Tricyclic imidazole antagonists of the Neuropeptide S Receptor

Article abstract of DOI:10.1016/j.bmcl.2010.04.016

A new structural class of potent antagonists of the Neuropeptide S Receptor (NPSR) is reported. High-throughput screening identified a tricyclic imidazole antagonist of NPSR, and medicinal chemistry optimization of this structure was undertaken to improve potency against the receptor as well as CNS penetration. Detailed herein are synthetic and medicinal chemistry studies that led to the identification of antagonists 15 and NPSR-PI1, which demonstrate potent in vitro NPSR antagonism and central exposure in vivo.

Full text of DOI:10.1016/j.bmcl.2010.04.016

Bioorganic & Medicinal Chemistry Letters 20 (2010) 4704–4708
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Tricyclic imidazole antagonists of the Neuropeptide S Receptor
b
b
c
c
B. Wesley Trotter a, , Kausik K. Nanda , Peter J. Manley , Victor N. Uebele , Cindra L. Condra ,
Anthony L. Gotter c, Karsten Menzel d, Martin Henault e, Rino Stocco e, John J. Renger c, George D. Hartman b,
Mark T. Bilodeau b
*
a Department of Medicinal Chemistry, Merck Research Laboratories, Boston, MA, USA
b Department of Medicinal Chemistry, Merck Research Laboratories, West Point, PA, USA
c Department of Neuroscience, Merck Research Laboratories, West Point, PA, USA
d Department of Drug Metabolism, Merck Research Laboratories, West Point, PA, USA
e In Vitro Sciences Group, Merck Canada, Montreal, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
A new structural class of potent antagonists of the Neuropeptide S Receptor (NPSR) is reported. High-
throughput screening identified a tricyclic imidazole antagonist of NPSR, and medicinal chemistry
optimization of this structure was undertaken to improve potency against the receptor as well as CNS
penetration. Detailed herein are synthetic and medicinal chemistry studies that led to the identification
of antagonists 15 and NPSR-PI1, which demonstrate potent in vitro NPSR antagonism and central
exposure in vivo.
Received 8 March 2010
Revised 5 April 2010
Accepted 7 April 2010
Available online 13 April 2010
Keyword:
Neuropeptide S Receptor
Ó 2010 Elsevier Ltd. All rights reserved.
The neurotransmitter Neuropeptide S (NPS) was first reported
in 2004 as a novel mediator of arousal and fear responses.1 It
was shown to activate the orphan G-protein coupled receptor
behavioral effects in preclinical models.9 Given the lipophilic nat-
ure of 2 (measured log P >3.8) we focused our efforts on increasing
both polarity and potency in order to modulate physical properties
and identify inhibitors with increased likelihood of specific, central
NPSR antagonism in vivo.
The tricyclic core structure could be readily assembled from
simple phenylimidazolines using a known lithiation/acylation/
dehydration sequence.10 In a representative example (Scheme 1),
phenylimidazoline 3 was treated with n-butyllithium and methyl-
4-bromobenzoate to give adduct 4. Heating this compound in acetic
acid resulted in elimination of water (possibly via the intermediacy
of an iminium 1,3-hydride shift)11 to provide 5. Deprotonation of
the benzylic position of 5 has been reported in the literature to pro-
ceed via addition of NaH and to result in air oxidation of this scaf-
fold. We found that alkylation of the sodium anion of 5 with
GPR154, subsequently termed the Neuropeptide
S Receptor
(NPSR). While NPSR is expressed throughout the central nervous
system, NPS precursor mRNA expression has been shown to be
localized to a few specific regions of the brain.2 Central administra-
tion of NPS has been shown to induce arousal and suppress sleep.1
Thus, there has been interest in identifying small molecule antag-
onists of NPSR, and reports of two distinct classes of NPSR antago-
nists have recently appeared.3–5
Here, we report a new class of potent NPSR antagonists identi-
fied at Merck and detail optimization efforts to improve potency
and CNS penetration in this series. Racemic imidazole 1 (Fig. 1),
identified in a high-throughput screening campaign, was a moder-
ately potent NPSR antagonist (NPSR IC50 380 nM).6 Resolution of 1
gave potent enantiomer 2 (FLIPR IC50 163 nM).7
In vitro assays demonstrated that 2 was not a substrate of the
human or rat PGP transporter and thus held promise as a starting
point for optimized CNS-penetrant antagonists.8 While the com-
pound exhibited a measurable free fraction in rat plasma (1%), it
displayed poor membrane permeability (Papp 10 ꢀ 10ꢁ6 cm/s). In
addition, broad counterscreening against a panel of unrelated en-
zymes and receptors indicated that 2 possessed a number of signif-
icant off-target activities that might confound interpretation of
Br
Br
N
N
N
N
2 (enantiopure
NPSR IC50 163 nM
PGP B-A/A-B 0.8
)
1
(racemate)
NPSR IC50 380 nM
* Corresponding author.
