Environ. Sci. Technol. 2004, 38, 5985-5992
Physicochemical Characteristics and Biological Activities of Seasonal Atmospheric Particulate Matter Sampling in Two Locations of Paris
induced peroxide production and cytokine release to the similar extent in the absence of cytotoxicity. In conclusion, whereas the PM2.5 samples differ by their PAH and metal composition, they induce the same biological responses likely either due to components bioavailability and/ or interactions between PM components.
A U G U S T I N B A U L I G , * ,† JEAN-JACQUES POIRAULT,‡ PATRICK AUSSET,§ ROEL SCHINS,| T I N G M I N G S H I , |,⊥ D E L P H I N E B A R A L L E , + PASCAL DORLHENE,+ MARTINE MEYER,‡ ROGER LEFEVRE,§ ARMELLE BAEZA-SQUIBAN,† AND FRANCELYNE MARANO† Laboratoire de Cytophysiologie et Toxicologie Cellulaire, Universite´ Paris 7, 2 Place Jussieu, Tour 53-54, 3e EÄ tage, Case Courrier 7073, 75251 Paris Ce´dex 05, France, Renault, Direction de l’Inge´nierie des Mate´riaux, Service 64120, TCR LAB 2 50, 1 Avenue du Golf, 78288 Guyancourt, France, Laboratoire Interuniversitaire des Syste`mes Atmosphe´riques, Universite´ Paris 12, 61 Avenue du Ge´ne´ral de Gaulle, 94010 Creteil, France, Renault, Direction de l’Inge´nierie des Mate´riaux, Service 64121, CTL L16 1 49, 1 Alle´e Cornuel, 91510 Lardy, France, and Particle Research, Institut fu ¨r Umweltmedizinische Forschung (IUF) an der Heinrich-Heine University Du ¨ sseldorf, Auf’m Hennekamp 50, 40225 Du ¨ sseldorf, Germany
Introduction
Fine particulate matter present in urban areas seems to be incriminated in respiratory disorders. The aim of this study was to relate physicochemical characteristics of PM2.5 (particulate matter collected with a 50% efficiency for particles with an aerodynamic diameter of 2.5 µm) to their biological activities toward a bronchial epithelial cell line 16HBE. Two seasonal sampling campaigns of particles were realized, respectively, in a kerbside and an urban background station in Paris. Sampled-PM2.5 mainly consist of particles with a size below 1 µm and are mainly composed of soot as assessed by analytical scanning electron microscopy. The different PM2.5 samples contrasted in their PAH content, which was the highest in the kerbside station in winter, as well as in their metal content. Kerbside station samples were characterized by the highest Fe and Cu content, which appears correlated to their hydroxyl radical generating properties measured by electron paramagnetic resonance. Particles were compared by their capacity to induce cytotoxicity, intracellular ROS production, and proinflammatory cytokine release (GM-CSF and TNF-R). At a concentration of 10 µg/cm2, all samples * Corresponding author phone: 01-44-27-60-62; fax: 01-44-27-6999; e-mail:
[email protected]. † Laboratoire de Cytophysiologie et Toxicologie Cellulaire, Universite´ Paris 7. ‡ Renault, Direction de l’Inge ´ nierie des Mate´riaux, Service 64120. § Laboratoire Interuniversitaire des Syste ` mes Atmosphe´riques, Universite´ Paris 12. | Institut fu ¨ r Umweltmedizinische Forschung (IUF) an der Heinrich-Heine University Du ¨ sseldorf. + Renault, Direction de l’Inge ´ nierie des Mate´riaux, Service 64121. ⊥ Present address: Jubei Provincial Centre for Disease Control and Prevention, Hubei, People’s Republic of China. 10.1021/es049476z CCC: $27.50 Published on Web 10/19/2004
2004 American Chemical Society
With anthropic activities development, atmospheric particulate pollution has become of great concern in relation to public health, in particular, in urban areas with a high traffic density. Urban particulate matter is rich in soot emitted during combustion processes including those produced by diesel vehicles and characterized by a carbonaceous core with many adsorbed organic compounds. Nowadays, the international regulation relative to air quality takes in consideration only the PM10, but PM2.5 could be more important as a human health risk because, due to their smaller size distribution, a larger fraction of particles can reach distal lungs after inhalation (1). Time-series epidemiological studies have highlighted that exposure to PM2.5 induces an increase of morbidity and mortality (2). In addition, a long-term study has shown that each 10 µg/m3 elevation in fine particulate air pollution was associated with approximately a 6% increased risk of cardiopulmonary mortality and 8% of lung cancer (3). PM2.5 have a variety of negative impacts on human health, but their underlying mechanisms are poorly understood. The lung, because of its interface with the environment, is a major target organ for particles injury that induces inflammatory processes. Rats exposed to PM2.5 exhibit an increase of the total number of bronchoalveolar lavages cells and a significant decrease of their viability (4). In vitro investigations have revealed that the PM-target respiratory cells (bronchial, alveolar epithelial cells, and macrophages) react to the release of various proinflammatory cytokines (5-9). This proinflammatory response results from activation of the MAPK signal transduction pathways and involves transcription factors