Regulation of intrapleural fibrinolysis by urokinase-α-macroglobulin complexes in tetracycline-induced pleural injury in rabbits

Andrey A. Komissarov, Andrew P. Mazar, Kathy Koenig, Anna K. Kurdowska, Steven Idell

Abstract

The proenzyme single-chain urokinase plasminogen activator (scuPA) more effectively resolved intrapleural loculations in rabbits with tetracycline (TCN)-induced loculation than a range of clinical doses of two-chain uPA (Abbokinase) and demonstrated a trend toward greater efficacy than single-chain tPA (Activase) (Idell S et al., Exp Lung Res 33: 419, 2007.). scuPA more slowly generates durable intrapleural fibrinolytic activity than Abbokinase or Activase, but the interactions of these agents with inhibitors in pleural fluids (PFs) have been poorly understood. PFs from rabbits with TCN-induced pleural injury treated with intrapleural scuPA, its inactive Ser195Ala mutant, Abbokinase, Activase, or vehicle, were analyzed to define the mechanism by which scuPA induces durable fibrinolysis. uPA activity was elevated in PFs of animals treated with scuPA, correlated with the ability to clear pleural loculations, and resisted (70–80%) inhibition by PAI-1. α-macroglobulin (αM) but not urokinase receptor complexes immunoprecipitated from PFs of scuPA-treated rabbits retained uPA activity that resists PAI-1 and activates plasminogen. Conversely, little plasminogen activating or enzymatic activity resistant to PAI-1 was detectable in PFs of rabbits treated with Abbokinase or Activase. Consistent with these findings, PAI-1 interacts with scuPA much slower than with Activase or Abbokinase in vitro. An equilibrium between active and inactive scuPA (kon = 4.3 h−1) limits the rate of its inactivation by PAI-1, favoring formation of complexes with αM. These observations define a newly recognized mechanism that promotes durable intrapleural fibrinolysis via formation of αM/uPA complexes. These complexes promote uPA-mediated plasminogen activation in scuPA-treated rabbits with TCN-induced pleural injury.

  • single-chain urokinase
  • tissue plasminogen activator

fibrinolytic therapy has been used to treat pleural loculation since the 1940s (39, 40) and remains a viable therapeutic option in clinical practice (14, 21). However, the role of fibrinolytic therapy for pleural loculation is undergoing reconsideration. In a large multicenter, randomized, and double-blind clinical trial, administration of intrapleural streptokinase did not improve clinical outcomes. Conversely, a number of clinical reports suggest that fibrinolytic therapy is an effective treatment of pleural loculation (14, 21). Intrapleural urokinase (2-chain form; tcuPA) was reported to be as effective as, and less costly than, video-assisted thoracoscopy for the treatment of childhood empyema (35). The agent now most commonly used for intrapleural administration in the U.S. is single-chain tissue plasminogen activator (sctPA) (14, 21), which is an active zymogen (30, 31). While systemic fibrinolysis does not usually occur after intrapleural administration of fibrinolysins (7, 9, 32), the use of active fibrinolysins such as tPA carries the risk of bleeding (24), and their safety profile remains uncertain (8). A recent meta-analysis of seven randomized controlled trials including a total of 761 patients suggested that fibrinolytic therapy facilitates pleural drainage and mitigates the requirement for surgery in some studies, but not others (8). The variable efficacy and safety profiles of fibrinolysins now available for intrapleural use provide a strong rationale to better understand how these agents work in pleural fluids (PFs) and to explore new pharmacological approaches.

In a series of recent studies, we reported that intrapleural administration of the proenzyme single-chain urokinase plasminogen activator (scuPA) effectively clears intrapleural loculation in rabbits with tetracycline (TCN)-induced pleural injury (15, 16, 18). These findings strongly suggest that the mechanism by which intrapleural scuPA generates intrapleural fibrinolytic activity differs from that of a mature enzyme such as low-molecular-weight (LMW) tcuPA (Abbokinase), and an active zymogen, sctPA (Activase). Commercial fibrinolysins are rapidly inactivated, mainly by plasminogen activator inhibitor-1 (PAI-1), in exudative PFs (17, 29). Since it is known that a number of proteinases form active complexes with α-macroglobulins (αM) in humans (2, 6, 36) and in rabbits (1, 13, 38), we hypothesized that uPA/αM complexes contribute to durable PA activity after intrapleural administration of scuPA.

In this study, mechanisms by which enzymatic activities of scuPA and commercially available fibrinolysins are regulated in PFs of rabbits with TCN-induced pleural injury are elucidated and compared. An advantage of this strategy is that the intrapleural processing of these agents can be assessed in the same temporal and pathophysiological context. We now report that intrapleural administration of scuPA selectively forms uPA complexes with αM, that these complexes are able to activate plasminogen, and that scuPA-generated αM but not scuPA-soluble uPAR complexes foster protracted intrapleural fibrinolytic activity. We demonstrate that the mechanism of inactivation of scuPA by PAI-1 differs from that for Activase and Abbokinase and that effective formation of scuPA generated αM/uPA complexes in vivo appears to be facilitated by an equilibrium between active and inactive species of the proenzyme.

MATERIALS AND METHODS

Proteins and reagents.

scuPA was provided as a generous gift by Dr. Jack Henkin (Abbott Laboratories, Chicago, IL). S356A recombinant catalytically inactive, endotoxin-depleted scuPA was generated and purified, as previously described (18). Lack of activity of the mutant scuPA was confirmed in the presence of plasmin activation using the chromogenic substrate S2444 (data not shown). Commercially available proteolytically active fibrinolysins included Abbokinase, the LMW active form of urokinase, uPA (Abbott Laboratories, Abbott Park, IL), and Activase (active recombinant tPA; Genentech, San Francisco, CA). Wt- and S338C (P9 Cys) PAI-1 was purified, labeled with N-{[2-(iodoacetoxy) ethyl]-N-methyl} amino-7-nitrobenz-2-oxa-3-diazole (NBD), and characterized as previously described (19, 20, 34). High-molecular-weight (HMW) tcuPA and HMW tcuPA activity standard (100,000 IU/mg) were obtained from Abbott Laboratories and American Diagnostica (Stamford, CT), respectively. Human recombinant soluble urokinase receptor (suPAR) was obtained from Attenuon (San Diego, CA). Human plasma α2-macroglobulin (α2M) was purchased from American Diagnostica or Biodesign (Saco, Maine). Plasminogen, plasmin, and fluorogenic plasmin substrate were purchased at Haematologic Technologies (HTI, Essex Junction, VT). A specific synthetic fluorogenic uPA substrate, tPA substrate, and chromogenic plasmin substrates were from Centerchem (Switzerland). The concentrations of proteins were calculated either from absorbance at 280 nm (using Mr of 54,000, 43,000, and 63,500, and ε280 of 1.36, 0.93, and 1.90 ml·mg−1·cm−1, for scuPA, PAI-1, and tPA, respectively) or from results of measurements with BCA protein assay kit (Pierce). All experiments were carried out either in 50 mM phosphate (pH 7.4) or in 0.05–0.1 M HEPES/NaOH buffer (pH 7.4, 20 mM NaCl).

