Early Response of Neurons and Glial Cells to Hypoxia in the Retina

Abstract

purpose. The present study was undertaken to examine the involvement of nitric oxide (NO) and excitotoxicity in the development of hypoxia-induced retinopathy in adult rats.

methods. Retinas of adult rats were examined at 3 hours to 14 days after hypoxia. The mRNA and protein expression of endothelial, neuronal, and inducible nitric oxide synthase (eNOS, nNOS, and iNOS, respectively), hypoxia-inducible factor-1α (HIF-1α), vascular endothelial growth factor (VEGF), N-methyl-d-aspartate receptor subunit 1 (NMDAR1), and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid glutamate (AMPA GluR2 and GluR3) receptors in the retina was determined by real-time RT-PCR, Western blot analysis, and immunohistochemistry. The response of retinal microglial cells to hypoxia was also studied by immunohistochemistry.

results. Hemorrhages were observed in the retina after hypoxia. Upregulated mRNA and protein expression of HIF-1α, NMDAR1, GluR2, GluR3, VEGF, eNOS, nNOS, and iNOS in the retina was observed in response to hypoxia. Complement type 3 (CR3) receptors and major histocompatibility complex (MHC) class I and II antigen expression on the microglial cells was increased after exposure to hypoxia.

conclusions. The findings of this study indicate that NO and excitotoxicity may produce damage to retina in response to hypoxia. Increased expressions of eNOS and VEGF in response to hypoxia are indicative of vasodilatation and increased permeability of retinal blood vessels. Increased phagocytosis by retinal microglial cells evidenced by increased expression of CR3 receptors may occur for the removal of hemorrhagic debris. Upregulation of MHC antigens indicates the readiness of these cells to participate in an immune response.

The neural retina has high metabolic demands and is very sensitive to hypoxia. ¹ Retinal hypoxia occurs acutely in central retinal artery occlusion and more chronically in ischemic central retinal vein thrombosis. Retinal ischemia from capillary closure or underperfusion is a well-recognized feature of diabetic retinopathy. As early as 1948, Michaelson ² postulated that retinal hypoxia caused the release of a vasoactive factor responsible for the neovascularization in such conditions as diabetic retinopathy. Patz ³ expanded on the role of ischemia and hypoxia in the genesis of diabetic proliferative retinopathy and the neovascular retinopathy of prematurity. Since then, there have been numerous reports implicating retinal hypoxia as playing an etiologic role in the sight-threatening complications of diabetic eye disease, including retinal and optic nerve head neovascularization. ^{4 5 6 7 8 9}  

Systemic causes of retinal hypoxia include the cardiovascular effects of chronic obstructive airways disease; the ocular ischemic syndrome associated with arterial obstructive conditions such as carotid artery stenosis, ¹⁰ Takayasu’s arteritis, ¹¹ and hyperviscosity syndromes ¹² ; or trauma ^{13 14} (Purtscher’s retinopathy). Hypoxia from retinal ischemia is characterized by retinal venous dilatation, retinal hemorrhages, retinal edema, capillary microaneurysm, cotton wool spots, and neovascularization. 

The pathophysiology of hypoxic damage to the retina is not fully understood. Nitric oxide (NO), synthesized from l-arginine by the family of nitric oxide synthase (NOS) enzymes ¹⁵ is known to mediate neuronal communication as well vasodilatation. NOS from neurons (nNOS) and endothelial cells (eNOS) are constitutively expressed enzymes the activities of which are stimulated by increases in intracellular calcium. ¹⁶ Inducible nitric oxide synthase (iNOS) is calcium-independent, and NO generated from this isoform is known to mediate immune functions. 

NO has also been implicated in the pathogenesis of brain and retinal injury from hypoxia-ischemia. ^{17 18} Because nNOS is known to be expressed under various stressful conditions or injuries to neurons in the central nervous system (CNS), ^{19 20} in the present study, we sought to examine its expression in the retina after the stress of hypobaric hypoxia, especially as the neural retina is an extension of the CNS. In addition to nNOS, we also examined the expression of eNOS and iNOS in the retina and with a nitric oxide colorimetric assay, determined the amount of NO produced. 

Activation of glutamate receptors particularly N-methyl-d-aspartate (NMDA) and amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) after ischemic or hypoxic insults has been described as a possible triggering mechanism for neuronal death ^{21 22 23 24} in the CNS and retina. ^{25 26} Earlier studies have shown that glutamate acting via NMDA receptors activates nNOS, ¹⁵ and the production of NO by nNOS is closely related to the activation of NMDA receptors. ²⁷ In view of these findings, we sought to examine whether, after hypobaric hypoxia exposure, there are any changes in the mRNA and protein expression of NMDA receptor subunit 1 (NMDAR1) and AMPA GluR2/3 receptors in the retina, as these may be involved in hypoxic damage to the retina. 

We also investigated the mRNA and protein expression of hypoxia-inducible factor (HIF)-1α and vascular endothelial growth factor (VEGF) in the retina after hypoxia, as studies have shown that hypoxia is a potent inducer of VEGF. ^{28 29} HIF-1α regulates transcription of hypoxia-responsive genes such as VEGF. ³⁰ In vitro and in vivo studies have shown that VEGF, induced as a result of low oxygen tension, is responsible for stimulating the proliferation of vascular endothelial cells that leads to new vessel formation. ³¹ VEGF is also known to increase permeability of blood vessels and is believed to be the prime regulator of hypoxia-induced angiogenesis. ^{28 29 31} The tissue concentration of VEGF in the retina was measured by an enzyme immunoabsorbent assay (EIA). 

