Erythropoietin and VEGF Promote Neural Outgrowth from Retinal Explants in Postnatal Rats

Abstract

purpose. Recent studies have reported neuroprotective effects of erythropoietin (EPO) and vascular endothelial growth factor (VEGF). The purpose of the present study was to clarify their influence on neurite outgrowth and regeneration of rat retinal ganglion cells (RGCs) in vitro and to elucidate the expression of corresponding receptors in the rat retina in vivo.

methods. Retinal explants from postnatal rats were stimulated with VEGF alone; VEGF in combination with anti-VEGF-receptor (VEGF-R)-2 antibody or T-type Ca²⁺ channel blocker ethosuximide (ESX); EPO alone; or EPO in combination with anti-EPO-receptor antibody or ESX. The presence of the corresponding receptors in the rat retina was assessed by reverse transcription-polymerase chain reaction (RT-PCR) and by immunohistochemistry.

results. EPO induced a stable improvement of neurite outgrowth of RGCs in a dose-dependent manner (5 × 10⁻¹⁵ M to 5 × 10⁻¹³ M) up to 169% (P < 0.05). Treatment of the explants with anti-EPO-R antibody (1:80 dilution) and with ESX (5 μM) totally inhibited EPO-mediated effects on RGCs. In comparison, VEGF (50 ng/mL), induced neurite outgrowth of retina explants up to 167% (P < 0.05), which again was inhibited in the presence of anti-VEGF-R2 antibody or ESX. Transcripts of EPO-R, VEGF-R1, and VEGF-R2 were detected by RT-PCR. Intense immunoreactivity for VEGF-R1, VEGF-R2, and EPO-R were found in the RGC layer of the retina.

conclusions. The data demonstrate for the first time that EPO and VEGF have a significant and specific biological effect on neurite regrowth of axotomized RGCs. Therefore, these results imply that EPO and VEGF have not only a neuroprotective but also a neuroregenerative role in ischemic retinal conditions.

Erythropoietin (EPO) is a hematopoietic cytokine with functions not only in erythrocyte development but also in the nervous system. ^{1 2} It is mainly produced by the kidney in adults ³ and by the liver in fetal stages. ⁴ In the human fetus, EPO expression is localized in the central nervous system (CNS), cerebrospinal fluid, adrenal cortex, and neural retina. ^{5 6 7}  

EPO exerts its biological functions in responsive cells in the CNS through a 66-kDa cell surface receptor (EPO-R), which is expressed in such areas of the brain as the cerebral cortex, midbrain, and hippocampus. ^{8 9} Astrocytes are able to synthesize EPO under insulin or insulin-like growth factor (IGF) stimulation ¹⁰ or under hypoxic conditions—for example, in ischemic insults. ¹¹ Therefore, EPO could be involved in providing a better oxygen supply to the CNS. Furthermore, EPO has been shown to protect primary cultured neurons from N-methyl-d-aspartate (NMDA) receptor–mediated glutamate toxicity and against ischemia-induced neuronal death. ^{9 12} In vivo evidence for a neuroprotective function of EPO has been demonstrated in different models of hypoxic-ischemic injury ^{11 12 13 14} and in an animal model of Parkinson disease. ¹⁵  

Comparable results have been reported in studies of vascular endothelial growth factor (VEGF). VEGF, known as a potent angiogenic factor, has similar neuroprotective properties. ^{16 17} The angiogenic action of VEGF involves an antiapoptotic effect on endothelial cells, which is mediated through the VEGF-receptor (VEGF-R)-2. In oxygen-injured rat retina, upregulation of the tyrosine kinase receptors VEGF-R1/flt-1 and VEGFR-2/flk-1 were demonstrated. ¹⁸  

Neuronal survival mediated by neuroprotective factors and neuronal regeneration in the CNS after injury are closely related. Despite the demonstrated benefits in neuroprotection, there are no data available concerning EPO-induced stimulation of neurite regeneration. Even though it has been reported that VEGF, besides its neuroprotective effects, has a neurotrophic influence, by stimulating axonal outgrowth from dorsal root ganglia or superior cervical ganglia in culture, ¹⁹ data describing the action of VEGF on axonal regeneration of central intrinsic neurons are still unavailable. 

