The discovery of a selective, small molecule agonist for the mas-related gene X1 receptor

Article abstract of DOI:10.1021/jm800962k

The novel 7-transmembrane receptor MrgX1 is located predominantly in the dorsal root ganglion and has consequently been implicated in the perception of pain. Here we describe the discovery and optimization of a small molecule agonist and initial docking s

Full text of DOI:10.1021/jm800962k

818
J. Med. Chem. 2009, 52, 818–825
The Discovery of a Selective, Small Molecule Agonist for the Mas-Related Gene X1 Receptor
Berthold Wroblowski, Mark J. Wigglesworth, Philip G. Szekeres, Graham D. Smith, Shahzad S. Rahman,
Neville H. Nicholson,* Alison I. Muir, Adrian Hall, Jag P. Heer, Stephen L. Garland, and William J. Coates
GlaxoSmithKline Pharmaceuticals, New Frontiers Science Park (North), Third AVenue, Harlow, Essex CM19 5AW, United Kingdom
ReceiVed July 30, 2008
The novel 7-transmembrane receptor MrgX1 is located predominantly in the dorsal root ganglion and has
consequently been implicated in the perception of pain. Here we describe the discovery and optimization of
a small molecule agonist and initial docking studies of this ligand into the receptor in order to provide a
suitable lead and tool compound for the elucidation of the physiological function of the receptor.
Introduction
The reversal of morphine tolerance was described over 30
years ago,24 and several mechanisms, including NMDA
receptors25,26 and PKA,27 GSK3,28 and PKC29 kinases, have
been invoked to explain the phenomenon that has also been
reported to emanate from the spinal column.30 Chen et al.31
have recently suggested that activation of rat MrgX receptors
induces spinal analgesia by suppressing NMDA receptor-
mediated activation of spinal dorsal horn neurons and increasing
NOS activity.
Stimulation of MrgX1 is an alternative mechanism for the
reversal of morphine tolerance and provides a target that is a
simple 7-transmembrane receptor that has no known function
other than its putative role in pain and is limited in its occurrence
to the dorsal root ganglion. This mechanism requires an MrgX1
agonist, whereas the reversal of morphine tolerance by direct
antagonism of the NMDA receptor requires a ligand with a high
affinity for a widely distributed receptor.
Pain, especially in the terminally ill, remains a tragic and
unresolved medical dilemma. Despite the plethora of approaches
to a solution1,2 encompassing ion channels,3-7 7-transmembrane
receptors,8,9 and holism,10,11 opiates such as morphine remain
the treatment of choice for the relief of acute pain.12,13 However,
the analgesic effect of morphine declines and adverse effects
increase with prolonged exposure, and this limits its use to
transitory, postoperative pain or, in escalating doses, to the final
days of disease.
The recently described Mas-related genes (Mrg) or sensory
neuron-specific receptors, are a family of 7-transmembrane
receptors that have been implicated in the perception of pain
through their expression in small afferent nerve fibers of the
dorsal root ganglion.14,15 Data from the rat have suggested that
painful stimuli are intensified by the presence of an Mrg agonist
in the cerebrospinal fluid and that an Mrg antagonist might
therefore give protection against painful stimuli.16 A member
of this family of receptors, MrgX1a has been identified in the
human, but the high frequency of significant changes within
these genes shows that they are undergoing rapid evolution17,18
and, as a consequence, the closest rat gene has only 60%
identity. Functional and structural homologues more closely
related to the human gene have been identified in the rhesus19
and macaque20 monkeys and screening for MrgX1 antagonists
was initiated21,22 to enable studies of the function of these
receptors, but there have been no more specific reports linking
MrgX1 antagonists to the amelioration of pain in higher
mammals.
These data have generated a consequent interest in novel,
pharmaceutically acceptable MrgX1 agonists,32 and we describe
here the discovery, synthesis, and optimization of a small
molecule agonist and initial docking studies of the ligand into
the receptor.
High Throughput Screening
The GlaxoSmithKline compound collection was screened
using HEK293 cells stably transfected with human MrgX1 in
an intracellular calcium mobilization assay with the fluorometric
imaging plate reader (FLIPR, Molecular Devices) essentially
as described by Sulivan et al.33 This gave 79 compounds
comprising several different chemotypes that had an intrinsic
activity greater than 0.25 of that of the standard BAM6-22.
Among the most active of these was a series of pyridazinones
represented by 1, which appeared to be readily synthesized and
amenable to variation using precursors immobilized on resin
or in solution phase.
The high throughput screening hit was resynthesized using
the route shown in Scheme 1. The 5-chlorine of commercially
available 4,5-dichloropyridazin-3-one 2 was displaced regiose-
lectively with piperazine to give fragment 3. 1H NMR and NOE
experiments were consistent with the proposed regiochemistry
of compound 3 but not definitive, so the remaining chlorine
was hydrogenated to give compound 3a, which was shown to
have the required regiochemistry. The amine 3 was attached to
Wang resin through a carbamate linkage to give the immobilized
reagent 4, which was reacted with 2-methyl-3-nitrobenzylbro-
mide in the presence of a strong base. An aliquot of the product
5 was liberated from the resin with TFA to provide the soluble
However, recent publications23 have shown that an Mrg
agonist is able to reactivate the analgesic effect of morphine
after it has ceased to be effective due to prolonged use
(tolerance). Intrathecal administration of bovine adrenal medulla
8-22 (BAM8-22), a selective small peptide agonist at Mrg
receptors, did not by itself directly affect acute thermal
nociception in the rat. However, when chronic morphine
administration was intermittently supplemented with BAM8-22,
the antinociception due to the morphine was prolonged.
Furthermore, intermittent administration of BAM8-22 consis-
tently reversed established morphine tolerance.
* To whom correspondence should be addressed. Phone and Fax: 0(44)
1440 730580. E-mail:nxnicholson@gmail.com.
a Abbreviations: MrgX1, mas-related gene X1; BAM, bovine adrenal
medulla; DRR, dorsal root receptor; NMDA, N-methyl-D-aspartate; 5-HT,
5-hydroxytryptamine; HOAt, 3H-[1,2,3]triazolo[4,5-b]pyridin-3-ol; IA,
intrinsic activity; DMSO, dimethylsulfoxide; TFA, trifluoroacetic acid; DMF,
dimethylformamide, THF, tetrahydrofuran; NMP, N-methylpyrollidinone.
