Experimental Eye Research 135 (2015) 47e58

Contents lists available at ScienceDirect

Experimental Eye Research journal homepage: www.elsevier.com/locate/yexer

Neuroprotective effect of memantine on the retinal ganglion cells of APPswe/PS1DE9 mice and its immunomodulatory mechanisms Lixiong Gao a, b, c, Xi Chen a, b, d, e, Yongping Tang a, b, Jinghui Zhao c, Qiyou Li a, b, Xiaotang Fan c, Haiwei Xu a, b, *, Zheng Qin Yin a, b, * a

Southwest Hospital/Southwest Eye Hospital, Third Military Medical University, Chongqing 400038, China Key Lab of Visual Damage and Regeneration & Restoration of Chongqing, Chongqing 400038, China Department of Developmental Neuropsychology, School of Psychology, Third Military Medical University, Chongqing 400038, China d School of Medicine, Nankai University, Tianjin 300071, China e Department of Ophthalmology, Chinese People's Liberation Army General Hospital, Beijing 100853, China b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 December 2014 Received in revised form 20 March 2015 Accepted in revised form 21 April 2015 Available online 23 April 2015

Besides the cognitive impairment and degeneration in the brain, vision dysfunction and retina damage are always prevalent in patients with Alzheimer's disease (AD). The uncompetitive antagonist of the Nmethyl-D-aspartate receptor, memantine (MEM), has been proven to improve the cognition of patients with AD. However, limited information exists regarding the mechanism of neurodegeneration and the possible neuroprotective mechanisms of MEM on the retinas of patients with AD. In the present study, by using APPswe/PS1DE9 double transgenic (dtg) mice, we found that MEM rescued the loss of retinal ganglion cells (RGCs), as well as improved visual impairments, including improving the P50 component in pattern electroretinograms and the latency delay of the P2 component in flash visual evoked potentials of APPswe/PS1DE9 dtg mice. The activated microglia in the retinas of APPswe/PS1DE9 dtg mice were also inhibited by MEM. Additionally, the level of glutamine synthetase expressed by Müller cells within the RGC layer was upregulated in APPswe/PS1DE9 dtg mice, which was inhibited by MEM. Simultaneously, MEM also reduced the apoptosis of choline acetyl transferase-immunoreactive cholinergic amacrine cells within the RGC layer of AD mice. Moreover, the phosphorylation level of extracellular regulated protein kinases 1 and 2 was increased in APPswe/PS1DE9 dtg mice, which was blocked by MEM treatment. These findings suggest that MEM protects RGCs in the retinas of APPswe/PS1DE9 dtg mice by modulating the immune response of microglia and the adapted response of Müller cells, making MEM a potential ophthalmic treatment alternative in patients with AD. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Alzheimer's disease Memantine Retinal ganglion cell Müller cell Microglia

1. Background Alzheimer's disease (AD), the most prevalent aging dementia, is characterized by progressive impairments of cognitive function and memory (Fan et al., 2014). Visual abnormalities and retinal

degeneration have also been reported in patients with AD (Bayer et al., 2002; Guo et al., 2010), and are characterized by the loss of retinal ganglion cells (RGCs) (Blanks et al., 1996), reduced thickness of the retinal nerve fiber layer (NFL) (Kesler et al., 2011), decreases in the amplitude of the P50 component in pattern

List of abbreviations: Ab, amyloid-b; AD, Alzheimer's disease; BSA, bovine serum albumin; ChAT, choline acetyl transferase; dtg, double transgenic; ERK1/2, extracellular signal-regulated protein kinases 1 and 2; FAD, familial Alzheimer's disease; fVEP, flash visual evoked potential; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GS, glutamine synthetase; HE, hematoxylin and eosin; Iba1, ionized calcium binding adaptor molecule 1; ILM, inner limiting membrane; INL, inner nuclear layer; i.p., intraperitoneal injection; IPL, inner plexiform layer; IR, immunoreactive; MEM, memantine; NeuN, neuronal nuclei; NFL, nerve fiber layer; NIH, National Institutes of Health; ntg, nontransgenic; OD, optical density; ONL, outer nuclear layer; OPL, outer plexiform layer; PBS, phosphate buffer solution; pERG, pattern electroretinaogram; RGCs, retinal ganglion cells; RGCL, retinal ganglion cell layer; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling; WT, wild type. * Corresponding authors. Southwest Eye Hospital, Southwest Hospital, Third Military Medical University, Chongqing 400038, China. E-mail addresses: [email protected] (L. Gao), [email protected] (X. Chen), [email protected] (Y. Tang), [email protected] (J. Zhao), [email protected] (Q. Li), [email protected] (X. Fan), [email protected] (H. Xu), [email protected] (Z.Q. Yin). http://dx.doi.org/10.1016/j.exer.2015.04.013 0014-4835/© 2015 Elsevier Ltd. All rights reserved.

48

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

electroretinograms (pERGs) (Krasodomska et al., 2010), and increases in the latency of the P2 component in flash visual evoked potentials (fVEPs) (Coburn et al., 2005). The APPswe/PS1DE9 double transgenic (dtg) mouse, a classic mouse model of AD, also shows pathological changes in the retina, including deposition of amyloid-b (Ab) within the NFL, RGC layer (RGCL), inner nuclear layer (INL) (Dutescu et al., 2009; Ning et al., 2008), and inner plexiform layer (IPL) (Perez et al., 2009). Particularly, the loss of RGCs was also found in APPswe/PS1DE9 dtg mice (Ning et al., 2008). As an uncompetitive N-methyl-D-aspartate receptor antagonist, memantine (MEM) improves the symptoms of patients with AD in many ways including preventing apoptosis from glutamate excitatory toxicity (Bormann, 1989; Hare and Wheeler, 2009) and decreasing the deposition of Ab (Alley et al., 2010). Nonetheless, MEM could also enhance neurotrophic factors release from the astroglia, and had novel anti-inflammatory effects including preventing microglial activation and protecting radial glial-like cells in the hippocampus (Sun et al., 2014; Wu et al., 2009). In the retina, MEM is able to protect RGC from apoptosis in glaucoma (Lipton, 2003). However, the exact effect of MEM within the retina of APPswe/PS1DE9 dtg mice is not clear. Müller cells in the retina support the function of neurons by maintaining the balance of water transportation and ionic metabolism (Francke et al., 1997; Newman and Reichenbach, 1996). After retinal injury, Müller cells are over-activated, which eventually results in the formation of glial scars (Bringmann et al., 2006; Bringmann and Wiedemann, 2012). Microglia are the innate immune cells of the retina (Karlstetter et al., 2010) and play a key role in clearing the necrotic tissue in the retina when activated (Ling and Wong, 1993; Liu et al., 2012). Following microglia activation, an adaptive Müller cell response can be found in the retina, which results in the interaction between microglia and Müller cells and might participate in the inflammation of the retina (Wang et al., 2011). During retinal degeneration, microglia are activated via the extracellular signal-regulated protein kinases 1 and 2 (ERK1/2) pathway hereafter releasing inflammatory factors, which cause the apoptosis of neurons and the proliferation of Müller cells. These processes may achieve through the activation of nuclear factorkappa B and cell adhesion molecules (Dong et al., 2014). In the present study, we treated APPswe/PS1DE9 dtg mice with MEM to explore its abilities to rescue RGCs and to improve retinal function. The characteristic changes in microglia, Müller cells, and cholinergic amacrine cells in the retinas of APPswe/PS1DE9 dtg mice and the effects of MEM were also investigated. Furthermore, as the ERK1/2 signaling pathway has been shown to play a critical role in mediating inflammation (Qian et al., 2008) and apoptosis (Guise et al., 2001), the potential involvement of ERK in the protective effects of MEM on the retinas of APPswe/PS1DE9 dtg mice was explored. 2. Materials and methods 2.1. Transgenic mice APPswe/PS1DE9 dtg mice co-expressing familial AD (FAD) mutant human PS1DE9 and a chimeric mouse-human APP695 harboring a human Ab domain and mutations (K595N, M596L) linked to Swedish FAD pedigrees (APPswe) have been described previously (Borchelt et al., 1996a, 1996b; Lee et al., 1997; Li et al., 2008). APPswe/PS1DE9 dtg mice and their age-matched nontransgenic (ntg) littermates at 13 months of age were used. All animals were raised in a specific-pathogen-free room and maintained on a 12-h light/dark cycle. Animal care and

