Malondialdehyde–deoxyguanosine and bulky DNA adducts in schoolchildren resident in the proximity of the Sarroch industrial estate on Sardinia Island, Italy

Abstract

Air quality is a primary environmental concern in highly industrialised areas, with potential health effects in children residing nearby. The Sarroch industrial estate in Cagliari province, Sardinia Island, Italy, hosts the world’s largest power plant and the second largest European oil refinery and petrochemical park. This industrial estate produces a complex mixture of air pollutants, including benzene, heavy metals and polycyclic aromatic hydrocarbons. Thus, we conducted a cross-sectional study to evaluate the prevalence of malondialdehyde–deoxyguanosine adducts in the nasal epithelium of 75 representative children, aged 6–14 years, attending primary and secondary schools in Sarroch in comparison with 73 rural controls. Additionally, the levels of bulky DNA adducts were analysed in a subset of 62 study children. DNA damage was measured by ³²P-postlabelling methodologies. The air concentrations of benzene and ethyl benzene were measured in the school gardens of Sarroch and a rural village by diffusive samplers. Outdoor measurements were also performed in other Sarroch areas and in the proximity of the industrial estate. The outdoor levels of benzene and ethyl benzene were significantly higher in the school gardens of Sarroch than in the rural village. Higher concentrations were also found in other Sarroch areas and in the vicinity of the industrial park. The mean levels of malondialdehyde–deoxyguanosine adducts per 10⁸ normal nucleotides ± standard error (SE) were 74.6±9.1 and 34.1±4.4 in the children from Sarroch and the rural village, respectively. The mean ratio was 2.53, 95% confidence interval (CI): 1.71–2.89, P < 0.001, versus rural controls. Similarly, the levels of bulky DNA adducts per 10⁸ normal nucleotides ± SE were 2.9±0.4 and 1.6±0.2 in the schoolchildren from Sarroch and the rural village, respectively. The means ratio was 1.90, 95% CI: 1.25–2.89, P = 0.003 versus rural controls. Our study indicates that children residing near the industrial estate have a significant increment of DNA damage.

Introduction

Air quality is a primary environmental concern in highly industrialised areas, especially because of potential health effects in children residing near sources of contamination (1–3). The Sarroch industrial estate (SIE) in Cagliari province, Italy, hosts the second largest European oil-refinery and petrochemical complex, with a refining capacity of ~15 million tons per year. The world’s largest integrated gasification combined cycle plant, which gasifies heavy refining residues for power generation, is also located in the industrial area. Combined, the SIE park produces a mixture of air pollutants, including volatile organic compounds (VOCs) such as benzene, ethyl benzene, formaldehyde, xylenes and toluene; heavy metals such as cadmium, chromium (VI), lead and nickel compounds; and polycyclic aromatic hydrocarbons (PAHs) such as benzo(a)pyrene [B(a)P] (4–6).

VOCs and heavy metals are relevant sources of oxidative stress and reactive oxygen species (ROS), which interact with DNA and lipids, leading to oxidation of deoxyribose in DNA and lipid peroxidation (LPO) (7,8). In particular, ROS-inducing chemicals cause the production of malondialdehyde (MDA)—or more correctly, β-hydroxyacrolein, a reactive LPO by-product (9,10)—and the generation of base propenal intermediates in DNA, structural analogues of the enol tautomer of MDA (11). As a result, increased levels of 3-(2-deoxy-β-D-erythro-pentafuranosyl)pyrimido[1,2-α]purin-10(3H)-one–deoxyguanosine (M₁–dG) adducts are generated (10,11). Plastaras et al. showed that base propenals are more reactive than MDA towards DNA to form M₁–dG adducts (11). The relevance of M₁–dG adducts in the carcinogenic process is due to their ability to block replication and induce base-pair and frameshift mutations in repeated sequences (12). The formation of M₁–dG adducts is also associated with loss of DNA methylation in the long interspersed nuclear element-1 (LINE-1) repeated elements and in the coding region of the inflammatory cytokine ‘interleukin-6’ gene (13). Both DNA mutations and altered patterns of DNA methylation are important hallmarks in carcinogenesis, and clinical studies indeed showed that enhanced levels of M₁–dG adducts can be linked to cancer development and tumour progression (14–17).

