1,2,4-Triazolyl 5-Azaspiro[2.4]heptanes: Lead Identification and Early Lead Optimization of a New Series of Potent and Selective Dopamine D3 Receptor Antagonists

Article abstract of DOI:10.1021/acs.jmedchem.6b00972

A novel series of 1,2,4-triazolyl 5-azaspiro[2.4]heptanes with high affinity and selectivity at the dopamine (DA) D3 receptor (D3R) is described. Some of these compounds also have high selectivity over the hERG channel and were characterized with respect to their pharmacokinetic properties both in vitro and in vivo during lead identification and early lead optimization phases. A few derivatives with overall favorable developability characteristics were selected for further late lead optimization studies.

Full text of DOI:10.1021/acs.jmedchem.6b00972

Subscriber access provided by United Arab Emirates University | Libraries Deanship
Article
1
,2,4-Triazolyl 5-Azaspiro[2.4]heptanes: Lead Identification
and Early Lead Optimization of a New Series of Potent
and Selective Dopamine D3 Receptor Antagonists
Fabrizio Micheli, Alessia Bacchi, Simone Braggio, Laura Castelletti, Palmina Cavallini, Paolo Cavanni,
Susanna Cremonesi, Michele Dal Cin, Aldo Feriani, Sylvie Gehanne, Mahmud Kajbaf, Luciano Marchió,
Selena Nola, Beatrice Oliosi, Annalisa Pellacani, Elisabetta Perdona', Anna Sava, Teresa Semeraro,
Luca Tarsi, Silvia Tomelleri, Andrea Wong, Filippo Visentini, Laura Zonzini, and Christian A. Heidbreder
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00972 • Publication Date (Web): 26 Aug 2016
Downloaded from http://pubs.acs.org on August 28, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 105
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
,2,4ꢀTriazolyl 5ꢀAzaspiro[2.4]heptanes: Lead
Identification and Early Lead Optimization of a New
Series of Potent and Selective Dopamine D3 Receptor
Antagonists.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
,
a
#
a
a
a
Fabrizio Micheli* , Alessia Bacchi , Simone Braggio , Laura Castelletti , Palmina Cavallini , Paolo
a
a
a
a
a
a
Cavanni , Susanna Cremonesi , Michele Dal Cin , Aldo Feriani , Sylvie Gehanne , Mahmud Kajbaf ,
#
a
a
a
a
Luciano Marchió , Selena Nola , Beatrice Oliosi , Annalisa Pellacani , Elisabetta Perdona’ , Anna
a
a
a
a
a
a
Sava , Teresa Semeraro , Luca Tarsi , Silvia Tomelleri , Andrea Wong , Filippo Visentini , Laura
a
b
Zonzini , and Christian Heidbreder .
a
b
Aptuit s.r.l., Via Fleming 4, 37135 Verona (Italy), Indivior Inc, The Fairfax Building, 10710
#
Midlothian Turnpike, Suite 430, Richmond VA 23235, USA, Dipartimento di Chimica, Universita’ di
Parma, Viale delle Scienze, 17A ꢀ Campus, Iꢀ43124 Parma, Italy
ABSTRACT. A novel series of 1,2,4ꢀTriazolyl 5ꢀAzaspiro[2.4]heptanes with high affinity and
selectivity at the Dopamine (DA) D3 receptor (D3R) is described. Some of these compounds also have
high selectivity over the hERG channel and were characterized with respect to their pharmacokinetic
properties both in vitro and in vivo during lead identification and early lead optimization phases. A few
derivatives with overall favorable developability characteristics were selected for further late lead
optimization studies.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 105
Introduction
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Longꢀlasting adaptations in the brains of individuals addicted to drugs support the concept of addiction
as a chronically relapsing disorder that is characterized by the compulsion to seek and take the drug, the
loss of control in limiting drug intake, and the emergence of a negative emotional state when access to
the drug is withdrawn. The inherent complexity of these neuroadaptations makes the therapeutic
management of substance use disorder (SUD) a major challenge for the discovery and development of
safe and efficacious pharmacotherapeutic strategies ensuring satisfactory patient compliance and
promoting drug abstinence and recovery.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Half a century of research has pointed towards activation of the mesolimbic dopamine (DA) system as a
common denominator in response to virtually all drugs of abuse that are voluntarily selfꢀadministered by
laboratory animals and humans. It is thus quite natural that DA receptors were considered as reasonable
targets for the development of SUD medications. Of particular interest is the restricted highꢀdensity
expression of the DA D3 receptor (D3R) subtype into key elements of the mesolimbic DA system such
as the ventral striatum, midbrain, and pallidum, which have all been associated with behaviors
1
ꢀ3
controlled by the presentation of drugꢀassociated cues . The strategic localization of the D3R together
with changes in its expression following acute, subꢀchronic or chronic exposure to drugs of abuse have
triggered research efforts leading to the discovery, synthesis and characterization of highly potent and
1
ꢀ4
selective D3R antagonists . Nonclinical evidence clearly supports the contention that selective D3R
antagonists show promise for the treatment of SUD as reflected by their potential to (1) regulate the
motivation to selfꢀadminister drugs under schedules of reinforcement that require an increase in work
demand, and (2) disrupt the responsiveness to drugꢀassociated stimuli that play a key role in
reinstatement of drugꢀseeking behavior triggered by reꢀexposure to the drug itself, reꢀexposure to
1
ꢀ3
environmental cues that had been previously associated with drugꢀtaking behavior, or stress.
ACS Paragon Plus Environment

Page 3 of 105
Journal of Medicinal Chemistry
5
,6
GlaxoSmithKline (GSK) reported compounds with high in vitro and in vivo selectivity at the D3R,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
which can attenuate or block the effects of drugs of abuse in a wide range of animal models of addiction.
7
Additional compounds were also disclosed by the pharmaceutical industry and academic centers .
The scope of the present work was to identify potential replacements of the “amino portion” of the
8
,9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
above mentioned scaffolds. Two attempts of “amino portion” substitution were recently disclosed
through a mixed computational and medicinal chemistry strategy that identified both morpholines and
octahydropyrrolo[2,3ꢀb]pyrroles.
Using a similar approach, the present work describes a new series of 5ꢀazaspiro[2.4]heptanes with a
detailed Structure Activity Relationship (SAR) in lead identification and early lead optimization phases;
a detailed in vitro and in vivo pharmacokinetic (PK) strategy to selected compounds for the subsequent
late lead optimization phase is also presented.
Results and discussion
Since the discovery of the D3R in 1990, both industrial and academic researchers have been trying to
4
,7
identify novel chemical entities (NCEs) that selectively modulate the D3R . Among the most
10
characterized derivatives, both in vitro and in vivo, SB-277011, 1 represents one of the prototypical
5
selective D3R antagonists. In 2010, GSK presented a full characterization of GSK598809A, 2 which
1
1
was progressed to clinical study in humans . Examples of structurally unrelated templates are
1
2a
13
PG01037, 3 the tranylcypromine substituted cisꢀhydroxycyclobutylnaphthamides (CJ-1882, 4) , and
1
4
9
YQA14 (5) , all shown in Figure 1. The octahydropyrrolo[2,3ꢀb]pyrrole derivatives series (6) is also
reported in the same figure.
Figure 1 goes here
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 105
All these structures can, at least theoretically, fit the general “pharmacophore model” for D3R
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
5
antagonists originally described by Hackling and Stark . In Figure 2, 1 was pictorially positioned along
such a model, which requires an aryl region, an Hꢀbond acceptor, a spacer, the “amino” region and
another aryl moiety.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2 goes here
The tridimensional representation of this theoretical space in a D3R model clearly highlights the
medicinal chemistry challenges associated with the identification of an appropriate “amino portion”.
Such a scaffold must be compliant with specific criteria including appropriate basicity to interact with
3
.32
16
Asp in the orthosteric binding site ; contribution to the overall selectivity of the molecule vs. the DA
D2 receptor (D2R), and novelty from an intellectual property perspective. Last, but not least, new
17
molecules must have the appropriate “developability” profile to be progressed towards clinical trials.
1
8
Mixed approaches based on the D3R crystal structure were attempted for the identification of the
8
,9
19
previously mentioned NCEs . As described more than 10 years ago , “scaffold hopping” is also used
by medicinal chemists to identify structurally novel molecules starting from known active compounds
and modifying their central core structure.
This approach can lead to completely different core structures that can then be characterized through
screening cascades. In this manuscript, our efforts led to the identification of the 5ꢀazaspiro[2.4]heptane
moiety as the novel “amino portion”.
8
,9
Based on our previous experience the new molecules identified in this exploratory process were
initially characterized for their binding activity at the human D3R and D2R, while a few selected
derivatives were also evaluated in functional assays to determine their agonist and antagonist properties
at the same receptors. The aim of this approach was to identify potent D3R antagonists (pKi ≥ 8),
endowed with high selectivity (ideally ≥ 100ꢀfold) vs. the D2R, high (≥ 100ꢀfold or hERG fpKi < 6.0)
selectivity vs. the hERG ion channel (for preventing potential cardiovascular liability) as well as a ≥ 50ꢀ
ACS Paragon Plus Environment

