Non-peptide αvβ3 antagonists. Part 4: Potent and orally bioavailable chain-shortened RGD mimetics

Article abstract of DOI:10.1016/S0960-894X(02)00396-7

Potent non-peptidic α_vβ₃ antagonists have been prepared where deletion of an amide bond from an earlier series of linear RGD-mimetics provides a novel series of chain-shortened α_vβ₃ antagonists with significantly improved oral pharmacokinetics. These chain-shortened α_vβ₃ antagonists represent structurally novel integrin inhibitors.

Full text of DOI:10.1016/S0960-894X(02)00396-7

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2463–2465
Non-Peptide ꢀvꢁ3 Antagonists. Part 4: Potent and Orally
Bioavailable Chain-Shortened RGD Mimetics
Paul J. Coleman,a,* Ben C. Askew,a,y John H. Hutchinson,a David B. Whitman,a
James J. Perkins,a George D. Hartman,a Gideon A. Rodan,b Chih-Tai Leu,b
Thomayant Prueksaritanont,c Carmen Fernandez-Metzler,c Kara M. Merkle,c
Robert Lynch,d Joseph J. Lynch,d Sevgi B. Rodanb and Mark E. Duggana
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bDepartment of Bone Biology and Osteoporosis Research, Merck Research Laboratories, West Point, PA 19486, USA
cDepartment of Drug Metabolism, Merck Research Laboratories, West Point, PA 19486, USA
dDepartment of Pharmacology, Merck Research Laboratories, West Point, PA 19486, USA
Received 12 March 2002; accepted 29 April 2002
Abstract—Potent non-peptidic avb3 antagonists have been prepared where deletion of an amide bond from an earlier series of linear
RGD-mimetics provides a novel series of chain-shortened avb3 antagonists with significantly improved oral pharmacokinetics.
These chain-shortened avb3 antagonists represent structurally novel integrin inhibitors. # 2002 Published by Elsevier Science Ltd.
Osteoporosis is a disease of the skeleton associated with
diminishing bone mass and an increased risk of debili-
tating fractures.1 In most instances, osteoporosis arises
from an imbalance between the activities of bone-resorbing
osteoclast cells and bone-depositing osteoblast cells.2 In
post-menopausal women, estrogen depletion results in
increased bone resorption due to the enhanced activity and
number of osteoclasts. Osteoclasts mediate bone resorption
by initial attachment to the bone surface followed by
secretion of matrix-degrading proteinases. Attachment of
osteoclasts to bone involves binding of the highly
expressed integrin avb3 to an RGD (arg-gly-asp) tripep-
tide sequence found in extracellular matrix proteins.3 The
integrin, avb3, is thought to be involved in not only
cellular adhesion of osteoclasts to the bone surface but
also in regulating their migration along the bone surface.
Antagonists that block the binding of RGD expressing
proteins to avb3 inhibit bone resorption in rodent models
and offer a potential therapy for the prevention and
treatment of osteoporosis.4
Replacement of the guanidine moiety of arginine with a
tetrahydro[1,8]naphthyridine (THN) provides integrin
selectivity for avb3 versus the fibrinogen receptor aIIbb3
(2, Fig. 1). Further enhancements in binding affinity
were realized by replacement of aspartic acid with a
3-aryl b-amino acid and inclusion of a pyrrolidinone
central constraint.5 Although receptor antagonists from
this class bound to human avb3 with sub-nanomolar
affinity, the oral pharmacokinetic profiles for these
zwitterionic moieties were generally poor. For example,
full-length antagonist 2 when dosed orally to dogs is
poorly bioavailable (10%) and rapidly cleared (Cl=20
mL/min/kg) from plasma. Herein, we describe a chain-
length truncation of these full-length avb3 antagonists
(represented by 2) to give a new class of chain-shortened
analogues with significantly improved pharmacokinetics.