Similarly, B(a)P and other PAHs form bulky DNA adducts (18), which, if unrepaired, induce mutations in, for instance, the TP53 gene (19). Bulky DNA adducts can be used as a biomarker of PAH exposure (20) and were found to be predictive of lung cancer risk (21,22). Recent studies showed that the environmentally induced loss of DNA methylation in the promoter region of TP53 and in LINE-1 repetitive sequences is mediated by the generation of bulky DNA adducts (13).

To identify potential health impacts associated with schoolchildren’s exposure to air pollutants, we recently conducted a pilot study in Sardinia, Italy, to evaluate the use of the ³²P-postlabelling assay in the detection of M₁–dG adducts, a biomarker of free radical-inducing chemical exposures (7–9), in 28 children, aged 12–14 years, residing near the SIE complex (23). Our findings suggested an association between residential vicinity to the industrial park and DNA damage; therefore, we decided to evaluate the prevalence of M₁–dG adducts in 148 children, 6–14 years of age, representative of the children attending primary and secondary schools in the municipality of Sarroch and Burcei, a rural village. The levels of bulky DNA adducts were also analysed in a subset of 62 children to characterise the effects of PAH emissions from the facilities located in the industrial estate. DNA damage was measured in the nasal epithelium by ³²P-postlabelling methodologies (8,24). The concentrations of two atmospheric VOCs, benzene and ethyl benzene, were measured in the school gardens of Sarroch and the rural village Burcei using diffusive samplers. Outdoor VOC measurements were also performed in other areas of Sarroch and in the proximity of the industrial estate. VOC concentrations were analysed by gas chromatography technique by Passam Ltd (Männedorf, Switzerland).

Materials and methods

Study population

The parents of 194 children, aged 6–14 years, representative of the children attending primary and secondary schools in the municipalities of Sarroch and Burcei, Cagliari province, Sardinia Island, Italy, were invited to participate in the cross-sectional study. The study population consisted of 96 schoolchildren from Sarroch and 98 rural controls. Sarroch is about 20 km south of Cagliari on the south-western coast of the Sardinia Island. The children from the community of Sarroch resided in the proximity of the industrial buffer zone of the SIE complex, specifically, near a national storage facility, constituted by 160 tank farms, containing raw materials and liquid products with an overall capacity of around 3.5 million cubic metres. The distance of the home residences from the industrial park ranged from 100 m to less than 1.6 km, in particular, the primary and the secondary schools were at about 5–600 m from the SIE complex. Conversely, the control children lived in Burcei (‘remote’ site), a rural village with limited traffic and without industrial settlements, ~40 km north-east of Cagliari. The parents were asked to complete a questionnaire aimed to collect age, gender and residence information. Signed informed consent to participate in the study was obtained from 181 participants. A self-administered questionnaire regarding smoking habit was completed by the children (12–14 years old) at school. The study was approved by the Institutional Review Board of the University of Cagliari, Cagliari, Italy.

Volatile organic compound analysis

The outdoor levels of benzene and ethyl benzene were analysed for 3 weeks, starting 1 week before biological sampling, with 19 diffusive samplers (Passam Ltd, Männedorf, Switzerland). Four diffusive samplers were located in the school gardens of Sarroch and Burcei. VOC measurements were also performed in other areas of Sarroch: one diffusive sampler was located in the city park, four in the city centre and seven in the residential areas. Three samplers were also positioned in the proximity of the industrial estate. The dosimeters were in accordance with the European Committee for Standardization (CEN—European Committee for Standardization, ‘Ambient air quality – Diffusive samplers for the determination of concentrations of gases and vapors—Requirements and test methods’, Brussels, 2000). VOC concentrations were analysed by gas chromatography technique by Passam Ltd (Männedorf, Switzerland).

Preparation of the reference M₁–dG adduct standards for ³²P-postlabelling and mass spectrometry analyses

Two reference adduct standards were prepared: calf-thymus (Ct) DNA and leucocyte DNA from a blood donor were treated with 10mM MDA (ICN Biomedicals, Irvine, CA, USA), as previously described (25). MDA-treated Ct-DNA was diluted with untreated Ct-DNA to obtain decreasing levels of the reference adduct standard to generate a calibration curve.

