Endogenous Fibronectin of Blood Polymorphonuclear Leukocytes: Stimulus-Induced Secretion and Proteolysis by Cell Surface-Bound Elastase

EXPERIMENTAL CELL RESEARCH ARTICLE NO.

233, 33–40 (1997)

EX973545

Endogenous Fibronectin of Blood Polymorphonuclear Leukocytes: Stimulus-Induced Secretion and Proteolysis by Cell Surface-Bound Elastase Rosalba Salcedo, Ken Wasserman, and Manuel Patarroyo1 Microbiology and Tumorbiology Center, Karolinska Institute, S 171 77 Stockholm, Sweden

a central role against infections and to mediate tissue injury [1]. They are highly specialized cells equipped with a battery of enzymes and other proteins which are stored in secretory granules [2]. During the acute inflammatory response, blood PMNs attach to the endothelial lining of postcapillary venules, move between the endothelial cells, and migrate through tissues toward the focus of inflammation [3]. Stimulated by chemoattractants, the PMNs rapidly undergo a variety of physiological responses such as oxidative metabolism, degranulation, and adhesion [1–3]. During the extravasation process, PMNs interact physically with the surface of other cell types and with extracellular matrix proteins such as collagens, laminin, and fibronectin [3, 4]. These adhesive interactions are largely mediated by integrin receptors [3, 5–7]. Fibronectin is a large adhesive protein able to interact with cells and other matrix components such as collagens [reviewed in 8]. Many cell types, including platelets [9], are able to secrete fibronectin. This molecule, which has a key role in cellular migration, is a common component of extracellular matrices but is also found in plasma and other body fluids. It is composed of two similar subunits covalently linked near their COOH-termini by disulfide bonds. Each subunit is made of different types of internal repeats designated I, II, and III. The a5b1 integrin, which is expressed on PMNs [10], recognizes the RGD sequence in the 10th type III repeat of the central cell-binding domain [11]. Other integrins, such as a4b1 and aVb3, also recognize fibronectin [12]. Exogenous fibronectin or its fragments have been reported to enhance chemotaxis, phagocytosis, degranulation, and oxidative metabolism in PMNs [13–16], and to mediate chemotaxis of monocytes [17], cells which follow PMNs during extravasation [18]. Moreover, blood PMNs adhere to plasma fibronectin by using the a5b1 integrin [10]. In an accompanying paper [19], we have characterized the endogenous fibronectin of blood PMNs. Intact fibronectin was found in the specific granules of resting cells. During the 1980s, blood PMN-secreted material reactive with anti-fibronectin antisera was reported

In an accompanying study, we described the presence of intact fibronectin, a large adhesive molecule, in the specific granules of blood PMNs. Secretion of fibronectin by blood PMNs is poorly understood, and the fate of this fibronectin is practically unknown. In the present study we demonstrate that nanomolar concentrations of phorbol ester or the chemoattractants fMLP, PAF, and LTB4 induce fibronectin secretion from blood PMNs. Phorbol ester induced secretion of approximately 85% of the total fibronectin content, as well as expression of small amounts on the cell surface of the activated PMNs. Secreted fibronectin was proteolytically cleaved and, after 20 min, four major fragments of 150, 120, 90, and 80 kDa containing a midchain epitope were identified by Western blot analysis. Kinetic studies indicated that fibronectin was rapidly secreted as an intact molecule and that proteolysis started within minutes and proceeded for at least 1 h. If cells were removed after 5 min TPA treatment, no further proteolysis of the secreted fibronectin was observed, indicating participation of cell-bound proteinases. From a cocktail of proteinase inhibitors, PMSF was the most active in suppressing fibronectin proteolysis. Studies with specific peptidyl inhibitors of human leukocyte elastase and cathepsin G, major serine proteinases of PMNs, demonstrated some inhibition with the cathepsin G inhibitor, while the human leukocyte elastase inhibitor almost completely abolished fibronectin proteolysis. A monoclonal antibody to the elastase had a similar effect. The results indicate that intact fibronectin is a secretory product of blood PMNs and that this endogenous adhesive molecule is within minutes extracellularly processed by cell surfacebound elastase. q 1997 Academic Press

INTRODUCTION

PMNs represent approximately 60% of the total circulating leukocytes in humans and are known to play 1 To whom correspondence and reprint requests should be addressed. Fax: 46-8-328878.