procedures were conducted according to the National Institutes of Health (NIH) guidelines for the Care and Use of laboratory animals and the Third Military Medical University policies for animal use. 2.2. MEM administration Drug treatment had been previously established (Sun et al., 2014). In brief, MEM (SigmaeAldrich, USA) was dissolved in 0.9% saline (2 mg/mL). Six transgenic mice and 6 ntg mice received an MEM intraperitoneal injection (i.p.) at a dosage of 10 mg/kg per injection daily for 8 days, and are referred to as the “AD þ MEM” and “WT þ MEM” groups, respectively. The remaining 12 mice received identical doses of saline via i.p. for 8 days (once per day), and are referred to as the “AD þ Saline” and “WT þ Saline” groups for the transgenic mice and ntg mice, respectively. 2.3. pERGs and fVEPs Visual electrophysiological tests were performed 2 weeks before the animals were sacrificed. A superior pERG recording protocol was established previously (Porciatti et al., 2007; Tao et al., 2013). Briefly, the mice were anesthetized by intramuscular injections of amine (100 mg/kg) and xylazine (12 mg/kg). A heating pad was positioned under the mice to keep their body temperature at a constant 37.0  C. One drop of tropicamide and phenylephrine were used to dilate each pupil. Gold wire loops were used as corneal pERG electrodes. Then, 0.9% saline was frequently applied on the cornea to prevent dehydration and increase the conductibility. Reference and ground electrodes were stainless steel needles respectively inserted into the scleral conjunctiva around the equator of the eyes and tail. Using the Retiscan system (Roland, Germany), pERGs were recorded by pattern stimuli, which were displayed at 1 Hz in the spatial phase on the video monitor. An fVEP recording protocol was also established in an optimized way (Ridder and Nusinowitz, 2006). Differences between the pERGs and fVEPs included the placement of the recording electrode and the stimuli. The fVEP recording electrode involved stainless steel screws inserted into the skull contralateral to the stimulated eye. The position was 2 mm lateral to the lambda suture. The stimuli were diffuse bright flashes on a rod-saturating background as described. 2.4. Tissue preparations and HE staining After being anesthetized by 1% pentobarbital (150 mL/kg), mice were perfused with 0.9% sodium chloride, followed by 4% paraformaldehyde in 0.01 M phosphate buffer solution (PBS) via the circulation system. Both eyes were removed rapidly, placed in 4% paraformaldehyde at room temperature for 30-min fixation. Then, anterior segments were removed microscopically, placed eyes back into the fixation fluid overnight at 4  C. The fixed eyes were dehydrated with a graded series of ethanol and chloroform, and then embedded in paraffin wax. Using a paraffin microtome, 5-mm thick paraffin sections that crossed the optic disc were collected. After drying for 2 days at 37  C, the sections were deparaffinized in xylene and rehydrated with a graded series of ethanol, and then stained with hematoxylin and eosin (HE) for 2 min and 20 s each. 2.5. Immunofluorescence and TUNEL staining The immunofluorescence of sections was described previously (Yang et al., 2014). After being deparaffinized and rehydrated, sections were incubating in 0.3% Triton X-100 and 3% bovine serum

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

49

Fig. 1. HE staining showed the entire retinal structure of the 13-month-old APPswe/PS1DE9 transgenic mouse. AeD: HE staining in the (A) WT þ Saline, (B) WT þ MEM, (C) AD þ Saline, and (D) AD þ MEM. E, F: The comparison of the thickness in each layer among the four groups stained by HE. E: Thickness of the nuclear layers including RGCL, INL and ONL. F: Thickness of plexiform layer including IPL and OPL. The results showed that there were no significant differences among the four groups. Scale bar: AeD 100 mm. HE: hematoxylin and eosin; RGCL: retinal ganglion cell layer; IPL: inner plexiform layer; INL: inner nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear layer; WT: wild type; S: saline; MEM: memantine; AD: Alzheimer's disease.

albumin (BSA) at room temperature for 30 min. Then sections were then incubated with the primary antibody anti-neuronal nuclei (NeuN) (1:500; Millipore, Germany), anti-glutamine synthetase (GS) (1:1000; Millipore), anti-choline acetyl transferase (ChAT) (1:500; Millipore), or anti-Ionized calcium binding adaptor molecule 1 (Iba1) (1:500; Wako, Japan) in 1% BSA at 4  C overnight. Secondary antibodies, cy3-or 488-conjugated (Jackson ImmunoResearch, West Grove, PA, USA), were then implemented (1:200; 3 h). Before examination with a confocal laser scanning microscope (Leica, Germany), sections were counterstained with 40 , 6diamidino-2-phenylindole (DAPI) (SigmaeAldrich, St. Louis, MO, USA). For testing the apoptosis in NeuN-staining and ChAT-staining, terminal deoxynucleotidyl transferased UTP nick end labeling (TUNEL) assays were performed according to the instructions of the manufacturer with the In Situ Cell Death Detection Kit (Fluorescein

or TMR, Roche Diagnostics, Germany). In brief, sections were incubated in a reaction mixture (Enzyme Solution 1 mL þ Label Solution 9 mL) at 37  C for 1 h after the secondary antibody implementation. Following DAPI counterstaining and PBS washing, sections were viewed and photographed under a Zeiss (Oberkochen, Germany) Axivert microscope equipped with a Zeiss Axio Cam digital color camera connected to the Zeiss AxioVision 3.0 system. 2.6. Western blot analysis Retinas were isolated from enucleated eyes on ice and homogenized in ice-cold RIPA Lysis buffer (Beyotime, Shanghai, China) containing proteinase inhibitor. After 5 min, 10,000 rpm/ min centrifugation (Thermo, USA) at 4  C was performed and the protein concentration was determined via the BAC test (Beyotime

