Endothelial-derived tissue factor pathway inhibitor regulates arterial thrombosis but is not required for development or hemostasis

Abstract

The antithrombotic surface of endothelium is regulated in a coordinated manner. Tissue factor pathway inhibitor (TFPI) localized at the endothelial cell surface regulates the production of FXa by inhibiting the TF/VIIa complex. Systemic homozygotic deletion of the first Kunitz (K1) domain of TFPI results in intrauterine lethality in mice. Here we define the cellular sources of TFPI and their role in development, hemostasis, and thrombosis using TFPI conditional knockout mice. We used a Cre-lox strategy and generated mice with a floxed exon 4 (TFPI(Flox)) which encodes for the TFPI-K1 domain. Mice bred into Tie2-Cre and LysM-Cre lines to delete TFPI-K1 in endothelial (TFPI(Tie2)) and myelomonocytic (TFPI(LysM)) cells resulted in viable and fertile offspring. Plasma TFPI activity was reduced in the TFPI(Tie2) (71% ± 0.9%, P < .001) and TFPI(LysM) (19% ± 0.6%, P < .001) compared with TFPI(Flox) littermate controls. Tail and cuticle bleeding were unaffected. However, TFPI(Tie2) mice but not TFPI(LysM) mice had increased ferric chloride-induced arterial thrombosis. Taken together, the data reveal distinct roles for endothelial- and myelomonocytic-derived TFPI.

Figure 1

TFPI targeting vector with 5′ DNA probe Southern blot screening. (A) Schematic representation of the introduction of the targeting vector and subsequent removal of the PGK-NEO cassette using flippase-mediated flippase recombination target site. (B) 129Sv/J embryonic stem cells were screened for successful targeting of TFPI exon 4 using Southern blot analysis that used a 5′ DNA probe differentiating wild-type and floxed alleles. (C) Germline transmission of the targeting vector was also assessed using Southern blot screening.

Figure 2

Transcriptional efficiency of Cre-mediated deletion. (A) Tie2-directed deletion of TFPI-K1 had high-efficiency deletion of TFPI-K1 in vascular endothelial cells, bone marrow cells, and circulating blood cells. (B) LysM-directed deletion resulted in partial deletion of TFPI-K1 in cells from whole blood and bone marrow.

Figure 3

Plasma TFPI activities. (A) TFPI activities were measured in plasma from mice that contain a floxed TFPI-K1 domain and either do not express Cre recombinase (TFPI^Flox) or express the Cre recombinase via the endothelial specific promoter for Tie2 (TFPI^Tie2) or the myelomonocytic specific promoter for LysM (TFPI^LysM). Animals that express Cre recombinase have a TFPI K1 domain deletion, which is the domain necessary for TFPI inhibition of the TF coagulation pathway. TFPI^Tie2 mice have significantly reduced circulating TFPI activity levels compared with TFPI^Flox (*P < .001) and TFPI^LysM mice (†P < .001). The TFPI^LysM mice have significantly lower circulating TFPI activities than TFPI^Flox littermate controls (*P < .001; panel A). (B) Heparin infusion fails to increase plasma TFPI activity in TFPI^Tie2 mice.

Figure 4

Murine model of thrombosis in TFPI^LysM mice. Thrombosis was examined using a FeCl₃ arterial injury model in TFPI^Flox and TFPI^LysM mice. A flow probe was used to determine when the carotid arteries of TFPI^Flox and TFPI^LysM mice became occluded. (A) No significant differences were observed in times to occlusion between TFPI^Flox and TFPI^LysM mice. (B) Percent patency was plotted against time using a Kaplan-Meier survival plot and χ² analysis revealed no significant differences of plots for TFPI^Flox and TFPI^LysM mice (P = .25).

Figure 5

Murine model of thrombosis in TFPI^Tie2 mice. Thrombosis was examined using a FeCl₃ arterial injury model in TFPI^Flox and TFPI^Tie2 mice. A flow probe was used to determine when the carotid arteries of TFPI^Flox and TFPI^Tie2 mice became occluded. (A) Times to occlusion were significantly reduced in TFPI^Tie2 mice compared with TFPI^Flox mice (*P < .001). (B) Percentage of patency was plotted against time using a Kaplan-Meier survival plot, and χ² analysis revealed a significant difference between plots for TFPI^Flox and TFPI^LysM mice (P = .007).