INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The development of inflammatory processes in the pleural space may result in increased pleural vascular permeability leading to the accumulation of fluid enriched in proteins, and to the recruitment of cells into the pleural space (i.e., the development of an exudative pleural effusion) (1). Although pleural effusions are common, very little information is available on the inflammatory and immune mechanisms that are involved in their development. In particular, it is unclear which cells and mediators are involved in the inflammatory processes, and whether resident immunocompetent cells may orchestrate the development of an inflammatory response.

Inflammation within the pleural space could be mediated by a number of proinflammatory molecules. One of these, interleukin-8 (IL-8), has been shown to promote neutrophil chemotaxis into the pleural space (2, 3). However, the murine equivalent for IL-8, neutrophil-activating protein-1 (NAP-1), is also a potent chemotactic agent for lymphocytes (4). Thus, there is the potential for IL-8 to contribute not only to neutrophil recruitment but also to lymphocyte recruitment in human disease.

IL-8 is produced by a wide variety of cells including endothelial cells (5), epithelial cells (6), mesangial cells (7), microglial cells (8), amnion (9), neutrophils (10), T cells (11), and mononuclear phagocytes (12). However, only a minority of these cell types are present in the quiescent pleural compartment. It is not known which cells within the pleura, other than mesothelial cells (13), can produce IL-8. Because cell recruitment in response to IL-8 follows a chemotactic gradient (3), the cellular source of IL-8 plays a critical role in the directed migration of inflammatory cells. It is not known whether in some diseases a chemotactic gradient is established into the pleural space, or whether the cells are sloughed into the pleural cavity as a consequence of a vigorous inflammatory process within the adjacent visceral or parietal pleura.

Chemokines are potent proinflammatory molecules, and their regulation must therefore be tightly controlled. Pleural macrophages (PleM) play a crucial role in local immunologic responses. In particular, it has been shown that these cells exert accessory activity for T-lymphocytes and actively produce tumor necrosis factor- (TNF-) and interleukin-1 (IL-1) (14). However, it has not been shown that PleM regulate the influx of inflammatory cells from the blood circulation into the inflamed pleural space.

The present study was performed: (1) to identify the source of IL-8 within the pleural space; (2) to evaluate the role of IL-8 in regulating lymphocyte chemotaxis into the pleural space; and (3) to investigate the role of PleM in modulating the production of IL-8 by other pleural cells.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pleural Fluid Collection

Pleural fluid was collected by thoracentesis from hospitalized patients with congestive heart failure (CHF) (n = 10; age range: 48 to 69 yr), tuberculosis (TB) (n = 9; age range: 26 to 70 yr), and cancer (n = 12; age range: 45 to 72 yr) following the granting of informed consent. The effusions were first classified as transudates or exudates according to at least one of the criteria described by Light and colleagues (15). The diagnosis for each patient was established by standard clinical, laboratory, and radiologic investigations. In detail, CHF effusions were defined as transudates associated with an enlarged heart, distended neck veins, and a cardiac gallop that improved with therapy for CHF; tuberculous effusions were defined as exudates that gave positive cultures for Mycobacterium tuberculosis or a positive smear for acid-fast bacilli typically found in tuberculous pleural fluids; malignant effusions were defined as exudates associated with a pathologic diagnosis of cancer upon cytologic examination of pleural fluids or pathologic examination of lung tissues. None of the patients was undergoing antiinflammatory or steroid therapy. The fluids were drawn into polypropylene tubes containing heparin (10 to 20 IU/ml). The total white cell and differential counts were obtained, and the fluids were subsequently centrifuged at 400 × g for 10 min. Cell-free fluids were immediately frozen at 70° C until they were analyzed for total protein and IL-8 concentrations.

Measurement of IL-8

The concentration of IL-8 was determined with an enzyme-linked immunosorbent assay (ELISA) (Quantikine; R&D Systems, Minneapolis, MN). All commercial reagents were used according to the manufacturers' specifications. The concentrations of IL-8 in pleural effusions were determined by reference to an IL-8 standard curve constructed with concentrations of 0 to 2,000 pg/ml of recombinant human IL-8. The lower limit of detection was 18 pg/ml.

