β₂-Glycoprotein I Is a Cofactor for t-PA–Mediated Plasminogen Activation

Abstract

Regulation of the conversion of plasminogen to plasmin by tissue-type plasminogen activator (t-PA) is critical in the control of fibrin deposition. While several plasminogen activators have been described, soluble plasma cofactors that stimulate fibrinolysis have not been characterized. Here, we report that the abundant plasma glycoprotein, β₂-glycoprotein I (β₂GPI), stimulates t-PA–dependent plasminogen activation in the fluid phase and within a fibrin gel. The region within β₂GPI responsible for stimulating t-PA activity is at least partially contained within β₂GPI domain V. β₂GPI bound t-PA with high affinity (K_d ~ 20 nM), stimulated t-PA amidolytic activity, and caused an overall 20-fold increase in the catalytic efficiency (k_cat/K_m) of t-PA–mediated conversion of Glu-plasminogen to plasmin. Moreover, depletion of β₂GPI from plasma led to diminished rates of clot lysis, with restoration of normal lysis rates following β₂GPI repletion. Finally, stimulation of t-PA–mediated plasminogen activity by β₂GPI was inhibited by monoclonal anti-β₂GPI antibodies, as well as by anti-β₂GPI antibodies from patients with antiphospholipid syndrome (APS). These findings suggest that β₂GPI may be an endogenous regulator of fibrinolysis. Impairment of β₂GPI-stimulated fibrinolysis by anti-β₂GPI antibodies may contribute to the development of thrombosis in patients with APS.

Introduction

β₂-glycoprotein I (β₂GPI) belongs to the complement control (CCP) superfamily and circulates in plasma at a concentration of ~ 4μM (1). Like other CCP superfamily members, which contain one or more characteristic short consensus repeats (SCR)(2), β2GPI contains five SCRs. However, domain V of β₂GPI consists of an atypical SCR containing a lysine-rich sequence (CKNKEKKC) that imparts a positive charge to this domain and mediates the binding of β₂GPI to phospholipid (3). Plasmin cleavage between Lys317-Thr318 of β₂GPI domain 5 abolishes phospholipid binding (4). β₂GPI is an important antigen in the antiphospholipid syndrome (APS), and anti-β₂GPI antibodies are an independent risk factor for thrombosis and recurrent pregnancy loss (5–7). β₂GPI may regulate both procoagulant and anticoagulant activities (7–9). Thrombin generation is impaired in the plasma of β₂GPI-deficient mice (10).

Impaired fibrinolysis may contribute to the development of thrombosis (11;12). Plasmin plays a central role in lysis of fibrin thrombi, and the conversion of plasminogen to plasmin by tissue-type plasminogen activator (t-PA) and urokinase-type plasminogen activator (u-PA) is precisely regulated (13). t-PA binds fibrin with high affinity and may be of particular importance for lysis of fibrin thrombi in the vasculature (13). Several stimulators of fibrinolysis, including insoluble proteins, protein aggregates, microfilaments, fibrin fragments, or collagen IV may stimulate fibrinolysis, primarily by promoting formation of a ternary t-PA–fibrin–plasminogen complex (14;15). Annexin A2 is an endothelial cell co-receptor for plasminogen and t-PA that accelerates t-PA–dependent plasminogen activation (16;17), and the annexin A2/S100A10 heterotetramer may be even more potent in this regard (18–20). However, to date there have been no soluble plasma cofactors identified that promote t-PA-dependent plasminogen activation.

Several studies have examined interactions of β₂GPI with the fibrinolytic system. Lopez-Lira et al (21) and Yasuda et al (22) reported low affinity binding of Glu-plasminogen to intact, or plasmin-cleaved β₂GPI, respectively. However, interactions of β₂GPI with t-PA have not been described. Here, we report that β₂GPI binds t-PA with high affinity, and enhances t-PA activity and t-PA–dependent plasminogen activation. Depletion of β₂GPI from plasma impairs the lysis of plasma clots, and anti-β₂GPI antibodies inhibit the ability of β₂GPI to stimulate fibrinolysis. Given the abundance of β₂GPI in plasma, these findings suggest that β₂GPI may be an endogenous regulator of fibrinolysis, and that impairment of fibrinolysis by anti-β₂GPI antibodies may contribute to APS-associated thrombosis.

