Plasmin and the thrombin–thrombomodulin complex both contribute to thrombin-activatable fibrinolysis inhibitor activation in whole blood model thrombi


Ann Gils, Laboratory for Pharmaceutical Biology, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Campus Gasthuisberg, O&N2, PB 824, Herestraat 49, B-3000 Leuven, Belgium.
Tel.: +32 16 32 34 36 ; fax: +32 16 32 34 60.

Activated thrombin activatable fibrinolysis inhibitor (TAFIa) is formed upon activation of TAFI by thrombin, either alone or in complex with thrombomodulin (T/TM complex), and by plasmin [1,2]. In vitro clot lysis experiments in human plasma revealed that the T/TM complex is the predominant activator of TAFI [3]. This finding was substantiated by an in vivo study in which TAFIa generation during Escherchia coli-induced sepsis in baboons was inhibited by a monoclonal antibody (MA) specifically inhibiting the T/TM-mediated activation of human TAFI [4]. Recent studies from our laboratory using specific MAs that mainly inhibit plasmin-mediated TAFI activation revealed that plasmin functions as an activator of TAFI during plasma clot lysis [5,6]. This was confirmed in an in vivo mouse thromboembolism model in which the analyzed MAs exhibited a strong profibrinolytic effect [5].

In this study, we further evaluate the contribution of TAFIa in the inhibition of fibrinolysis and in particular the importance of the different activators of TAFI. The effect of MAs that impair TAFI activation via different mechanisms was assessed in a whole blood thrombus lysis model [7,8]. The MAs analyzed were (i) MA-TCK11A9 that selectively inhibits the plasmin-mediated human TAFI activation, (ii) MA-T12D11 that selectively inhibits the T/TM-mediated human TAFI activation and (iii) MA-TCK27A4 that inhibits the thrombin-, T/TM- and plasmin-mediated human TAFI activation [6,9].

Model thrombi were formed in a Chandler loop system as previously described [7,8]. Briefly, FITC-labeled fibrinogen was added to citrated whole blood, followed by the addition of phosphate-buffered saline (PBS), potato tuber carboxypeptidase inhibitor (PTCI, 25 μg mL−1; Calbiochem, Beeston, Notts, UK) or MAs. The MAs were incorporated into forming thrombi at a five-fold molar ratio (65 μg mL−1) over TAFI (assuming a plasma content of 55% in blood and a TAFI concentration of 10 μg mL−1 in human plasma [10,11]). The mixture was recalcified and rotated at 30 rpm giving a shear rate of 428 s−1 for 90 min at room temperature to form model thrombi. The resulting thrombi were subsequently bathed in PBS containing 0.25 μg mL−1 human tissue plasminogen activator. Thrombi were incubated at 37 °C and samples removed at defined times to quantify the released fluorescence that reflects fibrin degradation and thus fibrinolysis. The effect of the inhibitory MAs was compared with PTCI, a small molecule TAFIa inhibitor.

All TAFI inhibitors analyzed in this study resulted in a significant increase (< 0.05; repeated measures anova) in fluorescence release (relative to no inhibitor; Fig. 1A). Interestingly, the MAs that inhibit either plasmin-mediated TAFI activation (MA-TCK11A9) or that inhibit activation through all TAFI activators (MA-TCK27A4) resulted in more rapid lysis than PTCI i.e. a 1.92-fold and 1.90-fold increase in fluorescence release for MA-TCK11A9 and MA-TCK27A4, respectively, compared with a 1.46-fold increase for PTCI. On the other hand, MA-T12D11 that exclusively inhibits the T/TM-mediated TAFI activation revealed only a 1.34-fold increase in fluorescence release. These data illustrate that MAs that impair TAFI activation are profibrinolytic when incorporated into thrombi. This suggests that TAFIa can also exert an antifibrinolytic effect within the environment of the thrombus and not just at the thrombus–plasma interface as previously described [7]. The difference in findings is most likely explained by the high specificity of the inhibitory MAs used here compared with the use of PTCI alone in Mutch et al. [7].

