1-sulfonyl-4-acylpiperazines as selective cannabinoid-1 receptor (CB1R) inverse agonists for the treatment of obesity

Article abstract of DOI:10.1021/jm900063x

A novel series of 1-sulfonyl-4-acylpiperazines as selective cannabinoid-1 receptor (CB1R) inverse agonists was discovered through high throughput screening (HTS) and medicinal chemistry lead optimization. Potency and in vivo properties were systematically

Full text of DOI:10.1021/jm900063x

2550
J. Med. Chem. 2009, 52, 2550–2558
1-Sulfonyl-4-acylpiperazines as Selective Cannabinoid-1 Receptor (CB1R) Inverse Agonists for
the Treatment of Obesity
Petr Vachal,*,† Joan M. Fletcher,† Tung M. Fong,‡ Cathy C. R.-R. Huang,‡ Julie Lao,‡ Jing C. Xiao,‡ Chun-Pyn Shen,‡
Alison M. Strack,§ Lauren Shearman,§ Sloan Stribling,§ Richard Z. Chen,‡ Andrea Frassetto,‡ Xinchun Tong,| Junying Wang,|
Richard G. Ball, Nancy N. Tsou, Gerard J. Hickey,§ Donald F. Thompson,§ Terry D. Faidley,§ Susan Nicolich,§
Joana Achanfuo-Yeboah,§ Donald F. Hora,§ Jeffrey J. Hale,† and William K. Hagmann†
Departments of Medicinal Chemistry, Metabolic Diseases, Pharmacology, Drug Metabolism/Pharmacokinetics, and Pharmaceutical Research
and DeVelopment, Merck Research Laboratories, Rahway, New Jersey 07065
ReceiVed January 16, 2009
A novel series of 1-sulfonyl-4-acylpiperazines as selective cannabinoid-1 receptor (CB1R) inverse agonists
was discovered through high throughput screening (HTS) and medicinal chemistry lead optimization. Potency
and in vivo properties were systematically optimized to afford orally bioavailable, highly efficacious, and
selective CB1R inverse agonists that caused food intake suppression and body weight reduction in diet-
induced obese rats and dogs. It was found that the receptor binding assay predicted in vivo efficacy better
than functional antagonist/inverse agonist activities. This observation expedited the structure-activity
relationship (SAR) analysis and may have implications beyond the series of compounds presented herein.
At least 4 of the top 10 causes of death in the U.S. and Europe
are medical conditions associated with obesity, including heart
disease, cancer, stroke, and type II diabetes.1 As the rate of the
overweight and obese population increases alarmingly world-
wide, so does the need to reduce the incidence of obesity, a
root cause of some of the more common morbidities. While
healthy food choices, exercise, and a balanced life style are, in
many cases, sufficient to prevent unhealthy weight gain, the
same paradigm does not apply to weight loss of already
overweight individuals. This paradox is mainly due to human
body homeostatic energy control, a complex system of signaling
and metabolic pathways responsible for promoting food intake
as a means of preventing starvation.2 To this end, a disruption
of the homeostatic energy control, in addition to a healthy life
style, is critical in reducing and maintaining reduced body
weight. One point of such intervention is regulation of cannab-
inoid-1 receptor (CB1Ra) function, which has been the subject
of intense study for several decades. While there is abundant
preclinical evidence linking CB1R inverse agonism/antagonism
to food intake suppression and body weight loss, this association
is most convincingly supported by the results of human clinical
trials of two CB1R inverse agonists, rimonabant3,4 and tarana-
bant.5 Herein, we describe the discovery and development of a
structurally diverse series of 1-sulfonyl-4-acylpiperidines as
CB1R inverse agonists.
assay that measures cyclic-AMP (cAMP) levels6,7 (EC50 ) 1
nM) but only modest activity in a CB1R competition binding
assay6,8 (IC50 ) 260nM). It had no in vivo efficacy in an
overnight diet-induced obese (DIO) rat assay.6,9 The disparity
between the binding and functional activities offered the
opportunity to explore which of the two in vitro assays might
be more predictive of the pharmacodynamic (PD) in vivo
outcome. While the absence of in vivo efficacy of 1 did not
allow for any firm conclusions because of its poor pharmaco-
kinetics (PK) in the rat,6,10 it was reasonable to assume that a
derivative of 1 with good PK might elucidate the disconnect
between the two in vitro assays. An SAR analysis of close
analogues of 1 validated the lead by establishing that its high
functional activity was not due to the 3,5-bis(trifluoromethyl)-
benzenesulfonyl subunit, which has produced false positives
from an HTS in other programs (Table 1, 2-6). These results
also confirmed that compounds 1-6 displayed the curious
separation between their functional and binding in vitro activi-
ties. Compounds 1-6 did not exhibit the desired PK properties
suitable for definitive evaluation in PD assays and thus for
further assessment.
The chemical synthesis of derivatives of 1 was accomplished
by reaction of 1-Boc-piperazine with the first coupling partner
(either sulfonyl chloride R1SO2Cl or carboxylic acid R2CO2H
activated by EDC), followed by deprotection of the Boc group
with hydrogen chloride and subsequent reaction with the second
coupling partner (either sulfonyl chloride R1SO2Cl or carboxylic
acid R2CO2H activated by EDC) as outlined in the scheme in
Table 1.
Initial optimization of 1 was primarily based on the two in
vitro CB1R assays. Particular emphasis was placed on the
binding assay, since many compounds were active in the
functional assay and provided no real distinction for the SAR.
