Secretion of tissue-type plasminogen activator and plasminogen activator inhibitor by cultured human endothelial cells: modulation by thrombin, endotoxin, and histamine

Primary cultures of pooled endothelial cells obtained from four to six human umbilical cord veins were grown to confluency in M 199 cell culture medium containing 20% human serum. The secretion of tissue-type plasminogen activator antigen (t-PA Ag) and fast-acting plasminogen activator inhibitor activity (PA-I activity) was measured in the conditioned medium after incubation of confluent cultures with serum-free medium for 24 hours (1.5 ml/25 cm2 cell surface). The baseline production of t-PA Ag was 2.7 +/- 1.4 ng/ml (mean +/- SD) and of PA-I activity 36 +/- 18 IU/ml. Stimulation with thrombin resulted in a dose-dependent and time-dependent increase of the secretion of both components. At 1 NIH U thrombin per milliliter, a fourfold increase of t-PA Ag and a twofold increase of PA-I activity were observed. This effect was dependent on a free active site in thrombin and specific protein synthesis (inhibited by cycloheximide and dactinomycin) but unrelated to prostacyclin synthesis (no effect of aspirin or indomethacin). No free t-PA activity could be demonstrated on fibrin plates in conditioned medium from either stimulated or nonstimulated endothelial cell cultures. Endotoxin stimulation resulted in a specific secretion of PA-I activity but not of t-PA Ag, whereas histamine stimulation at concentrations of 0.1 to 1 mumol/L caused a specific secretion of t-PA Ag but not of PA-I. These findings indicate that the secretion of t-PA Ag and PA-I activity are not necessarily coupled.(ABSTRACT TRUNCATED AT 250 WORDS)     

Inhibition of plasminogen activators by conditioned medium of human hepatocytes and hepatoma cell line Hep G2

Primary cultures of human hepatocytes and the human hepatocellular cell line Hep G2 are shown to produce fast-acting inhibitors of tissue-type plasminogen activator (tPA) and urokinase. The tPA inhibitory activities in conditioned medium of these liver cell types are very similar to those present in human endothelial cell conditioned medium. They are stable at pH 2.5, have similar dissociation constants with tPA (1.5 to 5 pmol/L), and are similar in thermostability. Addition of tPA to conditioned medium of Hep G2 and endothelial cells that has been depleted of tPA and urokinase reveals a 100 kilodalton tPA-inhibitor complex. The fast-acting tPA inhibitory activity in human plasma has comparable properties, and may originate from the liver or the vascular endothelium or both. After sodium dodecyl sulfate-polyacrylamide gel electrophoresis of conditioned medium from hepatocytes, Hep G2, and endothelial cells, additional fibrinolytic inhibition at 52 kilodaltons was visualized. This was not found with human plasma.     

Thrombin stimulates tissue plasminogen activator release from cultured human endothelial cells

The effect of thrombin on the release of tissue plasminogen activator from endothelial cells was studied in primary cultures of human umbilical vein endothelial cells. Tissue plasminogen activator concentration in conditioned medium was measured by a two-site radioimmunometric assay. The addition of increasing concentrations (0.01 to 10 U/ml) of thrombin to confluent cultures produced a saturable, dose-dependent increase in the rate of release of tissue plasminogen activator. A sixfold increase in tissue plasminogen activator concentration (from 2 to 12 ng/ml) occurred after the addition of 1 U/ml thrombin (8 X 10(-9) M) to cultures containing 5 X 10(4) cells/cm2. Enhanced release was not observed until 6 h after thrombin addition, reached a maximum rate of 1.3 ng/ml per h between 8 and 16 h, and then declined to 0.52 ng/ml per h after 16 h. The 6-h lag period before increased tPA release was reproducible and independent of thrombin concentration. Thrombin inactivated with diisopropylfluorophosphate or hirudin did not induce an increase in tissue plasminogen activator levels. A 50-fold excess of diisopropylfluorophosphate-treated thrombin, which inhibits binding of active thrombin to endothelial cell high affinity binding sites, did not inhibit the thrombin-induced increase. It is concluded that proteolitically active thrombin causes an increase in the rate of release of tissue plasminogen activator from cultured human endothelial cells. The 6-h interval between thrombin treatment and enhanced tissue plasminogen activator release may reflect a delaying mechanism that transiently protects hemostatic plugs from the sudden increase in the local concentration of this fibrinolytic enzyme.     

