Novel potent and?selective αvβ3αvβ5 integrin dual antagonists with reduced binding affinity for?human serum albumin

Article abstract of DOI:10.1016/j.ejmech.2006.03.008

The binding of lead compounds and drugs to human serum albumin (HSA) is a ubiquitous problem in drug discovery since it modulates the availability of the leads and drugs to their intended target, which is linked to biological efficacy. In our continuing e

Full text of DOI:10.1016/j.ejmech.2006.03.008

European Journal of Medicinal Chemistry 41 (2006) 847–861
http://france.elsevier.com/direct/ejmech
Original article
Novel potent and selective αvβ3/αvβ5 integrin dual antagonists
with reduced binding affinity for human serum albumin
Pierre Raboisson, Carl L. Manthey, Margery Chaikin, Jennifer Lattanze, Carl Crysler,
Kristi Leonard, Wenxi Pan, Bruce E. Tomczuk, Juan José Marugán *
Departments of Medicinal Chemistry, Structural Biology and Molecular Design and Informatics, Drug Discovery, Johnson &
Johnson Pharmaceutical Research and Development, L.L.C., 665 Stockton Drive, Exton, PA 19341, USA
Received in revised form 7 March 2006; accepted 9 March 2006
Available online 11 May 2006
Abstract
The binding of lead compounds and drugs to human serum albumin (HSA) is a ubiquitous problem in drug discovery since it modulates the
availability of the leads and drugs to their intended target, which is linked to biological efficacy. In our continuing efforts to identify small
molecule αVβ3 and αVβ5 dual antagonists, we recently reported indoles 2–4 as potent and selective αVβ3/αVβ5 antagonists with good oral bioa-
vailability profile. In spite of subnanomolar binding affinity of these compounds to human αVβ3 and αVβ5 integrins, high HSA binding (96.5–
97.3%) emerged as a limiting feature for these leads. Structure–activity HSA binding data of organic acids reported in the literature have demon-
strated that the incorporation of polar groups into a given molecule can dramatically decrease the affinity toward HSA. We sought to apply this
strategy by examining the effects of such modifications in both the central core constrain and the substituent β to the carboxylate. Most of these
derivatives were prepared in good yields through a cesium fluoride-catalyzed coupling reaction. This reaction was successful with a variety of
nitrogen-containing scaffolds (20, 33, and 43) and selected acetylenic derivatives (16, 19, and 34). Among the compounds synthesized, the 3-[5-
[2-(5,6,7,8-tetrahydro [1,8]naphthyridin-2-yl)ethoxy]indol-1-yl]-3-[5-(N,N-dimethylaminomethyl)-3-pyridyl]propionic acid (25) was found to be
the most promising derivative within this novel series with a subnanomolar affinity for both αvβ3 and αvβ5 (IC50 = 0.29 and 0.16 nM, respec-
tively), similar to our initial lead receptor antagonists 2–4, and exhibiting a low HSA protein binding (40% bound, Kd = 1.1 0.4 × 103 μM) and
an improved in vitro stability profile toward human and mouse microsomes (99.9% and 98.7% remaining after 10 min). Moreover, the selectivity
of 25 toward α5β1 and IIbIIIa integrins was perfectly maintained when compared to the parent leads 2–4. Thus, compound 25 was selected as a
new lead with improved drug-like properties for further evaluations in the field of oncology and osteoporosis.
© 2006 Elsevier SAS. All rights reserved.
Keywords: Integrin; αvβ3; αvβ5; Indole; Albumine; Protein binding
1. Introduction
cytoplasmic tail interacts with cytoskeletal filaments and pro-
teins to direct many important adhesive, migrational, prolifera-
tive and apoptotic events [1–3]. Particularly, the expression of
αvβ3/αvβ5 integrins, is known to be essential for the adherence
of newly-generated vascular cells to provisional extracellular
matrix, and their proliferation and survival [1–3]. Moreover,
it has been established that both αvβ3 and αvβ5 are expressed
most abundantly in tumor-associated blood vessels at sites of
neovascularization (vascular cell proliferation and vessel
sprouting) [4]. The recent expanded knowledge of the role of
intergrins in angiogenesis (the formation of new blood vessel)
has prompted the development of αvβ3/αvβ5 dual antagonists as
potential treatment of cancer because of their propensity to pre-
vent the “angiogenic switch”, which allows the endothelial
The integrin superfamily are membrane bound heterodi-
meric glycoproteins responsible for cell–cell and cell–matrix
interactions. Among the 26 different mammalian integrin re-
ceptors reported to date, αvβ3 and αvβ5 are the two most closely
related members: they contain the same αv subunit and their β
subunits are very similar [1]. The extracellular domain of αvβ3
and αvβ5 recognizes proteins with the tripeptide sequence argi-
nine–glycine–aspartic acid (RGD), e.g. vitronectin, whereas the
* Corresponding author.
E-mail address: Jmaruga1@prdus.jnj.com (J.J. Marugán).
0223-5234/$ - see front matter © 2006 Elsevier SAS. All rights reserved.
doi:10.1016/j.ejmech.2006.03.008

848
P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
cells to migrate, proliferate and invade the surrounding matrix.
Therefore, it has been suggested that an agent that is able to
block the adhesion function of αvβ3/αvβ5 receptors will render
vascular cells unresponsive to growth-promoting signals, there-
by inhibiting angiogenesis cascade and ultimately inhibit the
initial progression from the pre-malignant tumor to an invasive
cancer and the growth of dormant micrometastases into clini-
cally detectable metastatic lesions [4]. Recent works have va-
lidated this hypothesis, since the inhibition of αvβ3 and αvβ5
with monoclonal antibodies (LM-609) [5,6] or cyclic RGD
peptides (1, cilengitide) [7] was found to prevent angiogenesis
and tumor growth in several preclinical models [4–12]. Based
on this wide range of preclinical studies [5,6,9], clinical trials
have been initiated for the humanized version of LM609
(VitaxinTM) [5,6] as well as the cilengitide [7] in cancer ther-
apy [13].
Beside the implication of αvβ3 and αvβ5 in angiogenesis,
these integrins have been recognized for many years as attrac-
tive therapeutic targets for the treatment of thrombosis, vasooc-
clusion in sickle cell anemia, osteoporosis, diabetic retinopa-
thy, macular degeneration, and rheumatoid arthritis [2,9,11,
12,14–16]. Of these indications, osteoporosis seems to have
the most supporting evidence since RGD constraining peptides
and non-peptide αvβ3/αvβ5 antagonists have been shown to in-
hibit bone loss in the ovariectomized rat model [2,3].
In view of these promising results, many companies in-
itiated research to identify potent and selective non-peptidic
αvβ3/αvβ5 antagonists with improved oral bioavailability com-
pared to the parent drugs vitaxin [5,6] and cilengitide [7]. Be-
cause most of the compounds developed are RGD mimetics the
identification of non-prodrug αvβ3/αvβ5 antagonists with good
oral bioavailability remained an unmet objective for many
years due, in large part, to the polar nature of the zwitterionic
structures reported [3]. Recently, we and other have reported
non-peptidic αvβ3/αvβ5 antagonists with improved oral bioa-
vailable properties [3,17–27]. Although some of these novel
derivatives are potent αvβ3 and αvβ5 antagonists with nanomo-
lar-range activities, their efficacy in different preclinical studies
have been found to be limited when compared to vitaxin [5,6]
and the cilengitide [7]. This relative lack of potency in animal
models could, at least partially, be explained by the high affi-
nity of many integrin antagonists for human serum albumin
(HSA), which markedly affect the compound in vivo distribu-
tion and hinder them from reaching their target sites [28].
Thus, high daily doses are needed to achieve the desired free
blood concentrations, which can lead to undesirable side ef-
fects during acute or chronic therapy [24]. In this paper, we
report the design and synthesis of novel αvβ3/αvβ5 integrin dual
antagonists with reduced affinity for HSA, which may result in
compounds with significantly lower dosing levels and im-
proved in vivo tolerance.
2. Design of compound with reduced affinity for HSA
HSA is a highly soluble 66.5 kDa monomeric protein com-
prising a single chain of 585 residues organized in a series of
three structurally similar α-helical domains I–III, which are
further divided into subdomains A and B [29–32]. Beside its
important role in the maintenance of blood pH and colloidal
osmotic pressure [31], HSA possess an esterase-like activity
[33] and is also the major general transport protein in plasma
for a wide variety of endogenous and exogenous substances
with widely different structures. Thus, the effect of the binding
of several different categories of small molecules to bovine and
HSA has been studied for many years in both in vitro and in
vivo assays. In vitro, it was demonstrated that the receptor an-
tagonism activity of highly bound drugs can be decreased by
the addition of albumin [31]. In vivo, a high albumin protein
binding has been recognized to significantly impact the phar-
macokinetic and pharmacodynamic properties of drugs and
overall reduces their efficacy [31].
It is generally accepted that some substances bind to a lim-
ited number of specific high affinity binding sites on HSA,
while other (e.g. fatty acids) are bound less specifically in hy-
drophobic regions. The high affinity drug binding sites on
HSA have been classified into two well-characterized groups,
sites I and II [34]. Site I, classically associated with warfarin
and phenylbutazone binding, has been localized to subdomain
IIA. Site II, which binds diazepam and ibuprofen, is in subdo-
main IIIA [31]. Both sites (I and II) feature distinct hydropho-
bic and positively charged basic regions, consistent with their
binding of organic anions.
Within our current research program on the synthesis and
pharmacological evaluation of a variety of non-peptidic RGD
mimetics, we have recently reported a series of potent indole-
based αvβ3/αvβ5 integrin antagonists (e.g. compounds 2–4
Fig. 1. αvβ3 and αvβ5 integrin receptor antagonists.