33

AID

ECR 3545

/

6i20$$$$81

04-24-97 00:26:41

0014-4827/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.

eca

34

SALCEDO, WASSERMAN, AND PATARROYO

[20–22]. However, the fate of the secreted fibronectin is practically unknown. Since stimulated PMNs adhere to a large variety of both physiological and artificial surfaces [3, 5–7, 20], secreted fibronectin may constitute an endogenous adhesive ligand. In the present study we describe stimulus-dependent secretion of fibronectin and its extracellular proteolysis by cell surface-bound elastase. MATERIAL AND METHODS PMN isolation and stimulation. PMNs were isolated from citrated blood from healthy donors by discontinuous Percoll (Pharmacia, Uppsala, Sweden) gradient centrifugation [19]. More than 98% of the cells were PMNs as determined morphologically and by using CD markers [19]. This cell preparation was practically plateletfree [19]. To analyze fibronectin secretion and proteolysis, PMNs were resuspended at 20 1 106/ml in RPMI 1640 medium (GIBCO, Grand Island, NY) and stimulated at 377C for 20 min, unless otherwise stated, with either 12-O-tetradecanoylphorbol-13-acetate (TPA) (Sigma, St. Louis, MO), formyl-methionyl-leucyl-phenylalanine (fMLP) (Sigma), platelet activating factor (PAF) (Sigma) or leukotriene B4 (LTB4) (kindly provided by Dr. Jesper Haeggstro¨m, Karolinska Institutet, Stockholm), in the absence or the presence of proteinase inhibitors. After stimulation, cell suspensions were centrifuged and the supernatants collected, aliquoted, lyophilized, and subjected to analysis as described below. Antibodies and proteinase inhibitors. Monoclonal antibodies (mAbs) to human fibronectin, the approximate location of their epitopes, and their source are shown in Fig. 5 and in the accompanying paper [19]. mAb to human leukocyte elastase (clone NP57) was obtained from Dakopatts (Copenhagen, Denmark) and used as purified IgG. All mAbs are mouse IgG. Mouse IgG (Coulter, Hialeah, HI) was used as the negative control. Rabbit antibodies (purified IgG) to human fibronectin and the rabbit IgG control were obtained from Dakopatts. Purified human plasma fibronectin was obtained from Sigma. The cocktail of proteinase inhibitors used for biochemical studies contained 2 mM EDTA (Sigma), 2 mM leupeptin (Boehringer Mannheim, Mannheim, Germany), 2 mM pepstatin (Boehringer Mannheim), 1 mM PMSF (Boehringer Mannheim), and 1 mg/ml aprotinin (Boehringer Mannheim). N-Methoxysuccinyl-ala-ala-pro-val chloromethyl ketone (HLE/CMK) (Sigma) and Z-gly-leu-phe chloromethyl ketone (CG/CMK) (Enzyme Systems Products, Sierra LaneDublin, CA) [23] were dissolved in methanol as stock concentrations of 83 and 100 mM, respectively. Immunofluorescence flow cytometry. Indirect immunofluorescence was performed by exposing cells to saturating amounts of rabbit antibodies to human fibronectin. As the second antibody, fluorescein-conjugated F(ab)2 fragments of swine anti-rabbit immunoglobulin (Dakopatts) were used at a 1:20 dilution. After staining, the cells were analyzed in a FACScan flow cytometer (Becton Dickinson, Mountain View, USA). Rabbit IgG was used as the negative control. For cell-surface labeling, PMNs were stimulated for 20 min in medium, centrifuged, and, following removal of the supernatant, washed once with 10 mM EDTA/PBS and then fixed with 3% paraformaldehyde/PBS. Cell permeabilization was performed by treatment with paraformaldehyde (fixation) and then octyl-b-D-glucopyranoside (permeabilization), as previously described [19]. Gel electrophoresis, immunoblotting, and enhanced chemiluminescence. Protein samples were analyzed by SDS–PAGE using 7.5% polyacrylamide under reducing conditions. In Western blots, filters were blocked with 0.05% Tween X-100/5% dry milk in PBS. Peroxidase-linked anti-mouse or anti-rabbit immunoglobulin (Dakopatts) was used as secondary antibodies and ECL (Amersham, UK) as the developer. Filters were exposed to Kodak X-Omat AR film, and, when

AID

ECR 3545

/

6i20$$$$82

04-24-97 00:26:41

FIG. 1. Secretion and proteolysis of blood PMN fibronectin. Western blot analysis of cell lysate from unstimulated PMNs (1), basal secretion (2), and secretion after 20 min TPA of treatment (3). Rabbit anti-fibronectin antibodies (A) and monoclonal antibody FN 30-8 (B) were used for detection. Electrophoresis was run under reducing conditions.

