Synthesis, anticonvulsant properties and pharmacokinetic profile of novel 10,11-dihydro-10-oxo-5H-dibenz/b,f/azepine-5-carboxamide derivatives

Article abstract of DOI:10.1016/S0223-5234(01)01220-X

A series of novel derivatives of oxcarbazepine (5), 10,11-dihydro-10-oxo-5H-dibenz/b,f/azepine-5-carboxamide was synthesised and evaluated for their anticonvulsant activity and sodium channel blocking properties. The oxime 8 was found to be the most active compound from this series, displaying greater potency than its geometric isomer 9 and exhibiting also the highest protective index value. Importantly, the metabolic profile of 8 differs from the already established dibenz/b,f/azepine-5-carboxamide drugs such as 1 and 5 which undergo rapid and complete conversion in vivo to several biologically active metabolites. In contrast 8 is metabolised to only a very minor extent leading to the conclusion that the observed anti-convulsant effect is solely attributable to 8. It is concluded that 8 may be as effective as 1 and 5 at controlling seizures and that the low toxicity and consequently high protective index should provide the compound with an improved side-effect profile.

Full text of DOI:10.1016/S0223-5234(01)01220-X

Eur. J. Med. Chem. 36 (2001) 227–236
´
© 2001 Editions scientifiques et me´dicales Elsevier SAS. All rights reserved
PII: S0223-5234(01)01220-X/FLA
Original Article
Synthesis, anticonvulsant properties and pharmacokinetic profile of novel
10,11-dihydro-10-oxo-5H-dibenz/b,f/azepine-5-carboxamide derivatives
David A. Learmontha, Jan Benesa, Anto´nio Paradab, Dominik Hainzlb, Alexander Beliaeva,
Maria Joa˜o Bonifa´ciob, Pedro M. Matiasc, Maria A. Carrondoc, Jose´ Garrettb,1,
Patr´ıcio Soares-da-Silvab*
aDepartment of Research & Development, Laboratory of Chemistry, BIAL, 4785 S. Mamede do Coronado, Portugal
bDepartment of Research & Development, Laboratory of Pharmacology, BIAL, 4785 S. Mamede do Coronado, Portugal
cInstituto de Tecnologia Qu´ımica e Biolo´gica, Oeiras, Portugal
Received 11 October 2000; revised 8 January 2001; accepted 11 January 2001
Abstract – A series of novel derivatives of oxcarbazepine (5), 10,11-dihydro-10-oxo-5H-dibenz/b,f/azepine-5-carboxamide was
synthesised and evaluated for their anticonvulsant activity and sodium channel blocking properties. The oxime 8 was found to be
the most active compound from this series, displaying greater potency than its geometric isomer 9 and exhibiting also the highest
protective index value. Importantly, the metabolic profile of 8 differs from the already established dibenz/b,f/azepine-5-carboxamide
drugs such as 1 and 5 which undergo rapid and complete conversion in vivo to several biologically active metabolites. In contrast
8 is metabolised to only a very minor extent leading to the conclusion that the observed anti-convulsant effect is solely attributable
to 8. It is concluded that 8 may be as effective as 1 and 5 at controlling seizures and that the low toxicity and consequently high
´
protective index should provide the compound with an improved side-effect profile. © 2001 Editions scientifiques et me´dicales
Elsevier SAS
Anticonvulsant / dibenz/b,f/azepine-5-carboxamides / metabolism / sodium channel blockade / oxcarbazepine derivatives
1. Introduction
formation of the meso-epoxide 2 (figure 1). This ac-
tive main metabolite is further converted by a micro-
somal epoxide hydrolase into a pharmacologically
Carbamazepine (dibenz/b,f/azepine-5-carboxamide,
1) has become established as an effective agent in the
inactive mixture of (10S,11S)-trans-diol 3 and its
management of epilepsy, trigeminal neuralgia and af-
enantiomer 4 in the proportion of ca. 9/1 in humans
fective disorders [1]. However, administration of 1 in
[7].
humans is complicated by potent induction of hepatic
Oxcarbazepine 5, also structurally based upon the
oxidative enzymes, by adverse central nervous system
dibenz/b,f/azepine framework, was later found to
effects and frequent and serious idiosyncratic reac-
possess comparable antiepileptic potency but impor-
tions [2, 3]. Potent induction of hepatic microsomal
tantly provokes generally fewer side-effects. This ob-
enzymes [4] that cause self-induction of its own
servation is attributable to the very different
metabolic profile of 5 where oxidative attack leading
to the epoxide 2 is avoided and the predominant
pathway involves fast reduction of the ketone func-
metabolism [5] together with slow and erratic absorp-
tion makes medication and especially polymedication
with 1 more complicated [6]. General and neuronal
toxicity of 1 has been ascribed in part to the in vivo
tionality. However, a crucial consequence of the non-
stereoselective reduction of this apparently achiral
prodrug is the undesirable formation of a mixture of
two active enantiomeric alcohols (S)-6 and (R)-7 in
* Corresponding author
E-mail address: psoares.silva@bial.pt (P. Soares-da-Silva).
1 Deceased.

228
D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
the ratio of ca. 4/1, the anticonvulsant properties of
which have been described previously [8].