Mass spectrometry analysis

As fully described elsewhere (26,27), DNA adducts in MDA-treated Ct-DNA sample were analysed by mass spectrometry (Voyager DE STR from Applied Biosystems, Framingham, MA, USA) through the following sequence of steps: (i) reaction of DNA with sodium borohydride (NaBH₄) (28), followed by precipitation with isopropanol; (ii) digestion with snake venom phosphodiesterase and nuclease P₁ (NP₁); (iii) extraction of DNA adducts that are less polar than normal nucleotides on an OASIS cartridge (Waters Corp.); (iv) tagging with an isotopologue pair of benzoylhistamines (d₀ and d₄) in a phosphate-specific labelling reaction in the presence of carbodiimide (26,27); (v) removal of residual reagents by ion-exchange solid-phase extraction; (vi) resolution of tagged adducts by capillary reversed-phase high-performance liquid chromatography (HPLC) with a collection of drops onto a matrix-assisted laser desorption/ionisation (MALDI) plate; (vii) addition of matrix (α-cyano-4-hydroxycinnamic acid); and (viii) analysis by MALDI time-of-flight mass spectrometry (MALDI-TOF-MS).

Nasal brushing and DNA extraction and purification

Nasal epithelium samples were collected from the lower lateral nasal wall and the inferior turbinate in each nostril with a Papanicolaou (Pap) test cytobrush during school, after gently washing the nasal cavity of children with saline solution (0.9% NaCl) and cleaning it with a cotton swab. The brushing method described in this study was adopted because increased amounts of volatile particles are found on the head of the inferior turbinate due to air flow turbulence behind the nasal valve and limited mucociliary transport. Nasal cells were rinsed from the cytobrush into a tube containing saline solution and 10% acetyl cysteine and incubated for 30min (100 cycles/min shaking frequency) at room temperature. After centrifugation, the cell pellets were kept at −80°C until DNA extraction.

DNA was extracted and purified using a method that requires digestion with ribonuclease A, ribonuclease T₁ and proteinase K; successive extraction with saturated phenol, phenol/chloroform/isoamyl alcohol (25:24:1), chloroform/isoamyl alcohol (24:1); and finally, ethanol precipitation (29). DNA concentration and purity were determined using a spectrophotometer. Coded DNA samples were subsequently stored at –80°C until laboratory analyses.

DNA hydrolysis and nuclease P1 treatment

Nasal epithelium DNA (2–4 μg) was hydrolysed by incubation with micrococcal nuclease (21.45 mU/μl) and spleen phosphodiesterase (6.0 mU/μl) in 5.0mM Na succinate, 2.5mM calcium chloride, pH 6.0, at 37°C for 4.5h. Hydrolysed samples were treated with NP₁ (0.1U/μl) in 46.6mM sodium acetate, pH 5.0, and 0.24mM ZnCl₂ at 37°C for 30min (27). After NP₁ treatment, 1.8 μl of 0.16mM Tris base was added to the sample. The NP₁-resistant nucleotides were incubated with 7–25 μCi of carrier-free [γ-³²P]ATP (3000 Ci/mM) and polynucleotide kinase T₄ (0.75U/μl) to generate ³²P-labelled DNA adducts in bicine buffer (20mM bicine, 10mM MgCl₂, 10mM dithiotreithol, 0.5mM spermidine, pH 9.0) at 37°C for 30min (29).

M₁–dG adduct analysis.

³²P-labelled samples were applied on polyethyleneimine (PEI) cellulose thin layer chromatography (TLC) plates (Macherey-Nagel, Postfach, Germany). The chromatographic analysis of M₁–dG adducts was performed using a low-urea solvent system known to be effective for the detection of low-molecular-weight and highly polar DNA adducts (8): 0.35M MgCl₂ for the preparatory chromatography; and 2.1M lithium formate, 3.75M urea (pH 3.75) and 0.24M sodium phosphate, 2.4M urea (pH 6.4) for the two-dimensional chromatography. Detection and quantification of M₁–dG adducts and normal nucleotides (nn), i.e. diluted samples that were not treated with NP1, were performed by storage phosphor imaging techniques employing intensifying screens from Molecular Dynamics (Sunnyvale, CA, USA). The intensifying screens were scanned using a Typhoon 9210 (Amersham). Software used to process the data was ImageQuant (version 5.0) from Molecular Dynamics. After background subtraction, the levels of M₁–dG adducts were expressed as relative adduct labelling (RAL) = pixels in adducted nucleotides/pixel in nn. The levels of M₁–dG adducts were corrected across experiments based on the recovery of the MDA-treated ct-DNA adduct standard. Co-chromatography of nasal mucosa epithelium DNA samples together with the MDA-treated ct-DNA adduct standard was performed to identify the M₁–dG adducts detected in the chromatograms of the children using 2.1M lithium formate, 3.75M urea, pH 3.75, and 0.24M sodium phosphate, 2.4M urea, pH 6.4, or 0.24M sodium phosphate, 2.7M urea, pH 6.4.

