Several studies have reported the presence of DNA adducts derived from benzo(a)pyrene and other polyaromatics by 32P-postlabeling/TLC by measuring diagonal radioactive zones (DRZs) in lung tissues of human smokers. However, our experimental studies in rodent models, which used modified chromatographic conditions to obtain distinct adduct spots, suggested that cigarette smoke-related lipophilic DNA adducts may not be derived from polycyclic aromatic hydrocarbons (PAHs) or aromatic amines. In the present study, we have performed similar analysis of human lung tissues to study the chemical nature of DNA adducts. Fifty human lung tissues from cancer patients (ages 42-83 years) with active, ex-, or never-smoking status were analyzed for highly lipophilic DNA adducts by nuclease P1- and n-butanol enrichment-mediated 32P-postlabeling assay. All DNA samples yielded low to highly intense adduct DRZs when adducts were resolved by PEI-cellulose TLC in standard high-salt, high-urea solvents. Adduct burden ranged from 6.6 to 2930 per 10(10) nucleotides. However, when adducts were resolved in a different solvent system comprising of high-salt, high-urea in direction 3 and dilute ammonium hydroxide in direction 4, which retained adducts derived from PAHs and aromatic amines on the chromatograms, this yielded no detectable adducts from human lung DNAs. Furthermore, analysis of human lung DNAs mixed with reference adducted DNAs in multisolvent systems confirmed an absence of PAH- and aromatic amine-derived adducts in human smoker lung DNA. To determine the origin of cigarette smoke-associated DNA adducts, calf thymus DNA was incubated with formaldehyde and acetaldehyde, which are known to be present in cigarette smoke in significant quantities. Analysis of purified DNAs by 32P-postlabeling resulted in adduct DRZs in the aldehyde-modified DNAs when adducts were resolved in standard urea-containing solvents, but no adducts were detected when the ammonium hydroxide-based solvent was used, suggesting that even nonpolyaromatic electrophiles can result in adduct DRZs on the chromatograms similar to those from PAH metabolites. Taken together, our data demonstrate that cigarette smoke-associated lung DNA adducts appear on chromatograms as DRZs, consistent with the literature, but they are not related to PAHs and aromatic amines.

To use oxidative DNA lesions as biomarkers of exposure and disease in the human scenario, the assay should not only be sensitive but also applicable to small quantities of DNA that is obtained from human tissue biopsies. Previously, we reported a nonenrichment (32)P-postlabeling method for measuring the oxidative DNA lesion, 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxo-dG) [Devanaboyina, U., and Gupta, R. C. (1996) Carcinogenesis 17, 917-924]. Here we report a substantial improvement of the assay by enriching 8-oxo-dGp by TLC prior to (32)P-labeling. Thus, unmodified nucleotides were removed by 1-directional polyethyleneimine (PEI)-cellulose TLC in 0.2 M formic acid. 8-Oxo-dGp, retained close to the origin, was eluted in 2 M triethylammonium acetate (pH 7.0), lyophilized, (32)P-labeled, resolved by TLC, and quantitated. The 8-oxo-dG signal was found to increase linearly with increasing amount of DNA. Its recovery was found to be 55-70%, as determined by using synthetic 8-oxo-dGp. Since the modified assay allowed using larger quantities of DNA and higher specific activity [gamma-(32)P]ATP than those used in the nonenrichment procedure, the signal-to-noise ratio increased to 30-100:1 from 3-5:1. The assay has a sensitivity of detection of <1 8-oxo-dG/10(7) nucleotides using submicrogram to microgram DNA. The 8-oxo-dG levels (mean +/- SE; n = 5) in the liver, lung, heart, bladder, and trachea DNA of 3 month-old female Sprague-Dawley rats were found to be 1.05 +/- 0.24, 0.97 +/- 0.09, 0.75 +/- 0.15, 0.79 +/- 0.15, and 1.17 +/- 0.28 per 10(6) nucleotides, respectively.