Microsome-mediated oxidation of N-nitrosodiethanolamine (NDELA), a bident carcinogen

Article abstract of DOI:10.1021/tx000267b

N-Nitrosodiethanolamine (NDELA), an environmentally prevalent, potent carcinogen, undergoes competitive rat liver microsome-mediated oxidation at both the α (adjacent to N)-and β-positions of the 2-hydroxyethyl chains. The former process, α-hydroxylation, is detected by the formation of glycolaldehyde (determined as its 2,4-dinitrophenylhydrazone DNP) that is assumed to arise from the decomposition of the corresponding α-hydroxynitrosamine, which is also the progenitor of the 2-hydroxyethyldiazonium ion. This finding refutes prior published work that states that the α-hydroxylation of NDELA does not occur. Competitive microsomal oxidation at the β-position gives the hemiacetal N-nitroso-2-hydroxymorpholine (NHMOR) at a rate 1.5 times α-hydroxylation. Glycolaldehyde is oxidized in this system to glyoxal at a rate 39 times the conversion of NDELA to glycolaldehyde. The α-hydroxylation of NHMOR at either C-3 or C-5 to give glyoxal or glycolaldehyde, respectively, occurs at respective rates 3-6 times that of the α-hydroxylation of NDELA. Ethylene glycol, a hydrolysis product of the 2-hydroxyethyldiazonium ion is shown to undergo microsome mediate oxidation to glyoxal. Ethyl-2-hydroxyethylnitrosamine (NEELA) undergoes a similar set of microsome-mediated oxidations at α-position of the ethyl (fastest) and 2-hydroxyethyl groups, as well as β-oxidation of the 2-hydroxyethyl group, a process which is slightly more rapid than α-hydroxylation of the same chain. Comparisons of oxidations rates of these substrates, as manipulated by preinducers, isoniazid, streptozocin, and phenobarbital, and enzyme inhibitors diethyldithiocarbamate and 4-methylpyrazole, with that of dimethylnitrosamine, a substrate for cytochrome P450 2E1, strongly suggest that this isozyme is also responsible for the oxidations reported here. α-Deuteration of NDELA practically eliminates its α-hydroxylation by microsomes from isoniazid induced rats, but doubles β-oxidation, while β-deuteration of this substrate significantly reduces β-oxidation and enhances α-hydroxylation. Since both glyoxal-guanine and 2-hydroxyethyl-DNA base adducts are known to arise from the in vivo administration of NDELA and because this work demonstrates that these two fragments can come from the microsomal oxidation of a single nitrosamine molecule containing the 2-hydroxyethyl group, NDELA and related nitrosamines are bident (two-toothed) carcinogens, a process which is likely to enhance their carcinogenic potency.

Full text of DOI:10.1021/tx000267b

Chem. Res. Toxicol. 2002, 15, 457-469
457
Articles
Micr osom e-Med ia ted Oxid a tion of
N-Nitr osod ieth a n ola m in e (NDELA), a Bid en t Ca r cin ogen
Richard N. Loeppky* and Petra Goelzer
Department of Chemistry, University of Missouri, Columbia, Missouri 65211
Received December 29, 2000
N-Nitrosodiethanolamine (NDELA), an environmentally prevalent, potent carcinogen,
undergoes competitive rat liver microsome-mediated oxidation at both the R (adjacent to N)-
and â-positions of the 2-hydroxyethyl chains. The former process, R-hydroxylation, is detected
by the formation of glycolaldehyde (determined as its 2,4-dinitrophenylhydrazone DNP) that
is assumed to arise from the decomposition of the corresponding R-hydroxynitrosamine, which
is also the progenitor of the 2-hydroxyethyldiazonium ion. This finding refutes prior published
work that states that the R-hydroxylation of NDELA does not occur. Competitive microsomal
oxidation at the â-position gives the hemiacetal N-nitroso-2-hydroxymorpholine (NHMOR) at
a rate 1.5 times R-hydroxylation. Glycolaldehyde is oxidized in this system to glyoxal at a rate
39 times the conversion of NDELA to glycolaldehyde. The R-hydroxylation of NHMOR at either
C-3 or C-5 to give glyoxal or glycolaldehyde, respectively, occurs at respective rates 3-6 times
that of the R-hydroxylation of NDELA. Ethylene glycol, a hydrolysis product of the 2-hydroxy-
ethyldiazonium ion is shown to undergo microsome mediate oxidation to glyoxal. Ethyl-2-
hydroxyethylnitrosamine (NEELA) undergoes a similar set of microsome-mediated oxidations
at R-position of the ethyl (fastest) and 2-hydroxyethyl groups, as well as â-oxidation of the
2-hydroxyethyl group, a process which is slightly more rapid than R-hydroxylation of the same
chain. Comparisons of oxidations rates of these substrates, as manipulated by preinducers,
isoniazid, streptozocin, and phenobarbital, and enzyme inhibitors diethyldithiocarbamate and
4-methylpyrazole, with that of dimethylnitrosamine, a substrate for cytochrome P450 2E1,
strongly suggest that this isozyme is also responsible for the oxidations reported here.
R-Deuteration of NDELA practically eliminates its R-hydroxylation by microsomes from
isoniazid induced rats, but doubles â-oxidation, while â-deuteration of this substrate
significantly reduces â-oxidation and enhances R-hydroxylation. Since both glyoxal-guanine
and 2-hydroxyethyl-DNA base adducts are known to arise from the in vivo administration of
NDELA and because this work demonstrates that these two fragments can come from the
microsomal oxidation of a single nitrosamine molecule containing the 2-hydroxyethyl group,
NDELA and related nitrosamines are bident (two-toothed) carcinogens, a process which is
likely to enhance their carcinogenic potency.
notably in metalworking fluids (13-18) where concentra-
tions as high as 2-3% have been detected. To better
estimate the human health hazards of this carcinogen,
an understanding of its mechanism of activation is
necessary. In recent publications we have documented
some of the difficulties that have plagued searches for
the NDELA mechanism(s) (19, 20). While most nitro-
samines are activated by initial P450 mediated R-hy-
droxylation, this process is unknown for NDELA (20-
22). Nevertheless, recent studies in our laboratory utilizing
In tr od u ction
In developed countries, N-nitrosodiethanolamine (NDE-
LA),1 1, a potent animal carcinogen (1-5), is a ubiquitous
trace contaminant in many commercial products. It arises
from di- and triethanolamine, and their various deriva-
tives by inadvertent nitrosation processes that occur
during product formulation. It has been found in cosmet-
ics, shampoos, and personal care items (6-12), but most
1
Abbreviations: ADH, alcohol dehydrogenase; DCC, diethyldithio-
carbamate; DMN, dimethylnitrosamine; DNP, 2,4-dinitrophenyhydra-
zone; DNPH, 2,4-dinitrophenylhydrazine; MEPY, 1-methylpyrazole;
NDELA, N-nitrosodiethanolamine; NEEALA, N-nitrosoethyletha-
nalamine; NEELA, ethyl-2-hydroxyethylnitrosamine; NHMOR, N-ni-
troso-2-hydroxymorpholine; SSB, single strand breaks.
* To whom correspondence should be addressed. Phone: (573) 882-
4885. E-mail: LoeppkyR@missouri.edu.
10.1021/tx000267b CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/20/2002

