Hit-to-lead investigation of a series of novel combined dopamine D 2 and muscarinic M1 receptor ligands with putative antipsychotic and pro-cognitive potential

Article abstract of DOI:10.1016/j.bmcl.2012.05.048

We describe the discovery of a series of compounds based on 1-{3-[4-(2-oxo-2,3-dihydro-benzoimidazol-1-yl)-piperidin-1-yl]-propyl}-3, 4-dihydro-1H-quinolin-2-one (3), showing combined D₂ receptor affinity and M₁ receptor agonism. Based on a strategy of controlling log P, we herein describe a hit-to-lead investigation with the aim of retaining the combined D₂/M₁ profile, while removing the propensity of the compounds to inhibit the hERG channel, as well as at obtaining acceptable pharmacokinetic properties. Although a SAR was evident for all four parameters in question, it was not possible to separate hERG channel inhibition and D ₂ receptor affinity by this effort; whilst it was feasible to obtain compounds with M₁ receptor agonism, acceptable clearance, and weak hERG inhibition.

Full text of DOI:10.1016/j.bmcl.2012.05.048

Bioorganic & Medicinal Chemistry Letters 22 (2012) 5134–5140
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Hit-to-lead investigation of a series of novel combined dopamine D2
and muscarinic M1 receptor ligands with putative antipsychotic
and pro-cognitive potential
Anette Graven Sams a, , Krestian Larsen a, Gitte Kobberøe Mikkelsen a, Morten Hentzer a,
⇑
Claus Tornby Christoffersen a, Klaus Gjervig Jensen b, Kristen Frederiksen a, Benny Bang-Andersen a
a Neuroscience Drug Discovery Denmark, H. Lundbeck A/S, 9 Ottiliavej, DK-2500 Copenhagen-Valby, Denmark
b Drug ADME Research, H. Lundbeck A/S, 9 Ottiliavej, DK-2500 Copenhagen-Valby, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 26 March 2012
Revised 9 May 2012
Accepted 10 May 2012
Available online 18 May 2012
We describe the discovery of a series of compounds based on 1-{3-[4-(2-oxo-2,3-dihydro-benzoimidazol-
1-yl)-piperidin-1-yl]-propyl}-3,4-dihydro-1H-quinolin-2-one (3), showing combined D2 receptor affinity
and M1 receptor agonism. Based on a strategy of controlling logP, we herein describe a hit-to-lead inves-
tigation with the aim of retaining the combined D2/M1 profile, while removing the propensity of the com-
pounds to inhibit the hERG channel, as well as at obtaining acceptable pharmacokinetic properties.
Although a SAR was evident for all four parameters in question, it was not possible to separate hERG
channel inhibition and D2 receptor affinity by this effort; whilst it was feasible to obtain compounds with
M1 receptor agonism, acceptable clearance, and weak hERG inhibition.
Keywords:
Schizophrenia
Cognition
D2
Ó 2012 Elsevier Ltd. All rights reserved.
M1
Schizophrenia is a debilitating psychiatric disorder, which is
characterized by the presence of three major symptom clusters.
These include positive symptoms, such as delusions, hallucinations
and disordered speech; negative symptoms, including social with-
drawal and anhedonia; and cognitive symptoms. The cognitive
symptoms are evident in patients even before the onset of the first
psychotic episode, and have been suggested to be the core symp-
toms of the disease.1 The cognitive impairment affect several cog-
nitive domains, particularly working memory and executive
function (the ability to plan and execute tasks).2 The cognitive
impairment associated with schizophrenia (CIAS) negatively im-
pacts the patients’ quality of life, and are untreated by current anti-
psychotic drugs.3 Treatment of CIAS is regarded as a major unmet
medical need.