Reactive changes in the microglial cells are a hallmark of injuries or other pathologic conditions in the central nervous system. Microglial cells are resident immunocompetent cells in the retina ³² and are extremely sensitive to their microenvironment. These cells, under normal conditions, express complement type 3 receptors (CR3) and major histocompatibility complex (MHC) class I antigens that are involved in endocytosis and antigen presentation respectively. It is well documented that expression of CR3 receptors and MHC I antigens are upregulated under pathologic conditions. ³³ Another sign of activation of the microglial cells is the expression MHC class II antigens, which are not expressed on these cells under normal conditions. It is not known whether microglial cells in the retina are activated in response to hypoxia. The present study, therefore, sought to determine whether these cells are responsive to hypoxic stress in the retina. 

Materials and Methods

Eighty-four albino rats weighing 200 g each were used in the present study. They were kept in cages in groups of six or seven rats per cage with free access to food and water. Sixty-eight rats were exposed to hypobaric hypoxia by placing them in a decompression (hypobaric) chamber (model 16M; Environmental Tectonics Corp., International, Southampton, PA) at an atmospheric pressure of 360 mm Hg for 2 hours, the Po ₂ being 73 mm Hg in the inspired air (Po ₂ 159 mm Hg at 760 mm Hg). The rats were then allowed to recover under normobaric conditions for 3 and 24 hours and for 3, 7, and 14 days before death. Another group of 16 rats of similar weight kept under similar laboratory conditions but not exposed to low atmospheric pressure were used as the control. The handling and care of animals complied with the guiding principles for research in the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research. 

Real-Time RT-PCR

Retinas were removed from the eyes of rats at 3 and 24 hours and at 3, 7, and 14 days (n = 3 at each time point) after exposure to hypoxia and from control rats (n = 3). They were stored at −80°C until RNA extraction. Total RNA was isolated (RNeasy mini kit; Qiagen, Valencia, CA) and the quality and quantity of RNA were determined by a biophotometer (ratios of A260:A280 were >1.8; Eppendorf, Fremont, CA). 

Reverse Transcription

Quantitative mRNA expression analysis of target genes was performed with a real-time RT-PCR system. When two-step real-time RT-PCR was planned, RT was performed by adding 2 μg of total RNA to 25 μL of reaction mixture (containing 1× first-strand buffer, 1 U/μL RNasin, 10 mM dNTP mix, and 200U M-MLV reverse transcriptase (all reagents were purchased from Invitrogen, Carlsbad, CA), which allows reverse transcription for 50 minutes at 42°C. After the reaction was complete, the mixture was heated to 95°C for 5 minutes to inactivate the M-MLV reverse transcriptase and chilled at 4°C. 

Quantitative RT-PCR

Quantitative RT-PCR was performed on a thermocycler (Light Cycler 3, using the FastStart DNA Master^plus SYBR Green I kit; Roche Diagnostics, Indianapolis, IN), according to the manufacturer’s instructions. Forward and reverse primer sequence for each gene and their corresponding amplicon size are provided in Table 1 . The rat β-actin, as an internal control, was also amplified for specific primers. The RT-PCR products were loaded onto 1.5% agarose gel stained with ethidium bromide and visualized under UV using an image analyzer (Chemi Genius²; Syngene, Cambridge, UK). Gene expression was quantified with a modification of the 2^−ΔΔCt method, as previously described. ³⁴  

Western Blot Analysis

Retinas were removed from the eyes of rats 3 and 24 hours and 3, 7, and 14 days (n = 3 at each time point) after exposure to hypoxia, and from the eyes of control rats (n = 3). Retinas were homogenized in 20 mM HEPES buffer (pH 7.2) containing sucrose, EDTA, dithiothreitol (DTT), and protease inhibitors. All procedures were performed at 4°C. Homogenates were centrifuged at 14,000g for 15 minutes, and the supernatant was collected. Protein concentrations were determined by the method of Bradford, ³⁵ using bovine serum albumin (BSA) as a standard. Samples of supernatants were mixed 1:1 with 2× sample buffer and heated to 95°C for 3 minutes. SDS-polyacrylamide gel electrophoresis was performed in 10% gel with 5% stacking gel and with 0.25 M Tris-glycine (pH 8.3), as the electrolyte buffer (Mini-Protein II apparatus; Bio-Rad, Hercules, CA). Protein bands were electroblotted for 1 hour onto 0.45 μm polyvinylidene difluoride (PVDF) membranes (Bio-Rad) by means of a semidry transfer apparatus (Bio-Rad) for incubation with antibodies. Nonspecific binding sites were blocked for 1 hour at room temperature with 5% nonfat dried milk powder and 0.05% Tween-20 in Tris-buffered saline (TBS; pH 7.6). The membranes were then separately incubated with dilutions of the polyclonal VEGF (1:1000), nNOS (1:500), NMDAR1 (1:1000), GluR2/3 (1:1000), and monoclonal HIF-1α (1:500), eNOS (1:2500), and iNOS (1:3000) antibodies in blocking solution overnight at 4°C and then were incubated with the appropriate secondary antibodies: horseradish peroxidase (HRP) conjugated anti-rabbit (1:5000) for VEGF, nNOS, NMDAR1 and GluR2/3, and HRP-conjugated anti-mouse (1:5000; Amersham Biosciences, Little Chalfont, UK) for HIF-1α, eNOS, and iNOS. Specific binding was revealed by enhanced chemiluminescence (ECL kit; Amersham Biosciences), according to the manufacturer’s instructions. For loading control, the membrane was incubated with monoclonal mouse anti-actin (1:3000; Sigma-Aldrich, St. Louis, MO) and revealed as just explained. Precision prestained standards (Bio-Rad) were used as molecular weight markers. X-ray films (Amersham Biosciences) were scanned with a computer-assisted densitometer (model G-710; Bio-Rad) to quantify band optical density (Quantity One software; Bio-Rad). 