2. Chemistry
New derivatives of 10,11-dihydro-10-oxo-5H-
dibenz/b,f/azepine-5-carboxamide (5) were prepared
by either direct oximation of oxcarbazepine or by
condensation with free amino group-containing com-
pounds, both reactions satisfactorily achieved in gen-
erally good yields via standard procedures. The
parent oxime 8 subsequently underwent smooth acy-
lation using various selected anhydrides although in
certain cases the more readily available acid chlorides
were used instead without detriment. Surprisingly,
attempted O-alkylation of 8 (e.g. methylation, benzy-
lation) proceeded in poor yields under standard reac-
tion conditions used and condensation of 5 with the
appropriate alkyloxyamine hydrochloride was found
prudent in these cases. Interestingly, HPLC analysis
of the oxime 8 indicated in each batch the presence of
a minor product with similar retention time which
could not be removed by recrystallisation and which
was subsequently found to increase in concentration
when a solution of the oxime was allowed to stand in
daylight. This observation was confirmed when the
solution was irradiated under conditions simulating
exterior natural light and which facilitated isolation
of the impurity by column chromatography in suffi-
cient quantity for characterisation. The MS spectrum
of the isolated product displayed the same molecular
ion as for 8 leading to the suspicion that this sub-
stance was in fact the geometric isomer of the oxime.
Unfortunately the NMR spectra of the isolated
product and 8 were both conceivably compatible with
either the E- or Z-isomer. However, a study of the
results of NOE difference spectroscopy of the sepa-
rate compounds implied that 8 exists as the E-isomer
and the impurity 9 as the Z-isomer (figure 2).
We have recently described [9] the synthesis and
anti-convulsant profile of enantiomerically pure ester
derivatives of both (S)-6 and (R)-7, which were de-
signed to avoid such stereochemical complications
during the metabolic pathway, with the objective to
produce more effective alternative medications to 1
and 5 but endowed with the greater advantages of
increased safety-profile and reduced side-effects. As
part of an ongoing research programme concerning
the discovery of improved anticonvulsive agents, the
present study was conducted to evaluate potential
differences in anticonvulsant activity and metabolic
profile of novel derivatives of 5. Structural modifica-
tions of 5 were carried out, primarily entailing re-
placement of the carbonyl moiety by selected
substituted imino-groups (CꢀN–R) as a means to
influence the metabolic profile from that described for
5 whilst maintaining high activity. These new com-
pounds were tested in seizure models predictive of
anticonvulsant efficacy in generalised tonic–clonic
seizures, namely the maximal electroshock (MES)
test. The most active compound identified from this
series, the 10-hydroxyimino derivative 8 was there-
after further tested and compared with 1 and 5. The
putative mechanism of action of 8 was evaluated in
experiments designed to study interference with
voltage-gated sodium channels in rat brain mem-
branes. Furthermore, the metabolic profile was stud-
ied in several species and compared to that for 1 and
5.
Conclusive evidence was obtained by X-ray diffrac-
tion analysis of 8, which unequivocally confirmed the
E-configuration of this compound. The crystal struc-
ture contains two independent molecules of 8, which
are approximately related by a non-crystallographic
inversion centre. However, the crystal structure can-
not be centrosymmetric because the OH groups in the
two associated molecules of ethanol are not suitably
oriented as shown in figure 3. Both molecules of 8
have E- stereochemistry with respect to the oxime
groups: the torsion angle C(11)–C(10)–N(10)–O(10)
is −0.3(5) and −1.4(4)° in molecules 8a and 8b, re-
spectively, while the torsion angle C(5A)–N(5)–
C(5)–O(51) is −0.8(4) and −5.3(4)° in molecules 8a
Figure 1. Metabolism of compounds 1 and 5.

D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
229
Figure 2. Results of NOE difference spectroscopy of compounds 8 and 9.
Figure 3. Figure 2. ORTEP [21] stereo diagram of the unit cell contents of the 8·C2H6O crystal structure showing the non-crystallo-
graphic inversion centre between the two independent molecules of 8 (8a and 8b) and the relative orientation of the OH groups of
C2H6O which may be responsible for the lack of an crystallographic inversion centre in the crystal structure.
and 8b, respectively. The asymmetric unit of this
crystal structure further contains four solvent ethanol
molecules (C2H6O). Molecular diagrams presenting
the atomic numbering scheme and thermal ellipsoid
plots for both independent molecules of 8 are shown
in figure 4.
For the identification of metabolites, the synthesis
of the unsaturated nitro-derivative 10 was achieved
according to the literature procedure [10] which then
underwent facile reduction with sodium borohydride
to the saturated analogue 11. Yields and physical
properties of test compounds are given in table I.
