Activated microglia-induced neuroinflammatory cytokines lead to photoreceptor apoptosis in age-related macular degeneration-like mice model

Background Age-related macular degeneration (AMD) is mainly characterized by progressive deposits of drusen and photoreceptor apoptosis. Due to amyloid β (Aβ) is the main component of drusen, there is a great possibility that Aβ-induced activated microglia leads to inflammation, and plays a critical role in the pathogenesis of AMD. However, the relationship between activated microglia-mediated neuroinflammatory cytokines and photoreceptor apoptosis still remains unclarified. In this study, we demonstrated that subretinal injection of Aβ 1−42 induced the microglia activation and increased inflammatory cytokines, gave rise to photoreceptor apoptosis in mice. Our results were verified in vitro by co-culture of Aβ 1−42 activated primary microglia and photoreceptor cell line 661W, and we also performed that p38 MAPK signaling pathway was involved in Aβ 1−42 induced microglia activation and inflammatory cytokines release. Overall, our findings indicated that activated microglia-mediated neuroinflammatory cytokines contributed to photoreceptor apoptosis under the stimulation of Aβ 1−42 . Moreover, this study will provide a potential preventive and therapeutic approach for AMD treatment.

Inflammation response caused by microglia plays a critical role in the pathogenesis of AMD [14,15]. Microglial cells are the major resident immune cells in the central nervous system (CNS) and the retina, displaying a few of distinctive features and profoundly being influenced by the local microenvironment [16]. Under physiological conditions in retina, microglia are usually restricted in the inner retina, including the nerve fiber layer (NFL), ganglion cell layer (GCL), inner plexiform layer (IPL) and inner nuclear layer (INL), with a ramified morphology. Microglia could be activated once the retinal homeostasis was disturbed and migrate from the inner retina to the outer retina, including the outer nuclear layer (ONL), and even the subretinal space, with the characteristic amoeboid morphology, further releasing inflammatory cytokines to facilitate pathological destruction [17]. As a result, Aβ deposition in the retinas of patients with AMD may activate microglia and provoke its production of proinflammatory cytokines, which ultimately results in microglial-mediated neuroinflammation and retinal degeneration.
Photoreceptor degeneration is a remarkable pathological feature during AMD processing, while cell-autonomous signaling within photoreceptors are vital to AMD. However, few studies addressed the non-cell-autonomous manner between microglia-derived neuroinflammatory cytokine and photoreceptor death in the pathogenesis of AMD. In dry AMD, the activated microglia migrate to the ONL, where the cell body of photoreceptor localizes. Therefore, we wondered whether the non-cell-autonomous mechanism could mediate photoreceptor cell apoptosis by activated microglia.
In this study, we mimicked an inflammatory milieu in C57bl/6J mouse by subretinal injection of Aβ 1−42 and studied the relationship between microglia activation and photoreceptor apoptosis in retina. The effects were detected by using Aβ 1−42 treated primary microglial cells on photoreceptor cell line (661W). Meanwhile, the p38 mitogen-activated protein kinase (MAPK) pathway mediated neuroinflammatory cytokines release was also explored in microglia. Therefore, by modulating neuroinflammatory cytokines produced by microglial activation it represents potential as a possible preventive and therapeutic approach for AMD.