E-mail address: bwesley_trotter@merck.com (B.W. Trotter).
Figure 1. Tricyclic imidazole screening leads 1 and 2.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.04.016

B. W. Trotter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4704–4708
4705
Br
Br
a
H
N
N
N
96%
HO
MeO2C
H
N
O
N
N
a
N
N
MeO
MeO
3
4
b
H
b
N
Br
92%
Br
N
O
c
Scheme 3. Reagents and conditions: (a) 1,2-ethylenediamine, I2, K2CO3, t-BuOH, rt;
(b) 1,2-ethylenediamine, NBS, CH2Cl2, 0 °C to rt.
N
N
N
N
1
5
in situ generated aminals derived from the requisite aldehyde
and 1,2-ethylenediamine.12,13
Scheme 1. Reagents and conditions: (a) n-butyllithium, 50 °C, 3 h, then methyl-4-
bromobenzoate, rt, overnight, 55%; (b) AcOH, 125 °C, 5 h, 98%; (c) LDA, THF, ꢁ78 °C,
then benzyl bromide, ꢁ78 to ꢁ30 °C, 92%.
Variation of the benzyl substituent was targeted initially as a
synthetically viable approach to rapid modulation of physical
properties (Table 1).14 Addition of polar substituents to the aryl
ring met with limited success. While 2-CN analog 9a preserved po-
tency, no improvement in permeability or free fraction was evi-
dent. Evaluation of pyridine replacements identified 3-pyridyl
variant 9b, which displayed improved free fraction and permeabil-
ity. We found that the use of saturated alkyl groups in place of the
phenyl ring was tolerated (cyclohexane 9c and cyclopentane 9d).
Incorporation of oxygen to provide tetrahydropyran 9e resulted
in improved potency versus NPSR as well as increased free fraction
and permeability. A key limitation of 9e was its rapid plasma
benzyl bromide to provide 1 was competitive with oxidation and
proceeded in moderate yield. The yield of the alkylation process
could be increased and the reaction extended to a wider variety
of electrophiles by using LDA as the base at low temperature. This
modification allowed replacement of the benzyl moiety with satu-
rated alkyl groups as well as amide-containing electrophiles; elec-
trophiles from both of these classes proved useful in modifying
physical properties in this series.
Replacement of the benzyl group with amides and chain-
extended amide substituents was also readily accomplished via
the corresponding esters. As shown in Scheme 2, deprotonation
of 5 with LDA followed by treatment with Mander reagent, tert-
butylbromoacetate, or tert-butylbromopropionate provided esters
6a–c. Compound 6a was converted directly to the piperidine
amide 8a via trimethylaluminum-mediated transamidation. Com-
pounds 6b and 6c were deprotected under acidic conditions to
provide carboxylic acids 7b–c. Amide coupling reactions provided
compounds of type 8b–c.
