Introduction

Top Abstract Introduction Materials And Methods Results Discussion References

Epidemiologic studies have shown a strong association between cardiopulmonary morbidity and mortality and levels of ambient particulate matter with a diameter of less than 10 µm (PM₁₀) (1, 2). In spite of consistent evidence supporting this relationship, the biologic mechanisms responsible for it are unclear. Seaton and colleagues proposed that the inhalation of fine particles provokes a low-grade inflammatory response in the lung that causes an exacerbation of pre-existing lung disease and a change in blood coagulability that results in increased pulmonary and cardiovascular deaths (3). Studies from our laboratory have shown that both small inert carbon particles (4) and ambient air particles (5) induced a systemic inflammatory response that includes stimulation of the bone marrow when they were instilled into the lungs of rabbits. We have also shown that an acute exposure to air pollution causes a leukocytosis in humans (6) and proinflammatory cytokines in the blood collected from these subjects increased during and decreased after the Southeast Asia haze in 1997 cleared (7). These data suggest that inflammatory mediators released from lung are capable of initiating not only a local inflammatory response but also a systemic response when PM₁₀ are deposited in the lung.

When alveolar macrophages (AMs) and airway epithelial cells are directly exposed to inhaled atmospheric particles, these small particles are phagocytized by both cells types. (5, 8). Several studies have shown that PM₁₀ stimulates the production of reactive oxygen species and inflammatory mediators by AMs (7, 9, 10) and airway epithelial cells (8, 11, 12). Both cell types can synthesize a variety of proinflammatory cytokines that influence the airway inflammatory response and contribute to the airway lesions in asthma and chronic obstructive pulmonary disease (12). In vivo, the close proximity of AMs to airway epithelial cells allows interaction between these two cell types that could amplify cytokine production.

The aim of this study was to test the hypothesis that cocultures of AMs and human bronchial epithelial cells (HBECs) amplify the response to PM₁₀ exposure compared with AMs and HBECs alone. We used primary cultures of HBECs and human AMs freshly isolated from lobectomy or pneumonectomy specimens and measured the expression of inflammatory cytokines. An RNase protection assay (RPA) that can simultaneously quantify several mRNA species in a single sample of total RNA and enzyme-linked immunosorbent assay (ELISA) were used to measure cytokine expression. We further determined whether supernatants from HBEC monocultures or AM/ HBEC cocultures incubated with PM₁₀ stimulate the bone marrow.

	Materials And Methods

Top Abstract Introduction Materials And Methods Results Discussion References

PM₁₀ Particles

PM₁₀ particles (EHC-93) were collected by the Environmental Health Directorate, Health Canada, Ottawa. A detailed analysis of the EHC-93 has been presented elsewhere (13). Particles were suspended in hydrocortisone-free supplemented bronchial epithelial cell growth medium (BEGM; Clonetics, San Diego, CA) at a concentration of 1 mg/ml and sonicated three times for 1 min each at maximal power on a Vibra Cell VC-50 sonicator (Sonics and Materials Inc., Danbury, CT) before adding the cells. The endotoxin content of the PM₁₀ suspension after sonication was measured using a commercially available kit (Limulus Amebo- cyte Lysate Test, QCL-1000; Bio.Whittaker, Inc., Walkersville, MD) according to the manufacturer's instructions.

Isolation of HBECs and Human AMs

Bronchial tissue and bronchial alveolar lavage (BAL) fluid was obtained from a total of eleven patients who underwent lobectomy or pneumonectomy for small peripheral nodules at St. Paul's Hospital, Vancouver. Informed consent was obtained from all subjects, and these studies were approved by the Human Ethics Committee of the University of British Columbia. All subjects were current smokers with a mean age of 62.5 yr (range 48- 70 yr). Primary HBECs were isolated from bronchial tissues according to a previously described procedure (8).

