As a normal physiological compensatory response to a hypoxic environment, high-altitude (HA) migrants often experience increases in their number of red blood cells (RBCs), which improve the oxygen-carrying capacity of the blood (Windsor and Rodway, 2007). However, some migrants (~10–40%, depending on the altitude) have excessive erythropoiesis after several months or years, which results in high altitude polycythemia (HAPC), also known as chronic mountain sickness (CMS) or Monge's disease (Negi et al., 2013). Patients with HAPC experience increased blood viscosity and symptoms such as headache, dizziness, palpitations, and insomnia. Currently, there is no effective treatment method for HAPC because the pathogenesis of HAPC is complicated (Villafuerte, 2015). Although hypoxia-induced erythropoietin (EPO) is considered a major hormone that regulates erythropoiesis, the levels of EPO usually do not correlate well with erythrocyte production in many patients with HAPC (Gore et al., 2006; Painschab et al., 2015). It is necessary to explore other mechanisms that may be involved in erythrocytosis at high altitudes.

Iron has a variety of important physiological functions in the human body and is an important component of heme in hemoglobin. The increase in the number of RBCs under hypoxic conditions requires more iron utilization. Iron in the body comes from absorption of iron from the intestines (called exogenous iron), recycling of iron from the hemoglobin of damaged senescent erythrocytes and release of stored iron from the liver, muscle, and other tissues (called endogenous iron). Under HA hypoxic conditions, it is necessary to recruit more iron to provide enough materials for erythropoiesis (Cook et al., 2005; Robach et al., 2007; Gassmann and Muckenthaler, 2015). Theoretically, enhanced iron mobilization at high altitudes is controlled within a reasonable range to match the compensatory increase in RBCs to adapt to hypoxia. This raises the question of whether the status of an individual's iron metabolism (including iron storage, mobilization, and utilization) under conditions of a long-term hypoxic environment is one of the factors that limits or promotes the progression of HAPC.

Recent studies have revealed that hepcidin, which is mainly synthesized by the liver, plays a central and critical role in regulating iron homeostasis (Ganz, 2011; Deschemin et al., 2017; Sangkhae and Nemeth, 2017). Hepcidin can bind with ferroportin 1 (Fp1), a transmembrane protein in enterocytes, hepatocytes, and macrophages. After binding, Fp1 is internalized and degraded, leading to decreased exogenous intestinal iron uptake and endogenous iron release from hepatocytes and RBC phagocytic macrophages, which accurately regulates the intake and recycling of iron in the body (Nemeth et al., 2004). Studies have confirmed that disorders of hepcidin are a main contributing factor to chronic anemia in patients with infections and tumors (Reichert et al., 2017) and that hepcidin is involved in the pathological processes of hereditary hemochromatosis, thalassemia and iron deficiency anemia (Girelli et al., 2016; Daher et al., 2017).

It has been reported that exposure to high altitude hypoxic environments for several days or weeks significantly affect iron metabolism and hepcidin levels (Talbot et al., 2012; Goetze et al., 2013). It remains largely uncharacterized how iron status changes for residents who have lived at high altitudes for months or even years compared with their iron status during their early emigration period. Moreover, no literature has been reported on the features of iron metabolism and hepcidin for HAPC patients. Iron homeostasis is generally regulated by the levels of body iron load and iron demand. However, at high altitudes, many factors related to hypoxic environments may affect the regulation of iron, including chronic hypoxic stimuli (Hintze and McClung, 2011); EPO levels (Gammella et al., 2015); erythroid proliferation signals, such as growth differentiation factor 15 (Tantawy et al., 2014) and the newly discovered erythroferrone (Kautz et al., 2014); hypoxia-induced inflammatory cytokines, such as interleukin (IL)-1 and IL-6 (Hartmann et al., 2000; Lee et al., 2005; Schmidt, 2015). These may all affect the hepcidin level and the iron balance. Among the above mentioned factors, except for disordered EPO, hypoxia-induced inflammation or immune imbalance at high altitudes is a prominent problem with large individual differences in the population (Boos et al., 2016; Evstratova et al., 2016; He et al., 2016).

