Discovery of piperidine carboxamide TRPV1 antagonists

Article abstract of DOI:10.1016/j.bmcl.2008.07.035

A series of piperidine carboxamides were developed as potent antagonists of the transient receptor potential vanilloid-1 (TRPV1), an emerging target for the treatment of pain. A focused library of polar head groups led to the identification of a benzoxazinone amide that afforded good potency in cell-based assays. Synthesis and a QSAR model will be presented.

Full text of DOI:10.1016/j.bmcl.2008.07.035

Bioorganic & Medicinal Chemistry Letters 18 (2008) 4569–4572
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of piperidine carboxamide TRPV1 antagonists
Wing S. Cheung, Raul R. Calvo, Brett A. Tounge, Sui-Po Zhang, Dennis R. Stone, Michael R. Brandt,
*
Tasha Hutchinson, Christopher M. Flores, Mark R. Player
Johnson & Johnson Pharmaceutical Research and Development, Welsh and McKean Roads, Spring House, PA 19477, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of piperidine carboxamides were developed as potent antagonists of the transient receptor
potential vanilloid-1 (TRPV1), an emerging target for the treatment of pain. A focused library of polar
head groups led to the identification of a benzoxazinone amide that afforded good potency in cell-based
assays. Synthesis and a QSAR model will be presented.
Received 27 May 2008
Revised 8 July 2008
Accepted 10 July 2008
Available online 15 July 2008
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
TRPV1
VR1
Capsaicin
A variety of potentially damaging chemical, thermal, and
mechanical stimuli can elicit signals from peripheral nerve end-
ings, ultimately leading to the perception of pain. These nerve end-
ings are small diameter sensory neurons known as nociceptors.
The transient receptor potential vanilloid 1 (TRPV1) is found on
nociceptive primary afferents of C- and Ad-fibers, and is considered
to be a key polymodal integrator of noxious sensations.1 It is acti-
vated by heat, low pH, and endogenous lipidic mediators, such as
anandamide, as well as by pungent natural products such as capsa-
icin and resiniferatoxin.
TRPV1 agonists cause calcium influx, depolarization, and an ini-
tial sensation of pain. Activation over time causes physiologic
changes to the nociceptor resulting in desensitization and a subse-
quent period of analgesia. Thus, local exposure to capsaicin has
been shown to provide pain relief in humans, though the irritancy
of TRPV1 agonists to date limits their application to primarily top-
ical indications. Collectively, these findings have inspired intense
pharmaceutical interest in exploring the potential utility of orally
bioavailable TRPV1 antagonists as therapies for acute and chronic
pain. Presently, GlaxoSmithKline (SB705498), Neurogen/Merck
(NGD-8243/MK-2295), and Lilly/Glenmark (GRC-6211) are investi-
gating TRPV1 antagonists in phase II trials for acute pain.2 Herein,
we describe our findings on a carboxamide series of TRPV1
antagonists.
polar head, linker, and hydrophobic tail (Table 1). Keeping the lin-
ker fixed, we elected to explore the SAR of the polar head and
hydrophobic tail regions.
The various piperidine carboxamides were synthesized as
shown in Scheme 1.4 The piperidine acid (Method A) or ester
precursor (Method B) was subjected to Pd-mediated arylation.
Subsequent amide formation by either EDC (Method C) or acid
chloride-mediated coupling (Method D) afforded the desired com-
pounds, which were purified by silica gel chromatography.
We initially explored the addition of a number of different het-
erocycles in the polar head group region. The isomeric 2,3-dihydro-
1H-indoles 2–4 (Table 1) showed the sensitivity to changes in this
part of the polar region. Compound 3 exhibited activity but 2 and 4
were devoid of potency. Indole 5 showed increased potency but
isomer 6 was fourfold less potent. Indole isomer 7 was inactive,
illustrating the importance of examining amide substitution posi-
tion with each heterocycle explored. From this exercise, 4H-
benzo[1,4]oxazin-3-one 8 was found to be a moderately potent
antagonist in a functional assay with an IC50 of 180 nM and a Ki
of 65 nM in a hTRPV1 binding assay.5 Next, various substituents
were placed on the phenyl ring in the nonpolar tail region of 9 to
examine the effect on functional potency (Table 2). In order to bet-
ter understand the nature of the substituent effects, a QSAR model
was constructed.
