New bicyclic cannabinoid receptor-1 (CB1-R) antagonists

Article abstract of DOI:10.1016/j.bmcl.2005.10.019

A series of conformationally constrained bicyclic derivatives derived from SR141716 was prepared and evaluated as hCB₁-R antagonists and inverse agonists. Optimization of the structure-activity relationships around the 2,6-dihydro-pyrazolo[4,3-

Full text of DOI:10.1016/j.bmcl.2005.10.019

Bioorganic & Medicinal Chemistry Letters 16 (2006) 731–736
New bicyclic cannabinoid receptor-1 (CB1-R) antagonists
Philip A. Carpino,* David A. Griffith, Subas Sakya, Robert L. Dow, Shawn C. Black,
John R. Hadcock, Philip A. Iredale, Dennis O. Scott, Michael W. Fichtner, Colin R. Rose,
Robert Day, Joseph Dibrino, Mary Butler, Demetria B. DeBartolo, Darrin Dutcher,
Denise Gautreau, Jeff S. Lizano, Rebecca E. OꢀConnor, Michelle A. Sands,
Dawn Kelly-Sullivan and Karen M. Ward
Pfizer Global Research and Development—Groton Laboratories, Groton, CT 06340, USA
Received 8 August 2005; revised 5 October 2005; accepted 6 October 2005
Available online 2 November 2005
Abstract—A series of conformationally constrained bicyclic derivatives derived from SR141716 was prepared and evaluated as
hCB1-R antagonists and inverse agonists. Optimization of the structure–activity relationships around the 2,6-dihydro-pyrazol-
o[4,3-d]pyrimidin-7-one derivative 2a led to the identification of two compounds with oral activity in rodent feeding models (2h
and 4a). Replacement of the PP group in 2h with other bicyclic groups resulted in a loss of binding affinity.
Ó 2005 Elsevier Ltd. All rights reserved.
Endocannabinoids are endogenous fatty acid-derived
molecules (e.g., anandamide and 2-arachidonylglycerol)
that exhibit a wide range of pharmacological activities
such as stimulating appetite, decreasing locomotor
activity, increasing pain sensitivity, and enhancing func-
tion of the immune system.1,2 The biological properties
of these substances are mediated through two G-protein
coupled receptors—cannabinoid receptor type 1 (CB1-
R) which is expressed predominantly in the central ner-
vous system (amygdala, cortex, hippocampus, and
hypothalamus) and to a lesser extent in the periphery
(gastric tract and adipose tissue) and cannabinoid recep-
tor type 2 (CB2-R) which is found in the spleen and in
cells of the immune system (macrophages and T-lym-
phocytes).3 In the central nervous system, endocannabi-
noids behave as retrograde signaling messengers that
stimulate pre-synaptic receptors on neurons. CB1-R
activation results in inhibition of both adenylate cyclase
activity and voltage-gated ion channels. Over the past
years, numerous cannabinoid receptor modulators have
been identified including non-selective CB1-R/CB2-R
agonists such as CP-55940 and D9-tetrahydrocannabi-
nol (a constituent of Cannabis sativa L. or marijuana)4
and selective CB1-R antagonists and inverse agonists
such as SR141716 (1, rimonabant, Fig. 1).5–7
Keywords: Cannabinoid receptor antagonists; CB1-R; 2,6-Dihydro-
pyrazolo[4,3-d]pyrimidin-7-one; 2,6-Dihydro-pyrazolo[3,4-c]pyridin-7-
one; 2,6-Dihydro-pyrazolo[3,4-d]pyridazin-7-one; 1,9-Dihydro-purin-
6-one; Obesity; Conformational restriction.
