THE rhGM-CSF–EPO HYBRID PROTEIN MEN 11300 INDUCES ANTI-EPO ANTIBODIES AND SEVERE ANAEMIA IN RHESUS MONKEYS

A recombinant human GM-CSF-EPO hybrid protein named MEN 11300 was administered biweekly for a total of 6 weeks to rhesus monkeys in order to evaluate its pharmacokinetic behaviour, tolerability and immunogenicity. In this primate species a strong antibody response was induced which neutralized the in vitro biological activity of human EPO while no antibody response could be detected against human GM-CSF. A severe drop in reticulocyte counts at approximately 2 weeks after initiation of treatment was followed by a dramatic decrease in the number of erythrocytes. No effects were observed on GM-CSF-dependent hematopoietic lineages and the clinical chemistry analyses did not reveal signs of general toxicity. Reticulocyte and erythrocyte counts started to recover 3–4 weeks after discontinuation of treatment in concert with a decline in anti-EPO antibody titres. Nevertheless, cell numbers remained below basal levels up to 50 days after the last MEN 11300 administration. Haematological impairment indicates that the administration to non-human primate of human EPO fused to human GM-CSF, induces neutralizing autoantibodies to the self EPO. Present data do not allow prediction of the immunogenic potential of the fusion protein in humans and a dose-escalating phase I study should be addressed to investigate the safety of the product.     

Pharmacokinetic and immunogenic behavior of three recombinant human GM-CSF-EPO hybrid proteins in cynomolgus monkeys

MEN 11300, MEN 11301, and MEN 11303 are three recombinant human hybrid proteins that, as has recently been described, induce in vitro erythroid differentiation. This article provides data on their pharmacokinetic and immunogenic behavior after repeated iv administration to cynomolgus monkeys at 0.8 or 1.6 μg/kg doses. Pharmacokinetic data, obtained after the first administration, showed that the half-life (t _1/2) and clearance (CL) values are dose dependent, with no significant differences among the three hybrid proteins. After the tenth administration, MEN 11300 and MEN 11301, both at high and low dose, and MEN 11303 at high dose were undetectable in plasma, whereas MEN 11303 at the lower dose showed no alteration in its pharmacokinetic profile. Immunologic analyses of plasma provided an explanation for this different pharmacokinetic behavior. In fact, plasma samples from animals treated repeatedly with MEN 11300 and MEN 11301 showed specific antibody formation in response to both the high- and the low-dose regimens. These antibodies exerted in vitro a strong neutralizing activity of the hybrid proteins, with a predominant specificity for the erythropoietin (EPO) portion. By contrast, MEN 11303 at the lower dose did not induce a detectable antibody response whereas the antibodies observed on the high-dose regimen did not exert neutralizing activity against the hybrid proteins nor against granulocyte-macrophage colony-stimulating factor (GM-CSF) or EPO. Hematologic parameters were not affected by the treatments, thus indicating that the anti-EPO neutralizing antibody response does not cross react with the endogenous monkey cytokine. The overall immunogenicity data suggest that among the three fusion proteins, MEN 11303 could have a lower immunogenic potential.     

Structural characterization and independent folding of a chimeric glycoprotein comprising granulocyte-macrophage colony stimulating factor and erythropoietin sequences

MEN 11300 is a hybrid glycoprotein of 297 amino acids obtained by fusion of the cDNA encoding GM-CSF with the cDNA encoding EPO followed by transfection of the hybrid gene into CHO cells. The oligonucleotide construct incorporated a spacing sequence between the two individual cDNAs which encodes eight amino acids constituting a linker peptide intended to separate the GM-CSF and EPO moieties. The recombinant MEN 11300 protein was submitted to a detailed structural characterization including the verification of the entire amino acid sequence, the assignment of the disulfide bridges pattern, the identification of the glycosylation sites and the definition of the glycosidic moiety, including site-specificity. Partial processing of the C-terminal Arg residue and the occurrence of N-glycosylation sites at Asn27, Asn155, Asn169, Asn214 were established. Moreover, O-glycosylation at Ser257 and at the N-terminal region was also detected. A large heterogeneity was observed in the N-glycans due to the presence of differently sialylated and fucosylated branched complex type oligosaccharides whereas O-linked glycans were constituted by GalGalNAc chains with a different number of sialic acids. The disulfide bridges pattern was established by direct FABMS analysis of the proteolytic digests or by ESMS analysis of HPLC purified fractions. Pairing of the eight cysteine residues resulted in Cys54-Cys96, Cys88-Cys121, Cys138-Cys292, and Cys160-Cys164. This S-S bridges pattern is identical to that occurring in the individual natural GM-CSF and EPO, thus showing that the two protein moieties in MEN 11300 can independently acquire their native three-dimensional structure.      

