Exploring SAR features in diverse library of 4-cyanomethyl-pyrazole-3-carboxamides suitable for further elaborations as CB1 antagonists

Article abstract of DOI:10.1016/j.bmcl.2009.11.047

A chemically diverse library of secondary and tertiary 4-cyanomethyl-1,5-diphenyl-1H-pyrazole-3-carboxamides was synthesized to enable mapping of the SAR, in the eastern amide region, with regard to CB1 antagonist activity, This study was initiated as a prelude to the design and synthesis of possible CB1 antagonists that do not readily pass the blood-brain-barrier. In general a range of modifications were found to be tolerated in this part of the molecule, although polar and especially charged groups did to a degree reduce the CB1 antagonistic activity. Several compounds with single-digit or even sub-nanomolar potency, suitable for further elaboration of the nitrile moiety, were identified.

Full text of DOI:10.1016/j.bmcl.2009.11.047

Bioorganic & Medicinal Chemistry Letters 20 (2010) 26–30
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Exploring SAR features in diverse library of 4-cyanomethyl-pyrazole-3-
carboxamides suitable for further elaborations as CB1 antagonists
Martin Cooper, Jean-Marie Receveur, Emelie Bjurling, Pia K. Nørregaard, Peter Aadal Nielsen,
*
Niklas Sköld, Thomas Högberg
7TM Pharma A/S, Fremtidsvej 3, DK-2970 Hørsholm, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history:
A chemically diverse library of secondary and tertiary 4-cyanomethyl-1,5-diphenyl-1H-pyrazole-3-car-
boxamides was synthesized to enable mapping of the SAR, in the eastern amide region, with regard to
CB1 antagonist activity, This study was initiated as a prelude to the design and synthesis of possible
CB1 antagonists that do not readily pass the blood–brain-barrier. In general a range of modifications were
found to be tolerated in this part of the molecule, although polar and especially charged groups did to a
degree reduce the CB1 antagonistic activity. Several compounds with single-digit or even sub-nanomolar
potency, suitable for further elaboration of the nitrile moiety, were identified.
Received 24 October 2009
Revised 10 November 2009
Accepted 12 November 2009
Available online 15 November 2009
Keywords:
CB1 antagonists
Obesity
Ó 2009 Elsevier Ltd. All rights reserved.
CNS
Blood–brain-barrier
Chemical library
Plate synthesis
Endogenous cannabinoids (CB) or endocannabinoids, such as
anandamide, are signaling substances derived from arachidonic
control and lipid profile in type 2 diabetic patients (SERENADE)
as well as loss of visceral and hepatic fat in abdominally obese
patients (ADAGIO).6,7 These long term peripheral effects could
arise from, for example, reduced hepatic lipogenesis, enhanced
lipid oxidation, increased adipose lipolysis, and reduced insulin
resistance. Recently, blockade of the CB1 receptor has also been
shown, in rodents, to hold promise as a therapy for chronic liver
diseases such as liver fibrosis.8
acid that are functionally related to
D
9-tetrahydrocannabinol
(D
9-THC), the major psychotropic constituent in marijuana
(Fig. 1).1,2 Two types of cannabinoid receptors have been identified
and cloned. The CB1 receptor is widely expressed in the brain and
to a lower extent in several peripheral organs such as liver, adipose
tissue, gastrointestinal tract, vagal nerves, pancreas and skeletal
muscle.1–3 CB2 receptors exist mainly in the immune system,
although recent publications have reported their existence also in
the CNS.1–3 The endocannabinoids play a pivotal role, through
the CB1 receptor, in regulating food intake and energy expenditure,
in synergy with other anorexigenic and orexigenic regulatory path-
ways.1,2 CB1 receptor-deficient mice are resistant to diet-induced
obesity even though their total caloric intake is similar to that of
wild-type littermates.4 Correspondingly CB1 antagonists/inverse
agonists, such as rimonabant and taranabant (Fig. 1), have been
shown to inhibit food intake and reduce body weight in obese ani-
mals and humans.1,2,5 These effects are likely to be mediated by a
combination of central and peripheral target organ actions.2,6 Phar-
macological blockade of the peripheral CB1 system can elicit ben-
eficial effects on other parameters of the metabolic syndrome that
are independent of eating behavior and body weight.2,6 Thus
clinical trials with rimonabant showed improvement in glycemic
An issue associated with this drug class is the induction of cen-
tral nervous system (CNS) side-effects such as anxiety and depres-
sion.2,9 Effects evidently linked to functional CB1 receptors in brain
areas such as the frontal cortex, hippocampus and amygdala.10
A
plausible working hypothesis, to avoid the CNS effects seen with
rimonabant and taranabant but still achieve effects on body weight
and metabolic parameters, would be to design compounds not able
to pass the blood–brain-barrier (BBB).2,11 One can envisage several
possible ways to achieve this such as by considerably increasing
polar surface area and lowering log D, compared to the centrally
acing drugs, or by invoking efflux transporters located in the
BBB. As a first step to obtaining more polar compounds, we needed
to expand the known SAR for rimonabant-like compounds so as to
better understand which features drive potency and identify
regions where polar groups can be introduced without adversely
affecting CB1 antagonistic effects.
