Epidemiology and Cancer Prevention
DNA adduct levels and DNA repair polymorphisms in traffic-exposed workers and a general population sample
Version of Record online: 24 JUL 2001
Copyright © 2001 Wiley-Liss, Inc.
International Journal of Cancer
Volume 94, Issue 1, pages 121–127, 1 October 2001
How to Cite
Palli, D., Russo, A., Masala, G., Saieva, C., Guarrera, S., Carturan, S., Munnia, A., Matullo, G. and Peluso, M. (2001), DNA adduct levels and DNA repair polymorphisms in traffic-exposed workers and a general population sample. Int. J. Cancer, 94: 121–127. doi: 10.1002/ijc.1433
- Issue online: 23 AUG 2001
- Version of Record online: 24 JUL 2001
- Manuscript Accepted: 8 MAY 2001
- Manuscript Revised: 17 MAR 2001
- Manuscript Received: 20 DEC 2000
- European Union. Grant Numbers: SI2.146423 (99CVF2-034), SI2.296584/2000CVG2-014
- Associazione Italiana per la Ricerca sul Cancro (AIRC), Milan, Italy
- DNA adducts;
- DNA repair;
Peripheral blood DNA adducts have been considered an acceptable surrogate for target tissues and possibly predictive of cancer risk. A group of 114 workers exposed to traffic pollution and a random sample of 100 residents were drawn from the EPIC cohort in Florence, a population recently shown to present increased DNA adduct levels (Palli et al., Int J Cancer 2000;87:444–51). DNA bulky adducts and 3 DNA repair gene polymorphisms were analyzed in peripheral leukocytes donated at enrollment, by using 32P-postlabeling and PCR methods, respectively. Adduct levels were significantly higher for traffic workers among never smokers (p = 0.03) and light current smokers (p = 0.003). In both groups, urban residents tended to show higher levels than those living in suburban areas, and a seasonal trend emerged with adduct levels being highest in summer and lowest in winter. Traffic workers with at least 1 variant allele for XPD-Lys751Gln polymorphism had significantly higher levels in comparison to workers with 2 common alleles (p = 0.02). A multivariate analysis (after adjustment for age, season, area of residence, smoking, XPD-Lys751Gln genotype and antioxidant intake) showed a significant 2-fold association between occupational exposure and higher levels of adducts (odds ratio 2.1; 95% confidence interval 1.1–4.2), in agreement with recent pooled estimates of increased lung cancer risk for similar job titles. Our results suggest that traffic workers and the general population in Florence are exposed to high levels of genotoxic agents related to vehicle emissions. Photochemical pollution in warmer months might be responsible for the seasonal trend of genotoxic damage in this Mediterranean urbanized area. © 2001 Wiley-Liss, Inc.
The interaction between environmental factors and genetic susceptibility is considered to play a role in most human tumors.1 Recognized sources of environmental pollutants are residential and commercial space heating, motor vehicle exhausts, oil and coal fired power plants and industrial emissions.2 The incomplete combustion of coal and petroleum fuels results in the release of a complex mixture of airborne pyrolysis products, including polycyclic aromatic hydrocarbons (PAH), which are 1 of the major classes of airborne carcinogens capable of forming DNA adducts after metabolic activation by cytochrome P450.3 DNA adducts represent a reliable biological marker widely used to identify health hazards and to evaluate the dose–response relationship in humans exposed to low levels of carcinogens and mutagenic compounds.1 Recently, peripheral blood DNA adducts have been reported to be an acceptable surrogate for lung tissue,4 possibly predictive of lung cancer risk.5, 6 Several studies have reported higher levels of DNA adducts among subjects heavily exposed to air pollutants, such as police officers, bus drivers, news vendors and residents in heavily polluted areas.7–9 A recent meta-analysis of 13 32P-DNA postlabeling studies on occupational cohorts has shown that the association between DNA adducts and air pollution exposure is significant in heavily exposed industrial workers and in less severely exposed urban workers.10
Bulky DNA adducts, such as those induced by PAH, may be repaired by base excision repair (BER) and nucleotide excision repair (NER) pathways.11 The genetic polymorphism of DNA repair enzymes have been reported to influence DNA adduct levels.12 In particular, xeroderma pigmentosum group D (XPD, also known as ERCC2) and x-ray repair cross complementing group 1–3 (XRCC1 and XRCC3) genes have been reported to be polymorphic and might be involved in environmental carcinogenesis.12–18 The XRCC1 protein participates in the BER pathway,19 acting apparently as a scaffold protein, facilitating the repair reaction by binding DNA ligase III at its carboxy and DNA polymerase β to its amino terminus. XRCC3 participates in DNA double-strand break/recombination repair and is a member of an emerging family of Rad-51-related proteins that likely participate in homologous recombination to maintain chromosome stability and repair DNA damage.20XPD is involved in the NER pathway, which recognizes and repairs a wide range of structurally unrelated lesions such as bulky adducts and thymidine dimers.21XPD functions as an ATP-dependent 5′-3′ helicase joint to the basal transcription factor IIH complex.
