Cycloalkyl[b][1,4]benzodiazepinoindoles are agonists at the human 5-HT 2C receptor

Article abstract of DOI:10.1016/j.bmcl.2004.02.100

Evaluation of selected compounds from our Corporate Compound Library in a human 5-HT_2C receptor binding assay led to the discovery of WAY-629, a cyclohexyl[b][1,4]benzodiazepinoindole (K_i 56nM, E_max 90%), which is selectiv

Full text of DOI:10.1016/j.bmcl.2004.02.100

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2603–2607
Cycloalkyl[b][1,4]benzodiazepinoindoles are agonists at the human
5-HT2C receptor
Annmarie L. Sabb,a,* Robert L. Vogel,a Gregory S. Welmaker,a Joan E. Sabalski,a
Joseph Coupet,b John Dunlop,b Sharon Rosenzweig-Lipsonb and Boyd Harrisona
aChemical and Screening Sciences, Wyeth Research, CN 8000, Princeton, NJ 08543-8000, USA
bWyeth Neuroscience, Wyeth Research, CN 8000, Princeton, NJ 08543-8000, USA
Received 7 January 2004; accepted 18 February 2004
Abstract—Evaluation of selected compounds from our Corporate Compound Library in a human 5-HT2C receptor binding assay
led to the discovery of WAY-629, a cyclohexyl[b][1,4]benzodiazepinoindole (Ki 56 nM, Emax 90%), which is selective for the 5-HT2C
receptor versus other serotonin receptor subtypes, and dopamine, histamine, adrenergic, and muscarinic receptors. In addition,
WAY-629 was active in vivo in a rat model of feeding behavior. An SAR study based on WAY-629 led to compound 11 (Ki 13 nM,
Emax 102%).
Ó 2004 Elsevier Ltd. All rights reserved.
The serotonin 5-hydroxytryptamine 2C (5-HT2C)
receptor is a member of the seven transmembrane
spanning G-protein coupled 5-HT receptor family,
which has been divided into seven subfamilies (5-HT1–5-
tive, partial agonist (IL639, intrinsic activity 57%) has
been described (Fig. 1).3
Studies in several animal species and in humans have
shown that the nonselective 5-HT2B=2C receptor agonist,
meta-chlorophenylpiperazine (m-CPP) (Fig. 2) decreases
food intake.4 m-CPP is a metabolite of trazodone.
Furthermore, Tecott et al. have demonstrated that
HT7). The 5-HT2 receptor subfamily consists of 5-HT2A
,
5-HT2B, and 5-HT2C receptors, which are positively
coupled to the second messenger, inositol monophos-
phate. Although these receptor subtypes are similar in
amino acid sequence and pharmacology, there are dis-
tinct differences in their distribution. Whereas the
5-HT2A and 5-HT2B receptors are widely distributed in
the periphery and only sparingly in the CNS, the 5-HT2C
receptor has been found only in the CNS in humans.1
The presence of 5-HT2C receptors in many human brain
regions (cerebral cortex, cerebellum, and substania
nigra) and the spinal cord suggests that novel selective
5-HT2C agonists may provide therapeutic benefit with
reduced EPS liability in the treatment of CNS disorders.
N
N
CH3
N
H
N
H
N
O
O
O
N
N
CF3
CF3
CH3
F
SB-228357
OCH3
SB-243213
NH
Although the discovery of two 5-HT2C receptor anta-
gonists with selectivity versus 5-HT2A and 5-HT2B
receptors has been reported (SB-228357 and SB-
243213)2 the search for potent, efficacious, and selective
5-HT2C receptor agonists remains a major focus of
several research laboratories. Recently, a potent, selec-
F
F
S
IL639
* Corresponding author. Tel.: +1-732-274-4411; fax: +1-732-274-4505;
e-mail: sabba@wyeth.com
Figure 1.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.02.100

