Novel α-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) receptor antagonists of 2,3-benzodiazepine type: Chemical synthesis, in vitro characterization, and in vivo prevention of acute neurodegeneration

Article abstract of DOI:10.1021/jm0580003

Under pathophysiological conditions, α-amino-3-hydroxy-5-methyl-4- isoxazole propionate (AMPA) receptor activation is considered to play a key role in several disorders of the central nervous system. In the search for AMPA receptor antagonists, the synthe

Full text of DOI:10.1021/jm0580003

4618
J. Med. Chem. 2005, 48, 4618-4627
Novel r-Amino-3-hydroxy-5-methyl-4-isoxazole Propionate (AMPA) Receptor
Antagonists of 2,3-Benzodiazepine Type: Chemical Synthesis, in Vitro
Characterization, and in Vivo Prevention of Acute Neurodegeneration
Bernd Elger,*,† Andreas Huth,† Roland Neuhaus,† Eckard Ottow,† Herbert Schneider,† Bernd Seilheimer,† and
Lechoslaw Turski‡
Schering AG, 13342 Berlin, Germany, and Solvay Pharmaceuticals Research Laboratories, C.J. van Houtenlaan 36,
1381 CP Weesp, The Netherlands
Received January 7, 2005
Under pathophysiological conditions, R-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA)
receptor activation is considered to play a key role in several disorders of the central nervous
system. In the search for AMPA receptor antagonists, the synthesis and pharmacological
characterization of a series of novel compounds that are structurally related to GYKI 52466
(1), a well-known selective noncompetitive AMPA receptor antagonist, was performed. In vitro,
2,3-dimethyl-6-phenyl-12H-[1,3]dioxolo[4,5-h]imidazo[1,2-c][2,3]benzodiazepine (ZK 187638,
14a) antagonized the kainate-induced currents in cultured hippocampal neurons with an IC50
of 3.4 µM in a noncompetitive fashion. When tested in a clinically predictive rat model of acute
ischemic stroke, this noncompetitive AMPA receptor antagonist significantly reduced brain
infarction, indicating that it is neuroprotective after permanent focal cerebral ischemia.
Introduction
Chemistry
AMPA (R-amino-3-hydroxy-5-methyl-4-isoxazole pro-
pionate) receptors are a group of glutamatergic receptors
that are present in various cerebral cell types such as
neurons, oligodendrocytes, and microglia. AMPA recep-
tors are implicated as causative in the pathogenesis of
numerous acute and chronic neurological disorders.1-3
In the search for agents to treat these diseases competi-
tive AMPA receptor antagonists have been synthesized.
As an example, YM-872 ([2,3-dioxo-7-(1H-imidazol-1-yl)-
6-nitro-1,2,3,4-tetrahydro-1-quinoxalinyl]acetic acid mono-
hydrate, 2) is a competitive AMPA receptor antagonist
that has shown efficacy in rats after permanent MCA
occlusion and that is currently undergoing clinical
stroke trials.4 However, the risk that compounds that
inhibit AMPA receptors in a competitive way could fail
in clinical development for stroke therapy is potential-
ly high because the development of the compounds
2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline
(NBQX, 3) and [1,2,3,4-tetrahydro-7-morpholinyl-2,3-
dioxo-6-(trifluoromethyl)quinoxalin-1-yl]methylphospho-
nate (ZK 200775, 4) has been discontinued because of
unwanted side effects. Therefore, the search for neuro-
protective agents was extended to chemical structures
that inhibit AMPA receptors in a noncompetitive way.
Here, we describe the synthesis and pharmacological
characterization of 2,3-dimethyl-6-phenyl-12H-[1,3]di-
oxolo[4,5-h]imidazo[1,2-c][2,3]benzodiazepine (14a) and
its metabolite (18). Compound 14a emerged as the most
potent AMPA receptor inhibitor of a series of chemical
structures (Chart 1) that were synthesized as deriva-
tives of GYKI 52466 (1), the prototype of noncompetitive
AMPA receptor antagonists.5
The first step in the chemical synthesis of 2,3-
dimethyl-6-phenyl-12H-[1,3]dioxolo[4,5-h]imidazo[1,2-c]-
[2,3]benzodiazepine (14a) was the reduction of 5 with
LiAlH4 to alcohol 6, which was cyclized with benzalde-
hyde to isochromane 7a (Scheme 1). Oxidation of 7a
yielded benzoylphenylacetic acid 8a. Cyclization via the
mixed anhydride and subsequent reaction with hydra-
zine gave benzo-2,3-diazepine-4-one 9a, which was
converted to thione 7. Thione 10a could be cyclized by
reaction with aminoketal 12 (prepared by reductive
amination of 11), followed by treatment with acidic
EtOH to give 14a directly. Intermediate 13a could not
be isolated using the described reaction conditions. The
syntheses of 14b and 14c were accomplished in an
analogous manner (Table 1 and Scheme 1).
5-Phenyl-8-methyl-9-hydroxymethyl-11H-1,3-dioxolo-
[4,5-h]imidazo[1,2-c][2,3]benzodiazepine (18), which is
a metabolite of 14a, could be synthesized via the
monomethyl derivative 17 by reaction with aqueous
formaldehyde in the presence of sodium acetate and
glacial acetic acid as shown in Scheme 2. 17 was
synthesized from thione 10a in a manner analogous to
the preparation of 14a. In this case intermediate 16
could be isolated (Scheme 2).
The central intermediate for the synthesis of com-
pounds 24 and 27 was 5-(4-nitrophenyl)-11H-1,3-di-
oxolo[4,5-h]imidazo[1,2-c][2,3]benzodiazepine (21), which
could be obtained by reaction of thione 10d with
aminoacetal 19. Also in this case, the intermediate,
acetal 20, could be isolated. 24 was synthesized from
21 by reduction to 22, reductive diazotization to 23, and
subsequent chlorination of the imidazole ring by NCS
in acetonitrile. In the case of 27 it was necessary to first
brominate 21-25 to reduce the nitro group selectively
by Fe in acetic acid and finally to eliminate the amino
group by reductive diazotization, giving 27 (Scheme 3).
* To whom correspondence should be addressed. Phone: +49-30-
468-17253. Fax: +49-30-468-97253. E-mail: bernd.elger@schering.de.
† Schering AG.
‡ Solvay Pharmaceuticals Research Laboratories.
10.1021/jm0580003 CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/21/2005

Noncompetitive AMPA Receptor Antagonists
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4619
Chart 1. Noncompetitive and Competitive AMPA Receptor Antagonists of 2,3-Benzodiazepine-Type (1) and
Quinoxalinedione-Type (2-4), Respectively
Scheme 1
Pharmacology
addition, posttreatment of compound 14a was studied
in the rat model of proximal middle cerebral artery
(MCA) occlusion. This animal model has been shown to
have a predictive value for stroke in humans because
beneficial drug effects after permanent MCA occlusion
in rats were reflected in a positive outcome of a phase
III clinical trial.7,8 For appropriate dosing of 14a in the
in vivo studies, the pharmacokinetics properties of the
compound were evaluated. Finally, potential hypother-
The ability of the newly synthesized compounds to
inhibit AMPA receptors in a noncompetitive fashion was
determined in a functional in vitro test measuring
kainate-induced currents in primary cultured rat fetal
hippocampal neurons.
Selected compounds were screened in vivo for neuro-
protective efficacy using a pretreatment paradigm in the
mouse model of permanent focal cerebral ischemia.6 In

4620 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Elger et al.
Table 1. Inhibition of Kainate-Induced Whole-Cell Currents by
2,3-Benzodiazepine Derivatives in Cultured Rat Hippocampal
Neurons
without altering the potency of the agonist, thus behav-
ing as a noncompetitive antagonist. Competitive binding
of 14a and 18 to glutamate receptors, i.e., AMPA,
N-methyl-D-aspartate (NMDA), the kainate competitive
site, and the NMDA/glycine site, was inhibited by less
than 30% at 10 µM.
