p58IPK Is an Endogenous Neuroprotectant for Retinal Ganglion Cells

p58IPK is an endoplasmic reticulum (ER)-resident chaperone playing a critical role in facilitating protein folding and protein homeostasis. Previously, we have demonstrated that p58IPK is expressed broadly in retinal neurons including retinal ganglion cells (RGCs) and loss of p58IPK results in age-related RGC degeneration. In the present study, we investigate the role of p58IPK in neuroprotection by in vitro and in vivo studies using primary RGC culture and two well-established disease-relevant RGC injury models: retinal ischemia/reperfusion (I/R) and microbead-induced ocular hypertension. Our results demonstrate that in both in vivo models, p58IPK −/− mice exhibit significantly increased RGC loss compared to wild type (WT) mice. In vitro, p58IPK-deficient RGCs show reduced viability and are more susceptible to cell death induced by the ER stress inducer tunicamycin (TM). Overexpression of p58IPK by adeno-associated virus (AAV) significantly diminishes TM-induced cell death in both WT and p58IPK −/− RGCs. Interestingly, we find that loss of p58IPK leads to reduced mRNA expression, but not the protein level, of mesencephalic astrocyte-derived neurotrophic factor (MANF), a neurotrophic factor that resides in the ER. Treatment with recombinant MANF protein protects R28 retinal neural cells and mouse retinal explants from TM-induced cell death. Taken together, our study suggests that p58IPK functions as an endogenous neuroprotectant for RGCs. The mechanisms underlying p58IPK’s neuroprotective action and the potential interactions between p58IPK and MANF warrant future investigation.


INTRODUCTION
Accumulating evidence suggests that disturbed protein homeostasis in the endoplasmic reticulum (ER), or ER stress, is a significant contributing factor to neurodegeneration in the central nervous system including the retina . Prolonged ER stress activates pro-apoptotic genes such as C/EBP homologous protein (CHOP) resulting in apoptosis, while reducing ER stress alleviates cell death of retinal neurons, such as retinal ganglion cells (RGCs), induced by a variety of disease-provoking insults (Hu et al., 2012;Jing et al., 2012;Chen et al., 2014;Huang et al., 2015). In addition, ER stress reduces the production of neurotrophic factors but increases the secretion of pro-inflammatory cytokines, which further exacerbate neuronal death (Zhong et al., 2012;Kim et al., 2014;Cai et al., 2016). Thus, restoring ER protein homeostasis may provide a new strategy to prevent and treat retinal neurodegeneration.
The ER chaperone p58 IPK is an important component of the ER protein folding system and is upregulated by the highly conserved unfolded protein response (UPR) genes XBP1 and ATF6 (Lee et al., 2003;van Huizen et al., 2003). Activation of p58 IPK facilitates protein folding, reduces ER stress, and promotes cell survival (Rutkowski et al., 2007). In pancreatic β cells, deletion of p58 IPK results in CHOP-mediated apoptosis and reduced insulin production (Huber et al., 2013). In a previous study, we demonstrated that p58 IPK is present throughout the retina and mice lacking p58 IPK display increased susceptibility to retinal injury caused by glutamate toxicity and age-related cell death (Boriushkin et al., 2015). In R28 retinal neural cells, p58 IPK overexpression suppresses ER stress and improves cell survival under conditions of oxidative stress (Boriushkin et al., 2015). Furthermore, p58 IPK acts as a potent inhibitor of NLRP3 inflammasome and reduces IL-1β secretion in bone marrow macrophages (Boriushkin et al., 2016). These findings suggest that p58 IPK may function as a neuroprotectant and play an important role in maintaining the viability of retinal neurons such as RGCs in pathogenic conditions related to ischemic retinopathy and glaucoma.
In the present study, we tested this hypothesis in two commonly used RGC injury models induced by retinal ischemia/reperfusion (I/R) or ocular hypertension and explored the potential mechanism of p58 IPK neuroprotection and its interaction with mesencephalic astrocyte-derived neurotrophic factor (MANF), an ER stress responsive neurotrophic factor.

Animals
The generation and maintenance of p58 IPK knockout mice were previously described elsewhere (Ladiges et al., 2005;Boriushkin et al., 2015). All experimental procedures were performed in compliance with protocols approved by the Institutional Animal Care and Use Committees at the State University of New York at Buffalo.

