N-(indol-3-ylglyoxylyl)piperidines: High affinity agonists of human GABA-A receptors containing the α1 subunit

Article abstract of DOI:10.1016/S0960-894X(00)00245-6

A new class of N-(indol-3-ylglyoxylyl)piperidines are high affinity agonists at the benzodiazepine binding site of human GABA-A receptor ion-channels, with modest selectivity for receptors containing the α₁ subunit over α₂ and α₃. All three receptor subtypes discriminate substantially between the two enantiomers of the chiral ligand 10. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00245-6

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1381±1384
N-(Indol-3-ylglyoxylyl)piperidines: High Anity Agonists of
Human GABA-A Receptors Containing the ꢀ1 Subunit
Ian Collins,a,* William B. Davey,a Michael Rowley,a Kathleen Quirk,b
Frances A. Bromidge,b Ruth M. McKernan,b Sally-Anne Thompsonc
and Keith A. Wa€ordc
aDepartment of Medicinal Chemistry, Merck Sharp & Dohme Research Laboratories, Neuroscience Research Centre,
Terlings Park, Harlow, Essex CM20 2QR, UK
bDepartment of Biochemistry, Merck Sharp & Dohme Research Laboratories, Neuroscience Research Centre,
Terlings Park, Harlow, Essex CM20 2QR, UK
cDepartment of Pharmacology, Merck Sharp & Dohme Research Laboratories, Neuroscience Research Centre,
Terlings Park, Harlow, Essex CM20 2QR, UK
Received 5 January 2000; accepted 25 April 2000
AbstractÐA new class of N-(indol-3-ylglyoxylyl)piperidines are high anity agonists at the benzodiazepine binding site of human
GABA-A receptor ion-channels, with modest selectivity for receptors containing the a1 subunit over a2 and a3. All three receptor sub-
types discriminate substantially between the two enantiomers of the chiral ligand 10. # 2000 Elsevier Science Ltd. All rights reserved.
Compounds that act as agonists at the benzodiazepine
binding site of the GABA-A receptor-chloride ion channel
have proved a rich source of clinically e€ective drugs,
particularly anxiolytics, hypnotics and anticonvulsants.
Unfortunately the use of these classical benzodiazepine
receptor (BZR) ligands can be compromised by the di-
culty of separating their many pharmacological e€ects in
man: for example, the treatment of anxiety with the
agonist diazepam is accompanied by unwanted side-
e€ects, including sedation, ataxia and the risk of depen-
dence.1,2 Mammalian GABA-A receptors have been
shown to be heteropentameric assemblies of protein
subunits (a,b,g,d,e,y,p,r), each of which is expressed as a
number of subtypes, where the most abundant receptor-
ion channels have the composition (an)2(bn)2(g).3 5 The
apparent di€erential localisation of these subtypes
within the brain4,6 o€ers the exciting possibility that
subtype selective ligands will allow separation of the
many pharmacological e€ects of BZR agonists. Recent
pharmacological and behavioural studies on transgenic
mice in which the a1 subunit was rendered benzodiaze-
pine insensitive support this hypothesis, and imply that
a1 BZRs are important for sedation but not for anxio-
lytic e€ects.7 Further study in this area requires the
identi®cation of more selective compounds, of which
few are known.8
Although N-(indol-3-ylglyoxylyl)amines (Fig. 1) are a
well established class of BZR ligands,9 there is currently
no reported data on their anities or ecacies at subtypes
of human GABA-A receptors. In this communication we
describe the evolution of new N-(indol-3-ylglyoxylyl)
piperidine agonists with high anity and modest bind-
ing selectivity for a1b2g2 containing human GABA-A
receptors.
The pyrrolidine amide 1 was identi®ed as a weak, non-
selective BZR ligand from screening of the corporate
sample collection. Compounds related to 1 were prepared
by literature procedures10 12 (Schemes 1, 2 and 3) and
their binding anities were measured by displacement of
*Corresponding author. Tel.: +1-279-440606; fax: +1-279-440187;
e-mail: ian_collins@merck.com
Figure 1. Representative structure.
