The identification of a selective dopamine D2 partial agonist, D3 antagonist displaying high levels of brain exposure

Article abstract of DOI:10.1016/j.bmcl.2010.01.090

The identification of a highly selective D₂ partial agonist, D₃ antagonist tool molecule which demonstrates high levels of brain exposure and selectivity against an extensive range of dopamine, serotonin, adrenergic, histamine, and muscarinic receptors is described.

Full text of DOI:10.1016/j.bmcl.2010.01.090

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2013–2016
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The identification of a selective dopamine D2 partial agonist, D3 antagonist
displaying high levels of brain exposure
Ian P. Holmes a, , Richard J. Blunt a, Olivier E. Lorthioir a, Stephen M. Blowers a, Andy Gribble a,
Andrew H. Payne a, Ian G. Stansfield a, Martyn Wood a, Patrick M. Woollard a, Charlie Reavill a,
Claire M. Howes a, Fabrizio Micheli b, Romano Di Fabio b, Daniele Donati b, Silvia Terreni b,
Dieter Hamprecht b, Luca Arista b, Angela Worby c, Steve P. Watson a,
*
a GlaxoSmithKline, Neurosciences Centre of Excellence for Drug Discovery, New Frontiers Science Park, Harlow, UK
b Medicine Research Centre, Verona, Italy
c Molecular Discovery Research, New Frontiers Science Park, Harlow, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The identification of a highly selective D2 partial agonist, D3 antagonist tool molecule which demon-
strates high levels of brain exposure and selectivity against an extensive range of dopamine, serotonin,
adrenergic, histamine, and muscarinic receptors is described.
Received 7 December 2009
Revised 13 January 2010
Accepted 14 January 2010
Available online 25 January 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
DA D2
DA D3
Dopamine antagonist
Dopamine partial agonist
Aripiprazole
A major area of focus in the development of clinically effective
antipsychotics has been the study of compounds which display
high affinity binding, and low intrinsic activity (IA) partial agonism
at the dopamine (DA) D2 receptor.1
activity at 5-HT2A and 5-HT2C may contribute towards the im-
proved side effect profile of aripiprazole over high potency DA D2
antagonists.5
Herein, we report the identification of a highly selective DA D2
partial agonist, DA D3 antagonist tool molecule 3 (Fig. 2), suitable
for use in in vivo studies, to allow delineation of the dopamine as-
pects of the pharmacology of aripiprazole from activities at other
receptors.
Studies towards the identification of a DA D3 selective PET li-
gand6,7 also identified DA D2 partial agonist, DA D3 antagonist
ligands 4 and 5 (racemic) (Fig. 3, Table 1).
Taking 4 and 5 as starting points we explored the imidazolidi-
none template structure–activity relationships SAR) via synthesis
of structurally related compounds. In the first instance efforts were
focused on hexahydro azepine 4 where studies were undertaken to
vary the 3-Cl phenyl fragment, the imidazolidinone fragment and
the alkyl chain length.
Aryl and heteroaryl imidazolidinone fragments were prepared
using the procedure of Vidaluc, Imbert, and co-workers.8 Alkyl-
ation with 2-(hexamethyleneimino) ethyl chloride hydrochloride
using sodium hydride in DMF gave compounds 6–18 (Fig. 4).
Initial profiling of compounds 6–18 (Fig. 4, Table 2) revealed
that modest changes around the aryl fragment in this series had
a significant impact on the mode (agonist vs antagonist) of
compound action, affinity and selectivity. In fact, substitution in
Molecules such as aripiprazole 1 (Fig. 1) exhibit partial agonism
at the DA D2 receptor (IA = 0.31a), a profile which is postulated to
prevent effective blockade of DA D2 function from rising above
70% even when receptor occupancies approaching 100% are ob-
tained. This level of blockade falls within the anticipated therapeu-
tic window of greater than 65% required for clinical efficacy, and
below 80%, above which adverse events occur. Clinical investiga-
tions support this hypothesis with only minimal clinically limiting
side effects observed in patients receiving aripiprazole compared
with a high adverse event rate in patients receiving a comparably
efficacious dose of the potent DA D2 antagonist haloperidol 2.3
Whilst this interpretation of the data is attractive, in reality ari-
piprazole displays activities at a number of other receptors includ-
ing dopamine DA D3, serotonin 5-HT1A, serotonin 5-HT2A, serotonin
5-HT2C, histamine H1, and adrenergic
a
1 subtypes.2–4 It has been
suggested that the partial agonist activity at 5-HT1A and antagonist
* Corresponding author. Tel.: +44 1279 875004.