Therefore, we wanted to clarify our hypothesis that EPO has the potential to stimulate axonal regeneration of RGCs in vitro by acting through the EPO-R. Furthermore, we wanted to prove the existence of EPO-R’s expression in the retina. Additionally, we wanted to elucidate the effect of VEGF on axonal outgrowth of central neurons. 

Material and Methods

Animals

Fisher rats (F344) were used in all experiments according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were kept under conditions of constant temperature and humidity in a 12 hour light–dark cycle in the animal house of the University of Kiel. Animals were mated overnight, and the day after mating was counted as embryonic day (E)1. Pups were born after 22 days of gestation, and the day of birth was counted as postnatal day (P)0. 

Preparation of Retinal Explants and Tissue Culture

Rat retinal explants (P11) were prepared and cultured in a three-dimensional fibrin gel, as described. ²⁰ Briefly, the retinae were prepared under semisterile conditions, and explants were punched out of the retinae with a 400-μm syringe needle. The retinal explants were cultured at 37°C in humidified CO₂-air (7.5%–92.5%) in a fibrin gel in serum-free (control cultures) or drug-supplemented media for 3 days. 

Cytokine and Drug Treatment

EPO (the kind gift of Janssen-Cilag, Neuss, Germany) was freshly prepared by dissolving 84 μg (10,000 U) in 1 mL distilled water. Ethosuximide (ESX; Sigma, Munich, Germany), a T-type Ca²⁺ channel blocker, was dissolved in ethanol to yield a 10-mM stock solution. Polyclonal anti-EPO-R antibody (Upstate Biotechnology, Lake Placid, NY) was diluted 1:80 and added to the cultures. VEGF (Biosource, Solingen, Germany) was prepared from a 100-μg/mL stock solution in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA), and anti-VEGF-R2 antibody (R&D Systems, Minneapolis, MN) was used in a concentration of 1 μg/mL. 

Evaluation of Neurite Outgrowth

After 3 days in vitro, the explant cultures were fixed in 2.5% glutaraldehyde by microwave irradiation (10 seconds, 50°C), dehydrated in ethanol, stained with Sudan black (microwave irradiation, 1 minute, 50°C), and embedded in a mixture of glycerol and gelatin. Neurite outgrowth was assessed using a computerized image-analysis system (Analysis; SIS, Münster, Germany). The axonal domain of the retinal explants was plotted by connecting the tips of the outgrowing neurites. The area of this domain was then determined by measuring the complete area and subtracting the area of the solid retinal explant. The axonal domain was expressed as a percentage of control cultures (100%). 

Reverse Transcription–Polymerase Chain Reaction Analysis

Total RNA from retinae (P10) was isolated and reverse transcribed (RT) according to the manufacturer’s recommended protocol (Gibco BRL, Eggenstein, Germany). RNA (5 μg) was processed with reverse transcriptase (Superscript; Gibco BRL) and Oligo(dT) primer (Amersham Pharmacia Biotech, Freiburg, Germany). RT product (3 μL) was then amplified by polymerase chain reaction (PCR) using a specific EPO-R primer pair (5′-CTATGGCTGTTGCAACGCGA-3′, sense), and (5′-CCGAGGGCACAGGA GCTT AG-3′, antisense) as reported by Morishita et al. ⁹ To amplify the VEGF-R1 transcript, the primer pair 5′-CAAGGGACTCTACACTTGTC-3′ (sense) and 5-′CCGAATAGCAGGCCTTC-ACT-3′ (antisense) was used, yielding a 240-bp PCR product, and the primers 5′-GCCAA-TGAAGGGGAACTGAAGAC-3′ (sense) and 5′-TCTGACTGCTGGTGATGCTGTCC-3′ (antisense) were used for amplification of the VEGF-R2 transcript, yielding a 537-bp product. ²¹ cDNA products were subjected to agarose gel electrophoresis (1.5%), and fragments were visualized with ethidium bromide incorporation under UV light. The rat β-actin housekeeping gene was used as an internal control to demonstrate that equivalent amounts of RNA had been loaded. 