Bulky DNA adduct analysis

³²P-labelled samples were applied on different PEI cellulose TLC plates to analyse the formation of bulky DNA adducts. The chromatographic resolution of bulky DNA adducts was carried out using a high-urea solvent system capable of ensuring efficient detection of aromatic/hydrophobic DNA adducts, such as B(a)P-related DNA adducts (29): 1.0M sodium phosphate, pH 6.8, for the preparatory chromatography; 4.0M lithium formate; 7.5M urea, pH 3.5, and 0.65M lithium chloride, 0.45M Tris base, 7.7M urea, pH 8.0, for the two-dimensional chromatography; and 1.7M sodium phosphate (pH 5.0) for the clean-up chromatography. The levels of bulky DNA adducts were expressed as RAL = pixel in adducted nucleotides/pixel in nn. The values of DNA adducts were corrected across experiments based on the recovery of B(a)P–DNA adduct standard (kindly donated by Prof. F. A. Beland, National Center for Toxicological Research, Jefferson, AR, USA), which was analysed in parallel.

Statistical analysis

Non-parametric statistical analyses were conducted to analyse the differences in the airborne levels of benzene and ethyl benzene between the contaminated and the rural sites. We used multivariate statistical analyses using log-normal regression models, including age, gender and residence as predictive variables, to evaluate the association between residential proximity to the SIE complex and the levels of DNA damage in the nasal epithelium of schoolchildren. The regression parameters estimated from the models were interpreted as ratios [means ratio (MR)] between the means of DNA adducts of each level of the categorical variables with respect to the reference level, whereas for continuous variables, they were interpreted as the percentage change in DNA damage caused by increasing the continuous variable by one unit, adjusted by age, gender and residence, as appropriate. The MR was used as a measure of effect. A value of P < 0.05 (two-tailed) was considered significant. Data were analysed using SPSS 13.0 (SPSS, USA).

Results

Study population

Thirteen children did not consent to the biological sampling during school attendance. Additionally, six children who reported to be smokers were excluded from the study. Adequate biological samples for the ³²P-postlabelling analysis of the levels of M₁–dG adducts were obtained from 148 volunteers. The mean age ± SD of the study population was of 10.6±1.7 years. Further, 77 participants were girls and 71 boys. Among these, 75 children resided in the municipality of Sarroch and 73 were living in the rural area of Burcei.

Volatile organic compounds

Table I reports the results of the analyses of the VOC concentrations in the ambient air of Sarroch and Burcei. Levels of benzene and ethyl benzene distinguished the industrial estate, the nearby residential areas, and the remote site from each other. The outdoor levels of the two gaseous pollutants were significantly higher in the school gardens of Sarroch than in the rural village. Significantly higher concentrations of VOCs were also found in other areas of Sarroch. In the vicinity of the industrial park, a mean value of benzene that exceeded the regulatory limit was found.

Table I.

The air ambient concentrations of two volatile organic compounds, benzene and ethyl benzene, in various areas of the municipalities of Sarroch and Burcei, a rural village

Diffusive sampler locations	n	Volatile organic compounds (µg/m3)
		Benzene ± SD	P value	Ethyl benzenea ± SD	P value
Rural village
School gardensb	2	1.4±0.08	–	0.15±0.0	–
Sarroch municipality
School gardens	2	2.7±1.1	0.019	1.0±0.5	0.019
City park	1	3.8±0.6	0.060	1.5±0.4	0.028
City centre	4	3.6±1.0	0.007	1.5±0.5	0.007
Residential areas	7	3.1±0.5	0.003	1.2±0.2	0.002
Urban area	14	3.2±0.8	0.001	1.3±0.4	0.001
Near the industrial park	3	5.8±2.2	0.010	1.8±0.4	0.008

Measurements were performed over a period of 3 weeks.

