Structure-activity relationship of a series of non peptidic RGD integrin antagonists targeting α5β1: Part 2

Article abstract of DOI:10.1016/j.bmcl.2012.04.061

Potent antagonists of the integrin α₅β₁, which are RGD mimetics built from tyrosine are described. This paper describes the optimization of in vitro potency obtained by variation of two parts of the molecule, the central aromatic core and the amide moiety.

Full text of DOI:10.1016/j.bmcl.2012.04.061

Bioorganic & Medicinal Chemistry Letters 22 (2012) 4117–4121
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure–activity relationship of a series of non peptidic RGD integrin
antagonists targeting
a5b1: Part 2
Bénédicte Delouvrié a, , Katherine Al-Kadhimi b, Jean-Claude Arnould a, Simon T. Barry b,
Darren A.E. Cross b, Myriam Didelot a, Paul R. Gavine c, Hervé Germain a, Craig S. Harris a,
Adina M. Hughes b, David A. Jude b, Jane Kendrew b, Christine Lambert-van der Brempt a,
Jean-Jacques Lohmann a, Morgan Ménard a, Andrew A. Mortlock b, Martin Pass b, Claire Rooney b,
Michel Vautier a, Jennifer L. Vincent b, Nicolas Warin a
⇑
a AstraZeneca, Centre de Recherches, Z.I. La Pompelle, B.P. 1050, Chemin de Vrilly, 51689 Reims, Cedex 2, France
b AstraZeneca Pharmaceuticals, Alderley Park, Macclesfield, Cheshire SK10 4TG, United Kingdom
c AstraZeneca R&D, Innovation Center China, Shanghai 201203, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 21 April 2012
Potent antagonists of the integrin a5b1, which are RGD mimetics built from tyrosine are described. This
paper describes the optimization of in vitro potency obtained by variation of two parts of the molecule,
the central aromatic core and the amide moiety.
Keywords:
Ó 2012 Elsevier Ltd. All rights reserved.
Integrin
a5b1 antagonist
Zwitterion
RGD integrin
Angiogenesis
Integrins
a
5b1 and
a
vb3 are members of a complex family of
group, as well as looking at the best basic moieties to optimize
their interactions with the 5 subunit, we also worked on optimiz-
heterodimeric glycoprotein receptors that regulate a number of
cellular processes such as cell adhesion, migration and survival.
They interact with a family of matrix proteins including fibronec-
tin, vitronectin and fibrinogen, through specific motifs known as
RGD motifs. A key role is the regulation of angiogenesis, through
complex processes that are not yet fully clarified.1 Several integrins
a
ing the amide moiety and assessed a diversity of central cores. Our
SAR findings on the core and the amide are the topic of this letter.
With the exception of compound 23, the synthesis of com-
pounds 1–34 was described previously and is summarized on
Scheme 1.10 To study the influence of the amide, the basic chain
was introduced first and a peptidic coupling allowed the synthesis
of compounds 1–17 in a parallel synthesis fashion. On the contrary,
when the core was diversified, each synthesis was linear. The basic
chains were introduced either by Mitsunobu for O-linked chains or
by Heck coupling followed by hydrogenation for C-linked chains.
