Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

In general, eosinophilic effusion is defined as an effusion in which eosinophils account for more than 10% of the white blood cells. Pleural effusions secondary to pneumothorax, hemothorax, pulmonary embolism, asbestosis, drug reaction, parasitic disease, and Churg-Strauss syndrome are frequently eosinophilic (1). Previous studies showed that interleukin (IL)-5, granulocyte macrophage-colony stimulating factor (GM-CSF), and IL-3 are involved in eosinophilic effusions via stimulation of eosinophil survival (2). The platelet activating factor (PAF) levels in pleural fluids were reported to correlate with the numbers of eosinophils (3). However, PAF seemed to be produced by eosinophils themselves, suggesting that increased PAF levels in effusions may not be a cause but a result of eosinophilia. Thus, the mechanisms of eosinophilic accumulation in the pleural cavity are not well understood.

There are several chemokines that attract eosinophils, such as eotaxin (4) and regulated on activation, normal T cells expressed and secreted (RANTES) (5). Eotaxin was purified from bronchoalveolar lavage fluids of a rat asthma model in 1994 (4). It is now known to be released from a variety of cells, such as bronchial epithelial cells, nasal epithelial cells, fibroblasts, eosinophils, macrophages, and T cells (6). The C-C chemokine receptor 3 (CCR3) is the only receptor to bind eotaxin. RANTES was cloned from an activated T-cell cDNA library in 1988 (5). RANTES may be produced by T cells, platelets, monocytes, macrophages, fibroblasts, bronchial epithelial cells, vascular endothelial cells, and eosinophils. It attracts not only eosinophils but also monocytes and lymphocytes (6). Receptors for RANTES are CCR1, CCR3, and CCR5 (6, 7).

Although eosinophilic effusion itself is not an informative diagnostic finding, its presence is reported to be associated with good prognosis in a large prospective study (8). Thus, knowledge of the underlying mechanism responsible would be important for understanding the mechanisms of not only tissue eosinophilia but also a more favorable immune response to pleural disease. Recently, we found that eotaxin is easily detectable in almost all pleural effusions, and its level is significantly higher in eosinophilic effusions than in non-eosinophilic effusions (9). Furthermore, the number of eosinophils in pleural effusions was significantly correlated with the concentration of eotaxin. Because pleural eotaxin could be identified regardless of the etiology of the effusion, we speculated that resident cells (such as mesothelial cells), rather than infiltrating inflammatory cells, produce eotaxin. However, it is not known whether mesothelial cells can produce eotaxin.

To clarify this point, we purified normal pleural mesothelial cells (NPMC) from pleural effusions and investigated their production of eotaxin and RANTES. We also studied the regulatory mechanisms for production of chemokines from NPMC via cytokine networks.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

Human recombinant tumor necrosis factor (TNF)-, IL-4, interferon (IFN)-, eotaxin, and RANTES were purchased from Pepro Tech (London, UK). Monoclonal antihuman eotaxin antibody (IgG₁) and monoclonal antihuman RANTES antibody (IgG₁) were obtained from Genzyme-Techne (Cambridge, MA). Actinomycin D was obtained from Sigma (St. Louis, MO).

Primary Culture of NPMC

Pleural fluids were collected from donors with lung cancer and congestive heart failure. NPMC were purified by a previously described method (10). Briefly, pleural fluids were centrifuged at 1,500 rpm for 5 min, and cell pellets were washed three times using RPMI 1640 medium (Nikken Bio Medical Laboratory, Kyoto, Japan). The pellets were resuspended in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum (Gibco, Grand Island, NY), 100 U/ml of penicillin G (Banyu Pharmaceutical, Tokyo, Japan), 100 µg/ml of streptomycin (Meiji Seika, Tokyo, Japan), and 10 mM of Hepes (Sigma) (complete medium). The cells were placed in 100-mm Petri dishes (Nunc, Roskilde, Denmark) and cultured in a humidified atmosphere of 5% CO₂ at 37°C overnight. The dishes were gently washed using warmed medium to remove nonadherent cells. The cultured medium was replaced with fresh complete medium supplemented with 10 ng/ml of epidermal growth factor (HIH Biocenter, Tokyo, Japan) once a week and maintained for more than 4 wk. A small sample of cells was immunostained to determine the purity of NPMC by using an antikeratin (DAKO-CK-1; DAKO PATTS, Glostrup, Denmark) or an antivimentin (DAKO-Vimentin; DAKO-PATTS) antibody and Vectastain Elite ABC kits (Vector Laboratories, Burlingame, CA). The cultured NPMC were positive for both keratin and vimentin.

