Introduction
Retinal ischemia, often referred as “stroke of retina”, is an important cause of visual impairment in retinal vascular occlusion, diabetic retinopathy, glaucoma, and ocular trauma [
1‐
5]. It is caused by a reduction of the retinal blood supply that decreases the delivery of oxygen and other nutrients to various retinal layers. Reperfusion of blood following ischemia is associated with oxidative stress and inflammatory responses [
6]. In particular, resulting retinal ganglion cell (RGC) death is caused by a variety of cell death mechanisms including necrosis, apoptosis, necroptosis and autophagy after ischemia/reperfusion (I/R) injury [
7‐
9]. To better understand the pathophysiological mechanisms associated with retinal I/R injury, several different experimental approaches have been designed in rodent models [
8,
10‐
12].
The “pressure-induced retinal I/R model” involves cannulation of the ocular anterior chamber followed by raising intraocular pressure above systolic blood pressure. After a specified period of time, the cannula is removed allowing restoration of retinal blood flow. This model has been used for the investigation of ischemia-derived ocular pathologies such as glaucoma and diabetic retinopathy [
13,
14]. This model currently is the most widely used method to study ocular diseases related to retinal ischemia.
Retinal damage due to I/R injury is associated with the loss of neurons, morphological degeneration of the retina, loss of retinal function, and ultimately vision loss [
15‐
17]. Although degeneration times vary in different experimental conditions, I/R-induced injury and retinal degeneration is initially observed primarily in inner retinal layers [e.g. the inner plexiform layer (IPL) and inner nuclear layer (INL)] that are supplied by the central retinal artery, in contrast to the outer nuclear layer (ONL) that is generally less affected [
18‐
21]. Differently from inner retinal layers, the choroid supplies blood and nutrients to the photoreceptors and ONL [
22,
23]. This structural difference may influence the initiation stage of I/R injury. Furthermore, morphological changes are often associated with functional impairment of the retina [
24‐
26]. However, the correlation between morphological and functional changes with molecular mechanisms from different stages of pathogenesis associated with retinal I/R injury in retina is poorly characterized.
Several signaling pathways have been reported as key molecular events related to the degeneration or protection of RGCs and their axons in retinal I/R injury. Increased nuclear factor (NF)-κB p65 immunoreactivity was associated with retinal degeneration following retinal ischemia and reperfusion injury in mice [
27], and inactivation of astroglial NF-κB promotes survival of retinal neurons following ischemic injury [
28]. In addition to NF-κB, cyclooxygenease-2 appears to play a critical role in RGC death after transient ischemia [
29]. Activation of Stat3 protects retinal ganglion cell layer neurons in response to transient retina ischemia in mice [
30]. Several reports have indicated that inflammatory responses associated with the innate and adaptive immune system are major pathological processes in retinal I/R injury. These include toll-like receptor 4, complement component C3, tumor necrosis factor receptor, and surface molecule CD40 that are associated with functional impairment, retinal layer morphological changes, and/or loss of RGCs [
20,
31‐
33]. However, further studies are still required to define and better understand pathologic progression including temporal changes in retinal morphology, function, and molecular signaling.
In our study, we used a mouse model of pressure-induced retinal I/R injury and characterized morphological changes in retinal layers associated with retinal ERG functions. Using histological analysis and SD-OCT (spectral domain-optical coherence tomography) scanning, we showed progressive morphological changes of retinal layers at various times after I/R injury. In addition, we used scotopic electroretinography (ERG) to demonstrate functional deficits after I/R injury that corresponded to the morphological changes. We also profiled changes in retinal gene expression associated with I/R injury at various time points using cDNA microarray analysis to correlate the molecular mechanisms with morphological and functional changes.
Material and methods
Mice
Female C57BL6/J mice (8-10 weeks of age) were maintained in 12:12 light /dark cycle under optimal temperature and humidity controlled conditions. All studies were approved by University of North Texas Health Science Center’s Institutional Animal Care and Use Committee (IACUC) and complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.
Pressure induced retinal ischemia
Mice were anesthetized using a ketamine/xylazine cocktail (100/10 mg/kg). While anesthetized, mice were placed on a heating pad to prevent hypothermia. Mydfrin (Alcon, Inc. Fort Worth, TX) was topically administered to the test eye to dilate the iris. A 30-gauge needle connected to saline reservoir was inserted into the anterior chamber through the cornea of left eyes. Intraocular pressure was raised to 120 mmHg for 60 min. Retinal ischemia was confirmed by blanching of the retina using an ophthalmoscope. Contralateral right eyes served as controls. The needle was removed after 60 minutes to allow reperfusion of the retina. Tobrex (Alcon Inc. Fort Worth, TX) was topically administered to prevent ocular infection.
