Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

An important determinant of a community's quality of life is the air it breathes. Considerable effort has been spent in characterizing the biologic effects of gaseous pollutants, such as ozone and oxides of nitrogen and sulfur, and reducing their emission. Recently, particles, especially those smaller than 10 µ in diameter (PM₁₀), have also been shown to be an important hazardous component of air pollution. PM₁₀ exposure has been associated with excess death from cardiopulmonary disorders and lung cancer in studies of six cities (1) and 151 metropolitan areas (2) in the United States, with a 1% increase in daily mortality for each 10 µg/m³ increase in PM₁₀ (3). PM₁₀ pollution also leads to increased respiratory symptoms (4), more frequent use of bronchodilators (5), increased emergency room visits for asthma (9), increased hospitalization rates for respiratory disorders (10), and decreased lung function (5, 14). Chemically, PM₁₀ pollution is a complex mixture of organic and inorganic compounds, but the properties responsible for its health effects remain a mystery. One hypothesis relates the biologic effects of particles to their H⁺ content (15). Another suggests that the toxicity of some particulates resides in associated metals (16), such as iron (8) or vanadium (17, 18).

Epidemiologically, the Utah Valley is one of the best characterized areas plagued by excess particulate air pollution (5, 6, 10, 11, 14, 19). Provo is its major metropolitan area. The valley has a dry, desert, four-season climate, and is bordered on both the east and west by high mountain ranges. Frequent winter inversions trap air pollutants near the valley floor, often resulting in high PM₁₀ concentrations (6, 10). The particulate pollution in the Utah Valley is low in acid content (5, 19) and is thought to come from emissions from a local steel mill (10). Therefore, we speculated that toxicity of this particulate was related to iron content.

To test this hypothesis, we characterized the metal content of air pollution particulates from Provo, Utah, and then instilled the particulates into the lungs of rats. After finding that the inflammation produced was dependent on local lung production of the rat interleukin-8 (IL-8) homolog cytokine-induced neutrophil chemoattractant (CINC), we used BEAS-2B cells, an SV-40 transformed human airway epithelial cell line, to characterize futher the aspects of Utah particulate pollution responsible for cytokine stimulation. Surprisingly, we found that activation of the transcription factor nuclear factor-B (NF-B) and stimulation of cytokine production were related to a high particulate content of copper and were exacerbated by the copper-binding protein ceruloplasmin, a normal constituent of respiratory epithelial lining fluid that has been previously thought to protect the lung from oxidant stress (20).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Collection and Characterization of Particulates

We studied total suspended particulates (TSP) collected onto matted glass-fiber filters (Mine Safety Apparatus) during January through March of 1982 at the North Provo monitoring station in Provo, Utah, by the Utah Division of Air Quality, using an Andersen High Volume TSP Sampler (Grasby-Andersen Co., Smyrna, GA). The U.S. Environmental Protection Agency (EPA) has determined that TSP comprises 50% to 60% PM₁₀ for most sampling sites in the United States (3). Filters were agitated in deionized water for 96 h, and the aqueous extract was centrifuged at 1,200 × g for 30 min to remove suspended matter, lyophilized, and resuspended in deionized water or normal saline (36.5 mg/ml). Particulate content of iron, vanadium, nickel, copper, zinc, lead, and sulfate (SO₄⁼) was measured by inductively coupled plasma-emission spectroscopy (ICPES) (Model P40; Perkin-Elmer, Norwalk, CT). To determine the relative amounts of Fe²⁺ to Fe³⁺ and Cu¹⁺ to Cu²⁺, extracts were diluted 1:10 or 1:20 in Hanks' balanced salt solution (HBSS) containing 125 µg/ml of the chelator phenanthroline (23), with or without 40 mg/ml of Na₂S₂O₄ to reduce all Fe³⁺ to Fe²⁺ and Cu²⁺ to Cu¹⁺. The absorbance of the Fe²⁺-phenanthroline complex was read at 515 nm and compared with a standard curve prepared with 0.1 mM FeSO₄ · 7 H₂O to determine the concentration of Fe²⁺ with and without the addition of reductant. Particulate endotoxin content was measured using the Limulus amebocyte lysate assay (E-Toxate; Sigma, St. Louis, MO) according to the manufacturer's instructions.

For experiments to determine the effect of humate and iron-loaded humate on respiratory epithelium, a 10-mg/ml aqueous solution of sodium humate (Aldrich Chemicals, Milwaukee, WI) was incubated overnight with 1 mM FeCl₃ in 50-ml conical polystyrene tubes on an orbital agitator. The iron-complexed humic acid precipitate was centrifuged at 1,200 × g for 10 min, washed three times with distilled deionized water, and dried at 50° C.

