Novel thioamide derivatives as neutral CB1 receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2009.11.125

A novel class of cannabinoid-1 (CB1) receptor antagonists for the treatment of obesity is presented. The carboxamide linker in a set of 5,6-diaryl-pyrazine-2-amide derivatives was transformed into the corresponding thioamide, by using a one-pot synthesis. The structural series of thioamides not only showed retained CB1 potency (below 10 nM), but also showed improved solubility. In addition, the neutral antagonist 2c significantly reduced body weight in cafeteria diet obese mice.

Full text of DOI:10.1016/j.bmcl.2009.11.125

Bioorganic & Medicinal Chemistry Letters 20 (2010) 479–482
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel thioamide derivatives as neutral CB1 receptor antagonists
b
b
a
a
Jonas Boström a, , Roine I. Olsson , Joakim Tholander , Peter J. Greasley , Erik Ryberg ,
*
Henrik Nordberg c, Stephan Hjorth d, Leifeng Cheng b
a Lead Generation Department, AstraZeneca R&D Mölndal, S-431 83 Mölndal, Sweden
b Medicinal Chemistry Department, AstraZeneca R&D Mölndal, S-431 83 Mölndal, Sweden
c Dx DMPK Department, AstraZeneca R&D Mölndal, S-431 83 Mölndal, Sweden
d Bioscience Department, AstraZeneca R&D Mölndal, S-431 83 Mölndal, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel class of cannabinoid-1 (CB1) receptor antagonists for the treatment of obesity is presented. The
carboxamide linker in a set of 5,6-diaryl-pyrazine-2-amide derivatives was transformed into the corre-
sponding thioamide, by using a one-pot synthesis. The structural series of thioamides not only showed
retained CB1 potency (below 10 nM), but also showed improved solubility. In addition, the neutral antag-
onist 2c significantly reduced body weight in cafeteria diet obese mice.
Received 5 October 2009
Revised 19 November 2009
Accepted 21 November 2009
Available online 27 November 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Cannabinoid-1
Neutral antagonist
Thioamide
Lawesson’s reagent
Conformational energy
DIO mice
Matched-pairs
Obesity is a predisposing factor for the development of type-2
diabetes, hypertension, and cardiovascular disease.1 Currently
available anti-obesity therapeutics includes Orlistat2 and Sibutr-
amine.3 Orlistat works by inhibiting gastric and pancreatic lipases,
and Sibutramine is a CNS acting serotonin-norepinephrine re-up-
take inhibitor. The body weight reduction Orlistat and Sibutramine
offer is modest, and both drugs are associated with adverse ef-
fects.4 Alternative treatments are thus desirable. Blocking the can-
nabinoid-1 (CB1) receptor presents an interesting approach for
treating obesity.5 While the clinical effectiveness of this mecha-
nism was indeed borne out by the CB1 receptor antagonist Rimo-
nabant, the compound was rejected before launch by the FDA in
June 2007,6 and was withdrawn from all other markets late 2008
due to an increased risk of psychiatric disorders.7
In our quest for novel CB1 receptor antagonists as anti-obesity
agents we, and others, have previously described scaffold-hopping
activities. That is, the pyrazole ring-scaffold in Rimonabant has been
replaced by other scaffolds, such as pyrazine,8 thiazole,9 pyrrole,10
and pyridine11,12 rings. All these classes include ligands with high
potencies for the CB1 receptor. It is thus believed that the central scaf-
fold is mainlyessential for geometrical reasons, since these fragments
are electronically quite different.8,13 In contrast, modifications made
on the carboxamide linker have shown more influential effects. In
an interesting study by Reggio,14 a vinyl cyclohexyl analog (VCHSR)
of Rimonabant was presented (Fig. 1). Results indicated that the ana-
log, lacking hydrogen bonding ligand-receptor interaction possibili-
ties in the (carboxamide) linker region, was significantly less potent
at CB1. Intriguingly, the vinyl cyclohexyl analog was also reported
to act as neutral antagonists at CB1, while Rimonabant acts as an in-
verse agonist at this receptor. Moreover, there are a number of exam-
ples where differentiation between neutral and inverse agonism has
beenshowntohaveclinicalconsequences.15 Thisisofparticularinter-
est since CB1 neutral antagonists and inverse agonists may cause dif-
ferent degrees of unwanted side-effects at clinically relevant doses.16
We hypothesized that, by modifying the hydrogen bonding strength
in the linker region, potent and neutral CB1 antagonists that suppress
appetite, while avoiding undesired side-effects, could be designed.
The frequently occurring carboxamide linker8–13 was replaced
by a thioamide linker, using a one-step reaction, see Figure 2.