458 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Loeppky and Goelzer
phosphate, glucose-6-phosphate dehydrogenase, 2,4-dinitrophe-
nylhydrazine (DNPH), glycolaldehyde (dimer) and glyoxal (40%
solution) were purchased from Sigma Chemical (St. Louis, MO).
Ethylene glycol was obtained from Fisher (St. Louis, MO). The
inhibitors, 4-methylpyrazole and diethyldithiocarbamate were
obtained from Aldrich Chemical (Milwaukee, WI). Nitrosamines
(dimethylnitrosamine, ethylethanolnitrosamine, N-nitrosodi-
ethanolamine, ethylethanalnitrosamine, N-nitroso-2-hydroxy-
morpholine) are all known compounds prepared in our labora-
tory according to literature procedures. The isotopic purity of
the deuterated NDELA was 98% (24).
In cu ba tion s. Different substrate concentrations were incu-
bated with microsomes (1 mg protein/mL), 10 mM MgCl2, 7.5
mM G-6-P (glucose6-phosphate), 0.75 mM NADP, and 5 units/
mL glucose 6-phosphate dehydrogenase in a 100 mM phosphate
buffer, pH 7.4. Samples (volume 500 µL) were incubated in a
shaking thermomixer for 20 min to 1 h at 37 °C. The reaction
was terminated by the addition of 100 µL trichloroacetic acid
(10%), and then 500 µL of 2,4-dinitrophenyl hydrazine (2 mM
in 18% phosphoric acid) was added. After 5 min reaction time
the hydrazones and the remaining hydrazine were extracted into
500 µL of methylene chloride. The organic phase (200 µL) was
transferred into a new vial, evaporated with N2, and redissolved
in 200 µL of acetonitrile. To ensure an excess of hydrazine, only
an aliquot of each of the incubation mixtures containing
glycolaldehyde and NHMOR was reacted with DNPH. Inhibitors
were added at concentrations of 1 mM before the addition of
the microsomes.
A typical run involved the addition of the following substances
to give the final concentration noted: 20 µL of NDELA (0.5 M)
) 20 mM; 50 µL of microsomes (10 mg/mL) ) 1 mg/mL; 50 µL
of cofactor (20 mg of MgCl2 ) 100 mM, 20 mg of glucose 6
phosphate ) 75 mM, 6 mg of NADP ) 7.5 mM in 1 mL of water);
5 µL of glucose-6-phosphate dehydrogenase (500 units/mL) ) 5
units/mL; 375 µL of phosphate buffer, pH 7.4, 0.1 M.
HP LC. The samples were separated on a Zorbax C18 column
(4.6 mm × 25 cm), flow 1 mL/min, with a gradient of acetonitrile
(B) in water (A): 0 min 90% A/10% B, 34 min 60/40, 50 min
20/80, 65 min 20/80, 75 min 90/10. Detection: UV-vis at 390
nm.
specifically R- and â-deuterated isotopomers of NDELA
have produced evidence for in vivo DNA damage pro-
duced by the scission of the R-CH bond (breakage of the
CH bond adjacent to the N) (20). Because these data
strongly suggest the operation of an R-hydroxylation
pathway for NDELA, we have reinvestigated the research
which claims that this process does not occur (21, 22).
Numerous in vitro investigations of nitrosamine me-
tabolism by either S-9000 cell fractions or microsomes
have relied on the detection of aldehydes which are
produced by the decomposition of an unstable R-hydrox-
ynitrosamine intermediate 3, as is shown in Scheme 1.
Sch em e 1
The aldehyde 4 is commonly trapped as a semicarbazone
7 and then converted into a 2,4-dinitrophenylhydrazone
(DNP) 8, which is extracted with isooctane and detected
by HPLC-UV-vis in comparison with known standards
(23). In the case of NDELA, R-hydroxylation would result
in the generation of glycolaldehyde 11 (Scheme 2).
Sch em e 2
DNP s. Hydrazone standards were prepared by refluxing the
respective aldehydes (1 mmol) with 2,4-dinitrophenylhydrazine
(1 mmol) in ethanol, containing 10% phosphoric acid. The
precipitated hydrazones were filtered, washed and dried, and
dissolved in acetonitrile. To quantify the recovery, of the
respective hydrazone, a known amount of aldehyde was added
to a sample of heat deactivated microsomes, derivatized with
DNPH, and immediately extracted according to the procedure
described above. Microsomal and glycolaldehyde blanks were
performed in each case. Recoveries were as follows: glyoxal,
100%; glycolaldehyde, 32%, acetaldehyde and formaldehyde,
45%; NHMOR, 15%; and N-nitrosoethylethanalamine (NEEA-
LA), 45%. Low recoveries in some cases resulted from incomplete
derivatization and/or extraction, but were very reproducible.
Farrelly et al. attempted to detect the R-hydroxylation
of NDELA by both rat liver microsomes and rat hepato-
cytes using the analytical method above and failed to
observe formation of the DNP of glycolaldehyde or other
aldehydes (21, 22). Because our deuteration experiments
(20) with NDELA in vivo pointed toward an R-hydroxy-
lation pathway for this compound, we reconsidered the
applicability of the method shown in Scheme 1 to the
detection of glycolaldehyde, a very polar aldehyde, which
is also susceptible to osazone reactions with hydrazines
to give hydrazones of glyoxal. Isooctane is a very nonpolar
organic solvent, and the possibility that it would not
extract either the DNP of glycolaldehyde or glyoxal was
considered prior to the initiation of the work presented
here. We report that through a modification of the
analytical method shown in Scheme 1 we have been able
to observe the R-hydroxylation of NDELA, as well as
other significant biotransformations.
Resu lts
Mod ifica tion of th e r-Hyd r oxyla tion Assa y. In-
vestigation of the assay used by Farrelly et al. (23) and
Yoo et al. (25) to detect the R-hydroxylation of NDELA
showed that it could not detect glycolaldehyde or glyoxal.
As a result, several steps in these similar procedures
were changed. Difficulties with the assay arise because
the DNP of glycolaldehyde can be oxidized by DNPH
under acidic conditions (the osazone reaction, Scheme 3)
Ma ter ia ls a n d Meth od s
Gen er a l. Microsomes were purchased from Xenotech, Kansas
City, KS. Male Sprague-Dawley rats (8 weeks old) were induced
with isoniazid (200 mg/kg every day for 4 days), phenobarbital
(80 mg/kg every day for 4 days), and streptozocin (single
injection 100 mg/kg). Rat liver microsomes were prepared on
day 5 by differential centrifugation in a suspension of 250 mM
sucrose with 10 mg of protein/mL. NADP, NAD, glucose-6-
Sch em e 3

R- and â-Oxidation of N-Nitrosodiethanolamine
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 459
F igu r e 1. Products, glycoladehyde and NHMOR, from the
microsomal oxidation of NDELA are shown to increase as the
substrate concentration is increased. All incubations were
stopped at 20 min.
F igu r e 2. Time course of product formation from the incuba-
tion of NDELA (20 mM) with rat liver microsomes (isoniazid
induction).
to the di-DNP of glyoxal. Since glyoxal itself is a possible
product of the NDELA metabolism, selective determina-
tion of the two aldehydes is important. Therefore, the
addition of semicarbazide to trap the aldehydes im-
mediately in the incubation mixture was omitted because
this step significantly increased osazone formation. The
application of milder conditions (higher pH, and lower
temperatures) with the semicarbazide trapping did not
help, and resulted in a mixture because of incomplete
conversion of the semicarbazone to the DNP. Direct
addition of DNPH and a short reaction time of 5 min at
room-temperature minimized the oxidative conversion of
glycolaldehyde to glyoxal via the osazone reaction to
about 1%. Isooctane was found to be an ineffective solvent
for the extraction of the polar DNPs of glycolaldehyde
and glyoxal. It was replaced by the more polar methylene
chloride. With this solvent, the unreacted DNPH is
extracted too, but it can be separated from the products
by HPLC on RP-18 column using a gradient of acetoni-
trile and water. With these modifications, the recovery
for glycolaldehyde was about 30%.
Meta bolism of NDELA by Ra t Liver Micr osom es.
Incubation of NDELA with isoniazid-induced rat liver
microsomes and a NADPH generating system resulted
in the formation of glycolaldehyde, NHMOR, and glyoxal.
Concentrations ranging from 0.5 to 80 mM NDELA
produced increasing amounts of aldehyde finally reaching
saturation at 120 mM (Figure 1). Product formation was
nearly linear with time up to 30 min before leveling off
at about 60 min (Figure 2). At 20 mM NDELA, the rate
of NHMOR formation (1.8-2.0 nmol/mg/min) exceeded
glycolaldehyde formation (0.9-1.0 nmol/mg/min) by a
factor of 2 while glyoxal was only a minor product with
0.1 nmol/mg min. While we cannot be absolutely certain
at this stage, the reaction appears to be catalyzed by a
cytochrome P450 enzyme since controls without glucose-
6-phosphate dehydrogenase, as well as controls without
microsomes, showed no aldehyde formation. There was
no evidence for an ADH (alcohol dehydrogenase)-medi-
ated oxidation of NDELA (20, 26-30) under these condi-
tions as replacement of the cofactor NADPH by NAD did
not result in any NHMOR formation. In a control
reaction, we observed that microsomes from isoniazid-
induced rats, with the NADPH generating system, pro-
F igu r e 3. A comparison of the production of NHMOR, glyco-
laldehyde, and glyoxal from the incubation of NDELA with the
same microsomes that were used to separately produce form-
aldehyde from the incubation with dimethylnitrosamine (DM-
N).For each set of microsomes, each induced by a different agent,
the quantity of NDELA metabolites per milligram of protein is
represented by the three left bars and the right bar shows the
quantity of formaldehyde produced from DMN under identical
conditions.
duced formaldehyde and acetaldehyde without the ad-
dition of any specific substrate. Therefore, the possible
formation of small amounts of these compounds from
NDELA could not be detected. The same system also
showed formation of glyoxal, but the amounts were very
small and did not prevent the determination of glyoxal
in the sample.
To further investigate the possible involvement of
cytochrome P450, microsomes from animals that had
been induced by different drugs were compared. Isoniazid
and streptozocin, which are known to be inducers of 2E1
(25, 31-34), and phenobarbital an inducer of both 2E1
and 2B1 (32), were selected and the activity of the
respective microsomes was measured by determining the
demethylation of dimethylnitrosamine DMN (Figure 3).
The incubation time was extended to 1 h to increase
product formation. Normal microsomes formed formal-
dehyde at a rate of 62 µM/h/mg protein which is in the

