The Effects of Sonic Hedgehog on Retinal Müller Cells Under High-Glucose Stress

Abstract

Purpose.: To investigate sonic hedgehog (SHH) expression and its effects on retinal Müller cells in diabetic rats and in vitro culture systems under high-glucose stress.

Methods.: Diabetic rats were induced by intraperitoneal injection of streptozotocin. Primary rat retinal Müller cells were grown in medium containing 5.5 or 35 mM glucose with SHH and/or cyclopamine. Retinas' and primary Müller cells' expression of SHH pathway components protein and mRNA were determined by Western blot analysis and real-time PCR. The effects of exogenous SHH and its inhibitor cyclopamine on retinal ganglion cells (RGCs) survival after 3-month diabetes were examined by the counting of Brn-3a–labeled RGCs. Phosphoinositide 3-kinase (PI3K), extracellular signal regulated kinases 1 and 2 (ERK1/2), and P38 were detected by Western blot.

Results.: Both mRNA and protein expression of SHH, SMO, GLI1, and glial fibrillary acidic protein (GFAP) in the retinas of diabetic rats and high-glucose cultured Müller cells increased in a time-dependent manner. Exogenous SHH increased the mRNA and protein expression of SHH, SMO, and GLI1 and cyclopamine reversed that effect. Three months after onset of diabetes, administration of SHH inhibited gliosis significantly and promoted RGC survival. Exogenous SHH upregulated the phosphorylation of PI3K and downregulated ERK1/2, but did not affect the expression of P38.

Conclusions.: Sonic hedgehog signaling pathway was upregulated in diabetic rat retina and high-glucose cultured Müller cells, and SHH exerted neuroprotective effects on damaged RGCs in a rat diabetes model. The neuroprotective effects of SHH may act indirectly, via Müller cells, through PI3K or ERK1/2 pathways.

Diabetic retinopathy (DR) is the most common complication in patients with diabetes and the leading cause of blindness in working-age adults. Most research and treatment efforts have focused on the vascular changes in DR. However, many studies also indicate that alterations of neuroretinal structure and function occur prior to the clinically observable retinal vasculopathic lesions associated with DR.^1–3 Functional alterations, such as the impairment of color vision and contrast and alterations in the electroretinogram, have been detected in diabetic patients before vascular dysfunction.^4–7 Recent reports also show neural alterations at the cellular level, without vascular morphologic changes, in diabetic mice.^8,9 Retinal glial dysfunction causes neuronal dysfunction. Generally speaking, neurons are the ultimate effectors of the nervous system and their responses depend on the normal function of glial cells. Treatments designed to protect the retina from dysfunction and slow the progress of DR at its earliest stages could benefit patients, but these require further investigation.¹⁰ 

The retinal glial cells in the vertebrate retina are grouped into microglia, astrocytes, and Müller cells. Müller cells constitute the main glial cell population in the retina and have the most extensive contact with the bodies and dendrites of ganglion cells. They coordinate vascular responses to meet the metabolic demand of neurons, interchange metabolites, recycle neurotransmitters, and establish the extracellular microenvironment for retinal vessels and neurons.^11,12 Müller cell processes contact the extracellular clefts around each retinal neuron and ensheath neuronal somata and processes. Müller cells support the functioning and metabolism of retinal neurons and are active players in normal retinal function.¹³ They are also profoundly involved in the physiopathology of DR. The activation of retinal Müller cells, evidenced by overexpression of glial fibrillary acidic protein (GFAP), is one of the first histologic indicators of DR.^14,15 However, whether the activated retinal Müller cells exert protective or toxic effects on photoreceptors and neurons is still debated.^13,16,17 

