Novel approach for chemotype hopping based on annotated databases of chemically feasible fragments and a prospective case study: New melanin concentrating hormone antagonists

Article abstract of DOI:10.1021/jm8016199

A novel strategy for chemotype hopping, based on annotated databases of chemically feasible fragments and their oriented functionalization, is presented. A three-dimensional (3D) similarity analysis of projectoriented functionalized scaffolds provides a p

Full text of DOI:10.1021/jm8016199

2076
J. Med. Chem. 2009, 52, 2076–2089
Novel Approach for Chemotype Hopping Based on Annotated Databases of Chemically Feasible
Fragments and a Prospective Case Study: New Melanin Concentrating Hormone Antagonists
Julen Oyarzabal,*,†,§ Trevor Howe,# Jesu´s Alcazar, Jose Ignacio Andre´s, Rosa M. Alvarez,§, Frank Dautzenberg,
Laura Iturrino, Sonia Mart´ınez,§, and Ilse Van der Linden
Departments of Molecular Informatics and Medicinal Chemistry, Johnson & Johnson Pharmaceutical R&D, Jarama 75, 45007 Toledo, Spain,
and Department of Molecular Informatics and CNS Biology Department, Johnson & Johnson Pharmaceutical R&D,
Turnhoutseweg 30, 2340 Beerse, Belgium
ReceiVed December 20, 2008
A novel strategy for chemotype hopping, based on annotated databases of chemically feasible fragments
and their oriented functionalization, is presented. A three-dimensional (3D) similarity analysis of project-
oriented functionalized scaffolds provides a prioritized proposal for synthesis with the most appropriate
linkers and optimal regiochemistry on R-groups. This strategy maximizes the potential of proprietary and
commercially available compounds. A retrospective and prospective case study, on melanin concentrating
hormone (MCH) antagonists, showing the impact on the drug discovery process of this new strategy by
maintaining primary activity and improving key ADME/Tox property while enhancing intellectual property
(IP) position is demonstrated.
Scheme 1. Flow Chart of the Proposed Strategy with Two Main
Phases: 1, Generation of Annotated DBs of Chemically Feasible
Fragments; 2, Chemotype Hopping Based on DBs Previously
Generated
Introduction
A primary goal of drug discovery is to reach the clinical phase
with a candidate molecule having viable pharmacokinetic
parameters, a low toxicity profile, and a solid intellectual
property position. A failure in any one of the multifarious
requirements of a candidate drug may thwart or terminate this
process; therefore, it is preferable to pursue multiple chemotypes,
so bioisosteres to the main chemical series are often sought.
This is known as scaffold hopping. In recent years, the term
scaffold has been used extensively in the context of describing
the central core component of a molecule.1 According to
Schneider, scaffold hopping is defined as “identification of iso-
functional molecular structures with significantly different
backbones”,2 therefore keeping the key interactions between
scaffold and receptor while changing its chemical structure. The
assumption in this process is that the key interactions with the
receptor are preserved, maintaining binding affinity while
modifying drug-like properties and improving IP position. In
this work the scaffold hopping is focused on the central skeleton
replacement while maintaining the key pharmacophoric features
from a well-defined substitution pattern of R-groups transferred
to the new scaffold(s). Only a few methods3,4 address the central
skeleton as a fragment that can be replaced while keeping the
R-group substituents. Published 3D scaffold hopping methods
outperform bidimensional (2D) methods,5,6 and there is one
reported method successfully linking 3D molecular properties
of the scaffold with geometrical information and synthetic
feasibility.4 Ideally, screening methods that demonstrate suc-
cessful scaffold hopping should find not only a maximal number
of chemotypes but also a diverse set.7
To define a systematic approach ready to look for alternative
chemotypes, a general strategy was implemented as shown in
Scheme 1. The process is as follows: (a) build annotated
database (DB)a of chemically feasible fragments describing the
number of diversity points, regiochemistry, number of rings,
and origin; (b) generate a virtual library according to the
substitution pattern for the reference compound (only the atoms
closest to the main chemotype borne by the reference compound
* Author to whom correspondence should be addressed (telephone +34
91 7328000; fax +34 91 7328051; e-mail joyarzabal@cnio.es).
†
Department of Molecular Informatics, Toledo.
Present address: Spanish National Cancer Research Centre (CNIO),
§
Experimental Therapeutics Programme, Melchor Fernandez Almagro 3,
28029 Madrid, Spain.
#
a Abbreviations: MCH, melanin-concentrating hormone; VL, virtual
library; ET, electrostatics Tanimoto; TS, Tanimoto shape; EF, enrichment
factors; R, recall; TA, therapeutic area; DB, databases.
Department of Molecular Informatics, Beerse.
Department of Medicinal Chemistry, Toledo.
CNS Biology Department, Beerse.
10.1021/jm8016199 CCC: $40.75
2009 American Chemical Society
Published on Web 03/16/2009

Melanin Concentrating Hormone Antagonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7 2077
such as ClogP, molecular weight, and polar surface area (PSA) as
calculated by ACD software11 (Table 1).
Johnson & Johnson (J&J) Corporate Database. Using an
updated version of the previously reported J&J database,17 currently
more than 1.5 million compounds, the above-described program
was utilized to obtain Onion0 and Onion1 fragment databases.
Generation of each of those two databases took ∼15 h in a Linux
workstation (as CPU, Dual Core AMD Opteron 2.3 GHz and 8GB
RAM). Onion0 fragmentation, utilized in the case study reported
below, drove to a database with 42,499 unique fragments. Data
mining is required to archive only those fragments useful for
chemotype hopping; therefore, the following criteria were applied:
(i) fragments bearing at least one ring, (ii) molecular weight between
60 and 300, and (iii) number of diversity points from 1 to 4. Thus,
the Onion0 J&J corporate database contains around 12,585 unique
fragments fitting these criteria. An identical procedure has been
applied to our database of compounds accessible from external
suppliers, ready to be acquired; in this case, we have around 39,912
unique fragments for Onion0 meeting the criteria described above.
Finally, to define therapeutic area DB, in-house information from
J&J together with data from external databases, such as Integrity
by Prous Science,18 was utilized to build up CNS oriented fragments
DB, where Onion0 CNS DB has 4,686 unique annotated fragments
and there are around 10,872 Onion1 fragments. Once the fragment
databases for each of these, at least four, databases (corporate,
external suppliers, therapeutic area, or target family oriented and
MedChem) are ready to be used, then, project-oriented chemotype
hopping may start.
Fragments DBs in Project-Oriented Chemotype Hopping.
Different strategies are applied for Onion0 and Onion1 fragment
DBs, respectively, in this proposed approach for chemotype
hopping. Regarding Onion0, the first step is defining which
fragments from this database are used in each case; then, only those
scaffolds meeting certain criteria are selected, for example, a certain
number of rings and number of diversity points. This decision is
guided by the reference structure. The next step is the identification
of the closest key minimal substitution pattern around the reference
compound at one atom distance from the central core. Then, this
substitution pattern is utilized together with other potentially
interesting linkers or substitutions to build up a virtual library (VL)
using the annotated anchor points, their regiochemistry, from
previously selected scaffolds; this is illustrated below in the case
study. Regarding Onion1, all anchor points are capped as methyl
and the corresponding capped Onion1 fragment DB is further
utilized without any additional manipulation in any drug discovery
project where it is required (Table 2); thus, no virtual library is
built.
Once the focused virtual library from the Onion0 fragment DB
is built and Onion1 DB is properly capped, the corresponding
conformers for each of these two set of compounds are generated.
Software developed by OpenEye,19 Omega,20 is utilized to explore
the conformational space around each functionalized chemotype
within an energetic window of 25 kcal/mol, where conformers with
root of mean squared deviation (rmsd) >0.4 Å are stored until a
maximum number of 500.21 This is the most time-consuming step;
a library with 58,565 structures takes ∼19 h using the workstation
described above. Fragments from Onion1 fragmentation are not
further functionalized, just capped; therefore, once their corre-
sponding conformer database is constructed, they can be utilized
against any reference compound for any project without any
additional manipulation.
Similarity Analyses. Once all conformers, from Onion0 project-
oriented functionalized fragments and from Onion1 capped frag-
ments, have been generated, we have to align them with the
corresponding reference compound before doing any further
analysis. In this case we utilize the ROCS (rapid overlay of chemical
structures) package,22 by Open Eye, to perform the alignment
because this is a fast and accurate23 method for superimposing
molecules. Shape similarity can be determined, in part, by compar-
ing the shapes of those molecules. This effectively reduces to
calculating the overlap volume between two molecules and has been
Figure 1. Illustration of the fragmentation script applied to compound
a provides as result from “Onion0” the closest fragmentation around
each ring system, e.g., driving to fragment b, pyridine (color-coded in
blue), with its corresponding growing vectors. In this case, pyridine,
more detailed information is provided through “Onion1” fragmentation
c, color-coded in pink, including some additional functionalities. Then,
from compound a, phenyl and cyclopropyl are also considered as ring
systems, and the same fragmentation exercises, Onion0 and Onion1,
are performed around them.
are considered to construct the virtual library); (c) conformer
generation of the constructed virtual library; (d) alignment of
conformers to the reference structure and then prioritization
based on shape and electrostatic similarity in 3D, therefore
avoiding bidimensional chemical structure.
In addition to a detailed description of the new strategy for
chemotype hopping, a retrospective and prospective case study
is reported in this paper to validate and illustrate the impact of
this approach on the drug discovery process.
Method
Database Preparation. All fragments included in these databases
were extracted from compounds already synthesized and so, by
definition, chemically feasible. Compounds are extracted from
corporate databases, external supplier databases, therapeutic area
(TA) databases (in this case, central nervous system, CNS), target
families related ligands databases, and a database based on
MedChem experience. Before doing any fragmentation, rare ele-
ments and salts are removed. Structures are standarized, through
tautomer generation and their canonization. Duplicates are elimi-
nated by using a customized Pipeline Pilot8 protocol. Fragment
abstraction is performed at different levels from original compound
databases, by using a program coded in the scientific vector
language (SVL) of the software system MOE.9 The SVL script
utilized in this study was developed in collaboration with the
Chemical Computing Group (CCG) and is publically available. In
general, two fragmentation levels are utilized, Onion0 and Onion1,
so each database is in duplicate with fragments from each of those
two levels. Onion0 fragmentation level provides structures from
the closest fragmentation around each ring system contained within
compounds, plausible central scaffolds, driving to “naked” chemo-
types decorated only with their corresponding growing vectors or
anchor points. Onion1 fragmentation delivers a more elaborate
structure with not only the information for the atom at a distance
of one atom from each ring system but also information about the
functionality of that atom. If the atom belongs to a functional group
such as an amide or ester, for example, or is part of an aromatic
ring, its whole functionality is conserved as shown in Figure 1.
Sometimes functionalities close to the central core play a key driving
force in ligand-receptor interactions together with main chemotype.
Therefore, having these two fragmentation levels, Onion0 and
Onion1, provides an additional value to our fragment database in
comparison with those derived from existing approaches for
fragment generation.10
All fragments included in the databases have at least one ring
and are annotated with the most relevant information such as
regiochemistry, the position for anchor points from where chemo-
types can be functionalized, the number of diversity points, the
number of fused rings, molecular weight, source, phase (preclinical,
phase I, II, or III), and some other in-silico estimated properties

