Mechanisms of Erythropoietin-Mediated Protection Against Retinal Ganglion Cell Apoptosis

Introduction

Retinal ganglion cells (RGCs), as cells located in the innermost layer of the retina, are the last cells in the neural network of the eye to receive visual stimuli [1]. However, RGCs play a critical role in the visual information transmitted from the axon-formed optic nerve to the brain [2]. Therefore, the dysfunction or death of the RGCs can cause ocular disorders with irreversible blindness, such as glaucoma and diabetic retinopathy [3,4]. There are many reasons for the death of RGCs, including apoptosis, inflammation, and excitotoxicity; therefore, the discovery of a neuroprotective molecule or the development of a therapeutic method for inhibiting RGC death has become a potential treatment strategy for the associated eye diseases of RGCs and for preventing vision deterioration.

Erythropoietin (EPO), as a type of glycoprotein, could play a nerve-protective role in retinal disease and stimulate red blood cell production [3,5]. A previous study [6] reported that EPO could interact with its receptors and inhibit the apoptosis of erythrocyte progenitors. Moreover, EPO has been shown to have neuroprotective functions in retinal degeneration, stroke, axotomy, glaucoma, and glutamate-induced and N-methyl-d-aspartate (NMDA)-mediated toxicity [5,7–10]. However, whether EPO plays a critical role in the viability or apoptosis of RGCs, and the associated molecular mechanisms, are unclear. The signaling pathways and associated molecules involved in the effects of EPO on RGCs need to be clarified; therefore, it is critical to study the therapeutic targets for the treatment of retinal disease.

The PI3K/Akt signaling pathway is considered a crucial intracellular pathway that plays an important role in the cell cycle process [11]. The JAK/STAT signaling pathway has been shown to be an evolutionarily conserved transmembrane pathway that enables cells to communicate with the external environment [12]. The MAPK/ERK signaling pathway has a critical role in modulating apoptosis, proliferation, autophagy, and differentiation [13]. EPO and its receptor, EPOR, can activate the cascade signaling pathways JAK2/STAT5, MAPK, and PI3K/AKT4[14]. These signaling pathways transmit signals to each other and play an important role in cell survival and proliferation. We speculated that EPO might also play an important role in the regulation of cell viability and apoptosis through the above signaling pathways. Therefore, in this study, we aimed to investigate the effect of the EPO molecule on cell viability and apoptosis and examine the expressions of apoptosis-associated proteins and PI3K/Akt-JAK/STAT signaling pathway-associated molecules in RGCs. Resolving these problems will help further explore controversial issues in previous studies.

Material and Methods

ISOLATION, PURIFICATION, AND IDENTIFICATION OF PRIMARY RGCS:

The RGCs were isolated, purified, and cultured with an immunopanning method, as a previous study described [1], with a few modifications. Briefly, the retinas isolated from 3-day-old Sprague-Dawley rats (certificate no. SCXK (Yu) 2022-0010) were incubated in neurobasal medium (cat. no. 21103-049, Gibco, USA) containing 0.25% trypsin (cat. no. C0201, Beyotime Biotech, Shanghai, China). These cells were treated with rabbit anti-rat CD11b antibody (1: 1000, cat. no. ab133357, Abcam, UK) at a concentration of 1.3 μg/mL at 37°C for 20 min. The dissociated cell suspension was then treated on a panning plate coated with goat anti-rat IgG to remove the macrophages. The RGCs were purified with a panning plate coated with rabbit anti-rat Thy-1 antibody (1: 500, cat. no. DF4804, Affinity, USA) at a concentration of 2 μg/mL at 37°C for 30 min. The RGCs obtained were seeded at a density of 1×105 cells/mL and cultured in neurobasal medium with 10% fetal bovine serum in a humidified atmosphere and 5% CO2.

The isolated RGCs were identified using the immunohistochemical assay, by staining with rabbit anti-rat antibody (4°C, overnight) and horseradish peroxidase-labeled goat anti-rabbit (cat. no. A0208, Beyotime Biotech) at room temperature for 60 min, coloring with DAB for 1 to 10 min, and re-dyeing with hematoxylin for 2 times, 5 min/time.