Sonic hedgehog (SHH) is a glycoprotein molecule expressed widely throughout the central nervous system.¹⁸ It is important during neurodevelopment, particularly through its induction of endogenous neural precursor cells and neural stem cells.^19,20 Potential neuroprotective effects of modulation of the SHH signaling pathway have shown some promise as a treatment in animal models of CNS injury.¹⁹ Sonic hedgehog mediates its action via a receptor complex consisting of the transmembrane receptor protein patched and the G protein–coupled coreceptor smoothened.²¹ In the absence of the SHH ligand, Patched represses signal transduction from the smoothened coreceptor, thereby inhibiting the transcription of the SHH-targeted gene. However, after SHH ligand binding to patched, disinheriting leads to activation of a complex signaling cascade that results in transcription of GLI-proteins and other SHH-targeted genes.^22,23 Sonic hedgehog signaling pathways can also be activated in neurons under oxidative stress and play a neuroprotective role.²⁴ 

Mitogen-activated protein kinase (MAPK) have such important members as extracellular signal-regulated kinase (ERK), p-p38 kinase, and phosphoinositol 3-kinase (PI3K),^25,26 which are critical for the regulation of cell proliferation and migration by SHH signaling pathway.^27–29 These pathways play different roles, depending on some factors such as cell types, environmental conditions, and cellular stimuli. The phosphorylation of ERK1/2 is involved in the activation of Müller cells' prosurvival pathways.^30,31 It has been reported that inhibition of ERK can protect retinal neurons from apoptosis.³² And the activation of ERK1/2 increases the apoptosis by inhibiting PI3K/Akt pathway.³³ The PI3K/Akt pathway plays an important role in cell survival and is also crucial for SHH signaling.^29,34 However, the impact of these signaling pathways caused by SHH and mechanisms have not been clearly investigated until now. 

Whether SHH could be involved in the neuroprotective action of retinal Müller cells under high-glucose stress is still unclear. Thus, we investigated the effects of SHH, and its possible mechanisms, on retinal Müller cells and retinal ganglion cells (RGCs) using diabetic rats and primary-cultured retinal Müller cells under high-glucose stress. 

Materials and Methods

This research conformed to the ARVO Statement for Use of Animals in Ophthalmic and Visual Research and was approved by the Animal Ethics Committee of Zhongshan Ophthalmic Center of Sun Yat-sen University, Guangzhou, China. 

Diabetic Induction and Intravitreal Injections of SHH-N or Cyclopamine

We used 72 male Sprague-Dawley rats (6-weeks old; 200–250 g; Zhongshan Ophthalmic Center Animal Center, Guangzhou, China). Diabetes was induced in experimental rats (n = 63), and control rats (n = 9) did not have diabetes. All rats had free access to food and water and were maintained under a 12:12 light–dark cycle at 22°C to 25°C with relative humidity of 40% to 70%. 

Diabetes was induced by intraperitoneal injection of a single dose (60 mg/kg) of freshly prepared streptozotocin solution (Amresco Chemical Co., Solon, OH, USA). Rats with glucose levels higher than 16 mM were considered to be diabetic. Blood glucose was monitored biweekly. Part of diabetic rats were killed with an overdose of 10% chloral hydrate and eyes were removed for further investigation at 1 (n = 9), 2 (n = 9), and 3 months (n = 9) after onset of diabetes without any intervention. Other diabetic rats were assigned to the following conditions: SHH50 (n = 9) received 50 μg/mL concentrations of SHH-N; SHH100 (n = 9) received 100 μg/mL concentrations of SHH-N; cyclo (n = 9) received cyclopamine; PBS group (n = 9) received PBS. 

Recombinant mouse SHH amino-terminal peptide (SHH-N; R&D Systems, Minneapolis, MN, USA) was used to activate the SHH signaling pathway. Sonic hedgehog amino-terminal peptide was freshly prepared at concentrations of 50 or 100 μg/mL in 0.1 M PBS. Cyclopamine (Toronto Research Chemicals, North York, ON, Canada), a specific inhibitor of the SHH signaling pathway, was dissolved at a concentration of 5.0 μg/mL in dimethyl sulphoxide (DMSO; Sigma-Aldrich Corp., St. Louis, MO, USA). Intravitreal injections of SHH-N or cyclopamine were given 2 weeks after diabetes induction and repeated every 2 weeks thereafter. Depending on the condition group, 2 μL SHH-N (50 or 100 μg/mL), cyclopamine, or PBS was injected into the vitreous of the right eye at 2-mm posterior to limbus with a 30-G needle mounted on a Hamilton syringe (Sigma, Shanghai, China). After 3 months, rats were euthanized and right eyes were prepared. 