50

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

Fig. 2. MEM reduced retinal ganglion cell loss through anti-apoptosis in 13-month-old APPswe/PS1DE9 transgenic mice. AeD: Whole retinal montage of neuronal nucleiimmunoreactive (NeuN-IR) ganglion cells in the (A) WT þ Saline, (B) WT þ MEM, (C) AD þ Saline, and (D) AD þ MEM groups. EeH: NeuN and TUNEL double staining. Red: NeuN-positive RGCs; green: TUNEL-staining apoptotic cells. White oblique arrows: NeuN-positive RGCs; downward vertical white arrows: RGCs experiencing apoptosis. I: The statistic analysis of the whole retinal mounts of RGCs as well as ratio of apoptotic cells. The results showed that the amount of RGCs in the transgenic mice was decreased compared to the number in the WT groups. After treatment with MEM, the RGC loss decreased. *P < 0.05; **P < 0.01; Scale bar: AeD: 1000 mm; EeH: 50 mm. WT: wild type; MEM: memantine; AD: Alzheimer's disease; TUNEL: terminal deoxynucleotidyl transferase dUTP nick end labeling. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Institute of Biotechnology, Shanghai, China). As we previously reported (Yang et al., 2014), proteins were separated by 12% sodium dodecyl sulfate polyacrylamide gel and then transferred onto polyvinylidene fluoride membranes. The membranes were then blocked by 5% fat-free milk for 2 h at 37  C, followed by incubation with anti-ChAT (1:1000; Millipore), anti-GS (1:1000, Millipore), anti-b-Actin (1:1000; Santa Cruz Biotechnology, CA), anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (1:1000; CWbio, China), anti-phospho-ERK1/2 (1:1000; Cell Signaling Technology, USA), and anti-ERK1/2 (1:1000; Cell Signaling Technology) overnight at 4  C. The next day, the membranes were incubated with peroxidase-conjugated immunoglobulin G (1:2000; Santa Cruz Biotechnology). Finally, the membranes were scanned using the Odyssey infrared imaging system with the Odyssey Application software V1.2.15 (LI-COR, USA) for bands. Using b-Actin and GAPDH as internal controls, the quantification was analyzed by ImageJ software (NIH, Bethesda, MD, USA). The values of the phosphorylated ERK-bands and

respective ERK bands were subsequently compared. The relative densities of ChAT and GS were determined via normalization against GAPDH and b-Actin, respectively. 2.7. Cell counts and statistical analyses 2.7.1. Retinal thickness analysis Using six HE-stained sections in each mouse, which were cut using the same horizontal angle across the optic disc, the thickness of the RGCL, IPL, INL, OPL, and ONL were measured by the Zeiss AxioVision 3.0 system. 2.7.2. Optical density analysis and cell counts To perform the optical density (OD) measurements, 6 comparable retinal sections across the optic disc in each mouse were stained by the Müller cell marker GS. Quantification of the ILM and Müller cell bodies were performed using the Zeiss AxioVision 3.0 system. OD measurements were carried out as

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

51

Fig. 3. MEM partially recovered the impaired visual function in 13-month-old APPswe/PS1DE9 transgenic mice. A: The amplitudes of the P50 components in the pattern electroretinograms of four groups. B: The latency of the P2 components of flash visual evoked potentials in the four groups. C: Comparison of the P50 amplitude in (A). The results showed that the amplitude of P50 was decreased in the AD þ Saline group, but showed that MEM could partially improve this value compared to the WT þ Saline group. D: Comparison of the latency of P2 component in (B). The results showed that the latency was increased in the AD þ Saline group, and showed that MEM could partially decrease this value compared to the WT þ Saline group. *P < 0.05; **P < 0.01; WT: wild type; AD: Alzheimer's disease, MEM: memantine.

previously reported (Perez et al., 2009). In detail, using a highmagnification microscope (400), comparable 6 visual fields of each section were randomly selected (200e400 mm away from the optic disc, nasal side). Then, a 2000-mm2 and a 5000-mm2 rectangular area in the RGCL and INL respectively were used to quantitatively analyze the OD mentioned above. Six 300-mm2 rectangular areas were randomly selected at the blank space in each visual field and measured the ODs. These measurements were averaged and subtracted from the final ODs to erase background ODs. To perform the whole retina quantification, 6 sections across the optic disc and stained with NeuN, Iba1, and ChAT in each mouse were selected. For RGCs and microglia counting, the number of NeuN-immunoreactive (NeuN-IR) RGCs in the RGCL and the number of Iba1-immunoreactive (Iba1-IR) microglia cells within the whole section was counted respectively. For amacrine cells counting, the number of ChAT-immunoreactive (ChAT-IR) amacrine cells

both located in RGCL and INL within the whole section was counted. 2.7.3. Statistical analysis Using the Statistical Product and Service Solutions software V17.0 (SPSS, Chicago, IL, USA), data were analyzed by one-way analyses of variance followed by Fisher's protected least-significant difference post hoc tests. All data are expressed as the mean ± standard error. The significance level was set at 0.05. 3. Results 3.1. MEM protected the RGCs and improved the visual function of 13-month-old APPswe/PS1DE9 dtg mice 3.1.1. Effect of MEM on the retinal thickness of AD mice HE staining revealed the retinal lamellar structure (Fig. 1AeD).

52

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

Fig. 4. Comparison of cholinergic amacrine cells and Müller cells in 13-month-old APPswe/PS1DE9 transgenic mice. AeD: Whole retinal montage of Choline acetyl transferase-immunoreactive (ChAT-IR) amacrine cells and GS-IR Müller cell in the (A) WT þ Saline, (B) WT þ MEM, (C) AD þ Saline, and (D) AD þ MEM groups. A1-D1: ChAT staining montage; A2-D2: GS staining montage. EeH: Enlarged image of ChAT and GS double staining in (E) WT þ Saline, (F) WT þ MEM, (G) AD þ Saline, and (H) AD þ MEM groups. White arrows: ChAT-IR amacrine cells in both the retinal ganglion cell layer (RGCL) and inner nuclear layer (INL). I, J: Enlarged images of Müller cells which were interacting with dendrites of cholinergic amacrine cells in the AD þ Saline group. Parallel white arrows: dendrites of cholinergic amacrine cells. Oblique white arrow: dendrites of Müller cell. K: ChAT-IR cholinergic amacrine cell counts in total, in the INL and RGCL. L: Western blot analysis of ChAT in the four groups and the corresponding statistics ([A] WT þ Saline, [B] WT þ MEM, [C] AD þ Saline, [D] AD þ MEM). Oblique arrow in I, J indicated the interacting of Müller cells and synaptic bands. *P < 0.05; **P < 0.01; Scale bar: AeD: 1000 mm; A1-D1, A2-D2: 1000 mm; EeH: 100 mm; I, J: 60 mm. GS: glutamine synthetase; DAPI: 4ʹ,6-diamidino-2-phenylindole; IPL: inner plexiform layer; OPL: outer plexiform layer; ONL: outer nuclear layer; WT: wild type; AD: Alzheimer's disease; MEM: memantine.