required, the films were laser-scanned in a Personal Densitometer SI and analyzed using ImageQuant software (Molecular Dynamics, Sunnyvale, CA) with background subtraction. Secreted fibronectin was analyzed in the supernatant of stimulated PMNs, following removal of the cells by centrifugation and addition of the proteinase inhibitor cocktail. Total PMN fibronectin was obtained as the soluble fraction of unstimulated PMNs treated with lysis buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 0.02% sodium azide, 1% SDS, 1% Triton X100 plus the proteinase inhibitor cocktail) for 30 min on ice. ELISA. Fibronectin was measured by a sandwich enzymelinked immunoabsorbent assay using mAb FN 30-8 and purified rabbit antibodies to human fibronectin, as described in the accompanying paper [19].

RESULTS

Blood PMNs secrete fibronectin following stimulation. As previously described [19], intact fibronectin chains (230 kDa) were detected in the cell lysate of unstimulated blood PMNs by Western blot analysis (Fig. 1). In the supernatant of unstimulated PMNs, practically no fibronectin was found. However, after stimulation of the cells with 200 nM TPA for 20 min, immunoreactive polypeptides with apparent molecular weights between 160 and 80 kDa were readily detected in the supernatant by both rabbit antibodies and mAb FN 30-8 to fibronectin. By using a sandwich ELISA with these antibodies, the secreted fibronectin was quantified. TPA was a potent inducer of fibronectin secretion (Fig. 2A). The chemoattractants fMLP, PAF, and LTB4 were also able to induce fibronectin secretion, but to a lower extent. Dose–response studies indicated that nanomolar concentrations of the four stimuli were enough to exert maximal fibronectin secretion (Fig. 2A). Most secretion was already achieved 5 min after addition of the stimuli, as demonstrated by kinetic studies (Fig. 2B). Quantitative studies by both

eca

PMN FIBRONECTIN: SECRETION AND ENDOPROTEOLYSIS

35

fibronectin content of blood PMNs is expressed at the cell surface following stimulation. Monoclonal antibodies to fibronectin, including FN 30-8, which reacts with permeabilized cells [19], were unable to react with the cell surface of stimulated PMNs (data not shown). Proteolysis of secreted fibronectin. In contrast to the intracellular form, secreted fibronectin of PMNs was not intact (Fig. 1). After 20 min stimulation with TPA, several fibronectin fragments of 160 to 80 kDa were detected in the supernatant by rabbit antibodies in Western blot analysis. A similar pattern was obtained with mAb FN 30-8 against an epitope localized in the middle of the fibronectin chain, suggesting that this epitope was immunodominant in the rabbit. Four major fragments of 150, 120, 90, and 80 kDa were identified (Fig. 1). Similar fibronectin fragments were obtained when fMLP, PAF, or LTB4 were used as stimuli (data not shown). As an attempt to determine the localization of all fragments, a large panel of monoclonal antibodies to common frame epitopes distributed along the fibronectin molecule were tested (Fig. 5). Although all tested antibodies are able to recognize intact PMN fibronectin by Western blotting [19], only antibodies FN 12-8, FN 30-8, 3E1, 3E3, and 4B2 were clearly reactive with fragments (Fig. 5). These antibodies bind to a wide middle region of the fibronectin chain. Antibod-

FIG. 2. Dose response and kinetics of stimulus-induced PMN fibronectin secretion. (A) Dose response. Different concentrations of TPA, fMLP, PAF, and LTB4 were added to the cells for 20 min at 377C. Mean of 10 experiments for each stimulus. (B) Kinetics. Twenty million cells were stimulated with TPA (200 nM), fMLP (100 nM), PAF (10 nM), and LTB4 (100 nM) up to 1 h at 377C. Mean { SD of seven experiments. In (A) and (B) fibronectin amount was determined in the supernatant by sandwich ELISA.

sandwich ELISA and densitometric analysis of Western blots indicated that TPA induced secretion of most (85%) of the fibronectin content of blood PMNs (Fig. 3). Cell-surface expression of fibronectin following stimulation. Unstimulated blood PMNs do not express fibronectin at the cell surface [19]. By using purified rabbit antibodies and immunofluorescence flow cytometry, small amounts of fibronectin were detected on the cell surface of PMNs following stimulation with TPA, fMLP, PAF, or LTB4 for 20 min (Fig. 4). The average specific fluorescence intensity, 12.3, 9.1, 5.9, and 4.7, respectively, was proportional to the ability of each stimulus to induce secretion. Comparison with the specific fluorescence intensity of permeabilized unstimulated cells, 89.2, indicated that only a minority of the