Statistical Analysis

For Western blots and quantitative RT-PCR, data are reported as the mean ± SD. An independent Student’s t-test was used to determine the statistical significance of differences between control and hypoxic animals. 

NO Colorimetric Assay

The total amount of NO in the retina samples was assessed by Griess reaction with a colorimetric assay kit (U.S. Biologicals, Swampscott, MA) that detects nitrite (NO₂ ⁻), a stable reaction product of NO. Retinas from the eyes of rats at 3 and 24 hours and at 3, 7, and 14 days (n = 3 at each time point) after exposure to hypoxia and those of control rats (n = 3) were collected and homogenized in homogenizing buffer (T-PER; Pierce Biotechnology, Inc., Rockford, IL). Homogenates were centrifuged at 14,000g for 15 minutes, and the supernatant was collected. Briefly, 80 μL of the samples was added with 10 μL of enzyme cofactor followed by 10 μL of nitrate reductase, according to the manufacturer’s instructions, and incubated for 4 hours at room temperature. Griess reagents (100 μL) were mixed and added to the above solution, and the color was allowed to develop at room temperature for 10 minutes. The optical density of the samples was measured at 540 nm with a microplate reader (GENios; Tecan Austria GmbH, Grödig/Salzburg, Austria). The nitrite concentration (in micromolar) was determined from a nitrite standard curve. Results are expressed as the mean ± SD, Student’s t-test was used to determine significance, and P < 0.05 was considered significant. 

Analysis of VEGF by EIA

The amount of VEGF released in the rat retina control and hypoxic samples was determined using Chemikine VEGF EIA kit (Chemicon International Inc., Temecula, CA). Homogenates as described for the NO assay were prepared, and EIA measurements were performed according to the manufacturer’s protocol. Briefly, 50 μL of supernatant of retina samples was diluted with 50 μL of diluent, added to the precoated 96-well plate, and incubated with rabbit anti-VEGF antibody for 3 hours at room temperature. After this, 25 μL of VEGF conjugate was added to each well and further incubated for 30 minutes at room temperature. After thorough washing five times with washing buffer, 50 μL of diluted streptavidin-alkaline phosphatase was added to each well and incubated at room temperature for 30 minutes. Subsequently, equal volumes of color reagent A and color reagent B solution were added to each well and kept for 20 minutes at room temperature. The optical density was measured at 490 nm. The amount of VEGF (in nanograms per milliliter) detected in each sample was compared to a VEGF standard curve. Data are presented as the mean ± SD and significance established by Student’s t-test. 

A Pearson correlation analysis to assess the degree of correlation between NO production and tissue concentration of VEGF and was performed with a statistical package (SPSS ver. 13.0; SPSS Inc., Chicago, IL). 

Immunohistochemistry

Rats exposed to hypobaric hypoxia were subsequently killed at 3 and 24 hours and at 3, 7, and 14 days (n = 4 at each time point) after exposure, along with four control rats. They were anesthetized by intraperitoneal injection of 7% chloral hydrate and killed by transcardiac perfusion with a solution containing 2% paraformaldehyde in phosphate buffer (pH 7.4). The eyes were removed from the rats, immersed for 2 to 4 hours in the same fixative, and kept overnight in a solution of 15% sucrose in phosphate buffer. Frozen coronal sections of 40-μm thickness were cut with a cryotome (Frigocut; Leica, Wetzlar, Germany) and were processed by the avidin-biotin-peroxidase complex (ABC) technique to visualize VEGF, nNOS, eNOS, iNOS, GluR2/3, NMDAR1, OX-42, OX-18, and OX-6 immunoreactive sites. GluR2/3 antibody recognizes both GluR2 and GluR3 subunits of AMPA glutamate receptors. OX-42, -18, and -6 antibodies detect complement type 3 receptors (CR3 receptors), major histocompatibility class I and II (MHC I and -II) antigens, respectively, on the microglial cells. Sections were incubated for 30 minutes in phosphate-buffered saline (PBS; pH 7.4), containing 0.2% Triton X-100, and then separately with the antibodies at dilutions shown in Table 2 , in PBS/Triton X-100 overnight at room temperature. After several washes with PBS/Triton X-100, the sections were incubated with biotinylated goat anti-rabbit immunoglobulin (Vector Laboratories, Burlingame, CA) for VEGF, GluR2/3, NMDAR1, and nNOS and with biotinylated goat anti-mouse immunoglobulin (Vector Laboratories) for eNOS, iNOS, OX-42, OX-18, and OX-6 for 1 hour. After washing, the sections were incubated with peroxidase-linked ABC (Vector Laboratories) for 90 minutes. The peroxidase activity was demonstrated by nickel-enhanced 3,3-diaminobenzidine (DAB; Sigma-Aldrich). Sections were counterstained by 0.5% methyl green nuclear stain for 10 minutes, dehydrated by immersion in alcohol and then cleared with xylene before mounting in medium (Permount; Fisher Scientific, Pittsburgh, PA). Some sections were treated simultaneously without the primary antibodies to confirm the specificity of immunoreactivities. 

Double-Immunofluorescence Staining

Cellular localization of VEGF in the retina of control rats (n = 3) and in rats at 24 hours after hypoxia (n = 3) was examined by the double-immunofluorescence staining method, according to the following procedures. Briefly, the rats were killed as described earlier for immunohistochemistry, and 40-μm-thick frozen coronal sections of the eyes were cut and incubated with 0.3% H₂O₂ in methanol for 20 minutes to block nonspecific binding sites. After several washes with PBS, the sections were incubated in a humid chamber with a mixture of two primary antibodies (polyclonal anti-VEGF and monoclonal anti-GFAP or polyclonal anti-glutamine synthetase and monoclonal anti-GFAP [glial fibrillary acidic protein]) diluted with PBS. GFAP was used, as it is a specific marker for astrocytes and our earlier studies have shown that GFAP-positive astrocytes are colocalized with VEGF-positive astrocytes in the CNS (Kaur C, Sivakumar V, unpublished data, 2004). Glutamine synthetase (GS) is a specific marker for Müller cells. After they were washed in PBS, the sections were incubated with a mix of two fluorescent secondary antibodies, Cy3-conjugated goat anti-rabbit IgG, and FITC-conjugated sheep anti-mouse IgG for 40 to 60 minutes. After several washes with PBS, the sections were mounted with fluorescent mounting medium (DakoCytomation, Carpinteria, CA). Colocalization was observed by confocal microscopy (FV 500; Carl Zeiss Meditec, Inc. Dublin CA). Some sections were treated simultaneously without the primary antibodies, to confirm the specificity of the immunoreactivity. 

