Simultaneous Determination of Five Tobacco ... - ACS Publications

advantages in terms of reduced sample preparation and analytical .... A custom-written Access macro converted ... (SAS) software (SAS Institute, Cary,...
1 downloads 0 Views 93KB Size
Anal. Chem. 2003, 75, 4827-4832

Simultaneous Determination of Five Tobacco-Specific Nitrosamines in Mainstream Cigarette Smoke by Isotope Dilution Liquid Chromatography/Electrospray Ionization Tandem Mass Spectrometry Weijia Wu, David L. Ashley, and Clifford H. Watson*

Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, 4770 Buford Highway, Atlanta, Georgia 30341

Tobacco-specific nitrosamines (TSNAs) have been previously implicated as a source of carcinogenicity in tobacco and cigarette smoke. Accurate quantification of these chemicals is needed to help assess public health risk. We have developed and validated a specific and sensitive method to simultaneously measure five TSNAs in the particulate phase of mainstream tobacco smoke. Cigarette smoke particulate, produced using standardized machine smoking protocols, was collected on a Cambridge filter pad. The particulate matter was extracted using methylene chloride, back extracted into aqueous solution, further purified by solid-phase extraction, and analyzed by liquid chromatography/electrospray ionization tandem mass spectrometry using isotopically labeled analogues as internal standards. Limits of detection for this method ranged from 0.05 to 1.23 ng/mL using an injection volume of 20 µL. A linear calibration range spanning 2.52500 ng/mL was adequate to measure TSNA levels in cigarette smoke. The method achieved excellent reproducibility and accuracy. The identity of each TSNA was established by chromatographic retention time, analytespecific fragmentation patterns, and relative peak area ratios of two product/precursor ion pairs. This new method provides higher sensitivity, specificity, and throughput than earlier methods for TSNA determination. Cigarette smoking is widely acknowledged as the leading cause of lung cancer and related preventable diseases.1 It is well documented that tobacco smoke contains many mutagenic and carcinogenic chemicals.2 Of these, tobacco-specific nitrosamines (TSNAs) are among the most abundant carcinogens identified in tobacco and its smoke. Seven TSNAs, 4-(methylnitrosamino)-1(3-pyridyl)-1-butanone (NNK), N′-nitrosonornicotine (NNN), N′nitrosoanabasine (NAB), N′-nitrosoanatabine (NAT), 4-(methylni* Corresponding author. Tel: (770) 488-7638. Fax: (770) 488-0181, E-mail: [email protected]. (1) Blot, W. J.; Fraumeni, J. F., Jr. Cancers of the lung and pleura. In Cancer Epidemiology and Prevention; Schotenfield, D., Fraumeni, J., Jr.; Eds.; Oxford University Press: New York, 1996; pp 637-65. (2) Hecht, S. S. J Natl. Cancer Inst. 1999, 91 (14), 1194-210. 10.1021/ac030135y Not subject to U.S. Copyright. Publ. 2003 Am. Chem. Soc.

