Extensive Metabolic Activation of the Tobacco-Specific Carcinogen 4-(Methylnitrosamino)-1-(3-Pyridyl)-1-Butanone in Smokers

Numerous investigations indicate that tobacco-specific nitrosamines play an important role in cancer induction by tobacco products (reviewed in refs. 1, 2). 4-(Methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) is one of the most prevalent of these compounds present in both unburned tobacco and cigarette smoke, and is a remarkably effective lung carcinogen in laboratory animals, inducing lung tumors in rodents independent of the route of administration (1). NNK and polycyclic aromatic hydrocarbons are believed to be major causative agents for lung cancer in smokers (2-4). 4-(Methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL), a metabolite of NNK, is also a pulmonary carcinogen. Both compounds require metabolic activation to exert their carcinogenic effects.

NNK and NNAL metabolism, adduct formation, and detoxification have been investigated in several studies involving laboratory animals and humans and are now quite well understood (1). Figure 1 presents an overview of NNK and NNAL metabolism and DNA adduct formation. Although there are quantitative differences, many of the same reactions are observed in both laboratory animals and humans. Studies in rodents and primates show that NNK is rapidly metabolized and distributed to all tissues (5-9). Overall, three major routes of NNK metabolism are consistently observed in vivo: carbonyl reduction, pyridine oxidation, and α-hydroxylation (1). Reduction of the NNK carbonyl group produces NNAL, which can be partially converted back to NNK, the NNK-NNAL equilibrium favoring NNAL in rodents, primates, and humans (8-13). Major metabolic pathways and carcinogenicity of NNAL are similar to those of NNK, and therefore, it is not a detoxified metabolite of NNK (1, 14-16). A unique feature of NNAL metabolism is the formation of NNAL-glucuronides (NNAL-Glucs), which is an important detoxification pathway for NNK and NNAL (1, 9, 11). Glucuronidation of NNAL at the pyridine nitrogen forms NNAL-N-Gluc, whereas that at the carbinol oxygen gives NNAL-O-Gluc (17). Levels of NNAL-Glucs are generally higher than those of NNAL in human urine. Pyridine-N-oxidation results in formation of NNK-N-oxide and NNAL-N-oxide in rodents and primates (9, 11).

In humans, urinary levels of NNAL-N-oxide were <20% of those of NNAL-Glucs, whereas NNK-N-oxide and unchanged NNK have not been detected (13, 18). Metabolic activation of NNK and NNAL to DNA adducts proceeds via α-hydroxylation pathways (Fig. 1). The DNA adducts, if unrepaired, cause miscoding leading to gene mutations and cancer. 4-Oxo-4-(3-pyridyl)butanoic acid (keto acid) and 4-hydroxy-4-(3-pyridyl)butanoic acid (hydroxy acid) are the principal end products of the NNK α-hydroxylation pathways in rodents and primates (1, 9, 11, 19-21). Both of these metabolites can be measured in human urine; however, they cannot be used to evaluate the extent of NNK α-hydroxylation in smokers because they are also formed from nicotine, and levels of nicotine in cigarettes are much higher than those of NNK (22, 23).

In this study, we address the important question of the extent of NNK metabolic activation in humans, which, for the reasons discussed above, has never been previously reported. Our hypothesis is that individuals who activate NNK more extensively will be at higher risk for lung cancer. The key factor in the evaluation of NNK metabolic activation would be distinguishing urinary keto acid and hydroxy acid as formed from NNK versus nicotine. Because hydroxy acid is chiral, it was suggested that one enantiomer would be formed preferentially from NNK, whereas the formation of the other would result from nicotine, and studies in rats showed that this was plausible (22). However, in contrast to the studies in rats, hydroxy acid was found to be a more abundant nicotine metabolite in humans, and even the minor enantiomer, as formed from nicotine, was much higher in concentration than that which would be produced from NNK (23). Therefore, we took advantage of the commercial availability of a low-nicotine cigarette called Quest, which also has relatively low NNK levels (24). We added [pyridine-D₄]NNK (Fig. 2) to the tobacco of Quest cigarettes such that the total NNK concentration (unlabeled plus deuterated) in the tobacco of these cigarettes was comparable with that of a standard cigarette. We recruited smokers of standard brands to smoke the cigarettes containing [pyridine-D₄]NNK and then quantified [pyridine-D₄]NNAL, [pyridine-D₄]NNAL-Glucs, and the sum of [pyridine-D₄]hydroxy acid and [pyridine-D₄]keto acid in the urine of these individuals. Data from this study thus provide the first direct information on the extent of NNK metabolic activation in smokers.

