Fibrinolytic cross-talk: a new mechanism for plasmin formation

Abstract

Fibrinolysis and pericellular proteolysis depend on molecular coassembly of plasminogen and its activator on cell, fibrin, or matrix surfaces. We report here the existence of a fibrinolytic cross-talk mechanism bypassing the requirement for their molecular coassembly on the same surface. First, we demonstrate that, despite impaired binding of Glu-plasminogen to the cell membrane by epsilon-aminocaproic acid (epsilon-ACA) or by a lysine-binding site-specific mAb, plasmin is unexpectedly formed by cell-associated urokinase (uPA). Second, we show that Glu-plasminogen bound to carboxy-terminal lysine residues in platelets, fibrin, or extracellular matrix components (fibronectin, laminin) is transformed into plasmin by uPA expressed on monocytes or endothelial cell-derived microparticles but not by tissue-type plasminogen activator (tPA) expressed on neurons. A 2-fold increase in plasmin formation was observed over activation on the same surface. Altogether, these data indicate that cellular uPA but not tPA expressed by distinct cells is specifically involved in the recognition of conformational changes and activation of Glu-plasminogen bound to other biologic surfaces via a lysine-dependent mechanism. This uPA-driven cross-talk mechanism generates plasmin in situ with a high efficiency, thus highlighting its potential physiologic relevance in fibrinolysis and matrix proteolysis induced by inflammatory cells or cell-derived microparticles.

Figure 1. Identification of activators and cellular activation of plasminogen

A. Fibrin autography of platelets, cortical neurons, HMEC-1 and THP-1 cells. The samples were electrophoresed on 8% (w/v) polyacrylamide, SDS was then exchanged with 2.5% (w/v) Triton X-100 and the gel overlaid on a fibrin-agar indicator gel. The picture was taken after 4 h at 37 °C. The position of purified controls (Pn: plasmin, tPA and uPA) is indicated on top. The thin vertical line indicates assembly from the same gel. The thick vertical line separates two different gels. B. THP-1 cells (10⁵ cells/well) were incubated with varying concentrations of plasminogen (0 to 3 μM) and 0.75 mM CBS0065. Kinetics of plasmin formation (mOD/min) was followed by measuring the realease of p-nitroaniline. Data were fitted according to the Michaelis–Menten equation (Km = 492 nM).

Figure 2. Cellular activation of plasminogen: effect of LBS ligands

A. Cortical neurons (10⁵ cells/well) and B. THP-1 cells (10⁵/well), were incubated with 125 nM Glu-plasminogen supplemented with varying concentrations of ε-ACA (0 to 25 mM) and 0.75 mM CBS0065. Plasmin formation (●) was detected by measuring the release of p-nitroaniline. C. THP-1 cells (10⁵/well) were incubated with 125 nM Glu-plasminogen supplemented with of 0 to 1 μM anti-K1-LBS mAb 34D3 and 0.75 mM CBS0065. A, B, C. After detection of plasmin formation, the cells were washed twice with PBS and incubated with 0.75 mM CBS0065 to detect cell-associated plasmin (□). Results are expressed as a percentage (mean ± SD, n = 3) of plasmin formation or of cell-associated plasmin activity in the absence of ε-ACA or mAb.

Figure 3. Effect of ε-ACA and carboxipeptidase B on Lys- and Glu-plasminogen activation by cellular uPA

THP-1 cells (10⁵/well) were incubated with 500 nM Lys- (grey bars) or Glu-plasminogen (open bars) in medium alone or supplemented with ε-ACA (5 or 25 mM) and 0.75 mM CBS0065. Carboxypeptidase B, CpB (50 μg/mL), pre-treated THP-1 cells were incubated with plasminogen and CBS0065. Rate of plasmin formation (A) and amount of cell-associated plasmin (B) were detected as indicated in figure 2. Bars represent the amount of plasmin formed or associated to the cells (mOD/min) versus the concentration of ε-ACA or after CpB treatment (mean ± SEM, n = 3). Stars indicated significant changes as compare to Lys-Pg (*p < 0.05)/Glu-Pg (^§p < 0.05) alone.

