Synthesis and SAR development of novel P2X7 receptor antagonists for the treatment of pain: Part 2

Article abstract of DOI:10.1016/j.bmcl.2011.10.037

Novel P2X₇ antagonists were developed using a purine scaffold. These compounds were potent and selective at the P2X₇ receptor in human and rodent as well as efficacious in rodent pain models. Compound 15a was identified to have oral potency in several pain models in rodent similar to naproxen, gabapentin and pregabalin. Structure-activity relationship (SAR) development and results of pain models are presented.

Full text of DOI:10.1016/j.bmcl.2011.10.037

Bioorganic & Medicinal Chemistry Letters 21 (2011) 7287–7290
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and SAR development of novel P2X7 receptor antagonists
for the treatment of pain: Part 2
Stephanie Brumfield a, , Julius J. Matasi a, Deen Tulshian a, Michael Czarniecki a, William Greenlee a,
⇑
Charles Garlisi a, Hongchen Qiu a, Kristine Devito a, Shu-Cheng Chen a, Yongliang Sun a, Rosalia Bertorelli a,
Justin Ansell a, William Geiss b, Van-Duc Le b, Gregory S. Martin b, Samuel A. Vellekoop b, James Haber b,
Melissa L. Allard b
a Merck Research Laboratories, 2015 Galloping Hill Road, Kenilworth, NJ 07033-0539, USA
b Albany Molecular Research, Inc., 21 Corporate Circle, Albany, NY 12212-5098, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Novel P2X7 antagonists were developed using a purine scaffold. These compounds were potent and
selective at the P2X7 receptor in human and rodent as well as efficacious in rodent pain models.
Compound 15a was identified to have oral potency in several pain models in rodent similar to naproxen,
gabapentin and pregabalin. Structure–activity relationship (SAR) development and results of pain models
are presented.
Received 6 September 2011
Revised 7 October 2011
Accepted 11 October 2011
Available online 21 October 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
P2X7
Purine derivatives
Inflammatory pain
Neuropathic pain
Located primarily in lymphocytes, macrophages, and mast
cells,1 the P2X7 ligand-gated ion channel belongs to the super fam-
ily of adenosine 50-triphosphate (ATP)-sensitive (P2) receptors.2
The P2 receptor family has been shown to play a role in the mod-
ulation and transmission of nociceptive signals altered by various
forms of injury.3 Of the seven ionotropic P2X subtypes, P2X4 and
P2X7 have been associated with the release of proinflammatory
cytokines, such as Il-1b.4 This has been shown to facilitate pain
transmission.4,5 While ATP has been shown to be a non selective
agonist for the P2X family,2 it has been demonstrated in the liter-
ature, due to its unusual phenotypic characteristics6 that it is pos-
sible to synthesize small molecules which are selective P2X7
antagonists.7 It is also important to note that P2X7 receptor knock-
out mice did not have inflammatory or neuropathic hypersensitiv-
ity, however normal nociceptive processing was persevered.8 For
these reasons we believe that P2X7 is an attractive target in the
treatment of pain.
(MIA)-induced hyperalgesia model but had only marginal efficacy
in chronic constriction injury (CCI) rat model.11 In our quest to de-
velop a compound that would treat patients across the broad spec-
trum pains, we continued SAR efforts in this area. Our thoughts
were that a compound with good whole blood potency and sus-
tainable oral exposure would achieve our goal.
Our previous work has indicated that bulkier substitutions were
tolerated at the C-2 position of the core. To evaluate this in detail,
the 2-trifluoromethylpyridine substitution at the C-2 position was
replaced with a biphenyl and we investigated various fluorinated
benzyl substituents at N-1 position (Fig. 2). These initial set of
compounds were prepared using Scheme 1 and results are pre-
sented in Table 1.