After iv injection of 14a, the plasma clearance and
half-life were calculated to be 54 mL‚min-1‚kg-1 and 1.2
h for the terminal elimination phase. The volume of
distribution was 2.7 L/kg. LC/MS/MS analysis revealed
a compound in the plasma samples of rats that was
prominent besides 14a. The chemical structure of 18
(Scheme 2) has been attributed to this metabolite of 14a.
AMPA receptor
substituents
mediated currents
compd
R1
R2
R3
IC50 (µmol)
Neuroprotective efficacy of 14a and 18 was studied
after focal cerebral ischemia in mice. Following perma-
nent MCA occlusion in mice, the ischemic territory was
strictly ipsilateral (left side) and localized exclusively
in the temporoparietal cortex. The animal model was
validated for pharmacological testing using the competi-
tive AMPA receptor antagonist NBQX (3) as a reference.
Total cerebral infarct volume was significantly (p <
0.05) reduced by 21% from 23.2 ( 4.8 mm3 (mean ( SD)
in vehicle-treated controls (n ) 9) to 18.4 ( 5.2 mm3 in
mice (n ) 12) treated with three ip injections (each 30
mg/kg NBQX) 60, 70, and 85 min after permanent MCA
occlusion.
14a
18
Me
CH2OH
H
Cl
Br
Me
Me
Me
Me
H
H
3.4
4.6
H
23
H
6.5
24
27
14b
14c
Cl
H
>10
>10
10
10
9.0
Br
Me
Me
H
Cl
F
GYKI 52466 (1)
NH2
mic activity of 14a was assessed in rats, which could
eventually account for indirect neuroprotective effects.
Results
14a was tested as pretreatment in MCA-occluded
mice, which consisted of four intraperitoneal injections
in total. 14a was given 1 h before and 0, 1, and 2 h after
MCA occlusion. Twenty-four hours after permanent
MCA occlusion, total cerebral lesion volumes were
significantly (p < 0.05) reduced by 34% from 26.0 ( 4.5
mm3 in vehicle-treated controls (n ) 8) to 17.1 ( 5.2
mm3 in mice (n ) 10) treated with four ip injections of
14a (each injection with a dose of 10 mg/kg). Compound
18 was studied in MCA-occluded mice using the same
treatment paradigm as for 14a; however, no significant
effects on brain lesion volumes were observed (Table 2).
Therapeutic compound effects on infarct volume were
investigated 24 h following permanent MCA occlusion
in male Fischer-344 rats. Both cerebral cortex and
striatum were affected by ischemia on the ipsilateral
side. Total infarct volume was significantly (p < 0.05)
reduced from 190 ( 54 mm3 in vehicle-treated controls
(n ) 15) to 138 ( 62, 142 ( 51, and 132 ( 59 mm3 in
rats given four ip injections of 14a in doses of 3 (n )
12), 10 (n ) 15), and 30 (n ) 10) mg/kg, respectively,
starting 1 h after MCA occlusion (Table 3). Lesion
reduction was not dose-dependent, suggesting that a
maximal drug effect (30.5% reduction) was reached
under the given experimental conditions. Investigation
of the NMDA receptor antagonist MK-801 as a reference
compound yielded effects that were similar in magni-
tude to those of 14a in this stroke model. Infarct volume
was significantly (p < 0.05) reduced by 29% from 166
( 43 mm3 in controls (physiological saline, n ) 10) to
118 ( 25 mm3 in rats (n ) 6) given a single ip injection
of MK-801 (3 mg/kg) immediately after MCA occlusion.
As mentioned above, derivatives of the selective
noncompetitive AMPA receptor antagonist GYKI 52466
(1) were synthesized (Chart 1 and Table 1). The basic
structural differences of all of the derivatives 14a-c,
23, 24, and 27 to GYKI 52466 are (1) the replacement
of the 4-methyl substitution by a nitrogen-containing
heterocycle attached to the 3,4-position of the 2,3-
benzodiazepine ring system and (2) the replacement of
the amino group on the phenyl ring system. The
inhibitory effects of these compounds on induced
AMPA receptor-mediated currents were investigated
in primary cultured rat fetal hippocampal neurons.
Long-lasting inward currents due to activation of AMPA
receptors were induced by the nondesensitizing AMPA
receptor ligand kainate. The novel structure (5-
phenyl-11H-1,3-dioxolo[4,5-h]imidazo[1,2-c][2,3]benzo-
diazepine, 23) with an IC50 of 6.5 µM was slightly
more potent than 1 even though it lacks the amino
group on the phenyl ring (Table 1). However, the activity
was lost when halogen atoms were used as substituents
on the nitrogen-containing heterocycle (compounds 24
and 27). On the other hand, the potency of 23 was
further enhanced when the nitrogen-containing hetero-
cycle was methylated as in 14a. The IC50 of 14a for
inhibition of kainate-induced currents was 3.4 µM
(Figure 1). The IC50 for the reference compounds GYKI
52466 (noncompetitive) and NBQX (competitive) in
antagonizing kainate-induced currents were 9 and 0.03
µM, respectively, in this experimental setting. When a
halogen atom was introduced as a substituent on the
phenyl ring of 14a, the potency for AMPA receptor
inhibition was reduced (compounds 14b and 14c).
To investigate the mechanism of current inhibition,
the effect of compound 14a on kainate concentration-
response curves was studied in a separate set of experi-
ments (Figure 2). The antagonist caused a concentration-
dependent reduction of kainate-evoked currents at all
kainate concentrations (i.e., a reduction in efficacy)
Core body temperature in conscious rats decreased
significantly by 2 °C after injection of the reference
compound 8-OH-DPAT. However, no difference in body
temperature was found between vehicle-treated controls
and rats injected a high dosage (100 mg/kg ip) of 14a
(data not shown), indicating that cerebroprotection after

Noncompetitive AMPA Receptor Antagonists
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4621
Scheme 2
permanent focal cerebral ischemia is not mediated by
unspecific hypothermia.
NBQX is about 100-fold more potent in antagonizing
AMPA receptor mediated currents in cultured hippoc-
ampal neurons. These findings support the notion that
noncompetitive AMPA receptor antagonists may have
an advantage under conditions in which high glutamate
levels may render the competitive antagonists relatively
ineffective.5
Discussion
Derivatives of the selective prototype noncompetitive
AMPA receptor antagonist GYKI 52466 (1) were syn-
thesized, and the new 2,3-dimethyl-6-phenyl-12 H-[1,3]-
dioxolo[4,5-h]imidazo[1,2-c][2,3]benzodiazepine (14a) was
found to be about 3 times more potent than the parent
compound in antagonizing AMPA receptor mediated
currents in hippocampal neurons. Inhibition studies in
hippocampal neurons showed that 14a acts in a non-
competitive manner at the AMPA receptor. Moreover,
our in vitro examinations revealed that the effects of
14a were non-NMDA antagonistic and that there was
no relevant competitive inhibition of the AMPA recogni-
tion site, as has also been demonstrated for GYKI-
52466.9,10
When tested in vivo, 14a displayed about 30% reduc-
tion of total cerebral lesion volume in rats after perma-
nent MCA occlusion. The magnitude of lesion reduction
was similar to that of the NMDA receptor antagonist
MK-801, which is an established gold standard refer-
ence compound in this animal stroke model.11 The
cerebroprotection elicited by 14a did not result from
induction of unspecific systemic hypothermia because
administration of even higher dosages of 14a did not
reduce the body temperature of rats.