Results

Analysis of HIF-1α, VEGF, NMDAR1, GluR2, GluR3, nNOS, eNOS, and iNOS mRNA by Real-Time RT-PCR

The specificity of RT-PCRs was verified by checking that the PCR products were of the expected size by gel electrophoresis. The primer pair for each gene resulted in a single product with the desired length: HIF-1α (198 bp), VEGF (177 bp), GluR2 (539 bp), GluR3 (180 bp), NMDAR1 (333 bp), eNOS (243 bp), iNOS (179 bp), and nNOS (617 bp). 

Using quantitative RT-PCR, we were able to detect the expression of HIF-1α, VEGF, GluR2, GluR3, NMDAR1, eNOS, iNOS, and nNOS mRNA in the adult rat retina (Fig. 1) . The HIF-1α mRNA was significantly elevated up to 3 days (P < 0.05, Fig. 1A ). VEGF mRNA expression was significant (P < 0.05) in comparison with the control at all time intervals after hypoxia (Fig. 1A) . GluR2 mRNA was significantly (P < 0.05) higher at 3 and 24 hours and 3 and 7 days, compared with the control levels, and declined thereafter (Fig. 1A) . GluR3 mRNA increased up to 24 hours after hypoxia. NMDAR1 mRNA was elevated up to 3 days (Fig. 1B) . Hypoxia increased (P < 0.05) the iNOS and nNOS mRNA levels up to 7 days, the expression levels equaling the control levels at 14 days. The eNOS mRNA remained significantly elevated up to 3 days (P < 0.05, Fig. 1B ). 

Western Blot Analysis

The results of Western blot analysis of extracts of adult rat retina are shown in Figure 2 . Densitometry of the HIF-1α immunoreactive band of approximately 120 kDa increased significantly from 3 hours to 3 days after hypoxia, compared with control levels (Fig. 2A) . VEGF was detected as a major band of approximately 25 kDa, which showed a statistically significant increase from 3 hours to 14 days over control levels after hypoxia (Fig. 2A) . The GluR2/3 immunoreactive band of approximately 100 kDa in the supernatant adult rat retina extracts from the hypoxia group showed a significant increase from 3 hours to 3 days compared with the control. This was followed by a decline to a low level at 7 and 14 days (Fig. 2A) . NMDAR1 and eNOS immunoreactive bands, with molecular weights of approximately 110 (Fig. 2A)and 140 (Fig. 2B)kDa, increased significantly from 3 hours to 7 days after hypoxia compared with control levels. At 7 days after hypoxia, the NMDAR1 and eNOS protein quantity was comparable to control levels. The nNOS immunoreactive band, approximately 155 kDa, showed a significant difference between control animals and those at 3 hours to 14 days after hypoxia (Fig. 2B) . The iNOS immunoreactive band, approximately 130 kDa in the control, showed significantly elevated levels at 3 and 24 hours and 3 days after hypoxia (Fig. 2B) . 

NO Colorimetric Assay

The NO levels in the retina samples were significantly (P < 0.05) increased at 3 and 24 hours and 3 and 7 days after exposure to hypoxia when compared with the control (Fig. 3A). 

Enzyme Immunoassay

Analysis by EIA revealed that VEGF concentration in the retina increased significantly (P < 0.05) from 3 hours to 14 days after hypoxia, when compared with the control (Fig. 3B) . 

Pearson correlation analysis to determine the degree of correlation between NO production and VEGF concentration at various time points after hypoxia showed the values to be between −1 and +1 (r = 0.99, control; r = −0.80, 3 hours; r = −0.45, 24 hours; r = −0.72, 3 days; r = 0.97, 7 days; r = −0.45, 14 days) confirming that a true association exists between NO production and concentration of VEGF in the retina. 

Immunohistochemical Analysis

Hemorrhages were observed in retinas examined for immunohistochemical staining at 3 and 24 hours after hypoxia. Table 3summarizes the immunohistochemical labeling of cells in the retina after hypoxia compared with the control. 

nNOS.

In the control rats, neurons in the ganglion cell layer (GCL) expressed weak nNOS immunoreactivity (Fig. 4A). Expression of nNOS was absent in the other layers of the retina. After hypoxia, nNOS expression was intense in the neurons of the GCL at 3 hours to 7 days (Fig. 4B) , whereas it declined to control levels at 14 days. Occasional cells in the inner nuclear layer expressed nNOS immunoreactivity from 3 hours to 3 days, which was absent in this layer at longer time intervals. Nerve fibers in the inner plexiform layer (IPL) also expressed nNOS immunoreactivity at 3 hours to 7 days. No reaction was observed in sections incubated without the primary antibody. 

iNOS.

Expression of iNOS was absent in the GCL and other layers of the retina in control animals (Fig. 4C) , but occasional amacrine cells expressed iNOS immunoreactivity in control animals. After hypoxia, intense iNOS expression was observed in the amacrine cells at 3 hours (Fig. 4D) , whereas the ganglion cells showed an intense expression at 24 hours to 3 days (Figs. 4E 4F , respectively). Expression of iNOS was also observed in the inner plexiform, outer plexiform, photoreceptor, and inner nuclear layers (Figs. 4D 4F) . The expression declined at 7 days and was comparable to control levels at 14 days. 

eNOS.

