Tobacco Documents Online

: ; ~~. r I Tobacco Specific 1V Nitrosamines: Occurrence, Carcinogenicity, and Metabolism STEPIII:N S. HECHT, CHI-HONGB. CHEN; C: DAVID McCOY, and DIETRICH HOFFMANN Navlon Dana Institute for Disease Prevention, American Health Foundation, Valhallh, NY 10595 It is now widely accepted that cigarette smoking,causes lung, cancer (1,2). It is less widely known that smoking is also cor- reliated with an increased incidence of cancer of the oral cavity, esophagus, pancreas andibladd'er (2,3,4,5,6). Tobacco chewing can also cause oral cavity and esophageal cancer (3,4,7). In fact, oral cavity cancer is a major cancer among men in India, where the habit of chewing the betel quid containing tobacco is widespread (8). Cigarette smoke is known to contain tumor initiators such as the polynuclear aromatic hydrocarbons, and tumor promoters and cocarcinogens, such as catechol (9). These agents can explain many of the observed effects of cigarette smoke condensates in experimectal animals and almost certainly are involved im some of the human cancers associated with smoking. However, ni~trosamines may also be causative factors in the tobacco related cancers, especially in those organs which are remote from direct contact with tobacco or tobacco smoke. Thus it is known that nitrosamines canicause esophageal, pancreas and bladder cancer in experimental animals, as well as affecting the lung and oral cavity (10,11,12). Since tobacco and tobacco smoke have specific carcinogenic effects in man, it is tempti~ng,to speculate that there may be unique carcinogenic agents in tobacco and tobacco smoke. The tobacco specific nitrosamines are such a group. These nitros- amines are derived from the tobacco alkaloids (see Figure 1). The most prevalent alkaloid is nicotine, which occurs in general in concentrations of 1-2$,in commercial tobacco products. Both nico- tine and ncrnicotine could give rise to the prototype of tobacco specific nitrosamines, N'-nitrosonornilcotine (NNN). Nicotine could also be nitrosated to form 4-(N-methyl-N-nitrosami~no)-1- (3-pyridyl)-1-butanone (NNIf), or 4-(N-methyl-N-nitrosamino)-4- (3-pyridyl)butanal (NNA). In addition, N-nitrosopyrrolidine (NPy) could a1iso be derived from ni~cotine and nornicotine. Nitrosation of anabasine would give nitrosoanabasi~ne (NAB). The structures of these nitrosamines, which will be considered in this review, are shown in Figure 2. Of course, inspection of Figure 1 reveals other interesting possibiliti~es for nitrosation of the tobacco

126 N-NITROSAMINES NICOTINE 0 DEHYDRONICOTINE NICOTYRINE COTININE ANABASINE NORNICOTINE ANATABINE MYOSMINE BIPYRIDYL Figure 1. Common tobacco alkaloids in tobacco arul/or tobacco smoke 10 N-CH3 i I .., . 1LJ1i I cn3 - N - N°0 NNN NNK NNA NAB Figure 2. Some nitrosamines which can be derived from the tobacco alkaloids