Established mesothelial cells were cultured at a concentration of 5 × 10⁵/ml (10 ml per dish) in RPMI 1640 medium containing 0.1% bovine serum albumin (Sigma), 100 U/ml penicillin G, 100 µg/ml streptomycin, and 10 mM Hepes in the presence or absence of TNF-, IL-4, IFN-, TNF- plus IL-4, or TNF- plus IFN-. We added these cytokines at approximately 70% of full confluent condition. Culture supernatants were collected and frozen until use for determination of chemokine protein levels or eosinophil chemotactic activity. Cytokines were used at 10 ng/ml unless otherwise stated.

Determination of Protein Levels of Eotaxin and RANTES

We used a specific enzyme-linked immunosorbent assay (ELISA) to measure eotaxin as described previously (11). RANTES was measured using a commercially available kit purchased from Endogen (Boston, MA), according to the manufacturer's recommendation. The detection limits were 5 pg/ml and 2 pg/ml for eotaxin and RANTES, respectively.

cDNA Probes

Tissue total RNA was extracted from cells of each dish by the protocol of ISOGEN (Nippon Gene, Tokyo, Japan). Total RNA for chemokine cDNA probes was extracted from the human mesothelioma cell line EH-MES1. Reverse transcription was performed at 42°C for 60 min using a cDNA synthesis kit (Life Sciences, St. Petersburg, FL), and the polymerase chain reaction was performed at 94°C for an initial 5 min followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 58°C for 45 s, and extension at 72°C for 60 s. The primers for eotaxin were sense, 5'-CTT AAGCTTCCAACATGAAGGTCTCCGC-3'; antisense, 5'-ACA CTCGAGGCTCTGGTTTGGTTTCAA-3' and for RANTES were sense, 5'-CCACAGGTACCATGAAGGTCTCCGC-3'; antisense, 5'-TCAAGAGCTCTCCATCCTAGCTCATC-3'. The cDNA probes were labeled with [-³²P]dCTP by the random oligonucleotide primer technique using a Prime It II kit (Stratagene, La Jolla, CA).

Northern Blot Analysis

Denatured RNAs (10 µg) were size fractionated by gel electrophoresis on 1% agarose/5% formaldehyde gels containing 20 mM morpholinosulphonic acid and transferred to a nylon membrane (Hybond-N⁺; Amersham, Buckinghamshire, UK). The filter was prehybridized for 15 min at 68°C in QuickHyb Hybridization Solution (Stratagene) with 20 µg/ml salmon sperm DNA (Sigma) and hybridized in the same buffer containing 50 ng/ml heat-denatured -³²P-labeled probes at 68°C for 1 h. After hybridization, the membrane was washed twice in 2 × saline sodium citrate (SSC) (1 × SSC = 150 mM sodium chloride and 15 mM sodium citrate) with 0.1% sodium dodecyl sulfate (SDS) at room temperature for 15 min and in 0.1 × SSC with 0.1% SDS at 60°C for 30 min. The blots were exposed to X-ray film at 80°C for 48 h.

Chemokine mRNA Stability Assay

After treatment with TNF- (10 ng/ml) for 48 h, NPMC were exposed to fresh medium with or without IL-4 or IFN- in the presence of actinomycin D (10 µg/ml), and total cellular RNA was extracted from the cells after 0, 2, 4, 8, and 16 h of incubation. Chemokine mRNA expression was detected by Northern blotting. After quantitation by densitometry, relative levels of RNA loaded were normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA levels.

