Malondialdehyde–Deoxyguanosine Adducts among Workers of a Thai Industrial Estate and Nearby Residents

Background Humans living near industrial point emissions can experience high levels of exposures to air pollutants. Map Ta Phut Industrial Estate in Thailand is the location of the largest steel, oil refinery, and petrochemical factory complexes in Southeast Asia. Air pollution is an important source of oxidative stress and reactive oxygen species, which interact with DNA and lipids, leading to oxidative damage and lipid peroxidation, respectively. Objective We measured the levels of malondialdehyde–deoxyguanosine (dG) adducts, a biomarker of oxidative stress and lipid peroxidation, in petrochemical workers, nearby residents, and subjects living in a control district without proximity to industrial sources. Design We conducted a cross-sectional study to compare the prevalence of malondialdehyde-dG adducts in groups of subjects experiencing various degrees of air pollution. Results The multivariate regression analysis shows that the adduct levels were associated with occupational and environmental exposures to air pollution. The highest adduct level was observed in the steel factory workers. In addition, the formation of DNA damage tended to be associated with tobacco smoking, but without reaching statistical significance. A nonsignificant increase in DNA adducts was observed after 4–6 years of employment among the petrochemical complexes. Conclusions Air pollution emitted from the Map Ta Phut Industrial Estate complexes was associated with increased adduct levels in petrochemical workers and nearby residents. Considering the mutagenic potential of DNA lesions in the carcinogenic process, we recommend measures aimed at reducing the levels of air pollution.

Humans living in urban areas situated near industrial point emissions, such as steel facto ries and petrochemical complexes, can experi ence high levels of exposure to air pollutants (Kibble and Harrison 2005;Peluso et al. 2008), including a variety of known carcino gens (Cohen 2000). Map Ta Phut Industrial Estate (MIE) in Rayong Province, Thailand, is the location of the largest steel, oil refin ery, and petrochemical factory complexes in Southeast Asia. MIE complexes produce mix tures of air pollutants containing nitrogen dioxide, ozone, propylene, ethylene, benzene, and polycyclic aromatic hydrocarbons (PAH) together with particulate matter onto which compounds are absorbed (Bakker et al. 2000;Bergamaschi et al. 2005;Boogard and Van Sitter 1994;Kim et al. 2005;Sørensen et al. 2003;Tang et al. 2006;Yang et al. 2002).
We have recently undertaken a cross sectional study to evaluate the formation of bulky DNA adducts, a biomarker of PAH exposure (Peluso et al. 2001), in people living and working in Map Ta Phut (MTP) (Peluso et al. 2008). That study showed an increased formation of bulky DNA adducts in the MIE workers with respect to those living in Rayong Province. Furthermore, the subjects living near the MIE complexes experienced an excess of bulky DNA adduct formation relative to resi dents of a control district in the same province.
Air pollution is also an important source of oxidative stress and reactive oxygen species (ROS), which can interact with DNA and lip ids, leading to oxidative damage and lipid per oxidation (LPO), respectively (Buthbumrung et al. 2008;CalderónGarcidueñas et al. 1999;Chen et al. 2007;Møller et al. 2008;Singh et al. 2007;Sørensen et al. 2003).
Malondialdehyde (MDA) is a natural product of LPO, which is also formed dur ing prostaglandin E 2 biosynthesis via cyclo oxygenase (Marnett 1999). MDA is an aldehyde capable of interacting with DNA to form exo cyclic adducts, including 3(2deoxy βderythropentafuranosyl)pyrimido [1,2α] purin10(3H)one deoxyguanosine (M 1 dG). This adduct can be also generated through base propenal intermediates (Jeong and Swenberg 2005). The importance of M 1 dG adducts in carcinogenesis is emphasized by their ability to induce base pair mutations and cause frameshift mutations in reiterated sequences (VanderVeen et al. 2003).
A physiologic background of M 1 dG adducts has been reported in a number of human tissues, including breast, colon, and bronchial mucosa, which seems to be influ enced by individual susceptibility and envi ronmental factors, including dietary and lifestyle habits (Fang et al. 1996;Leuratti et al. 2002;Munnia et al. 2006;Wang et al. 1996). For instance, we and others found that the formation of M 1 dG adducts is associated with smoking habit in laryngeal, bronchial, and oral mucosa (Munnia et al. 2004(Munnia et al. , 2006Zhang et al. 2002). Furthermore, popula tionbased studies suggest that increased levels of M 1 dG adducts can be related to cancer develop ment and tumor progression (Munnia et al. 2004(Munnia et al. , 2006Wang et al. 1996).
