Discovery of P2X3 selective antagonists for the treatment of chronic pain

Article abstract of DOI:10.1016/j.bmcl.2012.01.124

Purinergic receptor P2X3 has been linked to analgesia in a number of pre-clinical models of pain, and is expressed in the human pain perception pathway. Only few P2X3-selective antagonists have been reported to date. This Letter describes the SAR and in vivo analgesic profile of a novel scaffold of selective P2X3 antagonists.

Full text of DOI:10.1016/j.bmcl.2012.01.124

Bioorganic & Medicinal Chemistry Letters 22 (2012) 2565–2571
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of P2X3 selective antagonists for the treatment of chronic pain
Louis-David Cantin a, , Malken Bayrakdarian a, Christophe Buon a, Eric Grazzini c, Yun-Jin Hu a,
⇑
Jean Labrecque c, Carmen Leung a, Xuehong Luo a, Giovanni Martino c, Michel Paré c,
Kemal Payza c, Nirvana Popovic a, Denis Projean b, V. Santhakumar a,
Christopher Walpole a, Xiao Hong Yu c, Mirosław J. Tomaszewski a
a Department of Medicinal Chemistry, AstraZeneca R&D, 7171 Frederick-Banting, Montréal, QC, Canada H4S 1Z9
b Department of Drug Metabolism and Pharmacokinetics, AstraZeneca R&D, 7171 Frederick-Banting, Montréal, QC, Canada H4S 1Z9
c Department of Biosciences, AstraZeneca R&D, 7171 Frederick-Banting, Montréal, QC, Canada H4S 1Z9
a r t i c l e i n f o
a b s t r a c t
Article history:
Purinergic receptor P2X3 has been linked to analgesia in a number of pre-clinical models of pain, and is
expressed in the human pain perception pathway. Only few P2X3-selective antagonists have been
reported to date. This Letter describes the SAR and in vivo analgesic profile of a novel scaffold of selective
P2X3 antagonists.
Received 16 December 2011
Revised 27 January 2012
Accepted 30 January 2012
Available online 9 February 2012
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
P2X3
Purinergic receptors
Analgesia
Pyrrolo-pyrimidinones
Nociceptive pain
Chronic pain can erode the quality of life of patients suffering
from conditions such as arthritis. Additionally peripheral nerve
damage caused by complications of diseases such as diabetes, viral
infections or cancer can also evolve into chronic neuropathic pain.
While current treatments can offer some pain relief, patients often
become non-responsive to these therapies, or need to bear with
incapacitating side-effects.1 For example, patients suffering from
osteoarthritis may only get partial pain relief from non-steroidal
anti-inflammatory drugs, while experiencing side-effects that
range from gastro-intestinal discomfort to ulcers, renal failure
and increased risks of myocardial infarction.
The P2X3 receptor is a homomeric trimer assembled from three
P2X3 monomers. Two units of P2X3 monomer can also form heter-
omeric trimers when combined with one P2X2 monomer, leading
to the P2X2/3 receptor.5 The latter has some overlapping expres-
sion with P2X3 receptor, complicating the assessment of the role
of P2X3 in vivo. Despite having access to P2X3 and P2X2 knock-
out animals, understanding the relative contributions of P2X3
and P2X2/3 to pain perception has been difficult since knock-out
animals lack both the homomer and heteromer.6 Initial reports
from an Abbott research group on the use of A-317491 (1, Fig. 1)
allowed for the characterization of unselective antagonists as effi-
cacious analgesic in pre-clinical models of pain.7
Seminal work of Burnstock2 and North3 has elucidated the role
of purinergic receptors in disease conditions and highlighted their
potential value in the development of novel therapeutics.
Purinergic receptors are classified into two sub-families: P2Y
receptors that are GPCRs, and P2X receptors that are ligand-gated
ion channels for which the endogenous ligand is ATP. The con-
served expression of P2X3 in the pain pathways (including dorsal
root and trigeminal ganglia)4 made this an attractive target for
our research team.
