1-Aminoindanes as novel motif with potential atypical antipsychotic properties

Article abstract of DOI:10.1016/j.bmcl.2007.11.106

As part of an on-going effort to investigate the chemical space requirements for D₂/5-HT_2A receptor antagonists as atypical antipsychotics, new 1-aminoindanes were synthesized. The replacement of the heterocycle (oxindole) in ziprasi

Full text of DOI:10.1016/j.bmcl.2007.11.106

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 489–493
1-Aminoindanes as novel motif with potential atypical
antipsychotic properties
James M. Graham,* Linda L. Coughenour, Bridget M. Barr, David L. Rock
and Sham S. Nikam
Pfizer Global Research and Development, 2800 Plymouth Road, Ann Arbor, MI 48105, USA
Received 5 October 2007; revised 26 November 2007; accepted 27 November 2007
Available online 3 December 2007
Abstract—As part of an on-going effort to investigate the chemical space requirements for D2/5-HT2A receptor antagonists as atyp-
ical antipsychotics, new 1-aminoindanes were synthesized. The replacement of the heterocycle (oxindole) in ziprasidone with a
carbocycle (indane) was well tolerated and was found to retain binding affinities for dopamine D2, serotonin 5-HT2A, and serotonin
5-HT1A. Such compounds hold promise as a new chemical motif with atypical antipsychotic properties for the treatment of schizo-
phrenia and related disorders.
Ó 2007 Elsevier Ltd. All rights reserved.
Schizophrenia is a chronic, debilitating disorder afflict-
ing more than 24 million people worldwide1 and is gen-
potent 5-HT2A antagonism together with relatively
weaker dopamine antagonism are principal features that
differentiate the side-effect profile of atypical antipsy-
chotics like clozapine from the first generation of treat-
ments such as haloperidol.5 Although the newer atypical
antipsychotics, exemplified by clozapine and olanzapine,
have brought improvements in treatment of negative
symptomology, chronic treatment with these medicines
can still lead to substantial weight gain, blood dyscra-
sias, and some motor dysfunctions such as EPS.6 With
a high refractive treatment incidence and low compli-
ance due to tolerability, the search continues for new
atypical antipsychotics that have a better balance be-
tween efficacy and side-effect profile to treat this com-
plex disease.
erally
characterized
by
positive
symptoms
(hallucinations, delusions, etc.) and negative symptoms
(cognitive deficits, social withdrawal, suicidality, etc.).2
It is thought that excess dopamine in the brain underlies
the positive symptoms of schizophrenia and that block-
ade of dopamine receptors is a requisite property for
antipsychotic activity.3 Older standards of treatment
such as chlorpromazine and haloperidol are potent D2
receptor antagonists, however, these treatment regimens
have neglected negative symptoms, produced motor def-
icits (extra-pyramidal side effects (EPS)), and are impli-
cated in the development of tardive dyskinesia, a long-
term movement disorder.4
The second generation antipsychotics are referred to as
atypical antipsychotics due to their lower propensity to
elicit EPS and their moderate efficacy toward negative
symptoms. Such compounds have potent activity at sev-
eral serotonin (5-HT) receptor subtypes that are thought
to lead to their improved efficacy and reduced liability
for motor side effects. This rich pharmacological profile
has led to the D2/5-HT2A hypothesis.5 It is thought that
N
OH
O
Cl
N
S
Cl
N
F
F
Haloperidol
Chlorpromazine
N
N
N
N
Cl
Keywords: Atypical antipsychotic; D2 antagonism; 5HT antagonism;
hERG; Indane.