P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
849
Fig. 1) with good oral bioavailability [17,18]. These com-
pounds possess low nanomolar range potencies (IC50-
αvβ3/αvβ5 = 1.0/0.68 nM, 0.25/0.21 nM and 0.38/0.50 nM, re-
spectively) with good selectivity toward αvβ3 and αvβ5 versus
α5β1 and IIbIIIa integrins. However, compounds 2–4 where
shown to be bind HSA 97.3%, 96.5%, and 97.2%, respec-
tively. This drawback adversely hampered the development of
these lead compounds as drug candidates. Thus, we became
interested in derivatives with lower affinity for HSA to im-
prove in vivo efficacy compared to the parent compounds.
Our approach was essentially qualitative, empirical and intui-
tive, and was guided by previous researches, which have effec-
tively illustrated the feasibility of altering the albumin protein
binding of some biological active agents.
of the propionic chain could dramatically affect the affinity of
the resulting molecules for HSA. Such an approach has already
been successfully applied to various organic acid drugs (Fig. 2).
For instance, the introduction of a more polar region (pipera-
zine moiety and an additional nitrogen in position 8) in the
structure of the highly albumin bound (92–97%) nalidixic acid
(6) resulted in the discovery of pipemidic acid (7), which ex-
hibited no significant HSA binding [36]. Similar results were
obtained with the analogues of diflunizal (compounds 8–11)
[37]. In view of these promising results, we decided to apply
the same strategy to our indole-based leads. Thus, compounds
25, 26, 38, 39, 42, and 49 (Table 1) have been synthesized and
evaluated for their affinity on both HSA and αvβ3/αvβ5 integ-
rins.
The chemical structures of our lead compounds (2–4, Fig. 1)
possess the same common feature of other known αvβ3/αvβ5
integrin antagonists (Fig. 1) which could be separated in three
pharmacophoric regions: (i) a hydrophobic area (indole bi-
cycle) bearing (ii) a guanidinium-like moiety (the tetrahydro
[1,8]naphthyridine) and (iii) a carboxylate group (propionate
moiety). The substitution of the propionate moiety in the β-po-
sition with an aryl or an heteroaryl group such as phenyl, 3-
pyridyl and 3,4-(methylenedioxy)phenyl (compounds 2–4, re-
spectively) has been shown to improved the antagonist potency
on αvβ3/αvβ5 integrins [3,18]. In the present case, it can be
assumed that both the negatively charged carboxylate and the
positively charged guanidine-like moiety of the integrin an-
tagonists probably interact electrostatically with charged resi-
dues of HSA, while the aromatic rings may participate in a
hydrophobic or specific stacking interaction with other hydro-
phobic and aromatic residues present in the HSA sequence.
Since it has been demonstrated that both the basic and the
acidic endings are required elements to achieve high affinity
for αvβ3/αvβ5 integrins [3], we decided to maintain this part
of the molecule unchanged, while exploring the possibility of
modifying the central molecular core (indole ring) as well as
the substitution in the β-position of the carboxylic acid side
chain (Fig. 1).
3. Chemical synthesis
The synthetic strategies used for the preparation of the target
compounds 25, 26, 38, 39, 42, and 49 are summarized in
Schemes 1–7.
The acetylenic 16 and 19 represent the starting material for
the synthesis of compounds 25 and 26 (Scheme 3). The ethyl
3-[5-(N,N-dimethylaminomethyl)-3-pyridyl]propiolate 16 was
prepared in four steps from the dibromopyridine 12 according
to reaction sequence depicted in Scheme 1. The reaction of 12
with lithium tributylmagnesium in toluene at low temperature
followed by the treatment with dimethylformamide afforded
the desired aldehyde 13 in 83% yield. Reductive amination
using dimethylamine and sodium cyanoborohydride in metha-
nol followed by the palladium-mediated cross-coupling reac-
tion of the bromo intermediate 14 and ethyl orthopropiolate,
gave the 3,3,3-triethoxypropynylpyridine 15, which was di-
rectly converted to the ethyl ester 16 (72% yield) by treatment
with diluted hydrochloric acid in acetonitrile. Similarly, the 4-
(N,N-dimethylamino)-4-oxotetrolic acid ethyl ester 19 was pre-
pared in 34% overall yield (Scheme 2) by palladium-mediated
coupling of dimethylcarbamyl chloride and the readily avail-
able ethyl orthopropiolate 17, followed by the hydrolysis of
the so-formed orthoester 18.
Because it is known that a combination of hydrophobic re-
gion and a negative charge is favorable for strong binding of
many acidic substance in both site I and II [35], it seemed
reasonable that the incorporation of more hydrophilicity in
both the central constraint (indole core) and in the β-position
The 1,5-disubstituted indole derivatives (25 and 26) were
prepared by treatment of the corresponding t-BOC-protected
5,6,7,8-tetrahydro [1,8]naphthyridyl indole 20 [18,38], with
the corresponding 3-substituted ethyl propiolates 16 and 19 in
Fig. 2. Chemical modifications on organic acids leading to significant lower HSA binding.

850
P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
Table 1
Integrin and HSA binding affinities for selected αvβ3/αvβ5 dual antagonists.
For esthetic and homogeneity reasons (cf. Table 2), can we move down the “Integrin IC50 (nM)a” and “Binding to HSAb” in Table 1?
Integrin IC50 (nM)a
Binding to HSAb
c
Compounds
2 [18]
3 [18]
X
O
O
O
R1
H
H
R2
Ph
3-pyridyl
αvβ3
1.0
0.25
0.38
αvβ5
0.68
0.21
0.50
α5β1
41
15
IIbIIIa
%
Kd (μM)
> 1000 97.3
> 1000 96.5
> 1000 97.2
18.7 0.32
24.6 2.1
19.8 0.9
4 [18]
H
0.024
51 [18]
25
CH2
O
H
H
Ph
23
0.29
15
0.16
1.1
5.3
> 1000
> 1000 40
–
–
(1.1 0.4).103
26
39
a
O
O
H
(CH3)2NCO-
Ph
37
190
21
45
6600
–
> 1000 81.3
> 1000 74
155
236 19.8
3
(CH3)2NCH2
The binding IC50 on integrins was determined by an αvβ3-vitronectin ELISA assay as previously reported in [39,40].
HSA binding (% of bound values and Kd) was determined by BIACORE SPR biosensors technology with full length HAS as previously reported in [41].
Percentage of bound values.
b
c
presence of cesium fluoride (Scheme 3). Palladium-catalyzed
hydrogenation of olefins 21 and 22 in presence of lithium hy-
droxide afforded the free acid indole derivatives 23 and 24 in
99% and 97% yield, respectively, as outlined in Scheme 3.
Subsequent removal of the t-BOC-protecting group with cop-
per(I) trifluoromethanesulfonate at 130 °C gave the final indole
products 25 and 26 in 95% and 58% yield, respectively.
The benzimidazole 38 was prepared according to the proce-
dure reported in Scheme 4. The nitro aniline 27 was first N-
diprotected leading to the tri-t-BOC aniline 28, which was se-
lectively O-deprotected using N,N-diethylethylenediamine. A
Mitsunobu reaction of phenol 29 with alcohol 30 [38], afforded
nitroaniline 31 in 51% yield. Selective deprotection of the ani-
line moiety using trifluoroacetic acid (reaction monitored by
TLC), followed by the treatment of free aniline 32 with formic
acid and acetic anhydride yielded N-formyl intermediate 33.
Subsequent alkylation of 33 with propiolate 34 using the ce-
sium fluoride procedure afforded olefin 35. Reduction of the
nitro group of compound 35 using iron in acetic acid at
80 °C afforded the cyclic benzimidazole 36 in 44% yield,
Scheme 1. Synthesis of propiolate 10. (a) i) n-Bu3MgLi, toluene, –10 °C; ii) DMF, 0 °C; (b) Me2NH, NaBH3CN, MeOH, r.t.; (c) HC≡C(OEt)3, PdCl2(PPh3)2, CuI,
Et3N, 55 °C; (d) HCl, CH3CN, r.t.
Scheme 2. Synthesis tetrolic ester 20. (a) PdCl2(PPh3)2, CuI, Et3N, 55 °C; (b) HCl, EtOH, r.t.

P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
851
Scheme 3. Synthesis of indoles 25 and 26. Reagents and conditions: (a) 16 or 19, CsF, DMF, 40–50 °C; (b) LiOH·H2O, Pd/C, H2, THF/MeOH/H2O, r.t.; (c) Cu(I)
triflate, toluene/DMF, 130 °C.
Scheme 4. Synthesis of benzimidazole 38. (a) (t-BOC)2O, DMAP, CH3CN, r.t.; (b) NH2(CH2)2N(C2H5)2, CH3CN, r.t.; (c) DIAD, PPh3, THF, 0 °C; (d) TFA,
CH2Cl2, r.t.; (e) HCOOH, (CH3CO)2O, THF, 50 °C; (f) CsF, DMF, 80 °C; (g) Fe, AcOH, 80 °C; (h) SmI2, HMPA, EtOH, THF; (i) LiOH·H2O, MeOH, THF, r.t.
which was successively reduced with samarium iodide and sa-
ponified to give the final target compound 38.
the reduction of intermediate 40 [17] with sodium cyanoboro-
hydride followed by the saponification of the ester 41 with
lithium hydroxide.
The 7-deazahypoxanthine derivative 49 was synthesized ac-
cording to the procedure shown in Scheme 7 starting from the 6-
chloro-7-deazapurine 43. Treatment of 43 with ethyl phenylpro-
piolate in presence of cesium fluoride followed by the hydroly-
The conversion of the previously reported indole 2 [18] to
the 3-(N,N-dimethylaminomethyl) 39 was readily achieved by
treatment of 2 with methyleneimmonium iodide in dichloro-
methane at room temperature as shown in Scheme 5, whereas
the 2,3-dihydroindole 42 was prepared in 33% overall yield, by

852
P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
Scheme 5. Synthesis of 3-(dimethylaminomethyl)indole 39. (a) H2C=N+(CH3)2, CH2Cl2, r.t., 12 h.
Scheme 6. Synthesis of 2,3-dihydroindole 42. (a) NaBH3CN, MeOH, r.t.; (b) LiOH·H2O, Pd/C, H2, THF/MeOH/H2O, r.t.
Scheme 7. Synthesis of 7-deazahypoxanthine 49. (a) Ethylphenylpropiolate, CsF, DMF, 75 °C; (b) HCl, H2O/EtOH, 37 °C; (c) 50, DIAD, PPh3, THF, r.t.; (d) Pd/C,
H2, MeOH, r.t.; (e) Cu(I)triflate, toluene/DMF, microwave, 130 °C; (f) LiOH·H2O, THF/MeOH/H2O, r.t.
sis of the 1,6-iminochloride 44 afforded deazahypoxanthine 45.
The introduction of the tetrahydronaphthyridyl moiety was
achieved by a Mitsunobu reaction between 45 and alcohol 50
[38] leading to intermediate 46 in 64% isolated yield. Reduction
of 46 by hydrogenation followed by the deprotection of the t-
BOC group with trifluoroacetic acid and the saponification of
the ester moiety of 48 furnished the target compound 49.
IC50 (αvβ3/αvβ5) = 0.25/0.21 nM) or a 3,4-methylenedioxyphenyl
moiety (4, IC50 (αvβ3/αvβ5) = 0.38/0.50 nM) at the β-position of
the propionic acid chain strongly improved the affinity of the
resulting compounds for both αvβ3 and αvβ3 integrins (Table 1).
However, these transformations did not significantly modify the
binding to HSA (96.5–97.3% HSA bound). To more dramati-
cally increase the polarity in the β-position and evaluate the ef-
fect on protein binding, a N,N-dimethylaminomethyl function
was introduced at the position 5 of the β-pyridyl moiety of in-
dole 3 leading to compound 25. To our satisfaction, this mod-
ification strongly diminished the HSA bound fraction (40%)
while maintaining the potency toward both αvβ3 and αvβ3 integ-
rins (25, IC50 (αvβ3/αvβ5) = 0.29/0.16 nM). Attempt to apply the
same strategy at the 3-position of the indole ring (compound 39)
significantly reduced the protein binding but also dramatically
decreased the affinity for αvβ3 and αvβ3 (39, IC50 (αvβ3/αvβ5)
= 190/45 nM). The replacement of the phenyl ring at the β-posi-
tion with a polar group was also envisioned to decrease the
HSA binding of this series of αvβ3 antagonists. The N,N-di-
methylaminocarbonyl moiety appeared to be a very interesting
candidate for the replacement of the β-phenyl ring of 2 because
it shares several desired molecular characteristics: (i) it is a polar
4. Pharmacology, results and discussion
All synthesized compounds were evaluated for their affinity
for αVβ3 and αVβ5 integrin receptors (ELISA binding assay)
[39,40] and for their affinity for HSA (BIACORE Surface Plas-
mon Resonance (SPR) biosensors technology) [41]. Structure
and binding data are shown in Tables 1 and 2. In our hands, the
reference compounds 52 (IC50 (αvβ3/αvβ5) = 0.19/0.18 nM) and
5 [42] (IC50 (αvβ3/αvβ5) = 1.5/9.6 nM), were found highly HAS
bound (87.9% and 99.6%, respectively), whereas 1 [7]
(IC50 (αvβ3/αvβ5) = 0.86/2.1 nM) was found to be exempt of af-
finity toward HSA (less than 3.3% bound).
As reported previously in [3,18], the replacement of the phe-
nyl ring of 2 (IC50 (αvβ3/αvβ5) = 1.0/41.7 nM) with a 3-pyridyl (3,

