Oxidative metabolism of 1-(2-chloroethyl)-3-alkyl-3-(methylcarbamoyl)triazenes: Formation of chloroacetaldehyde and relevance to biological activity

Article abstract of DOI:10.1021/tx9500639

(Methylcarbamoyl)triazenes have been shown to be effective cancer chemotherapeutic agents in a number of biological systems. Because of their chemical stability, it is likely that their activity in vivo is the result of a metabolic activation process. Previous studies have shown that 1-(2-chloroethyl)-3-methyl-3-(methylcarbamoyl)triazene (CMM) and 1-(2-chloroethyl)-3-benzyl-3-(methylcarbamoyl)triazene (CBzM) are metabolized by rat liver microsomes in the presence of NADPH to yield the ((hydroxymethyl)carbamoyl)triazene analogs of the parent compounds. The present studies show that both compounds are also oxidized at the chloroethyl substituent to yield chloroacetaldehyde and a substituted urea. In the case of CBzM metabolism, 47% of the metabolized parent compound was recovered as benzylmethylurea, 8% was recovered as benzylurea, and 26% was recovered as the ((hydroxymethyl)carbamoyl)-triazene and carbamoyltriazene metabolites. These results suggest that the chloroethyl group is the favored initial site of metabolism. In reaction mixtures containing intitial concentrations of 300 μM. CBzM, 78 μM chloroacetaldehyde was produced, as compared to 58 μM chloroacetaldehyde produced from the metabolism of 300 μM CMM. The formation of chloroacetaldehyde, a known mutagenic DNA alkylating agent, may explain the biological activity of these compounds.

Full text of DOI:10.1021/tx9500639

1
72
Chem. Res. Toxicol. 1996, 9, 172-178
Oxid a tive Meta bolism of
-(2-Ch lor oeth yl)-3-a lk yl-3-(m eth ylca r ba m oyl)tr ia zen es:
F or m a tion of Ch lor oa ceta ld eh yd e a n d Releva n ce to
Biologica l Activity
1
Carol A. Rouzer,*, Michelle Sabourin, Tricia L. Skinner, Erin J . Thompson,
†,‡
†
†
†
†
§
§
§
Thomas O. Wood, Gwendolyn N. Chmurny, J ohn R. Klose, J ohn M. Roman,
Richard H. Smith, J r., and Christopher J . Michejda‡
†,‡
Department of Chemistry, Western Maryland College, Westminster, Maryland 21157,
Molecular Aspects of Drug Design Section, ABL-Basic Research Program, NCI-Frederick Cancer
Research and Development Center, Frederick, Maryland 21702, and Chemical Synthesis and
Analysis Laboratory, SAIC, NCI-Frederick Cancer Research and Development Center,
Frederick, Maryland 21702
Received April 21, 1995X
(Methylcarbamoyl)triazenes have been shown to be effective cancer chemotherapeutic agents
in a number of biological systems. Because of their chemical stability, it is likely that their
activity in vivo is the result of a metabolic activation process. Previous studies have shown
that 1-(2-chloroethyl)-3-methyl-3-(methylcarbamoyl)triazene (CMM) and 1-(2-chloroethyl)-3-
benzyl-3-(methylcarbamoyl)triazene (CBzM) are metabolized by rat liver microsomes in the
presence of NADPH to yield the ((hydroxymethyl)carbamoyl)triazene analogs of the parent
compounds. The present studies show that both compounds are also oxidized at the chloroethyl
substituent to yield chloroacetaldehyde and a substituted urea. In the case of CBzM
metabolism, 47% of the metabolized parent compound was recovered as benzylmethylurea,
8
% was recovered as benzylurea, and 26% was recovered as the ((hydroxymethyl)carbamoyl)-
triazene and carbamoyltriazene metabolites. These results suggest that the chloroethyl group
is the favored initial site of metabolism. In reaction mixtures containing intitial concentrations
of 300 µM CBzM, 78 µM chloroacetaldehyde was produced, as compared to 58 µM chloro-
acetaldehyde produced from the metabolism of 300 µM CMM. The formation of chloroacetal-
dehyde, a known mutagenic DNA alkylating agent, may explain the biological activity of these
compounds.
1
In tr od u ction
The synthesis and chemical decomposition of 1,3-
3-(methylcarbamoyl)triazene (CMM) possessed remark-
ably low toxicity while effecting cures in the cases of the
P388 murine leukemia and the LOX melanoma (3).
These initial studies suggested that the acyltriazenes
might be a promising new class of cancer chemothera-
peutic agents. However, it was soon realized that,
because of their very high chemical stability under
physiological conditions, their biological activity could not
easily be explained by chemical decomposition alone. For
example, the estimated half-life of CMM at pH 7.5 and
37 °C is between 2 and 3 days (4). The possibility was
therefore considered that a metabolic activation step
might lead to more rapid chemical breakdown. Consis-
tent with this hypothesis were findings that acyltriazenes
bearing a carbethoxy substituent demonstrated a marked
increase in their ability to alkylate DNA if porcine liver
esterase was included in the reaction mixtures. The
esterase catalyzed the deacylation of the triazene, yield-
ing a highly unstable dialkyltriazene which then decom-
posed to form an mixture of alkanediazonium ions (6-
dialkyl-3-acyltriazenes have been extensively studied in
our laboratory (1-5). These compounds were originally
designed with the expectation that they would be DNA
alkylating agents through the formation of an alkanedia-
zonium ion intermediate. As predicted, under acidic and
neutral conditions, the acyltriazenes undergo heterolysis
of the N(2)-N(3) bond to yield the expected alkanedia-
zonium ion bearing the substituent originally at N(1) (4).