*
Corresponding author. Tel.: +1 860 441 5694; e-mail:
philip.a.carpino@pfizer.com
Figure 1. Structures of 1 and bicyclic derivatives 2–7.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.10.019

732
P. A. Carpino et al. / Bioorg. Med. Chem. Lett. 16 (2006) 731–736
Overactivity of the endocannabinoid system has been
implicated in the development of obesity, a disease of
excess body weight that is associated with multiple
co-morbidities such as hypertension, diabetes mellitus,
and dyslipidemia.8 Modulation of this system using
CB1-R antagonists/inverse agonists represents a promis-
ing new approach for reducing body weight, preventing
weight re-gain, and decreasing co-morbidities associated
with excess body fat.9 One selective CB1-R antagonist,
SR141716, has recently completed phase III clinical tri-
als, while several other antagonists are in earlier stages
of clinical development.
ane (LiN(SiMe3)2), followed by reaction with diethyl
oxalate (EtO2CCO2Et) yielded 9a. A three-step proce-
dure involving nitrosation, 2-chlorophenylhydrazine
addition, and cyclocondensation was used to convert
9a into pyrazole-3-carboxylic acid ethyl ester 10a.
Reduction of the nitroso group in 10a and treatment
of the product 11a with an appropriately substituted
amidine derivative afforded PP 12a. N-alkylation using
commercially available R2X provided compounds
2a,c–h,j–l and 4a–c. Triflate 13, prepared in three steps
from 1-benzoyloxypropan-2-one 13a12 (Scheme 2), was
used to synthesize 2i. Treatment of compound 11a with
N-aminopiperidine and trimethylaluminum (AlMe3)
and condensation of the resulting hydrazide intermedi-
ate with triethyl orthoformate yielded 2b. PP com-
pounds 3a,b were prepared analogously to 2a,h except
for 2-chloroacetophenone 8b, which was used in the first
step and 4-chlorophenylhydrazine in the formation of
pyrazole intermediate 10b.
Several conformationally constrained analogs of the pyr-
azole-3-carboxamide derivative, SR141716, have recently
been described.10 As part of a program to identify new
rigid analogs of SR141716 with improved CB1-R poten-
cies, a series of diaryl compounds containing different
bicyclic scaffolds was prepared. Carboxamide groups
were incorporated within the framework of various het-
erocyclic scaffolds, rather than appended to such systems.
These compounds were specifically designed to explore
alternative spatial arrangements of the key functional
groups in SR141716, not to mimic the low-energy
conformer. Herein, we describe the syntheses, struc-
ture–activity relationships, and biological activities of
compounds 2–7 containing 2,6-dihydro-pyrazolo[4,3-
d]pyrimidin-7-one (PP), 2,6-dihydro-pyrazolo[3,4-c]pyri-
din-7-one, 2,6-dihydro-pyrazolo[3,4-d]pyridazin-7-one,
and 1,9-dihydro-purin-6-one groups (Fig. 1).
Pyrazole-3-carboxylic acid ester 1413 was used to prepare
2,6-dihydro-pyrazolo[3,4-c]pyridin-7-one 5 (Scheme 3).
Hydrolysis of the ethyl ester group in 14, followed by
coupling of the resulting carboxylic acid moiety with
Bicyclic PP derivatives 2a,c–l and 4a–c were prepared
from 4-chloroacetophenone 8a (Scheme 1).11 Deproto-
nation of 8a with lithium 1,1,1,3,3,3-hexamethyldisilaz-
Scheme 2. Reagents: (a) (diethylamino)sulfur trifluoride, CH2Cl2,
EtOH (94%); (b) NaOH/MeOH; (c) Tf2O, DMAP, NEt3, CH2Cl2
(72% from 13b).
Scheme 1. Reagents: (a) LiN(SiMe3)2, EtO2CCO2Et, Et2O, ꢀ78 °C to rt (100%); (b) NaNO2, HOAc; (c) Y-PhNHNH2ÆHCl; (d) i-PrOH, H2SO4 (56–
71% from 9); (e) Na2S2O4 (3 equiv), EtOAc, H2O (57–75%); (f) R1C(@NH)NH2ÆHOAc, EtOH, reflux (75–96%); (g) R2X, Cs2CO3, DMF, 100–115 °C
(35–55%); (h) N-aminopiperidine (1-piperidinamine), 2 M AlMe3 in toluene, CHCl3 (97%); (i) CH(OEt)3, toluene, 0.45 h at 100 °C, 2 h at 120 °C
(63%).

