Conformationally constrained NR2B selective NMDA receptor antagonists derived from ifenprodil: Synthesis and biological evaluation of tetrahydro-3-benzazepine-1,7-diols

Article abstract of DOI:10.1016/j.bmc.2010.09.026

NR2B selective NMDA receptor antagonists with tetrahydro-3-benzazepine-1,7- diol scaffold have been designed by formal cleavage and reconstitution of the piperidine ring of the lead compound ifenprodil (1). The secondary amine 10 represents the central bu

Full text of DOI:10.1016/j.bmc.2010.09.026

Bioorganic & Medicinal Chemistry 18 (2010) 8005–8015
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Conformationally constrained NR2B selective NMDA receptor antagonists
derived from ifenprodil: Synthesis and biological evaluation
of tetrahydro-3-benzazepine-1,7-diols
Bastian Tewes a, Bastian Frehland a, Dirk Schepmann a, Kai-Uwe Schmidtke b, Thomas Winckler b,
Bernhard Wünsch a,
⇑
a Institut für Pharmazeutische und Medizinische Chemie der Westfälischen Wilhelms-Universität Münster, Hittorfstraße 58-62, D-48149 Münster, Germany
b Institut für Pharmazie der Friedrich-Schiller-Universität Jena, Semmelweisstraße 10, D-07743 Jena, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 July 2010
Revised 9 September 2010
Accepted 10 September 2010
Available online 24 September 2010
NR2B selective NMDA receptor antagonists with tetrahydro-3-benzazepine-1,7-diol scaffold have been
designed by formal cleavage and reconstitution of the piperidine ring of the lead compound ifenprodil
(1). The secondary amine 10 represents the central building block for the synthesis of more than 25
tetrahydro-3-benzazepin-1-ols. Generally 7-hydroxy derivatives display higher NR2B receptor affinities
than the corresponding 7-benzyloxy compounds. A distance of four atoms (five bond lengths) between
the basic amino group and the terminal aryl moiety led to highest NR2B affinity. 3-(4-Phenylbutyl)-
2,3,4,5-tetrahydro-1H-3-benzazepine-1,7-diol (WMS-1410, 25) represents the most promising NR2B
antagonist of this series showing a Ki-value of 14 nM. Compound 25 reveals excellent selectivity over
more than 100 further relevant target proteins, antagonizes glutamate induced excitotoxicity
(IC50 = 18.4 nM) and is metabolically more stable than ifenprodil. Up to a dose of 100 mg/kg 25 is well
tolerated by mice and it shows dose dependent analgesic activity in the late neuropathic pain phase of
the formalin assay.
Keywords:
NR2B selective NMDA receptor antagonists
Structure–activity relationships
3-Benzazepines
Neuroprotection
Neuropathic pain
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
concentration, which results in uncontrolled activation of various
Ca2+-dependent enzymes, neuronal damage and at the end cell
In the mammalian central nervous system (CNS), the N-methyl-
-aspartate (NMDA) receptor associated ion channel plays a central
death. This process of excitotoxicity occurs during many acute
adverse events (e.g., traumatic brain injury, stroke) and chronic
neurodegenerative disorders (e.g., Alzheimer‘s disease, Parkinson‘s
disease, Huntington’s disease). Due to its central role in these
processes the NMDA receptor represents an interesting target for
the development of novel drugs.1–3
D
role in physiological processes, such as neuronal development,
synaptic plasticity, learning, and memory. The NMDA receptor rep-
resents a ligand gated ion channel controlling the flow of Ca2+-,
Na+-, and K+-ions across neuronal cell membranes. According to
the concentration gradient Ca2+- and Na+-ions penetrate into the
neuron, whilst K+-ions leave the cells. The NMDA receptor is acti-
vated by simultaneous binding of the physiological agonist (S)-glu-
tamate and the coagonist glycine. In addition to binding of two
agonists, a predepolarization of the cell membrane is required be-
fore opening of the ion channel takes place. This depolarization can
be achieved by activation of further ionotropic glutamate recep-
tors, like AMPA or kainate receptors, in the vicinity of the NMDA
receptor. The depolarization opens the ion channel by removal of
Mg2+-ions, which are bound within the channel pore and block
the passage of cations in particular the influx of Ca2+-ions.1–3
Overactivation of the NMDA receptor caused by excessive gluta-
A functional NMDA receptor is formed by four subunits (het-
erotetramer).4 Three types of subunits have been identified: The
NR1 subunit with eight splice variants (NR1a-h), the NR2 subunit,
which exists in four distinct subtypes (NR2A-D) encoded by four
distinct genes, and the NR3 subunits NR3A and NR3B (two genes).
At least one NR1 subunit bearing the glycine binding site and one
NR2 subunit responsible for glutamate binding are required in a
functional NMDA receptor. Whereas the NR1 subunit is expressed
ubiquitously in the CNS, the density of the different NR2 subunits
varies depending on the region of the CNS. The NR2A subunit is
found throughout the whole CNS, but the NR2B subunit is highly
expressed only in the cortex and hippocampus with a rather low
density in the cerebellum and hypothalamus. The NR2C subunit
predominates in the cerebellum and both the NR2C and NR2D
subunits are preferentially formed in the brain stem and the
spinal cord. NR3 subunits are expressed predominantly in the
mate release leads to an unphysiologically high intracellular Ca2+
-
⇑
Corresponding author. Tel.: +49 251 8333311; fax: +49 251 8332144.
E-mail address: wuensch@uni-muenster.de (B. Wünsch).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.09.026

8006
B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
isoleucine, valine binding protein (LIVBP) as template, showed the
HO
NR2B receptor antagonist Ro 25-6981 docked into the modeled
binding pocket of the NR2B subunit and the corresponding crucial
interactions.17
OH
N
N R
H3CO
CH3
Originally, 1 has been developed as an a1 adrenoceptor antago-
HO
nist. However, it was shown that 1 also interacts with several other
2
1
types of receptors and ion channels, including 5-HT1A, 5-HT2 and
both
r1 and r
2 receptors.3,18–20 Altogether, 1 is a nonselective drug
HO
with side effects such as hypotension, cognitive problems and
neurotoxicity.20 In addition to the side effects, the bioavailability
of 1 is low due to the fast and extensive metabolic degradation.21,22
Formal cleavage and reconstitution of a bond in the piperidine
ring of 1 lead to tetrahydro-3-benzazepines 2 (Fig. 1). Since the
methyl moiety in the side chain of ifenprodil is not essential for
the interaction with NR2B receptors, it was omitted in the first ser-
ies of 3-benzazepines 2. Very recently we have reported on the
synthesis of tetrahydro-3-benzazepines 2 with a methoxy moiety
in position 7.23 The most potent compounds in this new class of
1
3
N R
HO
7
3
Figure 1. Ifenprodil (1) and tetrahydro-3-benzazepines 2 as lead compounds for
the development of novel NR2B selective NMDA receptor antagonists of type 3.
NR2B antagonists are bearing
x-arylalkyl residues with preferably
developing CNS and seem to be of low relevance in the adult
brain.5,6
four methylene moieties between the basic N-atom of the 3-
benzazepine ring and the terminal phenyl moiety. In particular,
the NR2B affinity of 2a (WMS-1405, R = 4-phenylbutyl, Ki = 5.4 nM)
is higher than the NR2B affinity of 1 (Ki = 10 nM).23 In contrast to 1,
the tetrahydro-3-benzazepine 2a displayed high selectivity against
Non-selective NMDA receptor antagonists, like ketamine, phen-
cyclidine, and CPP (3-(3-carboxypiperazin-1-yl)propane-1-phos-
phonic acid) have been reported to produce symptomatic relief
in a number of neuropathic disorders, including postherpetic
neuralgia, central pain caused by spinal cord injury and phantom
limb pain.7–10 However, at analgesic doses these drugs also induce
inacceptable side effects, including hallucinations, dysphoria, and
cognitive and motor dysfunctions, which limit their use.11
Due to the restricted expression of the NR2B subunit in the CNS,
NMDA receptor antagonists selectively addressing NR2B subunit
containing NMDA receptors should show an improved side effect
profile. In particular, expression of the NR2B subunit in the cerebel-
lum is rather low and, therefore, cognitive side effects should be
minimized with NR2B selective antagonists.11 Potential therapeutic
applications of NR2B selective NMDA receptor antagonists are
Parkinson‘s disease,12 traumatic brain injury,13 stroke,14 migraine,15
alcohol-withdrawal,16 and chronic and neuropathic pain.11
The lead compound for this project was ifenprodil (1, Fig. 1),
which blocks NMDA receptors by selective interaction with the ifen-
prodil binding site on the NR2B subunit. The ifenprodil binding site
is close to the polyamine binding site, but not completely identical
with it. Neither the NMDA receptor nor the NR2B subunit have been
crystallized so far to prove the exact structure of the binding site of
NR2B selective ligands. However a molecular modeling study using
the X-ray crystal structure of the similar bacterial protein leucine,
a1, 5-HT1A, 5-HT2, r1 and r2 receptors.