Alkylation of commercially available 3 with 3,4-difluorobenzyl
chloride provided compound 4 as a mixture of regio-isomers. The
Our recently published work9 described the discovery of com-
pound 1 (Fig. 1) as a prototype P2X7 antagonist. This compound
exhibited good affinity for P2X7 in human and rat with a moderate
pharmacokinetic (PK) profile the in rapid rat PK model.10 It showed
oral efficacy similar to that of naproxen in rat mono-iodacetate
O
N
N
N
N
h P2X7 IC50 = 50 nM
r P2X7 IC50 = 32 nM
h WB (IL-1β) IC50 = 82 nM
RR AUC = 2685 nM. hr
Conc.6h = 227
(C-2)
(N-1)
CF3
N
N
(C-7)
F
H
1
F
⇑
Corresponding author.
E-mail address: Stephanie.Brumfield@merck.com (S. Brumfield).
Figure 1. Previously disclosed P2X7 antagonist.9
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.10.037

7288
S. Brumfield et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7287–7290
Table 1
Initial SAR modifications
O
N
O
N
N
N
N
N
N
N
CF3
O
N
N
N
N
N
N
N
H
H
N
F
F
F
F
10
H
1
F
2
hwbIL-1β = 82 nM
hwbIL-1β = 178 nM
Entry
Fluorine substitution pattern on benzyl ring
hwb IL-1b12
IC50 (nM)
Figure 2. SAR of selected C-2 substitutions.
2
3,4
2
3
178
937
269
237
870
845
1036
107
10a
10b
10c
10d
10e
10f
10g
Cl
N
Cl
4
2,3
2,4
2,5
3,5
N
a
b
N
N
N
Cl
N
N
H
Cl
N
3
F
The IC50 and Ki data are an average of at least three measurements.
4
F
(10b and 10c). A similar drop in whole blood activity was also ob-
served with all disubstituted analogs (10b–f) which had fluorine at
the ortho position. However, 3,4 and 3,5-difluorobenzyl derivatives
(e.g., 2 and 10g) exhibited h-whole-blood potency similar to our
prototype compound 1. Both compounds (2 and 10g) were dosed
(10 mg/kg) in the rapid rat PK model10 to determine oral expo-
sures. Compound 10g exhibited an AUC of 1113 nM h with a
O
N
O
N
N
N
HN
c
d
N
N
Cl
N
Cl
F
F
93 nM concentration at 6 h, while
2
exhibited an AUC
6
5
F
F
(3763 nM h) with a concentration of 392 nM at 6 h. Compared to
compound 1, neither compound displayed superior oral exposure.
To this end we went back and explored the substitution around the
C-7 position with the biphenyl at C-2. All compounds were pre-
pared using methods described in Scheme 1 and the data is pre-
sented in Table 2.
O
N
O
N
N
N
N
N
N
e
f
N
From previous work9 we knew that small substitutions were
tolerated at C-7. Gem-dimethyl was the first substitution evalu-
ated. It showed an IC50 of 85 nM in h-whole-blood assay. When
this substitution was expanded to include a spirocycle (11b) a
twofold reduction in h-whole blood potency was observed. When
the ring was expanded by one carbon (11c) it provided a com-
pound equipotent to that of compound 2. We then began to
explore the substitution pattern around the benzyl ring to ensure
that the 3,4-difluoro substitution was optimal. We found that 11d
was equipotent to that of its isopropyl counterpart 10c (Table 1).
However, when fluorine was added to the ortho position (11e)
an increase in whole-blood potency was observed. Compound
11f was equipotent to that of its isopropyl counterpart (10g).
