Discovery and SAR development of 2-(phenylamino) imidazolines as postacyclin receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2003.10.070

On the basis of screening hits (1a,b), a series of selective, high affinity prostacyclin receptor antagonists was developed. The optimized lead compound 25d [(4,5-dihydro-1H-imidazol-2-yl)-[4-(4-isopropoxybenzyl)phenyl] amine] had analgesic activity in the rat.

Full text of DOI:10.1016/j.bmcl.2003.10.070

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1053–1056
Discovery and SAR development of 2-(phenylamino) imidazolines
as postacyclin receptor antagonists
Robin D. Clark,a,* Alam Jahangir,a,* Daniel Severance,a,y Rick Salazar,b
Thomas Chang,b David Chang,b Mary Frances Jett,b Steven Smithc and Keith Bleyb,{
aDepartment of Medicinal Chemistry, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
bDepartment of Pharmacology, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
cDepartment of Drug Metabolism and Pharmacokinetics, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
Received 3 September 2003; accepted 27 October 2003
Abstract—On the basis of screening hits (1a,b), a series of selective, high affinity prostacyclin receptor antagonists was developed.
The optimized lead compound 25d [(4,5-dihydro-1H-imidazol-2-yl)-[4-(4-isopropoxybenzyl)phenyl]amine] had analgesic activity in
the rat.
# 2003 Elsevier Ltd. All rights reserved.
Considerable evidence indicates that prostaglandin I2
(PGI2 or prostacyclin) is a major mediator of inflam-
mation and pain.1 Prostacyclin is one of several prosta-
noids released upon tissue damage and causes
sensitization or direct activation of sensory nerve end-
ings. Prostaglandin E2 has previously been assumed to
be the major prostanoid involved in these processes, but
recent evidence suggests that prostacyclin may play a
more important role than has heretofore been recog-
nized. For example, high concentrations of the prosta-
cyclin metabolite 6-keto-PGF1a have been found in
inflammatory exudates, for example from synovial fluid
from human arthritic knee joints2 and from the perito-
neal cavity fluid of mice injected with chemical irri-
tants.3 More compelling evidence for a major role of
prostacyclin in inflammatory pain is the altered pain
and inflammatory responses observed in transgenic mice
lacking the prostacyclin receptor (IP receptor). In these
IP receptor deficient animals, responses to pain and
inflammation (acetic acid-induced writhing and carra-
geenan-induced edema) were reduced to levels observed
when wild-type mice were treated with indomethacin.4
Thus it can be hypothesized that an IP receptor
antagonist could have therapeutic potential as an
analgesic/anti-inflammatory agent with the important
caveat that untoward cardiovascular and thrombotic
side effects would not be induced. Herein we describe the
discovery of the first-reported IP receptor antagonists.
Prostacyclin receptor affinity was determined by radio-
ligand binding utilizing displacement of the high-affi-
nity ligand [3H]iloprost from prostacyclin receptors
endogenously expressed in NG108-15 cells, a rodent
neuroblastoma cell line. Initial screening was performed
on a subset (460 compounds chosen for structural
diversity) of the Syntex compound library. Fortuitously,
arylguanidines 1a and 1b were found to have modest
affinity with pKi values of 6.6 and 6.2, respectively.
Structure–activity relationship studies were then under-
taken based on these leads with the goal of producing
higher affinity ligands.
Modification of the guanidine moiety of 1a and 1b
(Table 1) led to identification of the 2-(phenylamino)
imidazolines 4a,b and tetrahydropyrimidine 8 with
improved affinity.5 A degree of steric intolerance was
noted by the reduced affinity of ligands 2 and 6. In
* Corresponding author. E-mail: robin.clark5@verizon.net
y
Current address: Neurogen, 35 Northeast Industrial Park Road,
Branford, CT 06405, USA.
{
Current address: NeurogesX, San Carlos Business Park, 981F
Industrial Road, San Carlos, CA 94070-4117, USA.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.10.070

1054
R. D. Clark et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1053–1056
Table 1. Receptor bindings for compounds 1–16
Compd
X
R
pKia
Compd
X
R
pKia
Compd
X
R
pKia
1a
OMe
6.6
5
OMe
7.2
11
OMe
5.9
1b
2
Cl
OMe
Cl
6.2
<5
<5
7.3
6
7
OMe
Cl
6.2
6.3
7.1
5.9
5.3
12
13
14
15
16
OMe
OMe
OMe
OMe
OMe
6.5
6.7
3
8
OMe
OMe
OMe
<5
<5
5.0
4a
4b
OMe
Cl
9
7.1
10
a Displacement of [3H]iloprost in rodent NG-108-15 neuroblastoma cells, mean of two or more determinations ꢀ0.1.
addition, the 2-(amino)imidazoline was clearly preferred
over the methylene linked imidazoline group as in 10.
