Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Air pollution is now an important issue for human health in industrialized countries. Both gaseous and particulate emissions resulting from motor vehicle traffic contribute to this phenomenon. Epidemiologic studies suggested that exposure to air pollutants generated by petrol- and diesel-burning engines worsens the symptoms of lung diseases such as asthma and rhinitis (1, 2). Among the various environmental factors, diesel exhaust particles (DEP) must be taken into account as they are considered to be among the most abundant components of particulate matter with an aerodynamic diameter < 2.5 µm (PM 2.5), but their real impact on air quality has not been fully elucidated. In vivo experiments have revealed that nasal challenges of humans by DEP result in a local increase in immunoglobulin (Ig) E and cytokine production (3). DEP can therefore induce inflammation of the respiratory tract, but the cellular events and the physicochemical characteristics of DEP involved in the induction of this effect are still unclear.

DEP are composed of carbonaceous particles with adsorbed and condensed hydrocarbons and sulfur (6). This organic fraction could represent up to 60% of the total mass of the particle in the absence of an exhaust gas treatment system and contains polycyclic aromatic hydrocarbons (PAH). The small size of DEP allows their penetration into the lung and their potential accumulation at the bronchiolar and alveolar levels. However, airway epithelial cells and alveolar macrophages remove most of the particles by mucociliary clearance and phagocytosis, respectively. Macrophages and epithelial cells are therefore the first cell types encountering DEP, which could initiate the inflammatory response. In vitro investigations have shown that both cell types released cytokines involved in inflammation when exposed to DEP (7). We have previously shown that native DEP are taken up by airway epithelial cells and stimulate the release of proinflammatory cytokines interleukin (IL)-8, IL-1, and granulocyte macrophage colony-stimulating factor (GM-CSF), which are involved in allergic diseases (11). This process occurs at noncytotoxic concentrations and involves the regulation of gene expression. Reverse transcriptase/polymerase chain reaction studies have shown that DEP-induced GM-CSF release is accompanied by an increase of the GM-CSF messenger RNA (mRNA) level in human bronchial epithelial (16HBE) cells (12) and by the activation of nuclear factor (NF)-B (13, 14), a transcription factor known to regulate the gene expression of many proinflammatory cytokines such as the GM-CSF (15). DEP also trigger signaling pathways in epithelial cells such as the phosphorylation of mitogen-activated protein kinase (MAPK) Erk 1/2 and p38 (14, 16). However, the DEP properties responsible for these events have not yet been identified. One hypothesis is that reactive oxygen species (ROS) could explain the cellular and molecular effects triggered after DEP exposure (17).

In the present study, to clarify the contribution of each component of DEP in the induction of the inflammatory response, we systematically compared the effects induced by native DEP (nDEP) with those induced by organic extracts of DEP (OE-DEP) and the carbonaceous core, represented either by stripped DEP (sDEP) or by carbon black particles (CB). This comparison was performed in 16HBE cells on the various DEP-induced cellular events, such as GM-CSF release, NF-B activation, and MAPK phosphorylation. The involvement of ROS in the effects induced by each component was also analyzed. In addition, we investigated the expression of the CYP1A1 gene, which is responsible for PAH metabolism.

We present evidence that, even if the carbonaceous core is not totally devoid of effect, organic compounds are the main contributors to the DEP-induced GM-CSF release, IB degradation, NF-B DNA binding, and MAPK Erk1/2 phosphorylation. ROS are involved in the effects induced by organic compounds and nDEP. We also demonstrated, for the first time, that nDEP induce CYP1A1 gene expression.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Diesel particulate matter SRM 1650 was purchased from the National Institute of Standards and Technology (Gaithersburg, MD) and are characterized by a 20.2 ± 0.4% of dichloromethane extractable mass and 1.33 ± 0.35 µg/g of benzo(a)pyrene (BaP). CB (FR103, 95 nm diameter) was obtained from Degussa (Frankfurt, Germany). Stock solutions of particles were prepared by suspension in a 0.04% (wt/vol) solution of dipalmitoylphosphatidylcholine (DPPC) in distilled water and by ultrasonication twice for 5 min at maximum power (100 W) (Vibra-cell; Bioblock Scientific, Illkirch, France). Particles were used at 10 µg/cm². Concentrations are expressed in µg/cm² because particles rapidly sediment onto the culture. DEP were extracted by dichloromethane in a soxhlet apparatus, and the extract obtained was dried, then redissolved in dimethyl sulfoxide. The collected particles were extracted a second time to ensure total extraction. Benzo(a)pyrene, a typical PAH present on DEP, was tested at 250, 50, and 0.25 µg/ml in comparison with OE-DEP at 15 µg/ml. It was used at 250 µg/ml in the CYP1A1 experiments as a positive control. All chemicals were purchased from Sigma (Saint Quentin Fallavier, France) except when otherwise specified.

