Unfolded Protein Response Pathways Correlatively Modulate Endoplasmic Reticulum Stress Responses in Rat Retinal Müller Cells

Background Endoplasmic reticulum stress (ERS) in the retinal Müller cells is a key factor contributing to the retinal inflammation and vascular leakage in diabetic retinopathy (DR). This study was to investigate the underlying mechanisms through which the 3 main unfolded protein response (UPR) pathways regulate ERS and to examine the expression levels of vascular endothelial growth factor (VEGF) in Müller cells in vitro. Methods Rat Müller cell lines were stimulated with high glucose to mimic a diabetic environment in vitro. PKR-like endoplasmic reticulum kinase (PERK), inositol-requiring enzyme 1 (IRE1) and activating transcription factor 6 (ATF6) were downregulated or upregulated with shRNA or overexpression plasmids. The transfected Müller cells were cultivated in high glucose medium for 48 hours. Expression of glucose-regulated protein 78 (GRP78), activating transcription factor 4 (ATF4), X-box binding protein 1 (XBP1), ATF6, and VEGF was examined with immunofluorescence and western blot. Results Our data indicated that ERS was found in both high glucose and osmotic control groups. Overexpression or downregulation of UPR pathways effectively increased or reduced the production of GRP78, ATF4, XBP1, ATF6, and VEGF, respectively. These 3 signaling pathways had similar regulatory effects on VEGF. Conclusion The 3 UPR-mediated inflammatory pathways were dependent on each other. Inhibition any of these signaling pathways in UPR might be a potential therapeutic target for DR.


Introduction
Diabetic retinopathy (DR) is a severe complication of diabetes and one of the leading causes of binocular blindness [1]. Its incidence increased considerably along with diabetes during last decades [2]. e global incidence of DR is predicted to rise dramatically from an estimated 127 million people in the year 2010 to 191 million by 2030 [3]. Several treatments are available for DR, including photocoagulation [4], vitrectomy [5], and repeated intraocular injections of steroids [6]; however, adverse effects are common [7].
Müller cells are principal glial cells in the retina to maintain retinal homeostasis [8].
ey are rich in endoplasmic reticulum (ER) and express growth factors to nourish retinal neurons and capillary cells [9]. Vascular endothelial growth factor (VEGF) is produced by Müller cells [10] in the retina and activated at the early stages of DR [11]. Anti-VEGF therapy, as a major strategy to treat DR, was demonstrated to alleviate retinal inflammation and vascular leakage in diabetic patients [12][13][14]. However, inhibition of VEGF by anti-VEGF antibodies resulted in systemic adverse effects such as thromboembolic events, hemorrhage, gastrointestinal perforation, hypertension, and nephrotic syndrome [15]. erefore, investigation of the key mechanisms that regulate inflammation in Müller cells may elucidate new therapeutic targets to prevent retinal complications of diabetes.
e ER is the primary cellular organelle responsible for the synthesis and processing of proteins [16]. Protein folding in the ER may be disturbed by various physiologic and pathologic conditions, causing endoplasmic reticulum stress (ERS) [17]. To restore homeostasis under ERS, cells usually trigger the unfolded protein response (UPR), which alleviates the accumulation of misfolded or unfolded protein [18]. UPR is transduced by 3 main ER-resident stress sensors, PKR-like endoplasmic reticulum kinase (PERK) [19], inositol-requiring enzyme (IRE) 1 [20], and activating transcription factor (ATF) 6 [21]. At normal condition, these 3 proteins are bonded to glucose-regulated protein (GRP) 78 associated with the internal membrane. Upon ER stress, GRP78 is upregulated rapidly and then dissociates from the ER membrane to enter the lumen [22]. Previous studies demonstrated that high glucose induced ERS in Müller cells [23]. VEGF, as an important molecule signaling preinflammation, was apparently induced by ERS [10].
Taken together, high glucose induced retinal ERS reactions. However, the underlying mechanisms and signaling pathways in UPR-mediated stress are yet to be determined. In this study, we used a hyperglycemia cell model to simulate the diabetic environment [24], explored the UPR pathways, and investigated the mechanisms that activated ERS in rat retina Müller cells.

