Erythrocytes retain hypoxic adenosine response for faster acclimatization upon re-ascent AbstractFaster acclimatization to high altitude upon re-ascent is seen in humans; however, the molecular basis for this enhanced adaptive response is unknown. We report that in healthy lowlanders, plasma adenosine levels are rapidly induced by initial ascent to high altitude and achieved even higher levels upon re-ascent, a feature that is positively associated with quicker acclimatization. Erythrocyte equilibrative nucleoside transporter 1 (eENT1) levels are reduced in humans at high altitude and in mice under hypoxia. eENT1 deletion allows rapid accumulation of plasma adenosine to counteract hypoxic tissue damage in mice. Adenosine signalling via erythrocyte ADORA2B induces PKA phosphorylation, ubiquitination and proteasomal degradation of eENT1. Reduced eENT1 resulting from initial hypoxia is maintained upon re-ascent in humans or re-exposure to hypoxia in mice and accounts for erythrocyte hypoxic memory and faster acclimatization. Our findings suggest that targeting identified purinergic-signalling network would enhance the hypoxia adenosine response to counteract hypoxia-induced maladaptation. IntroductionInsufficient tissue oxygen (O2) availability (hypoxia) is a common and dangerous physiological and pathological condition. To survive hypoxia, our body undergoes a number of adaptive responses to promote O2 delivery to peripheral tissues to cope with this challenging condition1,2. For more than a century, substantial effort has focused on understanding the integrated physiological response to high altitude in normal individuals, including the hypoxia ventilator response, diuresis, increased cardiac output, improved oxygen-carrying capacity, cerebral blood flow and erythropoiesis3,4. In this way, our body gradually acclimatizes to high altitude with decreased acute mountain sickness (AMS), improved exercise performance and restored cognitive function5. The inability to adjust to high altitude may lead to pulmonary or cerebral oedema, poor cardiovascular function and even death6,7,8. An intriguing and consistent observation is that following descent to lower elevations, humans retain the acclimatization to high altitude and show a much faster acclimatization upon re-ascent for some time5. Surprisingly, the improved and faster acclimatization to high altitude upon re-ascent does not correspond to increased arterial oxygenation and increased erythropoiesis (CaO2 is lower), two common physiological responses associated with the initial adaptive response5, suggesting that other factors are responsible for the facile response upon re-ascent to high altitude.Like normal individuals facing high-altitude hypoxia, patients with cardiovascular diseases, respiratory diseases, haemolytic disorders and certain cancers are confronted with pathological hypoxia, which participates in disease progression, organ damage and failure3,6,7,8,9,10. As with high-altitude hypoxia in normal individuals, these patients are able to trigger an adaptive response to pathological hypoxic conditions to survive. Thus, adaptive responses to hypoxia are common in normal healthy individuals facing high-altitude hypoxia and patients facing pathological hypoxia to counteract tissue hypoxia for survival. It is extremely difficult to dissect out the adaptive response to hypoxia in patients because of complicated factors associated with time course of disease progression, with disease-specific tissue damage and variables including genetic predisposition and environmental factors. Thus, understanding cellular and molecular mechanisms through which altitude acclimatization occurs in normal humans may lead to new insights regarding adaption to hypoxia and identify potential targets to counteract the maladaptive effects of hypoxia.Extracellular adenosine levels are tightly controlled at multiple steps including its generation from ATP by ectonucleotidases (CD39 and CD73), degradation by adenosine deaminase (ADA) and elimination by equilibrative nucleotide transporters (ENTs). For the past 20 years, substantial studies have focused on extracellular adenosine generation under stress or hypoxic conditions and its function via activation of its specific surface receptors on multiple cell types11,12,13. For example, early studies showed that genetic deletion of CD73 or CD39 abolishes acute extracellular accumulation of adenosine and leads to severe hypoxic tissue damage14,15. Once extracellular adenosine is produced, it elicits multiple functions including anti-vascular leakage, anti-inflammation and vasodilation to protect tissue damage under acute hypoxia setting16,17,18. More recent studies have revealed a protective role of extracellular adenosine activating AMP-mediated protein kinase through ADORA2B receptor in the normal erythrocyte to induce 2,3-bisphosphoglycerate (2,3-BPG) production and subsequently promote oxygen delivery to counteract hypoxic tissue damage19. In contrast, due to the mutation of 尾-haemoglobin in sickle cell disease (SCD; HbS), elevated adenosine signalling via ADORA2B-induced production of 2,3-BPG in the SCD erythrocyte becomes detrimental because it triggers deoxygenated HbS, polymerization and eventually sickling, a central pathophysiology of SCD20. Besides SCD, numerous studies showed that sustained accumulated adenosine signalling via ADORA2B receptors contributes to pathophysiology of multiple chronic settings including chronic kidney diseases, pulmonary fibrosis, priapism, preeclampsia and chronic pain20,21,22,23,24,25. However, whether the hypoxia adenosine response is a common and key regulatory mechanism underlying initial acclimatization and subsequent retention during re-ascent remains unclear.Here by combining human high-altitude studies and mouse genetic studies, we discovered that CD73-depedent elevation of plasma adenosine signalling via ADORA2B-mediated protein kinase A (PKA) phosphorylation, ubiquitination and proteasome degradation of erythrocyte ENT1 is a novel feed-forward signalling network underlying initial hypoxic adaptation and retention upon re-exposure. These findings reveal significant new insight to the molecular basis underlying adaptation to physiological and pathological hypoxia and thereby open up novel therapeutic possibilities for the potential consequences of exposure to hypoxia.ResultsPlasma adenosine associates with altitude acclimatizationTo assess the hypoxic adenosine response at high altitude and upon re-ascent, we measured the levels of purinergic components in plasma from human volunteers participating in the AltitudeOmics study5 (Fig. 1a). Unexpectedly, we found that plasma adenosine levels were significantly higher upon re-ascent to 5,260鈥塵 for 1 day, after spending 7 or 21 days at 1,525鈥塵 (Post7/21), than on the first hypoxia exposure for 1 day (ALT1; Fig. 1b). Thus, plasma adenosine levels were induced by initial hypoxia in a time-dependent manner and remained at significantly higher levels upon re-ascent compared with the first hypoxic exposure (ALT1).Figure 1: Elevated circulating purinergic components are associated with acclimatization and subsequent hypoxic adenosine response upon re-ascent.(a) Schematic illustration of Human AltitudeOmics Study. (b) Plasma adenosine levels of samples collected at different time points (n=21, error bar: mean卤s.e.m., *P 0.05, t-test). (c) sCD73 activities of plasma samples collected at different time points (n=21, error bar: mean卤s.e.m., *P 0.05, **P 0.001, t-test). (d) Changes of plasma adenosine was significantly correlated with improved acclimatization judged by negative correlation with changes of AMS-C composite scores (acute mountain sickness; n=21, Pearson correlation r=鈭?.64, P=0.003). (e,f) Plasma adenosine is significantly associated with improved acclimatization at Post7/21 in comparison with ALT1 (e) associated with LLQ-Headache (ordered logistic regression, n=21, P 2.2e鈭?6, FDR 2.2e鈭?6), (f) associated with LLQ-Dizzy (ordered logistic regression, n=21, P 2.2e鈭?6, FDR 2.2e鈭?6). FDR=false discovery rate.Full size imageSoluble CD73 (sCD73) is an important purinergic component of the hypoxic adenosine response in acute hypoxia and functions to generate extracellular adenosine. Therefore, we measured sCD73 activity in normal individuals at sea level (SL), at high altitude and upon re-ascent. Intriguingly, sCD73 activity displayed the same trend as elevated extracellular adenosine, a significant time-dependent elevation at high altitude and a further induction upon re-ascent (Fig. 1c). No significant differences in plasma adenosine and sCD73 activity were seen between males and females upon first ascent to high altitude and following re-ascent (Supplementary Fig. 1a,b). Thus, we provide human in vivo evidence that sCD73 activity was induced during initial exposure to high altitude and retained its high activity upon re-ascent.Our discoveries of increased plasma adenosine and sCD73 activity in response to high altitude and their rapid and even greater elevation on re-ascent raise a possibility that these biomarkers are physiological signals that contribute to rapid physiological acclimatization upon re-ascent. To test this possibility, we evaluated the relationship between the increase in these two biochemical parameters (that is, plasma adenosine level and sCD73 activity) to the improvement of the physiological acclimatization on Post7/21 (day 1 upon re-ascent) relative to ALT1 (day 1 of the initial ascent). Specifically, we performed Pearson Correlation analysis to determine whether the elevation of plasma adenosine and sCD73 on Post7/21 relative to ALT1 were associated with the reduction of AMS-C composite score, which is commonly used to evaluate physiological adaptation to high altitude. We observed a significant correlation between the increase in plasma adenosine levels and the decrease in AMS-C composite score on Post7/21 compared with ALT1 (Pearson correlation r=鈭?.64, P=0.003; Fig. 1d). However, we did not observe a significant correlation between increased sCD73 activity with decreased AMS-C composite score on Post7/21 compared with ALT1 (Supplementary Fig. 2). These data support our hypothesis that the rapid rise in plasma adenosine upon re-ascent represents a physiological retention for rapid acclimatization upon re-ascent.Because elevated plasma adenosine was associated with reduced AMS-C composite score upon re-ascent, we used ordered logistic regression to further assess which specific physiological outcomes measured in the AMS-C composite score were most significantly associated with plasma adenosine levels on ALT1 and Post7/21, respectively. Among all of the categorical outcomes, including Lake Louise Questionnaire (LLQ)-Fatigue, LLQ-Headache, LLQ-Gastrointestinal, LLQ-Dizzy and LLQ-AMS Score in the AMS-C composite score, we found that LLQ-Headache (ordered logistic regression, P 2.2e鈭?6, false discovery rate (FDR) 2.2e鈭?6) and LLQ-Dizzy (ordered logistic regression, P 2.2e鈭?6, FDR 2.2e鈭?6) were two key physiological parameters measured in AMS-C composite score that significantly associated with plasma adenosine levels on ALT1 and Post7/21, respectively (Fig. 1e,f). Overall, our human studies revealed that increased plasma adenosine levels are correlated to initial acclimatization, and that higher levels of plasma adenosine are correlated to the rapid acclimatization upon re-ascent.eENT1 plays a major role in extracellular adenosine uptakeCirculating adenosine levels are governed by synthesis via CD73, degradation by ADA and cellular uptake by facilitated ENTs. However, the function and regulation of ENTs under acute hypoxia remains undetermined. To examine the role of ENTs in regulating extracellular adenosine, we translated our human finding to in vivo studies with intact animals to determine the role of erythrocytes in the uptake of plasma adenosine. Specifically, we used an in vivo C14-adenosine uptake assay to accurately trace the distribution of adenosine in mice at different time points. We found that injected C14-adenosine was rapidly transported inside erythrocytes, as quickly as 1鈥塵inute (min) (Fig. 2a). In addition, we found that the distribution of C14-adenosine among whole blood, plasma and red blood cells (RBCs) reached a maximum within 1鈥塵in and maintained steady levels for at least 10鈥塵in (Fig. 2a). We chose to compare the distribution of C14-adenosine at 5鈥塵in after injection. We found that major portion of injected C14-adenosine was retained in the circulation or whole blood (68%). Hence, 鈭?/span>30% of total injected adenosine was taken up by peripheral tissues. Intriguingly, 鈭?/span>54% of injected adenosine was taken up by erythrocytes, which is responsible for almost 90% of injected adenosine in whole blood cells. Approximately 10% of injected adenosine was taken up by other non-erythrocyte circulating cells and 5% of total injected adenosine was retained in plasma (Fig. 2a). Thus, these results indicate that erythrocytes are the major cell type responsible for the rapid uptake of extracellular adenosine in vivo.Figure 2: eENT1 is the major factor eliminating extracellular adenosine.(a) In vivo adenosine uptake assay, comparison is between plasma and blood, n=4, P 0.05. (b) In vitro adenosine uptake assay, *Comparison between RBC from ENT1鈭?鈭?/i> and WT, n鈮?, P 0.001; #comparison between dipyridamole or NBMPR and DMSO, n=4, P 0.001. (c) In vitro adenosine uptake assay using human erythrocytes, n=4, *P 0.05. (d) Plasma adenosine from hypoxia-treated Ent1flox/flox/EpoRCre or EpoRCre mice. *: between normoxia and 48-h hypoxia, n=4, P 0.05. **between 24- and 72-h hypoxia, n=4, P 0.05. +between normoxia and 24-, 48- or 72-h hypoxia, n=4, P 0.05. ++between EpoRCre and Ent1flox/flox/EpoRCre after 48- or 72-h hypoxia, n=4, P 0.05. (e) Reduced tissue hypoxia was observed in Ent1flox/flox/EpoRCre compared with EpoRCre after 72-h hypoxia treatment by HypoxiaProbe, red: HypoxiaProbe, blue: DAPI; scale bar, 200鈥壩糾. (f鈥?b>h) Deletion of eENT1 protects from acute lung injury indicated by (f) total cell number (n=4, *P 0.05), (g) albumin (n=4, *P 0.05) and (h) IL-6, (n=4, *P 0.05) in BALF from 72-h hypoxia-treated animals. (i) Pulmonary neutrophil accumulation by MPO assay in lung tissues from 72-h hypoxia-treated animals (n鈮?, *P 0.05). (j) H E-stained lung section from 72-h hypoxia-treated animals (edema, red arrows); scale bar, 500鈥壩糾. All error bars: mean卤s.d. and t-test.Full size imageNext, to investigate which ENT on erythrocytes is the predominant adenosine transporter, we carried out in vitro adenosine uptake assays using erythrocytes isolated from wild-type (WT C57BL/6) mice and normal human subjects in the presence or absence of a specific ENT1 inhibitor, nitrobenzylmercaptopurine (NBMPR), or an ENT general inhibitor, dipyridamole. NBMPR inhibits 95% of ENT1 activity at 0.1鈥壩糓 and has no effects on other ENTs at this concentration26,27. NBMPR and dipyridamole treatment significantly inhibited eENT activity in mouse and human compared with untreated cells (Fig. 2b,c). Moreover, there was no significant difference between the NBMPR- and dipyridamole-treated groups (Fig. 2b,c). Consistent with these pharmacological studies, adenosine uptake in the erythrocytes isolated from global ENT1-deficient mice (Ent1鈭?鈭?/i>) was almost completely abolished (Fig. 2b). In contrast, erythrocytes isolated from global Ent2鈭?鈭?/i> mice maintained similar ENT transporter activity as normal mouse erythrocytes (Fig. 2b). These results indicate that ENT1 on the erythrocytes is the major ENT responsible for uptake of extracellular adenosine in mice and humans.eENT1 regulates extracellular adenosine under acute hypoxiaTo evaluate the importance of erythrocyte equilibrative nucleoside transporter 1 (eENT1) in the uptake of adenosine in vivo, we generated erythrocyte-specific ENT1 and ENT2 knockouts by crossing Ent1flox/flox or Ent2flox/flox with EpoRCre-GFP (EpoRCre) mice (Supplementary Fig. 3a)28. Western blot analysis showed that most of the ENT1 or ENT2 protein was absent from the erythrocytes of Ent1flox/flox/EpoRCre-GFP (Ent1flox/flox/EpoRCre) or Ent2flox/flox/EpoRCre-GFP (Ent2flox/flox/EpoRCre) mice, respectively, but not in other tissues, such as the brain (Supplementary Fig. 3b). Moreover, we demonstrated that ENT transport activity was significantly decreased in the erythrocytes isolated from Ent1flox/flox/EpoRCre mice compared with EpoRCre mice, but not the erythrocytes from Ent2flox/flox/EpoRCre mice (Supplementary Fig. 3c). These results indicate that we successfully deleted eENT1 or eENT2 in mice.To determine the importance of eENT1 in the acute hypoxic adenosine response, we exposed EpoRCre (controls) and Ent1flox/flox/EpoRCre mice to hypoxia (8% oxygen, equivalent to an altitude of 7,500鈥塵) for up to 72鈥塰. Similar to human high-altitude studies, plasma adenosine levels were induced by hypoxia in both EpoRCre control and Ent1flox/flox/EpoRCre, while it increased much faster in Ent1flox/flox/EpoRCre than EpoRCre mice (Fig. 2d). Of note, we found that plasma adenosine levels were induced to higher levels in mice than in humans under hypoxia (Figs 1b and 2d). This difference is likely due to the mice being exposed to 8% normobaric hypoxia, equivalent to an altitude of 7,500鈥塵, which is higher than humans who were exposed to 5,260鈥塵 equivalent to 10% hypobaric hypoxia. These studies indicate that eENT1 plays an important role in the rapid elevation of extracellular adenosine under acute hypoxia in mice.Next, we assessed tissue hypoxia using HypoxiaProbe (pimonidazole, forming protein adducts in hypoxic cells when pO2 10鈥塵m鈥塇g) in mice29. A significant decrease in tissue hypoxia in the heart, kidney and lung was observed in hypoxia-exposed Ent1flox/flox/EpoRCre mice compared with EpoRCre control mice, indicating that rapid induction of extracellular adenosine is likely beneficial to prevent tissue hypoxia (Fig. 2e images; statistical results in Supplementary Fig. 4a鈥揷). Because lung damage is known to be induced by acute hypoxia30, we chose to further quantify pulmonary vascular leakage and inflammation in those mice. We found significantly reduced total cell numbers, albumin and interleukin (IL)-6 concentration in collected bronchoalveolar lavage fluid (BALF) in Ent1flox/flox/EpoRCre than EpoRCre mice under hypoxia (Fig. 2f鈥揾). Moreover, pulmonary myeloperoxidase (MPO) activity, a specific enzyme for primary neutrophils31 was significantly less induced in hypoxia-challenged Ent1flox/flox/EpoRCre compared with EpoRCre mice (Fig. 2i). Finally, haematoxylin and eosin (H E) staining showed a significant increase in perivascular fluid accumulation (oedema) in hypoxia-treated EpoRCre mice compared with Ent1flox/flox/EpoRCre (Fig. 2j, red arrows; Supplementary Fig. 4d). Our findings indicate that the loss of eENT1 leads to quicker and higher accumulation of extracellular adenosine under acute hypoxia resulting in reduced hypoxia-induced tissue inflammation and damage.Hypoxia induces eENT1 ubiquitination and degradationTo our surprise, we observed that extracellular adenosine levels in control mice eventually accumulated to levels similar to that of the eENT1-deficient mice after 72-h hypoxia (Fig. 2d). These findings raise an intriguing possibility that 72-h hypoxia exposure likely reduces eENT1 activity and thus allows the accumulation of extracellular adenosine in the control mice to a similar level as in eENT1-deficient mice. To test this possibility, we measured eENT activity in control mice. We found that under hypoxia, eENT activity was significantly reduced at 48鈥塰 and further reduced at 72鈥塰 (Fig. 3a). We also compared sCD73 activity and plasma ATP levels between controls and erythrocyte-specific ENT1 knockouts under normoxia and 72-h hypoxia. We found that sCD73 activity and plasma ATP levels were induced by hypoxia compared with normoxia in both EpoRCre and Ent1flox/flox/EpoRCre mice, with no significant difference in hypoxia-induced sCD73 activity and plasma ATP levels between them (Fig. 3b,c). Thus, in response to hypoxia, there is a coordinated increase in circulating sCD73 activity, plasma ATP levels and reduced ENT activity, which work together to promote a rapid rise in plasma adenosine.Figure 3: Hypoxia downregulates eENT1 through adenosine鈥揂DORA2B鈥搖biquitin signalling.(a) Hypoxia decreased eENT1 activity in erythrocytes from hypoxia-treated animals as judged by in vitro adenosine uptake assay, n鈮?, *comparison between normoxia and 48-h hypoxia-treated animals, P 0.05; **comparison between normoxia and 72-h hypoxia-treated animals, P 0.001, error bar: mean卤s.e.m. (b) sCD73 activity in Ent1flox/flox/EpoRCre or EpoRCre mice with or without hypoxia treatment, n=4, error bar: mean卤s.d., P 0.001, t-test. (c) Plasma ATP in Ent1flox/flox/EpoRCre or EpoRCre mice with or without hypoxia treatment, n=4, error bar: mean卤s.e.m., P 0.05. (d) Erythrocyte ENT1 is reduced during hypoxia treatment through ubiquitination by western blot against ENT1 (left panel, 尾-actin was used as loading control, n=3, error bar: mean卤s.d., +P 0.05). Right panel, IP ubiquitin followed by western blot against ENT1, *non-specific bands. (e) Hypoxia-mediated downregulation of eENT1 activity (left panel) and eENT1 protein (immunofluorescence against ENT1, red colour: ENT1, right panel; scale bar, 5鈥壩糾) were observed in C57BL/6 mice, but not CD73鈭?鈭?/i> or AODRA2B鈭?鈭?