(i) In vivo studies

Animals

The Working Group did not identify any publications that assessed the induction of mutations in experimental animals (e.g. rodents) experimentally exposed to outdoor air or samples derived from outdoor air. However, as summarized in Supplemental Table S24 (available online), Yauk & Quinn (1996) and Yauk et al. (2000) used multilocus DNA fingerprinting and pedigree analyses to assess the frequency of heritable genetic minisatellite mutations in herring gulls (Larus argentatus) collected from several locations affected by urban and/or industrial activities. The results obtained showed a significant > 2-fold increase in mutation rate at industrial sites compared with rural controls and, moreover, decreasing mutation rates with increasing distance from the highlighted industrial sources.

The follow-up studies used modified animal enclosures to expose mice to outdoor air in situ at selected urban/industrial locations (Somers et al., 2002, 2004; Yauk et al., 2008). Results were compared with matched exposures at rural control sites. This strategy offers the distinct advantage of real outdoor air exposures under semi-controlled conditions (e.g. animal housing and food); however, it is generally not possible to reliably estimate the actual dose (i.e. milligrams of PM or cubic metres of air per kilogram body weight [bw] per day). These studies have shown an increase in induced heritable mutations at expanded simple tandem repeat (ESTR) loci. High-efficiency particulate air (HEPA) filtration of outdoor air reduced heritable mutation rates (Somers et al., 2004; Yauk et al., 2008) (see also Section 4.3.3b for genotoxic effects on germ cells).

Plants

Studies in plants (see Supplemental Table S1, available online) assessed the ability of outdoor air or samples derived from outdoor air to induce genetic mutations or chromosomal damage. For instance, studies by Ferreira et al. (2000, 2003, 2007) reported the mutagenic activity of outdoor air at selected locations in or near the São Paulo (Brazil) metropolitan area.

(ii) In vitro studies

Table 4.1 provides a summary of studies that have used in vitro assays to assess the ability of outdoor air or samples derived from outdoor air to induce genetic mutations or gene conversion.

Table 4.1

Genetic mutations associated with outdoor air pollution in human or animal cells in vitro.

Human cells

Eight studies used human h1A1v2 cells to assess the induction of mutations at the thymidine kinase TK^+/− locus (Hannigan et al., 1997, 1998, 2005; Durant et al., 1998; Pedersen et al., 1999, 2004, 2005; Adonis & Gil, 2000).

The results showed that the mutagenic potency of PM extracts for urban sites expressed per unit of equivalent organic carbon (EOC) are generally less than 2-fold greater than values obtained for rural (control) sites (Hannigan et al., 1997, 2005; Pedersen et al., 1999, 2004, 2005). However, when expressed per equivalent cubic metre of outdoor air, the potency values for urban sites were 3–10-fold higher than those for rural sites.

Detailed source apportionment revealed that diesel and natural gas combustion emissions make a large contribution to the mutagenic activity of outdoor air PM (Hannigan et al., 1997, 2005). In addition, detailed extract fractionation revealed that polar, semipolar (e.g. nitro-PAHs, ketones, and quinones), and non-polar (e.g. PAHs) extract fractions can each account for a substantial portion of the observed mutagenic activity per unit of EOC. For example, in their analysis of organic extracts of standard reference mixture (SRM) 1649, Durant et al. (1998) noted that semipolar and non-polar extract fractions accounted for 70% of the mutagenic activity. In their analyses of samples collected in southern California, Hannigan et al. (1998) noted that aromatic substances can account for more than 50% of extract mutagenic activity. In their analyses of sites in the north-eastern USA, Pedersen et al. (2004) noted that moderate-molecular-weight PAHs can account for 4–38% of observed mutagenic activity and that polar compounds (e.g. organic acids and hydroxy-polycyclic aromatic compounds [hydroxy-PACs]) can account for 14–32% of the observed mutagenic activity. Pedersen et al. (1999, 2005) noted that mutagenic activity per cubic metre is approximately 2-fold higher in the winter relative to the summer across all sites investigated.

Adonis & Gil (2000) used the TK^+/− locus mutation assay in h1A1v2 cells to assess the mutagenic activity of an organic extract of PM collected from an urban, heavy-traffic site in Santiago, Chile. The reported mutagenic potency (expressed per cubic metre equivalent per assay millilitre) is more than 400-fold greater than that observed in Los Angeles (i.e. Hannigan et al., 1997, 1998, 2005). Comparisons based on potency expressed per unit of EOC revealed that extracts of PM from Santiago are approximately 2-fold more potent than extracts of PM from Los Angeles.

Animal cells

Several studies have demonstrated that organic extracts of outdoor air PM can induce significant dose-dependent increases in mutant frequency in cultured non-human mammalian cells. For instance, studies by Seemayer et al. (1987a, 1988) revealed that extracts of PM collected from the industrialized Rhine–Ruhr region in Germany induced a significant increase in Hprt mutant frequency. Similarly, Courtois et al. (1988) noted that organic extracts of PM collected from an urban location in Paris, France, induced a significant dose-dependent increase in Hprt mutant frequency. Source apportionment revealed noteworthy contributions by natural gas and diesel combustion emissions, and extract fractionation revealed substantial contributions from each of several chemical classes, including non-polar compounds (e.g. PAHs), semipolar compounds (e.g. nitro-PAHs and quinones), and polar compounds (e.g. organic acids and hydroxy-PACs).

Identification of putative mutagens

Bagley et al. (1992) and Gundel et al. (1993) noted substantial declines in mutagenic activity in TA98 strains deficient in nitroreductase (NR) (TA98NR). The results indicated a strong involvement of nitro-PAHs, in particular compounds such as the dinitropyrenes.

Many studies of organic extracts of PM from urban areas compared mutagenic potencies in the absence of S9 between TA98 and TA98NR to infer the involvement of nitroarenes in the observed responses. For example, analyses of extracts of PM collected in France, Japan, the Netherlands, and Sweden recorded a marked reduction in mutagenic activity in TA98NR without S9 relative to TA98 (Festy et al., 1984; de Raat & de Meijere, 1988; Takagi et al., 1992; Strandell et al., 1994).

A wide range of studies used the NR-deficient and O-acetyltransferase (OAT)-deficient versions of TA98 to examine the influence of nitroarenes in determining the mutagenic activity of organic PM extracts. These studies collectively examined extracts of PM collected from a wide variety of urban and/or industrial locations in Chile, Germany, Italy, Japan, Mexico, Norway, Spain, Sweden, and the USA (Alfheim et al., 1983; Tokiwa et al., 1983; Löfroth et al., 1985; Moriske et al., 1985; Wolff et al., 1986; Crebelli, 1989; De Flora et al., 1989; Takagi et al., 1992; Adonis & Gil, 1993, Espinosa-Aguirre et al., 1993; Gundel et al., 1993; Strandell et al., 1994; Casellas et al., 1995; Sato et al., 1995). Some of these studies noted dramatic reductions in mutagenic activity in the NR-deficient strain TA98NR relative to TA98. Some studies have noted that the relative decline in mutagenic potency of air samples in strain TA98NR relative to TA98 is seasonally variable (Erdinger et al., 2005).

Similarly, numerous studies have used metabolically enhanced Salmonella strains such as YG1021 (NR-enhanced) and YG1024 (OAT-enhanced) for comparative assessment of PM extracts. For example, studies of extracts of PM collected from urban and/or industrial areas in Brazil, Italy, Japan, Mexico, and Poland noted substantial increases in mutagenic activity in strains YG1021 and/or YG1024 without S9 (Espinosa-Aguirre et al., 1993; Yamaguchi et al., 1994; Jadczyk & Kucharczyk, 2005; Pereira et al., 2010; Traversi et al., 2011; Lemos et al., 2012). These studies frequently found that OAT enhancement contributed to larger relative increases in potency (without S9) compared with NR enhancement. For example, Yamaguchi et al. (1994) observed 3–4-fold increases in potency in YG1021 and 5–10-fold increases in potency in YG1024 for extracts of PM collected from a high-traffic site in Kobe, Japan.

In addition, numerous studies have used the TA98-derived frameshift strain YG1041, which overexpresses both NR and OAT, for analyses of extracts of PM from diverse regions (e.g. Brazil, the Czech Republic, Denmark, and Poland) (Binková et al., 2003; Sharma et al., 2007; Umbuzeiro et al., 2008; Piekarska et al., 2009). For example, Umbuzeiro et al. (2008) reported dramatic (> 30-fold) increases in the mutagenic activity in YG1041 relative to TA98 without S9 of organic extracts of PM collected from urban locations in São Paulo, Brazil. Extract fractionation confirmed that the highest activity in YG1041 without S9 was associated with nitro-PAHs. This study also indicated more modest, but nevertheless substantial, increases in mutagenic activity in OAT-enhanced strains in the presence of S9. This pattern of activity is thought to be associated with aromatic amines, including N-heterocyclics, as well as polar mutagens that, to date, have not been well characterized (e.g. oxy-PAHs) (Umbuzeiro et al., 2008).

Mutagenic potential and particle size

Studies from several geographical regions reported that the mutagenic activity of PM_2.5 extracts was greater per unit of extractable organic matter (EOM) or per milligram of PM than that of PM₁₀.

Several studies have documented a noteworthy increase in potency with decreasing particle size. For instance, Kawanaka et al. (2004) found that approximately 90% of the mutagenic activity associated with extracts of PM collected in the Tokyo, Japan, area was associated with fine particles. In addition, studies that examined extracts of PM from Germany (Massolo et al., 2002) and Italy (Pagano et al., 1996; Monarca et al., 1997) specifically reported enhanced potency for the fine (< 1.5 µm) and/or ultrafine (< 0.5 µm) PM fractions. Nonetheless, small PM size fractions make relatively small contributions to atmospheric mutagenic activity per cubic metre due to low atmospheric levels (i.e. mass concentration) of fine and ultrafine PM per cubic metre.

Mutagenic potential of semivolatile organic compounds

In addition to detailed analyses of PM extracts, several studies also evaluated extracted SVOCs. For example, the study by Ciganek et al. (2004), which examined both PM and extracts from urban sites in Brno, Czech Republic, noted that extracts accounted for 15–40% of the total mutagenic activity.

Several studies have indicated that the similarities and differences between SVOC concentrates and PM extracts are affected by season, ambient temperature, adsorbent type, and the presence of S9. For example, in their study of samples collected in the Flanders region of Belgium, Du Four et al. (2004) observed that polyurethane foam extracts were more potent per microgram of EOM in the summer, whereas PM extracts were more potent in the winter. In contrast, Tuominen et al. (1988) noted that XAD extracts from samples collected in Helsinki, Finland, during the winter were generally more potent than PM extracts, with no appreciable difference between XAD and PM extracts from samples collected during the summer.

Spatial and temporal patterns in atmospheric mutagenic activity

Many studies reported markedly higher mutagenic potencies per cubic metre of air for extracts of PM collected during the colder months (winter and autumn) compared with those from the warmer months (spring and summer). Numerous studies conducted in Asia compared extracts of PM samples collected during different seasons and observed that winter and/or autumn samples were markedly more mutagenic than samples collected in spring and/or summer (e.g. Goto et al., 1982; Shimizu et al., 1982; Takagi et al., 1992; Qian et al., 1997; Qian & Zhang, 1997; Vinitketkumnuen et al., 2002). Similarly, studies conducted in European countries found comparable elevations in potency during colder months (Møller & Alfheim, 1980; Wullenweber et al., 1982; Alfheim et al., 1983; Athanasiou et al., 1986; Morozzi et al., 1992; Crebelli et al., 1995; Černá et al., 1999; Binková et al., 2003; Du Four et al., 2004; Piekarska et al., 2009, 2011). Finally, several studies conducted in North America, South America, and New Zealand also indicated elevations in mutagenic potency during colder months (Crebelli, 1989; Daisey et al., 1980; Brown et al., 2005; Cavanagh et al., 2009; Müller et al., 2001; Török et al., 1989).

Nevertheless, some studies from diverse geographical regions failed to detect any appreciable seasonal trend in mutagenic potency or noted that the mutagenic potency levels of extracts of PM collected in the summer were higher relative to those of winter samples (Commoner et al., 1978; Ohtani et al., 1985; Athanasiou et al., 1987; Adonis & Gil, 1993; Greenberg et al., 1993; Kuo et al., 1998).

The trend towards an increased atmospheric burden of PM-associated mutagens during colder months is quite clear and well substantiated. Several studies have found that the root causes of the observed seasonal trends may not be evident. Although some authors have pointed towards contributions from fuel oil combustion for residential heating during winter months (Daisey et al., 1980), others have reported that the presence of atmospheric oxidants and the atmospheric transformation of nitroarenes are important determinants of seasonal fluctuation in PM mutagenic activity (Arey et al., 1988). Villalobos-Pietrini et al. (2006) noted the importance of a ground-level temperature inversion. Finally, other studies (Festy, 1980; Festy et al., 1984) support the contention that mutagens associated with winter PM are chemically different from mutagens associated with PM emitted during warmer months.