We have recently shown that dietary habits characterized by frequent consumption of fresh fruit and vegetables and high intake of antioxidants were strongly associated with reduced levels of DNA adducts in a large series of healthy subjects,22 randomly sampled among participants into a prospective study, the Italian section of the collaborative European project known as European Prospective Investigation into Cancer and nutrition (EPIC).23
The aim of our current study was to better evaluate the determinants of DNA adducts in peripheral leukocytes in Florence, the area where the highest adduct levels were found in the EPIC Italy cross-sectional study.22 We have compared the random sample of 100 residents, stratified by age and sex, originally selected from the local EPIC cohort in the previous cross-sectional study,22 with a group of 114 volunteers who, at enrollment, had reported an occupation consistent with a high exposure to traffic pollution. DNA adducts were determined in peripheral white blood cells using the nuclease P1 modification of the 32P-postlabeling assay, a technique mainly effective in detecting bulky hydrophobic DNA adducts, such as those formed by PAH,10, 24 whereas PCR methods were used to define 3 polymorphic DNA repair genes in genomic DNA samples obtained at enrollment. A better understanding of DNA adduct determinants might help to clarify the role of this biomarker in predicting cancer risk at target sites, as a surrogate measure of DNA damage resulting from the interplay of cumulative exposure to environmental genotoxic agents, over a specific time period, and DNA repair.
MATERIAL AND METHODS
EPIC Florence is 1 of the centers participating in the Italian section of the large European project EPIC;23 between January 1993 and March 1998, we enrolled approximately 3,500 men and 10,000 women, aged 35–64 years and residing in the Florence metropolitan area. A group of 100 volunteers (51 men, 49 women) were randomly sampled from this cohort22 and compared with a group of 114 volunteers (81 men, 33 women) who reported themselves as occupied in specific job titles. Briefly, all study participants in this center were originally classified according to International Labor Office (ILO) codes by 1 of us (G.M.), based on a short form collecting information on current occupation. A series of job titles consistent with high professional exposure to vehicle traffic pollution were defined a priori and a group of volunteers of reasonable size was selected among those reporting a job title included in the list.
All participants were classified according to the area of residence at the date of enrollment in the cohort, in 2 broad categories: urban (city of Florence, approximately 70% of the whole cohort) or suburban (in the surrounding area). A food frequency questionnaire and a standardized lifestyle questionnaire were filled by each participant.22
Exposed workers were significantly younger (46.2 vs. 50.2 years), more overweight (body mass index 26.6 vs. 25.0 kg/m2), less educated, more frequently of male sex and were more often residing in the suburban area than in the city of Florence (Table I).
