Abstract
Detoxication of the tobacco-specific carcinogen 4-methylnitrosamino-1-(3-pyridyl)-1-butanone (NNK) in humans is mainly due to carbonyl reduction to the chiral alcohol 4-methylnitrosamino-1-(3-pyridyl)-1-butanol (NNAL), which undergoes glucuronidation and excretion. NNAL has a carcinogenic potential with (S)-NNAL being more tumorigenic in the mouse. Therefore, the enantioselectivity of NNK reductases seems toxicologically relevant. NNAL enantiomers were measured by a novel high-performance liquid chromatography procedure. The aldo-keto reductases AKR1C1, 1C2, and 1C4 and carbonyl reductase purified from human liver cytosol produced NNAL with >90% (S)-enantiomer in accordance with the enantioselectivity of NNK reduction by cytosol from liver, placenta, and lung. In contrast, the (R)-NNAL content was 35% on NNK reduction with 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) purified from human liver microsomes, but around 70% with human microsomes. The selectivity for (R)-NNAL formation was still higher with microsomes from placenta (87%) and lung (89% in 10 of 11 surgical samples). Microsomes from lung of one patient reduced NNK at a much lower rate, with production of 14% (R)-NNAL. This points to predominant reduction in microsomes by an enzyme with selectivity for (R)-NNAL formation that was apparently absent from the lung of one patient. Experiments with 18β-glycyrrhetinic acid, a potent inhibitor of 11β-HSD1, also indicated a minor or no role for 11β-HSD1. Rat liver and lung microsomes produced NNAL with about 33% and 55% (R)-enantiomer and a mean contribution of 11β-HSD1 of 12% and 32%, respectively. Multiple enzymes seem to participate in NNK reduction in human and rat tissues.
The main biochemical pathways by which the tobacco-specific carcinogen 4-methylnitrosamino-1-(3-pyridyl)-1-butanone (NNK) is detoxicated are carbonyl reduction followed by conjugation with glucuronic acid and N-oxidation at the pyridine nitrogen. In humans, the major reaction is reduction, which results in the formation of the (R)- and (S)-enantiomers of 4-methylnitrosamino-1-(3-pyridyl)-1-butanol (NNAL) (Hecht, 1998). Due to isomerism at the N-nitroso group, NNK and the NNAL enantiomers exist in interconvertible geometric isomers with the (E)-isomer exceeding the (Z)-isomer in quantity (Fig. 1). Together with their glucuronides, the NNAL isomers can be detected in urine from present and past human tobacco users (Hecht, 1998; Hecht et al., 1999, 2002) and from individuals exposed to environmental tobacco smoke (Hecht et al., 2001). Interest in the enantiomeric composition of NNAL was stimulated by the finding of a higher tumorigenicity of (S)- compared with (R)-NNAL in the A/J mouse (Upadhyaya et al., 1999). Kinetic differences that may contribute to the phenomenon are a more extensive glucuronidation of (R)-NNAL in the mouse in vivo (Upadhyaya et al., 1999), in rat liver microsomes, by expressed rat UDP-glucuronosyltransferase 2B1 (Ren et al., 1999), and in the rat in vivo (Wu et al., 2002). (S)-NNAL is a better substrate for hydroxylation at the α-C position to the N-nitroso group (Upadhyaya et al., 1999) and for oxidation to NNK by cytochrome P450 species expressed in mouse, rat, and human lung (Jalas and Hecht, 2003; Jalas et al., 2003); in addition, it is stereoselectively retained in rat lung (Wu et al., 2002), and the same seems to apply in the human organism (Hecht et al., 2002).