Materials and methods

Purification of β₂GPI from human plasma

β₂GPI was purified from human plasma using a modification of previously described methods (23). Briefly, polyethylene glycol (final concentration 15%) was added to plasma, and the precipitated proteins were collected by centrifugation at 10,000 × G for 30 minutes. The precipitate was resuspended, and β₂GPI was isolated by sequential chromatography using heparin-Sepharose CL-6B (Sterogene Bioseparations Inc., Carlsbad, CA) and Source 15S (Amersham-Pharmacia Biotech, Uppsala, Sweden). The purity of isolated β₂GPI was confirmed using 12% SDS-polyacrylamide gel electrophoresis (SDS-PAGE).

Cloning, expression and purification of β₂GPI in mammalian cells

Full length β₂GPI cDNA was cloned into pDNR-LIB (Sanying Biotechnology, Beijing, China). For expression of recombinant β₂GPI, the β₂GPI coding sequence was amplified using the primers 5′ C ACC ATG GGA CGG ACC TGT CCC AAG 3′ and 5′GCA TGG CTT TAC ATC GGA TGC ATC A 3′ (24) and cloned into pcDNA3.1. β₂GPI-pcDNA3.1 was transfected into 293-T cells, and cell extracts prepared 48 hours after transfection were analyzed by immunoblotting using rabbit anti-human β₂GPI and anti-His (C-terminal) antibodies. The 6× His-tagged recombinant β₂GPI was purified using Ni-NTA superflow columns (Amersham-Pharmacia Biotech).

Preparation of recombinant β₂GPI domain V

The sequence encoding β₂GPI domain V (β₂GPI-V) was amplified from β₂GPI-pDNR-LIB using the primers 5′ CAC GGA TCC AAA GCA TCT TGT AAA GTA CC 3′ and 5′ CTG AAG CTT TTA GCA TGG CTT TAC ATC 3′, and the PCR product cloned into vector PQE30 (ZH Xie, Tsinghua University) using Bam HI and Hind III restriction sites (underlined). The domain V construct was transfected into competent E. coli M15 cells grown in 100 μg/mL ampicillin and 25 μg/mL kanamycin, and expression was induced by exposure of cells to 1 mM IPTG at 37°C for 4 hours. The 6× His-tagged β₂GPI-V was purified using Ni-NTA superflow, and the purity of the recombinant polypeptide was confirmed using 15% SDS-PAGE.

A domain V peptide corresponding to amino acids Gly274-Cys288 (GQKVSFFCKNKEKKC) of β₂GPI was synthesized by Eurogentec (San Diego, CA).

Monoclonal anti-β₂GPI antibodies

The monoclonal anti-β₂GPI antibody BD4 was raised against intact β₂GPI, using standard methods (25). This antibody had an affinity constant for β₂GPI of 2.04 × 10⁷M⁻¹. Murine IgG1 monoclonal antibody 1D2, also raised against intact β₂GPI, was obtained from ABCAM (Cambridge, UK). The epitopes for these antibodies have not been defined. Both antibodies are of the IgG₁ subclass.

The effect of β₂GPI on t-PA-dependent plasminogen activation

The effect of β₂GPI on t-PA-dependent plasminogen activation was measured using a fluorogenic plasmin substrate (I-1390, H-D-Val-Leu-Lys-AMC; Bachem Bioscience, King of Prussia, PA). Briefly, 10 nM t-PA was incubated with increasing concentrations of native or recombinant β₂GPI, β₂GPI-V, β₂GPI domain V peptide, or 0.5% BSA (as a control) at room temperature for 15 minutes. Fifty microliters of each reaction mixture was then transferred to 96-well microplates, to which 100 nM Glu-plasminogen (Enzyme Research Laboratories, South Bend IN) and 200 μM H-D-Val-Leu-Lys-AMC was added. After mixing, substrate hydrolysis was measured at regular intervals as relative fluorescence units (I_{360/465 nm}). Initial rates of plasmin generation were determined by linear regression analysis of plots of I_{360/465 nm} versus time² using Prism 4.0 (GraphPad Software, San Diego, CA) (17).

