RAGE mediated intracellular Aβ uptake contributes to the breakdown of tight junction in retinal pigment epithelium.

Intracellular amyloid beta (Aβ) has been implicated in neuronal cell death in Alzheimer's disease (AD). Intracellular Aβ also contributes to tight junction breakdown of retinal pigment epithelium (RPE) in age-related macular degeneration (AMD). Although Aβ is predominantly secreted from neuronal cells, the mechanism of Aβ transport into RPE remains to be fully elucidated. In this study, we demonstrated that intracellular Aβ was found concomitantly with the breakdown of tight junction in RPE after subretinal injection of Aβ into the mouse eye. We also presented evidence that receptor for advanced glycation end products (RAGE) contributed to endocytosis of Aβ in RPE. siRNA-mediated knockdown of RAGE prevented intracellular Aβ accumulation as well as subsequent tight junction breakdown in RPE. In addition, we found that RAGE-mediated p38 MAPK signaling contributed to endocytosis of Aβ. Blockade of RAGE/p38 MAPK signaling inhibited Aβ endocytosis, thereby preventing tight junction breakdown in RPE. These results implicate that intracellular Aβ contributes to the breakdown of tight junction in RPE via the RAGE/p38 MAPK-mediated endocytosis. Thus, we suggest that RAGE could be a potential therapeutic target for intracellular Aβ induced outer BRB breakdown in AMD.


INTRODUCTION
Age-related macular degeneration (AMD) is the leading cause of blindness in the elderly population and the prevalence is increasing in the world [1]. Drusen, a focal deposition of acellular debris between the retinal pigment epithelium (RPE) and Bruch's membrane, is the clinical hallmark and usually the initial clinical finding of AMD. Amyloid β (Aβ) is known to be found in drusen [2,3]. In addition, the deposition of Aβ in the subRPE space is also concomitant with several features of AMD in a mouse model of Alzheimer disease (AD) [4,5]. It has been suggested that Aβ in drusen may be an important contributor to the development of AMD.
Drusen in AMD and senile plaque cores in AD are similar in that deposition of Aβ is found in extracellular Oncotarget 35264 www.impactjournals.com/oncotarget space at later stages of both diseases. The role of extracellular Aβ in RPE is known to alter the tight junction without apoptosis [6]. Recent studies, on the other hand, suggest that accumulation of intraneuronal Aβ contributes neuronal apoptosis and may be an early event in the pathogenesis of AD [7,8]. In this regard, we recently reported that intracellular Aβ contributed to the breakdown of outer blood-retinal barrier (BRB) in 5XFAD mice, a mouse model of AD, and suggested that intracellular Aβ could be also a key contributor to the development of AMD [9]. However, the precise mechanism of intracellular uptake of Aβ in RPE remains to be elucidated.
Receptor for advanced glycation end products (RAGE), a member of the immunoglobulin superfamily, is a multi-ligand receptor capable of binding diverse range of molecules including Aβ [10]. The receptor is known to be highly expressed in the RPE and levels of RAGE are significantly elevated in AMD, especially in RPE adjacent to drusen [11,12]. RAGE is also known to participate in the uptake of Aβ from the circulation into the brain by endocytosis and transcytosis in endothelial cells [13,14]. Recently, it has been suggested that RAGE-mediated signaling contributes to intraneuronal transport of Aβ [15]. Thus, we hypothesized that RAGE could contribute to intracellular transport of Aβ in RPE. If Aβ uptake is mediated by RAGE in RPE, anti-RAGE therapy could inhibit intracellular accumulation of Aβ, and subsequently prevent the breakdown of tight junction in RPE. This can expand the application of anti-RAGE therapy from AD to AMD [16,17].
Herein, we demonstrate that subretinal injection of Aβ leads to intracellular uptake of Aβ and subsequent breakdown of tight junction in RPE in vivo. We also present evidence that extracellular Aβ gets translocated into intracellular space via RAGE-mediated endocytosis in RPE. The Aβ/RAGE-mediated p38 MAPK signaling contributes to endocytosis of Aβ. Blocking of RAGE inhibits intracellular Aβ-induced tight junction breakdown in RPE. These findings indicate that RAGE contributes to the intracellular transport of Aβ, resulting in increased tight junction breakdown in RPE. Therefore, we suggest that blocking of RAGE can be a potential therapeutic target in AMD.