|Characteristics||Random sample subjects||Traffic-exposed workers||p value1||p value2|
|n4||Mean5 ± SE||Median||% of positive6||n4||Mean5 ± SE||Median||% of positive6|
|Males||51||8.9 − 1.3||6.6||82.4||81||13.1 − 1.6||9.7||81.5||0.1|
|Females||49||13.2 − 1.9||9.7||87.8||33||15.3 − 2.2||13.5||90.9||0.003||0.4|
|Age group (years)|
|<44||32||11.7 − 1.7||9.6||87.5||50||13.1 − 1.8||10.8||86.0||0.9|
|45–54||34||11.7 − 2.4||6.9||91.2||45||16.7 − 2.3||13.9||86.7||0.08|
|>54||34||9.6 − 1.7||8.1||76.5||19||8.4 − 2.2||4.4||73.7||0.01||0.6|
|Period of blood drawing|
|Winter||31||9.3 − 2.0||7.1||83.9||40||12.5 − 1.7||10.8||87.5||0.1|
|Midseason||37||10.7 − 2.1||7.3||83.8||43||13.3 − 1.8||11.2||83.7||0.1|
|Summer||32||13.0 − 1.9||12.2||87.5||31||15.9 − 3.4||9.4||80.7||0.7||0.8|
|Year of blood drawing|
|1993–1994||34||13.0 − 1.8||10.5||94.1||35||14.4 − 2.1||12.9||91.4||0.7|
|1995–1996||42||10.4 − 1.9||7.3||85.7||39||12.6 − 2.0||9.4||79.5||0.3|
|1997–1998||24||9.1 − 2.4||2.5||70.8||40||14.2 − 2.6||11.1||82.5||0.2||0.1|
|Never smoker||38||8.7 − 1.1||7.2||84.2||29||14.9 − 2.3||14.1||82.8||0.03|
|Former||34||12.7 − 2.1||10.1||88.2||46||12.1 − 2.2||8.9||84.8||0.5|
|Current||28||11.9 − 2.9||8.3||82.1||39||14.6 − 2.1||10.5||84.6||0.1||0.2|
|Cigarettes per day|
|1–8||7||4.7 − 1.6||4.5||85.7||6||25.2 − 2.9||23.8||100.0||0.003|
|9–18||14||13.1 − 3.1||12.1||92.9||14||10.6 − 3.2||4.2||78.6||0.3|
|19+||7||16.5 − 9.8||3.1||57.1||13||13.1 − 4.0||9.0||76.9||0.4||0.7|
|Urban (Florence)||71||11.8 − 1.5||7.3||85.9||62||15.2 − 1.9||12.6||88.7||0.1|
|Suburban (Province)||29||9.0 − 1.5||8.1||82.8||52||11.9 − 1.6||9.2||78.9||0.01||0.4|
|Total||100||11.0 − 1.1||7.7||85.0||114||13.7 − 1.3||11.1||84.2||0.15|
Blood collection and storage
An informed consent form was signed by all subjects before enrollment. Blood samples were processed by centrifugation in a dedicated laboratory, in the same day of collection and divided into 28 aliquots of 0.5 ml each (12, plasma; 8, serum; 4, concentrated red blood cells; 4, buffy coat). The aliquots have been stored in liquid nitrogen tanks at −196°C. We retrieved 1 buffy coat straw for each subject included in this sample, and all available straws were shipped in dry ice to the study laboratory at the National Cancer Institute (IST) in Genoa for DNA extraction and DNA adduct analyses. DNA sample aliquots were then shipped to Turin University for DNA repair polymorphisms analyses. The exact date of blood drawing was used for analyses by calendar year and by season; 3 periods (winter, midseason and summer) were defined a priori according to earlier reports from Italy.22, 25
Leukocyte DNA was isolated and purified from stored buffy coats by enzymatic digestion of RNA and proteins followed by phenol-chloroform extractions.26 Leukocyte DNA adducts levels were measured using the nuclease P1 modification of the 32P-postlabeling technique, as described previously; the detection limit was 0.1 adduct per 109 normal nucleotides.22 The reproducibility of the 32P-postlabeling technique was verified analyzing approximately 20% of DNA samples with a second independent experiment and the results of the 2 analyses were in perfect agreement (r = 0.98). All the analyses were carried out blindly before decoding. One standard was routinely included in the analysis, that is, benzo(a)pyrene DNA adducts, from liver of mice treated intraperitoneally with 0.06 mg/kg B(a)P for 24 hr.27 The average levels of B(a)P DNA adducts were 51 ± 1.0 (SE) per 109 nucleotides.
