Probing the Mechanism of the Carcinogenic Activation of N

Nov 17, 1998 - Deuterium isotope effects on DNA SSB levels were inversely dependent on dose: α-D4NDELA, 3.22−1.37; and β-D4NDELA, 1.38−0.79. At ...
0 downloads 19 Views 149KB Size
1556

Chem. Res. Toxicol. 1998, 11, 1556-1566

Probing the Mechanism of the Carcinogenic Activation of N-Nitrosodiethanolamine with Deuterium Isotope Effects: In Vivo Induction of DNA Single-Strand Breaks and Related in Vitro Assays Richard N. Loeppky,*,†,‡ Anne Fuchs,§ Christine Janzowski,‡,§ Christina Humberd,† Petra Goelzer,† Heiko Schneider,§ and Gerhard Eisenbrand‡,§ Department of Chemistry, University of Missouri, Columbia, Missouri 65211, and Division of Food Chemistry and Environmental Toxicology, University of Kaiserslautern, 67663 Kaiserslautern, Germany Received July 21, 1998

A series of bioassays, including in vivo induction of DNA single-strand breaks (SSB) and cytotoxicity in cytochrome P450 2E1-transfected cells, were utilized with N-nitrosodiethanolamine (NDELA), its deuterated isotopomers (R-D4NDELA and β-D4NDELA), N-nitroso-2hydroxymorpholine (NHMOR), and two of its deuterated isotopomers (2-D-NHMOR and 5,5D2-NHMOR) to probe the mechanism of carcinogenic activation of NDELA and the role of its metabolite NHMOR. DNA samples, taken from the livers of male Wistar rats 4 h after the administration of NDELA, exhibited dose-dependent DNA SSB levels over the range of 0.080.75 mmol/kg (body weight), with the greatest SSB level at the highest dose. Deuterium isotope effects on DNA SSB levels were inversely dependent on dose: R-D4NDELA, 3.22-1.37; and β-D4NDELA, 1.38-0.79. At the lowest dose of 0.15 mmol/kg (body weight), 5,5-D2-NHMOR gave an isotope effect for DNA SSB of 2.8 while that for 2-D-NHMOR was 0.7. NDELA and β-D4NDELA were equally cytotoxic to human P450 2E1-transfected V79 Chinese hamster cells, while R-D4NDELA was not. Significant DNA SSB levels were observed in these cells for NDELA and β-D4NDELA but not for R-D4NDELA. A kinetic deuterium isotope effect of 2.6 for Vmax/Km was observed for the horse liver alcohol dehydrogenase-mediated oxidation of β-D4NDELA to NHMOR, while kH/kD for R-D4NDELA was 1.05. These data provide the first definitive evidence for the activation of NDELA by a pathway involving the scission of the R-CH bond and are consistent with P450 2E1-mediated R-hydroxylation of NDELA producing the corresponding reactive R-hydroxynitrosamine.

Introduction

Scheme 1 1

N-Nitrosodiethanolamine (NDELA) is a widespread (1) and potent liver and nasal cavity carcinogen (2-6) in several species of rodents. NDELA forms from the inadvertent nitrosation of triethanolamine, diethanolamine, and its amide derivatives (Scheme 1) (7). Because these precursors are used in so many different types of formulations, NDELA has been found in cosmetics and other personal care items (8, 9), pesticides (10), tobacco (11, 12) which has been treated with alkanolaminecontaining formulations such as herbicides, and most prominently in some metal-working fluids (13, 14), in which concentrations as high as 3% have been observed (15). NDELA can easily penetrate human skin and has been found in the urine of exposed metal workers (16). Despite its wide occurrence in the human environment and its carcinogenicity, relatively little is understood

regarding the mode of bioactivation of this important carcinogen. Here, we review the conflicting data regarding its carcinogenic bioactivation and present new findings produced from the application of deuterium isotope effects on NDELA-induced in vivo DNA single-strand

* To whom correspondence should be addressed. † University of Missouri. ‡ Principal authors. § University of Kaiserslautern. 1 Abbreviations: ADH, alcohol dehydrogenase; bw, body weight; gG, glyoxal guanine adduct; NDELA, N-nitrosodiethanolamine; NHEG, N-nitroso-N-(2-hydroxyethyl)glycine; NHMOR, N-nitroso-2-hydroxymorpholine; SSB, single-strand break.