Human hepatoma cell line plasminogen activator

A human hepatoma cell line, Hep G2, which synthesizes and secretes several components of the fibrinolytic system, was examined for the capacity to produce plasminogen activator (PA). PA activity was found to accumulate in medium conditioned by the cells but was not detected in cell extracts. Analysis of Hep G2 conditioned medium by SDS-polyacrylamide gel electrophoresis revealed multiple PAs of Mr 100,000, 60,000, and 52,000 daltons. In addition, a heterogeneous band of activity was present between the 100,000 and 60,000 dalton PAs. All detectable PA activity in conditioned medium fractionated at a single position after isoelectric focusing. The isoelectric point of this material (pI 8.6) was identical to that of purified urokinase. Anti-urokinase IgG, but not anti-tissue activator IgG, neutralized the activity of these PAs. Treatment of conditioned medium with 25 mM diisopropylfluorophosphate for 5 hr at 37 degrees C inactivated the 52,000 and 60,000 dalton PAs but had no effect on the 100,000 dalton PA or the heterogeneous zone of activity. The possibility that the absence of detectable PA activity in the cell extracts resulted from the presence of fibrinolytic inhibitors was considered. Addition of cell extract to purified urokinase resulted in a concentration-dependent reduction of urokinase-mediated fibrinolytic activity, consistent with the presence of such inhibitors. Acidification of the cell extracts to inactivate the inhibitor(s) revealed the presence of a latent, cell-associated PA activity. Thus, in addition to the other major components of the fibrinolytic system, Hep G2 cells produce multiple forms of a urokinase-like PA. Detection of these PAs may be obscured by the presence of cellular inhibitors.     

The addition of endothelial cell growth factor and heparin to human umbilical vein endothelial cell cultures decreases plasminogen activator inhibitor-1 expression.

Plasminogen activator inhibitor-1 (PAI-1) is a specific and rapid inhibitor of tissue plasminogen activator (tPA) and urokinase. Clinical studies suggest that PAI-1 may play a crucial role in the regulation of fibrinolysis. A number of factors modulate PAI-1 activity in endothelial cell culture, and the isolation of PAI-1 cDNA now allows study of PAI-1 regulation at the mRNA level. We examined the effect of endothelial cell growth factor (ECGF) and heparin on PAI-1 expression in human umbilical vein endothelial cell (HUVEC) culture. The addition of ECGF and heparin to HUVEC cultures results in a 3-10-fold decrease in the PAI-1 activity secreted into the conditioned media. This effect is mediated at the mRNA level. A decrease in PAI-1 is also seen with higher concentrations of ECGF alone, but is greatly enhanced by the addition of heparin. No significant change in tPA antigen or mRNA levels was observed.     

Thrombin upregulates production of plasminogen activator inhibitor type 1 in human peritoneal mesothelial cells.

OBJECTIVE: To determine the influence of thrombin, which is generated intraperitoneally during peritoneal dialysis, on the synthesis of fibrinolytic system components in human peritoneal mesothelial cells (HMC). METHODS: Confluently grown HMC, isolated from the omental tissue, were used in the experiments. Conditioned media were obtained by incubating cells with serum-free M199 containing the appropriate concentration of the test compound. Tissue type plasminogen activator (tPA) and plasminogen activator inhibitor type 1 (PAI-1) antigen concentrations were measured by ELISA. Northern blot analysis was conducted for mRNA expression experiments. To test thrombin specificity, we used the thrombin inhibitor hirudin. The protein kinase C (PKC) inhibitor Ro 31-8220 was inserted to examine whether the effect of thrombin depends on PKC activity. RESULTS: Thrombin increased PAI-1 antigen in the conditioned media of HMC in a time- and concentration-dependent manner. After 24 hours incubation, PAI-1 levels increased from 350+/-30 ng/10(5) cells in control conditions to 620+/-30 ng/10(5) cells in HMC exposed to 5 U/mL thrombin (n = 8, p < 0.05). In contrast, there was no effect of thrombin on tPA antigen levels. An increase of PAI-1 mRNA expression was also observed by Northern blot hybridization. Hirudin (10 U/mL) inhibited the thrombin-induced increase in PAI-1 synthesis. In addition, a complete inhibition of the stimulating effect of thrombin on PAI-1 synthesis was obtained by blocking PKC activity with Ro 31-8220 (3 micromol/L). CONCLUSIONS: Thrombin increases PAI-1 synthesis in HMC via a PKC-dependent mechanism.Thereby the synthesis of tPA is not affected. Thus, thrombin may not only promote fibrin formation in the peritoneal cavity, but may also inhibit fibrin degradation by release of free PAI-1 from HMC.     