AID

ECR 3545

/

6i20$$$$83

04-24-97 00:26:41

FIG. 3. Quantification of fibronectin content and secretion by blood PMNs. (A) Determination by sandwich ELISA. Mean of 20 experiments. (B) Densitometric analysis of Western blot from two donors. a,a*, cell lysate; b,b*, supernatant after 20 min TPA (200 nM) stimulation.

eca

36

SALCEDO, WASSERMAN, AND PATARROYO

FIG. 4. Cell surface expression of fibronectin in blood PMN following stimulation with various stimuli. Rabbit anti-fibronectin antibodies were used for flow cytometry. PMNs with no stimulus (rrrrrr), 200 nM TPA ( ), 100 nM fMLP (r r r r), 10 nM PAF (—), and 100 nM LTB4 (---) are shown. Permeabilized PMNs are shown for comparison (r-r-).

ies to NH2- and COOH-terminal regions of the molecule were unreactive. The pattern of reactivity of antibodies FN 12-8 and 3E3 was most similar. Since these antibodies are known to bind the cell-binding domain of fibronectin [24, 25], they recognize most likely the same epitope. Thus, four immunoreactive patterns, defined by antibodies FN 12-8/3E3, FN 30-8, 3E1, and 4B2, with fibronectin fragments could be detected. Fibronectin is secreted as an intact protein and cleaved pericellularly by cell-bound proteinases. Kinetic studies and Western blot analysis demonstrated that fibronectin was first secreted and then cleaved (Fig. 6A). Intact fibronectin was secreted at 1 min. Proteolysis started at 2 min and proceeded for at least 1 h. Practically no intact fibronectin was observed after 1 h, and only fragments of 90, 80, and 30 kDa were detected. Recent studies have demonstrated that certain proteinases of blood PMNs are not secreted, but instead are displayed at the cell surface following cell stimulation [26, 27]. To determine whether a similar mechanism was used by fibronectin-reactive proteinase(s), PMNs were stimulated with TPA for 5 min and thereafter removed by centrifugation. As demonstrated by Western blotting (Fig. 6B), practically no further proteolysis of the secreted fibronectin was observed, indicating participation of cell-bound proteinase(s). Arrest of the proteolysis was similarly observed when the cells were removed after 20 min TPA stimulation (data not shown).

AID

ECR 3545

/

6i20$$$$83

04-24-97 00:26:41

Fibronectin proteolysis is largely mediated by human leukocyte elastase. As a first attempt to identify the mechanism responsible for the proteolysis of fibronectin, PMNs were stimulated with TPA in presence of a cocktail of proteinase inhibitors commonly used in biochemical studies. This cocktail contains EDTA, pepstatin, leupeptin, aprotinin, and PMSF and covers metallo-, aspartate-, thiol-, and serine-proteinases. As detected by Western blotting (Fig. 7A), the cocktail largely inhibited the proteolysis of fibronectin. When the various components were tested separately, the serine proteinase inhibitor PMSF (1 mM) was identified as the active one. To identify the responsible serine proteinase(s), specific peptidyl inhibitors of cathepsin G (CG/CMK) and human leukocyte elastase (HLE/CMK), major serine proteinases of blood PMNs [26, 27], were tested (Fig. 7B). CG/CMK (300 mM) exerted limited inhibition, whereas HLE/CMK (500 mM) almost completely abolished the proteolysis of fibronectin. Even at 5 mM, HLE/CMK was equally effective (data not shown). A purified monoclonal antibody to the elastase had a similar effect (Fig. 7C), confirming the participation of this abundant neutral proteinase and its extracellular localization. DISCUSSION

In the present study, we unambiguously demonstrate that fibronectin is a secretory product of blood

eca

PMN FIBRONECTIN: SECRETION AND ENDOPROTEOLYSIS

FIG. 5. Reactivity of monoclonal antibodies to common frame epitopes with secreted and proteolytically cleaved fibronectin. PMNs were stimulated with 200 nM TPA for 20 min at 377C and the fibronectin fragments were analyzed by Western blotting under reducing conditions. (A) Pattern obtained with rabbit IgG (1), rabbit antihuman fibronectin (2); (B) Pattern obtained with mouse IgG (1), FN 12-8 (2), FN 30-8 (3), 3E1 (4), 3E3 (5), and 4B2 (6); (C) Schematic representation of fibronectin subunit domains and approximated localization of epitopes detected by fragment-reactive antibodies shown in B and nonreactive antibodies (underlined): FN 12-8 (2), FN 30-8 (3), 3E1 (4), 3E3 (5), 4B2 (6), FN 9-1 (7), FN 21-1 (8), 1935 (9), 1934 (10), FN 1-1 (11), IST-1 (12), and IST-2 (13).

