Discovery, synthesis and SAR of azinyl- and azolylbenzamides antagonists of the P2X7 receptor

Article abstract of DOI:10.1016/j.bmcl.2011.06.117

The discovery, of a series of 2-Cl-5-heteroaryl-benzamide antagonists of the P2X₇ receptor via parallel medicinal chemistry is described. Initial analogs suffered from poor metabolic stability and low Vd_ss. Multi parametric optimizat

Full text of DOI:10.1016/j.bmcl.2011.06.117

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5475–5479
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery, synthesis and SAR of azinyl- and azolylbenzamides antagonists
of the P2X7 receptor
⇑
Chakrapani Subramanyam , Allen J. Duplantier, Mark A. Dombroski, Shang-Poa Chang,
Christopher A. Gabel, Carrie Whitney-Pickett, David G. Perregaux, Jeff M. Labasi, Kwansik Yoon,
Richard M. Shepard, Michael Fisher
Neuroscience Chemistry, Pfizer Global R&D, Groton Laboratories, Eastern Point Road, Groton, CT 06340, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery, of a series of 2-Cl-5-heteroaryl-benzamide antagonists of the P2X7 receptor via parallel
Received 20 April 2011
Revised 25 June 2011
Accepted 27 June 2011
Available online 2 July 2011
medicinal chemistry is described. Initial analogs suffered from poor metabolic stability and low Vdss
.
Multi parametric optimization led to identification of pyrazole 39 as a viable lead with excellent potency
and oral bioavailability. Further attempts to improve the low Vdss of 39 via introduction of amines led to
analogs 40 and 41 which maintained the favorable pharmacology profile of 39 and improved Vdss after iv
dosing. But these analogs suffered from poor oral absorption, probably driven by poor permeability.
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
P2X7 antagonists
Parallel medicinal chemistry
Rheumatoid arthritis
Volume of distribution
Oral bioavailability
The P2X7 receptor is a member of the purinoreceptor superfam-
ily; this family includes both G-protein coupled P2Y receptors and
ionotropic P2X receptors, a group of ligand-gated ion channels.1
The P2X7 receptor has been implicated in a number of disease states
such as pain, neurodegeneration, rheumatoid and Oseteoarthritis,
making it a very attractive target for therapeutic intervention.2,3
Herein, we describe the discovery of azinyl and azolylbenzamides,
which are potent antagonists of the P2X7 receptor.
O
N
O
N
O
N
HO
HN
HN
N
N
N
N
O
O
MeO
O
Cl
H
N
H
N
H
N
OH
Cl
O
Cl
O
Cl
O
1
2
3
IL1β-IC50=0.82 μM
IL1β-IC50 = 0.3 μM
IL1β-IC50 = 0.0013 μM
High throughput screening of the Pfizer compound file identi-
fied a number of hits with compound 1 being the most attractive.
Medium speed analoging efforts led to the 2-chlorophenethyl
derivative 2 as a potent and orally bioavailable lead. Further opti-
mization of 2 led to the clinical candidate CE-224,535 (3) (Fig. 1).4
In spite of its excellent attributes the 6-azauracil series had an
intrinsic liability characterized by low Vdss in dog and monkey
which resulted in a short predicted half-life in humans. To over-
come these liabilities, we initiated parallel efforts to find alternate
modalities for the 6-azauracil ring of 2, which we felt was a
primary contributor to the observed short half-life.5 We envisioned
that a parallel medicinal chemistry approach would offer an opti-
mal opportunity to identify a replacement for the 6-azauracil in
our lead series. Retrosynthetically we hypothesized that a readily
accessible boronate such as 5 would allow for the parallel prepara-
tion of bi-aryl amides A via a Suzuki reaction (Scheme 1).
Figure 1. Lead structures and candidate CE-224535.
O
O
Ar(Het-Ar)
B
Br
Cl
H
N
H
N
OH
Cl
O
Cl
O
Cl
Cl
O
A
5
3
Scheme 1. Proposed synthesis of targets.
Thus reaction of acid 3 with o-Cl-phenethylamine under stan-
dard coupling conditions yielded the amide 4 in excellent yield.
⇑
Corresponding author.
