Optimisation of a novel series of selective CNS penetrant CB2 agonists

Article abstract of DOI:10.1016/j.bmcl.2011.05.063

A series of benzimidazole CB₂ receptor agonists were prepared and their properties investigated. Optimisation of the three benzimidazole substituents led to the identification of compound 23, a potent CB₂ full agonist (EC₅₀

Full text of DOI:10.1016/j.bmcl.2011.05.063

Bioorganic & Medicinal Chemistry Letters 21 (2011) 4284–4287
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Optimisation of a novel series of selective CNS penetrant CB2 agonists
⇑
Christine Watson , Dafydd R. Owen, Denise Harding, Kana Kon-I, Mark L. Lewis, Helen J. Mason,
Miyako Matsumizu, Takasuke Mukaiyama, Margarita Rodriguez-Lens, Akiko Shima,
Mifune Takeuchi, Isabelle Tran, Tim Young
Pfizer Global Research and Development, Ramsgate Road, Sandwich, Kent CT13 9NJ, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of benzimidazole CB2 receptor agonists were prepared and their properties investigated. Optimi-
sation of the three benzimidazole substituents led to the identification of compound 23, a potent CB2 full
agonist (EC50 2.7 nM) with excellent selectivity over the CB1 receptor (>3000-fold). Compound 23 dem-
onstrated good CNS penetration in rat. Further optimisation led to the identification of compound 34 with
improved selectivity over hERG and excellent CNS penetration in rat.
Received 13 April 2011
Revised 17 May 2011
Accepted 18 May 2011
Available online 25 May 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
CB2 agonist
CNS penetration
Benzimidazole
CB1 selective
Cannabinoid receptors are involved in a broad range of pro-
cesses including appetite, anxiety, memory, cognition, immune
regulation and inflammation.1 In particular, the association of can-
nabinoids with pain pathology has long been recognised and in-
deed agonists of the CB1 receptor are analgesic.2 However, their
use is severely limited by the associated hypomobility, hypother-
mia and catalepsy.3 CB2 receptors were first reported to be located
only in the periphery, mainly in immune cells, but it has now been
established that they are also expressed in glial cells in the central
nervous system (CNS). Indeed, in pre-clinical models of neuro-
pathic pain, CB2 receptors are upregulated in the CNS.4 While there
is evidence that CB2 agonists are efficacious in animal models of
pain,5 recent reports have highlighted that some highly selective
CB2 agonists are not efficacious in rat models of analgesia.6
The debate about the effectiveness of selective CB2 agonists as
analgesics prompted us to report our efforts in this field. Our aim
was to identify a centrally penetrant, selective CB2 agonist as a po-
tential analgesic agent devoid of the side-effects associated with
the CB1 receptor.
hCB2 EC50 12 nM
O
S
O
hCB1 Ki >19.4 µM
HLM Clint <7.0 µL/min/mg
RLM Clint <11.2 µL/min/mg
TPSA 84, clogP 2.8
Bioavailability in rat 21%
[brain]u:[plasma]u <1%
N
N
HO
1
N
Figure 1. Structure and properties of project lead.
high (Cl 75 mL/min/kg) and the oral bioavailability was a modest
21%. In addition, the compound exhibited very poor CNS penetra-
tion—the ratio of unbound drug in the brain to unbound drug in
the plasma ([brain]u:[plasma]u) was <1%.9
Our expectation was that hydrogen bonding groups in 1, in par-
ticular the hydrogen bond donor alcohol, had limited its CNS pen-
etration. Thus we re-evaluated our structure activity relationship
(SAR) data, looking for starting points with the potential for im-
proved brain penetration which also retained good CB2 potency
and selectivity over the CB1 receptor. By removing polar function-
ality, we expected to improve CNS penetration but risked compro-
mising metabolic stability through increased lipophilicity.
Compound 2, Figure 2, had an encouraging profile. It showed
good CB2 potency and selectivity and the lower TPSA10 was ex-
pected to result in improved CNS penetration. As expected, the in-
creased lipophilicity (measured by c log P) had increased the rate
of clearance in vitro, hence we embarked on a programme of SAR
exploration aimed at addressing this.
Previous efforts in our laboratories had resulted in the discovery
of compound 1 (Fig. 1) as a potent CB2 agonist.7 Compound 1 and
all the compounds in this communication proved to be full ago-
nists in our CB2 functional assay8 unless otherwise stated. 1 was
also selective over CB1 (CB1 binding Ki >10 lM) and was metabol-
ically stable in human and rat liver microsomes (HLM and RLM,
respectively). However, the in vivo clearance in rat proved to be
⇑
Corresponding author. Tel.: +44 1304646390; fax: +44 1304651817.
E-mail address: christine.watson100@gmail.com (C. Watson).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.05.063

