INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pleural effusions (PE) are defined as accumulations of free liquid in the pleural space caused by increased production or decreased clearance of pleural fluid (1). Several criteria are routinely used to distinguish between exudative PE and transudative PE, in general by analyzing the protein and lactate dehydrogenase (LDH) content in serum and pleural fluid (4- 6). During the development of pleural exudates, inflammatory and fibrinolytic mechanisms have been shown to contribute to the accumulation of fluid (7). The inflammatory processes associated with the development of pleural effusions are characterized by an influx of cells into the pleural space and by secretion of inflammatory mediators into the pleural fluid by resident cells (1, 8, 9). In this context, proteolytic and matrix-degrading enzymes have recently been shown to be an essential feature of inflammation (9, 10). Proteolytic processes within the pleural space may alter the integrity of the mesothelial cell layer and/or the underlying basement membrane, and therefore facilitate fluid influx into the pleural space. In this respect, several reports have underlined the importance of a fine balance between proteolytic and antiproteolytic enzymes during the pathogenesis of pleural disease (9).

Among the group of proteolytic enzymes, the matrix metalloproteinases (MMP) constitute a family of endopeptidases primarily responsible for the degradation and turnover of extracellular matrix (ECM) (12). Their proteolytic activity is targeted to all components of the ECM, such as collagens, elastin, laminin, and proteoglycans. In the lung, MMP are involved in the physiologic turnover of the ECM, which occurs at rates of > 10% per day of total collagen mass (17, 18). However, imbalances between the physiologic expression pattern of MMP and TIMP are observed in diseases such as lung fibrosis, lung cancer, and adult respiratory distress syndrome (ARDS) (15). In these diseases, increased proteolytic activity has been shown to be associated with tumor invasiveness and altered tissue architecture.

The MMP family consists of 14 different isoforms that are highly conserved across mammalian species, and which share significant structural homologies (12). However, each MMP isoform exhibits distinct substrate specificities in vitro and in vivo. Active MMP are generated from their secreted precursors by removal of a propeptide sequence (12, 19). Proteolytic activity of MMP is physiologically counterbalanced by specific endogenous inhibitors, the tissue inhibitors of metalloproteases (TIMP) (12). Three distinct TIMP isoforms have been characterized, all inhibiting corresponding members of the MMP family by a stoichiometric 1:1 inhibition. In vivo, the processes controlling ECM production and degradation are therefore tightly regulated by the balanced expression of activated MMP and free TIMP (12).

A previous study by Hurewitz and colleagues demonstrated gelatinolytic activity attributable to the presence of gelatinase-A and gelatinase-B in 32 pleural effusions (11). The aim of this study was to characterize the distinct expression pattern of four MMP and two TIMP isoforms in 88 pleural effusions of different origins. The expression and activity of MMP and TIMP isoforms were correlated with the etiology of the effusions, and the possible impact of certain isoforms on the pathogenesis of pleural diseases was evaluated.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials and Reagents

Calcium chloride, ethylenediaminetetraacetic acid (EDTA), gelatine, glycerol, phenylmethylsulfonyl fluoride (PMSF), sodium chloride, Tris-HCl, Triton X-100, and Tween-20 were supplied by Sigma Chemicals, Emmenbrücke, Switzerland. Brij-35, Pefabloc, and peroxidase-coupled secondary antibodies were from Boehringer Mannheim, Rotkreuz, Switzerland. Gradient 4% to 15% sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels were purchased from BIO-RAD, Zürich, Switzerland. Hybond-ECL nitrocellulose membranes were from Amersham, Zürich, Switzerland. Polyclonal antibodies to human interstitial collagenase, gelatinase-A, and gelatinase-B, and purified human interstitial collagenase, gelatinase-A, and gelatinase-B were obtained from ANAWA Trading SA, Wangen, Switzerland. Monoclonal antibodies to human neutrophil collagenase, TIMP-1, TIMP-2, and recombinant human TIMP-1 and TIMP-2 were purchased from Oncogene Science, Paris, France. Gel drying film was obtained from Promega Corporation, Zürich, Switzerland. Supernatants of activated human peripheral blood mononuclear cells (PBMC), as a positive control for neutrophil collagenase, were kindly provided by Dr. E. Lach of the Department of Research, University Hospital, Basel, Switzerland.

