly facili-
tating the lead optimization phase. In the case of 3-ami-
no-1-cyano-indolizines, the C-5 position of the
indolizine ring could also be varied easily by conducting
a subsequent SNAr substitution of the fluoro group on
the compound 2 (Scheme 1, step b).
As shown in Scheme 2, removal of the 2,4-dimethoxyb-
enzyl group in 11 with TFA afforded primary amine 9,
followed by N-acylation with benzoyl chloride and ace-
Table 1. Yield and inhibitory activity of 3-amino-1-cyano-indoli-
zines—variations at the R1 position
CN
N
CN
CN
R1
HN
a
R3
NC
CHO
R3
+
+
N
R2
N
b
Compound R1
Yield % ELISAc IC50 (lM)
2
X
HN
R
X
1
2
3
4
5
6
7
8
Cl
F
68a
65a
98b
95b
49b
17b
61b
29
na
na
na
na
79
49
42
X = F, Cl
CN
SCH2CO2Me
SCH2CO2H
SMe
R3
N
R1
2
HN
R
O(CH2)3NMe2
1-Morpholinyl
NH(CH2)2NMe2 48b
Scheme 1. Reagent and conditions: (a) 10% DBU, n-butanol, rt then
100 °C, 16 h; (b) methyl thioglycolate for 3, mercaptoacetic acid for 4,
methyl mercaptan for 5, TEA, DMF, 70 °C, 16 h; 3-dimethylamino-1-
propanol for 6, NaH, DMF, 70 °C, 16 h; morpholine for 7, unsym-
dimethylethylenediamine for 8, DMF, 70 °C, 16 h.
a Isolated yield based on cyano-2-methylene-pyridine.
b Scheme 1, step b.
c na, not active.

4000
K. Bedjeguelal et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3998–4001
Table 2. Yield and inhibitory activity of 3-amino-1-cyano-indoli-
zines—variations at the R2 position
Table 3. Yield and inhibitory activity of 3-amino-1-cyano-indolizines -
variations at the R3 position
CN
CN
R3
N
N
2
HN
HN
Cl
R
Cl
Compound R2
Yielda
%
ELISAd
IC50 (lM )
Compound R3
Yielda (%) ELISAb
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
H
(CH2)2NEt2
74b
7
na
na
na
na
na
80
62
60
47
43
19
13
9
IC50, (lM)
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Cyclohexyl
4-Quinolinyl
4-Piperidinyl
76
38
98
7
na
na
95
22
14
13
13
10
10
9
2,4-Dimethoxybenzyl
Benzoyl
Acetyl
73
9c
32c
11
11
7
2-(5-Methylthiazoyl)
4-tert-Butylphenyl
(CH2)2OH
3-Aminophenyl
(CH2)3NMe2
2,6-Dimethylphenyl
3
44
4-Trifluoromethoxyphenyl 54
4-Dimethylaminophenyl 18
(CH2)2NH2 17
CH2CONHCH2(2-furyl) 17
4-(Me2N(CH2)3O)-Phenyl
2-Thiophenyl
2-Toyl
8
2
8
3,4-Dichlorobenzyl
(CH2)2OMe
11
18
10
41
36
11
4-Pyridyl
3-Thiophenyl
12
2
6
6
(CH2)2N(Me)Benzyl
2-(4-Chlorophenyl)ethyl
2-(2-Chlorophenyl)ethyl
2-Phenylethyl
4
3,5-Dimethylphenyl
4-Fluorophenyl
4-Toyl
62
10
6
6
4
5
3
5
2
4-Methoxyphenyl
3-Pyridyl
5
5
a Isolated yield based on cyano-2-methylene-pyridine.
b Scheme 2, step a.
c Scheme 2, step b.
d na, not active.
45
10
13
11
15
13
24
21
46
12
6
4
4-Fluoronaphthyl
3-Chlorophenyl
4-Methylthiophenyl
2-Methoxyphenyl
2-Chlorophenyl
2-Fluorophenyl
2-Phenylethyl
2-Pyridyl
4
3
3
3
3
3
R3 optimization was addressed by using a variety
of commercial aldehydes (compounds 26–53, Table 3).
It appeared that aromatic moieties, presence were
related to activities, while higher ranking com-
pounds are characterized by substituted phenyl
rings. Nature and position of the aromatic subtitu-
ents proved to have little influence on potency
(compounds 39–43).
2
2
3-Methoxyphenyl
4-Chlorophenyl
3-Fluorophenyl
2
2
10
2
a Isolated yield base on cyano-2-methylene-pyridine.
b na, not active.
Table 4. Migration assay on PC3 cells with compounds 1, 19, 24, 25,
and 53
From the hit identified in the primary screening
campaign (compound 1), we were able to rapidly
improve potency by an order of magnitude to reach
low micromolar inhibitory activity within a four
months timeframe. To further validate the potential
relevance of cyano-indolizine for cancer treatment,
several compounds from the 3-amino-1-cyano-indoli-
zines series were selected and evaluated in an in vitro
assay for measuring their inhibitory potency against
migration of PC3 prostate cancer cell lines14 (see
Table 4).
Compound
ELISA IC50 (lM)
Migrationa IC50 (lM)
1
19
24
25
53
29
19
3
72
54
23
21
8
2
2
a All experiments were repeated atleast twice.
References and notes
Compounds 1, 19, 24, 25, and 53 inhibited migration of
PC3 cell lines in the range from 8 to 72 lM, while main-
taining a correlation between the biochemical and func-
tional efficacy.15
1. Soker, S.; Takashima, S.; Miao, H. Q.; Neufeld, G.;
Klagsbrun, M. Cell 1998, 92, 735.
