Inhibition of factor XIIIa-mediated incorporation of fibronectin into fibrin by pulmonary surfactant

Andreas Elssner, Gertraud Mazur, Claus Vogelmeier

Abstract

Intra-alveolar deposition of exudated plasma proteins is a hallmark of acute and chronic inflammatory lung diseases. In particular, fibrin and fibronectin may provide a primary matrix for fibrotic lung remodeling in the alveolar compartment. The present study was undertaken to explore the effect of two surfactant preparations on the incorporation of fibronectin into fibrin. We observed that surfactant phospholipids are associated with insoluble fibrin, factor XIIIa-cross-linked fibrin, and cross-linked fibrin with incorporated fibronectin. Factor XIIIa-mediated binding of fibronectin to fibrin was noticeably altered in the presence of surfactant. Coincubation with two different commercially available surfactants but not with dipalmitoylphosphatidylcholine alone resulted in a reduction of fibronectin incorporation into fibrin clots by approximately one-third. This effect was not dependent on the calcium concentration. We conclude that 1) factor XIIIa-cross-linked fibrin-fibronectin is able to incorporate surfactant phospholipids in amounts comparable to fibrin clots without fibronectin and2) the binding of fibronectin to fibrin is partially inhibited in the presence of pulmonary surfactant.

  • plasma proteins
  • surfactant phospholipids
  • adult respiratory distress syndrome
  • fibrosis

leakage of plasma proteins into alveoli and formation of hyaline membranes are commonly found in adult respiratory distress syndrome (ARDS) and chronic inflammatory lung diseases such as idiopathic pulmonary fibrosis (3, 7, 20). In particular, fibrin and fibronectin may provide a primary matrix for fibrotic lung remodeling (1, 2, 6, 8, 23, 41). The intra-alveolar exudate leads to a dysfunction of the surfactant system. It is therefore assumed that under such circumstances regulation of alveolar surface tension is disturbed (for a review see Ref. 35). Among the proteins capable of inhibiting surface tension-lowering properties of surfactant, fibrinogen and fibrin monomer surpass the potency of, e.g., albumin and hemoglobin (36). The predominant inhibitory mechanism may be the incorporation of lipophilic surfactant components into growing fibrin clots (34). The likelihood of intra-alveolar clot formation in ARDS is enhanced by a shift in the hemostatic balance toward a predominance of coagulation (10, 19, 29). An increased expression of procoagulative factors such as tissue factor and factor VII may be responsible for the extrinsic activation pathway of coagulation within the alveolus. Tissue factor and factor VII may be produced by alveolar epithelial cells (14) as well as by activated macrophages (10, 29). In addition, reduced fibrinolytic activity via an urokinase-type activator and elevated levels of plasminogen activator inhibitor-1 and α2-antiplasmin were found in inflammatory lung diseases (5, 9, 19). Whereas leakage from capillaries appears to be the predominant source of intra-alveolar plasma proteins, fibronectin may also be locally secreted by alveolar macrophages (32,33). Usually, covalent cross-linking of fibrin and the incorporation of 0.7% (percentage of clot mass) α2-antiplasmin and 4.4% fibronectin is mediated by factor XIIIa (11, 27). The fibronectin binding site of fibrin mediated by factor XIIIa is located in the carboxy-terminal region of the α-chain (26). The corresponding binding region on fibronectin is located in the amino terminus of the molecule (18). Because the fibrin split productd-dimer has been found in bronchoalveolar lavage fluid obtained from humans with acute lung injury (12), covalent cross-linking of fibrin may also take place in the alveolus. This hypothesis is supported by the occurrence of cross-linked fibrin in the terminal airways of rats with bleomycin-induced lung injury (31). With these findings as a background, the present study was conducted to evaluate the interaction of fibrin-fibronectin with pulmonary surfactant. We studied the influence of two commercially available surfactants, a bovine lung extract (Alveofact) and a porcine lung extract (Curosurf), on the incorporation of fibronectin into fibrin.

