Alpidem analogues containing a GABA or glycine moiety as new anticonvulsant agents

Article abstract of DOI:10.1016/S0928-0987(03)00002-2

Alpidem analogues containing a GABA (1-3) or glycine (4-6) moiety were synthesized and their interaction with the GABA/benzodiazepine receptor complex at central (CBR) and peripheral (PBR) level was evaluated. In particular, their ability to modulate the specific binding of [³H]-GABA to washed membrane preparations from the rat cerebral cortex, as well as their effects on human recombinant GABA_A receptors in Xenopus laevis oocytes, were assessed. Results from these in vitro assays showed that the most active compounds were 1 and 4. Intraperitoneal administration of compound 1 at a dose of 150 mg/kg significantly antagonized pentylenetetrazole-induced seizures in rats and the protective effects were evident for 90 min. However, compound 4 failed to interact with strychnine-sensitive Gly-binding sites. Consistent with these binding results, intraperitoneal administration of compound 4 at 150 mg/kg showed no effect against convulsions induced by strychnine, except for a prolonged time of the latency of convulsions. The results obtained suggest that compound 1 possesses interesting anticonvulsant activity and deserves further investigation as a novel lipophilic GABA derivative.

Full text of DOI:10.1016/S0928-0987(03)00002-2

European Journal of Pharmaceutical Sciences 18 (2003) 231–240
www.elsevier.com/locate/ejps
Alpidem analogues containing a GABA or glycine moiety as new
anticonvulsant agents
Giuseppe Trapania , , Andrea Latrofa , Massimo Franco , Antonio Carrieri , Saverio Cellamare ,
a
a
a
b
*
Mariangela Serrac, Enrico Sannac, Giovanni Biggioc, Gaetano Lisoa
a
`
`
Dipartimento Farmaco-Chimico, Facolta di Farmacia, Universita degli Studi di Bari, Via Orabona 4, 70125 Bari, Italy
b
`
Dipartimento Scienze del Farmaco, Universita degli Studi di Chieti ‘‘G. D’Annunzio’’, Via dei Vestini 31, I-66100, Chieti, Italy
c
`
Dipartimento di Biologia Sperimentale, Sezione di Neuroscienze, Universita di Cagliari, Cittadella Universitaria Monserrato,
SS 554 Km 4.5 Monserrato (Cagliari), Italy
Received 24 September 2002; accepted 30 December 2002
Abstract
Alpidem analogues containing a GABA (1–3) or glycine (4–6) moiety were synthesized and their interaction with the GABA/
benzodiazepine receptor complex at central (CBR) and peripheral (PBR) level was evaluated. In particular, their ability to modulate the
specific binding of [3H]-GABA to washed membrane preparations from the rat cerebral cortex, as well as their effects on human
recombinant GABAA receptors in Xenopus laevis oocytes, were assessed. Results from these in vitro assays showed that the most active
compounds were 1 and 4. Intraperitoneal administration of compound 1 at a dose of 150 mg/kg significantly antagonized
pentylenetetrazole-induced seizures in rats and the protective effects were evident for 90 min. However, compound 4 failed to interact
with strychnine-sensitive Gly-binding sites. Consistent with these binding results, intraperitoneal administration of compound 4 at 150
mg/kg showed no effect against convulsions induced by strychnine, except for a prolonged time of the latency of convulsions. The results
obtained suggest that compound 1 possesses interesting anticonvulsant activity and deserves further investigation as a novel lipophilic
GABA derivative.
2003 Elsevier Science B.V. All rights reserved.
Keywords: Blood–brain barrier; Inhibitory neurotransmitters; Imidazopyridine compounds
1. Introduction
with the strychnine-insensitive Gly binding sites of the
N-methyl-D-aspartate (NMDA) receptor complex poten-
tiates the excitatory effects of the neurotransmitter gluta-
mate (Aprison, 1990). A decrease in GABA-ergic and
Gly-ergic neurotransmissions appears to be involved in the
etiology of several neurological disorders, including anxie-
ty, pain and epilepsy. Several efforts have been devoted to
discovering and developing drugs capable of modifying the
GABA-ergic and Gly-ergic systems because of their
potential for treating these pathological states. For the
management of epilepsy, in particular, it is well known
that currently available drugs often do not provide
adequate control of seizures. Therefore, because about
20% of all patients remain refractory to current drugs or
suffer from several side effects, there is still a need for new
antiepileptic drugs (Lloyd and Morselli, 1983; Krogsgaard-
Larsen, 1988; McNamara et al., 1993).
g-Aminobutyric acid (GABA) and glycine (Gly) are the
main inhibitory neurotransmitters in the mammalian cen-
tral nervous system (CNS). Both Gly and GABA receptors
are coupled to membrane chloride channels, which have
very similar multiconductance properties (Krogsgaard-
Larsen, 1990). There is electrophysiological evidence that
GABA and Gly may share the same chloride conductance
channel (Krogsgaard-Larsen, 1990). Gly, however, ex-
hibits unique properties, being an inhibitory neurotrans-
mitter and, simultaneously, a co-agonist of the excitatory
neurotransmitter glutamate (Aprison, 1990). The inhibitory
effect of Gly is mediated through the strychnine-sensitive
Gly-receptor, a ligand-gated ion channel present mainly in
the brain and spinal cord. On the other hand, interaction
Penetration of the blood–brain barrier (BBB) by GABA
and Gly is poor, due to their hydrophilic nature and the
absence of active transport systems (Krogsgaard-Larsen,
*
Corresponding author. Tel.: 139-80-544-2764; fax: 139-805-442-
724.
E-mail address: trapani@farmchim.uniba.it (G. Trapani).
0928-0987/03/$ – see front matter
doi:10.1016/S0928-0987(03)00002-2
2003 Elsevier Science B.V. All rights reserved.

232
G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
1990; Geurts et al., 1998). Administration of large doses of
these neurotransmitters are necessary to produce neuro-
pharmacological effects (Toth et al., 1983).
tions as carrier groups in order to increase BBB penetra-
tion and delivery of the resulting molecules to the target
site. A number of advantages were expected: good lipo-
philicity, BBB penetration, recognition of and delivery
enhancement at the target site. Synthesis, binding data and
results of electrophysiological studies on compounds 1–6,
as well as the anticonvulsant potency of the most active
ones, are herein reported and discussed.
