Design, synthesis, and discovery of 5-piperazinyl-1,2,6,7-tetrahydro-5H-azepino[3,2,1-hi]indol-4-one derivatives: A novel series of mixed dopamine D2/D4 receptor antagonist

Article abstract of DOI:10.1016/S0960-894X(02)01056-9

5-Piperazinyl-1,2,6,7-tetrahydro-5H-azepino[3,2,1-hi]indol-4-one derivatives were designed, synthesized, and identified as a new series of mixed dopamine D₂/D₄ receptor antagonists. This series featured a rigid tricyclic ring system

Full text of DOI:10.1016/S0960-894X(02)01056-9

Bioorganic & MedicinalChemistry Letters 13 (2003) 701–704
Design, Synthesis, and Discovery of 5-Piperazinyl-1,2,6,7-
tetrahydro-5H-azepino[3,2,1-hi]indol-4-one Derivatives: A Novel
Series of Mixed Dopamine D2/D4 Receptor Antagonist
He Zhao,* Xiaoyan Zhang, Kevin Hodgetts, Andrew Thurkauf,
Jack Hammer, Jayaraman Chandrasekhar, Andrzej Kieltyka, Robbin Brodbeck,
Stanislaw Rachwal, Renee Primus and Charles Manly
Neurogen Corporation, 35 Northeast Industrial Road, Branford, CT 06405, USA
Received 13 September 2002; accepted 8 November 2002
Abstract—5-Piperazinyl-1,2,6,7-tetrahydro-5H-azepino[3,2,1-hi]indol-4-one derivatives were designed, synthesized, and identified as
a new series of mixed dopamine D2/D4 receptor antagonists. This series featured a rigid tricyclic ring system as an important
pharmacophore core structure for high binding affinity. Molecular modeling studies are also described.
# 2003 Elsevier Science Ltd. All rights reserved.
Dopamine has been implicated in the pathophysiology
of schizophrenia for decades. The traditionalanti-
psychotic agents provide very good correlation between
their clinical efficacies and binding affinities for dopa-
mine D2 receptor. These ‘typical’ antipsychotic agents
are used for treatment of positive symptoms of schizo-
phrenia, but their use is limited by disabling side effects
such as extrapyramidalsyndrome (EPS), tardive dyski-
D4 (<10 nM) and moderate D2 (<200 nM) affinities
which maintained a similar binding ratio to that of clo-
zapine, and the postulate has been supported by our
recent studies.6,7 A secondary criteria of our search
required lower binding affinity to a1 (>1000 nM) in
order to avert undesirable cardiovascular effects.
1
nesia, and hormonalside effects. On the other hand,
Design and Molecular Modeling
the ‘atypical’ antipsychotic agent clozapine has several
clinical advantages over classical antipsychotic agents.
This drug displays not only high effects in positive and
negative symptoms without producing side effects, but
also prevents psychosis in some patients who were either
refractory or intolerant to the effects of classic neuro-
leptics.2,3 The higher affinity of clozapine for D4 over D2
receptors (about 10-fold) sparked research efforts in the
D4 receptor as a potentialtarget for antipsychotic ther-
apy.4 Several laboratories have examined their highly
selective dopamine D4 antagonists in clinical trials, but,
to date there is no positive efficacy result for these
agents.5 Therefore, we hypothesized that the unique
profile of clozapine may be a result of a particular ratio
of D4 and D2 receptor affinities. Thus, we set out to
identify mixed D2/D4 receptor antagonists having high
In a previous paper,8 we identified a novelseries of
benzofused d-lactam piperazine mixed D2/D4 receptor
antagonists which were discovered through the sys-
tematic transformation of lead compound 2-[-4-(4-chloro
-benzyl)-piperazin-1-yl]-1-(2,3-dihydro-indol-1-yl)-etha-
none (1). A good example from this d-lactam series is
3-[4-(4-chloro-benzyl)-piperazin-1-yl]-1-ethyl-3,4-dihydro-
1H-quinolin-2-one (2) which showed high affinity for
both D2 (21 nM) and D4 (4 nM) receptors in a ratio as
that of clozapine. However, further studies of structure–
activity relationships indicated that seven-member ring
lactam containing compound 3-[4-(4-chloro - benzyl) -
piperazin-1-yl]-1-ethyl-1,3,4,5-tetrahydro-benzo[b]aze-
pine-2-one (3) had a less favorable profile. These results
suggested that a suitable conformationally constrained
structure is required for both D2 and D4 receptors
binding. Therefore, we decided to expand the con-
formationalSAR studies on the previous series, and
designed new compounds (e.g., 4) having a tricyclic ring
*Corresponding author. Tel.: +1-203-488-8201x3077; fax: +1-203-
483-7027; e-mail: hzhao@nrgn.com
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(02)01056-9

