Erythropoietin activates nitric oxide synthase in murine erythrocytes
Deyan Mihov, Johannes Vogel, Max Gassmann, and Anna BogdanovaInstitute of Veterinary Physiology, Vetsuisse Faculty and Zurich Center for Integrative Human Physiology, Universityof Zurich, Zurich, Switzerland
Submitted 22 October 2008; accepted in final form 9 June 2009
Mihov D, Vogel J, Gassmann M, Bogdanova A. Erythropoietinactivates nitric oxide synthase in murine erythrocytes. Am J PhysiolCell Physiol 297: C378–C388, 2009. First published June 10, 2009;doi:10.1152/ajpcell.00543.2008.—Erythropoietin (Epo) is the mainregulator of erythrocyte production and a potent cytoprotective factor.It was suggested that some of Epo cytoprotective properties are due toits regulation of nitric oxide (NO) production. Recently, functionallyactive endothelial type NO synthase (eNOS) was discovered in maturemurine and human red blood cells (RBC-eNOS). The goal of thepresent study was to characterize the effect of physiological andtherapeutic doses of Epo on RBC-eNOS function. We found thatrecombinant human Epo (rHuEpo) binds specifically to mouse eryth-rocytes. Epo binding sites are not equally distributed through the RBCpopulation but prevail in reticulocytes and young erythrocytes withabout 105 receptors/cell, compared with adult and old erythrocytescontaining 1–4 receptors/cell. The treatment of mouse erythrocyteswith rHuEpo resulted in a time- and dose-dependent upregulation ofNO production mediated via activation of the phosphatidylinositol-3-kinase /Akt pathway and RBC-eNOS phosphorylation at Ser-1177.Finally, when erythrocytes were incubated in L-arginine-free medium,rHuEpo treatment resulted in upregulation of superoxide radicalproduction with concomitant shifting of the cellular redox statetoward more oxidized state. Epo-induced changes in erythrocyteredox potential were absent in erythrocytes from eNOS-deficientmice.
red blood cells; phosphatidylinositol-3-kinase/Akt pathway; nitricoxide production; redox state
ERYTHROPOIETIN (Epo) is a major regulator of the red blood cellproduction, widely used in clinics to treat anemia (21, 47).Along with its erythropoietic properties, Epo nowadays isconsidered as a pleiotropic cytoprotective factor (22, 28).Binding of Epo to its receptor (EpoR) activates several survivalsignaling cascades, including phosphatidylinositol-3-kinase(PI3K)/Akt pathway in endothelial cells (4, 51) and cardiomyo-cytes (12). In both cell types studied, Epo upregulates theexpression of endothelial-type nitric oxide (NO) synthase(eNOS) and directly enhances the enzyme activity via PI3K/Akt-mediated phosphorylation of serine-1177 (Ser-1177) (4,12, 51). Recent reports suggested that Epo-induced regulationof NO production has a pivotal role in the cytoprotectiveeffects of Epo (12, 18, 29, 51).
The presence of EpoR on erythrocyte membranes is disput-able. The data found in the literature are controversial and theresponses of mature red blood cells (RBCs) to Epo treatmentare poorly investigated. The expression of EpoR during the lateerythroid development decreases exponentially, and matureerythrocytes were claimed to virtually lack EpoR and to be Epoinsensitive (9, 60). However, specific binding of Epo to rat and
human RBCs was reported (3, 45). Epo was shown to have aneffect on glucose transport (3, 23), antioxidant defense systems(13, 15), ion transport (5, 45), and rheological properties (26,58) of mature RBCs. At present, no data are available on theEpo-binding sites on erythrocyte membranes or their distribu-tion within RBC population.
As a putative downstream target of Epo-induced signaling inRBCs, we have chosen eNOS (RBC-eNOS), which presence inhuman and mouse erythrocytes was recently reported (33). Theactivity of the RBC-eNOS is comparable to that observed inconventional endothelium-derived eNOS (33). It is shown thatinsulin activates RBC-eNOS by phosphorylation of the enzymeat Ser-1177 via the PI3K/Akt pathway (33).
The aim of the present study was to characterize the bindingof Epo to murine erythrocytes and the effect of physiologicaland therapeutic doses of Epo on RBC-eNOS. Our data indicatethe presence of a single class Epo binding sites, similar inaffinity and downstream targets to a classical EpoR. We haveshown that Epo-binding sites are not equally distributed withinRBC population. Reticulocytes and young erythrocytes containmore Epo binding sites when compared with adult and oldcells. Epo treatment results in upregulation of erythrocyte NOproduction via activation of PI3K/Akt-signaling pathway andphosphorylation of RBC-eNOS at Ser-1177. Additionally, weprovide evidences for a direct link between the Epo-inducedregulation of RBC-eNOS activity and the maintenance of theredox state in mouse RBCs.
MATERIALS AND METHODS
All chemicals and kits used in this study were purchased fromSigma Aldrich, St. Louis, MO when not stated specifically.
Animals and RBC Preparation
In the present study we used male C57BL/6 mice (WT), 12 to 20wk old, and homozygous eNOS-deficient mice (eNOS�/�). Theanimals were raised in the sterile breeding facilities at the Institute ofVeterinary Physiology, University of Zurich. The mice were kept ona commercial diet as approved by the Veterinary Department ofCanton Zurich. All experiments were approved and performed inaccordance with the Swiss animal protection laws and institutionalguidelines. Animals were euthanized with CO2, and blood was col-lected immediately by cardiac puncture (0.8–1.2 ml) into heparinizedsyringes. The packed cells were washed three times in incubationmedium (in mM: 150 NaCl, 5 KCl, 1 CaCl2, 0.15 MgCl2, 10Tris-MOPS, 10 glucose, and 10 sucrose, pH 7.4). Buffy coat andplasma were discarded during the washing steps. The homogeneity oferythrocyte suspensions was monitored by FACS analysis. After thebuffy coat removal, the remaining nonerythroid cells (CD45-, CD16-,or CD36-positive cells) were �0.02% of the total cell population(supplemental Fig. S.1). Washed erythrocytes were subsequentlyresuspended to a hematocrit of about 10% and immediately used forfurther experiments. All studies were performed in the presence of 3mM L-arginine (L-Arg) unless stated otherwise.
Address for reprint requests and other correspondence: A. Y. Bogdanova.Institute of Veterinary Physiology, Zurich Univ., Winterthurerstrasse 260,CH-8057 Zurich, Switzerland (E-mail: [email protected]).
Am J Physiol Cell Physiol 297: C378–C388, 2009.First published June 10, 2009; doi:10.1152/ajpcell.00543.2008.
Binding of Epo to the mature murine erythrocytes was assessed byusing 125I-labeled recombinant human Epo (125I-Epo) with a molec-ular mass of 34,000 Da and a specific activity of 30 TBq/mmol(Amersham Biosciences, Freiburg, Germany) as previously described(45). Briefly, washed erythrocytes were resuspended at 5% hematocritin incubation medium containing 0.1% BSA. Aliquots from the cellsuspension (100 �l) were incubated for 3 h at 4°C with differentamounts of 125I-Epo (24–494 pM). Thereafter, nonbound radiola-beled Epo was removed by washing the cells three times with coldincubation medium, and cell-bound radioactivity was determined in agamma counter (Kontron Gamma Counting System). Specific bindingwas determined by subtracting the values of 125I-Epo in the presenceof at least 100-fold excess of unlabeled recombinant human Epo(rHuEpo/Eprex, Janssen-Cilag, Baar/Schweiz) from those in the ab-sence of nonlabeled rHuEpo. To investigate the nature of the Epo-binding sites, we tested the maximal specific binding of iodinated Epo(240 pM 125I-Epo, � excess of nonlabeled rHuEpo) in cells pretreatedfor 2 h with antibody (dilution 1:50) raised against EpoR (M-20, SantaCruz Biotechnology, Heidelberg, Germany). Finally, to determine thedistribution of Epo-binding sites within erythrocyte population,mouse RBCs were incubated with 240 pM 125I-Epo for 2 h and wereafterwards separated by density gradient centrifugation as describedfurther. Cell-bound radioactivity from the separated cells was deter-mined in a gamma counter or visualized autoradiographycally on anX-ray film.
