VGF nerve growth factor inducible is involved in retinal ganglion cells death induced by optic nerve crush

VGF nerve growth factor inducible (VGF) is a polypeptide that is induced by neurotrophic factors and is involved in neurite growth and neuroprotection. The mRNA of the Vgf gene has been detected in the adult rat retina, however the roles played by VGF in the retina are still undetermined. Thus, the purpose of this study was to determine the effects of VGF on the retinal ganglion cells (RGCs) of mice in the optic nerve crush (ONC) model, rat-derived primary cultured RGCs and human induced pluripotent stem cells (iPSCs)-derived RGCs. The mRNA and protein of Vgf were upregulated after the ONC. Immunostaining showed that the VGF was located in glial cells including Müller glia and astrocytes but not in the retinal neurons and their axons. AQEE-30, a VGF peptide, suppressed the loss of RGCs induced by the ONC, and it increased survival rat-derived RGCs and promoted the outgrowth of neurites of rat and human iPSCs derived RGCs in vitro. These findings indicate that VGF plays important roles in neuronal degeneration and has protective effects against the ONC on RGCs. Thus, VGF should be considered as a treatment of RGCs degeneration.

Thus, the purpose of this study was to determine the role played by VGF on the degeneration of the RGCs and optic nerve of mice after the optic nerve crush (ONC). In addition, we evaluated the effect of AQEE-30, a VGF peptide, on the retinal injury induced by the ONC and also on RGCs in culture.

Results
Level of expression of mRNA of Vgf gene and VGF protein after ONC. Initially, we determine the level of expression of the mRNA of the Vgf gene and the protein after the ONC. The level of the mRNA of Vgf was increased in the retina at 2, 3, and 5 days after the ONC (Fig. 1A). The VGF protein expression in retinas was also increased (Fig. 1B-D). Immunostaining showed that the expression was observed in the RNFL + GCL, IPL, INL, and OPL, and the degree of expression in the RNFL + GCL and IPL was higher than that in the INL and OPL at 7 days after the ONC (Fig. 1C,E,F).
Retinal site of VGF expression in sham group and in ONC group after 7 days. Next, we investigated the site of VGF expression in the different retinal layers by immunostaining with markers for nerve fibers, Müller glia, and astrocytes. This investigation was performed at 7 days after the ONC because the level of VGF is increased in this time. At 7 days, the immunoreactivity was increased throughout the RNFL + GCL and INL and especially around the cells in the GCL. VGF was co-localized with GS, a marker of Müller glia, in both the ONC and control groups ( Fig. 2A). VGF was partly co-localized with GFAP, a marker of astrocytes (Fig. 2B). On the other hand, VGF was not co-localized with NF-H, a marker of neuronal axons (Fig. 2C).
Expression and localization of VGF in optic nerve. We determined the level and site of expression of VGF in the crushed optic nerve at the anterior site, the crushed site, and the posterior site at 1, 2, 3, 5, and 7 days after the ONC (Fig. 3A-D). VGF was co-localized with GFAP in both the sham group and the ONC group 7 days at the anterior site, and the posterior site after the ONC (Fig. 3E).

AQEE-30, a VGF peptide, suppressed loss of RGCs induced by ONC.
We evaluated the effects of AQEE-30 on the loss of RGCs. Representative photomicrographs of fluorescence-labeled RGCs in flat-mounted retinas are shown in Fig. 4A. First, we found that the number of RGCs was significantly decreased in the vehicle-treated ONC group compared with the sham group (Fig. 4B,C). In addition, we used brimonidine which slightly suppresses the reduction of RGCs in the ONC model for the positive control 30 . As the result, we confirmed that brimonidine reduced loss of RGCs in this model. As the previous reports, administration of BDNF promote RGCs survival about 1.5 ~ 2.0 times in the optic nerve injury model 31,32 . Actually, our study showed that brimonidine prevented ONC-induced RGCs loss in the same way. In addition, brimonidine stimulates trophic factor productions including BDNF 33 . Thus, we considered that brimonidine is suitable as a positive control in this ONC model. Using this model, we found that AQEE-30 slightly suppressed the loss of RGCs after the ONC compared to the loss in the vehicle-treated group (Fig. 4B,C). In addition, AQEE-30 and brimonidine treatment did not change the number of FG-labeled microglia (Fig. 4D,E).