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00245-6

1382
I. Collins et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1381±1384
[3H]Ro15-1788 from recombinant human GABA-A
receptors (a1b2g2; a2b3g2; a3b3g2) stably expressed in
L(tk ) cells13 (Table 1). These subtypes were chosen
because of their major contribution to benzodiazepine
sensitive GABA-A receptor populations in brain tissue.4
Classical benzodiazepines such as diazepam do not bind
to a4- and a6-containing receptors.3,4 Although a5-con-
taining receptors are sensitive to diazepam, the dis-
tribution of this subtype within the brain is limited.4,6b
The isoindoline amide 6, although of moderate anity,
showed a good selectivity between receptors containing
a1-subunits and those comprised of a2 (ca. 7-fold) and a3
(ca. 100-fold). Comparison of 4, 5 and 6 suggested that
the a3 subtype of the BZR was more sensitive to small
changes in ligand structure than the a1 or a2 assemblies.
Replacement of the pyrrolidin-2-yl acetate with nipecotate
(7) was tolerated and, for synthetic convenience, optimi-
sation of this structure was pursued (Scheme 2). Increas-
ing the size of the ester substituent was generally bene®cial
for anity at all receptor subtypes in both the pyrrolidinyl
and piperidinyl amides, illustrated by the increased anity
of the benzyl ester 8 relative to 7. The anity at the a1
subtype was maintained in the secondary amide analogue
9 whilst an increase in selectivity against the a2 and a3
BZRs was seen. A further substantial increase in anities
was achieved on substitution of the amide nitrogen to give
the racemic N-methyl-N-benzyl amide 10, with poor
selectivity for the a1 BZR subtype.
Investigation of substitution around the indole nucleus
was largely fruitless; only groups in the 7-position were
tolerated (compounds 2 and 3) and these gave no sig-
ni®cant improvement in anity, although a trend towards
higher anities was noted with electron-rich 7-sub-
stituents. Notably, substitution patterns that were optimal
in previously reported series of indole glyoxylamides,9a
e.g., 5-NO2, 5-Cl, gave much lower anity (Ki>1 0 mM
at all subtypes) than the unsubstituted lead 1. Deletion
of the pendant ester of the pyrrolidine lead 1 was not
tolerated, but it could be replaced by a phenyl group, as
in 4. In addition to this the conformationally constrained
benzylamine analogues 5 and 6 also retained anity.
As the BZR was known to discriminate between di€er-
ent enantiomers of some chiral indole glyoxylamides,9b
Scheme 1. Reagents and conditions: (i) (COCl)2, Et2O, 0 ꢀC; (ii) HNR2R3, Et3N, CH2Cl2, rt.
Scheme 2. Reagents and conditions: (i) LiOH, THF-H2O, rt (76±93%); (ii) CDI, ROH, Et3N, THF, rt or BOPCl, Et3N, NHR2R3, CH2Cl2, rt.

I. Collins et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1381±1384
1383
Scheme 3. Reagents and conditions: (i) H2, PtO2, AcOH (58%); (ii) BOC2O, Et3N, CH2Cl2, rt (87%); (iii) silica chromatography; (iv) HCl, EtOAc,
rt (100%); (v) indole-3-glyoxylyl chloride, Et3N, MeCN, rt.
Table 1. Anities of selected compounds at subtypes of human
GABA-A receptorsa
receptors (a5b3g2) but showed no appreciable anity
(Ki>350 nM; <50% inhibition of radiolabel binding at
test concentration of 2 mM in two independent determi-
nations).