Regulation of the fibrinolytic system of cultured human vascular endothelium by interleukin 1.

We examined the effects of human interleukin 1 (IL-1) on the production of fibrinolytic components by cultured human vascular endothelium. Conditioned media collected from IL-1-treated (5 U/ml, 24 h) monolayers exhibited decreased tissue-type plasminogen activator (tPA) activity and increased plasminogen activator inhibitor (PAI) activity, as assessed by fibrin and reverse fibrin-autography. Quantitative immunological assays revealed a 35% decrease in tPA antigen and a 360% increase in active PAI antigen, after incubation for 24 h with 0.6 U/ml IL-1. Maximal effects (approximately 50% decrease in tPA antigen; 400-800% increase in active PAI antigen) were observed with 2.5-5 U/ml IL-1. Changes in tPA and PAI reached a maximum at approximately 24 h and persisted for greater than 48 h. IL-1 induction of endothelial procoagulant activity was more rapid and transient, peaking by 6 h and subsiding by 24 h. Natural monocyte-derived IL-1 and two species of recombinant IL-1 had comparable effects. Heat and polymyxin-B treatments differentiated IL-1 actions from those of endotoxin, which promoted similar endothelial alterations. IL-1 effects on endothelial procoagulant and fibrinolytic activities may contribute to the generation and maintenance of fibrin in pathophysiological settings in vivo.     

Detection and partial characterization of an inhibitor of plasminogen activator in human platelets

In this study, we demonstrate the presence of a previously undescribed fibrinolytic inhibitor in human serum. It has an apparent molecular weight of 50,000 and is not detected in serum derived from platelet-poor plasma, suggesting that it originates from platelets. This conclusion is supported by a number of observations. For example, extracts of washed, gel-filtered human platelets contain an inhibitor of similar activity and size, and physiological concentrations of thrombin induce its release from the platelets. Moreover, the kinetics and dose dependency of this release are similar to those observed for the release of platelet factor 4, and the release of both molecules is blocked by pretreating the platelets with prostaglandin E1 and theophylline. Mixing experiments, which were devised to investigate the specificity of the inhibitor, showed that the fibrinolytic activity initiated by both urokinase and tissue-type plasminogen activator was blocked by platelet releasate in a dose-dependent manner. In both cases, the amount of inhibition increased when the releasates were preincubated with the purified activators, indicating a direct interaction between the activators and an inhibitor(s). The inhibitory activity was removed by preincubating the releasates with antiserum prepared against an antiactivator purified from cultured bovine aortic endothelial cells. These results indicate that platelets contain an inhibitor which is released by thrombin, inhibits both urokinase and tissue-type plasminogen activator, and is immunologically similar to an inhibitor produced by endothelial cells. This molecule may represent a new class of inhibitors, the antiactivators, which function together with alpha 2-antiplasmin to regulate the fibrinolytic system of the blood. Its release from platelets by thrombin may protect the growing thrombus against premature dissolution initiated by plasminogen activators released by the endothelium.     

Effect of thrombin on the fibrinolytic activity of cultured bovine endothelial cells.

The vascular endothelium is a rich source of plasminogen activator (PA) and thus of blood vessel-associated fibrinolytic activity. Cultured bovine aortic endothelial cells were employed to determine if components of the coagulation system interact with the endothelium to modify expression of this activity. The addition of thrombin to these cultures led to a rapid decline in intracellular PA activity, with as little as 3 ng/ml, or 0.1 nM thrombin causing a 50% decrease within 30 min. Thrombin inactivated with diisopropylflurophosphate or hirudin did not elicit the response. Although control cultures secreted high levels of PA, no PA activity could be detected in the media surrounding the thrombin-treated cells. This loss of activity did not appear to result from direct inactivation of PA by thrombin. These observations indicate that the fibrinolytic potential of cultured endothelial cells is rapidly suppressed by trace amounts of thrombin. The generation of thrombin at sites of vascular injury may have a similar effect on the endothelium.     

