Structure-activity relationship of a series of non peptidic RGD integrin antagonists targeting α5β1: Part 1

Article abstract of DOI:10.1016/j.bmcl.2012.04.063

Potent antagonists of the integrin α₅β₁, which are RGD mimetics built from tyrosine are described. This letter describes the optimization of in vitro potency obtained by variation of two parts of the molecule, the basic group and the

Full text of DOI:10.1016/j.bmcl.2012.04.063

Bioorganic & Medicinal Chemistry Letters 22 (2012) 4111–4116
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 1
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
letter describes the optimization of in vitro potency obtained by variation of two parts of the molecule,
the basic group and the linker between the basic group and the phenyl central core.
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Integrin
a5b1 antagonist
Zwitterion
RGD integrin
Angiogenesis
The fibronectin receptor
family of heterodimeric glycoprotein receptors, which mediate
cell-cell and cell-extracellular matrix interactions. 5b1 is known
to play a key role in the early stage of angiogenesis and is overex-
pressed in human tumor vasculature (e.g., colon and breast carci-
nomas).1,2 Its overexpression has been associated with poor
prognosis.3 It was also demonstrated in animal models that inhibi-
a
5b1 is a member of the integrin super-
(Attenuon),11 currently in phase II trials. Here we describe the
SAR of a series of non peptidic 5b1 antagonists.
Minimal structural information has been reported on
receptor itself but a crystal structure of the extracellular domain
of
vb3 in complex with a cyclopeptide inhibitor is available.12 In
a
a
a5b1
a
the RGD integrins, the ligand binding is mediated through the
RGD recognition motif which makes two key and conserved inter-
tion of
growth.4,5 Moreover, the combination of a VEGFR antibody with
an 5b1 antibody proved more beneficial than each agent alone
to control tumor growth in several animal models, opening a
new avenue for a best use of
5b1 inhibitors in cancer.6
5b1 is a RGD integrin (like vb3,5,6 and IIbb3), in reference to
a
5b1 significantly reduced tumor angiogenesis and tumor
actions with both a and b subunits. The carboxylate group of the
aspartic acid coordinates with the metal-ion dependent adhesion
site (MIDAS) located in the b subunit (herein referred to as the
right hand side of the receptor), while the guanidine function of
arginine is engaged in a bidentate salt bridge with a highly con-
a
a
a
a
a
served aspartic acid residue in the
crystal structure of the cyclopeptide ligand bound to
that the distance between the basic and acidic moiety is about
13 Å. The strong sequence similarity between 5b1 and vb3
around the RGD binding site allowed us to build a useful homology
model of
5b1 to support our design rationale.13 The 2 key ionic
a
subunit (left hand side). The
the key recurrent arginine-glycine-aspartic acid (RGD) motif on
the natural peptidic ligand fibronectin. The very few non peptidic
avb3 reveals
small molecules reported to inhibit
a
5b1 all mimic elements of this
a
a
RGD peptidic sequence. Original spiro compounds of MW > 600 g/
mol from Dupont were followed first by a series of trisubstituted
pyrrolidines from Jerini, then more recently by a series of lower
MW.7–9 One compound of undisclosed structure, JSM6427, has re-
cently completed a phase I study for age-related macular degener-
ation (AMD) with an intravitreal injection formulation.5 The most
a
interactions oriented us to develop zwitterionic molecules repre-
sented by the general structure in Figure 1.
The central core was initially fixed as a phenyl ring, starting
materials being accessible as derivatives of tyrosine. Literature on
advanced
a
5b1 antagonists in clinic are the chimeric antibody
a
4b1 integrin suggested to use the 2,6-dichlorophenylamide group
Volociximab (PDL/Biogen),4,10 and the small peptide ATN-161
as a suitable right hand part of the molecule.14 Other groups also
described the importance of 2,6-disubstitution pattern for
a5b1
selectivity.9 Indeed, the 2,6-disubstitution twists the aromatic ring
out of the plane of the amide bond and enables an ideal orientation
⇑
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.2012.04.063

