Heterogeneities in Inflammatory and Cytotoxic ... - Toxicology

Inhalation Toxicology, 19:213–225, 2007 c Informa Healthcare Copyright ISSN: 0895-8378 print / 1091-7691 online DOI: 10.1080/08958370601067863

Heterogeneities in Inflammatory and Cytotoxic Responses of RAW 264.7 Macrophage Cell Line to Urban Air Coarse, Fine, and Ultrafine Particles From Six European Sampling Campaigns Pasi I. Jalava, Raimo O. Salonen, and Arto S. Pennanen National Public Health Institute, Department of Environmental Health, Kuopio, Finland

Markus Sillanpäa¨ Finnish Meteorological Institute, Air Quality Research, Helsinki, Finland

Arja I. Hälinen and Mikko S. Happo National Public Health Institute, Department of Environmental Health, Kuopio, Finland

Risto Hillamo Finnish Meteorological Institute, Air Quality Research, Helsinki, Finland

Bert Brunekreef Institute for Risk Assessment Sciences, Utrecht, and Julius Center for Health Sciences and Primary Care, University Medical Center, Utrecht, The Netherlands

Klea Katsouyanni University of Athens, Department of Hygiene and Epidemiology, Athens, Greece

Jordi Sunyer Institut Municipal d’Investigació Mèdica, Barcelona, Spain

Maija-Riitta Hirvonen National Public Health Institute, Department of Environmental Health, Kuopio, Finland

We investigated the cytotoxic and inflammatory activities of size-segregated particulate samples (particulate matter, PM) from contrasting air pollution situations in Europe. Coarse (PM10-2.5 ), fine (PM2.5-0.2 ), and ultrafine (PM0.2 ) particulate samples were collected with a modified Harvard high-volume cascade impactor (HVCI). Mouse RAW 264.7 macrophages were exposed to the samples for 24 h. Selected inflammatory mediators, nitric oxide (NO) and cytokines (tumor necrosis factor alpha [TNFα], interleukin 6 [IL-6], macrophage inflammatory protein-2 [MIP-2]), were measured together with cytotoxicity (MTT test), and analysis of apoptosis and cell cycle (propidium iodide staining). The PM10-2.5 samples had a much higher inflammatory

Received 31 July 2006; accepted 28 September 2006. The authors are grateful for funding by the EC-FP5 Quality of Life and Management of Living Resources Program (QLK4-CT-200100423), the Academy of Finland (contracts 201701 and 53307) and Tekes (contract 40715/01), and for conduction of the field campaigns by the PAMCHAR partners in the respective cities (www.pamchar.org). Cooperation and help in flow cytometric analyses by Piia Penttinen, MSc, and laboratory assistance by Arja Rönkkö and Reetta Tiihonen (National Public Health Institute, Kuopio, FI), are also highly appreciated. The PAMCHAR project belongs to the COST Action 633 on Particulate matter: properties related to health effects (http://www.cost.esf.org/index.php?id=206&action number=633). Address correspondence to Pasi Jalava, National Public Health Institute, Department of Environmental Health, PO Box 95, FI-70701 Kuopio, Finland. E-mail: [email protected]

213

214

P. I. JALAVA ET AL.

activity than the PM2.5-0.2 and PM0.2 samples, but the PM2.5-0.2 samples showed the largest differences in inflammatory activity, and the PM0.2 samples in cytotoxicity, between the sampling campaigns. The PM2.5-0.2 samples from traffic environments in springtime Barcelona and summertime Athens had the highest inflammatory activities, which may be related to the high photochemical activity in the atmosphere during the sampling campaigns. The PM0.2 sample from wintertime Prague with proven impacts from local coal and biomass combustion had very high cytotoxic and apoptotic activities and caused a distinct cell cycle arrest. Thus, particulate size, sources, and atmospheric transformation processes affect the toxicity profile of urban air particulate matter. These factors may explain some of the heterogeneity observed in particulate exposure-response relationships of human health effects in epidemiological studies.

Urban air thoracic particles (PM10 ; 50% cutoff diameter [D50 ] = 10 μm) and its subfraction, fine particles (PM2.5 ; D50 = 2.5 μm), have been consistently associated in epidemiological studies with morbidity and mortality among susceptible population groups (children, asthmatics, elderly cardiorespiratory patients) (WHO, 2003; Anderson et al., 2004; U.S. EPA, 2004). However, there have been regional heterogeneities in the PM10 exposure-response relationships with the short-term mortality (Samoli et al., 2005) and hospital admissions (Atkinson et al., 2001) in Europe. In addition, other epidemiological studies have shown stronger than average exposure-response relationships with the short-term mortality (Laden et al., 2000; Hoek et al., 2002) as well as cardiovascular (Lanki et al., 2006) and child respiratory (Janssen et al., 2003) effects in association with urban air particles derived from traffic. Moreover, there is evidence of a substantial increase in daily mortality in association with particles derived from small-scale coal combustion (Clancy et al., 2002) and of hospital admissions of children in association with particles derived from poorly controlled steel mill (Pope et al., 1991). Inflammation has been suggested as a major biological mechanism, mediating not only particulate-associated respiratory effects such as exacerbation of asthma and chronic obstructive pulmonary disease (Brunekreef & Holgate, 2002), but also cardiovascular effects such as atherosclerosis, increased blood coagulation, and dysrhythmias (Brook et al., 2004; Pope et al., 2004). It has also been suggested that urban air coarse thoracic particles (PM10-2.5 ; 10 μm > particle diameter [Dp ] > 2.5 μm) are associated even more strongly than PM2.5 with respiratory hospital admissions due to their stronger inflammatory activity (Brunekreef & Forsberg, 2005). Recent studies on cultured macrophages have shown that the inflammatory activity of urban air PM10-2.5 is higher than that of PM2.5 , which in turn seems to have a higher activity than ultrafine particles (D50 ≈ 0.1–0.2 μm) (Jalava et al., 2006; Becker & Soukup, 2003; Becker et al., 2003). The biological activity of urban air particles in different size ranges is likely to vary in a complex manner in relation to emission sources, geographical location, and meteorology, but there have been few toxicolog-

