Identification of modified atropaldehyde mercapturic acids in rat and human urine after felbamate administration

Article abstract of DOI:10.1021/tx960205e

3-Carbamoyl-2-phenylpropionaldehyde has recently been proposed [Thompson et al. (1996) Chem. Res. Toxicol. 9, 1225-1229] as a potential reactive metabolite of the anti-epileptic drug felbamate. This aldehyde was found to undergo rapid elimination to generate 2-phenylpropenal and reversible cyclization to generate 4-hydroxy-5-phenyltetrahydro-1,3-oxazin-2-one at physiological pH. 2-Phenylpropenal, an α,β-unsaturated aldehyde commonly termed atropaldehyde, is a potent electrophile and undergoes rapid conjugation with glutathione. We sought to demonstrate the formation of atropaldehyde in vivo through the identification of mercapturic acids in rat and human urine after felbamate administration. In this paper, we describe the identification of both the reduced (N-acetyl-S-(2-phenylpropan-3-ol)-L- cysteine) and oxidized (N-acetyl-S-(2-phenyl-3-propanoic acid)-L-cysteine) mercapturic acids of atropaldehyde in rat and human urine. The reduced species was the more abundant in human (~2:1) and rat (~6:1) urine. These findings establish the possibility that atropaldehyde is formed from felbamate in vivo, undergoes glutathione conjugation, and is ultimately excreted in urine in the form of mercapturic acids. Thus, the proposed pathway of felbamate biotransformation, if confirmed in patients, could contribute to our understanding of the toxicities observed during felbamate treatment.

Full text of DOI:10.1021/tx960205e

Chem. Res. Toxicol. 1997, 10, 457-462
457
Id en tifica tion of Mod ified Atr op a ld eh yd e Mer ca p tu r ic
Acid s in Ra t a n d Hu m a n Ur in e a fter F elba m a te
Ad m in istr a tion
Charles D. Thompson, Pamela H. Gulden, and Timothy L. Macdonald*
Chemistry Department, University of Virginia, Charlottesville, Virginia 22901
Received December 16, 1996X
3-Carbamoyl-2-phenylpropionaldehyde has recently been proposed [Thompson et al. (1996)
Chem. Res. Toxicol. 9, 1225-1229] as a potential reactive metabolite of the anti-epileptic drug
felbamate. This aldehyde was found to undergo rapid elimination to generate 2-phenylpropenal
and reversible cyclization to generate 4-hydroxy-5-phenyltetrahydro-1,3-oxazin-2-one at physi-
ological pH. 2-Phenylpropenal, an R,â-unsaturated aldehyde commonly termed atropaldehyde,
is a potent electrophile and undergoes rapid conjugation with glutathione. We sought to
demonstrate the formation of atropaldehyde in vivo through the identification of mercapturic
acids in rat and human urine after felbamate administration. In this paper, we describe the
identification of both the reduced (N-acetyl-S-(2-phenylpropan-3-ol)-L-cysteine) and oxidized
(N-acetyl-S-(2-phenyl-3-propanoic acid)-L-cysteine) mercapturic acids of atropaldehyde in rat
and human urine. The reduced species was the more abundant in human (∼2:1) and rat (∼6:
1) urine. These findings establish the possibility that atropaldehyde is formed from felbamate
in vivo, undergoes glutathione conjugation, and is ultimately excreted in urine in the form of
mercapturic acids. Thus, the proposed pathway of felbamate biotransformation, if confirmed
in patients, could contribute to our understanding of the toxicities observed during 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 and demon-
strated an excellent therapeutic index throughout its
clinical trials (1-3). However, subsequent to its release,
several toxicological issues, namely, aplastic anemia and
hepatotoxicity, were associated with felbamate use (4, 5).
The apparent idiosyncratic nature and severity of these
side effects prompted a recommendation by the U.S. Food
and Drug Administration to withdraw patients from
felbamate therapy unless the benefit of seizure control
outweighed the risk of the reported toxicities (6). Re-
cently, we proposed 3-carbamoyl-2-phenylpropionalde-
hyde 3 as a probable reactive metabolite of felbamate (7).