Density Separation of Mouse RBCs
To divide mouse erythrocytes according to their age, RBCs wereseparated on continuous Percoll density gradient (Percoll Plus, GEHealthcare, Freiburg, Germany) as described previously (40). Thecells were divided into three fractions: reticulocytes and young RBCs(low cell density), adult (medium cell density), and old (high celldensity) erythrocytes (supplemental Fig. S.2A). The age of the differ-ent erythrocyte fractions was confirmed by the ratio of band 4.1a/4.1bas described elsewhere (5) (supplemental Fig. S.2, B and C) Thequality of the cell fractionation and the homogeneity of the RBCpopulations were verified by repeated density separation (supplemen-tal Fig. S.3).
RBCs Treatment
Characterization of molecular mechanism and kinetics of Epoeffect on the RBC-eNOS function. Erythrocytes from WT mice wereincubated for 2 h at room temperature in the presence or absence of 1U/ml rHuEpo. Aliquots were taken at 0, 30, 60, 120 min for Westernblot analysis and for measurement of NO metabolites (NO2
�/NO3�) in
the medium and cells. As a negative control, RBCs from eNOS�/�
mice were subjected to the same experimental protocol and NO2�/
NO3� levels in the medium, and cells were estimated. In additional
experiments, erythrocytes from WT and eNOS�/� mice were incu-bated for 2 h in the presence or absence of L-Arg ,and the levels ofnonprotein thiols were measured.
Dose response of Epo on the RBC-eNOS function. WT RBC wereincubated with 0, 1, 10, 50, and 100 U/ml rHuEpo for 30 min at roomtemperature and RBC-eNOS activity was estimated.
Sensitivity of Epo-induced NO production to PI3K and PKB block-ers, EpoR antibody, and NG-monomethyl-L-arginine. WT RBCs werepretreated for 30 min with either 1 �M wortmannin (PI3K blocker),150 �M A6730 (Akt inhibitor), EpoR antibody (M-20; dilution 1:10),or 3 mM NG-monomethyl-L-arginine (L-NMMA). After the incuba-tion, RBCs were treated with 10 U/ml rHuEpo for 1 h and activity ofRBC-eNOS was measured.
Erythrocyte Ghost Preparation
After treatment, erythrocytes were centrifuged at 4,000 g for 5min at 4°C. Packed RBCs (about 500 �l) were hemolyzed in 2 mlof ice-cold lysis buffer (10 mM Tris �HCl, 1 mM EDTA, 10 �g/mlpepstatin A, 10 �g/ml leupeptin, 5 �g/ml aprotinin, 0.1 mMPMSF, 10 mM Na-pyrophosphate, and 10 mM NaF, pH 7.4).Membranes were pelleted at 47,000 g for 20 min at 4°C (Sorvallcentrifuge RC-5B; rotor SS-34; Thermo Electron, Franklin, MA).The supernatant was used as cytosolic extract, and membraneswere solubilized in lysis buffer containing 0.5% deoxycholate and1% Triton X-100. The protein concentration of the samples wasdetermined with BCA Protein Assay (Pierce; Rockford, IL) withBSA as a standard.
Western Blot Analysis
The proteins from both cytosolic and membrane fractions wereseparated by 7.5% or 10% SDS-PAGE (500 �g protein/lane) andtransferred to Protan BA83 nitrocellulose membranes (Schleicher andSchuell, Dassel, Germany). Protein transfer was controlled by Pon-ceau red staining. Membranes were blocked for 1 h at room temper-ature and incubated overnight at 4°C with primary antibodies againstphospho-eNOS (Ser-1177; Cell Signaling Technology, Danvers, MA)or phospho-Akt/PKB� (Thr-308; Upstate, Lake Placid, NY). Stainingwith antibody against total actin was used as a loading control. Afterwashing was completed, membranes were incubated for 1 h at roomtemperature with the corresponding horseradish peroxidase-conju-gated secondary antibodies (anti-mouse/anti-rabbit; Jackson Immu-noResearch Laboratories, West Grove, PA). The enhanced chemilu-minescence detection Western blotting system was used for signalvisualization.
Evaluation of RBC-eNOS Activity
One of the widely used markers for NO production in biologicalsystems are nitrite/nitrate (NO2
�/NO3�) levels because they are stable
oxidation products of NO (32). In our study, we measured NO2� in the
incubation medium and RBCs using triiodide-based chemilumines-cence assay described in detail elsewhere (19). Briefly, an aliquot of200 �l from the RBC suspensions was taken at given times andcentrifuged for 1.5 min at 13,200 g. One-hundred seventy microlitersof the medium were immediately separated from the packed cells andplaced on ice until the assay was performed. Erythrocytes werereconstituted with 70 �l fresh, cold medium and lysed with 25 �l ofnitrite preservation solution containing 800 mM K3Fe(CN)6, 100 mMN-ethylmaleimide, and 10% Nonidet-40. The samples were deprotein-ized by addition of 125 �l methanol (100%) and centrifuged for 10min at 13,200 g. Fifty microliters from the incubation medium orsupernatant from the cell lysates were injected in the preheated (65°C)reaction chamber containing acidic triiodide (I 3
� ) reagent. The reagentwas prepared fresh before the measurements by mixing 1.65 g KI,0.57 g I2, 15 ml ddH2O, and 200 ml glacial CH3COOH. The reactionchamber was purged with helium, and released NO was detected usingCLD-88 analyzer (ECO MEDICS, Durnten, Switzerland). The signalwas processed using PowerChrom 280 system (eDAQ Pty; Spech-bach, Germany). To measure nitrate in the samples, we reduced theNO3
� to NO2� using cadmium-copper-based reduction kit Nitralyzer-II
(World Precision Instruments, Sarasota, FL). The assay was per-formed according to the manufacturer instructions. After the reduc-tion, NO2
� was measured as described above. NO3� levels were
estimated by the subtraction of NO2� levels before the reduction from
those obtained after the conversion of NO3� to NO2
�.
Measurements of Nonprotein Thiols
The amount of reduced (GSH) and oxidized (GSSG) glutathionewas assayed in erythrocytes as described previously (6). Briefly,samples from RBC suspensions were mixed 1:10 with deproteinizing
C379Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
solution containing 1.67 g glacial metaphosphoric acid, 0.2 gNa2EDTA, 30 g NaCl,and 100 ml ddH2O. After centrifugation, GSHconcentration was determined in supernatants using 5,5�-dithiobis(2-nitrobenzoic acid) (Ellman’s reagent). Optical density of the coloredcomplex was measured photometrically at 412 nm. Simultaneously,aliquots from the same samples were incubated in the presence ofglutathione reductase and NADPH for reduction of GSSG to GSH,and total glutathione levels (GSH�GSSG) were determined. Thehalf-cell reduction potential (Ehc) was calculated according to Schaferand Buettner (52).