/i> mice when we treated these animals under hypoxia (8% oxygen) for 72鈥塰, n鈮?, error bar: mean卤s.d., P 0.05). (f) Adenosine signalling through ADORA2B downregulates eENT1 activity as judged by in vitro adenosine uptake assay (n=3, error bar: mean卤s.d., *P 0.05). t-test is used for all comparisons in this figure.Full size imageBecause erythrocytes have no nuclei, we speculated that reduced ENT1 transporter activity in response to hypoxia is likely due to reduced protein levels. We measured membrane ENT1 protein levels in the erythrocytes from EpoRCre mice at normoxia and under hypoxia for up to 72鈥塰. Intriguingly, as with ENT activity, membrane-anchored ENT1 protein levels in the erythrocytes were reduced at 48鈥塰 and further reduced after 72-h hypoxia challenge (Fig. 3d, left panel; Supplementary Fig. 5). To further determine whether hypoxia-mediated reduced membrane eENT1 is through ubiquitin鈥損roteasomal degradation, we performed immunoprecipitation (IP) followed with western blot analysis. Specifically, we pulled down ubiquitinated protein from erythrocyte lysates from normoxia- or hypoxia-treated animals using an antibody specific for ubiquitin, followed by western blot analysis against ENT1. We found that polyubiquitinated ENT1 ((Ub)n-ENT1)) levels in lysates were substantially elevated after 48-h hypoxia compared to normoxia and were degraded after 72-h hypoxia exposure (Fig. 3d, right panel; Supplementary Fig. 6). These results indicate that hypoxia-mediated reduced membrane eENT1 is due to ubiquitination-dependent proteasome degradation in vivo.Adenosine induces eENT1 degradation under acute hypoxiaGiven the fact that adenosine is a potent hypoxic-signalling sensor, we speculated that elevated sCD73-induced production of extracellular adenosine may function as a positively feed-forward loop to promote ubiquitination of eENT1, subsequent accumulation of extracellular adenosine to protect hypoxic tissue damages. To test this possibility, we took a genetic approach to expose WT and Cd73鈭?鈭?/i> mice to normoxia and hypoxia up to 72鈥塰. Consistently, we found that hypoxia-induced plasma adenosine levels in WT mice but not Cd73鈭?鈭?/i> mice19. Moreover, we found that eENT activity and membrane-anchored eENT1 levels were significantly reduced by hypoxia in WT mice but not in Cd73鈭?鈭?/i> mice (Fig. 3e). These findings provide in vivo mouse genetic evidence that sCD73-dependent elevation of plasma adenosine underlies hypoxia-mediated downregulation of membrane-anchored eENT1 levels and thus reduced eENT1 activity.Extracellular adenosine elicits multiple cellular functions by the activation of four adenosine receptors12,13. To determine whether adenosine signalling via its specific receptors underlies reduction of eENT1, we compared eENT1 activity in primary cultured erythrocytes (mature and enucleated erythrocytes) isolated from WT and four adenosine receptor-deficient mice in response to 5鈥?N-ethylcarboxamidoadenosine (NECA) treatment, a non-metabolized adenosine analogue32. We found that NECA directly and significantly reduced eENT1 activity in WT mouse erythrocytes (Fig. 3f). However, genetic deletion of Adora2b but not Adora1, Adora2a or Adora3 attenuated NECA-mediated reduction of ENT activity (Fig. 3f). Similar to CD73鈭?鈭?/i> mice, we confirmed our in vitro studies in vivo by showing that hypoxia-mediated reduction of membrane-anchored eENT1 and ENT activity were abolished in ADORA2B-deficient mice (Fig. 3e). Here we provided in vivo and in vitro genetic evidence that CD73-dependent production of plasma adenosine signalling via ADORA2B underlies hypoxia-mediated reduction of ENT activity by decreased membrane-anchored ENT1 levels in erythrocytes.ADORA2B is a Gs-coupled receptor and PKA activation is reportedly involved in ubiquitination33. Thus, it is possible that NECA-mediated reduction of eENT1 transporter activity and protein level is dependent on PKA-mediated phosphorylation and subsequent ubiquitination. To test this possibility, we first pulled down phosphorylated PKA substrate from erythrocyte lysates from normoxia- or hypoxia-treated animals using an antibody specific for phospho-PKA substrate, followed by western blot against ENT1 (Fig. 4a; Supplementary Fig. 7). We found that hypoxia treatment increased ENT1 phosphorylation by PKA and reached a peak at 48鈥塰, then decreased after 72-h hypoxia treatment that was the same trend as ubiquitinated ENT1 (Fig. 3d, right panel). Next, we treated primary cultured WT mouse erythrocytes with or without Bay60-6583, a specific ADORA2B agonist34, or forskolin, a specific PKA activator35, in the presence or absence of two structurally independent inhibitors of the 20S proteasome, the proteasome found in erythrocytes36, MG132 (ref. 37) or Bortezomib38, or a PKA inhibitor, H-89 (ref. 39). We found that H-89, MG132 and Bortezomib significantly attenuated Bay60-6583-mediated decrease in eENT1 activity in a dosage-dependent manner, respectively (Fig. 4b). Moreover, forskolin alone directly downregulated eENT1 activity and this downregulation was also attenuated by H-89, MG132 and Bortezomib in a dose-dependent manner (Fig. 4c). Finally, we also monitored ENT1 localization and ubiquitinated proteins following Bay60-6583 or forskolin treatment using immunofluorescence (IF) confocal image analysis. Both Bay60-6583 and forskolin treatments induced ubiquitination (green fluorescence) and subsequently ENT1 translocation from membrane to cytosol (red fluorescence), and finally degradation in a time-dependent manner. These events were blocked by Bortezomib or H-89 treatment (Fig. 4d). Notably, Bortezomib blocked eENT1 degradation but not translocation and H-89 blocked both ENT1 translocation and degradation, indicating that PKA-mediated phosphorylation is required for ADOAR2B-induced ubiquitination and subsequent proteasome degradation of ENT1 (Fig. 4d). Taken together, we provide both genetic and pharmacological in vivo and in vitro evidence that PKA functions downstream of ADOAR2B underlying adenosine-mediated reduction of eENT1 activity by inducing its phosphorylation, ubiquitination and proteasome degradation.Figure 4: ADORA2B鈥揚KA鈥搖biquitin鈥損roteasomal degradation of eENT1 in mice.(a) IP鈥搘estern blot using phospho-PKA subtract antibody pull-down proteins from RBC lysate followed by western blot against ENT1, 尾-actin as an input control, n=3, *compared with normoxia, P 0.05, t-test. (b,c) In vitro adenosine uptake assay. Activation of ADORA2B by Bay 60-6583 (B60) or activation of PKA by forskolin decreases eENT1 activity that can be blocked by H-89, MG132 or BT (Bortezomib). (b) By B60, n=3, *comparison between DMSO and B60 treatment, P 0.05; + and ++comparison between B60 and B60 plus inhibitor, +P 0.05, ++P 0.001. (c) By forskolin, n=3, *comparison between DMSO and forskolin treatment, P 0.05; + and ++comparison between forskolin and forskolin plus inhibitor, +P 0.05, ++P 0.001. (d) Immunofluorescence, activation of ADORA2B or PKA leads to eENT1 ubiquitination and degradation. B60 and forskolin induced protein ubiquitination in a time-dependent manner and the effect of B60 can be blocked by H-89 (middle panels). B60 and forskolin induced eENT1 translocation to the cytosol after 3-h treatment and degradation after 6-h treatment. H-89 blocked B60-induced translocation and degradation of eENT1. BT blocked B60- and forskolin-induced degradation of eENT1, but not the translocation (upper panels). Red: eENT1, green: ubiquitin, scale bar: 2鈥壩糾. Error bars: mean卤s.d., t-test.Full size imagePKA mediates eENT1 degradation in response to high altitudeTo validate our mouse studies in humans, we measured PKA phosphorylated ENT1, ubiquitinated ENT1 and ENT1 protein levels in erythrocytes isolated from the human volunteers participating in the AltitudeOmics study. We found that PKA-mediated phosphorylated ENT1 started to rise on ALT1, reached the peak on ALT7, began to decrease and maintained at similar but still elevated levels on ALT16 and on Post7 upon re-ascent compared with SL (Fig. 5a; Supplementary Fig. 8). Similar to the trend of PKA-mediated ENT1 phosphorylation, the polyubiquitination of ENT1 ((Ub)n-ENT1)) in the cell lysate started to rise on ALT1, reached the peak on ALT7, began to decrease and maintained at similar but still elevated levels on ALT16 and on Post7 upon re-ascent compared with SL (Fig. 5b; Supplementary Fig. 9). Supporting the mouse finding of PKA-mediated phosphorylation-dependent ubiquitination and proteasomal degradation of ENT1, we found that membrane ENT1 protein levels were significantly reduced at ALT7, further reduced on ALT16 and remained a similar low level upon re-ascent at Post7 (Fig. 5c; Supplementary Fig. 10). These findings provide strong human in vivo evidence that high-altitude hypoxia leads to PKA-mediated phosphorylation and ubiquitination-dependent proteasome degradation of eENT1, and that reduced eENT1 membrane protein is retained upon re-ascent. As such, we observed a much more rapid elevation of extracellular adenosine upon re-ascent and concurrent retention of physiological acclimatization (Fig. 1).Figure 5: PKA phosphorylation-mediated ubiquitination and porteasomal degradation of eENT1 in humans at high altitude and upon re-ascent.(a) IP phospho-PKA substrate followed by western blot against ENT1, (IP鈥搘estern blot is representative image from three experiments using three sets of RBC samples from the Human AltitudeOmics Study), P-ENT1: PKA phosphorylated ENT1. (b) IP ubiquitin followed by western blot against ENT1, n=3, (IP鈥搘estern blot was performed using RBC samples from three subjects), *: compared with SL, error bars: mean卤s.d., P 0.05, t-test. (c) Hypoxia-induced downregulation of eENT1 protein levels in erythrocyte samples judged by western blot against ENT1, n=3 (RBC samples from three subjects), *compared with SL, error bars: mean卤s.d., P 0.05, t-test.Full size imageHypoxic adenosine response is through eENT1 downregulationHere we revealed that ENT1 protein in human erythrocytes is reduced by initial hypoxia-induced proteasome degradation and remained low upon re-ascent to high altitude. Because erythrocytes contain no nuclei and do not carry out translation, they lack the ability to make new protein. Thus, our human in vivo high-altitude studies and mouse genetic studies support a novel but compelling hypothesis that erythrocytes, as the most abundant cells in our body, retain a hypoxic adenosine response for a certain time period until hypoxia-exposed RBCs are replaced by newly synthesized RBCs containing higher levels of ENT1 protein. To test this intriguing possiblity, we mimicked human high-altitude studies by exposing WT mice to hypoxia, bringing them to normoxia and re-exposing them to hypoxia (Fig. 6a). First, to trace ENT1 protein levels and activity in erythrocytes from first hypoxia to second hypoxia, we in vivo labelled erythrocytes at 24鈥塰 before first hypoxia by N-hydroxysuccinimide (NHS) biotin20. Because we found that 72-h hypoxia is long enough to induce eENT1 proteasomal degradation and reduce its activity in WT mice but not to significantly increase in erythropoiesis, we exposed WT mice to hypoxia for 72鈥塰 (8% oxygen balanced with nitrogen; Fig. 3a,d). After the first hypoxic challenge, mice were returned to normoxia (21% oxygen) for 3 days up to 50 days because the life span of mouse erythrocytes is 55 days. At the end of each time point, some of the mice were re-exposed to hypoxia for 24鈥塰 (at Post3 or Post50) and then killed for further analysis (Fig. 6a, illustration).Figure 6: Erythrocyte has hypoxic adenosine response through hypoxia-mediated downregulation of eENT1.