METHODS

Materials. Alveofact was kindly provided by Thomae (Biberach, Germany). Curosurf was a gift from Serono Pharma (Unterschleissheim, Germany). Purified bovine fibrinogen (plasminogen and factor XIII free, >95% clotting ability) was generously donated by Dr. H. Keuper (Behringwerke, Marburg, Germany). Human thrombin (specific activity 120 U/mg) and human fibronectin (pure; Boehringer Mannheim, Mannheim, Germany), human factor XIII (Fibrogammin, Centeon Pharma, Marburg, Germany), dipalmitoylphosphatidylcholine (DPPC; Sigma, Munich, Germany), 125I-labeled fibrinogen (200 μCi/mg), [14C]DPPC (154 μCi/mg; both from Amersham Buchler, Braunschweig, Germany),125I-fibronectin (220 μCi/mg), and Coomassie blue colloidal stain (both from ICN Biomedicals, Eschwege, Germany) were commercially available. Microcentrifuge tubes with 10-μm polypropylene mesh filters were obtained from Whatman (Maidstone, UK).

Preparation of reaction mixtures. For examination of phospholipid incorporation into fibrin, surfactant and plasma proteins (fibrinogen and fibronectin) were mixed and diluted with buffer (20 mM Tris, 0.9% NaCl, and 3 mM CaCl2) to a final sample volume of 0.5 ml. All samples contained Alveofact at a constant concentration of 2 mg/ml of phospholipids. Fibrinogen was added at different concentrations ranging from 0.1 to 2 mg/ml in either the presence or absence of 1 U/ml of factor XIII. Alternatively, in addition to factor XIII, 0.2 mg/ml of fibronectin was admixed. In addition, a fibrinogen-free preparation was made consisting of 2 mg/ml of surfactant, fibronectin in varying concentrations from 0.05 to 1 mg/ml, and 1 U/ml of factor XIII. In all samples except controls, coagulation was started by the addition of 0.5 U/ml of thrombin. The reaction temperature was 37°C, and the incubation time was 3 h. Before initiation of coagulation, the mixtures were vortexed for 1 min followed by a 10-min period of gentle shaking at 37°C.

To investigate the influence of surfactant on the incorporation of fibronectin into fibrin, 2 mg/ml of fibrinogen and 0.2 mg/ml of fibronectin (diluted with buffer as described) were coagulated by adding 1 U/ml of factor XIII and 0.5 U/ml of thrombin in the presence and absence of surfactant. Before admixture, the thrombin samples were vortexed and preincubated as described above. Surfactant stock solutions (Alveofact and Curosurf) were admixed to a final concentration of 5 mg/ml of phospholipids. The reaction time ranged from 5 to 60 min. In addition, dose-effect characteristics were studied by varying the surfactant concentrations from 0.1 to 5 mg/ml at an incubation time of 1 h. In this experiment, the effect of the two surfactants was compared with that of DPPC (0.1–5 mg/ml). In another experiment designed to study the effects of varying calcium concentrations, 2 mg/ml of fibrinogen, 0.2 mg/ml of fibronectin, and 5 mg/ml of Curosurf were incubated with factor XIII and thrombin for 3 h as described while CaCl2 was added at different concentrations ranging from 0 to 10 mM.

Determination of phospholipid, fibrinogen, and fibronectin content in the clotted material. Either [14C]DPPC (10 nCi/mg surfactant phospholipids),125I-fibrinogen, or125I-fibronectin (30 nCi/ml each) was added to the reaction mixtures. After the incubation period, the mixtures were spun at 300 g through a 10-μm polypropylene mesh filter to separate the clot from the clot liquor. 14C and125I counts were identified in the filtrate and related to the radioactivity assessed before the start of coagulation in each experiment. The cross-linking reaction was not affected by the radioactivity because pilot studies all showed similar results with different amounts of labeled and unlabeled proteins as well as of phospholipids. In other control experiments, similar results were observed by calculating the content of phospholipids and protein based on the measurement of organic phosphorus and photometric determination of absorbance.

SDS-PAGE. For gel electrophoresis, coagulation was stopped by adding an equal volume of SDS-PAGE sample buffer containing 4% SDS, 20 mM EDTA, and 4% 2-mercaptoethanol in 8 M urea and 0.02% bromphenol blue. After complete dissolution of the clot, 10 μl of the mixture were subjected to 7.5% polyacrylamide separation gels. The running buffer was 0.025 M Tris and 0.192 M glycine, pH 8.3, with 0.1% SDS. The gels were fixed with 40% methanol, 10% acetic acid, and 50% deionized water. Staining was performed with Coomassie blue colloidal stain and destaining with 10% acetic acid and 90% deionized water.

