On the fibrinolytic and thrombolytic properties of active-site p-anisoylated streptokinase-plasminogen complex (BRL 26921)

The fibrinolytic and thrombolytic properties of BRL 26921, a p-anisoylated streptokinase-plasminogen complex, were studied in vivo and in vitro. The in vivo studies were performed in rabbits with an experimental pulmonary embolus consisting of a 0.15 ml ¹²⁵I-fibrinogen labeled whole human blood clot. The degree of thrombolysis was calculated as the difference in isotope content of the thrombus before injection and after its recovery from the pulmonary artery of the animals sacrificed after 9 hours. In four control animals the extent of thrombolysis was 3.2 ± 1.5 (mean ± S.D.) percent. BRL 26921 given as a single bolus injection of 2,000, 10,000 or 60,000 streptokinase equivalent units in groups of three or four rabbits resulted in 10.3 ± 6.5, 29.4 ± 10.2 and 28.4 ± 9.3 percent lysis. A mixture of 60,000 IU of streptokinase and 1 mg of human plasminogen caused 11.2 ± 7.2 percent lysis and a tenfold higher amount 38.4 ± 16.9 percent. Injection of BRL 26921 or of the low dose of streptokinase and human plasminogen induced negligible systemic activation of plasminogen or consumption of fast-acting antiplasmin and no fibrinogen breakdown. The high dose of streptokinase and human plasminogen (600,000 units and 10 mg) resulted in systemic defibrination. Thus in this experimental model, BRL 26921 is a much more potent thrombolytic agent than an equivalent amount of streptokinase and human plasminogen. In addition BRL 26921 causes specific thrombolysis without systemic fibrinogen breakdown.The in vitro experiments were performed in a system composed of a radioactive human blood clot (¹²⁵I-fibrinogen labeled), hanging in circulating human or rabbit plasma. Significant thrombolysis (release of radioactivity) in the human plasma system was obtained with 50 units or more of BRL 26921 per ml plasma or with 500 or more units of streptokinase per ml. Specific thrombolysis without systemic plasminogen activation, α₂-antiplasmin consumption and fibrinogen breakdown was however not obtained with either agent.Replacement of the normal human plasma with that of a patient with high titers of streptokinase antibodies (3 weeks after streptokinase therapy) abolished the effects of both BRL 26921 and of streptokinase. In the rabbit plasma system significant thrombolysis was already observed with 10 to 50 units of BRL 26921 per ml plasma although this did not result in systemic fibrinolytic activation or fibrinogen breakdown. This was only observed with a concentration of 500 units per ml.The present findings indicate that in the rabbit system, BRL 26921 has a high specific thrombolytic and low fibrinogenolytic effect. In the human plasma system BRL 26921 has a higher specific activity than streptokinase but there is no differentiation between its thrombolytic and fibrinogenolytic effects.     

The level of extrinsic plasminogen activator (t-PA) during clotting as a determinant of the rate of fibrinolysis; inefficiency of activators added afterwards

Dissolution of washed fibrin clots in vitro was found to depend on the concentration of extrinsic (tissue-type) plasminogen activator (t-PA) during clotting. Washed fibrin clots prepared from t-PA-rich plasma lyse spontaneously within a few hours, but similar clots, prepared from t-PA-poor plasma, do not lyse within 24 hours, neither spontaneously nor on addition of t-PA, streptokinase (SK) or urokinase (UK). When washed fibrin clots were suspended in buffered saline to which t-PA, SK or UK was added, lysis was enhanced by addition of lys-plasminogen to the buffer. This observation may have important consequences for the interpretation of fibrinolytic experiments in vitro and in vivo, for the management of a variety of diseases complicated by fibrin depositions, and eventually for thrombolytic therapy with t-PA.     