Identification of Pleural Cells Able to Express Messenger RNA for IL-8

Cytocentrifuge preparations of pleural effusion cells were made, fixed in cold 4% paraformaldehyde, and stained according to the immunoalkaline phosphatase-antialkaline phosphatase (PAP) method as previously described (16). The following monoclonal antibodies (mAbs) were used: anti-CD68, specific for mononuclear phagocytes (Dakopatts, Glostrup, Denmark); anti-CD3, specific for T lymphocytes (Becton Dickinson, Mountain View, CA); anticytokeratin-6 and anticytokeratin-18, specific for epithelial and mesothelial cells (Dakopatts); antihuman epithelial membrane antigen (EMA) (Dakopatts); and anti-smooth-muscle, which served as a control mAb (Ortho Pharmaceuticals, Raritan, NJ).

To identify cells expressing IL-8, the same slides that were stained with mAbs were used for primed in situ labeling (PRINS), utilizing a probe specific for IL-8 according to a method described by Mogensen and colleagues, with minor modifications (17). Briefly, the same slides that had been used for immunochemical staining were immediately immersed in 4% paraformaldehyde/phosphate-buffered saline (PBS) for 5 min, treated with Triton X-100 (0.3% in PBS) for 10 min, and then hydrated with Tris-glycine buffer (0.2 M Tris-HCl, 0.1 M glycine, pH 7.4) for 10 min. Slides were then treated for 10 min at room temperature with 2× standard saline citrate (SSC) (1× SSC = 150 mM NaCl, 15 mM sodium citrate, pH 7.0) mixed with 50% of deionized formamide. The temperature was subsequently increased to 55° C and slides were kept at this temperature for an additional 15 min. Slides were then transferred to 0° C precooled nick-translation buffer (NTB) (50 mM Tris-HCl, 10 mM MgSO₄, 100 mM dithiothreitol, 150 µg bovine serum albumin [BSA]) and incubated for 30 min at 4° C. To avoid self-priming, slides were then preincubated (90 min, 42° C) with 5 nmol each of deoxyadenosine triphosphate (dATP), deoxycytosine triphosphate (dCTP), deoxyguanosine triphosphate (dGTP), and deoxythymidine triphosphate (dTTP), 0.25 U ribonuclease (RNase) inhibitor, and 0.5 U avian myeloblastosis virus (AMV) reverse transcriptase in a total volume of 10 µl of NTB. After preincubation, slides were washed in NTB and then incubated (90 min, 42° C) for sequence-specific reactions with 2 µg of oligo-IL-8 primer, 5 nmol each of dATP, dCTP, and dGTP; 0.5 nmol dTTP; 0.078 nmol digoxigenin-deoxyuridine triphosphate (dUTP), 0.25 U RNase inhibitor, and 0.5 U AMV reverse transcriptase in a total volume of 10 µl of NTB. After incubation, slides were washed twice in 0.5× SSC and twice in 0.1× SSC, respectively (10 min each at 42° C). Slides were then washed in tris-buffered saline (TBS) (150 mM NaCl, 50 mM Tris-HCl, pH 7.4) and incubated in 3% BSA/TBS (10 min) and in 1:500 to 1:300 dilutions of sheep polyclonal antidigoxigenin antiserum conjugated with alkaline phosphatase (<Dig>-AP) (Boehringer Mannheim, Mannheim, Germany) in TBS (1 h). Slides were again washed three times with TBS (15 min each wash) and incubated with equalization buffer (0.1 M HCl, 0.1 M NaCl, 50 mM MgCl₂, pH 9.5) (5 min). Color development of the slides was achieved by adding a freshly prepared substrate solution consisting of 0.175 mg X-phosphate-5-bromo-4-chloro-3-indolyl phosphate (BCIP) and 0.37 mg nitroblue tetrazolium (NBT) salt per milliliter of equalization buffer (for 20 to 40 min at room temperature). Slides were then transferred to TBS and washed in tap water, counterstained with hematoxylin for 5 s, and mounted with glycerol-gelatin. The following control groups were included in each experiment: (1) pretreatment with RNase A before PRINS; (2) omission of the specific primer; and (3) omission of the <Dig>-AP conjugate. Slides were evaluated under a light microscope (×40 magnification and ×100 oil immersion magnification). At least four fields per slide were evaluated. Cell counts were expressed as the percentage of IL-8 messenger RNA (mRNA)-positive cells. The human IL-8 oligonucleotide probe was complementary to nucleotides 378 through 398, and had the sequence 5'-TGA-ATT-CTC-AGC-CCT-CTT-CAA-3' (18).