Spectral domain-optical coherence tomography (SD-OCT)
Mouse retinas were scanned with an SD-OCT Ophthalmic Imaging System (Bioptigen Inc. Durham, NC). Briefly, mice were anesthetized using a ketamine/xylazine cocktail (100/10 mg/kg) and placed on the mouse holder to fix the mouse posture for scanning. The retina was scanned with rectangular scanning mode (1.2 mm diameter) consisting of 100 B scans/1000 A scans per B scan using InVivoVue software (Bioptigen Inc, Durham, NC). Superior, center and inferior images were imported and analyzed by ImageJ software (NIH) with four vertical calipers on each retinal layer (n = 4-5 mice per time point).
Flash scotopic electroretinogram (ERG)
Mice were dark adapted for 16 hrs. After anesthesia with Ketamine/Xylazine (100/10 mg/kg), mice were placed in a Gantzfield light chamber on the LKC electroretinogram system (LKC Technologies Inc., Gaithersburg, MD) with temperature control (37°C). Amplitude and implicit times of ERG waveforms were measured at a series of flash intensities (-30, -20, -10, 0, 5, 10, 15 dB) (n = 9 mice per time point).
Histology and cell counting in RGC layer
Eyes were harvested and fixed in neutral buffered 10% or 4% paraformaldehyde. After paraffin embedding, retinal cross sections were prepared (5 μm) followed by Hematoxylin-Eosin (H&E) staining for morphological observation of the retinal layers. Four retinal sections from ora serrata to ora serrata through the optic nerve head were chosen from each eye and the cells in the RGC layer were counted and averaged. Day 0 RGC counts from non-ischemic control eyes were set as 100%, and RGC cell counts in the rest of the eyes were compared to these controls.
Analysis of retinal thickness
H&E stained whole retina or individual retinal layer thicknesses were measured using ImageJ software (NIH). Individual or whole (RGCL to ONL) layer thickness from 4 retinal cross-sections per eye were measured at quarterly points for each retinal cross-section and averaged. For SD-OCT images, layer thickness at 2 different distances from optic nerve head (~0.35 and 0.55 μm) was determined and averaged. Averaged retinal thickness was converted to the percentage of the thicknesses from day 0 non-ischemic contralateral eyes.
RNA extraction and DNA microarray analysis
Total retinal RNA was extracted from control and I/R injured eyes. Retinas were collected and homogenized in Iso-RNA Lysis Reagent (5 PRIME Inc. Gaithersburg, MD). Total RNA was further extracted using an RNeasy Micro Kit (QIAGEN Science, Germantown, MD). RNA quality was controlled by determining RNA integrity number (RIN) using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA) with nano- or picochip systems. Total RNA with RINs >7 were selected and pooled from either control or experimental eyes at each time point for gene expression microarray analysis. Microarray hybridizations were performed at the University of Iowa DNA Core Facility. Briefly, 50 ng total RNA was converted to SPIA amplified cDNA using the WT-Ovation Pico RNA Amplification System, v2 (NuGEN Technologies, San Carlos, CA, Cat. #3302) according to the manufacturer’s recommended protocol. The amplified SPIA cDNA product was purified through a QIAGEN QIAquick PCR Purification column (QIAGEN Cat #28104) according to modifications from NuGEN. Five microgram samples were fragmented (average fragment size = 85 bases) and biotin labeled using the NuGEN FL-Ovation cDNA Biotin Module (NuGEN Technologies, Cat. #4200) per the manufacturer’s recommended protocol. The resulting biotin-labeled cDNA was mixed with Affymetrix eukaryotic hybridization buffer (Affymetrix, Inc., Santa Clara, CA), placed onto Affymetrix Mouse Gene 1.0 ST arrays (Part No. 901168) (Affymetrix Inc. Santa Clara, CA), and incubated at 45°C for 18 h with 60 rpm rotation in an Affymetrix Model 640 Genechip Hybridization Oven. Following hybridization, the arrays were washed, stained with streptavidin-phycoerythrin (Molecular Probes, Inc., Eugene, OR), and the signals were amplified with anti-streptavidin antibody (Vector Laboratories, Inc., Burlingame, CA) using the Affymetrix Model 450 Fluidics Station. Arrays were scanned with the Affymetrix Model 3000 scanner with the 7G upgrade and data were collected using the using the GeneChip operating software (GCOS) v1.4.