Acylated derivatives of streptokinase-plasminogen activator complex as thrombolytic agents in a dog model of aged venous thrombosis

A procedure is described for the establishment of experimental clots, aged in vivo for 72 hr, in the jugular vein of beagle dogs. Two acylated derivatives of streptokinase-human (lys) plasminogen activator complex with greatly differing deacylation rates under physiological conditions were compared as thrombolytic agents in the model. These were BRL 26921 (deacylation half-life, 40 min) and BRL 33575 (deacylation half-life c. 17 hr). The pharmacokinetic clearance rate of BRL 33575 from the circulation was studied and gave a clearance half-life of about 7 hr. BRL 33575 was found to be the superior agent in lysing 72 hr aged clots, being effective at a single bolus dose of 420 micrograms/kg or in three equal divided doses of 140 micrograms/kg given at 12 hr intervals. The single dose regime gave moderate systemic plasminogen activation, and the effect was significantly reduced with the divided dose regime. Infusion of freshly formed streptokinase-human plasminogen activator complex at 420 micrograms/kg over 15 hr gave little thrombolysis despite marked systemic plasminogen activation.     

Influence of intrinsic and extrinsic plasminogen upon the lysis of thrombi in vitro.

When the rate of lysis of artificial thrombi (prepared from plasma or whole blood) was expressed according to the concentration of tissue type plasminogen activator (t-PA) or single chain urokinase type plasminogen activator (sc-uPA) then bell-shaped dose response curves were obtained, low rates being observed at concentrations of activator greater than 500 units/ml. Bell-shaped dose response curves were not observed for rate of lysis of artificial thrombi over the concentrations of streptokinase tested (SK) or for the lysis of plasma gel clots by any of the activators tested. Further investigation indicated that the preponderant mechanism for dissolution of thrombi at 500 units/ml of t-PA was by activation of the plasminogen within the thrombus (intrinsic) since the plasminogen present in the plasma perfusing the thrombus (extrinsic) rapidly became depleted. On the other hand, at 50 units/ml t-PA the lysis was observed to be due preponderantly to the action of plasmin arising from extrinsic rather than intrinsic plasminogen. If "plasminogen enriched" thrombi were prepared in the presence of Lys plasminogen (Lys-Plg) faster rates of lysis occurred and bell-shaped biometric curves were not observed.     

The systemic fibrinolytic effect of BRL 26921 during the treatment of acute peripheral arterial occlusions

BRL 26921 is a new acylated streptokinase-plasminogen complex which may have a more specific local thrombolytic effect than streptokinase or urokinase. 34 patients with acute peripheral arterial occlusions were given eight hourly bolus injections of 5 mg BRL 26921 for up to 72 h. Systemic fibrinolysis was observed in all patients yet in only 24% was the occluding thrombus lysed. 44% of the patients had haemorrhagic complications and 24% suffered further thrombotic events during or soon after treatment. There was no correlation between the degree of systemic fibrinolysis produced and dissolution of the thrombi. The degree of systemic fibrinolysis did not affect the complication rate. There is no evidence from this study that BRL 26921 has a specific local thrombolytic effect.     

Factor XIII cross-linking and the rate of fibrinolysis induced by streptokinase and urokinase

The rate of lysis induced by streptokinase or urokinase of fibrin clots with varying degrees of Factor XIII cross-linking was investigated. Various techniques were used in which the plasminogen activator was uniformly distributed through the clot, or was present only at the surface. The clots were produced from purified fibrinogen solutions or from plasma. None of these ivestigations gave any evidence to indicate that lytic rates were decreased by the cross-linking of either the α or γ chains of fibrin.     

