Synthesis and SAR of 5,6-diarylpyridines as human CB1 inverse agonists

Article abstract of DOI:10.1016/j.bmcl.2004.11.031

Structure-activity relationship studies for two series of 2-benzyloxy-5-(4-chlorophenyl)-6-(2,4-dichlorophenyl)pyridines having either a 3-cyano or 3-carboxamide moiety resulted in the preparation of the 2-(3,4-difluorobenzyloxy)-3-nitrile analog 10d and

Full text of DOI:10.1016/j.bmcl.2004.11.031

Bioorganic & Medicinal Chemistry Letters 15 (2005) 645–651
Synthesis and SAR of 5,6-diarylpyridines as
human CB1 inverse agonists
Laura C. Meurer,a,* Paul E. Finke,a Sander G. Mills,a Thomas F. Walsh,a
Richard B. Toupence,a Mark T. Goulet,a Junying Wang,a Xinchun Tong,a Tung M. Fong,b
Julie Lao,b Marie-Therese Schaeffer,b Jing Chen,b Chun-Pyn Shen,b D. Sloan Stribling,c
Lauren P. Shearman,c Alison M. Strackc and Lex H. T. Van der Ploegb
aDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
bDepartment of Metabolic Disorders, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
cDepartment of Pharmacology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 8 October 2004; revised 10 November 2004; accepted 15 November 2004
Available online 2 December 2004
Abstract—Structure–activity relationship studies for two series of 2-benzyloxy-5-(4-chlorophenyl)-6-(2,4-dichlorophenyl)pyridines
having either a 3-cyano or 3-carboxamide moiety resulted in the preparation of the 2-(3,4-difluorobenzyloxy)-3-nitrile analog 10d
and the 2-(3,4-difluorobenzyloxy)-3-(N-propylcarboxamide) analog 16c, (hCB1 IC50 = 1.3 and 1.7 nM, respectively) as potent
and selective hCB1 inverse agonists. Their synthesis and biological activities are described herein.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
O
CN
Me
N
O
N
H
Preparations of Cannabis sativa, such as marijuana and
hashish, have been used for medicinal and recreational
purposes for centuries. Recently, the cannabinoid D9-
tetrahydrocannabinol (D9-THC) has received renewed
interest as a potential treatment for analgesia, nausea,
and appetite stimulation.1 Several related cannabinoid
compounds can serve as ligands for and activate the
G-protein coupled cannabinoid receptor type 1 (CB1),
located predominantly in the central nervous system
(CNS) but also in the peripheral nervous system, and/
or type 2 (CB2), expressed mostly by immune cells.
CB1 agonists, such as D9-THC, are known to stimulate
food intake, whereas CB1 receptor knockout mice are
lean and resistant to diet-induced obesity.2 Initially,
the search for CB1 antagonists/inverse agonists was
based on the structure of known agonists such as D9-
THC.1,3 The first potent and selective hCB1 inverse ago-
nist 1 (SR141716, Fig. 1) was reported in 1994 by
researchers at Sanofi-Synthelabo and belonged to a
new family of CB1 ligands based on a 1,5-diphenylpyr-
N
N
N
Cl
Cl
H3CO
N
Cl
1 (SR141716)
hCB1 IC50 = 6 nM
2
hCB1 IC50 = 530 nM
Figure 1. Structure of SR141716 and the Merck lead structure.
azole structure.3,4 In a 2-week rat feeding study, 1 was
found to cause an initial decrease in food intake and
then a sustained slower weight gain over the 14-day dos-
ing period.5 In addition, a more recent study showed
that 1 produced a marked and sustained reduction of
adiposity in diet-induced obese mice, which was more
than expected from the reduction of food intake.6 These
studies suggested that CB1 and its endogenous ligands
modulate energy balance via a dual mechanism of intake
reduction and increased energy expenditure. The CB1
agonist WIN-55212-2 at 1 lM was found to increase
lipoprotein lipase activity of mouse primary adipocytes,
which could be blocked by 1, although it was not clear
Keywords: Human CB1 receptor; Inverse agonist; Pyridine.
*
Corresponding author. Tel.: +1 732 594 4291; fax: +1 732 594
5790; e-mail: laura_meurer@merck.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.11.031

646
L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 645–651
whether this was mediated by the CB1 receptor.7 CB1
receptors are coupled to the Gi/o class of G-proteins,
which inhibit adenylate cyclase activity.8 CB1 agonists
such as CP-559409 and WIN-55,212-210 decrease for-
skolin-induced cAMP accumulation in tissue from
rodent and mammalian brain in a concentration-depen-
dent manner. However, 1 produced a dose-dependent
increase in forskolin-induced cAMP,11 thus suggesting
inverse agonism of the CB1 receptor.12
a 2- or 3-amido functionality as is present in 1 were also
prepared as described herein.
