5-Sulfonyl-benzimidazoles as selective CB2 agonists

Article abstract of DOI:10.1016/j.bmcl.2008.03.048

A novel series of benzimidazole CB2-receptor agonists was synthesized and the structure-activity relationship explored. The results showed agonistic activities with an EC₅₀ up to 0.5 nM and excellent selectivity (>4000-fold) over the CB1 receptor. The size of the substituent on the 2-position determined the level of agonism, ranging from inverse agonism to partial agonism to full agonism, which was more pronounced for the rat CB2 receptor. A wide variation of sulfonyl substituents at the benzimidazole 5-position was tolerated, which was used to optimize the drug-like properties. This resulted into lead compound 14j that can be used to investigate the potential of a selective, peripherically acting CB2 agonist. The in vitro profile of key compounds is displayed using pie bar charts (VlaaiVis).

Full text of DOI:10.1016/j.bmcl.2008.03.048

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2574–2579
5-Sulfonyl-benzimidazoles as selective CB2 agonists
Bie M. P. Verbist,a Michel A. J. De Cleyn,a Michel Surkyn,a Erwin Fraiponts,a
Jeroen Aerssens,b Marjoleen J. M. A. Nijsena and Harrie J. M. Gijsena,*
aJohnson & Johnson Pharmaceutical Research & Development, Turnhoutseweg 30, 2340 Beerse, Belgium
bTibotec BVBA, Generaal de Wittelaan L 11B3, 2800 Mechelen, Belgium
Received 9 February 2008; revised 14 March 2008; accepted 16 March 2008
Available online 20 March 2008
Abstract—A novel series of benzimidazole CB2-receptor agonists was synthesized and the structure–activity relationship explored.
The results showed agonistic activities with an EC50 up to 0.5 nM and excellent selectivity (>4000-fold) over the CB1 receptor. The
size of the substituent on the 2-position determined the level of agonism, ranging from inverse agonism to partial agonism to full
agonism, which was more pronounced for the rat CB2 receptor. A wide variation of sulfonyl substituents at the benzimidazole 5-
position was tolerated, which was used to optimize the drug-like properties. This resulted into lead compound 14j that can be used to
investigate the potential of a selective, peripherically acting CB2 agonist. The in vitro profile of key compounds is displayed using pie
bar charts (VlaaiVis).
Ó 2008 Elsevier Ltd. All rights reserved.
The endocannabinoid system has been known to medi-
ate a complex array of biological effects. These effects
are regulated through at least two distinct G-protein
coupled receptors, the CB1- and CB2-receptor.1 While
the effects mediated by CB1, mostly in the CNS, have
been thoroughly investigated, those mediated by CB2
are not well defined. CB2 selective agonists are claimed
to be effective in the treatment of pain,2 various inflam-
matory diseases,3 osteoporosis4, and atherosclerosis.5
Separation between the therapeutically undesirable psy-
chotropic effects mostly associated with the CB1 recep-
tor, and the clinically desirable ones are considered to
be essential for a cannabinoid agonist to be widely ac-
cepted as a useful drug. Two strategies can be envisaged
to avoid possible CB1-mediated central side effects of
cannabinoid receptor agonists. One could aim for com-
pounds without central penetration, only exerting their
pharmacological effects peripherically, or aim for selec-
tive CB2-compounds. Since cannabinoid ligands, in gen-
eral, are rather lipophilic molecules, the complete
avoidance of central penetration is not a straightforward
task. Most of the described CB2 selective ligands possess
a substantial potency for the CB1 receptor when tested
in our protocols, which can contribute to their analgesic
action.6 The discrepancies between the reported potency
values can be explained by differences in methodologies
used as pointed out by Huffman with WIN-55,212-2.7
Even combination experiments of CB2 agonist ligands
and CB1 antagonists cannot always mask or suppress
the CB1 mediated activity of the ligands due to the
incompatibility of their pharmacokinetic properties
and/or the degree of selectivity of the used antagonists.