Similarities of neuroprotective drug effects after
permanent MCA occlusion in rats and mice have been
repeatedly observed.6 Likewise, compound 14a produced
comparable infarct reductions after permanent MCA
occlusion in rats and mice. In comparison to the well-
known competitive AMPA receptor antagonist NBQX,
which was tested in MCA-occluded mice at an effective
total dose of 90 mg/kg as reported in the literature,12,13
our noncompetitive inhibitor 14a achieved significant
infarct reductions at a total dose of 40 mg/kg, although
The main difference in the chemical structures of 14a
and 23 compared to those of the other well-known
noncompetitive AMPA receptor antagonists GYKI 52466
and GYKI 53773 (LY300164, talampanel)9 is the re-
placement of the 4-methyl substitution by a nitrogen-
containing heterocycle attached to the 3,4-position of the
2,3-benzodiazepine ring system. The electrophysiological
results of the present study indicate that the potency
for AMPA receptor inhibition is maintained despite this
modification, which is in line with structure-activity
relationship investigations by others.9 In contrast to the
noncompetitive AMPA receptor antagonist GYKI 52466,
14a and 23 do not possess an amino group on the phenyl
moiety. Nevertheless, both compounds displayed higher
in vitro activity than GYKI 52466. This result confirms
the general conclusion from previous SAR studies with
2,3-benzodiazepine derivatives related to 1 that the
amino-substituent on the phenyl ring is not essential
for AMPA receptor antagonistic efficacy.9 In recently
published SAR studies on annelated 2,3-benzodiazepine
derivatives, it has been suggested that electron-with-
drawing substituents on the phenyl ring might be useful
for optimizing the pharmacological properties of AMPA
receptor antagonists.14 However, electrophysiological
tests of 14b and 14c in the present investigation
revealed that inhibition of AMPA receptor mediated
currents in cultured neurons is reduced in azolo-
condensed 2,3-benzodiazepines carrying a halogen atom
on the phenyl ring compared to 14a. Halogen atoms as
substituents on the nitrogen-containing heterocycle also
led to significantly decreased activities of compounds

4622 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Elger et al.
Scheme 3
are expressed in δ (ppm), and coupling constants (J) are in
hertz. The purity of the compounds was checked by HPLC on
a YMC Pro 18RS 5 µm column (150 mm × 4.6 mm), with a
gradient of 95% to 5% water (containing 0.1% ammonia) in
acetonitrile (1 mL/min) and detection at 254 nm. Reanalyses
of compounds by a different HPLC method was performed as
described in detail in the Supporting Information. The high-
resolution mass spectra were measured on a AutoSpec Q, V6-
Micromass at 70 eV.
24 and 27. Furthermore, our findings with 18 in MCA-
occluded mice show that the in vivo efficacy of 14a is
lost when the methyl group of the nitrogen-containing
heterocycle is hydroxylated.
Conclusions
From a series of azolo-condensed 2,3-benzodiazepines,
14a emerged as a potent noncompetitive AMPA receptor
antagonist that reduces ischemic brain lesions after
permanent middle cerebral artery occlusion in both mice
and rats. The compound is efficacious when treatment
is started before or 1 h after MCA occlusion. Since the
animal models were validated by use of reference
compounds that are considered as “gold standard”, these
studies indicate that the noncompetitive AMPA receptor
antagonist 14a has potential for therapy of acute stroke.
Compounds 5 and 11 are commercially available.
3,4-Methylenedioxyphenethyl Alcohol (6). To a stirred
ice cold suspension of lithium aluminum hydride (58 g, 1500
mmol) in Et2O (2.5 L, 0-10 °C internal temperature) under
argon was added a solution of 3,4-methylenedioxyphenylacetic
acid (191 g, 0.347 mol) in dry THF (579 mL) dropwise within
45 min. After being stirred at ambient temperature for 8 h,
the mixture was cooled to 0 °C and dilute aqeuous sodium
hydroxide (4 M, 58 mL) was added in such a manner to
maintain the temperature at 5-10 °C. Afterward, an amount
of 250 mL of water was added and the mixture was stirred
overnight at room temperature. The precipitate was removed
Experimental Section
1H NMR spectra were measured with a Bruker Aspekt 3000
(300 MHz) unless stated otherwise in the text. Chemical shifts

Noncompetitive AMPA Receptor Antagonists
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4623
Table 3. Effect of MK-801 and Compound 14a on Total Infarct
Volume Measured 24 h after Permanent MCA Occlusion in
Ratsa
treatment
dose (mg/kg)
total infarct volumeb (mm3)
vehicle
MK-801
vehicle
14a
166 ( 43 (10)
118 ( 25 (6)c
190 ( 54 (15)
138 ( 62 (12)c
142 ( 51 (15)c
132 ( 59 (10)c
3
4 × 3
14a
4 × 10
4 × 30
14a
a MK-801 was injected ip immediately after MCA occlusion, and
compound 14a was given ip 1, 2, 3, and 4 h after MCA occlusion.
b Data are the mean ( SD with numbers of animals in parenthe-
ses. c p < 0.05 vs corresponding vehicle-treated group.
5-Phenyl-1,3-dioxolo[4,5-h]-isochromane (7a). To a
stirred solution of 3,4-methylenedioxyphenethyl alcohol (191
g, 0.347 mol) in toluene (300 mL) at room temperature were
added sequentially benzaldehyde (14 g, 13.5 mL, 132 mmol)
and concentrated HCl (11.3 mL). Stirring was continued for 2
h before quenching with water (130 mL). The mixture was
extracted with EtOAc (3 × 150 mL), and the combined organic
layers were washed with water, dried, and concentrated in
vacuo. Trituration with hexane (400 mL) followed by suction
filtration gave 5-phenyl-1,3-dioxolo[4,5-h]isochromane (22.4 g,
75%) as light-ochre crystals (mp 102 °C). 1H NMR (DMSO), δ:
2.7 (m, 1H), 2.9 (m, 1H), 3.8 (m, 1H), 5.6 (s, 1H), 5.9 (s, 1H),
5.95 (s, 1H), 6.27 (s, 1H), 6.8 (s, 1H), 7.3 (m, 5H).
Figure 1. Concentration-dependent inhibition of kainate-
induced whole-cell currents in hippocampal neurons. Data
points represent the mean ( SD of five to six independent
measurements (cells) normalized to the current values re-
corded in the absence of antagonist. Compound 14a (ZK
187638) and kainate were applied by continuous-bath perfu-
sion. Antagonist effects were measured after 90 s of preincu-
bation. 14a was dissolved in pure DMSO and diluted in
recording solution to a final DMSO concentration of 0.3%.
DMSO was present in all solutions at the same concentration.
4,5-Methylenedioxy-2-benzoylphenylacetic Acid (8a).
Jones reagent was prepared by dissolving chromium(VI) oxide
(26.7 g) in concentrated sulfuric acid (23 mL) and adding water
to make up to a final volume of 100 mL. To a stirred solution
of 5-phenyl-1,3-dioxolo[4,5-h]-isochromane (7a, 18.6 g, 73.14
mmol) in acetone (1.17 L, pa) was added Jones reagent (110
mL, 2.66 M, 292.6 mmol) dropwise in such a manner that the
internal temperature did not exceed 15 °C. The reaction
mixture was warmed to room temperature, and stirring
continued for 30 min before recooling to 0 °C and quenching
with 2-propanol (22 mL). Stirring was continued for 5 min,
and the chromium salts were removed by suction filtration and
washed with acetone. The filtrate was evaporated to dryness,
treated with 100 mL of water, and filtered off to obtain 4,5-
methylenedioxy-2-benzoylphenylacetic acid (8a, 16.98 g, 81%)
as ochre crystals (mp 181.2 °C; lit.17 182-184 °C). 1H NMR
(DMSO), δ: 3.7 (s, 2H), 6.15 (s, 2H), 6.9 (s, 1H), 7.1 (t, J ) 7.5
Hz, 2H), 7.7 (m, 3H).
