Microglia Activation in Retinal Ischemia Triggers Cytokine and Toll-Like Receptor Response

Mechanisms and progression of ischemic injuries in the retina are still incompletely clarified. Therefore, the time course of microglia activation as well as resulting cytokine expression and downstream signaling were investigated. Ischemia was induced in one eye by transiently elevated intraocular pressure (60 min) followed by reperfusion; the other eye served as a control. Eyes were processed for RT-qPCR and immunohistochemistry analyses at 2, 6, 12, and 24 h as well as at 3 and 7 days. Already 2 h after ischemia, more microglia/macrophages were in an active state in the ischemia group. This was accompanied by an upregulation of pro-inflammatory cytokines, like IL-1β, IL-6, TNFα, and TGFβ. Activation of TLR3, TLR2, and the adaptor molecule Myd88 was also observed after 2 h. NFκB revealed a wave-like activation pattern. In addition, an extrinsic caspase pathway activation was noted at early time points, while enhanced numbers of cleaved caspase 3+ cells could be observed in ischemic retinae throughout the study. Retinal ischemia induced an early and strong microglia/macrophage response as well as cytokine and apoptotic activation processes. Moreover, in early and late ischemic damaging processes, TLR expression and downstream signaling were involved, suggesting an involvement in neuronal death in ischemic retinae.