The enantioselectivity of NNK reduction to NNAL has been investigated in various human and animal tissues using separation of the NNAL trimethylsilyl ethers on a chiral gas chromatography column (Carmella et al., 1999). The contribution of (S)-NNAL to total NNAL formation was reported to average 90% or more in cytosol and microsomes from mouse and rat lung and liver, and in human liver cytosol, and 64% in human liver microsomes (Upadhyaya et al., 2000). In bile and urine of NNK-treated rats, the glucuronide of (R)-NNAL by far exceeded that of the (S)-enantiomer in quantity (Wu et al., 2002), and this observation was ascribed to extensive reoxidation of (S)-NNAL to NNK and preferential glucuronidation of (R)-NNAL (Upadhyaya et al., 2000). Since the (R)/(S) ratio in NNAL glucuronide in bile and urine was nearly the same whether NNK or racemic NNAL was administered (Wu et al., 2002), carbonyl reduction must have led to comparable quantities of the two NNAL enantiomers; this finding is at variance with the findings in tissue fractions. In urine of humans who were current smokers or users of smokeless tobacco, the mean (R)/(S) ratios of NNAL and its glucuronide were between 0.34 and 0.85 (Carmella et al., 1999; Hecht et al., 2002) and, therefore, also did not reflect the ratio measured on incubation of human liver fractions (Upadhyaya et al., 2000).
No data are available on the enantioselectivity of NNK carbonyl reduction by individual enzymes. 11β-HSD1 isolated from mouse liver microsomes (Maser et al., 1996) and several oxidoreductases isolated from human liver cytosol were demonstrated to catalyze the reaction (Atalla et al., 2000). NNK reduction also takes place in human placenta microsomes (Collazo and Sultatos, 1995) and cytosol, with carbonyl reductase (CR) (EC 1.1.1.184) playing an important role in placenta cytosol (Atalla and Maser, 2001). NNK is further reduced in human lung cytosol (Maser et al., 2000) and microsomes (Smith et al., 1992, 1995, 2003), but the enantiomer composition of NNAL has not been measured. In the present investigation, a HPLC procedure was established for analysis of the enantiomers of NNAL that circumvents derivatization and is simpler than the earlier method (Carmella et al., 1999). It was applied to NNAL produced from NNK by individual oxidoreductases from human liver cytosol and by 11β-HSD1; in addition, the NNAL enantiomer composition was studied following NNK incubation with microsomes and cytosol from human lung, liver, and placenta and from rat lung and liver. Comparison of the results obtained in tissue fractions and purified enzymes, and experiments using an inhibitor of 11β-HSD1 were expected to provide insight into the nature of the enzymes predominantly catalyzing NNK reduction in vivo.
Materials and Methods
Chemicals. NNK and NNAL were purchased from Toronto Research Chemicals (Toronto, ON, Canada), (S)-MTPA-Cl and (R)-MBIC from Fluka (Seelze, Germany), 18β-glycyrrhetinic acid from Sigma (Deisenhofen, Germany), Ultima Gold scintillator from PerkinElmer Life and Analytical Sciences (Groningen, The Netherlands), biochemicals from Roche (Mannheim, Germany), and organic solvents of the highest purity from Merck (Darmstadt, Germany). [5-3H]NNK, 2.54 Ci/mmol (95% radiochemical purity), was obtained from Chemsyn through Campro Scientific (Berlin, Germany) and purified by thin-layer chromatography as described below for NNAL to a purity of >99.5%.
Tissue Samples. Human liver samples were either excess normal tissue obtained on partial hepatectomy for tumor metastases or parts of livers excluded from transplantation for medical reasons. They were cut into pieces of 5 to 10 g and stored at -70°C. Placental samples were obtained after full-term pregnancies terminated by vaginal delivery in five cases and by cesarean section in one. They were frozen within 30 min and stored at -70°C. Lung tissue (2-5 g) was excess normal material removed on bronchial tumor resection and frozen at -70°C within 15 to 30 min.
Adult male F344 rats and male and female Sprague-Dawley rats were purchased from Charles River Laboratories (Sulzfeld, Germany); they were fasted overnight and killed by decapitation. Livers of Sprague-Dawley rats were perfused with ice-cold saline before removal of livers and lungs. The tissues were cooled in saline and fractionated immediately.
Tissue Fractionation. From thawed samples of human liver or placenta, connective tissue was removed and the tissue was homogenized at 0-4°C by a tissue blender in 4 volumes of 250 mM sucrose containing 20 mM Tris and 5 mM EDTA-disodium and adjusted with HCl to pH 7.4 at 37°C. Cytosol and microsomes were obtained by differential centrifugation in the last step at 85,000g for 60 min, and microsomes were washed once. Human and rat lung tissue was freed from bronchi, homogenized in a glass-Teflon Potter-Elvehjem homogenizer in the above buffer, and fractionated by centrifugation to obtain cytosol and unwashed microsomes (Finckh et al., 2001). Rat liver was processed in the same way with microsomes being washed once. Microsomal preparations were suspended in buffer and stored at -70°C.