2604
A. L. Sabb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2603–2607
Table 2. WAY-629 is inactive at other receptors
Cl
H
a1
Adrener-
gic b1
Adrener-
gic b2
D1
H1
M
N
N
36%
15%
8%
0%
0%
9%
N
N
M is muscarinic.
H
mCPP
WAY-629
available starting materials or from intermediates that
can be prepared using literature procedures.
Figure 2.
transgenic mice lacking the 5-HT2C receptor eat more
and are heavier than Wild Type mice.5
Scheme 1 shows the preparation of such key interme-
diates and their use in the synthesis of analogues of this
series. According to Scheme 1, an isatoic anhydride is
allowed to react with a glycine ester hydrochloride in the
presence of pyridine or triethylamine to give an open
chain intermediate, which is then cyclized by heating in
the presence of acetic acid or sulfuric acid.9 The benzo-
diazepinediones were reduced with either lithium alu-
minum hydride or 1 M borane in tetrahydrofuran to
give the corresponding benzodiazepines. The benzylic
nitrogen was protected by treatment with acetic anhy-
dride in the presence of triethylamine to give the acetyl
benzodiazepines. Treatment of these intermediates with
aqueous sodium nitrite in acetic acid solution gave
the N-nitroso intermediates, which were reduced to the
hydrazines using powdered zinc and acetic acid.
In an effort to discover potent and selective 5-HT2C
receptor agonists for the treatment of obesity, evalua-
tion of selected compounds from our Corporate Data-
base was undertaken. From that screening, WAY-629
was identified as a potent 5-HT2C agonist6 (Fig. 2).
WAY-629 and its analogues were evaluated in a number
of receptor binding assays and functional activity mea-
sures done in CHO cells transfected with human
receptors. The transporter binding was done in the rat.7
As a lead molecule, WAY-629 showed the selectivity we
were seeking, having 45-fold selectivity for the 5-HT2C
receptor versus the 5-HT2A receptor in binding and 610-
fold selectivity for the 5-HT2C receptor versus the
5-HT2A receptor in phosphoinositide turnover. In order
to test for in vivo activity, a model of reduced feeding
behavior in rats was used.8 WAY-629 was found to
reduce feeding behavior in rats with an ED50 21 mg/kg,
ip. The drug was tested up to 30 mg/kg. At 30 mg/kg the
reduction in feeding behavior was 53%. Table 1 sum-
marizes the affinity and intrinsic activity of WAY-629 at
5-HT2C receptors and shows its selectivity versus a
number of other serotonin receptor subtypes. Affinity
for nonserotonin neurotransmitter receptors [a1,
adrenergic b1, adrenergic b2, dopamine D1, histamine
H1, and muscarinic] is found in Table 2. WAY-629 was
inactive at a number of ion channels (Kþ, ATP-Sensitive
Kþ, Ca2þ Act., Voltage Insensitive (each 10% at 1 lM)]
according to Nova Screen.
A Fischer indole synthesis with cycloalkanones gave the
acetyl protected cycloalkyl[b][1,4]diazepinoindoles. The
cyclohexyl through cyclooctyl derivatives were depro-
tected by stirring with 6 N HCl at room temperature
overnight. The cyclopentyl derivatives were acid sensi-
tive and were deprotected by heating with aqueous
sodium hydroxide in methanol.
Compounds 1–11 were evaluated for agonist binding
affinity at the human 5-HT2C and 5-HT2A receptors.
Compounds of interest were also evaluated for agonist
efficacy. The SAR study began by adding substitution to
the benzene ring of WAY-629. The results are summa-
rized in Table 4. Compounds 3, 4, and 5 show the effect
of substitution on the benzene ring. Little difference in
5-HT2C affinity was observed between the effect of an
electron donating group (3) or an electron withdrawing
group (4) in position 6 of the benzene ring. The reduc-
tion in affinity in both cases is most likely due to steric
effects. Replacement of the proton in position 7 with
fluorine (5) reduced 5-HT2C affinity by 2-fold versus the
unsubstituted WAY-629.
In this paper we report the synthesis and structure–
activity relationship (SAR) studies around WAY-629.
We discovered a series of cycloalkyl[b][1,4]benzodiaze-
pinoindoles, which are 5-HT2C agonists. One of these, 11
(WAY-470), is a more potent and efficacious agonist of
the human 5-HT2C receptor than WAY-629.
The cycloalkyl[b][1,4]benzodiazepinoindole compounds
shown in Table 3 were prepared from commercially
Keeping the benzene ring unsubstituted and adding
gem-dimethyl substituents to position 9 on the cyclo-
Table 1. Affinity, function, and selectivity of WAY-629
5-HT2c
5-HT2A
5-HT1A
5-HT6
5-HT7
Ki
EC50
Emax
Ki
EC50
Emax
@ 1 lM
Ki
Ki
a
a
56 nM
426 nM
90%
2530 nM
260,000 nM
60%
20%
1575 nM
815 nM
a % Of the serotonin response where serotonin is 100%.