Immunohistochemistry of EPO-R, VEGF-R1, and VEGF-R2 in the Rat Retina

The animals were killed by transcardial perfusion with 4% paraformaldehyde in PBS. The eyes were removed and embedded in paraffin after removal of the lens. Rehydrated sections (7 μm) were incubated in 10% horse serum for 20 minutes to block nonspecific binding, washed three times in PBS, and incubated for 1 hour with anti-EPO-R antibody (rabbit polyclonal IgG, clone M-20, 1:100 dilution), anti-VEGF-R1 antibody (rabbit polyclonal IgG, clone c-17, 1:80 dilution), or anti VEGF-R2 antibody (mouse monoclonal IgG, clone A-3, 1:80 dilution). After a wash in PBS, sections were incubated with goat anti-rabbit IgG (1:1000, 1 hour; Vector Laboratories, Burlingame, CA). Immunohistochemical staining was performed with the avidin-biotin complex technique according the manufacturer’s protocol (Vector). Negative controls were performed by omitting the primary antibodies and in parallel (for EPO-R) by preabsorbing the primary antibodies with a 10-fold (by mass) excess of a specific blocking peptide (Santa Cruz Biotechnology, Inc., Heidelberg, Germany). Sections were photographed under a microscope (Axioplan; Carl Zeiss Inc., Thornwood, NY) on TMX400 film (Eastman Kodak, Rochester, NY). 

Statistical Analysis

All results are expressed as the mean ± SEM. The significance of differences was tested with ANOVA and Bonferroni (Fisher) test. P < 0.05 was considered statistically significant. 

Results

Control Cultures

After 3 days in vitro, many regenerating but short neurites were visible, growing out from the whole circumference of the retinal explants, reaching growth distances up to 180 μm (Fig. 1A) . Their responsiveness to brain-derived neurotrophic factor (BDNF) ²² confirmed that the regenerating neurites were processes of RGCs (data not shown). 

Effects of EPO and VEGF on Neurite Outgrowth from RGCs

Compared with control explants (100%), EPO dose dependently enhanced the axonal domain with a half-maximum concentration (EC₅₀) of 7.8 × 10⁻¹⁵ M. In a concentration of 5 × 10⁻¹³ M (maximum concentration) EPO increased the neurite elongation up to 169% ± 18% (n = 18; Figs. 1D 2 ). Most of the RGCs extended their axons more than 400 μm—some of them, up to 650 μm. The addition of VEGF (50 ng/mL) elongated the neurites up to 167% ± 20% (n = 18; Figs. 1E 2 ). The length of the majority of axons was in the range of more than 350 μm. Single axons reached distances of up to 600 μm. 

Effects of Anti-EPO-R Antibody, the T-Type Ca2⁺ Channel Blocker ESX and Anti-VEGF-R2 Antibody on Neurite Outgrowth

In the presence of anti-EPO-R antibody, the EPO-induced neurite outgrowth was totally abolished (102% ± 8%, n = 18; Figs. 1B 2 ). Similar results were achieved in the presence of anti-VEGF-R2 antibody after VEGF stimulation (Fig. 2) . Coincubation with EPO (5 × 10⁻¹³ M) or VEGF (50 ng/mL) and the T-type Ca 2⁺ channel blocker ESX (5 μM) induced a reduction of the EPO- and VEGF-mediated regenerative response to control levels (EPO: 101% ± 11%, n = 18; VEGF: 108% ± 9%, n = 12; Figs. 1C 2 ). 