C. Watson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4284–4287
4285
Br
NO2
Br
Br
NO2
NH
Br
NH2
NH
hCB2 EC50 3.2 nM
O
S
O
a
b
N
N
hCB1 Ki >10.0 µM
HLM Clint 23 µL/min/mg
RLM Clint 144 µL/min/mg
TPSA 60, clogP 4.3
10
11
12
2
Br
R1 S
N
N
N
c
d
e
Figure 2. Structure and properties of lead compound 2.
N
13
14
A general route to these benzimidazoles is shown in Scheme 1.
The synthesis started with alkylation of the sulfide formed by
reduction of disulfide 3 followed by oxidation to give sulfone inter-
mediate 5. Elaboration to the diamine intermediate 8 generally
proceeded in good yield via nitration, chloride displacement and
nitro reduction. Benzimidazole ring formation was achieved by
reaction with the appropriate acid chloride then cyclisation using
p-toluenesulfonic acid.
Alternatively, late stage modification of the sulfone was enabled
using the route shown in Scheme 2. Starting from 1,4-dibromo-2-
nitrobenzene (10), introduction of the amine was followed by nitro
reduction and benzimidazole formation. A range of thiols could be
introduced by a palladium catalysed coupling.11 Oxidation to the
sulfone then delivered the final compound.
A range of substituents on the sulfone was prepared using these
routes and good CB2 potency could be readily achieved, as shown
in Table 1. Unfortunately, any gains in potency tended to be offset
by increased metabolic instability (as measured by HLM and RLM).
However, we were encouraged by the data for compound 16. This
appeared to be as potent as compound 2 but had lower c log P (4.0)
and an improved metabolic profile.
Optimisation of the C-2 position was enabled by library chem-
istry (Scheme 3) from diamine intermediate 21. Benzimidazole for-
mation was achieved directly in parallel from a range of acids using
either T3P12 or HATU to activate the acid, giving a synthesis success
rate of 54% for the library.
O
R1 S
O
N
N
9
Scheme 2. Reagents and conditions: (a) (aminomethyl)cyclopropane, EtOH, 80 °C,
18 h, 100%; (b) Fe, NH4Cl, EtOH, H2O, reflux, 4 h, 98%; (c) (i) tBuCH2COCl, DIPEA,
EtOAc, rt, 1 h; (ii) pTsOHÁH2O, reflux, 9 h, 49%; (d) R1SH, Pd2(dba)3, xantphos, 1,4-
dioxane, reflux, 21 h; (e) mCPBA, CH2Cl2, 0 °C, 1 h.
Table 1
Effect of C-5 substituent on hCB2 agonist activity
O
O
S
R1
N
N
Compound
R1
hCB2 EC50 nM
HLM Clint
RLM Clint
2
Isopropyl
t-Butyl
Ethyl
(CH3)2NCH2CH2–
Me2C(OH)CH2–
HOCH2–C(CH3)2–
PhCH2–
3.2
0.43
2.9
32
23
38
8.4
79
<7
144
ND
38
170
244
54
15
16
17
18
19
20
3.1
1.2
6.5
<7
276
369
The SAR for the C-2 position is shown in Table 2. Although a
large number of benzimidazoles with high structural diversity
could be synthesised in this way, the SAR for CB2 agonist activity
proved very tight. In addition, we found that extending the size
of R2 could impact on efficacy, for example, compound 31 had an
Emax of 58%,13 further limiting the scope for optimisation at this
position.
O
O
S
O
O
S
T3P, E t 3N, dioxane,
110oC, 16 h
NH2
O
N
R2
N
R2
+
NH
HO
or
HATU, Et 3N, CH2Cl2,
AcOH, 110oC, 16 h
21
22
In fact only tert-butyl retained similar potency to neopentyl
(compound 23 vs compound 16).
Scheme 3. Parallel modification of R2 substituent.
Encouragingly, 23 showed low levels of metabolism in HLM and
RLM. The physicochemical properties (c log P 3.5; TPSA 60) were
also expected to give improved CNS penetration relative to com-
pound 1.
Further profiling showed that 23 exhibited excellent selectivity
over the hCB1 receptor in binding and functional assays (Fig. 3). It
was therefore progressed to an in vivo pharmacokinetic (PK) study
in rat.
Rat PK (1 mg/kg i.v.) demonstrated that it was a low clearance
compound and oral dosing (2 mg/kg) led to excellent bioavailabil-
ity (F = 100%). Encouragingly, the CNS penetration in rat had also
improved significantly with a [brain]u:[plasma]u ratio of 30%.
Whilst compound 23 demonstrated that we could achieve our
desired potency, selectivity and PK profile, it suffered from signif-
O
R1 S
R1 S
S S
O
a
b
Cl
Cl
Cl
Cl
5
3
4
O
O
O
O
R1 S
R1 S
NO2
NO2
NH
c
d
e
Cl
6
7
O
R1 S
O
S
O
O
NH2
NH
N
R1
f
icant hERG ion channel activity (hERG patch clamp IC50 2.6 lM).
N
We now assessed the potential to optimise the cyclopropylm-
ethyl substituent (R3, Table 3) aiming to identify a compound with
a similar profile to 23 but reduced hERG affinity.
8
9
Scheme 1. Reagents and condition: (a) (i) NaBH4, EtOH, THF, 0 °C, 30 min; (ii) R1-I,
0 °C to rt, 3 h; (b) mCPBA, CH2Cl2, 0 °C to rt, 2 h; (c) KNO3, H2SO4, 80 °C, 2 h;
(d) (aminomethyl)cyclopropane, DIPEA, IPA, 75 °C, 18 h; (e) H2, 10% Pd/C, THF,
25 °C, 4 h; (f) (i) 3,3-dimethyl-butyryl chloride, EtOAc, rt, 1 h; (ii) p-TsOHÁH2O,
reflux, 3 h.
A broader range of substituents could be tolerated at R3 with
cyclic heterocycles, small alkyl groups and polar functionality all
showing promising activity. Compounds 32 and 34 looked promis-
ing—they both had good potency and low levels of metabolism in

