Factor XIIIa-dependent retention of red blood cells in clots is mediated by fibrin α-chain crosslinking

Abstract

Factor XIII(a) [FXIII(a)] stabilizes clots and increases resistance to fibrinolysis and mechanical disruption. FXIIIa also mediates red blood cell (RBC) retention in contracting clots and determines venous thrombus size, suggesting FXIII(a) is a potential target for reducing thrombosis. However, the mechanism by which FXIIIa retains RBCs in clots is unknown. We determined the effect of FXIII(a) on human and murine clot weight and composition. Real-time microscopy revealed extensive RBC loss from clots formed in the absence of FXIIIa activity, and RBCs exhibited transient deformation as they exited the clots. Fibrin band-shift assays and flow cytometry did not reveal crosslinking of fibrin or FXIIIa substrates to RBCs, suggesting FXIIIa does not crosslink RBCs directly to the clot. RBCs were retained in clots from mice deficient in α2-antiplasmin, thrombin-activatable fibrinolysis inhibitor, or fibronectin, indicating RBC retention does not depend on these FXIIIa substrates. RBC retention in clots was positively correlated with fibrin network density; however, FXIIIa inhibition reduced RBC retention at all network densities. FXIIIa inhibition reduced RBC retention in clots formed with fibrinogen that lacks γ-chain crosslinking sites, but not in clots that lack α-chain crosslinking sites. Moreover, FXIIIa inhibitor concentrations that primarily block α-, but not γ-, chain crosslinking decreased RBC retention in clots. These data indicate FXIIIa-dependent retention of RBCs in clots is mediated by fibrin α-chain crosslinking. These findings expose a newly recognized, essential role for fibrin crosslinking during whole blood clot formation and consolidation and establish FXIIIa activity as a key determinant of thrombus composition and size.

Figure 1

In the absence of FXIIIa activity, RBCs are extruded from the clot during clot contraction. (A-B) Clot formation and contraction were triggered by recalcification (10 mM, final) and addition of TF (1 pM, final) to whole blood spiked with octadecyl rhodamine B-labeled RBCs, Alexa Fluor 488-labeled fibrinogen, and ε-ACA (to inhibit fibrinolysis). Clot contraction was visualized on a Zeiss 710 NLO confocal laser scanning microscope with an incubation chamber and temperature stage (Carl Zeiss, Thornwood, NY), and monitored at 40× magnification. Representative frames of contracting clots formed in the (A) absence and (B) presence of T101. Times (in seconds) are indicated in each panel. Scale bar, 50 μm. (C) Clot formation and contraction was triggered by addition of thrombin (1 U/mL [10 nM], final) and recalcification (10 mM, final) to whole blood. Clot contraction was visualized at 10× with digital zoom on a Nikon Eclipse TE2000-U inverted microscope (Nikon Instruments, Melville, NY). Frames depict RBC extrusion from a contracting clot. Black arrowheads highlight a deforming RBC as it exits the clot. Numbers indicate successive frames. Scale bar, 10 μm.

Figure 2

FXIIIa does not crosslink fibrin to RBCs. Clotting was initiated in recalcified (5 mM, final) PRP (−RBCs) or whole blood (+RBCs) with thrombin (20 nM, final), in the absence or presence of T101 (200 μM, final). Clots were then dissolved and fibrin crosslinking patterns were analyzed by western blotting using a polyclonal anti-human fibrinogen antibody. Representative blot of n = 4 experiments.

Figure 3

FXIIIa does not crosslink FXIIIa substrates to RBCs. (A) Cy5-labeled A15 peptide (glutamine donor, 3 μM, final) or (B) biotinylated cadaverine (BC, lysine acceptor, 5 mM, final) was incubated with FXIII (20 μg/mL, final) and washed RBCs in the presence of thrombin (IIa, 5 nM, final) and CaCl₂ (10 mM, final) to probe for reactive lysine and glutamine residues, respectively, on the RBC surface. RBCs were then lysed, proteins were separated by SDS-PAGE, and labeling was visualized using Cy5 fluorescence (A15) or by transfer to a PVDF membrane and probing with Alexa Fluor 488-labeled streptavidin. Control reactions contained fibrinogen (0.5 mg/mL, final), FXIII, thrombin, CaCl₂, and A15 or BC. (C-D) For flow cytometry, intact cells were incubated with a phycoerythrin-labeled anti-human CD235a antibody and analyzed for (C) A15 and (D) BC labeling. Bars are means ± standard error (SE).

Figure 4

FXIIIa does not require α₂-antiplasmin, TAFI, or fibronectin to promote RBC retention in clots. Serum RBC content from ex vivo clot contraction assays using whole blood from (A) WT, α₂-antiplasmin–deficient (α₂-AP^−/−), and TAFI-deficient (TAFI^−/−) mice (n = 3) or (B) fibronectin-sufficient (pFn⁺) and -deficient (pFn⁻) mice (n = 6-9). Clotting was initiated with TF (1 pM, final) and recalcification (10 mM, final). WT and pFn⁺ blood was treated with T101 (5 μM, final) as positive controls. Bars are means ± SE.

Figure 5

Fibrin network density promotes RBC retention in clots, but FXIIIa mediates RBC retention independent of fibrin network density. (A-B) Recalcified (10 mM, final) plasma spiked with Alexa Fluor 647-labeled fibrinogen was clotted with the indicated TF concentrations in the (A) absence or (B) presence of T101. Images are representative confocal micrographs (z-projections of 30 individual slices) of clots visualized on a Zeiss LSM710 laser scanning confocal microscope with a 63× oil-immersion lens (Carl Zeiss). Scale bar, 30 μm. (C) Serum RBC content and (D) clot weight following clot contraction. Each dot represents an individual clot. Lines connect clots formed from the same blood donor. Horizontal dark lines indicate medians. (E-F) Representative scanning electron micrographs of clots formed in recalcified (10 mM, final) whole blood with the indicated TF concentrations in the (E) absence and (F) presence of T101 (10 μM, final). Clots were visualized at 5010× on a Zeiss Supra 25 Field Emission Scanning Electron Microscope (Carl Zeiss). Scale bar, 10 μm. Micrographs were used to measure fibrin network density (supplemental Methods), which was compared with the (G,I) serum RBC content or (H,J) clot weight following clot contraction in the (G-H) absence and (I-J) presence of T101.

Figure 6

FXIIIa inhibition does not reduce RBC retention in clots formed with Aα251 fibrinogen. Percent reduction in clot weight of contracted clots formed from TF-treated, recalcified fibrinogen-deficient plasma reconstituted with RBCs and platelets (2 million/μL and 200 000/μL, respectively), and γA/γA, γNNR, or Aα251 fibrinogen (0.25 mg/mL, final), in the absence or presence of 10 μM (final) T101 (n = 3-7 per fibrinogen). Bars are means ± SE.

Figure 7

RBC retention is reduced at concentrations of T101 that inhibit α-chain crosslinking. (A) Recalcified (10 mM, final) plasma was clotted with TF (1 pM) in the presence of increasing concentrations of T101. Clots were dissolved and analyzed by western blotting with polyclonal anti-human fibrinogen antibody. The second lane (Fgn) is unclotted plasma. (B) Normalized intensity of γ-γ dimer (open diamonds, short dashed line, n = 3) or HMW bands (open squares, long dashed line, n = 3) superimposed on normalized clot weight following clot contraction in the presence of T101 (closed circles, solid line, n = 5-7). Data are means ± SE.