5-HT2C antagonists based on fused heterotricyclic templates: Design, synthesis and biological evaluation

Article abstract of DOI:10.1016/j.bmcl.2006.10.034

Design, synthesis and properties of a new tricyclic series of selective 5-HT_2C receptor antagonists are reported. Conformational analysis of a 2-phenyl-dihydropyrrolone scaffold suggested that ring fusion, locking coplanarity between the rings

Full text of DOI:10.1016/j.bmcl.2006.10.034

Bioorganic & Medicinal Chemistry Letters 17 (2007) 424–427
5
-HT antagonists based on fused heterotricyclic templates:
Design, synthesis and biological evaluation
2C
a,
a
a
a
*
Dieter Hamprecht, Fabrizio Micheli, Giovanna Tedesco, Daniele Donati,
a
a
b
Marcella Petrone, Silvia Terreni and Martyn Wood
a
GlaxoSmithKline, Medicine Research Centre, Via Fleming, 4, 37135 Verona, Italy
GlaxoSmithKline Pharmaceuticals, Third Avenue, Harlow, Essex, CM19 5AW, UK
b
Received 30 August 2006; revised 10 October 2006; accepted 12 October 2006
Available online 17 October 2006
Abstract—Design, synthesis and properties of a new tricyclic series of selective 5-HT2C receptor antagonists are reported. Confor-
mational analysis of a 2-phenyl-dihydropyrrolone scaffold suggested that ring fusion, locking coplanarity between the rings of this
moiety, might be tolerated by the 5-HT2C receptor. An interesting effect of this is the change of the nature of the carbon–carbon
double bond of the lactam ring from vinylic to aromatic. The changes were found to result in a favourable profile at both, receptor
and in vivo level.
Ó 2006 Elsevier Ltd. All rights reserved.
5
The central role of the serotonergic neurotransmitter
system in controlling CNS function of importance in
the pathology of depression and anxiety disorders is well
established. The gold standard antidepressant treat-
ments, the selective serotonin reuptake inhibitors
In a previous communication we described a series of
compounds exemplified by 1 (Fig. 1) which combine
potency at the 5-HT2C receptor with good selectivity
for this target and a favourable profile in vivo.
(
SSRIs) exert their efficacy with some lag time via this
Given the large interest in compounds such as 1, we
started to analyse conformational aspects of the tem-
plate. Conformational analysis for the scaffold of 1
was carried out using program Macromodel within
Maestro, starting from a 3D structure generated with
7
Corina. One thousand steps of Low-Mode conforma-
system by modulating synaptic serotonin (5-hydroxy-
tryptamin, 5-HT) levels and thus effecting activity of
the 5-HT receptor subtypes in a non-specific way. More
selective action on 5-HT receptor subtypes may be
expected to avoid some of the shortcomings of SSRIs,
6
1
such as side effects and delayed onset of action. In this
tional search were carried out using the MMFFs force
field in combination with the GB/SA hydration model.
Energy minimisations were carried out using the trun-
cated Newton conjugate gradient algorithm (500 steps,
0.005 convergence threshold). Minimised structures with
energy higher than 50 kJ/mol above the global minimum
were discarded. Redundant conformations were re-
moved by heavy atom superimposition (max distance
context the 5-HT2C receptor has been recognised as a
target of potentially significant therapeutic interest.
Thus, for example, 5-HT2C receptor mediated responses
are decreased as a result of chronic SSRI treatment sug-
gesting that these receptors contribute to mediating the
2
therapeutic effect. Furthermore, the moderately selec-
tive 5-HT2C agonist meta-chlorophenylpiperazine
˚
(
mCPP) has anxiogenic properties, suggesting that
3
between atoms in equivalent structures <0.25 A).
5-HT2C antagonism would result in an anxiolytic effect,
as has indeed been demonstrated in animal models of
4
anxiety.
Cl
Cl
O
O
O
N
Keywords: Serotonin receptors; 5-HT2c receptors; 5-HT2c receptor
antagonists; Fused rings; Vinylic and aromatic double bonds; Antide-
pressant; Anxiolytic agent.
N
1
*
Corresponding author. Tel.: +390458219733; fax: +390458218196;
e-mail: dieter.w.hamprecht@gsk.com
Figure 1. Structure of a previously described 5-HT2C antagonist.
0
960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.10.034