Introduction
Ischemia is based on limited blood supply in a localized area, due to the blockage of blood vessels in the affected area. This in turn causes energy exhaustion and ultimately cell death. In the retina, ischemia develops as a result of capillary blockage, results in non-perfusion of this region, and leads to dysfunction as well as death of neuronal cells (Osborne et al. 2004). This is accompanied by glia cell activation followed by a secretion of cytotoxic mediators, including cytokines. Ischemic processes occur in several eye diseases, such as diabetic retinopathy or glaucoma (Schmidl et al. 2011;Sim et al. 2013;Terelak-Borys et al. 2012).
Glaucoma is the second most common cause of blindness worldwide, with a rising trend due to an aging society (EGS 2014;Quigley and Broman 2006). Furthermore, glaucoma not only leads to health problems but also results in economic damage. In Europe, the cost of treating glaucoma patients rose from 455 € to 969 € per person and, in the USA, the cost for treatment for this disease is estimated at $2.5 billion a year (Lazcano-Gomez et al. 2016;Traverso et al. 2005). In general, glaucoma is a multi-factorial disease characterized by a chronic loss of retinal ganglion cells (RGCs) and their axons (Casson et al. 2012;EGS 2014). Currently, the elevated intraocular pressure (IOP) is the only risk factor that can be handled by medical or surgical treatment. Unfortunately, in many cases, optic nerve and RGC degeneration as well as visual field loss continue on a long-term basis (Chang and Goldberg 2012;Pascale et al. 2012). Consequently, improved knowledge about the ischemia downstream processes can help to develop novel treatment options.
A common animal model to study pathological ischemic processes is the retinal ischemia/reperfusion model. Here, the IOP is temporarily increased by the infusion of fluid into the anterior chamber with a subsequent natural reperfusion. The consequence is a lower oxygenation capacity and supply of Electronic supplementary material The online version of this article (https://doi.org/10.1007/s12031-020-01674-w) contains supplementary material, which is available to authorized users. nutrients followed by the formation of oxidative stress during the recurring blood flow (Kaur et al. 2008;Kim et al. 2013;Minhas et al. 2012). It is known that RGCs and other cell types of the inner retinal layers are mainly affected in this model (Palmhof et al. 2019a;Schmid et al. 2014;Zheng et al. 2004). Previous studies on retinal ischemia have mostly analyzed only one or a few time points. Therefore, our research group recently investigated how cell types are affected over time in this model. At 3 days, a significant decrease in the total retinal thickness was observed, while changes in the RGC layer and a specific loss of RGCs were already noted after a few hours, which continued to increase over time (Palmhof et al. 2019a).
Microglia are the only permanent immune cells of the central nervous system (CNS) (Hanisch and Kettenmann 2007). In line, many molecules and conditions can trigger a transformation of resting or surveying microglia to alerted or reactive states. Therefore, they are the first line of defense against neuronal injury or ischemia (Lyons et al. 2000;Schmid et al. 2014). A few hours after ischemia, inflammation as well as an activation of glial cells and angiogenesis occurs. In consequence, proinflammatory cytokines, like interleukin (IL)-1β, IL-6, and γ-interferon, are activated (Berger et al. 2008;Rivera et al. 2013;Yoneda et al. 2001). In general, cytokines can affect not only the cells they are released from but also adjacent ones, which lead to critical misbalance in retinal diseases. Increased levels of IL-1 and other inflammatory markers were detected in vitreous samples of diabetic retinopathy patients (Reverter et al. 2009). Also, in retinal ischemia animal models, increased VEGF (Abcouwer et al. 2010;Chen et al. 2012;Hayashi et al. 1996) and IL-6 (Hangai et al. 1996;Sanchez et al. 2003;Wang et al. 2006) levels were observed after injury. IL-6 alterations are already detectable about 2 h after ischemia (Sanchez et al. 2003). A comparable early response occurs after ischemia/reperfusion brain injury (Berti et al. 2002). These effects could contribute to secondary cellular responses that lead to further retinal damage.
Toll-like receptors (TLRs) are pattern recognition receptors that play an important role in the initiation of the immune system. In line, TLRs generate signals which are passed through the NFκB signaling pathway as well as the MAP kinase pathway to recruit pro-inflammatory cytokines which in turn promote inflammatory response (Vidya et al. 2018). Interestingly, when microglia are triggered by a stimulus, such as oxidative injury, activated microglia express higher levels of TLRs in the CNS (Carpentier et al. 2008). Furthermore, previous studies showed a wide range of TLRs on microglia/ macrophages which can be stimulated by secreted cytokines. This in turn reinforces the proinflammatory environment as well as neuronal dysfunction and thus enhances neuronal cell death (Chen et al. 2017;Li et al. 2014).
RGCs are strongly affected in the retinal ischemiareperfusion model (Palmhof et al. 2019b). It is known that RGC cell death in this model is caused mainly by apoptosis (Lam et al. 1999;Selles-Navarro et al. 1996). Several initiator caspases (e.g. caspase 8 or 9) trigger both caspase pathways (intrinsic and extrinsic). Through effectors, like caspase 3, the common final track of the signaling cascade is activated (Erekat 2018; Kurokawa and Kornbluth 2009).
The aim of this project was to depict the timeline of microglia/macrophage response due to retinal ischemia. In addition, the importance of microglia associated cytokine expression as well as TLR signaling and apoptosis for retinal ischemia will be analyzed in detail. We could show that already 2 h after ischemia, microglia as well as their proinflammatory cytokines were enhanced. This was accompanied by TLR signaling and apoptosis.
The improved knowledge of the ischemia-related downstream processes and the importance of microglia gained from this study will then contribute to the development of new treatment options.

Ischemia/Reperfusion Model
Male brown Norway rats (7-8 weeks of age; Charles River Laboratories, Sulzfeld, Germany) were used for this ischemia/ reperfusion project. The study was approved by the animal care committee of North Rhine-Westphalia (Germany); all experiments were carried out in accordance with the ARVO statement for the use of animals in ophthalmic and vision research. Rats were housed under environmentally controlled conditions (12-h light-dark cycle) with free access to chow and water.
Retinal ischemia/reperfusion was induced as previously described (Joachim et al. 2017;Palmhof et al. 2018;Schmid et al. 2014). An elevated IOP of 140 mmHg was applied to the right eye of each animal for 60 min, followed by reperfusion (Fig. 1). The left, untreated eye served as a control.

Quantitative Real-time PCR Analysis
Retinae (n = 5/group/time point) were used for RNA preparation and cDNA synthesis as previously described (Osborne et al. 2004;Reinehr et al. 2019). A PikoReal 96 Real-time PCR System (Thermo Fisher Scientific) with SYBR Green (DyNAmo Flash SYBR Green qPCR Kit; Thermo Fisher Scientific) was used for RT-qPCR. Primer efficiencies of each primer set were calculated based on a dilution series of 5 to 125 ng cDNA. Ct values of the house-keeping genes β-actin (Actb) and Cyclophilin (Ppid) were applied for normalization and relative quantification of gene expressions (Table 1).