E-mail address: steve.p.watson@gsk.com (S.P. Watson).
Cancer Therapeutics CRC, Monash Institute of Pharmaceutical Sciences, 381 Royal
Parade, Parkville, VIC 3052, Australia.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.090

2014
I. P. Holmes et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2013–2016
F
H
O
N
O
Cl
N
N
Cl
N
O
OH
Cl
1
2
Figure 1. Aripiprazole 1 and haloperidol 2.
O
Table 3). In all such derivatives prepared a reduction in affinity
for both the DA D2 and D3 receptors was observed.
H
N
N
N
Variation in the length of the alkyl chain in the series was also
found to negatively impact affinity for both the DA D2 and D3
receptors. The C3 alkyl derivative 21 was prepared via alkylation
of 1-(3-chlorophenyl)-2-imidazolidinone8 with (3-bromoprop-
oxy)-t-butyldimethylsilane using sodium hydride in DMF. Subse-
quent treatment with tetra-n-butyl ammonium fluoride gave the
corresponding alcohol which was reacted with methanesulfonyl
chloride to give the corresponding mesylate. Reaction of the mes-
ylate with hexahydro azepine in the presence of triethylamine and
potassium iodide gave 21.
3
Figure 2. Selective DA D2 partial agonist, DA D3 antagonist tool molecule 3.
Cl
Cl
O
O
N
N
N
N
A substantial reduction in affinity for both the DA D2 and D3
receptors was observed for the three carbon linked compound 21
of this series with only weak antagonist activity at both receptors
(Fig. 6, Table 4).
N
N
4
5
Compounds obtained in the exploration around 4 (compounds
6–21) and which displayed high levels of DA D2 partial agonism,
were assessed to establish their broader selectivity and in vitro
pharmacokinetic profile. Unfortunately, significant off target activ-
ity at a range of aminergic receptors was identified including sero-
tonin 5-HT1A, serotonin 5-HT2A, and the histamine H3 receptor. The
intrinsic clearance in liver microsomes (Cli) was generally poor
across the series. With this data in hand it was decided to refocus
attention onto derivatives of 5 as the parent compound had dis-
played a superior selectivity profile to 4.
Figure 3. DA D2 partial agonist starting points.
Table 1
Profiling results for initial hits
10
10
10
Entry
D2 pEC50
D2 (IA)
D2 pKi
D3 pEC50
D3 pKi
4
5
9.0
8.6
0.75
0.48
8.0
7.8
<5.5
<5.5
8.3
8.4
IA aripiprazole1b = 0.60.
In the first instance des-methyl analogues of 5 were prepared in
which the position of the nitrogen within the piperidine ring was
varied. These compounds were prepared by reaction of 1-(3-chlo-
rophenyl)-2-imidazolidinone8 with the corresponding bromo-
methyl pyridine in the presence of sodium hydride in DMF.
Reduction of the pyridine ring using Adams catalyst in ethanol con-
taining acetic acid gave piperidines 22 and 23 (as racemates) and
24 (Fig. 7).
Of the three positional isomers prepared only 22 showed activ-
ity at the DA D2 and D3 receptors. Pleasingly, the level of intrinsic
activity observed for DA D2 partial agonism for 22 was comparable
to that of aripiprazole (Fig. 7, Table 5).