Compound
Ki (nM)b
a1b2g2
a2b3g2
a3b3g2
The evolution of the high anity ligand R-10 could be
viewed as mainly the optimisation of lipophilic binding
of the substituent distal to the indole nucleus, but the
conformational e€ects of the amide rotamers present in
this class of compound were also considered. With the
exception of the symmetrically substituted isoindoline
amide 6, compounds 1±10 existed as mixtures of rota-
tional isomers about the glyoxylyl amide bond at room
temperature. The compounds 10 showed an additional
barrier to rotation about the pendant amide. This was
seen clearly in the NMR spectra of the compounds and
variable temperature NMR experiments, (e.g., of 1)
showed no coalescence of signals at temperatures up to
353K, indicating a substantial barrier to interconversion.
To assess the degree to which the population of lower
anity conformations could compromise the binding of
these compounds, symmetrical disubstituted analogues
of the nipecotate 7 were prepared (Scheme 3).
1
2
3
4
5
6
890 (Æ60)
510 (Æ70)
460 (Æ43)
440 (Æ30)
380 (Æ33)
200 (Æ28)
430 (Æ63)
440 (Æ82)
270 (Æ12)
830 (Æ200)
680 (Æ120)
1300 (Æ310)
570 (Æ70)
500 (Æ73)
320 (Æ34)
960 (Æ160)
720 (Æ50)
19,000c
(13,600, 26,400)
990c
7
880c
390c
(710, 1100)
77 (Æ11)
300 (Æ22)
14 (Æ2.5)
140 (Æ8.9)
7.3 (Æ1.5)
400c
(330, 450)
180 (Æ25)
970 (Æ55)
40 (Æ8.4)
250 (Æ42)
19 (Æ3.3)
200c
(890, 1100)
91 (Æ14)
8
9
10
(S)-10
(R)-10
11
1,500 (Æ70)
96 (Æ15)
670 (Æ13)
50 (Æ6.5)
250c
(270, 598)
14 (Æ3.0)
(178, 215)
22 (Æ1.9)
(245, 247)
23 (Æ2.2)
12
Diazepamd
Flunitrazepamd
12 (Æ2.2)
6.6 (Æ0.9)
33 (Æ5.3)
3.9 (Æ0.9)
1.1 (Æ0.5)
5.9 (Æ1.5)
aDisplacement of [3H]Ro15-1788 from hGABA-A subtypes stably
expressed in L(tk ) cells.
bMean (ÆSEM) for n=4±6, data expressed to 2 signi®cant ®gures.
cMean of n=2 expressed to 2 signi®cant ®gures, individual determi-
nations in parentheses.
Diethyl 3,5-pyridinedicarboxylate was exhaustively
hydrogenated to give a mixture of the cis and trans
piperidines, which were N-protected and separated by
chromatography. The individual piperidine isomers were
deprotected and coupled to indole-3-glyoxylyl chloride
as before to give 11 and 12. Both disubstituted analo-
gues were of higher anity at all BZR subtypes than
the parent 7, but the cis compound 12 showed a parti-
cularly dramatic (20- to 60-fold) increase in anity.
This increase in anity could be due to additional spe-
ci®c binding of the second ester substituent to the BZR,
coupled with the e€ect of the partial symmetry of the
piperidine moieties, which would render either amide
rotamer of 11 and 12 equally able to occupy the binding
site e€ectively. The especially high anity of the cis iso-
mer 12 may result from the relatively ¯at overall con-
formation of the diequatorially substituted 6-membered
ring, compared to the trans isomer 11 where one ester is
axially oriented and may interact less favourably with
dMean (ÆSD) for n>6 .
compound 10 was prepared as separate enantiomers
starting from resolved ethyl nipecotate14 (Scheme 2).