The enzymes AKR1C1, 1C2, and 1C4, and CR I and II were purified to electrophoretic homogeneity from human liver cytosol (Breyer-Pfaff and Nill, 2000) and 11β-HSD1 from human liver microsomes (Maser et al., 2002).
In Vitro Metabolism. Nonradioactive incubations were carried out at 37°C in a medium containing 100 mM Tris-HCl adjusted to pH 7.4 at 37°C, 25 mM KCl, 8 mM MgCl2, 0.2 mM NADP+, 1.9 mM glucose 6-phosphate, 1.2 U/ml glucose-6-phosphate dehydrogenase, and 10 to 1000 μM NNK in a total volume of 1 to 4 ml. When NADH was the cofactor, it was added to a concentration of 0.2 mM and the NADPH-generating system was omitted. After preincubation for 5 min, the reaction was started by adding cytosol or microsomes to a concentration of 1 mg/ml protein or purified proteins as indicated. For incubations in a CO atmosphere, CO was bubbled through the solution for 1.5 min before the addition of NNK and protein, and the tubes were sealed. After 15 to 60 min, the samples were cooled on ice, mixed with 250 mg of NaCl, 0.1 ml of ammonium hydroxide (25%), and 5 μl of 10% sodium deoxycholate per ml of incubation mixture, and extracted three times with an equal volume of ethyl acetate. The combined extracts were evaporated at 35°C under a stream of nitrogen.
In samples with [5-3H]NNK, the substrate concentration was 1 μM and the tritium concentration 2.54 μCi/ml, total volume 1 ml. The reaction was terminated by addition of 2 ml of ice-cold acetonitrile, and protein was removed by centrifugation. The supernatant was concentrated to about 0.2 ml under a nitrogen stream and, if necessary, filtered through a cellulose acetate membrane of 0.2-μm pore size (Lida, Rochester, NY) prior to gradient HPLC.
Thin-Layer Chromatography and Derivatization. NNAL extracted from incubation mixtures with ethyl acetate was purified by thin-layer chromatography (TLC) on silica gel in ethyl acetate/acetone/acetic acid (3:0.8:0.2, v/v). When cytosol or microsomes had been incubated, lipids were removed by running toluene/acetone (4:1, v/v) to the upper edge before separation in the above solvent. The chromatogram was moistened by spraying water, and the UV-absorbing band at RF 0.26, containing NNAL, was removed, suspended in 0.2 ml of 2 N ammonium hydroxide, and extracted three times with 0.7 ml of ethyl acetate.
Diastereomeric derivatives were synthesized from 1 to 50 μg of NNAL by a slight modification of the procedure of Hecht et al. (1997) using (S)-MTPA-Cl or (R)-MBIC. The derivatives were purified by TLC on silica gel in toluene/acetone (4:1) followed, after drying, by di-isopropyl ether/acetone/ammonium hydroxide (25%) (4:2:0.1, v/v). MTPA-NNAL (RF 0.52) or MBICNNAL (RF 0.38) was isolated by distribution between 0.2 ml of 2 N ammonium hydroxide and three times 0.7 ml of tert-butylmethyl ether.
HPLC Analyses. Detector responses were registered by the MT2 integration program (Kontron Instruments, München, Germany).
Gradient HPLC with radioflow detection was modified from the method of Peterson et al. (1991). The 250 × 4.6 mm column of Prodigy 5-μm ODS 100 Å (Phenomenex, Hösbach, Germany) was run with a gradient of solvent A (20 mM sodium acetate, pH 4.5) and solvent B (methanol) at a rate of 1 ml/min: 5% B for 5 min, increase to 35% B within 30 min, hold 35% B for 5 min, gradient to 5% B in 5 min, hold 5% B for 10 min. UV absorption was monitored at 260 nm and tritium with a solid-phase yttrium-glass scintillator cell in the HPLC Radioflow Detector LB 509 of Berthold Technologies (Bad Wildbad, Germany). (E)- and (Z)-NNAL were eluted at 27.0 and 27.6 min, and (E)- and (Z)-NNK at 36.6 and 37.3 min, respectively. The fraction containing [5-3H]NNAL was collected and, after addition of 250 mg of NaCl and 0.1 ml of ammonium hydroxide (25%) per ml eluate, was extracted three times with an equal volume of ethyl acetate.