4286
C. Watson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4284–4287
Table 2
Table 4
Effect of C-2 substituent on hCB2 agonist activity
hERG activity in the optimisation of compound 23
O
O
Compound
hCB2 EC50 nM
hCB1 EC50 nM
hERG IC50 nM
c log P
S
N
N
R2
23
32
34
2.7
11
5.3
>10000
15000
ND
2600
63000
6200
3.5
2.7
3.4
Compound
R2
hCB2 EC50 nM
HLM Clint
RLM Clint
16
23
2.9
2.7
8.4
9
38
12
HLM Clint 12 µl/min/mg
RLM Clint 17 µl/min/mg
O
S
O
N
N
Rat PK (1mg/kg i.v):
Cl 27.1 ml/min/kg
Vd 1.0 L/kg
24
25
132
199
ND
43
ND
34
O
32
T
1/20.4h
425a
ND
ND
[brain]u:[plasma]u 1.0
26
Figure 4. PK properties of compound 32.
27
28
82
25
8.0
11
33
106
by weaker CB2 activity, resulting in a similar selectivity window for
both compounds. Pleasingly, compound 32 was a potent CB2 ago-
nist and selective with respect to CB1 with significantly reduced
hERG activity compared to 23. The in vivo clearance in rat was sim-
ilar to compound 23 and the oral bioavailability was once again
100% ( Fig. 4). Encouragingly, the [brain]u/[plasma]u in rat was
1:1, that is, 32 was fully CNS penetrant.
29
30
51
8.0
42
11
9.6
106
31
67b
ND
ND
In summary, the benzimidazole scaffold enabled identification
of novel, CNS penetrant CB2 agonists with excellent selectivity over
the CB1 receptor. We have identified compounds with similar
potency to lead compound 2 but with an improved PK profile.
Compound 23 is a useful tool compound to investigate the
pharmacology of brain penetrant selective CB2 agonists. Further
optimisation led to the discovery of compound 32, a CNS penetrant
selective CB2 agonist with reduced hERG activity. Further studies
are underway to assess the in vivo pharmacology of this compound
in pain models.
a
Emax 79%.
Emax 58%.
b
hCB2 EC50 2.7 nM
hCB1 Ki >20.0 µM
hCB1 EC50 >10.0 µM
Rat PK (1mg/kg i.v.):
Cl 20 mL/min/kg
Vd 3.2 L/kg
O
S
O
N
N
23
T
1/21.9h
Acknowledgments
Figure 3. Structure and properties of compound 23.
We would like to acknowledge the Pfizer Nagoya project team
for their discovery of the benzimidazole sulfone scaffold and their
work in this series.
Table 3
Effect of N substituent on hCB2 agonist activity
O
O
References and notes
S
N
N
1. Grotenhermen, F. Curr. Drug Targets: CNS Neurol. Disord. 2005, 4, 507.
2. Pertwee, R. G. Prog. Neurobiol. 2001, 63, 569.
R3
3. Chaperon, F.; Thiebot, M.-H. Crit. Rev. Neurobiol. 1999, 13, 243.
4. (a) Beltramo, M.; Bernardini, N.; Bertorelli, R.; Campanella, M.; Nicolussi, E.;
Fredduzzi, S.; Reggiani, A. Eur. J. NeuroSci. 2006, 23, 1530; (b) Zhang, J.; Hoffert,
