Discovery and SAR of small molecule PAR1 antagonists

Article abstract of DOI:10.1016/j.bmcl.2009.12.072

High-throughput screening resulted in the identification of a small molecule inhibitor of PAR1. Optimisation of the initial hit led to the discovery of compounds 34 and 49, which displayed antithrombotic activity in an arteriovenous shunt model in the rat after iv administration.

Full text of DOI:10.1016/j.bmcl.2009.12.072

Bioorganic & Medicinal Chemistry Letters 20 (2010) 903–906
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and SAR of small molecule PAR1 antagonists
a
b
a
Ian Rilatt a, , Etienne Mirabel , Bruno Le Grand , Michel Perez
*
a Department of Medicinal Chemistry 4, Centre de Recherche Pierre Fabre, 17, Avenue Jean Moulin, 81106 Castres Cedex, France
b Department of Cardiovascular Diseases 2, Centre de Recherche Pierre Fabre, 17, Avenue Jean Moulin, 81106 Castres Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
High-throughput screening resulted in the identification of a small molecule inhibitor of PAR1. Optimi-
sation of the initial hit led to the discovery of compounds 34 and 49, which displayed antithrombotic
activity in an arteriovenous shunt model in the rat after iv administration.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 12 November 2009
Revised 16 December 2009
Accepted 17 December 2009
Available online 23 December 2009
Keywords:
PAR1
Antagonist
Thrombosis
SAR
Platelet activation in response to vascular injury occurs via a
number of signalling pathways, for which a variety of agonists,
such as thrombin, ADP, thromboxane A2, collagen, or epinephrine
are known. Among these, thrombin is probably the most potent
platelet activator, whose action is mediated via proteolytic cleav-
age of specific cell surface receptors known as protease activated
receptors (PARs).1 In humans, thrombin-induced platelet aggrega-
tion is mediated by one subtype of these receptors, termed PAR1.2
Since inhibition of the PAR signalling pathway could reduce the
number of activated platelets without inhibiting thrombin’s roles
in the coagulation cascade, it is thought that PAR1 antagonists
may hold the key to antithrombotic therapies without the associ-
ated risk of bleeding side effects seen with current treatments. In-
deed, SCH- 530348, a novel PAR1 antagonist, is currently in phase
III clinical trials for the treatment of acute coronary syndrome.3
Herein we wish to report our studies towards the development
of selective PAR1 antagonists.
Thus, nucleophilic substitution of commercially available alpha-
bromo ketones was used to furnish thioethers in good yields. One-
pot alpha halogenation of the crude products with NCS in CCl4
followed by nucleophilic substitution of the intermediate chlorides
O
S
1
N
Cl
N
Figure 1. Initial PAR1 inhibitor hit.
O
O
O
R
R
R
SR2
SR2
a
b
Br
Cl
An in-house high-throughput screening program, based on inhi-
bition of intracellular calcium release induced by a PAR1-selective
agonist peptide (SFLLR) in CHO cells,4 led to the discovery of com-
c
pound 1 (Fig. 1), which was found to be a 96% antagonist at 10
Analogues of 1 were synthesised via the route depicted in Scheme
1. All compounds were tested at 10 M in the calcium release
lM.
OH
OR3
l
O
SR2
X
R
SR2
X
SR2
X
R
R
screening assay, and for compounds displaying greater than 75%
inhibition, we also determined the pKb for the inhibition of
SFLLR-induced aggregation of human platelets.
e
d
N
N
N
Z
Y
Z
Y
Z
Y
Scheme 1. Reagents and conditions: (a) R2SH, DIEA or NaH, THF, 0 °C, 90–100%; (b)
NCS, CCl4, rt, 30 min; (c) heterocycle, DCM, rt, 16 h, ca. 50% two steps; (d) NaBH4,
MeOH, rt, 70–80%; (e) R3Br, NaH, DMF, 60 °C, 16 h, 50–80%.
* Corresponding author.
E-mail address: Ian.Rilatt@Pierre-Fabre.com (I. Rilatt).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.12.072

904
I. Rilatt et al. / Bioorg. Med. Chem. Lett. 20 (2010) 903–906
Table 3
yielded alpha-disubstituted aryl ethanones in ca. 50% yield. So-
dium borohydride reduction of the carbonyl group followed by
alkylation of the resulting alcohols with alkyl bromides furnished
the desired products as mixtures of enantiomers.