10.1021/jm800962k CCC: $40.75
2009 American Chemical Society
Published on Web 01/15/2009

Mas-Related Gene X1 Agonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 3 819
Scheme 1a
a Reagents and conditions: (i) EtOH, 90 °C; (ii) 2-pyridylcarboxy-Wang resin; (iii) 2-methyl-3-nitrobenzylbromide, tetramethylguanine; (iv) SnCl2.2H2O,
NMP; (v) 4-biphenylcarboxylic acid, HOAt, DIC, DMF/MDC 1/1; (vi) 20% TFA /DCM.
Table 1. The Affinity of the Pyridazinones for Aminergic Receptorsa
pKi
pEC50
Mrg
X1
5-HT
R
ꢀ
dopamine
R1
R2
compd
1B
1D
2A
2B
2C
6
1B
2
D2
D3
H
H
CH3
H
CH3
H
8
9
1
6.28
6.11
6.23
6.38
6.13
6.34
7.22
8.67
7.28
6.81
7.52
7.01
7.52
8.41
6.52
6.31
6.08
5.47
6.2
6.25
5.84
5.64
5.62
<5.2
5.28
5.31
5.29
6.09
6.11
5.45
5.08
5.39
6.9
a The standard deviation was less than 0.35 for between 2 and 6 replicates. The cell line for 5-HT2 receptors was HEK293 for 5-HT6, HELA, and for
all other receptors CHO. Ligands were: 5-HT1B and D and 5-HT2B, 3H 5-HT; 5-HT6, 3H LSD; 5-HT2A, 3H ketanserin; 5-HT2C, 3H mesulergine; R1B, 3H
Prazosin; ꢀ2, 125I iodocyanopindolol; D2 and D3, 125I iodosulpride. pKi ) -log 10 Ki, where Ki is calculated using the Cheng-Prusoff equation IC50/1 +
([free radioligand]/[the dissociation constant]).
1
compound 5a, which was analyzed by H and 13C NMR to
ensure that alkylation had occurred at the required position. The
carbon shift of the benzyl CH2 at 53 ppm was consistent with
N-alkylation rather than O-alkylation and HMBC correlations
confirmed N2 as the site of alkylation. The nitro moiety of resin
bound 5 was reduced with tin chloride, and the resulting amine
6 was coupled with 4-phenylbenzoic acid to provide compound
7, which, on cleavage from the resin with TFA, gave the target
compound 1.
affinity for 5-HT1 B and D, R1B and ꢀ2 adrenoreceptors and
for dopamine D2 and D3 receptors (Table 1).
Receptor Modeling Studies
A model of the MrgX1 receptor was built in order to facilitate
chemical optimization of the ligand. The protein structure was
modeled by homology to the published bovine rhodopsin crystal
structure (PDB code 1f88) using the Modeler software.34 The
hit structure 1 and a series of closely related compounds were
docked in using FLO.35 This allowed conformational flexibility
in both the ligand and amino acid side-chains of the binding
pocket. The partial charges of the ligand had been precalculated
using Vamp36 to obtain a more realistic charge distribution than
the default.
In the modes with the highest docking scores, the compounds
exhibited well-defined interactions of the charged piperazine
nitrogen with the acidic residues E157 of TM4 and/or D177 of
TM5. The pyridazinone portion of the structure sat high up in
the receptor in an approximately horizontal orientation, contact-
ing residues from the second extracellular loop. The ligand then
turned about the benzyl amide section, with the biphenyl part
running vertically downward into a hydrophobic pocket between
TMs 3 and 6 (Figure 1).
Testing of the pure 1 confirmed its MrgX1 agonist activity
(pEC50 ) 6.9, intrinsic activity 0.7 relative to BAM3200), and
initial SAR of the core of the molecule was available from
compounds already in the GlaxoSmithKline collection. Excision
of the methyl from the central benzene ring to give compound
8 or moving it to the ortho-position on the alternate side of the
amide, compound 9, reduced potency 15-fold (Table 1). Placing
the amide in the ortho- or para- positions of the central benzene
ring caused a similar decrease in potency. Clearly, the preferred
agonist conformation required the amide side chain to be
extended with the amide carbonyl away from the methyl group
of 1. Consequently, this methyl, together with the rest of the
benzylpyridazinone core, were adopted to investigate the SAR
of the periphery of the molecule.
The rhodopsin structure is of the inactive state of the receptor.
We considered whether to invoke some specific changes in the
model to account for the fact that the ligands are agonists rather
than antagonists. Typically, this might involve movement of a
key W residue that changes conformation as part of the “toggle
Aminergic activity was clearly adumbrated by the positive
charge at physiological pH in the selected structures, and this
was confirmed for three members of the series that displayed
high affinities for 5-HT2 A, B, and C receptors and moderate

820 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 3
Wroblowski et al.
Figure 1. (a) Position of compound 22 (shown in orange space fill) as modeled in the MrgX1 receptor bundle. The transmembrane helices run
from TM1 shown in red through to TM7 shown in dark blue. The ligand is anchored on TMs 4 and 5 (yellow/green and green/blue) via two
charge-charge interactions and then forms a series of π-stacking interactions with aromatic residues of TMs 3 (yellow), 6 and 7 (light blue and
dark blue). (b) Close up view of compound 22 (shown in space fill) as modeled in the MrgX1 receptor bundle. Key interactions are formed with
Tyr 99, Tyr 106, Glu 157, Asp 177, Phe 232, Phe 236, and His 254.
switch” model37 of receptor activation. However, this cannot
be applied here. The “switch”, usually afforded by the W on
TM6 as part of a CWLP sequence, is replaced by a G in MrgX1.
Rather than tripping the switch to alleviate the steric constraint
and hence allow TMs 3 and 6 to come closer together, it may
be that the ligands form a series of favorable interactions with
residues along each side that serves to draw TM3 and TM6
closer together on the extracellular side. It is noticeable that
the biaryl portion contacts a number of aromatic residues from
TMs 3 (Y99, Y106) and 6 (F232, F236), and it is possible that

Mas-Related Gene X1 Agonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 3 821
Scheme 2a
a Reagents and conditions: (i) tetramethylguanine, THF; (ii) SnCl2 ·2H2O, EtOH; (iii) 4-[6-(methyloxy)-3-pyridinyl]benzoic acid, HOAt, DIC, DMF/
MDC 1/1; (iv) piperazine; (v) hydrazine, 10% Pd-C.
the interactions with these residues are improved in an activated
form of the receptor. The H254 residue from TM7 is also in
reasonable proximity to the ligands but may improve its
interaction if the top of the helix were to come closer to TM3.
