Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The idea that respirable air pollutant particles (PM) produce acute cardiovascular and respiratory effects is widely accepted (1). Less well established is the concept that chronic exposure to PM can produce chronic effects, including increased risks of lung cancer and chronic obstructive pulmonary disease (COPD) (2). Morphologic examination of lungs from persons living in high PM areas suggests that PM produces fibrosis of the walls of airways (5, 6; A. Churg, unpublished data), thus providing a pathologic basis for COPD. This observation implies that PM particles are probably fibrogenic, but the exact mediators of fibrogenicity are unknown.

Considerable study has been devoted to the properties that cause acute reactions to PM as well as occupationally encountered particles, and there is extensive evidence that the availability of transition metals (usually iron, less commonly vanadium, cobalt, or nickel) from the matrix or surface of the dust is key to many of these effects. Simeonova and Luster (7) showed that surface iron was crucial to the induction of tumor necrosis factor (TNF)- in alveolar macrophages by asbestos fibers. A TiO₂ sample, which they noted did not contain iron, was nonreactive in their system. They also found that scavengers of active oxygen species or iron chelators (deferoxamine) prevented asbestos-induced TNF- production. Ghio and coworkers (8) found that surface complexed iron on silicate particles increased inflammation, oxidant production, and LTB₄ release by macrophages. Costa and Dreher (9) found that the dose of transition metals rather than total particle mass correlated with the inflammatory response to PM particles in rats. Frampton and colleagues (10) reported that the closure of a steel mill reduced the iron, copper, and zinc content of locally collected PM₁₀ compared with PM₁₀ collected during years when the mill was open, and this metal-reduced PM₁₀ showed the least oxidant generation and cellular reactivity when applied to cultured BEAS-2B cells. Ghio and associates (11) observed that release of interleukin (IL)-8 by cultured respiratory epithelial cells correlated with the ionizable metal content of PM samples, as did lavage neutrophils and protein in rats instilled with the dusts. Zhang and coworkers (12) found that the inflammatory activity and toxicity of ultrafine cobalt and nickel was much greater than that of ultrafine TiO₂. Quay and colleagues (13) noted that the induction of IL-6 by residual oil fly ash (ROFA) in human airway epithelial cells was inhibited by both deferoxamine and the oxidant scavenger N-acetyl cysteine, again implicating surface metals and oxidants. Jimenez and associates (14) showed that PM₁₀ activated NF-B and that pretreatment of the PM with a combination of the chelators deferoxamine and ferrozine prevented this effect in A549 cells, implicating iron as the mediator of NF-B activation. By contrast, Shulka and coworkers (15) also reported that PM, in this instance PM_2.5, activated NF-B, but deferoxamine did not prevent activation (see DISCUSSION).

The studies have all dealt with short-term observations. In contrast, relatively little information is available about the role of iron and other transition metals in chronic dust effects in the lung. Ghio and colleagues (16) noted that amount of complexation of iron by humic substances in coal correlated with degree of fibrosis in the lung tissue. Mossman and associates (17) showed that administration of polyethylene glycol-conjugated catalase decreased inflammation and fibrosis in a rodent inhalation model of early asbestosis, suggesting that rapid removal of hydrogen peroxide, and/or prevention of formation of hydroxyl radical, was protective; the implication was that AOS were produced from surface iron-mediated reactions. Kamp and coworkers (18) reported that treatment of rats with the iron chelator, phytic acid, protected against asbestos-induced injury, thus implicating iron more explicitly. Bonner and colleagues (19) reported that Mexico City PM₁₀ was able to upregulate the platelet-derived growth factor (PDGF)-receptor  in cultured fibroblasts via a metal- related mechanism, suggesting that these particles may enhance responses to fibrogenic mediators. Zhang and associates (20) showed specifically that vanadium, which is an important component of ROFA, stimulates bronchial epithelial cells to produce EGF-like growth factors that are mitogenic for fibroblasts, and Bonner and coworkers (21) found that vanadium pentoxide instilled intratracheally produced airway wall fibrosis in rats. We (22) recently demonstrated that levels of surface iron correlated with the ability of amosite asbestos fibers to induce procollagen gene expression and actual collagen content, as well as PDGF-A and transforming growth factor (TGF)-₁ gene expression in rat tracheal explants; procollagen production could be abolished by treatment of the fibers with deferoxamine or treatment of the explants with scavengers of active oxygen species.

