Optimization of triaryl bis-sulfones as cannabinoid-2 receptor ligands

Article abstract of DOI:10.1016/j.bmcl.2007.04.028

Structure-activity relationship on our recently reported triaryl bis-sulfone class of cannabinoid-2 (CB2) receptor selective inverse agonists was explored. Modifications to the methane sulfonamide, substitutions to B and C phenyl rings, and replacements o

Full text of DOI:10.1016/j.bmcl.2007.04.028

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3760–3764
Optimization of triaryl bis-sulfones as cannabinoid-2
receptor ligands
a,
a
a
Brian J. Lavey,a,* Joseph A. Kozlowski, Bandarpalle B. Shankar, James M. Spitler,
*
Guowei Zhou,a De-Yi Yang,a Youheng Shu,a Michael K. C. Wong,a Shing-Chun Wong,a
Neng-Yang Shih,a Jie Wu,a Stuart W. McCombie,a Razia Rizvi,a
Ronald L. Wolina and Charles A. Lunnb
aDepartment of Chemistry, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033-0539, USA
bDepartment of Immunology, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033-0539, USA
Received 2 February 2007; revised 4 April 2007; accepted 5 April 2007
Available online 10 April 2007
Abstract—Structure–activity relationship on our recently reported triaryl bis-sulfone class of cannabinoid-2 (CB2) receptor selective
inverse agonists was explored. Modifications to the methane sulfonamide, substitutions to B and C phenyl rings, and replacements
of the C-ring were investigated. A compound with excellent CB2 activity, selectivity for CB2 over CB1, and in vivo plasma levels was
identified.
Ó 2007 Elsevier Ltd. All rights reserved.
We previously reported identification of potent CB2
selective compounds10 and their pharmacology.11,12
The compound with the best combination of potency,
selectivity, and rat exposure was 1.13 However, we felt
the plasma levels of 1 after oral dosing were not high
enough for further development (6 h AUC after oral
dosing at 10 mpk = 6331 nM h, and the plasma levels
decayed rapidly). Therefore, we selected 1 as a lead
for further optimization. We hoped to try and find a
compound with better plasma levels in vivo while
maintaining excellent CB2 affinity, and selectivity for
CB2 over CB1. Herein we describe SAR studies, on
the fluorophenyl ring, methanesulfonyl, and B-ring
substituent.
The CB2 receptor is a G-protein coupled receptor which
is located primarily in the spleen and other immune-re-
lated tissues.1 It is 44% homologous with CB1, and its
difference in homology with CB1 offers the possibility
to uncouple the immune-related effects of cannabinoids
from the psychoactive effects typically associated with
CB1.2 To support this possibility, CB2 activity has been
associated with modulation of B-cell differentiation,3 al-
tered immune cell migration4 or chemotaxis, and altered
antigen processing in macrophages.5 Also, cannabinoid
ligands are reported to be active in animal models of
rheumatoid arthritis,6,7 multiple sclerosis,8 and inflam-
matory pain9 reinforcing the idea that they possess
anti-inflammatory properties.
CH3
We pursued the synthetic strategy of assembling the A–
B ring system first, and then installing the C-ring group,
as shown in Scheme 1. Biarylsulfone 3 was prepared
from 2 by metalation with n-BuLi and trapping with
either a sulfonyl fluoride or disulfide. If a disulfide is
used, then in a second step the sulfur was oxidized with
m-chloroperoxy benzoic acid. Compound 3 was selec-
tively ortho-lithiated on the B-ring using two equivalents
of n-BuLi. Trapping of the anion with sulfur dioxide gas
and conversion to the sulfonyl chloride with N-chloro-
succinimide14 gave compound 4. Treatment of the sulfo-
nyl chloride 4 with various amines produced the
Cl
NHSO2CH3
B
A
SO2
1
SO2
C
F
Keywords: CB2 receptor ligands; Inverse agonists; Cannabinoid recep-
tors; Cannabinoid-2; CB2; Anti-inflammatory.