Whole blood clot lysis: In vitro modulation by activated protein C

Lysis of clots prepared from native or citrated whole blood as measured by release of ¹²⁵I fibrinogen degradation products was 10% or less at 20 hours. Lysis of these clots was accelerated by activated protein C in a dose-dependent manner (0.1 to 20 μg/ml) from less than 10% to 60–80% at 20 hours. Lysis of clots prepared from native or citrated platelet poor plasma across the same concentration range of activated protein C was less than 15%. Gla-domainless activated protein C was equally effective in accelerating clot lysis whereas DIP-activated protein C or factor Xa did not accelerate clot lysis. This suggested that this action of activated protein C was enzymatic and this this action was limited to protein C among the vitamin K dependent proteins.The unresponsiveness of platelet poor plasma to activated protein C was completely restored to that of whole blood by addition of mononuclear leukocytes. Addition of red corpuscles or platelets $\text{alone}$ had no effect on this response, while addition of polymorphonuclear leukocytes partially restored this response. Addition of metabolic inhibitors 2-deoxyglucose and oligomycin inhibited the response of whole blood and of plasma-mononuclear leukocytes to activated protein C. Reconstitution studies of platelet poor plasma made deficient in plasminogen activator and plasminogen showed that accelerated clot lysis produced by mononuclear leukocytes and activated protein C required the presence of plasminogen. We concluded, therefore, that activated protein C accelerates whole blood or plasma-leukoycte clot lysis by modulating activation of the plasminogen system by metabolically active leukocytes.     

Correlation between progressive adsorption of plasminogen to blood clots and their sensitivity to lysis

The binding of plasminogen to preformed human plasma clots immersed in citrated human plasma was measured and correlated with the sensitivity of these clots to lysis with recombinant tissue-type plasminogen activator (rt-PA), recombinant single-chain urokinase-type plasminogen activator (rscu-PA) or two chain urokinase-type plasminogen activator (tcu-PA, urokinase). When 0.15 ml plasma clots were compressed mechanically to about 1% of their original weight, and immersed in 0.15 ml plasma, 131I-labeled native plasminogen (Glu-plasminogen) adsorbed progressively from the plasma milieu onto the clot; binding was 3 +/- 1% (n = 10) after 1 h, 7 +/- 1% after 12 h and 12 +/- 1% after 48 h. This was associated with an increased sensitivity of the clot to lysis; 50% clot lysis in 4 h was obtained with 65 +/- 5 ng/ml (n = 3) rt-PA before and 30 +/- 5 ng/ml (n = 3) after 48 h preincubation in plasma (p less than 0.01), with corresponding values of 660 +/- 55 ng/ml (n = 3) and 280 +/- 25 ng/ml (n = 3) for rscu-PA, (p less than 0.01), and 800 +/- 85 ng/ml (n = 3) and 270 +/- 35 ng/ml (n = 3) for urokinase (p less than 0.01). Additional binding of plasminogen and increased sensitivity to lysis were reduced or abolished when the clot was preincubated in plasminogen-depleted or in t-PA-depleted plasma, or when 20 mM 6-aminohexanoic acid or 2,000 KIU/ml aprotinin were added.(ABSTRACT TRUNCATED AT 250 WORDS)     

The effect of flow on lysis of plasma clots in a plasma environment

Fibrinolysis initially generates channels in an occluding thombus which results in blood flow through the thrombus. Since the impact of flow along the surface of a thrombus on thrombolysis has not been investigated in detail, we studied in vitro how such a flow affects lysis. Compacted and noncompacted plasma clots were used as model thrombi. With compacted clots, fibrin-specific lysis induced by alteplase in the outer plasma was accelerated about 2-fold by strong flow (arterial shear rate). Non-fibrin-specific lysis induced either by a high concentration of alteplase or by streptokinase was slow, was accompanied by rapid depletion of plasminogen in the outer plasma, and was only slightly accelerated by flow. With noncompacted clots, similar acceleration factors were documented, when mild flow (venous shear rate) was applied. Strong flow further accelerated fibrin-specific lysis, up to 10-fold as compared to lysis without flow, but paradoxically retarded non-fibrin-specific lysis. The data suggest that flow accelerates lysis by enhancing transport of plasminogen from the outer plasma to the surface of the clot. Both opposite effects of the strong flow were mediated by forceful intrusion of the outer plasma into the noncompacted clot due to flow irregularities. In the case of non-fibrin-specific lysis this resulted in the replacement of the plasminogen-containing milieu by plasminogen-depleted outer plasma in certain areas of the clot turning them into virtually unlysable fragments. This flow-enforced "plasminogen steal" may contribute to the relatively high percentage of incomplete thrombolysis (TIMI-2 grade flow) documented in a number of trials for non-fibrin-specific thrombolytic agents. In the case of fibrin-specific lysis, the effect of flow on the speed of fibrinolysis is always beneficial.     