Initial optimization began with modification of the ether substi-
tuent (Table 1). In the screening assay, ethyl and methyl ethers
(compounds 2 and 3) were similar to hit structure 1, whilst larger
alkyl chains (compounds 5 and 6) and the alcohol functionality
(compound 4) were much less active. The platelet model revealed
that compounds 1 and 2 were equipotent whilst compound 3 was
slightly more active.
Screening of thioether analogues (Table 2) showed that modifi-
cation of the tert-butyl group led to a loss of potency. Surprisingly,
isopropyl thioether 9 was found to be much less active than the
corresponding tert-butyl thioether (compound 2).
We then proceeded to make changes to the aryl group (Table 3).
Amongst the compounds tested, only the 4-carboxymethyl- and 2-
nitro-substituted phenyl were not potent enough in the screening
assay not to be evaluated in the platelet model (compounds 17 and
19, respectively).
Antagonist activity of compounds 2 and 11–19
O
S
Ar
N
N
Compd
Ar
% antag.@ 10
l
Ma
Hum. Plat.b (pKb)
2
11
12
13
14
15
16
17
18
19
4-Cl-Ph
4-Br-Ph
2,4-Cl-Ph
Ph
Benzothiophene
4-CF3-Ph
Benzodioxane
4-CO2Me-Ph
3-OMe-Ph
2-NO2-Ph
99
100
100
92
5.05
5.06
5.01
4.46
5.20
5.38
4.79
—
98
100
88
37
76
33
0
—
a
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
l
M of SFLLR.
b
The latter test revealed that with respect to an unsubstituted
phenyl group (compound 13), halogenated phenyls were more po-
tent (compounds 2, 11 and 12), although the activity with respect
to hit structure 1 was the same. A 5-substituted benzothiophene
(compound 14) was more potent than compound 1, and a 4-trifluo-
romethylphenyl group (compound 15) even more so. A 6-substi-
tuted 1,4-benzodioxane (compound 16) was more active than the
unsubstituted phenyl ring, but less potent than the initial hit
structure.
intermediate in Scheme 1 with 1H-1,2,4-triazole led to a mixture
of 1- and 4-substituted triazole isomers, which were easily sepa-
rated by flash chromatography. Reaction with 4-chloro-1H-imidaz-
ole5 gave
a mixture of well-separated 4- and 5-substituted
imidazoles, but 4-methyl-1H-imidazole and 4-(trifluoromethyl)-
1H-imidazole6 led to predominately the 4-isomer. The alpha-
chloro intermediate was also reacted with 4,5-dimethyl-1H-imid-
azole7 in order to avoid the problem of regioisomers. The positions
of the substituent on the imidazole ring of the final products were
determined by 2D NMR.
We next turned our attention to modification of the imidazole
ring. Thus, nucleophilic substitution of the alpha-chloro ketone
Replacement or substitution of the imidazole had a large impact
upon antagonist activity (Table 4). In the screening test, pyrazole
20 was completely inactive. Potency was retained with triazole
21, whereas isomer 22 was a weak inhibitor. Potency was also re-
tained with a chlorine atom or a methyl group in the 4-position of
the imidazole ring (compounds 23 and 25) but not with chlorine in
the 5-position (compound 24). Compound 26, a trifluoromethyl
analogue of compound 25, was very weak inhibitor. Finally, a loss
of activity was found when the imidazole ring was substituted
with a methyl group in the 2-position (compound 27), or with
two methyl groups (compound 28). Out of the compounds tested
in the platelet model, only the 4-methylimidazole (compound
25) was found to be more potent than compound 1.
Table 1
Antagonist activity of compounds 1–6
OR
S
N
Cl
N
Compd
OR
% antag.@ 10
l
Ma
Hum. Plat. (pKb)b
1
2
3
4
5
6
OPr
OEt
OMe
OH
OBu
OBn
96
99
95
11
33
23
5.01
5.05
5.24
—
—
—
Table 4
a
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
l
M of SFLLR.