Immunofluorescence Staining of Frozen Retinal Sections and Whole-Mount Retinas

Rats were anesthetized by intraperitoneal injection of 10% chloral hydrate and perfused via the left ventricle with ice-cold PBS and then 1% paraformaldehyde (100 mL/kg). The fixed eyes were dehydrated in graded sucrose solutions. For retina sections, the anterior segments were removed and eye cups were embedded in optimal cutting temperature compound (Sakura Finetek, Torrance, CA, USA) and frozen in liquid nitrogen. Sections (8 μm) were obtained and stored at −20°C. For whole mounts, retinas were dissected by four radial cuts, which were postfixed for an additional hour. After incubation with 0.05% Triton X-100 in PBS at 37°C for 30 minutes, frozen retinal sections and whole-mount retinas were incubated with a blocking solution at 37°C for 30 minutes. 

Immunofluorescent staining was performed to identify immunoactive cells using mouse monoclonal anti-GFAP 1:100 (Cell Signaling Technology [CST], Beverly, MA, USA), mouse monoclonal anti-Brn-3a 1:500 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), goat polyclonal anti-SHH 1:50 (Santa Cruz Biotechnology), rabbit polyclonal anti-SMO 1:50 (Santa Cruz Biotechnology), and goat polyclonal anti-GLI1 1:50 (Santa Cruz Biotechnology) primary antibodies, respectively, at 4°C overnight. After the frozen sections and whole-mount retinas were rinsed three times with PBS, a secondary antibody conjugated to either Alexa 488 (1:500; Invitrogen, Carlsbad, CA, USA) or Alexa 594 (1:500; Invitrogen) were applied to the samples, which were then incubated for 1 hour at 37°C. Cell nuclei were counterstained with 4′-6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich Corp.) for 5 minutes, and slides were mounted with an antifading fluorescence medium (Vector Laboratories, Burlingame, CA, USA). Images were acquired with a laser confocal microscope (Carl Zeiss, Oberkochen, Germany). 

To quantify Brn-3a-labeled RGCs, retinas were divided into four eccentric zones with respect to the optic nerve head. Five fields in each quadrant (total 20) were selected for counting at ×200 final magnification, starting at the optic disc and progressing to the border at 500-μm intervals. Data were expressed as the relative percentage of RGC loss (mean ± SD) in PBS, SHH50, SHH100, and cyclo groups compared with that in the control group. 

SHH-N or Cyclopamine Treatment for Primary-Cultured Müller Cells Under High-Glucose Stress

Fifteen Sprague-Dawley rats (3–5 days old) were obtained from the Sun Yat-sen University Experimental Animal Center (Guangzhou, China). Retinas were dissected under sterile conditions and pooled together. Cells were dissociated by gentle trituration in a Pasteur pipette and collected by centrifugation at 500g for 5 minutes. They were then cultured in Dulbecco's modified Eagle's medium F12 (DMEM-F12) supplemented with 5.5 mM glucose, 20% fetal bovine serum, streptomycin (100 μg/mL), and penicillin (100 μg/mL) in 25 cm² culture flasks at 37°C in an incubator with 5% CO₂. The medium was changed every 2 to 3 days. Isolated cells were identified by immunofluorescence staining using rabbit polyclonal anti-glutamine synthetase (GS) 1:200 (CST). Cells at passages 3 or 4 were used for experiments. The medium was replenished 24 hours before glucose treatment. Some cells were treated with 35 mM glucose for 0, 3, 12, 24, 36, or 48 hours directly, and some cells were pretreated with SHH-N (3 μg/mL), cyclopamine (20 μg/mL), or both SHH-N and cyclopamine for 30 minutes before 35 mM glucose treatment for 48 hours. During the glucose treatment, cells were maintained in DMEM-F12 containing 2% fetal bovine serum and were treated in 35 mM glucose for up to 48 hours. The medium was changed every 24 hours to maintain a constant glucose level. 