In order to come closer to the structure of the lead compound 1
the methoxy group in position 7 of the 3-benzazepines 2 should be
replaced by a hydroxy group in a new series of NR2B antagonists.
The resulting phenols 3 contain the same H-bond-donor group as
1. Since an H-bond donor group in position 7 is described to be
essential for high NR2B affinity18 the phenols 3 should interact
with high affinity with the NR2B subunit.
2. Chemistry
The designed phenols 3 were synthesized starting from 7-meth-
oxy-3-benzazepin-1-one 423 (Scheme 1). At first the methyl ether
of 4 was cleaved with the Lewis acid AlCl3 in refluxing CH2Cl2 to
afford the phenol 5 in 68% yield. After reduction of the ketone 5
with NaBH4 the diol 6 was obtained in 93% yield. However, all at-
tempts to remove the tosyl protecting group of 6 failed to produce
the secondary amine 7 (e.g., Mg in methanol, reflux, decomposi-
tion). Since the cleavage of the tosyl group of the corresponding
methoxy derivative had been successful,23 we assumed that the
O
O
HO
6
HO
i)
ii)
N
Tos
N
Tos
N
Tos
NH
HO
H3CO
HO
HO
7
5
4
iii)
HO
O
HO
v)
iv)
N
Tos
N
Tos
NH
BnO
BnO
BnO
8
9
10
Scheme 1. Reagents and conditions: (i) AlCl3, CH2Cl2, reflux, 23 h, 68%; (ii) NaBH4, CH3OH, rt, 18 h, 93%; (iii) benzyl bromide, acetone, K2CO3, reflux, 4 h, 85%; (iv) NaBH4,
CH3OH, rt, 6 h, 98%; (v) Mg, CH3OH, reflux, 5 h, 83%.

B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
8007
unstable diol substructure of 6 (phenylogous hydrate) inhibits the
removal of the N-protecting group.
Therefore, the phenolic hydroxy group had to be supplied with
a new protecting group. For this purpose the benzyl group was se-
lected, because it can easily be attached, is stable during various
reaction conditions, and can be removed selectively upon hydrog-
enolysis. Reaction of phenol 5 with benzyl bromide in the presence
of K2CO3 provided the benzyl ether 8, which was reduced with
Table 1
HO
Synthesis of 7-benzyloxy-3-benzazepines by alkylation of 10 with RX
HO
Compd
R
Y
X
Yield (%)
iii)
N R
N R
11 (WMS-1418)
12
CH2
O
Cl
Br
69
79
HO
BnO
X
13
14
S
SO2
Cl
Cl
91
90
Y
25 - 28
i)
11 - 19
15
16
17
tert-Butyl
OCH3
F
Cl
Cl
Cl
74
76
90
R
10
ii)
Y
O
O
O
HO
HO
R1
R2
BnO
H3C
R2
R1
R2
R1
iii)
CH3
N
N
HO
18
19
–
–
Br
Cl
65
93
20 - 24
29 - 32
CH3
N
Scheme 2. Reagents and conditions: (i) R-X (see Table 1), CH3CN, K2CO3, Bu4NI,
reflux, 8–72 h; (ii) different ketones and an aldehyde (see Table 2), NaBH(OAc)3, 1,2-
dichloroethane, rt, 16 h; (iii) H2 (1 bar), Pd/C, CH3OH or THF, rt, 30 min–16 h (see
Table 3).
O
Table 2
Synthesis of 7-benzyloxy-3-benzazepines by reductive alkylation of 10 with R1C(@O)R2 and NaBH(OAc)3
NR2
O
Compd
Yield (%)
R1
R2
R1
O
R2
O
O
O
O
R2N
20
80
41
O
O
O
R2N
O
O
cis-21
O
O
trans-21
R2N
R2N
O
O
O
O
N
N
22
43
R2N
cis-23
71
14
trans-23
R2N
R2N
O
H
24
93
N
N

8008
B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
NaBH4 to obtain the alcohol 9. Treatment of the sulfonamide 9 with
magnesium in boiling methanol gave the secondary amine 10
(83%), which represents the central building block of this project
(Scheme 1).
In order to find reliable structure NR2B affinity relationships a
large number of tetrahydro-3-benzazepines with various N-sub-
stituents was prepared. Since the 4-phenylbutyl derivative 2a with
the same N-Ph-distance as the lead compound 1 (see Fig. 1)
Table 3
Synthesis of 7-hydroxy-3-benzazepines by hydrogenolysis of 7-benzyloxy-3-benzazepines
Educt
Product
R
Reaction time
30 min
Solvent
CH3OH
Yield (%)
67
11
25 (WMS-1410)
12
14
19
26
27
28
2 h
CH3OH
CH3OH
CH3OH
63
75
68
O
S
16 h
1 h
O
O
CH3
N
O
O
O
20
29
1 h
1 h
CH3OH
CH3OH
63
93
O
cis-21/trans-21 (1:1)
trans-30
O
O
N
22
31
2 h
2 h
CH3OH
CH3OH
60
74
cis-23
cis-32
trans-23
trans-32
2 h
1 h
THF
65
64
38
39
CH3OH
F
O
O
MesO
i)
O
HO
34
O
O
O
HO
N
O
BnO
ii)
iii)
33
21
O
O
O
35
Scheme 3. Reagents and conditions: (i) methanesulfonyl chloride, NEt3, CH2Cl2, rt, 16 h, 99%; (ii) Dess–Martin periodinane, CH2Cl2, rt, 3 h, 28%; (iii) 10, NaBH(OAc)3, 1,2-
dichloroethane, rt, 16 h, 41%.

B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
8009
showed the highest NR2B affinity in the series of 7-methoxytetra-
hydro-3-benzazepines,23 substituents with a spacer of 4–5 methy-
lene groups (or heteroatom equivalents) were preferably selected.
The tetrahydro-3-benzazepines 11–19 were synthesized by alkyl-
ation of 10 with various haloalkanes (Scheme 2 and Table 1). In
case of chloro derivatives Bu4NI was added to improve the conver-
sion by in situ generation of iodoalkanes (Finkelstein reaction).
Alternatively, tetrahydro-3-benzazepines 20–24 were synthe-
sized by reductive alkylation of secondary amine 10 with different
ketones and an aldehyde in the presence of NaBH(OAc)3 (Scheme 2
and Table 2).24 In order to increase the reactivity of the ketone ace-
tic acid was added to the reaction mixture during the reductive
alkylation of 10 with 4-phenylcyclohexanone. In this case two dia-
stereomers cis-23 and trans-23 were formed in the ratio 1:1, which
were separated by fc. The relative configuration of cis-23 and trans-
23 was unequivocally assigned by 1H NMR spectroscopy.
receptor binding assay recently developed in our group.28 In this
assay tritium labeled ifenprodil was employed as radioligand.
Membrane homogenates prepared upon ultrasonication of L(tk-)-
cells stably expressing recombinant human NR1a/NR2B receptors
served as receptor material.29 The high density of NMDA receptors
renders this system selective. The expression of NMDA receptors at
the cell surface was induced by addition of dexamethasone to the
growth medium. During this period cell death was inhibited by
addition of the NMDA receptor antagonist ketamine (phencycli-
dine binding site) to the growth medium.