The rimonabant template offers several possibilities for modifi-
cations and we have explored several of these. In this report we
present a strategy of introducing a nitrile group to the 4-methyl
* Corresponding author. Tel.: +45 3925 7760; fax: +45 3925 7776.
E-mail address: th@7tm.com (T. Högberg).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.11.047

M. Cooper et al. / Bioorg. Med. Chem. Lett. 20 (2010) 26–30
27
system. The changes in calculated log P values and polar surface
areas (PSA) of these target motifs are shown in relation to a nitrile
compound in Figure 2. Rimonabant displays the corresponding
values of c log P 6.47 and PSA 50.2 Å2 which reflects a favorable
BBB penetration. The pyrazole intermediate II was obtained from
2-chlorophenylhydrazine and the diketo compound I, according
to the literature procedure outlined in Scheme 1.12 The key inter-
mediate III was obtained by radical induced bromination of II
directly followed by crown ether facilitated substitution with
cyanide in acetonitrile at reflux overnight, or under the milder
conditions with cyanide in ethanol/water, and a final hydrolysis
of the ester. Finally the amides IV were obtained by coupling the
resulting acid III with amines under standard conditions, such as
1-hydroxy-benzotriazole and 1-ethyl-3(3-dimethyl-aminopro-
pyl)-carbodiimide in dichloromethane. Many of the amides were
made by plate chemistry using polystyrene (PS)-supported re-
agents and a diverse set of amine reagents and parallel purification
techniques.13
O
OH
OH
N
H
O
Δ9-THC
anandamide
CN
O
N
N
H
O
N
N
N
O
N
H
Cl
Cl
CF3
Cl
rimonabant
taranabant
Cl
Figure 1. Chemical structures of CB1 agonists (
D
9-THC and anandamide) and two
representative CB1 antagonists progressed to clinical development.
The compounds were screened against the human CB1 receptor
using a GTPcS antagonist assay with CP55940 as agonist and
membranes produced from CHO-K1 cells stably transfected with
recombinant human CB1 receptor. Selected compounds were also
investigated for inverse agonism by running the same assay with-
out the CP55940 agonist. As seen in Table 1, a wide range of amide
functionalities are tolerated in the eastern amide position; thus the
anilinic amides 1-6 have potencies in the range of 10–60 nM. It is
notable that the 3-pyridyl derivative 2 loses potency compared
group of the pyrazole core together with the exploration of a wide
range of eastern amide modifications (IV). The nitrile was chosen
as it can be elaborated into a variety of polar functional groups that
are either neutral (amides, amidoximes), positively charged (ami-
dines) or negatively charged (carboxylic acids, tetrazoles) at phys-
iological pH. The two chloro substituents on the phenyl rings were
chosen so as to maximise potency while limiting log P in the core
clogP 4.91
H2N
HO N
PSA 73.9
-1.11 (+19.4)
-1.06 (+26.1)
-0.21 (+34.9)
-0.18 (+30.7)
O
N
H2N
O
N
N
H2N
H
N
N
HN
HO
Cl
Cl
N
HN
-0.15 (+13.6)
N
N
O
Figure 2. Differences in c log P and PSA (in parentheses) by modifications of the nitrile group in 14 into amides, amidoximes, amidines, carboxylic acids, and tetrazoles.
H
N
NH2
1)
Li
O
O
O
1) LiHMDS
Cl
EtOH
COOEt
2) (COOEt)2
Cl
Cl
Cl
2) AcOH
O
I
O
OEt
OH
NC
1) NBS/AIBN
2) KCN/18-C-6
N
N
N
N
Cl
Cl
Cl
3) 1N NaOH
II
III
R1
R2
O
N
NC
1) EDAC/HOBt
or HOAT
N
R1
R2
N
Cl
Cl
2)
N
H
IV
Scheme 1.

