Pathways of Fibrin Turnover of Human Pleural Mesothelial Cells In Vitro Steven Idell, Christian Zwieb, Anuradha Kumar, Kathleen B. Koenig, and Alice R. Johnson Departments of Medicine, Molecular Biology, and Biochemistry, University of Texas Health Science Center, Tyler, Texas

The mesothelium contains both procoagulant and fibrinolytic activities. An imbalance between these activities could account for the abnormal fibrin turnover and pleural fibrin deposition that is characteristic of pleural inflammation. Procoagulant activity of human pleural mesothelial cells (HPMC) is in part due to tissue factor, and the prothrombinase complex can also assemble at the HPMC surface. HPMC express tissue plasminogen activator (tPA) but no detectable fibrinolytic activity in a fibrin plate assay. Inhibition of HPMC fibrinolytic activity is due, in part, to elaboration of plasminogen activator inhibitors-l and -2 (PAI-l and PAI-2) as well as antiplasmins. Synthesis of PAI-l and PAI-2 is inhibited by actinomycin D and cyclohexamide. HPMC PAI-l is increased by transforming growth factor-S (TGF-fj) and tumor necrosis factor-a (TNF-a), as is tPA release, while PAI-l mRNA is unchanged and tPA mRNA is increased. PAI-2 release is induced by TNF-a and TGF-{3. Because they are a rich source of PAI-l and PAI-2, HPMC may contribute to the high levels of these inhibitors in pleural exudates. Stimulation of HPMC by TNF-a or TGF-fj in vitro did not alter HPMC procoagulant activity nor the balance of elevated PAl and antiplasmins relative to PA, changes that collectively favor formation and persistence of pericellular fibrin.

with disease involving the pleural space. We recently reported that exudative pleural effusions from patients with infectious or neoplastic diseases contained markers of increased coagulation accompanied by decreased fibrinolytic activity (9). The fibrinolytic defect in these pleural exudates appears to be, in large part, attributable to an increase of plasminogen activator inhibitor-l (PAl-I). This inhibitor increases in pleural exudates to concentrations of about 1,000fold higher than those in plasma (9). PAI-2 was also found in pleural fluids and was increased in pleural exudates. The cellular sources of these inhibitors are unknown but may include mesothelial cells or fibroblasts. Mesothelial cells that line the pleural cavity and the lung surface may influence local fibrin deposition and clearance. Recent studies implicate the mesothelium in abnormalities of fibrin turnover. Mesocrin, a 46,000 M, glycoprotein secreted by human peritoneal mesothelial cells, was recently identified as PAI-l by analysis of its cloned DNA sequence (10). Human omental mesothelial cells express both tissue plas(Received in original form December 26, 1991 and in revised form May 5, minogen activator (tPA) and PAI-l and PAI-2, and tumor 1992) necrosis factor-a (TNF-a) decreases tPA while increasing Address correspondence to: Steven Idell, M.D., Ph.D., Professor of Mediboth PAIs (11). From these observations, we suspected that cine, University of Texas Health Science Center at Tyler, Routes 271 and 155, Tyler, TX 75710. human pleural mesothelial cells (HPMC) could promote fibrin deposition in the pleural space. In order to address this Abbreviations: fetal calf serum, FCS; Hanks' balanced salt solution, HBSS; human lung fibroblast(s), HLF; human pleural mesothelial cell(s), HPMC; . possibility, we characterized the procoagulant and fibrinolyplasminogen activator, PA; plasminogen activator inhibitor, PAl; polymertic activities of untreated HPMC and of cells exposed to ase chain reaction, PCR; sodium dodecyl sulfate, SOS; Tris-HCl EDTA transforming growth factor-S (ffiF-fj) and TNF-a. Our buffer, TE; transforming growth factor-S, TGF-,8; tumor necrosis factor-a, studies indicate that HPMC contribute to pleural fibrin TNF-a; tissue plasminogen activator, tPA; urokinase plasminogen activator, uPA. deposition by expression of procoagulant activity and by regulation of local pathways of fibrinolysis. Am. J. Respir. Cell Mol. BioI. Vol. 7. pp. 414-426, 1992 Extravascular, pleural fibrin deposition is a hallmark of pleural inflammation and occurs in a wide variety of pleural diseases, including parapneumonic effusions and hemothoraces (1-3). The fibrin gel is the end product of hemostasis, but the effects of fibrin formation (coagulation) and dissolution (fibrinolysis) influence several elements of the inflammatory response, including increased microvascular permeability and scarring (4,5). The importance of these events in human pleural injury is suggested by the incidence of pleural scarring and lung entrapment from large hemothoraces (1). Additionally, fibrinolytic agents, s'uch as streptokinase, can break up pleuralloculations and facilitate drainage of loculated effusions, such as those associated with serious infections (6, 7). Conversely, intrapleural instillation of blood has been used to achieve therapeutic pleurodesis (8). Pathways of fibrin turnover are disrupted in association