There were no significant differences in the thicknesses of the RGCL, NFL, IPL, INL, outer plexiform layer (OPL), and outer nuclear layer (ONL) (P > 0.05) (Fig. 1E, F) among the four groups. Additionally, there were no significant morphological differences between the groups (Fig. 1AeD). However, the distribution of cells in the RGCL of APPswe/PS1DE9 dtg mice was

sparser than that in the age matched wild type (WT) ntg mice (Fig. 1A, C). 3.1.2. Influences of MEM on the NeuN-IR RGCs of APPswe/PS1DE9 dtg mice NeuN was used as RGC marker in the retina (Niyadurupola et al.,

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

53

3.2. Displaced cholinergic amacrine cell degeneration in APPswe/ PS1DE9 dtg mice and the effect of MEM

Fig. 5. Cholinergic amacrine cell loss in 13-month-old APPswe/PS1DE9 transgenic mice via apoptosis. AeD: Double staining of terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and choline acetyl transferase (ChAT). In the AD þ Saline group, TUNEL and ChAT-immunoreactive cells were found in the retinal ganglion cell layer (C). The results showed cholinergic amacrine cells were lost through the process of apoptosis. Cell experiencing apoptosis was in a deformative shape (C). Scale bar: 25 mm GS: glutamine synthetase; DAPI: 4ʹ,6-diamidino-2-phenylindole; IPL: inner plexiform layer; WT: wild type; AD: Alzheimer's disease; MEM: memantine.

2013). The number of RGCs decreased significantly in the AD þ Saline group (202.33 ± 11.26) compared to that in the WT þ Saline group (243.00 ± 7.07) (about a 16.87% decrease, P < 0.01) (Fig. 2AeD, I). After MEM treatment, the number increased markedly (224.67 ± 12.47) compared to that in the AD þ Saline group (about a 11.04% increase, P < 0.05) (Fig. 2CeD, I). Following NeuN and TUNEL double staining, relative few appoptotic RGC was observed in the retina of mice in the WT þ Saline group (1.22% ± 0.32%) and WT þ MEM (1.38% ± 0.39%) group. There were significant apoptosis of RGCs in AD þ Saline group (8.55% ± 0.64%) compared to that in WT þ Saline group (P < 0.01) (Fig. 1G). MEM significantly decreased the number of appoptotic RGCs in the retina of 13-month-old APPswe/PS1DE9 dtg mice (4.29% ± 0.63%) (P < 0.01) (Fig. 1H, I).

3.1.3. Effect of MEM on the pERGs and fVEPs of APPswe/PS1DE9 dtg mice Both pERGs and fVEPs reflect visual function. pERGs represent the activity of RGCs and fVEPs test the signal transduction from ganglion cells towards the visual cortex (Porciatti, 2007; Ridder and Nusinowitz, 2006). The amplitude of the P50 component from the pERGs was significantly decreased in the AD þ Saline group (2.26 ± 0.11) compared to that in the WT þ Saline group (3.91 ± 0.34) (P < 0.01). After MEM treatment, although this value did not recover to the original level, it significantly increased to a level higher than that of the AD þ Saline group (2.96 ± 0.19) (P < 0.05) (Fig. 3A, C). We also observed a delay in the P2 component in the AD þ Saline group (75 ± 4.79) compared to that in the WT þ Saline group (62 ± 2.72) (P < 0.01). MEM treatment significantly reduced the latency and improved transduction (67 ± 3.21) (P < 0.05) (Fig. 3B, D).

Cholinergic neurons are known to be specifically damaged in the brains of patients with AD (Mesulam, 2004). Within the retina, cholinergic cells are ChAT-IR amacrine cells located in both the RGCL, called displaced amacrine cells, and the INL (Bernstein and Guo, 2011) (Fig. 4AeD, A1-D1). Applying ChAT immunofluorescence, we found that in the RGCL, ChAT-IR amacrine cells decreased significantly in the AD þ Saline group (34.00 ± 2.00) compared to that in the WT þ Saline group (46.33 ± 2.08) (P < 0.01) (Fig. 4EeH, K). Using TUNEL-ChAT double staining, we found that the decrease of the cell number might be caused by the apoptosis of ChAT-IR amacrine cells in which amacrine cells showed a long deformative shape (Fig. 5C), rather than the normal round shape (Fig. 5A, B, D). After MEM treatment, the number of ChAT-IR cell in the RGCL increased significantly compared with that in the AD þ Saline group (37.67 ± 1.53) (P < 0.05) (Fig. 4GeH, K). There were no significant differences in the number of ChAT-IR cells in the INL among the four groups (P > 0.05) (Fig. 4EeH, K). Besides, the results of the ChAT western blot also verified this finding. As shown in Fig. 4L, based on the expression level of bactin, the ChAT expression levels in the AD þ Saline group were significantly lower compared to the levels in the WT þ Saline group, but MEM treatment only demonstrated a trend towards recovering the expression level. 3.3. MEM inhibited the adapted response of Müller cells in the retina of APPswe/PS1DE9 dtg mice The GS immunofluorescence results showed that the inner limiting membrane (ILM), formed by the endfeet of Müller cells (Fig. 4AeD, A2-D2), increased in thickness in the AD þ Saline group (Fig. 6C) compared to that in the WT þ Saline (Fig. 6A). After MEM treatment, the thickened ILM decreased in thickness in the AD þ MEM group (Fig. 6D). In order to quantitatively describe this change, we performed optical density (OD) analysis of the GSstained area. The results showed that in the RGCL, the OD increased significantly in the AD þ Saline group (28.13 ± 2.00) (P < 0.05) compared to that in the WT þ Saline group (15.74 ± 1.19) (Fig. 6E), and MEM could partially decrease this value (21.29 ± 2.23) (P < 0.05) (Fig. 6E). There was no significant difference in OD within the INL for any of the groups (Fig. 6E). Western blot analysis of the GS protein showed that the expression was upregulated in the AD þ Saline group, but that this difference was not significant (Fig. 6F, G). In the retina, synapses of ChAT-IR amacrine cells appeared as two bands located in the IPL in retinal sections (Fig. 4I, J parallel white arrows). With ChAT and GS double immunofluorescence staining, we found that the synapses of ChAT-IR cholinergic amacrine cells overlapped with GS-IR Müller cells within the IPL in the AD þ Saline group (Fig. 4I, J oblique white arrow). 3.4. Activation of microglia in the retina of APPswe/PS1DE9 dtg mice and the modulation of MEM Iba1 was used as a specific marker of microglia (Sasaki et al., 2001). When activated, microglia change both their number and their cellular morphologies into amoeboid shapes (Ling and Wong, 1993). In the AD þ Saline group, the number of Iba1-IR microglia increased significantly (10.00 ± 1.00) compared to that in the WT þ Saline group (4.00 ± 1.00) (P < 0.01) (Fig. 7AeC, G). Furthermore, in the WT þ Saline group, microglia were in a small, ramified shape (Fig. 7A), while in the AD þ Saline group, microglia were activated, entered the characteristic amoeboid shape, and