We demonstrated the rapid formation of 2-phenylprope-
nal 7, commonly termed atropaldehyde, from 3 in vitro
and proposed this R,â-unsaturated aldehyde as a reactive
metabolite of felbamate that may play a role in the
observed toxicities.
The in vivo metabolism of felbamate has been studied
in several species including rat, rabbit, dog, and human
(8, 9). The primary excreted compound is unchanged
felbamate. In addition, four metabolites have been
identified. 2-Phenyl-1,3-propanediol monocarbamate 2
is the product of hydrolysis of one of the carbamate
moieties and has been observed in all of the species.
3-Carbamoyl-2-phenylpropionic acid 6, an oxidized sec-
ondary metabolite from 2, has been identified as a major
metabolite in humans [reported to be ∼12% of a dose in
4-8 h post-dose urine (9)] and also observed in dogs.
F igu r e 1. Metabolic transformation of felbamate 1 to mercap-
turic acids 9-11.
Felbamate represents a pro-chiral substrate for the
enzymes that produce the observed metabolites, and
presumably, these metabolites are chiral although their
exact stereochemistry has not been reported.
The mechanism for formation of 6 has not been
determined. This acid may result from the direct oxida-
* To whom correspondence should be addressed: e-mail:
tlm@virginia.edu.
X Abstract published in Advance ACS Abstracts, March 15, 1997.
S0893-228x(96)00205-6 CCC: $14.00 © 1997 American Chemical Society

458 Chem. Res. Toxicol., Vol. 10, No. 4, 1997
Thompson et al.
purified by chromatography on silica gel (chloroform/acetone)
to give a 63% yield with starting material and dicarbamate
accounting for the remainder. 1H NMR (DMSO-d6): δ 2.4 (bs,
1H, -OH); 3.1 (quin, 1H, J ) 6.2, benzylic); 3.84 (d, 2H, J ) 6.2,
-CH2OH); 4.40 (d, 2H, J ) 6.2, -CH2OCONH2); 4.8 (bs, 2H,
-NH2); 7.3 (m, 5H, Ph).
2-P h en yl-1,3-p r op a n ed iol Dica r ba m a te (1). Felbamate
used for in vivo metabolism studies was purchased from the
University of Virginia Medical Center Pharmacy and manufac-
tured by Wallace Laboratories. Formation of felbamate for use
as a standard for HPLC and LC/MS was achieved by treatment
of the isolated monocarbamate 2 with another equivalent of 1,1′-
carbonyldiimidazole, followed by ammonia as described above.
1H NMR (DMSO-d6); δ 3.16 (quin, 1H, J ) 6.3, benzylic); 4.1
(d, 4H, J ) 6.3, -CH2OCONH2); 6.4 (bs, 4H, -NH2); 7.3 (s, 5H,
Ph).
N-Acetyl-S-(2-p h en ylp r op a n a l)-L-cystein e (9). To 10 mg
of atropaldehyde in methanol (2 mL) was added 16 mg of
N-acetyl-L-cysteine (1.3 equiv) with stirring at room tempera-
ture. Two drops of 200 mM KPO4 buffer (pH ) 8) were added
for catalysis, and the reaction was stirred for 1 h. The solvent
was removed by vacuum evaporation and reconstituted in 700
µL of 25% C/75% D. This solution was applied to a C18 solid
phase extraction cartridge that was then washed with 3 mL of
100% D. The purified atropaldehyde mercapturic acid was then
eluted with 3 mL 50% C in 50% D. This fraction was then
quickly frozen and lyophilized to dryness in efforts to limit
decomposition as this conjugate was found to decompose under
acidic conditions. 1H NMR (D2O): δ 1.92 (bs, 3H, -COCH3);
2.64-3.0 (m, 4H, -CH2SCH2-); 3.05 (m, ∼0.8 H, benzylic); 3.89
(m, ∼0.2 H, benzylic); 4.42 (m, 1H, cysteine R); 5.14 (d, ∼0.8 H,
J ) 5.0, -CH(OH)2) 7.28 (m, 5H, Ph); 9.58 (s, ∼0.2 H, -COH).
LC/MS: MH+ ) 296. The peaks at 9.58 and 3.89 correspond to
the aldeydic and benzylic protons in the aldehyde state, while
the peaks at 5.14 and 3.05 correspond to the aldehydic and
benzylic protons in the gem diol state.
