Introduction

Top Abstract Introduction Material and Methods Results Discussion References

Macrophages are mononuclear phagocytes that develop in various tissues by maturation and differentiation from immigrating blood monocytes. Their major functions within the immune system are (1) to regulate local inflammatory reactions by releasing pro- and anti-inflammatory molecules, (2) to support a primary defense mechanism via phagocytosis and respiratory burst, and (3) to mediate immune responses by antigen processing and presentation. Macrophages exhibit a marked heterogeneity that appears to be tissue dependent. So far, several macrophage subpopulations have been described, such as peritoneal macrophages, Kupffer cells in the liver, pulmonary (alveolar macrophages, airway macrophages, interstitial macrophages, and intravascular macrophages) and thoracic macrophages (pleural macrophages [PLM] and lymph node macrophages) (1). While peritoneal macrophages, Kupffer cells, and alveolar macrophages are well characterized, little is known regarding the phenotype and function of PLM (2).

In humans, the pleural space is a 20-µm-wide gap between the visceral pleura and the parietal pleura that contains a small volume of fluid. Various cell types occur in the pleural fluid, including macrophages that account for about 50% of free cells (1). It is assumed that PLM play a role of defense in this compartment and that they may be involved in the development of several lung diseases (4).

Gjomarkaj and coworkers (3) demonstrated that PLM from rats are functionally and phenotypically different from alveolar macrophages but similar to peritoneal macrophages in regard to histochemical properties and expression of membrane antigens. They conclude that these cells might play an important role in cell-mediated immune reactions in the human pleural space. Gjomarkaj and associates (2) characterized PLM in malignant effusions from patients with primary bronchogenic carcinoma by analyzing surface proteins (CD11b and CD14) and cytokine expression (interleukin [IL]-1 and tumor necrosis factor [TNF]).

Some limited evidence has suggested that the PLM originate from peripheral blood monocytes that migrate across the mesothelial lining of the pleura (1). The aim of the current report was to compare the resting PLM and the two monocyte cell populations in blood by phenotypic and functional analysis. In this study, we demonstrate that the PLM form a unique type of tissue macrophage.

	Material and Methods

Top Abstract Introduction Material and Methods Results Discussion References

Patients

Sixteen patients with a central bronchial tumor not involving the pleura who underwent lobe resection were selected for a pleural lavage before surgery. In addition, we studied two patients with pulmonary chondroma, one with intrapulmonary hemorrhage, and one with pneumothorax. The 20 patients (16 male, four female) had an average age of 60.6 ± 11.4 yr. Both during surgery and in the analysis of lavage cells there was no evidence of involvement of the pleural space in any one of the patients. After opening the thorax by incision, the pleural space was instilled with 500 ml prewarmed, phosphate-buffered saline (PBS) solution, strictly avoiding contamination by blood. Thoracic surgery continued after this brief lavage procedure. Written informed consent was obtained for participation in this study, which had been approved by the Ethics Committee of the Medical Faculty, University of Munich.

Culture Medium

RPMI 1640 (Biochrom, Berlin, Germany) was supplemented with 2 mM L-glutamine (GIBCO, Grand Island, NY), 200 U/ml penicillin, 200 µg/ml streptomycin (GIBCO), nonessential amino acids 1-2× (GIBCO), 10 ml for 1 liter OPI supplement (contains oxalacetic acid, sodium pyruvate, and insulin) (Sigma, St. Louis, MO). After filtration through a Gambro ultrafilter U 2000 (Martinsried, Germany) to remove lipopolysaccharide (LPS), we added 10% fetal calf serum (FCS) that had been pretested for low levels of LPS.

Cells

Lavage cells were centrifuged for 10 min at 400 × g and resuspended in RPMI 1640 + 10% FCS. Total cells were counted in a Neubauer chamber. On average, of 20 patients we obtained 25.2 × 10⁶ ± 45.7 × 10⁶ pleural lavage cells out of the entire pleural lavage. Peripheral blood mononuclear cells (PBMC) were isolated from peripheral blood using density gradient centrifugation with Lymphoprep (Nycomed Pharma, Norway) according to the manufacturer's instructions.