PF samples from rabbits treated with intrapleural scuPA, tcuPA, or sctPA.

The studies described in this report were approved by the Animal Review Committee of The University of Texas Health Science Center at Tyler. PFs used in this study were from female 3.5–4 kg New Zealand white rabbits from Myrtle's Rabbitry (Thompson Station, TN) with TCN-induced pleural injury subsequently treated with intrapleural fibrinolysins described in the previous section and as we previously reported (16). Fresh, previously unthawed aliquots of PFs obtained from these animals (16) stored at −80°C were used in this study. Briefly, pleural injury was induced by intrapleural administration of TCN into the right pleural space, and the rabbits were allocated to the following groups in which pleural fluids were harvested at 96 h after intrapleural TCN, at which time loculation was assessed. The groups included rabbits treated with a clinical range of concentrations of commercially available uPA or tPA as well as scuPA and controls. The PFs we used in this study were obtained from 1) animals treated with a single dose of normal sterile saline, the vehicle for scuPA and other interventional agents, at 72 h after intrapleural TCN, as previously described (16); 2) animals treated with a single dose of scuPA (0.5 or 0.25 mg/kg) at 72 h after intrapleural TCN; the higher scuPA dose effectively cleared pleural loculations, whereas the lower dose did not (15, 16); 3) animals treated with commercially available LMW two-chain (tc) uPA (Abbokinase, Abbott Laboratories) at a low dose (1,429 IU/kg, approximating a clinical intrapleural unit dose of 100,000 IU in a 70-kg subject); 4) animals treated with Abbokinase at a high dose (3,577 IU/kg, approximating an intrapleural unit dose of 250,000 IU for a 70-kg subject); 5) animals treated with commercially available recombinant single-chain tPA (Activase, Genentech) at a low dose (0.1425 mg/kg, approximating a clinical intrapleural unit dose of 10 mg/70-kg subject); 6) animals treated with a single administration of high-dose Activase (0.5 mg/kg, equivalent to the scuPA effective dose; n = 5); and 7) animals treated with the recombinant catalytic site-inactivated scuPA mutant. These and all PFs were obtained from rabbits used in our previous report (16).

Measurement of uPA, tPA, and plasmin amidolytic activity.

Amidolytic uPA and tPA activity was determined from time traces of changes in the fluorescence emission at 440 nm (excitation 344 nm) of fluorogenic uPA and tPA substrates (Pefafluor uPA and tPA, respectively, Centerchem) in 0.1 M HEPES buffer (pH 7.4, 20 mM NaCl). Aliquots (5–20 μl) of samples were mixed with buffer (total vol 50 μl) in 96-well white flat-bottom Costar plates (Corning, NY). Equal volumes of 0.1 mM Pefafluor uPA or 0.2 mM Pefafluor tPA in the same buffer were added to each well and mixed. Increases in fluorescence emission were detected using a Varian Cary Eclipse fluorescence spectrophotometer equipped with a 96-well plate reader accessory (Varian, IL). A linear equation was fit to the results using Varian software. Plasmin activity was measured using either chromogenic (Centerchem) or fluorogenic (HTI) substrates. Samples were analyzed in 96-well flat-bottom plates, either Pro-Bind #353915 (Becton Dickinson, Franklin Lakes, NJ) or Costar, respectively. Plasmin activity was detected via changes in absorbance at 405 nm using a SpectraMax 96-well optical absorbance plate reader (Molecular Devices, Sunnyvale, CA) or via an increase in the fluorescence emission at 470 nm (excitation at 352 nm) using a Varian Cary Eclipse fluorescence spectrophotometer. Enzymatic activities were reported as arbitrary units (AU) that were calculated based on tcuPA (American Diagnostica), Abbokinase (Abbott Laboratories), or Activase (Genentech) standards of known activity.

Effects of human recombinant suPAR and human plasma α2M on inactivation of uPA by PAI-1.

scuPA, HMW tcuPA, or Abbokinase (LMW tcuPA) (30-100 nM) were incubated with either 100–250 nM suPAR (Attenuon) or with 0.5–2.5 μM α2M from human plasma (Meridian Life Science, Saco, ME) in 0.05 M HEPES/NaOH buffer (pH 7.4) at 37°C. Aliquots were withdrawn at 0, 2, and 7 h, and uPA amidolytic activity was determined using Pefafluor uPA before and after mixing of the aliquot with a 5–10 molar excess of PAI-1. At the end of the incubation period, complexes of uPA with suPAR and α2M were isolated by immunoprecipitation (IP) with 20 μg of MAb ATN-658 (Attenuon; ATN-658 does not affect amidolytic activity of uPA complexed with suPAR) or with 20 μg of polyclonal sheep anti-human α2M antibodies (Meridian Life Science), respectively. Antibodies and their complexes with antigens were precipitated with 40 μl of 1:1 slurry of protein A/G Plus Agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA). Samples were incubated with Agarose for 1–2 h at room temperature on a rocker and then centrifuged at 3,000–6,000 rpm using an AccuSpin Micro R centrifuge (Thermo Fisher, Freemont, CA). The resin was washed three times with cold HEPES buffer (0.6 ml), and 50 μl of the resin suspension (total vol 200–400 μl) was transferred to 96-well plates and used for measurements of enzymatic activity (with or without added excess of PAI-1, 80–200 nM).

Addition of exogenous enzymes to PFs and immunoprecipitation of enzymes complexed with αM and suPAR.

Immunoprecipitation of uPA or tPA-α2M complexes were done from PFs of rabbits treated with scuPA, Activase, Abbokinase, or with a control vehicle (sterile PBS or normal sterile saline), or from PFs supplemented with exogenous enzymes, PAI-1, or both. PFs were incubated with or without exogenous enzymes (50–100 nM scuPA, tcuPA, sctPA) for 0–18 h at 37°C in 100–200 μl of 50 mM HEPES/NaOH, pH 7.4. Aliquots of the reaction mixtures were withdrawn at 0, 2, 7, and 18 h, after which uPA and tPA activity were measured in the samples. To estimate the activity of α2M/enzyme complexes, enzymatic activity was measured in the presence of an excess of PAI-1 (final concentration in the reaction mixture 80–200 nM). At the end of the incubation, 20 (PFs with supplementation with enzymes) or 50 μg (native PFs) of anti-rabbit α2M (1.0 mg/ml) or anti-human α2M (experiments with pure α2M) IgY (1.0 mg/ml) was added, and reaction mixtures were incubated for 0.5–1.0 h at 4°C. Then 40 (PFs with supplementation with enzymes) or 80 μl (native PFs) of a slurry containing Agarose with immobilized anti-IgY goat polyclonal IgG (ICL, Newberg, OR) was added to each sample and incubated overnight (10 h) at 4°C on a rocker. Complexes bound to the resin were isolated, and uPA activity was measured with and without addition of 80–200 nM of PAI-1 as described in the previous section.