Subsequent reports from Roche,8 confirmed the role of P2X3–
P2X2/3 and disclosed the first orally active compounds (2, 3).9 Only
recently have reports of antagonists selective for P2X3 been pub-
lished (4), but the in vivo efficacy data has yet to be shared.10
Therefore, it remains unclear whether the analgesic effects seen
with P2X3 antagonists require blockade of P2X3 homotrimers or
P2X2/3 heterotrimers, or both. This Letter will discuss SAR explora-
tion and in vivo characterization of a novel scaffold for P2X3-selec-
tive antagonists.
With the goal of identifying orally available P2X3 antagonists,
we explored a variety of hit-finding options. Lacking meaningful
structural information about the target, we decided to undertake
an HTS screen,11 in which 800,000 compounds were assayed. This
⇑
Corresponding author.
E-mail addresses: Louis-david.cantin@astrazeneca.com, ldcantin@eckerdalumni
.com (L.-D. Cantin).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.01.124

2566
L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2565–2571
CO2H
CO2H
NH2
N
NH2
N
HO2C
O
O
OH
OH
O
O
N
O
N
NH2
O
N
N
H
I
I
O
1
2
3
N
O
S
N
N
H
N
N
HN
4
6
N
2
N
N
N
O
O
4
5
Figure 1. Antagonists of P2X3 and P2X2/3.
approach led to the identification of 445 hits, from which a small
cluster of pyrrolo-pyrimidinones was identified. For example, com-
pholine group could be replaced by a piperazine (11a) without
affecting potency, but dramatically improving solubility. A smaller
pyrrolidine 11b was also well tolerated, as was the opening of the
morpholine ring (11c). However, in both cases solubility was not
improved relative to that of 11a. Wanting to retain rigidity in this
section of the molecule, we decided to explore the piperazine
group further. Methylation of the piperazine (11d) caused a
fourfold decrease in potency with a slight improvement in solubil-
ity. However, for both 11a and 11d, the basic nature of the com-
pounds raised concerns with their eventual safety profile (e.g.,
risks to encounter PLD, hERG inhibition). Introduction of an acetyl
group (11e) increased potency twofold while reducing the pKa. A
sulfonamide group (11f) was also well tolerated, as was the
N,N-dimethyl-proprionamide group (11g). However, using an
N,N-dimethylamino acetyl group (11h) caused a 10-fold decrease
in potency indicating that a charged group was not tolerated.
This initial exploration of the SAR provided us with 11e, a com-
pound with sufficient potency to reliably assess selectivity over the
heteromer. Compound 11e proved very selective against rat P2X2/
pound 5 ( Fig. 1) displayed promising potency at hP2X3 (IC50
1.3 M), however the high lipophilicity (logD7.4 4.3) and low aque-
ous solubility (pH 7.4, 1 M) presented areas for improvements.
,
l
l
Synthesis12 of the desired analogs began with condensation of
orotic acid (6, Scheme 1) with paraformaldehyde to give the corre-
sponding lactone 7, which was in turn converted to the dichloropy-
rimidine intermediate 8. Introduction of phenylbenzylamine
occurred chemoselectively at the C4 position under mild condi-
tions to provide 9, an advanced intermediate ready for further
derivatization. The C2 substituent was introduced by reacting
chloro 9 with the appropriate amine in n-butanol at 140 °C. Subse-
quent lactam formation under more forcing conditions (195 °C)
afforded compounds 11a–h, albeit in low yield.
Our initial exploration of the lactam substituent (R3) showed
that this section of the molecule was intolerant to modifications;
the original isopropyl group provided the best potency (data not
shown). The C2 position provided more opportunities to introduce
structurally diverse substituents (Table 1). For example, the mor-
3, with an IC50 >2.5
lM. This level of selectivity should enable us to
O
OH
Cl
NH
b
c
a
NH
N
N
N
O
O
O
HO
N
H
O
N
OH
N
Cl
N
Cl
O
O
O
O
6
7
8
9
HN
HN
N
d
N
e
N
R3
N
O
R2
R2
N
N
N
R1
R1
O
O
10
11a-h
Scheme 1. Synthesis of pyrrolo-pyrimidinones through late-stage lactam formation. Reagents and conditions: (a) (CH2O)2, HCl, 85–95 °C, 77% yield; (b) POCl3, N,N-
diethylaniline, 110 °C, 98% yield; (c) phenylbenzylamine, DIPEA, CH2Cl2, rt, 47% yield. d) R1R2NH, n-butanol, 140 °C, 53–61% yield; (e) R3NH, HCl, methoxymethanol, 195 °C, 8–
23% yield.

