Synthesis and in vitro reactivity of 3-carbamoyl-2- phenylpropionaldehyde and 2-phenylpropenal: Putative reactive metabolites of felbamate

Article abstract of DOI:10.1021/tx9601566

We propose that 3-carbamoyl-2-phenylpropionaldehyde is an intermediate in the metabolism of felbamate, an anti-epileptic drug with a unique profile of therapeutic activity, and undergoes a cascade of chemical reactions responsible for the toxic properties of the parent drug. To test this hypothesis, we have synthesized 3-carbamoyl-2-phenylpropionaldehyde and evaluated its in vitro reactivity. This molecule was found to be highly unstable at physiological pH (t(1/2) ≤ 30 s) and to undergo facile elimination to 2-phenylpropenal, an α, β-unsaturated aldehyde commonly termed atropaldehyde. However, the predominant reaction pathway for 3- carbamoyl-2-phenylpropionaldehyde was reversible cyclization to generate 4- hydroxy-5-phenyltetrahydro-1,3-oxazin-2-one, a urethane that has a considerably longer half-life at physiological pH (t(1/2) ≤ 5 h) and may serve as a stable reservoir of the reactive aldehyde both in vitro and in vivo. Atropaldehyde is a potent electrophile and was found to exhibit cytotoxicity to cultured fibroblasts (50% growth inhibition (GI₅₀) = 4.1 ± 1.1 μM) comparable to the known unsaturated aldehyde toxins, 4-hydroxy-2- nonenal and acrolein. 3-Carbamoyl-2-phenylpropionaldehyde also exhibited significant cytotoxicity (GI₅₀ = 53 ± 8 μM), whereas 2-phenyl-1, 3- propanediol monocarbamate (GI₅₀ > 500 μM) and 3-carbamoyl-2- phenylpropionic acid (GI₅₀ > 500 μM) were nontoxic. We have additionally demonstrated the formation of a glutathione-atropaldehyde conjugate from the in vitro incubation of 3-carbamoyl-2-phenylpropionaldehyde with glutathione. Thus, the potent cytotoxicity and potential allergenicity of atropaldehyde implicate this unsaturated aldehyde as a possible causative agent in the toxicities observed with felbamate treatment.

Full text of DOI:10.1021/tx9601566

Chem. Res. Toxicol. 1996, 9, 1225-1229
1225
Communications
Syn th esis a n d in Vitr o Rea ctivity of
3-Ca r ba m oyl-2-p h en ylp r op ion a ld eh yd e a n d
2-P h en ylp r op en a l: P u ta tive Rea ctive Meta bolites of
F elba m a te
Charles D. Thompson,† Michael T. Kinter,‡ and Timothy L. Macdonald*,†
Chemistry Department, University of Virginia, and Departments of Microbiology and Pathology,
University of Virginia, Charlottesville, Virginia 22901
Received September 9, 1996X
We propose that 3-carbamoyl-2-phenylpropionaldehyde is an intermediate in the metabolism
of felbamate, an anti-epileptic drug with a unique profile of therapeutic activity, and undergoes
a cascade of chemical reactions responsible for the toxic properties of the parent drug. To test
this hypothesis, we have synthesized 3-carbamoyl-2-phenylpropionaldehyde and evaluated its
in vitro reactivity. This molecule was found to be highly unstable at physiological pH (t1/2
e
30 s) and to undergo facile elimination to 2-phenylpropenal, an R,â-unsaturated aldehyde
commonly termed atropaldehyde. However, the predominant reaction pathway for 3-carbamoyl-
2-phenylpropionaldehyde was reversible cyclization to generate 4-hydroxy-5-phenyltetrahydro-
1,3-oxazin-2-one, a urethane that has a considerably longer half-life at physiological pH (t1/2 g
5 h) and may serve as a stable reservoir of the reactive aldehyde both in vitro and in vivo.
Atropaldehyde is a potent electrophile and was found to exhibit cytotoxicity to cultured
fibroblasts (50% growth inhibition (GI50) ) 4.1 ( 1.1 µM) comparable to the known unsaturated
aldehyde toxins, 4-hydroxy-2-nonenal and acrolein. 3-Carbamoyl-2-phenylpropionaldehyde also
exhibited significant cytotoxicity (GI50 ) 53 ( 8 µM), whereas 2-phenyl-1,3-propanediol
monocarbamate (GI50 > 500 µM) and 3-carbamoyl-2-phenylpropionic acid (GI50 > 500 µM) were
nontoxic. We have additionally demonstrated the formation of a glutathione-atropaldehyde
conjugate from the in vitro incubation of 3-carbamoyl-2-phenylpropionaldehyde with glu-
tathione. Thus, the potent cytotoxicity and potential allergenicity of atropaldehyde implicate
this unsaturated aldehyde as a possible causative agent in the toxicities observed with felbamate
treatment.