Figure 2. Effect of 14a (ZK 187638) on kainate concentra-
tion-response curves. Data points represent the mean ( SD
of 4-10 independent measurements (cells) normalized to the
inward current obtained with 10 mM kainate in the absence
of antagonist.
5-Phenyl-7H-1,3-dioxolo[4,5-h][2,3]benzodiazepine-
8(9H)-one (9a). To a stirred solution of 4,5-methylenedioxy-
2-benzoylphenylacetic acid (8a, 12 g, 42 mmol) in dry THF (600
mL) under nitrogen at -20 °C (internal temperature) was
added sequentially triethylamine (4.3 g, 5.9 mL, 42 mmol) and
isobutylchloroformate (5.8 g, 5.5 mL, 42 mmol) in such a
manner that the internal temperature did not exceed -15 °C.
After being stirred for 30 min at -20 °C, the reaction mixture
was cooled to -30 °C, treated dropwise with hydrazine (100%,
7 mL) (instead of 100% hydrazine; hydrazine hydrate (98%,
10 mL, 210 mmol) can be used), warmed to room temperature,
and then stirred for a further 1.5 h. The reaction mixture was
filtered. The filtrate was concentrated to ca. 200 mL, concen-
trated hydrochloric acid (18 mL) was added, and the mixture
was stirred for 1 h. Water (200 mL) was then added, tetrahy-
drofuran was removed in vacuo, and the residue was extracted
three times with EtOAc. The combined organic layers were
washed with water, dried, and evaporated to dryness to obtain
5-phenyl-7H-1,3-dioxolo[4,5-h][2,3]benzodiazepine-8(9H)-one
(9a, 4.56 g, 38%) as ochre crystals (mp 182.9 °C; lit. 182-184,18
Table 2. Effects of NBQX (3) and Compound 14a and Its
Metabolite 18 on Brain Lesion Measured 24 h after Permanent
MCA Occlusion in Micea
treatment
dose (mg/kg)
total infarct volumeb (mm3)
vehicle
NBQX
vehicle
14a
vehicle
18
18
23.2 ( 4.8 (9)
18.4 ( 5.2 (12)c
26.0 ( 4.5 (8)
17.1 ( 5.2 (10)c
23.7 ( 7.2 (11)
20.6 ( 6.7 (9)
23.0 ( 8.9 (8)
3 × 30
4 × 10
4 × 10
4 × 30
a NBQX was injected ip 60, 70, and 80 min after MCA occlusion.
Compounds 14a and 18 were given ip 1 h before and again 0, 1,
and 2 h after MCA occlusion. b Data are the mean ( SD with
numbers of animals in parentheses. c p < 0.05 vs corresponding
vehicle-treated group.
1
by suction filtration and washed twice with methyl tert-butyl
ether, and the combined filtrates were extracted three times
with methyl tert-butyl ether. The combined organic layers were
washed with brine, dried, and concentrated in vacuo to obtain
3,4-methylenedioxyphenethyl alcohol15,16 (171 g, 95%) as an
184-186 °C 19). H NMR (DMSO), δ: 3.4 (s, 2H), 6.1 (s, 2H),
6.6 (s, 1H), 7.1 (s, 1H), 7.5 (m, 5H), 10.9 (broad s, 1H).
5-Phenyl-7H-1,3-dioxolo[4,5-h][2,3]benzodiazepine-
8(9H)-thione (10a). A stirred solution of 5-phenyl-7H-1,3-
dioxolo[4,5-h][2,3]benzodiazepin-8(9H)-one (9a, 20.5 g, 73.14
mmol) in dry pyridine (360 mL) under argon was heated to
110 °C (bath temperature) before treating with phosphorus
pentasulfide (12.2 g, 54.9 mmol). Stirring was continued at
1
oil. H NMR (DMSO), δ: 2.65 (t, J ) 7.5 Hz; 2H), 3.55 (dd, J
) 7.5, 5 Hz, 2H), 4.6 (t, J ) 5 Hz, 1H), 5.95 (s, 2H), 6.7 (d, J
) 7.5 Hz; 1H), 6.8 (m, 2H).

4624 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Elger et al.
this temperature for a further 1.5 h before cooling to room
temperature. The mixture was concentrated in vacuo to ca.
30 mL. To this well-stirred residue, water (800 mL) was added.
The resulting precipitate was isolated by suction filtration and
washed successively with water (150 mL), ethanol (150 mL),
and hexane (150 mL) to obtain after drying 5-phenyl-7H-1,3-
dioxolo[4,5-h][2,3]benzodiazepine-8(9H)-thione (10a, 21.1 g,
97.3%) as a reddish brown powder (mp 203.6 °C; lit. 201-203
°C20). [If yields exceed 100%, then the product contains
amounts of sulfur or phosphosulfur impurities. These impuri-
ties disturb the next reaction by the necessity of much higher
amounts of aminoketal (12).] 1H NMR (DMSO), δ: 3.8 (broad
d, 2H), 6.2 (s, 2H), 6.6 (s, 1H), 7.1 (s, 1H), 7.5 (m, 5H), 12.7 (s,
1H).
2-Amino-3,3-dimethoxybutane (12).21 To a stirred solu-
tion of 3,3-dimethoxybutan-2-one (11, 50 g, 378 mmol) in
methanol (1250 mL) under nitrogen were added sequentially
molecular sieves (3Å, 60 mL), ammonium acetate (290 g, 3.75
mol), and sodium cyanoborohydride (25 g, 0.41 mol). Stirring
was continued overnight. The mixture was filtered through
Celite, and the filtrate was concentrated in vacuo (bath
temperature of 30 °C) to ca. 300 mL. The residue was diluted
with dichloromethane (1.5 L) and washed successively with
dilute aqueous sodium hydroxyde solution (2 M, 2 × 300 mL),
water (2 × 200 mL), and brine. The organic layer was dried,
filtered, and evaporated to dryness to obtain 2-amino-3,3-
dimethoxybutane (12, 45.9 g, 91%).
5-Phenyl-8,9-dimethyl-11H-1,3-dioxolo[4,5-h]imidazo-
[1,2-c][2,3]benzodiazepine (14a). 5-Phenyl-7H-1,3-dioxolo-
[4,5-h][2,3]benzodiazepine-8(9H)-thione (10a, 6 g, 20.5 mmol)
[in the scale-up of the reaction, the yield dropped] was taken
up in ethylene glycol monomethyl ether (10 mL) and concen-
trated in vacuo (bath temperature of 50 °C, 20-30 mbar)
almost to dryness (ca. 3-5 mL). 2-Amino-3,3-dimethoxybutane
(12, 4 mL) was then added. The reaction mixture was heated
at 110 °C while a very slow stream of nitrogen was passed
continuously through it. The reaction mixture was cooled, and
further cycles of 2-amino-3,3-dimethoxybutane (12) addition
and heating were carried out according to a specified time
schedule. Finally, the reaction mixture was evaporated to
dryness and the residue was purified by chromatography on
silica gel (eluant: 800 mL of 50% EtOAc in hexane, then 50%
acetone in hexane). Further purification by trituration with
25% EtOAc in hexane gave pure 5-phenyl-8,9-dimethyl-11H-
1,3-dioxolo[4,5-h]imidazo[1,2-c][2,3]benzodiazepine (14a, 2.3 g,
35%) as cream-white crystals (mp 176.6 °C). 1H NMR (DMSO),
δ: 2.0 (s, 3H), 2.2 (s, 3H), 3.8 (s, 2H), 6.1 (s, 2H), 6.6 (s, 1H),
7.2 (s, 1H), 7.5 (m, 3H), 7.7 (m, 2H). Anal. Calcd: C) 72.49%;
H) 5.17%; N) 12.68%; O)9.66%. Obsd: C) 72.30%; H)
5.43%; O) 9.82%; N) 12.51%. HPLC: tR ) 9.03 min; purity,
99.78%. HR MS: Molpeak, C20H17N3O2; calcd, 331.132 077;
found 331.131 844. A second crop of slightly impure 14a (1.2
g) was obtained, which could be purified by chromatography.