P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
853
Table 2
Integrin and HSA binding affinities for selected αvβ3/αvβ5 dual antagonists.
Compounds
Formula
Integrin IC50 (nM)a
Binding to HSAb
Kd (μM)
> 2000
c
αvβ3
0.86
0.17
1.5
αvβ5
α5β1
14
IIbIIIa
> 1000
–
%
1 (Cilengitide) [7]
Echistatin
5 [42]
2.1
12
9.6
< 3.3
–
99.6
1.4
2.2
–
> 1000
2.5 0.5
52 [3]
0.19
13.6
0.18
10
3.5
> 1000
> 1000
87.9
82.9
92.1 1.68
139
38
> 10000
49
590
1300
>10000
> 1000
95.8
30.0 0.5
53 [17]
470
800
13000
1200
> 10000
14000
> 1000
> 1000
99.2
93
5.51 0.06
42
49
1
a
The binding IC50 on integrins was determined by an αvβ3-vitronectin ELISA assay as previously reported in [39,40].
HSA binding (% of bound values and Kd) was determined by BIACORE SPR biosensors technology with full length HAS as previously reported in [41].
Percentage of bound values.
b
c
functional group, (ii) it does not contain hydrogen bond donat-
ing moiety, and (iii) it has no net charge. For these common
characteristics, the N,N-dimethylaminocarbonyl moiety has
been fruitfully used in the design and identification of novel
surfaces resistant to protein adsorption [43]. Unfortunately,
although the replacement of the phenyl ring at the β-position
with the N,N-dimethylaminocarbonyl group significantly de-
creased the HSA binding (unbound fraction of 2 and
26 = 2.7% and 18.7%, respectively), which was our primary
goal, this modification was also associated with a loss of affinity
for both αVβ3 and αVβ5. Finally, modifications of the central
indole template by introducing heteroatoms (benzimidazole 38
and 7-deazahypoxanthine 49), or by increasing the polarity
throughout the saturation of the 2,3-indole bond (42) resulted
in a loss of binding potency on both HSA and αv intergrins
(compare 38 to 4, 49 to 51 and 42 to 53).
implicated in essential phenomena such as blood coagulation
[44], selectivity of αV antagonists is essential for their develop-
ment as drugs. Thus, the selectivity for αVβ3 and αVβ5 versus
α5β1 and IIbIIIa integrins was determined. In the α5β1-ELISA
assay compound 25 was found to be approximately 20-fold
less active for α5β1 than for αVβ3 and αVβ5 and was found
completely inactive against the fibrinogen receptor IIbIIIa.
Moreover, to predict first pass metabolism [45], compound
25 was tested in vitro using liver microsomes from human
and mouse and was found to be perfectly stable (99.9% and
98.7% remaining after 10 min, respectively). Thus, indole 25
appeared to be significantly more stable than the parent lead 3
toward human and mouse liver microsomes (75.1% and 55.2%
remaining after 10 min, respectively).
5. Conclusion
With the above promising data, which have demonstrated
the possibility of reducing the HSA binding affinity while
maintaining the potency toward αVβ3 and αVβ5 integrins, we
selected compound 25 for further biological evaluation. Be-
cause other integrins like the fibrinogen receptor IIbIIIa are
The main objective of the present study was the design and
the synthesis of potent and selective indole-based αvβ3 and
αvβ5 dual antagonists with reduced HSA protein binding.
Using an intuitive and empirical approach, a series of indole

854
P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
and indole bioisoster RGD mimetics have been synthesized
and evaluated for their affinity on both αv integrins and HSA.
Among the compounds synthesized, the 3-[5-[2-(5,6,7,8-tetra-
hydro [1,8]naphthyridin-2-yl)ethoxy]indol-1-yl]-3-[5-(N,N-di-
methylaminomethyl)-3-pyridyl]propionic acid (25) was found
to be the most promising derivative within this novel series
exhibiting a subnanomolar affinity for both αvβ3 and αvβ5
(IC50 = 0.29 and 0.16 nM, respectively) similar to our initial
leads 2–3. Moreover, 25 exhibited an unbound HSA-fraction
of 60% and an optimized profile of selectivity toward α5β1
and IIbIIIa integrins associated with an excellent in vitro stabi-
lity profile toward human and mouse microsomes. Thus, indole
25 has been selected for further in vitro and in vivo biological
evaluations as part of our continuing effort to discover non-
peptidic αVβ3 antagonists that may be administered orally in
human for the treatment and prevention of osteoporosis and
cancer.
tion of 6K; the ions were produced in a fast atom bombardment
source at 8 kV. Linear voltage scans were collected to include
the sample ion and two poly(ethylene glycol) ions, which were
used as internal reference standards.
6.1.2. 5-Bromopyridine-3-carboxaldehyde (13)
A solution of n-BuLi (20.0 ml, 2.5 M in hexanes) was
added under argon to toluene (117 ml). The mixture was
cooled to –10 °C and n-butylmagnesium chloride (13.0 ml,
2 M in Et2O) was added at –10 °C over 5 min. After 30 min,
a solution of 3,5-dibromopyridine (15.6 g, 66.0 mmol) in to-
luene (32 ml) was added at –10 °C over 20 min and the result-
ing solution was stirred for 3 h at the same temperature. After
3 h, DMF (33.0 ml, 426 mmol) was added at –10 °C over
5 min. Then the mixture was quenched with an ice-cold solu-
tion of citric acid (19.0 g, 99 mmol in 35 ml of H2O) and water
(20 ml). The mixture was extracted with ethyl acetate
(3 × 50 ml). The combined organic extracts were washed with
water (2 × 50 ml) and saturated sodium bicarbonate (50 ml),
dried (Na2SO4), and concentrated to give a yellow oil. Purifi-
cation by column chromatography (petroleum ether/ether; 5:1)
afforded the target product 13 as a brown solid (10.3 g, 83%).
1H NMR (400 MHz, CDCl3) δ 8.31 (s, 1H, ArH), 8.89 (s, 1H,
ArH), 9.2 (s, 1H, ArH), 10.12 (s, 1H, CH); m/z 256 (M + H)+.
6. Experimental section
6.1. Chemical synthesis
6.1.1. General
Reagents used for the synthesis were purchased from Sig-
ma-Aldrich (Milwaukee, WI, USA) and Lancaster (Windham,
NH, USA). All solvents were obtained from commercial sup-
pliers and used without further purification. Flash chromatogra-
phy was performed on Geduran® Silica gel Si 60 (40–63 μm,
Merck). Thin-layer chromatography was carried out using
plates Silica gel 60 F254 (Merck). The spots were visualized
either under UV light (λ = 254 nm) or by spraying with molyb-
date reagent (H2O/concentrated
6.1.3. 3-Bromo-5-(N,N-dimethylaminomethyl)pyridine (14)
A solution of 13 (10.2 g, 54.8 mmol), acetic acid (100 ml),
and dimethylamine (0.5 M in MeOH, 700 ml) was stirred at
room temperature for 18 h. Then, sodium cyanoborohydride
(4.17 g, 65.8 mmol) was added in portions over 5 min. After
5 h, the reaction was cooled to 0 °C and quenched with an ice-
cold solution of sodium hydroxide (2 N in H2O, 475 ml). The
resulting solution was extracted with ethyl acetate. The organic
layer was successively washed with brine, dried (Na2SO4), and
concentrated to dryness under reduced pressure. Then, the re-
sidue was partitioned between dichloromethane (100 ml) and
1 N hydrochloric acid in water (70 ml). The aqueous layer was
adjusted to pH 9 with 2 N sodium hydroxide and extracted
with CH2Cl2. The organic layer was dried (Na2SO4), then con-
centrated to dryness under reduced pressure to give compound
14 (5.75 g, 40%) as a colorless oil. 1H NMR (400 MHz,
CDCl3) δ 2.25 (s, 6H, 2 CH3), 3.41 (s, 2H, CH2), 7.85 (s,
1H, ArH), 8.43 (d, J = 1.5, 1H, ArH), 8.58 (d, J = 1.5, 1H,
ArH); m/z 283 (M + H)+.
H2SO4/(NH4)6Mo7O24·4H2O/(NH4)2–Ce(SO4)4·2H2O,
90:10:25:1, v/v/w/w) and charring at 140 °C for a few minutes.
All chemical yields are unoptimized and generally represent
the result of a single experiment.
1H NMR were recorded on
a Bruker B-ACS-120
(400 MHz) spectrophotometer at room temperature. Chemical
shifts are given in ppm (δ), coupling constants (J) are in Hertz
(Hz) and signals are designated as follows: s, singlet; d, doub-
let; t, triplet; q, quadruplet; quint., quintuplet; m, multiplet; br
s, broad singlet. The LC–MS/MS were determined using an
API-2000 triple quadrapole mass spectrometer using the Tur-
boIonspray source, a Shimadzu LC-10ADvp pumping system,
a CTC/Leap HTS PAL autosampler, and a C8 LC column
(Princeton Chromatography, 5 u, 50 × 3.0 mm). The LC–MS
data were recorded on a Waters ZQ electrospray mass spectro-
meter equipped with 4-channel MUX capabilities (Milford,
MA) with ELS detection using a Princeton SPHER HTS
60 Å, 5 μm column (3 × 50 mm) Princeton Chromatography
(Cranbury, NJ). Two mobile phases (A: 99.85% water, 0.1%
formic acid, 0.05% TFA; B: 99.9% acetonitrile, 0.1% formic
acid, 0.05% TFA) were employed as a gradient from 10% B to
100% B in 4 min with a flow rate of 1.2 ml/min. Accurate
mass determination was performed on an Autospec E high-re-
solution magnetic sector mass spectrometer tuned to a resolu-
6.1.4. Ethyl 3-[5-(N,N-dimethylaminomethyl)-3-pyridyl]
propiolate (16)
A mixture of 14 (5.75 g, 26.7 mmol), 3,3,3-triethoxypro-
pyne (12.1 g, 70 mmol), copper(I) iodide (510 mg, 2.68 mmol)
and dichlorobis(triphenylphosphine)palladium(II) (938 mg,
1.34 mmol) in triethylamine (30 ml) was heated at 55 °C for
36 h under nitrogen. The mixture was successively cooled to
room temperature, diluted with ethyl acetate (50 ml), filtered,
and concentrated under reduced pressure. Purification by flash
chromatography (heptane/EtOAc/MeOH, 3:2:0.2) afforded
5.95 g of a yellow oil. This oil was dissolved in acetonitrile