Published on Web 08/16/2003

trosamino)-1-(3-pyridyl)-1-butanol (NNAL), 4-(methylnitrosamino)4-(3-pyridyl)-1-butanol (iso-NNAL), and 4-(methylnitrosamino)4-(3-pyridyl)butanoic acid (iso-NNAC), have been identified both in tobacco filler and tobacco smoke (Figure 1). The levels of NNN, NNK and NAT generally are found to be higher than the other TSNAs. Of the seven TSNAs, NNN, NNK, and NNAL are considered the most carcinogenic.3 NNK and its primary metabolite, NNAL, are thought to be particularly important in the induction of adenocarcinoma, which is now the leading lung cancer type in the United States.4 Because of health implications, there has been a longstanding interest in analyzing TSNAs in tobacco products. Efforts to measure TSNA began as early as the 1960s. Boyland et al. attempted to analyze tobacco smoke for NAB and NNN in 1964,5 but their efforts were not successful. In 1974, Hoffmann et al. first reported a quantitative method measuring NNN in the tobacco smoke and chewing tobacco using gas chromatography (GC).6 Shortly thereafter, several additional analytical methods for determining NNN featuring high-performance liquid chromatography (HPLC),7 GC,8 and gas chromatography/ mass spectrometry (GC/MS)9 were published. The application of HPLC with UV detection identified the second TSNA, NNK, in 1978.10 The introduction of thermal energy analyzer (TEA), a chemiluminescence detector that is highly sensitive and relatively specific for nitrosamine compounds,11 provided a means of identifying and quantifying additional TSNAs. An HPLC method coupled with TEA led to the identification of third TSNA, NAT.12 Subsequent work (3) Hecht, S. S. Chem. Res. Toxicol. 1998, 11, 559-603. (4) Thun, M. J.; Lilly, C. A.; Flannery, J. T.; Calle, E. E.; Flanders, W. D. J. Natl. Cancer Inst. 1997, 89, 1580-6. (5) Boyland, E.; Roe, F. J. C.; Gorrod, J. W.; Mitchley, B. C. V. Br. J. Cancer 1964, 18, 265-70. (6) Hoffmann, D.; Hecht, S. S.; Ornaf, R. M.; Wynder, E. L. Science 1974, 186, 265-6. (7) Hecht, S. S.; Ornaf, R. M.; Hoffmann, D. Anal. Chem. 1975, 47, 2046-8. (8) Bharadwaj, V. P.; Takayama, S.; Yamada, T.; Tanimura, S. Gann 1975, 66, 585-6. (9) Munson, J. W.; Abdine, H. Anal. Lett. 1977, 10, 777-86. (10) Hecht, S. S.; Chen, C. B.; Hirota, N.; Ornaf, R. M.; Tso, T. C.; Hoffmann, D. J. Natl. Cancer Inst. 1978, 60 (4), 819-24. (11) Fine, D. H.; Rufeh, F.; Lieb, D.; Rounbehler, D. P. Anal. Chem. 1975, 47, 1188-91.

Analytical Chemistry, Vol. 75, No. 18, September 15, 2003 4827

Figure 1. Structures of tobacco-specific nitrosamines and their alkaloid precursors: NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone; NNA, 4-(methylnitrosamino)-4-(3-pyridyl)butanal; NNN, N′-nitrosonornicotine; NAB, N′-nitrosoanabasine; NAT, N′-nitrosoanatabine; NNAL, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol; iso-NNAL, 4-(methylnitrosamino)-4-(3-pyridyl)-1-butanol; iso-NNAC, 4-(methylnitrosamino)-4-(3pyridyl)butanoic acid.

with GC and TEA identified the fourth TSNA, NAB,13 and three additional TSNA compounds.14,15 Although TEA has been widely and effectively used, it is limited by its nonspecificity for individual nitroso compounds, large dead volumes within the TEA detector that limit the GC scan times, and relatively high detection limits. The use of LC/MS/MS instead of GC/TEA provides higher throughput with analyte-specific detection based on both retention time and structurally specific analyte fragmentation information. We report the development of a new two-step sample preparation and cleanup procedure followed by LC/MS/MS analysis. Five of the TSNAs extracted from tobacco smoke are analyzed with a high-throughput (8 min/ sample) and selective LC/MS/MS technique that provides excellent reproducibility and accuracy using isotope-labeled analogues as internal standards. This approach has several additional advantages in terms of reduced sample preparation and analytical run times over previous HPLC and GC/TEA methods that required fractionation of the tobacco extract and derivatization of selected TSNAs. EXPERIMENTAL SECTION Reagents. The TSNA analytical standards, NNN, NNK, NNAL, NAT, and NAB were purchased from Toronto Research Chemicals (Ontario, Canada). Isotopically labeled analogues, [2,2′,3,4,5,613C ]-NNN and [1,2′,3′,4′,5′,6′-13C ]-NNK, were obtained for use 6 6 as internal standards from Cambridge Isotope Laboratories (12) Hoffmann, D.; Adams, J. D.; Brunnemann, K. D.; Hecht, S. S. Cancer Res. 1979, 39, 2505-9. (13) Adams, J. D.; Brunnemann, K. D.; Hoffmann, D. J. Chromatogr. 1983, 256, 347-51. (14) Brunnemann, K. D.; Genoble, L.; Hoffmann, D. Carcinogenesis 1987, 8, 465-9. (15) Djordjevic, M. V.; Brunnemann, K. D.; Hoffmann; D. Carcinogenesis 1989, 10, 1725-31.