Cell Culture.
Rat retinal Müller cell line (rMC-1) was generously provided by the College of Life Sciences at Wuhan University (Wuhan, China). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 1 g/l glucose with 10% fetal bovine serum, 1% penicillin, and 1% streptomycin, at 37°C in a humidified atmosphere with 5% CO 2 . After washing with 0.25% Tris in the logarithmic growth phase, cells were seeded in 24-well polystyrene plates. e cells were then verified by immunocytochemistry of glial fibrillary acidic protein (GFAP; DAKO, Hamburg, Germany). For the studies of high glucose (HG), cells were starved in serum-free DMEM overnight and treated with or without normal glucose (5 mmol/L), high glucose (30 mmol/ L), or mannitol (25 mmol/L; as an osmotic control) for 48 hours.

Construction of pEGFP-RNA Plasmid.
Total RNA was extracted from rMCs using Trizol (Invitrogen, USA). e reverse transcription reaction was conducted using a reverse transcription kit ( ermo Fisher Scientific, USA) to obtain cDNA according to the manufacturer's protocols. e recombinant pEGFP-N1-ATF6 and pEGFP-N1-PERK plasmids were constructed at HindIII and XhoI restriction endonuclease sites and validated by Sangon Biotech (Shanghai). e pEGFP-IRE1 plasmid was purchased from Zoonbio Biotechnology (Nanjing, China).

Transfection.
Rat Müller Cells were transiently transfected in duplicates in 6-well plates using Lipofectamine 2000 (Invitrogen), following the manufacturer's protocol. e cells (2 × 10 5 per well) were incubated with fresh DMEM medium without antibiotics for 48 hours to obtain 80-90% confluency on the day of transfection. Mock-transfected cells (i.e., cells transfected with Lipofectamine reagent only) served as negative controls.

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR).
Total RNA was extracted from rMCs using Trizol. RNA (2 µg) was used for cDNA synthesis using a Revert Aid First Strand cDNA Synthesis Kit (Invitrogen). Quantitative real-time PCR was performed with Fast Start Universal SYBR Green Master Kit (Rox, USA) on a QuantStudio 5 Real-Time PCR system (Applied Biosystems, USA). e experiments were independently repeated 3 times. PCR primer sequences were as follows: e comparative C T method was used to calculate the mRNA expression levels. e expression of selected genes was normalized to that of the reference gene, β-actin, at each time point and converted to the relative levels as follows: fold expression � 2ΔΔC T , where ΔC T �average C T of target gene − average C T of endogenous control (β-actin) and ΔΔC T �average ΔC T of target sample − average ΔC T of the calibrator sample.

Statistical
Analysis. All data were presented as mean-± standard deviation (SD). Comparisons between 2 groups were conducted with Student's t-test. e average number of samples was compared using single-factor analysis of variance (ANOVA). Multiple comparisons between the groups were tested by the SNK test. P < 0.05 was considered as statistically significant.

Purity of Müller Cell
Culture. Immunocytochemistry was used to determine the purity of Müller cell cultures. e intermediate filament protein GFAP was reported to express in astrocytes and Müller cells [26]. In our study, the cultured cells had strong GFAP staining (Figure 1). e results showed 95% cells were GFAP-positive and derived from Müller cells.

Detection of ERS Markers and VEGF in rMCs Exposed to HG for 48 h.
Our preliminary studies demonstrated that ERS was not fully activated by HG at 24 hours, so we extended the intervention to 48 hours. At this time point, the protein levels of the major markers for ER stress were increased significantly compared with the rMCs exposed to normal glucose (p < 0.05, Figure 2). Moreover, VEGF was also markedly upregulated (p < 0.05), suggesting induced retinal neovascularization, vascular leakage, and perhaps macular edema in DR [27]. Interestingly, our results also indicated that rMCs grown in glucose (5 mmol/L) plus mannitol (25 mmol/L) showed a significant increase in the expression of ER stress markers and VEGF (p < 0.05).