54

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

Fig. 6. MEM regulated the adaptive responses of Müller cells in 13-month-old APPswe/PS1DE9 transgenic mice. AeD: Glutamine synthetase (GS) staining of Müller cells in the (A) WT þ Saline, (B) WT þ MEM, (C) AD þ Saline, and (D) AD þ MEM groups. White arrows respectively indicated the thickness of the internal limiting membrane (ILM) formed by the endfeet of Müller cells. E: Optical density analysis of GS-immunoreactive Müller cells in the INL and retinal ganglion cell layer (RGCL). F, G: Western blot analysis of GS protein and the related statistics ([A] WT þ Saline, [B] WT þ MEM, [C] AD þ Saline, [D] AD þ MEM). The result showed that the ILM was thicker in the AD þ Saline group (C), and MEM had the ability to make ILM return to a relatively thin state (D). Compared to the WT þ Saline group, the increased optical density in the AD þ Saline group indicating that Müller cells were stimulated into adaptive responses. Additionally, it decreased in the AD þ MEM group comparing to that in the AD þ Saline group, which corresponds with the results in GS staining. Though the western blots did not show significant result, we still observed an increasing trend in the AD þ Saline group (F, G). *P < 0.05; Scale bar: 100 mm. IPL: inner plexiform layer; INL: inner nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear layer; WT: wild type; AD: Alzheimer's disease; MEM: memantine.

showed a gathering trend toward the RGCL (Fig. 7C, F). After MEM treatment, the number of microglia was also significantly decreased (7.00 ± 1.70) (P < 0.05) compared to that in the AD þ Saline group (Fig. 7D, G), and the activated microglia not only recovered their ramified shape but also left RGCL (Fig. 7E, F). Moreover, with Iba1 and GS double staining, we also captured morphological contact between microglia and Müller cells (Fig. 7E) in the AD þ MEM group.

3.5. ERK1/2 pathway was involved in the immunomodulation effect of MEM in the retina of APPswe/PS1DE9 dtg mice We found that in the AD þ Saline group, the phosphorylation level of ERK1/2 significantly increased by 200% (P < 0.01) compared with the WT þ Saline group (Fig. 8A, B). MEM treatment was able to inhibit the upregulation by 7.5% significantly (P < 0.05) (Fig. 8A, B).

This result illustrated that MEM modulates the immune response via ERK1/2 pathway.

4. Discussion In the present study, we confirmed the loss of RGCs and visual dysfunction in APPswe/PS1DE9 dtg mice. Simultaneously, we found that treatment with MEM could improve visual function by regulating microglia, Müller cells, and by protecting RGCs and cholinergic amacrine cells. These findings suggest that MEM treatment, in addition to its convincing ability to relieve the cognitive damage symptoms in patients with AD (Araki et al., 2014), might contribute to the relief of their poor vision at the same time. It has been shown that the steady-state plasma level of 0.5e1 mM MEM produced significant neuroprotective effect in AD patients (Minkeviciene et al., 2008). Treating 10-month-old Tg2576

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

55

Fig. 7. MEM regulated inflammation reactions of microglia in APPswe/PS1DE9 transgenic mice. AeD: Double staining with ionized calcium binding adaptor molecule 1 (Iba1) and glutamine synthetase (GS) in the (A) WT þ Saline, (B) WT þ MEM, (C) AD þ Saline, and (D) AD þ MEM groups. White arrows indicated the Iba1-mmunoreactive microglia. E: An enlarged image of the Iba1-immunoreactive microglia and GS-immunoreactive Müller cells in the AD þ MEM group. White arrows in (E) showed the ramified processes of the microglia. F, G: comparison of the numbers of microglia in the RGCL and the whole section. In the WT groups, (A) and (B), microglia had small cell bodies with ramified processes. However, in the AD þ Saline group (C), the morphology of microglia became amoeboid shaped, together with an increase in cell number both in the RGCL (F) and the whole (G), which meant microglia were activated and gathered to the RGCL where injuries happened. After MEM treatment, microglia not only returned to their ramified shape (D, E), but also decreased in number compared to the AD þ Saline group. Especially, (E) showed a ramified microglia interacting with an Müller cell in the AD þ MEM group, which meant that there was an interaction between activated microglia and Müller cell in APPswe/PS1DE9 mice. After MEM treatment, the microglia returned to the resting-state (regeneration of ramified processes) and left the Müller cells. *P < 0.05; **P < 0.01; Scale bar: AeD 100 mm; E: 50 mm. DAPI: 40 , 6-diamidino-2-phenylindole; IPL: inner plexiform layer; OPL: outer plexiform layer; ONL: outer nuclear layer. WT: wild type; AD: Alzheimer's disease; MEM: memantine.

mice via subcutaneous injection at a dose of 10 mg/kg for 10 days, Unger et al. found that MEM reduced the total cortical levels of membrane-bound APP and Ab (Unger et al., 2006). However, treating 15-week-old APPswe/PS1DE9 mice via i.p. at a dose of 2.5 mg/kg for 8 days with MEM showed no significant effect on reducing the Ab level in APPswe/PS1DE9 mice (Alley et al., 2010).

On the other hand, using i.p. with a dose of 10 mg/kg for 8 days, MEM significantly reduced the accumulation of Ab in the cortex and hippocampus in APPswe/PS1DE9 mice (Sun et al., 2014). Considering these works and the toxic effect of long term MEM administration (Dong et al., 2008), we used the dose of MEM at 10 mg/kg/day i.p. for 8 days.

56

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

Fig. 8. MEM regulated the activation of ERK1/2 signaling pathway in APPswe/ PS1DE9 transgenic mice. A: Western blot analysis of ERK and phosphorylated ERK (pERK) protein ([A] WT þ Saline, [B] WT þ MEM, [C] AD þ Saline, [D] AD þ MEM). B: Statistical result of p-ERK/ERK. Phosphorylation of ERK was activated obviously in the AD þ Saline group, about 200%, compared with the WT þ Saline group. After MEM treatment, phosphorylation was downregulated in the AD þ MEM group. ERK 1/2: extracellular signal-regulated protein kinases 1 and 2; WT: wild type; AD: Alzheimer's disease; MEM: memantine.