Effects of VGF overexpression on RGCs after ONC.
We investigated whether the ONC-induced RGCs death would be suppressed in VGF-overexpressing mice. In preliminary studies, we confirmed that the expression of VGF protein in the retina of VGF-overexpressing mice was 2.3 times higher than that of WT mice (Fig. 5A).
The retinal structure of VGF-overexpressing mice was not different from that of WT mice ( Fig. 5B-D). Ten days after the ONC, the number of RGCs in the VGF-overexpressing mice was not significantly different from that in WT mice (Fig. 5E,G).
Effects of VGF peptide on RGCs in culture. Finally, we assessed the direct effects of AQEE-30 against RGCs by using rat-derived RGCs and human iPSCs-derived RGCs in culture. In this investigation, BDNF and CNTF were used as positive control because these factors are known to promote outgrowth of RGCs neurites. Similarly, the neurites outgrowth and the number of survival rat-derived RGCs was promoted by AQEE-30 treatment compared with the control group ( Fig. 6A-C). For the human iPSCs-derived RGCs, we evaluated whether AQEE-30 also promoted neurites outgrowth in a concentration dependent manner. As the result, AQEE-30 enhanced neurite outgrowth in a concentration dependent manner and suggesting that 1 µM AQEE-30 is the minimal effective concentration (Fig. 6D,E). Thus, it is estimated that 1 µM AQEE-30 would be needed to expect the neuroprotective effect in the injected retinas. In addition, high-concentration of AQEE-30 treatment (3 µM and 10 µM) did not show toxicity, suggesting that range of AQEE-30 concentration within 1-10 µM would not be toxic for RGCs.