Flow Cytometry

Isolated and washed PLM and whole blood samples were stained with specific monoclonal antibodies or the isotype control from the same manufacturer. The following antibodies (all fluorescein isothiocyanate labeled) were used: CD14, My4 (isotype: mouse immunoglobulin [Ig]G2b; Immunotech); CD16, Leu11c (isotype: mouse IgGl; Becton Dickinson, Franklin Lakes, NJ); CD32, FLI8.26 (isotype: mouse IgG2b; Pharmingen, San Diego, CA); CD64 (isotype: mouse IgG1; Immunotech); CD33, My9 (isotype: mouse IgG2b; Coulter); CD54, intercellular adhesion molecule-1 (isotype: mouse IgG1; Immunotech); CD11b, bear1 (isotype: mouse IgG1; Immunotech); human leukocyte-associated antigen-DR (HLA-DR), I2 (isotype: mouse IgG2a; Coulter). Monocytes were stained in whole blood using the CD14 antibody My4-PE (isotype: mouse IgG2b; Coulter) to distinguish between the subsets. Flow cytometry analysis in an EPICS XL (Coulter) flow cytometer was done directly for PLM and after lysis of the erythrocytes when using whole blood samples.

Phagocytosis

Sheep red blood cells (SRBC) (Dade Behring, Liederbach, Germany) were stained with a fluorescent dye (1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanin-perchlorate [DiI]; Molecular Probes, Leiden, The Netherlands) and coated with rabbit antisheep IgG antibodies (Ambozeptor; Dade Behring). In brief, 5 ml sheep blood was washed three times in LPS-free PBS and then separated over polymorphprep density gradient (Nycomed Pharma) to obtain leukocyte-free erythrocytes. For staining, 2 × 10⁹ SRBC were incubated 2 h at 37°C in 1 ml PBS with DiI (400 µg/ml final). SRBC were then washed in LPS-free PBS until the supernatant remained colorless. Half of the stained SRBC were additionally opsonized with rabbit antisheep antibodies. For this, 1 × 10⁹ SRBC in 500 µl PBS were incubated 1 h at 37°C with 1 µl ambozeptor (1:500 dilution final) and washed twice in 1 ml PBS. Both SRBC preparations showed comparable fluorescence intensity when analyzed by flow cytometry.

For phagocytosis, 4 × 10⁶ SRBC with or without antibody coating were incubated 2 h in 1 ml culture medium + 10% FCS with either 1 × 10⁶ PLM or PBMC. Cells were then washed with 10 ml PBS/2% FCS 10 min at 400 × g. Cell pellets were resuspended in 100 µl PBS/2% FCS and nonphagocytosed SRBC were lysed in a Coulter Q-Prep Immunology work station according to the manufacturer's instructions. Washed cells were analyzed immediately in a Coulter EPICS XL flow cytometer.

Reverse Transcriptase/Polymerase Chain Reaction

Quantitative polymerase chain reaction (PCR) for TNF messenger RNA (mRNA) was performed with an internal cRNA standard according to the method of Wang and colleagues (5). After cell lysis, 5 × 10⁵ copies of the cRNA standard pAW 108 (Perkin Elmer Cetus, Applied Biosystems, Weiterstadt, Germany) and 15 µg transfer RNA as carrier were added per sample. After isolation, the RNA was reverse transcribed with oligo(dT) as primer. PCR amplification was performed with the gene Amp RNA-PCR kit (Perkin Elmer Cetus). Thirty-six to 38 cycles for TNF and -enolase or 45 cycles for IL-10 (here using the TaqGold enzyme; Perkin Elmer) were performed with a Hybaid Touch Down Thermal Cycler (MWG, Ebersberg, Germany) with sets of 30 s at 94°C, 30 s at 60°C, and 40 s at 72°C for TNF and IL-10, and with sets of 30 s at 94°C, 40 s at 59°C, and 40 s at 72°C for -enolase. The PCR products were separated on a 1.4% (TNF) or 2% (IL-10 and -enolase) agarose gel containing ethidium bromide. Polaroid photographs with ultraviolet exposure were taken for analysis.

