Exposure to Heavy Metals and Polycyclic Aromatic Hydrocarbons and DNA Damage in Taiwanese Traffic Conductors

In urban environments across the globe, air pollution from traffic exhaust has been considered an important public health issue. Although there is accumulating epidemiologic evidence that exposure to traffic-related air pollutants plays a role in the etiology and prognosis of many illnesses, including cardiovascular disease and cancer, the role of specific causal agents and the underlying mechanisms for different health outcomes remain unknown (1). Among these pollutants are polycyclic aromatic hydrocarbons (PAH) and heavy metals adsorbed in the pores and surfaces of the particles could easily enter into the respiratory tract and reach the circulatory system (2).

Previous studies have reported that exposure to PAHs or metals could generate reactive oxidative species (ROS) to induce oxidative stress in vitro or in vivo (3, 4). Oxidative stress could play an important role in the initiation and promotion of carcinogenesis. However, few studies have explored the effect of co-exposure to PAHs and metals on oxidative stress in humans (5, 6).

Applying biomarkers of exposure and outcome could be useful both in the quantification of the association and in providing insight into the specific causal agents and underlying biological mechanism. A major product with a clear mutagenic potential, deoxyguanosine (dG) in urine, has been commonly used as a biomarker of oxidative stress in studies on ambient air pollution (7–10). Conversely, urinary 1-hydroxypyrene-glucuronide (1-OHPG), a PAH metabolite, has been used as a biomarker of exposure to PAHs (9, 11, 12). Urinary metals, including cadmium (Cd), arsenic (As), nickel (Ni), lead (Pb), and copper (Cu), indicate short-term or long-term levels of internal dose of exposure to metals (13). Most studies showed that exposure to either PAHs or metals could increase levels of urinary 8-oxodG in occupational or environmental fields (7–9, 14). Only a few studies have investigated the effect of their interaction on biomarkers of oxidative stress, but a joint effect between PAHs and metals was not found (5, 6).

Therefore, we carried out a study of traffic conductors to examine the relation between urinary biomarkers of PAHs and heavy metals and the effect of their interaction on DNA damage.

We recruited 91 traffic conductors as the exposed group and 53 indoor office workers as the reference group, between May 2009 and June 2011, in Taipei, Taiwan. All of the selected subjects were between 20 and 63 years of age. All participants were free of cancer and pulmonary disease, and they had all been working in their current position for at least 3 months.

The participants underwent health examinations and completed a self-administered questionnaire on demographic information, lifestyle habits, and history of previous and current diseases. Postshift urine was collected from participants on 2 consecutive days. This study was approved by the Institutional Review Board of the National Health Research Institutes in Taiwan, and written informed consent from participants was obtained before study enrollment.

The urine samples were collected within half an hour of the end of a work shift. We divided the urine samples into several small-volume aliquots and stored them at −80°C in a freezer until analysis.

Urinary 1-OHPG was measured using the assay developed by Strickland and colleagues in 1994 (15). Urine samples (2 mL) were treated with 0.1 N HCl (90°C) to hydrolyze acid-labile metabolites. The hydrolyzed samples were loaded onto Sep-Pak C18 cartridges (Waters) and washed with methanol (30% in water). The relatively nonpolar metabolites were eluted with methanol (80% in water; 4 mL) and the volume was reduced to 0.5 mL by a centrifugal and vacuum evaporator (Eyela CVE-100). The concentrated samples were diluted to 4 mL with 15 mmol/L PBS. Immunoaffinity columns were prepared using polyprep columns (0.8 × 4 cm) filled with CNBr-activated Sepharose 4B (0.8 mL) coupled with monoclonal antibody 8E11, which recognizes several PAH-DNA adducts and metabolites. Monoclonal antibody 8E11 was obtained from Trevigen. It was originally produced against benzo[a]pyrene-diolepoxide-modifed DNA, and has been shown to recognize 1-OHPG. After washing the columns twice with 4 mL of 15 mmol/L PBS, samples in PBS were loaded on columns and bound material was eluted with 2 mL of 40% methanol. Eluted fractions were quantified by synchronous fluorescence spectroscopy (SFS) using a Perkin-Elmer LS 50B luminescence spectrometer. The excitation–emission monochromators were driven synchronously with a wavelength difference of 34 nm. 1-OHPG, purchased from the National Cancer Institute (NCI) Chemical Carcinogen Repository (MRI), produces a characteristic fluorescence emission maximum at 381 nm (347 nm excitation). Fluorescence intensity was used to quantify 1-OHPG. The recovery of the assay was 91%. The coefficient of variation of the assay was less than 5% during sample analysis. The limit of detection was 0.05 ng/mL.