2. Chemistry16
The preparation of 3a and 3b employed the analogous
chemistry that had provided the original lead 2 (Scheme
1). Thus, commercially available benzyl 4-chlorophenyl
ketone (5a) or 4-chlorobenzyl 2,4-dichlorophenyl ket-
one17 (5b) were reacted with DMF dimethyl acetal to
afford 6a,b. Treatment with cyanoacetamide and sodium
hydride in DMF effected ring closure to the 5,6-diaryl-
3-cyanopyridin-2-ones 7a,b. Alkylation with benzyl
bromide in the presence of cesium carbonate afforded
predominantly the O-benzyl pyridines 3a (1H NMR
(CDCl3): d 5.64 (s, 2H, O–CH2Ph)) and 3b along with
the N-benzyl pyridin-2-ones 8a (1H NMR (CDCl3): d
5.23 (s, 2H, N–CH2Ph)), and 8b, which were readily sep-
arated by silica gel chromatography (ratio of 3 to
8 = 1.5–3:1).16 The methyl ether 9a was obtained from
7a via alkylation with methyl iodide. Based on the initial
binding results for 3a and 3b, the Sanofi type trichloro
series was selected for further SAR studies. The effect
of substitution on the benzyl ether was first explored
as indicated in Scheme 1 (10a–f; X1–2 = Cl, R = substi-
tuted benzyl) and listed in Table 1. Subsequent to these
benzyloxy substitution studies, several alternative 5- and
6-phenyl substitution patterns were prepared in the 3,4-
difluorobenzyloxy series (see Table 1, 10g–k) starting
with the appropriate benzyl phenyl ketone in Scheme 1.
Our interest in discovering novel anti-obesity agents
based on the CB1 hypothesis was initially pursued with
a high throughput screening of the Merck sample collec-
tion to identify potential leads for an hCB1 antagonist
program. The 5,6-diarylpyridine 2 (Fig. 1) was found
to have moderate affinity for the hCB1 receptor
(IC50 = 530 nM) and became a lead structure in this pro-
gram. Based on reported structure–activity relationship
(SAR)13 and modeling14 studies of 1, the pyridine ring
of 2 was viewed as being a possible alternative scaffold
to the pyrazole of the Sanofi compound with the pyri-
dine nitrogen being equivalent to N-2 of the pyrazole.
Other interesting features of 2 relative to the structure
of 1 were the possible overlap of the lipophilic benzyloxy
group with the N-piperidine amide and the role of the
3-cyano as a surrogate for the proposed amide carbonyl
interaction at the binding site. The moderate affinity of
2 was attributed to the inappropriate substitution on
the phenyls.13
Our initial efforts examined the more closely related 6-
(4-chlorophenyl) substituted pyridine 3a (Fig. 2). The
surprisingly poor binding on the hCB1 receptor (3a,
IC50 = 2800 nM)15 with this substitution suggested that
the chlorination pattern on the two phenyl moieties
would be critical. Therefore, the substitution pattern
corresponding to 1 having a 5-(4-chlorophenyl) and a
6-(2,4-dichlorophenyl) was subsequently examined as
in 3b. Interestingly, a greater than 200-fold increase in
hCB1 binding was observed (IC50 = 11 nM), confirming
this rationale. A subsequent functional activity analysis
of 3b indicated that it was acting as an inverse agonist
having an EC50 = 70 nM with a 52% increase in
cAMP.15 The role of the nitrile was also investigated
by replacement with an amide as shown in structure 4
and an extensive investigation of 5-(4-chlorophenyl)-6-
(2,4-dichlorophenyl)pyridines in both series was under-
taken. In addition, two des-benzyloxy series having just
The effect of converting the 3-cyano to an amide moiety
as in 1 was then investigated in the trichloro series
(Scheme 2). Hydrolysis of 7b in sulfuric acid provided
the acid 11, which was subsequently converted to
the methyl ester 12 by treatment with HCl/MeOH.