Eosinophil Separation

Discontinuous Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden) gradients were prepared according to a previously described method with some modification (12). Nine parts Percoll were mixed with one part 1.5 M NaCl. This solution and 0.15 M NaCl were then mixed to make mixtures of the following densities: 1.100, 1.090, 1.085, 1.080, and 1.070 g/ml. Heparinized venous blood was collected from healthy volunteers. Five parts blood were mixed with one part 6% Dextran T70 (Pharmacia Fine Chemicals) in 0.15 M NaCl and left at room temperature for 60 min to let the red cells sediment. The dextran-plasma was collected and centrifuged at 1,500 rpm for 8 min. The cells were washed twice in phosphate-buffered saline (PBS) supplemented with 2 mM ethylenediaminetetraacetic acid (EDTA) (PBS/2 mM EDTA) and suspended in Percoll solution (density 1.070 g/ml). The cell suspension was layered on top of the Percoll gradients with multiple densities in 15-ml tubes (Nunc) and centrifuged at 3,000 rpm for 20 min at room temperature. The eosinophil fraction was found at the density of 1.087-1.100 g/ml and collected. Red blood cells were removed by hypotonic lysis. The cells were added onto 50 µl of MACS CD16 MicroBeads (Miltenyi Biotech, Bergish Gladbach, Germany) per 5 × 10⁷ total cells, incubated for 30 min at 7°C, adjusted with PBS/2 mM EDTA to a final volume of 1 ml per 5 × 10⁷ total cells, and subjected to magnetic separation. Both purity and viability were more than 95% as determined by examination of slides stained with Hansel solution (Torii Pharmaceutical, Tokyo, Japan) and by the trypan blue dye-exclusion test, respectively.

Measurement of Eosinophil Chemotactic Activity

Eosinophil chemotactic activity was determined in 24-well microchemotaxis assemblies. The bottom wells (24-well tissue culture plates; Costar, Cambridge, MA) of the assembly were filled with 600 µl of fluid containing the chemotactic stimulus or media. Eosinophils (1 × 10⁶/ml) were added to the upper chamber (Transwell; Costar). After incubation for 2 h in a humidified atmosphere of 5% CO₂ at 37°C, the eosinophils that migrated across the filter and adhered to the bottom side of the filter were counted in 10 random high-power fields (×400) per well using light microscopy.

Statistical Analysis

Data are expressed as mean ± standard error of the mean. The differences among groups were tested by analysis of variance with post-hoc comparisons using the Scheffè procedure or Mann-Whitney U-test. Statistical significance was defined as P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Production of Eotaxin and RANTES by NPMC

Unstimulated NPMC produced a small amount of eotaxin but not RANTES. We evaluated the effects of IL-4, IFN-, and TNF- (Figure 1) and found that IL-4 was a potent inducer for eotaxin but not RANTES. IFN- induced small amount of RANTES but not eotaxin. TNF- induced both eotaxin and RANTES. These effects of IL-4 and TNF- were dose dependent (data not shown). Increased doses (up to 100 ng/ml) of IFN- did not induce eotaxin, and increased doses of IL-4 or IFN- did not induce RANTES production. As shown in Figure 1, there is some interindividual variability in terms of the amount of chemokines produced, but the effects of the cytokines are essentially the same among the individuals.

View larger version (26K): [in this window] [in a new window]   Figure 1.   Time-dependent production of eotaxin (left panels) and RANTES (right panels) from NPMC stimulated with cytokines. NPMC from three different donors (A, B, and C) were cultured with or without () TNF- (, 50 ng/ml), IL-4 (, 50 ng/ml), or IFN- (, 50 ng/ml) for the indicated time periods. The concentrations of chemokines in the supernatants were determined by ELISA. Data are expressed as mean ± SEM for triplicate samples. *P < 0.05, versus time point of 8 h. P < 0.05, versus supernatant without cytokine stimulation.