Measurement of Particulate Activity as a Fenton Catalyst

The capability of Provo particulate to support transitional-metal-dependent generation of reactive oxygen species (ROS) was measured by an assay that employs deoxyribose as a detector molecule. In this assay the pentose sugar 2-deoxy-D-ribose reacts with oxidants to yield a product that forms a pink chromophore on heating with thiobarbituric acid (TBA) at low pH (24). The reaction mixture, containing 1.0 mM deoxyribose as detector molecule, 1.0 mM H₂O₂ as substrate, 1.0 mM ascorbate as reducing agent for transitional metals, and Provo particulate, was incubated in saline at 37°C for 60 min with agitation, and then centrifuged at 1,200 × g for 10 min. One milliliter of both 1.0% (wt/vol) TBA and 2.8% (wt/ vol) trichloroacetic acid (TCA) was added to 1.0 ml of supernatant, heated at 100°C for 10 min, and cooled on ice. The chromophore concentration determined by its absorbance at 532 nm (A₅₃₂) was used as an index of oxidant generation by transition metals.

Measurement of Acute Lung Inflammation from Provo Particulate

Sixty-day-old male Sprague-Dawley rats (Charles River Laboratories, Raleigh, NC) were anesthetized with metafane (Pitman-Moore, Inc., Mundelein, IL) and then intratracheally instilled with either 500 µl of normal saline (NS) or Provo particulate in 500 µl NS. Twenty-four hours later, rats were again anesthetized with metafane, euthanized by exsanguination, and intratracheally lavaged with 35 ml/kg NS that was withdrawn after a 3-s pause and reinjected an additional two times with similar delays. Cell counts were quantitated with an automated cell counter (Coulter Instruments, Hialeah, FL), and differential counts were made by counting 200 cells per animal on modified Wright's-stained (Diff-Quik stain; American Scientific Products, McGaw Park, IL) smears observed at ×400 magnification. Lavage protein was measured with the Pierce Coomassie Plus Protein Assay Reagent (Pierce, Rockford, IL) modified for automated measurement (Cobas Fara II centrifugal analyzer; Roche Diagnostics, Montclair, NJ). Bovine serum albumin served as the standard.

To probe the relationship between particulate-induced neutrophil influx and induction of CINC expression, some rats were injected intratracheally with 1.5 mg goat polyclonal anti-CINC antibody (25) at the same time as with particulates. Preimmune goat serum was used as a negative control. The anti-CINC antibody used for these studies has been previously tested for specificity against the rat C-X-C chemokines macrophage inflammatory protein-2 (MIP-2) and CINC-2 in an assay involving in vitro neutrophil chemotaxis (26). In the chemotaxis assay, the activity of 10 ng/ml of CINC is completely blocked by 1 µg/ml of the antibody, whereas the activity of 10 ng/ml of CINC-2 or MIP-2 is blocked by 30 to 100 µg/ml of the antibody. Thus, the anti-CINC antibody is selective for CINC, but not totally specific. After 24 h, lungs were either lavaged as described previously or were inflation-fixed overnight in formalin at 20 cm H₂O and stained with hematoxylin and eosin (H&E).

Cell Culture

BEAS-2B cells (S6 subclone) were obtained from Drs. Curtis Harris and John Lechner of the National Institutes of Health, Bethesda, MD. BEAS-2B cells are an SV40-transformed human airway epithelial cell line (27) used previously to study the response of human airway epithelium to environmental, infectious and inflammatory insults (28). BEAS cells were grown to 90% to 100% confluence on uncoated plastic 12-well plates (Costar, Cambridge, MA) in keratinocyte growth medium (KGM), which is essentially MCDB 153 medium containing 0.15 mM calcium and supplemented with 30 µg/ml bovine pituitary extract, 5 ng/ml human epidermal growth factor (EGF), 500 ng/ml hydrocortisone, 0.1 mM ethanolamine, 0.1 mM phosphoethanolamine, and 5 ng/ml insulin as described previously (28). BEAS cells that had been passaged 60 to 80 times in our laboratory were used for the majority of studies. Medium was changed 24 h before experiments. These studies were interrupted by the government shutdown of 1996, necessitating the use of a different BEAS cell subpopulation for completion of the experiments than had been used prior to the work stoppage. Primary human tracheal and bronchial epithelial cells were obtained from healthy volunteers by cytologic brush biopsy of the lower airway. Primary cells were cultured in bronchoepithelial growth medium (BEGM; Clonetics) for a maximum of five passages.