Bovine αs1-casein gene sequences direct high level expression of human granulocyte-macrophage colony-stimulating factor in the milk of transgenic mice

The generation is reported of transgenic mice expressing human granulocyte-macrophage colony-stimulating factor (GM-CSF) or human erythropoietin (EPO) under the control of bovine s1- casein regulatory sequences. GM-CSF expression was specific to the mammary gland, and levels of human GM-CSF in transgenic mouse milk were in the range of mg ml–1. The specific activity of the milk GM-CSF was similar to that of the recombinant protein produced in Escherichia coli, and the glycosylation-derived size heterogeneity corresponded to that of the native human protein. In spite of the identical bovine regulatory sequences of the fusion genes, the levels of human EPO in transgenic mouse milk were 103--106 times lower than those of GM-CSF, ranging from 0.003 to 3 g ml–1. There appeared to be a positive correlation between the amount of EPO in the milk of lactating females and blood haematocrit values. In view of this, other type of constructs should be used to achieve more efficient EPO expression and to circumvent concomitantly-occurring adverse effects. In contrast, the high-level production of recombinant GM-CSF, its resemblance to the native mammalian protein, and mild adverse consequences of transgene expression imply that the current construct could be used for generation of larger GM-CSF transgenic anim als to produce this protein in quantities sufficient for therapeutic purposes     

人源抗EPO基因工程抗体重链可变区的克隆和鉴定

利用噬菌体抗体显示技术筛选 EPO的人源抗体 ,得到了抗 EPO的人源抗体的重链基因。此抗体基因在噬菌体表面呈现的抗体分子具有良好的抗体活性和特异性。为制备完整的、具有更高亲和力的抗体打下了基础。     

Kinetics of development and characteristics of antibodies induced in cancer patients against yeast expressed rDNA derived granulocyte macrophage colony stimulating factor (GM-CSF)

We have determined the presence and kinetics of granulocyte macrophage colony stimulating factor (GM-CSF) antibodies induced after repeated administration of a yeast expressed GM-CSF product in prostate cancer patients with minimal recurrent disease using a panel of assays for detection and characterization of antibodies. Results showed that all 15 prostate cancer patients treated with GM-CSF developed GM-CSF reactive antibodies during the course of therapy. Most patients (87%) developed GM-CSF reactive antibodies within 3 months while in other patients (13%), these antibodies were induced after additional cycles of GM-CSF treatment. For most patients, the timing of occurrence of these antibodies was the same regardless of whether the ELISA or surface plasmon resonance (SPR) assays were used for detection. However, in two patients, the recognition of GM-CSF reactive antibodies by SPR assays preceded their detection by ELISA. A significant number of patients (n=9, 60%) developed GM-CSF antibodies which neutralized the biological activity of GM-CSF in vitro in a cell-line based bioassay. These antibodies also recognized GM-CSF protein from different expression systems including the non-glycosylated protein from E. coli indicating that the antibody response is directed towards the amino acid backbone of the protein. A significant effect of GM-CSF antibodies on PSA modulation was not observed in this small cohort of patients despite an alteration in PSA levels in some treated patients. The study design used here did not allow conclusions regarding the relationship between neutralizing antibodies and the PSA levels which were used as a marker for clinical outcome. Implementation of a clinical strategy which permits monitoring for antibody development and for levels of a relevant pre-determined clinical marker at appropriate time-points is necessary for assessing the impact of antibody development on the therapeutic efficacy of the protein.     