In vitro effects of the acylated streptokinase-plasminogen activator complex BRL 33 575 incubated with normal human plasma

Thrombolysis with acylated streptokinase-plasminogen complexes is aimed to achieve fibrinolysis without systemic fibrinogenolysis. The p-aminobenzoyl-streptokinase-(Lys)-plasminogen-complex (BRL 33 575) should be particularly useful due to its slow deacylation rate. Unexpectedly, repeated doses of 10 mg of BRL 33 575 (corresponding to 310'000 streptokinase equivalent units) induced systemic effects in patients though less than streptokinase alone. In vitro incubation of normal human plasma with BRL 33 575 at concentrations used in patients resulted in nearly complete consumption of alpha 2-antiplasmin and plasminogen and significant fibrinogenolysis within 3 hr. This demonstrates that - despite of slow deacylation of BRL 33 575 - the small amounts of activator generated are highly efficacious in activating plasma plasminogen under conditions in which no physiological clearance of the free activator takes place. Simulating the calculated activator release from BRL 33 575 by infusing equivalent amounts of streptokinase into plasma resulted in less pronounced effects. This is probably explained by anti-streptokinase antibodies which will neutralize the initially infused streptokinase but will be bound by BRL 33 575. Our in vitro experiments indicate that further clinical studies should be done with lower doses of BRL 33 575 or prolonged dosage intervals.     

Inhibitory effects of lysine analogues on t-PA induced whole blood clot lysis

The lysine analogues epsilon-aminocaproic acid (EACA) and trans-4-amino-methyl cyclohexane carboxylic acid (AMCA) are used to prevent excessive bleeding in patients with coagulopathies, such as hemophilia and thrombocytopenia, or in those who have received tissue plasminogen activator (t-PA). However, their relative efficacy in inhibiting lysis of clots that have been formed in the presence of exogenous t-PA or that have been formed and then exposed to exogenous t-PA has not been well characterized. The present study utilized blood from normal volunteers and ¹²⁵I-fibrinogen in a dilute whole blood clot assay to determine the relative concentrations of lysine analogues required for inhibition of clot lysis induced by exogenous t-PA. AMCA (0.06 mM) and EACA (0.6 mM) were effective in prolonging clot lysis if (1) whole blood clots were formed and then exposed to a lysine analogue and exogenous t-PA or if (2) whole blood clots were formed in the presence of exogenous t-PA and a lysine analogue. However, their inhibitory effect was markedly reduced if clots were formed in the presence of t-PA and then exposed to either of the lysine analogues. The analogues did not inhibit the initial binding of t-PA to fibrin. They did inhibit binding of plasminogen to fibrin as well as the activation of plasminogen by t-PA in the absence of fibrin. The data suggest that lysine analogues, even at low concentrations, reduce the rate of t-PA induced whole blood clot lysis by several mechanisms.     

Fibrin-specificity of a plasminogen activator affects the efficiency of fibrinolysis and responsiveness to ultrasound: comparison of nine plasminogen activators in vitro