Other basic groups led to a significant decrease in affi-
nity; therefore, the 2-(phenylamino)-imidazoline group
was maintained for subsequent SAR development.
therefore, functional antagonism can be determined
using a classical cyclic AMP (cAMP) assay. When tested
for antagonism of carbaprostacyclin-stimulated cAMP
accumulation in human SH-SY5Y neuroblastoma cells,
4j was indeed found to be a competitive antagonist with
a pKB of 8.1. However, the moderate affinity of this
compound, coupled with poor oral bioavailability due to
slow absorption and high first-pass metabolism precluded
further development.
Compounds in Tables 2 and 3 were prepared by reac-
tion of the requisite aniline (17) with 2-chloroimidazo-
line6 as exemplified in Scheme 1. Transposition of the
para-thiophenyl group in compound 4a to the ortho and
meta positions (18 and 19) led to a dramatic decrease in
receptor affinity, as did constraint into the dibenzothio-
phene system (des-methoxy analogue 20) (Table 2).
Similarly, constraint of the imidazoline group into the
tetrahydro-imidazo[2,1-b] ring system (21) resulted in
loss of affinity.
Continued SAR development indicated that a modest
increase in affinity was obtained when the link between
the phenyl rings in prototype 4a was changed from sul-
fur to carbon as in bibenzyl analogue 25a (Table 4).
This prompted synthesis8 (Scheme 2) of a series of rela-
ted ethers which were tested for functional antagonism
(Table 5). Increased activity generally correlated with
the size of the ether moiety and antagonists with pKB
values in the nanomolar range were obtained. These
were subsequently differentiated on the basis of in-vitro
metabolic and in-vivo pharmacokinetic parameters and
the iso-propoxy derivative 25d was determined to have
the best overall profile. This compound had an oral
bioavailability of 85% in the dog with a T1/2 of 8 h.
The importance of the distal phenyl ring in 4 was
emphasized by the low affinity of analogues 22 and 23
(Table 2). Variation of the substituent pattern in this
phenyl ring indicated that, at least for a methoxy group,
the para position was optimal (Table 3). Although para-
alkoxy groups resulted in the highest affinity, a number
of other para substituents groups were tolerated.
The iso-propoxy derivative 4j, with a pKi of 7.5 at the
prostacyclin receptor was evaluated to assess whether
this prototypical compound was a functional antago-
nist, and hence a suitable candidate for in vivo eval-
uation. The prostacyclin receptor is a G-protein coupled
receptor that is positively linked to adenylyl cyclase;7
Selectivity of 25d over other receptors was evaluated
through several protocols. PGE2 stimulates cAMP for-
mation through the EP2 (and EP4) receptor; however,
25d failed to inhibit this response in SH-SY5Y cells
(pKB <5) indicating that this compound is not an EP2,4
receptor antagonist. Similarly, 25d did not display

R. D. Clark et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1053–1056
Table 2. IP receptor binding for compounds 18–23
1055
Compd
pKia
18
5.4
Scheme 1. (a) 4-Chloronitrobenzene, DMF, 120 ꢁC; (b) SnCl2, EtOH,
reflux; (c) 2-chloroimidazoline, THF, heat.
19
20
21
22
23
6.2
5.5
<5
<5
5.5
Scheme 2. (a) Nitrobenzoyl chloride, CS2, AlCl3, rt; (b) HBr, HOAc,
reflux; (c) RX, K2CO3, DMF, 60–70 ꢁC; (d) H2, 10% Pd/C, EtOH,
HCl; (e) 2-chloroimidazoline, i-PrOH, reflux.
Table 4. IP receptor binding for compounds 26–31
Compd
X
pKia
5a
24
25a
27
26
28
29
S
O
CH2
CO
SO2
NMe
(bond)
7.3
6.8
7.6
6.5
6.8
6.3
6.0
a Displacement of [3H]iloprost in rodent NG-108-15 neuroblastoma
cells, mean of two or more determinations ꢀ0.1.