Benzofuran 23 was made from an iodo tyrosine derivative used
for the synthesis of compound 24. The iodo phenol reacted with
an acetylenic derivative to form the benzofuran moiety in a palla-
dium catalyzed annulation reaction.11
have been associated with cancer, amongst which
a5b1 and avb3
are the most studied in clinic at present, with a diversity of inhib-
itors including small peptides (Cilengitide and ATN161) and anti-
bodies (Abegrin and Volociximab).2–5
RGD integrins such as a5b1 are challenging to inhibit, due to the
protein–protein nature of their interaction with their ligands. Pep-
tide–peptide interactions are notoriously challenging to mimic
with non peptidic small molecules and the first attempts to inhibit
a
5b1 resulted in relatively large molecules of MW >600 g/mol.6,7 An
additional complexity is that the RGD sequence to mimic contains
both an acidic and a basic function, which restrained most of the
research effort on zwitterionic approaches.8 We reported in a pre-
vious communication our effort to develop a series of orally avail-
Amide variation: Literature on a4b1 inhibitors reported that
ortho substituted phenyl amides with halogen or alkyl groups gave
optimal potency. In particular 2,6 di-ortho substituted phenyls
enabled an ideal orientation of the aromatic moiety inside the
hydrophobic pocket of the b1 subunit.12 Since the amide group is
modeled to be positioned in the b1 unit, it was logical to assume
able
a
5b1 inhibitors from a tyrosine based zwitterionic scaffold.9
During optimization of the linker between the core and the basic
that the knowledge of
a4b1 would transfer well to a5b1. Indeed,
ortho monosubstitution with a chloride in compounds 2 and 3 or
small alkyls in compounds 6–8 provided an average 10-fold
improvement of both binding and cell adhesion potencies
⇑
Corresponding author.
E-mail address: benedicte.delouvrie@astrazeneca.com (B. Delouvrié).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.04.061

4118
B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4117–4121
O
O
O
N
O
N
1. Mitsunobu
2. de-BOC
3. peptidic coupling
O
NH
OH
R
N
H
NH 4. saponification
O
O
H2N
N
H
O
O
O
O
1-17
1. Heck coupling
2. hydrogenation
3. de-BOC
OH
O
Br
chain
1. Mitsunobu
2. peptidic coupling
Cl
core
O
chain
core
O
Cl
O
core
O
core
OH
O
O
N
N
O
3. de-BOC
N
N
H
4. peptidic coupling
5. saponification
H
H
O
Cl
4. saponification
O
Cl
O
O
18-22, 24-26, 30, 32, 34
27-29, 31, 33
OH
OH
I
i
Cl
O
Cl
O
Cl
Cl
Cl
O
O
N
O
H
N
H
O
O
N
N
H
O
O
N
O
Cl
Cl
N
N
iv
v-vi
Cl
O
OH
Cl
O
N
H
Cl
Cl
Cl
O
O
N
H
O
Cl
O
O
Cl
23
H2N
ii-iii
N
N
Scheme 1. General synthesis of compounds 1–34. Reagents and conditions: (i) benzyltrimethylammonium dichloroiodate, DMF, 0–20 °C, 40%; (ii) 2,2,2-trichloroethylchlo-
roformate, NEt3, CH2Cl2, 0–20 °C, 29%; (iii) propargyl alcohol, di-t-butylazodicarboxylate, PPh3, THF, 20 °C, 74%; (iv) NEt3, cat. CuI, cat. PdCl2(PPh3)2, DMF, 65 °C, 23%; (v) Zn,
H2O/THF/AcOH, 20 °C, 61%; (vi) LiOH, H2O/DMF, 20 °C, 51%.
Table 1
Amide optimization on compounds 1–17
R
O
HN
O
N
COOH
N
Compound
R
Binding IC50
l
,
K562 cell adhesion no HSA IC50
l
,
K562 cell adhesion 630
Ma
l
M HSA IC50
,
% Free
human
% Free
rat
Ma
Ma
l
1
2
3
4
5
6
7
8
Ph
2-Cl Ph
2-Cl, 4-F Ph
2-Cl, 6-F Ph
0.049
0.003
0.002
<0.001
0.13
0.009
0.004
0.003
0.001
0.001
<0.001
<0.001
0.012
0.001
0.007
0.005
0.004
0.43
—
—
—
0.042
0.033
0.019
0.82
>26
0.4
0.2
0.6
—
—
—
—
—
—
1.0
2.4
0.1
—
—
0.25
—
0.5
0.3
0.5
—
11b
1.5
2-OMe, 4-F Ph
2-Me, 4-F Ph
2-Et, 4-F Ph
2-cPr, 4-F Ph
2,6-Di Me Ph
2,4,6-Tri Me Ph
2,6-Di Cl Ph
3,5-Di Cl 4-pyridyl
3-Cl 2-thiophenyl
3,5 Di-Me 4-isoxazolyl
2-Me cyclohexyl
1-Me cyclohexyl
1-MeO 4-
>100b
24b
0.036
0.024
0.039
<0.023
0.022
0.007
<0.004
0.052
<0.008
0.071
0.039
0.050
0.5
0.4
0.4
2.2
0.5
1.1
3.3
0.2
3.8
0.9
0.9
21.2
5.0b
7.7b
0.42b
1.5b
0.17b
0.055b
36b
9
10
11
12
13
14
15
16
17
1.1
29b
7.5b
0.48b
tetrahydropyranyl
a
In all tables, unless stated otherwise, numbers are a geometric mean of 2 or more values.