In tr od u ction
Felbamate (1, Figure 1) was approved in J uly 1993 for
the treatment of several forms of epilepsy including
Lennox-Gastaut syndrome, a pharmacologically refrac-
tory form of epilepsy in children (1-4). Felbamate
demonstrated an excellent therapeutic index throughout
preclinical and clinical trials with only relatively mild
side effects (5-7). However, subsequent to felbamate’s
widespread distribution, significant toxicological issues
have been realized, notably aplastic anemia and hepa-
totoxicity. The severity of the side effect profile prompted
a recommendation by the US Food and Drug Adminis-
tration to withdraw patients from felbamate therapy (8).
As of J anuary 1995, a total of 33 cases of aplastic anemia
with 10 fatalities had been reported to the manufacturer
(9). This represents an estimated incidence of approx-
imately 1 in 4000-10,000 exposed patients (as com-
pared to an incidence of aplastic anemia in the general
population of 2-6 in one million (10)). In addition, there
F igu r e 1. The metabolic disposition of felbamate (1) in humans
and experimental animals.
have been 36 cases of hepatic toxicity with 5 fatalities
associated with felbamate use (11, 12). To date, the
mechanisms underlying the observed toxicities of fel-
bamate remain unidentified and no hypotheses have been
advanced to rationalize this behavior.
* To whom correspondence should be addressed: Timothy L. Mac-
donald, Chemistry Department, McCormick Rd., Charlottesville, VA
22901. E-mail: tlm@virginia.edu.
† Chemistry Department, University of Virginia.
The metabolism of felbamate has been studied in
several species including rat, rabbit, dog, and human (13,
‡ Departments of Microbiology and Pathology, University of Virginia.
X Abstract published in Advance ACS Abstracts, November 1, 1996.
S0893-228x(96)00156-7 CCC: $12.00 © 1996 American Chemical Society

1226 Chem. Res. Toxicol., Vol. 9, No. 8, 1996
Communications
F igu r e 3. The syntheses of 3-carbamoyl-2-phenylpropional-
dehyde (6) and atropaldehyde (7) (see the Experimental Proce-
dures for details).
phile. Finally, we have assessed the cytotoxicities of
3-carbamoyl-2-phenylpropionaldehyde (6), atropaldehyde
(7), and other felbamate metabolites, to fibroblast cells
in culture.
Exp er im en ta l P r oced u r es
Ca u tion . Ammonia is highly toxic and should only be
employed with care. All operations should be done in an efficient
fume hood.
Ch em ica ls a n d In str u m en ts. All reagents were purchased
from either Aldrich Chemical Co. or Sigma Chemical Co. (except
as otherwise noted) and were of the highest quality available.
HPLC was performed on a Waters 600E gradient pump with a
Waters 484 tunable absorbance detector using a 5 µm Microsorb-
mv C18 (4.6 mm × 250 mm) column (Rainin, Emeryville, CA)
and a binary solvent system: solvent A: 95% acetonitrile, 4.93%
water, and 0.07% trifluoroacetic acid; solvent B: 0.1% trifluo-
roacetic acid and 99.9% water. NMR spectra were recorded on
a General Electric QE300 spectrometer at 300 MHz. LC/MS
was conducted on a Finnigan MAT TSQ70 spectrometer equipped
with an electrospray ionization source.
Syn th esis. Atr op a ld eh yd e (7). The synthesis of 7 was
accomplished following the procedure outlined by Crossland
(Figure 3) (16). Briefly, the addition to styrene (11) of dichlo-
rocarbene generated from chloroform in the presence of potas-
sium hydroxide and the phase transfer catalyst triethylbenzy-
lammonium chloride gave 1,1-dichloro-2-phenylcyclopropane
(12). Diethyl acetal 13 was then formed from 12 by reflux in
ethanol in the presence of base. Diethyl acetal 13 was converted
to 7 by hydrolysis in aqueous formic acid. Atropaldehyde (7)
was recrystallized from a 1:1 mixture of diethyl ether and
petroleum ether. The overall yield of 7 from 11 was 38% yield.