2078 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7
Oyarzabal et al.
Table 1. Representative Set of Compounds from Database Obtained after Applying Fragmentation Script at Two Levels, Onion0 and Onion1, for CNS
Reported Compounds
Table 2. Representative Set of Onion1 Fragments, for CNS Reported
Compounds, Capped
using ROCS, providing shape similarities versus reference com-
pound, and electrostatic potentials calculated using OpenEye’s ZAP
Poisson-Boltzmann solver implemented in EON. In this case the
ET is calculated using an external dielectric of 80, thus accounting
for dampening of electrostatic field by aqueous solvent, and may
better represent the field experienced upon binding a protein.25
A
normalized shape Tanimoto of unity indicates that the molecules
are identical, and the Tanimoto tends to zero as molecules become
less similar; the electrostatic Tanimoto ranges from -0.3 to +1,
indicating dissimilar and similar electrostatic fields.19 From each
original fragment library top-ranking compounds, in terms of shape
and electrostatics, may be selected.
Biological Target Utilized in the Described Case Study
and Reported Known Ligands for the Target. We were interested
in melanin-concentrating hormone (MCH) as a therapeutic target
in one of our projects at J&J. This was an outstanding opportunity
to validate this novel strategy for chemotype hopping from two
points of view: (a) retrospective validation, because there are several
published patents and papers describing different MCH-R1 antago-
nists (Table 1), and (b) prospective validation, which is the impact
of this new approach in a real scenario of a drug discovery project,
looking for a new chemotype in an active project. Thus, this was
our case study to check the value of this novel strategy. MCH is a
cyclic nonadecapeptide expressed in the lateral hypothalamus and
the natural ligand for the seven-transmembrane G-protein-coupled
receptors known as MCH-R1 and MCH-R2.26-28 Biological
function for MCH-R2 remains unclear to date,29 but many published
studies indicate that MCH-R1 is involved in biological processes
related to mammalian feeding behavior and energy expenditure.30,31
In addition, it has been suggested that MCH-R1 plays a key role
in anxiety and mood disorders because MCH-R1 antagonists have
been found to have anxiolytic and antidepressant effects in various
animal models.32,33 MCH-R1 is a very attractive target because an
orally active MCH-R1 antagonist could find useful applications in
two disease areas in high need for novel and effective therapies:
obesity and depression/anxiety disorders.29 Despite the large number
of drug discovery programs dedicated to finding small-molecule
MCH-R1 antagonists for the treatment of obesity and/or mood
previously reported.22 Once the overlap has been optimized, the
shape similarity may be computed by Tanimoto equation. However,
ROCS does not contain an accurate notion of charge distribution
and therefore is not a complete solution to the search for molecular
similarity. The electrostatic fields of molecules can be calculated
and the similarity between fields expressed as the electrostatic
Tanimoto. Electrostatic Tanimoto (ET) scores are calculated using
the EON program,24 by Open Eye. Molecules are superimposed

Melanin Concentrating Hormone Antagonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7 2079
Figure 2. Melanin-concentrating hormone receptor 1, MCH-R1, antagonists in the clinic. Compound 2 is our reference structure where in magenta
is the central scaffold and in blue its closest substitution pattern which is key to build the VL from Onion0.
Figure 3. Takeda compound and its key pharmacophoric features as a bidimensional representation (2D): in green, aromatic as chemical feature;
in red, hydrogen bond acceptor (in this case, projection points are represented); in blue, titratable nitrogen at physiological pH.
disorders (more than 15 pharmaceutical companies have been
actively looking for MCH-R1 antagonists), a very limited number
have progressed into the clinic; in fact, only three compounds have
entered phase I clinical trials: AMG-076 1,34 Amgen; GW803430
2,35 GlaxoSmithKline; and NGD-4715 3,36 Neurogen (Figure 2).
The first report of oral activity by MCH-R1 antagonist came from
Takeda, T226296 4,37 and this biphenyl amide served as a template
in the design of a number of series of MCH-R1 antagonists,
constrained analogues, from other companies; a well-defined set
of key pharmacophoric features (Figure 3) is kept constant among
all of them (5-19)38-52 (Figure 4). Beyond the common challenges
in drug design related to ADME and safety profiles, cardiovascular
risk involving the human ether-a-go-go related gene (hERG) binding
activity and drug-induced QTc prolongation has been the major
hurdle for a significant number of MCH-R1 research programs29,36
due to this frequently being associated with potential lethal
arrhythmias known as torsades de pointes (TdP), which has led to
the removal of several drugs from the market.53
Finding optimal small-molecule MCH-R1 antagonists is a real
challenge due to (a) target attractiveness, two potential therapeutic
applications; (b) cardiovascular risk associated with many reported
MCH antagonists; and (c) several pharmaceutical companies36
focused on the same aim; in fact, from 2000, there were more than
55 published patents before ours. Therefore, this was an optimal
case to utilize and validate this novel strategy for chemotype
hopping: looking for novel chemical series circumventing the
cardiovascular issue, keeping primary activity, and providing strong
IP position.
linkers and key substitutions illustrated in the reference sub-
structure, methyl and phenyl, because this may have an impact
on the electrostatics of the main chemotype of the reference
structure. In addition to these two key R-groups (blue-colored
in reference structure 20), N-methyl, S-methyl, and methoxy
were considered as potential alternatives to methyl; therefore,
eight potential virtual libraries were constructed. Considering
all scaffolds from the Onion0 fragment DB meeting the
previously reported criteria and the additional 15, the corre-
sponding virtual libraries (VL) were generated: (a) VLA, keeping
phenyl at position R1 and four different options at R2: methyl,
methoxy, thiomethyl, and methylamino; and (b) VLB, in which
those four different options will be borne, in this case, at R1
and phenyl at position R2 (Figure 6). Selection of R-groups to
build this virtual library was mainly based on the closest
substitution pattern around central cores from reported reference
compounds (Figures 2, 3 and 4); however, any R-group may
be included in this analysis. The most critical stage of this
approach is defining the closest substitution pattern for reference
scaffold as only one reference structure has to be utilized and
many options may be feasible.
These libraries are constructed using MOE software,9 and their
generation is very fast, just a few seconds. Then, to remove any
compound with unwanted chemistry (e.g., any nitrogen-oxygen
bond) a Pipeline Pilot8 protocol was built and run, obtaining the
final, project-based, virtual library from annotated fragments. In
total, this final VL contains 58,565 functionalized chemically
feasible fragments: 10,811 unique chemotypes. The corresponding
conformer database was constructed for this VL, using Omega,
and all conformers were aligned with the minimized reference
substructure, 20, using ROCS; the latest process took just 20 min.
Then, once overlap was optimized, 3D similarity analysis based
on electrostatic fields was computed by Tanimoto equation
implemented in EON, as described above. This final step is not
time-consuming either, only 55 min for the whole library.
Results and Discussion
Retrospective Analysis. Once the new strategy has been
defined and its corresponding infrastructure, fragment databases,
set up, the approach was validated. Taking advantage of the
considerable number of diverse chemotypes reported as MCH-
R1 antagonists, 15 different central cores (Figure 4), scaffold
hopping was performed. In this case, the central core from the
well-known MCH-R1 antagonist in the clinic, 2, together with
its closest substitution pattern (magenta- and blue-colored,
respectively) is utilized as reference structure (Figure 5).
According to the flowchart implemented with this new strategy
(Scheme 1), the first step for project-oriented chemotype hopping
is applied: those chemotypes that might fit better with reference
substructure, that is, fragments with one or two fused rings and
two diversity points, were selected from the Onion0 database.
Then, these fragments together with those 15 chemotypes from
reported MCH-R1 antagonists, in magenta at Figure 4, were
utilized to build up virtual libraries bearing as decoration those
3D similarity values, shape and electrostatics, were computed
for the whole virtual library (Figure 7A); therefore, we were
able to assess the performance of this analysis using the reported
known MCH-R1 antagonists as hits, 5-19. For this assessment
two different points of view were considered: (a) matching not
just the chemotype (colored in magenta) but also the substitution
pattern (colored in blue), therefore using as the total number of
potential samples the whole virtual library, 58,565 compounds;
and (b) identifying the chemotype, independent of its substitution

