D. C. Rijken, Erasmus University Medical Center Rotterdam, Department Hematology, Room Ee1393, Dr. Molewaterplein 50, 3015 GE Rotterdam, The Netherlands. Tel.: +31 10 70 44 723; fax: +31 10 70 44 745; e-mail: email@example.com
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 °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 μg 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.
The fibrinolytic system is responsible for the proteolytic degradation of fibrin and may therefore play a role in hemostasis and thrombosis [1,2]. The main components of the system include tissue-type plasminogen activator (tPA) and urokinase-type plasminogen activator (uPA), plasminogen and plasmin, and the inhibitors plasminogen activator inhibitor-1 (PAI-1), α2-antiplasmin (plasmin inhibitor) and thrombin-activatable fibrinolysis inhibitor (TAFI). Other factors, for example those from the contact activation system , might affect fibrinolysis as well.
Most proteins involved in fibrinolysis can nowadays be assayed by specific activity and antigen assays. Global assays for the whole fibrinolytic system, reflecting the net effect of all pro- and antifibrinolytic parameters, are however less well developed, in spite of the numerous global assays described in the literature. At least three major problems hamper the development of a reliable global fibrinolysis test. The first problem is that the apparent base-line fibrinolytic capacity of blood is very low. To solve this problem many global assays eliminate fibrinolytic inhibitors in the sample to be tested, thereby shifting the balance between pro- and antifibrinolytic proteins. Classic examples include the diluted blood clot lysis assay, in which the effectivity of inhibitors is partially eliminated by a tenfold dilution of the blood sample  and the euglobulin clot lysis time assay, in which a euglobulin fraction of plasma is prepared that is relatively poor in inhibitors . Another example is the whole blood clot lysis test in the presence of citrate or ethylenediaminetetraacetic acid (EDTA) , which block calcium-dependent inhibitory mechanisms such as factor (F)XIII/α2-antiplasmin and TAFI. The disadvantage of these approaches is that the outcome of the assays does not reflect the real balance that existed in the original blood sample. Recent global tests solve the problem of the low fibrinolytic capacity of blood by adding purified tPA to the sample to increase the lysis rate [7–10]. The disadvantage of this approach is that the outcome of the assays is relatively insensitive to fluctuations of tPA and PAI-1 in the original samples.
The second problem of many global fibrinolysis tests is that they use plasma samples to assess the fibrinolytic capacity in vivo. Although fibrinolytic proteins are indeed predominantly localized in the plasma compartment of blood, cells such as platelets and leukocytes also contain components of the fibrinolytic system . Platelets in particular contain high amounts of PAI-1. It is therefore preferable to assess the fibrinolytic capacity using whole blood.
The third problem that hampers the development of a reliable global fibrinolysis test is that endogenous tPA is rapidly inhibited by PAI-1 present in the sample. This may even take place during plasma preparation. Blood handling at a low temperature increases the stability of the fibrinolytic activity , but the most effective solution for this problem is to collect blood samples at low pH in so-called Stabilyte collection tubes . This low pH prevents tPA–PAI-1 complex formation in vitro and makes it possible to accurately measure tPA activity in the plasma, but hinders the development of a global clot lysis test.
This paper describes the development of a new test for the global fibrinolytic capacity. The principle of the test is the measurement of clot lysis after a 3-h incubation of a blood clot. The three aforementioned problems are solved by respectively (i) using a sensitive ELISA for the generated fibrin degradation products to determine the extent of lysis; (ii) preparing clots from non-anticoagulated whole blood samples rather than from plasma; and (iii) clotting the blood immediately after collection to prevent tPA inactivation before the test is started.
Materials and methods
Bovine serum albumin (BSA, Boseral) and bovine thrombin were obtained form Organon Teknika (Boxtel, The Netherlands), aprotinin (Trasylol) was from Bayer (Leverkusen, Germany), potato carboxypeptidase inhibitor (PCI) from Calbiochem (EMD Biosciences/Merck Biosciences, Darmstadt, Germany), recombinant tPA (Actilyse) from Boehringer Ingelheim (Ingelheim, Germany), rabbit anti-tPA IgG from Technoclone (Vienna, Austria) and goat anti-PAI-1 IgG from Biopool (Umea, Sweden). Rabbit anti-uPA IgG and control rabbit IgG were prepared as described earlier . Two-milliliter evacuated blood collection tubes (Vacuette, product number 454088) were obtained from Greiner (Kremsmuenster, Austria). Three-milliliter evacuated blood collection tubes containing 1.4 NIH U lyophilized thrombin (stat chemistry tubes) were from Becton Dickinson (Franklin Lakes, NJ, USA).