Although these studies support a role for iron and other transition metals in fibrogenesis, the iron (or other transition metal) content cannot be separated from the dusts; thus, other intrinsic fibrogenic properties of the dust may not be evident even though they also drive the fibrosing process. Indeed, whether bioactive metal content and generation of AOS is a requirement for a dust to be fibrogenic is itself unclear. Many fibrogenic processes appear to proceed via elaboration of fibrogenic mediators such as TGF- and PDGF without any clear requirement for involvement of metals or AOS (reviewed in Ref. 23); and Brody and colleagues have shown that chrysotile asbestos, a dust with little iron, nonetheless induces both TNF- and fibrogenic mediators in a rat inhalation model (23, 24).

In this study we use a rat tracheal explant system to further examine the role of iron by adding iron to an ordinarily nonfibrogenic model PM particle and investigating the effects of this iron-bearing dust on fibrogenesis.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Dusts and Iron Loading

Most of the experiments used TiO₂ (anatase, geometric mean diameter [GSD] = 0.12 [1.4]µ; Aldrich Chemicals, Milwaukee, WI). Some experiments used glass fiber MMVF10 (Thermal Insulation Manufacturer's Association, Stamford, CT). The dust was sterilized in an autoclave overnight to destroy endotoxin before use. The titanium dioxide was analyzed by inductively coupled plasma atomic emission spectroscopy. This revealed concentrations of: iron, not detected (DL < 3 ppm); vanadium, not detected (DL < 1 ppm); zinc, not detected (DL < 2 ppm); copper, not detected (DL < 2 ppm). To load iron onto the dust surface, the dust was treated overnight with various concentrations of a freshly prepared mixture of equimolar Fe(II)/Fe(III) chloride because this combination has previously been shown to provide the greatest amount of particle uptake by the epithelial cells (25). The dust suspension was then washed three times with saline to remove unbound iron, and resuspended in culture medium. To confirm that iron had been added to the surface, surface iron was determined according to the dithionite-citrate-bicarbonate method as previously reported (25). With loading concentrations of iron of none (native dust), 0.1, 1.0, and 10 mM, this produced measured surface concentrations of 0, 35 ± 4.2, 82 ± 13, and 295 ± 36 µM of iron/g of TiO₂, respectively.

Explant Preparation and Culture

Tracheal explants were prepared from 250 g male Sprague-Dawley rats as previously described (26). Each explant was ~ 2 × 2 mm. For dust exposure the explants were submerged, epithelial side up, in a 500 µg/cm² (see DISCUSSION) suspension of dust in Dulbecco's modified Eagle medium (DMEM) without serum for 1 h. Controls were exposed only to culture medium. At the end of this time the explants were transferred to Petri dishes containing DMEM in agarose supplemented with 1% glutamine, 1% penicillin/streptomycin/fungizone, 1 µg/ml insulin, 0.1 µg/ml hydrocortisone, 1.5× amino acids, and 10% chicken serum. Explants were maintained in air plus 5% CO₂ organ culture with basal feeding in an incubator at 37°C for 7 d. This procedure results in dust uptake by the epithelium and transport to the interstitial tissues (26). Previous investigation in our laboratory has shown that there are few changes in the level of procollagen expression in this model before 7 d (26), hence only 7 d cultures were examined.

Measurement of Hydroxyproline

Hydroxyproline (HP) was measured on individual explants by high-performance liquid chromatography using the method described in reference (26).