1H NMR (CDCl3): δ 6.17 (s, 1H), 6.63 (s, 1H), 7.45 (m, 5H), 9.82
(s, 1H). 13C NMR, (CDCl3): δ 128.6, 128.9, 129.3, 134.3, 136.3,
148.9, 193.5.
3-Ca r ba m oyl-2-p h en ylp r op ion a ld eh yd e (6). Aldehyde 6
was prepared in three steps from diethyl acetal 13 produced as
an intermediate in the synthesis of atropaldehyde (7) (Figure
3). Hydroboration and oxidation of the double bond were
achieved by standard methodology: treatment of diethyl acetal
13 (0.4 mol) with BH3 (1.1 equiv) in THF at room temperature
(r.t.)1 for 3 h was followed by quenching with water (2 mL). A
solution of sodium hydroperoxide [NaOH (6.2 g) and 30% H2O2
(20 mL)] was then added dropwise and the reaction stirred for
an additional 1 h. After quenching the excess peroxide with
aqueous sodium thiosulfate and ether extraction of the reaction
mixture, 14 was purified by chromatography on silica gel (ethyl
acetate/hexane) to give a 60% yield (based on recovered starting
material).
F igu r e 2. The proposed intermediacy of 3-carbamoyl-2-phe-
nylpropionaldehyde (6) in the biotransformation of 2-phenyl-
1,3-propanediol monocarbamate (4) to 3-carbamoyl-2-phenylpop-
ionic acid (5), and the established (solid lines) and possible
(dashed lines) chemical and metabolic conversions of 6.
14). The known metabolites are shown in Figure 1. 2-(4-
Hydroxyphenyl)-1,3-propanediol dicarbamate (2) and
2-hydroxy-2-phenyl-1,3-propanediol dicarbamate (3) are
presumably the products of cytochrome P450 hydroxy-
lation and have been demonstrated in all of the species
studied. 2-Phenyl-1,3-propanediol monocarbamate (4) is
the product of hydrolysis of one of the carbamate moieties
and has also been observed in all species. 3-Carbamoyl-
2-phenylpropionic acid (5), an oxidized secondary me-
tabolite proposed to be derived from the corresponding
alcohol 4, has been identified as the major metabolite in
humans (∼12% of a total dose) and also observed in dogs
(14). Thus, hydrolysis of felbamate to the alcohol 4
comprises the first step in the primary pathway of
felbamate metabolism in humans. None of the known
metabolites have anti-epileptic activities comparable to
the parent drug (13, 15), and there has been no direct
link between any of these metabolites and the observed
toxicities.
We propose that the toxicity observed with the use of
felbamate may be a consequence of the chemical reactiv-
ity of a metabolite, that has not been directly detected,
but whose formation can be inferred based on the known
metabolites. 3-Carbamoyl-2-phenylpropionaldehyde (Fig-
ure 2: 6) is a requisite intermediate in the two-step
oxidation of the alcohol 4 to the major human metabolite
5. We hypothesize that the aldehyde could undergo
â-elimination to form 2-phenylpropenal (7), commonly
termed atropaldehyde, as illustrated in Figure 2. This
R,â-unsaturated aldehyde is an extremely reactive elec-
trophile and is capable of facile reaction with tissue
nucleophiles, which may culminate in the observed
toxicities or in the initiation of allergic responses.
We report in this communication the synthesis of
3-carbamoyl-2-phenylpropionaldehyde (6) and an inves-
tigation of its in vitro chemical reactivity, including model
systems for the initiation of toxic and/or allergic re-
sponses involving its reactivity with a sulfhydryl nucleo-
Synthesis of carbamate acetal 15 was achieved by treatment
of the precursor alcohol 14 (13 mmol) with 1,1′-carbonyldiimi-
dazole (1.1 equiv) in THF at r.t. for 1 h. The temperature of
1Abbreviations: FBS, fetal bovine serum; GI50, concentration caus-
ing 50% growth inhibition; MEM, minimal essential media; MTS,
3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophe-
nyl)-2H-tetrazolium, inner salt; PMS, phenazine methosulfate; r.t.,
room temperature.

