The discovery of novel indole-2-carboxamides as cannabinoid CB1 receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2010.10.104

The discovery and structure-activity relationship of a novel series of indole-2-carboxamide antagonists of the cannabinoid CB₁ receptor is disclosed. Compound 26i was found to be a high potency, selective cannabinoid CB₁ antagonist.

Full text of DOI:10.1016/j.bmcl.2010.10.104

Bioorganic & Medicinal Chemistry Letters 21 (2011) 497–501
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The discovery of novel indole-2-carboxamides as cannabinoid
CB1 receptor antagonists
Phillip M. Cowley a, , James Baker c, David R. Barn a, Kirsteen I. Buchanan a, Ian Carlyle a, John K. Clark a,
⇑
Thomas R. Clarkson a, Maureen Deehan b, Darren Edwards a, Richard R. Goodwin a, David Jaap a,
Yasuko Kiyoi a, Chris Mort a, Ronald Palin a, Alan Prosser a, Glenn Walker b, Nick Ward c, Grant Wishart a,
Trevor Young a
a Department of Chemistry, MSD, Newhouse, Lanarkshire, ML1 5SH, UK
b Department of Molecular Pharmacology, MSD, Newhouse, Lanarkshire, ML1 5SH, UK
c Department of Pharmacology, MSD, Newhouse, Lanarkshire, ML1 5SH, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery and structure–activity relationship of a novel series of indole-2-carboxamide antagonists
of the cannabinoid CB1 receptor is disclosed. Compound 26i was found to be a high potency, selective
cannabinoid CB1 antagonist.
Received 28 September 2010
Revised 19 October 2010
Accepted 20 October 2010
Available online 26 October 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Cannabinoid
CB1
Antagonist
Indole
The identification of cannabinoid CB1 and CB2 G-protein cou-
pled receptors and their endogenous ligands in the early 1990s
stimulated further research efforts to identify novel ligands for
these receptors and thoroughly explore their therapeutic poten-
tial.1 One of the outcomes of these studies was the identification
of selective CB1 receptor antagonists, two of the prototypical com-
pounds being rimonabant (SR141716A, 1) and LY320135 (2).2,3
Profiling of CB1 antagonists in pre-clinical and clinical studies sug-
gest therapeutic roles in a variety of indications, with rimonabant
gaining approval for the treatment of obesity in Europe and other
territories outside the US.4 Unfortunately, psychiatric adverse
events were observed in clinical use and the compound was subse-
quently withdrawn from the market.
In this publication, the identification of a novel series of indole-
2-carboxamides as CB1 receptor antagonists and their subsequent
SAR is reported. Compound 3a (Fig. 1), derived from a hit resulting
from a medium throughput screening campaign, was identified as
a potential starting point for an optimisation program. However,
3a showed relatively weak potency as a CB1 receptor antagonist
and high lipophilicity (CB1 pIC50 = 6.37, c log P = 6.32).5 Therefore
the initial optimisation strategy was to explore the SAR around
the scaffold, increasing potency whilst attempting to improve
physico-chemical properties. Potencies of compounds as antago-
nists of the human CB1 receptor were determined in a CB1 reporter
gene assay as described previously.6 pIC50’s are reported as the
means of at least two independent experiments.
Radioligand binding experiments have shown that compounds
1 and 2 compete with [3H]CP55,940 at the human CB1 receptor.2,3
This suggests a common binding site for 1 and 2. In the absence of
preliminary binding data for compound 3a it was hypothesised
that 3a may also interact with the CB1 receptor at the same binding
site. Therefore, to assist the chemistry exploration a preliminary
superposition model of compounds 1, 2 and 3a was developed.
An appropriate low energy conformation of 1 consistent with con-
formational observations from the Cambridge Structural Database7
was selected as a rigid template. Flexible superposition of 2 and 3a
onto the fixed conformation of 1 afforded the model depicted in
Figure 2.8 All three molecules share two aromatic hydrophobes,
this is supported by previous literature reports highlighting the
importance of aromatic interactions in CB1 receptor recognition.9
Interestingly the nitrile moiety of 2 matches the indole C-5 posi-
tion of 3a perhaps suggesting that polar electron withdrawing in-
dole substituents may be tolerated. The amides of 1 and 3a
overlay, which is consistent with site-directed mutagenesis and
modelling studies suggesting the carbonyl of 1 plays an important
role in interacting with the CB1 receptor.10,11 However, compound
2 does not appear to share this feature with the carbonyl of 2
⇑
Corresponding author. Tel.: +44 1698 736105; fax: +44 1698 736187.