EPO-R, VEGF-R1, and VEGF-R2 Transcripts in the Rat Retina

The presence of EPO-R, VEGF-R1, and VEGF-R2 in rat retinae was analyzed by RT-PCR. Figure 3 shows an ethidium bromide–stained agarose gel of PCR products from rat retina. In lane 2 the expected 240-bp PCR product of VEGF-R1 and in lane 3 the 537-bp product of VEGF-R2 cDNA are visible. The band in lane 4 represents the expected 402-bp PCR product of EPO-R cDNA. Lane 5 shows β-actin (279 bp), and the DNA migration in a 100-bp DNA-ladder is seen in lane 1. 

Immunohistochemistry

To localize EPO-R, VEGF-R1, and VEGF-R2 in the adult rat retina, immunohistochemical studies were performed. Positive EPO-R immunostaining was detectable in different layers of the rat retina, with the most intensive immunostaining identifiable in the RGC layer. Less intensive labeling was depicted in the inner and outer plexiform layers and in the inner portion of the inner segments (Fig. 4A) . The EPO-R–specific staining was confirmed by coincubation of the anti-EPO-R antibody with the EPO-R–specific blocking peptide (Fig. 4B) . 

Positive immunolabeling of VEGF-R1 (Fig. 4C) and of VEGF-R2–bearing cells (Fig. 4D) showed comparable results in the RGC layer. 

Discussion

The hematopoietic cytokine EPO and the angiogenic factor VEGF have been demonstrated to exert neuroprotective effects (e.g., during hypoxia in vitro and in vivo ^{12 16} ) in α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)–mediated cytotoxicity ²³ and in an animal model of parkinsonism. ¹⁵  

In this article, we present evidence that EPO is also capable of stimulating neurite outgrowth in neurons of the rat retina in vitro—a hitherto unknown function of EPO in the CNS. We also complement and extend the findings on the neurotrophic potency of the angiogenic factor VEGF, ^{16 19} by demonstrating a robust axonal outgrowth of rat retinal ganglion cells (RGCs) stimulated with VEGF in vitro. Both peptides are induced under hypoxic conditions by the basic helix-loop-helix transcription factor HIF-1 (hypoxia-induced factor-1; see Ref. ²⁴ for review). 

In our three-dimensional tissue culture system of retinal explants, the ability of RGCs to extend their axons is comparable to that of other adult mammalian CNS neurons. ^{25 26} The limited regenerative capacity is reflected in the control experiments where the explants were cultured in serum-free medium, and only abortive axonal sprouts were visible. However, the addition of growth-promoting molecules, such as BDNF (10 ng/mL) or ANG II (5 × 10⁻⁶ M), ²² induces an enhanced outgrowth of axons from the whole circumference of the explants. Hence, if supplied with the appropriate neurotrophic substrates, postnatal RGCs are able to regenerate their axons in this system, which may serve as a reliable model for the investigation of potential neuroprotective–neuritogenic molecules for CNS neurons. An effect similar to 10 ng/mL of BDNF is elicited by the administration of either EPO (5 × 10⁻¹³ M) or VEGF (50 ng/mL). 

VEGF and Neurite Outgrowth of RGCs

Sondell et al. ¹⁹ recently reported that VEGF stimulates axonal outgrowth from mouse superior cervical ganglion and dorsal root ganglion cells through the flk-1 receptor (VEGF-R2). Their results of VEGF-mediated effect on axonal outgrowth can now be extended to a subset of CNS neurons: postnatal RGCs. We cannot deliver direct evidence that the neuritogenic effect of VEGF in RGCs is mediated through the flk-1 receptor, because the receptor antagonist SU5416 (Sugen, South San Francisco, CA) has not been made available yet, but we demonstrate that VEGF-induced neurite outgrowth could be blocked by an anti-VEGF-R2 antibody or by addition of ESX, a T-type calcium channel blocker. Furthermore, transcript and protein of the VEGF receptor flk-1 (VEGF-R2) are strongly expressed in the RGC layer of adult rats. Jin et al. ¹⁷ demonstrated that the VEGF-induced neuroprotection is mediated through a VEGF-R2–dependent activation of the phosphatidylinositol 3′ kinase (PI3-K)/Akt signaling pathway, which also promotes cell survival and proliferation of vascular endothelial cells. ²⁷ Because PI3-K/Akt has been shown to enhance axonal regeneration in motoneurons and directly mediates the neuroprotective effects of IGF-1 in RGCs, ²⁸ it is likely to hypothesize that activation of the PI3-K/Akt signaling pathway may be a key event in the demonstrated VEGF-evoked neurite outgrowth of rat RGCs. 