Immunohistochemistry
Six retinal cross sections (n = 7-8/group/time point) were stained per immunohistochemistry marker (Table 2) (Palmhof et al. 2019a). Therefore, the sections were first thawed and then rehydrated in PBS, followed by blocking in 10-20% appropriate serum with or without 1% BSA in 0.1% Triton X-100 in PBS. Specific first and secondary antibodies were applied (Table 2), followed by DAPI (4′,6-diamidin-2-phenylindol; Serva Electrophoresis, Heidelberg, Germany) to visualize cell nuclei. For each staining, negative controls were included, with secondary antibodies only. Four pictures per section were taken with a fluorescence microscope (Axio Imager M1 and M2; Carl Zeiss Microscopy, Jena, Germany). All images were transferred to Corel Paint Shop Photo Pro (V 13; Corel Corporation, Fremont, CA, USA), masked, and equal excerpts were cut out.

Statistical Analysis
RT-qPCR data are presented as median ± quartile + minimum/maximum and immunohistology data as mean ± standard error mean (SEM) with *p < 0.05, **p < 0.01, and ***p < 0.001. Relative expression variations in RT-qPCR analyses were done by REST© software (QIAGEN GmbH, Hilden, Germany) using a pairwise fixed reallocation and randomization test. Regarding immunohistochemistry, both groups were compared per point in time using Student's t test (Statistica V13.3; Dell; Tulsa, OK, USA).
Moreover, the total microglia/macrophage population and the ones in an active state were stained for not only 2, 6, 12, and 24 h but also 3 and 7 days after ischemia induction (Fig. 2d). Already at 2 h, ischemic retinae displayed more Iba1 + cells then control ones (p = 0.005; Table 3; Fig. 2e). These numbers further increased over time (6 h: p = 0.001, 12 and 24 h: p < 0.001). At 3 and 7 days, microglia/macrophage numbers in the ischemia group had more than doubled compared with controls (both: p < 0.001).
The relative expression levels of Tnfa were already upregulated 2 h after ischemia induction (10.95-fold expression, p = 0.004; Fig. 3c). Higher expression levels of Tnfa were also revealed at 6 (5.80-fold expression, p = 0.005), 12 (6.21-fold expression, p = 0.004), and 24 h (3.35-fold expression, p = 0.006) in ischemic retinae. At day 3, the relative Tnfa expression was still higher in the ischemia group than in the control retinae (1.90-fold expression, p = 0.017) but was already lower than the expression level measured at 24 h. At 7 days, no difference in the expression level of Tnfa was detectable between the ischemic and the control groups (0.82-fold expression, p = 0.49).
In addition to the RT-qPCR analyzes, immunohistochemical stainings were also carried out with an anti-TLR3 antibody (Fig. 4d). In contrast to mRNA levels, an early significant difference was detected in the number of TLR3 + cells in ischemic retina. Already at 2 h, more TLR3 + signals were noted in the ischemia group compared with controls (p = 0.029; Table 3; Fig. 4e). More TLR3 + signals were also present from 6 h up to 3 days (6 h: p = 0.038; 12 h: p = 0.001, 24 h: p < 0.001; 3 days: p = 0.042). However, at 7 days, both groups had a comparable number of TLR3 + signals (p = 0.128).