*
Corresponding authors. Tel.: +1 908 740 2329 (B.J.L.); tel.: +1 908
740 3488 (J.A.K.); fax: +1 908 740 7152; e-mail addresses:
brian.lavey@spcorp.com; joseph.kozlowski@spcorp.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.04.028

B. J. Lavey et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3760–3764
3761
Me
Me
NHCOCF3
X
NHCOCF3
a or b
Br
X
SO2
3
2
a or b
c, d
Me
Me
NHCOCF3
X
NHCOCF3
SO2
SO2
SO2Cl
4
7
SO2
f, g, h
e, f, g
Y
Me
NHR
Me
NHSO2R
X
X
SO2
SO2
R"
8
9
R = SO2CH3
R = SO2CF3
10 R = alkyl
5 R = CH3
R = CF3
SO2
SO2
N
6
R'
Y
F
F
F
NR1R2
C
=
i
11
12
Scheme 1. Reagents and conditions: (a) i—n-BuLi (2 equiv), THF, ꢀ78 °C; ii—sulfonyl fluoride (55–65%); (b) i—n-BuLi (2 equiv), THF, ꢀ78 °C;
ii—disulfide; iii—m-chloroperoxy benzoic acid, CH2Cl2, rt (30–50%); (c) n-BuLi (2 equiv), THF, ꢀ78 °C; ii—SO2 gas; (d) N-chlorosuccinimide (40%
over two steps); (e) (C2H5)3N, R0NHR00, CH2Cl2 (25–75%); (f) LiOH, dioxane (70–90%); (g) (C2H5)3N, CH3SO2Cl or ðCF3SO2Þꢁ (57–88%); (h) R-
2
OTf, Et3N (47–62%); (i) NHR1R2, CH3CN, D 31–87%).
corresponding sulfonamides. The trifluoroacetamide
group was removed by aqueous LiOH hydrolysis, fol-
lowed by sulfonylation of the amine to provide analogs
5 and 6.
Scheme 2 describes the preparation of analogs shown in
Table 3. The conversion of compounds 13–14 involves
either hydrogenation of the chlorine (analog 31), O-
demethylation of a methoxy group (analog 32), replace-
ment of chlorine with alkoxy (analogs 34, 35), or
replacement of chlorine with alkylamino (analogs 39–
41).
The anion of compound 3, formed from selective lithia-
tion of the B-ring, was also trapped with either an aryl-
sulfonyl fluoride, or disulfide. When disulfides were
used, the newly formed sulfide linker between the B
and C rings was oxidized to sulfone with m-chloroper-
oxy benzoic acid to produce 7. Removal of the trifluoro-
acetamide group (aqueous LiOH) and sulfonylation of
the amine provided 8 or 9. With compound 17 an addi-
tional oxidation step was required to form the N-oxide.
In compound 11, one of the two fluorines could be dis-
placed by heating the compound with an amine in ace-
tonitrile to produce 12. The cyclopropyl analog 38 was
made from 3, where X is cyclopropyl and 4-cyclo-
propylbenzenesulfonyl fluoride was prepared from
cyclopropylbenzene.16
Table 1 shows the outcome of modifying or replacing
the C-ring of compound 1.17 Simple alkyl and cycloalkyl
analogs 15 and 20 lost affinity relative to 1. Amino and
sulfonamide analogs 21–23 also lost CB2 affinity.
Among the heterocyclic analogs 16–19, the 2-pyridyl
analog 16 showed CB2 affinity and selectivity similar
to 1, however, it did not offer a clear advantage over
the 2-F-phenyl. Hence, the 2-F-phenyl was maintained
for further SAR modification.
We next explored the possibility of replacing the metha-
nesulfonyl by other groups. The results are shown in
Me
Me
X
Z
NHSO2CF3
NHSO2CF3
SO2
SO2
SO2
SO2
(a) X = Cl to Z = H
13
14
C
C
(b) X = OMe to Z = OH
(c) X = Cl to Z = OR
(d) X = Cl to Z = NRR1
F
F
Scheme 2. Reagents and conditions: (a) H2/Pd(C) 91%; (b) BBr3, CH2Cl2; 9% (c) NaOH, ROH (70–89%); (d) NRR0, Pd, Ref. 15 (22–40%).

