Background

A plateau is a special environment characterized by low atmospheric pressure and low partial oxygen pressure. Long-term exposure to plateau environments may lead to chronic mountain disease [1]. High-altitude polycythemia (HAPC) is a common and widespread chronic mountain sickness characterized by excessive erythrocytosis [2]. Hypobaric hypoxia (HH) is the main cause of erythrocytosis to alleviate hypoxic conditions in tissues under high-altitude (HA) exposure [3]. In a healthy organism, erythropoiesis in bone marrow is equivalent to eryptosis in the spleen [4]. Although HA/HH exposure can affect the balance between the formation and clearance of RBCs, a new physiological RBC steady state can be rapidly formed in normal individuals [5]. Nevertheless, RBC homeostasis is disturbed and never formed into a homeostasis status in HAPC patients, which finally results in a continuous increase in RBCs [4]. It is well known that HA/HH exposure may lead to erythropoiesis; however, the pathogenesis of eryptosis under HA/HH conditions remains unclear.

The spleen clears the altered RBCs in the circulation system, which helps to balance erythropoiesis and RBC clearance [6]. Moreover, red pulp macrophages (RPMs) of the spleen are the major scavengers that remove unhealthy, old, and malformed RBCs from the circulation to recover iron [7]. In brief, RPMs recognize the RBCs to be removed through the signal on their cell surface and trigger RBC endocytosis and digestion into heme in lysosomes [8]. Subsequently, the heme is decomposed by heme oxygenase-1 (HO-1) into biliverdin, carbon monoxide and iron [9]. Most of the iron recycling from heme is transported out of the cell through ferroportin (Fpn) and then participates in the formation of RBCs through the combination of transferrin (Tf) and transferrin receptor (TfR) on the cell membrane [10]. A portion of the released iron is loaded into cellular ferritin (Ft; including Ft-H and Ft-L). Ft-H oxidizes Fe²⁺ (ferrous ion) to Fe³⁺ (ferric ion) in the presence of O₂, and then Fe³⁺ can be stored on Ft-L [11]. When the body’s iron metabolism is vigorous, nuclear receptor coactivator 4 (NCOA4) can recognize Ft-H, bring Ft into the lysosomal pathway for degradation, and release iron in Ft for iron recycling in the body [12]. However, some iron is free in cells in the form of ferrous iron (Fe²⁺), which is well known to be an important initiator of free radical oxidation. Moreover, the accumulation of large amounts of Fe²⁺ ions in cells may cause ferroptosis [13].

Although erythropoiesis under HA/HH is well studied thus far, the precise effects of HA/HH on erythrophagocytosis in the spleen remain largely unexplored. Based on the fact that the spleen is the major organ for RBC processing and that RPMs play an important role in recycling iron from RBC clearance, we investigated the effects of HA exposure on spleen/splenic macrophages in an HH-exposed mouse model (6000 m exposure conditions). In the current study, we further explored whether HA/HH could regulate the erythrocyte disposal and iron recycling of macrophages, and specifically predict the damage of erythrophagocytosis after HA/HH exposure. We demonstrated that HH exposure induced ferroptosis in the spleen, especially in macrophages, which led to a decrease in the number of macrophages, followed by eryptosis and iron recycling in the spleen. These findings may be clinically relevant to continuous pathological erythrocytosis and HAPC progression under HA exposure.

Methods

HH mouse exposure model

Male C57BL/6 male mice (at 8 weeks of age) were obtained from the animal experimental centre of Nantong University. Mice were adapted to the facilities for 3 days before experiments. Mouse models of HH exposure were established by placing the mice in the decompression chamber (TOW-INT TECH, ProOx-810, China) for 1 d, 2 d, 3 d, 7 d, and 14 d. The parameters were set as follows: 6000 m above sea level, oxygen partial pressure of 11 kPa, lifting speed of 20 m/s, chamber pressure of 54 kPa, temperature of 25 ℃, humidity of 40%, and 12-h light/dark cycles. Tuftsin (1.5 mg/kg, MCE, HY-P0240, USA) was used to stimulate the phagocytosis of macrophages and was injected at 7 d and 11 d during the 14-d period of mice exposed to HH.

