Discovery of a potent, non-peptide bradykinin B1 receptor antagonist

Article abstract of DOI:10.1021/ja0353457

Bradykinin (BK) plays an important role in the pathophysiological processes accompanying pain and inflammation. Selective bradykinin B₁ receptor antagonists have been shown to be anti-nociceptive in animal models and could be novel therapeutic

Full text of DOI:10.1021/ja0353457

Published on Web 05/31/2003
Discovery of a Potent, Non-peptide Bradykinin B1 Receptor Antagonist
Dai-Shi Su,* M. Kristine Markowitz, Robert M. DiPardo, Kathy L. Murphy,‡ C. Meacham Harrell,‡
Stacy S. O’Malley,‡ Richard W. Ransom,‡ Raymond S. L. Chang,‡ Sookhee Ha, Fred J. Hess,‡
Douglas J. Pettibone,‡ Glenn S. Mason,§ Susan Boyce,§ Roger M. Freidinger, and Mark G. Bock
Departments of Medicinal Chemistry and Neuroscience, Merck Research Laboratories, West Point, Pennsylvania
19486, and Neuroscience Research Centre, Merck Research Laboratories, Terlings Park, Harlow,
Essex, CM20 2QR, England
Received March 27, 2003; E-mail: daishi_su@merck.com
Table 1. Receptor Binding Affinities of Dihydroquinoxalinone
Analogues
Bradykinin (BK) is an autacoid peptide produced by the catalytic
action of kallikrein enzymes on plasma and tissue precursors termed
kininogens. It plays an important role in the pathophysiological
processes accompanying pain and inflammation. Its biological
actions are mediated by two known G-protein-coupled receptors
named B1 and B2. The BK B2 receptor is constitutively expressed
in most cell types, whereas the BK B1 receptor is induced during
inflammatory insults.1 In animal models, BK B1 receptor agonists
produce hyperalgesia, an effect blocked by peptide BK B1 receptor
antagonists.2 Recent data from transgenic BK B1 receptor knockout
mice has implicated a role for the BK B1 receptor in inflammation
and algesia.3 Accordingly, selective and effective BK B1 receptor
antagonists hold promise as novel therapeutic agents for the
treatment of pain and inflammation. In this communication we
report the development of the first non-peptide BK B1 antagonist
with subnanomolar affinity for the BK B1 receptor and functional
potency in an animal pain model.
The discovery process was initiated with a high-throughput screen
(HTS) of the Merck sample collection which uncovered the
dihydroquinoxalinone lead compound, 1. In racemic form, 1 exhibits
a Ki value of 1.4 µM for binding to the human BK B1 receptor
with selectivity vs the human BK B2 receptor (Ki > 10 µM) (Table
1). To improve the potency and pharmacological parameters of this
screening lead we followed classical structure activity relationship
(SAR) trends. This optimization process was also facilitated by a
modeling study wherein we docked 1 into a homology model of
the BK B1 receptor constructed on the crystal structure of bovine
rhodopsin (Figure 1).4,5 The latter exercise suggested that 1 interacts
mainly with transmembrane spanning domains 3 (TM3) and TM7
of the receptor. The aromatic ring of the sulfonamide chain and
the dihydroquinoxalinone core are both accommodated by a
hydrophobic pocket formed by residues Ile97, Trp98, Trp103,
Ile113, and Phe302. Residue Gln 295 interacts with the sulfonamide
oxygen of 1 via a hydrogen bond. A weak interaction is also
observed between residue Asn114 and the N1 nitrogen of the
dihydroquinoxalinone ring in 1. On the basis of this analysis, it
was determined that the phenyl sulfonamide moiety plays a key
role in the substrate-receptor binding event and was therefore
targeted for initial modification.
a All Ki values are calculated for competition binding assays utilizing
cloned human BK B1 receptor and are the average of at least two
experiments. b The human BK B2 Ki values for all compounds are greater
than 10,000 nM.
Among the first analogues prepared which displayed measurable
advances over the lead compound 1 was derivative 2 (Table 1).6
The potency of racemic 1 was improved by increasing the
lipophilicity of the phenyl sulfonamide group and introducing a
2-methoxyl group in the phenyl amide ring. Resolution of 2
provided compound 3, the S-enantiomer, and 4, the corresponding
R-enantiomer. As anticipated, receptor binding affinity and the
Figure 1. Homology model of BK B1 receptor with 1 and 11 docked in
the putative binding site. Transmembrane domains and extracellular loops
are represented in green and yellow ribbons, respectively. Analogues 1 and
11 are colored in magenta and cyan, respectively.
configuration of the asymmetric center are linked. In this case, the
R-isomer is approximately 2.5-fold more potent than the S-isomer.