PCR followed by enzymatic digestion was used for the genotyping of the XRCC1-Arg399Gln (G28152A, exon 10), XRCC3-Thr241Met (C18067T, exon 7) polymorphisms and XPD-Lys751Gln (A35931C, exon 23).14, 18 All of the PCR reactions were performed in a total reaction volume of 20 μl containing 10 ng of genomic DNA, 0.4 units of Taq polymerase (PE applied biosystems) in PCR buffer 1×, 1.5 mM MgCl2, 50 mM dNTPs and 250 nM of each primer. Thermal cycling conditions were as follows: initial denaturation step at 95°C for 3 min, 35 cycles of PCR consisting of 95°C for 20 sec, 20 sec at the appropriate annealing temperature (55°C, 67°C and 60°C, respectively), 72°C for 20 sec and a final extension step at 72°C for 5 min. The following PCR primers and restriction enzymes were used to detect genotypes as described in Matullo et al.18: XRCC1-Arg399Gln, primer sense 5′-CAAGTACAGCCAGGTCCTAG-3′, antisense 5′-CCTTCCCTCA TCTGGAGTAC-3′ (Nci I, Promega); XRCC3-Thr241Met, primer sense 5′-GCCTGGTGGTCATCGACTC-3′, antisense 5′-ACAGGGCTCTGGAAGGCACTGCTCAGC TCACGCACC-3′(Nco I, Promega); XPD-Lys751Gln, primer sense 5′-CTGCTCAGCCTGGAGCAGCTAGA ATCAGAGGAGACGCTG-3′, antisense 5′-AAGACCTTCTAGCACCACCG-3′ (Pst I, Promega).
Methodologic validation included a comparison between PCR-RFLP, direct sequencing and denaturing high performance liquid chromatography (DHPLC).27 All uncertain PCR-RFLP typings have been reanalyzed with the same technique and usually 1 more test was sufficient to clarify any doubt. Moreover, 10% of the genotypings were randomly reanalyzed. Because of the small amount of DNA available for the subjects studied, we checked the accuracy of PCR-RFLP genotyping for the 3 polymorphisms by using the “primer extension” technique with DHPLC (Transgenomic, Santa Clara, CA) on a set of 50 individuals belonging to a cardiovascular disease study and a complete agreement was obtained. The primer extension technique28 is based on specific incorporation of the complementary dydeoxinucleotide into the base substitution; the rate of right incorporation completely overwhelms dydeoxinucleotide misincorporations and heterozygotes are easily detectable. The following primers flanking the base substitution were used: XRCC1-Arg399Gln, 5′-CGGCGGCTGCCCTCCC-3′; XRCC3-Thr241Met, 5′-GGCATCTGCAGTCCCTGGGGGCCA-3′; XPD-Lys751Gln, 5′- ATCTGCTCTATCCTCT-3′.
To investigate the relationship between DNA adduct levels and individual characteristics, we compared adduct values and the percentage of positive samples for different levels of these variables. Negative samples (below 0.1 adduct per 109 normal nucleotides, the threshold of detection of the 32P-postlabeling method) were assigned a value of 0.1. Mean levels of DNA adducts were compared between the 2 groups (random sample and traffic workers) for selected individual variables. Group comparisons were carried out by means of analysis of covariance to account for several confounders, introducing into each model terms for age, sex, smoking history and period of blood drawing. A multivariate logistic analysis was carried out among subjects with detectable levels of DNA adducts (>0.1 per 109 nucleotides) to quantify the association between the 2 exposure categories (traffic-exposed workers vs. the random sample of volunteers) and a dichotomous category for DNA adduct levels (high/low according to the median value) as the outcome variable, after adjustment for several confounders, including dietary intakes. Box and whisker plots were used to summarize the data for selected variables.