Control of enzymatic reactions.Thrombin was dissolved to a 10 U/ml stock solution and divided into aliquots. Thrombin activity was monitored by the use of a cromogenic substrate-based commercial assay. For all other proteins (fibrinogen, fibronectin, and factor XIII), solutions of the lyophilized materials were prepared from the same batches on the day of the experiment.

RESULTS

Incorporation of phospholipids into polymerizing fibrin. In these experiments, >95% of the fibrinogen had to be clotted to insoluble fibrin after the incubation time of 3 h. Fibrin was therefore retained on the filter, and a quasi fibrinogen-free clot liquor was released into the filtrate. As anticipated, when 2 mg/ml of [14C]DPPC-enriched surfactant (Alveofact) were added, the Alveofact was incorporated into the growing fibrin clot in a dose-dependent manner. Fibrin at a dose of 0.5 mg/ml provoked an ≈50% loss in [14C]DPPC from the soluble phase, whereas the polymerization of 2 mg/ml of fibrin resulted in >90% depletion of DPPC from the clot liquor (Fig.1 A). Adding factor XIII to produce covalently cross-linked fibrin caused a dose-dependent effect that paralleled that of noncovalently cross-linked fibrin (Fig. 1 B). No significant change was observed when fibronectin was admixed (Fig.1 C). In contrast, no retention of DPPC on the filter and full recovery in the soluble filtrate were achieved by incubating 2 mg/ml of Alveofact, fibronectin in varying concentrations from 0.05 to 1 mg/ml, and 1 U/ml of factor XIII in the absence of fibrin(ogen) (Fig. 1 D).

Fig. 1.

Dose-dependency of dipalmitoylphosphatidylcholine (DPPC;14C labeled) association with growing fibrin clots. Alveofact (2 mg/ml) was mixed with [14C]DPPC and increasing concentrations of fibrinogen. Complete conversion of fibrinogen to fibrin was achieved after incubation of mixture with 0.5 U/ml of thrombin for 3 h at 37°C. Clotting to fibrin was performed in absence of factor (F) XIIIa (A), in presence of 1 U/ml of F XIIIa to get covalently cross-linked fibrin (B), and in presence of F XIIIa and 0.2 mg/ml of fibronectin (C). As a control, fibronectin in concentrations ranging from 0.05 to 1 mg/ml was incubated with 0.5 U/ml of thrombin and 1 U/ml of F XIIIa in absence of fibrinogen (D). To separate clotted material from clot liquor, samples were filtered. Radioactivity was assessed in filtrate, and counts are given as a percentage of radioactivity initially present in incubation medium. Each point is mean ± SE of at least 4 independent experiments; error bars not shown are within symbol.

Inhibition of incorporation of fibronectin into fibrin by surfactant. An incubation time (5–60 min)-dependent rise in the amount of fibrin-fibronectin hybrids could be evidenced by SDS-PAGE (Fig. 2). In the clot filtration experiments, incubation of 2 mg/ml of fibrinogen with 0.2 mg/ml of fibronectin resulted in the incorporation of ≈35% of the fibronectin after 10 min and ≈45% after 60 min of incubation through the action of activated factor XIII (Fig.3). The amount of incorporated fibronectin could not be enhanced by increasing the incubation time to 24 h. With the addition of 5 mg/ml of Curosurf, the amount of fibronectin incorporated into fibrin was markedly reduced. After an initial incorporation of ≈30% in the first 10 min, no significant increase could be achieved by increasing the incubation time (Fig. 3). Similar results were observed with 5 mg/ml of Alveofact, although there was a slightly lower incorporation of fibronectin compared with that with Curosurf (Fig. 3). No substantial loss of fibronectin in the filtrate and, therefore, no incorporation into fibrin were noticed in the absence of factor XIII in the incubation mixture (Fig. 3). The same was true in the absence of calcium and in the absence of fibrinogen (data not shown in detail). Dose-effect characteristics of Curosurf and Alveofact were studied with concentrations ranging from 0.1 to 5 mg/ml of phospholipids during a 60-min incubation period. For Curosurf, a significant increase in fibronectin in the soluble phase (indicating nonincorporation into fibrin) was observed between 0.5 and 1 mg/ml of phospholipids, and the maximum amount of fibronectin in the filtrate occurred at 5 mg/ml of phospholipids. No significant inhibitory effect could be demonstrated at concentrations < 1 mg/ml. With Alveofact, similar results were obtained (Fig. 4). In contrast, with DPPC alone, no inhibitory effect could be found (Fig.4). The inhibition of factor XIIIa-mediated binding of fibronectin to fibrin was also evidenced by SDS-PAGE. Coincubation with 5 mg of Alveofact during coagulation resulted in a markedly reduced band for the fibronectin α-complex (Fig. 5). There was optimal fibronectin incorporation into fibrin in the presence and absence of 5 mg/ml of Curosurf at a 3–5 mmol/l calcium concentration (Fig. 6). Adding more calcium to the incubation mixture was not accompanied by an increase in fibronectin incorporation into fibrin whether or not the surfactant was present.

