Piperazinyl-glutamate-pyridines as potent orally bioavailable P2Y12 antagonists for inhibition of platelet aggregation

Article abstract of DOI:10.1016/j.bmcl.2009.06.075

Polymer-assisted solution-phase (PASP) parallel library synthesis was used to discover a piperazinyl-glutamate-pyridine as a P2Y₁₂ antagonist. Exploitation of this lead provided compounds with excellent inhibition of platelet aggregation as mea

Full text of DOI:10.1016/j.bmcl.2009.06.075

Bioorganic & Medicinal Chemistry Letters 19 (2009) 4657–4663
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Piperazinyl-glutamate-pyridines as potent orally bioavailable P2Y12
antagonists for inhibition of platelet aggregation
b
a
b
b
b
John J. Parlow a, , Mary W. Burney , Brenda L. Case , Thomas J. Girard , Kerri A. Hall , Ronald R. Hiebsch ,
Rita M. Huff b, Rhonda M. Lachance b, Deborah A. Mischke a, Stephen R. Rapp b, Rhonda S. Woerndle a,
Michael D. Ennis a
*
a Department of Medicinal Chemistry, Pfizer Global Research & Development, 700 Chesterfield Parkway West, Chesterfield, MO 63017, USA
b Department of Biology, Pfizer Global Research & Development, 700 Chesterfield Parkway West, Chesterfield, MO 63017, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Polymer-assisted solution-phase (PASP) parallel library synthesis was used to discover a piperazinyl-
glutamate-pyridine as a P2Y12 antagonist. Exploitation of this lead provided compounds with excellent
inhibition of platelet aggregation as measured in a human platelet rich plasma (PRP) assay. Pharmaco-
kinetic and physiochemical properties were optimized leading to compound (4S)-4-[({4-[4-(methoxy-
methyl)piperidin-1-yl]-6-phenylpyridin-2-yl}carbonyl)amino]-5-oxo-5-{4-[(pentyloxy)carbonyl]pipera-
zin-1-yl}pentanoic acid 22J with good human PRP potency, selectivity, in vivo efficacy and oral
bioavailability.
Received 9 April 2009
Revised 19 June 2009
Accepted 22 June 2009
Available online 25 June 2009
Keywords:
P2Y12 receptor
GPCR antagonist
Antiplatelet
Ó 2009 Elsevier Ltd. All rights reserved.
Antithrombotic
Cardiovascular disease
Polymer-assisted solution-phase
Cardiovascular and cerebrovascular diseases remain the most
common cause of mortality in the western world. Current therapy
for these diseases includes antiplatelet agents such as aspirin,
dipyridamole, glycoprotein IIb/IIIa antagonists, and thienopyri-
dines.1 PlavixÒ (clopidogrel), a thienopyridine approved for the
reduction of atherosclerotic events (stroke, myocardial infarction
(MI), and death) in patients with atherosclerosis documented by
a recent event (stroke, MI or established peripheral vascular dis-
ease), acts by blocking adenosine diphosphate (ADP)-stimulated
platelet aggregation. ADP is an important platelet agonist that in-
duces a primary aggregation response and contributes to second-
ary aggregation following release from platelet dense granules
following activation by other agonists. ADP-induced platelet aggre-
gation is mediated by a dual receptor system involving activation
of P2Y1 and P2Y12 receptors, both members of the G-protein cou-
pled receptor (GPCR) family.2 Experimental studies have demon-
strated that selective blockade of either receptor is sufficient to
inhibit platelet activation. However, P2Y1 has ubiquitous expres-
sion whereas P2Y12 is primarily a platelet specific receptor and
thus represents a more attractive therapeutic target for selective
modulation of ADP-induced platelet activation. Several groups
have research efforts aimed towards identifying inhibitors of the
P2Y12 receptor3 including some recent patents4 issued by Actelion
which inspired us to describe our effort.
The P2Y12 receptor has been identified as the molecular target
of clopidogrel.5 Clopidogrel is a prodrug, the active metabolite of
which irreversibly and selectively inhibits the P2Y12 receptor; thus
antiplatelet efficacy requires a loading dose and several days of
treatment to achieve its full effect.6 Once it is activated, the drug
becomes irreversibly bound to platelets. As a result, clopidogrel
has a slow onset and slow offset of pharmacological action. This
makes it less effective in acute settings and difficult to manage if
a patient bleeds, experiences a trauma, or requires emergency sur-
gery. In addition, there are subsets of individuals who either do not
metabolize the prodrug adequately or who are resistant to the ef-
fects of clopidogrel (ꢀ12–25% of patients).7 It is anticipated that a
direct acting P2Y12 inhibitor will not be associated with such diffi-
culties and will therefore exhibit a significant improvement in effi-
cacy while maintaining a better safety profile. Therefore, we
sought to discover and develop direct acting, reversible P2Y12
antagonists through parallel medicinal chemistry to address the
unmet medical need for safe and effective oral antiplatelet agents.
Quinoline derivatives I have been reported as antagonists of the
platelet P2Y12 receptor (Fig. 1).8 The lead compound from this class
is a pro-drug (R4 side-chain possesses ethyl ester). Metabolism
(systemic esterases) of the ester pro-drug to its corresponding acid
provides the active metabolite, which contains two carboxylic acid
* Corresponding author. Tel.: +1 636 247 3494.