O
R
N
N
N
Figure 4. SAR for alternate aryl substitution and replacements.
the 2-position (compound 8) resulted in a complete loss of dopa-
mine affinity, whereas substitution in the 3-position was generally
well tolerated resulting in several compounds showing a high level
of affinity for both the DA D2 and D3 receptors. It was notable that
compound mode of action could be controlled in this series by the
choice of 3-substituent with either an antagonist profile at DA D2
(compounds 11 and 12) or partial agonist profile with varying de-
grees of IA (compounds 4, 9, 13, 14 and 17). Interestingly, 3,5-di-
substitution with chlorine, which inferred partial agonism when
only 3-substituted (compound 4), resulted in an antagonist profile
(compound 7). Substitution in the 4-position in this series was less
well tolerated and all examples prepared showed a drop in affinity
compared with the corresponding 3-substituted derivatives. Addi-
tion of a 4-fluorine substituent to the potent 3-chloro partial ago-
nist template also resulted in a drop in affinity as well as a decrease
in the IA observed at DA D2 (compound 16).
Exploration of the SAR around the aryl fragment was then
undertaken for the 2-methylene piperidine system.
As reported in Table 6, 3-Cl phenyl 22, phenyl 3 and 3-pyridyl
26 all displayed DA D2 partial agonism and the 3-CF3 phenyl 25
showed an antagonist profile. Notably, the 3-pyridyl fragment in
this series displayed significantly reduced affinity at both DA D2
and D3 compared with the corresponding hexahydro azepine com-
pound 17; this was rationalized by the highly polar nature of 26
(Fig. 8).
Both enantiomers of 22 displayed DA D2 partial agonism with a
similar IA to aripiprazole, though one enantiomer showed 10-fold
greater affinity for the DA D2 and D3 receptors. As both enantio-
mers also displayed similar in vitro metabolic stability it was
decided to progress compounds 3 and 22 as racemic mixtures.
A number of molecules were prepared by introducing a struc-
turally different moiety in place of imidazolidinone, including imi-
dazol-2-one 19 and 1,2,5-thiadiazolidine 1,1-dioxide 20 (Fig. 5,

I. P. Holmes et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2013–2016
2015
Table 2
Profiling results for aryl substitution SAR
10
10
10
Entry
R
D2 pEC50
D2 (IA)
D2 pKi
D3 pEC50
D3 pKi
4
6
7
8
9
10
11
12
13
14
15
16
17
18
3-Cl Phenyl
4-Cl Phenyl
3,5-Di-Cl phenyl
2-Me Phenyl
3-Me Phenyl
4-Me Phenyl
3-OMe Phenyl
3-CF3 Phenyl
Phenyl
3-F Phenyl
4-F Phenyl
4-F, 3-Cl Phenyl
3-Pyridyl
4-Pyridyl
9.0
7.3
0.75
0.34
—
—
0.83
—
8.0
7.0
9.2
<6.2
7.6
7.3
7.4
8.8
8.1
8.6
7.2
7.8
7.0
6.2
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
<5.5
8.3
7.1
8.8
<6.2
7.1
6.9
7.3
9.0
7.2
7.8
6.6
7.6
7.1
<6.2
<5.1
<5.1
8.0
<5.1
<5.1
<5.1
9.2
9.1
8.0
8.0
8.1
—
—
0.88
0.83
0.64
0.48
0.69
—
<5.5
IA aripiprazole1b = 0.60.
Table 5
Cl
Cl
Profiling results for piperidine nitrogen positional isomers
O
O
N
O
10
10
10
S
Entry
D2 pEC50
D2 (IA)
D2 pKi
D3 pEC50
D3 fpKi
N
N
N
N
N
22
23
24
9.2
<5.1
<5.1
0.60
—
—
7.1
<6.2
<6.2
<5.5
<5.5
<5.5
8.4
<6.2
<6.2
19
20
IA aripiprazole1b = 0.60.