It is noteworthy that all compounds in this series exhibited
selectivity against CB2R with CB2R/CB1R in vitro activity
ratios always exceeding 100 but more typically more than 1000
(e.g., hCB2 IC50 values exceeded 10 µM for all key compounds
depicted in Table 1, 1-5, 33-36). The SAR of the CB1R in
vitro assays was found to be additive with respect to the
structural subunits of the series, namely, the sulfonyl group,
Compound 1 (Table 1) was identified through an HTS of
compounds from the Merck sample collection. Interestingly, 1
exhibited potent inverse agonist activity in a CB1R functional
* To whom correspondence should be addressed. Phone: +1-732-594-
6624. Fax: +1-732-594-2210. E-mail: petr_vachal@merck.com.
†
Department of Medicinal Chemistry.
Department of Metabolic Diseases.
Department of Pharmacology.
Department of Drug Metabolism/Pharmacokinetics.
‡
§
|
Department of Pharmaceutical Research and Development.
a Abreviations: CB1R, cannabinoid-1 receptor; HTS, high-throughput
screening; SAR, structure-activity relationship; DIO, dietary-induced obese;
EDC, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride; DI-
PEA, N,N-diisopropylethylamine; MED, minimal effective dose; BP, brain
penetration; BW, body weight; FI, food intake; cAMP, cyclic adenosine
monophosphate; PK, pharmacokinetic; PD, pharmacodynamic; RPHPLC,
reverse phase high-pressure liquid chromatography.
10.1021/jm900063x CCC: $40.75
2009 American Chemical Society
Published on Web 03/25/2009

1-Sulfonyl-4-acylpiperazines
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8 2551
Table 1. CB1R Binding and Functional Activities of 1-Sulfonyl-4-acylpiperazines
a All compounds were inverse agonists as described in ref 6.

2552 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8
Vachal et al.
Figure 1. Design strategy of conformationally constrained CB1R inverse agonist 38.
Figure 2. Effect of 38 on the overnight food intake and body weight change in DIO rats. A statistically significant food intake reduction of 59%
vs vehicle was observed along with statistically significant body weight change vs vehicle, consistent with a similar effect caused by taranabant.
Table 2. Correlation of CB1R in Vitro Binding and Functional Assay to the PD DIO Rat in Vivo Efficacy Assay: PD Efficacy Consistent with Binding
CB1R in Vitro Assay
39
36
38
human
rat
human
rat
human
rat
Binding
CP-55940,a IC50 (nM)
SR-141716,b IC50 (nM)
58
47
175
227
1.2
0.5
27
11
0.30
0.10
2.2
0.20
Functional Activity
agonist EC50 (nM) (% max)
1.3 (-98%)
172
4.4 (-115%)
1.1 (-151%)
334
0.6 (-145%)
5.8 (-106%)
156
3.4 (-125%)
antagonist MeAc (% inhibition at 2 µM)
MeAc (300 nM), IC50 (nM)
77
1.9
9.2
PD @ 3mpk (DIO rat)
food intake change vs vehicle (%)
absolute body weight change (g)
-14 (NSd)
-40
-2
-59
-14
+8 (NSd)
a Competition CB1R binding against CP-55940. b Competition CB1R binding against SR-141716. c Methanandamide. d NS ) not statistically significant.
the center piperidine, and the acyl group. As such, modifications
of the individual subunits did not alter the privileged structural
features identified for the remaining parts of the molecule,
rendering in vitro assays very predictable when key structural
features were interchanged. Structural features critical for CB1R
activity included the following:

1-Sulfonyl-4-acylpiperazines
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8 2553
Figure 3. Metabolism of 38. Two major metabolites were observed, 38M1 and 38M2.
Table 3. Derivatives of 40: Phase II Metabolism Studies in Human Hepatocytes (in Vitro)
compd
RO-
HO-
MeO-
CB1R IC50 (nM)
phases II metabolite 46 observed?a
40
41
42
43
44
45
6
yes, exclusively
trace
no
yes
yes
yes
0.5
0.3
8
8
2
i-PrO-
t-Bu-OCO-
EtOCO-
Me2NOCO-
a In vitro incubation in hepatocytes conducted across four species (rat, dog, monkey, and human), observation of phase II metabolite in human hepatocytes
indicated therein.
(1) Sulfonamide. It was not possible to substitute the
sulfonamide for conformationally less rigid functional groups
or replace it with rigid isosteres, such as carboxamide.
(2) Piperazine Ring. Replacement of the central piperazine
with piperidine, pyrrolidine, or either regioisomer of 3-ami-
noazetidine produced inactive compounds (Table 1, 7-9 vs 33).
It was found that the piperazine ring underwent metabolic
oxidation as assessed in liver microsome incubations. In an
attempt to improve the metabolic stability of the piperazine ring,
2- and 3-substituted analogues were prepared. While monosub-
stitution and disubstitution of the 3-position with relatively small
groups (10, 12) were tolerated, 2-substitution and 3,5-disubsti-
tution generally reduced CB1R activity (13-17). More impor-
tantly, there was no improvement in PK properties by piperazine
substitutions.
(3) Carboxamide. Replacement with methylene (i.e., N-
(cyclohexylmethyl) derivative) or complete omission of the
carboxamide functionality (i.e., N-alkyl- and N-arylpiperidine
derivatives) failed to display any CB1R activity in vitro. It is
also noteworthy that the carboxamide was not interchangeable
with sulfonamide, implying that its role extended beyond just
structural rigidity (analogue of 1 in which the carboxamide was
replaced with sulfonamide was inactive up to 10 µM in both in
vitro assays). Along with the need for presence of the sulfon-
amide on the second piperazine nitrogen, these data suggested
that the molecules bound to the CB1R in a specific orientation
and the two sides of the molecule were not interchangeable.