Treatment of Cells

Fresh stock suspensions of Provo particulate were prepared in KGM before each experiment and added to culture medium in volumes sufficient to expose BEAS-2B or primary human bronchial epithelial cultures to 100 to 500 µg/well. At the criterion value of 150 µg/m³, often exceeded in Provo during winter atmospheric inversions (10), 500 µg of PM₁₀ particulates will be found in 3.33 m³ of air, and 500 µg of total suspended particulates will be contained in approximately half this air volume. To determine whether biologic activity was soluble or resided in a fraction complexed to an insoluble chelator such as a soil silicate, an 18.25-mg/ml dilution of Provo particulate in KGM was microfuged, and the sediment was resuspended in an equal volume of KGM. BEAS cells were then cultured with sufficient volumes of the supernatant or resuspended sediment to approximate a 500-µg/well equivalent dose. To expose cells individually to concentrations of iron, zinc, copper, or lead found in Provo particulate, fresh stock solutions were prepared of Fe₂(SO₄)₃, ZnSO₄ · 7 H₂O, CuSO₄ · 5 H₂O, or PbSO₄ in KGM. The concentrations of particulate and metal-ion solutions used did not alter the pH of the culture medium. Cell viability following particulate or metal treatment was determined microscopically after 20 h by the exclusion of trypan blue and by measurement of lactate dehydrogenase (LDH) activity released into the supernatant. Trypan blue dye exclusion in each well was determined by computerized image analysis before and after addition of 40 µl of 1% Triton X100 to each well, and was expressed as a ratio of initial area staining with trypan blue dye to total nuclear area (% trypan blue dye uptake). LDH activity was measured in a Cobas Fara II clinical chemistry analyzer (Roche), using an adaptation of a previously published procedure (31). In some experiments, recombinant manganese superoxide dismutase (MnSOD; Boehringer-Ingelheim; a gift from Dr. Claude Piantadosi of Duke University, Durham, NC), catalase, antioxidant interventions, deferoxamine, bovine salivary mucin, human ceruloplasmin, or human apoferritin (all from Sigma) were added in the final concentrations indicated. In other experiments, BEAS cells were exposed to humate or iron-complexed humate prepared as described previously.

Measurement of Interleukin-6 and Interleukin-8 Protein Concentrations

Cell supernatants were collected 24 h after addition of particulate or individual metals, and were microfuged and stored at 20°C until measurement of IL-6 and IL-8 concentrations with enzyme-linked immunosorbent assay (ELISA) kits purchased from R&D Systems (Minneapolis, MN). We have previously demonstrated BEAS cell release of these cytokines in response to air pollutants (29).

Measurement of ICAM-1 Expression by Fluorescence Flow Cytometry

Confluent cultures were exposed for 18 h to particulate, washed twice with Ca²⁺- and Mg²⁺-free Dulbecco's modified phosphate-buffered saline (DPBS; JRH Biologicals, Lenexa, KS), and dissociated by brief treatment with trypsin-ethylenediamine tetraacetic acid (EDTA) in an incubator at 37°C under 5% CO₂. Trypsinization was terminated by adding one-fourth volume of 1 mg/ml soybean trypsin inhibitor (Sigma) dissolved in KGM. Dissociated cells were then washed once with 2% bovine serum albumin (BSA; Sigma) in DPBS (2% BSA/DPBS). Equal aliquots of cells (2 × 10⁵) were stained for 20 min on ice with phycoerythrin-conjugated anti-ICAM-1 (Becton-Dickinson, San Jose, CA) or phycoerythrin-conjugated mouse IgG_2b control antibody (Becton-Dickinson), washed once with 2% BSA/DPBS, fixed with 2% paraformaldehyde, and analyzed with a FACSort flow cytometer (Becton-Dickinson). ICAM-1 surface expression was estimated by subtracting the mean fluorescence intensity of cells stained with control antibody from that of anti-ICAM-1-stained cells from the same culture.

Reverse Transcriptase-Polymerase Chain Reaction

Cells were grown to confluence in six-well plates and exposed for 2 h or 24 h to particulate. Monolayers were washed twice with DPBS, and cells were lysed with 4 M guanidine thiocyanate (Boehringer Mannheim, Indianapolis, IN), 50 mM sodium citrate, 0.5% Sarkosyl, and 0.01 M dithiothreitol (DTT). After scraping, lysates were sheared with four passes through a 22-gauge needle. RNA was pelleted by ultracentrifugation through 5.7 M cesium chloride (Boehringer Mannheim) and 0.1 M EDTA. One hundred nanograms of RNA was reverse-transcribed using Moloney murine leukemia virus (MMLV) reverse transcriptase (Life Technologies). The resultant complementary DNA (cDNA) was amplified with the polymerase chain reaction (PCR) (30, 32) for 26 and 31 cycles for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and IL-8, respectively, using gene-specific sense and antisense primers based on the following sequences published in the GenBank human DNA database: GAPDH: 5'-CCATGGAGAAGGCTGGGG-3' and 5'-CAAAGTTGTCATGGATGACC-3'; IL-8: 5'-AACCCTCTGCACCCAGTTTTCCTT-3' and 5'-TCCACTCTCAATCACTCTCA-3'. PCR-amplified DNA was separated on 2% denaturing agarose gel, interchelated with ethidium bromide, and visualized and photographed under ultraviolet (UV) light. The resulting polaroid negative (Type 55 film; Polaroid Corp., Cambridge, MA) was quantitated with a Bio Image Analyzer (Bio Image, Ann Arbor, MI). The intensity of the GAPDH cDNA band (a housekeeping gene unaffected by the addition of air pollution particulates) for each sample was then used to normalize differences between samples. For each sample, the integrated optical densities of IL-8 bands were divided by that of the GAPDH cDNA band to rectify any errors in quantitation. The linearity of IL-8 mRNA measured in this fashion by semiquantitative PCR has been previously confirmed during at least four successive 10-fold serial dilutions (28).

