RETINAL ischemia is associated with vascular diseases that may result in significant visual loss. The retina’s blood and oxygen supply is decreased with atherosclerosis, diabetic retinopathy, central retinal artery or vein occlusion, and sickle cell retinopathy. An endogenous protective capacity in the rat retina, produced by ischemic preconditioning (IPC) can induce tolerance to retinal ischemia.1IPC, a brief period of ischemia, stimulates endogenous mechanisms that provide protection in the event of subsequent ischemia. IPC, and the subsequent ischemia, did not affect the contralateral retina.2,3Enhanced knowledge of the mechanisms of IPC could lead to therapeutic strategies for retinal ischemic injury, or ischemia in other central nervous system regions.2–4Earlier studies from our laboratory, some of which have been confirmed by others, indicated the roles in this neuroprotection of adenosine, protein kinase C, heat shock protein 27 (HSP27), reactive oxygen species, nitric oxide synthase, the opening of mitochondrial adenosine triphosphate–sensitive K+ channels, mitogen-activated protein kinases, and decreased retinal cell apoptosis.2,3,5–9Despite these studies, the molecular basis for IPC remains incomplete.
A potential signaling pathway in retinal IPC is the hematopoietic cytokine erythropoietin. Intriguingly, erythropoietin, in addition to its hematopoietic effects, protects neurons from ischemic damage, and it may decrease neuronal injury in stroke.10We previously demonstrated that retinal ischemia increased retinal protein levels of erythropoietin and decreased levels of erythropoietin receptor (EPO-R). Systemic injection of erythropoietin protected retinal neurons from ischemic injury, whereas blockade of erythropoietin by intravitreal administration of soluble EPO-R (sEPO-R) worsened recovery.11In mouse or rat models, erythropoietin protected against light-induced retinal injury and axotomy-induced neurodegeneration.12–16Watanabe et al. 17found elevated erythropoietin levels in the vitreous in diabetic retinopathy, and Morita et al. 18demonstrated that hyperoxia-normoxia in a murine retinopathy of prematurity model induced neovascularization in wild-type mice, but not in hypoxia-inducible factor-1α–like factor kD/kD mice, where erythropoietin levels were decreased.
In this study, we examined the hypothesis that erythropoietin was an essential signaling molecule in retinal IPC via production of increased levels of erythropoietin. We examined potential downstream effectors to erythropoietin in this ischemic neuroprotection.
Materials and Methods
Procedures7,8conformed to the Association for Research in Vision and Ophthalmology Resolution on the Use of Animals in Research and were approved by our Animal Care Committee (Division of Biologic Sciences, University of Chicago, Chicago, Illinois). Sprague-Dawley rats (200–250 g) from Harlan (Indianapolis, IN) were maintained on a 12 h on/12 h off light cycle. Rats were anesthetized with 450 mg/kg intraperitoneal chloral hydrate. For baseline and postischemic follow–up electroretinograms, rats were injected intraperitoneally with 35 mg/kg ketamine (Parke–Davis, Morris Plains, NJ) and 5 mg/kg xylazine (Miles, Shawnee Mission, KS). Corneal analgesia was achieved with 0.5% proparacaine (Allergan, Irvine, CA). Pupils were dilated with 0.5% tropicamide (Alcon, Ft Worth, TX) and cyclomydril (0.2% cyclopentolate HCl and 1% phenylephrine HCl, Alcon). Body temperature was maintained at 36.5–37.0°C with a servo–controlled heating blanket (Harvard Apparatus, Natick, MA).
For preconditioning, intraocular pressure was increased to 160 mmHg for 8 min using a pressurized 1000-ml plastic bag of sterile normal saline (Baxter, North Chicago, IL) connected to a 30-gauge needle in the center of the eye’s anterior chamber. For ischemia, performed 24 h after preconditioning, the intraocular pressure was increased to 110 mmHg for 45 min.9
Procedures have been described in detail previously.7,8In brief, the scotopic electroretinogram was recorded from rats dark-adapted for at least 1 h by using platinum needle electroencephalogram electrodes (Grass, Providence, RI) in contact with the corneal surfaces of both eyes. Responses to 10-μs white light flashes from a Nicolet Ganzfeld stimulator (Madison, WI) with the rat’s head centered 6 inches away were recorded on a Nicolet Spirit 486. Data are the average of three flashes at least 2 min apart. Wave amplitudes 7 days after ischemia were measured and reported as a percentage of the baseline, nonischemic wave amplitude.
