BACKGROUND. Tissue factor (TF) is critical for hemostasis. However, TF also plays a pathogenic role in a multitude of diseases, including: thrombosis, coagulation-associated inflammation, multi-organ failure associated with infections, tumor growth, angiogenesis and metastasis. Tissue Factor Pathway Inhibitor (TFPI), the regulator of TF actions, produces FXa-dependent feedback inhibition of the FVIIa/TF catalytic complex, directly inhibits FXa, and suppresses cellular signaling by FVIIa/TF/FXa complexes. In man, TFPI is expressed as isoforms TFPIα and TFPIβ. The pathophysiologic impacts of the individual α and β isoforms have yet to be elucidated.
Mouse models have been used extensively to provide insight into human physiology and disease, including abnormalities concerning hemostasis and TF-associated diseases. Like man, mice express TFPI α and β, but unlike man, mice express a third transcript encoding TFPIγ. In fact, we recently reported that the vast majority (98%) of TFPI in mouse plasma is derived from TFPIγ. We also reported that anti-TFPI antibodies enhance in vitro thrombin generation in hemophilia mouse plasma that contains only the TFPIγ isoform, which demonstrated that TFPIγ possesses anticoagulant activity.
RESULTS. We report herein that, while complete TFPI gene deletion is 100% embryonic lethal, TFPIγ only mice [TFPIαβ(-/-)] are viable and reproduce. Offspring genotypes from parents with intact TFPIγ(+/+) exons and TFPIαβ(+/-) were 33 TFPIαβ(+/+), 57 TFPIαβ(+/-) and 34 TFPIαβ(-/-) at weaning. Also, TFPIγ only mice produced similar numbers of pups per litter as wild type mice. The viability of TFPIγ only mice proves endogenous mouse TFPIγ possesses substantial functional activity. TFPIα is considered to be a much more potent anticoagulant than TFPIγ. Our preliminary thrombin generation assay data suggests that the sum total anticoagulant activity of the endogenous low level (0.07 nM) of TFPIα is similar to the sum total anticoagulant activity of the endogenous high level (3.5 nM) of TFPIγ.
The presence of active TFPIγ in mice, not found in man, confounds attempts to translate results from wild type mouse studies to the human condition. The extents to which TFPIγ influences the various wild type mouse models of human diseases are unknown. The presence of mouse TFPIγ also precludes our ability to properly distinguish the roles of TFPI α and β isoforms independently. Deletion of the TFPIγ exon eliminates this confounder. Fortunately, in TFPIαβ mice the levels, properties and location of TFPIα and TFPIβ in mice reflect those in man. Importantly, we can now report that our CRISPR-Cas9-generated TFPIγ KO strains - TFPIαβ mice, TFPIα only mice, and TFPIβ only mice are viable and reproduce.
To illustrate the potential influence of TFPIγ in a mouse model, we used a tail vein-rebleeding assay to assess bleeding in TFPI isoform-specific mice in a FVIII KO background. Reported as mean number of clots per 15 minutes, results were: 4.8 for hemophilia controls; 25.4 for controls with FVIII replacement; 4.2 for TFPIα KOs; and 13.5 for TFPIβγ KOs. (The S.E.M. of each group was <1.0 ). We find that deletion of TFPIα (including in platelets!) does not affect bleeding (P>0.5 vs controls), whereas deletion of TFPIβ and γ improves hemostasis ( P<0.001 vs controls). In hemophilia mice, suppressing TFPIα was without effect. We don't expect this observation to translate to man. Rather, suppressing TFPIα in man is anticipated to enhance coagulation in hemophilia patients. (Note, multiple trials show that TFPI antibodies that inhibit all TFPI forms enhance coagulation). Moreover, in the absence of the confounding influence of TFPIγ, we anticipate suppression of TFPIα will improve coagulation and reduce bleeding in TFPIγ deleted hemophilia mice. Further studies in our TFPIγ deleted strains are necessary to determine if this prediction is correct and could influence future anti-TFPI approaches for man.
CONCLUSIONS. 1.) Endogenous mouse TFPIγ possesses significant functional activity. 2.) TFPIγ confounds the extrapolation of results in wild type mouse models to human diseases. 3.) The TFPIγ KO mouse possesses a TFPIαβ profile that is similar to man. 4.) For essentially all studies of human TF-associated diseases in mouse models, it would be prudent to delete TFPIγ.
Disclosures
No relevant conflicts of interest to declare.
S2′-subsite variations between human and mouse enzymes (plasmin, factor XIa, kallikrein) elucidate inhibition differences by tissue factor pathway inhibitor -2 domain1-wild-type, Leu17Arg-mutant and aprotinin