Identification of a Spontaneous Noncoding Mutation on Chromosome 18 Which Suppresses Thrombosis in Factor V Leiden Homozygous, TFPI Deficient Mice

Factor V Leiden (FVL) is an incompletely penetrant thrombosis susceptibility variant common in humans. This incomplete penetrance suggests there are modifiers of FVL (F5^L) that alter the susceptibility to a thrombotic event. To identify modifiers, we performed a sensitized ethylnitrosourea (ENU) mutagenesis screen in mice to induce mutations suppressing the F5^L/LTfpi^+/- perinatal lethal thrombotic phenotype. Although we identified thrombosuppressor mutations in two of our 22 independent MF5L mouse lines thus far, the others are still unknown. Our mouse line MF5L16 has produced a large, highly penetrant (77.2%), multigenerational pedigree containing 136 viable F5^L/LTfpi^+/- mice. We performed the Prothrombin Time (PT) assay on plasma from a subset of these mice and observed a significantly longer PT in MF5L16F5^L/LTfpi^+/- mice compared to F5^+/LTfpi^+/- and Tfpi^+/- control mice (q<0.01). The aim of our study was to identify the MF5L16 thrombosuppressor. First, we analyzed four mice by whole genome sequencing and identified a total of 44 candidate ENU mutations. Sanger re-sequencing analysis determined that 22 were false-positive calls. Seven arose in our F5^L/L breeding colony and were introduced into MF5L16 by these mice. Three arose on the Tfpi^- mouse background. The remaining 12 candidates could not be analyzed as they were in repetitive/low complexity regions. Due to the lack of ENU-induced mutations segregating in MF5L16, we next analyzed the seven mutations entering the pedigree from the F5^L/L breeders in all 136 MF5L16 F5^L/LTfpi^+/- mice. We identified a significant association between a Chromosome 18 intergenic variant (Chr18 G to A, Chr18^A) and F5^L/LTfpi^+/- mouse survival (p=0.0031). Two additional mutations within intronic regions on Chromosomes 9 and 13 were also significantly associated with survival when co-inherited with the Chromosome 18 mutation (p=0.000113). Importantly, no coding variants were linked to these variants. To re-create the suppression of the lethal F5^L/LTfpi^+/- phenotype, we bred a F5^L/L Chr18^A/A mouse onto the Tfpi^- background to produce F5^+/L Chr18^+/ATfpi^+/- mice. We then bred these F5^+/L Chr18^+/ATfpi^+/- triple heterozygous mice to F5^L/L Chr18^A/ATfpi^+/+ mice to observe the effects of Chr18^A on F5^L/LTfpi^+/- mouse survival. The first litter of six mice from this cross yielded one F5^L/LTfpi^+/- mouse that was also heterozygous for Chr18 (F5^L/LTfpi^+/- Chr18^+/A: expected ratio ~1:8). This suggests that the Chr18^A variant suppresses F5^L/LTfpi^+/- lethal thrombosis and that noncoding genomic variants can act as potent thrombosuppressors. Our genetic analyses suggest that thrombosuppressor variants arose in our F5^L mouse colony due to positive selective pressure for the best F5^L breeders. In addition, the genomic interactions between the putative thrombosuppressor variants described here illuminate the complexities of thrombosis genomics and could provide insights into human thrombosis.