28
M. Cooper et al. / Bioorg. Med. Chem. Lett. 20 (2010) 26–30
Table 1
Table 1 (continued)
Antagonism on hCB1 and lipophilicity of secondary and tertiary 4-cyanomethyl-1,5-
diphenyl-1H-pyrazole-3-carboxamides
No.
R
c log P
IC50 GTP
71
c
Sb (nM)
(log D)a
O
NC
R
15
4.4
6.2
N
N
N
O
H
N
N
Cl
Cl
N
H
16
17
0.49
0.52
No.
R
c log P
IC50 GTP
c
Sb (nM)
(log D)a
N
H
5.5
Rimonabant
6.5 (4.7)
4.5
H
N
114
5.6 (4.2)
4.6 (3.3)
11
57
N
H
F
18
19
5.3 (2.9)
6.2
1110
1.6
N
N
H
2
3
N
N
H
CH3
N
H
CF3
5.6
14
18
H
N
N
20
21
6.0
2.9
2.2
31
N
H
H
N
H
4
5
5.3 (4.4)
O
Cl
N
N
S
CH3
H
O
H
N
6.0 (4.0)
19
N
22
23
5.4
4.9
1.2
7.3
N
H
COOMe
CN
6
7
8
4.9 (3.8)
5.8
55
16
15
N
H
H3C
CH3
H
N
N
H
COOMe
CH3
3.9
N
H
24
25
5.2
4.7
4450
6250
N
H
N
O
COOH
H3C
9
6.0
0.41
CH3
N
H
N
H
COOH
HO
10
11
4.9
4.9
7.8
1.7
26
27
5.8
6.3
13
N
N
N
N
N
H
HO
1.0
F
N
H
N
N
N
28
4.0
3.8
12
13
14
5.5
5.8
1.9
7.5
N
H
29
30
5.9
4.7
9.2
3.0
F
N
N
5.7 (3.5)
4.9 (3.6)
N
H
OH
F
N
H
N

M. Cooper et al. / Bioorg. Med. Chem. Lett. 20 (2010) 26–30
29
comparator 14 to rimonabant displays comparable activity (4.5
vs 7.5 nM) despite the drop in log P and log D induced by the nitrile
moiety. Removal of the hydrazine nitrogen (13) leads to an
increase in potency, whereas the introduction of oxygen leads to
a 10-fold less active morpholine 15. Ring expansion with or
without hydrazine gives compounds with sub-nanomolar potencies
(16, 17). Notably, the insertion of an ethylene linker that was
allowed in the phenyl case (7) is not allowed in the piperidine case
(18). The more bulky amines 19 and 20 show potencies equal to
the cyclohexylamine 13, but the more polar sulfonamide 21 loses
activity. Thus, there are several examples where the introduction
of additional acceptor sites on the eastern amide moiety leads to
a reduction in activity. To explore this aspect further, carboxylic
acids 24 and 25 were included in the library and these were found
to have considerably reduced potencies compared to the corre-
sponding esters 22 and 23.
The library also contained a set of tertiary amides (26–42),
which provide complementary SAR information. The piperazinyl-
pyrimidine 28, containing several acceptor sites, displays high
potency comparable with the piperidinylphenyl compounds, either
with or without hydroxyl groups, that is compounds 30 and 27,
respectively. However, the piperazine 26 is less active than the
more polar 28. Linkage with pyrrolidine (29) seems less attractive
while appending the phenyl system in the 2-position of the pyrrol-
idine (37) or piperidine (31) leads to a greater loss in potency.
Some ring annelated compounds show pronounced potency with
a positive influence from lipophilic electron withdrawing groups
(cf. 32 vs 33 and 34 vs 35). The methylsulfone 39 is comparable
to 26 whereas the more polarized acetyl derivative 40 is not toler-
ated. We also confirmed the finding of the deleterious effect of a
carboxyl moiety in this series (41 vs 42).
Table 1 (continued)
No.
R
c log P
IC50 GTPc
Sb (nM)
(log D)a
31
7.1
23
N
F
N
32
33
34
35
36
6.1
7.0
5.9
6.5
4.8
5.7
2.1
5.7
0.22
3.4
CF3
N
F
N
Br
N
N
OH
37
6.6
27
N
F
F
O
N
38
39
5.8
3.8
20
17
The substituent effects on a set of benzylic amides were
explored (Table 2). Compounds with a single halogen substituent,
either para (43–45), meta (46) or ortho (47), were found to be com-
parable. Unexpectedly the 2,4-dichloro substituted compound 49
showed no potency advantage over 44, whereas the 2,6-dichloro
substituted compound 50 was worse — presumably due to steric
effects. However, certain other disubstitution patterns (51, 52)
were found to be more favourable, with the significant difference
in potency between the para bromo compounds 45 and 51 being
attributable to the ortho fluoro group. In terms of more polar mo-
tifs, it was noteworthy to observe that the para cyano compound
O
N
S
CH3
O
O
N
N
40
41
3.3
6.2
199
2.2
CH3
MeOOC
HOOC
Table 2
Antagonism on hCB1 and lipophilicity of 4-cyanomethyl-1,5-diphenyl-1H-pyrazole-
3-carboxamides
N
42
4.8
2810
O
NC
N
R1
R2
a
c log P was calculated and log D chromatographically measured according to
Ref. 15.