Communications
Chem. Res. Toxicol., Vol. 9, No. 8, 1996 1227
the reaction flask was then lowered to -48 °C, and liquid
ammonia was introduced from a cold finger at -78 °C attached
to a tank of compressed NH3. After the introduction of ap-
proximately 20 drops of liquid ammonia from the cold finger,
the reaction was stirred for 3 h at -48 °C. The reaction was
then quenched by the addition of H2O (20 mL) and allowed to
stir for 1 h. Carbamate 15 was purified by chromatography on
silica gel (ethyl acetate/hexane) to give a 74% yield. 1H NMR
(acetone-d6): δ 0.89 (t, 3H, J ) 6.9 Hz), 1.04 (t, 3H, J ) 7.1
Hz), 3.11 (m, 1H), 3.5 (m, 2H), 3.3 (m, 2H), 4.17 (dd, 1H, J )
11.0, 8.3 Hz), 4.28 (dd, 1H, J ) 11.0, 4.8 Hz), 4.61 (d, 1H, J )
6.1 Hz), 5.6 (bs,2H), 7.18 (m,5H). 13C NMR (acetone-d6): δ 15.1,
15.1, 49.6, 62.7, 63.0, 65.0, 104.2, 126.9, 128.3, 129.5, 139.7,
157.3.
Carbamoyl aldehyde (6) was prepared by treatment of 15
(0.10 mmol, 20 mg) with Amberlyst 15 (10 mg) at r.t. in acetone
(2 mL) for 1 h. The mixture was then filtered to remove the
resin and the solvent removed by vacuum evaporation. Conver-
sion of 15 to 6 could routinely be achieved in 50-70% yield;
however, if the reaction was allowed to proceed longer than 1
h, substantial conversion to atropaldehyde (7) was observed.
Additionally, 6 could not be purified from the remaining starting
material by chromatography on silica gel as it decomposed to
atropaldehyde (7). Furthermore, 6 was found to decompose to
F igu r e 4. The half-life (t1/2) of 3-carbamoyl-2-phenylpropional-
dehyde (6) (solid line) and the percent formation of atropalde-
hyde produced during decomposition of 6 (dashed line) in 50
mM potassium phosphate buffer as a function of pH.
over a period of 30 min to GSH (500 µmol) in 0.5 M KPO4 buffer
(10 mL) at pH ) 8.0. The solution was allowed to stir for an
additional 30 min, after which it was extracted with chloroform
to remove side products. The aqueous layer was then lyophi-
lized to dryness and reconstituted in 0.2% acetic acid (1 mL)
and purified on a C18 solid phase extraction cartridge (Alltech,
Deerfield, IL). GSH and GSSG were eluted with 0.2% acetic
acid (v/v), and then the atropaldehyde-GSH conjugate was
eluted with 30% acetonitrile (v/v). 1H NMR (D2O): δ 2.02 (td,
2H, J ) 6.8, 5.2 Hz), 2.37 (t, 2H, J ) 6.8 Hz), 2.82 (dd, 1H, J )
14.2, 8.3 Hz), 2.99 (dd, 1H, J ) 14.2, 4.4 Hz), 3.69 (t, 1H, J )
5.2 Hz), 3.72 (m, 1H), 3.76 (d, 2H, J ) 1.5 Hz), 4.05 (d, 2H, J )
1.9 Hz), 4.48 (dd, 1H, J ) 8.3, 4.4 Hz), 7.3 (m, 2H), 7.5 (t, 2H,
J ) 7.6 Hz), 7.6 (t, 1H, J ) 7.8 Hz), 7.9 (d, 1H, J ) 7.8 Hz).
LC/MS: MH+ ) 438.7.
7 at r.t. in dimethyl sulfoxide (t1/2 ≈ 35 min), chloroform (t1/2
≈
20 min), and acetone (t1/2 ≈ 6.5 h). However, if the acetone was
removed immediately after the reaction, carbamoyl aldehyde 6
could be stored neat at -40 °C for several days without
significant decomposition. 1H NMR (acetone-d6): δ 3.90 (ddd,
1H, J ) 7.3, 6.2, 1.9 Hz), 4.23 (dd, 1H, J ) 11.2, 6.2 Hz), 4.52
(dd, 1H, J ) 11.2, 7.3 Hz), 5.79 (bs, 2H), 7.2 (m, 5H), 9.64 (d,
1H, J ) 1.9 Hz).