‡ Department of Neuroscience, Merck Research Laboratories.
§ Neuroscience Research Centre, Merck Research Laboratories.
9
7516
J. AM. CHEM. SOC. 2003, 125, 7516-7517
10.1021/ja0353457 CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
Replacement of trimethyl phenyl in 4 with the 3,4-dichloro phenyl
ring yielded a compound of equivalent potency (compound 5).
Capping the N1 nitrogen atom with a methyl group, which interrupts
the interaction between the N1 nitrogen and Asn114 in the TM3
region as suggested by our model, caused a loss of potency
(compound 6). The increase in electron density of the phenyl amide
ring in 5, by the addition of two methoxyl groups, resulted in a
further potency increase (compound 7). Other neutral functional
groups at various positions of the phenyl amide ring were examined
(data not shown), but the 2,3,4-trimethoxyl groups were preferred.
Replacement of the sulfonamide group with other functionalities
resulted in a significant decrease in potency (data not shown) which
supports the modeling observation of a hydrogen-bond interaction
between residue Gln 295 in TM7 and one of the oxygen atoms in
the sulfonamide group.
Analogue 11 is selective for the BK B1 versus the BK B2 receptor
(human BK B2, Ki > 10 µM), and it demonstrates excellent
functional antagonist potencies in a FLIPR assay10 that are in concert
with the receptor binding affinities. In addition, 11 is selective
versus a number of human opioid receptor subtypes (IC50: 7.6 µM
(µ), 3.2 µM (δ), and 7.3 µM (κ)), and in a panel of assays11
representing 170 enzymes, receptors, and transporters, 11 exhibited
over 5000-fold selectivity for the human BK B1 receptor.
To assess the in vivo antinociceptive efficacy of 11, it was
examined in a rabbit assay of inflammatory hyperalgesia.12 In this
model, 11 inhibits the spinal nociceptive reflex response to noxious
pinch of an inflamed paw in a dose-dependent manner. It is
efficacious at inhibiting the nociceptive reflex response to both low-
and high-intensity stimuli. Analogue 11 is more effective than
morphine at inhibiting low-intensity stimuli, indicating that it
possesses potent antinociceptive activity.
In close proximity to the hypothesized receptor binding site is
the extracellular domain 4 (EC4) containing two acidic residues,
Glu273 and Asp291. We speculated that the receptor binding
affinity of 7 could be further strengthened by exploiting a potential
binding interaction between the ligand and EC4. Therefore, a basic
amino group was introduced to give 8 and subsequently 9, which
displayed a further 50-fold boost in receptor binding potency. Since
there are two accessible acidic residues associated with EC4, a
logical extension was to incorporate a bidentate basic group.
Accordingly, the 4-ethylamine group in 8 was replaced with an
imidazoline ring, yielding compound 10. The imidazoline ring in
10 appears to be suitably positioned to simultaneously interact with
Glu273 and Asp291 by insertion of an ethyl linker between the
side chain amide nitrogen and the phenyl imidazoline ring system.
This modification afforded 11, the optimal compound in this study
which displays subnanomolar affinity for the human BK B1 receptor.
Receptor mutagenesis was utilized in an attempt to further
delineate regions of the BK B1 receptor that are involved with
binding 11. Mutation of amino acid residue Asn114 in TM3 of the
human BK B1 receptor to Ala results in an approximately 15-fold
loss in affinity for 11, with no loss in affinity for kinin peptides.
This amino acid residue is on the same helical face as Lys118, a
residue that has been implicated in binding the C-terminal carboxyl
group of the peptide agonist des-Arg10-kallidin.7 Significantly, the
binding affinity of 11 is not affected by the Lys118Ala mutation.
Therefore, 11 appears to be interacting with a region in TM3 that
is in close proximity, but is not identical, to that involved with
binding peptide agonists.
The compounds disclosed in this work represent the first
generation of dihydroquinoxalinones which are a useful base for
the design of BK B1 receptor antagonists. In particular, the
pharmacokinetic and physicochemical properties of these com-
pounds are suboptimal. Analogue 11 is the most prominent member
of the series to emerge from this study, and its further evaluation,
as well as those of its congeners, is in progress. The results from
these studies will be disclosed in due course.
Acknowledgment. We are pleased to acknowledge the efforts
of Drs. P. Kunapuli, S. M. Pitzenberger, C. W. Ross, Mrs. J. S.