N
N
b
Antagonistic activity in GTP
c
S assay using CP55940 as agonist and membranes
Cl
Cl
produced from CHO-K1 cells stably transfected with hCB1 receptor. All compounds
displayed efficacy above 100% indicating inverse agonist activity and all values are
mean of at least two determinations.
No.
R1
R2
c log P (log D)a
IC50 GTPc
Sb (nM)
43
44
4514
46
47
48
49
50
51
52
53
4-F
4-Cl
4-Br
3-Cl
2-F
4-CN
2,4-Cl2
2,6-Cl2
2-F, 4-Br
4-F, 3-Br
4-F
H
H
H
H
H
H
H
H
H
H
5.8 (4.1)
6.3 (4.4)
6.5 (4.4)
6.3 (4.4)
5.8 (4.1)
5.0 (3.8)
7.0 (4.6)
7.0 (4.6)
6.6 (4.5)
6.6 (4.4)
5.9 (4.2)
2.1
1.2
5.8
2.6
4.1
4.2
1.6
7.0
0.51
0.70
5.6
to 1 (previously described in Ref. 14) but the activity can be re-
gained by the introduction of a trifluoromethyl (3) or chloro (4)
group in para or ortho position, resepctively. The phenyl moiety
can be moved further by insertion of an ethylene linker as in 7
without loss in activity compared to 1. Also the more polar isoxaz-
ole compound 8 has a comparable activity. However, the rigid in-
dan system in 9 leads to high potency. Interestingly, one of the
enantiomeric hydroxyindan derivatives, that is, 11, displays activ-
ity comparable to 9 whereas the other (10) drops in activity. A
number of aliphatic amines were also investigated and the direct
CH3
a
b
and as in Table 1. Values are single or mean of double determinations.

30
M. Cooper et al. / Bioorg. Med. Chem. Lett. 20 (2010) 26–30
5. Medicinal chemistry reviews: (a) Lee, H.-K.; Choi, E. B.; Pak, C. S. Curr. Top. Med.
Chem. 2009, 9, 482; (b) Jagerovic, N.; Fernandez-Fernandez, C.; Goya, P. Curr.
Top. Med. Chem. 2008, 8, 205; (c) Muccioli, G. C.; Lambert, D. M. Curr. Med.
Chem. 2005, 12, 1361.
48 had the same potency as the halogen derivatives. Finally, com-
parison of 43 and 53 implied that the amide hydrogen is not an
essential interaction feature.
6. Scheen, A. J.; Paquot, N. Best Pract. Res. Clin. Endocrinol. Metab. 2009, 23,
103.
7. (a) Rosenstock, J.; Hollander, P.; Chevalier, S.; Iranmanesh, A. Diabetes Care
2008, 31, 2169; (b) Després, J. P. Arterioscler. Thromb. Vasc. Biol. 2009, 29, 416;
(c) Després, J. P. Curr. Pharm. Des. 2009, 15, 553.
8. Caraceni, P. Best Pract. Res. Clin. Endocrinol. Metab. 2009, 23, 65.
9. (a) Moreira, F. A.; Grieb, M.; Lutz, B. Best Pract. Res. Clin. Endocrinol. Metab. 2009,
23, 133; (b) Viveros, M. P.; deFonseca, F. R.; Bermudez-Silva, F. J.; McPartland, J.
M. Endocr. Metab. Immune. Disord. Drug Targets 2008, 8, 220.
10. Viveros, M. P.; Marco, E. M.; Llorente, R.; López-Gallardo, M. Neural Plast. 2007,
2007, 52908.
These compounds also act as inverse agonists, for example, the
bromo derivative 45 has an IC50 of 1.2 nM in the GTPcS assay. A
selectivity of greater than 1000-fold for CB1 over CB2 was demon-
strated, for a selection of compounds, by testing against the CB2
sub-type receptor in a GTPcS antagonist assay using CP55940 as
agonist (data not shown).