460 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Loeppky and Goelzer
Ta ble 1. Mich a elis-Men ten P a r a m eter s for th e Micr osom a la -Ca ta lyzed Oxid a tion of NDELA a n d Its P r od u cts
Vmax
(nmol/min/
mg protein)
Km
(mM)
substrate
NDELA
NDELA
NHMOR
NHMOR
glycolaldehyde
product
Vmax/Km
rel. rate
glycolaldehyde
NHMOR
glycolaldehyde
glyoxal
3.84 ( 0.07
10.4 ( 0.6
7.4 ( 0.6
1.5 ( 0.4
3.2 ( 0.5
15 ( 0.01
27 ( 2 0.8
5.2 ( 0.5
1.8 ( 0.5
0.32 ( 0.05
0.256 ( 0.005
0.38 ( 0.03
1.4 ( 0.3
0.8 ( 0.4
10 ( 3
1
1.5
5.6
3.3
39.1
glyoxal
a
Isoniazid induced rat liver microsomes.
F igu r e 5. Michelis-Menten plots of the formation of NHMOR
and glycolaldehyde from NDELA where catalysis is provided
by isoniazid induced microsomes. Each line represents the
average of three determinations.
formed glycolaldehyde at a rate of 0.58 nmol/mg/min
while its conversion rate to glyoxal was 0.26 nmol/min/
mg. Without the addition of microsomes concentrations
up to 2 mM NHMOR showed neither formation of
glycolaldehyde nor glyoxal within 60 min. Figure 4B
shows that glyoxal can also be generated by the oxidation
of glycolaldehyde under these conditions. Incubation of
0.4 mM glycolaldehyde with microsomes from isoniazid-
induced rats gave glyoxal at a rate of 1.36 nmol/mg/min.
This value was corrected for the nonenzymatic autoxi-
dation of glycolaldehyde (0.44 nmol/min). Enzymatic
oxidation of glycolaldehyde to glyoxal is about 10 times
faster than glyoxal formation from NHMOR (Table 1).
With microsomes from noninduced rats the reaction is
much slower for both NHMOR and glycolaldehyde. After
incubating 0.4 mM NHMOR for 1 h with these mi-
crosomes, 3 nmol/mg glyoxal and 3.3 nmol/mg glycolal-
dehyde form, compared to 22 nmol of glyoxal and 24 nmol
of glycolaldehyde, when microsomes from induced rats
are used. A similar result was obtained for glycolaldehyde
where microsomes from noninduced animals produced
only 15% of the glyoxal generated by the induced rat
microsomes.
F igu r e 4. Yield of oxidation products from the microsomal
oxidation of NDELA is shown as a function of time. Panel A
shows that both glycoladehyde and glyoxal are formed from the
isoniazid induced microsomal oxidation of NHMOR at a sub-
strate concentration of 0.4 mM. Panel B shows that glyoxal is
formed from both the microsomal and nonenzymatic oxidation
of glycolaldehyde (0.4 mM).
expected range (31, 32, 35). Phenobarbital induction
increased the formaldehyde production by a factor of 2
while isoniazid and streptozocin inductions resulted in
an increase of about 4 times. With NDELA, the various
microsomes showed an increase in aldehyde formation
pattern similar to the demethylation of DMN. Mi-
crosomes from isoniazid and streptozocin-induced rats
produced the highest yields of both glycolaldehyde and
NHMOR, while phenobarbital-induced rat microsomes
only showed an increase of a factor of 2 over those from
noninduced rats. Comparing the ratio of glycolaldehyde
to NHMOR, induction has a slightly stronger effect on
glycolaldehyde formation than on NHMOR.
Oxid a tion of Meta bolites fr om NDELA by Ra t
Liver Micr osom es. The metabolites NHMOR and gly-
colaldehyde themselves can be further oxidized by rat
liver microsomes. NHMOR is metabolized to glycolalde-
hyde and glyoxal (Figure 4A). At a concentration of 0.4
mM NHMOR, microsomes from isoniazid induced rats
Another product of the microsome-mediated oxidation
of glycolaldehyde is formaldehyde, which was formed at
a rate of 0.42 nmol/mg/min. Because formaldehyde is also
formed in the microsomal blank, we have estimated that
only about 50% of the total formaldehyde arises from the
microsomal oxidation of glycolaldehyde. Formaldehyde
was not found in the glycolaldehyde substrate blank, nor
was it present in the incubation of glyoxal with mi-
crosomes, showing that it is a result of enzymatic

R- and â-Oxidation of N-Nitrosodiethanolamine
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 461
F igu r e 6. Deuterium isotope effects on the microsomal oxidation of isotopomers of NDELA. In panel A the affect of specific deuterium
substitution on the rates of glycolaldehyde, NHMOR, and glyoxal formation from the isoniazid induced microsome mediated oxidation
are shown. In panel B rates are given as sets for each isotopomer and all of the induced rat liver microsomes used in this work.
(Inducers: Strep ) streptozocin, PB ) phenobarbital).
oxidation of glycolaldehyde. After 1 h incubation time,
45% of the glycolaldehyde was recovered unchanged, 28%
was converted to glyoxal and 10% to formaldehyde.
From other experiments, not reported here, we know
that ethylene glycol is a product of the decomposition of
the R-hydroxynitrosamine 9 through the hydrolysis of the
2-hydroxyethyldiazonium ion 10 (Scheme 2). Accordingly,
we have investigated the microsomal oxidation of ethyl-
ene glycol and shown that it produces glycolaldehyde.
Rates are less than that observed for the comparable
glycolaldehyde oxidation. Incubation of 2 mM ethylene
glycol with microsomes from isoniazid-induced rats re-
sults in the formation of glycolaldehyde at a rate of 120
nmol/mg/h. The NADPH-generating system was neces-
sary to catalyze this reaction, suggesting an involvement
of a cytochrome P 450.
The Michaelis-Menten parameters (Vmax and Km) for
the isoniazid-induced rat liver microsomal oxidation of
NDELA and its oxidation products are given in Table 1.
In all cases, data comparable to that shown for the
oxidation of NDELA in Figure 5 were obtained. Initial
rates were determined from data obtained in the first
20 min of reaction. Because of the different reaction rates,
variable concentrations were used and ranged from 1 to
120 mM for NDELA, 0.2-10 mM for NHMOR and
0.04-2 mM for glycolaldehyde, respectively.
Oxid a t ion of Sp ecifica lly Deu t er a t ed NDE LA
Su bstr a tes. As mentioned in the Introduction, the
impetus for this work was the finding that deuterium
substitution at the R-position of NDELA inhibited DNA
single strand breaks in vivo (20). It was therefore
important to determine whether the same phenomena
could be observed in the microsome-mediated oxidation
of NDELA. In general, rate reductions can be expected
upon deuterium substitution of hydrogen when the C-H
bond is broken in a rate-determining step. The rates of
production of the spectrum of metabolites arising from
NDELA were determined as a function of deuterium
substitution in the substrate and the results are shown
in Figure 6. With microsomes from isoniazid-induced
rats, all three compounds formed the same three me-
tabolites yet the ratio of metabolites depended on the
deuteration pattern. Deuteration at the C-2 reduced
NHMOR formation by a factor of 2 and increased the rate
of production of glycolaldehyde by more than a factor of
2 over the control. On the other hand, the C-1 deuteration
did not change the rate of NHMOR formation signifi-
cantly, but nearly eliminated the R-hydroxylation product
glycolaldehyde. Glyoxal formation was decreased for both
deuterated compounds because it requires oxidation at
both carbons. In Figure 6B, we compare the rates of
metabolite formation as a function of the inducer. Al-
though there are some differences in relative amounts,
the same general trends are observed.
Com p a r ison of NDELA w ith NEELA. To study the
influence of the hydroxyl group in the â-position (to the
N), we compared the metabolism of ethyl-2-hydroxyeth-
ylnitrosamine (NEELA 26) and NDELA. With mi-
crosomes from isoniazid-induced rats, NEELA formed
both glycolaldehyde 11 and acetaldehyde. Oxidation also
occurs at the â-position resulting in the formation of
ethylethanalnitrosamine (NEEALA) 29. The rate of ac-
etaldehyde formation (361 nmol/h/mg) was greater than
that of glycolaldehyde formation (139 nmol/h/mg) dem-
onstrating a preference for oxidation of the ethyl group
over the hydroxyethyl moiety. As was the case with
NDELA, the rate oxidation of the 2-hydroxyethyl group
is slightly more rapid (153 nmol/h/mg) at the â-position
to produce 29 than at the R-position which gives 11.
â-Oxidation of the ethyl chain would produce NDELA,
but this process, which is known for other nitrosamines,
cannot be detected by our assay because the product is
not an aldehyde.
As is shown in Figure 7, those inducers that are most
effective in increasing the rate of DMN oxidation are also
the ones that are most effective in catalyzing the R-hy-
droxylation of NEELA suggesting that the same isoen-
zyme is involved. Oxidation at the â-position of the
2-hydroxyethyl chain shows the same pattern.
Effect of En zym e In h ibitor s. Diethyldithiocarbam-
ate (DDC) and methylpyrazole (MEPY) are relatively
selective inhibitors for P450 2E1 in rat liver microsomes
(31, 36). Methylpyrazole is also an inhibitor for alcohol
dehydrogenase. With our isoniazid-induced microsomes
both inhibitors showed a strong inhibition of DMN
demethylation as described by Yamazaki et al. (31). The
data for the effect of these inhibitors on the generation
of aldehydes from NDELA and NEELA are shown in