Among the approaches aimed at enhancing the delivery
of biologically active substances to the brain, chemical
modification of the active compounds, yielding appropriate
prodrugs, is widely applied. Therefore, it may be useful to
manipulate the molecules of GABA and Gly to obtain
compounds with increased lipophilicity, and thereby allow
them to gain access to the CNS. In this paper, we describe
the synthesis and properties of compounds 1–3 and 4–6
(Table 1) which are characterized by a GABA- or Gly-
ethyl ester moiety linked by an amide bond to an appro-
priate phenylimidazopyridine portion. This moiety was
chosen for two reasons. The first is of a pharmacodynamic
nature and is based on the high affinity and selectivity for
the GABA-benzodiazepine receptor complex shown by
most phenylimidazopyridine compounds, as exemplified
by alpidem and zolpidem (Benavides et al., 1993; Durand
et al., 1992). The second reason is that the lipophilic
phenylimidazopyridine moiety could serve as a carrier for
the amino acids taking into account of the high BBB
crossing ability shown by most phenylimidazopyridine
derivatives (e.g. log Cbrain /Cblood of alpidem50.4, Durand
et al., 1992). We therefore attempted to demonstrate the
feasibility of utilizing the covalent attachment of inhibitory
neurotransmitter molecules to phenylimidazopyridine por-
2. Materials and methods
2.1. Chemistry
Melting points were determined in open capillary tubes
¨
with a Buchi apparatus and are uncorrected. IR spectra
were obtained on a Perkin-Elmer 283 spectrophotometer
(KBr pellets for solid or nujol for liquid). 1H-NMR spectra
were determined on a Varian 390 or Bruker 300 MHz
instrument. Chemical shifts are given in d values down-
field from Me4Si as internal standard. Mass spectra were
recorded on a Hewlett-Packard 5995c GC–MS low-res-
olution spectrometer. All compounds showed appropriate
1
IR, H-NMR and mass spectra. Elemental analyses were
carried out with a Carlo Erba model 1106 analyzer and
results were within 60.40% of the theoretical values.
Silica gel 60 (Merck 70–230 mesh) was used for column
Table 1
Structure, lipophilicity, penetration of BBB of compounds 1–6
Compound
R1
R2
R3
R4
R5
n
CLog Pa
RP-HPLC
log k
PSAc
(A )
Log BBd
2
9b
˚
1
2
3
4
5
6
Cl
Cl
Cl
Cl
Cl
Cl
H
Cl
H
Cl
Cl
H
H
H
H
H
H
H
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
3
3
3
1
1
1
4.32
4.32
5.03
3.73
3.74
4.45
0.74
0.64
0.86
0.56
0.55
0.79
61.2
61.7
61.8
55.4
55.5
55.4
20.11
20.12
20.01
20.11
20.11
20.01
Cl
Cl
H
Cl
Cl
Cl
Alpidem
GABA
Glycine
Cl
H
Cl
C3H7
C3H7
0
5.37
22.77
23.21
0.40e
a Calculated according to CLOGP software.
b Capacity factors (k9) of each solute calculated as: k95(tR 2t0)/t0, where tR is the retention time of the solute and t0 is the column dead time.
c Calculated according to MAREA software.
d Calculated according to Clark’s model.
e Durand et al., 1992.

G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
233
chromatography. All the following reactions were per-
2.1.2. General procedure for the preparation of N-
(ethoxycarbonyl-methyl) (2-phenylimidazo[1,2-
a]pyridine3-yl)-acetamides 4–6
formed under a nitrogen atmosphere. The starting 2-amino-
pyridine compounds are commercially available. The
preparation of compounds 9 and 10 has previously been
reported (Trapani et al., 1999).
These compounds were prepared as above by using Gly
ethyl ester.HCl instead of GABA ethyl ester.HCl.
N-(Ethoxycarbonyl-methyl) [2-(4-chlorophenyl)-6-chlo-
ro-imidazo[1,2-a]pyridine3-yl)]-acetamide (4): m.p. 177–
181 8C. Yield 42%. IR (KBr) 3260, 1740, 1670, 1640
2.1.1. General procedure for preparation of N-(3-
ethoxycarbonylpropyl) (2-phenylimidazo[1,2-a]pyridine3-
yl)-acetamides 1–3
1
cm21; H NMR (CDCl3) d: 1.27 (t, J56Hz, 3H, CH3),
3.93 (s, 2H, CH2), 4.08 (d, J56 Hz, 2H, CH2), 4.21 (q,
J56 Hz, 2H, CH2), 6.95 (br s, 1H, NH), 7.1–7.5 (m, 6H,
Ar), 8.06 (s, 1H, Ar); MS m/z 405 (M1, 8), 275 (base).
Anal. (C19H17Cl2N3O3) C, H, N.
To an ice cooled and stirred solution of the suitable ester
10 (1 g) in EtOH (15 ml), 1 M NaOH (3 ml) was added
dropwise. After the addition was completed, stirring was
prolonged at room temperature for 4 h. Evaporation of the
solvent under reduced pressure gave a residue which was
dissolved in water, acidified with dilute HCl (pH 1) and
the corresponding crude acid 11 was extracted with
chloroform (3330 ml) (90 yield%). To a solution of the
crude acid 11 (1.6 mmol) in dry THF (20 ml), CDI (1.8
mmol) was added portionwise and stirred at room tempera-
ture for after 15 min; then GABA ethyl ester.HCl (2.4
mmol) was added and stirring prolonged for 30 min.
Evaporation of the solvent under reduced pressure gave a
residue which was washed with water, extracted with ethyl
ether and evaporated. The crude product was purified by
silica gel column chromatography [light petroleum ether–
ethyl acetate (1:1, v/v) as eluent] to give the required
amide. Yields and physical properties of compounds 1–3
are summarized in Table 1.
N-(Ethoxycarbonyl-methyl) (2-phenyl-6,8-dichloro-im-
idazo[1,2-a]pyridine3-yl)-acetamide (5): m.p. 119 8C dec.