Both endogenous human uPA/rabbit uPAR and exogenous uPA/suPAR (10–50 nM) complexes were precipitated from PFs with anti-uPAR MAb #3936 (American Diagnostica), which has been shown to cross-react with rabbit uPAR (15). In control experiments, endogenous α2M/uPA complexes were precipitated from PFs using polyclonal sheep anti-human α2M antibodies (Meridian Life Science). Amidolytic uPA activity of the precipitated antibody/antigen complexes bound to the resin was measured in the presence or absence of excess of exogenous PAI-1.

Immunoblotting of uPA associated with α2M.

α2M/uPA complexes isolated from rabbit PFs and from the reaction between human α2M and tc- or scuPA were subjected to SDS-PAGE (4–12% gradient gel; NuPage Invitrogen) using the XCell SureLock Mini-Cell (Invitrogen). Proteins were transferred to a polyvinylidene difluoride membrane (Invitrogen) using the XCell II Blot Module (Invitrogen). A Chemiluminescent Western Blot Immunodetection Kit WesternBreeze (Invitrogen) with mouse monoclonal primary antibody #3689 (American Diagnostica) at a dilution of 1:5,000 was employed for uPA detection. Nonspecific mouse IgG was employed as a control for the primary antibody. The image was developed using anti-mouse goat IgG conjugated with alkaline phosphatase (AP). α2M was detected using anti-human α2M IgY conjugated with horseradish peroxidase (dilution 1:5,000) (IgY-HRP; Immunology Consultants Laboratory, Newberg, OR) and Immobilon Western Chemiluminescent HRP substrate (Millipore, MA). The membranes were incubated with chemiluminescent substrate (AP or HRP) and exposed to HXR film (Hawkins X-Ray Supply). Images were scanned using an HP 3210 scanner.

Kinetics of the reaction between NBD P9 PAI-1 and scuPA, tcuPA, Abbokinase, or Activase.

Stopped-flow fluorimetry was employed to examine kinetics of the reaction between proteinases and PAI-1. The rate of insertion of the reactive center loop (RCL) of PAI-1 which drives the reaction with the proteinase was monitored through the increase in the fluorescence emission of the NBD group attached to a cysteine residue in the Ser338Cys mutant variant of PAI-1 (NBD P9 PAI-1) (34). Time-dependent inactivation of scuPA, Abbokinase (LMW tcuPA), Activase, and HMW tcuPA was measured by incubating various concentrations of the enzyme (0.1–5.5 μM) with 20–40 nM NBD P9 PAI-1. A microvolume stopped-flow reaction analyzer (model SX-20, Applied Photophysics) equipped with a fluorescence detector (excitation at 490 nm, emission through 515 nm cutoff filter) was employed to monitor an increase in NBD fluorescence emission due to the reaction with the enzymes. Traces of changes in the fluorescence emission were measured in 50 mM phosphate buffer, pH 7.2, at 25°C. The values of observed first-order rate constants (kobs) were determined directly from the stopped-flow traces by fitting a double exponential equation, Ft = F + Amaxe−(kobs1) t + Amine−(kobs2) t, to observed time-dependent increase in the NBD fluorescence emission [Ft and F are current and final values of the NBD fluorescence emission, respectively; Amax and Amin are amplitudes (Amax/Amin > 7); and kobs1 and kobs2 are the first-order rate constants]. Pro-Data viewer software (Applied Photophysics) was used for fitting; the quality of the fit was estimated by visual analysis of plots of the residuals (deviation of the fitted function from the data). The values of kobs1, corresponding to the major increase in the NBD fluorescence emission, were plotted against proteinase concentration using SigmaPlot 11.0 (SPSS), and plots were fit by either a linear (kobs = klim for scuPA) or a hyperbolic {kobs1 = klim*[E]/(Km + [E]) for tcuPA, Activase, Abbokinase} equation, where klim is the first-order limiting rate constant of RCL insertion, and Km is the proteinase concentration ([E]) at half of saturation (kobs1 = klim/2).

Data analysis and statistics.

Amidolytic enzyme activity and stopped-flow fluorescence traces were analyzed employing Varian Eclipse Kinetic Software and Pro-Data Viewer, respectively. The quality of the fit was estimated by visual analysis of plots of the residuals (deviation of the fitted function from the actual data). Nonlinear least squares fitting employing the Levenberg-Marquardt algorithm (SigmaPlot 11.0 for Windows; SPSS) was used for calculation of the kinetic parameters. The values of kobs (average of 5–10 measurements; SE less than 10%) were plotted against enzyme concentration. Plots were fit by either hyperbolic or linear functions as described above. Correlation coefficients (r) calculated from curve fittings were also used as a parameter of goodness of fitting (r2 of the fit was greater than 0.98 for all the kinetic data). Statistical significance, as well as medians and values at 75 and 25% intervals, were determined using the Mann-Whitney Rank Sum Test as described previously (16). P < 0.05 was considered statistically significant. Box plots of the results were generated as previously described (16) using SigmaPlot 11.0 and SigmaStat 3.5 software (SPSS).

RESULTS

PFs of rabbits treated with intrapleural scuPA demonstrate increased uPA activity vs. those of rabbits treated with commercial fibrinolysins at clinically applicable doses.

Amidolytic uPA and tPA activities were measured in PFs collected from animals treated with the different PAs and in the samples from the control group in which rabbits received intrapleural injection of vehicle (Fig. 1). Saline and PBS vehicle controls all yielded comparable low levels of uPA activity and are illustrated as combined vehicle controls in the figure. Maximal uPA activity was found in the PFs of animals treated with a therapeutically effective dose (ED) of scuPA [correlated with nearly complete clearance of intrapleural loculation (16); 0.5 mg/kg, median: 2.8, range 0.7–7.8 AU, P < 0.05 vs. all other groups] (Fig. 1). Significantly lower levels of uPA activity were detected in PFs of animals that received half (0.25 mg/kg) of the effective intrapleural dose of scuPA, which did not clear intrapleural loculations (15). uPA activity levels were also significantly increased in PFs of ED scuPA-treated rabbits vs. those treated with either high- or low-dose intrapleural Abbokinase or the inactive S195A scuPA mutant. These treatments likewise did not clear intrapleural loculation induced by TCN (15, 16). In contrast to uPA, due to the low specificity of the tPA fluorogenic substrate, tPA activity was detectable in all of the PFs, with the highest activity in the samples treated with a dose of intrapleural scuPA (data not shown).