ical studies challenging this issue. Major seasonal variations in the inflammatory and cytotoxic activities of PM10 (Monn et al., 2003; Salonen et al., 2004) and of coarse and fine particles (Dybing et al., 2004) have been shown in cultured macrophages and respiratory epithelial cells. Our recent data have also revealed that the inflammatory and cytotoxic properties of sizesegregated particulate matter in macrophages vary with shortterm air pollution situations (Jalava et al., 2006). Moreover, road tunnel PM10 (Hetland et al., 2004) and Utah Valley total suspended particulates (Molinelli et al., 2002) have induced strong inflammatory effects on human respiratory epithelial cells. In the multinational PAMCHAR project (Chemical and biological characterisation of ambient air coarse, fine, and ultrafine particles for human health risk assessment in Europe), we investigate European contrasts in the toxicological characteristics of PM10-2.5 , fine (PM2.5-0.2 ; 2.5 μm > Dp > 0.2 μm) and ultrafine (PM0.2 ; D50 < 0.2 μm) particles. Six sampling sites across Europe were chosen to represent different source environments and seasons of public health interest. The present study focused on toxicological comparison of the size-segregated urban air particulate samples between the six sampling campaigns as follows: (1) cytotoxic (MTT test, cell cycle analysis) and inflammatory (interleukin-6 [IL-6], tumor necrosis factor alpha [TNFα], macrophage inflammatory protein-2 [MIP-2], NO) activities in the mouse RAW 264.7 macrophages, and (2) association of these activities with the previously identified main chemical components and sources of fine and coarse particulate matter in the present campaigns. METHODS Sampling Campaigns and Particulate Sources A series of 7-week sampling campaigns were conducted in six European cities: Duisburg (October 4–November 21, 2002), Prague (November 29, 2002–January 16, 2003), Amsterdam (January 24–March 13, 2003), Helsinki (March 21–May 12, 2003), Barcelona (March 28–May 19, 2003) and Athens (June 02–July 21, 2003). The periods of sampling campaigns included seasons with high particulate concentrations or suspected enhanced health effects in populations. The sampling sites were located in urban background areas as previously described in detail (Sillanpäa¨ et al., 2005). According to chemical mass closure, made on the basis of in-depth chemical analyses of virtual impactor samples collected in the present sampling campaigns (Sillanpäa¨ et al., 2006), the major components of PM2.5 in all sites are carbonaceous compounds (organic matter + elemental carbon) and secondary inorganic ions (SO4 2− , NO3 − , NH4 + ), whereas those in PM10-2.5 are crustal material, carbonaceous compounds, sea salt, and nitrate. The sources of the large contribution of organic matter (21–54% of total mass) to PM2.5 , have been analyzed with chemical tracers (Sillanpäa¨ et al., 2005). The impact of automotive traffic on the fine organic matter is highest in Duisburg and Barcelona followed by Amsterdam and Athens. In addition, there are substantial contributions

215

HETEROGENEITIES OF PARTICULATE TOXICITY

from metal industries (Duisburg > Barcelona), coal combustion (Prague > Duisburg > Amsterdam > Barcelona), fuel oil combustion most likely in ships at city harbors (Barcelona > Helsinki > Amsterdam), and biomass combustion (Prague > Amsterdam > Duisburg). Finally, atmospheric photochemical activity seems to have the highest impact on fine organic matter in Barcelona followed by Athens and Helsinki. The source contributions to the smaller but still substantial organic matter of PM10-2.5 (9–27% of total mass) have not been equally well determined as for PM2.5 (Sillanpäa¨ et al., 2005). However, the coappearance of higher contributions of elemental carbon in autumn and winter campaigns (2.3–5.5%) than in spring and summer campaigns (0.96–1.6%) suggests that part of coarse organic matter originates from local sources of incomplete coal and biomass combustion and that a smaller part was from traffic exhausts and oxidation of volatile organic compounds (Sillanpäa¨ et al., 2006). Other major sources include biological debris, pollens, fungal spores, etc. The particulate samples for toxicological studies were collected in alternating 3- and 4-d periods (n = 14) in each city. The collection was made with a modified Harvard high-volume cascade impactor (HVCI) to enable a high-efficiency, representative sampling in different particulate size ranges (Sillanpäa¨ et al., 2003). The PM10-2.5 and PM2.5-0.2 samples were collected on polyurethane foam (PUF; antistatic polyurethane foam 87035K13, McMaster-Carr, New Brunswick, NJ), while the PM0.2 samples were collected on glass fiber filters (Munktell MGA, Munktell Filter AB, Grycksbo, Sweden). A summary of air quality during the campaign periods is presented in Table 1. The average solar radiation and ozone concentration during the spring and summer campaigns of Helsinki, Barcelona and Athens were much higher than those during the autumn and winter campaigns of Duisburg, Prague and Amsterdam. Sample Preparation for Cell Studies The size-segregated particulate samples were prepared for the cell experiments using previously validated procedures (Jalava

et al., 2005, 2006). Briefly, the sampled PUF strips or quarters of glass fiber filter were extracted with methanol (J. T. Baker HPLC grade, Deventer, The Netherlands) 2 × 30 min in water bath sonicator (FinnSonic m20, FinnSonic Oy, Lahti, Finland) at 20◦ C. The methanol extracts from the particle-loaded substrates of each city were pooled together by size range, and the excess methanol was evaporated with a rotary evaporator. Thereafter, the methanol suspension containing PM0.2 particles was filtered (Schleicher & Schuell FP 30/0.2 CA-S filter, pore size 0.2 μm, Dassel, Germany) to remove glass fibers derived from backup filters. Finally, the concentrated suspension was divided into sample tubes on mass basis and the tubes were dried under nitrogen (99.5%) flow and stored at –20◦ C. The masses of the PM0.2 samples were calculated from the net filter weights but represented mostly the soluble fraction. A similar procedure to the size-segregated particulate samples was used in preparation of the corresponding pooled blanks (Jalava et al., 2006). Cell Culture A mouse macrophage cell line RAW264.7 obtained from American Type Culture Collection (ATCC, Rockville, MD) was cultured in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1% L-glutamine, and 1% penicillin– streptomycin (Gibco BRL, Paisley, UK). The cells were cultured at 37◦ C and 5% CO2 atmosphere. For the experiments, the cell suspension was diluted to 5 × 105 cells/ml. One hour before the experiments, 1 ml fresh medium (37◦ C) was changed on the 6well plates (Costar, Corning, NY). The cells were cultured for 24 h in the experiments. Experimental Setup Before the cell experiments, sample and blank tubes were sonicated for 30 min in a water bath sonicator (FinnSonic m03) to suspend the size-segregated particulate samples into water (Sigma W1503, St. Louis, MO) at a concentration of 5 mg/ml. The exposures were made on equal mass basis to all the particulate samples. The macrophage cell line was separately exposed to the PM10-2.5 , PM2.5-0.2 , and PM0.2 samples at 4 doses of 15, 50,

TABLE 1 Mean values of gaseous air pollutants and meteorological parameters during the 7-week sampling campaigns in 6 cities Dui-aut CO (mg/m3 ) NO2 (μg/m3 ) O3 (μg/m3 ) SO2 (μg/m3 ) T (◦ C) RH (%) Rain (mm) Radiation (W/m2 )

0.5 34 17 10 9 88 88 42

Pra-win 0.7 33 21 16 −2 87 47 n.a

Ams-win 0.6 43 22 5 4 82 60 68

Note. n.a., Not available; largest response in each parameter is given in boldface type.