Idell, Zwieb, Kumar et al.: Pleural Mesothelial Cell Fibrin Turnover

Materials and Methods Cell Culture and Preparation !IPMC were isolated from pleural fluids of patients undergomg thoracentesis for clinical indications. The material was harvested in accordance with guidelines developed by the Human Subjects Institutional Review Committee of the University of Texas Health Science Center at Tyler (Tyler, TX). A total of 27 HPMC cultures, all from different patients, were used in these studies. Cultures were derived from patients with parapneumonic effusions (n = 2), pleural effusions secondary to involvement of the pleura by carcinoma of the lung (n = 16), or congestive heart failure (n = 4), with all clinical diagnoses fulfilling previously reported criteria (9). Two additional patients had hemothoraces secondary to trauma, one had biopsy-proven pleural involvement by lymphoma, one had cirrhosis, and one had asbestos exposure but no diagnosis of the pleural effusion despite extensive evaluation. Mesothelial cells were cultured from pleural fluids by modification of published methods (12-14). The fluid was centrifuged to sediment the cells (1,000 X g for 10 min); the cells were resuspended in culture medium 199 supplemented with 15% heat-inactivated fetal calf serum (FCS) and 1% antibiotics and plated in tissue culture flasks. The cultures were incubated at 37° C in an atmosphere of 95% air and 5% CO 2 • The cells grew slowly from small adherent patches in the initial isolate to establish a contact-inhibited monolayer of polygonal , epithelioid cells. Medium was replaced 3 times weekly, and the cells were subcultured when they became confluent, usually within 1 to 2 wk after initiation of the culture. Although some exudates originally contained phagocytes and erythrocytes, these cells were eliminated as the cultures were passaged. The mesothelial cells in starting cultur~s were clearly distinguished from other cells both by their general appearance in culture and by reaction with antibodies to o-cytokeratin (12). Cells from a mesothelioma cell line, MS-1 (15) (a genero~s gift from Dr. H. S. Hsu of the University of Arkansas, LIttle Rock, AR), were cultured to compare with cells obtained from the patient fluids. The mesothelioma cells were incubated and passaged just as the cells from pleural fluids. Human lung fibroblasts were also compared with the mesothelial cells. Fibroblast cultures were initiated from the explants of lung tissue taken from surgical specimens. Small pieces of tissue (approximately 2 rnm') were placed in culture flasks and covered with culture medium containing 10% FCS and antibiotics; within 5 to 7 days, cells migrated from the tissue and formed clusters of elongated, spindle-shaped cells. The fibroblast cultures, also contact-inhibited monolayers, had a distinctly different appearance from mesothelial cells and were not recognized by antibodies to o-cytokeratin. Characterization of Cultured Cells with Antibodies Cultures of mesothelial cells and fibroblasts were grown on chamber slides, fixed with 100% methanol for 15 min, and reacted with commercial antibodies to o-cytokeratin and vimentin. Antibody binding was detected by a fluorescently labeled second antibody. The cultures were washed twice in Hanks' balanced salt solution (HBSS) containing 10% FCS, and incubated with unlabeled primary antibodies to o-cyto-

415

k~ratin. (Biologi~al Technologies Inc., Sloughton, MA) or vimentm (Boehrmger-Mannheim, Indianapolis, IN) for 30 m~n at ~oom tempera~ure. Controls were treated with appropnate Irrelevant antibodies. The cultures were washed 3 times with serum-containing HBSS, and fluorescently labeled secondary antibodies were applied. Incubation was continued for 30 min; the specimens were washed 3 times with medium containing 15% FCS, once with phosphatebuffered saline, and mounted. The cells were photographed through a 20 X objective using a 35-mm camera attached to an Olympus AH-2 microscope. In other experiments, HPMC were cultured on chamber slides. The medium was removed, and the chambers were washed once with HBSS. Pooled human plasma was added to duplicate chambers, and then CaCb was added to a final concentration of 1 mM. Additional chambers were treated only with HBSS. Once the plasma clotted (within 2 to 3 min), the plates were fixed with 100% methanol for 15 min at r~~m temperature. The clot was removed by suction, and a~dl~lOnal methanol was added to the cells for 15 min longer. ~lbnn .formed on ~nd. around the cel~s was identified by reacnon WIth a 1:50 dilution of goat anti-human fibrinogen antibody (Chemicon International, Temecla, CA) for 30 min at room temperature and developed with a fluorescently lab~led rabbit anti-goat gamma globulin (Zymed, San FranCISCO, CA) (1:20) for 30 min. The specimens were observed and photographed through an Olympus AH-2 fluorescent microscope.