462 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Loeppky and Goelzer
NDELA practically eliminates the formation of glycola-
ldehyde; (3) model chemistry experiments which show
that the hydrolysis of R-acetoxy NDELA, a progenitor of
R-hydroxy NDELA, generates glycoladehyde in addition
to other products (52); and (4) O6-hydroxyethylguanine
adducts, produced from the reaction of the diazonium ion
derived from R-hydroxy NDELA which were found in rat
liver DNA following the administration of NDELA (30,
53). Second, microsomes mediate the oxidation of NDELA
at the â-position to give NHMOR, and the rate of this
process is greater than that of the R-hydroxylation. Both
glycolaldehyde and NHMOR are substrates for further
microsomal oxidation yielding glyoxal and other electro-
philes that bind DNA. Thus, both R- and â-oxidation are
important in the carcinogenic activation of NDELA.
F igu r e 7. Rates of production of the aldehyde metabolites from
the rat liver microsome meditated oxidation of NEELA are
shown as a function of the inducing agent (STR ) streptozocin,
ISO ) isoniazid, PB ) phenobarbital, and N ) no induction
(control).
Sch em e 4
Figure 8. In incubations with NDELA, DDC and MEPY
inhibit the formation of glycolaldehyde, glyoxal, and
NHMOR, exhibiting a slightly greater effect on glycola-
ldehyde and glyoxal production. Similar results were
obtained with NEELA, where DDC and MEPY also
produced a greater inhibition of glycolaldehyde and
acetaldehyde formation than that observed for the â-ox-
idation as detected by production of 29. As can be seen
from inspection of panels A and B of Figure 8, the two
inhibitors produced similar effects with each of the
substrates. We also found both inhibitors to decrease the
rate of NDELA metabolism by microsmes from phe-
nobarbital-induced animals (data not shown), but MEPY,
in particular was a significantly less effective inhibitor
with these microsomes.
r-Hyd r oxyla tion of NDELA. The thinking that
R-hydroxylation was not involved in the bioactivation of
NDELA arose from three types of experimental data.
Numerous mutagen assay experiments involving S-9000
fractions as an activating media failed to show mutage-
nicity for NDELA (19, 20). Neither glycolaldehyde nor
glyoxal, the expected products from the R-hydroxylation
of NDELA or its metabolite NHMOR, respectively, could
be detected from the incubation of either rat liver
microsomes or hepatocytes with NDELA or NHMOR (22).
Using high specific activity [14C]NDELA, Lijinsky and
Farrelly found only low levels of radio-carbon incorpora-
tion into rat liver DNA and this observation is contrary
to that observed for numerous other nitrosamines for
which evidence of bioactivation by R-hydroxylation exists
(21). On the other hand, N-2-hydroxyethyl-N-nitrosogly-
cine is a urinary metabolite of NDELA and NHMOR, its
precursor, has been detected in the rat liver S-9000
mediated metabolism of NDELA (45, 46). Both of these
Discu ssion
The mechanism of bioactivation of NDELA 1 has been
in doubt for many years (19-22, 26-30, 37-51). The
work presented here reveals two major findings that
significantly clarify this problem and lead to a sound
hypothesis for the initiation of carcinogenesis by NDELA.
First, NDELA undergoes microsome-mediated R-hy-
droxylation to generate an unstable R-hydroxynitro-
samine 9, as shown in Scheme 4. Evidence for the
existence of the pathway is provided by (1) the detection
of the R-hydroxynitrosamine decomposition product, gly-
colaldehyde; (2) the demonstration that R-deuteration of
F igu r e 8. Effect of enzyme inhibitors, DCC and MEPY, on the percentage of each aldehyde, based on controls, formed during
the isoniazid induced rat liver microsome incubation with (A) NDELA and (B) NEELA, respectively, is displayed in the two
panels.

R- and â-Oxidation of N-Nitrosodiethanolamine
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 463
substances are â-oxidation products. NHMOR is a mod-
estly reactive compound that reacts with DNA in vitro
(54).
results in detoxification. Another detoxification pathway
is represented by path B, which involves an intramo-
lecular rearrangement via a 1,2-hydride shift (sigmatro-
pic transformation) to 17 and generates acetaldehyde 18.
Independent estimates show that this pathway accounts
for at least 35% of the chemistry of the 2-hydroxyethyl-
diazonium ion, based on product studies (52). Pathway
C is relatively minor (8-10%) and generates the rela-
tively unreactive ethylene oxide 19. It is unlikely that
the free carbocation 20 forms. Thus, the intrinsic chem-
istry of 10 severely limits its reaction with DNA, because
highly reactive electrophiles must be produced nearby
for reaction and the mutagenesis that results from that
process to occur. In summary, then, failure to observe
mutagenicity for NDELA in numerous experiments dur-
ing 1970s was likely due to the relatively inactive S-9000
toward nitrosamines, the poor substrate quality of
NDELA, and the “self-detoxifying chemistry” of a sig-
nificant fraction of the 2-hydroxyethyldiazonium ions
produced.
The most important evidence for the R-hydroxylation
of NDELA is the observation of the production of glycol-
aldehyde and that this transformation is significantly
inhibited by the R-deuteration of NDELA. As discussed
earlier in this paper, previous assays for the DNP of this
compound relied on the extraction of this very polar
compound into nonpolar isooctane in which it has very
low solubility and did not take into account the osazone
reaction of the hydrazine and the R-hydroxyaldehyde
which result in the even less soluble glyoxal di-DNP (an
osazone) (22, 23). Numerous investigations of the mi-
crosome-mediated metabolism of nitrosamines have em-
ployed the isooctane extraction of the poorly soluble
DNPs produced from the aldehydes resulting from the
decomposition of the R-hydroxynitrosamines. Unless,
careful recovery and control experiments are or were
performed, the extent of metabolic activation by this
pathway could be significantly underestimated. Here we
extracted the DNPs with CH2Cl2 in which they are much
more soluble. Our improvements in the assay procedure
are undoubtedly responsible for our ability to detect the
R-hydroxylation of NDELA and metabolites derived from
it.
Sch em e 5
If NDELA is activated by R-hydroxylation to generate
a reactive R-hydroxynitrosamine that then decomposes
to the 2-hydroxyethy-diazonium ion (see Scheme 2), a
putative potent electrophile, then why was no mutage-
nicity observed for NDELA in numerous Ames assays
and modifications thereof. In fairness, several groups
report low levels of mutagenicity for NDELA in S-9000
mediated Ames assays (51), but the vast majority of
reports show negative mutagenicity for this compound
(55). There are probably several reasons for the failure
to observe S-9000-mediated mutagenicity for NDELA
(56). As we have shown (Table 1), NDELA is not a
particularly good substrate even with the microsomes we
used. In an S-9000 preparation, the microsome quantity,
and thereby the concentration of the active enzyme(s),
is diluted compared to the preparations used by us.
Moreover, anecdotal reports of failures to observe mu-
tagenicity for various known mutagenic nitrosamines are
common. Much more is known now about the P450
enzymes in microsomal preparations and how to keep
them active than was the case when many of the
mutagenicity experiments were performed several de-
cades ago. Moreover, the compounds generated by the
R-hydroxylation of nitrosamines are often much less
stable than the electrophiles produced from other car-
cinogens. R-Hydroxynitrosamines have short half-lives
and the diazonium ions and carbocations derived from
them are even more reactive. Primary diazonium ions,
such as the methyl- or ethyldiazonium ion, mainly react
by very low activation energy Sn-2 nucleophilic displace-
ment reactions, rather than by carbocation generation
(57). In the case of the 2-hydroxyethyldiazonium ion,
other pathways also play an important role in the
chemistry of this species. These are depicted in Scheme
5. The two predominant reaction pathways for this
diazonium ion are A and B. Nucleophilic displacement
of N2 is principally governed by the relative concentration
of nucleophile, which in most cases is H2O, a transforma-
tion which generates ethylene glycol (15 NuH ) OH)
here, and, from the perspective of carcinogenesis, mostly
As mentioned in the Introduction, the level of DNA
single strand breaks in vivo is significantly reduced by
the R-deuteration of NDELA. The data of Figure 6
demonstrate that only very small amounts of glycolal-
dehyde are produced from the microsomal oxidation of
R-D4NDELA compared to controls. Thus, these two sets
of observations are complementary and support the
hypothesis that R-oxidation is involved in the carcino-
genic bioactivation of NDELA. The preinduction history
of the microsomes was relatively unimportant in this
context. R-Deuteration of NDELA reduced glycolaldehyde
production with all microsomes, although the levels of
metabolic transformation were a function of the inducing
drug. The rate of NHMOR formation (Figure 6A) is not
increased above the control when the R-position is
deuterated, as might be expected if oxidation at the two
sites were competitive in roughly the same time realm.
The data basically represents “one-point kinetic experi-
ments.” Evidence for mechanistic switching by R-deu-
teration, as evidenced by enhanced NHMOR production,
must await more detailed kinetic experiments utilizing
reconstituted enzyme preparations. On the other hand,
evidence for mechanistic switching as a result of â-deu-
teration is clearly evident. â-Deuteration of NDELA
enhances the R-oxidation of NDELA with microsomes
from isoniazid-induced animals by a factor of 2. While
quantitation is more problematic in the DNA SSB (single
strand break) work, we did observe a modest enhance-
ment in DNA SSB by â-D4NDELA (20) that is also
indicative of mechanistic switching in vivo as a result of
D for H substitution in NDELA.
A possible indicator of R-hydroxylation would be the
observation of the production of DNA hydroxyethyl