2. He, Z.; Tessier-Lavigne, M. Cell 1997, 90, 739.
3. Results to be published elsewhere.;
4. Various approaches have been tested in order to selectively
inhibit the VEGF pathway. Soluble VEGF receptors (i.e.
VEGF-Trap) have shown a powerful and selective antitu-
mor action, whereas VEGF specific antibody Avastin has
been successfully used as a therapeutic in several cancer
indications. Other strategies focus on inhibiting the kinase
In conclusion, the use of MCCs allowed us to conduct
expeditious lead optimization process even in the
challenging domain of protein–protein interaction
disruption historically addressed by more elaborate
peptidomimetics and natural compounds.

K. Bedjeguelal et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3998–4001
4001
´
9. See for instance: (a) Zhao, G.; Sun, X.; Bienayme, H.;
Zhu, J. J. Am. Chem. Soc 2001, 123, 6700; (b) Janvier, P.;
activity of VEGFR2 in order to annihilate the VEGF
effect.
´
Sun, X.; Bienayme, H.; Zhu, J. J. Am. Chem. Soc. 2002,
124, 2560.
5. (a) Vassilev, L. T.; Vu, B. T.; Graves, B.; Carvajal, D.;
Podlaski, F.; Filipovic, Z.; Kong, N.; Kammlott, U.;
Lukacs, C.; Klein, C.; Fotouhi, N.; Liu, E. A. Science
2004, 303, 844; (b) Grasberger, B. L.; Lu, T.; Schubert, C.;
Parks, D. J.; Carver, T. E.; Koblish, H. K.; Cummings, M.
D.; LaFrance, L. V.; Milkiewicz, K. L.; Calvo, R. R.;
Maguire, D.; Lattanze, J.; Franks, C. F.; Zhao, S.-Y.;
Ramachandren, K.; Bylebyl, G. R.; Zhang, M.; Manthey,
C. L.; Petrella, E. C.; Pantoliano, M. W.; Deckman, I. C.;
Spurlino, J. C.; Maroney, A. C.; Tomczuk, B. E.; Molloy,
C. J.; Bone, R. F. J. Med. Chem. 2005, 48, 909; (c) Parks,
D. J.; LaFrance, L. V.; Calvo, R. R.; Milkiewicz, K. L.;
Gupta, V.; Lattanze, J.; Ramachandren, K.; Carver, T. E.;
Petrella, E. C.; Cummings, M. D.; Maguire, D.; Grasber-
ger, B. L.; Lu, T. Biorg. Med. Chem. Lett. 2005, 15, 765.
6. (a) Degterev, A.; Lugovskoy, A.; Cardone, M.; Mulley, B.;
Wagner, G.; Mitchison, T.; Yuan, J. Nat. Cell Biol. 2001,
3, 173; (b) Tzung, S. P.; Kim, K. M.; Basanez, G.; Giedt,
C. D.; Simon, J.; Zimmerberg, J.; Zhang, K. Y.; Hocken-
bery, D. M. Nat. Cell Biol. 2001, 3, 183.
10. ELISA test: Neuropilin 1 was expressed in a baculovirus
system as a soluble form able to bind VEGF165, purified
from High Five insect cells, and stored at À80 °C in 50%
glycerol. VEGF165 was bought from R&D Systems and
stored at À80 °C. The screening procedure is based on an
ELISA test with Nrp1 coated on the plates and VEGF165
added for the binding assay. After several washing steps,
an anti-VEGF antibody is used to detect the level of
interaction between the two proteins. Detection is based
on luminescence generated by an enzyme (HRP) coupled
to the antibody.
´
11. (a) Bedjeguelal, K.; Bienayme, H.; Schmitt, P.; Grisoni, S.
PCT Int. Appl. WO 2,004,108,722, 2004;.; (b) Bedjeguelal,
´
K.; Bienayme, H.; Poigny, S.; Schmitt, P.; Tam, E. QSAR
Combi. Sci. 2006, in press.
12. For a similar mechanism-based three-component conden-
´
sation, see: Bienayme, H.; Bouzid, K. Angew. Chem. Int.
Ed. 1998, 37, 2234.
7. For additional examples of small molecules antagonizing
protein–protein interactions, see; (a) Gadek, T. R.; Nich-
olas, J. B. Biochem. Pharmacol. 2003, 65, 1; (b) Toogood,
P. L. J. Med. Chem. 2002, 45, 1543.
13. Bon, R. S.; Vliet, B. V.; Sprenkels, N. E.; Schmitz, R. B.;
de Kanter, F. J.; Stevens, C. V.; Swart, M.; Bickelhaupt,
F. M.; Groen, M. B.; Orru, R. V. A. J. Org. Chem. 2005,
70, 3542.
8. (a) Do¨mling, A.; Ugi, I. Angew. Chem. Int. Ed. 2000, 39,
´
14. Fong, C. J.; Sutkowski, D. M.; Kozlowski, J. M.; Lee, C.
Invasion Metastasis 1992, 12, 264.
3168; (b) Bienayme, H.; Hulme, C.; Oddon, G.; Schmitt,
P. Chem. Eur. J. 2000, 6, 3321; (c) Zhu, J. Eur. J. Org.
Chem. 2003, 7, 1133; (d) Zhu, J.; Bienayme, H. Multi-
component Reactions; Wiley-VCH: Weinheim, 2005.
15. These activities were not the consequence of a general
toxicity, as the compounds failed to show any toxicity in a
LDH-release assay.