The present research systematically examined the iron status and relevant regulators in patients with HAPC and compared them with subjects at sea-level and with healthy high-altitude migrants. Using experiments in mice under a simulated high-altitude hypoxic environment, we also studied the effects of different levels of iron and iron-regulatory cytokines on erythrocytosis. These results revealed that the individual regulatory ability of iron metabolism was a non-negligible factor in the pathogenesis of HAPC.

Our study found, for the first time, that increased iron availability was a notable feature of patients with HAPC and was associated with disordered IL-10 and IL-22. We performed experiments in mice and demonstrated that excessive obtainable iron can promote erythroid hematopoiesis, and the diminished hepcidin resulting from decreased IL-10 and IL-22 was an important cause of excess iron supply. These results from human and murine studies suggest that an imbalance in immune-inflammatory cytokines participate in promoting excessive RBC production in some patients with HAPC. With the increase in the RBC counts and hemoglobin levels among the high altitude migrants, serum sTfR released by the erythropoietic precursors significantly increased because of a higher speed of iron use (Beguin, 2003). Meanwhile, total serum iron and TIBC were elevated in the HA migrants compared with the sea-level residents, which revealed that more iron was prepared due to advancing erythropoietic demands (Northrop-Clewes and Thurnham, 2013). In patients with HAPC, the vigorous production of RBCs resulted in further enhanced iron utilization compared with healthy HA migrants. In parallel, we observed a higher level of ferritin in patients with HAPC, which reflects excessive iron storage from the continuous destruction of aging RBCs (Salvin et al., 2014). It is remarkable that the serum iron levels were much higher in patients with HAPC due to increases in iron mobilization from exogenous absorption and endogenous reuse. The simultaneous increase in iron storage and iron mobilization in HAPC patients suggests that the massive amount of iron coming from damaged RBCs constantly repeats the cycle of steps involved in storage and reuse. In addition, iron availability was significantly associated with the severity of HAPC. This result has not been reported in previous studies.

To clarify whether easy access to iron is a risk factor for or a stimulus of HAPC, we injected mice with different concentrations of iron, establishing mouse models with different levels of iron availability under high altitude hypoxic conditions, and then observed the effects on erythropoiesis. Similar to previous studies (Hartmann et al., 2013; Lu et al., 2013), we found that a high concentration of iron can inhibit hematopoiesis, particularly the BFU-E and CFU-E progenitor cells and Ter119^medCD71^high proerythroblasts, indicating iron toxicity to erythroid precursors of earlier differentiation stage. However, we also found that low and moderate levels of extra iron supply, as observed in HAPC patients, can promote erythroid precursors formation in the BM, especially the cells after CFU-E stage which begin to express the transferrin receptor (Suzuki et al., 2003). Unlike the BM of patients with aplastic anemia and myelodysplastic syndrome, which is more sensitive to iron toxicity as a result of poor BM status (Della Porta et al., 2015), the BM tissues of hypoxia-induced HAPC patients were not usually damaged. Thus, excessive iron within a certain range could be transformed into a favorable factor to promote erythroid hematopoiesis in HAPC, which contributed to the worsening of uncontrolled erythrocytosis and was related to the severity of HAPC.

Consistent with the results on short-term HA travelers published by others (Piperno et al., 2011), our results showed that hepcidin levels remained continuously low during chronic hypoxic exposure in long-term HA residents. Crucially, we found a remarkable decrease in hepcidin levels in patients with HAPC compared with healthy migrants who lived at the same altitude. Decreased hepcidin may be involved in the occurrence of excessive iron mobilization and subsequent runaway polycythemia among some HAPC patients, and this hypothesis is supported by the correlation analysis between the hepcidin levels and both serum iron concentrations and hemoglobin contents. The decrease in hepcidin is beneficial to the body if it is moderate and compatible with the physiological iron demand at high altitudes. However, greater decreases in hepcidin in HAPC may result in excessive iron absorption and the reuse and release of stored iron to exceed the iron demands of the body. It can be speculated that those without HAPC have controlled iron availability to synthesize hemoglobin. To a certain extent, iron deficiency limits the rate of erythrocyte formation and the development of HAPC. In contrast, it is easier for HAPC patients to obtain iron than others, particularly through the reuse of iron from damaged aging RBCs, as raw material for generating RBCs.