4828 Analytical Chemistry, Vol. 75, No. 18, September 15, 2003

(Andover, MA). Acetonitrile and methylene chloride were obtained from Burdick & Jackson (Muskegon, MI). All other chemicals were of analytical grade and obtained through Sigma (St. Louis, MO) unless otherwise indicated. Smoke Sample Analysis Procedure. Before smoking, each Cambridge filter pad (CFP) was conditioned at 22 °C and 60% humidity for at least 24 h. Before use, each pad was treated with 2 mL of methanol solution containing 40 mg of ascorbic acid and air-dried in a chemical fume hood. Individual pads were then placed inside the Cerulean filter holders (Milton Keynes) to collect mainstream smoke. Before smoking, cigarettes were conditioned at 22 °C and 60% humidity for at least 24 h. Cigarettes were smoked using a Cerulean (Milton Keynes) ASM 500 16-port smoking machine. Each cigarette was smoked at 1 puff/min, 2-s puff duration, and 35-mL puff volume. Standard and Sample Preparation Procedure. Standards and internal standard were dissolved in methanol at 2 mg/mL, diluted to intermediate stock solutions of 10, 1, and 0.1 µg/mL and stored at -20 °C. Intermediate stock solutions were used to prepare calibration standard solutions. All TSNA standards were stable at room temperature in the clear vial for at least 7 days as verified by LC/MS/MS analysis. Each pad with trapped smoke particulate was first spiked with 20 µL (1 µg/mL) of internal standard mixtures and then extracted with 14 mL of methylene chloride for 45 min using a Lab-line shaker (Melrose Park, IL) operated at 160 rpm. The extract was then dried using a Zymark Turbovap LV evaporator (Hopkinton, MA) and reconstituted with 500 µL of methylene chloride. An aliquot of 1 mL of 0.1 N HCl was added to each sample, and the combined solution was thoroughly vortexed. After allowing the phases to separate, the upper aqueous phase was transferred into a clean glass tube and combined with 120 µL of 1 N NaOH solution before being loaded onto the solid-phase extraction (SPE) column.

Table 1. Multiple Reaction Monitoring Analysis of Five Tobacco-Specific Nitrosamines and Its Internal Standards precursor ions

product ions

DPa

EPb

CEc

CXPd

NNAL

210.1

NNK

208.1

NNN

178.1

NAT

190.2

NAB

192.2

[13C6]-NNK [13C6]-NNN

214.2 184.1

180.1e 149.0f 178.1e 122.0f 148.1e 120.1f 160.1e 106.0f 162.2e 133.1f 128.0 154.1

45 45 50 50 33 33 43 43 41 41 50 45

9 9 8 8 8 8 10 10 8 8 8 8

15 20 12 17 15 26 14 24 15 31 18 15

11 9 10 10 8 10 9 9 9 12 7 9

analyte

a DP, declustering potential. b EP, entrance potential. c CE, collision energy. d CXP, collision cell exit potential. e Quantitation ion. f Confirmation ion.

SPE was performed with a Supelco Visiprep vacuum manifold (Bellefonte, PA). Waters Oasis HLB 60-mg SPE cartridges (Milford, MA) were prepared by washing with 2 mL of methanol followed by 2 mL of water. After the samples were loaded, the cartridges were washed with 2 mL of an aqueous solution containing 5% methanol. Analytes and their corresponding internal standards were eluted with 2 mL of 100% methanol. The eluant was dried down under Turbovap and reconstituted with 5% methanol in 5 mM ammonium acetate solution. The sample preparation time for 30 samples was ∼4 h. A 20-µL aliquot of this solution was injected onto the LC/MS for analysis. Instrumental Analysis. All samples were analyzed using an Agilent 1100 liquid chromatograph (Agilent Technologies, Wilmington, DE) coupled with an API 4000 triple quadruple mass spectrometer (Applied Biosystems, Foster City, CA). Separation of TSNA from the smoke extract was performed using reversedphase HPLC with a Waters Xterra C18 MS column (4.6 × 50 mm i.d. × 5 µm) at 60 °C and equilibrated with 95% solvent A (5 mM ammonium acetate) and 5% solvent B (5 mM ammonium acetate in acetonitrile). A flow rate of 1 mL/min with a linear gradient was used as follows: 5% solvent B for 1 min, 5-50% solvent B over the course of 1 min, and 50% solvent B for 3 min. The total run time was 8 min, including a 3-min equilibration time. An inline Upchurch Scientific solvent filter (Oak Harbor, WA) was used to extend the lifetime of the analytical column and improve the method’s robustness. Ionization of the TSNA analytes was achieved using the Turbospray ionization source operated in positive ion mode. The Turbospray settings were as follows: curtain gas (N2) at 12 psig, ion source gases 1 and 2 both at 40 psig, source temperature at 600 °C, ion spray voltage at 5000 V, and collision gas pressure at 5 psig. Mass spectral data on precursor and product ions were collected in multiple reaction monitoring (MRM) mode. The declustering potential, entrance potential, collision energy, and cell exit potential were optimized for each analyte (Table 1). The labeled NNK was used as internal standard for the NNAL and NNK quantification, and the labeled NNN was used as internal standard for NNN, NAB, and NAT quantification.