Figure 5. Sample imidazolidinone replacement derivatives.
Table 3
Profiling results for imidazolidinone replacement derivatives
Table 6
Profiling results for aryl substitution SAR
10
10
10
10
Entry
D2 pEC50
D2 (IA)
D2 pKi
D3 pEC50
D3 pKi
Entry
R
D2
D210 (IA) D210 pKi D3 pEC50 D3 fpKi
pEC50
19
20
6.9
6.9
0.73
0.34
6.4
6.7
<5.5
<5.5
<6.2
6.5
22
25
3-Cl Phenyl
3-CF3
Phenyl
9.2
<5.1
0.60
—
7.1
7.7
<5.5
<5.5
8.4
9.6
IA aripiprazole1b = 0.60.
3
26
Phenyl
3-Pyridyl
8.3
6.3
0.76
0.60
7.1
<6.2
<5.5
<5.5
7.6
<6.2
IA aripiprazole1b = 0.60.
Cl
Cl
O
O
N
N
O
N
N
N
N
H
R
N
N
N
4
21
Figure 8. SAR for alternate aryl substitution and replacement.
Figure 6. Variation in alkyl chain length.
Cl
O
Table 4
Profiling results for imidazolidinone replacement derivatives
O
H
N
N
N
H
10
10
10
N
N
N
Entry
D2 pEC50
D2 (IA)
D2 pKi
D3 pEC50
D3 pKi
4
21
9.0
<5.1
0.75
—
8.0
6.2
<5.5
<5.5
8.3
6.9
IA aripiprazole1b = 0.60.
3
22
Figure 9. Compounds selected for physiochemical and in vitro pharmacokinetic
profiling.
Cl
O
22 X=NH, Y=CH2, Z=CH2
23 X=CH2, Y=NH, Z=CH2
24 X=CH2, Y=CH2, Z=NH
Compounds 3 and 22 were screened to establish their physio-
chemical and in vitro pharmacokinetic profile (Fig. 9).
N
N
X
Y
Metabolic stability in vitro for both compounds was comparable
(Table 7) However, compound 3 showed a superior P450 inhibition
Z
profile to 22. viz compound 22 CYP2D6 IC50 = 1.0 lM as opposed to
Figure 7. Piperidine positional isomers.

2016
I. P. Holmes et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2013–2016
Table 7
Table 9
Physiochemical and in vitro pharmacokinetic data for 3 and 22
In vivo DMPK profile of 3 after sc administration
Entry Solubility
g/mL
log D pH Cli rat mL/
Cli human mL/ 2D6 IC50
Dose mg/kg
3.0
Br:Bl
5.5
Cmax (Br) ng/g
Cmax (Bl) ng/mL
Tmax (h)
l
7.4
min/g
min/g
(lM)
667
142
0.5
3
22
106
148
0.73
1.24
22.9
23.7
4.5
3.4
>10
1.0
Table 10
Protein and tissue binding profile of 3
Table 8
Sample selectivity data for 3 and 22
HSA (%)
41.8
Plasma rat (%)
21.5
BTB rat (%)
77.7
Entry 5-HT1A
pEC50
5-HT1D
pEC50
5-HT2A
fpKi
a
fpKi
1B
H1
fpKi
hERG
pIC50
3
22
<5.5
6.1
<5.5
6.3
<5.6
<5.8
6.5
<6.6
<5.7
<5.6
4.8
5.1
mine, and muscarinic receptors. Moreover on administration sc 3
achieves high levels of brain exposure with a high free fraction. De-
tails of the in vivo profiling of 3 will be published in due course.
greater than 10
potencies greater than 10
l
M for 3, both compounds showed inhibition
l
M at all other isoforms tested.9
Acknowledgments
Abir Khazragi.
Initial selectivity profiling against a panel of receptors and
channels, chosen to allow a comparison with aripiprazole, revealed
3 displayed a superior profile to 22 (Table 8).
