Blockade of Adenosine A2A Receptor Protects Photoreceptors after Retinal Detachment by Inhibiting Inflammation and Oxidative Stress

Purpose Adenosine A2A receptor (A2AR) signaling is neuroprotective in some retinal damage models, but its role in neuronal survival during retinal detachment (RD) is unclear. We tested the hypothesis that A2AR antagonist ZM241385 would prevent photoreceptor apoptosis by inhibiting retinal inflammation and oxidative stress after RD. Methods The A2AR antagonist ZM241385 was delivered daily to C57BL/6J mice for three days at a dose (3 mg/kg, i.p.) starting 2 hours prior to creating RD. A2AR expression, microglia proliferation and reactivity, glial fibrillary acidic protein (GFAP) accumulation, IL-1β expression, and reactive oxygen species (ROS) production were evaluated with immunofluorescence. Photoreceptor TUNEL was analyzed. Results A2AR expression obviously increased and accumulated in microglia and Müller cells in the retinas after RD. The A2AR antagonist ZM241385 effectively inhibited retinal microglia proliferation and reactivity, decreased GFAP upregulation and proinflammatory cytokine IL-1β expression of Müller cells, and suppressed ROS overproduction, resulting in attenuation of photoreceptor apoptosis after RD. Conclusions The A2AR antagonist ZM241385 is an effective suppressor of microglia proliferation and reactivity, gliosis, neuroinflammation, oxidative stress, and photoreceptor apoptosis in a mouse model of RD. This suggests that A2AR blockade may be an important therapeutic strategy to protect photoreceptors in RD and other CNS diseases that share a common etiology.


Introduction
Photoreceptor apoptosis because of physical separation of the photoreceptors from the retinal pigment epithelium (RPE) results in visual loss in a number of retinal diseases, including macular degeneration [1], retinopathy of prematurity [2], diabetic retinopathy [3], and retinal detachment (RD) [4,5]. Photoreceptors are extremely vulnerable and undergo apoptosis in various types of RD including rhegmatogenous, tractional, and exudative. Although surgical treatment is routinely carried out to reattach the retina, visual acuity is not always restored because of RD-induced photoreceptor apoptosis [4,5]. Currently, there are no effective therapies for protecting photoreceptors after RD. Therefore, the development of neuroprotective agents for photoreceptors is essential to provide visual stability for RD patients undergoing surgical treatment.
In RD, multiple pathways are known to be involved in the RD-induced photoreceptor apoptosis, including the caspase pathway, autophagy, inflammation and gliosis, and reactive oxygen species (ROS) [6][7][8][9]. Reactive microglia cells are prevalent in the detached retina where they play a principal role in RD-induced photoreceptor apoptosis [10]. Müller cells are central to the pathophysiology of RD, being involved in gliosis, initiation of inflammatory cascades, and proliferative responses [11]. Increased levels of proinflammatory mediator, IL-1β, are found in the retina and aqueous humor of RD patients and experimental RD [8,[12][13][14]. Excessive generation of ROS and the consequent induction of oxidative stress are one of the critical factors that trigger cellular response to RD [15] and are also a major cytotoxic factor for photoreceptor apoptosis [9,16].
Adenosine is a neuromodulator in the central nervous system acting through the activation of four receptors, A 1 , A 2A , A 2B , and A 3 [17]. The adenosine A 2A receptor (A 2A R), a key molecule in the neural network, has been located on photoreceptors and RGCs of zebrafish [18], rat [19] and mouse [20], rat Müller cells [21], retinal pigment epithelium and choriocapillaris of rat [19], dog endothelium [22], and microglia of human [23] and rat [24]. A 2A R participates in inducing and maintaining microglial reactivity [25], NO synthase-II expression [26], cyclooxygenase-2 (COX-2), and the synthesis and release of proinflammatory cytokines [26,27] through the activation of its G-protein-coupled receptor [28]. Previous studies have demonstrated that pharmacological inhibition of A 2A R affords profound neuroprotection in animal models of several cerebral diseases [29,30]. Recent studies show that A 2A R antagonists prevent photoreceptors and RGC apoptosis by modulating the inflammation and oxidative stress in both age-related macular degeneration (AMD) [31] and glaucoma [32]. However, whether A 2A Rmediated neuroprotection is applicable to RD-induced photoreceptor apoptosis has not been determined.
In the current study, we have found that the expression of A 2A R is obviously increased in microglia and Müller cells in the detached retina in a time-dependent manner, which is accompanied by enhanced microglia and Müller cell reactivity. Meanwhile, we also provided evidence that a selective A 2A R antagonist, ZM241385, effectively protected photoreceptors with concomitant suppression of microglia activation, GFAP and proinflammatory cytokine IL-1β expression and ROS production after RD. Thus, in this study, we point out that A 2A R is a potential therapeutic target for preventing RD-induced photoreceptor apoptosis.