Weak expression of eNOS was observed in blood vessels in the nerve fiber layer and the inner and outer plexiform layers of the retina in the control rats (Fig. 5A). Strong eNOS immunoexpression was induced in the blood vessels at 3 and 24 hours after hypoxia (Fig. 5B) . Along with the blood vessels, neurons in the GCL also expressed eNOS immunoreactivity which became stronger at 3 and 7 days (Figs. 5C 5D , respectively). At 14 days after hypoxia, the expression of eNOS was comparable to control levels. 

NMDAR1 and GluR2/3.

Weak expression of NMDAR1 and GluR2/3 was detected in the ganglion cells in control rats (Figs. 6A, 6C). The NMDAR1 and GluR2/3 was markedly enhanced inthe ganglion cells at 3 and 24 hours (Figs. 6B 6D) , and remained elevated up to 3 days after hypoxia. Besides the ganglion cells, GluR2/3 and NMDAR1 expression was also upregulated in the IPL at these time intervals. At 7 days, the NMDAR1 and GluR2/3 expression declined and was comparable to control levels. 

VEGF.

VEGF expression was observed in branched cells in the nerve fiber layer of the retina in control rats (Fig. 7A). Many of the VEGF-positive cells were located in the vicinity of the blood vessels. The cells were identified by double immunofluorescence to be astrocytes in the control rats as they were completely colocalized with GFAP-positive cells (Fig. 8) . However, after hypoxia, GFAP was also expressed by GS-positive Müller cells (Fig. 9) . Expression of VEGF in the astrocytes and Müller cells was upregulated at all time intervals from 3 hours to 14 days (Figs. 7B 7C 7D)after hypoxia. Astrocyte processes around the blood vessels expressed very strong VEGF immunoreactivity. At 24 hours and 3 and 7 days many of the VEGF-positive processes (Figs. 7C 7D)extended into the other layers (i.e., the inner plexiform, inner nuclear, outer plexiform, and outer nuclear layers). Immunofluorescence labeling with GS confirmed that these were Müller cell processes. At 14 days, the VEGF immunoreactivity was still stronger than that in the control but was less than at earlier time intervals. 

CR3 Receptors and MHC I and II Antigens.

In the control rats, ramified microglial cells in the nerve fiber layer and IPL of the retina expressed OX-42 (CR3 receptors) and OX-18 (MHC I antigens) immunoreactivity (Figs. 10A and 11 A), whereas OX-6 (MHC II antigens) immunoreactivity was not observed in these cells (Fig. 11E) . At 3 hours after hypoxia, the OX-42 and OX-18 immunoreactivity was downregulated in these cells (Figs. 10B 11B) . OX-42 and OX-18 expression was upregulated at 24 hours and 3 and 7 days (Figs. 10C 10D 11C 11D)and was comparable to control levels at 14 days after hypoxia. Often hypertrophic OX-18-positive cells were observed in the vicinity of the blood vessels at 3 days (Fig. 11C) . Occasional microglial cells expressed OX-6 immunoreactivity at 24 hours after hypoxia (Fig. 11F) . At 3 and 7 days, a large number of microglial cells were induced to express OX-6 immunoreactivity (Figs. 11G 11H , respectively) in the nerve fiber and IPL of the retina. Many of the cells expressing MHC II antigens were in the vicinity of blood vessels at 3 days, and these cells appeared hypertrophic, with short processes. OX-6 expression was not observed at 14 days. 

Discussion

The retina is known to be extremely sensitive to fluctuations in oxygen levels and hypoxia is known to cause the development of retinopathy. ³⁶ The present study has shown that after a hypoxic exposure the mRNA and protein expression of HIF-1α, NMDAR1, GluR2, GluR3, VEGF, nNOS, eNOS, and iNOS were upregulated in the retina. 

Excessive activation of NMDA and AMPA receptors has been described as a mechanism underlying cell death after ischemic conditions in the CNS ^{22 37 38} and retina. ²⁶ The increased expression of NMDAR1 and GluR2/3 in the retina in our study may be linked to an increase in the extracellular levels of glutamate. It has been reported that extracellular levels of glutamate are increased to toxic levels in the CNS in ischemic conditions ³⁹ and neurotoxicity associated with excitatory amino acids is mediated by the activation of NMDA receptors. ⁴⁰ An increase in the NMDAR1 expression may also be responsible for an increased influx of Ca²⁺ in the neurons of the GCL, as NMDA receptor channels are known to provide an important route for calcium to enter the neuron and activate intracellular calcium-dependent enzymes. The increased expression of NMDAR1 after hypoxic insult in the retina may lead to neuronal damage. A previous study ²⁵ has shown that NMDA receptor–mediated toxicity in the retinal ganglion cells is dependent on the influx of extracellular Ca²⁺. Increased expression of GluR3 has also been reported as facilitating increased Ca²⁺ loading during excitotoxicity. ⁴¹ However, GluR2 mRNA and protein upregulation may be a protective response to diminish glutamate-induced excitotoxic damage to retinal neurons. GluR2 is also known to impart low Ca²⁺ permeability to ganglion cells. ⁴² A downregulation of GluR2 renders the neurons permeable to Ca²⁺, whereas its upregulation protects against excessive Ca²⁺ influx. The GluR2/3 protein in the retina was upregulated up to 3 days, whereas GluR2 and GluR3 mRNA remained elevated up to 7 days after hypoxia. 

