Title: Discovery of thrombin activatable fibrinolysis inhibitor (TAFI)
Abstract: In the early 80s, we became interested in the profibrinolytic effects of activated protein C (APC) [1Comp P.C. Esmon C.T. Generation of fibrinolytic activity by infusion of activated protein C into dogs.J Clin Invest. 1981; 68: 1221-8Crossref PubMed Google Scholar, 2Taylor F.B. Lockhart M.S. Whole blood clot lysis: in vitro modulation by activated protein C.Thromb Res. 1984; 37: 639-49Abstract Full Text PDF Google Scholar]. In preliminary experiments, we had confirmed some of the observations of Taylor [2Taylor F.B. Lockhart M.S. Whole blood clot lysis: in vitro modulation by activated protein C.Thromb Res. 1984; 37: 639-49Abstract Full Text PDF Google Scholar] using a whole blood clot lysis system and we decided to combine the expertise of the Gaubius Institute (fibrinolysis) and the Haemostasis and Thrombosis Research Laboratory (coagulation) to find an explanation for these profibrinolytic effects of APC. In 1984, Nanneke de Fouw started to work on this project for her PhD and during the next 3 years, she identified the neutralization of PAI‐1 [3De Fouw N.J. Haverkate F. Bertina R.M. Koopman J. Van Wijngaarden A. Van Hinsbergh V.W. The cofactor role of protein S in the acceleration of whole blood clot lysis by activated protein C in vitro.Blood. 1986; 67: 1189-92Crossref PubMed Google Scholar, 4De Fouw N.J. Van Hinsbergh V.W. De Jong Y.F. Haverkate F. Bertina R.M. The interaction of activated protein C and thrombin with the plasminogen activator inhibitor released from endothelial cells.Thromb Haemost. 1987; 57: 176-82Crossref PubMed Scopus (0) Google Scholar, 5De Fouw N.J. De Jong Y.F. Haverkate F. Bertina R.M. Activated protein C increases fibrin clot lysis by neutralization of plasminogen activator inhibitor. No evidence for a cofactor role of protein S.Thromb Haemost. 1988; 60: 328-33Crossref PubMed Scopus (0) Google Scholar, 6De Fouw N.J. De Jong Y.F. Haverkate F. Bertina R.M. The influence of thrombin and platelets on fibrin clot lysis rates in vitro. A study using a clot lysis system consisting of purified proteins.Fibrinolysis. 1988; 2: 235-44Crossref Google Scholar] and the inhibition of thrombin formation [7De Fouw N.J. Haverkate F. Bertina R.M. Protein C and fibrinolysis: a link between coagulation and fibrinolysis.Adv Exp Med Biol. 1990; 281: 235-43Crossref PubMed Google Scholar, 8De Fouw N.J. Van Tilburg N.H. Haverkate F. Bertina R.M. Activated protein C accelerates clot lysis by virtue of its anticoagulant activity.Blood Coagulation and Fibrinolysis. 1993; 4: 201-10Crossref PubMed Google Scholar] as the mechanisms responsible for the profibrinolytic effects of APC. Nanneke hypothesized that thrombin inhibits fibrinolysis via one or more unknown components and that as a consequence each anticoagulant has profibrinolytic properties while each procoagulant would have antifibrinolytic properties [9De Fouw NJ. Activated protein C and fibrinolysis. Evidence for a profibrinolytic effect of an anticoagulant. PhD Thesis. Leiden: University of Leiden, 1988.Google Scholar]. The seminal observation was that APC accelerated tPA‐dependent clot lysis only when there was a simultaneous thrombin generation through the coagulation system. Inhibition of thrombin formation – either by depletion of the plasma of prothrombin or factor X, or by addition of APC (but not DIP‐APC)/Ca2+/phospholipids or heparin – always resulted in the same acceleration of the tPA‐dependent clot lysis of platelet‐free plasma. Nanneke de Fouw [9De Fouw NJ. Activated protein C and fibrinolysis. Evidence for a profibrinolytic effect of an anticoagulant. PhD Thesis. Leiden: University of Leiden, 1988.Google Scholar] published her PhD thesis in 1988 and some of the major results were published in the proceedings of a congress in Taiwan [7De Fouw N.J. Haverkate F. Bertina R.M. Protein C and fibrinolysis: a link between coagulation and fibrinolysis.Adv Exp Med Biol. 1990; 281: 235-43Crossref PubMed Google Scholar]. It took us, however, 5 years before a journal finally accepted the manuscript of de Fouw et al. [8De Fouw N.J. Van Tilburg N.H. Haverkate F. Bertina R.M. Activated protein C accelerates clot lysis by virtue of its anticoagulant activity.Blood Coagulation and Fibrinolysis. 1993; 4: 201-10Crossref PubMed Google Scholar], in which she reported that ‘APC accelerates clot lysis by virtue of its anticoagulant activity’. We had the feeling that reviewers just did not believe our explanation of the experimental data. After Nanneke de Fouw left our department, Nico van Tilburg continued with the identification of what we called the ‘de Fouw Factor’. His studies showed that thrombin dose dependently inhibited the tPA‐dependent clot lysis in bariumcitrate/Al(OH)3 adsorbed plasma but not in a system of purified components (fibrinogen, plasminogen, FXIII) [10Van Tilburg N.H. De Fouw N.J. Haverkate F. Bertina R.M. Thrombin inhibits t‐PA dependent clot lysis via an unidentified plasma component.Thromb Haemost. 