Tricyclic pyridones as functionally selective human GABA Aα2/3 receptor-ion channel ligands

Article abstract of DOI:10.1016/j.bmcl.2004.01.057

A series of tricyclic pyridones has been evaluated as benzodiazepine site ligands with functional selectivity for the α₃ over the α₁ containing subtype of the human GABA_A receptor ion channel. This investigation led to the

Full text of DOI:10.1016/j.bmcl.2004.01.057

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1679–1682
Tricyclic pyridones as functionally selective human GABAAꢀ2/3
receptor-ion channel ligands
James Crawforth,a,* John R. Atack,b Susan M. Cook,b Karl R. Gibson,a Alan Nadin,a
Andrew P. Owens,a Andrew Pike,a Michael Rowley,a Alison J. Smith,b Bindi Sohal,b
Francine Sternfeld,a Keith Waffordb and Leslie J. Streeta
aDepartment of Medicinal Chemistry, Merck Sharp & Dohme Research Laboratories, The Neuroscience Research Centre,
Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
bDepartment of Biochemistry, Merck Sharp & Dohme Research Laboratories, The Neuroscience Research Centre, Terlings Park,
Eastwick Road, Harlow, Essex CM20 2QR, UK
Received 30 June 2003; revised 16 January 2004; accepted 21 January 2004
Abstract—A series of tricyclic pyridones has been evaluated as benzodiazepine site ligands with functional selectivity for the a3 over
the a1 containing subtype of the human GABAA receptor ion channel. This investigation led to the identification of a high affinity,
functionally selective, orally bioavailable benzodiazepine site ligand that demonstrated activity in rodent anxiolysis models and
reduced sedation relative to diazepam.
# 2004 Elsevier Ltd. All rights reserved.
Neurotransmission by g-aminobutyric acid (GABA) is
the major inhibitory mechanism in the central nervous
system (CNS). GABA exerts its effects via the ligand
gated GABAA and GABAC receptor ion channels and
the metabotropic GABAB receptors. However, the
GABAA receptor has received most attention since
modulation of its function by a number of classes of
compounds such as barbituates, neurosteroids, ethanol
and most notably, the benzodiazepines (BZs), results in
a diverse range of responses, including changes in motor
activity, seizures, sedation, modulation of cognition and
mood, particularly anxiety states. GABAA receptors are
agonists are anxiolytic and inverse agonists are anxio-
genic, whereas antagonists have no effect. Classical BZ’s
such as Diazepam 1, have equivalent affinity and full
agonist efficacy at a1, a2, a3, and a5 containing GABAA
receptors. Although Diazepam 1 is efficacious in the
treatment of anxiety, it causes side effects2 such as
sedation, ataxia, potentiation with alcohol and risk of
tolerance and dependance with chronic use.
oligomeric assemblies of a large range of subunits (a1À6
,
b
1À3, g1À3, d, e, p, and y).1 Most GABAA receptors in
the brain contain a, b and g subunits, and most specifi-
cally those with an a1, a2, a3, or a5, in conjunction with
b and g subunits, possess a specific binding site for
benzodiazepines. Ligands at the BZ site are categorised
on the basis of their ability to modulate GABA-acti-
vated currents. BZ agonists, inverse agonists or antago-
nists increase, decrease or have no effect, respectively on
GABA induced currents. These different efficacies are
mirrored in opposing behavioural effects. Generally
The need for improved treatments which retain the
anxiolytic efficacy but have reduced side effect liability is
clear. It has been shown3,4 that the a1 subtype of the
GABAA receptor is responsible for the sedative/motor
effects of diazepam while the a2/a3 subtypes are pri-
marily responsible for the anxiolytic activity. In the light
of this, a subtype selective ligand (i.e., a1 antagonist and
a2/a3 agonist) should discriminate between sedation
Keywords: Tricyclic pyridones; GABAA.