administered by the oral route at a dose equimolar to 15
mg kg−1 of oxcarbazepine 5. Two hours after their
administration, the MES test was performed for each
compound. Initial evaluation of the anticonvulsant ac-
tivity of the new derivatives (table II) showed that the
unsubstituted oxime 8 displayed activity comparable to
or greater than the standards 1 and 5. Interestingly, the
acetate 12 derived from 8 displayed dramatically re-
duced activity which was inherently surprising since
enzymatic hydrolysis in vivo was predicted to give origin
to the parent oxime. In general the various acylated
forms of the oxime, designed to span a broad range of
hydrolytic and enzymatic stability exhibited reduced ac-
tivity, with only glutaroyl derivative 19 showing the
similar moderate activity of 12. Of the carbonate series
derived from 8 the t-butyl carbonate 22 was clearly
superior to the n-butyl isomer 21 and lower homologue
20, retaining activity comparable to 5. Substitution of
the free hydroxyl functionality of the oxime by a methyl
group (24) resulted in a compound possessing mar-
3. Results and discussion
3.1. Pharmacology
The compounds were tested for anticonvulsant activ-
ity using the procedures described previously [11, 12]. In
the first series of experiments, the compounds were

230
D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
producing almost no motor impairment, the non-linear
regressions for the determination of the TD50 could not
be performed (figure 5). Considering the Protective In-
dex (TD50 p.o./ED50 p.o.) as a measure of therapeutic
tolerability, this data indicates that compound 8 should
be better tolerated than 1 and 5.
ginally higher activity than most acylated derivatives but
introduction of the bulky O-benzyl group (25) abolished
activity completely. Further investigation revealed that
changing the substituent on the nitrogen atom of the
imino group from oxygen to nitrogen as in hydrazone
26 reduced activity substantially and introduction of a
phenyl substituent (27) had the same effect as 25 in the
carbonate series. Notably, of the metabolites of 8 iden-
tified from rats, the low activity associated with the
unsaturated nitro-compound 11 was completely extin-
guished on saturation to the analogue 10. The interest-
ing observation was also noted of the low
anti-convulsant activity of the pure (Z)-isomer 9, ca.
50% that of (E)-isomer 8. These results would indicate
that the observed anti-convulsant effect is exclusively
attributable to the oxime 8 itself and mainly to the
(E)-isomer.
The administration by gastric tube of compounds
listed in table III conferred a dose-dependent protection
of rats against MES induced seizures. Compound 8 was
found to be equally potent to 1 and more potent than 5
at 2 and 8 h after their administration. Four hours after
administration compounds 8 and 5 had similar potency
and both were less potent than 1.
3.2. Metabolism
The metabolism and pharmacokinetic profile of com-
pound 8 were evaluated in male Wistar rats, male CD-1
mice and female New Zealand rabbits (figure 6) and
appear different in each of the three species. While the
non-active nitro-derivative 11 was the major metabolite
detected in the rat, by contrast the active 10-hydroxy
compound 21 was found to be the most abundant
metabolite in the rabbit (table IV). Mice appear inca-
pable of metabolising compound 8 very effectively so
that even 8 h post dose only traces of other products are
detectable. The unsaturated nitro derivative 10 was only
detected in the rat and its role as a possible intermediate
between 11 and 5 has not yet been established.
3.3. Sodium channel blocking properties
The oral administration of 1 and 5 conferred a dose
and time dependent motor impairment of rats in the
rotating rod apparatus. Compound 8 caused less motor
impairment than 1 and 5 (table III). Due to the ex-
tremely low toxicity of 8, with even 2000 mg kg−1
At therapeutic concentrations, 1 has a specific action
to prevent seizures without diminishing normal electrical
activity in the brain. Although the mechanism by which
1 exerts its antiepileptic effect is not clear, it has been
proposed to inhibit voltage-dependent sodium channels
Figure 4. Figure 3. ORTEP [21] diagram of both independent molecules of 8 showing the atomic notation. The thermal ellipsoids of
the non-hydrogen atoms have been drawn at the 40% probability level.

D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
231
Table I. Structure, physical properties and methods of prepa-
ration of compounds.
pare the relative potency of these compounds as
anticonvulsants with their relative potency to bind to
sodium channels and to modulate sodium entry. All
dibenz/b,f/azepine-5-carboxamide derivatives listed in
table V, within the concentration range tested (3–1000
mM), displaced [3H]BTX (10 nM) binding to rat brain
membranes. As shown in this table, the relative potency
(IC50 in mM) of 1 was significantly lower than that of
compound 8. Compound 11, the major metabolite of 8,
had no significant effect on the binding of [3H]BTX to
its binding site (figure 7).
Recently, Unverferth et al. have suggested a pharma-
cophore model for structurally different anticonvulsants
containing aryl rings and electron donor and hydrogen
bond donor/acceptor functions [20]. Carbamazepine 1
was found to satisfy the demands of this model and
shares with 8 the dibenz/b,f/azepine-5-carboxamide nu-
cleus with the amide bond serving as the hydrogen bond
donor/acceptor. It could be envisaged that the nitrogen
atom of the hydroxyimino group of the oxime 8 replaces
the olefinic function of 1 as the electron donor so that
Compound R
Method M.p. (°C) Yield (%)
8
9
OH
OH
A
238–240 86
a
12
13
14
15
16
17
OCOMe
OCOn-Pr
OCOt-Bu
OCOPh
B
B
B
B
176–177 85
160–161 58
192–193 61
170–171 53
185–186 55
188–189 53
OCO–3-OMePh B b
OCO–2-
B
B
B
Naphthyl
OCO(CH2)2-
CO2H
OCO(CH2)3-
CO2H
Table II. Protection by 1, 5 and new 10,11-dihydro-10-oxo-
18
19
179–180 48
169–170 74
5H-dibenz/b,f/azepine-5-carboxamide
derivatives
against
MES-induced seizures a.