NNAL produced from unlabeled NNK was analyzed on an aliquot of 5 to 10% by isocratic HPLC on a 250 × 4 mm C18-silica column (Hibar LiChrospher 100 RP-18; Merck) eluted (1 ml/min) with 10 mM sodium phosphate, pH 7.4/acetonitrile (82:18, v/v). The retention times of (E)- and (Z)-NNAL were 5.8 and 6.2 min, and those of (E)- and (Z)-NNK, 11.0 and 11.8 min, respectively. Quantitation was based on peak areas in relation to those of standards containing 0.15 to 0.75 nmol of NNAL. Recovery experiments were carried out with addition of racemic NNAL (2-35 nmol/ml) to incubation mixtures of human liver microsomes. They were processed in the same way as samples containing NNK. NNAL recoveries amounted to 79 ± 4% (mean ± S.D.) in five samples without and three with incubation for 60 min.
Enantiomer separation of underivatized NNAL was achieved by a modification of a chiral HPLC procedure (Landsiedel, 1998). The 250 × 4 mm column of cellulose tris-(N-3,5-dimethylphenylcarbamate) (Chira Grom 2; Grom, Herrenberg, Germany) was eluted at 1 ml/min with heptane/tert-butylmethyl ether/2-propanol (106:43:7, v/v) and the eluate monitored at 230 or 240 nm. Samples were dissolved in 10 μl of ethanol and diluted with eluent to 100 μl. The four NNAL isomers were eluted between 30 and 70 min. Alternatively, a 250 × 2 mm column with the same material was run at 0.3 ml/min, samples being injected in 50 μl of eluent and resulting in similar retention times. (R)-NNAL percentage was calculated as the ratio (areas of the two last peaks)/(areas of all four peaks). In samples with racemic NNAL, the (R)-NNAL content was found to be 50.4 ± 0.4% (mean ± S.D., n = 8).
When [5-3H]NNAL obtained by gradient HPLC was to be separated, 1 to 2 nmol of unlabeled racemic NNAL was added before injection, and fractions containing the four isomers were collected according to UV absorption. After evaporation under nitrogen, scintillator was added and tritium was counted (LS 3801; Beckman Coulter, München, Germany).
The MTPA and MBIC derivatives of NNAL were separated into their isomers by HPLC on a silica gel column according to the method of Hecht et al. (1997) using slightly modified eluents.
Results
Human Tissues and Enzymes. On a chiral HPLC column, the isomers of NNAL could be separated to near-baseline resolution (Fig. 2A). The order of elution was the same as that of the MTPA-NNAL isomers synthesized with (S)-MTPA-Cl and separated on silica gel (Fig. 2E). The assignment of the first two peaks to (S)-NNAL results from the predominant production of this enantiomer in human liver cytosol (Upadhyaya et al., 2000) and is in accordance with the data of Hecht et al. (1997) when the correct absolute configuration is taken into consideration (Hecht et al., 2000).
The oxidoreductases AKR1C1, AKR1C2, and CR I and II from human liver cytosol all reduced NNK predominantly to (S)-NNAL. Enantiomer compositions derived from analyses of NNAL and of MTPA-NNAL were in excellent accordance (Fig. 2; Table 1). The highest enantioselectivity was present with AKR1C1, which produced 99% (S)-enantiomer (Fig. 2, C and G), whereas the selectivity of the other enzymes including AKR1C4 from cytosol was slightly less (Table 1).