Two compounds (11a and 11f) with good h-whole blood potency
were examined for oral exposure in the rapid rat PK model10 at
10 mg/kg dose. Compound 11a exhibited very good exposure with
an AUC of 6863 nM h and a plasma concentration of 1369 nM at
6 h. Based on its good exposure, compound 11a was evaluated in
rodent whole blood assay and showed a rodent whole-blood
potency of 121 nM. Based this, compound 11a was evaluated in
variety of rodent pain models encompassing both inflammatory
and neuropathic pain. Gabapentin (50 mg/kg) and naproxen
(10 mg/kg) were used as positive controls. All efficacy measure-
ment was made at 4 h post dose. In the Streptozocin (STZ)-induced
diabetic neuropathy model13 in rat, 11a exhibited efficacy compa-
rable to gabapentin at an oral dose of 10 mg/kg. Similarly, com-
pound 11a produced efficacy similar to gabapentin when dosed
orally (at 60 mg/kg) in chronic constriction injury (CCI) model of
neuroinflammatory pain. Compound 11a also produced a naprox-
en like effect when dosed at 50 mpk in the rat mono-iodacetate
(MIA)-induced hyperalgesia model.14 Compound 11a was the first
HN
N
F
OH
F
F
F
8
7
O
O
N
N
N
N
g
N
N
H
Br
N
N
N
N
F
F
H
F
F
9
2
Scheme 1. Reagents and conditions: (a) 3,4-Difluorobenzyl chloride, K2CO3, DMF, rt
24 h; (b) 0.2 N NaOH reflux, 24 h; (c) ethyl iodide, K2CO3, DMF, rt 24 h; (d) (R)-(ꢀ)-
2-amino-3-methyl-1-butanol, TEA, NMP 150 °C 24 h; (e) thionyl chloride, DCM, rt
1 h; (f) NBS, DCM rt 1 h; (g) 4-biphenyl boronic acid, Pd(PPh3)4, Na2CO3, CH3CN/H2O
(3:1), lk 130 °C, 30 min.
hydrolysis of the major regio-isomer of 4 in aq sodium hydroxide
afforded compound 5. Alkylation of 5 using ethyl iodide produced
6, which on displacement with (R)-(ꢀ)-2-amino-3-methyl-1-buta-
nol gave intermediate 7. Cyclization of 7 using thionyl chloride,
produced compound 8. Compound 8 was treated with NBS to give
compound 9. The synthesis of the desired targets was completed
by performing a Suzuki coupling to give compound 2. Other ana-
logs were prepared using this synthetic route.
As described in Table 1, substitution of fluorine at the ortho po-
sition (Entry 10a) was least tolerated and a fourfold loss in whole
blood activity was observed compared to meta and para positions

S. Brumfield et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7287–7290
7289
Table 2
Core SAR modifications
O
N
N
N
N
N
F
n
R1 R2
11
Entry
R1
R2
n
Substitution of fluorines on benzyl ring
hwb IL-1b12
IC50 (nM)
2
Isopropyl
Me
Cyclohexyl
Me
Me
Me
H
Me
0
0
0
1
0
0
0
3,4
3,4
3,4
3,4
4
178
85
11a
11b
11c
11d
11e
11f
388
108
276
700
91
Me
Me
Me
Me
2,4
3,5
Me
N
Cl
N
a
Cl
Table 3
Cl
Distal phenyl SAR modifications
13
12
b
N
O
O
O
B
R
N
N
N
14
N
N
N
Scheme 2. Reagents and conditions: (a) Phenyl boronic acid, Pd(PPh3)4, Na2CO3,
toluene/ethanol/water (2.4:1:1.4), 80 °C, 24 h; (b) bispinacolediborane, Pd(dba),
Xphos, KOAc, dioxane, 80 °C, 24 h.
15
F
F
Entry
R
hwb IL-1b12
IC50 (nM)
RR AUC10
nM h
Concn6 h
nM
compound in this series which showed efficacy similar to control
in all three pain models. This compound was screened against a
panel of ion channels and was found to have greater than 80% inhi-
15a
15b
15c
15d
15e
15f
15g
15h
15i
15j
15k
15l
15m
15n
H
24
52
101
25
104
28
110
39
137
45
50
1944
95
13670
2714
1621
231
1836
3559
ND
2638
537
224
0
395
897
ND
0
1569
897
0
ND
537
1263
2-Fluoro
2-Methoxy
3-Methyl
3-Trifluoromethyl
3-Fluoro
3-Trifluoromethoxy
4-Methyl
4-Trifluoromethyl
4-Fluoro
bition at 10 l
M for several different ion channels (Na+ Channel
(site 2), Ca+ Channel (L, diltiazem), Clꢀ Channel and Nor epineph-
rine transporter). In addition, 11a exhibited 55% inhibition of P2X4
at 10 lM.