1
Yield 15%. IR (KBr) 3260, 1740, 1670, 1640 cm21; H
NMR (CDCl3) d: 1.24 (t, J56Hz, 3H, CH3), 3.99 (s, 2H,
CH2), 4.02 (d, J56 Hz, 2H, CH2), 4.18 (q, J56 Hz, 2H,
CH2), 6.13 (br s, 1H, NH), 7.2–7.7 (m, 6H, Ar), 8.11 (s,
1H, Ar); MS m/z 405 (M1, 15), 275 (base). Anal.
(C19H17Cl2N3O3) C, H, N.
N-(Ethoxycarbonyl-methyl)
dichloro-imidazo[1,2-a]pyridine3-yl)]-acetamide (6): m.p.
190–193 8C. Yield 16%. IR(KBr) 1720, 1620 cm21 1H
[2-(4-chlorophenyl)-6,8-
;
NMR (CDCl3) d: 1.25 (t, J56Hz, 3H, CH3), 3.96 (s, 2H,
CH2), 4.02 (d, J56 Hz, 2H, CH2), 4.19 (q, J56 Hz, 2H,
CH2), 6.22 (br s, 1H, NH), 7.2–7.7 (m, 6H, Ar), 8.09 (s,
1H, Ar); MS m/z 439 (M1, 14), 311 (base). Anal.
(C19H16Cl3N3O3) C, H, N.
N-(3-Ethoxycarbonylpropyl)
[2-(4-chlorophenyl)-6-
chloro-imidazo[1,2-a]pyridine3-yl)]-acetamide (1): m.p.
2.1.3. Computational calculations of physicochemical
and permeability characteristics of compounds 1–6
Lipophilicity and log Cbrain /Cblood (log BB) values were
estimated by using CLOGP (v. 4, BioByte) and MAREA (v. 1.4
Department of Pharmacy, Uppsala, Sweden; Stenberg et
al., 1999) softwares. The lipophilicity indexes of com-
pounds 1–6, expressed as log k95log (tR 2 t0)/t0, were
obtained eluting these compounds on a reversed- phase
Symmetry C18 (25 cm33.9 mm; 5 mm particles) with
methanol–water (75:25, v/v) and a flux of 0.8 ml/min.
High-performance liquid chromatography (HPLC) analy-
ses were performed with a Water Associates Model 600
pump equipped with a Water 990 variable wavelength UV
detector. The column effluent was monitored continuously
at 254 nm. The HPLC mobile phase was prepared from
HPLC-grade methanol.
178–180 8C. Yield 45%. IR (KBr) 3280, 1720, 1640 cm21
;
1H-NMR (CDCl3) d: 1.19 (t, J56 Hz, 3H, CH3), 1.7–1.8
(m, 2H, CH2), 2.27 (t, J56Hz, 2H, CH2), 3.2–3.3 (m, 2H,
CH2), 3.92 (s, 2H, CH2), 4.03 (q, J56Hz, 2H, CH2), 6.22
(br s, 1H, NH), 7.1–7.6 (m, 6H, Ar), 8.10 (s, 1H, Ar); MS
m/z 433 (M1, 6), 275 (base). Anal. (C21H21Cl2N3O3) C,
H, N.
N-(3-Ethoxycarbonylpropyl) (2-phenyl-6,8-dichloro-im-
idazo[1,2-a]pyridine3-yl)-acetamide (2): m.p. 129–132 8C.
1
Yield 17%. IR (KBr) 3280, 1730, 1630 cm21; H NMR
(CDCl3) d: 1.18 (t, J56Hz, 3H, CH3), 1.7–1.8 (m, 2H,
CH2), 2.25 (t, J56Hz, 2H, CH2), 3.2–3.3 (m, 2H, CH2),
3.87 (s, 2H, CH2), 4.01 (q, J56 Hz, 2H, CH2), 6.53 (br s,
1H, NH), 7.2–7.6 (m, 6H, Ar), 8.09 (s, 1H, Ar); MS m/z
433 (M1, 11), 275 (base). Anal. (C21H21Cl2N3O3) C, H,
N.
N-(3-Ethoxycarbonylpropyl) [2-(4-chlorophenyl)-6,8-
dichloroimidazo[1,2-a]pyridine3-yl)]-acetamide (3): m.p.
2.2. Biology
145–149 8C. Yield 45%. IR (KBr) 3280, 1720, 1635 cm21
;
1H NMR (CDCl3) d: 1.19 (t, J56Hz, 3H, CH3), 1.7–1.8
(m, 2H, CH2), 2.28 (t, J56Hz, 2H, CH2), 3.2–3.3 (m, 2H,
CH2), 3.81 (s, 2H, CH2), 4.02 (q, J56 Hz, 2H, CH2), 6.83
(br s, 1H, NH), 7.2–7.6 (m, 6H, Ar), 8.05 (s, 1H, Ar); MS
m/z 467 (M1, 6), 311 (base). Anal. (C21H20Cl3N3O3) C,
H, N.
Adult male or female Sprague–Dawley CD rats (Charles
River, Como, Italy) with body masses of 200–250 g at the
beginning the experiments, were maintained under an
artificial 12-h light–dark cycle (lights on 08.00 to 20.00 h)
at a constant temperature of 2362 8C and 65% humidity.
Food and water were freely available, and the animals

234
G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
were acclimatized for .7 days before use. Experiments
were performed between 08.00 and 14.00 h. Animal care
and handling throughout the experimental procedure were
performed in accordance with the European Communities
Council Directive of 24 November 1986 (86/609/EEC).
The experimental protocol was approved by the Animal
Ethical Committee of the University of Cagliari.
specific binding was defined as binding in the presence of
10 mM unlabelled PK 11195 (Sigma).
2.3.3. [3H]-GABA binding assay
Fresh cerebral cortices were homogenized with a Poly-
tron PT 10 (setting 5 for 30 s) in 10 volumes of ice-cold
water and centrifuged 10 min at 48 000 g at 0 8C. The
pellet was washed once by resuspension and recentrifuga-
tion in 10 volumes of 20 mM K2HPO4 –KH2PO4 buffer
(pH 7.4) containing 50 mM KCl. The membranes were
stored at 220 8C until used 1–15 days later. On the day of
the assay the membranes were thawed and centrifuged.