Alkylation of 12 with benzyl bromide and cesium
(H3C)2N
CN
O
X1
O
O
X2
b
a
NH2
+
X2
X1
Cl
5a X1 = X2 = H
5b X1 = X2 = Cl
6a X1 = X2 = H
6b X1 = X2 = Cl
Cl
CN
CN
CN
O
O
O
NH
X2
R
+
N
N
c,d or e
X1
R
X2
X2
R1
X1
X1
N
O
N
R2
CN
O
Cl
O
Cl
Cl
8a X1 = X2 = H,
R = Bn
8b X1 = X2 = Cl,
R = Bn
7a X1 = X2 = H
7b X1 = X2 = Cl
3a X1 = X2 = H, R = Bn
3b X1 = X2 = Cl, R = Bn
9a X1 = X2 = H, R = Me
10a-f, X1 = X2 = Cl,
R = BnX4
N
Cl
X2
Cl
X1
See Table 1 for BnX4
Cl
Cl
Scheme 1. Reagents and conditions: (a) Me2NCH(OMe)2, DMF,
75 °C (100% crude); (b) MeOH, NaH, DMF (60–75% from 5); (c)
BnBr, Cs2CO3, DMF (50–65% 3, 18–35% 8); (d) MeI, Cs2CO3, DMF
(30%); (e) X4BnBr or X4BnI, Cs2CO3, DMF (40–50% 10a–f).
4
3a X1 = X2 = H; hCB1 IC50 = 2800 nM
3b X1 = X2 = Cl: hCB1 IC50 = 11 nM
Figure 2. Basic structures of two initial SAR series. See Tables 1 and 2.

L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 645–651
647
Table 1. Structures and hCB1 activities for 2-benzyloxy-3-cyano
derivatives 3 and 10
R'-O
O
O
X4
a, b
c, d or e
f
NH
Cl
CN
7b
O
Cl
N
11 R' = H
12 R' = Me
X2
Cl
X1
R'-O
O
N
MeO
O
O
N
X3
X2 X3 X4
O
Compound X1
No.
hCB1
IC50 (nM)a
SEMb
R
+
1
3a
6.2
2800
11
Cl
Cl
for R' = Me,
Cl
Cl
H
H
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
H
340
2
13a R = Bn
3b
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
F
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
H
4-F
13b R = 3,4-diFBn
13c R = n-Bu
14
10a
10b
10c
10d
10e
10f
10g
10h
10i
10j
10k
3.1
1.8
2.7
1.3
5.8
3.4
18
1.1c
0.8
0.9c
0.8c
2.2
0.5
4c
Cl
Cl
13d R = CH2(cyclohexyl)
15 R' = H, R = Bn
3,5-diF
2,4-diF
3,4-diF
3-Cl
g, h
or i
R2
N
O
4-CF3
3,4-diF
3,4-diF
3,4-diF
3,4-diF
3,4-diF
R1
4a-g, R = Bn
16a-e, R = 3,4-diFBn
17a-c, R = n-Bu
O
R
Cl
Cl
8
3
Cl
Cl
F
7.2
6.3
66
1.0c
1.0
18c
N
17d-e, R = CH2(cyclohexyl)
Me Cl
F
Cl
See Table 2 for
R1 and R2
Cl
F
a See Ref. 15 for assay details.
b n = 2, except as noted.
c n = 4.
Cl
Scheme 2. Reagents and conditions: (a) aq H2SO4, 140 °C; (b) HCl,
MeOH (69% from 7b); (c) BnBr, Cs2CO3, DMF (47% 13a, 48% 14); (d)
3,4-diFBnBr, Cs2CO3, DMF (40% 13b); (e) RI or RBr, Cs2CO3, DMF
(67% 13c; 29% 13d); (f) NaOH, MeOH (75–85%); (g) oxalyl chloride,
DMF (cat), CH2Cl2; (h) HNR1R2, Et3N, CH2Cl2 (55–90%); (i)
HNR1R2, Py–BOP, DIPEA, CH2Cl2 (40–85%).
carbonate afforded equal amounts of the O-benzyl 13
and the N-benzyl 14 derivatives.16 Hydrolysis of the
methyl ester 13 gave the acid 15. Reaction with oxalyl
chloride followed by addition of various amines or simul-
taneously with Py–BOP activation provided the amides
4a–g (Table 2). As a result of the potent affinity obtained
with the 3,4-difluorobenzyl ether 10d (Table 1), a hybrid
series of simple alkyl amides 16a–e was examined as listed
in Table 2. Also, an investigation of replacements for the
benzyl ether was initiated with the preparation of a series
of 2-alkoxyl analogs 17a–e (Table 2) via alkylation of 12
and subsequent conversion to the amides as above.
Based on analogy to the Sanofi pyrazole and our pyri-
dine scaffold, it was also of interest to prepare a series
of 2-amido-5-(4-chlorophenyl)-6-(2,4-dichlorophenyl)
pyridines (29a–i, Table 4) to see if the 2-substitution im-
proved binding compared to the 3-amido series 20a–d.18
In an independent synthesis (Scheme 4), 4-chlorocin-
namic acid (21) was reduced to the alcohol 22 via forma-
tion of the mixed anhydride and subsequent reduction
with sodium borohydride.19 Swern oxidation of 22 affor-
ded the aldehyde 23, which was treated with ethyl azido
acetate 24 to provide the ethyl 2-azido-5-phenylpenta-
2,4-dienoate 25.20 Conversion to the pyridine ring was
accomplished by stirring 25 overnight with triphen-
ylphosphene to provide phosphazene 26, which was sub-
sequently reacted with 2,4-dichlorobenzaldehyde (27) at
60 °C to provide the desired 2-substituted pyridine scaf-
fold as the methyl ester intermediate 28.21 Utilizing pre-
viously described chemistry, a series of amides 29a–i
were similarly prepared (Table 4).