LENTIVIRAL VECTORS CONSTRUCTION, LENTIVIRUS PRODUCTION, AND INFECTION:

In this study, the EPO-overexpressed lentiviral vector was constructed. In summary, the open reading frame of the EPO gene (NCBI reference sequence: NM_017001.2, gene ID: 24335) was amplified and cloned into the lentiviral Plvx-mCMV-zsGreen-Puro vector and named as the lentiviral vector “Ov-EPO”. The lentiviral vector Ov-EPO was co-transfected with the enveloping plasmid pMD2.G and the packaging plasmid psPAX2 in 293 T using NDE3000 in Opti-MEM. Virus particles (Ov-EPO lentivirus) were harvested 24 h after transfection and infected RGCs at a multiplicity of infection of 50. Finally, the obtained Ov-EPO lentivirus was concentrated and titrated for the following experiments.

EXPERIMENTAL GROUPING AND TREATMENT:

The RGCs in this study were divided into 7 groups, including the RGC, blank, Ov-EPO, H₂O₂, NMDA, H₂O₂+Ov-EPO, and NMDA+Ov-EPO groups. RGCs in the RGC group were treated without any reagents, while RGCs in the blank group were treated with blank lentivirus. RGCs in the Ov-EPO group were treated with Ov-EPO lentivirus. When cell density reached 60%, RGCs in the H₂O₂ group and H₂O₂+Ov-EPO group were administered with 200 μmol/L of H₂O₂ and 200 μmol/L of H₂O₂ plus Ov-EPO lentivirus, respectively. The RGCs in the NMDA group and in the NMDA+Ov-EPO group were administered with 100 μmol/L of NMDA and 100 μmol/L of NMDA plus Ov-EPO lentivirus, respectively.

CELL-COUNTING KIT-8 ASSAY:

The viability of the RGCs was determined using a cell-counting kit (CCK-8) kit (cat. no. C0039, Beyotime Biotech), according to the manufacturer’s protocol. Briefly, RGCs were seeded in 96-well plates at a concentration of 0.5×10⁴ cells/well, and cultured for 48 h (achieving a cell density of 80%) at 37°C. The RGCs were then treated using 10 μL of CCK-8 solution at 37°C for 1 h. Finally, a Varioskan LUX ELISA reader (Thermo Fisher, Waltham, MA, USA) was applied at a wavelength of 450 nm to record the optical density of cells in different groups.

REAL-TIME QUANTITATIVE POLYMERASE CHAIN REACTION ASSAY:

Total RNAs were extracted from RGCs that underwent different treatments with RNAiso plus (cat. no. 9109Q, Takara, Japan). Complementary DNAs (cDNAs) of the RNAs extracted above were synthesized with the Hifair III 1st cDNA synthesis kit (cat. no. 11139ES10, Yeasen Biotech, Shanghai, China). The Hieff UNICON Universal Blue Qpcr SYBR Green Master Mix (cat. no. 11184ES03, Yeasen Biotech) was used to perform the real-time quantitative polymerase chain reaction (RT-qPCR) assay and amplify targeting genes. The reaction conditions of the RT-qPCR were listed as follows: 95°C for 2 min, 40 cycles at 95°C for 10 s and 60°C for 30 s, and 72°C for 30 s. In this study, targeting genes, including EPO, Bcl-2, alpha serine/threonine-protein kinase (AKT1), signal transducer and activator of transcription 5A (STAT5A), and mitogen-activated protein kinase (MAPK), were amplified, and the primers are listed in Table 1. GAPDH was used as internal control to normalize the above target genes. The relative transcription of the target genes was calculated with the 2−ΔΔCt method and analyzed using GraphPad software.