Parallel studies were performed using an identical approach, but substituting urokinase-type plasminogen activator (u-PA) for t-PA.

The effect of β₂GPI on t-PA-dependent plasminogen activation in a fibrin gel

Fibrin-agarose gels were prepared by mixing 4 mL of agarose (1%, 55° C), 4 mL fibrinogen (3 mg/mL), 320 μL Glu-plasminogen (500 nM) and 320 μL thrombin (1 U/ml) in 10 mL polypropylene tubes. The contents of each tube were added to 90 mm plates and allowed to gel. After incubation at 4 C for 1 hour, 10 μl of reaction mixtures containing 10 nM t-PA alone or in the presence of β₂GPI, β₂GPI-V, β₂GPI domain V peptide or 0.5% BSA, were placed in individual wells prepared in the gels using glass pipettes. After 28 hours at 37 C, the area of lysis caused by each sample was measured.

Effect of plasma β₂GPI depletion on clot lysis

The fibrinolytic activity of control and β₂GPI-depleted plasma from three normal donors was compared. First, rabbit preimmune or anti-human β₂GPI IgG was conjugated to Affigel-10 (Bio-Rad) at a concentration of 5 mg IgG per ml of beads. The beads were placed in separate columns, and 2 ml of plasma was passed through each. β₂GPI was not detectable by immunoblotting of plasma passed through the anti-β₂GPI column, and the first 1 ml flow-through fraction from this column was used for subsequent studies. The corresponding flow-through fraction from the column containing preimmune IgG was used as a control.

The fibrinolytic activity of control and β₂GPI-depleted plasma was determined using a 96-well microplate clot lysis assay described by Nagashima et al (26). Briefly, 15 μl of β₂GPI-depleted or control plasma was mixed with 3 μl of 2 μM thrombomodulin and 60 μl of 0.04 M HEPES, pH 7.0, containing 0.15 M NaCl and 0.01% Tween-80. This mixture was added to another well containing 4 μl of 75 U/ml thrombin, 2 μl of 1 M CaCl₂ and 4 μl of 0.5 μg/ml t-PA. The total volume was adjusted to 120 μl with water. The absorbance of the clotted sample at 405 nm (A₄₀₅) was monitored every 15 minutes for 180 minutes. Maximal absorbance occurred at 30 minutes, and the extent of lysis was defined as the difference between A₄₀₅ at 30 minutes (T = 30), and 180 minutes (T = 180).

To confirm that any difference in clot lysis between control and β₂GPI-depleted plasmas was actually due to β₂GPI depletion, a β₂GPI-depleted plasma sample was replenished with 100 nM or 200 nM (final concentrations) purified human β₂GPI. Clot lysis rates were then determined using the protocol described in the preceding paragraph.

The effect of β₂GPI on t-PA amidolytic activity

t-PA activity was determined in the absence or presence of β₂GPI using a fluorogenic t-PA substrate (I-1195, Glutaryl-Gly-Arg-AMC; Bachem). Assays were performed in 96-well plates containing mixtures of t-PA (10 nM) and increasing concentrations of β₂GPI (0–5 μM), or 0.5% BSA. Reaction mixtures were incubated at room temperature for 15 minutes, at which point 200 μM I-1195 was added and substrate hydrolysis measured at regular intervals (I_{360/465 nm}). Initial rates of hydrolysis were calculated by linear regression of plots of I_{360/465 nm} versus time² (K) using Prism 4.0 (17).

Binding of β₂GPI to t-PA

Binding of β₂GPI to t-PA was initially analyzed using a microplate binding assay. High-binding ELISA plates (Costar, Cambridge, MA) were coated with 10 μg/mL single chain t-PA by overnight incubation at 4° C. After washing, wells were incubated with 100 μL of increasing concentrations of β₂GPI at 25° C for 4 hours. Bound β₂GPI was detected using rabbit anti-human β₂GPI (0.33 μg/mL) followed by peroxidase-conjugated goat anti-rabbit IgG and tetramethylbenzidine. Binding specificity was determined by assessing the ability of fluid phase t-PA to inhibit the binding of 40 nM β₂GPI.