Splenic macrophage culture and treatments

After isoflurane anaesthesia, mice were transcardially perfused with precooled PBS, and then splenic tissue (approximately 1mm³ in volume) was collected and separated in cold, phenol red-free Hanks’ balanced salt solution (HBSS; Gibco, Grand Island, NY, USA). The cell extracts were centrifuged at 500 rpm for 5 min at 4 ℃. The pellets were then resuspended and incubated with ammonium-chloride-potassium (ACK) lysing buffer (Thermo Fisher Scientific, USA) for 3 min at room temperature. Next, cell extracts were purified in a four-phase Percoll (Sigma‒Aldrich, St. Louis, MO, USA) gradient: 0, 30, 40, and 50% in HBSS. After centrifugation at 500 rpm for 20 min at 4°C, the macrophage-enriched fraction was collected at the interface between the 30% and 40% phases. Cells were washed in two rounds with HBSS and then resuspended in AIM V medium (Thermo Fisher Scientific). The cells in this medium were cultured at 37 ℃ and 5% CO₂ for 12 h. After the cells were treated with ferrostatin-1 (Fer-1, 2 μm; Selleck Chemicals, S7243, USA) 1 h in advance, the cells were placed in a hypoxia workstation (Invivo2, Ruskinn, UK) for 24 h.

Blood smear and hematological indices

The collected blood was gently mixed with EDTA (15 g/L) to obtain anticoagulant blood, which was detected by a hematology analyser (XE-5000, Sysmex Corporation, Japan). Blood smear preparation method: a blood sample is dropped onto a slide, and another slide is used to push the blood sample into a uniform thin layer at a constant speed. After the blood samples were dried, Wright’s stain solution was added and incubated for 1 min. Then, the slide was rinsed with water. Blood smears were photographed with a DM4000B microscope (Leica, Germany).

Western blot

Tissue and cell samples were homogenized on ice after cell lysates and protease inhibitors were added. Centrifugation was performed at 12,000 rpm for 15 min at 4°C to collect the supernatant. Protein concentration was determined by the BCA detection method. A sample containing 30 μg protein was loaded and run in each well of SDS‒ PAGE gels. The membranes were incubated with the following antibodies: HO-1 (1:2000, Abcam, ab13243, USA); Ft-L (1:1000, Abcam, ab69090, USA); Ft-H (1:1000, Novus, NBP1-31944, USA); NCOA4 (1:1000, Santa Cruz, sc-373739, USA); TfR (1:1000, Thermo Fisher, 13-6800, USA); Fpn (1:1000, Novus, NBP1-21502, USA); ACSL4 (1:1000, Santa Cruz, sc-271800, USA); xCT (1:1000, Proteintech, 26864-1-AP, USA); Gpx4 (1:1000, Abcam, ab125066, USA); CD206 (1:1000, RD, AF2535, USA); CD16 (1:2000, RD, AF1960, USA); and β-actin (1:1000, Sigma‒Aldrich, A5316, USA). The secondary antibodies were as follows: goat anti-mouse (1:10000, Jackson, USA) and goat anti-rabbit (1:10000, Jackson, USA). The gray value of specific blots was scanned and analysed using ImageJ software (National Institute of Health, USA).

RT‒PCR analysis

Total RNA extraction and RT‒PCR analysis were performed essentially as described previously [14]. Primer sequences were as follows:

GEO analysis

GSE46480 samples were found in the GEO database of NCBI with the keyword “High Altitude” search. The data set was divided into two groups, one for the plain (Base Camp) and the other for the plateau (Altitude), with a sample size of 98 for each group. Blood samples were collected from the same person at McMurdo Station (48 m) and immediately transferred to Amundsen-Scott South Pole Station (2835 m) on the third day. R language was used for data analysis, and GraphPad Prism was used for statistics. Malondialdehyde (MDA), Cysteine (Cys) and Glutathione (GSH) content detection MDA was detected by a Micro Malondialdehyde Assay Kit (Solarbio, BC0025, China). Cys was detected by a Micro Cysteine Assay Kit (Solarbio, BC0185, China). GSH was detected by a Micro Reduced Glutathione Assay Kit (Solarbio, BC1175, China).

Immunofluorescence

Frozen sections of the spleen (9 μm) were stored at −80°C. Sections were placed at room temperature for 30 min before immunofluorescence staining, incubated in 10% BSA-PBS for 10 min, incubated overnight with F4/80-PE (1:200, Biolegend, 123110, USA), washed with 0.05% PBST three times, incubated with DAPI for 10 min, washed with PBS three times, sealed with 50% glycerine-PBS, and photographed with a confocal laser scanning microscope (SP8, Leica Microsystems, Wetzlar, Germany).