P. A. Carpino et al. / Bioorg. Med. Chem. Lett. 16 (2006) 731–736
733
Scheme 3. Reagents and conditions: (a) KOH, EtOH; (b) N-benzyl-N-(2,2-diethoxyethyl)amine, EDCÆHCl, EtN-i-Pr2, CH2Cl2; (c) p-TsOH, toluene,
reflux (16%); (d) POCl3, reflux, 120 h (77%); (e) 3 M HCl, THF, 60 °C (80%); (f) CF3CH2OTf, Cs2CO3, DMF (79%).
N-benzyl-N-(2,2-diethoxyethyl)-amine, yielded the pyra-
zole-3-carboxamide derivative 15. Intramolecular cyclo-
condensation of 15 under acidic conditions, followed
by removal of the N-benzyl protecting group and N-al-
kylation, afforded 5.
pyrazole-3-carboxylic acid 20 first with oxalyl chloride
and then with benzylamine dihydrochloride provided
the bicyclic pyrazolo[3,4-d]pyridazin-7-one derivative
21. Palladium-catalyzed biaryl cross-coupling gave
intermediate 22. Removal of the N-benzyl protecting
group in 22 and N-alkylation yielded 6.
2,6-Dihydro-pyrazolo[3,4-d]pyridazin-7-one 6 was pre-
pared from diethylacetylene dicarboxylate (17) as shown
in Scheme 4. Reaction of 17 with 2-chlorophenylhydr-
azine afforded the ethyl ester 18, which was formylated
using DMF in phosphorus oxychloride to give 19.
Hydrolysis of the ester group in 19 and reaction of the
5-Amino-4,6-dichloropyrimidine (24) was used to pre-
pare 1,9-dihydro-purin-6-one 7 (Scheme 5). Displace-
ment of a chloro group in 24 with 4-chloro-aniline,
followed by acylation of the resulting diamino derivative
with 2-chlorobenzoyl chloride, gave intermediate 25.
Scheme 4. Reagents and conditions: (a) 2-Cl-PhNHNH2, K2CO3, EtOH (87%); (b) POCl3, DMF (41%); (c) LiOHÆH2O, 2:1 MeOH/H2O (100%);
(d) ClCOCOCl, DMF; (e) PhCH2NHNH2Æ2HCl, NEt3, CH2Cl2 (74% from 20); (f) 4-Cl-PhB(OH)2, Pd(P(Ph3)4, aqueous K2CO3, toluene/EtOH,
110 °C (97%); (g) AlCl3, PhNO2, 80 °C (90%); (h) CF3CH2I, NaH, DMF, rt (23%).
Scheme 5. Reagents and conditions: (a) 4-Cl-PhNH2, concd HCl, H2O, EtOH, 19 h at 82 °C, 60 h at rt (95%); (b) 2-Cl-PhCOCl, CH3CONMe2, 0 °C
to rt (100%); (c) H2SO4, i-PrOH, 8 h at reflux, 16 h at rt (90%); (d) CF3CH2I, Cs2CO3, DMF, 100 °C (50%).

734
P. A. Carpino et al. / Bioorg. Med. Chem. Lett. 16 (2006) 731–736
The synthesis of 7 was completed from 25 by a three-
step, two-pot procedure that involved cyclocondensa-
tion, hydrolysis, and N-alkylation.
Binding affinities of bicyclic derivatives 1–7 were deter-
mined with standard competitive binding assays in
HEK-293 cells transfected with the human CB1-R
(hCB1-R) using [3H]-SR141716 (Amersham) as
a
radioligand and in CHO-K1 cells overexpressing human
CB2-R (hCB2-R) using [3H]-5-(1,1-dimethyl-heptyl)-2-
[5-hydroxy-2-(3-hydroxy-propyl)-cyclohexyl]-phenol as
a radioligand. Functional activities of 1–7 were deter-
mined in GTP-c-35S binding assays using membranes
from CHO-K1 cells stably transfected with hCB1-R
cDNA.