Human AMs were harvested from BAL fluid obtained from a noninvolved segment or lobe of the resected lung. These cells were washed twice in cold RPMI 1640 containing 10% fetal bovine serum (Life Technologies, Rockville, MD), antibiotics (100 U/ml of penicillin and 100 µg/ml of streptomycin; Sigma, St. Louis, MO) and Fungizone (1 µg/ml; GIBCO-BRL, Gaithersburg, MD). AM monocultures and AM/HBEC cocultures were suspended in BEGM. BEGM used throughout these experiments was without hydrocortisone. The BAL fluid cells were > 90% viable (trypan blue exclusion method) and consisted of 90-95% AMs and < 2% neutrophils. The remaining cells were lymphocytes.

Exposure of Cells to PM₁₀

Primary HBECs from the third or fourth passage were cultured to 90-100% confluence in 100-mm cell culture dishes (~ 2.5-3.0 × 10⁶ cells/dish) then exposed for 2 and 24 h to fresh stock suspensions of 100 and 500 µg/ml PM₁₀ (EHC-93) prepared in BEGM.

AMs (1.0 × 10⁷) were placed in 100-mm cell culture dishes and allowed to adhere to the plastic dish for 30 min in a humidified incubator (37°C and 5% CO₂). The nonadherent cells (< 1.0 × 10⁶) were then removed by rinsing twice with BEGM, and adherent AMs (> 98% AMs as assessed by Wrights-Giemsa stain) were incubated in 10 ml of BEGM with or without 100 µg/ml of PM₁₀ for 2 and 24 h.

In coculture experiments, freshly prepared AMs (5.0 × 10⁶) were directly placed on the confluent HBEC monolayers, which were grown in 100-mm cell culture dishes. The AMs were allowed to adhere to HBECs, and the nonadherent cells were removed by washing twice with BEGM. The AM/HBEC cocultures cells were incubated in 10 ml of BEGM with or without 100 µg/ml of PM₁₀ for 2, 8, and 24 h. Cell viability was determined following PM₁₀ exposure for 24 h in all experiments using the trypan blue exclusion method and neutral red assay for HBECs and AMs, respectively, as previously described (14).

Stimulation of HBECs by AM-Conditioned Medium

After AMs had been incubated with or without PM₁₀ (100µg/ml) for 24 h, the conditioned medium (CM) was collected, centrifuged, filtered through a syringe filter (0.22 µm; Corning, Cambridge, MA), and stored at 70°C. Primary HBECs, isolated from the same patients as the AMs, were grown to confluence in 60-mm cell culture dishes (Becton Dickinson, Rutherford, NJ) in BEGM, then incubated for 2 h in a mixture of 1 ml of AM-CM and 1 ml of BEGM. In a separate experiment, the additional effect of particles on HBECs was examined by adding 1 ml of 200 µg/ml of PM₁₀ suspension was added to 1 ml of AM-CM.

Immunocytochemistry

To demonstrate that AMs are in contact with HBECs when cocultured, HBECs grown to confluence on coverslips in 6-well plates (Corning) were seeded with AMs, cocultured as described above, and incubated with 100 µg/ml of PM₁₀ for 24 h. Cocultured cells were fixed with acetone for 10 min, and immunocytochemistry was performed by the alkaline phosphatase anti-alkaline phosphatase method using mouse anti-human CD68 monoclonal antibody (Dako, Glostrup, Denmark) to identify AMs.