Compound
Protection (%)
20
21
22
23
24
25
26
27
28
OCOOEt
OCOOnBu
OCOOtBu
OCOOCH2Ph
OMe
OCH2Ph
NH2
NHPh
NH(2,4-di-
NO2Ph)
B b
B b
B c
B b
A
A
C
C
C
189–190 72
167–168 72
179–180 51
189–190 34
158–160 87
161–162 58
209–211 46
220–221 72
244–245 45
1
5
100.090.0
68.3920.3
90.699.4
19.9920.8
1.899.9
8
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
11
10
9
3.596.5
14.594.8
−6.399.0
13.495.5
8.799.1
29
30
NHCONH2
NHCSNH2
C
C
247–249 68
238–240 61
20.6922.8
14.0911.6
13.398.0
64.3922.1
8.2911.7
36.7921.4
−8.792.7
23.6921.4
−2.498.6
−13.198.5
−1.394.6
−8.699.7
18.2917.7
36.0920.9
40.2924.5
a See Section 5.
b Corresponding chloride used instead of anhydride.
c Di-t-butyl dicarbonate used for acylation.
[13], to inhibit calcium channels and to interact with
adenosine receptors [14] amongst other mechanisms [15].
In fact, sodium channel inhibition has been a natural
candidate for the mechanism of action of 1 based on the
reported interaction with receptor sites involved in the
activation of sodium channels [16], and on the modula-
tion of sodium entry [17] and sodium currents [18, 19].
To confirm that sodium channels are a primary target
for the action of the new dibenz/b,f/azepine-5-carbox-
amide derivatives, it was considered necessary to com-
a All compounds were given by gastric tube at 15 mg kg−1
.
Values are mean9S.E.M. for five to eight rats per group.

232
D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
Table III. ED50, TD50 and protective index values for 1, 5 and 8 a.
Compound
Time (h)
MES, ED50 (mg kg−1
)
Rotarod, TD50 (mg kg−1
)
Protective index b
1
2
4
8
2
4
8
2
4
8
591
1291
1892
1092
2093
2495
1292
2092
1793
251910
390930
591991
998996
1273985
1498980
\2000
50.2
32.5
32.8
99.8
63.7
5
8
62.4
\166
\100
\117
\2000
\2000
a All compounds were given by gastric tube. Data points are mean9S.E.M. for five to eight rats per group. For the calculation of
the ED50 and TD50 values the parameters of the logistic equation were fitted to the experimental data.
b TD50/ED50.
this compound also meets the requisites of the proposed
model thereby correlating with the high observed
potency.
on an Electrothermal Model 9100 hot stage apparatus
and are uncorrected. NMR spectra were recorded on a
Bruker Avance DPX (400 MHz) Spectrometer with
solvent used as internal standard, and data are reported
in the order: chemical shift (ppm), multiplicity (s, sin-
glet; d, doublet; t, triplet; m, multiplet; br, broad),
number of protons, approximate coupling constant in
Hertz and assignment of a signal. IR spectra were
measured with a Bomem Hartmann & Braun MB Series
FTIR spectrometer using KBr tablets. Analytical HPLC
was performed on a Gilson System equipped with a
Model 305 pump and 117 UV detector, LiChrospher
100 RP-18 EcoCART 125-3 cartridges (Merck) in com-
bination with acetonitrile/water mixtures. Analytical
TLC was performed on precoated silica gel plates
(Merck 60 Kieselgel F 254) and visualised with UV
light. Preparative chromatography was done on Merck
60 Kieselgel (0.063–0.2 mm). Elemental analyses were
performed on a Fisons EA 1110 CHNS instrument and
all analyses are consistent with theoretical values to
within 0.4% unless indicated. Solvents and reagents
were purchased from Aldrich or E. Merck. Standard
work-up refers to washing of a reaction mixture in
water-immiscible solvent consecutively with ice-cold 0.1
N aqueous hydrochloric acid, saturated aqueous sodium
bicarbonate solution and brine, followed by drying with
sodium sulphate and evaporation at 45°C at water
aspirator vacuum. X-ray crystallography-data collec-
tion: reflection intensities for 8·C2H6O were measured
with an Enraf-Nonius CAD-4 diffractometer and Ni-
4. Conclusion
A series of novel derivatives of 10,11-dihydro-10-
oxo-5H-dibenz/b,f/azepine-5-carboxamide 5 were syn-
thesised. Oxime 8 was found to exhibit greatest
activity from this new family through screening in
vivo. Most notable however is the fact that chemical
modification of 5 has given rise to a compound ex-
hibiting a unique metabolic profile when compared to
the already established dibenz/b,f/azepine-5-carbox-
amide drugs such as 1 and 5 which undergo rapid and
complete conversion in vivo to several biologically
active metabolites. In contrast, 8 is metabolised to
only a very minor extent in some species, leading to
the conclusion that the high observed anti-convulsant
effect is directly attributable to 8. It is suggested that
8 may be as effective as 1 and 5 at controlling seizures
and that the low toxicity and consequently high pro-
tective index should provide the compound with an
improved side-effect profile. A tentative mechanism of
action for this compound is proposed based on the
observed interaction with sodium channels in rat
brain membranes.