Carboxamides derived from aromatic carboxylic acids were
found to exhibit inferior binding compared to their aliphatic
counterparts (e.g., Table 1, compare 1 vs 18). On the other hand,
CB1R active aliphatic carboxamides represented a metabolic
liability as they underwent rapid oxidation of their aliphatic side
chains. This liability was overcome by introducing metabolically
more robust aromatic substituents separated from the carbonyl
functionality with a methylene spacer. The resulting hybrid
derivative (e.g., 19) proved much more stable (both in vitro in
liver microsomes and in vivo PK) while maintaining strong
CB1R functional properties in vitro. Metabolic stability of the
benzylic R-carboxamide methylene could be further enhanced
by substitution (20, 21) or by deactivation with electron-
withdrawing functionalities (22-24). The latter also proved
beneficial for CB1R binding and functional activity (compare
24 vs 19). Electron-withdrawing groups in the 4-position were
found to be superior relative to 2- and 3-substitutions (24 vs 25
and 26).
(4) Arylsulfonamide. While good in vitro activity was limited
to arylsulfonamides, some degree of flexibility was observed
with a range of aryl substituents (Table 1, 1-6). Initially, 3,5-
disubstitution of the aryl ring proved to be superior to other
mono-, bis-, and trisubstitution patterns (e.g., 6 vs 27-31; 1 vs
32). Generally, electron-deficient arylsulfonamides were slightly
more active in vitro and also enhanced piperazine ring oxidative
stability, thus providing improved overall PK profiles as
evidenced by 33 (rat PK: Clp ) 37 (mL/min)/kg, Vdss ) 9.6
L/kg, t1/2 ) 6.1 h, F ) 20%). To this end, a single 3-trifluoro-
methyl substituent of arylsulfonyl ring caused the aromatic ring
to be sufficiently stable to rapid oxidation. Of several 5-posi-
tional derivatives of 33, exemplified by 34 and 35, the best
combination of potent in vitro activity with modest in vivo
stability was observed for the 5-cyclopropyl-3-trifluoromethyl
derivative 36 (Tables 1 and 2).
The additive nature of the SAR for individual subunits of
the lead structure was further exploited to advance the series.
Initially, while compound 36 was the best example of balanced
functional and binding CB1R in vitro activities, its overall
properties were still suboptimal, as it afforded only a modest
food intake reduction in the DIO rat assay. Second, the role of
R-methylene substitution was evaluated by resolving the racemic
R-methyl compound 20 (Table 1). The two resulting enanti-
omers, 20a and 20b, displayed vastly different activities in vitro
(Figure 1). Finally, compound 37 exhibited in vitro profile
similar to 20a, suggesting that a two carbon spacer might be
superior to the one carbon spacer. The inherent potential of the
ethylene linker might also have been somewhat compromised
by its flexibility. The combination of these three features
(namely, the 3,5-disubstitution pattern in 36, the configuration

2554 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8
Vachal et al.
Table 4. Derivatives with the Potential for Formation of Phase II
Conjugates in Hepatocytes
to vehicle treated animals. This result finally validated the
potential of this series of compounds (Figure 2). The ability of
38 to reduce body weight was further demonstrated in a DIO
dog model (2% overall body weight reduction at 2 (mg/kg)/
day po after 5 days; data not shown).6
While the magnitude of the functional and binding CB1R in
vitro assays converged for some of the most active compounds
of the series (Table 1, 34-36), the disparity of the two assays
observed for the majority of examples listed in Table 1 remain-
ed of interest. An effort was devoted to connect one of the two
often divergent assays to the primary PD assay and thus to in
vivo efficacy. Table 2 outlines a series of three closely related
analogues (36, 38, and 39) of which most in vitro and in vivo
characteristics were found to be remarkably similar. All three
compounds had similar functional activity in vitro (EC50) and
were similarly potent inverse agonists. They had similar rat PK
properties and similar brain penetration (brain/plasma ratio, or
b/p), an important attribute given the strong CB1R expression
in the brain. The three compounds differ only in the values of
the in vitro binding potency (IC50) in competition assays
employing different ligands ([3H]CP-55,940 or [3H]SR141716).
The significant differences among the compounds enabled a
convincing comparison of otherwise remarkably similar com-
pounds, differing only in their binding. The effect on body
weight decrease and food intake suppression in the DIO rat
model increased with increased binding potency: 38 > 36 > 39,
respectively. The correlation was consistent with the binding
assay being more reliably predictive of in vivo efficacy. This
conclusion not only rationalized the lack of efficacy of the many
strongly functionally active compounds outlined in Table 1 but
it also expedited subsequent SAR analysis, enabling us to focus
on binding rather than functional activity as the more reliable
predictor of compounds which may be more suitable for further
in vivo evaluation.
Isolation of metabolites in rats and dogs established oxidation
of the piperazine ring as the primary metabolism of 38 (Figure
3). Two major metabolites, 38M1 and 38M2, were observed,
neither of which had any significant activity against CB1R in
vitro. While plasma half-life of the acid 38M2 was short-lived
and cleared rapidly, the primary sulfonamide 38M1 accumulated
in circulation in both rats and dogs. Comparison of human
metabolism to rat and dog in vitro suggested that a similar
profile, including long-lived 38M1, was likely to be observed
in humans. The potential of encountering serious issues associ-
ated with a long-lived metabolite such as 38M1 in late stage
development prompted a focused effort to alter the PK profile
of the lead by substituting the arylsulfonylamide with a
metabolically labile functional group which might prevent
metabolite accumulation by increasing its rate of clearance. It
was clear from the SAR that such a functionality should be
introduced in the 5-position of arylsulfonamide, replacing the
cyclopropyl residue, in order to maintain both CB1R activity
and overall in vivo PK stability. Both 3,5-substitution with at
least one being a CF3 group had been established (see discussion
of Table 1). Initially, derivatives bearing functionality capable
of undergoing phase II metabolism in vitro (i.e., forming a
glucuronide adduct) were surveyed.
a hCB2 IC50 exceeded 10 µM for all compounds reported in Table 4.
b Acute overnight body weight change in DIO rat at 3 mpk. c Not statistically
significant vs vehicle. d MK0499 binding assay.