Increased production of NO up to 7 days after hypoxia was observed in the present study. Activation of NMDA receptors in ischemic conditions leads to increased activity of nNOS and, hence, production of NO. ⁴³ The enhanced mRNA and protein levels of nNOS in the present investigation indicate excess production of NO through this isoform. Excess production of NO through nNOS is believed to mediate neuronal injury elicited by glutamate acting at NMDA receptors. ^{44 45} Increased production of nNOS mRNA has been detected in ischemic lesions as early as 15 minutes after occlusion of the middle cerebral artery ⁴⁶ and in hypoxic conditions. ⁴⁷  

Expression of iNOS has been shown to occur in neurons after oxygen and glucose deprivation, ⁴⁸ and the NO generated from this isoform plays an important pathogenic role in the tissue damage that occurs after cerebral ischemia. The mRNA and protein expression of iNOS increased at 3 hours to 7 days in response to hypoxia in the present study. Intermittent hypoxia has been shown to cause time-dependent induction and increased expression of iNOS which leads to excessive NO production and functional losses in the cerebral cortex. ⁴⁹  

Vasodilation occurs in hypoxic–ischemic conditions, the excess production of NO through the eNOS isoform being responsible for this effect. It has also been suggested that NO contributes to increased retinal blood flow in hypoxic conditions. ⁵⁰ In the present study, increased expression of eNOS in the blood vessels in the nerve fiber layer and the inner and outer plexiform layers of the retina was observed after hypoxia. This correlated with the mRNA and protein expression of eNOS in the retina. Besides the blood vessels, ganglion cells also expressed eNOS immunoreactivity. Neurons expressing eNOS immunopositivity have been localized in the visual cortex. ⁵¹ Expression of eNOS has also been reported on neurons in the nodose ganglion in response to hypoxia, ⁵² thus adding to excess NO production. In ischemic conditions, eNOS expression has been reported to have a neuroprotective role. ⁵³ In addition to vasodilation, the expression of eNOS in the retina may have a similar neuroprotective function in hypoxic conditions. 

NO produced through the eNOS isoform is also known to activate VEGF gene transcription ⁵⁴ and eNOS has been described as playing a predominant role in VEGF-induced angiogenesis and vascular permeability. A reduced angiogenic response was observed in mice deficient in the eNOS gene. ⁵⁵ The elevated mRNA and protein expression of eNOS in the retina after hypoxia may be an early step in the angiogenesis process. NO is known to influence neovascularization, and there is increasing evidence of eNOS being a proangiogenic factor. ⁵⁶  

Hypoxia induces the expression of VEGF via HIF-1 α. ³⁰ Increased gene and protein expression of HIF-1α was observed up to 3 days, and expression of VEGF mRNA and protein levels was observed up to 14 days after hypoxia in the present study. It has been postulated that exposure to systemic hypoxia leads to activation of VEGF gene transcription in the CNS leading to increased VEGF protein levels and hence increased vascular permeability. ⁵⁷ Increased expression of VEGF has been shown to occur in the retina in hypoxic ^{36 30} and ischemic conditions. ⁵⁸ VEGF is a key molecular regulator of angiogenesis. ⁵⁹ Both hypoxia and NO have been described to contribute to activation of VEGF gene transcription in the brain after subarachnoid hemorrhage. ⁶⁰ A strong correlation was observed between NO production and VEGF concentration in the retina after hypoxia in the present study. Leakage of serum-derived substances can thus occur in the retina through increased permeability of the blood vessels, leading to the development of retinal exudates and edema. In addition to its role in angiogenesis, VEGF has been reported to be neuroprotective in CNS. ⁵⁹ It has also been postulated to have a role in lesion-induced neurogenesis which, coupled with angiogenesis, could promote repair of neural tissue. ⁶¹ A role for VEGF as a linking element in neuronal growth and angiogenesis as well as an antiapoptotic agent has also been proposed. ⁵⁹ After hypoxia, the upregulated VEGF production as observed in the present study, may have a similar neuroprotective role besides initiating increased vascular permeability and angiogenesis. 

VEGF appears to be released by the Müller cells and astrocytes under hypoxic conditions. Double immunofluorescence labeling with GFAP, a marker for astrocytes, and VEGF showed complete colocalization in the retina of control rats, indicating that under normal conditions, VEGF is produced by astrocytes. GS-positive Müller cell processes were not labeled with GFAP in control retinas but showed colocalization with GFAP after hypoxia. Müller cells do not express GFAP under normal physiological conditions but are known to express GFAP under pathologic conditions. ^{62 63}  

The expression of CR3 receptors and MHC I antigens on the microglial cells is related to their phagocytic activity and antigen-presenting function. The expression of CR3 receptors and MHC I on the microglial cells was downregulated at 3 hours after hypoxia but was upregulated at 24 hours and 3, 7, and 14 days when hypertrophied microglial cells were observed in the IPL and in the nerve fiber layer. The initial downregulation of CR3 receptors and MHC I indicates a suppression of phagocytic activity as well as antigen-presenting capacity of the microglial cells. Hypoxia has been recognized as a specific stimulus that can modulate macrophage activity, resulting in decreased phagocytosis. ⁶⁴ Our earlier study showed that hormonal changes in response to the stress of hypoxia can also result in decreased CR3 receptor expression on macrophages/microglia in the adenohypophysis. ⁶⁵ The subsequent upregulation of CR3 receptors on the microglia, a feature of increased phagocytic activity, indicates the readiness of these cells to phagocytose the debris of any degenerating elements in the retina. Microglial activation and response in experimentally induced neurodegeneration is a well-documented phenomenon in the CNS. ^{33 66} The upregulation of MHC I antigens and de novo expression of MHC II antigens indicates that these cells are ready to participate in a potential immune response. 