Our initial work identified a class of piperidine carboxamides
such as 1 (Table 1), which behaved as weak antagonists of the
TRPV1 receptor. Similar to other prototypical piperazine carbox-
amides,3 the structure could be divided into three components:
Compounds 8–29 were used to construct the QSAR model. The
compounds were first subjected to Ligprep to generate an initial
set of 3D coordinates.6 Next, a set of low energy conformations
was generated for each ligand using the Monte Carlo Multiple Min-
imum (MCMM) method in MacroModel (OPLS-2005, water,
e = 1,
40 kJ/mol energy window, 1000 steps).7 Using these conformations
as input, a Phase common pharmacophore was built for the top 4
* Corresponding author. Tel.: +1 610 458 6980; fax: +1 610 458 8258.
E-mail address: mplayer@its.jnj.com (M.R. Player).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.07.035

4570
W. S. Cheung et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4569–4572
Table 1
Human TRPV1 functional potencies of polar head substituents
Polar
Head
Nonpolar
Tail
Linker
O
N
R1
R NH
Compound
R
R1
Functional
potency
IC50 (nM)
1
2
3
4
5
6
7
8
Quinolin-3-yl
3-CF3
3-CF3
3-CF3
3-CF3
3-CF3
3-CF3
3-CF3
3-CF3
600
5610
800
>9000
230
1070
>9000
180
1-Acetyl-2,3-dihydroindol-7-yl
1-Acetyl-2,3-dihydroindol-6-yl
1-Acetyl-2,3-dihydroindol-5-yl
Indol-4-yl
Indol-5-yl
Indol-6-yl
Figure 1. Overlay of the twenty-two model compounds as obtained from the Phase
pharmacophore model. Red ? hydrogen bond acceptor, Blue ? hydrogen bond
donor, Orange ? aromatic, and Green ? hydrophobic.
3-Oxo-3,4-dihydrobenzo[1,4]oxazin-6-yl
calculated. These included clogP and the set of descriptors from
QuickProp.9,10
Ar2
NH
To assess the best descriptors to use in terms of fitting the
TRPV1 functional activity data (pIC50 = ꢀ1.0 ꢁ log (Functional
Activity IC50)), a stepwise multiple linear regression run was per-
formed using the JMP statistics package.11 One compound (29) in
the training set could not be fit to the same trends as the remaining
21 ligands, so it was removed from the final model. The final QSAR
model obtained was limited to a 3-parameter fit:
O
OR
O
OH
O
Methods A, B
30-60%
Methods C, D
65-80%
N
H
N
Ar1
N
Ar1
R = H, Et
pIC50 ¼ 29:4007 þ 0:5893 ꢁ c log P ꢀ 0:2214 ꢁ FISA;
where FISA is the hydrophilic solvent accessible surface area as cal-
culated in QuickProp. The model yielded an r2 = 0.71 and a root
mean squared error (RMSE) = 0.22 (Fig. 2).
ð1Þ
Scheme 1. Reagents and conditions: Method
A (R = H): Pd(OAc)2, 2-(tBu)2P-
biphenyl, NaOtBu, Ar1X, THF, 65 °C; Method B (R = Et): (a) Pd(P(tBu)3)2, K3PO4,
Ar1X, DME, 100 °C; (b) MeOH, LiOH, 80 °C, 3 h; Method C: EDC, HOBt, Et3N, DMF,
Ar2NH2, rt; Method D: oxalyl chloride, CH2Cl2, Et3N, Ar2NH2, 0 to 25 °C.
The QSAR model contains two terms that capture the hydropho-
bic/hydrophilic nature of these TRPV1 compounds. Compounds
with a large clogP values (i.e., more hydrophobic) exhibited higher
functional potency. Since, in the ligand training set, we have fo-
cused on changes in the tail region (the generally more hydrophilic
head is fixed), this means that the tail region in particular favors
hydrophobic moieties. While the 2D clogP descriptor captures a
large part of the variation seen in the TRPV1 functional activity,
the 3D conformation-dependent FISA term improves the fit mark-
edly. We see in this term a penalty for any solvent-exposed hydro-
philic surface area. For example, this can be seen in the methoxy-
substituted phenyl ring of 13.
