Exploring subtype selectivity and metabolic stability of a novel series of ligands for the benzodiazepine binding site of the GABAA receptor

Article abstract of DOI:10.1016/j.bmcl.2010.12.107

A novel series of agonists at the benzodiazepine binding site of the GABA_A receptor was prepared by functionalizing a known template. Adding substituents to the pyrazolone-oxygen of CGS-9896 led to a number of compounds with selectivities for e

Full text of DOI:10.1016/j.bmcl.2010.12.107

Bioorganic & Medicinal Chemistry Letters 21 (2011) 1523–1526
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Exploring subtype selectivity and metabolic stability of a novel series
of ligands for the benzodiazepine binding site of the GABAA receptor
Samuel Hintermann a, , Konstanze Hurth a, Joachim Nozulak a, Marina Tintelnot-Blomley a,
⇑
Reiner Aichholz b, Joachim Blanz b, Klemens Kaupmann a, Johannes Mosbacher a,
a Novartis Institutes for BioMedical Research, Neuroscience Research, CH-4002 Basel, Switzerland
b Novartis Institutes for BioMedical Research, Metabolism and Pharmacokinetics, CH-4002 Basel, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of agonists at the benzodiazepine binding site of the GABAA receptor was prepared by func-
tionalizing a known template. Adding substituents to the pyrazolone-oxygen of CGS-9896 led to a num-
ber of compounds with selectivities for either a2- or a1-containing GABAA receptor subtypes offering an
entry into indications such as anxiety and insomnia. In this communication, structure–activity relation-
ship and efforts to increase in vitro stabilities are discussed.
Received 23 November 2010
Revised 21 December 2010
Accepted 22 December 2010
Available online 28 December 2010
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
GABAA ligands
Subtype selectivity
Metabolic stability
Benzodiazepines (BZs) as positive modulators of
c
-amino buty-
This evidence encouraged us to identify novel compounds with im-
proved selectivity profiles for the treatment of anxiety
disorders.
We decided to use CGS-9896 (1) as starting point for derivatiza-
tion efforts to identify subtype selective compounds. The oxygen-
atom of the dihydro-pyrazolone moiety was used as attachment
point for further substituents.
ric acid type A (GABAA) receptors have been for years the most
widely used drugs for disorders of the nervous system such as anx-
iety and insomnia. Despite their fast onset of action and relatively
high efficacy, they are no longer first-line treatment choice for
most types of anxiety spectrum disorders because of their adverse
side effect profile.1
a2
Benzodiazepines bind to GABAA receptors at the interface of
and subunits that form together with b subunits most of the cen-
tral pentameric receptor channels. Out of the six known subunits,
1, 2, 3 and 5 are part of receptors carrying the benzodiazepine
(BZ) binding site.
a
c
Cl
Cl
OMe
a
a
a
a
a
N
N
N
N
N
N
O
O
O
Recent molecular genetic studies have led to the hypothesis
that subtype-specific GABAA receptor modulators can alleviate
anxiety without unwanted side-effects such as sedation, memory
N
N
N
H
H
H
impairment or abuse liability.2–4 Mice with mutations in the
a1
CGS-9896 (1)
CGS-17867A
CGS-20625
subunit that render those receptors insensitive to BZs showed no
sedation upon treatment with diazepam, whereas the drug was
no longer anxiolytic in mice carrying a benzodiazepine-insensitive
Compounds were routinely screened for their in vitro efficacy at
1- and 2-containing GABAA receptor subtypes. The assay was
a2 subunit. CGS-17876A, a BZ-site ligand with anxiolytic but much
a
a
less sedative activity in rodents was shown a posteriori to have low
performed at recombinant GABAA receptors expressed in CHO cells
using a membrane potential sensitive fluorometric read-out.8 The
compound-induced potentiation of the response to a sub-maximal
GABA concentration was measured and data are reported for a sin-
gle defined concentration of test compounds. The intended target
efficacy at
a
1-carrying GABAA receptors.5 In fact, the structurally
related compound CGS-20625 was evaluated in clinical trials6
based on its interesting anxio-selective profile in pre-clinical tests.7
profile was partial to full agonism at
a
2 and/or
a
3 and antagonism
⇑
Corresponding author. Tel.: +41 616968456; fax: +41 616962455.
at
a
1. For selected compounds, affinity for
a1- and a
2-containing
E-mail address: samuel.hintermann@novartis.com (S. Hintermann).
Present address: Actelion Pharmaceuticals Ltd, Gewerbestrasse 16, CH-4123
Allschwil, Switzerland.
recombinant GABAA receptors was determined using [3H]-flu-
mazenil as radioligand9 to confirm binding to the BZ binding site.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.12.107