PMNs. Our results agree with earlier reports describing detection of immunoreactive material released from blood PMNs by ELISA and immunofluorescence [20– 22]. We also show that, as in platelets [9], fibronectin secretion is stimulus-dependent. Both pharmacological and physiological stimuli were found to be active at nanomolar concentrations. That fibronectin is a secretory product of PMNs is also supported by the fact that, when TPA is used as stimulus, almost all fibronectin content is secreted. Interestingly, small but significant amounts of fibronectin were expressed on the cell surface following stimulation of the PMNs. The exact mechanism responsible for this cell surface expression is currently unknown. It could be due to fusion of the membrane of specific granules with the plasma membrane or, alternatively, binding of secreted fibronectin to the cell surface by integrins, elastase/cathepsin G, or other cellsurface structures of activated PMNs. Significant secretion of endogenous fibronectin and minor expression at the cell surface should be taken into account in functional studies with stimulated blood PMNs, such as

AID

ECR 3545

/

6i20$$$$83

04-24-97 00:26:41

37

adhesion to and migration on exogenous matrix proteins. In vitro, PMNs isolated from synovial fluid of rheumatoid arthritis patients synthesize fibronectin and secrete larger amounts of this protein, compared to blood PMNs, without addition of exogenous stimulus [21]. These ‘‘long-term’’ activated and extravasated cells differ in several respects from blood PMNs. They have a high rate of protein synthesis and express fibronectin at the cell surface [28, 29]. Kinetic studies and the effect of specific proteinase inhibitors indicated that fibronectin was first secreted as an intact molecule and then digested by cell-surface elastase and, to a minor extent, cathepsin G, arguing against intracellular proteolysis. Human leukocyte elastase appeared to exert most of the fibronectin proteolysis. Complementary participation of cathepsin G was indicated by total abolishment of the proteolysis when both HLE/CMK and CG/CMK were present. In contrast to fibronectin, which is localized in the specific granules [19], human leukocyte elastase and cathepsin G are stored in the azurophil granules [2], which are formed earlier during myelopoiesis [30]. These neutral and serine proteinases are abundant in blood PMNs. One million cells contain approximately 1 mg of each enzyme [27], compared to 8 ng of fibronectin [19]. Inter-

FIG. 6. Kinetics of endogenous fibronectin proteolysis in the presence or absence of PMNs. (A) In the presence of PMNs. Blood PMNs were stimulated with 200 nM TPA at 377C. Supernatants were collected at different time points (min) and analyzed by Western blotting using rabbit anti-human fibronectin. (B) In the absence of PMNs. Blood PMNs were stimulated with 200 nM TPA at 377C for 5 min. After the cells were removed by centrifugation, the supernatants were collected at different time points (min) and analyzed by Western blotting using rabbit anti-human fibronectin.

eca

38

SALCEDO, WASSERMAN, AND PATARROYO

FIG. 7. Effect of proteinase inhibitors on fibronectin endoproteolysis. Blood PMNs were stimulated with 200 nM TPA for 20 min at 377C in the presence or absence of inhibitors. Supernatants were analyzed by Western blotting using antibody FN 30-8. (A) Effect of proteinase inhibitor cocktail and its components. (a) No inhibitor; (b) cocktail with 10 mM EDTA, 2 mM leupeptin, 2 mM pepstatin, 1 mM PMSF, and 1 mg/ml aprotinin; (c) 10 mM EDTA (d); 2 mM leupeptin; (e) 2 mM pepstatin; (f) 1 mM PMSF; (g) 1 mg/ml aprotinin. (B) Effect of inhibitors of leukocyte elastase and cathepsin G. (a) No inhibitor; (b) proteinase inhibitor cocktail; (c) 1 mM PMSF; (d) 300 mM cathepsin G inhibitor Z-gly-leu-phe chloromethyl ketone CG/ CMK; (e) 500 mM human leukocyte elastase inhibitor N-methoxysuccinyl-ala-ala-pro-val chloromethyl ketone HLE/CMK; (f) 300 mM CG/ CMK plus 500 mM HLE/CMK. (C) Effect of mAb anti-leukocyte elastase. (a) Mouse IgG control; (b) mAb anti-human leukocyte elastase (12 mg/ml final concentration).