Gel-Shift Assay for NF-B

To examine whether TiO₂ or iron-loaded TiO₂ activated NF-B in our explant system, additional explants were treated with the various dusts and incubated for 7 d. Explants were snap-frozen and then homogenized in 0.1% Triton X-100, 150 mM NaCl, 10 mM HEPES pH 7.5, 1 mM EDTA, 0.5 mM AEBSF, 1 mg/ml leupeptin, 1 µg/ml aprotinin, 10 µg/ml soybean trypsin inhibitor, and 1 µg/ml pepstain A. The homogenate was incubated on ice for 5 min and then centrifuged at 5,000 rpm for 5 min. The pelleted nuclei were resuspended in 100-500 µl of 25% glycerol, 20 mM HEPES, 420 mM NaCl, 1.2 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, 0.5 mM AEBSF, 1 µg/ml leupeptin, 1 µg/ml aprotinin, 10 µg/ml soybean trypsin inhibitor, 1 µg/ml pepstain A, and left on ice for a 30-min high salt extraction of the nuclear proteins. The lysed nuclei were centrifuged at 2,000 rpm for 15 s and a protein assay was carried on the supernatant. Single-stranded NF-B consensus oligonucleotide (5'-AGT TGA GGG GAC TTT CCC AGG C-3') was end-labeled with [-³²P] ATP. Binding reactions containing equal amounts of protein (7 µg) and 6.7 pmol of oligonucleotide were performed for 20 min in binding buffer (10 mM Tris HCl, 50 mM NaCl, 1 mM EDTA, 4% glycerol, 67 µg/ml poly[dI-dC]). Reaction products were separated in a 5% polyacrylimide gel in 0.25× TBE buffer and analyzed by autoradiography and densitometry. To obtain a reliable signal, 25 to 30 explants were combined for each data point.

Supershift Assay to Detect NF-B Components

Nuclear extracts were isolated as described above and preincubated with antibodies to NF-B components p65 or p50 (Santa Cruz Biotechnology, Santa Cruz, CA) for 20 min at room temperature before addition of the other reaction components.

Analysis of Cytoplasmic IB and IB Levels by Western Blot

Snap-frozen explants were homogenized in a lysis solution of 0.1% Triton X-100, 150 mM NaCl, 10 mM HEPES pH 7.5, 1 mM EDTA, 0.5 mM AEBSF, 1 µg/ml leupeptin, 1 µg/ml aprotinin, 10 µg/ml soybean trypsin inhibitor, and 1 µg/ml pepstain A. The homogenate was incubated on ice for 5 min and then centrifuged at 5,000 rpm for 5 min. The nuclear pellet was collected for gel shift assay (see above). A protein assay was performed on the supernatant and equal amounts of protein were used for each sample. The samples were separated in a 12% polyacrylimide gel and transferred to nitrocellulose membranes. The membranes were incubated in tris-buffered saline (TBS) containing 0.05% Tween-20 and 5% skim milk powder for 18 h at 4°C. The membranes were then incubated in a 1:400 dilution of rabbit polyclonal anti-IB or IB (Santa Cruz Biotechnologies) in TBS with 0.05% Tween-20 and 5% skim milk powder. The membranes were washed three times in TBS with 0.05% Tween-20 and incubated in a 1:1,000 dilution of horseradish peroxidase-conjugated goat anti-rabbit IgG (ICN Biochemicals, Cleveland, OH). Detection was by chemiluminescence and densitometry was performed on the films. A total of 25-30 explants was combined for each data point.

Analysis of IB-Phosphoserine 32/36 and IB-Phosphotyrosine

Cytoplasmic extracts were incubated with rabbit polycloncal anti IB (Santa Cruz Biotechnology) at 4°C for 1 h, followed by incubation with protein A-sepharose (5ng/ml) at 4°C for 30 min. The antigen-antibody complexes were pelleted and washed three times with ice-cold lysis solution (see above), centrifuged at 13,000 × g for 1 min, then redissolved and run on an SDS-PAGE gel for Western blotting with anti IB phosphoserine 32/36 antibody (New England Biolabs, Beverly, MA), antiphosphotyrosine (Upstate Biotechnology, Lake Placid, NY), or, in some instances, with anti-IB for IB detection. A total of 25-30 explants was combined for each data point.

NF-Prolyl Hydroxylase Probe

To investigate the possibility that NF-B translocation leads to increased binding to the prolyl-4-hydroxylase promoter, a probe with the sequence 5' TGG GAA TCT GTG TCC ATT ACT 3' was constructed and used in a gel shift assay as described above. This probe is derived from GenBank entry AF197928 and is a composite of the portion of the NF-B consensus sequence found in the promoter of the -chain of prolyl-4-hydroxylase and 12 base pairs downstream of the NF-B sequence. A total of 25-30 explants was combined for each data point.

Expression of Type I Procollagen and -Chain of Prolyl-4-Hydroxylase by RT-PCR

After 7 d in organ culture, explants were harvested and RNA extracted by the method of Chomczynski and Sacchi (27). Because most of the explant by weight is cartilage and the amount of tissue that actually contributes RNA is extremely small, three explants were used to prepare RNA for each data point for RT-PCR analysis. We have previously shown that this procedure provides a reliable signal (26). Freshly prepared explants from several different animals were used for each experiment and mixed to ensure that all explants for a given data point did not come from the same animal. Each test group for RT-PCR consisted of three data points.

