Iron and copper were the most abundant transition metals of ASPM samples. The concentration of iron and copper in the ASPM samples were 37.4 ± 2.4 μg/mg and 2.1 ± .016 μg/mg, respectively (mean ± SD, n = 5). In comparison, there is much lower content of iron and copper in the SRM1650 (0.0031 and 0.020 μg/mg, respectively) and SRM2975 (0.0009 μg/mg and below limit of detection, respectively) 22. It has not been possible to make a thorough assessment of the total concentration of PAH in the ASPM samples because of limited material. However, a recent collection of similar samples at the same location found that the concentration of 8 different PAH was 40 ng/mg 25. The concentration of same PAH in the SRM1650 and SRM2975 preparations has been reported to be 47 and 29 ng/mg, respectively 2021. Collectively, the most pronounced difference between the SRMs and ASPM is the high concentration of transition metals in the latter.

The data on the cytotoxicity of PM, determined as release of LDH, are depicted in figure 1. The figure shows a dose-dependent increase in LDH enzyme activity in the cell culture supernatant at 25, 100 and 250 μg/ml for SRM1650 and at 100 and 250 μg/ml for SRM2975 (p < 0.05, ANOVA). The SRM1650 preparation was more cytotoxic to the cells than SRM2975 at 250 μg/ml (p < 0.01, Mann-Whitney U test) and more cytotoxic than both SRM2975 and ASPM at 25 μg/ml (p < 0.05, Kruskal-Wallis test). The ASPM preparation (all filters pooled) showed no significantly increased cytotoxicity at any doses compared to the control. The cytotoxicity at 250 μg/ml for ASPM was not measured. It should be noted that exposure to 100 μg/ml of ASPM was associated with approximately 20% increased LDH release that was not statistically significant at 5% level, but a similar level of LDH release was statistically significant for the SRM2975 samples. This suggests that the ASPM samples could be associated with cytotoxicity to a similar degree as the SRM2975 preparation.

Overall, the data from the LDH assay suggest that SRM1650 was more cytotoxic than SRM2975 at equal mass concentration and it cannot be ruled out that the ASPM preparations are slightly less cytotoxic.

The results from the colony forming assay depicted in figure 2 demonstrated a moderate loss of ability to form colonies. For SRM1650, the loss was significant for doses at 100 and 250 μg/ml (p < 0.05, Kruskal-Wallis test). For SRM2975, the loss was significant for 2.5 and 250 μg/ml (p < 0.05, Kruskal-Wallis test). Considering the non-significant effects of SRM2975 at 25 and 100 μg/ml, the effect observed in cell cultures exposed at the dose of 2.5 μg/ml could be a chance finding. The ASPM preparation (all filters pooled) showed no significant loss of the ability to form colonies.

Figure 3 depicts the results of the generation of 8-oxodG in calf thymus DNA by the PM preparations. The SRM1650 and SRM2975, with or without presence of H₂O₂, did not generate 8-oxodG in incubations lasting 30 minutes (figure 3A and 3B). Increasing the incubation time to 120 minutes did not show any SRM1650 or SRM2975 mediated oxidation of DNA with or without H₂O₂ (data not shown).

In contrast to the SRM preparations, the ASPM generated 8-oxodG in calf thymus DNA in a dose-depended manner and were more potent in the sense that it required much lower concentration of H₂O₂ to yield effect (10 μM versus 5 mM H₂O₂). Figure 3C shows concentration-dependent increases in the level of 8-oxodG after 30 minutes exposure to the 5 different samples of ASPM. No statistical analysis was made to differentiate the filters because each point represents 1–2 determinations. The results suggest that the different filters had similar ability to generate 8-oxodG and any minor differences did not appear to be associated with the small variation in iron or copper content.

Figure 4 depicts the dose-response and time curves of SB and FPG sites in A549 cells exposed to SRM1650 and SRM2975. There were significantly increased levels of SB and FPG sites after 3, 24 and 48 hours exposure to SRM2975 at doses 25, 100 and 250 μg/ml (p < 0.01, ANOVA). The level of SB was also significantly increased after 48 hours by SRM2975 at dose 2.5 μg/ml (p < 0.05, ANOVA). In cells exposed to SRM1650, the level of SB and FPG sites after 3, 24 and 48 hours were significantly increased at doses 100 and 250 μg/ml (p < 0.01, ANOVA, except SB at 24 hours; p < 0.01, Kruskal-Wallis test). The levels of SB after 48 hours and FPG sites after 24 hours were significantly increased by SRM1650 at dose 25 μg/ml (p < 0.01 and p < 0.05, respectively, ANOVA). The lowest effective SRM1650 dose in the 3 hours exposure experiments was 2.5 μg/ml for both SB and FPG sites (p < 0.05, ANOVA).