Consequently, a series of 1,3-dimethyl-3-acyltriazenes
and 1-(2-chloroethyl)-3-methyl-3-acyltriazenes have been
synthesized to act as methylating and chloroethylating
agents, respectively (1, 2). Biological studies revealed
that all of these triazenes possessed cytotoxic activity in
the MTT-microculture tetrazolium assay, with the (2-
chloroethyl)triazenes generally showing higher potency
than the dimethyltriazenes (3). In vivo, several of the
compounds were also found to be active against selected
tumor xenografts implanted in nude mice. Of particular
interest were the findings that 1-(2-chloroethyl)-3-methyl-
8
). The mechanism of action of carbethoxytriazenes in
vivo could therefore easily be explained; however, the
1
Abbreviations: 1-(2-chloroethyl)-3-methyl-3-(methylcarbamoyl)-
*
Address correspondence to this author at the Department of
triazene (CMM), 1-(2-chloroethyl)-3-methyl-3-carbethoxytriazene (CMC),
1-(2-chloroethyl)-3-benzyl-3-(methylcarbamoyl)triazene (CBzM), 1-(2-
chloroethyl)-3-benzyl-3-((hydroxymethyl)carbamoyl)triazene (CBzM-
met1), 1-(2-chloroethyl)-3-benzyl-3-carbamoyltriazene (CBzM-met2),
1-(2-chloroethyl)-3-methyl-3-((hydroxymethyl)carbamoyl)triazene (CMM-
met1), (dinitrophenyl)hydrazone of chloroacetaldehyde (CAA-DNP),
(dinitrophenyl)hydrazone of methoxyacetaldehyde (MAA-DNP).
Chemistry, Western Maryland College, 2 College Hill, Westminster,
MD 21157. Phone: 410-857-2492; FAX: 410-857-2729.
†
Western Maryland College.
ABL, Frederick Cancer Research and Development Center.
SAIC, Frederick Cancer Research and Development Center.
Abstract published in Advance ACS Abstracts, December 1, 1995.
‡
§
X
0
893-228x/96/2709-0172$12.00/0 © 1996 American Chemical Society
+
+

Metabolism of (Methylcarbamoyl)triazenes
Chem. Res. Toxicol., Vol. 9, No. 1, 1996 173
high toxicity of these compounds (3) precluded further
biological studies.
9.6 Hz), 8.33 (1H, d of d, J ) 2.6 Hz, J ) 9.7 Hz), 8.96 (1H, d,
J ) 2.5 Hz), 11.04 (1H, broad). Electron impact mass spec-
trometry indicated an accurate mass of 258.0143 amu (calcu-
During the course of these investigations, a striking
difference in toxicity was noted between CMM and its
corresponding carbethoxytriazene analog, 1-(2-chloro-
ethyl)-3-methyl-3-carbethoxytriazene (CMC). This led to
the hypothesis that these compounds might have very
different fates in vivo. Studies were therefore initiated
to determine the metabolic pathway of the (methylcar-
bamoyl)triazenes, focusing on 1-(2-chloroethyl)-3-benzyl-
lated value of 258.0155 amu for molecular formula C
The compound decomposed above 152 °C.
8 7 4 4
H N O Cl).
The (dinitrophenyl)hydrazone of methoxyacetaldehyde (MAA-
DNP) was prepared as described above for CAA-DNP, except
that 4.5 mmol of methoxyacetaldehyde dimethyl acetal (neat)
replaced the chloroacetaldehyde solution, and no recrystalliza-
1
tion was performed. H NMR of the MAA-DNP derivative
yielded the following: (CD CN, Me Si) δ 3.38 (3H,s), 4.15 (2H,
3
4
3
-(methylcarbamoyl)triazene (CBzM) as a model com-
d, J ) 5.0 Hz), 7.71 (1H, t, J ) 5.0 Hz), 7.91 (1H, d, J ) 9.6 Hz),
8.31 (1H, d of d, J ) 2.6 Hz, J ) 9.6 Hz), 8.96 (1H, d, J ) 2.6
Hz), 11.00 (1H, broad). Electron impact mass spectrometry
indicated an accurate mass of 254.0626 amu (calculated value
pound. Results indicated that degradation of this
compound was oxidative rather than hydrolytic and was
most likely catalyzed by one or more isozymes of cyto-
chrome P450. Similar results were obtained in prelimi-
nary studies of the metabolism of CMM. During the
course of CBzM breakdown, two triazene metabolites
were formed which were isolated and identified as 1-(2-
chloroethyl)-3-benzyl-3-((hydroxymethyl)carbamoyl)tria-
zene (CBzM-met1) and 1-(2-chloroethyl)-3-benzyl-3-
carbamoyltriazene (CBzM-met2). A single triazene
metabolite isolated from the incubation of CMM with
microsomes was identified as 1-(2-chloroethyl)-3-methyl-
9 10 4 5
of 254.0650 amu for molecular formula C H N O ).
Benzylmethylurea was synthesized by slowly adding 25 mL
of a solution of 0.5 M methyl isocyanate in methylene chloride
to 25 mL of a solution of 0.6 M benzylamine in methylene
chloride with constant stirring at 0 °C. When the addition was
complete, the reaction mixture was stirred for an additional 2
h at room temperature. After extraction twice with 50 mL of
0.1 M HCl, the methylene chloride solution was dried over
anhydrous Na
solid was recrystallized from an ethanol/water solution.
NMR revealed the following: (CD CN, Me Si) δ 2.62 (3H, d, J
4.8 Hz), 4.25 (2H, d, J ) 6.2 Hz), 4.94 (1H, s), 5.45 (1H, s),
.28 (5H, m). Electron impact mass spectrometry indicated an
accurate mass of 164.0943 amu (calculated value of 164.0949
amu for molecular formula C O).
2 4
SO and evaporated to dryness. The resulting
1
H
3
-((hydroxymethyl)carbamoyl)triazene
(CMM-met1).
3
4
)
7
These results suggested that the primary site of attack
on these compounds was the methylcarbamoyl substitu-
ent, leading to its hydroxylation and, in the case of CBzM,
its subsequent demethylation. In CBzM incubations,
both the parent compound and the metabolites eventually
completely disappeared, and the lack of appearance of
new compounds bearing the triazene chromophore indi-
cated that further processes led to the destruction of the
triazene moiety. In the case of CMM, metabolism ceased
with significant quantities of both the parent compound
and triazene metabolite still present in incubation mix-
tures.
9
12 2
H N
NADP, glucose 6-phosphate, glucose 6-phosphate dehydro-
genase, and benzylurea were purchased from Sigma Chemical
Co. (St. Louis, MO). Methoxyacetaldehyde dimethyl acetal and
chloroacetaldehyde were purchased from Aldrich (Milwaukee,
WI).