Page 5 of 105
Journal of Medicinal Chemistry
fold selectivity vs. muscarinic M1ꢀ3 receptors (based on previous liability shown by some D3R
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
8
,9
antagonists reported in literature ). Exceptions to these general rules were applied in case of need of
additional SAR information both in vitro and in vivo.
To ensure appropriate developability, all NCEs went through in vitro pharmacokinetic (PK) assays such
as CYPEX bactosome P450 inhibition and rat and human in vitro clearance (rCli and hCli, respectively)
in liver microsomes early in the screening cascade.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The synthesis of the new scaffolds is described in General Synthetic Scheme, while all the new
compounds identified (7ꢀ85a) are reported in Tables 1 to 11. Compound 1 was used as internal standard
5
for biological assays, while literature data were reported for 2.
General Synthetic Scheme goes here
Since the 5ꢀazaspiro[2.4]heptane template contains two stereogenic centers, the relative stereochemistry
of the compounds can lead to the isolation of both “cis” and “trans” isomers. Early in the process, the
racemate was submitted to test and promising results resulted in each single enantiomer being tested
after chromatographic separation. Later on in the exploration, only specific isomers were tested; the Xꢀ
rays of a few key intermediates were obtained as well as described later on.
The exploration of the newly designed “amino portion” started with an unsubstituted phenyl ring (see
Table 1 for derivatives 7 and 8). Derivative 8b, which showed nanomolar affinity at the D3R, had about
140ꢀfold selectivity vs. the D2R and more than 500ꢀfold selectivity vs. the hERG channel. This
compound was further profiled and proved to be fully inactive as an agonist at the D3R; it behaved as an
antagonist (fpKi = 8.18 ± 0.09) with 100ꢀfold selectivity vs. the D2R (fpKi = 6.17 ± 0.16) when tested in
a functional DA D2 assay (see Table 8 for a comparison between binding and functional data). In
addition, derivative 8b was completely inactive (fpKi < 4.0) at the DA D1 receptor (D1R) and DA D4
receptor (D4R)), whereas 100ꢀfold selectivity was also observed at the M1 and M3 receptors (fpKi =
6.13 ± 0.11 and 5.69 ± 0.25, respectively) suggesting that the new scaffold could be a promising starting
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 105
point for exploration. Considering its in vitro PK properties, compound 8b showed IC values greater
5
0
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
than 50 ꢀM on all CYP P450 isoforms (namely CYP1A2, CYP2C9, CYP2C19, CYP3A4 DBOMF and
CYP3A4 7BQ) tested with the exception of CYP2D6 (IC = 0.13 ꢀM).
5
0
The introduction of a –CF moiety on the para position of the phenyl ring (9, 10, Table 1) of the “amino
3
2
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
portion” increased the lipophilicity (cLogP) of the overall molecule of about one unit. This increase
led to an improvement in the affinity at the D3R of about one log unit on the best enantiomers both in
the trans (9a) and in the cis (10a) configurations of the 5ꢀazaspiro[2.4]heptane moiety. The difference in
affinity between the two enantiomers was greater for the cis vs. the trans isomer. The substitution had
no effect on the functional activity of the compounds, which continued to behave as full antagonists
(Table 8). No modification on the muscarinic profile of compound 10a was observed (fpKi = 6.51 ±
0
.04 and 6.56 ± 0.11 respectively for M1 and M3 receptors, Table 11) maintaining the selectivity at
about 400ꢀfold over these receptors. Selectivity vs. the D2R was increased by about 350ꢀfold, while
selectivity vs. the hERG channel was greater than 600ꢀfold, even if the affinity for the channel was also
slightly increased to 0.4 ꢀM.
Both compounds were also tested for in vitro PK properties (Table 9). Compound 9a showed IC50
values greater than 17 ꢀM on all CYP P450 isoforms tested with the exception of CYP2D6 (IC = 1.3
5
0
ꢀ
M), while compound 10a showed IC values greater than 17 ꢀM on all CYP P450 isoforms tested
50
with the exception of CYP2D6 and CYP3A4 7BQ (IC50 = 1.3 ꢀM and 0.30 ꢀM, respectively).
Derivative 9a showed relatively low human intrinsic clearance (hCli) on microsomes (22 µL/min/mg
protein) and high clearance in rat (rCli) microsomes (161 µL/min/mg protein). However, derivative 10a
showed relatively high Cli in both species (hCli= 86 and rCli= 83 µL/min/mg protein, respectively). The
1
7e
last in vitro parameter to be considered (based on the need for brain penetration) was the blood and
tissue (rodent) protein binding. Both derivatives had excellent fraction unbound (Fu) both in brain (Fubr)
and blood (Fu ). For 9a the Fu was 4.8%, while its Fu was 12%. Derivative 10a showed Fu equal to
bl
br
bl
br
9.5% and Fu equal to 10.2%. Altogether these data suggested that these molecules are endowed with in
bl
ACS Paragon Plus Environment

Page 7 of 105
Journal of Medicinal Chemistry
vitro properties that are suitable for further optimization. Selected PK data are reported in Tables 9 and
0.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
9
,16
In agreement with previously reported procedures , derivative 10a was fitted in a D3R model based on
21
xꢀray crystal structure using MOE . As clearly shown in Figure 3, derivative 10a (red) can adopt similar
conformations to compound 2. One may appreciate, with the obvious limitation of reproducing in two
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
dimensions a 3D interaction, that the basic nitrogen of the new “amino portion” can interact with
3,32
Asp . Both derivative 10a and derivative 2 (in yellow) formed a salt bridge with the aspartic acid in
the receptor. Furthermore, the aromatic portion (the oxazole in both compounds) clearly pointed towards
1
6
the extracellular loop (EL) of the receptor in agreement with literature data . Finally, the new template
was able to direct the –CF moiety of the phenyl ring of derivative 10a in a similar hydrophobic region
3
common with compound 2.
Figure 3 goes here
Keeping constant the pꢀCF substitution on the phenyl ring and shortening the linker chain (from C3 to
3
C2, derivatives 11 and 12) led to a reduction in affinity at the D3R; an increase in length (from C3 to
C4) also had a detrimental effect on the affinity at the D3R (13). The change in the linker length (12, 13,
Table 1) did not modify the functional efficacy of the compounds as reported in Table 8.
To verify whether the rotation of the phenyl group of the “amino portion”, i.e. whether a change in the
amount of that dihedral angle, had any effect on the affinity of the compounds both at the D3R receptor
and at the hERG channel, the –CF substituent was added in ortho position (without changing the
3
cLogP, 14, 15, Table 1). Although the impact on affinity was not obvious for the trans isomer, basically
maintaining nanomolar affinity at the D3R, a more pronounced effect was observed with the cis isomer
with a reduction in affinity of about 10ꢀfold at the D3R. The affinity of the best compound at the hERG
channel was not affected.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 105
To verify if the cLogP itself had a role in driving the affinity at the D3R, the –CF moiety was replaced
3
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
by a single –F atom (16, 17, Table 1; cLogP lowered from 3.4 to 2.7). Comparing the most potent
compound in each series (i.e. 10a vs. 17b) no major difference was observed in terms of affinity and
selectivity vs. the D2R and hERG targets. Furthermore, the functional profile (full antagonism, Table 8)
remained unchanged. The in vitro PK profiling (Table 9) for 17b showed IC values greater than 17 ꢀM
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
50
on all CYP P450 isoforms tested with the exception of CYP2D6 and CYP3A4 7BQ (IC = 0.7 ꢀM and
5
0
3.4 ꢀM, respectively). The compound showed relatively low hCli on microsomes (37 µL/min/mg
protein) and moderate rCli (73 µL/min/mg protein). Finally, both Fu and Fu were exceptionally high
br
bl
with values of 27% and 38%, respectively supporting a promising profile to interact with the D3R in the
brain.
The introduction of a second –F atom (18, 19, Table 1) led to a slight change in the overall cLogP, but
had no major impact on the affinity at the D3R and the selectivity profile of the best enantiomer (19b).
Furthermore, the functional profile remained unchanged as reported in Table 8.
The in vitro PK profile (Table 9) of 19b revealed IC values greater than 17 ꢀM on all CYP P450
5
0
isoforms tested with the exception of CYP2D6 (IC = 1.2 ꢀM). Although hCli on microsomes was
5
0
relatively high (53 µL/min/mg protein), rCli was definitely higher (161 µL/min/mg protein). Both Fubr
and Fu were high with values of 28% and 33%, respectively, suggesting again a promising profile to
bl
interact with the D3R in the brain.
A slight increase in cLogP with respect to compound 10a was achieved with the introduction of a single
–F on the already –CF substituted phenyl ring (cLogP from 3.4 to 3.6 in derivatives 20ꢀ23, Table 1).
3
The introduction of the –F in position 2, keeping the –CF in position 4 of the phenyl ring, did not
3
change significantly the affinity of the best enantiomer (21b) at the D3R and at the hERG channel with
still more than 500ꢀfold selectivity vs. the D2R while increasing selectivity vs. the hERG channel to
more than 1250ꢀfold. The cis enantiomer showed greater affinity at the D3R compared to the trans
enantiomer. As reported in Table 8, functional antagonism was confirmed for compounds 20b and 21b.
ACS Paragon Plus Environment