As it was, we found that the ligands fitted well in the receptor
bundle as modeled based on the available crystal structure
without any explicit change to account for modality.
Table 2. The Effect of Variation of the Amide Substituent on MrgX1
Activitya
The modeling was used to identify structural changes that
would increase MrgX1 affinity and also to consider issues
around selectivity. The acidic side chains at the tops of TMs 4
and 5 in MrgX1 are not preserved in the aminergic receptors.
The equivalent positions in 5-HT2C, for example, are both I
residues. Hence, the ligands probably do not bind in the same
mode at aminergic receptors and are perhaps more likely to
anchor on the key D residue of TM3 usually implicated in small
molecule binding. Thus, rather than focus on specific differences
between MrgX1 and the aminergic receptors for selectivity, we
sought opportunities to improve the potency at the target receptor
in the expectation that this would also serve to drive divergent
SAR. In particular, these studies suggested the incorporation
of heteroatoms into the distal ring of the biphenyl moiety might
enhance binding to the MrgX1 receptor by forming hydrogen
bonding interactions, particularly in the region of the hydroxyl
group of Y106. A series of compounds was synthesized using
biaryl acids to investigate this possibility.
Optimization of the Biaryl Moiety. The route shown in
Scheme 1 was employed; 4-biphenyl carboxylic acid being
replaced in the penultimate step by a series of acids to produce
the compounds shown in Table 2.
The results indicate that small alkyl (compound 10) or small
aryl (compounds 11 and 12) amides provide an inadequate
hydrophobic surface to bind to MrgX1 and that the high
concentration at which the ligands bind gives poor intrinsic
activity. Increasing the bulk of the hydrophobe (compound 13)
improves the potency but not the intrinsic activity, and this can
be rationalized in our model by an improvement in the binding
of the ligand. However, extending the hydrophobic moiety has
a positive effect on the potency and the removal of bulky
substituents close to the amide increases intrinsic activity
(compound 14).
a The standard deviation of any of the four replicates tested on two
separate occasions was less than 0.34. pEC50 ) -log 10[EC50(M)], where
EC50 was determined from four parameter fit dose-response curves within
an automated fitting program.
Variation of predominantly hydrophobic substituents has
some effect (compounds 15 and 16), but altering the vector of
the hydrophobe (compounds 17 and 18) is detrimental to

822 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 3
Wroblowski et al.
Table 3. The Effect of Variation of the Amide Substituent on Binding
Table 4. The Effect of Variation of the Piperazine Substituent on
to Aminergic Receptorsa
MrgX1 Activitya
receptor
pEC50
pKi
compd
MrgX1
5-HT2A
5-HT2B
5-HT2C
22
23
1
7.6
7.3
6.9
5.53
<5.13
7.28
5.77
<5.32
7.01
5.81
<5.22
6.52
a The standard deviation of the agonist activity at MrgX1 over the six
replicates tested on three separate occasions was <0.1. The standard
deviation of the binding affinity of the compounds at the 5-HT2 receptors
was <0.3 over at least 3 replicates.
potency. Hydrogen bond acceptors close to the meta- position of
the distal aromatic ring are beneficial to both potency and intrinsic
activity (compounds 19-21), but the most effective substituents
contain both hydrogen bond acceptors and donors in this region
(compounds 22 pEC50 ) 7.6 and 23 pEC50 ) 7.3).
The specificity of the hydrogen bond donor/acceptor interac-
tion for MrgX1 is demonstrated by the reduced affinity of the
new ligands for aminergic receptors (Table 3). Whereas the HTS
hit (1) is equipotent at MrgX1 and 5-HT2, the agonists 22 and
23 show at least 15-fold selectivity for MrgX1 over the 5-HT2
receptors. Docking of the agonist 22 into the MrgX1 model is
shown in Figure 1, where the principal putative interactions with
the receptor have been highlighted.
Examination of the Piperazine Module. The resin based
synthesis was adapted to solution phase (Scheme 2) so that a
variety of amines could be added to the 4-position of the
1
pyridazinone ring as the final step in the synthesis. H and 13
C
NMR analysis of the final product showed that the modifications
to the synthesis had not altered the regioselectivity of the
aromatic substitution or alkylation reactions.
It was rapidly apparent from the results of testing the products
(Table 4) that the original piperazine ring was the best
substituent at the 4-position of the pyridazinone ring.
a The standard deviation of the two replicates for each of the above
determinations was <0.4.
Table 5. Aminergic Activity of the Dechlorinated MrgX1 Agonistsa
Homopiperazine was tolerated (compound 27), and methy-
lation, bridging, or opening of the piperazine ring were only
slightly detrimental to activity (compounds 28-30), but bulky
substitution around the basic amine markedly reduced activity
(compounds 31 and 32) as did replacement of the piperazine
with 3-methylaminopyrrolidine (compound 33). Removal of the
strong positive charge destroyed MrgX1 activity (compounds
34 and 35). These results suggested that further efforts to
improve the piperazine substituent were unlikely to be rewarded
and other aspects of the ligand docked into the receptor should
be examined as candidates for modification.
The optimal docking mode for compound 22 into MrgX1
shows the chlorine on the pyridazinone ring pointing away from
and making no significant interaction with the receptor. Sub-
stitution of hydrogen for this chlorine should not affect
hydrophobic binding of the ligand to the receptor and will
increase the positive charge on the adjacent nitrogen at physi-
ological pH, strengthening its interaction with the acidic residues
of the receptor. The chlorine was removed from compounds
22 and 1 by catalytic hydrogenation (Scheme 2) to provide
compounds 36 and 37, and the structure and regiochemistry of
compound 36 was confirmed by 1H and 13C NMR experiments
employing HMBC and NOE correlations. In both cases, the
removal of the chlorine atom from the pyridazinone ring
improved the potency of the compounds for MrgX1 more than
5-fold (Tables 3 and 5) and further reduced their affinity for
5-HT2 receptors. Compound 36 is a full agonist at MrgX1 with
a potency in the tens of nanomolar and only weak interaction
with 5-HT2B in the micromolar range. This compound was
a The standard deviation of the agonist activity of these compounds at
MrgX1 in any of the six replicates tested on three separate occasions was
<0.13. The standard deviation of the binding affinity of compounds 36
and 37 at the 5-HT2 receptors was <0.3 over 6 and 3 replicates, respectively.
tested against a panel of 50 receptors (CEREP) and found to
be inactive at all of them up to 1 µM concentration. When tested
against the rat DRR1 receptor compound 36 together with
compounds 1, 22, and 26 were found to have no detectable
agonist activity.