The pellet was washed three additional times by resuspen-
sion and recentrifugation in ice-cold buffer. The tissue was
resuspended in 50 volumes of the same buffer and 700 ml
of membrane suspension (500–600 mg of protein) was
added to plastic minivials. Drugs were dissolved in DMSO
and serial dilutions were made up in DMSO and added in
5-ml aliquots. The total incubation volume was 1 ml.
Nonspecific binding was defined as binding in the presence
of 1 mM GABA. The incubation (10 min at 4 8C) was
started by the addition of 40 nM [3H]-GABA (spec. act. 89
Ci/mmol) and was stopped by centrifugation of the
incubation mixture at 48 000 g for 10 min. The supernatant
was discarded and the pellet was washed gently twice with
4 ml of ice-cold distilled water and was then dissolved in
3 ml of scintillation fluid (Ultima Gold, Camberra Pac-
kard).
2.3. In vitro receptor binding assays
2.3.1. [3H]-Flunitrazepam binding
The cerebral cortex was homogenized with a Polytron
PT 10 in 50 volumes of ice-cold 50 mM Tris–HCl buffer
(pH 7.4) and centrifuged twice at 20 000 g for 10 min. The
pellet was reconstituted in 50 volumes of Tris–HCl buffer
and was used for the binding assay. Aliquots of 400 ml of
tissue homogenate (0.4–0.5 mg of protein) were incubated
in presence of [3H]-flunitrazepam (spec. act. 7.4 Ci/mmol,
New England Nuclear) at a final concentration of 0.5 nM,
in a total incubation volume of 1000 ml. Drugs were
dissolved in dimethylsulfoxide (DMSO) and serial dilu-
tions were made up in DMSO and added in 100 ml
aliquots. After a 60-min incubation at 0 8C, the assay was
determined by rapid filtration through glass-fiber filter
strips (Whatman GF/B). The filters were rinsed with 2- to
4-ml portions of ice-cold Tris–HCl buffer as described
above. Radioactivity bound to the filters was quantitated
by liquid scintillation spectrometry (Ultima Gold, Camber-
ra Packard). Nonspecific binding was determined as bind-
ing in the presence of 5 mM diazepam, and represented
about 10% of total binding.
2.3.4. [3H]-Strychnine binding assay
Rats were killed by guillotine and the spinal cord was
rapidly dissected and homogenized on ice with PTFE
pestle and a glass homogenizer in 10 volumes of 0.32 M
saccharose. The homogenate was centrifuged at 1000 g for
10 min at 4 8C and the supernatant was collected and
centrifuged at 18 000 g for 20 min at 4 8C. The resulting
pellet was resuspended by homogenization with a Polytron
PT 10 (setting 3 for 30 s) in five volumes of ice-cold
distilled water and centrifuged at 8000 g for 20 min at
4 8C. The whole supernatant was collected and centrifuged
at 40 000 g for 20 min at 4 8C. The final pellet was then
reconstituted in 1 ml of an ice-cold solution containing (in
2.3.2. [3H]-PK 11195 binding
After sacrifice, brain and ovary were rapidly removed,
the cerebral cortex was dissected and both tissues were
stored at 280 8C until assay. The tissues were thawed and
homogenized in 50 volumes of Dulbecco’s phosphate
buffered saline (PBS) pH 7.4 at 4 8C with a Polytron PT
10 (setting 5, for 20 s). The homogenate was centrifuged at
40 000 g for 30 min, and the pellet was resuspended in 50
volumes of PBS and recentrifugated. The new pellet was
resuspended in 20 volumes of PBS and used for the assay.
[3H]-PK 11195 binding was determined in a final volume
of 1000 ml tissue homogenate (0.15–0.20 mg protein), 100
ml of [3H]-PK 11195 (spec. act. 85.5 Ci/mmol, New
England Nuclear) at final assay concentration of 1 nM, 5
ml of drug solution or solvent and 795 ml of PBS buffer
(pH 7.4 at 25 8C). Drugs were dissolved in DMSO and
serial dilutions were made up in DMSO and added in 5-ml
aliquots. Incubations (0 8C) were initiated by addition of
membranes and were terminated 90 min later by rapid
filtration through glass-fiber filter strips (Whatman GF/B),
which were rinsed with two 4 ml of ice-cold PBS buffer
using Cell Harvaster filtration manifold (Brandel). Filter
bound radioactivity was quantified by liquid scintillation
spectrometry (Ultima Gold, Camberra Packard). Non-
mM):
4-(2-hydroxyethyl)-piperazine-1-ethanesulphonic
acid (HEPES)–Tris, 10; CaCl2, 1; NaCl, 145; KCl, 5;
MgCl2, 2.13; glucose, 10 (pH 7.5) and stored at 280 8C
until use (1–15 days later). On the day of the assay, the
membranes were thawed and reconstituted with a Polytron
PT 10 (setting 3 for 30 s) in buffer solution to a final
protein concentration of ¯1 mg/ml and used for the
binding assay. Protein concentration was determined by
the method of Lowry et al. (1951).
[3H]-Strychnine binding was studied by the procedure
described by Pfeiffer and Betz (1981) with minor modi-
fications. [3H]-strychnine binding was determined in a final
volume of 500 ml consisting of 200 ml of membrane
suspension, 50 ml of 8 nM [3H]-strychnine (specific

G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
235
activity 23 Ci/mol, New England Nuclear), 50 ml of drug
solution or solvent, and buffer solution (pH 7.5) to
volume. Drugs were dissolved in DMSO and serial dilu-
tions were made up in DMSO and added in 5-ml aliquots.
Nonspecific binding was defined as binding in the presence
of 100 mM strychnine. Incubations (0 8C) were initiated by
addition of membranes and terminated after 60 min by
rapid filtration through glass-fiber strips (Whatman GF/B).