Coagulation Assays Recalcification times of HPMC or supernatants were measured in normal pooled plasma. Briefly, the recalcification times of equal volumes of sample, normal pooled plasma, and 25 mM CaCl 2 were determined using a fibrometer (Fibrosystems, Orangeburg, NY) (9). Neutralization of procoagulant activity was performed as previously described and the results of each culture studied were confirmed in duplicate experiments (9). Indirect binding studies reflected the ability of the extrinsic activation complex (tissue factor associated with factor VII) and prothrombinase complex (factors Xa and Va and prothrombin) to assemble at the HPMC surface. These studies were performed using amidolytic assays of factor X activating activity and prothrombin-activating activity, as previously described (9, 17). In additional experiments, replicate samples of 105 HPMC in 24-well plastic Falcon plates were incubated at room temperature with 1 ml of normal pooled plasma (George King, Overland Park, KS) and 1 ml of 25-mM CaCI2 , and the recalcification time of replicate samples was determined by tilt testing. Fibrin Plate Radioassay, Fibrin Gel Enzymography, Reverse Fibrin Gel Enzymography, and Measurement of Antiplasmin Activity Plasminogen activator (PA) activity and plasminogen indepen.dent fibrino~ytic activity were measured in a fibrin plate radioassay, WhICh quantitates the ability of a sample to cleave ra~iolabeled fibrin (17). Fibrin gel enzymography, reverse fibnn gel enzymography, and antibody neutralizations of the PA and PA inhibitor (PAl) activities were performed

416

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 7 1992

as previously described (9). These techniques involve electrophoresis of samples after which zones of fibrinolysis (enzymography) or inhibition of fibrinolysis (reverse enzymography) are identified using a fibrin gel overlay. In some experiments, the fibrinolytic activity of HPMC was assessed in the fibrin plate assay using buffer supplemented with 1% sodium dodecyl sulfate (SDS) (Sigma Chemical Co., St. Louis, MO). This mixture activates any latent PAl-I, as described by Levin (16). For concentration and time course experiments, it was determined that the full effect of SDS on HPMC fibrinolytic activity was achieved at a final sample concentration of 1% and incubation for 10 min at 37° C. Samples were first treated with 1% SDS (controls were untreated), incubated for 10 min at 37° C, adjusted to 5 % Triton X-100, after which the samples were diluted in 0.1 M sodium phosphate-5 % Triton X-100, and assayed by the fibrin plate method. Antiplasmin activity was determined by measuring the residual plasmin activity present in a 1:1 (vol:vol) mixture of sample and a 1.25 j-tg/ml plasmin standard incubated at 37° C for 10 min (9, 17). Immunoassays tPA, urokinase plasminogen activator (uPA), PAl-I, and PAI-2 were measured by commercially available enzyme immunoassays. Assays for tPA, uPA, and PAI-1 were obtained from American Diagnostica (Greenwich, CT) and were sensitive to 0.07,0.1, and 1 ng/ml, respectively. According to the manufacturer, the PAI-1 assay measures both free (both active and latent) PAI-1 and that bound to tPA. The tPA assay measures both free tPA and that bound to PAL The PAI-2 immunoassay, which is sensitive to 6 ng/ml, was obtained from Biopool (Umea, Sweden). TNf-a concentrations in pleural fluids were measured using the TNF-a Quantikine immunoassay kit (Research and Diagnostics Systems, Minneapolis, MN). We measured TNF-a in eight pleural fluids from which HPMC were cultured, including two randomly selected from each group (patients with congestive heart failure, pleural involvement by lung cancer, or parapneumonic effusions). Twopleural fluids from patients with empyema (9) were also studied but HPMC were not isolated from these fluids. Influence of Protein Synthesis Inhibitors and Lipopolysaccharide on HPMC Function Procoagulant and fibrinolytic . activities of HPMC were evaluated after treatment with cyclohexamide and actinomycin D. Confluent monolayers were treated for 18 h with serumless medium (Newman-Tydall) alone or with actinomycin D (2 j-tg/ml) or cyclohexamide (10 j-tg/ml). The cells were detached with Puck's EDTA, counted in an electronic cell counter (Coulter Electronics), and adjusted to the same cell concentration. One-stage coagulation, fibrinolysis, and immunochemical assays were done as described. HPMC were also treated with lipopolysaccharide (lipopolysaccharide B Escherichia coli 055:B5; Difco Laboratories, Detroit, MI), at concentrations of 1 and 10 j-tg/ml in serumless medium for 18 h. The cells were detached, counted, and assayed as described above. Cell Treatments with Cytokines HPMC were treated with 1, 5, and 10 ng/ml TGF-13 (Biomedical Technologies) or 100, 200, and 400 V/ml TNF-a (Gen-