Cell Culture and Stimulation

An SV-40 large T antigen transformed human airway epithelial cell line, 16HBE14o- (16HBE), kindly provided by Dr. D. C. Gruenert (18) (Colchester, VT), was cultured in Dulbecco's modified Eagle's medium-nutrient mixture F12 Ham (DMEM/F12) (GIBCO BRL, Cergy Pontoise, France) supplemented with penicillin (100 U/ml), streptomycin (100 µg/ml), glutamine (1%), fungizone (0.125 µg/ml; GIBCO BRL), and ultroserG (2%; GIBCO BRL). Cells were cultured on collagen (type I, 4 µg/cm²). At the time of treatment, ultroserG was not added to DMEM/F12. Cultures were incubated in humidified 95% air with 5% CO₂ at 37°C. Different inhibitors and antioxidants were tested on the 16HBE cells. The antioxidants dimethylthiourea (DMTU) and N-acetylcysteine (NAC) were dissolved in water and used at 10 and 10 mM, respectively. PD98059, an MAPK/extracellular regulated kinase kinase inhibitor (Tebu, Le Perray-en-Yvelines, France), and SB203580, a p38 kinase inhibitor (Calbiochem, Meudon, France), were prepared in DMSO and used at 10 and 1 µM, respectively. The antioxidants and inhibitors were used for a 1-h pretreatment and were kept in the culture medium during treatment of the cells. DMSO at 0.1% was added to the control and to the samples when necessary.

GM-CSF Assay

After 24 h of treatment, the supernatants of subconfluent cells of the 16HBE cell line cultured in 12-well dishes were centrifuged before being frozen at 80°C until use. The GM-CSF concentration released into the culture supernatant was measured with the human GM-CSF enzyme-linked immunosorbent assay kit (Amersham Life Science, Les Ulis, France).

Electrophoretic Mobility Shift Assay

After treatment, nuclear extracts were isolated from subconfluent 16HBE cells (3 × 10⁶) cultured in 25-cm² flasks as described by Staal and coworkers (19). Briefly, cells were washed and removed by scraping in Tris-buffered saline (TBS) (25 mM Tris-HCl, 136 mM NaCl, 2.7 mM KCl, pH 7.4), and pelleted. The pellets were resuspended in 400 µl of ice-cold hypoosmotic buffer (10 mM Hepes, 10 mM KCl, 2 mM MgCl₂, 0.1 mM ethylenediaminetetraacetic acid [EDTA], pH 7.8) supplemented with 0.5 mM dithiothreitol (DTT), 0.5 mM phenylmethylsulfonylfluoride (PMSF), 1 µg/ml antipain, 0.3 µg/ml leupeptin, 0.5 µg/ml pepstatin. Nuclei were spun down at 16,000 × g for 30 s after the addition of a 10% Nonidet P-40 solution, then resuspended in 40 µl of a hyperosmotic buffer (50 mM Hepes, 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 10% glycerol, pH 7.8) supplemented with 0.5 mM DTT, 0.5 mM PMSF, 1 µg/ml antipain, 0.3 µg/ml leupeptin, 0.5 µg/ml pepstatin. Nuclear proteins were extracted, incubating the nuclear suspension for 30 min at 4°C, with a slow rotation followed by a 16,000 × g centrifugation for 10 min. The supernatants that contain the nuclear extracts were stored at 80°C until used. They were complexed with radiolabeled double-stranded oligonucleotides containing a consensus NF-B site (Promega, Lyon, France) and loaded onto 5% polyacrylamide gel as described by Baeza-Squiban and associates (13).