Tissue iron staining (DAB-enhanced Perls’ staining)

DAB-enhanced Perls’ staining for the spleen paraffin sections was performed as described previously [15]. Briefly, the sections were washed with PBS and cultured in freshly prepared Perls’ solution (1% potassium ferricyanide in 0.1 M hydrochloric acid buffer). The slides were then immersed and stained with DAB. All slides were counterstained with hematoxylin and visualized under a DM4000B microscope (Leica, Germany). Data were collected from three fields of view per mouse and semiquantitatively analysed with ImageJ software. Quantitative results of iron staining were finally normalized to the NN control group.

Flow cytometry

1) The steps of reticulocyte ratio detection were as follows: The experimental tube and negative control tube were prepared, 125 μL normal saline, 4 μL anticoagulant and 125 μL TO working solution (1 μg/mL, Sigma, 390062, USA) were added to each tube, and 250 μL normal saline and 4 μL anticoagulant were added to each tube of the negative control tube. After incubation at room temperature for 1 h, flow cytometry analysis was carried out by using the FL1 (488 nm/525 nm) channel.

2) The steps for the determination of intracellular divalent iron content and lipid peroxidation level were as follows: The spleen was made into a single cell suspension, and then the RBCs were lysed. This step can be omitted from cell samples. A total of 1 × 10⁶ cells were incubated with 100 μL of BioTracker Far-red Labile Fe²⁺ Dye (1 mM, Sigma, SCT037, USA) for 1 h or C11-Bodipy 581/591 (10 μM, Thermo Fisher, D3861, USA) for 30 min. After the cells were washed with PBS twice, flow cytometry analysis was carried out by using the FL6 (638 nm/660 nm) channel for determination of intracellular divalent iron content or the FL1 (488 nm/525 nm) channel for determination of lipid peroxidation level.

3) The steps for detecting the mortality of spleen cells and peritoneal macrophages were as follows: 1 × 10⁶ cells were incubated at room temperature for 40 min with 2 μM Calcein AM and 8 μM PI. Flow cytometry analysis was carried out by using FL1 (488 nm/525 nm, Calcein AM) and FL3 (488 nm/620 nm, PI) channels.

4) The steps for detecting the number of M1/M2 macrophages in the spleen were as follows: 1 × 10⁶ cells were incubated with 2% mouse serum-PBS for 10 min, incubated with F4/80-PE (1:200), CD86-PE/Cyanine7 (1:20, Biolegend, 105014, USA), and CD206-Alexa 647 (1:80, BD, 565250) for 30 min, and washed with 2% mouse serum-PBS twice. Flow cytometry analysis was carried out by using FL2 (488 nm/575 nm, F4/80-PE), FL5 (488 nm/755 nm, CD86-PE/Cyanine7) and FL6 (638 nm/660 nm, CD206-Alexa 647) channels.

5) The steps for detecting the number of monocytes in blood, spleen and bone marrow were as follows: 1 × 10⁶ cells were incubated with 2% mouse serum-PBS for 10 min, incubated with F4/80-PE, CD11b-PE/CY7 (1:2000, BD, 552850, USA), and Ly6C-APC (1:2000, Thermo Fisher, 17-5932-82) for 30 min, and washed with 2% mouse serum-PBS twice. Flow cytometry analysis was carried out by using FL2 (488 nm/575 nm, F4/80-PE), FL5 (488 nm/755 nm, CD11b-PE/CY7) and FL6 (638 nm/660 nm, Ly6C-APC) channels.

Phagocytosis of E. coli and RBCs

1. E. coli was labelled with Cy5.5 (5 mg/mL), and RBCs were labelled with NHS-biotin (20 mg/mL). Macrophages (1 × 10⁶) were coincubated with E. coli-Cy5.5 or RBC- Biotin for 30 min and washed with 2% mouse serum-PBS twice. Flow cytometry analysis was carried out by using FL6 (638 nm/660 nm) to determine the phagocytosis of spleen by E. coli. Macrophages (coincubated with Biotin-RBCs) were incubated with streptavidin-FITC for 3 h and washed twice with 2% mouse serum PBS. FL1 (488 nm/525 nm) was used for flow cytometry analysis to determine the phagocytosis of RBCs in the spleen.