Activation of the TLR/MyD88/NFκB Signaling Pathway
The relative gene expression levels of the TRL adaptor molecule Myeloid differentiation primary response 88 (Myd88) were also significantly higher in the ischemic animals in comparison to control retinae at all investigated time points (Fig. 5a). Already 2 h after ischemia induction, a significantly increased Myd88 mRNA level was noted (2.02-fold expression, p = 0.028). Also, at 6 (4.20-fold expression, p = 0.001), 12 (6.45-fold expression, p = 0.003), and 24 h (2.05-fold expression, p = 0.01), a significant upregulation of the Myd88 mRNA expression was visible in ischemic retinae. At both late time points, 3 (2.28-fold expression, p = 0.024) and 7 days (2.79-fold expression, p = 0.001), the Myd88 expression was still higher in the ischemia group. The ubiquitously expressed transcription factor nuclear factor κB (NFκB), as a downstream regulator of the TLRs signaling cascade, is involved in inflammatory and apoptotic processes (Hayden et al. 2006;Trahtemberg and Mevorach 2017). Therefore, it mediates either cell survival or cell death events. To determine the Nfκb expression, specific primers were analyzed via RT-qPCR (Fig. 5b). No change in the Nfκb mRNA expression was found between ischemic and control retinae after 2 (0.63-fold expression, p = 0.100) and 24 h (0.57-fold expression, p = 0.066) as well as at 3 days (1.51-fold expression, p = 0.111). Increased expression levels of Nfκb were detected at 6 (2.76-fold expression, p = 0.002) and 12 h (2.83-fold expression, p = 0.002) in ischemic retinae.
Moreover, at 7 days, a significant upregulation in the expression level of Nfκb was detectable in the ischemic group (1.73fold expression, p = 0.001).

Extrinsic and Intrinsic Caspase-Dependent Apoptotic Pathways
Caspase 3 is a member of the cysteine-aspartic acid protease (caspase) family which interacts with caspase 8 and caspase 9. The Casp3 expression was determined via specific primers through RT-qPCR analysis (Fig. 6a). No difference in the Casp3 expression was observed at 2 (0.91-fold expression, p = 0.700) and 24 h (0.98-fold expression, p = 0.973) as well as at 3 days (1.04-fold expression, p = 0.891). However, a significant upregulated Casp3 expression was detected at the early time points, namely 6 (1.94-fold expression, p = 0.003) and 12 h (2.16-fold expression, p < 0.001). At the latest time point of 7 days after ischemia induction, an increase of Casp3 was also detectable (2.04-fold expression, p = 0.021).

Time-Dependent Upregulation of Microglia and Macrophages by Retinal Ischemia
Ischemic processes are involved in many CNS diseases and also in retinal ischemia. An activation of microglia and a recruitment of macrophages are noted in numerous neurological disorders. This can be beneficial or detrimental for disease progression (Batchelor et al. 2002;Krady et al. 2005;London et al. 2013;Nakagawa and Chiba 2015;Prinz and Priller 2014). The activation of microglia/macrophages was therefore intensely studied in various retina degeneration models, like glaucoma (O'Koren et al. 2016;Rojas et al. 2014) or ischemia/reperfusion models (Cho et al. 2011;Halder et al. 2013;Santos et al. 2010;Zhang et al. 2005a). However, the role of microglia/macrophages in the response to an ischemic damage in the retina has not yet been clarified. Hence, we analyzed the response of resident and active microglia/macrophages at several time points after ischemic injury. Iba1 mRNA expression was upregulated at late stages (at 3 and 7 days). Comparing these data to our immunohistochemical analyses, we saw a clear difference. We already noted a significantly higher number of Iba1 + cells at 2 h until 7 days after ischemia. Thus, a difference on protein level was noted earlier than on mRNA level. Nevertheless, we should keep in mind that Iba1 + cells were counted in the inner retinal layers (ganglion cell layer and inner plexiform layer), while the mRNA of the whole retina was analyzed. Moreover, this inconsistency could be explained with posttranscriptional and translational regulation.
However, a significant difference in the Cd68 mRNA expression was observed already 12 h after ischemic damage, which was still present at 7 days. To differentiate between microglia and macrophages, we examined Tmem119, which is exclusively expressed in microglia (Bennett et al. 2016;Satoh et al. 2016). Interestingly, the regulation of the Tmem119 mRNA showed a timedependent expression pattern. At the early time points (2 and 6 h), we discovered a significant downregulation in ischemic retinae. However, at 12 and 24 h, no differences between the groups were noted. Interestingly, a significantly higher Tmem119 expression was observed in the ischemia group at the two late time points, 3 and 7 days. This result was surprising, since an activation of microglia was expected due to the retinal injury. Nevertheless, the exact function of Tmem119 in microglia is not known yet (van Wageningen et al. 2019). In brains affected by Alzheimer's disease, TMEM119 was documented to be expressed on microglia with ramified and amoeboid morphologies (Satoh et al. 2016), indicating a late infiltration by microglia due to time of development of the ischemic damage. The results of van Wageningen et al. showed an unexpected downregulation of the Tmem119 immunoreactivity in white matter lesions of multiple sclerosis patients. They explained this effect with the presence of lymphocytes and i nflammatory mediators (van Wageningen et al. 2019). The inflammatory recruitment of microglia/macrophages seems to need time to respond to the ischemic damage.