E-mail address: Chakrapani.Subramanyam@pfizer.com (C. Subramanyam).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.06.117

5476
C. Subramanyam et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5475–5479
Table 1
Br
Br
SAR for Aryl analogs of compound 2
Cl
a, b
H
N
90%
Het
Cl
CO2H
Cl
O
Cl
Cl
H
N
3
4
c
71%
O
A
O
O
Ar(Het)
B
a
d
Compound
Ar=
c log P
Monocyte IC50
0.11
(lM)
Cl
Cl
H
N
H
N
50-60%
N
N
6
4.42
Me
Cl
Cl
O
O
A
5
7
4.42
1.60
Scheme 2. Synthesis of 2-Cl-5-aryl benzamides. Reagents: (a) SOCl2/D
(b) o-Cl-phenythylamine, DIEA, DCE; (c) bis-pinacolatodiboron, KOAc, PdCl2(DPPF),
DMF, ; (d) ArX, PdCl2(DPPF), 2.0 M Na2CO3, DMF.
;
N
D
8
9
4.21
4.21
5.44
4.8
5.2
4.6
N
N
Conversion of 4 to the boronate 5 was accomplished by reaction
10
CF3
with bis-pinacolatodiboron in presence of PdCl2(DPPF).6 Finally
Me
Suzuki–Miyaura coupling6 of
5 with a variety of aryl and
N
N
11
12
4.92
4.62
0%inh@1.0
0%inh@1.0
l
M
heteroaryl halides led to the desired targets in modest isolated
yield after high throughput purification via reverse phase HPLC
(Scheme 2).7
l
M
Analogs from this array were evaluated for the ability to inhibit
the release of lL-1b from monocytes stimulated by ATP,8 The activ-
Me
N
N
ity spanned a range from 500 nM to double digit lM. The 2-pyridyl
analog 6 was the most potent, with activity similar to that of the
lead 2. Removal of the methyl group from 6 (compound 7) led to
a 10-fold drop in potency. The 3-pyridyl and 4-pyridyl analogs (8
and 9) were much less potent. Replacement of the methyl group
in 6 with CF3 (compound 10) led to considerable loss of activity.
Walking the methyl group around the pyridyl ring in 6 (com-
pounds 11–13) also led to a substantial drop in potency. Replacing
the methyl group of 6 with either chloro or primary amide (14 and
15) led to 5- to 7-fold loss in activity. Replacing the primary amide
with other electron withdrawing groups such as cyano or methyl
ketone (16 and 17), however, led to a considerable loss in activity.
Reduction of the ketone to the secondary alcohol (18) restored
most of the activity. The pyridazine 19 and 20 were 2- to 3-fold less
potent than 6 while the C-linked 6-azauracil analog 21 was 10-fold
less potent (Table 1).
Although we were able to find some alternatives for the 6-aza-
uracil portion of 2, these analogs all suffered from relatively high
lipophilicity. This was reflected in their high in vitro microsomal
clearance.9 To address these issues, we embarked on an analoging
effort focused on reducing the lipophilicity of these analogs and at
the same time improving the activity. To lower lipophilicity, we
borrowed from the work in the 6-azauracil series, and replaced
the o-Cl-phenethyl side chain of the potent analogs described in
Table 1 with 1-hydroxycycloheptane.3 Simultaneously we also ex-
plored varying the heterocycle portion of compound 6. The synthe-
sis of these analogs is depicted in Schemes 3 and 4. Thus the readily
accessible boronate 25 was converted to the desired targets by a
sequence that involved Suzuki reaction with the heteroaryl halide
followed by saponification of the ester and amide coupling with
the cycloheptanol amine (Scheme 3).
13
4.92
0%inh@1.0
l
M
Me
Cl
14
15
16
17
5.21
3.57
4.04
4.25
0.46
N
0.72
CONH2
CN
N
N
N
0%inh@1.0
0%inh@1.0
l
M
M
l
COCH3
OH
Me
18
3.69
0.668
N
N
N
N
19
20
3.73
3.89
0.25
0.49
Me
OMe
O
NH
NH
21
2.51
1.05
O
N
a
Inhibition of the stimulation of the release of lL-1b from monocytes by ATP
Data is for n = 2.
b
As predicted, a number of these analogs either maintained or
improved the activity of the corresponding o-Cl phenethyl amides
and also improved human microsomal clearance. The pyridyl ana-
log 27 was equipotent to the analog with a 6-azauracil head
group.3 Pyridazinone 32, which mimics the 6-azauracil in com-
pound 2 was threefold less potent. The five-membered analogs
(28–30, 36–39) in general were more potent than the six-mem-
bered heterocyclic analogs probed. N-Methylation of pyrazole 36
led to a drop in activity with the N-2 isomer 37 more potent than
the N-1 analog 38. The regioisomeric analog where the C5-position
has been methylated (compound 39) provided an unexpected
Preparation of the pyrazole targets 36–38 involved intermedi-
acy of the methyl ketone 34. Thus Sonogashira reaction10 of
bromide 23 with trimethylsilyl acetylene followed by acid cata-
lyzed hydration with formic acid11 led to the methyl ketone 34
which was converted to pyrazoles 36 and 39 via the enamine 35.