The filters were rinsed with two 4-ml portions of ice-cold
buffer in a Cell Harvester filtration manifold (Model M-24;
Brandel Instruments, Gaithersburg, MD, USA), and filter-
bound radioactivity was quantified by scintillation spec-
trometry (Ultima Gold, Camberra Packard).
tration that induces a Cl– current with an amplitude of
2065% of the maximal response evoked by 1 mM GABA
and which was determined for each oocyte at the begin-
ning of the experiment) was used to produce the control
response.
2.3.6. Pharmacology
Compounds 1 and 4 were suspended in distilled water
with a drop of Tween 80 per 5 ml and injected in a volume
of 3 ml, per kilogram of body weight. Control rats received
an equivalent volume of vehicle. Rats were injected with
pentylenetetrazole (55 mg/kg) or strychnine (5 mg/kg) at
the indicated times after injection of vehicle, compounds 1
and 4, respectively. The animals were observed for 15 or
30 min after injection of pentylentetrazole or strychnine,
respectively, during which time of the latency of seizures
and death were recorded. Data were analyzed by Fisher’s
exact probability test or Student’s t test.
2.3.5. Functional in vitro studies: electrophysiological
studies using Xenopus oocytes
Complementary DNAs encoding the human a1, b2 and
,
g2L GABAA receptor subunits were subcloned into the
pCDM8 expression vector (Invitrogen, San Diego, CA,
USA). The cDNAs were purified with the Promega Wizard
Plus Miniprep DNA Purification System (Madison, WI,
USA) and then resuspended in sterile distilled water,
divided into portions, and stored at 220 8C until used for
injection. Stages V and VI oocytes were manually isolated
from sections of Xenopus laevis ovary, placed in modified
Barth’s saline (MBS) containing 88 mM NaCl, 1 mM KCl,
10 mM Hepes–NaOH (pH 7.5), 0.82 mM MgSO4, 2.4
mM NaHCO3, 0.91 mM CaCl2, and 0.33 mM Ca(N03)2
and treated with 0.5 mg/ml of collagenase Type IA
(Sigma) in collagenase buffer (83 mM NaCl, 2 mM KCl,
1 mM MgCl2, 5 mM Hepes–NaOH, pH 7.5) for 10 min at
room temperature, to remove the follicular layer. A
mixture of GABAA receptor a1, b2 and g2 subunit cDNAs
(1.5 ng/30 nl) were injected into the oocyte nucleus using
a 10-ml glass micropipette (10–15 mm tip diameter). The
injected oocytes were cultured at 19 8C in sterile MBS
supplemented with streptomycin (10 mg/ml), penicillin
(10 U/ml), gentamicin (50 mg/ml),0.5 mM theophylline,
and 2 mM sodium pyruvate. Electrophysiological record-
ings began approximately 24 h following cDNA injection.
Oocytes were placed in a 100-ml rectangular chamber and
continuously perfused with MBS solution at a flow-rate of
2 ml/min at room temperature. The animal pole of oocytes
was impaled with two glass electrodes (0.5–3 MV) filled
with filtered 3 M KCl and the voltage was clamped at 270
mV with an Axoclamp 2-B amplifier (Axon Instruments,
Burlingame, CA, USA). Currents were continuously re-
corded on a strip-chart recorder. Resting membrane po-
tential usually varied between 230 and 250 mV. Drugs
were dissolved in DMSO and diluted to the final con-
centration with MBS. In control experiments, DMSO at
concentrations of 1–3% did not significantly affect GABA-
evoked Cl– currents. Oocytes were exposed to drugs for 20
s (7–10 s were required to achieve equilibrium in the
recording chamber). Intervals of 5–10 min were allowed
between drug applications. GABA at the EC20 (concen-
3. Results and discussion
3.1. Chemistry
According to Scheme 1, compounds 1–6 were prepared
by condensing the appropriate carboxylic acid 11 with
GABA- or Gly-ethyl esters.HCl in the presence of car-
bonyldiimidazole (CDI) as dehydrating agent. Compound
11, in turn, was obtained by alkaline hydrolysis of esters
10 synthesized by condensation of suitably substituted
2-aminopyridines 8 with the appropriate bromoketoesters
9. By these synthetic procedures, compounds 1–6 were
prepared in moderate to good yield and were characterized
by IR, 1H-NMR and mass spectroscopy. Table 1 reports the
structures and some physicochemical characteristics of
compounds 1–6.
3.2. Lipophilicity and computational approach to
estimate BBB penetration by compounds 1–6
To be useful, compounds acting at the CNS level should
possess good BBB penetration properties. Such a charac-
teristic is roughly related to the compound’s overall
lipophilicity, which is reflected by its partition coefficient
(log P) (Bodor and Buchwald, 1999). Therefore, the log P
values between n-octanol and water of compounds 1–6
were calculated by using the CLOGP computer program and
measuring retention in RP-HPLC (Table 1). A satisfactory
correlation was found between the calculated log P values
and the capacity factors log k9, and this analysis indicated
a linear relationship with acceptable statistics (n56, r25
0.89, s50.18). Log P values of 2.1 are generally consid-
ered optimal for penetration of the BBB (Norinder and
Haeberlein, 2002). As can be seen from Table 1, com-
pounds 1–6 possess log P values higher than 2.1 (i.e. in
the range 3.73 –5.03) and appeared to be lipophilic enough

236
G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
Scheme 1.
to cross the BBB. To gain further information on the BBB
penetration properties of compounds 1–6 by passive
transport, additional computational studies were carried
out. Recently, several attempts to correlate BBB penetra-
tion with physicochemical parameters have been reviewed
(Norinder and Haeberlein, 2002). It has been made clear
that the octanol–water partition coefficient is an important
factor, although in itself it correlates poorly with the log
Cbrain /Cblood (log BB). This ratio is a measure of the
degree of BBB penetration. Experimental values of log BB
published to date cover a range of about 22.00 to 11.00.
