Protease-Activated Receptor 4 (PAR4) Plays a Critical Role in Fetal Loss and Peripheral Thrombosis in a Mouse Model of Thrombophilia

Abstract 373

We generated a mouse model of thrombophilia by combining two common risk factors for thrombosis in humans; transgenic for high levels of FIX (TgFIX⁺) and activated protein C resistance due to factor V Leiden mutation (FVL^Q). Breedings were used in which both parents were heterozygous for the FVL (FVL^R/Q) mutation, but only one carried the TgFIX⁺(resulting in ∼5-fold higher of normal levels). More than 200 newborns were obtained but surprisingly no animal of FVL^Q/Q/ TgFIX⁺ genotype was identified. When pregnancies were interrupted at embryonic age of E9.5 to E16.5, no deviation from the expected/observed embryo genotypes ratios was observed. However, the number of reabsorbed embryos increased significantly from 13% (E9.5) to 33% (E.16.5), p<0.02. All FVL^Q/Q/ TgFIX⁺ embryos exhibited findings consistent with disseminated intravascular coagulation. Placental analysis (n=18) of these embryos harvested at E12.5 or later demonstrated prothrombotic abnormalities such as excessive fibrin deposition and innumerous apoptotic cells, determined by TUNEL assay, localized mainly in the labyrinthine trophoblast zone. The maternal genotype clearly influenced the pregnancy outcome since the number of newborns per litter was 2-fold lower when the mother carried the TgFIX⁺versus the paternal (average 4.3 vs. 8.3, respectively; p<0.02). No other genotypes analyzed (n=112) presented similar abnormalities in either embryos or in placental tissue. Maternal anticoagulant therapy with heparin or specific thrombin or factor Xa inhibitors reverted the prothrombotic placental phenotype, although a few newborns reached to term, none survived beyond day 1. Although FVL^R/Q/ TgFIX⁺mice developed normally when reaching 8–10 months of age we observed a high rate of spontaneous peripheral thrombotic episodes (3/40; 7.5%) and histological analysis confirmed venous thrombosis. Thus, these mice represent a novel murine model of venous thrombosis. Moreover, this complication was also present in 5.2% of the FVL^R/R/ TgFIX⁺ whereas no other age-matched mice (n=74) developed such complications. There was an age-dependent increase in TAT and D-Dimer levels in both prothrombotic strains by comparing samples collected at 2 and 10 months time points. In the FVL^R/Q/ TgFIX⁺ or FVL^R/R/ TgFIX⁺, levels of TAT (70ng/ml, n=28/group) and D-dimers (230ng/ml, n=17/group) were ∼ 2-fold higher (P<0.04), when compared to mice (TAT: 30ng/ml and D-Dimer: 100ng/ml) of other genotypes. Similarly, the highest fibrin deposition determined by Western blot analysis of tissues harvested was found in the FVL^R/Q/ TgFIX⁺ followed by FVL^R/R/ TgFIX⁺ mice when compared to the other animals (P<0.05). Thus, this model provides a unique opportunity to test the effect of modulators of thrombosis in a chronic thrombophilia status rather than simply upon chemical- or laser-induced acute vascular injury. Since thrombin-mediated mouse platelet activation is PAR4 dependent in an acute vascular injury model, we sought to determine whether PAR4 could modulate the prothrombotic phenotype in our chronic thrombophilia model. Using triallelic crosses FVL^Q/Q/TgFIX⁺/PAR4^(−/−) and FVL^Q/Q/TgFIX⁺/PAR4^(+/−) animals were generated. Interestingly, we were able to rescue the fetal mortality of FVL^Q/Q/ TgFIX⁺by introducing just one PAR4 null allele. Notably, PAR4 deficiency in heterozygous or homozygous mice of FVL^Q/Q/ TgFIX⁺allowed them to reach term and develop normally. Kaplan-Meier analysis of both FVL^Q/Q/TgFIX⁺/PAR4^(−/−)and FVL^Q/Q/TgFIX⁺/PAR4^(+/−)mice showed survival rates of 87.5% and 65%, respectively, at the 10-month time point (ongoing observation). Notably, these rescued mice showed no evidence of peripheral thrombosis. In conclusion, we have characterized a mouse model of chronic hypercoagulability and have shown that in this model PAR4 plays a central role in modulating thrombophilic complications.