We also analyzed the binding of β₂GPI to t-PA in real time using an IAsys plus biosensor system (Affinity Sensors, Paramus, NJ). β₂GPI was immobilized on a carboxymethyl dextran biosensor using amine coupling, and the amount of ligand bound to the β₂GPI-coated surface was assessed following injection of t-PA(0–14 nM). Iasys FASTfit software was used to determine the association and dissociation rate constant (K_ass and K_diss). The binding of plasminogen to β₂GPI was assessed using an identical approach.

Anti-β₂GPI antibodies inhibit plasminogen activation

IgG fractions were purified from 83 plasma samples (supplied by the Guangdong Family Planning Science and Technology Institute) using protein G agarose. These samples were from 40 healthy controls and 43 patients who met Sapporo criteria for definite APS (27). The 43 patients with APS were positive for either IgG/M anticardiolipin antibody (aCL) (32 patients), IgG/M anti-β₂-glycoprotein 1 (β₂GP1) antibody (24 patients), and/or or lupus anticoagulant (LA) (11 patients). Twenty-two samples contained anti-β₂GPI but not anti-tPA antibodies, while 21 APS samples contained neither anti-β₂GPI nor anti-tPA antibodies.

We also assessed the effect of affinity-purified anti-β₂GPI antibodies, isolated from patient plasma as we have previously described (28), on the ability of β₂GPI to enhance t-PA-mediated plasminogen activation.

The effect of anti-β₂GPI antibodies on β₂GPI-dependent enhancement of plasminogen activation was determined by measuring plasmin amidolytic activity following incubation of plasminogen with t-PA and β₂GPI in the absence or presence of monoclonal anti-β₂GPI antibodies, patient-derived anti-β₂GPI⁺/anti-tPA⁻ or anti-β₂GPI⁻/anti-tPA⁻ IgG fractions or affinity-purified anti-β₂GPI IgG. Initial rates of plasmin generation were calculated by linear regression analysis of plots of I_{360/465 nm} versus time² using Prism^® 4.0.

Statistical analysis

Stastical analyses were performed using the Student’s two-tailed T-test or the Mann-Whitney U test. A P value of < 0.05 was considered significant.

Results

β₂GPI stimulates t-PA-mediated plasminogen activation

Native β₂GPI (β₂GPI-n), recombinant β₂GPI (β₂GPI-r), and β₂GPI domain V (β₂GPI-V) accelerated the initial rate of plasminogen activation by t-PA in a concentration-dependent manner, as measured by cleavage of the plasmin substrate H-D-Val-Leu-Lys-AMC (Figure 1A, B). In contrast, the β₂GPI domain V peptide GQKVSFFCKNKEKKC did not significantly enhance plasminogen activation (Figure 1A).

Figure 1. β₂GPI enhances t-PA–dependent Glu-plasminogen activation.

(A) t-PA-dependent plasminogen activation was measured either alone (in the presence of BSA) or in the presence of 1 μM human β₂GPI (β₂GPI-n), recombinant β₂GPI (β₂GPI-r), β₂GPI domain V (β₂GPI-V), or β₂GPI domain V peptide (β₂GPI-p). Substrate hydrolysis was measured as relative fluorescence units (I_{360/465 nm}) as a function of time² (T)² (37). (B) Initial rates of plasmin generation were determined as in (A) in the presence of increasing concentrations of β₂GPI-n, β₂GPI-r, β₂GPI-V, or BSA.

β₂GPI was similarly effective when a fixed concentration of t-PA (10 nM) was employed to activate increasing concentrations of plasminogen (0–200 nM). Analysis of this data using Lineweaver-Burke plots revealed that β₂GPI and β₂GPI domain V conferred a 19.7-fold and 5.9-fold increase, respectively, in the catalytic efficiency (k_cat/K_m) of t-PA–mediated glu-plasminogen activation (Table 1). The observed increase in catalytic efficiency resulted from a decrease in the K_m (9.8-fold) and enhancement of V_max (2-fold). By comparison, Cesarman et al (17) reported that annexin A2 caused a 63-fold increase in the catalytic efficiency of t-PA-mediated Glu-plasminogen activation, reflecting a 9-fold decrease in the K_m and a 5-fold increase in V_max (17).