In an attempt to shorten the synthesis of the amides in
Scheme 2, the hydrolysis of the cyano group of 3b was
attempted to prepare the acid intermediate 15. However,
the vigorous conditions needed to hydrolyze the cyano
group also lead to cleavage of the benzyl moiety to give
back the prior acid 11. However, it was discovered that
reaction of 11 with oxalyl chloride provided the interme-
diate 2-chloropyridine acid chloride 18a, which on sub-
sequent treatment with several amines afforded a series
of 2-chloro-3-amidopyridines 19a–d (Scheme 3, Table
3). It was then envisioned that selective hydrogenation
of 18b might afford a synthesis of simpler amide analogs
more closely related to 1. Thus, methyl ester 12 was trea-
ted with oxalyl chloride to give the 2-chloropyridine 18b.
Mild hydrogenation at atmospheric pressure afforded a
26% yield of desired des-chloro 18c along with unre-
acted starting material and a mixture of dechlorination
by-products. Fortunately, 18c was readily isolated by
silica gel chromatography and was subsequently con-
verted as in Scheme 2 to the amides 20a–d (Table 3).
3. Results and discussion
Binding affinities on recombinant human CB1 receptor
expressed in Chinese Hamster Ovary (CHO) cells for
compounds in Tables 1–4 were determined with a stan-
dard binding assay using [3H]CP-55940 as the radioli-
gand.15 Interesting compounds were further evaluated

648
L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 645–651
Table 2. Structures and hCB1 activities of 2-benzyloxy- and 2-alkoxy-3-ester derivatives 13 and 3-amido derivatives 4, 16, and 17
R1
R'-O
O
N
N
O
N
R2
O-R
Cl
O-R
Cl
Cl
Cl
Cl
Cl
13
4, 16, 17
Compound No.
R
13 R0
4, 16, 17 NR1R2
hCB1 IC50 (nM)a
SEMb
4a
4b
Bn
Bn
––
––
––
––
––
––
––
Me
Me
Me
Me
––
––
––
––
––
––
––
––
––
––
NH2
18
4
NHMe
NMe2
NHEt
7.0
22
2c
4c
4d
Bn
Bn
5c
1.4
6
0.5
1.3
4
4e
Bn
Bn
NH(n-Pr)
NH(cyclopentyl)
NH(piperidin-1-yl)
––
4f
4g
18
41
Bn
Bn
18
13a
13b
13c
13d
16a
16b
16c
16d
16e
17a
17b
17c
17d
17e
3.5
1.3
1.9
2.9
3.1
1.9
1.7
1.5
1.8
32
1.5
0.2c
0.9
0.1
1.0
1.0
0.4c
0.1
0.2
16
3,4-diFBn
n-butyl
––
––
CH2(cyclohexyl)
3,4-diFBn
3,4-diFBn
3,4-diFBn
3,4-diFBn
3,4-diFBn
n-butyl
––
NHMe
NHEt
NH(n-Pr)
NH(CH2)2F
NH(i-Pr)
NH2
n-butyl
n-butyl
NHMe
NH(n-Pr)
NHMe
NH(n-Pr)
17
21
7
2
CH2(cyclohexyl)
CH2(cyclohexyl)
3.4
5.2
1.5
1.1
a See Ref. 15 for assay details.
b n = 2, except as noted.
c n = 4.
O
NR1R2
R4
O
R3
N
Table 3. Structures and hCB1 activities of 3-amido derivatives 19 and 20
b
a
R4
Cl
R1
3b
11
N
c or
O
N
N
R2
b, d
e,b,c
Cl
R4
Cl
12
Cl
Cl
Cl
19a-d R4 = Cl
Cl
Cl
18a R3 = R4 = Cl
20a-d
R4 = H
18b R3 = OMe, R4 = Cl
18c R3 = OMe, R4 = H
See Table 3
Cl
for R1 and R2
Compound
No.
R4
R3
hCB1
IC50 (nM)a
SEMb
Scheme 3. Reagents and conditions: (a) 50% H2SO4, 140 °C (quant.);
(b) oxalyl chloride, CH2Cl2 (82% for 18b); (c) HNR1R2, Et3N, CH2Cl2
(52% 19a from 3b); (d) 5% Pd/C, H2 (26%); (e) NaOH, MeOH (90%).
