Analysis of 4-aminobiphenyl-DNA adducts in human urinary bladder and lung by alkaline hydrolysis and negative ion gas chromatography-mass spectrometry.

Analysis of carcinogen-DNA adducts has been regarded as a useful means of assessing human exposure to chemical carcinogens. We have established a method for quantitation of 4-aminobiphenyl (4-ABP)-DNA adducts by alkaline hydrolysis and gas chromatography with negative ion chemical ionization mass spectrometry (GC-NICI-MS). Aliquots of DNA (typically 100 micrograms/ml) were spiked with an internal standard, d9-4-ABP, and were hydrolyzed in 0.05 N NaOH at 130 degrees C overnight. The liberated 4-ABP was extracted with hexane and derivatized using pentafluoropropionic anhydride in trimethylamine for 30 min at room temperature prior to GC-NICI-MS. With in vitro [3H]N-hydroxy-4-ABP modified DNA standards, we observed 59 +/- 7% (n = 9) recovery of the 4-ABP and a linear correlation between hydrolyzed 4-ABP and the adduct levels ranging from about 1 in 10(8) to 1 in 10(4) nucleotides (r = 0.999, n = 9). The method was further validated by comparison of the results with that obtained by the 32P-postlabeling method. There was excellent agreement (r = 0.994, p < 0.001) between the two methods for quantitation of the adduct in eight samples of Salmonella typhimurium DNA treated with 4-ABP and rat liver S9, although the 32P-postlabeling method gave slightly higher values. The DNA adducts in 11 human lung and 8 urinary bladder mucosa specimens were then determined by our GC-NICI-MS method. The adduct levels were found to be < 0.32 to 49.5 adducts per 10(8) nucleotides in the lungs and < 0.32 to 3.94 adducts per 10(8) nucleotides in the bladder samples.(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
The covalent modification of DNA by chemical carcinogens appears to be a critical event in chemical carcinogenesis. As a result, the analysis of carcinogen-DNA adducts has been considered to be a useful means of assessing human exposure to chemical carcinogens. 4-Aminobiphenyl (4-ABP) is a potent human urinary bladder carcinogen and is present in significant quantities in tobacco smoke (1). This aromatic amine is metabolically activated to several electrophilic intermediates, which then covalently bind to cell macromolecules. The reaction of activated 4-ABP with DNA in vitro and in vivo in laboratory ani-This paper was presented at the Fifth International Conference on Carcinogenic and Mutagenic N-Substituted Aryl Compounds held 18-21 October 1992 in Wurzbrug, Germany.
Address correspondence to Fred F. Kadlubar, Office of Research (HFT-100), National Center for Toxicological Research, Jefferson, AR 72079. mals is known to result in formation of three to four nucleoside adducts. The predominant adduct is N-(deoxyguanosin-8-yl)-ABP (dG-C8-ABP), which accounts for about 70% of the bound radioactivity (2,3). Evidence for human exposure to 4-ABP and its metabolic activation was initially obtained from the detection of 4-ABP-hemoglobin adducts in humans, with higher levels being observed in cigarette smokers versus nonsmokers (4)(5)(6) and in patients with a history of urinary bladder cancer versus controls (7). With recent advances in analytical methodology, 4-ABP-DNA adducts have been detected either by immunochemical or 32P-postlabeling-HPLC techniques in human lung or urinary bladder at levels ranging from one adduct per 106 to one adduct per I09 nucleotides (8,9). More recendy, 4-ABP-DNA adducts were found as smoking related adducts in human urinary bladder biopsies using 32P-postlabeling methods (10). These findings suggested that 4-ABP is likely to be an important factor in the etiology of tobacco-related human cancers. However, since these methods may not always provide sufficient specificity and precision, a sensitive, structure-specific, and relatively simple method was warranted for molecular dosimetry of human exposure to this carcinogen. Furthermore, a specific methodology for confirmation of sample analyses was also highly desirable.