Construction of shRNA and Overexpression Plasmid.
In order to investigate the ATF6 response, 3 different ATF6-shRNAs were designed and their effects were examined by quantitative real-time PCR. All of the 3 shATF6s downregulated ATF6 mRNA levels (Figure 3(a)) in transiently transfected rMC-1, compared with the control group. (p < 0.05). Moreover, shATF6 682 had the strongest downregulated effects. Similarly, 3 shPERKs were transfected into rMCs, and shPERK 1216 showed significant reduction (p < 0.05) on the PERK mRNA levels compared with the control group (Figure 3(b)). e shIRE1 1173 group showed the lowest IRE1 expression followed by the shIRE1 2219 and shIRE1 2732 groups (Figure 3(c)). e recombinant pEGFP-N1-ATF6 and pEGFP-N1-PERK plasmids were used to overexpress ATF6 in rMCs (p < 0.05, Figures 3(d) and 3(e)) compared with the normal group, and elevated levels of IRE1 mRNA were detected in rMCs transfected with pEGFP-N1-IRE1 (Figure 3(f )).

Regulation of ERS Signal Pathways Modulated ATF4 Expression in rMCs.
To examine the effects of the 3 ERS pathways on Müller cells, we treated rMCs with 3 different specific shRNA constructs. Both the immunofluorescence and protein levels of ATF4, a downstream target of PERK, were decreased significantly in HG-exposed in cells transfected with PERK-shRNA (p < 0.05, Figure 4). We next investigated whether inhibition the ATF6 and IRE1 pathways downregulated PERK. e results of immunoblotting analysis indicated that knockdown of ATF6 or IRE1 significantly decreased ATF4 expression in Müller cells under HG condition (p < 0.05). In order to identify the effects of overexpression of these proteins, rMCs exposed to HG were transfected with pEGFP-ATF6, pEGFP-PERK, and pEGFP-IRE1, respectively. rMCs transfected with pEGFP-PERK showed the highest expression of ATF4 after HG treatment (Figures 4(c) and 4(e)). To gain more insight into the 3 UPR channels, we inhibited 2 of the 3 pathways in different combinations. Inhibition of both ATF6 and IRE1 resulted in a significant decrease of ATF4 compared with PERK-shRNA alone (p < 0.05, Figures 4(a) and 4(b)). e results of WB demonstrated that ATF4 protein levels in the shPERK + shATF6 group were significantly decreased compared with shPERK alone. However, the expression of ATF4 in the shPERK + shIRE1 and shATF6 + shIRE1 groups showed no significant difference with the shPERK group (p > 0.05, Figures 4(c) and 4(f)).

Regulation of ERS Signal Pathways Modulated XBP1
Expression in rMCs. Similar results were observed for XBP1 expression. XBP1 expression in cells with shIRE1 transfection was evaluated by WB and immunofluorescence after exposure to HG for 48 hours. XBP-1 was downregulated ( Figures 5(a) and 5(b)). However, rMCs with shRNA transfection exhibited higher levels of these markers compared with normal glucose. Both shPERK and shATF6 inhibited XBP1 (p < 0.05), but there was no difference between these groups (Figures 5(c) and 5(d)). rMC transfected with pEGFP-IRE1 showed the highest expression of XBP1 in HG (Figures 5(c) and 5(e)). XBP1 in the shIRE1 group was significantly upregulated compared with the shIR-E1 + shPERK and shIRE1 + shATF6 groups (Figures 5(c) and 5(f )).