First-strand cDNA was synthesized using superscript Rnase H reverse transcriptase (GIBCO-BRL, Grand Island, NY) according to the manufacturer's instruction. Briefly, 5 µg RNA were added to a reaction mixture of 1× first-strand buffer, 200 ng oligo(dT)12-18 primer, 0.5 mM each dATP, dTTP, dGTP and dCTP, 0.1 M DTT, plus water to 49 µl. 200 U superscript RT was added and the reaction incubated at 42°C for 1 h. PCR reactions contained 1 µM primers, 1.5 mM Mg++, 200 µM deoxynucleotide triphosphates, reaction buffer, 2.5 U of Taq DNA polymerase (Perkin-Elmer Cetus Instruments, Norwalk, CT) and 1 or 5 µl of cDNA in a final volume of 20 µl. The PCR temperature profile consisted of 25 or 28 cycles of denaturation at 94°C for 45 s, primer annealing at 60°C for 45 s, and extension at 72°C for 1.25 min, followed by an additional 5-min final extension at 72°C. The PCR products were size fractionated on 1.5% agarose gel and quantified from this ethidium bromide-stained gel using a gel Documentation System (Bio-Rad Laboratories, Hercules, CA).

Expression of malate dehydrogenase was used as control (housekeeper) gene. The primer sequences were:

Prolyl Hydroxylase -chain F: ATT CTA CTT CCT CAG TGT TCA GC

(GenBank AF197928) R: TCA CTC CAC TCG GTG TTC AGA CG

Procollagen type I F: CCA ATC TGG TTC CCT CCC AC

(GenBank M27208) R: GTA AGG TTG AAT GCA CTT

Malate Dehydrogenase F: CAA GAA GCA TGG CGT ATA CAA

(GenBank NM031151; Ref. 38) R: TTT CAG CTC AGG GAT GGC CTC

Treatment with Inhibitors and Chelators

1. 1. Iron chelator deferoxamine: TiO₂ or iron loaded TiO₂ was incubated overnight with 10 mM deferoxamine (DFX) (Desferal, Ciba-Geigy) and excess DFX was removed by washing the dust in saline before use. Dust was then resuspended in culture medium and used as above.
2. 2. Proteasome inhibitor MG-132: Explants were submerged in 0.5 µM MG-132 (Peptide Institute Inc., Osaka, Japan) in 0.1% DMSO/DMEM for 1 h and were then exposed to TiO₂ as above, but with MG-132 in the medium. 0.5 µM MG-132 was also included in the agarose culture medium for the 7-d incubation period.
3. 3. Cell-permeable AOS scavenger tetramethylthiourea (TMTU): Explants were exposed for 1 h to 10 mM TMTU (Sigma) and then to TiO₂ with TMTU in the medium. A quantity of 10 mM TMTU was also included in the agarose culture medium for 7 d.
4. 4. Tyrosine kinase inhibitor genistein: Explants were exposed for 1 h to 74µM genistein (Sigma) and then to TiO₂ with genistein in the medium. Genistein was also included in the agarose culture medium for 7 d.

Mobilization of Iron from the Dust with Citrate

Titanium dioxide was preincubated for 24 h in a 1-mM mixture of equal parts freshly prepared ferrous and ferric chlorides in saline, then centrifuged at 2,000 rpm for 10 min to remove the iron- saline solution. The dust was suspended again in 50 mM NaCl (pH 7.5) solution containing 1 mM citrate with stirring overnight in the dark at room temperature. The titanium dioxide was sedimented by centrifugation. The amount of total iron in the supernatant was determined according to the dithionite-citrate-bicarbonate method (25). The tracheal explants were treated with supernatant-DMEM containing a final concentration of 50 µM mobilized iron for 1 h, then incubated on supernatant-agarose medium culture containing 50 µM iron at 37°C for 7 d.