A. L. Sabb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2603–2607
Table 3. Cycloalkyl[b][1,4]benzodiazepinoindoles10
2605
10
11
10
R1
2
N
1
R1
2
1
9
{
N
3 H
{
9
N
N
H
3
8
n
8 R2
4
4
5
7
5
7
R
R
6
6
I
II
Compound
I or II
N
R
R1
R2
1
I
Na
Na
1
H
H
H
2
I
6-CH3
6-CH3
6-Cl
7-F
H
H
H
3
II
II
II
II
II
II
II
II
II
H
H
4
1
H
H
5
1
H
H
9,9-DiCH3
8,8,10,10-TetraCH3
6
1
H
7
1
H
H
8
1
H
2-S-CH3
2-R-CH3
H
H
H
H
H
9
1
H
10
11
2
H
3
H
H
O
O
R1
O
R1
NH
R1
OEt
b
a
O
OEt
HCl
R
N
R
R
H2N
H
+
O
O
N
H
N
NH2
O
O
H
O
N
O
NH
CH3
R1
d
CH3
c
R
R1
e
N
R
N
H
R
R1
N
H
N
N
O
O
O
CH3
R1
CH3
R1
N
f
g
N
R
R
N
O
N
H2N
R2
R2
{
{
n
{
n
{
NH
h
R
R1
N
R2
{
{
n
Scheme 1. Reagents and conditions: (a) pyridine, reflux, 24 h, 40–50% yield; (b) acetic acid or sulfuric acid 24 h, 40–75% yield; (c) LAH or 1 M BH3 in
THF, 24 h, aq HCl workup 65–75% yield; (d) acetic anhydride, TEA, ether, rt 24 h, 90% yield; (e) NaNO2, aq HCl; (f) powdered Zn in aqueous acetic
acid, 25–35 °C; (g) cycloalkanone, acetic acid, 1.5 h, 75–90 °C, 20–60% yield; (h) NaOH, aq MeOH (free base) or concd HCl (isolated as HCl salt).
hexyl ring, compound 6, reduced 5-HT2C affinity sug-
gesting that disubstitution in the 9-position was not
tolerated. However, the tetramethyl analogue of WAY-
629 (7) was nearly equipotent to WAY-629 at the
5-HT2C receptor. Addition of a methyl group at position
2 of the diazepine ring of WAY-629 reduced affinity at
the 5-HT2C receptor in each enantiomer but the R iso-
mer 9 is better tolerated than the S isomer 8.