Conclusion
Investigation into the physiological role of MrgX1 in humans
has been hampered by the lack of a close orthologue in lower
species. Of the 27 mouse and 9 rat Mrg genes, the rat DRR1
(RN2) is the closest equivalent to the human gene but shares

Mas-Related Gene X1 Agonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 3 823
mL) and shaken with tetramethylguanidine (2.89 mL, 23 mmol)
and 2-methyl-3-nitrobenzyl chloride (8.5 g, 45.8 mmol) at room
temperature overnight. The resin was filtered off and washed twice
with each of the following solvents: dichloromethane, THF,
dichloromethane-methanol (2:1, 1:1), methanol, dichloromethane,
ether, dichloromethane.
only 60% of its sequence. Evidence for a shared function
between human MrgX1 and rat DRR1 comes from their location
on similar cells comprising small afferent fibers of the dorsal
root ganglion and from their activation by the same small peptide
agonist, BAM8-22.16 Similar difficulties are encountered with
all small mammals in which a pain model has been established
and which might otherwise be employed to define the role of
MrgX1 in the human.
A sample of this resin (20 mg) was cleaved with 20% TFA in
dichloromethane for 1.5 h. The reaction mixture was evaporated
to dryness and azeotroped with toluene to give 5a as a yellow oil
1
Our approach to this conundrum has been to design a potent
and selective small molecule agonist of MrgX1. The SAR
derived from this exercise can be used to define the atomic
interactions of the agonist with the receptor that will facilitate
comparison of the human with the myriad lower mammal
receptors and help to define their functional similarity. Further
characterization of the human receptor using this agonist together
with site directed mutagenesis will be published in due course.
BAM8-22 is a potent agonist of the rat receptor DRR1 but
(5 mg). H NMR (400 MHz, CD3OD): δ 2.48 (3H, s), 3. 41 (4H,
t, J ) 4.8 Hz), 3.73 (4H, t, J ) 4.8 Hz), 5.46 (2H, s), 7.34 (1H, t,
J ) 8 Hz), 7.46 (1H, d, J ) 7.6 Hz), 7.69 (1H, d, J ) 8 Hz), 7.97
(1H, s). 13C NMR (100 MHz, DMSO-d6): δ 14.5, 43.3 (2C),
46.1(2C), 52.7, 116.0, 123.3, 127.2, 130.2, 132.8, 133.0, 137.9,
147.5, 151.5, 157.6; m/z 364 [M + H]+. HMBC correlations
confimed N2 as the site of alkylation.
1-(2-(2-Methyl-3-aminobenzyl)-4-chloro-3-oxopyridazin-5-yl)-
piperazine 4-Carboxy Wang Resin 6. 1-(2-(2-Methyl-3-nitroben-
zyl)-4-chloro-3-oxopyridazin-5-yl)-piperazine 4-carboxy Wang resin
5 from the previous preparation was suspended in NMP (100 mL)
and shaken with SnCl2 ·2H2O (14.4 g, 64 mmol) at room temper-
ature overnight. The resin was filtered off and washed twice with
each of the following solvents: NMP, dichloromethane, dioxan,
dioxin-water (2:1, 1:1, 1:2), dioxan, and dichloromethane and dried
to give 5.7 g of resin. (1.2 mmol/g).
A sample of this resin (29 mg) was cleaved with 20% TFA in
dichloromethane and the supernatant fluid was evaporated and
azeotroped with toluene. 1H NMR (400 MHz, CD3OD): δ 2.37 (3H,
s), 3. 39 (4H, t, J ) 5.2 Hz), 3.69 (4H, t, J ) 5.2 Hz), 5.40 (2H,
s), 7.04 (1H, d, J ) 7.6 Hz), 7.12-7.23 (2H, m), 7.93 (1H, s); m/z
334 [M + H]+.
1-(2-(2-Methyl-3-(4-phenylbenzoylamido)-benzyl)-4-chloro-3-ox-
opyridazin-5-yl)-piperazine 4-Carboxy Wang Resin 7. 1-(2-(2-
Methyl-3-aminobenzyl)-4-chloro-3-oxopyridazin-5-yl)-piperazine
4-carboxy Wang resin 6 (500 mg, 0.5 mmol) was suspended in
DMF-dichloromethane (1:1, 50 mL) and reacted by shaking with
4-phenylbenzoic acid (300 mg, 1.5 mmol), HOAt (204 mg, 1.5
mmol), and DIC (0.234 mL, 1.5 mmol) at room temperature
overnight. The resin was separated from the supernatant fluid and
was again reacted with the acylating mixture overnight. The resin
was filtered off and washed twice with each of the following
solvents: dichloromethane-DMF (1:1), dichloromethane, dichloro-
methane-methanol (1:1), methanol, dichloromethane, to give 529
mg of dried resin bound title compound 7.
1-(2-(2-Methyl-3-(4-phenylbenzoylamido)-benzyl)-4-chloro-3-ox-
opyridazin-5-yl)-piperazine 1. 1-(2-(2-Methyl-3-(4-phenylbenzoy-
lamido)-benzyl)-4-chloro-3-oxopyridazin-5-yl)-piperazine 4-carboxy
Wang resin (7) (0.25 mg) was reacted with 20% TFA in dichlo-
romethane for 1 h. The supernatant fluid was separated and
evaporated to dryness and azeotroped with toluene. Purification of
this solid by semipreparative reverse phase chromatography gave
the title compound (16 mg). 1H NMR (400 MHz, CD3OD): δ 2.35
(3H, s), 3. 34 (4H, t, J ) 5.2 Hz), 3.66 (4H, t, J ) 5.2 Hz), 5.44
(2H, s), 7.14 (1H, d, J ) 6.8 Hz), 7.21 (1H, t, J ) 7.6), 7.30 (1H,
d, J ) 8.0 Hz), 7.40 (1H, t, J ) 7.2), 7.48 (2H, t, J ) 7.4), 7.69
(2H, d, J ) 7.2), 7.79 (2H, d, J ) 8.4), 7. 94 (1H, s), 8.05 (2H, d,
J ) 8.4); m/z 514 [M + H]+, m/z [C29H28ClN5O2H]+ found
514.2002, error 1.5 ppm.
only a weak agonist of the human receptor MrgX1 (pEC50
)
6.1). By contrast, the selective MrgX1 agonist, which we have
developed, has potent activity against the human receptor and
could be used to test the effects of an MrgX1 agonist in
preparations of human dorsal root ganglia or as an adjunct to
morphine therapy in higher mammals that had developed a
degree of tolerance to morphine.