E-mail address: john.j.parlow@pfizer.com (J.J. Parlow).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.06.075

4658
J. J. Parlow et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4657–4663
R4
initial library synthesis in which an ethyl carbamate and a m-tolyl
ring system were substituted on the terminal piperazine nitrogen
with either a glycine or S-glutamic acid as the amino acid. The aryl
acid inputs 3 were initially biased towards fused ring systems
mimicking the quinoline and naphthalene. The amines 2 were cou-
pled to the aryl acids 3 using polymer-bound carbodiimide 4 as the
coupling agent with hydroxybenzotriazole to afford the crude
amide products. Upon completion of the reaction, a mixed-bed
containing polymer-bound amine 6 and polymer-bound isocya-
nate 5 was added to the reaction mixture to sequester hydroxyben-
zotriazole and any unreacted acid 3 or amine starting material 2
followed by filtration and evaporation to afford purified products.
Deprotection of the t-butyl ester using TFA in DCM provided the
desired products 7.
O
O
R1
N
N
NH
N
R2
R
I
Figure 1. Berlex quinoline core structures.
groups. It would be desirable to eliminate the need for a pro-drug
and improve upon the physical chemical properties as well as po-
tency. Interestingly, only quinolines and naphthalenes were re-
ported as the right-hand aryl moieties. We approached analog
synthesis by considering the molecule as being composed of three
parts: the piperazine, the amino acid, and the quinoline. Parallel
syntheses allowing for variations of all three parts of the molecule
were then developed.
Several hundred compounds were prepared by this synthesis
and screened in a human P2Y12 receptor binding assay at 5–
10 lM concentration, and Ki’s were determined for those com-
pounds with >50% inhibition. These compounds were also tested
as antiplatelet agents by measuring their inhibitory action on the
in vitro aggregation of human platelet rich plasma (PRP) stimu-
Initial efforts were focused at replacing the quinoline ring with
other heteroaryl and aryl ring systems. It was envisioned that a
polymer-assisted solution-phase (PASP) synthesis approach for
the aryl carboxamides could be accomplished through a simple
amide bond-forming reaction (Scheme 1). Thus, reacting aryl car-
boxylic acids with the amine would afford analogs with alternate
aryl ring systems. Three amine templates 2a–c were selected for
lated by 20
lM ADP using a turbidimetric method. The compounds
were initially assayed at 100 or 50
lM and IC50’s were determined
for compounds with >50% inhibition of platelet aggregation.
Emphasis, in regard to the SAR, was placed on the PRP potency
cHex
i)
N
O
O
O
O
C
4
R1
N
N
R1
R
N
N
N
+
NH
NH2
R
OH
R2
R2
2a-c
3
7a-c
ii)
N C O
5
a R1 = EtOCO, R2 = H
R'
N
b R1 = EtOCO, R2 = (CH2)2CO2tBu
c R1 = m-tolyl, R2 = (CH2)2CO2tBu
H
6
TFA/DCM
iii)
Scheme 1. Polymer-assisted solution-phase (PASP) library synthesis. Reagents and conditions: (i) 1.5 equiv 4, NMM, HOBt, DCM/DMF; (ii) excess 5 and 6, additional DCM;
(iii) 10% TFA/DCM.
Table 1
Binding and PRP activity data for selected library compounds9
R4
X3 X5
O
X1 R6
O
NH
N
R2
N
R1
R4
R6
X1
X3
X5
R1
R2
Kia (nM)
IC50
(l
M)
b
Compound
7A
7B
7C
7D
7E
7F
7G
7H
7I
Ph
H
Ph
H
Ph
H
Ph
H
H
Ph
H
Ph
H
Ph
H
H
N
N
N
N
CH
CH
CH
N
N
N
CH
CH
CH
CH
CH
CH
CH
CH
N
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
N
EtOCO
EtOCO
EtOCO
EtOCO
EtOCO
EtOCO
EtOCO
EtOCO
EtOCO
EtOCO
EtOCO
m-Tolyl
m-Tolyl
(CH2)2CO2H
(CH2)2CO2H
(CH2)2CO2H
(CH2)2CO2H
(CH2)2CO2H
(CH2)2CO2H
(CH2)2CO2H
(CH2)2CO2H
(CH2)2CO2H
H
15
76
>5400
>5700
209
373
3640
484
356
1190
733
1610
5.7
>5700
>100
>100
71
>100
>100
>50
>100
74
>50
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
N
CH
CH
CH
CH
CH
7J
7K
7L
7M
N
N
N
H
>100
>100
>100
(CH2)2CO2H
(CH2)2CO2H
a
Membranes from CHO cells expressing recombinant human P2Y12 receptors incubated with 33P ADP and compound. Ki values are corrected from IC50 using Cheng and
Prusoff equation and are the geometric mean of n = 2 or greater.
b
IC50 values are from human PRP incubated with 20 lM ADP.