Next, we evaluated the effects of combinations of cytokines. Production of eotaxin induced by IL-4 was further increased in a dose-dependent manner by the addition of TNF- (Figure 2, left panels). For RANTES, production induced by TNF- was synergistically increased in the presence of IL-4 or IFN- (Figure 2, right panels). Addition of IFN- suppressed the production of eotaxin induced by IL-4 or TNF- (Figure 3A). In contrast, IFN- increased, in a dose-dependent manner, RANTES production induced by TNF- (Figure 3B).

View larger version (29K): [in this window] [in a new window]   Figure 2.   Effect of TNF- on eotaxin (left panels) and RANTES (right panels) production by NPMC. NPMC from three different donors (A, B, and C) were cultured with or without () IL-4 (, 10 ng/ml) or IFN- (, 10 ng/ml) in combination with serially diluted TNF- for 48 h. At the end of the incubation period, the concentration of chemokines in each supernatant was determined by ELISA. Data are expressed as mean ± SEM for triplicate samples. *P < 0.05, versus cultures without TNF-. P < 0.05, versus cultures without IL-4 or IFN-.

View larger version (13K): [in this window] [in a new window]   Figure 3.   Effect of IFN- on eotaxin (A) and RANTES (B) production from NPMC. NPMC were cultured with or without () TNF- (, 10 ng/ml) or IL-4 (, 10 ng/ ml) in combination with serially diluted IFN- for 48 h. At the end of the incubation period, the chemokines in each supernatant were determined by ELISA. Data are expressed as mean ± SEM for triplicate samples. A representative result out of three independent experiments with similar results is shown. *P < 0.05, versus cultures without IFN-. P < 0.01, versus cultures without TNF- or IL-4.

Chemokines mRNA Expression by NPMC

As observed at the protein level, IL-4 or TNF- alone enhanced the expression of eotaxin mRNA, and IFN- or TNF- alone increased RANTES mRNA expression (Figure 4). The combination of TNF- and IL-4 markedly increased eotaxin mRNA, whereas the addition of IFN- decreased eotaxin mRNA expression. TNF- strongly augmented, and IL-4 mildly enhanced, RANTES mRNA induced by IFN-.

View larger version (47K): [in this window] [in a new window]   Figure 4.   Northern blot analysis of eotaxin and RANTES expression. Total RNA of NPMC was extracted following culture (48 h) with or without (lane 1) TNF- (10 ng/ml, lane 2), IL-4 (10 ng/ml, lane 3), IFN- (10 ng/ml, lane 4), TNF- plus IL-4 (lane 5), IL-4 plus IFN- (lane 6), TNF- plus IFN- (lane 7), or TNF- plus IL-4 plus IFN- (lane 8). A representative result out of three independent experiments with similar results is shown.

Effect of Cytokines on Chemokine mRNA Stability

To further investigate the mechanisms for the combined effects of cytokines on chemokine production, we investigated the effect of IL-4 or IFN- on the stability of TNF--induced mRNA. Although the stability of eotaxin mRNA was not affected by either cytokine, that of RANTES mRNA was obviously increased by IFN- but not IL-4 (Figure 5).

View larger version (31K): [in this window] [in a new window]   Figure 5.   Effect of IL-4 and IFN- on the stability of eotaxin or RANTES mRNA. After treatment with TNF- (10 ng/ml) for 48 h, NPMC were exposed to fresh medium without () or with IL-4 () or IFN- () in the presence of actinomycin D (10 µg/ml). Total RNA was extracted from the cells after the indicated time periods. Chemokine mRNA expression was examined by Northern blotting. After normalization of expression using the GAPDH mRNA level, data are presented as percent changes in reference to the 0-time level that was set to 100%. A representative result out of three separate experiments with similar results is shown.

Eosinophil Chemotactic Activity of Culture Supernatant of NPMC

As shown in Figure 6, chemotactic activity for eosinophils was observed in culture supernatants of NPMC stimulated with TNF- alone and TNF- plus IL-4 or IFN-. A cytokine alone or combinations of cytokines did not induce the migration of eosinophils. Although supernatants from unstimulated NPMC induced the migration of eosinophils, this effect was proven by checkerboard analysis (data not shown) to be due to chemokinesis but not chemotaxis.