Hz, 1H), 3.25 (d, J ) 12 Hz, 1H), 3.9 (m, 4H), 6.05 (s, 1H), 6.1
(s, 1H), 6.5 (t, J ) 6 Hz, 1H), 6.6 (s, 1H), 6.85 (s, 1H), 7.4 (m,
3H), 7.55 (m, 2H).
5-Phenyl-8-methyl-11H-1,3-dioxolo[4,5-h]imidazo[1,2-
c][2,3]benzodiazepine (17). A stirred solution of (2-methyl-
[1,3]dioxolan-2-ylmethyl)-(5-phenyl-7,9-dihydro-1,3-dioxa-6,7-
diazacyclohepta[f]inden-8-ylidene)amine (16, 402 mg, 1.06
mmol) in ethanol (8 mL) was treated with concentrated
hydrochloric acid (8 mL) and heated to 100 °C for 2.5 h. After
dilution with 20 mL of water the ethanol was distilled off. The
aqueous layer was brought to pH 11 by addition of 5 M NaOH
and extracted three times with EtOAc. The organic layer was
washed, dried, filtered, and concentrated to dryness. The
residue was triturated with a mixture of ethyl acetate and
hexane to obtain 5-phenyl-8-methyl-11H-1,3-dioxolo[4,5-h]-
imidazo[1,2-c][2,3]benzodiazepine (17, 258 mg, 77% in 2
1
batches (mp 214.3 °C). H NMR (DMSO), δ: 2.3 (s, 3H), 3.9
(s, 2H), 6.1 (s, 2H), 6.6 (s, 1H), 6.65 (s, 1H), 7.2 (s, 1H), 7.5 (m,
3H), 7.7 (m, 2H).
5-Phenyl-8-methyl-9-hydroxymethyl-11H-1,3-dioxolo[4,5-h]-
imidazo[1,2-c][2,3]benzodiazepine (18). 5-Phenyl-8-methyl-
11H-1,3-dioxolo[4,5-h]imidazo[1,2-c][2,3]benzodiazepine (17,
2.7 g, 8.54 mmol) was treated with acetic acid (3.6 mL), sodium
acetate (3.4 g, 41.9 mmol, dry), and formalin (28.6 mL, 37%
in water, 94.1 mmol). This mixture was heated for 5 h to 110
°C. After diluting with water to 100 mL, it was made alkaline
by addition of potassium carbonate and extracted three times
with ethyl acetate. The organic layer was washed, dried,
filtered, and evaporated to dryness. The residue was purified
by chromatography (silica gel, eluent CH2Cl2/EtOH ) 10:1).
Further purification by trituration with EtOAc gave 5-phenyl-
8-methyl-9-hydroxymethyl-11H-1,3-dioxolo[4,5-h]imidazo[1,2-
c][2,3]benzodiazepine (18, 695 mg, 23%, mp 229.4 °C). 1H NMR
(DMSO), δ: 2.3 (s, 3H), 3.9 (s, 2H), 4.25 (d, J ) 5 Hz, 2H), 4.7
(t, J ) 5 Hz, 1H), 6.1 (s, 2H), 6.55 (s, 1H), 6.65 (s, 1H), 7.2 (s,
1H), 7.5 (m, 3H), 7.7 (m, 2H). HPLC: tR ) 6.80 min; purity,
98.75%; HR MS: Molpeak, C20H17N3O3; calcd, 347.126 992;
found, 347.125 081.
2,2-Dimethoxyethyl-5-phenyl-7,9-dihydro-1,3-dioxa-
6,7-diazacyclohepta[f]inden-8-ylidene)amine (20). A mix-
ture of 5-(4-nitrophenyl-7H-1,3-dioxolo[4,5-h][2,3]benzodiazepine-
8(9H)-thione (3 g, 8.8 mmol) and 2,2-dimethoxyethylamine in
2-methoxyethanol (82 mL) was heated at 110 °C for 3 h, while
a slow stream of argon was passed through it. When the
mixture was cooled and evaporated to dryness, the residue was
triturated by a mixture of EtOAc and hexane to give 2,2-
dimethoxyethyl-5-phenyl-7,9-dihydro-1,3dioxa-6,7diazacyclo-
hepta[f]inden-8-ylidene)amine (20, 2.6 g, 72%).
5-(4-Nitrophenyl)-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]benzo-
diazepine (21). A mixture of 2,2-dimethoxyethyl-5-phenyl-
7,9-dihydro-1,3dioxa-6,7-diazacyclohepta[f]inden-8-ylidene)am-
ine (20,1.6 g, 3.8 mmol), ethanol (50 mL), and concentrated
hydrochloric acid (50 mL) was refluxed for 3 h. On cooling,
the mixture was concentrated in vacuo and the residue was
treated at 0 °C with 3 M NaOH and extracted three times with
EtOAc. The organic layer was washed with brine and evapo-
rated to dryness. The residue was purified by chromatog-
raphy on silica gel (eluant CH2Cl2/ethanol ) 95:5) to obtain
5-(4-nitrophenyl)-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]benzo-
Compounds 14b and 14c were synthesized in the same
manner as 14a (Scheme 1).
1
14b. H NMR (DMSO), δ: 2.0 (s, 3H), 2.2 (s, 3H), 3.85 (s,
2H), 6.1 (s, 2H), 6.6 (s, 1H), 7.2 (s, 1H), 7.6 (d, J ) 7.5 Hz,
2H), 7.75 (d, J ) 7.5 Hz, 2H). HPLC: tR ) 10.18 min; purity,
99.44%. HR MS: Molpeak, C20H16N3O237Cl; calcd, 367.090 155;
found, 367.089 925.
1
diazepine (21, 1.04 g, 80%). H NMR (DMSO), δ: 4.1 (s, 2H),
6.1 (s, 2H), 6.65 (s, 1H), 6.9 (d, J ) 1.5 Hz, 1H), 7.2 (s, 1H),
7.5 (d, J ) 1.5 Hz, s, 1H), 7.9 (d, J ) 10 Hz, 2H), 8.4 (d, J )
10 Hz, 2H).
14c. HPLC: tR ) 9.05 min; purity, 95.75%. HR MS:
Molpeak, C20H16N3 O2F; calcd, 349.122 655; found, 349.121 815.
(2-Methyl[1,3]dioxolan-2-ylmethyl)-(5-phenyl-7,9-dihy-
dro-1,3-dioxa-6,7-diazacyclohepta[f]inden-8-ylidene)am-
ine (16). A stirred solution of 5-phenyl-7H-1,3-dioxolo[4,5-
h][2,3]benzodiazepine-8(9H)-thione (10a, 400 mg, 1.35 mmol)
in ethanol (1 mL) was treated with C-(2-methyl[1,3]dioxolan-
2-yl)methylamine22 (1 mL) while a very slow stream of argon
was passed continuously through it. The mixture was then
refluxed for 3.5 h. After addition of diisopropyl ether the
precipitate was filtered off to obtain (2-methyl[1,3]dioxolan-
2-ylmethyl)-(5-phenyl-7,9-dihydro-1,3-dioxa-6,7-diaza-cyclohepta-
[f]inden-8-ylidene)amine (16, 402 mg, 78.5%) as a dark powder
(mp >300 °C). 1H NMR (DMSO), δ: 1.1, s, 3H), 2.8 (d, J ) 12
5-(4-Aminophenyl)-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]-
benzodiazepine (22). 5-(4-Nitrophenyl)-11H-1,3-dioxolo[4,5-
h][1,2-c][2,3]benzodiazepine (21, 1.28 g, 3.7 mmol) was treated
under argon with HOAc (50 mL) and iron powder (3.6 g, 64.8
mmol). This mixture was heated for 15 min in an oil bath,
which was preheated at 90 °C. After filtration of the hot
mixture through Celite and subsequent washing with HOAc,
the filtrate was concentrated to dryness. The residue was
treated with 2 M NaOH and extracted three times with EtOAc.