Table 1 shows the level of SB and FPG sites in A549 cells exposed for 24 hours to the ASPM samples. There was a general tendency of dose-dependent elevation of DNA damage, but when tested for each filter this only reached statistical significance for filter number 5 (p < 0.05, Kruskal-Wallis test). The level of DNA damage for all filters reached statistical significance at 100 μg/ml (p < 0.05, randomized block ANOVA). Due to limited amount of ASPM, we excluded the highest dose of 250 μg/ml. The exposure time of 24 hours was chosen because 3 hours exposures of SRMs had a narrow linear dose-response relationship and the cell cultures treated with SRM for 48 hours were nearly confluent.

Figure 5 shows the level SB and FPG sites generated by ASPM (all filters), SRM1650 and SRM2975 after 24 hours exposure. For the purpose of comparing the genotoxicity of ASPM and SRMs, the data from figure 4 (250 μg/ml is excluded) and table 1 were baseline adjusted. Comparisons of the effect at doses 25 and 100 μg/ml did not reveal significant differences in SB and FPG sites between the types of PM preparations (p > 0.05, Kruskal-Wallis test).

Samples of SRM1650 and SRM2975 were obtained from the National Institute of Standards and Technology (Gaitersburg, MD, USA). The SRM1650 and SRM2975 are samples of diesel PM collected from the engine of heat exchangers and filtering system of an industrial diesel-powered forklift, respectively 2021. ASPM was collected on Millipore type RA membrane filters, consisting of a gridded membrane filter made from mixed esters of cellulose, as total suspended particulates at a monitoring station on Jagtvej in Copenhagen on five weekdays in September 2000. During this period the annual TSP concentration was <50 μg/m³ at Jagtvej. Based on measurements from the PM10 fraction was has been estimated that 11 μg/m³ of PM₁₀ concentration originated from traffic and the rest (22 μg/m³) derive from road dust, tire wear, road salt and long-range transport of particles. Approximately 75% and 25% of the traffic-generated PM₁₀ originated from diesel and petrol powered vehicles, respectively 37. The particles were collected and analyzed as a part of the Danish Air Quality Monitoring Programme (conducted by the Danish National Environmental Research Institute) and stored in dark containers at room temperature. The collection volume was 60 m³ and the total collected material was 56.5 ± 13.2 μg/cm³ (mean ± SD, n = 5). The elemental composition (Al, Si, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Ga, As, Se, Br, Rb, Sr, Zr, Mo, Sn, Sb, Ba and Pb) was measured by proton induced X-ray emission analysis at the National Environmental Research Institute, Denmark 38. Each filter was placed in a vial with 5 ml of ultrapure water and subjected to ultrasonication for 5 min at room temperature to liberate particles. Stock suspensions of these ASPM samples were stored at -20°C until used in experiments. The concentration of the ASPM stock solutions (μg/ml) was calculated from the total concentrations on the filters, assuming that all material was liberated from the filters. The size distribution was not defined. Stock suspensions with a concentration of 2500 μg/ml of SRM1650 and SRM2975 were prepared in ultrapure water. The suspensions were ultrasonicated and vortexed prior to dilution with ultrapure water or cell medium to the final concentrations.

A549 cells (American Type Culture Collection, Manassas, VA, USA), kindly provided by the National Research Centre for the Working Environment, Denmark, were grown in F12 nutrient mixture (HAM) supplemented with 10% heat inactivated foetal bovine serum, 1% L-glutamine and 1% penicillin-streptomycin (incubation at 37°C, 5% CO₂). All were Gibco^® cell culture products purchased from Invitrogen™, Denmark.

Plates of different sizes have been used in the experiments of LDH, colony forming ability, and comet assay. Exposures are reported as concentrations (μg/ml). For comparison between the experiments in terms of cell area, the highest concentration (250 μg/ml) in the 24 hour incubations corresponds to 132 μg/cm² (comet assay), 78 μg/cm² (colony forming assay), and 166 μg/cm² (LDH assay).