Western Blot Analysis

Total proteins were collected by scraping the 16HBE cell cultures (3 × 10⁶, 25-cm² flask) supplemented with solubilization buffer (pH 7.4, 0.06 M Tris, 3% sodium dodecyl sulfate [SDS], 5% -mercaptoethanol, 10% glycerol, 1 mM PMSF), boiled, and run on 12% SDS-polyacrylamide gel electrophoresis gels at 150 V. Prestained molecular-mass markers (Bio-Rad, Hercules, CA) were run on adjacent lanes. Samples were normalized for protein contents before loading with the Bicinchoninique Acid (BCA) kit (Pierce, Rockford, IL). Electrophoresed proteins were electroblotted onto nitrocellulose membrane, and the blots were blocked with 5% milk in TBS, pH 7.6 (0.02 M Tris, 0.135 M NaCl) containing 0.1% Tween-20. The membranes were incubated for 2 h with the primary antibody in 1% milk in TBS-0.1% Tween-20. Polyclonal rabbit antibodies against phospho-Erk 1/2, Erk 1/2, and phospho-p38 (New England Biolabs, Beverly, MA) were used at a dilution of 1/2,000. The mouse monoclonal antibody 10B against IB was kindly provided by Dr. R. Hay (Edinburgh, UK) and used at a dilution of 1/100. The rabbit polyclonal antibody against -tubulin (Sigma) was used as an internal reference to control equal protein loading. Horseradish peroxidase-conjugated antirabbit or mouse IgG was used as a secondary antibody (New England Biolabs) at a dilution of 1/5,000. Bands were detected with chemiluminescence reagents and film according to the manufacturer's instructions (NEN, Boston, MA).

RNA Isolation and Northern Blot Analysis

Total cellular RNA was isolated from subconfluent cells (10 × 10⁶), cultured in 75-cm² flasks, by using Tri Reagent according to the manufacturer's instructions. The amount of RNA in aqueous solution was determined by absorbance at 260 nm. Equal amounts (30 µg) of total cellular RNA were size separated by 0.65 M formaldehyde-agarose (0.8%) gel electrophoresis and transferred onto Hybond-Plus membranes (Amersham Life Science) using 10× saline sodium citrate (SSC) buffer (1×: 150 mM NaCl, 15 mM sodium citrate). The blot was then baked for 2 h at 80°C and hybridized to specific probes. The blots were prehybridized at 65°C in Rapid Hybrid solution (Amersham Life Science) and hybridized overnight at 65°C with 1 to 5 × 10⁷ cpm/ml of ³²P-labeled complementary DNAs (cDNAs). Probes were synthesized from cDNAs with the Megaprime DNA labeling kit (Amersham Life Science) according to the manufacturer's instructions. After hybridization, the blots were washed for 30 min at 65°C with 2× SSC, then 1× SSC, and then 0.5× SSC, in the presence of 0.1% SDS. The membranes were exposed and the radioactivity was quantified with a PhosphorImager and ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

The cDNA probe of the CYP1A1 gene was a generous gift from Dr. I. De Waziers (Paris, France) (20).

Dichlorofluorescein Staining

Dichlorofluorescein-diacetate (DCFH-DA) is a nonfluorescent compound that enters the cells and is trapped by removal of the diacetate group. Upon interaction with peroxides, dichlorofluorescein (DCF) is converted into a fluorescent product. 16HBE cells cultured in 12-well dishes (1 × 10⁶) were incubated with DCFH-DA (20 µM) in Hanks' balanced salt solution (HBSS) for 20 min at 37°C in the dark. DCFH-DA was removed and the cells were then washed with HBSS and treated. Cells were collected by trypsination. Propidium iodide (3 µg/ml) was added to the samples, which were then immediately subjected to flow cytometric analysis.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Role of Organic Compounds in DEP-Induced GM-CSF Release

GM-CSF release was used as a biomarker of the proinflammatory response induced by DEP in 16HBE cells (12) in order to determine which component of nDEP is the most critical one for this response. For this purpose, nDEP- induced GM-CSF release was compared with that induced by OE-DEP, BaP, one of the numerous DEP organic compounds used as a model of PAH, sDEP, and CB that were used as surrogates of the carbonaceous core of DEP, which are known to be endocyted to the same extent as nDEP (11).