Fig. 1.

Amidolytic urokinase plasminogen activator (uPA) activity in pleural fluids (PFs) of rabbits with tetracycline (TCN)-pleural injury treated with intrapleural single-chain (sc)uPA, Abbokinase, or Activase. Amidolytic uPA activity was measured in PFs as described in materials and methods. Activity is expressed as slopes of the linear dependencies of changes in the fluorescence emission at 440 nm (excitation 340 nm) of 50 μM fluorogenic substrate Pefafluor uPA (Centerchem) in 0.05 M HEPES/NaOH buffer pH 7.4. uPA activity was significantly increased in the pleural fluids of rabbits with TCN-induced injury treated with an effective dose (ED) of intrapleural scuPA vs. all other groups (P < 0.05), including rabbits treated with intrapleural Abbokinase [Abb; at low (LD) or high (HD) dose] or Activase (Act; LD or HD). The 0.5 ED of scuPA = 0.25 mg/ml did not effectively clear intrapleural loculation (16). S195A, inactive mutant variant of scuPA. uPA activity is indicated as arbitrary units (AU). Data are illustrated in a box plot format in which the 25 and 75% quartiles are indicated in the box, and median values are shown as horizontal lines within the box, as previously described (16).

uPA activity in PFs from scuPA-treated rabbits is protected from inactivation by PAI-1.

To determine the susceptibility of uPA and tPA activity to inhibition by PAI-1, 200 nM exogenous PAI-1 was added to the PFs of rabbits treated in vivo with either intrapleural scuPA, Abbokinase, or Activase before the addition of the fluorogenic uPA substrate. Elevated uPA activity found in the PFs from rabbits treated with intrapleural scuPA was markedly resistant to inhibition by PAI-1 (Fig. 2). While most of the uPA activity of PFs of rabbits treated with scuPA was resistant to PAI-1, a proportionate fraction, 20–30%, was susceptible to PAI-1. To determine if the therapeutic effect of scuPA correlates with the fraction of uPA activity resistant to inhibition by PAI-1, the activities in the samples of PFs of animals treated with either ED or half of the therapeutic dose (0.5 ED, which does not clear loculations, Ref. 18) of scuPA were also compared (Fig. 2). Total uPA activity and the fraction susceptible to PAI-1 were significantly increased in PFs of rabbits that received the ED vs. 0.5 ED of intrapleural scuPA.

Fig. 2.

PFs of rabbits treated with intrapleural scuPA exhibit uPA activity that resists inhibition by plasminogen activator inhibitor-1 (PAI-1). The effects of exogenous PAI-1 on uPA amidolytic activity in the PFs of animals treated with either an effective therapeutic dose (ED) of scuPA or an ineffective dose (ED 0.5) are illustrated. Amidolytic uPA activity was determined before (gray) and after (white) addition of an excess of PAI-1 as described in materials and methods. Vehicle: PFs from PBS vehicle-treated controls for the scuPA HD- or scuPA LD-treated rabbits. Data are illustrated in the box plot format as described in Fig. 1 legend. While uPA activity in the presence of PAI-1 was lower than that without serpin in all samples, there was no statistical difference between the groups lacking PAI-1 (gray) vs. those supplemented with excess PAI-1 (white boxes, P ≥ 0.1). *Statistically significant differences (P < 0.05) between corresponding groups.

scuPA or tcuPA but not sctPA in equivalent exogenous concentrations generate enzymatic activity in PFs that is protected from inhibition by PAI-1.

To determine whether or not fibrinolysins added to pleural fluids in vitro are similarly protected from inactivation by PAI-1, 50 nM of either scuPA, sctPA (Activase), or tcuPA were incubated for 0–18 h at 37°C with PFs from rabbits treated with the same respective enzyme in vivo. The effects of these agents were also determined in vitro in PFs obtained from rabbits treated with intrapleural normal saline (vehicle). Amidolytic uPA and tPA activity was then measured over time and after the addition of an excess of active PAI-1 at the end (18 h) of the incubation period (Fig. 3) to detect enzymatic activity resistant to PAI-1. The activity of scuPA was low immediately after addition to the pleural fluids (0 time interval) and increased significantly 2 h later (Fig. 3A), reflecting conversion of at least part of the scuPA to the mature, more active two-chain enzyme (3, 23, 28). While uPA activity decreased between 2 and 18 h of incubation in both scuPA-treated and control vehicle-treated PFs, the resulting activity at 18 h was almost completely insensitive to PAI-1 (Fig. 3C). The enzymatic activity of Activase was conversely maximal at the beginning of the incubation and slowly decreased with time (Fig. 3B). In contrast to scuPA, PAI-1 added at the end of incubation neutralized nearly all of the Activase enzymatic activity. These observations show that exogenous single- and two-chain forms of uPA can form complexes resistant to PAI-1 when added to rabbit pleural fluids in equivalent concentrations, whereas tPA does not. The same trends are observed in PFs from rabbits treated with the added enzyme or in untreated animals, but the readout enzymatic activity is, as anticipated, augmented in the PFs of rabbits treated with the same enzymes in vivo.

Fig. 3.

Time-course evidence that exogenous scuPA (A) generates uPA activity that is resistant to PAI-1, that tPA (B) does not generate PAI-1-resistant tPA activity, and that tcuPA in equivalent dose (C) can generate PAI-1-resistant uPA activity in PFs. Box plots of time dependencies of the activity of the mixtures of (A) 50 nM scuPA with PFs of a rabbit treated either with the effective dose of scuPA (gray) or saline control vehicle (white); randomly selected samples per group are n = 7 and 3, respectively. B: 50 nM single-chain tPA (sctPA; Activase) with PFs of animals treated either with high dose of Activase (gray) or saline vehicle (white); n = 5 and 3, respectively, randomly selected samples/group. C: 50 nM tcuPA added to PFs of animals treated with high dose of Abbokinase (LMW tcuPA; gray) or with buffer (white); n = 3 randomly selected samples/group. Mixtures were incubated at 37°C, in 0.05 M HEPES/NaOH, pH 7.4, for 18 h. At the end of the incubation, an excess of PAI-1 (200 nM) was added to each reaction mixture to determine the fraction of the amidolytic activity that is resistant to PAI-1. The data are illustrated using the box plot format described in Fig. 1 legend. *Significant differences (P ≤0.02) between the groups.