Phenol 40 underwent rapid glucuronidation when incubated
in hepatocyte homogenates from several species (rat, dog,
monkey, and human) to produce glucuronide 46 as the major
metabolite (Table 3). While phenol 40 could not be considered
as a development candidate for numerous reasons, including
poor oral bioavailability in rats, derivatives of 40 that would
release the parent phenol 40 in vivo as well as any primary
preference in 20a, and the two carbon spacing of 37) led to the
design of conformationally constrained 38 which provided
optimal in vitro binding and functional properties and signifi-
cantly improved PK and brain penetration in the rat. Compound
38 caused a robust overnight food intake reduction in DIO rats
resulting in a very significant body weight decrease compared

1-Sulfonyl-4-acylpiperazines
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8 2555
Figure 4. Overlay of compound 50 (yellow) with MK0499 (green) in a model of the potassium hERG channel indicated a significant degree of
similarity of the two compounds consistent with a strong hERG activity observed for 50.
Scheme 1. Synthesis of 65 via Distereoselective Cyclopropanation and Direct 1,3,4-Triazole Formationa
a Conditions: (a) EDC, DIPEA; (b) Me3S(O)I, NaH, DMSO, 40°C, E/Z > 50:1; (c) 4 M HCl in dioxane; (d) sulfonyl chloride (see scheme for its structure),
aq NaHCO3, DCM; (e) chiral separation, Chiracel AD, 35% yield (five steps), single enantiomer; (f) Zn(CN)2, Pd2(dba)3, dppf, DMF, 115°C, 74%; (g)
RCONHNH2, DMF/NMP (1/1), 140 °C.
sulfonamide metabolite analogous to 38M1 (Figure 3) were
prepared. Simple alkyl ethers (41, 42), which were affording a
robust body weight reduction in DIO rats, did not undergo
efficient dealkylation to produce 40. On the other hand,
carbonates 43 and 44 and carbamate 45 all rapidly produced
40 and thus glucuronide 46 but were only marginally efficacious
in the DIO rat model.
Alcohols 47-50 represented another class of derivatives with
the potential to form phase II conjugates (Table 4). Trifluoro-
methyl alcohol 50 produced trace amounts of a glucuronide
adduct in hepatocyte incubations and was efficacious in the acute
DIO rat model at 3 mg/kg po. Counterscreening revealed a very
strong affinity of 50 to potassium channel (hERG) and as such
was not considered as a viable lead. It served, however, as an

2556 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8
Vachal et al.
Novel chemistry was developed for the synthesis of 65. A
reaction of nitriles with hydrazides should theoretically produce
the desired 1,3,4-triazoles upon heating. However, we found
no example of such reactivity described in the literature.
Intrigued by the potential simplicity of this single-step process,
we screened several solvents and found that heating formylhy-
drazide with nitrile 73 in a DMF/NMP mixture produced the
desired 65 in 81% isolated yield (Scheme 1). This novel
reactivity extended to other substrates as exemplified by reaction
of 73 with acethydrazide to afford 66. While the reaction is
likely limited to strongly electron-deficient nitriles, it does
provide an attractive solution for derivatives of 65 serving as
the primary synthetic route to multigram quantities of the
compound (Scheme 1). Another noteworthy chemical reactivity
arose from the application of the Corey-Chaykovsky cyclo-
propanation11 to amides. Cyclopropanation of esters, for which
the Corey-Chaykovsky cyclopropanation was developed and
mostly utilized, proceeds generally with relatively low levels
of distereomeric control. When applied to cyclopropanation of
methyl cinnamates in the context of our series, an E/Z-
cyclopropane ratio of 2/1 was obtained. However, when
cyclopropanation was applied to amide 69, the reaction pro-
ceeded with excellent selectivity forming the desired E-isomer
exclusively (E/Z > 50/1). This discovery significantly expedited
preparation of a larger quantity of 65 necessary for full in vivo
and in vitro characterization (Scheme 1). Finally, absolute
stereochemistry of 65 was determined as (R,R) by means of
single X-ray crystallography analysis (Figure 5).
Figure 5. X-ray crystallography analysis determined the absolute
stereochemistry of the cyclopropane ring of 65 as (R,R).
important tool in establishing key features responsible for its
high affinity for hERG channel. Figure 4 depicts a computational
docking model comparing 50 to MK0499, a strong ligand of
hERG used as a standard in many hERG binding assays. This
modeling exercise suggested qualitatively that a polar R
substituent might disrupt the high affinity for hERG resulting
from the apparent binding interaction of the trifluoromethyl
group of 50 with lipophilic residues of hERG.
Heterocycles containing polar functionalities represent at-
tractive targets by having the potential for exhibiting both phase
II conjugation and lowering the observed hERG activity.
Numerous heterocyclic derivatives were prepared and found to
differ vastly in their ability to interact with CB1R (Table 4,
51-66). Compounds active in the binding assay were tested
for their ability to form phase II conjugates by incubations in
human hepatocytes. Contrary to our expectations, the vast
majority of heterocycles failed to produce phase II conjugates.
Only four heterocyclic derivatives that displayed both in vitro
binding to CB1R and the potential for glucuronidation were
identified (56, 63-65). Of these four, only 56 and 65 proved
efficacious in the acute overnight DIO rat model, causing robust
body weight change and food intake suppression at a dose of 3
mg/kg po. The 1,3,4-triazole derivative 65 had a superior profile
demonstrating CB1R potency, activity in DIO rats, and reduced
hERG activity (IC50 > 1 µM).