J. J. Parlow et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4657–4663
4659
i, ii
R
O
O
H
H
N
N
N
H
N
+
6
N
N
E
N
R1
HN
O
N
O
8
9
10
iii
X
R1Br (Cl)
R1NCO
O
Ot-Bu
O
OH
R1NSO
R1COCl
R1SO2Cl
R1OCOCl
Scheme 2. Polymer-assisted solution-phase (PASP) synthesis of substituted piperazine analogs 10. Reagents and conditions: (i) 4 equiv electrophile 9, TEA, DCM; (ii) excess 6,
DCM; (iii) 20% TFA/DCM.
as this was an indication of the functional activity taking into ac-
count the effect of protein binding. Many different heterocyclic
ring systems were evaluated, and an apparent trend was observed.
Compounds with a heteroatom, preferably a nitrogen, ortho to the
amide position and a phenyl ring ortho to the heteroatom exhibited
P2Y12 binding activity. A second iteration with the heteroaryl acids
possessing this type of substitution pattern led to the discovery of
compound 7A possessing a 4,6-diphenylpyridine ring system with
mides, and amides were inactive. Small aliphatic ureas and
thioureas exhibited good binding activity and moderate PRP po-
tency, while the aliphatic carbamates were optimal with respect
to both binding and PRP activity. A representative set of carbamate
analogs is shown in Table 2. The SAR clearly shows that carbamates
with aliphatic straight chains of four to six carbon length are pre-
ferred. When unsaturation, branching, or a heteroatom is intro-
duced into the carbamate side-chain, PRP activity is lost.
a P2Y12 binding Ki of 15 nM and a PRP IC50 of 76
l
M. In an effort to
Additional chemistry efforts explored replacing the 4-phenyl
substituent on the pyridine ring of lead 7A with other groups such
as substituted amines and ethers. The 4-chloropyridine intermedi-
ate 11 allows for substitution at the 4-position with nitrogen
nucleophiles as shown in Scheme 3. The 4-chloro was displaced
by heating intermediate 11 with an excess of amine 12 and trieth-
ylamine in DMSO at 100 °C to afford the desired 4-aminopyridine
analogs 13. Displacement of the chlorine at the 4-position afforded
clean products with either the free carboxylic acid or the t-butyl
ester group present. Attempts to displace the 4-chloro with oxygen
nucleophiles provided limited success. As a result, 4-hydroxypyri-
dine intermediate 14 was prepared and under Mitsunobu condi-
tions or direct alkylation provided 4-oxygen analogs 17 as shown
in Scheme 4. Direct alkylation of the 4-hydroxypyridine 14 was
accomplished using 2 equiv of the electrophile 16 with cesium car-
bonate as the base and a catalytic amount of potassium iodide in
DMF to afford the desired 4-ether analogs 17. Alternatively,
employing Mitsunobu conditions with 2 equiv of the alcohol 15
using DEAD and triphenylphosphine in THF provided the 4-ether
follow-up on the lead compound 7A, a focused library was pre-
pared which explored elementary structural changes around the
pyridine ring system. A representative set of compounds from this
library is shown in Table 1. The unsusbstituted pyridine ring, com-
pound 7B, was devoid of any type of activity. Replacing the 6-phe-
nyl ring with a hydrogen, compound 7C, also resulted in a loss of
activity. However, when replacing the 4-phenyl ring with a hydro-
gen, compound 7D, PRP activity was maintained. Removing the
nitrogen from the pyridine ring system, regardless of the phenyl
substitution, resulted in loss of binding activity with no PRP activ-
ity (7E,F). An additional nitrogen at the 3-position with a 6-phenyl
substitution, compound 7I, maintained PRP activity, where as the
additional nitrogen at the 5-position, compound 7H, resulted in
no PRP activity.
The SAR trends across the different amine templates were con-
sistent. The glycine analogs (7J,K, R2 = H) had reduced binding
activity and no PRP activity while the S-glutamic acid piece
(R2 = (CH2)2CO2H) was required for PRP activity. The ethyl carba-
mate group on the piperazine nitrogen was preferred over the m-
tolyl for PRP activity. While the m-tolyl substitution provided com-
pounds with excellent binding affinity, 7L, no PRP activity was ob-
served with any of these analogs, presumably due to high protein
binding.9
Table 2
Binding and PRP activity data for selected straight chain carbamate analogs 109
Upon discovery of the pyridine ring system, modifications to
other parts of the molecule were explored. Efforts were focused
on derivatization of the piperazine nitrogen. With no difference
in PRP activity of the 4,6-diphenyl (7A) versus 6-phenylpyridine
(7D) ring system, the 6-phenylpyridine ring was used as the start-
ing template for piperazine derivatization. Scheme 2 shows the
synthesis of substituted piperazine analogs. The amine 8 was re-
acted with an excess of electrophile 9 in the presence of triethyl-
amine. The electrophiles used (alkyl halides, chloroformates, acid
halides, isocyanates, thioisocyanates, and sulfonyl halides) were
selected to represent a diverse set, affording products 10 contain-
ing alkyl, carbamate, amide, urea, thiourea, or sulfonamide func-
tionality. Upon completion of the reaction, polymer-bound amine
6 was added to the reaction mixture to sequester any unreacted
electrophile 9 followed by filtration and evaporation to afford puri-
fied products. Deprotection of the t-butyl ester using TFA in DCM
provided the desired substituted piperazine products 10.