The following primers were used: TNF: product length, 325 bp; 5' primer, 5' CAG AGG GAA GAG TTC CCC AG 3'; 3' primer, 5' CCT TGG TCT GGT AGG AGA CG 3'; IL-10: product length, 352 bp; 5' primer, 5' ATG CCC CAA GCT GAG AAC CAA GAC CCA 3'; 3' primer, 5' TCT CAA GGG GCT GGG TCA GCT ATC CCA 3'; -enolase: product length, 619 bp; 5' primer, 5' GTT AGC AAG AAA CTG AAC GTC ACA 3'; 3' primer, 5' TGA AGG ACT TGT ACA GGT CAG 3'.

Assays for Cytokine Protein

For determination of TNF and IL-10 protein, 1 × 10⁶ PBMC or PLM, respectively, were stimulated with 1 µg LPS/ml (from Salmonella minnesota, Sigma) in 1 ml of culture medium. To obtain optimal cytokine production, supernatants were harvested after 4 h for TNF and after 16 h for IL-10 measurement. TNF and IL-10 were determined by an enzyme-linked immunosorbent assay kit (PeliKine-compact human cytokine ELISA kit; CLB, Amsterdam, The Netherlands) according to the manufacturer's instructions.

Statistics

For statistical analysis, Student's t test was employed.

	Results

Top Abstract Introduction Material and Methods Results Discussion References

Surface Molecule Expression

When analyzing the pleural lavage cells by light scatter in flow cytometry, a population of large macrophage-type cells that account for 36.1 ± 24.3% of all cells and a population of small lymphocyte-type cells that account for 51.8 ± 22.8% of all cells can be distinguished. The latter cells are predominantly CD2-positive T cells (data not shown). When gating on the large cells, we found 63.8 ± 19.2% to be CD14-positive macrophages (Figure 1). Hence, the pleural lavage cell population contained 23% of CD14-positive macrophages, which is very similar to the 15.2 ± 4.3% CD14-positive monocytes in blood. PLM expression of CD14 (44.5 ± 14% relative to the expression on CD14⁺⁺ blood monocytes set at 100%) was lower than that of regular blood monocytes but higher than that of the CD14⁺CD16⁺ blood monocytes (Figure 2A) with the latter showing a broad distribution of CD14 expression. Also, the majority of the PLM expressed CD16, the low affinity Fc receptor (Figure 1). The staining intensity for CD16 on PLM was intermediate between the two blood monocyte subpopulations in that it was higher compared with control CD14⁺⁺ cells, which are CD16 negative, and lower compared with the CD14⁺CD16⁺ cells (7.7 ± 4.5% of the level on CD14⁺CD16⁺ cells).

View larger version (24K): [in this window] [in a new window]   Figure 1.   Immunofluorescence of peripheral blood monocytes and PLM. 1 × 106 PLM and 100 µl washed whole blood, respectively, were stained for 20 min with monoclonal antibodies against CD14 and CD16, and analyzed in a flow cytometer. The histograms show the expession of CD14 (left column) and CD16 (right column) on the two blood monocyte populations CD14++ (upper panels) and CD14+CD16+ (middle panels), and on PLM (lower panels) (x-axis: fluorescence intensity; y-axis: cell count). The PLM reveal an intermediate phenotype concerning the expression of both antigens.

View larger version (30K): [in this window] [in a new window]   View larger version (24K): [in this window] [in a new window]   Figure 2.   (A-C ) Staining pattern of PLM compared with blood monocytes. PLM and whole peripheral blood, respectively, were stained with the monoclonal antibodies indicated below the bars. The two blood monocyte populations were identified via the differential expression of CD14 by costaining with My4-PE. For all surface molecules, the delta mean fluorescence intensity for CD14++ regular monocytes was set at 100% except for the expression of CD16, where the value for the CD14+CD16+ cells was set at 100%. CD14++ = black bars; CD14+CD16+ = dark gray bars; PLM = light gray bars. Values indicated on the bars represent the mean of eight experiments (*P < 0.05 compared with CD14++ monocytes; **P < 0.05 compared with the CD14+CD16+ monocytes).