It has been suggested that activation of muscarinic M1 receptors
could play a role in the treatment of CIAS.4–7 For example, musca-
rinic M1 receptor expression is decreased in cortical regions of
schizophrenic patients,8,9 and M1 receptor knock-out mice re-
vealed deficits in hippocampus and prefrontal cortex dependent
working and episodic memory tasks.10,11
We hypothesised that compounds with a combined pharmaco-
logical profile of D2 receptor antagonism and M1 receptor agonism
could have antipsychotic as well as a pro-cognitive potential.3 In-
deed, the major human metabolite of the antipsychotic drug cloza-
pine, N-desmethyl clozapine (NDMC), with a broad pharmacological
profile including D2 receptor antagonism and M1 receptor agonism,
was pursued as a putative antipsychotic with pro-cognitive poten-
tial. NDMC was progressed into phase 2 clinical trials, although it
failed to show antipsychotic efficacy as mono-therapy.12,13 Accord-
ingly, we compiled a focused library by a virtual in silico screen of
the corporate compound file, using a D2 receptor pharmacophore
model. Approximately 10,000 compounds were selected and subse-
quently screened in a D2 receptor binding assay.14 The same com-
pound collection was screened for functional agonism at the
human M1 receptor.15 From this screen, compounds 1 and 2, having
the desired combined D2 and M1 receptor profile, were identified
(Chart 1). Compound 1 (Oxiperomid, R4714) displayed high affinity
for the D2 receptor (Ki = 5.5 nM), for comparison, the antipsychotic
drug haloperidol (Chart 1) exerts a similar level of D2 receptor affin-
ity (Ki = 3.0 nM). Furthermore, it was shown that compound 1 was a
partial M1 receptor agonist with an EC50 value of 400 nM and
Emax = 78%, relative to an Emax of 100% for acetylcholine (Ach). The
EC50 of Ach in the assay is 1.1 nM. For reference, we have recently
demonstrated a pro-cognitive effect of the allosteric M1 receptor
agonist Lu AE51090 (4), which has an EC50 value of 61 nM and an
Emax = 83% in the same assay.15 Compared to 1, hit compound 2
⇑
Corresponding author at present address: LEO Pharma A/S, Industriparken 55,
2750 Ballerup, Denmark. Tel.: +45 41772172.
E-mail address: anette.sams@leo-pharma.com (A.G. Sams).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.05.048

A. G. Sams et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5134–5140
5135
N
O
N
O
O
N
N
N
N
N
N
H
O
N
H
N
H
O
, Oxiperomid
pKa = 7.4
2
3
1
pKa = 9.0
logP = 4.1
pKa= 7.6
logP = 3.4
logP = 2.8
O
N
F
O
O
O
N
Cl
N
N
H
OH
4 (Lu AE51090)
pKa = 7.9
Haloperidol
logP = 2.4
Chart 1.
showed a weaker D2 receptor affinity (Ki = 42 nM), but a similar pro-
file at the M1 receptor, being a partial agonist with an EC50 value of
160 nM and Emax = 70%.
was mainly cleared by hepatic metabolism in rat. The similar clear-
ance rates obtained from rat microsomes (predominantly phase I
enzymes) and hepatocytes (both phase I and phase II enzymes),
indicated predominant phase I clearance of 3. Thus, HLM incubation
was used for monitoring clearance during investigation of the ser-
ies.17 Furthermore, 3 was found to inhibit the hERG channel with
Incubation of 1 and 2 with human liver microsomes (HLM) sug-
gested that
1 had an acceptable intrinsic hepatic clearance
(hClint = 1.4 L/min), while 2 showed a high intrinsic hepatic clear-
ance (hClint = 9 L/min). For comparison, the human liver–blood
flow (LBF) is 1.4 L/min. Compounds 1 and 2 were also tested for
inhibition of Rb+ flux through the hERG channel,16 and were found
an IC50 value of 7.3
to 3, shows comparatively weaker hERG channel inhibition, with an
IC50 value of 20 M, and for a series of analogues of 4, we have
lM. Compound 4, which is structurally related
l
to inhibit the channel with IC50 values of 2.1
respectively.
l
M and 4.4
l
M,
shown that the pharmacokinetic properties can be modulated.15
These observations indicated to us that hERG inhibition and meta-
bolic stability might be modulated within a series of compounds
based on 3. Compound 3 was further profiled in functional hM2–
hM5 assays,15 and showed no agonism at any of these muscarinic
receptor subtypes. This was encouraging, as moderate selectivity
for hM1 receptors over the other muscarinic receptor subtypes,
especially hM2 and hM3 receptors, has been suggested to cause
the peripheral cholinergic side effects observed with orthosteric
M1 receptor agonists in clinical investigations.5,18,19 The structural
similarity of 3 and 4 suggested that 3 was an allosteric hM1 agonist
in analogy with 4.15 To investigate this in more detail, compound 3
was assayed for activity at the hM1 Y381A receptor mutant, in
which the essential tyrosine residue 381 located in the transmem-
We decided not to invest further in hit compound 2, mainly due
to the suggested poor pharmacokinetic properties. The similarities
between 1 and 2 inspired us to design compound 3, combining
structural features from both initial hits. Compound 3 retained
the combined D2/M1 receptor profile of its progenitors, displaying
a high D2 receptor affinity (Ki = 9.5 nM) as well as M1 receptor par-
tial agonism (EC50 = 110 nM, Emax = 53%). The intrinsic clearance in
HLM incubations was high (hCLint = 3.9 L/min). High in vivo clear-
ance (3 L/h/kg) was also observed the rat, in good agreement with
high intrinsic clearances measured in vitro using rat liver micro-
some incubations (Clint = 130 ml/min, rat LBF=20 ml/min) and rat
hepatocyte incubations (Clint = 160 ml/min). This suggested that 3
Table 1
Functional agonism of 3 at muscarinic receptorsa
Compound
hM1 Y381A
hM2
hM3
EC50 (nM)
hM4
hM5
EC50 (nM)
Emax (%)
EC50 (nM)
Emax (%)
Emax
%
EC50 (nM)
Emax (%)
EC50 (nM)
Emax (%)
ACh
3
27000
6.3
100
89
220
n.d.