In the present study, we evaluated whether air pollution emitted from MIE complexes was associated with levels of M 1 dG adducts, a biomarker of oxidative stress and LPO. Our approach consisted of a crosssectional study to compare the prevalence of DNA damage in groups of subjects experiencing various degrees of air pollution exposures (Peluso et al. 2008). This was done by meas uring the amount of M 1 dG adducts in MIE workers, nearby residents, and subjects living in a con trol district in the same province (Rayong) but without proximity to industrial sources. The levels of M 1 dG adducts were measured using the 32 PDNA postlabeling technique (Munnia et al. 2006).

Materials and Methods
Study subjects. Study subjects working in MIE complexes were identified and recruited by the industrial health service. MTP participants who were nearby residents or living in a con trol district from the Rayong Province without proximity to industrial sources were contacted and recruited by local health personnel. Only control subjects without occupational history in industries entailing exposure to known or suspected carcinogens were eligible in the MTP study. The study population comprised three groups of subjects: a) MIE workers, b) nearby residents, and c) residents in a control district. volume 118 | number 1 | January 2010 • Environmental Health Perspectives Ninetyfive percent of the eligible residents in each study group participated in the MTP study. Written consent to participate in the MTP study was given by all subjects fulfill ing inclusion criteria after they were given a detailed description of sampling procedures and the aims of the project. The MTP study was approved by the relevant ethical commit tee. A questionnaire concerning occupational history, smoking habit, and diet was adminis tered to study subjects before blood sampling.
Preparation of the reference M 1 dG adduct standards. We prepared two reference adduct standards using MDA or hydrogen peroxide (H 2 O 2 ). For MDA, calf thymus (CT) DNA or leukocyte DNA from a blood donor was treated with 10 mM MDA as described by Sun et al. (2004), yielding M 1 dG (Leuratti et al. 1998;Sun et al. 2004;Vaca et al. 1995). For H 2 O 2 , the epithelial lung carcinoma cell line A549 was exposed to 100 µM H 2 O 2 . MDAtreated CT DNA was then diluted with untreated CT DNA to obtain decreasing levels of the reference adduct standard to generate a calibration curve (R 2 = 0.99).

Measurement of M 1 dG adducts in peripheral leukocytes.
We measured the levels of M 1 dG adducts using our previously described method (Munnia et al. 2006) with minor modifications. DNA (1-2 µg) was digested by micro coccal nuclease and spleen phospho diesterase. Hydrolyzed samples were treated with nuclease P1 (2.5 µg) for 30 min at 37°C. The nuclease P1treated samples were incubated with 15-25 µCi [γ 32 P]ATP and T4polynucleotide kinase (0.75 U/µL) to gen erate labeled M 1 dG adducts. Samples were applied to the origin of chromatograms and developed with 0.35 MgCl 2 up to 2.0 cm on a filter paper wick. Plates were developed in the opposite direction with 2.1 M lithium for mate, 3.75 M urea (pH 3.75), and then run at the right angle to the previous development with 0.24 M sodium phosphate, 2.4 M urea (pH 6.4).
We detected and quantified M 1 dG adducts and total nucleotides using the storage phosphor imaging technique, which employs intensifying screens (Molecular Dynamics, Sunnyvale, CA, USA) for 0.20-48 hr. The screens were scanned using a Typhoon 9210 (Amersham, Little Chalfont, Buckinghamshire, UK). To pro cess the data, we used ImageQuant (Molecular Dynamics, Sunnyvale, CA, USA). After background subtraction, the levels of M 1 dG adducts were expressed as relative adduct label ing [screen pixels in adducted nucleotides ÷ screen pixels in total normal nucleotides (NN)]. To calculate the levels of total NN, ali quots of hydrolyzed DNA were appropriately diluted and reacted in the mixtures used for M 1 dG adduct labeling. The 32 Plabeled total nucleotides obtained were separated on Merck PEIcellulose TLC plates using 280 mM ammonium sulfate, 50 mM sodium phosphate. The values meas ured for the M 1 dG adducts were corrected across experi ments based on the recovery of the internal standard after the 32 PDNA post labeling assay.
Statistical analysis. All statistical analyses were performed on logtransformed data to stabilize the variance and normalize the distri bution of M 1 dG adducts. MIE workers and nearby residents were grouped according to tertiles for duration of employment and years spent at residence before statistical analyses.