The organic layer was washed with brine and evaporated to
dryness. The residue was purified by chromatography on silica

Noncompetitive AMPA Receptor Antagonists
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4625
gel (eluant CH2Cl2/ethanol ) 95:5) to obtain 5-(4-aminophenyl)-
11H-1,3-dioxolo[4,5-h][1,2-c][2,3]benzodiazepine (22, 787 mg,
67%).
Membrane currents were recorded in the whole-cell config-
uration of the patch clamp technique from cells cultured for
6-14 days. The external recording medium contained the
following (mM): NaCl 150; KCl 5; CaCl2 1.8; MgCl2 2; HEPES
10; pH adjusted to 7.4 with NaOH. The patch-pipet solution
contained the following (mM): CsCl 150; MgCl2 1; EGTA 10;
HEPES 10; Mg-ATP 2; pH adjusted to 7.4 with NaOH. For
recording sodium currents, the external medium and pipet
solution contained 10 mM TEA-Cl, and 0.1 mM CdCl2 was
added to the external medium. Compounds were dissolved in
pure DMSO and diluted in the recording medium to a final
concentration of 0.3% DMSO. The nondesensitizing AMPA
receptor agonist kainate was dissolved in recording medium.
Kainate and 14a were applied by bath perfusion. If not
otherwise indicated in the text, the holding potential was set
to -60 mV. Inhibition curves were fitted to the data according
to the logistic equation
5-Phenyl-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]benzodia-
zepine (23). 5-(4-Aminophenyl)-11H-1,3-dioxolo[4,5-h][1,2-c]-
[2,3]benzodiazepine (22, 820 mg, 2.6 mmol) in THF (35 mL)
was treated under argon with 5.7 mL (43.4 mmol) n-pentyl
nitrite. This mixture was refluxed for 2.5 h. On cooling, the
mixture was concentrated in vacuo and the residue was
purified by chromatography on silica gel (eluant CH2Cl2/
ethanol ) 95:5) to obtain 5-phenyl-11H-1,3-dioxolo[4,5-h][1,2-
c][2,3]benzodiazepine (23, 500 mg, 57%). 1H NMR (Bruker
Avance 300 (300 MHz), DMSO), δ: 3.9 (s, 2H), 5.95 (s, 2H),
6.5 (s, 1H), 6.9 (d, J ) 2.5 Hz, 1H), 7.2 (s, 1H), 7.45 (d, J ) 2.5
Hz, s, 1H), 7.5 (m, 3H), 7.7 (dd, J ) 2.5 Hz, 7.5 Hz, 2H);
HPLC: tR ) 7.88 min; purity, 99.26%. HR MS: Molpeak,
C18H13N3O2; calcd, 303.100 777; found, 303.099 342.
5-Phenyl-8,9-dichloro-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]-
benzodiazepine (24). A mixture of 5-phenyl-11H-1,3-dioxolo-
[4,5-h][1,2-c][2,3]benzodiazepine (23, 100 mg, 0.3 mmol) and
N-chlorosuccinimide (85 mg, 0.64 mmol) in acetonitrile (4 mL)
was refluxed under argon for 2 h. On cooling, the mixture was
diluted with water and extracted with EtOAc. The organic
layer was washed with brine and evaporated to dryness. The
residue was purified by chromatography on silica gel (eluant
CH2Cl2/ethanol ) 95:5) to obtain 5-phenyl-8,9-dichloro-11H-
1,3-dioxolo[4,5-h][1,2-c][2,3]benzodiazepine (24, 63 mg, 53%).
1H NMR (Bruker Avance 300 (300 MHz), DMSO), δ: 4.0 (s,
2H), 5.9 (s, 2H), 6.6 (s, 1H), 7.2 (s, 1H), 7.55 (m, 3H), 7.5 (d, J
100
I )
n
IC50
1 +
(
)
[14a]
Kainate concentration-response curves were fitted according
to the logistic equation
Imax
I )
n
EC50
1 +
(
)
[kainate]
) 1.5 Hz, s, 1H), 7.7 (dd, J ) 2.5 Hz, J ) 7.5 Hz, 2H); HPLC:
37
tR ) 11.56 min; purity, 99.44%. HR MS: Molpeak, C18H11N3O2
Cl2; calcd, 375.016 932; found, 375.018 669.
-
except for 100 µM 14a, where the EC50 was forced to 0.15 mM
(I is the normalized inward current (%), Imax is the maximal
inward current obtained at 10 mM kainate, and n is the slope).
Ionotropic Glutamate Receptor Binding Studies. Com-
petitive binding studies were performed at 10 µM in rat
forebrain membranes using the specific ligands [3H]AMPA,
[3H]CGP 39653, [3H]kainic acid, [3H]MDL-105,519.23-26
In Vivo Experiments. Male rats and male mice were
maintained under standard conditions (12 h day/night cycle,
22 °C, 50% humidity) with free access to food and tap water.
Before use for the experiments, the animals were allowed to
recover from transportation for about 1 week. All experiments
were performed in accordance to the requirements of the
German laws for the protection of animals (Deutsches Tiers-
chutzgesetz) and in compliance with the European Community
Directive 86/609.
Pharmacokinetics of 14a. Rat plasma was taken at nine
time points (5 min to 7 h) after iv injection of 14a (2.5 mg/kg).
Blood samples were collected in heparinized tubes and cen-
trifuged to separate the plasma, which was stored at -20 °C
until analysis. Plasma was precipitated with acetonitrile (1:
5), and the supernatant was taken directly for LC/MS/MS
quantification, using calibration curves in the matrix and a
related compound as internal standard. The separation was
performed on a XTerra MS C18 column (10 cm × 2.1 mm i.d.,
3.5 µm particle size, precolumn 1 cm × 2.1 mm) at 25 °C. The
mobile phase was A (H2O + 0.1% acetic acid) and B (acetoni-
trile + 0.1 acetic acid): flow rate 0.3 mL/min, gradient 75% A
to 5% A in 5 min. The retention time for 14a was 3.2 min,
and that for the internal standard was 3.1 min. Mass spec-
trometric detection was by TIS Pos MS.
5-(4-Nitrophenyl)-8,9-dibromo-11H-1,3-dioxolo[4,5-h]-
[1,2-c][2,3]benzodiazepine (25). A mixture of 5-(4-nitrophe-
nyl)-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]benzodiazepine (21, 420
mg, 1.2 mmol) and N-bromosuccinimide (475 mg, 2.64 mmol)
in dry DMF (10 mL) was stirred for 2 h at ambient temper-
ature. The mixture was partitioned between EtOAc and water.
The organic layer was washed with brine and evaporated to
dryness. The residue was purified by chromatography on silica
gel (eluant EtOAc/hexane ) 1:1) to obtain 5-(4-nitrophenyl)-
8,9-dibromo-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]benzodiazepine (25,
1
432 mg, 82%). H NMR (DMSO), δ: 4.1 (s, 2H), 6.15 (s, 2H),
6.7 (s, 1H), 7.3 (s, 1H), 8.0 (d, J ) 10 Hz, 1H), 8.4 (d, J ) 10
Hz, 2H).
5-(4-Aminophenyl)-8,9-dibromo-11H-1,3-dioxolo[4,5-h]-
[1,2-c][2,3]benzodiazepine (26). In a manner analogous to
that for 5-(4-nitrophenyl)-8,9-dibromo-11H-1,3-dioxolo[4,5-h]-
[1,2-c][2,3]benzodiazepine (25, 1.16 g, 2.3 mmol) was obtained
5-(4-nitrophenyl)-8,9-dibromo-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]-
benzodiazepine (26, 930 mg, 85%).