As such, 3 was cross screened against an extensive panel of
aminergic receptors (including dopamine, serotonin, adrenergic,
histamine, and muscarinic receptors9) and liability targets which
revealed only modest off-target activity at DA D1 fpKi = 5.3, a1B
adrenergic fpKi = 6.5 and, hERG pIC50 = 4.8. Notably, 3 was com-
pletely selective against the DA D4 receptor (agonist and antagonist
assays). More extensive screening by CEREP against a diverse range
of targets revealed no further liabilities.
In light of these results compound 3 was selected for in vivo
testing in the rat. As the rat Cli data had suggested, extensive first
pass metabolism resulted in low exposure of 3 via the oral route,
thus the compound was formulated in labrasol for subcutaneous
(sc) administration, from which good brain exposure was achieved
(Table 9).
The DMPK profile of 3 showed high brain penetration (brain to
blood ratio of 5.5) with high concentrations in the brain (667 ng/g)
after a single sc administration at 3 mg/kg. The compound exhib-
ited a Tmax of 0.5 h making it suitable for use in behavioural mod-
els, moreover, human serum albumin binding (HAS) was moderate
the rat brain tissue binding (BTB) was comparatively low leading to
a fraction unbound of 22.3% (Table 10).
In summary, we have prepared a DA D2 partial agonist, DA D3
antagonist tool molecule which shows a high degree of selectivity
over an extensive range of dopamine, serotonin, adrenergic, hista-
References and notes
1. Ohlsen, R. I.; Pilowsky, L. S. J. Psycopharmacol. 2005, 19, 408. (a) Typical IA value
in native tissue preparations. (b) In the recombinant assay used for SAR studies
(Ref. 10) Aripiprazole displayed an IA of 0.6, against which other compounds
were benchmarked.
2. Burris, K. D.; Molski, T. F.; Xu, C.; Ryan, E.; Tottori, K.; Yocca, F. D.; Molinoff, P. B.
J. Pharmacol. Exp. Ther. 2002, 302, 381.
3. Goodnick, P. J.; Jerry, J. M. Expert Opin. Pharmacother. 2002, 3, 1773.
4. Wood, M.; Reavill, C. Expert Opin. Investig. Drugs 2007, 16, 771.
5. Jordan, S.; Koprivica, V.; Chen, R.; Tottori, K.; Kikuchi, T.; Altar, C. A. Eur. J.
Pharm. 2002, 441, 137.
6. Holmes, I. P.; Micheli, F.; Gaines, S.; Lorthioir, O. E.; Watson, S. P.; Di Fabio, R.;
Gentile, G.; Heidbreder, C.; Savoia, C.; Worby, A. Bioorg. Med. Chem. Lett. 2009,
19, 4799.
7. Micheli, F.; Holmes, I. P.; Arista, L.; Bonanomi, G.; Braggio, S.; Di Fabio, R.;
Donati, D.; Gentile, G.; Hamprecht, D.; Heidbreder, C.; Savoia, C.; Terreni, S.;
Manolo, C.; Worby, A. Bioorg. Med. Chem. Lett. 2009, 19, 4011.
8. Mayer, P.; Brunel, P.; Chaplain, C.; Piedecoq, C.; Calmel, F.; Schambel, P.;
Chopin, P.; Wurch, T.; Pauwels, P. J.; Marien, M.; Vidaluc, J.-L.; Imbert, T. J. Med.
Chem. 2000, 43, 3653.
9. Aminergic receptor profiles, In vitro metabolic stability (Cli) and CYPEX
bactosome P450 inhibition profiles and protein binding were generated using
the procedures described in: Micheli, F. et al J. Med. Chem. 2007, 50, 5076.
10. The functional activity, pEC50 and pKi values, for the human D2 and D3
receptors were determined using stably transfected CHO cell lines and a 384-
well GTPg35S assay with Scintillation Proximity Affinity Detection (SPA).
Quinperol was used as the agonist. All values quoted are a mean generated
from a minimum of two determinations.