P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
855
(50 ml) and a solution of 0.1 N hydrochloric acid in acetonitrile
(212 ml) was added over 5 min, under argon. After 30 min, the
reaction mixture was quenched with potassium carbonate
(5.0 g, 36.2 mmol), filtered, and evaporated to afford 16
(quint, J = 6.2, 2H, CH2), 2.21 (s, 6H, 2 CH3), 2.73 (t, J = 6.6,
2H, CH2), 3.20 (t, J = 6.9, 2H, CH2), 3.46 (s, 2H, CH2), 3.76
(t, J = 4.0, 2H, CH2), 4.11 (q, J = 7.1, 2H, CH2), 4.39 (t,
J = 4.7, 2H, CH2), 6.24 (s, 1H, CH), 6.52 (d, J = 3.3, 1H,
ArH), 6.72 (d, J = 1.3, 1H, ArH), 6.78–6.88 (m, 1H, ArH),
6.93 (d, J = 7.6, 1H, ArH), 7.05 (d, J = 3.4, 1H, ArH), 7.09–
7.13 (m, 1H, ArH), 7.32 (d, J = 7.6, 1H, ArH), 7.64 (br s, 1H,
ArH), 8.59 (dd, J = 21.1, J = 2.0, 2H, ArH); m/z 626 (M + H)+.
1
(4.2 g, 72%) as a yellow oil. H NMR (400 MHz, CDCl3) δ
1.36 (t, J = 7.1, 3H, CH3), 2.25 (s, 6H, 2 CH3), 3.44 (s, 2H,
CH2), 4.31 (q, J = 7.13, H, CH2), 7.80 (s, 1H, ArH), 7.88 (s,
1H, ArH), 8.50 (s, 1H, ArH); m/z 369 (M + H)+.
6.1.5. 4,4,4-Triethoxy-N,N-dimethyltetrolamide (18)
6.1.8. (Z,E)-3-[5-[2-[8-tert-Butyloxycarbonyl(5,6,7,8-
tetrahydro [1,8]naphthyridin-2-yl)] ethoxy]indol-1-yl]-4-oxo-
4-(N,N-dimethylamino)crotonic acid ethyl ester (22)
A mixture of dichlorobis(triphenyl phosphine)palladium(II)
(1.78 g, 2.5 mmol), dimethylcarbamyl chloride (5.37 g,
50 mmol), 3,3,3-triethoxypropyne (10 ml, 53 mmol), and CuI
(0.95 g, 5 mmol) in triethylamine (60 ml) was heated at 55 °C
for 3 h. The mixture was cooled at room temperature, diluted
with diethyl ether (300 ml) and filtered and concentrated under
reduced pressure. Purification by column chromatography (pet-
roleum ether/EtOAc, 1:1) afforded othoformate 18 as a yel-
A mixture of 20 (3.42 g, 8.7 mmol), 19 (2.5 g, 14.8 mmol),
and cesium fluoride (2.25 g, 14.8 mmol) in anhydrous di-
methylformamide (8.2 ml) was stirred under nitrogen at
40 °C for 2 h. The mixture was cooled to room temperature,
then diluted with ethyl acetate (40 ml) and washed with water
(3 × 50 ml). Purification by column chromatography (heptane/
EtOAc/MeOH, 2:1:0.15) gave 22 (3.43 g, 70%) as white solid
1
low–brown oil (6.19 g, 48%). H NMR (400 MHz, CDCl3) δ
1
1.22 (t, J = 7.3, 9H, 3 CH3), 3.05 (s, 3H, CH3), 3.24 (s, 3H,
(mixture of E/Z derivatives 3/1): H NMR (400 MHz, CDCl3)
CH3), 7.71 (q, J = 7.3, 6H, 3 CH2); m/z 244 (M + H)+.
(E)-22 δ 1.29 (t, J = 7.1, 3H, CH3), 1.52 (s, 9H, 3 CH3) 1.91–
1.97 (m, 2H, CH2), 2.73 (t, J = 6.6, 2H, CH2), 2.87 (s, 3H,
CH3), 3.14 (s, 3H, CH3), 3.19-3.24 (m, 2H, CH2), 3.76 (t,
J = 6.0, 2H, CH2), 4.23 (q, J = 7.1, 2H, CH2), 4.40 (t,
J = 7.0, 2H, CH2), 6.14 (s, 1H, CH), 6.57 (d, J = 3.6, 1H,
ArH), 6.93 (s, 1H, ArH), 6.96 (s, 1H, ArH), 7.08–7.12 (m,
1H, ArH), 7.20 (d, J = 3.6, 1H, ArH), 7.33 (d, J = 7.6, 1H,
ArH), 7.60 (d, J = 9.1, 1H, ArH); (Z)-22 δ 1.12 (t, J = 7.1, 3H,
CH3), 1.52 (s, 9H, 3 CH3) 1.91–1.97 (m, 2H, CH2), 2.46 (s,
3H, CH3), 2.73 (t, J = 6.6, 2H, CH2), 2.95 (s, 3H, CH3), 3.19–
3.24 (m, 2H, CH2), 3.76 (t, J = 6.0, 2H, CH2), 4.11 (q, J = 7.1,
2H, CH2), 4.38 (t, J = 7.0, 2H, CH2), 6.00 (s, 1H, CH), 6.55 (d,
J = 3.6, 1H, ArH), 6.82–6.93 (m, 2H, ArH), 7.08–7.12 (m, 1H,
ArH), 7.20 (d, J = 3.6, 1H, ArH), 7.28 (d, J = 3.5, 1H, ArH),
7.60 (d, J = 9.1, 1H, ArH); m/z 563 (M + H)+.
6.1.6. 4-(N,N-Dimethylamino)-4-oxotetrolic acid ethyl ester
(19)
A solution of 18 (6.14 g, 25 mmol) and hydrochloric acid
(0.05 N in ethanol, 64 ml) was stirred at room temperature for
5 h. Then, potassium carbonate (5 g, 36.1 mmol) was added.
After 30 min, the reaction mixture was filtered, washed with
ethanol (10 ml) and concentrated to afford a brown–yellow oil,
which was purified by flash chromatography (petroleum
ether/Et2O, 1:1) to give compound 19 (3.3 g, 70%) as a yellow
1
oil. H NMR (400 MHz, CDCl3) δ 1.31 (t, J = 7.3, 3H, CH3),
3.03 (s, 3H, CH3), 3.22 (s, 3H, CH3), 4.28 (q, J = 7.3, 2H,
CH2); m/z 170 (M + H)+.
6.1.7. (Z,E)-3-[5-[2-[8-tert-Butoxycarbonyl(5,6,7,8-tetrahydro
[1,8]naphthyridin-2-yl)] ethoxy]indol-1-yl]-3-[5-(N,N-
6.1.9. 3-[5-[2-[8-tert-Butoxycarbonyl-(5,6,7,8-tetrahydro [1,8]
naphthyridin-2-yl)]ethoxy] indol-1-yl]-3-[5-(N,N-
dimethylaminomethyl)-3-pyridyl]acrylic acid ethyl ester (21)
A mixture of 20 [18,38] (4.2 g, 18.1 mmol), 16 (20.0 g,
50.8 mmol), and cesium fluoride (6.75 g, 44.4 mmol) in anhy-
drous dimethylformamide (20 ml) was stirred under nitrogen at
40 °C for 2 h. The mixture was cooled to room temperature,
then diluted with ethyl acetate (100 ml) and washed with water
(3 × 100 ml). Purification by column chromatography (hep-
tane/EtOAc/MeOH, 2:1:0.15) gave compound 21 (4.8 g,
72%) as white solid (mixture of E/Z derivatives 3/1): 1H
NMR (400 MHz, CDCl3): (E)-21 δ 1.04 (t, J = 7.1, 3H,
CH3), 1.52 (s, 9H, 3 CH3), 1.94 (quint, J = 6.2, 2H, CH2),
2.18 (s, 6H, 2 CH3), 2.73 (t, J = 6.6, 2H, CH2), 3.20 (t,
J = 6.9, 2H, CH2), 3.39 (s, 2H, CH2), 3.76 (t, J = 4.0, 2H,
CH2), 4.03 (q, J = 7.1, 2H, CH2), 4.36 (t, J = 4.7, 2H, CH2),
6.24 (s, 1H, CH), 6.59 (d, J = 3.3, 1H, ArH), 6.72 (d, J = 1.3,
1H, ArH), 6.93 (d, J = 7.6, 1H, ArH), 7.05 (d, J = 3.4, 1H,
ArH), 7.09–7.13 (m, 2H, ArH), 7.32 (d, J = 7.6, 1H, ArH),
7.51 (br s, 1H, ArH), 8.53 (dd, J = 31.3, J = 1.9, 2H, ArH);
(Z)-21 δ 1.16 (t, J = 7.1, 3H, CH3), 1.52 (s, 9H, 3 CH3), 1.94
dimethylaminomethyl)-3-pyridyl]propionic acid (23)
A mixture of 21 (4.0 g, 7.3 mmol), lithium hydroxide
monohydrate (504 mg, 12.0 mmol), and 10% Pd/C (2.0 g) in
tetrahydrofuran/methanol/water, 2:1:0.1 (50 ml) was shaken in
a hydrogenation apparatus under 45 psi pressure at room tem-
perature for 36 h. The catalyst was removed by filtration,
washed with tetrahydrofuran, and the filtrate was concentrated
to dryness. The residue was dissolved in water (50 ml) and the
pH was adjusted to 3 with HCl 1 N. The solution was extracted
with ethylacetate (3 × 100 ml), dried (Na2SO4), and evaporated
to give compound 23 (3.8 g, 99%) as a white powder. 1H NMR
(400 MHz, DMSO-d6) δ 1.41 (s, 9H, 3 CH3), 1.79–1.84 (m,
2H, CH2), 2.05 (s, 6H, 2 CH3), 2.69 (t, J = 6.6, 2H, CH2),
2.98–3.08 (m, 4H, 2 CH2), 3.31 (s, 2H, CH2), 3.62 (t,
J = 6.00, 2H, CH2), 4.28 (t, J = 6.88, 2H, CH2), 6.00 (t,
J = 7.48, 1H, CH), 6.35 (d, J = 3.1, 1H, ArH), 6.67 (dd,
J = 8.9, J = 2.4, 1H, ArH), 6.98–7.03 (m, 2H, ArH), 7.28 (d,
J = 9.0, 1H, ArH), 7.43 (d, J = 7.6, 1H, ArH), 7.51 (br s, 1H,