n = 1 value.
b
compared to the unsubstituted phenyl amide 1 (Table 1). However,
an ortho-OMe did not provide a similar advantage (5 vs 1). Since
X-ray structures in the Cambridge Crystallography Structural Data-
base show that both ortho-Cl and ortho-OMe phenyl amides main-
tain the ortho substituent on the same side than the NH of the
amide and in the same plane, the potency difference between Cl
and OMe is presumably due to an electronic effect rather than a
conformational effect. Moreover, the difference of potency be-
tween the ethyl 7 or the cyclopropyl 8 and the methoxy 6 sug-
gested an electronic effect rather than a steric effect. Further
potency was gained with di-ortho substitution, which is known
to force the aryl ring to twist by approximately 80°. Such

B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4117–4121
4119
11 vs 2). This gain of potency was even more noticeable in the cell
adhesion assay with HSA, because di ortho substituted compounds
were both more potent and less bound to albumin than their mono
ortho substituted analogs (see 11 vs 2).13 Para substitution had lit-
tle influence on potency (see for instance 3 vs 2) but could offer
interesting PK modulations as will be described in a forthcoming
paper. Heteroaryl amides were also explored. Again, ortho substi-
tution with chloride or small alkyl groups was found beneficial
on both 5- and 6-membered heteroaryls (compounds 12–14), with
some of these examples being among our most potent compounds
in cell. Cycloalkyls were also well tolerated, in particular those
with steric hindrance close to the amide carbonyl (2-substitution
in 15; 1-substitution in 16 and 17).
The amide SAR transferred when the basic chain on the left
hand side of the molecules was varied. This is represented by a plot
of matched pairs with 3 different basic chains and 9 different
amides (Fig. 1).
Core variation: Due to the flexible linker and based on the
homology model allowing some flexibility, one could have ex-
pected to maintain reasonable activity when moving from the
para-tyrosine to the meta-tyrosine core. However, comparison of
compounds 19 with 18 and 21 with 20 clearly illustrated a signif-
icant drop off in potency with a meta linker (Table 2). Compound 22
with a longer linker was also poorly active. The benzofurane 23, a
cyclized version of compound 20, was more potent than the meta-
tyrosine examples, but still 35-fold less potent than compound 20
in cells.
Substitution on the phenyl ring was tolerated and guided by the
homology model which suggested that there was limited room to
add substituents. Mono Me or F examples 24 and 25 remained
potent, although showing a 5- to 10-fold drop off in cell potency
compared to compound 11. Di-Me substitution was less tolerated
(26 vs 11).