The PLM surface density of CD11b (127.7 ± 38% of CD14⁺⁺ monocytes) was nearest to that of CD14⁺⁺ blood monocytes (Figure 2A). With regard to the CD32 and CD64 Fc receptors, a markedly elevated level was found for CD32 (200.0 ± 120.0%), whereas CD64 was more similar to the regular CD14⁺⁺ monocytes than the CD14⁺CD16⁺ subpopulation. The surface expression of CD33 on PLM was comparable to regular CD14⁺⁺ blood monocytes (73.6 ± 29%) and higher than that on CD14⁺CD16⁺ cells. Very interestingly, we found an extremly high expression of HLA-DR (1,906 ± 1,576%) on the PLM (Figure 2C). In addition, there was also a high density in CD54 molecules (299 ± 61%) (Figure 2B).

Phagocytosis

Because the flow cytometry analysis revealed a prominent expression of all three Fc receptors (CD16, CD32, and CD64) on the PLM, we asked whether these macrophages are able to phagocytose antibody-coated SRBC. As an example, the results of PLM (upper panel) and PBMC (lower panel) of one patient are shown in Figure 3. The left diagrams reveal an increase in forward scatter caused by the uptake of SRBC. The right histograms show the increase of fluorescence intensity gated on PLM or blood monocytes. The capacity of phagocytosis was calculated by subtracting the fluorescence intensity of the control staining from the value obtained after the incubation with fluorescence-labeled and antibody-coated SRBC. The controls with non-antibody-coated erythrocytes showed no increase in forward scatter and only very little increase in fluorescence intensity (data not shown). In five of six patients, we found the PLM to be more potent in phagocytosis than the blood monocytes; in one case we measured equal values in both cell populations. On average, the PLM could phagocytose 1.7 ± 0.4 times better than could blood monocytes (nonsignificant with P = 0.058).

View larger version (21K): [in this window] [in a new window]   Figure 3.   Phagocytosis of antibody-coated SRBC. A total of 1 × 106 isolated PLM and PBMC, respectively, were incubated without or with 4 × 106 fluorescence-labeled and antibody-coated SRBC for 2 h. Nonphagocytosed SRBC were lysed in a Coulter EPICS Q-Prep workstation and washed cells analyzed in a flow cytometer. The left column shows the scatter analysis ± SRBC (upper panels: PLM; lower panels: PBMC), the right column reveals the fluorescence intensity gated on monocytes and PLM, respectively. Shown is one of six experiments.

Cytokine Expression

Because PLM showed a high expression level for CD14, the LPS receptor, we asked whether these cells are also able to produce cytokines after stimulation with LPS. We therefore incubated 1 × 10⁶ pleural lavage cells and freshly isolated PBMC per milliliter from the same donor without and with 1 µg LPS/ml. For the measurement of TNF, we harvested the supernatants after 4 h. For IL-10 production, we harvested after 16 h because kinetics for IL-10 are slower compared with TNF. The results for TNF and IL-10 production in specific enzyme-linked immunosorbent assay (ELISA) systems are shown in Figure 4. The upper panel (Figure 4A) compares the TNF production after 4 h ± 1 µg LPS in PBMC (left) and PLM (right). The average constitutive TNF production in PBMC is 5.7 ± 4.2 pg/ml versus 229 ± 251 pg/ml in PLM. After stimulation with LPS, the TNF production in PBMC increased to 834 ± 368 pg/ml, whereas in PLM the mean value rose to 5,226 ± 3,000 pg/ ml. A similar pattern of results was obtained for IL-10 protein after 16 h incubation ± 1 µg LPS/ml. There was no constitutive IL-10 in PBMC (0 pg/ml) (sensitivity of the assay: 3 pg/ml), whereas PLM constitutively produced 68 ± 34 pg/ml. Stimulation with LPS led to 888 ± 625 pg/ml in PBMC and to 2,191 ± 1,507 pg/ml in PLM. In conclusion, the cytokine production for both TNF and IL-10 was stronger in PLM.