Measurements of Superoxide Anion Production
To evaluate the influence of L-Arg deprivation and Epo on gener-ation of superoxide anions (O2
•�) by erythrocytes, we used chemilu-minescence Superoxide Anion Assay Kit (CS1000). Erythrocytes(5 � 106) were incubated for 2 h in the presence or absence of L-Argor in the presence of 1 U/ml rHuEpo without eNOS substrate in themedium. The assay was performed according to the producer’s guide-line.
Statistical Analysis
All data are based on at least six experiments and are presented asmeans � SE. The comparison between the experimental groups wasperformed using ANOVA and two-tailed Student’s t-test for unpairedsamples (GraphPad Instat.V3.05). The level of statistical significancewas set at P � 0.05.
RESULTS
Epo Binding to Murine Erythrocytes
Treatment of WT mouse erythrocytes with 125I-Epo in thepresence and in the absence of excessive amounts of nonradioac-tive Epo revealed a specific binding of Epo to the RBC mem-branes (Fig. 1A). The equilibrium binding isotherm (Fig. 1B)and scatchard analysis (Fig. 1B, inset) indicated the presence ofa single-type Epo-binding site with high affinity to Epo (Kd 58.3 � 11.1 pmol/l). We further performed separation ofmouse erythrocytes pretreated with 125I-Epo according to theirage. It is known that RBC density increases with cell aging(40). Using Percoll density gradient, RBCs were divided intothree major fractions: reticulocytes � young RBCs (low den-sity), adult erythrocytes (medium density), and old cells (highdensity), as shown in supplemental Fig. S.2A. The relative agedifferences between the RBC fractions were confirmed by theband 4.1a-to-4.1b ratio (supplemental Fig. S.2, B and C), whichincreases with the cell aging (56). Our data revealed unequal125I-Epo binding to the cells of the different erythrocyte frac-tions. Reticulocytes and young erythrocytes, representing2.24 � 0.14% of the RBC population (supplemental Fig.S.2D), had markedly higher ability to bind 125I-Epo (Fig. 1C).The number of Epo-binding sites per cell was respectively
Fig. 1. Binding of 125I-labeled erythropoeitn(125I-Epo) to erythrocytes. A: erythrocytesfrom wild-type (WT) mice were exposed to24 or 240 pM of 125I-Epo (molecular mass,34,000 Da; specific activity, 30 TBq/mmol)for 3 h at 4°C. The presence of excess (atleast 100�) of nonlabeled Epo reduced thebinding (shaded bars) compared with thesamples treated only with radioactive Epo(solid bars). *P � 0.05 compared with redblood cells (RBCs) treated only with 125I-Epo; n 6; means � SE. B: specific bindingof 125I-Epo to the RBCs membranes in thepresence of different amounts of labeledEpo. Erythrocytes were incubated for 3 h at4°C. B, inset: scatchard analysis of the datafrom B. C: binding of 125I-Epo to the differ-ent RBC fractions, measured by gammacounter. *P � 0.05 compared with fraction-1(reticulocytes � young RBCs), #P � 0.05compared with fraction-2 (adult RBCs); n 6; means � SE. D: autoradiography of den-sity-separated erythrocytes. The Percoll den-sity gradient of 125I-Epo-treated RBCs ispresented on the right with the correspondingX-ray image on the left.
C380 Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
105 � 8 for low-density, 4 � 1 for medium-density, and 2 �1 for high-density RBCs. The results obtained by using thegamma counter were confirmed autoradiographycally. Afterthe separation of mouse erythrocytes, pretreated with radioac-tive Epo, X-ray film was placed next to the centrifuge tube forabout 6 h. A specific band at the level of low-density erythro-cyte fraction was observed (Fig. 1D). Unfortunately, no bandswere detected on the film from the adult and old erythrocytepopulations, most probably due to the weaker and diffusesignal.
For further characterization of the Epo binding sites inmouse erythrocytes, the cells were pretreated with an antibodyagainst Epo-R (M-20, dilution 1:100). This resulted in a re-duction of the specific 125I-Epo binding from 427.5 � 77.6 to47.11 � 136.1 dpm/100 �l RBCs. Obtained results suggestthat murine erythrocytes specifically bind Epo and that itsbinding sites are recognized by EpoR-specific antibodies.
In addition, Epo binding to mouse erythrocytes was visual-ized using biotinylated human Epo and concomitant stainingwith the avidin conjugated with fluorescein. Double-stainingusing antibodies against transferrin receptors allowed us tospecifically assess interaction of Epo with reticulocytes. Asshown in Fig. S.4 reticulocytes showed strong Epo-positivestaining, whereas the Epo binding to mature erythrocytes wasmuch weaker and could not be detected in all cells but in some.These data are in line with those obtained using 125I-Eposuggesting that Epo preferably interacts with reticulocytes.Furthermore, there are some cells showing Epo binding but notransferrin receptor and some erythrocytes where Epo-positivestaining is below the detection limits.
Epo Triggers Activation of the Akt andeNOS Phosphorylation
We have further monitored the activity of the PI3K/Aktcascade known as a common EpoR-sensitive signaling path-way (21, 22, 28, 47). The cells were incubated in the absenceor presence of 1 U/ml rHuEpo for the period of 2 h, andaliquots of cell suspensions were collected after 30 min, 1 hand 2 h of treatment with the cytokine. Since Kleinbongardet al. (33) reported the presence of eNOS in both cytosolic andmembrane fractions in human erythrocytes, we have used bothcytosolic and membrane protein extracts for the assessment ofEpo effect on the Akt and eNOS phosphorylation state. West-ern blot analysis revealed an increased phosphorylation of Aktin the cytosolic extracts from RBCs treated with 1 U/mlrHuEpo compared with the Epo-free controls (Fig. 2A). Thephosphorylation peaked in the cells exposed to Epo for 30 minand decreased thereafter. The activation of Akt was followedby an increase in the phosphorylation of eNOS, one of itsdownstream targets. As shown in Fig. 2B, exposure of RBCs to1 U/ml rHuEpo caused phosphorylation of the RBC-eNOS atSer-1177. Upregulation of Ser-1177 phosphorylation could bedetected in the membrane fractions already 30 min after cyto-kine administration reaching its maximum within 1 h of Epotreatment. Phosphorylation of eNOS at Ser-1177 is known toactivate the enzyme (43). Therefore, we followed the kineticsand dose dependence of Epo action on the NO production inmurine erythrocytes.
Kinetics of Epo effect on RBC-eNOS activity. We usedNO2
�/NO3� levels in the cells and incubation medium as mark-
ers of NO production (32) by erythrocytes treated with variousdoses of rHuEpo for various time periods. The exposure ofmouse erythrocytes to 1 U/ml rHuEpo triggered upregulationof NO production. The rate of NO2
� accumulation in theincubation medium was doubled in the presence of 1 U/mlrHuEpo (Fig. 3A; 2.27 � 0.39 compared with 0.87 � 0.13�mol/h � l cells in the Epo-free control). The intracellular NO2
�
levels in nontreated cells transiently decreased during theincubation, whereas Epo administration resulted in a slowaccumulation of NO2
� in the RBCs (Fig. 3B). The NO3� cellular
content in control erythrocytes was practically constant over2 h of incubation but increased profoundly in the Epo-treatederythrocytes already after 30 min of incubation (Fig. 3D). Ourdata suggest that most of de novo produced NO and resultingNO2
� were either removed from the cells or rapidly oxidized tonitrate when in the cytosol. The resulting steady-state levels ofNO3
� in Epo-treated RBCs were threefold exceeding those incontrol cells (Fig. 3D). Concomitantly, the rate of NO3
� accu-mulation in the incubation medium was threefold higher in thepresence of Epo (Fig. 3C; 179.9 � 25.5 compared with63.04 � 18.1 �mol/h � l cells in the Epo-free control).