702
H. Zhao et al. / Bioorg. Med. Chem. Lett. 13 (2003) 701–704
system by either adding two carbons between the a
position of the amide and the corresponding carbon of
the phenylring of compound 1, or simply connecting
the ethyland phenylgroup of compound 3 as shown in
Figure 1.
both the substituents on the piperazinylring are die-
quatorialand the cholrophenylring is oriented in the
same way. The removalof the indoinl e fusion in
1
allows the ethyl substituent on the nitrogen to swing out
of plane (j3) but does not result in any other distortion
of the molecule. The additional methylene unit in the
lactam 2 is incorporated in the twist conformation
without disrupting the conjugation between the aro-
matic ring and the amide bond. All the potential phar-
macophore points are nearly superimposable in the low-
energy conformations of 1 and 2.
To seek further support on this drug design strategy and
predict the biological profiles for the new tricyclic lac-
tam compounds, we subsequently performed molecular
modeling studies. The low energy conformers of 1 and 2
have many common features. The amide bond is copla-
nar with the aromatic ring (j1, Table 1), the carbonyl
and C–N (piperazine) bonds are nearly eclipsed (j2),
When 2 is expanded by one methylene group to form 3,
the additionalflexibitly in the seven-membered ring
leads to substantial geometric changes. In the more
stable conformers, the ring is calculated to prefer a boat
conformation in which the arene ring and the amide
bond are not fully in conjugation. The corresponding
dihedralangel ( j1) is 34ꢀ, substantially higher than in 1
and 2 (Table 1). This twist is noted in all the low energy
conformers within 20 kJ/molof the golbalminimum.
The relative positions of the arene ring, the amide bond
and the piperazinylring are therefore quite different
from those in 1 and 2. Interestingly, fusing the amide
nitrogen to the arene unit through a five-membered ring
(4) acts as a strong restraint and, as a result, the amide
bond is nearly coplanar with the arene ring (Table 1).
The key geometric features of 4 are therefore nearly the
same as in 1 and 2 as shown in Figure 2.
In order to estimate the energetic cost involved in
retaining perfect conjugation between the aromatic ring
and the amide bond, additional calculations were car-
ried out on 1 and 3 in which the dihedralangel j1 was
constrained to be 0ꢀ but all other geometric parameters
were fully optimized. For 1, the constrained structure
was computed to be only 0.2 kJ/mol higher in energy
than the fully optimized form. In contrast, forcing the
dihedralangel j1 to be 0ꢀ was computed to result in an
energy penalty of 19.8 kJ/mol for 3. The calculations
confirm that the relative spatial disposition of the dif-
ferent functionalgroups of 1, 2 and 4 is similar but that
of 3 is quite different. The overall shapes of these com-
pounds may therefore account for the trends in the
observed D4 binding. If the coplanarity of the aryl ring
Figure 1.
Table 1. Calculated dihedral angles in the lowest energy conformers
of 1–4a
Compd
j1
j2
j3
1
2
3
4
7.3
1.6
34.0
7.3
18.8
13.2
26.6
3.2
14.6
91.7
105.0
12.7
aDetails of calculations: A Monte Carlo conformational search was
performed on compounds 1–4 using Macromodel’s Batchmin program
(v7.2). The ‘mixed MCMM/low mode’ method was used with the
MMFF94s force field and water as solvent. For each compound, 1000
structures were generated and optimized using the conjugate-gradient
energy minimization method. These searches were run on a SGI
Octane workstation with an R12000 CPU.
Figure 2. Low energy confirmations of structures 1–4.