IL-6 plays an essential role in neutrophilia under inflammation

In the present study, we explored the involvement of interleukin-6 (IL-6) in neutrophilia under inflammatory conditions. The neutrophil count in the peripheral blood was high in arthritic monkeys, and anti-IL-6 receptor antibody reduced neutrophil counts to normal levels. IL-6 injection into normal monkeys significantly increased neutrophil counts in the blood 3h after injection. The expression of cluster of differentiation (CD) 162 on circulating neutrophils was reduced by IL-6 injection. IL-6 treatment in vitro did not affect CD162 expression on neutrophils from human blood. In IL-6-treated monkeys, IL-8 and granulocyte-macrophage colony-stimulating factor (GM-CSF) levels in plasma were clearly elevated. IL-8 and GM-CSF treatment in vitro reduced cell-surface CD162 expression on human neutrophils, and moreover, increased soluble CD162 expression in the cell supernatant. The addition of IL-6 into human whole peripheral blood induced IL-8 production and reduced CD162 expression on neutrophils. Furthermore, IL-8 and GM-CSF augmented mRNA expression of a disintegrin and metalloprotease like domain 10 (ADAM10) in neutrophils. Knock-down of ADAM10 by siRNA in neutrophil-like HL-60 cells partially reversed the expression of CD162 reduced by GM-CSF and IL-8 on HL-60 cells. In conclusion, IL-6 induced neutrophilia and reduced CD162 expression on neutrophils in inflammation.     

GM-CSF triggers a rapid,glucose dependent extracellular acidification by TF-1 cells: Evidence for sodium/proton antiporter and PKC mediated activation of acid production

The extracellular acidification rate of the human bone marrow cell line, TF-1, increases rapidly in response to a bolus of recombinant granulocyte-macrophage colony stimulating factor (GM-CSF). Extracellular acidification rates were measured using a silicon microphysiometer. This instrument contains micro-flow chambers equipped with potentiometric sensors to monitor pH. The cells are immobilized in a fibrin clot sandwiched between two porous polycarbonate membranes. The membranes are part of a disposable plastic “cell capsule” that fits into the microphysiometer flow chamber. The GM-CSF activated acidification burst is dose dependent and can be neutralized by pretreating the cytokine with anti-GM-CSF antibody. The acidification burst can be resolved kinetically into at least two components. A rapid component of the burst is due to activation of the sodium/proton antiporter as evidenced by its elimination in sodium-free medium and in the presence of amiloride. A slower component of the GM-CSF response is a consequence of increased glycolytic metabolism as demonstrated by its dependence on D-glucose as a medium nutrient. Okadaic acid (a phospho-serine/threonine phosphatase inhibitor), phorbol 12-myristate 13-acetate (PMA, a protein kinase C (PKC) activator), and ionmycin (a calcium ionophore) all produce metabolic bursts in TF-1 cells similar to the GM-CSF response. Pretreatment of TF-1 cells with PMA for 18 h resulted in loss of the GM-CSF acidification response. Although this treatment is reported to destroy protein kinase activity, we demonstrate here that it also down-regulates expression of high-affinity GM-CSF receptors on the surface of TF-1 cells. In addition, GM-CSF driven TF-1 cell proliferation was decreased after the 18 h PMA treatment. Short-term treatment with PMA (1–2h) again resulted in loss of the GM-CSF acidification response, but without a decrease in expression of high-affinity GM-CSF receptors. Evidence for involvement of PKC in GM-CSF signal transduction was obtained using calphostin C, a specific inhibitor of PKC, which inhibited the GM-CSF metabolic burst at a subtoxic concentration. Genistein and herbimycin A, tyrosine kinase inhibitors, both inhibited the GM-CSF response of TF-1 cells, but only at levels high enough to also inhibit stimulation by PMA. These results indicate that GM-CSF activated extracellular acidification of TF-1 cells is caused by increases in sodium/proton antiporter activity and glycolysis, through protein kinase signalling pathways which can be both activated and down-regulated by PMA. © 1993 Wiley-Liss, Inc.     

粒细胞-巨噬细胞集落刺激因子/白细胞介素-3融合蛋白的表达研究

利用PCR扩增得到粒细胞巨噬细胞集落刺激因子(GM-CSF)、白细胞介素-3(IL-3)完整基因片段,将其分别克隆至pGEM-T,构建成GMCSF/IL-3融合蛋白基因,DNA序列与设计预期一致。将得到的融合蛋白基因克隆至T7RNA聚合酶表达载体pT7zz,得到表达质粒pFu,经转化至表达宿主E.coli BL21(DE3),在IPTG诱导下获得融合蛋白目的产物的直接表达。经SDS-PAGE电泳鉴定扫描分析,目的基因产物表达量占菌体总蛋白量的30%以上,目的基因表达产物以包涵体的形式表达。Westernblot鉴定表明,该表达产物可以与GM-CSF抗体及IL-3抗体特异性结合。目的基因表达产物经过包涵体变性、透析复性及柱层析纯化,用GM-CSF、IL-3依赖细胞株TF-1检测,具有明显的生物学活性。     