5. Experimental
,
filtered Cu Ka radiation (u=1.5418 A) using the ꢀ–2q
scan technique. Data reduction included corrections for
background, Lorentz and polarisation effects. No ab-
sorption corrections were deemed necessary (v=0.748
mm−1). The structure was determined by application of
5.1. Chemistry
Melting points were measured in open capillary tubes

D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
233
direct methods (SHELXS-97) [21] and refined by full-ma-
trix least-squares on F2 for all observations (SHELXL-97)
[22]. The non-hydrogen atoms were refined with an-
isotropic thermal motion parameters. The hydrogen
atoms were placed in calculated positions and refined
according to the riding or rotating group models. The
geometric calculations were performed with the program
Figure 6. Metabolites of 8 formed in different species.
SHELXL-97 [22]. The OSCAIL [23] interface was used as a
front end to all calculations.
5.1.1. 10-(E-Hydroxyimino)-10,11-dihydro-
dibenz/b, f/azepine-5-carboxamide (8)
General method A: oxcarbazepine (5) (1.40 g, 5.56
mM) and hydroxylamine hydrochloride (0.772 g, 11.1
mM) were suspended in ethanol (15 mL), and pyridine
(0.757 g, 9.57 mM) was added. The reaction mixture
was stirred and heated at reflux for 4 h. The solvent was
evaporated under reduced pressure and water (20 mL)
was added to the white precipitate that was then allowed
to stand at room temperature overnight. The crystalline
product was filtered off and washed by water (20 mL)
and cold ethanol (10 mL) and then dried at 60°C under
vacuum to afford the product as colourless crystals,
1
1.275 g, (86%), m.p. 238 to 240°C, H-NMR (DMSO-
d6) l 11.7 (br s, 1H, O–H), 7.7 (dd, 1H, J=8Hz, J=1.5
Hz, C9–H), 7.45–7.2 (m, 7H, Ar–H), 5.9 ( br s, 2H,
–NH2), 4.3 (d, 1H, J=10 Hz, C10–H), 3.8 (d, 1H,
J=18 Hz, C10–H); 13C-NMR (DMSO-d6) l 32.0 (C10,
128.0, 128.1, 128.8, 128.9, 130.1, 130.4, 131.1, 132.5,
135.5 (C arom.), 142.2, 144.0 (C4a,C5a arom.), 152.8
(CꢀO), 156.8 (CꢀN). Anal. (C15H13N3O2) C, H, N.
5.1.2. 10-(Z-Hydroxyimino)-10,11-dihydro-
dibenz/b, f/azepine-5-carboxamide (9)
Oxime (8) (0.5 g, 2.0 mM) was dissolved in acetoni-
trile/water (1/1, 500 mL) and the solution was irradiated
in a Sunset CPS Apparatus (filter>290 nm, 765 W m−2
,
three 1-h cycles). The solvent was evaporated under
reduced pressure and the residue was dissolved in ethyl
acetate. The Z-isomer was separated by column chro-
matography (silica gel, ethyl acetate) as a slower-moving
Figure 5. Dose response curves (p.o. administration) for 8
(ꢀ), 1 (ꢁ), 5 (ꢂ) in the rotating rod test at 2 h (upper panel),
4 h (middle panel) and 8 h (lower panel). Symbols are means
and vertical lines show S.E.M.

234
D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
Table IV. Levels of 8 and metabolites in rat, mouse and rabbit plasma (in mg mL−1) 1 h after an oral dose of 80 mg kg−1 (rat and
rabbit) and 50 mg kg−1 (mouse), respectively a.
8
11
10
31
5
3
Rat
Rabbit
Mouse
4.7890.60
5.0890.63
3.0490.14
3.3590.44
0.4190.03
0.0190.00
0.7290.14
0.0090.00
0.0090.00
0.2490.03
0.9890.06
0.0090.00
0.6290.09
0.4990.03
0.0090.00
0.0090.00
0.0690.02
0.0090.00
a Mean9S.E.M., with n=4 (rat, mouse) or 3 (rabbit).
product (0.199 g, 72.5%) as yellow crystals of m.p.
220–221°C, 1H-NMR (DMSO-d6) l 9.7 (br s, 1H,
NHAr), 7.9 (m, 1H, C9–H), 7. 5–7.2 (m, 11H, Ar–H),
6.8 (tt, 1H, J=6.5, 1.8 Hz, ArH), 6.0 (br s, 2H, –NH2),
4.1 (d, 1H, J=18 Hz, C10–H), 3.9 (d, 1H, J=18 Hz,
C10–H); 13C-NMR (DMSO-d6) l 34.1 (C10, 128.1, 128.6,
128.7, 128.9, 129.2, 129.3, 129.8, 130.8 135.2, 135.5 (C
arom.), 141.5, 144.2 (C4a,C5a arom.), 140.0 (CꢀN), 156.9
(CꢀO). Anal. (C21H18N4O) H, N; C: Calc., 73.66;
Found, 72.83%.
band. The chromatographically homogeneous fractions
were pooled, evaporated under reduced pressure and the
residue was crystallised from ethanol, affording 84 mg
(16.8%) of the pure Z-isomer as colourless crystals, m.p.