2606
A. L. Sabb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2603–2607
Table 4. Affinity and function of cycloalkyl[b][1,4]benzodiazepinoin-
dole analogues at 5-HT2C and 5-HT2A receptors
rat model of feeding behavior, showing reduce feeding
behavior with an ED50 of 21 mg/Kg, ip. Substitution on
any ring of the WAY-629 structure reduced affinity for
the 5-HT2C receptor.
Ki, nM agonist binding
7;a
Compound
Function: (%Emax)
5-HT2C
5-HT2A
The cyclopentyl, cycloheptyl, and cyclooctyl analogues
of WAY-629 (1, 10, and 11) are also 5-HT2C agonists.
The cyclooctyl analogue 11 (WAY-470) was found to
be the most potent and efficacious (Ki 13 nM, Emax
102%) 5-HT2C agonist in this series.
1
97 (90)
985
599
922 (9)
2531
1231
Ndb
2
3
4
358
111
5
Ndb
6
243
79
Ndb
152
7
8
452
189
Ndb
References and notes
9
1559
199
36 (80)c
10
11
38 (73)
13 (102)
1. Hoyer, D.; Clarke, D. E.; Fozard, J. R.; Hartig, P. R.;
Martin, G. R.; Mylecharane, E. J.; Saxena, P. R.;
Humphrey, P. P. A. Pharmacol. Rev. 1994, 46, 157.
2. Robichaud, A. J.; Largent, B. L. Annu. Rep. Med. Chem.
2000, 5, 11.
a % Emax not listed was not done.
b Nd ¼ not done.
c EC50 64,000 nM.
3. Fitzgerald, L. W.; Ennis, M. D. Annu. Rep. Med. Chem.
2002, 37, 21.
4. Martin, G. E.; Elgin, R. J., Jr.; Mathiasen, J. R.; Davis, C.
B.; Kesslick, J. M.; Baldy, W. J.; Shank, R. P.; DiStefano,
D. L.; Fedde, C. L.; Scott, M. K. J. Med. Chem. 1989, 32,
1052.
5. Tecott, L. H.; Sun, L. M.; Akana, S. F.; Strack, A. M.;
Lowenstein, D. H.; Dallman, M. F.; Julius, D. Nature
1995, 374, 542.
6. Kim, D. J. Heterocycl. Chem. 1975, 12, 1323; Kim, D.
U.S. Patent 3,914,250, 1975, CAN 84:31150.
Subsequently we examined the effect of the size of the
cycloalkyl ring on 5-HT2C affinity. Reducing the cyclo-
hexyl ring of WAY-629 to cyclopentyl gave a somewhat
less potent 1 (Ki 97 nM), whereas increasing the size of
the cycloalkyl ring from cyclohexyl to cycloheptyl
improved the 5-HT2C affinity of 10 (Ki 38 nM), although
the compound was a weaker agonist than WAY-629
(73% efficacy for 10 vs 90% efficacy for WAY-629).
7. Pharmacology: Affinity and function: Receptor binding
studies and functional activity measures were done in
CHO cells stably transfected with human receptors, with
the exception of the transporter binding, which was also
done in rat.
The cyclooctyldiazepinoindole analogue 11 was found
to have potent affinity at the 5-HT2C receptor (Ki 13 nM)
and to be a full agonist (102% efficacy). The selectivity
profile of 11 versus 5-HT1A, 5-HT2A 5-HT2B, dopamine
D3, D4, a1, and the rat 5-HT transporter is summarized
in Table 5. Although the binding affinity of 11 is sepa-
rated by 3-fold from binding to the 5-HT2A receptor, a
372-fold separation is achieved in function. The affinity
of 11 to the 5-HT2B receptor is >5000 nM. In addition,
11 (WAY-470) is selective in binding versus other
receptors tested and versus the rat serotonin transporter.
Affinity: Inhibition of [125I] DOI binding was used to
measure affinity to 5-HT2C (VNI isoform) and 5-HT2A
receptors; Inhibition of 8-OH-DPAT binding was used to
measure affinity to 5-HT1A receptors; Inhibition of
[3H]LSD binding was used to measure affinity to 5-HT6
and 5-HT7 receptors; Inhibition of [3H]spiperone binding
to used to measure affinity; Inhibition of [3H]paroxetine
binding was used to measure affinity to the serotonin
transporter; Inhibition of a1 adrenergic receptor binding
was used to measure a1 binding.
In summary, a series of cycloalkyl[b][1,4]benzodiaze-
pinoindoles has been discovered, which are agonists at
the human 5-HT2C receptor. Our lead molecule, WAY-
629, 1,2,3,4,8,9,10,11-octahydro[1,4]diazepino[6,5,4-jk]-
carbazole is a 5-HT2C agonist selective for the 5-HT2C
receptor versus several other serotonin receptor sub-
types, the serotonin transporter, and other nonserotonin
receptors and ion channels. WAY-629 is also active in a
Function: Stimulation of [3H]inositol monophosphate
production was used to measure the degree of agonism
in 5-HT2C and 5-HT2A receptors.
8. Feeding behavior in rats: Eight (8) male Sprague–Dawley
rats weighing 150–180 g were separated into individual
cages and acclimated to a powdered diet for 2 weeks.
During this period and throughout the test procedure, the
food cup and the animals were weighed daily. Following
the acclimation period, animals were fasted for 24 h and
then injected with either vehicle or one of 4 doses of the
test compound. Food intake was assessed at 2 and 24 h
following compound administration. Compounds to be
evaluated were injected 1–2 ꢀ per week until all animals
had received all doses of the test compound. The order of
doses were chosen using a modified Latin Square design.
Additional studies may be conducted in satiated rats at the
start of the dark cycle. Compounds were injected ip, sc, or
given po. At the end of the study effects of the test
compound on food intake were evaluated using a repeated
measures ANOVA. Data collected were for 2 h food
intake (g). Data were subjected to a one-way ANOVA
with posthoc t-test to assess group differences. Where
Table 5. Profile of 11 (WAY-470)7
H
N
N
Inhibition @ 1 lM
5-HT1A
4%
5-HT2B (Ki)
>5000 nM
r5-HT
14%
D3
D4
a1 (Ki)
37%
47%
665 nM

A. L. Sabb et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2603–2607
2607
appropriate, ED50 values were calculated. The ED50
value is the dose that produces a 50% reduction in food
intake.
10. All new compounds gave consistent 400 MHz 1H NMR
spectra, mass spectral data and combustion analysis. For
experimental procedures, see: Sabb, A. L.; Vogel, R. L.;
Welmaker, G. S.; Sabalski, J. E. WO 02/36596 A2, 2002;
CAN 136:355252.
9. Certain intermediates cyclized without the necessity of
further heating in acid during the test period.