Early Cytokine Release
In this study, we analyzed the mRNA expression of various cytokines, like Il1b, Il6, Tnfα, and Tgfβ. Regarding Il1b, Il6, and Tnfα, they already reacted 2 h after ischemia with an elevated expression. This upregulation was constant during all investigated time points for Il1b. This could also be demonstrated on protein level using an anti-IL-1β antibody. Il6, on the other hand, was not upregulated at 3 days and Tnfα was not higher expressed at 7 days. Hangai et al. identified RGCs, endothelial cells, and recruited neutrophils in the retina as the source of Il1b (Hangai et al. 1995). They suggested that IL-1β plays an essential role in ischemic injuries. In their ischemia rat model, they induced the damage via a ligation of the optic nerve. Three hours after the induction, they documented an early upregulation in the Il1b expression with two Fig. 4 Time-dependent activation of toll-like receptors. a On mRNA level, a significant upregulation of Tlr2 mRNA expression could be detected in the ischemic group at all analyzed time points (2 h: p = 0.004; 6 h: p = 0.007; 12 h: p = 0.004; 24 h: p = 0.015; 3 days: p = 0.002; 7 days: p = 0.003). b In contrast, Tlr3 mRNA expression level was unaltered between 2 and 24 h but was upregulated 3 (p = 0.035) and 7 days (p = 0.009) after ischemia/reperfusion. c No altered Tlr4 expression could be noted at all time points via RT-qPCR. d Exemplary images of retinal cross sections labeled with an anti-TLR3 antibody (red) and cell nuclei were counterstained with DAPI (blue). e The evaluation of the amount of TLR3 + cells revealed a significant higher number in ischemic retinae compared with control at all time points except 7 days (p = 0.128). Already at 2 h after ischemia, a significant difference was noted (2 h: p = 0.028; 6 h: p = 0.005; 12 h: p = 0.004; 24 h: p > 0.001; 3 days: p = 0.042). f Retinal cross sections were labeled with an anti-TLR4 antibody (red) and cell nuclei were counterstained with DAPI (blue). (G) More TLR4 + cells were already noted 2 h after ischemia (p = 0.020). Counts were still higher at 6 (p = 0.003), 12, and 24 h (both: p < 0.001). At 3 days (p = 0.016), an upregulation was still detectable, but not at 7 days. GCL ganglion cell layer, IPL inner plexiform layer. Values are median ± quartile + maximum/minimum for RT-qPCR and mean ± SEM for immunohistology; RT-qPCR: n = 5/group; immunohistology: n = 8/ group. *p < 0.05, **p < 0.01, ***p < 0.001. Scale bar 20 μm peaks, 12 and 96 h after ischemia (Hangai et al. 1996). The analyzed cytokines Il1b, Il6, Tnfa, and Tgfb all play important roles in various neurological disorders and have multiple functions involving cell differentiation, growth, and survival (Gruol and Nelson 1997;Yafai et al. 2014  The expression level of the TLR downstream transcription factor Nfκb revealed a wave-like shape after ischemia. No changes in the Nfκb expression were detectable after 2 and 24 h as well as after 3 days. In contrast, at 6 (p = 0.002) and 12 h (p = 0.002) as well as at 7 days (p = 0.001), a significant upregulation in Nfκb expression was detected. c Retinal cross sections were labeled with an anti-NFκB antibody (green), and cell nuclei were stained with DAPI (blue). d More NFκB + cells were visible in ischemic retinae at all investigated time points (2 h: p = 0.032; 6, 12, and 24 h: p < 0.001; 3 days: p = 0.005; 7 days: p = 0.047). GCL ganglion cell layer, IPL inner plexiform layer. Values are median ± quartile + maximum/minimum for RT-qPCR and mean ± SEM for immunohistology; RT-qPCR: n = 5/group; immunohistology: n = 8/group. *p < 0.05, **p < 0.01, ***p < 0.001. Scale bar 20 μm a neuroprotective effect on inner retinal neurons including RGCs after retinal ischemia (Loddick et al. 1998;Sanchez et al. 2003). Sanchez et al. confirmed also that microglial/phagocytic cells are the major source of retinal IL-6 after an injury (Sanchez et al. 2003). A deleterious role after ischemia/reperfusion injury was postulated for TNFα (Berger et al. 2008). Berger et al. noted that a neutralization of TNFα improved the retinal function in an in vivo model based on p55 and p75 knockout mice and rats with an elevated IOP of 120 mmHg for different durations from 30 to 60 min (Berger et al. 2008). This demonstrated a pivotal role of TNFα in the immune response to an ischemic damage. Striking was the wave-like expression of Tgfb in our study. An elevated mRNA level was only notable at 6 and 12 h and at the late time points, 3 and 7 days after ischemia. However, at 2 and 24 h, no significant difference was visible between the groups.