1 8. xEcirr ET wi.. Tobacco Specific N-Nitrosamines 127 alkaloids; some of these will be the subject of future studies. Occurrence And Formation Of Tobacco Specific Nitrosamines The prototype of the tobacco specific nitrosamines, NNN, has been detected in both tobacco smoke and unburned tobacco. Various analytical methods have been used including gas chromatography (GLC) (1!3,114,15,16) combined GLC-mass spectrometry (L7), thin, layer chromatography (118),high pressure Liqui&chromatography' (HPLC) (19,20), andicombined'HPLC-thermaL energy analysis (21). NNN levels in cigarette smoke typically range from 140-240 ng/cig in a typical American 85mm non-filter cigarette. Surprisingly high lievels of NNN were found in unburned tobacco (0.3-9.0 ppm in, cigarette tobacco, 3.0-45.3 ppm in cigar tobacco, 3.5-90.6 ppm in chewing tobacco, and 12.1-29.1i ppm in snuff). These levels are among the highest for an environmental nitrosamine in terms of occurrence and human exposure (22). Thus, rather detailed studies were carried out to determine the origins of NNN in tobacco andi tobacco smoke. To study the formation of NNN in tobacco, plants were an- alyzed at various stages of growth and curiing,(23). NNN was not detected prior to harvest or in freshly harvested Burley tobacco but only during and after air curing (0.5-1.1 ppm). Since either nicotine or nornicotine could have been al~recursor to NNN in tobacco, tobacco leaves were fed nicotine-2'- C or nornico- tilne-2'-14C and:cured (24). The cured leaves were then analyzed for NNN-2'-14C. The yield of NNN frominicotine was 0.009% and from nornicotine, 0.007%. These results showed that both nico- tine and nornicotilne could be precursors to NNN in tobacco. How- ever, the greater abundance of nicotine in tobacco leaf (20-100 times the concentration of nornicotine) favore&nicotine as the major precursor of NNN in tobacco. The transfer of NNN from cigarette tobacco to mainstream smoke was studied (20). For this purpose, NNN-2''-14C was added to cigarettes and the smoke was analyzed. The transfer rate was found to be 11.3%. Since, in this experiment,, the tobacco column smoked contained 974 ng NNN,, 110 ng were transfered to the main- stream smoke. Analysis of the mainstream smoke revealed 238 ng NNN; thus the remaining 128 ng were formed duri~ng,smoking. There- fore, about 50% of the NNN ih mainstream smoke originated by transfer from tobacco while the remainder was formed during smok- ing. Either nicotine or nornicotine could be a precursor to NNN formed during smoking. To examine this question, nicotine or nornicotine was added to cigarettes and the smoke was analyzed: for NNN (13),. In each case, NNN concentration in smoke increased indicating that both alkaloids are precursors to NNN formed dur- ing smoking. However, nicotine is considered:the more important precursor due to its higher concentration in tobacco. The results of these studies on the formation of NNN during curing, its

i, 128 W nicotine (26,29,30,31). The possibility that tobacco bacteria ~ could nitrosate nicotine via this pathway is currently under in- 0 vestigation. ~ N-NITROSAT4INES transfer to smoke, and its formation during smoki~ng,are summarized' in Figure 3. In tobacco samples examined so far, the 1'evels of NAB were significantly less than those of NNN. In fact, NAB has not yet been detected with certainty in unburned tobacco (15). These findings are iniline with the major role of nicotine rather than nornicotine as a precursor to NNN'since kinetic studies showed that nornicotine and anabasine were nitrosated at similar rates (25). These rates are relatively high, which suggests that the formation of NNN and NAB could be favored in vivo. When chewing tobacco was i~ncubated:with human saliva for 3 hours at 37° and the mixture analyzed for NNN, the concentrations of NNN increased by 44% over that in the chewing tobacco, presumably as a result of further nitrosation (15). Thus, in vivo formation of NNN and NAB could constitute an additional exposure of smokers or chewers to these tobacco specific nitrosam,ines. Since nicotine is the major precursor to NNN in tobacco and tobacco smoke, the reaction of nicotine with sodium nitrite was studied to provide information on formatiomof other tobacco specific nitrosamines, especially NNK and NNA, which could~arise by oxidative cleavage of the 1'-2' bonds or 1'-5' bond of nicotine followed by nitrosation (26). The reaction-was investigated under a variety of conditions as summarized im Table L. All three nitrosamines were formed when the reaction was done under rela- tivelly mild conditions (17 hrs, 200). The yields are typical of the formation,of nitrosamines from tertiary amines (27). At 90°, with a five fold excess of nitrite, only NNN and NNK were detected. Under these conditions, both NNK and NNA gave secondary products. NNK was nitrosated a to the carbonyl to yield 4-(N-methyl-N- nitrosamino)-2-oximi~no-1-(3-pyridyl')-1i-butanone while NNA under- went cyclization followed:by oxidation, decarboxylation and de- hydration to give 1-methyl-5-(3=pyridyl)pyrazole, as shown in Figure 4. Extensive fragmentation and oxidation of the pyrroli- dine ring was also observed under these conditions. The products of the reaction of nicotine and nitrite at 90° are summarized in Table II. The formation of NNN, NNK, and NNA from nicotine probably in- volived the intermediacy of cyclic iminium salts, as shown in Fig- ure 5 (28). These salts can undergo hydrolysis to the free amines which are niitrosated:, or at near neutral pH, can be directly nitrosated to give nitrosamines. The formation of nitrosamines from iminiumisalts under neutral conditions has been demonstrated in at least two studi~es and is of interest because iminium salts are known to be intermediates in the mammalian metabolism of The formation studies encouraged and tobacco smoke. of NNK and NNA from nicotine in these model us to search for these nitrosamines in tobacco In studies undertaken so far, NNK but not NNA