The 16 data CEL files for all 8 time points (0, 6 hr., 1, 3, 7, 14, 21 and 28 days) after I/R injury were imported into the Partek Genomics Suite 6.6 software (Partek Inc., Louis, MO) and normalized based on robust multi-array averaging (RMA). At each time point, the retinal I/R sample was compared to the contralateral control samples to calculate the microarray ratios and log
2 fold change values. Using Excel, the selective filter of ≥1.5 fold change was utilized to identify up-regulated genes and ≤ -1.5 for the down-regulated genes per time point for the retina datasets. The genes were further analyzed using the publicly available bioinformatics software DAVID (Database for Annotation, Visualization and Integrated Discovery). Gene ontology (GO) based cluster analysis was performed to identify possible enrichment of genes (GO enrichment score calculated using a χ
2 test) using filtered genes from each time point. The Fishers Exact p value is calculated by DAVID to identify GO enrichment based clusters, and p values <0.05 were considered to be significant in the enriched annotation category based on the Benjamini multiple test correction[
34,
35]. Clusters of genes were identified at specific time points and changes in gene expression were graphed temporally for each GO category.
Real-time reverse transcription (RT)-polymerase chain reaction (PCR)
Complementary DNA (cDNA) was generated from the individual retinal RNA samples (n = 4-5) that were pooled for the microarray study. 2 μg of total RNA was used for reverse transcription (RT) at 25°C (10 min), 37°C (120 min) and 85°C (5 min) using a Multiscribe reverse transcriptase kit (Applied Biosystems, Life Technologies Corp., Grand Island, NY) and a PTC-100 thermal cycler (MJ Research Inc. Waltham, MA). Gene expression of cryaa, cryba1, ccl12, c3 and gapdh was analyzed using Taqman gene expression assay primer sets (Applied Biosystems, Life Technologies Corp., Grand Island, NY) with Taqman Fast Advanced Master Mix (Applied Biosystems, Life Technologies Corp., Grand Island, NY) on a CFX96 real-time system (Bio-Rad, Life Science Research, Hercules, CA) with recommended thermal cycles by manufacturer (hold 50°C (2 min), hold 95°C (20 sec), 40 cycles of denaturation (3 sec) and annealing /extension (30 sec)). The expression of the housekeeping gene gapdh was used as an internal control to normalize target gene expression between samples. Differences in target gene expression were calculated using the following formula: ΔΔCT = ΔCT (target gene) – ΔCT (gapdh). The ΔΔCT value of cDNA amplification from the I/R eye was normalized to the control right eye. 0 hr was set as 1 and data from other time points were expressed as fold difference compared to 0 hr data.
Immunohistochemistry
Sections of paraffin embedded retinas were subjected to antigen retrieval in citrate buffer (pH 6.0), and the slides were washed twice in PBS. Slides were blocked for 1 hr using SuperBlock Blocking Buffer (Thermo Scientific, Rockford, IL), and then incubated with mouse anti-Brn3a (1:50) (Cat # MAB 1585, EMD Millipore Corp. Billerica, MA), mouse anti-NeuN (1:1000) (Cat # MAB 377, EMD Millipore Corp. Billerica, MA), or rabbit anti-GFAP (1:200) (glial fibrillary acidic protein) (Abcam plc. cat.# ab7779, Cambridge, MA) for 16 hr. at 4°C, followed by washing and then incubation with goat anti-mouse antibody (1:500) or goat anti-rabbit antibody (1:500) conjugated with Alexaflour 488 (Cat.# A11001 or A11008, Invitrogen, Life Technologies, Grand Island, NY) for 1 hr. at room temperature. Cover slips were mounted using ProlongGold anti-fade reagent with DAPI (Molecular probes, Life Technologies, Grand Island, NY) for nuclear visualization. Images were acquired using a Nikon Eclipse Ti inverted microscope (Nikon Instruments Inc. Melville, NY) and a CRi Nuance FX multispectral imaging system (Caliper Life Sciences, Hopkinton, MA). Autofluorescence was subtracted using the Nuance 3.0 software.
Discussion
Ischemic damage to different tissues such as brain and kidney share many similar pathologies [
41,
42]. In particular, retinal I/R results in neuronal degeneration associated with visual impairment and irreversible destruction of many layers of the structurally complex retina. Common morphological features in rodent models of retinal I/R include morphological degeneration of retinal layers, RGC death, and impairment of retinal function [
18‐
21,
24‐
26]. In order to discover overall pathological mechanisms, we observed pathological progression of I/R injury over 35 days, studying impairments in retinal morphology, function, and temporal changes in gene expression. There was a significant increase in thickness of inner retinal layers 3 days after I/R injury, most likely due to initial retinal edema. This histological finding was confirmed using non-invasive SD-OCT scanning. SD-OCT is now frequently used clinically and experimentally to detect morphological features of retina [
43‐
47]. This technique allows live monitoring of the retina without mechanical invasion or damage. In contrast, traditional histological assessment is terminal and can introduce artifacts due to tissue processing. SD-OCT scanning allowed us to detect retinal detachment 3 days post injury in all mice, which was confirmed by histology (Figure
1). Both techniques confirmed time-dependent changes in retinal morphology after I/R injury.