Early TLR2 and TLR3 Activation in Ischemic Retinae
Next, we investigated different members of the evolutionarily conserved TLR family, which can be divided into different subfamilies based on their individual recognition pattern (Akira and Takeda 2004;Beutler 2004;Janeway and Medzhitov 2002). Whereas TLR2 recognizes lipids and peptidoglycans, TLR4 recognizes different ligands like lipopolysaccharides, fibronectin, and heat shock proteins (Akira et al. 2006;Kielian 2006). In human glaucomatous retinae, TLR2, TLR3, and TLR4 were overexpressed. Thereby, TLR2 and TLR4 were distinctively noted in microglia cells, while TLR3 expression was prominently seen in astrocytes (Luo et al. 2010). TLR2 and TLR4 are located on the cell surface, while TLR3 is found solely in intracellular compartments. Adaptor proteins, such as MyD88, are essential for the mediation of the TLR downstream signaling pathway, except for TLR3 (Akira and Takeda 2004).
TLRs respond to pathogens, multiple cytokines, and environmental stress through a quick modulation. We noted a significant increased Tlr2 mRNA expression starting at 2 h until 7 days. Interestingly, we found a significant upregulation of the Tlr3 expression only at the both late time points (3 and 7 days). In addition, we noted a different regulation on protein level via TLR3 staining. A significantly increased number of TLR3 + cells was revealed at all investigated time points in ischemic compared with control retinae, except at the latest time point of 7 days. In contrast, no difference in the Tlr4 expression pattern was detected throughout the study. These results suggest that the immune response to the retinal damage is mediated via TLR2 starting at 2 h. In addition, the Myd88independent pathway via TLR3 seems to be activated. TLR4 and the associated downstream pathway on the other hand seem to be not exclusive in the immune response to retinal damage.
TLR2 is one of the most intensely investigated members of the TLRs relating to neurodegenerative disorders (Guillot-Sestier and Town 2018). However, previous studies mainly focused on TLR4 after ischemic injury as it is known that necrotic cells release endogenous ligands, which can be recognized by TLRs and activating an innate immune response (Dvoriantchikova et al. 2010a;Dvoriantchikova et al. 2010b;Fujita et al. 2009;Piccinini and Midwood 2010). It is worth noting that there seems to be a certain contradiction to the results of other groups, who also examined the role of TLR2 and TLR4 in different ischemia/reperfusion models (Dvoriantchikova et al. 2010b;Ishizuka et al. 2013;Qi et al. 2014). Qi et al. found a significant upregulation of Tlr4 starting 1 h after injury, while they could not find an altered Tlr2 expression (Qi et al. 2014). Therefore, they suggested that the TLR4 signaling pathway activates the NLRP3 inflammasomes. Nevertheless, their results were obtained from a different ischemia model. Our model is based on IOP-raising, while Qi et al. generated damage by ligating the central retinal artery (Qi et al. 2014). Ishizuka et al. used a retinal ischemia mouse model, where they ligated the arteries. Via Western blot analysis and immunohistochemistry, they observed more TLR4 1 and 19 h after ischemia. Additionally, they noted reduced damage after ischemic injury in TLR4 knockout mice (Ishizuka et al. 2013). The TLR4 staining of our study supports these findings, as we noted an upregulation of TLR4 starting at 2 h until 3 days after ischemia. Comparable findings of a TLR2 dominance over TLR4 under stress conditions like ischemia/reperfusion damage were also seen by Bagchi et al. in cardiac tissue of wild-type and IL-10 −/− mice and rats (Bagchi et al. 2017). They induced the ischemia via cannulation of the aorta. Interestingly, they measured a differential expression of TLR2 and TLR4 and proved a TLR2 increase contrary to TLR4 in IL-10 − / − mice, indicating a more prominent role of TLR2 in the immune response under ischemic conditions which needs to be analyzed further.