Table 1.

Effects of native β₂GPI and β₂GPI domain V kinetics of t-PA–dependent activation of Glu-plasminogen^*

	Km (μM)	Kcat (s−1)	Kcat/Km (μM−1s−1)
Plasminogen + BSA	0.4538	0.1149	0.253
Plasminogen + β2GPI	0.0464	0.2310	4.983
Plasminogen + β2GPI-V	0.3148	0.4718	1.499

^*V_max and K_m were calculated using nonlinear regression. k_cat was calculated as V_max/E_0, where E₀ represents the t-PA concentration (10 nM). V_max was derived from a standard curve of plasmin amidolytic activity.

Neither human albumin nor control human IgG affected t-PA-dependent plasminogen activation. Parallel studies demonstrated that β₂GPI did not affect the ability of single-chain urokinase to activate plasminogen.

β₂GPI stimulates fibrinolysis in a fibrin gel

β₂GPI caused a 2.6 and 3.1-fold increase in lysis of a fibrin gel at concentrations of 1 μM and 5 μM, respectively (Figure 2).β₂GPI-V (5 μM) caused 2.6-fold stimulation of fibrinolysis, though the Gly274-Cys288 β₂GPI-V peptide had no effect. Control proteins, BSA and IgG, also had no activity. While the relative amount of fibrin digestion in this assay seems proportionally small compared to the extent of plasminogen activation in amidolytic assays, the kinetic parameters governing these assays are not comparable.

Figure 2. Native β₂GPI and β₂GPI domain V enhance fibrinolysis in fibrin gels.

Lysis of fibrin gels was assessed as described in Materials and Methods. (A) A representative experiment: (1) 10 nM t-PA + BSA; (2) 10 nM t-PA + 5 μM β₂GPI-V; (3) 10 nM t-PA + 5 μM native β₂GPI. (B) t-PA activity (U/ml) (as determined from the size of lytic areas) in the absence or presence of increasing amounts of native β₂GPI-n, β₂GPI-V, or β₂GPI peptide (β₂GPI-p), calculated from a standard curve.

Effect of plasma β₂GPI depletion on clot lysis

Plasma that had been perfused over an anti-β₂GPI column contained no immunologically-detectable β₂GPI, as determined by immunoblotting (not shown). Plasma from the same donors was perfused over a column containing preimmune (control) rabbit IgG, to control for dilution. Both plasma samples were clotted and their clot lysis rates determined as described in Materials and Methods.

Table 2 depicts the extent of clot lysis using control and β₂GPI-depleted plasma from three donors. Though the maximal A₄₀₅ and extent of lysis varied, the decrease in absorbance between 30 and 180 minutes, reflecting clot lysis (26), was significantly diminished in the absence of β₂GPI (P = 0.0260 when comparing the means of triplicate points).

Table 2.

Effect of β₂GPI depletion on clot lysis

		A405 (T = 30)	A405 (T = 180)	ΔA405	% lysis
Pt 1	Anti-β2GPI	.595	.509	.086	14.4
	Control IgG	.368	.204	.164	44.5
Pt 2	Anti-β2GPI	1.09	.81	.28	25.6
	Control IgG	.92	.47	.45	48.9
Pt 3	Anti-β2GPI	1.29	1.30	−0.01	0
	Control IgG	1.19	.70	.49	41.1

Figure 3A depicts a representative experiment in which endogenous clot lysis rates in control and β₂GPI-depleted plasma is compared. Figure 3B demonstrates that the reconstitution of β₂GPI-depleted plasma with purified β₂GPI (100 and 200 nM final concentrations) led to complete restoration of lytic activity in the β₂GPI-depleted plasmas.

Figure 3. Clot formation and lysis in the absence and presence of β₂GPI.