estingly, eosinophils appear to lack elastase [31]. Recently, Owen et al. reported that stimulation of blood PMNs induces cell surface expression of catalytically active human leukocyte elastase and cathepsin G [26, 27]. In the present study, arrest of fibronectin proteolysis by removal of the cells indicated a similar mechanism, namely participation of cell-associated proteinases. In addition, since fibronectin was first secreted as an intact molecule and the proteolysis was largely inhibited by an antibody to elastase, the proteinases must be cell-surface bound. In accordance, Owen et al. reported that no human leukocyte elastase activity was detected in cell-free supernatants of PMNs primed with LPS and stimulated with fMLP [26, 27]. Several groups have previously reported proteolysis of exogenous (plasma) fibronectin by stimulated blood PMNs. While three studies concluded major, if not sole, participation of human leukocyte elastase [26, 32, 33], a fourth study concluded, by using different inhibitors, that cathepsin G was more important than human leukocyte elastase [34]. The later enzyme has been identi-

AID

ECR 3545

/

6i20$$$$83

04-24-97 00:26:41

fied as the proteinase in wound fluid of burn patients responsible for degradation of fibronectin [35]. Indeed, purified human leukocyte elastase and cathepsin G are able to digest purified plasma fibronectin [36, 37]. The former enzyme generates fragments of 140, 60, and 29 kDa [36]. The present study describes, for the first time, proteolysis of endogenous fibronectin of PMNs. Epitopes of the few monoclonal antibodies reactive against the fibronectin fragments, including FN 30-8 [24], are located in a wide middle region of the fibronectin chain. All antibodies against the most NH2- and COOH-terminal regions were unreactive. Since practically all tested antibodies recognize intact PMN fibronectin by Western blotting [19], it is likely that epitopes are destroyed during the proteolysis. Certainly, a number of fragments, not visualized by the few reactive antibodies, should also be present. Four different patterns were observed with the reactive antibodies. Although only sequencing would give a definitive answer, these four patterns also contribute to identifying the fragments. Based on these antibody reactivities, we proposed a model of proteolysis. In this model, which is in accordance with the kinetics of proteolysis, almost simultaneous cleavages at COOH- and NH2-termini occur first, followed by additional cleavages at the termini of the large middle fragment. This sequential proteolysis from the extremes to the center of the fibronectin chain gives rise to the central fragments of 150, 120, 90, and 80 kDa. According to the reactivity with mAb FN 12-8 against the 10th type III repeat [24], the RGD sequence would be present in the three larger fragments. Secreted fibronectin or its fragments from a blood PMN could be used by the same cell, other PMNs, or other cell types such as monocytes. Several fibronectinbinding integrins are expressed by blood PMNs, namely a5b1, aVb3, a4b1 [10, 38, 39] and the less wellcharacterized leukocyte response integrin [40]. All, but a4b1, recognize the RGD sequence [10, 12, 38–40]. Whether b2 integrins, the most abundant integrins of blood PMNs, recognize fibronectin is controversial [3, 5–7, 10, 12, 41–43]. Interestingly, CD11b/CD18, the major b2 integrin of blood PMNs, mediates adherence of these cells to plastic and other artificial surfaces [5– 7]. We have found that fibronectin secreted from blood PMNs binds to polystyrene beads (unpublished results). Thus, it is feasible that CD11b/CD18 binds endogenous fibronectin or its fragments. Studies are in progress in our laboratory to examine this issue. Recently, Cai and Wright reported binding of human leukocyte elastase to CD11b/CD18 and participation of the elastase in PMN detachment [44]. Since exogenous fibronectin or its fragments are known to enhance chemotaxis, phagocytosis, degranulation, and oxidative metabolism in PMNs [13–16] and to mediate chemotaxis of monocytes [17], similar functions could be at-