H. Zhao et al. / Bioorg. Med. Chem. Lett. 13 (2003) 701–704
703
and the amide bond is the principaldeterminant of
activity, distalmodifications on the benzylring of 4 may
not have too much of an impact on D4 binding affi-
nities. Structure 4 would then represent an attractive
core for fine-tuning D2/D4 selectivity.
[3H]prazosin as the competitive ligand. The con-
formationally restricted e-lactam subunit tremendously
changes the compound biological properties. In parti-
cular, compared with compound 1, all nine e-lactam
compounds (4, 12–19) showed lower binding affinities
for a1. Among them, compounds 4 and 12 display
binding affinities and affinity ratio in the desired range.
Compound 4 displays 6-fold greater potency for D2 and
3-fold lower for D4 than compound 1. The 4-methyl-
Synthesis, Biological Results and Discussion
Scheme 1 depicts the synthesis of 5-[4-(4-chloro-benzyl)-
piperazin-1-yl]-1,2,6,7-tetrahydro-5H-azepino[3,2,1-
hi]indol-4-one 4. Acylation of indoline 5 with succinic
anhydride 6 in the presence of triethylamine in di-
chloromethane gave amide acid 7, which was then con-
verted to keto-lactam 8 by intramolecular Friedel–
Crafts cyclization in low yield (15%). Several reaction
conditions were examined for the cyclization, but no
improvement could be achieved, possibly due to fea-
tures of the seven-membered ring. Hydrogenation of
compound 8 yielded e-lactam 9, followed by silylation
with iodotrimethylsilane and iodination to give 5-iodo-
1,2,6,7-tetrahydro-5H-azepino[3,2,1-hi]indol-4-one 10.
Finally, compound 4 was obtained in high yield by
refluxing of compound 10 and 1-(4-chloro-benzyl)-
piperazine 11 with potassium carbonate in acetonitrile.
In addition, a number of methylindoinl e and sub-
stituted benzylpiperazine containing compounds have
been prepared using the same synthetic pathway.
benzylcompound
4-chlorobenzyl compound 4.
12 showed a profile similar to
Compounds were also assessed as to their functional
activity both at the D2 and D4 receptors. D2 functional
activity was assessed via compound reversalof quinpir-
ole inhibited, forskolin stimulated cAMP production
from whole cells, while D4 functionalactivity was
assessed via inhibition of quinpirole stimulated
GTPg35S binding from cell membranes. Functional
assessment of compound 4 at both the D2 and D4
receptors indicates no agonist properties up to 10 mM,
while demonstrating functional Ki values of 62 nM at
the D2 receptor and 3 nM at the D4 receptor.
In conclusion, with the assistance of molecular
modeling studies, a new series of mixed dopamine
D2/D4 receptor antagonist 5-piperazinyl-1,2,6,7-tetra-
hydro-5H-azepino[3,2,1-hi]indol-4-one derivatives were
designed and synthesized. As a result of SAR stud-
ies, the highly conformationally restricted tricyclic
compounds 4 and 12 displayed a D2 and D4 affinity
ratio similar to that of clozapine while being free of
the liabilities caused by high a1 affinity. These two
representative compounds from the new tricyclic
series are currently under further pharmacological
evaluation.
The binding affinity data for D2, D4 and a1 are sum-
marized in Table 2. Affinities at D2 and D4 receptors
were determined via standard competitive displacement
assays using human D2 and D4 clones with [3H]YM
09151 as the competitive ligands. Affinity at the a1
receptor was determined via standard competitive dis-
placement assays using rat brain homogenate with
Table 2. Binding affinities
Compd
R1
R2
Ki (nM )
D4
D2
113
690
21
>1000
5
209
9
a1
Clozapine
1
2
3
4
12
13
14
15
16
17
18
19
—
—
—
—
H
H
Me
—
—
—
—
4-Cl116
4-Me
4-Cl139
4-Me
17
1.6
4
1511
4
88
1265
2678
2284
4
1361
1000
1000
10
1735
Me
26
Scheme 1. Reagents and conditions: (i) TEA, DCM, rt, 16 h, 78%; (ii)
oxalyl chloride, DMF (cat), DCE, rt, 3 h; then 2 equiv anhydrous
AlCl3, DCE, 0 ꢀC to rt, 4 h, additional2 equiv anhydrous AlCl 3, DCE,
60 ꢀC, 16 h, 15%; (iii) H2, 10% Pd/C, 50 psi, HOAc, rt, 24 h, 98%; (iv)
TMSI, TMEDA, DCM, 0 ꢀC, 30 min; then iodine, 0 ꢀC, 40 min, 74%;
(v) K2CO3, CH3CN, reflux, 18 h, 90%.
di-Me
di-Me
di-Me
di-Me
di-Me
4-Cl201
4-Me
19
165
952
313
220
12
34
65
29
490
653
2-OMe-4-Me
2-OMe-5-Me
5-Cl-2-OMe
983
1056

704
H. Zhao et al. / Bioorg. Med. Chem. Lett. 13 (2003) 701–704
6. Zhao, H.; Thurkauf, A.; He, X.; Hodgetts, K.; Zhang, X.;
References and Notes
Rachwal, S.; Kover, R. X.; Hutchison, A.; Peterson, J.; Kiel-
tyka, A.; Brodbeck, R.; Primus, R.; Wasley, J. Bioorg. Med.
Chem. Lett. 2002, 12, 3105.
7. Zhao, H.; He, X.; Thurkauf, A.; Hoffman, D.; Kieltyka,
A.; Brodbeck, R.; Primus, R.; Wasley, J. Bioorg. Med. Chem.
Lett. 2002, 12, 3111.
1. Reynolds, G. P. Trends Pharmacol. Sci. 1992, 13, 116.
2. Kane, J.; Honigfeld, G.; Singer, J.; Meltzer, H. Arch. Gen.
Psychiatry 1988, 45, 789.
3. Kane, J.; Honigfeld, G.; Singer, J.; Meltzer, H. Psycho-
pharmacol. Bull. 1988, 24, 62.
4. Kulagowski, J. J.; Patel, S. Curr. Pharm. Des. 1997, 3, 355.
5. Faraci, W. S.; Zorn, S. H.; Sanner, M. A.; Fliri, A. Curr.
Opin. Chem. Biol. 1998, 2, 535.
8. Zhao, H.; Thurkauf, K.; Braun, J.; Brodbeck, R.;
Kieltyka, A. Bioorg. Med. Chem. Lett. 2000, 10,
2119.