N-Acetyl-S-(2-p h en ylp r op a n -3-ol)-L-cystein e (10). To a
mixture of the crude atropaldehyde mercapturic acid 9, formed
as above, was added an excess of NaBH4. This solution was
stirred for 1 h at room temperature followed by quenching with
1 mL of 0.2% acetic acid. The solvent was reduced by vacuum
evaporation and then reconstituted to a total volume of 1 mL
with 0.2% acetic acid. HPLC analysis (30% A/70% B) indicated
complete conversion of 9. The reduced conjugate was then
purified by solid phase extraction as described above. 1H NMR
(D2O): δ 1.91 (bs, 3H, -COCH3); 2.66-3.0 (m, 6H, -CH2SCH2-
CHPhCH2OH); 3.72 (t, 1H, J ) 5.5, benzylic); 4.39 (m, 1H,
cysteine R); 7.27 (m, 5H, Ph). LC/MS: MH+ ) 298.
N-Acetyl-S-(2-p h en ylp r op a n oic a cid )-L-cystein e (11). To
9 mg of atropic acid in methanol (5 mL) was added 18 mg of
N-acetyl-L-cysteine (NaC) (1.3 equiv) with stirring at room
temperature. The addition of several drops of 200 mM KPO4
buffer (pH ) 8.0) resulted in a cloudy suspension that was
stirred over night. The solvent was removed by vacuum
evaporation, and the residue was reconstituted in 1 mL of 100%
D for purification by solid phase extraction as described above.
1H NMR (D2O): δ 1.94 (bs, 3H, -COCH3); 2.75-3.23 (m, 4H,
-CH2SCH2-); 3.84 (t, 1H, J ) 5.7, benzylic); 4.41 (m, 1H, cysteine
R); 7.32 (m, 5H, Ph). LC/MS: MH+ ) 312.
Ra t in Vivo Meta bolism . Six adult male Sprague-Dawley
rats (∼250 mg) were housed in individual metabolic cages under
a 12-h light/dark cycle and provided with free access to food
and water throughout the experiment. Urine was collected in
plasticware cups placed beneath the cages at room temperature.
The rats were allowed to acclimate for 36 h, after which time
the background urine (designated RU-A) was collected, pooled,
and stored frozen at -60 °C. Each rat was administered 800
mg/kg felbamate (formulation, Felbatol oral suspension, 600 mg/
mL, lot no. 5LO6Q) via gavage and returned to their individual
cages. Urine from 1-18 h post-dose (designated RU-B) was
collected, pooled, and stored frozen at -60 °C.
tion of 3 by an aldehyde dehydrogenase. Alternatively,
6 may be produced from the selective hydrolysis of
tetrahydro-1,3-oxazin-2,4-dione 5, the presumed product
from oxidation of 4-hydroxy-5-phenyl-tetrahydro-1,3-ox-
azin-2-one 4. The cyclic carbamate 4 is the predominant
product formed, in a reversible process, from 3 under
physiological pH in vitro but has not yet been observed
in vivo. Regardless of the mechanism leading to forma-
tion of acid 6, it is clear that hydrolysis of felbamate to
the alcohol 2 comprises the first step in the primary
pathway of felbamate metabolism in humans.
We sought to demonstrate the formation of atropalde-
hyde from felbamate in vivo. If formed, atropaldehyde
would be expected to react with biological nucleophiles,
such as GSH. We have demonstrated the formation of
atropaldehyde-GSH conjugate 8 in the non-enzyme-
catalyzed reaction of atropaldehyde and GSH in vitro (7).
GSH conjugates are known to be processed in vivo
through a multi-step, enzyme-catalyzed process to N-
acetyl-L-cysteine conjugates, commonly termed mercap-
turic acids (10). These mercapturic acids are excreted
in urine. Herein, we report the identification by LC/MS
of mercapturic acids 10 and 11 in rat and human urine
after felbamate administration.