The receptor affinities of 7-benzyloxy-3-benzazepines are sum-
marized in Table 4. Most of the 7-benzyloxy-3-benzazepines display
only moderate affinity to the ifenprodil binding site of the NMDA
receptor. For example, the NR2B affinity of the 4-phenylbutyl deriv-
ative 11 (Ki = 187 nM) is about 35-fold lower than the NR2B affinity
of the corresponding methoxy derivative 2a (Ki = 5.4 nM).
For the synthesis of the 2-phenyl-1,3-dioxan-5-yl substituted 3-
benzazepine 21 the 1,3-dioxan-5-ol 33 was converted into mesy-
late 34 (Scheme 3). However, all attempts to react the secondary
amine 10 with the mesylate 34 failed to give the substitution prod-
uct 21. Therefore, the alcohol 33 was oxidized with Dess–Martin
Periodinane25 to produce the ketone 35, which reacted smoothly
with the secondary amine 10 and NaBH(OAc)3 to yield the dioxanyl
derivative 21 as 1:1-mixture of cis/trans-diastereomers.
In the final step the benzyl ether of the tetrahydro-3-benzaze-
pines was cleaved by hydrogenolysis with H2 (1 bar) and Pd/C to
yield the phenols 25–32 (Table 3). After hydrogenolysis of the dia-
stereomeric mixture cis-21 and trans-21 (1:1) the 1H NMR spec-
trum of the product shows only one set of signals for trans-30.
The trans-configuration was proved by NOE difference spectros-
copy. The fact that after hydrogenolysis of a 1:1 mixture of diaste-
reomers cis-21 and trans-21, only the trans-isomer trans-30 was
isolated in a yield of 93%, can only be explained by a Pd catalyzed
isomerization of the acetalic substructure to end up with the ther-
modynamically more stable trans-isomer trans-30.
The bioisosteric replacement of the methylene moiety in posi-
tion 4 of the 4-phenylbutyl side chain of 11 with an O- (12), S-atom
(13) or a SO2-group (14) led to increased N2RB affinity. In case of
the sulfone 14 (Ki = 37 nM) the NR2B affinity is fivefold increased
compared with the affinity of 4-phenylbutyl derivative 11.
In the 4-phenylcyclohexyl derivatives 23 a conformational
restriction of the flexible side chain of the 4-phenylbutyl derivative
11 is realized. Both diastereomers cis-23 (Ki = 139 nM) and
trans-23 (Ki = 389 nM) show similar NR2B receptor affinities as
the flexible 4-phenylbutyl derivative 11 do, although cis-23 is
slightly more potent than trans-23. The comparable affinities of
11 and 23 indicate that a conformationally constrained side chain
is well accepted by the ifenprodil binding site of the NR2B subunit.
The receptor affinities of phenolic tetrahydro-3-benzazepines
are summarized in Table 5. In general tetrahydro-3-benzazepines
with a phenolic OH moiety in position 7 show considerably higher
NR2B affinities than the corresponding benzyl ethers (Table 4) and
methyl ethers.23 This observation is in good accordance with re-
ported results,18 showing that an H-bond donor group, for exam-
ple, a OH group, is favorable for high NR2B affinity.
In order to increase the diversity of substituents, the acetals
15–17 were hydrolyzed with dilute HCl to yield the ketones
36–38. Hydrogenolytic cleavage of the benzyl protecting group of
38 led to phenol 39 (Scheme 4). In particular the haloperidol like
N-residue of 38 and 39 was selected, because haloperidol also
interacts with NR2B containing NMDA receptors.26,27
The NR2B affinities of the 4-phenylbutyl derivative WMS-1410
(25, Ki = 14 nM) and its bioisosteric phenoxypropyl derivative 26
(Ki = 20 nM) are in the same range as the NR2B affinity of ifenprodil
(Ki = 10 nM). Surprisingly the methyl ether 2a (Ki = 5.4 nM) is
slightly more potent than the phenols 25 and 26.
Whereas introduction of a sulfonyl moiety (27) or a rigid amide
substructure (28) into the N-side chain led to considerable loss of
NR2B affinity, a carbonyl moiety in the side chain (39) is tolerated
by the ifenprodil binding site. The N-residue of 29 without a phe-
nyl group decreased the NR2B affinity.
3. Receptor affinity
3.1. Affinity toward NR2B containing NMDA receptors
The NR2B receptor affinities of the synthesized tetrahydro-3-
benzazepines 11–32 and 36–39 were determined in a competitive
The most promising NR2B receptor antagonists of this series of
compounds are the diastereomeric 4-phenylcyclohexyl derivatives
HO
HO
ii)
i)
15 - 17
N
N
HO
BnO
F
R
O
O
39
36 - 38
Compound
15, 36
R
tert-Butyl
OCH3
F
16, 37
17, 38
Scheme 4. Reagents and conditions: (i) 1 M HCl, Et2O, rt, 16 h, 69% (36), 80% (37), 76% (38); (ii) H2 (1 bar), Pd/C, CH3OH or THF, rt, 1 h, 64% (see Table 3).

8010
B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
Table 4
Affinities of 7-benzyloxy-3-benzazepin-1-ols to the ifenprodil binding site of NR2B containing NMDA receptors, the PCP binding site of the NMDA receptor as well as to
receptors
r1 and r2
HO
N
R
O
Compd
R
Ki SEM (nM)
NR2B
PCP
r1
r2
11
12
13
14
–(CH2)4Ph
187 72
93 27
41 2.6
37 8.9
>1
>1
>1
>1
l
M
M
M
M
293 58
193
376
1050
945
653
–(CH2)3OPh
–(CH2)3SPh
–(CH2)3SO2Ph
l
l
l
349
3040
CH3
CH3
CH3
O
15
16
>1
lM
>1
>1
l
M
M
>1
>1
l
M
M
900
O
H3C
H3C
CH3
OCH3
O
661
l
l
>1
l
M
O
CH3
O
F
17
1420
>1
l
M
>1
l
M
>1
>1
l
M
M
O
H3C
CH3
–CH2-Ph-3-Ph
18
19
1210
377 53
>10
>1
l
M
>1
>1
lM
M
l
–CH2CON(CH3)CH2Ph
lM
l
2960
O
20
1120
1000
>1
l
M
183
753
O
O
21
>1
lM
186 74
>1
>1
l
M
M
O
cis/trans (50:50)
O
N
22
>1
lM
>10
l
M
>1 lM
l
cis-23
139 29
>1
>1
l
M
10 0.8
311
trans-23
388 130
lM
386
>1
l
M

B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
8011
Table 4 (continued)
Compd
R
Ki SEM (nM)
NR2B
PCP
r1
r2
24
1310
>1
lM
604
6430
N
36
37
38
–(CH2)3CO-Ph-4-tert-butyl
–(CH2)3CO-Ph-4-OCH3
–(CH2)3CO-Ph-4-F
–(CH2)4Ph 7-H3CO instead of 7-BnO
Ifenprodil
644 202
>1
>1
>1
>10
—
l
l
l
M
M
M
>1
>1
293 58
182 38
125 24
—
l
l
M
M
>1 lM
4500
1050
554 127
98.3 34
—
>1
lM
151 61
5.4 0.4
10 0.7
13 2.5
—
2a (WMS-1405)
1
l
M
Eliprodil
Dexoxadrol
—
32 7.4
—
—
Haloperidol
Di-o-tolylguanidine (DTG)
—
—
—
—
6.3 1.6
89 29
78 2.3
58 18
At a concentration of 1
lM or even 10
l
M low-affinity compounds did not compete significantly with the radioligand. Therefore, the Ki-value is only estimated to be greater
than 1 lM.
Table 5
tolerated by NR2B containing receptors. However, this system
Affinities of tetrahydro-3-benzazepine-1,7-diols to the ifenprodil binding site of NR2B
containing NMDA receptors, the PCP binding site of the NMDA receptor as well as to
r1 and r2 receptors
appeared to be rather sensitive, since replacement of two methylene
moieties of the cyclohexane ring of trans-32 by two O-atoms (trans-
30) led to almost complete loss of NR2B affinity. The high electro-
negativity and/or the additional H-bond acceptor properties of the
O-atoms may be responsible for the reduced affinity of trans-30.