The various types of particles were used at 10 µg/cm². At this concentration, nDEP are not cytotoxic and clearly induce a GM-CSF release (11, 12). nDEP and sDEP from standard reference DEP (SRM 1650), and CB exhibited different activities after 24 h of treatment (Figure 1). nDEP induced a 4.7-fold increase in GM-CSF release compared with the control, whereas sDEP had a lower but significant effect and CB had no effect, suggesting a role of organic compounds in this response. OE-DEP tested at 15 µg/ml, which is in the range of the concentration of organic compounds present on 10 µg/cm² of DEP, induced a marked increase of GM-CSF release (3.7 fold), suggesting their participation in the induction of this proinflammatory response. Their effect was compared with that of BaP. BaP at 50 and 0.25 ng/ml had a slight but significant effect on GM-CSF release, which became stronger at 250 ng/ml, but always remained lower than that induced by OE-DEP.

View larger version (23K): [in this window] [in a new window]   Figure 1.   GM-CSF release induced by human bronchial epithelial cells (16 HBE 14o-) treated or not with nDEP, OE-DEP, sDEP, CB, and BaP for 24 h. Particles (nDEP, sDEP, CB) were used at 10 µg/cm2. OE-DEP were used at 15 µg/ml and BaP at 250, 50, or 0.25 µg/ml. Controls for BaP contain DMSO at 0.1%. Values were mean ± standard error (SE) of triplicate cultures from a representative experiment. *Significantly different (P < 0.05, corrected for multiple comparison) compared with control cultures.

These data highlight the role of organic compounds in DEP-induced GM-CSF release.

Role of Organic Compounds on DEP-Induced NF-B Activation

nDEP-induced GM-CSF release in 16HBE cells has been previously shown to be associated with increased GM-CSF gene transcription (12), and nDEP has been shown to induce a time- and dose-dependent increase in NF-B DNA binding (13, 14). The contribution of organic compounds to NF-B activation was investigated by studying both degradation of the inhibitory subunit IB by Western blot and NF-B DNA binding by electrophoretic mobility shift assay (EMSA). In the cytoplasm of nonstimulated cells, NF-B binds to IB and is maintained in an inactive form. IB degradation after stimulation allows translocation of NF-B to the nucleus and its subsequent binding to DNA. A kinetic study of the degradation of IB (Figure 2) was performed in 16HBE cells with or without treatment by either nDEP or OE-DEP. Western blot of total cellular proteins was performed using a specific antibody against IB (Mouse antibody [Mab] 10) or a control antibody against -tubulin and revealed that nDEP induced IB degradation after 2 to 4 h of exposure. OE-DEP also induced a similar IB degradation.

View larger version (64K): [in this window] [in a new window]   Figure 2.   Effect of nDEP (10 µg/cm2) or OE-DEP (15 µg/ml) on IB degradation in 16HBE cells analyzed by Western blot. Cells were exposed for indicated times. Blots with a specific antibody to IB were stripped and reprobed with antibody to -tubulin. A representative experiment with its respective densitometric analysis by Bio1D software.

The time-course study of NF-B DNA binding induced by nDEP and OE-DEP revealed that the kinetics of activation of NF-B DNA binding was different for these compounds (Figure 3). Indeed, whereas OE-DEP activated NF-B DNA binding after only 1 h of exposure, the nDEP effect clearly increased after 2 and 4 h of treatment. The carbonaceous core could play a role in this response as an increase of NF-B DNA binding was observed with CB (Figure 3) and with sDEP (data not shown). Taken together, these findings suggest that the organic compounds play an important but not exclusive role in NF-B activation.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Time-course study of NF-B DNA binding activity in 16HBE cells exposed either to nDEP or CB at 10 µg/cm2, OE-DEP at 15 µg/ml, or BaP at 250 µg/ml. Nuclear extracts were prepared after 1, 2, or 4 h of treatment and incubated with 32P-labeled oligonucleotides, encompassing NF-B consensus motif followed by analysis with EMSA. A representative experiment with its densitometric analysis (Bio1D) expressed as a percentage of NF-B DNA binding increase in comparison with control is shown.