Response of RGC Axons to EPO

Our results suggest that the EPO-induced growth response is mediated directly through EPO-Rs on RGCs. Transcribed mRNA of the EPO-R gene is present in the adult retina, and a predominant expression of EPO-R in the RGC layer was visualized by immunohistochemistry. Coincubation with EPO-R–blocking antibodies totally abolished the effects on neurite outgrowth of the retinal explants. 

Besides its well-known action on the hematocrit, EPO has been implicated in the process of angiogenesis (e.g., Ref. ² ), pointing toward a possible role of an improved blood supply for the neuroprotective actions of EPO during ischemic insults and diabetic macular retinopathy. ²⁹ Furthermore, EPO has been reported to act as an anti-apoptotic agent ³⁰ or as an antioxidant, ³¹ both of which could explain the neuroprotective effects described so far. Although vascularization or erythropoietic effects of EPO can be excluded in the assay system we used, we cannot completely rule out the possible action of EPO on glial cells stimulating the release of other neurotrophic factors that may act on RGCs in a paracrine manner. In contrast, immunohistochemical staining did not reveal a distinct expression of EPO-R on Müller cells in the adult retina, making EPO-responsiveness of only the glial elements in the retina rather unlikely. 

Intracellular Signaling through EPO-R

The intracellular signaling pathways of EPO-R in neuronal cells is not yet fully understood. From the studies of other cell populations and a study by Assandri et al. ³² in human neuroblastoma cells, it is tempting to speculate that the EPO-mediated regulation of calcium fluxes might be involved in eliciting the axonal growth response. Calcium has been shown to be pivotal in the processes of neurite extension and growth cone guidance (for review, see Ref. ³³ ). As reported in a paper by Archer et al., ³⁴ the positive effects of the FGF receptor and L1 on axonal outgrowth is blocked when calcium channel blockers are administered. In our hands, the T-type calcium-channel blocker ESX blocked the neuritogenic effect both of EPO and VEGF. In rat pheochromocytoma cells (PC12) a calcium-dependent increase in mitogen-activated protein kinase (MAPK) activity has been demonstrated on EPO-stimulation. ³⁵ These mechanisms may also be involved in the calcium-dependent growth response elicited by EPO-R or VEGF-R in the retinal explants. 

In recent studies, it has been shown that another crucial element of EPO-R signaling consists of the activation of transcription factors of the signal transducer and activator of transcription (STAT) family. ³⁶ The EPO-R associates with JAK2, a member of the Janus-protein kinases, thereby converting latent cytoplasmic STAT-transcription factors into their active form through tyrosine phosphorylation. The ability to inhibit apoptotic signals in erythroid progenitor cells through STAT5-dependent induction of bcl-xL gene expression, ^{37 38} taken together with our results that caspase-1 inhibition mediates neurite outgrowth from retinal explants, ³⁹ shows a potential contribution of the JAK/STAT pathway in EPO-mediated neuroprotection–growth stimulation of CNS axons. Ciliary neurotrophic factor (CNTF), a known neurotrophic factor for RGCs, activates STAT3 in RGCs and Müller cells, when directly administered into the vitreous body, supporting a neuroprotective–regenerative role of STAT factors. ⁴⁰ The direct involvement of the JAK/STAT pathway in EPO signaling in axotomized RGCs is currently under investigation. 