Isolation and Culture of Mouse RGCs
RGCs were isolated from retinas of postnatal day 6 neonatal mice using the Miltenyi Biotec magnetic cell sorting (MACS) system following a published protocol (Huang et al., 2003;Jiao et al., 2005). Immediately after isolation, cells were incubated with adeno-associated virus (AAV)-GFP or AAV-p58 IPK (Vector Biolabs, Malvern, PA, USA) at 10 ∧ 12 GC/ml following standard procedures approved by the Institutional Biosafety Committee at the University at Buffalo. After 24 h, transduced cells were treated with 1 µg/ml tunicamycin (TM) or a vehicle control (0.05% DMSO) for an additional 16 h. Cell viability was examined using the live/dead cytotoxicity assay (Molecular Probes) following the manufacturer's protocol. Images were analyzed for live cells (green) and dead cells (red), blind to treatment or genotype (see Supplementary Material).

Retinal Ischemia/Reperfusion (I/R) Mouse Model
The anterior chamber of anesthetized mice was canulated with a 30g needle attached to a reservoir of sterile PBS. Retinal ischemia was induced by elevating the reservoir to generate a hydrostatic pressure of 90 mmHg for 60 min. After 7 days, mice were sacrificed and retinal whole-mounts were prepared for Tuj1 immunostaining to visualize RGCs (see Supplementary Material).

R28 Cell Treatment and Analysis
R28 cells were cultured as described previously (Boriushkin et al., 2015). At approximately 60%-70% confluence, cells were pretreated with 50 ng/ml recombinant human MANF (hMANF; PeproTech) or vehicle for 16 h and then treated with 5 µg/ml TM or vehicle for 24 h. Cells were labeled using the live/dead cytotoxicity assay, photographed, and analyzed (see Supplementary Material).

Quantitative RT-PCR
Total RNA was isolated with Trizol (Invitrogen) per manufacturer instructions. cDNA was made from 500 ng of total RNA using Bio-Rad cDNA kit followed by quantitative RT-PCR (qPCR) with Bio-Rad iQ Sybr Green Supermix. The following primers were used: MANF f1: 5'-CAC CAG CCA CTA TTG AAG AAG A-3' with MANF r1: 5'-AGC ATC ATC GTG GTC CAA-3'; MANF f2: 5'-AAA GAG AAT CGG TTG TGC TAC T-3' with MANF r2: 5'-CCA GGA TCT TCT TCA GCT CTT T-3'; 18sF 5'-GTA ACC CGT TGA ACC CCA TT-3' with 18sR 5'-CCA TCC AAT CGG TAG TAG CG-3'. Retinal cryosections (20 µM) from 2-month-old wild type (WT) and p58 IPK−/− mice labeled with primary antibodies against various retinal markers and a fluorescent secondary antibody (red). Anti-Pax6 labels amacrine cells and some RGCs; anti-PKCα labels rod bipolar cells and their dendrites as well as a subset of amacrine cells; anti-Ribeye labels ribbon synapses between bipolar cells and photoreceptors in the OPL, as well as synapses between amacrine cells and RGCs in the IPL; anti-calretinin labels three synaptic lamina within the IPL. There is no clear difference between WT and p58 IPK−/− retina in labeling for any of these markers. All sections are counterstained with DAPI (blue) to label nuclei. n = 3 mice per group, scale bar = 50 µm. (B) Retinal whole mounts from WT and p58 IPK−/− mice at 7 days after retinal I/R were stained with RGC marker, Tuj1. Representative images are shown on the left. Graph shows that the percentage of RGC loss, as determined from wholemount retinal Tuj1 staining, is significantly higher in p58 IPK−/− mice compared to WT. Mean ± SD, n = 5 mice. * p < 0.05 (Student's t-test). Scale bar = 100 µm. (C) Graph depicts the average intraocular pressure (IOP) of sham control eyes and microbead-injected eyes in WT (+/+) and p58 IPK−/− mice at the time points as indicated. (D) Retinal whole mounts from WT and p58 IPK−/− mice at 14 days after microbead injection were stained with RGC marker, Tuj1. Representative images are shown on the left. Bar graph shows the percentage of RGC loss is significantly greater in p58 IPK−/− mice compared to WT. Mean ± SD, n = 4 mice. * p < 0.05 (Student's t-test).