100
0%
3.2
n.d.
100
0%
10
n.d.
100
8%
1.2
n.d.
100
0%
a
Data are means of two experiments Activity at 10 lM test concentration tested is given when an EC50 could not be established. n.d.: not determined.

5136
A. G. Sams et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5134–5140
Table 2
M1 receptor agonism, D2 receptor affinity, hCLint and hERG inhibition data for compounds 1–3, and 5–14, and reference compounds
R1
N
R 2
R1
R2
hM1
a
Compounds
EC50 nM
400
Emax (%)
hD2 Ki (nM)b
hClint (L/min)
hERG IC50
2.1
(lM)
D
MOE logPc
N
1
2
78
5.5
1.4
9
—
O
O
N
N
H
Indol-5-yl
160
110
70
53
42
4.4
7.3
—
—
O
N
O
3
5
9.5
3.9
2.8
N
N
O
N
H
N
12
69
2200
>50
À1.3
O
N
N
6
7
8
n.d.
n.d.
20
0%d
7%d
78
1100
55%
600
9.4
3.6
1.3
>50
>50
6.3
À1.2
À1.2
À0.8
N
O
N
O
N
N
O
O
N
9
73
58
16
55
5.1
2.3
6.6
9.3
À0.4
À0.3
O
O
10
n.d.
31%
N
O
O
N
N
O
11
12
1700
1200
30
40
1250
3.5
0.9
13
À1.2
À1.6
N
O
N
H
N
N
>9500
>50
O
N
H

A. G. Sams et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5134–5140
5137
Table 2 (continued)
a
Compounds
R1
R2
hM1
EC50 nM
n.d.
Emax (%)
hD2 Ki (nM)b
hClint (L/min)
hERG IC50
46
(lM)
D
MOE logPc
N
13
14
4%
1350
1.7
À1.6
O
N
N
H
N
n.d.
8%
650
2.6
22
À1.2
O
N
H
N
Lu AE51090 (4)
Haloperidol
ACh
—
—
—
—
—
—
61
n.d.
1.1
83
n.d.
100
>5000
3.0
n.d.
3.4
0.8
n.d.
20
1.1
n.d.
—
—
—
a
Data are means of two experiments. When an EC50 could not be established, activity is given as the percent activity at a compound test concentration of 10
indicated. n.d.: not determined.
l
M, as
b
Data are means of two experiments. Affinities are expressed as Ki values (nM) or as percent displacement of radioligands at a compound test concentration of 10
l
M, as
indicated.
c
D
n = 1.
MOElogP values are calculated relative to the corresponding MOElogP value for 3.
d
Table 3
M1 receptor agonism, D2 receptor affinity, hCLint and hERG data for compounds 15–22
5
X
N
6
R1
7
O
8
5
4
N
6
R2
N
7
N
H
O
R1
R2
X
hM1
hD2 Ki (nM)b
a
Compounds
EC50 nM
Emax (%)
hClint (L/min)
hERG IC50 (lM)
15
16
17
18
19
20
21
22
5-F
6-F
7-F
8-F
H
H
H
H
H
H
H
H
4-F
5-F
6-F
7-F
CH2
CH2
CH2
CH2
O
O
O
O
68
72
60
6.5
90
95
41
14
2.5
15
4.2
6.2
7.8
5.5
12
8.7
8.8
13
5.3
3.4
5.7
7.9
6.3
4.4
1.9
3.4
n.d.
n.d.
300
n.d.