C.; Vu, H. K.; Groblewski, T.; Ahmad, S.; O’Donnell, D. Eur. J. NeuroSci. 2003, 17,
2750.
5. (a) Ibrahim, M. M.; Rude, M. L.; Stagg, N. J.; Mata, H. P.; Lai, J.; Vanderah, T. W.;
Porreca, F.; Buckley, N. E.; Makriyannis, A.; Malan, T. P., Jr. Pain 2006, 122, 36;
(b) Bingham, B.; Jones, P. G.; Uveges, A. J.; Kotnis, S.; Lu, P.; Smith, V. A.; Sun, S.
C.; Resnick, L.; Chlenov, M.; He, Y.; Strassle, B. W.; Cummons, T. A.; Piesla, M. J.;
Harrison, J. E.; Whiteside, G. T.; Kennedy, J. D. Br. J. Pharmacol. 2007, 151, 1061;
(c) Beltramo, M. Mini-Rev. Med. Chem. 2009, 9, 11.
6. (a) Trotter, B. W.; Nanda, K. K.; Burgey, C. S.; Potteiger, C. M.; Deng, J. Z.; Green,
A. I.; Hartnett, J. C.; Kett, N. R.; Wu, Z.; Henze, D. A.; Della Penna, K.; Desai, R.;
Leitl, M. D.; Lemaire, W.; White, R. B.; Yeh, S.; Urban, M. O.; Kane, S. A.;
Hartman, G. D.; Bilodeau, M. T. Bioorg. Med. Chem. Lett. 2011, 21, 2354; (b)
Manley, P. J.; Zartman, A.; Paone, D. V.; Burgey, C. S.; Henze, D. A.; Della Penna,
K.; Desai, R.; Leitl, M. D.; Lemaire, W.; White, R. B.; Yeh, S.; Urban, M. O.; Kane,
S. A.; Hartman, G. D.; Bilodeau, M. T.; Trotter, B. W. Bioorg. Med. Chem. Lett.
2011, 21, 2359.
Compound
R3
hCB2 EC50 nM
HLM Clint
12
RLM Clint
17
O
32
33
11
106
8.0
8.5
34
35
5.3
39
7.0
8.0
19
OCF3
N
8.5
7. Kon-I, K.; Matsumizu, M.; Shima, A. U.S. 2006/0094750.
HLM and RLM. Their selectivity over CB1 and hERG is shown in
Table 4.
Compound 34 had similar a c log P to compound 23 and
although the hERG activity was slightly reduced, this was negated
8. Ryckmans, T.; Edwards, M. P.; Horne, V. A.; Correia, A. M.; Owen, D. R.;
Thompson, L. R.; Tran, I.; Tutt, M. F.; Young, T. Bioorg. Med. Chem. Lett. 2009, 19,
4406.
9. Kalvass, J. C.; Maurer, T. S.; Pollack, G. M. Drug Metab. Distrib. 2007, 35, 660.

C. Watson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4284–4287
4287
10. Hitchcock, S. A.; Pennington, L. D. J. Med. Chem. 2006, 49, 7559.
11. Itoh, T.; Mase, T. Org. Lett. 2004, 24, 4587.
Pan, L.; Ke, H.-A.; Brown, T. A.; Thompson, D. D.; DaSilva-Jardine, P. Bioorg. Med.
Chem. Lett. 2005, 15, 5562.
12. Chesworth, R.; Wessel, M. D.; Heyden, L.; Mangano, F. M.; Zawistoski, M.;
Gegnas, L.; Galluzzo, D.; Lefker, B.; Cameron, K. O.; Tickner, J.; Lu, B.;
Castleberry, T. A.; Petersen, D. N.; Brault, A.; Perry, P.; Ng, O.; Owen, T. A.;
13. Similar Emax SAR has been reported in Verbist, B. M. P.; De Cleyn, M. A. J.;
Surkyn, M.; Fraiponts, E.; Aerssens, J.; Nijsen, M. J. M. A.; Gijsen, H. J. M. Bioorg.
Med. Chem. Lett. 2008, 18, 2574.