O
O
N
10
NH
N
O
N
R1X
O
OH
b
Compound
R1
Kia (nM)
IC50 (lM)
10A
7D
OMe
OEt
OPr
2440
209
27
29
11
43
7.0
14
>100
71
10C
10D
10E
10F
10G
10H
62
28
15
19
OBu
OPent
OHex
OHept
OOct
56
>100
a
Membranes from CHO cells expressing recombinant human P2Y12 receptors
incubated with 33P ADP and compound. Ki values are corrected from IC50 using
Cheng and Prusoff equation and are the geometric mean of i = 2 or greater.
Many substituted piperazine compounds were prepared to pro-
vide a diverse set of analogs. In general the aliphatics, sulfona-
b
IC50 values are from human PRP incubated with 20 lM ADP.

4660
J. J. Parlow et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4657–4663
Cl
NR1R2
i
O
O
R1R2NH
O
O
N
+
N
NH
NH
13
11
12
N
N
O
O
N
N
EtO
EtO
O
O
OH
OH
Scheme 3. Synthesis of 4-aminopyridine analogs 13. Reagents and conditions: (i) 1.5–20 equiv 12, Et3N, DMSO, 100 °C.
OH
OR
ROH
15
O
O
1) i or ii
2) iii
O
O
N
+
N
or
NH
NH
17
14
N
N
RX
O
O
N
N
16
EtO
EtO
X= Br, Cl, I, etc.
O
O
OtBu
OH
Scheme 4. Synthesis of 4-ether pyridine analogs 17. Reagents and conditions: (i) 2 equiv 16, Cs2CO3, KI, DMF, rt to 100 °C; (ii) 2 equiv 15, DEAD, PPh3, THF; (iii) 10% TFA/DCM.
analogs 17. Using both procedures allowed for a greater diversity
of monomer inputs 15 and 16, ultimately providing a wide array
of 4-ether analogs. Deprotection of the t-butyl ester using TFA in
DCM afforded the desired final ether products 17.
Table 3
Binding and PRP activity data for a representative set of 4-ether and 4-aminopyridine
analogs9
OR/NR1R2
Kia (nM)
IC50
(l
M)
b
Compound
Table 3 shows the data for a representative set of compounds of
both the 4-ether and 4-aminopyridine libraries of 17 and 13
respectively. Substituting the 4-position with either an amine or
oxygen substituent led to a favorable enhancement in both the
P2Y12 binding and PRP activity. In the ether series, substituting
the oxygen with alkyl chains of varying length attenuated the
binding activity, but the PRP activity remained equivalent to that
of the methoxy analog 17A (data not shown). Comparable PRP
activity was observed with the benzyl ether analog 17B, and rela-
tively flat SAR was observed across various substituted benzyl
17A
17B
17C
17D
17E
17F
OMe
OCH2Ph
O(CH2)2OMe
O(CH2)3OH
O(CH2)2NH2
O(CH2)3NH2
96
28
92
68
328
214
39
31
5.8
9.0
7.5
3.8
17G
17H
68
76
1.9
4.0
O
NH
NH
O
ether analogs with PRP activity in the range of 9–39 lM (data
not shown). Extending a hydroxy or alkyl ether two to four carbon
atoms away from the 4-oxygen atom increased the PRP activity, as
seen with compounds 17C and 17D. A similar trend was observed
by extending a primary or secondary amine two to four carbon
atoms away from the 4-oxygen atom. In particular, the 4-piperi-
dine ether analogs, 17G and 17H, were two of the most potent
compounds of the series. The racemic mixture of 17I also exhibited
good PRP activity.
NEt2
17I
81
3.0
O
O
13A
13B
13C
13D
13E
NHMe
NHPr
NEt2
35
64
23
99
114
45
28
>100
14
NH(CH2)2OMe
NH(CH2)2OH
8.1
Similar trends were observed in the 4-aminopyridine series.
Small secondary alkyl amines (13A, B) were preferred over tertiary
amines (13C). As the alkyl chain was lengthened (>4 carbons) or
branched, activity was lost (data not shown). Compounds with
an oxygen heteroatom 13D–F, as a hydroxy or ether, as part of
the secondary amine group (several atoms removed from the
nitrogen) increased PRP activity relative to the corresponding alkyl
group. Extending an amine group two to four carbon atoms away
from the 4-nitrogen atom increased PRP activity. This trend was
most notable with substituted piperazines 13J and piperidines
13K–L, as they are the most potent compounds of the 4-aminopyr-
idine series.
The initial efforts described above at optimizing both the 4-po-
sition of the pyridine ring and the piperazine carbamate occurred
simultaneously. Each alteration led to compounds with good PRP
potency. The next obvious step was to make both changes within
the same molecule to determine whether an additive effect could
be achieved. A multi-step PASP synthesis was devised allowing
13F
455
21
HN
O
13G
13H
NHCH2Ph
NH(CH2)2NMe2
26
270
81
36
NH
NH
13I
13J
13K
190
47
8.5
NMe2
2.3
2.0
2.8
N
N
N
34
N
13L
27
N
NH2
a
Membranes from CHO cells expressing recombinant human P2Y12 receptors
incubated with 33P ADP and compound. Ki values are corrected from IC50 using
Cheng and Prusoff equation and are the geometric mean of n = 2 or greater.
b
IC50 values are from human PRP incubated with 20 lM ADP.