Hel-spr

Bar-spr

Ath-sum

0.3 24 63 4 4 67 40 157

0.6 43 49 3 15 64 11 225

0.5 37 85 4 29 43 0 320

216

P. I. JALAVA ET AL.

150, and 300 μg/ml for 24 h. Three independent experiments were made in duplicate. After adding the sample suspension, the volume of each well was adjusted to 2 ml by adding fresh medium. To be sure that the particulate collection materials were inert, each plate had also an untreated cell control and a PUF or filter blank at a volume corresponding to the dose of 150 μg/ml of the actual particulate samples. After exposing the macrophages to the particulate samples for 24 h, the cells were resuspended into cell culture medium by scraping. Viability of the cells was measured with the MTT test from cell suspension (2 × 100 μl) of each well. The remaining cell suspension was centrifuged (5 min, 8000 rpm, +4◦ C) (Heraeus Biofuge Fresco, Kendro, Osterode, Germany) to separate the cells and the medium. Supernatants were stored at −80◦ C for cytokine analysis (Jalava et al., 2005). The cells were washed, suspended in phosphate-buffered saline (PBS), and fixed in 70% (v/v) ethanol for further propidium iodide staining (Penttinen et al., 2005).

fate (SDS) extraction buffer (100 μl) was added and incubated overnight at 37◦ C. The MTT test measures the ability of the cells to transform MTT to formazan, which can be spectrophotometrically detected at a wavelength of 570 nm with the microplate reader. Viability of the cells was calculated as percentage by comparing absorbances from cell suspensions exposed to sizesegregated particulate samples with those from corresponding control cell suspensions.

Nitric Oxide Analysis Nitric oxide was measured from the centrifuged (10,000 rpm for 3 min at +4◦ C, Heraeus Biofuge Fresco) cell culture medium as the stable metabolite nitrite (NO2 − ) by the Griess method (Green et al., 1982). Griess reagent, which forms a colored chromophore with nitrite, was mixed 1:1 with 50 μl cell culture medium in wells of a 96-well plate, excluding blank wells. Absorbances were spectrophotometrically measured with a microplate reader (iEMS reader MF, Thermo Labsystems, Helsinki, Finland) at wavelength 540 nm. Results were quantified by comparing measured absorbances with a standard curve of sodium nitrite (Jalava et al., 2005). The possible interference of the particulate samples per se (e.g., their nitrate content) with the spectrophotometric nitrite analysis was tested from particulate suspensions in cell culture medium. Even at the highest particulate mass dose (300 μg/ml), the readings remained below or around the control levels measured from a 24-h incubation of blank samples with macrophages.

Statistical Analysis The results from exposures to actual particulate samples were tested against corresponding blanks as well as with regard to particle dose. The dose-response trends were mathematically tested in automatic fitted regressions of the measured values. The equations giving the best fits were compared between the sampling campaigns within each particulate size range. Levene’s test for equality of variances was used before analyzing the data with the analysis of variance (ANOVA). Paired comparisons between the different particulate doses within each size range and sampling campaign were made by Dunnett’s test considering first level as control. ANOVA and Tukey’s test were used for comparison of the differences between the sampling campaigns. Correlations between the different toxic responses (NO, cytokines, cell viability, and apoptosis) were tested at the best representative particulate dose of 150 μg/ml using Pearson’s correlation. All the differences were regarded as statistically significant at p < .05. The data were analyzed using the SPSS version 13.0 (SPSS, Inc., Chicago).

Cytokine Analysis The proinflammatory cytokines, TNFα and IL-6, and the chemokine MIP-2 were immunochemically analyzed from cell culture medium, using commercial enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems, Minneapolis, MN) according to manufacturer’s instructions as earlier described in detail (Jalava et al., 2005).

RESULTS Size-Segregated Particulate Mass Concentrations The mean HVCI-PM mass concentrations in three size ranges during the sampling campaigns are shown in Figure 1. The highest HVCI mass concentrations of PM10-2.5 and PM0.2 were measured in the Athens campaign, while the PM2.5-0.2 concentration was highest in Prague. The lowest PM2.5-0.2 and PM0.2 concentrations were measured in the Helsinki campaign, but the PM10-2.5 concentration was lowest in Prague. The wet and cool autumn and winter seasons in Duisburg, Prague, and Amsterdam favored a high PM2.5-0.2 -to-PM10-2.5 ratio (2.08–4.25), whereas the ratio was low (0.62–0.65) during the warmer and drier spring and summer campaigns in Helsinki, Barcelona, and Athens.

Cell Viability Analysis Viability of the macrophages was detected with the MTT test. Twenty-five microliters of MTT-solution was added to 100 μl cell culture medium in wells of a 96-well plate. The plate was incubated for 2 h at 37◦ C. Thereafter, socium dodecyl sul-

Analysis of Apoptosis and Cell Cycle DNA content was analyzed by propidium iodide (PI) staining of permeabilized cells, in which the cells containing fragmented DNA can be identified as apoptotic cells (SubG1). This method provides also information about the cell cycle of nonapoptotic cells. After PI staining, 10,000 cells per sample were analyzed with a flow cytometer (CyAn ADP, DakoCytomation, CO), using Summit software version 4.2. (Penttinen et al., 2005).


217

FIG. 1. Mean mass concentrations (μg/m3 ) of urban air particulate matter in three HVCI size ranges during the sampling campaigns in six European cities. Inflammatory and Cytotoxic Responses to Size-Segregated Particulate Samples The responses to particulate samples are described by size range for each parameter. Blank samples had negligible activity in all the measured parameters. NO production. Figure 2 shows the NO responses to the six-city particulate samples in macrophage cell line. All the samples in three size ranges induced a statistically significant NO production in macrophages when compared to the correspond-

ing blank samples from each city. In PM10-2.5 , the average NO production across the 6 sampling campaigns was 3.4-fold and 2.9-fold when compared to PM2.5-0.2 and PM0.2 , respectively. The dose-response trends followed quadratic regressions in all the size ranges, resulting in bell-shaped dose-response curves of variable steepness. In PM10-2.5 , the NO responses were smallest to the Helsinki sample (up to 6.3 ± 1.3 μM). When compared to this, all except for the Duisburg sample induced significantly larger

FIG. 2. Nitric oxide (NO)-production (μM) in RAW 264.7 macrophages after a 24-h incubation with four particulate doses (15, 50, 150, and 300 μg/ml) or blank in three size ranges. Each bar shows mean ± SEM of three exposures in duplicate. Asterisk indicates statistically significant difference from control (Dunnett, p < .05). Letters above dose-response bars indicate the other sampling campaigns with significantly smaller mean response to all the doses (Tukey, p < .05): Duisburg (D), Prague (P), Amsterdam (A), Helsinki (H), Barcelona (B), and Athens (G).