Day-to-day and diurnal variability

Marked day-to-day variability in mutagenic potency has been reported. For instance, a study in Sagamihara, Japan (Takagi et al., 1992), found lower potency on Sundays and holidays and concluded that vehicular emissions were significant contributors to the mutagenic activity of atmospheric PM. Other studies noted differences in mutagenic activity during the day compared with evenings (Møller et al., 1982; Gupta et al., 1996). A study by Kameda et al. (2004) of PM extracts collected in Osaka, Japan, reported peaks in potency in the early morning and late evening; potency corresponded with peaks in atmospheric levels of nitrogen oxide (NO), carbon monoxide (CO), and 1-nitropyrene. Some studies indicated that diurnal patterns varied with the season. For example, Shimizu et al. (1982) observed that daytime PM potency per cubic metre for samples collected in the centre of Tokyo, Japan, exceeded night-time PM potency for winter samples only. Conversely, Sakitani & Hayashi (1986) found that daytime potency exceeded nighttime potency for summer and autumn samples only.

Temporal trends

Two studies investigated temporal trends in PM-associated mutagenicity across an extended period of time. Matsumoto et al. (1998) monitored PM-associated mutagenic activity in Sapporo, Japan, between 1974 and 1992 and indicated a modest 44–50% temporal decline in mutagenic activity with exogenous metabolic activation. This corresponded with a marked 75–80% temporal decline in benzo[a]pyrene (B[a]P) adsorbed to PM (nanograms per cubic metre). Mutagenic activity without exogenous metabolic activation did not change over time. Similarly, Poli et al. (1999) examined PM-associated mutagenic activity in Parma, Italy, between 1991 and 1998 and reported a marked decline in mutagenic activity between 1992 and 1998.

Spatial variability

Many studies examined the spatial variability in the atmospheric burden of PM-associated mutagenic activity. The most common comparisons concern site-specific conditions related to urbanization, industrial activities, and/or traffic density. For example, many studies conducted in a wide range of locations (Brazil, the Czech Republic, Germany, Greece, Italy, Japan, the Netherlands, Poland, Saudi Arabia, Taiwan [China], Thailand, and the USA) have observed that the mutagenic potency per cubic metre of extracts of PM collected from urban sites near roadways and/or sites described as high-traffic sites is markedly higher relative to more rural reference sites (Preidecker, 1980; Athanasiou et al., 1986; de Raat & de Meijere, 1988; Yu et al., 1989; Wei et al., 1991; Vellosi et al., 1994; Sato et al., 1995; Černá et al., 1999; Vinitketkumnuen et al., 2002; Erdinger et al., 2005; Elassouli et al., 2007; Piekarska et al., 2011).

Elevated mutagenic activity per cubic metre was reported at residential areas located downwind of urban/industrial locations (e.g. de Raat, 1983). Some studies, including those conducted in China (Kong et al., 1994; Zhao et al., 1996), the Netherlands, (van Houdt et al., 1987), and Chile (Gil et al., 1997), observed substantial levels of PM-associated mutagenic activity per cubic metre at control sites, such as a suburban park.

Few studies have compared the mutagenic activity per cubic metre of PM extracts collected from different elevations. Comparisons of extracts of PM collected from ground level with samples collected from the same location at an elevated site, such as a rooftop, reported reduced potency at the site with higher elevation (e.g. Alfheim et al., 1983).

Effect of combustion

Viau et al. (1982) found an increase in the PM-associated mutagenic activity with S9 activation during “smoky” conditions caused by a forest fire in Kentucky, USA. de Andrade et al. (2011) reported that increased levels of PM-associated mutagenic activity were associated with cane-burning activities near São Paulo, Brazil. Similarly, al-Khodairy et al. (1998) noted that increased levels of mutagenic activity were associated with oil well fires in Kuwait. Nevertheless, several studies that examined levels of PM-associated mutagenic activity close to a suspected source (e.g. a municipal waste incinerator or an aluminium smelting operation) were unable to detect any appreciable influence of the source (e.g. Alfheim et al., 1984; Watts et al., 1989).

Several studies conducted fairly rigorous source apportionment and found that a substantial portion of PM-associated atmospheric mutagenicity is from mobile-source emissions (e.g. Israël & Busing, 1983; Lee et al., 1994; Hannigan et al., 2005). Other studies highlighted emissions from wood smoke as more important than mobile-source emissions (e.g. Claxton et al., 2001). Daisey et al. (1980) indicated that in their study in New York City, 50% of PM-associated atmospheric mutagenic activity was from fuel oil combustion for residential heating.

Effect of atmospheric pollutants

Numerous studies from diverse geographical regions have reported positive associations with atmospheric pollutants. These include lead, CO, nitrogen oxides (NO_x, NO and NO₂), SO₂, PAHs, and non-methane hydrocarbons. Associations have been found between PM-associated mutagenic activities and atmospheric lead, a pollutant associated with metal refining and founding and municipal waste incineration, or, for studies conducted before the mid-1990s (UNEP, 1999), with gasoline engine emissions (Flessel et al., 1985; Pitts et al., 1985). Significant associations between PM-associated mutagenicity and NO_x, an indicator of mobile-source emissions, have also been reported (Morris et al., 1995). Several studies have also highlighted associations with SO₂, an atmospheric pollutant associated with combustion of coal, residential fuel oil, and heavy fuel oils, such as marine fuel oil (Israël & Busing, 1983; Wolff et al., 1986; Morris et al., 1995). Finally, numerous studies have documented associations between atmospheric mutagenic activity and levels of atmospheric PAHs, including several known mutagens and/or mutagenic carcinogens (Viras et al., 1990; Černá et al., 1999).

Effect of meteorological conditions

Several studies have observed that meteorological conditions, such as wind speed and direction, temperature, precipitation, humidity, and solar penetration, can influence the levels of PM-associated atmospheric mutagenicity.

Several studies reported that PM-associated atmospheric mutagenicity per cubic metre was negatively affected by precipitation (rain or snow).

Because moving air masses can contain urban/industrial combustion emissions, it is perhaps not surprising that several studies have highlighted the role of wind speed and direction in determining levels of PM-associated atmospheric mutagenicity per cubic metre (e.g. Commoner et al., 1978; Wang et al., 1980). Detailed analyses (e.g. as conducted by Alink et al., 1983; de Raat et al., 1985; de Raat & de Meijere, 1988, and Morris et al., 1995) specifically identified wind directions that were associated with increased levels of atmospheric mutagenicity. For example, studies conducted in the Netherlands found increased levels of PM-associated mutagenicity for easterly or southerly winds, from Germany and Belgium.

Post-emission formation of potent mutagens: atmospheric reactions

Studies that investigated the effects of meteorological conditions on atmospheric mutagenic activity are congruent with those of Arey et al. (1988, 1992) regarding the post-emission formation of potent mutagens derived from combustion emissions. Arey et al. showed that atmospheric reactions can contribute to the formation of potent nitro-PAHs. Such observations are consistent with mutagen formation during airborne movement from an urban/industrial area to a less-congested area, and with some atmospheric transformation products being mutagenic.

Conclusions

In summary, several fundamental conclusions can be drawn from the more than 250 studies that used the Salmonella reverse mutation assay to examine samples derived from outdoor air (e.g. PM extracts) collected from locations on five continents over the past 30 years.

1. Outdoor air PM extracts, including samples of total suspended particles (TSP), PM₁₀, and PM_2.5, evaluated for mutagenicity yielded a significant positive response; however, the mutagenic potency values, expressed per cubic metre, ranged over 5 orders of magnitude. Thus, some atmospheric samples clearly contain very little mutagenic activity, whereas others carry a high burden of activity.
2. The mutagenic potency of outdoor air is positively associated with outdoor air PM levels, reflecting an overall correspondence between declines in air quality and increased levels of PM-associated mutagenic activity. Increased PM-associated mutagenic activity is positively associated with other measures of impaired air quality, such as increased NO_x, lead, PAHs, CO, nitro-PAHs, and SO₂.
3. Samples derived from outdoor air PM collected during colder seasons (winter) are generally more mutagenic (per cubic metre) than those from PM collected during warm seasons (summer). This is likely due to a combination of factors that include changes in source contributions, meteorological changes, and seasonal land-use changes. Concomitantly, studies find that the PM-associated mutagenic potency per cubic metre of outdoor air is inversely related with air temperature.
4. Bioassay-directed fractionation studies confirm that much of the mutagenic activity associated with the particulate portion of outdoor air is found in the moderately polar and/or polar organic fractions, and includes a wide range of acids, bases, and neutral compounds, confirming a significant role for numerous classes of organic compounds. Although several noteworthy mutagens associated with outdoor air PM have been identified (e.g. nitro-PAHs), in most cases the putative mutagens have not been well characterized.
5. Samples derived from outdoor air PM are generally more mutagenic (per cubic metre) during workdays relative to non-workdays (e.g. weekends) and are generally more mutagenic during daytime than at night. The higher mutagenic potencies during workdays and daytime are associated with higher outdoor concentrations of lead, CO, and NO_x, reflective of mobile-source emissions.
6. Studies conducted over 7 years in Parma, Italy, showed a decline (63–76%) in outdoor air mutagenic activity (per cubic metre), reflecting the potential benefits of increasingly stringent emission controls for mobile combustion sources (Poli et al., 1999). A similar study conducted over 18 years in Sapporo, Japan, noted a modest (44–50%) decline only for mutagenic activity commonly associated with PAHs (i.e. not nitro-PAHs) (Matsumoto et al., 1998).
7. Samples derived from outdoor air PM are generally more mutagenic if sampled at ground level, relative to higher elevations. Mutagenic activity is also greatly affected by wind direction and other meteorological conditions. Precipitation events reduce the levels of PM-associated mutagenic activity (per cubic metre).
8. Smaller particles (PM_2.5) are generally more mutagenic per mass of particle than larger particles (PM₁₀); maximum mutagenic activity is associated with particles of 0.1–1.2 µm.
9. The main contributors to PM-associated outdoor air mutagenicity appear to be urbanization, industrial activity, and traffic density. In some cases, wood smoke and/or emissions associated with residential heating have been highlighted as important sources of PM-associated mutagens, and these latter sources can exceed contributions from mobile-source emissions. Many studies have associated outdoor air mutagenicity with SO₂, which is associated with the combustion of coal, residential fuel oil, and heavy fuel oils, including marine fuel oil.

(i) Chromosomal aberrations

Table 4.2 summarizes the studies in which CAs were evaluated as a biomarker of outdoor air exposure versus subjects who worked or spent the majority of their time indoors. The studies reviewed cover several categories of workers exposed to outdoor air, and nearly all showed an association between CAs and this exposure, although this was the case for only two exposure groups when the data were stratified by genotype/phenotype (Knudsen et al., 1999). Four studies included exposure assessments, and all of them found higher levels of exposure among the subjects exposed to outdoor air than among the controls (Burgaz et al., 2002; Cavallo et al., 2006; Srám et al., 2007; Zidzik et al., 2007).

Table 4.2

Chromosomal aberrations in peripheral blood lymphocytes of humans exposed to polluted outdoor air.

Among the studies included in Table 4.2, three studies (Anwar & Kamal, 1988; Knudsen et al., 1999; Cavallo et al., 2006) cultured the lymphocytes for 48 hours; however, three studies (Burgaz et al., 2002; Beskid et al., 2007; Sree Devi et al., 2009) cultured the cells for 72 hours, and one study for 69 hours (Chandrasekaran et al., 1996). Culturing cells for more than 48 hours can result in increased frequencies of CAs formed during the extended period of growth in culture. However, only one study with long culturing times (Sree Devi et al., 2009) appears to have an elevated frequency of CAs among the controls. Nonetheless, all of the studies in Table 4.2 reported significantly higher frequencies of CAs among the exposed relative to control populations.

Ten studies found increased frequencies of CAs among traffic police compared with their respective control populations (Table 4.2). Thus, traffic police in Cairo, Egypt, had higher frequencies of CAs compared with police trainers (Anwar & Kamal, 1988), as did those in Ankara, Turkey, compared with office workers (Burgaz et al., 2002), as did traffic police in Hyderabad, India, compared with subjects who did not work in traffic (Sree Devi et al., 2009), as did traffic police in Hebei City, Henan, China, compared with police who worked in offices (Chen et al., 1999). The study in Turkey included an exposure assessment, which found higher concentrations of urinary 1-OHP among the traffic police relative to the control population (Burgaz et al., 2002). The study by Beskid et al. (2007) showed that police officers who worked outdoors in Prague (Czech Republic), Košice (Slovakia), or Sofia (Bulgaria) had higher frequencies of CAs as determined by fluorescence in situ hybridization (FISH) compared with subjects who were indoors at least 90% of the time. However, when traditional cytogenetic analyses were used, only the Sofia, Bulgaria, traffic police had elevated CA frequencies (Zidzik et al., 2007). An exposure assessment of these three sets of populations showed that the police who worked outdoors had higher exposures to carcinogenic PAHs compared with the controls (Zidzik et al., 2007). Using FISH, Srám et al. (2007) showed that traffic police in Prague, Czech Republic, had higher CA frequencies in January, when air pollution (PM₁₀) was high, than in March, when the PM₁₀ concentration was significantly lower. However, no differences in CA frequencies were found when traditional cytogenetic methods were used.

One study performed in the Czech Republic (Rubes et al., 2005) evaluated the association between exposure of men to polluted outdoor air and CAs in their sperm, and found no association between aneuploidy in the sperm and outdoor air pollution. A study in traffic police in Prague, Czech Republic, showed that the police had higher frequencies of CAs in sperm when sampled in January, when the PM₁₀ concentrations were high, than in March, when the PM₁₀ concentrations were low (Srám et al., 1999) (see Section 4.3.3a).

The frequencies of CAs were higher in taxi drivers in Ankara, Turkey, compared with office workers (Burgaz et al., 2002; Table 4.2), and the taxi drivers had higher concentrations of urinary 1-OHP compared with the control subjects.