NNK, [pyridine-D₄]NNK, NNAL, and [pyridine-D₄]NNAL are carcinogenic and mutagenic and should be handled with extreme care using appropriate protective clothing and ventilation at all times.

NNAL and 5-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (C₅-NNAL) were purchased from Toronto Research Chemicals, Inc. NNK, 5-(methylnitrosamino)-1-(3-pyridyl)-1-pentanone (C₅-NNK), [¹³C₆]NNAL, [pyridine-D₄]NNAL, and hydroxy acid were synthesized as previously described (19, 22, 25-29). β-Glucuronidase (type IX-A from Escherichia coli) was purchased from Sigma-Aldrich Chemical Co. [Pyridine-D₄]ethyl nicotinate was purchased from CDN Isotopes. Methyl-5-methyl nicotinate was obtained from Lancaster (Alfa Aesar). NaBH₄ and all other chemicals were purchased from Sigma-Aldrich Chemical.

[Pyridine-D₄]NNK. The synthesis of [pyridine-D₄]NNK was carried out essentially as previously described (28) but starting with [pyridine-D₄]ethyl nicotinate. Properties of [pyridine-D₄]NNK: mp 66-68°C, literature (ref. 30), 63-65°C; NMR δ (CDCl₃) 4.3 (t, 2H, CH₂-N-N = O, E-isomer), 3.8 (s, CH₃-N-N = O, Z isomer), 3.7 (t, 2H, CH₂-N-N = O, Z-isomer), 3.07 (s, CH₃-N-N = O, E-isomer), 3.05 (t, O = C-CH₂, E-isomer), 2.9 (t, O = C-CH₂, Z-isomer), 2.2 (m, O = C-CH₂-CH₂, E-isomer), 1.9 (m, O = C-CH₂-CH₂, Z-isomer). Positive ion electrospray-mass spectrometry (ESI-MS) m/z 212 [M + H]⁺, MS/MS of m/z 212, m/z 182 [M − N = O]⁺, 126 [M − C₃H₅N]⁺.

A concentrated stock of 1.0 μg/μL [pyridine-D₄]NNK in ethanol was prepared. Spiking solution (15.0 ng/μL [pyridine-D₄]NNK in H₂O) was prepared by adding 15 μL of stock to 985 μL high-performance liquid chromatography (HPLC)-grade H₂O.

4-Hydroxy-4-(5-Methyl-3-Pyridyl)Butanoic Acid (5-Methylhydroxy Acid). This was prepared from 5-methyl methyl nicotinate essentially as described (29). Spectral properties: NMR δ[(CD₃)₂SO]8.27 (d, 2H, Pyr-2H, Pyr-6H), 7.5 (s, 1H, Pyr-4H), 4.5 (m, 1H, CH-OH), 2.8 (s, 3H, Pyr-CH₃), 2.2 (t, 2H, CH₂-COOH), 1.8 (t, 2H, CH₂- CH₂-COOH). Positive ion ESI-MS m/z (relative intensity) 196 [M + H]⁺ (63), 178 [M + H − H₂O]⁺ (100), 132 [M − CH₃O₃]⁺ (30).

[Pyridine-D₄]Hydroxy Acid. [Pyridine-D₄]hydroxy acid was synthesized as described for 5-methylhydroxy acid, except that [pyridine-D₄]ethyl nicotinate was used as the starting material. HPLC purification was carried out on a 250 mm × 10 mm, C-18 Vydac 201TP 10 μm column (Separations Group) eluted at 3 mL/min using isocratic elution with 0.15% trifluoroacetic acid in H₂O for 20 min, giving [pyridine-D₄]hydroxy acid, retention time of 7 min. NMR δ[(CD₃)₂SO]4.8 (m, 1H, CH-OH), 2.3 (t, 2H, CH₂-COOH), 1.9 (m, 2H, CH₂-CH₂-COOH). Positive ion ESI-MS m/z (relative intensity) 186 [M + H]⁺ (40), 168 [M + H − H₂O]⁺ (100), 122 [M − CH₃O₃]⁺ (63).