The obtained results suggest that treatment of mouse eryth-rocytes with 1 U/ml rHuEpo caused a threefold increase in theRBC-eNOS activity that could be detected already 30 min afterthe cytokine administration. Therefore, 30-min incubation pe-riod was chosen for the assessment of the dose response of Epoaction on the eNOS function.
Dose Dependence of Epo Effect on RBC-eNOS Activity
The concentration of Epo (1 U/ml, rHuEpo) that we usedwhen studying kinetics of Epo-induced RBC-eNOS activationreflects the endogenous cytokine plasma levels after physio-
Fig. 2. Epo triggers Akt and RBC-endothelial nitric oxide synthase (eNOS)phosphorylation in mouse erythrocytes. Erythrocyte lysates from WT micewere divided into cytosolic (cyt) and membrane (mem) fractions and proteinsseparated by SDS-PAGE. A: Western blot analysis revealed that Epo treatmentresulted in increased phosphorylation of Akt. The signal was maximal at 30min time point and gradually decreased thereafter. B: Epo induced phosphor-ylation of RBC-eNOS at Ser-1177. A specific signal at 132 kDa was detected30 min after Epo administration. Maximum in RBC-eNOS phosphorylationwas observed after 1 h of incubation and than the signal decreased. Specificsignal was observed predominantly in the membrane fractions.
C381Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
logical stimulation by hypoxia or decreased blood hemoglobin(1, 21, 28). Therapeutic Epo doses used in the clinics howevervary between 150 and 40,000 U/kg (21, 28, 47). Thus Epoplasma levels in patients reach 10–100 U/ml. To characterizethe effects of physiological and therapeutic doses of Epo on theRBC-eNOS activity, we incubated WT erythrocytes with in-creasing amount of rHuEpo ranging between 1 and 100 U/ml.
Epo-induced changes of NO2�/NO3
� levels in the erythro-cytes and the medium over 30 min of incubation were dosedependent, with a peak at 10 U/ml rHuEpo (Fig. 4). Whenhigher Epo doses were applied (50–100 U/ml), NO2
� levelswere somewhat lower than those at 10 U/ml. Of note, similarbiphasic dose-dependent induction of NO production by Epowas reported in endothelial cell lines (4).
Lack of Epo Effect on NO2�/NO3
� Levels in RBCs FromeNOS�/� Mice
To confirm the specific effect of Epo on RBC-eNOS, westudied the Epo-induced responses on NO2
�/NO3� levels in
erythrocytes obtained from eNOS�/� mice. Basal NO2�/NO3
�
levels in the medium during the incubation of eNOS-deficientRBCs were significantly lower compared with the valuesobtained for WT erythrocytes (Fig. 5, A and C). The intracel-lular NO2
�/NO3� levels in eNOS�/� erythrocytes decreased
with time (Fig. 5, B and D) as NO2� oxidation to NO3
� alongwith NO2
�/NO3� leakage from the cells occurred. The treatment
of eNOS�/� RBCs with 10 U/ml rHuEpo failed to affect theNO2
�/NO3� levels in the medium and erythrocytes (Fig. 5).
Further experiments with WT RBCs revealed that in absenceof eNOS substrate L-Arg, Epo failed to affect NO2
�/NO3�
content in the cells or incubation medium (supplemental Fig.S.5). These results strongly suggest that the observed Epo-induced changes in NO2
�/NO3� levels in WT erythrocytes and
incubation medium are due to the specific action of Epo on theeNOS present in the murine RBCs.
White Blood Cells Do Not Contribute to Epo-InducedRegulation of RBC-eNOS
In the next set of experiments we assessed the possibleinfluence of white blood cell contamination of the erythrocytessuspension on Epo-induced NO production. Whole blood sam-ples and buffy coat-free erythrocytes suspension were exposedto 10 U/ml Epo for 1 h and the NO2
�/NO3� accumulation in the
cells and the medium compared with those for the Epo-freesamples. There was no difference in the rates of Epo-inducedNO2
� accumulation (NO2�, 0.67 � 0.13 �mol/h � l RBC with
and 0.69 � 0.18 �mol/h � l RBC without buffy coat) or in theEpo-induced intracellular NO3
� accumulation (NO3�, 387.6 �
86.5 �mol/h � l RBC with and 370 � 49.3 �mol/h � l RBCwithout buffy coat) between the whole blood and the buffycoat-free samples. Obtained results confirm the specific, direct,
Fig. 3. Kinetics of Epo-induced regulationof the RBC-eNOS activity. RBC-eNOS ac-tivity was estimated by monitoring nitrite/nitrate (NO2
�/NO3�) levels in the incubation
medium and the cells. Erythrocytes fromWT mice were incubated for 0–120 min inthe presence (closed circles) or absence(open circles) of 1 U/ml recombinant humanEpo (rHuEpo). A: accumulation rate andsteady state of NO2
� in the incubation me-dium were elevated by Epo. B: transientdepletion of basal NO2
� cellular levels wasobserved. Epo treatment resulted in a mod-erate increase of the intracellular NO2
� lev-els. C: both accumulation rate and steadystate for NO3
� in the incubation mediumwere elevated by Epo. D: Epo treatmentresulted in pronounced accumulation ofNO3
� in the RBCs. *P � 0.05 compared withthe respective time matching control, #P �0.05 compared with the NO2
� levels in con-trol RBCs at the beginning of the incubation,$P � 0.05 compared with NO2
� levels inEpo-treated RBCs at the beginning of theincubation; n 10; means � SE.
C382 Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
and acute modulatory effect of Epo on the erythrocyte NOproduction.
Blocking of Epo-Induced Activation of RBC-eNOS
To further characterize the mechanism behind Epo-inducedactivation, we blocked different components of the putativesignaling pathway and monitored the NO2
�/NO3� levels in the
presence or absence of Epo. NO production by Epo wascompletely abolished by the pretreatment of RBCs with 1 �MPI3K inhibitor wortmannin, 150 �M Akt blocker A6730 orEpoR antibody (M-20, dilution 1:100) (Fig. 6). In addition, thedirect blocking of de novo NO production by pretreatment ofthe cells with L-NMMA, prevented Epo-induced increase ofintra- and extracellular NO2
�/NO3� levels (Fig. 6). These results
are in agreement with the data obtained for the Epo-induced
phosphorylation of Akt and RBC-eNOS, suggesting that Eporegulates erythrocyte NO production via PI3K/Akt-mediatedRBC-eNOS activation.
L-Arg Availability and Epo-Treatment Modulate the RedoxState of RBCs
Under certain conditions, such as substrate deficiency, eNOSis capable of generation of superoxide anions (O2
•�) instead ofNO (14, 44), causing oxidative stress. We tested whether Epomay induce an oxidative stress in RBCs deprived of L-Arg. Theeffect of Epo on NO production and cellular redox state inRBCs incubated in L-Arg-containing and L-Arg-free mediumwas monitored. Epo failed to increase NO2
�/NO3� levels in the
erythrocytes and incubation medium in the absence of L-Arg(data not shown). Moreover, the omission of L-Arg triggered a
Fig. 4. Dose dependence of Epo effect onRBC-eNOS activity. Erythrocytes from WTmice were incubated in the presence of 0–100U/ml rHuEpo for 30 min and NO2
�/NO3� lev-
els in the medium and cells were measured.Epo-induced activation of erythrocyte NOproduction had a bell-shape dose dependenceprofile with maximal response observed at 10U/ml rHuEpo. A: dose dependence of Epo-induced changes of NO2
� levels in the incu-bation medium. B: dose dependence of Epoeffect on intracellular NO2
� levels. C: Epo-induced accumulation of NO3
� in the incuba-tion medium as a function of Epo concentra-tion. D: dose dependence of the Epo-inducedincrease in cellular NO3
�. *P � 0.05 com-pared with untreated cells (open bars); n 10; means � SE.