218

P. I. JALAVA ET AL.

FIG. 3. Tumor necrosis factor (TNFα) production (ng/ml) in RAW 264.7 macrophages after a 24-h incubation with four particulate doses (15, 50, 150, and 300 μg/ml) or blank in three size ranges. Each bar shows mean ± SEM of three exposures in duplicate. Asterisk indicates statistically significant difference from control (Dunnett, p < .05). Letters above dose-response bars indicate the other sampling campaigns with significantly smaller mean response to all the doses (Tukey, p < .05) compared to other sampling campaigns: Duisburg (D), Prague (P), Amsterdam (A), Helsinki (H), Barcelona (B), and Athens (G). responses (e.g. Athens up to 12 ± 0.5 μM). Also in PM2.5-0.2 , the Helsinki sample induced the smallest NO production (up to 2.5 ± 0.1 μM), and the samples from Prague (up to 4.1 ± 0.3 μM) and Athens (up to 2.7 ± 0.1 μM) produced significantly larger responses. In PM0.2 , the NO responses were smallest for the Barcelona sample (up to 2.6 ± 0.2 μM), and the samples from Prague (up to 5.5 ± 0.5 μM) and Helsinki (up to 3.4 ± 0.5 μM) induced significantly larger responses (Figure 2).

Cytokine production. Figures 3–5 show the production of cytokines induced by the six-city samples. In PM10-2.5 , the average cytokine production across the 6 sampling campaigns was 7.8-fold and 83-fold for TNFα, and 4.4-fold and 530-fold for MIP-2, when compared to PM2.5-0.2 and PM0.2 , respectively. The corresponding difference in IL-6 between the PM10-2.5 and PM2.5-0.2 samples was 27-fold, while there were no significant IL-6 responses to the PM0.2 samples. The dose-response trends

FIG. 4. Macrophage inflammatory protein 2 (MIP-2) production (ng/ml) in RAW 264.7 macrophages after a 24-h incubation with 4 particulate doses (15, 50, 150, and 300 μg/ml) or blank in three size ranges. Each bar shows mean ± SEM of three exposures in duplicate. Asterisk indicates statistically significant difference from control (Dunnett, p < .05). Letters above dose-response bars indicate the other sampling campaigns with significantly smaller mean response to all the doses (Tukey, p < .05) compared to other sampling campaigns: Duisburg (D), Prague (P), Amsterdam (A), Helsinki (H), Barcelona (B), and Athens (G).


219

FIG. 5. Interleukin 6 (IL-6) production (ng/ml) in RAW 264.7 macrophages after a 24-h incubation with four particulate doses (15, 50, 150 and 300 μg/ml) or blank in two size ranges. The responses to PM0.2 samples were negligible. Each bar shows mean ±SEM of three exposures in duplicate. Asterisk indicates statistically significant difference from control (Dunnett, p < .05). Letters above dose-response bars indicate the other sampling campaigns with significantly smaller mean response to all the doses (Tukey, p < .05) compared to other sampling campaigns: Duisburg (D), Prague (P), Amsterdam (A), Helsinki (H), Barcelona (B), and Athens (G). for the cytokines followed either quadratic bell-shaped (TNFα, IL-6, MIP-2), quadratic inclining (IL-6), or linear (MIP-2) patterns. All the samples in the PM10-2.5 size range induced statistically significant increases in production of all the cytokines when compared to the blank values. The smallest TNFα production was equally induced by the Duisburg (up to 18 ± 3.1 ng/ml) and Helsinki (up to 19 ± 3.8 ng/ml) samples, but only the Prague (up to 40 ± 13 ng/ml) and Barcelona (up to 45 ± 6.6 ng/ml) samples caused significantly larger responses. Moreover, the smallest MIP-2 production was induced by the Helsinki sample (up to 68 ± 13 ng/ml), but only the Prague sample (up to 301 ± 82 ng/ml) caused a significantly larger response. Also, the lowest IL-6 production was induced by the Helsinki sample (up to 12 ± 2 ng/ml), whereas all, except for the Duisburg sample (up to 20 ± 1.9 ng/ml), caused significantly larger IL-6 responses. There was much more heterogeneity in the PM2.5-0.2 induced cytokine responses between the sampling campaigns than there was with the PM10-2.5 samples. All the PM2.5-0.2 samples, except that from Prague caused a statistically significant increase in production of all the cytokines. The Prague sample was least active in production of the cytokines, reaching the level of statistical significance only in MIP-2 production. All the other samples (highest Barcelona up to 6.8 ± 1.3 ng/ml) in-

duced a significantly larger TNFα production than the Prague sample (up to 0.6 ± 0.03 ng/ml). In MIP-2 production, the samples from Barcelona and Athens induced significantly larger responses than the Prague sample. Moreover, in IL-6 production the response to the Prague sample (up to 0.1 ± 0.01 ng/ml) was significantly smaller than the responses to the Duisburg, Athens, and Barcelona (up to 1.5 ± 0.3 ng/ml) samples. In the PM0.2 size range, only the samples from Amsterdam and Athens caused a small, statistically significant TNFα production in macrophage cell line, whereas a significant MIP-2 production was detected only after exposure to the samples from Prague and Helsinki. The IL-6 responses to the PM0.2 samples were negligible. Cytotoxicity. Figure 6 shows the viability of macrophages assessed with the MTT test after a 24-h exposure to the sizesegregated 6-city samples. All the PM0.2 samples, except for those from Helsinki and Athens, decreased significantly the cell viability when compared to unexposed controls. The average cytotoxicity of the PM10-2.5 and PM2.5-0.2 samples was roughly equal, but the PM0.2 samples were less cytotoxic with the exception of the Prague sample. The dose-response trends to most of the samples were linearly declining. In contrast, in the PM0.2 size range, only one or two largest doses caused a significant decrease in cell viability.