(A) Plasma from a normal donor was divided into equal aliquots and subjected to chromatography on a column to which either anti-β₂GPI antibodies or preimmune-rabbit IgG had been conjugated. After passing through the former column, plasma was completely immunodepleted of β₂GPI as determined by immunoblotting. The flow-through from both columns was placed in 96 well microplates and clotted as described in Materials and Methods. Clots were monitored by measurement of A₄₀₅. Clotting led to an increase in A₄₀₅ which was maximal at approximately 30 minutes, followed by a progressive decrease as clots lysed. Normal plasma (black diamonds) did not undergo chromatography. (B) The identical experiment as depicted in (A) was performed, but β₂GPI-depleted plasma was analyzed either directly (squares, solid lines) or after the addition of 100 μM β₂GPI (triangles) or 200 μM β₂GPI (inverted triangles). The addition of β₂GPI to the β₂GPI-depleted plasma restored its fibrinolytic activity.

β₂GPI stimulates t-PA amidolytic activity

β₂GPI enhanced the t-PA-dependent cleavage of I-1105, a t-PA specific substrate, in a concentration-dependent manner. A concentration of 0.3 μM β₂GPI caused a 4-fold increase in the rate of substrate hydrolysis (Figure 4A). A stimulatory effect of β₂GPI (1 μM) was also observed in the presence of increasing concentrations of t-PA (0–50 nM) (Figure 4B).

Figure 4. β₂GPI stimulates t-PA amidolytic activity.

(A) t-PA amidolytic activity in the presence of increasing amounts of native β₂GPI and β₂GPI domain V. (B) The effect of 1 μMβ₂GPI on t-PA amidolytic activity in the presence of increasing concentrations of t-PA.

β₂GPI binds to t-PA

β₂GPI bound to t-PA–coated microplates in a concentration dependent manner (Figure A) that was competitively inhibited by fluid phase t-PA (IC₅₀ ~ 310 nM) (Figure 5B). Binding was not inhibited by the β₂GPI domain V-derived peptide (Gly274-Cys288) previously reported to mediate the binding ofβ₂GPI to anionic phospholipid (not shown), suggesting that different molecular interactions facilitate these binding events.

Figure 5. Binding of β₂GPI to t-PA.

(A) Binding of β₂GPI to t-PA using a microplate assay was performed as described in Materials and Methods. (B) Inhibition of the binding of 40 nM β₂GPI to immobilized t-PA by fluid phase t-PA. β₂GPI binding is expressed as the percentage of β₂GPI bound in the presence of fluid-phase t-PA relative to that bound in its absence. (C) Binding of t-PA to β₂GPI using an optical biosensor. The concentration of t-PA used is depicted to the right of the curve. (D) Binding of t-PA to β₂GPI was inhibited by increasing concentrations of NaCl.

Optical biosensor analysis confirmed a specific interaction between β₂GPI and t-PA (Figure 5C). Analysis of real time binding curves yielded a K_ass as 1.13 ± 0.07 × 10⁵ M⁻¹S⁻¹ and K_diss as 6.19 ± 0.10 × 10⁻³ S⁻¹. Accordingly, K_D and K_A were determined to be 2.0 × 10⁻⁸ M and 5.0 × 10⁷ M⁻¹, respectively.

We observed only weak binding between Glu-plasminogen and intact β₂GPI (K_D > 1 μM) using the same optical biosensor system (not shown), consistent with the findings of Lopez-Lira et al (21).

High concentrations of NaCl inhibited the binding between β₂GPI and t-PA, suggesting a strong ionic component to the binding interaction (Figure 5D).

Anti-β₂GPI antibodies inhibit the stimulatory effect of β₂GPI on t-PA-dependent plasminogen activation