8. HECHT ET AL. Tovacco'S)2ecirtc N-Nitrosamines 129 FRESH HARVESTED TOBACCO NICOTINE 0.009°h 0.007% NORNICOTINE NITRITE \ / NITRITE CURED TOBACCO MAINSTREAM CIGARETTE SMOKE NNN International Agency for Research on Cancer Figure 3. Origins of NNN, in tobacco and tobacco smoke (22) Table I Buffer systems: pH2, KC1-HC1; pH 3.4-7, citrate-phosphate. cND=not detected. Formation of NNNNNK, and NNA from Nicotine and NaNO2 [NaN021, Conditions Yields ($)a' [Nicotine) pH-b T(°C) t(hrs) NNN! NNK NNA 1.4 2.0 20 17 0.1 AIDc 0.2 1.4 3.4 20 17 0.5 0.1 2.8 1.4 4.5 20 17 0.5 0.5 2.3 1.4 7.0 20 17 0.2 0.1 0.1 5.0 3.4-4.2 90 0.3 8.0 0.7 ND 5.0 3.4-4.2 90 3.0~ 8.8 2.3 ND 5.0 3.4-4.2 90 6.0 8.0 1.5 ND 5.0 5.4-5.9 90 0.3 9.0 2.7 ND 5.0 5.4-5.9 90 3.-0~ 13.5 4.3: ND 5.0 5.4-5.9 90 6.0 11.7 2.6 ND 5.0 7.0-7.3 90 0.3 1.3 0.1 ND 5.0 7.0-7.3 90 3.0: 4.5 0.2' ND 5.0 7.0-7.3 90 6.0 5.5 0.2 ND aDeterm,ined by GC and based on starting nicotine. b NtCOTiNE NNN OXIDES ~ OXIDES NITROGEN NORNICOTINE NITROGEN'N

130 N-NITROSAMINES. Table II Products Formed in the Reaction of Nicotine and NaNOZa,b Product Yield($)c , Method of d I~dentification NNN' 8.8 A NNK 2.3 A 4-(N-methyl-N-nitrosamino)- 4.0 C,A 2-oximino-l^{3'-pyridyl)- 1-butanone 1-methyl-5-(3-pyridyl)- 2.1 CC pyrazole cis and trans 3-Pyr-CH=CHCN 19.0 A 3-Pyr-CONHCH3 6.2 B 3-Py r-COOH 4.0 B cotininee 0.6 A 3-Pyr-CH=CH-COOH 0i.5 B 3-Pyr-COCH3 0.5 B 3-Pyr-CN 0.5 B 3-Pyr-CO2CH3 0.3 B 3-Pyr-CHO 0.2 B myosminef 0.1 A 3-Pyr-CH2CN 0.1 B aReaction of I equivalent nicotine with 5 equivalents NaNO2 at 90°.3 h, pH 3.4-4.2' b15-25$ nicoti~ne was unreacted. cBased:on starting nicotine. d A; comparison of GC or HPLC retention times and mass spectra to independently synthesize&standard5; B, compari~son to commer- cially available standards; C, spectral properties. e1-methyl-5-(3-pyridyl)-2-pyrroli~dinone. f2-(3-pyridyl)-li-pyrroline.