Concomitant secretion of prourokinase and of a plasminogen activator-specific inhibitor by cultured human monocytes-macrophages

The plasminogen activator (PA) produced by freshly purified human monocytes-macrophages and histiocytic, lymphoma-derived U 937 cells was analyzed by zymography after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and found to migrate with an apparent Mr of 55,000, identical to that of urokinase (Uk). By immunoprecipitation with antibodies specific for the two different types of PA, the enzyme was shown to be immunologically related to urokinase, and not to tissue PA. Urokinase was secreted in the form of the inactive Mr 55,000 zymogen prourokinase , and could be converted to the active Mr 55,000 enzyme by limited proteolysis with plasmin. Conditioned media from cultures of U 937 cells and monocytes-macrophages inhibited the fibrinolytic activity of exogenously added urokinase. Using [125I]- labeled urokinase we observed the formation of an enzyme-ligand complex, which was not dissociated by boiling in SDS and migrated with an apparent Mr 40,000 daltons higher than the free enzyme; since complexed urokinase was functionally inactivated as a PA, the ligand is an inhibitor of urokinase. This inhibitor is different from fibroblast- produced protease- nexin , in that it did not interact with thrombin. These results suggest that plasminogen activation by mononuclear phagocytes can be modulated through the secretion of both (pro)enzyme and a specific inhibitor.     

Endothelial cells and vascular hemostasis]

Procoagulant, anticoagulant, and fibrinolytic activities are associated with endothelial cells and involve the production, secretion, and receptor mediated binding of proteins involved in these processes. The procoagulant aspect of endothelial cells function involves the production and release of von Willebrand Factor(vWF), the production of tissue factor, and the presence of Factor IX/IXa receptors on the cell surface. Secretion of vWf will promote the initial steps in thrombus formation by supporting platelet-platelet interaction and platelet-subendothelial matrix adhesion. Tissue factor which is undetectable in resting cells appears after exposure to various cytokines and initiates factor VIIa activation of factors IX and X. Receptors of Factor IX/IXa are also present and mediate the assembly of the prothrombinase complex on the endothelial cell surface. The anticoagulant pathway involves the cell surface protein thrombomodulin, protein C and its cofactor protein S. Thrombomodulin binds thrombin which activates protein C which in the presence of protein S cleaves and inactivates Factors V and VIII. Inactivation of these two coagulation cofactors halts the coagulation. Finally, endothelial cells also play a pivotal role in the fibrinolytic system. Production and regulated secretion of tissue plasminogen activator creates a profibrinolytic state in the endothelial cell environment. In addition, receptors for plasminogen and urokinase are also present, constituting a cell surface mediated fibrinolytic pathway. Plasminogen activator inhibitor type I, the primary inhibitor of tPA, is also produced by endothelial cells. Thus endothelial cells can promote and inhibit fibrinolysis, depending on the prevailing environmental conditions.     

Plasminogen activator activity and urokinase inhibitor activity in human amniotic fluid and fetal membranes

This study reports on the presence of latent plasminogen activator (PA) activity in human amniotic fluid (HAF). To measure PA, HAF was incubated with plasminogen, and the formation of plasmin was followed by its ability to cleave globin. The latent proenzyme in HAF was converted to active PA by treatment with sodium dodecyl sulphate (SDS) but not by tryptic digestion. The level of SDS-activatable PA activity in HAF increased with increasing gestational age. In an alternative, direct assay of PA based on its amidolytic activity upon L-pyroglutamyl-glycyl-L-arginine-p-nitroanilide (S-2444), HAF PA activity could be demonstrated even without prior exposure to SDS. Medium conditioned with either chorion or amnion produced PA activity suggesting that HAF PA is derived from the fetal membranes. Treatment of the conditioned medium with SDS or trypsin further increased the enzyme activity. The fetal membranes also produce inhibitory activities towards exogenous trypsin, plasmin, and urokinase. The inhibition of plasmin could be separated from the inhibitory activities towards trypsin and urokinase by DEAE-sephadex ion-exchange chromatography. The function of PA in the normal physiology and in pathological processes involving HAF and the fetal membranes remains to be elucidated.     

Interleukin 1 and lipopolysaccharide induce an inhibitor of tissue-type plasminogen activator in vivo and in cultured endothelial cells

Human IL-1, recombinant murine IL-1 and E. coli LPS were found to be potent inducers of plasminogen activator (PA)-inhibitor activity, both in vivo, in rats, as well as in cultured human endothelial cells. In vivo, LPS rapidly and dose-dependently (0.01-1,000 micrograms/kg) increased plasma PA-inhibitor activity. Infusion of IL-1 into rats resulted in a small but significant increase in PA-inhibitor activity in rat plasma. Likewise, in cultured human umbilical vein endothelial cells, LPS and IL-1 induced increased synthesis of PA-inhibitor. We suggest that the induced rat plasma inhibitor might be of endothelial origin.     