Statistics

All experiments were repeated. Where it was possible to analyze and run multiple samples/experiments on a single gel, for example, for RT-PCR, then formal ANOVA was performed and standard deviations are shown. Comparisons for hydroxyproline content were made among treatment groups by ANOVA. Values of P < 0.05 were considered significant. Where the complexity of the experiment precluded running all samples together, representative data from single experiments are shown.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Figure 1 shows the effects of iron loading on type I procollagen gene expression. TiO₂ by itself does not affect procollagen gene expression, but adding increasing amounts of iron to the TiO₂ leads to progressively increased expression. This figure also shows the effects of iron loading on tissue hydroxyproline levels. Again, this particular TiO₂ by itself does not cause increases in hydroxyproline levels, but iron-loaded TiO₂ does. This effect is completely prevented by treatment with the cell-permeable AOS inhibitor, TMTU. To see how specific the phenomenon of iron loading in inducing fibrosis might be, we iron loaded the standard glass fiber MMVF10. Figure 1 shows that, by itself, MMVF10 does not induce procollagen gene expression, but it does do so when iron is added to the surface.

Figure 2 illustrates gel shifts to show NF-B translocation into the nuclei. Iron-loaded TiO₂ causes an approximately doubling of nuclear NF-B levels, whereas TiO₂ by itself has essentially no effect. The supershift assay in the same gel confirms that p65 is specifically translocated into the nucleus; we were not able to demonstrate translocation of p50. Figure 3 illustrates IB levels at 7 d of culture. Iron loading leads to decreases in IB but increases in IB. Figure 4 illustrates cytoplasmic levels of IB-phosphoserine 32/36. Relative levels are increased by iron loading but not by TiO₂ alone compared with non-dust-treated controls. Figure 4 also shows levels of IB-phosphotyrosine. Again, fine TiO₂ by itself does not affect phosphotyrosine levels, but addition of surface iron produces a marked increase in relative levels.

View larger version (47K): [in this window] [in a new window]   Figure 2.   Gel shift and densitometry to show effects of iron loading on nuclear NF-B levels after 7 d organ culture. Iron loading (lane 3) leads to a distinct increase in NF-B levels; TiO2 by itself (lane 2) has no effect compared with non-dust-loaded control (lane 1). Supershift assay using antibody against p65 (lane 4) shows that p65 has been translocated into the nuclei after loading of the dust with iron; however, we were not able to detect p50 (lane 5). Data from a representative experiment.

View larger version (15K): [in this window] [in a new window]   Figure 3.   IB (top) and IB (bottom) levels on Western blot and corresponding densitometry. Iron loading leads to relatively decreased levels of IB and increased IB. Lane 1, control; lane 2, TiO2; lane 3, TiO2 plus iron. Data from representative experiments.

View larger version (11K): [in this window] [in a new window]   Figure 4.   IB-phosphoserine 32/36 (top) and IB-phosphotyrosine (bottom) levels by Western blot. Iron loading increases the relative amounts of both IB forms. This finding may imply that different mechanisms are proceeding in different anatomic compartments. Lane 1, control; lane 2, TiO2; lane 3, TiO2 plus iron. Data from representative experiments.

Figure 5 shows the effects of various inhibitors and chelators on procollagen gene expression. Addition of TMTU to the medium abolishes the effects of iron loading, indicating the involvement of AOS. A similar result is obtained by chelating iron with the non-redox-active chelator deferoxamine (DFX), confirming that iron is playing a role. Preventing NF-B activation by preventing proteasome degradation with MG-132 prevents iron-induced increases in procollagen. The tyrosine kinase inhibitor genistein also prevents the iron effect on procollagen expression (Figure 5). Lastly, application of a citrate extract of iron from the iron-loaded dust leads to marked increases in procollagen gene expression.

View larger version (17K): [in this window] [in a new window]   Figure 5.   Effects of various inhibitors and chelators on procollagen gene expression. TMTU and the non-redox-active iron chelator, deferoxamine, abolish the increased gene expression seen with iron loading, as does treatment of the explants with the proteasome inhibitor MG132 or the tyrosine kinase inhibitor, Genistein. Application of a citrate extract of the iron-loaded dust leads to marked increases in procollagen gene expression. Values are mean ± SD.

Figure 6 shows a gel shift run using nuclear extracts incubated with a probe matching the portion of the NF-B consensus sequence found in the prolyl-4-hydroxylase promoter and adjacent base pairs, as described in MATERIALS AND METHODS. Although signals were not as strong as those observed with the usual consensus sequence probe, a small increase was consistently found with iron-loaded TiO₂. This figure also shows RT-PCR data for gene expression of the -chain of prolyl hydroxylase. Again, titanium dioxide by itself has no effect on prolyl hydroxylase expression, but iron loading produces a marked increase and this effect is largely blocked by deferoxamine.