Antagonist activity of compounds 2 and 20–28
b
O
S
Table 2
Antagonist activity of compounds 2 and 7–10
N
Cl
X
Z
Y
O
SR
Compd
X
Y
Z
% antag.@ 10
l
Ma
Hum. Plat.b (pKb)
2
20
21
22
23
24
25
26
27
28
CH
N
N
CH
CH
CCl
CH
CH
C–CH3
C–CH3
CH
CH
CH
N
CCl
CH
C–CH3
C–CF3
N
N
CH
N
N
N
N
N
N
CH
N
99
0
5.05
—
4.64
—
4.60
—
5.36
—
N
Cl
99
34
96
59
92
20
63
64
N
Compd
SR
% antag.@ 10
l
Ma
Hum. Plat. (pKb)b
2
7
8
9
10
StButyl
SnButyl
SPentyl
SiPropyl
SCyclohexyl
99
76
19
40
3
5.05
4.58
—
—
—
—
—
C–CH3
a
a
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
l
M of SFLLR.
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
l
M of SFLLR.
b
b

I. Rilatt et al. / Bioorg. Med. Chem. Lett. 20 (2010) 903–906
905
Table 5
OH
b
Antagonist activity of compounds 29–38
a
O
Br
X
Ar
OR
Ar
S
c
N
O
O
d
e
39 : X = C, Ar = 4-Cl-Ph
40 : X = O, Ar = 4-Cl-Ph
41 : X = S, Ar = 4-Cl-Ph
Br
X
N
Ar
Ar
Compd
Ar
OR
% antag.@ 10
l
Ma
Hum. Plat.b (pKb)
29
30
31
32
33
34
35
36
37
38
2-Cl–Ph
2-Cl–Ph
3-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Br–Ph
4-Br–Ph
2,4-Cl–Ph
2,4-Cl–Ph
OEt
OPr
OEt
OMe
OEt
OPr
OEt
OPr
OEt
OPr
100
97
100
96
92
99
100
76
93
5.21
5.14
4.96
4.42
5.36
5.71
5.73
5.99
5.84
5.95
O
O
X
f
X
Ar
Ar
X2
N
N
g
h
88
42 : X = C, Ar = 4-Cl-Ph
43 : X = O, Ar = 4-Cl-Ph
2 : X = S, Ar = 4-Cl-Ph
OH
OEt
a
i
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
l
M of SFLLR.
X
X
b
Ar
Ar
N
N
44 : X = SO, Ar = 4-Cl-Ph
45 : X = SO2, Ar = 4-Cl-Ph
N
N
Following these results, a library of compounds was synthesised
using the optimal substituents. The most interesting results are
shown in Table 5. Activity was found to be consistently higher with
the 4-methylimidazole compounds than the non-substituted ana-
logues (results not shown).
An additional pharmacological assay was then performed in
order to better evaluate the antithrombotic activity of lead
compounds. Thus, an in vivo model consisting of an arteriovenous
extra-corporal shunt in anaesthetised rats was performed. The
Scheme 2. Reagents and conditions: (a) (1) Mg, Et2O, reflux, 2 h; (2) 4-chloro-
benzaldehyde, rt, 16%; (b) (1) sBuLi, tBuOK, À78 °C; (2) 4-chlorobenzaldehyde,
À78 °C to rt, 49%; (c) PCC, DCM, 100%; (d) X = S: tBuSH, NaH, THF, 0 °C to rt, 16 h,
100%; (e) X = C, X2 = Br: Br2, DCM; X = O, X2 = Cl: SO2Cl2; X = S, X2 = Cl: NCS, CCl4; (f)
imidazole, DCM (g) NaBH4, MeOH, 63% three steps when X = C, 42% three steps
when X = O, 49% three steps when X = S; (h) EtBr, NaH, DMF, 60 °C, 16 h, 60–70%; (i)
mCPBA, DCM, 16% X = SO, 16% X = SO2.
Further changes to the core structure were then performed in
order to improve the metabolic stability against animal micro-
somes. Optimisation began with replacement of the thioether
functionality of compound 2 with CH2, O, SO and SO2. The synthetic
pathway to these compounds is shown in Scheme 2.