Previous studies have suggested diverse molecular events promoting or attenuating I/R-induced retinal damage. A number of neuroprotective approaches have been tested in retinal I/R models. One major effort was to define protective mechanisms using pharmacological approaches. Ueda et al. reported that retinal neuronal injury occurred by both necrosis and apoptosis, which were inhibited by treatment with the cognition-enhancer, Nefiracetam [
48]. Sun et al. also demonstrated that Cobalt protophorphyrin-induced Heme Oxygenase-1 attenuated I/R induced injury in retina [
49]. The hypolopodemic drug Simvastatin also improved retinal ganglion cell survival in mouse model of retinal I/R injury [
50]. More recently, Fujita et al. showed that pharmacological blocking of Angiotensin II type 1 receptor signaling produced neuroprotection via suppression of ROS production [
51]. Genetic disruption of specific genes in mice has also been used to identify pathogenic and neuroprotective pathways. Aquaporin-4 null mice were protected from ischemia-induced retinal functional impairment and cell death [
52]. Deletion of complement component C3 also induced retinal protection against I/R injury [
32].
It is also important to examine molecular mechanisms associated with progression of retinal injury. cDNA microarray analysis is often used to provide extensive mRNA expression data [
53‐
56]. Gene arrays provide profiles of functional gene clusters using a variety of bioinformatic approaches [
34,
35]. Youshimura et al. previously reported the temporal and spatial expression of immediate early genes in retinal neurons after retinal I/R injury [
57]. Hollborn et al., also specifically identified inflammatory and immune-response-related genes activated in the early stage of experimental retinal detachment [
58]. In addition, Kamphuis et al. evaluated changes in retinal gene expression following ischemic preconditioning [
59]. In our study, we performed cDNA microarray analysis at 8 different time points after retinal I/R injury and verified their differential expression with real-time RT-PCR (Figures
8 and
9). Based on our gene clustering data, we observed temporal changes of several genes related with signaling pathways, structure/cellular stress and inflammation, based on their relation with ischemic diseases in the retina or other tissues. For example, Stat3 is protective in various ischemic diseases including retinal I/R injury [
30]. Modulation of glutathione peroxidase (
gpx) expression has been reported under ischemic environments in various tissues [
60‐
63].
bcl6 was originally known as a modulator of STAT-dependent interleukin-4 (IL-4) response in B cells [
64].
bcl6 is induced in circulated leukocytes after ischemic stroke, but its precise role in this condition is unknown [
65]. Caspase8 is a major molecule in the apoptotic cascade involved in ischemia-induced cell death [
66,
67]. Although crystallins were originally known as structural proteins in lens [
68‐
70] crystallins also are molecular chaperones structurally similar to small heat shock protein (hsp) with the ability to prevent protein aggregation [
70‐
73]. Altered expression of the crystallins has been observed in various ocular diseases such as diabetic retinopathy, uveitis and glaucoma [
74‐
78]. However, no clear changes in crystallin expression has been shown previously in retinal I/R injury.
One of the most interesting changes was the up-regulation of various inflammatory genes including
c3,
c4b,
ccl12, and
gfap (Figures
8,
9,
10). In particular, several researchers discovered that complement components play important roles in eye development and ocular pathology such as glaucoma [
79]. In addition, genetic disruption of
c3 protected mice against retinal I/R injury [
32]. Both
c3 and
c4b genes encode C3 and C4b proteins, essential for the classical complement cascade [
80,
81]. Expression of these proteins is temporally regulated and may play differential roles at different times during I/R injury. Retinal GFAP expression is mainly observed in Müller cells during retinal injury [
39,
82,
83]. Like other glial cells in neuronal system, Müller cells play a pivotal role to maintain retinal neuron homeostasis, such as scavenging neurotransmitter/waste products, supplying energy for retinal neurons, and other protective and maintenance roles for neurons [
84,
85]. Under pathological conditions, Müller cells are activated, undergoing functional and morphological changes associated with gliosis [
39,
82]. Hirrlinger et al. demonstrated that transient retinal ischemia in mice induced Müller cell gliosis accompanied by altered protein expression and changes in membrane properties [
86]. Our data provide further support for Müller cell-dependent retinal gliosis, involving changes in gene and protein expression as well as Müller cell morphological changes that correlate with progression of I/R injury.