Should our in vitro findings be verified by comparable in vivo studies, the results would further elucidate the novel view of extensive crosstalk between the hematopoietic cells and the CNS, where neural stem cells can repopulate depleted bone marrow, ⁴¹ and hematopoietic cytokines are able to stimulate axonal regeneration (this study) or the differentiation patterns of expanded CNS precursors. ⁴²  

Possible Therapeutic Relevance

Retinal diseases such as glaucoma, central artery occlusion, and diabetic retinopathy, are characterized by progressive damage to RGC axons and RGC death. They are among the leading causes of blindness, and, at present, neuroprotective therapies for these debilitating diseases remain unavailable. Based on the findings described herein, we propose that the EPO/EPO-R system (and possibly the VEGF/VEGF-R system) could provide an endogenous mechanism that may protect the retina against damages of different origin and may also promote axonal regeneration in CNS lesions. However, reports of a potentially harmful role of VEGF, through VEGF-R2, in the pathogenesis of neovascularizing retinal diseases must be carefully considered. ^{43 44} Because a proangiogenic role has been described, not only for VEGF, but also for EPO, further elucidation of the respective signaling pathways seems an absolute prerequisite to molecular dissection of the stimulation of axonal outgrowth from potentially harmful angiogenic effects. 

Considering recent findings that could demonstrate neurotrophic actions of a truncated EPO peptide through EPO-R without the untoward effects of stimulating erythropoiesis and raising the hematocrit, ⁴⁵ a selectively designed “neuro-EPO” might provide a basis for novel, receptor-directed therapeutic strategies for a plethora of retinal ischemic diseases. 

 Supported by a grant from the Hensel-Foundation, University of Kiel (SB-M) and by a grant of the Research Center of Integrative Neurosciences (FZIN), University of Kiel (RL, SB-M).

 Submitted for publication August 13, 2001; revised January 15, 2002; accepted January 30, 2002.

 Commercial relationships policy: F.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Retinal (P10) explants after 3 days in vitro. (A) Many regenerating but short neurites were detected in control culture. (D) In the presence of EPO, the number and length of neurites were increased up to 169% ± 19%. Coincubation of EPO with anti-EPO-R antibody (B) or the T-type Ca channel blocker ESX (C) totally abolished the EPO-induced axonal outgrowth. (E) Addition of VEGF to retinal explants enhanced the neurite growth up to 167% ± 17%. Bar, 100 μm.

Figure 2.

 

Effect of EPO and VEGF on neurite outgrowth of retinal explants (P10). EPO increased the axonal regeneration in comparison with control culture. The effect of EPO on axonal elongation was blocked by anti-EPO-R antibody (Ab) or by the T-type calcium blocker ESX (5 μM). VEGF (50 ng/mL) exerted a similar effect on axonal regeneration, and this effect was blocked by anti-VEGF-R2 antibody or by the T-type calcium blocker ESX (5 μM). In all experimental groups, n = 18, except VEGF+ESX, n = 12. Data are expressed as the mean ± SEM; *P < 0.05 compared with the control.

Figure 3.

 

Detection of EPO-R, VEGF-R1, and VEGF-R2 mRNA in the P10 rat retina. RT-PCR products were visualized with agarose gel electrophoresis in the presence of ethidium bromide. Electrophoresis of RT-PCR products yielded bands of expected sizes: Lane 1: DNA migration on a 100-bp ladder; lane 2: 240 bpm, VEGF-R1; lane 3: 537 bp, VEGF-R2; lane 4: 402 bp, EPO-R; lane 5: 279 bp, β-actin.

Figure 4.

 

Immunohistochemicaldetection of EPO-R (A), VEGF-R1 (C), and VEGF-R2 (D) in the rat retina. (A) Note the intensive immunoreactivity of the ganglion cell layer (GCL) as well as the moderate staining of the inner (IPL) and outer plexiform layers (OPL) and the inner portion of the inner segment (IIS). (B) Negative control: EPO-R immunoreactivity preabsorbed with blocking peptide. Strong immunolabeling of VEGF-R1 (C) and VEGF-R2 (D) in the GCL. Original magnification, ×400.

The authors thank Sibille Piontek for her expert technical assistance and Christian Bauer (Zurich, Switzerland) for helpful comments.