Cell Viability Assay

The cell viability was examined by 3-(4, 5-dimethylthiazole-2-yl)-2, 5-dipenyltetrazolium bromide (MTT) assay. Cells were plated onto a 96-well microplate and cultured for 24, 48, and 72 hours at concentrations of 5.5, 15, 25, 35, 45, or 55 mM glucose. All cells were then treated with MTT at a final concentration of 0.5 mg/mL for 4 hours. After the supernatants were discarded, DMSO was added to dissolve the MTT crystals, and absorbance was measured at a wavelength of 570 nm by a microplate reader (Elx800; Bio-TEK, Winooski, VT, USA). 

Quantitative Real-Time PCR

For mRNA detection, control and diabetic rats were anesthetized with an overdose of 10% chloral hydrate, retinas and primary Müller cells were gathered. Total RNA was extracted using TRIzol reagent (Invitrogen Life Technologies, Grand Island, NY, USA) and reversely transcribed to cDNA using a PrimeScript RT reagent kit (TaKaRa, Dalian, China). Real-time quantitative PCR was run using an ABI Prism 7000 system with SYBR Green PCR kit (TaKaRa). Polymerase chain reaction was performed by denaturing at 95°C for 5 minutes, followed by 40 cycles of denaturation at 95°C (10 seconds), annealing at 60°C (10 seconds), and extension at 72°C (10 seconds). Data were analyzed using the comparative threshold cycle (Ct) method, and results were expressed as the fold difference normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The specific sense and antisense sequences of primers are listed in the Table. 

Table

 

Primers Used for Real-Time PCR

 

Western Blot Analysis

After an overdose of 10% chloral hydrate, control and experimentally treated retinas were freshly dissected and immediately frozen and placed into a lysis buffer (BD Biosciences, San Jose, CA, USA) supplemented with protease inhibitors (Calbiochem; EMD Biosciences, La Jolla, CA, USA) and sonicated. The lysate was centrifuged and the supernatant was collected. 

Müller cells were cultured, pretreated, and glucose treated in the same manner used for the retinas of newborn rats. After the 48-hour glucose treatment, the cells were homogenized in an ice-cold mixture of lysis buffer and protease inhibitors; the homogenate was then centrifuged for 10 minutes (10,000g; 4°C) to remove cell debris. Protein concentration was determined using the BCA assay kit (Biocolor BioScience & Technology Company, Shanghai, China). Subsequently, 30 μg total protein in each group was separated by SDS-PAGE and electroblotted to polyvinylidene fluoride membranes (Millipore, Billerica, MA, USA). After blocking with 5% skim milk, proteins were identified by incubating the membrane with goat anti-SHH (Santa Cruz Biotechnology), anti-GLI1 (Santa Cruz Biotechnology), rabbit anti-SMO (Santa Cruz Biotechnology), extracellular signal regulated kinases 1 and 2 (ERK1/2; CST), anti-p-ERK1/2, (CST), anti-p-PI3K, PI3K (CST), P38 (CST), anti-p-P38 (CST), mouse anti-GFAP (CST), or anti-GAPDH (CST) at 4°C overnight, followed by 1-hour incubation in horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology). Signals were visualized with chemiluminescence (CST) according to the manufacturer's protocol. Western blot bands were subjected to densitometric analysis with ImageJ software (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA). 

Statistical Analysis

Data were expressed as mean ± SD and statistically analyzed with SPSS 14.0 software (SPSS, Inc., Chicago, IL, USA). Multiple groups were compared using a one-way ANOVA with Duncan's multiple pairwise comparison tests. A value of P less than 0.05 was considered significant. Graphs were made using GraphPad Prism 4 (GraphPad Prism, Inc., La Jolla, CA, USA). 