(a) Experimental illustration for acclimatization hypoxic adenosine response. (b) Erythrocyte life span was indicated by biotin NHS labelling and revealed by flow cytometry. (c) eENT1 changes during hypoxia treatment as judged by immunofluorescence, red: ENT1, green: Biotin; scale bar, 5鈥壩糾. (d) eENT1 activity (n=4, *P 0.05). (e) Plasma adenosine levels. n=4 for all groups, *between normoxia and first hypoxia, P 0.05; +between Post3 normoxia and Post3 re-hypoxia, P 0.05; ++between Post50 normoxia and Post50 re-hypoxia, P 0.05; **between first hypoxia and Post3 re-hypoxia, P 0.05. Error bars: mean卤s.d., t-test.Full size imageAs expected, flow cytometry analysis of biotinated erythrocytes showed that they started to decrease on Post3 and almost completely disappeared by Post50, indicating that old labelled erythrocytes were replaced by newly synthesized erythrocytes by Post50 (Fig. 6b). Follow-up co-immunoflurosence staining of eENT1 and biotin showed that membrane-anchored eENT1 staining was significantly reduced after first hypoxic challenge (72-h hypoxic challenge; Fig. 6c, upper panel). Moreover, eENT1 membrane protein was still reduced to the similar level at 72鈥塰 following first hypoxia in the biotinated erythrocytes after mice were brought back to normoxia (Post3 normoxia group) and followed by 1 day exposure to second hypoxia (Post3 re-hypoxia group; Fig. 6c). In contrast, eENT1 protein levels almost returned to the normal levels in the non-biotinylated and newly synthesized erythrocytes in mice after 50 days of normoxia (Post50 normoxia) and followed by 1 day exposure to second hypoxia, indicating that old erythrocytes had been replaced by newly synthesized erythrocytes and thus eENT1 expression levels were back to normal levels (Fig. 6c, Post50 re-hypoxia). Functionally, we validated that the eENT activities also followed the same trend as eENT1 protein levels (Fig. 6d). Similar to human high-altitude studies, plasma adenosine levels were rapidly induced to significantly higher levels after 1 day second hypoxia following 3 days at normoxia (Fig. 6e, Post3). In contrast, the faster and higher increases of plasma adenosine during second hypoxia challenge went away when the mice were brought back to normoxia for 50 days and re-exposed to hypoxia (Fig. 6e, Post50). Taken together, our mouse findings recapitulate human high altitude observations and provide strong evidence that erythrocytes retain a hypoxic adenosine response through downregulation of eENT1 protein and activity until hypoxia-challenged erythrocytes are replaced by newly synthesized erythrocytes.DiscussionHere we report both human in vivo high-altitude studies and mouse genetic hypoxic findings that eENT1 is a key purinergic-signalling component responsible for hypoxia-induced plasma adenosine. Mechanistically, we revealed that increased plasma adenosine subsequently signals via ADORA2B, inducing PKA-dependent phosphorylation, ubiquitination and proteasomal degradation of eENT1 protein mediated by the initial hypoxia challenge and in turn results in decreased uptake of extracellular adenosine and accumulation of plasma adenosine. The reduced eENT1 resulting from the initial hypoxic exposure establishes an erythrocyte hypoxic adenosine response for a second hypoxic exposure, and underlies a faster and improved acclimatization upon re-ascent associated with high levels of circulating adenosine (Fig. 7).Figure 7: Working model.Our findings support a working model that elevated sCD73-mediated increased adenosine signalling via ADORA2B results in PKA-mediated phosphorylation, ubiquitination and proteasomal degradation of ENT1 in the erythrocytes of humans at high altitude and mice exposed to hypoxia. Thus, purinergic-signalling components play an important role in enhancing accumulation of plasma adenosine and thus promote initial adaptation to hypoxia in a feed-forward manner. Moreover, these newly identified erythrocyte purinergic components retain hypoxic adenosine response to promote quicker and higher elevation of plasma adenosine and thereby allow for rapid acclimatization upon re-ascent. Our findings offer multiple innovative therapies to counteract hypoxic challenge under physiological and pathological conditions.Full size imageAscending to high altitude is challenging due to limited oxygen supply. Consequently, an acute and integrative physiological response to protect against hypoxic tissue damage is essential to allow adaptation and survival. Faster and improved acclimatization is a consistent and compelling observation seen in humans upon re-ascent. Although elevation of adenosine is widely considered beneficial under acute hypoxic settings, changes in plasma adenosine during high-altitude acclimatization and upon re-ascent were unrecognized until we measured circulating purinergic components in healthy human lowland volunteers involved in the AltitudeOmics study5. Here we demonstrated that sCD73 activity and plasma adenosine levels were concurrently induced by initial exposure to high altitude and rapidly accumulated to much higher levels upon re-ascent following 7 days at a much lower elevation. Functionally, plasma adenosine levels were positively associated with initial altitude acclimatization and subsequent rapid acclimatization upon re-ascent.The erythrocyte is the most abundant circulating cell type and it has been speculated that erythrocytes regulate extracellular adenosine uptake40. The ENT (SLC29) family contains four members, namely, ENT1 through ENT4 that are widely expressed in mammalian tissue with different expression levels41. Early studies demonstrated that the inhibition of ENT1 is beneficial for acute liver and kidney damage induced by ischaemia and reperfusion in experimental models42. However, the function of eENT1 in response to hypoxia could not be determined without our newly developed genetically manipulated mice with erythrocyte-specific ablation of ENT1 or ENT2. These mice represent important genetic tools used to demonstrate that ENT1 but not ENT2 on the erythrocyte plays a key role in adenosine uptake. The ablation of eENT1 but not ENT2 in mice results in much faster and higher accumulation of plasma adenosine under acute hypoxia. We further demonstrated that genetic deletion of eENT1 results in the rapid accumulation of extracellular adenosine and is beneficial to protect against acute hypoxia-induced tissue damage including vascular leakage and inflammation in the lung tissues compared with the control mice. Taken together, we have provided both human and mouse evidence that eENT1 is critical to uptake exogenous extracellular adenosine and that reduced eENT1 is a previously unrecognized regulatory mechanism to enhance adenosine hypoxia response to counteract tissue hypoxia.Multiple molecules including hypoxia-inducible factor-1伪 and nitric oxide are associated with physiological and pathological hypoxic responses43,44. Of note, hypoxia-inducible factor-1伪 regulates transcription of multiple purinergic-signalling components including CD73, ADOAR2B, ADORA2A and ENTs45,46,47. Recent studies showed that alveolar ADORA2B crosstalks with ENT2 to dampen acute lung injury48. However, erythrocytes have no nuclei and therefore regulation of eENT1 is likely independent of transcription and translation. Here we demonstrated that erythrocyte ENT1 is regulated by ubiquitination-dependent proteasomal degradation machinery in humans under high altitude and mice under hypoxia. Thus, the hypoxia-mediated degradation of eENT1 lowered plasma adenosine uptake and promoted the accumulation of extracellular adenosine. To our surprise, we found that elevated sCD73-induced production of plasma adenosine signalling via ADOAR2B-induced ubiquitination and degradation of ENT1 in both human and mouse erythrocytes. Moreover, we revealed that PKA-mediated phosphorylation of ENT1 is required for ADOAR2B-mediated ubiquitination and degradation of ENT1 in both human and mouse erythrocytes. Among four adenosine receptors, ADORA2B has a lowest affinity and it is usually activated by adenosine at micromolar concentration49. Thus, erythrocyte ADORA2B-mediated PKA phosphorylation and ubiquitination of ENT1 do not occur until adenosine is elevated to the micromolar range following 48鈥塰 of hypoxia in mice and day 7 at high altitude in humans. As such, CD73-ADORA2B-mediated erythrocyte proteasomal degradation of ENT1 functions synergistically as a feed-forward purinergic-signalling component to adapt to hypoxia by promoting adenosine production and reducing its elimination and thus enhancing accumulation of adenosine in the plasma. Moreover, besides erythrocytes, multiple-nucleated cells including endothelial cells also uptake extracellular adenosine50. Our finding about adenosine signalling via ADORA2B reducing eENT1 activity is likely not limited to erythrocytes and may also regulate ENT1 activity and extracellular adenosine in other cell types. This notion is strongly supported by a recent study showing that ADORA2A modulates drug efflux transporter, P-glycoprotein in the blood鈥揵rain barrier by ubiquitination51. Thus, our current findings add a new function for erythrocytes in the hypoxic adenosine response and reveal novel mechanisms regulate purinergic-signalling components.Because erythrocytes do not have nuclei, hypoxia-induced proteasomal degradation of ENT1 in erythrocytes exposed to hypoxia cannot be replenished by new protein synthesis until new erythrocytes are generated. Moreover, we found that sCD73 activity was also further induced upon re-ascent to high altitude. Thus, the initial hypoxia-mediated elevation of sCD73 activity in the circulation and the proteosomal degradation of eENT1 work collaboratively to retain hypoxic adenosine response for acclimatization to high altitude resulting in a higher and quicker elevation of plasma adenosine and faster adaptation to second hypoxia. Extending from these findings, we further