Cultured bovine endothelial cells produce both urokinase and tissue-type plasminogen activators

Cell extracts and conditioned media (CM) from cultured bovine aortic endothelial cells (BAEs) were fractionated by PAGE in the presence SDS, and plasminogen activator (PA) activity was localized by fibrin autography. Multiple molecular weight forms of PA were detected in both preparations. Cell-associated PAs had Mr of 48,000, 74,000, and 100,000 while secreted PAs showed Mr of 52,000, 74,000, and 100,000. A broad zone of activity (Mr 80,000-100,000) also was present in both cellular fractions. In addition, PAs of Mr 41,000 and 30,000 appeared upon prolonged incubation or repeated freezing and thawing of the samples, and probably represent degradation products of higher molecular weight forms. This complex lysis pattern was not observed when CM was subjected to isoelectric focusing. Instead, only two classes of activator were resolved, one at pH 8.5, the other at 7.6. Analysis of focused samples by SDS PAGE revealed that the activity at pH 8.5 resulted exclusively from the Mr 52,000 form; all other forms were recovered at pH 7.6. The activity of the Mr 52,000 form was neutralized by anti-urokinase IgG but was not affected by antitissue activator IgG indicating that it is a urokinaselike PA. The activities of the Mr 74,000-100,000 forms were not affected by anti-urokinase. They were blocked by antitissue activator suggesting that all the forms in this group were tissue-type PAs. The multiple forms of PA were differentially sensitive to inactivation by diisopropylfluorophosphate (DFP). Treatment of CM with 10 mM DFP for 2 h at 37 degrees C only partially inhibited the 52,000-dalton form. However, it completely inactivated the 74,000-dalton partially inhibited the 52,000-dalton form. However, it completely inactivated the 74,000-dalton PA. The activity of the Mr 100,000 form was not affected by this treatment, or by treatment with 40 mM DFP. Thus, cultured BAEs produce multiple, immunologically distinct forms of PA which differ in size, charge, and sensitivity to DFP. These forms include both urokinaselike and tissue-activator-like PAs. The possibility that one of these forms is a zymogen is discussed.     

Spectrophotometric method to quantify and discriminate urokinase and tissue-type plasminogen activators.

Plasminogen activator and urokinase are often used as biological markers of cell activation. However, the methods currently used are cumbersome, make no discrimination between tissue-type plasminogen activator and urokinase, and do not allow expression of the results of the overall reaction in International Units. The one-step method described in this paper lacks these drawbacks. Moreover, we propose use of H-D-Val-Phe-Lys-4-nitroanilide as substrate which has a lower Km than the standard H-D-Val-Leu-Lys-4-nitroanilide which is commercially available. Low concentrations of sodium dodecyl sulfate in the reaction mixture dramatically and preferentially accelerate the reaction catalyzed by tissue-type plasminogen activators. Identical results are obtained under kinetic or fixed-time assay conditions using either a photometer or 96-well plate reader. The corresponding formulae are provided.     

Binding of mutant tissue-type plasminogen activators to human endothelial cells and their extracellular matrix

We previously demonstrated that tissue-type plasminogen activator (t-PA) specifically bound to its receptor (t-PAR) on human umbilical vein endothelial cells (HUVEC). In addition to analyses of t-PA binding to plasminogen activator inhibitor-1 (PAI-1) in the extracellular matrix (ECM) and to the t-PAR, we further evaluated the binding of three t-PA mutants, deltaFE1X t-PA lacking finger (F), epidermal growth factor-like (E) domains and one sugar chain at Asn177 thus comprising two kringles (K1 and K2) and protease (P) domains, deltaFE3X t-PA with three glycosylation sites deleted at Asn117, 184, and 448, and deltaFEK1 t-PA comprising K2 and P domains without glycosylation. Wild-type t-PA bound to ECM with high affinity, which was completely blocked by anti-PAI-1 IgG. Wild-type t-PA, deltaFE1X t-PA and deltaFEK1 t-PA bound to two classes of binding sites with high and low affinities on monolayer HUVEC. However, all t-PAs bound to a single class of binding site in the presence of anti-PAI-1 IgG. DeltaFEK1 t-PA bound t-PAR maximally among these t-PAs. These results suggested that the high affinity binding of t-PA mainly occurred with PAI-1 on ECM while the low affinity binding was with t-PAR. The deletion of F, E domains and sugar chains had no effect on binding with t-PAR. However, since only K1-missing t-PA (deltaFEK1) exhibited significantly increased binding sites among these t-PAs, it was suggested that the binding to t-PAR was mediated mainly by K2 domain and that the increase of binding was due to direct exposure of K2 domain.     