J. J. Parlow et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4657–4663
4661
Cl
NR1R2
R1R2NH
i)
O
O
H
H
12
N
N
N
18
N
N
N
ii)
N C O
19
O
O
N
O
N
O
5
O
O
O
Ot-Bu
iii
Ot-Bu
O
O
iv)
R
NR1R2
NR1R2
O
Cl
R
21
O
O
H
H
v)
N
H
6
N
N
N
22
N
20
N
N
O
O
O
N
HN
vi
O
O
OH
O
Ot-Bu
R
Scheme 5. PASP synthesis of 4-aminopyridine-piperazine carbamate analogs 22. Reagents and conditions: (i) 1.5 equiv 12, Et3N, DMSO, 100 °C; (ii) excess 5, DCM; (iii)
(Ph3P)4Pd, morpholine, MeCN; (iv) 21, Et3N, DCM; (v) excess 6, DCM; (vi) 10% TFA/DCM.
for substitution at both positions as shown in Scheme 5. Com-
pound 18 was selected as the starting template with the piperazine
nitrogen protected as the Alloc and the acid group protected as the
t-butyl ester. The protecting groups were chosen to allow for
orthogonal deprotection in a parallel format. The first step of the
synthesis involved reaction of the starting template 18 with excess
amine 12 at 100 °C. After total consumption of the template, poly-
mer-bound isocyanate 5 was added to sequester the unreacted
amine 12. Simple filtration and rinsing with dichloromethane
yielded a filtrate whereupon evaporation of the solvents left highly
purified product 19 from each parallel reaction. Selective deprotec-
tion of the Alloc group was accomplished using 1 equiv of morpho-
line with tetrakis(triphenylphosphine)palladium (0) followed by
filtration through CeliteÒ to afford the amine intermediate 20.
The morpholine derivative side-product was carried on to the next
step and could serve as the base during acylation. Deprotection
using polymer-bound piperazine and palladium on carbon was
unsuccessful. The next step in the synthesis called for derivatiza-
tion of the piperazine nitrogen. The amine intermediate 20 was re-
acted with an excess of chloroformate 21 in the presence of
triethylamine. Upon completion of the reaction, polymer-bound
amine 6 was added to the reaction mixture to sequester any unre-
acted chloroformate 21 followed by filtration and evaporation to
afford purified products. Deprotection of the t-butyl ester using
TFA in DCM provided the desired bi-substituted products 22.
A focused library was designed in which ethyl, pentyl, 2-cyclo-
pentylethyl, and hexyl chloroformates 21 were selected based on
some of the most active carbamate analogs (Table 2). The amine in-
puts 12 included substituted piperidines and N-methyl-2-(methyl-
sulfonyl)ethanamine. The data for this library is shown in Table 4
and demonstrates the additive effects of optimizing the piperazine
carbamate and the 4-position of the pyridyl ring. Improved PRP po-
tency was observed, with the pentyl and hexyl carbamates pre-
ferred in most cases. The 4-aminomethylpiperidine analog 22N
had exceptional potency with a PRP IC50 of 770 nM.
sumption of the 4-hydroxy template 23, polymer-bound amine 6
was added to sequester the unreacted alkyl halide 24. It was ob-
served that with some reactions, despite excess electrophile or
prolonged reaction times, the reaction would not go to completion.
In these cases the polymer-bound amine 6 was sufficiently basic
enough to also sequester the unreacted 4-hydroxy template 23.
Simple filtration and rinsing with dichloromethane yielded a fil-
trate whereupon evaporation of the solvents left purified product
25. Alloc deprotection was attempted with tetrakis(triphenylphos-
phine)palladium (0) as previously used, but with an oxygen at the
4-position, triphenylphosphine and triphenylphosphine oxide by-
products tended to carry through the synthesis, resulting in impure
final products even after chromatography. Thus, an anthracene-
tagged palladium catalyst 28 was employed to allow for easy
sequestration of any phosphine by-products.10 The reaction was
conducted using the anthracene-tagged palladium reagent 28
and excess DEA in dichloromethane. Upon complete deprotection
of the Alloc group, evaporation removed the DEA and allyldiethyl-
Table 4
Binding and PRP activity data for focused 4-aminopyridine library9
b
Compound
R
NR1R2
Kia (nM)
IC50
(l
M)
22A
22B
22C
22D
Et
Pent
Hex
48
15
36
80
7.6
4.1
17.7
17.5
N
SO2Me
2-Cyclopentylethyl
22E
22F
22G
22H
Et
Pent
Hex
42
11
24
40
9.8
2.4
6.4
N
N
N
OH
2-Cyclopentylethyl
57.8
22I
22J
22K
22L
Et
Pent
Hex
18
15
33
—
9.9
3.9
7.6
—
OMe
2-Cyclopentylethyl
22M
22N
22O
22P
Et
Pent
Hex
14
7
18
69
3.3
0.77
1.6
14.1
A similar multi-step PASP synthesis was designed to allow for
various carbamates at the piperazine nitrogen with oxygen substit-
uents at the pyridine 4-position as shown in Scheme 6. Compound
23 was selected as the starting template with the hydroxy at the 4-
position allowing for the preparation of ether analogs. The first
step of the synthesis was alkylation of the 4-hydroxy using 2 equiv
of electrophile 24 with cesium carbonate as the base. After con-
NH2
2-Cyclopentylethyl
a
Membranes from CHO cells expressing recombinant human P2Y12 receptors
incubated with 33P ADP and compound. Ki values are corrected from IC50 using
Cheng and Prusoff equation and are the geometric mean of n = 2 or greater.
b
IC50 values are from human PRP incubated with 20 lM ADP.