44
24%
23%
59
22%
67
66
110
67
51
a
Data are means of two experiments. When an EC50 could not be established, activity is given as the percent activity at a compound test concentration of 10
lM, as
indicated. n.d.: not determined.
b
Data are means of two experiments. Affinities are expressed as Ki values (nM) or as percent displacement of radioligands at a compound test concentration of 10
lM, as
indicated.
brane domain 6 is mutated to alanine.20 The Y381 residue is critical
for orthosteric binding to the M1 receptor, and mutation to alanine
has been shown to decrease ACh binding affinity 30-fold and dras-
tically reduce its potency about 3,000-fold without affecting the
efficacy,13 making the hM1 Y381A receptor mutant a useful tool
to detect allosteric agonists. Compound 3 was a potent agonist at
the hM1 Y381A receptor mutant, while ACh showed greatly reduced
potency (>20,000-fold reduction), suggesting that 3 is activating the
hM1 receptor through interaction with different amino acid resi-
dues than Ach, that is an allosteric binding site. The data are sum-
marized in Table 1.
We decided to initiate a hit-to-lead investigation with the aim
of concluding whether 3 would be a workable hit, with particular
focus on metabolic stability and safety profile. Our strategy was
to investigate whether convergent structure–activity relationships
(SARs) would allow us to modulate the D2/M1 receptor ratio,
abolish hERG channel inhibition, and obtain acceptable clearance
within a compound series based on 3. Hence, in a small synthetic
program, we decided to independently monitor the SAR for D2
receptor affinity, M1 receptor agonism, hClint and hERG inhibition.
A body of examples exist in the literature of successful medicinal
chemistry programs aimed at overcoming hERG channel inhibition.
The strategy behind these approaches have been summarized as
(a) lowering of logP; (b) reducing basicity; (c) small structural
changes; or (d) formation of zwitterions.21,22 Also, in more general
terms, control of logP has been shown to be of great importance in
order to manage the pharmacokinetic properties as well as lower-
ing the general risks of toxicity.23–25 The logP and pKa values for 1–
4 are given in Chart 1. The pKa values of compounds 1, 3 and 4 were
7.4, 7.6 and 7.9, respectively, which is weakly basic, and further-
more appeared to be uncorrelated to the measured hERG channel
inhibition of the individual compounds. We therefore decided to

5138
A. G. Sams et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5134–5140
H
N
Y
z
R1
S
N
L
O
a
N
+
3, 9-22
T
O
U
N
H
V
X
23a:Y=Z=L=CH2, R1=H, X=Br
b:Y=O, Z=L=CH2, R1=H, X=Br
c:Y=Z=CH2, L=O, R1=H, X=Cl
d:Y=Z=L=CH2, R1=5-F, X=Br
e:Y=Z=L=CH2, R1=6-F, X=Cl
f:Y=Z=L=CH2, R1=7-F, X=Br
g:Y=Z=L=CH2, R1=8-F, X=Br
24a: S=T=U=V=CH
b: S=CF, T=U=V=CH
c: T=CF, S=U=V=CH
d: U=CF, S=T=V=CH
e: V=CF, S=T=U=CH
f: S=N, T=U=V=CH
g: T=N, S=U=V=CH
h: U=N, S=T=V=CH
i: V=N, S=T=U=CH
Scheme 1. Reagents and conditions: (a)DIPEA, toluene/DMF (1:5), 80 °C.
base our hit-to-lead investigation of 3 on lowering logP. This can
generally be obtained in two ways: (1) through introduction of
polarity in the compounds, such as by introduction of heteroatoms,
or (2) by removal of lipophilic groups. We opted for the first strat-
egy, and performed a heteroatom scan of compound 3 by synthesis
of compounds 5–14. The results are summarized in Table 2. The
experimental logP of compound 3 was 2.8, while the calculated
logP for 3 was 3.3 and 2.7 using the clogP26 or MOE logP27 algo-
rithms, respectively. Based on this, the MOE logP algorithm was se-
lected as the best method to estimate the logP values of
compounds 5–14. Generally, MOE logP calculations suggested
With compounds 3 and 9 as starting points, we decided to flu-
oro-scan the aromatic positions to identify any potential metabolic
hot-spots, as a complement to the SAR investigation based on low-
ering logP. The results are summarized in Table 3. None of the
compounds 15–18 and 19–22 showed any improvement of hClint
compared to 3 or 9, respectively. Introduction of fluorine into the
aromatic positions of 3 or 9 led to differential effects at the M1
or D2 receptors. At M1 receptors, introduction of fluorine in the
6- or 7-positions of the dihydroquinolinone ring system (16 and
17) led to complete loss of potency compared to 3. This was in con-
trast to previous results with analogues of 4.15 Also, a complete
loss of M1 receptor potency resulted from 4-fluoro substitution at
the benzimidazolone ring system (19) compared to 9. Fluorine in
other aromatic positions had little or no effect on M1 receptor po-
tency (15 and 18 vs. 3, as well as 20 and 21 vs 9). Fluorination in
the dihydroquinolinone ring system lowered the affinity of the
compounds to the D2 receptor (15, 17, and 18) compared to 3,
while the 6-fluoro dihydroquinolinone analogue 16 was equipo-
tent to 3. Fluorination of the benzimidazolone moiety led to differ-
ential effects at the D2 receptors, where a slightly lower affinity
compared to 9 resulted from 4-fluorination (14), while 6-fluorina-
tion increased the potency compared to 9 (compound 21). In con-
trast, fluorination in the 5- or 7-positions had no effect on the D2
receptor affinity (20 and 22).