DNA adducts were detected in 85.0% and 84.2% of the samples in the 2 groups, respectively (Table I). Overall, the adjusted mean levels of adducts tended to be higher in the group of traffic-exposed workers than in the random sample subjects (13.7 vs. 11.0 adducts per 109 normal nucleotides, respectively), but did not reach statistically significance (p = 0.10). Exposed workers showed significantly higher levels among subjects classified as never smokers (p = 0.03) and light current smokers (fewer than 9 cigarettes per day reported at interview, p = 0.003); differences were also evident among men, urban residents (Fig. 1) and subjects enrolled in winter and midseason months (p < 0.10). In both groups, a similar seasonal trend emerged with highest levels of DNA adducts in samples collected in summer months and lowest levels in winter (Fig. 2).
Figure 3 shows 4 representative DNA adduct profiles of traffic-exposed workers detected using the nuclease P1 modification of the 32P-postlabeling technique. A stronger diagonal radioactive zone (DRZ) or distinct DNA adduct spots were more frequently detected in the chromatograms of traffic-exposed worker DNA samples collected in summer (Fig. 3a,b), than in those obtained in midseason or winter season (Fig. 3c,d). However, DNA adduct patterns (e.g., DRZ or distinct spots) detected in the chromatograms of smoking and nonsmoking traffic-exposed workers were similar. A faint DRZ or a single adduct spot (mostly situated near the origin) was generally found in adduct-positive samples obtained from residents (data not shown), with a similar seasonal variation.
A moderately decreasing trend over calendar time emerged among the 100 volunteers included in the random sample, but failed to reach statistical significance (p = 0.10); among exposed workers, adduct levels were stable during the 5-year study period.
The group of workers selected for comparison included volunteers reporting different job titles: 41 drivers (30 bus and truck, 11 taxi drivers), 26 garbage collectors, 17 police officers, 17 news and street vendors and 13 other vehicle-related occupations (including mechanics, gas station and car parking attendants). Crude mean levels of DNA adducts in separate categories are shown in Table II, and tended to be higher among garbage collectors and workers more directly exposed to vehicle exhaust. A comparison of the means by covariance analysis, corrected for multiple testing, did not show any difference between mean levels of DNA adducts across these job categories. The combined distribution of individual DNA adduct levels above the 90th percentile in the total series of 181 positive samples, ranked by descending order, is presented in Table III. Among 18 subjects included in this highest decile (higher than the 90th percentile), 5 had been originally drawn in the random sample and 13 reported an occupation consistent with an increased professional exposure. All 5 subjects residing in suburban Florence area included in this highest decile were traffic-exposed workers.
|Job title||DNA adducts||% of positive1||% of never smokers|
|n||Crude mean ± SE||Median|
|Bus and truck drivers||30||11.1–2.2||8.9||73.3||20.0|
|News and street vendors||17||13.9–2.5||13.9||100.0||41.2|
|Car services and others2||13||18.3–4.6||14.8||84.6||23.1|
|Rank #||Occupation||Sex||Age (years)||Smoking history||Area of residence||DNA adducts per 109 nucleotides|
|Traffic-exposed workers||Random sample subjects|
|11||Gas station attendant||M||53||Current||Urban||37.0|
DNA repair polymorphisms
Polymorphic alleles of 2 DNA repair genes (XRCC1-Arg399Gln and XRCC3-Thr243Met) were not significantly related to DNA adduct levels, although subjects with at least an XRCC1-399 Gln or an XRCC3-243 Met variant allele tended to show higher levels within each study group. However, the relatively large group of exposed workers with at least a Gln variant allele for XPD-751 polymorphism (71 of 114 subjects, 62.3%) showed significantly increased mean levels of DNA adducts (Table IV) in comparison to the other 41 workers with no polymorphic allele (14.8 vs. 10.9 DNA adducts per 109 normal nucleotides, p = 0.02); a similar difference was suggested when the former group was compared with the 69 subjects with the same variant genotype but no professional exposure (14.8 vs. 11.5 DNA adducts per 109 normal nucleotides, p = 0.07).