Test development was performed with blood freshly collected from apparently healthy blood donors (laboratory personnel). The clinical study was performed with 47 patients (37 males and 10 females; average age 61 years, range 33–81 years) who underwent percutaneous transluminal coronary angioplasty (PTCA) at the Leiden University Medical Center. Patients with coronary heart disease were eligible for inclusion if they had been successfully treated with PTCA for stable angina, non-ST-elevation acute coronary syndromes, or silent ischemia. Patients treated for acute ST-elevation myocardial infarction were excluded. Blood was drawn before the PTCA started. The study protocol conforms to the Declaration of Helsinki and was approved by the ethics committee of the medical center. Written informed consent was obtained from each participant prior to the intervention. The same was true for a group of 30 apparently healthy volunteers (15 males and 15 females), with an average age of 50 years (range 18–72 years).
Blood collection and sample preparation
Blood was collected between 09.00 and 12.00 h in the morning, both for the global fibrinolytic capacity test (see below) and for the preparation of citrated plasma and Stabilyte plasma (Stabilyte tubes; Biopool, Umea, Sweden). Platelet-poor plasma was obtained by centrifugation at 2000 × g for 20 min at 4 °C and was stored frozen at −80 °C.
Global fibrinolytic capacity test
Two-milliliter evacuated plastic blood collection tubes were prefilled with 100 μL 0.15 M NaCl, 1 mg mL−1 BSA, 6 NIH U mL−1 thrombin with or without 4000 KIU mL−1 aprotinin by injecting this solution through the rubber stoppers of the tubes. When indicated, varying concentrations of thrombin were used. After blood collection, the contents of the tubes were immediately mixed by inverting them ten times and then incubated for 3 h at 37 °C. The clots were released from the tube wall with a plastic spatula and centrifuged at 2000 × g for 20 min at 4 °C. Aliquots of 500 μL serum (which was slightly hemolysed) were collected, supplemented with 10 μL 10 000 KIU mL−1 aprotinin and stored frozen if the analysis for fibrin degradation products (FnDP) could not be performed on the same day. FnDP was determined using an enzyme immunoassay  (Fibrinostika FbDP; Organon Teknika, Boxtel, The Netherlands). A small number of samples was analyzed with a latex agglutination assay (Auto Dimer, Biopool). These results were multiplied by 1.19 to correct for the slightly lower response of the agglutination assay. The global fibrinolytic capacity (GFC) was determined by subtracting the FnDP level obtained in the blood sample that was incubated in the presence of aprotinin from the FnDP level obtained in the blood sample that was incubated in the absence of aprotinin and was expressed in μg/mL.
The activity levels of tPA and PAI-1 were assayed in Stabilyte plasma, the other levels in citrated plasma. The activity of tPA was determined using a bioimmunoassay , tPA antigen was determined using an enzyme immunoassay (Zymutest tPA antigen; Hyphen BioMed, Andrésy, France), PAI-1 activity was determined using a bioimmunoassay (Chromolize PAI-1; Biopool, Umea, Sweden), TAFI antigen was determined using an enzyme immunoassay with sheep polyclonal antibodies (Affinity Biologicals, Ontario, Canada) as previously described , plasminogen activity was determined in a chromogenic substrate assay using streptokinase and H-D-Val-Leu-Lys-pNA , α2-antiplasmin activity was determined in a chromogenic substrate assay using H-D-Val-Leu-Lys-pNA  and fibrinogen was determined in a clotting rate assay according to Von Clauss . The activity of uPA was determined using a bioimmunoassay as essentially described by Binnema et al.. Citrated plasma samples were incubated in microtiter plates that were coated with a mixture of two monoclonal antibodies, UK 2.1 and UK 26.15 . Immobilized uPA was treated with plasmin (Chromogenix, Instrumentation Laboratory, Milano, Italy) and quantitated with a mixture of plasminogen, ε-aminocaproic acid and the fluorogenic plasmin substrate H-D-Val-Leu-Lys-AMC (Bachem, Bubendorf, Switzerland), using recombinant single-chain uPA (Saruplase; Gruenenthal, Aachen, Germany) as a standard.
Results were given as mean ± SD (statistical significance determined using the Student’s t-test) or as median with range (statistical significance determined using the Mann–Whitney U-test). Pearson’s correlation coefficients between the GFC and other hemostasis factors were determined after log transformation of the GFC data. The coefficients were adjusted for other variables by calculating partial correlation coefficients. Non-parametric correlations were analyzed according to Spearman. All tests were performed with SSPS 11.5 (SSPS Inc., Chicago, IL, USA).