eca

PMN FIBRONECTIN: SECRETION AND ENDOPROTEOLYSIS

tributed to secreted PMN fibronectin or its fragments. In a recent study [45], Parekh et al. reported that transforming growth factor-induced PMN chemotaxis was mediated through recognition of endogenous fibronectin by the a5b1 integrin. In accordance, the authors detected fibronectin in the supernatant of PMNs. Thus, the secreted and deposited fibronectin could provide a trail for cell adhesion and migration. In addition, the ability of fibronectin to bind collagen and certain microorganisms [8, 46, 47] may contribute to PMN migration and phagocytosis, respectively. What could be the physiological significance of fibronectin proteolysis? Fragmentation may expose cryptic sites within the molecule. In accordance, proteolytic digests of fibronectin, but not the intact molecule, were found to promote PMN chemotaxis and degranulation [13, 15]. Similarly, fibronectin fragments, but not intact fibronectin, have been shown to be chemotactic for monocytes [17], cells which follow PMNs during extravasation [18]. These effects appear to involve the RGD sequence of the cell-binding domain. Thus, proteolytic fragments of fibronectin possess biological properties not inherent to intact fibronectin. Interestingly, a fibronectin fragment from the gelatin-binding domain has been reported to inhibit several leukocyte activities in vitro [48], and proteinase inhibitors have been found to inhibit PMN chemotaxis [49, 50]. Moreover, synthetic peptides corresponding to the RGD cell-binding domain and other regions of fibronectin have been demonstrated to suppress leukocyte adhesion and recruitment and to inhibit inflammation [51]. To our knowledge, this is the first report on processing of endogenous PMN fibronectin by the cell’s own proteolytic machinery. Further studies will be necessary to define the functional significance of the secreted and proteolytically cleaved PMN fibronectin. This project was supported by the Swedish Cancer Society and the Karolinska Institute. We thank Drs. Erkki Ruoslahti (La Jolla Cancer Research Foundation, La Jolla, CA) and Niels Borregaard (The National University Hospital, Copenhagen, Denmark) for comments on the manuscript.

2. Borregaard, N., Lollike, K., Kjeldsen, L., Segelov, H., Bastholm, L., Nielsen, M. H., and Bainton, D. F. (1993) Eur. J. Haematol. 51, 187–198. 3. Carlos, T. M., and Harlan, J. M. (1994) Blood 84, 2068–2101. 4. Pakianathan, D. R. (1995) J. Leuk. Biol. 57, 699–702. 5. Arnaout, M. A. (1990) Blood 75, 1037–1050. 6. Springer, T. A. (1990) Nature 346, 425–434. 7. Patarroyo, M., Prieto, J., Rincon, J., Timonen, T., Lundberg, C., ˚ sjo¨, B., and Gahmberg, C. G. (1990) Immunol. Lindbom, L., A Rev. 114, 67–108.

/

6i20$$$$83

9. Zucker, M. B., Mosesson, M. W., Broekman, M. J., and Karplan, K. L. (1979) Blood 54, 8–12. 10. Bohnsack, J. F., and Zhow, X. N. (1992) J. Immunol. 149, 1340– 1347. 11. Ruoslahti, E., and Pierschbacher, M. D. (1987) Science 238, 491–497. 12. Hynes, R. O. (1992) Cell 69, 11–25. 13. Odekon, L. E., Frewin, M. B., Del Vecchio, P., Saba, T. M., and Gudewicz, P. W. (1991) Immunology 74, 114–120. 14. Proctor, R. A. (1987) Rev. Infect. Dis. 9, S412–S419. 15. Wachtfogel, Y. T., Abrams, W., Kucich, U., Weinbaum, G., Schapira, M., and Colman, R. (1988) J. Clin. Invest. 81, 1310– 1316. 16. Nathan, C., Srimal, S., Farber, C., Sanchez, E., Kabbash, L., Asch, A., Gailit, J., and Wright, S. D. (1989) J. Cell Biol. 109, 1341–1349. 17. Clark, R. A. F., Wikner, N. E., Doherty, D. E., and Norris, D. E. (1988) J. Biol. Chem. 263, 12115–12123. 18. Doherty, D. E., Downey, G. P., Worthen, G. S., Harlett, C., and Henson, P. M. (1988) Lab. Invest. 59, 200–213. 19. Salcedo, R., Segura, C., Szekely, L., de Mesquita, R., Biberfeld, P., and Patarroyo, M. (1997) Exp. Cell Res. 233, 25–32. 20. Hoffsstein, S. T., Weissmann, G., and Pearlstein, E. (1981) J. Cell Sci. 50, 315–327. 21. Menard, C., Beaulieu, A. D., Audette, M., Corbeil, J., and Latulippe, L. (1985) Clin. Exp. Immunol. 60, 347–354. 22. Marino, J. A., Davis, A. H., and Spagnuolo, P. J. (1987) Biochem. Biophys. Res. Commun. 146, 1132–1138. 23. Powers, J. C., Gupton, B. F., Harley, A. D., Nishino, N., and Whitley, R. J. (1977) Biochem. Biophys. Acta 485, 156–166. 24. Katayama, M., Hino, F., Odate, Y., Goto, S., Kimizuka, F., Kato, I., Titani, K., and Sekiguchi, K. (1989) Exp. Cell Res. 185, 229– 236. 25. Pierschbacher, M. D., Hayman, E. G., and Ruoslahti, E. (1981) Cell 26, 259–267. 26. Owen, C. J., Campbell, M. A., Sannes, P. L., Boukedes, S. S., and Campbell, E. J. (1995) J. Cell Biol. 131, 775–789. 27. Owen, C. A., and Campbell, E. J. (1995) Semin. Cell Biol. 16, 367–376. 28. Beaulieu, A. D., Lang, F., Belles-Issles, M., and Poubelle, P. (1987) J. Rheumatol. 14, 656–661. 29. Beaulieu, A. D., Audette, M., Menard, C., Parent, C., and Duval, M. (1985) Clin. Exp. Immunol. 60, 339–346.