Isolation of Macrophages from Pleural Effusions

Macrophages were harvested from pleural fluids of patients with CHF according to a previously described method (14). Briefly, after filtration through two layers of sterile surgical gauze, the cells were washed three times in Hanks' balanced salt solution (HBSS) (300 × g, 10 min). One million cells were resuspended in each milliliter of complete medium (CM) (RPMI 1640 + 10% fetal calf serum + 25 mM 4-(2-hydroxyethyl)-1-piperazine-N'-2-ethanesulfonic acid [Hepes] + 2 mM glutamine + 100 ng/ml penicillin + 100 ng/ml streptomycin) (all from GIBCO, Paisley, UK). Cells were centrifuged on a Ficoll-Hypaque gradient in 50-ml plastic tubes, and mononuclear cells were recovered and allowed to adhere to plastic dishes (37° C, 5% CO₂). Cells that remained adherent to dishes were > 90% macrophages as confirmed by nonspecific esterase staining and a specific mAb (CD68).

PleM were resuspended in CM at 1 × 10⁶ cells/ml and incubated at 37° C under 5% CO₂ in the absence or in the presence of 1 mg/ml lipopolysaccharide (LPS) from Escherichia coli (serotype 0111:B4; Sigma Chemical Co., St. Louis, MO) for 4 h (14). After incubation, tubes were centrifuged (600 × g for 10 min) and culture supernatants were collected and stored at 70° C until they were used in subsequent experiments.

Modulation of IL-8 mRNA and IL-8 Production

To evaluate the ability of PleM to modulate IL-8 expression, total pleural cells were cultured in CM for 24 h (37° C, 5% CO₂) in the presence or in the absence of PleM supernatants, using polypropylene round-bottom tubes (Becton Dickinson, Lincoln Park, NJ). Recombinant TNF- (1,000 IU/ml) or sufficient polyclonal anti-TNF- antibody to neutralize 1,000 IU of TNF- as indicated by the manufacturer (polyclonal rabbit anti-human TNF-; Genzyme Diagnostics, Cambridge, MA) were used as controls. After incubation, tubes were centrifuged (600 × g, 10 min) and cells were recovered. Cytocentrifuge preparations were made with cells from different experimental conditions, and were used for IL-8 mRNA and protein detection.

The expression of mRNA for IL-8 was detected by PRINS, as described earlier but with minor modifications. In particular, an antidigoxigenin antiserum conjugated with peroxidase (<Dig>-POD) (Boehringer Mannheim), followed by treatment with a Tyramide Signal Amplification kit (NEN Life Science Products, Boston, MA), was used to amplify the chromogenic reaction, rather than an antidigoxigenin antiserum conjugated with alkaline phosphatase (<Dig>-AP).

The expression of cell-associated IL-8 antigen was detected by immunocytochemical staining with a mouse antihuman IL-8 mAb (Genzyme), using the horseradish peroxidase method as previously described (6). A Tyramide Signal Amplification kit (NEN Life Science Products) was used according to the manufacturer's specifications to amplify the chromogenic reaction.

Lymphocyte Chemotaxis Assay

Lymphocyte-enriched fractions were obtained from the peripheral blood of healthy adults by Ficoll-Hypaque centrifugation followed by fractionation on a one-step Percoll (Pharmacia, Uppsala, Sweden) gradient. The lymphocyte-enriched fraction was passed through a nylon wool column to further purify the T-cell population (1 h, 37° C, 5% CO₂) as previously described (6). The recovered T-cell fraction was > 95% CD3-positive as assessed with cytofluorimetric analysis.