Urinary 8-oxodG concentrations were measured using liquid chromatography/tandem mass spectrometry (LC/MS-MS), as described previously (16). Briefly, 20 μL of urine was diluted 10-fold with 5% methanol containing 0.1% formic acid. After the addition of 40 μL of ¹⁵N₅-8-oxodG solution (20 μg/L in 5% methanol per 0.1% formic acid) as internal standard, 100 μL of prepared urine sample was directly injected into the on-line solid-phase extraction LC/MS-MS. After automatic sample cleanup, LC/MS-MS analysis was done using a PE Series 200 HPLC system interfaced with a PE Sciex API 3000 triple quadrupole mass spectrometer with electrospray ion source. The samples were analyzed in the positive ion multiple-reaction-monitoring mode, and the transitions of the precursors to the product ions were as follows: m/z 284 → 168 (quantifier ion) and 284 → 140 (qualifier ion) for 8-oxodG and m/z 289 → 173 (quantifier ion) and 289 → 145 (qualifier ion) for ¹⁵N₅-8-oxodG. A detection limit of 5.7 ng/L was obtained using 7 repeated analyses of deionized water. The coefficients of variation in inter- and intraday tests were less than 5%. Mean recovery of 8-oxodG in urine was between 99% and 102%. Analyses of field blanks found no significant contamination. Each participant's urinary 1-OHPG and 8-oxodG concentrations were normalized to urinary creatinine.

Urinary metal concentration was analyzed by inductively coupled plasma-mass spectrometer (Agilent 7700X series ICP-MS; Agilent Technologies), as previously described (17). ICP-MS was used to determine the Cd, Ni, As, Pb, and Cu concentrations in the urine. Briefly, the frozen urine samples were first moved to a refrigerator and stored at 4°C for several hours, and then thawed at room temperature. Two milliliters of thawed urine were diluted 15 times with 1% HNO₃ diluent containing 1 μg/L yttrium (Y) internal standard in the final sample dilution, then mixed well, and finally aspirated from plastic containers into the ICP-MS. Standard solutions of 0.1, 0.5, 1, 2, 5, 10, 20, and 50 μg/L were used for sample quantification. The spike recoveries using 5 μg/L urinary matrix were: Cd, 94.1 ± 2.1; Ni, 100.5 ± 1.4; As, 99.7 ± 4.6; Pb, 99.3 ± 4.8; and Cu, 98.5 ± 3.0. The method detection limits were: Cd, 0.075; Ni, 0.022; As, 0.184; Pb, 0.095; and Cu, 0.023 μg/L. The outcome figures below the limit of detection (LOD) were set to LOD/√2.

The levels of urinary metals, urinary 1-OHPG, and urinary 8-oxodG were compared between the exposed and reference groups by mixed-model repeat measure analysis (Proc mixed). This model was used to investigate the correlations among urinary metals, urinary 1-OHPG, and 8-oxodG after adjusting for fixed covariates (such as age, sex, education level, smoking habit, season of data collection, and exposure status). These models treated the participants as random effects, and model selections were based on Akaike's Information Criterion. The compound symmetry and variance components were constructed as the covariance structures. The dependent variables were transformed by the natural logarithm (Ln). Residual and influence analyses were conducted.

We also assessed the effect of interaction from exposure to PAHs and metals on urinary 8-oxodG in nonsmoking workers. We categorized the exposures as greater or less than the 50^th percentile among nonsmoking workers. We compared the geometric mean (GM) levels of urinary 8-oxodG according to the combination of these exposures in the model. A 2-sided P value less than 0.05 was considered statistically significant. All statistical analyses were carried out using SAS (version 9.1.3; SAS Institute Inc., Carry, NC).

Certain characteristics of the exposed and reference groups of the study population were compared (Table 1). The mean age was 49.2 years (SD 9.17) in the exposed group and 42.9 years (SD 8.86) in the reference group. The exposed group had a lower percentage of women and a lower education level than the reference group. The most common season for data collection was spring for the exposed group, while it was spring and winter for the reference group. The distributions of lifestyle factors, such as smoking habits, alcohol consumption habits, and cooking habits, were similar between the exposed and reference groups.