The benzamide 31 (Ki = 204 nM) with changed geometry and
N-Ph-distance displays moderate NR2B affinity.
With respect to NR2B receptor affinity compounds 25
(Ki = 14 nM), 26 (Ki = 20 nM), cis-32 (Ki = 3.3 nM), trans-32
(Ki = 9 nM), and 39 (Ki = 36 nM) represent the five most potent
tetrahydro-3-benzazepines of this series.
HO
N
R
HO
Compd
R
Ki SEM (nM)
NR2B
PCP
r1
r2
25
26
27
28
–(CH2)4Ph
–(CH2)3OPh
–(CH2)3SO2Ph
14 1.5
20 2.2
>1
>1
>1
>1
l
l
l
l
M
M
M
M
194
18,000
170 51 5.9 1.7
>1
172
3.2. Receptor selectivity
>1
>1
l
l
M
M
l
M
653
1580
–CH2CON(CH3)CH2Ph
O
In order to learn more about the receptor selectivity of this new
class of NR2B antagonists (see also Ref. 23), the affinities of all test
compounds to the phencyclidine (PCP) binding site of the NMDA
29
>1
>1
l
M
M
>1
>1
l
l
M
M
>1
>1
l
M
658
O
receptor30 and to both
r
receptor subtypes (r1 and r2 recep-
tor)31–33 were investigated systematically in receptor binding stud-
ies with radioligands. In Tables 4 and 5 the PCP, 1 and 2 receptor
trans-
O
O
l
l
l
M
M
>1
98
lM
30
r
r
affinities of the tetrahydro-3-benzazepines 11–32 and 36–39 are
compared with their NR2B receptor affinities.
O
N
31
204 95 >1
l
l
M
M
>1
The 7-benzyloxy-3-benzazepines as well as the corresponding
phenols do not interact significantly with the PCP binding site of
the NMDA receptor. This result indicates a high selectivity for the
ifenprodil binding site over the phencyclidine binding site of the
NMDA receptor, at least for the most potent NR2B ligands.
The r1 receptor affinity of the tetrahydro-3-benzazepines has
to be discussed differentially. Most of the compounds have low
cis-32
3.3 1.0 >1
123 11 8.6 2.5
to moderate r1 affinity. However, the Ki-values (r1) of the ben-
zyl ethers 11, 12, and trans-23 are very similar to their Ki-values
at the NR2B receptor indicating low receptor selectivity. More-
over, for the cis-configured phenylcyclohexyl derivative cis-23 a
r1 affinity of 10 nM was found, which is 14-fold higher than
its NR2B affinity. On the other hand the most potent NR2B li-
gands with phenolic substructure, that is, 25, 26, cis-32, trans-
32, and 39, show a 10–40-fold selectivity for the NR2B receptor
over the r1 receptor.
trans-
32
9.0 3.7 >1
36 6.1 >1
l
l
M
M
200 36 8.9 1.5
1050 33 12
39
(CH2)3CO-Ph-4-F
At a concentration of 1
with the radioligand. Therefore, the Ki-value is only estimated to be greater than
M.
lM low-affinity compounds did not compete significantly
1
l
cis-32 (Ki = 3.3 nM) and trans-32 (Ki = 9.0 nM) binding in the low
nanomolar range. As described for the benzyl ethers 23 the
cis-isomer is slightly more potent than the trans-isomer. Obviously
the conformationally restricted 4-phenylcyclohexyl residue is well
The r2 receptor affinities of most of the test compounds are
very low. Unfortunately, the r2 receptor affinities of just the most
potent NR2B ligands are remarkably high. Whereas the Ki-values
(r2) of trans-32 and 39 are identical with their NR2B Ki-values,

8012
B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
the r2 affinity of 26 even exceeds its NR2B affinity threefold (see
Table 5). The most potent NR2B ligand of this series, that is,
trans-32, shows only a slight preference for the NR2B receptor
(2.5-fold). Apparently, conformational restriction of the flexible
N-butyl side chain by a cyclohexane ring does not only increase
Surprisingly, cis-32, which represents the most potent ligand in
the competition experiments, was considerably less potent in the
functional assay (IC50 = 12.1 lM). However, the structure of the
curve and the dose/activity relationship proves the antagonistic
activity of cis-32 as well.
the NR2B affinity but also the
r
2 (and
r
1) affinity, losing selectivity
The same experiment was performed using L12-G10 cells stably
expressing NR1a/NR2A receptors. The inhibition of excitotoxicity
against receptors. Compound 25 with the rather flexible 4-phen-
r
ylbutyl side chain is the only potent NR2B ligand of this series
was lower than 10%, when a concentration of 10 lM of 25, cis-32,
showing an excellent selectivity over the r2 receptor.
and 2a was used. Obviously, 25, cis-32, and 2a did not block NMDA
Taking NR2B affinity and selectivity over related receptor sys-
tems into account, the phenylbutyl derivative 25 (Ki = 14 nM) rep-
resents the most promising ligand of this study. Therefore, further
pharmacological properties of 25 were investigated.
receptors containing NR2A subunits instead of NR2B subunits.
5. Metabolic stability
At first a receptor binding profile of 25 was performed. In compe-
tition experiments with radioligands the interaction of 25 with more
Studies on metabolic stability were carried out using micro-
somes prepared from mouse, rat and human livers.37 The half-life
of phenol 25 was 28 min (mouse), <3 min (rat) and 21 min (hu-
man), respectively. The same set-up resulted in half-life of ifenpro-
dil (1) of <5 min (mouse), <1 min (rat), and 5 min (human),
respectively. Obviously the metabolic degradation of 25 by CYP en-
zymes is reduced compared to the lead compound 1 indicating a
higher metabolic stability of 25.
than 100 relevant receptors, transporters and enzymes including a1
,
D1, D2, -opioid, -opioid, d-opioid, 5-HT1A, 5-HT2, 1, and 2 recep-
j
l
r
r
tors, glutamate, glycine and PCP bindings sites of the NMDA recep-
tor, histamine and noradrenalin transporter, cyclooxygenase,
phosphodiesterase and some kinases was investigated. The com-
plete list of considered targets and corresponding radioligands is gi-
ven in Supplementary data. At a concentration of 1–3 lM 25 did not
compete significantly with the employed radioligands. The only
exception is a 56% inhibition of radioligand binding to the calcito-
nine gene-related peptide CGRP1 at a compound concentration of
6. Animal test
3 lM and, furthermore, an IC50-value of 560 nM at hERG chan-
At first the tolerability of phenol 25 was investigated in the
Irwin test with mice.38 In this assay increasing doses of 25
(1–100 mg/kg body weight) were injected intraperitoneally and
then the behavior of the mice was observed for 24 h. The mice
did not show unusual reactions after ip application of the phenol
25. Only diarrhea was observed in some animals 30 min after ip
injection of the highest dose of 100 mg/kg body weight. The phenol
25 seems to be well tolerated by the mice.
Next, the potential of 25 for the treatment of neuropathic pain
was investigated with the formalin test.39 In this assay different
doses (10, 30, and 100 mg/kg) of the NR2B antagonist 25 were in-
jected intraperitoneally to mice. After 15 min, an inflammatory
process was induced upon injection of formalin into the hind
paw. The animal tried to reduce the acute pain (0–5 min) by licking
of the paw. The number of licking activities per second was taken
as a correlate for pain intensity. After 5 min the acute pain phase
decreased and a phase of neuropathic pain developed. In this situ-
ation pain stimuli were set by pressing a von Frey filament on the
inflamed paw of the mice. The reactions of the mice were recorded
for 10–35 min after formalin injection.
nel34,35 was estimated for 25. Altogether, this screening revealed
very high selectivity of 25 for the ifenprodil bindingsite of NR2B con-
taining NMDA receptors over the investigated systems.