Ra t Liver Micr osom es. Rat liver microsomes were pre-
pared as previously described (9). The preparations contained
approximately 40 mg/mL protein as determined by the method
of Bradford (10).
The above results indicated that the antitumor efficacy
of the (methylcarbamoyl)triazenes may be attributable
to a unique mechanism that does not involve hydrolytic
deacylation or the formation of an alkanediazonium ion.
However, the ultimate fate of these compounds remained
unknown. In this report, we present evidence that
oxidation of both CMM and CBzM also occurs at the
In cu ba tion s of Tr ia zen es w ith Ra t Liver Micr osom es.
Incubation mixtures contained 0.1 M Tris-HCl (pH 7.5), 1 mM
ethylenediaminetetraacetic acid, 1 mM NADP , 5 mM glucose
+
6
3
-phosphate, 0.2 unit/mL glucose 6-phosphate dehydrogenase,
mg of microsomal protein/mL, and the desired final concentra-
tion of triazene added as a stock solution in acetonitrile. In
general, all components of the reaction mixture were combined
except for the microsomes. Samples were warmed to 37 °C, and
the microsomes were added to initiate the reaction. Aliquots
were then removed at desired time points for asssay of triazene,
benzylurea, benzylmethylurea, or chloroacetaldehyde concentra-
tion.
2
-chloroethyl substituent. This oxidation leads to loss
of the 2-chloroethyl group as chloroacetaldehyde. The
resulting unstable triazene breaks down with loss of
nitrogen to yield the corresponding urea.
Assa y of Tr ia zen e Meta bolism . The concentrations of
triazenes were determined by high performance liquid chroma-
tography (HPLC) as described previously (9). In the case of
CBzM, CMM was added to the samples as an internal standard.
When CMM was to be assayed, the internal standard was 1-(2-
chloroethyl)-3-methyl-3-acetyltriazene. The HPLC effluent was
monitored by absorbance at 245 nm, and concentrations of the
parent compounds were calculated from the area of the corre-
sponding peak relative to the internal standard peak. The
concentrations of CBzM-met1, CBzM-met2, and CMM-met1
were similarly estimated, based on the assumption that the
molar absorption coefficients of these compounds would not
differ significantly from those of their parent triazenes. Details
of the eluting solvents and retention times of the compounds
have been published (9).
Ma ter ia ls a n d Meth od s
An im a ls. Young adult male Fischer 344 rats (8-10 weeks)
from the colony of the Frederick Cancer Research and Develop-
ment Center were maintained on standard laboratory diet and
given water ad libitum.
Ch em ica ls. Triazenes were prepared, purified, and char-
acterized according to the procedure of Smith et al. and were
>
99% pure as determined by NMR and UV/vis spectroscopy and
high performance liquid chromatography (1, 2). Triazene stock
solutions were 100 or 300 mM in acetonitrile and were stored
at -20 °C.
The (2,4-dinitrophenyl)hydrazone of chloroacetaldehyde (CAA-
DNP) was prepared by dissolving 4.5 mmol of (2,4-dinitrophen-
yl)hydrazine in 4.5 mL of concentrated H
2
SO
4
, and diluting with
Assa y of Ben zylu r ea a n d Ben zylm eth ylu r ea F or m a tion .
Mixtures containing 300 µM initial concentrations of CBzM were
incubated as described above. At desired time points, aliquots
(100 µL) were combined on ice with 50 µL each of saturated
barium hydroxide and 5% (w/v) zinc sulfate. The precipitates
were removed by centrifugation, and the samples were subjected
to HPLC on a column (4.6 × 250 mm) of Microsorb C18 (Rainin
6
.0 mL of water and 21.0 mL of 95% (v/v) ethanol. Chloroac-
etaldehyde (4.5 mmol) was then added as a 50% (w/v) aqueous
solution. The resulting precipitate was collected by filtration
and recrystallized from 95% (v/v) ethanol. H NMR of the CAA-
1
DNP derivative yielded the following: (CD
3
CN, Me
4
Si) δ 4.35
(2H, d, J ) 5.7 Hz), 7.73 (1H, t, J ) 5.7 Hz), 7.92 (1H, d, J )
+
+

1
74 Chem. Res. Toxicol., Vol. 9, No. 1, 1996
Rouzer et al.
Instrument Co.) eluted at a flow rate of 1 mL/min with
methanol/water (30:70 v/v). The retention times for benzylurea
and benzylmethylurea were 9.1 and 12.2 min, respectively. The
effluent was monitored by absorbance at 250 nm. Quantitation
of each urea was based on the area of the corresponding peak
as compared to a standard curve prepared from the injection of
samples containing 0.5-25 nmol of the authentic compounds.
The response of the HPLC detector was linear with respect to
urea quantity over this concentration range.
with a linear gradient beginning with methanol/water contain-
ing 5 mM ammonium acetate and 6 mM ammonium hydroxide,
pH 9.0 (35:65 v/v), and ending with 100% methanol in a total
volume of 60 mL. The MAA-DNP eluted with a retention time
of 28.0 min. The purified material was evaporated to dryness
under reduced pressure and dissolved in deuterated acetonitrile
for 1H NMR analysis.
An a lyses. NMR spectra were obtained on a Varian VXRS
500 MHz spectrometer equipped with a SUN 4/110 data station.
Assa y of Ch lor oa ceta ld eh yd e F or m a tion . Mixtures (3.5
mL total volume) containing an initial concentration of 300 µM
CBzM or CMM were prepared as described above and incubated
at 37 °C for 3 h. For quantitation standards, identical mixtures
were prepared, except that the triazenes were replaced with
solvent, and known concentrations (100-300 µM) of chloro-
acetaldehyde were added after 1.5 h of incubation. At the
beginning and end of the incubation period, duplicate 100 µL
samples were removed for assay of triazene and metabolite
concentration. To the remaining (3 mL) reaction mixture was
added 1.5 mL each of saturated barium hydroxide and 5% (w/
v) zinc sulfate. After removal of the precipitates by centrifuga-
tion, the resulting supernatants were combined with 5.4 mL of
water and 1.0 mL of a solution of 12.5 mM (2,4-dinitrophenyl)-
hydrazine in 6 M HCl. Samples were mixed for 5 min at room
temperature, at which time 5 mL of isooctane was added. After
mixing for an additional 25 min, the isooctane layers were
removed and extracted with 1 mL of acetonitrile. The aceto-
nitrile extracts were evaporated to dryness, and the residues
were dissolved in 100 µL of acetonitrile for HPLC analysis (11).