Conclusions

We conclude from the findings of this study that hypoxia leads to upregulation of NMDAR1 and GluR2/3 receptors. Although NMDAR1 activation is known to produce excitotoxic damage through excessive influx of Ca²⁺ and nNOS activation, upregulation of GluR2 may be a neuroprotective response. Production of NO through nNOS and iNOS isoforms may produce neuronal damage. Upregulated eNOS and VEGF expression is involved in vasodilatation and increased permeability of blood vessels in the retina. Activation of microglial cells, as shown by an upregulated CR3 receptor and MHC I antigen expression as well as de novo expression of MHC II antigens, may serve to clear debris of retinal hemorrhages and any degenerating elements and to mount a possible immune response. We postulate that a balance between damaging and protective factors exists in the retina at early time intervals after hypoxia. This balance may be upset at longer time intervals, leading to hypoxic damage to the retina and neovascularization. 

 Supported by Research Grant R181-000-065-112 from the National University of Singapore.

 Submitted for publication April 27, 2005; revised August 24 and October 28, 2005; accepted January 20, 2006.

 Disclosure: C. Kaur, None; V. Sivakumar, None; W.S. Foulds, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Sequence of Specific Primers

Primer	Forward	Reverse	Amplicon Size (bp)
HIF-1α	tcaagtcagcaacgtggaag	tatcgaggctgtgtcgactg	198
VEGF	agaaagcccaatgaagtggtg	actccagggcttcatcattg	177
NMDAR1	aacctgcagaaccgcaag	gcttgatgagcaggtctatgc	333
GluR2	ctatttccaaggggcgctgat	cagtccaggattacacgccg	539
GluR3	cgcagagccatctgtgttta	gttgccacaccatagccttt	180
eNOS	tggcagccctaagacctatg	agtccgaaaatgtcctcgtg	243
iNOS	ccttgttcagctacgccttc	ggtatgcccgagttctttca	179
nNOS	aacctgcagaaccgcaag	gcttgatgagcaggtctatgc	617
β-Actin	tcatgaagtgtgacgttgacatccgt	cctagaagcatttgcggtgcaggatg	285

Table 2.

 

Antibodies Used for Immunohistochemistry and Immunofluorescence

Antibody	Host	Source	Dilution
VEGF	Rabbit polyclonal	Santa Cruz Biotechnology, Santa Cruz, CA	1:200
GluR2/3	Rabbit polyclonal	Chemicon International Inc., Temecula, CA	1:200
NMDAR1	Rabbit polyclonal	Chemicon International Inc., Temecula, CA	1:200
eNOS	Mouse monoclonal	BD Biosciences, San Diego, CA	1:250
iNOS	Mouse monoclonal	BD Biosciences, San Diego, CA	1:1000
nNOS	Rabbit polyclonal	BD Transduction Labs, San Jose, CA	1:500
GFAP	Mouse monoclonal	Chemicon International Inc., Temecula, CA	1:1000
OX-42	Rat monoclonal	Harlan Sera-Lab, Ltd. Loughborough, UK	1:100
OX-18	Rat monoclonal	Harlan Sera-Lab, Ltd. Loughborough, UK	1:100
OX-6	Rat monoclonal	Harlan Sera-Lab, Ltd. Loughborough, UK	1:100
Glutamine synthetase	Rabbit polyclonal	Sigma-Aldrich, Ann Arbor, MI	1:600

Figure 1.

 

(A) RT-PCR analysis of HIF-1α, VEGF, GluR2, and GluR3 gene expression in the retina of adult rats after hypoxia. Left: ethidium bromide stained 1.5% agarose gel with RT-PCR products of the above- mentioned mRNA in the retinas of control rat (lane 2) and of rats at 3 (lane 3) hours and 24 (lane 4) hours and 3 (lane 5), 7 (lane 6), and 14 (lane 7) days after hypoxia. Lanes 1 and 8: 100-bp DNA ladder. Right: changes (x-fold) quantified by normalization to β-actin as an internal control. Bars, mean ± SD. The mRNA levels were significantly (*P < 0.05) altered in retinas of hypoxic rats when compared with those of control subjects. (B) RT-PCR analysis of NMDAR1, eNOS, iNOS, and nNOS gene expression in the retina of adult rats after hypoxia. Left: ethidium bromide stained 1.5% agarose gel with RT-PCR products of mRNA in the retinas of control and hypoxic rats. Lanes and graphs are as described in (A). *mRNA level significantly different from the control (P < 0.05).

Figure 2.

 

(A) Western blot analysis of HIF-1α, VEGF, GluR2/3, and NMDAR1 protein expression in the retina supernatants from control rats (lane 1) and rats subjected to hypobaric hypoxia and killed at 3 (lane 2) and 24 (lane 3) hours and 3 (lane 4), 7 (lane 5), and 14 (lane 6) days after the exposure. Top: HIF-1α (120 kDa), VEGF (25 kDa), GluR2/3 (100 kDa), and NMDAR1 (110 kDa) immunoreactive bands. Bar graphs representing HIF-1α, VEGF, GluR2/3, and NMDAR1 show significant changes in the optical density after hypoxia (mean ± SD). *P < 0.05; **P < 0.01 compared with the control. (B) Western blot analysis of eNOS, iNOS, and nNOS protein expression in the retina supernatants from control rats and rats submitted to hypobaric hypoxia. Lanes are as in (A). Top: eNOS (140 kDa), iNOS (130 kDa), and nNOS (155 kDa) immunoreactive bands. Bar graphs representing eNOS, iNOS, and nNOS show significant changes in the optical density after hypoxia (given as mean ± SD). *P < 0.05; **P < 0.01 compared with the control.

Figure 3.

 

Determination of NO by colorimetric assay (A) and VEGF by enzyme immunoassay (B) in the retina of control rats (C) and at 3 and 24 hours and 3, 7, and 14 days after exposure to hypoxia. Data represent the mean ± SD of the amount of NO and VEGF (n = 3). *Significant increases in NO and VEGF production between control and hypoxic groups (P < 0.05).