C383Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
significant upregulation of O2•� generation by the RBCs,
which was further facilitated by Epo (Fig. 7A).Both, NO and O2
•� generation contribute substantially to themaintenance of the cellular redox state. We measured intracel-lular reduced (GSH) and oxidized (GSSG) glutathione levelsafter 2 h of incubation with 1–100 U/ml rHuEpo in thepresence and absence of L-Arg and expressed the data as ahalf-cell redox potential for the GSH:GSSG couple (52). In-cubation of the RBCs in L-Arg-free medium shifted the half-cell reduction potential to more oxidized (Fig. 7B). Treatmentof the erythrocytes with 1 U/ml rHuEpo in absence of L-Argfurther facilitated GSH oxidation (Fig. 7B). In L-Arg-contain-ing medium the redox potential was preserved in cells incu-bated with 1 U/ml rHuEpo for at least 2 h. Acute treatment ofRBCs with a therapeutic dose of 100 U/ml rHuEpo in thepresence of L-Arg caused a statistically significant decrease inthe half-cell reduction potential, indicating a shift to morereduced state. Our data suggest that the influence of Epo-induced regulation of RBC-eNOS on the cellular redox statedepends on the presence of L-Arg. Depending on the L-Argavailability, Epo may play a dual role as pro- or antioxidant.
Finally, we tested whether Epo and L-Arg will affect thecellular redox state of the eNOS�/� erythrocytes. As expected,the availability of the eNOS substrate was without an effect onthe half-cell redox potential of eNOS-deficient RBCs(�291.7 � 5.6 mV in the presence and �290.2 � 3.3 mV in
the absence of 3 mM L-Arg in the medium). Moreover, 100 U/ml rHuEpo failed to affect the cellular redox state in thepresence or absence of L-Arg (�287.6 � 13.4 in the absenceand �292.4 � 8.5 in the presence of 3 mM L-Arg in themedium).
DISCUSSION
Our study is the first to characterize the interaction of Epowith murine RBCs. The obtained data imply that eNOS is oneof the targets of Epo action in mouse erythrocytes. By activat-ing RBC-eNOS, Epo may have a profound effect on a numberof cellular properties of which maintenance of the redoxpotential is of key importance.
Characteristics of Epo Binding Sites on Murine Erythrocytes
We have detected the presence of Epo binding sites onmurine erythrocyte membrane. Skatchard analysis of the dataon the dose response of 125I-Epo binding (Fig. 1B and inset)reveals a single binding site class with a Kd 58.3 � 11.1pmol/l, a value within the range of affinity of classical EpoR toEpo reported in the literature (30 to 330 pmol/l) (9, 45, 54, 60).The wide scatter of the Kd values reveals a broad tissue- andspecies-specific variability of the EpoR properties. Interactionof Epo with the erythrocyte membrane could be abolished bypretreatment of the cells with an antibody against mouse EpoR,
Fig. 5. Lack of effect of Epo on NO2�/NO3
�
production in erythrocytes from eNOSknockout (eNOS�/�) mice. Erythrocytesfrom eNOS�/� mice were incubated for 2 hin the presence (closed triangles) or absence(open triangles) of 10 U/ml rHuEpo. Thecorresponding basal values obtained fromWT erythrocytes (open circles) are includedfor comparison. A: basal NO2
� levels in themedium of eNOS�/� erythrocytes were sig-nificantly lower compared with WT. Epoadministration did not cause any change inthe extracellular NO2
� levels. B: Epo admin-istration was without an effect on the intra-cellular NO2
� content in eNOS�/� RBCs.C: basal NO3
� levels in the incubation me-dium of eNOS�/� erythrocytes were lowercompared with WT RBCs. Epo treatmentdid not alter the extracellular NO3
� levels.D: addition of rHuEpo to eNOS�/� erythro-cytes failed to affect cytosolic NO3
� levels.*P � 0.05 WT control compared witheNOS�/� control, #P � 0.05 WT controlcompared with eNOS�/� RBCs treated with10 U/ml rHuEpo; n 10; means � SE.
C384 Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
which was claimed to target the receptor specifically (20).Furthermore, Epo binding to the cells activated PI3K/Aktsignaling cascade (Figs. 2 and 6), a well-characterized down-stream signaling pathway of a common EpoR in erythroidprecursor cells (21, 22, 28, 47). Taken together, these obser-vations suggest that Epo interacts with a common Epo receptorpresent on erythrocyte membranes. The low number of recep-tors as well as the absence of reliable highly specific antibodiesmakes the direct molecular identification technically demand-ing.
Previous studies using 125I-Epo to track Epo binding tohuman erythrocytes reported the presence of 5–6 Epo bindingsites per cell, based on the assumption that the number ofspecific Epo binding sites is equal for each erythrocyte (45). Ifthe same assumption is applied to our data, similar numbersmay be obtained (5–6 Epo binding sites per mouse erythro-cyte). However, the number of binding sites per cell variesbetween 100 and 2 binding site per cell, being most abundant
Fig. 7. Effect of L-Arg deprivation and Epo on cellular redox state.A: superoxide anion (O2
•�) production of WT RBCs was measured during 2 hof incubation period using luminescent assay in the presence or absence of 3mM L-Arg. L-Arg deprivation resulted in pronounced increase of O2
•� gener-ation. Administration of 1 U/ml rHuEpo in the absence eNOS substrate wasaccompanied by further increase in O2
•� production. B: WT erythrocytes wereincubated for 2 h in the presence or absence of 3 mM L-Arg. At the beginningof the incubation, RBCs were separated in several groups and treated witheither saline, 1 U/ml, or 100 U/ml rHuEpo. At the end of the incubation period,samples were taken for measuring of reduced and oxidized glutathione (GSH/GSSG). Half-cell reduction potential (Ehc) was calculated as previously de-scribed (52). eNOS substrate deprivation resulted in increased Ehc comparedwith the L-Arg-containing control thus shifting cellular redox state to moreoxidized. Epo treatment led to further oxidation of GSH. In contrary, in thepresence of L-Arg low Epo levels did not affect Ehc. Moreover, therapeuticdoses of Epo (100 U/ml) exerted antioxidative effect by decreasing the Ehc andrespectively altering redox potential of the RBCs to more reduced. *P � 0.05compared with L-Arg containing control, #P � 0.05 compared with L-Arg-deficient control; n 6; means � SE.
Fig. 6. Blocking of Epo-induced activation of RBC-eNOS. WT RBCs werepretreated with either 1 �M wortmannin, 150 �M A6730, 3 mM NG-monomethyl-L-arginine (L-NMMA) or Epo receptor antibody (M-20, dilution1:100) for 30 min to block phosphotidylinositol-3-kinase (PI3K), Akt, eNOS,or Epo binding to the cells. Furthermore, RBC suspensions were incubatedwith 10 U/ml rHuEpo for 1 h. Control experiments with the blockers alonewere performed and NO2
�/NO3� levels were taken as zero (basal) levels. The
presented values were obtained by subtraction of the basal/zero levels from thecorresponding values for the Epo-treated samples. These results hence repre-sent the Epo-sensitive fraction of the intra- and extracellular nitrite/nitrateproduction, termed as NO2
�/NO3�. A: Epo-induced increase of cellular and
extracellular NO2� levels was abolished by pretreatment of the RBCs with
wortmannin, A6730, L-NMMA, or M-20. B: Epo failed to affect intra- andextracellular NO3
� levels in mouse RBCs pretreated with wortmannin, A6730,L-NMMA, or M-20. *P � 0.05 compared with basal level, $P � 0.05compared with NO2
�/NO3� in the medium of cells treated only with Epo
(first solid bars on the left), #P � 0.05 compared with NO2�/NO3
� in RBCstreated only with Epo (first shaded bars on the left); n 10; means � SE.