View larger version (19K): [in this window] [in a new window]   Figure 6.   Effect of culture supernatants of NPMC on eosinophil migration. Mesothelial cells from three different donors (A, B, and C) were preincubated in the presence or absence of TNF- (10 ng/ml), IL-4 (10 ng/ml), IFN- (10 ng/ml), TNF- plus IL-4, or TNF- plus IFN- for 48 h. The lower wells of the chemotaxis chambers contained supernatants of these cultures (closed bars) or the cytokine(s) at the same concentration (negative control: open bars). The migrated cells were counted and presented as the number of eosinophils per high-power field. *P < 0.05, compared with eosinophil migration induced by medium alone (spont). P < 0.05, compared with eosinophil migration of negative controls.

In our system, both recombinant eotaxin and RANTES induced eosinophil migration in a dose-dependent manner. Neutralizing antibody against eotaxin showed a dose-dependent inhibition of eosinophil migration induced by recombinant eotaxin (50 ng/ml) but not RANTES (50 ng/ml), and the maximum inhibition (52%) was observed at 1-10 µg/ml of antibody (data not shown). Antibody against RANTES inhibited, in a dose-dependent manner, eosinophil migration induced by recombinant RANTES but not eotaxin, and the maximum inhibition (87%) was observed at 10 µg/ml of antibody (data not shown). Eosinophil migration induced by the culture supernatant of NPMC stimulated with TNF- plus IL-4 was inhibited by a neutralizing antibody against eotaxin but not RANTES, and that induced by the culture supernatant of cells stimulated with TNF- plus IFN- was attenuated by an antibody against RANTES but not eotaxin (Figure 7).

View larger version (27K): [in this window] [in a new window]   Figure 7.   Effect of neutralizing antibodies on eosinophil migration induced by culture supernatants of mesothelial cells from three different donors (A, B, and C) stimulated with TNF- plus IL-4 (left panels) or TNF- plus IFN- (right panels). The lower wells of the chemotaxis chambers contained supernatants of these cultures with neutralizing antibodies against either eotaxin (), or RANTES (), or both (). The migrated cells were counted and presented as the percentage (A 20.4 and 12.1; B 21.9 and 10.7; and C 21.4 and 12.2 eosinophils/hpf was considered 100%, respectively) of the number of eosinophils migrated without neutralizing antibodies. *P < 0.05, compared with eosinophil migration without neutralizing antibodies.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We demonstrated that NPMC has a potency to produce biologically active eotaxin and RANTES. The production of eotaxin and RANTES was regulated by different cytokines and different mechanisms. A single cytokine, TNF-, induced both eotaxin and RANTES. IL-4 stimulated eotaxin but not RANTES production. IFN- induced small amount of RANTES but not eotaxin. The most powerful stimulation for eotaxin was TNF- plus IL-4, and that for RANTES was TNF- plus IFN-. The effect of IFN- on TNF--induced RANTES production was at least in part due to the increased stability of RANTES mRNA. On the other hand, IFN- inhibited eotaxin induced by TNF- or IL-4. IFN- did not affect the stability of eotaxin mRNA, indicating that this inhibition was induced at transcriptional level.

The above observations indicate that NPMC can produce different eosinophil chemokines in response to Th1 and Th2 cytokines (i.e., they produce eotaxin in a Th2-dominant inflammatory response, and they produce RANTES in a Th1-dominant response). Such relationships have also been observed in fibroblasts (13), bronchial epithelial cells (16), and airway smooth muscle cells (17). Similar to NPMC, these cells produce eotaxin upon stimulation with IL-4, and the production of eotaxin is inhibited by IFN- (13, 18). These cells can produce RANTES by stimulation with TNF- plus IFN-, and the production is inhibited by IL-4 (17, 19, 20). Such an inhibitory effect of IL-4 was not observed in NPMC, and IL-4 to some extent augmented RANTES production in NPMC.