Fig. 2.

F XIIIa-mediated incorporation of fibronectin into fibrin as documented by SDS-PAGE under reducing conditions. Fibrinogen (2 mg/ml) was incubated with 0.2 mg/ml of fibronectin, 1 U/ml of F XIIIa, and 0.5 U/ml of thrombin. Coagulation was stopped by adding sample buffer (seemethods) after incubation periods of 5 (lane 1), 10 (lane 2), 30 (lane 3), and 60 (lane 4) min.

Fig. 3.

Incorporation of fibronectin into fibrin is dependent on incubation time in absence of F XIII (A), in presence (+) of Alveofact (B), in presence of Curosurf (C), and in absence of surfactant (D). Except for control where F XIII was left out, 2 mg/ml of fibrinogen were mixed with fibronectin (0.2 mg/ml, 125I labeled) and coagulation occurred over different time intervals after 0.5 U/ml of thrombin and 1 U/ml of F XIII were added. Alternatively, either Alveofact or Curosurf (5 mg of phospholipids each) was added to initial solution. Each point is mean ± SE of at least 4 independent experiments.

Fig. 4.

Influence of Curosurf, Alveofact, and DPPC alone on fibronectin (0.2 mg/ml) incorporation into fibrin. Complete conversion of fibrinogen (2 mg/ml) into insoluble fibrin was achieved by incubation with 1 U/ml of F XIII and 0.5 U/ml of thrombin for 1 h at 37°C. Surfactants and DPPC were added in different doses ranging from 0 to 5 mg/ml. Samples were filtered, and radioactivity was assessed in filtrate. Each point is mean ± SE of at least 4 independent experiments; error bars not shown are within symbol.

Fig. 5.

SDS-PAGE (reducing conditions) of fibronectin incorporation into fibrin by F XIIIa in presence and absence of pulmonary surfactant.Lane 1, fibronectin and fibrinogen (α-, β-, and γ-chains) before start of coagulation (2 mg/ml fibrinogen and 0.2 mg/ml fibronectin); lane 2, coagulation of fibrin with incorporated fibronectin in presence of γ-γ-cross-links between γ-chains of fibrin and development of complexes between fibronectin and α-chains of fibrin;lane 3, same protocol except for addition of 5 mg/ml of Curosurf before start of coagulation by F XIIIa.

Fig. 6.

Dependence of fibronectin incorporation into fibrin on calcium concentration in presence and absence of 5 mg/ml of surfactant (Curosurf). Fibrinogen (2 mg/ml) and fibronectin (0.2 mg/ml) were clotted by 0.5 U/ml of thrombin and 1 U/ml of F XIII at different calcium concentrations from 0 to 10 mmol/l. Clotted material was separated from clot liquor by filtering samples. Radioactivity was assessed in filtrate. Each point is mean ± SE of at least 4 independent experiments.

DISCUSSION

In the present work, we observed incorporation of surfactant phospholipids into a growing fibrin network. This was demonstrable in factor XIII-free fibrinogen coagulated by thrombin and factor XIIIa-cross-linked fibrin as well as in cross-linked fibrin with incorporated fibronectin. In all dose-effect experiments, ∼50% of the DPPC present at a surfactant concentration of 2 mg/ml of phospholipids was retained in the clotted material at a fibrin concentration of 0.5 mg/ml, whereas >90% of the DPPC was retained in the clot when 2 mg/ml of fibrin were used. Interestingly, the presence of two different commercially available surfactants (Alveofact and Curosurf) during the process of coagulation led to an inhibition of fibronectin incorporation into fibrin.