RPA

Total RNA was isolated from cell cultures 2 and 8 h after addition of PM₁₀ or AM-CM using a single-step phenol/chloroform extraction procedure (Trizol; Life Technologies, Inc., Grand Island, NY). Cytokine mRNA levels were determined using the RPA, RiboQuant multiprobe system (PharMingen, San Diego, CA), following the instructions of the supplier. The customized template set for human RANTES, tumor necrosis factor (TNF)-, granulocyte macrophage colony-stimulating factor (GM-CSF), IL-1, monocyte chemotactic protein (MCP)-1, IL-6, leukemia inhibitory factor (LIF), stem cell factor, and oncostatin M (OSM) was used. Human IL-8 mRNA was determined using a separate template set. Internal controls included ribosomal protein (L32) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). In brief, 10 µg of total cellular RNA was hybridized overnight to the [³²P]UTP-labeled riboprobes, which had been synthesized from the supplied template sets. Single-stranded RNA and free probe remaining after hybridization were digested by an RNase A and T1 mixture. The protected RNA was then phenolized, precipitated, and analyzed on a 5% denaturing polyacrylamide gel. Following electrophoresis, the gel was dried and subjected to autoradiography, and the quantity of protected labeled RNA was determined using densitometry and the NIH image 1.62 software (National Institutes of Health, Bethesda, MD). Results were normalized to the expression of the internal control, GAPDH.

ELISA Measurements

Cell culture supernatants were collected 24 h after addition of 100 or 500 µg/ml of PM₁₀ suspension, centrifuged, and stored at 70°C. TNF-, GM-CSF, IL-1, IL-6, and IL-8 levels were measured by the Cytokine Core Laboratory (Baltimore, MD) using an ELISA based on a biotin-strepavidin-peroxidase detection system as previously described (8), and values were corrected for cell number. All measurements were done in triplicate, and values reported are the mean of five measurements.

Effect of Mediators from AM/HBEC Cocultures on the Bone Marrow

All animal studies were approved by the University of British Columbia Committee on Animal Care (Vancouver, BC). Female New Zealand white rabbits (weight 1.8-2.8 kg; n = 24) were injected with 5'-bromo-2'-deoxyuridine (BrdU) (100 mg/kg; Sigma) by intravenous infusion through the marginal ear vein at a concentration of 10 mg/ml in normal sterile saline over a period of 5 min. The rabbits were anesthetized 24 h later, and supernatants (1.0 ml/kg) from HBEC monocultures stimulated with 500 µg/ml of PM₁₀ for 24 h, or the supernatants from AM/HBEC cocultures (0.6 ml/kg) stimulated with 100 µg/ml of PM₁₀ for 24 h, were instilled intrabronchially under fluoroscopy as previously described (10). Peripheral blood samples were obtained from the central ear artery just before the BrdU injection and 3, 6, 9, 12, 24, 36, 48, 72, 96, 120, 144, and 168 h after the intrabronchial instillation. One milliliter of blood was collected in standard Vacutainer tubes containing ethylenediaminetetraacetic acid (Becton Dickinson) for leukocyte counts, and an additional 1 ml was collected in tubes containing acid citrate-dextrose for the detection of BrdU-labeled polymorphonuclear leukocytes (PMN^BrdU). Fentanyl (0.02 mg/kg) and droperidol (1.0 mg/kg) were injected subcutaneously for sedation before each time point of blood collection.

Total white blood cell counts were determined on a model SS80 Coulter Counter (Coulter Electronics, Hialeah, FL). Differential counts were obtained by counting 200 leukocytes in randomly selected fields of view on Wright-stained blood smears, and 100 PMNs were evaluated in randomly selected fields of view to determine the changes in the number of band cells. Blood collected in acid citrate-dextrose was used to obtain leukocyte-rich plasma as previously described (15). Leukocyte-rich plasma was cytospun onto 3-aminopropyltriethoxysilane-coated slides by cytocentrifugation at 180 × g for 5 min in a Cytospin 2 (Shandon Lab Products, Cheshire, UK). The cytospun specimens were air-dried and stained with the alkaline phosphatase-anti-alkaline phosphatase method to determine the fraction of PMN^BrdU in each specimen, as previously described (15). PMN^BrdU were divided into three groups (G1 to G3) according to the intensity of nuclear staining, using an arbitrarily designated grading system: weakly positive (staining of less than 5% of nucleus: G1), moderately positive (staining of 5 to 80% of the nucleus: G2), and highly positive (staining of more than 80% of nucleus: G3). The transit time of G3 cells represents the transit time of PMN through the postmitotic pool in the marrow, and the transit time of G1 cells represents the transit time of PMN through both the mitotic and the postmitotic pool in the marrow (15). All slides were coded and examined without knowledge of the group or sampling time, fields were selected in a randomized fashion, and 200 cells were evaluated per specimen to calculate the transit times of PMNs through the different pools in the marrow, as previously described (15).