Figure 2. Superimposition of minimized structures 1 and 2b.14
Conformationally constrained analogs 2a,b with bicyclic
PP rings were slightly less active in the hCB1-R binding
assay than SR141716 (Table 1). The two PP analogs,
which have only hydrogen bond accepting capability,
exhibited surprisingly high affinities despite the poor
overlap of both the core 2,6-dihydro-pyrazolo[4,3-d]pyr-
imidin-7-one rings and the R2 side chains with the corre-
sponding pyrazole-3-carboxamide and 1-piperidinyl
groups in the low-energy conformer of SR141716 (see
Fig. 2). These results suggest that the PP analogs may
bind differently than SR141716 in hCB1-R, with the
non-aryl R1 side chains in 2a,b occupying a separate
hydrophobic pocket, distinct from that for the 1-pipe-
ridinyl group in SR141716 and the hydrogen bond-
accepting groups (i.e., oxygen atoms in the C7 carbonyl
moiety) forming different networks of hydrogen bond
interactions. It is also possible the rigid PP compounds
induce conformational changes in hCB1-R that allow
the receptor to accommodate the R1 side chains in the
same hydrophobic pocket occupied by the 1-piperidinyl
group in SR141716 (ꢁinduced fitꢀ model of binding). The
PP derivatives 2a,b contain alternative spatial arrange-
ments of the key pharmacophoric groups in SR141716
that do not result in any significant loss of binding
activity.
Structure–activity relationships around 2a were investi-
gated. The isomeric dichloro analog 3a was less potent
than 2a. Trichloro compounds with the identical substi-
tuent pattern on the diaryl groups in SR141716 were not
prepared. Replacement of the N6 cyclohexyl group with
branched alkyl groups had no effect on activity (2c,d).
Introduction of small alkyl (i.e., ethyl and n-propyl),
fluorinated alkyl, or benzyl groups increased affinity
Table 1. Binding and functional activities of compounds 1–7
Compound no.
R1
R2
hCB1-R binding Ki (nM)a,b
hCB2-R binding Ki (nM)a
GTP-c-35S binding Ki (nM)a
e
1
—
H
—
c-Hex
2.1 1.6c
1
e
2a
2b
2c
2d
2e
2f
2g
2h
2i
20 10c
>10,000
>10,000
H
1-Piperidinyl
CH(CH2CH3)2
CH(CH3)2
CH2CH3
CH2CH2CH3
CH2Ph
12 8c
20
e
e
H
26
21
e
e
H
H
4
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
4
H
2
2
H
5
4
H
CH2CF3
CH2CF2CH3
CH2CH2CF3
CH2CHF2
CH2CH2F
c-Hex
0.3 0.1d
0.2
0.1
0.1
4
H
2j
H
1.8 0.6c
0.6
2k
2l
H
1
H
3
38
5
e
3a
3b
4a
4b
4c
5
H
e
H
CH2CF3
CH2CF3
CH2CF3
CH2CF3
—
28
0.6 0.1c
Me
Et
i-Pr
—
—
—
1
3
1
6
1
8
1
13
1
6
—
—
4
7
1
a See Ref. 15 for details.
b Data are mean of one independent determination performed in triplicate, except as noted.
c Data are means of two independent determinations performed in triplicate.
d Data are means of three independent determinations performed in triplicate.
e Not measured.

P. A. Carpino et al. / Bioorg. Med. Chem. Lett. 16 (2006) 731–736
735
(2e–g). Compounds 2h,i with 20,20,20-trifluoroethyl and
Table 2. Pharmacokinetic properties of 2h and 4a
Compound 2h Compound 4a
45 32
9.2 7.7
2.6
62
0.54
10
223
20,20-difluoropropyl groups were 10-fold more potent
than the corresponding ethyl and n-propyl alkyl deriva-
tives 2e and f and 60-fold more potent than 2a, the
benchmark compound. The position of the fluoro substi-
tution on the N6 alkyl chain was important since no in-
crease in affinity was observed when the n-propyl group
in 2f was replaced with a 30,30,30-trifluoropropyl group as
in 2j. Removal of one or two fluorine atoms in 2h led to a
decrease in affinity (2k,l). The fluorine atoms in 2h,i, and
k may favorably modulate the electronic character of the
PP ring or act as weak hydrogen bond-accepting groups
in hCB1-R. No increase in binding affinity was observed
when the N6 cyclohexyl group in 3a was replaced with a
20,20,20-trifluoroethyl group (3b).