Endothelial cell-mediated conversion of Glu-plasminogen to Lys-plasminogen. Further evidence for assembly of the fibrinolytic system on the endothelial cell surface.

Lysine-plasminogen (Lys-PLG), the plasmin-modified form of native glutamic acid-plasminogen (Glu-PLG), displays enhanced binding affinity for fibrin and also enhanced activation by urokinase and tissue plasminogen activator. We previously demonstrated high-affinity, specific, and functional binding of Glu-PLG as well as tissue plasminogen activator to cultured human umbilical vein endothelial cells (HUVEC). In the present study, we demonstrate binding of Lys-PLG to HUVEC, as well as conversion of Glu-PLG to Lys-PLG at the cell surface. Binding of Lys-PLG to HUVEC was saturable, reversible, epsilon-aminocaproic acid-sensitive, and involved two saturable sites with Kd's of 142 pM and 120 nM, respectively. Upon incubation with Glu-PLG, HUVEC, as well as endothelium in situ, partially converted the ligand to a Lys-PLG-like species. Conversion by HUVEC was blocked by diisopropyl-fluorophosphate, but not by other serine protease inhibitors, including alpha 2-plasmin inhibitor. Eluates of intact umbilical cord vessels contained Lys-PLG by immunoblot analysis. Lys-PLG was also identified immunohistochemically on the endothelial surface of vessels from a variety of normal and inflamed tissues. Thus, endothelial cells appear to actively modify circulating Glu-PLG, bind Lys-PLG to their surface, and thus enhance the fibrinolytic potential of the blood vessel wall.     

Blood clearance and activity of erythrocyte-coupled fibrinolytics

Conjugating tissue-type plasminogen activator (tPA) to red blood cells (RBCs) endows it with features useful for thromboprophylaxis. However, the optimal intensity and duration of thromboprophylaxis vary among clinical settings. To assess how the intrinsic properties of a plasminogen activator (PA) affect functions of the corresponding RBC/PA conjugate, we coupled equal amounts of tPA or Retavase (rPA; a variant with an extended circulation time, lower fibrin affinity, and greater susceptibility to PA inhibitors). Conjugation to RBC markedly prolonged the circulation of each PA in rats and mice, without detrimental effects on carrier RBC. The initial blood clearance of RBC/tPA was faster than RBC/rPA, yet it exerted greater fibrinolytic activity, in part due to greater resistance of tPA and RBC/tPA to plasma inhibitors versus rPA and RBC/rPA observed in vitro. Soluble and RBC-coupled tPA and rPA exerted the same amidolytic activity, yet RBC/tPA lysed fibrin clots more effectively than RBC/rPA, notwithstanding comparable fibrinolytic activity of their soluble counterparts. Conjugation to RBC suppressed rPA's ability to be activated by fibrin, whereas the fibrin activation of RBC-coupled tPA was not hindered. Therefore, the functional profile of RBC/PA is influenced by: pharmacokinetic features provided by carrier RBC (e.g., prolonged circulation), intrinsic PA features (e.g., clearance rate, resistance to inhibitors), and changes imposed by conjugation to RBC (e.g., loss of cofactor stimulation). These factors, different from those guiding the design of soluble PA for lysis of existing clots, can be exploited in the rational design of RBC/PA tailored for specific prophylactic indications.     

Development of a new test for the global fibrinolytic capacity in whole blood.

BACKGROUND: The development of global tests for the fibrinolytic capacity in blood is hampered by the low base-line fibrinolytic activity in blood, by the involvement of both plasmatic components and blood cells in the fibrinolytic system and by the loss of fibrinolytic activity as a result of the action of plasminogen activator inhibitor-1 (PAI-1). OBJECTIVE: To develop a new test for the global fibrinolytic capacity (GFC) of whole blood samples. METHODS AND RESULTS: Collection of blood in thrombin increased the subsequent generation of fibrin degradation products. This was ascribed to rapid clot formation and concomitant reduction of in vitro neutralization of tissue-type plasminogen activator (tPA) by PAI-1. On the basis of this observation, the following test was designed: blood samples were collected in thrombin with and without aprotinin and clots were incubated for 3 h at 37 degrees C. The GFC was assessed from the difference between the fibrin degradation products in the two sera. The assay was applied to blood samples from patients and healthy subjects. Other hemostasis parameters were determined in plasma samples taken simultaneously. The GFC varied considerably (normal range 0.13-13.6 microg mL(-1)); physical exercise strongly increased the GFC. Statistically significant correlations were found with tPA activity, PAI-1 activity and fibrinogen level. A mixture of antibodies against tPA and urokinase-type plasminogen activator (uPA) completely inhibited the GFC. An inhibitor of activated thrombin-activatable fibrinolysis inhibitor (TAFI) accelerated fibrinolysis 8-fold. CONCLUSION: The new test represents a global assessment of the main fibrinolytic factors in plasma and potentially those associated with blood cells.     