4112
B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4111–4116
*
*
H
N
N
N
*
*
tyrosine derivatives
O
Cl
X
*
N
H
N
Cl
α4β1 literature
linker
core
HOOC
base
R1
H
N
N
*
αvβ3 literature
key interaction with MIDAS cation
Figure 1. Rational design of a5b1 inhibitors.
of the aromatic moiety inside the hydrophobic pocket of the b1
subunit. Alternative literature exploring RGD mimics, especially
compounds, it is to be understood that more examples exist to sup-
port the SAR described.
those targeting
a
vb3, described extensive optimization of the basic
Linker variation: A preliminary docking study of compounds 1–3
left hand part, suggesting that derivatives of 2-aminopyridines
were suitable replacements of the fibronectin’s arginine.15 Having
those fragments of the molecules set up, the nature and length of
the linker needed to be optimized, driven by the homology model.
This letter describes the SAR observed when varying the linker and
in our a5b1 homology model suggested that the ideal distance be-
tween the aminopyridine group and the carboxylic acid group to
catch the two interactions with Asp227 and MIDAS would be ob-
tained with the five atom linker compound 2 (Fig. 2). Indeed, both
enzyme and cell IC50s confirmed this hypothesis, 2 being signifi-
cantly more potent than the shorter and longer linkers, respec-
tively 1 and 3 (Table 1). Analogs of compounds 1–3 were made
with a methylene linker and an alkyne linker, the best potency
being obtained with the five atom linkers 4 and 5.
The O-linker proved more potent than the methylene linker, as
illustrated by the pairs 2–5, 6–7, 9–10. This trend was observed on
18 matched pairs during our research program, as highlighted in
Figure 3, with up to 20-fold better cell adhesion potency for the
O-linked compounds. Although it is tempting to put forward the
reduced conformational freedom of the O-linker, resulting in a de-
creased entropic penalty compared to the methylene linker to
adopt the binding conformation, we cannot discount that the
receptor would better accommodate a more hydrophilic O atom
in this particular position.
the basic group of these new a5b1 inhibitors.
The synthesis of compounds 1–22 was described previously
and is summarized on Scheme 1.16 The linker was introduced by
a coupling reaction: Mitsunobu for X = O, Y = CH2, Heck coupling
for XY = HCCH, Sonogashira coupling for XY = CC and peptide cou-
pling for X/Y = N, Y/X = C(O). These coupling reactions led to the
methyl esters of general formula 23. When required, a deprotec-
tion step of the basic nitrogen (most typically protected as a
BOC) and/or a reduction step of unsaturated bonds could precede
a saponification of 23 to afford the final carboxylic acids 1–22.
Compounds were routinely evaluated using a binding assay and
an adhesion assay in K562 cells.16 For clarity of discussion, only a
limited set of data on representative compounds is used to de-
scribe the SAR. When trends are exemplified by a single pair of
O
Cl
+
OH
O
H
N
H
base
MeOOC
Cl
X=O, Y=CH2
Mitsunobu
O
Cl
Cl
+
2N
H
COOH
O
Cl
N
base
X
Y
XY=NHC(O) or C(O)NH
peptidic coupling
N
MeOOC
Cl
H
base
MeOOC
23
Cl
or
O
+
HOOC
NH2
XY=CC
Sonogashira
XY=HCCH
Heck
N
base
MeOOC
Cl
O
Cl
+
Br
N
H
Cl
base
base
MeOOC
O
Cl
O
Cl
1) reduction/deprotection steps
2) saponification
X
Y
X
Y
N
H
N
H
Cl
base
HOOC
MeOOC
23
Cl
1-22
Scheme 1. General synthesis.

