Evidence for a Factor IX-Independent Role for Factor XI in Thrombin Generation

Abstract 2244

In the widely used activated partial thromboplastin time (aPTT) assay, fibrin formation is induced by a series of sequential activations of the plasma protease zymogens factor (f) XII, fXI, fIX, fX and prothrombin, in that order. Conversion of prothrombin to the protease α-thrombin results in fibrin formation. α-Thrombin also enhances its own generation through activation of the cofactors fV and fVIII. While the linear sequence of reactions in the aPTT implies that loss of any single protease should have a comparable deleterious effect on the system, it is recognized that complete deficiency of a protein near the start of the sequence (e.g. fXII or fXI) results in greater aPTT prolongation than deficiency of proteins further down the sequence (e.g. fIX). This implies that proteases activated early in the process have multiple plasma substrates. For example, fXIa was recently reported to activate fVIII and fV (JTH 8;1532:2010), in addition to its role in fIX activation. Here, we present evidence that fXIa contributes to α-thrombin generation in the absence of fIX through activation of fX and/or fV.

We noted that an anti-fXI antibody (O1A6) prolonged the aPTT of plasma from a patient with severe hemophilia B (fIX antigen undetectable) or plasma immunodepleted of fIX. This observation held even when an anti-fIX antibody was added to neutralize potential traces of fIX. Addition of activated fXI (fXIa - 3 nM) directly to fIX-deficient recalcified plasmas induced clot formation, and the time to clot formation was prolonged by O1A6. To further exclude the possibility that traces of fIX were contributing to thrombin generation, we confirmed the results using plasma from mice with combined complete deficiencies of fXII, fXI, and fIX. We tested the capacity of fXIa to cleave/activate fX and fV, the protease zymogen and cofactor, respectively, immediately downstream of fIX in the coagulation cascade. FX, the zymogen of the protease fXa, is evolutionarily related to fIX. SDS-PAGE analysis confirmed that fXIa cleaves fX. FX cleaved by fXIa demonstrated fXa activity in a chromogenic substrate assay, and converted prothrombin to α-thrombin in the presence of fVa and phospholipid. As previously reported, fXIa readily cleaved fV. The cleavage pattern differed from that generated by α-thrombin, however, formation of the fVa light chain was clearly evident. In a plasma clotting assay designed to measure either fXa or fVa activity, fX or fV pre-incubated with fXIa significantly shortened the clotting time of fIX-deficient plasma, while fX or fV pre-incubated with vehicle did not.

In thrombin generation assays, fXIa (1.25 to 15 nM) induced thrombin generation in fIX-deficient plasma supplemented with anti-fIX antibody in a concentration dependent manner. FXIa did not induce thrombin generation in plasma lacking fV, or in fIX-deficient plasma containing the fXa inhibitor apixaban. This indicates that fXIa is working at the level of fX/fV in this assay, and is not directly converting prothrombin to α-thrombin. A recombinant variant of fXIa lacking the major fIX-binding exosite (fXIaPKA3, J Biol Chem 1996;271:29023) demonstrated a marked defect, compared to wild type fXIa (fXIaWT), in its capacity to induce thrombin generation in normal plasma. However, in fIX-deficient plasma, fXIaPKA3 and fXIaWT are comparable in their ability to enhance thrombin generation, supporting the premise that fXIa is acting through activation of fX and/or fV in the absence of fIX.

Previously, we observed that fXI deficient mice and fIX deficient mice are comparably resistant to carotid artery thrombosis induced by exposure of the vessel to ferric chloride, despite having very different propensities to bleed. The animals were uniformly resistant to thrombosis with 5% FeCl₃, and some were resistant at 7.5% FeCl₃. All experienced vessel occlusion with 10% FeCl₃. This is consistent with fXIa contributing to thrombosis in this model through fIX activation. However, we observed that some mice with combined fIX and fXI deficiency were resistant to FeCl₃ concentrations up to 12.5%, implying that fXIa was contributing to thrombosis in a fIX-independent manner, as well. These results are consistent with those from the in vitro assays described above, and support the hypothesis that fXIa contributes to thrombin generation through fIX-dependent and fIX-independent processes.