464 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Loeppky and Goelzer
adducts in vivo. There are two relevant literature reports.
Farrely et al. reported evidence for the formation O6- and
N7-2-hydroxyethylguanine in DNA and RNA obtained
from the livers of rats administered high specific activity
[14C]NDELA. For DNA, the counts eluting at the same
place as the standards were so low as to raise questions
about the validity of the observation (21). RNA, however,
yielded higher levels of coeluting adducts upon acidic
hydrolysis. Through the use of an imuno-slot blot analy-
sis, Scherer et al. showed that O6-2-hydroxyethylguanine
formed in a dose responsive manner upon administration
of NDELA (30). It was also demonstrated that this same
adduct was formed in rat liver DNA following the
administration of methyl- or ethylethanolnitrosamine. In
both cases, however, the authors attributed these obser-
vations to a metabolic activation path other than the
R-hydroxylation of NDELA.
â-Oxid a tion . Others and we have shown that NDELA
undergoes â-oxidation, but the focus of this transforma-
tion has been on alcohol dehydrogenase (ADH) mediated
processes (20, 26-30, 58). The data presented here
clearly show that microsomes are also capable of mediat-
ing the â-oxidation of NDELA to NHMOR (see Scheme
4). Evidence that ADH is not responsible for this phe-
nomenon rests on the requirement for NADPH and not
NAD, the required cofactor of ADH enzymes. Moreover
ADHs are found in the cytosol and should not be in the
microsomal preparations used here. As is evident from
Table 1, and Figures 1, 2, and 6, the rate of microsomal
â-oxidation of NDELA to NHMOR is 1.5 times more rapid
than the R-hydroxylation of NDELA. While we have not
made kH/kD measurements for the microsomal oxidation
for either the â- or R-oxidation of NDELA, the data of
Figure 6 clearly show that NHMOR production is reduced
by about 50% by â-deuteration of NDELA. This effect is
not as great as that observed for R-deuteration, however.
We did not observe a significant isotope effect on the DNA
SSB produced by â-D4NDELA, but there are several
possible explanations for these observations. NDELA is
not a particularly good substrate for horse liver ADH,
but a comparison of â-oxidation rates for NDELA by ADH
and the P450 enzymes, likely responsible for the microso-
mal oxidation reported here must await further work.
Ethanol is oxidized to both acetaldehyde and acetic acid
by P450 2E1 (59, 60), so it is not unusual that the alcohol
functional group of NDELA should be oxidized by mi-
crosomes. As is discussed below, this â-oxidation process
leads to further DNA damaging species.
still higher than those used by others in some of the 2E1
work. Nevertheless, we believe the good correlations of
activity observed here, are meaningful. In all cases the
levels of glycolaldehyde and NHMOR formation cor-
related with the amount of formaldehyde formation from
DMN. With respect to the preinduction protocol, the
conversion level was in the order isoniazid ≈> strepto-
zocin > phenobarbital > no preinduction. This order and
correlation with DMN metabolism strongly suggests that
cytochrome P450 2E1 is the microsomal enzyme respon-
sible for the metabolism of NDELA at both the R- and
â-positions. Further support for this conclusion is ob-
tained from our work on the effect of inhibitors. DCC and
MEPY are both effective inhibitors of cytochrome P450
2E1. Both of these inhibitors were demonstrated to
effectively inhibit the metabolism of DMN in our mi-
crosomes and they also inhibited the metabolism of
NDELA. Because the â-oxidation of NDELA was not
inhibited as much as was the R-hydroxylation, it is
possible that other enzymes or P450s may be involved
in this process as well.
Oxid a tion of Rela ted Su bstr a tes. The data pre-
sented in Figure 3 and Table 1 shows that glyoxal is
produced in small amounts from the microsomal oxida-
tion of NDELA. Incubation of guanine nucleosides, nucle-
otides, or DNA in vitro with NHMOR produces glyoxal-
guanine (gG) adducts by means of an unknown mechanism
which does not involve any other activation process (29,
53, 58, 63, 64). Moreover, as is demonstrated in the
following paper in this issue (64), gG adducts are formed
in rat liver DNA upon administration of either NDELA
or NHMOR, as well as structurally related nitrosamines
(53). Glyoxal has been known to be an effective mutagen
for many years, but it is not a carcinogen, probably
because its oral administration results in its metabolic
detoxication. But, because it is a significant DNA adduct
from the carcinogenic NDELA and because it is found in
small amounts in the microsome-mediated transforma-
tions of NDELA, we have investigated its formation from
other substrates. The formation of glyoxal from the
microsomal metabolism of N-nitrosomorpholine, which
also gives NHMOR has been reported (65, 66), but it is
also known and we show here that the air oxidation of
glycolaldehyde produces glyoxal. A comparison of the rate
of this process, however, with the microsomal oxidation
of the same substrate (see Figure 4B and Scheme 6)
Sch em e 6
Na tu r e of th e En zym e. Dimethylnitrosamine is
known to be R-hydroxylated by cytochrome P450 2E1 (25,
35, 61, 62). This enzyme can be induced in rat liver by
the preadministration of isoniazid or streptozocin (31).
Using our assay we compared the activation of DMN with
that of NDELA and other nitrosamines by different
microsome preparations (no induction, isoniazid, strep-
tozocin, or phenobarbital-induced). We have taken the
level of product formation (aldehyde) to be indicative of
microsome/enzyme(s) capacity to metabolize the sub-
strate. Comparisons of formaldehyde production from
DMN with aldehyde production from NDELA are shown
in Figure 3. Some experimentation supports the idea that
there are two DMN metabolizing enzymes present in
microsomes, a low Km enzyme presumed to be 2E1 and a
higher Km enzyme. While the concentrations of DMN
used in our experiments are significantly lower than the
concentration of NDELA, the DMN concentrations are
shows that it is slow. The microsomal oxidation of
NHMOR also results in the generation of glycoladehyde
and glyoxal (Figure 4A). As shown in Scheme 7, R-hy-
Sch em e 7

R- and â-Oxidation of N-Nitrosodiethanolamine
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 465
droxylation of NHMOR at C-3 leads to the direct forma-
tion of glyoxal 22 through the decomposition of the
R-hydroxynitrosamine 23. This process also results in the
formation of the potent alkylating agent 10. On the other
hand, R-hydroxylation at C-5 produces the unstable
R-hydroxynitrosamine 24, the decomposition of which
results in the formation of glycolaldehyde 11 and the
diazonium ion 25 (Scheme 8). The rate of glycolaldehyde
2-hydroxyethyl group. The microsomal oxidation of NEE-
LA produces the aldehydes shown in Scheme 9, as has
been discussed above. Like NDELA, NEELA is oxidized
at both the R- and â-positions of the 2-hydroxyethyl chain.
The former process gives the R-hydroxynitrosamine 28
while the latter gives the nitrosamino aldehyde 29. It is
also R-hydroxylated on the ethyl chain to give the
R-hydroxynitrosamine 27 that decomposes to acetalde-
hyde 30 and the diazonium ion 10. Compared to NDELA,
Sch em e 8
Sch em e 9
formation from NHMOR is greater than the rate of
formation of glyoxal, but these data can only be consid-
ered to be “apparent rates”. The reason for this is also
revealed in the data of Table 1, where we see that the
rate of microsomal oxidation of glycoladehyde to glyoxal
is more than 10 times faster than the conversion of
NHMOR to this aldehyde. In short, glycoladehyde pro-
duced from either NDELA or NHMOR can be further
oxidized to glyoxal. Not only that, the hydrolysis of the
diazonium ions 10 or 25 resulting from the R-hydroxyni-
trosamines 9, 23, or 24 produced from the microsomal
oxidation of NDELA or NHMOR can lead to glyoxal by
oxidation of glycoladehyde. This is even true for ethylene
glycol 21 (Scheme 6), the hydrolysis product of the
diazonium ion 10. We demonstrated that 21 is oxidized
to glycoladehyde and that this process requires NADPH
not NAD. In fact, based on our work with inhibitors and
preinducers all of the oxidations discussed in this para-
graph appear to be mediated by a P450 enzyme, probably
2E1 (Figure 8). The properties of the diazonium ion 25
have not been explored in our laboratory yet, but its
hydrolysis is expected to give glycolaldehyde, and thence,
by further oxidation, glyoxal.
NEELA is a better substrate for R-hydroxylation (acetal-
dehyde and glycolaldehyde) by a factor of 5 as measured
by product yields under comparable conditions. As is
evident from Figure 7, R-hydroxylation on the ethyl chain
predominates. Although glycolaldehyde, produced by
R-hydroxylation on the other chain, is formed only from
one chain, rates are still nearly twice the glycolaldehyde
formation from NDELA. This demonstrates the fact that
this more lipophilic nitrosamine is a much better sub-
strate for P450.
The data of Figures 7 and 8 provide good evidence that
P450s responsible for the metabolism of NEELA and
NDELA are likely the same since the effect of inducers
and inhibitors are similar and once again implicate
cytochrome P450 2E1 as the major enzyme responsible
for the oxidative metabolism. The processes depicted in
Scheme 9 provide evidence for pathways leading to DNA
damage produced upon the in vivo administration of 26
to rats. It has been shown to give both gG and O6-2-
hydroxyethylguanine adducts.
A reading of the literature on the mechanism of
ethylene glycol toxicity leads one to believe that the
primary bio-oxidative path for glycolaldehyde, presumed
to arise principally from ADH mediated metabolism, is
its oxidation to glycolic acid through aldehyde dehydro-
genase catalysis (67). The possible role of glyoxal in
ethylene glycol toxicity has been ignored. But, the
enzymatic oxidation of ethylene glycol through glycola-
ldehyde to glyoxal has been reported recently and is the
basis of an industrial process (68, 69). We have been
unable to find any previous reports of the microsomal
oxidation of glycolaldehyde to glyoxal. Our work demon-
strates that microsomal oxidation of ethylene glycol, and
glycolaldehyde to glyoxal may be important in the toxic
action of these compounds.
Sim u lta n eou s P r od u ction of Tw o DNA-Bin d in g
F r a gm en t s fr om t h e Sa m e Molecu le (A Bid en t
Ca r cin ogen ). A simple nitrosamine such as dimethylni-
trosamine or ethylmethylnitrosamine, can undergo R-hy-
droxylation at either carbon next to the nitrogen, but
decomposition of these R-hydroxynitrosamines will give
nonbindings aldehydes (formaldehyde or acetaldehyde)
and a single diazonium ion which leads to DNA adducts.
The data described and discussed above clearly show that
all four of the carbon atoms from a single molecule of
NDELA can become bound to DNA as two adducts. One
two-carbon fragment binds as glyoxal to form gG adducts.
The other two-carbon fragment, arises from the 2-hy-
droxyethyldiazonium ion, which alkylates various sites
in DNA, giving, for example, O6-2-hydroxyethylguanine
adducts. We define a bident (two-toothed) carcinogenic
compound as one which simulataneaously gives two
DNA-binding fragments from the same molecule upon
metabolic activation. In the case of NDELA, and related
â-oxidized nitrosamines, the process requires both R- and
â-oxidation of one of the two carbon fragments. This can
Micr osom a l Oxid a tion of Eth yleth a n oln itr osa m -
in e. Ethylethanolnitrosamine NEELA 26, a compound
that differs from NDELA only in the absence of a
terminal hydroxyl group, is a potent rat renal carcinogen,
yet, despite numerous carcinogenesis experiments, rela-
tively little is known about its mode of bioactivation. In
this study, it has served as a convenient comparison to
NDELA and provides us with valuable knowledge about
the activation processes of nitrosamines carrying the