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 (at 214 nm) using a 5
µm microsorb-mv C18 (4.6 mm × 250 mm) column (Rainin,
Emeryville, CA) and a binary solvent systemsSolvent A: 95%
acetonitrile, 4.93% water, and 0.07% trifluoroacetic acid; solvent
B: 0.1% trifluoroacetic acid and 99.9% water. Solid phase
extraction was accomplished using C18 solid phase extraction
cartridges (Alltech, Deerfield, IL) and a binary solvent systems
Solvent C: 100% acetonitrile; solvent D: 0.2% acetic acid. 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. 2-P h en ylp r op en a l (7). The synthesis of 7 was
accomplished as described (7).
2-P h en yl Acr ylic Acid . Oxidation of 7 by the method of
Dalcanale and Montanari (11) afforded 2-phenyl acrylic acid,
commonly known as atropic acid. Atropaldehyde (0.3 mmol) was
stirred in acetonitrile (5 mL) at 10 °C. H2O2 (100 µL 35%) and
0.6 M NaH2PO4 (2 mL, pH ) 2.0 adjusted with HCl) were added
followed by 700 µL of 1 M NaClO2 in 50 µL aliquots at 5-min
intervals. After 1 hr, the reaction was quenched by the addition
of sodium bisulfite and 3 M HCl (2 mL). Acid-base extraction
with CHCl3 afforded the pure 2-phenyl acrylic acid in 70% yield.
1H NMR (acetone-d6): δ 5.88 (s, 1H, HCdCR2); 6.25 (s, 1H,
HCdCR2); 7.3 (m, 5H, Ph); 13.5 (bs, 1H, -CO2H).
2-P h en yl-1,3-pr opan ediol. Reduction of commercially avail-
able diethyl phenylmalonate to 2-phenyl-1,3-propanediol with
lithium aluminum hydride was achieved using established
methods (12).
2-P h en yl-1,3-p r op a n ed iol Mon oca r ba m a te (2). The syn-
thesis of 2 was achieved by treatment of the diol (2.6 mmol)
with 1,1′-carbonyldiimidazole (1 equiv) in THF at room tem-
perature for 1 h. The temperature of 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 approximately 30 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 and allowed to stir for 1 h. The monocarbamate 2 was
Hu m a n in Vivo Meta bolism . A healthy adult male volun-
teer was administered a single dose of 600 mg (∼9 mg/kg)
felbamate (formulation, Felbatol 600 mg tablet lot no. 5EO5N).
Pre-dose urine (designated HU-A) had been collected, pooled,

Atropaldehyde Mercapturic Acids from Felbamate
Chem. Res. Toxicol., Vol. 10, No. 4, 1997 459
F igu r e 2. Schematic of the preparation of urine samples (see
Experimental Procedures for details).
and stored frozen at -60 °C the previous day. The 0-8 h post-
dose urine (designated HU-B) was collected and stored frozen
at -60 °C.
P r ep a r a tion of Ur in e Sa m p les. (See Figure 2.) Urine
samples were removed from the freezer and allowed to thaw at
room temperature after which they were centrifuged (4000g, 10
min) to remove precipitated material. The supernatant was
decanted away, and the precipitate was discarded. Urine (3 mL)
was combined with 1 mL of 4% acetic acid and placed on ice for
5 min followed by centrifugation at 4000g to remove the
precipitated material. The supernatant was lyophilized to near
dryness and reconstituted to a total volume of 1 mL with 0.2%
acetic acid. The samples were then fractionated using a C18
solid phase extraction cartridge. The cartridges were individu-
ally prepared just prior to use by the sequential elution of 3
mL of 90% C/10% D and 6 mL of 100% D. The sample was then
applied, and the cartridge was washed with 3 mL of 100% D
(fraction 1) followed by 4 mL of 10% C/90% D (fraction 2). The
samples for further analysis were then eluted with 3 mL of 30%
C/70% D (fraction 3). Fraction 3 samples were lyophilized to
F igu r e 3. Comparison of (A) pre-dose rat urine (RU-A-F3) to
(B) post-dose rat urine (RU-B-F3). The chromatograms show
percent relative abundance of ion current versus m/z value for
summed mass spectra over the elution range of felbamate (m/z
) 239) and the mercapturic acids (m/z ) 298 and 312).
of the parent ion, was scanned at a rate of 500 Da/s, allowing
the individual fragments to pass through to the detector.