Experimental Section
4-Chloro-5-piperazin-1-yl-pyridazin-3-one hydrochloride 3·HCl.
4,5-Dichloro-3-hydroxypyridazine 2, (11 g, 67 mmol) and pipera-
zine (5.7 g, 66 mmol) were heated in ethanol solution (150 mL) at
90 °C overnight under a reflux condenser. Filtration yielded the
1
solid 3·HCl (13 g, 92%). H NMR (400 MHz, CDCl3): δ 3.03
(4H, t, J ) 4.8 Hz), 3.42 (4H, t, J ) 4.8 Hz), 7.26 (1H, s), 7.64
(1H, s). 13C NMR (100 MHz, DMSO-d6): δ 46.5, 51.3, 115.9, 131.8,
148.4, 163.2; m/z 215 [M + H]+.
5-(1-Piperazinyl)-3(2H)-pyridazinone 3a. Palladium (10%) on
carbon (50% H2O, 0.1 g) was suspended in a solution of 4-chloro-
5-piperazin-1-yl-pyridazin-3-one hydrochloride 3, (0.5 g, 2.3 mmol)
in ethanol (25 mL) and treated at 50 °C with hydrazine hydrate (1
mL, 13 mmol). The suspension was stirred under reflux conditions
for 1 h when the hot reaction mixture was filtered through a bed of
Kieselguhr and the filtrate evaporated to dryness and azeotroped
with water. The residue was dissolved in water and adjusted to pH
11 with NaOH and the resulting solution evaporated to dryness
and triturated with a mixture of dichloromethane-MeOH (8:2). The
organic layer was separated from the solid and evaporated to dryness
1
to provide the title compound as a white solid (360 mg, 87%). H
NMR (400 MHz, DMSO-d6): δ 2.93 (4H, t, J ) 4.8 Hz), 3.46
(4H, t, J ) 4.8 Hz), 5.77 (1H, d, J ) 2.8 Hz), 6.2-7.2 (1H, brs),
7.92 (1H, d, J ) 2.8 Hz), 12.10-12.33 (1H, brs).
1-(4-Chloro-3-oxopyridazin-5-yl)-piperazine 4-Carboxy Wang
Resin 4. 4-Chloro-5-piperazin-1-yl-pyridazin-3-one hydrochloride
3·HCl was suspended in water (50 mL), which was adjusted to
pH 11 with 40% NaOH and evaporated to dryness. The solid was
extracted with dichloromethane-methanol (8:2) and separated from
the liquid by filtration. The filtrate was evaporated to dryness and
azeotroped with DMF and toluene and dried in a desiccator to
provide 3 (2.4 g, 11.2 mmol). This solid was dissolved in
dichloromethane (150 mL) and shaken under argon at room
temperature with 2-pyridylcarboxy-Wang resin (4 g, 1.7 mmol/g)
overnight. The resin was filtered off and washed three times with
eachofthefollowingsolvents:dichloromethane,THF,methanol-water
(1:1), THF, and dichloromethane.
A small portion of the dried resin was treated with 20% TFA in
dichloromethane to provide 3 as the trifluoroacetate salt; m/z 215
[M + H]+.
1-(2-(2-Methyl-3-nitrobenzyl)-4-chloro-3-oxopyridazin-5-yl)-pip-
erazine 4-Carboxy Wang Resin 5. 1-(4-Chloro-3-oxopyridazin-5-
yl)-piperazine 4-carboxy Wang resin 4 was suspended in THF (150
Biphenyl-4-carboxylic acid [3-(6-oxo-4-piperazin-1-yl-6-pyridazin-
1-ylmethyl)-2-methyl-phenyl]-amide 37. Biphenyl-4-carboxylic acid
[3-(5-chloro-6-oxo-4-piperazin-1-yl-6-pyridazin-1-ylmethyl)-2-methyl-
phenyl]-amide, 1 (136 mg, 0.27 mmol) in ethanol (200 mL) was
stirred at 70 °C with 10% Pd-C-H2O (1:1) (50 mg) and hydrazine
hydrate (1 mL) for 7 h. The catalyst was removed by filtration and
washed with warm ethanol, and the combined filtrates were
evaporated to dryness. Trituration of the solid residue with ether
gave substantially pure product as a white solid (130 mg), a portion
(40 mg) of which was purified by reverse phase HPLC to give the
title compound 37 (10 mg). 1H NMR (400 MHz, CD3OD): δ 2.33
(3H, s), 3. 34 (4H, t, J ) 5.2 Hz), 3.66 (4H, t, J ) 5.2 Hz), 5.35
(2H, s), 6.09 (1H, d, J ) 2.8 Hz), 7.05 (1H, d, J ) 7.2 Hz), 7.19

824 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 3
Wroblowski et al.
(1H, t, J ) 7.8), 7.29 (1H, d, J ) 7.2 Hz), 7.39 (1H, t, J ) 7.6),
7.48 (1H, t, J ) 7.2), 7.48 (1H, t, J ) 7.6), 7.70 (2H, d, J ) 8.8),
7.78 (2H, d, J ) 8.4), 8.02 (1H, d, J ) 2.8 Hz), 8.05 (2H, d, J )
8.4 Hz); m/z 502.5 [M + Na]+.
2-(3-Nitro-2-methylbenzyl)-4,5-dichloropyridazin-3-one 24. Tet-
ramethylguanidine (15 mL, 120 mmol) was added slowly to a
solution of 2-methyl-3-nitrobenzyl chloride (12.5 g, 67 mmol) and
4,5-dichloropyridazin-3-one (10 g, 61 mmol) in THF (200 mL).