As the native immunocytes within the retina, microglia provide significant and efficient immune response (Karlstetter et al., 2010). Similarly, in our study, the activation of microglia in APPswe/ PS1DE9 dtg mice was observed, which caused the degeneration of RGCs and cholinergic amacrine cells, possibly through the release of inflammatory factors and reactive oxygen species (Karlstetter et al., 2010). Consistent with the immune regulating effect of MEM on microglia in the brain (Wu et al., 2009), MEM treatment was able to modulate the activation of microglia in the retina. GS, mainly expressed in Müller cells, catalyzes the amidation of glutamate to glutamine and is the major component of the extracellular glutamate clearing system in retinas (Ehinger, 1977; Hampton and Redburn, 1983; Linser et al., 1984; Newman and Reichenbach, 1996). It is harmful of the high-level extracellular glutamate, since the RGCs death in glaucoma had been previously hypothesized to be an outcome of long-term elevation of extracellular glutamate (Dreyer et al., 1996). At the same time, intravitreal glutamate injection could cause the up-regulation of GS expression in rats (Shen et al., 2004). In the present study, the increase in GS expression in RGCL was observed, which illustrated that extracellular glutamate was increased. As the classical NMDA receptor antagonist, MEM is able to block the opened channel and reduce the excitotixity, once the extracellular glutamate was elevated (Hare and Wheeler, 2009). By NeuN staining, decrease in RGCs loss was discovered in AD dtg mice. However, as MEM showed its effect in reducing the deposition of Ab (Sun et al., 2014), and treating cultured neurons with Ab significantly increased the concentration of extracellular glutamate (Song et al., 2008). It is possible for MEM to lower the level of extracellular glutamate. In the present study, decrease in GS expression of AD dtg mice after MEM treatment was also observed, which informed us that MEM was able to reduce the level of extracellular glutamate in AD tg mice. Meanwhile, we captured the morphological contact between microglia and Müller cells in the AD þ MEM group. Unlike the previously described typical gliosis changes in Müller cells, which include proliferation, hypertrophy, and upregulation of glial fibrillary acidic protein and vimentin (Bringmann et al., 2009), our results showed that Müller cells only upregulated the expression of GS in the ILM and interacted with microglia, suggesting an adaptive response of Müller cells following microglial activation. This difference may be due to the relatively short-term and weak

degeneration stimuli within the retina of APPswe/PS1DE9 dtg mice (Wang et al., 2011). After MEM treatment, we found a significant decrease in both the thickness of the ILM and the expression of GS. These results suggest that MEM may relieve the adaptive responses of Müller cells. Furthermore, as interneurons in the retina, cholinergic amacrine cells are mainly charged with integrating signals, especially direction selecting signals, from bipolar cells to RGCs (Masland, 2005). Our current findings also revealed that ChAT-IR cholinergic amacrine cells were lost in APPswe/PS1DE9 dtg mice. The loss of these cells indicates a decrease in signal transduction and synaptic transmission between ChAT-IR amacrine cells and RGCs. As signal integration and upstream-derived signal transduction may be crucial in the function of RGCs (Bernstein and Guo, 2011), the loss of these cells may have a negative impact on the survival of RGCs. It has been reported that Ab deposition within the IPL may disturb the dendritic network of ChAT-IR amacrine cells (Perez et al., 2009). Our study showed that adaptive responses of Müller cells also formed contacts with the dendrites of amacrine cells, which might aggravate the disorder in their dendritic network. After treating the cells with MEM, a significant increase in the ChAT-IR amacrine cell number in the RGCL was observed. Meanwhile, western blots demonstrated that the level of ChAT was significantly decreased in APPswe/PS1DE9 mice, which was blocked by MEM. As a well-studied signal pathway, ERK1/2 was shown to be activated in the hippocampus of APPswe/PS1DE9 mice, resulting in further damage (Li and Liu, 2010). Activation of ERK was also regarded as a non-specific marker of activation in Müller cells (Bringmann et al., 2006). In the present study, the activation of microglia, the adapted response of Müller cells and the apoptosis of cholinergic amacrine cells accompanying the changes of phosphorylation of ERK1/2 was observed, which demonstrates that the ERK1/2 signaling pathway plays an important role in the pathological changes within the retinas of AD mice. MEM decreased the phosphorylation level of ERK, implying that the protective effects of MEM might be achieved by modulating ERK1/2 signaling pathway. As an integral part of the central nervous system, the retina also undergoes pathological changes similar to those in the brains of patients with AD (Kesler et al., 2011). The most remarkable advantage of the retina is that we can make relatively thorough iconography and functional inspections of the retina through the eyeball (Koronyo et al., 2012). The retina can also serve as a window for assessing the therapeutic effects of medications for AD (Koronyo et al., 2012). Therefore, studying the early pathological changes in the retina of AD is essential for the early diagnosis and therapeutic assessment of patients with AD. 5. Conclusions In summary, except for the deposition of Ab (Perez et al., 2009), loss of RGCs (Ning et al., 2008), and activation of microglia (Perez et al., 2009), the pathological changes within the retina of APPswe/PS1DE9 dtg mice also include the adapted response of Müller cells and apoptosis of ChAT-IR amacrine cells in the RGCL. MEM is able to recover the loss of RGCs, as shown by fVEPs and pERGs, mainly through its immunomodulating effect, which improves the entire living environment of RGCs and breaks the vicious cycle formed by the interactions of Müller cells and microglia, and the interactions of Müller cells and cholinergic amacrine cells. Competing interests The authors declare that they have no competing interests.