Table 1
Benzyl group SAR
Br
R
N
N
Variation of the fused aryl ring was most readily accomplished
by synthesis of substituted arylimidazolines followed by recapit-
ulation of the known lithiation/acylation/dehydration chemistry.
Preparation of the requisite arylimidazolines is detailed in
Scheme 3. This was generally accomplished via oxidation of
Entry
R
NPSR IC50
(nM)a
Plasma free
fraction (%)b
PGP B–A/
A–Bc
Papp
(10ꢁ6 cm/s)d
9a
9b
369
351
<1
1.7
2.1
17
31
NC
N
Br
2.1
Br
Br
Me
O
a
b
O
N
9c
254
321
O
N
N
N
N
N
N
9d
5
6a
8a
c
O
O
9e
9f
92
19
1.2
31
Br
Br
Br
R1
O
O
O
N
t-BuO
HO
d
e
R2
n
n
n
N
N
N
154
N
N
N
6b,c
7b,c
8b,c
a
Values represent the numerical average of at least two experiments. Interassay
variability was 30%.
Scheme 2. Reagents and conditions: (a) LDA, THF, ꢁ78 °C, then methylcyanofor-
mate, 67%; (b) piperidine, trimethylaluminium, toluene, 65%; (c) LDA, THF, ꢁ78 °C,
then, for 6b (n = 1), tert-butylbromoacetate, 81%; for 6c (n = 2), tert-butylbromo-
propionate, 95%; (d) trifluoroacetic acid, CH2Cl2, quantitative; (e) HNR1R2, EDC,
HOAt, i-Pr2EtN.
b
Measured via ultracentrifugation following incubation with rat serum.
Rat MDR1 directional transport ratio (B–A)/(A–B). Values represent the average
c
of three experiments and interassay variability was 20%.
d
Passive permeability.

4706
B. W. Trotter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4704–4708
Table 3
Substituent combinations
Table 2
Aryl group SAR
R
R2
N
R
N
N
Entry R1
R2
NPSR
IC50
Plasma free PGP
Papp
fraction
B–A/
(10ꢁ6
cm/s)d
N
(nM)a
(%)b
A–Bc
O
CN
F
Entry
R
NPSR IC50 (nM)a
336
11a
11b
11c
635
660
134
F
10a
O
O
Cl
Me
10b
10c
10d
358
1325
387
Cl
1.8
1.4
31
CN
Br
11d
11e
851
145
1.3
1.1
1.3
0.6
25
22
N
CN
CN
10e
10f
1725
3166
Br
CN
Br
11f
114
1.6
17
10g
1500
2179
N
a
Values represent the numerical average of at least two experiments. Interassay
variability was 30%.
Br
b
Measured via ultracentrifugation following incubation with rat serum.
Rat MDR1 directional transport ratio (B–A)/(A–B). Values represent the average
10h
c
N
of three experiments and interassay variability was 20%.
d
a
Passive permeability.
Values represent the numerical average of at least two experiments. Interassay
variability was 30%.
In order to further modulate physical and pharmacokinetic
properties, incorporation of amide functionality into the alkyl sub-
stituent connective framework was investigated (Figure 2). While
the initially targeted analog 8a resulted in a significant loss of po-
tency, synthesis of a series of methylene-linked amides revealed a
strong dependence of potency on the length of the linker region.
Propionamide 13 was >10-fold more potent than 8a. Importantly,
chain-extended amides 12 and 13 were not substrates of the PGP
transporter (Figure 2).
Optimization of potency and physical properties in this amide
series was pursued via a rapid analog approach. Simple amide cou-
pling enabled a survey of piperidine replacements. Table 415 shows
representative results from a library based on the 4-chlorophenyl
substituted scaffold. We observed that relatively small cyclic and
acyclic dialkyl amides were potent NPSR antagonists. Plasma free
fraction was found to vary inversely with susceptibility to PGP;
thus a balance between these two factors had to be achieved. A
leading example of this was that incorporation of a pyrrolidinyl-
ethyl substituent on the piperidine (14c) retained potency and
clearance in rat (Cl 48 mL/min/kg), which was expected to impede
achievement of adequate CNS exposure in vivo.