Chromatography was performed on the column described above,
eluted at 1 mL/min with a linear gradient beginning with
methanol/water containing 5 mM ammonium acetate and 6 mM
ammonium hydroxide (pH 9.0) (35:65 v/v) and ending with 100%
methanol in a total volume of 30 mL. The column effluent was
monitored for absorbance at 340 nm. The retention time for
the peak corresponding to CAA-DNP was 19.4 min. Quantita-
tion was based on the area of the peak corresponding to CAA-
DNP obtained from triazene incubation mixtures relative to that
of CAA-DNP quantitation standard mixtures. All incubations
were performed in triplicate. The assay was linear with respect
to chloroacetaldehyde concentations from 0 to 300 µM.
Mass spectra were obtained on a VG 70-250 mass spectrom-
eter (VG Analytical, Danvers, MA). The mass spectrometer was
operated in the positive ion detection mode. Ion formation was
by electron impact at a trap current of 100 µA and an electron
energy of 70 eV. The ion source temperature was 180 °C. The
instrument was scanned from m/ z 650 to m/ z 40 at 3 s/decade
with a 0.5 s magnet reset time and a resolving power of 3000.
The accelerating voltage was 6000 v. Accurate masses were
obtained by peak matching using the mass spectrometer data
system and perfluorokerosene as the source of reference peaks.
Samples were introduced via the solid probe.
HPLC was performed on an instrument composed of a SCL-
6B system controller driving two LC-6A pumps, an SPD6AV
UV/vis spectrophotometric detector, and a CR601 Chromatopac
data integrator, all obtained from Shimadzu Scientific Instru-
ments.
Resu lts
F or m a tion of Ben zylu r ea a n d Ben zylm eth ylu r ea
d u r in g CBzM Meta bolism . Previously reported stud-
ies had shown that oxidation of the methylcarbamoyl
substituent led to the formation of the ((hydroxymethyl)-
carbamoyl)triazene and the carbamoyltriazene analogs
of CBzM. However, it was also shown that these
compounds were further metabolized oxidatively in a
process that appeared to result in the destruction of the
triazene moiety (9). Consideration of the structure of
CBzM led to the hypothesis that a likely second site for
metabolism might be the chloroethyl group. Hydroxy-
lation of the carbon attached to N(1) of CBzM should
result in the loss of the chloroethyl substituent as
chloroacetaldehyde, and the breakdown of the triazene
to yield nitrogen and benzylmethylurea. The same
oxidation of the chloroethyl group of CBzM-met1 should
produce (hydroxymethyl)benzylurea, whereas oxidation
of CBzM-met2 should produce benzylurea (Scheme 1).
Consequently, the elution times of commercially available
benzylurea and synthetic benzylmethylurea on reverse-
phase HPLC were ascertained in order that the presence
of these potential metabolites could be investigated in
CBzM reaction mixtures.
Isola t ion of Ben zylm et h ylu r ea fr om CBzM R ea ct ion
Mixtu r es. A mixture (20 mL total volume) containing a 300
µM initial concentration of CBzM was incubated at 37 °C for 3
h. The reaction was terminated by the addition of 10 mL each
of saturated barium hydroxide and 5% (w/v) zinc sulfate. The
precipitate was removed by filtration, and the supernatant was
extracted twice with 40 mL of methylene chloride. The meth-
ylene chloride solution was dried over anhydrous sodium sulfate
and evaporated to dryness. The residue was dissolved in 300
µL of methanol and purified by HPLC on the column described
above eluted at 1 mL/min with methanol/water (40:60 v/v). The
column effluent was monitored for absorbance at 250 nm, and
the retention time for benzylmethylurea was 8 min. The
partially purified samples from three such preparations were
combined, concentrated, and subjected to a second purification
by HPLC, in which the column was eluted with methanol/water
When incubation mixtures containing 300 µM CBzM,
rat liver microsomes, and NADPH were incubated for 3
h at 37 °C, two new compounds were formed which
coeluted on HPLC with benzylmethylurea and benzyl-
urea. Coelution was tested by injecting the authentic
compounds combined with the CBzM reaction mixtures.
Single peaks were obtained for each compound in two
separate solvent systems (methanol/water, 30:70 v/v, and
acetonitrile/water, 20:80 v/v). Neither of these com-
pounds was formed when microsomes, NADPH, or CBzM
was omitted from the reaction mixtures, confirming that
they were the products of oxidative metabolism of the
triazene. Sufficient quantity of the compound that
coeluted with benzylmethylurea was produced to allow
isolation and purification from the reaction mixture. The
(30:70 v/v). The purified material (retention time 12.2 min) was
concentrated under reduced pressure and dissolved in deuter-
1
ated acetonitrile for H NMR and mass spectral analysis.
Isola tion of MAA-DNP fr om CBzM Rea ction Mixtu r es.
A mixture (20 mL) was prepared, incubated, and treated with
barium hydroxide and zinc sulfate as described for the isolation
of benzylmethylurea. Following removal of the precipitate by
filtration, the supernatant was combined with 7 mL of a solution
of 12.5 M (2,4-dinitrophenyl)hydrazine in 6 M HCl. After mixing
for 5 min at room temperature, the sample was overlaid with
3
5 mL of isooctane and mixed for an additional 25 min. The
isooctane layer was then extracted twice with 7 mL of acetoni-
trile. The acetonitrile extracts from two such preparations were
dried over sodium sulfate and evaporated to dryness. The
residues were taken up in a total volume of 500 uL of acetonitrile
for purification by HPLC. The column was eluted at 1 mL/min
1
H NMR spectrum of the isolated metabolite was identi-
cal to that of synthetic benzylmethylurea (see Materials
and Methods), and mass spectral analysis revealed the
+
+

Metabolism of (Methylcarbamoyl)triazenes
Chem. Res. Toxicol., Vol. 9, No. 1, 1996 175
Sch em e 1. P r op osed P a th w a y for th e Oxid a tive Meta bolism of CBzM
correct accurate mass of 164.0954 amu. It should be
noted that a third compound having a retention time
prior to both benzylurea and benzylmethylurea also
appeared in CBzM reaction mixtures. This compound
may have been (hydroxymethyl)benzylurea; however, the
lack of authentic standard and the minute amount of this
product precluded identification.