1991; 65: 662Google Scholar, 11Bertina R.M. Van Tilburg N.H. De Fouw N.J. Haverkate F. Thrombin, a link between coagulation activation and fibrinolysis.Ann NY Acad Sci. 1992; 667: 239-48Crossref PubMed Scopus (0) Google Scholar]. Thrombin needed its active site and fibrinogen/fibrin was needed for the expression of the thrombin‐induced antifibrinolytic effect, but the thrombin‐binding site on fibrin seemed not to be involved [12Van Tilburg N.H. Haverkate F. Bertina R.M. Anti‐fibrinolytic action of thrombin in t‐PA dependent plasma clot lysis.Fibrinolysis. 1992; 6: 75Google Scholar]. By using selectively depleted plasmas (e.g. pro‐urokinase deficient) and plasma of patients with congenital deficiencies (e.g. factor XIII deficiency, α2‐antiplasmin deficiency) in the clot lysis experiments, we could exclude that one of the known coagulation or fibrinolysis proteins was the ‘de Fouw factor’. At the same time, we were adding new components in physiological concentrations (α2‐antiplasmin, pro‐urokinase, PAI‐1, VWF, fibronectin, IgG, albumin) to our purified clot lysis system, which, however, remained insensitive to thrombin. During this period, we followed the publications of Bajzar and Nesheim [13Bajzar L. Fredenburgh J.C. Nesheim M. The activated protein C‐mediated enhancement of tissue‐type plasminogen activator‐induced fibrinolysis in a cell‐free system.J Biol Chem. 1990; 265: 16948-54Abstract Full Text PDF PubMed Google Scholar, 14Bajzar L. Nesheim M. The effect of activated protein C on fibrinolysis in cell‐free plasma can be attributed specifically to attenuation of prothrombin activation.J Biol Chem. 1993; 268: 8608-16Abstract Full Text PDF PubMed Google Scholar] not only with great interest, but also with increasing confusion. They reported a profibrinolytic effect of APC in plasma, which disappeared after bariumcitrate adsorption, and were able to reconstitute the APC sensitivity of t‐PA‐dependent clot lysis by adding two different fractions of a fractionated barium citrate eluate to the purified fibrinogen and plasminogen. We did not find a profibrinolytic effect of APC in plasma unless we had added phospholipids, and we found that the thrombin activatable fibrinolysis inhibitor (or the ‘de Fouw factor’) was present in bariumcitrate‐adsorbed plasma (and not in the eluate). Finally, we never could demonstrate inhibition of t‐PA‐dependent clot lysis by thrombin in a system of purified components. Reading the interesting report of Nesheim and Bajzar [15Nesheim M. Bajzar L. The discovery of TAFI.J Thromb Haemost. 2005; 3: 2139-46Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar] on the history of the discovery and characterization of TAFI in their laboratory did help a lot in clarifying the confusion of that time. Apparently, they did not use platelet‐free plasma like we did, while their plasminogen preparations contained TAFI in contrast to ours, which were TAFI‐free. Two small details of which the importance was not recognized at that time, but together they satisfactorily explained the apparent discrepancies. In 1995, Bajzar et al. [16Bajzar L. Manuel R. Nesheim M.E. Purification and characterisation of TAFI, a thrombin‐activable fibrinolysis inhibitor.J Biol Chem. 1995; 270: 14477-84Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar] reported the purification of the Thrombin Activable Fibrinolysis Inhibitor and its identification as procarboxypeptidase B. After Eaton et al.'s [17Eaton D.L. Malloy B.E. Tsai S.P. Henzel W. Drayna D. Isolation, molecular cloning, and partial characterization of a novel carboxypeptidase B from human plasma.J Biol Chem. 1991; 266: 21833-8Abstract Full Text PDF PubMed Google Scholar] report in 1991, that procarboxypeptidase B can be activated by thrombin, we hypothesized that the ‘de Fouw factor’ could be procarboxypeptidase B. To test this hypothesis, we studied the effect of the carboxypeptidase B inhibitor GEMSA on the thrombin‐induced inhibition of tpA‐dependent clot lysis in Al(OH)3‐adsorbed plasma. We did not find any effect of 0.5–2.0 mm GEMSA on the clot lysis time, while porcine carboxypeptidase B was almost completely inhibited at 10‐fold lower concentrations. Therefore, we concluded that procarboxypeptidase B was not the ‘de Fouw factor’. We should have been more careful! Today, we know that a reversible inhibitor of TAFIa like GEMSA has both stimulatory and inhibitory effects on tPA‐dependent plasma clot lysis [18Walker J.B. Hughes B. James I. Haddock P. Kluft C. Bajzar L. Stabilisation versus inhibition of TAFIa by competitive inhibitors in vitro.J Biol Chem. 2003; 278: 8913-21Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar, 19Schneider M. Nesheim M. Reversible inhibitors of TAFIa can both promote and inhibit fibrinolysis.J Thromb Haemost. 2003; 1: 147-54Crossref PubMed Scopus (0) Google Scholar]. We had been just unlucky in selecting these particular GEMSA concentrations.