466 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Loeppky and Goelzer
occur by the processes depicted in Schemes 6, 7, and 8
or combinations thereof. The route to the “alkylating”
fragments differs depending upon whether NDELA is
R-hydroxylated or â-oxidized to give the aldehyde
(NHMOR) first. R-Hydroxylation gives 10 which decom-
poses to the 2-hydroxyethyldiazonium ion and glycolal-
dehyde, which we have shown to be more rapidly oxidized
to glyoxal by microsomes than the parent nitrosamine is
R-hydroxylated. Hydrolytic decomposition of the 2-hy-
droxyethyldiazonium gives ethylene glycol, which we
have shown can be oxidized to glyoxal by microsomal
preparations. Microsomal or ADH-mediated oxidation of
NDELA at the primary alcohol produces NHMOR. This
compound can be R-hydroxylated at either of two posi-
tions. R-Hydroxylation at C-3 (Scheme 7) leads ultimately
to the 2-hydroxyethyldiazonium ion and glyoxal after
decomposition of the R-hydroxynitrosamine. If oxidation
occurs at C-5 (Scheme 8), then the R-hydroxynitrosamine
will give glycolaldehyde and the 2-oxoethyldiazonium ion
25. We have not yet determined whether this latter
pathway is operative, but the hydrolysis of 25 gives
glycolaldehyde that can be further oxidized to glyoxal.
Thus, there are a number of biochemical routes leading
to the formation of glyoxal from NDELA.
tered dose is excreted unchanged in the urine (47, 49,
71-74). Yet several dose response studies have shown
NDELA to be a very potent carcinogen (1, 3, 4). The
“double alkylating” capacity of NDELA could be respon-
sible for its “bident carcinogenic” character and its high
effective potency. The two teeth (bident) in this case are
the two DNA adducts arising from the metabolic conver-
sion of each two-carbon fragment into a DNA binding
moiety. Our work with NEELA 26, shows that this bident
carcinogenic capacity is not limited to NDELA or NHMOR.
Reference to Scheme 9 shows that the pathways going
from 26 f 28 f 11 + 31, and from 26 f 29 (followed by
R-hydroxylation on the 2-oxoethyl group) can be bident,
but the more prevalent process which involves the
R-hydroxylation of the ethyl chain (26 f 29) cannot. In
principle, any nitrosamine carrying a 2-hydroxyethyl
chain (an ethanol nitrosamine) bound to N can be a
bident carcinogen. This is also true of N-nitrosomorpho-
line, because it is known to be metabolized in part to
NHMOR (65, 66, 75).
The emphasis of this discussion has been on the
microsomal oxidation pathways that lead to the DNA
binding of carbon fragments from NDELA, NHMOR, and
NEELA. But, we have demonstrated through in vitro
experiments that NHMOR and other R-nitrosaminoal-
dehydes, e.g., 25 can deaminate the amino groups of
guanine, adenine, and cytosine bases in DNA (54, 76) by
a process which is believed to involve NO transfer (37,
58). This process does not lead to adducts, per se, but
modified bases which will mispair. At present, we have
no method available to assess the importance of this
pathway in vivo compared to the metabolic transforma-
tions that we have delineated here.
While bident DNA-binding capacity has been designed
into several well-known anti-cancer drugs such as bis-
chloroethylnitrosourea, certain nitrogen mustards, and
other compounds, the “double alkylation”, bident carci-
nogenic, capacity of NDELA, and related â-oxidized
nitrosamines, has been observed for one other carcinogen
that we know of, the highly potent carcinogen methylvi-
nylnitrosamine 31 (Scheme 10). Good evidence has been
Of course, further oxidative or reductive metabolism
of the molecules generated from the microsomal oxidation
of NDELA can enter detoxification pathways. N-Nitroso-
2-hydroxyethylglycine, the oxidation product of the masked
aldehyde group of NHMOR, does not appear to be
genotoxic and is excreted in the urine of animals fed
NDELA. NHMOR is also known to undergo ADH medi-
ated reduction to NDELA. We can reasonably expect that
both glycolaldehyde and glyoxal, once they diffuse or are
transported to the cytosol, will be substrates for aldehyde
dehydrogenase enzymes which will convert them into
their more easily excreted carboxylic acids. Additionally
further microsomal oxidation of these compounds to
nontoxic products is highly likely. P450 2E1 has been
shown to catalyze the oxidation of ethanol to acetalde-
hyde and further to acetic acid without loss of acetalde-
hyde from the enzyme pocket (59, 60). A similar process
could be occurring with NHMOR arising from NDELA,
but competitive oxidation at carbons 3 or 5 is likely to
yield DNA binding fragments as shown here. The latter
paths are expected to lead to carcinogenesis. Such a
phenomenon, that is, further oxidation of NHMOR at
carbons 3 or 5 before it escapes from the P450 active site,
could explain why Hecht et al. failed to observe carcino-
genic activity for orally administered NHMOR (77). It is
possible that most of this material is detoxified by
cytosolic enzymes before it reaches the endoplasmic
reticulum where high concentrations of P450 are located.
Although the carcinogenesis experiments can be criticized
because of the relatively low doses used, competitive
detoxifying oxidation of NHMOR and its excretion as
N-nitroso-2-hydroxyethylglycine could prevent its reach-
Sch em e 10
presented for its P450-mediated oxidation to the epoxide
32 (70). Attack of a nucleophilic nitrogen atom in a DNA
base will result in the formation of an unstable DNA
bound R-hydroxynitrosamine 33. Decomposition will lead
to etheno adducts and the methyldiazonium ion 34 which
will readily alkylate other nucleophilic sites in DNA. This
chemistry has been demonstrated in vitro by Okazaki et
al. (70). Even if the epoxide is hydrolyzed (32 f 36) prior
to reaching DNA it still has the capacity of being a bident
carcinogen because the R-hydroxynitrosamine 36 can
ultimately give rise to both glyoxal and the methyldia-
zonium ion. The metabolite 36 is produced by the
R-hydroxylation of N-nitrosomethylethanolamine 35, as
well. The bident carcinogenic nature of methylvinylnit-
rosamine may be responsible for its extreme carcinoge-
nicity.
An interesting feature of the carcinogenicity of NDELA
compared to other nitrosamines is that, even at the
lowest dose studied, a large percentage of the adminis-