View larger version (19K): [in this window] [in a new window]   Figure 6.   Top: Gel shift run using a probe containing the portion of the NF-B consensus sequence found in the prolyl-4-hydroxylase promoter plus 12 adjacent bases from the promoter. There is a small but consistent increase in the signal (arrow) from iron-treated explants. Data from a representative experiment. Bottom: Gene expression of prolyl-4-hydroxylase -chain by RT-PCR. Iron loading increases gene expression and DFX blocks this effect. These findings suggest that iron-mediated NF-B activation may directly drive prolyl hydroxylase gene expression. Values are mean ± SD.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

As described earlier, the role of transition metals, especially iron, in fibrogenesis is uncertain, although individual measurements at the cellular level suggest that such metals have the potential to directly or indirectly upregulate fibroblast proliferation. In this study we have taken an ordinarily nonfibrogenic and minimally reactive model PM particle, fine (0.12 µ) TiO₂, and simply added iron to the surface. By doing so we convert the TiO₂ into a fibrogenic particle that causes increases in procollagen gene expression, tissue hydroxyproline content, and NF-B activation. We see the same fundamental phenomenon using a fibrous glass, MMVF10; this again is nonfibrogenic but becomes so when loaded with redox-active iron on the surface. These observations support our hypothesis that surface iron or iron mobilized from the matrix plays an important role in particle fibrogenesis. However, the limitations of this theory must also be stressed: as noted, some dusts with low levels or no iron, for example chrysotile asbestos, are fibrogenic, as are many forms of ultrafine particles that, by analysis, have very little iron (29)

Our model needs some comment. The explants offer the advantage that both epithelial and interstitial cells are present in their normal anatomic arrangement, and that dust is taken up by the epithelial cells and transported to the interstitium (26), as it is in vivo, allowing realistic levels of mediator production by the cells in each compartment and realistic interactions between compartments; the explants thus provide a good approach to examining both mechanism(s) and ultimate effects (fibrosis). We have chosen to use TiO₂ as a model PM particle because it is easy to load with iron in a reproducible fashion. Iron-loaded carbon black aggregates would in theory be a better model, but in our experience they are difficult to load in a reproducible fashion. The use of TiO₂ as a model is well established in the literature (29), but some care nonetheless needs to be taken with translating from the model to actual PM particles.

Iron may enter the lung on the surface or as part of the matrix of particles, but, as recently reported by Ghio and coworkers (30), iron levels on the particle are not static: iron may be removed, but particles in the lung also tend to accumulate iron from cellular sources and thus become foci of reactivity. This latter process probably is minimal here because there is no blood and no circulating iron transport proteins; certainly the control (nonloaded) particles show no effects.

It appears that activation of procollagen gene expression and production of hydroxyproline proceed, in our iron-loading model, via activation of NF-B. As alluded to earlier, activation of NF-B by mineral particles and air pollutant particles is claimed to occur via a variety of mechanisms. We (22) previously reported that amosite asbestos, a dust containing large amounts of iron, activated NF-B through a proteasome-dependent pathway, using a similar explant model. As in the present studies, chelation of surface iron with deferoxamine or scavenging of AOS with TMTU was protective. Shulka and colleagues (15) found that Vermont PM_2.5, again an iron-containing dust, also activated NF-B, but chelation of the iron with deferoxamine did not prevent NF-B activation, suggesting that an alternative pathway was involved. These authors demonstrated the presence of oxidative stress when PM_2.5 was applied to pulmonary epithelial cells, and treatment with catalase, which scavenges H₂O_2, was effective in preventing NF-B activation. Jimenez and associates (14) reported that Edinburgh PM₁₀ activated NF-B in A549 cells and that this process could be prevented by deferoxamine, or even better, by the combination of the iron chelators deferoxamine and ferrozine. However, Western blots failed to show evidence of IB degradation, implying that iron-generated AOS were driving an alternate pathway. In this regard it is of interest to note the recent report of Kang and coworkers (31). They showed that crystalline silica activated NF-B in a macrophage cell line via an oxidant-driven pathway that involved tyrosine phosphorylation of IB but not serine phosphorylation; this process was independent of proteasome-mediated degradation of IB. Janssen-Heininger and colleagues (32) have proposed that AOS in general activate NF-B through a proteasome-independent pathway that does not involve IB degradation but instead involves mitogen-activated protein kinase (MAPK) phosphorylation, whereas TNF- activates NF-B through a proteasome-dependent pathway.