1524
S. Hintermann et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1523–1526
Initially, we pursued a medicinal chemistry program around
further investigations with the main goal to improve potency and
selectivity.
benzyl-substituted analogs. They were prepared by alkylation of
CGS-989610 under basic conditions in
(Scheme 1).
a
microwave reactor
Then, the influence of substituents on the benzyl group with re-
spect to potency and subtype selectivity was investigated (Table 2).
Placing the chloro-substituent on different positions was detri-
mental for selectivity. But replacing the 2-chloro with the 2-fluoro
substituent gave a compound (11) with a very promising func-
This straightforward route was used for the initial SAR work
even though the isolated yield was usually rather low. After chro-
matographic removal of N-alkylated side products, the pure O-
alkylated isomers were typically isolated in yields between 10%
and 30%.
Table 1 summarizes results of a set of representative com-
pounds. The 2-chloro-benzyl substituent was initially kept con-
stant in order to explore the influence of core modifications10,11
in this novel series of modulators. Variations of substituents on
the phenyl pyrazolone moiety (cf. 2–4) showed mixed results.
tional selectivity for the
a2 subtype. Also the trifluoro-methyl
derivative 12 was very potent and reached a similar maximal effi-
cacy as midazolam, the standard benzodiazepine used in the assay.
Unfortunately, this potent compound was completely unselective.
All these modifications usually had limited impact on the affin-
ity for the BZ binding site. Only the 4-fluoro analog 13 showed an
increased affinity for the
the functional assay. Replacing the phenyl by a cyclopropyl substi-
tuent (14) was tolerated only on 1.
a2 subtype but without any selectivity in
The corresponding compounds were very weak agonists at
and some behaved as antagonists at 1. The 2-fluoro analog 3
had highest affinity but only low efficacy.
a2
a
a
In parallel to the ongoing derivatization program, compound 5
Next, the influence of changes in the quinoline core was inves-
tigated by looking at modifications which had previously been re-
ported for members of the CGS-9896 family.5 Partial saturation of 2
was well tolerated and led to the equipotent compound 5 with
was chosen for further profiling. It was anticipated to test whether
the weak a2 selectivity in vitro would translate into a measurable
effect in vivo. Unfortunately, the properties of compound 5 pre-
cluded it from testing in anxiety models. Especially, the clearance
functional preference for
full agonist at both subtypes whereas incorporation of polar oxy-
gen (7) yielded a weak but 2-selective agonist. Finally, incorpora-
tion of a gem-dimethyl substituents in the 8-positon of the core (8)
yielded compound with preference for 1.
Based on these results, compound 5 combining both functional
a
2. Expanding the ring-size (6) led to a
was very high (in vitro: CLint rat: 128.3
l
l/minÃmg). Therefore, a
detailed in vitro metabolism study with rat liver microsomes12 to
identify major pathways and block weak spots in next generation
compounds was performed.
a
a
This study identified two main degradation pathways for
compound 5 (Fig. 1). Extensive hydroxylation of the aliphatic part
of the core was observed (pathway A) but the exact positions of
preferences for
a2 and adequate potency was chosen as a basis for
Cl
Table 2
Cl
Affinity and efficacy at a1 and a2 GABAA receptor subtypes for differently substituted
analogs of 5
i
N
N
N
N
Cl
O
O
Cl
Cl
N
H
N
N
N
N
N
1
2
O
O
R
Scheme 1. Reagents and conditions: (i) K2CO3, acetone, 1-bromomethyl-2-chloro-
benzene, microwave, 80 °C, 30 min, 20% yield.
N
N
5, 9-13
14
Entry
R
% Efficacya
Affinityb (nM)
Table 1
Efficacy and affinity for selected CGS-9896 analogs at
receptor subtypes
a
1- and a2-containing GABAA
a
1
a
2
IC50
a
1
IC50 a2
5
9
2-Cl
3-Cl
4-Cl
2-F
2-CF3
4-F
25
38
20
31
84
24
37
50
50
42
83
94
41
36
49
271
n.d.
177
154
n.d.
57
n.d.
157
64
n.d.
50
10
11
12
13
14
Cl
R
N
N
N
N
Cl
Cl
—
74
472
O
O
X
a
Determined at 3 lM compound concentration (n = 2) and normalized to the
maximal effect of midazolam (300 nM).
N
N
2 - 4
5 - 8
b
n.d. = not determined.
Compds
R
X
% Efficacya
Affinityb (nM)
a
1
a
2
IC50
a
1
IC50 a2
Cl
1
2
3
4
5
6
7
8
38
13
47
22
17
19
50
96
32
23
1
64
1
50
49
242
n.d.
67
4
265
2
155
271
605
461
213
PathwayA: CYP450
mediated aliphatic
C-hydroxylation
Pathway B: Back conversion
to pyrazolone (enzymatic or
non-enzymatic degradation)
4-Cl
2-F
3,4-(OCH2O)
À8
À15
25
114
4
N
N
O
CH2
(CH2)2
O
N
Cl
CMe2
48
5
a
Determined for single compound concentration (3 lM [except 1: 1 lM]; n = 2)
and normalized to the maximal effect of midazolam (300 nM).
Figure 1. The main pathways A and B which led to high in vitro clearance of
compound 5.
b
n.d. = not determined.