Procedures were those we used in previous studies.8,9,11Briefly, retinas were solubilized in 9 M urea, 4% Nonidet P-40, and 2% 2-mercaptoethanol (pH 9.5). Protease inhibitor cocktail (P8340; Sigma, St Louis, MO) consisting of 4-(2-aminoethyl) benzenesulfonyl fluoride, pepstatin A, bestatin, leupeptin, and E-64 prevented protease activity.
Equal amounts of protein per lane (40 μg) were loaded onto gels for sodium dodecyl sulfate polyacrylamide gel electrophoresis (4–20% or 16%; Invitrogen, Carlsbad, CA) or ν-polyacrylamide gel electrophoresis 4–12% bis-Tris gels (Invitrogen). Proteins were electroblotted to polyvinylidene difluoride (Immobilon-P; Millipore, Bedford, MA). Nonspecific binding was blocked with 5% nonfat dry milk in Tween-Tris-buffered saline. Membranes were incubated overnight at 4°C with rabbit polyclonal anti-erythropoietin (1:2000; Santa Cruz Biotechnology, Santa Cruz, CA), rabbit polyclonal anti-EPO-R (1:500; Santa Cruz), rabbit polyclonal antiphosphorylated Ser82 HSP27 (1:300; Upstate, Lake Placid, NY), rabbit polyclonal anti-HSP25 (1:500; Assay Design, Ann Arbor, MI), rabbit polyclonal antiphosphorylated Thr202/Tyr204 extracellular-signal-regulated kinase (ERK, 1:2000; Cell Signaling, Danvers, MA), rabbit polyclonal anti-ERK (1:1000; Cell Signaling), rabbit polyclonal antiphosphorylated Ser473 protein kinase B (Akt, 1:300; Cell Signaling), and rabbit polyclonal anti-Akt (1:500; Cell Signaling).
Anti-rabbit horseradish peroxidase-conjugated (goat IgG; Jackson Immuno-Research, West Grove, PA) secondary antibody was applied at 1:20,000. Chemiluminescence was developed with Super Signal West Pico (Pierce, Rockford, IL). Protein bands were digitally imaged with a CCDBIO 16SC Imaging System (Hitachi Genetic Systems/MiraiBio, Alameda, CA) and semiquantitated by densitometry (Gene Snap and Gene Tools; Hitachi). Equal protein loading was checked by Ponceau S red staining and by immunoblotting with rabbit polyclonal anti-opsin (1:250; Santa Cruz Biotechnology) or with antitotal Akt, antitotal ERK, or anti-HSP25 antibodies.
Peptide Competition Assay
To confirm specificity, EPO-R primary antibody was incubated overnight with either EPO-R blocking peptide (50-X vs. antibody by weight; Santa Cruz) in phosphate-buffered saline (PBS) or PBS alone. Supernatants of the mixtures, collected after centrifugation (10,000 rpm for 15 min at 4°C) were then brought up to a volume of 2 ml in 5% dry milk in PBS for subsequent Western blot analysis on whole retinal homogenates.
Fluorescent TUNEL and Imaging
Fluorescent terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) was performed, and cells were identified using a Fluorescein FragEL DNA fragmentation detection kit (Calbiochem, La Jolla, CA) on 10-μm-thick retinal cryosections as described previously.9,19We used a fluorescence microscope (inverted Olympus IX81; Olympus, Center Valley, PA), a Fast firewire Retiga EXi chilled CCD camera (QImaging, Pleasanton, CA), and a 40× oil lens. Excitation/dichroic/emission settings were 480/40 nm–505 LP–535/30 nm for rhodamine, and 530–550 nm–570 Dm–590 LP for fluorescein.
Eyes enucleated on the seventh day after ischemia were immediately fixed in 4% paraformaldehyde in PBS for 48 h, transferred to buffered formalin for 24 h, and stored in PBS at 4°C. Paraffin-embedded 5-μm-thick sections were stained with hematoxylin and eosin and examined by light microscopy, and retinal ganglion cell counts were quantified as described.1,9,11,19
Whole retinal homogenates were collected 1, 6, and 24 h after IPC for Western blotting. To test the role of EPO-R in IPC, either 2 μl of 20-ng soluble EPO-R (R&D Systems, Minneapolis, MN) or 20-ng denatured soluble EPO-R (boiled for 30 min then allowed to cool to body temperature) was injected into the mid-vitreous 15 min before IPC, as described previously.11Doses injected into the vitreous were estimated assuming an approximately 30-μl vitreous volume as determined from earlier experiments; thus the concentration after 2-μl injection was assumed to be diluted 15-fold.8In addition, previous experiments indicated that this intravitreal dose of EPO-R was effective in blunting effects of EPO and was not toxic.11To examine the relationship between EPO and the downstream activated phosphorylated ERK, HSP27, and Akt, we injected denatured soluble EPO-R (dEPO-R) or soluble EPO-R (sEPO-R) into the vitreous 15 min before IPC, and we removed retinas 24 h later and examined them by Western blot.