Statistical Analysis

Data are expressed as mean values ± SE. The minimum number of replicates for all measurement was at least three. For RPA and ELISA, differences between matched pairs (control versus PM₁₀-treated) were compared by the Wilcoxon signed ranked test. For all rabbit experiments and for ELISA where cocultures exposed to PM₁₀ were compared with the sum of the monocultures of AMs and HBECs, unmatched pairs were compared by the Mann-Whitney U test. Differences between multiple groups were compared by one-way analysis of variance (ANOVA). The post hoc test for multiple comparison was the Bonferonni/Dunn method. Significance was assumed at P < 0.05.

	Results

Top Abstract Introduction Materials And Methods Results Discussion References

AM/HBEC Cocultures and PM₁₀

Figure 1 shows a coculture of AMs and HBECs immunocytochemically stained with CD68 to identify AMs. Most AMs and HBECs were in contact with each other. Both types of cells had internalized PM₁₀ particles, with many cells containing more than one particle. The PM₁₀ were not toxic to HBECs, and > 90% of cells were viable after a 24-h exposure of up to 500 µg/ml PM₁₀, as assessed by the trypan blue exclusion method. When the viability of AMs incubated with and without 100 µg/ml PM₁₀ for 24 h was compared using the neutral red assay, no significant difference was found (n = 3, P = 0.499). The EHC-93 is known to contain small amounts of lipopolysaccharide (9) and the endotoxin content of the PM₁₀ suspension of 100 µg/ml was 6.4 ± 1.8 EU/ml (n = 4) or less than 3.0 ng/ml. This dose of lipopolysaccharide has been shown not to activate either AMs or HBECs to produce cytokines or to activate the bone marrow when instilled into the lung of rabbits (8, 10).

View larger version (131K): [in this window] [in a new window]   Figure 1.   Photomicrograph of cocultured primary HBECs and human AMs incubated with 100 µg/ml of PM10 (EHC-93) for 24 h showing particles internalized by both HBECs and AMs. Cells were cocultured on coverslips, and immunocytochemistry was performed using mouse anti-human CD68 monoclonal antibody to identify AMs. The bar represents 10 µm.

Effect of PM₁₀ on Expression of Cytokine mRNA

Representative autoradiographs of cytokine mRNA expression by AM/HBEC cocultures after 2 and 8 h incubation in medium alone (control) or a 100 µg/ml of PM₁₀ suspension are shown in Figure 2. TNF-, GM-CSF, IL-1, IL-6, LIF, OSM, and IL-8 mRNA expression by AM/ HBEC cocultures were increased after incubation with PM₁₀ for 2 h. MCP-1 mRNA was constitutively expressed. Similar results were shown after 8 h incubation. Densitometric analysis of four RPAs confirmed that the mRNA levels of these seven cytokines at 2 h were significantly higher compared with control values (Figure 3).

View larger version (74K): [in this window] [in a new window]   Figure 2.   The mRNA expression of cytokines after incubation of primary HBECs and human AMs for 2 h and cocultures for 2 and 8 h in medium alone (control) or medium with 100 µg/ml PM10. Representative autoradiographs of the RPA 2 h after exposure to PM10 show increased mRNA expression of IL-1 and IL-8 in AM/HBEC cocultures, AM monocultures and HBEC monocultures; TNF-, IL-6 and OSM in AM/HBEC cocultures and AM monocultures; LIF in AM/HBEC cocultures and HBEC monocultures. GM-CSF mRNA expression was induced within 2 h of exposure to PM10 in AM/HBEC cocultures only and remained elevated after 8 h of exposure. RANTES mRNA was not detected in any of the cultures. MCP-1 mRNA was constitutively expressed and not induced by PM10 exposure in either AM/HBEC cocultures or AM monocultures. L32 and GAPDH were used as controls for lane loading conditions.