Table 3.

 

Summary of Immunohistochemical Reactions in the Retina

Antibody	Control	Time Intervals following Hypoxic Exposure
				3h	24h	3d	7d	14d
NMDAR1	+−	++	++	++	+	+−
GluR2/3	+−	++	++	++	+	+−
eNOS	+	++	+++	+++	++	+
iNOS	−	++	+++	++	+	−
nNOS	+−	++	+++	+++	++	+−
VEGF	+	++	+++	+++	++	+
OX-42	++	+−	++	+++	+++	++
OX-18	+	+−	+	++	++	+
OX-6	−	−	+−	++	+	−

 NMDAR1, GluR2/3, iNOS and nNOS immunoreaction was observed in the ganglion cells. Expression of eNOS was seen in the blood vessels in the nerve fiber layer, inner and outer plexiform layers, and on the ganglion cells. VEGF expression was observed in the astrocytes and Müller cells. OX-42, OX-18, and OX-6 immunoreaction was observed in the microglial cells located in the nerve fiber layer and inner plexiform layer. Very strong reaction, +++; strong reaction, ++; moderate reaction, +; weak reaction, +−; no reaction, −.

Figure 4.

 

(A) Coronal section of the retina in a control rat showing extremely weak expression of nNOS in the neurons (arrows) of the ganglion cell layer (GCL). (B) Expression of nNOS was upregulated in the neurons (arrows) at 24 hours after hypoxia. (C) Retina of a control rat showing absence of iNOS in the neurons in GCL. (D) Retina of a rat at 3 hours after hypoxia showing expression of iNOS in neurons (arrowheads) in the GCL, IPL, and amacrine cells (arrows). Expression of iNOS was also induced in the photoreceptor layer (*). Expression of iNOS was upregulated in the neurons in the GCL (arrowheads) at 24 hours (E) and 3 days (F) after hypoxia. Many cells (arrows) in the inner nuclear layer also expressed iNOS (F). Scale bar, 10 μm.

Figure 5.

 

(A) Retina of a control rat showing eNOS expression in blood vessels (arrows) in the nerve fiber layer and outer plexiform layer (*). (B) eNOS was expressed in many blood vessels (arrows) in the nerve fiber layer at 24 hours after hypoxia. Weak expression was present in the neurons of the GCL at this time interval. Blood vessels in other layers of the retina (arrowheads) also showed eNOS expression. At 3 (C) and 7 (D) days after hypoxia, intense eNOS expression was present in the neurons (arrows) of the GCL. Blood vessels expressed intense eNOS immunoreactivity (arrowheads). Scale bar: (A, C, D) 10 μm; (B) 50 μm.

Figure 6.

 

Weak expression of NMDAR1 in the neurons (arrows) in the GCL of a control retina (A) compared with strong NMDAR1 expression (arrows) at 3 hours after hypoxia (B). Expression of GluR2/3 in the neurons (arrows) in the GCL in the control rat (C) and at 3 hours after hypoxia (D) showed an intense difference. Scale bar, 10 μm.

Figure 7.

 

VEGF expression in branched cells (arrows) in the nerve fiber layer in a control rat (A) and in rats 24 hours (B), 3 days (C), and 7 days (D) after hypoxia showing VEGF expression to be much stronger after hypoxia. (B) Strong VEGF expression was present in cells surrounding a blood vessel (arrow). (C, D) Long processes (arrows) of VEGF-positive cells extend throughout the thickness of the retina. Scale bar, 10 μm.

Figure 8.

 

Confocal images showing the distribution of GFAP (A, green) and VEGF (B, red) immunoreactive cells in the retina of a control rat. (C) Colocalized expression of VEGF and GFAP. GFAP (D, green), VEGF (E, red) and colocalized expression of GFAP and VEGF (F) at 24 hours after hypoxia. Scale bar, 50 μm.

Figure 9.

 

Confocal images showing the distribution of (A, green) GFAP- and (B, red) GS-immunoreactive cells in the retina of a control rat. (C) Image of colocalized expression of GS and GFAP shows that GFAP was not completely colocalized with GS in the processes in the control. GFAP (D, green), GS (E, red), and colocalized expression of GFAP and GS (F) at 24 hours after hypoxia. GS and GFAP were completely colocalized (F). Scale bar, 50 μm.

Figure 10.

 

Sections of retina showing expression of CR3 receptors, detected by the OX-42 antibody, in ramified microglia (arrows) in the nerve fiber layer, GCL, and IPL of retina in a control rat (A). CR3 expression was downregulated in rats at 3 hours after hypoxia in the microglial cells ((B, arrows) and was upregulated at 3 (C) and 7 (D) days. (C, D) Hypertrophied microglial cells expressed intense OX-42 immunoreactivity (arrows). Scale bar, 10 μm.

Figure 11.

 

Expression of MHC I antigens, detected by OX-18 antibody, on the microglial cells (arrows) in the retina of a control rat (A) and in rats at 3 hours (B) and 3 (C) and 7 (D) days after hypoxia. MHC I expression was downregulated at 3 hours but was strong at 3 and 7 days. MHC I-expressing cells were present in the vicinity of blood vessels (C, *) in the IPL at 3 days. The cells appeared hypertrophic at 3 (C) and 7 (D) days. Expression of MHC II antigens, detected by OX-6 antibody, was absent in the retina of a control rat (E) but was induced in a few branched cells (arrows) at 24 hours (F) and in many cells (arrows) around a blood vessel (BV) at 3 (G) and 7 (H) days after hypoxia. Scale bar, 10 μm.

The authors thank Pee Chai Lim, Cheng Hai Yeo, and Eng Siang Yong for providing technical assistance and Eng Ang Ling for providing the glutamine synthetase antibody.