Deuterium is nonradioactive and nonhazardous, and its substitution in the pyridine ring of NNK is not expected to have an effect on NNK metabolic activation or carcinogenicity (15, 28). The use of these cigarettes was approved by the U.S. Food and Drug Administration.

Quest 1 cigarettes were ordered in four batches through the SmokePlaza³

The study was approved by the University of Minnesota Research Subjects' Protection Programs Institutional Review Board: Human Subjects Committee. Subjects were daily “light,” nonmenthol cigarette smokers recruited through newspaper advertisements. They were initially screened over the telephone, and potential subjects were scheduled for a screening visit at the University of Minnesota Tobacco Use Research Center. At the screening visit, they signed a consent form and completed several questionnaires on demographics and smoking and health history. Subjects were excluded if they had poor or unstable physical or mental health that would affect study participation. Eligible subjects were asked to collect a 24-h urine sample the day before their next scheduled visit. At that visit, they were given the study cigarettes and instructed to smoke only those cigarettes over the following 7-d period. Beginning on the 5th day of smoking the study cigarettes, subjects were asked to collect their 24-h urine on each of the next 3 d and to keep each 24-h urine collection in a separate container. After completing the collection of each day, subjects were required to bring the sample to the Tobacco Use Research Center.

To prepare tobacco samples, three cigarettes were cut open, the tobacco was mixed, and two 500-mg samples were analyzed.

Total NNK. Before adding [pyridine-D₄]NNK, the tobacco of Quest 1 cigarettes was analyzed for NNK using a model 5890 gas chromatograph (Hewlett-Packard) interfaced with a model 610 Thermal Energy Analyzer (Orion Research) as described elsewhere (24). The same method was used to analyze total NNK (NNK plus [pyridine-D₄]NNK) content in the study cigarettes after their addition of [pyridine-D₄]NNK.

[Pyridine-D₄]NNK. To distinguish between nondeuterated NNK and [pyridine-D₄]NNK, we used gas chromatography-MS/MS. Tobacco samples were prepared in the same manner as for analysis of nondeuterated NNK (24), and 2 μL of the prepared sample were injected into the gas chromatography-MS/MS system. The analysis was carried out with a Finnigan TSQ-7000 instrument operated in the positive ion electron ionization mode. Daughter ion scans were done to monitor the transitions m/z 177 → m/z 146 and m/z 177 → m/z 118 for NNK, m/z 181 → m/z 150 and m/z 181 → m/z 122 for [pyridine-D₄]NNK, and m/z 191 → m/z 134 and m/z 191 → m/z 162 for C₅-NNK, operating Q3 in the selected ion monitoring mode at a scan rate of 0.5 scan per second. The gas chromatograph (HP Model 5890) was equipped with a 15 m × 0.25 mm inner diameter DB 1301 column (0.25 μm film thickness) from J & W Scientific connected to a 2 m × 0.32 mm inner diameter deactivated precolumn. The injection mode was splitless, the constant flow rate was 2.0 mL/min He, and the injection port temperature was 225°C. The temperature program was as follows: 80°C for 2 min, then 20°C per minute to 155°C, then 2°C per minute to 190°C, and then 20°C per minute to 250°C. The final temperature was held for 5 min.