B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4111–4116
4113
Table 1
Linker optimization on compounds 1–12
Cl
Cl
R
HN
O
COOH
Compound R
Binding K562 cell
IC50 adhesion
Ma
Measured
basic pKa
,
l
IC50
, l
Ma
N
N
N
N
O
1
2
0.082
0.011
2.5
6.7
7.1
O
0.12
Figure 2. Compound 2 docked in
a5b1 homology model revealing the expected
electrostatic interactions between the carboxylate group and the MIDAS site (b1
unit) and between the bidentate salt bridge and Asp227 (a5 unit).
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
O
3
0.054
0.069
0.024
0.71
2.1
The O- and CH2-linkers were always significantly more potent
than the alkyne linker analog (4 vs 2 and 5), possibly suggesting
that a partially rigidified linker was not able to position the basic
group to interact optimally with Asp227 when the flexible linear
linkers could. However, not all flexible linear linkers were equally
potent. Indeed, compound 6 (O-linker) was much more potent
than compound 8 (CH2O linker) and the O- and CH2-linkers were
more potent than the amide linkers (11–12 vs 9–10). This
illustrated that in addition to the conformational and flexibility
requirements, the electronic nature of the linker needs to be subtly
adjusted to fit well with the surrounding amino acids residues of
4
6.7
7.2
7.0
7.6
6.4
7.6
5
1.1
O
6
<0.001 0.007
7
0.008
0.22
0.046
1.4b
the integrin. Similar observations were reported with
avb3
inhibitors.17
O
8
Basic group variation: The 6-methylamino-pyridin-2-yl (MAP)
compounds 6–7 and the 5,6,7,8-tetrahydro[1,8]naphthyridin-2-yl
(THN) compounds 9–10 were much more potent (up to >100-fold)
than their 2-aminopyridinyl analogs 2 and 5 (Table 1). The small
differences of pKa (measured pKas were spread only on ꢀ1 unit
in Table 1) might not explain the significant differences of potency.
The salt bridge interaction with Asp227 seems stronger when the
non aromatic N atom is positioned towards the end of the a5
pocket rather than inserted into the linker chain.
O
9
<0.001 <0.003
<0.001 <0.008
10
11
N
O
0.003
0.005
0.084
0.12
O
N
In addition to aminopyridine derivatives (2, 6 and 9), we also
investigated other basic groups (Table 2). A basic aminoheterocycle
appeared to be a pre-requisite for getting strong affinity for the
N
N
12
6.8
receptor. This was reported on related
a
5b1 inhibitors.9 Amino-
benzimidazole with the appropriate linker also gave an interesting
level of potency as shown with compound 13. The aminobenzim-
idazole proved more potent than the non basic aminobenzothia-
zole analog (13 vs 14). The pKa influence could also explain the
decreased potency between the aminopyridine 2 and the non basic
urea 15. This could however be specific to a5 since potent urea
a
Unless stated otherwise, numbers are a geometric mean of 2 or more values.
n = 1 value.
b
of the groove is occupied by a Thr212. In
same location at the entry of the pocket than Asp218 in
the other ( v)Asp150 is replaced by a small non polar Ala159 and
v)Thr212 in the bottom is changed for a larger Gln221.This de-
creases the negatively charged surface of the 5 groove and slightly
shortens it. The potential of ( 5)Gln221 to provide extra H-bond
remains unclear. In IIb, the conserved Asp224 lies at the bottom
a5, Asp227 occupies the
av, but
inhibitors of
of monodentate basic groups were poor inhibitors of
examplified by the azetidine 16. This is clearly different from the
a
vb3 were reported.18 We also noticed that a variety
5b1, as
a
(
a
a
a
a
a
IIbb3 inhibitors, where for instance piperidines have reached high
levels of in vitro potency.19
a
of the groove, conferring a deeper character to this pocket. This
affords both monodentate aliphatic and bidentate basic groups to
bind, provided the monodentate base presents the N–H bond in
the appropriate direction. Another consequence is that the optimal
Figure 4 offers a possible explanation for these results. Compar-
ison of the X-ray crystal structures of
model of 5b1 revealed subtle differences in the location of the
conserved Asp in the b propeller groove of the respective
subunits.12,20 In
v, the conserved Asp218 is located on the side,
avb3, aIIb b3, and our 3D
a
a
distance between the basic and acidic moieties is longer in
than in vb3 or
aIIbb3
a
a
a
5b1 (ꢀ16 Å vs ꢀ13 Å).
at the entry of a shallow groove. Another Asp150 is present on
the opposite side of this pocket, offering an extra H-bond interac-
tion opportunity. Both Asp residues can indeed interact with the
arginine, providing a particularly strong interaction. The bottom
The piperazinopyridine 17 and the morpholinopyridine 18 were
well tolerated, although less potent than the THN analog 9, proba-
bly due to their lower pKa. Derivatives of the N-Methyl aminopyr-

4114
B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4111–4116
Figure 3. Binding and K562 cell adhesion IC50s for all C- versus O-linker matched pairs.
Table 2
Basic group variation on compounds 2,6,9 and 13–22
Cl
Cl
R
HN
O
COOH
Compound
R
Binding
IC50
Ma
K562 cell
adhesion
Measured
basic pKa
,
l
IC50, l
Ma
O
O
O
N
N
N
N
N
N
2
6
9
0.011
0.12
7.1
7.0
7.6
<0.001
<0.001
0.007
<0.003
0.16
N
S
13
14
0.008
0.97
7.2
O
O
N
N
N
H
2.4b
4.3
—
N
O
N
N
15
16
0.15
10
O
N
0.36b
58b
O
N
N
O
17
0.001
<0.008
7.1
6.0
N
N
O
O
N
18
19
0.001
0.010
0.009
O
H2N
N
<0.001