19a
19b
19c
19d
20a
20b
20c
20d
Cl
Cl
Cl
Cl
H
NMe2
NH(n-Pr)
NH(n-hexyl)
1300
52
400
4
58
6
NH(piperidin-1-yl)
piperidin-1-yl
NH(n-hexyl)
400
460
65
250
50
6
in a functional assay also using recombinant human
CB1 receptors expressed in CHO cells.15 Binding affini-
ties were also routinely measured for the CB2 receptor
expressed in CHO cells and using [3H]WIN-55,212-2
as radioligand.15,22 While the initial result with the 6-
(4-chlorophenyl) derivative 3a was disappointing
(IC50 = 2800 nM), the 200-fold improvement in the
binding affinity of 3b to an IC50 = 11 nM validated the
hypothesis that a pyridine could replace the pyrazole
of 1. In addition, 3b was found to be >200-fold selective
H
H
NH(cyclohexyl)
NH(piperidin-1-yl)
210
480
80
170
H
a See Ref. 15 for assay details.
b n = 2.
for hCB1 over hCB2 (IC50 > 2000 nM (34%)) in the
binding assay. Also encouraging was the finding that

L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 645–651
649
Table 4. Structures and hCB1 activities of 2-amido derivatives 29
The affinity of 3b was further explored with a benzyl
substitution SAR study (Table 1). The 4-fluoro and 4-
trifluoromethyl derivatives 10a and 10f and the 3-chloro
analog 10e showed improved binding (IC50 = 3.1, 3.4,
and 5.8 nM). 3,5-Disubstitution as in the 3,5-difluoro
10b (IC50 = 1.8 nM) was also beneficial while the 2-sub-
stitution of the 2,4-difluoro 10c (IC50 = 2.7 nM) did not
afford any apparent improvement over the 4-fluoro 10a.
The optimum substitution appeared to be obtained with
the 3,4-difluoro compound 10d (IC50 = 1.3 nM), which
also had potent inverse agonist functional activity
(EC50 = 8 3 nM with an average ꢀ110% maximal re-
sponse, n = 4) and was now 400-fold selective for
hCB1 over hCB2 (IC50 = 1.3 vs 520 nM) in the binding
assays.
O
R1
N
N
R2
Cl
Cl
Cl
Compound No. –NR1R2
hCB1 IC50 (nM)a SEMb
29a
29b
29c
29d
29e
29f
29g
29h
29i
piperidin-1-yl
340
43
13
11
11
13c
38
10
18
40
2
NH(n-hexyl)
NH(4-heptyl)
NH(cyclopentyl)
NH(cyclohexyl)
NH(cycloheptyl)
NH(piperidin-1-yl)
NHPh
32
53
95
35
56
As already seen with 3a, the 4-chloro substitution on the
5-phenyl was critical and even replacement of the 5-(4-
chlorophenyl) with a 4-fluoro- or 4-methylphenyl as in
10i and 10j reduced affinity 5-fold (IC50 = 7.2 and
6.3 nM). The effect of substitution on the 6-phenyl ring
was also examined in the 3,4-difluorobenzyloxy series
with the preparation of the 6-(4-chlorophenyl) 10g and
6-(2-chlorophenyl) 10h analogs which, as expected,13a
afforded reduced binding compared to 10d (IC50 = 18,
8.0, and 1.3 nM, respectively). Since the trifluoro analog
10k resulted in a 50-fold loss of affinity (IC50 = 66 nM),
the trichloro substitution pattern of 1 was determined to
be optimal for binding in our pyridine series.
210
31
NHBn
a See Ref. 15 for assay details.
b n = 2.
c n = 3.