The assignment of HPLC peaks to NNAL isomers is further confirmed by comparison with chromatograms of MBIC derivatives of NNAL. These were better separated than in a previous publication (Hecht et al., 1997), each one of the isomers resulting in a separate peak, although baseline separation was not achieved (Fig. 3A). The first and third peaks representing (S)-(Z)- and (S)-(E)-NNAL, respectively, were the major ones when NNK had been reduced by CR II (Fig. 3B).
Cytosol from human liver (n = 4) and placenta (n = 3) reduced 250 μM NNK at rates of (mean ± S.D.) 730 ± 130 and 74 ± 34 pmol/min/mg of protein, respectively. (S)-NNAL was formed preferentially, the mean contribution of the (R)-enantiomer being 5% in both materials. When 1 μM [5-3H]NNK was incubated with placenta cytosol, (R)-NNAL was increased to 12%.
11β-HSD1 from human liver microsomes led to a (R)/(S)-NNAL ratio of about 1:2 (Table 1). In contrast, seven of eight human liver microsomal preparations produced more (R)- than (S)-NNAL on incubation with 1 to 250 μM NNK under carbon monoxide (Table 2; Fig. 3C). Incubation time did not influence the ratio since, when 250 μM NNK was incubated with microsomes for 20, 40, and 60 min with conversion of 8 to 25% of the substrate to NNAL, the (R)-enantiomer contributed 82, 81, and 80%, respectively, to total NNAL. NNK reduction rates were similar in liver microsomes and cytosol. In view of about 3- to 4-fold higher quantities of cytosolic protein, enzymes in cytosol are expected to make a proportionally larger contribution to total NNK reduction in liver. Microsomal preparations were routinely incubated in an atmosphere of CO to inhibit alternative metabolism of NNK and further metabolism of the NNAL enantiomers formed. Incubation under air instead of CO reduced the NNAL quantity recovered by 11 to 35% (mean 24%), but the enantiomer composition was unchanged (Table 3). 18β-Glycyrrhetinic acid at 0.1 μM, a concentration assumed to inhibit 11β-HSD1 selectively, produced a small nonsignificant decrease of the reduction rate and a small significant increase of the (R)-NNAL percentage; both changes were more marked and highly significant with 10 μM glycyrrhetinic acid (Table 4). With NADH as a cofactor, human liver microsomes reduced NNK at 5% of the rate measured with NADPH to NNAL containing 16 to 25% (R)-NNAL.
Using placental microsomes, the NNK reduction rate was about 5-fold lower than that with liver microsomes, and the proportion of (R)-NNAL was larger (Table 2). 18β-Glycyrrhetinic acid (10 μM) did not reduce the reduction rate significantly, and it only slightly increased the selectivity in favor of (R)-NNAL (Table 4). The rate of NNK reduction in placenta microsomes was nearly 80% higher per milligram of protein than in cytosol, but the recovery of microsomal protein was only 2 to 3 mg/g of tissue and, thus, about 20-fold lower than that of cytosolic proteins. The activity of the latter should therefore predominate in the organ.
In 10 of 11 lung samples, 250 μM NNK was reduced by microsomal enzymes with a high selectivity in favor of (R)-NNAL (89%) similar to that in placental microsomes (Table 5). Deviant results were obtained with lung sample 6, which originated from a cigar smoker who differed from the other patients neither with regard to clinical course nor laboratory findings, including histology. Microsomes from this lung exhibited a conspicuously low total reductase activity and produced NNAL with only 14.4 and 14.5% (R)-enantiomer in duplicate experiments. Very similar results were obtained on incubation of 1 μM [5-3H]NNK from which 19% (R)-NNAL was produced by lung 6 microsomes, whereas those from three other lungs resulted in a high proportion of (R)-enantiomer (mean 84%). Reduction rates were always about 250-fold lower at 1 μM than at 250 μM NNK, pointing to apparent Km values in the millimolar range. The influence of incubation time was studied for lung 5 microsomes, which produced 89.6% (R)-NNAL within 20 as well as within 60 min. Results were the same whether lung microsomes were incubated under air or under CO (Table 3).18β-Glycyrrhetinic acid (0.1 μM) failed to affect the NNK reduction rate and the enantiomer composition of NNAL (Table 4).