E-mail address: phillip.cowley@merck.com (P.M. Cowley).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.10.104

498
P. M. Cowley et al. / Bioorg. Med. Chem. Lett. 21 (2011) 497–501
N
O
N
N
Cl
O
N
H
N
N
H
O
N
Cl
Cl
O
O
O
Cl
3a
2
1
Figure 1. Structures of rimonabant (1), LY320135 (2) and early lead 3a.
OEt
OEt
O
1
a
1
R
R
N
H
N
O
Ar
4
9
b
2
H
N R
OH
O
1
1
c
R
R
N
O
N
Ar
Ar
10
Figure 2. Superposition model of 1 (green), 2 (orange) and 3a (magenta).
11-23, 25-26
21 R1=5-CN
24 R1=5-CH2NH2
d
positioned in the region of the pyrazole N-2 of 1. The piperidine of
1, the 4-methoxyphenyl of 2 and the phenethyl side chain of 3a
form a general hydrophobic region perhaps suggesting that this
may be a region of tolerance that would be suitable for exploration
with diverse substituents.
The synthesis of indole-2-carboxamide analogues of compound
3a is depicted in Schemes 1–3, all starting materials and reagents
were purchased from commercial sources unless otherwise stated.
Ethyl 5-chloroindole-2-carboxylate (4a) was either benzylated by
deprotonation with NaH followed by addition of the bromobenzyl
or arylated by heating the indole and bromobenzene with K2CO3,
CuBr, Cu(OAc)2 and KOH (Scheme 1). Hydrolysis gave the acid
Scheme 2. Reagents and conditions: (a) NaH, BrCH2Ar; (b) aq NaOH, EtOH; (c)
R2NH2, EDCI, HOBt, Et3N, CH2Cl2, rt; (d) Raney Ni, EtOH, 5 bar H2.
(6), which was converted to the acid chloride before coupling with
the amine to afford the desired amides 3a–l. The acid was reduced
with borane–THF complex to the alcohol (7), oxidation to the alde-
hyde with pyridineÁSO3 was followed by reductive amination to af-
ford the desired amine 8.
Substituted indoles 4 were deprotonated with sodium hydride
and addition of the corresponding bromobenzyl afforded the
Cl
Cl
Cl
O
O
f
a or b
(R1 = CH2Ph)
N
OH
N
H
OEt
N
R
OR
1
4a
5 R = Et
6 R = H
c
7
d, e
g, h
2
R
N R
3
Cl
Cl
N
R
N
O
N
1
H
3a-l
8
Scheme 1. Reagents and conditions: (a) R1Br, NaH, DMF, 60 °C; (b) PhBr, K2CO3, CuBr, KOH, Cu(OAc)2, 150 °C; (c) KOH, EtOH/H2O; (d) SOCl2, CH2Cl2; (e) R2R3NH, Et3N, CH2Cl2;
(f) BH3ÁTHF, THF, reflux; (g) pyridineÁSO3, Et3N, DMSO; (h) PhCH2CH2NH2, NaBH4, AcOH, AcONa, Na2SO4, EtOH.

P. M. Cowley et al. / Bioorg. Med. Chem. Lett. 21 (2011) 497–501
499
N
a
b
c
c
O
O
O
Si
Br
Si
H2N
Si
28
29
27
F
F
F
F
F
F
N
N
N
NH
NH2
30
31
32
Scheme 3. Reagents and conditions: (a) Isobutyronitrile, LDA, THF, À70 °C; (b) bromoacetonitrile, CaCO3, DMF, rt; (c) 1 M LiAlH4 in THF, Et2O, 0 °C.