A parallel investigation surveyed variants of the 4-bromophenyl
substituent. Replacement of the bromo group with a number of
heterocycles abrogated potency, indicating that there was little
room for additional functionality at this position. However,
replacement of the bromine with smaller substituents was less del-
eterious (10a–10c), and cyano analog 10d preserved potency while
significantly improving plasma free fraction and permeability (rat
free fraction 17%, B–A/A–B 0.9, Papp 33 ꢀ 10ꢁ6 cm/s). 3-Bromophe-
nyl antagonist 10e displayed reduced potency as did the corre-
sponding 3-CN analog 10f. Bromopyridine analogs 10g and 10h
were also significantly less potent than lead 2.
Straightforward additivity was not observed upon combining
modifications of the benzyl and phenyl substituents (Table 3).
Combinations of the methylene-linked THP group found in 9e with
tolerated phenyl groups such as 4-cyanophenyl (11a) and 4-fluoro-
phenyl (11b) resulted in significant losses in potency. 4-Chloro-
phenyl analog 11c proved to be one of few suitable
combinations, with good NPSR potency and higher free fraction
and permeability than 2. A broadened survey of substituent combi-
nations provided some unexpected results, including cyanophenyl
cyclopentyl antagonist 11e. This antagonist was among the most
potent combinations despite no evident potency benefit from
either of these groups individually (9d, Table 1, and 10d, Table
2). Free fraction in the range of 1–2%, lack of PGP susceptibility,
and good permeability were generally observed in this group of
saturated analogs (Table 3), but no significant improvements in
rat pharmacokinetic properties were realized.
N
Br
Br
Br
O
O
N
N
O
N
N
N
N
N
N
12
13
8a
Figure 2. Amide antagonists.

B. W. Trotter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4704–4708
4707
Table 4
Amide optimization
NR1R2
Cl
O
N
N
Entry
NR1R2
NPSR IC50 (nM)a
131
Plasma free fraction (%)b
1.1
PGP B–A/A–Bc
1.6
Rat clearance (mL/min/kg)d
N
14a
20
26
Et
Me Me
N
Me
Me
14b
14c
96
126
138
0.8
8.1
0.7
1.4
34.2
2.3
Me
N
38
31
N
N
Me
Me
Me
14d
14e
95
92
90
0.6
0.6
2.2
1.7
1.5
2.1
34
36
38
N
Me
14f
N
Me
Me
14g
N
N
Me
Me
F
N
14h
239
2.9
2.0
9
F
a
b
c
Values represent the numerical average of at least two experiments. Interassay variability was 30%.
Measured via ultracentrifugation following incubation with rat serum.
Rat MDR1 Directional Transport Ratio (B–A)/(A–B). Values represent the average of three experiments and interassay variability was 20%.
d
IV clearance measured in cassette dosing experiments. Clearance for 14g is quoted for the racemic form.
substantially increased free fraction but also transformed the com-
pound into a PGP substrate. Modulation of amine basicity via fluo-
rination resulted in PGP non-substrate 14h. While generally
improved relative to precursor compounds, plasma clearance in
this series was significantly dependent on the identity of the amide
group. The simple diethyl amide 14g16 balanced these factors well,
combining >2% free fraction with low PGP susceptibility and mod-
erate clearance. However, 14g exhibited low exposure when dosed
Plasma exposure of each compound when dosed orally to rats
was relatively low (Table 5), but we found that good peripheral
and central exposure could be achieved by IP dosing. 15 achieved
cerebrospinal fluid (CSF) levels (representative of unbound com-
pound levels in the brain)21 of 594 nM (30 min) and 153 nM (2 h)
when dosed at 75 mpk to rats as a solution in 25% HPB-cyclodex-
trin. CSF/plasma ratios of 1% were observed at each time point
and corresponded to the observed plasma free fraction. Thus the
compound exhibited CNS exposure well above its IC50 for NPSR
antagonism.
orally to rats (AUC0–1 0.18 lM-h at 10 mg/kg).