Use of enzyme microassay to detect eosinophil potentiating activity of cytokines in sheep

Abstract Microplate assays for eosinophil peroxidase (EPO) and arylsulphatase (EAS) have been established as indices of eosinophil survival/proliferation in sheep bone marrow cell (SBMC) cultures. Cell specificity was confirmed using density-fractionated and differentially depleted SBMC populations. Several recombinant cytokines including interleukins 3 (IL-3) and 5 (IL-5), and granulocyte macrophage-colony stimulating factor (GM-CSF), but not macrophage-CSF (M-CSF), had demonstrable eosinophil-potentiating activity on the basis of enhanced EPO and EAS activities in treated, compared with untreated, SBMC cultures. Effects of IL-5 were abrogated in the presence of a specific monoclonal anti-IL-5 antibody. The results indicate that measurement of EPO and EAS in cultured SBMC offers a simple and effective method for detecting eosinophil potentiating activity in the ovine.     

Role of specific IgE antibodies in peroxidase (EPO) release from human eosinophils

After the demonstration of cytophilic IgE immunoglobulins (Ig) on human blood and lung eosinophils, their role in cell activation was studied by eosinophil peroxidase (EPO) assay. Hypodense human eosinophils from filariasis-infected patients were activated by anti-human Ig or various antigens. A selective release of EPO occurred after incubation with anti-human IgE, but not with anti-human IgG. The activation by antigens showed a strict antibody specificity of cytophilic IgE antibodies. The direct involvement of IgE antibodies in activation by the specific antigen was evidenced by inhibition experiments with aggregated human IgE myeloma protein. Circulating IgE antibodies exhibiting the same specificity and able to induce EPO release were detected in the sera from filariasis patients by a passive sensitization assay. Only the hypodense eosinophils were able to release EPO after IgE-dependent activation both in the direct assay and in the passive sensitization test, confirming the functional heterogeneity of human eosinophils. These results suggest that the interaction between IgE antibodies and human eosinophils can play a role both in protective immunity and pathology by releasing active pharmacologic mediators.     

Endothelium-derived GM-CSF influences expression of oncostatin M

During and after transendothelial migration, neutrophils undergo a number of phenotypic changes resulting from encounters with endothelium-derived factors. This report uses an in vitro model with human umbilical vein endothelial cells and isolated human neutrophils to examine the effects of two locally derived cytokines, granulocyte (G)-macrophage (M) colony-stimulating factor (GM-CSF) and G-CSF, on oncostatin M (OSM) expression. Neutrophils contacting activated HUVEC expressed and released increased amounts of oncostatin M (OSM), a proinflammatory cytokine known to induce polymorphonuclear neutrophil adhesion and chemotaxis. Neutrophil transendothelial migration resulted in threefold higher OSM expression and protein levels compared with nontransmigrated cells. Addition of anti-GM-CSF neutralizing antibody reduced OSM expression level but anti-G-CSF was without effect. GM-CSF but not G-CSF protein addition to cultures of isolated neutrophils resulted in a significant increase in OSM protein secretion. However, inhibition of β(2) integrins by neutralizing antibody significantly reduced GM-CSF-induced OSM production indicating this phenomenon is adhesion dependent. Thus cytokine-stimulated endothelial cells can produce sufficient quantities of GM-CSF to influence in an adhesion-dependent manner, the phenotypic characteristics of neutrophils resulting in the latter's transmigration. Both transmigration and adhesion phenomenon lead to increased production of OSM by neutrophils that then play a major role in inflammatory response.     