Materials and Methods
2.1. Experimental Animals. We followed the methods of Su et al. [33]. All animal experiments followed the guidelines of the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by Shanghai Jiao Tong University School of Medicine Animal Care and Use Committee. Male, 7-9 weeks old, C57BL/6J mice were allowed free access to water in a climate-controlled room with a 12 h light/12 h dark cycle.
2.2. Induction of RD. RD was induced as described previously [34]. Briefly, the mice were anesthetized with sodium pentobarbital (30 mg/kg, i.p.), and pupils were dilated with phenylephrine (5%) and tropicamide (0.5%). The temporal conjunctiva at the posterior limbus was incised. A 30-gauge needle (BD) was used to create a sclerotomy 1 mm posterior to the limbus. A corneal puncture was made with a 30-gauge needle to lower intraocular pressure. A 33-gauge needle connected to a Hamilton 10 μL syringe was inserted into the subretinal cavity. Then, 4 μL of 1% sodium hyaluronate (Provisc; Alcon) was injected, separating approximately 60% of the neurosensory retina from the underlying RPE. Eyes with subretinal hemorrhage or unsuccessful detachment were excluded from analysis. We detected retinas only within the area detached after RD by immunofluorescence staining.
2.3. Drug Administration. The mice were given the following treatment, as once daily i.p. injections of ZM241385 (S8105; Selleckchem, USA) at a dose of 3 mg/kg or vehicle for three days. The injected volumes did not exceed 0.2 mL per animal. The doses of the drugs were based on previous studies [35,36].
2.5. Oxidative Stress Assay. Superoxide dismutase (SOD) activity and malondialdehyde (MDA) were performed as previously reported [38]. Briefly, fresh retinal tissue was converted to 100 g/L of retina homogenates in a homogenizer filled. The homogenates were centrifuged at low temperature for 15 min at a speed of 3500 r/min. Proper amount of supernatant was given to perform tissue protein quantification. Levels of SOD and MDA were determined in accordance with the specifications of the SOD kit (Dojindo Molecular Technologies, Japan) and the MDA kit (Nanjing Jiancheng Bioengineering Institute, China). The protein concentration of the samples was determined using a BCA protein assay kit.
2.6. TUNEL Assay. The eyes were fixed in 4% paraformaldehyde overnight, then embedded in paraffin and sectioned at a thickness of 10 μm. The TUNEL assay was performed using the In Situ Cell Death Detection Kit (11684795910; Roche Diagnostics GmbH, Germany) according to the manufacturer's instructions on the frozen sections as previously described. The number of TUNEL + cells in the outer nuclear layer (ONL) was calculated in a masked fashion. The density of TUNEL + cells in the ONL was counted using ImageJ 1.48v 2 Oxidative Medicine and Cellular Longevity software. These measurements were carried out without knowledge of the treatment.