Triazole 65 was evaluated in a series of in vivo and in vitro
experiments (Figure 6). The minimal effective dose (MED)
causing statistically significant body weight reduction and food
intake suppression in DIO rat was determined to be 0.3 mpk.
1,3,4-Triazole 65 exhibited a very clean profile in a broad
counterscreen, displaying no off-target activities (IC50 g 1 µM).
The strong signal of the primary glucuronide of 65 formed in
hepatocyte homogenates across four species was anticipated to
provide a sufficient clearance mechanism for 65 and its potential
sulfonamide metabolites. Pharmacokinetic studies of 65 estab-
lished good profile in rats and dogs but revealed an unexpectedly
low oral bioavailability in rhesus monkey (Figure 6). Metabolic
stability of 65 in liver microsomal incubations revealed that the
Figure 6. In vitro and in vivo characterization of 65. Primary PD assay established MED as 0.3 mpk. Off-target activity was assessed by PAN
laboratory screen, CYP panel inhibition assays, and hERG (vs MK499). In vivo characterization included PK across three preclinical species and
brain PK in the rat.

1-Sulfonyl-4-acylpiperazines
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8 2557
added sequentially with stirring. The reaction mixture was stirred
at room temperature for 2 h, concentrated, and purified using
RPHPLC to afford 1-(cyclohexylcarbonyl)-4-(1-naphthylsulfo-
nyl)piperazine 4 as colorless oil (0.73 mmol, 66% yield): LCMS
stability of 65 in human microsomes fell between that of dogs
and rhesus monkeys. As such, these results raised some
ambiguity about the stability of 65 in human. The potential risk
of poor PK in humans ultimately rendered the final decision to
discontinue further investigation of the piperazine series as a
treatment of obesity in human.
In conclusion, a novel series of highly efficacious and target
specific CB1R inverse agonists, exemplified by 65, was
developed. Numerous compounds in this series caused signifi-
cant body weight reduction and food intake suppression in two
preclinical species, DIO rats and dogs. During the development
process, an important causality link between the primary binding
and in vivo efficacy was established, not only effectively
advancing the series described herein but providing a basis for
expedited SAR optimization for other CB1R inverse agonist
leads. Careful profiling of in vitro and in vivo characteristics,
including receptor potency, off-target activity, and designed
metabolic pathways for clearance, led to the identification of a
unique series of potent inhibitors of the CB1R with many
desirable properties.
1
(method A) m/z (MH)+ ) 387.1 at 2.19 min; H NMR (CDCl3,
500 MHz) δ 8.8 (d, J ) 8.5 Hz, 1H), 8.2 (d, J ) 7.5 Hz, 1H), 8.1
(d, J ) 8.3 Hz, 1H), 7.9 (d, J ) 8.0 Hz, 1H), 7.7 (t, J ) 8.4 Hz,
1H), 7.6 (m, 2H), 4.7 (m, 1H), 3.6 (m, 4H), 3.5 (m, 4H), 3.2 (m,
4H), 1.4 (m, 4H), 1.2 (m, 6H).
Detailed Experimental Procedure of the Key Compound
65. 69. N,N-Diisopropylmethylamine (0.3 mol) was added to a
mixture of (2E)-3-[4-(trifluoromethyl)phenyl]acrylic acid (67, 21.6 g,
0.10 mol), 1-(1-naphthylsulfonyl)piperazine (68, 0.10 mmol), N-(3-
dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (edc
hydrochloride, 2.0 mmol) in anhydrous dichloromethane (300 mL)
via syringe at room temperature over 10 min and allowed to stir at
ambient temperature for additional 30 min. The mixture was
transferred into a separatory funnel, diluted with additional dichlo-
romethane (300 mL), and washed sequentially with 1 M aqueous
hydrochloric acid (500 mL), 1 M aqueous sodium hydroxide (500
mL), and brine (500 mL), dried over sodium sulfate, filtered, and
concentrated yielding 69 as 37 g of a white solid used in the
subsequent step without further purification (0.097 mol, 97%, >97%
pure by LCMS): LCMS (method A) m/z (MH)+ ) 285.1, 329.2 at
3.42 min.
70. Sodium hydride (60% w/w in mineral oil, 0.15 mol) was
added to a solution of trimethylsulfoxonium iodide (0.15 mol) in
DMSO (600 mL) at 40 °C over 5 min under an inert atmosphere
of nitrogen in oven-dried glassware. The reaction mixture was
stirred at ambient temperature for 30 min at 40 °C, and 69 (37 g,
0.097 mol) was added in a single portion. The resulting mixture
was heated to 40 °C for additional 15 h, poured onto ice (1 L), and
extracted with ether (1 L). The aqueous layer was extracted with
ethyl acetate (2 × 200 mL). Combined organic layers were washed
sequentially with 1 M aqueous hydrochloric acid (500 mL), 1 M
aqueous sodium hydroxide (500 mL), and brine (500 mL), dried
over sodium sulfate, filtered, and concentrated, yielding 70 as 36.8 g
of a white solid used in the subsequent step without further
purification (90% pure13 by LCMS): LCMS (method A) m/z (MH)+
) 299.2 at 3.54 min.
Experimental Section
General. Unless specified otherwise, all materials were purchased
1
from commercial sources and used as received. H and 13C NMR
spectra were recorded using a Bruker 400 MHz instruments.