Regardless of whether the workers were nonsmoking or smoking, the geometric mean levels of urinary Ni, Cu, 1-OHPG, and 8-oxodG among the exposed group were significantly higher than those among the reference group (Table 2). No differences of geometric mean levels of urinary Cd, As, and Pb were observed between the exposed and reference groups.

Table 3 presents the relations of urinary heavy metals and 1-OHPG levels with urinary 8-oxodG levels. After controlling for the fixed covariates, we found that the urinary Cd and 1-OHPG levels were positively associated with urinary 8-oxodG levels in all workers and nonsmoking workers. Due to small sample size in smokers, further analysis was limited. We also examined this association using data separately for each day (Supplementary Table S1A). The conclusion is the same with similar β and P values.

For further analysis of the interaction of exposure to both PAHs and Cd, we grouped participants according to the median levels of urinary 1-OHPG and Cd among nonsmoking workers. Workers with high urinary 1-OHPG and high urinary Cd levels had, on average, 60.3% higher levels of urinary 8-oxodG than did those with low urinary 1-OHPG and low urinary Cd levels in the adjusted model. Workers with high urinary 1-OHPG and low urinary Cd levels and workers with low urinary 1-OHPG and high urinary Cd levels had, on average, 25.4% and 36.3% higher levels of urinary 8-oxodG, respectively, than workers with low urinary 1-OHPG and low urinary Cd levels. The P value for the interaction term of both exposures (P = 0.001) is consistent with an additive interaction between PAHs and Cd on urinary 8-oxodG levels. Figure 1 shows GM and 95% confidence interval (CI) of adjusted urinary 8-oxodG levels by urinary 1-OHPG and urinary Cd concentrations.

To our understanding, this is the first study to indicate that levels of urinary 1-OHPG and Cd had significant associations with levels of urinary 8-oxodG. With increasing levels of urinary 1-OHPG and Cd, elevated levels of 8-oxodG in urine were observed. These results provide evidence that co-exposure to PAHs and Cd could increase levels of oxidative stress.

Concentrations of urinary Cd could be regarded as a well-recognized biomarker of chronic Cd exposure (18). Cd elicits oxidative stress by inducing the generation of ROS, reducing the antioxidant defense systems of cells by depleting glutathione, and decreasing the activities of cellular antioxidant enzymes, leading to mitochondrial injuries and increasing the susceptibility of cells to oxidative attack by altering membrane integrity and fatty acid composition (18, 19). Consequently, it is plausible that the toxic effect from Cd can be partially responsible for the impaired balance between oxidants and antioxidants. The strong association observed between urinary Cd levels and oxidative stress markers in the third U.S. National Health and Nutrition Survey suggests that Cd could increase the burden of oxidative stress (20). Bae and colleagues, in 2010, also reported that exposure to Cd from air pollution was associated with oxidative stress in schoolchildren (5). However, the lack of correlation between exposure to these metals, including Ni, As, Pb, and Cu, and oxidative stress could be attributed to a low biologically relevant dose in the study population. Although there were consistent negative associations between other metals (Ni, As, and Pb) and urinary 8-oxodG, these negative associations might be generated by statistical random variations.

We found no differences in urinary Cd levels between the traffic conductors and indoor offices workers. Although air pollution from traffic exhausts was identified as a possible source of Cd (21), contribution from diets cannot be ruled out. Dietary intake is one of the important sources of exposure to Cd in the general nonsmoking population, especially in consumption of some foods such as rice, other crops, and shellfish (22–25). The urinary Cd concentrations in our study population were higher than those reported in the third U.S. National Health and Nutrition Survey (geometric mean: 0.37 μg/g creatinine; ref. 20) and similar to those reported for the Japanese population (geometric mean 0.8 μg/g creatinine; ref. 26). Overall, most participants (99%) had urinary Cd levels of less than 5 μg/g creatinine, the World Health Organization (WHO) standard for urinary Cd (27).