D. Hamprecht et al. / Bioorg. Med. Chem. Lett. 17 (2007) 424–427
425
The thus obtained global minimum conformer shows a
N
O
rather small dihedral angle between the lactam and the
adjacent LHS phenyl ring (37°). We observed that a
similar (coplanar) conformation could be locked by
introducing cyclisation from the lactam beta position
to the phenyl ortho-position by a ca. 1-atom bridge such
as a substituted nitrogen atom. Taken together this sug-
gested that, if the conformation most relevant for bind-
ing at the target receptor was similar to the minimum
energy conformer, tricyclic indole derivatives such as 5
might be able to bind to the 5-HT2C receptor (Fig. 2).
We furthermore hypothesized that inclusion of the car-
bon–carbon double bond of the lactam in an aromatic
ring, by profoundly changing the chemical nature of this
fragment, could modulate the overall properties of the
resulting compounds in a useful way.
.
HCl
H N
O
2
CO Et
CO Et
2
2
a
4
b
N
N
Cl
2
3
O
O
N
O
N
N
.
HCl
5
Scheme 1. Reagents and conditions: (a) NCS (1.1 equiv), (PhCO
0.05 equiv), CCl
, reflux (94%); (b) DMF, dioxane, 2,6-lutidine, 95 °C
(41%); then Me O, rt (87%).
Al, DCM, 25 °C; then HCl, Et
2 2
)
(
4
To test this concept we prepared the tricyclic indole
derivative 5 as a prototype of this new class of ligands.
Radical halogenation of ethyl 1,2-dimethylindole-3-car-
boxylate 2 gave the 2-chloromethyl intermediate 3 with
the indole 2- and 3-substituents already in the desired
3
2
powder/ammonium chloride gave mainly the N-hydroxy
indole derivative besides only minor amounts of 9. This
problem was overcome when TiCl was employed as the
reducing agent to give indole 9 in good yield (Scheme 2).
N-Methylation, radical chlorination and condensation
with aniline 4 as outlined in Scheme 1 gave the desired
dichlorinated product 10. It is interesting to note that
tert-butyl ester analogues of intermediate 8, prepared
by arylation of acetyl acetate tert-butyl ester, were less
suitable substrates for the final condensation reaction
oxidation stages. Heating of compound 3 with aniline
5
3
4
resulted in N-alkylation of the anilinic nitrogen. Cyc-
lisation with condensation was subsequently achieved in
the presence of trimethylaluminum to provide indolyl
dihydropyrrolone 5 in moderate yield (Scheme 1).
We were pleased to find that compound 5 displayed po-
tent affinity for the 5-HT2C receptor (pK = 8.2) with
i
selectivity over the 5-HT
8
(pK = 6.3) and 5-HT
i
pK = 6.7) subtypes. Encouraged by these results we
i
11
2
A
2B
with aniline 4.
(
decided to explore substitution of the indolyl phenyl
ring, initially focussing on chloro substitution that had
been found to be optimal in 1. We adapted the elegant
The route outlined in Scheme 2 allowed the preparation
of a range of substituted indolyl dihydropyrrolone ana-
logues. As we had previously demonstrated improved
properties when a methyl substituent was introduced
9
oxindole synthesis described by Gruda to the synthesis
of the required 2-alkyl indole 3-carboxylate intermedi-
ates, as recently also reported by Bunce and cowork-
5
into the 4-position of the piperidine moiety a number
of examples with this modification were included.
Results are summarised in Table 1. It becomes apparent
that 6,7-dichloro substitution in compound 10 does not
increase affinity at the 5-HT2C receptor in this series.
From the mono chloro substituted examples a slight
preference can be seen for the 7-substituent in 12 which
results in an increase in potency and better selectivity
over the 5-HT2B isoform. In the fluoro substituted
1
0
ers.
Thus, ortho-fluoronitrobenzene derivative 6
reacted smoothly with methyl acetylacetate (7) to give
intermediate 8. Reduction of the nitro group with iron
CO Me
2
Cl
Cl
F
Cl
Cl
O
O
a
+
OH
NO
OMe
NO2
2
6
7
8
CO Me
2
Cl
Cl
b
steps
10
N
H
9
Figure 2. Molecular overlay of truncated forms of 1 and 5 showing
high similarity in the spatial display of key features and a high degree
of shape similarity (superimposition of topologically correspondent
heavy atoms: RMSD = 0.51).
Scheme 2. Reagents and conditions: (a) 7 (2.1 equiv), NaH (2.1 equiv),
DMF, 0 °C to rt (85%); (b) TiCl in aq HCl, HOAc, dioxane, 90 °C
3
(79%).