zyme, Cambridge, MA) in serumless medium for 18 h at 37° C, unless otherwise specified. Controls were incubated with serumless medium and assayed as described above. Amplification and Detection of PAl-I, PAI-2, tPA, and uPA mRNA Using the Polymerase Chain Reaction (PCR) Oligonucleotides specific for the coding region of tPA, uPA, PAl-I, and PAI-2 were synthesized on an Applied Biosystem (PCR-mate) DNA synthesizer (trityl-on) using {j-cyanoethylphosphoramidite chemistry. Oligonucleotide primers for the reaction are shown in Table 1. Forward and reverse oligonucleotides were purified and detritylated on cartridges supplied by the manufacturer. Total cytoplasmic RNA was isolated from cells using standard techniques (34). cDNAs were synthesized in a 100-j-t1 reaction containing PCR buffer (10 mM Tris-HCI [pH 8.3], 50 mM KCI, 1.5 mM MgCI2 , 0.001% gelatin), 11 mM dithiothreitol, 200 j-tM of each dNTPs, 10 V RNasin (Promega), 20 pMol antisense primer, 1 j-tg total cellular RNA, and 100 V MoMuLV reverse transcriptase (BRL). Incubation was for 1 hat 37° C after which the enzyme was inactivated by incubation for 5 min at 95 ° C. Amplification of cDNAs was carried out in a 100-j-t1 reaction volume in PCR buffer, 200 j-tM of each dNTPs, 100 pMol of forward and reverse primers, various concentrations of cDNA, and 2.5 V of Thermus aquaticus polymerase (Amplitaq'" DNA polymerase; Perkin-Elmer Cetus). The solution was overlaid with 100 j-tl mineral oil, and the cDNA was amplified in a Perkin-Elmer Cetus temperature cycler. The PCR cycles were 30 s at 95° C to denature the template DNA, 2 min at 55° C to allow the primers to anneal, and 2 min at 72° C for 30 cycles for DNA extension. After amplification, 150 j-tl of 10 mM Tris-HCI (pH 7.5), 1 mM EDTA (TE) was added and the water phase was removed and mixed with 200 j-tl ofTE-saturated chloroform. After brief centrifugation, the water phase was collected, extracted with icecold ethanol containing 3 M Na acetate (pH 6.0). The DNA precipitated during a 20-min incubation at -80° C and was collected by centrifugation. The pellet was washed with 300 j-tl of 80% ethanol, centrifuged for 5 min, dried, and dissolved in a small volume ofTE. An aliquot of the sample was analyzed by electrophoresis on a 2 % agarose or 8 % acrylamide gel, and the DNA was visualized by staining with ethidium bromide. The PAI-l product is 274 bp long; its authenticity was verified by digestion of the DNA with SAlI restriction endonuclease generating 227- and 47-bp fragments, as analyzed by agarose gel electrophoresis. The PAI-2 product is a 415-bp DNA that yields 240- and 175-bp products when digested with XbaI. When digested with EcoR I and BamHI, uPA DNA yielded products of 295 and 2,900 bp. When digested with EcoR I and BgI II, tPA DNA yielded the expected products of 260, 1,170, and 1,700 bp. This method easily detected the uPA, tPA, PAl-I, and PAI-2 mRNA from 1Q4 cells.

Results Characterization of HPMC Small numbers of epithelioid cells were initially obtained from each of the pleural fluids. These grew to confluence in 4 to 6 days, forming a monolayer. The cells could be propagated over four to six passages, after which time growth

Idell, Zwieb, Kumar et al.: Pleural Mesothelial Cell Fibrin Thrnover

417

TABLE 1

Oligonucleotide probes used to detect mRNAs for tPA, uPA, PAl-I, and PAl-2 tPA Forward primer: Reverse primer: uPA Forward primer : Reverse primer: PAI-l Forward primer : Reverse primer: PAI-2 Forward primer : Reverse primer:

CATAAGGAATTCGATGATGACACTTACGAC GACAGATCTGTTAAGTAAATGTTGTGATGTG AAAGAGAATTCTACCGACTATCTCTATCCCG TGCGGATCCAGGGTAAGAAGTGTGAGAC GGCTGGT ACCCCTCCTGGTTCTGCCCAA GAGGATCCGTCTGTCCATGATGATCTCCTC GGGGTACCAGGATGGTCCTGGTGAAT CTCCAGGATCCTGCTTCTCAGAATGGATC

declined and the cells became senescent. All studies reported below were performed using HPMC from the first three passages. The HPMC could be distinguished from human lung fibroblasts (HLF) by growth characteristics and general ap-

pearance. HPMC had typical immunocytochemical staining (n = 9 HPMC, n = 5 HLF cultures studied), and the appearance of the HPMC and HLF as shown in Figure I is typical of the cells used in this study. Antibody to cytokeratin (Figure lA) recognizes HPMC but not HLF (Figure lB). As

A

c

D

Figure 1. (A) HPMC and HLF staining after exposure to antibodies to a-cytokeratin and vimentin. Representative findings of nine HPMC can be distinguished from fibroblasts (n = 5 cultures studied) by general appearance and immunostaining. The cells were fixed with methanol and reacted with antibodies to a-cytokeratin or vimentin. Antibody binding was detected with FITC-labeled second antibody. Antibody to cytokeratin (top panels) recognizes HPMC (A) but not HLF (B). Both types of cells stain with antibody to vimentin (C and D) .

418

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 71992

A

B

c

Figure 2. HPMC shown in phase contrast (A) stained positively for fibrin(ogen) after treatment with pooled human plasma in the presence of calcium. (B) Cells that were reacted with a goat antibody to fibrinogen and developed with a fluorescently labeled rabbit anti-goat IgG. (C) Lack of staining in the absence of fibrin formation (cells treated only with HBSS). Cells reacted with an irrelevant primary antibody before addition of the fluorescent secondary antibody were also unstained (not shown).

reported by others, HPMC also stained with vimentin, as did HLF (Figures 1C and 10) (11). Demonstration of TNF-a in Pleural Fluids TNF-a was not detected in the pleural fluids of patients with congestive heart failure or lung cancer (n = 2 samples/ group) but was found in both fluids of patients with parapneumonic effusions (184.3 and 496.0 pg/ml) and in one of the fluids from a patient with empyema (33.4 pg/ml) .