^aValues of half the limit of detection were assigned for ethyl benzene concentrations below the limit of detection.

^bReference level.

In detail, the outdoor levels of benzene (mean ± SD) were 2.7±1.1 and 1.4±0.08 µg/m³ in the school gardens of Sarroch and Burcei, respectively (P = 0.019). Even higher concentrations of benzene were found in the city park: 3.8±0.6 µg/m³; in the centre of city: 3.6±1.0 µg/m³; and in the residential areas: 3.1±0.5 µg/m³, with P = 0.060, P = 0.007 and P = 0.003, respectively, versus the remote site. The average concentration observed for benzene in the urban area of Sarroch was 3.2±0.8 µg/m³, P = 0.001, versus the remote site. The highest level of benzene was detected in the proximity of the industrial estate: 5.8±2.2 µg/m³, P = 0.010, versus the remote site.

The outdoor levels of ethyl benzene (mean ± SD) were 1.0±0.5 and 0.15±0.0 µg/m³ in the school gardens of Sarroch and the rural village, respectively (P = 0.019). Comparable levels of ethyl benzene were measured in other areas of Sarroch: 1.5±0.4 µg/m³ in the city park, 1.5±0.5 µg/m³ in the city centre, and 1.2±0.2 µg/m³ in the residential areas; P = 0.028, P = 0.007 and P = 0.002, versus the remote site, respectively. The highest level of ethyl benzene was found in the vicinity of the industrial estate (1.8±0.4 µg/m³, P = 0.008, versus the remote site). The mean level of ethyl benzene in the urban area of Sarroch was 1.3±0.4 µg/m³ (P = 0.001, versus the remote site).

M₁–dG adduct standard by ³²P-postlabelling analyses and mass spectrometry

The mean levels of M₁–dG adducts standard error (SE) were 5.0 0.6 and 0.06 0.01 adducts per 10⁶ nn in MDA-treated and untreated control Ct-DNA, respectively, as assessed by the ³²P-postlabelling assay. The average levels of M₁–dG adducts ± SE were 2.2 0.6 and 0.02 0.01 adducts per 10⁶ nn in MDA-treated and untreated leucocyte DNA, respectively, as assessed by the ³²P-postlabelling technique. Thus, the levels of M₁–dG adducts were significantly increased in MDA-treated Ct-DNA and MDA-treated leucocyte DNA with respect to untreated DNA (P < 0.001). A calibration curve was obtained by the analysis of decreasing concentrations of the MDA-treated Ct-DNA (R² = 0.99).

The presence of six MDA–DNA adducts in MDA-treated NaBH₄-reduced Ct-DNA was confirmed after mass tagging by mass spectrometry, based on agreement of the accurate (measured) with exact (true) masses of the compounds. Accurate masses were assigned by using known matrix peaks to calibrate the spectra. The structures of five of the adducts were originally reported by Goda and Marnett (28), who introduced the M nomenclature for these adducts that we have used in Figure 1, and the structure of the remaining one (M3mdC) was assigned by considering the structure of M3dC. M3mdC is seen in a later MALDI spot than M3dC (from reversed-phase HPLC) because M3mdC is less polar.

Fig. 1.

Detection of MDA–DNA adducts in MDA-treated, NaBH₄-reduced calf-thymus DNA by mass tagging-MALDI-TOF-MS. The M nomenclature of Goda and Marnett (28) is used to define the structures of the adducts. In the inset is shown the detection of M3mdC (an MDA adduct of 5-methyldeoxycytosine) that was present in a later MALDI spot from the reversed-phase HPLC separation relative to a spot that enabled the detection of the other adducts.