In a number of cases, thrombolytic therapy fails to re-open occluded blood vessels, possibly due to the occurrence of thrombi resistant to lysis. We investigated in vitro how the lysis of hardly lysable model thrombi depends on the choice of the plasminogen activator (PA) and is accelerated by ultrasonic irradiation. Lysis of compacted crosslinked human plasma clots was measured after addition of nine different PAs to the surrounding plasma and the effect of 3 MHz ultrasound on the speed of lysis was assessed. Fibrin-specific PAs showed bell-shaped dose-response curves of varying width and height. PAs with improved fibrin-specificity (staphylokinase, the TNK variant of tissue-type PA [tPA], and the PA from the saliva of the Desmodus rotundus bat) induced rapid lysis in concentration ranges (80-, 260-, and 3,500-fold ranges, respectively) much wider than that for tPA (a 35-fold range). However, in terms of speed of lysis, these three PAs exceeded tPA only slightly. Reteplase and single-chain urokinase were comparable to tPA, whereas two-chain urokinase, anistreplase, and streptokinase were inferior to tPA. In the case of fibrin-specific PAs, ultrasonic treatment accelerated lysis about 1.5-fold. For streptokinase no acceleration was observed. The effect of ultrasound correlated with the presence of plasminogen in the outer plasma, suggesting that it was mediated by facilitating the transport of plasminogen to the surface of the clot. In conclusion, PAs with improved fibrin-specificity induce rapid lysis of plasminogen-poor compacted plasma clots in much wider concentration ranges than tPA. This offers a possibility of using single-or double-bolus administration regimens for such PAs. However, it is not likely that administration of these PAs will directly cause a dramatic increase in the rate of re-opening of the occluded arteries since they are only moderately superior to tPA in terms of maximal speed of lysis. Application of high-frequency ultrasound as an adjunct to thrombolytic therapy may increase the treatment efficiency, particularly in conjunction with fibrin-specific PAs.     

Acyl-enzymes as thrombolytic agents in a rabbit model of venous thrombosis

A derivative of human lys-plasmin in which the active site has been reversibly acylated (BRL 26920; p-anisoyl human lys-plasmin) has been examined as a fibrinolytic agent in a previously described rabbit model of venous thrombosis and shown to be significantly more active and less fibrinogenolytic than free plasmin. A p-anisoylated derivative of a streptokinase (SK)-activated plasmin preparation was significantly less fibrinogenolytic in vivo than the non-acylated enzyme. Acylation increased the fibrinolytic activity of preparations of SK-plasmin activator complexes. BRL 26921, the active site anisoylated derivative of the primary 2-chain SK-plasminogen complex was the most potent fibrinolytic agent studied. SK-Val442-plasminogen complexes, free or acylated, were biologically inactive in this model and confirm the essential nature of fibrin binding processes for effective thrombolysis in vivo.     

The dissolution of human cross-linked plasma fibrin clots by the equimolar human plasmin-derived light (B) chain-streptokinase complex. The acceleration of clot lysis by pretreatment of fibrin clots with the light (B) chain.

The equimolar human light (B) chain·streptokinase activator complex was shown to be an effective fibrinolytic agent in the dissolution of human cross-linked plasma fibrin clots, in vitro; its activator activity was similar to that of the equimolar human plasmin·streptokinase complex. ¹²⁵I-radiolabelled preparations of the human plasmin-derived light(B) and heavy(A) chains, and Lys-plasminogen were shown to be adsorbed to cross-linked plasma fibrin clots invitro to the same extent. Clot lysis by the light(B) chain·streptokinase complex, the plasmin·streptokinase complex, streptokinase and urokinase was enhanced by a factor of at least 10-fold when the fibrin clots were pretreated with either the light(B) chain or Lys-plasminogen, but was not enhanced when the fibrin clots were pretreated with the heavy(A) chain. The activator activities of the light(B) chain·streptokinase complex, the plasmin·streptokinase complex and streptokinase were similar on these pretreated cross-linked fibrin clots. Limited proteolysis of fibrinogen by the light(B) chain and by plasmin showed the same specificity giving Fragments X, Y, D and E, but the rate of hydrolysis of fibrinogen by the light(B) chain was much slower.     