220

P. I. JALAVA ET AL.

FIG. 6. The decrease in viability of RAW 264.7 macrophages assessed with the MTT test after a 24-h incubation with four doses (15, 50, 150 and 300 μg/ml) of particulate samples or blank in three size ranges. Each bar shows mean ± SEM of three exposures in duplicate. Asterisk indicates statistically significant difference from control (Dunnett, p < .05). Letters above dose-response bars indicate the other sampling campaigns with significantly smaller mean response to all the doses (Tukey, p < .05) compared to other sampling campaigns: Duisburg (D), Prague (P), Amsterdam (A), Helsinki (H), Barcelona (B), and Athens (G). In the PM10-2.5 size range, the Helsinki sample had the lowest cytotoxicity (viability down to 43 ± 2.3% of control), and all except for the Duisburg sample were significantly more cytotoxic. In PM2.5-0.2 , the Amsterdam sample was the least cytotoxic (viability down to 62 ± 3.2% of control), but only the Athens sample had a significantly higher cytotoxicity (viability down to 33 ± 2.2% of control). In PM0.2 , the Athens sample had the lowest cytotoxicity (viability down to 77 ± 6.7% of control), but only the sample from Prague was significantly more cytotoxic (viability down to 26 ± 5.8% of control).

Figure 7 shows the cell cycle effects and apoptosis induced by a 24-h exposure to a single dose of 150 μg/ml of sizesegregated particulate samples in the macrophage cell line. All the samples significantly increased apoptosis when compared to control. This was indicated by a significantly increased portion of cells in the SubG1 phase (control 1.9 ± 0.4%). There were substantial differences between the sampling campaigns with regard to the cell cycle effects and apoptosis induced by the PM10-2.5 and PM0.2 samples, whereas the differences were small for the PM2.5-0.2 samples. In the PM10-2.5 size range, the samples

FIG. 7. The different phases of cell cycle in RAW 264.7 macrophages presented as percentage of 10,000-cell counts after a 24-h incubation with 150 μg/ml of the particulate samples in three size ranges. Each bar shows mean ± SEM of three exposures. +, Statistically significant positive difference when compared to corresponding cell cycle phase in control cells (Dunnett, p < .05); –, statistically significant negative difference when compared to corresponding cell cycle phase in control cells (Dunnett, p < .05). Unexposed controls are the same for all the size ranges.


from Duisburg (5.5 ± 0.8% of cells in SubG1 phase) and Prague (6.6 ± 0.9%) had the lowest apoptotic activity, while the Athens sample induced the most pronounced apoptosis (25.7 ± 2.0%). In the PM2.5-0.2 size range, the portion of apoptotic cells varied between 10.7 ± 1.1% in Duisburg and 18.5 ± 2.7% in Helsinki In the PM0.2 size range, the Helsinki sample had the lowest apoptotic activity (17.5 ± 1.3%), while the Barcelona sample was the most active (30.2 ± 1.1%) affecting also the synthesis and mitosis of the cells. Most interestingly, the PM0.2 sample from Prague caused not only apoptosis but also a cell cycle arrest in the macrophage cell line. Thus, there were greatly increased portions of cells in the SubG1 (22.4 ± 2.6%) and premitotic/mitotic G2/M (25.0 ± 0.9%) phases, and a substantially lower portion of cells in the G1 phase (22.1 ± 1.1%) than with the other PM0.2 samples (range 38.4–48.9%) (Figure 7). Correlations Between Different Toxic Responses There was a statistically significant positive correlation of IL-6 production with both TNFα (r = .88) and NO (r = .91) productions, and of TNFα with MIP-2 production (r = .92), in the PM10-2.5 size range (n = 6). In the PM2.5-0.2 size range, there was a significant correlation of IL-6 with TNFα (r = .92; n = 6). In the PM0.2 size range, the correlations could not be calculated for cytokines due to minimal responses. There was no statistically significant correlation of NO or cytokine production with cytotoxicity (MTT test) in any particulate size range (r = .20 . . . 0.80; n = 6). Apoptotic activity did not correlate significantly with any other measured parameter. DISCUSSION This study chose both the inflammatory activity and cytotoxicity as potentially important endpoints in health effects of outdoor air particles. We observed that the PM10-2.5 samples were more active per unit of mass than the PM2.5-0.2 samples in induction of cytokine production in the mouse macrophage cell line, whereas the particulate size range was not as important in the induced NO production or cytotoxicity assessed by the MTT test. This is a consistent feature that agrees with our previous study on size-segregated particles collected in the late summer– early autumn campaign of 2002 in Helsinki (Jalava et al., 2006). Moreover, the inflammatory responses to the PM2.5-0.2 samples showed larger heterogeneity between the sampling campaigns than the corresponding responses to the PM10-2.5 samples. In contrast, the largest heterogeneity in the cytotoxic activity assessed by the MTT test, and in the apoptotic activity, was observed between the samples in PM0.2 size range. Particulate Size-Related Effects Particulate size range was a major determinant with regard to the inflammatory activity of the collected samples in macrophage cell line. However, the dose-response trend curves within the PM10-2.5 and PM2.5-0.2 size ranges varied both in shape and maximum response between the sampling campaigns. These

221

variations were probably due to differences in the physicochemical characteristics of the complex mixture samples (e.g., finer particulate size distribution, solubility, chemical composition). The PM0.2 samples represented mostly the soluble extract due to the fact that after extraction with methanol the PM0.2 suspension was filtered in order to remove glass fiber fragments released from the backup filters. We could see under the microscope that the insoluble carbon-like particles were attached to the glass fibers and thus were removed from the suspension. However, some PM0.2 samples induced larger NO responses and higher cytotoxicity than the PM2.5-0.2 samples from the same campaigns. In our previous study on four particulate size ranges from Helsinki (Jalava et al., 2006), not only the PM0.2 samples, collected on the backup filter, but also the accumulation size range particles (PM1-0.2 ; 1 μm > Dp > 0.2 μm), collected on PUF, had a much lower inflammatory activity than the larger intermodal size range (PM2.5-1 ; 2.5 μm > Dp > 1 μm) and PM10-2.5 samples. All the comparisons between toxic activities of particulate samples cited in this study have been made on a mass basis. The PM10-2.5 samples induced the largest production of NO in the mouse macrophage cell line. There was up to a fourfold difference when the responses to the PM10-2.5 samples were compared with those of the roughly equipotent PM2.5-0.2 and PM0.2 samples. The present NO responses in macrophages are well in line with our earlier findings on PM10 samples (Salonen et al., 2004) and size-segregated particulate samples (Jalava et al., 2006) from Helsinki. The particulate size range had the largest impact on MIP-2 production, as shown in the difference between the average responses to the PM10-2.5 and PM2.5-0.2 samples. The corresponding differences in TNFα and IL-6 production were smaller. The PM0.2 samples induced negligible cytokine production. These findings are supported by several previous in vitro studies showing that coarse particles induce larger cytokine production than fine particles in macrophages and respiratory epithelial cells (Hetland et al., 2005; Dybing et al., 2004; Becker & Soukup, 2003; Becker et al., 2002, 2003; Monn & Becker, 1999). Furthermore, the in vitro cytokine responses to fine particles have been larger than those to ultrafine particles in human alveolar macrophages (Becker et al., 2003) and human airway epithelial cells (Becker et al., 2005). Our finding that fine and coarse particles induce a larger production of MIP-2 than TNFα in macrophages is also supported by previous studies (Jalava et al., 2006; Imrich et al., 2000). Larger productions of TNFα than IL-6 have also been found by Pozzi et al. (2003) in RAW 264.7 macrophages, and by Osornio-Vargas et al. (2003) in murine monocytes. Altogether, the cytokine responses in our present study are in good agreement with those reported previously. All the PM10-2.5 samples were more cytotoxic than the particulate samples in smaller size ranges, but there was much more heterogeneity between the sampling campaigns with regard to the cytotoxic activities of the PM2.5-0.2 and PM0.2 samples. Our present results on PM10-2.5 samples being more cytotoxic than