Page 9 of 105
Journal of Medicinal Chemistry
A slightly lower selectivity window with high D2R functional potency was observed for 21b (Table 8).
With regards to in vitro PK properties (Table 9), 20b showed an improved profile at CYP2D6 (IC = 5.7
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
50
ꢀM) with IC50 values greater than 17 ꢀM on all CYP P450 isoforms tested with the exception of
CYP1A2 (IC = 0.8 ꢀM). Compound 21b showed IC values greater than 17 ꢀM on all CYP P450
5
0
50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
isoforms tested with the exception of CYP2D6 (IC = 0.6 ꢀM). Considering a marked difference in the
5
0
Cli of the two derivatives, with 20b showing low hCli and rCli (20 and 23 µL/min/mg protein,
respectively) and 21b showing high hCli and moderate rCli (107 and 59 µL/min/mg protein,
22, 23
respectively), the two compounds were selected to validate in vitro data with rodent in vivo data
.
2
3
Therefore, both compounds were tested in a portal veinꢀcannulated rat model . Compound 20b showed
a fraction absorbed (F %) of 70% with moderate distribution volume (V = 4.9 L/kg) and moderate
a
ss
clearance in blood as expected from the in vitro parameters (Cl = 32.9 mL/min/kg) leading to moderate
b
halfꢀlife (T1/2 = 2.3 h) and good bioavailability (F% = 49%) given the relatively low hepatic extraction
(
E = 0.31). Compound 21b had a fraction absorbed (F %) of 8.6% with low distribution volume (V =
H a ss
0
.49 L/kg), clearance in blood in relative agreement with the in vitro parameters (Cl = 15.1 mL/min/kg)
b
although lower than expected. The halfꢀlife was short (T = 0.41 h) and bioavailability was low (F%=
1
/2
4
%) because of the poor fraction absorbed and a doubled E (0.56). These data are reported in Table 10.
H
This head to head comparison therefore demonstrated a relatively good reliability of the Cli parameters,
but also highlighted a striking difference in terms of the in vivo PK properties of the diastereoisomers.
Brain penetration was also assessed and a brain to blood (B/B) ratio was achieved using the area under
the curve (AUC) values. Both compounds had good ratios, namely 1.6 and 0.8 for 20b and 21b,
respectively (Table 10).
Interestingly enough, when the –F was kept in position 4 of the phenyl ring and the –CF moiety was in
3
position 2, the most active enantiomer resulted in the trans derivative 22a (Table 1); neither the overall
affinity of this compound at the D3R nor its selectivity vs. offꢀtargets changed significantly. The in vitro
PK parameters of this derivative did not change either (Table 9). 22a showed a Fu = 8.9% and a Fu =
br
bl
1
2.3%, while both hCli and rCli were moderate (65 and 51 µL/min/mg protein, respectively). In
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 105
contrast, derivative 23b showed Fu and Fu values that were comparable to the other geometric isomer
br
bl
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(8.6 and 13.5% respectively), both hCli and rCli significantly increased (320 and 151 µL/min/mg
protein, respectively) suggesting a role of the 2ꢀCF moiety in the relative conformation of the phenyl
3
ring, potentially through a change in dihedral angle in the cis derivative to a more metabolically unstable
situation.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The introduction of a hydroxyl moiety in the side chain (C3 side chain, derivative 24, Table 1) to look
for additional interaction within the receptor and to reduce the cLogP ever further, was detrimental for
1
2b
the affinity profile, similarly to what was also reported for a different series by Newman et al.
Finally, the introduction of a 3,5 diꢀchloro substitution on the phenyl (25,26, Table 1) had no major
effect on the affinity at the D3R, but was clearly detrimental for the selectivity vs. the D2R and vs. the
hERG channel.
Based on the preliminary results of this early exploration, it was decided to maintain the p–CF phenyl
3
moiety of the “amino portion”, but to replace the methyl oxazole ring of 10 with different aromatic or
heteroaromatic moieties.
The results for 5ꢀmembered heteroaromatics are reported in Table 2. The introduction of a methyl
2
0
thiazole (27, 28) despite an increase in cLogP and in PSA led to a slight decrease in affinity at the D3R
when compared to the racemic cis derivatives (28 vs. 10). The removal of the methyl group and a
change in relative orientation of the heteroatoms (29, 30) led to equipotent derivatives (30 vs. 10) with
about 300ꢀfold selectivity over the D2R and the hERG channel, but with more than one log unit increase
in cLogP. The introduction of one extra nitrogen atom in the system (31, 32), despite a significant
increase in PSA, left the primary affinity almost unchanged, but produced an unexpected drop in affinity
at the hERG channel in the cis racemate (32), which did not match the significant increase in affinity
observed for the trans derivative (31). The functional profile of 32 was unchanged (Table 8).
Unfortunately, and in contrast with other compounds in this series, 32 was not progressed since it
showed a significantly reduced selectivity vs. the M1 receptor (fpKi= 6.77 ± 0.18, Table 11).
ACS Paragon Plus Environment