With the development of hyaline membranes with an intra-alveolar accumulation of plasma proteins, a well-known morphological feature of inflammatory lung diseases (3, 7), Balis et al. (4) first reported a coagulative type of surfactant depletion. It has been shown that this depletion is a result of an association of surfactant phospholipids with polymerizing fibrin (34). Plasma proteins and, in particular, fibrinogen and fibrin monomer cause a deterioration of the biophysical properties of pulmonary surfactant (35, 36). The “trapping” of surfactant phospholipids, mainly DPPC but also surfactant apoprotein (SP) B (A. Elssner and W. Seeger, unpublished data), results in a loss of surface activity that surpasses the effect of fibrinogen and fibrin monomer by more than two orders of magnitude (34). In plasma coagulation under physiological conditions, ∼4–5% fibronectin is incorporated into the clot due to the action of activated factor XIII (27). In agreement with these findings, the calculated fibronectin content within our fibrin clots was 4–5%. In addition, we chose a ratio of fibronectin to fibrin (0.2:2 mg/ml) that is found physiologically in human plasma (26, 27).

Admixture of either Alveofact or Curosurf reduced the incorporation rate of fibronectin to approximately one-third compared with coagulation in the absence of surfactant. Whereas the time-dependent fibronectin incorporation into fibrin was slightly reduced with Alveofact, a significant difference between Alveofact (a calf lung extract) and Curosurf (a pig lung extract) could not be demonstrated. This finding is not surprising considering the similarities in the chemical composition of the two surfactants, which both predominantly contain polar lipids (mainly DPPC) and ≈1% hydrophobic SP-B and SP-C (information provided by the manufacturers). In summary, when surfactant was admixed, a distinct reduction in fibronectin incorporation into fibrin was reproducible, whereas a difference between the two surfactant extracts used could not be found. Interestingly, the addition of DPPC alone showed no significant effect on the binding of fibronectin to fibrin. Thus SP-B and SP-C, which are present in minor amounts in both surfactants used, are probably involved in the observed inhibitory effect. The interaction of surfactant lipids and SPs may lead to the change in the composition of the clot.

In agreement with earlier observations, we found no influence of surfactant on the generation of fibrin by thrombin- or factor XIII-induced cross-linking of fibrin. This implies that the enzymatic activity of thrombin and factor XIII themselves are not compromised by surfactant. The activation of factor XIII and, therefore, covalent incorporation of fibronectin into fibrin are strongly calcium dependent (25). Under the conditions we used, the optimum calcium concentration for fibronectin incorporation was 5 mM, which was independent from the presence or absence of surfactant. The inhibition of fibronectin incorporation into fibrin due to the admixed surfactant could not be overcome by enhancing the calcium concentration. Hence it is unlikely that the inhibitory effect of surfactant was due to the binding of calcium ions to negatively charged surfactant components. Thus other possible mechanisms need to be discussed.1) The alterations in the mechanical properties of fibrin caused by embedded surfactant may lead to changes in the three-dimensional arrangement of the fibrin network, with concomitant changes in the binding sites for fibronectin.2) Surfactant components may directly interfere with the binding sites of fibronectin. This could be due to electrostatic or hydrophobic interactions in these parts of the macromolecular fibrin network. 3) Surfactant components may inhibit the activity of factor XIIIa at the site of the fibronectin molecule itself. Surfactant also interferes with enzymatic reactions important for fibrinolysis: synthetic surfactant inhibits cleavage of fibrinogen by plasmin (15) and natural surfactant can block plasmin-induced fibrin lysis (37). Another in vitro study (16) showed that incorporation of a bovine surfactant extract into fibrin provokes retardation of clot lysis through the inhibition of plasmin, trypsin, or elastase.