1
215–217°C, H-NMR (DMSO-d6) l 11.0 (s, 1H, O–H),
7.8 (br d, 1H, J=7.9 Hz, C9–H), 7.45–7.2 (m, 7H,
Ar–H), 5.75 ( br s, 2H, –NH2), 3.9 (br s, 2H, C10–H);
13C-NMR (DMSO-d6) l 38.5 (C10, 127.9, 128.2, 128.4,
129.9, 130.7, 131.16, 131.22, 131.6, 135.4 (C arom.),
141.8, 142.1 (C4a,C5a arom.), 150.7 (CꢀO), 157.1 (CꢀN).
Anal. (C15H13N3O2) C, H, N.
Table V. Displacement of [3H]BTX (10 nM) binding to rat
brain membranes by 1, 8 and 11, the major metabolite of 8 a.
5.1.3. 10-Acetoxyimino-10,11-dihydro-
dibenz/b, f/azepine-5-carboxamide (12)
Compounds
IC50 (mM)
General method B: a suspension of oxime (8) (0.50 g,
1.87 mmol) in dichloromethane (25 mL) was treated
with pyridine (0.72 g, 9.16 mmol) and acetic anhydride
(0.57 g, 5.61 mmol). The resulting mixture was stirred at
room temperature overnight and then diluted with
dichloromethane (10 mL). After standard workup and
crystallisation from a mixture of dichloromethane and
1
8
11
390 (250, 608)
418 (314, 557)
2291 (1458, 3598)
a Data points are means with 95% confidence values for four
to five determinations in triplicate. For the calculation of the
IC50 values the parameters of the equation for one site binding
were fitted to the experimental data [26].
ethyl acetate the product was obtained as white crystals,
1
0.491 g, (85%), m.p.176–177°C, H-NMR (CDCl ) l
3
8.0 (dd, 1H, J=7.9 Hz, J=1.5 Hz, C9–H), 7.6–7.25
(m, 7H, Ar–H), 4.9 (br s, 2H, –NH2), 4.4 (d, 1H, J=18
Hz, C10–H), 4.2 (d, 1H, J=18 Hz, C10–H), 2.4 (s, 3H,
COCH3); 13C-NMR (CDCl3) l 20.3 (CH3), 33.5 (C10,
128.1, 128.7, 129.0, 129.2, 129.3, 129.8, 130.0, 131.0,
132.3, 135.5 (C arom.), 141.8, 142.5 (C4a,C5a arom.),
156.4 (NCON), 160.5 (CꢀN), 169.1 (OCO). Anal.
(C17H15N3O3) C, H, N.
5.1.4. 10-(Phenyl-hydrazono)-10,11-dihydro-
dibenz/b, f/azepine-5-carboxamide( 26)
General method C: a mixture of 5 (0.2 g, 0.8 mmol),
phenylhydrazine (0.5 g, 4.6 mmol) and sodium acetate
(0.5 g, 6 mmol) in a mixture of water (5 mL), ethanol (5
mL) and three drops of concentrated hydrochloric acid
was heated at 60°C for 30 min and then allowed to cool
to room temperature. The precipitate was then filtered
and washed with cold water and ethanol to give the
Figure 7. Displacement of [3H]-BTX binding by 8 (ꢀ), 1 (ꢁ)
and 11 (ꢂ) in % of maximum. Symbols are means of four to
five experiments; S.E.M. were less than 10% of the correspon-
dent mean.

D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
235
5.1.5. 10-Nitro-10,11-dihydro-
dibenz/b, f/azepine-5-carboxamide (11)
described above. All compounds used were suspended in
0.5% carboxymethylcellulose (4 mL kg−1) and given by
stomach tube.
10-Nitro-dibenz[b,f]azepine-5-carboxamide (10) [10]
(10.60 g, 37.7 mM) was dissolved in a warm mixture of
THF and methanol (1/1, 380 mL) and then cooled to 5°C.
Sodium borohydride (3.58 g, 94.3 mM) was added over
30 min at 5–7°C with stirring, and the reaction mixture
was reacted for another 40 min at the same temperature.
The resulting solution was concentrated on rotary evap-
orator under vacuum until crystallisation began and then
kept overnight at 4°C. Filtration, washing with water and
ethanol and drying gave 7.70 g (72%) of white crystals,
5.2.2. Rotarod test
Rats were examined for motor toxicity in the rotating
rod apparatus (Accelerator Rota-Rod [Jones & Roberts]
7750; Ugo Basile). Naive rats were trained to hold onto
the 5 cm diameter neoprene rubberised cylinder until
achieving to maintain the equilibrium for 3 min while
rotating at 6 rpm. The day after, rats were given i.p. the
test compound dissolved in DMSO 2 mL kg−1 and 15 min
later were placed on the rotating rod at a speed of 6 r.p.m.
In a compound treated rat the neurological deficit is
indicated by the inability of the rat to maintain equi-
librium for 1 min in each of three trials [24].