O
OH
a
b
OH
Cl
Cl
21
22
O
O
O
c
Replacement of the 3-cyano of 3b with an amide moiety
as in 1 was then investigated, first in the unsubstituted
benzyl ether series with 4a–g and subsequently, after
the discovery of 10d, in the 3,4-difluorobenzyl series
16a–e (Table 2). The SAR indicated that a secondary
amide was preferred as seen with the series of primary
4a, secondary 4b, and tertiary 4c amides (IC50 = 18, 7,
and 22 nM, respectively). The observation that the
methyl esters 13a–d were equipotent or even better than
the corresponding NHMe amides 4b, 16a, 17b, and 17d,
respectively, implied that there is not a direct N–H inter-
action, only steric and/or lipophilic effects of the N-alkyl
moiety. These effects were apparently optimum with the
N-ethyl derivative 4d (hCB1 IC50 = 1.4 nM; functional
activity EC50 = 21 2 nM with an average ꢀ60% ago-
nist response, n = 3), while n-propyl 4e or larger again
gave poorer binding. Incorporating the piperidine moi-
ety of 1 into 4g or the comparably sized cyclopentyl of
4f also reduced affinity (IC50 = 18 and 41 nM) implying
that the benzyloxy was in fact occupying the same area
of the receptor where the piperidine of 1 binds. Analogs
with the potency enhancing 3,4-difluorobenzyl were sub-
sequently prepared (16a–e, Table 2). While some
improvement was observed for 16a and 16c compared
to 4b and 4e, this change did not afford an additional
10-fold improvement as had been obtained in the 3-cya-
no series (16a–c; IC50 = 3.1, 1.9, and 1.7 nM). Both 16b
and 16c were shown to be inverse agonists
(EC50 = 13 2 and 50 20 nM with an average ꢀ60
and ꢀ73% maximal response, respectively, n = 2) and
were both >400-fold selective for hCB1 (hCB2
IC50 = 820 and 815 nM) in the binding assays. Addition
of a fluorine as in 16d or branching of the alkyl as in 16e,
H
OEt
e
+
OEt
24
N3
Cl
Cl
+
N3
23
25
O
Cl
O
d
OEt
N=PPh3
H
Cl
26
27
Cl
O
O
OMe
NR1R2
N
N
f,g,h
Cl
Cl
Cl
Cl
28
Cl
29a-i
Cl
Scheme 4. Reagents and conditions: (a) ClCO2Et, Et3N, THF, then
NaBH4, H2O (79%); (b) oxalyl chloride, DMSO, then DIPEA, CH2Cl2
(88%); (c) NaOCH3, NaOH (100%); (d) PPh3, Et2O, 0 °C; (e) CH3CN,
60 °C (22% from 23); (f) NaOH, MeOH (96%); (g) oxalyl chloride,
DMF (cat), CH2Cl2; (h) HNR1R2, Et3N, CH2Cl2 (35–75%).
the hCB1 functional assay indicated both 3a (up to
ꢀ67% @ 3000 nM) and 3b (ꢀ52% maximal response
with an EC50 = 28 20 nM, n = 3) were interacting with
the hCB1 receptor as inverse agonists. The diminished
binding for the methyl ether 9a was also in agreement
with the benzyl taking the place of the piperidine amide.
The N-benzyl pyridin-2-one by-products 8a and 8b
(Scheme 1, IC50 = 6000 and 2000 nM, respectively) were
relatively inactive, as were all other N-substituted ana-
logs in both the 3-cyano and 3-amido series and all inter-
mediates, and will not be further mentioned.

650
L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 645–651
did not further improve binding (IC50 = 1.5 and
1.8 nM). In the hope of improving solubility, the 2-benz-
yloxy group was also replaced with two alkoxy moieties
as shown in Table 2. Interestingly, the smaller n-butyl
group (17a–c) gave only moderately diminished results
and the isosteric cyclohexylmethyl showed comparable
or better results than the benzyl compound (17b,d, and
4b; IC50 = 17, 3.4, and 7 nM). This implied that the
receptor interaction in this area was simply binding to
a lipophilic pocket without a direct aromatic interaction.
reduction in food intake at any of the doses as seen with
1. However, by 18 h post-dosing, the 10 mg/kg dose
afforded a cumulative, but non-significant (p > 0.05),
22% food intake reduction and a dose-dependent weight
loss of 1 g compared to an 8 g gain for vehicle treated
animals (p < 0.05). These results suggested that the slow
CNS exposure, as evident from the B/P ratio experi-
ment, prevented an immediate food intake effect. A simi-
lar feeding study with the slightly less potent 16c
afforded a cumulative 16% reduction in food intake at
10 mg/kg (p > 0.05), but did not give a net weight loss
(5 g gain vs an 11 g gain for vehicle). The results with
Removal of the 2-benzyloxy moiety was studied in the
2-chloro (19a–d, Table 3) and des-chloro-3-amido
(20a–d, Table 3) series and resulted in overall decreased
affinity. The 3-propyl and 3-hexyl amides 19b,c, and 20b
were moderately active (IC50 = 52, 58, and 65 nM,
respectively); however, the N-piperidinyl amides 19d
and 20d and the isosteric cyclohexyl amide 20c
(IC50 = 400, 480, and 210 nM, respectively) were sur-
prisingly inactive compared to 1 and the benzyloxy
derivatives discussed above. The secondary amides 19a
and 20a were also comparatively inactive.
16c were consistent with its poorer PK profile (AUCnom
=
5.2 lM h kg/mg; Clp = 5.8 mL/min/kg, Vdss = 0.4 L/kg;
t1/2 = 2.5 h), lower oral absorption (F = 12%), and
slower brain penetration (B/P ratio = 0.01–0.06 at
0.25–4 h).