S. Hintermann et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1523–1526
1525
Next, we focused our efforts on the stabilization of the benzylic
side chain. Replacement of benzyl by phenyl (18) led to a largely
improved stability but at the expense of on-target activity. Here,
the cyclopropyl-methyl group was an almost equipotent replace-
ment for the 2-chloro-benzyl group with a nice functional selectiv-
Cl
Cl
i, ii
N
N
N
N
O
O
ity for
a2 (17). In addition, it showed significantly increased
N
N
H
in vitro metabolic stability but core metabolism was still consid-
ered too high for in vivo testing.
14
1
Finally, the cyclopropyl replacement was applied to the quin-
oline core to combine both the optimal metabolic properties of
both the core and the side chain. The imidazole substituted core
was chosen based on a previous observation that it could influ-
ence the core stability beneficially (data not shown). And, indeed,
cyclopropyl ether 19 showed an excellent in vitro metabolic sta-
bility with almost complete parent retention under assay condi-
Scheme 2. Reagents and conditions: (i) POCl3, sealed tube, 135 °C, 30 min (97%);
(ii) NaH 60% in mineral oil, cyclopropylmethanol, THF, reflux, 2.5 h (66%).
core-hydroxylation could not be identified. In addition, significant
amounts of the parent pyrazolone were observed (pathway B). This
back-conversion could be due to oxidative cleavage or simple
chemical instability of the benzylic position of the side-
chain under test conditions. Based on these results, we designed
the next series of compounds with the aim to lower the in vitro
clearance.
tions. Surprisingly, this stable analog showed
functional assay and from the binding assay
also evident. In the binding assay 19 showed unusual displace-
ment characteristics at both 1 and 2 with a Hill coefficient
a1 selectivity in the
a
1 selectivity was
a
a
of only 0.5–0.6 which could be a hint for binding of multiple
A novel synthesis was developed to gain easy access to a diverse
set of side-chain analogs. The optimized route is exemplified for
compound 14 in Scheme 2. Pyrazolone 1 was chlorinated with
phosphoryl chloride followed by nucleophilic displacement with
the sodium salt of the corresponding alcohol. This led to the de-
sired ethers in good overall yields and proved to be general for a
wide variety of alcohols.
A first series of compounds with modifications in the tetrahydro-
quinoline was prepared to address the metabolic stability of the
core. Indeed, incorporation of an oxygen (15) or methyl-substituents
(16) in the 8-position led to decreased core-hydroxylation. This
effect was however partially offset by increased degradation
through pathway B and efficacy at the target was considerably lower
(Table 3).
ligands.
The high functional efficacy and binding selectivity of 19 for a1
offers an interesting starting point for further optimization of the
affinity of this chemotype to identify candidates for the treatment
of insomnia.
In conclusion, a novel series of agonists at the BZ binding site
of the GABAA receptor based on modifications of a known tem-
plate was described. The optimization work surprisingly yielded
both
a2 and a1 preferring agonists depending on the substitu-
tion patterns. Low metabolic stability of initial derivatives was
improved by both core and side-chain modifications. Currently,
further work is ongoing to identify compounds with low-
nanomolar affinity, selectivity and stability for the treatment of
either anxiety or insomnia.
Table 3
Affinity and efficacy at a1 and a2 GABAA BZ receptor subtypes and metabolic data for core and side-chain modified analogs
Cl
Cl
N
N
N
N
R
X
R
X
N
N
5, 15-18
19
Entry
X
R
Parent [%]+
Pathway Aa (%)
P
Pathway B (%)
Efficacyb
1 (%)
Affinityc (nM)
IC50 a2
of metabolites
a
a2 (%)
IC50
46
a1
5
CH2
O
4
82
11
16
25
4
50
32
271
15
41
41
n.d.
461
16
17
18
CMe2
CH2
2
35
45
66
55
49
26
10
0
4
23
67
94
213
185
18
64
CH2
n.m.
n.m.
>3000
134
>3000
2114
19
96
1
1
80
35
a
b
c
Relative peak areas assuming identical response factors in MS detection; further minor metabolites not reported (n.m. = not meaningful).
Determined at 3 M compound concentration (n = 2) and normalized to the maximal effect of midazolam (300 nM).
n.d. = not determined.
l

