Abstract
Oxygen homeostasis is tightly controlled by the number of red blood cells (RBCs). Hypoxia increases RBCs by enhanced erythropoiesis mediated by hypoxia-inducible factors (HIFs). Upon return to normoxia, the increase in RBCs is overcorrected by preferential destruction of young RBCs, a process termed neocytolysis. Neocytolysis was first described in astronauts and later in people descending from high altitude. The molecular mechanism of neocytolysis is obscure. We hypothesized that neocytolysis occurs because of rapid transient changes of HIF levels, resulting in increased reactive oxygen species (ROS) from mitochondria in reticulocytes and defective antioxidant protection of young RBCs generated in hypoxia.
We developed a neocytolysis model by exposing mice to 12% oxygen (equivalent to 4500m altitude) for 10 days. This model recapitulates the RBC changes observed in humans exposed to hypoxia for ~30 days. Upon return to normoxia, ROS were markedly increased in reticulocytes and mature RBCs, but not in neutrophils, B- or T-cells, or monocytes. Reduction of ROS by antioxidant (N-acetyl-L cysteine) treatment attenuated hemolysis and decreased hematocrit.
To test whether HIFs (transcription factors regulating hypoxic response) contribute to neocytolysis, we repeated these experiments using Chuvash mice (which bear a VHLR200Wmutation, resulting in constitutively high HIF). These mice also showed attenuated hemolysis and decreased hematocrit; in addition, their reticulocyte half-life was higher (36.6 vs.17.8 hours in wild type). Similar findings were also observed with treatment of mice with dimethyloxallyl glycine (DMOG), an inhibitor of prolyl hydroxylase (another negative regulator of HIF). These experiments indicate the essential role of HIF pathways in neocytolysis.
Mitochondria are a major source of ROS in cells. During terminal differentiation of RBCs, mitochondria are removed from reticulocytes by mitophagy. After hypoxia treatment, mitochondria mass increased in reticulocytes concomitant with the reduction of HIF-regulated Bnip3L (a mediator of mitophagy) transcripts. These increased ROS were of mitochondrial origin, as detected by Mito-Sox staining.
To pursue the mechanism behind the preferential destruction of young RBCs, we investigated antioxidant enzymes. After hypoxia treatment, catalase decreased by 30%, but not glutathione peroxidase, superoxide dismutase or NADH oxidase. The decreased catalase in RBCs produced during hypoxia was unexpected, as it was shown previously that catalase is regulated by HIF2 (as we also show, to a lower degree by HIF1 in our Hif1a-/-embryo) suggesting alternate negative regulator(s) of catalase in hypoxia. Several hypoxia-regulated microRNAs (miRs) are reported to control oxidative stress; we found that miR-451, miR-205 and miR-21 were expressed in erythroid progenitors and reticulocytes and induced after 10 day-hypoxia exposure. To verify whether these miRs regulate catalase expression, we overexpressed and downregulated these miRs in K562 and HEL erythroid cell lines, and found that only miR-21 regulated catalase. Further, we found increased miR-21 after 10 day-hypoxia exposure, with a concomitant decrease of catalase transcripts and activity resulting in impaired ROS scavenging.
We conclude that neocytolysis is mediated by excessive generation of ROS from increased mitochondrial mass due to reduced Bnip3L in reticulocytes upon return to normoxia. The reticulocyte ROS then interact with hypoxia-produced young RBCs having miR-21-downregulated catalase, resulting in their preferential destruction. We show that increased mitochondrial ROS and miR-21-downregulated catalase provide the molecular basis of neocytolysis.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.