induced by reactive oxygen species (ROS) (10-12). A widely accepted hypothesis is that particles toxicity depends on their composition (13). Among the biologically active components, organic compounds seem to play an important role in the proinflammatory response, as highlighted by studies performed on diesel exhaust particles (DEP) (14-17). Some organic compounds adsorbed on PM likely become bioavailable and are metabolized, especially the polycyclic aromatic hydrocarbons (PAH) (16). ROS can be produced during this metabolisation (18, 19) as well as by a quinoı¨d redox cycling (20, 21), resulting in oxidative stress. Oxidative stress however, suspected to play an important role in the generation of the proinflammatory response, can also be due to ROS generated by metals contained within PM or due to ultrafine particle surfaces reactivity (22-25). The transition metal content of PM is considered to be responsible for hydroxyl radical production (26) inducing activation of transcription factors (27), leading to cytokine production by human airway epithelial cells as exemplified with residual oil fly ash (ROFA) metal rich particles (28). Whereas the PM composition can influence particle toxicity, another level of complexity in analyzing the particle effect is brought by the variation of particle composition according to seasons and locations (29, 30) that produce different biological effects (31). The aim of this study was to associate physicochemical characteristics of PM2.5 to their biological activities in a human lung epithelial cell line. Urban PM2.5 recovered with a high VOL. 38, NO. 22, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5985
volume air sampler was collected during two different seasons and in two different locations in Paris, characterized by their contrasting heavy traffic load. An attempt was made to determine whether the particle monitoring and the physicochemical analysis data could be predictive of the biological activities of PM. The different types of particles were morphologically and chemically characterized, and their ability to generate hydroxyl radicals in abiotic conditions was measured. The PM2.5 biological effects were investigated using a human bronchial epithelial cell line (16-HBE). The different samples of PM2.5 were comparatively evaluated for their cytotoxicity, their ability to produce intracellular ROS, and their ability to induce a proinflammatory response characterized by the release of the granulocyte macrophage colony stimulating factor (GM-CSF) and tumor necrosis factor R (TNF-R) cytokines.
Materials and Methods Particle Collection and Preparation. Urban atmospheric particulate matter of 2.5 µm sampled with 50% efficiency (PM2.5) was collected with a high volume sampler machine (DA-80, Megatec, Paris, France), equipped with a PM2.5 selective-inlet head, in two locations of Paris: (a) a school playground at Vitry-sur-Seine, a suburb of Paris, classified by Airparif (the association that monitored the air quality in Paris) as an urban background station and (b) a site at Porte d’Auteuil adjacent to a major highway, which is a ring road of Paris, classified as a kerbside station. The machine operated at a flow rate of 30 m3/h, and particles were recovered on 150 mm diameter nitrocellulose filters (HAWP, Sartorius, La Ferte´sous-Jouarre, France) automatically changed when saturated. Particles were detached from filters by sonication in water. All particles recovered from the different filters of a same campaign were pooled to have a homogeneous batch of particles representative of the sampling period. The recovered suspension was frozen in liquid nitrogen and lyophylized at -80 °C (Christ Alpha 2-4, Bioblock Scientific, Illkirch, France). Then, stock solutions were prepared at the concentration of 2 mg/mL, by sonication in 0.04% dipalmitoyl lecithin (DPL) (Sigma, St. Quentin-Fallavier, France) in water. Particles were generally used at 10 µg/cm2 (respectively, 50 µg/mL) or from 1 to 30 µg/cm2 in a dose response study. Negative controls were made by using 0.04% DPL in water. As a positive control, Diesel exhaust particles (DEP, standard reference materiel 1650a purchased from the National Institute of Standards and Technology, Gaithersburg, MD) was used at 10 µg/cm2. These particles were collected directly from the heat exchangers of a dilution tube facility following the diesel engine. Finally, a filter without particles has been prepared in the same conditions. No biological effects have been detected with this control. In parallel, total suspended particles (TSP) were collected by a five-stage cascade impactor (DGI, Ecomesure, France). The flow rate of the sampler was 55 L/min. The sampler separates the particles into five granulometric fractions within the 3-0.25 µm aerodynamic diameter range. The stage sampler cutoff diameters are 3, 1.2, 0.6, and 0.25 µm, respectively. For each stage, the particles were collected on Teflon filters (FHLP, Millipore, St. Quentin en Yvelines, France) supported on stainless steel disks, except for the absolute filtration stage where a glass fiber Teflon filter (Pallflex TX40, Pall Gellmann Laboratory, Ann Arbor, USA) was used. Scanning Electron Microscopy Observations. Morphological characterization and individual elemental chemical analysis of particles were performed with an analytical scanning electron microscope (ASEM, JSM 6301F, JEOL, Japan) fitted with an energy dispersion X-ray spectrometer LINK ISIS (Oxford Instruments, U.K.). Particles in powder 5986