A clear SAR was evident for all four monitored parameters in the
present series of compounds. A general and considerable lowering
of the inhibition of the hERG channel resulted from lowering logP
by introduction of nitrogen atoms in the aromatic positions of 3,
with the exception of 8-aza dihydroquinolinone derivative 8. Intro-
duction of oxygen (9 and 10) was calculated to lower the logP to a
less extent than the aza analogues, and inhibition of the hERG chan-
nel was largely unaffected by these permutations when compared
to 3. In contrast, improvement of intrinsic clearance was not
achieved by a general lowering of logP, but rather seemed to de-
pend on introduction of heteroatoms into specific positions, as
exemplified by the 8-aza dihydroquinolinone derivative 8, or the
5- or 6-aza benzimidazolone derivatives 12 and 13, respectively.
However, a fluoro-scan of the aromatic positions of 3 did not sug-
gest any particular metabolic hot-spots, including the correspond-
ing direct fluorinated analogues of the compounds with improved
intrinsic clearance in the aza-scan (18 vs 8, 20 vs 12, and 21 vs
13). The potency at the M1 receptor was also influenced by hetero-
atoms in specific positions, and could be either increased, lowered
or completely abolished; and the same was true for substituting
D
logP values to be in the order of 1–1.5log units for CH to N trans-
formations, and to be dependent on the regio-isomeric transforma-
tion (Table 2). For CH2 to O transformations, the calculated changes
in MOE logP relative to 3 were significantly smaller, and below
0.5log unit (Table 2). Aza-scanning the aromatic positions of the
dihydroquinolinone ring of 3 gave compounds with significantly
reduced D2 receptor affinity (5–8) when compared to 3. In contrast,
the effect on M1 receptor potency was highly dependent on the
specific position of the nitrogen atom, with the 5- and 8-aza ana-
logues 5 and 8 showing improved M1 receptor potency relative
to 3, as well as increased efficacy. In contrast, the 6- and 7-aza iso-
mers 6 and 7 were inactive at the M1 receptor. The hClint was af-
fected specifically by the position of the introduced nitrogen.
Thus, for the 6-aza isomer 6, the hClint was significantly worsened,
while an improvement in clearance was obtained with the 8-aza
isomer 8. The introduction of nitrogen in this region of the mole-
cules generally abolished inhibition of the hERG channel, except
for the 8-aza isomer 8, in which the degree of hERG inhibition
was unaltered compared to 3. Exchange of the dihydroquinolinone
moiety of 3 for the oxygen containing benzoxazinone (9), had little
effect on either the D2 or M1 receptor profile, hClint or hERG inhibi-
tion. We also prepared the hydroxamic ester derivative 10 in order
to polarize the propyl linker region of 3. Interestingly, this permu-
tation completely removed the M1 receptor potency and lowered
the D2 receptor affinity, while hClint and hERG channel inhibition
remained unchanged relative to 3. With 9 as a starting point, we
next performed an aza-scan of the aromatic positions of the benz-
imidazolone moiety, leading to compounds 11-14. A nitrogen atom
in any position in this part of the molecule drastically lowered the
activity at the primary targets, and also removed the hERG channel
inhibition, particularly for the 5-, 6-, and 7-aza isomers 12-14. The
effect on hClint followed the same pattern, with 12 and 13 display-
ing acceptable intrinsic clearances.