The results of the present study suggest that the origin of pleural eotaxin is NPMC. In our previous study, we observed that eotaxin is detectable in almost all effusions, and the pleural levels of eotaxin are correlated with number of eosinophils in the effusion (9). As mentioned above, Th1 or Th2 cells are key players in the regulation of the production of chemokines. Upon stimulation in the pleural cavity, infiltrating T helper cells produce IL-4 or IFN-, which induce the secretion of chemokines from resident mesothelial cells. This process may also occur in other sites that contain mesothelial cells, such as the abdominal, pericardial, and spinal cavities. It may also occur in other tissues such as in the bronchial wall, where other resident mesenchymal cells such as fibroblasts and resident epithelial cells may perform the role of mesothelial cells.

Multiple chemokines are involved in pleural inflammation, such as tuberculous effusion. In murine mesothelial cells, bacillus Calmette-Guérin can induce macrophage inflammatory protein-1 and monocyte chemoattractant protein-1 (21), which are chemotactic for mononuclear cells. The production of these chemokines is augmented by the addition of IFN-, similar to RANTES production in the present study. Because tuberculous effusions contain large amounts of TNF- and IFN-, eosinophils should, in theory, accumulate via RANTES in such effusions. However, tuberculous effusions are usually non-eosinophilic. The reason for this is not clear. Although RANTES cannot attract eosinophils in the guinea pig, human RANTES induces migration of eosinophils. In fact, by using neutralizing antibodies, we further showed that the RANTES protein produced by NPMC was biologically active. The reason for the non-eosinophilic nature of tuberculous effusions may be due to the fact that RANTES is not selective for eosinophils. Furthermore, RANTES may not be chemotactic for eosinophils in vivo. Airway-specific expression of RANTES in transgenic mice showed neutrophilic but not eosinophilic inflammation, possibly due to other chemokines induced by RANTES (22). It may also be due to decreased expression of particular adhesion molecules that are necessary for the migration of eosinophils and that are reported to be different from the vascular cell adhesion molecule-1 very late antigen-4 system (23). Further studies are needed to clarify these points.

The supernatants of cultures stimulated with TNF- (10 ng/ml) plus IL-4 (10 ng/ml) for 48 h usually contained < 1 ng/ml of eotaxin. The eosinophil chemoattractant activity of the supernatant was equal to > 5 ng/ml recombinant eotaxin, which is much more than actual content of eotaxin protein in the supernatant. Therefore, the chemotactic activity of eotaxin is augmented in some way, or other chemokines in the culture supernatant increase the activity. IL-5 promotes the eotaxin responsiveness of eosinophils (24, 25). However, we could not detect IL-5 in the supernatants. Mesothelial cells are reported to produce another eosinophil-activating cytokine, GM-CSF (26). We also detected GM-CSF in supernatants from NPMC stimulated with TNF- (unpublished observation). On the other hand, such supernatants contained more RANTES protein. However, the chemotactic activity was almost exclusively due to eotaxin alone. Although eotaxin is known to be more potent than RANTES in terms of eosinophil chemotactic activity, it does not seem to be due to this because there is not such a big difference between these chemokines in our experimental system (data not shown). Furthermore, the level of RANTES produced in this condition is enough to induce migration of eosinophils. Therefore, there may be inhibitory factor(s) for the activity of RANTES and other chemotactic or chemokinetic factors that work with eotaxin but not RANTES, although these notions have not been examined in our experimental system.

In conclusion, this study demonstrated that NPMC can produce eotaxin, and NPMC can produce different types of chemokines by different mechanisms, depending on Th1 and Th2 responses. IL-4, a representative Th2 cytokine, favors eotaxin production. On the other hand, IFN-, a representative Th1 cytokine, increases RANTES at least in part by increasing the stability of RANTES mRNA, and IFN- also acts to decrease eotaxin production without influencing mRNA stability.