* Corresponding author. Tel.: +44-1279-440436; fax: +44-1279-
440187; e-mail: james_crawforth@merck.com
0960-894X/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.01.057

1680
J. Crawforth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1679–1682
and the anxiolytic properties, providing an improved
anxiolytic therapy. Continuing from our work on 3-
heteroaryl-2-pyridones5 in this publication we describe
work on a series of tricyclic pyridone GABAA BZ site
ligands with functional selectivity for the a2 and a3
subtypes over the a1 containing subtype. The com-
pounds showed no selectivity between the a2 and a3
subtypes, binding affinities and efficacy being compar-
able. Our studies indicated that it was possible to obtain
compounds that display good affinity at the a1 and a3
GABAA subtypes as exemplified by the 5-phenyl deri-
vative 2a6 (Table 1). In general compounds in this series
displayed no binding selectivity between a1, a2 and a3
subtypes (a1 0.7 nM; a2 0.3 nM; a3 0.4 nM). However,
compound 2a, demonstrated functional selectivity for a3
over a1 subtypes (efficacy a1 0.10; a3 0.32). 2a suffered
from a poor pharmacokinetic profile with high turnover
in dog liver microsomes and low oral bioavailability in
dogs. As a continuation of our studies we sought to
investigate the 5-position on the pyridone core in order
to improve the pharmacokinetic profile, whilst main-
taining the functional selectivity. The derivatives 2b–t
were prepared as outlined in Schemes 1 and 2.7,8
Scheme 2. Reagents: (i) LDA, ethylene oxide, À20 ꢀC, 78%; (ii)
TBDMSCl, imidazole, DCM, rt, 98%; (iii) Pd(OAc)2, Ph2P(CH2)3PPh2,
EtNiPr2, CO, MeOH, DMF, 95 ꢀC, 88%; (iv) KOTMS, Et2O, rt, 89%;
(v) CDI, DMF, 50 ꢀC, then ArCH2CO2Me, NaH, 0 ꢀC–rt, 24–70%;
(vi) NaCl, H2O, DMSO, 150 ꢀC, 50–76%; (vii) DMF.DMA, rt, 60–
75%; (viii) 4, NaH, DMF, 50 ꢀC, 1N HCl, 45–75%; (ix) DEAD, PPh3,
THF, rt, 23–70%.
Base-catalysed cyclisation of 37 and the thiazole acet-
amide 4 gave the 2-pyridone 5. N-alkylation of 5 with
3-bromo-4-(3-bromo-propyl)-pyridine using Curran’s
LiBr-mediated procedure7 followed by radical cyclisa-
tion6 gave the tricycle 6. Subsequent deprotection of the
alcohol with boron tribromide, formation of the triflate,
and Pd(0)-catalysed arylation via either Suzuki coupling
of arylboronic acids or Stille coupling of the corres-
ponding arylstannane gave 2b–g, i–k, n, s. An alternative
synthetic sequence by Gibson et al.9 could be employed
as outlined in Scheme 2. Commercially available 3-bromo-
4-methylpyridine was lithiated and alkylated with eth-
ylene oxide.10 Protection and carbonylation of the 3-
bromo group gave the methyl ester 7. The latter was
hydrolysed to the nicotinic acid, activated as the imida-
zolide and condensed with the appropriate methyl
arylacetate, and decarboxylated under Krapcho condi-
tions.11 Reaction of the ketone with dimethylforma-
mide-dimethylacetal yielded the dimethylaminopropen-
2-ones 8. Base-catalysed condensation of the thiazole
acetamide 4 and 8, deprotection of the alcohol, and
closure of the seven membered ring under Mitsunobu
conditions gave the tricyclic 2-pyridones 2h, l, m, p–r, t.
Table 1 summarises the affinities and efficacies at both
a1 and a3 containing GABAA receptors for substituted
phenyl groups at the 5-position of the pyridone core.