Total NNAL and Total [Pyridine-D₄]NNAL. Urine (3 mL) was placed in a 10 mL glass centrifuge tube, the pH was adjusted to 6 to 7 if necessary, and 20 pg [¹³C₆]NNAL was added as internal standard. β-Glucuronidase (10,000 units in 500 μL HPLC-grade H₂O) was added to each sample, and the mixture was incubated at 37°C overnight. The next day, the mixture was applied to a 5-mL ChemElut cartridge (Varian) and eluted with 2 × 8 mL CH₂Cl₂ into a 15-mL glass centrifuge tube, and the combined eluants were concentrated to dryness (SpeedVac concentrator). The dry residue was redissolved in 1 mL H₂O, adjusted to pH 2 to 3 by adding 100 μL 1 N HCl, and the mixture was applied to a 60-mg Oasis MCX cartridge (Waters Corp.) activated with 5 mL CH₃OH and equilibrated with 10 mL H₂O. The cartridges were washed with 5 mL 1 N HCl, 5 mL CH₃OH, and 5 mL H₂O:CH₃OH:NH₄OH (90:5:5) and these washings were discarded. Then, the analytes were eluted from the Oasis MCX cartridges with 5 mL H₂O:CH₃OH:NH₄OH (30:65:5) and the eluant was concentrated to dryness. The dry residues were transferred to autosampler vials with CH₃OH, dried, and stored at −20°C until analysis by LC-ESI-MS/MS. Before analysis, samples were redissolved in 10 μL of 2% CH₃OH in H₂O, and 5 μL were injected.

LC-ESI-MS/MS was carried out with a Finnigan TSQ Quantum Discovery Max instrument (Thermo Electron Corp.) interfaced with an Agilent Model 1100 capillary HPLC system and a Model 1100 micro autosampler (Agilent). The HPLC was fitted with a 150 × 0.5 mm ZORBAX SB C18 RR 3.5 μm column (Agilent) eluted isocratically with 35% methanol in H₂O for 20 min at a flow rate of 10 μL/min. The column was maintained at 25°C. MS/MS variables were as follows: positive ion electrospray mode with selected reaction monitoring for m/z 210 → m/z 180 ([M+H]⁺ → [M+H - NO]⁺) for NNAL, m/z 214 → m/z 184 for [pyridine-D₄]NNAL, and m/z 216 → m/z 186 for [¹³C₆]NNAL at 0.5 a.m.u. scan width. The collision gas was Ar at a pressure of 1 mTorr, with a collision energy of 10 eV. The quadrupoles were operated at a resolution of 0.7 a.m.u.

Free NNAL and Free [Pyridine-D₄]NNAL. The method was practically identical to that described for total NNAL and total [pyridine-D₄]NNAL, except that the samples were not treated with β-glucuronidase before their extraction on ChemElut cartridges.

NNAL-Glucs and [Pyridine-D₄]NNAL-Glucs. These were determined by subtracting amounts of free NNAL and free [pyridine-D₄]NNAL from total NNAL and total [pyridine-D₄]NNAL, respectively.

This was done by a modification of the previously described method (23), as follows.

Reduction of [Pyridine-D₄]Keto Acid to [Pyridine-D₄]Hydroxy Acid. Urine (3 mL) was treated with NaBH₄ overnight at room temperature, as previously described (23). The next day, the excess NaBH₄ was destroyed by dropwise addition of 1 N HCl, and the samples were then analyzed as described for [pyridine-D₄]hydroxy acid.