R- and â-Oxidation of N-Nitrosodiethanolamine
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 467
(12) Westin, J . B., Spiegelhalder, B., Preussmann, R., and Shani, J .
(1990) Assay of suntan lotions for the carcinogenic, non-volatile
N-nitrosamine N-nitrosodiethanolamine. Cancer Lett. 50, 157-
160.
ing the same cell sites which permit its microsomal
oxidation to produce reactive electrophiles.
(13) Rappe, C., and Zingmark, P. A. (1978) Formation of N-nitro-
samines in cutting fluids. Int. Agency Res. Cancer Sci. Publ. 19,
213-217.
Con clu sion
(14) Fan, T. Y., Morrison, J ., Rounbehler, D. P., Ross, R., Fine, D. H.,
Miles, W., and Sen, N. P. (1977) N-Nitrosodiethanolamine in
synthetic cutting fluids: a part-per-hundred impurity. Science
196, 70-71.
(15) Monarca, S., Scassellati Sforzolini, G., Spiegelhalder, B., Pasquini,
R., and Fatigoni, C. (1993) Monitoring nitrite, N-nitrosodietha-
nolamine, and mutagenicity in cutting fluids used in the metal
industry. Environ. Health Perspect. 101, 126-128.
(16) Keefer, L. K., Goff, U., Stevens, J ., and Bennett, E. O. (1990)
Persistence of N-nitrosodiethanolamine contamination in Ameri-
can metal-working lubricants. Food Chem. Toxicol. 28, 531-
534.
(17) Fadlallah, S., Cooper, S. F., Fournier, M., Drolet, D., and Perrault,
G. (1990) Determination of N-nitroso compounds in metalworking
fluids. J . Chromatogr. Sci. 28, 517-523.
(18) Fuchs, J ., Burg, J ., Hengstler, J . G., Bolm-Audorff, U., and Oesch,
F. (1995) DNA damage in mononuclear blood cells of metal
workers exposed to N-nitrosodiethanolamine in synthetic cutting
fluids. Mutat. Res. 342, 95-102.
(19) Loeppky, R. N. (1999) The mechanism of bioactivation of N-
nitrosodiethanolamine. Drug Metab. Rev. 31, 175-193.
(20) Loeppky, R. N., Fuchs, A., J anzowski, C., Humberd, C., Goelzer,
P., Schneider, H., and Eisenbrand, G. (1998) Probing the mech-
anism of the carcinogenic activation of N-nitrosodiethanolamine
with deuterium isotope effects: In vivo induction of DNA single
strand breaks and related in vitro assays. Chem. Res. Toxicol.
11, 1556-1566.
(21) Farrelly, J . G., Thomas, B. J ., and Lijinsky, W. (1987) Metabolism
and cellular interactions of N-nitrosodiethanolamine. IARC Sci-
entific Pub. 84, 87-90.
(22) Farrelly, J . G., Stewart, M. L., and Lijinsky, W. (1984) The
metabolism of nitrosodi-N-propylamine, nitrosodiallylamine and
nitrosodiethanolamine. Carcinogenesis (London) 5, 1015-
1019.
(23) Farrelly, J . G. (1980) A new assay for the microsomal metabolism
of nitrosamines. Cancer Res. 40, 3241-3244.
(24) Loeppky, R. N., and Xiong, H. (1994) The synthesis of deute-
rium-labeled N-nitrosodiethanolamine and N-nitroso-2-hydroxy-
morpholine. J . Labelled Compd. Radiopharm. 24, 1099-
1110.
(25) Yoo, J . S. H., Ishizaki, H., and Yang, C. S. (1990) Roles of
cytochrome P450IIE1 in the dealkylation and denitrosation of
N-nitrosodimethylamine and N-nitrosodiethylamine in rat liver
microsomes-9. Carcinogenesis (London) 11, 2239-2243.
(26) Eisenbrand, G., Denkel, E., and Pool, B. (1984) Alcohol dehydro-
genase as an activating enzyme for N-nitrosodiethanolamine
(NDELA): in vitro activation of NDELA to a potent mutagen in
Salmonella typhimurium. J . Cancer Res. Clin. Oncol. 108, 76-
80.
(27) Denkel, E., Pool, B. L., Schlehofer, J . R., and Eisenbrand, G. (1986)
Biological activity of N-nitrosodiethanolamine and of potential
metabolites which may arise after activation by alcohol dehydro-
genase in Salmonella typhimurium, in mammalian cells, and in
vivo. J . Cancer Res. Clin. Oncol. 111, 149-153.
(28) Denkel, E., Sterzel, G., and Eisenbrand, G. (1987) In-vivo activa-
tion of N-nitrosodiethanolamine and other N-nitroso-2-hydroxy-
alkylamines by alcohol dehydrogenase and sulfotransferase. IARC
Scientific Pub. No. 84, 83-86.
(29) Loeppky, R. N., Tomasik, W., Denkel, E., and Eisenbrand, G.
(1987) R-Nitrosaminoaldehydes: Highly reactive metabolites.
IARC Sci. Publ. 84, 94-99.
(30) Scherer, G., Ludeke, B., Kleihues, P., Loeppky, R. N., and
Eisenbrand, G. (1991) Mutagenicity, DNA damage and DNA
adduct formation by N-nitroso-2-hydroxyalkylamine and corre-
sponding aldehydes. IARC Sci. Publ. 105, 339-342.
(31) Yamazaki, H., Oda, Y., Funae, Y., Imaoka, S., Inui, Y., Guenger-
ich, F. P., and Shimada, T. (1992) Participation of rat liver
cytochrome P 450 2E1 in the activation of N-nitrosodimethy-
lamine and N-nitrosodiethylamine to products genotoxic in an
acetyltransferase-overexpressing Salmonella typhimurium strain
(NM2009). Carcinogenesis 13, 979-983.
Contrary to all previous reports, we have unequivocally
demonstrated that NDELA undergoes microsome-medi-
ated R-hydroxylation producing glycolaldehyde and pre-
sumably the short-lived 2-hydroxyethyldiazonium ion.
This process competes slightly less favorably with the
â-oxidation of the alcohol group to give the masked
aldehyde NHMOR, a hemiacetal. In this system, glyco-
laldehyde is further oxidized to glyoxal, a known DNA-
binding molecule. Since NDELA is known to produce both
2-hydroxyethyl- and gG DNA adducts in vivo, we propose
that it is a bident carcinogen where all four carbons of
the molecule can become bound to DNA as two separate
adducts and this process may account for the potency of
this carcinogen. NHMOR has also been shown to undergo
microsome-mediated oxidation to give glyoxal and gly-
colaldehyde in addition to diazonium ions. Thus, this
compound can also act as a bident carcinogen, as can any
alkyl-2-hydroxyethylnitrosamine, as we have shown here
with NEELA.
Ack n ow led gm en t. We gratefully acknowledge the
support of this research by a grant, RO1 ES03953, from
the National Institute of Environmental Health Sciences
(NIH, PHS).
Refer en ces
(1) Preussmann, R., Habs, M., Habs, H., and Schmaehl, D. (1982)
Carcinogenicity of N-nitrosodiethanolamine in rats at five differ-
ent dose levels. Cancer Res. 42, 5167-5171.
(2) Lijinsky, W., Reuber, M. D., and Manning, W. B. (1980) Potent
carcinogenicity of nitrosodiethanolamine in rats. Nature (London)
288, 589-590.
(3) Lijinsky, W., and Reuber, M. D. (1984) Dose-response study with
N-nitrosodiethanolamine in F344 rats. Food Chem. Toxicol. 22,
23-26.
(4) Hoffmann, D., Rivenson, A., Adams, J . D., J uchatz, A., Vinchkoski,
N., and Hecht, S. S. (1983) Effects of route of administration and
dose on the carcinogenicity of N-nitrosodiethanolamine in the
Syrian golden hamster. Cancer Res. 43, 2521-2524.
(5) Hilfrich, J ., Schmeltz, I., and Hoffmann, D. (1978) Effects of
N-nitrosodiethanolamine and 1,1-diethanolhydrazine in Syrian
golden hamsters. Cancer Lett. 4, 55-60.
(6) Fan, T. Y., Goff, U., Song, L., Fine, D. H., Arsenault, G. P., and
Biemann, K. (1977) N-Nitrosodiethanolamine in cosmetics, lotions
and shampoos. Food Cosmet. Toxicol. 15, 423-430.
(7) Havery, D., C, and Chou, H. J . (1994) Nitrosamines in sunscreens
and cosmetic products. In Nitrosamines and related N-nitroso
compounds: Chemistry and biochemistry (Loeppky, R. N., and
Michejda, C. J ., Eds.) pp 20-33, American Chemical Soc.,
Washington, DC.
(8) Gettings, S. D., Dressler, W. E., Franz, T. J ., Kelling, C. K., Howes,
D., and Walters, K. A. (1995) Assessing risk from N-nitrosamines
in cosmetic products. Cosmet. Toiletries 110, 72-80.
(9) Fine, D. H. (1983) HPLC-TEA determination of N-nitrosodietha-
nolamine (NDELA) and similar compounds in cosmetics. Inter-
national Agency for Research in Cancer Scientific Publication, p
45.
(10) Rollmann, B., Lombart, P., Rondelet, J ., and Mercier, M. (1981)
Determination of N-nitrosodiethanolamine in cosmetics by gas
chromatography with electron capture detection. J . Chromatogr.
206, 158-163.
(11) Sommer, H., Blankart, M., and Eisenbrand, G. (1989) Determi-
nation of N-nitrosodiethanolamine in cosmetics and in alkanola-
mines. Results of collaborative studies. Z. Lebensm. Unters.
Forsch. 189, 144-146.
(32) Shu, L., and Hollenberg, P. F. (1997) Alkylation of cellular
macromolecules and target specificity of carcinogenic nitrosodi-
alkylamines: metabolic activation by cytochromes P450 2B1 and
2E1. Carcinogenesis 18, 801-810.