Data Analysis. Peak area determinations for all samples, blanks, standards, and quality control (QC) materials were processed using the Analyst software version 1.2 of the API 4000. Each ion of interest in the chromatogram was automatically selected and integrated. The peak integrations were manually inspected for errors and re-integrated if necessary. Eight standards were prepared in the range of 2.5-2500 ng/mL with internal standards added at a concentration of 200 ng/mL. Least-squares linear regression of the response factors (RF, peak area ratio of native to internal standard) versus standard analyte amount provided calibration constants. Quantitative calibrations were accomplished using the standard Analyst software, and the resultant data were transferred to an Access database (Microsoft, Bellevue, WA). The integrated peak areas for each analyte and retention times were saved in a report file in Analyst software. A custom-written Access macro converted the Analyst report file into a text report file, which was imported into Access database specifically created for this analysis. Statistical analysis including QC characterization was carried out in a custom program developed using the Statistical Analysis System (SAS) software (SAS Institute, Cary, NC). RESULTS AND DISCUSSION Method Development. Caldwell and Conner have suggested that artifact formation via nicotine nitrosation could occur during smoke collection16 and skew TSNA quantification. To examine this possibility, we measured the effect of ascorbic acid treatment of the CFP. Mainstream smoke particulate matter, from individual cigarettes, was collected on five untreated CFPs and five CFPs treated with ascorbic acid. The smoke particulate trapped on treated CFPs was 25-35% lower in TSNAs (p < 0.05 by means of a two-tailed, two-sample unequal variance t-test) than the untreated CFPs. In addition, the variation of TSNA level, which may reflect different degrees of nitrosation between the replicate cigarettes, is also higher without ascorbic acid treatment. These results suggest that some artifact formation via nitrosation occurs on the untreated CFPs and ascorbic acid treatment is warranted. Varying the amount of ascorbic acid in the range of 20-80 mg on the CFP produced similar results, so we adopted 40 mg of ascorbic acid for regular use. To optimize the extraction of TSNAs from the CFPs, a series of 1R4F cigarettes were smoked and the CFPs were extracted with methylene chloride at 0.75, 1.5, 3, and 6 h, respectively. The extraction efficiency of TSNAs at each time point was calculated as the ratio of the TSNA response factor obtained from spiking the smoke sample with the isotopically labeled standard after and before the methylene chloride extraction process, respectively. All the samples went through the same preparation procedure, and TSNAs levels were determined using LC/MS/MS. The timedependent study showed that increased extraction time provided improved recovery for NAB and NAT, consistent recovery for NNN and NNK, and a slightly decreased recovery for NNAL. Typically, NNAL has lower abundance than the other TSNAs in smoke. Therefore, 0.75 h was selected as an optimum extraction time for NNAL while providing adequate signal strength for the remaining TSNAs. Because the tobacco smoke matrix contains (16) Caldwell, W. S.; Conner, J. W. J. Assoc. Off. Anal. Chem. 1990, 73, 783-9.