Procoagulant Activity of HPMC When incubated for 10 min in the presence of recalcified normal pooled plasma at room temperature, HPMC invested themselves in a coagulum and assembled fibrin at the cell surface (Figure 2). In three cultures studied, the mean recalcification time by tilt testing (105 cells in 1 ml plasma) was 105.2 s versus more than 13 min in the absence of HPMC. Tissue factor was recently identified as the major procoagulant present in human pleural fluids (9). Antibody to tissue factor (50 JLg/ml of the IgG fraction of a monoclonal antibody) extended the median (fibrometer-measured) recalcification time of HPMC to 179.9 s (range, 103.4 to 195.0 s) versus 39.3 s (range, 25.4 to 44 .0 s) in an identical volume of diluent buffer or 37.4 s (range, 25.9 to 43.4 s) in the presence of control mouse IgG (n = 3 cultures , 104 celIs/assay). These data indicate that tissue factor procoagulant activity is expressed by HPMC. Tissue factor was likewise released into the serum-free media by HPMC; the median recalcification time of medium incubated with anti-tissue factor IgG was 232.9 s versus 87.4 s or 88.5 s in the 0.15 M NaCl, 0.02 M Tris-HCl, ph 7.4 (lCB)/BSA and mouse IgG controls, respectively. Recalcification times of the buffer and antibodies in recalcified plasma without cells exceeded 300 s. Mesothelioma (MS-l) cells also expressed tissue factor ; the recalcification time of 10' cells in the presence of anti-tissue factor IgG was 359.2 s, while that in lCB/BSA was 128.1 s and that in the presence of control mouse IgG was 152.9 s. HPMC (n = 7 cultures studied) expressed factor X activating activity and were next studied in neutralization experiments. In seven HPMC cultures studied, factor X activating activity in the presence of control antibody, 50 JLg/ml normal mouse IgG (median, 35.1; range, 2.0 to 58.9 ng/min/ml), could be inhibited ~ 95 % in the presence of the same concentrations of either mouse monoclonal anti-human tissue factor IgG (median , 0; range, 0 to 6.6 ng/min/ml) or rabbit anti-human factor VII IgG (median, 0; range, 0 to 9.2 ng/ min/ml) . The factor X activating activity in the presence of control antibody was the same as that in lCB/BSA (data not shown). These data indicate that tissue factor is expressed by HPMC, and the factor X activating activity expressed by these cells is related to the extrinsic activation complex, tissue factor associated with factor VII . Indirect Binding Studies We next evaluated the ability of HPMC to express extrinsic activation complex activity by indirect studies of the binding of factor VII to the surface of live HPMC. Figure 3 shows data from 11 HPMC cultures grown from pleural effusions from two patients with congestive heart failure, eight patients with pleural spread of lung cancer, and one patient with a parapneumonic effusion. All cultures expressed increased factor X activating activity in the presence of increasing concentrations of factor VII. These data indicate that tissue factor, in part associated with factor VII, is abundant at the HPMC surface and that factor VII is rate-limiting for expression of the procoagulant activity of the extrinsic activation complex. MS-1 cells and HLF exhibited similar properties (data not shown) . To determine whether the prothrombinase complex assembled at the HPMC surface, we sequentially added components of this complex, including factors II (prothrombin),

Idell, Zwieb, Kumar et al.: Pleural Mesothelial Cell Fibrin Turnover

800,--------------------,

419

200

C

~

'0 E

]= '0

...E

160

~

c~

.S

600

~

120

E 1.0

0

~

.:

80

,Q

E 0 1.0

.c:

o

~

0

~

~

II

40

Eo-
1.5 X the outermost values in the box. The standard was diluted in assay buffer (the lysate diluent) or serumless medium and is indicated by the horizontal solid line.

424

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 7 1992

PA I-1

PAI-2 uPA

tPA

1 2 3 4 5 6 7 8 9

Figure 8. Polymerase chain reaction analysis of HPMC mRNAs for PAl-I, PAI-2, uPA, and tPA. mRNAs isolated from total cellular RNA using specific oligonucleotide probes were converted to the DNAs using reverse transcriptase and amplified by PCR. Lane 1: DNA standards; lanes 2, 4, 6, and 8: mRNA from control HPMC for PAl-I, PAI-2, uPA, and tPA, respectively; lanes 3, 5, 7, and 9: mRNA from HPMC treated with 10 ng/ml TGF-t3 for 24 h at 37° C.