[Pyridine-D₄]Hydroxy Acid. 5-Methylhydroxy acid internal standard (50 ng) was added to the neutralized urine that had been treated with NaBH₄. Each sample was applied to an activated 200 mg Strata X cartridge (Phenomenex, Inc.), the [pyridine-D₄]hydroxy acid was eluted with 6 mL of 20% CH₃OH, and the eluant was concentrated to dryness. CH₃OH (1 mL) was added to the dry residue, and the sample was concentrated to dryness. This was repeated twice to ensure complete removal of H₂O. The dry residue was dissolved in 1 mL of freshly prepared 3% concentrated H₂SO₄ in CH₃OH, and the mixture was allowed to stand overnight at room temperature. This step converts very polar [pyridine-D₄]hydroxy acid to its methyl ester, which further can be extracted into a nonpolar organic solvent and subjected to normal phase extraction, thus leading to less chemical noise and lower ion suppression on sample analysis by LC-ESI-MS/MS. The conversion is nearly quantitative (22). The next day, the mixture was neutralized by addition of 2.0 mL 5% (w/v) NaHCO₃ and loaded on a 5 mL ChemElut cartridge. The methyl ester was eluted with 2 × 8 mL CH₂Cl₂, and the eluant was concentrated to dryness. The dry residue was immediately redissolved in 500 μL CH₂Cl₂ and loaded on a 500 mg Sep Pak Silica cartridge (Waters) preequilibrated with 4 mL CH₂Cl₂. The cartridge was washed with 4 mL CH₂Cl₂ and 4 mL CH₂Cl₂/EtOAc (50:50); these fractions were discarded. The methyl ester of [pyridine-D₄]hydroxy acid was eluted with 8 mL of EtOAc:CH₃OH (90:10), and the eluant was concentrated to dryness. The dry residue was immediately redissolved in 1 mL H₂O and 100 μL 1 N HCl, and the mixture was incubated at 40°C for 2 h. This step hydrolyzed the methyl ester back to [pyridine-D₄]hydroxy acid, which was further purified on a 60-mg Oasis MCX cartridge activated with 5 mL CH₃OH and preequilibrated with 10 mL H₂O. The cartridge was washed with 3 mL 1 N HCl and 3 mL of 80:20 CH₃OH/1 N HCl, which were discarded, and [pyridine-D₄]hydroxy acid was eluted with 6 mL of CH₃OH:H₂O:NH₄OH (20:75:5). The eluant was concentrated to dryness, and the residue was transferred in two 100-μL aliquots of acetonitrile to a 45-μm nylon Spin-X LC filtration vial. The filtrate was transferred to a glass microinsert autosampler vial, concentrated to dryness, and stored at −20°C until analysis by LC-ESI-MS/MS.

LC-ESI-MS/MS was carried out with a Finnigan TSQ Quantum Discovery Max instrument (Thermo Electron) interfaced with an Agilent Model 1100 capillary HPLC system and a Model 1100 micro autosampler. The HPLC was fitted with a 100 × 0.5 mm ZORBAX SB C18 1.8 μm column (Agilent) maintained at 35°C and eluted over the course of 35 min at a flow rate of 9 μL/min. The following gradient program was used: 10 mmol/L ammonium acetate for 2 min, then a linear gradient over the course of 5 min to 95% 10 mmol/L ammonium acetate/5% acetonitrile, which was held for 5 min; then a linear gradient over the course of 10 min to 65% 10 mmol/L ammonium acetate/35% acetonitrile; and then return to 100% 10 mmol/L ammonium acetate in 5 min and hold for 8 min to equilibrate the column for the next injection. Based on the MS of the analytes, the following MS/MS variables were used: positive ion electrospray mode with selected reaction monitoring for m/z 182 → m/z 164 plus m/z 182 → m/z 118 for hydroxy acid, m/z 186 → m/z 168 plus m/z 186 → m/z 122 for [pyridine-D₄]hydroxy acid, and m/z 196 → m/z 178 plus m/z 196 → m/z 132 for 5-methylhydroxy acid at 0.2 a.m.u. scan width. The collision gas was Ar at a pressure of 1 mTorr, with a collision energy of 10 eV. The quadrupoles were operated at a resolution of 0.7 a.m.u.

The Pearson correlation was used to measure the association between the ratios of [pyridine-D₄]NNAL-Glucs/free [pyridine-D₄]NNAL and NNAL-Glucs/free NNAL, both on the log scale.

Quest cigarettes are available in three varieties: Quest 1, low nicotine; Quest 2, extra low nicotine; and Quest 3, nicotine-free. Our previous studies showed that levels of NNK are similar in the tobacco of Quest 1 and 2 and lower in the nicotine-free Quest 3 (24). Because the nicotine content in Quest 1 is the highest of the three Quest varieties, we chose it for the trial because we expected that our smokers would not try to compensate for the lack of nicotine by smoking their usual cigarettes or by increasing the number of study cigarettes smoked per day (thus increasing their uptake of NNK).

The mean (±SD) NNK content in four batches of Quest 1 cigarettes was 0.225 (±0.07) μg/g tobacco, and the average tobacco weight per cigarette was 0.623 g (Table 1). Based on the reported average NNK levels in popular commercial cigarette brands (0.750 μg/g tobacco; ref. 24), we added 0.315 μg [pyridine-D₄]NNK to each study cigarette so that the amount of total (deuterated plus unlabeled) NNK in these cigarettes would be ∼0.750 μg/g tobacco. The results of the analyses of the study cigarettes before and after adding are summarized in Table 1. The spiked cigarettes from batch 2 were also analyzed for [pyridine-D₄]NNK. They contained 0.527 μg/g tobacco or 0.323 μg/cigarette.