desired N-benzyl compounds 9 (Scheme 2). Hydrolysis of the ester
in aq NaOH in ethanol gave the acid (10) which was coupled with
amines to afford the desired amides 11–26. The 5-aminomethyl
Table 1
analogue 24 was prepared from the corresponding nitrile (21) by
hydrogenation in the presence of Raney nickel. Only two of the
Cl
R1
amines were not commercially available. 4-(tert-Butyldimethylsi-
lyloxy)-2,2-dimethylbutan-1-amine (29) was synthesised from
isobutyronitrile and (2-bromoethoxy)-tert-butyldimethylsilane
N
R2
(27), followed by reduction of nitrile 28 (Scheme 3). The tert-butyl-
dimethylsilyl group was removed during the work up of the reac-
tion between 29 and the carboxylic acid 10 to afford the desired
compound 25g (Scheme 2). 2-(4,4-Difluoropiperidin-1-yl)ethan-
amine (32) was prepared from 4,4-difluoropiperidine (30) and bro-
moacetonitrile (Scheme 3) followed by reduction of the nitrile 31.
Any benzyl bromides that were not commercially available were
prepared from the appropriate benzylalcohol using PPh3 and CBr4.
Table 1 shows the SAR around the initial modifications to the
carboxamide side chain and indole N-1 position. Shortening the
phenethyl moiety (3b) and replacement by cyclohexylethyl (3d)
resulted in modest decreases in potency. However, the chain ex-
tended analogue 3c showed similar potency to the parent com-
pound 3a. Tertiary amide 3e exhibited an order of magnitude
decrease in potency whereas removal of the carbonyl (8) had a les-
ser effect. Replacement of the indole N-1 benzyl by phenyl (3f) or
cyclohexylmethyl (3g) also resulted in decreased potency. Trans-
posing the hydrazide side chain of 1 to the current series affording
compound 3h gave a fivefold drop in potency with the ethylpiper-
idine analogue 3i showing a further decrease in potency. This sug-
gests the amide substituents of 1 and 3a are not completely
overlapping which is supported by the superposition model. Di-
methyl substituted alcohol 3l was effectively equipotent to 3a
although the alcohol 3k and ether 3j both showed reduced
potency.
The N-(3-hydroxy-2,2-dimethylpropyl)-carboxamide substitu-
ent offered moderate CB1 potency and reduced lipophilicity (c log P
4.78), therefore compound 3l was selected for further exploration
around the indole ring and indole N-1 benzyl substituent (Table 2).
Both ortho and meta-substitution of the benzyl ring resulted in in-
creased potencies however, para-substitution (data not shown)
was not tolerated. 2-Methyl substitution (11) afforded a 14-fold
increase in CB1 potency over 3l and was more effective than triflu-
oromethoxy or chloro. Substitution at the meta-position with 3-
methyl, 3-chloro and 3-cyano also resulted in increased potencies.
However, 3-trifluoromethoxy (17) exhibited a 17-fold increase in
potency versus 3l and was deemed to be optimal. Maintaining
the 3-trifluoromethoxybenzyl moiety substitution of the indole
ring was pursued. The unsubstituted analogue (18) and the 5-
methyl (20) derivative afforded potencies similar to the 5-chloro
compound (17). Moving the chloro substituent to the C-6 position
(19) was not well tolerated and gave a 31-fold decrease in potency.