Table 2
Human TRPV1 binding affinities and functional potencies of nonpolar tail substituents
O
NH
O
N
R1
O
NH
Compound
R1
H
3-Fluoro
3-Chloro
3-Bromo
3-Methoxy
3-Methyl
2-Fluoro
2-Chloro
Binding affinity
Ki (nM)
Functional potency
IC50 (nM)
9
—
1810
520
190
330
1420
490
670
220
160
380
150
150
180
130
100
60
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
84
97
—
—
225
1470
93
2-Bromo
2-Trifluoromethyl
2-Nethyl
71
68
61
2-Ethyl
125
78
2-Isopropyl
2-Bromo-3-fluoro
2-Methyl-3-chloro
2,3-Dimethyl
4-t-Butyl
3,4-Difluoro
3,4-Dimethyl
3,5-Difluoro
3,4,5-Trifluoro
81
100
28
142
56
314
137
49
90
850
380
370
59
compounds based on functional activity (23–25, 29).8 This model
was then used to align all the molecules in the QSAR training set
(Fig. 1). Finally, for each aligned pose several descriptors were
Figure 2. Final QSAR model for the functional activity of compounds 7–26
(r2 = 0.71, RMSE = 0.22). Compound 29 could not be fit (red triangle in above plot).

W. S. Cheung et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4569–4572
4571
Table 3
Table 4
Stability in human and rat liver microsomes
Human TRPV1 functional potencies of optimized polar head substituents
Compound
HLM Stability t1/2 (min)
RLM Stability t1/2 (min)
F
O
8
17.3
27.7
11.4
15.6
>60
44.3
81.7
58.9
2.8
4.4
9
7.2
6.5
15.5
19.4
63.8
F
N
11
21
22
24
26
28
29
R NH
F
R
Functional Potency % inh.
or IC50 ( M)
l
H
N
30 5-yl
31 7-yl
32 8-yl
9% inh. at 1 lM
O
O
55% inh. at 1
lM
Five selected benzoxazinone derived carboxamides (8, 11, 21,
22, and 24) were then subjected to in vitro metabolic stability tests
in human and rat liver microsomes (HLM, RLM) (Table 3). All com-
pounds demonstrated moderate to excellent stability in human
microsomes but were markedly less stable in rat microsomes. It
was also observed that fluorination on the nonpolar phenyl tail im-
proved rat and human liver microsomal stability (26, 28, and 29).
In fact, 29 not only possessed the best human and rat liver micro-
somal stabilities, but was also the most potent compound contain-
ing a 4H-benzo[1,4]oxazin-3-one polar head (Table 2).
We next turned our attention to additional modifications of the
4H-benzo[1,4]oxazin-3-one-containing polar head in order to in-
crease functional potency. First, the 5-, 7-, and 8-position regioi-
somers (30–32) were prepared. None of these were superior to
29. In fact, only the 7-isomer 31 showed any appreciable potency.
The importance of a free oxazine NH (29) was confirmed by the
complete eradication of activity by N-methylation (33). Replace-
ment of the oxazine ring oxygen with sulfur also reduced func-
tional activity by 33-fold (Table 4).
Inactive
O
X
N
33 X = Me, Y = O
34 X = H, Y = S
Inactive
1.96
lM
Y
35 X = Me, Y = CO2Et
36 X = Me, Y = CO2H
37 X = H, Y = CO2Et
0.87
l
M
Inactive
Inactive
Inactive
H
N
O
38 X = H, Y = CO2H
X
Y
39 X = H, Y = CH2CO2Me
40 X = H, Y = CH2CO2H
41 X = H, Y = (CH2)2OH
0.89
5% inh. at 1
0.005
lM
O
lM
l
M
In summary, a series of piperidine carboxamides were synthe-
sized and tested, resulting in the identification of the lead candi-
date 41. Due to potential dissolution-limited absorption and
possible efflux, further optimization is underway to explore the po-
tential of piperidine carboxamide TRPV1 antagonists for the treat-
ment of pain.