Study Group

Pleural effusion samples from 88 consecutive patients (40 males and 48 females; aged 22 to 93 yr [mean age: 67.3 ± 15.4 yr]) undergoing thoracentesis for diagnostic or therapeutic reasons were analyzed. Pleural fluid was designated as exudate (n = 67) or transudate (n = 21) according to Light's criteria as described in the literature (3). Pleural fluid was considered to be an exudate if one of the following two criteria was met: (1) the pleural fluid protein concentration divided by the serum protein concentration was greater than 0.5; or (2) the pleural fluid LDH concentration was greater than 280 IU/L, the concentration that represents two-thirds the upper limit of normal serum LDH and serves as a cutoff value distinguishing exudates from transudates. Exudates were subdivided into groups of paraneoplastic (n = 46) and parainfectious (n = 21) origin (for group characteristics see Table 1). Paraneoplastic exudates either contained malignant cells or occurred in combination with previously diagnosed malignant tumors or metastases. Parainfectious exudates developed in association with pneumonia, without evidence of malignancies.

Gelatine Zymography

Pleural fluid obtained by thoracentesis was centrifuged immediately at 3,500 × g for 20 min at 4° C. Aliquots were analyzed directly or stored at 70° C until used. MMP activity in each sample was determined through zymographic analysis under denaturing but nonreducing conditions, as previously described (20, 21). In brief, each sample was diluted 1:10, and 5-µl aliquots were applied to a denaturing, 8% SDS-polyacrylamide gel containing 0.1% gelatine. Electrophoresis was performed at a constant current of 25 mA for 2 h at room temperature (RT), followed by equilibration in twice-distilled water containing 2.5% Triton X-100 for 1 h in order to remove SDS. The gel was then incubated in enzyme buffer containing 50 mM Tris-HCl (pH 7.3), 200 mM NaCl, 5 mM CaCl₂, and 0.02% Brij-35 for 18 h at 37° C. Bands of enzymatic activity were visualized by negative staining with standard Coomassie brilliant blue dye solution. Molecular sizes of bands displaying enzymatic activity were identified by comparison with prestained standard proteins, as well as with purified gelatinase-A or gelatinase-B. The character of proteolytic bands was analyzed by incubating identical zymograms in: (1) 10 mM EDTA, a selective MMP inhibitor; (2) 0.1 mg/ml PMSF, a serine protease inhibitor; or, (3) 2 mM Pefabloc, an irreversible serine protease inhibitor (20, 21).

Western Blot Analysis

Expression of different MMP and TIMP isoforms was determined at the protein level by Western blot analysis, using gradient 4% to 15% SDS-PAGE gels. Aliquots of each sample were applied to the gels and size-fractionated by electrophoresis according to a modified method of Laemmli (22). Proteins were electroblotted on Hybond-ECL nitrocellulose membranes for 90 min at 1 mA/cm². Membranes were blocked in 5% skim milk in Tris-buffered saline (TBS)-Tween (10 mM Tris, 150 mM NaCl, 0.05% Tween 20, pH 8.0) for 1 h at room temperature (RT). After blocking, membranes were incubated at 4° C overnight with the specific antibodies to interstitial collagenase, neutrophil collagenase (at a titer of 1:500 each), gelatinase-A, gelatinase-B (1:1,000 each), TIMP-1, or TIMP-2 (1:200 each). After three washes, membranes were incubated with the secondary peroxidase-coupled antibody at a dilution of 1:5,000 for 1 h at RT. Membranes were washed three times in TBS-Tween, and specific bands were visualized using the ECL system from Amersham according to the manufacturer's instructions.

Analysis of Gelatinase-B/Gelatinase-A Ratio

Zymographies of transudates and exudates were processed identically, fixed in 1% glycerol between two sheets of gel drying film, and dried overnight at RT. Fixed zymograms were then scanned on a digital camera setup, and intensities of gelatinolytic bands due to the presence of gelatinase-A and gelatinase-B were quantified on a disk operating system (DOS)-based image analysis system installed by Raytest, Straubenhardt, Germany. Ratios of gelatinase-B to gelatinase-A activity were then plotted according to the origin of the pleural fluid, and the significance of the results was analyzed with the Mann-Whitney U-test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Gelatinolytic Activity in Pleural Effusions