Electrophoretic Mobility Shift Assay

Cells grown to confluence on 75-cm² plastic petri dishes were stimulated 1 h with particulate or the treatments noted, and electrophoretic mobility shift assays (EMSAs) were performed as described (33). Briefly, nuclear extracts were prepared from 10 × 10⁶ cells per treatment. Adherent cultures were washed twice with cold DPBS and equilibrated for 10 min on ice with 1 ml of cold cytoplasmic extraction buffer (CEB: 10 mM Tris-HCl, pH 7.9; 60 mM KCl; 1 mM EDTA; 1 mM DTT) with protease inhibitors (PI: 1 mM Pefabloc, 50 µg/ml antipapain, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 40 µg/ml bestatin, 3 µg/ml E-64, 100 µg/ml chymostatin; Boehringer Mannheim). The detergent NP-40 was added to a final concentration of 0.1%, and cells were dislodged with a cell scraper. Nuclei were pelleted by centrifugation and washed with CEB/PI. Nuclei were then incubated for 10 min on ice in nuclear extraction buffer (NEB: 20 mM Tris-HCl, pH 8; 400 mM NaCl; 1.5 mM MgCl₂; 1.5 mM EDTA; 1 mM DTT; 25% glycerol) with PI, spun briefly to clear debris, and stored at 80°C. Using wild-type (WT: 5'-GGCTGGGGATTCCCCATCT-3') and mutant (MT: 5'-GGCTGcGGATTCCgCATCT-3') class I major histocompatibility complex (MHC) enhancer sequences for NF-B loci (34), DNA protein-binding reactions were performed with 2 µg of nuclear protein (as determined by Bradford dye binding; Bio-Rad, Richmond, CA) and 0.3 ng of ³²P-end-labeled, double-stranded DNA probe incubated for 10 min at room temperature in 10 mM Tris-HCl, pH 7.9; 50 mM NaCl; 2.5 mM EDTA; 1 mM DTT; 5 µg BSA; 0.1 µg poly deoxyinosine-deoxycytosine (dI-dC); and 4% Ficoll. Competition experiments were performed with unlabeled wild-type (10×) and mutant (50×) oligonucleotide sequences. In parallel EMSAs, specific NF-B subunits were identified in shifted complexes by addition to each reaction mixture of 1 µg of supershifting antibodies (Santa Cruz) against the p50, p65, and c-rel components of human NF-B. Samples were electrophoresed on a 5% nondenaturing polyacrylamide gel in 0.5% Tris, glycine, EDTA (TGE) buffer. Gels were dried and analyzed by exposure to a phosphorimaging screen (Molecular Dynamics, Sunnyvale, CA).

Statistical Analysis

Data are expressed as mean values ± SE. The minimum number of replicates for all measurements was four. Differences between multiple groups were compared through one-way or two-way analysis of variance (ANOVA) (35). The post hoc test used was Duncan's multiple range test. Two-tailed tests of significance were employed. Significance was assumed at P < 0.05. 

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Particulate air pollution collected in Provo contained: Zn, 1.01 ± 0.02 mg/g; Ni, 0.04 ± 0.01 mg/g; Fe, 2.22 ± 0.04 mg/g; Vn, 0.01 ± 0 mg/g; Cu, 1.37 ± 0.07 mg/g; Pb, 1.68 ± 0.04 mg/g; and SO₄⁼, 748 ± 16 mg/g, respectively. The high content of iron and sulfate was predicted for a sampling station near a steel mill, but the relatively elevated concentrations of copper, lead, and zinc were an unexpected surprise. More than 87% of the iron and copper was in the oxidized Fe³⁺ or Cu²⁺ state. Particulate endotoxin content was 0.6 EU/mg. The transitional metals in Provo particulate readily generated ROS in an ascorbate-driven Fenton reaction, catalyzing oxidation of the detector molecule deoxyribose (A₅₃₂ of 1.037 ± 0.016 for 500 µg Provo particulate/well, versus 0.003 ± 0.009 for none). Deoxyribose oxidation was completely inhibited by the ·OH scavenger dimethylthiourea (DMTU) (A₅₃₂ without DMTU = 1.037 ± 0.016, versus A₅₃₂ + 1 mM DMTU = 0.070 ± 0.011 for 500 µg Provo particulate/well). When instilled intratracheally into animals, Provo particulate produced profound dose-dependent airways inflammation, characterized by dramatic increases of protein and polymorphonuclear leukocytes (PMN) in bronchoalveolar lavage fluid (BALF) (Figure 1), proliferation of bronchiolar epithelium (Figure 2B), and intra-alveolar hemorrhage and influx of PMN into lung parenchyma (Figure 2B). PMN influx and lung injury were probably dependent in part on local lung production of the rat chemotaxin CINC, a chemokine similar to human IL-8 (25, 26, 36), because anti-CINC antibody decreased lung hemorrhage and significantly reduced PMN influx (Figures 2C and 2D).

View larger version (17K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 1.   Provo particulate caused airways inflammation in rats. Sixty-day-old male Sprague-Dawley rats were intratracheally instilled with either normal saline (NS) or Provo particulate in NS, and protein and neutrophils were quantitated in BALF 24 h later. (A) Lavage protein concentration. (B) Total neutrophils in lavage (n = 6 per treatment; *P < 0.05 versus saline).