congestive heart failure, pleural exudates demonstrate increased procoagulant activity and depressed fibrinolytic activity (9). These abnormalities individually and collectively favor pleural fibrin deposition. The fibrinolytic defect in the pleural exudates was associated with high concentrations of both PAI-l and PAI-2, relative to plasma or pleural fluids from patients with congestive heart failure (9). Agrenius and colleagues reported that fibrinolytic activity in pleural effusions from patients with pleural involvement by cancer decreased with inflammation induced by quinacrine pleurodesis (25). Concentrations of PAI-l in these pleural fluids rose concurrently with the decrease in fibrinolytic activity. In an animal model of adhesive pleuritis, procoagulant activity was retained but fibrinolytic activity was undetectable (26). These studies confirm the association between augmented pathways of fibrin formation and/or impaired clearance and fibrin deposition in pleural disease. The proximity of mesothelial cells suggests that they could contribute to fibrin turnover in the pleural compartment. To our knowledge, the procoagulant pathways of HPMC have not been previously elucidated. Our data indicate that HPMC in vitro can initiate coagulation by the assembly of the extrinsic activation complex, tissue factorfactor VII, at the cell surface, and the availability of factor VII is rate-limiting for expression of factor X activating ac- . tivity. In addition, distal procoagulants factor Xa, factor VIVa, and prothrombin assemble on the HPMC surface as the prothrombinase complex. This complex can potentiate the process of coagulation. Other cell types, such as fibroblasts, may express surface procoagulant activity and could

likewise initiate fibrin formation in the pleural space. The contribution of these cell types to pleural fibrin formation could be relatively greater in diseases associated with mesothelial denudation. There appears to be little or no difference in expression of procoagulant activities by HPMC from patients with congestive heart failure compared with those with exudative processes such as pleural carcinomatosis. The shortened recalcification times of normal pooled plasma in contact with HPMC implicates the mesothelium as a likely site for initiation of coagulation. Availability of extravasated coagulation substrates to interact with HPMC would promote pleural fibrin deposition associated with increased microvascular permeability in pleural injury. Mesothelioma cells retain similar procoagulant properties, which may influence the spread of these tumors, as reviewed elsewhere (5). Irrespective of the underlying disease process, we found that HPMC or HPMC conditioned media uniformly express no fibrinolytic activity in our fibrin plate assay. HLF or cultured MS-l mesothelioma cells, however, express readily detectable plasminogen-dependent fibrinolytic activity. The lack of detectable HPMC fibrinolytic activity relates in part to the expression of PAI-l and PAI-2 and in part to antiplas mins. Our data indicate that HPMC are rich in both PAls, that PAI-2 is largely cell associated, and that large quantities of PAI-l are released by HPMC. Our data also implicate HPMC as an important source of PAI-2, PAl-I, and antiplasmins (other than a-2-antiplasmin) in pleural effusions, since these inhibitors are found in high concentrations in these fluids (9). Others investigated the fibrinolytic potential of mesothelial cells elsewhere in the body. Human omental mesothelial cells express tPA (but not uPA), PAl-I, and PAI-2 (11) . By contrast, we found that HPMC contained mRNA for uPA and that TGF-{3 or TNF-a induced expression of the protein. By fibrin enzymography, we also demonstrated that HPMC elaborate tPA, which , by virtue of its migration at M, 97,500, appears to represent tPA bound to either PAl-lor PAl-2. The tPA inhibitor complex retains its fibrinolytic capacity in this gel system, as reported by others (27). Donaldson and colleagues reported that rat pleural mesothelial cells expressed fibrinolytic activity in a fibrin plate assay (28, 29) . This observation likely reflects species differences in the elaboration of PAversus PAl, as rat pleural macrophages and mast cells rather than mesothelial cells were found to be important sources of PAl. One previous description of fibrinolytic activity by HPMC relied on an imprint technique in which HPMC were directly apposed to a fibrin indicator gel (30). This method is similar to that of fibrin enzymography and demonstration of HPMC fibrinolytic activity is consistent with our findings. HPMC respond to stimulation by growth factors and cytokines. Epidermal growth factor, platelet-derived growth factor, or TGF-{3l induce DNA synthesis in cultured HPMC, indicating that these cells have growth-regulatory properties similar to connective tissue cells (14). We chose to study the effects of TGF-{3 and TNF-a on HPMC as these cytokines are released by other cell types found in the pleural space and might influence HPMC function. Both of these cytokines are potent modulators whose effects have been studied in a vari-