Finally alkylation of pyrazole NH led to the N-methyl analogs 37
and 38 (Scheme 4).

C. Subramanyam et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5475–5479
5477
As shown in Table 2, the activity in this assay was 2- to 15-fold less
potent than that observed in low serum monocyte assay.
O
O
Compound 39 emerged as a viable lead from the non 6-
azauracil series and was profiled further (Fig. 2). It showed good
oral bioavailability in rats. The exquisite selectivity of compound
39 for human receptor versus murine/rat precluded its evaluation
in standard rodent models of inflammation. However, an in vivo
model was developed that allowed a pharmacodynamic assess-
ment to be conducted. In this model, LPS-activated human
monocytes are implanted into the peritoneal cavity of Balbc mice.
Subsequently, these animals are dosed orally with test compound,
followed 60-min later by an ip injection of ATP to promote IL-1b
posttranslational processing. IL-1b released extracellularly is mea-
sured using a human-specific ELISA. In this model, 39 is a very
effective inhibitor of ATP-induced IL-1b output (Fig. 3) after an
oral dosing of 4 and 20 mpk. Unfortunately, compound 39 suf-
fered from the same low Vdss and short t1/2 observed for the
6-azauracil analogs 2. We believed that these shortcomings could
be addressed by improving overall polarity of the molecule. Spe-
cifically we hypothesized that introduction of strategically placed
basic amino substituent could overcome the low volume of distri-
bution liability of 39.13 In silico tools predicted such amino
substituents would provide a 4- to 6-fold increase in Vdss while
maintaining the favorable clearance profiles. Synthesis of the ami-
no substituted analogs is shown in Scheme 5. Compound 39 was
alkylated with chiral epoxy nosylate using a phosphazene base14
followed by opening of the intermediate epoxide with ammonia
and provided the desired amines 40, 41 (Scheme 5).
Br
Br
B
a
b
CO2H
CO2Me
CO2Me
Cl
Cl
Cl
23
24
25
c
Het
Het
d, e
H
N
CO2Me
OH
Cl
O
Cl
26
OMe
O
N
N
HN
N
f, e
H
N
CO2Me
OH
Cl
O
Cl
31
32
Scheme 3. Preparation of cycloheptanol analogs. Reagents and conditions: (a) HCl,
MeOH, 98%; (b) bis-pinacalotodiboron, KOAc, PdCl2(DPPF), DMF, , 60%; (c) Het-X,
Pd(PPh3)4, Cs2CO3, 4 Å sieves, dioxane, 80 °C, 50–60%; (d) NaOH, MeOH, 100%; (e) 1-
(aminomethyl)cycloheptanol, HOBT, EDCI, TEA, DMF, 60–70%; (f) HBr,
Amino alcohols 40 and 41 were very potent in inhibiting the re-
lease of IL1-b from monocytes both in the presence and absence of
human blood (IC50 = <0.005 lm). In addition, introduction of the
D
D
.
polarity also improved their microsomal clearance (HLM t1/2
>120 min). Unfortunately, these modifications also led to poorly
permeable analogs as evidenced by their poor flux in Caco-2 cells
and this was reflected in poor oral bioavailability of 40 in rats
(Fig. 4). We did, however, establish that introduction of the amino
groups improved Vdss and t1/2 of these analogs upon iv dosing.
Modulating the basicity of the terminal amine and reducing the
number of H-bond donors15 in 40 may provide a path forward to
improve the oral bioavailability of these analogs.