Blood collection in thrombin
The new test for the global fibrinolytic capacity in blood was based on a measurement of the generation of fibrin degradation products (FnDP) during in vitro incubation of clotted whole blood for 3 h at 37 °C. In pilot experiments, we noticed that the generation of FnDP was higher when blood clotted immediately after venipuncture by collecting the samples in thrombin than when blood clotted slowly by contact activation. Thrombin had no effect on the assay of FnDP. The increase was donor dependent. Figure 1 shows the dependency of the fibrin degradation on the thrombin concentration. Blood collection in saline without thrombin (clotting time about 14 min) resulted in the generation of 0.7 μg mL−1 FnDP in 3 h. The fibrin degradation reached an optimum of about 3 μg mL−1 FnDP in 3 h at a thrombin concentration between 0.1 and 1 NIH U/ml (clotting time less than 30 s), and decreased again at higher concentrations. A concentration of 0.3 NIH U/ml of blood was chosen for further experiments.
Mechanism of acceleration of fibrin degradation by thrombin
We hypothesized that the observed acceleration of fibrin degradation by thrombin was as a result of rapid clot formation and strong reduction of the time period during which tPA could be inhibited by PAI-1 in a clot-free milieu. To further study this hypothesis, citrated blood was collected from a donor with a high plasma PAI-1 activity of 239% compared with pooled normal plasma (subject 1 in Fig. 2) and from a donor with a low plasma PAI-1 activity of 59% (subject 2 in Fig. 2). Blood samples were pre-incubated without or with anti-PAI-1 IgG to block PAI-1 activity and then supplemented with tPA to restore a physiological tPA activity and immediately clotted at 37 °C with calcium only (clotting time 12 min) or with calcium and thrombin (clotting time 1–2 min). Figure 2 shows the generation of FnDP after 3 h in the samples. In subject 1 with high PAI-1 activity and no added anti-PAI-1 IgG, rapid clotting with thrombin increased the FnDP generation 2.95-fold (P < 0.001). In subject 1 with added anti-PAI-1 IgG, rapid clotting with thrombin did not increase the FnDP generation (1.06-fold, P = 0.083). In subject 2 with low PAI-1 activity and no added anti-PAI-1 IgG, rapid clotting with thrombin increased the FnDP generation significantly but only 1.36-fold (P < 0.001). In subject 2 with added anti-PAI-1 IgG, rapid clotting with thrombin did not increase the FnDP generation (1.08-fold, P = 0.252). These results showed that the presence of PAI-1 activity in plasma determined whether or not thrombin accelerated the fibrin degradation. This also explained the observation that the acceleration was donor dependent. The accelerating effect of anti-PAI-1 IgG, even in subject 2 with a low plasma PAI-1 activity, may be partly ascribed to the neutralization of platelet PAI-1 released during or after clot formation.
Test for global fibrinolytic capacity
A test for the GFC in blood was set up by collecting one blood sample in thrombin as indicated above, and a parallel sample in thrombin plus aprotinin to block in vitro fibrin degradation by plasmin. Both samples were then incubated for 3 h at 37 °C. FnDP levels obtained in the presence of aprotinin were subtracted from the FnDP levels obtained in the absence of aprotinin to correct for the FnDP levels already present in the blood before the venipuncture. The corrected FnDP level was defined as the GFC. Figure 3 shows the results obtained from duplicate measurements in 47 patients with coronary artery disease. The median FnDP level obtained in the absence of aprotinin was 2.55 μg mL−1 (range 0.24–32.7 μg mL−1). The median FnDP level obtained in the presence of aprotinin was about tenfold lower: 0.24 μg mL−1 (range 0.10–0.80 μg mL−1). The median GFC was 2.17 μg mL−1 (range 0.10–32.1 μg mL−1). The GFC significantly correlated with the FnDP levels obtained in the presence of aprotinin (Spearman’s rho = 0.60, P < 0.001). The detection limit of the GFC test was 0.1 μg mL−1 and the coefficient of variation, calculated from the duplicate measurements, was 22%.
The GFC was also determined in a group of 30 apparently healthy subjects. The median GFC was 2.06 μg mL−1 (range 0.13–13.6 μg mL−1). The coefficient of variation, calculated from the duplicate measurements, was 24%. The GFC in this group did not differ from the GFC in the patient group (P = 0.86, Mann–Whitney U-test). However, it should be noted that the two groups were not properly matched.
Three subjects were tested before and after strenuous physical exercise (running for 20 min). The GFC increased significantly from 3.3 to 320 μg mL−1 (ninety-sevenfold), from 1.7 to 9.5 μg mL−1 (sixfold) and from 5.3 to 490 μg mL−1 (ninety-twofold), respectively.