A. G. Sams et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5134–5140
5139
N
COOEt
N
CHO
a
b
a,b
N
O
O
N
H
N
O
NHBoc
NHBoc
OH
29
Cl
N
c
25
23c
N
H
O
Scheme 2. Reagents and conditions: (a)Na2WO4 Á 2 H2O, 30% H2O2, MeOH, 0–20 °C
(37% yield); (b) NaH, Br(CH2)2Cl, THF, 0–20 °C (29% yield).
27a
Scheme 5. Reagents and conditions: (a) NaH, (EtO)2P(O)CH2CO2Et, THF, 25 °C (73%
yield); (b) 5% Pd/C, H2, EtOH, 25 °C (67% yield); (c) TFA, DCM, 20 °C (74% yield).
NH2
NO2
H
N
F
F
a,b
N
O
while 24c–e28 and 24f–i29 were prepared as described in the cited
references. The 3-fluoro benzimidazolone 24b was prepared simi-
larly, as outlined in Scheme 3. Reaction of 1,2-difluoro-3-nitro-
benzene with 4-amino piperidine-1-ethylcarbamate, followed by
catalytic hydrogenation, afforded intermediate 26, and 24b was
obtained in 42% overall yield after carbonylation and ring closure
using carbonyl diimidazole, followed by deprotection under basic
conditions. Compounds 5–8 were synthesized as detailed in
Scheme 4. The intermediates 28a–d were obtained from 27a–d
by alkylation with 3-chloropropylacetate and subsequent saponifi-
cation, followed by oxidation of the intermediary alcohol with 2-
iodoxybenzoic acid (IBX). Reductive amination with 24a yielded
the final products. The intermediates 27b–d were commercially
available, while the preparation of 27a is described in Scheme 5.
Sequential Horner-Wadsworth-Emmons reaction between 2-for-
myl-pyridin-3-yl-tert-butyl carbamate and (diethoxy-phos-
phoryl)-acetic acid ethyl ester, and catalytic hydrogenation
afforded 29 in 49% overall yield, which spontaneously cyclized
upon deprotection with TFA to yield 27a (74% yield).
In conclusion, we designed compound 3, which had the desired
combined D2 and M1 receptor profile. However, the compound had
a high clearance and inhibited the hERG channel. Based on a strat-
egy of lowering logP, a hit-to-lead effort was undertaken to clarify
whether compounds with combined D2 and M1 receptor activity
and good pharmacokinetic and safety properties could be identi-
fied. Although a SAR was evident for all four parameters in ques-
tion, it was not convergent towards the aim of the investigation.
It was possible within this series of compounds to modulate the ra-
tio of activity at the D2 and M1 receptors, and it was feasible to ob-
tain compounds with M1 receptor agonism, acceptable clearance,
and weak hERG inhibition. However, a separation between hERG
channel inhibition and D2 receptor affinity was not possible by this
effort. Based on these results, we decided to stop any further
investigations of the series based on 3, and the project was
prompted to pursue other avenues.
F
O
26
H
N
O
c, d
N
NH
F
24b
Scheme 3. Reagents and conditions: (a) Na2CO3, KI, 4-amino-piperidine-1-carbox-
ylic acid ethyl ester, DMF, 50 °C; (b) 5% Pd/C, H2, MeOH, 20 °C; (c) CDI, MeCN, 40 °C;
(d) NaOH, EtOH, reflux. (42% overall yield).
G
K
G
a-c
d
5-8
K
Q
W
N
O
Q
W
N
H
O
OHC
27a:G=N, K=Q=W=CH
b:K=N, G=Q=W=CH
c:Q=N, G=K=W=CH
d:W=N, G=K=Q=CH
28a-d
Scheme 4. Reagents and conditions: (a) NaH, DMSO, 60 °C, then Cl(CH2)3OAc,
20 °C; (b) MeOH, KOtBu, 20 °C; (c) IBX, THF, reflux; (d) 24a, NaBH(OAc)3, THF, AcOH,
20 °C.
fluorine into specific positions of 3, yet in an uncorrelated manner.
In contrast, the D2 receptor affinity was generally lowered by low-
ering of the logP, but could be improved by a fluorine at the benz-
imidazolone 6-position (21).
Acknowledgement
The authors wish to thank Dr. Klaus Gundertofte for computa-
tional assistance.
Compounds 3 and 9–22 were prepared by a simple parallel syn-
thesis protocol as described in Scheme 1, by alkylation of piperidyl
benzimidazolones 24a–e, or piperidyl pyridoimidazolones 24f–i
with intermediates 23a–g in the presence of di-isopropylethyl
amine (DIPEA), in a mixed solvent of DMF /toluene (1:5) at 80 °C.