4662
J. J. Parlow et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4657–4663
OH
OR
OR
O
H
O
O
N
i, ii
O
O
O
iii, iv
N
N
N
25
N
RX
+
24
HN
NH
NH
O
23
26
N
N
O
O
N
N
O
OtBu
O
+
O
O
O
OtBu
OtBu
R
O
Cl
21
v, ii, vi
OR
O
O
NH
H
N
NHR'
N
N
N
O
Ph
P
O
N
O
6
PdCl2
O
2
28
O
R1
Ph
29
O
OH
Scheme 6. PASP synthesis of 4-ether pyridine-piperazine carbamate analogs 27. Reagents and conditions: (i) 2 equiv 24, Cs2CO3, KI, DMF, rt to 100 °C; (ii) 10 equiv 6; (iii)
0.1 equiv 28, 20 equiv DEA, DCM; (iv) 10 equiv 29, microwave, 80 °C, dioxane; (v) 1.5 equiv 21, 5 equiv TEA, DCM; (vi) 10% TFA/DCM.
amine by-product. The resulting mixture was dissolved in dioxane
and incubated with polymer-bound maleimide 29 under micro-
wave conditions to effect complete sequestration of the catalyst
28 and any dissociated phosphine and phosphine oxide ligand.
Simple filtration and evaporation of the solvent left purified prod-
ucts 26. Derivatization of the piperazine nitrogen intermediate 26
was accomplished with excess chloroformate 21 followed by
sequestration of the remaining chloroformate 21 with polymer-
bound amine 6 to afford the t-butyl ester products. Deprotection
of the t-butyl ester using TFA in DCM provided the desired bi-
substituted products 27.
A focused library was designed in which ethyl, isopropyl, butyl,
and pentyl chloroformates 21 were selected with a variety of alkyl
halides 24 (Table 5). Again, an additive effect was observed provid-
ing analogs with excellent PRP potency with the ethyl, butyl and
pentyl carbamates preferred in most cases. This focused library
provided 4-piperidine analog 27T having exceptional potency with
a PRP IC50 of 550 nM.
weight. The core templates 10E and 10F have molecular weights
over 500. Substituting at the pyridine 4-position only adds to the
molecular weight, potentially resulting in the loss of good bioavail-
ability. Despite high molecular weight and, in numerous cases,
poor Caco-2 permeability values, many compounds had good bio-
availability [and probably have transporter mediated absorption,
as exemplified by 27I (molecular weight 570, Papp 1.1 Â 10À6).]
Unfortunately, most of the initial compounds profiled had high
clearance values despite good in vitro metabolic stability. The 4-
piperidine analogs typically provided the more potent compounds
and also allowed for optimizing the AMDE properties. It was found
that compounds with the 4-piperidine substituted with a basic
amine (22N) usually had high clearance values. Replacing the
Table 5
Binding and PRP activity data for focused 4-ether pyridine library9
b
Compound
R1
OR
Kia (nM)
IC50
(l
M)
Several hundred compounds were prepared with nitrogen or
oxygen substituents at the 4-position of the pyridine ring and vary-
ing carbamate chain lengths on the piperazine terminal nitrogen.
The SAR clearly showed that butyl, pentyl, and hexyl carbamates
were optimal and enhancement of potency and other properties
could be obtained by variations at the 4-position of the pyridine,
providing compounds with submicromolar PRP levels of activity.
The potency optimization described above was conducted with
full consideration of maximizing reasonable drug-like properties.
In general this class of compounds has good solubility and is chem-
ically stable. These inhibitors also show excellent metabolic stabil-
ity in both the rat and human microsomal assays. Many
compounds were evaluated for their in vivo pharmacokinetic char-
acteristics in the rat and a representative set of pharmacokinetic
profiles is shown in Table 6. Compounds 7D, 10E and 10F, each
possessing a hydrogen at the 4-position with an ethyl, pentyl and
hexyl carbamate, respectively, were selected to profile the general
template and test if the carboxylic acid would be detrimental to
oral absorption. Gratifyingly, all of the compounds were bioavail-
able with high bioavailabilities for the pentyl and hexyl carbamate
analogs. However, all three of the compounds had high clearance
values. Another concern with oral absorption was the molecular
27A
27B
27C
27D
27E
Et
157
203
145
21
7.6
11.1
25.6
7.8
Allyl
i-Pr
Bu
O(CH2)2OH
Pent
32
11.5
27F
27G
27H
27I
Et
92
95
189
18
5.8
13.3
20.1
4.4
Allyl
i-Pr
Bu
O(CH2)2OMe
27J
Pent
10
4.3
27K
27L
27M
27N
27O
Et
39
99
146
36
33.9
61.7
—
26.9
28.4
Allyl
i-Pr
Bu
OBu
Pent
49
27P
27Q
27R
27S
27T
Et
68
93
157
12
1.8
9.4
—
1.1
0.55
Allyl
i-Pr
Bu
O
NH
Pent
3.8
a
Membranes from CHO cells expressing recombinant human P2Y12 receptors
incubated with 33P ADP and compound. Ki values are corrected from IC50 using
Cheng and Prusoff equation and are the geometric mean of n = 2 or greater.
b
IC50 values are from human PRP incubated with 20 lM ADP.

