Abstract
Late in erythropoiesis, nuclei are expelled from erythroblasts and 2×1011 anucleated new red blood cells are daily delivered in the peripheral blood. Expelled nuclei expose phosphatidylserine (PS) on their surface, which is used as an “eat me” signal for their engulfment by macrophages located in the blood island. The two PS opsonins, milk-fatglobule EGF8 (MFG-E8) and Growth arrest-specific gene 6 product (GAS6) together with their respective receptors αvβ5/αvβ3 and TAM (TYRO3, AXL and MER), are involved in the phagocytosis of apoptotic cells, but their role in the phagocytosis of expelled nuclei from erythroblasts is not determined. Because fetal liver and bone marrow macrophages do not express MFG-E8, the GAS6-MER pathway might constitute a crucial pathway for the engulfment of nuclei expelled from erythroblasts. To test this hypothesis, we isolated nuclei from late-stage erythroblasts from spleens of phlebotomized mice, and studied nuclei internalization capacity of bone marrow derived macrophages (BMDM) from mice deficient either in GAS6 (GAS6−/−), AXL (AXL−/−) or TYRO3 (TYRO3−/−), or lacking MER kinase domain (MERkd). Released nuclei were identified by flow cytometry according to their size and their double positive staining for the erythroid lineage marker Ter119 and Annexin V for PS. Purity of the preparation was checked by morphological examination of cytospin preparations. In vitro phagocytosis assays show that GAS6−/− BMDM cleared 30% less nuclei than wild-type (WT) BMDM. We observed a slight decrease of internalization capacity for AXL−/− BMDM, whereas TYRO3−/− BMDM engulfed the nuclei as efficiently as WT BMDM. In contrast, MER deficiency nearly abolished nuclei phagocytosis. AXL−/−/TYRO3−/− and AXL−/−/MERkd BMDM were tested and did not show any cumulative effects when compared to WT and single knockouts. We also investigated the signalling pathway downstream of MER in BMDM. In particular, we assessed the expression of the activated form of Rac1, which is crucial for the cytoskeletal reorganization in phagocytosis. Activation of Rac1 after the initiation of the phagocytosis was delayed for 45 minutes in MERkd as compared to WT BMDM. In vivo, we found an accumulation of nuclei in MERkd mice 4 days post phlebotomy, when erythropoiesis is increased in response to anemia. Nuclei circulated in the blood of MERkd mice at a level of 0.08 ± 0.042 G/L and were identified on peripheral blood smears of these mice whereas they were undetectable in the blood of WT mice. We demonstrated an increase of a double labelled Ter119/AnnexinV population corresponding to nuclei in BM (2-fold) and spleen (1.5-fold) of MERkd mice as compared to WT mice. The augmentation of this double labelled population in the MERkd mice translated the phenotype of splenomegaly of these mice. Hematocrit and reticulocyte levels were comparable between WT and MERkd as previously reported (JCI118:583–596, 2008). Thus, MER was critical for in vitro phagocytosis of nuclei from erythroblasts whereas the role of AXL and TYRO3 appeared to be negligible. GAS6 binding to nuclei exposing PS on their surface might form a bridge between PS and MER receptor on BMDM, allowing nuclei clearance. In vivo, the absence of MER caused an accumulation of nuclei in BM and spleen and their appearance in circulating blood due to their inefficient elimination during erythropoietic response to anemia. In conclusion, we postulate that GAS6 and its receptor MER were involved in late erythropoiesis when nuclei are expelled from the erythroblasts and engulfed by BMDM in the blood island, through Rac1 activation.
Correction to: Dragonfies of Dragon’s Blood Island: Atlas of the Odonata of the Socotra Archipelago (Yemen)