Slight degradation of the enantiopurity of the materials
occurred during the synthesis, presumably in the basic
hydrolysis step, and R-10 and S-10 were each isolated
with 91% enantiomeric excess as determined by chiral
analytical HPLC. Nevertheless, the R enantiomer was
clearly some 10- to 20-fold more active in the binding
assay than the S enantiomer, providing a compound
with excellent binding anity at the a1 BZR subtype,
although with modest selectivity over a2- and a3-con-
taining receptors. The high anity compounds 8, 10
and (R)-10 from this study were tested for their ability
to displace [3H]Ro15-1788 from human a5-containing

1384
I. Collins et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1381±1384
the receptor. Models of the binding of indole glyoxylyl
amides to the BZR emphasize the apparent planarity of
the binding site.9 Nonetheless, compound 11 still showed
an increase in anity relative to the parent 7 as a result of
both rotamers being able to present at least one sub-
stituent in the optimal orientation for binding.
subunit over a2 and a3. All three receptor subtypes dis-
criminate substantially between the two enantiomers of
the chiral ligand 10, and it is possible to overcome the
e€ect of low anity rotational conformers in this series
by appropriate symmetrical or pseudosymmetrical sub-
stitution of the piperidine ring.
Despite optimisation of the binding of these compounds
for the a1 BZR subtype, the selectivities observed
between the a1-, a2- and a3-containing receptors were at
best modest. There is high sequence homology between
the BZR subtypes,3 and this may account for the di-
culty in achieving high selectivities. The benzodiazepine
binding site is probably located between the a- and g-
subunits in the receptor complex.3 Although speci®c
residues on the a-subunits have been show to be critical
for benzodiazepine binding,15 it is also known that
amino acids in the g-subunits contribute to the binding
pocket.16
Acknowledgement
We thank Sharon Penn for VT NMR experiments on 1
and analogues.
References
1. Ashton, H. A. Drugs 1994, 48, 25.
2. In Martindale: The Complete Drug Reference; Par®tt, K.,
Ed.; Pharmaceutical Press: 1999; pp 661±668.
3. Stephenson, F. A. Biochem. J. 1995, 310, 1 and references
therein.
The benzodiazepine binding site is an allosteric mod-
ulatory site on the GABA-A receptor, and BZR ligands
can be classi®ed according to their ecacy; i.e., the abil-
ity to enhance (agonists), diminish (inverse agonists) or
leave unchanged (antagonists) the response of the chlo-
ride ion channel to the endogenous neurotransmitter g-
aminobutyric acid (GABA). Selected compounds were
therefore tested for their ability to a€ect the GABA-
induced chloride current in Xenopus oocytes transiently
expressing human GABA-A a1b2g2 receptors using two-
electrode voltage-clamp electrophysiology17 (Table 2).
4. McKernan, R. M.; Whiting, P. J. Trends Neurosci. 1996,
19, 139 and references therein.
5. Farrar, S. J.; Whiting, P. J.; Bonnert, T. P.; McKernan, R.
M. J. Biol. Chem. 1999, 274, 10100.
6. (a) Korpi, E. R.; Luddens, H. Br. J. Pharmacol. 1997, 120,
741. (b) Sur, C.; Fresu, L.; Howell, O.; McKernan, R. M.;
Atack, J. R. Brain Res. 1999, 822, 265.
7. 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.
8. (a) Atack, J. R.; Smith, A. J.; Emms, F.; McKernan, R. M.
Neuropsychopharmacology 1999, 20, 255. (b) Quirk, K.; Blur-
ton, P.; Fletcher, S.; Leeson, P.; Tang, F.; Mellilo, D.; Ragan,
C. I.; McKernan, R. M. Neuropharmacol. 1996, 35, 1331. (c)
Huang, Q.; Cox, E. D.; Gan, T.; Ma, C.; Bennett, D. W.;
McKernan, R. M.; Cook, J. M. Drug Des. Discovery 1999, 16,
55. (d) He, X.; Huang, Q.; Yu, S.; Ma, C.; McKernan, R.;
Cook, J. M. Drug Des. Discovery 1999, 16, 77.