MyD88 Predominantly Activates Wave-Like NFκB-Mediated Response
As mentioned before, Myd88 is an adaptor molecule, which is critical to mediate the induction of inflammatory cytokines via NFκB stimulated by TLR pathways (Sugiyama et al. 2016). The results of our study indicate an activation of Myd88 via TLR2 signaling. This activation takes place simultaneously with the upregulation of TLR2. Thus, we noted a higher Myd88 mRNA expression from 2 h to 7 days.
An activation of the transcription factor NFκB was documented after ischemic injury in the CNS (Chen et al. 2003;Ridder and Schwaninger 2009;Zeng et al. 2008;Zhang et al. 2005b). NFκB, a key regulator of cell survival and inflammatory responses, is expressed by neuronal and glial cells (Oeckinghaus and Ghosh 2009). We observed timedependent changes regarding Nfκb mRNA levels after ischemic injury. At the early time points, a significant upregulation was followed by a gradual decrease. At 3 days, Nfκb mRNA levels in both groups were comparable, while we observed another distinct increase at 7 days. Nevertheless, we did not note an early upregulation of Nfκb mRNA in contrast to Wang et al. (2006). They also used an ischemia/reperfusion model with high IOP, but only applied 110 mmHg. However, we observed a wave-like decrease from 6 to 24 h, reaching a peak at 12 h. Wang et al. found comparable Nfκb mRNA levels, though they claimed to see the peak at 4 h after ischemia, followed by gradual decrease towards 24 h (Wang et al. 2006).