Chemotaxis was effected as previously described (19). Briefly, cell migration was assessed with a modification of the Boyden chamber assay, using a microchemotaxis chamber (Costar; Neuro Probe Inc., Cabin John, MD). T cells (10 × 10⁶/ml in RPMI 1640) were loaded into the upper well of the chamber and 30 µl of pleural fluid were placed in the bottom chamber. The two wells were separated by a polycarbonate filter paper with a pore size of 8 µm, and the chamber was incubated at 37° C for 3 h. At the end of incubation, the filter was fixed, stained, and mounted on a glass microscope slide. Migration was assessed by counting the number of cells that had migrated beyond a certain depth into the filter (50 µm). Ten to 15 cells/high power field (hpf) were counted in the negative control wells (HBSS alone), and represented the random migration (chemokinesis) of lymphocytes. Chemotactic bioactivity was expressed as percentage of random cell migration (negative control) normalized to 0% in all experiments, as previously described (20). To demonstrate that IL-8 was responsible for T-lymphocyte migration, blocking experiments were performed by mixing the pleural fluid with an anti-IL-8 mAb (Genzyme) (1 µg/ml, a concentration sufficient to neutralize the bioactivity of 50% of chemotactic activity induced by 2 nM recombinant IL-8) for 30 min at 37° C prior to loading the chemotaxis chamber.

Statistics

Data are expressed as mean ± SD. Spearman's test was used for correlations. The Mann-Whitney U test was used for comparisons of patient groups, and comparisons of differential experimental conditions were evaluated through analysis of variance (ANOVA) corrected with Bonferroni's test. A value of p < 0.05 was accepted as statistically significant for Spearman's test and the Mann-Whitney U test, and a value of p < 0.005 was accepted as statistically significant for ANOVA and Bonferroni's test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IL-8 Is Detected in Pleural Fluid

To determine whether pleural effusions from patients with CHF, TB, and cancer contained the chemotactic cytokine IL-8, the concentration of IL-8 in pleural fluid was determined by ELISA. IL-8 was detected in both TB and malignant effusions, whereas almost no IL-8 was detected in effusions caused by CHF (Figure 1). The concentrations of IL-8 were low in transudative effusions (in CHF the ratio of total protein in effusion/blood was 0.4 ± 0.2, and the ratio of lactate dehydrogenase [LDH] in effusion/blood was 0.45 ± 0.1), and were significantly higher in exudative effusions (in TB the ratio of total protein in effusion/blood was 0.78 ± 0.16, and the ratio of LDH in effusion/blood was 2.5 ± 1.0, whereas in cancer the ratio of total protein in effusion/blood was 0.67 ± 0.14 and the ratio of LDH in effusion/blood was 1.2 ± 0.6). Although the concentrations of IL-8 tended to be higher in cancer patients' effusions than in TB patients' effusions, no significant differences between the two study groups were detected.

View larger version (12K): [in this window] [in a new window]   Figure 1.   IL-8 is present in the pleural effusions of patients with TB and cancer. Pleural fluid was obtained from CHF patients (stippled bar, n = 8), TB patients (open bar, n = 9), and cancer patients (closed bar, n = 9). IL-8 concentrations were measured with an ELISA and are expressed as pg/ml (mean ± SD). TB versus CHF, p < 0.03; cancer versus CHF, p < 0.02; TB versus cancer, p = NS.

Correlation of IL-8 Concentrations with the Number of Cells in Pleural Fluid

Prior data suggested that IL-8 was an important contributor to the inflammatory response. To determine the possible role of IL-8 in cell recruitment, the concentration of IL-8 was correlated with the number of neutrophils and lymphocytes in pleural effusions. Surprisingly, IL-8 concentrations did not correlate with the number of neutrophils when the data for all the three study groups were pooled (Figure 2A; Rho = 0.2, p = 0.356), but did correlate with the number of lymphocytes (Figure 2B; Rho = 0.7, p = 0.012).

View larger version (15K): [in this window] [in a new window]   Figure 2.   The concentration of IL-8 did not correlate with the number neutrophils (A), but did correlate with the number of lymphocytes in pleural effusions (B). The total cell count and differential were determined for patients with cancer (n = 7), CHF (n = 4), and TB (n = 3). IL-8 concentration was determined with an ELISA. Each point represents an individual patient.