Discussion
The results showed that VGF was associated with the retinal and optic nerve degeneration induced by the ONC, and that AQEE-30 had neuroprotective effects against the ONC. We found that the expression of the mRNA of the Vgf gene and the VGF protein were upregulated after the ONC. In addition, immunohistochemistry showed that VGF was expressed in the glial cells but not in neuronal axons. Moreover, AQEE-30 slightly suppressed the reduction of the number of RGCs after the ONC, and it also promoted neurite outgrowth in rat-derived RGCs and human iPSCs-derived RGCs. Neurotrophic factors are known to be secreted peptides that influence neuronal survival, development, and function. Although previous studies have shown that neurotrophic factors such as BDNF, NGF, CNTF, glial cell-line derived neurotrophic factor (GDNF), and basic fibroblast growth factor (bFGF) were found to promote RGCs survival [34][35][36][37][38] , the application of these factors as medications has not been realized. The difficulties with these factors are the low bioavailability and the necessity of frequent administration to have an effect due to their large molecular weight and short half-life 37 . Similar to BDNF, the effects of VGF peptides including those promoting neuronal survival and regulating synaptic plasticity are associated with BNDF-TrkB mediated signaling pathway in brain neuronal cells 38,39 . Conversely, the molecular weight of VGF peptides is much lower than that of BDNF. Therefore, VGF should be considered as a potential agent to treat axonal injury. However, the molecular function of VGF has not been determined in retina although the function of VGF in brain is known. In this study, the mRNA of Vgf was increased in the retina at 2, 3, and 5 days after the ONC. Furthermore, the expression of the VGF protein was upregulated 7 days after the ONC especially in the RNFL + GCL. The results of previous studies have indicated that the mRNA levels of both BDNF and bFGF were increased at 3 days after optic nerve transaction 40 rapidly from 3 to 7 days after the ONC 42 . Therefore, these findings suggest that the mRNA of Vgf and VGF protein increased with the loss of RGCs induced by the ONC as well as the neurotrophic factors. VGF is expressed in neuronal cells such as the hippocampal neurons and dorsal root ganglion neurons in the central nerve systems (CNS) [43][44][45] . According to these reports, VGF also seemed to be localized in the retinal neurons including the RGCs before investigation. Interestingly, VGF was localized in glial cells including the Müller glia and astrocytes but not in neuronal axons. Glial cells such as astrocytes and Müller glia are known to provide trophic support for retinal neurons 46,47 . For example, Müller glia produce neurotrophic factors including BDNF 48 . Importantly, glial cells undergo morphological changes and alterations of their gene expression after injury or disease which is called reactive gliosis 49,50 . Genomic analysis showed that the gene of Vgf and Bdnf were upregulated after reactive gliosis in the mouse brain 51 . Additionally, microglia were reported to be the source of BDNF for neurons and contribute to neuroprotection and functional recovery 52,53 . In the ONC model, reactive gliosis occurred in the retina 54 , and an activation of astrocytes and migration of microglia were observed in the optic nerve after the ONC 55 . We also observed changes of the astrocyte-like immunoreactivity in the retina and the optic nerve after the ONC along with an increase in the VGF expression after the ONC. Thus, these findings suggest that an increase of VGF occurred with the reactive gliosis. In Fig. 3A-D, VGF increased except for 3 days in the optic nerve after the ONC. Previous reports indicated that astrocytes, which expressed VGF protein, started to disappear from the crush site as early as 2 days after the ONC 56 . On the other hand, activation of astrocytes and scar forming was occurred at 7 days after the ONC in the optic nerve 13 . Therefore, we estimated that VGF might not increase at 3 days after the ONC and might increase from 5 days to 7 days after the ONC. Although the level of endogenous VGF was increased, the number of RGCs was reduced. The reason for this is that the time when the expression of VGF was upregulated after the ONC was considered to be too late for protecting RGCs. Because this time coincides with the late stage of reducing the number of RGCs induced by the ONC, a supply of VGF is thought to be required at the early stage. In fact, although it was reported that CNTF To administrate VGF, we considered what VGF peptide is appropriate. The processed VGF peptides can be divided into two groups; one is the peptides including TLQP-21 and AQEE-30 that are generated mainly by cleavage of VGF C-terminal, and the other including NERP-1 and NERP-2. It has been reported that NERP-1 and NERP-2 regulate water homeostasis and secretion of vasopressin in the hypothalamus 25,58 . On the other hands, TLQP-21 and AQEE-30 are involved in neuronal activity 59,60 . It was recently shown that AQEE-30 but not TLQP-21 has neuroprotective effects against the Huntington's disease model 27 . Huntington's disease is a neurodegenerative disease that results in the death of neuronal cells within the basal ganglia and the cerebral cortex. In primary retinal cells, the neuroprotective effect of VGF against RGCs has been shown 29 . Therefore, AQEE-30 is most likely involved in the neuroprotective effects on the RGCs in the ONC model. As this prediction, VGF treatment by an intravitreal injection of AQEE-30, inhibited the reduction of the RGCs. Additionally, AQEE-30 treatment promoted the outgrowth of neurites in RGCs cultures. According to these findings, it can be inferred that AQEE-30 has direct effects on RGCs. Consequently, AQEE-30 has promise for protecting RGCs, although the neuroprotective mechanism of AQEE-30 has not been revealed. To explore the potency of AQEE-30 for a neuroprotective agent, more studies are needed. For example, AQEE-30 treatment did not change the number of FG-labeled microglia in this study. Generally, the number of FG-labeled microglia, which approximately correlate with phagocytic activity of microglial cells and FG-labeled RGCs are inverse correlation at 10 days from ONC 61 . Considering this study, FG-labeled microglia are expected to decrease in the VGF and brimonidine treatment group because VGF and brimonidine inhibited loss of RGCs. However, FG-labeled microglia did not change in these groups compared with the vehicle group. We did not determine why FG-labeled microglia did not change by administration of VGF and brimonidine in this study. As one of several possibilities, we possibly speculate the involvement of BDNF on microglia. The reason why is that VGF and brimonidine have the potential to activate BDNF-TrkB receptor-mediated signaling pathways 33,38,39,62 . Moreover, BDNF enhances microglial phagocytic activity 63,64 . Thus, we cannot exclude the possibility that VGF and brimonidine would enhance microglial phagocytic activity via increasing BDNF production.
In spite of the neuroprotective activity of AQEE-30 for RGCs, VGF overexpression did not protect RGCs death induced by the ONC. In the VGF-overexpressing mice, the retinal VGF expression was about 2.4 fold higher than that of WT mice. Regarding the amount of endogenous VGF, previous reports have indicated that the concentration of VGF peptides in mouse plasma was 5 pmol/ml 65 . According to this report, it is assumed that VGF peptides expression in the VGF-overexpressing mouse plasma was about 11 pmol/ml. In contrast, the treating concentration of AQEE-30 was 50 µmol/L. For this reason, the endogenous VGF in VGF-overexpressing mice retina was quiet low. As a result, the retinal expression levels of VGF in VGF-overexpressing mice might not be enough to protect the RGCs.
In conclusion, we identified the close association between VGF and neuronal degeneration induced by the ONC and neuroprotective effects of VGF peptides. Therefore, treatments targeting the VGF should be considered in RGCs degeneration. Animals. Eight-week-old male ddY albino mice (35-40 g body weight) and adult Sprague-Dawley rats (250-350 g body weight) were purchased from Japan SLC (Hamamatsu, Japan). Neonatal rats were obtained from breeding the purchased rats. Fifteen-weeks-old male VGF-overexpressing mice (25-40 g body weight) were generated as described in detail 67 . Animals were housed under 12 h:12 h light:dark lighting conditions with a standard diet (CLEA Japan, Inc., Tokyo, Japan). Optic nerve crush model. The ONC was performed as described in detail 68,69 . Briefly, mice were anesthetized with a mixture of 120 mg/kg ketamine (Daiichi-Sankyo, Tokyo, Japan) and 6 mg/kg xylazine (Bayer Health Care, Tokyo, Japan) for the surgery. The optic nerve of the left eye was exposed and crushed at 0.5-1.5 mm posterior to the eye. In sham group, only the exposure of optic nerve was performed. After the ONC, ofloxacin ointment (Daiichi-Sankyo) was applied topically.