222

P. I. JALAVA ET AL.

PM2.5-0.2 samples are supported by some previous studies (Monn & Becker, 1999; Osornio-Vargas et al., 2003). In the study of Hetland et al. (2004), urban air coarse and ultrafine particles were more cytotoxic than fine particles in both the human and rat alveolar type 2 cells. Moreover, particulate samples in all these size ranges induced apoptosis in human cells. Also, urban air total suspended particulates have induced apoptotic responses (Upadhyay et al., 2003). Interestingly, the highest apoptotic activity in the macrophage cell line was connected to the PM0.2 samples, although the average decrease in cell viability (MTT test) was somewhat smaller for these samples when compared to the PM2.5-0.2 and PM10-2.5 samples. It has been recently shown on bioaerosol particles in the RAW 264.7 cells that the MTT test results do not correlate with apoptosis (Penttinen et al., 2005). Moreover, apoptosis in the human alveolar type 2 cell line A549 has not correlated well with the decrease in cell viability assessed with fluorescence microscopy after exposure to urban air particles (Hetland et al., 2004). Thus, it is obvious that various cell viability tests, like the MTT test, measure nonspecifically both apoptosis and necrosis. Particulate doses in the present study may seem high when compared to average surface doses per square centimeter in the

human lungs after usual short-lasting urban air exposures. However, the surface doses within the lungs can exhibit larger than thousandfold variations in respiratory patients due to uneven particle deposition (Phalen et al., 2006). We emphasize that the use of this kind of doses is required to make it possible to show statistically significant differences in inflammatory activity between the particulate samples in a reasonable number of experiments. Comparison Between Sampling Campaigns and Associations with Sources The cytotoxic and inflammatory activities of the sizesegregated particulate samples showed heterogeneities between the sampling campaigns, which most likely reflected impacts on the complex ambient air mixtures by differing particulate source contributions and seasonal meteorological phenomena. In Table 2, the relative inflammatory and cytotoxic activities of the best representative dose (150 μg/ml) from the dose-response curves are compared between the sampling campaigns per unit of particulate mass as well as per cubic meter of urban air. The latter approach connects the present toxicological results to the average particulate pollution situation in each campaign. This is of

TABLE 2 Size-segregated particulate mass concentrations in urban air and relative toxic responses to particulate samples collected in three size ranges from 6 sampling campaigns

PM0.2

PM2.5-0.2

PM10-2.5

Campaign

PM conc.

Dui Pra Ams Hel Bar Ath Dui Pra Ams Hel Bar Ath Dui Pra Ams Hel Bar Ath

2.8 4.7 3.8 2.7 4.5 6.7 15.8 25.1 22.8 8.3 14.3 18.9 7.6 5.9 9.8 12.8 22.9 29.6

Relative response per μg of PM mass/per m3 of air NO 1.2 1.7 1.1 1.5 1.0 1.2 1.0 1.4 1.0 1.0 1.1 1.2 1.3 1.4 1.9 1.0 1.9 1.9

IL-6 1.0 2.4 1.2 1.2 1.3 2.4 1.9 4.2 2.8 1.0 1.9 2.8 1.2 1.0 2.2 1.5 5.2 6.8

n.a n.a n.a n.a n.a n.a 6.7 1.0 5.9 3.3 17.6 17.6 2.0 2.7 3.5 1.0 3.1 2.8

TNFα n.a n.a n.a n.a n.a n.a 4.2 1.0 5.3 1.1 10.0 13.2 1.2 1.3 2.6 1.0 5.6 6.5

n.a n.a n.a n.a n.a n.a 3.7 1.0 3.4 5.0 7.8 7.8 1.0 2.3 2.5 1.0 2.4 1.8

n.a n.a n.a n.a n.a n.a 2.3 1.0 3.1 1.6 4.4 5.9 1.0 1.8 3.2 1.7 7.1 6.9

MIP-2 n.a n.a n.a n.a n.a n.a 2.2 1.0 1.8 4.3 3.5 9.0 1.0 3.1 2.6 1.2 2.3 2.4

n.a n.a n.a n.a n.a n.a 1.4 1.0 1.6 1.4 2.0 6.8 1.0 2.4 3.3 2.0 6.9 9.3

Cytotoxicity 15.5 32.8 5.4 1.0 4.6 1.1 1.0 2.2 1.3 1.8 2.5 3.7 1.2 1.4 1.2 1.0 1.2 1.1

22.5 47.7 7.8 1.0 6.7 1.0 1.1 3.8 2.0 1.0 2.4 4.8 1.1 1.0 1.5 1.6 3.3 4.1

Note. In the left column for each studied parameter, the value 1 is given to the smallest response to an equal mass dose (150 μg/ml) of coarse (PM10-2.5 ), fine (PM2.5-0.2 ) and ultrafine (PM0.2 ) particulate samples. In the right column for each parameter, a similar comparison of toxic activities is made per cubic meter of urban air. n.a., Not applicable for this comparison due to negligible response. Values in boldface are the largest response within each studied parameter and particulate size range.