Results

Expression of Upregulated SHH, SMO, GLI1, and GFAP in the Retinas of Diabetic Rats

To quantify the expression of SHH, SMO, GLI1, and GFAP in the retinas of diabetic rats during the process of diabetes, we performed real-time PCR and Western blot analysis of retinal samples at 1, 2, and 3 months after onset of diabetes. In diabetic retinas, both mRNA expression (Figs. 1A–D) and protein expression (Figs. 1E–I) of SHH, SMO, GLI1, and GFAP in the retinas of diabetic rats increased in a time-dependent manner after onset of diabetes and peaked at 2 to 3 months (P < 0.05). We also confirmed the protein expression of SHH, SMO, GLI1, and GFAP in the retinas of both normal and diabetic rats using immunofluorescence staining. In normal retinas, SHH, SMO, and GLI1 were mainly detected in the ganglion cell layer (GCL), but were also weakly present in the inner plexiform layer (IPL). However, the expression of SHH, SMO, and GLI1 was much higher in the GCL, IPL, and outer plexiform layer (OPL) of 3-month diabetic rats than in those of normal rats (Fig. 2). In control retinas, GFAP was not expressed in retinal Müller cells, but its expression was significantly increased in activated Müller cells (in all three retinal layers) of 3-month diabetic rats. 

Figure 1

 

Upregulated expression of SHH, SMO, GLI1, and GFAP in the retinas of diabetic rats. Quantitative analysis of the mRNA expression of SHH (A), SMO (B), GLI1 (C), and GFAP (D) in the retinas of normal and diabetic rats at 1, 2, and 3 months after onset of diabetes. There was a significant increase in SHH, SMO, GLI1, and GFAP mRNA in diabetic retinas compared with controls. (E) The protein expression of SHH, SMO, GLI1, and GFAP in the retinas of normal and diabetic rats. Western blot quantitative data shows significant increases in SHH (F), SMO (G), GLI1 (H), and GFAP (I) proteins in diabetic retinas compared with controls. n = 6 per group. ^★★P < 0.01, compared with control group. Fold changes are shown as mean ± SD.

Figure 2

 

Immunofluorescence staining showing increased expression of SHH, SMO, GLI1, and GFAP in retinas of diabetic rats 3 months after diabetes onset. The expression of SHH, SMO, and GLI1 was much higher in the retinas of 3-month diabetic rats than those of control rats. An increased activation of GFAP was observed in 3-month diabetic retinas found in the GCL, IPL, and INL. Glial fibrillary acidic protein was not detected in the retinal Müller cells in control rats. n = 3 per group. Sonic hedgehog, GLI1, and GFAP staining (green); SMO staining (red); DAPI (blue). Scale bars: 50 μm. INL, inner nuclear layer; ONL, outer nuclear layer.

Effects of SHH-N and Cyclopamine on the Expression of SHH, SMO, GLI1, and GFAP in the Retinas of Diabetic Rats

Real-time PCR (Figs. 3A–C) and Western blot analysis (Figs. 3D–G) revealed that exogenous SHH increased mRNA and protein expression of SHH, SMO, and GLI1 in the retinas of diabetic rats, and exogenous cyclopamine reversed that effect. The opposite result was found for GFAP: SHH-N decreased mRNA and protein expression in the retinas of 3-month diabetic rats and cyclopamine increased them (Figs. 4A–C). In the normal rat retina, exogenous SHH increased protein expression of SHH, SMO, and GLI1, and exogenous cyclopamine reversed that effect (Supplementary Figs. S1A–D). As described earlier, diabetes caused glia activation. Sonic hedgehog amino-terminal peptide treatment reduced the irregular shape and disorderly arrangement of Müller cells, while cyclopamine treatment aggravated the abnormality of glial structures and gliosis. After SHH-N injection, GFAP staining intensity significantly decreased in a dose-dependent manner compared with that in diabetic rats. In contrast, cyclopamine enhanced GFAP expression in the activated retinal glia cells (Fig. 4D). 