Preparation ofDNA Samples
[3H]N-Hydroxy-4-ABP modified DNA standard was prepared by the reaction of radiolabeled N-hydroxy-4-ABP with calf thymus DNA in 100 mM potassium citrate/ 0.1 mM EDTA, pH 5, as described previously (2,3). The adduct levels were determined by enzymatic hydrolysis, HPLC, and scintillation counting. The Salmonella typhimurium TAI 538 and TA98 DNA was isolated from the bacteria treated with various concentrations of 4-ABP ranging from 5 to 24 pg/ml in the presence of Aroclortreated rat liver S9 under the conditions described (13). The bacteria were digested for 10 min with 0.1% SDS, 50 mM Tris-HCl buffer (pH 7.9), 10 mM EDTA, and 0.28 mg/ml proteinase K; DNA was purified from the lysates as described (14). Human lung and urinary bladder tissues obtained by surgery or autopsy were from the John A. McClellan Memorial Veterans Hospital in Little Rock, AR. The DNA was isolated by homogenization of the tissue, lysis of the cells with SDS, digestion with protease, and extraction with phenol and chloroform/isoamyl alcohol; traces of RNA were removed by digestion with RNase and DNA was precipitated with ethanol (15). The smoking histories of the lung donors are shown in Table 1; however, the smoking histories of the urinary bladder donors were not available.
Alkaline Hydrlysis of4-ABP from DNA Aliquots of DNA solution (typically 100 pg DNA/ml) in screw-capped tubes (13 x100 mm) were brought to 0.05 N with respect to NaOH and extracted three times with 2 ml of hexane before spiking with internal standard, d9-4-ABP, (2 to 200 pg in 10 pl methanol, for GC-NICI-MS analysis) or 4-ABP (as UV marker for HPLC analysis). The solution was then heated at 130°C overnight (18 hr) in a Temp-Blok module heater (Lab-Line Instruments Inc., Melrose, IL) that had been filled with sea sand. The samples were then cooled to room temperature and the released 4-ABP was extracted into hexane (2 ml, twice). For HPLC analysis, [3H]N-hydroxy-ABP-modified DNA was hydrolyzed. After extraction, hexane was removed and the residue redissolved in methanol. The 4-ABP was separated by HPLC on a Waters pBondapak C,8 column (3.9 x 300 mm). Elution was carried out with a linear gradient starting from 20% methanol in water to 100% methanol over 25 min at a flow rate of 2 ml/min. Under these conditions, the retention time of 4-ABP was 15.3 min. The eluate was collected at 1-min intervals and mixed with scintillation fluid for counting radioactivity in a Tracor Model 688 Analytic Marker III scintillation counter (Elk Grove, IL). For GC-NICI-MS analysis, the extracted 4-ABP and the internal standard were derivatized as described below.
Derivatization of4-ABP 4-ABP was derivatized on the basis of the procedure described previously (4). Briefly, 10 pl of trimethylamine in hexane was added to the combined hexane extracts obtained as described above in a test tube, followed by 10 p1 of PFPA. After reaction for 30 min at room temperature, the solvent and excess reagents were removed by vacuum Retention time (min)   Figure 1A.
The yield of 4-ABP from the 4-ABPadducted DNA was found to be 59 ± 7% m the parallel (n=9) and this result was consistent with -ABP-modified that obtained by HPLC analysis. An exceladduct levels lent correlation (r = 0.999, n = 9) was 05 nucleotides, observed between hydrolyzed 4-ABP r with a corre-amounts and the adduct levels in the DNA (Figure 2). The detection limit is 10 fmoles 4-ABP per mg DNA (0.32 adduct per 108 nucleotides) due to the amount of about 0.006 pg of 4-ABP present in the blank. Reproducibility was assessed by sampling e performed to in triplicate for nine different samples, and BP hydrolyzed the coefficient of variation was estimated to kBP-adducted be 9.7%. hydrolysis con-To validate this method for DNA with kBP-modified unknown levels of the adduct, eight DNA I]dG-C8-ABP samples, isolated from S. typhimurium .ing a variety of strains TA1538 and TA98 incubated with   Application ofthe Method to Hwnan Thsue DNA Eleven human lung and eight human urinary bladder mucosa DNA samples were analyzed for dG-C8-ABP adducts. The results, summarized in Table 1 and Figure 5, indicate that 10 of 11 lung DNA samples and 5 of 8 urinary bladder mucosa DNA samples were positive. The GC-NICI-MS traces of one of these positive DNA samples are shown in Figure 1 B and 1 C, respectively. The adduct levels in lung DNA varied among the individuals, ranging from <0.32 to 49.5 adducts in 108 nucleotides. However, the adduct levels did not seem to be directly related to the numbers of cigarettes smoked per day or the duration of smoking ( Table 1). The adduct levels in the eight urinary bladder mucosa DNA ranged from <0.32 to 3.9 adducts in 108 nucleotides ( Figure 5), indicating that the adduct levels are comparable in both the lungs and the urinary bladder.