Page 11 of 105
Journal of Medicinal Chemistry
The introduction of a 2 thiophene (33, Table 2) further increased lipophilicity while reducing PSA. This
had a beneficial effect on the affinity at the D3R leading to a subꢀnanomolar derivative on the racemate,
but this change also had a detrimental effect on the hERG affinity. Its regioisomers (34) was equivalent
both in terms of potency and selectivity. The same behavior was also observed in the case of the furane
moiety (35, 36), but high affinity at the hERG channel was observed for derivative 36. The introduction
of a pyrazole moiety (37, 38) led to compounds with high affinity at the D3R and high selectivity.
Derivative 37 showed about 200ꢀfold selectivity vs. the D2R and the hERG channel, while 38 increased
such window to more than 450ꢀfold. To complete the 5ꢀmembered preliminary heterocyclic
replacement, the isoxazole 39 was prepared. In this case, the affinity at the D3R was maintained, but
with a slight reduction of selectivity as well as an increase in the hERG affinity. None of these
derivatives were tested for in vitro PK properties because of their racemic nature and their suboptimal
profile for further progression through the screening cascade.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The exploration of the oxazole replacement (10) was pursued with substituted phenyl rings, so that some
of the original oxazole polarity/charge distribution was present (see Table 3 with only the best
enantiomer of the derivatives with cis configuration).
Derivative 40a (pꢀCONH phenyl, Table 3) achieved a 2200ꢀfold selectivity over the D2R and a 5000ꢀ
2
fold selectivity over the hERG channel, thus supporting the potential of this subꢀseries. The compounds
showed no activity at the D1R and D4R (pKi < 4.50) and a more than 2500ꢀfold selectivity was also
achieved vs. M1 and M3 receptors (fpKi= 6.34 ± 0.05 and 6.37 ± 0.08, respectively, Table 11). The in
vitro PK profile (Table 9) revealed IC values greater than 5 ꢀM on all CYP P450 isoforms tested. hCli
5
0
on microsomes was high (155 µL/min/mg protein), rCli was moderate (41 µL/min/mg protein); Fu was
br
equal to 4.4% and Fu resulted 14.1%. The modelling studies performed on the compound (Figure 4)
bl
suggested the possibility for the amide to positively interact with additional groups present in the
1
6
secondary binding pocket (SBP) of the D3R.
Moving the amide portion to position 3 of the aromatic ring (mꢀCONH phenyl; 41a, Table 3) led to
2
about half a log unit reduction in the affinity at the D3R, but also led to a similar reduction in affinity at
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 105
the other offꢀtargets. Consequently, selectivity vs. the D2R was greater than 1300ꢀfold and selectivity
over the hERG channel was 6000ꢀfold; more importantly the absolute hERG value was 3 ꢀM compared
with a 0.4 nM affinity at the D3R. The compound showed no affinity at the D1R and only low affinity at
the D4R (fpKi= 5.23 ± 0.03); affinity at the M1 and M3 receptors was also moderately low (fpKi= 6.58
± 0.16 and 6.42 ± 0.10, respectively, Table 11) leading to more than 550ꢀfold selectivity at these
receptors. When tested for in vitro PK properties (Table 9), 41a showed IC values greater than 3 ꢀM
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
on all CYP P450 isoforms tested. hCli on microsomes was comparable to derivative 40a (187
µL/min/mg protein), while rCli was higher (92 µL/min/mg protein) suggesting an increased metabolic
transformation of the now unprotected position 4 of the phenyl ring in rodents. Conversely, both Fubr
and Fu were quite similar to the previous compound (3.7 and 6.7%, respectively). Both 40a and 41a
bl
were tested in a functional D3R assay, confirming an unchanged antagonist behavior (Table 8).
The insertion of the amide moiety in position 2 of the phenyl ring (oꢀCONH phenyl; 42a, Table 3) led
2
to a significant decrease in affinity at the D3R (about 100ꢀfold when compared to 40a) as well as a
decrease in selectivity vs. the D2R. It was hypothesized that the change in the dihedral angle induced by
the oꢀamide insertion disfavored the interactions with key residues within the D3R and the SBP, leaving
the residues within the D2R unchanged. Alternatively, one may hypothesize that such a substituent may
have generated steric clashes preventing the best fitting within the binding pocket. To find additional
hints which could help to explain better the observed SAR, all the three structures were modeled within
9
the D3R model (Figure 4) as previously reported , and the GBVI/WSA dG scoring function as
21
implemented in MOE was used. Subsequently the “lowest score” poses were submitted to
conformational search using LowModeMD (MOE) and the lowest energy conformations obtained were
further minimized. The ligandꢀprotein affinity was reꢀcalculated and the analysis of the values observed
(
data in the experimental part), suggested a trend in line with the binding affinity of the derivatives (40a,
1a, 42a), even if the absolute difference in energy did not fully justify the differences observed in the
binding affinities for 42a.
4
ACS Paragon Plus Environment

Page 13 of 105
Journal of Medicinal Chemistry
1
2
Figure 4 goes here
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The introduction of a pꢀSO NH phenyl (43a, Table 3) slightly decreased cLogP, but greatly increased
2
2
the PSA value; the affinity at the D3R was maintained and the selectivity over the D2R and the hERG
channel remained high (about 700ꢀfold). Unfortunately, the absolute affinity at the hERG channel was
about 0.2 ꢀM and the compound was not progressed. The replacement of the sulphonamide group with
an acetyl moiety (44a) increased cLogP reducing the PSA value. The affinity at the D3R was unaltered
leading to a selectivity vs. the D2R greater than 2000ꢀfold. Unfortunately, the fpKi observed at the
hERG channel was similarly high. The nitrile derivative 45a behaved similarly with very high affinity at
the D3R and high selectivity vs. the D2R, but with a comparably high hERG affinity. In contrast, the
introduction of a p–CH CONH (46a) group on the phenyl group led to a slight decrease in affinity at
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
the D3R with respect to derivative 10a. Selectivity vs the D2R remained greater than 500ꢀfold and the
selectivity vs. the hERG channel was increased to about 8900ꢀfold with an absolute value of 7.5 ꢀM,
leading to one of the lowest hERG affinity up to that point of the exploration. Accordingly, in vitro PK
data were generated (Table 9) and the compound showed IC values greater than 7 ꢀM on all CYP
5
0
P450 isoforms tested; hCli on microsomes was moderate (58 µL/min/mg protein), while rCli was low
(
22 µL/min/mg protein). In terms of free fraction, both Fu and Fu were relatively high (5.2 and 14.1
br bl
%, respectively).
In light of the promising results achieved with derivatives 40a and 41a, a minimal bioꢀisosteric
replacement of the primary amide was attempted and the oxazole derivatives 47a and 48a were
prepared. Derivative 47a reached subꢀnanomolar affinity (0.07 nM) at the D3R with a more than 5300ꢀ
fold selectivity window vs. the D2R and more than 2200ꢀfold selectivity over the hERG channel. The
compound showed no affinity at the D1R and D4R; the affinity at the M1 and M3 receptors was also
moderately low (fpKi= 6.17 ± 0.01 and 6.52 ± 0.11, respectively, Table 11) leading to about 4500ꢀfold
selectivity at these muscarinic receptors. The functional activity confirmed the antagonistic activity at
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 105
the D3R (Table 8). Despite a relatively high cLogP, the in vitro PK profile (Table 9) of 47a showed IC50
values greater than 3 ꢀM on all CYP P450 isoforms tested. In line with our expectations, however, this
high value was also reflected in a reduction of free compound both in brain and blood (Fu and Fu =
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
br
bl
1
.0 and 1.2 %, respectively). In addition, Cli values reflected the increased lipophilicity: hCli and rCli
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
were increased to 352 and 276 (µL/min/mg protein), respectively. Comparable results were obtained
with derivative 48a (Table 9). No further bioꢀisosteric replacement exercise was performed at that stage
of the project and this activity was postponed to a subsequent lead optimization phase. Functional data
for 48a are reported in Table 8.
A rapid exploration of the tolerability of the scaffold to basic 6ꢀmembered heteroaromatics was
performed using racemic mixtures (and consequently both trans and cis isomers were also assayed). The
results of this exploration are reported in Table 4. For the pyridine moiety (49ꢀ54), the cis pyridinꢀ2ꢀyl
substituent (54) showed the best affinity at the D3R and a 660ꢀfold selectivity vs. the D2R. The cis
derivative 52 (pyridineꢀ3ꢀyl moiety) showed a more balanced profile with increased selectivity at the
hERG up to 680ꢀfold and a 1.2 ꢀM absolute hERG value vs. a 0.2 ꢀM hERG value for derivative 54.
Although the introduction of a second nitrogen atom (55ꢀ62) in the sixꢀmembered ring system proved to
be efficient for the reduction in overall lipophilicity, it failed to show any major advantage in terms of
affinity at the D3R and overall selectivity. One of the most balanced derivatives was the pyridazineꢀ4ꢀyl
(62), which showed (once again on the cis isomer) good affinity at the D3R and 390ꢀ and 630ꢀfold
selectivity over the D2R and the hERG channel, respectively (the absolute hERG value of 2.5 ꢀM was
observed for 62).
In light of the overall balanced profile of the cis 3ꢀpyridinyl derivative 52, the unsubstituted derivative
6
3 was prepared (Table 5). The direct comparison between the two racemic compounds clearly show
that the removal of the 4ꢀCF group in the phenyl derivative of the “amino portion” led to a decrease in
3
affinity greater than one log unit. For compound 64b (the best cis single enantiomer), such substitution
was reꢀintroduced adding an additional 2ꢀF substitution. The result was a derivative showing almost 1.2
nM affinity at the D3R and a selectivity greater than 600ꢀfold at the D2R and hERG channel. The in
ACS Paragon Plus Environment