View larger version (19K): [in this window] [in a new window]   Figure 3.   TNF-, GM-CSF, IL-1, IL-6, LIF, OSM, and IL-8 mRNA expression in AM/HBEC cocultures after 2 h of exposure to PM10. The density of the bands representing the cytokine mRNA was compared with that of the GAPDH mRNA band in the same lane, and the resulting ratio (filled bar) is shown as the percentage change from control (open bar) values. Values are means ± SEM of four experiments. *P < 0.01, **P < 0.05 compared with control.

The expression of TNF- (P < 0.05), IL-1 (P < 0.05), and IL-6 (P < 0.05) mRNA by AM monocultures was also significantly increased by 100 µg/ml of PM₁₀ stimulation for 2 h compared with control values (Figure 2 and data not shown). Although higher than control values, OSM and IL-8 mRNA expression were not significantly increased at 2 h. MCP-1 mRNA was constitutively expressed.

HBECs stimulated by 100 µg/ml of PM₁₀ for 2 h increased mRNA expression of IL-1 (P < 0.05) and LIF (P < 0.05), with a small nonsignificant increase in IL-8 mRNA expression (P = 0.193) (Figure 2 and data not shown).

Figure 4A compares cytokine mRNA expression in HBECs cultured in control unconditioned medium to cells exposed for 2 h to CM from AMs incubated for 24 h with 100 µg/ml of PM₁₀ (AMs-CM_PM10) or without PM₁₀ (AMs-CM_cont). When HBECs were exposed to AMs-CM_PM10, IL-1, LIF, and IL-8 mRNA expression were increased compared with both control or AMs-CM_cont exposure. TNF- and GM-CSF mRNA were weakly expressed after exposure to AMs-CM_PM10. IL-1, IL-6, LIF, and IL-8 mRNA expression were enhanced further by addition of 100 µg/ml PM₁₀ particles to the control unconditioned or either conditioned media, with no change in TNF- or GM-CSF mRNA expression (Figure 4B). The experiments were repeated twice and showed similar results (data not shown).

View larger version (58K): [in this window] [in a new window]   Figure 4.   A shows a representative RPA of cytokine mRNA expression in HBECs exposed for 2 h to unconditioned medium (control) and to CM from AMs incubated for 24 h without PM10 (AMs-CMcont) or with 100 µg/ml PM10 (AMs-CMPM10). IL-1, LIF, and IL-8 mRNA expression was increased by AMs-CMPM10 exposure compared with control or to AMs-CMcont exposure. B shows a representative RPA of cytokine mRNA expression of HBECs exposed for 2 h with the same conditioned as in A except 100 µg/ml PM10 was added to the medium (control + PM10, AMs-CMcont + PM10, AMs-CMPM10 + PM10). IL-1, IL-6, LIF, and IL-8 mRNA expression was further enhanced by addition of PM10 particles.

Effect of PM₁₀ on Release of Cytokines from AMs and HBECs

Figure 5 shows the GM-CSF, IL-6, IL-8, IL-1, and TNF- protein levels in supernatants of AM/HBEC cocultures, AM monocultures, and HBEC monocultures incubated for 24 h with medium alone (control) or with 100 µg/ml of PM₁₀ and, in addition, HBEC monocultures incubated with 500 µg/ ml of PM₁₀. GM-CSF, IL-6, IL-8, IL-1, and TNF- production by AM/HBEC cocultures and AM monocultures stimulated with 100 µg/ml of PM₁₀ suspension increased significantly compared with controls. GM-CSF, IL-8, and IL-1 production by HBEC monocultures stimulated by 500 µg/ ml of PM₁₀ was increased over control levels and appeared to increase in a PM₁₀ dose-dependent manner. IL-6 or TNF- were not increased in HBEC monocultures.