10.1021/tx9801716 CCC: $15.00 © 1998 American Chemical Society Published on Web 11/17/1998

Mechanism of N-Nitrosodiethanolamine Bioactivation Scheme 2

Scheme 3

breaks (SSB), and on in vitro P450-transfected cell cytotoxicity. The new data require revisions of hypotheses for the bioactivation of NDELA. Careful dose-response studies have shown that NDELA is a potent carcinogen in several species of animals (3, 5). NDELA has the unusual property, for a nitrosamine, of being excreted unchanged to the extent of 9060%, even at the lowest doses (17-19). Oral administration of [14C]NDELA to rats showed that only a small fraction of it is converted to CO2 (17). These same studies showed the formation of as many as three urinary metabolites, but they were not identified. The use of very high-specific activity [14C]NDELA revealed only low levels of 14C binding to DNA (20). Numerous investigations of NDELA mutagenicity have been reported, and while controversial, the vast majority of these have shown that NDELA is not a mutagen in the standard S-9-mediated Ames assay (21, 22). The failure to observe mutagenicity for NDELA in the Ames assay has been attributed to a lack of microsome-, and thus S-9-, mediated R-hydroxylation, a process which is intimately connected to the bioactivation of many nitrosamines (23). In one case, low levels of S-9-meditated mutagenicity were shown to be blocked by DMSO (22), a known inhibitor of some P450 enzymes, and a solvent used in some mutagenicity studies of NDELA. The mechanism of S-9-mediated mutagenicity (and carcinogenicity) of dialkylnitrosamines is presumed to arise from DNA alkylation by highly unstable diazonium ions produced from the facile decomposition of R-hydroxynitrosamines which result from microsomal oxidation (23, 24). This pathway (1 f 2, Scheme 2), however, has never been shown to exist for the bioactivation of NDELA. Farrelly et al. (25) compared the metabolism of three nitrosamines in microsomes and hepatocytes. While two of the nitrosamines, dipropylnitrosamine and diallylnitrosamine, were metabolized in both systems, no metabolites were observed for NDELA in microsomes and [14C]NDELA produced very low yields of several metabolites in hepatocytes but none that could be attributed to R-hydroxylation. The lack of S-9-mediated mutagenicity and the failure to observe microsomal R-hydroxylation of NDELA stimulated research on alternative bioactivation pathways. Airoldi et al. (26, 27) demonstrated that N-nitroso-N-2hydroxyethylglycine (NHEG) is a urinary metabolite of NDELA in rats (Scheme 3). Loeppky et al. (28-30) showed that R-nitrosamino aldehydes, possible interme-

Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1557

diates in the β-oxidation of ethanol nitrosamines, are unstable, transfer their NO group to other amines, form diazonium ions, and generate glyoxal equivalents from the ethanol chain. These unusual chemical properties were proposed to be capable of damaging DNA in several ways (path A of Scheme 4). N-Nitroso-2-hydroxymorpholine (NHMOR), the cyclic hemiacetal form of N-(2hydroxyethyl)-N-nitrosoethanal, the immediate oxidation product of NDELA, is produced both from rat liver S-9 incubations of NDELA (27, 31) and from the alcohol dehydrogenase (ADH)-catalyzed oxidation of NDELA (29, 32). NHMOR is an intermediate in the metabolic formation of NHEG and is a direct-acting mutagen (30, 31), and it reacts with guanosine to form 1,N2-cyclic “glyoxalguanine” (gG) adducts (30, 33, 34). Mutagenicity for NDELA is observed when it is incubated in the Ames assay with ADH in lieu of S-9 (32). NHMOR and other R-nitrosamino aldehydes deaminate guanine (30), adenine, and cytosine nucleotides, singly and in DNA, generating in vitro xanthine, hypoxanthine, and uracil residues, respectively, and produce gG adducts (34). Airoldi (35) observed a correlation between the the extent of β-oxidation of NDELA, as measured by production of metabolites NHMOR and NHEG, and its carcinogenicity in four different animal species. Both NDELA and NHMOR are effective producers of DNA SSB in rat liver DNA (36-38). The DNA SSB produced by NDELA or NHMOR are either blocked or significantly modulated by inhibitors of ADH, P450, sulfotransferase, and probably other enzymes (36-40). While the inhibitors used in these experiments are likely to have effected more than a single enzyme system, the experiments were indicative of and supportive of the β-oxidation path. While all of the observations described in the preceding paragraph substantiate the β-oxidation pathway for NDELA, other observations suggest the possible involvement of other pathways or raise serious questions regarding the viability of this route to the genotoxicity of NDELA. The mutagenicity of NHMOR is enhanced by rat liver S-9, suggesting an additional activation step possibly involving P450 (31, 38). On the basis of their experiments with inhibitors of DNA SSB produced by NDELA, NHMOR, and related nitrosamines, the Eisenbrand group suggested that NDELA first underwent ADH-mediated oxidation to NHMOR, but that this was followed by sulfation of the OH for generating an electrophilic intermediate capable of damaging DNA (path B of Scheme 4) (37, 38). By using various enzyme inhibitors to modulate the host-mediated mutagenesis of Escherichia coli produced by NDELA in rats, it also has been concluded that the activation of this carcinogen may involve more than one enzyme system (41, 42). The steps shown in path C of Scheme 4 could account for many of these observations as well as for the DNA products known to arise from NDELA in vivo. Michejda and coworkers have postulated that (2-hydroxyethyl)methylnitrosamine is activated by sulfation of the hydroxy group followed by the generation of an electrophilic 1,2,3oxadiazolinium ion akin to that shown in path D of Scheme 4 (for a review, see ref 40), but little evidence has been produced to indicate a role for such a process in the activation of NDELA. More recent and sensitive DNA adduct studies involving the in vivo administration of NDELA and NHMOR to rats show that NDELA produces O6-(hydroxyethyl)guanine adducts (36) but that NHMOR produces none

1558 Chem. Res. Toxicol., Vol. 11, No. 12, 1998

Loeppky et al. Scheme 4

(Scheme 4). In vivo metabolism experiments with NHMOR reveal both its oxidation to NHEG and its reduction to NDELA, as well as the formation of unknown metabolites (43). These experiments could be interpreted to suggest an undetected R-oxidation pathway for NDELA which results in the formation of O6(hydroxyethyl)guanine DNA adducts. The most significant challenge to the β-oxidation pathway comes from carcinogenesis experiments utilizing NHMOR. Lijinsky and Hecht found that NHMOR is not carcinogenic in F344 rats and only weakly carcinogenic in A/J mice (44). While it can be argued that administration of NHMOR in the drinking water does not result in its arrival at the proper compartment in the liver to elicit its carcinogenic action, its lack of carcinogenicity, and additional confounding factors presented above, raise doubts regarding the β-oxidation hypothesis and the mode of NDELA carcinogenic activation. Janzowski et al. have recently reported an examination of the cytotoxic action of NDELA in hamster V79 cells transfected with either P450 2E1 or P450 2B1. NDELA had no effect on control V79 cells or those transfected with P4502B1, but did produce a dose-dependent (10-