C385Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
in reticulocytes (Fig. 1, C and D; supplemental Fig. S.4). Themarked decrease in Epo-binding capability with cell aging ismost likely achieved as Epo receptor is released in vesiclesduring cell maturation (39) or undergoes internalization andproteasomal degradation similar to transferrin receptor (25).EpoR internalization with a following degradation or partialrecycling was reported for erythroid precursor cells (24, 59)but never before in the circulating erythrocytes despite thepresence of proteasomal machinery in mature RBCs (30).
The results we obtained suggest that all erythrocytes have apotential to be Epo sensitive, but the degree and significance ofthe cellular response may be age dependent. Unfortunately, nofunctional studies can be performed on the Epo-induced NOproduction in subfractions obtained by Percoll density gradientcentrifugation, since the contact with the Percoll rendered thecells Epo insensitive (data not shown). Since released, NO maydiffuse affecting all the neighboring cells when in bloodstream. Therefore, the Epo effects on NO production that wehave observed in nonseparated RBCs population are physio-logically relevant even when generated by relatively few cells.
Epo-Induced Regulation of Erythrocyte NO Production
We have shown that Epo triggers an increase in the activityof the eNOS in mouse erythrocytes by promoting phosphory-lation of the enzyme at Ser-1177 by Akt (Fig. 2). This adds Epoto the list of physiological regulators of the RBC-eNOS alongwith insulin, acting through the same signaling cascade (33).RBC-born NO, in addition to the upregulation of NO produc-tion by endothelial cells and cardiac myocytes, would facilitatethe cytoprotective effect of Epo (4, 12, 42, 51).
Since Epo is strongly considered as a potential cardiopro-tective agent (7, 12, 22, 28), the dose dependence of the NOrelease is of particular importance. The acute upregulation ofNO production in RBCs could be detected after administrationof 1 U/ml rHuEpo (Fig. 3). This Epo concentration exceeds theone found in normoxic human or mouse blood (0.001 to 0.027U/ml) (21, 28, 47) by several orders of magnitude but might bereached in vivo in response to hypoxia or anemia (1–10 U/ml)(1). We cannot exclude that the long-term effects of Epo onRBC-eNOS may be observed at much lower doses than thoseused in our study. Epo doses used in the clinics are however inthe range between 150 and 40,000 U/kg corresponding to10–100 U/ml in plasma (21, 28, 47). Our data indicate that theincrease in Epo dose above 10 U/ml does not cause furtherincrease in the NO production (Fig. 4). In fact the Epo-inducedregulation of RBC-eNOS was biphasic and resembled the doseresponse for Epo-induced NO generation reported in the en-dothelial cells (4).
Role of Epo in Controlling the Cellular Redox State
eNOS is actively involved in the regulation of intracellularredox state. Under conditions of substrate or cofactor defi-ciency and in response to atherogenic stimuli, eNOS generatessuperoxide anions along or instead of NO (14, 44). We haveshown that this is the case when Epo-induced activation ofRBC-eNOS occurs in L-Arg-free medium (Fig. 7). The ob-tained results imply that Epo can be pro- or anti-oxidantdepending on the conditions. Oxidative stress is one of thefactors triggering RBCs senescence (37, 38). Thus Epo mayaccelerate or attenuate senescence depending on the L-Arg
availability (2). Of note, our experiments were performed withRBCs equilibrated with 20% O2 and thus fully O2 saturated. Inthe organism, where hemoglobin O2 saturation varies fromalmost 100% in lung arterioles to about 75% in venous blood,O2
•� production may be reduced compared with that we reporthere. However, under pathological conditions, Epo-inducedchanges in cellular redox state may become significant.
Our results may have particularly important consequencesfor the pathophysiology of rHuEpo therapy. Increased oxida-tive stress is a feature characteristic in patients with chronicrenal failure undergoing hemodialysis (HD) (27, 36, 48). Sev-eral publications reported that Epo treatment of uremic patientson chronic HD is accompanied by further increase of oxidativestress and requires coapplication of antioxidants (17, 35, 46). Apossible cause for these observations could be L-Arg deficiencyand uncoupling of eNOS activity via inhibition of L-Arg uptake(10, 41, 62) or changes in extracellular L-Arg content (10, 11,61). Our data suggest that coapplication of Epo and L-Argwould help to avoid probable side effects of Epo treatment.However, our observations were made for mouse erythrocytesand thus cannot be directly transferred to human RBCs. Furtherexperiments are needed to assess the relevance of our findingsfor human erythrocytes.
Physiological Relevance of Epo-Induced Activation of theeNOS in RBCs
Our findings along with the previous reports raise an impor-tant question that remains to be answered: what is the role ofRBC-born NO and Epo-induced regulation of RBC-eNOSactivity? When Epo plasma levels increase under hypoxicconditions, together with an augment in RBC mass, Epo-induced NO production could improve tissue’s blood supplydue to NO-mediated vasodilatation (16, 31, 53). The relativecontribution of RBC-eNOS activity to the total NO poolremains to be characterized. The increase in RBC-derived NOwould reduce the scavenging of NO produced by endothelialcells, thereby facilitating NO-induced vasodilatation. Underhypoxic conditions, the NO2
� to NO conversion may be cata-lyzed by membrane-bound deoxyhemoglobin (16, 19, 31, 53).Under O2 saturating conditions, RBC-eNOS synthase willcontribute most to the RBC-derived NO production. Apart ofcontrolling vascular tone, RBC-derived NO pool serves thelocal needs; e.g., increasing membrane fluidity and RBC de-formability (8, 33) thus improving capillary perfusion andreducing sheer stress (58). Furthermore, oxygen release fromhemoglobin in hypoxic conditions is also enhanced by NO asit reduces the oxygen affinity of hemoglobin, thus facilitatingthe oxygen transfer from erythrocytes to the tissues (34, 55).
Another interesting aspect that draws attention is the unequaldistribution of Epo binding sites within RBC population and itsrole for the cells. According to our findings reticulocytes andyoung erythrocytes contain much more Epo receptors com-pared with the adult and old ones, suggesting that young RBCsare more sensitive to changes in Epo plasma levels. Over thelast years several papers appeared describing a process ofselective lysis of relatively young erythrocytes called neocy-tolysis (2, 50, 57). Neocytolysis occurs in individuals acclima-tized to high altitude on descent to sea level or astronautsascending into space and is considered as a physiologicalprocess for downregulation of excessive RBC mass (2). The
C386 Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
exact mechanism of neocytolysis is still unclear, but it wasfound that it is accompanied by decrease of Epo plasma levelsand selective removal of relatively young erythrocytes, gener-ated over the previous 10–11 days (49). Our data suggest thatthe number of Epo-binding sites may be the key factor forselecting and survival of the different erythrocyte fractions.The greater number of Epo-binding sites on the young RBCscould explain their preferential removal when Epo plasmalevels drop below a certain threshold.