An investigation of outdoor airport workers at the international airport in Rome, Italy, found higher frequencies of CAs in this population compared with the frequencies found among airport office workers (Cavallo et al., 2006; Table 4.2). Exposure assessments found higher concentrations of PAHs in the air outdoors compared with in the offices, but there was no difference in the concentrations of urinary 1-OHP among the exposed and control groups.

In Denmark, a strategic environmental health programme, including studies of exposures and biomarkers related to traffic-generated air pollution, was carried out with bus drivers, letters carriers, and post office workers. A study of bus drivers categorized the exposure groups as high for drivers within the city of Copenhagen, medium for drivers in the suburbs, and low for drivers in the countryside (Knudsen et al., 1999; Table 4.2). When stratified by genotype/phenotype, those bus drivers who were glutathione S-transferase M1 (GSTM1) null and had the NAT2 slow acetylator genotype exhibited an exposure-related increase in CAs compared with those with the NAT2 fast acetylator and GSTM1-positive genotypes (Knudsen et al., 1999). [The role of genotype/phenotype is discussed later in this Monograph (see Section 4.4).] A separate study of bus drivers in Sofia, Bulgaria, found increased frequencies of CAs in that population relative to office workers when analysed either by traditional cytogenetic methods (Zidzik et al., 2007) or by FISH (Beskid et al., 2007).

A comparison of mail carriers in Copenhagen, Denmark, with office workers found that the mail carriers who were NAT2 slow acetylators had higher frequencies of CAs compared with those who were NAT2 fast acetylators (Knudsen et al., 1999; Table 4.2). This result suggested that the NAT2 genotype may influence responses to other common exposures or may influence the baseline frequencies of CAs (Knudsen et al., 1999).

Using white blood cells from 55 children attending a school in a rural area of Thailand and from 91 children attending an urban school in Bangkok, Thailand, Tuntawiroon et al. (2007) and Ruchirawat et al. (2007) exposed the cells to 100 cGy of ionizing radiation from a ¹³⁷Cs source at a dose rate of 5 Gy/minute and then determined the frequency of CAs. The authors found significantly higher frequencies of deletions/metaphase among the urban schoolchildren (0.45 ± 0.01) than among the rural schoolchildren (0.26 ± 0.01). Exposure analyses showed that the urban schoolchildren had higher concentrations of urinary 1-OHP and blood benzene compared with the rural schoolchildren. In addition, the air in the urban area had higher concentrations of PAHs and benzene than that from the rural area. Together, these studies indicated that the global DNA repair system of the urban schoolchildren was less effective at repairing the DNA damage after a challenge by ionizing radiation compared with that of the rural schoolchildren.

Cui et al. (1991) determined the frequency of CAs in the chorionic villi of 2698 women (aged 25–35 years) having abortions for non-medical reasons at less than 10 weeks of pregnancy from three cities with different levels of air pollution. The three cities were Shenyang, which had heavy pollution due to industry and coal combustion (811 women), Zhengzhou, which had moderate air pollution (1060 women), and Dalian, which was the least polluted, with light industry (827 women). The incidences of polyploidy, trisomy, and chromosome structural abnormalities in the women in Shenyang were 2.3, 3.4, and 16 times, respectively, those in the women in Dalian. Similarly, the cytogenetic frequencies of polyploidy, trisomy, and chromosome structural abnormalities in the women in Zhengzhou were 3.9, 1.3, and 4.9 times, respectively, those in the women in Dalian. The data for structural abnormalities are shown in Table 4.2. The results suggested that there was a positive correlation between the incidence of numerical and/or structural CAs and the severity of air pollution, especially SO₂ concentrations (Cui et al., 1991).

(ii) Micronuclei

The studies reviewed here cover a variety of exposure situations (Table 4.3), and of those that included exposure assessments, all found higher levels of exposure to air pollutants among the group exposed to outdoor air compared with the unexposed controls. Three studies evaluated MN in buccal cells (Karahalil et al., 1999; Hallare et al., 2009; Sellappa et al., 2010), and the rest evaluated MN in lymphocytes; however, one evaluated MN in both cell types (Cavallo et al., 2006), and one evaluated MN in maternal lymphocytes and cord blood (Pedersen et al., 2009). Most studies found increased frequencies of MN in the outdoor versus indoor exposure settings, or in populations living in urban/industrial areas versus rural areas. All lymphocyte studies except that of Zhao et al. (1998) used the cytokinesis-block version of the MN assay; those studies that used buccal cells did not (Karahalil et al., 1999; Cavallo et al., 2006; Hallare et al., 2009; Sellappa et al., 2010).

Table 4.3

Micronuclei in humans exposed to polluted outdoor air.

Six studies investigated the induction of MN in traffic police relative to controls not exposed chronically to traffic (Table 4.3). Traffic police in Ankara, Turkey, had higher frequencies of buccal cell MN compared with the controls (details of controls not specified) (Karahalil et al., 1999; Table 4.3). Likewise, higher frequencies of buccal cell MN were found in traffic police in Manila, Philippines, compared with other residents of Manila (Hallare et al., 2009; Table 4.3). Increased frequencies of MN were found in buccal cells of traffic police in Lanzhou, China, compared with the frequencies found in police who worked in offices (Zhao et al., 1998; Table 4.3).

A study of MN in lymphocytes in traffic police in Genoa, Italy, found increased frequencies in this group relative to a group of indoor workers (Merlo et al., 1997; Table 4.3). Exposure assessments found a 30-fold higher concentration of PAHs in the air outdoors compared with that in the office space. Another study in lymphocytes of traffic police in Genoa found no increase in MN frequencies in traffic police compared with a group of laboratory workers (Bolognesi et al., 1997a; Table 4.3). Nonetheless, exposure assessments showed higher levels of B[a]P in the air outdoors compared with that in the laboratories. Higher frequencies of MN were found among traffic police in Hebei, China, compared with police who worked in offices (Bai et al., 2005; Table 4.3), and exposure assessments found increased concentrations of inhaled particles in the air breathed by the exposed compared with the control subjects. Increased concentrations of NO_x, CO, hydrocarbons, and lead were also found among the exposed versus the control populations.

Unlike the previous studies of traffic police, the study by Rossnerova et al. (2009) evaluated traffic police in Prague, Czech Republic, but compared the MN frequencies among the police as measured in a more-polluted month (February) versus a less-polluted month (May). Exposure assessments had shown that the air in February had higher concentrations of carcinogenic PAHs, B[a]P, and various VOCs (benzene, ethylbenzene, and o-xylene). Traffic police had frequencies of MN that were higher in the more-polluted month (February) compared with those in the less-polluted month (May) (Table 4.3).

Filling station attendants in Manila, Philippines, had higher frequencies of buccal cell MN compared with residents of Manila (Hallare et al., 2009; Table 4.3). The frequencies of buccal cell MN were higher in filling station attendants in Coimbatore City, India, compared with other residents of Coimbatore City (Sellappa et al., 2010; Table 4.3). The frequencies of buccal cell MN were higher in taxi drivers in Ankara, Turkey, compared with control subjects (Karahalil et al., 1999; Table 4.3). [The Working Group noted that this human study addressed an occupational situation and might not be broadly applicable.]

No increases in either buccal or lymphocyte MN frequencies were found among outdoor airport workers at the international airport in Rome, Italy, compared with airport office workers (Cavallo et al., 2006; Table 4.3). Exposure assessments found that the concentration of PAHs was higher outdoors than in the offices, but concentrations of urinary 1-OHP were not different between the two groups of workers.

One study found higher frequencies of MN in lymphocytes of mothers and umbilical cord blood from those mothers who lived in high- versus low-traffic areas of Denmark (Pedersen et al., 2009; Table 4.3).

Several studies evaluated subjects living in urban/industrial areas versus rural areas. For example, Ishikawa et al. (2006) showed that residents of an industrial district of Shenyang, China, had higher frequencies of MN compared with residents of a rural district of the same city (Table 4.3). Likewise, Pedersen et al. (2006) found that young children living in an urban area (Teplice, Czech Republic) had higher frequencies of MN compared with young children living in a rural area (Prachatice, Czech Republic) (Table 4.3).

Another comparison of subjects working in two different environments was performed by Peng & Ye (1995), who measured the frequency of MN in bus drivers or on-site bus ticket officers a route that runs through a tunnel in Shanghai, China, and compared the results with those obtained in officers who worked in the Shanghai Botanical Garden. Daily average TSP concentrations in the tunnel were extremely high (1.86 mg/m³) compared with the established standard of 0.15 mg/m³. The bus drivers and on-site bus ticket officers had higher MN frequencies compared with the officers in the botanical garden (Table 4.3).

(iii) Sister chromatid exchanges

SCEs have been used extensively as a biomarker of genotoxicity; however, unlike CAs and MN, SCEs have not turned out to be predictive of cancer risk (Norppa et al., 2006). Nonetheless, SCEs are a sensitive indicator of exposure to a variety of genotoxic agents; as reviewed below, this includes exposure to outdoor air pollution.

There were 11 studies in which SCEs in lymphocytes were investigated as a biomarker associated with exposure to outdoor air pollution (Table 4.4). Of these 11 reports, all but two found increased frequencies of SCEs in the exposed population compared with the control subjects. Among the 11 reports, four exposure groups were studied; three of the 10 studies included exposure assessments, all of which found a difference between the exposure levels of the exposed and control populations. All of the studies cultured the lymphocytes for 72 hours, except for Sree Devi et al. (2009), where the cells were cultured for 70 hours, and Cavallo et al. (2006), where the cells were cultured for 48 hours.

Table 4.4

Sister chromatid exchanges in lymphocytes of humans exposed to polluted outdoor air.

There were eight studies of SCEs in traffic police, all but one of which found higher frequencies of SCEs in the traffic police relative to the control populations (Table 4.4); the one negative study included an exposure assessment. Traffic police in Cairo, Egypt, had higher frequencies of SCEs compared with police trainers (Anwar & Kamal, 1988); the same was true for traffic police in Madras, India, compared with subjects not working in traffic (Chandrasekaran et al., 1996). Traffic police in Lanzhou, China, had higher frequencies of SCEs compared with police who worked in offices (Zhao et al., 1998); the same was true for traffic police in Hyderabad, India (Sreedevi et al., 2006), or Chennai City, India (Anbazhagan et al., 2010), compared with office workers. Traffic police in Bangkok, Thailand, had higher frequencies of SCEs compared with university students, who were used as the control population (Soogarun et al., 2006). Although traffic police in Genoa, Italy, did not have elevated frequencies of SCEs compared with laboratory workers, used as controls (Bolognesi et al., 1997b), an exposure assessment found that the concentration of B[a]P and other PAHs in the outdoor air was higher than that in the laboratory spaces. Traffic police in Hebei, China, had higher frequencies of SCEs compared with police who worked in offices (Bai et al., 2005; Table 4.4), and exposure assessments showed that there were higher concentrations of particles, NO_x, CO, hydrocarbons, and lead in the air for the exposed populations compared with the controls.

Outdoor workers at the international airport in Rome, Italy, had higher frequencies of SCEs compared with indoor workers at the airport (Cavallo et al., 2006; Table 4.4), and exposure assessments found that the outdoor air had higher concentrations of PAHs than the indoor air. Nonetheless, there was no difference in the urinary concentration of 1-OHP between the outdoor and indoor workers. The frequencies of SCEs were not higher among tunnel workers in the Umbrian Apennine Mountains, Italy, compared with outdoor workers away from traffic (Villarini et al., 2008; Table 4.4).

A comparison of subjects working in two different environments was performed by Peng & Ye (1995), who measured the frequency of SCEs in bus drivers or on-site bus ticket officers on a route that runs through a tunnel in Shanghai, China, versus that of officers who worked in the Shanghai Botanical Garden. Daily average TSP concentrations in the tunnel were extremely high (1.86 mg/m³) compared with the established standard of 0.15 mg/m³. The bus drivers and on-site bus ticket officers had higher SCE frequencies compared with the officers in the botanical garden (Table 4.4).

In summary, two types of studies were performed to evaluate CAs, MN, and SCEs in humans exposed to outdoor air pollution. One type studied subjects whose work involved being outside (frequently in or near traffic) for most of the workday (e.g. traffic police, street vendors, and toll booth operators) compared with subjects who worked primarily indoors (e.g. office workers). Another type of study compared subjects who lived or worked in more- versus less-polluted airsheds. Nearly all studies showed that polluted outdoor air induced significantly higher cytogenetic effects relative to either indoor air or less-polluted outdoor air. These studies covered 10 countries for CAs, seven for MN, and five for SCEs. Two of these end-points (CAs and MN) are associated with increased risk of cancer, highlighting the importance of these genotoxicity biomarker studies.

(i) In vivo

Animals

See Table 4.5.

Table 4.5

Cytogenetic damage associated with outdoor air pollution in experimental animals in vivo.

Chromosomal aberrations

Several studies have examined the effect of outdoor air pollution or samples derived from it on cytogenetic abnormalities in experimental animals in vivo. Only one study examined the frequency of cytogenetic abnormalities in animals exposed in situ at locations highlighted for poor air quality. Rubeš et al. (1997) investigated cytogenetic effects in peripheral blood lymphocytes of dairy cattle in the Teplice district of the Czech Republic, an industrialized area with severe air pollution, and the Prachatice district, an agricultural area with lower levels of air pollution. The results revealed a significantly higher percentage of aberrant cells (chromatid aberrations or CAs) in animals in Teplice relative to Prachatice. The CAs included chromatid breaks, isochromatid breaks, acentric fragments, and translocations (Rubeš et al., 1997).