Twenty subjects were recruited for the study. Eleven (55%) were male and 15 (75%) were white. Their average age was 34 years (range, 20-56). The average (±SD) number of study cigarettes smoked per day during the trial, as reported by the participants, was 28 (±9). Six subjects reported smoking a total of 1 to 19 cigarettes of their own brand during the trial period, in addition to the reported number of study cigarettes.

Deuterated biomarkers were not detected in the baseline urine samples. The mean level of unlabeled total NNAL in the baseline urine was 2.99 ± 1.75 nmol/24 h. In the urine collected during days 5 to 7 of smoking the spiked Quest 1 cigarettes, these levels slightly decreased, averaging 2.23 ± 1.23 nmol/24 h. Typical LC-MS/MS traces from the analysis of NNAL, [pyridine-D₄]NNAL, hydroxy acid, and [pyridine-D₄]hydroxy acid in the urine of a study participant after 5 days of smoking the spiked Quest 1 cigarettes are illustrated in Figs. 3 and 4. The results of the analyses of the deuterated biomarkers are summarized in Table 2. Despite the quite variable amounts of deuterated metabolites among the trial participants, the sum of the products of [pyridine-D₄]NNK α-hydroxylation pathways was the prevalent biomarker in all subjects and all urine samples analyzed. Thus, the total amount of [pyridine-D₄]keto acid plus [pyridine-D₄]hydroxy acid averaged 4.00 ± 2.49 nmol/24 h (geometric mean, 2.87 nmol/24 h), whereas the average amount of total [pyridine-D₄]NNAL was 0.511 ± 0.368 nmol/24 h (geometric mean, 0.407 nmol/24 h). There was also a significant day-to-day variability of the deuterated biomarkers of α-hydroxylation pathways within subjects (average coefficient of variation, 44.5%), whereas the levels of total [pyridine-D₄]NNAL showed lower intraindividual variation (average coefficient of variation, 23.1%).

To further investigate the α-hydroxylation pathways of [pyridine-D₄]NNK, we did separate analysis of [pyridine-D₄]keto acid and [pyridine-D₄]hydroxy acid in three randomly selected subjects. The analysis for [pyridine-D₄]hydroxy acid was accomplished by excluding the NaBH₄ reduction step from the sample preparation procedure. The amount of [pyridine-D₄]keto acid was determined by subtracting [pyridine-D₄]hydroxy acid from the total amount of [pyridine-D₄]keto acid plus [pyridine-D₄]hydroxy acid detected with NaBH₄ reduction. Nondeuterated keto acid and hydroxy acid were also analyzed in the same subjects. The results are summarized in Table 3. [Pyridine-D₄]keto acid accounted for an average of 28.2% of the total amount of the deuterated products of α-hydroxylation pathways, which was quite similar to the 22.5% average contribution of nondeuterated keto acid to the total amount of unlabeled keto acid plus hydroxy acid found in the same urine samples.

Important ratios between the measured biomarkers are summarized in Table 4. The average ratios of [pyridine-D₄]NNAL-Glucs to free [pyridine-D₄]NNAL (1.90 ± 0.95) and nondeuterated NNAL-Glucs to free NNAL (2.01 ± 0.91) were similar and were highly correlated in the trial participants (r = 0.72; P < 0.001). The ratio of the sum of [pyridine-D₄]keto acid plus [pyridine-D₄]hydroxy acid to total [pyridine-D₄]NNAL ranged from 0.99 to 33.4, averaging 10.8 ± 9.16 (geometric mean, 7.04). The ratio of the sum of deuterated α-hydroxylation biomarkers to the product of [pyridine-D₄]NNK detoxification, [pyridine-D₄]NNAL-Glucs, varied from 1.35 to 62.8, averaging 18.8 ± 16.8 (geometric mean, 11.7). Frequency plots of these ratios are illustrated in Fig. 5.