Western Blot Analysis.
Retinas from experimental eyes were dissected from the RPE-choroid. The samples were homogenized and lysed in RIPA buffer. And protein concentrations were calculated by the BCA protein assay kit (P0009; Beyotime, China). The samples were resolved on 10% SDS-PAGE gels and transferred onto PVDF membranes (EMD Millipore Corporation, US). The membranes were incubated overnight at 4°C with primary antibodies against A 2A R (05-717; Millipore, USA) and β-actin (CW0096S; CWBiotech, China). After being washed, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Inc., USA). The protein expression level was determined by densitometric analysis and normalized to the level of β-actin.
2.8. Quantitative RT-PCR Analysis. mRNA expression of A 2A R was analyzed by qRT-PCR as previously reported [37]. Briefly, after total RNA was extracted from the retinas using the TRIzol Reagent (Invitrogen, Carlsbad, CA, USA); total RNA was used to synthesize complementary DNA (cDNA) with a reverse transcription kit (Takara, Japan).
The reaction system was composed of specific primers, cDNA, and the SYBR Green qPCR Mix (Takara, Japan). Primers were synthesized by a private company (Sangon Biotech Co. Ltd., China). β-Actin served as the reference gene, and the expression of target genes was calculated as 2 -△△Ct . Primer sequences used were designed as follows: murine A 2A R forward, 5 ′ -AGAGCAAGAGGCAGGTATCTC-3 ′ and reverse, 5 ′ -CCCAAAGGCTTTCTCACGGA-3 ′ ; 2.9. Statistical Analysis. The data are presented as the mean ± standard deviation (mean ± SD). The differences among the groups were analyzed by Student's t-test or one-way ANOVA according to the normal distribution. All statistical analysis was performed using GraphPad software (Prism 8; GraphPad Software, Inc.). Values of P < 0:05 were considered statistically significant.

A 2A R Expression Is Increased after RD.
To investigate the expression of A 2A R in RD, we collected the mouse retinas from 12 h to day 7 after induction of RD (Figure 1(a)) and examined the mRNA and protein expression of A 2A R using quantitative real-time PCR and Western blotting, respectively. Compared to the control, the mRNA expression of A 2A R in retinas after RD increased. RT-PCR analysis revealed that there was a more than 6-fold induction of A 2A R mRNA expression in the RD model at day 1 ( Figure 1(b)). Meanwhile, immunoblots also showed high expression of A 2A R protein in retinas after RD at 12 h, day 1, and day 3 ( Figure 1(c)). Furthermore, to detect the distribution of A 2A R in detached retina, the immunofluorescence staining was performed ( Figure 1(d)). A 2A R immunoreactivity was increased with time and very strong at day 1 in the detached retina, whereas there was minimal staining in the control. The A 2A R staining was obviously detected in the (a)

A 2A R Staining Is Colocalized with Microglia and Müller
Cells after RD. To verify the identity of the A 2A R + cells, double immunofluorescence was performed from 12 h to day 1 after RD by using antibodies against A 2A R and Iba-1, a marker of microglia cell (Figure 2(a)). Microglia was significantly increased in retina after RD, which is consistent with previous research. The A 2A R and Iba-1 signals colocalized in microglia cells in the IPL, the inner nuclear layer (INL), the OPL, and the ONL (Figure 2(a), arrows). To verify the identity of the A 2A R + cells, double immunofluorescence was also performed after RD using antibodies against A 2A R and GFAP, a Müller cell marker (Figure 2(b)). After RD, simultaneously upregulation of GFAP was detected with the increment of A 2A R. Compared to the restricted expression of GFAP that was observed in the nerve fiber layer of the retina from the control eyes, the GFAP immunoreactivity was marked and extended across the entire neural retina to the ONL after RD. The A 2A R and GFAP signals colocalized in Müller cells in the GCL and IPL (Figure 2(b), arrows). These data demonstrate that A 2A R expression dramatically increased in the microglia and Müller cells in the retina after RD.