These results are of particular interest, since the low receptor
selectivity is one of the major drawbacks of the lead compound 1
(see Section 1). Obviously, the reduced conformational flexibility
of the ethylene spacer in the 3-benzazepine system of 25 signifi-
cantly reduces interaction with competing a1 (IC50 = 5.3 lM), r1
(Ki = 194 nM), r2 (Ki > 10 lM), and PCP (no affinity) receptors. A
similar result had been obtained with the methyl ether 2a.23
4. Functional activity
In order to verify the antagonistic activity of the potent NR2B
selective ligands 25 and cis-32 the inhibition of the excitotoxicity
induced by overactivation of NMDA receptors was investigated.
In this assay L(tk-)-cells stably expressing NR1a/NR2B receptors
were employed. After addition of (S)-glutamate and glycine the cell
damage (excitotoxicity) was determined by measuring the amount
of lactate dehydrogenase (LDH) released from the cells into the cul-
ture supernatant. Different concentrations of the test compounds
25 and cis-32 were added 30 min before addition of (S)-glutamate
and glycine and the released amount of LDH was measured.36
In Figure 2 the inhibition of glutamate/glycine induced excito-
toxicity at different concentrations of phenols 25, cis-32 and
methyl ether 2a is compared. The phenylbutyl derivative 25
reduced the excitotoxicity with an IC50-value of 18.4 nM. This re-
sult clearly shows that 25 is not only binding at the ifenprodil
binding site of NR2B containing NMDA receptors but is also
blocking the NMDA associated ion channel and producing antag-
onistic effects. We speculate that the H-bond donor properties of
the phenolic OH-moiety of 25 are responsible for the ion channel
blockade due to conformational changes of the receptor protein.
The corresponding methyl ether 2a, which is even more potent
in the binding assay, showed an IC50-value of only 360 nM in
the excitotoxicity assay. This result indicates that the methyl
ether 2a is well recognized by the receptor protein, presumably
by a slightly different binding mode, but the ion channel blocking
activity is considerably reduced due to the missing H-bond donor
group.
The phenol 25 was virtually inactive in the acute early pain
phase at all tested doses. However in the late neuropathic pain
phase a dose dependent analgesic activity starting at 30 mg/kg of
25 was observed. At a dose of 10 mg/kg body weight 25 did not
show any analgesic effects. The lowest active dose of 30 mg/kg is
relatively high considering the high in vitro potency of 25, but
the formalin model might not be ideal for NR2B antagonists.
7. Conclusion
Receptor binding studies with [3H]ifenprodil have provided five
phenolic 3-benzazepines interacting with low nanomolar affinity
with NR2B containing NMDA receptors. Four of these compounds
revealed very low selectivity over
r2 receptors. Therefore, only the
4-phenylbutyl substituted derivative 25 was considered for further
pharmacological characterization. In addition to high selectivity
against r1 and r2 receptors 25 did not interact with more than
100 further relevant target systems. Its antagonistic activity
(IC50 = 18.4 nM) was proved in the excitotoxicity assay using
L(tk-)-cells stably expressing NR1a/NR2B receptors. Despite the

B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
8013
Figure 2. Inhibition of NMDA receptor-mediated cell toxicity by compounds 2a, 25, and cis-32. Different concentrations of the potent and selective ligands 2a, 25, and cis-32
were applied to L13-E6 cells and NMDA receptors were activated by addition of (S)-glutamate and glycine. The excitotoxicity was determined by the amount of released
lactate dehydrogenase (LDH). Compounds 2a, 25, and cis-32 have IC50 values of 360 nM, 18.4 nM, and 12.1 lM.
corresponding methyl ether 2a and the conformationally restricted
4-phenylcyclohexyl derivative cis-32 showed similar receptor
affinities, their antagonistic activities were considerably reduced.
tosyl and 5-H tosyl), 7.19 (d, J = 8.4 Hz, 1H, 9-H), 7.31 (d, J = 8.3 Hz,
2H, 2-H tosyl and 6-H tosyl). A signal for the OH-proton is not seen.
Incubation with microsome preparations indicated
a
higher
8.1.2. 7-Benzyloxy-3-(4-tosyl)-2,3,4,5-tetrahydro-3-benzazepin-
1-one (8)
metabolic stability of 25 compared with 1 and in the first animal
studies 25 was well tolerated by mice. In the formalin assay 25
showed dose dependent analgesic activity starting at a relatively
high dose of 30 mg/kg. Altogether these results render 25 a very
promising starting point for the development of potent and selective
NR2B receptor antagonists with improved metabolic stability and
tolerability.
A mixture of 5 (0.61 g, 1.83 mmol), K2CO3 (1.01 g, 7.32 mmol),
benzyl bromide (0.38 g, 2.20 mmol) and acetone (60 mL) was
heated to reflux for 4 h. After removal of K2CO3 by filtration the
solution was concentrated in vacuum. The residue was purified
by fc (5.5 cm, n-hexane/ethyl acetate 7:3, 65 mL, Rf = 0.31) to afford
8 (0.66 g, 85%) as colorless solid, mp 136 °C. C24H23NO4S (421.1).
1H NMR (CDCl3):
d (ppm) = 2.33 (s, 3H, Ph-CH3), 2.94 (t,
J = 6.9 Hz, 2H, 5-H), 3.67 (t, J = 6.7 Hz, 2H, 4-H), 4.17 (s, 2H, 2-H),
5.10 (s, 2H, O–CH2-Ph), 6.71 (d, J = 2.5 Hz, 1H, 6-H), 6.80 (dd,
J = 8.6/2.5 Hz, 1H, 8-H), 7.08 (d, J = 8.1 Hz, 2H, 3-H tosyl and 5-H to-
syl), 7.36–7.45 (m, 8H, 9-H, 2-H tosyl, 6-H tosyl, 2-H phenyl, 3-H
phenyl, 4-H phenyl, 5-H phenyl and 6-H phenyl).
8. Experimental section
8.1. Synthesis general
Unless otherwise noted, moisture sensitive reactions were con-
ducted under dry nitrogen. Flash chromatography (fc): silica gel 60,
8.1.3. ( )-7-Benzyloxy-3-(4-tosyl)-2,3,4,5-tetrahydro-1H-3-
benzazepin-1-ol (9)
40–64 lm (Merck); parentheses include: diameter of the column,
eluent, fraction size, Rf value. 1H NMR (400 MHz), 13C NMR
(100 MHz): Unity Mercury Plus 400 spectrometer (Varian); d in
ppm related to tetramethylsilane; coupling constants are given
with 0.5 Hz resolution. According to HPLC analysis the purity of
all test compounds is greater than 95%. Elemental analyses: Vario
EL (Elementaranalysesysteme GmbH). In addition to the HPLC
analysis the purity of key compounds was proved by elemental
analysis; all values are within 0.4%.
NaBH4 (0.21 g, 5.64 mmol) was added in several portions to a
suspension of 8 (0.40 g, 0.94 mmol) in CH3OH (25 mL) and the mix-
ture was stirred for 6 h at rt. Then H2O (30 mL) was added and the
mixture was extracted with CHCl3 (4 Â 30 mL). The organic layer
was washed with H2O (3 Â 40 mL), dried (Na2SO4) and concen-
trated in vacuum. The residue was purified by fc (3 cm, n-hex-
ane/ethyl acetate 5:5, 10 mL, Rf = 0.59) to afford 9 (0.39 g, 98%) as
colorless solid, mp 132 °C. C24H25NO4S (423.1). 1H NMR (CDCl3):
d (ppm) = 2.32 (s, 3H, Ph-CH3), 2.71–2.77 (m, 2H, OH and 5-H),
3.00–3.06 (m, 1H, 4-H), 3.16–3.21 (m, 2H, 2-H and 4-H), 3.47–
3.60 (m, 2H, 2-H and 5-H), 4.75 (br t, J = 5.6 Hz, 1H, 1-H), 4.94 (s,
2H, O–CH2-Ph), 6.63 (d, J = 2.6 Hz, 1H, 6-H), 6.67 (dd, J = 8.3/
2.6 Hz, 1H, 8-H), 7.14 (d, J = 8.4 Hz, 1H, 9-H), 7.20 (d, J = 8.4 Hz,
2H, 3-H tosyl and 5-H tosyl), 7.31–7.33 (m, 5H, 2-H phenyl, 3-H
phenyl, 4-H phenyl, 5-H phenyl and 6-H phenyl), 7.58 (d,
J = 8.3 Hz, 2H, 2-H tosyl and 6-H tosyl).