40 mM) cytotoxicity for P450 2E1-transfected cells (45). Experiments using HPRT mutagenesis as an end point in these cell lines demonstrated similar results (45). NDELA was only active in those cells transfected with P450 2E1. These results suggest that P450 2E1 may be involved in the bioactivation of NDELA. To clarify the activation pathway(s) for NDELA, we have chosen to examine the effect of deuterium substitution in NDELA and NHMOR on DNA SSB in vivo. Deuterium isotope effects have previously been observed in nitrosamine carcinogenesis (46, 47), and a kH/kD has been determined for the P450-mediated oxidation of dimethylnitrosamine (48, 49). Although the DNA SSB assay is a relatively crude assay, sometimes viewed more as a marker of acute toxicity rather than as an indicator of genotoxicity, at the outset of this investigation, there were few other bioassays available for NDELA due to its apparent low degree of DNA binding. We have prepared (50) deuterated isotopomers of NDELA and NHMOR for in vivo administration to rats. The simple initial hypothesis is that deuterium isotope effects will be observed for NDELA and NHMOR-induced DNA SSB. If R-hydroxylation is important, then deuterium substitution at

Mechanism of N-Nitrosodiethanolamine Bioactivation

that position should diminish the extent of DNA SSB. Given the body of data supporting β-oxidation, we expected to observe a significant deuterium isotope effect for substitution at that position. On the other hand, if the bioactivation is solely dependent upon sulfation, then no isotope effect should be observed because no C-H bonds are broken in this activation scheme.

Materials and Experimental Methods Materials. Alcohol dehydrogenase from equine liver, crystallized and lyophilized with 99.5%. The isotopic purity of the deuterated nitrosamines was >98%. Materials for cytotoxicity assays utilized DMEM (Gibco, Rockville, MD), supplemented with 10% NCS (52), newborn calf serum (NCS) (Gibco), sodium chloride (Roth, Farmingdale, NY), and 0.33% neutral red in PBS (Sigma), diluted 5-fold in sodium chloride (0.9%) before use. Caution: NDELA is a potent animal carcinogen. NHMOR is a mutagen, and all nitrosamines should be treated as carcinogenic unless known otherwise. To destroy nitrosamines, aqueous solutions are treated with hydrogen formed from small pieces of aluminum foil added to the alkaline solution. Organic solutions are treated with 10% HBr in glacial acetic acid. DNA SSB Analysis. (1) In Vivo Experiments with Wistar Rats. Male Wistar rats (Central Institute for Research Animal Science, Hannover, Germany), which were 8-10 weeks old (200 g), were kept fasting for 18 h. Then the nitrosamines [0.08, 0.15 (0.25), 0.37, and 0.75 mmol/kg of body weight (bw) dissolved in water] were administered by gavage, and the animals were sacrificed after 4 h. The liver was flushed with Dulbecco’s PBS (pH 7.4) containing 0.53 mM Na2EDTA, removed, and homogenized in a hand potter. The suspension was filtered through gauze (30 µm), and nuclei were counted with trypan blue. Aliquots containing 1.5 × 106 nuclei were loaded on 2 µm polycarbonate filters and submitted to alkaline filter elution using a slight modification of the procedure of Sterzel et al. (53). After the lysis, the filters were washed with 20 mL of EDTA buffer, and the DNA was eluted with an alkaline buffer adjusted to pH 12.15. Ten fractions were collected, and the DNA was analyzed with an automated system using fluorochrome 33258. Prior to the initiation of the DNA SSB assay, the highest in vivo dose of nitrosamine which was not significantly toxic to hepatocytes was determined as follows. Hepatocytes were isolated as published previously (54), following a protocol similar to that described above except that collagenase IV (50 mg/100 mL) was used. Viability was determined with trypan blue, and the hepatocytes were also used for alkaline filter elution. The highest working dose (millimoles per kilogram of bw) for NDELA (0.775) and NHMOR (0.37) was established by this means. (2) Treatment of Data. Two different animals were utilized on different days for each dose level and each nitrosamine. The amount of eluted DNA in the 10 fractions (portions) was summed and added to the amount of DNA remaining on the filter. The fraction (amount) of eluted DNA at each time point was subtracted from the total amount of DNA to give the fraction (amount) of DNA on the filter. Four determinations were made for each run and the data averaged. Linear regression of log(fraction of DNA on the filter) versus the fraction number (volume of eluent) for fractions 3-9 yielded a slope indicative of the degree of DNA SSB. This method of data treatment is essentially that of Elia et al. (55). (3) In Vitro Experiments with P450-Transfected V79 Cells. Parental, P450 2E1- and P450 2B1-transfected hamster V79 cells were cultivated as described previously (52). For the

Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1559 incubations, 5 × 105 cells were transferred into a Petri dish forming a monolayer in 2 days. The nitrosamine dissolved in medium (10-40 mM) was added and the mixture incubated for 24 h at 37 °C. The medium was removed and the monolayer treated with trypsin, and the released cells were adjusted to 1.5 × 106/mL and submitted to alkaline filter elution. Determination of Kinetic Deuterium Isotope Effects for the ADH-Mediated Oxidation of NDELA. (1) Determination of Vmax and Km. The instrument that we used was a HP 8452 diode array spectrophotometer. The standard method for monitoring the kinetics of ADH-mediated transformations involves the spectrophotometric determination of NADH concentrations at 340 nm. The determination of the ADH-mediated oxidation of NDELA to NHMOR by this methodology is compromised by two factors. Both the substrate and the product have a significant extinction coefficient at 340 nm, and NDELA is a relatively poor substrate for ADH (at pH 7), requiring its use at higher concentrations. At the concentrations used here (20-55 mM), the extent of absorption at 340 nm exceeds the limit of the instrument. This necessitated the monitoring of NADH production at 370 nm. This reduced the extinction coefficient of NADH to about 50% compared to that at 340 nm. Even at 370 nm, the dynamic range is low and extreme care had to be taken to obtain reproducible measurements. A 2.7 mL solution containing 80 mM phosphate buffer (pH 7), 3.5 mM NAD, and 20-55 mM NDELA was preincubated in a cuvette in a heated cell holder at 37 °C. After a blank measurement was taken, 300 µL of ADH (3.3 units, filtered before use) was added and the increase in the level of absorption was recorded every second at 370 nm for 100 s. A reference wavelength of 550 nm was used. The initial slope of the reaction was determined. Commercially available ADH contains small amounts of ethanol, resulting in a slow reaction without substrate. Therefore, a blank of ADH was run without nitrosamine, and the reaction rate of 0.39 nmol of NADH (mg of protein)-1 min-1 was subtracted. Measurements for each concentration and each substrate were performed in at least triplicate, and the average initial slope at each concentration was determined. Vmax and Km were determined by linear regression of Lineweaver-Burk plots for NDELA, R-D4NDELA, and β-D4NDELA. (2) Analysis of NADH and NHMOR by HPLC. Because of the different method used for the determination of the kinetic parameters, the formation of NADH and NHMOR was established. A solution of NAD+ (3.5 mM), NDELA (50 mM), and 1.1 units of ADH/mL in 60 mM phosphate buffer (pH 7) was incubated for 4 min at 37 °C. The sample was divided. For the NADH analysis, the sample was directly injected onto a Waters HPLC system: Column Merck Purospher C18 column, 250 mm × 4 mm; eluent A was 50 mM phosphate buffer (pH 7) with 2 mM tetrabutylammonium dihydrogen phosphate and eluent B acetonitrile for generating a gradient of 5 to 20% B over the course of 25 min at a flow rate of 1 mL/min, utilizing a UV detector at 340 nm. For the NHMOR analysis, the sample was extracted with an equal volume of ethyl acetate and the organic extract was separated by HPLC on a Lichrosorb Si-60 column (100 mm × 3 mm); eluent A was methylene chloride and eluent B 2-propanol for generating a gradient of 0 to 3% B from 0 to 3 min, 3 to 20% B from 3 to 13 min, and 20% B from 13 to 25 min, with a flow rate of 0.7 mL/min utilizing a UV detector at 240 nm. The recovery of NHMOR, added to a control sample without ADH, was 45%. A nitrosamine blank without ADH was subtracted. Cytotoxicity Assay Utilizing Hamster V79MZh2EI Cells. P450 2E1 activity of the cells was confirmed by 6-hydroxylation of chlorzoxazone and by using dimethylnitrosamine as a positive control. Cytotoxicity testing was performed as described by Schmalix et al. (52) with the following modifications. Cells, 500 per well, were seeded in 96-well microtiter plates and cultivated in DMEM medium for 24 h. Then medium was exchanged against the medium containing the test compound. After incubation for 24 h, medium with test compound was replaced