Statistical Analysis
The quantitative data were expressed as mean ± SD. Statistical analyses were performed using Student's t-test for two-group comparisons and one-way ANOVA with Bonferroni post hoc test for three groups or more. Statistical differences were considered significant at a p value of less than 0.05.

p58 IPK Deletion Does Not Affect Retinal Morphology in Young Adult Mice
We first examined overall retinal morphology of young adult p58 IPK−/− and wild type (WT) mice using a variety of retinal cell type specific markers. We found that there was no significant difference in retinal expression of PKCα, Pax6, ribeye, or calretinin in p58 IPK−/− and WT mice ( Figure 1A). The retinal layer thickness was identical in these mice (not shown). We conclude that the retinas of p58 IPK−/− mice were morphologically indistinguishable from WT at the ages we performed the retinal injury interventions described below.

p58 IPK Deficiency Exacerbates RGC Loss Induced by Retinal I/R
Retinal I/R is a widely used model for investigating the mechanisms and treatment of inner retinal neuron injury caused by acute ischemia. Seven days after I/R, RGC numbers were examined by immunostaining for Tuj1, a putative RGC marker, in retinal whole mounts. We found a 36.6% RGC loss in WT mice and a 52.0% RGC loss in p58 IPK−/− mice ( Figure 1B). This indicates that I/R induced a significantly higher rate of RGC loss (a 42% increase) in p58 IPK−/− mice than in WT, suggesting a protective role of p58 IPK in RGCs.

Loss of p58 IPK Increases RGC Damage in Microbead-Induced Ocular Hypertension
We used a second model that is highly relevant to glaucoma to validate the role of p58 IPK in RGC protection. We found no significant difference in IOP elevation after microbead injection in p58 IPK−/− and WT mice ( Figure 1C). Two weeks after induction, WT retinas with increased IOP exhibited a 20% loss of RGCs; the percentage of RGC loss increased significantly to over 35% in p58ipk −/− mice ( Figure 1D). This again FIGURE 2 | Overexpressing p58 IPK by adeno-associated virus (AAV) protects RGCs from endoplasmic reticulum (ER) stress-induced cell death. (A) Cultured primary mouse RGCs isolated from p58 IPK−/− and WT retinas were infected with AAV-GFP or AAV-p58 IPK 24 h prior to a 16 h-treatment with tunicamycin (TM; 1 µg/ml) or control. Cells are labeled with the live/dead cytotoxicity assay (calcein-AM to label live cells (green) and ethidium homodimer-1 to label dead cells (red)). Representative images of live/dead assay are shown above. Graph depicts the survival percentage of RGCs from WT and p58 IPK−/− mice. Under control conditions, p58 IPK−/− RGCs infected with AAV-GFP have a significantly lower survival percentage than the matched WT group. RGCs from both WT and p58 IPK−/− mice infected with AAV-GFP show a significant decrease in survival percentage after TM treatment compared to control treatment. The survival percentage is significantly increased for RGCs infected with AAV-p58 IPK , compared to infection with AAV-GFP for both WT and p58 IPK−/− mice. Mean ± SD, n = 3 independent experiments. * p < 0.05 (one-way ANOVA with Bonferroni post hoc test). Scale bar = 80 µm. (B) Quantitative RT-PCR (qPCR; top) using total RNA from whole retina from 2-month-old WT and p58 IPK−/− mice. Two sets of independent primers specific for mesencephalic astrocyte-derived neurotrophic factor (MANF) reveal a mean significant 50% reduction in mRNA level in p58 IPK−/− retina compared to WT. All qPCR data are normalized to 18 s. Mean ± SD, n = 3 mice per genotype. * p < 0.05 (Student's t-test). Western blot analysis (bottom) of whole retina lysate from WT or p58 IPK−/− mice reveal that the level of MANF protein is not statistically different for whole retina (n = 6 mice per group). (C) R28 cells in culture exposed to 16-h pre-treatment with 50 ng/ml MANF or vehicle, followed by 5 µg/ml TM or control treatment for 24 h. Cells are labeled with the live/dead cytotoxicity assay as in (A). Quantification of the survival of R28 cells reveals that approximately 60% of R28 cells survive TM treatment in the absence of MANF. With MANF pre-treatment the survival percentage is significantly increased to over 85%. Mean ± SD, n = 4 independent experiments; * p < 0.05, * * p < 0.01 (one-way ANOVA with Bonferroni post hoc test). Scale bar = 100 µm.
indicates that a loss of p58 IPK results in an increase in RGC death.

Overexpression of p58 IPK Protects RGCs From ER Stress-Induced Cell Death
Next, we examined whether overexpressing p58 IPK by AAV is sufficient to protect primary mouse RGCs isolated from p58 IPK−/− and WT mice. Under control conditions, we found that the percentage of RGCs survival in culture was significantly decreased in p58 IPK−/− RGCs compared to WT (Figure 2A), although this decrease was not rescued by acute overexpression of p58 IPK . TM exposure caused a significant loss of cells in both p58 IPK−/− and WT RGCs infected with control virus (Figure 2A). Overexpressing p58 IPK led to a significant recovery to near control levels of survivability in both WT and p58 IPK−/− RGCs.
Reduced Retinal mRNA Expression of MANF in p58 IPK−/− Mice MANF is a member of a newly identified ER-localized neurotrophic factor family and is upregulated during ER stress (Glembotski et al., 2012). We investigated whether MANF expression is altered in p58 IPK−/− retinas. Using qPCR, we found that MANF mRNA expression was significantly reduced in the retina of p58 IPK−/− mice ( Figure 2B). However, the protein level appeared to be insignificantly different ( Figure 2B). To further explore whether there is a reciprocal regulation between p58 IPK and MANF, we transduced R28 cells with adenovirus to overexpress p58 IPK and then treated the cells with TM to induce ER stress. We found that both MANF and p58 IPK levels were significantly increased after TM treatment (Supplementary Figure S1A). Overexpression of p58 IPK did not alter MANF protein expression in unstimulated cells, but significantly reduced the induction of MANF by TM (Supplementary Figure S1A).