5. Conclusions
The finding that both the 2- and 3-amido pyridine ana-
logs were only moderately potent hCB1 receptor inverse
agonists was disappointing based on the similarity to the
Sanofi compound 1. This might be a result of the addi-
tional ring atom slightly changing the relative orienta-
tion of the three pyridine substituents relative to the
proposed nitrogen interaction, and/or the pyridine nitro-
gen did not have an equivalently effective receptor inter-
action as the pyrazole nitrogen due to electronic
differences, and/or the effect of the neighboring methyl
of 1. However, the excellent binding affinity of the
six-membered ring 3-cyano-2-(3,4-difluorobenzyloxy)
pyridine derivative 10d and the equivalent 3-(N-propyl-
carboxamido) derivative 16c demonstrated the possibili-
ty that the amide moiety of the five-membered pyrazole
1 could be split into a lipophilic portion and a polar
functionality. This finding might allow more flexibility
in selection of derivatives in the future with enhanced
PK and CNS exposure profiles and improved physical
and off-target properties than could be achieved with a
single amide functionality. Further SAR studies will be
published elsewhere in the near future.
The poor binding of the 3-amido compounds led to the
investigation of the 2-amido series where somewhat
more encouraging results were obtained (Table 4). The
N-piperidinyl amide 29g analogous to 1 was found to
be modestly active (IC50 = 56 nM). The corresponding
cyclohexyl derivative 29e, as well as the cyclopentyl
29d and cycloheptyl 29f, were about the same, as were
the noncyclic n-hexyl 29b and 4-heptyl 29c and the benz-
yl amide 29i. The secondary piperidine amide 29a and
aniline derivative 29h had very weak binding. Although
these amides were not particularly potent, their overall
functional data appeared to indicate the desired inverse
agonist activity at high concentrations (data not shown).
There are several possible reasons for the relatively poor
binding of these 2- and 3-amido pyridine derivatives
compared to 1. The divergent affinity might be related
to the ring size giving a slightly different relative orienta-
tion of the three pyridine substituents, and/or the basic-
ity difference of the nitrogen, and/or possibly the effect
of a lack of a neighboring methyl, which would force
the amide moiety of 1 to be perpendicular to the pyraz-
ole ring, although in the 3-carboxamide series the neigh-
boring chloro in 19d did not have an effect compared to
20d.
References and notes
1. For general reviews of cannabinoid receptor ligands and
potential therapeutic applications, see: (a) Di Carlo, G.;
Izzo, A. A. Expert Opin. Invest. Drugs 2003, 12, 39; (b)
Pertwee, R. G. Expert Opin. Invest. Drugs 2000, 9, 1553;
(c) Xiang, J.-N.; Lee, J. C. Ann. Rep. Med. Chem. 1999, 34,
199.
2. (a) Trillou, C. R.; Delgorge, C.; Menet, C.; Arnone, M.;
Soubrie, P. Int. J. Obesity 2004, 28, 640; (b) Zimmer, A.;
Zimmer, A. M.; Hohmann, A. G.; Herkenham, M.;
Bonner, T. I. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
5780.
3. Barth, F.; Rinaldi-Carmona, M. Curr. Med. Chem. 1999,
6, 745.
4. Rinaldi-Carmona, M.; Barth, F.; Heaulme, M.; Shire, D.;
Calandra, B.; Congy, C.; Martinez, S.; Maruani, J.;
Neliat, G.; Caput, D.; Ferrarra, P.; Soubrie, P.; Breliere,
J. C.; Le Fur, G. FEBS Lett. 1994, 350, 240.
4. In vivo studies
Based on their favorable hCB1 binding and functional
assay results, 10d and 16c were selected for further in
vivo evaluation. Preliminary pharmacokinetic (PK)
studies in male Sprague–Dawley rats (1 mg/kg i.v.,
2 mg/kg p.o.) with 10d indicated good i.v. pharmacoki-
netic properties (AUCnom = 9.3 lM h kg/mg; Clp = 3.6
mL/min/kg; Vdss = 0.8 L/kg; t1/2 = 3.6 h) and moderate
oral absorption (F = 27%), but slow brain penetration
and a low brain-to-plasma ratio (B/P ratio = 0.03–0.26
at 0.25–4 h).23 In a food intake and body weight loss
study using diet-induced obese rats fed ad libitum, 10d
at 1, 3, and 10 mg/kg p.o. did not show an immediate

L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 645–651
651
5. Colombo, G.; Agabio, R.; Diaz, G.; Lobina, C.; Reali, R.;
Gessa, G. L. Life Sci. 1998, 8, 113.