Four studies conducted in China examined the ability of extracts of airborne PM to induce various cytogenetic abnormalities in murine bone marrow. The study by Wang & Zhang (1984) was conducted in the northern Chinese city of Harbin, which experiences a marked reduction in air quality in colder months due to residential heating by coal. The authors reported that mice exposed orally by gavage to methanol extracts of TSP collected from a residential site in the winter showed a dose-dependent increase in aneuploidy and CAs. The CAs included chromosome breaks, fragments, dicentric chromosomes, and ring chromosomes (Wang & Zhang, 1984).

The other three studies were conducted in Taiyuan, an industrialized city in north-western China that contains chemical production facilities and coal-fired electricity generation facilities. Yang & Wu (1984) found that mice (strain not specified) treated with a single intraperitoneal injection of methanol extracts of TSP collected from several locations had dose-dependent increases in CAs in bone marrow cells. Marked increases were noted for samples from industrial, commercial, and residential sites, and the CAs included chromatid or chromosome breaks for the industrial or high-traffic sites and ring chromosomes for the commercial or residential sites. The authors observed that the frequency of induced CAs was positively associated with PM level (Yang & Wu, 1984).

A similar study in Taiyuan involved intraperitoneal exposures of mice to inorganic (i.e. nitric acid) PM extracts, or PM extracts prepared using simulated lung fluid. The results showed significant dose-dependent increases in CA frequency and increased responses for particles smaller than 2.5 µm. For the inorganic extract, the authors reported that the observed CA frequency was correlated with concentrations of lead, manganese, chromium, nickel, and cadmium (Lei et al., 1993). The study by Sun et al. (1995) investigated CAs in male germ cells of Kunming mice treated intraperitoneally with dichloromethane (DCM) extracts of TSP collected downwind of a coal-fired electricity generation facility. The results revealed dose-dependent increases in sperm abnormalities, CAs in spermatogonia and primary spermatocytes, and frequencies of meiotic MN in early spermatids (Sun et al., 1995) (see also Section 4.3.3).

Micronuclei

Eleven published studies examined the ability of outdoor air or samples derived from outdoor air to induce MN in vivo. The majority of these studies (8 of 11) investigated the frequency of MN in haematopoietic tissues (bone marrow or peripheral blood) in mice exposed intraperitoneally to a single acute dose or to repeated (2–5) consecutive daily doses. One study examined the frequency of MN in peripheral blood of Balb/c mice exposed in situ for up to 120 days to urban air in São Paulo (Soares et al., 2003). The results revealed a significant increase in MN frequency relative to a rural control location, and MN frequency was positively correlated with atmospheric levels of CO, NO₂, and PM₁₀ (Soares et al., 2003).

The study by Izzotti et al. (1996) examined the ability of cyclohexane extracts of TSP collected in Sicily, Italy, from urban and rural locations to induce MN in rat alveolar macrophages and pulmonary epithelial cells after five consecutive daily intratracheal instillations. The results showed significant 3.4-fold and 4.5-fold increases in MN frequency in pulmonary alveolar macrophages and epithelial cells, respectively, of rats treated with extracts from urban locations relative to the control (Izzotti et al., 1996). Zhao et al. (2001) noted that four consecutive gavage doses of DCM extracts of TSP collected at a heavy-traffic location in Lanzhou, an industrialized city in north-western China, induced a significant dose-dependent increase in MN frequency in bone marrow of Kunming mice (Zhao et al., 2001).

Two studies in Europe investigated the ability of PM extracts to induce increases in MN frequency in mouse bone marrow after intraperitoneal injection. Motykiewicz et al. (1990, 1996) showed that benzene extracts of PM collected from the heavily industrialized region of Upper Silesia, Poland (which has coke production, metal refining, steel foundries, etc.), induced a significant increase in micronucleated polychromatic erythrocytes (PCEs) in Balb/c mice after two consecutive intraperitoneal injections. Crebelli (1989) reported that two consecutive intraperitoneal doses of DCM extract from PM collected in Rome, Italy, failed to elicit significant increases in micronucleated PCEs in Swiss mice (Crebelli et al., 1988).

Six studies investigated the ability of organic extracts of PM collected in urban centres in China (Beijing, Shanghai, and Taiyuan) to induce significant increases in MN frequency in bone marrow of Kunming mice exposed via intraperitoneal injection. For example, Wang et al. (1991) noted that nanomaterial extracts of TSP from several sites in Beijing induced dose-dependent increases in MN frequencies and, moreover, that MN frequencies were markedly higher for samples from industrial or commercial areas relative to those from residential areas (Wang et al., 1991). Similarly, studies by Yao et al. (1993) and Zhao et al. (2002) revealed that DCM extracts of PM collected from a variety of locations in Shanghai induced significant dose-dependent increases in MN frequency for 10 of the 13 locations examined, with a maximum response approximately 5-fold above the control (Yao et al., 1993).

Studies by Bai et al. (1999) and Zhang et al. (2002) investigated the ability of DCM extracts of TSP collected in Taiyuan to induce a significant increase in MN frequency. Bai et al. (1999) found dose-dependent increases in MN frequency, and extract fractionation showed no induction of MN by the aliphatic hydrocarbon fraction but induction of high MN frequencies by fractions containing organic acids, polar aromatics, basic organics, and PAHs (Bai et al., 1999).

Zhang et al. (2002) studied a site near the Taiyuan steel foundry and compared results with those obtained from samples from a less-contaminated site at Yangqu. They found a significant induction of MN by air sample extracts from both sites, but there was a marked increase for extracts of TSP from the foundry area. The authors observed that the results correspond to a higher incidence of lung cancer in the Taiyuan foundry area relative to the control site (Zhang et al., 2002).

A single study in Japan reported a significant dose-related increase in micronucleated PCEs in Balb/c mice exposed to a methanol extract of PM₁₀ collected in Tokyo (Sakitani & Suzuki, 1986).

Sister chromatid exchanges

A single study that investigated the ability of organic extracts of PM collected in West Virginia, USA, to induce SCEs in bone marrow and spleen cells of CD1 mice exposed via single intraperitoneal or oral administration failed to show a significant increase relative to control (Krishna et al., 1986).

In summary, polluted outdoor air, outdoor air PM, or samples derived from outdoor air PM are capable of inducing significant increases in cytogenetic damage in animals exposed in situ or exposed experimentally via a variety of routes of administration. Exposures of experimental animals via oral, intraperitoneal, or intratracheal administration show a clear dose-dependent induction of cytogenetic damage recorded as CAs or MN. [The Working Group expressed concerns about intraperitoneal injections and their relevance to human cancer risk.]

Plants

Plant assays (see Supplemental Table S10, available online) have also been used to assess the ability of outdoor air pollution or samples derived from it to induce cytogenetic damage (Ma et al., 1994). In particular, many studies have examined the induction of MN in meiotic pollen mother cells (i.e. tetrads formed after the second meiotic division) of the sterile Tradescantia clone 4430 or isolates of T. paludosa or T. pallida exposed to outdoor air in situ for extended periods (e.g. several months) or in the laboratory to extracts of airborne PM.

In their review of the mutagenicity and carcinogenicity of outdoor air pollution, Claxton & Woodall (2007) observed that although the dynamic range of plant genotoxicity assays varies across plants and end-points, the dynamic range of the induced responses, relative to the control, for the popular Tradescantia assays ranges between 2-fold for the stamen-hair mutation assay and 20-fold for the MN assay. In their review of mutagens in contaminated soils, White & Claxton (2004) also critically examined the utility of the Tradescantia genotoxicity assays and noted the limited dynamic range and lack of sensitivity, particularly for short-term exposures.

(ii) In vitro

Cytogenetic effects induced by extracts of airborne particulates were assessed in cultured human lymphocytes, human cell lines, cultured animal primary cells, and animal cell lines. The cytogenetic effects included CAs, aneuploidy, MN, and SCEs. The results of these in vitro cytogenetic studies are summarized in Table 4.6.

Table 4.6

Cytogenetic damage associated with outdoor air pollution in human and animal cells in vitro.

Chromosomal aberrations

Human cells

A series of studies showed significant dose-related increases in the frequency of chromosome and chromatid breaks in cultured human lymphocytes exposed to organic extracts of airborne PM from the Rhine–Ruhr region of Germany (Hadnagy et al., 1986, 1989; Hadnagy & Seemayer, 1987; Seemayer et al., 1989). Acetone extracts of outdoor air PM collected in West Virginia, USA, also induced CAs in human lymphocytes in a dose-dependent manner (Krishna et al., 1984). Three studies in China also showed that extracts of outdoor air particles induced CAs in human lymphocytes. Organic extracts of airborne TSP samples collected at five sites in Lanzhou, a city heavily contaminated by coal combustion and automobile exhausts, all induced CAs in cultured human lymphocytes obtained from umbilical cord blood. The major CAs included chromatid gaps, chromosome gaps, chromatid breaks, chromosome breaks, and fragments. Potency was correlated with the degree of air pollution (Ding et al., 1999). Tan et al. (2002) examined water extracts of airborne TSP collected in a tunnel in Shanghai and noted significant induction of CAs, including fragments and dicentric chromosomes, in cultured human lymphocytes compared with controls. The authors reported that metals such as lead, zinc, manganese, and iron likely contribute to the observed increase in the frequency of CAs. Wei & Meng (2006a) examined water extracts of PM_2.5 samples collected from Baotou (an industrial city in Inner Mongolia) and Wuwei (an agricultural city in Gansu province) and noted dose-dependent increases in the frequencies of CAs in cultured human lymphocytes. Samples were collected during sandstorms as well as on non-storm control days, and CA frequencies were higher in Baotou compared with Wuwei, for non-storm conditions only. CAs included chromatid breaks, chromosome breaks, acentric fragments, dicentric chromosomes, and gaps (Wei & Meng, 2006b).

Animal cells

A study by Motta et al. (2004) showed significant increases in CA frequencies in Chinese hamster epithelial liver cells, which maintained metabolic competence, exposed to extracts of airborne PM from Catania, Italy. However, negative results were seen in Chinese hamster ovary cells that required exogenous metabolic activation (Motta et al., 2004). Alfaro Moreno et al. (1997) reported a dose-related increase in anaphase aberrations in murine Balb/c 3T3 cells exposed to a suspension of atmospheric dust collected in Mexico. A study by Zwanenburg (1988) reported significant induction of CAs in Chinese hamster V79 lung cells exposed to extracts of particles collected from several sites after a large industrial fire in Switzerland in the presence of S9 (Zwanenburg, 1988). Extracts of PM from some urban sites elicited positive responses 4–5 months after the fire, and the authors could not provide convincing evidence that the fire resulted in the release of clastogenic substances. A study conducted in Poland, which examined fractions of organic extracts of airborne PM from high-pollution locations, showed significant increases in CA frequencies in Chinese hamster V79 lung cells by fractions containing polar aromatics, monophenols, and basic N-heterocyclics (Motykiewicz et al., 1988). A study by Ares et al. (2000) failed to show significant increases in CA frequency in primary F344 rat hepatocytes treated with extracts of airborne PM from Patagonia (Ares et al., 2000). Two studies reported significant increases in aneuploidy in Chinese hamster V79 lung cells exposed to extracts of airborne PM from Poland (Motykiewicz et al., 1991) and Germany (Hadnagy & Seemayer, 1991).

Micronuclei

Human cells

Water or organic solvent extracts of airborne PM from five cities in China were tested for induction of MN in cultured human lymphocytes. In Lanzhou, DCM extracts of TSP from five sites with varying degrees of air pollution all caused dose-dependent increases in MN frequency in cultured human lymphocytes. The samples from sites with heavy traffic or close to petroleum industries showed more potent induction of MN compared with samples from moderately contaminated sites or relatively clean sites (Ding et al., 1999). In Shanghai, saline extracts of airborne PM from Taopu, an industrial region, caused a dose-dependent increase in MN in cultured human lymphocytes (Tan et al., 2004). Wei & Meng (2006a) and Wei et al. (2006) compared MN induction by organic and inorganic extracts of PM_2.5 collected during a sandstorm or in non-storm conditions from the industrial city of Baotou (in Inner Mongolia) and the agricultural city of Wuwei (in Gansu province). The results indicated that organic and saline PM suspensions, collected during storm and non-storm conditions, showed dose-dependent increases in MN frequency in cultured human lymphocytes (Wei et al., 2006). DCM extracts of TSP and PM₁₀ samples from Guangzhou also significantly increased MN frequency in cultured human lymphocytes, and extract fractionation showed significant MN induction by the aromatic hydrocarbon fraction of PM₁₀ (Xu & Wang, 2008). A study by Yuan et al. (1999a) examined acid and organic solvent extracts of airborne PM of different sizes (< 1.1 µm, 1.1–2.0 µm, 2.0–3.3 µm, 3.3–7.0 µm, and > 7.0 µm) that were collected from a residential area in Taiyuan, and they observed dose-dependent increases in frequencies of MN in cultured human lymphocytes (Yuan et al., 1999a). Acid extract studies showed that the smaller the particulate size, the higher the MN frequency; the MN frequencies were also positively correlated with the concentrations of metals in the PM extracts (Yuan et al., 1999b). In addition, a study conducted in Flanders, Belgium, also reported a significant dose-related (i.e. equivalent cubic metres per millilitre) increase in MN frequency in cultured human lymphocytes exposed to organic extracts of urban air PM (Brits et al., 2004).