8.1.1. 7-Hydroxy-3-(4-tosyl)-2,3,4,5-tetrahydro-3-benzazepin-1-
one (5)
AlCl3 (4.91 g, 36.8 mmol) was added to a solution of methyl ether
423 (1.06 g, 3.07 mmol) in CH2Cl2 (50 mL). The suspension was
heated to reflux for 23 h. After addition of water (60 mL) under
ice-cooling the mixture was stirred for 1 h. The aqueous layer was
separated and extracted with a mixture of CH2Cl2 and CH3OH
((8:2) 3 Â 30 mL). The combined organic layers were dried (Na2SO4)
and concentrated in vacuum. The residue was purified by fc (3 cm, n-
hexane/ethyl acetate 7:3, 30 mL, Rf = 0.14) to afford 5 (0.69 g, 68%) as
colorless solid, mp 198 °C (decomposition). C17H17NO4S (331.4). 1H
NMR (CD3OD): d (ppm) = 2.26 (s, 3H, Ph-CH3), 2.88 (t, J = 6.6 Hz, 2H,
5-H), 3.54 (t, J = 6.7 Hz, 2H, 4-H), 4.15 (s, 2H, 2-H), 6.47 (d, J = 2.4 Hz,
1H, 6-H), 6.51 (dd, J = 8.4/2.4 Hz, 1H, 8-H), 7.06 (d, J = 7.9 Hz, 2H, 3-H
8.1.4. ( )-7-Benzyloxy-2,3,4,5-tetrahydro-1H-3-benzazepin-1-ol
(10)
Mg (144.2 mg, 5.93 mmol) was added to a solution of 9 (114 mg,
0.27 mmol) in CH3OH (10 mL) and the mixture was heated to reflux
for 5 h. Then concentrated H2SO4 (0.33 mL) was added under

8014
B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
ice-cooling and the mixture was filtered. The pH value was adjusted
to pH 9 by addition of NaOH. The aqueous layer was extracted with
CH2Cl2 (5 Â 15 mL), the organic layer was dried (Na2SO4) and
concentrated in vacuum. The residue was purified by fc (2 cm,
CH2Cl2/CH3OH 9.5:0.5 and 2% NH3, 10 mL, Rf = 0.11) to afford 10
(72.7 mg, 83%) as colorless solid, mp 137 °C. C17H19NO2 (269.3). 1H
NMR (CDCl3): d (ppm) = 2.64 (dd, J = 15.5/5.8 Hz, 1H, 5-H), 2.77 (br
t, J = 11.4 Hz, 1H, 4-H), 2.86 (d, J = 12.6 Hz, 1H, 2-H), 3.19–3.33 (m,
3H, 2-H, 4-H, 5-H), 4.58 (d, J = 6.1 Hz, 1H, 1-H), 5.08 (s, 2H,
Ph-CH2–O), 6.70–6.74 (m, 2H, 6-H and 8-H), 7.12 (d, J = 8.0 Hz, 1H,
9-H), 7.30–7.43 (m, 5H, 2-H benzyl, 3-H benzyl, 4-H benzyl, 5-H ben-
zyl and 6-H benzyl). Signals for the OH- and NH-protons are not vis-
ible. Anal. (C17H19NO2) C, H, N.
8.1.7. ( )-3-(4-Phenylbutyl)-2,3,4,5-tetrahydro-1H-3-
benzazepine-1,7-diol (25, WMS-1410)
A suspension of 11 (39 mg, 0.10 mmol) and Pd/C (37 mg, 10%) in
abs. CH3OH (5 mL) was stirred under H2-atmosphere (1 bar) for
30 min at rt. The catalyst was removed by filtration and the solvent
was evaporated in vacuum. The residue was purified by recrystalli-
zation from diisopropyl ether to afford 25 (20.1 mg, 67%) as pale
yellow solid, mp 116 °C. C20H25NO2 (333.4). 1H NMR (CDCl3): d
(ppm) = 1.42–1.62 (m, 4H, N–CH2–CH2–CH2–CH2-Ph), 2.32 (t,
J = 12.0 Hz, 1H, 4-H), 2.43 (d, J = 12.0 Hz, 1H, 2-H), 2.46–2.58 (m,
5H, N–CH2–CH2–CH2–CH2-Ph and 5-H), 2.85–2.91 (m, 1H, 4-H),
3.03–3.15 (m, 2H, 5-H and 2-H), 4.49 (d, J = 6.7 Hz, 1H, 1-H), 7.75
(br s, 1H, 7-OH), 6.43–6.45 (m, 2H, 6-H and 8-H), 8.82 (d,
J = 6.9 Hz, 1H, 9-H), 7.09–7.13 (m, 3H, 2-H phenyl, 4-H phenyl and
6-H phenyl), 7.18–7.23 (m, 2H, 3-H phenyl and 5-H phenyl). A sig-
nal for the OH-proton (C-1) is not visible. Anal. (C20H25NO2) C, H, N.
8.1.5. ( )-7-Benzyloxy-3-(4-phenylbutyl)-2,3,4,5-tetrahydro-1H-
3-benzazepin-1-ol (11)
1-Chloro-4-phenylbutane (39.5 lL, 0.24 mmol) was added to a
mixture of 10 (42.9 mg, 0.16 mmol), CH3CN (5 mL), Bu4NI
(59.1 mg, 0.16 mmol) and K2CO3 (176.9 mg, 1.28 mmol) and the
reaction mixture was heated to reflux for 72 h. Afterwards the
K2CO3 was filtered off and the solvent was evaporated in vacuum.
The residue was purified by fc (2 cm, n-hexane/ethyl acetate 7:3
and 1% N,N-dimethylethylamine, 10 mL, Rf = 0.30) to afford 11
(44.3 mg, 69%) as colorless oil. C27H31NO2 (401.2). 1H NMR (CDCl3):
d (ppm) = 1.43–1.54 (m, 2H, N–CH2–CH2–CH2–CH2-Ph), 1.55–1.64
(m, 2H, N–CH2–CH2–CH2–CH2-Ph), 2.32 (t, J = 11.7 Hz, 1H, 4-H),
2.43 (d, J = 12.0 Hz, 1H, 2-H), 2.51–2.59 (m, 5H, N–CH2–CH2–
CH2–CH2-Ph and 5-H), 2.92 (ddt, J = 12.3/6.2/2.0 Hz, 1H, 4-H),
3.09 (ddd, J = 12.0/6.8/1.8 Hz, 1H, 2-H), 3.15–3.22 (m, 1H, 5-H),
4.49 (d, J = 6.7 Hz, 1H, 1-H), 4.96 (s, 2H, O–CH2-Ph), 6.63–6.66
(m, 2H, 6-H and 8-H), 7.02 (d, J = 9.0 Hz, 1H, 9-H), 7.10–7.13 (m,
3H, CH phenyl), 7.19–7.36 (m, 7H, CH phenyl). A signal for the
OH-proton is not visible.