To simulate conditions likely to prevail in the pleural space of animals with TCN injury where PAI-1 is continuously replenished, the experimental protocol shown in Fig. 3 was next repeated with the modification that excess exogenous active PAI-1 was added to the pleural fluids immediately after the addition of 50 nM scuPA, sctPA, or tcuPA (Fig. 4). To mitigate the effects of in vivo pretreatment with the same enzymes, PFs used in these experiments were obtained from control vehicle-treated animals with TCN-induced pleural effusions (n = 3, 4, 5/group for tcuPA, scuPA, and sctPA, respectively). In the presence of supplemental active PAI-1, scuPA generated relatively more uPA activity over the 18-h time course. An early decrement in scuPA activity at the 2-h interval suggests that a proportion of the “active” form of scuPA, which possesses relatively low activity, was in part neutralized by PAI-1. Conversely, the concurrent addition of excess PAI-1 to PFs to which equivalent concentrations of either Activase or tcuPA were added did not yield appreciable enzymatic activity resistant to PAI-1 (Fig. 4). While the data shown in Fig. 4 were obtained with control PFs (animals treated with vehicle), similar trends were observed with samples from animals treated with each intrapleural fibrinolysin (data not shown). It is notable that uPA activity in the scuPA-PAI-1-treated samples increased with time during the incubation, and the final values of uPA amidolytic activity were actually comparable to that observed without addition of excess PAI-1 (Figs. 3A and 4). Incubation of PAI-1 at pH 7.4 with a twofold excess of scuPA for 2 h at 37°C resulted in only partial (70–80%) inactivation of the serpin due to slow interaction with scuPA and spontaneous transition to inactive latent conformation (data not shown).

Fig. 4.

Effect of concurrent administration of excess PAI-1 on activity of 50 nM exogenous scuPA, sctPA, or tcuPA added to PFs of control vehicle-treated animals. Box plots as described in Fig. 1 legend illustrate time dependencies of the activity of the mixtures of 100 nM PAI-1 (added to the PF immediately after the enzyme) and 50 nM exogenous sctPA (white boxes, n = 5), tcuPA (gray boxes; n = 3), and scuPA (hatched boxes; n = 4) to PFs vehicle-treated rabbits. Mixtures were incubated at 37°C in 0.05 M HEPES/NaOH, pH 7.4, for 18 h. Aliquots (6 μl) were withdrawn at the indicated times to measure the enzymatic activity using a fluorescence spectrophotometer. At the end of the incubation, an excess of PAI-1 (200 nM) was added to each reaction mixture to determine the fraction of the amidolytic activity that is resistant to PAI-1. There was a statistically significant (P ≤0.03) increase in uPA activity resistant to PAI-1 seen only in the PFs to which scuPA was added at 18 h after incubation (*).

Human plasma α2M protects uPA from inactivation by PAI-1, whereas human recombinant suPAR does not.

We next used an immunoprecipitation strategy to confirm that scuPA generates complexes with αM that resist inactivation by PAI-1 and compared the ability of suPAR to likewise form such complexes (Fig. 5). After 7 h of incubation, the reaction mixtures were immunoprecipitated with a polyclonal antibody to sheep anti-human α2M or with the anti-uPAR MAb ATN-658 (Fig. 5). As expected, scuPA, HMW, or LMW tcuPA formed complexes with α2M that resist inhibition by a 10-fold excess of PAI-1. Conversely, uPA activity in suPAR complexes with scuPA (Fig. 5) or HMW tcuPA (data not shown) that were isolated by immunoprecipitation was completely inactivated by PAI-1. The immunoprecipitation of α2M-uPA complexes that resist PAI-1 was independently confirmed using a chicken anti-rabbit α2M IgY antibody (data not shown).

Fig. 5.

Incubation of uPA with human plasma α2-macroglobulin (α2M), but not with human recombinant suPAR, results in protection from inactivation by PAI-1. scuPA (50 nM) was incubated with 150 nM suPAR, and 50 nM scuPA, HMW tcuPA, or Abbokinase (LMW tcuPA) were incubated with 2.0 μM α2M from human plasma in 0.05 M HEPES/NaOH buffer (pH 7.4) at 37°C for 7 h. At the end of incubation, complexes of uPA with suPAR or α2M were isolated from each reaction mixture by IP with 20 μg of MAb ATN-658 (ATN-658 does not affect amidolytic activity of uPA complexed with suPAR) and with 20 μg of polyclonal sheep anti-human α2M antibodies. Reaction mixtures were incubated with 40 μl of 1:1 slurry of protein A/G Plus Agarose for 2 h at room temperature on a rocker. Reaction mixtures were centrifuged at 3,000–6,000 rpm using an AccuSpin Micro R centrifuge. The resin was washed 3 times with 0.6 ml of cold HEPES buffer, and 50 μl of the resin suspension (total vol 200 μl) was transferred to 96-well plates. Amidolytic uPA activity in precipitated samples was measured with (white bars) or without (gray bars) 80 nM PAI-1. The composition of each reaction mixture, antibodies used for IP, and presence of PAI-1 are shown in the table under the plot.

Exogenous scuPA generates uPA amidolytic activity in PFs via complexes with αM but not with suPAR.

To determine whether or not exogenous scuPA forms complexes with rabbit αM in vivo, α2M was precipitated from PFs treated with supplemental exogenous scuPA using chicken anti-rabbit α2M and nonspecific IgY as described in materials and methods (Fig. 6A). α2M immunoprecipitated from samples of PFs incubated with scuPA contained uPA amidolytic activity, which resisted exogenous PAI-1 (Fig. 6A). Immunoprecipitation with IgY specific to rabbit α2M was next done to test the possibility that αM from PFs of animals treated with an effective dose of scuPA possess relatively greater uPA amidolytic activity than PFs from rabbits treated with Abbokinase or Activase in vivo. The results of experiments illustrated in Fig. 6B confirm that PFs from scuPA-treated animals have relatively more uPA activity that can be immunoprecipitated with an antibody to α2M compared with the PFs from rabbits treated with other intrapleural fibrinolysins. The level of uPA activity in precipitates with nonspecific IgY was low for every PF studied (Fig. 6B).

Fig. 6.