Treatment with GM-CSF and IL-2 in patients with metastatic colorectal carcinoma induced high serum levels of neopterin and sIL-2R,an indicator of immune suppression

Cytokines may enhance the effect of therapeutic monoclonal antibodies (mAb). Granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-2 (IL-2) have been shown to increase ADCC levels. GM-CSF may augment the induction of an idiotypic network response (anti-tumour immunity). The clinical anti-tumour effect of a combination of mouse mAb17-1A-1A [anti-colorectal carcinoma (CRC)], and GM-CSF was, however, not enhanced by the addition of IL-2. In the present study, some immune functions considered to be involved in mAb-mediated tumour cell killing were analysed in patients receiving GM-CSF and GM-CSF/IL-2 respectively together with the mAb17-1A-1A. Ten patients received mAb17-1A and GM-CSF, and ten patients mAb17-1A with GM-CSF and IL-2. During a 10- day cytokine treatment period, a significantly higher increase in white blood cell counts was noted in the GM-CSF/IL-2 treatment group as compared to GM-CSF-treated patients. In the GM-CSF/IL-2 group, significantly higher serum concentrations of neopterin and soluble IL-2 receptor (sIL-2R) respectively were induced as compared to GM-CSF-treated patients. However, the ADCC of peripheral blood mononuclear cells (PBMC) against a CRC cell line was significantly higher in the GM-CSF group than in the GM-CSF/IL-2 group. The frequencies of patients developing human anti-mouse antibodies (HAMA) and anti-idiotypic antibodies were the same in both groups, while serum concentrations were significantly lower in the GM-CSF/IL-2 group as compared to the GM-CSF group. GM-CSF/IL-2 therapy seems to induce an immune suppressive stage compared to GM-CSF alone affecting cytotoxic mononuclear cells and B cells, which might be mediated through the neopterin metabolic pathway or other inducible immune suppressive factors such as reactive oxygen and nitrogen intermediates.     

Erythropoietin and hypoxia increase erythropoietin receptor and nitric oxide levels in lung microvascular endothelial cells

Acute lung exposure to low oxygen results in pulmonary vasoconstriction and redistribution of blood flow. We used human microvascular endothelial cells from lung (HMVEC-L) to study the acute response to oxygen stress. We observed that hypoxia and erythropoietin (EPO) increased erythropoietin receptor (EPOR) gene expression and protein level in HMVEC-L. In addition, EPO dose- and time-dependently stimulated nitric oxide (NO) production. This NO stimulation was evident despite hypoxia induced reduction of endothelial NO synthase (eNOS) gene expression. Western blot of phospho-eNOS (serine1177) and eNOS and was significantly induced by hypoxia but not after EPO treatment. However, iNOS increased at hypoxia and with EPO stimulation compared to normal oxygen tension. In accordance with our previous results of NO induction by EPO at low oxygen tension in human umbilical vein endothelial cells and bone marrow endothelial cells, these results provide further evidence in HMVEC-L for EPO regulation of NO production to modify the effects of hypoxia and cause compensatory vasoconstriction.     

A recombinant human G-CSF/GM-CSF fusion protein from E. coli showing colony stimulating activity on human bone marrow cells

Granulocyte-Macrophage colony stimulating factor (GM-CSF) and Granulocyte colony stimulating factor (G-CSF) are cytokines involved in the differentiation of bone marrow progenitor cells into myeloid cells. They also activate mature myeloid cells to mediate a variety of antimicrobial activities and inflammatory responses. Recombinant GM-CSF and G-CSF proteins have been used to treat various diseases including cancer and hematopoietic diseases and to isolate peripheral blood progenitor cells for bone marrow transplantation. A plasmid construct expressing recombinant human G-CSF/GM-CSF fusion protein has now been prepared by linking the human G-CSF and GM-CSF coding regions and the recombinant fusion protein has been successfully expressed in E. coli. The recombinant human G-CSF/GM-CSF fusion protein was extracted and purified from the cellular inclusion and refolded into the biologically active form to show colony stimulating activity. The recombinant fusion protein exhibited colony stimulating activity on human bone marrow cell cultures, indicating that the linkage of GM-CSF and G-CSF by a linker peptide may not interrupt activities of the cytokines in the fusion protein. The colony forming unit of the fusion protein was also higher than those of the cultures treated with the same molar numbers of the recombinant human GM-CSF and G-CSF separately, which suggests that the fusion protein presumably retains both G-CSF and GM-CSF activities.     