Retinal functional deterioration and photoreceptor degeneration after subretinal injection of Aβ 1−42 in mice
To testify whether visual function was impaired by Aβ 1−42 , ERG was performed in mice on day 1, day 3 and day 7 after subretinal injection of Aβ 1−42 . Compared with those in PBS-treated mice, the scotopic a-wave and b-wave ERG amplitudes were decreased in a time-dependent manner in Aβ 1−42 injected groups. The average amplitude of a-wave was decreased by 15.8% on day 1, 61.4% on day 3, and 67.0% on day 7; and the average amplitude of b-wave was decreased by 20.4% on day 1, 51.4% on day 3, and 64.1% on day 7, respectively ( Fig. 1A). The lower in ERG wave amplitudes generally implies the lasting deterioration in photoreceptor and RPE.
To detect the apoptosis of photoreceptor, TUNEL assay and the thickness of ONL measurement were performed in mice after subretinal injection of Aβ 1 − 42 on day 1, day 3 and day 7. One day after subretinal injection of Aβ 1 − 42 , isolated TUNEL-positive photoreceptors were detected in the ONL, there was no difference between Aβ 1 − 42 induced mice and PBS-injected control mice (Fig. 1B). At day 3, the amounts of photoreceptor apoptotic nuclei significantly increased in the ONL,and was about 4.5-fold of that than in control. At day 7, the number of photoreceptor apoptotic nuclei was fallen down to relatively normal level. The thickness of ONL was decreased by time during this process, and it was obviously decreased to about 70.1% on day 3 (Fig. 1B).
To further confirm the apoptosis after subretinal of Aβ 1 − 42 in mice, we performed Western blot to examine the protein levels for early indicators of apoptosis, including the total and cleaved caspase 3 (a substrate of activated caspase-3) protein expressions on day 3, and the data showed both total and cleaved caspase 3 protein levels were increased significantly in retina in Aβ 1 − 42 injected mice compared with that in control, i.e., about 1.9-and 1.6-fold for total and cleaved caspase 3 (Fig. 1C). Altogether, these data suggest that the subretinal injection of Aβ1-42 could induce to visual function deterioration and photoreceptor apoptosis aggravation, which led to AMD-like pathology in mice.

Aβ 1 − 42 induced microglia activation and pro-inflammatory cytokines expression in mice
To access whether Aβ 1 − 42 deposition could induce microglia activation after subretinal injection, we performed microglia immunostaining (Iba-1 positive) on both retinal cryosections and flat-mount. Compared with control, the expression of Iba-1positive cells was elevated abundantly at day 3 after subretinal injection. However, the activation of retinal microglia was down regulated in Aβ 1 − 42 induced AMD mice at day 7. As shown in Fig. 2A, from retinal cryosections, the amount of microglia was increased significantly on day 3. Importantly, while Iba1 positive cells were barely found in the outer retina in control mice, we could observe the distribution of microglia population partly migrated from the inner retina towards the ONL and subretinal space after subretinal injection of Aβ 1 − 42 on day 3 ( Fig. 2A). The results were confirmed on retinal flat-mounts (Fig. 2B), in the control mice, most of the Iba-1 positive cells had a ramified appearance and their processes were retracted through the retina. While in Aβ 1 − 42 induced mice, the appearance of Iba-1 positive cells were converted to amoeboid type with morphological features, including: with larger bodies (about 3.2 fold) and extension of processes across the ONL to intercalate to photoreceptor somata. (Fig. 2B-C).
To further characterize the activation of microglia, we detected the TSPO, CD86 and inflammatory factor in the retinas of Aβ 1−42 induced AMD model mice compared with control group on day 3 using Western blot. The protein levels of TSPO and CD86 were dramatically higher in the retina of Aβ 1−42 induced AMD mice, which were about 1.9-and 1.5-fold of that in the control (Fig. 2D), indicating the activation of microglia. Besides, the proinflammatory factors such as IL-1β and cyclooxygenase-2 (COX-2) were also detected, since these factors were reported to be produced by activated microglia. Compared with the controls, Aβ1-42 treated group resulted in significant increases in the IL-1β and COX-2 expressions on day 3, i.e., 1.5-and 2.5-fold of the control (Fig. 2D). The data showed that subretinal injection of Aβ 1 − 42 induced microglia activation and migration from inner to outer retina, accompanying with increased pro-inflammatory cytokines expression and photoreceptor apoptosis on day 3.