468 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Loeppky and Goelzer
(33) Chen, T. L., Chen, S. H., Tai, T. Y., Chao, C. C., Park, S. S.,
Guengerich, F. P., and Ueng, T. H. (1996) Induction and sup-
pression of renal and hepatic cytochrome P450-dependent mo-
nooxygenases by acute and chronic streptozotocin diabetes in
hamsters. Arch. Toxicol. 70, 202-208.
(34) Sohn, O. S., Ishizaki, H., Yang, C. S., and Fiala, E. S. (1991)
Metabolism of azoxymethane, methylazoxymethanol and N-
nitrosodimethylamine by cytochrome P450IIE1. Carcinogenesis
12, 127-131.
(35) Yang, C. S., Yoo, J . S. H., Ishizaki, H., and Hong, J . (1990)
Cytochrome P450IIE1: roles in nitrosamine metabolism and
mechanisms of regulation. Drug Metab. Rev. 22, 147-159.
(36) Lauriault, V. V., Khan, S., and O’Brien, P. J . (1992) Hepatocyte
cytotoxicity induced by various hepatotoxins mediated by cyto-
chrome P-450IIE1: protection with diethyldithiocarbamate ad-
ministration. Chem.-Biol. Interact. 81, 271-289.
(37) Loeppky, R. N., Tomasik, W., Kovacs, D. A., Outram, J . R., and
Byington, K. H. (1984) Alternative bioactivation routes for â-hy-
droxynitrosamines. Biochemical and chemical model studies. In
N-Nitroso Compd: Occurrence, Biol. Eff. Relevance Hum. Cancer
(Bartsch, H., O’Neill, I., and VanBorstal, Eds.) pp 429-436, IARC,
Lyon.
(38) Loeppky, R. N., Srinivasan, A., and Erb, E. (1992) Putative
proximate carcinogens derived from ethanolnitrosamines: Chemi-
cal properties. In Nitroso Compounds: Biological Mechanisms,
Exposures and Cancer Etiology (Bartsch, H., and O’Neill, I., Eds.)
pp 42a-42b, IARC, Lyon.
(39) Eisenbrand, G., and J anzowski, C. (1994) Potential mechanism
of action of nitrosamines with hydroxy, oxo, or carboxy groups.
In Nitrosamines and Related N-nitroso Compounds: Chemistry
and Biochemistry (Loeppky, R. N., and Michejda, C., J , Eds.) pp
179-194, American Chemical Society, Washington, DC.
(40) Sterzel, W., and Eisenbrand, G. (1986) N-Nitrosodiethanolamine
is activated in the rat to an ultimate genotoxic metabolite by
sulfotransferase. J . Cancer Res. Clin. Oncol. 111, 20-24.
(52) Park, M., Gu, F., and Loeppky, R. N. (1998) The synthesis of
highly reactive multifunctional R,â-epoxy- and R-acetoxy-nitro-
samines. Tetrahedron Lett. 39, 1287-1290.
(53) Loeppky, R. N., Cui, W., Goelzer, P., Park, M., and Ye, Q. (1999)
Glyoxal-guanine DNA adducts: Detection, stability and formation
from nitrosamines. In Cyclic DNA Adducts II (Singer, B., and
Bartsch, H., Eds.) pp 155-168, IARC, Lyon.
(54) Park, M., and Loeppky, R. N. (2000) In Vitro DNA Deamination
by R-Nitrosaminoaldehydes Determined by GC/MS-SIM Quan-
titation. Chem. Res. Toxicol. 13, 72-81.
(55) Rao, T. K., Young, J . A., Lijinsky, W., and Epler, J . L. (1979)
Mutagenicity of aliphatic nitrosamines in Salmonella typhimu-
rium. Mutat. Res. 66, 1-7.
(56) Lijinsky, W., and Andrews, A. W. (1983) The superiority of
hamster liver microsomal fraction for activating nitrosamines to
mutagens in Salmonella typhimurium. Mutat. Res. 111, 135-
144.
(57) McGarrity, J . F., and Smyth, T. (1980) Hydrolysis of diazometh-
anesKinetics and mechanism. J . Am. Chem. Soc. 102, 7303-
7308.
(58) Loeppky, R. N., Tomasik, W., and Kerrick, B., E. (1987) Nitroso
transfer from R-nitrosamino aldehydes: Implications for carcino-
genesis. Carcinogenesis (London) 8, 941-946.
(59) Bell-Parikh, L. C., and Guengerich, F. P. (1999) Kinetics of
cytochrome P450 2E1-catalyzed oxidation of ethanol to acetic acid
via acetaldehyde. J . Biol. Chem. 274, 23833-23840.
(60) Bell, L. C., and Guengerich, F. P. (1997) Oxidation kinetics of
ethanol by human cytochrome P450 2E1. Rate-limiting product
release accounts for effects of isotopic hydrogen substitution and
cytochrome b5 on steady-state kinetics. J . Biol. Chem. 272,
29643-29651.
(61) Tu, Y. Y., Peng, R., Chang, Z. F., and Yang, C. S. (1983) Induction
of a high affinity nitrosamine demethylase in rat liver microsomes
by acetone and isopropanol. Chem.-Biol. Interact. 44, 247-
260.
(41) Michejda, C. J ., Koepke, S. B., Kroeger, M. B., and Hernandez,
L. (1994) Activation of â-hydroxyalkylnitrosamines. Evidence for
involvement of a sulfotransferase. In Nitrosamines and related
N-nitroso compounds: Chemistry and Biochemistry (Loeppky, R.
N., and Michejda, C. J ., Eds.) pp 195-210, Am. Chem. Soc.,
Washington, DC.
(42) Koepke, S. T. R., Kroeger-Koepke, M. B., Born, W., Thomas, B.
J ., Alvord, W. G., and Michejda, C. J . (1988) Alkylation of DNA
in rats by N-nitrosomethyl-(2-hydroxyethyl)amine: Dose respose
and persistence of the alkylated lesions in vivo. Cancer Res. 48,
1537-1542.
(43) Kroeger-Koepke, M. B., Koepke, S. R., Hernandez, L., and
Michejda, C. J . (1992) Activation of a â-hydroxyalkylnitrosamine
to alkylating agents: Evidence for the involvement of a sul-
fotransferase. Cancer Res. 52, 3300-3305.
(44) Koepke, S. R., Cresia, D. R., Knutsen, G. L., and Michejda, C. J .
(1988) Carcinogenicity of hydroxyalkylnitrosamines in F344
rats: Contrasting behavior of â- and γ-hydroxyalkyl nitrosamines.
Cancer Res. 48, 1533-1536.
(45) Airoldi, L., Bonfanti, M., Benfenati, E., Tavecchia, P., and Fanelli,
R. (1983) Identification of an acidic metabolite of N-nitrosodi-
ethanolamine isolated from rat urine. Biomed. Mass Spectrom.
10, 334-337.
(46) Airoldi, L., Bonfanti, M., Fanelli, R., Bove, B., Benfenati, E., and
Gariboldi, P. (1984) Identification of a nitrosamino aldehyde and
a nitrosamino acid resulting from â-oxidation of N-nitrosodietha-
nolamine. Chem.-Biol. Interact. 51, 103-113.
(47) Airoldi, L., Macri, A., Bonfanti, M., Bonati, M., and Fanelli, R.
(1984) Kinetics and bioavailability of N-nitrosodiethanolamine
after intravenous and cutaneous administration to rats. Food
Chem. Toxicol. 22, 133-138.
(48) Bonfanti, M., Castelli, M. G., Fanelli, R., and Airoldi, L. (1985)
Enterohepatic recycling of N-nitrosodiethanolamine in rats. Food
Chem. Toxicol. 23, 1011-1013.
(62) Tu, Y. Y., and Yang, C. S. (1983) High-affinity nitrosamine
dealkylase system in rat liver microsomes and its induction by
fasting. Cancer Res. 43, 623-629.
(63) Chung, F. L., and Hecht, S. S. (1985) Formation of the cyclic 1,-
N2-glyoxal-deoxyguanosine adduct upon reaction of N-nitroso-2-
hydroxymorpholine with deoxyguanosine. Carcinogenesis (Lon-
don) 6, 1671-1673.
(64) Loeppky, R. N., Ye, Q., Goelzer, P., Chen, Y., and Li, E. (2002)
Detection of glyoxal-guanine DNA adducts produced in vivo from
â-oxidized nitrosamines by 32P-postlabeling. Chem. Res. Toxicol.
15, 470-482.
(65) Manson, D., Cox, P. J ., and J arman, M. (1978) Metabolism of
N-nitrosomorpholine by the rat in vivo and by rat liver mi-
crosomes and its oxidation by the Fenton system. Chem.-Biol.
Interact. 20, 341-354.
(66) J arman, M., and Manson, D. (1986) The metabolism of N-
nitrosomorpholine by rat liver micrsomes and ists oxidation by
the Fenton system. Carcinogenesis 7, 559-565.
(67) Wiener, H. L., and Richardson, K. E. (1988) The metabolism and
toxicity of ethylene glycol. Res. Commun. Subst. Abuse 9, 77-
87.
(68) Isobe, K. (1995) Oxidation of ethylene glycol and glycolic acid
by glycerol oxidase. Biosci., Biotechnol., Biochem. 59, 576-
581.
(69) Isobe, K., and Nishise, H. (1995) A new enzymic method for
glycolaldehyde production from ethylene glycol. J . Mol. Catal. B:
Enzymol. 1, 37-43.
(70) Okazaki, O., Persmark, M., and Guengerich, F. P. (1993) N-
Nitroso-N-methylvinylamine: Reaction of the epoxide with guanyl
and adenyl moieties to yield adducts derived from both parts of
the molecule. Chem. Res. Toxicol. 6, 168-173.
(71) Lethco, E. J ., Wallace, W. C., and Brouwer, E. (1982) The fate of
N-nitrosodiethanolamine after oral and topical administration to
rats. Food Chem. Toxicol. 20, 401-406.
(49) Bonfanti, M., Magagnotti, C., Fanelli, R., and Airoldi, L. (1986)
Comparative in vitro and in vivo â-oxidation of N-nitrosodietha-
nolamine in different animal species. Chem.-Biol. Interact. 59,
203-210.
(72) Preussmann, R., Wuertele, G., Eisenbrand, G., and Spiegelhalder,
B. (1978) Urinary excretion of N-nitrosodiethanolamine admin-
istered orally to rats. Cancer Lett. 4, 207-209.
(50) Suzuki, E., Iiyoshi, M., and Okada, M. (1981) Metabolic fate of
N-alkyl-N-(3-hydroxypropyl and 2- hydroxyethyl)nitrosamines in
the rat in relation to the induction of bladder cancer by N-butyl-
N-(4-hydroxybutyl)nitrosamine and its homologs. Gann 72, 254-
258.
(73) Preussmann, R., Spiegelhalder, B., Eisenbrand, G., Wuertele, G.,
and Hofmann, I. (1981) Urinary excretion of N-nitrosodiethano-
lamine in rats following its epicutaneous and intratracheal
administration and its formation in vivo following skin application
of diethanolamine. Cancer Lett. 13, 227-231.
(51) Mori, Y., Yamazaki, H., and Konishi, Y. (1987) Mutagenicity of
N-nitrosodiethanolamine in the Salmonella/microsome test. Mu-
tat. Res. 192, 91-94.
(74) Spiegelhalder, B., Eisenbrand, G., and Preussmann, R. (1982)
Urinary excretion of N-nitrosamines in rats and humans. IARC
Sci. Publ. 41, 443-449.

R- and â-Oxidation of N-Nitrosodiethanolamine
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 469
(75) Hecht, S. S. (1984) N-Nitroso-2-hydroxymorpholine, a mutagenic
metabolite of N-nitrosodiethanolamine. Carcinogenesis (London)
5, 1745-1747.
(76) Loeppky, R. N., Lee, M. P., and Mueller, S. (1993) DNA modifica-
tion by R-nitrosamino aldehydes in The Chemistry of DNA
Adducts (Hemminki, K., and Bartsch, H., Eds.) pp 429-432,
International Agency for Research on Cancer.
(77) Hecht, S. S., Lijinsky, W., Kovatch, R. M., Chung, F. L., and
Saavedra, J . E. (1989) Comparative tumorigenicity of N-nitroso-
2-hydroxymorpholine, N-nitrosodiethanolamine, and N-nitroso-
morpholine in A/J mice and F344 rats. Carcinogenesis (London)
10, 1475-1477.
TX000267B