View larger version (125K): [in this window] [in a new window]   View larger version (122K): [in this window] [in a new window]   View larger version (119K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   Figure 2.   The chemokine CINC mediated airways inflammation from Provo particulate in rats. Sixty-day-old male Sprague-Dawley rats were intratracheally instilled with 0 or 500 µg particulate in normal saline (NS). Twenty-four hours later, lungs were lavaged or inflation-fixed and stained with H&E. To probe the relationship between particulate-induced neutrophil influx and induction of CINC expression, some rats were injected intratracheally with 1.5 mg goat polyclonal anti-CINC antibody (24, 33) at the same time as with particulates. Preimmune goat serum was used as a negative control. (A) Saline-treated control rat lung (H&E-stained section). (B) Rat lung injected intratracheally with 500 µg Provo particulate. Note the proliferation of bronchial epithelium, hemorrhage, and PMN influx into airways and alveoli. (C) Rat lung injected with 500 µg particulate but treated with anti-CINC antibody. PMN influx and hemorrhage were greatly reduced. (D) Provo particulate (500 µg) increased total PMN in rat BALF. Lung PMN influx was significantly attenuated by anti-CINC antibody, suggesting a role for this IL-8-like rat chemokine in the pathogenesis of lung inflammation from particulate pollution (n = 6 per treatment; *P < 0.05 versus saline; +P < 0.05 versus particulate + goat serum).

To probe further mechanisms leading to airways inflammation, we studied the effect of Provo extract on cultured BEAS-2B cells. Provo particulate stimulated a dose-dependent increase in the concentration in the medium of the chemotactic cytokine IL-8 at 24 h (Figure 3A) yet did not increase LDH release in supernatants or uptake of trypan blue dye by cells. The content in the medium of IL-6 and the cell-surface expression of ICAM-1 were also increased (Figures 4 and 5). Cellular IL-8 mRNA was elevated at 2 h and further increased at 24 h (Figure 3B). When BEAS cells were stimulated with Provo particulate partitioned into soluble or sediment fractions, both fractions had some activity, but IL-8 secretion was markedly greater in response to the soluble fraction (1,519 ± 64 pg/ ml for supernatant, versus 262 ± 45 and 29 ± 2 pg/ml for sediment and medium controls, respectively; P < 0.05). Provo particulate also stimulated IL-8 secretion in primary cultures of human bronchial epithelium (Figure 6).

View larger version (17K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   View larger version (23K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   Figure 3.   (A) Provo particulate stimulated IL-8 secretion by BEAS-2B cells. Particulate was added to media and IL-8 protein concentrations were measured in the media after 24 h (n = 6 for each dose; *P < 0.05 as compared with cells not treated with particulate). Cells of Passage 80 were used for these experiments. (B) Provo particulates increased BEAS-2B content of IL-8 mRNA. Cells were grown to confluence in six-well plates and exposed for 2 h or 24 h to particulate. RNA was harvested, and mRNA for IL-8 was assessed through semiquantitative PCR. For each sample, the integrated optical densities of IL-8 cDNA bands were divided by that of GAPDH, a housekeeping gene unaffected by air pollution particulates, to normalize differences between samples. Results were then expressed as percent of values for media controls. Insert shows ethidium bromide-stained gels of IL-8 and GAPDH cDNA for media control cells (lanes 1 to 4) and cells stimulated with 200 µg particulate for 24 h (lanes 5 to 8) (n = 4 for each experimental time point; *P < 0.05 compared with media control cells). (C) Stimulation with Provo particulate of BEAS-2B IL-8 secretion is reproduced by its copper component, but not by component lead, zinc, or even iron. Fe2(SO4)3, ZnSO4 · 7 H2O, CuSO4 · 5 H2O, or PbSO4 in KGM was added to BEAS-2B cultures in amounts necessary to stimulate cells individually with concentrations of iron, zinc, copper, or lead found in Provo particulate. IL-8 protein concentrations were measured in the media after 24 h (*P < 0.05 as compared with cells not treated with metals). (D) Provo particulate stimulation of BEAS-2B IL-8 secretion is inhibited by some antioxidant scavengers and metal chelators. Confluent cells were exposed to 200 µg Provo particulate in the presence or absence of the O2 scavenger MnSOD (2,000 U/ml), the thiol antioxidant NAC (1 mM), the transition-metal chelator deferoxamine (Def, 100 µM), or the ·OH scavenger dimethylthiourea (DMTU, 1 mM). IL-8 protein concentrations were measured in the media after 18 h (n = 4 to 6 for each treatment; *P < 0.05 as compared with cells stimulated with particulates alone). Cells of Passage 60 were used for these experiments.

View larger version (13K): [in this window] [in a new window]   Figure 4.   Provo particulate stimulated secretion of IL-6 by BEAS-2B cells. Particulate was added to media and IL-6 protein concentrations were measured in the media after 24 h (n = 6 per treatment; *P < 0.05 as compared with cells treated with no particulate).

View larger version (17K): [in this window] [in a new window]   Figure 5.   Provo particulate enhances ICAM-1 expression by BEAS-2B cells. Cultures were exposed for 18 h to particulate and were then analyzed for ICAM-1 expression by fluorescence flow cytometry. The specific mean fluorescence intensity is shown as mean ± SE for n = 3 (*P < 0.05 as compared with cells not treated with particulate).