Single-cell RNA Sequencing

The spleen tissues were surgically removed and stored in MACS Tissue Storage Solution (Miltenyi Biotec, Bergisch Gladbach, Germany) until processing. Single-cell dissociation was performed by the experimentalists at the GENECHEM laboratory (Shanghai, China). Dissociated single cells were then stained for viability assessment using Calcein-AM (BD Biosciences, USA) and Draq7 (BD Biosciences). The BD Rhapsody system was used to capture transcriptomic information from single cells. Single-cell capture was achieved by random distribution of a single-cell suspension across >200,000 microwells through a limited dilution approach. Beads with oligonucleotide barcodes were added to saturation so that a bead was paired with a cell in a microwell. The cells were lysed in the microwell to hybridize mRNA molecules to barcoded capture oligos on the beads. Beads were collected into a single tube for reverse transcription and ExoI digestion. Upon cDNA synthesis, each cDNA molecule was tagged on the 5′ end (that is, the 3′ end of an mRNA transcript) with a unique molecular identifier (UMI) and cell barcode indicating its cell of origin. Whole transcriptome libraries were prepared using the BD Rhapsody single-cell whole-transcriptome amplification (WTA) workflow, including random priming and extension (RPE), RPE amplification PCR and WTA index PCR. The libraries were quantified using a High Sensitivity D1000 ScreenTape (Agilent) and High Sensitivity D1000 Reagents (Agilent) on a 4150 TapeStation System (Agilent, Palo Alto, CA, USA) and the Qubit High Sensitivity DNA assay (Thermo Fisher Scientific). Sequencing was performed on an Illumina sequencer (Illumina Nova Seq 6000, San Diego, CA) in a 150 bp paired-end run.

Statistical analysis

Statistical analysis was performed with GraphPad Prism 8.0 software. Two-tailed Student’s t test or one-way analysis of variance (ANOVA) with Tukey’s post hoc test. P < 0.05 was considered statistically significant.

Results

HH exposure promotes erythrocytosis in mice

To mimic 6000 m HA exposure, we placed C58BL/6 mice in an animal hypobaric oxygen chamber and detected the blood indices in the blood of the mice after HH exposure for different times. The blood smear showed that the number of RBCs was increased from 3 to 14 days after HH exposure compared with the normobaric normoxia (NN) group (Figure 1A). Routine blood tests further confirmed the results in blood smears, which showed that the RBC number (Figure 1B), HGB content (Figure 1C) and HCV value (Figure 1D) were all increased significantly to varying degrees, while MCH was not changed after HH exposure (Figure 1E). We further performed flow cytometry by TO staining to detect reticulocytes after 7 and 14 days of HH treatment. The results showed that at both HH 7 and 14 d of exposure, the reticulocyte proportion was increased (Figure 1F). In addition, the number of CD47⁺ RBCs in the blood was also significantly increased at both 7 (Figure 1G and H) and 14 days (Figure 1G and I) after HH exposure. These results suggested that the HH-treated mouse model mimics HA exposure well, which promotes erythropoiesis and results in erythrocytosis in mice following HH exposure.

Fig. 1

Spleen inhibits the immoderate increase in RBCs under HH conditions

To determine the roles of the spleen in RBC homeostasis under HA/HH, we investigated the effects of HH on the morphology, volume and weight of the spleen as well as erythrocyte indices. As shown in Figure 2, the spleen volume and weight were decreased significantly after HH exposure for 1 day compared to NN treatment (Figure 2A-C). However, the spleen was obviously enlarged from 2 to 14 days after HH exposure (Figure 2A-C). The results indicated that the spleen contracted, the stored RBCs in the spleen were released into the blood at 1 day, and the RBCs were produced and/or retained in the spleen from 2 to 14 days after HH exposure. In addition, we also investigated the effect of the spleen on RBC homeostasis under HH conditions and observed whether its clearance effect on RBCs in HH was compensated by the liver or other mononuclear macrophage systems. We removed the spleen from mice using splenectomies and then performed HH exposure for 14 days to detect the RBC counts and blood deposition. The results showed that compared with the splenectomy group mice under NN conditions or the sham group mice exposed to HH, erythrocyte deposition (Figure 2D) and counts (Figure 2E), HGB (Figure 2F) and HCV (Figure 2G) levels were all significantly increased 14 days after HH splenectomy, while MCH did not change (Figure 2H). However, these indices were not changed in the mice with or without splenectomies under NN conditions (Figure 2D and E). It is suggested that the liver and other mononuclear macrophages did not play an effective compensatory role under HH exposure conditions. In addition, the spleen is essential to maintain the homeostasis of RBCs after HH exposure, which can inhibit the excessive proliferation of RBCs after HH exposure.