Resu lts
Synthetic standards of the mercapturic acids 9-11
were prepared as described in Experimental Procedures
1
and their structures confirmed by H NMR and LC/MS.
These standards were then used to develop the protocol
for sample preparation (Figure 2). This protocol was
optimized for the analysis of these mercapturic acids in
the presence of felbamate. However, the relative recov-
eries of the other metabolites and the parent drug,
relative to the mercapturic acids, have not been deter-
mined.
The mass spectra shown in Figures 3 and 4 represent
the summed mass spectra recorded during the period
when the mercapturic acids and felbamate elute from the
microcapillary column into the mass spectrometer, as
determined by their relative position to an internal
standard. Figure 3 compares pre-dose rat urine (RU-A-
F3) in panel A to pooled 1-18 h rat urine after a single
600 mg/kg administration of felbamate by gavage in
panel B. The parent molecular ion peaks for felbamate
1 (m/z ) 239), the NaC-alcohol 10 (m/z ) 298), and the
NaC-acid 11 (m/z ) 312) can be observed in the post-
dose sample. Similarly, in Figure 4, pre-dose human
urine (HU-A-F3, panel A) is compared to 0-8 h human
urine (HU-B-F3, panel B) after a single 600 mg (∼9 mg/
kg) oral dose of felbamate. Again, the peaks correspond-
ing to the appearance of 1, 10, and 11 can be observed
dryness and then reconstituted in 1 mL of 10% C/90% D.
A
1:10 dilution into 0.2% acetic acid afforded the samples for LC/
MS analysis.
Mass Spectr om etr y An alysis of P r epar ed Ur in e Sam ples.
Samples were analyzed by tandem mass spectrometry as
described previously (13). Samples were loaded into 75 µm i.d.
× 185 µm o.d. fused silica tubing (SGE, Austin, TX), packed
with 12 cm of 10 µm C18 beads (YMC Gel, Kyoto, J apan). The
samples were eluted into a TSQ 70 (upgraded with the electron-
ics of a TSQ 700) triple quadrupole mass spectrometer (Finni-
gan-MAT, San J ose, CA) equipped with a Finnigan API ESI/CI
source, at a flow rate of 0.5 µL/min with a 12-min gradient of
0-80% acetonitrile in 0.1 M acetic acid. For molecular weight
data, spectra were recorded every 1.5 s over a mass range of
190-1300 amu. The resulting spectrum contained ions of the
type [M + nH]n+, where M is molecular mass and n is the
number of attached protons. Fragmentation data were recorded
similarly, where the first quadrupole mass filter was set to pass
all ions within a 3-Da window centered around the mass of
interest. These ions were transmitted to the second quadrupole
(collision cell), where they fragment upon collision with argon
atoms. These fragments were then analyzed with the third
quadrupole. A mass range, from 50 to 5 amu above the mass

460 Chem. Res. Toxicol., Vol. 10, No. 4, 1997
Thompson et al.
in Figure 4B. We were unable to identify the appearance
of a peak at m/z ) 296 corresponding to the mercapturic
acid of atropaldehyde 9 in these samples. However, our
experience with synthetic NaC-atropaldehyde 9 has
demonstrated that it is unstable in a biological matrix,
such as urine (data not shown).
In order to demonstrate that the peaks corresponding
to the m/z of 10 and 11 in post-dose samples were in fact
the modified mercapturic acids of atropaldehyde, we
conducted co-elution studies with synthetic standards
(Figure 5). Figure 5A shows the percent relative abun-
dance versus scan number of 1, the putative 10, and the
putative 11 from the post-dose human urine sample HU-
B-F3. As indicated by the percent relative abundance
value, the putative 10 was approximately twice as
abundant at the putative 11; whereas, in post-dose rat
urine the ratio of 10 to 11 is approximately 6:1 (data not
shown). Figure 5B shows the elution profile of the same
sample with the synthetic NaC-alcohol 10 doped into the
sample. As indicated in the middle panel of Figure 5B,
the synthetic standard and the putative 10 from the urine
sample co-elute, giving rise to an increase in the percent
relative abundance. This demonstrates that this analyte
in the urine sample and the synthetic standard of 10 have
identical m/z and elution characteristics.