2080 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7
Oyarzabal et al.
Figure 4. Patented MCH-R1 antagonists bearing different chemotypes (in magenta) but keeping the same pharmacophoric features as compound
in clinical phases, 2, which is a constrained analogue of Takeda’s structure, 4.35b The optimal substitution pattern we will look for during our
scaffold hopping approach, in addition to the chemotype, is in blue.
Figure 6. Virtual library generated from those scaffolds from Onion0
Figure 5. (A) Key substructure 20, from reference compound 2,
as sum of VLA and VLB. For R-groups, attachment points are defined
utilized as template for 3D similarity analysis; (B) electrostatic map,
by pink balls.
obtained with EON, for the reference substructure 20 utilized for
chemotype hopping.
were estimated at different percentages of total number of
pattern; in this case, the total number of samples to consider in
the statistical analysis is 10,811 unique chemotypes.
Analyzing similarity values, according to their ranking
position based on shape and electrostatics similarity values, we
identified at which levels from the whole library those 15 known
hits were placed. Thus, corresponding enrichment factors (EF)
analyzed samples. These metrics (EF) were utilized not just to
assess the value of the approach, retrospectively in this case,
but also, together with its corresponding recall (R), to define
the optimal percentage of the library to be analyzed in a
prospective analysis in order to have higher probability to find
hits with a reduced number of false positives. In general, queries

Melanin Concentrating Hormone Antagonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7 2081
Figure 7. (A) 3D similarity scores: Tanimoto shape and electrostatics values after comparison of the whole virtual library versus reference substructure
20. (B) Boost plots of the percent hits identified versus percentage of the 3D similarity: score-ordered list selected considering both hit definitions.
Figure 8. Six hits, chemotypes (in magenta), with proposed optimal substitution pattern (in blue).
with high EF, equivalent to high precision (P), and low recall
(R) are likely to be overtrained and therefore poor in prediction,
whereas a higher recall and lower precision query may be more
appropriate in a virtual screening context;55 therefore, combina-
tions of these two parameters, EF and R, were considered to
get the optimal balance and, thus, prediction. Formulas for the
EF and R are usually expressed as follows (eqs 1 and 2
respectively):55,56
Tanimoto (ET) is >0.62 and for Tanimoto shape (TS) >0.75.
However, considering as hits those compounds just fitting the
main chemotype (magenta-colored substructures in Figure 4;
case b, reported above), the results are slightly worse (magenta
line in Figure 7B). Thus, using identical criteria as above for
ET and TS, 0.62 and 0.75, the top 0.9% ranking of all scaffolds
is analyzed: 93 chemotypes of 10,811. For this setup the EF is
23.2; 3 hits were identified, so its corresponding recall is 0.2.
To identify optimal balance, and therefore prediction, between
enrichment factor and recall, different combinations for those
3D similarity values reported in Figure 7A have been evaluated
as thresholds. In those cases where the threshold value for ET
is >0.56 and TS >0.75, the top 1.3% of the whole library (777
compounds; EF ) 30.1 and R ) 0.4) or the top 2.5% of the
whole set of chemotypes (275 scaffolds; EF ) 15.7 and R )
0.4), a more balanced situation between EF and R values is
obtained. The most important point to highlight from these
results, together with this improvement in the balance between
EF and R, is the structural diversity provided by those six hits
identified in both scenarios: among them and with reference
substructure 20. It is noteworthy that each of these six hits has
been patented by a different pharmaceutical company (Figure
8): four hits contain two fused rings, two of those hits contain
two fused six-member rings; therefore, they are different from
the reference substructure. In addition, there was one hit where
the chemotype was just one ring, an aromatic six-member ring,
and, finally, one hit with an aliphatic five-member ring as
scaffold. This diversity degree within the top-ranked hits, where
hit means identifying a different chemotype (no analogue),
provides a very important additional value to those metrics (EF
and R) utilized for the assessment of this approach.
EF ) (L × VA)/(V × LA)
R ) VA/LA
(1)
(2)
Here L and LA mean the total number of compounds in a library
and the number of experimentally verified bioactive compounds
in the library, respectively. V and VA mean the number of
compounds predicted as active and the number of known
bioactive compounds contained in V, respectively. An enrich-
ment factor of 1 is equally good as random selection. Recall
indicates how many of those known active molecules are
selected from the whole set of evaluated structures; a recall value
of 0.5 means 50% of hits are identified.
The blue line represented in Figure 7B reports those results
(enrichment factor) obtained after considering as hits those
compounds from the virtual library that fit chemotypes and
substitution pattern (and their regiochemistry) as they are
described in Figure 4, magenta and blue substructures (case a,
reported above). In this scenario, the highest EF is obtained
after analyzing 0.4% of the whole library (234 compounds),
where 3 hits, of 15, were found: EF ) 50.1 but R ) 0.2; this
EF is similar to the best value obtained in a previously reported
scaffold hopping where a broad set of different virtual screening
methods, 2D and 3D, were evaluated: EF ) 50.0, after analyzing
2% of the DB.6 In this case, the threshold value for electrostatic
The lowest limit considered to get a good balance between
enrichment factor and recall is defined by ET similarity values

2082 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7
Oyarzabal et al.
Figure 9. (A) Three additional hits, chemotypes (in magenta), with proposed optimal substitution pattern (in blue). (B) Heat map computing 2D
structural similarity, where dark blue means Tanimoto fingerprint (TF) value of <0.3, light blue for TF < 0.5, and green for TF g 0.5.
>0.50 and TS >0.75. In the similarity space defined by these
two values the enrichment factor is >10, and recall has a good
value, too, 0.6. Looking at those compounds placed at this area,
the chemical diversity of those three additional identified hits
must be highlighted (Figure 9A): One hit is based on a scaffold
built by two fused rings, a seven-member ring and a phenyl, an
additional hit is a six-member aliphatic ring, and, finally, we
have a five-member ring, a cyclic urea, as the main chemotype
for the last hit. The heat map show in Figure 9B, based on 2D
similarity Tanimoto values, quantifies 2D structural similarity
among the chemotypes (in magenta) from these nine identified
hits and reference substructure 20, confirming, objectively, the
degree of 2D structural diversity among those scaffolds borne
by the identified hits. A detailed analysis around the reference
chemotype 20, the box defined by a bright green line, shows
that only two of those nine identified chemotypes have a 2D
similarity >30%, always lower than 42% (Figure 9B). Two-
dimensional structural similarity for these chemotypes (in
magenta) together with their proper substitution pattern (in blue)
was generated as well to analyze the 2D similarity degree among
them; results indicate that only three structures have 2D
similarity values >30%, always lower than 50% (this plot is
available as Supporting Information, Figure S1). This compari-
son highlights the poorer sensitivity of the 2D similarity; most
of the compounds are collapsed in a range between 0 and 0.3,
thus stressing the use of these 3D approaches to perform a proper
prioritization among potential bioisosteric replacements, as was
described in previous papers.5,6 Two-dimensional structural
similarity analyses are based on circular molecular fingerprints
using ECFP_6 descriptors,57 implemented in Pipeline Pilot,8 that
define molecular structure using radial atom neighborhoods.
Retrospective analysis, to validate this new stategy for
chemotype hopping, provided us three important keys: the first
two assess (a) a clear metric, very good enrichment factors (in
a scenario where no analogue is possible), and (b) an additional
qualitative added value, confirmed by the heat map, a good
degree of structural diversity in top-ranked identified chemo-
types; the third key to be applied in a prospective analysis is
(c) identification of optimal thresholds, for similarity values (ET
and TS), to achieve good balance between enrichment factor
(EF) and recall (R) and, thus, prediction.
Prospective Analysis. Finding a small molecule, from novel
chemical series, that is a MCH-R1 antagonist circumventing
the cardiovascular issue, keeping its primary activity, and
providing strong IP position is a real challenge. However, the
positive results obtained from this retrospective analysis encour-
age us to proceed and try to identify new scaffolds by using
this new strategy for chemotype hopping. Therefore, the aim
for this prospective analysis, applied to an active project, was
replacing reference substructure 20 that comes from a compound
in clinical phases, 2, by a new chemotype meeting those criteria
described above. Thus, taking advantage of the setup and
analysis done in the retrospective analysis, we do not need any
additional virtual library from Onion0 and, on the other hand,