Table 3. Receptor binding for compounds 4a–r
a Displacement of [3H]iloprost in rodent NG-108-15 neuroblastoma
cells, mean of two or more determinations ꢀ0.1.
Table 5. IP receptor antagonism for compounds 25a–o
Compd
X
pKia
a
Compd
R
pKB
SEM
4a
4b
4c
4d
4e
4f
4g
4h
4i
4j
4k
4l
4m
4n
4r
4p
4o
4q
4-OMe
4-Cl
2,6-Cl2
H
2-OMe
3-OMe
7.3
7.1
6.2
6.1
5.4
6.4
6.1
6.2
7.4
7.5
7.0
7.3
6.6
6.4
7.3
<5
6.5
7.1
25a
25b
25c
25d
25e
25f
25g
25h
25i
25j
25k
25l
25m
25n
25o
Me
Et
n-Pr
i-Pr
n-Bu
i-Bu
n-Hexyl
7.7
8.2
8.8b
8.8
9.0b
8.8
8.2
7.7b
8.7
8.0
8.0
8.9
8.6
8.4b
8.5
0.2
0.2
0.1
2,5-(OMe)
3,4-(OCH2CH2O)
4-OEt
0.3
0.1
CH2Ph
CH2CF3
4-O-i-Pr
4-Me
4-CF3
4-i-Pr
4-t-Bu
4-CO2Et
4-CO2H
4-NO2
0.1
0.2
0.2
0.1
0.2
CH2CH2OH
CH2CH2OMe
Cyclopentyl
Cyclohexyl
4-Tetrahydropyranyl
4-Tetrahydropyranylmethyl
0.2
a Antagonism of carbaprostacylin-stimulated c-AMP accumulation
in human SH-SY5Y neuroblastoma cells, mean of two or more
determinations ꢀSEM unless otherwise noted.
b n=1.
4-MeSO2NH
a Displacement of [3H]iloprost in rodent NG-108-15 neuroblastoma
cells, mean of two or more determinations ꢀ0.1.

1056
R. D. Clark et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1053–1056
significant affinity for cloned human EP1, EP3, FP, and
TP receptors as determined by radioligand binding. In a
standard profiling panel, this compound did not display
significant affinity for 30 other receptors and ion chan-
nels with the exception of the human a2A receptor (pKi
6.5) and, perhaps not suprisingly given its chemical
structure, the imidazoline I2 binding site (pKi 8.7). The
potential pharmacological ramifications of the latter
result are not clear.9
In conclusion, relatively simple structural modification
of the screening hits 1a,b led to a two-log order
improvement in prostacyclin receptor affinity. The high
affinity antagonist 25d demonstrated analgesic activity
in the rat and would appear to have sufficient selectivity
and bioavailability for determining the therapeutic
potential of this class of agents.
References and notes
Thus 25d was considered to be a sufficiently selective,
bioavailable prostacyclin receptor antagonist to warrant
in vivo testing in analgesic assays. In the rat acetic acid-
induced writhing assay, 25d caused a significant reduc-
tion in mean number of writhes over a dose range of 1–
10 mg/kg, ip with an ED50 of 4 mg/kg. The effect of 25d
on carrageenan-induced paw hyperalgesia in the rat was
evaluated and a significant reduction in mean with-
drawal pressure was observed over the dose range of
0.3–10 mg/kg, iv (ED50 2.8 mg/kg) and 1–100 mg/kg po
(ED50 18 mg/kg). Interestingly, 25d had no effect on
bleeding time in the mouse at a dose of 100 mg/kg, po,
nor were there significant cardiovascular effects in the
rat at analgesic doses.
1. Bley, K. R.; Hunter, J. C.; Eglen, R. M.; Smith, J. A. M.
Trends Pharmacol. Sci. 1998, 19, 141.
2. (a) Bombardieri, S.; Cattani, P.; Ciabattoni, G.; Di
Munno, O.; Pasero, G.; Patrono, C.; Pinca, E.; Pugliese,
F. Br. J. Pharmacol. 1981, 73, 893. (b) Brodie, M. J.;
Hensby, C. N.; Parke, A.; Gordon, D. Life Sci. 1980, 27,
603.