r 8. HECHT ET AL. Tobacco Specific NLNitrosamines O' 131 CHO C02H N H IN O N. -OH P _ - O, N OH --+ N CH3 -CH2 N I CH3 2 . 1 CH3, Figure 4. Formation of l-methyl-5•(3-pyridyl)pyrazoie from NNA - HNO~ r---% - HW 1 CHj N n- 'N -(E) . (P CHJ NO Hl0 HNO o ---NNA - NNK RJ HONO NH HONOi I CH„ n 'N ® 11 / OHi \ IJ/Hi0 \\,NO}O R=J.pyr H HONO NNN International Agency tonResearchon Cancer Figure 5, Formation of tobacco specific nitrosamines from nicotine (22)

132 N-NITROSAMINES has been detected. NNK was most readily analyzed by combined FPLC-TEA, although conventional HPLC methods have also been used (21,24Y. Levels of NNN and NNK in tobacco and mainstream cigar- ette smoke are summarized in Table III. During these studies by HPLC-TEA, we also identified N'-•nitrosoanatabine in tobacco (0.44- 3.2 ppm) and mainstream (.0.33-4.6 Pg/cig) and si~destream cigar- ette smoke (0.15-1.5ug/cig). The analytical studies on NNN dis- cussed in this section were done using NNN-2''-14C as internal standard. Tobacco was extracted with aqueous ascorbic acid and smoke was collected in traps containing ascorbic aci&to prevent artifactual formation of nitrosamines. Table III Non-VoLatile N-Nitrosamines in Tobacco And Tobacco Smoke Ma}nstream (pg/cig) Sidestream (ug/cig) Tobacco (ppm) Product NNN NNK NNN NNK NNN NNK Burley,NF 3.7 0.32 6.1 0.66 7.0 N.D. Bright,NF 0.62 0.42 1.7 0.50 0.22 0.37 Commercial,NF 0.24 0.11 1.7 0.41 1_7 0.74 Commercial,F 0.31 0.119 0.15 0.19 1.4 0.70 Kentucky,lR1,NF 0.39 0.116 0.21 0.24 0.63 0.13 Little Cigar,F 5.5 4.2 0.88 0.81 45.3 35.4 Columbia Cigar 3.2 1.9 116.6 16.7 10.7 1.1 (5.7g)1 N.D. = Not detected Carcinogenicity Of Tobacco Specific Nitrosami~nes The earliest studies on the carcinogenicity of NAB and NNN were done by Boyland and co-workers, who demonstrated that NAB caused esophageal tumors in rats and that NNN,induced lung ade- nomas in mice (32,33). NAB was administere&to rats in the drink- ing water (total dose, 7.9-11.5 mmoles/rat) an&25 of 32'rats treated developed tumors of the esophagus with the tumors appear- ing after 50-70 weeks of treatment. NNN was injected in mice (total dose, 0.5 mmoi'/mouse) and 7 of 40 mice d'evei'oped pulmonary adenomas, compared to 1 of 30 mice in the control groups. In our own studies, the carcinogenicity of NNN and NAB'was first compared in male Fischer rats (34). Each compound was administered in the driinking water for 30 weeks (total dose; 3.3 tiamles NAB, 3.6 mmlies NNN/rat) to a group of 20 animals. After 48: weeks, the experiment was terminated. In the NNN group, 14 of 20 animals developed tumors; these were mainly esophageal papil- lomas and carcinomas. One pharyngeal tumor and 3 nasal cavity carcinomas were also observed. By contrast, NAB at this dose gave only 2 of 20 tumor bearing animals. Thus NNN was a moderately.