Melanin Concentrating Hormone Antagonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7 2083
Figure 10. (A) Plausible alternative chemotype, pyrazinone, as central core from these top-ranked structures based on their 3D similarity values
vs reference substructure 20. (B) Impact of linkers regiochemistry on electrostatics (ET), comparison between 24 and 22. (C) Graphical comparison,
based on electrostatic maps, between 20 and 21. (D) 3D alignment of compound 2, in magenta, with 37, which bears structure 23: Omega and
ROCS were utilized to perform this alignment.
Scheme 2a
a (i) NaH (60%), cyclohexanol, 1,2-dimethoxyethane, reflux, 30 min; (ii) NaOH (50%), DMSO, reflux, 1.5 h; (iii) 1-[2-(4-bromo-2-methoxyphenoxy)-
ethyl]pyrrolidine, CuI, N,N′-dimethylethylenediamine, potassium phosphate, dioxane:DMF 5:1, µW 180 °C, 15 min; (iv) POCl3, 1,2-dichloroethane, µW,
150 °C, 30 min; (v) NaH (60%), phenethyl alcohol, 1,2-dimethoxyethane, µW, 80 °C, 10 min.
the Onion1 conformer database is always ready to be launched.
In this case, we were focused on Onion0 and performed a
detailed analysis for those top-ranking hits from the virtual
library previously described, 58,565 compounds.
R ) 0.6 (Figure 10). Comparing electrostatic similarity values
between virtual compounds 22 and 24 highlights the impact of
this approach to identify not just the most appropriate linkers
around the proposed chemotype but also their optimal regio-
chemistry and, therefore, to address medicinal chemistry efforts.
Once pyrazinone was identified as a potential alternative
scaffold to 20, to check the hypothesis, the synthesis of three
tool compounds bearing this chemotype and the optimal linkers
with the appropriate regiochemistry, 21-23, was evaluated.
Regarding R-groups around chemotype, the structural informa-
tion described by all MCH-R1 patented compounds, 2 and
4-19, which fit the pharmacophoric features described in Figure
3, was taken into account; thus, this knowledge was transferred
to the selected new chemotype, with these three linkers, leading
to compounds meeting pharmacophoric criteria. Therefore,
compounds 30, 34, and 37 were proposed and synthesized
according to strategies depicted in Schemes 2-4. Two-
dimensional differences between compound 2 and proposed
compounds bearing this new chemotype are not spectacular, as
Figure 10D shows, but drove us to an optimal starting point
Considering the information obtained in the previous analysis,
we were initially focused on those top-ranking compounds with
scores for electrostatic Tanimoto (ET) >0.62 and shape >0.75;
therefore, we looked at the area (Figure 7A) with highest values
for 3D similarity to reference substructure 20, where we obtained
the highest enrichment factors and a moderate recall. Among
those 234 compounds, 93 unique chemotypes, there is an
interesting virtual compound, 21, built from a chemically
feasible scaffold (in magenta), pyrazinone, that might provide
an interesting IP position. Further analyses around this scaffold
showed that there were more functionalized fragments bearing
this chemotype, among those structures within the constructed
virtual library, placed at those areas where there is still a good
enrichment factor and the recall is clearly optimized: ET values
greater than 0.56 or 0.50, where the latter is the lower limit
considered to obtain a reasonably good balancesEF > 10 and

2084 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7
Oyarzabal et al.
Scheme 3a
a (i) trans-3-Phenylpropen-1-yl-boronic acid, Pd(Ph3P)4, NaHCO3 (1 M), dioxane, µW, 150 °C, 0.5 h; (ii) H2, Pd/C 10%, MeOH, RT, 4 h; (iii) HCl (6
N)/THF 2:1 µW 140 °C, 1 h; (iv) 1-[2-(4-bromo-2-methoxyphenoxy)ethyl]pyrrolidine, CuI, N,N′-dimethylethylenediamine, potassium phosphate, dioxane:
DMF 9:1, µW 175 °C, 30 min.
Scheme 4a
a (i) Phenethylamine, DBU, acetonitrile, µW, 180 °C, 20 min; (ii) NaOH (50%), DMSO, µW, 150 °C, 30 min; (iii) 1-[2-(4-bromo-2-methoxyphenoxy)-
ethyl]pyrrolidine, CuI, N,N′-dimethylethylenediamine, potassium phosphate, dioxane:DMF 9:1, µW 175 °C, 20 min.
(hit), improving key ADME/Tox property while keeping primary
activity and enhancing IP position through a systematic approach.
To prepare the target compounds in a fast and efficient way,
microwave irradiation was used in an extensive way.58 For the
synthesis of target compound 30 (Scheme 2), cyclohexanol was
used as protecting group. This group allowed us to hydrolyze
selectively the chlorine atom and obtain pyrazinone 27. Sub-
sequent copper coupling with the corresponding bromoarene
under microwave irradiation afforded compound 28. Cleavage
of the protecting group and subsequent chlorination was
achieved in one step using phosphorus oxychloride under
microwave irradiation to afford compound 29. Finally, target
compound 30 was obtained by nucleophilic aromatic substitution
of intermediate 29 with the sodium salt of phenethyl alcohol.
The key step for the preparation of target compound 34 was
the formation of the carbon-carbon bond at position 3 of the
pyrazinone (Scheme 3). This was achieved by Suzuki reaction
of commercially available 2-chloro-3-methoxypyrazine 31 with
trans-3-phenylpropen-1-ylboronic acid. Further reduction of the
double bond by catalytic hydrogenation and acidic hydrolysis
of the methoxy group of the intermediate obtained afforded
compound 33, which was coupled with the bromoarene under
microwave irradiation conditions to afford target compound 34.
Preparation of target compound 37 was fast and efficient using
a sequence of three steps under microwave irradiation with a
total reaction time of 70 min. The first step was the introduction
of phenethylamine to get intermediate 35. Then hydrolysis of
the chlorine atom and subsequent coupling with the bromoarene
afforded the desired compound 37.
Once these three compounds, 30, 34, and 37, were synthe-
sized, they were tested in a cell-based FLIPR assay54 measuring
their functional antagonism. In all cases, these newly synthesized
compounds were equipotent to the reference compound 2. In
addition, due to the cardiovascular risk associated with many
reported MCH antagonists, hERG activity59 was measured as
well driving to in vitro binding affinities >5 times lower than
the reference compound in clinical phases, 2 (Table 3).
Through this prospective analysis, real case study, the impact
that this new strategy for chemotype hopping may have in drug
discovery programs has been clearly exemplified. In this
particular case, looking for MCH-R1 antagonists, this new
strategy for scaffold hopping drove us to (a) compounds from
a new chemically feasible series, where primary functional
activity in cells was kept constant and equipotent, (b) a
chemotype with good IP position, in fact, a new patent was
filled around it,60 and (c) changes in the ADME profile; in this
case, their affinity for hERG was reduced, >5 times lower. Thus,
this case perfectly fits with scaffold hopping definition described
in the Introduction: alternative chemical structure, ideally with
optimal IP position, preserving the original profile as ligand and
binding affinity and improving its drug-like properties.
In addition to the pyrazinone ring, among those 93 unique
chemotypes from the top-ranking 234 compounds, there are
chemotypes without the amide moiety, which is present in most

Melanin Concentrating Hormone Antagonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7 2085
Table 3. Experimental IC50 Values, MCH-R1 FLIPR and hERG, for Reference Compound 2 and Newly Synthesized Structures 30, 34, and 37
a All values are mean values (SD in parentheses) and are derived from at least three independent experiments. b The binding activity of compounds is
represented as means of two independent experiments. Only differences in pIC50 up to 0.6 (SD < 0.5) were considered as reproducible and were maintained.
c Not determined.
of the scaffolds from the reported 2 and 5-19 MCH-R1
antagonists (in magenta). These additional scaffolds, not
included in this paper, provide interesting, less obvious alterna-
tives to reference chemotype 20 and pyrazinone. Experimental
data from compounds based on these additional scaffolds will
be reported in due course.
keeping primary activity, and improving ADME profile); in this
case, in comparison with reported compound in clinical phase,
2, selectivity over hERG was increased. In addition, this new
chemical series is chemically feasible and the optimal regio-
chemistry for the required substitution pattern is proposed,
therefore making the corresponding synthetic effort more
efficient.
With regard to this new chemical series, pyrazinones, further
details about improvements for the synthetic strategy, compound
optimization, etc. will be reported at a later date.
Conclusion
Novel strategy presented herein for chemotype hopping
provides an excellent platform to be routinely utilized in drug
discovery projects: (1) systematic evaluation of chemically
feasible chemotypes from different sources (corporate, external
suppliers, therapeutic area, target family related ligands, and
MedChem experience); (2) the potential of corporate libraries
and commercially available compounds is increased through the
virtual customization of scaffolds, close to existing structures
but one step beyond real compounds; (3) regiochemistry is
annotated, providing project-oriented functionalization that
delivers more rigorous comparisons for electrostatic maps and
shape; (4) knowledge about these growing vectors delivers
optimal construction and assessment of final compounds, once
chemotype is identified, based on pharmacophoric features
(ligand-based approach, as is the case) or based on structural
information for the receptor; and (5) considering different
fragmentation (e.g., Onion1) levels may lead to identify exo-
cyclic, not trivial, key driving ligand-receptor interactions from
which a new set of compounds can be built.
Experimental Section
Chemistry. General Procedure. Microwave-assisted reactions
were performed in a single-mode reactor, an Initiator Sixty
microwave reactor (Biotage) (description of the instrument can
be found at www.biotage.com), and in a multimode reactor, a
MicroSYNTH Labstation (Milestone, Inc.) (description of the
instrument can be found at www.milestonesci.com). Thin layer
chromatography (TLC) was carried out on silica gel 60 F254 plates
(Merck) using reagent grade solvents. Flash column chroma-
tography was performed on silica gel, particle size 60 Å, mesh
) 230-400 (Merck), under standard conditions. Automated flash
column chromatography was performed using ready-to-connect
cartridges from Merck, on irregular silica gel, particle size )
15-40 µm (normal phase disposable flash columns) on an SPOT
or FLASH system from Armen Instrument (www.armen-instru-
1
ment.com). H NMR spectra were recorded on a Bruker DPX-
400 and on a Bruker AV-500 spectrometer with standard pulse
sequences, operating at 400 and 500 MHz, respectively. Chemi-
cal shifts (δ) are reported in parts per million (ppm) downfield
from tetramethylsilane (TMS), which was used as internal
standard. HPLC-MS analysis was performed using an Agilent
1100 series system comprising a quaternary pump with degasser,
an autosampler, a column oven, and a diode array detector
(DAD) operating within a range between 200 and 450 nm to
determine the purity of all compounds reported in this paper.
Application of this new approach drove to excellent results:
(1) Not only were outstanding enrichment factors, numbers of
chemotypes identified, but also excellent quality in terms of
chemical diversity covered by those hits in the retrospective
analysis was achieved; (2) in the prospective analysis, the
challenge for an ongoing project, MCH-R1, was achieved
(finding new chemical series, novel IP in a crowded area,