The synthesis of the avb3 antagonists described in this
paper involved the coupling of 3-aryl-b-aminoesters6 to
the tetrahydro[1,8]naphthyridine pentanoic acid 5b
(Scheme 1). The preparation of carboxylic acid 5b began
with commercially available 5-acetylvaleric acid 3a.
Fisher esterification of 3a gave ketoester 3b in quanti-
tative yield. Friedlander condensation of ketoester 3b
with 2-amino-3-formylpyridine gave a mixture of two
naphthyridine isomers 4a and 4b (ratio 3:1). Although
separation of naphthyridine regioisomers was feasible by
flash chromatography, separation after naphthyridine
Previously, several reports from this laboratory have
detailed the design and synthesis of novel, high affinity
non-peptidic RGD mimetics as avb3 antagonists.5
*Corresponding author. Tel.: +1-215-652-4618; fax: +1-215-652-
7310; e-mail: paul_coleman@merck.com
yCurrent address: Amgen, Inc., Thousand Oaks, CA 91320, USA.
0960-894X/02/$ - see front matter # 2002 Published by Elsevier Science Ltd.
PII: S0960-894X(02)00396-7

2464
P. J. Coleman et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2463–2465
Figure 1. Evolution of full-length RGD mimetic.
Scheme 2. Chain-shortened RGD mimetics.
receptor binding assay to the full-length analogues 10
and the chain-shortened counterparts 11. Although the
potency of 11 is modestly compromised in comparison
to 10, the chain-shortened analogue 11 exhibits an
excellent pharmacokinetic profile when dosed orally to
dogs with good bioavailability (F=65%), low clearance
(Cl=5.8 mL/min/kg), and moderate plasma half-life
(T1/2=2.4 h). In contrast, our earlier series of full-length
analogues generally had much poorer pharmacokinetics
(Fꢀ10%; Cl>20 mL/min/kg). Chain-length was fur-
ther investigated in this amide-deleted series by increas-
ing the chain-length of 11 by one methylene unit (12;
SPAV3=5.9 nM) and by two methylene units (13;
SPAV3=5.8 nM).9 From this study, we concluded that
the chain-shortened 11, with eight atoms separating the
THN N-terminus and carboxylic acid, possessed a
potency-optimized chain length in the amide-deleted
series.
Scheme 1. Synthesis of avb3 antagonists. Reagents and conditions:
(a) MeOH, ClCH2CH2Cl, H2SO4; (b) 2-amino-3-formylpyridine,
proline, EtOH; (c) PtO2, H2 (1 atm), 34% (three steps); (d) LiOH,
EtOH, 79%; (e) 3-aryl-b-aminoester, EDC, HOBt, morpholine; (f)
LiOH, EtOH.
reduction proved to be more facile. Catalytic hydro-
genation of naphthyridines 4a and 4b was achieved in a
smooth fashion with platinum oxide to give the desired
tetrahydronapthyridine 5a. Saponification gave the
intermediate acid 5b. Carboxylic acid 5b was trans-
formed to receptor antagonists 7 by standard carbodii-
mide coupling to 3-aryl-b-aminoesters followed by ester
hydrolysis.
From previous studies, we discovered that 3-aryl moi-
eties have significant potential to affect receptor binding
affinity and to modify physicochemical properties of the
derived antagonists.5 Indeed, replacement of the 3-qui-
nolinyl substituent in 11 with either benzodioxole 14 or
6-dihydrobenzofuranyl 15 yielded potent chain-shor-
tened antagonists with reduced lipophilicity (Table 1).
The 5-pyrimidinyl substituent 17 provided a less potent
analogue with increased hydrophilicity. The benzox-
azolidone 18 is an appreciably less potent avb3 antago-
nist with an IC50 in the SPAV3 assay of 13 nM. None of
these avb3 antagonists had significant activity in our
platelet aggregation assays.