TUNEL ASSAY:

TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) staining was conducted with a colorimetric TUNEL apoptosis assay kit (cat. no. C1098, Beyotime Biotech), according to the manufacturer’s protocol. Briefly, RGCs were seeded in 96-well plates at a concentration of 0.5×10⁴ cells/well, and cultured for 1 day and 7 days. The cultured and treated RGCs were then fixed with 4% polyformaldehyde for 30 min at 4°C, and permeabilized with 0.3% Triton X-100 in phosphate-buffered saline solution for 5 min at room temperature. The RGCs were then incubated with biotin labeling solution for 60 min, reaction termination solution for 10 min, and streptavidin-horseradish peroxidase solution for 30 min at 37°C. Finally, the RGCs were stained with DAB solution to color the apoptotic cells. TUNEL-positive cell images were observed and captured with a DMC5400 microscope (Leica, Germany), with a magnification of 400× in a visual field. The TUNEL-positive stained cells were analyzed using Image J software (NIH, USA). The staff member observing the TUNEL-positive cells was unaware of the cell grouping.

STATISTICAL ANALYSIS:

Data were defined as mean±standard deviation, and data analysis was performed with SPSS 26.0 software (IBM Corp, Armonk, NY, USA). The assumption of normality and homogeneity for the variance comparisons was analyzed using Kolmogorov-Smirnov and Leven tests, respectively (all data in accordance with normality and homogeneity in the findings). Differences among groups were analyzed using a one-way analysis of variance (ANOVA) test validated by Tukey’s post hoc test. P<0.05 was defined as a significant difference.

Results

THE OV-EPO LENTIVIRUS SHOWED A HIGHER EXPRESSING EFFICACY:

In this study, RGCs were successfully isolated from retinal tissues and identified by staining the Thy-1 molecule (Figure 1A). According to the results of the gene sequencing, the EPO gene was successfully cloned, and the Ov-EPO lentivirus was established (Figure 1B). The RT-qPCR results showed that RGCs cells infected with the Ov-EPO lentivirus expressed significantly higher EPO that did the RGC group and blank group (Figure 1C, both P<0.001). Therefore, Ov-EPO lentivirus was successfully established in this study.

EPO OVEREXPRESSION INHIBITED APOPTOSIS OF RGCS:

To evaluate the effect of EPO overexpression on RGC apoptosis, the TUNEL assay was conducted. The results of the TUNEL stain indicated that the TUNEL-positive stained cells in the Ov-EPO group were less than those of the RGC group and blank group at 1 day (both Ov-EPO vs RGC and Ov-EPO vs blank: P<0.05) and 7 days (both Ov-EPO vs RGC and Ov-EPO vs blank: P<0.05) after culture (Figure 2). In addition, there were no differences for TUNEL-positive stained cells between the RGC group and blank group, both at 1 day and 7 days after culture (Figure 2). This result suggests that EPO overexpression could inhibit apoptosis of RGCs.

EPO OVEREXPRESSION RESISTED OXIDATIVE DAMAGE AND EXCITATORY TOXICITY:

In this study, we used H2O2 and NMDA to induce oxidative damage and excitatory toxicity of RGCs. The findings of the CCK-8 assay showed that 200 μmol/L of H2O2 and 100 μmol/L of NMDA were the appropriate concentrations to induce oxidative damage and excitatory toxicity, respectively (Figure 3A). Furthermore, under the microscope, the cell counts in the 200 μmol/L H2O2 group (P<0.05) and 100 μmol/L NMDA group (P<0.01) were significantly more than that in the RGC group (Figure 3B), which confirmed the results of the CCK-8 assay above.

As shown in Figure 4A, the cell viability of the RGCs in the H2O2 group and NMDA group was significantly lower than that of the blank group (both P<0.05). However, EPO overexpression not only upregulated the viability of RGCs (Figure 4A, P<0.05), but also markedly reversed the oxidative damage induced by H2O2 and excitatory toxicity induced by NMDA of RGCs (Figure 4A, both P<0.05). EPO overexpression upregulated the transcription of the Bcl-2 gene in RGCs, compared with those in RGCs without Ov-EPO treatment, with a 1.5-fold increase in the transcription of the Bcl-2 gene (Figure 4B, P<0.001), showing that the increase in Bcl-2 transcription is EPO overexpression dependent. Furthermore, EPO overexpression could also significantly reverse the H2O2 or NMDA caused by the reduction of Bcl-2 gene transcription (Figure 4B, P<0.01 and P<0.001, respectively). Therefore, overexpression of EPO could resist the death of RGCs caused by oxidative damage and excitatory toxicity.