In addition to studies in cultured human lymphocytes, the induction of MN by suspended PM or extracts of PM was also investigated in human cell lines. Oh et al. (2011) found significant induction of MN in BEAS-2B human lung bronchial epithelial cells exposed to organic PM extracts and extract fractions from a high-traffic area in the Republic of Korea (Oh et al., 2011). Fractionation showed significant increases in MN for aliphatic, aromatic (i.e. PAHs), and slightly polar (i.e. nitro-PAHs and quinones) fractions. Significant increases in MN frequency were also induced in A549 human alveolar adenocarcinoma cells by water and organic extracts of industrial and residential particles from Mexico City (Roubicek et al., 2007), and in HS 27 human skin fibroblasts (Poma et al., 2002).

Animal cells

Significant increase in MN frequency were also observed in RAW 264.7 mouse macrophages exposed to suspensions of airborne PM from L’Aquila, Italy (Poma et al., 2006). Moreover, an organic extract of PM from Beijing, China, induced a significant dose-related increase in MN frequency in Balb/c 3T3 cells (Zhang et al., 2003).

Sister chromatid exchanges

Human cells

Twenty-five studies investigated the induction of SCEs in cultured human lymphocytes and a variety of cultured animal cells exposed to organic PM extracts, and most of the studies showed significant increases in SCE frequency. Organic PM extracts assessed in cultured human lymphocytes include samples derived from PM collected in Lexington, Kentucky, USA (Viau et al., 1982), Silesia, Poland (Motykiewicz et al., 1990), Lanzhou, China (Zhang & Li 1994; Wang & Ding 1998), West Virginia, USA (Krishna et al., 1984), and many sites in the Rhine–Ruhr region of Germany (Seemayer et al., 1984, 1987a, b, 1988, 1989, 1990a, 1990b; Hadnagy et al., 1986, 1989; Hadnagy & Seemayer, 1987). Seasonal PM samples from Mexico City induced significant dose-related increases in SCE frequency, with the highest frequencies produced by samples taken in November and the lowest by samples taken in April (Calderón-Segura et al., 2004). Significant increases in SCE frequencies were induced in HS 27 human skin fibroblasts exposed to suspensions of PM from L’Aquila, Italy (Poma et al., 2002), and in A549 human alveolar adenocarcinoma cells and human BEAS-2B cells exposed to organic extracts of PM collected from the Rhine–Ruhr region of Germany (Seemayer et al., 1989; Hornberg et al., 1998).

Animal cells

A series of in vitro cytogenetic studies of organic extracts of PM from several locations within the Rhine–Ruhr region of Germany showed significant dose-dependent increases in SCE frequencies in primary tracheal epithelial cells from Syrian golden hamsters or Wistar rats (Seemayer et al., 1994; Hornberg & Seemayer, 1995; Hornberg et al., 1996, 1997). Significant dose-related increases in SCE frequencies were also induced in primary bone marrow and spleen cells exposed to extracts of PM from West Virginia, USA (Krishna et al., 1986). Three independent studies showed significant increases in SCE frequencies in Chinese hamster ovary cells exposed to extracts of PM from the Netherlands (de Raat, 1983), a coastal area in Finland (Pyysalo et al., 1987), and Athens, Greece (Athanasiou et al., 1987). Two independent studies showed significant increases in SCE frequency in Chinese hamster V79 cells exposed to organic extracts of PM from the Netherlands (Alink et al., 1983) and Paris, France (Courtois et al., 1988). The study in Finland noted that concentrates of SVOCs collected on XAD-2 resin consistently elicited stronger responses compared with PM extracts (Pyysalo et al., 1987).

A study by Yang et al. (1994) showed that extracts of airborne PM of various sizes (< 1.1 µm, 1.1–2.0 µm, 2.0–3.3 µm, 3.3–7.0 µm, and > 7.0 µm) collected from industrial, residential, and suburban districts in Taiyuan, China, induced significant dose-related increases in SCEs in Chinese hamster lung cells (Yang et al., 1994). The authors also observed that a greater response was produced by extracts of smaller-sized PM and that SCE induction was positively correlated with PAH concentrations.

In summary, substantial evidence consistently shows that organic extracts, water extracts, or suspensions of outdoor air PM from urban or industrial areas induce significant dose-related cytogenetic effects (CAs, aneuploidy, MN, and SCEs) in cultured human lymphocytes, human cell lines, cultured animal primary cells, or animal cell lines in vitro.

(i) Humans

Studies on DNA adducts in humans after exposure to polluted outdoor air are summarized in Table 4.7.

Table 4.7

DNA adducts in humans exposed to polluted outdoor air.

A systematic review (Demetriou et al., 2012) evaluated DNA adducts as one of several biomarkers with the potential to contribute an intermediate end-point in the association between air pollution and lung cancer and graded DNA adducts in leukocytes as A for evidence, A for replication, and B for bias.

In an early study of the effects of outdoor air pollution, male residents of an industrial and highly polluted city in Poland (Gliwice) were compared with men from a rural part of the country (Biała Podlaska) (Perera et al., 1992). Both summer and winter samples were analysed by enzyme-linked immunosorbent assay (ELISA) and by ³²P-postlabelling for PAH–DNA adducts and aromatic DNA adducts. The exposed residents of Gliwice had significantly increased PAH–DNA and aromatic adducts compared with rural residents. For the winter samples, the Gliwice values were significantly greater than the rural values by ELISA only, and the same was found for the summer samples. The Gliwice winter values were also significantly greater than the Gliwice summer values by both methods of analysis. Other comparisons that were statistically significant were by ³²P-postlabelling: control winter values were greater than exposed summer values, and exposed winter values were greater than control summer values (see Table 4.7).

Other studies in Poland have focused on mother–newborn pairs. Analysis by immunoassay (ELISA) for PAH–DNA adducts of maternal and cord white blood cells of mothers and newborns from Cracow found significant correlations between adduct levels and outdoor air pollution levels (PM₁₀) close to their places of residence (Whyatt et al., 1998). In a subsequent study, the Cracow cohort was compared with a rural cohort from Limanowa, and differences in adduct levels between the urban and rural groups were not significant (Perera et al., 1999); however, it was noted that there was heavier use of coal for home heating in the rural district than in the city. Among non-coal users only, adduct levels in Cracow maternal samples were significantly higher than those in Limanowa maternal samples.

In a subsequent study of a larger group of Cracow women, PAH exposure was estimated from personal air monitors worn during pregnancy (Kelvin et al., 2009). There was a significant interaction between prenatal exposure to PAHs and the levels of cord blood B[a]P–DNA adducts, determined by high-performance liquid chromatography (HPLC)/fluorescence analysis. This association was stronger in babies with low blood levels of α-tocopherol and carotenoids.

An early study measured DNA adducts by ³²P-postlabelling in the lymphocytes of taxi drivers, urban bus drivers, suburban bus drivers, and controls (hospital workshop workers) in Stockholm, Sweden (Hemminki et al., 1994). The adduct levels in the taxi drivers and the suburban bus drivers were significantly higher than those in the controls (P < 0.01 and P < 0.001, respectively), but the adduct levels in urban bus drivers were not significantly different from those in the controls.

Another study investigated DNA adducts in bus drivers in Copenhagen, Denmark (Nielsen et al., 1996a). Significantly higher DNA adduct levels were found in the drivers in central Copenhagen compared with those driving in outer areas, and all driver groups had significantly higher levels than controls consisting of rural dwellers or the general population.

A study on rickshaw drivers in Dhaka City, Bangladesh, used ELISA to detect PAH–DNA adducts in white blood cells (Rahman et al., 2003). A higher proportion of the drivers, who were not shielded or protected from exposure to traffic pollution, had detectable DNA adducts than a control group of unexposed people (19/46 vs 11/48; P = 0.06), and the mean adduct level was significantly higher in the rickshaw drivers than in the controls (overall, P = 0.04; for those with detectable adducts, P = 0.01).

Studies in the Czech Republic have focused on comparisons of residents of a highly industrialized and polluted region, Northern Bohemia, with those of a relatively unpolluted rural part of the country. When women in Teplice (in the industrialized region) were compared with women in Prachatice (rural control), significant associations between bulky DNA adducts in white blood cells and individual levels of exposure to PAHs (measured by personal air monitors) were found (Binková et al., 1996). A study of the placentas of mothers in Teplice detected bulky DNA adducts by ³²P-postlabelling and PAH–DNA adducts by immunohistochemistry that were unrelated to the smoking status of the women (Pratt et al., 2011).

One study of male police officers working outdoors in the downtown area of Prague found a correlation between bulky DNA adducts in their lymphocytes and air levels of PAHs at the various sampling times (Topinka et al., 2007), but another study that compared male police officers in Prague with residents of the city did not find a difference in overall adduct levels between the two groups, although the level of a B[a]P-like adduct was significantly higher in the exposed group (Binková et al., 2007).

Comparisons of Prague residents with those of an area of higher pollution, the industrialized region of Ostrava, revealed that levels of B[a]P-like adducts in lymphocytes were positively affected by B[a]P exposure levels for the Ostrava residents but not for Prague residents (Rossner et al., 2013a). Although B[a]P concentrations were higher in Ostrava, levels of B[a]P-like adducts were higher in Prague. For total bulky adducts, levels in both cohorts were negatively associated with B[a]P and pollution levels.

A study that monitored residents of Prague (Czech Republic), Košice (Slovakia), and Sofia (Bulgaria) compared several biomarkers of exposure, including DNA adducts detected by ³²P-postlabelling (Singh et al., 2007a). Levels of total bulky DNA adducts, and also of B[a]P-like adducts, negatively correlated with levels of the oxidatively generated lesion 8-oxodG in DNA.

In a study that compared non-smoking men in Athens, Greece, with men in urban and rural areas of Denmark, the median adduct levels in white blood cells (Athens) or lymphocytes (Denmark) were significantly different in the three groups; levels in Athens were higher than those in urban Denmark, and levels in urban Denmark were higher than those in rural Denmark (Nielsen et al., 1996b). However, in another study of students living in Copenhagen, Denmark, there was no significant association between levels of bulky DNA adducts in lymphocytes and two exposure markers – levels of PM_2.5 and black smoke – measured by personal exposure monitors (Sørensen et al., 2003a). Also, among mother–newborn pairs living in Copenhagen, adduct levels were significantly elevated in maternal and cord blood of those living in medium-traffic-density areas, but not of those living in high-traffic-density areas, relative to those living in low-traffic-density areas (Pedersen et al., 2009)

Several studies in Italy have shown a positive association between exposure to outdoor air pollution and DNA adduct levels. Traffic police had significantly higher levels of bulky DNA adducts in white blood cells than age-matched urban residents (Merlo et al., 1997). Police officers in Genoa had a significantly higher median DNA adduct level than office workers (Peluso et al., 1998). Traffic-exposed workers in Florence had significantly higher DNA adduct levels than urban residents (Palli et al., 2001), with a significant correlation between adduct levels and ozone concentrations (cumulative exposure) (Palli et al., 2004). DNA adduct levels in white blood cells in traffic-exposed workers in Florence correlated with average levels of exposure to PM₁₀ (Palli et al., 2008). When urban (Pisa) and suburban (Cascina) residents were compared for antibodies to B[a]P diol epoxide (BPDE)–DNA adducts in serum, there was a significant excess prevalence of antibody positivity among the urban residents (Petruzzelli et al., 1998). However, the same measurement carried out among police officers in Rome found only a non-significant increase (P = 0.095) among traffic police (10/134 positive for antibodies) compared with those with office duties (1/60 positive) (Galati et al., 2001). An earlier study of newspaper vendors, in whom bulky DNA adducts were measured in lymphocytes, did not find a difference between those working in high-traffic areas and those working in low-traffic areas of Milan (Yang et al., 1996).

In Thailand, schoolchildren in Bangkok were found to be exposed to levels of airborne PAHs 3.5-fold higher than those in a rural area (Ruchirawat et al., 2007, Tuntawiroon et al., 2007); in the same study, mean levels of bulky DNA adducts in blood lymphocytes in the Bangkok schoolchildren were 5 times those in the rural schoolchildren. Bangkok traffic police had significantly higher levels of bulky DNA adducts than office-based police (Ruchirawat et al., 2002). Another Thai study, of residents near an industrial estate, found higher bulky adduct levels in white blood cells (Peluso et al., 2008) and also lower methylation of the p53 gene, an epigenetic effect, associated with increased levels of DNA adducts (Peluso et al., 2012).

A study of placental DNA samples from Ukraine (exposed group) compared them with samples from a rural area of Poland (control group) for DNA adducts measured by immunoassay (Obolenskaya et al., 2010). A higher proportion of the Ukrainian group had detectable levels of PAH–DNA adducts compared with the Polish group, and those newborns with the most compromised health status also had the highest adduct levels. Among residents of Mexico City, whose white blood cell DNA was monitored by the same immunoassay technique as for the Ukrainian and Polish placental samples, it was found that the seasonal variation in the mean adduct level correlated with airborne concentrations of PM₁₀ and PM_2.5; all parameters were higher in the dry season than in the rainy season (García-Suástegui et al., 2011).

A single study in Africa found similar results to those of the studies in Europe and Asia; in Benin, levels of bulky DNA adducts in lymphocytes were significantly higher among urban residents than among people living in suburban or village environments (Ayi-Fanou et al., 2011).