Pleural Effusions Are Chemotactic for Lymphocytes

The finding that the concentration of IL-8 in pleural effusions correlated best with the number of lymphocytes in such effusions, together with the fact that IL-8 is chemotactic for T lymphocytes (4), prompted us to test the chemoattractant activity of pleural fluids on human peripheral T cells. We compared the chemoattractant activity of pleural fluids from patients with cancer, TB, and CHF. Exudative pleural effusions from both cancer and TB patients exerted a more potent chemoattractant activity for T lymphocytes than did transudative effusions from CHF patients (Figure 3). Moreover, the chemoattractant activity of effusions from cancer patients was significantly greater than the activity exerted by effusions from TB patients.

View larger version (19K): [in this window] [in a new window]   Figure 3.   Pleural fluid is chemotactic for T lymphocytes. Pleural fluids from patients with CHF (n = 4), TB (n = 4), and cancer (n = 5) were used to stimulate chemotaxis of peripheral blood T lymphocytes isolated from healthy adults. The experiments were performed in the absence and in the presence of an anti-IL-8 mAb. Data are expressed as percent of control (mean ± SD) (see METHODS for details). Open bars represent chemotaxis in the absence of anti-IL-8 mAbs; closed bars represent chemotaxis in the presence of anti-IL-8 mAbs. An irrelevant control antibody (isotype-matched mouse IgG1, crosshatched bars) was included for one patient in each group. (*p < 0.05 compared with the corresponding group without anti-IL-8 mAb; TB versus CHF, p < 0.05; cancer versus CHF, p < 0.05; TB versus cancer, p = NS.)

To determine whether IL-8 was responsible for the recruitment of T lymphocytes, the ability of an anti-IL-8 mAb to neutralize the chemoattraction of T lymphocytes was tested. The anti-IL-8 antibody completely suppressed T-lymphocyte chemotaxis in TB effusions, but only partly abrogated this activity in cancer patients (Figure 3). This suggests that IL-8 is responsible for T-lymphocyte recruitment in TB, but that in cancer effusions other soluble factors in addition to IL-8 may contribute to the influx of T lymphocytes.

Identification of Cells Able to Produce IL-8

Having established that IL-8 was present in pleural fluid, and that it was capable of recruiting T lymphocytes, we performed experiments to determine whether the cells within the pleural space expressed IL-8 or whether IL-8 was undetectable and therefore probably derived from the cells lining the pleural space. To evaluate whether cells recovered from pleural effusions can express IL-8, we performed PRINS for IL-8 mRNA on the total cell populations recovered from the three study groups. In accord with the observations described earlier, we detected no mRNA for IL-8 in cells recovered from transudative effusions. Cells from malignant effusions expressed mRNA for IL-8 (12.7 ± 7.8% of cells) (Figure 4A); by contrast, mRNA for IL-8 was undetectable in cells from TB effusions (Figure 4C).

View larger version (47K): [in this window] [in a new window]   Figure 4.   Cells from malignant effusions express IL-8 mRNA, whereas cells from tuberculous and CHF pleural effusions do not. Cells were obtained from malignant (A and B), tuberculous (C ), and CHF (D) effusions. Cells were labeled through PRINS for IL-8 (A, C, and D), or with an irrelevant probe (B). Cells that were labeled through PRINS for IL-8 are indicated with arrows. Original magnification: ×40 (A and B); ×16 (C and D).

To identify the phenotype of the cells producing mRNA for IL-8, we performed immunocytochemical staining and PRINS on cells recovered from malignant effusions. Interestingly, cancer cells, which were labeled with an antibody to the epithelial membrane antigen, in addition to mesothelial cells, which label with anti-cytokeratin antibody, actively expressed mRNA for IL-8 (Figures 5A and 5C).

View larger version (47K): [in this window] [in a new window]   Figure 5.   Malignant effusions contain both cancer cells and mesothelial cells that express IL-8 mRNA. (A) Cells were labeled with anti-cytokeratin-6 and anti-cytokeratin-18 antibodies and were labeled in situ for IL-8. The arrow indicates a cell that is a doubly positive mesothelial cell. (B) Labeling with anti-cytokeratin-6 and anti-cytokeratin-18 antibodies alone. The arrow indicates a cytokeratin-positive cell. (C ) Cells labeled with anti-EMA antibody and labeled in situ for IL-8. The arrow indicates a malignant cell that is doubly positive. (D) Cells were labeled with anti-EMA antibody alone. The arrow indicates a malignant cell that is positive for EMA. Original magnification: ×25 (A and B); ×100 oil immersion (C and D).