The avb3 antagonists prepared in this study were tested
for their ability to inhibit the in vitro binding of a high
affinity radioligand to human avb3 immobilized on
scintillation proximity beads (SPAV3).7 Selectivity for
the vitronectin receptor (avb3), versus the fibrinogen
receptor (aIIbb3) was determined by measuring inhibi-
tion of the rate of ADP-stimulated platelet aggrega-
tion.8 Full-length receptor antagonists typified by 8
(Scheme 2) have a linear backbone that spans ten atoms
between the terminal carboxylic acid and tetra-
hydro[1,8]naphthyridine (THN). In an effort to probe
the effects of amide deletion and chain length alterations
within this series of RGD mimetics, analogue 9 which
lacks the central asp-gly amide was prepared and eval-
uated. The chain-shortened antagonist 9 which has a
chain-length of eight atoms between the carboxylic acid
and THN terminus, has an IC50 in the SPAV3 assay
that is approximately 10-fold less potent than the cor-
responding full-length analogue 8. This result prompted
us to incorporate a potency enhancing 3-quinolinyl
substituent which imparted significant potency in our
Pharmacokinetic evaluations of these avb3 antagonists
were made following their oral and iv administration to
dogs (Table 1). Compared to the quinoline lead 11, the
dihydrobenzofuran analogue 15 displayed
a sig-
nificantly improved oral profile. Within this series of
compounds, it was evident that plasma clearance and
oral bioavailability were closely related to measured
Log P. Indeed, the more polar benzoxazolidone
antagonist 18 and 5-pyrimidinyl 17 suffer decreased
bioavailability and higher clearance relative to 11 and
14. From this series, dihydrobenzofuran 15 represents

P. J. Coleman et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2463–2465
2465
Table 1. Aryl substitution on chain-shortened backbone: potency and dog pharmacokineticsa
Entry
11
Ar
SPAV3 (IC50, nM)
1.6
Plaggin (IC50, nM)
4440 nM
F (%)
65
CI (mL/min/kg)
5.8
T1/2 (h)
Log P
0.76
2.4
14
15
16
17
1.5
3.0
>10 mM
>10 mM
>10 mM
>10 mM
94
99
81
10
1.2
1.2
2.0
26
2.4
2.5
5.5
2.4
0.47
0.53
23.0
3.3
0.29
ꢁ1.21
18
13.1
>10 mM
25
37
1.2
ꢁ0.24
aCompounds were dosed at 1 mpk po and 0.2 mpk iv to dogs (n=2).
Rodan, G. A.; Leu, C.-T; Prueksarinont, T.; Fernandez-Metzler,
C.; Ma, B.; Libby, L. A.; Merkle, K. M.; Stump, G. L.;
Wallace, A. A.; Lynch, J. J.; Lynch, R.; Duggan, M. E.
Bioorg. Med. Chem. Lett. 2002, 12, 31.
an optimal compound in terms of receptor potency,
pharmacokinetics, and physicochemical properties.
In summary, this series of chain-shortened avb3
antagonists provides high affinity, integrin-selective
ligands with good pharmacokinetics in dogs. This finding
represents the first demonstration that the molecular
framework of avb3 RGD mimetics can be dramatically
modified by the removal of two atoms between the
carboxylic acid and guanidine surrogate while preserving
integrin affinity and selectivity.10,11 Deletion of the arg-gly
amide from these tripeptide mimetics provides antagonists
with lower plasma clearance and improved bioavail-
ability. In keeping with our observations of full-length
antagonists, 3-aryl substituents provide a significant
enhancement in receptor affinity. Additionally, the 3-
aryl substituent is capable of modifying lipophilicity of
the derived antagonists. Within this series of chain-
shortened antagonists, measured Log P correlates with
lower plasma clearance. Further studies on these struc-
turally novel chain-shortened avb3 antagonists will be
reported in due course.
6. For the preparation of 3-aryl-aminoesters: (a) 5-Benzo-
dioxole: Zablocki, J. A.; Rico, J. G.; Garland, R. B.; Roger, T.