revealed that eENT1 contributes to initial adaption and retains a hypoxic adenosine response upon re-ascent until newly synthesized RBCs containing high ENT1 levels replace old RBCs with reduced ENT1-levels mediated by initial hypoxia. Thus, erythrocyte hypoxic purinergic memory is transient and is gradually lost when the newly synthesize erythrocytes replace the old ones. These findings immediately suggest that targeting CD73, ENT1 and ADORA2B will likely enhance the adenosine hypoxic response and prolong this response in normal individuals ascending to high altitude for quicker acclimatization, in athletes with strenuous exercise to improve performance for gold medals and patients frequently facing hypoxia for better adaptation.In conclusion, very little was known about the function and regulation of the erythrocyte hypoxic adenosine response before our current studies. Here we provide both human and mouse genetic evidence that CD73-depedent elevation of plasma adenosine signalling via ADORA2B-mediated phosphorylation, ubiquitination and degradation of ENT1 in the erythrocytes is a novel feed-forward signalling network underlying initial hypoxic adaptation and retention upon re-exposure to enhance the hypoxic adenosine response (Fig. 7). These findings add significant new insight to our understanding of molecular mechanisms of adaptation to hypoxia under physiological and pathological conditions and thereby open up novel therapeutic avenues to treat and prevent hypoxia.MethodsHuman studyThe current study was a part of the human AltitudeOmics project examining the integrative physiology of human responses to high-altitude hypoxia5. In brief, all procedures conformed to the Declaration of Helsinki and are approved by the Universities of Colorado and Oregon Institutional Review Boards and the US Department of Defense Human Research Protection office. All participants are healthy (mean卤s.d. age, 21卤1 years; stature, 1.78卤0.10鈥塵; body mass, 69卤11鈥塳g), non-smokers, cardiorespiratory disease free, born and raised at 1,500鈥塵, and had not travelled to elevations 1,000鈥塵 in 3 months before the investigation52. Blood samples from each subject were drawn near SL at Eugene, OR, USA (baseline, 130鈥塵, Barometric pressure =749鈥塵m鈥塇g) and also on the first (ALT1), seventh (ALT7) and sixteenth (ALT16) day at Mt Chacaltaya, Bolivia (5,260鈥塵, mean Barometric pressure =406鈥塵m鈥塇g) and upon reasceding to high altitude for 1 day after descending to low altitude at 1,525鈥塵 for one (Post7, n=14) or 3 weeks (Post21, n=7; Fig. 1a). The detail experimental protocols were described previously5.AMS was accessed by Environmental Symptom Questionnaire (AMS-C) and LLQ5. AMS-C composite score is a self-reported 11-question inventory and indicative of AMS was considered when a score 鈮?.7 (ref. 5). LLQ is a self-reported assessment of AMS symptoms.Mouse strainsAll animal protocols were in accordance with the University of Texas, Medical School at Houston guidelines. C57BL/6 mice were purchased from Jackson Lab. Ent1鈭?鈭?/i>, Ent2鈭?鈭?/i>, Ent1flox/flox and Ent2flox/flox were obtained from Dr. Eltzschig鈥檚 lab, University of Colorado. Ent1flox/floxEpoRCre+ and Ent2flox/floxEpoRCre+ mice were generated by cross Ent1flox/flox, Ent2flox/flox mice with EpoRCre-GFP mice28 and the genotype of these are Ent1flox/flox+/+EpoRCre+ and Ent2flox/flox+/+EpoRCre+. Adora1鈭?鈭?/i>, Adora2a鈭?鈭?/i>, Adora2b鈭?鈭?/i> and Adora3鈭?鈭?/i> have been previously described20. Eight to 12 weeks, age- and sex-matched (including both male and female) C57BL/6 and mutant mice were used for experiments.ReagentsNECA (catalogue #: 1691), Bay60-6583 (catalogue #: 4472) and 2鈥?deoxycoformycin (DCF; catalogue #: 2033) were from TOCRIS Bioscience, Bristol, UK. Dipyridamole (catalogue #: D9766), adenosine (catalogue #: A9251) and biotin NHS (catalogue #: H1759) were from Sigma-Aldrich, St. Louis, MO, USA. Forskolin (catalogue #: 3828) and dihydrochloride (H-89, catalogue #: #9844) were from Cell Signalling, Danvers, MA, USA. MG132 (catalogue #: sc-201270) and Bortezomib (catalogue #: sc-217785) were from Santa Cruz Biotechnology, Dallas, TX, USA. C14-adenosine (Moravek Biochemicals, Inc. Brea, CA; #MC-1499; specific activity: 280鈥塵Ci鈥塵mol鈭?) and Percoll (GE Healthcare Life Sciences, catalog#: 17-0891-01, Piscataway, NJ, USA). ENT1 antibodies were from Santa Cruz Biotechnology (catalogue #: sc-377283. 200鈥塶g鈥塵l鈭? of antibody was used for western blot, and 1鈥壩糶鈥塵l鈭? of antibody was used for IP) and Abcam (catalogue #: ab135756, Cambridge, MA, USA; 250鈥塶g鈥塵l鈭? of antibody was used for western blot). ENT2 antibody was from Abcam (catalogue #: ab48595). Phospho-PKA substrate (RRXS*/T*) antibody used for IP and western blot was from Cell Signalling (catalogue #: 9624, the antibody was diluted at 1:1,000 in LI-COR blocking buffer with 0.1% Tween 20 for western blot). Ubiquitin antibodies used for western blot, IP and IF were from Santa Cruz Biotechnology (catalogue #: sc-9133 and sc-8017, for western blot: 200鈥塶g鈥塵l鈭?; for IP: 1鈥壩糶鈥塵l鈭?; for IF: 1鈥壩糶鈥塵l鈭?). HypoxiaProbe antibody used for IF was from HypoxiaProbe (catalogue #: 2627, 1:100 ratio of rabbit antisera was used for IF).Plasma adenosine extraction and measurementPlasma sampling, adenosine extraction and adenosine measurement were performed as described previously20,52.For human plasma sampling, whole blood samples were collected following 30鈥塵in of rest from a catheter placed in an antecubital vein. Blood samples were spun for 20鈥塵in at 800鈥?i>g at room temperature. After separation, plasma samples were stored on ice for 10鈥塵in and then stored at 鈭?0鈥壜癈 for further analysis52.For mouse plasma sampling, mice were anesthetized with 2.5% avertin and blood was collected, in a collection tube with EDTA anticoagulant containing 1鈥壩糽 of 10鈥塵mol鈥塴鈭? dipyridamole and 1鈥壩糽 of 10鈥塵mol鈥塴鈭? ADA inhibitor deoxycoformycin (DCF), centrifuged at 2,400鈥?i>g for 5鈥塵in to isolate plasma, plasma samples were stored at 鈭?0鈥壜癈 before adenosine extraction and analysis.For plasma adenosine extraction, 500鈥壩糽 of plasma was added to 500鈥壩糽 of 0.6鈥塵ol鈥塴鈭? cold perchloric acid on ice and vortexed. The homogenate was centrifuged at 20,000鈥?i>g for 10鈥塵in at 4鈥壜癈. The supernatant (568鈥壩糽) was transferred to a new tube and neutralized with 40.9鈥壩糽 of 3鈥塵ol鈥塴鈭? KHCO3 in 3.6鈥塏 KOH. Phenol red (2鈥壩糽 of 0.2鈥塵g鈥塵l鈭?) was added as pH indicator. The sample was acidified with 5.7鈥壩糽 of 1.8鈥塵ol鈥塴鈭? ammonium dihydrogen phosphate (pH 5.1) and 13.2鈥壩糽 phosphoric acid (30%). Finally, the sample was centrifuged at 20,000鈥?i>g for 5鈥塵in and the supernatant was transferred to a new tube and stored at 鈭?0鈥壜癈. Before high-performance liquid chromatography (HPLC) assay, the sample was thawed on ice, and then centrifuged at 20,000鈥?i>g for 10鈥塵in. The supernatant was transferred to a new tube for HPLC analysis.For plasma adenosine measurement, plasma concentration was measured using HPLC. Two hundred microlitre of adenosine sample was loaded in the HPLC and the flow rate was set at 1.5鈥塵l鈥塵in鈭?. The representative peaks were identified and quantitated by running external adenosine standard curves53.In vivo adenosine uptake assayThe following protocol was approved by the Animal Welfare Committee of the University of Texas, Medical School at Houston. Mice (10鈥?2 weeks old) were anesthetized and 100鈥壩糽 C14-Adenosine master mix (20鈥壩糾ol鈥塴鈭? adenosine, 20鈥壩糃i C14-adenosine and 200鈥壩糾ol鈥塴鈭? DCF, in 0.9% saline) was injected through tail vein to achieve a final concentration of 鈭?/span>1鈥壩糾ol鈥塴鈭? adenosine, 0.1鈥壩糃i C14-adenosine, and 10鈥壩糾ol鈥塴鈭? DCF in the circulation system. After injection, 60鈥壩糽 of blood was withdrawn through the right ventricle at different time points and immediately transferred to a 1.5-ml Eppendorf tube with heparin anticoagulant containing 1鈥壩糽 of stop cocktail (10鈥塵mol鈥塴鈭? DCF and 10鈥塵mol鈥塴鈭? dipyridamole), stored on ice. Twenty microlitre of whole blood was taken and lysed in 60鈥壩糽 water and the lysate was spread on a glass microfiber filter (GE Healthcare Life Sciences, catalogue number: 1825-025), heat dried for counting of C14 isotope using a scintillation counter (LKB WALLAC 1209 EACKBETA Liquid Scintillation Counter, LKB Instruments, Victoria, Australia). The rest of the blood was centrifuged at 2,400鈥?i>g for 5鈥塵in at 4鈥壜癈. Twenty microlitre of plasma was spread on a glass microfiber filter for C14 isotope counting. After all plasma was removed, 20鈥壩糽 of RBC was transferred into a new 1.5-ml Eppendorf tube with 80鈥壩糽 0.9% saline with 10鈥壩糾ol鈥塴鈭? dipyridamole and 10鈥壩糾ol鈥塴鈭? DCF; then the mixture was loaded on the top of 70 over 75% Percoll gradient, centrifuged at 2,400鈥?i>g for 5鈥塵in at 4鈥壜癈, the supernatant was then carefully and sequentially removed. Finally, the RBCs were lysed with 60鈥壩糽 water for C14 isotope counting as mentioned previously.In vitro adenosine uptake assay using RBCsBlood was collected with heparin as an anticoagulant and centrifuged at 2,400鈥?i>g for 5鈥塵in at room temperature, followed by aspiration of plasma and buffy coat. Packed RBCs were then washed three times with pre-wormed culture medium at 25鈥壜癈 and resuspended to 4% haematocrit using same medium. One mililitre of RBC suspension was added to the wells of 12-well plates and cultured for 1鈥塰 at 25鈥壜癈. The uptake assay started with transferring 54鈥壩糽 of RBC suspension to an Eppendorf tube with 6鈥壩糽 C14-adenosine master mix (10鈥壩糾ol鈥塴鈭? adenosine with 1鈥壩糃i ml鈭? C14-adenosine in PBS) to get a final adenosine concentration of 1鈥壩糾ol鈥塴鈭?. The uptake was performed for 30鈥塻 and stopped by adding 100鈥壩糽 cold stop solution (0.9% saline with 10鈥壩糾ol鈥塴鈭? dipyridamole and 10鈥壩糾ol鈥塴鈭? DCF). The stopped reaction was loaded on the top of 70 over 75% percoll gradients, centrifuged at 2,400鈥?i>g for 5鈥塵in. The supernatant was withdrawn and the RBC pellet was processed for scintillation counting as mentioned above. Also, 54鈥壩糽 of RBC suspension (washed and resuspended to 4% haematocrit as mentioned above) was aliquoted for total protein measurement using Pierce BCA Protein Assay kit (Thermo Scientific, catalogue #: 23225, Rockford, IL, USA).RBC membrane protein extraction and western blotBlood was collected with EDTA as anticoagulant. Two hundred microlitre of blood was loaded on the top of 2鈥塵l cold 70% over 75% percoll gradient and centrifuged at 2,400鈥?i>g for 5鈥塵in at 4掳鈥塁. The RBC pellet was washed three times with cold PBS, lysed with 540鈥壩糽 cold H2O with proteinase and phosphatase inhibitors, and incubated on ice for 10鈥塵in with occasional vortexing. A volume of 60鈥壩糽 of 10 脳 PBS was added and the rest of lysate was spun at 20,000鈥?i>g for 20鈥塵in at 4鈥壜癈. The pellet was washed three times with cold PBS and dissolved in 200鈥壩糽 RIPA buffer with proteinase inhibitors (cOmplete Protease inhibitor cocktail tablets, Roche, catalogue #: 11697498001, Mannheim, Germany) and phosphatase (PhosSTOP Phosphatase inhibitor cocktail tablets, Roche, catalogue #: 04 906 837 001, Mannheim, Germany) inhibitors, and stored at 鈭?0掳鈥塁 for further analysis. Protein concentration was determined using the Pierce BCA Protein Assay kit. Protein samples were boiled in Laemml Sample Buffer (Bio-Rad, catalogue #: 161-0737, Hercules, CA) for 10鈥塵in at 95鈥壜癈, separated using 10% home-made SDS鈥揚AGE gel. Then, proteins were transferred to a nitrocellulose membrane. Western blot was performed and scanned using Odyssey Imaging System (LI-COR, Lincoln, NE) following the guidelines from the manufacture. All primary antibodies were used at 1:1,000 ratio or as indicated in Odyssey Blocking Buffer (LI-COR, catalogue #: 927-40000) with 0.2% Tween 20 and incubated at 4掳鈥塁 overnight. The secondary antibodies were obtained from LI-COR (IRDye 800CW Donkey anti-Rabbit IgG (H+L), catalogue #: 926-32213; IRDye 680RD Donkey anti-Mouse IgG (H+L), catalogue #: 926-68072 and all secondary antibodies were used at 100鈥塶g鈥塵l鈭? in LI-COR blocking buffer supplied with 0.2% Tween 20.In vivo hypoxic challengeHypoxia treatmentHypoxic challenge was performed using hypoxia chamber (BioSpherix) with continued monitoring of oxygen concentration (BioSpherix, ProIx-110). Oxygen concentration was 8% for all hypoxic challenge that are 鈭?