J. J. Parlow et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4657–4663
4663
Table 6
nificant toxicological findings. The potency, selectivity, safety, and
overall pharmacokinetic profile for 22J support advancement for
clinical evaluation. Reports from our backup program where efforts
at refining the SAR for more potent analogs with improved phar-
macokinetic properties will be forthcoming. In addition, alternate
aryl ring systems in place of the pyridine ring have been synthe-
sized and will be reported in due course.
a,b
Rat pharmacokinetic profiles of selected P2Y12 antagonists
Compound
CL (mL/min/kg)
Vdss (L/kg)
T1/2
,
(h)
Foral (%)
eff
7D
98
89
29
49
133
69
70
3
15.5
10.3
1.7
1.5
3.1
1.3
2.6
0.3
1.8
1.3
0.7
0.4
0.3
0.2
0.4
0.7
34
70
>95c
77
1
—
0
89
10E
10F
27I
27T
22B
22N
22J
Supplementary data
a
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.06.075.
Male Sprague–Dawley rats (n = 2–4 rats).
Dose: iv infusion at 2 mg/kg; po at 5 mg/kg (n = 2–4 rats).
Foral (n = 3) 92%, 95%, and 181%.
b
c
References and notes
1. Horiuchi, H. Ann. Med. 2006, 38, 162.
2. Gachet, C. Thromb. Haemost. 2001, 86, 222.
amine group with a methoxy resulted in compounds with accept-
able clearance values, although a slight loss in potency was ob-
served. These studies led to the discovery of compound 22J with
a low clearance of 3 mL/min/kg (1 mL/min/kg, dog) and good bio-
availability (>90%, dog). Clearance of 22J in rats appears to be med-
iated through extensive biliary excretion, with roughly 70% of an iv
dose (2 mg/kg) recovered in the bile within 4 h.
Further evaluation of compound 22J showed good solubility
and chemical stability with the crystalline form of the compound.
Compound 22J was found to have acceptable margins versus hERG
activity and a satisfactory CYP inhibition profile. Further evaluation
on the potency of compound 22J was determined by measuring the
3. (A) For reviews see: (a) Angiolillo, D. J. Am. J. Cardiovasc. Drugs 2007, 7, 423; (b)
Cattaneo, M. Ex. Rev. Cardiovasc. Ther. 2007, 5, 45; (c) Storey, R. F. Curr. Pharm.
Des. 2006, 12, 1255; (d) Boeynaems, J. M.; van Giezen, H.; Savi, P.; Herbert, J. M.
Curr. Opin. Invest. Drugs 2005, 6, 275; (B) For papers see: (a) Springthorpe, B.;
Bailey, A.; Barton, P.; Birkinshaw, T. N.; Bonnert, R. V.; Brown, R. C.; Chapman,
D.; Dixon, J.; Guile, S. D.; Humphries, R. G.; Hunt, S. F.; Ince, F.; Ingall, A. H.; Kirk,
I. P.; Leeson, P. D.; Leff, P.; Lewis, R. J.; Martin, B. P.; McGinnity, D. F.;
Mortimore, M. P.; Paine, S. W.; Pairaudeau, G.; Patel, A.; Rigby, A. J.; Riley, R. J.;
Teobald, B. J.; Tomlinson, W.; Webborn, P. J. H.; Willis, P. A. Bioorg. Med. Chem.
Lett. 2007, 17, 6013; (b) Husted, S.; Emanuelsson, H.; Heptinstall, S.; Sandset, P.
M.; Wickens, M.; Peters, G. Eur. Heart J. 2006, 27, 1038; (c) Van Giezen, J. J. J.;
Humphries, R. G. Sem. Thromb. Hemost. 2005, 31, 195; (d) De Marco, A.; de
Candia, M.; Carotti, A.; Cellamare, S.; De Candia, E.; Altomare, C. Eur. J. Pharm.
Sci. 2004, 22, 153; (e) Yang, J.; Hua, W.; Wang, F.; Wang, Z.; Wang, X. Bioorg.
Med. Chem. 2004, 12, 6547; (f) De Candia, M.; Summo, L.; Carrieri, A.; Altomare,
C.; Nardecchia, A.; Cellamare, S.; Carotti, A. Bioorg. Med. Chem. 2003, 11, 1439;
(g) Yang, S. W.; Buivich, A.; Chan, T. M.; Smith, M.; Lachowicz, J.; Pomponi, S. A.;
Wright, A. E.; Mierzwa, R.; Patel, M.; Gullo, V.; Chu, M. Bioorg. Med. Chem. Lett.
2003, 13, 1791; (h) Xu, B.; Stephens, A.; Kirschenheuter, G.; Greslin, A. F.;
Cheng, X.; Sennelo, J.; Cattaneo, M.; Zighetti, M. L.; Chen, A.; Kim, S. A.; Kim, H.
S.; Bischofberger, N.; Cook, G.; Jacobson, K. A. J. Med. Chem. 2002, 45, 5694.