2086 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7
Oyarzabal et al.
Purity for all assayed target compounds 34, 34, 37, and 2 is
>95%, as described in the Supporting Information. An ACE C-18
column, 30 × 4.6 mm i.d., 3.0 µm, from Advanced Chroma-
tography Technologies was used with a flow rate of either 2.0
or 1.5 mL/min and a generic gradient of either 80:10:10 AcONH4
0.05%/CH3CN/MeOH to 50:50 CH3CN/MeOH in 5.2 min to
100% CH3CN at 5.5 min and equilibrated to initial conditions
at 5.8 min until 7.0 min (method 1) or 80:10:10 AcONH4 0.05%/
CH3CN/MeOH to 50:50 CH3CN/MeOH in 6.5 min to 100%
CH3CN at 7.0 min and equilibrated to initial conditions at 7.5
min until 9.0 min (method 2). Flow from the column was split
to a single-quadrupole mass spectrometer from Waters for mass
spectra acquisition (either a Platform series II, or a ZQ or an
LCT Time of Flight) configured with an electrospray ionization
source (ESI) using MassLynx-Openlynx software.
2-Chloro-3-cyclohexyloxypyrazine (26). A solution of cyclo-
hexanol (5.4 mL, 0.05 mol) in 1,2-dimethoxyethane (15 mL) was
added dropwise to a mixture of NaH 60% (2.0 g, 0.05 mol) in
1,2-dimethoxyethane (10 mL), stirred at 0 °C. The mixture was
stirred for 10 min at 0 °C. Then a solution of 2,3-dichloropyrazine
(5.0 g, 0.034 mol) in 1,2-dimethoxyethane (25 mL) was added,
and the resultant reaction mixture was stirred and refluxed for 30
min. Water was added. This mixture was extracted with EtOAc.
The separated organic layer was dried (Na2SO4) and filtered, and
the solvent was evaporated, affording 9.0 g of crude compound 26
(quantitative yield; used in next reaction step without further
purification): MS m/z 213 [M + H]+.
1-[3-Methoxy-4-(2-pyrrolidin-1-ylethoxy)phenyl]-3-phenethyloxy-
1H-pyrazin-2-one (30). Phenethyl alcohol (32.5 µL, 0.27mmol) was
added to a mixture of NaH 60% (17 mg, 0.43mmol) in 1,2-
dimethoxyethane (0.5 mL), stirred at 0 °C. The mixture was stirred
for 30 min at room temperature. Then a solution of compound 29
(100 mg, 0.28 mmol) in 1,2-dimethoxyethane (1.5 mL) was added,
and the resultant reaction mixture was heated in a microwave oven
for 10 min at 80 °C. Water was added. This mixture was extracted
with CH2Cl2. The separated organic layer was dried (Na2SO4) and
filtered, and the solvent was evaporated. The residue was purified
by column chromatography over silica gel (CH2Cl2/MeOH 97.5:
2.5), affording 32 mg of compound 30: 1H NMR (400 MHz CDCl3)
δ 1.77-1.86 (m, 4 H), 2.59-2.69 (m, 4 H), 2.97 (t, J ) 6.43 Hz,
2 H), 3.18 (t, J ) 7.46 Hz, 2 H), 3.86 (s, 3 H), 4.19 (t, J ) 6.43
Hz, 2 H), 4.54 (t, J ) 7.67 Hz, 2 H), 6.83-6.90 (m, 3 H), 6.92-6.99
(m, 2 H), 7.19-7.25 (m, 1 H), 7.27-7.36 (m, 4 H); MS m/z 436
[M + H]+.
2-Methoxy-3-(3-phenylpropenyl)pyrazine (32). A mixture of
2-chloro-3-methoxypyrazine (300 mg, 21 mmol), trans-3-phenyl-
propen-1-ylboronic acid (336 mg, 21 mmol), and Pd(Ph3P)4 (243
mg, 0.21 mmol) in dioxane (3 mL) and NaHCO3 aqueous 1 M (1
mL) was stirred under microwave irradiation at 150 °C for 0.5 h.
The solvent was concentrated under reduced pressure. The residue
was purified by automated flash column chromatography over silica
gel (eluent: CH2Cl2 and CH2Cl2/MeOH(NH3); gradient from 100
to 97:3), affording 120 mg of desired compound 32 as a colorless
oil (25%): MS m/z 227 [M + H]+.
3-(3-Phenylpropyl)-1H-pyrazin-2-one (33). To a solution of 32
(120 mg, 0.5 mmol) in MeOH (20 mL) was added Pd/C 10%
(catalytic amount), and the mixture was stirred at room temperature
under hydrogen atmosphere for 4 h. The reaction was filtered, and
the filtrate was evaporated. The residue was dissolved in a mixture
of HCl 6 N (2 mL) and THF (1 mL) and stirred under microwave
irradiation at 140 °C for 1 h. The mixture was evaporated, affording
94 mg of desired compound 33 (100%): MS m/z 215 [M + H]+.
3-Cyclohexyloxy-1H-pyrazin-2-one (27). A mixture of com-
pound 26 (9.0 g, 0.034 mol) in NaOH (25 mL) and DMSO (25
mL) was heated for 90 min at 120 °C. Water was added. This
mixture was extracted with EtOAc. A saturated aqueous NH4Cl
solution was added to the separated organic phase. This mixture
was extracted with EtOAc. The combined organic layers were dried
(Na2SO4) and filtered, and the solvent was evaporated. The residue
was purified by column chromatography over silica gel (eluent:
CH2Cl2 and EtOAc). The product fractions were collected, and the
solvent was evaporated. The residue was treated with diisopropyl
ether, then filtered off, and dried, affording 2.4 g of compound 27
1-[3-Methoxy-4-(2-pyrrolidin-1-ylethoxy)phenyl]-3-(3-phenyl-
propyl)-1H-pyrazin-2-one (34). To a solution of 33 (100 mg, 0.5
mmol) in a mixture of dioxane/DMF 9:1 (2 mL) flushed with
N2 was added 1-[2-(4-bromo-2-methoxyphenoxy)ethyl]pyrroli-
dine (210 mg, 0.7 mmol) followed by CuI (94 mg, 0.5 mmol),
N,N′-dimethylethylenediamine (50 µL, 0.5 mmol) and K3PO4
(210 mg, 0.9 mmol). The reaction mixture was stirred under
microwave irradiation at 175 °C for 0.5 h. The solvent was
evaporated, and the residue was purified by automated flash
column chromatography over silica gel (eluent: CH2Cl2 and
CH2Cl2/MeOH(NH3); gradient from 100 to 95:5), affording 20
mg the desired compound. The compound was precipitated with
a mixture of DIPE/heptane (2:1), leading to 10 mg of desired
1
(36% two steps): H NMR (500 MHz CDCl3) δ 1.20-1.31 (m, 1
H), 1.35-1.47 (m, 2 H), 1.55-1.67 (m, 3 H), 1.79-1.87 (m, 2 H),
2.04-2.14 (m, 2 H), 5.03 (tt, J ) 10.11, 4.05 Hz, 1 H), 6.88 (d, J
) 4.34 Hz, 1 H), 6.95 (d, J ) 4.33 Hz, 1 H), 13.01 (br s, 1 H); MS
m/z 195 [M + H]+.
3-Cyclohexyloxy-1-[3-methoxy-4-(2-pyrrolidin-1-ylethoxy)phe-
nyl]-1H-pyrazin-2-one (28). A mixture of compound 27 (2.3 g,
0.012 mol), 1-[2-(4-bromo-2-methoxyphenoxy)ethyl]pyrrolidine61
(4.5 g, 0.015 mol), CuI (2.3 g, 0.012 mol), N,N′-dimethylethyl-
enediamine (2.55 mL, 0.024 mol), and K3PO4 (5.1 g, 0.024 mol)
in dioxane/DMF 5:1 (35 mL) was heated for 15 min at 180 °C in
a microwave oven. CH2Cl2 was added. The solid was filtered off
through a Celite pad, and to the filtrate was added a 32% NH3
solution. The organic layer was separated, washed with brine, dried
(Na2SO4), and filtered, and the solvent was evaporated. The residue
was purified by column chromatography over silica gel. The product
fractions were collected, and the solvent was evaporated, affording
4.2 g of compound 28 (85%), used in next reaction step without
further purification: MS m/z 414 [M + H]+.
3-Chloro-1-[3-methoxy-4-(2-pyrrolidin-1-ylethoxy)phenyl]-1H-
pyrazin-2-one (29). POCl3 (5.2 mL, 0.0564 mol) was added to a
mixture of compound 28 (4.67, 0.0112 mol) in 1,2-dichloroethane
(45 mL), and the resultant reaction mixture was heated for 30 min
at 150 °C in a microwave oven. The reaction was poured in an
ice-water bath and then basified with a saturated aqueous Na2CO3
solution. This mixture was extracted with CH2Cl2, then dried
(Na2SO4), and filtered, and the solvent was evaporated. The residue
was purified by column chromatography over silica gel (eluent:
CH2Cl2/MeOH(NH3) 95:5). The product fractions were collected,
and the solvent was evaporated, affording 3 g of compound 29
(77%), used in next reaction step without further purification: MS
m/z 350 [M + H]+.
1
compound 34 as a foam (5%): H NMR (400 MHz, CDCl3) δ
1.77-1.85 (m, 4 H), 2.05-2.15 (m, 2 H), 2.59-2.69 (m, 4 H),
2.75 (t, J ) 7.46 Hz, 2 H), 2.92 (t, J ) 7.67 Hz, 2 H), 2.97 (t,
J ) 6.43 Hz, 2 H), 3.87 (s, 3 H), 4.19 (t, J ) 6.43 Hz, 2 H),
6.86 (dd, J ) 8.50, 2.28 Hz, 1 H), 6.92 (d, J ) 2.49 Hz, 1 H),
6.97 (d, J ) 8.50 Hz, 1 H), 7.07 (d, J ) 4.56 Hz, 1 H), 7.13-7.19
(m, 1 H), 7.20-7.31 (m, 5 H); MS m/z 434 [M + H]+.
2-Chloro-3-phenethylaminopyrazine (35). A mixture of phen-
ethylamine (3.5 mL, 0.028mol), 2,3-dichloropyrazine (5 g,
0.034mol), and 1,8-diazabicyclo(5.4.0)undec-7-ene (DBU) (5
mL, 0.034 mol) in CH3CN (60 mL) was heated in a microwave
oven for 20 min at 180 °C. The solvent was evaporated. The
residue was purified by column chromatography over silica gel
(eluent: CH2Cl2). The product fractions were collected, and the
solvent was evaporated, affording 5.2 g of compound 35
(quantitative yield; used in next reaction step, without further
purification): MS m/z 234 [M + H]+.
3-Phenethylamino-1H-pyrazin-2-one (36). A mixture of com-
pound 35 (5.2 g, 0.028 mol) in NaOH (25 mL) and DMSO (25
mL) was heated in a microwave oven for 30 min at 150 °C. Water
was added. This mixture was extracted with EtOAc. The organic
layer was dried (Na2SO4) and filtered, and the solvent was
evaporated. The residue was purified by column chromatography