8. HECfrr Er AL. Tobacco Speci fic N-Nitrosamines 133 active carcinogen whereas NAB was only weakly active. The lower tumor yield for NAB in this experiment compared to Boyland's work was probably due to the lower dose of NAB and the shorter lifetime of the animals. The carcinogenicity of NNN in Sprague-Dawley rats was examin- ed by Singer and Taylor (35):. NNN was given in the drinking water for 44 weeks (total dose 8.8 mmoles/rat) to a group of 15 female rats. All the rats were dead by 46 weeks and all 15 animals had adenocarcinomas of the olfactory epithelium. In a parallel: study, the carcinogenicity of NPy was examined in male and female rats of the same strain (36). NPy was added to the drinking water for 50 weeks (total dose, 10.0 mmoles/rat). The females were dead after 85 weeks and the males, after 104 weeks. NPy induced hepato- cellular tumors in 13 of 14 males and in 14 of 15 females. Thus NNN was a stronger carcinogen than NPy, when judged by time until death. However, the target organs were different in each case. The tumorigenic activities of NNN and NAB were also compared' in Syrian Golden hamsters (37). In this experiment, NNN and NAB were each given.by subcutaneous injection for a period of 25 weeks (total dose; 2 mmoles/hamster). Within 83 weeks, 12 of 19 ham- sters given NNN developed tracheal tuaars and:1 had a carcinoma of the nasal cavity. In the same period, none of the animals treated with NAB developed tumors. Nitrosopiperidine was included as a positive control and induced tracheal tumors in all the animals after a total dose of 1.3 mmole/hamster. Thus, substitution of a pyridine ring adjacent to the ring nitrogen of nitrosopiperidine to give NAB resulted in a significant reduction in carcinogenic activity; this effect was also observed in Boyland's experiments on rats (32Y. Such an effect was not observed when NNN and NPy were compared (35,36)~. This is of interest when considering the mechanism of action of these compounds. The tumorigenic activities of NNN, NNK, and NNA were compared in strain A mice (24). Each compound was injected over a period of 4even weeks with a total dose of 0.1 mmole/mouse. For reasons of solubil'ity, NNK was injected as a suspension in trioctanoin while NNA was injected in sali~ne. For comparison, NNN was in- jectediboth in saline and trioctanoin. The positive control was urethan. The results are summarized in Table IV. As judged by multiplicity of lung tumors, both NNN and NNK showed significant ~' activity (P < 0.05) compared to controls and NNK was significantly Q more active tumorigenic ( P< 0.05), than NNN. NNA did not show significant activity. in this strain of mice 0 The greater tumorigenicity of NNK than NNN Q is indicative of potentialliy higher car- O-A cinogenicity in other rodent species; ly in progress. these bioassays are current- N Metabolic Studies On NPy, NNN, and NNK ~ (D 0 Nitrosamines, like many other classes of chemical carcinogens must undergo metabolic transformation to be converted into elect-

134 Table IV N-NITROSAMINES Bioassay in Strain A Mice of Nitrosamines Derived from Nicotine Lung % Lung Lung, Effective Adenoma Adenoma Adenomas Experimental Group No. of Animal Bearing Bearing s Animals Animals per Animal Others 1. Untreated 25 1 4 Q.D4 Control 2. Vehicle 25 3 12 0.24 Control(Saline). 3. Vphi~cle 24 5 21 01.20 Control (trioctanoin) 4. Urethane in 25 25(6) * 100 14.80 Saline 5. NNN in Saline 21 16 76 1.74 Undiffer- entiated Carcinoma of Salivary Glands . NNN in 3 12(1)* 7 .87 1 (Metasta- sis: Lungs, Pleura) Undiffer- Trioctanoin . NNA in Saline 5 9 6 .44 entiated Carcinoma of Salivary Glands 1,. Malignant Lymphoma 1 ~ 8. NNK in Trioctanoin 23 20, 87 2.61 C ~ *( ) Adenocarcinoma &