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

Authors' contributions Gao, Chen, Xu, and Fan contributed to manuscript preparation. Gao, Xu, Zhao, and Fan analyzed the data. Gao, Tang, and Zhao conducted the experiments including immunofluorescence staining, western blots, and so on. Gao and Li performed ERG and VEP testing. Yin, Xu, and Fan contributed to the design of the project and discussed the results. Acknowledgments This study was supported by the National Nature Science Foundation of China (No.31271051, No.81371197). References Alley, G.M., Bailey, J.A., Chen, D., Ray, B., Puli, L.K., Tanila, H., Banerjee, P.K., Lahiri, D.K., 2010. Memantine lowers amyloid-beta peptide levels in neuronal cultures and in APP/PS1 transgenic mice. J. Neurosci. Res. 88, 143e154. Araki, T., Wake, R., Miyaoka, T., Kawakami, K., Nagahama, M., Furuya, M., Limoa, E., Liaury, K., Hashioka, S., Murotani, K., Horiguchi, J., 2014. The effects of combine treatment of memantine and donepezil on Alzheimer's Disease patients and its relationship with cerebral blood flow in the prefrontal area. Int. J. Geriatr. Psychiatry 29, 881e889. Bayer, A.U., Ferrari, F., Erb, C., 2002. High occurrence rate of glaucoma among patients with Alzheimer's disease. Eur. Neurol. 47, 165e168. Bernstein, S.L., Guo, Y., 2011. Changes in cholinergic amacrine cells after rodent anterior ischemic optic neuropathy (rAION). Invest. Ophthalmol. Vis. Sci. 52, 904e910. Blanks, J.C., Schmidt, S.Y., Torigoe, Y., Porrello, K.V., Hinton, D.R., Blanks, R.H., 1996. Retinal pathology in Alzheimer's disease. II. Regional neuron loss and glial changes in GCL. Neurobiol. Aging 17, 385e395. Borchelt, D.R., Davis, J., Fischer, M., Lee, M.K., Slunt, H.H., Ratovitsky, T., Regard, J., Copeland, N.G., Jenkins, N.A., Sisodia, S.S., Price, D.L., 1996a. A vector for expressing foreign genes in the brains and hearts of transgenic mice. Genet. Anal. 13, 159e163. Borchelt, D.R., Thinakaran, G., Eckman, C.B., Lee, M.K., Davenport, F., Ratovitsky, T., Prada, C.M., Kim, G., Seekins, S., Yager, D., Slunt, H.H., Wang, R., Seeger, M., Levey, A.I., Gandy, S.E., Copeland, N.G., Jenkins, N.A., Price, D.L., Younkin, S.G., Sisodia, S.S., 1996b. Familial Alzheimer's disease-linked presenilin 1 variants elevate Abeta1-42/1-40 ratio in vitro and in vivo. Neuron 17, 1005e1013. Bormann, J., 1989. Memantine is a potent blocker of N-methyl-D-aspartate (NMDA) receptor channels. Eur. J. Pharmacol. 166, 591e592. Bringmann, A., Iandiev, I., Pannicke, T., Wurm, A., Hollborn, M., Wiedemann, P., Osborne, N.N., Reichenbach, A., 2009. Cellular signaling and factors involved in Müller cell gliosis: neuroprotective and detrimental effects. Prog. Retin. Eye. Res. 28, 423e451. Bringmann, A., Pannicke, T., Grosche, J., Francke, M., Wiedemann, P., Skatchkov, S.N., Osborne, N.N., Reichenbach, A., 2006. Müller cells in the healthy and diseased retina. Prog. Retin. Eye. Res. 25, 397e424. Bringmann, A., Wiedemann, P., 2012. Müller glial cells in retinal disease. Ophthalmologica 227, 1e19. Coburn, K.L., Amoss, R.T., Arruda, J.E., Kizer, L.D., Marshall, Y.S., 2005. Effects of flash mode and intensity on P2 component latency and amplitude. Int. J. Psychophysiol. 55, 323e331. Dong, H., Yuede, C.M., Coughlan, C., Lewis, B., Csernansky, J.G., 2008. Effects of memantine on neuronal structure and conditioned fear in the Tg2576 mouse model of Alzheimer's disease. Neuropsychopharmacology 33, 3226e3236. Dong, N., Chang, L., Wang, B., Chu, L., 2014. Retinal neuronal MCP-1 induced by AGEs stimulates TNF-a expression in rat microglia via p38, ERK, and NF-kB pathways. Mol. Vis. 20, 616e628. Dreyer, E.B., Zurakowski, D., Schumer, R.A., Podos, S.M., Lipton, S.A., 1996. Elevated glutamate levels in the vitreous body of humans and monkeys with glaucoma. Arch. Ophthalmol. 114, 299e305. Dutescu, R.M., Li, Q.X., Crowston, J., Masters, C.L., Baird, P.N., Culvenor, J.G., 2009. Amyloid precursor protein processing and retinal pathology in mouse models of Alzheimer's disease. Graefes Arch. Clin. Exp. Ophthalmol. 247, 1213e1221. Ehinger, B., 1977. Glial and neuronal uptake of GABA, glutamic acid, glutamine and glutathione in the rabbit retina. Exp. Eye Res. 25, 221e234. Fan, X., Sun, D., Tang, X., Cai, Y., Yin, Z.Q., Xu, H., 2014. Stem-cell challenges in the treatment of Alzheimer's disease: a long way from bench to bedside. Med. Res. Rev. 34, 957e978. Francke, M., Pannicke, T., Biedermann, B., Faude, F., Wiedemann, P., Reichenbach, A., Reichelt, W., 1997. Loss of inwardly rectifying potassium currents by human retinal glial cells in diseases of the eye. Glia 20, 210e218. Guise, S., Braguer, D., Carles, G., Delacourte, A., Briand, C., 2001. Hyperphosphorylation of tau is mediated by ERK activation during anticancer druginduced apoptosis in neuroblastoma cells. J. Neurosci. Res. 63, 257e267. Guo, L., Duggan, J., Cordeiro, M.F., 2010. Alzheimer's disease and retinal neurodegeneration. Curr. Alzheimer Res. 7, 3e14.