3762
B. J. Lavey et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3760–3764
Table 1. SAR: changes to the C-ring of 1
Table 3. SAR: changes to the B-ring of 24
Me
Me
Cl
SO2CH3
SO2CF3
X
N
N
A
A
B
H
B
H
SO2
SO2
SO2
SO2
C
C
C
F
Compound
Ki
Selectivity
(CB1 Ki/CB2 Ki)
(CB2, nM)
Compound
X
H
Ki Selectivity
(CB2, nM) (CB1 Ki/CB2 Ki)
1
15
16
1.3
308
2.9
4387
57
31
32
33
34
35
36
37
38
39
40
41
25
1.3
5
426
1571
525
F
–OH
–OCF3
–OCH2CH3
–CH2CH2CH3
0.9
1964
1602
2080
2287
2239
170
4403
–OCH2CH2OCH3 2.5
N
–CF3
–CH3
2.5
1.1
1.4
10
–Cyclopropyl
–NH2
17
18
19
20
119
129
204
8.9
840
538
31
N
+ O-
–NH-cyclopropyl 4.9
–N-Piperazine 1099
66
31
O
Now by fixing the B-ring group as chlorine and the
amine substitution as trifluoromethanesulfonyl, we
revisited C-ring modifications to identify the best combi-
N
345
Table 4. SAR: changes to the C-ring of 24
21
22
23
–NHCH3
–N(CH3)2
–NHSO2CH3
1251
1273
2749
64
58
36
Me
SO2CF3
Cl
N
H
A
B
SO2
Table 2 where an amide, sulfonamide, urea, or amine
was prepared. Alkylated amines 29 and 30 showed good
CB2 affinity, but it was the trifluoromethanesulfonamide
24 that showed the best combination of CB2 affinity and
selectivity.
SO2
C
Compound
C
Ki Selectivity
(CB2, nM) (CB1 Ki/CB2 Ki)
11
42
2,6-Difluorophenyl
5.8
3.6
642
Table 3 shows results where the chlorine group on 24
was replaced. Although analogs 34, 37, and 38 showed
good CB2 affinity and selectivity, none had a profile sig-
nificantly better than 24.
1152
N
N
N
43
44
45
46
47
48
21
859
1191
212
688
60
F
Table 2. SAR: changes to the methanesulfonate of 1
4.1
NH2
F
Me
Cl
D
22
A
B
NHCH3
F
SO2
SO2
F
14.4
C
NH
F
Compound
D
Ki Selectivity
(CB2, nM) (CB1 Ki/CB2 Ki)
115
N(CH3)2
F
24
25
26
27
28
29
30
–NH–SO2CF3
–NHCOCH3
–NHCOCF3
–NHCONHEt
–NH2
2
3393
250
784
—
60.7
3.7
140
18
N
O
>10,000
3953
7
14
49
50
51
–NH–t-Bu
–N-Piperidine
–N-Morpholine
136
58
123
231
99
–NCH2CF3
325
105
–NCH2CH2CF3 17
192

B. J. Lavey et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3760–3764
3763
1553; (d) Huffman, J. W. Curr. Pharm. Des. 2000, 6, 1323;
(e) Goya, P.; Jagerovic, N. Exp. Opin. Ther. Patents 2000,
10, 1529, and references cited therein.
nation, shown in Table 4. The 2-pyridyl analog 42 main-
tained reasonable CB2 affinity and selectivity, but we
found that none of the changes in the C-ring improved
CB2 affinity or selectivity over 24.
2. For a review of cannabinoids and the immune system, see:
Croxford, J. L.; Yamamura, T. J. Neuroimmunol. 2005,
166, 3.