Preparative reverse phase liquid chromatography (RP HPLC) was
performed using Waters MS directed purification system consisting
of 2525 binary gradient pump, 2767 injector/collector, and 2996
PDA UV detector, mobile phase consisting of a gradient of water
and acetonitrile (each cont. 0.1% TFA), and a Waters Xterra column
(50 mm × 3 mm, 3.5 µm packing material). The purity of 1-72
has been determined by HPLC analysis (electrospray positive
ionization, Micromass ZQ single quadrupole)12 as g95% for 1-69
and 72 and as 90% for 70 and 71 (see the detailed experimental
procedure for the discussion of the purity of 70 and 71). Preparation
of compound 4 represents a typical procedure used for the synthesis
of compounds disclosed therein:
Sulfonamide Formation. To a solution of 1-naphthylsulfonyl
chloride (300 mg, 1.32 mmol) in anhydrous acetonitrile (25 mL),
triethylamine (2.6 mmol) and tert-butylpiperazine-1-carboxylate (2.6
mmol) were added sequentially with stirring. The reaction mixture
was stirred at room temperature for 2 h, concentrated, and the crude
oily residue was dissolved in ethyl acetate (100 mL). The solution
was washed sequentially with an aqueous solution of hydrochloric
acid (0.25 M, 100 mL) and aqueous solution of sodium hydroxide
(1 M, 50 mL), dried with sodium sulfate, and concentrated to afford
tert-butyl 4-(1-naphthylsulfonyl)piperazine-1-carboxylate as a white
solid (456 mg, 1.21 mmol, 91% yield): LCMS m/z (MH - Boc)+
) 277 at 2.21 min; 1H NMR (CDCl3, 500 MHz) δ 8.8 (d, J ) 8.7
Hz, 1H), 8.2 (d, J ) 7.5 Hz, 1H), 8.1 (d, J ) 8.0 Hz, 1H), 7.9 (d,
J ) 8.0 Hz, 1H), 7.6 (t, J ) 7.1 Hz, 1H), 7.6 (m, 2H), 3.5 (m, 4H),
3.2 (m, 4H), 1.4 (s, 9H).
Boc Deprotection. tert-Butyl 4-(1-naphthylsulfonyl)piperazine-
1-carboxylate (400 mg, 1.22 mmol) was dissolved in a 20% (v/v)
solution of trifluoroacetic acid in dichloromethane (10 mL) and the
resulting solution stirred at room temperature for 2 h and concen-
trated. The resulting crude residue was dissolved in dichlo-
romethane, washed with 1 M solution of sodium hydroxide, dried
with sodium sulfate, and concentrated to afford 1-(1-naphthylsul-
fonyl)piperazine as a white solid (256 mg, 1.12 mmol, 92%): LCMS
m/z (MH)+ ) 277; 1Η ΝMR (CDCl3, 500 MHz) δ 8.8 (d, J ) 8.4
Hz, 1H), 8.2 (d, J ) 7.3 Hz, 1H), 8.1 (d, J ) 8.2 Hz, 1H), 7.9 (d,
J ) 7.8 Hz, 1H), 7.6 (t, J ) 6.9 Hz, 1H), 7.5 (m, 2H), 3.2 (m, 4H),
2.9 (m, 4H).
Carboxamide Formation (4). To a solution of 1-(1-naphthyl-
sulfonyl)piperazine (250 mg, 1.1 mmol) in anhydrous dichlo-
romethane (10 mL), cyclopropylcarboxylic acid (2.0 mmol) and
N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (edc
chloride, 2.0 mmol) and diisopropylethylamine (5.0 mmol) were
71. A 4 M solution of hydrogen chloride in dioxane (400 mL)
was added to 70 (36.8 g 90% purity) and the resulting mixture
was stirred at ambient temperature for 0.5 h, concentrated, and
flushed with toluene (0.5 L), yielding 71 as a clear oil: LCMS
(method A) m/z (MH)+ ) 299.2 at 2.23 min.
3-Bromo-5-(trifluoromethyl)benzenesulfonyl chloride (0.11 mmol)
was added to a vigorously stirred mixture of 71, dichloromethane
(500 mL), and saturated aqueous solution of sodium bicarbonate
(500 mL). The resulting mixture was stirred at ambient temperature
for 1 h, the organic layer was seprated, and the aqueous layer was
extracted with dichloromethane (200 mL). Combined organic layers
were dried over sodium sulfate, filtered, and concentrated. The crude
product was purified by a filtration through a silica gel column
(eluent: gradient of ethyl acetate in heptane 0-50%) and concen-
trated, yielding 42 g of white solid. Chiral separation was performed
using Chiracel AD column (eluent: 12% solution of 2-propanol in
heptane); the desired (R,R)-enantiomer 71 was collected as the
second eluted isomer (20.2 g, 0.035 mol, >95% purity by LCMS,
35% overall yield from 67). LCMS (method A) m/z (MH)+ ) 585,
1
587 at 3.98 min; H NMR (CDCl3, 500 MHz) δ 1.37 (m, 1H),
1.66 (m, 1H), 1.90 (m, 1H), 2.49 (m, 1H), 3.22 (m, 4H), 3.60 (m,
4H), 7.20 (d, J ) 16 Hz, 2H), 7.50 (d, J ) 16 Hz, 2H), 8.50 (s,
1H), 8.68 (s, 1H), 8.80 (s, 1H).
72. A mixture of 71 (2.93 g, 5 mmol), zinc(II) cyanide (5 mmol),
1,1′-bis(diphenylphosphine)ferrocene (0.27 mmol), tris(dibenzylide-
neacetone)dipalladium (0.11 mmol), N,N-dimethylformamide (99
mL), and water (1 mL) was combined, and the resulting mixture
was degassed with a stream of nitrogen bubbled through the solution
at ambient temperature with stirring for 1 h and then heated to 115
°C for 30 min. The resulting solution was combined with ether
(200 mL) and washed with deionized water (2 × 100 mL), dried

2558 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8
Vachal et al.