View larger version (19K): [in this window] [in a new window]   Figure 5.   GM-CSF, IL-6, IL-8, IL-1, and TNF- protein levels in supernatants of AM/HBEC cocultures, AM monocultures, and HBEC monocultures incubated for 24 h with medium alone (control; white bars) or 100 µg/ml (black bars) and HBEC monocultures incubated for 24 h with 500 µg/ml of PM10 (gray bars). GM-CSF, IL-6, IL-8, IL-1, and TNF- production by AM/HBEC cocultures and AM monocultures stimulated by PM10 (100 µg/ml) increased significantly compared with controls. GM-CSF, IL-8, and IL-1 production by HBEC monocultures stimulated by 500 µg/ml of PM10 suspension were also increased. GM-CSF and IL-6 production by AM/HBEC cocultures significantly exceeded the sum of the cytokines produced by these cells exposed separately. Values are means ± SEM of five experiments. *P < 0.01, **P < 0.05 compared with control. P < 0.05, AM/HBEC cocultures compared with the sum of HBEC monocultures plus AM monocultures.

The GM-CSF and IL-6 produced by AM/HBEC cocultures were more than the sum of these produced by AM and HBEC monocultures alone, suggesting a synergistic effect in production of these cytokines (P < 0.05). This synergistic effect was not found in the production of IL-8, IL-1, and TNF-.

Bone Marrow Stimulation by Supernatants of AMs and HBECs

Instillation of the supernatants from AM/HBEC cocultures incubated with 100 µg/ml PM₁₀ (Coculture-PM₁₀ group) caused an increase in circulating band cell counts at 6 and 24 h (Figure 6A) compared with the instillation of control supernatants (Coculture-control group).

View larger version (25K): [in this window] [in a new window]   Figure 6.   The number of band cells (A), all BrdU-labeled PMNs (B), G3 cells (see text) (C), and G1 cells (see text) (D) in the circulation of rabbits after instillation of supernatants from AM/HBEC cocultures with (Coculture-PM10 group; n = 6) and without (Coculture-control group; n = 6) PM10. White arrow: BrdU labeling, black arrow: supernatant instillation. Values are means ± SEM. **P < 0.05 compared with AM/HBEC cocultures without PM10 (Coculture control group).

Figure 6B shows the number of all PMN^BrdU in the circulation over time. The number of PMN^BrdU in the peripheral blood of the Coculture-PM₁₀ group was higher at 6, 12, and 24 h compared with the Coculture-control group. Figure 6C shows the G3 cells (representing the transit of PMNs through the postmitotic pool), and Figure 6D shows the G1 cells (representing the transit of PMNs through the mitotic plus postmitotic pools). Both G3 (Figure 6C) and G1 (Figure 6D) cells were released earlier by supernatants from the Coculture-PM₁₀ group compared with those from the Coculture-control group.

Table 1 shows the calculated transit times of all the different populations of PMN^BrdU (all PMN^BrdU and G3, G2, and G1 cells). Supernatants from PM₁₀-stimulated cocultures shortened the transit time of PMNs through the bone marrow (all PMN^BrdU). The transit time of PMNs through the mitotic plus postmitotic (G1) pools was shorter in Coculture-PM₁₀ group.

Supernatants from HBEC monocultures stimulated with 500 µg/ml of PM₁₀ for 24 h did not increase band cell counts (data not shown) or affect the transit time of the four population of PMNs compared with the control supernatants (Table 1).