Labeling and histological analysis of retinal ganglion cells (RGC).
To determine the number of RGCs surviving after the ONC, the RGCs were retrogradely labeled with fluorogold three days before the ONC as described in detail 68 . Briefly, 1 µL of 4% fluorogold (Biotium, Hayward, CA, USA) was injected at a depth of 2 mm from the brain surface into the superior colliculus under anesthesia.
Ten days after the ONC, mice were euthanized by cervical dislocation, and the eyes were enucleated. The eyes were fixed in 4% paraformaldehyde (PFA) at 4 °C, and after 6 to 7 h, the retinas were isolated from the retinal pigment epithelium, and flat-mounted on glass slides. To quantify the number of RGCs, the retinas were examined with a BX50 fluorescence microscope and photographed with a DP30BP camera (Olympus, Tokyo, Japan). Fluorescent images were produced using the MetaMorph software (Universal Imaging Corp., Downingtown, PA, USA). To determine the percentage of RGCs surviving, the number of stained RGCs was counted manually at 0.5 mm (central), 1.5 mm (middle) and at 2.5 mm (peripheral) from the optic nerve head. We distinguished FG-labeled RGCs from FG-labeled microglia as the previous reports 70- 73 . Briefly, FG-labeled elongated or irregularly shaped cells are regarded as microglia. These cells are both larger and smaller than RGCs. The averages of 12 counts/retina were used for the statistical analyses.

RNA extraction and real-time PCR.
Mice were killed on 1, 2, 3, 5, and 7 days after the ONC, and the eyes were quickly removed. The retinas were carefully isolated and frozen in liquid nitrogen. The isolation of the RNA from the retinas and quantification of gene expression were performed as reported in detail 63 . The expression of glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used for the internal control. The sequences of the primers were; for Vgf, 5′-CAGGCTCGAATGTCCGAAAG-3′ (forward) and 5′-CTTGGATAAGGGTGTCAAAGTCTCA-3′ (reverse) for Gapdh, 5′-TCTGCAAGAGACTTCCATCCAGT-3′ (forward) and 5′-TCTGCAACTGCATCATCGTTGT-3′ (reverse).
Immunostaining. Sections were prepared and stained as reported in detail 74  Histological analysis. Eyes were isolated from VGF-overexpressing and wildtype (WT) mice, and immersed in 4% PFA for 24 h at 4 °C. The fixed eyes were dehydrated in ethanol, cleared with xylene, and embedded in paraffin. Six 5 μm sections were cut through the optic nerve head, and the sections were stained with hematoxylin and eosin. Light microscope images were photographed (Micro Publisher 5.0 RTV; QIMAGING, Surrey, BC, Canada). The cells in the RGC layer and the thicknesses of the IPL, inner nuclear layer (INL), outer plexiform layer (OPL), and outer nuclear layer (ONL) were measured between 375 and 625 μm from the optic disc. The results from 3 sections which were selected randomly from the six sections were averaged for each eye.
Immunocytochemistry and neurite outgrowth evaluation. The plated cells were fixed in 4% PFA for 15 min at room temperature. The cells were blocked with non-immune serum for 30 min at room temperature and then incubated with rabbit anti-class III β-tubulin (Tuj-1) monoclonal antibody (1:1000; Biolegend, San Diego, CA, USA) overnight at 4 °C. On the following day, the cells were labeled with Alexa Fluor ® 546 donkey anti-rabbit IgG (1:1000) for 1 h. The nuclei were stained with Hoechst 33342, and the sections were mounted in Fluoromount.
To evaluate the rat-derived RGCs, 9 photographs/well were taken at random sites with a fluorescence microscope (BZ-X710; Keyence, Kyoto, Japan). For this assay, calcein-AM and Tuj-1 positive-RGC were defined as live RGCs, and the neurite length of these cells were measured semi-automatically with the BZ-X analyzer (Keyence).
For the analysis of human iPSCs-derived RGC, 5 images/well were taken with a fluorescence microscope (BZ-9000; Keyence), and the lengths of the neurites of the Tuj-1 positive cells were measured semi-automatically.
Statistical analysis. Data are presented as the means ± standard error of the means (SEMs). Statistical comparisons were made by Student's t-tests or Dunnett's tests with the SPSS Statistics software (IBM, Armonk, NY, USA). A P value < 0.05 was taken to be statistically significant.