To improve selectivity of 11a against the ion channels, we
decided to investigate the incorporation of pyridine in the interior
phenyl ring of the C-2 biphenyl. The rational behind this was prior
knowledge that pyridine substitutions at C-29 provided com-
pounds that were potent in the whole-blood assay. Scheme 2 pro-
vided the synthetic route to achieve this substitution pattern.
Compound 14 was synthesized via a Suzuki coupling of 12 and
commercially available boronic acids. Compound 13 was then con-
verted to the boronic ester and taken through the final Suzuki cou-
pling as described in Scheme 1. This combined schemes provided
access to the compounds presented in Table 3.
Replacement of the interior phenyl ring of 11a with a pyridine
ring gave 15a. This compound had very good potency in human
whole blood assay. A variety of substituents were investigated at
the distal phenyl moiety. Substitution around the distal phenyl
ring was also well tolerated with most compounds exhibiting
potencies under 100 nM in the whole-blood assay. However, once
trifluoromethyl and trifluoromethoxy ( 15e, 15g, 15i and 15l)
groups were substituted, a drop in whole-blood potencies was ob-
served. This is most dramatically demonstrated by 15l. Further dif-
ferentiation of these compounds came when their oral exposures
were measured in the rapid rat PK model. Methyl substitution
(15d and 15h) produced compounds with extremely low expo-
sures. Although compounds 15b, 15f, and 15j showed significant
exposures, compound 15a had best sustained levels in plasma.
The in vivo potency of this compound was measured in several
inflammatory and neuropathic pain models as described for
0
5677
3559
0
ND
4137
7538
4-Methoxy
4-Trifluoromethoxy
2,4-Difluoro
3,5-Difluoro
55
compound 11a. The potency of compound 15a (10 mg/kg dose)
was similar to that of 11a in STZ-induced diabetic neuropathy
model, but was much more potent (1–3 mg/kg dose) in both MIA
and CCI models in rat. In addition, potency of 15a was also mea-
sured in the spinal nerve ligation (SNL) model15 for neuropathic
pain in rat. Pregabalin (10 mg/kg) was used as a positive control.
It exhibited efficacy comparable pregabalin at a dose between 10
and 30 mg/kg given via the oral route. All efficacy measurements
were carried out at 4 h post dose. Compared to 11a, compound
15a was also found to be significantly more selective when
screened for its inhibitory activity against other ion channels.
Excited with these results, we further studied compound 15a
for oral exposures in higher species. To our disappointment the
presence of glutathione adducts was observed in dog urine. These
adducts were determined to form through enzymatic epoxidation
of the distal phenyl ring. The glutathione formation was mitigated
by the addition of electron withdrawing groups on the distal phe-
nyl ring. Unfortunately these substitutions resulted in significant

7290
S. Brumfield et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7287–7290
6. Surpenant, A.; Rassendren, F.; Kawashima, E.; North, R. A.; Buell, G. Science
1996, 272, 735.
drop in potency in the whole blood assay. Work is currently in pro-
gress to identify a compound with oral efficacy similar to that of
15a without this glutathione issue.
7. Representative examples not limited to: (a) Chambers, L. J.; Stevens, A. J.;
Moses, A. P.; Michel, A. D. S.; Walter, D.; Davies, D. J.; Livermore, D. G.; Fonfria,
E.; Demont, E. H.; Vimal, M.; Theobald, P. J.; Beswick, P. J.; Gleave, R. J.; Roman,
S. A.; Senger, S. Bioorg. Med. Chem. Lett. 2010, 20, 3161; (b) Chen, X.; Pierce, B.;
Naing, W.; Grapperhaus, M. L.; Phillion, D. P. Bioorg. Med. Chem. Lett. 2010, 20,
3107; (c) Nelson, D. W.; Gregg, R. J.; Kort, M. E.; Perez-Medrano, A.; Voight, E.
A.; Wang, Y.; Grayson, G.; Namovic, M. T.; Donnelly-Roberts, D. L.; Niforatos,
W.; Honore, P.; Jarvis, M. F.; Faltyneck, C. R.; Carroll, W. A. J. Med. Chem. 2006,
49, 3659; (d) Florjancic, A. S.; Peddi, S.; Perez-Medrano, A.; Li, B.; Namovic, M.