Numerous other modifications were made to the 4H-
benzo[1,4]oxazin-3-one polar head, namely the addition of addi-
tional polar functionality at the 2-position. The majority of these
were inactive, although esters 35 and 39 demonstrated relatively
weak functional potency. Compound 41, however, demonstrated
remarkable potency (IC50 = 5 nM).
Compound 41 was stable in RLM (t1/2 = 79 min) and in HLM (t1/2
= 55 min), comparable with the parent molecule 29. In addition, 41
was not a potent inhibitor of recombinant CYP enzymes (rCYP3A4
References and notes
1. (a) Suh, Y.-G.; Oh, U. Curr. Pharm. Design 2005, 11, 2687; (b) Ferrer-Montiel, A.;
Garcia-Martinez, C.; Morenilla-Palao, C.; Garcia-Sanz, N.; Fernandez-Carvajal,
A.; Fernandez-Ballester, G.; Planells-Cases, R. Eur. J. Biochem. 2004, 271, 1820;
(c) Owsianik, G.; D’hoedt, D.; Voets, T.; Nilius, B. Rev. Physiol. Biochem.
Pharmacol. 2006, 156, 61; (d) Appendino, G.; Szallasi, A. Prog. Med. Chem.
2006, 44, 146.
IC50 = 3.7 lM and rCYP2D6 IC50 > 10 lM). However, there was a
potential for PGP-mediated efflux, as 41 demonstrated apical to
basolateral transfer (A ? B) across a Caco-2 cell monolayer with
a Papp = 1.5 ꢂ 10ꢀ6 cm/s, but with a B ? A/A ? B ratio of 10.6.
Compound 41 was evaluated for in vivo efficacy in a rodent
model of thermal hyperalgesia.12 In the rat at an oral dose of
30 mg/kg, 41 produced a small but non-significant decrease in
radiant heat latency at 30 min post-dose, as compared to 0.5%
hydroxymethylcellulose vehicle (Fig. 3).
2. Szallasi, A.; Cortwright, D. A.; Blum, Charles A.; Eid, S. R. Nat. Rev. Drug Disc.
2007, 6, 357.
3. (a) Valenzano, K. J.; Grant, E. R.; Wu, G.; Hachicha, M.; Schmid, L.; Tafesse, L.;
Sun, Q.; Rotshteyn, Y.; Francis, J.; Limberis, J.; Malik, S. J. Pharm. Exp. Ther. 2003,
306, 377; (b) Swanson, D. M.; Dubin, A. E.; Shah, C.; Nasser, N.; Chang, L.; Dax, S.
L.; Jetter, M.; Breitenbucher, J. G.; Liu, L.; Mazur, C.; Lord, B.; Gonzales, L.; Hoey,
K.; Rizzolio, M.; Bogenstaetter, M.; Codd, E. E.; Lee, D. H.; Zhang, S.-P.; Chaplan,
S. R.; Carruthers, N. I. J. Med. Chem. 2005, 48, 1857.
4. Two general preparations of carboxylic acids: Method A. Sodium t-butoxide
(2.2 mmol) was added to a solution of phenyl halide Ar1X (1.0 mmol), Pd(OAc)2
(0.01 mmol), and 2-(di-t-butylphosphino)-biphenyl (0.02 mmol) in 5 ml of
THF. After 5 min of stirring, isonipecotic acid (1.0 mmol) was added to the
solution. The solution was heated to 65 °C for 18 h under an argon atmosphere,
then cooled to rt and partitioned between EtOAc and water. The aqueous layer
was separated and acidified using 3 N HCl and then extracted with EtOAc. The
organic layer was washed with brine, dried over MgSO4, filtered and
HPMC
14
41
13
12
11
10
9
concentrated to give the product.Method B. To
a solution of ethyl
isonipecotate (6.36 mmol) in 10 mL of anhydrous DME were added phenyl
halide Ar1X (9.54 mmol), bis(tri-t-butylphosphine)palladium(0) (0.636 mmol),
and potassium phosphate (12.7 mmol). The reaction mixture was heated to
100 °C and stirred for 14 h. Solids were filtered off from the resulting mixture,
and the collected filtrate was concentrated under reduced pressure to provide
the crude product, which was then redissolved into 10 mL of methanol. Two
milliliters of 1 N lithium hydroxide was added to hydrolyze the ester at 80 °C
for 3 h. Solvent was evaporated, and the basic residue was acidified by 2 N HCl.