	Discussion

Top Abstract Introduction Materials And Methods Results Discussion References

AMs and epithelial cells play a key role in processing inhaled particulate matter. We have previously shown that both human AMs and bronchial epithelial cells produce cytokines when they phagocytose atmospheric particles (7, 8). The present study extends these observations by showing that human AMs and HBECs interact when they were cocultured ex vivo and exposed to atmospheric PM₁₀. This results in the upregulation of TNF-, GM-CSF, IL-1, IL-6, and IL-8 mRNA expression, protein secretion by both cell types, and a synergistic increase in the production and secretion of GM-CSF and IL-6. These two cytokines are known to be involved in the bone marrow response that follows PM₁₀ exposure (7). This is supported by studies where supernatants from the cocultured AMs and HBECs exposed to PM₁₀ accelerated the transit time of PMNs through the bone marrow and stimulated the release of immature PMNs into the circulation when instilled into the lungs of rabbits. The immature PMNs released from the marrow during this response enhance the inflammatory response in the lung (16, 17) and contribute to endothelial damage. The data presented here show that the interaction between AMs and HBECs during exposure to atmospheric particles amplifies the production of mediators that initiate the systemic inflammatory response and suggest that these mediators can also modulate the local inflammatory response in the lung.

Studies from our laboratory, both in vitro and in vivo, have shown that AMs are crucially important in producing the mediators responsible for the bone marrow response induced by exposure to PM₁₀ (7, 10, 18). We (8) and others (11, 12, 19) have also shown that HBECs exposed to PM₁₀ produced proinflammatory mediators. Because of the close proximity of AMs and epithelial cells in the lung, we postulated that interactions between these two cell types will enhance the proinflammatory response elicited by PM₁₀ exposure. There are several studies that used transwell systems to coculture lung cells to avoid cell adherence responses between airway epithelial cells and AMs or other lung cells (20, 21). Previous studies using AM-CM to stimulate airway epithelial cells (22, 23) showed that the secretion of cytokines, such as IL-1 and TNF-, by AMs could amplify the expression of IL-6 and IL-8 by airway epithelial cells. Our results show that CM from PM₁₀- exposed AMs caused a small increase in HBECs' mRNA expression of cytokines, suggesting that cell-cell contact between AMs and epithelial cells is of critical importance in generating inflammatory mediators in response to PM₁₀ exposure. Several studies have suggested that cellular contact between lung cells (e.g., epithelial cells, endothelial cells, and fibroblast) is necessary for activation and cytokine expression (24, 25). Using a transwell system, Arnold and colleagues reported that IL-8 release by cocultured human pulmonary type II-like epithelial cells (A549) and peripheral blood mononuclear cells produced only 30% of the IL-8 released compared with when direct cell-to-cell contact occurred (26). In our experiments, we allowed direct cell-to-cell contact between human AMs and bronchial epithelial cells of the same subject before exposure to ambient particles to simulate in vivo contact of these cells. These cocultured AMs and HBECs produced several cytokines, including TNF-, GM-CSF, IL-1, IL-6, and IL-8. GM-CSF, IL-1, and IL-6 have been implicated in the systemic inflammatory response observed during an acute episode of air pollution in humans (6, 7).

The dose of PM₁₀ we have used in these experiments (~ 23 pg/epithelial cell) is similar to a calculated 24-h exposure of subjects during the South East Asia forest fires of 1997 (6) exposed to 150 µg/m³. To directly extrapolate the magnitude of the response observed in our cell culture system to a response of the lung cells in vivo is not possible and should be tested in human or animal exposure experiments. However, our study supports the notion that a qualitatively equivalent response could be initiated in vivo with deposition of inhaled air pollution particles in the lung.