To investigate gelatinolytic activity present in pleural effusions, substrate gel zymography was performed. This method allows the detection of the two metalloproteases that exhibit significant gelatinolytic activity (gelatinase-A and -B). As shown in Figure 1a, zymographic analyses revealed the existence of two major bands of gelatinolytic activity in pleural effusions, migrating at approximately 90 kDa and 70 kDa, respectively. In the presence of EDTA, a selective inhibitor of MMP, these gelatinolytic bands disappeared (Figure 1c). In contrast, bands of gelatinolytic activity were unaffected by the addition to the enzyme buffer of PMSF, an inhibitor of serine proteases (Figure 1b), or Pefabloc, an irreversible serine protease inhibitor (data not shown). This inhibition profile clearly attributes gelatinolytic activity in pleural effusions to the presence of MMP. Comparison of these two gelatinolytic bands with prestained standard proteins and with purified gelatinase-A (MMP-2) and gelatinase-B (MMP-9) clearly identified the two MMPs constituting the bands as gelatinase-A (72 kDA) and gelatinase-B (92 kDa) (Figure 1a and b).

View larger version (92K): [in this window] [in a new window]   Figure 1.   Three zymograms were loaded and processed identically, as described in METHODS. To characterize the nature of gelatinolytic bands, gels were developed in normal enzyme buffer (a), enzyme buffer + 5 mM PMSF (b), or enzyme buffer + 5 mM EDTA (c). Sizes of prestained molecular-weight markers are indicated on the right. Lane 1: purified gelatinase-A (MMP-2); Lane 2: representative transudate; Lanes 3 through 5: representative paraneoplastic exudates; Lanes 6 through 8: representative parainfectious exudates; Lane 9: purified gelatinase-B (MMP-9).

Gelatinase-A was detected in all samples analyzed, and its amount in pleural effusions did not seem to vary among the three groups of pleural effusions (Figure 2). In contrast, expression of gelatinase-B demonstrated a high variability among the three groups of pleural effusions. In transudates, clear expression of gelatinase-B was observed in only four of 21 samples (19%) (Figure 2a). In exudates, gelatinase-B was detected in 19 of 21 samples of parainfectious origin (90%) (Figure 2b) and in 43 of 46 samples of paraneoplastic origin (93%) (Figure 2c). However, gelatinase-B seemed to be expressed at higher amounts in paraneoplastic than in parainfectious exudates (Figure 2b and c). Only single bands of gelatinase-B, at 90 kDa, were observed in all samples investigated, which indicated that the activated, lower-molecular-weight form of gelatinase-B (86 kDa) was not present in pleural effusions. In contrast, the activated, lower molecular weight form of gelatinase-A (66 kDa) was detected as a double band in paraneoplastic and parainfectious exudates, but not in transudates (Figures 1a and 2).

View larger version (93K): [in this window] [in a new window]   Figure 2.   Zymograms were prepared with samples from representative transudates (a), representative parainfectious exudates (b), or representative paraneoplastic exudates (c). Gels were loaded and processed under the same conditions as described in METHODS. Sizes of prestained molecular-weight markers are indicated on the right. Lane 1: purified gelatinase-B (MMP-9); Lanes 2 through 7: six representative transudates (a); six representative parainfectious exudates (b); or six representative paraneoplastic exudates (c); Lane 9: purified gelatinase-A (MMP-2).

Immunoblotting

To investigate the expression of MMP isoforms that cannot be detected by zymographic analysis, immunoblotting under reducing conditions was performed with antibodies specific for gelatinase-A, gelatinase-B, interstitial collagenase, or neutrophil collagenase.

As shown in Figure 3, immunoblotting confirmed the identity and expression pattern of gelatinase-A (Figure 3a) and gelatinase-B (Figure 3b). In addition, interstitial collagenase, migrating at the same size as the latent form of purified interstitial collagenase (55 kDa; Figure 3a), exhibited a constant pattern of expression. This expression of interstitial collagenase did not differ in effusions of different origins. In addition, low expression levels of neutrophil collagenase were detected in most samples analyzed, and correlated with the number of neutrophils present in the effusions (Figure 3b).

View larger version (61K): [in this window] [in a new window]   View larger version (43K): [in this window] [in a new window]   Figure 3.   Aliquots of pleural effusions were applied to 4% to 15% gradient SDS-PAGE gels, run at 25 mA for 90 min, and electroblotted onto nitrocellulose membranes. Filters were then hybridized with antibodies specific for human interstitial collagenase (MMP-1, 1:500) and gelatinase-A (MMP-2, 1:1,000) (a), or with antibodies specific for human neutrophil collagenase (MMP-8, 1:500) and human gelatinase-B (MMP-9, 1:1,000) (b). Lane 1: controls of purified MMP as indicated; Lanes 2 and 3: transudates; Lanes 4 through 6: paraneoplastic exudates; Lanes 7 and 8: parainfectious exudates.