Enhanced Caspase Activation
It is known that RGCs are the most sensitive cell type to ischemic injury. Sellés-Navarro et al. induced ischemia over several periods, from 30 to 120 min, by increasing the IOP in rat eyes. They performed examinations at 5, 7, 14, and 30 days after injury to investigate the survival of RGCs (Selles-Navarro et al. 1996). They noted a loss of RGCs already after 5 days. In a previous study by our group, we also noted that RGCs are in particularly damaged after the ischemic insult, with a progressive degeneration (Palmhof et al. 2019b).
Apoptosis via activation of caspases leads to rapid cell death. Caspase 3 is one member of the caspase execution subfamily (Degterev et al. 2003). In our study, significantly more cleaved caspase 3 + cells were noted in the GCL of ischemic retinae at all investigated time points. Also, mRNA levels of Casp3 were found to be upregulated in early and late stages. Yang et al. observed more caspase 3 protein 6 h after ischemia in rats (Yang et al. 2011). Also in other organs, like in the kidney, ischemia-induced injury activates apoptosis through caspase 3 activity (Meng et al. 2020).
TLR4 can lead to an increased caspase 8 expression in a retinal ischemia-reperfusion model (Chi et al. 2014), which is in accordance with our study results. The authors postulate that a subsequent production of IL-1β occurs. Caspase 8 activation can also be engaged by inflammasome components. Inflammasomes are cytosolic complex that process cytokines, like IL-1β, in their active forms (Tummers and Green 2017).We noted a higher Il1b mRNA expression in ischemic retinae as well as more IL-1β + cells throughout the study. Knockout of TLR4 signaling suppressed caspase 8 activation Fig. 6 Apoptotic processes after ischemic damage. a The relative gene expression level of Casp3 was significantly upregulated in the ischemia group at 6 (p = 0.03) and 12 h (p < 0.001) as well as 7 days (p = 0.002) after ischemia. Nevertheless, no changes in the Casp3 expression were visible after 2 and 24 h as well as at 3 days. b No changes in the Casp8 expression were detectable after 2 and 24 h as well as 3 days. In contrast, at 6 (p = 0.006) and 12 h (p = 0.016) as well as at 7 days (p = 0.009) a significant upregulation in Casp8 expression was discovered. c No differences in the relative Casp9 expression were measured between control and ischemia samples at all investigated time points. d Retinal cross sections were labeled with an anti-cleaved caspase 3 antibody (green), while cell nuclei were stained with DAPI (blue). e More cleaved caspase 3 + cells were seen 2 (p < 0.001), 6 (p < 0.001), 12 (p < 0.001), and 24 h (p < 0.001) as well as 3 (p = 0.001), and 7 days (p = 0.026) after ischemia. f Sections were stained with anti-caspase 8 antibody (red). Cell nuclei were visualized with DAPI (blue). g More caspase 8 + cells were observed in ischemic retinae at 2 h (p < 0.001), 6 h (p < 0.001), 12 h (p = 0.003), and 24 h (p = 0.001). At 3 and 7 days, counts in both groups were comparable. GCL ganglion cell layer, IPL inner plexiform layer.
Values are median ± quartile + maximum/minimum for RT-qPCR and mean ± SEM for immunohistology; RT-qPCR: n = 5/group; immunohistology: n = 7/group. *p < 0.05, **p < 0.01, ***p < 0.001. Scale bar 20 μm Fig. 7 Schematic overview of study results. Retinal ischemia/reperfusion led to a progressive microglia activation. At early time points (2 to 24 h) after an ischemic injury, the immune response seemed to be mediated mainly by TLR2. The following cascade included MyD88, which activated NFκB. This transcription factor led to the release of proinflammatory cytokines, such as IL-1β, IL-6, TNFα, and TGFβ. TLR3 was also upregulated and a MyD88-independent signaling pathway was activated in the immune response via NFκB. An activation of the initiator caspase 8 in retinal ganglion cells was noted at early time points, while the effector caspase 3 was enhanced throughout the study and RGC death (Chi et al. 2014). Shabanzadeh et al. observed that caspase 8 inhibition decreases the neuropathological consequences of cerebral or retinal infarction in a retinal stroke model (Shabanzadeh et al. 2015).

Conclusion
In summary, we could demonstrate a strong, progressive activation of microglia/macrophages throughout this study. Cytokine regulations were already noted a few hours after ischemia and continued until 7 days. Furthermore, the damage was associated with an early activation of TLR2, TLR3, and their downstream signaling cascade. These findings were accompanied by apoptotic mechanisms throughout the study (Fig. 7).
Acknowledgments The authors would like to thank Thomas Biniasch, Anna Janus, and Pascal Rappard for excellent technical assistance.
Code Availability Not applicable.
Authors' Contributions NW and SR performed experiments, analyzed data, and wrote the manuscript. MP, DS, VF, and ES performed experiments and analyzed data. TT wrote the manuscript. GS performed experiments. HBD revised the manuscript. SCJ revised the manuscript and designed the study. All authors commented on previous versions of the manuscript; they read and approved the final version of the manuscript.
Funding Information Open Access funding enabled and organized by Projekt DEAL. This project was supported in part by Bayer Vital GmbH. MP was supported by Wilhelm und Günter Esser Stiftung.
Data Availability Data is included in the manuscript.

Compliance with Ethical Standards
Competing Interests The authors declare that they have no competing interests.
Ethics Approval The study was approved by the animal care committee of North Rhine-Westphalia (Germany; AZ 84-02.04.2016. A141).

Consent for Publication Not applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.