8.1.8. ( )-cis-3-(4-Phenylcyclohexyl)-2,3,4,5-tetrahydro-1H-3-
benzazepine-1,7-diol (cis-32)
A suspension of cis-23 (340 mg, 0.80 mmol) and Pd/C (200 mg,
10%) in abs. CH3OH (30 mL) was stirred under H2-atmosphere
(1 bar) for 2 h at rt. The catalyst was removed by filtration over Celite
535Ò and the solvent was evaporated in vacuum. The residue was
purified by fc (2 cm, n-hexane/ethyl acetate 6:4 and 1% N,N-dimeth-
ylethylamine, 10 mL, Rf = 0.19) to afford cis-32 (199.3 mg, 74%) as
colorless solid, mp 67 °C. C22H27NO2 (337.5). 1H NMR (CDCl3):
d (ppm) = 1.56–1.66 (m, 4H, CH2 cyclohexyl), 1.67–1.76 (m, 2H,
CH2 cyclohexyl), 2.12–2.17 (m, 2H, CH2, cyclohexyl), 2.40–2.57 (m,
3H, 2-H, 4-H and 5-H), 2.71 (br s, 1H, 1-H cyclohexyl), 2.88 (quint,
J = 4.6 Hz, 1H, 4-H cyclohexyl), 2.97–3.17 (m, 3H, 2-H, 4-H and
5-H), 4.51 (d, J = 6.7 Hz, 1H, 1-H), 6.43 (d, J = 2.2 Hz, 1H, 6-H), 6.46
(dd, J = 8.0/2.4 Hz, 1H, 8-H), 6.88 (d, J = 8.0 Hz, 1H, 9-H), 7.09–7.14
(m, 1H, 4-H phenyl), 7.22–7.27 (m, 4H, 2-H phenyl, 3-H phenyl,
5-H phenyl and 6-H phenyl). Signals for the OH-protons are not
visible.
8.1.6. ( )-cis- and ( )-trans-7-(Benzyloxy)-3-(4-phenyl
cyclohexyl)-2,3,4,5-tetrahydro-1H-3-benzazepin-1-ol (cis-23
and trans-23)
NaBH(OAc)3 (496 mg, 2.34 mmol) was added to a solution of 10
(350 mg, 1.30 mmol), 4-phenylcyclohexanone (272 mg, 1.56 mmol)
8.2. Pharmacological studies
8.2.1. Materials and general procedures
and CH3CO2H (105 lL) in 1,2-dichloroethane (25 mL). The reaction
Centrifuge: High-speed cooling centrifuge model Sorvall RC-5C
plus (Thermo Finnigan). Filter: Printed Filtermat Type B (Perkin–El-
mer), presoaked in 0.5% aqueous polyethylenimine for 2 h at rt be-
mixture was stirred for 16 h at rt. A saturated NaHCO3 solution
(20 mL) and H2O (20 mL) were added, the layers were separated
and the aqueous layer was extracted with CH2Cl2 (3 Â 40 mL). The
combined organic layers were dried (Na2SO4) and the solvent was
evaporated in vacuum. The residue was purified by fc (3 cm, n-hex-
ane/ethyl acetate 9:1 and 1% N,N-dimethylethylamine, 10 mL, Rf
(trans-23) = 0.30, Rf (cis-23) = 0.36).trans-23 (Rf = 0.30): Colorless
solid, mp 147 °C, yield 80.6 mg (14%). C29H33NO2 (427.6). 1H NMR
(CDCl3): d (ppm) = 1.40–1.55 (m, 4H, CH2 cyclohexyl), 1.87–1.99
(m, 4H, CH2 cyclohexyl), 2.40–2.49 (m, 1H, 4-H cyclohexyl), 2.59–
2.68 (m, 4H, 2-H, 4-H, 5-H and 1-H cyclohexyl), 3.10 (ddt, J = 12.3/
5.6/1.9 Hz, 1H, 4-H), 3.21–3.29 (m, 2H, 2-H and 5-H), 4.54 (d,
J = 6.7 Hz, 1H, 1-H), 5.02 (s, 2H, O–CH2-Ph), 6.71 (dd, J = 8.0/2.3 Hz,
1H, 8-H), 6.72 (d, J = 1.9 Hz, 1H, 6-H), 7.09 (d, J = 8.0 Hz, 1H, 9-H),
7.15–7.20 (m, 3H, CH aromatic), 7.26–7.33 (m, 3H, CH aromatic),
7.34–7.42(m, 4H, CH aromatic). A signal for theOH-protonis not vis-
ible.cis-23 (Rf = 0.36): Colorless oil, yield 391.6 mg (71%). C29H33NO2
(427.6). 1H NMR (CDCl3): d (ppm) = 1.57–1.70 (m, 4H, CH2 cyclo-
hexyl), 1.72–1.82 (m, 2H, CH2 cyclohexyl), 2.16–2.24 (m, 2H, CH2
cyclohexyl), 2.43 (t, J = 11.8 Hz, 1H, 4-H), 2.47 (d, J = 12.2 Hz, 1H,
2-H), 2.60–2.68 (m, 2H, 5-H and 1-H cyclohexyl), 2.92 (quint,
J = 4.8 Hz, 1H, 4-H cyclohexyl), 3.08 (ddt, J = 12.4/5.8/2.1 Hz, 1H, 4-
H), 3.17–3.26 (m, 2H, 2-H and 5-H), 4.54 (d, J = 6.7 Hz, 1H, 1-H),
5.01 (s, 2H, O–CH2-Ph), 6.68–6.70 (m, 2H, 6-H and 8-H), 7.06 (d,
J = 8.9 Hz, 1H, 9-H), 7.16–7.21 (m, 1H, 4-H phenyl), 7.28–7.41 (m,
9H, C–H aromatic). A signal for the OH-proton is not visible.
fore use. The filtration was carried out with
a MicroBeta
FilterMate-96 Harvester (Perkin–Elmer). The scintillation analysis
was performed using Meltilex (Type A) solid scintillator (Perkin–
Elmer). The scintillation was measured using a MicroBeta Trilux
scintillation analyzer (Perkin–Elmer). The overall counting effi-
ciency was 20%.
8.2.2. Cell culture and preparation of membrane homogenates
for the NR2B assay28
In the assay mouse L(tk-)-cells stably transfected with the dexa-
methasone inducible eukaryotic expression vectors pMSG NR1a,
pMSG NR2B in a 1:5 ratio were used. The transformed L(tk-)-cells
were grown in Modified Earl’s Medium (MEM) containing 10% of
standardized FCS (Biochrom AG, Berlin, Germany). The expression
of the NMDA receptor at the cell surface was induced after the cell
density of the adherent growing cells had reached approximately
90% of confluency. For the induction, the original growth medium
was replaced by growth medium containing 4
lM dexamethasone
and 4 M ketamine (final concentration). After 24 h the cells were
l
harvested by scraping and pelleted (10 min, 5000 g, Hettich Rotina
35R centrifuge, Tuttlingen, Germany).
For the binding assay, the cell pellet was resuspended in phos-
phate buffer saline (PBS) buffer and the number of cells was deter-

B. Tewes et al. / Bioorg. Med. Chem. 18 (2010) 8005–8015
8015
mined using an improved Neubauer’s counting chamber (VWR,
Darmstadt, Germany). Subsequently, the cells were lysed by soni-
cation (4 °C, 6 Â 10 s cycles with breaks of 10 s, device: Soniprep
150, MSE, London, UK). The resulting cell fragments were centri-
fuged with a high performance cool centrifuge (20,000 g, 4 °C,
Sorvall RC-5 plus, Thermo Scientific). The supernatant was dis-
carded and the pellet resuspended in a defined volume of PBS
yielding cell fragments of approximately 500,000 cells/mL. The
suspension of membrane homogenates was sonicated again (4 °C,
2 Â 10 s cycles with a break of 10 min) and stored at À80 °C.
compound 25. We are also grateful to Professor Dr. D. Steinhilber,
Department of Pharmacy, University of Frankfurt, for donating us
the L(tk-)-cells stably expressing NR1a/NR2B receptor proteins.
Supplementary data
Purity data of all test compounds; physical and spectroscopic
data of all new compounds; screening list of all considered recep-
tors, transporters and enzymes; general chemistry methods;
details of the pharmacological assays. Supplementary data associ-
ated with this article can be found, in the online version, at
doi:10.1016/j.bmc.2010.09.026.