1560 Chem. Res. Toxicol., Vol. 11, No. 12, 1998

Loeppky et al.

Figure 1. Alkaline elution profiles for rat liver DNA following administration of the indicated nitrosamine at a dose of 0.37 mM/kg (bw). The values of the slopes of the regression lines shown in panel B were multiplied by -1 to obtain the values listed in Table 1. by DMEM and incubation was continued until controls approached confluency (5 days). Then relative viability of cells was determined photometrically via the extent of neutral red uptake. Cells were incubated with diluted neutral red solution for 90 min at 37 °C in the incubator. Excess dye was removed from the cells with NaCl (0.9%). After extraction of the dye with ethanol/citrate buffer/HCI, the extent of neutral red absorption in the wells was measured at 546 nm in a plate reader (BioRad, Hercules, CA). Twelve replicate wells were monitored for each concentration, and outliers were identified and removed. Relative viability was calculated using an untreated set of cells (12 replicate wells) as a control (relative viability of 100%).

Results Deuterium Isotope Effects on Nitrosamine-Induced DNA SSB in Vivo in Rat Liver. With the goal of determining which C-H bonds in NDELA are involved in its genotoxic action, we initiated a DNA SSB assay with NDELA, its deuterated isotopomers, and NHMOR. In the first set of experiments, care was taken to isolate rat hepatocytes after high-dose nitrosamine treatment and to determine their viability as measured by the extent of trypan blue exclusion. An NDELA dose of 0.775 mmol/kg (bw) and an NHMOR dose of 0.37 mmol/kg (bw) were found not to induce significant hepatotoxicity (>70% viability). Each of the compounds was administered by gavage to rats. Two complete runs were performed. In a single run, one animal was used for each concentration of an individual compound. After 4 h, the animals were sacrificed, the livers were homogenized, and the nuclei were analyzed for SSB by alkaline filter elution using four filters per animal. With the apparatus running at a constant elution rate of volume per minute, the fraction of DNA on the filter at 10 time points (fractions) was determined. Typical data are presented in Figure 1A. The percentage of DNA remaining on the filter is plotted versus the fraction number. It is readily apparent that both NDELA and its two isotopomers induce significant DNA SSB in comparison to controls. The extent of strand breakage for NDELA and β-D4NDELA is significantly greater than that observed for R-D4NDELA. This phenomenon was observed at every dose level (see Table 1). To present the data in a more quantitative format, we have followed the practice of Elia et al. (55) and plotted

Table 1. Rat Liver DNA SSB Induced by NDELA, Its Deuterated Isotopomers, and NHMOR dosea

compound

slopeb

SD

kH/kDc

SDd

0.75

NDELA R-D4NDELA β-D4NDELA NDELA R-D4NDELA β-D4NDELA NHMOR NDELA R-D4NDELA β-D4NDELA NHMOR NDELA R-D4NDELA β-D4NDELA NHMOR control