special importance, since the size-segregated particulate concentrations were highly different between the sampling campaigns. The PM0.2 samples from Prague and Duisburg had very high relative cytotoxic activities in the MTT test when compared to the other sampling campaigns (Table 2). In both of these campaigns, the fine particulate organic matter contents have been relatively high (54% for Prague and 31% for Duisburg) in parallel low-volume samples (Sillanpäa¨ et al., 2006). In addition, tracers of specific sources have indicated that the highest source contributions to the particulate organic matter of these campaigns were by biomass (levoglucosan) and coal (As) combustion (Sillanpäa¨ et al., 2005). In contrast to all the other samples, the PM0.2 sample from wintertime Prague caused also a cell cycle arrest in the G2/M phase. This may be due to a much higher contribution of PAH compounds to the particulate mass in Prague than in the other campaigns, resulting from prevalent local combustion of the solid fuels (Sillanpäa¨ et al., 2004). This interpretation is supported by several previous studies showing effects of PAH compounds on the cell cycle in various cell types. Dibenzo[a,l]pyrene has caused cell cycle arrest in the G2/M phase of human mammary carcinoma MCF-7 cells, allowing the DNA repair mechanisms (Baird et al., 1999), and dibenzo[a,l]pyrene and benzo[a]pyrene have shown delay on the S phase of human lung fibroblast HEL-cells (Binkova et al., 2000). Moreover, polar organic and aromatic compounds in particulate matter have caused mitochondria-dependent apoptosis in the RAW264.7 cells via oxidative stress (Xia et al., 2004). In line with these findings, it has been previously reported that high PAH content of air particles is associated with high genotoxicity and oxidative stress (Binkova et al., 2003; Farmer et al., 2003). The PM2.5-0.2 samples showed large differences in their activities to induce cytokine production. The samples from Athens and Barcelona were several times more active than the other PM2.5-0.2 samples, and these samples showed also the highest cytotoxicity, along with the Prague sample, which, however, had the lowest inflammatory activity (Table 2). According to chemical tracers, the distinct sources of fine particulate organic matter in the Athens and Barcelona samples have been atmospheric photochemical processes (secondary organic acids). In Barcelona, there have also been contributions from automotive traffic (elemental carbon, Cu), and fuel oil (Ni, V) and coal (As) combustion (Sillanpäa¨ et al., 2005). The PAH contents of particles from these Mediterranean cities have been very low compared to the contents in the Prague and Duisburg samples (Sillanpäa¨ et al., 2004). This may be due to a rapid photooxidation of these compounds to more reactive compounds such as quinones in the warm and sunny Mediterranean atmosphere, as reviewed by the WHO IPCS (WHO, 1998). In two previous studies with samples from different sampling sites (Rome, Oslo, Amsterdam, Lodz) and seasons in Europe, there have been heterogeneities in IL-6 and TNFα production by both the fine and coarse particulate samples in rat alveolar macrophages (Hetland et al., 2005) and rat alveolar type 2 cells (Dybing et al., 2004). The cytokine responses have been generally higher to the spring-

223

and summertime samples than to the wintertime samples, but these studies do not contain information on associations with different sources of particulate mass. The inflammatory (NO, cytokines) and overall cytotoxic activities (MTT test) of the PM10-2.5 samples per unit of mass had much smaller differences between the sampling campaigns than what was observed for the PM2.5-0.2 samples (Table 2). The only consistent feature was the lowest activity of the Helsinki PM10-2.5 sample in nearly all measured parameters. However, due to the highest actual mass concentrations of PM10-2.5 , the Mediterranean coarse particles were assessed to have the highest inflammatory and cytotoxic activities per cubic meter of urban air in the present study. It has been shown on human alveolar macrophages that the fall season coarse particles induce larger IL-6 production than coarse particles from other seasons in Chapel Hill, NC. However, the production of reactive oxygen species has been shown as highest in summer (Becker et al., 2005). It has also been observed that coarse particles collected during the summer season in different locations of Europe are, in most cases, more active inducers of MIP-2 production than coarse particles collected during the winter season (Dybing et al., 2004). Thus, it remains uncertain whether atmospheric photochemical activity could enhance the inflammatory activity of coarse particulate samples. The PM10 samples from four Swiss sampling sites and seasons have shown heterogeneities with regard to induction of NO and cytokine production and cytotoxicity in rat alveolar macrophages (Monn et al., 2003). The filter extracts from winter season samples have, in most cases, been less active NO and TNFα producers than the extracts from the spring, summer, and autumn season samples. However, the cytotoxicity induced by samples from two urban sites has been highest in winter but the winter samples from their Alpine site, Montana, have been least cytotoxic. In the study of Salonen et al. (2004), it has been observed that the spring PM10 in Helsinki is a much more active inducer of cytokines than the winter PM10 but there is no such difference in cytotoxicity. CONCLUSIONS AND IMPLICATIONS Our data presented here showed considerable heterogeneities between the sampling campaigns with regard to the inflammatory activity induced by the fine particulate samples as well as the cytotoxicity and apoptosis induced by the ultrafine particulate samples. It should be noted, however, that these findings do not specifically represent any city or season but rather the prevailing sources of size-segregated particles. Interestingly, the fine and coarse particulate samples from springtime Barcelona and summertime Athens had the highest inflammatory activities, which may be related to their transformed chemical properties due to the high photochemical activity and ozone concentration in the atmosphere during the sampling campaigns. These atmospheric features have also been associated with stronger than average PM10 exposure-response relationships for short-term mortality (Samoli et al., 2005) and hospital admissions (Atkinson et al.,

224

P. I. JALAVA ET AL.

2001) in Europe. In our Mediterranean campaigns, there were high impacts on particulate sample chemical compositions by local traffic. The ultrafine particulate samples, collected in Prague during the heating season with proven impacts from local coal and biomass combustion, had very high cytotoxic and apoptotic activities and caused a distinct cell cycle arrest. This gives insight into the biological mechanisms of local coal combustion derived particles suggested to have stronger than average particulate effects on cardiovascular and respiratory mortality (Clancy et al., 2002). Thus, particulate size, sources, and atmospheric transformation processes affect the toxicity profile of urban air particulate matter in a macrophage cell line. These factors may help explaining some of the heterogeneity observed in particulate exposure-response relationships of human health effects in epidemiological studies. REFERENCES Anderson, H. R., Atkinson, R. W., Peacock, J. L., Marston, L., and Konstantinou, K. 2004. Meta-analysis of time-series studies and panel studies of particulate matter (PM) and ozone (O3 ). Report EUR/04/5042688 of a WHO task group. Copenhagen, Denmark: WHO Regional Office for Europe. Atkinson, R. W., Anderson, H. R., Sunyer, J., Ayres, J., Baccini, M., Vonk, J. M., Boumghar, A., Forastiere, F., Forsberg, B., Touloumi, G., Schwartz, J., and Katsouyanni, K. 2001. Acute effects of particulate air pollution on respiratory admissions: Results from APHEA 2 project. Am. J. Respir. Crit. Care Med. 164:1860– 1866. Baird, W. M., Kaspin, C., Kudla, K., Seidel, A., Greim, H., and Luch, A. 1999. Relationship of dibenzo[a,l]pyrene–DNA binding to the induction of p53, p21(WAF1) and cell cycle arrest in human cells in culture. Polycycl. Aromat. Comp. 16:119–129. Becker, S., and Soukup, J. M. 2003. Coarse (PM2.5-10 ), fine (PM2.5 ) and ultrafine air pollution particles induce/increase immune costimulatory receptors on human blood-derived monocytes but not on alveolar macrophages. J. Toxicol. Environ. Health A 66:847–859. Becker, S., Soukup, J. M., and Gallagher, J. E. 2002. Differential particulate air pollution induced oxidant stress in human granulocytes, monocytes and alveolar macrophages. Toxicol. In Vitro 16:209– 218. Becker, S., Soukup, J. M., Sioutas, C., and Cassee, F. 2003. Response of human alveolar macrophages to ultrafine, fine, and coarse urban air pollution particles. Exp. Lung Res. 29:29–44. Becker, S., Dailey, L. A., Soukup, J. M., Grambow, S. C., Devlin, R. B., and Huang, Y.-C. 2005. Seasonal variations in air pollution particleinduced inflammatory mediator release and oxidative stress. Environ. Health Perspect. 113:1032–1038. ˇ Binková, B., Giguère, Y., Rössner, P., Dostál, M., and Sram, R. J. 2000. The effect of dibenzo[a,l]pyrene and benzo[a]pyrene on human diploid lung fibroblasts: The induction of DNA adducts, expression of p53 and p21WAF1 proteins and cell cycle distribution. Mutat. Res. 471:57–70. ˇ a, M., Pastorovká, A., Jel´ınek, R., Beneˇs, I., Novák, Binková, B., Cern´ ˇ am, R. J. 2003. Biological activities of organic compounds J., and Sr´ absorbed onto ambient air particles: Comparison between the cities of Teplice and Prague during the summer and winter seasons 2000– 2001. Mutat. Res. 525:43–59.