Therefore, in order to investigate the true potential of a
CB2 agonist, there is a need for drug-like CB2 ligands
with an increased selectivity over the CB1 receptor.
A high throughput screening assay was initiated for the
development of novel selective CB2 agonists. The benz-
imidazole derivative 1 was obtained as an initial hit
compound. The pharmacological data are presented in
a pie bar chart (VlaaiVis)8 (Fig. 1 and Table 1).
The pie bar chart displays the profile of a compound and
allows for an easy comparison of compounds. In this
type of visualization, the preferred profile (Table 1) is
represented as the grey outer rim and determines the
properties of the dataset. Each property is represented
by a colored wedge in the circle, and the width of these
wedges hints at the relative importance of the property.
An increase in coloring of the wedge corresponds to an
improved value for the respective parameter: the fuller
Keyword: Cannabinoid receptor CB2 agonist benzimidazole VlaaiVis
species difference.
*
Corresponding author. Tel.: +32 14 60 6830; fax: +32 14 60 5344;
e-mail: hgijsen@prdbe.jnj.com
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.03.048

B. M. P. Verbist et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2574–2579
2575
matic substitution making use of R1NH2, mostly
tetrahydro-4-pyranylmethylamine, to give 5. Reduction
of the nitro group to aniline 6 was followed by acylation
to introduce substituent R2. The desired benzimidazoles
7 were obtained via final condensation in hot acetic acid.
This synthesis route was ideal for the exploration of the
SAR around R1 and R2 since they were introduced at a
late stage of the synthesis.
H
N
Cl
N
S
O
Cl
1
A synthetic route where the R5-substituent was intro-
duced in one of the last steps started from 5-chloro-2-ni-
Figure 1. Pharmacological data: EC50 CB2 = 25 nM; ratio EC50 CB2/
CB1 = 354.
tro-aniline
8
(Scheme 2). 4-Methoxybenzylthiol
(PMBSH) was introduced as a protecting group via
nucleophilic aromatic substitution and subsequent reac-
tion with pivaloylchloride gave 9. Reduction of the nitro
group followed by reductive amination with tetrahydro-
pyran-4-carbaldehyde introduced the most favorable
R1-substituent. Condensation to the benzimidazole un-
der acidic conditions and subsequent deprotection of
the para-methoxybenzyl group with TFA gave interme-
diate 11. In the next step, the various R5 groups could be
introduced. Alkyl substituents were introduced via
nucleophilic substitution making use of R5-Br or R5-
Cl. (Hetero)aromatic substituents were introduced via
Pd-catalyzed reactions.10 Finally, the sulfides were oxi-
dized to sulfones 12–14 using meta-chloroperoxybenzoic
acid.
Table 1. Preferred profile as used in the pie bar chart representation
Properties
Preferred
profile
Bluea
hCB2-agonism-EC50
Ratio EC50 CB2/CB1
<1 nM
>1000
P100%
<50%
hCB2-agonism-% efficacy
CYP-450 1A2/2C19/2C9/2D6
3A4 BFC, BQ and DBF
inhibition9
Green
%Metabolism in human
liver microsomes after 15 min
Solubility at pH 7.4 and pH 4
Rule of five violation
<50%
>20 lM
61
Yellow
a Start from the first blue value and read clockwise.
The first exploration of the SAR around 1 focused on
the variation of the substituents at the benzimidazole
2- and 5-positions, but did not lead to any improvement
in potency or selectivity for the CB2 receptor.
the better. The blue value represents the in vitro phar-
macological activity of a compound. The green wedges
contain the ADME data and the yellow bar corresponds
to in silico calculations.
´
A recent publication by Page et al. on benzimidazole
derivatives as selective CB2 inverse agonists illustrated
the importance of a substituent on the benzimidazole-
N1-nitrogen.11 In their series, a cyclohexylmethyl or a
tetrahydro-4-pyranylmethyl substituent at this position
was well tolerated. When implemented to our initial
set of compounds, we realized their added value as dem-
onstrated in Table 2.