The present study, which uses organ culture rather than single cell monolayer culture, suggests that both pathways of NF-B activation are involved in mediating the effects of iron-bearing dust on procollagen production. The demonstration that p65 is translocated into the nuclei is also important, because it indicates that this translocated material is active NF-B (33). The fact that the proteasome inhibitor MG-132 blocks procollagen gene expression, and that total cytoplasmic IB is decreased by iron loaded dust while, at the same time, relative levels of IB-phosphoserine 32/36 are increased, indicates that this process is proceeding via the usual classic pathway of IB degradation (33). However, the marked increase in IB-phosphotyrosine and the ability of the tyrosine kinase inhibitor, genistein, to block increases in procollagen production indicate that the alternate pathway is also involved. The most likely explanation for the set of observations in our dust model is that we are seeing effects from two different anatomic compartments (i.e., interstitium versus epithelium) or, conceivably, two cell types in the same compartment. This in one sense is a disadvantage of complex systems such as organ cultures, but in another sense provides information that is not obtained from monolayer cultures; namely that in a "real" system in which there are several cell types with modulation of effects in one anatomic compartment by the cells of the other anatomic compartment, multiple pathways are involved.

Our findings also support the role of both AOS and iron in increased procollagen production, because this process is blocked by the cell permeable AOS scavenger TMTU or the non-redox-active iron chelator, deferoxamine. It is interesting in this context that removal of iron from the particle into solution with citrate, a chelator that retains the redox activity of iron, reproduced the increases in procollagen gene expression seen with the iron-loaded dust. This process is chelator dependent, because we previously tried including a mixture of iron(II)-iron(III) in the agarose medium (without dust) and found it had no effect on procollagen gene expression (22). Whether the effects of iron-bearing dusts are mediated by iron bound to the dust surface or by iron leached into the cytoplasm has been an area of uncertainty (34), and many of the reports cited earlier have shown that a soluble component of the dust produces most or all of its effects. In regard to asbestos, Gardi and coworkers (35) showed that application of a citrate extract of crocidolite asbestos (another high iron dust) to cultured fibroblasts resulted in increased collagen production. Our observations support the importance of iron leached from the particle.

How iron might act to increase collagen production in the lung is not known. The data of Gardi and colleagues (35) imply that iron can act directly on fibroblasts. More data is available in the liver where iron in general acts to increase fibrosis, often in synergy with other insults such as alcohol (36). In hemochromatosis models, iron overload increases collagen synthesis by hepatocytes. This is associated with increased gene expression of type I procollagen as well as increased prolyl-4-hydroxylase (PH) chemical activity (37). One possible mechanism in our model would involve PH, a crucial enzyme in collagen biosynthesis. PH requires ferrous iron for activity and thus more available iron (from the dust) might simply increase prolyl hydroxylase activity. However, this theory would not explain why we can block increased procollagen expression with AOS scavengers or a proteasome inhibitor. We have examined gene sequences in GenBank and found that the promoter for the -subunit of PH in rat contains the NF-B consensus sequence TGG GAA TCT at position 601 (GenBank AF197928). Thus it is possible that NF-B activation leads to increased transcription and translation of prolyl hydroxylase. To further examine this possibility, we created an NF-B probe combining this portion of the consensus sequence and an additional 12 contiguous bases from the promoter. Use of this probe in a gel shift assay consistently showed an increase in binding activity with nuclei from iron-loaded TiO₂ treatment. In addition, RT-PCR analysis showed that gene expression of the -subunit was increased by iron loading and this effect was largely prevented by DFX. These findings thus support the idea that iron- driven NF-B activation might directly lead to increased transcription of PH mRNA.

In summary, these experiments show that an ordinarily nonfibrogenic particle, one that can serve as a model, relatively inert, fine PM particle, becomes fibrogenic when iron is added to the surface. These findings suggest that PM particles with bioavailable iron are all potentially fibrogenic, and might explain the morphologic abnormalities noted in the small airways in individuals living in regions of chronically high PM (5, 6). Because fibrogenic airway remodeling is associated with chronic airflow obstruction, our observations may provide a mechanistic basis for chronic airflow obstruction in such individuals.