B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4111–4116
4115
Table 2 (continued)
Compound
R
Binding
K562 cell
adhesion
Measured
basic pKa
IC50
,
l
Ma
IC50, l
Ma
O
N
N
20
21
0.14
2.7b
O
N
N
0.003
0.083
<0.041
0.32b
O
N
N
22
6.8
a
Unless stated otherwise, numbers are a geometric mean of 2 or more values.
n = 1 value.
b
Figure 4. Interactions in the av
,
a5 and aIIb subunits.
Table 3
Cell adhesion assays +/À human serum albumin
Compound
% free in rat
Binding IC50
,
l
Ma
K562 cell adhesion assay no HSA IC50
,
l
Ma
K562 cell adhesion assay 630 lM HSA IC50, l
Ma
9
10
17
0.7%
0.6%
1.5%
<0.001
<0.001
0.001
<0.003
<0.008
<0.008
0.029
0.45
<0.056
a
All numbers are a geometric mean of 2 or more values.
idine 6 were made to explore the steric and electronic scope on the
a5 subunit. The des-methylated compound 19 was almost as po-
tent as 6. On the contrary, the N,N-dimethyl analog 20 showed a
significant decrease of potency. This could either be due to a steric
clash, or reflect the importance of a NH for a good binding to
Asp227. Some steric hindrance could be tolerated, as examplified
by compound 21, but to a certain limit as shown by the potency
decrease of cyclopentyl 22. This is in agreement with the presence
of the bulky Gln221 at the bottom of the groove.
When targeting a specific RGD integrin, one has to evaluate
selectivity versus other RGD integrins. It quickly appeared that
these compounds were inactive against
aIIbb3 (IC50 >10 lM, data
not shown). This can be explained by the distance difference high-
lighted on Figure 4. To assess the selectivity of our compounds
against
entiate selective from dual inhibitors.21 The A375 cell adhesion as-
say with fibrinogen reflects binding mediated solely through vb3.
avb3, we developed cell adhesion assays that could differ-
a
Our series showed some inhibition in this assay (Table 4). It should
During the course of the project, it appeared that our most po-
tent compounds were reaching the lower limit of the cell adhesion
assay run without human serum albumin (HSA). To better discrim-
inate between potent compounds, we developed the same assay
however be noted that the level of potency was at least 10 fold
lower than the Merck MK-0429
adhesion assay with fibronectin reflects binding mediated by both
5 and v integrins. The activity of our series for both 5b1 and vb3
was confirmed using the A375 cell adhesion assay in presence of
fibronectin without and with an vb3 inhibitor. Results showed
modest ratios for our series (1.6–5.9) suggesting dual 5b1/ vb3
inhibition. In comparison, Jerini’s compound was less potent
a
vb3 inhibitor.22 The A375 cell
a
a
a
a
with HSA (630 lM concentration). Since typical binding to albumin
across species was ꢀ99% in this zwitterionic series, clear differ-
ences were seen between the cell assays with and without serum,
as shown in Table 3. Compounds 9, 10 and 17 were difficult to rank
in the no serum assay, but 9 and 17 clearly appeared more potent
than 10 in the full serum assay.
a
a
a
against
versus
a
5b1 (K562 IC50 of 0.586
lM) but more selective for a5b1
a
vb3 (ratio of 15).