Page 15 of 105
Journal of Medicinal Chemistry
vitro PK (Table 9) revealed IC values greater than 7 ꢀM on all CYP P450 isoforms tested with the
5
0
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
exception of CYP2D6, which was greater than 1 ꢀM. Both Fu and Fu were relatively high with 7.7 %
br
bl
and 16.1 %, respectively. As clearly expected for a nonꢀsubstituted 3ꢀpyridine, the in vitro clearance was
characterized by high values (namely, hCli and rCli= 485 and 278 µL/min/mg protein, respectively).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A 2,4ꢀF diꢀsubstitution, i.e. the replacement of the 4ꢀCF group with a –F atom (65b, Table 5), led to a
3
slight reduction in the affinity at the D3R with the advantage of a more than half one log unit decrease in
cLogP. While maintaining a 1.2 nM affinity at the D3R and a 400ꢀfold selectivity vs. the D2R, this
compound increased selectivity vs. hERG to over 2100ꢀfold with an absolute 2.8 ꢀM hERG affinity. Its
functional behavior was also confirmed (Table 8). A slight improvement was observed with respect to
the CYP P450 profile, while a large increase in the free fraction was seen both in brain and blood (Fubr
and Fu = 25 and > 50 %, respectively, Table 9). Similar to the previous compound, Cli data were high
bl
(namely, hCli and rCli = 258 and 253 µL/min/mg protein, respectively). Finally, the decision to leave a
single 4ꢀF substituent (66b, Table 5) on the aromatic ring further reduced cLogP almost to the original
value of the unsubstituted derivative. This reduction had no major impact on the affinity at the D3R,
keeping the selectivity profile comparable to the previous derivatives. The fraction unbound (Table 9) of
the compound was high (Fu and Fu = 26 and > 50 %, respectively) and, most probably because of the
br
bl
decrease in lipophilicity and despite the presence of an unsubstituted 3 pyridine, Cli values were
markedly reduced with respect to the previously analyzed compounds leading to hCli value equal to 91
µL/min/mg protein.
Given the results of this very rapid exploration and to maintain homogeneous comparison with the
previously synthesized compounds, a further analysis of the aromatic pattern of substitution of this part
of the molecule was postponed to the lead optimization phase. The 4ꢀCF group was kept as a reference
3
while substituting the 3ꢀpyridine with groups that, based on Cli data, were hypothesized to prevent the
metabolic degradation of the ring.
The insertion of a methyl group in position 2 of the pyridine system (67a, Table 6) was slightly less
potent than the corresponding unsubstituted racemate (52), but was also less active at offꢀtargets leading
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 105
to 450ꢀ and 1900ꢀfold selectivity vs. the D2R and the hERG channel; derivative 67a showed a 6.3 ꢀM
hERG activity. According to the working hypothesis, the methyl group should have increased the
dihedral angle formed by the pyridine ring with the triazole moiety, hence reducing the conjugation of
the two rings. Compound 67a was inactive at the D1R and D4R and the activity at the M1 and M3
muscarinic receptors was quite reduced leading to a selectivity greater than 350ꢀfold at both receptors
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(fpKi= 5.78 ± 0.14 and 5.91 ± 0.07, respectively, Table 11). The impact on the CYP P450 profile was
positive with IC values greater than 7 ꢀM on all CYP P450 isoforms tested with the exception of
50
CYP3A4 7BQ, which was greater than 2 ꢀM. Both Fu and Fu were moderate with values of 6.2 and
br
bl
1
8.6 %, respectively. A clear reduction in Cli was also observed with hCli and rCli equivalent to 80 and
103 µL/min/mg protein, respectively, Table 9).
When this methyl was replaced by a trifluoromethyl group (68a, Table 6) cLogP was clearly increased; a
decrease in affinity at the D3R and a reduced selectivity profile were also observed. In contrast, this
manipulation had a positive effect on the CYP P450 profile; on all the CYP P450 isoforms tested,
compound 68a had IC50 values greater than 6 ꢀM. However, the increased lipophilicity had a
detrimental effect on free fraction (Fu and Fu = 2.2 and 7.5 %, respectively, Table 9) as well as on
br
bl
intrinsic clearance; hCli and rCli reverted to 302 and 114 (µL/min/mg protein), respectively.
A reduction in cLogP and a slight increase in PSA were calculated for the replacement of the –CF with
3
an electronꢀdonating group such as the methoxy one (69a, Table 6). Potency and selectivity vs. the D2R
were comparable to the methyl derivative 67a, while the affinity at the hERG channel increased by one
log unit.
When derivative 70a (cyano derivative) was prepared to probe the other position close to the pyridine
nitrogen, an increase in affinity at the D3R was observed and a good selectivity vs. the D2R was
maintained. The affinity at the hERG channel, however, also increased and it was decided not to
progress this compound along the screening cascade. The transformation of the nitrile into a primary
amide (71a) produced a one log unit increase in the affinity at the D3R and the selectivity vs. the
secondary targets was greater than 1600ꢀfold. The in vitro PK profile (Table 9) of 71a revealed IC50
ACS Paragon Plus Environment

Page 17 of 105
Journal of Medicinal Chemistry
values greater than 17 ꢀM on all CYP P450 isoforms and it was therefore decided to investigate its
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
profile further. The reduction in lipophilicity (cLogP= 3.4) had a beneficial effect on free fraction (Fubr
and Fu = 2.2 and 18.4 %, respectively) and on intrinsic clearance (hCli and rCli were 106 and 29
bl
µL/min/mg protein, respectively). Compound 71a behaved as a competitive antagonist at the D3R
(Table 8) with no activity at the D1R and the D4R, and a 550ꢀfold selectivity over muscarinic M1 and
M3 receptors (fpKi= 6.65 ± 0.05 and 6.72 ± 0.04 on M1 and M3, respectively, Table 11); this compound
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2,23
was therefore tested in vivo
to validate correlation with in vitro PK data. Compound 71a had a
relatively good fraction absorbed (F = 40%, Table 10) with moderate distribution volume (V = 3.2
a
ss
L/kg), and relatively low clearance in blood as expected from the in vitro parameters (Cl = 19.4
b
mL/min/kg) leading to moderate halfꢀlife (T1/2 = 2.2 h) and relatively good bioavailability (F%= 40%)
given low hepatic extraction (E = 0.35). As expected from increased PSA values, the B/B ratio was
H
slightly lower than the previously tested compounds (B/B= 0.3). The overall parameters for this
compound constituted an excellent starting point for further exploration in the lead optimization phase.
The exploration of the substitution of the 3 pyridine ring continued with the introduction of two methyl
groups (72a, Table 6) to assess whether the contemporary substitution of the two positions alpha to the
nitrogen was tolerated. No major differences with the mono substituted methyl derivative 67a were
observed, suggesting that the second methyl was relatively “neutral” to the overall molecule in terms of
affinity. The slight increase in lipophilicity had no major impact on the CYP P450 profile (all isoforms
had IC values greater than 9 ꢀM) or on free fraction (Fu and Fu = 4.5 and 12.6 %, respectively), but
5
0
br
bl
affected Cli with increased values (hCli and rCli were 299 and 98 µL/min/mg protein, respectively,
Table 9).
Replacing one of the two methyls with the primary amide of 71a led to the synthesis of compound
7
3a(Table 6). This derivative showed a balanced affinity profile with 1.2 nM affinity at the D3R and
9
00ꢀfold and 2150ꢀfold vs. the D2R and the hERG channel, respectively. Moreover, the absolute affinity
at the hERG channel was about 3 ꢀM. The compound had a “clean” CYP P450 profile showing IC50
values greater than 10 ꢀM on all isoforms tested. The rCli was greatly reduced with respect to 72a
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 105
(
rCli= 38 µL/min/mg protein. Table 9), while hCli values remained unexpectedly high (hCli = 299
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
µL/min/mg protein). High free fraction values were also observed (Fu and Fu = 7.8 and 19 %,
br
bl
respectively).
The compound was further profiled in vitro for selectivity and functional activity was also assessed. The
compound was an antagonist with 0.7 nM affinity at the D3R thus increasing the selectivity vs. the
hERG channel to more than 4000ꢀfold; no activity at the D1R and the D4R was observed and a 270ꢀfold
selectivity was achieved over the muscarinic M1 and M3 receptors (fpKi= 6.29 ± 0.08 and 6.73 ± 0.08
on M1 and M3, respectively, Table 11). Given the balanced profile of 73a, its PK profile was assessed
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2, 23
in vivo
. Compound 73a showed a good fraction absorbed (F = 54%, Table 10) with moderateꢀhigh
a
distribution volume (V = 5.5 L/kg) and low clearance in blood as expected from the in vitro parameters
ss
(Cl = 21.1 mL/min/kg). Given the low hepatic extraction (E = 0.28), the parameters showed a good
b
H
halfꢀlife (T1/2 = 4.3 h) and relatively good bioavailability (F%= 53%). Brain penetration was also
assessed showing a B/B ratio equal to 1.5. Similar to 71a, this compound was selected as a starting point
for further characterization.
Finally, the transformation of the primary amide into the corresponding carboxylic acid (74a, Table 6)
caused an overall reduction in affinity across all targets, and IC on all CYP P450 isoforms tested
5
0
greater than 38 ꢀM. Quite unexpectedly for a free carboxylic acid, very high free fractions were
observed (Fu and Fu = 22.9 and 29.9 %, respectively, Table 9) and Cli values were both very low (r,
br
bl
hCli < 11 µL/min/mg protein). Further exploration of this compound during the lead optimization phase
of this project was deemed worthy based on its overall selectivity that was greater than 100ꢀfold vs. both
the D2R and hERG channel (10 ꢀM affinity).
The exploration of the pyridine moieties 50 and 52 was continued (Table 7) with the introduction of a
hydroxyl substituent on the ring leading to the pyridone derivatives 75 and 76. Only data on the most
active cis stereoisomers are reported. Both compounds were originally prepared as racemates showing
promising selectivity profiles and nanomolar affinity at the D3R. Since 75 had an IC at the hERG
5
0
channel greater than 10 ꢀM, its most active enantiomer (75a) was further profiled showing selectivity
ACS Paragon Plus Environment