Figure 1 shows the time course of the metabolism of
CBzM and the formation of CBzM-met1, CBzM-met2,
benzylmethylurea, and the compound that coeluted with
benzylurea. Benzylmethylurea, CBzM-met1, and CBzM-
met2 come directly from the oxidation of CBzM. There-
fore, these compounds appear immediately, as the con-
centration of CBzM decreases. In contrast, benzylurea
should be formed from the oxidation of CBzM-met2.
Consistent with this, we find that the peak coeluting with
benzylurea does not appear until significant levels of
CBzM-met2 have formed. It should be noted that the
efficiency of oxidative metabolism of acyltriazenes is
directly related to their lipophilicity (9). Because CBzM-
met2 is more polar than CBzM, it should not compete
effectively as a substrate for metabolism until its con-
centration reaches reasonable levels.
F or m a tion of Ch lor oa ceta ld eh yd e d u r in g th e
Meta bolism of CBzM. If the metabolic pathway de-
picted in Scheme 1 is correct, then one would expect to
detect chloroacetaldehyde in addition to the substituted
ureas in CBzM reaction mixtures. In order to test this
hypothesis, an assay was devised that could detect and
quantitate chloroacetaldehyde as its (2,4-dinitrophenyl)-
hydrazone derivative. Reaction mixtures containing 300
µM CBzM, microsomes, and NADPH were incubated for
3
h at 37 °C. Following reaction with (2,4-dinitrophenyl)-
hydrazine, a compound was detected in these mixtures
that coeluted with CAA-DNP synthetic standard on
HPLC. This substance was not formed if CBzM, mi-
crosomes, or NADPH was omitted from the incubation.
In order to verify the identity of the putative CAA-
DNP, the compound was purified from CBzM reaction
1
mixtures. The H NMR spectrum of the isolated material
was similar, but not identical to that of synthetic CAA-
DNP derivative (Materials and Methods). In particular,
instead of a doublet at 4.35 with a coupling constant of
5
.7 Hz as found in the CAA-DNP spectrum, the spectrum
of the isolated metabolite showed a doublet at 4.15 with
a coupling constant of 5.0 Hz. In addition, the metabolite
spectrum showed that the coupling constant of the triplet
at 7.73 was 5.0 Hz instead of 5.7 Hz as found in the CAA-
DNP spectrum, and it contained a singlet corresponding
to 3 protons at 3.38 not found in the spectrum of CAA-
DNP. Furthermore, mass spectral analysis yielded a
mass of 254.0638 amu for the metabolite, as opposed to
258.0155 amu expected for CAA-DNP. Clearly, this
compound was not CAA-DNP.
F igu r e 1. Time course for the metabolism of CBzM and the
formation of CBzM metabolites. An incubation mixture contain-
ing an initial concentration of 300 µM CBzM was prepared as
described in Materials and Methods. At the indicated times,
duplicate samples were removed for assay of CBzM, CBzM-
met1, and CBzM-met2 (CBzM-met, quantitated together), ben-
zylmethylurea (BMU), and the compound coeluting with ben-
zylurea (BU). Each data point represents the mean and standard
deviation for three separate experiments.
Since chloroacetaldehyde was a highly likely metabo-
lite from the breakdown of CBzM, we considered the
possibility that CAA-DNP formed in the reaction mix-
+
+

1
76 Chem. Res. Toxicol., Vol. 9, No. 1, 1996
Rouzer et al.
tures was being altered during the purification procedure.
To test this hypothesis, we subjected samples of synthetic
CAA-DNP to the HPLC purification followed by NMR
analysis. The results showed that the NMR spectrum
of the recovered compound was identical to that of the
material isolated from CBzM incubation mixtures, and
not that of CAA-DNP. Furthermore, simply incubating
CAA-DNP with solutions of methanol and water at
concentrations used in the HPLC purification, followed
by evaporation under reduced pressure, led to a similar
change in NMR spectrum. These findings along with the
spectral data suggested the hypothesis that the com-
pound recovered from the HPLC was the (dinitrophenyl)-
hydrazone of methoxyacetaldehyde (MAA-DNP), result-
ing from nucleophilic substitution of the chloro substituent
by methanol. This was confirmed through the prepara-
tion of synthetic MAA-DNP standard, which coeluted
exactly with CAA-DNP on HPLC and gave NMR and
mass spectral data identical to those of the compound
isolated from CBzM incubation mixtures.
Ta ble 1. Meta bolites P r od u ced d u r in g th e Br ea k d ow n of
CBzM
concentration (µM)a
CBzM- CBzM-
CBzMb
BMUc BUd CAAe totalf
expt
met1
met2
1
22.9
50.8
(1.6)
29.5
(5.8)
46.3
35.0
(3.2)
41.3
(4.4)
17.3
(4.6)
133
(4)
126
(7)
143
(24)
22.7
(4.1)
21.5
(8.3)
21.5
(8.3)
67.0
(5.8)
78.3
(1.3)
87.9
(3.0)
264
240
234
(3.7)
2
3
21.5
(4.7)
5.8
(1.1)
(16.8)
meang 16.7
9.5)
42.2
(11.2)
31.2
(12.4)
134
(8)
21.9
(0.7) (10.5)
77.7
246
(
a
Values represent the mean and (standard deviation) of
b
triplicate incubation mixtures for each experiment. Values rep-
resent remaining CBzM concentration at the end of the incubation.
c
Benzylmethylurea. d Benzylurea. e Chloroacetaldehyde. f Total of
CBzM accounted for as the sum of remaining CBzM, CBzM-met1,
CBzM-met2, benzylmethylurea, and benzylurea. g Mean and (stan-
dard deviation) of three separate experiments.