Several different intermediates were synthesized to en-
able variation at particular sites of the molecule. For
the variation of the right-hand side of the molecule syn-
thesis began with 4-chlorothiophenol 2 where the de-
sired R5-substituent was introduced via nucleophilic
substitution (Scheme 1). The resulting sulfide was oxi-
dized to sulfone 3 using meta-chloroperoxybenzoic acid.
Nitration under standard conditions afforded com-
pound 4, then R1 was introduced via a nucleophilic aro-
In general, the cyclohexylmethyl group (17) led to the
most potent compounds but the tetrahydro-4-pyranylm-
HNO3
H2SO4
1. Cs2CO3, acetone
ethylbromide
Cl
Cl
Cl
S
O
S
O O
NO2
HS
2.
m
-CPBA, CHCl3
100%
O
2
3
4
1. R2COCl
CH2Cl2, pyr
O
NH2
O
or R2COOH, DIPEA
HBTU, CH2Cl2
O
N
N
NH
R2
K2CO3
2. HOAc, Δ
S
O
S
O
R
diox or
i-PrOH
O
O
91%
7
5 R = NO2
6 R = NH2
H2, Pd/C
MeOH (69%)
Scheme 1.

2576
B. M. P. Verbist et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2574–2579
1. Fe, aq. HOAc, Δ
O
O
1. PMBSH
H
2.
NH
NH
EtOH, KOH
NO2
NH2
NO2
O
Δ
PMB
PMB
S
Cl
S
NH
O
2. PivCl
pyridine
CH2Cl2
HOAc, CH2Cl2
NaBH3CN
O
10
69%
48%
8
9
1a. R5Br, Cs2CO3,
acetone
O
O
or 1b. Pd2(dba)3, diox,
1. HOAc, Δ
Xantphos, Cs2CO3,
N
N
R5
N
2. TFA
70%
S
O
2.
m-CPBA,
N
HS
CHCl3
O
11
12-14
Scheme 2.
Table 2. Influence of benzimidazole-N1-nitrogen substituent and oxidation state of sulfur
R1
N
N
Y
Cl
S (a)
SO (b)
SO2 (c)
Y
OEt
R1
H
15
EC50: 2 lM
EC50: >10 lM
O
EC50: 4 nM
CB2/CB1: 75
%Metaba: 98%
EC50: 20 nM
CB2/CB1: 484
%Metab: 95%
EC50: 20 nM
CB2/CB1: 537
%Metab: 66%
16
17
EC50: 125 nM
CB2/CB1: 1.2
%Metab: 85%
EC50: 6.3 nM
CB2/CB1: 8.2
EC50: 5 nM
CB2/CB1: 50
%Metab: 74%
a %Metabolism as measured after 15 min incubation with human liver microsomes (hLM).
ethyl group (16) showed the best selectivity window with
only a slight decrease in potency. In Table 2, the influ-
ence of the oxidation state of the sulfur atom is also
O
N
demonstrated. Although not yet optimal, the sulfone
derivatives 16c–17c show the highest metabolic stability
as well as the best CB2/CB1 selectivity.
N
S
O
O
O
Removal of the chlorine atom at the 6-position led to
compound 7a (Fig. 2). With an improved potency, selec-
tivity, and stability, 7a became our new starting point
for the development of a highly selective CB2 agonist.
7a
Figure 2. Pharmacological data: EC50 CB2 = 4.0 nM; ratio EC50 CB2/
CB1 = 1597; metabolism_hLM = 34%.