426
D. Hamprecht et al. / Bioorg. Med. Chem. Lett. 17 (2007) 424–427
8
Table 1. Affinity results
O
O
7
R
N
6
X
O
N
3
N
5
X
Compound
R = H
Compound
R = Me
pK
i
5-HT2C
pK
i
5-HT2A/B
pK
i
5-HT2C
i
pK 5-HT2A/B
6
6
7
5
6
7
7
3
,7-Bis Cl
10
11
8.2
8.3
8.6
6.3/6.7
6.5/6.8
6.8/<6.0
n.d.
15
-Cl
-Cl
-F
8.4
8.5
8.6
7.8
8.6
6.6/6.7
6.9/7.4
6.6/7.5
6.6/6.6
6.8/7.2
12
16
17
n.d.
13
-F
-F
7.9
7.5
<5.0/<5.0
6.0/6.1
18
19
n.d.
14
-OMe
-Me
n.d.
20
n.d.
8.3
<5.0/<5.2
n.d. = not done.
examples 6-substitution (e.g., compound 18) is disfa-
voured, suggesting electronic effects may be of impor-
tance. In contrast, fluorination of the 5- or 7-position
CO Me
2
a
S
S
(
17 and 19) results in examples with high 5-HT2C affin-
ity, though at the expense of slightly reduced selectivity
with respect to the 5-HT2B receptor. In this context it is
noteworthy that introduction of a methyl group in the
dihydropyrrolone moiety (20, racemic) results in signifi-
cantly improved selectivity over both, 5-HT2A and
21
2
2
O
O
O
steps
R
N
N
S
5
-HT2B receptors.
2
2
3: R = H
4: R = Me
Substitution of the exposed indolyl phenyl ring was
expected to have an influence on pharmacokinetic
(
PK) parameters. Indeed, in a standard rat iv/po cross-
1
2
over PK study we found that while compound 5 had
limited oral bioavailability and was rapidly cleared from
the blood stream (Fpo 15%, CLb 53 mL/min/g), the 7-Cl
O
O
CHO
CHO
b
N
O
N
derivative 12 showed improved parameters (F 21%,
po
1
3
CLb 39 mL/min/g). A significant further improvement
was achieved when moving to the 4-methyl piperidine
series where compound 16 was found to have much im-
proved metabolic stability correlating with improved
oral bioavailability (Fpo 44%, CLb 23 mL/min/g).
Importantly, the compound furthermore demonstrated
good brain permeability as expressed by a brain–blood
25
26
Scheme 3. Reagents and conditions: (a) 1—(COCl)
2
3
, AlCl , 1,2-
dichloroethane, 0 °C; then (COCl) , DCM, cat. DMF, 0 °C to rt; then
MeOH, rt to 60 °C (19%); (b) 4, DMF, HCl (0.2 equiv), 120 °C (5%).
2
1
3
plasma ratio of 1.0.
due to its more linear shape, though the effect of dif-
ferent electronic properties may also be of importance.
We became interested in the importance of the indolyl
fragment of the indolyl dihydropyrrolone moiety for
potent and selective affinity to the 5-HT2C receptor.
To investigate this aspect we prepared two additional
scaffolds which maintain the dihydropyrrolone feature
Examples from this series were tested for agonist and
antagonist properties at the human 5-HT2A, 5-HT2B
and 5-HT2C receptors. All compounds lacked agonist
activity at concentrations up to 10 lM but blocked the
1
5
(
Scheme 3). Replacement of the indolyl pyrrole with a
thiophene resulted in fused benzothiophenes 23 and
4. Fused naphthalene 26 was obtained directly from
naphthalene dicarbaldehyde 25 and aniline 4, albeit
effects of 5-HT (5-HT2C affinity estimates: pK 16, 8.6;
b
18, 7.5; 19, 8.6; 24, 8.7).
2
In conclusion we report design, synthesis and properties
of a new series of tricyclic 5-HT2C selective antagonists.
The newly introduced ring fusion, locking coplanarity
between the rings of the 2-phenyl-dihydropyrrolone
moiety, is apparently well tolerated by the 5-HT2C
receptor. The concurrent change of the nature of the
carbon–carbon double bond of the lactam ring (from
vinylic to aromatic) is compatible with a favourable pro-
1
4
in low yield. Both benzothiophene based examples
showed potent and selective binding to the 5-HT2C
receptor (23: pK 5-HT
i
= 8.5/6.4/<6.2; 24: pKi
2C/2A/2B
8
5
-HT2C/2A/2B = 8.8/<6.1/6.7). This suggests that simi-
lar binding modes may be accessible to a wider range
of heterotricycles. Fused naphthalene 26 however was
found to be of reduced affinity and selectivity, possibly