Protraction of whole blood and plasma clot lysis in patients with high levels of an inhibitor of tissue-type plasminogen activator

The influence of the newly discovered, fast-acting inhibitor of tissue-type plasminogen activator (t-PA) on the lysis time of plasma clots was studied by visual observation of lysis of clotted citrated plasma after addition of purified t-PA. To a series of plasma samples with various concentrations of naturally occurring PA-inhibitor purified t-PA was added to a final concentration, which in pooled normal plasma is sufficient to induce clot lysis within a few hours. In those plasma samples with a high free inhibitor level, determined by measuring the recovery of the activity of added purified t-PA, clot lysis was retarded. Whole blood clots were made by clotting freshly collected non-anticoagulated blood with thrombin after admixture of a trace amount of radiolabeled fibrinogen and a fixed amount of t-PA. Lysis rate, read from the appearance of radioactivity in the serum after centrifugation, was significantly lower in clots obtained from subjects with a high free inhibitor level than in those with a low inhibitor level. It is concluded that the PA-inhibitor protracts clot lysis and may be relevant for physiological fibrinolysis.     

Fibrin clot lysis by thrombolytic agents is impaired in newborns due to a low plasminogen concentration.

Although thrombolytic drugs have been extensively used in adults, there is sparse information on their effectiveness in newborns whose fibrinolytic system differs significantly from adults. The purpose of this study was to determine if low plasma levels of plasminogen in cord plasma limited the therapeutic effectiveness of thrombolytic agents. Urokinase (UK), streptokinase (SK) and tissue plasminogen activator (TPA) were compared for their ability to lyse washed 125I-labelled adult or cord fibrin clots suspended in cord or adult plasma. 125I-labelled fibrin clots were prepared by recalcifying cord or adult plasma spiked with labelled fibrinogen and then placed into cord or adult plasma which contained either saline or differing amounts of a specific thrombolytic agent. After a 60 min incubation, the remaining 125I-fibrin in clots released 125I-fibrin fragments, and concentrations of fibrinogen, alpha 2-antiplasmin, and plasminogen in the bathing plasma were measured and compared to starting values. Cord fibrin clots were more resistant than adult fibrin clots to all thrombolytic drugs tested (p less than 0.001). On average, the cord system retained 27% more 125I-fibrin in clots, and released 32% less 125I-fibrin fragments into plasma. Fibrinogenolysis was also decreased in cord plasmas compared to adult plasmas. The degree of fibrinolysis and fibrinogenolysis in cord plasma increased to adult values when plasminogen concentrations were increased in the bathing plasma. Thus, cord fibrin clots have an impaired response to thrombolytic agents secondary to low levels of plasminogen.(ABSTRACT TRUNCATED AT 250 WORDS)     

Dissolution of thrombi by tissue plasminogen activator, urokinase and streptokinase in an artificial circulating system

The ability of porcine tissue plasminogen activator, urokinase and streptokinase to dissolve whole blood clots has been compared in an artificial circulating system, the Chandler loop. On a molar basis the tissue activator was more effective in lysing both partially and totally crosslinked thrombi than urokinase or streptokinase. The latter drugs required 30–50 times higher concentrations to achieve a fibrinolytic response, similar to that of tissue activator. As demonstrated by analysis of fibrinogen, ₂-antiplasmin and plasmin-₂-antiplasmin complex there was no activation of plasminogen in plasma by the tissue activator in the absence of fibrin. Both urokinase and streptokinase gave under the same conditions rise to a significant dose dependent fibrinogenolysis. In the presence of fibrin, tissue activator induced a weak fibrinogenolysis (about 20%) in plasma, whereas both urokinase and streptokinase caused a much higher degree of fibrinogenolysis (50 and 60 % respectively).     