Compound 24 was dosed orally at 10 mpk in the rapid
rat assay18 and showed a 6 h AUC of 3581 nM h, which
was worse than the 6331 nM h observed for compound
1. However, we observed that the compound was a sus-
pension in the methylcellulose dosing solution and sus-
pected that poor solubility might be a factor in the
lower plasma levels. Taking advantage of the increased
acidity of the NH of NHSO2CF3, over the NH of
NHSO2CH3 (pKa approximately 6.5 vs ꢂ12) we pre-
pared and dosed several salt forms of compound 24.
In the same rapid rat paradigm, we found the potassium
(5405 nM h), and calcium (5075 nM h) salts gave AUCs
comparable to the neutral compound, and the zinc
(297 nM h) salt was lower. However, the sodium salt
AUC (21,022 nM h) was significantly improved over
the neutral compound.
3. Carayon, P.; Marchand, J.; Dussossoy, D.; Derocq, J. M.;
Jbilo, O.; Bord, A.; Bouaboula, M.; Galiegue, S.; Mon-
diere, P.; Penarier, G.; Fur, G. L.; Defrance, T.; Casellas,
P. Blood 1998, 92, 3605.
4. Sacerdote, P.; Massi, P.; Panerai, A. E.; Parolaro, D.
J. Neuroimmunol. 2000, 109, 155.
5. McCoy, K. L.; Matveyeva, M.; Carlisle, S. J.; Cabral, G.
A. J. Pharmacol. Exp. Ther. 1999, 289, 1620.
6. Zurier, R. B.; Rossetti, R. G.; Lane, J. H.; Goldberg, J.
M.; Hunter, S. A.; Burstein, S. H. Arthritis Rheum. 1998,
41, 163.
7. Malfait, A. M.; Gallily, R.; Sumariwalla, P. F.; Malik, A.
S.; Andreakos, E.; Mechoulam, R.; Feldman, M. Proc.
Natl. Acad. Sci. 2000, 97, 9561.
8. Baker, D.; Pryce, G.; Croxford, J. L.; Brown, P.; Pertwee,
R. G.; Huffman, J. W.; Layward, L. Nature 2000, 404, 84.
9. Quartilho, A.; Mata, H. P.; Ibrahim, M. M.; Vanderah, T.
W.; Porreca, F.; Makriyannis, A.; Malan, T. P. Anesthe-
siology 2003, 99, 955.
Because of its potency, selectivity, and improved plasma
levels in the rapid rat screening assay, the DMPK profile
of the sodium salt of compound 24 was investigated fur-
ther. In rats, the bioavailability of 24 was 25%, the t1/2
was 1.6 h, the clearance was 11.4 mL/min/kg, and the
volume of distribution was 0.7 L/kg. When dosed orally
in cynomolgus monkeys at 10 mpk, it had a 0–24 h AUC
of 45,400 nM h, bioavailability of 50%, a t1/2 of 3.7 h,
clearance of 8.1 mL/min/kg, and a volume of distribu-
tion of 1.3 L/kg. In dogs, the 0–24 h AUC was
173,800 nM h after dosing at 10 mpk orally. The bio-
availability was 66%, the clearance was 1.2 mL/min/kg,
and the Vd was 1.1 L/kg.
10. Lavey, B. J.; Kozlowski, J. A.; Hipkin, R. W.; Gonsiorek,
W.; Lundell, D. J.; Piwinski, J. J.; Narula, S.; Lunn, C. A.
Bioorg. Med. Chem. Lett. 2005, 15, 783.
11. Lunn, C. A.; Fine, J. S.; Rojas-Triana, A.; Jackson, J. V.;
Fan, X.; Kung, T. T.; Gonsiorek, W.; Schwarz, M.; Lavey,
B. J.; Kozlowski, J. A.; Narula, S. K.; Lundell, D. J.;
Hipkin, R. W.; Bober, L. A. J. Pharacol. Exp. Ther. 2006,
316, 780.
12. Gonsiorek, W.; Hesk, D.; Kinsley, D.; Chen, S.-C.; Fine,
J. S.; Jackson, J. V.; Bober, L. A.; Deno, G.; Lunn, C. A.;
Kozlowski, J. A.; Lavey, B.; Piwinski, J.; Narula, S. K.;
Lundell, D. J.; Hipkin, R. W. J. Biol. Chem. 2006, 281,
28143.