Lung cytosol reduced 1 and 250 μM NNK preferentially to (S)-NNAL, the mean (R)-enantiomer content being 19 and 14%, respectively. NNK reduction per milligram of protein occurred at about an 8-fold lower rate in cytosol than in microsomes (Table 5). Since the protein quantity recovered in cytosol was about 8-fold that in microsomes, similar contributions of the two cell fractions to NNK reduction in human lung can be assumed.
Rat Tissues. Lung microsomes from male and female Sprague-Dawley rats converted NNK to NNAL with an (R)/(S) ratio around 1, whereas those from male F344 rats produced a small excess of (R)-NNAL. In lung cytosol, the (S)-enantiomer clearly predominated in both strains and sexes (Table 6). On average, the reduction rate was 6- to 8-fold higher per milligram of microsomal protein than per milligram of cytosolic protein from the same tissue, but in view of an approximately 6-fold higher content of cytosolic proteins, the latter have an equally important part. In the absence of CO, the NNAL production by lung microsomes was nearly unchanged (Table 3). 18β-Glycyrrhetinic acid, even at a low concentration, markedly lowered the NNAL quantity and increased its (R)-enantiomer content (Table 4).
NNK reduction by NADPH proceeded at a high rate in liver microsomes from female Sprague-Dawley and male F344 rats and was even significantly faster in those from Sprague-Dawley males. The resulting NNAL consisted of the (R)-enantiomer by about one-third (Table 6). When NADH was the cofactor, the mean reduction rates were similar (2600 and 2200 pmol/min/mg of protein in microsomes from male and female Sprague-Dawley rats, respectively), but the product contained on average only 1.8 and 0.7% (R)-NNAL, respectively. Incubations of liver microsomes from Sprague-Dawley males with NADPH in the absence of CO resulted in NNAL quantities that were on average 19% lower than in a CO atmosphere, whereas the (R)-NNAL content was slightly reduced (Table 3). The 0.1 μM concentration of 18β-glycyrrhetinic acid led to a small decrease of the reduction rate without changing the NNAL composition, whereas at 10 μM, the reaction was markedly inhibited and the percentage of (R)-NNAL increased (Table 4).
Rat liver cytosol reduced NNK at a low rate with production of 11 to 17% (R)-enantiomer (Table 6). In contrast to human liver, NNK reduction in rat liver seems to be predominantly catalyzed by microsomal proteins. These exhibited on average 74-, 25-, and 21-fold higher reaction rates than cytosolic proteins from male and female Sprague-Dawley and male F344 rat liver, respectively. Quantities of cytosolic proteins exceeded those of microsomal proteins by a factor of about 4, such that the contribution of cytosol to NNK reduction in liver can be estimated to amount to 5 to 16%.
Discussion
The results of the present study are in accordance with published data inasmuch as NNK reduction to NNAL is catalyzed by the microsomal and cytosolic fractions of human and rodent liver and lung (Smith et al., 1992, 1995, 2003; Atalla and Maser, 1999; Maser et al., 2000; Upadhyaya et al., 2000). Preferential formation of (S)-NNAL in cytosol from the two organs in human and rat (Upadhyaya et al., 2000) could also be confirmed. In the case of human liver, this is in accordance with the enantioselectivity of individual oxidoreductases (Table 1), which are the major NNK-reducing enzymes in cytosol (Atalla et al., 2000). For unknown reasons, cytosol produced a somewhat larger fraction of (R)-NNAL than did the purified enzymes. CR present in human placenta cytosol (Westbrook et al., 1978; Wermuth et al., 1988) has, according to inhibition experiments, an important role in NNK reduction (Atalla and Maser, 2001). This is compatible with the high proportion of (S)-NNAL found in the product.
On the other hand, discrepancies became apparent on measuring the enantioselectivity of the reaction in microsomes. The only microsomal enzyme known to catalyze NNK reduction is 11β-HSD1 (Maser et al., 1996). However, the enzyme purified from human liver microsomes (Maser et al., 2002) produced NNAL with 35% (R)-enantiomer, whereas about 70% (R)-NNAL was found on reduction by microsomes (Table 2). Provided that the enantioselectivity of the enzyme is not altered on isolation, the major part of microsomal NNK reduction must be due to one or more additional enzymes that produce more (R)- than (S)-NNAL. This is further substantiated by the experiments with 18β-glycyrrhetinic acid, which inhibits 11β-HSD1 with a Ki value of about 10 nM (Maser et al., 2002). A concentration of 0.1 μM should inhibit the enzyme nearly completely, but the mean decrease of the NNAL production rate in human liver microsomes was only 10% (Table 4). Higher concentrations led to a more pronounced inhibition, pointing to a loss of specificity for 11β-HSD1.