Aβ 1 − 42 induced primary microglial activation and pro-inflammatory cytokine expression in vitro
To further investigate whether Aβ 1 − 42 was the activator of microglia and modulated pro-inflammatory cytokines secretion, primary microglia cells were employed and treated with Aβ 1 − 42 in vitro. First, to test the cellular compatibility of the primary microglia treated with different concentrations (0.5 µmol/L to 10 µmol/L) of Aβ1-42 , we monitored the viability of the cultured primary microglia using the CCK-8 assay. The cell viability was dosedependently decreased by Aβ 1 − 42 , when the concentration of Aβ 1 − 42 reached 4 µmol/L, the number of viable cells declined significantly, which was about 59.7% of the control (Fig. 3A). Based on this result, 2 µmol/L Aβ 1−42 was selected as the optimal concentration in our subsequent experiments. The morphology of microglia displayed differently between control and Aβ 1−42 treated group. In control group, the cellular body of the basal homeostatic microglia was relatively smaller with ramified morphology after 24 hours' culture. Conversely, Aβ 1−42 -treated microglial cells converted to larger cellular body with amoeboid morphology (Fig. 3B). The morphological features suggested that primary microglia may be activated to adopt an amoeboid shape by Aβ 1−42 stimulation.
To further confirm our findings, we examined the protein expressions of TSPO. TSPO protein expression was increased within 24 hours of Aβ 1−42 treatment, which was about 2.3-fold of the control. To verify the inflammatory production for Aβ 1−42 stimulation on the primary microglia, western blot analysis indicated that the levels of IL-1β and COX-2 were also upregulated in microglial cells for 24 hours of Aβ 1−42 treatment in vitro, i.e., 1.7-and 3.2-fold of the control ( Fig. 3E-3F). All of these results suggested that microglial activation be induced by Aβ 1−42 stimulation, and activated microglia released proinflammatory cytokines in this process.

The inflammatory cytokines produced by Aβ 1−42 -treated primary microglial cells increased the apoptosis of 661W cells
Both in vivo and in vitro data demonstrated the activation of microglia, the increased inflammatory cytokines production as well as the apoptosis of photoreceptors in mice retina under the treatment of Aβ 1 − 42 . It deserved the exploration that whether the apoptosis of photoreceptors could be induced directly by the activated microglia derived inflammatory cytokines. To do this, we collected the conditioned medium (MCM) from the supernatant of the primary microglia cell treated with Aβ 1 − 42 , and then cultured 661W with it. We detected cellular morphology and analyze the apoptosis of 661W cells after 24 hours MCM culture, control medium was applied by the supernatant of microglial cells without adding to Aβ 1−42 . Given the direct impact of Aβ 1−42 on 661W culture, we also tested whether 661W apoptosis was changed in basal medium (BM) adding to 2 µmol/L Aβ 1−42 . On light microscope, 661W displayed similar-appearing morphological features in both CM and BM with Aβ 1−42 cultured 661W for 24 hours, however, the cell density of 661W incubated with MCM was significantly reduced from 18 hours to 24 hours (Fig. 4A). Western blot analysis showed after 24 hours cultured of 661W, the level of caspase 3 in 661W was significantly increased in MCM-treated group, compared with CM group or BM with Aβ 1−42 group ( Fig. 4B-C). The above data indicated that photoreceptor apoptosis might be caused by the activated microglia-derived inflammatory cytokines.