1526
S. Hintermann et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1523–1526
7. Williams, M.; Bennett, D. A.; Loo, P. S.; Braunwalder, A. F.; Amrick, C. L.; Wilson,
Acknowledgments
D. E.; Thompson, T. N.; Schmutz, M.; Yokoyoma, N.; Wasley, J. W. F. J.
Pharmacol. Exp. Ther. 1989, 248, 89.
The authors thank Richard Felber, Sandrine Liverneaux, Kerstin
Barker, Marcel Reck, Christine Blumer, Christina Wittmann and
Jakob Heid for excellent technical assistance. Kevin McAllister,
Graeme Bilbe, Jürgen Wagner and Yves Auberson are acknowl-
edged for critical input and stimulating discussions.
8. Joesch, C.; Guevarra, E.; Parel, S. P.; Bergner, A.; Zbinden, P.; Konrad, D.;
Albrecht, H. J. Biomol. Screen 2008, 13, 218.
9. The flumazenil radioligand binding assay was done using membranes from
CHO cells stably overexpressing GABAA
a
1 or
a
2 (and b2
l
c2) subunits. Assay
mixtures (200 l; 50 mM Tris–Cl pH 7.4, 40–60
l
g membranes, 2 nM 3H-Ro-
15-1788; NEN PerkinElmer) were incubated for 45 min at RT in 96 wells and
subsequently filtered using 96-Unifilter GF/C (Packard) or MAHFC1H60
(Millipore) plates. Non-specific binding was determined in the presence of
References and notes
1 lM flumazenil. Data were analyzed by non-linear curve fitting using
GraphPad Prism 4 software (triplicate determination SEM).
10. Yokoyama, N. Patent application EP 22078.
1. Nemeroff, C. B. J. Clin. Psychiatry 2003, 64, 3.
2. Rudolph, U.; Crestani, F.; Benke, D.; Brunig, I.; Benson, J. A.; Fritschy, J.-M.;
Martin, J. R.; Bluethmann, H.; Mohler, H. Nature 1999, 401, 796.
3. Low, K.; Crestani, F.; Keist, R.; Benke, D.; Beunig, I.; Benson, J. A.; Fritschy, J.-M.;
Ruicke, T.; Bluethmann, H.; Mohler, H.; Rudolph, U. Science 2000, 290, 131.
4. McKernan, R. M.; Rosahl, T. W.; Reynolds, D. S.; Sur, C.; Wafford, K. A.; Atack, J.
R.; Farrar, S.; Myers, J.; Cook, G.; Ferris, P.; Garrett, L.; Bristow, L.; Marshall, G.;
Macaulay, A.; Brown, N.; Howell, O.; Moore, K. W.; Carling, R. W.; Street, L. J.;
Castro, J. L.; Ragan, C. I.; Dawson, G. R.; Whiting, P. J. Nat. Neurosci. 2000, 3, 587.
5. Mitchinson, A.; Atack, J. R.; Blurton, P.; Carling, R. W.; Castro, J. L.; Curley, K. S.;
Russell, M. G. N.; Marshall, G.; McKernan, R. M.; Moore, K. W.; Narquizian, R.;
Smith, A.; Street, L. J.; Thompson, S.-A.; Wafford, K. Bioorg. Med. Chem. Lett.
2004, 14, 3441.
11. Yokoyama, N. Patent application US4647566.
12. The in vitro metabolism was determined in rat liver microsomes (Gentest
452501, mixed gender). Incubations were conducted at 37 °C for 60 min using
0.3 mg/mL microsomal protein, NADPH regenerating system (isocitrate-
dehydrogenase (1 U/mL), NADP (1 mmol/L), isocitrate (5 mmol/L), UDP-
glucuronic acid (2.4 mM) and 5 lM of the respective parent compound. For
detection of potential reactive intermediates 1.5 mM of glutathione was
supplemented in a separate experiment. Metabolites were characterized by
capillary HPLC/high resolution MSn analysis. The data obtained are in
accordance with elemental composition, MS/MS fragmentation and number
of exchangeable hydrogen atoms (replacement of H2O by D2O in the mobile
phase).
6. Hirschberg, Y.; Oberle, R. L.; Ortiz, M.; Lau, H.; Markowska, M. J. Pharmacokinet.
Biopharm. 1995, 23, 11.