Compounds with log BB .0.3 cross the BBB readily,
while compounds with log BB,21.0 are poorly distribut-
ed to the brain. In addition to log P, the importance of a
molecular size descriptor has been shown, as well as the
need to include a descriptor relating to hydrogen bond
formation (Van de Waterbeemd et al., 1998). To readily
estimate BBB penetration by compounds 1–6, we used the
model of Clark (1999) to predict the log BB of compounds
heteroatoms) and calculated log P (Clog P) according to
the equation: log BB520.0148 (60.001)PSA10.152
(60.036)Clog P10.139 (60.073). Calculation of the
polar surface area of compounds 1–6 was performed using
the MAREA software (Stenberg et al., 1999). As can be seen
from the data reported in Table 1, on the basis of this
computational approach, significant brain penetration can
be predicted for compounds 1–6.
3.3. In vitro studies
3.3.1. Radioligand binding assays
GABAA receptors are allosterically modulated by sever-
al types of compounds, including ethanol, benzodiaze-
pines, barbiturates and neuroactive steroids (Majewska et
al., 1986). Moreover, there is overwhelming evidence that
activation of peripheral benzodiazepine receptors (PBR)
stimulates the synthesis of neuroactive steroids (Giesen-
Crouse, 1993). Apart from their pharmacological and
biological functions, PBRs also differ from central benzo-
diazepine receptors (CBR) in their subcellular location in
the brain, being present in glial cells (Giesen-Crouse,
1–6. This model relates log BB to polar surface area (PSA,
2
˚
defined as the surface area in A occupied by nitrogen and
oxygen atoms and polar hydrogens connected to these

G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
237
1993). Nevertheless, PBR and CBR receptors share a
common role in the CNS, where both modulate GABA-
ergic transmission, PBRs doing so indirectly, by influenc-
ing neurosteroidogenesis, while CBRs act directly. It
should be noted that alpidem interacts with both CBRs and
PBRs. Previous works from these laboratories (Trapani et
al., 1997, 1999) pointed out that substitution of the
alpidem at position 8 with lipophilic substituents and the
presence of one chlorine atom at the para position of the
phenyl ring at C(2) are key structural features for high
binding affinity and selectivity toward PBRs.
Evaluations to assess the interaction of compounds 1–6
with the GABA/benzodiazepine receptor complex at cen-
tral (CBR) and peripheral (PBR) level were performed by
measuring their ability to inhibit [3H]-flunitrazepam and
[3H]-PK 11195 (Le Fur et al., 1983) binding to membrane
preparations arising from the cerebral cortex. The reference
compound in these binding assays was zolpidem and the
measured binding affinities for CBR and PBR are shown in
Table 2. As can be seen, compounds 1–6 displayed low
affinity and selectivity for PBR. Only compounds 1 and 4
(i.e. the 8-unsubstituted imidazopyridine conjugates of
GABA and Gly, respectively) possess a significant binding
affinity towards CBR. It should also be noted that the acid
11 (R1 5R3 5Cl, R2 5H) displayed very low ability to
displace [3H]-flunitrazepam from the cerebral cortex (IC50
7 mM). These binding results suggest that the most active
compounds (i.e. 1 and 4) interact at the central (neuronal)
but not at the peripheral (glial) receptors.
in [3H]-GABA binding modulation. Among the Gly-ana-
logues, compound 4 showed a significant increase in [3H]-
GABA binding, even though in a concentration indepen-
dent manner. Compounds 5 and 6 displayed little signifi-
cant increase in [3H]-GABA binding. Compounds 1 and 4,
however, did not show any direct action (GABA-mimetic)
on chloride currents even at high concentrations (100
mM), thus confirming that their effects are modulatory in
nature. On the whole, results from this in vitro test are
consistent with those observed in benzodiazepine receptor
binding assays, pointing to analogues 1 and 4 as the most
active compounds.
Among the Gly-analogues, compound 4 was chosen to
evaluate its ability to interact with the strychnine-sensitive
Gly-binding sites. It was found that this compound failed
to inhibit [3H]-strychnine binding at the two concentrations
tested (i.e. 1 mM and 10 mM).
3.3.2. Electrophysiology in Xenopus oocytes
Compounds 1–6 were evaluated both for their effects at
human recombinant GABAA receptors, in order to de-
termine their capacity to interact at CBR. As summarized
in Table 3, Cl2 currents elicited by GABA at GABAA
receptors were enhanced by compounds 1 and 4, with
maximal effects observed at 10 mM (111.8 and 80,
respectively). Compounds 2, 3, 5 and 6 failed to produce
statistically significant changes in the amplitude of GABA-
elicited currents. It was also proved that acid 11 (R1 5
R3 5Cl, R2 5H) displayed a much lower ability to enhance
GABA-evoked currents compared to compound 1 (data not
shown). The ability of compounds 1 and 4 to inhibit
GABA-induced Cl2 currents was flumazenil-sensitive
(Table 3), indicating that the effect is mediated by the
benzodiazepine recognition sites.
As shown in Table 2, compounds 1–6 were also
evaluated for their ability to modulate the specific binding
of [3H]-GABA to washed membrane preparations from the
rat cerebral cortex. The reference compound in this binding
assay was abecarnil and the measured binding affinities for
[3H]-GABA are shown also in Table 2. Among the
GABA-analogues 1–3, only compound 1 enhanced the
specific binding of [3H]-GABA (about 30% above the
control value), whereas the reference compound showed no
effect (Table 2). Compounds 2 and 3 displayed a decrease
3.4. In vivo studies
The next aspect of these studies was to examine some in
vivo properties of compounds 1 and 4. Compound 1
Table 2
Affinities of compounds 1–6 for CBR and PBR and modulatory action at [3H]-GABA binding
Compound
IC50 (nM)a
[3H]-GABA bindinga
CBR
PBR
2600
57 000
5500
1080
6700
1600
–
1 mM
10 mM
30 mM
100 mM
130
1
2
3
4
5
6
60
.104
.104
150
117
270
13063b
251
127
215
216
225
218
12668.8c
116
13464.5c
112
14660b
112
.104
.104
7000
26
114
15
118
11 (R1 5R3 5Cl, R2 5H)
Zolpidem
–
Abecarnil
25
22
213
a Mean of two determinations. Relative standard deviation (R.S.D.) values were less than 8%.
b Mean of two experiments6standard deviation.
c Mean of three experiments6standard deviation.