Having identified both the THP-methyl and diethylamino-pro-
pionate subunits as motifs with free fraction and PGP properties
suitable for CNS penetration, we attempted to further improve
potency and oral exposure via tuning of substitution on the tricy-
clic core. Among substituents surveyed on the fused benzo ring,17
few were tolerated, and the methoxy group was unique in its
improvement of potency versus NPSR. Both the amide and THP-
containing antagonists gained potency via methoxy substitution
at both the 6- and 7-positions. Compounds 15 and NPSR-PI118
were identified as particularly potent THP and amide antagonists
with >1% plasma free fraction and minimal PGP transport.19 Both
compounds were counterscreened against broad panels of en-
zymes, receptors and ion channels and found to exhibit much im-
proved profiles relative to lead 2.20
NPSR-PI1 was dosed at 30 mpk IP to rats. Plasma levels mea-
sured at 30 min, 1 h, and 2 h were 7.8
respectively. These concentrations correspond to unbound expo-
sures of 101 nM, 66 nM, and 39 nM, at or above the NPSR IC50
lM, 5.1 lM, and 3.0 lM,
.
The exposures observed for both 15 and NPSR-PI1 relative to IC50
are comparable to exposures demonstrated to result in NPSR recep-
tor occupancy in vivo in a separately reported antagonist series.3
In summary, a new structural class of NPSR antagonists has
been identified. Optimization provided compounds that were not
substrates of rat PGP and exhibited significant plasma free frac-
tions, enabling in vivo CNS exposure upon IP dosing. This new class
of NPSR antagonists provides additional tools for in vivo studies of
the relevance of NPSR to behavioral modification.

4708
B. W. Trotter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4704–4708
Table 5
Optimal core substituents
Entry
Structure
NPSR IC50 (nM)a
Plasma free fractionb
PGP B–A/A–Bc
Rat AUC (PO, dose)d
O
Cl
15
82
1.6
1.3
0.114
0.425
l
M-h (2 mpk)
N
MeO
MeO
N
O
Cl
16
17
46
50
0.8
0.7
na
lM-h (10 mpk)
N
N
Me
N
Me
Cl
O
2.2
0.115
lM-h (2 mpk)
N
MeO
Et
N
Et
N
Cl
O
NPSR-Pl1
43
1.3
2.9
1.15 lM-h (10 mpk)
MeO
N
N
a
b
c
Values represent the numerical average of at least two experiments. Interassay variability was 30%.
Measured via ultracentrifugation following incubation with rat serum.
Rat MDR1 Directional Transport Ratio (B–A)/(A–B). Values represent the average of three experiments and interassay variability was 20%.
Dosing vehicle: 20% aqueous Vitamin E/TPGS.
d
14. Except where noted, compounds were synthesized and tested in racemic form.
15. Except where noted, compounds were synthesized from racemic acids of type
7. Stereogenic amines were also introduced as racemates to afford
diastereomeric mixtures; in order to accelerate screening, diastereomers
were not separated.
16. Single enantiomer, obtained from the enantiopure carboxylic acid precursor,
which was separated via chiral HPLC (analytical: SFC-OD-H column
(4.6 ꢀ 250 mm), MeOH/CO2 (20:80, 4 mL/min), retention time 5.7 min
(slower eluting peak). Absolute stereochemistry has not been determined.
17. Methyl substitution was tolerated but did not improve physical properties.
Cyano groups as well as larger alkoxy substituents led to loss of NPSR
antagonist potency. Pyridinyl replacements for the fused benzo ring also
resulted in potency decreases. Benzyl variant i (NPSR IC50 1458 nM) is a typical
example.