The aqueous mixture was extracted with EtOAc. The combined organic layers
were washed with brine, dried over MgSO4, filtered, and concentrated under
reduced pressure to afford the crude product, which was then purified by flash
column chromatography with CH2Cl2 and CH3OH to yield the product. Two
general preparations for amide coupling: Method C. A solution of aniline Ar2NH2
8
7
6
5
(-24h )
0
30
60
100
180
Time (min) after oral administration
Figure 3. CFA radiant heat latency time course of 41.

4572
W. S. Cheung et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4569–4572
(0.54 mmol) carboxylic acid (0.18 mmol), hydroxybenzotriazole (HOBt)
membranes. Both assays were carried out under conditions previously
reported. Jetter, M. C.; Youngman, M. A.; McNally, J. J.; Zhang, S. -P.; Dubin,
A. E.; Nasser, N.; Dax, S. L. Bioorg. Med. Chem. Lett. 2004, 14, 3053–3056.
6. LigPrep, version 2.1, Schrödinger, LLC, New York, NY, 2005.
7. MacroModel, version 9.5, Schrödinger, LLC, New York, NY, 2007.
8. Phase, version 2.5, Schrödinger, LLC, New York, NY, 2007.
9. clogP, Daylight Chemical Information Systems, Inc., Aliso Viejo, CA, 2007.
10. QuickProp, version 3.0, Schrödinger, LLC, New York, NY, 2005.
11. JMP, version 6. SAS Institute Inc., Cary, NC, 1989–2005.
12. Complete Freunds adjuvant (CFA; 100 mL emulsion of saline and heat-killed
Mycobacterium tuberculosis in mineral oil) was injected into a single hind paw
of male Sprague–Dawley rats. Each rat was placed in a test chamber directly on
a warm glass surface and acclimated for 10 min. Response latencies to the heat
stimulus were recorded for each animal before and 24 h after the injection of
CFA. Directly following the post-CFA latency measurement, test compound or
hydroxymethylcellulose vehicle was administered orally to the animals. Post-
treatment withdrawal latencies were assessed at 0, 30, 60, 100, and 180 min
post-dose.
(0.54 mmol) and triethylamine (0.54 mmol) in 1 mL N,N-dimethylformamide
was treated with ethyl-3-(3 dimethylaminopropyl)-carbodiimide (EDC)
(0.54 mmol). The solution was stirred at room temperature for 18 h. 5 mL of
10% K2CO3(aq) were added to the solution, which was then filtered. The solid
was sequentially washed with water, CH3OH and then dried in vacuo to afford
the product. Method D. To a solution of carboxylic acid (0.122 mmol) in 5 mL of
CH2Cl2 was added oxalyl chloride (0.244 mmol) and a catalytic amount of DMF.
The reaction mixture was stirred for 2 h. Solvent was removed in vacuo. The
residue was redissolved in 5 mL of CH2Cl2. Aniline Ar2NH2 (0.122 mmol) and
triethylamine (0.183 mmol) were added sequentially into the acyl chloride
solution at 0 °C. The reaction mixture was warmed to 25 °C and stirred for
20 min. Concentration of the solvent provided the residue, which was then
purified by flash column chromatography with hexanes and ethyl acetate to
afford the product. Spectral data have been reported: Calvo, Raul R.; Cheung,
Wing S.; Player, Mark R. WO 06/058338.
5. The compounds were assayed for their ability to antagonize capsaicin-induced
calcium flux using a Ca2+-sensitive fluorescent dye and FLIPRTM technology
(Molecular Devices Inc.). The [3H]-RTX binding assay used hVR1/HEK293 cell