0.081 0.059 0.102 0.096 0.051 0.088 0.077 0.040 0.016 0.031 0.034 0.029 0.009 0.021 0.0120 0.008

0.007 0.006 0.009 0.01 0.004 0.013 0.003 0.004 0.0008 0.004 0.001 0.002 0.0002 0.001 0.0001

1.37 0.79

.02 0.001

1.88 1.09

0.05 0.01

2.5 1.29

0.1 0.04

3.22 1.38

0.2 0.03

0.37

0.15

0.08

a In millimoles per kilogram of bw. b Data are reported as the slope of the regression line times -1. c kH/kD values are determined by division of the slopes for NDELA by those of R-D4NDELA or β-D4NDELA. d Errors are determined by propagation of the error (standard deviation) in the slopes.

the log of the fraction of DNA remaining on the filter for fractions 3-9 versus the fraction number. Linear regression of these data then gives a slope which is used as a marker of the compound’s ability to induce DNA SSB. A typical plot of our data in this format is presented in Figure 1B. It is easily observed that the slope of the curve for R-D4NDELA is less than that for either NDELA or β-D4NDELA. The slopes of all plots and their standard deviations are presented in Table 1. Only the data for one complete run are presented, as it was observed that the results were reproduced in the second run. Deuterium isotope effects for DNA SSB were calculated by dividing the slopes of the NDELA or NHMOR control by the slope of the appropriate deuterated NDELA or NHMOR plot at each dose and are presented in Table 1. Both the slopes of the strand break plots and the resulting isotope effects are a function of dose level as is shown in Figure 2A,B. The data suggest that a saturation phenomenon has been achieved at the highest dose levels. The isotope effects are most pronounced at the lowest dose levels. If the isotopic substitution at a given

Mechanism of N-Nitrosodiethanolamine Bioactivation

Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1561

Figure 2. (A) Dose response of NDELA, its deuterated isotopomers, and NHMOR on the slopes of the DNA SSB regression lines times (-1). (B) Dose response of deuterium isotope effects (kH/kD) on DNA SSB induced by R-D4NDELA and β-D4NDELA. Table 2. DNA SSB Induced by NHMOR and Its Deuterated Isotopomers dosea

compound

slopeb

error

kH/kDc

0.37

NHMOR 5,5-D2-NHMOR 2-D-NHMOR NHMORd 5,5-D2-NHMOR NHMORe 2-D-NHMOR NHMOR 5,5-D2-NHMOR 2-D-NHMOR control

0.157 0.096 0.138 0.141 0.111 0.152 0.146 0.073 0.026 0.107 0.008

0.004 0.007 0.001 0.002 0.003 0.002 0.002 0.003 0.001 0.002

1.6 1.1

0.25

0.15

1.27 1.04 2.8 0.7

a In millimoles per kilogram of bw. b Data are reported as the slope of the regression line times -1. c kH/kD values are determined by division of the slopes for NHMOR by those of its isotopomers. d Control for 5,5-D -NHMOR. e Control for 2-D-NHMOR. 2

position has no effect, then the isotope effect should be 1. It can be seen that the isotope effects for R-D4NDELA at all doses are significantly above 1. The isotope effects observed for β-D4NDELA increase with decreasing dose levels nearly linearly from a value of 0.79 to 1.38 at the lowest dose. While these values are close to the no effect level, we believe they are significant and point to a distinct isotope effect at the lowest dose. NHMOR, the oxidation product of NDELA, also causes DNA SSB in rat liver after in vivo application. To better understand the underlying activation mechanism, DNA SSB induced by 2-DNHMOR and 5,5-D2NHMOR was compared with that generated from the application of NHMOR. The induction of DNA SSB from all three compounds is observed (Table 2). At the lowest dose (0.15 mmol/kg), significant isotope effects are seen for both isotopomers, although that observed for 2-D-NHMOR (0.7) is an inverse isotope effect (1 are observed at low doses, but a kH/kD of