Correlations between GFC and other hemostasis factors
GFC values obtained from the patients were compared with various hemostasis parameters determined in plasma. Figure 4 shows scatter plots and Table 1 the correlation coefficients. The unadjusted correlation coefficients were statistically significant for tPA activity (r = 0.51), PAI-1 activity (r = −0.54) and fibrinogen levels (r = 0.36). The correlation with tPA activity became weaker (r = 0.31) but remained significant after adjustment for PAI-1 activity. Similarly, the correlation with PAI-1 activity became weaker (r = −0.40) but remained significant after adjustment for tPA activity. After adjustment for tPA or PAI-1 activity, the significant correlation between GFC and fibrinogen levels (r = 0.36) hardly changed.
Table 1. Correlations between the global fibrinolytic capacity (GFC) in whole blood and other hemostasis parameters determined in plasma of the 47 patients shown in Fig. 4
Adjusted for tPA activity
Adjusted for PAI-1 activity
Adjusted for fibrinogen
The correlation coefficients were determined with a Pearson’s correlation test after log transformationof the GFC data. ND, not determined, because uPA activity was not normally distributed. A Spearman correlation test between GFC and uPA activity yielded a non-significant rho of 0.02 (P = 0.90). ***P < 0.001, **P < 0.01, *P < 0.05. tPA, tissue-type plasminogen activator; PAI-1, plasminogen activator inhibitor-1; uPA, urokinase-type plasminogen activator; TAFI, thrombin-activatable fibrinolysis inhibitor.
No significant correlations were found with tPA antigen, uPA activity, TAFI antigen, plasminogen and α2-antiplasmin levels. Interestingly, the correlation with tPA antigen became stronger (r = 0.30) and significant after adjustment for PAI-1 activity, and the correlation with α2-antiplasmin became stronger (r = 0.29) and significant after adjustment for fibrinogen. Two patients had clearly elevated uPA activity levels (Fig. 4C, patients 19 and 36). These two patients also showed a relatively high GFC, which is particularly evident in the GFC vs. tPA activity scatter plot (Fig. 4A). This suggested that at least elevated uPA activities contribute to the GFC.
The potential involvement of tPA and uPA was further studied by incorporating quenching antibodies into blood clots prepared from three apparently healthy donors. Anti-uPA IgG slightly inhibited fibrin degradation in two out of three donors. Anti-tPA IgG strongly inhibited fibrin degradation in all donors. A combination of the antibodies fully inhibited the generation of fibrin degradation products (Fig. 5).
The role of TAFI in fibrin degradation
In order to study whether or not TAFI inhibits fibrin degradation under the experimental conditions of the fibrinolysis test, multiple blood samples from a single donor were collected either in thrombin or in thrombin plus potato carboxypeptidase inhibitor (PCI, 15 μg mL−1 blood), a specific inhibitor of activated TAFI. During the incubation period of 3 h at 37 °C, 1.12 ± 0.18 μg mL−1 FnDP (mean ± SD, n = 3) were generated in the absence of PCI and 8.8 ± 1.8 μg mL−1 FnDP (mean ± SD, n = 3) in the presence of PCI. The 8-fold increase in fibrin degradation in the presence of PCI indicated that TAFI played a significant role.
We developed a new test for the global fibrinolytic capacity in which we used a sensitive ELISA to quantify the fibrin degradation products generated during a 3-h incubation of a whole blood clot that was formed in vitro immediately after blood collection.
An essential observation for the development of our test was that whole blood clot lysis increased significantly when blood was collected in thrombin (Fig. 1). This could possibly be ascribed to inactivation of PAI-1 by thrombin . However, ancrod which is unable to inactivate PAI-1 , showed a similar effect (not shown). Therefore it appeared more likely that rapid clot formation stabilized the fibrinolytic activity, as has already been suggested some five decades ago . We provided evidence that the fibrinolysis promoting effect of blood collection in thrombin involved PAI-1 (Fig. 2). Reduction of clotting time from about 14 min to less than 1 min may indeed substantially reduce the inactivation of tPA by PAI-1 in the fluid phase, in particular at an elevated PAI-1 concentration. After the clot has formed, tPA is bound to fibrin and is to some extent protected from inactivation by PAI-1 [25,26]. The decrease of whole blood clot lysis at a high thrombin concentration (Fig. 1) could possibly be ascribed to the well-known antifibrinolytic actions of elevated thrombin levels, such as the induction of a fine fibrin network, enhanced activation of FXIII and TAFI, increased inactivation of single-chain uPA and stimulated release of PAI-1 from platelets .