Generally, a wide variety of small substituents are well
tolerated in terms of binding affinity. There is, however,
a size limitation in this part of the molecule as illu-
strated by the 4-tert-butylphenyl 2f and 2-napthyl 2g
derivatives, which both lose affinity relative to 2a. Only
the 4-chlorophenyl 2b retains the binding affinity of the
parent 2a, but exhibits a poor functional selectivity
profile. It is also clear from Table 1, that while sub-
stitution is broadly tolerated for binding affinities, a
wide spread of functional efficacy is observed at both
the a1 and a3 subtypes, indicating that this could be an
effective way of moderating the efficacy to give the
desired subtype selective profile.
If we only consider the compounds that display close to
the desired subtype selectivity, (i.e., an antagonist at a1
and agonist at a3), then from Table 1 only the 2-
methylphenyl 2c and 4-cyanophenyl 2e derivatives
approach this profile. Unfortunately the 2-methylphenyl
analogue 2c, despite good binding affinity and func-
tional selectivity, showed high turnover when exposed
to dog liver microsomes (all test compound incubated at
1 mM at a protein concentration of 0.4 mg/mL for 15
min at 37 ꢀC). However, for the 4-cyanophenyl 2e the
turnover in dog liver microsomes was only 28% as
compared to 67% for 2a, suggesting that plasma clear-
ance might also be reduced by modifications in this
Scheme 1. Reagents: (i) NaH, MeOH, DMF, 70 ꢀC, 73%; (ii) NaH,
LiBr, 3-bromo-4-(3-bromopropyl)-pyridine, DME, DMF, 75 ꢀC, 82%;
(iii) AIBN, Bu3SnH, benzene, 80 ꢀC, 50%; (iv) BBr3, DCM, 25 ꢀC,
89%; (v) Tf2O, pyr., DCM, À78–0 ꢀC, 73%; (vi) Pd(PPh3)4,
ArB(OH)2, CsCO3, or ArSnBu3, LiCl, CuI, 24–76%.

J. Crawforth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1679–1682
1681
Table 1. Substituted phenyl derivatives
Table 2. 5-Membered heterocycles
No.
2a
Ar
Ki (nM)a
Efficacyb
No.
2a
2h
2i
Ar
Ki (nM)a
Efficacyb
a1
a3
a1
a3
a1
a3
a1
a3
0.7
(Æ0.2)
0.4
(Æ0.1)
0.10
(Æ0.05)
0.32
(Æ0.03)
0.7
(Æ0.2)
0.4
(Æ0.1)
0.10
(Æ0.05)
0.32
(Æ0.03)
2b
0.2
(Æ0.0)
0.4
(Æ0.1)
0.17
(Æ0.2)
0.26
(Æ0.01)
14.3
(Æ4.3)
16.8
(Æ8.6)
À0.33
À0.01
(Æ0.05)
(Æ0.06)
3.9
(Æ1.0)
4.4
(Æ0.15)
0.04
(Æ0.05)
0.05
(Æ0.07)
2c
1.5
(Æ0.1)
1.2
(Æ0.1)
0.08
(Æ0.0)
0.36
(Æ0.1)
2j
4.6
(Æ0.3)
3.1
(Æ0.6)
À0.11
0.41
(Æ0.04)
2d
2e
2f
5.0
(Æ0.2)
3.0
(Æ0.6)
0.56
(Æ0.07)
0.48
(Æ0.1)
(Æ0.07)
2k
2l
7.8
(Æ0.1)
4.2
(Æ0.5)
À0.21
0.41
(Æ0.03)
(Æ0.06)
5.4
(Æ0.8)
3.1
(Æ0.3)
À0.04
0.39
(Æ0.07)
(Æ0.03)
4.0
(Æ0.1)
3.0
(Æ0.0)
0.00
(Æ0.02)
0.46
(Æ0.03)
>33
>33
—
—
—
—
2m
10.0
(Æ2.9)
20.0
(Æ3.1)
0.29
(Æ0.04)
0.47
(Æ0.04)
2g
>33
>33
a Affinity was determined by measuring the displacement of [3H]Ro15-
1788 from human recombinant GABAA receptors containing b3g2
plus either a1 or a3 stably expressed in L(tkÀ) cells. Values are the
(meanÆSD) of 2–10 separate determinations.