Figure 1. Transposition of amide SAR for different basic chains.
compounds (4, 9, 10, 11) consistently reached a threefold improve-
ment in cell potency in the adhesion assay without human serum
albumin (HSA) compared to monosubstituted analogs (for instance
Table 2
Core optimization on compounds 18–33
COOH
O
Cl
R
N
H
Cl
Compound
R
Binding IC50
,
l
Ma
K562 cell adhesion IC50
2.5
, l
Ma
O
18
0.082
N
*
N
N
N
19
5.9
>290b
O
*
N
N
N
O
20
21
22
0.011
—
0.12
20b
83b
*
*
O
N
N
N
1.9
O
*
O
N
N
23
11
0.89
4.0
*
*
N
N
O
<0.001
0.007
(continued on next page)

4120
B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4117–4121
Table 2 (continued)
Compound
R
Binding IC50
,
l
Ma
K562 cell adhesion IC50
0.081
, l
Ma
N
N
N
N
O
24
25
26
0.003
*
F
O
O
0.002
0.021
0.047
0.61
*
*
N
N
N
N
27
28
29
0.008
0.002
0.18
0.046
0.031
0.40
*
*
S
N
N
N
N
*
*
S
S
N
N
N
N
N
O
O
30
31
32
33
0.11
0.51b
0.11
0.009
0.001
0.005
0.006
*
*
*
*
N
N
N
N
N
N
0.026
0.076
0.11
N
N
N
O
34
a
In all tables, unless stated otherwise, numbers are a geometric mean of 2 or more values.
n = 1 value.
b
Replacement of the phenyl core with a thiophene core was ex-
plored with the two possible thiophene isomers. From examination
of the geometries, it was anticipated that the 2,5-thiophene would
better position the basic chain than the 2,4-thiophene (Fig. 2). And
indeed, compound 28 was much more potent than compound 29.
28 was also slightly more potent than the corresponding phenyl
analog 27. Moreover, the benefit in potency which was observed
with a phenyl core when moving from a carbon linker to an oxygen
linker was not observed for 2,4 thiophene (30 vs 29). The thio-
phene core was key to improve the DMPK parameters in this series.
Eventually, azatyrosines were explored. With a carbon linker,
the two isomers of azatyrosine 31 and 33 were within twofold
equipotent with the tyrosine analog 27. A larger drop off in po-
tency was observed for an oxygen linked pyridines versus the phe-
nyl analog 11, with a preference for isomer 32 compared to 34.
To assess the selectivity of our compounds against
avb3, we
developed cell adhesion assays that could differentiate selective
from dual inhibitors.14 The A375 cell adhesion assay with fibrino-
gen is reflecting binding mediated through
some inhibition in this assay but the level of potency was at least
10-fold lower than the Merck MK-0429
vb3 inhibitor.15 The
avb3. Our series showed
Figure 2. Geometric parameters of phenyl and thiophenes central cores.
a

B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4117–4121
4121
Table 3
References and notes
RGD integrin selectivity dataa
1. (i) Avraamides, C. J.; Garmy-Susini, B.; Varner, J. A. Nat. Rev. Cancer 2008, 8, 604;
(ii) Silva, R.; D’Amico, G.; Hodivala-Dilke, K. M.; Reynolds, L. E. Arterioscler.
Thromb. Vasc. Biol. 2008, 28, 1703; (iii) Cox, D.; Brennan, M.; Moran, N. Nat. Rev.
Drug Disc. 2010, 9, 804.
Compound A375 cell adhesion
A375 cell adhesion fibronectin IC50, l
Mb
without/with
fibrinogen IC50
,
l
Mb
a
vb3 inhibitor (ratio)c
1
6.7
7.4/5.4 (1.4)
2. (i) Reardon, D. A.; Neyns, B.; Weller, M.; Tonn, J. C.; Nabors, L. B.; Stupp, R.
Future Oncol. 2011, 7, 339; (ii) Mas-Moruno, C.; Rechenmacher, F.; Kessler, H.