9. (a) Martini, C.; Gervasio, T.; Lucacchini, A.; Da Settimo,
A.; Primo®ore, G.; Marini, A. M. J. Med. Chem. 1985, 28,
506. (b) Primo®ore, G.; Marini, A. M.; Da Settimo, F.; Mar-
tini, C.; Bardinelli, A.; Giannaccini, G.; Lucacchini, A. J.
Med. Chem. 1989, 32, 2514. (c) Bianucci, A. M.; Da Settimo,
A.; Da Settimo, F.; Primo®ore, G.; Martini, G.; Martini, C.;
Giannaccini, G.; Lucacchini, A. J. Med. Chem. 1992, 35, 2214.
(d) Primo®ore, G.; Da Settimo, F.; Marini, A. M.; La Motta,
C.; Martini, C.; Senatore, G.; Lucacchini, A. Il Farmaco 1995,
50, 5. (e) Da Settimo, A.; Primo®ore, G.; Da Settimo, F.;
Marini, A. M.; Novellino, E.; Greco, G.; Martini, C.; Gian-
naccini, G.; Lucacchini, A. J. Med. Chem. 1996, 39, 5083.
10. Shaw, K. N. F.; McMillan, A.; Gudmundson, A. G.;
Armstrong, M. D. J. Org. Chem. 1958, 23, 1171.
11. Mandell, L.; Roberts, E. C. J. Het. Chem. 1965, 2, 479.
12. Beak, P.; Lee, W. K. J. Org. Chem. 1993, 58, 1109.
13. Hadingham, K. L.; Wingrove, P.; Le Bourdelles, B.; Pal-
mer, K. J.; Ragan, C. I.; Whiting, P. J. Mol. Pharmacol. 1993,
43, 970.
14. Magnus, P.; Thurston, L. S. J. Org. Chem. 1991, 56, 1166.
15. McKernan, R. M.; Farrar, S.; Collins, I.; Emms, F.;
Asuni, A.; Quirk, K.; Broughton, H. B. Mol. Pharmacol. 1998,
54, 33.
The lead pyrrolidine 1 was a low ecacy partial agonist
at the a1-containing BZR. Both the a1-selective iso-
indolinamide 6 and the high anity diester 12 were also
partial agonists, with somewhat higher ecacy. In con-
trast, the high anity, selective N-methyl-N-benzyl-
amide (R)-10 was a full agonist with similar ecacy to
the benzodiazepines diazepam and ¯unitrazepam. The
association of higher ecacy with appropriately placed
aromatic or other lipophilic groups is a common feature
of many BZR pharmacophore models.18
In summary, a new class of N-(indol-3-ylglyoxylyl)piper-
idines are high anity agonists at the benzodiazepine
binding site of human GABA-A receptor ion-channels,
with modest selectivity for receptors containing the a1
Table 2. Ecacy of selected compounds at human a1b2g2 GABA-A
receptors
Compound
Test concentration
(mM)
Modulation of GABA EC20
current at a1b2g2 receptors (%)a,b
Diazepam
Flunitrazepam
1
6
(R)-10
12
1
1
30
3
1
1
+157 (Æ10)
+121 (Æ9)
+28 (Æ4)
+64 (Æ10)
+110 (Æ2)
+56 (Æ5)
16. Wingrove, P. B.; Thompson, S. A.; Wa€ord, K. A.;
Whiting, P. J. Mol. Pharmacol. 1997, 52, 874.
17. Wa€ord, K. A.; Whiting, P. D.; Kemp, J. A. Mol. Phar-
macol. 1992, 43, 240.
18. Zhang, W.; Koehler, K. F.; Zhang, P.; Cook, J. M. Drug
Des. Discovery 1995, 12, 193 and references therein.
aMean ( Æ SEM) for n=3±5.
bThe GABA EC20 (concentration eliciting 20% of maximum GABA
current) was determined for each oocyte (range 4±30 mM). The mod-
ulatory e€ects of each compound were then determined on coapplica-
tion with GABA at the EC20
.