Y. Iwata et al. / Bioorg. Med. Chem. Lett. 25 (2015) 236–240
237
intermediate 7. Removal of the Boc protecting group and subse-
quent functionalization of the nitrogen atom using standard proce-
dures delivered the various analogues shown in Table 1.
BocN
n
SH
Br
n
N
N
m
Br
NO2
F
6
e, f
a, b, c, d
In the first cycle of the SAR study, the substituents at the 1- and
2-positions of benzimidazole were held constant to cyclopropylm-
ethyl and neopentyl, respectively (Table 1). Interestingly, slight
changes in the ring size and the length of a linker had a significant
impact on CB2 agonistic activity (compounds 12 vs 20, 13 vs 21, 14
vs 22, 15 vs 23, 12 vs 16 and 13 vs 17). Azetidine was found to be
the more favored cyclic amine at this position for in vitro potency
(compounds 12–15). And in the piperidine subseries, keeping the
piperidine basic resulted in lower potency compared to com-
pounds where the nitrogen was part of an amide or urea moiety
(compounds 16, 17, 24, 25 vs 18, 19, 26, 27).
4
5
Boc
N
n
O
O
RN
n
O
O
S
m
S
m
N
N
N
N
n
n
g
7
8:R
= H
h
9:R = Me
10:R
i
j
= Ac
11:R = CONH2
We next explored the SAR at the 1-position of benzimidazole
with the 5-position held constant to an azetidinosulfonyl group.
The in vitro agonistic activities on the CB2 and CB1 receptors and
HLM clearance of these analogues are shown in Table 2. While lipo-
philic groups such as cyclopropylmethyl or trifluoromethoxyethyl
afforded highly potent compounds 13, 15, 28 and 33, HLM stability
of these analogues was decreased. On the other hand, introduction
of a polar group was effective for improving HLM stability as dem-
onstrated by compounds 29–32 and 34–37. The dimethylamino-
ethyl group exhibited the best balance between potency,
selectivity and metabolic stability (35).
With these findings, we proceeded to further optimize the sub-
stituent on the nitrogen atom of the azetidine ring (Table 3).
Although methyl (30) as an R group demonstrated a good profile
with respect to agonistic activity on CB2 receptor, HLM stability
and membrane permeability as measured in parallel artificial
membrane permeability assay (PAMPA), the compound resulted
in some activation of the CB1 receptor. R groups such as ethyl
(38), trifluroethyl (39) and methanesulfonyl (42) led to decreased
Scheme 1. Reagents and conditions: (a) cyclopropylmethylamine (1.2 equiv), Et3N
(1.2 equiv), EtOH, reflux; (b) Fe (8 equiv), 2 N HCl (5 equiv), reflux; (c) 3,3-
dimethylbutanoyl chloride (1 equiv), Et3N (1.5 equiv), CH2Cl2, 0 °C; (d) 5 N HCl
(3 equiv), MeCN, reflux, 69% from 4; (e)
6 (1.3 equiv), Pd2dba3 (0.025 equiv),
xantphos (0.05 equiv), Et3N (2 equiv), 1,4-dioxane, microwave irradiation, 170 °C,
1 h, 66–99%; (f) 30% H2O2, Na2WO4, MeOH, 0 °C–rt; (g) trimethylsilyl chloride
(4 equiv), MeOH, reflux, 72%-quant; (h) (CH2O)n (3 equiv), NaBH(OAc)3 (12 equiv),
AcOH (6 equiv), MeOH, rt; (i) AcCl (1.3 equiv), Et3N (3.6 equiv), CH2Cl2, rt; (j)
trimethylsilyl isocyanate (2 equiv), CH2Cl2, rt.
potency. Azetidine and piperidine were selected as the cyclic
amine moieties.
The syntheses of target compounds were conducted as outlined
in Scheme 1.8 Benzimidazole intermediate 5 was prepared accord-
ing to a conventional reaction sequence via addition of cyclopro-
pylmethylamine to the fluoronitrobenzene 4, reduction of the
nitro group, acylation, then cyclization. Cross coupling reaction
using a palladium catalyst was performed under microwave
irradiation, followed by oxidation with sodium tungstate to give
Table 1
In vitro agonistic activity of compounds 12–27 toward human CB2 receptor
O
R
O
S
N
N
Compound
R
hCB2
EC50 (nM)
Compound
R
hCB2
EC50 (nM)
a
a
HN
Me
12
13
14
8.9
2.0
9.6
20
21
22
104
32
HN
Me
N
N
N
Ac
73
O
H2N
HN
N
H2N
HN
Me
N
15
1.2
23
13
O
16
17
62
37
24
25
278
209
N
N
Me
Ac
Ac
N
18
6.2
26
7.7
N
O
H2N
N
H2N
N
19
6.3
27
0.99
O
EC50 values based on inhibition of forskolin-stimulated cAMP production in CHO cells expressing CB2 receptor.4
a