1
m.p. 188–189°C, H-NMR (DMSO-d6) l 9.7 (br s, 1H,
NHAr), 7.9 (m, 1H, C9–H), 7. 5–7.2 (m, 11H, Ar–H),
6.8 (tt, 1H, J=6.5, 1.8 Hz, ArH), 6.0 ( br s, 2H, –NH2),
4.1 (d, 1H, J=18 Hz, C10–H), 3.9 (d, 1H, J=18 Hz,
C10–H); 13C-NMR (DMSO-d6) l 34.1 (C10, 128.1, 128.6,
128.7, 128.9, 129.2, 129.3, 129.8, 130.8 135.2, 135.5 (C
arom.), 141.5, 144.2 (C4a,C5a arom.), 140.0 (CꢀN), 156.9
(CꢀO). Anal. (C15H13N3O3) C, H, N.
5.2.3. Blockade of voltage-sensitive sodium channels
Blockade of voltage-sensitive sodium channels was
studied by investigating [3H]batrachotoxinin A 20-a-ben-
zoate ([3H]BTX) displacement binding to rat brain mem-
branes. Membrane preparation and binding assays were
performed as described [25], with slight modifications.
Binding studies were performed by incubation of
[3H]BTX for 30 min at 37°C with 200 mg of membrane
protein in a final volume of 200 mL in a solution
containing choline chloride (130 mM) KCl (5.5 mM),
MgSO4 (0.8 mM), glucose (5.5 mM), HEPES/TRIS (50
mM, pH 7.4), scorpion toxin (2 mM), tetrodotoxin (1 mM),
[3H]BTX 10 nM, bovine serum albumin (1 mg mL−1), and
varying concentrations of competing compounds (3–1000
mM), as previously described [16]. Nonspecific binding
was determined in the presence of 300 mM veratridine.
The binding reactions, performed in 96 wells EIA/RIA
plates (Costar) with 300 mL capacity, were initiated by the
addition of 50 mL of the membrane suspension to 150 mL
of the reaction mixture. Plates were then incubated at
37°C for 1 h in a water bath with agitation. The binding
reactions were stopped by vacuum filtration (Brandel 96
Harvester) through glassfibre filter mats A (1450–21 from
Wallac) and washing of the filters and incubation tubes
with 200 mL of an ice-cold wash solution consisting of
choline chloride (130 mM), CaCl2 (1.8 mM), MgSO4 (0.8
mM), bovine serum albumin (1 mg mL−1) and 5 mM
HEPES/TRIS (pH 7.4). The filter mats were dried,
impregnated with MeltiLex™ a scintillation mixture
(Wallac), inserted into plastic sample bags (Wallac), and
the radioactivity was counted in a 1450 MicroBetaTM
spectrophotometric detector for 2 min with an efficiency
5.2. Pharmacological methods
All compounds were tested using male Wistar rats
(Harlan-Interfauna Ibe´rica, Barcelona, Spain) ranging
from 6 to 8 weeks old.
5.2.1. MES test
MES stimulation was applied for 0.2 s, using a Ugo
Basile ECT unit 7801, with a frequency of 100 Hz, pulse
width of 0.6 ms and a current of 150 mA through bipolar
corneal electrodes. A drop of electrolyte/anaesthetic,
oxibuprocaine chloride, was applied in the eyes of all
animals immediately before placement of corneal elec-
trodes. Abolition of the hindleg tonic extensor component
was used as the endpoint. These experimental conditions
produced tonic–clonic convulsions in 97% of animals
tested and only rats showing typical tonic–clonic convul-
sions were used [11]. All rats were submitted to a
maximum of 3 MES sessions: the first MES session was
performed to screen the animals and select those rats
presenting a typical convulsive behaviour. The day after,
rats were given the compounds to be tested or the vehicle
and submitted to a second MES session 2, 4 or 8 h after
the administration of test compounds. The evaluation of
the anticonvulsive profile of test compounds was based on
the duration of the tonic phase (in seconds) being each
rat its own control (internal control) as obtained in the
first MES session. An external control group was also
studied; in this particular case, rats were given the vehicle
and submitted to the three MES sessions procedure as
of 55–60%. [3H]BTX (specific activity: 39.9 Ci mmol−1
was from NEN Life Sciences. Tetrodotoxin, veratridine
)

236
D.A. Learmonth et al. / European Journal of Medicinal Chemistry 36 (2001) 227–236
and scorpion toxin from Leiurus quinquestriatus he-
braeus were from Sigma. Scorpion toxin was resus-
pended in ice-cold distilled water (1 mg mL−1), and
incubated for 1 h at 0°C. The mixture was then cen-
trifuged at 12 000×g for 10 min and protein determined
in the supernatant.
References
[1] Loiseau P., Duche´ P., in: Levy R.H., Mattson R.H., Meldrum
M.S. (Eds.), Antiepileptic Drugs, Raven Press, New York,
1995, pp. 555–566.
[2] Shorvon S.D., Epilepsia 37 (Suppl. 2) (1996) 1–3.
[3] Pellock J.M., Epilepsia 28 (Suppl. 3) (1987) S64–S70.
[4] Eichelbaum M., Tomson T., Tybring G., Bertilsson L., Clin.
5.3. Metabolism studies
Pharmacokinet. 10 (1985) 80–90.
[5] Bertilsson L., Tomson T., Clin. Pharmacokinet. 11 (1986)
177–198.