DEP-Induced CYP1A1 mRNA Level

The prevalent role of organic compounds in the DEP-induced effects could be owing to their PAH content, known to induce the CYP1A1 gene. The relative amount of the CYP1A1 mRNA was determined by Northern blot analysis of total RNA from 16HBE cells with or without treatment by nDEP (10 µg/cm²), OE-DEP (15 µg/ml), sDEP (10 µg/cm²), or BaP (250 µg/ml). The CYP1A1 mRNA level was markedly increased in nDEP- and OE-DEP-treated cells in comparison with their respective control cells (Figure 4). This increase was slightly lower than that with the positive control BaP. sDEP had a very weak effect. The faint band observed could represent the effect of a small amount of PAH that could not be extracted from the DEP. It is likely that the induced CYP1A1 contributes to the metabolism of these PAH.

View larger version (57K): [in this window] [in a new window]   Figure 4.   Induction of CYP1A1 gene expression in 16HBE cells treated or not with nDEP or sDEP (10 µg/cm2), OE-DEP (15 µg/ ml), or BaP (250 µg/ml). DPPC and DMSO are controls for nDEP or sDEP, and OE-DEP or BaP, respectively. RNA (30 µg) were extracted from 16HBE cells after 6 h of treatment, electrophoresed, Northern blotted, and then incubated with a 32P-labeled cDNA probe for CYP1A1 mRNA or 18S RNA. (Top panel) A representative experiment. (Bottom panel) The quantification was performed with a phosphorimager and analyzed by ImageQuant software on three separate experiments, represented by three different symbols.

Involvement of ROS in a DEP-Induced Effect

Because the metabolism of organic compounds may lead to the production of ROS and because NF-B is a redox-sensitive transcription factor (19), we investigated whether ROS were involved in the nDEP- or OE-DEP-induced effects. For this purpose, we studied the effect of radical scavengers DMTU (10 mM) and NAC (10 mM). GM-CSF release induced by nDEP, OE-DEP, and CB in 16HBE cells was markedly decreased by DMTU and NAC (Figure 5A). A similar degree of inhibition was observed for nDEP and extracts. NF-B DNA binding induced by nDEP or OE-DEP was attenuated by DMTU (Figure 5B). However, the DMTU appeared to be less effective on OE-DEP- induced NF-B DNA binding than on nDEP-induced NF-B DNA binding, suggesting that other pathways could be involved in NF-B activation.

View larger version (21K): [in this window] [in a new window]   Figure 5.   Effect of antioxidants on GM-CSF release (A) and NF-B DNA binding induced (B) in 16HBE cells. (A) Cells were treated or not with nDEP or CB (10 µg/cm2), and OE-DEP (15 µg/ml) for 24 h. Cultures were pretreated or not with NAC (10 mM) or DMTU (10 mM) for 1 h. Controls contain DMTU or NAC when necessary. Values were mean ± SE of triplicate cultures from a representative experiment. *Significantly different (P < 0.05, corrected for multiple comparison) compared with control cultures. oSignificantly different compared with treatment without antioxidant. (B) Nuclear extracts were prepared from 16HBE cells pretreated or not with DMTU (10 mM) for 1 h, and treated or not with nDEP (10 µg/cm2) or OE-DEP (15 µg/ml) for 4 h. They were incubated with 32P-labeled oligonucleotide, encompassing NF-B consensus motif followed by analysis with EMSA. Arrowhead indicates the position of the specific complex. (Top panel) A representative experiment. (Bottom panel) The densitometric analyses of three separate experiments, represented by three different symbols, were obtained with Bio1D and expressed as a percentage of NF-B DNA binding increase in comparison with control.

nDEP and OE-DEP Induced ROS Production

Because nDEP- and OE-DEP-induced GM-CSF release and NF-B activation were inhibited by antioxidants, ROS generation was measured using the DCFH-DA probe, which yields a fluorescent product by interaction with peroxides. Treatment of 16HBE cells with nDEP and OE-DEP generated a 4.6- and 4.7-fold increase, respectively, in the mean intensity of DCF fluorescence after 4 h of exposure (Table 1).