Murphy, Mr. C. F. Homnick, and Mrs. J. F. Kaysen. We are grateful
to Drs. N. J. Anthony, B. D. Dorsey, S. D. Kuduk, M. R. Wood,
and D-M. Feng for useful discussions.
Supporting Information Available: Assay protocols, all experi-
mental details for 11, and characterization of 3-11 (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
References
(1) For excellent reviews, see: (a). Couture, R.; Harrisson, M.; Vianna, R.
M.; Cloutier, F. Eur. J. Pharmacol. 2001, 429, 161-176. (b). Bock, M.
G.; Longmore, J. Curr. Opin. Chem. Biol. 2000, 4, 401-406. (c). Marceau,
F. Immunopharmacology 1995, 30, 1-26. (d). Regoli, D.; Barabe, J.
Pharmacol. ReV. 1980, 32, 1-46.
(2) Rupniak, N. M. J.; Longmore, J.; Hill, R. G. In Molecular Basis of Pain
Induction; Wood, J., Ed.; John Wiley Press: 2000; 149-174.
(3) (a) Pesquero, J. B.; Araujo, R. C.; Heppenstall, P. A.; Stucky, C. L.; Silva,
J. A., Jr.; Walther, T.; Oliveira, S. M.; Pesquero, J. L.; Paiva, A. C. M.;
Calixto, J. B.; Lewin, G. R.; Bader, M. Proc. Natl. Acad. Sci. U.S.A. 2000,
97, 8140-8145. (b). Ferreira, J.; Campos, M. M.; Araujo, R.; Bader, M.;
Pesquero, J. B.; Calixto, J. B. Neuropharmacology 2002, 43, 1188-1197.
(4) Palczewski, K., Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima, H.;
Fox, B. A.; Trong, I. L.; Teller, D. C.; Okada, T.; Stenkamp, R. E.;
Yamamoto, M.; Miyano, M. Science 2000, 289, 739-745.
Previous mutagenesis studies have also implicated TM7 of the
BK B1 kinin receptor in binding peptide agonists.8 We found that
mutation of Gln295, in TM7, to Ala results in an approximately
26-fold loss in affinity for 11. In contrast, the affinity of Gln295Ala
for kinin peptides is unaltered. This result further supports the
contention that the amino acid residues that bind 11 are distinct
from those involved with binding peptide agonists.
(5) A manuscript with detailed coordinates of this model is in preparation.
(6) See Supporting Information for the synthetic route and preparation of
compounds.
(7) Fathy, D. B.; Mathis, S. A.; Leeb, T.; Leeb-Lundberg, L. M. F. J. Biol.
Chem. 1998, 273, 12210-12218.
Analogue 11 exhibits excellent binding affinities across BK B1
receptors in different species (Ki: 0.034 nM (human), 0.05 nM
(rabbit), 1.28 nM (dog), and 62.0 nM (rat)). Relative to the human
B1 receptor, the affinity of the dog and rat BK B1 receptors for 11
is reduced 38- and 1800-fold, respectively. The EC4 domain of
the human BK B1 receptor has been implicated in determining the
species selectivity for kinin peptides.9 We tested the affinity of 11
for mutant chimeric receptors, in which this region of the human
BK B1 receptor was replaced with that of either the dog or rat BK
B1 kinin receptor. The affinity of the dog EC4 chimera is reduced
6-fold for 11 and the rat EC4 chimeric receptor affinity is reduced
30-fold. It can reasonably be concluded that EC4 confers a portion
of the species selectivity observed for 11.
(8) Bastian, S.; Pruneau, D.; Loillier, B.; Robert, C.; Bonnafous, J.-C.; Paquet,
J.-L. J. Biol. Chem. 2000, 275, 6107-6113.
(9) (a) Hess, J. F.; Hey, P. J.; Chen, T.-B.; Pettibone, D. J.; Chang, R. S. L.
Int. Immunopharmacol. 2002, 2, 1747-1754. (b). Fathy, D. B.; Kyle, D.
J.; Leeb-Lundberg, L. M. F. Mol. Pharmacol. 2000, 57, 171-179.
(10) Fluorescence imaging plate reader, IC50: 0.18 nM (human), 0.26 nM
(rabbit), 16.30 nM (dog), and 163.5 nM (rat). See Supporting Information
for assay protocol.
(11) Panlabs (MDS Pharma Services, Bothell, WA).
(12) Intravenous administration, ID50: 3.5 µg/kg (low-intensity stimuli) and
16.4 µg/kg (high-intensity stimuli). For morphine, ID50: 299 µg/kg (low-
intensity stimuli). See Supporting Information for results and the assay
protocol.
JA0353457
9
J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003 7517