Page 19 of 105
Journal of Medicinal Chemistry
greater than 4300ꢀfold vs. the hERG channel. No affinity at the D1R and the D4R was observed whereas
selectivity vs. muscarinic M1 and M3 receptors was higher than 200ꢀfold (fpKi= 6.31 ± 0.01 and 6.01 ±
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
.16 on M1 and M3, respectively, Table 11). Given its IC greater than 5 ꢀM on all CYP P450 isoforms
5
0
tested, its high free fraction (Fu and Fu = 8.0 and 19.1 %, respectively) and low Cli both in rat and
br
bl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
human (hCli= 28 and rCli = 14 µL/min/mg protein, respectively), it was decided to assess its PK profile
2
2,23
24
in vivo
. Unfortunately, because of its low membrane permeability (assessed in Cacoꢀ2 cell lines)
the fraction absorbed of 75a was negligible (F ≈ 1%, Table 10); this problem, associated with high Clb
a
(48 mL/min/kg) and very high distribution volume (V = 9.2 L/kg), led to negligible bioavailability
ss
(F%≈ 2%). For this reason, compound 76 (Table 7) was not even separated in its enantiomers and the Nꢀ
methylated version of both compounds was prepared (77a, 78a). Both derivatives showed an excellent
affinity and selectivity profile, especially at the hERG channel showing 1800ꢀ and 1300ꢀfold hERG
selectivity, respectively. Their overall in vitro properties were similar: IC values on all CYP P450
5
0
isoforms tested was greater than 9 ꢀM and greater than 17 ꢀM for 77a and 78a, respectively. High free
fraction was observed for both compounds (Fu and Fu = 10.8 and 25 %, respectively for 77a, and Fu
br
br
bl
and Fu = 14.6 and 16.7 %, respectively for 78a, Table 9 ). Both compounds also showed comparable
bl
Cli profiles (see Table 9). The medium permeability in the Cacoꢀ2 assay for both 77a and 78a was
22, 23
slightly better than 75a. The latter observation was critical considering the in vivo
PK profile. As
expected by the in vitro results reported in Table 9, Clb values (reported with the full profiles in Table
10) were almost identical with moderate clearance in rat, as well as V and T . The difference in
ss
1/2
bioavailability of the two compounds (10 vs. 30%) perfectly reflected and matched their difference in
fraction absorbed (17 vs. 40%). Functional data for 78a are available in Table 8.
The full removal of the aromaticity from the pyridine ring led to the piperidinyl derivatives 79a and 80a
(Table 7). The compounds both showed affinity at the D3R of about 10 nM with 100ꢀfold selectivity vs.
the D2R and a 1000ꢀfold selectivity vs. the hERG channel (10 ꢀM). The introduction of a lactam ring,
especially on derivative 79a, reduced cLogP and PSA values, which translated into the CYP P450
profile (IC of 79a greater than 21 ꢀM on all isoforms tested). The free fraction (Fu and Fu = 17.8
50
br
bl
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 105
and 23 %, respectively) and Cli (hCli= 25 and rCli < 11 µL/min/mg protein. Table 9) had benefited from
this reduction. Given its low permeability the compound was not progressed in that phase of the project.
To pursue the exploration, the carbonyl group of the lactams (79a, 80a, Table 7) was made exocyclic by
preparing derivative 81a; the system was then further elongated with the cyclohexyl amide 82a. Neither
of the two corresponding “free” amines were prepared and tested since previous explorations on related
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
8
,9
25
compounds and literature data demonstrated that diꢀbasic compounds, especially on lipophilic
scaffolds, usually suffer from very high volumes of distribution, long halfꢀlife and high total plasma
clearance.
Compound 81a (Table 7) showed a 7 nM affinity at the D3R with more than 100ꢀfold selectivity vs. the
D2R and about 400ꢀfold selectivity vs. the hERG channel (3 ꢀM). 81a had IC values greater than 17
5
0
ꢀ
M on all CYP P450 isoforms tested with high free fraction both in brain and blood (Fu and Fu = 11.8
br bl
and 20.0 %, respectively; intrinsic clearance values were moderate to low (hCli= 61 and rCli= 17
µL/min/mg protein, respectively, Table 9). A change in the nature of the thiotriazole substituent did not
change the functional profile of the compound (Table 8). Based on medium cLogP and PSA and
medium permeability in the Cacoꢀ2 assay, this compound was selected to further assess its PK
2
2,23
properties in vivo
. 81a showed good fraction absorbed (F = 28%, Table 10) with moderateꢀhigh
a
distribution volume (V = 6.0 L/kg) and moderate clearance in blood as expected from the in vitro
ss
parameters (Cl = 39.8 mL/min/kg). Hepatic extraction was unexpectedly low (E = 0.0) leading to
b
H
moderate halfꢀlife (T = 2.5 h) and a relatively good bioavailability (F%= 33 %). Brain penetration was
1
/2
also assessed with a B/B ratio equal to 0.5.
The acetamide 82a (Table 7) was well tolerated by the D3R showing an affinity of about 3 nM at the
D3R and a very good selectivity vs. the hERG channel (2500ꢀfold). Although lipophilicity was not
significantly increased, IC50 values on the CYP P450 2D6 decreased with respect to the previous
compounds (IC = 2 ꢀM; all other isoforms > 6 ꢀM).
5
0
Following the observation that hydrophilicity was well tolerated in this region of the receptor, the
piperidine moiety was replaced by a tetrahydropyrane template (83a). This compound showed an
ACS Paragon Plus Environment