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 22, 2004
were fixed on an adhesive tape and metalized with gold. ASEM operating conditions were the following: accelerating voltage, 20 kV; intensity, 6 × 10-10 to 3 × 10-12 A; working distance, 15 mm; analytical spectrum, 0-20 keV. Polycyclic Aromatic Hydrocarbons Analysis. After lyophylization, dichloromethane organic extracts of PM2.5 were realized with an accelerated solvent extractor (ASE 200, Dionex) and concentrated by a vortex system on nitrogen atmosphere (turbovap II, Zymark). Then, the PAH quantitation was made using high-pressure liquid chromatography (HPLC) (HP 1100, Hewlett-Packard) coupled with a fluorescence detector (HP 1100, Hewlett-Packard). Metals Analysis. After lyophilization, particles in powder were digested. The procedure consists of a microwave sample acidic total digestion by a HNO3/HCl/HF mixture and subsequent analysis by inductively coupled plasma-mass spectrometry (ICP-MS) for Cd, Pb, Mn, Ni, and Cu; inductively coupled plasma-atomic emission spectrometry (ICP-AES) for Al, Zn, Mg, Ti, Ca, and Fe; high-resolution inductively coupled plasma-mass spectrometry (HR ICP-MS) for As, Cr, and V; and finally, flame photometry (FP) for Na and K. Measurement of OH• Generation Using EPR. Hydroxyl radical formation by the particle suspensions was evaluated by electron paramagnetic resonance (EPR) as described previously (26). Particle suspension (100 µL) was mixed with the spin trap 5,5-dimethyl-1-pyrroline-N-oxide in the presence of 125 mM H2O2 or without the addition of H2O2. The mixture was incubated at 37 °C in a shaking water bath for different times and filtered through a 0.2 µm filter (15 mm syringe filter, Satorius AG, Goettingen, Germany). The filtrate was immediately transferred to a capillary and measured with a Miniscope EPR spectrometer (Magnettech, BerlinGermany). The ESR spectra were recorded at room temperature using the following instrumental conditions: microwave frequency, 9.39 GHz; magnetic field, 3360 G; sweep width, 100 G; scan time, 30 s; number of scans, 3; modulation amplitude, 1.8 G; and receiver gain, 1000. Quantification was carried out on first derivation of EPR signal as the sum of the total amplitude, and outcomes are expressed as the total amplitude in arbitrary units (AU). Cell Culture Conditions. Dr. D. C. Gruenert (32) (Colchester, VT) kindly provided the human bronchial epithelial cell subclone 16HBE14o-. The cell line was cultured in DMEM/ F12 culture medium supplemented with penicillin (100 U/mL), streptomycin (100 µg/mL), glutamine (1%), fungizone (0.125 µg/mL, Invitrogen, Cergy-Pontoise, France), and UltroserG (UG) (2%, Invitrogen). Cells were cultured on collagen (type I, 4 µg/cm2) coated 25 or 75 cm2 flasks, 6 or 96 well plates (Costar, Cambridge, MA) at 20 000 cells/cm2. At the time of treatment, UG was not added to DMEM/F12. Cultures were incubated in humidified 95% air with 5% CO2 at 37 °C. Analysis of Cytotoxicity. Cytotoxicity was studied on subconfluent cultures by release of lactate deshydrogenase (LDH). After a 24 h treatment, the concentration of LDH was measured using a kit purchased from Sigma according to the manufacturer protocol. Results were expressed as percentage of LDH release. Cytokine Assay. Subconfluent cultures were exposed for 24 h to particles. Culture supernatants were recovered and frozen at -80 °C until use. The concentration of granulocytemacrophage colony-stimulating factor (GM-CSF) and tumor necrosis factor-R (TNF-R) released into the culture supernatants were measured with human GM-CSF and TNF-R Duoset ELISA development system, respectively (R&D systems Europe, Abingdom, U.K.). Color development was measured at 450 nm with a microplate photometer MR 5000 (Dynatech Laboratories). Analysis of DCF Fluorescence. Intracellular peroxide levels were assessed using H2DCF-DA (Molecular Probes,
TABLE 1. Characteristics of Particulate Matter Sampling Campaigns Realized in Two Locations of Paris during Winter and Summer code
sampling date
winter summer
PM-VW PM-VS
03/04/2002 to 04/09/2002 06/10/2002 to 07/21/2002
winter summer
PM-AW PM-AS
01/20/2003 to 02/18/2003 08/02/2002 to 08/26/2002
sampling volume (m3)
mean pressure (mb)
mean temperature (°C)
rain (mm)
mean speed of wind (km/h)
Vitry-sur-Sei ne 39 411 347 18 380 147
1008 1009
11 18
57 74
11 11
Porte d’Auteuil 17 724 424 12 341 185
1010 1006
4 20
42 33
11 9
FIGURE 1. Comparison of the granulometric repartition of total suspended particles between Vitry-sur-Seine, the urban background station (V) and Porte d’Auteuil, the kerbside station (A) in winter (W) and summer (S). Left panel: site comparison. Right panel: season comparison.