Figure 4. Proteolytic cross-talk: activation of fibrin- and fibronectin-bound plasminogen by cellular microparticles

A. Native or recombinant active-site inactivated Glu-plasminogen (Glu-Pg, r-Pg-Ala⁷⁴¹) at 1 μM were bound to fibrin-coated beads for 1h at 37°C. Fibrin-coated beads with bound plasminogen were then incubated with 10⁶ endothelial microparticles (EMP) in a final volume of 200 μL. After overnight incubation at 22°C, the fibrin-coated beads were sedimented by centrifugation and resuspended in 10 mM Tris-HCL pH 6.8 containing 10 % SDS to elute fibrin-bound plasminogen derivatives. The supernatant was electrophoresed under non-reducing conditions, proteins were transferred to PVDF membranes and revealed with a HRP-conjugated mAb (150 ng/mL) directed against plasminogen K1. The Western blot shows Glu-plasmin formaton by EMP. Purified plasmin is shown as reference. B, C. Glu-plasminogen (1 μM) was bound to fibrin (B) or fibronectin (C) surfaces. After 3 washes with PBS, EMP were added at varying concentrations. D. Glu-Plasminogen (1 μM) was bound to fibrin surfaces (main graph) or to fibrin-coated beads (inset). THP-1 cells were then added to fibrin surfaces at varying concentrations (main graph) and 2.5×10⁵ fibrin-coated beads were incubated with 10⁵ adherent monocytes or neurons (inset). The formation of plasmin was detected by measuring the release of p-nitroaniline from the chromogenic substrate CBS0065 added at 0.75 mM. Bars represent the amount of plasmin formed (mOD/min, mean ± SEM, n = 3) by THP-1 cells on fibrin (B, D) and fibronectin (C), and by adherent monocytes or neurons on fibrin-coated beads (Inset to D). Amil: amiloride, anti-uPA, IgG: antibody against uPA and its non-immune IgG control. Stars indicated significant changes as compare to activation without THP-1 (A) or EMP (B, C) or activation on neurons (*p < 0.05); § indicated changes with inhibitors as compare to activation at 5×10⁵ EMP (^§p < 0.05).

Figure 5. Proteolytic cross-talk: activation of platelet-bound plasminogen by cells bearing uPA (monocytes) or tPA (neurons)

Glu-plasminogen (1 μM) was bound to platelets as indicated in Methods. A. After treatment with 50 μg/ml CpB, monocytes (●) or neurons (■) were incubated with plasminogen-bearing platelets at varying concentrations (0 to 1.510⁶/well) in the presence of 0.75 mM CBS0065. The formation of plasmin (mOD/min) was detected by measuring the release of p-nitroaniline. Inset. Detection of plasmin formation on platelets (1.8×10⁶) in the absence of cells or incubated with monocytes or neurons (10⁵ cells). Results are expressed as rate of plasmin formation (mean ± SEM, n = 3). B. Glu plasminogen (1 μM) was bound to platelets as indicated in Methods. After treatment with 50 μg/ml CpB, endothelial microparticles (EMP) were incubated with 5×10⁶ platelets bearing plasminogen (●) or with 3 nM plasminogen (concentration equivalent to that bound to platelets), or with buffer (▲) in the presence of 0.75 mM CBS0065. The formation of plasmin (mOD/min) was detected by measuring the release of p-nitroaniline. Results are expressed as rate of plasmin formation (n = 3). A representative experiment is shown.

Figure 6. Efficiency and specificity of the plasminogen cross-talk

Glu-plasminogen (1 μM) was bound to platelets as indicated in Methods. A. After treatment with 50 μg/mL CpB, adherent monocytes were incubated with 5×10⁶ platelets bearing plasminogen (cross-talk) or with plasminogen (concentration equivalent to plasminogen bound to platelets) (same surface) in the presence of 0.75 mM CBS0065. In a parallel experiment, uPA (concentration equivalent to uPA bound to monocytes) was incubated with 5×10⁶ platelets bearing plasminogen (free uPA) in the presence of 0.75 mM CBS0065. Results are expressed as rate of plasmin formation (mean ± SEM, n = 3, triplicates). NS: non significant (p = 0.078) *p <0.005, ^#p<0.014. B. Monocytes were incubated with 1 nM of native uPA and varying concentrations of a non-active mutant uPA (r-scuPA-Gly¹⁵⁹). Monocytes were then incubated with 5×10⁶ platelets bearing plasminogen, in the presence of 0.75 mM CBS0065. The formation of plasmin (mOD/min) was detected by measuring the release of p-nitroaniline (mean ± SEM, n = 2, triplicates). A representative experiment is shown (p = 0.023, 0 vs 5 nM r-scuPA-Gly¹⁵⁹).

Figure 7. Model of fibrinolytic and proteolytic cross-talk

In inflammatory processes of the vascular wall, activation of coagulation leads to fibrin deposits. Within this setting and according to data presented here, the fibrinolytic cross-talk mechanism could be an intermediary pathway for fibrinolysis and pericellular proteolysis. Fibrinolytic effects (upper part): plasminogen (Pg) bound to fibrin or to platelets could be activated by either monocyte- or microparticle-borne uPA (EMP or monocyte-derived microparticles, MoMP). In extracellular matrix (ECM) remodeling in the vascular wall (lower part), plasminogen bound to matrix components could be activated by microparticle-borne uPA (EMP or MoMP) or by macrophage-borne uPA.