Data (mean ± SEM) were analyzed as previously described, with ANOVA and post hoc t testing using Stata version 6.0 (College Station, TX).7,8Results between paired eyes were compared using paired t tests and between time-matched groups from different animals using an unpaired t test. P < 0.05 was used for statistical significance.
IPC Effects on Erythropoietin and Erythropoietin Receptor Protein Levels
IPC did not significantly change erythropoietin protein levels at 1 h (100 ± 5%; n = 4), 6 h (108 ± 14%; n = 4), and 24 h (96 ± 6%; n = 4) after IPC compared to normal paired-control eyes (fig. 1). However, IPC significantly increased the levels of EPO-R (fig. 2). Western blot analysis of EPO-R resulted in multiple bands confirmed to be EPO-R protein by peptide competition assay (fig. 2C). The 200-kD (139 ± 15%, P = 0.03; n = 7) and the 115-kD (127 ± 12%, P = 0.05; n = 7) EPO-R bands significantly increased at 24 h after IPC compared to normal paired-control eyes. There were no significant changes in the EPO-R protein levels at both 1 and 6 h after IPC.
sEPO-R Blocks IPC
Intravitreal injection of sEPO-R 15 min before IPC significantly attenuated IPC neuroprotection (a-wave 31 ± 12%, P = 0.007; b-wave 17 ± 7%, P = 0.006; n = 5; fig. 3) compared to the control intravitreal injection of dEPO-R (a-wave 86 ± 6%; b-wave 51 ± 7%; n = 12). Moreover, histologic examination of retinal sections showed that sEPO-R significantly decreased retinal ganglion cell numbers (66.4 ± 11.2% of normal; P = 0.04; n = 3), whereas denatured sEPO-R did not (86.5 ± 17.6% of normal; n = 5; table 1). Fluorescent TUNEL showed that sEPO-R significantly increased apoptotic retinal ganglion cells in ischemic retina (10.7 ± 2.9% vs. 1.7 ± 1.0% in normal; P = 0.03; n = 4), but dEPO-R had no effect (ischemic retina 5.6 ± 2.4% vs. 3.1 ± 2.4% in normal; P = 0.49; n = 4; fig. 4and table 2).
EPO Activates Downstream Effectors
We examined the activation of downstream proteins by injection of dEPO-R or sEPO-R 15 min before IPC. Western blot analysis of retina 24 h after IPC showed that phosphorylated ERK and HSP27, but not Akt, increased in dEPO-R but not in sEPO-R–treated eyes (fig. 5). The 42-kD phosphorylated ERK (corrected for total ERK protein levels) significantly increased to 152 ± 19% of normal (P = 0.05; n = 3) with injection of dEPO-R before IPC compared to 98 ± 17% (P = 0.91; n = 4) for sEPO-R. The 44-kD phosphorylated ERK protein did not change (119 ± 10% of normal for dEPO-R and 105 ± 7% for sEPO-R). Phosphorylated HSP27 (corrected for HSP25 protein levels) significantly increased to 165 ± 11% of normal (P = 0.01; n = 3) with injection of dEPO-R before IPC compared to 92 ± 23% (P = 0.73; n = 4) for sEPO-R. The protein levels of phosphorylated Akt did not differ between the two treatment groups. Phosphorylated Akt protein levels (corrected for total Akt protein levels) significantly increased to 193 ± 11% of normal (P = 0.01; n = 3) with injection of dEPO-R before IPC, and the levels trended to increase to 204 ± 50% (P = 0.08; n = 4) for sEPO-R.
Our results show the role of erythropoietin and EPO-R in retinal neuroprotective pathways. We demonstrated that (1) EPO-R but not erythropoietin protein levels are increased with IPC, (2) soluble EPO-R attenuates the protective effects of IPC on retinal function, retinal ganglion cell loss, and apoptosis after ischemia, and (3) activation of ERK and HSP27, potential neuroprotective pathways, occur downstream of upregulation of the EPO-R after IPC.