Two monoclonal anti-β₂GPI antibodies (1D2 and BD4) inhibited t-PA-dependent plasminogen activation to a significantly greater extent than control murine IgG1 (P<0.0001, n=4) (Figure 6A). Purified IgG fractions from patients with APS also inhibited plasminogen activation in a concentration-dependent manner (Figure 6A, B). To determine whether this effect was associated with anti-β₂GPI antibodies or anti-t-PA antibodies (29), we used two approaches. First, we affinity-purified anti-β₂GPI IgG from three patients, and tested the ability of the purified anti-β₂GPI IgG to inhibit the ability of β₂GPI to enhance t-PA-dependent plasminogen activation. As depicted in Figure 6C, the three affinity-purified anti-β₂GPI antibody preparations inhibited plasminogen activation in a concentration-dependent manner, while a control human IgG preparation (IgG1) had not effect. We subsequently isolated whole IgG fractions from 22 patients with APS whose plasma contained IgG anti-β₂GPI but not anti-t-PA antibodies (β₂GPI⁺/t-PA⁻), and 21 patients whose plasma contained neither anti-β₂GPI nor anti-t-PA antibodies (β₂GPI⁻/t-PA⁻), and measured their ability to inhibit β₂GPI-mediated stimulation of t-PA-dependent plasminogen activation. At a concentration of 80 μg/ml (533 nM), IgG fractions from β₂GPI⁺/t-PA⁻ patients caused significantly more inhibition of plasminogen activation than those from anti-β₂GPI⁻/anti-t-PA⁻ patients or normal controls (Figure 6D).

Figure 6. Anti-β₂GPI antibodies inhibit β₂GPI-mediated enhancement of plasminogen activation.

Plasminogen activation was determined as described in Materials and Methods. Percent inhibition by anti-β₂GPI antibodies was calculated by comparing initial reaction rates in the absence or presence of antibodies. (A) IgG76, IgGC, IgG35 and IgG58 are anti-β₂GPI antibody-containing IgG fractions from patients with APS, BD4 and ID2 are monoclonal anti-β₂GPI antibodies, and IgG1 is normal human IgG. (B) Inhibition of t-PA-dependent plasminogen activation by increasing concentrations of anti-β₂GPI IgG (IgG76), or control IgG (IgG1). (C) Anti-β₂GPI antibodies were affinity purified from the plasma of 3 patients with APS and anti-β₂GPI antibodies (abGPI-11, abGPI-17, abGPI-27), and their ability to inhibit β₂GPI-dependent t-PA-mediated plasminogen activation determined. Each of the purified anti-β₂GPI antibody preparations inhibited plasminogen activation in a concentration-dependent manner. (D) The effect of IgG fractions from 40 normal controls, 22 APS patients with anti-β₂GPI, but not anti-t-PA antibodies (anti-β₂GPI⁺/anti-tPA⁻), or 21 APS patients with neither anti-β₂GPI nor anti-t-PA antibodies (anti-β₂GPI⁻/anti-tPA⁻) on t-PA-dependent plasminogen activation in the presence of β₂GPI. One outlying point in which an IgG fraction from a patient with anti-β₂GPI caused marked inhibition of plasminogen activation (~60% activity) was removed for this analysis.

Discussion

Our studies demonstrate that β₂GPI binds t-PA with high affinity and stimulates t-PA–dependent plasminogen activation. β₂GPI enhanced the catalytic efficiency (k_cat/K_m) of t-PA-dependent plasminogen activation by approximately 20-fold, as a result of a decrease in the K_m and an increase in V_max (Table 1). The enhancement of plasmin generation and stimulation of fibrinolysis by β₂GPI was also demonstrated in a fibrin gel (Figure 4). Moreover, the depletion of plasma β₂GPI led to delayed clot lysis, which was restored upon repletion of purified β₂GPI, suggesting that these findings are relevant in plasma.

The activity of full-length recombinant β₂GPI expressed in 293T cells was similar to that of purified plasma β₂GPI (Figure 1), suggesting that isolated plasma β₂GPI was not denatured, and thus that the activity of β₂GPI was not attributable to non-specific effects of denatured protein (30). The ability of recombinant β₂GPI domain V to stimulate t-PA-dependent plasminogen activation was approximately 50% that of intact β₂GPI, suggesting thatβ₂GPI-V as well as other regions of β₂GPI contribute to this activity. However, theβ₂GPI Gly274-Cys288 peptide neither enhanced t-PA–dependent plasminogen activation nor significantly inhibited the binding of β₂GPI to immobilized t-PA. Thus, if this region is involved in the interaction of β₂GPI with t-PA, it likely requires conformational restraints imposed in the context of intact β₂GPI. Additional studies employing site-directed mutagenesis and β₂GPI mutants lacking specific β₂GPI domains may be of value in further defining the role of this region, if any, in the interactions of β₂GPI with t-PA.