P38 MAPK signaling pathway was involved in microglia activation and photoreceptor apoptosis under Aβ 1−42 treatment
Since p38 MAPK signaling pathway was one of signaling pathway involved in the process of microglia, thus, we explored whether p38 MAPK signaling pathway exerts its function in photoreceptor apoptosis via Aβ 1 − 42 induced microglia activation in retina. Both total and phosphorylated p38 MAPK protein levels in retina were examined on day 3 after subretinal injection in mice with Aβ 1−42 or with PBS as control. As shown in Fig. 5A, the ratio of phosphorylated to total level of p38 was accelerated greatly in Aβ 1−42 treated group, which was about 2.3-fold of that in control. This result was also confirmed in vitro and the expression of p38 MAPK was about 3.1fold of that in the control (Fig. 5B). These effects implicated the possibility that p38 MAPK signaling pathway was associated with Aβ 1 − 42 induced microglia activation.
Taking into account that our preliminary data revealed that p38 MAPK protein expression was elevated in Aβ 1−42 induced microglial cells both in vivo and vitro, we wondered if p38 MAPK favors its downstream the inflammatory cytokines secretion in microglia, as inflammatory cytokines initiating the apoptosis of photoreceptor. To address this issue, we used p38 MAPK inhibitor to evaluate the role of p38 MAPK on microglial cell secretion of inflammatory cytokines with Q-PCR. 10 µ mol/L BIRB 796 was added to microglia cultures for 1 h before treatment with 2 µ mol/L Aβ 1 − 42 for 24 hours. As expected, the mRNA expressions of inflammatory cytokines in Aβ 1−42 treated group, such as IL-1β, COX-2, TNF-α and iNOS, were dramatically increased, which was about 5.1-, 6.1-, 3.2-, and 2.7-fold, respectively, of that in control. BIRB 796 largely abolished such increases and phosphorylated p38 MAPK protein level decreased in Aβ 1−42 induced microglia cultures correspondingly ( Fig. 5B-C). These results are in accordance with our previous hypothesis that the p38 MAPK signaling pathway may have a vital role in Aβ 1−42 induced microglial activation and its secretion of inflammatory cytokines, all of which may play an important role in photoreceptor apoptosis.