Analytical Chemistry, Vol. 75, No. 18, September 15, 2003

4829

Figure 2. Typical multiple reaction monitoring chromatograms of TSNAs in the smoke extract of 1R4F: (A) NNAL (retention time, RT ) 3.02 min); (B) NNN (RT ) 3.12 min); (C) [C13]-NNN (RT ) 3.12 min); (D) NNK (RT ) 3.22 min); (E) [C13]-NNK (RT ) 3.22 min); (F) NAT (RT ) 3.38 min); (G) NAB (RT ) 3.43 min)

about 4800 compounds,17 we performed a one-step liquid-liquid extraction of the smoke particulate to reduce organic interferences. To further clean up smoke extracts, solid-phase extraction was performed following liquid-liquid extraction. The Oasis HLB, 30 and 60 mg, Oasis MCX 30-mg SPE cartridges were evaluated to obtain optimal recovery. We found that 60-mg Oasis HLB had the best overall recovery for all the TSNAs. Hecht et al.18 have reported that some TSNA standards contain E and Z isomers resulting in the shoulder effect and peak splitting on HPLC separation. We observed similar results under our HPLC separation conditions. We found all the TSNAs can exhibit some degree of shoulder effects or split peaks. However, this behavior was pH dependent over the range of 1-12. When we dissolved standards into different pH buffer solutions for LC/MS/MS analysis, we obtained different split peak ratios, suggesting that one isomer could be favored over another under certain pH condition. With careful column selection and optimization, we achieved good peak shapes without splitting at a pH of 5. To achieve maximum sensitivity and optimize peak shape, we evaluated different types of HPLC columns for the separation of TSNAs, including Waters Symmetry C18, Xterra C18, Xterra C18 MS, Xterra phenyl, and Thermo Hypersil-Keystone Betamax Base columns. After careful evaluation of these columns under selected mobile-phase conditions, we successfully optimized the TSNA peaks shapes while maintaining good separation using a Waters (17) Green, C. R.; Rodgman, A. Recent Adv. Tob. Sci. 1996, 22, 131-304 (18) Hecht, S. S.; Spratt, T. E.; Trushin, N. Carcinogenesis 1997, 18 (9), 18514.

4830 Analytical Chemistry, Vol. 75, No. 18, September 15, 2003

Symmetry C18 column (4.6 × 50 mm 3.5 µm) or Xterra C18 MS column (4.6 × 50 mm 5 µm) operated at 60 °C. We achieved optimal performance using gradient separation with 5 mM ammonium acetate as solvent A and acetonitrile containing 5 mM ammonium acetate as solvent B. Although the Symmetry column achieved slightly higher chromatographic resolution, the Xterra column was selected for robustness since the LC separation was done at elevated temperature (60 °C). Figure 2 shows the MRM traces for the internal standards and TSNAs obtained from the mainstream smoke particulate smoke of a 1R4F cigarette. Method Validation. The accuracy of the method was established by spiking known amounts of the TSNAs on CFP containing smoke samples from 1R4F cigarettes. Analytes were added at two concentrations: one at half the amount of in 1R4F cigarette smoke and another at double the amount in 1R4F smoke. The accuracy is calculated as the mean of the experimentally determined concentration from replicate analysis divided by the nominal concentration. The precision of the method was determined by calculating the relative standard deviations (RSD) of the replicate measurements. All RSD were less than 15% for both intra- and inter-assay precision (Table 2). Accuracy for all analytes was excellent except for NAT (Table 2). Labeled NNN does not appear to be completely effective in characterizing the effect of chemical interference on NAT in the smoke-imbedded CFP. Fortunately, the accuracy appears reasonably consistent, allowing for a bias correction to these measurements. We expect the accuracy of NAT would improve if an isotopically labeled analogue were available for use as internal standard.

Figure 3. Typical calibration curve for NNN (A) and NNK (B) based on the analysis of eight standard solutions across the linear range. The linear regression curve through all the standard points produced excellent linear correlation coefficients including the lowest levels as shown in the figure inset.

Table 2. Precision and Accuracy of the Measurement of Tobacco-Specific Nitrosamines in Smoke analyte NNAL NNN NNK NAT NAB

NNAL NNN NNK NAT NAB

nominal concn (ng)

mean calcd concn (ng)a

accuracy (%)b

precision (%)c

2.5 5.0 50 100 25 50 50 100 5.0 10

Intraassay (n ) 5) 2.23 5.05 47.27 100.87 28.70 57.40 30.55 73.35 4.34 10.06

89.0 101.0 94.6 100.9 105.0 114.8 61.1 73 86.9 100.7

14.3 5.67 9.31 2.88 3.72 3.89 4.33 4.85 3.49 4.33

5.0 100 50 100 10

Interassay (n ) 10) 4.89 102.2 55.3 75.3 10.8

97.7 102.2 110.5 75.3 108.2

12.5 5.70 6.66 3.93 4.04

a From the line regression standard curve. b Calculated as (mean calculated concentration)/(nominal concentration) × 100%. c Expressed as % RSD.

