In characterizing mice with targeted disruption of the SerpinB2 gene, we observed animals that were small at birth with delayed growth and decreased life expectancy. Although this phenotype cosegregated with homozygosity for the inactive SerpinB2 allele, analysis of homozygous SerpinB2-deficient mice derived from two additional independent embryonic stem (ES) cell clones exhibited no growth abnormalities. Examination of additional progeny from the original SerpinB2-deficient line revealed recombination between the small phenotype (smla) and the SerpinB2 locus. The locus responsible for smla was mapped to a 2.78-Mb interval approximately 30 Mb proximal to SerpinB2, bounded by markers D1Mit382 and D1Mit216. Sequencing of Irs1 identified a nonsense mutation at serine 57 (S57X), resulting in complete loss of IRS1 protein expression. Analysis of ES cell DNA suggests that the S57X Irs1 mutation arose spontaneously in an ES cell subclone during cell culture. Although the smla phenotype is similar to previously reported Irs1 alleles, mice exhibited decreased survival, in contrast to the enhanced longevity reported for IRS1 deficiency generated by gene targeting. This discrepancy could result from differences in strain background, unintended indirect effects of the gene targeting, or the minimal genetic interference of the S57X mutation compared with the conventionally targeted Irs1-KO allele. Spontaneous mutations arising during ES cell culture may be a frequent but underappreciated occurrence. When linked to a targeted allele, such mutations could lead to incorrect assignment of phenotype and may account for a subset of markedly discordant results from experiments independently targeting the same gene.insulin receptor substrate protein | mutant strain | plasminogen activator inhibitor 2 | knockout mice | embryonic stem cell mutation

Factor V Leiden, (FVL) is the most common known inherited thrombotic risk factor and is present in approximately 5% of most Western populations and 25–50% of patients presenting with venous thrombosis. However, FVL is incompletely penetrant, with only approximately 10% of FVL carriers developing thrombosis in their lifetimes. Though interactions between FVL and other known prothrombotic mutations have been documented in a few cases, the genetic factors responsible for the incomplete penetrance of FVL remain largely unknown. We previously reported a remarkable synthetic lethality in mice carrying the FVL mutation and partial deficiency of a key coagulation component, tissue factor pathway inhibitor (TFPI). Complete TFPI deficiency in mice is embryonic lethal, whereas heterozygosity is compatible with normal survival. However, homozygosity for FVL (FvQ/Q) in the context of heterozygosity for TFPI (Tfpi+/−) is uniformly lethal due to disseminated perinatal thrombosis. In order to identify potential modifier genes contributing to FVL penetrance, we have utilized this lethal genetic interaction as a phenotyping tool for a sensitized ENU mutagenesis screen in laboratory mice. We hypothesize that dominant mutations in key components of the coagulation system will improve hemostatic balance and allow survival in mice carrying the lethal FvQ/Q Tfpi+/− genotype combination. As an example, we propose that loss of one tissue factor allele might compensate for reduced TFPI and rescue FvQ/Q Tfpi+/− . To test this hypothesis, we bred tissue factor heterozygous mice (Tf+/−) with FvQ/Q Tfpi+/− mice and observed complete rescue, with normal survival and the expected number (8 of 57) of FvQ/Q Tfpi+/− Tf+/− mice from a FvQ/+ Tfpi+/− Tf+/−x FvQ/Q cross. In order to identify candidate modifier genes, we performed a whole genome mutagenesis screen. In this screen, male FvQ/Q mice were mutagenized with ENU and bred to FvQ/+ Tfpi+/− double heterozygous females. DNAs from surviving offspring were PCR assayed to identify rescued mice with the FvQ/Q Tfpi+/− genotype. Analysis of 2250 offspring, corresponding to approximately half genome coverage, has identified 15 mice that survived to weaning. Heritability was demonstrated for the 5 mutant lines subjected to progeny testing to date. Genetic crosses are in progress to map the mutant genes in 3 of the 5 progeny tested lines. These preliminary results demonstrate the feasibility of this sensitized screen for the identification of dominant suppressors of thrombosis. Based on our data, we estimate that there are likely 10–20 mammalian genes for which a <50% reduction in expression could result in a major shift in hemostatic balance sufficient to rescue the lethal thrombosis associated with the FvQ/Q Tfpi+/− lethal genotype. Each of these loci represent a candidate for a human modifier gene in patients with FVL and other thrombophilic mutations. Finally, the biologic pathways uncovered by these studies should provide new insights into the overall regulation of hemostatic balance and identify potential new targets for therapeutic intervention.