As shown in the pie bar chart (Fig. 2), compound 7a ful-
fills most of the criteria as set in our preferred profile;
however it acts as a partial agonist as indicated with
the third blue wedge. The percentage of agonism efficacy
reached only 44%.
ported as inverse agonists.11 Another article describing
a 2-oxoquinoline series as CB2 receptor inverse agonists
uses similar para-alkoxybenzyl substituents to elicit a
full CB2 inverse agonist response.12 We decided to focus
A benzimidazole series with a 2,3-dihydrobenzo-fura-
nylmethyl substituent at the 2-position has been re-

B. M. P. Verbist et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2574–2579
Table 3. Influence of substituent in position 2 on partial agonisma
EC50 Human CB2 Rat CB2
2577
Table 4. 5-Alkylsulfonyl substitution
EC50 Ratio EC50 %Metabolism
X
(nM) agonism
%Eff
agonism
%Eff
O
(nM) CB2/CB1
in hLM
(15 min)
N
N
N
N
R2
R5
O O
R5
S
S
O
O
12
7
R2
a CH2C6H4OEt
b CH3 (X = CH2)
a CF3(CH2)3
b CF3(CH2)2
c CH3CH2
1
285
394
36
13
0
4.0
31
44
1.3
1.2
1.3
4.0
2.5
10
551
121
c CH(CH3)2 (X = CH2)
d C(CH3)3 (X = CH2)
e C(CH3)3
1.6
0.8
1.2
4.0
4.0
>104
101
104
120
101
109
99
d c-PropCH2
e CN(CH2)2
f HO(CH2)2
g H2N(CH2)2
h EtHN(CH2)3
i AcHN(CH2)3
14
24
6
2114
3846
912
102
43
0
f CH2CH(CH3)2
g c-pentyl
h CH2-benzo[1,3]dioxoleb
158
100
4.3
109
a X is equal to oxygen unless mentioned.
b 7h EC50 = 6.3 nM; human CB2 inverse agonism %Eff = 92.13
position. First, the influence of an alkylsulfonyl group
was investigated (Table 4) (12a–i).
on the 2-position to understand the partial agonist
nature of compound 7a (Table 3). The influence of the
substituent at the 2-position was retrieved from a set
of compounds 7a–h possessing a cyclohexylmethyl or a
tetrahydro-4-pyranylmethyl group in the benzimidazole
1-position.
There was no significant effect on the parameters by
changing the chain length as demonstrated with a
trifluoroalkyl group in 12a–b. Simple alkyl groups in
12a–d showed good activity but moderate selectivity.
Introduction of polar groups such as cyano- (12e) and
hydroxyl groups (12f) improved the selectivity but
diminished the potency toward the CB2 receptor. Amine
(12g, h) or amide substituents (12i) decreased the
potency considerably. Further SAR exploration pointed
toward benzylsulfonyl substituted compounds 13 (Table
5). This variation led to a favorable potency but suffered
severely from metabolic cleavage of the benzylic moiety
within 13a. Difluorination as in 13b could solve this, but
again selectivity remained an issue.
Increasing the bulkiness of the moiety at the 2-substitu-
ent resulted in increased CB2 potency as demonstrated
by comparing compounds 7b, c, and d. Any further in-
crease in bulkiness led to partial agonism as observed
for 7a or even inverse agonism as demonstrated by
7h.13 Further SAR studies around this template enabled
us to conclude that the effect of bulkiness on agonism is
even more pronounced for the rat CB2-receptor. Com-
pound 7f showed partial agonism in rat and 7g showed
no agonism at all, while both retained full agonism for
the human receptor. Similar differences between rat
and human CB2-receptors have been observed regard-
ing binding potencies.14 Any effort to increase the polar-
ity on this site of the molecule resulted in a dramatic
decrease of potency. Overall, the tert-butyl group had
the best balance between percentage effective agonism
and potency.
Further SAR studies were focused on aryl- and hetero-
arylsulfonyl substituents in 5-position (14a–l, Table 6).
Substitution on the aryl group improved the potency,
independent of the site of substitution (14b–d). In the
case of meta-substituents (14c) an excellent selectivity
was obtained. Unfortunately, ortho- and meta-substi-
tuted phenylsulfones displayed poor metabolic stability.