The solution went green immediately and was allowed to stir at
room temperature overnight.
The solid precipitate was filtered off and washed with THF, and
the combined filtrates were evaporated to dryness. The residue was
dissolved in ethyl acetate, washed with 0.5 M HCl, an aqueous
solution of NaCl and water, dried with magnesium sulfate, and
evaporated to dryness. The crude product was recrystalised from
t-butylmethylether to give 7.93 g of the title compound 24 as a
highly crystalline yellow solid. 1H NMR (400 MHz, CDCl3): δ 2.54
(3H, s), 5.43 (2H, s), 7.30 (1H, t, J ) 8 Hz), 7.57 (1H, d, J ) 7.7
Hz), 7.71 (1H, d, J ) 8 Hz), 7.80 (1H, s).
2-(3-Amino-2-methylbenzyl)-4,5-dichloropyridazin-3-one 25. St-
annous chloride dihydrate (18.6 g, 82.7 mmol) was added to a
solution of 2-(3-nitro-2-methylbenzyl)-4,5-dichloropyridazin-3-one,
24 (5.2 g, 16.5 mmol), in ethanol (400 mL) and the mixture was
stirred overnight. A further aliquot of stannous chloride was added
(5 g, 22 mmol) and the solution was stirred for a further 24 h when
the reaction mixture was evaporated to dryness. The residue was
shaken with dichloromethane and aqueous sodium bicarbonate, and
the organic layer was separated from the resulting emulsion and
evaporated to dryness to give the title compound 25 as a yellow
solid (4 g). 1H NMR (400 MHz, CDCl3): δ 2.19 (3H, s), 3.64 (2H,
br s) 5.36 (2H, s), 6.66 (1H, d, J ) 7.6 Hz), 6.75 (1H, d, J ) 7.6
Hz), 6.99 (1H, t, J ) 7.6 Hz), 7.76 (1H, s); m/z 285 [M + H]+,
306 [M + Na]+.
Pd-C/H2O (1:1, 20 mg). The catalyst was removed by filtration
of the warm reaction mixture and, after evaporation of the solvent,
the residue was purified by reversed phase chromatography.
Trituration of the resulting solid with ether gave the title compound
1
36 (7 mg). H NMR (500 MHz, DMSO-d6): δ 2.20 (3H, s), 2.43
(1H, br. s.), 2.76 (t, J ) 4.9 Hz, 4 H), 3.26 (t, J ) 5.1 Hz, 4 H),
3.92 (s, 3 H), 5.20 (s, 2 H), 5.86 (d, J ) 2.8 Hz, 1 H), 6.88 (d, J
) 7.5 Hz, 1 H), 6.96 (d, J ) 8.6 Hz, 1 H), 7.17 (t, J ) 7.8 Hz, 1
H), 7.24 (d, J ) 7.7 Hz, 1 H), 7.84 (d, J ) 8.4 Hz, 2 H), 8.03 (d,
J ) 2.9 Hz, 1 H), 8.08 (d, J ) 8.3 Hz, 2 H), 8.12 (dd, J ) 8.6, 2.6
Hz, 1 H), 8.60 (d, J ) 2.5 Hz, 1 H), 10.00 (s, 1 H). 13C NMR (126
MHz, DMSO-d6): δ 13.5 (s), 44.9 (s), 46.6 (s), 50.6 (s), 53.3 (s),
99.2 (s), 110.6 (s), 125.3 (s), 125.6 (s), 126.1 (s), 126.1 (s), 128.2
(s), 128.3 (s), 129.8 (s), 132.3 (s), 133.0 (s), 136.5 (s), 136.5 (s),
137.6 (s), 139.8 (s), 145.0 (s), 150.0 (s), 160.2 (s), 163.4 (s), 164.8
(s); m/z 533 [M + Na]+, 511 [M + H]+; m/z found 511.24508,
C29H31N6O3 requires 511.24522.
Measurement of MrgX1 Activation by Putative Agonists.
Adherent HEK293 cells stably expressing the MrgX1 gene, were
seeded into black walled clear-base 384 well plates (Costar UK) at
a density of 20000 cells per well in minimum essential medium
with Earle’s salts and L-glutamine (Gibco 31095-029), supplemented
with 10% fetal bovine serum (Gibco 16000-044), 1% nonessential
amino acids (Gibco 11140-035) and 50 mg/mL Geneticin (Gibco
10131-027) and cultured overnight. The cells were then incubated
with media containing the cytoplasmic calcium indicator, Fluo-
4AM (2 µM) and extracellular quenching dye and 2.5 mM
probenecid at 37 °C for 60 min. The cells were washed twice with,
and finally resuspended in, Tyrode’s medium containing 2.5 mM
probenecid, The plates were then placed into a FLIPR (Molecular
Devices, UK) to monitor cell fluorescence (ex ) 488 nM, EM )
540 nM) before and after the addition of the test solutions.
Responses were measured as the maximum fluorescent count
reached minus the averaged baseline before compound addition.
Data was analyzed using a 0.25 IA level for single shot hits or had
curves fitted using a four-parameter logistic equation for dose
response experiments.
N-[3-(4,5-Dichloro-6-oxo-6H-pyridazin-1-ylmethyl)-2-methyl-
phenyl]-4-(6-methoxy-pyridin-3-yl)-benzamide 26. 2-(3-Amino-2-
methylbenzyl)-4,5-dichloropyridazin-3-one, 25 (1 g, 3.5 mmol) was
added to a mixture of HOAt (1.36 g, 10 mmol), dicyclohexylcar-
bodiimide (1.08 mL, 6.9 mmol), and 4-(6-methoxypyridin-3-
yl)benzoic acid (2.4 g, 10.4 mmol), which had been stirred in a
mixture of dichloromethane-DMF (1:1; 50 mL) for 1 h. After
stirring overnight, the reaction mixture was diluted with ether (400
mL) and the title compound 26 (1 g) was filtered off and washed
Intrinsic activity was calculated by taking an average of 16 maximal
BAM6-22 responses and expressing the maximal response observed
by the test compounds as a fraction of this value between 0 and 1.
Acknowledgment. We thank Mark Vine for expert NMR
acquisition and interpretation.