856
P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
ArH), 7.61 (d, J = 3.2, 1H, ArH), 8.27 (d, J = 1.9, 1H, ArH),
8.37 (d, J = 1.9, 1H, ArH); m/z 600 (M + H)+.
to dryness under reduced pressure. The crude material was pur-
ified by column chromatography on silica
(CH2Cl2/MeOH/NH4OH, 7:1:0.1) yielded compound 26
1
(1.14 g, 58%) as white powder: H NMR (400 MHz, CDCl3)
6.1.10. 3-[5-[2-[8-tert-Butyloxycarbonyl(5,6,7,8-tetrahydro
[1,8]naphthyridin-2-yl)]ethoxy] indol-1-yl]-4-(N,N-
dimethylamino)-4-oxobutyric acid (24)
δ 1.90 (quint, J = 5.7, 2H, CH2), 2.72 (t, J = 6.0, 2H, CH2),
2.80–2.93 (m, 7H, 2 CH3, CH), 3.02 (t, J = 5.9, 2H, CH2),
3.21-3.27 (m, 1H, CH), 3.38–3.48 (m, 2H, CH2), 4.14 (t,
J = 6.0, 2H, CH2), 5.69–5.75 (m, 1H, CH), 6.40 (d, J = 3.2,
1H, ArH), 6.48 (d, J = 7.3, 1H, ArH), 6.72 (dd, J = 8.9,
J = 2.3, 1H, ArH), 6.99 (d, J = 2.3, 1H, ArH), 7.17
(d, J = 3.2, 1H, ArH), 7.29–7.32 (m, 2H, ArH); m/z 437
(M + H)+. HRMS (ESI) m/z calcd for C24H28N4O4
436.21106, found 437.218881 (M + H+).
A mixture of 22 (2.7 g, 4.8 mmol), lithium hydroxide
monohydrate (252 mg, 6.0 mmol), and 10% Pd/C (2.55 g) in
tetrahydrofuran/methanol/water, 2:1:0.1 (64 ml) was shaken in
a hydrogenation apparatus under 45 psi pressure at room tem-
perature. After 20 h, additional lithium hydroxide (110 mg,
2.64 mmol) was added and the reaction was continued over-
night. The catalyst was removed by filtration and washed with
tetrahydrofuran, and the filtrate was concentrated to dryness
1
6.1.13. N,N-di-tert-Butoxycarbonyl-4-(tert-
yielded compound 24 (2.5 g, 97%) as a white solid: H NMR
butoxycarbonyloxy)-2-nitroaniline (28)
(400 MHz, CDCl3) δ 1.5 (s, 9H, 3 CH3), 1.88–1.95 (m, 2H,
CH2), 2.62 (s, 3H, CH3), 2.70–2.74 (m, 5H, CH2 + CH3),
2.85–3.10 (m, 2H, CH2), 3.19 (t, J = 6.7, 2H, CH2), 3.73 (t,
J = 5.5, 2H, CH2), 4.32 (t, J = 6.7, 2H, CH2), 5.71 (dd,
J = 10.5, J = 3.6, 1H, CH), 6.33 (d, J = 3.1, 1H, ArH), 6.82
(dd, J = 8.9, J = 2.2, 1H, ArH), 6.96 (d, J = 7.6, 1H, ArH),
7.04 (d, J = 2.3, 1H, ArH), 7.08 (d, J = 3.1, 1H, ArH), 7.3–
7.38 (m, 2H, ArH); m/z 537 (M + H)+.
A mixture 27 (4.30 g, 16.9 mmol), di-tert-butyl dicarbonate
(9.24 g, 42.3 mmol), and 4-(N,N-dimethylamino)pyridine
(2.07 g, 16.9 mmol) in acetonitrile (100 ml) was stirred at room
temperature for 12 h. After evaporation of the solvent, the re-
sidue was purified by chromatography on silica (EtOAc/hex-
anes, 1:5) to afford compound 28 (5.23 g, 68%) as a colorless
1
solid: H NMR (400 MHz, CDCl3) δ 1.41 (s, 18 H, 6 CH3),
1.60 (s, 9H, 3 CH3), 7.33 (d, J = 8.4 Hz, 1H, ArH), 7.49 (dd,
J = 8.4 Hz, J = 2.4 Hz, 1H, ArH), 7.98 (d, J = 2.4 Hz, 1H,
ArH).
6.1.11. 3-[5-[2-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
ethoxy]indol-1-yl]-3-[5-(N,N-dimethylaminomethyl)-3-pyridyl]
propionic acid ammonium salt (25)
6.1.14. N,N-di-tert-Butoxycarbonyl-4-hydroxy-2-nitroaniline
(29)
A solution of 23 (3.7 g, 6.2 mmol), copper(I) trifluorometha-
nesulfonate benzene complex (2.94 g, 5.9 mmol) in 30 ml of
toluene and 3 ml of dimethylformamide was heated at 130 °C
for 5 h. The solvents were evaporated to dryness and the residue
was triturated in dichloromethane/methanol/ammonium hydro-
xide 5:1:0.1 (300 ml) for 30 min, then concentrated to dryness
under reduced pressure. The crude material was purified by col-
umn chromatography on silica (CH2Cl2/MeOH/NH4OH, 7:1:0.1)
A solution of 28 (5.20 g, 11.4 mmol) and N,N-diethylethy-
lenediamine (3.80 ml, 12.6 mmol) in acetonitrile (50 ml) was
stirred at room temperature for 1 h. After evaporation of the
solvent, the residue was diluted with EtOAc (50 ml), washed
with 1 N HCl (20 ml) and cold water (20 ml), dried (Na2SO4),
then concentrated to dryness under reduced pressure to give 29
1
1
yielded compound 25 (3.1 g, 95%) as white powder: H NMR
as a colorless solid (3.8 g, 94%): H NMR (400 MHz, CDCl3)
(400 MHz, DMSO-d6) δ 1.71–1.75 (m, 2H, CH2), 2.49 (s, 6H, 2
CH3), 2.62 (t, J = 6.1, 2H, CH2), 2.90 (t, J = 6.7, 2H, CH2),
3.20–3.25 (m, 2H, CH2), 3.33–3.53 (m, 2H, CH2), 3.95 (s, 2H,
CH2), 4.20 (t, J = 6.7, 2H, CH2), 6.00–6.10 (m, 1H, CH), 6.39–
6.43 (m, 2H, ArH), 6.48 (s, 1H, ArH), 6.72 (dd, J = 2.3, J = 8.9,
1H, ArH), 7.03 (d, J = 2.3, 1H, ArH), 7.12 (d, J = 7.3, 1H, ArH),
7.45 (d, J = 9.0, 1H, ArH), 7.67 (d, J = 3.2, 1H, ArH), 7.82 (s,
1H, ArH), 8.48 (s, 1H, ArH), 8.61 (s, 1H, ArH); m/z 500 (M
+ H)+. HRMS (ESI) m/z calcd for C29H33N5O3 499.25834,
found 500.26616 (M + H+).
δ 1.40 (s, 18 H, 6 CH3), 7.12 (s, 1H, ArH), 7.13 (d, J = 2.4 Hz,
1H, ArH), 7.59 (d, J = 2.4 Hz, 1H, ArH); m/z 355 (M + H)+.
6.1.15. N,N-di-tert-Butoxycarbonyl-4-[2-(5,6,7,8-tetrahydro
[1,8]naphthyridin-2-yl)ethoxy]-2-nitroaniline (31)
Diisopropyl azodicarboxylate (2.07 ml, 10.6 mmol) was
added at 0 °C under argon to a solution of 29 (2.5 g,
7.05 mmol), triphenylphosphine (2.77 g, 10.6 mmol), and 30
[18,38] (1.96 g, 7.05 mmol) in tetrahydrofuran (30 ml). After
12 h, the reaction mixture was allowed to warm to room tem-
perature, and then concentrated to dryness under reduced pres-
sure. Chromatography on silica (EtOAc/hexanes, 1:4) afforded
6.1.12. 3-[5-[2-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
ethoxy]indol-1-yl]-4-(N,N-dimethyl amino)-4-oxobutyric acid
(26)
A solution of 24 (2.4 g, 4.47 mmol), copper(I) trifluoro-
methanesulfonate benzene complex (1.96 g, 3.9 mmol) in
30 ml of toluene and 3 ml of dimethylformamide was heated at
130 °C for 5 h. The solvents were evaporated to dryness and
the residue was triturated in dichloromethane/methanol/ammo-
nium hydroxide 5:1:0.1 (300 ml) for 30 min, then concentrated
1
the compound 31 (2.22 g, 51%) as a white powder: H NMR
(400 MHz, CDCl3) δ 1.39 (s, 18H, 6 CH3), 1.51 (s, 9H, 3
CH3), 1.90–1.97 (m, 2H, CH2), 2.74 (t, J = 6.6 Hz, 2H,
CH2), 3.22 (t, J = 6.8 Hz, 2H, CH2), 3.77 (t, J = 6.0 Hz, 2H,
CH2), 4.44 (t, J = 6.8 Hz, 2H, CH2), 6.91 (d, J = 7.6 Hz, 1H,
ArH), 7.15 (d, J = 2.4 Hz, 1H, CH), 7.17 (s, 1H, ArH), 7.34 (d,
J = 7.6 Hz, 1H, ArH), 7.56 (d, J = 2.8 Hz, 1H, ArH); m/z 615
(M + H)+.

P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
857
6.1.16. 4-[2-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
ethoxy]-2-nitroaniline (32)
nate, washed with water, dried (Na2SO4), and evaporated. The
crude material was purified by column chromatography on si-
lica (CH2Cl2/MeOH, 95:5) to give compound 40 mg (44%) as
a white powder (m/z 513 (M + H)+), which was dissolved in
absolute ethanol (58 μl) and hexamethylphosphoramide
(250 μl). To the resulting solution, samarium(II) iodide
(0.1 M in THF, 5.0 ml) was slowly added. After 24 h at room
temperature, the reaction mixture was quenched with ammo-
nium chloride (1 N, 10 ml), diluted with water and extracted
with ethyl acetate. Organic layer was washed with water, dried
(Na2SO4), and concentrated to dryness under reduced pressure.
Chromatography on silica (AcOEt/CH2Cl2/EtOH, 40:50:10) af-
forded the desired product 37 (23 mg, 40%) as a white solid: 1H
NMR (400 MHz, CDCl3) δ 1.14 (t, J = 7.1, 3H, CH3), 1.90
(quint, J = 4.8, 2H, CH2), 2.68–2.73 (m, 2H, CH2), 3.05 (t,
J = 6.8, 2H, CH2), 3.28 (ddd, J = 16.0, J = 6.8, J = 8.8, 2H,
CH2), 3.39-3.44 (m, 2H, CH2), 4.07 (q, J = 7.1, 2H, CH2),
4.30 (t, J = 6.8, 2H, CH2), 5.88 (dd, J = 6.8, J = 8.8, 1H,
CH), 5.94 (s, 2H, CH2), 6.47 (d, J = 7.6, 1H, ArH), 6.66 (s,
1H, ArH), 6.75 (s, 2H, ArH), 6.87 (dd, J = 8.8, J = 2.4, 1H,
ArH), 7.12 (t, J = 7.6, 1H, ArH), 7.26 (s, 1H, ArH), 7.96 (s,
1H, ArH); m/z 515 (M + H)+.
Trifluoroacetic acid (1.76 ml, 22.8 mmol) was slowly added
to a stirred solution of 31 (1.40 g, 2.28 mmol) in dichloro-
methane (20 ml). After 1 h, the mixture was diluted with ethyl
acetate (200 ml), washed with 1 N sodium bicarbonate
(250 ml) and ice-cold water (200 ml), dried (Na2SO4), and
concentrated to dryness under reduced pressure. Chromatogra-
phy on silica (EtOAc/hexanes, 1:6) afforded compound 32
(375 mg, 40%) as a white powder: 1H NMR (400 MHz,
CDCl3) δ 1.51 (s, 9H, 3 CH3), 1.92 (quint, J = 6.4 Hz, 2H,
CH2), 2.74 (t, J = 6.8 Hz, 2H, CH2), 3.17 (t, J = 6.6 Hz, 2H,
CH2), 3.76 (t, J = 6.0 Hz, 2H, CH2), 4.32 (t, J = 6.8 Hz, 2H,
CH2), 5.86 (br s, 2H, NH2), 6.72 (d, J = 8.8 Hz, 1H, ArH),
6.90 (d, J = 7.6 Hz, 1H, ArH), 7.05 (dd, J = 8.8 Hz,
J = 2.8 Hz, 1H, ArH), 7.32 (d, J = 7.2 Hz, 1H, ArH), 7.55 (d,
J = 2.8 Hz, 1H, ArH); m/z 415 (M + H)+.
6.1.17. 4-[2-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
ethoxy]-N-formyl-2-nitroaniline (33)
A solution of formic acid (28 μl, 0.744 mmol) and acetic
anhydride (58 μl, 0.620 mmol) in tetrahydrofuran (2 ml) was
heated at 50 °C for 3 h. After the solution was cooled to room
temperature, 32 (300 mg, 0.725 mmol) in tetrahydrofuran
(2 ml) was added and the resulting solution was stirred at room
temperature for 5 h. The solvent was evaporated to dryness and
the residue was partitioned between ethylacetate (50 ml) and
1 N sodium bicarbonate (30 ml), washed with water, dried
(Na2SO4), and concentrated under reduced pressure. The crude
material was purified by column chromatography on silica
(EtOAc/hexanes, 1:4) to give compound 33 (200 mg, 62%)
6.1.19. 3-[5-[2-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
ethoxy]benzimidazol-1-yl]-3-(3,4-methylenedioxyphenyl)
propionic acid (38)
A solution of 37 (23 mg, 0.045 mmol) and lithium hydro-
xide monohydrate (1 M in H2O, 54 μl) in methanol (1 ml) and
tetrahydrofuran (2 ml) was stirred for 36 h at 20 °C. The mix-
ture was evaporated to dryness and the residue was redissolved
in water (2 ml). The resulting solution was neutralized with
1 N acetic acid, then evaporated to dryness. The residue was
chromatographed on silica (AcOEt/CH2Cl2/EtOH, 3:6:1) to
give a white solid, which was recrystallized from ethanol and
diethyl ether to give compound 38 (18 mg, 83%) as a white
solid. 1H NMR (400 MHz, CDCl3) δ 2.70 (t, J = 6.0, 2H,
CH2), 2.97 (t, J = 7.0, 2H, CH2), 3.06–3.27 (m, 2H, CH2),
3.37 (t, J = 5.8, 2H, CH2), 4.23 (t, J = 7.0, 2H, CH2), 5.95 (d,
J = 5.2, 2H, CH2), 5.95 (t, J = 8.4, 2H, CH2), 6.47 (d, J = 7.2,
1H, ArH), 6.75 (d, J = 8.0, 1H, ArH), 6.79–6.85 (m, 3H, ArH),
7.13–7.16 (m, 2H, ArH), 7.28 (d, J = 9.2, 1H, ArH), 8.33 (br s,
1H, ArH); m/z 487 (M + H)+. HRMS (ESI) m/z calcd for
C27H26N4O5 486.19032, found 487.19814 (M + H+).
1
as a white powder: H NMR (400 MHz, CDCl3) δ 1.44 (s,
9H, 3 CH3), 1.86 (quint, J = 6.3 Hz, 2H, CH2), 2.68 (t,
J = 6.8 Hz, 2H, CH2), 3.13 (t, J = 6.6 Hz, 2H, ArH), 3.70 (t,
J = 6.0 Hz, 2H, CH2), 4.34 (t, J = 6.8 Hz, 2H, CH2), 6.84 (d,
J = 7.2 Hz, 1H, ArH), 7.17 (dd, J = 9.6 Hz, J = 3.2 Hz, 1H,
ArH), 7.27 (d, J = 7.2 Hz, 1H, ArH), 7.62 (d, J = 2.8 Hz, 1H,
ArH), 8.44 (s, 1H, ArH), 8.58 (d, J = 9.2 Hz, 1H, ArH), 9.96
(br s, 1H, NH); m/z 443 (M + H)+.
6.1.18. Ethyl 3-[5-[2-(5,6,7,8-tetrahydro [1,8]naphthyridin-2-
yl)ethoxy]benzimidazol-1-yl]-3-(3,4-methylenedioxyphenyl)
propionate (37)
A mixture of 33 (200 mg, 0.452 mmol), 34 [18,38]
(0.148 mg, 0.678 mmol), and cesium fluoride (102 mg,
0.678 mmol) in anhydrous dimethylformamide (1.5 ml) was
stirred under nitrogen at 80 °C for 24 h. The mixture was suc-
cessively cooled to room temperature, diluted with ethyl acet-
ate (40 ml), and washed with water (3 × 50 ml). Purification by
column chromatography (EtOAc/hexanes, 1:1) gave compound
132 mg (52%) of a white solid (m/z 561 (M + H)+), which was
dissolved in acetic acid (1 ml). To this solution, iron powder
(100 mg, 1.79 mmol) was added and the resulting solution was
heated to 80 °C for 3 h. After the solution was cooled to room
temperature, the solvent was evaporated in vacuo. The residue
was partitioned between ethyl acetate and 1 N sodium bicarbo-
6.1.20. 3-[5-[2-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
ethoxy]-3-(N,N-dimethylamino methyl)indol-1-yl]-3-
phenylpropionic acid (39)
A solution of 2 [18] (100 mg, 0.23 mmol) and dimethyl
methyleneimmonium iodide (46 mg, 0.25 mmol) in dichloro-
methane (3 ml) was stirred at room temperature for 12 h. Then,
the solvent was evaporated under reduced pressure. Chromato-
graphy on silica (CH2Cl2/EtOH, 9.5:0.5) followed by recrystal-
lization from ethyl alcohol and diethyl ether yielded compound
1
35 (92 mg, 54%) as colorless prisms: H NMR (400 MHz,
DMSO-d6) δ 0.88 (t, J = 6.8, 2H, CH2), 1.96 (quint, J = 6.10,
2H, CH2), 2.77 (t, J = 6.2, 2H, CH2), 2.91 (s, 6H, 2 CH3), 3.18