Incubation of scuPA with rabbit PFs forms αM complexes possessing uPA amidolytic activity that resists PAI-1. A: IP of uPA activity by anti-rabbit α2M IgY from PFs of rabbits treated with intrapleural ED of scuPA after preincubation with 50 nM scuPA for 5 h at 37°C in 50 mM HEPES/NaOH, pH 7.4. Samples were immunoprecipitated with α2M IgY or control nonspecific IgY, and uPA amidolytic activity was measured with (white) or without (gray) added exogenous PAI-1 (80 nM) as described in materials and methods and in Fig. 5. uPA amidolytic activity was significantly (*P < 0.05) increased in both groups of PFs from rabbits IP with anti-rabbit α2M vs. those IP nonspecific IgY. B: bar plot of the uPA activity in the precipitates from PFs of animals treated with the ED of scuPA or HD of Activase (Act), Abbokinase (Abb), or vehicle (n = 2 randomly selected samples/each paired group). The precipitation was carried out with 40 μg of either anti-rabbit α2M (filled bars) or nonspecific IgY (hatched bars). The complexes were captured with 80 μl of Agarose slurry and analyzed as described in Fig. 5. C: anti-α2M but not anti-uPAR antibodies precipitate uPA activity from PFs of rabbits, treated with ED of scuPA. IP of uPA activity by polyclonal sheep anti-human α2M antibodies and anti-human uPAR MAb (cross-reacting with rabbit αM and rabbit uPAR, respectively) from PFs of rabbits treated with ED of scuPA. PFs were incubated with 40 μg of antibodies for 2 h in 50 mM HEPES/NaOH, pH 7.4, either with anti-uPAR MAb or with anti-α2M antibodies. In control experiments, PFs were supplemented with exogenous mixtures of 150 nM suPAR with either 50 nM HMW or LMW tcuPA before anti-uPAR MAb was added. Another set of controls contained PFs without addition of antibodies. The composition of each reaction mixture and antibodies used for IP are shown in the table under the plot. Complexes were precipitated with 40 μl of Protein A/G Plus Agarose, and uPA amidolytic activity was measured with (white) or without (gray) 80 nM PAI-1 added, as described in Fig. 5 legend. uPA amidolytic activity was significantly (P < 0.05) increased in PFs incubated with sheep anti-human α2M antibodies and samples supplemented with HMW tcuPA complexed with suPAR (activity without PAI-1) vs. all other samples.

To test the possible contribution of rabbit suPAR/uPA complexes to uPA-related amidolytic activity in PFs of animals treated with ED scuPA, samples were immunoprecipitated with an anti-uPAR MAb that cross-reacts with rabbit uPAR (16). α2M/uPA complexes were precipitated from the same samples using sheep anti-human α2M polyclonal antibodies (which also cross-react with rabbit α2M). While PAI-1 resistant complexes were precipitated with anti-α2M antibodies, no uPA activity was precipitated with the anti-uPAR MAb (Fig. 6C). In control experiments, uPAR was precipitated from the same PFs supplemented with exogenous human suPAR complexed either with HMW tcuPA or with Abbokinase (LMW tcuPA, which lacks the ability to bind suPAR). As expected, only HMW tcuPA bound to suPAR was precipitated, but unlike α2M complexes, all of the uPA activity was neutralized by PAI-1 (Fig. 6C).

Western blot analysis of proteins immunoprecipitated from PFs of animals treated with intrapleural ED scuPA directly confirmed formation of αM/uPA-positive HMW complexes (Fig. 7). PFs were also supplemented (Fig. 7) with 50 nM exogenous scuPA (37°C over 4 h), and the reaction mixtures were immunoprecipitated either with anti-rabbit α2M or with nonspecific IgY. As shown in Fig. 7A, most α2M-positive bands were found at the level of 260 kDa and higher, which corresponds to a mixture of monomers and oligomers of αM. Bands positive for both α2M and uPA, which migrated faster than 180–200 kDa, likely reflect proteolytic degradation of α2M and α2M/uPA complexes (Fig. 7). Some nonspecific binding at the same level (∼120–140 kDa) was detected in both anti-α2M and anti-uPA developments. Notably, all uPA-positive bands (Fig. 7B) migrated much more slowly than the 50-kDa scuPA (Fig. 7B), and uPA-positive bands ≥260 kDa comigrated with those related to α2M, consistent with uPA-αM complex formation.

Fig. 7.

Western blot analysis of α2M (A) and uPA (B) antigens in complexes immunoprecipitated from PFs of rabbits treated with intrapleural scuPA. Western blot analysis of proteins precipitated by anti-rabbit α2M (lanes 1–3) or nonspecific (lane 4) IgY from PFs of rabbits treated with ED scuPA. PFs were incubated with (lanes 2 and 4) or without (lanes 1 and 3) exogenous scuPA (50 nM) and precipitated with anti-rabbit α2M IgY as described in materials and methods and Fig. 5 legend. Samples of the resin were heated with SDS loading buffer (100°C, 2 min) and subjected to a 4–12% gradient SDS-PAGE (NuPage, Invitrogen) under nonreducing conditions. Positions of molecular weight markers (Novex, Invitrogen) are indicated to the left of the gel to allow assessment of the molecular weights of the bands. Lane 5: scuPA standard. Proteins were transferred to the PVDF membrane (Invitrogen), and membranes were incubated with anti-uPA MAb and developed with anti-human-α2M IgY-HRP conjugate (A) using ECL substrate (Pierce) and with goat anti-mouse IgG-alkaline phosphatase conjugate (B) using WesternBreeze (Invitrogen). Each set of 5 lanes (A and B) represents data obtained from the same gel.

PAI-1 interacts with scuPA orders of magnitude slower than with tcuPA, Abbokinase, or Activase.

To better understand the ability of scuPA to resist inactivation by PAI-1, we next sought to explore the mechanism by which scuPA resists rapid inactivation by PAI-1 in PFs. To identify the mechanism of interaction between scuPA and PAI-1, the kinetics of the reaction between NBD P9 PAI-1 and scuPA was compared with that for HMW tcuPA, Abbokinase (LMW tcuPA), and Activase (sctPA) (Fig. 8). The rate of the reaction between PAI-1 and HMW tcuPA, Abbokinase, or Activase was found to be faster with an increase in the enzyme concentration (Fig. 8). The dependencies of the observed rate constant (kobs) on the concentration of the enzyme for HMW tcuPA, Abbokinase, and Activase were hyperbolic, reflecting a two-step interaction between the enzyme and NBD P9 PAI-1 (Fig. 8). The values of klim and Km for HMW tcuPA, Abbokinase, and Activase were 23.2 ± 1.1, 20.1 ± 3.3, and 3.1 ± 0.2 s−1, and 3.3 ± 0.3, 4.9 ± 1.0, and 1.9 ± 0.2 μM, respectively. In contrast to tcuPA and sctPA, there was no significant change in kobs (kobs = 0.0012 ± 0.0006 s−1; Fig. 8), whereas the concentration of scuPA was increased 25 times, from 0.2 to 5.0 μM. Since less than a 0.07 molar equivalent of PAI-1 was required for complete inhibition of initial amidolytic activity of scuPA (data not shown), we reasoned that scuPA, unlike tcuPA, Activase, and Abbokinase exists in equilibrium between active and inactive forms, which favors inactive species of the proenzyme (kon = kobs).