1
with ether. H NMR (400 MHz, DMSO-d6): δ 2.22 (3H, s), 3.92
(3H, s) 5.37 (2H, s), 6.96 (1H, d, J ) 8.4 Hz), 7.01 (1H, d, J ) 7.2
Hz), 7.19 (1H, t, J ) 8.0 Hz), 7.29 (1H, d, J ) 7.2 Hz), 7.84 (2H,
d, J ) 8.4 Hz), 8.08 (2H, d, J ) 8.4 Hz), 8.12 (1H, d, J ) 8.8 Hz),
8.27 (1H, s), 8.60 (1H, d, J ) 2.4 Hz) 10.00 (1H, s); m/z 495, 497
[M + H]+, 517, 519 [M + Na]+.
Supporting Information Available: NMR data for compounds
3a, 5a, and 36. Purity data for all resin unbound compounds.
CEREP screening data for compound 36. MrgX1 computational
model with bound ligand compound 22. This material is available
free of charge via the Internet at http://pubs.acs.org.
N-[3-(5-Chloro-6-oxo-4-piperazin-1-yl-6H-pyridazin-1-ylmethyl)-
2-methyl-phenyl]-4-(6-methoxy-pyridin-3-yl)-benzamide 22. N-[3-
(4,5-Dichloro-6-oxo-6 H-pyridazin-1-ylmethyl)-2-methyl-phenyl]-
4-(6-methoxy-pyridin-3-yl)-benzamide 26 (300 mg, 0.55 mmol) and
piperazine (57 mg, 0.66 mmol) were dissolved in DMF (50 mL)
and stirred overnight. The reaction mixture was evaporated to
dryness and the residue was triturated with ether (200 mL) to give
the crude product 22 (400 mg). An aliquot of this material (40 mg)
was purified by reverse phase chromatography to provide the title
compound 22 (12 mg). 1H NMR (400 MHz, DMSO-d6): δ 2.19 (3
H, s) 3.02-3.18 (4 H, m) 3.48-3.61 (4 H, m) 3.89 (3 H, s) 4.0-5.5
(2H, br s, exchange) 5.28 (2 H, s) 6.89-6.98 (2 H, m) 7.17 (1 H,
t, J ) 7.6 Hz) 7.31 (1H, d, J ) 8.8 Hz) 7.80 (2H, d, J ) 8.33 Hz)
7.97 (1H, s) 8.04 (2H, d, J ) 8.22 Hz) 8.07 (1H, dd, J ) 2.52 and
8.8 Hz) 8.28 (1H, br s) 8.54 (1H, d, J ) 2.3 Hz) 10.09 (1H, s,
References
(1) Butera, J. A. Current and Emerging Targets to Treat Neuropathic Pain.
J. Med. Chem. 2007, 50, 2543–2546.
(2) Kennedy, J. D. Neuropathic Pain: Molecular Complexity Underlies
Continuing Unmet Medical Need. J. Med. Chem. 2007, 50, 2547–
2556.
(3) Childers, W. E.; Baudy, R. B. N-Methyl-D-Aspartate Antagonists and
Neuropathic Pain: The Search for Relief. J. Med. Chem. 2007, 50,
2557–2562.
(4) Field, M. J.; Li, Z.; Schwarz, J. B. Ca2+ Channel R2-δ Ligands for
the Treatment of Neuropathic Pain. J. Med. Chem. 2007, 50, 2569–
2575.
(5) Munro, G.; Dalby-Brown, W. Kv7 (KCNQ) Channel Modulators and
Neuropathic Pain. J. Med. Chem. 2007, 50, 2576–2582.
(6) Kyle, D. J.; Ilyin, V. I. Sodium Channel Blockers. J. Med. Chem.
2007, 50, 2583–2588.
(7) Westaway, S. M. The Potential of Transient Receptor Potential
Vanilloid Type 1 Channel Modulators for the Treatment of Pain.
J. Med. Chem. 2007, 50, 2589–2596.
(8) Schkeryantz, J. M.; Kingston, A. E.; Johnson, M. P. Prospects for
Metabotropic Glutamate 1 Receptor Antagonists in the Treatment of
Neuropathic Pain. J. Med. Chem. 2007, 50, 2563–2568.
exchange); m/z 567 [M
C29H30ClN6O3 requires 545.20624.
+
Na]+; m/z found 545.20614,
N-[3-(6-oxo-4-Piperazin-1-yl-6H-pyridazin-1-ylmethyl)-2-methyl-
phenyl]-4-(6-methoxy-pyridin-3-yl)-benzamide 36. N-[3-(5-Chloro-
6-oxo-4-piperazin-1-yl-6H-pyridazin-1-ylmethyl)-2-methyl-phenyl]-
4-(6-methoxy-pyridin-3-yl)-benzamide, 22 unpurified from the
previous preparation (20 mg, ca. 0.36 mmol), was stirred at 70 °C
for 3 h as a suspension with hydrazine hydrate (0.2 mL) and 10%

Mas-Related Gene X1 Agonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 3 825
(9) Lai, J.; Luo, M-c.; Chen, Q.; Porreca, F. Pronociceptive actions of
dynorphin via bradykinin receptors. Neurosci. Lett. 2008, 437, 175–
179.
(10) Coa, H.; Zhang, Y-Q. Spinal glial activation contributes to pathological
pain states. Neurosci. BiobehaV. ReV. 2008, 32, 972–983.
(11) Burns, J. W.; Quartana, P. J.; Bruehl, S. Anger inhibition and pain:
conceptualizations, evidence and new directions. J. BehaV. Med. 2008,
31, 259–279.
(12) Guindon, J.; Walczak, J-S.; Beaulieu, P. Recent Advances in the
Pharmacological Management of Pain. Drugs 2007, 67, 2121–2133.
(13) Newsome, S.; Frawley, B. K.; Argoff, C. E. Intrathecal analgesia for
refractory cancer pain. Curr. Pain Headache Rep. 2008, 12, 249–256.
(14) Dong, X.; Han, S.; Zylka, M. J.; Simon, M. I.; Anderson, D. J. A
diverse family of GPCRs expressed in specific subsets of nociceptive
sensory neurons. Cell 2001, 106, 619–632.
