Multiple Thrombosis Modifier Loci Identified Using a Sensitized Mutagenesis Screen

Thrombosis is a leading cause of mortality in the developed world. Of the natural anticoagulants, congenital deficiencies of antithrombin III (AT3) lead to the highest risk of venous thromboembolism. However, the lifetime risk of thrombosis for these patients is only ~30%, presumably due to unknown modifier genes. Danio rerio (zebrafish) demonstrate a high degree of conservation of the coagulation cascade. Combined with their high fecundity, optically transparent embryos and well-characterized genetics, they offer an ideal model for low-cost high-throughput in vivo screens. Previously, we generated and described a knockout of at3 in zebrafish using genome editing. Homozygous mutant adults exhibit a decreased lifespan due to massive intracardiac thrombosis by 6 months of age, yet embryos survive spontaneous thrombosis with a consumptive coagulopathy characteristic of disseminated intravascular coagulation (DIC). This embryonic phenotype results in fibrinogen depletion and they are unable to form occlusive thrombi following laser-mediated venous endothelial injury.

Here we use our at3^-/- mutant line as a sensitized hypercoagulable background to identify novel modifiers of thrombosis. We performed a genome-wide N-ethyl-N-nitrosourea (ENU) mutagenesis screen to discover dominant suppressor mutations that lead to increased survival. ENU treated at3^+/- males were crossed with at3^+/- females and offspring genotyped at 2-4 months of age. 53 at3^-/- mutants survived beyond 7 months, 23 produced offspring, and 5 established permanent lines. One mutant, frost, also demonstrated reversal of embryonic DIC.

Genomic DNAs from three frost mutants were evaluated by whole exome sequencing. An initial search of 54 known platelet and coagulation factor loci revealed a candidate missense mutation, C559F, in a highly conserved residue of the prothrombin (F2) heavy chain. The predicted deleterious mutation was not present in wild type stocks, zebrafish variant databases or the original mutagenized male. By 3 months of age, at3^-/- juveniles heterozygous for the C559F allele demonstrated significantly improved survival compared to wild type siblings (p<0.0001). At 2 years of age, all at3^-/- survivors were heterozygous for C559F (n=22). In order to definitively prove the f2 locus as the underlying cause of frost, we used TALEN genome editing to produce a separate knockout allele f2^Δ, and bred that into the non-mutagenized at3^-/- background. We found a significant increase in f2^+/Δ;at3^-/- survivors compared to f2^+/+;at3^-/- and f2^Δ/Δ;at3^-/- (p=0.006), demonstrating that partial knockdown of thrombin activity rescues at3^-/- adult lethality, and confirming f2 as the locus underlying the frost mutation. Among f2^+/+;at3^-/- embryos 82% exhibit DIC, but this decreases to 51% in f2^+/C559F;at3^-/-. However, f2^Δ does not have the same effect in the non-mutagenized at3^-/- background with 85% of f2^+/Δ;at3^-/- embryos demonstrating DIC. This suggests that thrombin deficiency is necessary to unmask a phenotypic rescue due to a putative second modifier in the frost mutant background.

Taken together, these data provide proof of principle for the validity of this screen and evidence that we have the power to identify novel dominant modifiers of prothrombin expression or thrombin function, including molecules outside of the canonical coagulation cascade. Furthermore, our data provide evidence for a second unknown locus modifying the embryonic DIC phenotype and acting epistatic to thrombin activity. Detection and characterization of this underlying gene may provide insight into the variable penetrance of thrombosis in patients with thrombophilic risk factors and offer new therapeutic options for treatment of DIC.