View larger version (13K): [in this window] [in a new window]   Figure 6.   Provo particulate stimulates IL-8 secretion by primary cultures of human bronchial epithelium. Particulate was added to media, and IL-8 protein concentrations were measured in the media after 12 h (n = 3 per treatment; *P < 0.95 as compared with cells not treated with particulate).

When BEAS-2B cultures were stimulated with amounts of individual metals contained in Provo particulate, zinc, lead, and ferric iron had no effect on cytokine production (Figure 3C). It is possible that to be catalytically reactive, iron must be complexed to a soluble chelator such as humic acid, a common byproduct of combustion found in air pollution particulates (37). To test this possibility, we exposed BEAS cultures for 24 h to humate or iron-loaded humate. Neither increased IL-8 secretion (9 ± 4 pg/ml for the medium control; 7 ± 1 pg/ml for 500 µg/ml humate; and 7 ± 0 pg/ml for 500 µg/ml iron-loaded humate). To our surprise, copper dramatically stimulated IL-8 secretion (Figure 3C). Copper did not provoke LDH release or increase trypan blue dye exclusion.

Provo particulate appeared to cause cytokine secretion in part through an oxidant mechanism. Stimulation of IL-8 was significantly reduced at 18 h by the O₂ scavenger superoxide dismutase (SOD), the thiol antioxidant N-acetylcysteine (NAC), and the metal chelator deferoxamine (Figure 3D), suggesting that IL-8 secretion was promoted in part by particulate-induced ROS. In contrast to its inhibition of Provo extract-induced deoxyribose oxidation, the ·OH scavenger DMTU failed to inhibit IL-8 secretion by BEAS cells in culture (Figure 3C).

When inhaled, air pollution particulates first interact with components of the overlying mucus layer and respiratory lining fluid that might modify the response of the underlying epithelium. We therefore studied the effect of adding mucus or the major iron- or copper-transporting protein components of epithelial lining fluid, transferrin and ceruloplasmin (20), on particulate-induced IL-8 secretion. Mucin significantly suppressed IL-8 secretion in response to Provo particulate (Figure 7). Conversely, in concentrations previously reported in epithelial lining fluid, ceruloplasmin dramatically increased IL-8 secretion when added with Provo particulate, and even increased IL-8 secretion when added to cultures alone (Table 1). The stimulatory effects of ceruloplasmin were not altered by addition of transferrin in concentrations previously reported in normal epithelial fluid, either as unsaturated apo- or iron-loaded holotransferrin (data not shown).

View larger version (11K): [in this window] [in a new window]   Figure 7.   Mucin reduces IL-8 secretion by BEAS-2B cells stimulated with Provo particulate. Cultures were stimulated with 200 µg particulate in the presence and absence of 1 mg/ml bovine salivary mucin. IL-8 protein concentrations were measured in media after 18 h (*P < 0.05 as compared with Provo particulate alone).

Secretion of IL-6 and IL-8 (38) and expression of ICAM-1 (39) are regulated in part by activation of transcription factors, prominently NF-B. We therefore performed EMSAs to determine whether exposure to Provo particulate activates NF-B in BEAS-2B cells. Figure 8A shows that compared with control (lane 1), Provo particulate caused a modest increase in NF-B activation (lane 3). Densitometry confirmed this increase to be 41 ± 9% above control for 500 µg of particulate (n = 3). Enhanced NF-B activation was eliminated by addition of unlabeled wild-type, but not mutant, oligonucleotide sequences (lanes 4 and 5). Copper equivalent to that found in 500 µg of particulate caused even greater NF-B activation than did particulate (lane 6). Ceruloplasmin treatment of cells produced a dramatic increase in NF-B activation (lane 7), and in some experiments the exposure of cells to the combination of ceruloplasmin plus 100 µg of particulate appeared to increase activation even further. Antibody to the p50 (Figure 8B, lane 3) and p65 (lane 4) components of NF-B, but not to c-rel (lane 5) or c-jun (lane 6), produced supershifts, suggesting activated complexes of p50 and p65 components.