4116
B. Delouvrié et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4111–4116
7. (i) Jadhav, P. K.; Smallheer, J. M. WO1997/033887, 1997.; (ii) Smallheer, J.;
Table 4
RGD integrin selectivity dataa
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.
Compound
A375 cell adhesion
fibrinogen IC50
Mb
A375 cell adhesion
fibronectin IC50
Mb
vb3 inhibitor
,
, l
8. Stragies, R.; Osterkamp, F.; Zischinsky, G.; Vossmeyer, D.; Kalkhof, H.; Reimer,
U.; Zahn, G. J. Med. Chem. 2007, 50, 3786.
l
without/with
(ratio)c
a
9. (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.
2
6
9
11
12
17
18
0.084
0.007
0.001
0.049
0.35
0.023
0.017
1.2
2.5/0.61 (4.0)
0.23/0.097 (2.4)
0.009/0.002 (4.5)
0.86/0.21 (4.0)
8.4/1.4 (5.9)
0.25/0.15 (1.6)
0.27/0.13 (2.1)
3.2/0.20 (15)
10. (i) Tamaskar, I. Drugs Future 2010, 35, 16; (ii) Kuwada, S. K. Curr. Opin. Mol. Ther.
2007, 9, 92.
Jerini 1st
seriesd,e
MK-0429e
11. (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.
12. (i) Xiong, J. P.; Stehle, T.; Zhang, R.; Joachimiak, A.; Frech, M.; Goodman, S. L.;
Arnaout, M. A. Science 2002, 296, 151; (ii) Gottschalk, K.-E.; Kessler, H. Angew.
Chem., Int. Ed. 2002, 41, 3767; (iii) Tagaki, J.; Strokovich, K.; Springer, T. A.;
0.0003
0.81/2.2 (0.4)
a
b
c
All numbers are a geometric mean of 2 or more values.
Assay performed in presence of Mg2+
The Merck vb3 inhibitor MK-0429 was used at 1
3-[2-[1-(Benzyloxycarbonyl)-5(S)-(pyridin-2-ylaminomethyl)pyrrolidin-3(R)-
-alanine.
The Jerini and Merck compounds showed IC50s in the K562 cell adhesion assay
of 0.586 and 4.6 M respectively.
.
a
lM concentration.
d
yloxy]acetamido]-N-(2,4,6-trimethylbenzoyl)-
L
Waltz, T. Embo J. 2003, 22, 4607. An
a5b1 homology model was described by
e
Jerini in ref. 7iii.
l
13. Our a5b1 homology model was built using the package Prime from Schrödinger
(i) Jacobson, M. P.; Pincus, D. L.; Rapp, C. S.; Day, T. J. F.; Honig, B.; Shaw, D. E.;
Friesner, R. A. Proteins 2004, 55, 351; (ii) Jacobson, M. P.; Friesner, R. A.; Xiang,
Z.; Honig, B. J. Mol. Biol. 2002, 320, 597.
In conclusion, we have developed potent inhibitors of
which also carry activity against vb3. The homology model of
5b1 and the structural data available on vb3 and IIbb3 have
brought some rationale to the biological data while providing a
better understanding on key elements of selectivity across the
RGD integrins. SAR describing the influence of the core and the
amide is described in the corresponding part 2 of this letter.
a5b1
14. Davenport, R. J.; Munday, J. R. Drug Discovery Today 2007, 12, 569.
15. Cacciari, B.; Spalluto, G. Curr. Med. Chem. 2005, 12, 51.
16. Arnould, J.-C.; Delouvrié, B.; Ducray, R.; Lambert-Van Der Brempt, C. M. P.
WO2007/141473, 2007.
17. Leonard, K.; Pan, W.; Anaclerio, B.; Gushue, J. M.; Guo, Z.; DesJarlais, R. L.;
Chaikin, M. A.; Lattanze, J.; Crysler, C.; Manthey, C. L.; Tomczuk, B. E.; Marugan,
J. J. Bioorg. Med. Chem. Lett. 2005, 15, 2679.
18. (i) Urbahns, K.; Harter, M.; Albers, M.; Schmidt, D.; Stelte-Ludwig, B.;
Bruggemeier, U.; Vaupel, A.; Gerdes, C. Bioorg. Med. Chem. Lett. 2002, 12, 205;
(ii) Urbahns, K.; Harter, M.; Vaupel, A.; Albers, M.; Schmidt, D.; Bruggemeier,
U.; StelteLudwig, B.; Gerdes, C.; Tsujishita, H. Bioorg. Med. Chem. Lett. 2003,
1071, 13.
a
a
a
a
Acknowledgments
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 and
Dr. Kin Tam.
19. Scarborough, R. M.; Gretler, D. D. J. Med. Chem. 2000, 43, 3453.
20. (i) Xiao, T.; Takagi, J.; Coller, B. S.; Wang, J. H.; Springer, T. A. Nature 2004, 432,
59; (ii) Springer, T. A.; Zhu, J.; Xiao, T. J. Cell Biol. 2008, 182, 791.
21. 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
a
a
a
a
References and notes
a
a
a
1. Yang, T. J.; Rayburn, H.; Hynes, R. O. Development 1993, 119, 123.
2. Kim, S.; Bell, K.; Mousa, S. A.; Varner, J. A. Am. J. Pathol. 2000, 156, 1345.
3. Adachi, M.; Taki, T.; Higashiyama, M.; Kohno, N.; Inufusa, H.; Miyake, M. Clin.
Cancer Res. 2000, 6, 96.
4. Bhaskar, V.; Fox, M.; Breinberg, D.; Wong, M. H-L.; Wales, P. E.; Rhodes, S.;
DuBridge, R. B.; Ramakrishnan, V. Invest. New Drugs 2008, 26, 7.
5. Umeda, N.; Kachi, S.; Akiyama, H.; Zahn, G.; Vossmeyer, D.; Stragies, R.;
Campochiaro, P. A. Mol. Pharmacol. 2006, 1820, 69.
a
a
a
a
a
a
22. The Merck compound MK-0429 demonstrated Proof of Concept in a phase II
clinical trial against osteoporosis.
6. Chuntharapai, A.; Plowman, G.; Tessier-Lavigne, M.; Wu, Y.; Ye, W. WO2008/
060645, 2008.