A 2A R Blockade Inhibits Microglia Proliferation and
Reactivity after RD. Microglia cells, the resident tissue macrophages of the retina, play a critical role in damage processes and have both neuroprotective and neurotoxic effects during retinal damage. To investigate the effects of A 2A R on microglial response after RD, the immunofluorescence colocalization of Iba-1 (labeling both resting and active microglial cells) and CD11b (for active microglial cells staining) [39] was assessed in detached retina at day 3 (Figure 3(a)). Retinal Iba-1 + cell counts were significantly increased after RD, whereas Iba-1 + cell counts were strongly suppressed by ZM241385 compared with the control vehicle treatment after RD (Figure 3(c)). Significant accumulation of Iba-1 intensity was detected in detached retina compared to controls. However, the administration of ZM241385 obviously decreased Iba-1 intensity after RD (Figure 3(d)). Additionally, this blockade of microglial reactivity is illustrated by the colocalization of Iba-1 and CD11b. Iba-1 + CD11b + cell counts were dramatically inhibited by ZM241385 after RD, compared with the control vehicle treatment (Figure 3(e)).
In order to better detect microglia reactivity, double labeling was also performed using primary antibodies to Iba-1 and major histocompatibility complex class II (MHC-II, highly expressed in reactive microglia) (Figure 3(b)).

Oxidative Medicine and Cellular Longevity
Iba-1 + MHC-II + cells were considered reactive microglia. As expected, microglia reactivity was obviously increased in the retina after RD when compared with the retinas of control. ZM146385 administration showed a significant decrease in the percentage of reactive microglia cells compared to vehicle after RD, whereas the administration of ZM146385 to control animals did not change the number of reactive microglia (Figure 3(f)). These data suggest that A 2A R is critical for the   5 Oxidative Medicine and Cellular Longevity microglia proliferation and reactivity after RD and for the subsequent structural and functional disruption of these retinal layers.

A 2A R Blockade Reduces Reactive Gliosis and
Inflammatory Response after RD. Activation of gliosis represents Müller glial remodeling in response to RD-induced retinal damage and contributes to tissue inflammation. Additionally, IL-1β, an important inflammatory cytokine, has previously been reported to contribute to the pathogenesis of photoreceptor apoptosis after RD [40]. Moreover, the blockade of A 2A R has been shown to effectively prevent inflammatory responses within various injury models [41]. To investigate whether the A 2A R blockade contributes to inhibiting reactive gliosis and inflammatory responses after RD, we evaluated the role of A 2A R in the RD-induced GFAP and IL-1β expression. After RD, the level of GFAP was obviously upregulated, whereas the increased GFAP expression in detached retina was dramatically ameliorated by the addition of ZM241385 (Figures 4(a) and 4(b)). The expression of IL-1β, colocalized in Müller cells with GFAP signal, was detected in the GCL, the IPL, and the OPL at day 3 after RD ( Figure 4(a)). The ZM146385 administration obviously suppressed the upregulation of IL-1β expression in Müller cells induced by RD (Figure 4(c)). Taken together, these results indicate that the A 2A R blockade can inhibit RDinduced reactive gliosis and inflammatory response in Müller cells.

A 2A R Blockade Suppresses the Oxidative Stress after RD.
Silencing of the caffeine-antagonized A 2A R significantly reduced ROS production in THP-1 macrophages [42] and on UV-induced skin damage in mice [43]. We therefore hypothesized that A 2A R blockade might protect neuronal cells from RD by reducing oxidative stress. To test the hypothesis, oxidative stress was measured in detached retina after ZM241385 administration. Immunofluorescence results showed that the upregulation of ROS in the ONL was inhibited dramatically by ZM241385 at day 3 after RD ( Figure 5(a)). Moreover, we further observed that ZM241385 markedly restrained RD-induced MDA ( Figure 5(b)), whereas the decreased activity of SOD was   Oxidative Medicine and Cellular Longevity largely restored ( Figure 5(c)). These results demonstrate that A 2A R blockade effectively decreases RD-induced overproduction of oxidative stress.

A 2A R Blockade Prevents Photoreceptor Apoptosis after RD.
Photoreceptor apoptosis was quantified at day 3 after RD which is the peak time point [44,45]. To investigate whether A 2A R is involved in RD-induced photoreceptor apoptosis, the TUNEL assay was performed among different groups (Figure 6(a)). In the absence of RD, the general appearance of the retina was similar at vehicle and ZM241385 groups. ZM241385 almost completely suppressed the appearance of TUNEL + cells in the ONL after RD, whereas the control vehicle treatment had no effect (Figure 6(b)).