L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2565–2571
2567
Cl
OH
OH
a,b
c
N
N
N
N
N
HO
N
Cl
N
OH
N
OH
O
O
O
d
13
6
12
R5
N
R4
R5
Cl
R4
N
N
N
N
e
N
N
N
N
O
N
O
O
15a-y
14
Scheme 2. Synthesis of pyrrolo-pyrimidinones through Mannich-type condensation. Reagents and conditions: (a) i-PrNH2, (CHO)n, ethanol, reflux; (b) methoxyethanol,
reflux, 70% yield (over two steps); (c) POCl3, N,N-diethylaniline, 110 °C, 65% yield; (d) R1R2NH, DIPEA, CH2Cl2, rt; (e) acetylpiperazine, n-butanol, 160 °C, microwave reactor,
70–90% yield over two steps.
Table 1
design of future analogs during our exploration of the SAR at the
C4-position.
Exploration of the C2 position of the pyrrolo-pyrimidinone scaffold (R3 = i-Pr)
Ex
NR1R2
hP2X3a (IC50
)
Solubilityb (pH7.4,
lM)
Having identified suitable substituents at the 2- and 6-positions
of the pyrrolo-pyrimidinone core, we decided devise of a more
efficient synthetic route. Introduction of the isopropyl group at
the 6-position was achieved through a Mannich-type condensation
between orotic acid 6, formaldehyde and isopropylamine; subse-
quent lactam formation to produced 12. Careful control of the
reaction conditions allowed us to selectively chlorinate the
C2- and C4-positions using POCl3, without affecting the lactam.
With 13 in hand, we were poised to explore the C4 position. Reac-
tion of selected amines with 14 under mild conditions, followed by
introduction of the acetyl piperazine moiety provided compounds
15a–w in good yield for a wide variety of compounds.
N
5
1360
<1
O
N
N
11a
11b
618
327
1000
<1
NH
H
N
O
OMe
N
H
We began our exploration of the C4 group (NR4R5) by removing
one of the aromatic rings (15a, Table 2), which resulted in a more
soluble yet inactive compound.
11c
11d
255
<1
7
N
1947
N
Lengthening the substituent by one methylene unit (15b) pro-
vided marginal antagonism, which was only improved by the addi-
tion of lipophilic substituents to the aromatic ring. Indeed, either a
4-Cl (15c) or a 4-Me (15d) group provided compounds with micro-
molar potency and measurable aqueous solubility. Restoring lipo-
philicity in the linker area increased potency, as exemplified by
the methyl-substituted compound (15e). However, adding a sec-
ond methyl group (15f) erased this gain in potency. Surprisingly,
using a cyclobutyl group (15g) restored potency while moving
the gem-dimethyl group close to the C4-amino group (15h) pro-
vided the most potent compound yet.
N
11e
11f
N
N
N
268
398
<1
<1
O
O
N
N
O
S
Despite having poor aqueous solubility and poor stability in rat
liver microsomes (rCLint 83 lL/min/mg), we seized the opportu-
11g
11h
440
3.5
nity to profile 15h in our pre-clinical models of pain to ascertain
the contribution of P2X3 antagonism to analgesia. While 15h
shows a slight decrease in potency at rat P2X3 (IC50 280 nM) com-
pared to human P2X3, this compound shows a greater than 10-fold
NMe2
O
N
N
O
5689
50
selectivity against rat P2X2/3 (IC50 >2.5 l
M, FLIPR)13 and was de-
NMe2
FLIPRÒ, mean IC50 hP2X3, RLE cells loaded with FLUO-2 and stimulated with
100 lM (a,b)me-ATP; IC50 values are the mean of at least three experiments per-
formed in triplicates, with a SD of 20%.
