Tissue Factor Pathway Inhibitor Gamma Deletion Corrects the Bleeding Phenotype in Hemophilia Mice

Tissue factor (TF) is a transmembrane receptor that initiates coagulation and, via protease-activated receptors, enhances inflammation. Tissue factor pathway inhibitor (TFPI) is the endogenous regulator of TF activities. Anti-TFPI antibodies show promise as a hemophilia treatment but must be tittered to avoid potential thrombogenic side effects. From a single gene, through alternative splicing, humans express TFPI transcripts that encode for α and β isoforms; while mice express transcripts that encode for α and β, and also a third isoform γ, which is the most abundant. It is hypothesized that TFPIα is the dominant isoform anti-coagulant in man, whereas cell-anchored TFPIβ is the dominant anti-inflammatory. The main goal of our study is to further understand the impact of the different TFPI isoforms in hemophilic mice.

To assess the functions of the α, β, and γ mouse-specific TFPI isoforms, using CRISPR/Cas9 technology, we generated TFPI α-, β-, γ-, α/β-, α/γ-, and β/γ- deleted mouse strains. Each knock-out strain is viable, and the offspring were born with expected Mendelian ratios. To evaluate the hemostatic function of the individual isoforms, all TFPI mice strains were bred on a hemophilia (FVIII-null) background and evaluated with a tail-vein rebleeding assay (TVRB). In this TVRB, a lateral tail vein is transected to cause bleeding; time to clot formation is recorded, then the clot is immediately dislodged, and the process is repeated over 15 minutes. In this assay, C57/Bl6 mice clot on average 26 +1.3 (mean+SD) times, whereas F8 null mice clot 5 + 2.6 times on average. Surprisingly, in the F8 null background, deletion of TFPIγ (14 + 6.5), alone or in combination as α/γ-deleted or β/γ-deleted, significantly reduced bleeding (p < 0.001), whereas TFPIα deletion had no impact on hemophilia bleeding (5 + 3).

We next performed ex-vivo thrombin generation assays (TGA). The blood was collected via the inferior vena cava into corn trypsin inhibitor plus citrate syringes. Platelet-poor plasma was extracted and added into 96-well plates in addition to TF, phospholipids, calcium, and substrate. Endogenous thrombin potential was calculated from the area under the curve for generated thrombin for each TFPI mice strain on a hemophilia background. In agreement with the in vivo model, TFPIγ-deletion on the F8 null background significantly improved endogenous thrombin potential (P=0.037), while additional deletion of TFPIα (α/γ-deletion) did not provide any additional procoagulant effect.

In conclusion, both in vivo and ex vivo hemostasis mouse models showed that, rather than three Kunitz TFPIα or cell-anchored TFPIβ, soluble two-Kunitz domain TFPIγ is the dominant mouse plasma isoform with respect to anti-coagulant function. Studies using mouse models that investigate the role of TFPI in hemostasis, thrombosis, and inflammation must carefully consider and address the impact of mouse TFPIγ on outcomes before extrapolating findings to human TFPI in health and disease.

No relevant conflicts of interest to declare.