Discussion
In this study, we demonstrate that Aβ 1−42 induces the microglia activation and inflammatory cytokines secretion both in early AMD mice model and in primary microglial cells, leading to the disfunction of retina in vivo and 661W apoptosis in vitro. During these processes, the trend of photoreceptor apoptosis is consistent with the level of activated microglia and inflammatory cytokines. More importantly, we uncover that inflammatory cytokines produced by activated microglia initiate the photoreceptor apoptosis, otherwise than the direct toxic effect of Aβ 1−42 , few prior studies were focused on the non-cell-autonomous microenvironmental interactions between microglia and photoreceptor in AMD. Furthermore, we show that p38 MAPK signaling pathway is involved in modulating inflammatory cytokines secretion in Aβ 1−42 induced microglia, inhibition of the phosphorylated p38 MAPK effectively blocks inflammatory cytokines released by the activated microglia.
In clinic, AMD displays the accumulation of Aβ in the drusen beneath the RPE, Aβ-targeted agents and antiinflammatory therapies are still under evaluation for the treatment of AMD [18][19][20]. Accordingly, it is required to elucidate the mechanisms of AMD is of great importance to understand its pathogenesis and the effective treatments. To detect the harmful effect of Aβ 1−42 in retina, we examined ERG and photoreceptor apoptosis after Aβ 1−42 injection in mice model. Results showed that the amplitudes of both a-and b-waves were significantly reduced, while the number of apoptotic photoreceptors nuclei was increased significantly, accompanying the thinning of ONL on day 3 in Aβ 1−42 treated mice (Fig. 1). These ultrastructural alterations correspond to retinal functional changes in AMD process, as quantified by ERG declining and the progressive retinal degeneration.
Microglial activation is an important hallmark of neuroinflammatory diseases including AMD [14,21,22], as resident immunizing cells in retina, microglia sustain microenvironmental equilibrium and its dysfunction initiates a series of retinal degenerative performance [17,[23][24][25][26]. Therefore, considered that there was no wellestablished AMD animal models to mimic both the etiological and morphological properties of AMD [27,28], we focused on microglial activation by subretinal injection of Aβ 1−42 in mice and studied its indirect effect on photoreceptors by releasing the inflammatory cytokines. Our observations showed that Aβ 1−42 initiated microglial activation and inflammatory response after subretinal injection in mice model (Fig. 2). Activated microglia displayed the amoeboid morphology with enlarged cellular bodies and infiltrated the ONL from inner retina ( Fig. 2A-B), which indicate that activated microglia may contribute to retinal neuronal dysfunction and photoreceptor apoptosis in AMD.
Since neuroinflammation is a remarkable feature involved in the early phase of AMD procession [29], both in vivo and in vitro, the alterations in neuroinflammatory cytokine expression profiles were similar (Fig. 2D, Fig. 3C-F) in our study, the increased inflammatory cytokines like IL-1β and COX-2 were coincided with the published ones. Combining the ultrastructure and functional deterioration in our study, the Aβ 1−42 induced subretinal injection mice model may generally be used to as an early AMD mice model in further research. To the best of our knowledge, this is a novel model to mimic the pathogenesis of AMD and study the related mechanisms both in vivo and in vitro.
As the dry AMD progresses, visual loss is generally regarded as a consequence of photoreceptor degeneration, so it is critical to explore the mechanism of photoreceptor dysfunction or death in the early onset of AMD. As we known, microglial activation plays a vital role in the etiology and development of AMD,it is interesting to study the underlying relationship between microglia activation and photoreceptor apoptosis. As observed in Aβ 1−42 induced mice model, activated retinal microglia infiltrated the ONL and photoreceptor cells underwent apoptosis, and the quantitative analysis demonstrated photoreceptor apoptosis in the ONL performed consistent trend with microglia activation and migration via dynamic processes, peaking on day 3 and falling to a lower level on day 7. Furthermore, it showed the increased apoptosis of 661W cells under the treatment of MCM derived from activated microglia in vitro, while Aβ 1−42 had no direct function of inducing cell death in 661W cells basal medium ( Fig. 4A-C). Therefore, our study indicated that the cellular microenvironment produced by Aβ 1−42 activated microglia may have a critical function on photoreceptor apoptosis, which means Aβ 1−42 activated microglia and its related inflammatory cytokines may have a detrimental effect on photoreceptor. In other words, photoreceptor apoptosis was partly due to the increased inflammatory cytokines by activated microglia during the early onset of AMD.
It has been reported that there is an elevated phagocytic uptake of apoptotic photoreceptor fragments by activated microglia/macrophage whey acidic protein (AMWAP)-treated microglia [21], we infer that microglia may recognize and phagocytose apoptotic photoreceptor debris directly [30], however, the causal effect of neuroinflammatory cytokines released by activated microglia on photoreceptor death cannot be excluded. To prove this, activated derived MCM containing the inflammatory factors such as IL-1β, COX2, and etc. could induce photoreceptor apoptosis in our culture system, which confirmed the non-cell-autonomous interactions between the activated microglia and photoreceptors. To further clarify the pathways mediated inflammatory cytokines release by activated microglia, the p38 MAPK signaling pathway was studied. Accumulated studies reported that p38 MAPK signaling pathway was critical in many age-related degenerative diseases [31][32][33], and it was also a major pathway underlying activation of microglia required in AMD [34]. By attenuating the activation of p53 and p38 in a Müller cell ablation transgenic mice model, photoreceptor degeneration was prevented obviously [35], so we tested whether p38 MAPK expression was upregulated in Aβ 1−42 induced mice in vivo.
Furthermore, with BIRB 796, the inhibitor of p38 MAPK phosphorylation, we found that the mRNA levels of inflammatory cytokines, including IL-1β, COX-2, TNF-α and iNOS, were decreased significantly under the treatment of BIRB 796 in microglia (Fig. 5C). Our results were consistent with previous studies [36][37][38] that p38 MAPK signaling was vital in activated microglia to release inflammatory cytokines, these findings also indirectly confirmed our hypothesis that activated microglial may lead to the apoptosis of photoreceptor by releasing inflammatory cytokines, which provides a plausible explanation about the mechanism between Aβ 1−42 induced microglia and photoreceptor apoptosis in AMD (Fig. 6).
Interestingly, inflammatory cytokines produced by microglial activation leads to the apoptosis of photoreceptor.
Reviewing the related studies about the mechanisms of photoreceptor apoptosis in retinal degenerative diseases, RPE degeneration was the primary candidate due to its supportive functions [39][40][41], such as secretion of growth factors, phagocytosis of residual photoreceptor outer segments, spatial buffering of ions, etc. In this study, however, we uncovered activated microglia-derived neuroinflammatory cytokines also played a detrimental role in photoreceptor apoptosis during the early stage of AMD, which might add a new mechanism of photoreceptor apoptosis in the pathogenesis of dry AMD.