References and notes
1. Xu, Y. L.; Reinscheid, R. K.; Huitron-Resendiz, S.; Clark, S. D.; Wang, Z. W.; Lin, S.
H.; Brucher, F. A.; Zeng, J. A.; Ly, N. K.; Henriksen, S. J.; Lecea, L. C.; Civelli, O.
Neuron 2004, 43, 487.
2. Xu, Y. L.; Gall, C. M.; Jackson, V. R.; Civelli, O.; Reinscheid, R. K. J. Comp. Neurol.
2007, 500, 84.
3. Melamed, J. Y. et al Bioorg. Med. Chem. Lett. 2010, 20, 4700.
4. Okamura, N.; Habay, S. A.; Zeng, J.; Chamberlin, A. R.; Reinscheid, R. K. J.
Pharmacol. Exp. Ther. 2008, 325, 893.
5. Zhang, Y.; Gilmour, B. P.; Navarro, H. A.; Runyon, S. P. Bioorg. Med. Chem. Lett.
2008, 18, 4064.
6. A dual sequence FLIPR (Fluorometric Imaging Plate Reader) Ca2+ mobilization
assay was used to evaluate compounds for agonist and antagonist activity on
the Neuropeptide S Receptor expressed on recombinant CHOK1 cells plated in
Br
384 well PDL coated plates 24 h prior to experiment (15 K cells/well/20 lL).
Following washing with buffer and loading with Fluo4AM dye for 1 h at 37 °C
in 5% CO2, cell plates were placed in the FLIPR and incubated with titrated
compounds for 4 min during sequence 1 data collection. In sequence 2, an EC80
of human NPS was added to the assay plates and data collected for another
4 min. Percent stimulatory activity (sequence 1) or inhibition (sequence 2) was
calculated relative to Ca2+ mobilization in NPS stimulated control wells (EC100
or EC80 NPS). Compound potencies were determined from plots of
concentration response curves using 4p logistical fits and reported as
means S.D. of multiple determinations.
i
N
N
N
18. Single enantiomer, obtained via chiral HPLC (SFC-AD-H column (2 ꢀ 25 cm),
EtOH/CO2 (15:85, 70 mL/min), retention time 5.2 min (faster eluting peak).
Absolute stereochemistry has not been determined.
19. An HPLC log D of 3.2 was measured for 15.
7. ent-2 was significantly less potent (NPSR IC50 1987 nM).
8. Miller, D. S.; Bauer, B.; Hartz, A. M. S. Pharmacol. Rev. 2008, 60, 196.
9. IC50 values ranging from 1 to 8 lM were reported for several GPCRs, ion
channels, and enzymes with potential CNS relevance (MDS Pharma Services).
10. Houlihan, W. J.; Parrino, V. A. J. Heterocycl. Chem. 1981, 18, 1549.
11. Cook, A. G.; Switek, K. A.; Cutler, K. A.; Witt, A. N. Lett. Org. Chem. 2004, 1, 1.
12. Fujioka, H.; Murai, K.; Ohba, Y.; Hiramatsu, A.; Kita, Y. Tetrahedron Lett. 2005,
46, 2197.
20. Against a panel of 163 enzymes, receptors and ion channels, 15 showed IC50
values <10
rabbit monoamine transporter, 7
a panel of 36 CNS-relevant enzymes and receptors, NPSR-PI1 showed an IC50
value <10 M in only one assay: CB1 receptor, 6 M.
l
M in only four assays: thromboxane A2, 7
l
M, melatonin, 1.3
lM,
l
M, and sodium channel site 2, 4
lM. Against
l
l
21. Liu, X. R.; Smith, B. J.; Chen, C.; Callegari, E.; Becker, S. L.; Chen, X.; Cianfrogna,
J.; Doran, A. C.; Doran, S. D.; Gibbs, J. P.; Hosea, N.; Liu, J. H.; Nelson, F. R.; Szewc,
M. A.; Van Deusen, J. Drug Metab. Dispos. 2006, 34, 1443.
13. Ishihara, M.; Togo, H. Tetrahedron 2007, 63, 1474.