The intermediates 23a,b,d–f were prepared as described in ref.
15, while 23c was prepared outlined in Scheme 2. Tetrahydroquin-
oline was treated with sodium wolframate and hydrogenperoxide
to yield N-hydroxy dihydroquinolinone 25 in 37% yield, which was
alkylated with 1-bromo-3-chloropropane in a subsequent step to
give 23c in 29% yield. Compound 24a was commercially available,
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.05.048.
References and notes
1. Elvevåg, B.; Goldberg, T. E. Crit. Rev. Neurobiol. 2000, 14, 1.

5140
A. G. Sams et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5134–5140
2. Elliot, R.; McKenna, P. J.; Robbins, T. J.; Sahakian, B. I. Cogn. Neuropsych. 1998, 3,
45.
3. Arnt, J.; Bang-Andersen, B.; Dias, R.; Bøgesø, K. P. Drugs Future 2008, 33, 777.
4. Fischer, A. Neurotherapeutics 2008, 5, 433.
incubator at 37 °C and 5% CO2 for 2–3 h. The cells were washed once in 350 lL
Atomic Absorbtion Spectrometry (AAS) buffer (5.4 mM KCl, 137 mM NaCl,
8.1 mM Na2HPO4–2H2O, 1.5 mM KH2PO4, 0.49 mM MgCl2–6H2O, 0.90 CaCl2–
2H2O, pH 7.2–7.5), and the AAS-buffer was removed. To each well was added
5. Langmead, C. J.; Watson, J.; Reavill, R. Pharmacol. Ther. 2008, 117, 232.
6. Sellin, A. K.; Shad, M.; Tamminga, C. CNS Spect. 2008, 13, 985.
7. Sur, C.; Kinney, G. G. Curr. Neuropharmacol. 2005, 3, 63.
8. Dean, B.; McLeod, M.; Keriakous, D.; McKenzie, J.; Scarr, E. Mol. Psychiatry 2002,
7, 1083.
120
or 10 mM) diluted in the high KCl buffer (80.6 mM NaCl, 61.8 mM KCl) to final
concentrations in the range: 0.082, 0.205, 0.512,, 1.28, 3.2, 8, 20, 50 M. After
incubation for 30 min at room temperature, 100 L of the supernatant was
collected from each well. The cells were washed once in AAS buffer, and lysis
buffer (AAS buffer + 1% Triton X100) was added and allowed to stand for
lL of a solution of test compounds, prepared from DMSO stock solutions (2
l
l
9. Mancama, D.; Arrantz, M. J.; Landau, S.; Kerwin, R. Am J. Med. Gene. B.
Neuropsych. Gene. 2003, 119, 2.
15 min, whereafter 100 lL of lysate was collected from each well. For each
10. Anagnostaras, S. G.; Murpht, G. G.; Hamilton, S. E.; Mitchell, S. L.; Rahnama, N.
P.; Nathanson, N. M.; Silva, A. J. Nat. Neurosci. 2003, 6, 51.
11. Wess, J. Annu. Rev. Pharmacol. Toxicol. 2004, 44, 423.
12. Lameh, J.; Burstein, E. S.; Taylor, E.; Weiner, D. M.; Vanover, K. E.; Bonhaus, D.
W. Pharmacol. Ther. 2007, 115, 223.
13. Sur, C.; Mallorga, P. J.; Wittmann, M.; Jacobson, M. A.; Pascarella, D.; Williams,
J. B.; Brandish, P. E.; Pettibone, D. J.; Scolnick, E. M.; Conn, P. J. Proc. Natl. Acad.
Sci. U.S.A. 2003, 100, 13674.
well, concentrations of Rb+ in the supernatant sample and in the cell lysate
sample were determined by Atomic Absorption Spectrometry (Thermo
Elemental; Solaar; from Thermo Fisher Scientific), and the inhibition of the
Rb+-flux was calculated from the ratio between concentrations of Rb+ in the
supernatant and in the lysis buffer. Concentration–response curves were
obtained from 8 concentrations of test compounds, and IC50 values for Rb+ flux
inhibition were calculated using the Hill equation with variable slope
(equation: Sigmoidal dose–response (variable slope): Y = Bottom+(Top-
14. Human D2-receptor binding assay: The assay was performed as a SPA-based
competition-binding assay in a 50 mM Tris pH 7.4 assay buffer containing
120 mM NaCl, 5 mM KCl, 4 mM MgCl2, 1.5 mM CaCl2, 1 mM EDTA. 1.5 nM [3H]-
raclopride (Perkin Elmer, NET 975) was mixed with test compound before
Bottom/1+10^((LogIC50-X)⁄HillSlope)) where
X is the logarithm of the
concentration, is the response; this equation is identical to the ‘four
Y
parameter logistic eguation’.