Molecular alignments showed that the carboxamide linker in the
pyrazine-2-carboamide derivatives could be exchanged to a thio-
amide, while retaining the proposed bioactive conformation.8 Fig-
ure 3 illustrates a shape-based overlay of compound 1b over 2b.
The geometries of the proposed bioactive conformations are virtu-
ally identical. The ROCS program was used for the shape overlay.17
Thioamides are typically prepared by treating amides with
reagents such as phosphorus pentasulfide. In the current work
* Corresponding author.
E-mail address: dr.jonas.bostrom@gmail.com (J. Boström).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.11.125

480
J. Boström et al. / Bioorg. Med. Chem. Lett. 20 (2010) 479–482
Figure 1. Structures of Rimonabant and VCHSR.
Figure 2. Structures of the carboxamide 1a and the corresponding thioamide analog 2a.
A GTPcS CB1 functional assay was used to assess the potency of
the pyrazine derivatives described above.8 Receptor inhibition
potencies for compounds 1a–c and 2a–c, are reported in Table 1.
For all carboxamide thioamide pairs (1a vs 2a, 1b vs 2b and 1c
vs 2c) the CB1 affinity was found to be equipotent.
CB1 receptor antagonists are in general very lipophilic com-
pounds, with low solubilities. The compounds in Table 1 all show
experimentally determined log D values greater than 6.5.20 Unex-
pectedly, when considering the thioamide analogs of the carbox-
amide structures it was observed that two of the three pairs
showed improved solubility relative to the parent carboxamide
compound.21 To establish if this was a general effect, a search of
the AstraZeneca corporate database identified seven additional
such pairs which differed only by the substitution of a thioamide
for a carboxamide. In all such cases the measured solubility for
these ‘matched-pairs’22 showed that the carboxamide derivative
was the most soluble partner. Thus the solubilities of the pyrazine
structures reported in Table 1 appear to be anomalous in their
behavior with respect to substitution of carboxamide with
thioamide.
Figure 3. A molecular overlay of 1b and 2b.
Lawesson’s reagent18 was used to prepare compounds 2a–c from
compounds1a–c, see Figure2. All compoundsshowed 1H NMR spec-
tra consistent with their structure and were >95% pure by HPLC.19
Table 1
Structure and activity for compounds 1a–c and 2a–c
Compound
X
R1
R2
R3
GTP S IC50 valuesa (nM)
Solubilityb
(lM)
c
Rimonabant
2.1
5.3
4.5
16.6
5.8
2.9
<1.1
<0.1
6.6
<0.1
6.1
1a
2a
1b
2b
1c
2c
O
S
O
S
O
S
H
H
H
H
F
F
H
H
F
F
F
H
H
F
CH2OCH3
CH2OCH3
<0.1
<0.1
F
F
2.4
The experimental potency data are given as IC50 values.
a
The experimental potency data are given as IC50 values, for the CB1 antagonists.
The aqueous solubility data was measured in a consistent way from stored DMSO solutions.
b

J. Boström et al. / Bioorg. Med. Chem. Lett. 20 (2010) 479–482
481
**
Figure 4. Two torsional scans using density functional theory (B3LYP/6-31G ). The
thioamide (h) and the carboxamide (Â) structures both show a strong preference
for an internal hydrogen bond. That is, they both reach a geometry optimum at zero
degrees (N1–C2–C3–N4 = 0).
Thioamide derivatives are analogous to carboxamides, but they
exhibit greater multiple bond character along the amide (C–N)
bond, resulting in a larger rotational barrier.23 We performed de-
tailed density functional theory calculations to investigate if the
geometry preference for thioamides and carboxamides attached
to 2-pyrazines differs. The relaxed coordinate scan implemented
in the Jaguar program24 was employed for this purpose, using
Figure 6. (a–b) The in vivo treatment effect on body weight in cafeteria diet fed
mice of 1c, 2c and Rimonabant is shown. Vertical grey lines depicts end of
treatment/start of washout phase.
B3LYP25 and basis set 6-31G , in PBF solvent. The energy profile
**
for torsion N1–C2–C3–N4 is shown in Figure 4. To speed up the cal-
culations they were performed on a reduced system.