View larger version (34K): [in this window] [in a new window]   Figure 4.   Cytokine production: ELISA. A total of 1 × 106 PLM and PBMC per ml, respectively, were incubated in the presence or absence of 1 µg LPS/ml. For the measurement of cytokines, (A) supernatants were harvested after 4 h for TNF or (B) after 16 h for IL-10. Supernatants were then tested with commercial ELISA kits. The amount of produced cytokines is given in pg/ml. The left bars indicate control without LPS, the right bars represent LPS-stimulated cells. Given is the average of five experiments (*P < 0.05 as compared with stimulated PBMC or **P < 0.05 compared with unstimulated PBMC).

TNF and IL-10 mRNA levels in PBMC and PLM were analyzed by reverse transcriptase (RT)-PCR in fresh cells at time 0 and in cells cultured with and without LPS stimulation. The results from one of five patients are shown in Figure 5. In the upper panel the ethidium bromide-stained agarose gel for TNF in PBMC and PLM with internal standard and in the lower panel the gels for -enolase as external control are shown in Figure 5A. Constitutive TNF mRNA was very low in both PBMC and PLM and strongly inducible with LPS, comparable to the protein data. A striking finding is shown in the lower panel (Figure 5B) for the IL-10 mRNA: PLM showed very high amounts of constitutive IL-10 mRNA at time 0 and in unstimulated control cells cultured for 16 h. Stimulation with LPS could only minimally increase this strong signal (Figure 5B). By contrast, in PBMC there was no constitutive IL-10 mRNA and strong induction by LPS.

View larger version (52K): [in this window] [in a new window]   Figure 5.   Cytokine production: RT-PCR. PCR analysis for TNF and IL-10 mRNA in PBMC and PLM. Comparison of TNF (A) and IL-10 (B) in PBMC and PLM. Cells from the same patient with central bronchial tumor were either left untreated or stimulated for 4 h with LPS (1 µg/ml). A total of 2 × 104 cells was lysed and mRNA was isolated and amplified by PCR. Amplification for -enolase was used as an external control (one of five experiments).

Patients with Nonmalignant Disease

Four patients in the study population were not suffering from bronchial carcinoma but from nonmalignant disease (Table 1). When comparing data from those patients with the rest of the study population, the phagocytosis capacity and the staining pattern of the PLM for the different markers were similar to the samples from patients with nonmalignant disease (data not shown). The same was true for the unique finding of constitutive IL-10 mRNA. As shown in Figure 6, lanes 1 and 2, there is strong constitutive IL-10 mRNA in PLM from two patients with nonmalignant disease, which is similar (lane 1) or higher (lane 2) than the constitutive level seen in a patient operated for a non-small cell lung cancer (NSCLC) (lane 3). Also, shown in lane 5 (Figure 6) is the absence of constitutive IL-10 mRNA in PBMC.

View larger version (41K): [in this window] [in a new window]   Figure 6.   IL-10 mRNA expression in PLM from patients with nonmalignant disease. Comparison of IL-10 (upper panel) and -enolase mRNA as external control (lower panel) in PLM and PBMC. Cells were either left untreated or stimulated 16 h with 1 µg LPS/ml. Lanes 1 and 2 represent PLM of two different patients with nonmalignant disease. Lanes 3-6 show data from one patient with nonmalignant disease.

	Discussion

Top Abstract Introduction Material and Methods Results Discussion References

In our aim to characterize normal human PLM concerning surface antigen expression, capacity for phagocytosis, and cytokine production, we found PLM to represent a cell-type intermediate of regular CD14⁺⁺ and CD14⁺CD16⁺ monocytes, two well-defined subsets in human peripheral blood. Besides high HLA-DR and ICAM-1 expression, the most striking finding was a high constitutive IL-10 mRNA production in PLM that could not be substantially increased by LPS stimulation.

Tissue macrophages show a great heterogeneity concerning their immunologic phenotype and function. They belong to the monocyte/macrophage system and develop in various tissues by maturation and differentiation from immigrating blood monocytes. A phenotypic and functional heterogeneity of monocytes has also been described in human peripheral blood in which CD14⁺CD16⁺ and CD14⁺⁺ monocyte subpopulations could be defined (6). The CD14⁺CD16⁺ cells are functionally distinct from the classic monocytes because they efficiently produce the proinflammatory cytokine TNF but they are unable to produce the anti-inflammatory IL-10 (7). This proinflammatory functional pattern of the CD14⁺CD16⁺ monocytes fits well with the expansion of these cells in various inflammatory diseases (8). It is unclear at this point whether these CD14⁺CD16⁺ monocytes are committed to differentiate toward specific types of tissue macrophages.