N
N
N
H
O
O
N
*
Corresponding author. Tel.: +1 734 649 8971; e-mail:
jamesgraha@gmail.com
Clozapine
Risperidone
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.106

490
J. M. Graham et al. / Bioorg. Med. Chem. Lett. 18 (2008) 489–493
In this paper, we describe our recent efforts to discover
novel templates for potential use as treatments of
schizophrenia expanding on the D2/5-HT2A hypothesis
and understanding how minor modifications in structure
can lead to significant changes in pharmacological
profile.7,8 In addition to the D2/5-HT2A receptor antag-
onism, we wanted to retain affinity for serotonin 5-HT1A
receptors, an additional efficacy driver for superior
treatment of cognitive and affective symptoms of the dis-
order. Clinical trials have shown that addition of tando-
spirone, a partial agonist at 5-HT1A receptors, to a first
generation antipsychotic produced an improvement in
cognitive performance in schizophrenics.9
cular liability, mainly orthostatic hypotension.
However there is evidence that this pharmacology may
add to the efficacy of antipsychotics by playing a sympa-
thetic role in attenuating dopamine levels in the brain.10
Therefore, the impact of the SAR studies on a1A was
monitored using radioligand receptor binding. Blockade
of the hERG (Ikr) K+ channel has been shown to pro-
duce drug-induced QTc prolongation, an activity that
has been associated with Torsade de point and cardiac
arrest.11 hERG channel affinity was assessed by mea-
surement of the ability to inhibit binding of dofetilide
(DOF), a class III anti-arrhythmic, to recombinant
hERG channels. Dofetilide is a potent, selective inhibi-
tor of the potassium (Ikr) channel mainly responsible
for cardiac repolarization12,13 and has been correlated
with hERG functional activity.13 The high throughput
DOF binding assay was a useful early discovery indica-
tor of hERG activity for this series of compounds.
An on-going challenge in developing molecules with this
structural and pharmacological profile is to maintain a
favorable profile around cardiovascular risk. We
achieved this by monitoring affinity for the hERG K+
channel and adrenergic a1A receptors. Potent a1A recep-
tor antagonism is an indication of potential cardiovas-
The overall strategy was to use the structure of ziprasi-
done as a starting point, and replace the heterocyclic
B-ring with a carbocycle having an exocyclic amide
(Fig. 1) hoping to achieve compounds with low a1A
and DOF affinity.
O
O
R
HN
NH
N
N
N
S
N
N
Cl
S
N
B-Ring
The compounds made in this program include unique 5-
and 6-linked 1-aminoindane scaffolds (Scheme 1). The
chemistry was amenable to rapid diversification from
aminoindanes 10 and 11. Acylation of commercially
available (+/À) 1-aminoindane with trifluoroacetic
anhydride proceeded smoothly under common condi-
tions. The trifluoro acetate functionality serves as a ro-
bust protecting group that stands up nicely to strongly
acidic conditions in subsequent steps. The trifluoroace-
tamide compound 2 was acylated under standard Fri-
edel–Crafts conditions in high yield to give a 55:45
Ziprasidone
Assay Ki, nM
D2
5
0.08
4
5-HT2A
5-HT1A
α1A
0.5
NA
DOF
Figure 1. Ziprasidone structure and binding affinities at relevant
receptors. Binding data for DOF are not available. Proposed structure
changes primarily in the B-ring.
O
O
CF3
O
HN
H
O
CF3
O
Cl
N
i
NH2
ii
+
NH
CF3
Cl
qnt.
90%
2
1
3
4
F3C
O
O
HN
iv. 85%
HN
iii
CF3
Cl
+
6
N
N
S
N
+
9
NH
85%
N
S
N
5
8
7
R
O
H2N
HN
v
vi
90%
+
11
N
N
N
N
S
S
N
qnt.
N
(+/-)
12a-n
10
H
R
N
N
N
O
S
N
(+/-)
13a-h
Scheme 1. Reagents and conditions: (i) TFAA, Et3N, THF; (ii) chloroacetyl chloride, AlCl3, CH2Cl2; (iii) Et3SiH, TFA; (iv) MeCN, K2CO3, lk
150 °C; (v) K2CO3, MeOH/H2O; (vi) acylating agent, Et3N, THF.

J. M. Graham et al. / Bioorg. Med. Chem. Lett. 18 (2008) 489–493
491
mixture of isomers that were easily separated by flash
chromatography into racemic regioisomers 3 and 4.