void of activity against a panel of more than 200 targets. Following
subcutaneous administration of 15h, we observed an analgesic
effect (Fig. 2a) that was dose responsive in a model of chronic noci-
ceptive pain (FCA-treated rats).14
Analysis of the plasma concentration-response relationship re-
vealed an EC50 of 199 nM (free plasma concentration, Emax = 90%)
in-line with the in vitro IC50 of 15h. Further, the plasma levels re-
mained well below the in vitro IC50 of 15h against P2X2/3, suggest-
ing that the observed efficacy was due solely to antagonism of
P2X3. Given the low free brain exposure levels of 15h in rats (Br/
Pl = 0.02, unbound fraction in brain homogenate ꢀ3%), we postu-
lated that there would not be significant central contribution to
the analgesic effects we observed. To investigate this hypothesis
a
b
Thermodynamic solubility measured in a pH 7.4 buffer using HPLC with UV
detection for quantification.
obtain an unambiguous assessment of analgesic potential of
P2X3-selective compounds. However, compound 11e still suffered
from poor solubility, despite having lower lipophilicity relative to
5. Improving aqueous solubility therefore became a focus for the

2568
L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2565–2571
Table 2
Exploration of the C4 position of the pyrrolo-pyrimidinone scaffold
Ex
NR4R5
hP2X3a (IC50
)
Solubilityb
(lM)
11e
268
<1
HN
HN
15a
15b
>30,000
13,950
230
140
HN
HN
HN
HN
Cl
Me
Cl
15c
15d
1665
1311
28
28
Figure 2b. Analgesic effects of 15h in the 96h FCA rat model upon ip and it dosing.
discovery work to identify P2X3-selective compounds that would
not require necessarily high CNS exposure.
15e
15f
441
2057
262
61
32
12
46
24
Having clarified the site of action and demonstrated that a
P2X3-selective compound could produce robust analgesic effects,
we continued our exploration of the lead series. Profiling 15h fur-
ther revealed that improvements in physicochemical and DMPK
properties (e.g., in vitro metabolic stability, in vivo clearance) were
Me
Cl
Cl
HN
HN
still needed, and that CYP2C9 inhibition (IC50
improvement.
2 lM) required
We therefore decided to re-investigate the C4 benzyl substitu-
ent by introducing structurally diverse substituents to modulate
electronic properties. Adding a lipophilic group such as Cl (15i, Ta-
ble 3) increased potency, which was further enhanced by using
more electron withdrawing groups such as trifluoromethyl (15j)
and trifluoromethoxy (15k). Indeed, compound 15k combined
promising potency with good stability in human liver microsomes
(hCLint) and acceptable solubility. The addition of an electron
donating group at the ortho-position (15l) produced a substantial
improvement in potency; however solubility and microsomal sta-
bility decreased. We decided to apply the lessons learned from the
C4-homobenzyl amine SAR and introduced a small substituent in
15g
15h
Cl
HN
FLIPRÒ, mean IC50 hP2X3, RLE cells loaded with FLUO-2 and stimulated with
100 lM (a,b)me-ATP; IC50 values are the mean of at least three experiments per-
formed in triplicates, with a SD of 20%.
Thermodynamic solubility measured in a pH 7.4 buffer using HPLC with UV
detection for quantification.
a
b
the methylene
group15 (15m) provided an eightfold increase in potency. Combi-
nation of the -methyl with other 4-subtituents revealed interest-
a to the nitrogen. We observed that an S-methyl
a
ing SAR. Indeed, a methyl substituent (15m) was substantially less
potent than the corresponding trifluoromethoxy group (15n) as
was the methoxy group (15o); however an ethoxy group proved
most potent (15p). Having lowered lipophilicity (logD7.4 3.0) rela-
tive to 15e, 15p showed substantially improved solubility and sta-
bility in human microsomes. The CYP inhibition profile was also
improved over previous analogs; however compound 15p only
showed moderate permeability in the Caco-2 assay. Further assess-
ment revealed that this compound suffered from significant efflux
in Caco-2 cells and was a substrate of a drug transporter, P-glyco-
protein. Inhibition studies with a P-gp inhibitor (quinidine) in
Caco-2 cells further confirmed that this mechanism was responsi-
ble in limiting its in vitro permeability.
Figure 2a. Analgesic effects of 15h in the 96h FCA rat model upon sc dosing.