Retroviral gene transfer of tissue-type plasminogen activator targets thrombolysis in vitro and in vivo

Patients usually have serious complications of thrombosis and bleeding by eating anticoagulation medicine for their residual lives after mechanical valve replacement operation. Tissue-type plasminogen activator (tPA) could target thrombolysis by activating plasminogen to fibrinolysin. In this study, we recombined a retroviral vector pLEGFP-N1-tPA and cultured purified packaging cells PT67/pLEGFP-N1-tPA to produce high-titer retrovirus. In vitro, two target cells, endothelial cell of umbilical vein (ECUV) 304 and heart muscle cell (HMC) that consist of endocardium and heart muscle, were infected by pLEGFP-N1-tPA. The results demonstrated that exogenous tPA was successfully transferred into ECUV304 and HMC. tPA in the two cells shows significant thrombolysis in plasma plate and the activity and content of tPA were high. Furthermore, in vivo, no thrombus was seen on the surface of Dacron patches (the same material making up a ring of mechanical valve) by tPA locally transferring around Dacron patches that were transplanted in the inferior caval veins of rabbits. tPA was successfully transferred into the local inferior caval vein. Activity and content of tPA were high in local tissue and blood and thrombolysis was effectively demonstrated by tPA rapidly, efficiently and long expressing. This laid the foundation for study and appliance of the tPA gene valve.     

A monokine stimulates production of human erythroid burst-promoting activity by endothelial cells in vitro.

Conditioned media were prepared from human peripheral blood monocytes and human umbilical vein endothelial cells. These media were assayed for erythroid burst-promoting activity (BPA) using human peripheral blood monocyte-depleted mononuclear cells as targets and assessing the stimulatory effect of the conditioned media on growth of early erythroid progenitor cells. Both monocytes and endothelial cells produced modest amounts of detectable BPA. Addition of varying concentrations of media conditioned by monocytes to plateau concentrations (5-10%) of media conditioned by endothelial cells had no additive effect. Endothelial cells incubated in the presence of 50% monocyte-conditioned medium produced 2.5- to 6.6-fold more BPA than did endothelial cells incubated only in control tissue culture medium. In contrast, endothelial cell conditioned medium did not stimulate increased BPA production by monocytes. Neither neutrophil- nor marrow fibroblastoid cell-conditioned medium stimulated BPA production by endothelial cells. Therefore, both monocytes and endothelial cells produce BPA. Moreover, monocytes produce a monokine that, in turn, stimulates the production of BPA by endothelial cells. Inasmuch as a monokine also has been shown to stimulate production of granulocyte-macrophage colony-stimulating activity, we propose that monocytes play a critical role in regulating the production of humoral regulators of the very early stages of hemopoietic cell differentiation.     

Thrombin induction of plasminogen activator inhibitor mRNA in human umbilical vein endothelial cells in culture.

We have previously reported that incubation of human umbilical vein endothelial cells (HUVECs) with human alpha-thrombin causes a time- and concentration-dependent increase in secreted plasminogen activator inhibitor type 1 (PAI-1) activity (Gelehrter TD, Sznycer-Laszuk R. J Clin Invest 1986;77:165-9). Here we report that the regulation of PAI-1 activity by thrombin is secondary to the thrombin-induced increase in PAI-1 mRNA accumulation. Incubation of HUVECs for 6 to 24 hours with 0.3 to 1.0 U/ml thrombin causes a 1.8-fold to 10-fold increase in PAI-1 activity and a 1.5-fold to threefold increase in accumulation of both the 3.2 and the 2.2 kilobase PAI-1 mRNAs. These effects are prevented by the prior addition of hirudin, a specific thrombin inhibitor. Inhibition of RNA synthesis with actinomycin D blocks the thrombin induction of PAI-1 mRNA accumulation. The protein synthesis inhibitor, cycloheximide, which itself markedly stimulates the accumulation of PAI-1, appears to prevent the induction by thrombin, suggesting that thrombin may act by inducing another effector such as interleukin-1. Consistent with this hypothesis is our observation that simultaneous addition of antibodies to interleukin-1-alpha prevents the thrombin induction of PAI-1 activity and mRNA.