Page 21 of 105
Journal of Medicinal Chemistry
excellent affinity at the D3R (1.7 nM), 480ꢀfold selectivity vs. the D2R and 1700ꢀfold selectivity vs. the
hERG channel. Marked reduction in PSA and high permeability in Cacoꢀ2 cell lines were observed. The
affinity of 83a was also confirmed functionally (Table 8) with no activity at both the D1R and the D4R.
A good selectivity over muscarinic M1 and M3 receptors (> 100ꢀfold) was also achieved (fpKi= 6.60 ±
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
.09 and 6.05 ± 0.06 on M1 and M3, respectively, Table 11). The CYP P450 profile was excellent with
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
IC values greater than 10 ꢀM on all isoforms tested; free fraction was high (Fu and Fu = 9.7 and
5
0
br
bl
1
8.8 %, respectively, Table 9). Intrinsic clearance values in rat were higher than the ones in humans
(hCli= 34 and rCli= 59 µL/min/mg protein, respectively). Therefore, the compound was selected to
22, 23
profile its PK properties in vivo
. 83a had good fraction absorbed (F = 57%, Table 10) with moderate
a
distribution volume (V = 3.7 L/kg), and high clearance in blood as expected from the in vitro
ss
parameters (Cl = 87.3 mL/min/kg). Hepatic extraction was high (E = 0.8) leading to short halfꢀlife
b
H
(
T = 0.6 h) and low bioavailability (F%= 11 %). Brain penetration was also assessed with B/B ratio
1/2
equal to 0.3, potentially dictated by decreased PSA.
In an attempt to reduce Clb values masking the carbons alpha to the oxygen to prevent their oxidation,
the bridged derivative 84a (Table 7) was prepared. No major differences were observed in the
affinity/selectivity profile, while permeability was slightly lower. The free fraction was also slightly
reduced, and the rCli was reduced as expected (from 59 to 21 µL/min/mg protein, Table 9). To quickly
22,23
test the working hypothesis, the compound was administered intravenously (i.v.) in vivo
to assess
blood clearance in rat. Unfortunately, the difference in Cli was not reflected in a difference in Clb, and
the values for the two compounds were comparable (Table 10) and relatively high.
Finally, to complete the exploration, the oxygen of 83a was replaced by a carbon atom, leading to the
cyclohexyl derivative 85a (Table 7). A substantial increase in cLogP was calculated and translated into
an increase in affinity at the DA D3 receptor. As expected, such increase was also observed on the off
targets even if an excellent level of selectivity was still maintained (more than 2000ꢀfold vs. the D2R
and 1000ꢀfold vs. hERG). The high lipophilicity led to reduced free fraction and very high Cli (Table 9)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 105
as expected for such derivative. For both 84a and 85a, functional data were generated too (Table 8),
confirming the antagonist profile of the series.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Some of the compounds reported in Tables 6 and 7 were submitted to an “ExpresSProfile ꢀ P1” at
26
CEREP , a panel consisting of 55 receptors, ion channels and transporters at the 1 ꢀM concentration. A
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
great selectivity was observed for the compounds tested on the majority of the offꢀtargets (data not
reported) and the previously identified hERG/M1/M3 potential liability points were also highlighted,
confirming their correct inclusion in the screening cascade.
Finally, to attribute unequivocally the stereochemistry of the compounds and to revise the results of the
computational work, an appropriate strategy was set up to determine their absolute configuration. The
strategy was divided into two parts and comprised as a first step the preparation of the chiral
intermediate, (1R,3S)‐[4‐(trifluoromethyl)phenyl]‐5‐azaspiro[2.4]heptane (86) by preparation of
(1R,3S/1S,3R)‐[4‐(trifluoromethyl)phenyl]‐5‐azaspiro[2.4]heptane and resolution of the racemic
mixture by using chiral HPLC procedure as reported in the experimental section. The assignment of the
absolute configuration of the title compound was determined by a single crystal Xꢀray structure obtained
from
a
crystal
of
5ꢀ(4ꢀmethylbenzenesulfonyl)ꢀ(1R,
3S)ꢀ[4ꢀ(trifluoromethyl)phenyl]ꢀ5ꢀ
azaspiro[2.4]heptane (87) derived from the desired enantiomer and crystallized in EtOH as solvent to
obtain a single crystal. For those compounds synthesized from 86 or from its (1S,3R)‐enantiomer (87)
(with known absolute stereochemistry based on Xꢀray structure) a common trend was recognized
between absolute configuration of the 5‐azaspiro[2.4]heptane moiety and measured binding activity at
the D3R for each pair of enantiomers (i.e. the most active enantiomer of each pair of enantiomers
generally does have the same absolute configuration). Figure 5 reports the crystal structures for
derivatives 88 and 89.
Figure 5 goes here
ACS Paragon Plus Environment

Page 23 of 105
Journal of Medicinal Chemistry
Given the results achieved in this early part of the exploration and reported in Tables 1ꢀ7 some hints
about the structureꢀactivity of the new 5ꢀazaspiro[2.4]heptane moiety were identified. On this specific
scaffold, and coupled to the thiotriazole system to provide D3R antagonism, the relative configuration of
the system played a role in determining the affinity at the primary target. Both stereoisomers were active
and selective, and the cis stereoisomer resulted often more active than the trans stereoisomer (e.g. 10a
vs. 9a; 19b vs. 18b; 21b vs. 20a). Nonetheless, the steric hindrance of the substituents on the aromatic
ring of this amino portion also played an important role on the affinity and was able to reverse this
general trend (e.g. 22a vs. 23b; 14 vs. 15; 25 vs. 26). Even if this was not specifically noticed in the
molecular modelling calculations performed to support the medicinal chemists, this fact might suggest
that a specific dihedral angle in the cis configuration is required for a more appropriate fitting of the
whole molecule within the primary binding site in the D3R pocket.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
An additional observation was related to the very high level of selectivity achieved for all the derivatives
prepared in the different series. Differently from other series reported in the literature, a selectivity
greater than 1000ꢀfold over the hERG channel was observed for many of the compounds reported in the
manuscript. High cLogP values (e.g. 33, 44a, 45a, 47a) led to subꢀnanomolar affinity at the D3R, but
also caused an increase of potency at the hERG channel. Nonetheless, an optimal range of cLogP to
achieve both high D3R affinity and an “ideal” selectivity window with respect to the hERG channel was
5
identified (3 < cLogP < 4.5), also in agreement with previously reported data .
The introduction of additional Hꢀbond capabilities in the molecule (e.g. 40a, 44a, 47a) led to an increase
in the affinity at the D3R suggesting the possibility of additional interactions within the secondary
16
binding pocket (SBP) of the D3R, potentially in the region facing the extracellular loop 2 (EL2) as it
might be hypothesized by the docking of compound 40a reported in Figure 4.
The presence in the scaffold of an additional moderately basic center (49ꢀ74) was generally well
tolerated from the affinity point of view and had no major impact on the distribution volume point of
view. This, in addition to the previously mentioned Hꢀbond acceptor presence and to appropriate
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 105
protection of the molecule from the metabolic point of view, led to a series of derivatives with good
CYP P450 profile and excellent developability properties (Table 9 and 10).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 25 of 105
Journal of Medicinal Chemistry
Conclusions
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In summary, a combined medicinal chemistry and computational chemistry strategy led to the
identification of a new series of 1,2,4ꢀtriazolyl 5ꢀazaspiro[2.4]heptanes with high affinity and selectivity
at the D3R.
Despite the limited exploration of substituents in this early phase of the project, the newly identified
“amino portion” provided a wellꢀdefined SAR. The affinity at the D3R reached subꢀnanomolar values
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(e.g. 47a) with a large selectivity window, both on the D2R but especially on the hERG channel (e.g.
7
5a, 5000ꢀfold selective, or 46a, 10000ꢀfold selective).
The observed pharmacokinetic properties (both in vitro and in vivo) clearly predict good developability
characteristics for these molecules even though additional optimization work together with further in
vivo testing is now required. Some selected derivatives, which showed an overall balanced in vitro /in
vivo profile (e.g. 20b, 21b, 71a, 73a, 77a, 78a, 81a, and 83a), were considered for further progression
along the screening cascade for subsequent lead optimization, which will be reported in due course.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 105
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Experimental section
Biological Test Methods: In vitro studies
3
[
H]ꢀSpiperone filtration binding assay on membranes from hD3ꢀCHO cells.
This assay provided in vitro affinities (pIC , pKi) of compounds at the human D3R expressed in
5
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CHOꢀK1 cells. The filtration binding involved the incubation of hD3ꢀCHO cell membranes with the
3
radioligand [ H]ꢀSpiperone in the presence or absence of test compounds. Displacement binding was
measured by separation of bound from free radioligand using filtration through Glass Microfiber Filters.
The assay was performed at room temperature (23 °C). Equilibrium was reached after 80 min and
remained stable for at least 120 minutes. Bound radioactivity was measured using a TopꢀCount reader
(for competition experiments) or ßꢀCounter reader for saturation experiments), while radioligand
concentration was determined by liquid scintillation counting.
Expression plasmids
The human D3R (DRD3), transcript variant (NM_000796.2) was purchased from OriGene
Technologies, Inc. (cat # SC124083). The cDNA was cloned into pCMV6ꢀXL4 plasmid vector
containing CMV promoter for in vivo expression in mammalian cells. The molecular clone sequence
data was matched to the reference identifier above as a point of reference. The complete sequence of the
molecular clone may differ from the sequence published for this corresponding reference, e.g., by
representing an alternative RNA splicing form or single nucleotide polymorphism (SNP).
Large scale cell transfection and cell pellet preparation
CHOꢀK1 cells were maintained in F12K medium supplemented with 10% dialyzed fetal bovine
serum, in a humidified incubator at 37 °C with 5% CO . For transfection, cells were plated in Corning
2
2
Dish 500cm (20x106 cells /dish) and cultured overnight. Cells were then transfected with hDRD3ꢀ
pCMV6ꢀXL4 plasmid, following the manufacture's protocol with FuGENE HD with a ratio of
ACS Paragon Plus Environment