Annexin A2 and the annexin A2-S100A10 heterotetramer are endothelial cell receptors for plasminogen and t-PA (16;31). Binding of t-PA and plasminogen to annexin A2 enhances plasmin generation by facilitating co-assembly of these reactants, and lowering the K_M for their interaction (16;17). Cesarman-Maus et al. have described anti-annexin A2 antibodies in patients with APS that inhibit the fibrinolytic activity of this complex (32). While our laboratory has demonstrated a high-affinity interaction between β₂GPI and annexin A2 (33), we have not yet directly assessed the effects of β₂GPI on the ability of annexin A2 to enhance t-PA-mediated plasminogen activation either in isolation or on cell surfaces. However, Lopez-Lira et al reported an increase in plasmin generation on the surface of a human microvascular endothelial cell line (HMEC-1) following the addition of β₂GPI (21), and we have observed that fluid-phase annexin A2 andβ₂GPI stimulate t-PA-mediated plasminogen activation in an additive manner (not shown). Additional studies will clearly be required to better characterize the effects of β₂GPI on cell surface plasminogen activator activity, and better define the roles of specific receptors in these effects. Such studies will require a full consideration of the complex nature of conformation-dependent interactions of β₂GPI with phospholipids and proteins, such as apoER2 (34), factor XI (35), annexin A2 (33) and Lp(a) (36).

Lopez-Lira et al have reported thatβ₂GPI binds glu-plasminogen and stimulates streptokinase-mediated plasminogen activation (21). Our preliminary studies also suggest low affinity binding (K_d > 1 μM) between glu-plasminogen and β₂GPI, although the nature of the interactions between these two proteins requires further investigation. However, t-PA was not present in the system of Lopez-Lira, and thus our findings extend this work by demonstrating a direct interaction of β₂GPI with t-PA. Taken together, however, these results suggest that in the presence of fibrin, β₂GPI may potentially promote fibrinolysis through multiple pathways including 1) direct stimulation of t-PA amidolytic activity, 2) lowering of the K_M for t-PA-mediated plasminogen activation, and 3) binding to fibrin and providing additional binding sites for plasminogen and t-PA at the fibrin surface (21).

Yasuda et al have reported that nicked β₂GPI, which has been cleaved by plasmin between Lys317-Thr318 in domain 5, but not intact β₂GPI, binds with low affinity to glu-plasminogen (Kd = 0.37 μM) (22). These investigators also observed that nicked β₂GPI suppressed plasmin generation in the presence of t-PA, plasminogen and fibrin. Taken together with our findings, these results suggest that intact β₂GPI may stimulate t-PA-mediated plasminogen activation and subsequently undergo cleavage by newly-formed plasmin. In turn, cleaved (“nicked”) β₂GPI, if generated in sufficient concentrations in plasma, might limit additional plasmin generation (22).

The pathogenic effects of many “antiphospholipid” antibodies may be mediated through interactions with β₂GPI. Indeed, Takeuchi et al observed diminished fibrinolysis in euglobulin fractions from APS patients, attributing this effect to impaired factor XII activation and activity (11).β₂GPI has also been reported to protect t-PA from inhibition by plasminogen activator inhibitor type 1 (PAI-1), an activity blocked by monoclonal antiphospholipid antibodies (12). In this report, we have demonstrated that monoclonal anti-β₂GPI antibodies and affinity-purified anti-β₂GPI antibodies from APS patients inhibit the ability of β₂GPI to enhance t-PA-mediated plasminogen activation in a concentration-dependent manner. Morevoer, IgG fractions from APS patients containing anti-β₂GPI antibodies caused significantly more inhibition of t-PA-mediated plasminogen activation than those from control patients or APS patients without anti-β₂GPI antibodies. This activity was not due to anti-t-PA antibodies (29), and thus appears attributable to anti-β₂GPI antibodies, suggesting another potential mechanism by which heterogeneous APL may contribute to the development of thromboembolic disease (7).