Surprisingly Rapid Onset of Estrogen-Induced Venous Thrombosis in an Animal Model

Venous thrombosis is a well-known complication of sex hormone therapy, with onset typically within weeks to months after initiation. Worldwide, more than 100 million pre-menopausal women use combined oral contraceptives, thus tens to hundreds of thousands develop thrombosis annually, resulting in significant morbidity and mortality. Estrogens are thought to be the major thrombotic risk, although the impact of progestins is less clear. It has been hypothesized that progestin effects may be secondary to mitigation of the estrogen effect. Although it is known that estrogens and progestins can alter expression of coagulation factors, the pathways and mechanisms that connect the two systems, as well as the proteins involved in progression to thrombosis, are poorly understood. Identification of these mediators are central to any comprehensive understanding of hormone-induced pathophysiology, could help ascertain patients at higher risk for thrombosis, and might also pinpoint future therapeutic targets. The zebrafish is a powerful genetic model in which the hemostatic system is nearly entirely conserved with humans. Its external development, ability to generate thousands of offspring at low cost, and optical transparency all make it a powerful tool to study the genetics of coagulation disorders. We previously produced a transgenic line (fabp-fgb-egfp) that generates GFP tagged fibrinogen beta (Fgb), which is incorporated into induced and spontaneous fibrin-rich thrombi. Here we show surprisingly rapid onset of estrogen-induced thrombosis. We exposed 5-6 day old transgenic zebrafish larvae to a range of concentrations of estradiol over 4-24 hours. We discovered green fluorescent deposits co-localizing to the posterior cardinal vein (PCV, orthologous to the inferior vena cava), but not other vessels. This occurred within 4 hours in a dose-dependent fashion. Confocal imaging demonstrated that the fluorescence was intravascular and deposited along the endothelium. Pre-incubation with warfarin and the direct oral anticoagulants rivaroxaban and dabigatran inhibited the fluorescent signal, confirming that the deposits were fibrin thrombi. This phenotype was similar to spontaneous PCV thrombosis observed in zebrafish mutants deficient in antithrombin, which also demonstrate a consumptive coagulopathy with absence of venous occlusion in response to endothelial injury. Therefore we treated 3 dpf larvae with estradiol for 4 hours, followed by laser-mediated endothelial injury and measurement of the time to occlusion (TTO) in the PCV by a blinded observer. Although the median TTO only increased from 12.5 to 15 seconds with treatment, this was highly statistically significant (p=0.0001, Mann Whitney U test, n = 81-98 for each group), consistent with a consumptive coagulopathy. To evaluate the role of progestins, 6 dpf larvae were treated with estradiol and levonorgestrel, desogestrel, or norethisterone, alone or in combination, and quantitatively scored for the presence of thrombosis by a blinded observer. The rates of thrombosis were 72% for estradiol, ranged 15-31% for the progestins, and ranged 66-70% for the combined estrogen/progestin treatments. In summary, our results are nearly identical to loss of antithrombin in zebrafish, although the consumptive coagulopathy is not as potent. These data are consistent with human estrogen-induced venous thromboembolism, but it was completely unexpected and concerning to see the production of thrombi in such a rapid fashion. Our data also suggest that progestins do not mitigate nor potentiate this effect. Future screens using this model are ideally posed to identify the mediators of sex hormone-induced thrombosis, and the results could potentially lead to innovative preventive or therapeutic modalities.