Figure 1.

 Activated thrombin activatable fibrinolysis inhibitor (TAFIa) in model thrombi. (A) Lysis of model thrombi upon incorporation of TAFI inhibitors. Potato tuber carboxypeptidase inhibitor (PTCI; 25 μg mL−1) or monoclonal antibody (MA) directed towards TAFI (MA-TCK11A9, MA-T12D11, MA-TCK27A4; 65 μg mL−1) was incorporated in blood prior to thrombus formation. Thrombi were bathed in phosphate-buffered saline (PBS) containing 0.25 μg mL−1 tissue plasminogen activator (tPA). Lysis was monitored as fluorescence release. Duplicate thrombi were prepared from three donors that gave blood on at least two separate occasions. Data are expressed as mean ± standard deviation (SD). (B) Western blot analysis of thrombus extracts. Recombinant human TAFI (21 nmol L−1) partially activated by T/TM (1/0.25 nmol L−1; lane 1) and thrombus extracts of one representing donor are shown (lane 2–6). PTCI or MA was incorporated into blood prior to thrombus formation. Arrows indicate the migration of zymogen TAFI (56 kDa) and active TAFIa (36 kDa). The Western blot depicted in this figure is representative for thrombi extracts of two out of the three donors.

Our observations also indicate a difference in fluorescence release and thus the profibrinolytic effect between the evaluated TAFI inhibitors; with more rapid lysis observed on inclusion of MAs that inhibit plasmin-mediated TAFI activation (MA-TCK11A9 and MA-TCK27A4) compared with the MA that exclusively inhibits T/TM-mediated TAFI activation (MA-T12D11). This suggests that during model thrombus formation, TAFI activation is established to a significant extent via plasmin with the T/TM complex being a less prominent activator. This is an important finding in the current discussion regarding the relative contribution of the various TAFI activators (reviewed in [12,13]).

The presence of TAFIa within thrombi is supported by Western blot analysis. Extracts of model thrombi from three donors were prepared as described previously [14] and were subjected to SDS-PAGE followed by Western blot analysis with a monoclonal antibody that detects intact and activated TAFI (MA-T3D8 [15]). In all thrombi formed in the absence of any TAFI inhibitor both zymogen TAFI (56 kDa) and TAFIa (36 kDa) were present (Fig. 1B). In a previous study, only zymogen TAFI was detected in whole blood model thrombi [7]; however, the antibody used weakly recognized TAFIa. The presence of TAFI(a) in thrombi is also supported by the previous observation that TAFI is cross-linked to fibrin via the transglutaminase factor XIIIa, thereby contributing to clot stability [16,17]. Thrombi formed in the presence of PTCI also showed clear bands of TAFI and TAFIa. In the presence of MA-TCK27A4, one of the most potent TAFI inhibitors, the TAFIa band is virtually absent. However, in the presence of MA-TCK11A9, the other TAFI inhibitor with a high profibrinolytic activity, TAFIa is still detectable. We therefore cannot exclude the possibility that MA-TCK11A9 exerts additional effects on TAFI(a) besides the inhibition of the plasmin-mediated TAFI activation, for example destabilization of TAFIa or interference in the binding of TAFIa to fibrin. Surprisingly, the addition of the TAFI inhibitor with the lowest potency in our model, MA-T12D11, reduced the presence of TAFIa in two out of three donor thrombi; thus indicating a role for T/TM in TAFI activation. In the third donor there is a less pronounced effect of MA-T12D11, however, interpretation of this result is hampered by the presence of a number of additional bands (data not shown). It is highly unlikely that the MAs induce any off-target effect as none of the MAs have an effect on clot lysis in TAFI-depleted plasma [6].

To conclude, this study highlights that TAFIa exerts an anti-fibrinolytic function within the local environment of whole blood model thrombi. Furthermore, we provide clear evidence that plasmin and the T/TM complex both contribute to TAFI activation in this setting.

Disclosure of Conflict of Interest

The authors state that they have no conflict of interest.