In conclusion, we have explored the CB1 antagonist SAR features
for a chemically diverse library of secondary and tertiary 4-cyano-
methyl-1,5-diphenyl-1H-pyrazole-3-carboxamides. In general
a
11. Kunos, G.; Osei-Hyiaman, D.; Bátkai, S.; Sharkey, K. A.; Makriyannis, A. Trends
Pharmacol. Sci. 2009, 30, 1.
12. Lan, R.; Liu, O.; Fan, P.; Lin, S.; Fernando, S. R.; McCallion, D.; Pertwee, R.;
Makriyannis, A. J. Med. Chem. 1999, 42, 769.
range of amide substitutions are tolerated, but polar and especially
charged groups in this part of the molecule reduce the CB1 antago-
nistic activity. Some very potent eastern modifications have been
identified which will form the basis of further work to identify less
lipophilic compounds that do not readily pass the blood–brain-
barrier.
13. The parallel library synthesis was conducted according to either of two
protocols (a) or (b), as follows. (a) Larger scale using 9 parallel tubes: 1-
Hydroxybenzotriazole (HOBt, 47 mg, 1.3 equiv) and 1-ethyl-3(3-dimethyl-
aminopropyl)carbodiimide (EDAC, 77 mg, 1.5 equiv) were added to a stirred
solution of acid III (100 mg, 0.27 mmol, 1 equiv) in 15 ml CH2Cl2. After 20 min,
the appropriate amine (1.2 equiv) and diispropylethylamine (DIPEA, 77 mg,
1.3 equiv) were added and stirring continued overnight. Saturated NaHCO3
solution was added and the reaction mixture extracted with CH2Cl2. The
organic phase was washed with brine, dried (MgSO4) and evaporated by
vacuum centrifuge. Final purification was made using semi-preparative
reverse-phase HPLC or normal-phase column chromatography over silica; (b)
Plate chemistry using 1.4 ml tubes in 96 well format: Acid III (0.075 ml of
0.1 M CH2Cl2 solution, 1.5 equiv), PS-DCC (8.1 mg, 2 equiv), HOBt (0.085 ml of
0.1 M CH2Cl2 solution, 1.7 equiv) and PS-DIPEA (2.6 mg, 2 equiv) were mixed in
0.80 ml CH2Cl2. The plate was shaken for 25 min, the appropriate amines
(0.050 ml of 0.1 M N-methylpyrrolidone solution, 1 equiv) were added and
shaking continued for 3 days. Additional CH2Cl2 (0.15 ml) and PS-trisamine
(6.0 mg, 5 equiv) were added to each tube and shaking continued for a further
24 h. The resins were filtered and washed with CH2Cl2. Combined liquors were
evaporated and residues redissolved in DMSO (0.25 ml). Following LC–MS
analysis, the solutions were diluted with further DMSO so as to provide 10 mM
stock solutions for screening.
Acknowledgements
The authors thank Rokhsana Andersen and Helle Iversen for
excellent technical assistance.
References and notes
1. Endocannabinoid reviews: (a) Di Marzo, V. Pharmacol. Res. 2009, 60, 77; (b) De
Petrocellis, L.; Di Marzo, V. Best Pract. Res. Clin. Endocrinol. Metab. 2009, 23, 1;
(c) Di Marzo, V.; Ligresti, A.; Cristino, L. Int. J. Obesity 2009, 33, S18; (d) deKloet,
A. D.; Woods, S. C. Endocrinology 2009, 150, 2531.
2. CB1 reviews: (a) Janero, D. R.; Makriyannis, A. Expert Opin. Emerging Drugs
2009, 14, 43; (b) Di Marzo, V. Drug Discovery Today 2008, 13, 1026; (c) Akbas, F.;
Gasteyger, C.; Sjödin, A.; Astrup, A.; Larsen, T. M. Obes. Rev. 2009, 10, 58.
3. (a) Di Marzo, V. Diabetologia 2008, 51, 1356; (b) Mackie, K. Neuroendocrinology
2008, 20, 10.
14. Amengual, R.; Marsol, C.; Mayeux, E.; Sierra, M.; Wagner, P. Patent Appl. PCT
Int. Appl. WO 2006133926, 2006.
4. Ravinet Trillou, C.; Delgorge, C.; Menet, C.; Arnone, M.; Soubrie, P. Int. J. Obes.
2004, 28, 640.
15. Valko, K.; Du, C. M.; Bevan, C.; Reynolds, D. P.; Abraham, M. H. Curr. Med. Chem.
2001, 8, 1137.