17. For a description of DMPK assays and in vivo protocols, see: Sams, A. G.;
Mikkelsen, G. K.; Larsen, M.; Langgård, M.; Howells, M. E.; Schrøder, T. J.;
Brennum, L. T.; Torup, L.; Jørgensen, E. B.; Kreilgård, M.; Bang-Andersen, B. J.
Med. Chem. 2011, 54, 751.
addition of 20
with the human D2L receptor and 0.25 mg SPA beads (WGA RPNQ 0001,
Amersham) in a total volume of 80 L. The assay plates were incubated for
lg of membranes purified from CHO cells stably transfected
l
60 min at room temperature with agitation, and subsequently counted in a
scintillation counter (TriLux, Wallac). The total binding, which comprised app
15% of added radioligand, was defined using assay buffer whereas the non-
18. Fischer, A. CNS Drugs 1999, 12, 197.
19. Heinrich, J. N.; Butera, J. A.; Carrick, T.; Kramer, A.; Kowal, D.; Lock, T.; Marquis,
K. L.; Paush, M. H.; Popiolek, M.; Sun, S.-C.; Tseng, E.; Uveges, A. J.; Mayer, S. C.
Eur. J. Pharmacol. 2009, 605, 53.
specific binding was defined in the presence of 10 lM haloperidol. The non-
specific binding constituted app 10% of the total binding. Data points are
expressed in percent of the specific binding of [3H]-raclopride and the IC50
values are determined by non-linear regression analysis using a sigmoidal
variable slope curve fitting. The dissociation constant (Ki) were calculated from
the Cheng-Prusoff equation (Ki = IC50/(1+(L/KD)), where the concentration of
20. Spalding, T. A.; Trotter, C.; Skjærbæk, N.; Messier, T. L.; Currier, E. A.; Burstein,
E. S.; Li, D.; Hacksell, U.; Brann, M. R. Mol. Pharmacol. 2002, 61, 1297.
21. Waring, M. J.; Johnstone, C. Bioorg. Med. Chem. Lett. 2007, 17, 1759.
22. Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. J. Med. Chem. 2006, 49, 5029.
23. Gleeson, M. P. J. Med. Chem. 2008, 51, 817.
24. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug. Disc. 2007, 6, 881.
25. Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.; DeCrescenzo, G. A.; Devraj, R. V.;
Ellsworth, E.; Fobian, Y. M.; Gibbs, M. E.; Gilles, R. W.; Greene, N.; Huang, E.;
Krieger-Burke, T.; Loesel, J.; Whiteley, L.; Zhang, Y. Bioorg. Med. Chem. Lett.
2008, 18, 4872.
26. Daylight, version 4.8.
27. MOE version 2009.10.
28. Gustin, D. J.; Sehon, C. A.; Wei, J.; Cai, H.; Meduna, S. P.; Khatuya, H.; Sun, S.; Gu,
Y.; Jiang, W.; Thurmond, R. L.; Karlsson, L.; Edwards, J. P. Bioorg. Med. Chem.
Lett. 2005, 15, 1687.
free radioligand
L
is approximated to the concentration of added [3H]-
raclopride in the assay. The KD of [3H]-raclopride was determined to 1.5 nM
from two independent saturation assays each performed with triplicate
determinations.
15. Sams, A. G.; Hentzer, M.; Mikkelsen, G. K.; Larsen, K.; Bundgaard, C.; Plath, C.;
Christoffersen, C. T.; Bang-Andersen, B. J. Med. Chem. 2010, 53, 6386.
16. Inhibition of hERG channel Rb+ flux: CHO cells stably expressing the hERG
channels were plated out in 96 well plates 2 days before the experiment with a
cell density of 250.000 cells/ml and 100
testing, the media was removed and the cells were added 100
l
L volume per well. At the day of
L RbCl buffer
l
(5.4 mM RbCl, 137 mM NaCl, 8.1 mM Na2HPO4–2H2O, 1.5 mM KH2PO4, 0.49
MgCl2–6 H2O, 0.90 mM CaCl2, pH 7.2–7.5), and placed in an IGO-Jouan cell
29. Burgey, C. S., Stump, C. A., Williams, T. M. WO 2005/013894 A2.