Thus, condensation of 4-chlorobenzaldehyde with a Grignard
reagent derived from 1-bromo-3,3-dimethylbutane9 gave an inter-
mediate alcohol, which was oxidised to ketone 39 using PCC. Sim-
ilarly, condensation of the same aldehyde with an organometallic
resulting from the deprotonation of tert-butylmethyl ether10 gave
an intermediate alcohol which furnished ketone 40 after PCC oxi-
dation. Ketone 41 was synthesised via nucleophilic substitution
of an alphabromoketone with tert-butylthiol. The imidazole was
introduced via halogenation of the alpha keto position, followed
compounds were compared with
a potent PAR1 antagonist,
F16618, recently published by our group8 (Table 6).
Compounds 34, 36 and 38 are all potent antagonists of PAR1 in
the human platelet aggregation model, although only compound
34 was found to be active in vivo. F16618 was more potent than
compound 34 in the rat shunt model, despite its weaker activity
in the human platelet model.
Further tests revealed that the metabolic stability of com-
pounds 34, 36 and 38 was reasonable against human liver micro-
somes but poor against rat liver microsomes (Table 7). This
might partly explain the non-consistent results between the two
models, although other pharmacokinetic parameters are likely to
be involved.
Table 6
Antagonist activity of F16618 and compounds 34, 36 and 38
Table 8
Compd
% antag.@ 10
l
Ma
pKb
Hum. Plat. (pKb)b
AV Shuntc (iv,%)
a
Antagonist activity and metabolic stability of compounds 2 and 42–45
F16618
34
36
99
99
76
88
—
5.51
—
5.30
5.71
5.99
5.95
58
45
0
O
X
38
—
7
N
a
b
c
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
% increase in occlusion time of an arteriovenous shunt in the rat (1.25 mg/kg iv).
lM of SFLLR:% at 10 lM or pKb.
Cl
N
Compd
X
% antag.@ Hum.
Microsomal
stabilityc
(human)
Microsomal
10
l
Ma
plat.b
stabilityc (rat)
(pKb)
Table 7
Metabolic stability of F16618 and compounds 34, 36 and 38
2
42
43
44
45
S
C
99
74
85
15
5.05
4.12
4.54
—
41
57
74
59
35
2
3
25
39
47
Compd
Microsomal stabilitya (human)
Microsomal stabilitya (rat)
O
SO
SO2 13
F16618
34
36
48
28
23
28
27
0
0
—
a
b
c
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
% of parent compound remaining after 60 min incubation with rat or human
lM of SFLLR.
38
2
a
% of parent compound remaining after 60 min incubation with rat or human
hepatic microsomes.
hepatic microsomes.

906
I. Rilatt et al. / Bioorg. Med. Chem. Lett. 20 (2010) 903–906
Table 9
Table 10
Antagonist activity of compounds 43 and 46–50
Antagonist activity and metabolic stability of F16618 and compounds 43, 49 and 50
Compd
%
Hum.
plat.b
(pKb)
AV
Microsomal
stabilityd
(human)
Microsomal
stabilityd
(rat)
R2
O
antag.@
10
shuntc
(iv,%)
l
Ma
O
Ar
F16618 99
5.30
4.54
5.73
5.32
58
0
43
0
27
74
23
57
48
25
3
N
43
49
50
85
81
98
N
R
3
a
b
c
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
% increase in occlusion time of an arteriovenous shunt in the rat (1.25 mg/kg iv).
% of parent compound remaining after 60 min incubation with rat or human
lM of SFLLR.
Compd
Ar
R
R2
% antag.@ 10 l
Ma
Hum. plat. (pKb)b
43
46
47
48
49
50
4-Cl–Ph
3,4-Cl–Ph
4-F–Ph
4-CF3–Ph
3,4-Cl–Ph
2,4-Cl–Ph
H
H
H
Me
Me
Me
Et
Pr
Pr
Et
Pr
Et
85
93
18
98
81
98
4.54
5.04
—
5.20
5.73
5.32
d
hepatic microsomes.
a
compound 34. Despite its metabolic stability, compound 43 was
inactive in the shunt model. Although this compound was a much
weaker antagonist in the platelet model, the lack of in vivo activity
is likely due to other pharmacokinetic parameters.