Independently, chloroketones 3 and 4 were reduced with
triethylsilane in TFA to give chloroethyl compounds 5
and 6, respectively, providing coupling substrates for
the benzisothiazole piperazine 7.
were seen between aryl and alkyl amide motifs. Targets
with aryl amides (12f–n) had up to 4- to 6-fold higher
affinity for D2 than the aliphatic amides 8, 9, and 12a–e.
These targets also show mixed 5-HT affinities with more
than half having 5-HT2A potencies of 1–5 nM giving rise
to a less favorable D2/5-HT2A ratio, which is considered
important to mitigating EPS. The aryl amides show a
trend toward lower affinities for 5-HT1A and a1A. Aryl
amides from this series are hampered by very potent
dofetilide binding with compounds 12i and 12m having
78 nM and 35 nM binding affinities, respectively.
The conditions for the coupling reaction had to be opti-
mized as under conventional conditions (reflux, 72 h)
the piperazine alkylates in a variable 25–40%. Micro-
wave conditions (MeCN, Et3N, 150 °C) gave signifi-
cantly improved and consistent yields (ꢀ85%) of
piperazinyl derivatives 8 and 9 along with shorter reac-
tion times (ꢀ15 min). Details of this procedure and
method development will be described in a subsequent
publication. Regioisomers 8 and 9 were efficiently
deprotected under mild conditions to unveil the free
amines 10 and 11, respectively, which were used as con-
venient substrates for rapid analog preparation. From
these common substrates, the appropriate acid chlorides
were employed to yield 6-linked targets 12a–n and 5-
linked targets 13a–g under standard conditions.
The alkyl amides 9 and 12a–e have excellent D2/5-HT2A
ratios, high affinities for 5-HT1A, and a good a1A profile
relative to D2, although DOF binding was generally
quite potent. Of all the 6-linked targets compounds 9
and 12a showed the best overall profile including re-
duced affinity for the dofetilide binding site with Ki val-
ues of 616 nM and 885 nM, respectively.
Table 2 shows the in vitro binding data for 5-substi-
tuted aminoindane amides 11 and 13a–g. Changes in
binding affinities at the D2 receptor in this series are
less subtle, with higher homologs and branched alkyl
amides (13b–e) showing a pronounced lack of affinity
for D2 receptors while they retain higher binding affin-
ities for 5-HT2A, 5-HT1A, and a1A. Binding at a1A is
particularly high with several targets showing pico-
molar affinity. This series also retain the high potency
at the dofetilide binding site also seen in the 6-linked
isomers 12a–l. Only compounds 13a, 13f, and 13g
from this series have reasonable overall binding pro-
files, although compounds 13f and 13g have potent
interactions with the dofetilide site. Of the 5-substi-
tuted aminoindane amides, 13a is distinguished as
the superior compound with excellent potencies at
The affinities of target compounds were evaluated in
in vitro binding assays using radiolabeled reference li-
gands [3H]N-methylspiperone for D2; [3H]ketanserin
for 5-HT2A; [3H]8-OH-DPAT for 5-HT1A; [3H]prazosin
for adrenergic a1A; and [3H]dofetilide for dofetilide
binding site in the hERG K+ channel.14
The in vitro binding data for 6-linked 1-aminoindane
amides, as well as the primary amine compound 10,
are detailed in Table 1. Overall, the binding affinities
across D2, 5-HT2A, 5-HT1A, and a1A were in the desired
range, however, dofetilide binding was high and ranged
in affinity from 35 nM to 885 nM. Subtle differences
Table 1. 6-Linked 1-aminoindane amides
R
HN
O
S
N
N
N
Compound
R
D2 Ki, nMa
5HT2A Ki, nMa
5HT1A Ki, nMa,b
a1A Ki, nMa
DOF Ki, nMa
8
10
CF3
c
10
21
22
34
27
21
67
22
15
5
0.35
0.09
0.01
0.01
0.45
0.30
0.75
4.75
3.00
0.10
2.45
0.95
0.04
0.95
0.55
0.08
4.90
2.0
1.92
2.72
1.34
3.03
3.96
2.22
3.61
10.91
13.54
5.64
10.26
3.31
5.27
10.33
1.38
1.88
101
616
885
164
196
35
12a
12b
12c
12d
12e
12f
12g
12h
12i
12j
12k
12l
12m
12n
Me
Et
7.5
5.0
n-Pr
c-Pr
5.5
12.0
12.0
41.8
28.1
56.0
66.7
4.5
i-Pr
Ph
278
272
194
166
78
4MePh
4ClPh
4FPh
6
3-Pyridyl
CH2Ph
(CH2)2Ph
2-Furyl
5-(1,2-Oxazoyl)
6
358
201
511
35
26
33
5
20.9
37.2
7.9
3
6.5
150
a Values are geometric means of at least three experiments.