It is clear that both pyrazine-2-carboxamides and pyrazine-2-
thioamide will, without a doubt,26 adopt planar geometries. Thus,
both systems show a preferred energy minimum at zero degrees
(N1–C2–C3–N4 = 0). The main reason for the steep energy profile
is the stabilizing internal hydrogen bond (N4–H–N1). It is worth
noting that the thioamide N–H is a stronger hydrogen bond donor,
as compared to the carboxamide N–H, but the sulfur it is a weaker
hydrogen bond acceptor than the carboxamide oxygen.27
A few neutral CB1 antagonists (i.e., ligands that lack the ability
to attenuate CB1 receptor constitutive activity) have been reported
in the literature.16 For example, the high affinity CB1 antagonist O-
2050 is devoid of agonist effects. Nevertheless, O-2050 still shows
anorectic effects similar to Rimonabant.28 Figure 5a shows the
antagonistic effects of compounds 1c, 2c and Rimonabant, as mea-
sured using antagonism of CP-55940 activation in GTPcS CB1 func-
tional assay whilst Figure 5b demonstrates their inverse agonist
properties. The thioamide 2c display the typical characteristics of
a neutral antagonist; inhibiting agonist mediated activity whilst
not effecting the intrinsic activity of the CB1 receptor, whereas
its carboxamide partner 1c, as well as Rimonabant, shows typical
inverse agonist behavior.
Peroral administration of 1c, 2c and Rimonabant reduced body
weight relative to vehicle in cafeteria diet induced obese mice.29
The body weight after 1–3 weeks dropped to about 85% of the ini-
tial body weight for Rimonabant (20
lmol/kg; Fig. 6a–b), as com-
Figure 5. The antagonistic (a) and inverse agonistic (b) effects of compounds 1c (o),
2c (+) and Rimonabant (
D
), as measured in a GTP
c
S CB1 functional assay.
pared to 85–90% for 2c (20–60 mol/kg; Fig. 6a–b) after 1 week
l

482
J. Boström et al. / Bioorg. Med. Chem. Lett. 20 (2010) 479–482
Table 2
Plasma and brain tissue levels of 1c and 2c in diet induced obese mice at different time points, after the last administration of 2ca
Time
Compound
2c
1c
Rimonabant
mol/L)
Plasma (
lmol/L)
Brain (lmol/g)
Plasma (
l
mol/L)
Brain (
lmol/g)
Plasma (
l
Brain (lmol/g)
0.5 h
2 h
6 h
7d
0.403
0.604
0.137
0.032
0.18
0.32
0.14
0.05
0.136
0.153
0.148
0.024
0.05
0.08
0.07
NDb
1.36
1.95
0.90
NDb
1.11
1.69
0.68
NDb
a
20
l
mol/kg, PO administration (see Fig. 6a).
b
ND stands for not determined; below limit of quantification.
3. Heal, D. J.; Aspley, S.; Prow, M. R.; Jackson, H. C.; Martin, K. F.; Cheetham, S. C.
Int. J. Obes. Relat. Metab. Disord. 1998, 22, S18.
4. Leung, W. Y.; Neil, T. G.; Chan, J. C.; Tomlinson, B. Clin. Ther. 2003, 25, 58.
5. Pertwee, R. G. Addict. Biol. 2000, 5, 37.
6. http://www.fda.gov/ohrms/dockets/ac/07/minutes/2007-4306m1-final.pdf.
7. http://www.emea.europa.eu/htms/human/withdraw/withdraw.htm.
8. Boström, J.; Berggren, K.; Elebring, T.; Greasley, P. J.; Wilstermann, M. Bioorg.
Med. Chem. 2007, 15, 4077.
9. Berggren, A. I. K.; Boström, S. J.; Elebring, T. S.; Fällefors, L.; Wilstermann, J. M.;
Greasley, P. WO2004058255, 2004.
10. Berggren, A. I. K.; Boström, S. J.; Cheng, L.; Elebring, S. T.; Greasley, P.; Någård,
M.; Wilstermann, J. M.; Terricabras, E. WO 2004058249, 2004.
11. Finke, P. E.; Meurer, L. C.; Debenham, J. S.; Toupence, R. B.; Walsh, T. F. WO
2003082191, 2003.
12. Barth, F.; Martinez, S.; Rinaldi Carmona, M. WO 2003084930, 2003.
13. Lange, J. H. M.; Kruse, C. G. Curr. Opin. Drug Discovery 2004, 7, 498.
14. Reggio, P. H. Curr. Pharm. Des. 2003, 9, 1607.
and approximately 75% after 3 weeks. Treatment with 1c (2–
20 mol/kg) resulted in body weight reductions to 90–95% after
6 days (Fig. 6b). Interestingly, the body weight reduction effect re-
mains for 1c and 2c after a 1–2 week washout period. This is not
the case for Rimonabant (Fig. 6a–b).