The cytotoxicity of SRMs and ASPM was measured as LDH activity in cell medium by the Cytotoxicity Detection Kit from Roche Applied Science, Penzberg, Germany. An increase in the number of dead or cell membrane-damaged cells increases the LDH activity in the cell culture supernatant. In this assay the well area was 0.3 cm² (96 well culture plates) and the recommended amount of solution was 200 μl in each well.

The ability of a single cell to form a colony of cells after exposure to SRMs and ASPM was determined by the colony forming assay as described previously 39. 3 × 10⁵ cells were seeded into 6 well culture plates (9.6 cm²/well) in 3 ml cell culture medium. After 48 hours of incubation, cells were washed with phosphate buffer saline, and 3 ml SRM1650, SRM2975 or ASPM samples with concentrations corresponding to doses at 0, 2.5, 25, 100 and 250 μg/ml was added. In this assay, suspensions of ASPM were pooled from different filters, because of limited amount of material. Hereafter 200 cells were seeded into 6 well culture plates with 3 ml medium and incubated for 5 days in 5% CO₂ and 37°C. The cells were treated in methanol for 10 minutes and stained with trypan blue for 30 minutes. The number of colonies with more than 50 cells was counted using a magnifying glass.

DNA damage was measured as the formation of SB and FPG sites in the DNA by the comet assay as previously described 1540. For experiments, 10⁵ cells were seeded into 24 well culture plates (1.9 cm²/well), and grown for 48 hours (~80% confluence). The cells were treated with 1 ml SRM stock suspension diluted with cell culture medium to final concentrations; 0, 2.5, 25, 100 and 250 μg/ml or with 1 ml ASPM solution diluted with cell culture medium to final concentrations; 0, 2.5, 25, and 100 μg/ml. The cells were harvested after 3, 24 or 48 hours exposure with trypsin-EDTA (Invitrogen™, Denmark) and finally centrifuged at 3000 rpm for 6 minutes before addition of agarose gel.

The cell culture experiments were carried out on three different days for both SRMs and ASPM. Cells were exposed to the SRMs and ASPM in triplicates/day (N_total = 9) and duplicates/day (N_total = 6), respectively.

The samples were coded before visual scoring, where the level of SB and FPG sites was obtained by scoring 100 nuclei using a five-class scoring system (arbitrary score range: 0–400). The primary comet assay endpoints in arbitrary units were transformed into lesions/10⁶ base pair (bp), using an investigator-specific X-ray calibration curve and assuming that human diploid cells contain 4 × 10¹² Dalton DNA (corresponding to 6 × 10⁹ bp) as reported previously 41.

Different concentrations of ASPM were added to solutions of calf thymus DNA (150 μg/ml) in a total volume of 0.5 ml with 10 μM of H₂O₂ and incubated for 30 min at 37°C. For the determination of 8-oxodG generation by SRM1650 and SRM2975, SRM preparations were added to solutions of calf thymus DNA (300 μg/ml) in a total volume of 1 ml with or without 5 mM of H₂O₂ and incubated for 30 or 120 minutes at 37°C. The exposed DNA was precipitated, subjected to enzyme digestion and high performance liquid chromatography analysis with electrochemical detection (HPLC-EC) as described previously 42. The results are presented as the ratio between 8-oxodG and the normal dG.

All variables were tested for normal distribution using the Shapiro-Wilks test. If groups did not show homogeneity of variance by Levene's test, they were either log-transformed or cubic root transformed before further statistical analysis. Data that did not show homogeneity variance were analysed by the non-parametric Kruskal-Wallis test, with post-hoc Tukey-type multiple comparison test or the Mann-Whitney U test. Data that had homogeneity of variance were tested with ANOVA with post-hoc Fisher test. The dose-response relationship in DNA damage detected by the comet assay of ASPM samples was tested by a randomized block ANOVA analysis; only the mean values for each treatment was included in the test (corresponding to N = 20). The data on comet assay endpoints displayed inter-experiment variation that made it difficult to compare different samples directly. Therefore, the raw data are baseline-adjusted; the mean of the zero dose culture was subtracted from each data point. All variables were tested at the 5% probability significance level. P-values refer to post-hoc tests. The statistical analysis was performed in Statistica 5.5 (StatSoft, Inc., Tulsa, USA).