Studies of non-smoking mother–newborn pairs in Tongliang, China, a city whose principal source of air pollution was a coal-fired power plant, have also measured the effect on birth outcomes and subsequent child development. When measured by HPLC/fluorescence, B[a]P–DNA adduct levels in cord blood, but not in maternal blood, were associated with decreased weight at up to 30 months and decreased head circumference at birth (Tang et al., 2006). Increased adduct levels in cord blood were also associated with several physical and cognitive scores measured at age 2 years (Tang et al., 2008). The subsequent closure of the power plant yielded the opportunity to make comparisons between infants born before the closure and those born after the closure. The associations between elevated adduct levels in cord blood and deficiencies in development seen in the earlier cohort were not seen in the post-closure cohort of infants, suggesting the benefit of reduced exposure to air pollution of children prenatally and/or postnatally (Perera et al., 2008).

Collectively, the majority of these studies demonstrate the presence of elevated levels of DNA adducts in adults occupationally exposed to outdoor air pollution, relative to comparable groups in environments with lower levels of pollution. Seasonal differences were also observed. Studies in children and newborns have found similar differences.

(ii) Experimental studies – in vivo systems

See Table 4.8.

Table 4.8

DNA adducts in animals in vivo exposed to outdoor air pollution or extracts of air particles.

When cyclohexane extracts of air particles from rural and urban areas of Sicily, Italy, were instilled intratracheally in rats for 5 consecutive days, they were found to lead to the formation of lung DNA adducts, detected by ³²P-postlabelling and synchronous fluorescence spectroscopy (Izzotti et al., 1996). Higher levels of adducts were found in the animals treated with the urban extract than in those treated with the rural sample.

Feral pigeons were caught at four different locations in the Netherlands and analysed for their DNA adduct levels in kidney, lung, and liver (Schilderman et al., 1997). Although the levels of PAHs in the particulate samples reflected the density of traffic at each location, no differences were found in the tissue adduct levels (measured by ³²P-postlabelling) in pigeons from the different sites.

Extracts of airborne particles from Shanghai, China, when tested on mouse skin, gave rise to DNA adducts, detected by ³²P-postlabelling, in the skin, liver, and kidney, but not in the lung, of the animals (Zhao et al., 2003). Most of the genotoxic activity of the fractions was attributed to the PAH component of the extracts.

When mice were exposed to diesel exhaust particles by inhalation, increased levels of bulky DNA adducts (detected by ³²P-postlabelling) were formed in their lungs (Dybdahl et al., 2004). In another study, oral exposure of pregnant mice to diesel exhaust particles resulted in detectable levels of bulky DNA adducts in the embryos (Reliene et al., 2005).

In a study investigating the germline mutagenicity of outdoor air pollution, a group of mice was exposed in situ in the vicinity of steel mills and a major highway (Yauk et al., 2008). When mouse samples were analysed by ³²P-postlabelling, bulky DNA adducts were detected in lung tissue but were below the limit of detection in testis tissue.

(iii) Experimental studies – in vitro systems

See Supplemental Table S11 (available online).

Human cells

Air samples from Prague (Czech Republic), Košice (Slovakia), and Sofia (Bulgaria) were compared in metabolically competent human hepatoma HepG2 cells, human diploid lung fibroblasts (HEL), and the human monocytic leukaemia cell line THP-1 (Sevastyanova et al., 2007). DNA adduct formation was highest in HepG2 cells, followed by HEL and then THP-1 cells. Winter samples were more active than summer samples; for the winter samples, the activity was highest for Prague, followed by Sofia and then Košice, but for the summer samples, the order was reversed, with the highest activity for Košice, followed by Sofia and then Prague. However, when the activities were related to the extractable content per cubic metre of air, the Sofia samples had the highest genotoxic activity regardless of the sampling period.

Animal cells

SRM 1649a was extracted with DCM and fractionated before testing for DNA adduct-forming activity in rat liver epithelial WB-F344 cells (Andrysík et al., 2011). When analysed by ³²P-postlabelling and HPLC, the crude extract formed only one major adduct peak, which corresponded with the (+)-anti-BPDE–dG adduct. When analysed by ³²P-postlabelling and thin-layer chromatography, the crude extract and the non-polar fraction (containing PAHs, methylated PAHs, polychlorinated biphenyls, and polychlorinated dibenzodioxins/furans) gave rise to detectable adducts, but the polar fraction (containing oxygenated derivatives of PAHs) did not.

Air samples from the Czech Republic have also been tested in cellular assays for DNA adducts. Crude and fractionated DCM extracts of air samples from Teplice (industrialized area) and Prachatice (rural area) were incubated with cultured rat hepatocytes and Chinese hamster V79NH lung cells expressing nitroreductase activity (Topinka et al., 2000). In hepatocytes, the highest DNA adduct-forming activity was found in the fractions containing most of the PAHs and nitro-PAHs, and in V79NH cells, the highest levels were caused by the fraction containing only nitro-PAHs. Winter samples had 3–4-fold higher binding potential than summer samples.

Acellular systems

Several studies have investigated the effect of extracts of air samples from areas of the Czech Republic with different levels of air pollution on calf thymus DNA in the presence of S9. The extent of DNA adduct formation, detected by ³²P-postlabelling, was determined.

Overall, these studies demonstrated the potential of organic extracts of air sample particulates to form DNA adducts when incubated with DNA in the presence of an acellular metabolizing system, and when incubated with mammalian cells. Samples collected in the winter were generally more active than samples collected in the summer.

(i) Humans – in vivo studies

Associations between air pollution and biomarkers of oxidative stress, including DNA strand breaks, have been assessed in a variety of biomonitoring studies, including controlled exposure, panel, and cross-sectional studies. The controlled exposure studies have typically been better than panel and cross-sectional studies because of better control for possible confounders. Table 4.9 provides an overview of studies that have assessed the association between air pollution exposure and DNA strand breaks in cells from humans.

Table 4.9

DNA strand breaks in blood cells from humans exposed to outdoor air pollution.

In Belgium, a cross-sectional study showed that subjects in locations with heavy industry had increased levels of DNA strand breaks in leukocytes compared with subjects in low-pollution areas (Staessen et al., 2001). The levels of DNA strand breaks in leukocytes correlated with ozone levels as well as urinary excretion of 1-OHP and benzene metabolites (trans,trans-muconic acid [t,t-MA] and o-cresol) in univariate models (Koppen et al., 2007; Staessen et al., 2001). A later cross-sectional study in Belgium of subjects in areas with different types of air pollution showed that the highest levels of DNA strand breaks in leukocytes were observed in subjects living closest to air pollution sites, whereas there was no correlation between exposure markers (t,t-MA and 1-OHP) and levels of DNA strand breaks (De Coster et al., 2008; Ketelslegers et al., 2008).

A study in Benin investigated the association between four groups of exposed subjects, encompassing taxi-moto drivers in the city of Cotonou, subjects living near roads with heavy traffic or in the suburbs, and village controls. Exposure to air pollution was determined by urinary excretion of benzene metabolites and the number concentration of ultrafine particles (UFP) at specific sites in Cotonou or in the village (midday 1-hour average, 6961–265 145 UFP/cm³). In addition, the personal exposure level of benzene was assessed as S-phenyl mercapturic acid (S-PMA) excretion in urine. The authors showed a positive relationship between the levels of DNA strand breaks in peripheral blood mononuclear cells (PBMCs) and air pollution levels in terms of either the outdoor air concentration of UFP or the S-PMA concentrations in urine (Avogbe et al., 2005).

In a study in Brazil, subjects living at a location near an oil refinery plant had higher levels of DNA strand breaks in lymphocytes compared with subjects from a city that was characterized as having little traffic and industry (Coronas et al., 2009). Another study in Brazil, with the same type of study design, showed no difference in the level of DNA strand breaks between subjects from urban industrialized and non-industrialized areas (Pereira et al., 2013). It should be noted that the results might be biased because it is difficult to separate the effect of air pollution exposure from the other variables that differ between subjects from different locations.

A study of traffic police and a matched group of reference subjects (officers working mainly indoors) in Shanghai, China, assessed exposure by personal monitoring of PM_2.5 and measured levels of DNA strand breaks in lymphocytes by the comet assay. The subjects were smokers or had stopped smoking for more than 6 months (including family members) before the study. The personal monitoring data indicated that traffic police were exposed to higher levels of PM_2.5 (115.4 ± 46.2 µg/m³) compared with officers working indoors (74.9 ± 40.1 µg/m³). Traffic police had a higher percentage of lymphocytes with a comet tail compared with officers working indoors. In addition, the levels of DNA strand breaks, assessed as the average tail moment, were reported to be higher in lymphocytes from the group of traffic police compared with the group of officers working indoors (Li et al., 2010). [The Working Group noted that the statistical analysis of the reported results was based on the total number of cells from all the subjects, 100 scored nuclei per subject times the number of subjects, giving rise to group sizes of more than 10 000 data points for approximately 100 subjects per group. This is at odds with the standard procedure of statistical analysis for the comet assay.] Another study of police officers from eight districts in Guangzhou, China, showed that traffic police had significantly higher frequencies of DNA strand breaks in lymphocytes, measured as the comet tail length (4.2 µm; 95% confidence interval [CI], 3.98–4.42 µm), compared with officers working indoors (3.23 µm; 95% CI, 2.82–3.70 µm) (P < 0.001, F = 9.23, t-test). Smoking was identified as a confounding factor, although the results showed that traffic exhaust exposure was the main factor for the level of DNA strand breaks in lymphocytes (Zhu et al., 2003).

Several studies in the Czech Republic have assessed DNA strand breaks in people with different occupations or living in areas characterized by high or low air pollution levels. The early studies focused on differences in air pollution exposures between the regions of Teplice (industrial site) and Prachatice (low-pollution area). The Teplice area has higher air pollution levels than the Prachatice area. For instance, the PM_2.5 levels were 122 µg/m³ in Teplice and 44 µg/m³ in Prachatice during the winter of 1993 (Srám et al., 1996). During the summer of 1993, the levels were 29 µg/m³ in Teplice and 18 µg/m³ in Prachatice (Srám et al., 1996). In the subsequent years, the levels of air pollution were higher in Teplice than in Prachatice, although the differences were less dramatic than during the winter of 1993. During the summer of 1993 and the winter of 1998, the typical PM₁₀ levels were 40–60 µg/m³ in Teplice and 20–40 µg/m³ in Prachatice (Srám et al., 1999). These early studies showed that personal exposures to PAHs in respirable particles correlated with levels of DNA strand breaks in lymphocytes (Binková et al., 1996). Mothers and children from the Teplice area and the Prachatice area had the same levels of DNA strand breaks in leukocytes (Srám et al., 1998). In Prague, police officers with personal exposure to PAHs had the same level of DNA strand breaks in lymphocytes as controls, although there was a difference in exposure. The personal PAH levels for the police officers and the controls were 6.5 ng/m³ and 12.4 ng/m³, respectively, in February and 3.7 ng/m³ and 16.7 ng/m³, respectively, in June (Cebulska-Wasilewska et al., 2005). Another study of police officers showed higher levels of DNA strand breaks in lymphocytes in the season with a high level of air pollution exposure (January; PM_2.5 = 33 µg/m³), whereas there was no effect in the season with a low level of air pollution exposure (September; PM_2.5 = 15 µg/m³) (Novotna et al., 2007). A study of bus drivers, garage workers, and office workers (controls) showed increased levels of DNA strand breaks in lymphocytes of workers exposed to air pollution (Bagryantseva et al., 2010).

A panel study of students who were living in the centre of Copenhagen, Denmark, showed no association between levels of DNA strand breaks in lymphocytes and personal exposure to PM_2.5 in the range of 10–24.5 µg/m³ (Sørensen et al., 2003a, b). Another study of residents of Copenhagen used benzene as marker of urban air pollution exposure and also showed no association between urinary excretion of S-PMA and levels of DNA strand breaks in lymphocytes (Sørensen et al., 2003c). The effect of personal exposure to UFP in air pollution was investigated in people bicycling for approximately 90 minutes in the laboratory or on traffic-heavy streets in Copenhagen. This study showed no association between personal exposure to UFP and levels of DNA strand breaks in PBMCs (Vinzents et al., 2005). A later study on controlled exposure to air from a busy street in Copenhagen showed a correlation between particles in the size mode with a median diameter of 57 nm (representing carbonaceous soot) and levels of DNA strand breaks in PBMCs, whereas the size mode with a median diameter of 23 nm (representing SVOCs of diesel exhaust) was not associated with elevated levels of DNA strand breaks (Bräuner et al., 2007).

Non-smoking subjects in Athens, Greece, had elevated levels of DNA strand breaks in lymphocytes compared with subjects in a rural area; there was no difference in levels of DNA strand breaks in lymphocytes between smokers in Athens and those in the rural area (Piperakis et al., 2000).

A study in Florence, Italy, showed a positive association between urban ozone concentrations (~75 µg/m³ in June and ~17 µg/m³ in January) and levels of DNA strand breaks in nasal epithelial cells, and the residents of Florence had higher levels of DNA strand breaks and ozone exposure compared with people living in a city with a low air pollution level (45 µg/m³ in June) in Sardinia (Pacini et al., 2003). Another study of subjects in Florence showed a positive association between ozone concentrations and levels of DNA strand breaks in lymphocytes (Giovannelli et al., 2006). Police officers from Rome, Italy, had unaltered levels of DNA strand breaks in leukocytes compared with a control group of office workers, despite a large difference in benzene exposure (9.5 µg/m³ vs 3.8 µg/m³ as measured by personal air sampling during a work shift) between the groups (Carere et al., 2002).