6. Trillou, C. R.; Arnone, C.; Gonalons, N.; Keane, P.;
Maffrand, J. P.; Soubrie, P. J. Physiol. Regul. Integr.
Comp. Physiol. 2003, 284, 345.
7. Cota, D.; Marsicano, G.; Tschop, M.; Grubler, Y.;
Flachskamm, C.; Schubert, M.; Auer, D.; Yassouridis,
A.; Thone-Reineke, C.; Ortmann, S.; Tomassoni, F.;
Cervino, C.; Nisoli, E.; Linthorst, A. C. E.; Pasquali, R.;
Lutz, B.; Stalla, G. K.; Pagotto, U. J. Clin. Invest. 2003,
112, 423.
within 50% of the average value based on a standard
sample. For the assay conditions for both the binding and
functional assays, see: Felder, C. C.; Joyce, K. E.; Briley,
E. M.; Mansouri, J.; Mackie, K.; Blond, O.; Lai, Y.;
Ma, A. L.; Mitchell, R. L. Mol. Pharmacol. 1995, 48,
443.
16. 1H NMR and/or LC–MS data were in agreement with the
assigned structures. For full experimental and spectral
details, see: Finke, P. E.; Meurer, L. C.; Debenham, J. S.;
Toupence, R. B.; Walsh, T. F. WO Patent 03/082191 A2,
2003; Chem. Abstr. 2003, 139, 307686f.
8. Pertwee, R. G. Pharmacol. Ther. 1997, 74, 129.
9. Koe, B. K.; Milne, G. M.; Weissman, A.; Johnson, M. R.;
Melvin, L. S. Eur. J. Pharmacol. 1985, 109, 201.
10. DÕAmbra, T. E.; Estep, K. G.; Bell, M. R.; Eissenstat, M.
R.; Josef, K. A.; Ward, S. J.; Haycock, D. A.; Baizman, E.
R.; Casiano, F. M.; Beglin, N. C. J. Med. Chem. 1992, 35,
124.
17. Newman, M. S.; Reid, D. F. J. Org. Chem. 1958, 23,
665.
18. An EPO patent from Sanofi-Synthelabo related to this
same 2-amido pyridine structure has recently been pub-
lished, see: Barth, F.; Martinez, S.; Rinaldi-Carmona, M.
WO Patent 03/084930 A1, 2003; Chem. Abstr. 2003, 139,
307689j.
11. Mato, S.; Pazos, A.; Valdizan, E. M. Eur. J. Pharmacol.
2002, 443, 43.
19. Ishizumi, K.; Koga, K.; Yamada, S. Chem. Pharm. Bull.
1968, 16, 492.
12. (a) Landsman, R. S.; Burkey, T. H.; Consroe, P.; Roeske,
W. R.; Yamamura, H. I. Eur. J. Pharmacol. 1997, 334R1–
334R2; (b) Bouaboula, M.; Perrachon, S.; Milligan, L.;
Canat, X.; Rinaldi-Carmona, M.; Portier, M.; Barth, F.;
Calandra, B.; Pecceu, F.; Lupker, J.; Maffrand, J. P.; Le
Fur, G.; Casellas, P. J. Biol. Chem. 1997, 272, 22330.
13. (a) Lan, R.; Liu, Q.; Fan, P.; Lin, S.; Fernando, S. R.;
McCallion, D.; Pertwee, R.; Makriyannis, A. J. Med.
Chem. 1999, 42, 769; (b) Wiley, J. L.; Jefferson, R. G.;
Grier, M. C.; Mahadevan, A.; Razdan, R. K.; Martin, B.
R. J. Pharm. Exp. Ther. 2001, 296, 1013.
20. Boukou-Poba, J. P.; Farnier, M.; Guilard, R. Tetrahedron
Lett. 1979, 19, 1717.
21. Barluenga, J.; Ferrero, M.; Palacios, F. J. Chem. Soc.,
Perkin Trans. 1 1990, 2193.
22. CB2 binding affinities were routinely measured for all
compounds15 and, generally, were less sensitive to these
SAR changes than were the CB1 affinities. Higher CB2 to
CB1 IC50 ratios were observed for the more active
benzyloxy compounds of Tables 1 and 2 (50–500) than
for the less active amides of Table 3 (0.2–10) and Table 4
(5–30).
14. Shim, J.-Y.; Welsh, W. J.; Cartier, E.; Edwards, J. L.;
Howlett, A. C. J. Med. Chem. 2002, 45, 1447.
15. The reported binding results ( SEM) were an average of
2–4 independent determinations, which were normally
23. Ritschel, W. A.; Kearns, G. L. In Handbook of Basic
Pharmacokinetics Including Clinical Applications; Grau-
bart, J. I., Ed., 5th ed.; American Pharmaceutical Asso-
ciation, 1999.