|DNA repair genotypes||Random sample subjects||Traffic-exposed workers||p value1||p value2|
|n4||Mean5 ± SE||% positive6||n4||Mean5 ± SE||% positive6|
|Arg/Gln or Gln/Gln||54||13.1–1.9||85.2||58||15.1–1.9||86.2||0.6||0.4|
|Thr/Met or Met/Met||62||12.4–1.7||87.1||74||13.6–1.6||84.2||0.6||0.5|
|Lys/Gln or Gln/Gln||69||11.5–1.5||85.5||71||14.8–1.5||87.3||0.3||0.07|
Finally, all study subjects with detectable levels of DNA adducts (85 volunteers and 96 workers) were classified as high/low according to the median level of DNA adducts in the combined series (11.2 DNA adducts per 109 normal nucleotides), and a multivariate logistic analysis showed that exposed workers were at significantly higher risk of presenting a high level of the study biomarker (odds ratio 2.1, 95% confidence interval 1.1–4.2; Table V). A 2-fold increased risk of higher DNA adduct levels was also shown by samples collected in summer, but failed to reach statistical significance. A tendency of residents in the suburban area to show lower levels was confirmed by this multivariate analysis, although with wide confidence intervals.
|DNA adduct level1||OR2||95% CI|
|Low n (%)||High n (%)|
|Period of blood drawing|
|Winter||35 (38.4%)||26 (28.9%)||1#||—|
|Midseason||33 (36.3%)||34 (37.8%)||1.40||0.66–2.97|
|Summer||23 (25.3%)||30 (33.3%)||2.12||0.93–4.86|
|Area of residence|
|Urban (City of Florence)||56 (61.5%)||60 (66.7%)||1#||—|
|Suburban (Province)||35 (38.5%)||30 (33.3%)||0.72||0.35–1.49|
|Random sample||47 (51.6%)||38 (42.2%)||1#||—|
|Traffic workers||44 (48.4%)||52 (57.8%)||2.11||1.06–4.21|
A multivariate analysis, comparing a group of workers exposed to traffic pollution and a group of local residents, all identified among EPIC volunteers enrolled in the metropolitan area of Florence over a 5-year period, showed that higher levels of DNA adducts were significantly related to occupational exposure. Differences were particularly evident among never and light current smokers. Higher levels tended also to be related to a polymorphism in XPD, a DNA repair gene involved in the NER pathway, particularly among exposed workers. These results, overall, suggest that the main determinants of DNA adducts in residents of the Florence area are related to vehicle traffic exhaust and, possibly, to photochemical pollution in warmer months. The photodegradation and chemical modification of urban traffic PAH emissions occurring in the atmosphere, particularly in summer, might induce the formation of highly reactive products, such as B(a)P and B(a)P-7,8-dihydrodiol-9,10-epoxide quinones, capable of forming DNA adducts. This hypothesis is supported from the detection of more complex aromatic DNA adduct patterns in the chromatograms of summer samples (Fig. 3a,b) of traffic-exposed workers, compared to samples collected in winter (Fig. 3c,d).
In addition, these results confirm that the exposure to genotoxic agents, derived directly or indirectly from compounds present in vehicle exhaust, is not restricted to selected groups of workers heavily exposed during the daily work shift (with a cumulative exposure of approximately 30–40 hr per week). The general population of urbanized areas (including the workers themselves) is exposed to the same air pollutants, although at lower levels, 24 hr per day. Individual levels of DNA adducts tend to reflect, overall, the interplay of cumulative exposure to these agents and DNA repair, over a time period before the date on which samples were obtained. This specific time window has yet to be clarified.
Our current results suggest that geographic differences across EPIC Italy22 were mostly a result of differences in the cumulative exposure to genotoxic agents derived from traffic pollution. Because of the EPIC study design (a prospective follow-up of a large series of apparently healthy subjects, enrolled over a 5-year period), no measurement of the exposure to air pollutants was available at the individual level, in any time (days or weeks) immediately before the drawing of the blood samples. Air pollution in the Florence area, however, is monitored regularly and the seasonal trend in both study groups is consistent with a similar trend shown by ozone concentration.30 Ozone levels are particularly high in this area because of local geographic and meteorologic characteristics: Florence is surrounded by hills, except on the western side; winds are usually weak; and, during summer, average temperatures tend to be high and rainfall is modest. The higher levels of DNA adducts that tended to be evident among residents of the urban area in comparison to suburban residents are in agreement with differences in environmental measurements of air pollutants (including PAHs).