Role of PleM in the Modulation of IL-8 Production

The finding that cancer cells as well as mesothelial cells actively expressed IL-8 mRNA prompted us to evaluate whether PleM can modulate IL-8. To determine whether factors released from PleM could modulate IL-8, we assessed IL-8 mRNA and cell-associated IL-8 protein expression. PleM were stimulated with LPS because it activates these cells (14), and although it is not known whether PleM produce cytokines that stimulate pleural cells to produce IL-8, it has been shown that alveolar macrophages stimulate epithelial cells to produce IL-8 (6). Supernatants of LPS-stimulated PleM cultures induced both IL-8 mRNA and cell-associated IL-8 protein expression by cells recovered from pleural effusions (Figures 6 and 7). There was a significant increase in the percentage of pleural cells expressing IL-8 after treatment with stimulated macrophage supernatants. Anti-TNF- mAb significantly reduced the percentage of cells expressing both IL-8 mRNA and cell-associated IL-8 protein, suggesting that production of TNF- was responsible for the expression of IL-8. Additionally, recombinant TNF- alone was able to induce both IL-8 mRNA and cell-associated IL-8 protein expression by pleural cells. We found no significant induction of IL-8 by LPS alone at the concentration tested in these experiments (Figure 6). Thus, pleural cells produce IL-8 in response to the cytokines produced by pleural macrophages.

View larger version (15K): [in this window] [in a new window]   Figure 6.   Supernatants from stimulated macrophages induce IL-8 expression in pleural effusion cells. Cells were isolated from pleural effusions and cultured as described in METHODS. IL-8 mRNA expression was evaluated through PRINS (A) and cell-associated IL-8 protein was evaluated with immunocytochemistry (B). Data are expressed as percent positive cell (mean ± SD of four experiments). (Macrophage culture supernatants versus macrophage culture supernatants + anti-TNF- mAb, p < 0.005; macrophage culture supernatants versus LPS alone, p < 0.005; macrophage culture supernatants versus TNF- alone, p = NS; TNF- versus LPS, p < 0.005.)

View larger version (48K): [in this window] [in a new window]   Figure 7.   Activated PleM stimulate pleural effusion cells to express IL-8 mRNA and protein. Cells recovered from pleural effusions were labeled prior to (A and C ) or after stimulation with LPS-stimulated PleM culture supernatants (B and D). The cells were labeled to detect IL-8 mRNA (A and B) and cell-associated IL-8 protein (C and D). Original magnification: ×16.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During the first steps in the development of inflammatory processes, the migration and activation of leukocytes within the pleural space may be effects of soluble mediators (eicosanoids, vasoactive amines, and cytokines) released by "resident" cells that are activated by various stimuli (21). IL-8 is a chemotactic cytokine produced by immunocompetent (10) and structural (5, 6, 22) cells that promotes the development of phlogistic events by attracting neutrophils and lymphocytes (4, 23). The present study was done to determine the contribution of IL-8 to inflammatory processes in the pleural space. We showed that IL-8 concentrations are greater in pleural effusions from patients with cancer and TB than in effusions from patients with CHF. Our findings confirm previous observations (24) that IL-8 concentrations are increased in exudative pleural effusions caused by inflammatory processes and not in transudative pleural effusions caused by excess hydrostatic pressure.

