Oxidative Stress in Blood Cells Associated with Obstructive Sleep Apnea Contributes to Absence of Hypoxia-Induced Polycythemia

Obstructive sleep apnea (OSA) is a common condition associated with cardiovascular and endocrine comorbidities, as well as increased cancer risk and neurocognitive impairment. It has not been determined if OSA is accompanied by any hematological abnormalities. Although severe intermittent hypoxia is a unifying feature of OSA, the prevalence of hypoxia-induced polycythemia (erythrocytosis) has not been ascertained, nor its pathophysiology defined. We found that of 527 OSA patients studied, only 9 (1.7%) had polycythemia (Gangaraju, et al, abstract at this meeting); we hypothesized that excessive hemolysis associated with transitions from hypoxia to normoxia (neocytolysis) prevents polycythemia in OSA.

Neocytolysis is a mechanism that controls red cell mass. During hypoxia, a hypoxia-inducible factor (HIF), HIF2, increases erythropoietin (EPO). After return to normoxia, the high red cell mass is overcorrected by preferential destruction of hypoxia-made young red blood cells (RBCs) containing decreased catalase, downregulated by a hypoxia-upregulated miR-21. Neocytolysis is caused by excessive mitochondria-generating reactive oxygen species (ROS) from expanded mitochondria of reticulocytes mediated by decreased levels of BNIP3L expression. BNIP3L, a HIF regulated gene, causes mitophagy. Upon normoxic return, its expression decreases and reciprocally correlates with mitochondrial mass expansion and increase of miR-21 and ROS (Song et al, JMM, 2015). The ROS diffuse to plasma and cause preferential oxidative damage in those RBCs with low catalase. Thus we hypothesized that rapid cycling of hypoxia/normoxia in OSA may lead to neocytolysis and the resulting hemolysis precludes the development of polycythemia.

We analyzed 31 OSA patients without chronic cardiopulmonary disease before and after 3 months of treatment with continuous positive airway pressure (CPAP). Sleep studies (polysomnography) showed an average of 92.2 minutes spent with oxygen saturations (SpO₂) of <89%, with a 4% oxygen desaturation index of 23/hour and mean of the lowest SpO₂ value of 75%. EPO levels before CPAP treatment were correlated with periods with SpO₂ <89%; those with severe OSA had higher EPO levels. In this cohort, EPO decreased after CPAP, while the group with lower EPO (<8 mU/mL) had increased EPO (p=0.03); however, hematocrit levels were not changed after CPAP. In some patients, we documented hemolysis by increased end-tidal CO in exhaled air (intrinsic CO is generated by catabolism of hemoglobin). Mitochondria and ROS levels in reticulocytes and RBCs correlated with time spent below SpO₂ of 89% and decreased after CPAP. Transcript levels of CAT (encoding catalase) and BNIP3L in reticulocytes were increased and levels of miR-21 reversely-correlated after CPAP in those with higher EPO. In contrast to normoxic return from chronic hypoxia, wherein increased ROS and mitochondrial mass were seen only in reticulocytes, ROS levels were also increased in OSA nonerythroid cells and decreased after CPAP. CAT levels also increased after CPAP in granulocytes and platelets regardless of severity of OSA and EPO concentration. We conclude that patients with severe OSA have hemolysis mediated by excessive mitochondria-driven increase of ROS, possibly also from other cells than reticulocytes, which prevents polycythemia in most OSA. Evaluation of mitochondrial mass in nonerythroid cells is ongoing.

We measured HIF activity by quantifying transcripts of HIF target genes and found that HIF regulates gene transcripts in a tissue-specific manner, resulting in a heterogeneous effect in different blood cell lineages.

We demonstrate that neocytolysis plays a role in OSA via increased mitochondria-driven ROS in reticulocytes and nonerythroid cell interactions with RBCs with low catalase, thus preventing polycythemia in OSA. We also show that the severity of OSA correlates with the rate of erythropoiesis, increased oxidative stress, and decreased HIF target genes, suggesting that severe OSA causes more hemolysis than mild OSA. We will evaluate the role of increased erythropoiesis and inflammation in OSA by hepcidin, erythroferrone and C-reactive protein levels that may contribute to abrogation of polycythemia.

The increased ROS in nonerythroid blood cells that resolve with CPAP suggests that oxidative injury may contribute to the pathophysiology of OSA.