5-Phenyl-8,9-dibromo-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]-
benzodiazepine (27). In a manner analogous to that for 5-(4-
aminophenyl)-8,9-dibromo-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]-
benzodizepine (26, 490 mg, 1 mmol) was obtained 5-phenyl-
8,9-dibromo-11H-1,3-dioxolo[4,5-h][1,2-c][2,3]benzodiazepine (27,
1
316 mg, 75%). H NMR (DMSO), δ: 4.05 (s, 2H), 6.1 (s, 2H),
6.6 (s, 1H), 7.2 (s, 1H), 7.6 (m, 3H), 7.75 (dd, J ) 7.5 Hz, J )
1.5 Hz, 2H), 8.4 (d, J ) 10 Hz, 2H). HR MS: Molpeak,
C
18H11N3O281Br2; calcd, 462.917 707; found, 462.918 205.
HPLC: tR ) 10.98 min; purity, 93.5%.
In Vitro Studies: Functional Inhibition of AMPA
Receptors. Electrophysiological studies were carried out on
primary cultured hippocampal neurons. Hippocampi were
isolated from fetal Wistar rats (E19), dissected into small
pieces (about 1 mm2), and incubated in Mg/Ca-free Hank’s
balanced salt solution containing 0.25% trypsin at 37 °C for
15 min. Cells were dissociated by trituration and transferred
into medium (Neurobasal/B27, GIBCO) with 5% fetal calf
serum, then plated on poly-L-lysine-coated dishes (35 mm i.d.,
7.5 × 105 cells/dish). Dishes were stored in a humidified
incubator at 37 °C, 5% CO2. Medium was changed on days 3
and 6. On day 6 the medium was supplemented with 10 µM
cytosine-arabinoside.
Permanent MCA Occlusion in Rats. Male Fischer-344
rats weighing 250-300 g were anesthetized with 4% halothane
for induction and 1.5% halothane for maintenance in 30% O2/
70% N2O via a face mask. The left proximal middle cerebral
artery was occluded by a slight modification of the technique
of Tamura et al.27 Briefly, a vertical 2 cm skin incision was
made between the left eye and the left ear. Under an operating
microscope (Wild AG, Heerbrugg, Switzerland) with low-power
magnification, the temporalis muscle was divided and re-
tracted in order to expose the zygoma and the squamosal bone.
Parts of the temporalis muscle were removed, but the zygo-
matic bone was left intact. The left middle cerebral artery was
exposed through a burr hole craniectomy (2 mm diameter)

4626 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Elger et al.
drilled close to the foramen ovale. After the dura was opened,
the proximal portion of the middle cerebral artery was elec-
trocauterized with bipolar forceps (Erbotom T130, Erbe, Ele-
ktromedizin GmbH, Tu¨bingen, Germany) and cut through with
scissors. The burr hole was covered with Lyostypt (Braun
Dexon), and the skin was sutured. During surgery, the rectal
temperature of the anesthetised animals was maintained at
37 ( 0.5 °C by a feedback-regulated heating pad.
Permanent MCA Occlusion in Mice. Male mice (NMRI
strain) weighing 23-30 g were anesthetized with Avertin
(tribromoethanol dissolved in 10% cremophor), 240 mg/kg ip
in a volume of 10 mL/kg. Under an operating microscope (Wild
AG, Heerbrugg, Switzerland) with low-power magnification,
a 5-10 mm incision of the skin was made vertically on the
left side between the ear and the orbit. The parotid gland and
surrounding soft tissue were removed by electrocoagulation.
Parts of the underlying temporal muscle were removed until
the MCA became visible through the surface of the skull. A
small burr hole (1 mm) was made using a high-speed microdrill
through the outer surface of the semitranslucent skull just over
the visibly identified MCA. The inner layer of the skull was
removed with fine forceps. The MCA was electrocoagulated
with a bipolar diathermy above the level of the olfactory tract;
thus, lenticulostriate arteries were left intact. Afterward, the
wound was sutured. The animals did not sustain blood loss,
and all surgeries were completed within 7-10 min.
invaluable technical assistance. We also thank Stuart
Ince for proofreading of the manuscript. Appreciation
is expressed to Sandor Solyom of IVAX Drug Research
Institute for his expert collaboration in the design of
the compounds.
Supporting Information Available: Table on cycles of
2-amino-3,3-dimethoxybutane (12) addition and heating in the
synthesis of 14a, yields, mass spectra analyses, and HPLC
reanalyses. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) Dirnagl, U.; Iadecola, C.; Moskowitz, M. A. Pathobiology of
ischaemic stroke: an integrated review. Trends Neurosci. 1999,
22, 391-397.
(2) Groom, A. J.; Smith, T.; Turski, L. Multiple sclerosis and
glutamate. Ann. N.Y. Acad. Sci. 2003, 993, 229-275.
(3) Lees, G. J. Pharmacology of AMPA/kainate receptor ligands and
their therapeutic potential in neurological and psychiatric
disorders. Drugs 2000, 59, 33-78.
(4) Takahashi, M.; Ni, J. W.; Kawasaki-Yatsugi, S.; Toya, T.; Ichiki,
C.; Yatsugi, S.-I.; Koshiya, K.; Shimizu-Sasamata, M.; Yamagu-
chi, T. Neuroprotective efficacy of YM872, an R-amino-3-hydroxy-
5-methylisoxazole-4-propionic acid receptor antagonist, after
permanent middle cerebral artery occlusion in rats. J. Pharma-
col. Exp. Ther. 1998, 287, 559-566.
(5) Donevan, S. D.; Rogawski, M. A. GYKI 52466, a 2,3-benzodiaz-
epine, is a highly selective, noncompetitive antagonist of AMPA/
kainate receptor responses. Neuron 1993, 10, 51-59.
(6) Backhauss, C.; Karkoutly, C.; Welsch, M.; Krieglstein, J. A
mouse model of focal cerebral ischemia for screening neuropro-
tective drug effects. J. Pharm. Methods 1992, 27, 27-32.
(7) Elger, B.; Hornberger, W.; Schwarz, M.; Seega, J. MRI study on
delayed ancrod therapy of focal cerebral ischaemia in rats. Eur.
J. Pharmacol. 1997, 336, 7-14.
(8) Sherman, D. G.; Atkinson, R. P.; Chippendale, T.; Levin, K. A.;
Ng, K.; Futrell, N.; Hsu, C. Y.; Levy, D. E. For the STAT
participants. Intravenous ancrod for treatment of acute ischemic
stroke. JAMA, J. Am. Med. Assoc. 2000, 283, 2395-2403.
(9) Solyom, S.; Tarnawa, I. Non-competitive AMPA antagonists of
2,3-benzodiazepine type. Curr. Pharm. Des. 2002, 8, 913-939.
(10) Grasso, S.; Micale, N.; Zappala, M.; Galli, A.; Costagli, C.;
Menniti, F. S.; De Micheli, C. Characterization of the mechanism
of anticonvulsant activity for a selected set of putative AMPA
receptor antagonists. Bioorg. Med. Chem. Lett. 2003, 13, 443-
446.
Body Temperature. The core body temperature was
determined in conscious rats using a rectal probe (eight
animals/group). Rats were placed separately in Perspex boxes.
The temperature was repeatedly measured up to 3 h after drug
or vehicle injection. The method was validated using the
5-HT1A receptor agonist 8-OH-DPAT (1 mg/kg sc), which is
known to reduce body temperature in rats.28
Measurement of Lesion Volume Using 2,3,5-Triph-
enyltetrazolium Chloride (TTC). Twenty-four hours after
MCA occlusion, animals were anesthetized using Nembutal.