A study among subjects in Mexico City showed an association between levels of ozone and numbers of nasal epithelial cells with DNA strand breaks among adults from different locations in the city and in a low-pollution Pacific coastal town (Calderón-Garcidueñas et al., 1996). Embedded in the same study was also an assessment of the effect in young adults who moved to Mexico City from low-pollution small towns; the number of nasal cells with DNA strand breaks in this group of subjects increased during the first 2 weeks after arrival (Calderón-Garcidueñas et al., 1996). The same group of authors also showed that children in Mexico City had more nasal cells with DNA strand breaks compared with children in a low-pollution Pacific coastal town (Calderón-Garcidueñas et al., 1996, 1997). A seasonal variation was observed; samples of nasal epithelial cells that were collected during the autumn (with high air pollution levels) from a population chronically exposed to this atmospheric pollution had higher levels of DNA strand breaks compared with samples collected during the summer (with low air pollution levels) (Fortoul et al., 2010). A study of students in Mexico City showed that subjects living at a location with high outdoor air concentrations of ozone had elevated levels of DNA strand breaks in exfoliated tear duct cells compared with subjects from a location with lower ozone concentrations (Rojas et al., 2000). [This study had some limitations as judged by the standards that are used for comet assay analysis today. These include that the samples from exposed subjects and controls might have been collected and analysed at different times, without control for period effects, and that the results were reported as percentages of cells with DNA strand breaks rather than as numbers of lesions in the cells.] A later study by the same group investigated genotoxicity in nasal biopsies from children living in areas with different levels of exposure (a Pacific coastal town vs Mexico City) and showed positive associations between ozone exposure and elevated levels of DNA strand breaks in nasal cells (Calderón-Garcidueñas et al., 1999).

A study in Thailand with a relatively large benzene exposure gradient (8–50 µg/m³) among traffic police and office-based police showed no association with levels of DNA strand breaks in leukocytes, whereas there was a correlation between the levels of 1,3-butadiene and levels of DNA strand breaks (Arayasiri et al., 2010). A series of publications from studies of schoolchildren in Bangkok, compared with children in a provincial area (Chonburi), showed that the children exposed to air pollution had higher levels of DNA strand breaks in leukocytes (Buthbumrung et al., 2008; Ruchirawat et al., 2006, 2007). Schoolchildren in Bangkok had higher levels of DNA strand breaks in lymphocytes compared with children from a low-pollution area in Thailand (Tuntawiroon et al., 2007).

(ii) DNA strand breaks in the respiratory system of animals in vivo

Several studies have assessed the level of DNA strand breaks in the lungs of animals after pulmonary exposure to air pollution particles (Table 4.10). Three studies in China have observed increased levels of DNA strand breaks in lung tissue after intratracheal instillation of relatively high doses of air pollution particles (7.5 mg/kg bw and 37 mg/kg bw) (Lin et al., 2009; Meng & Zhang, 2006b, 2007; Zhang et al., 2011). Another study in China collected TSP from a residential area in Minqin county, Gansu province, where sandstorms occurred frequently. Wistar rats were exposed to these sandstorm particles in suspension by intratracheal instillation at doses of 0, 1.5, 7.5, 37.5 mg/kg bw. The TSP exposure caused a dose-dependent increase in the level of DNA strand breaks; the highest levels of DNA strand breaks were observed at 12 hours, and the effects were reduced at 24 hours after the exposure. The lowest dose that caused significantly increased levels of DNA strand breaks was 1.5 mg/kg bw (Xu et al., 2008b). However, another study on intratracheal instillation of SRM 1649 (i.e. urban dust from Washington, DC, USA) showed that 0.5 mg/kg bw administered twice during 24 hours did not increase levels of DNA strand breaks in lung tissue in mice (Vesterdal et al., 2014). Researchers in Brazil studied native rodents (Ctenomys minutus) and showed a correlation between environmental exposure to automobile emission and levels of DNA strand breaks in blood leukocytes (Heuser et al., 2002). Another study showed that dogs from different locations in São Paulo, Brazil, which had similar levels of PM₁₀, also had the same levels of DNA strand breaks in cells from the olfactory or respiratory epithelium (Kimura et al., 2010).

Table 4.10

DNA strand breaks in lungs of animals in vivo.

(iii) Human and mammalian cells in vitro

Table 4.11 lists studies that have assessed levels of DNA strand breaks in cultured cells. Several studies have shown that suspensions of PM samples or EOM of PM samples generate the same levels of DNA strand breaks in cultured cells (Brits et al., 2004; Carreras et al., 2013; Gutiérrez-Castillo et al., 2006; Healey et al., 2005; Jayasekher, 2009; Perrone et al., 2013). In addition, SRM 1649 particles retained the ability to generate DNA strand breaks in human fibroblasts after extraction in different solvents, including hexane, acetone, DCM, dimethyl sulfoxide (DMSO), and water (Karlsson et al., 2004). Another study on SRM 1648 (i.e. urban dust collected from St. Louis, Missouri, USA) showed that washed particles and the DCM extract generated lower levels of DNA strand breaks in THP-1 and A549 cells compared with the pristine particles (Don Porto Carero et al., 2001). [The concentration–response relationship was unclear.] Other studies have shown that suspensions of particles from cities in China generated DNA strand breaks in cells, as did the water and DCM extract of the particles (Meng & Zhang, 2007; Yi et al., 2014). Organic extracts of airborne particles with various sizes (< 1.1 µm, 1.1–2.0 µm, 2.0–3.3 µm, 3.3–7.0 µm, and > 7.0 µm) were also collected in a residential area in Taiyuan, China. Concentration-dependent responses of levels of DNA strand breaks in human lymphocytes were observed for airborne particles; small particles generated the highest levels of DNA strand breaks. The lowest effect level for particles smaller than 3.3 µm was 25 µg/mL (Zhang et al., 2004). The air pollution in Taiyuan consisted mainly of emissions from coal combustion, whereas the air pollution in Beijing was a mixture of coal combustion emissions and automobile exhausts. In Guangzhou, China, TSP and PM₁₀ samples were collected in a residential area in spring; organic extracts were separated into three fractions by chromatography and used to study the generation of DNA strand breaks. TSP or PM₁₀ extracts induced DNA strand breaks in human lymphocytes in a concentration-dependent manner. The aromatic hydrocarbon fraction of the TSP or PM₁₀ extract also induced a concentration-dependent increase in DNA strand breaks in human lymphocytes (Xu & Wang, 2008). In addition, the water extracts of PM_2.5 from Guangzhou on days with haze during summer and winter generated a concentration-dependent increase in DNA strand breaks (Qin et al., 2012).

Table 4.11

DNA strand breaks in mammalian cells in vitro.

Suspension solutions of PM_2.5 collected in Taiyuan, China, during the heating season caused a concentration-dependent increase in levels of DNA strand breaks in rat alveolar macrophage cells (Meng & Zhang, 2005). In another study, organic extracts and water extracts of PM_2.5 samples collected in Wuwei and Baotou, China, during normal weather or sandstorms caused a concentration-dependent increase in levels of DNA strand breaks in rat alveolar macrophage cells (Meng et al., 2006a). Zhang et al. (2003) reported that organic extracts of PM_2.5 collected at Beijing University caused a concentration-dependent increase in levels of DNA strand breaks in Balb/c 3T3 cells. The lowest concentration of the organic extract that caused a significantly increased level of DNA strand breaks was 1 m³ equiv/mL, whereas an aqueous extract of PM_2.5 had no significant effect on generation of DNA strand breaks (Zhang et al., 2003).

Several studies have found no clear difference in the potency to generate DNA strand breaks of particles collected at different locations (Alfaro-Moreno et al., 2002; Danielsen et al., 2011; Dellinger et al., 2001; De Vizcaya-Ruiz et al., 2006; Sharma et al., 2007; Shi et al., 2006; Xu & Zhang, 2004). Similarly, PM₁₀ collected from a busy street in the centre of Stockholm, Sweden, had the same potency on a mass basis as particles collected when running a road simulator (Karlsson et al., 2006). In contrast, aqueous extracts of PM_2.5 samples from industrial sites showed a higher induction of DNA strand breaks in A549 cells (Bonetta et al., 2009). One study reported that samples of PM_0.4, PM₁, PM_2.5, and PM₁₀ that were collected during the winter at a background site in Milan, Italy, were more potent than the same fractions collected during the summer (Longhin et al., 2013), whereas three studies reported no temporal variation in PM samples in regard to ability to generate DNA strand breaks (Danielsen et al., 2008; Danielsen et al., 2009, 2011). Studies on differences related to particle size indicated that the urban background PM_2.5 fraction was more potent than PM₁₀ on a mass basis (Gualtieri et al., 2011; Perrone et al., 2013). In general, consistency has been observed in studies showing increased levels of DNA strand breaks by aqueous suspensions of particles from various locations, times of the year, and size fractions. Other studies that have assessed the effect of particles from only a single site also indicated increased levels of DNA strand breaks (Dwivedi et al., 2012; Karlsson et al., 2005; Upadhyay et al., 2003), which could be reduced only slightly by treatment with deferoxamine (DFO) (Di Pietro et al., 2009; Knaapen et al., 2002). This suggests that the content of soluble transition metals such as iron was not the most important constituent in particles for the formation of DNA strand breaks.

It has been shown that acetonitrile-extracted material from PM_2.5 that was collected from a location close to heavy traffic had higher potency in generating DNA strand breaks in fibroblasts compared with PM_2.5 samples from other urban zones (Abou Chakra et al., 2007). A comparative investigation of organic extract of PM_2.5 from samples collected at a highway site (with high traffic intensity) showed higher levels of DNA strand breaks in A549 cells compared with extracts from an urban site (with medium traffic intensity) and an industrial site near a foundry (Bonetta et al., 2009). Also, samples collected in the industrial area of Kaifaqu district, Dalian, China, showed higher potency in generating DNA strand breaks compared with samples from three other areas in China (Jiang et al., 2011). DCM extracts of PM₁₀ samples from urban air in Prague (Czech Republic), Košice (Slovakia), and Sofia (Bulgaria) increased the generation of DNA strand breaks in HepG2 cells, whereas there were no clear spatial or temporal differences in potency (Gábelová et al., 2004, 2007). Another study found a difference between DCM extracts of PM₁₀ from different locations in Saudi Arabia (Elassouli et al., 2007). EOM (hexane and methanol) of TSP and PM₁₀ from an urban background site (a park) in Rome, Italy, had similar potency in generating DNA strand breaks in mononuclear blood cells (Fabiani et al., 2008). Another study in Parma, Italy, showed that EOM of PM_2.5 was more potent than TSP and PM₁₀ in generating DNA strand breaks (Buschini et al. 2001).

Organic extracts of PM_2.5 or PM₁₀ from samples that were collected in three French metropolitan areas in the winter had higher potential to generate DNA strand breaks than extracts collected in the summer (Abou Chakra et al., 2007). Organic extracts of PM₁₀ samples from an industrial area in China were more potent in generating DNA strand breaks in HepG2 cells when collected during the winter (Jiang et al., 2011). The same was shown for EOM of PM_2.5 collected in Hong Kong Special Administrative Region, China; the samples from the winter were more potent than those from the summer (Hsiao et al., 2000). EOM of PM₁₀ particles collected in Teplice, Czech Republic, increased the levels of DNA strand breaks in HepG2 and Caco-2 cells, and the samples collected during the summer had a stronger effect than those collected during the winter (Lazarová & Slamenová, 2004). In addition, EOM of PM_2.5 collected on days with haze had higher potential to generate DNA strand breaks compared with extracts of samples collected on days without haze (Xu et al., 2008a). The association between EOM of PM from China, France, and the Republic of Korea has also been shown in other studies, albeit without assessment of temporal, spatial, or particle size differences (Chen et al., 2013; Oh et al., 2011; Shang et al., 2013; Tarantini et al., 2009).

(iv) Acellular test systems

Studies in acellular test systems are summarized in Supplemental Table S13 (available online). The studies on the ability of PM to generate strand breaks in DNA have typically used relaxation of plasmid or bacteriophage DNA, in which the supercoil structure is relaxed by introduction of DNA strand breaks. The literature on studies of air pollution particles generally shows that PM from air pollution is associated with relaxation of supercoiled DNA. Early studies showed that PM₁₀ from Edinburgh, United Kingdom, increased the relaxation of supercoiled DNA and that this was reduced by treatment with an antioxidant (mannitol) or a metal chelating agent (DFO) (Donaldson et al., 1997; Gilmour et al., 1996). The strand-breaking potential of PM_2.5 samples from Baton Rouge, Louisiana, USA, was reduced in the presence of superoxide dismutase or catalase (Dellinger et al., 2001). The role of iron mobilization was demonstrated by studies showing that SRM 1648 and SRM 1649 were associated with strand breakage in DNA only in the presence of ascorbate, which functions as reductant (Smith & Aust, 1997). One study on coal fly ash also found increased generation of DNA strand breaks (Dwivedi et al., 2012). Particles collected in London, United Kingdom, in 1958 from an unknown site generated strand breaks in a concentration-dependent manner (Whittaker et al., 2004). Studies on different particle size fractions have produced mixed results, showing both higher potency in supercoil relaxation of small particles (Healey et al., 2005; Koshy et al., 2009; Lingard et al., 2005; Reche et al., 2012; Shao et al., 2006) and higher potency of coarse particles than fine particles (Greenwell et al., 2002). In addition, it has been shown that the potency of PM samples collected from a location near a busy motorway and steelworks depended on the wind direction, with the highest potency of strand scission activity observed for PM samples when the wind came from the motorway (Moreno et al., 2004). Another study showed that PM_2.5 from an urban site in Shanghai, China, was more potent in plasmid DNA supercoil relaxation compared with samples from a suburban site and that samples collected during the winter were more potent those collected during the summer (Senlin et al., 2008). Collectively, the studies indicate that aqueous suspensions of PM and water-soluble constituents have the ability to generate DNA strand breaks in naked DNA, which is driven mainly by production of ROS by transition metals.