K. M.; Lao, J.; Yu, H.; Feng, Y.; Xiao, J. C.; Van der Ploeg, L. H. T.;
Goulet, M. T.; Hagmann, W. K.; Lin, L. S.; Lanza, T. J.; Jewell, J. P.;
Liu, P.; Shah, S. K.; Qi, H.; Tong, X.; Wang, J.; Xu, S. S.; Francis,
B.; Strack, A. M.; MacIntyre, D. E.; Shearman, L. P. Antiobesity
efficacy of a novel cannabinoid-1 receptor inverse agonist, N-[(1S,2S)-
3-(4-chlorophenyl)-2-(3-cyanophenyl)-1-methylpropyl]-2-methyl-2-
{[5-(trifluoromethyl)pyridin-2-yl]oxy}propanamide (MK-0364), in
rodents. J. Pharmacol. Exp. Ther. 2007, 321, 1013–1022.
with sodium sulfate, concentrated, and purified using column
chromatography on silica gel (Biotage 40S, 10-100% ethyl acetate
in hexane) to afford 72 as a white solid (3.7 mmol, 74%): LCMS
1
(method A) m/z (MH)+ ) 532.1 at 3.58 min; H NMR (CDCl3,
500 MHz) δ 1.34 (m, 1H), 1.60 (m, 1H), 1.97 (m, 1H), 2.57 (m,
1H), 3.20 (m, 4H), 3.80 (m, 4H), 7.15 (d, J ) 16 Hz, 2H), 7.53 (d,
J ) 16 Hz, 2H), 8.55 (s, 1H), 8.71 (s, 1H), 8.80 (s, 1H).
(7) In vitro CB1R and CB2R functional activity of a test compound was
measured by incubating recombinant CB1R-CHO cells with various
concentrations of the test compound in the presence of 10 mM
forskolin, 200 mM phosphodiesterase inhibitor, 3- isobutyl-1-meth-
ylxanthine (IBMX) in the assay buffer (Earle’s balanced salt solution
supplemented with 5 mM MgCl2, 10 mM HEPES, pH 7.3, 1 mg/mL
BSA) at room temperature for 30 min. Cells were lysed by boiling,
and intracellular cAMP level was determined using the cAMP SPA
kit (Amersham). When adenylyl cyclase is activated by forskolin,
activation of Gi by CB1R in the presence of agonist (such as CP55940)
will lead to an inhibition of the forskolin-stimulated cAMP increase,
and inverse agonist will lead to a further increase of the forskolin-
stimulated cAMP increase. The maximal CP55940-mediated inhibition
of forskolin-stimulated cAMP increase is defined as 100% agonist
efficacy, and the intrinsic activity of all other compounds is relative
to the efficacy of CP55940. Negative efficacy denotes inverse agonism.
(8) In vitro CB1R and CB2R binding assays were performed by incubating
various concentrations of a test compound with 0.5 nM [3H]CP55940,
1.5 g of recombinant human CB1R-CHO membranes (or 0.1 mg of
human CB2R-CHO membranes) in 50 mM Tris-HCl, pH 7.4, 5 mM
MgCl2, 2.5 mM EDTA, 0.5 mg/mL fatty acid free bovine serum
albumin (BSA), 1× proteinase inhibitor mix (P8340, Sigma), and 1%
DMSO. After 1 h of incubation at 37°C, the reaction was stopped by
filtration and bound radioligand was separated from free radioligand
by washing the filter plate. Total specifically bound radiolabel was
10% or less of the total added radiolabel. Inhibitory IC50 values were
calculated through nonlinear curve fitting.
(9) Food intake assays. For automated food intake system, Sprague-
Dawley CD and DIO (dietary-induced obese Sprague-Dawley CD)
rats (obtained from Charles River Labs; Wilmington, DE) were caged
individually in Nalgene cages with metabolism feeders attached to
them. The food cups were external to the feeder and were placed on
individual balances. Each balance was connected to a central computer
that collects readings every 5 min to measure grams of food consumed.
Rats were transferred into the AFIS for at least 3 days before the
experiment to allow for acclimation. For overnight food intake (FI)
and body weight (BW) assay, male DIO rats were orally gavaged at
1 h prior to dark cycle onset with vehicle (10% Tween-80 in water)
or the CB1R inverse agonist. Rats were fed milled MHF diet during
acclimation to the caging and during the experiment. Overnight food
intake and body weight changes were measured. Significance level
was set at P < 0.05.
65. A mixture of 73 (1.60 g, 3 mmol), formylhydrazide (12
mmol), N,N-dimethylformamide (10 mL), and N-methylpyrolidine
(10 mL) was heated to 140 °C for 15 h, transferred to a silica gel
colum, and purified by a column chromatography (Biotage 40S,
10-100% ethyl acetate in hexane) to afford 65 as a white solid
(1.40 g, 2.43 mmol, 81%): LCMS (method A) m/z (MH)+ ) 574.2
at 3.42 min; 1H NMR (CDCl3, 500 MHz) δ 1.34 (m, 1H), 1.66 (m,
1H), 1.92 (m, 1H), 2.53 (m, 1H), 3.20 (m, 4H), 3.79 (m, 4H), 7.15
(d, J ) 16 Hz, 2H), 7.53 (d, J ) 16 Hz, 2H), 8.07 (s, 1H), 8.54 (s,
1H), 8.63 (s, 1H), 8.75 (s, 1H). X-ray analysis of 65 (Figure 5):
C24H21F6N5O3S, Mr ) 573.530, monoclinic, P21, a ) 4.8910(7)
Å, b ) 21.155(3) Å, c ) 14.292(2) Å, ꢀ ) 99.123(2)°, V )
1460.1(4) Å3, Z ) 2, Dx ) 1.304 g cm-3, monochromatized
radiation λ(Mo) ) 0.710 73 Å, µ ) 0.18 mm-1, F(000) ) 588, T
) 100 K. Data were collected on a Bruker CCD diffractometer to
a θ limit of 27.81° which yielded 19 071 reflections. There are 6900
unique reflections with 5126 observed at the 2σ level. The structure
was solved by direct methods (SHELXS-97; Sheldrick, G. M. Acta
Crystallogr., 1990, A46, 467-473) and refined using full-matrix
least-squares on F2 (Sheldrick, G. M. SHELXL-97. Program for
the Refinement of Crystal Structures, University of Go¨ttingen,
Germany). The final model was refined using 363 parameters and
all 6900 data. All non-hydrogen atoms were refined with anisotropic
thermal displacements. The final agreement statistics are as follows:
R ) 0.087 (based on 5126 reflections with I > 2σ(I)), wR ) 0.224,
S ) 1.15 with (∆/σ)max ) 0.03. The maximum peak height in a
final difference Fourier map is 1.402 e Å-3 and possibly associated
with disordered and unresolved solvent. CCDC contains the
supplementary crystallographic data for this paper (CCDC deposi-
tion number 721736). These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.ccdc.