Most members of the tissue macrophage family, such as peritoneal macrophages, alveolar macrophages, or Kupffer cells, are well-characterized, but very little is known about normal human PLM. Most of the available data concerning the PLM refer to cells obtained from patients with pleural effusions (2). Our aim was therefore to characterize normal human PLM from patients without pleural effusion. Because evidence has been reported that PLM might derive from blood monocytes (1), we compared the features of PLM with the previously mentioned CD14⁺CD16⁺ monocytes and the regular CD14⁺⁺ blood monocytes to describe common features or tissue-specific characteristics and differences.

A major common marker for cells of the monocyte/ macrophage lineage is the CD14 antigen, the receptor for LPS. CD14 is expressed at different densities on the various types of macrophages and even on blood monocyte subsets (9). Regular blood monocytes show a strong expression of CD14 antigen (indicated by "⁺⁺"), whereas it is lower on the CD14⁺CD16⁺ subset (indicated by "⁺"). Also, among the tissue macrophages CD14 is differentially expressed. Alveolar macrophages are weaker in CD14 expression compared with blood monocytes, whereas peritoneal macrophages have upregulated CD14 (9). When using CD14 as a marker to identify monocytes and macrophages, then the pleural lavage population contains about 20% PLM, which is very similar to what is found for monocytes among the PBMC in blood. The recently described blood monocyte subset CD14⁺CD16⁺ coexpresses the CD16 antigen (low-affinity Fc receptor) on its surface, which is absent on regular blood monocytes, but also detectable on alveolar macrophages. In Figure 1, we compared the expression of CD14 and CD16 on PLM with CD14⁺⁺ regular blood monocytes and the CD14⁺CD16⁺ subset. Regarding the staining pattern for CD16, the PLM exhibit an intermediate level of expression, whereas concerning the CD14 expression, PLM tend to be more similar to regular blood monocytes. Furthermore, we tested the expression of various monocyte/macrophage-associated surface antigens such as the other two types of Fc receptors CD32 and CD64, the myelomonocytic marker CD33, the adhesion molecules CD11b and CD54, and finally the class II antigen HLA-DR. Compared with the two blood monocyte types, the expression on PLM for CD11b, CD64, and CD33 was more similar to the regular CD14⁺⁺ monocytes, whereas we found a higher density of CD32 and CD54 on the PLM. The most striking difference was the extremely strong expression of HLA-DR class II molecules (about 20 times higher than on regular CD14⁺⁺ monocytes) on the PLM. One might assume that the PLM exhibit a high antigen-presenting capacity because of their extremely high density of class II antigens and the cellular adhesion molecule CD54 (ICAM-1), but this remains to be elucidated in further studies. Taken together, the PLM represent an intermediate cell type between the regular CD14⁺⁺ blood monocytes and the CD14⁺CD16⁺ subset in regards to their immunologic staining pattern. They share many common features, but there are also striking differences that may be due to the tissue-specific functions of the PLM in the pleural space.

Gjomarkaj and coworkers (2) also reported an expression of CD14 and HLA-DR in PLM from effusions, but they did not address receptor density and no comparison to an internal control such as blood monocytes was provided. Therefore, it is unclear whether the PLM from effusions analyzed by Gjomarkaj and colleagues have a phenotype similar to the resting PLM described herein.

Because of the prominent expression of the Fc receptors on the PLM, we performed a phagocytosis assay with antibody-coated SRBC to compare the capacity of PLM and blood monocytes to internalize opsonized particles via their Fc receptors. We modified the method described by Ziegler-Heitbrock and associates (10) with fluorescent- labeled SRBC for analysis in flow cytometry. The advantages of this method are (1) that SRBC outside the cell can be lysed by formic acid treatment and (2) that internalized SRBC can be detected by an increase in forward scatter and at the same time by an increase in fluorescence intensity. Compared with the entire blood monocyte population, the PLM were able to more efficiently phagocytose the opsonized SRBC. This effect may be due to the stronger expression of all three types of Fc receptors.