Similar considerations on the NNK-reducing enzymes apply to microsomes from human lung (with the exception of lung 6) and placenta, the NNAL produced by them containing around 90% (R) enantiomer. The unusually low (R)/(S)-NNAL ratio in microsomes from lung 6 (Table 5) is probably due to deficiency of an enzyme with predominant (R)-NNAL production. This deficiency profoundly decreased NNK reduction rate and therefore may be a risk factor for lung tumor initiation by NNK. Although 11β-HSD1 could be shown to be expressed in placental tissue (Atalla and Maser, 2001), 18β-glycyrrhetinic acid, even at 10 μM, failed to reduce the NNK reduction rate significantly and increased the (R)-NNAL percentage only slightly (Table 4); this further points to a minor role, at best, of the enzyme in placenta. Also, in lung microsomes, the quantity and (R)-enantiomer content of NNAL were not changed by 18β-glycyrrhetinic acid (Table 4), a result not compatible with a significant contribution of 11β-HSD1. In agreement with this, NNAL formation was not inhibited by metyrapone (Smith et al., 1995), a substrate of 11β-HSD1 (Maser and Bannenberg, 1994). The failure of CO in the atmosphere to substantially affect the NNAL quantity is in accordance with previous data (Smith et al., 1992, 2003). There is also agreement with regard to intrinsic clearances, which can be calculated as ratios of reaction rates and substrate concentrations when these are sufficiently below Km. Values given in Table 5 for lung microsomes result in 2.8 ± 1.0 μl/min/mg of protein, whereas data published previously range between 0.55 and 4.1 μl/min/mg of protein (Smith et al., 1992, 1995, 2003). Equal contributions of microsomes and cytosol to NNK carbonyl reduction in lung would result in NNAL with an (R)/(S) ratio around 1. Since lung is an extremely heterogeneous organ comprising at least 40 different cell types (Sorokin, 1970), the local formation of NNAL with one of the enantiomers predominating by far cannot be excluded.
According to the data in Table 2, human liver microsomes formed about twice as much (R)- as (S)-NNAL, whereas the reverse was reported by Upadhyaya et al. (2000). The higher substrate concentrations and longer incubation times used in the present study in the majority of experiments could be excluded as causing the discrepancy. Data in rats also disagreed, since for the product of NNK reduction in lung microsomes, a value of 96% (S)-NNAL was published previously. In contrast, less than 50% resulted in the present investigation in Sprague-Dawley rats as well as in F344 rats (Table 6), the strain used previously. Analogously, rat liver microsomes now were found to produce 30 to 37% (R)-NNAL, whereas a value of 97% (S)-NNAL was given by Upadhyaya et al. (2000). Discrepancies also became apparent on comparing the intrinsic clearances for NNK reduction in rat liver fractions. For liver microsomes and cytosol from male F344 rats, the present data give mean ratios of (reaction rate)/(substrate concentration) of 5.6 and 0.26 μl/min/mg of protein, whereas values of 310 and 47 μl/min/mg of protein, respectively, can be calculated from the previous data (Upadhyaya et al., 2000). On the other hand, data by Guo et al. (1992) result in 13 μl/min/mg of protein for liver microsomes from male Sprague-Dawley rats and are therefore in good accordance with the value of 10.4 μl/min/mg of protein found currently in the same strain.
Comparison of NNK reduction experiments under air and under CO revealed higher NNAL recoveries in the presence of CO (Table 3) in accordance with a pronounced inhibition of NNK oxidation by CO in rat liver microsomes (Guo et al., 1992). Higher NNK-reducing activities in microsomes from male than from female rat liver parallel data on reduction of 2-(2-amino-5-bromobenzoyl)-pyridine (Hara et al., 1987) or acetohexamide (Imamura et al., 1995). Of the microsomal oxidoreductases acting on steroids, 11β-HSD1 (Low et al., 1994) and 3β-hydroxysteroid dehydrogenase/Δ5-Δ4-isomerase (Couet et al., 1992) are expressed more highly in male than in female rat liver.