Subretinal injection of Aβ leads to intracellular Aβ uptake and subsequent breakdown of tight junction in RPE
We have recently demonstrated that Aβ, endogenously generated from 5 mutant transgenes, accumulated in intracellular space at the RPE layer with thickened Bruch's membrane and Aβ deposits, thereby attenuating tight junction of RPE in aged 5XFAD mice [9]. Exogenous oligomeric Aβ (OAβ) 42 injection into subretinal space is also known to break tight junctions of RPE and result in disorganized and irregular staining of both ZO-1 and occludin on days 3 to 15 post-injection in both young and aged mice [6]. These findings led us to probe whether extracellular Aβ translocates into intracellular space and subsequently induces breakdown of tight junction of RPE in vivo.
To evaluate cellular uptake of Aβ in RPE, we demonstrated intracellular Aβ uptake after subretinal injection of exogenous OAβ 42 . Seven days after subretinal OAβ 42 injection, intracellular Aβ 42 was found concomitantly with the breakdown of tight junction ( Figure 1). While typical hexagonal pattern of ZO-1 was not disrupted in vehicle injected mice ( Figure 1A), the disrupted and irregular pattern of ZO-1 was found in OAβ 42 injected mice which indicated a breakdown of tight junction in RPE ( Figure 1B). These specific changes of tight junction were shown in details ( Figure 1C and 1D). We also confirmed the intracellular uptake of Aβ using FITC-labeled OAβ 42 ( Figure 1E). These data suggest that RPE uptakes exogenous Aβ, and this, in turn, leads to tight junction breakdown in vivo.

Extracellular Aβ translocates into intracellular space via RAGE-mediated endocytosis in RPE
In order to determine the mechanism how extracellular Aβ gets translocated into intracellular space in RPE, we demonstrated intracellular Aβ uptake with plasma membrane and RAGE after exogenous OAβ 42 treatment. First, we assessed intracellular distribution of Aβ and biotinylated-cell surface protein after Aβ treatment. After biotinylation of cell surface protein, cells were incubated with vehicle or OAβ 42 for 1 h, and were treated with 2-mercaptoethanesulfonate (MesNa) to remove any remaining biotin on the cell surface. While control cells exposed to vehicle alone showed some biotinylated-proteins in cytosol (Figure 2A), RPE cells exposed to OAβ 42 showed significantly increased biotinylated-proteins in cytosol with intracellular Aβ ( Figure 2B). Intriguingly, intracellular Aβ was overlapped with some internalized biotinylated-proteins ( Figure 2B, inset) suggesting that Aβ could bind to cell surface protein and translocate into intracellular space together.
To access possible colocalization of RAGE and Aβ, cells were incubated with vehicle or OAβ 42 for 1h, and the distribution of RAGE and Aβ was examined under confocal microscope ( Figure 2C and 2D). Indeed, the colocalization of RAGE and Aβ was detected in RPE exposed to OAβ 42 ( Figure 2D). These data suggested that Aβ interacts with RAGE and is internalized into RPE.
To analyze the effect of intracellular Aβ on tight junction integrity in RPE, we studied tight junction of RPE Oncotarget 35265 www.impactjournals.com/oncotarget cells exposed to vehicle or OAβ 42 for 24 h ( Figure 2E and 2F). Intracellular Aβ was concomitant with disrupted and disorganized ZO-1 expression ( Figure 2F).

siRNA-mediated knockdown of RAGE suppresses Aβ uptake in RPE
To investigate the role of RAGE on Aβ uptake in RPE cells, we performed in vitro study in RPE cells treated with RAGE siRNA. siRNA-mediated knockdown effectively decreased AGER mRNA (0.16 ± 0.02 fold induction) compared to negative siRNA (1.00 ± 0.13, p < 0.05; Figure 3A). In accordance with this result, it effectively decreased RAGE expression in RPE ( Figure  3B). Then, we examined intracellular Aβ uptake in RPE cells with RAGE siRNA. Interestingly, intracellular Aβ uptake was decreased in RPE with RAGE siRNA compared to RPE with negative siRNA under confocal Oncotarget 35266 www.impactjournals.com/oncotarget microscope ( Figure 3C). Consistent with these data, western blot results also showed that intracellular Aβ was significantly decreased in RPE with RAGE siRNA compared to RPE with negative siRNA ( Figure 3D and 3E). These data showed that siRNA-mediated knockdown of RAGE suppressed Aβ uptake in RPE. This implied that RAGE played an important role in intracellular Aβ uptake in RPE

RAGE-mediated p38 MAPK signaling contributes to endocytosis of Aβ in RPE
RAGE is known to activate multiple downstream signaling pathways as a signal transduction receptor [18]. Based on the previous study [15], we demonstrated that Aβ-induced RAGE signaling activation might lead to Aβ uptake into RPE. We first examined the effect of Aβ treatment on phosphorylation of p38 MAPK. Treatment   Oncotarget 35268 www.impactjournals.com/oncotarget of OAβ 42 for 30 min showed a dose-dependent increase in phosphorylated p38 MAPK although it did not affect total protein levels of p38 MAPK ( Figure 4A).