3. (a) Berkenkopf, J. W.; Weichman, B. M. Prostaglandins
1988, 26, 693. (b) Doherty, N. S.; Beaver, T. H.; Chan,
K. Y.; Coutant, J. E.; Westrich, G. L. Br. J. Pharmacol.
1987, 39.
4. Murata, T.; Ushikubi, F.; Matsuoka, T.; Hirata, M.;
Yamasaki, A.; Sugimoto, Y.; Ichikawa, A.; Aze, Y.;
Tanaka, T.; Yoshida, N.; Ueno, A.; Oh-ishi, S.;
Narumiya, S. Nature 1997, 388, 678.
5. Compounds 4 and 8 have been previously described and
4b is specifically claimed as a fungicide: Jpn. Kokai
Tokkyo Koho, JP 54063080, 1979; Chem. Abstr. 1979, 91,
157761.
6. Trani, A.; Bellasio, E. J. Het. Chem. 1974, 11, 257.
7. Ushikubi, F.; Hirata, M.; Narumiya, S. J. Lipid Media-
tors Cell Signalling 1995, 12, 343.
8. Bley, K. R.; Clark, R. D.; Jahangir, A.; Kowalczyk, B. A.;
Lopez-Tapia, F. J.; Muehldorf, A. V.; O’Yang, C.; Sun,
T. W. U.S. Patent 6,472,536, 2002. Chem. Abstr. 2002,
130, 252356.
9. (a) Eglen, R. M.; Clark, R. D. Exp. Opin. Invest. Drugs
1995, 4, 663. (b) Brown, C. M.; MacKinnon, A. C.;
Redfern, W. S.; Williams, A.; Linton, C.; Stewart, M.;
Clague, R. U.; Clark, R.; Spedding, M. Br. J. Pharmacol.
1995, 116, 1737.
10. This situation is not completely without precedent in the
prostanoid field as the leukotriene LTB4 receptor antago-
nist CGS-25019C is a benzamidine: Brooks, C. D. W.;
Summers, J. B. J. Med. Chem. 1996, 39, 2629.
11. (a) Boie, Y.; Rushmore, T. H.; Darmon-Goodwin, A.;
Grygorczyk, R.; Slipetz, D. M.; Metters, K. M.;
Abramovitz, M. J. Biol. Chem. 1994, 269, 12173. (b)
Nakagawa, O.; Tanaka, I.; Usui, T.; Harada, M.; Sasaki,
Y.; Itoh, H.; Yoshimasa, T.; Namba, T.; Narumiya, S.;
Nakao, K. Circulation 1994, 90, 1643.
12. Sitham, J.; Stojanovic, A.; Merenick, B.; O’Hara, A.;
Hwa, J. J. Biol. Chem. 2003, 278, 4250.
13. Ilopost: pKI=8.5 (wild-type) and 8.1 (mutant), 25f: pKI=8.7
(wild-type) and 5.7 (mutant), 25l: pKI=8.7 (wild-type)
and 5.5 (mutant).
A priori, the structures of these guanidine containing
prostacyclin receptor antagonists appear rather unusual
compared to the structure of the native ligand
prostacyclin which contains a carboxylic acid group as a
primary structural feature and recognition element for
receptor binding.10 The molecular biology of prosta-
noid receptors is well described7 and all prostanoid
receptors sequenced thus far, including the human
prostacyclin receptor,11 contain an arginine in the
seventh transmembrane domain (TM7) which is pro-
posed to be the carboxylic acid binding site. In the
human prostacyclin receptor, this arginine is at position
279 in the second helical loop of TM7.12 We hypothe-
sized that the basic ligands bound to the aspartic acid
residue at position 288, which is below Arg-279 in TM7
(i.e., closer to the intracellular side) in the fourth
helical loop. Site-directed mutagenesis of the human
prostacyclin receptor wherein Asp-288 was replaced
with serine provided confirmation of this hypothesis. In
the D288S mutant receptor, binding of the basic
ligands 25f and 25l was decreased by three orders of
magnitude, whereas [3H]iloprost binding was not affec-
ted.13 Thus, it would appear that binding of these gua-
nidine containing ligands at Asp-288 interferes with
binding of prostacyclin, the primary locus of which is
presumed to be at Arg-279. The nature of this antag-
onism remains unclear although allosteric modulation
could be hypothesized.