Urokinase-type and tissue-type plasminogen activators are essential for in vitro invasion of human melanoma cells

This study evaluates the contribution of two types of plasminogen activators (PAs; tissue-type PA (tPA) versus urokinase-type PA (uPA) toward the invasiveness of human melanoma cells in a novel in vitro assay. We identified two human melanoma cell lines, MelJuso and MeWo, expressing uPA or tPA as shown at mRNA, protein, and enzyme activity level. MelJuso cells produced uPA as well as plasminogen activator inhibitor-1 (PAI-1). The latter was, however, not sufficient to neutralize the cell-associated or secreted uPA activity. MeWo cells secreted tPA, but the enzyme was not found to be cell-associated. PAI-1 production by these cells was not detectable. Plasminogen activation and fibrinolytic capacity of both cell lines were reduced by anticatalytic monoclonal antibodies specific for the respective type of PA or by aprotinin. In a novel in vitro invasion assay, antibodies to PA as well as aprotinin decreased the invasiveness of both cell lines into a fibrin gel, Matrigel, or intact extracellular matrix. Our results confirm the importance of uPA-catalyzed plasminogen activation in tumor cell invasiveness. Furthermore, we provide evidence that tPA, beyond its key role in thrombolysis, can also be involved in in vitro invasion of human melanoma cells.     

A highly sensitive chromogenic microtiter plate assay for plasminogen activators which quantitatively discriminates between the urokinase and tissue-type activators

A simple and highly sensitive chromogenic microtiter plate assay for plasminogen activators is described. The assay is based on plasmin cleavage of the synthetic tripeptide plasmin substrate H-D-norleucyl-hexahydrotyrosyl-lysine p-nitroaniline, which yields the yellow chromophore p-nitroanilide. Production of the latter compound is then quantitated spectrophotometrically at 405 nm on an ELISA plate reader. Linearity of the assay can be achieved over at least four orders of magnitude in a single experiment (0.01-100 milliPloug units) with appropriate incubation times. Capitalizing on tissue-type plasminogen activator's dependence on fibrin for enzymatic activity, the selective use of soluble fibrin products allows discrimination between urokinase and tissue-type activator. The utility of this aspect of the assay for the analysis of complex samples containing both types of plasminogen activators is demonstrated.     

Different receptors mediate the hepatic catabolism of tissue-type plasminogen activator and urokinase.

Tissue-type plasminogen activator (t-PA) and urokinase (u-PA) are proteins with partial structural similarity and which are of importance in the therapy of thrombotic diseases. Both are known to be cleared from the circulation in vivo by uptake in the liver. The present study investigated whether the hepatic catabolism of u-PA and t-PA is mediated by a common receptor system. Four experimental protocols of increasing complexity were used: hepatocyte plasma membranes, isolated primary hepatocytes, liver perfusion and whole animals. For t-PA, a specific high-affinity binding site to hepatocytes and plasma membranes could be defined with a mean Kd of 4 +/- 3 nM, whereas the Kd for u-PA was less than 300 nM. Binding of t-PA could not be competed for by u-PA, and vice versa. Furthermore, clearance of t-PA in isolated perfused rat livers and in rabbits in vivo was 3-fold higher than that of u-PA, and a 50-100-fold molar excess of u-PA failed to inhibit clearance of t-PA in either system, and vice versa. Taken together, the results imply that hepatic elimination of t-PA and u-PA is mediated by distinct receptor systems of differing affinity.     