4-Hydr oxy-5-ph en yltetr ah ydr o-1,3-oxazin -2-on e (8). Ure-
thane 8 was prepared by stirring aldehyde 6 (50 µmol, 10 mg)
in H2O (5 mL) at rt for 30 min. During this time, 6 decomposes
primarily to 8 with some elimination to atropaldehyde (7).
Extraction of the aqueous layer with chloroform removed all of
the organics, except for the urethane which remained in the
aqueous layer. The aqueous layer was then lyophilized to
dryness, after which the urethane could be reconstituted in
acetone for the in vitro studies. Urethane 8 is a mixture of two
epimers (the cis- and trans- 4-hydroxy-5-phenyl species) and,
since 6 represents a racemic mixture, exists as a pair of
diastereomers. The epimers 8 were resolved in the NMR
spectrum and were found to be in a ratio of ≈5:2 by NMR
integration. 1H NMR (acetone-d6): δ (major epimer) 2.7 (bs,
1H), 2.97 (ddd, 1H, J ) 5.9, 4.3, 3.8 Hz), 4.26 (dd, 1H, J ) 11.3,
5.9 Hz), 4.42 (dd, 1H, J ) 11.3, 3.8 Hz), 5.02 (d, 1H, J ) 4.3
Hz), 6.7 (bs, 1H), 7.3 (m, 5H); (minor epimer) 2.7 (bs, 1H), 3.23
(ddd, 1H, J ) 10.7, 4.3, 3.2 Hz), 4.16 (ddd, 1H, J ) 12.4, 4.3,
1.6 Hz), 4.68 (dd, 1H, J ) 12.4, 10.7 Hz), 4.93 (dd, 1H, J ) 3.2,
1.6 Hz), 6.7 (bs, 1H), 7.2 (m, 5H). 13C NMR (1:1 D2O/acetone-
d6): δ 44.0, 68.2, 77.9, 128.3, 128.8, 129.7, 136.5, 156.1.
Cell Cu ltu r e. Chinese hamster fibroblasts designated HA1
(17-19) were cultured in RPMI medium (Life Technologies,
Gaithersburg, MD) supplemented with 10% (v/v) fetal bovine
serum (FBS) and 2 mM L-glutamine in an atmosphere of 95%
air and 5% CO2 at 37 °C. Cell cultures were routinely passed
twice a week and monitored for the absence of mycoplasma
contamination.
Cytotoxicity Assa ys. HA1 cells were plated at a density of
7 × 103 cells/well in 96-well flat-bottom microtitration plates.
Cells were incubated at 37 °C for 24 h. Immediately prior to
application of drug solutions to the cells, the average cell density
was determined (from control wells) by treatment with trypsin
to provide a single cell suspension which was enumerated using
a hemacytometer. Drug stock solutions were made up in
acetone and diluted into MEM (Life Technologies) just prior to
application. The cells were treated with various concentrations
of drug and returned to the incubator. The final concentration
of acetone was maintained at a constant 0.2%, which was found
to have no effect on cell survival. After 3 h, the drug solutions
were removed and the wells washed with RPMI media followed
by the addition of fresh 10% FBS with no drug or acetone. The
plates were returned to the incubator for 24 h, after which the
MTS assay was performed as outlined by the supplier (Promega,
Madison, WI). Briefly, a solution of 3-(4,5-dimethylthiazol-2-
yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazo-
lium inner salt (MTS) and phenazine methosulfate (PMS) (final
concentrations: 330 µg/mL MTS and 25 µM PMS) in MEM was
added to each well. After application of the dye solution, the
plates were returned to the incubator for 60 min and then the
absorbance was read on a Titerteck multiscan plate reader at
492 nm. In the presence of PMS, MTS is converted to a water-
soluble formazan by metabolically active cells (20). The con-
centration of metabolite resulting in a 50% inhibition of cell
growth (GI50) were determined graphically from survival curves
plotting cell survival vs concentration. The concentrations
indicated are normalized to a cell density of 2 × 104 cells/well
at the time of drug application.