Conclusions
In summary, our study indicated the feasibility of creating a nurturing environment for AMD mice model with subretinal injection of Aβ 1−42 , which might be considered as a valuable research tool for the early stage of AMD.
Through modulating p38 MAPK signal pathway, activated microglia can proliferate, migrate and produce neuroinflammatory cytokines towards the impaired sites to initiate tissue damage. Non-cell-autonomous microglia-photoreceptor interactions may play a detrimental role in the AMD progression, the inflammatory cytokines released by active microglia could contribute to photoreceptor apoptosis. Hopefully, our preliminary results may represent a potential preventive and therapeutic approach for AMD treatment.

Animals And Subretinal Injection
C57bl/6J mice (Slaccas Laboratory Animal, Shanghai, China) were obtained at room temperature under a 12-hour light-dark cycle. All the animal studies were conducted according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The protocol was approved by the Committee on the Ethics of Animal Experiments of the tenth people's hospital affiliated Tongji University (Permit Number: SHDSYY-2020-2938). At least three individuals were performed for each experiment.
The subretinal injection was performed according to the previous studies [18]. Twelve-week-old C57bl/6J mice were anesthetized by an intraperitoneal (IP) injection of 4% chloral hydrate, and the pupils were dilated with 1% tropicamide (Alcon, USA). One eye was injected with 2 µL Aβ 1−42 solution (100 µmol/L), and the other eye was injected with 2 µL PBS (vehicle). Briefly, a 30-gauge needle puncture the sclera 2-mm behind the limbus to make a tunnel. Then 2 µL Aβ 1−42 solution or PBS was subretinal injected using a 33-gauge blunt needle (Hamilton Company, Reno, NV, USA).

Electroretinography (ERG)
RETIscan System (Roland Consult, Brandenburg, Germany) was used for scotopic ERG with signal stimulation, amplification and transportation. The protocol was followed as previous described [43], the ERG was carried out on day 1, day 3, and day 7 post-injection. Briefly, the mice were dark-adapted over 16 hours before performing ERG, mice were general anesthetized with an IP of 4% chloral hydrate under dim red illumination. Pupils of mice were dilated with 1% tropicamide for at least 20 minutes. After 0.5% Proparacaine hydrochloride (Alcon) applied for topical anesthesia of the cornea, contact lens electrodes with Gold wire loops were fixed on the surface of two corneas and the ground electrode was fixed in the tail, the parameters of scotopic ERG were measured. The amplitude of both a-wave and b-wave were recorded and analyzed using built-in software of ERG system.

TUNEL Assay
Apoptosis was detected with the kit of terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) according to the manufacturer instruction (YEASEN, China). Apoptotic cells were analyzed using Image J software.

Immunofluorescence
The cryosections of the retina and the primary microglia cells were permeabilized, blocked and incubated the primary antibody of Iba-1 (1:500, Wako) overnight at 4 °C. After washing thrice with PBS, the cryosections were incubated with Goat anti-Rabbit IgG conjugated to Alexa-568 (1:200, Invitrogen) for 1 hour at room temperature. After extensive wash, then the sections were counterstained with DAPI.
For the retinal whole-mount preparations, mice eyeballs were enucleated and fixed with 4% paraformaldehyde (PFA) for 2 hours, and then washed with PBS for three times. The retinas were isolated under dissecting microscope and were cut into four parts. Following the permeabilization and blocking step, the flat-mounts were incubated with the primary antibody of Iba-1 (1:500, Wako) overnight at 4 °C. After washing with PBS, the retina flat-mounts were incubated with the secondary antibody labeled with Alexa-568 (1:200, Invitrogen) for 2 hours and counterstained with DAPI.
The immunofluorescence of the cryosections, retinal flat-mounts and the primary cells were examined with NIKON A1 + confocal microscope (NIKON, Japan). The pictures were analyzed using ImageJ.