Recombinant MANF Protects R28 Cells or Retinal Explants From ER Stress Induced Cell Death and Caspase-3 Activation
To examine whether MANF is important for neuronal survival under ER stress, we pretreated R28 cells with 50 ng/ml recombinant hMANF protein or vehicle (0.1% PBS) followed by treatment with 5 µg/ml TM, or vehicle, for 24 h. We find TM treatment induces cell death in 43% of R28 cells in the absence of MANF treatment ( Figure 2C). With MANF pretreatment, this number was reduced to 16%, which represents a recovery to 91% of the survival rate of control R28 cells treated with vehicle ( Figure 2C). In addition, MANF treatment significantly reduced TM-elicited caspase-3 activation in ex vivo cultured retinal explants (Supplementary Figure S1B).

DISCUSSION
The findings from the present study provide strong evidence for an endogenous role of p58 IPK in protection of RGCs in conditions of ER stress, acute ischemia and ocular hypertension. Previous studies, including our own, have implicated p58 IPK in a broad range of neurodegenerative diseases and diabetes (Ladiges et al., 2005;Boriushkin et al., 2014Boriushkin et al., , 2015Boriushkin et al., , 2016Han et al., 2015;McLaughlin and Zhang, 2015). Clinically, patients bearing loss-of-function mutations in the DNAJC3 gene, which encodes p58 IPK , suffer from diabetes and multisystemic neurodegeneration (Synofzik et al., 2014). In animal models, loss of p58 IPK results in increased ER stress and oxidative stress, decreased β cell function, progressive cell death and diabetes in male p58 IPK KO mice (Ladiges et al., 2005;Han et al., 2015). In contrast to the relatively thoroughly studied role of p58 IPK in diabetes, how p58 IPK regulates neuronal function remains unexplored. Herein, we examined the impact of p58 IPK deficiency on RGC survival in disease-relevant models. Our data clearly indicate an important role of p58 IPK in RGC survival during acute retinal ischemia and in ocular hypertension. Notably, we used young adult p58 IPK KO mice that demonstrate no significant morphological defects in the retina compared to their age-matched littermate controls. In addition, most of our experiments were carried out in female mice at 2 months of age, which do not experience increased blood glucose levels as some male p58 IPK KO mice do with age. Our in vitro results suggest a therapeutic potential of overexpressing p58 IPK or enhancing its function for neuroprotection of RGCs, which warrants an in vivo study in disease models. Another interesting finding is that there is likely a reciprocal regulation between p58 IPK and MANF in retinal cells undergoing ER stress. Recent studies show that MANF protects CNS neurons (Voutilainen et al., 2017) and markedly improves RGC survival in a rat glaucoma model (Gao et al., 2017). Our data support these findings and demonstrate that MANF inhibits ER stress-induced cell death and caspase-3 activation in retinal cells and explants. In R28 cells, both p58 IPK and MANF are upregulated during ER stress; interestingly, overexpressing p58 IPK reduces ER stress-stimulated MANF upregulation. This suggests a potential interaction between the two chaperones, which both bind to GRP78 and regulate the UPR, although it remains unclear whether MANF secretion is affected by p58 IPK manipulation. Future studies will investigate how MANF and p58 IPK are regulated in RGCs and whether overexpression of p58 IPK and MANF would synergistically promote RGC survival and function.

AUTHOR CONTRIBUTIONS
TM designed and performed the experiments, analyzed the data and drafted the manuscript. ND and JL performed the experiments, analyzed the data and revised the manuscript. JW and SZ conceived and designed the study, reviewed all the data, drafted and revised the manuscript. All authors read and approved the final manuscript.

FUNDING
This work was supported by National Institutes of Health/National Eye Institute (NIH/NEI) grants EY019949 and EY025061, and an Unrestricted Grant to the Department of Ophthalmology, SUNY-Buffalo, from Research to Prevent Blindness.