Ta ble 2. Meta bolites P r od u ced d u r in g th e Br ea k d ow n of
CMM
We believe that the most likely explanation for the
isolation of MAA-DNP from CBzM incubation mixtures
is that CAA-DNP is converted to MAA-DNP either on
the HPLC column or during the lengthy evaporation
procedure after HPLC. The finding that this material
is not formed in incubation mixtures in the absence of
CBzM, microsomes, or NADPH indicates that it is the
result of CBzM metabolism; however, there is no obvious
reason to expect the direct formation of methoxyacetal-
dehyde from oxidative metabolism of CBzM. Since CAA-
DNP is converted to MAA-DNP upon incubation with
methanol/water mixtures, and since CBzM reaction
mixtures are not exposed to methanol prior to HPLC
injection, a conversion of CAA-DNP to MAA-DNP after
the HPLC step is reasonable. Attempts to use different
solvent systems and different purification techniques
have not yet resulted in the isolation of CAA-DNP from
CBzM reaction mixtures at an adequate level of purity
to allow spectral confirmation. Nevertheless, we believe
that the similarity in the behavior of the CBzM metabo-
lite compared to that of the CAA-DNP synthetic standard
provides strong evidence that the metabolite that we are
measuring is chloroacetaldehyde.
Qu a n tita tion of CBzM Meta bolites. A series of
experiments were performed in order to quantitate CBzM
and all metabolites formed during a 3 h incubation,
beginning with a starting concentration of 300 µM CBzM.
The results are summarized in Table 1. It is notable that,
in this series of experiments, the same pattern of
metabolites was obtained as was observed in the experi-
ments shown in Figure 1, but the actual quantities of
some of the metabolites differed. This variability was
not uncommon with different preparations of microsomes.
F or m a tion of Ch lor oa ceta ld eh yd e d u r in g th e
Meta bolism of CMM. Initial studies of the metabolism
of CMM had shown that, like CBzM, oxidation at the
methylcarbamoyl substituent could occur, yielding CMM-
met1 (9). However, it seemed quite likely that CMM
should also be subject to metabolism at the chloroethyl
substituent. Because the expected urea metabolites of
CMM, 1,3-dimethylurea, methylurea, and 1-(hydroxy-
methyl)-3-methylurea are not easily detected on HPLC
due to the absence of a chromophore, we searched instead
for the presence of chloroacetaldehyde in CMM reaction
mixtures. Results indicated that, when 300 µM CMM
was incubated for 3 h at 37 °C with rat liver microsomes
and NADPH, a compound was formed that coeluted with
concentration (µM)a
expt
CMMb
CMM-met1
CAAc
totald
292
1
214
(8)
199
45.5
(1.2)
42.8
(2.2)
49.5
(1.7)
91.2
(7.4)
32.3
(4.6)
34.0
(1.3)
49.3
(6.0)
116
2
3
4
276
324
324
(3)
226
(7)
117
(14)
(11)
meane
189
(49)
57.2
57.9
304
(24)
(22.8)
(39.5)
a
Values represent the mean and (standard deviation) of
triplicate incubation mixtures for each experiment. Values rep-
b
resent remaining CMM concentration at the end of the incubation.
c
Chloroacetaldehyde. d Total of CMM accounted for as the sum
e
of remaining CMM, CMM-met1, and chloroacetaldehyde. Mean
and (standard deviation) of the four separate experiments.
synthetic CAA-DNP (and MAA-DNP) on HPLC. This
compound did not form if CMM, microsomes, or NADPH
was omitted from the reaction mixtures. Table 2 sum-
marizes results from a series of experiments in which the
concentrations of CMM and its metabolites were quan-
titated in 3 h incubation mixtures. Consistent with
earlier studies (9), CMM was not as extensively metabo-
lized as CBzM under these conditions. Thus, high
concentrations of CMM and CMM-met1 remained at the
end of the incubation. However, significant levels of
chloroacetaldehyde (58 µM) were detected.
Discu ssion
The results presented here indicate that the (methyl-
carbamoyl)triazenes CBzM and CMM are subject to
oxidative metabolism at a minimum of two sites. These
sites are the methylcarbamoyl and the chloroethyl sub-
stituents. Hydroxylation of the methylcarbamoyl group
may lead to demethylation, at least in the case of CBzM,
but does not directly result in breakdown of the triazene
or the generation of an obvious alkylating agent (9). In
contrast, hydroxylation of the chloroethyl group results
in destruction of the triazene moiety, with formation of
chloroacetaldehyde, a substituted urea, and presumably
molecular nitrogen. The oxidation of the chloroethyl
group of CBzM (and its methylcarbamoyl metabolites)
is supported by the following evidence. Chloroacetalde-
+
+

Metabolism of (Methylcarbamoyl)triazenes
Chem. Res. Toxicol., Vol. 9, No. 1, 1996 177
hyde has been identified as its (2,4-dinitrophenyl)hydra-
zone in CBzM reaction mixtures by its coelution with
synthetic CAA-DNP standard on HPLC, and by its
quantitative conversion to MAA-DNP under conditions
of analysis, as also observed for the CAA-DNP standard.
The expected urea metabolite of CBzM, benzylmethyl-
urea, has been purified and positively identified by
chloroacetaldehyde was added to incubation mixtures
and subjected to the derivatization and extraction pro-
cedure, only 20% of it was recovered as CAA-DNP on
HPLC analysis. The low recovery could have been due
to chemical reaction of the chloroacetaldehyde with
reaction mixture components, to metabolism of chloro-
acetaldehyde to chloroacetic acid (12), and/or to partial
losses during sample processing. In order to compensate
for this problem, we attempted to simulate the environ-
ment of endogenously generated chloroacetaldehyde in
our quantitation standards by adding known concentra-
tions of the compound to microsome incubation mixtures
at the midpoint of the incubation period (see Materials
and Methods). It is not known, however, whether this
technique truly produces standards that accurately re-
flect the fate of the endogenously generated molecule.
Therefore, the exact quantities of chloroacetaldehyde
produced during the metabolism of CBzM cannot be
certain. Interestingly, in the CMM metabolism studies,
the measured quantity of chloroacetaldehyde accounted
fully for the quantity of CMM that was degraded, but
not accounted for as CMM-met1 (Table 2).