238
G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
Table 3
Modulatory action of compounds 1–6 at human a1b2g2L GABAA receptors expressed in Xenopus Laevis oocytesa
Compound
0.1 mM
26.969.5
0.3 mM
1 mM
3 mM
10 mM
1
41.2618.4
59.2610.5
91.2610.9
(13.663)b
12.065.0
0.4861.2
70.968.7
(5630)b
111.8617.7
2
3
4
15.066.1
20.1468.6
24.95618.9
1064.0
4.8465.7
15.8612.7
6.166.0
3.564.7
44.9611.5
3.263.2
9.964.0
81.866.49
5
6
23.3613.8
213.7618.6
3.968
20.763.1
2.562.3
9.362.4
11.469
21.264.1
6.3612.3
15.867.8
a Values represent the potentiation of Cl2 currents induced by GABA and are means (n55–8 different oocytes)6standard deviation.
b Values observed in the presence of flumazenil at 1 mM.
significantly antagonized pentylenetetrazde-induced sei-
zures in rats. The time course of its anticonvulsant activity
is shown in Table 4. Compound 1 administered 30 min
before pentylenetetrazole decreased the number of animals
showing convulsions, and completely protected animals
from death. These protective effects were still evident,
although not significant, 90 min after administration of
compound 1, while no differences were found between
animals that had received the drug 120 min before
pentylenetetrazole and controls. As for the times of the
latency of seizures they were highest at 90 min after
administration of compound 1, and it is indicative of a
prolonged pharmacological action. As expected, adminis-
tration of GABA at a dose of 150 mg/kg, 30 min before
pentylenetetrazole, failed to antagonize convulsions.
Administration of compound 4 at the dose of 150 mg/kg
failed to antagonize convulsions induced by administration
of strychnine (Table 4), and this is consistent with the
corresponding binding assay. The times of the latency of
convulsions were highest at 120 min after administration
of compound 4 suggesting again a prolonged action in this
pharmacological test.
and 4, two hypotheses may be put forward: (i) a mecha-
nism whereby these alpidem analogues remain intact and
act as true benzodiazepine receptor ligands regulating
GABA interaction by allosteric modulation; (ii) a prodrug
mechanism whereby, once at the brain, 1 and 4 undergo
chemical or enzymatic hydrolysis at the ester group to give
compound 12 and 13, respectively, or both at the ester and
amide group to give the corresponding neurotransmitters
and 11 (Scheme 2). Due to the fact that the amide bond is
known to be quite stable in vivo, formation of 11 might be
slow (Scheme 2). Both the proposed mechanisms also may
work in combination. The prodrug mechanism may be
proposed taking into account that 1 and 4, although stable
enough in solution buffered at pH 7.4, were found to
undergo a rapid cleavage in dilute rat serum at 37 8C (data
not shown). The delayed pharmacological effects observed
for compounds 1 and 4 may be explained by the prodrug
hypothesis. This last approach, indeed, is a well-estab-
lished strategy for obtaining sustained release (Wermuth et
al., 1996). On the other hand, N-benzoyl or N-pivaloyl, or
N,N-phthaloyl derivatives of GABA as well as conjugation
products of valproic acid and GABA or Gly are examples
of compounds able to penetrate the BBB, most of them
To explain the effects of the most active compounds 1
Table 4
Anticonvulsant activity of 1 and 4 in the pentylenetetrazole and strychnine tests over time after i.p. administration of 150 mg/kg of compound to rats
Test
Compound
Test
time
(min)
Convulsions
latencya
(s)
Convulsions
no. of
animals
Death
latency
(s)a
Death
no. of
animals
Pentylenetetrazole
(55 mg/kg)
Vehicle
1
1
1
9366
107612
139626
115613
9568
15/15
2/6b
5/8
8/8
8/8
351671
8/15
0/6c
1/8
2/8
6/8
30
90
120
30
240
23961
331655
GABA
Strychnine
(5 mg/kg)
Vehicle
4
4
4
166634
185646
166637
343677
197626
12/12
6/6
5/6
6/6
6/6
5586108
672666
4206114
450696
4866102
10/12
6/6
5/6
6/6
6/6
30
90
120
120
Glycine
Animals were treated with 150 mg/kg (i.p.) of 1 or 4 at the indicated times before pentylenetetrazole (55 mg/kg, i.p.) or strychnine (5 mg/Kg) and
observed for at least 15 min (for compound 1) and 30 min (for compound 4).
a Data are means of two determinations6S.D.
b P,0.01.
c P,0.025 (Fisher’s exact probability test).

G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
239
Scheme 2.
exhibiting anticonvulsant effects (Wermuth et al., 1996;
Bialer, 1999; Usifoh et al., 2001).
Acknowledgements
This work was supported by a grant from Ministero
`
dell’Universita e della Ricerca Scientifica e Tecnologica
(MURST). We thank Giovanni Dipinto for skillful techni-
cal assistance in recording mass spectra and determining
microanalytical data. Thanks are also due to Department of
Farmaceutics of Uppsala for providing us the MAREA
computer program.
4. Conclusion
The main goal of the present paper was to show that
alpidem analogues containing a GABA or Gly moiety,
prepared by covalently coupling these neurotransmitters to
an appropriate transport moiety (carrier group) represented
by a lipophilic phenylimidazopyridine moiety, may gain
access to the CNS by passive transport through the BBB
and interact with their target site. Taken together, the in
vitro results suggest that compounds 1 and 4 increases the
function of the GABA–ionophore complex by interaction
with the receptor. The observed in vivo anticonvulsant
activity of compound 1 led us to conclude that it may be
considered as a novel lipophilic GABA derivative, which
deserves further consideration. This study demonstrated
the feasibility of synthesizing useful anticonvulsants by
coupling the amine function of the GABA with a
phenylimidazopyridine portion. Moreover, the results pre-
sented in this article provide information regarding the
subsequent design of phenylimidazopyridine derivatives of
poorly BBB permeable drugs. In a companion paper
(manuscript in preparation) we show the results of further
studies aimed at elucidating both the chemical and en-
zymatic stability of compounds 1 and 4 and their brain
distribution by microdialysis, an in vivo model for study-
ing the penetration of BBB. These data provide insight into
the mechanism behind the effects shown by 1 and 4.