We thus decided to continue exploring the SAR with a focus on
increasing permeability without affecting the overall profile of 15p
through small structural modifications. Using an
a-ethyl (15q) or
further, we administered 15h via an intrathecal route to FCA-trea-
ted rats, which did not produce any analgesic effects. Differently,
intraplantar administration of similar doses of 15h produced sig-
nificant analgesic effects (Fig. 2b). Taken together, these results al-
lowed us to more clearly chart a path forward in our drug
an -CF3 (15r) group resulted in substantial decreases in potency
a
and no gains in permeability. Addition of substituents on the aryl
ring, as exemplified by 15s, was tolerated, but led to an erosion
of microsomal stability. Replacing the phenyl ring by a pyridine
(15t) led to a 15-fold drop in potency, whereas extension of the

L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2565–2571
2569
Table 3
Exploration of the C4 position of the pyrrolo-pyrimidinone scaffold
Ex
NR4R5
hP2X3a (IC50, nM)
Sol.b
30
(
lM)
hCLintc
6
(l
L/min/mg)
CYP2C9d (IC50
,
lM)
Caco-2e (1E-6 cm/s)
HN
15i
4214
2.4
Cl
HN
HN
15j
1271
324
17
84
31
4
CF3
OCF3
15k
2.4
2.9
OMe
HN
15l
67
613
64
10
260
61
31
OCF3
Me
Me
HN
HN
15m
15n
15o
15p
15q
15r
15s
Me
22
4
7.3
2.4
>10
4.5
OCF
3
Me
Me
Et
HN
HN
HN
HN
HN
2523
69
84
OMe
OEt
OEt
120
86
10
10
5
2.4
212
115
38
CF3
Me
Me
72
14
11
2
OEt
F
220
17
0.8
OEt
F
HN
HN
HN
15t
521
24
58
3
7
OEt
N
N
15u
11
16
0.1
15v
24
22
590
16
19
7
7
0.6
2.9
N
N
15w
1100
N
a
FLIPRÒ, mean IC50 hP2X3, RLE cells loaded with FLUO-2 and stimulated with 100
lM (a,b)me-ATP; IC50 values are the mean of at least three experiments performed in
triplicates, with a SD of 20%.
b
Thermodynamic solubility measured in a pH 7.4 buffer using HPLC with UV detection for quantification.
Metabolic stability performed in human liver microsomes in presence of NADPH using test compound at 1
P450 inhibition screen using recombinant CYP2C9 and a fluorescent probe; IC50 values are the mean of at least three experiments performed in duplicates, with a SD of
c
lM.
d
20%.
e
Transwell assay. Apparent permeability (apical to basolateral) across a caco-2 cell monolayer measured under pH gradient pH 6.5–7.4 using test compound at 10 lM
applied on the apical side.
ring system to a quinoline (15u) led to one of the most potent com-
hCLint was in a favorable range, however solubility and permeabil-
pounds in this series. CYP2C9 inhibition was also improved and
ity were poor. Using an isomeric isoquinoline (15v) increased sol-

2570
L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2565–2571
Table 4
Pharmacokinetic profile of compounds 15h and 15p
Ex
Dose
m/kg)
Cmax
M)
AUC
M h)
T1/2
(h)
CLb
(ml/min/kg)
F(%)
(oral)
(
l
(
l
(
l
15h (iv, rat)
15h (sc, rat)
15p (iv, rat)
15p (iv, rat)
15p (po, rat)
15p (po, rat)
15p (iv, dog)
2.2
5.9
2.2
11
12
58
—
1.0
—
0.7
2.8
0.4
8.05
0.4
1.8
4.8
0.9
54
0.5
2
88
22
—
0.1
6.9
—
16%a
62%a
2.2
14.5
8
a
Calculated using the 2.2
Rat liver blood flow = 72 mL/min/kg; dog liver blood flow = 55 mL/min/kg.
l
mol/kg iv dose.
b
ubility substantially, however permeability remained unimproved.
We speculated that removing the hydrogen bond donor NH would
lead to an increase in permeability, which proved to be the case for
the N-Me substituent (15w). However, decreased microsomal sta-
bility prevented further progression of these compounds.
while improving permeability and reducing CYP2C9 inhibition. This
series of compounds provided us with useful tools, such as (15h),
that demonstrated that analgesic effects, in the rat FCA model, of
small molecule antagonists selective for P2X3 and added further
support for a peripheral site of action in this model.