Guided by the superposition model a polar electron withdrawing
nitrile substituent was introduced at the indole C-5 position (21)
and resulted in almost an order of magnitude increase in CB1
Compd
R1
R2
CB1 pIC50
H
N
*
3a
CH2Ph
6.37
5.93
O
O
H
N
*
3b
3c
CH2Ph
CH2Ph
H
N
*
*
6.57
O
O
H
N
3d
3e
3f
CH2Ph
CH2Ph
Ph
5.83
5.35
<5
N
*
*
*
*
O
O
O
O
H
N
H
N
3g
3h
CH2cHex
CH2Ph
5.65
5.67
H
N
N
H
N
*
*
*
*
*
N
3i
3j
3k
3l
8
CH2Ph
CH2Ph
CH2Ph
CH2Ph
CH2Ph
5.37
5.55
<5
O
O
O
O
H
N
O
H
N
OH
OH
H
N
6.57
5.80
H
N

500
P. M. Cowley et al. / Bioorg. Med. Chem. Lett. 21 (2011) 497–501
Table 2
Table 3
H
N R1
NC
OH
H
N
N
O
R1
N
O
R2
OCF3
Compd
R1
CB1 pIC50
7.96
Compd
R1
R2
CB1 pIC50
11
12
13
14
15
16
17
18
19
20
21
22
23
24
5-Cl
5-Cl
5-Cl
5-Cl
5-Cl
5-Cl
5-Cl
H
6-Cl
5-CH3
5-CN
5-SO2CH3
5-CONH2
5-CH2NH2
2-CH3
2-Cl
2-OCF3
3-CH3
3-Cl
7.71
6.80
7.21
7.42
7.64
7.23
7.81
7.64
6.32
7.82
8.74
6.67
6.29
6.13
25a
OH
OH
*
*
25b
25c
7.90
7.42
OH
*
3-CN
3-OCF3
3-OCF3
3-OCF3
3-OCF3
3-OCF3
3-OCF3
3-OCF3
3-OCF3
OH
O
25d
25e
7.80
7.38
*
*
25f
8.82
OH
*
25g
25h
25i
7.12
6.38
7.01
*
*
OH
OH
potency. However, sulfone 22 and carboxamide 23 failed to dem-
onstrate similar increases in potency and together with primary
amine 24 showed significantly decreased potency upon compari-
son with the 5-chloro compound 17.
*
OH
25j
8.37
An extensive SAR exploration around the amide moiety of 21
was subsequently carried out, a small selection of analogues are
reported in Table 3. Monomethyl analogues 25a–c all showed a
decrease in potency as did the propyl alcohol 25d. Reduced po-
tency was also observed for the methyl ether analogue of 21,
compound 25e. However, the cyclobutyl derivative 25f main-
tained a potency similar to that of 21. Modification of the length
of the side chain afforded reduced potency as observed for com-
pounds 25g–i, although in the case of 25i much of the potency
loss can be recovered via the cyclopentyl derivative 25j. In an ef-
fort to improve physico-chemical properties the incorporation of
a basic amine was investigated. Primary amine 25k and second-
ary amine 25l showed profound 170- and 219-fold decreases in
potency, respectively, upon comparison to 21. However, the sig-
nificantly less basic 4,4-difluoropiperidine analogue 25m exhib-
ited good potency.
*
*
OH
25k
25l
6.51
6.40
NH2
N
*
*
N
25m
8.66
F
F
Table 4
OH
H
NC
N
N
A final optimisation cycle around the benzyl substituent of 21
was performed (Table 4). Replacing the 3-trifluoromethoxy group
of 21 by trifluoromethyl (26a) gave a small decrease in potency,
with methoxy (26b) and fluoro (26c) analogues showing further
decreases. Mono 2-substituted compounds 26d and 26e main-
tained some potency as CB1 antagonists although significantly less
than their 3-substituted counterparts. The 2-chloro derivative 26f
showed similar potency to the 3-fluoro compound 26c. As ex-
pected 4-substitution was not tolerated, this was exemplified by
the low potency of 26g. The SAR clearly indicates that 2 and 3-
substitution of the benzyl moiety is tolerated with the 3-position
being more favourable. Therefore in a final effort to further in-
crease potency several di-substituted analogues were synthesised.
The 2,5-substituted compounds 26i and 26j showed modest in-
creases in potency compared to 21 but 2,3-substituted 26h was
disappointing.
O
R1
Compd
R1
3-CF3
3-OCH3
3-F
2-CF3
2-OCH3
2-Cl
CB1 pIC50
26a
26b
26c
26d
26e
26f
26g
26h
26i
8.44
8.02
7.14
6.86
7.68
7.21
6.66
6.46
8.93
9.17
4-F
2-OCH3,3-CF3
2-OCH3,5-OCF3
2-OCH3,5-CF3
26j
Selected compounds were advanced to in vivo evaluation in a
reversal of mouse hypothermia induced by the cannabinoid
agonist WIN 55,212-2 (Table 5).12 Compounds 21 and 26i showed
potent reversal of WIN 55,212-2 induced hypothermia comparable
to that of rimonabant 1 (ID50 po 1.29 lmol/kg (0.40–4.17)). Com-
pounds 25f, 25m and 26j were significantly less potent or inactive
in vivo when dosed orally. The affinities of compounds 21 and 26i

P. M. Cowley et al. / Bioorg. Med. Chem. Lett. 21 (2011) 497–501
501
In summary compound 26i14 has been identified as a novel po-
tent and selective CB1 antagonist. The compound displays reversal
of the hypothermia induced by a cannabinoid agonist in vivo, with
potency comparable to that of rimonabant. Further assessment of
compound 26i and related compounds will be reported in due
course.