The CD14 molecule is a common marker for cells of the monocyte/macrophage lineage and represents the receptor for LPS. Via the CD14 receptor, the cells are stimulated to produce pro- and anti-inflammatory cytokines. Because the CD14 expression is high on the PLM, we tested their capacity to release cytokines constitutively and after stimulation with LPS at the protein and the mRNA level. As a proinflammatory cytokine, we chose TNF, as an anti-inflammatory, IL-10. For cytokine production, we incubated PLM and PBMC from the same donors in LPS-free culture medium and then stimulated them without and with LPS (1 µg/ml) for different incubation times. The measurement of TNF and IL-10 at the protein level revealed no constitutive production in PBMC but considerable amounts in cultured PLM. When stimulating with LPS, both cell types secreted large amounts of cytokines, but the levels in PLM were higher for both TNF and IL-10. The larger constitutive amounts of TNF and IL-10 in PLM might be due to a possible activation of the cells during the process of pleural lavage or to the PBS solution used, but all reagents were tested for the absence of endotoxin by Limulus amebocyte assay and the PBMC were exposed to the same type of reagents during the separation procedure that is comparable in time with the pleural lavage. Another argument against preactivation of the PLM is the high IL-10 mRNA production at time point zero, which is absent in PBMC. Given the slow kinetics of IL-10, several hours of incubation at 37°C would be required in order to reach this level of IL-10 transcripts. Hence, we conclude that the constitutive IL-10 mRNA found in the PLM is, in fact, present in vivo.

One also could imagine that in PLM with these large amounts of preformed mRNA, LPS only has to trigger translation in order to allow for rapid expression of IL-10 protein. The larger amounts of LPS-induced cytokines in PLM could also be due to a higher percentage of CD14-positive cells among the pleural lavage cells, but we detected comparable amounts of CD14-bearing cells in PBMC and pleural lavage. Hence, we conclude that PLM are able to secrete larger amounts of both TNF and IL-10 upon LPS exposure.

It is surprising to find such a high constitutive expression of IL-10 mRNA in resting PLM. Most cytokines, including IL-10, are not expressed constitutively but their production is induced by immunologic stimuli in the different leukocytes. It is unclear what molecular mechanisms are responsible for the constitutive IL-10 mRNA expression. It may involve signals provided by molecules that are specifically expressed in the pleural space and that trigger transcription factors like SP-1 or Stat3 (11, 12). In contrast, the differentiation of monocytes to PLM may lead to permanent upregulation of transcription factors that control the IL-10 gene. Further analysis of gene expression in PLM is required in order to resolve these questions.

With abundant IL-10 mRNA being present there is also some constitutive protein production, and large amounts of IL-10 protein can be induced by LPS stimulation. IL-10 is an anti-inflammatory cytokine that downregulates the immune response by suppressing other cytokines such as TNF and immunoreceptors such as major histocompatibility complex class II (13). Hence, the local production of IL-10 may serve to dampen an otherwise excessive immune response in the pleural space upon infection or malignancy.

Gjomarkaj and coworkers (2) also studied cytokine expression in PLM and they found that effusion macrophages are capable of producing IL-1 and TNF. The question whether pleural effusion macrophages have a higher cytokine production as compared with blood monocytes was not addressed, and the anti-inflammatory cytokine IL-10 was not studied. Based on our current finding of a high constitutive and inducible cytokine production by resting PLM, it will be of interest to analyze expression of these cytokines in malignant and infectious pleural effusions in comparison to resting cells.

One might assume that the characteristics found for PLM in this study could be influenced by the malignant tumor despite the fact that malignancy was restricted to the bronchus and did not extend to the pleural space. When comparing the features of cells from patients with malignant and nonmalignant disease, we could not detect any striking differences. This indicates that the findings reported herein, in fact, apply to resting PLM. Taken together, this first characterization of normal, noneffusion human PLM can form the basis for a better interpretation of findings in malignant and inflammatory exsudates.