Interestingly, the uncleaved form of gelatinase-A was clearly present in all effusions, independent of the origin of the effusion. However, only in parainfectious and, to a greater extent, in paraneoplastic exudates, did a lower-molecular-weight form of gelatinase-A appear at a size of about 20 kDa (Figure 3a). This band migrated at the same size under reducing and nonreducing conditions (data not shown). It thus seemed to be a free proteolytic fragment of gelatinase-A rather than a fragment of gelatinase-A complexed to TIMP-isoforms and migrating at the same size as TIMP-1 or TIMP-2 (20 kDa).

The biologic impact of expressed MMP isoforms may be interpreted only if data on the coexpression of their endogenous inhibitors, TIMP, are available. We therefore screened for the expression of the two TIMP isoforms TIMP-1 and TIMP-2 in pleural effusions by immunoblotting. As shown in Figure 4, TIMP-1 was highly expressed in all pleural effusions that we analyzed. This high expression level of TIMP-1 was similar in all effusions, and therefore seemed to be constitutive. In contrast, TIMP-2 protein was barely detectable in pleural effusions as compared with TIMP-1 (Figure 4). However, the amount of TIMP-2 clearly varied among different effusions. It was higher in parainfectious exudates than in either paraneoplastic exudates or transudates. In the latter, TIMP-2 was almost undetectable with the methods employed.

View larger version (40K): [in this window] [in a new window]   Figure 4.   Samples were applied to 4% to 15% gradient SDS-PAGE gels, run at 25 mA for 90 min, and electroblotted onto nitrocellulose membranes. Filters were then hybridized with antibodies specific for human TIMP-1 or TIMP-2 as indicated, at a titer of 1:200 each. Lane 1: recombinant human TIMP-1 or TIMP-2 as indicated; Lanes 2 through 5: transudates; Lanes 6 through 9: parainfectious exudates; Lanes 10 through 13: paraneoplastic exudates. Molecular weights are indicated on the right.

Correlation of Gelatinase-B/Gelatinase-A Expression Ratio to the Origin of the Effusion

Gelatine zymography and immunoblotting revealed that expression of gelatinase-A was similar in pleural effusions of different origin, but that expression of gelatinase-B varied significantly. Using a computerized image analysis system, we determined the ratio of the expression of gelatinase-B to gelatinase-A. As shown in Figure 5, the expression ratio of these two MMP isoforms revealed a characteristic pattern when transudates, parainfectious exudates, and paraneoplastic exudates were compared. The ratios determined in 21 transudates were clearly the lowest (0.01 to 1.6; median: 0.12), whereas paraneoplastic exudates revealed the highest gelatinase-B/gelatinase-A ratio (0.05 to 7.17; median: 1.39) (p < 0.001). Parainfectious exudates (0.05 to 8.72; median: 0.59) also demonstrated a significantly higher gelatinase-B/gelatinase-A ratio than did transudates (p < 0.001), but a lower ratio than did paraneoplastic exudates (p < 0.05).

View larger version (15K): [in this window] [in a new window]   Figure 5.   Box-and-whiskers plot of gelatinase-B versus gelatinase-A expression. Samples were identically processed in zymographic analyses, as described in METHODS. Bands of gelatinase-B and -A activity were then scanned on a DOS-based image analysis system with integrated background substraction. Ratios of activity of gelatinase-B to gelatinase-A were then plotted according to the origin of the effusion (transudates, n = 21; parainfectious exudates, n = 21; paraneoplastic exudates, n = 46). The box content includes values between the 25th and 75th percentiles; the dark line within the box represents the median. The whiskers represent all values except outliers. Significances were analyzed with the Mann-Whitney U test (*p < 0.001, +p < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study described here characterized the presence of MMP and TIMP isoforms in 88 pleural effusions of different origin. It demonstrated the constitutive expression of interstitial collagenase, gelatinase-A, and TIMP-1 in all pleural effusions, whereas neutrophil collagenase and TIMP-2 could only be detected in trace amounts. Gelatinase-B was found mainly in exudates, and was absent in 81% of transudates. The ratio of the expression of gelatinase-B to gelatinase-A was correlated with the origin of pleural effusions, and showed a significant difference among transudates, parainfectious exudates, and paraneoplastic exudates.