Table 5
Compd
Hypothermia ID50
lmol/kg (95% CI)
CB1 pKi
CB2 pKi
21
3.42 (0.97–12.04)
>10
>100
0.88 (0.58–1.34)
9.10
<5
25f
25m
26i
26j
9.40
<5
ꢀ10
Acknowledgements
We would like to thank our Analytical Chemistry colleagues for
structure and purity determination.
Table 6
References and notes
21
26i
Rat
Mouse
Rat
Mouse
Dog
1. Howlett, A. C.; Barth, F.; Bonner, T. I.; Cabral, G.; Casellas, P.; Devane, W. A.;
Felder, C. C.; Herkenham, M.; Mackie, K.; Martin, B. R.; Mechoulam, R.; Pertwee,
R. G. Pharmacol. Rev. 2002, 54, 161.
2. Rinaldi-Carmona, M.; Barth, F.; Héaulme, M.; Shire, D.; Calandra, B.; Congy, C.;
Martinez, S.; Maruani, J.; Néliat, G.; Caput, D.; Ferrara, P.; Soubrié, P.; Brelière, J.
C.; Le Fur, G. FEBS Lett. 1994, 350, 240.
3. Felder, C. C.; Joyce, K. E.; Briley, E. M.; Glass, M.; Mackie, K. P.; Fahey, K. J.;
Cullinan, G. J.; Hunden, D. C.; Johnson, D. W.; Chaney, M. O.; Koppel, G. A.;
Brownstein, M. J. Pharmacol. Exp. Ther. 1998, 284, 291.
CL (mL/min/kg)
Vss (L/kg)
t1/2 (h)
po dose (mg/kg)
Cmax (ng/mL)
tmax (h)
64
6.5
1.5
10a
11
2
3.4
0.5
1.7
5a
665
2
104
9.9
1.5
5a
16
1
16.9
1.7
7.0
3.5
10.3
2.4c
400
0.8
1.6
4.6b
237
1.3
F (%)
2
13
8
30
67
4. Akbas, F.; Gasteyger, C.; Sjödin, A.; Astrup, A.; Larsen, T. M. Obes. Rev. 2009, 10,
58.
a
b
c
5% (v/v) DMSO, 5% (v/v) cremophor EL in 5% (w/v) mannitol (aq).
5% (v/v) mulgofen in saline.
5% (v/v) DMSO, 5% (v/v) cremophor EL in 5% (v/v) mulgofen in saline.
5. c log P 4.3, BioByte Corp. 201 W. 4th St. #204 Claremont, CA 91711-4707, USA.
6. Price, M. R.; Baillie, G. L.; Thomas, A.; Stevenson, L. A.; Easson, M.; Goodwin, R.;
McLean, A.; McIntosh, L.; Goodwin, G.; Walker, G.; Westwood, P.; Marrs, J.;
Thomson, F.; Cowley, P.; Christopoulos, A.; Pertwee, R. G.; Ross, R. A. Mol.
Pharmacol. 2005, 68, 1484.
at the cannabinoid receptors were determined by radioligand
binding experiments performed at least in duplicate. Affinity for
CB1 was measured by competition with [3H]SR141716A binding
to membranes prepared from CHO cells expressing the human
CB1 receptor. Affinity for CB2 was measured by competition with
[3H]CP55,940 binding to recombinant human CB2 receptors ex-
pressed in Sf9 cell membranes. Both 21 and 26i showed high affin-
ity for the CB1 receptor and greater than four orders of magnitude
selectivity over the CB2 receptor.