Production of IL-13 in spleen cells by IL-18 and IL-12 through generation of NK-like cells

Treatment of Nylon wool-passed cells (NWC) prepared from the spleen of C57BL/6 mice with IL-18 and IL-12, but not with IL-18 alone, resulted in induction of IFN-gamma, a Th1 cytokine, and GM-CSF at 24 h, and IL-13, a Th2 cytokine at 72 h. The induction of IL-13 was suppressed by anti-GM-CSF antibody, indicating involvement of GM-CSF in IL-13 production. When NWC incubated with IL-18 and IL-12 for 72 h ("primary treatment") were treated again with the same cytokines ("secondary treatment"), IL-13 was induced much more quickly than observed in the primary treatment. Flow cytometric analysis of NWC after the primary treatment showed marked increases in the CD4(-)CD8(-) non-T cell population bearing CD25(+), CD45RB(super high) and CD122(+). These cells were positive for CD49b but negative for NK1.1, indicating that they were not typical but NK-like cells. The NK-like cells produced IL-13 in response to the treatment with IL-18 alone, indicating that the generation of these cells in the primary treatment likely accounts for the quick production of IL-13 in the secondary treatment. These results show that IL-18 and IL-12 generates the NK-like cells in NWC by a process mediated by GM-CSF that are ready for producing IL-13.     

Inoculation of baboons and macaques with simian immunodeficiency virus/Mne, a primate lentivirus closely related to human immunodeficiency virus type 2

A primate lymphotropic lentivirus was isolated on the human T-cell line HuT 78 after cocultivation of a lymph node from a pig-tailed macaque (Macaca nemestrina) that had died with malignant lymphoma. This isolate, originally designated M. nemestrina immunodeficiency virus (MnIV) and now classified as simian immunodeficiency virus (SIV/Mne), was inoculated intravenously into three juvenile rhesus monkeys (Macaca mulatta), three juvenile pig-tailed macaques (M. nemestrina), and two juvenile baboons (Papio cynocephalus). All six macaques became viremic by 3 weeks after inoculation, whereas neither of the baboons developed viremia. One pig-tailed macaque died at 15 weeks with suppurative peritonitis secondary to ulcerative, necrotizing colitis. Immunologic abnormalities included a marked decrease in CD4+ peripheral blood lymphocytes. Although five macaques mounted an antibody response to SIV/Mne, the animal that died at 15 weeks remained antibody negative. Three other macaques (two rhesus and one pig-tailed) died 66 to 87 weeks after inoculation after exhibiting progressive weight loss, anemia, and diarrhea. Histopathologic findings at necropsy included various manifestations of immune deficiency, nephropathy, subacute encephalitis, pancreatitis, adenocarcinoma, and lymphoid atrophy. SIV/Mne could be readily isolated from the spleens and lymph nodes of all necropsied macaques, and from the cerebrospinal fluid, brains, bone marrow, livers, and pancreas of some of the animals. SIV antigens were localized by avidin-biotin immunohistochemistry to pancreatic islet cells and to bone marrow endothelial cells. The data suggest that African baboons may be resistant to infection by SIV/Mne, whereas Asian macaques are susceptible to infection with this pathogenic primate lentivirus.     

Up-regulated Endogenous Erythropoietin/Erythropoietin Receptor System and Exogenous Erythropoietin Rescue Retinal Ganglion Cells after Chronic Ocular Hypertension

Aims Recent studies have showed that erythropoietin (EPO) is a neuroprotectant for central nerve system neurons in addition to being a hematopoietic cytokine in response to hypoxia. In this study, we investigate the role of the EPO/EPO receptor (EPOR) system in the rat retina after ocular hypertension injury that mimics glaucoma. Methods Elevated intraocular pressure was induced by laser coagulation of the episcleral and limbal veins. Expression of EPO and EPOR in the normal and glaucomous retinas was investigated by immunohistochemistry and Western blot. To examine the effects of endogenous EPO on the survival of retinal ganglion cells (RGCs) subjected to hypertensive injury, soluble EPOR was directly injected into the vitreous body. Recombinant human EPO was both intravitreally and systemically administrated to study the effect of exogenous EPO on the survival of RGCs after ocular hypertension injury. Results Immunohistochemistry studies identified Müller cells as the main source of EPO in the normal retina. Expression of EPO and EPOR proteins was increased significantly 2 weeks after ocular hypertension. RGCs, amacrine and bipolar cells all demonstrated an increased expression of EPOR after ocular hypertension. Neutralization of endogenous EPO with soluble EPOR exacerbated ocular hypertensive injury, suggesting a role of the EPO/EPOR system in the survival of RGCs after injury. Similarly, topical and systemic administration of recombinant human EPO rescues RGCs after chronic ocular hypertension. Conclusions These results indicate that an endogenous EPO/EPOR system participates in intrinsic recovery mechanisms after retina injury and RGCs might be rescued by exogenous administration of EPO.