Genetic Duplication of Tissue Factor Leads to Partial Specialization of Function in Zebrafish

Tissue factor (TF) is a critical factor for hemostasis in response to tissue injury. Among mouse knockouts of procoagulant factors, those lacking TF have the most severe phenotype, with complete lethality by midgestation. Furthermore, complete loss of TF has never been described in humans. Together, these suggest additional roles in embryonic development beyond coagulation. Zebrafish are a small freshwater teleost fish with a well described hemostatic system, including conservation of the coagulation cascade. Zebrafish are prolific breeders that reproduce through external fertilization, with subsequent rapid and transparent development, allowing studies not possible in mammals. Due to an ancient genomic event, 30-40% of the teleost genome is duplicated, resulting in two TF paralogs (TFa and TFb) with unknown functions. Here we use CRISPR/Cas9 to produce null alleles of TFa and TFb and uncover partial subspecialization of these duplicates. It has been shown previously that both TFs are expressed before the initiation of blood circulation, between 24 and 48 hours post fertilization, yet complete loss of TFa and TFb yielded no gross abnormalities. Embryos and larvae were able to develop normally through juvenile stages but succumbed to hemorrhage by early adulthood at 9 weeks of age. Surprisingly, a single allele of either TFa or TFb was able to rescue survival in the context of complete loss of the other gene. To evaluate for hemostatic effects of TF deficiency, laser-mediated endothelial injury was used in the venous and arterial systems at 3 and 5 days post fertilization (dpf), respectively. Loss of TFb alone at 3 dpf resulted in no observable hemostatic defects. Conversely, loss of TFa led to a 50% increase in the time to venous occlusion (TTO), which was exacerbated by concomitant loss of one allele of TFb. Total TF deficiency led to a complete inability to form occlusive venous thrombi, indicating that both TFs can trigger coagulation but TFa is able to completely compensate for the loss of TFb. Concordant with these data, loss of TFb resulted in transcriptional upregulation of TFa but not vice versa. Interestingly, the roles are reversed in the arterial vasculature. Loss of TFa had no effect, loss of TFb lead to a 60% reduction in the number of occlusive thrombi, and complete deficiency resulted in no arterial occlusion. Combined with the venous results, these data point to differentiated roles of TFa and TFb in the venous and arterial systems. In order to test whether these differences were functional, recombinant TF (rTF) molecules were expressed in E. coli, purified, and incorporated into 80% phosphatidylcholine/20% phosphatidylserine liposomes. Ex vivo tube-tilt clotting assays were performed by using each rTF to activate citrated plasma from lake trout. rTFa triggered stable clot formation within 1-2 minutes of recalcification. rTFb usually failed to induce clot formation, with occasional delayed fibrin thrombi that appeared to be grossly disorganized and were easily disrupted following agitation. Taken with the in vivo data, this hints at an altered kinetic profile, with TFa being a more potent cofactor for factor VIIa in low flow (venous) settings. The laboratory is an artificially safe environment, so a synthetic chemical stress test was performed on 3 dpf larvae. Prolonged treatment with cortisol and epinephrine led to the development of cardiac tamponade in larvae with complete TF deficiency (61%), but similar results were only found at low levels in wild type siblings (2-5%). The same assay in prothrombin mutants also revealed a high rate of tamponade (75%), but lower levels in fibrinogen-deficient larvae (20%). These data suggest an extrahemostatic risk factor for tamponade that is modified by prothrombin and tissue factor levels and is independent of fibrin formation. Our results intimate that TFa and TFb have overlapping procoagulant functions but differential kinetic profiles in venous vs arterial systems. We also find that the duplication provides a layer of quantitative regulation and creates a titratable level for regulation of hemostatic and extrahemostatic roles of TF. Overall, this novel model provides new structural and physiologic information about TF function in vivo, including potential previously unknown roles in perivascular development, cardiovascular stability, remodeling and/or regeneration.