Melanin Concentrating Hormone Antagonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7 2087
(10) (a) Lewell, X. Q.; Judd, D. B.; Watson, S. P.; Hann, M. M.
RECAPsretrosynthetic combinatorial analysis procedure: a powerful
new technique for identifying privileged molecular fragments with
useful applications in combinatorial chemistry. J. Chem. Inf. Comput.
Sci. 1998, 38, 511–522. (b) Degen, J.; Wegscheid-Gerlach, C.; Zaliani,
A.; Rarey, M. On the art of compiling and using “drug-like” chemical
fragment spaces. ChemMedChem 2008, 3, 1503–1507.
(11) Advanced Chemistry Development Inc., 110 Yonge St., 14th floor,
Toronto, ON, Canada, ACD/PhysChem Batch 10.00; available from
ACD Inc. at http://www.acdlabs.com.
(12) Needleman, P. Method of using cyclooxygenase-2 inhibitors in the
treatment and prevention of dementia. Patent WO 9843648 A1; 26
pp.
(13) Zhou, P.; Kelly, M. G. Preparation of 2-(1-phenylsulfonylindol- or
-indazol-4-yloxy)ethanamines as 5-HT6 receptor ligands. Patent WO
2002 059088 A1; 58 pp.
(14) Ohkawa, S.; Fujii, N.; Kato, K.; Miyamoto, M. Condensed heterocyclic
compounds (benzo[b]cyclopenta[e][1,4]diazepinones and analogs),
their production, and use as GnRH antagonists. Patent WO 9529900
A1; 220 pp.
(15) Duggan, M. E.; Lindsley, C. W.; Wisnoski, D. D. Preparation of
benzamide modulators of metabotropic glutamate receptors. Patent WO
2004087048 A2; 65 pp.
(16) Bromidge, S. M.; Lovell, P. J.; Goodacre, C. Preparation of cyclic
urea derivatives with 5-HT2c receptor activity. Patent WO 2003057220
A1; 36 pp.
(17) Engels, M. F. M.; Gibbs, A. C.; Jaeger, E. P.; Verbinnen, D.; Lobanov,
V. S.; Agrafiotis, D. K. A cluster-based strategy for assessing the
overlap between large chemical libraries and its application to a recent
acquisition. J. Chem. Inf. Model. 2006, 46, 2651–2660.
over silica gel (eluent: CH2Cl2/EtOAc 4:1 and 1:1). The product
fractions were collected, and the solvent was evaporated, affording
4.4 g of compound 36 (73% two steps): 1H NMR (500 MHz CDCl3)
δ 2.95 (t, J ) 7.08 Hz, 2 H), 3.69 (q, J ) 6.94 Hz, 2 H), 6.10 (t,
J ) 5.49 Hz, 1 H), 6.50 (d, J ) 4.33 Hz, 1 H), 6.96 (d, J ) 4.33
Hz, 1 H), 7.19-7.27 (m, 3 H), 7.28-7.36 (m, 2 H), 11.73 (br s, 1
H); MS m/z 216 [M + H]+.
1-[3-Methoxy-4-(2-pyrrolidin-1-ylethoxy)phenyl]-3-phenethy-
lamino-1H-pyrazin-2-one (37). A mixture of compound 36 (100
mg, 0.46 mmol) 1-[2-(4-bromo-2-methoxyphenoxy)ethyl]pyrroli-
dine (160 mg, 0.55 mmol), CuI (80 mg, 0.46 mmol), N,N′-
dimethylethylenediamine (98 µL, 0.92 mmol), and K3PO4 (190 mg,
0.92 mmol) in dioxane/DMF (3.5 mL; 9:1) was heated for 20 min
at 175 °C. CH2Cl2 was added. The whole was filtered through a
Celite pad, and the filtrate was treated with an aqueous NH4Cl
solution. The organic layer was separated, dried (Na2SO4), and
filtered, and the solvent was evaporated. The residue was purified
by column chromatography (eluent: CH2Cl2 and CH2Cl2/CH3OH
97.5:2.5, 95:5, and 90:10). The product fractions were collected,
and the solvent was evaporated, affording 130 mg of compound
37 (65%): 1H NMR (400 MHz CDCl3) δ 1.78-1.89 (m, 4 H), 2.69
(br s, 4 H), 2.98 (t, J ) 7.05 Hz, 2 H), 3.00 (t, J ) 6.22 Hz, 2 H),
3.71 (q, J ) 7.05 Hz, 2 H), 3.86 (s, 3 H), 4.21 (t, J ) 6.43 Hz, 2
H), 6.33 (t, J ) 5.81 Hz, 1 H), 6.55 (d, J ) 4.77 Hz, 1 H),
6.87-6.93 (m, 3 H), 6.97 (d, J ) 8.29 Hz, 1 H), 7.18-7.35 (m, 5
H); MS m/z 435 [M + H]+.
(18) Prous Science S. A. U., Provenza 388, 08025 Barcelona, Spain;
available from http://integrity.prous.com/.
Acknowledgment. We thank the members of the screening
group led by Dr. Danielle Peeters for performing the reported
hERG binding assays. We thank also the members of the
Analytical Department in Toledo for analytical and spectral
determinations, as well as Alcira Del Cerro for synthetic
assistance. We also thank Dr. Jan Linders and Dr. Gary
Tresadern for helpful discussions and Dr. Guido Kirsten,
Chemical Computing Group (CCG), for collaboration in the
development of the fragmentation script.
(19) OpenEye Scientific Software Inc., 9 Bisbee Ct., Suite D, Santa Fe,
NM 87508; available from http://www.eyesopen.com.
(20) Bostro¨m, J. Reproducing the conformations of protein-bound ligands:
a critical evaluation of several popular conformational searching tools.
J. Comput.-Aided. Mol. Des. 2001, 15, 1137–1152.
(21) Kirchmair, J.; Wolber, G.; Laggner, C.; Langer, T. Comparative
performance assessment of the conformational model generators omega
and catalyst: a large-scale survey on the retrieval of protein-bound
ligand conformations. J. Chem. Inf. Model. 2006, 46, 1848–1861.
(22) Grant, J. A.; Gallardo, M. A.; Pickup, B. T. A fast method of molecular
shape comparison. A simple application of Gaussian description of
molecular shape. J. Comput. Chem. 1996, 17, 1653–1666.
(23) Bostro¨m, J.; Greenwood, J. R.; Gottfries, J. Assessing the performance
of OMEGA with respect to retrieving bioactive conformations. J. Mol.
Graphics Modell. 2003, 21, 449–462.
(24) (a) Nicholls, A.; MacCuish, N. E.; MacCuish, J. D. Variable selection
and model validation of 2-D and 3-D molecular descriptors. J. Com-
put.-Aided Mol. Des. 2004, 18, 451–474. (b) Nicholls, A.; Grant, A. J.
Molecular shape and electrostatics in the encoding of relevant chemical
information. J. Comput.-Aided Mol. Des. 2005, 19 (9-10), 661–686.
(25) Jennings, A.; Tennant, M. Selection of molecules based on shape and
electrostatic similarity: proof of concept of “electroforms”. J. Chem.
Inf. Model. 2007, 47, 1829–1838.
(26) Saito, Y.; Nothacker, H.-P.; Wang, Z.; Lin, S. H. S.; Frances, L.;
Civelli, O. Molecular characterization of the melanin-concentrating
hormone receptor. Nature 1999, 400, 265–269.
(27) Chambers, J.; Ames, R. S.; Bergsma, D.; Muir, A.; Fitzgerald, L. R.;
Hervieu, G.; Dytko, G. M.; Foley, J. J.; Martins, J.; Liu, W.-S.; Park,
J.; Ellis, C.; Ganguly, S.; Konchar, S.; Cluderay, J.; Leslie, R.; Wilson,
S.; Sarau, H. M. Melanin-concentrating hormone is the cognate ligand
for the orphan G-protein-coupled receptor SLC-1. Nature 1999, 400,
261–265.
Supporting Information Available: Purity and retention times
for reported synthetic intermediates 26-29, 32, 33, 35, and 36 as
well as HPLC traces for assayed target compounds 30, 34, 37, and
2; an additional figure, S1, is also included. This material is available
free of charge via the Internet at http://pubs.acs.org.
References
(1) Brown, N.; Jacoby, E. On scaffolds and hopping in medicinal
chemistry. Mini ReV. Med. Chem. 2006, 6, 1217–1229.
(2) Schneider, G.; Neidhart, W.; Giller, T.; Schmid, G. “Scaffold-hopping”
by topological pharmacophore search: a contribution to virtual
screening. Angew. Chem., Int. Ed. 1999, 38, 2894–2896.
(3) (a) Bohl, M.; Dunbar, J.; Gifford, E. M.; Heritage, T.; Wild, D. J.;
Willett, P.; Wilton, D. J. Scaffold searching: automated identification
of similar ring systems for the design of combinatorial libraries. Quant.
Struct.-Act. Relat. 2002, 21, 590–597. (b) Lauri, G.; Bartlett, P. A.
CAVEAT: a program to facilitate the design of organic molecules.
J. Comput.-Aided Mol. Des. 1994, 8, 51–66.
(28) Sailer, A. W.; Sano, H.; Zeng, Z.; McDonald, T. P.; Pan, J.; Pong,
S.-S.; Feighner, S. D.; Tan, C. P.; Fukami, T.; Iwaasa, H.; Hreniuk,
D. L.; Morin, N. R.; Sadowski, S. J.; Ito, M.; Ito, M.; Bansal, A.; Ky,
B.; Figueroa, D. J.; Jiang, Q.; Austin, C. P.; MacNeil, D. J.; Ishihara,
A.; Ihara, M.; Kanatani, A.; Van der Ploeg, L. H. T.; Howard, A. D.;
Liu, Q. Identification and characterization of a second melanin-
concentrating hormone receptor, MCH-R2. Proc. Natl. Acad. Sci.
U.S.A. 2001, 98, 7564–7569.
(29) Mendez-Andino, J. L.; Wos, J. A. MCH-R1 antagonists: what is
keeping most research programs away from the clinic? Drug DiscoV.
Today 2007, 12, 972–978.
(30) Schwartz, M. W.; Woods, S. C.; Porte, D., Jr.; Seeley, R. J.; Baskin,
D. G. Central nervous system control of food intake. Nature 2000,
404, 661–671.
(4) Bergmann, R.; Linusson, A.; Zamora, I. SHOP: Scaffold HOPping
by GRID-based similarity searches. J. Med. Chem. 2007, 50, 2708–
2717.
(5) Jenkins, J. L.; Glick, M.; Davies, J. W. A 3D similarity method for
scaffold hopping from known drugs or natural ligands to new
chemotypes. J. Med. Chem. 2004, 47, 6144–6159.
(6) Zhang, Q.; Muegge, I. Scaffold hopping through virtual screening using
2D and 3D similarity descriptors: ranking, voting, and consensus
scoring. J. Med. Chem. 2006, 49, 1536–1548.
(7) Schneider, G.; Schneider, P.; Renner, S. Scaffold-hopping: how far
can you jump? QSAR Comb. Sci. 2006, 12, 1162–1171.
(8) Scitegic. Pipeline Pilot, 9665 Chesapeake Dr., Suite 401, San Diego,
CA 92123-1365; available from SciTegic Inc. at http://www.scitegic.
com.
(31) Shearman, L. P.; Camacho, R. E.; Sloan Stribling, D.; Zhou, D.;
Bednarek, M. A.; Hreniuk, D. L.; Feighner, S. D.; Tan, C. P.; Howard,
A. D.; Van der Ploeg, L. H. T.; MacIntyre, D. E.; Hickey, G. J.; Strack,
A. M. Chronic MCH-1 receptor modulation alters appetite, body
weight and adiposity in rats. Eur. J. Pharmacol. 2003, 475, 37–47.
(9) Chemical Computing Group Inc. Molecular Operating Environment
(MOE), 1010 Sherbrooke St. W, Suite 910, Montreal, QC, Canada;
available from Chemical Computing Group Inc. at http://www.
chemcomp.com or http://svl.chemcomp.com.