DEP-Induced MAPK Activation

To study the signal transduction pathways triggered by nDEP, OE-DEP, and sDEP, we investigated the effect of Erk1/2 and p38 inhibitors (PD98059 [10 µM] and SB203580 [1 µM], respectively) on GM-CSF release, and the phosphorylation of Erk1/2 and p38 after various treatment times (30 min, 2 h, and 4 h). GM-CSF release induced by nDEP and OE-DEP was markedly decreased by PD98059 (Figure 6A). In contrast, SB203580 had no effect on nDEP-induced GM-CSF release, but slightly reduced OE-DEP-induced GM-CSF release.

View larger version (40K): [in this window] [in a new window]   Figure 6.   nDEP-, OE-DEP- or sDEP-induced MAPK activation. (A) Effect of MAPK Erk1/2 and p38 inhibitors on GM-CSF release. Cells were treated or not with nDEP or sDEP (10 µg/cm2), and OE-DEP (15 µg/ml) for 24 h. Cultures were pretreated or not with PD98059 (10 µM) and SB203580 (1 µM) for 1 h. Controls contain 0.1% DMSO. Values were mean ± SE of triplicate cultures from a representative experiment. *Significantly different (P < 0.05, corrected for multiple comparison) compared with control cultures. oSignificantly different compared with treatment without inhibitor. Significantly different compared with treatment without inhibitor, with P < 0.15. Western blot analysis of the phosphorylated form of the MAPK Erk1/2 (B) and p38 (C). To confirm an equal loading of proteins, the membrane was reprobed with an antibody against Erk1/2 used as a standard reference. 16HBE cells were stimulated or not with nDEP and sDEP (10 µg/cm2), or OE-DEP (15 µg/ml), and harvested at the time intervals indicated.

nDEP triggered Erk1/2 phosphorylation (Figure 6B) in a time-dependent manner (30 min up to 4 h). The increased Erk1/2 phosphorylation induced by OE-DEP appeared to be always lower than that induced by nDEP, but was higher than that induced by sDEP at 2 and 4 h of exposure. Increased p38 phosphorylation (Figure 6C) was already detected after 30 min of treatment with nDEP, OE-DEP, or sDEP, increased after 2 h, and decreased after 4 h of treatment. nDEP and OE-DEP therefore appeared to mainly induce phosphorylation of both MAPK Erk1/2 and p38, whereas sDEP predominantly induced p38 phosphorylation.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

To clarify the properties of DEP associated with inflammatory effects, a systematic comparison of the cellular effects induced by nDEP, OE-DEP, and sDEP or CB was performed on a human bronchial epithelial cell line, the 16HBE cell line. For this purpose, different molecular targets, typical of DEP-induced inflammatory response, were studied (i.e., GM-CSF release, IB degradation, NF-B DNA binding, and MAPK activation).

In the present report, we demonstrated the critical contribution of organic compounds to DEP-induced effects: (1) OE-DEP always induced approximately the same response as nDEP; (2) sDEP and CB, used as surrogates of carbonaceous core, exerted either no effect or markedly reduced effects; (3) BaP, a typical PAH present in DEP, is able to induce the same type of responses as OE-DEP.

Because OE-DEP caused the greatest effect in the nDEP-induced responses, it is possible that organic compounds could be desorbed from DEP to become bioavailable for the cells. This might occur in cells after the phagocytosis of DEP (11), as an aqueous solution of DPPC, a phospholipid used to prepare the DEP suspension, had no extraction effect on the PAH adsorbed onto the surface of the particles (11). In the present study, we show for the first time in human bronchial epithelial cells that nDEP and OE-DEP can rapidly induce the expression of the CYP1A1 gene, which could be related to the desorption of PAH from nDEP and their metabolism. Our experiment clarifies the experiments conducted by Sato and colleagues (21), who recently reported the induction of the CYP1A1 gene in the lung homogenates of Big Blue rats exposed, by inhalation, to whole DEP fumes. CYP1A1 induction is dependent on the binding of PAH to the aryl hydrocarbon receptor (AhR) in the cytosol. This complex translocates to the nucleus and associates with the AhR nuclear translocator, resulting in an active transcription factor that binds to the promoter of the CYP1A1 gene (22). Extracts of DEP have been shown to act as activators of the AhR (23) in murine hepatoma or human breast cancer cells.