AntiCancer Agents Med. Chem. 2010, 10, 753.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
20
24
25
26
28
32
34
0.52
0.44
0.28
8.3
0.51
0.34
0.52
0.40
0.37
0.23
0.13
2.9
0.54/0.24 (2.2)
0.44/0.21 (2.1)
0.28/0.16 (1.7)
8.3/6.5 (1.3)
0.51/0.47 (1.1)
0.34/0.20 (1.7)
0.51/0.25 (2.0)
0.40/0.090 (4.5)
0.38/0.057 (6.7)
0.23/0.097 (2.3)
0.13/0.052 (2.6)
2.9/2.7 (1.1)
0.20/0.13 (1.5)
0.96/0.42 (2.3)
0.83/0.48 (1.8)
0.71/0.44 (1.6)
2.5/0.61 (4.2)
0.64/0.18 (3.6)
0.33/0.076 (4.3)
8.8/5.8 (1.5)
0.80/0.18 (4.4)
0.51/0.084 (6.1)
1.2/0.57 (2.1)
0.81/2.2 (0.4)
3. (i) Khalili, P.; Arakelian, A.; Chen, G.; Plunkett, M. L.; Beck, I.; Parry, G. C.;
Donate, F.; Shaw, D. E.; Mazar, A. P.; Rabbani, S. A. Mol. Cancer Ther. 2006, 5,
2271; (ii) Cianfrocca, M. E.; Kimmel, K. A.; Gallo, J.; Cardoso, T.; Brown, M. M.;
Hudes, G.; Lewis, N.; Weiner, L.; Lam, G. N.; Brown, S. C.; Shaw, D. E.; Mazar, A.
P.; Cohen, R. B. Br. J. Cancer 2006, 94, 1621; (iii) Donate, F.; Parry, G. C.; Shaked,
Y.; Hensley, H.; Guan, X.; Beck, I.; Tel-Tsur, Z.; Plunkett, M. L.; Manuia, M.;
Shaw, D. E.; Kerbel, R. S.; Mazar, A. P. Clin. Cancer Res. 2008, 14, 2137.
4. (i) Landen, C. N.; Kim, T.-J.; Lin, Y. G.; Merritt, W. M.; Kamat, A. A.; Han, L. Y.;
Spannuth, W. A.; Nick, A. M.; Jennings, N. B.; Kinch, M. S.; Tice, D.; Sook, A. K.
Neoplasia 2008, 10, 1259; (ii) Delbaldo, C.; Raymond, E.; Vera, K.;
Hammershaimb, L.; Kaucic, K.; Lozahic, S.; Marty, M.; Faivre, S. Invest. New
Drugs 2008, 26, 35; (iii) Mulgrew, K.; Kinneer, K.; Yao, X.-T.; Ward, B. K.;
Damschroder, M. M.; Walsh, B.; Mao, S.-Y.; Gao, C.; Kiener, P. A.; Coats, S.;
Kinch, M. S.; Tice, D. A. Mol. Cancer Ther. 2006, 5, 3122.
0.2
0.96
0.83
0.71
2.6
0.60
0.29
8.8
0.79
0.62
1.2
5. (i) Tamaskar, I. Drugs Fut. 2010, 35, 16; (ii) Kuwada, S. K. Curr. Opin. Mol. Ther.
2007, 9, 92.
6. (i) Jadhav, P. K.; Smallheer, J. M. WO1997/033887, 1997.; (ii) Smallheer, J.;
Weigelt, C. A.; Woerner, F. J.; Wells, J. S.; Daneker, W. F.; Mousa, S. A.; Wexler, R.
R.; Jadhav, P. K. Bioorg. Med. Chem. Lett. 2004, 14, 383; (iii) Marinelli, L.; Meyer,
A.; Heckmann, D.; Lavecchia, A.; Novellino, E.; Kessler, H. J. Med. Chem. 2005, 48,
4204.
7. Stragies, R.; Osterkamp, F.; Zischinsky, G.; Vossmeyer, D.; Kalkhof, H.; Reimer,
U.; Zahn, G. J. Med. Chem. 2007, 50, 3786.