Discussion
Increased A 2A R expression has been reported in several retinal disease models, including oxygen-induced retinopathy [22], diabetic retinopathy [46], glaucoma [41], and light-induced retinal degeneration [47]. Therefore, A 2A R may be a critical factor for the inflammatory response during various acute and chronic retinal diseases. In the light-induced retinal degeneration model, A 2A R upregulation is detected in the GCL and INL, coinciding with the massive apoptosis of photoreceptors [47]. In the current study, A There is a controversy on the effects mediated by A 2A R in pathological conditions, since A 2A R activation in peripheral immune cells is anti-inflammatory [48,49], and in chronic conditions of the central nervous system, the blockade of the A 2A R confers protection [28]. Recently, A 2A R blockade has been shown to selectively reduce avascular areas and neovascularization, with the decreased cellular apoptosis and proliferation, and increased astrocyte and tip cell functions in OIR [50]. And the administration of A 2A R blockade, ZM241385, has been illustrated to reduce microglia activation and decrease the proinflammatory factor expression to improve RGC survival in experimental glaucoma [41]. Moreover, Boia et al. [24] have demonstrated that treatment with A 2A R antagonist KW6002 and caffeine intake could obviously inhibit microglia reactivity and effectively protect retina against transient ischemic damage. A 2A R selective antagonist SCH58261 has been revealed to decrease GFAP expression in rat brain astrocyte cell line with ischemia-like injury [51]. However, whether A 2A R blockade could play a neuronal protective role in RD is still unclear. In the present study, we observed that ZM241385, a selective A 2A R antagonist, inhibited microglia reactivity after RD, accompanied by reduced proliferation of microglia. We found that ZM241385 also decreased GFAP expression and alleviated expression of inflammatory cytokine IL-1β. Furthermore, our results revealed that ZM241385 reduced obviously ROS production and attenuated the increase in MDA concentration after RD, while SOD activity increased in detached retinas. This was evident in a reduction of oxidative stress induced by RD, through administration of A 2A R blockade.
Mounting evidence indicates that the microglia activation may contribute to the secondary injury to neurons and result in the chronic neuroinflammation [52]. The activation of microglia leads to the excessive releasement of proinflammatory cytokines, such as IL-1β, IL-6, and TNF-α [53]. Although the release of these proinflammatory cytokines is intended to protect the central nervous system tissue from further damage, they can impair simultaneously neurons [54]. Interestingly, the blockade of A 2A R affords neuroprotection in several models of neurodegeneration, including in the retina [28,41,[55][56][57]. One of the mechanisms that explains the protective properties of A 2A R antagonists is the control of microglia-mediated neuroinflammation [28,56]. It has also been illustrated that A 2A R blockade confers neuroprotection by controlling microglia reactivity in vivo [31] and in glaucoma [58]. Likewise, we found that ZM241385 could dramatically inhibit the increase of Iba-1 intensity and count of activated microglial cells after RD in the current study. Our finding that A 2A R blockade could alter microglia reactivity is consistent with those in other experimental models of retinal damage [32].
GFAP is a marker of gliosis and is increased in reactive Müller glia cells in various retinal disorders. Increasing studies have reported that GFAP levels in the central nervous system are elevated under pathological conditions and that this process plays an important role in neural injury [59,60]. However, the pathological mechanisms of Müller activation and GFAP upregulation after RD are largely unknown. As a consequence of RD, a "mechanical" damage to the retina, GFAP expression increases obviously in Müller cells. In addition, the hypertrophy of Müller both within the retina and on the photoreceptor is accompanied with GFAP upregulation [61]. Previous studies have shown that the deletion of GFAP can prevent RD-induced gliosis and rescue photoreceptor degeneration, which highlights the key role of Müller cells in regulating retinal damage [62]. In this study, we found that ZM241385 effectively suppressed the upregulation of GFAP expression in Müller cells. The results suggest that A 2A R blockade could prevent Müller activation after RD. However, improvement is needed in order to assess the effect of A 2A R regulation in a therapeutic setting using a modified model of A 2A R antagonist administration after the induction of RD.