The purpose of this study was to determine the extent of NNK metabolic activation in smokers, as measured by the sum of the major end products of this pathway, keto acid and hydroxy acid, relative to total NNAL. To specifically identify NNK-derived keto acid and hydroxy acid, which are also metabolites of nicotine, we added [pyridine-D₄]NNK to cigarettes that were originally low in NNK, and measured the deuterium-labeled metabolites in the urine of people who smoked these cigarettes. The results show for the first time that the tobacco-specific lung carcinogen NNK is extensively metabolically activated in smokers.

Metabolic activation of NNK occurs largely via cytochrome P450–catalyzed α-hydroxylation (i.e., hydroxylation of the carbons adjacent to the N-nitroso group), which does not involve modifications in the pyridine ring (1). Therefore, deuterium substitution in the pyridine ring of NNK is not expected to have an effect on NNK metabolic activation or carcinogenicity. Similarly, this substitution is not expected to affect the formation of [pyridine-D₄]NNAL via carbonyl reduction of [pyridine-D₄]NNK.

The levels of unlabeled total NNAL in the baseline urine were similar to the earlier reported mean total NNAL in 27 smokers (3.14 nmol/24 h; ref. 31). After switching to the spiked Quest 1, which is relatively low in unlabeled NNK, urinary excretion of unlabeled total NNAL slightly decreased; however, it remained higher than that of total [pyridine-D₄]NNAL in the same urine samples. Thus, the average (±SD) ratio of unlabeled to labeled total NNAL in the urine collected during days 5 to 7 of smoking [pyridine-D₄]NNK–added cigarettes was 5.92 (±3.88), whereas it should have been ∼0.4 based on the amounts of NNK and [pyridine-D₄]NNK in the spiked cigarettes (Table 1). There could be several factors that contributed to the relatively high levels of unlabeled NNAL. First, it is possible that during the trial period, our smokers were compensating for the lack of nicotine in Quest cigarettes by smoking their regular brands in addition to the study cigarettes. Another possibility is that NNAL formed before the beginning of the study was still being excreted in the urine of our smokers and contributed to the total unlabeled NNAL measured during the trial period (31). Exposure to secondhand smoke could also contribute to the higher levels of unlabeled NNAL compared with deuterated NNAL. All studies reported to date show significantly higher amounts of total NNAL in the urine of secondhand smoke–exposed humans than unexposed controls (reviewed in ref. 32), and one recent publication has presented results of a study that showed that additional amounts of NNK can be formed in the secondhand smoke (33). Despite the unexpected relative amounts of deuterated and unlabeled NNAL, the ratio of [pyridine-D₄]NNAL-Glucs to free [pyridine-D₄]NNAL was very close to that of NNAL-Glucs to free NNAL in each urine sample (Table 3), supporting the concept that deuterated and nondeuterated NNK are metabolized in a similar way in humans.

The most important result of this study is the quantitation of the sum of urinary metabolites that resulted from [pyridine-D₄]NNK α-hydroxylation pathways—[pyridine-D₄]hydroxy acid and [pyridine-D₄]keto acid. The average contribution of [pyridine-D₄]keto acid to the sum of these metabolites was 28.2% (Table 3), which is consistent with the previously reported extensive conversion of keto acid to hydroxy acid in humans, and 15% keto acid contribution to the total keto acid and hydroxy acid in the urine of smokers (23). The relative amounts of the sum of α-hydroxylation biomarkers and total [pyridine-D₄]NNAL observed here (Table 2) indicate that α-hydroxylation is the major pathway of NNK metabolism in smokers.

This provides further insights into the overall NNK metabolic profile in human urine and its differences from those observed in studies with laboratory animals. Relative amounts of the urinary biomarkers of NNK major metabolic pathways in mice, rats, primates, and humans are compared in Table 5. Because of differences in routes of administration and doses applied in different studies, these data are not exactly comparable; however, they still provide some useful information. A notable similarity in the urinary NNK metabolic profiles in humans and animals is the prevalence of the products of α-hydroxylation pathways: the sum of keto acid and hydroxy acid accounts for 58%, 54%, 67.5%, and 86% of the urinary excretion of NNK metabolites in mice, rats, primates, and humans, respectively (Table 5). However, the relative contributions of keto acid and hydroxy acid to the total amount of these biomarkers are different. Thus, keto acid accounts for 42%, 74%, 37%, and 28% of total keto acid plus hydroxy acid in mice, rats, primates, and humans, respectively, suggesting that carbonyl reductase activity for keto acid is greater in humans than in the studied animals.