MK-0429d
0.0003
8. (i) Zischinsky, G.; Osterkamp, F.; Vossmeyer, D.; Zahn, G.; Scharn, D.;
Zwintscher, A.; Stragies, R. Bioorg. Med. Chem. Lett. 2010, 20, 65; (ii)
Zischinsky, G.; Osterkamp, F.; Vossmeyer, D.; Zahn, G.; Scharn, D.;
Zwintscher, A.; Stragies, R. Bioorg. Med. Chem. Lett. 2010, 20, 380; (iii)
Heckmann, D.; Meyer, A.; Marinelli, L.; Zahn, G.; Stragies, R.; Kessler, H.
Angew. Chem., Int. Ed. 2007, 46, 3571; (iv) Heckmann, D.; Meyer, A.; Laufer, B.;
Zahn, G.; Stragies, R.; Kessler, H. ChemBioChem 2008, 9, 1397.
a
b
c
All numbers are a geometric mean of 2 or more values.
Assay performed in presence of Mg2+
The Merck vb3 MK-0429 was used at 1
The Merck compound showed an IC50 of 4.6
.
a
l
M concentration.
d
l
M in our K562 cell adhesion assay.
9. Delouvrié, B.; et al. Bioorg. Med. Chem. Lett. 2012. http://dx.doi.org/10.1016/
j.bmcl.2012.04.063.
activity of our series for both
the A375 cell adhesion assay in presence of fibronectin without
and with an vb3 inhibitor. Results in Table 3 illustrated that our
a
5b1 and
a
vb3 was confirmed using
10. Arnould, J.-C.; Delouvrié, B.; Ducray, R.; Lambert-Van Der Brempt, C. M. P.
WO2007/141473, 2007.
11. Larock, R. C.; Yum, E. K.; Doty, M. J.; Sham, K. K. C. J. Org. Chem. 1995, 60, 3270.
12. Davenport, R. J.; Munday, J. R. Drug Disc. Today 2007, 12, 569.
13. Leach, A. G.; Jones, H. D.; Cosgrove, D. A.; Kenny, P. W.; Ruston, L.; MacFaul, P.;
Wood, M.; Colclough, N.; Law, B. J. Med. Chem. 2006, 49, 6672.
a
series showed modest ratios (1.1–6.7) suggesting dual
a
5b1/avb3
inhibition.
14. To determine the ability of our compounds to inhibit
adhesion assay A375M cells was employed. A375M melanoma cells express a
range of integrins including 5b1, vb3 and vb5, however in the presence of the
physiological cation Mg2+ binding to fibrinogen is mediated by
vb3 integrin.
This assay was validated with an av integrin blocking antibody (L230, active)
and an 5b1 integrin blocking antibody (M200, inactive).
A dual fibronectin adhesion assay was used to determine compound selectivity
for 5b1 over
vb3. In the presence of Mg2+, cell adhesion to fibronectin is
mediated by both v and 5b1 integrins. Inhibition of one integrin alone should
not affect adhesion to fibronectin, as the other integrin can compensate. To
eliminate adhesion via vb3 integrin, these assays were performed in the
presence or absence of an vb3 inhibitor. The larger the shift in IC50 of cell
adhesion to fibronectin from the absence to the presence of the vb3 inhibitor,
the more 5b1 selective the compound is.
avb3, a fibrinogen
In conclusion, we have developed potent inhibitors of
which also carry some activity against
a
5b1
a
vb3. The in vivo properties
a
a
a
a
of this zwitterionic series will be described in a forthcoming paper.
a
a
a
Acknowledgments
a
a
We would like to acknowledge the following scientists for their
contribution in the synthesis or characterization of the com-
pounds: Dominique Boucherot, Christian Delvare, Dr Delphine
Dorison Duval, Patrice Koza, Jacques Pelleter, Fabrice Renaud, Dr
Kin Tam.
a
a
a
a
15. The Merck compound MK-0429 demonstrated Proof of Concept in a phase II
clinical trial against osteoporosis.