In summary, our findings indicate that mouse erythrocytespossess classical EpoRs, the number of which per cell dependson the RBC age. Epo activates RBC-eNOS and regulates thedistribution of NO-bioactive species between the RBCs and themedium. Furthermore, the Epo-induced regulation of eNOSactivity affects redox state of the RBCs by generation of NO orO2
•� depending on L-Arg bioavailability. Pro-oxidative poten-tial of rHuEpo should be seriously considered when using thisdrug in clinics.
ACKNOWLEDGMENTS
We thank Dr. Hans U. Lutz, Institute of Biochemistry, ETH Zurich forhelpful comments and suggestions.
GRANTS
This work was supported by Swiss National Science Foundation (3100B0–112449 to A. Bogdanova) and Zurich Center for Integrative Human Physiol-ogy.
REFERENCES
1. Abbrecht PH, Littell JK. Plasma erythropoietin in men and mice duringacclimatization to different altitudes. J Appl Physiol 32: 54–58, 1972.
2. Alfrey CP, Rice L, Udden MM, Driscoll TB. Neocytolysis: physiolog-ical down-regulator of red-cell mass. Lancet 349: 1389–1390, 1997.
3. Baciu I, Ivanof L, Pavel T, Marina C, Zirbo M, Zdrenghea C.Erythropoietin binding to the red cell membranes. Physiologie 22: 227–231, 1985.
4. Beleslin-Cokic BB, Cokic VP, Yu X, Weksler BB, Schechter AN,Noguchi CT. Erythropoietin and hypoxia stimulate erythropoietin recep-tor and nitric oxide production by endothelial cells. Blood 104: 2073–2080, 2004.
5. Bogdanova A, Mihov D, Lutz H, Saam B, Gassmann M, Vogel J.Enhanced erythro-phagocytosis in polycythemic mice overexpressingerythropoietin. Blood 110: 762–769, 2007.
6. Bogdanova AY, Ogunshola OO, Bauer C, Gassmann M. Pivotal role ofreduced glutathione in oxygen-induced regulation of the Na(�)/K(�)pump in mouse erythrocyte membranes. J Memb Biol 195: 33–42, 2003.
7. Bogoyevitch MA. An update on the cardiac effects of erythropoietincardioprotection by erythropoietin and the lessons learnt from studies inneuroprotection. Cardiovasc Res 63: 208–216, 2004.
8. Bor-Kucukatay M, Wenby RB, Meiselman HJ, Baskurt OK. Effects ofnitric oxide on red blood cell deformability. Am J Physiol Heart CircPhysiol 284: H1577–H1584, 2003.
9. Broudy VC, Lin N, Brice M, Nakamoto B, Papayannopoulou T.Erythropoietin receptor characteristics on primary human erythroid cells.Blood 77: 2583–2590, 1991.
10. Brunini T, Resende A, Moss M, Soares de Moura R, and MendesRibeiro A. L-Arginine availability as a pathological mechanism in essen-tial hypertension, chronic renal and heart failure. Vasc Dis Prev 2: 37–51,2005.
11. Brunini TM, Roberts NB, Yaqoob MM, Ellory JC, Mann GE, andMendes Ribeiro AC. Activation of L-arginine transport in undialysedchronic renal failure and continuous ambulatory peritoneal dialysis pa-tients. Clin Exp Pharmacol Physiol 33: 114–118, 2006.
12. Burger D, Lei M, Geoghegan-Morphet N, Lu X, Xenocostas A, FengQ. Erythropoietin protects cardiomyocytes from apoptosis via up-regula-tion of endothelial nitric oxide synthase. Cardiovasc Res 72: 51–59, 2006.
13. Chakraborty M, Ghosal J, Biswas T, Datta AG. Effect of erythropoietinon membrane lipid peroxidation, superoxide dismutase, catalase, and
glutathione peroxidase of rat RBC. Biochem Med Metab Biol 40: 8–18,1988.
14. Channon KM, Guzik TJ. Mechanisms of superoxide production inhuman blood vessels: relationship to endothelial dysfunction, clinical andgenetic risk factors. J Physiol Pharmacol 53: 515–524, 2002.
15. Chattopadhyay A, Choudhury TD, Bandyopadhyay D, Datta AG.Protective effect of erythropoietin on the oxidative damage of erythrocytemembrane by hydroxyl radical. Biochem Pharmacol 59: 419–425, 2000.
16. Crawford JH, Isbell TS, Huang Z, Shiva S, Chacko BK, SchechterAN, Darley-Usmar VM, Kerby JD, Lang JD Jr, Kraus D, Ho C,Gladwin MT, and Patel RP. Hypoxia, red blood cells, and nitrite regulateNO-dependent hypoxic vasodilation. Blood 107: 566–574, 2006.
17. Cristol JP, Bosc JY, Badiou S, Leblanc M, Lorrho R, Descomps B,Canaud B. Erythropoietin and oxidative stress in haemodialysis: benefi-cial effects of vitamin E supplementation. Nephrol Dial Transplant 12:2312–2317, 1997.
18. d’Uscio LV, Smith LA, Santhanam AV, Richardson D, Nath KA,Katusic ZS. Essential role of endothelial nitric oxide synthase in vasculareffects of erythropoietin. Hypertension 49: 1142–1148, 2007.
19. Dejam A, Hunter CJ, Pelletier MM, Hsu LL, Machado RF, Shiva S,Power GG, Kelm M, Gladwin MT, Schechter AN. Erythrocytes are themajor intravascular storage sites of nitrite in human blood. Blood 106:734–739, 2005.
20. Elliott S, Busse L, Bass MB, Lu H, Sarosi I, Sinclair AM, Spahr C, UmM, Van G, Begley CG. Anti-Epo receptor antibodies do not predict Eporeceptor expression. Blood 107: 1892–1895, 2006.
21. Fisher JW. Erythropoietin: physiology and pharmacology update. ExpBiol Med 228: 1–14, 2003.
22. Ghezzi P, Brines M. Erythropoietin as an antiapoptotic, tissue-protectivecytokine. Cell Death Differ 11 Suppl 1: S37–S44, 2004.
23. Ghosal J, Chakraborty M, Biswas T, Ganguly CK, Datta AG. Effect oferythropoietin on the glucose transport of rat erythrocytes and bonemarrow cells. Biochem Med Metab Biol 38: 134–141, 1987.
24. Gross AW, Lodish HF. Cellular trafficking and degradation of erythro-poietin and novel erythropoiesis stimulating protein (NESP). J Biol Chem281: 2024–2032, 2006.
25. Harding C, Heuser J, Stahl P. Receptor-mediated endocytosis of trans-ferrin and recycling of the transferrin receptor in rat reticulocytes. J CellBiol 97: 329–339, 1983.
26. Hassan K, Roguin N, Kaganov Y, Hasan S, Kristal B. Effect oferythropoietin therapy on red cells filterability and left ventricular mass inpredialysis patients. Ren Fail 27: 177–182, 2005.
27. Jacob HS, Eaton JW, Yawata Y. Shortened red blood cell survival inuremic patients: beneficial and deleterious effects of dialysis. Kidney Int 2:139–143, 1975.
28. Jelkmann W. Erythropoietin after a century of research: younger thanever. Eur J Haematol 78: 183–205, 2007.
29. Jie KE, Verhaar MC, Cramer MJM, van der Putten K, GaillardCAJM, Doevendans PA, Koomans HA, Joles JA, Braam B. Erythro-poietin and the cardiorenal syndrome: cellular mechanisms on the cardio-renal connectors. Am J Physiol Renal Physiol 291: F932–F944, 2006.