EPO OVEREXPRESSION ACTIVATED PI3K/AKT AND JAK/STAT SIGNALING PATHWAY IN RGCS:

The key molecules of the PI3K/Akt signaling pathway (AKT1) and the JAK/STAT signaling pathway (STAT5A), were examined using the RT-qPCR assay in RGCs of different groups. The results showed that the gene transcriptions of AKT1 (Figure 5A, approximately 2.3-fold) and STAT5A (Figure 5B, approximately 1.6-fold) in RGCs of the Ov-EPO group were significantly higher than that in the blank group (Figure 5A, both P<0.001). Furthermore, when the H2O2- and NMDA-treated RGCs were infected with Ov-EPO lentivirus, the transcriptions of AKT1 (Figure 5A, both P<0.001) and STAT5A (Figure 5B, both P<0.01) genes were increased markedly compared with that of the H2O2 and NMDA groups. Therefore, EPO overexpression could activate the JAK-STAT signaling pathway in RGCs.

EPO OVEREXPRESSION WAS ASSOCIATED WITH THE MAPK/ERK SIGNALING PATHWAY:

In this study, the key molecule of the MAPK/ERK signaling pathway, MAPK, was also evaluated with the RT-qPCR assay. MAPK gene transcription in EPO-overexpressed RGCs was significantly increased compared with that in RGCs of the blank group, with an increase of 3-fold (Figure 6, P<0.01), while EPO overexpression in RGCs undergoing H2O2 and NMDA treatment significantly promoted MAPK gene transcription, compared with that of the H2O2 and NMDA groups (Figure 6, both P<0.001). These results suggest that the MAPK/ERK signaling pathway was triggered in EPO-overexpressed RGCs.

Discussion

In recent years, EPO has been shown to play a critical role in the treatment of ocular diseases. In this study, we identified the protective effect of EPO on the viability of RGCs, especially in terms of apoptosis associated with the mitochondrial pathway. Furthermore, the PI3K/Akt, JAK/STAT, and MAPK/ERK signaling pathways were involved in the viability and apoptosis of RGCs.

EPO is known to have a neuroprotective role and a hematopoietic effect in the retina, which has attracted a lot of attention in recent years [15,16]. A previous study [17] reported that EPO could protect RGCs and promote regeneration of the axon when the optic nerve was damaged in an animal model. Meanwhile, the protective effect of EPO on RGCs has been previously reported in the contexts of axotomy [7], glaucoma [8], NMDA and TNF-α [5], and high glucose [9]. Qi et al [10] showed that EPO could inhibit apoptosis of high-glucose-induced retinal ganglion cells through the JNK pathway. However, the specific mechanism of the protective effect of EPO on RGCs is still controversial. In this study, we synthesized Ov-EPO lentivirus and infected the RGCs. Although RGCs could also express EPO protein, EPO overexpression significantly promoted cell viability and inhibited the apoptosis of RGCs. Fu et al [18] also reported that Muller cells are the main source of EPO in normal retina, which is consistent with the present findings. Actually, cells could also be treated by adding exogeneous EPO. Al-Sarraf et al [19] reported that the application of exogenous EPO has an obvious neuroprotective effect on cerebral ischemia. However, under physiological conditions, the expression of EPO in cells is strictly regulated. Compared with exogenous addition, overexpression can better simulate the production and mode of action of endogenous EPO in cells, reduce nonspecific effects that can arise from exogenous addition, and provide a more reliable model for studying the role of EPO in neuroprotection. Furthermore, the expression level of EPO in the EPO overexpression RGCs was over 380-fold higher than that in normal RGCs. Therefore, in this study, we chose to overexpress the EPO molecule in RGCs. In this study, according to our RT-qPCR results, EPO overexpression obviously reversed H2O2- or NMDA-induced reduction in transcription of the Bcl-2 gene, suggesting that EPO overexpression triggered the decrease in apoptosis by increasing the transcription of the Bcl-2 gene and inhibiting oxidative damage and excitatory toxicity. As is known, Bcl-2 is a key biomarker of the mitochondria-associated apoptosis signaling pathway [20]; therefore, overexpression of EPO could inhibit apoptosis through the mitochondrial signaling pathway.