The relatively small sample size of subgroups of traffic-exposed workers probably explains the lack of statistical significance of comparisons by specific job titles. Garbage collectors emerged as a group with very high levels of DNA adducts, probably because of the physical activity required by most operations (hauling containers, sweeping), combined with direct exposure to vehicle exhaust. Overall, occupational exposures to motor vehicle exhausts, including diesel engine fumes, have been related with increased lung cancer risk in bus and truck drivers and other traffic-exposed workers,31, 32 with 2 recent meta-analyses reporting pooled estimates around 1.5.33, 34
Our results suggest that XPD, a polymorphic gene involved in nucleotide excision DNA repair,14 might play an important role in repairing bulky DNA adducts in human leukocytes, particularly those induced from traffic environmental carcinogens, in keeping with a previous study.11 A similar, although not statistically significant, effect was possibly suggested also for XRCC1, in agreement with the studies by Lunn et al.12 and Duell et al.16 A still weaker effect was shown for XRCC3-241 polymorphism for which, however, evidence of its possible involvement in DNA-adducts repair has been provided.18 Approximately two-thirds of our study subjects had at least a Gln variant at XPD-751 polymorphic locus, and tended to show higher levels of DNA adducts than subjects with 2 normal alleles. This difference was evident among traffic-exposed workers, suggesting a role for this polymorphism at higher levels of exposure. At the same time, differences in DNA adduct levels between the 2 study groups (exposed workers and volunteers included in the random sample) tended to be restricted to subjects with at least a variant allele. Our results, apparently contrast with a small study by Lunn et al.,15 whereas Sanford et al.35 had not found a higher chromatid aberration frequency in cells containing a rare XPD mutation. Amino acid variants in different domains of XPD may affect different protein interactions, resulting in the expression of different phenotypes.36 We studied the same XPD-Lys751Gln polymorphism: this might have divergent effects in different DNA repair pathways, or other types of DNA damage, when compared to hydrophobic bulky DNA adducts we have measured. Another recent study, of relatively small size, found no effect of XPD, but some association between XRCC1 genotype and polyphenol DNA adducts.16 Nevertheless, all these results must be confirmed on larger samples and on populations of different origins; actually in our sample from central Italy, the Gln751 allele has a frequency similar to the study carried out by Duell et al.16 in the United States (0.42 vs. 0.39), but higher than 2 other American studies (average frequency 0.27).14, 15 This discrepancy may be because the latter study groups were much smaller (12 and 31 subjects, respectively) in comparison to our and Duell's study (214 and 76 individuals, respectively).
Overall, our previous and current findings indicate that DNA adduct risk, and possibly cancer risk, attributable to air pollution exposure might be modulated from other factors, including diet,22 individual genetic susceptibility, residential and occupational history in downtown or suburban areas, synergistic effects of smoking and formation of high electrophilic photodegradation products of vehicle emissions in summer time. Further studies are needed to investigate if there is any direct correlation between individual levels of DNA adducts and the average concentration of ozone and related air pollutants in the EPIC Florence study area in the period immediately prior to blood drawing.
We thank all study participants for their cooperation, all collaborators of EPIC Italy Study Group (in particular P. Vineis, Turin, and V. Krogh, Milan), Dr. L. Sommani (ASL10, Florence) and Dr. D. Grechi (ARPAT, Florence) for helpful comments, and Chiara Zappitello for editorial assistance. EPIC is coordinated at the international level by E. Riboli (IARC, Lyon).
- 2Cancers of the lung and pleura. In: SchottenfeldD, FraumeniJF, Jr., eds. Cancer epidemiology and prevention. New York: Oxford University Press, 1996. 637–65., .
- 3IARC. Monographs on the evaluation of carcinogenic risks to humans. Polynuclear aromatic compounds. Part 1. Chemical, environmental and experimental data. vol. 32. Lyon: IARC, 1984.
- 30Agenzia Regionale per la Protezione Ambientale della Toscana (ARPAT). Urban mobility and environmental assessment: the Florence Town Council abatement and protection policy. Florence, Italy: ARPAT, 1999. 1–63.