7. Allen, F. H. Acta Crystallogr., Sect. B 2002, 58, 380.
8. All modelling was performed within MOE2008.10. Superposition was
performed with the Flexible Alignment tool within MOE2008.10 as
distributed by Chemical Computing Group, Inc., 1010 Sherbrooke Street
West, Suite 910, Montreal, Quebec, Canada.
9. McAllister, S. D.; Rizvi, G.; Anavi-Goffer, S.; Hurst, D. P.; Barnett-Norris, J.;
Lynch, D. L.; Reggio, P. H.; Abood, M. E. J. Med. Chem. 2003, 46, 5139.
10. Hurst, D. P.; Lynch, D. L.; Barnett-Norris, J.; Hyatt, S. M.; Seltzman, H. H.; Zhong,
M.; Song, Z.-H.; Nie, J.; Lewis, D.; Reggio, P. H. Mol. Pharmacol. 2002, 62, 1274.
11. Hurst, D.; Umejiego, U.; Lynch, D.; Seltzman, H.; Hyatt, S.; Roche, M.;
McAllister, S.; Fleischer, D.; Kapur, A.; Abood, M.; Shi, S.; Jones, J.; Lewis, D.;
Reggio, P. J. Med. Chem. 2006, 49, 5969.
Pharmacokinetic studies were performed in male MF1 mice,
male Wistar rats and female beagle dogs using either 1 or 2 mg/
kg for iv dosing and with doses and vehicles for oral dosing as
shown in Table 6. Clearance for both compounds was much higher
in rat than mouse, leading to lower rat oral bioavailability. Volume
of distribution showed high inter-species variability, which was
not explained by differences in plasma protein binding (both com-
pounds >99% bound in all species). Compound 26i showed low
clearance in dog and an oral bioavailability of 67%. Furthermore
26i (c log P = 4.76) showed significantly reduced lipophilicity com-
pared to 1 (c log P = 6.47) and starting point 3a (c log P = 6.32).5
Compound 26i also exhibited low affinity for the hERG channel
in a [3H]dofetilide binding assay (pKi 5.1).13 In wider selectivity
12. The antagonist or vehicle (5% mulgofen in saline, 10 mL/kg) was administered
po 75 min before rectal temperature was measured. WIN 55,212-2 mesylate
(10 lmol/kg, 10 mL/kg) was administered s.c. 60 min prior to the rectal
temperature measurement. The 60 min pre-treatment with WIN 55,212-2
mesylate corresponded to the maximal hypothermia attained by this agonist.
Rectal temperature was measured using a metal probe with a Fluke 51K/J
Thermometer. The probe was covered in
a lubricant (vaseline) and was
inserted approximately 1.5 cm into the rectum. The highest temperature stable
for 10 s was recorded. Following completion of the test animals were
humanely terminated.
13. The affinity of the test drugs for the hERG cardiac K+ channel was determined
by their ability to displace tritiated dofetilide in membrane homogenates from
HEK-293 cells expressing the hERG channel.
14. Analytical data for compound 26i: ESI-MS: m/z = 476.0 [M+H]+. 1H NMR
(400 MHz, CDCl3): d 8.02 (d, J = 1.5 Hz, 1H), 7.48 (dd, J = 1.5 and 8.5 Hz, 1H),
7.40 (d, J = 8.5 Hz, 1H), 7.05 (dd, J = 8.5 and 2.5 Hz, 1H), 7.01 (s, 1H), 6.85 (d,
J = 8.5 Hz, 1H), 6.81 (br t, J = 6.0 Hz, 1H), 6.28 (d, J = 2.5 Hz, 1H), 5.82 (s, 2H),
3.89 (s, 3H), 3.27 (d, J = 6.5 Hz, 2H), 3.24 (t, J = 6.5 Hz, 1H), 3.13 (d, J = 6.5 Hz,
2H), 0.90 (s, 6H). Elemental analysis: measured C, 60.71; H, 5.16; N, 8.77; F,
11.48. Theory C, 60.63; H, 5.09; N, 8.84; F, 11.99.
profiling 26i showed no appreciable affinity at 10 lM for a panel
of 62 molecular targets comprising of GPCRs, ion channels, trans-
porters, enzymes and nuclear receptors.