In vivo, the expression of proteases and antiproteases, such as MMP and TIMP, represents a highly efficient system that is used by a variety of cell types to degrade and remodel the ECM (12). Physiologically, the expression of MMP and TIMP isoforms is tightly coordinated and thereby responsible for tissue integrity. As seen in the case of severe inflammation or metastasized tumors, an imbalance between MMP and TIMP enzymes dramatically affects normal tissue architecture (12). Increased expression of MMP-isoforms has been associated with the disrupted ECM structure seen in bronchiectasis (23, 24), emphysema (25), and cystic fibrosis (26), thus emphasizing the biologic significance of these enzymes in the lung. In addition, the increased alveolocapillary permeability seen in acute lung injury or ARDS is accompanied by an increase in MMP expression (27, 28).

A previous study by Hurewitz and colleagues (11) demonstrated the presence of two MMP isoforms, gelatinase-A and gelatinase-B, in 32 pleural effusions, but could not correlate their pattern of expression with the origin of pleural disease. As in the case of our results, this study revealed a high expression of gelatinase-A in pleural effusions. Strong expression of gelatinase-A was not found in the patients' sera, and was thus thought to be the result of local secretion of gelatinase-A by resident cells (11, 29). In this respect, our data indicate that the expression of gelatinase-A is a constitutive feature of resident pleural mesothelial cells in vivo, since we found high levels of gelatinase-A in all 88 samples investigated. In this regard, it is of special interest that pleural mesothelial cells constitutively express gelatinase-A in vitro (30). Marshall and colleagues showed that neither lipopolysaccharide (LPS) nor phorbol myristate acetate (PMA) affected the constitutive expression of gelatinase-A by pleural mesothelial cells (30). In contrast, gelatinase-B was not present in media of unstimulated cells, but was rapidly upregulated by stimuli such as PMA (30).

These in vitro observations strongly support our hypothesis that gelatinase-A is constitutively expressed in vivo in effusions of different origin, even in transudates that are commonly considered as filtrates of blood plasma. In contrast, gelatinase-B expression is significantly upregulated in exudates of parainfectious and paraneoplastic origin. Since neither gelatinase-A nor gelatinase-B or TIMP-1 expression is correlated with cell numbers in pleural effusions (data not shown), resident mesothelial cells might be primarily responsible for the secretion of MMP and TIMP isoforms into the pleural space in vivo. Interestingly, this pattern of MMP expression is clearly different from the pattern we observed in bronchoalveolar lavage fluid (BALF) from lung cancer and control patients without pulmonary diseases. In BALF, gelatinase-B is the constitutive isoform expressed in all patients, and gelatinase-A is induced only in lung cancer patients (31).

TIMP isoforms have thus far not been demonstrated in pleural effusions; however, it has been shown that pleural mesothelial cells are capable of secreting significant amounts of TIMP into culture media in vitro (30). The available data obtained from lung tissue and BALF support the idea of a high constitutive expression of TIMP-1 in vivo (26, 28). In lung tissue of baboons, Minoo and coworkers demonstrated high expression levels of TIMP-1 only after birth, thus suggesting that TIMP-1 expression is an essential feature of postnatal adaptation of the lung (32).

Our studies extend these observations and show that TIMP-1 is also present in high amounts in the pleural space, whereas TIMP-2 could be detected only in trace amounts in exudates, and not at all in transudates. This observation is especially interesting in that TIMP-2 is considered to form complexes with gelatinase-A, whereas TIMP-1 is thought to complex gelatinase-B (12). Because this complex formation inhibits MMP proteolytic activity, the lack of TIMP-2 in cases of high gelatinase-A expression may essentially augment proteolytic activity within pleural effusions. In addition, TIMP-1 may be of value in preserving homeostasis in the pulmonary and pleural compartment, and along with gelatinase-A and interstitial collagenase may essentially contribute to the integrity of the pleural space. Thus, MMP are probably important in preventing fibrin clots and the development of adhesions at sites of focal pleural injury.

In summary, our data suggest that interstitial collagenase, gelatinase-A, and TIMP-1 are constitutively expressed in pleural fluid of any origin. Ratios of gelatinase-B to gelatinase-A may be useful in determining the cause of the fluid, but further studies are required to define the spectrum of diseases associated with the various isoforms of matrix-degrading enzymes and their inhibitors.