858
P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
(t, J = 6.2, 2H, CH2), 3.53 (t, J = 5.8, 2H, CH2), 4.51 (t,
J = 6.4, 2H, CH2), 4.60 (dd, J = 29.6, J = 13.2, 2H, CH2),
5.97 (dd, J = 9.2, J = 6.0, 1H, CH), 6.60 (d, J = 7.2, 1H,
ArH), 6.76 (dd, J = 9.2, J = 2.4, 1H, ArH), 7.19–7.28 (m, 6H,
ArH), 7.42 (d, J = 7.2, 1H, ArH), 7.62 (d, J = 2.4, 1H, ArH),
8.01 (s, 1H, ArH); m/z 499 (M + H)+. HRMS (ESI) m/z calcd
for C30H34N4O3 498.26309, found 498.26215 (M + H+).
J = 7.0, 3H, CH3), 4.13 (q, J = 7.0, 2H, CH2), 6.63 (s, 1H,
CH), 6.93 (d, J = 4.0, 1H, ArH), 7.28–7.52 (m, 6H, ArH),
1
8.80 (s, 1H, ArH); m/z 328 (M + H)+. Compound (E)-44: H
NMR (400 MHz, CDCl3) δ 1.09 (t, J = 7.0, 3H, CH3), 4.02 (q,
J = 7.0, 2H, CH2), 6.63 (s, 1H, CH), 6.81 (d, J = 4.0, 1H,
ArH), 7.20 (d, J = 4.0, 1H, ArH), 7.28–7.52 (m, 5H, ArH),
8.64 (s, 1H, ArH); m/z 328 (M + H)+.
6.1.21. Ethyl 3-[5-[3-[(2-pyridyl)amino]propoxy]-2,3-
6.1.24. (Z,E)-3-[1-[3-[8-tert-Butoxycarbonyl(5,6,7,8-
dihydroindol-1-yl]propionate (41)
tetrahydro [1,8]naphthyridin-2-yl)]prop-2-enyl]-7-
deazahypoxanthin-9-yl]-3-phenylacrylic acid ethyl ester (46)
A solution of 44 (1.00 g, 3.05 mmol) and concentrated hy-
drochloric acid (100 μl, 3.7 mmol) was heated at 37 °C for
8 days, then evaporated to dryness. Chromatography on silica
(AcOEt) afforded 690 mg of a colorless solid (m/z 310
(M + H)+), which was use without further purification in the
next step. This solid was added to a solution of 50 [18,38]
(690 mg, 2.39 mmol) and triphenylphosphine (935 mg,
3.56 mmol) in tetrahydrofuran (20 ml). Then, diisopropyl azo-
dicarboxylate (702 μl, 3.56 mmol) was added at 0 °C under
argon. After 5 min, the reaction mixture was allowed to warm
to room temperature. After 1 h, the reaction mixture was con-
centrated to dryness under reduced pressure. Chromatography
on silica (EtOAc/hexanes, 1:1) afforded the compounds 46
(953 mg, 50%) as a white powder (mixture of 2,3-E/Z deriva-
Sodium cyanoborohydride (43 mg, 0.69 mmol) was added
to a solution of 40 [17] (84 mg, 0.23 mmol) in methanol
(1.5 ml). Then, hydrochloric acid (4 M in dioxane, 57 μl)
was added. The resulting solution was stirred at room tempera-
ture for 16 h. The solvent was evaporated under reduced pres-
sure, and the residue was partitioned between dichloromethane
and water. Organic layer was dried (Na2SO4) and evaporated.
Chromatography on silica (CH2Cl2/AcOEt, 1:1) afforded com-
pound 41 (57 mg, 62%) as a colorless oil: 1H NMR (400 MHz,
CDCl3) δ 1.24 (t, J = 7.2, 3H, CH3), 2.03 (quint, J = 6.4, 2H,
CH2), 2.58 (t, J = 7.0, 2H, CH2), 2.86 (t, J = 8.2, 2H, CH2),
3.24 (t, J = 8.2, 2H, CH2), 3.29–3.32 (m, 2H, CH2), 3.45 (t,
J = 7.0, 2H, CH2), 3.99 (t, J = 6.0, 2H, CH2), 4.13 (q,
J = 7.2, 2H, CH2), 6.47 (d, J = 8.4, 1H, ArH), 6.56–6.65 (m,
3H, ArH), 6.73–6.74 (m, 1H, ArH), 7.46–7.50 (m, 1H, ArH),
7.86–7.88 (m, 1H, ArH); m/z 370 (M + H)+.
1
tives in a 3/4:1/4 proportion). 2,3-(Z)-46: H NMR (400 MHz,
CDCl3) δ 1.12 (t, J = 7.1, 3H, CH3), 1.52 (s, 9H, 3 CH3), 1.92
(quint, J = 6.2, 2H, CH2), 2.75 (t, J = 6.2, 2H, 2 CH2), 3.66–
3.77 (m, 2H, CH2), 4.08 (q, J = 7.1, 2H, CH2), 4.85 (d, J = 7.0,
2H, CH2), 6.52 (s, 1H, CH), 6.53 (d, J = 3.5, 1H, ArH), 6.58
(d, J = 3.5, 1H, ArH), 7.28–7.48 (m, 9H, ArH), 8.00 (s, 1H,
ArH); m/z 582 (M + H)+. 2,3-(E)-46: 1H NMR (400 MHz,
CDCl3) δ 1.08 (t, J = 7.1, 3H, CH3), 1.50 (s, 9H, 3 CH3),
1.92 (quint, J = 6.2, 2H, CH2), 2.75 (t, J = 6.2, 2H, 2 CH2),
3.66–3.77 (m, 2H, CH2), 4.06 (q, J = 7.1, 2H, CH2), 4.83 (d,
J = 7.0, 2H, CH2), 6.52 (s, 1H, CH), 6.86 (d, J = 3.5, 1H,
ArH), 6.89 (d, J = 3.5, 1H, ArH), 7.28–7.48 (m, 9H, ArH),
7.86 (s, 1H, ArH); m/z 582 (M + H)+.
6.1.22. 3-[5-[3-[(2-Pyridyl)amino]propoxy]-2,3-dihydroindol-
1-yl]propionic acid (42)
A solution of 41 (49 mg, 0.13 mmol) and sodium hydroxide
(21 mg, 0.53 mmol) in methanol (2 ml) was stirred at room
temperature for 16 h. The solvent was evaporated under re-
duced pressure, and the residue was redissolved in water. The
resulting solution successively neutralized with HCl 1 N, ex-
tracted with ethylacetate, dried (Na2SO4) and evaporated under
reduced pressure. Chromatography on silica (CH2Cl2/MeOH,
9.5:0.5) afforded compound 42 (24 mg, 53%) as a colorless
1
solid: H NMR (400 MHz, CDCl3) δ 2.06 (quint, J = 6.4, 2H,
CH2), 2.55 (t, J = 7.0, 2H, CH2), 2.85 (t, J = 8.0, 2H, CH2),
3.23–3.32 (m, 4H, 2 CH2), 3.49 (t, J = 7.0, 2H, CH2), 4.00 (t,
J = 5.9, 2H, CH2), 6.48 (d, J = 8.5, 1H, ArH), 6.62–6.73 (m,
3H, ArH), 6.80 (d, J = 8.9, 1H, ArH), 7.63–7.68 (m, 1H, ArH),
7.81–7.89 (m, 1H, ArH); m/z 342 (M + H)+. HRMS (ESI) m/z
calcd for C19H23N3O3 341.17394, found 342.18177 (M + H+).
6.1.25. 3-[1-[3-[8-tert-Butyloxycarbonyl(5,6,7,8-tetrahydro
[1,8]naphthyridin-2-yl)]propyl]-7-deazahypoxanthin-9-yl]-3-
phenylpropionic acid ethyl ester (47)
A mixture of 46 (310 mg, 0.53 mmol) and 10% Pd/C
(100 mg) in absolute methanol (8 ml) was shaken in a hydro-
genation apparatus under atmospheric pressure at room tem-
perature for 1 h. The catalyst was removed by filtration and
washed with methanol, and the filtrate was concentrated to dry-
ness to give compound 47 (295 mg, 95%) as white powder: 1H
NMR (400 MHz, CDCl3) δ 1.11 (t, J = 7.2, 3H, CH3), 1.54 (s,
9H, 3 CH3), 1.94 (quint, J = 6.3, 2H, CH2), 2.23 (quint,
J = 7.0, 2H, CH2), 2.73–2.79 (m, 4H, 2 CH2), 3.37 (ddd,
J = 16.0, J = 8.8, J = 6.8, 2H, CH2), 3.79 (t, J = 6.0, 2H,
CH2), 4.06 (d, J = 7.2, 2H, CH2), 4.15–4.23 (m, 2H, CH2),
6.31 (dd, J = 8.8, J = 6.8, 1H, CH), 6.71 (d, J = 3.6, 1H,
ArH), 6.86 (d, J = 7.6, 1H, ArH), 6.93 (d, J = 3.6, 1H, ArH),
7.28–7.44 (m, 6H, ArH), 8.25 (s, 1H, ArH); m/z 586 (M + H)+.
6.1.23. (Z,E)-3-(6-Chloro-7-deazapurin-9-yl)-3-phenylacrylic
acid ethyl ester (44)
A mixture of 43 (1.00 g, 6.51 mmol), ethylphenylpropiolate
(1.61 ml, 9.77 mmol), and cesium fluoride (1.97 g, 13 mmol)
in anhydrous dimethylformamide (12 ml) was stirred under ni-
trogen at 75 °C for 1 h. The mixture was successively cooled
to room temperature, diluted with ethyl acetate (40 ml), and
washed with water (3 × 50 ml). Purification by column chro-
matography (EtOAc/hexanes, 1:1) gave 44 (2.08 g, 97%) as
white solid (mixture of E/Z derivatives in a 3/4:1/4 proportion).
Compound (Z)-44: 1H NMR (400 MHz, CDCl3) δ 1.18 (t,