References
Aprison, M.H., 1990. In: Ottersen, O.P., Storm-Mathisen, J. (Eds.), The
Discovery of the Neurotransmitter Role of Glycine in Glycine
Neurotransmission. Wiley, New York, pp. 1–23.
Benavides, J., Peny, B., Ruano, D., Vitorica, J., Scatton, B., 1993.
Comparative autoradiographic distribution of central omega (benzo-
diazepine) modulatory site subtypes with high, intermediate and low
affinity for zolpidem and alpidem. Brain Res. 604, 240–250.
Bialer, M., 1999. Pharmacokinetic consideration in the design of better
and safer new antiepileptic drugs. J. Controlled Rel. 62, 187–192.
Bodor, N., Buchwald, P., 1999. Recent advances in the brain targeting of
neuropharmaceuticals by chemical delivery systems. Adv. Drug Deliv.
Rev. 36, 229–254.
Clark, D.E., 1999. Rapid calculation of polar molecular surface area and
its application to the prediction of transport phenomena. 2. Prediction
of blood–brain barrier penetration. J. Pharm. Sci. 88, 815–821.
Durand, A., Thenot, J.P., Bianchetti, G., Morselli, P.L., 1992. Compara-
tive pharmacokinetic profile of two imidazopyridine drugs: zolpidem
and alpidem. Drug Metab. Rev. 24, 239–266.
Geurts, M., Poupaert, J.H., Scriba, G.K.E., Lambert, D.M., 1998. N-
(Benzyloxycarbonyl)glycine esters and amides as new anticonvulsants.
J. Med. Chem. 41, 24–30.
Giesen-Crouse, E., 1993. In: Giesen-Crouse, E. (Ed.), Peripheral Benzo-
diazepine Receptors. Academic Press, London.

240
G. Trapani et al. / European Journal of Pharmaceutical Sciences 18 (2003) 231–240
Krogsgaard-Larsen, P., 1990. Amino acid receptors. In: Hansch, C.,
Sammes, P.G., Taylor, B.J. (Eds.), Comprehensive Medicinal Chemis-
try. Pergamon Press, Oxford, pp. 493–531, Chapter 12.7.
Krogsgaard-Larsen, P., 1988. GABA synaptic mechanisms: Stereochem-
ical and conformational requirements. Med. Res. Rev. 8, 27–56.
Le Fur, G., Perrier, M.L., Vaucher, N., Imbant, F., Flamier, A., Uzan, A.,
Renault, C., Dubroeucq, M.C., Gueremy, C., 1983. Peripheral binding
sites: effect of PK 11195, 1-(2-chlorophenyl)-N-(1-methylpropyl)-3-
isoquinolinecarboxamide. I. In vitro studies. Life Sci. 32, 1839–1847.
Lloyd, K.G., Morselli, P.L., 1983. Biochemistry, anatomy, and pharma-
cology of GABA neurons. In: Melzer, H.Y. (Ed.), Psychopharmacolo-
gy: The Third Generation of Progress. Raven Press, New York, pp.
183–201.
of endothelin receptor antagonists using three theoretical methods of
increasing complexity. Pharm. Res. 16, 1520–1526.
Toth, E., Lajhta, A., Sarhan, S., Seiler, N., 1983. Anticonvulsant effects
of some inhibitory neurotransmitter amino acids. Neurochem. Res. 8,
291–302.
Trapani, G., Franco, M., Ricciardi, L., Latrofa, A., Genchi, G., Sanna, E.,
Tuveri, F., Cagetti, E., Biggio, G., Liso, G., 1997. Synthesis and
binding affinity of 2-phenylimidazo[1,2-a]pyridine derivatives for both
central and peripheral benzodiazepine receptors. A new series of
high-affinity and selective ligands for the peripheral type. J. Med.
Chem. 40, 3109–3118.
Trapani, G., Franco, M., Latrofa, A., Ricciardi, L., Carotti, A., Serra, M.,
Sanna, E., Biggio, G., Liso, G., 1999. Novel 2-phenylimidazo[1,2-
a]pyridine derivatives as potent and selective ligands for peripheral
benzodiazepine receptors. Synthesis, binding affinity, and in vivo
studies. J. Med. Chem. 42, 3934–3941.
Lowry, O.M., Rosenbrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein
measurement with the Folin phenol reagent. J. Biol. Chem. 193,
265–275.
Majewska, M.D., Harrison, N.L., Schwartz, R.D., Barker, J.L., Paul,
S.M., 1986. Steroid hormone metabolites are barbiturate-like
modulators of the GABA receptor. Science 232, 1004–1007.
McNamara, J.O., Bonhaus, D.W., Shin, C., 1993. In: Schwartzkroin, P.A.
(Ed.), Epilepsy: Models, Mechanisms and Concepts. University Press,
Cambridge, pp. 27–47.
Norinder, U., Haeberlein, M., 2002. Computational approaches to the
prediction of the blood–brain distribution. Adv. Drug Deliv. Rev. 54,
291–313.
Pfeiffer, F., Betz, H., 1981. Solubilization of the glycine receptor from rat
spinal cord. Brain Res. 226, 273–279.
Stenberg, P., Luthman, K., Ellens, H., Lee, C.P., Smith, P.L., Lago, A.,
Elliott, J.D., Artursson, P., 1999. Prediction of the intestinal absorption
Usifoh, C.O., Lambert, D.M., Wouters, J., Scriba, G.K., 2001. Synthesis
and anticonvulsant activity of N,N-phthaloyl derivatives of central
nervous system inhibitory amino acids. Arch. Pharm. (Weinheim) 334,
323–331.
Van de Waterbeemd, H., Camenish, G., Folkers, G., Chretien, J.R.,
Raevsky, O.A., 1998. Estimation of blood–brain barrier crossing of
drugs using molecular size and shape, and H-bonding descriptors. J.
Drug Targeting 6, 151–165.
Wermuth, C.G., Gaignault, J.C., Marchandeau, C., 1996. Designing
prodrugs and bioprecursors I: carrier prodrugs. In: Wermuth, C. (Ed.),
The Practice of Medicinal Chemistry. Academic Press, New York, pp.
672–696.