The corresponding experiment was done for the NaC-
acid 11 as shown in Figure 5C. In this case, the bottom
panel demonstrates that the addition of a synthetic
standard of 11 to the urine sample results in the elution
of a single peak with a corresponding increase in its
percent relative abundance.
In order to further demonstrate that the urine analytes
in question unambiguously corresponded to the modified
mercapturic acids of atropaldehyde 10 and 11, we col-
lected collision-activated dissociation spectra for the
analytes in a urine sample (Figures 6B and 7B) and
compared them to the CAD spectra of the synthetic
standards (Figures 6A and 7A). Figure 6 compares the
F igu r e 4. Comparison of (A) pre-dose human urine (HU-A-
F3) to (B) post-dose human urine (HU-B-F3). The chromato-
grams show percent relative abundance of ion current versus
m/z value for summed mass spectra over the elution range of
felbamate 1 (m/z ) 239) and the mercapturic acids 9 and 11
(m/z ) 298 and 312).
F igu r e 5. Co-elution of mercapturic acids in post-dose human urine (HU-B-F3) with synthetic standards. (A) On-line microcapillary
HPLC elution of HU-B-F3 showing m/z ) 239 (for felbamate 1), m/z ) 298 (for NaC-alcohol 9), and m/z ) 312 (for NaC-acid 11). (B)
On-line microcapillary HPLC elution of HU-B-F3 plus added NaC-alcohol synthetic standard. (C) On-line microcapillary HPLC elution
of HU-B-F3 plus added NaC-acid synthetic standard.

Atropaldehyde Mercapturic Acids from Felbamate
Chem. Res. Toxicol., Vol. 10, No. 4, 1997 461
F igu r e 6. Collision-activated dissociation of NaC-alcohol from
(A) synthetic standard and (B) post-dose human urine (HU-B-
F3).
F igu r e 7. Collision-activated dissociation of NaC-acid from (A)
synthetic standard and (B) post-dose human urine (HU-B-F3).
of 4 has not been demonstrated in vivo, and its metabolic
fate, if formed, is unknown. The “model” for this possible
behavior of 4 is 4-hydroxycyclophosphamide, the meta-
bolically activated form of cyclophosphamide. Our pro-
posed pathway leading from 4 to atropaldehyde 7 par-
allels the known pathway from 4-hydroxycyclophos-
phamide to acrolein (14). The breadth to which this
similarity applies is the subject of ongoing investigation.
We have examined the potential for an alternative
process for the formation of the mercapturates 10 and
11 involving â-elimination of acid carbamate 6 to 2-phe-
nylacrylic acid and subsequent GSH conjugation. How-
ever, we have found 6 to be stable to elimination at
physiological pH (data not shown) and the derived atropic
acid to be slow (relative to atropaldehyde) in undergoing
uncatalyzed GSH conjugation. Moreover, it is unlikely
that alcohol adduct 10 would be derived from the acid
11. A possibility for the formation of mercapturates 10
and 11 that we have not examined is the glutathione
transferase-mediated conversion from the corresponding
metabolites 2 and 6; however, this potential pathway
would also be anticipated a priori to generate the
corresponding mono-GSH and subsequent mercapturate
adduct of felbamate itself.
CAD spectra of the NaC-alcohol 10 standard, panel A,
to that of the putative NaC-alcohol from a human post-
dose urine sample (HU-B-F3), panel B. As shown, these
two entities have identical fragmentation patterns.
The CAD spectra of the NaC-acid 11 standard and the
putative 11 from HU-B-F3 are shown in Figure 7. Again,
the CAD of the standard, panel A, and of the putative
mercapturic acid from the urine sample, panel B, dem-
onstrate identical fragmentation patterns.
Discu ssion
We have identified both the reduced (10) and oxidized
(11) mercapturic acids of atropaldehyde 7 in rat and
human urine after felbamate administration. Their
presence provides support for the hypothesis that atro-
paldehyde 7 is generated during the course of felbamate
metabolism. Analogous to our in vitro studies (7), 7 is
presumably produced in vivo via â-elimination of 3, a
presumed intermediate in the enzymatic oxidation of
alcohol 2 to carboxylic acid 6. We have reported that the
half-life of 3 is very short (<30 s) at physiological pH (7).