2088 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7
Oyarzabal et al.
(32) Borowsky, B.; Durkin, M. M.; Ogozalek, K.; Marzabadi, M. R.;
DeLeon, J.; Heurich, R.; Lichtblau, H.; Shaposhnik, Z.; Daniewska,
I.; Blackburn, T. P.; Branchek, T. A.; Gerald, C.; Vaysse, P. J.; Forray,
C. Antipressant, anxiolytic and anorectic effects of a melanin-
concentrating hormone-1 receptor antagonist. Nat. Med. 2002, 8, 825–
830.
(49) Meyers, K. M.; Mendez-Andino, J. L.; Colson, A.-O.; Warshakoon,
N. C.; Wos, J. A.; Mitchell, M. C.; Hodge, K. M.; Howard, J. M.;
Ackley, D. C.; Holbert, J. K.; Mittelstadt, S. W.; Dowty, M. E.;
Obringer, C. M.; Reizes, O.; Hu, X. E. Aminoethyl tetrahydronaph-
thalene ketopiperazine MCH-R1 antagonists-increasing selectivity over
hERG. Bioorg. Med. Chem. Lett. 2007, 17, 819–822.
(33) (a) Chaki, S.; Funakoshi, T.; Hirota-Okuno, S.; Nishiguchi, M.;
Shimazaki, T.; Iijima, M.; Grottick, A. J.; Kanuma, K.; Omodera, K.;
Sekiguchi, Y.; Okuyama, S.; Tran, T.-A.; Semple, G.; Thomsen, W.
Anxiolytic- and antidepressant-like profile of ATC0065 and ATC0175:
nonpeptidic and orally active melanin-concentrating hormone receptor
1 antagonists. J. Pharmacol. Exp. Ther. 2005, 313, 831–839. (b) Chaki,
S.; Yamaguchi, J.-I.; Yamada, H.; Kanuma, K.; Sekiguchi, Y. Melanin-
concentrating hormone receptor 1 antagonists for the treatment of
depression and anxiety. Drug DeV. Res. 2005, 65, 278–290.
(34) (a) Chen, X.; Chen, X.; Fan, P.; Jaen, J.; Li, L.; Mihalic, J. T. Patent
WO 04/043958 A1. (b) Andersen, D.; Storz, T.; Liu, P.; Wang, X.;
Li, L.; Fan, P.; Chen, X.; Allgeier, A.; Burgos, A.; Tedrow, J.; Baum,
J.; Chen, Y.; Crockett, R.; Huang, L.; Syed, R.; Larsen, R. D.;
Martinelli, M. Stereoselective synthesis of a MCH-R1 antagonist. J.
Org. Chem. 2007, 72, 9648–9655.
(50) Guo, T.; Hunter, R. C.; Zhang, R.; Greenlee, W. J. Microwave assisted
synthesis of isothiazolo-, thiazolo-, imidazo- and pyrimido-pyrimidi-
nones as novel MCHR1 antagonists. Tetrahedron Lett. 2007, 48, 613–
615.
(51) (a) Giordanetto, F.; Inghardt, T. Thienoheterocycles, processes for
preparing them, pharmaceutical compositions containing them, and
their use in the treatment of obesity, psychiatric and neurological
disorders. Patent WO07011286 A1; 34 pp. (b) Giordanetto, F.;
Inghardt, T.; Nordberg, P. Preparation of thieno[3,2-d]pyrimidin-4(3H)-
one derivatives as MCH agonists. Patent WO07011284 A1; 44 pp.
(c) Giordanetto, F.; Inghardt, T. Benzimidazolyl-substituted thieno-
heterocycles, processes for preparing them, pharmaceutical composi-
tions containing them, and their use in the treatment of obesity,
psychiatric and neurological disorders. Patent WO07011285 A1; 30
pp.
(35) (a) Hertzog, D. L.; Al-Barazanji, K. A.; Bigham, E. C.; Bishop, M. J.;
Britt, C. S.; Carlton, D. L.; Cooper, J. P.; Daniels, A. J.; Garrido,
D. M.; Goetz, A. S.; Grizzle, M. K.; Guo, Y. C.; Handlon, A. L.;
Ignar, D. M.; Morgan, R. O.; Peat, A. J.; Tavares, F. X.; Zhou, H.
The discovery and optimization of pyrimidinone-containing MCH-
R1 antagonists. Bioorg. Med. Chem. Lett. 2006, 16, 4723–4727. (b)
Hertzog, D. L.; Wittig, D. R. Biphenyl amides and isosteres as MCH
R1 antagonists. Curr. Top. Med. Chem. 2007, 7, 1455–1470.
(36) Rokosz, L. L. Discovery and development of melanin-concentrating
hormone receptor 1 antagonists for the treatment of obesity. Expert
Opin. Drug DiscoV. 2007, 2, 1301–1327.
(37) Takekawa, S.; Aami, A.; Ishihara, Y.; Terauchi, J.; Kato, K.;
Shimomura, Y.; Mori, M.; Murakoshi, H.; Kato, K.; Suzuki, N.;
Nishimura, O.; Fujino, M. T-226296: a novel, orally active and
selective melanin-concentrating hormone receptor antagonist. Eur.
J. Pharmacol. 2002, 438, 129–135.
(38) Armstrong, S. A.; Hamprecht, D. W.; Jones, M.; Witty, D. R.; Al-
Baranznji, K. A.; Tadayyon, M. Lactam derivatives as antagonists for
human 11CBY receptors. Patent WO03033480; 87 pp.
(39) Ammenn, J.; Gillig, J. R.; Heinz, L. J.; Hipskind, P. A.; Kinnick, M. D.;
Lay, Y.-S.; Morin, J. M.; Nixon, J. A.; Ott, C.; Savin, K. A.; Schotten,
T.; Slieker, L. J.; Snyder, N. J.; Robertson, M. A. Multicyclic
compounds for use as melatonin concentrating hormone antagonists
in the treatment of obesity and diabetes. Patent WO03097047; 592
pp.
(40) Lehmann-Lintz, T.; Stenkamp, D.; Lenter, M.; Wieland, H.-A.; Rudolf,
K.; Mu¨ller, S. G.; Lotz, R. R. H.; Arndt, K. and Lustenberger, F. Novel
carboxamide compounds having an MCH-antagonistic effect medica-
ments containing said compounds, and methods for the production
thereof. Patent WO04024702 A1; 344 pp.
(41) Moriya, M.; Sakakmoto, T.; Kishino, H.; Kanatani, A. Preparation of
2-aminobenzoimidazole derivatives as antagonists of melanin-
concentrating hormone receptor. Patent WO04031177 A1; 93 pp.
(42) Schwink, L.; Stengelin, S.; Gossel, M.; Boehme, T.; Hessler, G.; Rosse,
G.; Walser, A. Preparation of 1,3-dihydro-1,3-diphenyl-2Himidazol-
2-ones and related compounds as MCH receptor modulators for the
treatment of obesity. Patent WO04011438 A1; 113 pp.
(43) Otake, N.; Haga, Y.; Naya, A.; Mizutani, S.; Kanatani, A. Preparation
of pyridone and related compounds as melanin concentrating hormone
receptor antagonists. Patent WO05085200 A1.
(44) Goodfellow, V.; Dyck, B. P.; Zhao, L.; Tamiya, J.; Grey, J.; Wade,
W.; Rowbottom, M.; Zhang, M.; Vickers, T. D. Preparation of
aminopyrrolidinyl pyridines as MCH receptor antagonists. Patent
WO05103039 A1; 64 pp.
(52) (a) Balavoine, F.; Nicolaie, E.; Sartori, E. Preparation of hydantoin
derivatives and their use as MCHR-1 antagonists. Patent WO07012661
A1; 92 pp. Balavoine, F.; Nicolaie, E.; Sartori, E. Preparation of