Despite its moderate permeability, 15q showed a balanced
DMPK profile and we had confidence in achieving good oral expo-
sure (rCLint 58 ll/min/mg; 60% rat liver blood flow). We decided to
Acknowledgments
progress it into in vivo DMPK and into in vivo pre-clinical models
chronic pain.
When administered to rats, (Table 4), the compound showed
high systemic clearance (higher than the rat liver blood flow), short
half-life and low oral bioavailability. The high clearance could not
be explained completely by the measured in vitro metabolism
The authors would like to recognize the contributions of Joanne
Butterworth, Angelo Filosa, Claude Godbout, Vincent Kennedy,
Giuseppe Molinaro Julie Pelland, Stéphane St-Onge, Pascal Turcotte
and Huy Khang Vu to this work.
References and notes
either in rat microsomes (rCLint 17 lL/min/mg; 60% rat liver blood
flow) or in rat hepatocytes. Because 15p bears an amide function
we speculated that this compound may be hydrolyzed in plasma,
which proved not to be the case. Additionally, no preferential dis-
tribution was found in blood cells and therefore could not explain
the observed discrepancy. Further exploration of the clearance
mechanism revealed that only very small amounts of parent com-
pound were recovered unchanged in rat urine, however a signifi-
cant amount could be found in feces (>15% of the dose).
Administration of higher doses intravenously led both to a de-
crease in plasma clearance and excretion in feces (<3% of the dose).
Similarly, a higher oral dose led to an increase in bioavailability.
Taken together, these results suggest that metabolism and addi-
tional saturable mechanisms (e.g., efflux in bile, secretion directly
from blood to the gut mucosa) are involved in the disposition of
this molecule in rat.
1. (a) Turk, D. C.; Wilson, H. D.; Cahana, A. Lancet 2011, 377, 2226; (b) Portenoy, R.
K. Lancet 2011, 377, 2236.
2. Chen, C. C.; Akopian, A. N.; Sylviotti, L.; Colquhoun, D.; Burnstock, G.; Wood, J.
N. Nature 1995, 377, 428.
3. Lewis, C.; Neidhart, S.; Holy, C.; North, R. A.; Buell, G.; Surprenant, A. Nature
1995, 377, 432.
4. Vulchanova, L.; Reidl, M. S.; Shuster, S. J.; Buell, G.; Suprenant, A.; North, R. A.;
Elde, R. Neuropharmacology 1997, 36, 1229; (b) Xu, G. Y.; Huang, L. Y. J. Neurosci.
2002, 22, 93; (c) Chen, Y.; Li, G. W.; Wang, C.; Gu, Y.; Huang, L. Y. Pain 2005, 119,
38.
5. (a) North, R. A. Curr. Opin. Cell Biol. 1996, 8, 474; Virginio, C.; Robertson, G.;
Surprenant, A.; North, R. A. Mol. Pharmacol. 1998, 53, 969.
6. (a) Honore, P.; Kage, K.; Mikusa, J.; Watt, A. T.; Johnston, J. F.; Wyatt, J. R.;
Faltynek, C. R.; Jarvis, M. F.; Lynch, K. Pain 2002, 99, 11; Cockayne, D. A.;
Hamilton, S. G.; Zhu, Q. M.; Dunn, P. M.; Zhong, Y.; Novakovic, S.; Malmberg, A.
B.; Cain, G.; Berson, A.; Kassotakis, L.; Hedley, L.; Lachnit, W. G.; Burnstock, G.;
McMahon, S. B.; Ford, A. P. Nature 2000, 407, 1011; (c) Cockayne, D. A.; Dunn, P.
M.; Zhong, Y.; Rong, W.; Hamilton, S. G.; Knight, G. E.; Ruan, H. Z.; Ma, B.; Yip,
P.; Nunn, P.; McMahon, S. B.; Burnstock, G.; Ford, A. P. J. Physiol. 2005, 567, 621;
(d) Souslova, V.; Cesare, P.; Ding, Y.; Akopian, A. N.; Stanfa, L.; Suzuki, R.;
Carpenter, K.; Dickenson, A.; Boyce, S.; Hill, R.; Nebenius-Oosthuizen, D.; Smith,