Signaling mechanisms responsible for lysophosphatidic acid-induced urokinase plasminogen activator expression in ovarian cancer cells

Lysophosphatidic acid (LPA) enhances urokinase plasminogen activator (uPA) expression in ovarian cancer cells; however, the molecular mechanisms responsible for this event have not been investigated. In this study, we used the invasive ovarian cancer SK-OV-3 cell line to explore the signaling molecules and pathways essential for LPA-induced uPA up-regulation. With the aid of specific inhibitors and dominant negative forms of signaling molecules, we determined that the G(i)-associated pathway mediates this LPA-induced event. Moreover, constitutively active H-Ras and Raf-1-activating H-Ras mutant enhance uPA expression, whereas dominant negative H-Ras and Raf-1 block LPA-induced uPA up-regulation, suggesting that the Ras-Raf pathway works downstream of G(i) to mediate this LPA-induced process. Surprisingly, dominant negative MEK1 or Erk2 displays only marginal inhibitory effect on LPA-induced uPA up-regulation, suggesting that a signaling pathway distinct from Raf-MEK1/2-Erk is the prominent pathway responsible for this process. In this report, we demonstrate that LPA activates NF-kappaB in a Ras-Raf-dependent manner and that blocking NF-kappaB activation with either non-phosphorylable IkappaB or dominant negative IkappaB kinase abolished LPA-induced uPA up-regulation and uPA promoter activation. Furthermore, introducing mutations to knock out the NF-kappaB binding site of the uPA promoter results in over 80% reduction in LPA-induced uPA promoter activation, whereas this activity is largely intact with the promoter containing mutations in the AP1 binding sites. Thus these results suggest that the G(i)-Ras-Raf-NF-kappaB signaling cascade is responsible for LPA-induced uPA up-regulation in ovarian cancer cells.     

Urokinase-type and tissue-type plasminogen activators have different distributions in cultured bovine capillary endothelial cells

The cell extracts and conditioned medium from cultured bovine capillary endothelial (BCE) cells were examined to determine the types of plasminogen activator (PA) present in each of these two fractions. The fractions were first analyzed by fibrin autography after sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis. The cell extracts contained two species of PA of Mr 48,000 and 28,000. Multiple forms of PA were detected in the conditioned medium: variable amounts of the Mr 48,000 and 28,000 forms and a broad band of activity with Mr in the range of 67,000-93,000. The major fraction of the Mr 48,000 form was in the cell extract. Treatment of the cells with 12-0-tetradecanoyl phorbol-13-acetate or with a preparation containing angiogenic activity resulted in a proportionate increase in the levels of all forms. The Mr 48,000 form was demonstrated to be a urokinase-like PA, since it was immunoprecipitated with antibodies to urokinase. When conditioned medium or cell extracts from biosynthetically labelled BCE cells were incubated with antiserum to urokinase, the Mr 48,000 form was immunoprecipitated only from the cell extract. The Mr 67,000-93,000 forms were demonstrated to be tissue-type PAs, since they were immunoprecipitated with antibodies to tissue PA. When the same conditioned medium or cell extracts were incubated with antiserum to tissue-type PA, the Mr 67,000-93,000 forms were immunoprecipitated only from the conditioned medium. Therefore, BCE cells are able to produce both tissue-type PA, which is primarily secreted, and urokinase-type PA, which remains primarily cell associated.     

Extracellular matrix produced by cultured corneal and aortic endothelial cells contains active tissue-type and urokinase-type plasminogen activators

Incubation of plasminogen with the subendothelial extracellular matrix (ECM) synthesized by cultured bovine corneal and aortic endothelial cells resulted in generation of fibrinolytic activity, indicated by proteolysis of ¹²⁵I-fibrin in a time-and dose-dependent manner. Both tissue-type plasminogen activator (t-PA) and urokinase-type plasminogen activator (u-PA) were identified in the ECM by fibrin zymography, immunoblotting, and inhibition of plasminogen activation by anti-u-and anti-t-antibodies. Most of the ECM-resident plasminogen activator (PA) activity did not originate from intracellular PA release occurring when the endothelial cells were lyzed and the ECM exposed, since a comparable amount of PA was associated with the ECM when the cells were lyzed with Triton X-100 or removed intact by treatment with 2 M urea. Active u-PA and t-PA were released from ECM by treatment with heparanase (endo-β-D-), indicating that some of the ECM-resident PA activity is sequestered by heparan sulfate side chains. These results indicate that both u-PA and t-PA produced by endothelial cells are firmly sequestered in an active form by the subendothelial ECM. It is suggested that ECM-resident plasminogen activators participate in sequential matrix degradation during cell invasion and tumor metastasis. PA activity may also function in release of ECM-bound growth factors (i.e., basic fibroblast growth factor) and activation of proenzymes (i.e., prothrombin), resulting in modulation of the ECM growth-promoting and thrombogenic properties. © 1993 Wiley-Liss, Inc.     

