The principal function of red blood cells (RBCs) is to deliver oxygen via hemoglobin resulting in the formation of reactive oxygen species (ROS), which mandates the existence of an ROS protective mechanism. Hypoxia increases RBC mass by increasing hypoxia-inducible factors (HIFs) levels. This is mainly mediated by HIFs-regulated erythropoietin. Upon return to normoxia, RBC mass is overcorrected by the destruction of newly formed RBCs, a process termed neocytolysis. Neocytolysis was first described in astronauts and people descending from high-altitude, and likely accounts for the rapid transition from polycythemia to anemia in neonates upon leaving the fetal hypoxic environment. The molecular mechanism of neocytolysis has not yet been elucidated. Reticulocytes newly released from the marrow contain ribosomes and mitochondria which are removed by multiple mechanisms, including mitochondrial loss by autophagy regulated by HIFs-controlled Bnip3L gene. FOXO3a is a HIF-regulated transcription factor that controls ROS-scavenging enzymes. Mature RBCs cannot respond to rapid changes of ROS levels because of their inability to produce new proteins. We hypothesized that after hypoxic exposure and return to normoxia, there is an increase in reticulocytes' ROS (mediated by Bnip3L), causing preferential destruction of the hypoxia-born young RBCs having decreased activity of FOXO3a-controlled ROS-scavenging enzymes.
We developed a mouse model of neocytolysis that replicates RBC changes observed in humans to test our hypothesis, with awareness that RBCs live ∼120 days in humans and ∼45 days in mice. C57/BL6 mice were maintained at 12% oxygen (equivalent to 4500 m altitude) for 10 days followed by return to ambient air. We measured the following parameters: hematocrit (hct), reticulocyte count, RBC survival, mitochondrial mass, mRNA and the activity of ROS-scavenging enzymes, Bnip3L and FOXO3a mRNA, and ROS content in peripheral blood cells before and after hypoxic exposure.
Hypoxic exposure increased hct by 22%; upon return to normoxia, hct reached normal levels at post-hypoxia days (PHD) 4–7, then continued to decrease by 28% at PHD 7–10, while plasma volume did not significantly change until its decrease at PHD 10 (p=0.049). Reticulocyte count increased by 35% following hypoxia, then decreased by 33% at PHD 4 compared to pre-hypoxic levels. Reticulocytes recovered by 50% by PHD 7, then normalized at PHD 14. To measure RBCs survival, we labeled them with biotin prior to hypoxic exposure and found preferential destruction of newly formed RBCs in the initial phase of normoxic return. This was accompanied by a dramatic increase of ROS level at PHD 0 and PHD 4 confined only to RBCs (and preferentially to young RBCs, i.e. 2.6 fold higher at PHD 4 in CD71+/TER119+ compared to CD71−/TER119+ in older RBCs) but not to platelets, neutrophils, B and T cells, or monocytes. We demonstrated that the ROS changes were accompanied by a commensurate increase of mitochondrial mass that was limited to reticulocytes; i.e. mitochondrial mass increased in reticulocytes by 40%, with an inverse relation to reticulocytes' Bnip3L transcript levels (reduced by 35%). Furthermore, we show the transcripts and activity of FOXO3a-regulated ROS scavenging enzyme catalase were reduced at PHD 0 and PHD 4 by 30% and recovered at PHD 10, accompanied by commensurate changes in FOXO3a mRNA. However, there were no similar changes in glutathione peroxidase and superoxide dismutase. These data demonstrate that the increase in ROS is due to both excessive generation in mitochondria and insufficient scavenging due to a decrease in catalase activity.
We conclude that there are three different phases in the response of RBC mass to transient hypoxia: 1) Hypoxia increases erythropoiesis, resulting in polycythemia; 2) Upon return to normoxia, the increased RBC mass is overcorrected below the initial levels by neocytolysis; 3) After 10–14 days, RBC mass normalizes. We demonstrate that neocytolysis is mediated by ROS in reticulocytes, with increased mitochondrial mass due to downregulation of Bnip3L. These increased ROS cannot be scavenged by young RBCs with decreased catalase activity. These two processes lead in concert to neocytolysis. These studies reveal a novel role for mitochondria-derived ROS in the regulation of red cell mass.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.