A decrease in activity was observed by introducing
polarity (14e–f) in line with the alkyl substituents,
though this was less pronounced. Heteroaryl substitu-
ents (14g–i) were mostly well tolerated. Pyridine 14g
The right-hand side of the molecule was now fully opti-
mized and rather lipophilic. Substituents at the 1-posi-
tion should fit a large hydrophobic pocket on the
receptor and only a hydrogen acceptor moiety such as
an ether functionality is tolerated. A tert-butyl group
at the 2-position seemed to be the most optimal substi-
tuent. Compound 7e emerged as the preferred candi-
date, but despite a reasonable CB2/CB1 EC50 ratio of
ꢀ500 it still induced the typical psychotropic effects
when dosed in rat, commonly referred to as the tetrad
test.15
Table 5. 5-Benzylsulfonyl substitution
EC50 Ratio EC50 %Metabolism
(nM) CB2/CB1
in hLM
(15 min)
N
N
R5
O
S
O
We aimed for a CB2/CB1 selectivity ratio of at least
1000 to avoid these psychotropic effects. In order to im-
prove the selectivity, we also explored the left-hand side
of the molecule. As learned from our initial series, there
was a certain degree of freedom for variation at the 5-
13
R5
a p-CN-Ph-CH2
b p-CN-Ph-CF2
0.5
1.6
436
62
73
28

2578
B. M. P. Verbist et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2574–2579
Table 6. 5-(Hetero)aryl substitutiona
EC50 Ratio EC50 %Metabolism
(nM) CB2/CB1
X
in hLM
(15 min)
O
O
N
N
R5
O
R5
N
N
N
S
S
O
O
O
14
14j
a Ph (X = CH2)
b 2,6-Cl-Ph
c 3-CN-Ph
1.2
0.4
0.5
0.6
5.0
1.6
1.2
0.8
0.4
0.3
0.8
1.3
14
292
1462
191
363
550
215
4
23
96
61
42
Figure 3. ADME-data: inhibition of Cyp-450 isoforms 2C19 = 76%
inhibition at 10À5 M and 2C9 = 94% inhibition at 10À5 M.
d 4-CN-Ph
e 4-H2NCH2-Ph
f 4-AcNMeCH2-Ph
g 4-Pyridine
50
28
25
Acknowledgment
h 2-thiazole
i 2-furan
We thank Dr. P. Stanislawski for the useful comments
during the preparation of this manuscript.
40
j 3-OEt-4-pyridine
k 3-F-4-pyridine
l 3-Me-4-pyridine
4266
495
1622
48
40
40
Supplementary data
a X is equal to oxygen unless mentioned.
The details of the functional assay and the synthesis of
14j are described. Supplementary data associated with
this article can be found, in the online version, at
doi:10.1016/j.bmcl.2008.03.048.
gave a reasonable potency but suffered from the known
problem of severe CYP-interactions.16 However, these
interactions could be reduced to acceptable levels by
the introduction of a pyridine ortho-substituent (14j–l),
and this also increased the selectivity for the CB2 recep-
tor in analogy with 14c. In addition, compared to meta-
substituted arylsulfone 14c, the metabolic stability of
ortho-substituted pyridines was improved to a reason-
able level which led to acceptable plasma concentrations
when dosed orally in vivo (Table 7).
References and notes
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. Guindon, J.; Hohmann, A. G. Br. J. Pharmacol. 2008,
153, 319.
Overall, 2-tert-butyl-5-(2-ethoxypyridin-4-sulfonyl)-1-
14j
3. Whiteside, G. T.; Lee, G. P.; Valenzo, K. J. Curr. Med.
Chem. 2007, 14, 917.
4. Ofek, O.; Karsak, M.; Lecler, N.; Fogel, M.; Frenkel, B.;
Wright, K.; Tam, J.; Attar-Namdar, M.; Kram, V.;
Shohami, E.; Mechoulam, R.; Zimmer, A.; Bab, I. Proc.
Natl. Acad. Sci. U.S.A. 2006, 103, 696.
tetrahydropyran-4-ylmethyl)-1H-benzimidazole
showed an optimal potency, selectivity and ADME-pro-
file as illustrated in Figure 3.