Construction and expression of hybrid plasminogen activators prepared from tissue-type plasminogen activator and urokinase-type plasminogen activator genes

Recent data from several studies have suggested that the non-protease domains in tissue-type plasminogen activator (t-PA) and urokinase-type plasminogen activator (u-PA) determine their biological specificities, including binding to fibrin clots and survival in the circulatory system (Van Zonneveld, A.-J., Veerman, H., and Pannekoek, H. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 4670-4674; Rijken, D. C., and Emeis, J. J. (1986) Biochem. J. 238, 643-646). Structural manipulations (e.g. deletions, additions, or substitutions) in these domains can thus be utilized to maximize the desired biological effects. Using recombinant DNA technology, we constructed a number of hybrid molecules from the t-PA and u-PA genes. In hybrid A, the epidermal growth factor and finger domains of t-PA (residues 1-91) were replaced by the epidermal growth factor and kringle of u-PA (residues 1-131). In hybrids B and C, the u-PA kringle (residues 50-131) was inserted either before (residue 92) or after (residue 261) the double-kringle region of t-PA. All these hybrid PAs containing three kringles were expressed in mouse fibroblast cells (C-127). The hybrid proteins were synthesized in predominantly a single-chain form with molecular weights of 70,000-80,000 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and were enzymatically active as assayed by the fibrin-agar plate method. In vitro studies on the binding of hybrid PAs to fibrin showed that hybrid B, like t-PA, possesses affinity toward fibrin, while hybrid A shows lower binding. This suggests that the finger domain, which is not present in hybrid A, plays a role in conferring fibrin affinity to the hybrid PAs. The enzymatic activities of the hybrids were compared with that of recombinant t-PA (rt-PA) expressed in the same vector/host system and found to be similar in activity toward a chromogenic peptide substrate. In addition, plasminogen activation with all the hybrid-PAs, as with rt-PA, was stimulated by fibrin, with the order of activity being rt-PA greater than or equal to hybrid B greater than hybrid C greater than hybrid A. This study shows the feasibility of shuffling functional domain(s) of known specificity in plasminogen activators which may lead to the design of a superior thrombolytic agent.     

Hormonal control of urokinase plasminogen activator secretion by sheep ovarian surface epithelial cells

Secretion of urokinase plasminogen activator (uPA) by ovarian surface epithelium (OSE) adjacent to the preovulatory ovine follicle has been implicated in apical tissue degradation and follicular rupture. In vitro experiments were designed to test the hypothesis that uPA release by OSE is under direct hormonal control. Epithelial cells were isolated from the ovarian surface of sheep using a polytetrafluorethylene scraper designed to dislodge adherent cells from culture flasks. Amidolytic cleavage of a uPA-specific chromogen (carbobenzoxy-L-gamma-glutamyl [alpha-ot-but]-glycyl-arginine-p-nitroanilide monoacetate) was used as a measure of enzymatic bioactivity in OSE-conditioned incubation media. Secretion of uPA by OSE suspensions from proestrous ewes was stimulated by exposure (2 h) to a preovulatory surge-like concentration of LH. OSE cells obtained during the luteal phase or anestrus were not responsive to LH. Baseline rates of uPA secretion and expression of estradiol receptors (in situ immunofluorescence detection) were not affected by reproductive status. Induction of uPA secretion by anestrous OSE was attained after priming (6 h) with estradiol-17beta; responsiveness was attributed to gonadotropin receptor (ligand binding) up-regulation. Monolayers of OSE established on polyethylene membranes secreted uPA predominately in a basal (i.e., toward the substratum) direction. We suggest that OSE in juxtaposition with the (hyperemic) wall of the preovulatory follicle is perfused by surge levels of LH, invoking uPA release into underlying ovarian tissues.     

Teratocarcinoma F9 cells induced to differentiate with sodium butyrate produce both tissue-type and urokinase-type plasminogen activators.