13. Shankar, Bandarpalle B.; Lavey, Brian J.; Zhou, Guowei;
Spitler, James A.; Tong, Ling; Rizvi, Razia; Yang, De-Yi;
Wolin, Ronald; Kozlowski, Joseph A.; Shih, Neng-Yang;
Wu, Jie; Hipkin, R. William; Gonsiorek, Waldemar;
Lunn, Charles A. Bioorg. Med. Chem. Lett. 2005, 15, 4417.
14. To a THF (40 mL) solution of compound 3, X = Cl
(6.5 mmol), at ꢀ78 °C, n-BuLi (2.0 M, 6.6 mL) was added
drop wise. After stirring for 45 min at ꢀ78 °C, SO2 gas
was bubbled gently for 45 min. The reaction turned light
yellow and then it was allowed to warm to room
temperature. The reaction mixture was concentrated to
dryness and redissolved in CH2Cl2 (5 mL). The resulting
lithium sulfinate salt was precipitated by adding cold
hexanes and collected by suction filtration. The salt was
dissolved in CH2Cl2 (100 mL) and treated with N-chloro-
succinimide (1.05 g, 7.8 mmol) and the reaction mixture
was stirred overnight at room temperature. The organic
layer was washed with brine and water, dried over
Na2SO4, filtered, and concentrated to provide crude 4,
which was purified by silica gel chromatography (10%
EtOAc/Hexanes), resulting in 1 g of 4 (40%) as a solid.
15. For conditions to convert X = Cl to X = NRR0, see:
Wolfe, J. P.; Tomori, H.; Sadighi, J. P.; Yin, J.; Buchwald,
S. L. J. Org. Chem. 2000, 65, 1158.
In an effort to optimize the profile of compound 1,
we expanded the SAR, by making structural changes
to several points of the molecule. These changes
included the addition of polar groups to the 2-fluor-
ophenyl ring, replacement of the fluorophenyl ring
with heterocycles and other groups, and changing
the NHSO2CH3 group into a NHSO2CF3 thereby
allowing the formation of salts to improve solubility.
Compound 24 was found to have the best overall
profile as a sodium salt and was selected for further
progression.
Supplementary data
Experimental details of the determination of Ki values,
the synthesis of cyclopropyl benzene sulfonyl fluoride,
and the preparation of compound 24 can be found in
the online version at doi:10.1016/j.bmcl.2007.04.028.
16. For the preparation of cyclopropylbenzene sulfonyl fluo-
ride, see the supplementary material of this article.
17. In all cases, individual data points for determinations of Ki
for CB1 and CB2 were carried out in triplicate, in two
separate assays. For experimental details of the determi-
nation of Ki values, see the supplementary material. For
additional details on the syntheses, see: Kozlowski, J. A.;
References and notes
1. For general reviews of the cannabinoid field, see: (a)
Mackie, K. Annu. Rev. Pharmacol. Toxicol. 2006, 46, 101;
(b) Hertzog, D. L. Expert. Opin. Ther. Patents 2004, 14,
1430; (c) Pertwee, R. G. Exp. Opin. Invest. Drugs 2000, 9,

3764
B. J. Lavey et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3760–3764
Shih, N.-Y.; Lavey, B. J.; Rizvi, R. K.; Shankar, B. P;
18. For experimental details of the rapid rat assay, see:
Korfmacher, W. A.; Cox, K. A.; Ng, K. J.; Veals, J.;
Hsieh, Y.; Wainhaus, S.; Broske, L.; Prelusky, D.;
Nomeir, A.; White, R. E. Rapid Commun. Mass Spectrom.
2001, 15, 335.
Spitler, J. M.; Tong, L.; Wolin, R.; Wong, M. K.
WO02062750, and Kozlowski, J. A. et. al. US2003232859.
All CB2 inhibitors gave satisfactory analytical data as
1
indicated by H NMR and LC–MS.