We initially performed a descriptive analy sis to explore the relationship between indi vidual variables and M 1 dG adducts. In the univariate setting, the mean levels of DNA adducts across the levels of each variable, that is, age, sex, smoking habit, residence, employ ment, duration of employment (among petro chemical workers), and years of residence (among nearby residents), were compared by analysis of variance. Post hoc Dunnett tests were performed for multiple comparisons among such variable levels.
The multivariate analysis was then per formed using lognormal regression models including terms for type of exposure, age, sex, smoking habit, years spent at the residence, and duration of employment to estimate the effect of each variable on the outcome, adjusting for the concomitant effect of the other variables included in the model. The regression parameters estimated from the model are interpreted as ratios between the means of M 1 dG adducts of each level of the study variables with respect to the refer ence level, adjusted by age, sex, and smoking habit. A pvalue of < 0.05 (twotailed) was considered significant for all of the tests. The data were analyzed using SPSS 13.0 (SPSS, Chicago, IL, USA).

Results
Reference M 1 dG adduct standards. We first evaluated whether MDA treatment was capable of inducing the formation of M 1 dG adducts in CT DNA in vitro. A sta tistically significant increased formation of M 1 dG adducts was found in MDAtreated DNA rela tive to control DNA (p < 0.001). The mean levels of M 1 dG adducts per 10 6 NN ± SE were 0.32 ± 0.06, 1.6 ± 0.23, and 5.0 ± 0.45 in 1 mM, 4 mM, and 10 mM MDAtreated DNA, respectively, while a mean of 0.06 M 1 dG adducts per 10 6 NN (± 0.01) was detected in untreated DNA.
In a subsequent experiment, we analyzed whether MDA treatment was capable of caus ing adduct formation in human leukocyte DNA. We found a statistically significant increase in M 1 dG adducts ( Figure 1A) rela tive to control DNA (p < 0.001). The mean level of M 1 dG adducts per 10 6 NN ± SE was 2.2 ± 0.6 and 0.02 ± 0.01 in MDAtreated and untreated DNA, respectively. This adduct spot was previously identified as M 1 dG using different techniques (Leuratti et al. 1998;Sun et al. 2004;Vaca et al. 1995). The presence of a background adduct spot in the untreated samples is in keeping with previous results reporting background levels of M 1 dG adducts in control DNA (Leuratti et al. 1998;Sun et al. 2004).
Next, we analyzed whether free radicals were capable of inducing the same DNA lesion in an in vitro system by incubating epithelial lung carcinoma cell line A549 with H 2 O 2 . Our findings showed that treatment with 100 µM H 2 O 2 induced a statistically significant increase in M 1 dG adducts in the lung carcinoma cells relative to the unexposed cells (p < 0.001). The average level of M 1 dG adducts per 10 6 NN was 0.25 ± 0.09 and 0.07 ± 0.01 in H 2 O 2 treated and untreated cells, respectively. M 1 dG adducts in the leukocytes of MTP study subjects. We measured the formation of M 1 dG adducts in the leukocyte DNA of 173 subjects: 38.73% were MIE work ers, 33.53% were nearby residents, and 27.74% were living in a control district of the Rayong Province. The mean age of MIE workers, MIE residents, and district con trols was 31.6 ± 6.7 years, 36.2 ± 8.5 years, and 34.3 ± 6.8 years, respectively. There was a higher proportion of males than females 2.5 cm (80.3%); 38.2% of participants were classified as non smokers, 4.0% as former smokers, and 57.8% as current smokers, based on smoking during the 3 months before blood sampling. Figure 1 shows the pattern of M 1 dG adduct spots in the chromatograms of study participants. The intensity of the adduct spot was stronger in the plates of MIE workers and nearby residents compared with chromato grams of subjects living in the control district. Table 1 reports the distributions of demo graphic characteristics and mean levels of M 1 dG adducts according to exposure group. Adduct levels were higher in MIE work ers and nearby residents than in the control group. The mean levels of M 1 dG adducts of current smokers were increased, but without reaching statistical significance, in nearby resi dents and in the control group. No effect of smoking was observed in MIE workers. Table 2 shows the mean levels of M 1 dG adducts per 10 8 NN ± SE in the MTP study according to residence and employment, years of employment (among petro chemical work ers), and years of residence (among nearby residents) and for sex, age, and smoking habit for the whole population plus adjusted ratios of mean adduct levels based on the multi variate regression model.