While the product distribution for the decomposition of
3 in vitro favors cyclic carbamate 4 over atropaldehyde
7 (≈8:1), the formation of 4 is reversible, enabling the
possibility of reversion to 3 and subsequent elimination
to atropaldehyde 7.
Significantly, our in vitro studies indicate that cyclic
carbamate 4 has a considerably longer half-life relative
to 3 or 7 (hours versus seconds). If 4 exhibited a similarly
long half-life in vivo, then it could migrate to sites distal
to its formation where it could undergo reversion to 3
and subsequent â-elimination. However, the formation
Identification of mercapturic acids 10 and 11 in post-
dose urine is consistent with the hypothesis that atro-
paldehyde is formed in vivo and that it reacts with thiol
nucleophiles. While the mercapturic acids are derived
from addition to GSH, any similar nucleophile (i.e.,
protein thiols or DNA bases) would be expected to
undergo similar conjugation to atropaldehyde. Toxicity
after exposure to this type of alkylating agent might
manifest itself through a variety of pathways. Disruption

462 Chem. Res. Toxicol., Vol. 10, No. 4, 1997
Thompson et al.
light of the severe nature of the toxicities observed for
this important anti-epileptic agent.
of enzyme activity or alkylation of DNA may lead to cell
death through an apoptotic mechanism. Alternatively,
protein alkylation can precipitate an immune response
resulting in T-cell-mediated cell death as has been shown
for halothane and other agents (15, 16).
Ack n ow led gm en t. The authors wish to thank Dr.
W. Kline Bolton for access to metabolic cages for the rat
in vivo studies. Additionally, we wish to thank Dr.
Donald F. Hunt for making the LC/MS analysis possible.
The extent of atropaldehyde formation in vivo will be
dependent on several variables. The formation of atro-
paldehyde is dependent on the hydrolysis of felbamate
to the alcohol carbamate 2. Thus, factors that modulate
this pathway relative to the competing pathways of para-
hydroxylation [resulting in 2-(4-hydroxyphenyl)-1,3-pro-
panediol dicarbamate] and 2-hydroxylation (resulting in
2-phenyl-2-hydroxy-1,3-propanediol dicarbamate) will
impact on the amount of atropaldehyde formed. The
later two hydroxylated species are known human me-
tabolites of felbamate (9) and are presumably products
of P450-mediated metabolism. Consequently, factors
that modulate P450 activity (including enzyme induction
or inhibition) would be expected to modulate the relative
contribution of hydrolysis, resulting in the monocarbam-
ate 2 to the overall disposition of felbamate.
The alcohol 2 is the precursor of the labile aldehyde 3,
which represents an additional critical branch point in
the potential initialization of toxic responses to felbamate.
Aldehyde 3 may undergo oxidation by an unidentified
aldehyde dehydrogenase to the (apparently) innocuous
carboxylic acid 6, reversible cyclization to the cyclic
carbamate 4, or â-elimination to atropaldehyde with
potential toxic consequences. Thus, inhibition of the
enzyme responsible for conversion of 3 to 6 would
increase flux through the â-elimination pathway with
concomitant implications for the toxicity of felbamate.
Conjugation of atropaldehyde with GSH presumably
represents a critical detoxification mechanism. There-
fore, factors that modulate GSH levels, such as co-
administration of other electrophilic or pro-electrophilic
species (i.e., acetaminophen), may impact on the avail-
ability of GSH for detoxification of atropaldehyde and
contribute importantly to the toxicity of felbamate.
However, glutathione conjugates and mercapturic acids
of some other agents are known to be toxic (17, 18).
Importantly, the mercapturic acids of acrolein are thought
to be nephrotoxic (19), and there is the possibility that a
similar potential exists for mercapturic acids of atropal-
dehyde.