5-benzylideneimidazolidine-2,4-dione derivatives and their use as
MCHR-1 antagonists. Patent WO07012670 A1. (c) Balavoine, F.;
Malabre, P.; Alleaume, T.; Rey, A.; Cherfils, V.; Jeanneton, O.;
Seigneurin-Veninb, S.; Revah, F. Design and synthesis of novel
hydantoin-containing melanin-concentrating hormone receptor an-
tagonists. Bioorg. Med. Chem. Lett. 2007, 17, 3754–3759.
(53) (a) Scheneider, J.; Hauser, R.; Andreas, J. O.; Linz, K.; Jahnel, U.
Differential effects of human ether-a-go-go-related gene (hERG)
blocking agents on QT duration variability in conscious dogs. Eur.
J. Pharmacol. 2005, 512, 53–60. (b) Shah, R. R. The significance of
QT interval in drug development. Br. J. Clin. Pharmacol. 2002, 54,
188–202. (c) Haverkamp, W.; Breithardt, G.; Camm, A. J.; Janse,
M. J.; Rosen, M. R.; Antzelevitch, C.; Escande, D.; Franz, M.; Malik,
M.; Moss, A.; Shah, R. The potential for QT prolongation and pro-
arrhythmia by non-anti-arrhythmia drugs: clinical and regulatory
implications: report on a policy conference of the European Society
of Cardiology. CardioVasc. Res. 2000, 47, 219–232.
(54) Ca2+ mobilization experiment for the MCH-1 receptor, FLIPR assay:
24 h before the experiment, MCH-1 receptor-expressing CHO cells
are seeded in 20 µL (5000 cells per well) into 384-well black-wall,
clear-bottom microtiter plates (Costar). On the day of the experiment,
20 µL per well calcium assay kit containing 10 mM probenicide
(Molecular Devices) is added. Cells are loaded for 90 min at 37 °C
and 5% CO2 in a cell culture incubator. After loading, 20 µL of serial
dilutions of the test compound are added and cells are further incubated
for 20 min at room temperature in the dark. After 20 min, 20 µL of
a submaximal MCH concentration is added and changes in intracellular
calcium are recorded directly in a FLIPR III apparatus (Molecular
Devices).
(55) Birchall, K.; Gillet, V. J.; Harper, G.; Pickett, S. D. Evolving
interpretable structure-activity relationship models. 2. Using multiob-
jective optimization to derive multiple models. J. Chem. Inf. Model.
2008, 48, 1558–1570.
(56) Brooijmans, N.; Kuntz, I. D. Molecular recognition and docking
algorithms. Annu. ReV. Biophys. Biomol. Struct. 2003, 32, 335–373.
(57) (a) Bender, A.; Glen, R. C. A discussion of measures of enrichment
in virtual screening: comparing the information content of descriptors
with increasing levels of sophistication. J. Chem. Inf. Model. 2005,
45, 1369–1375. (b) Bender, A.; Mussa, H. Y.; Reiling, S.; Glen, R. C.
Similarity searching of chemical databases using atom environment
descriptors (MOLPRINT 2D): evaluation of performance. J. Chem.
Inf. Model. 2004, 44, 1708–1718. (c) Glen, R. C.; Bender, A.; Arnby,
C. H.; Carlsson, L.; Boyer, S. and; Smith, J. Circular fingerprints:
flexible molecular descriptors with applications from physical chem-
istry to ADME. IDrugs 2006, 49, 199–204.
(45) Amegadzie, A. K.; Beck, J. P.; Gardinier, K. M.; Hembre, E. J.; Ruble,
J. C.; Savin, K. A.; Wakefield, B. D. Preparation of thiazolopyridinones
as MCH receptor antagonists for treating and preventing symptoms
assocd. With obesity and related diseases. Patent WO06066174 A1;
154 pp.
(58) For a review of the application of microwave chemistry to medicinal
chemistry, see: (a) Alcazar, J.; Diels, G.; Schoentjes, B. Microwave
assisted medicinal chemistry. Mini-ReV. Med. Chem. 2007, 7, 345–
369. (b) Alcazar, J.; Diels, G.; Schoentjes, B. Applications of the
combination of microwave and parallel synthesis in medicinal
chemistry. Comb. Chem. High Throughput Screen. 2007, 10, 918–
932.
(46) Ishihara, Y.; Suzuhi, N.; Takekawa, S. Melanin-concentrating hormone
anatagonists. Patent JP2001226269 A; 88 pp.
(47) Dyck, B.; Zhao, L.; Tamiya, J.; Pontillo, J.; Hudson, S.; Ching, B.;
Heise, C. E.; Wen, J.; Norton, C.; Madan, A.; Schwarz, D.; Wade,
W.; Goodfellow, V. S. Substituted chromones and quinolones as potent
melanin-concentrating hormone receptor 1 antagonists. Bioorg. Med.
Chem. Lett. 2006, 16, 4237–4242.
(48) Dyck, B.; Markison, S.; Zhao, L.; Tamiya, J.; Grey, J.; Rowbottom,
M. W.; Zhang, M.; Vickers, T.; Sorensen, K.; Norton, C.; Wen, J.;
Heise, C. E.; Saunders, J.; Conlon, P.; Madan, A.; Schwarz, D.;
Goodfellow, V. S. A thienopyridazinone-based melanin-concentrating
hormone receptor 1 antagonist with potent in vivo anorectic properties.
J. Med. Chem. 2006, 49, 3752–3756.
(59) [3H]Astemizole binding assays in HEK293 cells expressing HERG K+
channels were performed following the procedure disclosed in the
literature: Chiu, P. J. S.; Marcoe, K. F.; Bounds, S. E.; Lin, C.-H.; Feng,
J.-J.; Lin, A.; Cheng, F.-C.; Crumb, W. J.; Mitchell, R. Validation of a
[3H]astemzole binding assay in HEK293 cells expressing HERG K+
channels J. Pharmacol. Sci. 2004, 95, 311-319.

Melanin Concentrating Hormone Antagonists
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 7 2089
(60) Andres-Gil, J. I.; Alcazar-Vaca, M. J.; Alvarez-Escobar, R. M.;
Oyarzabal-Santamarina, J.; Dautzenberg, F. M.; Macritchie, J.;
Simpson, D. and Martinez Gonzalez, S.Preparation of aryl and
heteroaryl substituted pyrazinone derivatives for use in melanin-
concentrating hormone mediated diseases. Patent WO07071646 A1;
139 pp.
Lin, H.-S.; Lynch, M. P.; Mccowan, J. R.; Palkowitz, A. D.; Richett,
M. E.; Sall, D. J.; Smith, G. F.; Takeuchi, K.; Tinseley, J. M. and
Zhang, M.Preparation of [(pyrrolidinoalcoxy)phenyl+benzothiphenes
and analogues as thrombin inhibitors. Patent WO9725033 A1; 43
pp.
(61) Bastian, J. L.; Chirgadze, N. Y.; Denney, M. L.; Foglesong, R. J.;
Harper, R. W.; Johnson, M. G.; Klimkowsky, V. J.; Kohn, T. J.;
JM8016199