In the current study, no significant change in erythropoietin protein expression was observed after IPC and 1, 6, or 24 h of reperfusion. These findings contrast with previously published studies in mouse retina and brain. In mouse retina, erythropoietin messenger ribonucleic acid (mRNA) expression significantly increased after hypoxic preconditioning and persisted for at least 1h after reperfusion.13In mouse brain, erythropoietin mRNA significantly increased after hypoxic preconditioning and 1 h of reperfusion, returning to baseline by 4 h of reperfusion. Erythropoietin protein levels were elevated at 1 h after preconditioning and increased further by 48 h.20The contrasting results observed for erythropoietin protein after ischemic and hypoxic preconditioning may be related to differences in the two preconditioning models. Hypoxic preconditioning was performed for at least 5 h before reperfusion in both mouse studies, whereas IPC was only performed for 8 min in our study. The molecular response to hypoxia involves the upregulation of hypoxia inducible factor-1, which regulates the transcription of hypoxia-responsive genes, including erythropoietin. IPC is characterized in part by hypoxia, but other components of this model may also influence the expression of erythropoietin. Alternatively, hypoxic preconditioning may have a broader effect on a wider range of retinal cell types, whereas IPC may affect fewer retinal cells. For example, a change in erythropoietin expression in a select group of retinal cell types, such as retinal ganglion cells, may not be detectable above baseline expression levels. Finally, the time course tested in the current study may not capture an increase in erythropoietin protein expression if it occurs after 24 h of reperfusion. However, in our previous studies, levels of multiple mediators of IPC were altered within the 24-h IPC time window. In the current study, we did not test the levels of erythropoietin mRNA, which would be a potentially more sensitive test. But measuring mRNA levels might also not be sensitive enough if the change occurs in a limited set of retinal cells.
We observed that EPO-R protein expression increased after IPC. In previous studies, several authors reported multiple protein bands with various EPO-R antibodies. In some instances, the presence of multiple bands appeared to be tissue-specific; multiple bands might be due to protein glycosylation or other posttranslational modifications.21,22In any event, the peptide competition experiment proved that the bands were all specific for EPO-R. While there are no previous studies that examined EPO-R mRNA and protein expression after preconditioning, several studies documented EPO-R upregulation after ischemia. In the mouse brain, EPO-R mRNA levels increased at 12 and 24 h after focal cerebral ischemia, and EPO-R protein levels were elevated after 24 h, peaking after 3 days.23In the rat retina, the same temporal induction of EPO-R protein was observed after ischemia as reported in mouse brain.11In neonatal rat brain, EPO-R protein expression was significantly increased at 6 and 12 h after focal cerebral ischemia.24Induction of EPO-R may play a key role in neuronal survival after ischemia, and it may also contribute to the endogenous protective mechanisms activated by IPC. Upregulation of EPO-R may result in more available binding sites for endogenous erythropoietin, thereby increasing signaling through EPO-R and activation of downstream pathways that contribute to neuronal survival and protection. Based on these previous studies, such a conclusion is both rational and biologically relevant; however, our current results have to be considered in light of the possibility that erythropoietin protein changed as well but could not be detected.
Among the possible downstream mechanisms for action of erythropoietin are Akt, ERK, and HSP27. These proteins are intermediaries as well in IPC. Transgenic mice overexpressing erythropoietin were protected against ischemic injury after middle cerebral artery occlusion, with erythropoietin activating janus-activated kinase-2, ERK-1/-2, and Akt pathways in the ischemic brain. Retinal ganglion cells in these transgenic rats were also protected from degeneration after axotomy.12,25Upregulation of Akt was neuroprotective by stabilizing mitochondrial function.26
ERK, activated after IPC in the rat retina,9is a cell survival factor that is believed to operate in part by increasing Mn-superoxide dismutase.27We showed that HSP27 was upregulated after retinal IPC in a time course consistent with that of IPC neuroprotection, and upregulation by cobalt chloride protected rat retina from ischemic injury in vivo .6,28HSP27 interacts with cytochrome c to negatively regulate cell death.29In a neonatal rat hypoxia-ischemia model, HSP27 was increased after administration of erythropoietin, and TUNEL-positive cells had decreased HSP27, suggesting that HSP27 may be increased by erythropoietin.30Considering our current results, upregulation of EPO-R by IPC may lead to enhanced availability of erythropoietin to activate ERK and HSP27 and enhance cell survival when the retina becomes ischemic. However, with no effect of sEPO-R on Akt activation after IPC in our model, it appears that there are other, erythropoietin-independent mechanisms that activate Akt.
Results of the current study demonstrate that EPO-R upregulation is a potential mechanism of IPC in the rat retina, with the neuroprotection afforded by erythropoietin involving activation of downstream proteins, including ERK and HSP27.