Sodium butyrate (NaB) can induce teratocarcinoma cell differentiation as retinoic acid (RA). However, the function of these two agents seems to be a little different [Kosaka et al., Exp Cell Res, 192:46-51, 1991]. F9 cells treated with NaB synthesize both tissue-type (tPA) and urokinase-type (uPA) plasminogen activator, though RA induces only tPA production. Urokinase-type PA is demonstrated to exist in association with membrane and to localize its activity to the close environment of the cell surface. This may cause the specific cell morphology and characteristics of differentiated F9 cells induced with NaB.     

Production of human plasminogen activators by cell culture

Plasminogen activators are a group of enzymes which play a crucial role in the breakdown of blood clots. The plasminogen activators currently used in medicine to remove clots from veins and arteries suffer from several disadvantages, but new cell culture techniques or cloned genes could make available safer and cheaper enzymes.     

Sequence of formation of molecular forms of plasminogen and plasmin-inhibitor complexes in plasma activated by urokinase or tissue-type plasminogen activator.

Total cellular polyadenylated RNA [poly(A)+ RNA] was prepared after guanidinium thiocyanate extraction of frozen brain tissue from age-matched normal and Down's-syndrome (trisomy 21) human foetuses. Poly(A)+ RNA populations were analysed by translation in vitro, followed by two-dimensional gel analysis by using both isoelectric focusing (ISODALT system) and non-equilibrium pH-gradient electrophoresis (BASODALT system) as the first-dimension separation. The relative concentrations of poly(A)+ RNA species coding for seven translation products were significantly altered in Down's syndrome, as determined by both visual comparisons of translation-product fluorograms from normal and Down's-syndrome samples and by quantitative radioactivity determination of individual translation products. The relative concentrations of mRNA species coding for two proteins (68 kDa and 49 kDa) were increased in Down's syndrome and may represent genes located on chromosome 21. The relative concentrations of mRNA species coding for five proteins (37 kDa, 35 kDa, 25.5 kDa, 24.5 kDa, 23 kDa) were decreased in Down's syndrome, these probably representing secondary effects of the trisomy. Six Down's-syndrome-linked translation products (49 kDa, 37 kDa, 33 kDa, 25.5 kDa, 24.5 kDa, 23 kDa) did not migrate with appreciable amounts of cellular proteins on two-dimensional gels and hence may represent either proteins of high turnover rates or those that are post-translationally modified in vivo. One translation product (68 kDa) comigrated with a major cellular protein species, which was identified as a 68 kDa microtubule-associated protein by limited peptide mapping. The significance of these changes is discussed in relation to the mechanisms whereby the Down's-syndrome phenotype is expressed in the human brain.     

Oncostatin M induces tissue-type plasminogen activator and plasminogen activator inhibitor-1 in Calu-1 lung carcinoma cells

Oncostatin M (OSM) is a glycoprotein cytokine that is produced by activated T-lymphocytes, monocytes, and macrophages. In a DNA synthesis assay, OSM reduced tritiated thymidine incorporation by 53% in Calu-1 lung carcinoma cells. Radiolabeled cDNAs from untreated Calu-1 cells and 30-h OSM-treated cells were used to probe duplicate nylon membrane cDNA expression arrays. This study revealed OSM-mediated expression of mRNAs encoding tissue-type plasminogen activator (tPA) and plasminogen activator inhibitor-1 (PAI-1). Northern blot analysis showed that the steady-state level of tPA mRNA is nearly undetectable in Calu-1 cells. Exposure of these cells to OSM for 30 h increased tPA mRNA expression by 20-fold and PAI-1 mRNA expression by 5-fold. Exposure of these cells to other gp130 receptor family cytokines, including leukemia inhibitory factor (LIF), interleukin-6 (IL-6), and IL-11, do not significantly affect DNA synthesis or induction of tPA/PAI-1. Western blot studies demonstrated that OSM mediates a marked increase in secretion of the tPA protein. Secreted tPA was present in the conditioned medium almost exclusively as tPA/PAI-1 complexes. Inhibitor studies demonstrated that OSM-mediated induction of tPA and PAI-1 mRNAs is largely dependent upon activation of the MEK1/2 pathway. The JAK3/STAT3 pathway potentially serves a secondary role in these regulatory events.     