Figure 3

 

Changes in expression of SHH, SMO, and GLI1 in retinas of 3-month diabetic rats treated with SHH-N or cyclopamine. Real-time PCR was used to analyze mRNA expression of SHH (A), SMO (B), and GLI1 (C) in retinas of 3-month diabetic rats after treatment with SHH-N or cyclopamine. (D) Western blot analysis of SHH, SMO, and GLI1 expression in diabetic retinas after SHH-N or cyclopamine treatment shows that expression of each was elevated by SHH-N treatment but decreased by cyclopamine treatment. Quantitative analysis of protein expression of SHH (E), SMO (F), and GLI1 (G). n = 6 per group. ^★★P < 0.01, ^★P < 0.05, compared with PBS-treated diabetic rats.

Figure 4

 

The changed activation of GFAP in retinas of SHH-N– or cyclopamine-treated diabetic rats. (A) The effect of SHH-N on mRNA expression of GFAP in the retinas of 3-month diabetic rats treated by SHH-N or cyclopamine. The mRNA level of GFAP was decreased by SHH-N treatment but elevated by cyclopamine treatment. (B) Representative Western blot analysis of GFAP in the retinas of 3-month diabetic rats treated by SHH-N or cyclopamine. (C) Quantitative analysis of the change in GFAP expression in the retinas of 3-month diabetic rats treated by SHH-N or cyclopamine. The protein level of GFAP was decreased by SHH-N treatment but elevated by cyclopamine treatment. (D) Whole-mount retinas were labeled for GFAP. Sonic hedgehog amino-terminal peptide treatment reduced the activation of retinal Müller cells in 3-month diabetic rats, but cyclopamine treatment aggravated the activation of Müller cells. Cross sections demonstrated that GFAP staining was mainly present in GCL, IPL, and INL. Glial fibrillary acidic protein expression was decreased by SHH-N treatment and increased by cyclopamine treatment. n = 4 per group. Scale bars: 50 μm. ^★★P < 0.01, ^★P < 0.05.

Effect of SHH-N on the Survival of RGC in Diabetic Rats

To determine whether the upregulated SHH expression influences RGC survival, Brn3a was used to label RGCs in the whole-mount retinas of diabetic rats. As shown in Figure 5, the PBS group lost significantly more RGCs after 3 months than did the SHH50 (P = 0.066) and SHH100 (P < 0.01) groups and the effect of SHH-N on RGC survival was dose-dependent (percent mean loss ± SD: PBS = 36.84% ± 7.01%; SHH50 = 23.79% ± 7.47%; SHH100 = 9.85% ± 5.56%). In contrast to SHH-N, cyclopamine exacerbated RGC loss after diabetes (52.81% ± 8.53%, P < 0.01). 

Figure 5

 

The effect of SHH-N and cyclopamine on RGC survival. Retinal whole-mount immunofluorescence staining of Brn3a. Brn3a+-labeled cells (green) show surviving RGCs in whole-mount retinas at 3 months after diabetes. The central (left) and peripheral (right) regions clearly show a marked reduction in RGC loss in SHH-treated diabetic rats. The graph shows the percentage of surviving RGCs in diabetic rats after SHH-N or cyclopamine treatment. This percentage was calculated with respect to the number of Brn3a+ RGCs counted in control retinas, which was considered 100%. n = 3 per group. Scale bars: 50 μm. ^★★P < 0.01.