31. Wiedow, O., Muhle, K., Streit, V., and Kameyoshi, Y. (1996) Biochem. Biophys. Acta 1315, 185–187.

1. Smith, J. A. (1994) J. Leuk. Biol. 56, 672–686.

ECR 3545

8. Hynes, R. O. (1990) Fibronectins (Rich, A., Ed.), Springer-Verlag, New York.

30. Borregaard, N., Sehested, M., Nielsen, B. S., Sengelov, H., and Kjeldsen, L. (1995) Blood 85, 812–817.

REFERENCES

AID

39

04-24-97 00:26:41

32. Daudi, I., Gudewicz, P. W., Saba, T. M., Cho, E., and Vincent, P. (1991) J. Leukoc. Biol. 50, 331–340. 33. Camba, A., Afford, S. C., Stockley, R. A., and Burnett, D. (1991) Am. J. Cell Mol. Biol. 4, 330–337. 34. Kubes, P., Smith, R., Grihman, M. D., and Granger, D. N. (1993) Inflammation 17, 321–332. 35. Grinnel, F., and Zhu, M. (1994) J. Invest. Dermatol. 103, 155– 161. 36. McDonald, J. A., and Kelley, D. G. (1980) J. Biol. Chem. 255, 8848–8858. 37. Vartio, T. (1981) J. Biol. Chem. 256, 471–477.

eca

40

SALCEDO, WASSERMAN, AND PATARROYO

38. Hendey, B., Lawson, M. A., Marcantonio, E. E., and Maxfield, F. R. (1996) Blood 87, 2038–2048. 39. Issekutz, T. B., Miyasaka, M., and Issekutz, A. C. (1996) J. Exp. Med. 183, 2175–2184. 40. Gresham, H. D., Goodwin, J. L., Alle, P. M., Anderson, D. C., and Brown, E. J. (1989) J. Cell Biol. 108, 1935–1943. 41. Thompson, H. L., and Matsushima, K. (1992) Clin. Exp. Immunol. 90, 280–285. ˚ kerlund, E., Berger, E., and Arfors, K. E. (1992) J. 42. Lundgren-A Leuk. Biol. 51, 603–608. 43. Owen, C. A., Campbell, E. J., and Stockley, R. A. (1992) J. Leuk. Biol. 51, 400–408.

44. Cai, T.-Q., and Wright, S. D. (1996) J. Exp. Med. 184, 1213– 1223. 45. Parekh, T., Saxena, B., Reibman, J., Cronstein, B. N., and Gold, L. I. (1994) J. Immunol. 152, 2456–2466. 46. Engvall, E., and Rouslahti, E. (1977) Int. J. Cancer 20, 1–5. 47. Proctor, R. A., Prendergast, E., and Mosher, D. F. (1982) Blood 59, 681–687. 48. Easter, D. W., Hoyt, D. B., and Ozkan, A. N. (1988) J. Surg. Res. 45, 370–375. 49. Hornebeck, W., Soleilhac, J. M., Tixier, J. M., Moczar, E., and Robert, L. (1987) Cell Biochem. Function 5, 113–122. 50. Barna, J. B., and Kew, R. R. (1995) Inflammation 19, 561–574. 51. Walh, S. M., Feldmamn, G. M., and McCarthy, J. B. (1996) J. Leuk. Biol. 59, 789–796.

Received November 14, 1996 Revised version received February 21, 1997

AID

ECR 3545

/

6i20$$$$84

04-24-97 00:26:41

eca