The overall method recovery of TSNA from cigarette smoke was calculated as RFa/RFb, where RFa and RFb are the TSNA response factor obtained from spiking the smoke sample with the isotopicallylabeled standard after and before the extraction process, respectively. The relative recovery for NNAL, NNN, NNK, NAT, and NAB is calculated as 50, 62, 58, 61, and 66%, respectively. The absolute internal standard recoveries, based on mean detector response for [C13]-NNN and [C13]-NNK of the extracted samples versus the pure standards, were 20 and 18% for NNN and NNK, respectively. The lower internal standard recovery, compared to the method recovery, suggests the signal suppression of ion current by other constituents in the smoke matrix may occur, although the MRM background for each ion trace appeared relatively free of interferences. These data confirm the necessity

of labeled internal standards in order to achieve the accuracy reported in Table 2. The calibration curves ranged from 2.5 to 2500 ng/mL. Excellent linearity was obtained with typical values for the correlation coefficient (R2) between 0.99 and 1.00. Typical calibration curves for NNN and NNK are shown in Figure 3. The calibration range could be extended to higher and lower amounts, but this was not routinely done since this working calibration range is sufficiently wide to measure smoke TSNAs from current cigarette samples. Standard deviations calculated from multiple measurements of standard concentrations were used to estimate the limit of detection (LOD), set equal to 3S0,19 where S0 is the standard deviation of the analyte concentration at zero concentration. The estimated LODs for NNAL, NNN, NNK, NAT, and NAB were 1.23, 0.18, 0.64, 0.05, and 0.25 ng/mL, respectively. The LOD in the matrix could be lowered using a smaller diameter HPLC column; however, the smaller diameter column degraded the chromatographic peak shape. Our time-dependent extraction experiments suggest that there could be a binding interaction of the TSNAs and the glass fiber CFP. Disrupting the nonspecific binding of TSNAs on the CFP or using CFP made of different material could increase recovery efficiency and result in improved detection limits. Extracts prepared from a blank CFP were routinely analyzed for any coeluting interferences or sample carry-over from previous TSNA-containing samples. No false positive responses were observed for any of the TSNA analytes in the sample blanks. Measurements also were made to determine how much native analyte contributed to the isotope-labeled internal standard and vice versa.20 No significant cross interferences were observed. The levels of NNN, NNK, NAT, and NAB in the smoke from a 1R4F research cigarette have been previously reported (Table 3).16,21 The NNN levels measured in 1R4F cigarette smoke from (19) Taylor, J. K. Quality Assurance of Chemical Measurement; Lewis Publishers: Chelsea: MI, 1987, 79-82. (20) Colby, B. N.; McCaman M. W. Biomed. Mass Spectrom. 1979, 6, 225-30.

Analytical Chemistry, Vol. 75, No. 18, September 15, 2003

4831

Table 3. Comparison of Levels of Tobacco-Specific Nitrosamine in the Mainstream Smoke of the 1R4F Cigarette concn (µg/cigarette) method

NNNa

NNKb

NATc

NABd

GC/TEAe GC/TEAf LC/MS/MS

0.090 0.085 0.091

0.081 0.030 0.054

0.111 0.158 0.096g

0.012 0.010 0.012

a NNN, N′-nitrosonornicotine. b NNK, 4-(methylnitrosamino)-1-(3pyridyl)-1-butanone. c NAT, N′-nitrosoanatabine. d NAB, N′-nitrosoanabasine. e Caldwell, W. S.; Conner, J. W. J. Assoc. Off. Anal. Chem. 1990, 73, 783-789. f Brunnemann, K. D.; Hoffmann, D. Crit. Rev. Toxicol. 1991, 21, 235-240. g Value corrected based on the 75% accuracy recovery of NAT.