a Affinity was determined by measuring the displacement of [3H]Ro15-
1788 from human recombinant GABAA receptors containing b3g2 plus
either a1 or a3 stably expressed in L(tkÀ) cells. Values are the
(meanÆSD) of 2–10 separate determinations.
b Modulation of chloride ion flux in cells expressing b3g2 plus either a1
or a3 produced by an EC20 equivalent concentration of GABA in the
presence of an approximate 1000ÂKi concentration of test com-
pound. Efficacy is expressed relative to the full agonist chlorodiazep-
oxide (relative efficacy=1.0), from at least seven independent
experiments.12
b Modulation of chloride ion flux in cells expressing b3g2 plus either a1
or a3 produced by an EC20 equivalent concentration of GABA in the
presence of an approximate 1000ÂKi concentration of test compound.
Efficacy is expressed relative to the full agonist chlorodiazepoxide
(relative efficacy=1.0), from at least seven independent experiments.12
area. Table 2 summarises the effects of five membered
heterocycles at the 5-position of the pyridone core. As
with the substituted phenyl groups, this position proved
tolerant in terms of binding affinity, of a wide variety of
functional groups.
although lower than the lead compound 2a, was still
considered too high to be taken further.
Table 3 summarises data for the 6-ring heterocycles.
The 4-pyrimidyl analogue 2p displays good functional
selectivity but has poor binding affinity. Pleasingly, the
4-pyridyl compound 2t14 retained the binding affinity at
the a3 subtype as compared to 2a, and also exhibits
excellent functional selectivity, with the desired antag-
onism at a1, and an agonist profile at a3. Binding selec-
tivity of 2t for a3 over a4 (27.9 nM), a5 (>33.3 nM), and
a6 (65.9 nM), containing subtypes was achieved, but no
binding or functional selectivity over a2 (0.65 nM, effi-
cacy 0.37) was observed. Compound 2t also showed an
improvement in turnover in dog liver microsomes
(14%). This reduced turnover translated into improved
oral bioavailability in dog of 20% as compared to 5%
for the starting lead 2a. Compound 2t was shown to
have 28% oral bioavailability in rat. In line with the
programme hypothesis, compound 2t was anxiolytic in
Although none of the five membered heterocycles
retained the subnanomolar binding affinity of 2a, a wide
range of functional selectivities was observed, again
indicating this strategy as a good way of modifying the
functional efficacy of this class of compounds. The most
active compounds were the isoxazole 2j and 1-methyl-
pyrrole 2l which both maintained good affinity. Isox-
azole 2j displayed a good overall selective efficacy for a3
over a1 GABAA receptors. The 1-methylpyrrole 2l also
showed good binding to both a1 and a3 subtypes and
encouragingly also displayed the desired functional
selectivity profile (Table 2). Compound 2l was again
screened in dog liver microsomes as a quick method for
predicting in vivo clearance. The turnover (53%),

1682
J. Crawforth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1679–1682
Table 3. 6-Membered heterocycles
Acknowledgements
We would like to thank Dr. Jason Elliott and Dr. Jose
L. Castro for their help during the preparation of this
manuscript.
References and notes
No.
2a
2n
2p
2q
2r
Ar
Ki (nM)a
Efficacyb
a1
a3
a1
a3
1. Sieghart, W.; Fuchs, K.; Tretter, V.; Ebert, V.; Jechlinger,
M.; Hoger, H.; Adamiker, D. Neurochem. Int. 1999, 34,
379.
2. Korpi, E. R.; Mattila, M. J.; Wisden, W.; Luddens, H.
Ann. Med. 1997, 29, 275.
0.7
(Æ0.2)
0.4
(Æ0.1)
0.10
(Æ0.05)
0.32
(Æ0.03)
21.1
(Æ4.9)
9.0
(Æ3.0)
À0.03
0.16
(Æ0.02)
3. 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. Nature Neuroscience 2000,
3, 587.