In spite of the difficulties with chloroacetaldehyde
quantitation, this compound is clearly produced from the
chloroethyl substituent of CBzM and CMM, whereas we
have found no evidence for the formation of the 2-chlo-
roethanediazonium ion during metabolic breakdown.
These findings are consistent with preliminary in vivo
studies which failed to show significant levels of chloro-
ethylation or methylation of DNA in the livers of rats
treated with single doses of either CBzM or CMM, when
CMC caused measurable DNA alkylation under the same
conditions. Threrefore, it now appears unlikely that the
formation of an alkanediazonium ion can account for the
documented antitumor efficacy of CMM.
1
comparison of its H NMR and mass spectrum with those
of authentic compound. The urea expected from CBzM-
met2, benzylurea, has been tentatively identified on the
basis of coelution on HPLC with standard. An assay for
the urea expected from CBzM-met1, (hydroxymethyl)-
benzylurea, is not yet available. However, it is unlikely
that large quantitities of this compound are formed, as
we observe only a very minor peak that demonstrates
the expected HPLC behavior. It should be noted that
formation of (hydroxymethyl)benzylurea from CBzM-
met1 might be followed by demethylation, in which case
it would be converted to and quantitated with benzylurea.
In the experiments summarized in Table 1, 287 µM
out of the 300 µM CBzM was actually metabolized over
the 3 h incubation. We were able to account for 81% of
the metabolized compound as follows: CBzM-met1 and
CBzM-met2 (26%), benzylmethylurea (47%), and benzyl-
urea (8%). It is interesting to note that, if the pathway
in Scheme 1 is correct, the formation of benzylmethylurea
(47%) represents oxidation of CBzM first at the chloro-
ethyl group, whereas the formation of the other metabo-
lites combined (34%) represents oxidation first at the
methylcarbamoyl group. The preponderance of benzyl-
methylurea suggests that oxidation of CBzM at the
chloroethyl group is favored over oxidation at the meth-
ylcarbamoyl group. It is not known whether the 19% of
the triazene that has not been accounted for represents
experimental error, or whether it indicates that some of
the parent compound decomposes by an as yet unknown
pathway. Alternatively, it is possible that the present
assays simply do not detect all of the metabolites formed,
including (hydroxymethyl)benzylurea. In addition, the
ureas themselves may be subject to still further metabo-
lism such as aromatic or benzylic hydroxylation, although
the stability in the level of benzylmethylurea from 90 to
The critical question arising from these results is
whether chloroacetaldehyde formation during CMM me-
tabolism respresents a simple mechanism for breakdown
and excretion of the triazene, or an alternative explana-
tion for CMM’s antitumor effects. Chloroacetaldehyde
is a known metabolite of both vinyl chloride, an important
industrial carcinogen, and isophosphamide, a clinically
useful cancer chemotherapeutic agent (13-16). However,
the exact role of CAA, if any, in the bological effects of
these compounds remains uncertain. CAA is a known
mutagen presumably due to its reaction primarily with
1
80 min of incubation (Figure 1) does not support this.
If the suggested pathway in Scheme 1 is correct for
CBzM oxidation, then one would expect to detect chlo-
roacetaldehyde in a concentration equal to the sum of
the concentrations of the substituted urea metabolites.
This was not the case, as we detected approximately 156
µM total urea, but only 77.7 µM chloroacetaldehyde in
the experiments shown in Table 1. This discrepancy may
be simply the result of error compounded from multiple
assays, but it also raises two critical questions. First,
we must consider the possibility that a portion of the urea
metabolites is formed by simple chemical breakdown
involving heterolysis of the N(2)-N(3) bond, which is
known to occur at neutral pH, and which would result
in the formation of the 2-chloroethanediazonium ion (4).
However, as mentioned previously, the rate of this
reaction is very slow relative to our 3 h incubation period,
and the fact that we do not detect any of the urea
metabolites in the absence of NADPH or microsomes
suggests that enzymatic oxidation is required for their
formation under our conditions. The second important
question that we must consider is the possibility that not
all chloroacetaldehyde formed during the incubation is
detected in our assay. This is a distinct possibility, as
indicated by our finding that, when a known quantity of
6
adenine and cytosine bases of DNA to form 1,N -
4
ethenoadenine and 3,N -ethenocytosine, respectively (17-
22). DNA cross-linking by chloroacetaldehyde has also
been reported (23). However, the kinetics of the reaction
of CAA with DNA are slow compared to the kinetics of
its reaction with protein (16).
Although further work will be necessary to determine
the exact mechanism for the antitumor action of CMM,
the fact remains that CMM has demonstrated a remark-
ably low degree of toxicity relative to its efficacy. This
is in marked contrast to CMC, which breaks down to
yield a highly reactive alkanediazonium ion in vivo, and
which is highly toxic. Thus, one may ask whether CMM
may represent a prototype for a new class of alkylating
agents in which the ultimate active species is of lower
reactivity than previously thought necessary for antitu-
mor activity, and from which a major benefit of reduced
toxicity may be achieved.
Ack n ow led gm en t. Research sponsored by the Na-
tional Cancer Institute, DHHS, under Contract NO1-CO-
+
+

1
78 Chem. Res. Toxicol., Vol. 9, No. 1, 1996
Rouzer et al.
4
6000 with ABL, and (C.A.R., in part) by a Cottrell
degradation of 1,3-dialkyl-3-acyltriazenes. Biochem. Pharmacol.
4
6, 165-173.
College Science Award of the Research Corp. The
contents of this publication do not necessarily reflect the
views or policies of the Department of Health and Human
Services, nor does mention of trade names, commercial
products, or organizations imply endorsement by the U.S.
government. We are grateful to Dr. Marilyn Kroeger
Smith, Dr. J oseph Saavedra, and Dr. J ames Farrelly for
invaluable advice and discussion.
(
10) Bradford, M. (1976) A rapid and sensitive method for the
quantification of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72, 248-254.
11) Farelly, J . G. (1980) A new assay for the microsomal metabolism
of nitrosamines. Cancer Res. 40, 3241-3244.