IL-1β is a critical inflammatory cytokine involved in various retinal diseases [58,[63][64][65]. Previous studies have suggested IL-1β plays an important role in inducing photoreceptor apoptosis, using models of retinopathy of prematurity [66] and age-related macular degeneration [40,67]. Kataoka et al. [40] showed that IL-1β partially contributed to photoreceptor cell apoptosis after RD utilizing caspase-1 inhibitor or IL-1β neutralizing antibody, which is in contrast to previous studies that have shown that IL-1β administration into the subretinal space does not increase photoreceptor apoptosis during RD [68]. Interestingly, Zhao and colleagues reported that caffeine could inhibit the increased production of IL-1β by suppressing A 2A R signaling to prevent LPSinduced THP-1 macrophage activation [42]. In this study, we found that IL-1β expression was significantly increased 8 Oxidative Medicine and Cellular Longevity and located in the Müller cells after RD as it was previously observed [40,68], whereas the upregulation of IL-1β dramatically decreased after ZM241385 administration. The results suggest that A 2A R blockade could effectively prevent RDinduced upregulation of IL-1β expression in Müller cells. The cellular and molecular mechanisms inducing photoreceptor cell death have been partially revealed. Recent studies have shown that RIP kinase-mediated necrotic signaling [69] and FAS-mediated apoptosis pathway [70] contribute to photoreceptor death after RD. However, accumulating evidence suggests that the rapid increase of oxidative stress is currently considered to be a critical event for irreversible cellular damage in RD [14]. Previous studies have demonstrated an increased generation of ROS after RD, whereas photoreceptor cell death can be prevented after RD by suppressing ROS [71,72]. The overproduction of ROS is known to interact with various inflammatory cytokines, including IL-1β, TNF-α, and CCL2, which suggests an important role for ROS in mediating the stress response [73]. Concurrent with this, previous studies have demonstrated that microglia A 2A R blockade suppresses elevated pressure-induced oxidative stress in retina [32]. Furthermore, modulation of α-adrenoceptor signaling protects photoreceptors from apoptosis after RD by inhibiting ROS production [14]. In our work, we showed that ZM241385 was able to prevent photoreceptor loss from ROS overproduction triggered after RD, further reinforcing its role in controlling retinal neuroinflammation. We first found that MDA concentration increased in the detached retina, while the SOD activity decreased significantly after RD. Secondly, we observed that accompanied by a decrease in MDA and an increase in SOD, ZM241385 inhibited photoreceptor apoptosis after RD. This suggests that oxidative stress plays a key role in nerve injury after RD. Therefore, pharmacological therapies targeting oxidative stress may be critical for inhibiting RD-induced photoreceptor apoptosis.

Conclusions
In conclusion, we demonstrate that A 2A R expression significantly upregulate and its colocalization with microglia and Müller cells in the retina after RD. In addition, A 2A R blockade could provide effective protection against photoreceptor apoptosis in a mouse model of experimental RD for the first time. Meanwhile, we further demonstrate that the neuroprotective effects of the A 2A R antagonist ZM241385 is related to the amelioration of RD-induced environmental stress that leads to microglia proliferation and reactivity, reactive gliosis, upregulation of proinflammatory cytokine, and activation of oxidative stress. Our results suggest that A 2A R blockade may present novel therapeutic targets for strategies aimed at preserving visual acuity in patients with RD.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare no conflict of interest.

Authors' Contributions
Xi Shen and Sha Gao conceived and designed the experiments. Sha Gao, Na Li, and Yanuo Wang performed the experiments. Sha Gao and Xi Shen analyzed the data. Xi Shen and Yisheng Zhong contributed reagents/materials/analysis tools. Sha Gao wrote the paper. Sha Gao and Na Li contributed equally to this work.