Idell, Zwieb, Kumar et al.: Pleural Mesothelial Cell Fibrin Turnover

ety of other cell types. We were able to demonstrate TNF-a in pleural fluids, and lGF-t3, as well as TNF-a, likely appears in the pleural space under conditions of inflammation. lGF-t3 is released by lung fibroblasts, which may enter the pleural space, and lung tissues are rich in lGF-t3 (31, 32). TNF-a, as well as lGF-t3, derives from macrophages, which are also found in the pleural space. TNF-a induces angiogenesis, and it appears to playa central role in acute respiratory failure associated with sepsis (31, 33). The effects of these cytokines on HPMC and fibrin turnover have not, to our knowledge, been studied previously. Both lGF-t3 and TNF-a increased levels of PAI-l in HPMC as well as the amount released into conditioned media. PAI-2 was mainly cell associated, but release of this inhibitor was also increased by TNF-a and lGF-t3. Production of HPMC PAI-l and PAI-2 was inhibited by cyclohexamide and actinomycin D, indicating that protein synthesis and transcriptional regulation are required for expression of these PAL tPA release by HPMC was also increased by lGF-t3 or TNF-a. The changes in HPMC fibrin turnover induced by TGF-t3 or TNF-a did not change the overall balance of PA versus PAl, since HPMC (or conditioned media) fibrinolytic activity, as measured by the fibrin plate assay, remained suppressed. While the PCR analysis is affected by the efficiency of the polymerization and amplification of the signal, qualitative differences may be detected using this technique. HPMC mRNA for PAI-l and PAI-2 analyzed at the same treatment intervals did not appear changed after exposure to identical concentrations oflGF-t3, but mRNA for tPA was increased. Although uPA was not detectable in our gel systems, HPMC contain mRNA for uPA which is likewise inducible by exposure to this cytokine. uPA may be expressed by unstimulated HPMC at levels beneath the sensitivity of our assay systems. The exact relationship between cytokine stimulation, message for tPA, uPA, PAl-I, and PAI-2, and the expression of these proteins by HPMC will only be determined by quantitative, parallel time course studies. In summary, cultured HPMC express procoagulant activity due to tissue factor that binds factor VII at the cell surface to initiate coagulation. The prothrombinase complex assembled at the HPMC surface can potentiate pericellular coagulation. HPMC contain tPA and uPA, but no fibrinolytic activity is expressed by these cells in a fibrin plate assay. The elaboration and release of large quantities of PAI-l and relatively smaller concentrations of PAI-2 by HPMC in part explains these findings. HPMC fibrinolytic activity is also inhibited by antiplasmin activity. Our data implicate the mesothelium as the source of the high concentrations of PAI-l (and relatively lower concentrations of PAI-2) found in pleural fluids. Increased expression of PAI-l as well as release of PAI-2, tPA, and uPA is induced by stimulation of HPMC by lGF-t3 or TNF-a while antiplasmin activity is maintained. Thus, HPMC, upon exposure to appropriate coagulation zymogens, are able to initiate and potentiate fibrin formation while pathways of fibrin clearance are suppressed by PAIs induced by proinflammatory cytokines. Antiplasmins also contribute to suppression of fibrinolysis. Taken together, our observations indicate a dynamic role for

425

the mesothelium in control of fibrin deposition within the pleural space. Acknowledgments: The writers wish to thank the members of the Pulmonary Division of the University of Texas Health Science Center at Tyler for' assisting with the collection of pleural fluids for mesothelial cell culture and Ms. Kim Chittenden for expert technical assistance. This work was supported by NIH HL-37770 and HL-45018 and the Gina Sabatasse Research Grant Award.

References 1. Bignon, J., and J. B. L. Gee. 1985. Pleural fibrogenesis in the pleura in health and disease. In The Pleura in Health and Disease. J. Chretien, J. Bignon, and A. Hirsch, editors. Marcel Dekker, New York. 417-443. 2. Mitchell, M. E., W. M. Alberts, K. W. Chandler, and A. L. Goldman. 1989. Intrapleural streptokinase in management of parapneumonic effusions. J. Fla. Med. Assoc. 76:1019-1022. 3. Aye, R. W., D. P. Froese, and L. D. Hill. 1991. Use of purified streptokinase in empyema and hemothorax. Am. J. Surg. 161:560-562. 4. Sueshi, K., S. Nannon, and K. Tanaka. 1981. Permeability enhancing and chemotactic activities of lower molecular weight degradation products of human fibrinogen. Thromb. Haemost. 45:90-94. 5. Dvorak, H. F. 1982. Tumors: wounds that do not heal. N. Engl. J. Med. 315:1650-1659. 6. Tillet, W. S., S. Sherry, and C. T. Read. 1951. Use of streptokinasestreptodornase in treatment of post-pneumonic empyema. J. Thorac. Surg. 21:275-297. 7. Bergh, N. P., R. Ekroth, S. Larsson, and P. Nagy. 1977. Intrapleural streptokinase in the treatment of hemothorax and empyema. Scand. J. Thorac. Cardiovasc. Surg. 11:265-268. 8. Dumire, R, M. M. Crabbe, F. G. Mappin, and L. J. Fontenelle. 1992. Autologous "blood patch" pleurodesis for persistent pulmonary air leak. Chest 101:64-66. 9. Idell, S., W. Girard, K. B. Koenig, J. McLarty, and D. S. Fair. 1991. Abnormalities of pathways of fibrin turnover in the human pleural space. Am. Rev. Respir. Dis. 144: 187-194. 10. Ciclia, G. T., T. M. O'Connell, W. C. Hahn, andJ. G, Rheinwald. 1989. Cloned eDNA sequence for the human mesothelial protein "mesocrin" discloses its identity as a plasminogen activator inhibitor (PAl -1) and a recent evolutionary change in transcript processing. J. Cell Sci. 94: 1-10. 11. van Hinsbergh, V. W. M., T. Kooistra, M. A. Scheffer, J. Hago van Bockel, and G. N. P, van Muigen. 1990. Characterization and fibrinolytic properties of human omental tissue mesothelial cells, comparison with endothelial cells. Blood 75:1490-1497. 12. Wu, Y.-J., L. M. Parker, N. E. Binder et al. 1982. The mesothelial keratins: a new family of cytoskeletal proteins identified in cultured mesothelial cells and non-keratinizing epithelia. Cell 31:693-703. 13. Niedbala, M. J., K. Crickard, andR. J. Bernacki. 1986. Adhesion, growth and morphology of human mesothelial cells on extracellular matrix. J. Cell Sci. 85: 133-147. 14. Gabrielson, E. W., B. I. Gerwin, C. C. Harris, A. B. Roberts, M. B. Sporn, and J. F. Lechner. 1988. Stimulation of DNA synthesis in cultured primary human mesothelial cells by specific growth factors. FASEB J. 2:2717-2721. 15. Hsu, S.-M., P.-L. Hsu, X. Zhao, C. S. Kao-Shan, and J. Whang-peng. 1988. Establishment of human mesothelioma cell lines (MS-l, -2) and production of a monoclonal antibody (anti-MS) with diagnostic and therapeutic potential. Cancer Res. 48:5228-5236. 16. Levin, E. G. 1986. Quantitation and properties of the active and latent plasminogen activator inhibitors in cultures of human endothelial cells. Blood 67: 1309-1313. 17. Idell, S., K. K. James, E. Levin et al. 1989. Local abnormalities in coagulation and fibrinolytic pathways predispose to alveolar fibrin deposition in the adult respiratory distress syndrome. J. Clin. Invest. 84:695-705. 18. Springers, E. D., H. M. G. Princen, T. Kooistra, and V. W. M. van Hinsbergh. 1985. Inhibition of plasminogen activators by conditioned medium from human hepatocytes and hepatoma cell line Hep G2. J. Lab. Clin. Med. 105:751-758. 19. Rennard, S. I., M.-C. Jaurand, J. Bignon et al. 1984. Role of pleural mesothelial cells in the production of the submesothelial connective tissue matrix of lung. Am. Rev. Respir. Dis. 130:267-274. 20. Harvey, W., and P. L. AmIot. 1983. Collagen production by human mesothelial cells in vitro. J. PathoZ. 139:337-347. 21. Aalto, M., E. Kulonen, R. Penttinen, and S. Renvall. 1981. Collagen synthesis in cultured mesothelial cells. Acta Chir. Scand. 147:1-6. 22. Antony, V. B., C. L. Owen, and K. J. Hadley. 1989. Pleural mesothelial cells stimulated by asbestos release chemotactic activity for neutrophils in vitro. Am. Rev. Respir. Dis. 139: 199-206.