Rat PKa
Cl (ml/min/kg)
Vd (l/kg)
Half life, t1/2 (h)
F (%)
4.7
7
Dog PKb Cl (ml/min/kg)
10
Vd (l/kg)
t1/2 (h)
17
39
a See Ref. 18 for details.
b See Ref. 19 for details.
1.5 time point following a dose of 1 mg/kg po = 7.24;
B/P ratio for 4a at the 1.5 h time point following a
3 mg/kg po dose = 1.92).
Substitution at the C5 position in PP analog 2h was
examined. Introduction of a C5 methyl group did not af-
fect binding affinity (4a). However, incorporation of
larger alkyl groups (Et and i-Pr) resulted in a small loss
of activity (4b,c).
Pharmacokinetic (PK) properties of 2h and 4a were
measured in rats and dogs (Table 2). Amorphous solids
were used in intravenous formulations, crystalline solids
for oral delivery. For both compounds, plasma clearanc-
es (Cl) were moderate in the rat, but low in the dog. Vol-
umes of distribution (Vd) were high, attributed primarily
to the presence of lipophilic 20,20,20-trifluoroethyl
groups. Bioavailability (F) in the rat was moderate for
2h, but very low for 4a. Good plasma exposure of the
amorphous form of 4a was observed at the 1.5 h time
point in the food intake studies (i.e., 209 ng/mL at
3 mg/kg po), suggesting the poor bioavailability of the
crystalline form was due to solubility-limited absorp-
tion, rather than low intestinal permeability. Introduc-
tion of a methyl substituent at the C5 position in 2h
resulted in minimal effects on most PK properties, the
only exception being bioavailability.
The core bicyclic PP ring in 2h was replaced with other
heterocyclic scaffolds. The 2,6-dihydro-pyrazolo[3,4-
c]pyridin-7-one 5 and the 2,6-dihydro-pyrazolo[3,4-
d]pyridazin-7-one 6 were, respectively, 3- and 26-fold
less potent in the binding assay than 2h. The weakly ba-
sic derivative, 1,9-dihydro-purin-6-one 7, was equipo-
tent to 5. The decreases in potency between these new
heterocyclic analogs and 2h were attributed to differenc-
es in the electronic characteristics and lipophilic proper-
ties of the bicyclic rings.
Rigid PP analogs 2b,e–l and 4a–c and heterocyclic deriv-
atives 5–7 were shown to be potent antagonists in the
GTP-c-35S assay, with Ki values less than 20 nM (Table
1). As with binding affinity, the presence of a hydrogen
bond donor group was not critical for functional activ-
ity. All compounds inhibited basal GTP-c-35S binding,
thus demonstrating inverse agonism activity (data not
shown). Compounds 2a,b,e–l, 4a–c, and 5–7 were highly
selective for hCB1-R over hCB2-R.
In conclusion, conformationally constrained derivatives
of SR141716 were prepared and evaluated as hCB1-R
antagonists and inverse agonists. Two 2,6-dihydro-
pyrazolo[4,3-d]pyrimidin-7-ones 2a,b exhibited high
binding affinities despite showing poor overlap with
SR141716. These compounds may bind differently than
SR141716 in hCB1-R or may induce conformational
changes in the receptor that then allow binding. Analogs
of 2a showed good anorectic activities in a fasting-in-
duced re-feeding model in rats following oral adminis-
tration. Novel compounds containing other bicyclic
ring systems will be the subject of future publications.
Analogs 2h and 4a were selected for further evaluation
in animal models. These two compounds were first as-
sessed for their ability to reverse CB1-R agonist-stimu-
lated analgesia and hypothermia in mice, two of the
four assays that constitute the ꢁtetradꢀ.16 In these chal-
lenge assays, activity is influenced by a combination of
factors such as in vitro binding/functional potency, plas-
ma protein binding, plasma half-life, and brain penetra-
tion. Both rigid compounds were able to fully reverse
CP-55940-stimulated responses at 1 mg/kg sc. Com-
pounds 2h and 4a were also examined in a fasting-in-
duced re-feeding model in Sprague–Dawley rats to
assess CB1-R mediated anorectic activity. Both com-
pounds decreased food intake over a 1.5 h time period
in a dose-dependent manner (2h, ꢀ23%, ꢀ32%, and
ꢀ78% change relative to vehicle-treated control animals
at 0.3, 1, and 3 mg/kg po; 4a, ꢀ35%, ꢀ42%, and ꢀ59%
change relative to vehicle-treated control animals at 1, 3,
10 mg/kg po).17 These PP analogs showed good brain
penetration (brain to plasma or B/P ratio for 2h at the
Acknowledgments
The authors thank Marlys Hammond and Rich Hank
for optimizing the synthesis of compound 11a and Zach-
ary S. Hendsch for the overlap of 1 and 2b.