Independent regulation of matrix metalloproteinases and plasminogen activators in human fibrosarcoma cells

Serine proteases and matrix metalloproteinases have been shown to often cooperate in multiple physiological and pathological processes associated with changes in the extracellular matrix (ECM). We have examined the interaction between the plasminogen activator (PA)-plasmin system and matrix metalloproteinases (MMPs) in HT1080 human fibrosarcoma cells treated with 12-O-tetradecanoyl-phorbol-13-acetate (TPA). While TPA treatment evoked a temporary increased expression of urokinase type PA (uPA), the production of both types of plasminogen activator inhibitors (PAI) was induced and sustained over 12 h by TPA treatment shifting the protease-protease inhibitors balance in favor of the inhibitors. TPA treatment of HT1080 cells induced the expression of interstitial collagenase (MMP-1) and increased the expression of gelatinase B (MMP-9), tissue inhibitor of metalloproteinases-1 (TIMP-1), and MT-MMP, a membrane-bound activator of progelatinase A (proMMP-2), while MMP-2 and TIMP-2 expression were decreased. Increased MT-MMP expression by TPA treatment was associated with increased activation of proMMP-2. These data show that the regulation of PA-plasmin and metalloproteinase and their specific inhibitors is uncoordinated. In addition, inhibition of the PA-plasmin system by PAI-2 or aprotinin did not prevent the activation of proMMP-2 by TPA, suggesting that plasmin is not involved in MT-MMP-mediated activation of proMMP-2. © 1996 Wiley-Liss, Inc.     

Platelet-derived growth factor D is activated by urokinase plasminogen activator in prostate carcinoma cells

Platelet-derived growth factor (PDGF) protein family members are potent mitogens and chemoattractants for mesenchymal cells. The classic PDGF ligands A and B are single-domain protein chains which are secreted as active dimers capable of activating their cognate PDGF receptors (PDGFRs). In contrast to PDGFs A and B, PDGF D contains an N-terminal complement subcomponent C1r/C1s, Uegf, and Bmp1 (CUB) domain and a C-terminal PDGF domain. PDGF D must undergo extracellular proteolytic processing, separating the CUB domain from the PDGF domain, before the PDGF domain can stimulate beta-PDGFR-mediated cell signal transduction. Here, we report that prostate carcinoma cells LNCaP and PC3 autoactivate latent full-length PDGF D into its active form under serum-independent conditions and that this autoactivation is inhibited by PAI-1, a urokinase plasminogen activator (uPA)/tissue plasminogen activator (tPA) inhibitor. Interestingly, uPA, but not the closely related protease tPA, is capable of processing recombinant latent PDGF DD into the active form. We identify the uPA cleavage site between the CUB and PDGF domains of the full-length PDGF D by mutational analysis and show that PDGF D and uPA colocalize in human prostate carcinoma. This evidence provides a direct link between uPA- and PDGF D-mediated cell signaling, which may contribute to the progression of prostate cancer.     

A synthetic DNA encoding a modified human urokinase resistant to inhibition by serum plasminogen activator inhibitor.

Using DNA synthesis technology we constructed two synthetic DNAs, designated syn-uPA-DNA and mut-uPA-DNA. Syn-uPA-DNA contains the complete coding sequence of human presecretion form of single-chain u-PA. Mut-uPA-DNA was derived from syn-uPA-DNA by deleting 18 base pairs coding for amino acids Arg179-Ser184. Each synthetic DNA was inserted into a bovine papilloma viral genome-based expression vector to obtain expression in mouse cells. The results indicate that both syn-uPA and mut-uPA proteins are secreted predominantly in single-chain form. The single-chain form of both enzymes can be completely converted to two-chain form by treatment with plasmin. Both enzymes are as active as natural urokinase (std-uPA) isolated from urine. Analysis of enzymatic activity indicates that under conditions where syn-uPA and std-uPA are completely inhibited by endothelial-type plasminogen activator inhibitor (PAI-1), mut-uPA retains 90% activity. In identical experiments with placental-type PAI (PAI-2), mut-uPA retains 80% activity. Syn-uPA is capable of forming a approximately 100-kDa complex with PAI, whereas mut-uPA can not. PAI-treated mut-uPA has kinetic properties similar to untreated syn-uPA or std-uPA. Overall, the data indicate that amino acids Arg179-Ser184 function at least in part as a binding site for PAI. Resistance to PAI inhibition may increase the potency of mut-uPA as a thrombolytic agent.