Airborne particles from many cities in China have been reported to induce plasmid DNA relaxation in acellular conditions. Samples collected during sandstorms were less potent than non-sandstorm samples (Shi et al., 2004). Water extracts or particle suspensions of PM₁₀ samples collected in four seasons in Lanzhou showed the ability to generate strand breaks in plasmid DNA. Average values of TD20 (the level causing 20% of DNA damage) were 17, 625, 56, and 260 µg/mL in the winter, spring, summer, and autumn, respectively, for PM₁₀ suspensions. Water extracts caused slightly lower induction of DNA strand breaks, with higher TD20. Suburban PM₁₀ samples showed higher TD20 values than samples from Lanzhou. Similar to the study in Beijing, particles were collected during dust storm episodes or after days with rain. The results showed lower ability to generate DNA strand breaks (TD20 > 1000 µg/mL) compared with the PM that was collected in Lanzhou, although the TD20 values correlated negatively with metal concentration (Xiao et al., 2009). In Macao Special Administrative Region, China, PM₁₀ samples from three sites (Sun Yat Sen Municipal Park, Avenida de Horta e Costa, and Macao University on Taipa Island) showed that whole PM₁₀ suspension samples caused formation of DNA strand breaks with values of TD30 (the level causing 30% of DNA damage) of 3, 10, and 20 µg/mL, respectively, for the three sites. Water extracts showed slightly higher TD30 values (Shen et al., 2009).

In summary, the majority of human studies have shown positive associations between exposure to particulate air pollution and elevated levels of DNA strand breaks in leukocytes as well as nasal epithelial cells. In animal studies, elevated levels of DNA strand breaks in the lung have been noted in studies on doses of PM by instillation, whereas lower doses have not increased levels of DNA strand breaks. Studies in cultured cells and acellular conditions have provided supporting mechanistic evidence for the ability of outdoor air PM to generate DNA strand breaks.

(i) Humans

The associations between air pollution and biomarkers of oxidatively damaged nucleobases in human leukocytes have been assessed in a variety of biomonitoring studies, including controlled exposure, panel, and cross-sectional studies. A previous assessment of studies measuring oxidized nucleobases highlighted that approximately half of the published studies had either suboptimal study design or measurement of 8-oxodG by unspecific methods (Møller & Loft, 2010). The discussion of the biomonitoring studies adheres to this critical assessment of the studies. The studies on associations between exposure to air pollution particles and levels of oxidatively damaged DNA in cells from humans are listed in Table 4.13.

Table 4.13

Exposure to air pollution and oxidatively damaged DNA in human leukocytes.

The first of two studies in Benin recruited taxi-moto drivers in the city of Cotonou, which has high levels of outdoor air pollution, as determined by assessments including total PAHs (35–103 ng/m³) and urinary excretion of benzene metabolites (S-PMA, 6.8–9.3 µmol/mol creatinine), and a control group in a village with low air pollution (PAHs, 7.3 ng/m³; S-PMA, 4.2 µmol/mol creatinine). This revealed that the taxi-moto drivers had higher levels of 8-oxodG in lymphocytes (21 lesions/10⁶ dG) compared with controls in the village (11 lesions/10⁶ dG) (Ayi-Fanou et al., 2006). [The high background levels of 8-oxodG suggest spurious oxidation of the DNA during the HPLC-electrochemical detection (ECD) measurement, and the study design with comparison of subjects in the city and village is not optimal.] A subsequent study also used taxi-moto drivers in Cotonou and village controls, as well as groups of subjects with intermediate exposure to air pollution as determined by urinary excretion of benzene metabolites and the number concentration of UFP (midday 1-hour average, 6961–265 145 UFP/cm³). There were clear gradients in both air pollution levels (assessed as S-PMA) and levels of formamidopyrimidine DNA glycosylase (FPG)-sensitive sites in PBMCs between the subjects living in areas with different air pollution levels (Avogbe et al., 2005).

A panel study of students who were living in the centre of Copenhagen, Denmark, showed a positive association between personal exposure to PM_2.5 (10–24.5 µg/m³) and levels of 8-oxodG in lymphocytes, whereas the exposure did not correlate with levels of FPG-sensitive sites in lymphocytes (Sørensen et al., 2003a). In addition, there was a correlation between the levels of 8-oxodG in lymphocytes and the concentration of water-soluble transition metals in PM_2.5 that was collected over a 2-day period for each subject (Sørensen et al., 2005). The same study also showed that there was no correlation between background mass concentration of PM_2.5 measured at a stationary monitoring station or personal exposure to NO₂ and levels of 8-oxodG in lymphocytes. Another study of residents of Copenhagen used benzene as a marker of urban air pollution exposure and showed an association between urinary excretion of S-PMA and levels of 8-oxodG in lymphocytes, whereas the levels of endonuclease III (ENDOIII)/FPG sites were unaltered in lymphocytes (Sørensen et al., 2003c). The effect of personal exposure to UFP in air pollution was investigated in people bicycling for approximately 90 minutes in the laboratory or on traffic-heavy streets in Copenhagen. This study showed a positive association between personal exposure to UFP and levels of FPG-sensitive sites in PBMCs (Vinzents et al., 2005). The same group of researchers also studied controlled exposure to air from a busy street in Copenhagen and reported correlations between particles in the size mode with a median diameter of 23 nm (representing SVOCs of diesel exhaust) and the size mode with a median diameter of 57 nm (representing carbonaceous soot) and levels of FPG-sensitive sites in PBMCs (Bräuner et al., 2007).

A study in Florence, Italy, showed no correlation between outdoor ozone concentrations (days 3–30 before sampling) and levels of FPG-sensitive sites in lymphocytes from healthy subjects (Giovannelli et al., 2006). In a subsequent study, a positive correlation was shown between ozone levels (days 60–90 before sampling) and levels of FPG-sensitive sites in lymphocytes of traffic-exposed workers (Palli et al., 2009).

A study in Bangkok, Thailand, with a relatively large benzene exposure gradient (8–50 µg/m³) between traffic police and office-based police showed no association with levels of 8-oxodG, whereas there was a correlation between personal 1,3-butadiene exposure and levels of 8-oxodG in leukocytes (Arayasiri et al., 2010). Another study in Thailand on malondialdehyde–deoxyguanosine (M₁dG) adducts, a biomarker of oxidative stress and lipid peroxidation, showed that residents living in a location near steel, oil refinery, and petrochemical factories had higher levels of DNA adducts in leukocytes (3.7 ± 0.4 lesions/10⁸ nucleotides) than subjects in a location with a low air pollution level (2.9 ± 0.4 lesions/10⁸ nucleotides) (Peluso et al., 2010; Peluso et al., 2012). One study on schoolchildren in Bangkok, compared with children in a provincial area (Chonburi), showed that the children exposed to air pollution had higher levels of 8-oxodG in leukocytes (Buthbumrung et al., 2008). In another study on healthy subjects living in traffic-congested areas in Bangkok, levels of 8-oxodG in leukocytes were significantly correlated with concentrations of individual exposure to PM_2.5 (Vattanasit et al., 2014).

Several studies in the Czech Republic have assessed biomarkers of oxidatively damaged DNA in people with different occupations or living in areas characterized by high or low air pollution levels. A study of police officers showed higher levels of ENDOIII/FPG sites in lymphocytes in the season with a high level of air pollution exposure (PM_2.5, 33 µg/m³), whereas there was no effect in the season with a low level of air pollution exposure (PM_2.5, 15 µg/m³) (Novotna et al., 2007). Studies of bus drivers, garage workers, and office workers (controls) indicated no associations between air pollution measures and levels of ENDOIII/FPG sites in lymphocytes from exposed subjects and controls (Bagryantseva et al., 2010). There was no difference in placental levels of 8-oxodG between mothers living in Teplice (urban area) and those living in Prachatice (rural area), and there was a lack of association between air pollution exposure levels and levels of 8-oxodG in multivariable-adjusted models (Rossner et al., 2011a). This research group also participated in the EXPAH project on associations between air pollution exposures in Prague (Czech Republic), Košice (Slovakia), and Sofia (Bulgaria) and biomarkers of genotoxicity in samples from male police officers, bus drivers, and office workers (Taioli et al., 2007). This study showed that police officers in Košice had higher levels of 8-oxodG in lymphocytes compared with controls, whereas there was no effect in police officers in Prague (Singh et al., 2007b). [The Working Group noted that there were very high levels of 8-oxodG in the reference group (i.e. 54 lesions/10⁶ nucleotides, corresponding to 244 lesions/10⁶ dG), indicating spurious oxidation during the measurement of 8-oxodG.] This study also showed that PAH-exposed subjects from Sofia had higher levels of lipid peroxidation-derived M₁dG adducts in lymphocytes, measured by immunoslot blot, compared with controls from the same city (Singh et al., 2007b).

One study showed that nasal biopsies from children living in areas with a low level of air pollution (a Pacific coastal town) had lower levels of immunostaining intensity for 8-oxodG compared with biopsies from children living in Mexico City, with a high level of air pollution (Calderón-Garcidueñas et al., 1999).

In summary, a substantial number of studies from humans (11 out of 13 studies) have reported positive associations between exposure to air pollution and levels of oxidatively damaged DNA in leukocytes.

(ii) Experimental systems

Few studies have assessed the level of oxidatively damaged DNA in lungs of animals after exposure to air pollution (see Supplemental Table S14, available online). Studies in rodents have reported no effect of intratracheal instillation of PM on levels of oxidatively damaged DNA in lung tissue. Studies in cultured cells (see Supplemental Table S15, available online) and in acellular test systems (see Supplemental Table S16, available online) have examined the induction of oxidatively damaged nucleobases by PM. Aqueous suspensions of PM from urban areas, mainly in Europe, have generated increased levels of oxidatively damaged DNA in cultured cells. There is also some evidence showing that organic extracts of PM are associated with generation of oxidatively damaged DNA in cultured cells. Aqueous suspensions of PM from urban areas, mainly in Europe, have generated increased levels of 8-oxodG in acellular conditions.

(i) In vivo

Several studies evaluated gene expression in lung, brain, or adipose tissue after inhalation exposure of rodents to outdoor air or to concentrated air particles (CAPs) (André et al., 2006; Heidenfelder et al., 2009; Bos et al., 2012; Soberanes et al., 2012; Tablin et al., 2012; Ljubimova et al., 2013; Mendez et al., 2013; Rowan-Carroll et al., 2013) (see Supplemental Table S18, available online). Other studies involved evaluation of gene expression in lung tissue of rats after intratracheal instillation of PM from urban air (Kooter et al., 2005; Wise et al., 2006).

(ii) In vitro

In vitro studies are summarized in Supplemental Table S19 (available online).

(i) DNA methylation

Collectively, the available studies generally show an association between DNA methylation and outdoor air pollution (see Table 4.15). Although air pollution enhances methylation of some genes, it reduces methylation of others. It is likely that this reflects different pathways and involves different components of air pollution (metals, PAHs, VOCs, PM, etc.). Nonetheless, the relatively consistent association between methylation of DNA and exposure to air pollution, resulting in altered gene expression, indicates that this is another mechanism by which air pollution may influence risk of cancer.

Table 4.15

Changes in DNA methylation in humans exposed to polluted outdoor air.

(ii) Leukocyte telomere length

Three studies have evaluated leukocyte telomere length relative to exposure to outdoor air pollution (see Table 4.16).

Table 4.16

Effects on leukocyte telomere length in humans exposed to polluted outdoor air.

Truck drivers in Beijing, China, had longer telomeres than office workers (Hou et al., 2012). For both the truck drivers and the office workers, an increase in telomere length was associated with personal PM_2.5 concentration, personal elemental carbon concentration, and outdoor PM₁₀ concentration on the day that blood was drawn from the subjects. However, shorter telomere length was associated with the PM₁₀ concentration averaged over the 2 weeks before the blood draw. These results indicate that longer telomere length is associated with short-term exposure to outdoor air PM, consistent with an effect of PM on telomeres during acute inflammatory responses. In contrast, long exposures to PM may shorten telomeres due to extended exposures to pro-oxidants.

A population living in Massachusetts, USA, was evaluated for telomere length, which was compared with modelled exposure to carbon black as a marker for traffic-related particles (McCracken et al., 2010). This study found that telomere length shortened as carbon black exposure increased. This result is consistent with the study in China, showing that prolonged exposure to airborne particles is associated with shortened telomeres.

A third study of telomere length found shorter telomeres among traffic officers in Milan, Italy, compared with office workers (Hoxha et al., 2009). Among the traffic officers, the adjusted mean telomere length was shorter in subjects working in high traffic density compared with low traffic intensity. An additional exposure assessment found that telomere length decreased with increasing concentrations of personal exposure to benzene and toluene. Collectively, these studies show that leukocyte telomere length is shortened in subjects chronically exposed to air pollution.