A. J. H.; Kidd, E. J.; Wood, J. N. Nature 2000, 407, 1015.
7. (a) Jarvis, M. F.; Burgard, E. C.; McGaraughty, S.; Honore, P.; Lynch, K.; Brennan,
T. J.; Subieta, A.; van Biesen, T.; Cartmell, J.; Bianchi, B.; Niforatos, W.; Kage, K.;
Yu, H. X.; Mikusa, J.; Wismer, C. T.; Zhu, C. Z.; Chu, K.; Lee, C. H.; Stewart, A. O.;
Polakowski, J.; Cox, B. F.; Kowaluk, E.; Williams, M.; Sullivan, J.; Faltynek, C.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 17179; (b) McGaraughty, S.; Wismer, C. T.;
Zhu, C. Z.; Mikusa, J.; Honore, P.; Chu, K. L.; Lee, C.; Faltynek, C. R.; Jarvis, M. F.
Br. J. Pharmacol. 2003, 140, 1381.
The physicochemical attributes of 15p, combined with the fact
that the compound is a substrate of drug transporters (P-gp), likely
explains some of these unusual findings.16 Interestingly, this phe-
nomenon did not carry in higher species like the dog, where com-
pound showed a lower clearance, in-line with the measured
metabolic stability in dog hepatocytes (2.5
liver blood flow).
l
L/min/106 cells, 20%
Despite possessing an overall favorable in vitro selectivity
profile (including against a panel of 100 targets) and suitable PK
properties, the observation of compound-related (reversible)
side-effects when tested in animals at higher doses precluded fur-
ther progression of 15p and its use as a tool compound. Indeed, in
contrast to 15h, 15p displayed side-effects that prevented accurate
measure of analgesic effects in behavior-based pre-clinical models
of pain.
8. Ford, A.P. Purinergic Signalling, published online 18 Nov. 2011 (DOI 10.1007/
s11302-011-9271-6).
9. (a) Carter, D. S.; Alam, M.; Cai, H. Y.; Dillon, M. P.; Ford, A. P. D. W.; Gever, J. R.;
Jahangir, A.; Lin, C.; Moore, A. G.; Wagner, P. J.; Zhai, Y. S. Bioorg. Med. Chem.
Lett. 2009, 19, 1628; (b) Gever, J. R.; Soto, R.; Henningsen, R. A.; Martin, R. S.;
Hackos, D. H.; Panicker, S.; Rubas, W.; Oglesby, I. B.; Dillon, M. P.; Milla, M. E.;
Burnstock, G.; Ford, A. P. Br. J. Pharmacol. 2010, 160, 1387; (d) Jahangir, A.;
Alam, M.; Carter, D. S.; Dillon, M. P.; Du Bois, D. J.; Ford, A. P. D. W.; Gever, J. R.;
Lin, C.; Wagner, P. J.; Zhai, Y. S.; Zira, J. Bioorg. Med. Chem. Lett. 2009, 19, 1632.
10. Brotherton-Pleiss, C. E.; Dillon, M. P.; Ford, A. P. D. W.; Gever, J. R.; Carter, D. S.;
Gleason, S. K.; Lin, C. J.; Moore, A. G.; Thompson, A. W.; Villa, M.; Zhai, Y. Bioorg.
Med. Chem. Lett. 2010, 20, 1031.
In summary, we have described our work in developing structur-
ally novel P2X3-selective antagonists based on a pyrrolo-pyrimidi-
11. HTS assay measured inhibition of the increase in intracellular calcium induced
none core. Starting from
a weakly active hit with poor
by (a,b)-Me-ATP in cells expressing human P2X3.
12. Synthetic procedures and characterization data can be found in the following
patent application: Bayrakdarian, M.; Buon, C.; Cantin, L.-D.; Hu, Y.-J.; Luo, X.;
physicochemical properties, we identified areas of the molecule
useful to increase potency and improve physicochemical properties,

L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2565–2571
2571
Santhakumar, V.; Tomaszewski, M.J. WO2008136756. Some of the
compounds can be prepared either using the route described in Scheme 1
or Scheme 2.