View larger version (61K): [in this window] [in a new window]   Figure 8.   Provo particulate increases DNA binding of the transcription factor NF-B. (A) Electrophoretic mobility shift assays (EMSAs) of BEAS-2B cells incubated for 1 h without (lane 1) or with 100 µg/ml (lane 2) or 500 (lane 3) µg/ml Provo particulate. In lanes 4 and 5, respectively, nuclear extracts from cells treated with 500 µg/ml Provo particulate were incubated with labeled oligonucleotides in the presence of excess wild-type (10×, lane 4) or mutant (50×, lane 5) sequences. In lanes 6 and 7, BEAS cells were incubated for 1 h with the amount of copper equivalent to that in 500 µg/ml particulate (lane 6), with 25 µg/ml human ceruloplasmin (lane 7), or with 25 µg/ml ceruloplasmin plus 100 µg/ml Provo particulate (lane 8). Positions of the major (I) and minor (II) specific DNA-protein complexes are indicated by the arrowheads. A third, unidentified complex (III) appears to be constitutively present and competitively displaced by complexes I and II. (B) EMSAs of BEAS cells incubated for 1 h without (lane 1) or with (lanes 2 to 6) 25 µg/ml ceruloplasmin and 100 µg/ml Provo particulate. Specific NF-B subunits were identified in shifted complexes by addition of antibodies against the p50 (lane 3), p65 (lane 4), and c-rel (lane 5) components of human NF-B, or antibody to c-jun (lane 6). Gels shown are representative of at least three experiments (ns = nonspecific bands).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have studied air pollution particles collected in Provo, Utah, for two important reasons. First, particulate air pollution in this area is relevant to human public health. Increased particulate air pollution in the Utah Valley has been carefully documented to be associated with excess mortality and a worsening of respiratory symptoms and function (5, 6, 10, 11, 14, 19). The problem has not only been well studied epidemiologically, but has also attracted considerable attention in the local press (40). Second, air pollution particulates in this region contain high levels of metals, but as a desert with a low-moisture climate, the region should have less contribution to air pollution particulates from biologic sources. Thus, samples from the Utah Valley offer a relevant paradigm for studying the natural contribution of metals to the biologic effects of particulate air pollution on respiratory epithelium.

Provo particulates caused pronounced lung inflammation in vivo (Figures 1 and 2B). PMN influx and lung hemorrhage were reduced by intratracheal instillation of anti-CINC antibody (Figures 2C and 2D), suggesting that particulate- induced injury was due in part to local induction of this chemotactic member of the C-X-C family. Ozone-induced rat airways inflammation and hyperresponsiveness (26), and lipopolysaccharide (LPS)-induced influx of PMN into rat airway (36) have previously been shown to be inhibited by intratracheal instillation of anti-CINC antibody, but to our knowledge this is the first demonstration that acute neutrophilic inflammation from air pollution particulates is mediated by local lung production of this murine relative to human IL-8. Provo particulates also stimulated dose-dependent secretion of the proinflammatory cytokines IL-6 (Figure 4) and IL-8 (Figures 3A and 6), increased IL-8 mRNA (Figure 3B), and enhanced ICAM-1 expression (Figure 5) by cultured respiratory epithelium. Provo particulates readily catalyzed generation of ROS in vitro. That some but not all antioxidant strategies reduced IL-8 secretion in response to application of particulates to cultured BEAS-2B cells (Figure 3D) indicates that metal-catalyzed generation of ROS may play a role in cytokine stimulation. Provo particulates could stimulate cytokine secretion in part through activation of the transcription factor NF-B (Figure 8A), although we cannot rule out a contribution from particulate-induced stabilization of message. The biologic effects observed are unlikely to have come from endotoxin present in particulates. We have previously found that BEAS-2B cells cultured in KGM medium are unresponsive to endotoxin stimulation without addition of human serum, because the plasma components lipopolysaccharide-binding protein and soluble CD14 are required for activation of epithelial cells by endotoxin (41).

We originally hypothesized that the location of a steel mill in the midst of the Utah Valley would result in particulate pollution that produced biologic effects due to iron. The iron content of Provo particulates is elevated, but exposure of cultured epithelium to amounts of ferric iron, lead and zinc found in these particulates had no effect on cytokine production (Figure 3C). To our surprise, copper dramatically stimulated IL-8 secretion (Figure 3C) and enhanced activation of the transcription factor NF-B (Figure 8A). Copper, like iron, is a transitional metal that readily supports generation of ROS (42) and free-radical- mediated oxidation of low-density lipoproteins (LDLs) (43, 44), and the metal chelator deferoxamine can protect against the cytotoxic effects of copper-mediated lipid peroxidation (45, 46). Other metals, including nickel and cobalt (47), have been reported to activate NF-B, but to our knowledge this is the first report of such an effect from copper alone. The source of the increased copper content in particulates is unclear. The substantially greater response of BEAS-2B cells to soluble as opposed to sediment fractions from particulates suggests that any metals responsible for biologic activity are unlikely to be complexed to suspended soil silicates. Although copper in Provo particulates could arise from slag during steel manufacturing at the local plant, the copper content of particulates from Salt Lake City (37), located to the north in an adjoining valley, is 26 times higher than in Provo. This suggests that copper in Provo could also possibly originate in emissions drifting down valley corridors from a large copper smelter northwest of Salt Lake City. Concentrations of copper comparable to that measured in Provo particulate have recently been demonstrated in a standardized urban air pollution particulate (SRM-1648, from the National Institute of Standards and Technology, Gaithersburg, MD) and in ultrafine air pollution particulates (PM_2.5) collected in Canadian cities from the Great Lakes basin (48). In this latter case (48), cellular induction of the stress genes metallothionein IIa and heat-shock protein 70 in vitro was closely correlated with particulate concentration of copper, suggesting that copper may represent a potentially injurious component of particulate air pollution in regions beyond the intermountain American west.