Nasal mucosa epithelium M₁–dG adducts

A typical ³²P-postlabelling M₁–dG adduct spot pattern was detected in the chromatograms of all samples of the study children. Co-chromatography confirmed the presence of M₁–dG adducts in the nasal epithelium DNA of the study population and the intensity of the M₁–dG adduct spots was generally stronger in the chromatograms of the schoolchildren from Sarroch compared with those of the rural controls. The mean levels of M₁–dG adducts in 148 schoolchildren from Sarroch and Burcei are reported in Table II. Increased levels of M₁–dG adducts, a marker of exposures to ROS-inducing chemicals (7–9), were detected in 75 children living in the residential areas around the industrial complex with respect to the 73 residing in Burcei, the rural village. In detail, the average levels of M₁–dG adducts ± SE were 74.6±9.1 and 34.1±4.4 adducts per 10⁸ nn in the schoolchildren from Sarroch and rural controls, respectively. We used multivariate statistical analyses to examine the effect of residential proximity to the industrial estate on the levels of M₁–dG adducts (Table II). Our findings show that, after adjusting for potential confounding factors, the MR of M₁–dG adducts for the children drawn from the community of Sarroch was 2.53, 95% confidence interval (CI): 1.71–2.89, P < 0.001, compared with rural controls. No effects of age and gender were observed.

Table II.

Mean levels of MDA–deoxyguanosine and bulky DNA adducts per 10⁸ normal nucleotides ± SE in the olfactory epithelium of schoolchildren resident near the Sarroch industrial estate and in a rural village according to age, gender and residence

	MDA–deoxyguanosine adducts					Bulky DNA adducts
	n	Mean ± SE	MR	CI	P value	n	Mean ± SE	MR	CI	P value
Age per unit	148	54.6±5.3	0.99	0.88–1.12	0.917	62	2.2±0.2	0.99	0.85–1.39	0.507
Gender
Girlsa	77	54.6±7.0	1	–	–	33	2.1±0.3	1	–	–
Boys	71	54.7±8.2	1.02	0.69–1.51	0.912	29	2.2. ± 0.4	1.06	0.69–1.64	0.781
Residence
Rural villagea	73	34.1±4.4	1	–	–	34	1.6±0.2	1	–	–
Sarroch	75	74.6±9.1	2.53	1.71–2.89	<0.001	28	2.9±0.4	1.90	1.25–2.89	0.003

^aReference levels.

Nasal mucosa epithelium bulky DNA adducts

It is known that PAH exposures induce aromatic/hydrophobic DNA adducts that are very different from those generated from oxidative stress and by free radical exposure; therefore, the levels of bulky DNA adducts, a biomarker of PAH exposures (20), were analysed in a subset of 62 children for whom nasal epithelium DNA was available. A diagonal radioactive zone (DRZ) was detected in DNA-adduct-positive samples. This DRZ pattern was more frequently found in the chromatograms of the schoolchildren living in the residential areas surrounding the industrial estate than those of the rural controls. Table II reports the average levels of DNA adducts detected in 62 schoolchildren from the municipalities of Sarroch and Burcei. In detail, the levels of DNA adducts ± SE were 2.9±0.4 and 1.6±0.2 adducts per 10⁸ nn in 28 children from Sarroch and 34 rural controls, respectively. The results of the subsequent multivariate statistical analyses using log-normal regression models showed a significant effect of the residential proximity to the industrial park on the levels of PAH-related DNA damage (Table II). The MR, adjusted for age and gender, of DNA adducts for the schoolchildren from Sarroch was 1.90, 95% CI: 1.25–2.89, P = 0.003 compared with the rural controls. Age and gender were not associated with the levels of DNA adducts.

Discussion

The outdoor concentrations of benzene and ethyl benzene, two VOCs reported to be carcinogenic and possibly carcinogenic for humans (30,31), were used as markers of the gaseous pollutants emitted from the facilities in the industrial estate. The air-monitoring results indicate that the outdoor levels of benzene and ethyl benzene were significantly increased in the school gardens of Sarroch with respect to those of the rural village. An exposure gradient and highest concentrations of VOCs were found in the proximity of the SIE complex, intermediate VOC levels in the nearby residential areas and the lowest levels in the remote site. Furthermore, the airborne levels observed for benzene in the vicinity of the industrial park exceeded the regulatory limit for air quality standards. On the whole, the air-sampling data showed that VOC pollutants were ubiquitous and widespread in the outdoor air of Sarroch, which is in line with previous reports from environmental protection agencies (5,6). Our measurements from diffusive samplers are not well representative of individual exposures due to spatial and temporal variations, but they provide evidence of children’s (and adults’) exposure via air. Recently, Svecova et al. (32) measured the concentrations of VOCs in a large study aimed to identify the impacts of air pollution on human health in the Czech Republic. Interestingly, the personal levels of VOCs were associated with the outdoor levels of VOCs and different life style factors, such as cooking, home heating and time spent outdoors. Unfortunately, we could not assess the potential relationships between VOC measurements and life style parameters in the present study.