426

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 7 1992

23. Kuwahara, M., M. Kuwahara, K. E. Bijwaard, D. M. Gersten, C. A. Diglio, and E. Kagan. 1991. Mesothelial cells produce a chemoattractant for lung fibroblasts: role of fibronectin. Am. J. Respir. Cell Mol. Biol. 5:256-264. 24. Light, R. W. 1990. Parapneumonic effusions. In Pleural Diseases. 2nd ed. Lea & Febiger, Philadelphia. 129-149. 25. Agrenius, V., J. Chmielewska, O. Widstrom, and M. Blomback. 1989. Pleural fibrinolytic activity is decreased in inflammation as demonstrated in quinacrine pleurodesis treatment of malignant pleural effusion. Am. Rev. Respir. Dis. 140:1381-1385. 26. Good, J. T., D. A. Taryle, T. M. Hyers, and S. Sahn. 1978. Clotting and fibrinolytic activity of pleural fluid in a model of pleural adhesions. Am. Rev. Respir. Dis. 118:903-908. 27. Vassalli, J. D., J. M. Dayer, A. Wohlwend, and D. Belin. 1984. Concomitant secretion of prourokinase and of a plasminogen activator-specific inhibitor by cultured human monocytes-macrophages. J. Exp. Med. 159:1653-1668. 28. Li, X. Y., G. M. Brown, D. Lamb, and K. Donaldson. 1990. Secretion

29. 30. 31. 32. 33. 34.

of plasminogen activator inhibitor by normal rat pleural leukocytes in culture. Lung 168:309-322. Donaldson, K., G. B. Brown, R. E. Bolton, and J. M. G. Davis. 1988. Fibrinolysis by rat mesothelial cells in vitro: the effect of mineral dusts at non-toxic doses. Br. J. Exp. Pathol. 69:487-494. Whitaker, D., J. M. Papadimitriou, and M. N. I. Walters. 1982. The mesothelium: its fibrinolytic properties. J. Pathol. 136:291-299. Kelly,J.1990. Cytokinesofthelung.Am. Rev. Respir. Dis. 141:765-788. Kelly, J., J. P. Fabisiak, K. Hawes, andM. Absher. 1991. Cytokinesignaling in the lung: transforming growth factor-S secretion by lung fibroblasts. Am. J. Physiol. 260(Lung Cell. Mol. Physiol. 4):L123-L128. Tracey, K. J., S. F. Lowry, and A. Cerami. 1988. Cachectin/TNF-a in septic shock and septic adult respiratory distress syndrome. Am. Rev. Respir. Dis. 138:1377-1379. Chirgwin, J. M., A. E. Pryzbyla, R. J. McDonald, and W. J. Rutter. 1979. Isolation of biologically active ribonucleic acids from sources enriched in ribonucleases. Biochemistry 18:5294-5299.