Inhibition of calcium release induced by 1
Inhibition of SFLLR-induced human platelet aggregation.
l
M of SFLLR.
b
by reaction of the crude intermediate chloride or bromide with
imidazole. Sodium borohydride reduction of the carbonyl group
followed by alkylation of the resulting alcohol with bromoethane
yielded products 2, 42 and 43 as mixtures of enantiomers. mCPBA
oxidation of product 2 led to a mixture of sulfoxide 44 and sulfone
45. The results of the modifications upon the activity and meta-
bolic stability of the compounds are shown in Table 8.
In conclusion, we have identified a series of novel PAR1 antag-
onists through screening of our in-house library. SAR studies re-
sulted in an increase in potency, and compounds 34 and 49 were
found to display strong antithrombotic activity in vivo. Although
the compounds lacked metabolic stability against rat microsomes,
the compounds were found to be sufficiently stable against human
liver microsomes. Work is underway in order to further optimise
these compounds.
Replacement of sulfur (compound 2) by carbon (compound 42)
resulted in a loss of potency and no increase in stability against rat
microsomes. The oxygen analogue (compound 43) was also less ac-
tive than compound 2, although more potent than compound 42.
However, a large increase in metabolic stability was observed. Oxi-
dation of the thioether to a sulfoxide or a sulfone (compounds 44
and 45, respectively) resulted in almost inactive molecules, but
which displayed increased stability against rat microsomes.
We then proceeded to synthesise a small library of oxygen ana-
logues to see if the potency could be improved with retention of
metabolic stability. The compounds were synthesised by the same
route as depicted in Scheme 2. The results are shown in Table 9.
As with the sulfur compounds, a methyl group in the 4-position
of the imidazole ring gave more potent antagonists with respect to
the unsubstituted analogues, best shown by the platelet model for
compounds 46 and 49. Unfortunately, the metabolic stability
observed with compound 43 was not retained (Table 10). Never-
theless, compound 49 was found to possess potent antithrombotic
activity in vivo, and its overall profile was similar to that of
References and notes
1. Ossovskaya, V. S.; Bunnett, N. W. Physiol. Rev. 2004, 84, 579.
2. Voss, B.; McLaughlin, J. N.; Holinstat, M.; Zent, R.; Hamm, H. E. Mol. Pharmacol.
2007, 71, 1399.
3. Chackalamannil, S.; Wang, Y.; Greenlee, W. J.; Hu, Z.; Xia, Y.; Ahn, H.; Boykow,
G.; Hsieh, Y.; Palamanda, J.; Agans-Fantuzzi, J.; Kurowski, S.; Graziano, M.;
Chintala, M. J. Med. Chem. 2008, 51, 3061.
4. Andrade-Gordon, P.; Maryanoff, B.; Derian, C. K.; Zhang, H.-C.; Addo, M. F.;
Darrow, A. L.; Eckardt, A. J.; Hoekstra, W. J.; McComsey, D. F.; Oksenberg, D.;
Reynolds, E. E.; Santulli, R. J.; Scarborough, R. M.; Smith, C. E.; White, K. B. Proc.
Natl. Acad. Sci. U.S.A. 1999, 96, 12257.
5. Hayakawa, Y.; Kimoto, H.; Cohen, L. A.; Kirk, K. L. J. Org. Chem. 1998, 63, 9448.
6. Kimoto, H.; Fujii, S.; Cohen, L. A. J. Org. Chem. 1982, 47, 2867.
7. D’Sa, A.; Cohen, L. A. J. Heterocycl. Chem. 1991, 28, 1819.
8. Perez, M.; Lamothe, M.; Maraval, C.; Mirabel, E.; Loubat, C.; Planty, B.; Horn, C.;
Michaux, J.; Marrot, S.; Letienne, R.; Pignier, C.; Bocquet, A.; Nadal-Wollbold, F.;
Cussac, D.; de Vries, L.; Le Grand, B. J. Med. Chem. 2009, 52, 5826.
9. Chen, X.; Hortelano, E. R.; Eliel, E. L.; Frye, S. V. J. Am. Chem. Soc. 1992, 114, 1778.
10. Corey, E. J.; Eckrich, T. M. Tetrahedron Lett. 1983, 24, 3165.