While pharmacokinetic factors may contribute, available data
are consistent with the suggestion that the body weight-decreas-
ing effect of 2c in the diet induced obese mouse model stems pri-
marily from its neutral antagonist properties. It might be argued
that bioconversion of 2c to 1c accounts for the observed in vivo ef-
fect. However, following repeat administration of compound 2c
l
(20 lmol/kg, PO; Fig. 6a), the observed plasma and brain concen-
trations of this agent clearly exceeds that of compound 1c in the
same samples (Table 2). Notably, no levels of 1c were detectable
in the 7 day washout samples from 2c-treated mice despite the
sustained in vivo effect on body weight at this time point
(Fig. 6a). Moreover, the neutral CB1 antagonist 2c is expected to
block the effects of agonists as well as inverse agonists at CB1. Gi-
ven the equipotency and the relative concentrations of 1c versus 2c
in the biophase, the neutral antagonism of 2c would therefore be
expected to prevail over any inverse agonist effects of 1c.
To conclude, a linker hopping approach has been exploited for
the design of a novel class of CB1 antagonists for the treatment
of obesity. The synthesis and CB1 antagonistic activities of a new
series of 5,6-diaryl-pyrazine-2-thioamide derivatives have been
described. Several compounds showed in vitro potency below
10 nM for the CB1 receptor. This approach not only showed re-
tained CB1 activity, but also, of particular interest in the CB1 area,
is the improved solubility of the thioamide derivates, as compared
to the corresponding carboxamides. A significant reduction in body
weight for cafeteria fed mice was observed for compound 2c. It is
hypothesized that the anorectic effects of 2c stems primarily from
its neutral antagonist properties. In future work, the identified car-
boxamide/thioamide pairs can be used to better understand what
different mode of actions mean.
15. Greasley, P. J.; Clapham, J. C. Eur. J. Pharmacol. 2006, 553, 1.
16. Pertwee, R. G. Life Sci. 2005, 76, 1307.
17. ROCS: OpenEye Scientific Software, Inc., 9 Bisbee Court, Suite D, Santa Fe, NM
87508, US.
18. (a) Lecher, H. Z.; Greenwood, R. A.; Whitehouse, K. C.; Chao, T. H. J. Am. Chem.
Soc. 1956, 78, 5018; (b) Jesberger, M.; Davis, T. P.; Barner, L. Synthesis 2003, 13,
1929.
19. In a typical experiment compound 1a (0.15 mmol) and Lawesson’s reagent
(0.31 mmol) were mixed in toluene (5 mL) and allowed to react at 150 °C for
10 min in the microwave reactor (single node heating). Toluene was removed
under reduced pressure and the residue was diluted with dichloromethane
(30 mL) and then washed with water (2 Â 10 mL) and brine (10 mL). The organic
layer was dried (MgSO4), filtered and concentrated under reduced pressure. The
crude product was purified by column chromatography on silica gel eluting with
25–80% ethyl acetate in heptane. The product fraction was concentrated under
reduced pressure to give 2a in 54% yield.
20. All compounds in Table 1 were found to be outside the calibration range (i.e.,
log D >6.5) at pH 7.4. The lipophilicity (log D) values were determined by a
HPLC LC/MS method described by Valko et al. (Valko, K.; Bevan, C.; Reynolds, D.
Anal. Chem. 1997, 69, 2022).
21. The aqueous solubility data was measured ina consistent way from stored DMSO
solutions.
22. Leach, A. G.; Jones, H. D.; Cosgrove, D. A.; Kenny, P. W.; Ruston, L.; MacFaul, P.;
Wood, J. M.; Colclough, N.; Law, B. J. Med. Chem. 2006, 49, 6672.
23. Wiberg, K. B.; Rablen, P. R. J. Am. Chem. Soc. 1995, 117, 2201.
24. Jaguar, version 7.5; release 207, Schrödinger, LLC: New York, NY, 2008.
25. (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648; (b) Lee, C.; Yang, W.; Parr, R. G.
Phys. Rev. B: Condens. Matter 1988, 37, 785.
26. Boström, J.; Norrby, P.-O.; Liljefors, T. J. Comput. Aided Mol. Des. 1998, 12,
383.
27. Lee, H.-J.; Choi, Y.-S.; Lee, K.-B.; Park, J.; Yoon, C.-J. J. Phys. Chem. 2002, 106,
7010.
References and notes
28. Gardner, A.; Mallet, P. E. Eur. J. Pharmacol. 2006, 530, 103.
29. The C57Bl/6 mice were purchased from Harlan Laboratories (Horst, the
Netherlands).
1. Sowers, J. R. Am. J. Med. 2003, 115, 37.
2. Bodkin, J.; Humphries, E.; McLeod, M. Aust. J. Chem. 2003, 56, 795.