Fibrin Crosslinking Is Required For Retention Of Red Blood Cells In Venous Thrombi,

Venous thrombi contain high levels of red blood cells (RBC) and fibrin, but little is known about the mechanisms regulating venous thrombus formation and composition. Fibrin acts as a scaffold for leukocytes and platelets that mediate thrombus formation, and cross-linked fibrin networks promote clot stability via their extraordinary extensibility and elasticity. Previous studies have shown that a region of the fibrinogen γ-chain (residues γ390-396) is critical for engagement of the leukocyte integrin receptor αMβ2. Mice expressing a mutant form of fibrinogen with residues 390-396 mutated to a series of alanines (termed Fibγ^390-396A) exhibit normal fibrin polymerization and normal hemostasis. However, Fibγ^390-396A mice exhibit blunted inflammatory responses and protection from the development of numerous inflammatory diseases (e.g., arthritis, neuroinflammatory disease, colitis). In this study, we analyzed the role of this region of fibrinogen in a stasis-induced model of venous thrombosis. Surprisingly, following inferior vena cava ligation – a model that produces thrombi independent of leukocyte tissue factor activity – Fibγ^390-396A mice had 50% smaller thrombi than wild type (WT) mice (9.7±1.3 vs 19.5±1.7 mg, P<0.0003). Reduced thrombus weight was not due to reduced thrombin generation (i.e., TAT levels), and total neutrophil, platelet, and fibrin content within thrombi were similar between groups. Strikingly, Fibγ^390-396A thrombi had significantly fewer RBCs than WT thrombi (16.0±4.1 vs 52.9±5.8 arbitrary units [AU], P<0.0001), and thrombus RBC content correlated positively (R=0.90) and significantly (P<0.001) with overall thrombus weight. To determine the mechanism of decreased RBC presence in thrombi from Fibγ^390-396A mice, we developed an ex vivo whole blood clot retraction assay. Interestingly, although retraction of platelet-rich plasma clots was indistinguishable for Fibγ^390-396A and WT mice (91±1 vs 92±1 %), retraction of whole blood clots resulted in dramatically reduced RBC retention (37.0±8.4 vs 79.0±8.0 % of initial RBCs, P<0.03) and smaller clots (11.6±1.6 vs 53.8±3.6 mg, P<0.003) for Fibγ^390-396A mice compared to WT. Reconstitution experiments showed the mechanism of decreased RBC retention was not due to abnormal RBC function. Microfluidic-based adhesion analyses indicated RBCs adhered similarly to both Fibγ^390-396A and WT purified fibrinogen (65.7±6.4 vs 76.7±14.8 % cell adhesion, respectively, P=0.55), indicating that RBC extrusion did not result from decreased RBC binding to Fibγ^390-396A clots. To test the hypothesis that the Fibγ^390-396A mutation disrupts a specific interaction with the fibrin-stabilizing transglutaminase, factor XIII (FXIII), we analyzed levels of FXIII that co-precipitated with WT and Fibγ^390-396A fibrinogen. Interestingly, despite normal circulating levels of FXIII in Fibγ^390-396A mice, FXIII co-precipitated with WT fibrinogen, but not with Fibγ^390-396A fibrinogen. Compared to WT, plasma clots from Fibγ^390-396A mice exhibited slower FXIII activation (58.3±19.2 vs 12.2±2.5 AU/min [x10^-3], P<0.05) and consequently, slower fibrin crosslinking (γ-γ dimers: 114±14.2 vs 19.5±2.5 AU/min [x10^-3], P<0.0001; α polymers: 278±52 vs 46.1±11.4 AU/min [x10^-3], P<0.002) and reduced elastic modulus (14.1±0.2 vs 7.8±0.6 G’ [Kd/sec], P<0.0005). Provocatively, whole blood from FXIII-deficient mice and humans phenocopied Fibγ^390-396A clots, with reduced RBC retention following clot retraction. Taken together, these studies suggest a critical, yet previously un-described, role for FXIII in mediating RBC retention within clots. Further, these data identify critical residues in fibrinogen that mediate FXIII activation and fibrin crosslinking, and reveal that FXIII-mediated fibrin crosslinking is required for the retention of RBCs in venous thrombi.