In vivo plasma kinetics study (Table 7) for 14j in rat
(10 mg/kg po) showed an absolute oral bioavailability
of 43% and a brain to plasma ratio of 0.1, indicating
poor blood brain barrier penetration. Compound 14j
did not show any CB1 mediated side effects in the tetrad
test (40 mg/kg po). This highly selective and almost
exclusive peripheral acting compound will be used to
investigate the true potential of a peripheric CB2 ago-
nist. Full in vivo characterization will be published in
pharmacological journals in due course.
5. Steffens, S.; Veillard, N. R.; Arnaud, C.; Pelli, G.; Burger,
F.; Staub, C.; Karsak, M.; Zimmer, A.; Frossard, J.-L.;
Mach, F. Nature 2005, 434, 782.
6. For a general overview of existing CB2 ligands we refer to
the following review Raito, K. H.; Salo, O. M. H.;
Nevalainen, T.; Poso, A.; Ja¨rvinen, T. Curr. Med. Chem.
2005, 12, 1217.
7. Huffman, J. W. Curr. Pharm. Des. 2000, 6, 1323.
8. (a) Agrafiotis, D. K.; Ales, S.; Dai, H.; Derkinderen, A.;
Farnum, M.; Gates, P.; Izrailev, S.; Edward, P.; Konstant,
P.; Leung, A.; Lobanov, V. S.; Marichal, P.; Martin, D.;
Rassokhin, D. N.; Shemanarev, M.; Skalkin, A.; Stong, J.;
Tabruyn, T.; Vermeiren, M.; Wan, J.; Xu, X. Y.; Yao, X.
J. Chem. Inf. Model. 2007, 47, 1999; (b) Howe, T. J.;
Mahieu, G.; Marichal, P.; Tabruyn, T.; Vugts, P. Drug
Discov. Today 2007, 12, 45. The software to create a pie
bar chart and a more detailed explanation is available as
freeware at www.vlaaivis.com.
Although compound 14j possesses a promising profile,
the solubility and the inhibition of CYP-450 isoforms
2C9 and 2C19 remain an issue, and further optimization
is currently ongoing.
9. Stresser, D. M.; Blanchard, A. P.; Turner, S. D.; Erve, J.
C. L.; Dandeneau, Z. Z.; Miller, V. P.; Crespi, C. L. Drug
Metab. Dispos. 2000, 28, 1440.
Table 7. Plasma kinetics after oral administration (10 mg/kg) of 14j in
male Sprague–Dawley rat
10. Itoh, T.; Mase, T. Org. Lett. 2004, 24, 4587.
´
11. Page, D.; Brochu, M.-C.; Yang, H.; Brown, W.; St.-Onge,
S.; Martin, E.; Salois, D. Lett. Drug Des. Discov. 2006, 3,
Cmax
Tmax
t1/2
AUCinf
(ng h/ml)
Fabs
(%)
(ng/ml)
(h)
(h)
14j
731 353
0.5
3.4 0.4
2501 706
61
298.

B. M. P. Verbist et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2574–2579
2579
12. Raitio, K. H.; Savinainen, J. R.; Vepsa¨la¨inen, J.; Laitinen,
J. T.; Poso, A.; Ja¨rvinen, T.; Nevalainen, T. J. Med. Chem.
2006, 49, 2022.
14. Mukherjee, S.; Adams, M.; Whiteaker, K.; Daza, A.;
Kage, K.; Cassar, S.; Meyer, M.; Yao, B. B. Eur. J.
Pharmacol. 2004, 505, 1.
13. The same signal transduction protocol is used as for the
agonism test, only the concentration of forskolin, which
activates the adenylate cyclase is 6.25 times less.
15. Pertwee, R. G. Br. J. Pharmacol. 1972, 46, 753.
16. Riley, R. J.; Parker, A. J.; Trigg, S.; Manners, C. N.
Pharm. Res. 2001, 5, 652.