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 . Sun et al. also demonstrated that Cobalt protophorphyrin-induced Heme Oxygenase-1 attenuated I/R induced injury in retina . The hypolopodemic drug Simvastatin also improved retinal ganglion cell survival in mouse model of retinal I/R injury . More recently, Fujita et al. showed that pharmacological blocking of Angiotensin II type 1 receptor signaling produced neuroprotection via suppression of ROS production . 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 . Deletion of complement component C3 also induced retinal protection against I/R injury .
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 . Hollborn et al., also specifically identified inflammatory and immune-response-related genes activated in the early stage of experimental retinal detachment . In addition, Kamphuis et al. evaluated changes in retinal gene expression following ischemic preconditioning . 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 . 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 . bcl6 is induced in circulated leukocytes after ischemic stroke, but its precise role in this condition is unknown . 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 . In addition, genetic disruption of c3 protected mice against retinal I/R injury . 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 . 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 . 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 .
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 . 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 . 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.