(15) Lembo, P. M. C.; Grazzini, E; Groblewski, T.; O’Donnell, D.; Roy,
M-O.; Zhang, J.; Hoffert, C.; Cao, J.; Schmidt, R.; Pelletier, M.;
Labarre, M.; Gosselin, M.; Fortin, Y.; Banville, D.; Shen, S. H.; Strom,
P.; Payza, K.l; Dray, A.; Walker, P.; Ahmad, S. Proenkephalin A gene
products activate a new family of sensory neuron-specific GPCRs.
Nat. Neurosci. 20025, 201–209.
(16) Grazzini, E.; Puma, C.; Roy, M-O.; Yu, X. H.; O’Donnell, D.; Schmidt,
R.; Dautrey, S.; Ducharme, J.; Perkins, M.; Panetta, R.; Laird, J. M. A.;
Ahmad, S.; Lembo, P. M. C. Sensory neuron-specific receptor
activation elicits central and peripheral nociceptive effects in rats. Proc.
Natl. Acad. Sci. U.S.A. 2004, 101, 7175–7180.
(17) Yang, S.; Liu, Y.; Lin, A. A.; Cavalli-Sforza, L. L.; Zhao, Z.; Su, B.
Adaptive evolution of MRGX2, a human sensory neuron specific gene
involved in nociception. Gene 2005, 352, 30–35.
(18) Zylka, M. J.; Dong, X.; Southwell, A. L.; Anderson, D. J. Atypical
expansion in mice of the sensory neuron-specific MrgG protein-
coupled receptor family. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
10043–10048.
(19) Burstein, E. S.; Ott, T. R.; Feddock, M.; Ma, J-N.; Fuhs, S.; Wong,
S.; Schiffer, H. H.; Brann, M. R.; Nash, N. R. Characterization of the
Mas-related gene family: structural and functional conservation of
human and rhesus MrgX receptors. Br. J. Pharmacol. 2006, 147, 73–
82.
(24) Teitelbaum, H.; Catravas, G. N.; McFarland, W. L. Reversal of
morphine tolerance after medial thalamic lesions in the rat. Science
1974, 185, 449–451.
(25) Guo, R-x.; Cui, Y.; Mo, L-q.; Liu, W.; Feng, J-q. Changes of the
spinal δ opioid receptors and NMDA receptors during the development
of morphine tolerance. Jiepouxue Yanjiu 2007, 29, 88–92.
(26) Mendez, I. A.; Trujillo, K. A. NMDA receptor antagonists inhibit
opiate antinociceptive tolerance and locomotor sensitization in rats.
Psychopharmacology (Berlin, Ger.) 2008, 196, 497–509.
(27) Dalton, G. D.; Smith, F. L.; Smith, P. A.; Dewey, W. L. Alterations
in brain protein kinase A activity and reversal of morphine tolerance
by two fragments of native protein kinase A inhibitor peptide (PKI).
Neuropharmacology 2005, 48, 648–657.
(28) Parkitna, J. R.; Obara, I.; Wawrzczak-Bargiela, A.; Makuch, W.;
Przewlocka, B.; Przewlocki, R. Effects of glycogen synthase kinase
3 and cyclin-dependent kinase 5 inhibitors on morphine-induced
analgesia and tolerance in rats. J. Pharmacol. Exp. Ther. 2006, 319,
832–839.
(29) Smith, F. L.; Javed, R.; Elzey, M. J.; Welch, S. P.; Selley, D.; Sim-
Selley, L.; Dewey, W. L. Prolonged reversal of morphine tolerance
with no reversal of dependence by protein kinase C inhibitors. Brain
Res. 2002, 958, 28–35.
(30) Adam, F.; Dufour, E.; Le Bars, D. The glycine site-specific NMDA
antagonist (+)-HA966 enhances the effect of morphine and reverses
morphine tolerance via a spinal mechanism. Neuropharmacology 2008,
54, 588–596.
(31) Chen, T.; Hu, Z.; Quirion, R.; Hong, Y. Modulation of NMDA
receptors by intrathecal administration of the sensory neuron-specific
receptor agonist BAM8-22. Neuropharmacology 2008, 54, 796–803.
(32) Olsson, R.; Knapp, A. E.; Eskildsen, J. Preparation of benzimidazole
derivatives, especially N2-derivatized argininamides containing a
substituted benzimidazole unit as modulators of MrgX receptors for
the treatment of pain and methods of screening them. Patent
WO2008052072, 2008.
(33) Sullivan, E.; Tucker, E. M.; Dale, I. L. Measurement of [Ca2+] using
the fluorometric imaging plate reader (FLIPR). Methods Mol. Biol.
1999, 114, 125–133.
(34) Sali, A.; Blundell, T. L. Comparative protein modeling by satisfaction
of spatial restraints. J. Mol. Biol. 1993, 234, 779–815.
(35) McMartin, C.; Bohacek, R. S. QXP: powerful, rapid computer
algorithms for structure-based drug design. J. Comput.-Aided Mol. Des.
1997, 11, 333–344.
(36) Rauhut, G.; Clark, T. Multicenter point charge model for high-quality
molecular electrostatic potentials from AM1 calculations. J. Comput.
Chem. 1993, 14, 503–509.
(37) Schwartz, T. W.; Frimurer, T. M.; Holst, B.; Rosenkilde, M. M.; Elling,
C. E. Molecular mechanism of 7TM receptor activationsa global
toggle switch model. Annu. ReV. Pharmacol. Toxicol. 2006, 46,
481–519.
(20) Zhang, L.; Taylor, N.; Xie, Y.; Ford, R.; Johnson, J.; Paulsen, J. E.;
Bates, B. Cloning and expression of MRG receptors in macaque,
mouse and human. Mol. Brain Res. 2005, 133, 187–197.
(21) Kunapuli, P.; Lee, S.; Zheng, W.; Alberts, M.; Kornienko, O.; Mull,
R.; Kreamer, A.; Hwang, J. I.; Simon, M. I.; Strulovici, B. Identifica-
tion of small molecule antagonists of the human mas-related gene-
X1 receptor. Anal. Biochem. 2006, 351, 50–61.
(22) Tomaszewski, M. ; Discovery of pain targetssGPCR target class
approach, XVIII International Symposium on Medicinal Chemistry,
Copenhagen, Denmark, August, 2004.
(23) Cai, Q.; Jiang, J.; Chen, T.; Hong, Y. Sensory neuron-specific receptor
agonist BAM8-22 inhibits the development and expression of
tolerance to morphine in rats. BehaV. Brain Res. 2007, 178, 154–159.
JM800962K