IL-8 is somewhat resistant to proteolytic enzymatic cleavage than are other chemotactic factors (23), suggesting that this molecule may be involved in more prolonged inflammatory cell influx and may have additional biologic functions other than just neutrophil recruitment, which is often a rapid event. It has been shown that pleural fluid can induce neutrophil chemotaxis in vitro, and that this effect is due to IL-8 (24). However, our observations demonstrate that the concentration of IL-8 did not correlate with the number of neutrophils in pleural effusions caused by persistent diseases such as TB and cancer. Because IL-8 has been shown to promote lymphocyte influx into the alveolar compartment in asthma (20), we were prompted to evaluate whether pleural IL-8 might be responsible for the influx of T lymphocytes into the pleural space. The findings that the level of IL-8 in pleural effusions from both TB and cancer patients tended to correlate with the burden of lymphocytes, and that an anti-IL-8 mAb inhibited the ability of the pleural fluid to stimulate peripheral lymphocyte chemotaxis in vitro, strongly suggest that IL-8 plays a role in the influx of lymphocytes into the inflamed pleural space. We speculate that high concentrations of IL-8 as well as other specific chemoattractants promote the accumulation of neutrophils during the first steps of pleural inflammation, and that subsequently, lower concentrations of IL-8 contribute to T-lymphocyte recruitment. In this regard, it has been demonstrated that T-lymphocytes are from two to 10 times more sensitive to IL-8 than are neutrophils, and that T-cell migration is due to a concentration gradient of this cytokine (chemotaxis) (4). However, T-lymphocyte chemoattractants other than IL-8 might be present in pleural effusions. Our data showing that the addition of anti-IL-8 mAb completely reversed the T-cell chemoattractant activity of TB effusions but only partially inhibited this activity of cancer effusions strongly support this hypothesis, and suggest that at least in cancer, other soluble factors in addition to IL-8 stimulate T-cell recruitment.

It has been shown that the accumulation of IL-8 in the pleural space is due to local production of this cytokine and not to increased vascular permeability, as reflected by the IL-8/ protein ratio (24). In the present study, the combined application of immunocytochemistry and PRINS to pleural cells freshly isolated from pleural effusions of cancer patients allowed us to demonstrate that mesothelial and cancer cells can express the mRNA for IL-8, suggesting that these two cell types might contribute to IL-8 production within the pleural space during malignancy.

To distinguish mesothelial cells from cancer cells, we used mAbs directed against epithelial membrane antigen and cytokeratin, because previous studies had shown that epithelial membrane antigen is the most useful marker for distinguishing cancer cells from reactive mesothelial cells (25), whereas cytokeratin identifies mesothelial cells and is responsible for staining of these cells with high intensity (26). We found that mesothelial cells isolated from exudative pleural effusions (cancer), but not mesothelial cells isolated from transudative pleural effusions, express the mRNA for IL-8. That mesothelial cells from transudative effusions express IL-8 only after activation (13, 27) suggests that the IL-8 that we observed was due to activation in vivo rather than to nonspecific activation in vitro.

Although the fluid from pleural effusions of TB patients had high concentrations of soluble IL-8, these effusions did not contain cells expressing mRNA for IL-8. This might have been due to the lack of mesothelial cells in these effusions. Mesothelial cells are usually absent in tuberculous effusions; indeed, the absence of mesothelial cells, together with the presence of increased numbers of lymphocytes in pleural effusions, is currently considered a diagnostic criterion for TB (28, 29). The high concentrations of soluble IL-8 in these effusions might be due to production of this cytokine by mesothelial cells that remain in the pleural lining and are not shed into the effusion. Although there was no significant difference between IL-8 concentrations in TB and malignant pleural effusions in our study, there was a trend toward higher IL-8 levels in malignancy (Figure 1). This phenomenon might be at least partly related to the presence of IL-8-producing cancer cells, which may represent an additional source of IL-8 not present in tuberculous effusions.

Our observation that cancer cells and mesothelial cells in exudative pleural effusions express IL-8 mRNA raised the possibility that a cytokine network for activating these cells is present in the inflamed pleural space. The findings that PleM actively stimulate IL-8 production by pleural effusion cells, and that this phenomenon is mediated by the release of TNF-, strongly suggest that PleM play a crucial role in regulating leukocyte recruitment. Conceptually, this is similar to a previous report that alveolar macrophages release the monokines IL-1 and TNF-, which in turn induce gene expression and production of IL-8 by pulmonary epithelial cells, resulting in the recruitment of leukocytes (6). Because macrophages play a crucial role in regulating inflammatory processes in different organs, spaces, and cavities (30) it is conceivable that PleM regulate the recruitment of leukocytes, at least in part, by modulating IL-8 production.

In summary, our findings suggest that soluble IL-8 in the pleural space is an important phlogistic agent, and that IL-8 is produced by pleural structural and cancer cells after activation by monokine stimulation. PleM, by responding to a primary signal that leads to the elaboration of chemotactic factors and subsequent lymphocyte influx, would appear to play a crucial role in the directed migration of T lymphocytes to the pleural space.