4. (a) Caroff, E.; Hilpert, K.; Meyer, E. WO 050301, 2008.; (b) Caroff, E.; Hilpert, K.;
Meyer, E. WO 044217, 2008.; (c) Caroff, E.; Fretz, H.; Hilpert, K.; Houille, O.;
Hubler, F.; Meyer, E. WO 114774, 2006.
IC50 in 13 human donors using 20
gation with the 4-well Chronolog PRP aggregometry assay and
found the average IC50 to be 1.8
M.11,12 Compound 22J was orally
lM ADP-stimulated PRP aggre-
l
efficacious in the rat ferric chloride thrombosis model in which it
dose-dependently prevented thrombus formation in this model
of injury induced thrombosis.13
In vitro receptor binding, signaling, and functional studies have
shown that 22J is a high affinity, selective, and competitive antag-
onist at P2Y12 receptors. Compound 22J is more than 340-fold
selective for P2Y12 over the other purinergic receptors tested
including the closest homolog, P2Y13 (48% homology to P2Y12),
and a second platelet purinergic GPCR, P2Y1 (19% homology to
P2Y12). The Kb (8.5 nM) for functional antagonism in the P2Y12 sig-
naling assay is in good agreement with the Ki (15 nM) for the
receptor. The Kb in human PRP platelet aggregation assays
(300 nM) is higher, due to the presence of plasma proteins. Schild
analysis indicates the interaction of 22J with the receptor in the
aggregation assays is competitive with ADP.
In summary, we have utilized PASP parallel library synthesis to
identify a P2Y12 lead. This lead was optimized with generation of
structure–activity relationships leading to highly potent P2Y12
antagonists. With sufficient levels of potency attained, pharmaco-
kinetic and physiochemical properties were modulated through
various substituents at the 4-pyridine position. Fine-tuning the
PK properties led to 22J, an orally bioavailable, direct acting,
reversible P2Y12 antagonist. This compound is a selective antago-
nist of the human P2Y12 receptor and demonstrates oral antiplate-
let and anti-thrombotic efficacy in preclinical species. No safety
issues were observed with 22J as it was inactive in both the Ames
and Micronucleus assays, and has been evaluated in both rat and
dog safety studies (single dose and 7-day repeat dose) without sig-
5. Hollopeter, G.; Jantzen, H. M.; Vincent, D.; Li, G.; England, L.; Ramakrishnan, V.;
Yang, R. B.; Nurden, P.; Nurden, A.; Julius, D.; Conley, P. B. Nature 2001, 409,
202.
6. (a) Meadows, T. A.; Bhatt, D. L. Circ. Res. 2007, 100, 1261; (b) Savi, P.; Pereillo, J.
M.; Uzabiaga, M. F.; Combalbert, J.; Picard, C.; Maffrand, J. P.; Pascal, M.;
Herbert, J. M. Thromb. Haemost. 2000, 84, 891.
7. Gurbel, P. A.; Bliden, K. P.; Hiatt, B. L.; O’Connor, C. M. Circulation 2003, 107,
2908.
8. (a) Bryant, J.; Post, J. M.; Alexander, S.; Wang, Y. X.; Kent, L.; Schirm, S.; Tseng, J.
L.; Subramanyam, B.; Buckman, B.; Islam, I.; Yuan, S.; Sullivan, M. E.; Snider, M.;
Morser, J. Thromb. Res. 2008, 122, 523; (b) Post, J. M.; Alexander, S.; Wang, Y. X.;
Vincelette, J.; Vergona, R.; Kent, L.; Bryant, J.; Sullivan, M. E.; Dole, W. P.;
Morser, J.; Subramanyam, B. Thromb. Res. 2008, 122, 533; (c) Wang, Y. X.;
Vincelette, J.; da Cunha, V.; Martin-McNulty, B.; Mallari, C.; Fitch, R. M.;
Alexander, S.; Islam, I.; Buckman, B. O.; Yuan, S.; Post, J. M.; Subramanyam, B.;
Vergona, R.; Sullivan, M. E.; Dole, W. P.; Morser, J.; Bryant, J. A. J. Thromb.
Haemost. 2007, 97, 847; (d) Bryant, J. A.; Buckman, B. O.; Islam, I.; Mohan, R.;
Morrissey, M. M.; Wei, G. P.; Xu, W.; Yuang, S. WO 052366, 2004.; (e) Bryant, J.
A.; Buckman, B. O.; Islam, I.; Mohan, R.; Morrissey, M. M.; Wei, G. P.; Xu, W.;
Yuang, S. WO 098856, 2002
9. These compounds showed significantly reduced binding affinity when 0.4%
human serum albumin was added to the binding assay. In most cases, the
difference in Ki value from the binding assay to the IC50 value of the functional
assay was attributed to protein binding.
10. Lan, P.; Berta, D.; Porco, J. A.; South, M. S.; Parlow, J. J. J. Org. Chem. 2003, 68,
9678.
11. Born, B. V. R. Nature 1962, 194, 927.
12. The R-enantiomer of 22J has a Ki of 510 nM and is inactive in the PRP assay,
IC50 > 50
lM.
13. Kurz, K. D.; Main, B. W.; Sadusky, G. E. Thromb. Res. 1990, 60, 269.