Associations between occupational and environmental air pollution and subsequent increment of biomarkers of exposure and cancer risk are apparent (8,23,33–36). Subsequently, we investigated whether children living in residential communities close to industrial areas also exhibited increased levels of biomarkers, which could indicate potential health risks in later life. Hence, we analysed the effects associated with the air pollutants emitted from the Sarroch industrial estate on the schoolchildren residing nearby. Our most striking result shows that the children living in the residential community around the industrial estate had increased levels of M₁–dG adducts in comparison with the rural controls who lived 68 km from the SIE complex. Even though the chemical exposure to VOCs from the industrial estate was below the regulatory limits, the multivariate regression analysis showed that the frequency of M₁–dG adducts in the nasal epithelium of the children residing nearby was significantly enhanced compared with those in children of the reference area (MR = 2.53, 95% CI = 1.71–2.89, P < 0.001). Some of the air pollutants, such as benzene, ethyl benzene, formaldehyde and heavy metals, emitted from the industrial estate (4–6) play important roles in the production of oxidative stress and free radicals (7). For instance, benzene is metabolised in the liver by the CYP₂E₁ protein to phenol, hydroquinone and catechol, reactive metabolites with the ability of redox cycling, a reaction that generates superoxide, hydrogen peroxide and hydroxyl radical. Ethyl benzene is metabolised to dihydroxylated metabolites such as ethyl hydroquinone and 4-ethyl catechol, compounds that also cause the production of ROS and oxidative DNA damage by redox cycling (37). Formaldehyde is a substrate for the CYP₂E₁ protein and is oxidised by peroxidase, aldehyde oxidase and xanthine oxidase, with subsequent free radical generation (25). Heavy metals, such as iron and copper, can lead to inflammation and to a consequent excess of oxidative stress (7). During the inflammatory response, activated macrophages and neutrophils can generate a variety of highly reactive oxidants, such as hydrogen peroxide and hypochlorite acid, which can induce LPO and oxidation of deoxyribose in DNA (9–11). Enhanced ROS induction increases the background levels of M₁–dG adducts in humans (8–11).

Our findings are in line with other studies reporting increased levels of oxidative stress in children resident near industrial complexes using a biomarker approach (38,39). Wong et al. (38) measured the levels of oxidative stress to DNA in children, 10–12 years of age, residing near the Taichung thermal power plant in Taiwan and exposed to environmental carcinogenic metals. Schoolchildren with the highest urinary chromium and arsenic levels had significantly increased urinary levels of 8-hydroxy-2′-deoxyguanosine adducts with respect to controls. Additionally, Vujovic et al. (39) used biomarkers of oxidative stress to evaluate the relationship between industrial proximity and oxidative stress in schoolchildren 12–15 years of age from Pancevo, which hosts the largest petrochemical park in Serbia. Children who resided near the petroleum refinery complex showed significantly higher levels of biomarkers of oxidative stress, including plasma MDA concentrations and decreased superoxide dismutase activity, in comparison with rural controls.

M₁–dG adducts have been found in several tissues such as liver, breast, colon, bronchi, gastric mucosa and peripheral leucocytes from healthy human beings at levels ranging from 1 to 160 per 10⁸ nn (14–17,24,25,40,41). The levels of M₁–dG adducts observed in the nasal mucosa of children of the rural village were lower than those detected in healthy liver, colon and gastric mucosa and higher than those recorded in leucocytes, bronchi, and breast (14–17,24,25,40,41). The high capacity for xenobiotic metabolism found in the nasal epithelium, compared with respiratory mucosa (42), may account, in part, for the high amounts of DNA damage detected in the nasal mucosa of the rural controls.