The chests were opened, and the brains were perfused by
intracardiac infusion of 1 mL of a saline solution containing
10% TTC. The brains were left in situ for 10-15 min, removed,
and stored in fixative (4% paraformaldehyde and 2% sucrose
in 0.1 M sodium phosphate) for 3 days. The brains were sliced
into 1 mm coronal sections with a matrix (Harvard Bioscience).
The slices were captured using a color camera combined with
a Macintosh computer (Apple, Cupertino, CA). The areas of
the infarcts were measured using the NIH Image software
package (freeware), and the infarct volumes were calculated
according to the Cavalieri principle.29 This procedure has been
shown to be reliable for distinction between ischemic and
nonischemic tissue.30
Drugs. MCA-occluded rats and mice received four ip injec-
tions of a solution of 14a containing 10% cremophor EL
(Sigma). The injections were spaced by 1 h intervals. Pretreat-
ment, starting 1 h before MCA occlusion was used in mice,
and posttreatment beginning 1 h after MCA occlusion were
used in rats. The NMDA receptor antagonist MK-801 (RBI)
and the AMPA receptor antagonist NBQX (Novo Nordisk) were
dissolved in physiological saline (0.9% NaCl) and studied as
reference compounds using treatment paradigms that have
been efficacious in animal models of global as well as focal
cerebral ischemia.11-13 MK-801 (3 mg/kg) was administered
ip immediately after MCA occlusion, and NBQX (30 mg/kg)
was injected ip 60, 70, and 85 min after MCA occlusion.
Statistical Analyses. All results are given as the mean
and standard deviation (SD). For each experimental group the
deviation from the Gaussian distribution was tested by the
Kolmogorov-Smirnov test. All the data passed the normality
test. Statistical comparisons between control and treatment
groups were made using analysis of variance followed by post-
hoc Dunnett’s multiple comparison test.
(11) Sauter, A.; Rudin, M. Strain-dependent drug effects in rat middle
cerebral artery occlusion model of stroke. J. Pharmacol. Exp.
Ther. 1995, 274, 1008-1013.
(12) Sheardown, M. J.; Nielsen, E. O.; Hansen, A. J.; Jacobsen, P.;
Honore, T. 2,3-Dihydroxy-6-nitro-7-sulfamoyl-benzo(f)quinoxa-
line: A neuroprotectant for cerebral ischemia. Science 1990, 247,
571-574.
(13) Wood, N. I.; Rothaul, A. L.; Meakin, J. E.; Hunter, A. J. NBQX
reduces lesion volume in a rat model of focal cerebral ischemia.
Br. J. Pharmacol. 1993, 108 (Suppl.), 265P.
(14) Gitto, R.; Orlando, V.; Quartarone, S.; De Sarro, G.; De Sarro,
A.; Russo, E.; Fereri, G.; Chimirri, A. Synthesis and evaluation
of pharmacological properties of novel annelated 2,3-benzodi-
azepine derivatives. J. Med. Chem. 2003, 46, 3758-3761.
(15) Fujii, T.; Tashiro, M.; Ohara, K.; Kunai, M. The preparation of
3,4-dimethoxy- and 3,4-methylendioxy-phenethylalcohol. Chem.
Pharm. Bull. 1960, 8, 266-269.
(16) Weinstein, C.; Craig, A. R. A synthetic approach to the cepha-
lotoxine skeleton. J. Org. Chem 1976, 41, 875-878.
(17) Bradshaw, D. P.; Jones, D. W.; Tideswell, J. 2-Benzopyran-3-
ones, stabilised by alkoxy-substituents, J. Chem. Soc., Perkin
Trans. 1 1991, 169-173.
(18) De Sarro, A.; de Sarro, G.; Giotto, R.; Grasso, S.; Silvana, N.;
Zappala, M. Synthesis and anticonvulsant activity of new 2,3-
benzodiazepines as AMPA receptor antagonists. Farmaco 1999,
54 (3), 178-187.
(19) Wang, Y.; Konkoy, C. S.; Ilyin, V. I.; Vanover, K. E.; Carter, R.
B.; Weber, E.; Keana, J. F. W.; Woodward, R. M.; Cai, S. X.
Synthesis of 7,8-(methylenedioxy)-1-phenyl-3,5-dihydro-4H-2,3-
benzodiazepin-4-ones as novel and potent noncompetitive AMPA
receptor antagonists. J. Med. Chem. 1998, 41 (14), 2621-2625.
(20) Grasso, S.; de Sarro, G.; de Sarro, A.; Nicale, N.; Zappala, M.;
Pina, G.; Baraldi, M.; de Micheli, C. Synthesis and anticonvul-
sant activity of novel and potent 2,3-benzodiazepine AMPA/
kainate receptor antagonists, J. Med. Chem. 1999, 42 (21),
4414-4421.
Acknowledgment. The authors thank Karola Behn,
Peter Boettcher, Kay Do¨rre, Martin Gieseler, Tanja
Jung, Hatu Lam, Oliver Schmuecker, Ingrid Schumann,
Christopher Kossmann, and Astrid Seltz for their
(21) Reader, V. A. An efficient synthesis of 2-(methylaminomethyl)-
4,5-dialkyl-1H-imidazoles. Synlett 1998, 10, 1077-1078.

Noncompetitive AMPA Receptor Antagonists
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4627
(22) Adachi, J.; Sato, N. Studies on pyrazines. I. The synthesis of
2,3-dihydropyrazines and their derivatives. J. Org. Chem. 1972,
37, 221-225.
early neuropathological consequences following middle cerebral
artery occlusion. J. Cereb. Blood Flow Metab. 1981, 1, 53-60.
(28) Elger, B.; Seega, J.; Raschack, M. Oedema reduction by leve-
mopamil in focal cerebral ischemia of spontaneously hyperten-
sive rats studied by magnetic resonance imaging. Eur. J.
Pharmacol. 1994, 254, 65-71.
(29) Gundersen, H. J. G.; Bendtsen, T. F.; Korbo, L.; Moller, A.;
Nielsen, K.; Nyengaard, J. R.; Pakkenberg, B.; Sorensen, F. B.;
Vesterby, A.; West, M. J. Some new, simple and efficient
stereological methods and their use in pathological research and
diagnosis. Acta Pathol., Microbiol. Immunol. Scand. 1988, 96,
379-304.
(23) Baron, B. M.; Harrison, B. L.; Kehne, J. H.; Schmidt, C. J.; Van
Giersbergen, P. L.; White, H. S.; et al. Pharmacological charac-
terization of MDL 105,519, an NMDA receptor glycine site
antagonist. Eur. J. Pharmacol. 1997, 323, 181-192.
(24) Hall, R. A.; Kessler, M.; Lynch, G. Kainate binding to the AMPA
receptor in rat brain. Neurochem. Res. 1994, 19, 777-782.
(25) Lehmann, J.; Hutchinson, A. J.; McPherson, S. E.; Mondadori,
C.; Schmutz, M.; Sinton, C. M.; et al. CGS 19755, a selective
and competitive N-methyl-D-aspartate-type excitatory amino
acid receptor antagonist. J. Pharmacol. Exp. Ther. 1988, 246,
65-75.
(30) Sager, T. N.; Hansen, A. J.; Laursen, H. Correlation between
N-acetylaspartate levels and histopathologic changes in cortical
infarcts of mice after middle cerebral artery occlusion. J. Cereb.
Blood Flow Metab. 2000, 20, 780-788.
(26) Morgan, R. C.; Merger, L. D.; Cincotta, M.; Beart, P. M. Binding
of [3H]AMPA to non-chaotrope, non-detergent treated rat syn-
aptic membranes: Characteristics and lack of effect of barbitu-
rates. Neurochem. Int. 1991, 18, 75-84.
(27) Tamura, A.; Graham, D. I.; McCulloch, J.; Teasdale, G. M. Focal
cerebral ischaemia in the rat: 1. Description of technique and
JM0580003