Kin etic a n d P r od u ct Stu d ies of th e Decom p osition of
Ca r ba m oyl Ald eh yd e (6). Carbamoyl aldehyde (6; 2.5 mM)
was added at 37 °C to 50 mM KPO4 buffer controlled at various
pHs (Figure 4). Aliquots were removed from the incubation at
appropriate time points, and the remaining concentration of 6
was determined by HPLC (1 mL/min 45% A/55% B) using the
diethyl acetal 15 as an internal standard. The half-life (t1/2
)
was determined using a first-order approximation by plotting
the natural logarithm of 6 concentration (ln [6]) vs time to reveal
a slope equal to -k. The t1/2 is equal to 0.693/k. The values
shown in Figure 4 represent the mean ( standard deviation of
three replicate determinations at each pH.
The formation of 7 and 8 from the decomposition of 6 at
various pHs was determined using the HPLC protocol described
above and could be shown to constitute 100% of the products.
The percentage of atropaldehyde 7 formed during the decom-
position of 6 presented in Figure 4 represents the mean ( the
standard deviation of three replicate determinations at each pH.
F or m a tion of a n Atr op a ld eh yd e-Glu ta th ion e Con ju -
ga te. Atropaldehyde (400 µmol) in acetone (5 mL) was added

1228 Chem. Res. Toxicol., Vol. 9, No. 8, 1996
Communications
Ta ble 1. Con cen tr a tion s of F elba m a te (1) a n d Its
Meta bolites a n d An a logs Requ ir ed for th e In h ibition of
) 53 ( 8 µM). This observation presumably reflects the
formation of the urethane 8, which may not fully convert
to atropaldehyde 7 during the 3 h time course of the
cellular exposure to the agent. Alternatively, the depres-
sion of cytotoxicity of 6 relative to 7 could be a conse-
quence of the cellular biotransformation of 6 to the
nontoxic carboxylic acid 5. It is additionally possible that
the urethane 8 exhibits independent metabolic and/or
chemical processing in cells, that was not observed in
vitro (see below).
th e Gr ow th of Cu ltu r ed HA1 Cells by 50% (GI50
)
compound
GI50 (µM)
felbamate (1)
>500
2-phenyl-1,3-propanediol monocarbamate (4)
3-carbamoyl-2-phenylpropionic acid (5)
3-carbamoyl-2-phenylpropionaldehyde (6)
atropaldehyde (7)
acrolein
4-hydroxy-2-nonenal
>500
>500
53 ( 8
4.1 ( 1.1
4.9 ( 0.9
5.2 ( 1.2
423 ( 57
2-phenylpropionaldehyde
The formation of urethane 8 has important implica-
tions for the toxicology of 6. It should be noted that
urethane 8 exists as mixture of two epimers (≈5:2) of two
enantiomers. Although the epimers appear to rapidly
equilibrate in vitro, it is possible that these isomers have
independent pharmacological profiles. We have been
unable to isolate urethane 8 in sufficient purity (i.e., with
no atropaldehyde contamination) for evaluation of its
cytotoxicity; however, we anticipate that urethane 8
would exhibit significantly depressed toxicity relative to
atropaldehyde 7. As is noted above and illustrated in
Figure 2, urethane 8 may act as a stable “reservoir” of
3-carbamoyl-2-phenylpropionaldehyde (6), enabling dif-
fusion to sites distant from its site of production, equili-
bration with 6, and subsequent â-elimination to the toxic
electrophilic aldehyde 7. An additional possibility is that
urethane 8 may undergo biooxidation to the correspond-
ing dioxo derivative 10 (see Figure 2). If produced,
tetrahydro-1,3-oxazine-2,4-dione (10) could have inde-
pendent toxicological properties. For example, 10 re-
sembles thalidomide and its analogs, which have under-
gone extensive evaluation for their ability to modulate
the production of TNFR (21). Moreover, dioxo species 10
could be a/the precursor for the observed carboxylic acid
5 through selective hydrolysis.
Resu lts a n d Discu ssion
Our data establish that 6, a presumed intermediate
in the metabolic disposition of felbamate, exhibits excep-
tional intrinsic chemical reactivity. In addition to its
biotransformation into propionic acid 5, 6 undergoes two
facile and highly pH dependent unimolecular reaction
pathways: â-elimination to generate atropaldehyde (7)
and cyclization to form urethane (8) (Figures 2 and 4).