P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
859
6.1.26. 3-[1-[3-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
propyl]-7-deazahypoxanthin-9-yl]-3-phenylpropionic acid
ethyl ester (48)
mixed with 40 μg/ml human GPIIbIIIa (Enzyme Research La-
boratories) in TACTB buffer, and 100 μl per well of these so-
lutions were incubated for 1 h at 37 °C. The plate was then
washed five times with PBST buffer, and 100 μl per well of
a monoclonal anti-GPIIbIIIa antibody in TACTB buffer (1 μg/
ml, Enzyme Research Laboratories) was incubated at 37 °C for
1 h. After washing five times with PBST buffer, 100 μl per
well of goat anti-mouse IgG conjugated to horseradish perox-
idase (Kirkegaard & Perry) was incubated at 37 °C for 1 h
(25 ng/ml in PBST buffer), followed by a sixfold PBST buffer
wash. The plate was developed by adding 100 μl per well of
0.67 mg o-phenylenediamine dihydrochloride per ml of
0.012% H2O2, 22 mM sodium citrate, 50 mM sodium phos-
phate, pH 5.0 at room temperature. The reaction was stopped
with 50 μl per well of 2 M H2SO4, and the absorbance at
492 nm was recorded. Percent (%) inhibition was calculated
from an average of three separate determinations relative to
buffer controls (no test compound added), and a four parameter
fit was used to estimate the half maximal inhibition concentra-
tion (IC50).
A mixture of 47 (240 mg, 0.41 mmol) and copper(I) tri-
fluoromethanesulfonate benzene complex (50 mg, 0.10 mmol)
in toluene (5 ml) and dimethylformamide (500 μl) was heated
at 130 °C for 13 min in a multimods Smith synthesizer® mi-
crowave oven (irradiation at 500 W). For this purpose, stan-
dard sealed tube flask was used. The resulting yellow solution
was cooled to room temperature and filtered. Chromatography
on silica (EtOAc/CH2Cl2, 1:1) afforded compound 48 (180 mg,
1
90%) as a colorless solid: H NMR (400 MHz, CDCl3) δ 1.12
(t, J = 7.0, 3H, CH3), 1.94 (quint, J = 5.9, 2H, CH2), 2.17
(quint, J = 7.4, 2H, CH2), 2.67–2.74 (m, 4H, 2 CH2), 3.35
(ddd, J = 16.0, J = 9.2, J = 6.8, 2H, CH2), 3.44–3.46 (m, 2H,
CH2), 4.02–4.09 (m, 4H, 2 CH2), 6.33 (dd, J = 9.2, J = 6.8,
1H, CH), 6.46 (d, J = 7.2, 1H, ArH), 6.70 (d, J = 3.6, 1H,
ArH), 6.94 (d, J = 3.6, 1H, ArH), 7.17 (d, J = 7.2, 1H, ArH),
7.26–7.39 (m, 5H, ArH), 7.95 (s, 1H, ArH), 8.04 (br s, 1H,
NH); m/z 486 (M + H)+.
6.1.27. 3-[1-[3-(5,6,7,8-Tetrahydro [1,8]naphthyridin-2-yl)
propyl]-7-deazahypoxanthin-9-yl]-3-phenylpropionic acid (49)
A mixture of 48 (180 mg, 0.37 mmol) and lithium hydro-
xide monohydrate (17 mg, 0.41 mmol) in tetrahydrofuran
(3 ml), methanol (1 ml), and water (300 μl) was heated at
100 °C for 2 min in a multimods Smith synthesizer® micro-
wave oven (irradiation at 500 W). For this purpose, standard
sealed tube flask was used. The resulting solution was cooled
to room temperature and evaporated. Chromatography on silica
(CH2Cl2/EtOH, 9.5:0.5) followed by recrystallization from
ethyl alcohol and diethyl ether yielded compound 49 (92 mg,
6.2.2. αvβ3-Vitronectin assay
The assay was based on the method of Niiya et al. [39], and
all steps were performed at room temperature. Costar 9018 flat-
bottom 96-well ELISA plates were coated overnight at room
temperature with 100 μl per well of 0.4 μg/ml human αvβ3
(Chemicon) in TS buffer (20 mM Tris–HCl pH 7.5, 150 mM
NaCl, 1 mM CaCl2, 1 mM MgCl2, 1 mM MnCl2). Plates were
subsequently emptied and blocked for 2 h with 150 μl per well
of TS buffer containing 1% BSA (TSB buffer), and washed
three times with 300 μl per well of PBST buffer. Controls or
test compound (0.0001–20.0 μM) were mixed with 1 μg/ml of
human vitronectin (Chemicon) that had been biotinylated in-
house with sulfo-NHS-LC-LC-biotin (Pierce, 20:1 molar ratio),
and 100 μl per well of these solutions (in TSB buffer) were
incubated for 2 h. The plate was then washed five times with
PBST buffer, and 100 μl per well of 0.25 μg/ml NeutrAvidin-
horseradish peroxidase conjugate (Pierce) in TSB buffer was
added to the plate and incubated for 1 h. Following a fivefold
PBST buffer wash, the plate was developed and results were
calculated as described for the IIbIIIa-fibrinogen assay.
1
54%) as colorless prisms: H NMR (400 MHz, DMSO-d6) δ
1.73 (quint, J = 5.7, 2H, CH2), 1.93 (quint, J = 7.4, 2H, CH2),
2.43 (t, J = 7.6, 2H, CH2), 2.59 (t, J = 6.0, 2H, CH2), 3.19–
3.53 (m, 4H, 2 CH2), 3.95–3.99 (m, 2H, CH2), 6.18 (t,
J = 8.0, 1H, CH), 6.27 (d, J = 7.6, 1H, ArH), 6.31 (br s, 1H,
NH), 6.50 (d, J = 3.6, 1H, ArH), 7.01 (d, J = 7.6, 1H, ArH),
7.23–7.36 (m, 5H, ArH), 7.45 (d, J = 3.6, 1H, ArH), 8.24 (s,
1H, ArH); m/z 458 (M + H)+. HRMS (ESI) m/z calcd for
C26H27N5O3 457.21139, found 458.21921 (M + H+).
6.2. Pharmacology
6.2.3. αvβ5-Vitronectin assay
The assay is similar to the αvβ3-vitronectin assay, and all
steps were performed at room temperature. Costar 9018 flat-
bottom 96-well ELISA plates were coated overnight at room
temperature with 100 μl per well of 1 μg/ml human αvβ5 (Che-
micon) in TS buffer. Plates were blocked for 2 h with 150 μl
per well of TSB buffer, and washed three times with 300 μl per
well of PBST buffer. Controls or test compound (0.0001–
20 μM) were mixed with 1 μg/ml of human vitronectin (Che-
micon) that had been biotinylated in-house with sulfo-NHS-
LC-LC-biotin (Pierce, 20:1 molar ratio), and 100 μl per well
of these solutions (in TSB buffer) were incubated for 2 h. The
plate was then washed five times with PBST buffer, and 100 μl
per well of 0.25 μg/ml NeutrAvidin-horseradish peroxidase
6.2.1. IIbIIIa-fibrinogen assay
The assay is based on the method of Dennis et al. [40].
Costar 9018 flat-bottom 96-well ELISA plates were coated
overnight at 4 °C with 100 μl per well of 10 μg/ml human
fibrinogen (Calbiochem) in 20 mM Tris–HCl pH 7.5,
150 mM NaCl, 2 mM CaC12, 0.02% NaN3 (TAC buffer).
Plates were subsequently emptied and blocked for 1 h at
37 °C with 150 μl per well of TAC buffer containing 0.05%
Tween 20 and 1% bovine serum albumin (TACTB buffer).
After washing three times with 300 μl per well of 10 mM
Na2HPO4 pH 7.5, 150 mM NaCl, 0.01% Tween 20 (PBST
buffer), controls or test compound (0.027–20.0 μM) were

860
P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
conjugate (Pierce) in TSB buffer was added to the plate and
incubated at for 1 h. Following a fivefold PBST buffer wash,
the plate was developed and results were calculated as de-
scribed for the IIbIIIa-fibrinogen assay.
6.2.5.3. Interaction analysis of drug concentration series. In-
teraction analysis of drugs binding to two differing capacity
surfaces of HSA was conduced at 25 °C in HBS + 3% DMSO
(v/v) at pH 7.4. Three compounds were analyzed per micro
titer plate. From each plate, 13 buffer injections initiated the
run followed by a set of 10 buffer samples, containing different
DMSO concentrations within the range 3.3–2.7% for the pur-
pose of constructing a calibration plot (see below). Stocks of
compounds supplied at 10 mM in 100% DMSO were diluted
3:100 (v/v) in DMSO-free running buffer (HBS pH 7.4).
Further dilutions were made serially in running buffer (HBS
+ 3% DMSO) to generate concentrations spanning a wide
(500-fold) range (typically 200–0.39 μM) in twofold incre-
ments. Less soluble compounds were analyzed at lower con-
centrations. Each sample was dispensed into duplicate 180 μl-
aliquots and placed in a single-use well of a 96-well micro titer
plate. Each concentration series was analyzed in increasing
concentration. Samples containing only buffer were inter-
spersed in the assay for the purpose of double referencing bio-
sensor responses during data processing (see below). Associa-
tion and dissociation phases were monitored for 1 and 2 min,
respectively, at 90 μl/min. An extra wash was included after
each sample injection.
6.2.4. α5β1-Fibronectin assay
Costar 9018 flat-bottom 96-well ELISA plates were coated
overnight at room temperature with 100 μl per well of 3 μg/ml
human α5β1 (Chemicon) in TS buffer. Plates were subse-
quently emptied and blocked for 2 h at 30 °C with 150 μl per
well of TSB buffer, and washed three times with 300 μl per
well of PBST buffer. Controls or test compound (0.0001–
20 μM) were mixed with 1 μg/ml of human fibronectin (Che-
micon) that had been biotinylated in-house with sulfo-NHS-
LC-LC-biotin (Pierce, 20:1 molar ratio), and 100 μl per well
of these solutions (in TSB buffer) were incubated for 2 h at
30 °C. The plate was then washed three times with PBST buf-
fer, and 100 μl per well of 0.25 μg/ml NeutrAvidin-horseradish
peroxidase conjugate (Pierce) in TSB buffer was added to the
plate and incubated at for 1 h at 30 °C. Following a sixfold
PBST buffer wash, the plate was developed and results were
calculated as described for the IIbIIIa-fibrinogen assay.
6.2.5. HSA binding
6.2.5.4. Data processing and analysis. Subtracting the
responses generated across an unmodified reference spot
referenced biosensor data. A DMSO calibration plot was
constructed from buffer samples containing different concen-
trations of DMSO and used to correct the data for mismatched
solvent (excluded volume effect). Buffer responses were aver-
aged and subtracted from the data in a process known as dou-
ble referencing. Equilibrium binding responses were plotted
against compound concentration and fit to a binding isotherm
to obtain affinities using an appropriate reaction mechanism. If
data did not fit to a single site model (A + B = AB, where
A = injected drug and B = binding site on immobilized HSA),
the highest affinity interaction was resolved using a model that
accounted for multi-site binding (A + B = AB, A + C = AC,
where B and C denote different binding sites on HSA). Percent
bound conversions were calculated based on the highest affi-
nity interaction by assuming a drug concentration of 10 μM (to
mimic a physiological concentration in man). Off rate analysis
was based on a single exponential decay (a simple reaction
model). Off rates were determined from dissociation phase data
of curves corresponding to the highest affinity interaction,
when multiple sites appeared to be populated (from affinity
determinations). Complex half-lives (t1/2) were calculated from
the relationship t1/2 = ln(2/Kd).
HSA binding (% of bound values and Kd) was determined
by BIACORE SPR biosensors technology with full length
HAS as previously reported in [41].
6.2.5.1. General. Interaction analysis was conducted at 25 °C
on an S51 BIACORE optical biosensor equipped with an S
series CM5 sensor chip. HBS (10 mM Hepes, 150 mM NaCl,
pH 7.4) was used as the running buffer for the immobilization.
It was supplemented with 3% DMSO (v/v) for the drug analy-
sis. To ensure purity, DMSO was drawn from single-use vials
(Sigma Hybri-max® D2650, sterile-0.2 μm filtered).
6.2.5.2. HSA immobilization. The fluidic system was equili-
brated in freshly filtered (0.2 μm) and degassed HBS running
buffer and the detector normalized using 70% v/v glycerol
water to maximize sensitivity (automated procedure). Buffer
was then flowed at 10 μl/min to immobilize HSA by a standard
amine coupling method. In separate cycles, each spot on flow
cell 2 was activated for 7 min using 50 mM NHS:200 mM
EDC. HSA (Sigma A6784) was diluted to 40 μg/ml (0.7 μM)
in 10 mM sodium acetate buffer at pH 5.2 and injected across
spots 1 and 2 using contact times of 7 and 2 min, respectively,
to create different capacity surfaces. A 7-min pulse of 1 M so-
dium ethanolamine at pH 8.5 was used to deactivate excess
esters. Final immobilization levels on spots 1 and 2 of flow
cell 2 were 12580 and 7390 RU, respectively. Replicate buffer
injections from two 96-well micro titer plates filled with HBS
(two sequential runs) stabilized freshly immobilized HSA in
preparation for interaction analysis with tested compounds.
During this stabilization experiment, association and dissocia-
tion phases were monitored for 1 and 2 min, respectively, at
90 μl/min.
References
[1] R.O. Hynes, Cell 110 (2002) 673–687.
[2] G.D. Hartman, M.E. Duggan, Exp. Opin. Invest. Drugs 9 (2000) 1281–
1291.
[3] Recent reviews: (a) M.E. Duggan, J.H. Hutchinson, Exp. Opin. Ther. Pa-
tents 10 (2000) 1367–1383; (b) S.A. Mousa. Med. Res. Rev. 23 (2003),
190–199; (c) J.S. Kerr, A.M. Slee, S.A. Mousa. Expert. Opin. Investig.