these different methods are consistent. The variation of NNK level among those methods could be attributed to one or more factors. For example, the 1R4F cigarettes were manufactured in 1983 and may have undergone various aging processes. Another source of variability is found in the filter ventilation, which ranged from 19 to 31% in the 1R4F cigarettes tested here. Our method uses isotopically labeled analogues of NNK and NNN as analytical references to provide similar response in terms of sample preparation loss, ionization efficiency, and suppression in the matrix. We expect that a method based on an isotopically labeled standard provides more rugged, reproducible, and accurate results than other methods that lack an internal standard or use a surrogate compound as an internal standard. CONCLUSIONS We have developed and validated a HPLC/MS/MS method for simultaneous determination of five TSNAs from cigarette smoke. The resolving power of tandem mass spectrometry has the inherent advantage over the current, widely used TEA method for TSNA quantitation in terms of the ease of sample preparation, selectivity, and sensitivity. In addition to unambiguous identification of five TSNAs, we achieved excellent reproducibility and accuracy with the use of isotope-labeled analogues as internal standard. TSNAs are among the 33 carcinogens in tobacco products including 6 known human carcinogens and 69 carcinogens in cigarette smoke including 11 known human carcinogens.22 The overall pathway leading to the initiation of cancer-carcinogen (21) Brunnemann, K. D.; Hoffmann, D. Crit. Rev. Toxicol. 1991, 21, 235-40. (22) Hoffmann, D.; Hoffmann, I.; El-Bayoumy, K. Chem. Res. Toxicol. 2001, 14 (7), 767-90. (23) Vector Tobacco Inc. Newsletter, 2001. (24) Fisher, B. Tob. Rep. 2001, 128 (12), 22-25. (25) Youth Tobacco SurveillancesUnited States; 2000. Surveillance Summaries 2001, 50 (SS04), 1-84. (26) Centers for Disease Control and Prevention (CDC). Morbidity Mortality Weekly Rep. 1999, 48, 796-9.

4832 Analytical Chemistry, Vol. 75, No. 18, September 15, 2003

exposure and uptake, metabolic activation, and DNA adduct formation has been demonstrated for TSNAs in humans. These data, together with the overwhelming epidemiological evidence that tobacco products cause cancer in humans, strongly suggest the relevance of these compounds in human cancer etiology.3 Effort has recently been focused by manufacturers on reducing the TSNA levels in tobacco products and their smoke such as by the modification of cigarette components and the tobacco curing process. Recently, two new cigarette brands with claims of reduced TSNA levels have been introduced. One, Omni, manufactured by Vector Tobacco Inc., claims a 24% reduction of NNN and 53% reduction of NNK compared with the traditional cigarettes when smoked using the Federal Trade Commission’s (FTC) machine smoking protocol.23 Advance Light, manufactured by Brown & Williamson Tobacco Corp., is another recently introduced cigarette brand. Advance Lights combines two new technologies, a special three-part Trionic filter and patented tobacco-curing process that claims to deliver reduced levels of TSNAs and other toxins in smoke.24 The TSNA levels in these and future modified tobacco products require stringent examination to assess these and other product claims. To facilitate such studies it is necessary to develop and maintain sensitive and reliable methods to measure toxic and carcinogenic agents in tobacco smoke. In addition to these new technology cigarette products, other alternative cigarette products such as bidi cigarettes and clove cigarettes are being used by adolescents.25 One of the major reasons that teenagers may prefer to smoke bidi cigarettes is because they believe that bidis are “more natural” and therefore less harmful than traditional cigarettes.26 Analyses of TSNA levels in the mainstream smoke from bidi, clove, new technology, and traditional cigarettes are currently under investigation in our laboratories. Detailed examination of these products should help provide insights into the relative carcinogen yield associated with their usage. Such studies are timely and needed to address how use of such products impacts public health in the United States. However, the only effective means to reduce the risk associated with smoking and other forms of tobacco use is to quit. ACKNOWLEDGMENT We acknowledge financial support for this study from the National Center for Chronic Disease Prevention and Health Promotion and the Office of Smoking and Health of the Centers for Disease Control and Prevention (CDC). The use of trade names and commercial sources is for identification only and does not imply endorsement by the CDC or the U.S. Department of Health and Human Services.

Received for review April 2, 2003. Accepted July 2, 2003. AC030135Y