Evaluation of High-Glucose Stress in the Primary-Cultured Retinal Müller Cells

The primary-cultured retinal Müller cells were identified by immunofluorescence staining of GS, which is a specific marker of Müller cells (Figure 6A). Glutamine synthetase–positive Müller cells show red fluorescence. Retinal Müller cells were treated with 5.5, 15, 25, 35, 45, and 55 mM glucose for 24, 48, and 72 hours to introduce the high-glucose stress, respectively. The best condition of high glucose stress induced to the cultured retinal Müller cells was by the exposure to 35 mM glucose for 48 hours (Fig. 6B). The exposure to 35 mM glucose for 48 hours was selected for the further experiments. We detected the SHH signaling pathway in primary Müller cell cultures under high-glucose stress. The protein levels of SHH, SMO, GLI1, and GFAP in the Müller cells were significantly increased 12 hours after cultured under 35 mM glucose stress (Fig. 6C). 

Figure 6

 

High glucose upregulated the expression of SHH, SMO, GLI1, and GFAP in primary-cultured retinal Müller cells. (A) Identification of retinal Müller cells by immunofluorescence staining of GS. (B) Dose-dependent effect of glucose on Müller cells viability at 24, 48, and 72 hours. Cell viability was measured by MTT assay. (C) Western blot shows the changed protein expression of SHH, SMO, GLI1, and GFAP in Müller cells under 35 mM glucose stress.

Effect of SHH-N and Cyclopamine Pretreatment in Primary-Cultured Retinal Müller Cells

Pretreatment with exogenous SHH-N (3 μg/mL) activated the SHH signaling pathway by increasing mRNA (Figs. 7A–C) and protein (Figs. 7E–H) expression of SHH, SMO, and GLI1. Cyclopamine made the opposite effect on the mRNA and protein expression of SHH, SMO, and GLI1, and the same in 5.5-mM glucose cultured primary Müller cells (Supplementary Figs. S1E–H). Glial fibrillary acidic protein mRNA and protein expression of was reduced by SHH-N pretreatment (P < 0.01) but was increased by cyclopamine (20 μg/mL) pretreatment (P < 0.05; Figs. 7D, 7E, 7I). 

Figure 7

 

Effects of SHH-N or cyclopamine on the expression of SHH, SMO, GLI1, and GFAP in primary-cultured retinal Müller cells. Real-time PCR analysis shows mRNA expression of SHH (A), SMO (B), GLI1 (C), and GFAP (D) in 35 mM glucose-cultured retinal Müller cells after SHH-N or cyclopamine treatment. (E) Western blot analysis of SHH, SMO, GLI1, and GFAP expression in 35 mM glucose-cultured retinal Müller cells after SHH-N or cyclopamine treatment. Quantitative analysis of SHH (F), SMO (G), GLI1 (H), and GFAP (I) in 35 mM glucose-cultured retinal Müller cells after SHH-N or cyclopamine treatment. The protein levels of SHH, SMO, or GLI1 in the high glucose-cultured Müller cells was elevated by SHH treatment but decreased by cyclopamine treatment. In contrast, the mRNA and protein expression of GFAP was reduced by SHH-N treatment but increased by cyclopamine. ^★★P < 0.01, compared with PBS-treated high-glucose group.

Effect of SHH-N and Cyclopamine on PI3K, ERK1/2, and P38 in Müller Cells Under High-Glucose Stress

Western blotting conducted 48 hours after cells were cultured with 35 mM glucose treatment, showed that SHH-N pretreatment inhibited p-ERK activation (Fig. 8). The effect of exogenous SHH-N was partly reversed by cyclopamine. In contrast, SHH-N pretreatment upregulated PI3K activation, and cyclopamine inhibited PI3K activation. Neither SHH-N nor cyclopamine treatment affected P38 activation. 

Figure 8

 

Sonic hedgehog amino-terminal peptide upregulated the activation of PI3K but inhibited the activation of ERK induced by high glucose. (A) Western blot images of the total protein and phosphorylated protein of ERK1/2, PI3K, and P38. (B–D) The quantitative data of Western blot from three independent experiments are shown. ^★★P < 0.01, ^★P < 0.05.