8.2.3. Performing of the NR2B binding assay28
The competitive binding assay was performed with the radioli-
gand [3H]ifenprodil (60 Ci/mmol; Perkin Elmer) using standard
96-well-multiplates (Diagonal, Muenster, Germany). The thawed
References and notes
1. Kennedy, M. B. Science 2000, 290, 750.
2. Bräumer-Osborne, H.; Egebjerg, J.; Nielsem, E. O.; Madson, U.; Krogsgaard-
Larsen, P. J. Med. Chem. 2000, 43, 2609.
cell membrane preparation (about 20 lg protein) was incubated
with 6 different concentrations of test compounds, 5 nM [3H]ifen-
prodil, and TRIS/EDTA-buffer (5 mM/1 mM, pH 7.5) in a total vol-
3. Stark, H.; Graßmann, S.; Reichert, U. Pharm. UZ. 2000, 29, 228.
4. Furukawa, H.; Singh, S. K.; Mancusco, R.; Gouaux, E. Nature 2005, 438, 185.
5. Paoletti, R.; Neyton, J. Curr. Opin. Pharmacol. 2007, 7, 39.
6. Goebel, D. J.; Poosch, M. S. Mol. Brain Res. 1999, 69, 164.
7. Kristensen, J. D.; Svensson, B.; Gordh, T., Jr. Pain 1992, 51, 249.
8. Eide, K.; Stubhaug, H.; Oye, I.; Breivik, H. Pain 1995, 61, 221.
9. Knox, D. J.; McLeod, B. J.; Goucke, C. R. Anaesth. Intensive Care 1995, 23, 620.
10. Max, M. B.; Byas-Smith, M. G.; Gracely, R. H.; Bennett, G. J. Clin.
Neuropharmacol. 1995, 18, 360.
11. Boyce, S.; Wyatt, A.; Webb, J. K.; O‘Donnell, R.; Mason, G.; Rigby, M.;
Sirinathsinghji, D.; Hill, R. G.; Rupniak, N. M. J. Neuropharmacology 1999, 38,
611.
12. Wessell, R. H.; Ahmed, S. M.; Menniti, F. S.; Dunbar, G. L.; Chase, T. N.; Oh, J. D.
Neuropharmacology 2004, 47, 184.
ume of 200 lL for 120 min at 37 °C. The incubation was
terminated by rapid filtration through the presoaked filtermats
by using the cell harvester. After washing each well five times with
300 lL of water, the filtermats were dried at 95 °C. Subsequently,
the solid scintillator was placed on the filtermat and melted at
95 °C. After 5 min, the solid scintillator was allowed to solidify at
rt. The bound radioactivity trapped on the filters was counted in
the scintillation analyzer. The non-specific binding was deter-
mined with 10 l
M unlabeled 1. The Kd-value of 1 is 7.6 nM.28
13. Yurkewicz, L.; Weaver, J.; Bullock, M. R.; Marshall, L. F. J. Neurotrauma 2005, 22,
1428.
14. Di, X.; Bullock, R.; Watson, J.; Fatouros, P.; Chenard, B.; White, F.; Corwin, F.
Stroke 1997, 28, 2244.
15. Menniti, S. F.; Pagnozzi, J. M.; Butler, P.; Chenard, B. L.; Jaw-Tasi, S. S.; White, W.
F. Neuropharmacology 2000, 39, 1147.
8.2.4. Affinity towards r1 and r2 receptors and the PCP binding
site of the NMDA receptor
The receptor binding studies were performed as previously
described.30–33
16. Nagy, J.; Horvath, C.; Farkas, S.; Kolok, S.; Szombathelyi, Z. Neurochem. Int.
2004, 44, 17.
8.2.5. Excitotoxicity assay36
17. Gitto, R.; De Luca, L.; Ferro, S.; Occhiuto, F.; Samperi, S.; De Sarro, G.; Russo, E.;
Ciranna, L.; Costa, L.; Chimirri, A. Chem. Med. Chem. 2008, 3, 1539.
18. Borza, I.; Domány, G. Curr. Top. Med. Chem. 2006, 6, 687.
19. Williams, K. Mol. Pharmacol. 1993, 44, 851.
Glutamate-induced cytotoxicity was determined as described
recently.36 Briefly, L(tk-)-cells stably expressing either NR1-1a/
NR2A (L12-G10 cells) or NR1-1a/NR2B (L13-E6 cells) were preincu-
20. Scatton, B.; Avenet, P.; Benavides, J.; Carter, C.; Duverger, D.; Oblin, A.; Perrault,
G.; Sanger, D. J.; Schoemaker, H. Direct Allosteric Control Glutamate Recept. 1994,
139.
21. Kew, J. N. C.; Kemp, J. A. Psychopharmacology 2005, 179, 1039.
22. Montastruc, J. L.; Rascol, O.; Senard, J. M. Neurosci. Biobehav. Rev. 1997, 21, 477.
23. Tewes, B.; Frehland, B.; Schepmann, D.; Schmidtke, K.-U.; Winckler, T.;
Wünsch, B. Chem. Med. Chem. 2010, 5, 687.
24. Abdel-Magit, A. F.; Mehrmann, S. J. Org. Process Res. Dev. 2006, 10, 971.
25. Boeckmann, R. K.; Shao, P.; Mullins, J. J. Org. Synth. 2000, 77, 141.
26. Coughenour, L. L.; Cordon, J. J. J. Pharmacol. Exp. Ther. 1997, 280, 584.
27. Ilyin, V. I.; Whittemore, E. R.; Guastella, J.; Weber, E.; Woodward, R. M. Mol.
Pharmacol. 1996, 56, 1541.
28. Schepmann, D.; Frehland, B.; Lehmkuhl, K.; Tewes, B.; Wünsch, B. J. Pharm.
Biomed. Anal. 2010, 53, 603.
29. Steinmetz, R. D.; Fava, E.; Nicotera, P.; Steinhilber, D. J. Neurosci. Methods 2002,
113, 99.
30. Wirt, U.; Schepmann, D.; Wünsch, B. Eur. J. Org. Chem. 2007, 462.
31. Maier, C. A.; Wünsch, B. J. Med. Chem. 2002, 45, 438.
32. Oberdorf, C.; Schepmann, D.; Vela, J. M.; Diaz, J. L.; Wünsch, B. J. Med. Chem.
2008, 51, 6531.
33. Holl, R.; Schepmann, D.; Fröhlich, R.; Grünert, R.; Bednarski, P. J.; Wünsch, B. J.
Med. Chem. 2009, 52, 2126.
bated with compounds (concentration range 1 nM–10
solved in DMSO) for 30 min. Then a mixture of (S)-glutamate and
glycine (10 M each) was added and the cells were further incu-
bated for 4 h. Excitotoxicity was determined by detection of LDH
release from the cells into the culture supernatants. In this assay
0% excitotoxicity is defined as LDH release in the absence of NMDA
lM, dis-
l
receptor agonist and in the presence 100 lM ketamine) and a
positive control (100% excitotoxicity is defined as LDH release after
addition of (S)-glutamate/glycine in the absence of ketamine).
Experiments were repeated at least three times with six replicates
per test concentration. IC50 values were calculated using GraphPad
Prism v5.0. A direct inhibitory effect of compound 25 and cis-32 on
LDH activity was excluded by performing the excitotoxicity assay
in the presence of test compound as described above, however,
all cells were lysed by addition of 1% triton X-100 prior to quanti-
fication of LDH activity.36 No effect of 25 and cis-32 was observed
under this condition. Under standard assay conditions 25 and cis-
32 had no effect on L(tk-)-untransfected cells.
34. Plásek, J.; Sigler, K. J. Photochem. Photobiol. 1996, 33, 101.
35. Warnke, J. W.; Ganetzky, B. Proc. Nat. Acad. Sci. 1994, 91, 3438.
36. Rook, Y.; Schmidtke, K.-U.; Gaube, F.; Schepmann, D.; Wünsch, B.; Heilmann, J.;
Lehmann, J.; Winckler, T. J. Med. Chem. 2010, 53, 3611.
37. Wiese, C.; Große Maestrup, E.; Schepmann, D.; Vela, J. M.; Holenz, J.;
Buschmann, H.; Wünsch, B. J. Pharm. Pharmacol. 2009, 61, 631.
38. Irwin, S. Psychopharmacology 1968, 13, 222.
Acknowledgments
We are grateful to thank Schwarz Pharma AG, Monheim for
supporting this work and performing the receptor screening of
39. Tsuda, M.; Ueno, S.; Inoue, K. Br. J. Pharmacol. 1999, 128, 1497.