Potentiation by Lys-plasminogen of clot lysis by single or two chain urokinase-type plasminogen activator or tissue-type plasminogen activator

Study has been made of the influence of addition of human NH2 terminal glutamic acid plasminogen (Glu-Plg) or human NH2 terminal lysine plasminogen (Lys-Plg) to normal citrated plasma upon the rate of lysis of fully crosslinked plasma clots in the presence of single or two chain urokinase type plasminogen activator (scu-PA/tcu-PA) or tissue plasminogen activator (t-PA). The specificity of any thrombolytic property was evaluated by measurement of plasma fibrinogen levels. Lys-Plg added to a concentration of 20% of normal plasma plasminogen caused 5 to 6 fold increase in the extent of lysis observed at 6 hours by 100 units/ml of scu-PA and with a small increase in fibrinogenolysis. Glu-Plg added at 20% of normal level had no influence on thrombolysis but at 50% of normal caused increased thrombolysis with rapid depletion of plasma fibrinogen. An apparently synergistic effect of addition of tcu-PA on scu-PA activity was increased by addition of plasminogen (e.g. addition of 20% Lys-Plg increased the lysis rate 4 to 5 fold over the first hour equivalent to an increase of potency of approximately three to four fold). Addition of plasminogen up to double the normal plasma concentration was observed to have no influence on clot lysis in the presence of t-PA. Plasminogen potentiated the rate of lysis by scu-PA/t-PA synergic mixtures with an approximately 1.5 to 1.9 fold increase in potency. Potentiation occurred without increase in the depletion of plasma fibrinogen.(ABSTRACT TRUNCATED AT 250 WORDS)     

Incidence of cerebral hemorrhage after treatment with tissue plasminogen activator or streptokinase following embolic stroke in rabbits [corrected

We studied thrombolysis in an animal model of embolic stroke to determine the safety of tissue plasminogen activator and streptokinase. We occluded the middle cerebral arteries of 137 rabbits with radiolabeled blood clots and administered tissue plasminogen activator (n = 49), streptokinase (n = 40), or saline (n = 48) at various times after embolization. We assessed the rate of thrombolysis and cerebral hemorrhage 24 hours later. Both drugs were very effective in producing thrombolysis. Compared with saline, streptokinase caused a significant increase in the rate of cerebral hemorrhage (p less than 0.05), but tissue plasminogen activator did not. We conclude that thrombolytic therapy for acute stroke should be safer with tissue plasminogen activator than with streptokinase.     

Turbulent axially directed flow of plasma containing rt-PA promotes thrombolysis of non-occlusive whole blood clots in vitro

The rate of thrombolysis markedly decreases after a thrombosed vessel is partly recanalized and the remaining clot poses serious risk for rethrombosis. We studied in vitro how thrombolysis depends on penetration of plasma containing thrombolytic agents - 0.2 micro g/ml rt-PA or 250 IU/ml streptokinase (SK) - and the magnetic resonance contrast agent Gd-DTPA (at 1 mmol/l) into non-occlusive clots under conditions of fast (turbulent) or slow (laminar) axially directed flow. Cylindrical non-retracted (fresh) or retracted (aged) whole blood clots were pierced lengthways and connected to a perfusion system. Dynamical spin-echo MRI was used for measuring the penetration of labeled plasma into clots and for assessing the remaining clot size. In both types of clots fast flow enhanced the penetration of Gd-DTPA-labeled plasma into clots in comparison to slow flow. In non-retracted clots, lysis with rt-PA and to a lesser extent also lysis with SK followed the path of plasma penetration into clots. After 40 minutes of fast axially directed flow rt-PA resulted in almost complete lysis and SK left only about a third of the clot undissolved, whereas with slow flow lysis was much slower (undissolved clot: 86 +/- 5 % with rt-PA and 95 +/- 1 % with SK). In retracted clots, substantial lysis was possible only with rt-PA and rapid flow (53 +/- 28% of the clot undissolved after 60 min), whereas the use of SK or slow flow precluded meaningful lysis. We conclude that rapid (turbulent) axially directed flow of plasma along non-occlusive blood clots causes forceful exchange of serum inside the clot with outer plasma which enhances both fibrin-specific and non-fibrin-specific lysis of fresh clots. Dissolution of non-occlusive retracted (aged) clots occurs only under fibrin-specific conditions combined with adequate transport of rt-PA into clots.