A previous study [1] described that ciliary neurotrophic factor protects RGCs through the JAK/STAT, PI3K/AKT, MAPK/ERK signaling pathways. The present study also discussed the roles of these 3 signaling pathways involved in EPO overexpression-mediated cell viability and apoptosis of RGCs. The key molecules of the PI3K/Akt signaling pathway (AKT1) and the JAK/STAT signaling pathway (STAT5A) [21], were determined in the EPO-overexpressed RGCs. The results showed that EPO overexpression significantly increased the expression of AKT1 (2.3-fold increase) and STAT5A (1.6-fold increase) in RGCs, compared with that in the RGCs group and the blank group, suggesting that the increase of both molecules could be EPO overexpression-dependent. Meanwhile, markedly inhibited EPO overexpression also reversed the H2O2- and NMDA-induced decreased expression of AKT1 and STAT5A, compared with that of the H2O2 and NMDA groups. These results suggest that EPO overexpression inhibits apoptosis of RGCs by modulating the PI3K/Akt and JAK/STAT signaling pathway. The MAPK molecule as a marker of the MAPK/ERK signaling pathway [22] was also identified in EPO-overexpressed RGCs. The result indicated that EPO overexpression could enhance MAPK expression in both the RGCs and the H2O2- and NMDA-treated RGCs; therefore, the MAPK/ERK signaling pathway is involved in the EPO-overexpression downregulated apoptosis of RGCs. Furthermore, these results suggest that EPO overexpression could suppress the oxidative damage and excitatory toxicity of RGCs. Therefore, EPO could play a protective role in RGCs by inhibiting the apoptosis, oxidative damage, and excitatory toxicity of RGCs, and could involve the PI3K/Akt-JAK/STAT-MAPK/ERK signaling pathway, which provides a potential mechanism distinguished from that of previous studies [5,7–10].

Although the present study showed some significant results, there are also a few limitations. First, in this study, we analyzed only the signaling cascades in gene transcription level, but did not detect the associated proteins involved in the above signaling pathways. Second, the EPO-overexpressing lentivirus was established and infected RGCs; however, no experiments were conducted to decrease or knockdown the EPO gene to validate the conclusion of this study. Third, the observed effects of EPO overexpression on RGCs were not verified in any retinal disease model at the in vivo level (animal experiments) in this study. Fourth, we evaluated only the apoptosis of RGCs using the TUNEL staining method; however, the apoptotic stimulus of RGCs has not been determined, and the results have not been confirmed using other different methods. Fifth, in this study, we overexpressed only the EOP molecule in RGCs by lentiviral vector infection; however, the assays in un-transfected cells directly exposed to EPO treatment were not conducted, to validate the results obtained. Sixth, the present results did not account for the activation of these signaling elements (PI3K/Akt, JAK/STAT, MAPK/ERK) upon exposure to EPO (such as phosphorylation) nor for their participation in the action mechanism of EPO. Therefore, administered inhibitors of the PI3K/Akt-JAK/STAT-MAPK/ERK signaling pathways could improve the present findings.

Conclusions

The present results indicated that EPO plays a protective role in the cell viability of RGCs, and Bcl-2 plays a critical role in this process. EPO inhibited the apoptosis, oxidative damage, and excitatory toxicity of RGCs, which involved the potential PI3K/Akt-JAK/STAT-MAPK/ERK signaling pathway (Figure 7). The results of this study can provide a basis for the development of potential therapeutic targets and the exploration of personalized medicine as a therapeutic agent for retinal-associated diseases.