As previously mentioned, retinal thickness, especially in the inner retina, significantly increased within 3 days of retinal I/R injury. It is likely that this increased retinal thickness was due to retinal edema. Others have shown that the increased retinal thickness in retinal ischemia is due to retinal edema, which may be mediated by ET-1 in endothelial cells [
12]. Retinal edema plays a major role in the pathogenesis of other types of retinal injury, including retinal vessel occlusion and diabetic retinopathy. In addition, progressive retinal degeneration following edema in our model was also strongly correlated with significant cell loss in the RGCL, including RGC and displaced amacrine cells. Immunohistological assessment using RGC specific antibodies to Brn3a and NeuN showed RGC loss induced by I/R injury. Interestingly, a similar cell loss ratio from histology data (~30%) was also shown in our immunohistochemistry data (~30%) at 28 days after I/R injury. These data suggest that I/R injury caused cell loss of both RGC and displaced amacrine cells in the RGCL. In support, Kim et al., previously showed that retinal I/R injury induced apoptotic cell death to both RGCs and displaced amacrine cells [
87].
One of our novel observations is I/R-induced retinal detachment (Figure
1). Retinal detachment is a major cause of vision loss in various ocular pathologies, including age-related macular degeneration (AMD) [
88‐
90]. We first observed the retinal detachment in all ischemic eyes by SD-OCT scanning at days 3 and 7 after I/R injury, but this detachment disappeared at 14 days. We took advantage of SD-OCT scanning to monitor the real-time morphological status of the retina without sacrificing mice. Interestingly, Zeng et al. developed a novel mouse model of retinal detachment using a similar cannulation method [
91]. In contrast to our data, retinal detachment completely recovered within 24 hrs in their model. In support, Uckermann et al. also suggested that transient retinal ischemia in rabbits can cause exudative detachment of the retina through days 3 and 8, which was accompanied by changes in Müller cell K
+ conductance [
92]. They suggested fluid-mediated retinal detachment as a novel ischemia-mediated damage to the ONL. Our findings also support retinal detachment as another potential pathologic mechanism for temporal retinal dysfunction and degeneration. Our ERG a-wave data, which is associated with outer retinal photoreceptor function, also support this finding (Figure
6). Retinal detachment was associated with ERG a-wave amplitude deficits from days 21-28 after I/R injury. Interestingly, a-wave amplitudes recovered at 35 days, correlating with recovery of retinal detachment. These results suggest that early retinal detachment causes delayed outer retina ERG deficits that are reversible after retinal reattachment. Therefore, retinal detachment in the mouse model of retinal I/R may contribute to overall ischemia-mediated retinal damage.
Interestingly, all of our data showed strong temporal correlations. We detected thickening of the whole retina, especially the inner retina, 3-7 days after I/R injury. Fourteen days post I/R injury, four significant changes simultaneously occurred including: (1) decreased inner retinal layer thickness, (2) significant loss of cells in RGC layer, (3) significant changes in gene expression profiles, and (4) increased GFAP immunostaining (gliosis) throughout entire retina. Functional impairment (i.e. decreased ERG responses) began 7 days after I/R injury, suggesting decreased retinal function was due to early retinal edema and/or damaged retinal cells prior to morphological degeneration. Our findings are the first to demonstrate temporal morphological changes accompanied with functional and molecular changes associated with progression of retinal I/R injury.
In conclusion, transient I/R induced morphological changes mainly in the inner retina that were strongly associated with functional impairment as well as temporal changes in retinal gene expression. Our data also indicated that retinal detachment was induced by retinal ischemia in the early stages of injury. Our characterization of temporal retinal changes produced by retinal ischemia will lead to a better understanding of molecular pathogenesis associated with this injury as well as suggest novel therapeutic approaches to mitigate this retinal damage. Future studies will identify cellular and molecular mechanisms associated with I/R damage to the optic nerve and visual axis in the brain. This will lead to the discovery of new neuroprotective strategies and agents for the treatment of the retina, optic nerve, and visual axis in the brain associated with retinal I/R injury.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
BK performed all experiments except microarray procedure and analyzed all data. BK also drafted manuscript including text and figures. TB participated in the microarray studies. RW provided guidance in the design and coordination of the study. AC designed and coordinated the study and finalized manuscript. All authors read and approved the final manuscript.