recovered mass (mg)
FIGURE 2. SEM micrographs of soot present in atmospheric samples from Paris. Soot consists generally at their emission in the atmosphere as single and approximately spherical particles of a few 10 nm in diameter (A) or organized in a chain or clusters of a few 100 nm (B). After time passed in the atmosphere, soot coalescence in large aggregates of few micrometers (C and D).
Eugene, OR), an oxidative-sensitive fluorescent probe, as previously described (18). The fluorescence analysis was performed with an EPICS-Elite-ESP flow cytometer (Coultronics-France). A 15 mW, argon-ion laser tuned at 488 nm was used for DCF fluorescence. DCF and PI fluorescence were, respectively, collected through a 525 and 620 nm bandpass-filter. Statistical Analysis. For in vitro experiments, all data were expressed as the mean ( standard error of the mean (SEM) of three cultures from a representative experiment. Means were compared by analysis of variance. The equal variance test is significant with R ) 0.05 (p < 0.001). All pairwise multiple comparisons were made with the Student-Newman-Keuls method (t-test, p < 0.05).
Results and Discussion We report that urban PM2.5 sampled in two locations of Paris and during two seasons were found to exhibit physicochemical differences but did not differ in their ability to induce an oxidative stress and proinflammatory cytokines release by human bronchial epithelial cells in vitro. It is the first study with urban Paris PM2.5 attempting to compare the biological effects induced by physicochemically different PM2.5. Four sampling campaigns were performed between March 2002 and March 2003 (Table 1). The determination of granulometric repartition in TSP revealed that for all samples, 95% of the particles have a size lower than 3 µm (Figure 1). Consequently, and due to the biological relevance of PM2.5, a PM2.5 sample device was chosen. Particles were collected with a high volume sampler to collect large amounts of particles sufficient for both physicochemical
FIGURE 3. SEM micrographs of different particles present with soot in atmospheric samples. (A and B) Silico-aluminated fly ash emitted during coal combustion. They are characterized by their size (5-10 µm), spherical form, smooth surface, and their characteristically chemical composition evaluated by X-ray energy dispersion spectrometry and showing the presence of Si and Al and in smaller proportion K, Ca, P, Fe, and Ti. (C) Pollen. (D) Sulfured particle probably produced by pneumatic wear. analysis and biological investigations. A particle recovery method from filters that does not modify particles as assessed by morphological observations and chemical analysis (data not shown) has been set up, and all particles from the same campaign were pooled to avoid day-to-day variations due to traffic and meteorological conditions. Morphological analysis revealed that soot is the most important particles in all the samples (Figure 2). This can be VOL. 38, NO. 22, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5987
TABLE 2. Comparison of the Soluble Organic Fractions (SOF) Percentage Contained in the Different Particle Samples
FIGURE 4. Micrographs of a terrigenous particle covered by mixture of salts such as (A) gypsum (CaSO4, 2H2O) or (B) halite (NaCl), contained in the air sampling of Vitry-sur-Seine and Porte d’Auteuil. explained by the fact that the greater Paris area comprises a running fleet of approximately 6 million motor vehicles, of which 47% are diesel powered. In addition, there is a power plant using charcoal located near the sampling site of Vitrysur-Seine. The elementary analysis by ASEM only revealed carbon (data not shown), characteristic of the soot’s core (graphite) on which organic compounds and other components are adsorbed. However, Paris PM2.5 also contains (i) some fly ash (Figure 3A) resulting from incomplete combustion of mineral impurities such as clay minerals and pyrites naturally contained in coal, (ii) biological materials including the 15 µm whole pollen as shown in Figure 3C, (iii) many particles that could result from wear process like the sulfured particle of 5 µm length as shown in Figure 3D, and (iv) particles of terrigeneous origin (Figure 4) modified by atmospheric reactions, generally bigger than the others. In the kerbside station Porte d’Auteuil, independent of the sampling season, more than 50% of particulate were closed to the ultrafine range (