Fig. 2

HH exposure causes a decrease in the number of macrophages in spleen

Since macrophages are the main cell population for processing RBCs in the spleen under physiological conditions, we next investigated the counts and activity of macrophages after 7 or 14 days of HH exposure via single-cell sequencing and flow cytometry. Single-cell sequencing of the spleen showed that splenic macrophages were significantly decreased after 7 days of HH exposure (Figure 3A-C). Moreover, the flow cytometry of Calcein/PI double staining showed that the living cells of spleen were significantly decreased while the dead cells were significantly increased after HH exposure for 7 and 14 days (Figure 3D-F). We further found that both CD16 (M1 macrophage marker) and CD206 (M2 macrophage marker) expression were decreased at both 7 and 14 days of HH exposure (Figure 3G-L). The flow cytometry results were also consistent with the Western blot results, which showed that mature macrophages (F4/80⁺/CD11b⁺), including M1-type (F4/80⁺/CD86⁺) and M2-type (F4/80⁺/CD206⁺) macrophages, in the spleens of mice decreased after HH exposure for 14 days (Figure 3M). Next, to identify the migration and differentiation of monocytes from bone marrow to spleen, we further analysed the expression of chemokines CCL2, CCL7, Csf1 and Csf2 in spleen by qPCR and detected the number of monocytes in bone marrow and spleen by flow cytometry. The results showed that CCL2 and CCL7 (Figure 3N) and Csf1 and Csf2 (Figure 3O) expression were significantly decreased in the spleen after HH exposure at 7 and 14 days. Moreover, monocytes (Ly6C⁺/CD11b⁺) in the bone marrow (Figure 3P-R) and spleen (Figure 3P and S-T) also declined after HH for 7 and 14 days. Then, we assessed macrophage depletion in the spleen under HH via detection of the expression and distribution of HO-1 and F4/80. Figure 3U shows that both HO-1 and F4/80 were distributed in the red pulp of the spleen, and their expression was also decreased after HH exposure. These results suggested that the number of splenic macrophages decreased after HH exposure, resulting in the impairment of the processing ability of splenic erythrocytes.

Fig. 3

HH exposure decreases erythrophagocytosis and the iron processing capacity of the spleen

We further investigated the effects of HH exposure on erythrocyte phagocytosis and heme iron recycling in splenic macrophages. We used Cy5.5 dye to label E. coli (Figure 4A) and incubated them with splenic cells from HH-exposed mice in vitro, and the phagocytosis of splenic cells was detected using flow cytometry (Figure 4B). Figure 4B shows that the phagocytic ability of splenic cells in vitro was significantly decreased after 7 days of HH exposure. In addition, we used biotin to label autologous RBCs and injected them into mice followed by HH treatment in vivo (Figure 4C and D), and the phagocytosis of splenic cells was detected using flow cytometry. At 7 and 14 days after HH, the number of biotin-labelled RBCs in blood and spleen cells decreased significantly (Figure 4E-H). Although the labelled RBC decay declined automatically with time, the degree of RBC decay in the HH exposure group was weaker than that in the NN group in both the blood (Figure 4E and G) and spleen (Figure 4F and H). Compared with blood, the phagocytic ability of mouse spleen macrophages to RBCs decreased more significantly after HH exposure, especially on the 14th day (Figure 4G and H). Furthermore, we injected Tuftsin, which can stimulate phagocytosis of macrophages, and performed immunofluorescence and iron staining to study the heme iron recycling ability of the spleen under HH conditions (Figure 4I). As shown in Figure 4J, compared with the NN group, F4/80 expression was decreased significantly as well as Fe³⁺ deposition in the red pulp after HH exposure. However, Tuftsin induced F4/80 expression and Fe³⁺ deposition in the red pulp of the spleen under HH (Figure 4K and L). These results suggested that the erythrocyte phagocytosis and heme iron recycling ability of the spleen were suppressed, which was mainly due to the decrease in macrophages in red pulp under HH conditions.

Fig. 4

HH exposure promotes iron mobilization and induces ferroptosis in the spleen

To clarify the precise mechanisms of macrophage reduction caused by HH, we examined the expression of iron metabolism-and ferroptosis-related proteins in mice treated with HH and analysed the expression of the corresponding genes in peripheral blood mononuclear cells (PBMCs) of healthy people acutely exposed to an HA environment from GEO data (No. GSE46480). The results showed that compared to the NN group, the expression levels of HO-1, Ft-L, Ft-H, NCOA4, and xCT were decreased, while the expression levels of Fpn, TfR and ACSL4 were significantly increased both at 7 and 14 days following HH exposure (Figure 5A-B, D-E; Figure S1, A-D). Except for NCOA4, the expression changes in iron metabolism and ferroptosis genes in PBMCs were consistent with our WB results (Figure 5C and F). There were no significant changes in the expression of the GPX4 gene and protein in human PBMCs and mouse spleens. The changes in iron metabolism-and ferroptosis-related genes in PBMCs can fully reflect the protein changes in the spleen under HH exposure. The levels of Fe²⁺ and lipid ROS in the spleen were detected by flow cytometry, and the contents of MDA, GSH and Cys were detected by biochemical detection kits. As shown in Figure 5G and H; Figure S1I-K, the levels of Fe²⁺ (Figure 5G; Figure S1E and F) and lipid ROS (Figure 5H; Figure S1G and H) in the spleen were significantly increased after HH exposure. In addition, the content of MDA (Figure 5I; Figure S1I) in the spleen was increased, while Cys (Figure 5J; Figure S1J) and GSH (Figure 5K; Figure S1K) were decreased significantly after HH exposure. All these results suggested that iron mobilization in the spleen was enhanced and ferroptosis was then induced after 7 days of HH exposure.