Blockade of RAGE inhibits intracellular Aβinduced tight junction breakdown in RPE
To determine whether inhibition of intracellular Aβ uptake could prevent Aβ-induced tight junction breakdown in RPE, we analyzed tight junction expression in RPE using western blot. We demonstrated that Oncotarget 35269 www.impactjournals.com/oncotarget intracellular uptake of exogenous OAβ 42 was decreased in RPE with siRNA-mediated knockdown of RAGE ( Figure  4B). Indeed, siRNA-mediated knockdown of RAGE diminished the uptake of Aβ which subsequently inhibited intracellular Aβ-induced tight junction breakdown in RPE ( Figure 5A and 5B). Consistent with these data, treatment of p38 MAPK inhibitor (SB 203580) and anti-RAGE neutralizing antibody also showed protective effect on Aβinduced decrease of occludin expression in RPE ( Figure  5C).
To determine the role of RAGE on barrier function of RPE after Aβ treatment, we measured transepithelial resistance (TER) in RPE for 48 h. The mean initial TER of controls was 46±4.5 ohms·cm2 when they reached a plateau. Anti-RAGE neutralizing antibody was sufficient to inhibit a decrease of TER during 48 h after OAβ 42 treatment (p < 0.05; Figure 5D). In addition, tight junction breakdown with ZO-1 disruption after OAβ 42 was also attenuated when ARPE-19 cells were treated with either p38 MAPK inhibitor or anti-RAGE neutralizing antibody ( Figure 5E). These data suggested that intracellular Aβ uptake by RAGE contributed to breakdown of tight junction in RPE.

In this study, we demonstrated that subretinal injection of Aβ led to intracellular uptake of Aβ and subsequent breakdown of tight junction in RPE in vivo.
We also showed that intracellular Aβ contributed to breakdown of tight junction in RPE via the RAGE/p38 MAPK-mediated endocytosis. RAGE contributed to intracellular uptake of Aβ which resulted in the increase of tight junction breakdown in RPE. Our study of the mechanism of intracellular Aβ uptake in RPE might help to investigate the role of intracellular Aβ in the pathogenesis of dry AMD. Furthermore, we also suggested that the research field of Aβ should be expanded from AD to AMD, and from extracellular Aβ to intracellular Aβ.
A growing body of evidence suggests that intraneuronal Aβ contributes neuronal apoptosis in the early pathogenesis of AD [7,8]. Intracellular Aβ 42 is known to accumulate in 5XFAD mice brain prior to plaque formation [19]. We hypothesized that intracellular Aβ was also important in the early pathogenesis of AMD. However, the role of intracellular Aβ in RPE was only described in our previous study suggesting that intracellular Aβ in RPE could contribute to the development of AMD [9]. Aβ is predominantly secreted from the neuronal cells, but the mechanism of Aβ transport into the RPE remains to be fully elucidated in AMD.
Regarding to Aβ uptake by RPE cells, the mechanism of intracellular uptake of exogenous Aβ is not established. Although the main role of RPE is phagocytosis, RPE phagocytosis showed a remarkable specificity toward ROS via αvβ5 integrin [20,21]. There were also several endocytotic pathways suggested as possible uptakes of Aβ in neurons and endothelial cells, including α7-NAch receptor [22], NMDA receptor [23], LDL receptor-related protein 1 [24], and RAGE [14]. Among these candidates, we demonstrated that RAGE contributed to intracellular Aβ uptake in RPE. RAGE is also important in retinal aging [25]. It is known that RAGE distributed at apical membrane of the RPE [11,26]. In line with our results, Aβ/RAGE-mediated p38 MAPK signaling contributes to Aβ transport and neuronal dysfunction [15]. Activation of endocytosis is essential in regulating the p38 MAPK activity in endothelial cells [27]. Despite the evident role of RAGE in endocytosis of Aβ in this study, phagocytosis or other receptor mediated endocytosis could also contribute to intracellular uptake of Aβ. In addition, we also hypothesized that RAGE-mediated signaling itself (i.e. GSK-3β or NF-κB related signaling pathway) could affect tight junction or Aβ-related pathology in RPE. Thus, those Aβ-related mechanisms on RPE pathology should be investigated in further study.
The limitation of the study is that ARPE19 cells generate less effective tight junction than human fetal RPE (hfRPE) cells for in vitro study. It is also known that the staining with occludin in ARPE-19 cells is not good as that in hfRPE [28]. Thus, it is better to show the distribution of ZO-1 tight junction protein in immunocytochemistry for tight junction evaluation in ARPE-19 cells. Nonetheless, this study for the first time demonstrated that intracellular Aβ uptake was mediated by RAGE via the RAGE/p38 MAPK-mediated endocytosis in RPE.
In conclusion, intracellular Aβ uptake was mediated by RAGE in RPE. Intracellular Aβ contributed to breakdown of tight junction in RPE via the RAGE/ p38 MAPK-mediated endocytosis. Blockade of RAGE with anti-RAGE antibody could block the intracellular accumulation of Aβ, and subsequently prevent the breakdown of tight junction in RPE. Thus, we suggest that blockade of RAGE can be a potential therapeutic target in AMD. Further study is warranted to develop new modality of RAGE-specific inhibitor for the treatment of AMD.