As the leading hormone that regulates body iron metabolism under high altitude hypoxic conditions, hepcidin was regulated in response to a series of signaling pathways related to hypoxic exposure, including signals involved in increased iron demand, such as transferrin (Bartnikas et al., 2011); signals related to enhanced erythropoiesis, such as GDF-15 (Ronzoni et al., 2015); and erythroid growth factors, such as EPO (Ribeiro et al., 2016). EPO and inflammation are the two most prominent features under high altitude conditions that may influence adaptation to the hypoxic environment (Lemos Vde et al., 2013). Our results showed that although EPO, IL-6, and IL-1β were easily increased in HA migrants, the possibility that these cytokines induced hepcidin downregulation was ruled out because no substantial difference was observed between the HA subjects with and without HAPC.

The crucial findings of this study were the roles of IL-10 and IL-22 in influencing the incidence and severity of HAPC in a hypoxic HA environment. Hypoxia usually leads to oxidative stress and inflammation, and some HA migrants could not launch an effective anti-inflammatory defense, which is involved in the pathogenesis of acute and chronic HA diseases. IL-10 is an anti-inflammatory cytokine that is produced by monocytes, type 2 T helper cells and CD4+, CD25+ and Foxp3+ regulatory T cells (Tregs), and lower IL-10 is associated with susceptibility to acute mountain sickness (Liu et al., 2017). IL-22 is produced by activated NK and T cells, and its role involves innate immune responses and Th17 function. Although IL-22 has a pro-inflammatory effect in the acute inflammatory response, it often plays a protective role in the process of chronic inflammation (Wolk et al., 2011). Recent studies have suggested that hypoxia may inhibit the expression of IL-22 and its activity (Budda et al., 2016, 2017). Our observations suggested that the progression of HAPC was accompanied by a reduction in these two protective cytokines and an a higher inflammatory susceptibility, which were also supported by the detection result of other pro-inflammatory cytokines such as higher IL-17A and IL-21 in HAPC subjects. We speculate that the immune balance of Treg and Th17 in HAPC patients may be broken, showing a trend toward Th17 polarization and a predominance of proinflammatory power.

We excluded the possibility that IL-10 and IL-22 affected erythroid production through direct stimulation of progenitors and erythroblasts. The decrease of these two cytokines should affect the erythroid hematopoiesis in an indirect way. IL-10 and IL-22 have been shown to trigger the expression of hepcidin in previous studies (Smith et al., 2013; Chang et al., 2014; Huang et al., 2014; Wallace and Subramaniam, 2015). IL-10 and IL-22 share common receptors, and both can activate the STAT3 signaling pathway of the target cells; in turn, STAT3 can upregulate the transcription of hepcidin in various cells (Lu et al., 2016; Kong et al., 2018). During the inflammatory process caused by infections, elevated IL-22 can limit iron availability to bacteria through inducing hepcidin production, which leads to suppression of bacterial growth (Sakamoto et al., 2017). In contrast, the immune activation caused by HA hypoxia is non-pathogenic, and IL-10 and IL-22 have different expression patterns and biological effects. Although hypoxia itself inhibits the production of hepcidin directly or indirectly, our results from the mice confirmed that reduced IL-10 and IL-22 were beneficial to the maintenance of hepcidin at a lower level during long-term hypoxia. This explains why HAPC patients with lower levels of IL-10 and IL-22 had lower hepcidin levels and more available iron for RBC production.

The present study confirms that the individual control of iron metabolism is not well matched with a reasonable compensatory increase in RBCs in some HAPC patients. The abnormal hepcidin caused by dysregulation of IL-10 and IL-22 is involved in the progression of HAPC. Our data suggest a new theory of pathogenesis and provide some prevention and treatment ideas for HAPC. For those HAPC patients with excessive iron load, the actual effects of iron removal therapy and immunomodulatory treatment can be explored in the future.

PL designed the study and drafted the report; Y-SL contributed to samples collection and detection; HH and S-MZ contributed the laboratory experiments; HT did the statistical analysis. All authors revised the manuscript critically for important intellectual content and approved the final manuscript.