Background: Protein S (PS) and growth arrest-specific gene 6 (Gas6) are vitamin K-dependent proteins. PS is an anticoagulant whose pivotal role is illustrated by purpura fulminans (PF), a life-threatening condition characterized by purpuric skin lesions, disseminated intravascular coagulation and thrombosis. Unlike PS, Gas6 displays a procoagulant effect. Its complete deficiency has anti-hemostatic, pro-inflammatory and anti-erythropoietic effects. In tissues, PS and Gas6 have an impact on cellular functions by binding to and activating receptor tyrosine kinases of the TAM family. PS complete deficiency-induced PF is thought to result from the imbalance between procoagulant and anticoagulant forces. However, its pathophysiological mechanism has not yet been fully elucidated. Previous studies showed that Pros1-/-mice die in utero with a PF phenotype similar to the one observed in human. We therefore considered Pros1-/- mice as a suitable model of the human PF. The aim of this work was to investigate if Gas6 deficiency could rebalance hemostasis in PS deficient mice and rescue them from PF.

Methods & results:Pros1+/-Gas6-/- mice were intercrossed to obtain Pros1-/-Gas6-/- mice. Embryonic mortality was higher in Pros1+/-Gas6-/- than in Pros1+/- mattings (6% vs 3 % at E14, 21% vs 18 % at E16, 30% vs 20 % at E17, respectively). About 7.5% of dead embryos from Pros1+/-Gas6-/- mattings were found necrotic and macerated versus only 3.8% inPros1+/- mattings. Macroscopically, 74% of Pros1-/-Gas6-/- embryos displayed maximal bleeding scores (3) corresponding to intracranial and major body bleedings compared to 47% in Pros1-/- embryos (p<0.02). Microscopically, all Pros1-/-Gas6-/- embryos had a bleeding score of 3, whereas 50% of Pros1-/- embryos displayed a milder bleeding score (2). Major blood vessels of E16 Pros1-/- and Pros1-/-Gas6-/- embryos contained a high number of non-phagocyted red blood cells (RBC) expelled nuclei (38% and 58%, respectively) while only few of them were found in Pros1+/+Gas6+/+ embryonic vessels (5%), indicating that altered phagocytosis due to the lack of PS was further aggravated by Gas6 deficiency (p<0.003).

More immature RBC were found in Pros1-/- than in Pros1+/+Gas6+/+ circulating blood (17% vs 12%, p<0.004), suggesting increased erythropoiesis in Pros1-/- embryos in response to bleeding. These data were confirmed using FACS analysis of CD71/Ter119 stained liver single cell suspensions. E14 Pros1-/- embryos had more immature RBC compared to Pros1+/+Gas6+/+ embryos (2.3% vs 1.4%, p<0.08). However, Pros1-/-Gas6-/- embryos displayed less immature RBC in their circulating blood than Pros1-/- embryos. The number of erythroid burst-forming units was lower in Pros1-/-Gas6-/- fetal livers compared with Pros1-/- fetal livers (5±1.3 vs 8±2, p<0.08) confirming reduced erythropoiesis in Pros1-/-Gas6-/- embryos. Perls staining showed increased iron stores in fetal liver and numerous intra-vascular hypochromic RBC in Pros1-/-Gas6-/- embryos pointing to a functional iron deficiency. We also investigated E15 embryonic dorsal skin by immunofluorescence. We found RBC extravasation (VE-cadherin and Ter119 stainings), vascular network outgrowth impairment with fewer vessel branches (CD31 staining), massively enlarged lymphatic vessels and increased macrophages infiltration (Lyve1 and F4/80 stainings) in Pros1-/-Gas6-/- and Pros1-/- indicating an ongoing inflammatory process.

Conclusion: Gas6 deficiency did not rebalance hemostasis in Pros1-/-embryos and prevent PF. Notably, Gas6 and PS combined deficiency leads to a more dramatic phenotype than PS deficiency alone with a higher mortality rate, more major bleedings, phagocytosis impairment, more inflammation and an erythropoietic defect compatible with anemia of inflammation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.