Fig. 5

Hypoxia promotes ferroptosis of primary splenic macrophages

To study whether hypoxia exposure can induce ferroptosis in macrophages, changes in Fe²⁺ and lipid ROS levels, cell viability, phagocytosis and ferroptosis-related protein expression in primary splenic macrophages were detected under hypoxia in the presence of a ferroptosis inhibitor (Fer-1). As shown in Figure 6, compared with the normoxia control group (Nor), the levels of Fe²⁺ (Figure 6A-C) and lipid ROS (Figure 6D-E) were significantly increased, and the number of viable cells (Figure 6G-H) and phagocytosis ability (Figure 6I) were also significantly increased after 24 h of hypoxia (Hyp) treatment. In addition, with prolonged hypoxia time, the expression of ACSL4 gradually increased, while the expression of xCT and GPX4 decreased (Figure 6F). Fer-1 inhibited the increase in Fe²⁺ content (Figure 6J-K) and reversed the expression changes in ACSL4, xCT and GPX4 caused by hypoxia exposure (Figure 6L-O). At the same time, Fer-1 reversed the levels of MDA, Cys and GSH induced by hypoxia (Figure 6P-R). These results confirmed that hypoxia induced ferroptosis in primary splenic macrophages.

Fig. 6

Discussion

The purpose of this study was to explore the effects and mechanisms of HA/HH exposure on erythrophagocytosis and iron circulation in mouse spleens. Here, we used an HH chamber to simulate 6000 m HA exposure and found that HH exposure induced iron mobilization and activated ferroptosis in the spleen, especially in red pulp macrophages (RPMs), which subsequently inhibited the phagocytosis and clearance of RBCs. Finally, chronic exposure to HA/HH may promote the retention of RBCs in the spleen, cause splenomegaly, advance RBC production, and promote the occurrence and development of HAPC (Figure 7).

Fig. 7

RBCs/erythrocytes are the carriers of oxygen and the most abundant cell type in the body. Erythrocytes are rapidly increased by triggering splenic contraction in acute HA/HH exposure [16] and stimulating erythropoiesis in subsequent continuous or chronic HA/HH exposure. Changes in spleen morphology and size are closely related to the spleen’s ability to recover RBCs [17], and studies have shown that the spleen is in a contraction state after short-term exposure to HA/HH, and approximately 40% of the increase in RBCs is due to the contraction of the spleen [16, 18, 19]. In the present study, mice were treated under HH to mimic HA exposure (mainly mimicking the low oxygen partial pressure environment of 6000 m HA) for different times according to other studies [20, 21]. Our results showed that HH exposure not only significantly increased the RBC and HGB contents but also significantly increased CD47⁺ cells in blood. However, short-term (1 d) HH exposure caused spleen contraction and induced splenomegaly after 3 d of HH treatment. These results are consistent with other research results; that is, HH or HA exposure affects spleen morphology and further affects RBC counts and HGB levels [22–24].

HAPC is a common chronic HA disease characterized by excessive proliferation of RBCs caused by HH conditions [25]. Under physiological conditions, RBC homeostasis is maintained by balancing erythropoiesis and erythrocyte clearance [4]. The ability of RBC disposal in the spleen for iron recycling under HA/HH exposure was important to RBC regeneration in bone marrow (BM) [26]. Thus, we hypothesized that erythrophagocytosis and iron recycling in the spleen were altered by HA/HH exposure and further disturbed RBC homeostasis and affected the progression of HAPC. We found that compared with sham group mice, HH significantly increased the contents of RBCs and HGB in the blood of splenectomy mice. This strongly verified that the spleen is a key organism that maintains RBC magnitude within a certain range of physiological statuses under HA/HH conditions.