Collection of plasma for determination of compound levels in behavioral assay: A
separate group of naïve rats, satellite animals, not subjected to nociceptive
testing were used for plasma collection in order to establish the PK/PD
relationship. At the appropriate time point, satellite animals were injected
with drug using the same dosing solution and route of administration as the
tested animals. Blood was collected by decapitation. Whole blood was
transferred to heparinized tubes and centrifuged at 3000g for 5 min. Plasma
supernatant was then collected and frozen (À80 °C). The measurement of
compound concentration was carried out on a generic HPLC coupled to a mass
spectrometer with an electrospray ionization source in the positive mode using
multiple reaction monitoring. Chromatographic separations were performed
13. 15h is a P2X3 antagonist non-competitive with (a,b)-Me-ATP. Data in the FLIPR
assays is consistent with that generated using PatchXpress or manual patch
clamp experiments using either cells overexpressing P2X3 or native rodent
tissues.
14. Methods for characterization of analgesic effects of compounds in the 96h FCA
rat model:
Animals: The animal experiments were approved by the AstraZeneca Animal
Care Committee and were conducted in accordance with the guidelines
provided by the Canadian Council on Animal Care and the Association for the
Assessment and Accreditation of Laboratory Animal Care. Male Sprague–
Dawley rats (Charles River, St-Constant, Qc, CAN) weighing 200–225 g were
utilized for the animal experiment studies. The animals were group-housed in
polycarbonate, ventilated cages (filter top) in a controlled environment room
(12-h light/dark cycle, 20.5–23.5 °C, relative humidity: 40–70%) with food (14%
Protein Rodent Maintenance Diet, Harlan Teklad, Madisson, WI, USA) and
water ad libidum.
using a generic reverse phase HPLC C18 column (2.1 Â 50 mm, 3
l). The mobile
phase consisted of 0.1% formic acid in purified water (solvent A) and 100%
acetonitrile containing 0.1% formic acid (solvent B). The following gradient was
used: 5% to 95% solvent B in 2 min. The flow rate was constant at 0.750 mL/min
and the column temperature was set at 45 °C. Data processing was performed
using the LC/MS/MS software. Rat plasma protein binding was determined
in vitro using an equilibrium dialysis technique run at 10 lM. [Kalvass, J.C.;
Maurer, T.S. Biopharmaceutics Drug Disposition 2002, 23, 327–338.] The
degree of brain binding was assessed by equilibrium dialysis and samples were
analyzed by LC–MS/MS.
Inflammatory pain model: Inflammation was induced by injection of a single
40 ll Freund’s complete adjuvant (FCA) intra-plantar injection into the rat’s
left hindpaw. All experiments were conducted 96h after FCA administration.
Animals were sacrificed immediately after data acquisition.
Data analysis: Data are presented as a mean data SEM. Statistical significance
was determined using one-way measures ANOVA followed by a Holm-Sidack
post hoc test using SigmaStat statistical package (Systat Software Inc., ver.
3.11). A p-Value of less than 0.05 was regarded as significant. Free plasma EC50
values in behavioral studies were determined from 4-point concentration-
response curves using non-linear regression and a sigmoidal variable slope
logistic model (Prism 4.03, GraphPad Software Inc., USA).
Behavioral measurements of nociception: Twenty-four hours prior to behavioral
testing animals were brought to the laboratory for acclimatization to the new
environment. Mechanical hyperalgesia was assessed using the Ugo Basile
analgesy meter (Ugo Basile, Comerio, Italy). Animals were gently restrained,
and a steadily increasing pressure was applied to the dorsal surface of a hind
paw via a probe with a dome-shaped tip (diameter of 1 mm). The pressure
required to elicit paw withdrawal was determined. An assay cut off was set at
295 g. Two trials were conducted with 5 min intervals between each trial. Paw
withdrawal thresholds were calculated as the mean of the two values.
Animals were randomised and allocated to treatment groups to achieve a
minimum statistical power of 80%. In all cases the experimenter was blind to
the treatment received.
15. We synthesized
a number of enantiomeric pairs, and in all cases the S-
enantiomer was the only one active. The R-enantiomer was consistently
inactive.
16. (a) Huang, L.; Berry, L.; Ganga, S.; Janosky, B.; Chen, A.; Roberts, J.; Colletti, A. E.;
Lin, M.-H. J. Drug Metab. Dispostion 2010, 38, 223; (b) Harrison, A.; Betts, A.;
Fenner, K.; Beaumont, K.; Edgington, A.; Roffey, S.; Davis, J.; Comby, P.; Morgan,
P. Drug Metabo. Disposition 2004, 32, 197.