Brook, R. D., Franklin, B., Cascio, W., Hong, Y., Howard, G., Lipsett, M., Luepker, R., Mittleman, M., Samet, J., Smith, S. C., and Tager, I. 2004. Air pollution and cardiovascular disease: A statement for healthcare professionals from the expert panel on population and prevention science of the American Heart Association. Circulation 109:2655–2671. Brunekreef, B., and Forsberg, B. 2005. Epidemiological evidence of effects of coarse airborne particles on health. Eur. Respir. J. 26(2):309– 318. Brunekreef, B., and Holgate, S. T. 2002. Air pollution and health. Lancet 360:1233–1244. Clancy, L., Goodman, P., Sinclair, H., and Dockery, D. W. 2002. Effect of air-pollution control on death rates in Dublin, Ireland: An intervention study. Lancet 360:1210–1214. Dybing, E., Løvdal, T., Hetland, R. B., Løvik, M., and Schwarze, P. E. 2004. Respiratory allergy adjuvant and inflammatory effects of urban ambient particles. Toxicology 198:307–314. Farmer, P. B., Singh, R., Kaur, B., Sram, R. J., Binkova, B., Kalina, I., Popov, T. A., Garte, S., Taioli, E., Gabelova, A., and CebulskaWasilewska, A. 2003. Molecular epidemiology studies of carcinogenic environmental pollutants: Effects of polycyclic aromatic hydrocarbons (PAHs) in environmental pollution on exogenous and oxidative DNA damage. Mutat. Res. 544:397–402. Green, L. C., Wagner, D. A., Glogowski, J., Skipper, P. L., Wishnok, J. S., and Tannebaum, S. R. 1982. Analysis of nitrate, nitrite and [15 N] nitrate in biological fluids. Anal. Biochem. 126:131–138. Hetland, R. B., Cassee, F. R., L˚ag, M., Refsnes, M., Dybing, E., and Schwarze, P. E. 2005. Cytokine release from alveolar macrophages exposed to ambient particulate matter: Heterogeneity in relation to size, city and season. Part. Fibre Toxicol. 2:4. Hetland, R. B., Cassee, F. R., Refsnes, M., Schwarze, P. E., L˚ag, M., Boere, A. J. F., and Dybing, E. 2004. Release of inflammatory cytokines, cell toxicity and apoptosis in epithelial lung cells after exposure to ambient air particles of different size fractions. Toxicol. In Vitro 18:203–212. Hoek, G., Brunekreef, B., Goldbohm, S., Fischer, P., and van den Brandt, P. A. 2002. Association between mortality and indicators of traffic-related air pollution in the Netherlands: A cohort study. Lancet 360:1203–1209. Imrich, A., Ning, Y. Y., and Kobzik, L., 2000. Insoluble components of concentrated air particles mediate alveolar macrophage responses in vitro. Toxicol. Appl. Pharmacol. 167:140–150. Jalava, P., Salonen, R. O., Hälinen, A. I., Sillanpäa¨ , M., Sandell, E., and Hirvonen, M.-R. 2005. Effects of sample preparation on chemistry, cytotoxicity and inflammatory responses induced by air particulate matter. Inhal. Toxicol. 17:107–117. Jalava, P. I., Salonen, R. O., Hälinen, A. I., Pennanen, A. S., Sillanpäa¨ , M., Hillamo, R., and Hirvonen, M.-R. 2006. In-vitro inflammatory and cytotoxic effects of size-segregated particulate samples collected during long-range transport of wildfire smoke to Helsinki. Toxicol. Appl. Pharmacol. 215:341–353. Janssen, N. A. H., Brunekreef, B., van Vliet, P., Aarts, F., Meliefste, K., Harssema, H., and Fischer, P. 2003. The relationship between air pollution from heavy traffic and allergic sensitization, bronchial hyperresponsiveness, and respiratory symptoms in Dutch schoolchildren. Environ. Health Perspect. 111:1512–1518. Laden, F., Neas, L. M., Dockery, D. W., and Schwartz, J. 2000. Associations of fine particulate matter from different sources with daily mortality in six U.S. cities. Environ. Health Perspect. 108:941–947.

HETEROGENEITIES OF PARTICULATE TOXICITY Lanki, T., de Hartog, J. J., Heinrich, J., Hoek, G., Janssen, N. A. H., Peters, A., Stölzel, M., Timonen, K. L., Vallius, M., Vanninen, E., and Pekkanen, J. 2006. Can we identify sources of fine particles responsible for exercise-induced ischemia on days with elevated air pollution? The ULTRA study. Environ. Health Perspect. 114:655–660. Molinelli, A. R., Madden, M. C., McGee, J. K., Stonehuerner, J. G., and Ghio, A. J. 2002. Effect of metal removal on the toxicity of airborne particulate matter from Utah Valley. Inhal. Toxicol. 14:1069– 1086. Monn, C., and Becker, S. 1999. Cytotoxicity and induction of proinflammatory cytokines from human monocytes exposed to fine (PM2.5 ) and coarse particles (PM10-2.5 ) in outdoor and indoor air. Toxicol. Appl. Pharmacol. 155:245–252. Monn, C., Naef, R., and Koller, T. 2003. Reactions of macrophages exposed to particles