Observed (unadjusted) mean lev els of M 1 dG adducts per 10 8 NN ± SE of MIE workers (6.0 ± 0.5) and nearby resi dents (3.7 ± 0.4) were significantly higher than those of subjects living in the control district (2.9 ± 0.4; pvalues < 0.001 and 0.014, respectively). Steel factory work ers had the highest levels of M 1 dG adducts (6.4 ± 0.7;p < 0.001). When recent smoking habit (within 3 months) was considered, the adduct levels of current smokers (4.8 ± 0.4) were significantly higher that those of non smokers (3.7 ± 0.4, p = 0.027). A nonsignifi cant increase in adduct levels was associated with duration of employment. No other uni variate effects were observed.
The results of the multivariate analysis confirm that the levels of M 1 dG adducts were significantly higher among MIE workers with respect to the control groups. Moreover, adduct levels of nearby residents were sig nificantly higher than those of subjects living in the control district. The highest adduct level was observed in the steel factory work ers. The formation of DNA damage tended to be associated with tobacco smoking, but without reaching statistical significance. A non significant increase in M 1 dG adducts was associated with 4-6 years of employment, but adduct levels were not associated with longer duration of employment. The adduct forma tion of MIE workers seems to reach some kind of saturation point for longer duration. No association with the duration of nearby residence was found.

Discussion
One of the largest steel, oil refinery, and petro chemical factory complexes in Southeast Asia has been located in MTP (Thailand) since 1988. Coal power plants and oil power plants, capable of using several byproducts including petroleum coke derived from oil refinery coker units or other cracking processes, are located at the MIE site for power generation. Complex mixtures of air pollutants, including benzene, toluene, benzo(a)anthracene, benzo(a)chrysene, and transition metals are produced from such petro chemical installations and energy plants (Bakker et al. 2000;Bergamaschi et al. 2005;RomaTorres et al. 2006). Steel factories are able to melt tons of metal per year, generating high PAH emissions (Yang et al. 2002).
In the present study, average levels of M 1 dG adducts were significantly higher in MIE workers than in control groups. In addi tion, adduct levels were highest among the individuals working in factories with indus trial processes characterized by emissions of PAH and transition metals, such as steel facto ries. Most likely, steel workers have increased exposure to iron, which can promote Fenton chemis try leading to oxidative stress. Finally, the mean level of M 1 dG adducts among nearby residents was significantly higher than the mean level among subjects living in a control district without proximity to industrial sites.
The industrial emissions from MIE com plexes are the most important source of air pollutants in the MTP area and are therefore likely to be involved in the increased adduct levels observed among MTP residents. Indeed, industrial air pollution constituents can induce DNA damage in a number of ways, including  CI, confidence interval. For the parameter estimates, the effect of each variable (means ratio) is the ratio between the mean adducts of each level of study variables with respect to the reference level, adjusted by age, sex, and smoking habit. a Some figures do not add up to the total because of missing values. b Levels of adducts per 10 8 NN. c Separate models were used to estimate associations according to residence and type of employment, duration of employment (among industrial workers only; n = 54), and duration of residence near the industrial complex (among nearby residents only; n = 58), with adjustment for sex, age, and smoking. d Reference level.
volume 118 | number 1 | January 2010 • Environmental Health Perspectives through production of ROS, which can initi ate LPO and cause an intra cellular excess of MDA. ROS can also induce the production of M 1 dG adducts through deoxyribose oxida tion. For instance, benzenederived quinones are reactive intermediates with the ability to redoxcycle, which is a reaction that gener ates super oxide, H 2 O 2 , and hydroxyl radicals (Bolton et al. 2000;Sørensen et al. 2003). An alternative mechanism by which air pol lutants can induce DNA damage involves the action of transition metals, such as iron, cop per, and chromium, on the surface of the par ticulate matter, which produces ROS through the Fenton reaction (Sørensen et al. 2003). Finally, air pollution exposure can activate macrophages and neutrophils that release ROS, such as H 2 O 2 and hypochlorite acid. Several studies have showed higher levels of oxidative DNA damage among subjects exposed to air pollution (Sørensen et al. 2003). Increased LPO, such as measured by MDA or 8isoprostaglandinF 2α , was detected in individuals exposed to air pollutants (Chen et al. 2007;Sørensen et al. 2003;Yang et al. 2007). High levels of M 1 dG adducts were also reported in urban workers from Sofia, Bulgaria, but not in the police officers from Kosice, Slovac Republic, and Prague, Czech Republic (Singh et al. 2007). Admittedly, we realize that our study could ideally have been conducted at the level of bronchi, which are more exposed to airborne carcinogens and more competent in terms of metabolic activa tion. However, although leukocytes are not the direct target of environmental carcino gens, the level of carcinogen-DNA adducts in leuko cytes correlates with carcinogeninduced damage in human lung tissues (Peluso et al. 2005;Tang et al. 2001;Wiencke et al. 1995). In addition, cells from peripheral blood that migrate and circulate through the lung can be exposed to accumulated unmetabolized toxic compounds in this tissue (Wiencke 2002).