T.; Grayson, G.; Donnelly-Rpbberts, D. L.; Jarvis, M. F.; Carroll, W. A. Bioorg. Med.
Chem. Lett. 2008, 18, 2089; (e) Baxter, A.; Bent, J.; Bowers, K.; Bradok, M.;
Brough, S.; Fagura, M.; Lawson, M.; McInally, T.; Mortimore, M.; Roberts, M.;
Weaver, R.; Webborn, P. Bioorg. Med. Chem. Lett. 2003, 13, 4047.
8. Chessell, I. P.; Hatcher, J. P.; Bountra, C.; Michel, A. D.; Hughes, J. P.; Green, P.;
Egerton, J.; Murfin, M.; Richardson, J.; Peck, J. L.; Grahames, C. B. A.; Casula, M.
A.; Yaingou, Y.; Birch, R.; Anand, P.; Buell, G. B. Pain 2005, 114, 386.
9. Matasi, J. J.; Brumfield, S.; Tulshian, D.; Czarniecki, M.; Greenlee, W.; Garlisi, C.
G.; Qiu, H.; Devito, K.; Chen, S.-C.; Sun, Y.; Bertorelli, R.; Geiss, W.; Le, V.-D.;
Martin, G. S.; Vellekoop, S. A.; Haber, J.; Allard, M. L. Bioorg. Med. Chem. Lett.
2011, 21(12), 3805.
In conclusion, current efforts identified compound 15a with
sustained oral exposures and very good potency in the human
whole blood assay. This compound was very efficacious in inflam-
matory (MIA), inflammatory/neuropathic (CCI) and neuropathic
(STZ and SNL) pain models in rat. Its efficacy was similar or better
than positive controls (naproxen, gabapentin and pregabalin) used
in these studies. Unfortunately, it also formed glutathione adducts
in higher species. Our current SAR development is focused on
resolving this issue and results will be discussed in further
publications.
Acknowledgments
The authors also acknowledge Emily Freeman, Geoffrey Giarmo,
Christian Holst and Lisa Peterson, all from Albany Molecular
Research Inc., for their contributions towards this work.
10. Mei, H.; Korfmacher, W.; Morrison, R. AAPS J. 2006, 8(3), E49.
11. (a) Bennett, G. J.; Xie, Y. K. Pain 1988, 33, 87; (b) Hybter, J. C.; Gogas, K. R.;
Hedley, L. O.; Fontana, D. J. Eur. J. Pharmacol. 1997, 324, 153.
12. Compounds were assessed for their ability to inhibit ATP-mediated IL-1b
production in LPS-stimulated whole blood cultures. Compounds were added to
whole blood for 30 min followed by LPS for 3 h. ATP was added and IL-1b
concentrations were measured using commercially available ELISA kits. Each
compound was tested at least twice and the average values are provided. For
this assay, the average coefficient of variance was 37%.
13. (a) Courteix, C.; Eschalier, A.; Lavarenne, J. Pain 1993, 53, 81–88; (b) Ahlgren, S.
C.; Levine, J. D. Neuroscience 1993, 52, 1049.
14. Bove, S. E.; Calcaterra, S. L.; Brooker, R. M.; Huber, C. M.; Guzman, R. E.; Juneau,
P. L.; Schrier, D. J.; Kilgore, K. S. Osteoarthritis Cartilage 2003, 11, 821–830.
15. Kim, S. H. T.; Chung, J. M. Pain 1992, 50, 355.
References and notes
1. (a) Burnstock, G. Brit. J. Anaesth. 2000, 84, 476; (b) Burnstock, G. Parmacol. Rev.
1972, 24, 509.
2. McGaraughty, S.; Jarvis, M. F. Drug Dev. Res. 2006, 6, 376.
3. Liu, X.; Slater, . M. Curr. Opin. Investig. Drugs 2005, 6, 65.
4. Ledeboer, A.; Sloane, E. M.; Milligan, E. D.; Frank, M. G.; Mahony, J. H.; Maier, S.
F.; Watkins, L. R. Pain 2005, 115, 71.
5. Sweitzer, S.; Martin, D.; Deleo, J. A. Neuroscience 2001, 103, 529.