E.; Williams, K.; Schretzman, L. A.; Rao, S. A.; Bovy, P. R.;
Tjoeng, F. S.; Lindmark, R. J.; Toth, M. V.; Zupec, M. E.;
McMakins, D. E.; Adams, S. P.; Miyano, M.; Markos, C. S.;
Milton, M. N.; Paulson, S.; Herin, M.; Jacqmin, P.; Nichol-
son, N. S.; Panzer-Knodle, S. G.; Haas, N. F.; Page, J. D.;
Szalony, J. A.; Taite, B. B.; Salyers, A. K.; King, L. W.;
Campion, J. G.; Feigen, L. P. J. Med. Chem. 1995, 38, 2378.
(b) 3-Quinoline: Duggan, M. E.;Meissner, R. S.; Perkins, J. J.
WO 99/30709; Chem. Abstr. 1999, 404839. (c) Dihy-
drobenzofuranyl: Coleman, P. J.; Hutchinson, J. H.; Hunt, C.
A.; Lu, P.; Delaporte, E.; Rushmore, T. Tetrahedron Lett.
2000, 41, 5803. (d) 5-Pyrimidinyl: Bovy, P. R.; Rico, J. G.
Tetrahedron Lett. 1993, 34, 8015.
7. (a) The SPAV3 assay utilizes purified human recombinant
avb3 and 4-[2-(2-aminopyridin-6-yl)ethyl]benzoyl-2-(S)-4–
125iodo phenylsulfonylamino-b-alanine. For preparation of
the radioligand see: Hamill, T. G.; Duggan, M. E.; Perkins,
J. J. J. Label. Comp. Radiopharm. 2001, 44, 55.
8. Marguerie, G. A.; Plow, E. F.; Edgington, T. S. J. Biol.
Chem. 1979, 254, 5357.
9. Homologated analogues 12 and 13 were prepared by a
sequence analogous to that shown in Scheme 1.
References and Notes
10. A detailed study of constrained cyclic RGD-containing
pentapeptides where the Arg-Asp distances were varied and
analyzed has been reported by Kessler et al. This study indi-
cated that a bent conformation of the RGD sequence should
have higher affinity binding to avb3 versus aIIbb3: Haubner,
R.; Gratias, R.; Diefenbach, B.; Goodman, S. L.; Jonczyk, A.;
Kessler, H. J. Am. Chem. Soc. 1996, 118, 7461.
11. (a) Chain-shortened aIIbb3 antagonists have been repor-
ted: Kamiyama, T.; Satoh, T.; Umino, T.; Katoh, H.; Uemara,
F.; Nakamura, Y.; Yokose, K. J. Antibiot. 1995, 48, 1226. (b)
Hartman, G. D.; Duggan, M. E.; Hoffman, W. F.; Meissner,
R. J.; Perkins, J. J.; Zartman, A. E.; Naylor-Olsen, A. M.;
Cook, J. J.; Glass, J. D.; Lynch, R. J.; Zhang, G.; Gould, R. J.
Bioorg. Med. Chem. Lett. 1999, 9, 863.
1. Graul, A. I.; Sorbera, L. A.; Prous, J. R. Drugs Today 2001,
37, 703.
2. (a) Ershler, W. B. Front. Biomed. 2000, 41. (b) Eastell, R.
N. Engl. J. Med. 1998, 338, 736.
3. Duong, L. T.; Rodan, G. A. Frontiers Biosci. 1998, 3, d757.
4. (a) For reviews see: Miller, W. H.; Keenan, R. M.; Willette,
R. N.; Lark, M. W. Drug Discov. Today 2000, 5, 397. (b)
Hartman, G. D.; Duggan, M. E. Exp. Opin. Invest. Drugs
2000, 9, 1281. (c) Duggan, M. E.; Hutchinson, J. H. Exp.
Opin. Ther. Pat. 2000, 10, 1367.
5. (a) For Part 3 of this series and references to preceding
papers, see: Coleman, P. J.; Brashear, K. M.; Breslin, M. J.;
McVean, C. A.; Askew, B. C.; Hartman, G. D.; Rodan, S. B.;