Animals
All animal experiments in this study were in strict agreement with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and the guidelines of the Seoul National University Animal Care and Use Committee. Six-week-old, pathogen-free male C57BL/6J mice were purchased from Central Lab (Seoul, Korea).

Cell cultures and siRNA transfection
ARPE-19 cells (American Type Culture Collection, Manassas, VA, USA) were used for human RPE cells. The cells were routinely maintained in DMEM/F12 containing 10% FBS, 100 U/mL penicillin, and 100 μg/ mL streptomycin. RAGE siRNA and negative siRNA (50 nM) were transfected using Lipofectamine RNAi max in opti-MEM ® media according to the manufacturer's instructions. After reaching confluent, ARPE-19 cells were maintained in DMEM/F12 containing 1% FBS to make a polarization for the experiments [29].

Preparation of oligomeric Aβ42 solution
OAβ 42 solution were generated as the previously described method [9]. Aβ 42 and FITC labeled Aβ 42 (62-0-80B and 62-0-82B) were dissolved in hexafluoro-2propanol (Sigma Aldrich, St. Louis, MO, USA) to a final concentration of 1 mg/ml at RT for 3 days. The peptide was aliquoted and dried under vacuum for 1 h. The aliquoted peptide was dissolved in DMSO to a final concentration of 2 mM. The protein concentration was measured using a BCA protein assay kit (Pierce Biotechnology Inc.). The Aβ 42 stock in DMSO was diluted directly into DMEM/F12 at 10 μM, and incubated for 24 h at 4 o C to make OAβ 42 .

Biotinylation of membrane proteins
Biotinylation of cell membrane proteins was performed as previously described method [15]. ARPE-19 cells plated in 8-well culture slides were washed with PBS and incubated with PBS containing 300 μg/mL sulfo-NHS-SS-biotin at RT for 30 min. The biotinylation of membrane proteins was quenched using 1 M glycine in PBS at 4 o C for 15 min. After additional two brief wash with 100 mM glycine in PBS to remove the residual biotin, cells were incubated with OAβ 42 at 37 o C for 1 h. Then, to terminate uptake of biotinylated proteins, cells were washed with cold NT buffer (150 mM NaCl, 1mM EDTA, 0.2% BSA, and 20 mM Tris, pH 8.6). Then, cells were incubated with 50 mM MesNa/NT buffer at 4 o C for 30 min.

Immunocytochemistry
ARPE-19 cells with confluence were incubated with OAβ 42 (10 μM) in serum free DMEM/F12 medium for 24 h. After removal of the medium, cells were intensively washed with warm PBS and fixed www.impactjournals.com/oncotarget with 4% paraformaldehyde at RT for 15 min. For the permeabilization, 0.2% Triton X-100 in PBS was treated for 10 min. After washing, the cells were incubated with 1% BSA in PBS at RT for 1 h. Cells were incubated overnight at 4 o C with rabbit anti-ZO-1 (1:1000), mouse anti-Aβ (1:1000) and goat anti-RAGE (1:500). After washing, cells were incubated at RT for 1 h with secondary antibodies (Alexa Fluor 488, 594, 647 anti-mouse, rabbit, goat IgG, 1:500). Nucleus was counterstained with DAPI at RT for 10 min. After washing, the slides were mounted and observed under confocal microscope (Leica TCS STED).

Transepithelial electrical resistance measurement
The measurement of TER was performed by impedance analysis using EVOM2 TER (World Precision Instruments). Briefly, the cell covered transwell filters (0.4 μm, Corning Inc., NY, USA) coated with laminin were placed in this setup using 12-well plate. After stabilization of TER value on plateau, TER was measured with 10 μM OAβ 42 and anti-RAGE neutralizing antibody (20 μg/ml) for 48 h. The TER value just after treatment was normalized to 100% for relative analysis.

Real-time PCR
Total RNA was isolated from RPE cells using TRIzol reagent according to the manufacturer's instructions. cDNA was prepared with High Capacity RNA-to-cDNA kit. Real-time PCR was performed with StepOnePlus Real-time PCR System with TaqMan ® Fast Advanced Master Mix and specific Gene Expression Assays (AGER and GAPDH). All procedures were performed in accordance with the MIQE guidelines. All material and machine used in Real-time PCR was purchased from Life Technologies.

Statistical analysis
Statistical analyses were performed using SPSS software version 18.0 (SPSS Inc., Chicago, IL, USA). Two-tailed unpaired T-test was used. P values less than 0.05 were considered to be statistically significant. Data and figures are depicted as mean ± SEM.