20-fold boost in activity while maintaining a favorable t1/2 in the
HLM assay (Table 2). Selected compounds were also tested for their
ability to inhibit uptake of YOPRO-1 by ATP-treated P2X7R over-
expressing HEK293 cells to provide direct evidence of P2X7R im-
pact.12 A number of these analogs were evaluated for their ability
to inhibit LPS/ATP-induced IL-1b release in diluted human blood.8
SiMe3
O
O
NMe2
Br
R
a
b
c
CO2Me
CO2Me
CO2Me
CO2Me
Cl
Cl
Cl
Cl
24
33
34
35
d,e,f
R
N
N
N
N
NH
N
g
+
H
N
H
N
H
N
OH
OH
OH
Cl
O
Cl
O
Cl
O
37
38
36, R = H
39, R = Me
Scheme 4. Synthesis of pyrazoles. Reagents: (a) Trimethylsilylacetylene, PdCl2(PPh3)2, PPh3, CuI, TEA, 98%; (b) 98% formic acid,
D
, 75%; (c) N,N-dimethylformamide dimethyl
acetal (or) N,N-dimethylacetamide dimethyl acetal,
% over three steps; (g) NaH, MeI (for R = H in 36).
D, 90%; (d) hydrazine hydrate, THF, EtOH; (e) KOH, MeO; (f) 1-(aminomethyl)cycloheptanol, HOBT, EDCI, TEA, DMF, 60–70

5478
C. Subramanyam et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5475–5479
Table 2
Activity of cycloheptanol analogs
Het
H
N
OH
Cl
M)
O
Compound
Ar=
c Log P
4.13
Monocyte IC50
0.095
(
l
P2X7R Yo-Pro IC50
0.031
(lM)
Whole blood IC50
(lM)
HLM t1/2 (min)
N
27
28
0.514
NT
34.7
85
Me
NH
N
N
2.96
0.51
29
3.96
0.28
0.006
NT
NT
S
N
30
31
32
36
3.39
3.10
1.71
3.26
0.24
0.28
0.93
0.18
NT
22.3
77
O
N
N
0.028
0.18
0.51
NT
OMe
O
N NH
NT
12
0.012
0.61
NH
N
N
N
37
38
3.22
3.22
0.238
>1.0
0.44
NT
3.7
NT
Me
N
NMe
Me
NH
0.266
0.003
39
3.53
0.007
0.121
75
N
Monocyte IC50 = 0.007 μM
Yo-Pro IC50 = 0.003 μM
Whole Blood IC50 = 0.121 μM
HLM t1/2 = 75 min
NH2•HCl
NH2•HCl
OH
NH
N
H
N
N
N
N
HO
O
H
H
N
Papp(X10-6/s) = 26
H
H
N
N
OH
Rat in-vivo PK:
Cl
O
OH
OH
Cl
Cl
O
VDss = 0.7 L/kg
39
t
1/2 = 0.7 h
40
41
%F = 59
logP = 2.6 logD = 1.6
logP = 2.6 logD = 1.6
Figure 2. In vitro and PK profile of 39.
IL-1β IC50 (μM) = 0.002
Yo-Pro IC50 (μM) = 0.0005
Whole Blood IC50 (μM) = 0.003
HLM = >120min
Caco-2 A→B = 2.98x10-6 cm/sec
MDR A→B = 0.26x10-6 cm/sec
MDCK B→A = 2.39x10-6 cm/sec
MDCK/MDR1: 11.5
IL-1β IC50 (μM) = 0.007
Yo-Pro IC50 (μM) = 0.003
Whole Blood IC50 (μM) = 0.01
HLM = >120 min
Caco-2 A→B = 10.5x10-6 cm/sec
MDR A→B = 0.26x10-6 cm/sec
MDCK B→A = 2.2x10-6 cm/sec
MDCK/MDR1: 10.5
Rat in-vivo PK:
VDss = 11.7, t1/2 = 8.36 h, %F = 9
Figure 4. Profile of amino alcohols 40 and 41.
In summary, a new series of potent P2X7 receptor antagonists
has been discovered. The neutral analogs in this series possessed
good properties and oral bioavailability, but were hampered by a
low volume of distribution coupled with a moderate clearance
which led to a short half-life in vivo. Attempts to overcome these
issues by introduction of amino substituents improved these attri-
butes. Unfortunately this change led to analogs with less than opti-
mal oral absorption, presumably due to the negative changes in
permeability. Future endeavors will need to address the balance
of these properties for successful identification of a development
candidate.16
Figure 3. In vivo activity of 39 in murine implant model.

C. Subramanyam et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5475–5479
5479
N
N
ONs
1.
P
N
N
NH2•HCl
, DMF. 80 oC
O
,
R1
NH
N
N
N
R
2. 7.0N NH3 in MeOH, 65 oC
3. HCl in Dioxane
H
N
H
N
OH
OH
Cl
O
Cl
O
39
1
40:
R =OH, R = H
41: R1 = H, R = OH
Scheme 5. Preparation of amino alcohols.