b All compounds are partial agonists with 60–80% intrinsic functional activity.
c Free primary amine as seen in Scheme 1.

492
J. M. Graham et al. / Bioorg. Med. Chem. Lett. 18 (2008) 489–493
Table 2. 5-Substituted aminoindane amides
H
R
N
O
N
N
S
N
Compound
R
D2 Ki, nMa
5HT2A Ki, nMa
5HT1A Ki, nMa,b
a
1A, Ki, nMa
DOF, Ki, nMa
9
CF3
Me
8
15
0.15
0.33
0.25
0.50
0.45
0.65
1.65
0.68
1.0
1.1
2.0
2.0
2.0
1.0
7.8
7.7
0.30
3.18
143
2202
184
269
137
219
135
202
13a
13b
13c
13d
13e
13f
13g
Et
21% at 100 nM
21% at 100 nM
27% at 100 nM
17% at 100 nM
10
0.45
0.90
n-Pr
c-Pr
0.73
0.49
i-Pr
4FPh
3-Pyridyl
15.58
17.37
5
a Values are means of at least three experiments.
b All compounds are partial agonists with 60–80% intrinsic functional activity.
4. Martin, A. R. In Burger’s Medicinal Chemistry and Drug
Discovery, 5th ed.; Wolf, M. E., Ed.; Wiley-Interscience:
New York, 1997; Vol. 5, p. 195.
5. Meltzer, H. Y.; Matsubara, S.; Lee, J. J. Pharm. Exp.
Ther. 1989, 251, 238.
6. Cunningham Owens, D. G. Drugs 1996, 51, 895; Wirshing,
D. A.; Wirshing, W. C.; Kysar, L.; Berisford, M.;
Goldstein, D.; Pashdag, M. A.; Mintz, J.; Marder, S.
J. Clin. Psychiatry 1999, 60, 358.
7. Roth, B. D.; Sheffler, D. J.; Kroeze, W. K. Nature Rev.
2004, 3, 353.
8. Cunningham Owens, D. G. Drugs 1996, 51, 895; Wirshing,
D. A.; Wirshing, W. C.; Kysar, L.; Berisford, M.;
Goldstein, D.; Pashdag, M. A.; Mintz, J.; Marder, S.
J. Clin. Psychiatry 1999, 60, 358; Green, B. Curr. Med.
Res. Opin. 1999, 15, 79.
9. Meltzer, H. Y.; Li, Z.; Kaneda, Y.; Ichikawa, J. J. Prog.
Neuro-Psychopharm. Biol. Psychiatry 2003, 27, 1159;
Sumiyoshi, T.; Matsui, M.; Nohara, S.; Yamashita, I.;
Kurachi, M.; Sumiyoshi, C.; Jayathilake, K.; Meltzer, H.
Y. Am. J. Psychiatry 2001, 158, 1722.
10. Svensson, T. H. Prog. Neuro-Psychopharm. Biol. Psychi-
atry 2003, 27, 1145; Praxcik, M. T.; Perez, D. M.
J. Pharmacol. Exp. Ther. 2001, 298, 403.
11. Brown, A. M. Card. Saf. Noncard. Drugs 2005, 67;
Sanguinetti, M. C.; Tristani-Firouzi, M. Nature 2006,
440, 463.