The new test was applied to an arbitrary group of patients (in this case patients with coronary artery disease) in order to evaluate whether the test might be usable in a clinical environment. A specific question was whether the in vitro generated FnDP levels were sufficiently high in comparison with the FnDP levels already present in vivo (and measured in the blood collection tube containing aprotinin). The results showed that this was the case, in spite of the fact that a substantial number of patients showed enhanced FnDP levels in vivo (i.e. above 310 ng mL−1 ). The in vitro-generated FbDP levels were about 10 times higher than the in vivo levels (Fig. 3). A subsequent study included apparently healthy volunteers and established the normal range of the GFC test. Physical exercise, known to increase the fibrinolytic capacity in blood, increased the outcome of the test 5- to one hundredfold in three healthy volunteers. The normal range was wide (0.13–13.6 μg mL−1), which is in line with the wide variation in fibrinolytic activity, as measured with other assays, but which will limit the applicability of the test. Nevertheless, the utility of the GFC test was recently demonstrated in a study to the existence of hyperfibrinolysis in liver cirrhosis .
In blood, FnDP or D-dimer levels are frequently used in the literature either as an indicator of the GFC in blood or as an indicator of the amount of fibrin present in the vasculature, for instance in the diagnosis of pulmonary embolism. The two approaches are in theory not fully justified, because FnDP levels are probably determined by both the fibrinolytic capacity and the amount of fibrin present. Our observation that the GFC correlated with the FnDP levels obtained in the presence of aprotinin (Spearman’s rho = 0.60, P < 0.001) suggested a significant contribution of the global fibrinolytic capacity to the FnDP levels in the circulation of this group of patients.
The significant correlations between the GFC and a number of hemostasis factors in the plasma of the patients (Fig. 4 and Table 1) suggested, not unexpectedly, that tPA activity and PAI-1 activity as well as (to a lesser extent) tPA antigen and α2-antiplasmin affected the GFC. The significant correlation with fibrinogen levels was possibly related to the use of the endogenous fibrinogen in the blood sample for the preparation of the fibrin substrate in the test. At higher fibrinogen levels more fibrin is available and higher FnDP levels are generated. To obtain an estimate of the global fibrinolytic capacity which is independent of fibrinogen, the GFC values could be divided by the fibrinogen levels. Other hemostasis factors will probably also affect the GFC, but the variations of these factors in this group of patients were too small to demonstrate significant correlations with the GFC. Studies in patient populations with larger variations may reveal additional correlations with the GFC.
uPA plays a major role in non-hemostatic functions of the plasminogen/plasmin system but its role in blood clot lysis is frequently questioned. Therefore it is interesting that the two patients with an elevated uPA activity level also showed an elevated GFC (Fig. 4A,C), suggesting that plasma uPA could also be involved in clot lysis. The involvement of uPA in our test system might be as a result of the use of whole blood, because the presence of platelets seems to be essential for the contribution of plasmatic uPA activity . We were unable to identify any clinical parameter that could explain the elevated uPA levels in the two patients.
The role of tPA and uPA was further substantiated by the use of antibodies against tPA and uPA. The GFC of three healthy donors was partially inhibited by the separate antibodies and fully inhibited by a combination of the antibodies.
The role of TAFI in various model systems is not yet fully clear. Therefore we have specifically addressed this question by incorporating PCI, a specific inhibitor of activated TAFI, into the blood clot of our test system. The 8-fold increase in fibrin degradation pointed to a significant role and was in line with a recent study which showed by the use of thromboelastography that inhibition of TAFI augments tPA-induced fibrinolysis in whole blood . Because TAFI is activated by thrombin which is generated during and after clot formation, the significant role of TAFI suggests that the thrombin-forming capacity of blood samples affects the outcome of the assay. This has still to be studied experimentally. The absence of a significant correlation between the GFC and TAFI antigen in the patient group is a bit surprising. This may be ascribed to the genotype-dependency of the ELISA . However, a comparable observation was recently described by Lisman et al. .
The important role of the plasminogen/plasmin system in the new test does not exclude the possibility that other enzymes, such as leukocyte elastase and cathepsin G, could contribute to blood clot lysis. However, the design of the test (i.e. difference in lysis with and without aprotinin) implies that only lysis by aprotinin-sensitive proteases is detected. Elastase and cathepsin G are not efficiently inhibited by aprotinin  and probably do not contribute to the outcome of the test. Furthermore, the ELISA for FnDP is highly specific for plasminogen activator-induced fibrin degradation products . Taken together, our new test can be considered as a global test for plasminogen/plasmin-mediated blood clot lysis.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.