cam.ac.uk/data_request/cif.
Supporting Information Available: Description of biological
assays, characterization of compounds 1-72, synthetic procedures,
and X-ray crystallographic analysis data of 65. This material is
available free of charge via the Internet at http://pubs.acs.org.
(10) Pharmacokinetic assays. Male Sprague-Dawley rats (Taconic Farms,
Germantown, NY) were dosed intravenously at 1 mpk and orally at 2
mpk for pharmacokinetic (PK) evaluations. The blood samples were
collected at various time points into lithium heparin tubes and
centrifuged. The plasma samples were kept at-70 °C until analysis.
For brain penetration (BP) study, the male Sprague-Dawley rats were
administered intravenously at 1 mpk and 1 h later euthanized by CO2
asphyxiation. The whole brain was collected, frozen on dry ice, and
stored at-70 °C. Blood was collected via cardiac puncture into lithium
heparin tubes and treated the same way as PK study. The plasma
samples were extracted by protein precipitation and analyzed by
LCMS2. The whole rat brain was homogenized, and was then extracted
by protein precipitation prior to LCMS2 analysis.
References
(1) Center of Disease Control Web site: http://www.cdc.gov/nchs/
FASTATS/lcod.htm.
(2) Morton, G. L.; Cummings, D. E.; Baskin, D. G.; Barsh, G. S.;
Schwartz, M. W. Central nervous system control of food intake and
body weight. Nature (London) 2006, 443, 289–295.
(3) (a) Di Marzo, V.; Goparaju, S. K.; Wang, L.; Liu, J.; Bitkai, S.; Jarai,
Z.; Fezza, F.; Miura, G. I.; Palmiter, R. D.; Sugiura, T.; Kunos, G.
Leptin-regulated endocannabinoids are involved in maintaining food
intake. Nature (London) 2001, 410, 822–825. (b) Van Gaal, L. F.;
Rissanen, A. M.; Scheen, A. J.; Ziegler, O.; Roessner, S. Effects of
the cannabinoid-1 receptor blocker rimonabant on weight reduction
and cardiovascular risk factors in overweight patients: 1-year experi-
ence from the RIO-Europe study. Lancet 2005, 365, 1389–1397.
(4) Rimonabant was approved for the treatment of obesity in the European
Union in 2006. For more information, see http://en.sanofi-aventis.com/
events/event1/en/about.asp.
(5) Addy, C.; Li, S.; Agrawal, N.; Stone, J.; Majumdar, A.; Zhong, L.;
Li, H.; Yuan, J.; Maes, A.; Rothenberg, P.; Cote, J.; Rosko, K.;
Cummings, C.; Warrington, S.; Boyce, M.; Gottesdiener, K.; Stoch,
A.; Wagner, J. Safety, tolerability, pharmacokinetics, and pharmaco-
dynamic properties of taranabant, a novel selective cannabinoid-1
receptor inverse agonist, for the treatment of obesity: results from a
double-blind, placebo-controlled, single oral dose study in healthy
volunteers. J. Clin. Pharmacol. 2008, 48, 418–427.
(11) Corey, E. J.; Chaykovsky, M. Dimethyloxosulfonium methylide and
dimethylsulfonium methylide. Formation and application to organic
synthesis. J. Am. Chem. Soc. 1965, 87, 1353–1364.
(12) Mass Spectrometer: Micromass ZQ single quadrupole, electrospray
positive ionization, full scan mode (150-750 amu in 0.5 s). HPLC:
Agilent 1100, binary pump. DAD UV detector: hardware/software
Waters/Micromass MassLynx 4.0. Column: Waters Xterra, 2.1 mm
width, 20 mm length, 3.5 µm packing material. Run time: 4 min. Flow
rate: 1.0 mL /min. Mobile phase A ) water + 0.05% TFA. Mobile
phase B ) acetonitrile + 0.05% TFA. Gradient, time/% A/% B: 0.00/
95/05, 3.00/2/98, 3.25/2/98, 3.26/95/5, 4.00/95/5.
(13) The final crude product 70 contains approximately 10% of tert-butyl
4-{3-[4-(trifluoromethyl)phenyl]butanoyl}piperazine-1-carboxylate
byproduct which does not negatively affect the subsequent steps; E/Z
ratio of 70 is >50/1.
(6) For a detailed discussion of biological assays, see the following: Fong,
T. M.; Guan, X.-M.; Marsh, D. J.; Shen, C.-P.; Stribling, D. S.; Rosko,
JM900063X