Defective TAFI Activation in Hemophilia Worsens Joint Bleeding

Background: Joint bleeds are common in hemophilia A or B and cause hemophilic arthropathy. It is clinically well recognized that patients with acquired hemophilia generally do not suffer from joint bleeding, but the molecular mechanisms responsible for the difference in joint bleeding tendency between acquired and congenital hemophilia are ill defined. FVIII deficiency causes defective thrombin generation, impaired coagulation, and increased fibrinolysis. The latter is caused by impaired activation of thrombin activatable fibrinolysis inhibitor (TAFI). Our previous plasma-based analyses showed that clotting and thrombin generation were readily inhibited by an anti-FVIII antibody, whereas a 10-fold higher antibody concentration was required to inhibit thrombin-mediated TAFI activation. We hypothesize that residual TAFI activation occurring in acquired hemophilia, but not in congenital hemophilia, protects against joint bleeding. Here, we determine whether TAFI activation prevents joint bleeding in a mouse model of acquired hemophilia.

Methods and results: A transient (anti-FVIII) acquired hemophilia A model was set up to compare joint bleeding in wild type (WT) vs. TAFI^-/- mice. Joint bleeding was induced by a subpatellar needle puncture in the right knee. This model caused considerable joint bleeding in FVIII^-/- mice as evidenced by the decreased hematocrit (Hct) 2 days post injury (D2 Hct) (D2 Hct= 29 ± 11 % (n= 9) vs. baseline Hct (46 ± 2 %); p< 0.0001). A single injection of the FVIII inhibiting antibody (GMA-8015; 0.25 mg/kg) in WT mice caused acquired hemophilia for up to 72 hours as evident from increased tail bleeding similar to that observed in FVIII^-/- mice. Consistent with clinical findings, only minimal joint bleeding was observed in inhibitor-treated WT mice (D2 Hct= 44 ± 4 % (n= 15) for BALB/c and 40 ± 4 % (n= 17) for C57Bl/6J). Significant joint bleeding (D2 Hct= 36 ± 9% (n= 12) for C57Bl/6J; p< 0.05) could be induced by a higher dose of inhibitor (1 mg/kg), however bleeding remained considerably less severe than that observed in FVIII^-/-mice.

In vitro, the FVIII inhibitor readily inhibited thrombin generation but was relatively ineffective in inhibiting TAFI activation. Therefore, we tested our hypothesis that continued TAFI activation prevented severe joint bleeding in the inhibitor-treated WT mice. Indeed, administration of the FVIII inhibitor (0.25 mg/kg) in TAFI^-/-mice resulted in excessive joint bleeding (D2 Hct= 25 ± 8 %; n= 14; p< 0.0001). Similarly, joint bleeding in WT mice was increased significantly when the FVIII inhibitor was co-administered with an inhibitory antibody against TAFI (D2 Hct= 34 ± 7 %; n= 13; p< 0.01). In contrast, TAFI deficiency did not increase tail bleeding with or without FVIII inhibitor, as determined by acute blood loss, 24-hour mortality, and Hct of the survivors at 24 hours post tail resection. These data clearly demonstrate that different vascular beds empower different mechanisms to curb bleeding and suggest that the protective effects of TAFI are specifically relevant for the vascular beds of the synovial joint.

Activated TAFI (TAFIa) conveys multiple functions, including anti-fibrinolytic effects and numerous anti-inflammatory activities. Interestingly, tranexamic acid (TXA), a Lys analogue and potent anti-fibrinolytic agent, added at 50 mg/ml to the drinking water, did not reduce joint bleeding in FVIII^-/- mice or TAFI^-/- mice with the FVIII inhibitor, whereas TXA did correct tail bleeding in these mice. This suggests that the protective effects of TAFI on joint bleeding were independent of its anti-fibrinolytic effects and may result from its anti-inflammatory activities. This is supported by histological analysis at day 7 showing increased stromal proliferation and inflammatory cell recruitment in the joints of TAFI^-/-mice.

Conclusions:TAFI activation is impaired in congenital hemophilia but not in acquired hemophilia. Abrogation of TAFIa activity, either genetically or pharmaceutically, increased joint bleeding in mice with acquired hemophilia, indicating that TAFI may be responsible for the difference in joint bleeding tendency between acquired and congenital hemophilia. Protective effects of TAFI were vascular bed specific and independent of its anti-fibrinolytic effects, suggesting that one or more of TAFIa's other substrates promote hemophilic joint bleeding.