Defective TAFI Activation in Hemophilia Exacerbates Vascular Remodeling in Hemophilic Arthropathy

Introduction: Patients with hemophilia A or B develop hemophilic arthropathy (HA), characterized by inflammation, cell proliferation, and hypervascularity in the synovial joint tissue. The hypervascular response in HA results in aberrant vascular remodeling and the formation of enlarged, irregularly shaped and leaky vessels, which are prone to bleeding. Recently, we showed that after a single joint bleed in FVIII^-/- mice changes in soft tissue proliferation, neo-angiogenesis, and cartilage destruction were reversible, whereas vascular remodeling in the joint became progressively worse over time (Bhat et al. 2015). Activated thrombin activatable fibrinolysis inhibitor (TAFIa) conveys anti-fibrinolytic and anti-inflammatory functions by removal of C-terminal (CT) Lys or Arg residues from fibrin and pro-inflammatory mediators, such as C3a, C5a, bradykinin, and thrombin-cleaved osteopontin that may be relevant to HA. Because the activation of TAFI is defective in hemophilia, we determined whether TAFI deficiency contributes to aberrant vascular remodeling in HA using a mouse knee injury model.

Methods and Results: Joint bleeding was induced by a subpatellar needle puncture in WT and TAFI^-/- mice after a single injection of a FVIII inhibiting antibody to evoke sufficient joint bleeding. Inhibitor-treated TAFI^-/- mice bled to the same extent as FVIII^-/- mice, however inhibitor-treated WT mice bled minimally. Combining both inhibitory anti-TAFI and anti-FVIII antibodies resulted in bleeding in WT mice comparable to that in FVIII^-/- and inhibitor-treated TAFI^-/-mice as indicated by decreased hematocrits 2 days post injury (27±9%, 28±10%, 24±7%) and a 4-fold impaired weight-bearing of the injured hind leg. The inhibitors were cleared after 4 days permitting analyses of TAFI-specific effects on long-term vascular remodeling after joint bleeding.

At week 2 post injury, all mice showed evidence of increased vascularity in the injured joint. Power Doppler Ultrasound (PD) signals, indicating flow in the joint, were increased 3.5-fold in FVIII^-/-, TAFI^-/- and WT mice. Accordingly, Safranin O-Fast Green analysis indicated induction of neoangiogenesis as mean vessel count increased from 40 at baseline to 101, 119 and 71 in FVIII^-/-, TAFI^-/- and WT mice, respectively. At week 4 post injury, however, WT mice diverged from FVIII^-/- and TAFI^-/- mice. PD signals of FVIII^-/- and TAFI^-/- mice remained high (3.5-fold increase over baseline, p< 0.001), whereas those of WT mice were significantly reduced compared to week 2 (p< 0.05). Persistent vascular remodeling occurred in FVIII^-/- and TAFI^-/- mice, but not in WT mice. While total vessel counts decreased in all mice, the percentage of vessels ≥ 20 µm, considered the upper limit of normal vessel diameter in the joint, was increased 2-fold in FVIII^-/- and TAFI^-/- mice (p< 0.001) but was similar to baseline (11%) in WT mice. Similarly, a 2-fold increase of αSMA expression in the medial meniscal area in TAFI^-/-mice support ongoing vascular remodeling in these mice (p< 0.01). Thus, the impaired TAFI function in HA promotes vascular remodeling.

In search of proangiogenic mediators implicated in the pathogenesis of HA and susceptible to inactivation by TAFIa, we identified stromal cell-derived factor 1α (SDF1α; CXCL12), a well-known endothelial progenitor recruitment factor. SDF1α contains a functionally requisite CT Lys that is needed for CXCR4 activation. Histological analysis using a CT-Lys-specific anti-SDF1α revealed pronounced expression of intact SDF1α on the joint vessels in TAFI^-/- mice at week 4 post injury. In vitro,TAFIa rapidly cleaved the CT-Lys of SDF1α as determined by a CT-Lys specific W-blot and ELISA. In addition, TAFIa-treated SDF1α failed to induce migration of endothelial cells in a transwell cell migration assay.

Conclusion: Vascular remodeling after joint bleeding was perpetuated in TAFI^-/- mice to the same extent as in FVIII^-/- mice. Based on the identification of SDF1α as a novel substrate for TAFIa, we propose that the defective activation of TAFI in hemophilia results in impaired inactivation of SDF1α, allowing functional SDF1α to accumulate on synovial vessels in the injured joint thereby exacerbating the recruitment of endothelial progenitor cells to the joint. This in combination with the ongoing angiogenesis in the joint ultimately results in irreversible vascular remodeling that underlies the development of HA.