4. Low, K.; Crestani, F.; Keist, R.; Benke, D.; Brunig, I.;
Benson, J. A.; Fritschy, J.-M.; Rulicke, T.; Bluethmann,
H.; Mohler, H.; Rudolph, U. Science 2000, 290, 131.
5. Collins, I.; Moyes, C. R.; Davey, W. B.; Rowley, M.;
Bromidge, F. A.; Quirk, K.; Atack, J. R.; McKernan,
R. M.; Thomson, S.; Wafford, K.; Dawson, G. R.; Pike,
A.; Sohal, B.; Tsou, N. N.; Ball, R. G.; Castro, J. L. J.
Med. Chem. 2002, 45, 1887.
(Æ0.04)
12.2
(Æ1.6)
8.5
(Æ1.0)
0.03
(Æ0.08)
0.32
(Æ0.08)
8.3
(Æ0.7)
5.0
(Æ2.1)
À0.41
0.20
(Æ0.03)
(Æ0.05)
26.1
(Æ2.1)
17.3
(Æ2.8)
À0.33
À0.44
(Æ0.07)
(Æ0.1)
2s
5.3
(Æ1.1)
2.4
(Æ0.5)
À0.16
À0.02
(Æ0.07)
(Æ0.08)
6. Harrison, T.; Moyes, C. R.; Nadin, A.; Owens, A. P.;
Lewis, R. T. PCT Int. Appl. (1998) WO9850384.
7. Nadin, A.; Harrison, T. Tetrahedron Lett. 1999, 40, 4073.
8. Liu, H.; Ko, S. B.; Josien, H.; Curran, D. P. Tetrahedron
Lett. 1995, 36, 8917.
9. Gibson, K. R.; Hitzel, L.; Mortishire-Smith, R. J.; Ger-
hard, U.; Jelley, R. A.; Reeve, A. J.; Rowley, M.; Nadin,
A.; Owens, A. P. J. Org. Chem. 2002, 67, 9354.
10. Mink, K.; Bracher, F. Liebigs Ann. 1995, 645.
11. Krapcho, P. A.; Lovey, A. J. Tetrahedron Lett. 1973, 14,
957.
2t
1.3
(Æ0.2)
0.5
(Æ0.1)
0.00#
(Æ0.02)
0.45#
(Æ0.1)
a As Tables 1 and 2. For compounds marked # efficacy measurement
of effect of test compound on a current at GABA EC20 using whole-
cell patch-clamp electrophysiological recording.13
the elevated plus maze3,15 at a dose of 1 mg/kg po. In
addition, 2t demonstrated no sedative effects up to 30
mg/kg po and no potentiation of ethanol interaction up
to 10 mg/kg as shown by the mouse rotarod perfor-
mance test.3
12. Smith, A. J.; Alder, L.; Silk, J.; Adkins, C.; Fletcher, A. E.;
Scales, T.; Kerby, J.; Marshall, G.; Wafford, K. A.;
McKernan, R. M.; Atack, J. R. Mol. Pharmacol. 2001, 59,
1108.
13. Whiting, P. J.; McAllister, G.; Vassilatis, D.; Bonnet,
T. P.; Heavens, R. P.; Smith, D. W.; Hewson, L.;
O’Donnell, R.; Rigby, M. R.; Sirinathsinghji, D. J. S.;
Marshall, G.; Thompson, S. A.; Wafford, K. A. J. Neu-
roscience 1997, 17, 5027.
14. Crawforth, J. M.; Gibson, K. R.; Rowley, M. U.S. Pat.
Appl. Publ. (2001) U.S 20010053776.
15. Collinson, N.; Dawson, G. R. Psycopharmacology 1997,
132, 751.
In summary, replacement of the phenyl moiety in 2a
with a 4-pyridyl group gave 2t; a high affinity benzo-
diazepine site ligand with a3-subtype functional selec-
tivity for the human GABAA receptor-ion channel. The
compound was efficacious in animal models of anxiety,
and showed no sedation or potentiation of ethanol
effects.