As originally proposed by Metchnikoff in the 19th century, macrophages, especially RPMs, play a pivotal role in the regulation of RBCs and iron homeostasis [27, 28]. We detected the macrophage population in the spleen and found that HH exposure for 7 and 14 days reduced the total macrophage and M1-and M2-type macrophage proportions in the spleen. This result is supported by another study, that is, hypoxia reduces M1-type macrophages in human gastric cancer [29]. We next observed whether monocyte migration and differentiation from the BM to the spleen replenish reduced macrophages and sustain macrophage homeostasis after HH treatment. It is well known that splenic erythropoiesis is a dynamic process [4]. Upon phagocytosing erythrocytes in macrophages, spleen tissue (including macrophages and fibroblasts) produces the chemokine C–C motif chemokine ligand 2 and 7 (CCL2 and CCL7) to recruit blood monocytes to the spleen [30, 31]. Unlike most studies in tumours, which mainly suggested that macrophage accumulation in hypoxic areas was derived by enhanced chemokine expression in tumours [32, 33], CCL2, CCL7 and Csf1 expression was decreased in the spleen after HH exposure for 7 and 14 days in our study. Furthermore, we found that HH exposure inhibited Ly6C^hi monocyte migration from the BM to spleen, where these cells subsequently differentiated into RPMs. The combination of decreased splenic RPMs, along with the reduced BM-derived Ly6C^hi monocytes, suggests that the homeostasis of the RPM population of circulating monocytes was also inhibited after the RPM depletion induced by HH exposure.

HO-1 is a major antioxidant and cytoprotective enzyme that catalyzes the degradation of heme to provide heme homeostasis and protect against free heme-induced toxicity [34]. In addition, it was supposed that mice deficient in HO-1 are largely devoid of RPMs and suggested that HO-1 is critical for RPM development and survival after erythrocyte clearance [35]. The reduced HO-1 protein of red pulp in our study also implied decreased macrophage accounts induced by HH exposure, although HO-1 was considered to be upregulated by hypoxia-inducible factor 1 (HIF-1) under hypoxia [36, 37]. Along with the decreased macrophage amount induced by HH, the phagocytic ability of macrophages in the spleen was also inhibited both in vitro and in vivo. Meanwhile, the injection of tufftsin strongly confirmed the enhancement of heme iron processing in the spleen under HH exposure. Nevertheless, our data conflict with the study of Anand et al, which reported that hypoxia causes an increase in phagocytosis by macrophages in a HIF-1α-dependent manner [38]. This may be because they performed intermittent hypoxia, while we used consistent hypoxia in the present study. The central role played by macrophages in supporting erythrocyte homeostasis stems, in part, from digesting the HGB content of clearance RBCs and recycling the iron back to erythroid progenitors for heme synthesis and HGB production [39]. In mammals, the majority of the body iron is in the form of heme, and the largest pool of heme consists of HGB [40]. HGB contains four prosthetic hemoglobins, and each mature RBC is thought to contain approximately 1.2 × 10⁹ heme moieties [39]. As previously mentioned, the majority of iron required to sustain erythropoiesis is derived from recycled RBCs. In general, RPMs are equipped with molecular machinery capable of neutralizing the toxic effects of heme and metabolizing iron [41, 42]. Nevertheless, defects in erythrocytophagy in macrophages probably lead to aberrant iron metabolism, including anemia and iron overload [41]. The release of heme upon processed RBCs constitutes a permanent and considerable threat for iron cytotoxicity in macrophages and may eventually result in a specific form of programmed cell death termed ferroptosis [43], which may cause decreased cell counts [44, 45].