Primary Microglial Cell And 661W Cell Culture
Mouse primary microglial cells were isolated from the brains of C57bl/6J mice at postnatal day 4 or day 5 according to previously described [44]. After detaching from brain tissues, microglial cells were maintained in 75 cm 2 flasks containing DMEM supplemented with 10% fetal bovine serum (FBS, Gibco), 10 ng/mL GM-CSF, 50 U/mL penicillin and 50 mg/mL streptomycin in a 37•C incubator. The microglial cells culture medium was replaced every 3 days and was identified with Iba-1immunostaining (1:500, WAKO). The microglia were plated at 5 × 105 cells/mL in twelve-well plate for further study.

Cell Viability Assay
Cell counting kit-8 (CCK-8) assay (MCE, USA) was used to detect microglial cells viability according to manufacturer's instruction. Microglial cells were seeded in 96-well plates with a density of 1.0 × 104 cells per well. To determine the optimal dose of Aβ 1 − 42 on microglia, Aβ 1 − 42 with the final concentration of 0, 0.5, 1, 2, 4, 8,10 µmol/L was added into each well and cultured at 37 °C for 24 hours. After that, 10 µL CCK-8 reagent was added to each well. The microglial cells were incubated for 2 hours, and the absorbance value of each well was measured at 450 nm. According to the standard curve, the corresponding cell viability was calculated.

Microglial Conditioned Medium (MCM)
Mouse primary microglial cells were cultured with 2 µmol/L Aβ 1 − 42 for 24 hours, then the supernatant was collected as MCM, centrifuged to remove cellular debris (5 minutes, 3,000 g), MCM was applied to 661W medium, then 661W were plated at 3 × 105 cells/mL in twelve-well plate for 24 hours for Western blot detection. The supernatant of primary microglial cells without adding to Aβ 1 − 42 was served as control medium (CM).

Protein Kinase Inhibitor For P38 MAPK
The protein kinase inhibitor for p38 MAPK (BIRB 796) was obtained from Selleck Co. (USA), dissolved in DMSO and diluted with DMEM to a concentration of 10 µmol/L. Microglial cells were pretreated with inhibitor (10 µmol/L BIRB 796) 1 hour before with 2 µmol/L Aβ1-42 treatment in medium.

Reverse-transcription Quantitative Polymerase Chain Reaction (RT-qPCR)
Total RNA of primary microglia was isolated with RNAiso reagent (Yeasen Biotech, Shanghai, China) according to the manufacturer's instruction. Reverse transcribed using a PrimeScript® RT Reagent Kit (Yeasen Biotech, Shanghai, China). Quantitative PCR was performed with a CFX96 Real-Time PCR System (Bio-Rad, Hercules, USA) with the primers information provided in Table 1. Relative changes in gene expression were calculated using the 2 −ΔΔCt method. Table 1 The primer information used in this study.  Figure 1 Retinal functional deterioration and accelerated photoreceptor apoptosis in C57BL/6J mice treated with subretinal injection of Aβ1-42 or PBS. (a) Representative waveforms of the maximal ERG response and the amplitudes of a-wave and b-wave of the scotopic ERG responses on day 1, 3 and 7 after treatment of Aβ1-42. (b) Apoptosis detection with TUNEL and ONL thickness measurement in retinas with or without Aβ1-42treatment on day 1, 3 and 7. (c) Western blot detection of caspase 3 and cleaved caspase 3 proteins expression in mouse retina treated with Aβ1-42 on day3. C, control; ONL, outer plexiform layer. Histograms represent the mean ± S.E.M. N = 3, * P<0.05, ** P<0.01, *** P<0.001 via Student's t-tests and one-way ANOVA, Bonferroni's test was analyzed for post hoc comparisons.      Schematic illustration of p38 MAPK signaling pathway and its downstream neuroinflammatory cytokines in activated microglia leading to photoreceptor apoptosis.