Discussion

Diabetic retinopathy is recognized as a neurodegenerative disease that can be affected by progressive degeneration of RGCs.³⁵ Thus, researchers are focusing on evaluating possible neuroprotective strategies that could slow progressive visual dysfunction during diabetes.³⁶ Because of their essential role in normal retinal functioning,¹³ strategies that focus on Müller cells and manipulation of retinal glia represent a promising avenue of therapeutic intervention, as glia cell participation has been reported in almost every retinal vascular disease.³⁷ 

The most sensitive nonspecific response to retinal diseases and injuries is the upregulation of GFAP,³⁸ which occurs early in the course of DR and precedes the onset of overt vascular changes. Though normal retinal Müller cells do not express GFAP, an increased GFAP expression results from oxidative stress or injury.^14,39 Glial fibrillary acidic protein's role in the development of DR, whether protective or harmful, is still debated.¹⁷ In our study, GFAP was upregulated in the retina of diabetic rats and in high glucose–cultured primary retinal Müller cells. The effect was inhibited by SHH but promoted by cyclopamine. These findings raise the possibilities that SHH might alter the function of retinal Müller cells under high-glucose stress and that glia-derived SHH might have a neuroprotective effect. 

Sonic hedgehog derived from RGCs plays a critical role in cell–cell and axoglial interactions and the development of normal glial cells.⁴⁰ Our results revealed that SHH was expressed at a low level in the retinas of normal adult rats and was upregulated after diabetes onset. Hyperglycemia upregulated the expression of SHH, SMO, and GLI1 in a time-dependent manner, but RGC loss was aggravated during the process of DR. Whether endogenous SHH is sufficient to protect the damaged RGCs of diabetic rats is not yet known. However, we found that exogenous SHH prevented RGC loss and cyclopamine, which inhibits SHH signaling, significantly increased RGC loss. This indicates that SHH might have a novel neuroprotective function in mature RGCs. Further studies will be necessary to understand how widely this key molecule works. 

Our study provides evidence that SHH exerts neuroprotective effects in an experimental model of DR; however, the molecular mechanisms by which SHH exerted its beneficial actions require explanation. Recent studies have reported an important role of MAPK pathways, including ERK1/2, p-p38, and PI3K, in diabetic complications.⁴¹ However, the roles of SHH and MAPK signaling pathways in neuroprotection, and the interactions between them, have not been clearly investigated until now. 

We speculated that the SHH pathway might crosstalk with MAPKs during diabetic retinal neurodegeneration. Our study found that exogenous SHH reduced the phosphorylation of ERK1/2, had no effect on p-p38, which indicated that SHH neuroprotection from high-glucose stress was independent of the p38 MAPK pathway, but was associated with inhibition of the ERK pathway. Our results also showed that PI3K was upregulated by exogenous SHH and inhibited by cyclopamine treatment. Therefore, the changed phosphorylation of PI3K and ERK1/2 in the retina of diabetic rats after SHH treatments might be related to cell death in the early stage of diabetes. Our findings suggest that Müller cell activation of SHH signaling has potential therapeutic value in the pathogenesis of retinal degeneration in DR. In previous studies, retinal neurons⁴⁰ and astrocyte⁴² were proved to be a source of SHH production. The deep relationship between them under injuries needs further investigation. 

Taken together, our results indicate that high-glucose stress can induce Müller cells to increase SHH production and that SHH may be involved in neuroprotective processes by regulating the PI3K and ERK pathways. However, further study is needed to elucidate the underlying mechanisms by which SHH modulates PI3K activation and ERK1/2. 

Acknowledgments

Supported by grants from the National Basic Research Development Program of China (973 program: 2013CB967000; Guangzhou, Guangdong, China) and the National Natural Science Foundation of China (LL [81170863], YL [81371020], and YHL [81371019]; Guangzhou, Guangdong, China). 

Disclosure: X. Zhao, None; Y. Li, None; S. Lin, None; Y. Cai, None; J. Zhang, None; X. Yu, None; H. Yang, None; L. Yang, None; X. Chen, None; Y. Luo, None; L. Lu, None