Accordingly, we investigated the iron metabolism status and explored ferroptosis in spleen/macrophages after HH/hypoxia treatment and found that HH exposure prompted iron mobilization, especially in ferritinophagy and lipid peroxidation in the spleen. Interestingly, we found that except for NCOA4, all the gene expression changes in the PBMCs of humans after acute HA exposure were similar to the changes in the spleens of mice after HH exposure for 7 and 14 days. This is probably because NCOA4 in the spleen mediates iron release from Ft-L and facilitates iron reuse, while PBMCs do not possess this function in vivo. These results not only indicated that HH exposure induces spleen ferroptosis but also implied that the gene expression changes in PBMCs under HA/HH may reflect RBC processing functions in the spleen and further indicate the iron metabolism status in the clinic. We further found that the exact mechanism of macrophage ferroptosis induced by HH exposure was caused by increased Fe²⁺ and decreased antioxidative system expression, which finally resulted in lipid peroxidation of macrophages. It has been proposed that heme catabolism by HO-1 protects macrophages from oxidative stress [41]. The enhanced ROS and lipid peroxidation in vivo were also consistent with the decreased HO-1 expression in the spleen. In addition, 1% hypoxia treatment induced ferroptosis in vitro, which was reduced by treatment with ferrostatin-1, a ferroptosis inhibitor. However, our data were inconsistent with the study of Fuhrmann et al., which reported that hypoxia inhibits ferritinophagy and protects against ferroptosis [46]. We hypothesized that the enhanced iron demand for new RBC regeneration in BM caused by HH leads to an increase in NCOA4 and Ft-L protein expression in the spleen. On the other hand, hypoxia inhibited anti-ferroptosis system expression, including reduced Gpx4 expression and increased lipid ROS production. The results were supported by the group of Youssef LA et al, who reported that increased erythrophagocytosis induces ferroptosis in RPMs in a mouse model of transfusion [47]. However, as most studies have reported, ferroptosis is not only characterized by increased lipid ROS but also significantly shrinking mitochondria [48]. We never found shrinking mitochondria in RPMs after HH exposure (data not shown). Gao et al. proposed that mitochondria played a crucial role in cysteine deprivation-induced ferroptosis but not in that induced by inhibiting glutathione peroxidase-4 (GPX4), the most downstream component of the ferroptosis pathway [49]. Interestingly, in our in vivo study, Gpx4 expression was also not changed after HH exposure. It is still not clear whether mitochondria are involved in ferroptosis. Whether HH treatment caused mitochondrial swelling directly and the exact mechanism involved in this process still need to be further investigated.

Based on the above experiments, we first put forward the “spleen theory” of HAPC (Figure 7). It hypothesizes that HH exposure induced spleen ferroptosis, especially RPMs, which further inhibited erythrophagocytosis, RBC clearance, HGB processing and iron recycling. Therefore, chronic exposure to HA/HH promotes the retention of RBCs in the spleen, leading to splenomegaly and the continuous production of RBCs. These findings may be clinically relevant to the pathological conditions of HAPC progression.

Abbreviations

BM: bone marrow; Cys: Cysteine; GSH: Glutathione; Ft-L: Ferritin light chains; Fpn: Ferroportin; Ft: Ferritin; HH: Hypobaric hypoxia; HA: High-altitude; HAPC: High-altitude polycythemia; HO-1: Heme oxygenase-1; HCT: Haematocrit; HGB: Haemoglobin; HIF-1a: Hypoxia inducible factor 1 a; MCH: Mean corpuscular haemoglobin; MDA: Malondialdehyde; NN: Normobaric normoxia; NCOA4: nuclear receptor coactivator 4; PBMCs: peripheral blood mononuclear cells; RBC: Red blood cell; RPM: Red pulp macrophage; TfR1: Transferrin receptor 1; Tf: Transferrin; TfR: Transferrin receptor.

Author contributions

QQL and GHW conceived, organized, and designed the study. YQG supervised the work. WPY, MQL, JD and JL performed the experiments. WPY, GW and BL contributed to the analysis of data. QQL and GHW prepared, wrote and revised the manuscript. All authors contributed to read, and approved the submitted version.

Funding

This work was supported by Natural Science Foundation of China (Grants 32271228, 82171190, and 81873924), Open Cooperation Program from Key Laboratory of Extreme Environmental Medicine, Ministry of Education (KL2019GY011), China Postdoctoral Science Foundation (2020M673649), High-level Innovation and Entrepreneurship Talents Introduction Program of Jiangsu Province of China, and Nantong Municipal Science and Technology Project (MS12021020 and MS22021010).

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All animal care and experimental protocols were carried out according to the Chinese Animal Management Rules of the Ministry of Health and were authorized by the Animal Ethics Committees of Nantong University research program protocol #S20190219-011. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

Consent for publication

Not applicable.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.

Author details

¹Department of Physiology and Hypoxic Biomedicine, Institute of Special Environmental Medicine and Co-innovation Center of Neuroregeneration, Nantong University, 9 Seyuan Road, Chongchuan District, Nantong, Jiangsu 226019, China. ²College of High-Altitude Military Medicine, Institute of Medicine and Hygienic Equipment for High Altitude Region, Army Medical University, Chongqing 400038, China. ³Key Laboratory of Extreme Environmental Medicine and High-Altitude Medicine, Ministry of Education of China, Chongqing 400038, China. ⁴Department of Neurosurgery, Southwest Hospital, Army Medical University, Chongqing, Chongqing 400038, China.