Fig. 8.

Different mechanisms of interaction of PAI-1 with scuPA vs. Activase, Abbokinase, and tcuPA. Dependencies of the observed first order rate constant (kobs) for the reaction of NBD P9 PAI-1 on concentration of scuPA (●), HMW tcuPA (■), Abbokinase (HMW tcuPA) (○), and sctPA (Activase) (▴). Values of kobs were calculated by fitting a double exponential equation to the time traces of changes in NBD-fluorescence emission. The scuPA tracing corresponds to the best fit of a linear equation kobs = 0.0012 s−1. The lines for HMW tcuPA, Abbokinase, and Activase correspond to the best fits (r2 > 0.99) of a hyperbolic equation kobs = klim*[enzyme]/(Km + [enzyme]) to the data shown. The values of klim (the limiting rate of RCL insertion) and Km (concentration of the enzyme at kobs = klim/2) for tcuPA, Abbokinase, and Activase were 23.2 ± 1.1, 20.1 ± 3.3, and 3.1 ± 0.2 s−1, and 3.3 ± 0.3, 4.9 ± 1.0, and 1.9 ± 0.2 μM, respectively.

Rabbit αM/uPA complexes activate plasminogen in vitro.

To confirm that αM/proteinase complexes are able to affect fibrinolytic activity, their effects on plasminogen activation in vitro were next studied as described in materials and methods. Complexes formed after incubation of exogenous scuPA with PFs in the presence of an excess of PAI-1 (Fig. 3C) were precipitated using anti-rabbit α2M IgY. The α2M/uPA complexes precipitated from these pleural fluids activated plasminogen (data not shown). Similar experiments were next performed to detect endogenous α2M complexes in the PFs of animals treated with the effective dose of scuPA, a high dose of Activase (which clears pleural loculation in TCN-injured animals), a high dose of Abbokinase (lacking an effective therapeutic effect), and vehicle saline (control, no effect) (16). The results shown in Fig. 9A demonstrate that only α2M complexes immunoprecipitated from PFs of animals treated with intrapleural scuPA possess plasminogen-activating activity.

Fig. 9.

Rabbit α2M/uPA complexes immunoprecipitated from PFs activate plasminogen and increase uPA activity when exposed to plasmin. A: α2M complexes were isolated from PFs of animals treated with ED scuPA, high dose of Activase, Abbokinase, and with buffer saline vehicle by immunoprecipitation with anti-rabbit α2M IgY (gray) or nonspecific IgY (white). The resin was transferred into 96-well Pro-Bind plates and incubated with plasminogen (0.2 μM) and 0.2 mM chromogenic plasmin substrate (Centerchem, Switzerland) at 37°C. Plasminogen activation was determined as an increase in plasmin (PL) activity in time using SpectraMax Plus (Molecular Devices). The data are presented in the box plot format as described in Fig. 1 legend. Plasminogen activating activity was significantly increased in the PFs of rabbits treated with intrapleural scuPA (P < 0.05). B: effect of plasmin on uPA activity of α2M/uPA complexes precipitated with anti-rabbit α2M IgY from pleural fluids. PFs of animals treated with scuPA were incubated with or without a mixture of 50 nM scuPA and 100 nM PAI-1 in HEPES buffer, pH 7.4, for 5 h at 37°C. α2M/uPA complexes were precipitated with anti-rabbit IgY and Agarose with immobilized anti-IgY goat polyclonal IgG. Amidolytic uPA activity in the resin samples was measured with fluorogenic substrate Pefafluor uPA with (gray) or without (white) preincubation with 1 nM plasmin. *Significant differences (P < 0.05) between the groups.

Finally, we tested whether or not uPA bound to rabbit α2M is activated by plasmin. When scuPA is activated to the mature tcuPA, its activity increases more than three orders of magnitude (3, 23, 28). Since PAI-1 retards activation of scuPA in PFs (Figs. 3A and 4), we speculated that uPA-rabbit α2M complexes might represent a sequestered source of PA activity. To address this possibility, amidolytic uPA activities of complexes of rabbit α2M immunoprecipitated from either PFs of animals treated with ED scuPA or from PFs preincubated with scuPA in the presence of PAI-1 were measured before and after treatment with plasmin (Fig. 9B). A moderate but significant increase in the uPA activity of the complexes in the presence of plasmin was found in the immunoprecipitates from the scuPA-supplemented PFs. A similar trend was observed in the PFs of rabbits treated with intrapleural scuPA.

DISCUSSION

While the use of intrapleural PAs remains an accepted alternative to surgical intervention at this time, currently available agents are reported to be variably effective and their safety profiles remain uncertain (8). These observations provide a strong rationale to better understand how these agents are processed in pleural fluids to optimize their utilization and to seek new approaches that are effective but mitigate the risk of bleeding. In this study, we identified complexes of αM/uPA in PFs of rabbits treated with intrapleural recombinant scuPA and showed that these complexes retain uPA activity and activate plasminogen. The findings represent a newly recognized, clinically relevant mechanism by which scuPA generates durable intrapleural fibrinolytic activity in rabbits with TCN-induced pleural loculation.

We found that the concentration of αM complexed with endogenous uPA is usually low in PFs of rabbits with TCN-induced injury. Intrapleural administration of scuPA in a dose that effectively clears intrapleural loculation (0.5 mg/kg) (16) promotes formation of αM/uPA complexes at significantly greater levels than those found in untreated animals or those treated with either intrapleural Abbokinase or Activase using the reported clinical dosing range for each agent (16). While it is clear that levels of fibrinolytic activity in pleural fluid correlate with lysis of loculations in TCN-induced injury, pleural fluid D-dimer concentrations do not correlate with effectiveness (data not shown; Ref.16), possibly relating to altered intrapleural clearance of this fibrin catabolite (16). These findings raise the alternative possibility that active complexes of PA with endogenous inhibitors could contribute to the sustained fibrinolytic activity and effectiveness in clearing intrapleural loculations. We found this to be the case and are unaware of any prior reports that define the ability of scuPA to form PA-generating complexes with αM in PFs or that demonstrate the ability of these complexes to contribute to intrapleural fibrinolysis in the setting of evolving loculation.