30. Kakhniashvili DG, Bulla LA Jr, and Goodman SR. The human eryth-rocyte proteome: analysis by ion trap mass spectrometry. Mol Cell Pro-teomics 3: 501–509, 2004.
31. Kim-Shapiro DB, Schechter AN, Gladwin MT. Unraveling the reac-tions of nitric oxide, nitrite, and hemoglobin in physiology and therapeu-tics. Arterioscler Thromb Vasc Biol 26: 697–705, 2006.
32. Kleinbongard P, Dejam A, Lauer T, Rassaf T, Schindler A, Picker O,Scheeren T, Godecke A, Schrader J, Schulz R, Heusch G, Schaub GA,Bryan NS, Feelisch M, Kelm M. Plasma nitrite reflects constitutive nitricoxide synthase activity in mammals. Free Radic Biol Med 35: 790–7962003.
33. Kleinbongard P, Schulz R, Rassaf T, Lauer T, Dejam A, Jax T,Kumara I, Gharini P, Kabanova S, Ozuyaman B, Schnurch HG,Godecke A, Weber AA, Robenek M, Robenek H, Bloch W, Rosen P,Kelm M. Red blood cells express a functional endothelial nitric oxidesynthase. Blood 107: 2943–2951, 2006.
34. Kosaka H, Seiyama A. Increased oxygen dissociation by nitric oxidefrom RBC. Adv Exp Med Biol 428: 349–354, 1997.
35. Linde T, Sandhagen B, Danielson BG, Wikstrom B. Impaired erythro-cyte fluidity during treatment of renal anaemia with erythropoietin. J IntMed 231: 601–606, 1992.
36. Loughrey CM, Young IS, Lightbody JH, McMaster D, McNamee PT,Trimble ER. Oxidative stress in haemodialysis. QJM 87: 679–683, 1994.
C387Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
37. Lutz HU. Innate immune and non-immune mediators of erythrocyteclearance. Cel Mol Biol (Noisy-le-Grand, France) 50: 107–116, 2004.
38. Lutz HU, Fasler S, Stammler P, Bussolino F, Arese P. Naturallyoccurring anti-band 3 antibodies and complement in phagocytosis ofoxidatively-stressed and in clearance of senescent red cells. Blood Cells14: 175–203, 1988.
39. Lutz HU, Liu SC, Palek J. Release of spectrin-free vesicles from humanerythrocytes during ATP depletion. I. Characterization of spectrin-freevesicles. J Cell Biol 73: 548–560, 1977.
40. Lutz HU, Stammler P, Fasler S, Ingold M, Fehr J. Density separationof human red blood cells on self forming Percoll gradients: correlationwith cell age. Biochim Biophys Acta 1116: 1–10, 1992.
41. Maarsingh H, de Boer J, Kauffman HF, Zaagsma J, Meurs H. Heparinnormalizes allergen-induced nitric oxide deficiency and airway hyperre-sponsiveness. Br J Pharmacol 142: 1293–1299, 2004.
42. Marzo F, Lavorgna A, Coluzzi G, Santucci E, Tarantino F, Rio T,Conti E, Autore C, Agati L, Andreotti F. Erythropoietin in heart andvessels: focus on transcription and signalling pathways. J Thromb Throm-bolysis 26: 183–187, 2008.
43. Mounta P, Kempc B, Powera D. Regulation of endothelial and myocar-dial NO synthesis by multi-site eNOS phosphorylation. J Mol Cell Cardiol42: 271–279, 2006.
44. Munzel T, Daiber A, Ullrich V, Mulsch A. Vascular consequences ofendothelial nitric oxide synthase uncoupling for the activity and expres-sion of the soluble guanylyl cyclase and the cGMP-dependent proteinkinase. Arterioscler Thromb Vasc Biol 25: 1551–1557, 2005.
45. Myssina S, Huber SM, Birka C, Lang PA, Lang KS, Friedrich B,Risler T, Wieder T, Lang F. Inhibition of erythrocyte cation channels byerythropoietin. J Am Soc Nephrol 14: 2750–2757, 2003.
46. Nemeth I, Turi S, Haszon I, Bereczki C. Vitamin E alleviates theoxidative stress of erythropoietin in uremic children on hemodialysis.Pediatr Nephrol 14: 13–17, 2000.
47. Ng T, Marx G, Littlewood T, Macdougall I. Recombinant erythropoietinin clinical practice. Postgrad Med J 79: 367–376, 2003.
48. Nguyen-Khoa T, Massy ZA, De Bandt JP, Kebede M, Salama L,Lambrey G, Witko-Sarsat V, Drueke TB, Lacour B, Thevenin M.Oxidative stress and haemodialysis: role of inflammation and duration ofdialysis treatment. Nephrol Dial Transplant 16: 335–340, 2001.
49. Rice L, Ruiz W, Driscoll T, Whitley CE, Tapia R, Hachey DL,Gonzales GF, Alfrey CP. Neocytolysis on descent from altitude: a newly
recognized mechanism for the control of red cell mass. Ann Int Med 134:652–656, 2001.
50. Risso A, Turello M, Biffoni F, Antonutto G. Red blood cell senescenceand neocytolysis in humans after high altitude acclimatization. Blood CellsMol Dis 38: 83–92, 2007.
51. Santhanam AVR, Katusic ZS. Erythropoietin and cerebral vascularprotection: role of nitric oxide. Acta Pharmacol Sinica 27: 1389–1394,2006.
52. Schafer FQ, Buettner GR. Redox environment of the cell as viewedthrough the redox state of the glutathione disulfide/glutathione couple.Free Radic Biol Med 30: 1191–1212, 2001.
53. Singel DJ, Stamler JS. Chemical physiology of blood flow regulation byred blood cells: the role of nitric oxide and S-nitrosohemoglobin. AnnuRev Physiol 67: 99–145, 2005.
54. Spivak JL, Avedissian LS, Pierce JH, Williams D, Hankins WD,Jensen RA. Isolation of the full-length murine erythropoietin receptorusing a baculovirus expression system. Blood 87: 926–937, 1996.
55. Stepuro TL, Zinchuk VV. Nitric oxide effect on the hemoglobin-oxygenaffinity. J Physiol Pharmacol 57: 29–38, 2006.
57. Trial J, Rice L. Erythropoietin withdrawal leads to the destruction ofyoung red cells at the endothelial-macrophage interface. Current PharmaDes 10: 183–190, 2004.
58. Vogel J, Kiessling I, Heinicke K, Stallmach T, Ossent P, Vogel O,Aulmann M, Frietsch T, Schmid-Schonbein H, Kuschinsky W, Gas-smann M. Transgenic mice overexpressing erythropoietin adapt to exces-sive erythrocytosis by regulating blood viscosity. Blood 102: 2278–2284,2003.
60. Wickrema A, Krantz S, Winkelmann J, Bondurant M. Differentiationand erythropoietin receptor gene expression in human erythroid progenitorcells. Blood 80: 1940–1949, 1992.
61. Wu G, Flynn NE, Flynn SP, Jolly CA, Davis PK. Dietary protein orarginine deficiency impairs constitutive and inducible nitric oxide synthe-sis by young rats. J Nutr 129: 1347–1354, 1999.
62. Yang Z, Kaye DM. Endothelial dysfunction and impaired L-argininetransport in hypertension and genetically predisposed normotensive sub-jects. Trends Cardiovasc Med 16: 118–124, 2006.
C388 Epo ACTIVATES eNOS IN MOUSE ERYTHROCYTES
AJP-Cell Physiol • VOL 297 • AUGUST 2009 • www.ajpcell.org