/span>7,500鈥塵 altitude equivalent (normobaric hypoxia).RBC IFTen microlitre fresh blood was fixed in 1鈥塵l cold methanol for 15鈥塵in on ice. Cells were then washed two times with cold PBS and stored in PBS at 4掳鈥塁. Fixed cells were blocked in 3% bovine serum albumin (BSA) in PBS for 1鈥塰 at room temperature and then incubated in primary antibody (anti-ENT1, Santa Cruz Biotechnology, 1鈥壩糶鈥塵l鈭? in 3% BSA-PBS) at 4掳鈥塁 overnight. After washing three times with PBS, cells were incubated in secondary antibody (Alexa 594 Donkey anti-mouse, 2鈥壩糶鈥塵l鈭? in BSA-PBS) for 1鈥塰 at room temperature. After washing, cells were spread on slides and counterstained with ProLong Gold Antifade Reagent with 4,6-diamidino-2-phenylindole (Cell Signalling, catalogue #: 8961). Images were acquired using a 脳 63 oil immersion objective (numerical aperture 1.4) of a Leica TCS SP5 confocal microscope (Leica Microsystems, Wetzlar, Germany).Assess tissue hypoxia status by hypoxia probeTo assess tissue hypoxia status, hypoxia probe-1 (Hypoxiaprobe, Inc, Burlington, MA) 50鈥塵g鈥塳g鈭? weight was injected to mice at 30鈥塵in before euthanizing. Tissues were isolated and fixed in 10% paraformaldehyde in PBS for at least 24鈥塰. Fixed tissues were rinsed in PBS, dehydrated through graded ethanol washes and embedded in paraffin. Sections (5鈥壩糾) were rehydrated, permeabilized with 0.5% Triton x-100 in PBS for 20鈥塵in, blocked in 3% BSA in PBS and incubated in anti-hypoxia probe antibody (raised in rabbit, Hypoxiaprobe, Inc, Burlington, MA, 1:100 of 2,627 rabbit antisera in 3% BSA-PBS) at 4掳鈥塁 overnight. Goat anti-rabbit IgG (H+L) Alexa Fluor 594-conjugate antibody was used as a secondary antibody at 2鈥壩糶鈥塵l鈭? for 1鈥塰 at room temperature. After washing, the slide was counterstained with ProLong Gold Antifade Reagent with 4,6-diamidino-2-phenylindole.BALF collection and analysisFor BALF collection, before sample collection, mice were anesthetized with 2.5% avertin and blood was collected. Then, lungs were lavaged four times with 0.3鈥塵l PBS, 鈭?/span>1鈥塵l fluid was collected. The BALF was centrifuged and the supernatants were stored at 鈭?0鈥壜癈 for further analyses53.For BALF total cell number counting, after BALF was collected, 40鈥壩糽 BALF was aliquoted for total cell number counting using a hemacytometer under light microscopy.BALF albumin was measured using Albuwell M kit (Exocell, Philadelphia, PA, USA) under the protocol suggested by the manufacturer.For BALF interleukin (IL)-6 measurement, BALF IL-6 measurement was performed using Mouse IL-6 ELISA Set (BD Biosciences, San Diego, CA, USA) and according to the manufacturer鈥檚 instructions. In brief, the plates were coated overnight with the 100鈥壩糽 capture antibody diluted in coating buffer at 4鈥壜癈 and washed three times with 300鈥壩糽 washing buffer. The wells then were blocked with 200鈥壩糽 assay diluent for 1鈥塰 at room temperature. After the plates were washed three times, 100鈥壩糽 of BALF sample or standard was added and incubated for 2鈥塰 at room temperature. Then, the plates were aspirated and washed five times. One hundred microlitre working detector (detection antibody and Streptavidin-HRP) was added to each well and incubated 1鈥塰 at room temperature, and then washed another seven times. A volume of 100鈥壩糽 of substrate solution was added and incubated for 30鈥塵in at room temperature in the dark. At last, 50鈥壩糽 of stop solution was added and optical density was read at 450鈥塶m within 30鈥塵in. IL-6 concentration was calculated based on a standard curve.Lung MPO assayPulmonary neutrophil sequestration was quantified using MPO assay as described previously54. In brief, animals were killed and lungs were perfused with 10鈥塵l of PBS through the right ventricle. Lungs were snap-frozen in liquid nitrogen and stored at 鈭?0鈥壜癈. Lung was homogenized in MPO assay buffer and spun at 10,000鈥?i>g for 10鈥塵in at 4鈥壜癈. The protein concentration of the supernatant was determined using the BAC assay kit and the supernatant was used for MPO assay using the Myeloperoxidase Colorimetric Activity Assay kit (Sigma-Aldrich, #MAK068) following the manufacturer鈥檚 directions.Lung histology and quantification of perivascular oedemaMice were killed. Lung tissue was isolated and fixed in 10% paraformaldehyde in PBS for at least 24鈥塰. Fixed lung tissue was rinsed in PBS, dehydrated through graded ethanol washes and embedded in paraffin. Paraffin-embedded lung was sectioned (5鈥壩糾). Sections were rehydrated and stained with Hematoxylin and Eosin (H E staining).For the quantification of perivascular oedema, the quantification was performed as described previously55. Perivascular oedema was defined as the fluid accumulation in the perivascular space (Fig. 2j, arrows). For the quantification of perivascular oedema, five to eight pictures of arteries from each lung section of left caudal, right caudal and right median lobes were captured at 脳 10 magnification on H E-stained slides. For each perivascular oedema, the distance between adventitia and external elastic lamina was determined by measuring and averaging four measurements along the cross lines through the centre of the oedema. Then, the distance in arbitrary unit was converted to 渭m according to the scale bar.Plasma CD73 enzyme activity measurementCD73 enzyme activity was measured by quantifying the conversion of etheno-AMP (E-AMP) to ethenoadenosine (E-ADO) as described previously56. In brief, 20鈥壩糽 plasma sample (collection method is described above) was incubated at room temperature with 200鈥塶M DCF (an inhibitor of ADA) in buffer (0.1鈥塎 HEPES (pH 7.4) and 50鈥壩糓 MgCl) with or without 伪,尾-methylene ADP (APCP, Sigma-Aldrich, a specific inhibitor of CD73). Next, reactions were incubated at 37鈥壜癈 for 30鈥塵in in the presence of 100鈥壩糓 E-AMP. Plasma CD73 activity was determined by the changes of E-ADO concentration between reactions with or without 伪,尾-methylene ADP. E-ADO concentration was measurement using reversed phase HPLC as describe before57,58.IP鈥搘estern blotRBCs were lysed in H2O at 1:10 volume ratio in the presence of protease inhibitor cocktail and proteasome inhibitor (N-ethylmaleimide, at 20鈥塵M), then 10 脳 PBS was added to the lysate to neutralize buffer. IP鈥搘estern blot was performed using 2鈥塵g total RBC lysate in 0.5鈥塵l IP solution (lysate and IP buffer, 50鈥塵M (Tris-HCl, pH 7.5), 150鈥塵M NaCl and 1% triton x-100) along with proteinase inhibitor and 20鈥塵M N-ethylmaleimide). Lysate was pre-cleaned with 50鈥壩糽 Pierce Protein A/G Agarose beads (catalogue #: 20421, ThermoFisher Scientific) at room temperature for 2鈥塰. Ubiquitin (mouse raised, Santa Cruz, catalogue #: sc-8017, 1鈥壩糶鈥塵l鈭? of antibody was used for IP) and phospho-PKA substrate (rabbit raised, Cell Signaling, catalogue #: 9624S, antibody was used at 1:200 ratio), incubation was performed at 4鈥壜癈 overnight. After primary antibody incubation, 50鈥壩糽 Protein G Sepharose beads (catalogue #: 17-0618-01, GE Healthcare Life Sciences, Pittsburgh, PA, USA) was added and incubated at room temperature for 1鈥塰. IP proteins were eluted by boiled in 30鈥壩糽 2 脳 Laemmli Sample Buffer (catalogue #: 1610737, Bio-Rad) for 10鈥塵in and resolved on 10% SDS鈥揚AGE gel for western blot.In vivo RBC biotin NHS labellingRBCs were labelled in vivo using NHS biotin. Specifically, 50鈥塵g鈥塊g鈭? of NHS biotin was injected into the retroorbital plexus of animals (prepared in 100鈥壩糽 sterile saline just before injection; initially dissolved at 50鈥塵g鈥塵l鈭? in N,-N,-dimethylacetamide) Blood samples (5鈥壩糽) were collected after first day of biotin injection from tail vein by venipuncture to determine the percentage of RBCs labelled with biotin. The percentage of biotinylated RBCs was calculated by determining the fraction of peripheral blood cells labelled with Ter-119 (to identify RBCs) that were also labelled with a streptavidin-conjugated fluorochrome by flow cytometry.Acclimatization hypoxic adenosine response experimentAs illustrated in Fig. 6a, 24鈥塰 after NHS biotin injection, mice were treated with hypoxia (8% O2) for 72鈥塰, then were kept at normoxia (21% O2) for either 3 days or 50 days and were given hypoxia (8% O2) treatment again for 24鈥塰. After treatment, blood samples were collected for flow cytometry, measurements of eENT1 activity, IF and plasma adenosine analysis as described previously.Statistical analysisFor the human high-altitude study, we evaluated the relationship between the increase in these two biochemical parameters (that is, plasma adenosine level and sCD73 activity) to the improvement of the physiological acclimatization on Post7/21 (day 1 upon re-ascent) relative to ALT1 (day 1 of the initial ascent). Specifically, we performed Pearson correlation analysis to determine whether the elevation of plasma adenosine and sCD73 on Post7/21 relative to ALT1 was associated with the reduction of AMS-C composite score, which is commonly used to evaluate physiological adaptation to high altitude. Then, we used ordered logistic regression to further assess which specific physiological outcomes measured in the AMS-C composite score were significantly associated with plasma adenosine levels on ALT1 and Post7/21, respectively, including LLQ-Fatigue, LLQ-Headache, LLQ-Gastrointestinal, LLQ-Dizzy and LLQ-AMS Score. Finally, we applied multiple test corrections to control the FDR, and considered FDR 0.05 as significant. All other data were presented as mean卤s.e.m. and analyzed statistically using GraphPad Prism 5 software (GraphPad Software). The significance of differences among two groups was assessed using two-tailed Student鈥檚 t-test. Differences between the means of multiple groups were compared by one-way analysis of variance, followed by a Turkey鈥檚 multiple comparisons test. Comparison of the data obtained at different time points was analyzed by two-way analysis of variance repeated measurements, followed by the Bonferroni post hoc test. A P value of 0.05 was considered significant. Error bars in graphs are presented as mean卤s.d., unless otherwise indicated.Data availabilityAll relevant data presented are available from the authors.Additional informationHow to cite this article: Song, A. et al. Erythrocytes retain hypoxic adenosine response for faster acclimatization upon re-ascent. Nat. Commun. 