Page 27 of 105
Journal of Medicinal Chemistry
transfection complex of 3:1 (300 ꢀL FuGENE: 100 ꢀg plasmid/dish). After 24 h incubation at 37 °C and
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
+
2+
5% CO , the medium was removed and cells were washed with 25mL of DPBS (no Ca , no Mg ).
2
Versene was added over the cells and incubated at 37 °C for 5 min. Cells were scraped with a "cell
scraper" and the cell suspension was collected in a Costar tubes prior to being put on ice. Cell
suspensions were centrifuged in a Beckman GS6R centrifuge for 10 min, 4 °C, 1200 rpm (330 rcf), the
supernatant removed and the cell pellets washed and collected in a single tube by reꢀsuspension in PBS
and centrifugation cycles. The pellet was weighed and frozen down at ꢀ80 °C.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CHOꢀhDRD3 membrane preparation
The frozen pellet was thawed and homogenized in 10 volumes (w/v) in Membrane Preparation Buffer
1 using an Ultraꢀturrax (3 times for 10 s each cycle). The homogenate was centrifuged for 20 min, 4 °C,
at 18500 rpm (40000 rcf) in a SLꢀ50T Sorvall rotor and the pellet reꢀsuspended in 10 volumes w/v in
Buffer 1 and reꢀhomogenized as before. After centrifugation, the pellet was reꢀsuspended in 5 volumes
of Membrane Preparation Buffer 2. The resulting suspension was aliquoted and frozen down at ꢀ80 °C.
Protein concentration was determined according to the instructions provided within the BioRad reagent
using a BSA standard curve. (Buffer 1= HEPES 20 mM, EDTA 2 mM pH 7.4, ice cold; Buffer 2=
HEPES 20 mM pH 7.4, NaCl 100mM, MgCl 10mM, EDTA 1 mM, ice cold).
2
Competition binding experiments
Competition binding experiments were performed in a 96ꢀdeep well plate at room temperature (23 °C)
with a final assay volume of 500 ꢀL/well according to the following protocol: a) 300 ꢀL of binding
3
buffer were dispensed into each well of the compound plate; b) [ H]ꢀSpiperone stock was diluted in
3
3
binding buffer solution to obtain the 5x [ H]ꢀ Spiperone solution (1.5 nM); c) 100 ꢀL of 1.5 nM [ H]ꢀ
Spiperone solution were dispensed into each well of the compound plate; d) the competition reaction
was started by adding 100 ꢀL of hDRD3ꢀCHO membrane suspension in binding buffer. The final
3
membrane concentration/well was 3.5 ꢀg and the final [ H]ꢀSpiperone concentration was 0.3 nM. The
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 105
plate was then incubated on a shaker at 23 °C for 90 min. The reaction was terminated by rapid filtration
through Unifilterꢀ96 GF/B filter plates preꢀsoaked for one hour in Polyethylenimine (PEI) 0.5% (w/v)
solution and washed with 1.0 mL of ice cold 0.9% NaCl before the filtration using a Packard cell
Harvester. The filter plate was washed 4 times with 1.0 mL iceꢀcold 0.9% NaCl and then left to dry for
at least one hour at 40 °C. The plate was sealed with a backꢀseal; 50 ꢀL of Microscintꢀ20 were added to
each well and the plate was sealed with a topꢀseal. Bound radioactivity was measured using a Microplate
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
TopCount. Radioligand concentration was determined as follows: 100 ꢀL of [ H]ꢀSpiperone solution
(5x) and 3mL Filter Count were mixed in the total added vial and read in βꢀCounter TriCarb 2900.
Saturation binding experiments
Saturation binding experiments were performed similarly to the competition binding experiments,
3
with the following deviation: [ H]ꢀSpiperone concentrations were chosen from 0.015 to 4.0 nM in a
concentration response curve with 12 points. The reaction was terminated by rapid filtration through
GF\B paper filter preꢀsoaked for one hour in Polyethylenimine (PEI) 0.5% (w/v) solution and washed
with 1.0 mL of ice cold 0.9% NaCl before the filtration using a Brandel Harvester. The filter was
washed 4 times with 1.0 mL iceꢀcold 0.9% NaCl. The filter was put into picoꢀvial (PerkinElmer 600252)
and 4 mL of Filter count were added. Bound radioactivity was measured using a βꢀCounter. Samples of
working radioligand solution were taken and measured by traditional liquid scintillation counting in
order to determine the actual concentration of radiolabel added.
3
[
H]ꢀSpiperone filtration binding assay on membranes from hD2ꢀCHO cells.
This assay provided in vitro affinities (pIC , pKi) of compounds at the human DA D2 receptor
5
0
expressed in CHOꢀK1 cells. The filtration binding involved the incubation of hD2ꢀ Gα16ꢀCHO cell
3
membranes with the radioligand [ H]ꢀSpiperone in the presence or absence of test compounds.
Displacement binding was measured by separation of bound from free radioligand using filtration
through Glass Microfiber Filterplates. The assay was performed at room temperature (23 °C).
ACS Paragon Plus Environment

Page 29 of 105
Journal of Medicinal Chemistry
Equilibrium was reached after 90 min and remained stable for at least 150 minutes. Bound radioactivity
was measured using a TopꢀCount reader (for competition experiments) or sꢀCounter reader (for
saturation experiments), while radioligand concentration was determined by liquid scintillation
counting.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Competition binding experiments
Competition binding experiments were performed in a 96ꢀdeep well plate at room temperature (23 °C)
with a final assay volume of 1000 ꢁL/well, according to the following protocol: 800 ꢁL of binding
3
buffer were dispensed into each well of the compound plate. [ H]ꢀSpiperone stock was diluted in
3
binding buffer solution to obtain the 10x [ H]ꢀ Spiperone solution (0.8 nM). 100 ꢁL of 0.8 nM [3H]ꢀ
Spiperone solution were dispensed into each well of the compound plate. The competition reaction was
started by adding 100 ꢁL of hD2ꢀ Gα16ꢀCHO membrane suspension in binding buffer. The final
3
membrane concentration/well was 2 ꢁg and the final [ H]ꢀSpiperone concentration was 0.08 nM. The
plate was then incubated on a shaker at 23 °C for 120 min. The reaction was terminated by rapid
filtration through Unifilterꢀ96 GF/B filter plates preꢀsoaked for at least one hour in Polyethylenimine
(PEI) 0.5% (w/v) solution and washed with 1.0 mL of ice cold 0.9% NaCl before the filtration using a
Packard cell Harvester. The filter plate was washed 4 times with 1.0 mL iceꢀcold 0.9% NaCl and then
left to dry for at least one hour at 40 °C. The plate was sealed with a backꢀseal; 50 ꢁL of Microscintꢀ20
were added to each well and the plate was sealed with a topꢀseal. Bound radioactivity was measured
3
using a Microplate TopCount. Radioligand concentration was determined as follows: 100 ꢁL of [ H]ꢀ
Spiperone solution (5x) and 3 mL Filter Count were mixed in the total added vial and read in βꢀ Counter
TriCarb 2900.
Saturation binding experiments
Saturation binding experiments were performed similarly to the competition binding experiments,
3
with the following deviation: [ H]ꢀSpiperone concentrations were chosen from 0.011 to 3.0 nM in a
ACS Paragon Plus Environment