The present data are compatible with those on in vivo metabolism of NNK in bile duct-cannulated rats, which resulted in a 14% recovery of the dose as (R)-NNAL glucuronide in bile and another 3% in urine (Wu et al., 2002). As can be deduced from Table 6, (R)-NNAL represents about one-fourth of the NNK reduction product in rat liver and lung. In conjunction with preferential (R)-NNAL glucuronidation (Ren et al., 1999), this finding can explain the relatively high recovery of the conjugate. In urine from humans exposed to tobacco, (S)-NNAL quantities exceed those of the (R)-enantiomer 1.2-fold and in the form of glucuronides, up to 3-fold (Carmella et al., 1999; Hecht et al., 2002), which is in good agreement with the ratios found in vitro. What can be inferred on the nature of the microsomal carbonyl-reducing enzymes acting on NNK? 11β-HSD1 seems to have an appreciable role in rat lung, a minor one in human and rat liver, and no quantifiable role in human lung and placenta. Microsomes from all tissues except rat liver, when incubated with 18β-glycyrrhetinic acid at 1 or 10 μM, produced NNAL with a mean (R)-enantiomer content of 83 to 91% (Table 4). Possibly, they contain the same enzyme with a high enantioselectivity for (R)-NNAL and a low sensitivity toward inhibition by 18β-glycyrrhetinic acid. Attempts to purify it have hitherto not been successful. However, fractionation of NNK-reducing enzymes in human liver microsomes led to partial purification of a protein producing (S)-NNAL with high selectivity (Martin and Maser, unpublished results). Therefore, at least three enzymes seem to participate in NNK reduction in human liver microsomes. The same applies to rat liver microsomes, because with NADH as a cofactor, they formed (S)-NNAL nearly exclusively.
In conclusion, the present data point to similar contributions of microsomal and cytosolic enzymes to NNK carbonyl reduction by human liver and lung and a predominance of cytosolic enzymes in placenta. The enantioselectivity in cytosol is compatible with a dominant role of CR, AKR1C1, AKR1C2, and AKR1C4, which produce 93 to 99% (S)-NNAL. On the other hand, microsomal 11β-HSD1 isolated from human liver reduces NNK to about twice as much (S) as (R)-NNAL, in contrast to complete microsomes, which result in about 70% (R)-NNAL. Still higher selectivities in favor of (R)-NNAL were observed with microsomes from placenta and from 10 of 11 lung samples. Thus, an NNK-reducing enzyme other than 11β-HSD1, that forms (R)-NNAL selectively, must occur in microsomes. In rat liver, NNK reduction is predominantly catalyzed by enzymes in microsomes with production of 30 to 37% (R)-NNAL.
Acknowledgments
Human placental tissue was kindly provided by Dr. R. Kurek, Department of Gynecology and Obstetrics, and liver tissue by Dr. W. Lauchart, Department of Surgery, University Clinic of Tübingen. We are grateful to Karl Nill for expert technical assistance in HPLC analyses.
Footnotes
-
Supported by Grant BR 478/14-1 from the Deutsche Forschungsgemeinschaft.
-
ABBREVIATIONS: NNK, 4-methylnitrosamino-1-(3-pyridyl)-1-butanone; AKR, aldo-keto reductase; CR, carbonyl reductase; HPLC, high-performance liquid chromatography; 11β-HSD1, 11β-hydroxysteroid dehydrogenase type 1; (R)-MBIC, (R)-(+)-α-methylbenzylisocyanate; (S)-MTPA, (S)-(+)-α-methoxy-α-trifluoromethylphenylacetyl; NNAL, 4-methylnitrosamino-1-(3-pyridyl)-1-butanol; TLC, thin-layer chromatography.
- Received March 24, 2004.
- Accepted May 24, 2004.
- The American Society for Pharmacology and Experimental Therapeutics