The decomposition of 6 appears to require the presence
of a protic solvent and to occur faster at higher pHs in
aqueous solution (Figure 4). Interestingly, the distribu-
tion of initial products 7 (9 ( 2%) and 8 (91 ( 2%) is
essentially independent of pH in phosphate buffer (Figure
4). As noted below and illustrated in Figure 2, the
generation of 7 is presumably an irreversible chemical
transformation, whereas the formation of the urethane
(8) is reversible, protecting against the immediate elimi-
nation of 6 to the highly toxic atropaldehyde (7). Thus,
urethane (8) may provide a “reservoir” from which the
reactive propionaldehyde 6 is released for ultimate
oxidation to the acid 5 or elimination to the unsaturated
aldehyde 7. Such a stable source of aldehyde 6 may
enable the distribution of felbamate-derived reactive
species to sites distant from the generation of 6. There-
fore, the modulation of the elimination and cyclization
modes of propionaldehyde 6 decomposition in vivo may
have important toxicological implications.
In conclusion, 3-carbamoyl-2-phenylpropionaldehyde
(6) is a likely intermediate in the metabolism of felbamate
(1) to its principal metabolite, propionic acid 5. We have
found 6 to be highly unstable at physiological pH (t1/2
e
30 s at pH ) 7.4) and to undergo facile â-elimination to
atropaldehyde 7 and reversible cyclization to generate
urethane 8sa species with a considerably longer half-
life at physiological pH (t1/2 g 5 h) that may serve as a
stable source of the reactive aldehyde both in vitro and
in vivo. Atropaldehyde (7) is a potent electrophile and
was found to exhibit significant cytotoxicity to cultured
fibroblasts. It is not clear how the toxicity of atropalde-
hyde might manifest itself in vivo. Atropaldehyde (7)
may act directly by inactivating crucial enzymes, ulti-
mately leading to cell death. Alternatively, the toxicity
of 7 may be a consequence of an immune response to
atropaldehyde-protein adducts, similar to that seen for
the anesthetic halothane (22). However the toxicity is
manifested, it seems likely that atropaldehyde plays a
prominent role. Thus, it is imperative for a thorough
understanding of the toxicology of felbamate that the
chemistry and pharmacology of propionaldehyde (6),
atropaldehyde (7), and urethane (8) be further explored.
Atropaldehyde 7 is highly electrophilic and was shown
to undergo rapid conjugate addition with GSH (t1/2 ≈ 1
min). The GSH-atropaldehyde adduct 9 was obtained
in in vitro incubations of 6 in phosphate buffer in the
presence of GSH, supporting our finding that 6 undergoes
elimination. Under the conditions employed for the
incubation, the t1/2 of 6 was too short to determine if GSH
catalyzed the decomposition of 6. However, GSH did not
appear to significantly alter the initial ratio of cyclization
to â-elimination, since the ratio of “kinetic” products
generated from the decomposition of 6 in the presence of
GSH (8/9 ≈ 8/1) was similar to the ratio of cyclization to
elimination (8/7 ≈ 10/1) observed in the absence of GSH
(see Figure 4). Significantly, with time (≈15 h at pH )
7.4) the urethane 8 was completely converted to the GSH
adduct 9, presumably through equilibration with 6, which
underwent elimination to 7 and reacted with GSH. Thus,
GSH and related biological nucleophiles and/or peptides
may either selectively catalyze the â-elimination pathway
or perturb the equilibrium between 6 and 8 by Le
Chatelier’s principle.
Ack n ow led gm en t. The authors wish to acknowledge
the support of the NIH through Grant NS 64378.
Atropaldehyde 7 exhibited cytotoxicity in our assay
comparable to the well-known R,â-unsaturated aldehyde
toxins, acrolein and hydroxynonenal (Table 1). However,
6, although 10-fold more toxic than its des-carbamoyl
counterpart, 2-phenylpropionaldehyde, was less toxic
than atropaldehyde 7 (7, GI50 ) 4.1 ( 1.1 µM, vs 6, GI50
Refer en ces
(1) The Felbamate Study Group in Lennox-Gastaut Syndrome. (1993)
Efficacy of felbamate in childhood epileptic encephalopathy
(Lennox-Gastaut syndrome). N. Engl. J . Med. 328, 29-33.

Communications
Chem. Res. Toxicol., Vol. 9, No. 8, 1996 1229
(2) Burdette, D. E., and Sackellares, J . C. (1994) Felbamate phar-
macology and use in epilepsy. Clin. Neuropharmacol. 17, 389-
402.
(3) Fraser, A. D. (1996) New drugs for the treatment of epilepsy. Clin.
Biochem. 29, 97-110.