9. Most of these analogs had in vitro clearance of >100 ll/min/mg.
10. Chinchilla, R.; Najera, C. Chem. Rev. 2007, 107, 874.
Acknowledgments
11. Lee, C. S.; Allwine, D. A.; Barbachyn, M. R.; Grega, K. C.; Dolak, L. A.; Ford, C. W.;
Jensen, R. M.; Seest, E. P.; Hamel, J. C.; Schaadt, R. D.; Stapert, D.; Yagi, B. H.;
Zurenko, G. E.; Genin, M. J. Bioorg. Med. Chem. 2001, 9, 3243.
12. In general these compounds were 3–5Â more potent in the Yo-Pro assay than
in the monocyte assay. For assay protocols, see: Duplantier, A. J.;
Subramanyam, C. WO 2,003,042,191, 2003.
The authors would like to thank Christopher J. O’Donnell, and
Mark Mitton-Fry carefully reading early drafts of this manuscript
and providing useful suggestions to improve it.
13. (a) Lombardo, F.; Obach, R. S.; Shalaeva, M. Y.; Gao, F. J. Med. Chem. 2002, 45,
2867; (b) Lombardo, F.; Obach, R. S.; Shalaeva, M. Y.; Gao, F. J. Med. Chem. 2004,
47, 1242; (c) Lombardo, F.; Obach, R. S.; DiCapua, F. M.; Bakken, G.; Lu, J.;
Potter, D. M.; Gao, F.; Miller, M. D.; Zhang, Y. J. Med. Chem. 2006, 49, 2262; (d)
Berellini, Giuliano; Springer, Clayton; Waters, Nigel J.; Lombardo, Franco J. Med.
Chem. 2009, 52, 4488.
14. (a) Schwesinger, R. Chimia 1985, 39, 269; (b) Xu, W.; Mohan, R.; Morrissey, M.
M. Bioorg. Med. Chem. Lett. 1998, 8, 1089.
15. Veber, D. F.; Johnson, S. R.; Cheng, H.-Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D.
J. Med. Chem. 2002, 45, 2615.
References and notes
1. Ralevic, V.; Burnstock, G. Pharmacol. Rev. 1998, 50, 413.
2. (a) Romagnoli, R.; Baraldi, P. G.; CruzLopez, O.; LopezCara, C.; Preti, D.; Borea, P.
A.; Gessi, S. Expert Opin. Ther. Targets 2008, 12, 647; (b) Solle, M.; Labasi, J.;
Perregaux, D. G.; Stam, E.; Petrushova, N.; Koller, B. H.; Griffiths, R. J.; Gabel, C.
A. J. Biol. Chem. 2001, 276, 125.
3. Guile, Simon D.; Alcaraz, Lilian; Birkinshaw, Timothy N.; Bowers, Keith C.;
Ebden, Mark R.; Furber, Mark; Stocks, Michael J. J. Med. Chem. 2009, 52, 3123.
4. Duplantier, A. J.; Dombroski, M. A.; Subramanyam, C.; Beaulieu, A. M.; Chang,
S.-P.; Gabel, C. A.; Jordan, C.; Kalgutkar, A. S.; Kraus, K. G.; Labasi, J. M.; Mussari,
C. r; Perregaux, D. G.; Shepard, R.; Taylor, T. J.; Trevena, K. n A.; Whitney-
Pickett, C.; Yoon, K. Bioorg. Med. Chem. Lett. 2011, 21, 3708.
16. A recent publication from Pfizer describing work completed after the work
described herein disclosed
a series of orally and brain available P2X7
antagonists which began with compound 27. See: Chen, X.; Pierce, B.; Naing,
W.; Grapperhaus, M. L.; Phillion, D. P. Bioorg. Med. Chem. Lett. 2010, 20, 3107.
5. We hypothesized that the weakly acidic nature of compound 2 may be the
primary contributor to the observed low Vdss
.
6. Ishiyama, T.; Murata, M.; Miyaura, N. J. Org. Chem. 1995, 60, 7508.
7. Dombroski, M. A.; Duplantier, A. J.; Subramanyam, C. WO 2,004,099,146, 2004.
8. Analogs were evaluated as antagonists of the P2X7 R by assessing whether they
inhibited the release of IL-1b from LPS activated human monocytes maintained
in the presence of low serum. Promising compounds were re-evaluated in an
analogous assay, employing human whole blood samples sequentially treated
with LPS and ATP. For assay protocols, see: Duplantier, A. J.; Subramanyam, C.
WO 2,003,042,191, 2003.