P. Raboisson et al. / European Journal of Medicinal Chemistry 41 (2006) 847–861
861
Drugs 11 (2002), 1765–1774; C. Henry, N. Moitessier, Y. Chapleur.
Mini Rev. Med. Chem. 2 (2002), 531–542 and cited references.
[4] R. Max, R.R. Gerritsen, P.T. Nooijen, S.L. Goodman, A. Sutter, U. Keil-
holz, D.J. Ruiter, R.M. De Waal, Int. J. Cancer 71 (1997) 320–324.
[5] P.C. Brooks, A.M.P. Montgomery, M. Rosenfeld, R.A. Reisfeld, T. Hu,
G. Klier, D.A. Cheresh, Cell 79 (1994) 1157–1164.
[6] M. Friedlander, P.C. Brooks, R.W. Shaffer, C.M. Kincald, J.A. Varner,
D.A. Cheresh, Science 270 (1995) 1500–1502.
[7] F. Mitjans, T. Meyer, C. Fittschen, S. Goodman, A. Jonczyk, J.F. Mar-
shall, G. Reyes, J. Piulats. Intern. J. Cancer 87 (2000) 716–723.
[8] M. Skobe, P. Rockwell, N. Goldstein, S. Vosseler, N. E. Nature Med. 3
(1997) 1222–1227.
R.C. Haltiwanger, C.J. Gress, M.W. Lark, S.M. Hwang, I.E. James, D.J.
Rieman, R.N. Willette, T.L. Yue, L.M. Azzarano, K.L. Salyers, B.R.
Smith, K.W. Ward, K.O. Johanson, W.F. Huffman, Bioorg. Med.
Chem. Lett. 13 (2003) 1483–1486.
[23] K. Urbahns, M. Harter, A. Vaupel, M. Albers, D. Schmidt, U. Brugge-
meier, B. Stelte-Ludwig, C. Gerdes, H. Tsujishita, Bioorg. Med. Chem.
Lett. 13 (2003) 1071–1074.
[24] A. Kling, G. Backfisch, J. Delzer, H. Geneste, C. Graef, W. Hornberger,
U.E. Lange, A. Lauterbach, W. Seitz, T. Subkowski, Bioorg. Med.
Chem. 11 (2003) 1319–1341.
[25] M. Bubenik, K. Meerovitch, F. Bergeron, G. Attardo, L. Chan, Bioorg.
Med. Chem. Lett. 13 (2003) 503–506.
[9] C.P. Carron, D.M. Meyer, J.A. Pegg, V.W. Engleman, M.A. Nickols,
S.L. Settle, W.F. Westlin, P.G. Ruminski, G.A. Nickols, Cancer Res. 58
(1998) 1390–1395.
[10] P.C. Brooks, S. Stromblad, R. Klemke, D. Visscher, F.H. Sarkar, D.A.
Cheresh, J. Clin. Invest. 96 (1995) 1815–1822.
[11] J. Luna, T. Tobe, S.A. Mousa, T.M. Reilly, P.A. Campochiaro, Lab. Inv.
75 (1996) 563–573.
[12] C.M. Storgard, D.G. Stupack, A. Jonczyk, S.L. Goodman, R.I. Fox, D.A.
Cheresh, J. Clin. Invest. 103 (1999) 47–54.
[13] M. Cristofanilli, C. Charnsangavej, G.N. Hortobagyi, Nature Rev. 1
(2002) 415–426.
[14] D.K. Kaul, H.M. Tsai, X.D. Liu, M.T. Nakada, R.L. Nagel, B.S. Coller,
Blood 95 (2000) 368–374.
[15] B.P. Feuston, B.P. Feuston, J.C. Culberson, M.E. Duggan, G.D. Hart-
man, C.T. Leu, S.B. Rodan, J. Med. Chem. 45 (2002) 5640–5648.
[16] K.P. McHugh, K. Hodival-Dilke, M.-H. Zheng, N. Namba, J. Lam, D.
Novack, X. Fang, H.P. Ross, R. Hynes, S.L. Teitelbaum, J. Clin. Invest.
105 (2000) 433–440.
[17] J.J. Marugán, C. Manthey, B. Anaclerio, L. LaFrance, T. Lu, T. Marko-
tan, C. Crysler, S. Eisennagel, M. Dasgupta, B. Tomczuk, J. Med. Chem.
48 (2005) 926–934.
[18] K. Leonard, W. Pan, B. Anaclerio, J.M. Gushue, Z. Guo, R.L. DesJar-
lais, J. Lattanze, C. Crysler, C.L. Manthey, B.E. Tomczuk, J.J. Marugan,
Bioorg. Med. Chem. Lett. 15 (2005) 2679–2684.
[19] P. Vianello, P. Cozzi, A. Galvani, M. Meroni, M. Varasi, D. Volpi, T.
Bandiera, Bioorg. Med. Chem. Lett. 14 (2004) 657–661.
[20] J.H. Hutchinson, W. Halczenko, K.M. Brashear, M.J. Breslin, P.J. Cole-
man, T. Le Duong, C. Fernandez-Metzler, M.A. Gentile, J.E. Fisher,
G.D. Hartman, J.R. Huff, D.B. Kimmel, C.T. Leu, R.S. Meissner, K.
Merkle, R. Nagy, B. Pennypacker, J.J. Perkins, T. Prueksaritanont, G.A.
Rodan, S.L. Varga, G.A. Wesolowski, A.E. Zartman, S.B. Rodan, M.E.
Duggan, J. Med. Chem. 46 (2003) 4790–4798.
[21] M.J. Breslin, M.E. Duggan, W. Halczenko, C. Fernandez-Metzler, C.A.
Hunt, C.T. Leu, K.M. Merkle, A.M. Naylor-Olsen, T. Prueksaritanont,
G. Stump, A. Wallace, S.B. Rodan, J.H. Hutchinson, Bioorg. Med.
Chem. Lett. 13 (2003) 1809–1812.
[26] R.M. Van Well, H.S. Overkleeft, G.A. Van der Marel, D. Bruss, G. Thi-
bault, P.G. De Groot, J.H. Van Boom, M. Overhand, Bioorg. Med.
Chem. Lett. 13 (2003) 331–334.
[27] C. Zechel, G. Backfisch, J. Delzer, H. Geneste, C. Graef, W. Hornberger,
A. Kling, U.E. Lange, A. Lauterbach, W. Seitz, T. Subkowski, Bioorg.
Med. Chem. Lett. 13 (2003) 165–169.
[28] H. Mao, P.J. Hajduk, R. Craig, R. Bell, T. Borre, S.W. Fesik, J. Am.
Chem. Soc. 123 (2001) 10429–10435.
[29] X.M. He, D.C. Carter, Nature 358 (1992) 209–215.
[30] D.C. Carter, J.X. Ho, Adv. Protein Chem. 45 (1994) 152–203.
[31] R.E. Olson, D.D. Christ, Chapter 33, Annu. Rep. Med. Chem. 31 (1996)
327–336.
[32] B. Meloun, L. Morávek, V. Kostka, FEBS Lett. 58 (1975) 134–137.
[33] H. Watanabe, S. Tanase, K. Nakajou, T. Maruyama, U. Kragh-Hansen,
M. Otagiri, Biochem. J. 349 (2000) 813–819.
[34] G. Sudlow, D.J. Brikket, D.N. Wade, Mol. Pharmacol. 11 (1975) 824–
832.
[35] J.B. Swaney, I.M. Klotz, Biochemistry 9 (1970) 2570–2574.
[36] T. Izumi, T. Kitagawa, Chem. Pharm. Bull. (Tokyo) 37 (1989) 742–745.
[37] H. Mao, P.J. Hajduk, R. Craig, R. Bell, B. Borre, S.W. Fesik, J. Am.
Chem. Soc. 43 (2001) 10429–10435.
[38] T. Lu, L.V. Lafrance, T.P. Markotan, J.J. Marugan, V.J. Marder, D.C.
U’Prichard, B.M. Anaclerio, Z. Guo, W. Pan, K.A. Leonard (2002)
WO2002060438.
[39] K. Niiya, E. Hodson, R. Bader, V. Byers-Ward, J.A. Koziol, E.F. Plow,
Z.M. Ruggeri, Blood 70 (1987) 475–483.
[40] M.S. Dennis, P. Mazur, M.S. Dennis, J.L. Seymour, R.A. Lazarus, Pro-
teins 15 (1993) 312–321.
[41] R.L. Rich, Y.S.N. Day, T.A. Morton, D.G. Myszka, Anal. Biochem. 296
(2001) 197–207.
[42] T.E. Rogers, P.G. Ruminski (2000) US 6013651.
[43] R.G. Chapman, E. Ostuni, S. Takayama, R.E. Holmlin, L. Yan, G.M.
Whitesides, J. Am. Chem. Soc. 122 (2000) 8303–8304.
[44] J. Samanen, Z. Jonak, D. Rieman, T.L. Yue, Curr. Pharm. Des. 3 (1997)
545–584.
[22] W.H. Miller, P.J. Manley, R.D. Cousins, K.F. Erhard, D.A. Heerding, C.
Kwon, S.T. Ross, J.M. Samanen, D.T. Takata, I.N. Uzinskas, C.C. Yuan,
[45] R.S. Obach, J.G. Baxter, T.E. Liston, M. Silber, B.C. Jones, F. MacIn-
tire, D.J. Rance, P. Wastall, J. Pharmacol. Exp. Ther. 283 (1997) 46–58.