Regulation of ERS Signal Pathways Modulated VEGF and GRP78 Expression in rMCs.
Interestingly, the protein levels of GRP78, an upstream target of UPR, were increased significantly in rMCs transfected with the overexpression plasmids after exposure to HG for 48 hours (Figures 7(a) and  7(b)). e expression of GRP78 was significantly reduced. e strongest inhibition of GRP78 was observed in the shATF6 group. When any 2 UPR pathways were suppressed, the expression of GRP78 was lower than that of single inhibition groups. However, the expression of GRP78 among the 2 pathway inhibition groups was not significantly different (p > 0.05, Figures 7(a) and 7(c)). Similar results were obtained for the expression of VEGF. Overexpression of any ERS pathway in rMCs induced a statistically significant increase of Müller cell-derived VEGF in response to HG. No significant difference of VEGF levels was observed between the pEGFP-IRE1 and pEGFP-ATF6 groups (p > 0.05). pEGFP-PERK induced significantly less increase than the other pathways (Figures 7(a) and 7(d)). e VEGF protein levels were reduced to the same extent (p > 0.05) by inhibition of any of the 3 signaling pathways (Figures 7(a) and 7(e)).

Discussion
Compelling evidence suggests that ERS plays an important role in chronic inflammatory conditions such as diabetes [27]. However, how ERS in rMCs under HG promotes VEGF expression remains poorly understood. Previous studies showed that Müller cells are capable of basal VEGF secretion and of VEGF synthesis in HG [23,28,29].
In the presented study, we developed Müller cell lines in vitro and set up a hyperglycemia model to simulate the diabetic environment. Many papers showed that ERS signals and inflammatory factors significantly increased in Müller cells at HG for 48 or 72 hours [18,19]. Our results showed that the protein levels of GRP78, ATF6, XBP1, and ATF4, as well as VEGF, were increased in HG at 48 hours, indicating the presence of ERS. Of note, the rMCs exposed to HG for 48 hours appeared to exhibit higher activation of the ATF6 than PERK and IRE1 pathways (as measured by ATF4 and IRE1, respectively). In addition, overexpressing any pathway caused more increase in ATF6 compared with PERK and IRE1. ese data indicated that the ATF6 pathway was more sensitive to HG.
To examine the relevance of glucose in the hypertonic solution, experiments were performed using mannitol (25 mmol/L) as a hypertonic compound with normal glucose. e results demonstrated that the high-mannitol did also induce a rapid upregulation of ERS markers, suggesting that the ERS was merely induced by the hypertonic solution and was only partially due to glucose in Müller cells. It is an interesting finding, and further consideration was required.
To investigate the individual effects of the 3 UPR pathways, we constructed PERK, ATF6, and IRE1 shRNAs and overexpression plasmids for Müller cells. Our results showed that the 3 UPR pathways were activated at the same time when exposed to HG for 48 hours. Previous studies showed that activated PERK inhibited the initiation step of mRNA translation via phosphorylating eukaryotic initiation factor 2α. Activated ATF6 translocated to the golgi complex to regulate the expression of molecules involved in protein quality control. In addition, ATF6 promoted IRE1amediated splicing of X-box binding protein 1 and its expression [30]. Indeed, we founded that the role of the 3 signaling cascades was not independent after exposure to HG for 48 hours.
In addition, we observed significantly increased levels of GRP78 protein in the overexpression groups, suggesting a positive feedback loop. Furthermore, significant upregulation of PERK, IRE1, and ATF6 pathways might induce accumulation of unfolded or misfolded proteins in the endoplasmic reticulum, leading to the endoplasmic reticulum stress response and increasing GRP78 protein levels. An interesting finding of our study was that UPR branches mediated by IRE-1 had a substantially greater effect on GRP78 expression compared with PERK and ATF6 pathways. Concurrently, GRP78 expression was further decreased when any 2 signaling pathways were suppressed.
Moreover, we demonstrated that regulation of the 3 UPR signaling pathways effectively modulated the production of VEGF. Moreover, our study showed that transfection with PERK, ATF6, and IRE1 shRNAs markedly attenuated VEGF expression in Müller cells compared with the control group, while no significant differences in the levels of VEGF between the 3 shRNA groups. ese observations indicated that the sensitivity of the ERS-triggered pathways was quite similar.

Conclusion
In summary, our data demonstrated that the high glucose or high osmotic pressure produced by HG activated ERS in Müller cells. Regulation of the 3 UPR signal pathways effectively modulated the other 2 signaling cascades and the production of VEGF in Müller cells. Inhibition of these 3 UPR pathways might be a potential therapeutic target for DR.