Hypoxia alters cellular metabolism and chemosensory endocrine regulation. Hypoxia increases production of erythrocytes in order to improve oxygen delivery to tissues. Organisms respond differentially to duration and patterns of hypoxia. Acute hypoxia stabilizes transcription factors' a-subunits of hypoxia inducible factors (HIFs; HIF-1a and HIF-2a) while in chronic hypoxia HIF-2a continues to be augmented, but HIF-1a decreases (1, 2). There are two clinical patterns of hypoxia: chronic sustained hypoxia (CSH) and chronic intermittent hypoxia (CIH). CSH is present in people living at high altitudes, patients with respiratory diseases, and Eisenmenger complex. CIH is characterized by repeated series of hypoxia followed by normoxia exposure and best exemplified by obstructive sleep apnea (OSA). The hematological responses to CSH and CIH has not been compared and studied in detail.
We studied 2 separate hypoxic mouse cohorts: 1) CSH (12% O2 for 24 hours/day for 33 days), and 2) CIH emulating human OSA (8% O2 for 30 sec followed by 21% O2 for 90 sec; 8 hours/day for 33 days) and nonhypoxic controls. Hematocrits (Hct) were higher in both CSH and CIH compared to controls. In CSH, Hct increased by 22.0% (p <0.0001) and by 7.6% in CIH (p=0.0115). To evaluate the effect of CSH and CIH on erythron we measured erythropoietin (Epo) and reticulocytes. Reticulocytes were lower in CSH and CIH compared to controls; however, CIH tended to have higher reticulocytes than CSH (p=0.0607). Epo levels decreased in CSH and CIH but not significantly in contrast to increased Epo detected at 3 days after acute hypoxia (2). These Epo levels were reflected by decrease of Epo transcripts in kidney after 33 days of CSH (p=0.0001). In contrast, there was no decrease of Epo kidney transcripts in CIH.
Hepcidin (encoded by Hamp) is the principal regulator of iron metabolism; its expression is decreased by hypoxia, iron deficiency, and augmented erythropoiesis. It suppresses erythropoiesis by inhibiting release iron from macrophages. We measured Hamp transcript in the liver. Despite presence of hypoxia, both CSH and CIH had higher Hamp transcripts. These results are consistent with co-existent suppression of erythropoiesis in both CSH and CIH, but more so in CSH. Since inflammation induces hepcidin expression, we measured inflammatory marker transcripts in granulocytes and found increased Il-6 transcripts in both CSH and CIH; Il-1a levels increased only in CSH. As Epo transcripts were decreased in the kidney, we also measured inflammatory markers` transcripts in kidney. The Ccl2, Il-10 and Il-6 transcripts were higher in both CSH and CIH. These results suggest that increased inflammation in CSH and CIH suppresses erythropoiesis via increase of hepcidin.
We also measured reactive oxygen species (ROS) in blood cells. The amount of mitochondrial ROS and the number of mitochondrial ROS positive reticulocytes increased in CSH and CIH and correlated with mitochondrial mass. We previously showed that excessive ROS associated with increased mitochondrial mass is due to reduced Bnip3L (the mediator of mitophagy) and decreased catalase upon normoxic return from CSH (2). Bnip3L and Cat (encodes catalase) were also decreased in both CSH and CIH in reticulocytes, which correlated with their increased mitochondrial mass and ROS.
We conclude that in CIH there is an increased hematocrit in mouse OSA model which contrasts with human OSA where Hct is not increased. Our data support that the increased Hct occurs in CSH and mouse CIH at the early stages of hypoxia, but that further increases of Hct is prevented by inflammation and hepcidin. The high Hct is further contributed by recent demonstration of prolonged mouse erythrocyte survival in hypoxia (4). Further, in this OSA mouse model increased ROS is restricted to reticulocytes but not present in leukocytes, unlike in human OSA (3). This demonstrates some limitation of using mouse as a model for human OSA, but also some similarities. Excessive ROS in reticulocytes due to increased mitochondrial mass and decreased catalase (due to a decrease of HIF-1a) may explain some features of pathophysiology of CSH and CIH (2).
1. Uchida et al, JBC 2004
2. Song et al, JMM 2015
3. Song et al," Normal Hemoglobin Concentrations in OSA", abstract this meeting
4. Song et al, High Alt Med Biol. 2019
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