8, 14108 doi: 10.1038/ncomms14108 (2017).Publisher鈥檚 note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References1Michiels, C. Physiological and pathological responses to hypoxia. Am. J. Pathol. 164, 1875鈥?882 (2004).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 2Sen, C. K. Wound healing essentials: let there be oxygen. Wound Repair Regen. 17, 1鈥?8 (2009).Article聽 PubMed聽 PubMed Central聽Google Scholar聽 3Petousi, N. Robbins, P. A. Human adaptation to the hypoxia of high altitude: the Tibetan paradigm from the pregenomic to the postgenomic era. J. Appl. Physiol. 116, 875鈥?84 (2014).CAS聽 Article聽 PubMed聽Google Scholar聽 4Liang, R. Ghaffari, S. Advances in understanding the mechanisms of erythropoiesis in homeostasis and disease. Br. J. Haematol. 174, 661鈥?73 (2016).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 5Subudhi, A. W. et al. AltitudeOmics: the integrative physiology of human acclimatization to hypobaric hypoxia and its retention upon reascent. PLoS ONE 9, e92191 (2014).ADS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 6Hackett, P. H. Roach, R. C. High-altitude illness. N. Engl. J. Med. 345, 107鈥?14 (2001).CAS聽 Article聽 PubMed聽Google Scholar聽 7Anderson, J. D. Honigman, B. The effect of altitude-induced hypoxia on heart disease: do acute, intermittent, and chronic exposures provide cardioprotection? High Alt. Med. Biol. 12, 45鈥?5 (2011).Article聽 PubMed聽Google Scholar聽 8Bartsch, P. Swenson, E. R. Acute high-altitude illnesses. N. Engl. J. Med. 369, 1666鈥?667 (2013).PubMed聽Google Scholar聽 9Bartsch, P. Gibbs, J. S. Effect of altitude on the heart and the lungs. Circulation 116, 2191鈥?202 (2007).Article聽 PubMed聽Google Scholar聽 10Hackett, P. H. et al. High-altitude pulmonary edema in persons without the right pulmonary artery. N. Engl. J. Med. 302, 1070鈥?073 (1980).CAS聽 Article聽 PubMed聽Google Scholar聽 11Zhang, Y. Xia, Y. Adenosine signaling in normal and sickle erythrocytes and beyond. Microbes Infect. 14, 863鈥?73 (2012).CAS聽 Article聽 PubMed聽Google Scholar聽 12Fredholm, B. B., IJzerman, A. P., Jacobson, K. A., Klotz, K. N. Linden, J. International Union of Pharmacology. XXV. Nomenclature and classification of adenosine receptors. Pharmacol. Rev. 53, 527鈥?52 (2001).CAS聽 PubMed聽Google Scholar聽 13Fredholm, B. B., IJzerman, A. P., Jacobson, K. A., Linden, J. Muller, C. E. International Union of Basic and Clinical Pharmacology. LXXXI. Nomenclature and classification of adenosine receptors--an update. Pharmacol. Rev. 63, 1鈥?4 (2011).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 14Eckle, T. et al. Cardioprotection by ecto-5鈥?nucleotidase (CD73) and A2B adenosine receptors. Circulation 115, 1581鈥?590 (2007).CAS聽 Article聽 PubMed聽Google Scholar聽 15Eltzschig, H. K. et al. Central role of Sp1-regulated CD39 in hypoxia/ischemia protection. Blood 113, 224鈥?32 (2009).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 16Idzko, M., Ferrari, D., Riegel, A. K. Eltzschig, H. K. Extracellular nucleotide and nucleoside signaling in vascular and blood disease. Blood 124, 1029鈥?037 (2014).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 17Ho, M. F. Rose\'Meyer, R. B. Vascular adenosine receptors; potential clinical applications. Curr. Vasc. Pharmacol. 11, 327鈥?37 (2013).CAS聽 Article聽 PubMed聽Google Scholar聽 18Eltzschig, H. K., Rivera-Nieves, J. Colgan, S. P. Targeting the A2B adenosine receptor during gastrointestinal ischemia and inflammation. Expert Opin. Ther. Targets 13, 1267鈥?277 (2009).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 19Liu, H. et al. Beneficial Role of Erythrocyte Adenosine A2B Receptor-Mediated AMP-Activated Protein Kinase Activation in High-Altitude Hypoxia. Circulation 134, 405鈥?21 (2016).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 20Zhang, Y. et al. Detrimental effects of adenosine signaling in sickle cell disease. Nat. Med. 17, 79鈥?6 (2011).CAS聽 Article聽 PubMed聽Google Scholar聽 21Iriyama, T. et al. Elevated placental adenosine signaling contributes to the pathogenesis of preeclampsia. Circulation 131, 730鈥?41 (2015).CAS聽 Article聽 PubMed聽Google Scholar聽 22Dai, Y. et al. A2B adenosine receptor-mediated induction of IL-6 promotes CKD. J. Am. Soc. Nephrol. 22, 890鈥?01 (2011).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 23Karmouty-Quintana, H. et al. Deletion of ADORA2B from myeloid cells dampens lung fibrosis and pulmonary hypertension. FASEB J. 29, 50鈥?0 (2015).CAS聽 Article聽 PubMed聽Google Scholar聽 24Ning, C. et al. Excess adenosine A2B receptor signaling contributes to priapism through HIF-1alpha mediated reduction of PDE5 gene expression. FASEB J. 28, 2725鈥?735 (2014).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 25Hu, X. et al. Sustained elevated adenosine via ADORA2B promotes chronic pain through neuro-immune interaction. Cell Rep. 16, 106鈥?19 (2016).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 26Nagai, K., Nagasawa, K., Kyotani, Y., Hifumi, N. Fujimoto, S. Mouse equilibrative nucleoside transporter 2 (mENT2) transports nucleosides and purine nucleobases differing from human and rat ENT2. Biol. Pharm. Bull. 30, 979鈥?81 (2007).CAS聽 Article聽 PubMed聽Google Scholar聽 27Ward, J. L., Sherali, A., Mo, Z. P. Tse, C. M. Kinetic and pharmacological properties of cloned human equilibrative nucleoside transporters, ENT1 and ENT2, stably expressed in nucleoside transporter-deficient PK15 cells. Ent2 exhibits a low affinity for guanosine and cytidine but a high affinity for inosine. J. Biol. Chem. 275, 8375鈥?381 (2000).CAS聽 Article聽 PubMed聽Google Scholar聽 28Heinrich, A. C., Pelanda, R. Klingmuller, U. A mouse model for visualization and conditional mutations in the erythroid lineage. Blood 104, 659鈥?66 (2004).CAS聽 Article聽 PubMed聽Google Scholar聽 29Arteel, G. E., Thurman, R. G., Yates, J. M. Raleigh, J. A. Evidence that hypoxia markers detect oxygen gradients in liver: pimonidazole and retrograde perfusion of rat liver. Br. J. Cancer 72, 889鈥?95 (1995).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 30Tuder, R. M., Yun, J. H., Bhunia, A. Fijalkowska, I. Hypoxia and chronic lung disease. J. Mol. Med. 85, 1317鈥?324 (2007).Article聽 PubMed聽Google Scholar聽 31Pulli, B. et al. Measuring myeloperoxidase activity in biological samples. PLoS ONE 8, e67976 (2013).ADS聽 CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 32Klotz, K. N. Adenosine receptors and their ligands. Naunyn Schmiedebergs Arch. Pharmacol. 362, 382鈥?91 (2000).CAS聽 Article聽 PubMed聽Google Scholar聽 33Hoang, T. et al. cAMP-dependent protein kinase regulates ubiquitin-proteasome-mediated degradation and subcellular localization of the nuclear receptor coactivator GRIP1. J. Biol. Chem. 279, 49120鈥?9130 (2004).CAS聽 Article聽 PubMed聽Google Scholar聽 34Alnouri, M. W. et al. Selectivity is species-dependent: characterization of standard agonists and antagonists at human, rat, and mouse adenosine receptors. Purinergic. Signal. 11, 389鈥?07 (2015).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 35Pattabiraman, D. R. et al. Activation of PKA leads to mesenchymal-to-epithelial transition and loss of tumor-initiating ability. Science 351, aad3680 (2016).Article聽 PubMed聽 PubMed Central聽Google Scholar聽 36Neelam, S., Kakhniashvili, D. G., Wilkens, S., Levene, S. D. Goodman, S. R. Functional 20S proteasomes in mature human red blood cells. Exp. Biol. Med. 236, 580鈥?91 (2011).CAS聽 Article聽Google Scholar聽 37Han, Y. H., Moon, H. J., You, B. R. Park, W. H. The effect of MG132, a proteasome inhibitor on HeLa cells in relation to cell growth, reactive oxygen species and GSH. Oncol. Rep. 22, 215鈥?21 (2009).CAS聽 PubMed聽Google Scholar聽 38Hirai, M. et al. Bortezomib suppresses function and survival of plasmacytoid dendritic cells by targeting intracellular trafficking of Toll-like receptors and endoplasmic reticulum homeostasis. Blood 117, 500鈥?09 (2011).CAS聽 Article聽 PubMed聽Google Scholar聽 39Lochner, A. Moolman, J. A. The many faces of H89: a review. Cardiovasc. Drug. Rev. 24, 261鈥?74 (2006).CAS聽 Article聽 PubMed聽Google Scholar聽 40Plagemann, P. G., Wohlhueter, R. M. Kraupp, M. Adenosine uptake, transport, and metabolism in human erythrocytes. J. Cell Physiol. 125, 330鈥?36 (1985).CAS聽 Article聽 PubMed聽Google Scholar聽 41Baldwin, S. et al. The equilibrative nucleoside transporter family, SLC29. Pflugers Arch. 447, 735鈥?43 (2004).CAS聽 Article聽 PubMed聽Google Scholar聽 42Zimmerman, M. A. et al. Equilibrative nucleoside transporter (ENT)-1-dependent elevation of extracellular adenosine protects the liver during ischemia and reperfusion. Hepatology 58, 1766鈥?778 (2013).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 43Eltzschig, H. K. Carmeliet, P. Hypoxia and inflammation. N. Engl. J. Med. 364, 656鈥?65 (2011).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 44Janocha, A. J. et al. Nitric oxide during altitude acclimatization. N. Engl. J. Med. 365, 1942鈥?944 (2011).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 45Synnestvedt, K. et al. Ecto-5鈥?nucleotidase (CD73) regulation by hypoxia-inducible factor-1 mediates permeability changes in intestinal epithelia. J. Clin. Invest. 110, 993鈥?002 (2002).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 46Poth, J. M., Brodsky, K., Ehrentraut, H., Grenz, A. Eltzschig, H. K. Transcriptional control of adenosine signalling by hypoxia-inducible transcription factors during ischemic or inflammatory disease. J. Mol. Med. 91, 183鈥?93 (2013).CAS聽 Article聽 PubMed聽Google Scholar聽 47Morote-Garcia, J. C., Rosenberger, P., Nivillac, N. M., Coe, I. R. Eltzschig, H. K. Hypoxia-inducible factor-dependent repression of equilibrative nucleoside transporter 2 attenuates mucosal inflammation during intestinal hypoxia. Gastroenterology 136, 607鈥?18 (2009).CAS聽 Article聽 PubMed聽Google Scholar聽 48Eckle, T. et al. Crosstalk between the equilibrative nucleoside transporter ENT2 and alveolar Adora2b adenosine receptors dampens acute lung injury. FASEB J. 27, 3078鈥?089 (2013).