(4) Leppik, I. E. (1996) Felbamate. In The Treatment of Epilepsy
(Shorvon, S., Dreifuss, F., Fish, D., and Thomas, D., Eds.) pp 421-
428, Blackwell Science, New York.
(13) Yang, J . T., Adusumalli, V. E., Wong, K. K., Kucharczyk, N., and
Sofia, R. D. (1991) Felbamate metabolism in the rat, rabbit, and
dog. Drug Metab. Dispos. 19, 1126-1134.
(14) Adusumalli, V. E., Choi, Y. M., Romanyshyn, L. A., Sparadoski,
R. E., Wichmann, J . K., Wong K. K., Kucharczyk, N., and Sofia,
R. D. (1993) Isolation and identification of 3-carbamoyloxy-2-
phenylpropionic acid as a major human urinary metabolite of
felbamate. Drug Metab. Dispos. 21, 710-716.
(5) Swinyard, E. A., Sofia, R. D., and Kupferberg, H. J . (1986)
Comparative anticonvulsant activity and neurotoxicity of felbam-
ate and four prototype antiepileptic drugs in mice and rats.
Epilepsia 27, 27-34.
(6) Sachdeo, R., Kramer, L. D., Rosenberg A., and Sachdeo, S. (1992)
Felbamate monotherapy: controlled trial in patients with partial
onset seizures. Ann. Neurol. 32, 386-392.
(7) Bourgeois, B., Leppik, I. E., Sackellares, J . C., Laxer, K., Lesser,
R., Messenheimer, J . A., Kramer, L. D., Kamin, M., and Rosen-
berg, A. (1993) Felbamate: a double-blind controlled trial in
patients undergoing presurgical evaluation of partial seizures.
Neurology 43, 693-696.
(8) Nightingale, S. L. (1994) From the food and drug administration.
Recommendation to immediately withdraw patients from treat-
ment with felbamate. J AMA, J . Am. Med. Assoc. 272, 995.
(9) Pennell, P. B., Ogaily, M. S., and Macdonald, R. L. (1995) Aplastic
(15) Swinyard, E. A., Sofia, R. D., and Kupferberg, H. J . (1986)
Comparative anticonvulsant activity and neurotoxicity of felbam-
ate and four prototype antiepileptic drugs in mice and rats.
Epilepsia 27, 27-34.
(16) Crossland, I. (1981) Atropaldehyde. Org. Synth. 60, 6-10.
(17) Yang, S. J ., Hahn, G. M., and Bagshaw, M. A. (1966) Chromosome
aberrations induced by thymidine. Cell Res. 42, 130-135.
(18) Kinter, M. T., and Roberts, R. J . (1996) Glutathione consumption
and glutathione peroxidase inactivation by 4-hydroxy-2-nonenal
in cultured fibroblasts. Free Radical Biol. Med. 21, 457-462.
(19) Spitz, D. R., Kinter, M. T., and Roberts, R. J . (1995) Contribution
of increased glutathione content to mechansims of oxidative stress
resistance in hydrogen peroxide resistant hamster fibroblasts. J .
Cell. Phys. 165, 600-609.
(20) Cory, A. H., Owen, T. C., Barltrop, J . A., and Cory, J . G. (1991)
Use of an aqueous soluble tetrazolium/formazan assay for cell
growth assay in culture. Cancer Commun. 3, 207-212.
(21) Niwayama, S., Turk, B. E., and Liu, J . O. (1996) Potent inhibition
of tumor necrosis factor-R production by tetrafluorothalidomide
and tetrafluorophthalimides. J . Med. Chem. 39, 3044-3045.
(22) Pohl, L. R., Kenna, J . G., Satoh, H., Christ, D. D., and Martin, J .
L. (1989) Neoantigens associated with halothane hepatitis. Drug
Metab. Rev. 20, 203-217.
anemia in
a patient receiving felbamate for complex partial
seizures. Neurology 45, 456-460.
(10) Young, N. S., and Alter, B. P. (1994) Aplastic anemia: acquired
and inherited, W. B. Saunders, Philadelphia.
(11) Brodie, M. J ., and Pellock, J . M. (1995) Taming the brain
storms: felbamate updated. Lancet 346, 918-919.
(12) O’Neil, M. G., Perdun, C. S., Wilson, M. D., McGown, S. T., and
Patel, S. (1996) Felbamate-associated fatal acute hepatic necrosis.
Neurology 46, 1457-1459.
TX9601566