The Gluthation peroxidase (GPX) enzymes are part of the protective system against lipid peroxydation that includes prevention of oxydation and reduction of already oxidized lipid through enzymatic reactions catalyzed by GSH. GPX4 is one of the five GPX able to incorporate selenium. It is also the only GPX able to directly reduce in the membrane the oxidized fatty acids and cholesterol. Recent reports identified GPX4 as the central inhibitor of ferroptosis, a process during which iron-induced peroxydation of membrane lipids causes a specific cell death that can be reverted by lipophilic antioxydants or by iron chelators. GPX4 has recently been involved during mice erythropoiesis: GPX4-/-mice present a hemolytic anemia and a high apoptotic rate in spleen erythroid progenitors. Although transcriptomics and proteomics found it expressed in human erythroid precursors, its role during human erythropoiesis has not been described. Using an in-vitroerythroid differentiation protocol from CD34+cells obtained from apheresis, we confirmed that GPX4 expression was induced at RNA and protein level during differentiation. RSL3, a specific GPX4 inhibitor, didn't affect early steps of erythropoiesis (i.eclonogenic potential and progenitor amplification) nor the early maturation of erythroid precursors (assessed by sequential CD49d/CD235/CD71 staining) but led to a significant decrease in the enucleation rate as assessed by Hoechst staining using flow cytometry (74%±9 DMSO versus 35%±6 RSL3, p<0.01) and by cytology after MGG staining (58%±10 DMSO versus 27% RSL3±5, p<0.05). This effect was not related to ferroptosis since (i) FIN56 and erastin, two other ferroptosis activators, didn't impact enucleation and (ii) the enucleation defect was not reverted by tocopherol or ferrostatin. Of note, tocopherol reverted lipid peroxydation at cell surface, as shown by Bodipy-C11 staining, showing that enucleation defect and lipid peroxydation were not directly related. These data argue for a specific GPX4 role in the enucleation process, independently to its well-described function in ferroptosis control.
Using Western Blot, we observed that RSL3 exposure induced a strong GPX4 depletion in erythroid progenitors while it didn't affect GPX1, another member of the GPX family expressed in erythroid cells. In order to confirm that enucleation defect was related to GPX4 knockdown, we used an Sh-RNA strategy that allowed a 62%±6 GPX4 decrease at RNA level and a 46%±5 at protein level. We observed a significant defect in terminal enucleation in the cells transduced with shGPX4-lentiviruses in comparison with sh-Scramble (59%±5 Sh-Scr versus 39%±6 Sh-Gpx4, p<0.05). Isopentylpyrophosphate, an intermediate component of cholesterol synthesis that also acts as a substract of selenoprotein synthesis, restored partially both GPX4 expression and erythroblast enucleation.
We investigated then whether GPX4-knock-down affected quantitatively or qualitatively the membrane lipid content, which was shown to be involved in the enucleation process. Addition of Cholesterol to RSL3 in the medium partially restored the enucleation rate. However, lipidomics failed to show any significant difference in the total membrane lipid content (and particularly in the cholesterol content) after RSL3 exposure in comparison to DMSO. Since cholesterol is particularly abundant in the lipid rafts, we investigated whether the lipid distribution was qualitatively altered within the cell membrane. We observed a disruption of membrane lipid rafts when cells were exposed to RSL3, as shown by a 60%±12 decrease in the mean cholera toxin fluorescence intensity. GPX4 presence in lipid rafts was confirmed using immunofluorescence showing their co-localization at cell surface in human primary erythroblasts. Since lipid rafts play a role in the contractile ring that separates pyrenocyte from reticulocyte, we evaluated the myosin-light chain phosphorylation using flow cytometry and Western Blot and found it drastically decreased in GPX4-knockdown conditions.
In summary, we identified GPX4 as a new actor of human terminal erythroid differentiation, independently to its function in ferroptosis control. We described its interaction at cell surface with lipid rafts that are required for the assembly of the contractile ring and cytokinesis leading to the nucleus extrusion.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.