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 49Feoktistov, I. Biaggioni, I. Adenosine A2B Receptors. Pharmacol. Rev. 49, 381鈥?02 (1997).CAS聽 PubMed聽Google Scholar聽 50Casanello, P. et al. Equilibrative nucleoside transporter 1 expression is downregulated by hypoxia in human umbilical vein endothelium. Circ. Res. 97, 16鈥?4 (2005).CAS聽 Article聽 PubMed聽Google Scholar聽 51Kim, D. G. Bynoe, M. S. A2A adenosine receptor modulates drug efflux transporter P-glycoprotein at the blood-brain barrier. J. Clin. Invest. 126, 1717鈥?733 (2016).Article聽 PubMed聽 PubMed Central聽Google Scholar聽 52Ryan, B. J. et al. AltitudeOmics: rapid hemoglobin mass alterations with early acclimatization to and de-acclimatization from 5260 m in healthy humans. PLoS ONE 9, e108788 (2014).ADS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 53Davies, J. et al. Adenosine promotes vascular barrier function in hyperoxic lung injury. Physiol. Rep. 2, e12155 (2014).Article聽 PubMed聽 PubMed Central聽Google Scholar聽 54Eckle, T., Grenz, A., Laucher, S. Eltzschig, H. K. A2B adenosine receptor signalling attenuates acute lung injury by enhancing alveolar fluid clearance in mice. J. Clin. Invest. 118, 3301鈥?315 (2008).CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 55Quintana, H. K. et al. Identification with MRI of the pleura as a major site of the acute inflammatory effects induced by ovalbumin and endotoxin challenge in the airways of the rat. Am. J. Physiol. Lung Cell Mol. Physiol. 291, L651鈥揕657 (2006).CAS聽 Article聽 PubMed聽Google Scholar聽 56Zhang, W. et al. Elevated ecto-5鈥?nucleotidase-mediated increased renal adenosine signaling Via A2B adenosine receptor contributes to chronic hypertension. Circ. Res. 112, 1466鈥?478 (2013).ADS聽 CAS聽 Article聽 PubMed聽Google Scholar聽 57Yegutkin, G. G., Samburski, S. S., Mortensen, S. P., Jalkanen, S. Gonzalez-Alonso, J. Intravascular ADP and soluble nucleotidases contribute to acute prothrombotic state during vigorous exercise in humans. J. Physiol. 579, 553鈥?64 (2007).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 58Yegutkin, G. G., Wieringa, B., Robson, S. C. Jalkanen, S. Metabolism of circulating ADP in the bloodstream is mediated via integrated actions of soluble adenylate kinase-1 and NTPDase1/CD39 activities. FASEB J. 26, 3875鈥?883 (2012).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 Download referencesAcknowledgementsThis work was supported by National Institute of Health Grants HL119549 (to Y.X.), DK083559 (to Y.X.) and HL113574 (to Y.X.). Funding for the overall AltitudeOmics study was provided, in part, by grants from the U.S. Department of Defense (W81XWH-11-2-0040 Telemedicine Advanced Technology Research Center (TATRC) to R.C.R. and W81XWH-10-2-0114 to A.T.L.); Cardiopulmonary Respiratory Physiology Laboratory, University of Oregon; and the Charles S. Houston Endowed Professorship at the Altitude Research Center, School of Medicine, University of Colorado. L.H. was supported by Cancer Prevention Research Institute of Texas RR150085.Author informationAffiliationsDepartment of Biochemistry and Molecular Biology, The University of Texas Health Science Center at Houston, Houston, 77030, Texas, USAAnren Song,聽Yujin Zhang,聽Leng Han,聽Hong Liu,聽Kaiqi Sun,聽Jessica Li,聽Harry Karmouty-Quintana,聽Takayuki Iriyama,聽Tingting Weng,聽Shushan Zhao,聽Wei Wang,聽Hongyu Wu,聽Mikhail Bogdanov,聽William Dowhan,聽Jianping Jin,聽Rodney E. Kellems,聽Michael Blackburn聽 聽Yang XiaMedicity Research Laboratory, University of Turku, Turku, 20520, FinlandGennady G. YegutkinGraduate School of Biomedical Sciences, University of Texas Health Science Center at Houston, Houston, 77030, Texas, USAHong Liu,聽Kaiqi Sun,聽Rodney E. Kellems,聽Michael Blackburn聽 聽Yang XiaDepartment of Biochemistry and Molecular Genetics, University of Colorado, Aurora, 80045, Colorado, USAAngelo D鈥橝lessandro聽 聽Kirk C. HansenDepartment of Obstetrics and Gynecology, Faculty of Medicine, The University of Tokyo, Tokyo, 113-8655, JapanTakayuki IriyamaDepartment of Orthopedics, Xiangya Hospital, Central South University, Changsha, 410008, Hunan, ChinaShushan ZhaoDepartment of Nephrology, Xiangya Hospital, Central South University, Changsha, 410008, Hunan, ChinaWei Wang聽 聽Yang XiaDepartment of Emergency Medicine University of Colorado School of Medicine, Altitude Research Center, Aurora, 80045, Colorado, USATravis Nemkov,聽Andrew W. Subudhi,聽Sonja Jameson-Van Houten,聽Colleen G. Julian,聽Andrew T. Lovering聽 聽Robert C. RoachDepartment of Pathology, MD Anderson Cancer Center, Houston, 77030, Texas, USAHong ZhangDepartment of Anesthesiology, Organ Protection Program, University of Colorado School of Medicine, Aurora, 80045, Colorado, USAHolger K. EltzschigAuthorsAnren SongView author publicationsYou can also search for this author in PubMed聽Google ScholarYujin ZhangView author publicationsYou can also search for this author in PubMed聽Google ScholarLeng HanView author publicationsYou can also search for this author in PubMed聽Google ScholarGennady G. YegutkinView author publicationsYou can also search for this author in PubMed聽Google ScholarHong LiuView author publicationsYou can also search for this author in PubMed聽Google ScholarKaiqi SunView author publicationsYou can also search for this author in PubMed聽Google ScholarAngelo D鈥橝lessandroView author publicationsYou can also search for this author in PubMed聽Google ScholarJessica LiView author publicationsYou can also search for this author in PubMed聽Google ScholarHarry Karmouty-QuintanaView author publicationsYou can also search for this author in PubMed聽Google ScholarTakayuki IriyamaView author publicationsYou can also search for this author in PubMed聽Google ScholarTingting WengView author publicationsYou can also search for this author in PubMed聽Google ScholarShushan ZhaoView author publicationsYou can also search for this author in PubMed聽Google ScholarWei WangView author publicationsYou can also search for this author in PubMed聽Google ScholarHongyu WuView author publicationsYou can also search for this author in PubMed聽Google ScholarTravis NemkovView author publicationsYou can also search for this author in PubMed聽Google ScholarAndrew W. SubudhiView author publicationsYou can also search for this author in PubMed聽Google ScholarSonja Jameson-Van HoutenView author publicationsYou can also search for this author in PubMed聽Google ScholarColleen G. JulianView author publicationsYou can also search for this author in PubMed聽Google ScholarAndrew T. LoveringView author publicationsYou can also search for this author in PubMed聽Google ScholarKirk C. HansenView author publicationsYou can also search for this author in PubMed聽Google ScholarHong ZhangView author publicationsYou can also search for this author in PubMed聽Google ScholarMikhail BogdanovView author publicationsYou can also search for this author in PubMed聽Google ScholarWilliam DowhanView author publicationsYou can also search for this author in PubMed聽Google ScholarJianping JinView author publicationsYou can also search for this author in PubMed聽Google ScholarRodney E. KellemsView author publicationsYou can also search for this author in PubMed聽Google ScholarHolger K. EltzschigView author publicationsYou can also search for this author in PubMed聽Google ScholarMichael BlackburnView author publicationsYou can also search for this author in PubMed聽Google ScholarRobert C. RoachView author publicationsYou can also search for this author in PubMed聽Google ScholarYang XiaView author publicationsYou can also search for this author in PubMed聽Google ScholarContributionsA.S. designed and conducted the mouse hypoxia treatment experiments, in vitro erythrocyte treatment experiments, western blot and IP and function of ENT measurement and analyzed all the experimental data, drew the figures and wrote the manuscript; G.G.Y. measured soluble CD3 activity in humans; H.L. measured sCD73 activity in mice. Y.Z. helped with biotin NHS injection and in vivo mouse hypoxia treatment; K.S. measured plasma ADA activity and helped with in vivo mouse hypoxia treatment; A.D. helped with adenosine determinations; J.L. helped to generate Ent1flox/floxEpoRCre-GFP mouse; H.K.-Q. and T.W. helped with BALF collection and related assay; T.I. helped with H.K.E. staining and tissue processing; S.Z. measured BALF IL-6 concentration; W.W. and H.W. helped with in vivo animal experiments, tissue collection and processing; T.N., A.W.S., S.J.-V.H., C.G.J. and A.T.L. helped with experiment design and blood sample collection in human high-altitude study; K.C.H. helped with adenosine determinations; M.B. and W.D. helped with in vitro C14-adenosine uptake assay and experiment design; L.H. performed statistical data analysis; J.J. helped with ubiquitin- and proteasome-related experiment design; H.Z. helped with IF and histological analysis; R.E.K. provided expertise in adenosine signalling and manuscript proofreading; H.K.E. and M.B. provided expertise in adenosine signalling and lung function and histological studies; R.C.R. led human high-altitude studies; Y.X. oversaw the design of experiments and interpretation of results, the writing and the editing of the manuscript.Corresponding authorCorrespondence to Yang Xia.Ethics declarations Competing interests The authors declare no competing financial interests. Supplementary information Supplementary InformationSupplementary Figures (PDF 674 kb) Peer Review File (PDF 345 kb)Rights and permissions This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article鈥檚 Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ Reprints and PermissionsAbout this articleCite this articleSong, A., Zhang, Y., Han, L. et al. Erythrocytes retain hypoxic adenosine response for faster acclimatization upon re-ascent. Nat Commun 8, 14108 (2017). https://doi.org/10.1038/ncomms14108Download citationReceived: 23 June 2016Accepted: 29 November 2016Published: 07 February 2017DOI: https://doi.org/10.1038/ncomms14108 M. Marlinge, D. Vairo, A. Bertaud, C. Vernet, M. Chefrour, L. Bruzzese, M. C. Chaptal, G. Mottola, A. Boussuges, J. J. Risso, M. Blot-Chabaud, M. Coulange, R. Guieu F. Joulia European Journal of Applied Physiology (2021) Supriya Saini, Praveen Vats, Susovon Bayen, Priya Gaur, Koushik Ray, Krishna Kishore, Meerim Sartmyrzaeva, Almaz Akunov, Abdirashit Maripov, Akpay Sarybaev, Bhuvnesh Kumar Shashi Bala Singh Functional Integrative Genomics (2019) Zhenwei Tang, Wenrui Ye, Haotian Chen, Xinwei Kuang, Jia Guo, Minmin Xiang, Cong Peng, Xiang Chen Hong Liu Purinergic Signalling (2019) C. Strewe, R. Zeller, M. Feuerecker, M. Hoerl, S. Matzel, I. Kumprej, A. Crispin, B. Johannes, T. Debevec, I. B. Mekjavic, O. Eiken, M. Thiel, G. Schelling A. Chouk猫r Scientific Reports (2018) CommentsBy submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate. Sign up for the Nature Briefing newsletter 鈥?what matters in science, free to your inbox daily.