The question is whether the demonstrated inhibitory effect of surfactant on fibronectin incorporation has biological relevance. Based on data obtained from studies with bronchoalveolar lavage fluid, the surfactant concentrations we used are within or even below the estimated concentration in the alveolar lining layer (35, 38, 40). Therefore, we and other investigators (17) believe that intra-alveolar coagulation takes place in an highly surfactant-enriched milieu. However, some skepticism is warranted. The lowest ratio of surfactant phospholipids to fibrinogen (mg/mg) at which we saw an effect was 0.5. Based on data from rat lungs (42), the concentration of surfactant phospholipids in a lung that is flooded to functional residual capacity (FRC) with fluid would be 0.33 mg/ml (estimated FRC of ∼3 ml). In cases where the locally produced amount of fibronectin can be neglected in acute lung injury, the intra-alveolar fibrinogen and fibronectin must be of plasma origin and the maximum concentrations of the proteins in the alveolar lining layer should be in the range of the physiological concentrations in plasma. Thus the maximum concentration of fibrinogen would be around 3 mg/ml. Hence alveoli that are flooded to their FRC with plasma would contain material with a surfactant phospholipid-to-fibrinogen ratio of 0.1 mg/mg. This is five times less than the lowest ratio we showed to have an effect. However, this calculation does not take into account that diffuse microatelectasis due to collapsed alveoli is a typical feature of the acute phase of ARDS. Thus the leakage of smaller amounts (<3 mg/ml) of fibrinogen may be sufficient to result in deterioration of the alveolar surfactant function and consequently contribute to the development of ARDS. In summary, our data may have relevance in diseases with an intra-alveolar accumulation of fibrin such as ARDS. On the other hand, under the condition of plasma-flooded alveoli, the surfactant phospholipid-to-fibrinogen ratio may be lower than the ratio at which we saw an effect on the incorporation of fibronectin into fibrin.

What can be expected when incorporation of fibronectin into fibrin is inhibited by surfactant in vivo within the alveolus, e.g., in ARDS? It is known that the addition of fibronectin to fibrin in physiological amounts increases fibrin fiber thickness and network permeability (28,30). In addition, altered mechanical properties of fibrin clots were recently described when coagulation takes place in a surfactant-enriched milieu. The fibrin network with incorporated surfactant displayed a decrease in the elastic modulus of arising fibrin polymers and an increased hydraulic conductivity in conjunction with an increased pore size, suggesting an altered architecture (17). Thus it could be speculated that these mechanical alterations of fibrin clots in the presence of surfactant are partially attributable to reduced fibronectin incorporation into fibrin, considering the minor amounts of fibronectin and factor XIII in the fibrinogen preparation used in that study. Fibrin and fibronectin may provide a provisional matrix for ongoing repair in lung injury. Type II cells are believed to reepepithelialize injured alveoli through integrin-mediated adherence to fibrin-fibronectin matrices or basement membranes (22, 39). On the other hand, a fibrin-fibronectin matrix may initiate fibrotic remodeling via an invasion of fibroblasts (1, 2, 6, 8, 23, 41), leading to a so-called collapse induration (7). Because fibroblasts bind more tightly to fibronectin when it is cross-linked to fibrin (13), surfactant may have an antifibroproliferative effect. Another aspect is that the addition of fibronectin to fibrin was shown to enhance macrophage binding in vitro (21). Therefore, the presence of surfactant within fibrin-fibronectin matrices could reduce the clearance of clots by alveolar macrophages within alveoli. This may prevent further plasma leakage, bleeding, and more lung damage due to persistent hyaline membranes. On the other hand, the reduced clearance of clots by surfactant may inhibit rapid restoration of gas exchange. Fibrin matrices that had been depleted of factor XIII and/or fibronectin were shown to be a superior matrix for macrophage migration over fibrin matrices with incorporated fibronectin (24). Consequently, surfactant may facilitate the movement of macrophages within damaged alveoli. In summary, our findings support the hypothesis that coagulation within the surfactant-containing alveolar milieu is different from clotting of plasma proteins in other compartments.

Footnotes

  • Address for reprint requests and other correspondence: A. Elssner, Dept. of Internal Medicine I, Div. for Pulmonary Diseases, Klinikum Grosshadern, Ludwig-Maximilians-Univ. of Munich, Marchioninistrasse 15, 81377 Munich, Germany (E-mail:Andreas.Ellsner{at}med1.med.uni-muenchen.de).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

REFERENCES

View Abstract