D2, 5-HT2A, 5-HT1A, and a1A as well as having the
lowest binding affinity for DOF with micromolar
activity.
Interesting differences between the 5- and 6-substituted
1-aminoindane amides were clearly seen in D2 binding
with additional differences seen in the serotonin and
adrenergic receptors as well. Of particular note in both
series, the acetamide analogs stood out as having supe-
rior pharmacological profiles including a reduced affin-
ity for dofetilide binding.
In summary, a series of novel 5- and 6- substituted
aminoindanes were synthesized and evaluated for
in vitro affinity at D2 and 5-HT (2A and 1A subtype)
receptors, as well as adrenergic a1A and hERG K+ chan-
nel receptors. This series of compounds possess potent
activity at 5-HT2A providing ample separation from
D2 activity suggesting the potential for wide therapeutic
index between efficacy and EPS. Addition of potent
binding at 5-HT1A may provide an additional efficacy
driver and lead to better treatment of cognitive and
affective symptomology in schizophrenia. Although
both 5- and 6- substituted aminoindanes showed varying
pharmacologic profiles across these receptors, the
methyl amides 12a and 13a from each series stood out
as leading candidates for further development. Work is
in progress to separate and fully characterize the stereo-
isomers of these leads. Future SAR around the 1-amin-
oindane scaffold should focus on small amides or amide
isosteres and avoid higher homologs and branched alkyl
substrates. This SAR indicates that changing the B-ring
heterocycle to a carbocycle with an external amide is a
viable strategy toward identifying new chemical space
with atypical antipsychotic pharmacology.
12. Finlayson, K.; Pennington, A. J.; Kelly, J. S. Eur. J.
Pharm. 2001, 412, 203.
13. Ward, K. J.; Gill, J. S. Exp. Opin. Invest. Drugs 1997, 6,
1269; Greengrass, P. M.; Stewart, M.; Wood, C. M.
WO2003021271.
14. Standard filtration receptor binding techniques were
used to evaluate the affinity of compounds for
recombinant receptors. The dopamine D2 receptor
binding assay utilized cell membranes prepared from
CHO cells expressing human dopamine D2L receptors
radiolabeled with [3H]spiperone. The 5-HT2
receptor
A
binding assay utilized cell membranes prepared from
Swiss 3T3 cells expressing human 5-HT2A receptors
radiolabeled with [3H]ketanserin. The 5-HT1A receptor
binding assay utilized cloned human 5-HT1A receptor
radiolabeled with [3H]8-OH-DPAT. The adrenergic a1A
receptor binding assay utilized cloned rat alpha1a in
fibroblast cell membranes radiolabeled with [3H]prazo-
sin. The hERG potassium channel binding assay utilized
HEK-293 cells that contained the hERG channel radi-
olabeled with [3H]dofetilide. Membrane preparations
and radioligand were incubated in the presence, or
absence, of varying concentrations of the compound of
References and notes
1. World Health Organization. WHO Fact Sheet Number
265; The World Health Organization: Geneva, Switzer-
land, 2001.
2. Capuano, B.; Crosby, I. T.; Lloyd, E. J. Curr. Med. Chem.
2002, 9, 521.
3. Seeman, P.; Chau-Wong, M.; Tedesco, J.; Wong, K.
Nature 1976, 261, 717.

J. M. Graham et al. / Bioorg. Med. Chem. Lett. 18 (2008) 489–493
493
interest until steady state was reached. Nonspecific
binding was determined by addition of high concentra-
tions of haloperidol, unlabeled ketanserin, phentolamine,
WAY 100135, and unlabeled dofetilide for the D2L, 5-
HT2A, 5-HT1A, adrenergic a1A, and hERG potassium
channel assays, respectively. Unbound radiolabel was
separated from bound radiolabel by vacuum filtration.
The amount of bound radiolabel was determined by
liquid scintillation spectrophotometry. The relationship
between concentration and percent inhibition of specific
binding was fit using nonlinear regression to determine
IC50 concentration. IC50 concentrations were converted
to Ki affinity constants using the Cheng–Prusoff
equation.