57

Hampton, C.K., Redburn, D.A., 1983. Autoradiographic analysis of 3H-glutamate, 3Hdopamine, and 3H-GABA accumulation in rabbit retina after kainic acid treatment. J. Neurosci. Res. 9, 239e251. Hare, W.A., Wheeler, L., 2009. Experimental glutamatergic excitotoxicity in rabbit retinal ganglion cells: block by memantine. Invest. Ophthalmol. Vis. Sci. 50, 2940e2948. Karlstetter, M., Ebert, S., Langmann, T., 2010. Microglia in the healthy and degenerating retina: insights from novel mouse models. Immunobiology 215, 685e691. Kesler, A., Vakhapova, V., Korczyn, A.D., Naftaliev, E., Neudorfer, M., 2011. Retinal thickness in patients with mild cognitive impairment and Alzheimer's disease. Clin. Neurol. Neurosurg. 113, 523e526. Koronyo, Y., Salumbides, B.C., Black, K.L., Koronyo-Hamaoui, M., 2012. Alzheimer's disease in the retina: imaging retinal Ab plaques for early diagnosis and therapy assessment. Neurodegener. Dis. 10, 285e293.  ski, W., Potemkowski, A., Honczarenko, K., 2010. Pattern Krasodomska, K., Lubin electroretinogram (PERG) and pattern visual evoked potential (PVEP) in the early stages of Alzheimer's disease. Doc. Ophthalmol. 121, 111e121. Lee, M.K., Borchelt, D.R., Kim, G., Thinakaran, G., Slunt, H.H., Ratovitski, T., Martin, L.J., Kittur, A., Gandy, S., Levey, A.I., Jenkins, N., Copeland, N., Price, D.L., Sisodia, S.S., 1997. Hyperaccumulation of FAD-linked presenilin 1 variants in vivo. Nat. Med. 3, 756e760. Li, D., Tang, J., Xu, H., Fan, X., Bai, Y., Yang, L., 2008. Decreased hippocampal cell proliferation correlates with increased expression of BMP4 in the APPswe/ PS1DeltaE9 mouse model of Alzheimer's disease. Hippocampus 18, 692e698. Li, N., Liu, G.T., 2010. The novel squamosamide derivative FLZ enhances BDNF/TrkB/ CREB signaling and inhibits neuronal apoptosis in APP/PS1 mice. Acta. Pharmacol. Sin. 31, 265e272. Ling, E.A., Wong, W.C., 1993. The origin and nature of ramified and amoeboid microglia: a historical review and current concepts. Glia 7, 9e18. Linser, P.J., Sorrentino, M., Moscona, A.A., 1984. Cellular compartmentalization of carbonic anhydrase-C and glutamine synthetase in developing and mature mouse neural retina. Brain Res. 315, 65e71. Lipton, S.A., 2003. Possible role for memantine in protecting retinal ganglion cells from glaucomatous damage. Surv. Ophthalmol. 48, S38eS46. Liu, S., Li, Z.W., Weinreb, R.N., Xu, G., Lindsey, J.D., Ye, C., Yung, W.H., Pang, C.P., Lam, D.S., Leung, C.K., 2012. Tracking retinal microgliosis in models of retinal ganglion cell damage. Invest. Ophthalmol. Vis. Sci. 53, 6254e6262. Masland, R.H., 2005. The many roles of starburst amacrine cells. Trends Neurosci. 28, 395e396. Mesulam, M., 2004. The cholinergic lesion of Alzheimer's disease: pivotal factor or side show? Learn Mem. 11, 43e49. Minkeviciene, R., Banerjee, P., Tanila, H., 2008. Cognition-enhancing and anxiolytic effects of memantine. Neuropharmacology 54, 1079e1085. Newman, E., Reichenbach, A., 1996. The Müller cell: a functional element of the retina. Trends Neurosci. 19, 307e312. Ning, A., Cui, J., To, E., Ashe, K.H., Matsubara, J., 2008. Amyloid-beta deposits lead to retinal degeneration in a mouse model of Alzheimer disease. Invest. Ophthalmol. Vis. Sci. 49, 5136e5143. Niyadurupola, N., Sidaway, P., Ma, N., Rhodes, J.D., Broadway, D.C., Sanderson, J., 2013. P2X7 receptor activation mediates retinal ganglion cell death in a human retina model of ischemic neurodegeneration. Invest. Ophthalmol. Vis. Sci. 54, 2163e2170. Perez, S.E., Lumayag, S., Kovacs, B., Mufson, E.J., Xu, S., 2009. Beta-amyloid deposition and functional impairment in the retina of the APPswe/PS1DeltaE9 transgenic mouse model of Alzheimer's disease. Invest. Ophthalmol. Vis. Sci. 50, 793e800. Porciatti, V., 2007. The mouse pattern electroretinogram. Doc. Ophthalmol. 115, 145e153. Porciatti, V., Saleh, M., Nagaraju, M., 2007. The pattern electroretinogram as a tool to monitor progressive retinal ganglion cell dysfunction in the DBA/2J mouse model of glaucoma. Invest. Ophthalmol. Vis. Sci. 48, 745e751. Qian, L., Wei, S.J., Zhang, D., Hu, X., Xu, Z., Wilson, B., El-Benna, J., Hong, J.S., Flood, P.M., 2008. Potent anti-inflammatory and neuroprotective effects of TGFbeta1 are mediated through the inhibition of ERK and p47phox-Ser345 phosphorylation and translocation in microglia. J. Immunol. 181, 660e668. Ridder 3rd, W.H., Nusinowitz, S., 2006. The visual evoked potential in the mouseeorigins and response characteristics. Vis. Res. 46, 902e913. Sasaki, Y., Ohsawa, K., Kanazawa, H., Kohsaka, S., Imai, Y., 2001. Iba1 is an actincross-linking protein in macrophages/microglia. Biochem. Biophys. Res. Commun. 286, 292e297. Shen, F., Chen, B., Danias, J., Lee, K.C., Lee, H., Su, Y., Podos, S.M., Mittag, T.W., 2004. Glutamate-induced glutamine synthetase expression in retinal Muller cells after short-term ocular hypertension in the rat. Invest. Ophthalmol. Vis. Sci. 45, 3107e3112. Song, M.S., Rauw, G., Baker, G.B., Kar, S., 2008. Memantine protects rat cortical cultured neurons against beta-amyloid-induced toxicity by attenuating tau phosphorylation. Eur. J. Neurosci. 28, 1989e2002. Sun, D., Chen, J., Bao, X., Cai, Y., Zhao, J., Huang, J., Huang, W., Fan, X., Xu, H., 2014 Sep 9. Protection of radial glial-like cells in the Hippocampus of APP/PS1 mice: a novel mechanism of memantine in the treatment of Alzheimer's disease. Mol. Neurobiol. [Epub ahead of print]. Tao, Z., Dai, J., He, J., Li, C., Li, Y., Yin, Z.Q., 2013. The influence of NaIO(3)-induced retinal degeneration on intra-retinal layer and the changes of expression profile/morphology of DA-ACs and mRGCS. Mol. Neurobiol. 47, 241e260.

58

L. Gao et al. / Experimental Eye Research 135 (2015) 47e58

Unger, C., Svedberg, M.M., Yu, W.F., Hedberg, M.M., Nordberg, A., 2006. Effect of subchronic treatment of memantine, galantamine, and nicotine in the brain of Tg2576 (APPswe) transgenic mice. J. Pharmacol. Exp. Ther. 317, 30e36. Wang, M., Ma, W., Zhao, L., Fariss, R.N., Wong, W.T., 2011. Adaptive Müller cell responses to microglial activation mediate neuroprotection and coordinate inflammation in the retina. J. Neuroinflammation 8, 173. Wu, H.M., Tzeng, N.S., Qian, L., Wei, S.J., Hu, X., Chen, S.H., Rawls, S.M., Flood, P.,

Hong, J.S., Lu, R.B., 2009. Novel neuroprotective mechanisms of memantine: increase in neurotrophic factor release from astroglia and anti-inflammation by preventing microglial activation. Neuropsychopharmacology 34, 2344e2357. Yang, Y., Tang, Y., Xing, Y., Zhao, M., Bao, X., Sun, D., Tang, X., Wu, Y., Xu, H., Fan, X., 2014. Activation of liver X receptor is protective against ethanol-induced developmental impairment of bergmann glia and purkinje neurons in the mouse cerebellum. Mol. Neurobiol. 49, 176e186.