This idea is supported by a comparative study of several carbonyl-containing drugs, which showed more efficient ketone reduction in human than in mouse and rat liver (34). The extent of carbonyl reduction, glucuronidation, and pyridine-N-oxidation also varies in rodents, primates, and humans. For instance, neither NNAL nor NNAL-Glucs was detected in the urine of mice or rats treated with low NNK doses (11), indicating that pyridine-N-oxidation predominates over glucuronidation or urinary excretion of NNAL in rodents. Conversely, the results obtained for primates and humans suggest that carbonyl reduction and glucuronidation are more important metabolic pathways than pyridine-N-oxidation, glucuronidation being an important detoxification pathway for NNAL (Table 5).

Our finding that total NNAL is present in human urine in relatively low amounts compared with NNK-derived keto acid and hydroxy acid does not diminish the importance of total NNAL as a biomarker of human NNK uptake. Most investigations to date show a significant correlation between total NNAL and cotinine in human urine (1, 18, 35-37). Cotinine, a major nicotine metabolite in humans, is widely used as a biomarker of human nicotine uptake, although its urinary levels account for only 10% to 15% of the nicotine dose (38). We also have shown a significant positive correlation between cigarettes per day and cotinine, and between cigarettes per day and NNAL in the urine of smokers (39). Therefore, total NNAL can serve as an indicator of NNK uptake, just as cotinine serves as an indicator of nicotine uptake from tobacco exposure.

Considering the 23% transfer rate of NNK from tobacco to cigarette smoke reported elsewhere (40) and an average of 28 spiked cigarettes smoked per day, our smokers were exposed to ∼9.6 nmol [pyridine-D₄]NNK per day. Thus, the average amount of total [pyridine-D₄]NNAL excreted by the study participants over the 24-h period accounts for ∼5%, whereas the average amount of [pyridine-D₄]hydroxy acid accounts for ∼42%, of the estimated average [pyridine-D₄]NNK exposure. Calculations based on the geometric means for the corresponding biomarkers (Table 2) change these numbers to 4% and 30%, respectively. These calculations, although quite approximate and potentially greatly affected by the misreported number of spiked cigarettes smoked by the study participants, as well as by the individual patterns of smoking, show a reasonably consistent relationship between the estimated dose of [pyridine-D₄]NNK and the amounts of deuterated metabolites detected in the urine of our subjects.

The more than 10-fold variation in the levels of the sum of [pyridine-D₄]keto acid plus [pyridine-D₄]hydroxy acid, more than 30-fold variation in the [pyridine-D₄]keto acid plus [pyridine-D₄]hydroxy acid/total [pyridine-D₄]NNAL ratio, and more than 45-fold variation in the [pyridine-D₄]keto acid plus [pyridine-D₄]hydroxy acid/[pyridine-D₄]NNAL-Glucs ratio among the 20 smokers strongly support our hypothesis that some smokers activate NNK more extensively than others and that the ratio between biomarkers of metabolic activation and detoxification at a given dose of NNK could be a potential indicator of cancer risk. We are planning studies in which these smokers could be identified by smoking cigarettes to which [pyridine-D₄]NNK has been added.

In summary, the results of this study show for the first time that NNK metabolic activation is a quantitatively significant pathway in smokers, accounting for ∼86% of total urinary excretion of NNK metabolites. In the future, [pyridine-D₄]NNK could be used in a biomarker strategy to identify smokers who activate NNK more efficiently and, therefore, might be at an increased risk of developing lung cancer.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank David Wen and Bradley Lieberman for technical assistance, Dr. Peter W. Villalta for help with mass spectrometry techniques, Siyi Zhang for help with NMR analysis, Brad Hochalter for valuable advice in the method development, Bruce Lindgren for statistical analyses, and Bob Carlson for editorial assistance.