References and notes
1. De Petrocellis, L.; Cascio, M. G.; Di Marzo, V. Br. J.
Pharmacol. 2004, 141, 765.
2. Wilson, R.; Nicoll, R. A. Science 2002, 296, 678.
3. Herkenham, M. In Cannabinoid Receptors; Pertwee, R.
G., Ed.; Academic Press: New York, 1995; pp 145–166.

736
P. A. Carpino et al. / Bioorg. Med. Chem. Lett. 16 (2006) 731–736
4. Devane, W. A.; Dysarz, F. A., III; Johnson, M. R.; Melvin,
L. S.; Howlett, A. C. Mol. Pharmacol. 1988, 34, 605.
14. 3D structures were generated with Concord and mini-
mized in Sybyl with the MMFF94s forcefield.
´
5. Rinaldi-Carmona, M.; Barth, F.; Heaulme, M.; Shire, D.;
15. Pertwee, R. G. Curr. Med. Chem. 1999, 6, 635.
16. Smith, P. B.; Compton, D. R.; Welch, S. P.; Razdan, R.
K.; Mechoulam, R.; Martin, B. R. J. Pharmacol. Exp.
Ther. 1994, 270, 219.
17. Overnight fasted male CD rats were dosed the following
morning with suspensions of compounds 2h and 4a in
0.5% methylcellulose. Food was returned 0.5 h after
dosing, and cumulative food intake was assessed 1.5 h
later.
18. Male Sprague–Dawley rats were dosed intravenously with
2h and 4a via a jugular vein cannula (1 mg/kg). Compound
2h was formulated in 20% EtOH/80% CD, compound 4a
in in Transcutol P/CD. For oral administration, com-
pounds 2h and 4a were formulated in 0.5% methylcellu-
lose. Plasma concentrations were determined by LC–MS/
MS following protein precipitation.
Calandra, B.; Congy, C.; Martinez, S.; Maruani, J.;
Neliat, G.; Caput, D.; Ferrara, P.; Soubrie, P.; Breliere,
J. C.; Le Fur, G. FEBS Lett. 1994, 350, 240.
6. Smith, R. A.; Fathi, Z. Idrugs 2005, 8, 53.
7. Lange, J. H. M.; Kruse, C. G. Curr. Opin. Drug Disc.
2004, 7, 498.
8. Harrold, J. A.; Williams, G. Br. J. Nutr. 2003, 90, 729.
9. Van Gaal, L. F.; Rissanen, A. M.; Scheen, A. J.; Ziegler,
O.; Roessner, S. Lancet 2005, 365, 1389.
10. Stoit, A. R.; Lange, J. H. M.; den Hartog, A. P.; Ronken,
E.; Tipker, K.; Van Stuivenberg, H. H.; Dijksman, J. A.
R.; Wals, H. C.; Kruse, C. G. Chem. Pharm. Bull. 2002,
50, 1109.
11. All compounds prepared in this work gave satisfactory
spectral data (1H NMR and MS).
12. Pero, R. W.; Babiarz-Tracy, P.; Fondy, T. P. J. Med.
Chem. 1977, 20, 644.
13. Katoch-Rouse, R.; Pavlova, O. A.; Caulder, T.; Hoffman,
A. F.; Mukhin, A. G.; Horti, A. G. J. Med. Chem. 2003,
46, 642.
19. Male beagle dogs were dosed intravenously with 2h and 4a
(0.5 mg/kg iv). Compound 2h was formulated in 6.7%
EtOH/60% CD/33.3% PEG 400, compound 4a in
Transcutol P. Plasma concentrations were determined by
LC–MS/MS following protein precipitation.