The identification of N-acetylcysteine adducts 10 and
11 as significant metabolites of felbamate in humans
implicates atropaldehyde as a potential contributor to the
toxicities observed during felbamate treatment. How-
ever, the apparent idiosyncratic expression of severe
toxicities (i.e., aplastic anemia and hepatotoxicity) sug-
gests that there are other contributing factors. Polymor-
phism and variable expression levels of the enzymes that
modulate the levels of atropaldehyde formation would be
expected to result in inter-individual differences in
atropaldehyde formation. These factors combined with
as yet unidentified environmental contributions that alter
the production or detoxification of the reactive species
associated with felbamate metabolism may result in
individuals that exhibit heightened production of atro-
paldehyde or susceptibility to the toxicity of this reactive
aldehyde. These issues need to be explored further in
Refer en ces
(1) Fraser, A. D. (1996) New drugs for the treatment of epilepsy. Clin.
Biochem. 29, 97-110.
(2) 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.
(3) 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.
(4) Pennell, P. B., Ogaily, M. S., and Macdonald, R. L. (1995) Aplastic
anemia in
a patient receiving felbamate for complex partial
seizures. Neurology 45, 456-460.
(5) 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.
(6) Nightingale, S. L. (1994) From the food and drug administration.
Recommendation to immediately withdraw patients from treat-
ment with felbamate. J . Am. Med. Assoc. 272, 995.
(7) Thompson, C. D., Kinter, M. T., and Macdonald, T. L. (1996)
Synthesis and in vitro reactivity of 3-carbamoyl-2-phenylpropi-
onaldehyde and 2-phenylpropenal: Putative reactive metabolites
of felbamate. Chem. Res. Toxicol. 9, 1225-1229.
(8) 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.
(9) 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.
(10) Vermeulen, N. P. E. (1989) Analysis of mercapturic acids as a
tool in biotransformation, biomonitoring and toxicological studies.
Trends Pharmacol. Sci. 10, 177-181.
(11) Dalcanale, E., and Montanari, F. (1986) Selective oxidation of
aldehydes to carboxylic acids with sodium chlorite-hydrogen
peroxide. J . Org. Chem. 51, 567-569.
(12) Brown, H. C., Narasimhan, S., and Choi, Y. M. (1982) Selective
reductions. 30. Effect of cation and solvent on the reactivity of
saline borohydrides for reduction of carboxylic esters. Improved
procedures for the conversion of esters to alchohols by metal
borohydrides. J . Org. Chem. 47, 4702-4708.
(13) Gulden, P. H., Fischer, P., III, Sherman, N. E., Wang, W.,
Engelhard, V. H., Shabanowitz, J ., Hunt, D. F., and Pamer, E.
G. (1996) A listeria monocytogenes pentapeptide is presented to
cytolytic T lymphocytes by the H2-M3 MHC class 1b molecule.
Immunity 5, 73-79.
(14) Fraiser, L. H., Kanekal, S., Kehrer, J . P. (1991) Cyclophosphamide
toxicity characterising and avoiding the problem. Drugs 42, 781-
795.
(15) Bourdi, M., Chen, W., Peter, R. M., Martin, J . L., Buters, T. M.,
Nelson, S. D., and Pohl, L. R. (1996) Human cytochrome P450
2E1 is a major autoantigen associated with halothane hepatitis.
Chem. Res. Toxicol. 9, 1159-1166.
(16) Uetrecht, J . P. (1996) Drug-induced lupus: possible mechanisms
and their implications for prediction of which new drugs may
induce lupus. Exp. Opin. Invest. Drugs 5, 851-860.
(17) Koga, N., Inskeep, P. B., Harris, T. M., and Guengerich, F. P.
7-Guanyl)ethyl]glutathione, the major DNA adduct
(1986) S-[2-(N
formed from 1,2-dibromoethane. Biochemistry 25, 2192-2198.
(18) Monks, T. J ., Anders, M. W., Dekant, W., Stevens, J . L., Lau, S.
S., van Bladeren, P. J . (1990) Glutathione conjugate mediated
toxicities. Toxicol. Appl. Pharmacol. 106, 1-19.
(19) Ramu, K., Fraiser, L. H., Mamiya, B., Ahmed, T., and Kehrer, J .
P. (1995) Acrolein mercapturates: synthesis, characterization,
and assessment of their role in the bladder toxicity of cyclophos-
phamide. Chem Res. Toxicol. 8, 515-524.
TX960205E