(
(
12) Bartsch, H., Malaveille, C., Barbin, A., Bresil, H., Tomatis, L.,
and Montesano, R. (1976) Mutagenicity and metabolism of vinyl
chloride and related compounds. Environ. Health Perspect. 17,
1
93-198.
(
13) Hefner, R. E., Watanabe, P. G., and Gehring, P. J . (1975)
Preliminary studies of the fate of inhaled vinyl chloride monomer
(VCM) in rats. Ann. N.Y. Acad. Sci. 246, 135-149.
(
14) G o¨ the, R., Callerman, C. J ., Ehrenberg, L., and Wachmeister, C.
A. (1974) Trapping with 3,4-dichlorobenzenethiol of reactive
metabolites formed in vitro from the carcinogen vinyl chloride.
Ambio 3, 234-236.
Refer en ces
(
1) Smith, R. H., Mehl, A. F., Hicks, A., Denlinger, C. L., Kratz, L.,
Andrews, A. W., and Michejda, C. J . (1986) 1,3-Dimethyl-3-acyl-
triazenes: Synthesis and chemistry of a novel class of biological
methylating agents. J . Org. Chem. 51, 3751-3757.
2) Smith, R. H., Mehl, A. F., Shantz, D. L., Chmurny, G. N., and
Michejda, C. J . (1988) Novel cross-linking alkylating agents, 1-(2-
chloroethyl)-3-methyl-3-acyltriazenes. J . Org. Chem. 53, 1467-
471.
3) Smith, R. H., Scudiero, D. A., and Michejda, C. J . (1990)
,3-Dialkyl-3-acyltriazenes, a novel class of antineoplastic alky-
(15) Norpoth, K. (1976) Studies on the metabolism of isophosphamide
(NSC-109724) in man. Can. Treat. Rep. 60, 437-443.
(
16) Guengerich, F. P. (1992) Roles of the vinyl chloride oxidation
products 2-chlorooxirane and 2-chloroacetaldehyde in the in vitro
formation of etheno adducts of nucleic acid bases. Chem. Res.
Toxicol. 5, 2-5.
(
1
(
17) McCann, J ., Simmon, V., Streitwieser, D., and Ames, B. N. (1975)
Mutagenicity of chloroacetaldehyde, a possible metabolic product
of 1,2-dichloroethane (ethylene dichloride), chloroethanol (ethyl-
ene chlorohydrin), vinyl chloride, and cyclophosphamide. Proc.
Natl. Acad. Sci. U.S.A. 72, 3190-3193.
(
1
lating agents. J . Med. Chem. 33, 2579-2583.
4) Smith, R. H., Wladkowski, B. D., Herling, J . A., Pfaltzgraff, T.
D., Pruski, B., Klose, J ., and Michejda, C. J . (1992) 1,3-Dialkyl-
(
(
18) Huberman, E., Bartsch, H., and Sachs, L. (1975) Mutation
induction in Chinese hamster V79 cells by two vinyl chloride
metabolites, chloroethylene oxide and 2-chloroacetaldehyde. Int.
J . Cancer 16, 639-644.
3
-acyltriazenes: Products and rates of decomposition in acidic and
neutral solutions. J . Org. Chem. 57, 654-661.
(
5) Smith, R. H., Wladkowski, B. D., Herling, J . A., Taylor, J . E.,
Pfaltzgraff, T. D., Thompson, E. J ., Pruski, B., Klose, J ., and
Michejda, C. J . (1992) Novel triazenes and triazolines from the
base-catalyzed hydrolysis of 1,3-dialkyl-3-acyltriazenes. J . Org.
Chem. 57, 6448-6454.
6) Kroeger-Koepke, M. B., Michejda, C. J ., and Smith, R. H. (1992)
Alkylation of DNA by 1,3-dialkyl-3-acyltriazenes: Correlation of
biological activity with chemical behavior. Chem. Res. Toxicol. 5,
(
19) Bignami, M., Conti, G., Crebelli, R., Misuraca, F., Puglia, A. M.,
Randazzo, R., Sciandrello, G., and Carere, A. (1980) Mutagenicity
of halogenated aliphatic hydrocarbons in Salmonella typhimu-
rium, Streptomyces coelicolor and Aspergillus nidulans. Chem.-
Biol. Interact. 30, 9-23.
(
(
(
(
(
(
(
(
20) J acobsen, J . S., Perkins, C. P., Callahan, J . T., Sambamurti, K.,
and Humayun, M. Z. (1989) Mechanisms of mutagenesis by
chloroacetaldehyde. Genetics 121, 213-222.
21) Barrio, J . R., Secrist, J . A., and Leonard, N. J . (1972) Fluorescent
adenosine and cytidine derivatives. Biochem. Biophys. Res. Com-
mun. 46, 597-604.
22) K u´ smierek, J . T., and Singer, B. (1982) Chloroacetaldehyde-
treated ribo- and deoxyribopolynucleotides. 1. Reaction products.
Biochemistry 21, 5717-5722.
23) Spengler, S. J ., and Singer, B. (1988) Formation of interstrand
cross-links in chloroacetaldehyde-treated DNA demonstrated by
ethidium bromide fluorescence. Cancer Res. 48, 4804-4806.
5
41-547.
7) Smith, R. H., Denlinger, C. L., Kupper, R., Mehl, A. F., and
Michejda, C. J . (1986) Decomposition of 1,3-dialkyltriazenes in
aqueous buffers: Kinetic and mechanistic studies. J . Am. Chem.
Soc. 108, 3726-3730.
8) Smith, R. H., Wladkowski, B. D., Mehl, A. F., Cleveland, M. J .,
Rudrow E, A., Chmurny, G. N., and Michejda, C. J . (1989)
,3-Dialkyltriazenes: Tautomeric equilibria and rates and prod-
ucts of decomposition. J . Org. Chem. 54, 1036-1042.
9) Rouzer, C. A., Thompson, E. J ., Skinner, T. L., Heavner, P. A.,
Bartolini, W. P., Mitchell, K., Kurz, E., Smith, R. H., and
Michejda, C. J . (1993) An unexpected pathway for the metabolic
1
TX9500639
+
+