A limitation of the present study is the lack of a suitable control population from an urban area. At first sight, we cannot exclude that the environment, in addition to the industrial estate, might have contributed to the observed increment of DNA damage in the children living in the surrounding areas of the SIE complex. However, the concentration of benzene in Sarroch (3.2 µg/m³) is comparable to that observed in metropolitan European cities (43), such as Birmingham (3.3 µg/m³), London (3.5 µg/m³), Liverpool (2.9 µg/m³), Copenhagen (2.9 µg/m³) and Barcelona (3.5 µg/m³) (33). Furthermore, the levels of benzene found in the school gardens (2.7 µg/m³) and in the city park (3.8 µg/m³) were higher than those measured outside several schools (2.2 µg/m³) located along different motorway stretches in the Netherlands (44). In the study of Janssen (44), the concentrations of benzene measured outside the schools close to motorways were associated with the intensity of truck traffic but not with that of car traffic. Hence, considering that Sarroch is a small city of 5327 inhabitants without heavy traffic on roads, it would be reasonable to assume that the industrial estate is the major source of air pollution in the Sarroch area.

Oil refineries and power generation plants include industrial processes that produce complex mixtures of PAHs (4,45–47). For instance, 150 metric tons/h of tar residues are gasified from the integrated gasification combined cycle plant (4). Nevertheless, stationary measurements of PAHs in the atmosphere of the residential areas of Sarroch were not available for this study. PAHs emitted from the facilities in the industrial park cause a type of DNA damage very different from those generated by VOCs and heavy metals. Therefore, we decided to evaluate the schoolchildren’s exposure to environmental PAHs using the nuclease P1 modification of the ³²P-postlabelling assay (48), a method that is effective for all types of hydrophobic bulky DNA adducts (48). Our next result indicates that the children resident in the proximity to the industrial park have an excess of bulky DNA adducts in their nasal epithelium compared with those who lived without such environmental contamination, and the multivariate regression analysis showed that this increment was statistically significant (MR = 1.90, 95% CI = 1.25–2.89, P = 0.003). Such DNA adducts are considered to be a biomarker of exposure and lung cancer risk (20–22).

Our findings are in line with previous studies reporting higher amounts of urinary 1-hydroxypyrene (1-OHP), a biomarker of PAH exposure (49), in children living near point industrial sources of PAHs (50–52). The study of Wilhelm (50) showed that children, 5–9 years of age, living close to a coke oven plant in North Duisburg, Germany, had increased levels of 1-OHP and other PAH metabolites in urine. A positive association between external and internal exposure was also reported between B(a)P in ambient air and the urinary levels of 1-OHP in children resident in the Duisburg North hot spot. Lee et al. (51) measured the urinary excretion of 1-OHP in children, 7–15 years of age, resident near a steel mill in the southern part of Korea. The residential proximity to the steel mill was associated with significantly increased levels of 1-OHP. Mucha et al. (52) reported enhanced levels of 1-OHP in urine of children, 2.5–3.5 years of age, living near a steel mill and a coke facility in Mariupol, Ukraine.

Adverse health impacts associated with exposure to emissions from SIE complex among nearby residents are of concern (23,53,54). Previous studies in such areas showed increased risks of respiratory diseases and lung cancer (53). In particular, a standardised mortality rate of 1.24 for lung cancer was found in males resident in the municipality of Sarroch, with a statistically significant departure from the expected value. An important cluster of pleural cancer deaths was identified (54). A decrease in lung function was reported in children living in the surrounding residential area (23). Herein, we observed that schoolchildren resident in the proximity of the industrial estate had increased levels of M₁–dG adducts and bulky DNA adducts in their nasal epithelium, when compared with the rural controls. Our study seems to suggest that the industrial estate might be the causative factor of the observed increment of DNA damage in the children living in the surrounding areas; however, further studies, possibly, with suitable controls from urban areas, are necessary to confirm our findings. Given that higher levels of DNA adducts early in life might be associated with higher risk for adverse health effects later in life, initiatives for reducing the levels of air pollution in Sarroch should be encouraged.

Funding

Associazione Italiana per la Ricerca sul Cancro; United States National Institute of Environmental Health Sciences (P42ES017198) and Municipality of Sarroch, Progetto Sarroch Ambiente e Salute.

Conflict of interest statement: None declared.