We had also hypothesized that the biologic response to air pollution particulates might be modified by inherent components of airway lining fluid. Mucus significantly attenuated IL-8 secretion stimulated by particulates (Figure 7). This effect is not surprising, and suggests that chronic bronchitis may be a protective adaptation to respired metals. Mucus can scavenge ROS (49) and bind iron (50), and could possibly complex copper and heavy metals, like other sulfated polysaccharides (51). However, the profound exacerbation of Provo-induced IL-8 secretion by ceruloplasmin (Table 1), and the activation of NF-B (Figure 8) and stimulation of IL-8 secretion (Table 1) by ceruloplasmin alone, were unexpected findings. Although we cannot yet be certain that these results were caused by ceruloplasmin rather than by a contaminant in the commercial preparation, this is the first report, to our knowledge, of activation of NF-B by ceruloplasmin. The 132-kD glycoprotein ceruloplasmin is an acute-phase reactant and the major copper-carrying protein in human plasma, containing approximately 7 copper atoms complexed over three 42 to 45-kD domains. Several activities of ceruloplasmin have been described, including copper transport, oxidation of organic amines, ferroxidase oxidation of Fe²⁺ to Fe³⁺, and antioxidant activity against lipid peroxidation (42). A major protective antioxidant function has been proposed for the ferroxidase activity of ceruloplasmin in lung epithelial lining fluid (20), but this same ferroxidase activity can facilitate lipid peroxidation under certain conditions (52). This possibility may account for the marked enhancement of cytokine secretion when ceruloplasmin was added to culture media with Provo particulates (Table 1). Ceruloplasmin has recently been shown to facilitate oxidation of LDL in vitro (53) and by macrophages (54), smooth-muscle cells (55), and endothelium (55), and to induce expression of tissue factor activity by human monocytic THP-1 cells (56). These proinflammatory events have been hypothesized to be pathophysiologically important in promoting atherosclerosis (53). Our finding that ceruloplasmin increases IL-8 secretion (Table 1) and activates NF-B (Figure 8) in cultured respiratory epithelium questions the proposed protective role of an increased ceruloplasmin concentration found in BALF of patients with adult respiratory distress syndrome (ARDS) (21, 22). Also, ceruloplasmin levels are increased in BALF from asthmatic individuals (57). The reported susceptibility of asthmatic individuals to Provo air pollution particulates (5, 6) might be explained in part if an interaction occurs in vivo between an increased airway luminal ceruloplasmin concentration and air pollution particulates (Table 1), leading to increased local cytokine production by respiratory epithelium in asthmatic as compared with normal individuals.

We were surprised that the iron component of Provo particulates failed to stimulate cytokine secretion by BEAS cells. However, we recently have found that in response to stimulation with metals, lung cells can greatly increase their in vitro production of ferritin (60) and lactoferrin (manuscript in preparation) to transport and store iron with its coordination sites fully complexed. These events might serve to protect cells from stress by iron added to the culture medium. To determine whether complexation of iron was required for it to be catalytically reactive, we exposed cultures of BEAS cells to iron complexed with humic acid, a common byproduct of combustion found in air pollution particulates (37). However, neither humate nor iron-loaded humate stimulated IL-8 secretion. Conversely, our failure to elicit a biologic response with Fe³⁺ may have been due to the absence of sufficient reductant in the growth medium to reduce it to Fe²⁺. The commercial KGM medium used in our experiments contained 30 µM ascorbate, but the concentration of ascorbate in human BALF is 10-fold higher (61). In the presence of both ascorbate and ceruloplasmin, iron can undergo active redox cycling, with ascorbate reducing Fe³⁺ to Fe²⁺ and ceruloplasmin oxidizing Fe²⁺ back to Fe³⁺ (62). Such a reaction, even with the relatively low concentration of ascorbate in KGM medium, could provide an alternate explanation for the enhancement of particulate-induced cytokine secretion by ceruloplasmin. The in vitro biologic effects of Provo particulates in the presence of ascorbate concentrations similar to those found in the lung in vivo (61) will be the subject of future experiments. As a further possibility, the biologic effects of particulates may also be the result of complex interactions among the individual metal components. As an example, zinc blocks absorption of copper in the intestine (63) and isolated cells (64) by induction of metallothionein, an effect used therapeutically to treat Wilson's disease (64). This interaction may explain the stronger activation of NF-B in BEAS cells by the copper equivalent in Provo particulates (Figure 8A, lane 6) than with intact particulates containing increased levels of both copper and zinc (Figure 8A, lane 3).

Despite its prominent presence in particulates from the intermountain American west (37), in standardized fine (PM_2.5) urban air particles from the National Institute of Standards and Technology (48), and in PM_2.5 samples from multiple sources in the Great Lakes basin (48), copper is currently not regulated as an air pollutant by the Clean Air Act (15). However, copper is toxic for the lung when instilled into the airway (65, 66). Brass polishers have an increased smoking-adjusted incidence of chronic bronchitis (67), and women living near copper smelters in the United States suffer increased mortality from acute respiratory diseases (68). Although additional work is needed to identify environmental sources of copper and to understand fully the mechanisms of its airway toxicity, our studies suggest that copper in particulate air pollution represents a biologically important environmental problem for some regions of the United States.