The urinary 1-OHPG levels in traffic conductors were significantly higher than those in indoor office workers, which were similar to the findings in toll-station workers (9), bus drivers (7), and police officers (28). These results suggest that traffic sources have been recognized as major contributors to PAHs and have an effect on urinary 1-OHPG levels. Conversely, the urinary 1-OHPG concentrations in traffic conductors (range, 0.08–0.11 μmol/mol) were lower than those reported in toll-station workers (range, 0.12–0.16 μmol/mol; ref. 9), painters in shipyards (range, 0.99–1.4 μmol/mol; ref. 12), and incinerator workers (mean, 0.24 μmol/mol; ref. 11), indicating a moderate exposure to PAHs in the present study. The main pathways of metabolic activation of PAHs are the formation of anti- and syn-diol epoxides, the formation of radical cations, and the formation of o-quinones (3). The radical cations can generate ROS and cause oxidative stress by redox cycling (29, 30). Epidemiologic studies found that exposure to PAHs from air pollution could be positively associated with urinary 8-oxodG levels (9, 31, 32).

Urinary Cd levels and urinary 1-OHPG levels as independent predictors of urinary 8-oxodG levels indicate the hypothesis that Cd and PAHs could induce oxidative stress through different mechanisms. Environmentally relevant pollutants seldom occur alone. Little is known on the exact mechanism of carcinogenesis of two or more pollutants when they are present together. Our results support the assumption that the concept of additivity is operative on moderate- or low-level exposures to chemical mixtures. Valavanidis and colleagues in 2005 showed that transition metals, redox cycling quinoids, and PAHs act synergically to produce ROS, to increase the burden of oxidative stress (33). Cd inhibits several enzymes involved in DNA repair, and this has been identified as a major mechanism underlying the carcinogenic potential of Cd (19, 34). Thus, the DNA damage might be induced by oxidative stress from exposure to Cd and/or PAHs, and further enhanced by impaired repair from one or both. In this way, the effect of interaction was observed although the concentrations were moderate. Future studies are required to clarify these findings.

The urinary excretion of products of damaged nucleotides in cellular nucleotide pools or in DNA may be important biomarkers of exposure to relevant carcinogens and may predict cancer risk. Urinary 8-oxodG in steady state reflects products of DNA damage and repair (35), and could also be regarded as biomarkers of oxidative stress in occupational or environmental fields (6–10). Our results also showed that traffic conductors could have higher levels of urinary 8-oxodG than indoor office workers, which were in agreement with previous studies (7, 9). Air pollution from traffic exhaust could increase levels of oxidative DNA damage. Other factors, such as cell death and diets, might contribute to increase urinary 8-oxodG levels. However, the evidence was limited (36, 37). Although genetic background (such as DNA repair capability) from each participant could confound the estimated effect, two measurements on the consecutive days in the present study could minimize these effects.

An early prospective study had reported that elevated levels of urinary 8-oxodG were significantly associated with increasing risk of lung cancer among never-smokers (38). Recently, a nested case-control study of lung cancer and diesel exhaust indicated that workers in highly polluted cities over a lifetime had at least a 50% increased risk of developing lung cancer (39). Exposures to PAHs and Cd are known as risk factors of many illnesses. Therefore, traffic conductors in urban environments could be more susceptible to injury and disease caused by exposure to PAHs and metals.

In summary, our results indicate that co-exposure to urinary 1-OHPG and Cd could increase the levels of urinary 8-oxodG, suggesting that the additive interaction between exposure to environmental PAHs and Cd could enhance the burden of oxidative stress.

No potential conflicts of interest were disclosed.

Conception and design: S.-H. Liou, C.-H. Lai, S.-L.J. Wang

Development of methodology: C.-J. Wang, S.-H. Liou, C.-H. Lai, S.-L.J. Wang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): G.-W. Chen, Y.-Y. Lin, C.-H. Lai, S.-L.J. Wang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H.-B. Huang, Y.-Y. Lin, S.-H. Liou, C.-H. Lai, S.-L.J. Wang

Writing, review, and/or revision of the manuscript: H.-B. Huang, C.-H. Lai, S.-L.J. Wang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G.-W. Chen, C.-J. Wang, C.-H. Lai, S.-L.J. Wang

Study supervision: C.-H. Lai, S.-L.J. Wang

This work was supported by funding from the National Research Program for Genomic Medicine, Taiwan (grant nos. DOH97-TD-G-111-032, DOH98-TD-G-111-021, DOH99-TD-G-111-018, and NSC100-2325-B400-010).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.