Since scuPA can bind soluble uPAR in pleural fluids, we sought to determine if uPA/suPAR complexes could likewise contribute to the protracted uPA activity found in rabbits treated with ED scuPA. The results shown in Figs. 5 and 6C demonstrate that under experimental design used in this study, human recombinant suPAR does not protect uPA from inactivation either in buffer solution or in PFs. Moreover, immunoprecipitation of uPAR complexes from PFs of rabbits treated with ED of scuPA (Fig. 6C) demonstrate that endogenous suPAR/uPA complex concentrations in PFs are low. The data indicate that rabbit suPAR/uPA complexes do not contribute significantly to uPA activity in PFs of animals treated with ED scuPA. While protection of uPA bound to uPAR from inactivation by PAI-1 remains controversial (3, 12), our present findings are consistent with our prior report that intrapleural complexes of scuPA with suPAR do not confer therapeutic benefit over scuPA alone (18).

Incubation of exogenous scuPA with PFs of vehicle-treated rabbits results in the formation of αM/uPA complexes that resist inhibition by PAI-1 (Figs. 3 and 4). In the vasculature, α2M/proteinase complexes are rapidly cleared, but our data suggest that these complexes may serve as a depot of plasminogen-activating/fibrinolytic activity in scuPA-treated animals with TCN-induced pleural injury (Fig. 9). Supporting this possibility, we found that both of the components of uPA activity within PFs of rabbits treated with intrapleural scuPA, the PAI-1-resistant and PAI-1-susceptible activities, correlate with the ability to clear pleural loculation in a dose-dependent manner (Fig. 2).

Our findings need to be considered in the context of the differences between αMs in humans and rabbits. Rabbits have two αMs (type 1 and type 2), both of which are able to interact with proteinases in a manner similar to that for human α2M (1, 38). Therefore, α1M, which is considered an acute phase protein (10), may also contribute to the formation of molecular cage-type complexes with uPA. Humans, by contrast express only α2M. Our data show that purified human α2M forms complexes with uPA and that these complexes resist inhibition by PAI-1 (Fig. 5). While we cannot exclude the possibility that both α-macroglobulins contribute to the PAI-1-resistant pool of uPA present within PFs of animals treated with scuPA, our data suggest that α2M contributes to the effect.

The presence of αM/uPA complexes isolated by immunoprecipitation from PFs of scuPA-treated rabbits was confirmed by Western blot analyses (Fig. 7). Formation of complexes (mol wt >200 kDa) of uPA and human α2M (either pure or plasma-derived) have previously been detected (25, 41), but complexes of uPA with human α2M lose most activity against high-molecular-weight substrates (37), whereas retaining ∼1% of their uPA activity in vitro (27). α2M/uPA complexes have been isolated from plasma (11), but their contribution to fibrinolysis in vivo has not, to our knowledge, previously been studied. Our data clearly show that αM-uPA complexes increase in PFs from rabbits treated with an effective dose of scuPA and that they retain the ability to activate plasminogen in the presence of PAI-1 (Figs. 1, 4, and 5).

The results of the present study strongly suggest that the slower induction of relatively low levels of fibrinolytic activity we observed after intrapleural administration of scuPA vs. Abbokinase or Activase at clinically relevant doses (16) is likely to relate to the relatively greater ability of ED scuPA to generate complexes with αM. Since the PFs used in this study were from rabbits treated with a single intrapleural dose of each fibrinolysin, their processing can be compared under virtually identical conditions and without the potential effect of differential accumulation with repeated intrapleural administration (16). Degradation of complexes containing proteinases (33) have been previously detected in sulfur mustard-injured skin lesions in rabbits (10), consistent with our findings. Increments of uPA-related activity derived from plasmin-mediated proteolysis of αM/uPA complexes were modest but were reproducibly and significantly enhanced by preincubation of PFs with exogenous scuPA (Fig. 9B). These novel observations support the concept that activity of these complexes is subject to proteolytic regulation that could affect intrapleural fibrinolysis.

The intrapleural dosing of scuPA is difficult to normalize to that of Abbokinase or Activase as scuPA is a proenzyme with relatively low levels of intrinsic activity. Conversely, Abbokinase and Activase are active serine proteases that exhibit much greater activity (5, 30). Both agents generate robust intrapleural fibrinolytic activity that is subject to relatively rapid inactivation by serpins, especially PAI-1 (Figs. 3 and 4) (16). Figure 3C shows that tcuPA also forms PAI-1-resistant complexes possessing uPA amidolytic activity when given at sufficient doses and raises the possibility that the paucity of uPA activity found in PFs of rabbits treated with Abbokinase is related to rapid inactivation by PAI-1 and/or the relatively low activity generated by doses normally used in clinical practice.

Kinetic studies of the interaction between PAI-1 and HMW tcuPA, Activase, or Abbokinase (LMW tcuPA, Fig. 8) clearly demonstrate high second order rate constants for the reaction with NBD P9 PAI-1 (7.0, 4.1, and 1.6 μM−1s−1, respectively), which may also contribute to the paucity of fibrinolytic activity in pleural fluids of TCN-injured rabbits treated with Abbokinase and to the rapid inactivation of Activase in these fluids. In clear contrast to HMW tcuPA, Abbokinase, or Activase, we found that the rate of the reaction between scuPA and PAI-1 is relatively much slower and does not depend on the enzyme or the serpin concentration. These results support the existence of an equilibrium between active and inactive species of scuPA (kon = 4.3 h−1), as proposed for zymogens of serine proteinases (4, 26). Assuming that the value of the equilibrium constant between active and inactive species of scuPA ranges from 250 (22) to 1,000 (3), the value of koff for transition of active to inactive scuPA is ∼0.3–1.2 s−1.

Administration of intrapleural sctPA (Activase) did not produce αM complexes capable of generating plasminogen-activating activity. These findings are of particular interest as Activase is now most commonly used to achieve intrapleural fibrinolysis in the U.S., and our findings demonstrate that its mode of action differs from that of intrapleural scuPA when used in identical doses and under comparable preclinical conditions. We speculate that intrapleural use of scuPA might provide a clinically useful safety advantage by virtue of its relatively slow intrapleural activation and durable activity related at least in part to complex formation with α2M.

In summary, we found that αM forms complexes with scuPA in rabbit PFs that retain uPA activity and activate plasminogen. These complexes are detectable within the pleural fluids of rabbits treated with exogenous scuPA and provide a source of residual uPA activity that can be increased by plasmin. The complexes between αM and uPA resist inhibition by PAI-1 and correlate with therapeutic efficacy. Equilibrium between active and inactive scuPA in the presence of PAI-1 may favor formation of the αM/uPA complexes. These newly appreciated relationships help explain the ability of scuPA to generate relatively slow, durable, effective intrapleural fibrinolysis in pleural fluids from rabbits with loculation induced by TCN. These properties may be advantageous for treatment of loculation associated with complicated parapneumonic effusions or empyema.

GRANTS

This work was supported by National Institutes of Health Program Project Grant PO1-HL-076406 and funding from the Texas Lung Injury Institute.

REFERENCES

View Abstract