Discussion
Based on the previous work on the hydrolysis of carcinogenic arylamine-nucleoside adducts (11,12), we have investigated the alkaline hydrolysis of 4-ABP-modified DNA. We found that 59% of 4-ABP bound to DNA could be liberated by hydrolysis with 0.05 N NaOH at 130°C overnight. The efficiency of the hydrolysis of the adducts and the minimal interference in the samples, as seen in Figure 1, has enabled application of the GC-NICI-MS method to quantitate trace amounts of 4-ABP hydrolyzed by the procedure. With [3H]N-hydroxy-ABP-modified DNA standards, the method is sensitive to levels as low as 10 fmole 4-ABP per mg DNA (0.32 adduct/103 nucleotides using 100 j'g DNA). The method has been used to quantify adduct levels up to one adduct per 104 nucleotides, and the curve is linear in this range (r = 0.999, Figure 2). These results indicate that this method provides both high selectivity and sensitivity required for detection and quantification of the adduct at levels found in human biological samples (10,18). We also compared this GC-NICI-MS method to the most widely used 32P-postlabeling technique in analysis of dG-C8-ABP in the same DNA samples. The results show that these two methods are highly correlated, although the 32P-postlabeling method gave slightly higher absolute values (Figure 3). This suggests that the GC-NICI-MS method is reliable and comparable to the 32P-postlabeling method for the determination of 4-ABP-DNA adducts at the levels analyzed. Tobacco smoking is implicated as an etiologic factor in human lung and urinary bladder cancers (19)(20)(21)(22). 4-ABP, among other aromatic carcinogens in tobacco smoke, is likely to play a role in these dis- eases because 4-ABP-DNA adducts were detected in smokers' lungs and urinary bladder (8)(9)(10). Using the GC-NICI-MS method, we have confirmed these previous findings. The adduct was found in almost all lungs and five of eight of the urinary bladders and the levels in both lungs and urinary bladder were comparable. Lack of correlation of the adduct levels in smokers' lungs to the number of cigarettes smoked may reflect limited sample size as well as the differences in actual intake of 4-ABP, which can be affected by smoking habits, such as brand of cigarette used (1,23) and the extent of smoke inhalation. Moreover, the biologic uptake and distribution of the carcinogens, metabolic phenotype, and DNA repair mechanisms are well known to vary greatly from individual to individual (23,24), and these factors undoubtedly  (25,26). These lung tissue specimens were not sampled from the same anatomic location. Therefore, the result does not necessarily mean that the adducts are not relevant to cigarette smoke. 4-ABP-DNA adducts also can be quantitated by the 32P-postlabeling method and immunochemical assays (9,10). While immunochemical assays may provide sensitivity required for human biomonitoring, the accuracy of the method can be compromised by cross-reactivity of the antibodies with unknown contaminants. The 32P-postdabeling technique is a widely used method and has been applied to analysis of 4-ABP-DNA adducts in human DNA in several studies carried out in this laboratory and others (10,18). This method is highly sensitive and requires only low microgram amounts of DNA for analysis. However, the quantitation is compromised when the authentic adduct is not available for use as a chromatographic standard, and accurate measurements are difficult when it is applied to human tissue DNA, especially smokers' lung DNA, because the 4-ABPadduct spots are usually located in the diagonal zone of radioactivity. Attempts were made in the current study to correlate the 32P-postlabeling method to the GC-NICI-MS method for quantitation of 4-ABP adducts in human urinary bladder DNA, but it was not successful due to the difficulty in precisely measuring 32P radioactivity in half the cases in which the adduct spots were located in the diagonal zone (data not shown). Under our experimental conditions, however, the GC-NICI-MS method obviated this problem for the adduct analysis and offered not only relatively high selectivity and sensitivity but also accuracy.
In conclusion, we have shown that 4-ABP-DNA adducts can be detected reliably by alkaline hydrolysis and GC-NICI-MS analysis. This method is sensitive, structure-selective, and appears to be useful for molecular dosimetry of human exposure to the carcinogen 4-ABP.