Blood taken from the vena cava (rat and mouse) or
the ear vein (rabbit) was stabilised with one drop of
heparin in the syringe and kept on ice until centrifuged
at ca. 3000×g for 15 min (4°C). Of the supernatant
plasma phase 1 mL was added to 1 mL of 0.1 M sodium
phosphate buffer (pH 4.6), spiked with 2 ppm 10,11-di-
hydrocarbamazepine as internal standard.
[6] Faigle J.W., Feldmann K.F., in: Levy R.H., Mattson R.H.,
Meldrum M.S. (Eds.), Antiepileptic Drugs, Raven Press, New
York, 1995, pp. 499–513.
[7] Bellucci G., Berti G., Chiappe C., Lippi A., Marioni F., J.
Med. Chem. 30 (1987) 768–773.
[8] Dickinson R.G., Hooper W.D., Dunstan P.R., Eadie M.J., Eur.
J. Clin. Pharmacol. 37 (1989) 69–74.
These samples were then added to LiChrolut RP-18
SPE columns (3 mL, 200 mg, Merck) which were condi-
tioned with 2 mL water–ACN (95:5) and 2×2 mL water
(Milli-Q). After eluting the sample, the columns were
washed with 2 mL water and 2 mL water–ACN (95:5)
and dried under airflow for 5 min. The remainder was
eluted with 2×250 mL acetone, the solvent evaporated
(Speedvac, 35–40°C) and the residue reconstituted in 1
mL HPLC mobile phase and 20 mL injected into the
HPLC.
[9] Benes J., Parada A., Figueiredo A.A., Alves P.C., Freitas A.P.,
Learmonth D.A., Cunha R.A., Garrett J., Soares-da-Silva P., J.
Med. Chem. 42 (1999) 2582–2587.
[10] E. Aufderhaar, In EP0 028 028 A2.
[11] Swinyard E.A., Brown W.C., Goodman L.S., J. Pharmacol.
Exp. Ther. 106 (1952) 319–330.
[12] Krall R.J., Penry J.K., White B.G., Kupferberg H.J., Swinyard
E.A., Epilepsia 19 (1978) 409–428.
[13] Willow M., Gonoi T., Catterall W.A., Mol. Pharmacol. 27
(1985) 549–558.
[14] Marangos P.J., Post R.M., Patel J., Zander K., Parma A.,
Weiss S., Eur. J. Pharmacol. 93 (1983) 175–182.
Sample analysis was performed using LC-MS (HP
1100 Series, Agilent) with AP-ES ionisation (positive ion
detection). At the same time, UV spectra were recorded
with a diode array detector (monitoring wavelength 210
nm). The separation was performed on a RP-18 column
(LiChroCART 125-4, Superspher 100 RP-18, 5 mm,
Merck) with a 4×4 mm guard column containing the
same stationary phase. Water–acetonitrile (73:27) as
mobile phase (flowrate 1 mL min−1) was freshly pre-
pared, filtered (0.22 mm, NL 16, Schleicher & Schuell)
and degassed (USB).
Selected ion monitoring (SIM) with one mass for each
compound of interest was used for quantification with
the fragmentor energy set to 100 V for maximal detec-
tion sensitivity for these compounds. Further settings
were: capillary voltage 3500 eV, nebulizer gas tempera-
ture 350°C and nebulizer pressure 40 psi.
[15] Rogawski M.A., Porter R.J., Pharmacol. Rev. 42 (1990) 223–
286.
[16] Willow M., Catterall W.A., Mol. Pharmacol. 22 (1982) 627–
635.
[17] Yoshimura R., Yanagihara N., Terao T., Minami K., Abe K.,
Izumi F., Naunyn-Schmiedeberg’s Arch. Pharmacol. 352 (1995)
297–303.
[18] Hershkowitz N., Dretchen K.L., Raines A., J. Pharmacol. Exp.
Ther. 207 (1978) 810–816.
[19] McLean M.J., MacDonald R.L., J. Pharmacol. Exp. Ther. 238
(1986) 727–733.
[20] Unverferth K., Engel J., Hofgen N., Rostock A., Gunther R.,
Lankau H.-J., Menzer M., Rolfs A., Liebscher J., Muller B.,
Hofman H.-J., J. Med. Chem. 41 (1998) 63–73.
[21] G.M. Sheldrick, Patterson (Eds.), Superposition and Ab-Initio
Phasing, Academic Press, New York, 1997.
[22] G.M. Sheldrick, T.R. Schneider (Eds.), SHELXL: High-Resolu-
tion Refinement, Academic Press, New York, 1997.
[23] McArdle P., J. Appl. Crystallogr. 28 (1995) 65.
[24] Rogawski M.A., Yamaguchi S., Jones S.M., Rice K.C.,
Thurkauf A., Monn J.A., J. Pharmacol. Exp. Ther. 259 (1991)
30–37.
Acknowledgements
[25] Shimidzu T., Itoh Y., Tastumi S., Hayashi S., Ukai Y.,
Yoshikuni Y., Kimura K., Naunyn-Schmiedeberg’s Arch. Phar-
macol. 355 (1997) 601–608.
The authors would like to thank Dr Ana P. Freitas
for providing both NMR spectra and elemental analyses
and Paula C. Alves for technical assistance.
[26] H.G. Motulsky, P. Spannard, R. Neubig, 1.0 ed., GraphPad
Prism Software Inc., San Diego, 1994.