D. Hamprecht et al. / Bioorg. Med. Chem. Lett. 17 (2007) 424–427
427
file at both, receptor and in vivo level. Within this series
good systemic exposure in preclinical species is achiev-
able after oral administration. Further progress in the
development of 5-HT2C antagonists will be published
in a subsequent paper.
Rudolph, C.; Gasteiger, J. Anal. Chim. Acta 1992, 265,
33.
. pK data were determined using [ H]-ketanserin, [ H]-5-
2
3
3
8
i
3
HT and [ H]-mesulergine displacement, respectively, in
HEK293 cells with stable expression of human 5-HT2A, 5-
HT2B and 5-HT2C receptors, in analogy to the procedure
described in Wood, M. D.; Reavill, C.; Trail, B.; Wilson,
A.; Stean, T.; Kennett, G. A.; Lightowler, S.; Blackburn,
T. P.; Thomas, D.; Gager, D. L.; Riley, G.; Holland, V.;
Bromidge, S. M.; Forbes, I. T.; Middlemiss, D. N.
Neuropharmacology 2001, 41, 186.
Acknowledgments
We would like to thank Computational, Analytical and
Structural Sciences and Screening and Compound Pro-
filing colleagues for technical support and providing
in vitro data.
9. Gruda, I. Can. J. Chem. 1972, 50, 18.
1
0. Bunce, R. A.; Randall, M. H.; Applegate, K. G. Org.
Prep. Proced. Int. 2002, 34, 493.
1. While N-alkylation of 4 proceeded without difficulty
1
cyclisation of the resulting intermediate using AlMe
resulted in very low conversion yielding traces only of
product. Also, a two-step protocol via the carboxylic acid
3
References and notes
(removing the tert-butyl group using trifluoroacetic acid)
and subsequent acid activation (oxalyl chloride, then
1
. See for example, (a) Nutt, D. J.; Forshall, S.; Bell, C.;
Rich, A.; Sandford, J.; Nash, J.; Argyropoulos, S. Eur.
Neuropsychopharmacol. 1999, 9(Suppl. 3), S81; (b) Mur-
phy, D. L.; Andrews, A. M.; Wichems, C. H.; Li, Q.;
Tohda, M.; Greenberg, B. J. Clin. Psychiatry 1998,
NEt ; or EDCI) provided only low yields.
3
12. The research complied with national legislation and with
company policy on the Care and Use of Animals and with
related codes of practice.
5
9(Suppl. 15), 4.
13. For iv administration, the compound was dissolved in 25%
(v/v) PEG400 aq containing 5% (v/v) DMSO at a
concentration of 0.5 mg free base/mL and administered
(2 mL/kg) as a bolus to male rats (n = 3) at a nominal dose
level of 1 mg free base/kg. Brain penetration was evaluated
at 1 h. For po administration, the compound was
dissolved in Methocel 0.5% w/v containing 5% DMSO at
a concentration of 0.6 mg free base/mL and administered
(5 mL/kg) by gavage to male rats (n = 3) at a nominal dose
level of 3 mg free base/kg.
2
. Serretti, A.; Artioli, P.; De Ronchi, D. Expert Opin. Ther.
Targets 2004, 8, 15.
. Kahn, R. S.; Wetzler, S. Biol. Psychiatry 1991, 30, 1139.
. (a) Kennett, G. A.; Wood, M. D.; Bright, F.; Cilia, J.;
Piper, D. C.; Gager, T.; Thomas, D.; Baxter, G. S.;
Forbes, I. T.; Ham, P.; Blackburn, T. P. Br. J. Pharmacol.
3
4
1
996, 117, 427; (b) Kennett, G. A.; Wood, M. D.; Bright,
F.; Trail, B.; Riley, G.; Holland, V.; Avenell, K. Y.; Stean,
T.; Upton, N.; Bromidge, S.; Forbes, I. T.; Brown, A. M.;
Middlemiss, D. N.; Blackburn, T. P. Neuropharmacology
14. Thiele, J.; Schneider, J. Justus Liebigs Ann. Chem. 1909,
369, 287.
1
997, 36, 609.
5
. Micheli, F.; Pasquarello, A.; Tedesco, G.; Hamprecht, D.;
Bonanomi, G.; Checchia, A.; Jaxa-Chamiec, A.; Damiani,
F.; Davalli, S.; Donati, D.; Gallotti, C.; Petrone, M.;
Rinaldi, M.; Riley, G.; Terreni, S.; Wood, M. Bioorg.
Med. Chem. Lett. 2006, 16, 3906.
. Schrodinger Inc., Portland. http://www.schrodinger.com.
. (a) Gasteiger, J.; Rudolph, C.; Sadowski, J. Tetrahedron
Computer Methodology 1990, 3, 537; (b) Sadowski, J.;
15. Compounds were tested for agonist and antagonist prop-
erties at the human 5-HT2A, 5-HT2B and 5-HT2C receptors
using mobilisation of intracellular calcium as described in,
and pKb values determined according to, Jerman, J. C.;
Brough, S. J.; Gager, T.; Wood, M.; Coldwell, M. C.;
Smart, D.; Middlemiss, D. N. Eur. J. Pharmacol. 2001,
414, 23, with the exception that the receptors were
expressed in HEK293 cells as described in Ref. 8.
6
7