In the present study, the formation of M 1 dG seems to saturate among the petro chemical workers that have been exposed for several years to the air pollutants emitted from the petrochemical complexes. Indeed, the levels of M 1 dG adducts tended to increase after 4-6 years of employment and to reach a steadystate level after 7 years. Conversely, we did not observe any effects with duration of nearby residence. Higher exposure levels may be necessary to induce similar effects in environmentally exposed residents. Table 3 compares the levels of M 1 dG adducts found in the present study with those detected by different laboratory methods in DNA of leukocytes as reported in the litera ture. The levels of M 1 dG adducts detected in the present study are in keeping with those determined in the leukocytes of human volun teers by immuno affinity purification gas chro matography/electrochemical detection negative chemical ionization/mass spectrometry (ima fin/GCMS NCI/MS), immunoenriched 32 P postlabeling, and immuno slot blot techniques (Leuratti et al. 1998;Rouzer et al. 1997;Sun et al. 2004). Higher M 1 dG adduct levels were reported in a multi center occupational study in Eastern Europe and in Finnish volunteers by an earlier 32 Ppostlabeling/reversephase high performance liquid chromatography (HPLC) technique and an immunoslot blot method (Singh et al. 2007;Vaca et al. 1995).
Current smokers inhale a broad range of carcinogens and ROS derived from tobacco pyrolysis products, which can lead to M 1 dG adduct formation in a number of ways. In addition, the relatively long halflife of M 1 dG adducts of 12.5 days (Marnett 1999) renders it a potentially interesting biomarker of oxida tive stress and carcinogen exposures, including recent smoking. Thus, we analyzed the associa tion between levels of M 1 dG adducts and smok ing. We found a significant difference between the levels of DNA damage of smokers and non smokers using univariate analysis. However, the effect was less evident in the multi variate analysis. The association between smoking habit and M 1 dG adduct levels may have been con founded by air pollution or other factors. This may explain why an effect of tobacco smoking on adduct levels was no longer evident in results from the adjusted regression model. We previously examined the relation ship between smoking and endogenous DNA adducts and found higher values of bron chial and laryngeal DNA damage in smok ers (Munnia et al. 2004(Munnia et al. , 2006. The effects of smoking on bronchial DNA adducts also persisted when urinary thio cyanate was used to meas ure the extent of exposure to tobacco smoke (Munnia et al. 2006). Similar relation ships have been reported in the oral mucosa of tobacco smokers (Zhang et al. 2002), but other investigators reported no differences for smoking habit in breast and colon mucosa (Leuratti et al. 2002;Wang et al. 1996).
The potential effect of MIE emissions on the health of nearby residents has been widely studied in recent years. An increased can cer incidence was reported in the MTP area (Peluso et al. 2008). Agestandardized inci dence rates for all cancers in 1997-2001 were 181.0 in men and 183.9 in women in the MTP area compared with 122.6 and 116.8 in the rest of the Rayong Province. An excess risk of respiratory diseases was also found in nearby residents (Jadsri et al. 2006). We recently showed that MIE workers and nearby residents can experience an excess formation of bulky DNA adducts, a biomarker of PAH exposure possibly related to lung cancer risk (Peluso et al. 2005). In the present study, we found that levels of M 1 dG adducts (a biomarker of oxidative stress and LPO) also were increased in MIE workers and nearby residents.
Oxidative stressinduced DNA damage is an important marker of air pollution exposure (Sørensen et al. 2003). Furthermore, oxidative stress and LPO are thought to underlie the etiology of many cancers (Leuratti et al. 2002;Munnia et al. 2004Munnia et al. , 2006Wang et al. 1996). For instance, we observed that higher levels of endogenous DNA adducts were increased in lung cancer cases compared with controls, but only in smokers (Munnia et al. 2006). In addition, lung cancer cases with levels of MDA-DNA adducts above the popu lation median had reduced survival, although not statistically significant, after adjusting for age, sex, and smoking habit (Munnia et al. 2006).

Conclusions
Air pollution exposure was associated with increased formation of M 1 dG adducts, a bio marker of oxidative stress, and LPO in MIE workers and nearby residents compared with individuals living in a nonindustrial control district. Thus, considering the mutagenic potential of the DNA lesions in the carcino genic process, we recommend measures of air pollution control aimed at reducing the levels of air pollutants in the MTP area.