The cytokines produced by cocultures of AMs and HBECs stimulated with PM₁₀ could be released into the circulation and either stimulate the bone marrow directly or stimulate other effector cells in the lung and distal organs (27). GM-CSF is an important stimulant of the turnover and release of granulocytes and monocytes from the bone marrow (27). It also activates circulating leukocytes (27) and prolongs their survival in the circulation (28). IL-6 stimulates the bone marrow to produce platelets and the liver to produce coagulation factors (27), and these prothrombotic factors could contribute to the cardiovascular events associated with exposure to particulate air pollution. Our laboratory has reported that IL-6 accelerates the transit time of PMNs through the bone marrow and promotes their sequestration in the lung microvascular beds (29). IL-8 is one of the most potent activators and chemoattractants for neutrophils and contributes to neutrophilic inflammation in chronic bronchitis and emphysema (30). It is also a potent chemotactic factor for eosinophils and T lymphocytes and could influence the pathophysiology of allergic airway disease (31). Although IL-8 is known to stimulate the bone marrow to release leukocytes, previous studies from our laboratory have shown that IL-8 caused a rapid release of PMNs residing in the bone marrow venous sinusoids without affecting the transit time of these cells through the bone marrow (18). These data may explain why supernatants from HBECs alone, containing a considerable quantity of IL-8, failed to shorten the PMNs' transit time through the bone marrow. IL-1 and TNF- promote neutrophilic and eosinophilic inflammation, because they can stimulate a number of different cell types to increase the expression, synthesis, and release of several cytokines and cell adhesion molecules (12, 32). IL-6, IL-8, and GM-CSF expression are enhanced by IL-1 and TNF- stimulation of HBECs (33). IL-1 and TNF- generated by AMs following PM₁₀ exposure may stimulate HBECs in a paracrine fashion, resulting in the production of cytokines that could contribute to the stimulation of the bone marrow. This is supported by our finding of synergy between AMs and HBECs in producing GM-CSF and IL-6 (Figure 5), two cytokines that are capable of stimulating the bone marrow (29).

Airway epithelial cells exposed to PM₁₀ produce several cytokines (8, 12, 34). To determine the contribution of cytokines produced by HBECs on the systemic inflammatory response induced by PM₁₀, we instilled the supernatants of HBECs exposed to these particles into lungs of rabbits and measured the bone marrow response. These studies showed that supernatants of HBECs exposed to high doses of PM₁₀ (500 µg/ml) failed to stimulate the marrow (Table 1). This suggests that the proinflammatory mediators released from HBECs when exposed to PM₁₀ alone do not contribute to the systemic inflammatory response induced by exposure to ambient particles. However, supernatants of HBECs cocultured with human AMs instilled into the lung significantly shortened the transit time of PMNs through the marrow. This effect was present with a five times lower dose of PM₁₀ (100 µg/ml). Studies from our laboratory show that the rabbit is a useful model to study the systemic inflammatory response induced by mediator release from human AMs exposed to PM₁₀ (10). Collectively, these results suggest that AMs are an important source of inflammatory mediators responsible for the systemic response after PM₁₀ exposure. The synergistic production of GM-CSF and IL-6 by cocultured AMs and HBECs suggests that cell-cell interactions contribute to the systemic response to PM₁₀ exposure. The cellular source of the increased IL-6 and GM-CSF in cocultures is unclear. The reduced capacity of HBECs to produce IL-6 points to the AMs as the principal source of IL-6 in the cocultures, in contrast to GM-CSF that can be produced by either AMs or HBECs. We suspect that IL-1 produced by HBECs following exposure to PM₁₀ (8) stimulates AMs to enhance their production of IL-6 and GM-CSF (35). Collectively, our results suggest that HBECs exposed to PM₁₀ produce inflammatory mediators that are involved in modulating the local inflammatory response in the lung, but these mediators also have the potential to augment the mediator production from AMs.

In this study we have used cocultures of AMs and HBECs to model the environment in the lung and provide insight into the inflammatory response induced by particulate matter air pollution. Exposure of cocultures of AMs and HBECs to ambient particles increases their inflammatory cytokine production measured at both the mRNA and protein levels: specifically, GM-CSF, IL-6, IL-8, IL-1, and TNF-. GM-CSF and IL-6 were produced in a synergistic fashion. These mediators also stimulate the bone marrow to release immature PMNs into the circulation, and we postulate that they augment the local inflammatory response in the lung. The results of this study suggest that both AMs and HBECs contribute to the pathogenesis of pulmonary and cardiovascular disease associated with particulate air pollution.