Time-course studies demonstrated different kinetics of action for nDEP and OE-DEP: OE-DEP seemed to activate NF-B earlier than did nDEP. This result tends to highlight the different behaviors of organic compounds that are either immediately available (OE-DEP), or require desorption after phagocytosis (nDEP). In this latter case, the carbonaceous core could be considered mostly as a vector allowing the entry of organic compounds into the cells and their slow diffusion leading to sustained stimulation of the cells, as nDEP-induced NF-B DNA binding started later but was more persistent than that induced by OE-DEP. However, the carbonaceous core is not completely devoid of effects: although the effects induced by the carbonaceous core are fairly low compared with those induced by nDEP, they are not negligible, suggesting that they could act in combination with organic compounds. In particular, the markedly increased phosphorylation of p38 induced by sDEP revealed that the carbonaceous core could trigger a specific signaling pathway. Various studies have also emphasized the potential role of carbon black in inflammation, particularly when occuring as ultrafine particles (24).

This study also demonstrated the involvement of ROS in nDEP- and OE-DEP-induced effects, by the use of antioxidant molecules and direct measurement of peroxides using the fluorescent probe DCFH, confirming experiments performed in a human alveolar macrophage cell line (25). DEP have been shown to generate ROS, leading to the transcription of antioxidant genes (such as heme oxygenase-1 [HO-1]), which are regulated by the antioxidant responsive element (26, 27). The main generation of ROS could proceed from organic compounds contained in DEP that could produce ROS (1) immediately by oxidation- reduction cycling of quinones or quinonelike compounds or (2) subsequently after metabolization of PAH. The catalytic activities of cytochrome P450 are known to produce ROS directly and also generate biologic reactive intermediates, including quinones, which produce ROS by redox cycling (28). These two different mechanisms could induce two waves of ROS production. ROS generation could also proceed, to a lesser extent, from the carbonaceous core. CB exhibit oxidative properties as they deplete the antioxidant defenses in the epithelial lining fluid (29) and induce DNA strand scission in plasmidic DNA (30).

MAPK pathways are triggered by many extracellular stimuli such as environmental stresses. They could interact with the regulation of the activity of transcription factors such as NF-B (31) and they could be involved in the transduction of oxidative stress to the gene-regulating machinery of the cell (32). Erk, Sapk/JNK, and p38 are the three major MAPK described today. PD98059, an inhibitor of the Erk pathway, significantly reduces GM-CSF release induced by nDEP or OE-DEP. However, PD98059 has been described as an AhR antagonist (33), suggesting that its inhibitory action could be owing to the inhibition of DEP metabolism. Nevertheless, a marked increase in the phosphorylated form of Erk1/2 was shown with nDEP and OE-DEP. Although the p38 inhibitor induced little or no reduction of nDEP- or OE-DEP-induced GM-CSF release, nDEP, OE-DEP, and sDEP markedly induced p38 phosphorylation. In another human bronchial epithelial cell line (BEAS-2B), only p38 phosphorylation was observed with nDEP (16). This discrepancy with our results could either reflect cell-specific responses or be owing to the shorter kinetics as this study was only conducted over 2 h, whereas Erk phosphorylation was most striking after 4 h of exposure in our experiment.

The involvement of organic compounds in DEP-induced cellular effects will require additional characterization of the organic components: either quinones or PAH or both, as they are DEP components known to have toxicologic effects. A possible role of quinones in the DEP-induced oxidative effects has already been suggested (27, 34). Because DEP-PAH are possibly desorbed and further bioactivated by the induced CYP1A1 gene, they may generate oxygenated PAH, including quinones. Quinones could therefore be the key molecules in the DEP-induced effects. Because of their ability to generate ROS, they could participate in the regulation of many genes controlled by redox-sensitive transcription factors. Li and coworkers (27) recently reported that HO-1 expression, induced by OE-DEP in macrophages, essentially involved the polar extracted fraction of DEP, which contained quinones and oxygenated PAH.

In conclusion, we have shown that the DEP-induced inflammatory response in 16HBE cells mainly involves organic compounds such as PAH, which induce CYP1A1 gene expression. However, the carbonaceous core also exhibits a slight effect. Understanding the respective contributions of the various components of DEP in its cellular effects is important to vehicle manufacturers so that they are able to improve their exhaust gas post-treatment technologies. In addition, the present determination of the cellular signaling pathways triggered by DEP and the involvement of ROS in these events could provide new insight into therapeutic strategies.