Endocytosis of Tight Junctions Caveolin Nitrosylation Dependent Is Improved by Cocoa Via Opioid Receptor on RPE Cells in Diabetic Conditions

Abstract

Purpose.: Retinal pigment epithelium cells, along with tight junction (TJ) proteins, constitute the outer blood retinal barrier (BRB). Contradictory findings suggest a role for the outer BRB in the pathogenesis of diabetic retinopathy (DR). The aim of this study was to investigate whether the mechanisms involved in these alterations are sensitive to nitrosative stress, and if cocoa or epicatechin (EC) protects from this damage under diabetic (DM) milieu conditions.

Methods.: Cells of a human RPE line (ARPE-19) were exposed to high-glucose (HG) conditions for 24 hours in the presence or absence of cocoa powder containing 0.5% or 60.5% polyphenol (low-polyphenol cocoa [LPC] and high-polyphenol cocoa [HPC], respectively).

Results.: Exposure to HG decreased claudin-1 and occludin TJ expressions and increased extracellular matrix accumulation (ECM), whereas levels of TNF-α and inducible nitric oxide synthase (iNOS) were upregulated, accompanied by increased nitric oxide levels. This nitrosative stress resulted in S-nitrosylation of caveolin-1 (CAV-1), which in turn increased CAV-1 traffic and its interactions with claudin-1 and occludin. This cascade was inhibited by treatment with HPC or EC through δ-opioid receptor (DOR) binding and stimulation, thereby decreasing TNF-α–induced iNOS upregulation and CAV-1 endocytosis. The TJ functions were restored, leading to prevention of paracellular permeability, restoration of resistance of the ARPE-19 monolayer, and decreased ECM accumulation.

Conclusions.: The detrimental effects on TJs in ARPE-19 cells exposed to DM milieu occur through a CAV-1 S-nitrosylation–dependent endocytosis mechanism. High-polyphenol cocoa or EC exerts protective effects through DOR stimulation.

Introduction

Diabetic retinopathy (DR) is the most serious complication of diabetic (DM) eye disease and is one of the most common causes of irreversible blindness worldwide. ¹ The RPE is a monolayer of pigmented cells that separates the neural retina from a network of fenestrated vessels called the choriocapillaris, which serves as the major blood supply for the photoreceptors, and therefore the RPE constitutes the outer blood-retinal barrier (BRB). Impairment of the outer BRB is increasingly recognized to play an important role in the initiation and progression of early DR. ^2,3 The outer BRB is responsible for transport of nutrients, ions, and water; absorption of light and protection against photo-oxidation; the visual cycle; phagocytosis of shed photoreceptor membranes; and secretion of essential factors for preservation of the structural integrity of the retina. It also contributes to the immune-privileged status of the eye. ⁴  

Apart from these functions, the RPE stabilizes the ion composition in the subretinal space, which is crucial for the maintenance of photoreceptor excitability. Any disturbance in function of these cells necessarily has detrimental consequences for the retina. ² Defects in RPE function are well documented to underlie a number of sight-threatening conditions, such as AMD, ⁵ proliferative vitreoretinopathy, ⁶ and DR. ^2,3,7,8 The functioning of the RPE layer, measured by c-waves in ERGs, is substantially reduced in experimental DM models. ^9,10 However, direct data for the effects of high glucose (HG) or hyperglycemia on the tight junction (TJ) integrity and transport functions at the outer BRB are not completely understood. Tight junctions expressed in the outer BRB control fluid and solute entry into the retina, and this sealing function, which is essential to retinal homeostasis, is impaired in DR. ¹¹ Our previous work showed that ARPE-19 cells exposed to HG displayed a decrease in claudin-1 expression, ¹² but the mechanisms were not addressed. 

Tumor necrosis factor-α (TNF-α) was shown to induce a focal intrajunctional concentration of occludin followed by caveolin-1(CAV-1)–dependent endocytosis in the intestinal epithelial cells. ¹³ Caveolin-1, the main scaffolding protein of caveolae, consists of a lipophilic, hairpin-shaped, helical sequence embedded in the inner leaflet of the plasma membrane, together with both N- and C-terminal cytoplasmic domains. The N-terminus binds to signaling molecules that are required for CAV multimerization. Caveolae have been implicated in endocytosis, transcytosis, calcium signaling, and numerous other signal transduction events. An understanding of CAV trafficking and caveola formation is therefore crucial to understanding the possible roles of CAV and caveolae. ¹⁴  

Tight junctions are relatively cholesterol-rich, ¹⁵ and the cholesterol-binding protein, CAV-1, was identified as a component of TJ membrane microdomains more than a decade ago by Nusrat and coworkers. ¹⁶ Many studies have provided compelling evidence that CAV-1 is involved in regulating endothelial permeability. ¹⁷ Cavelon-1 can be precipitated and it binds independently to claudin-2 and occludin in MDCK II cells, suggesting a potential mechanism for selective retrieval of TJ components. ¹⁸ Thus, CAV-1 might have a more general role in regulating cell junctions, but its molecular regulation of epithelial cell adhesion and barrier function needs to be defined. 

The RPE contributes to the immune-privileged status of the eye as part of the blood-eye barrier and by the secretion of immunosuppressive/inflammatory factors inside the eye. ¹⁹ Rat RPE cells express inducible nitric oxide synthase (iNOS) and produce nitric oxide (NO) in response to inflammatory cytokines and activated T cells. ^20,21 In addition, 48 hours of HG exposure causes an increase in iNOS expression in ARPE-19 cells. ²² Recently, activation of opioid receptors, particularly the δ-opioid receptor (DOR), was demonstrated to block proinflammatory cytokines, such as TNF-α in the retina under ischemia/reperfusion conditions. ²³  

Epicatechin (EC), the predominant flavonoid present in dark chocolate, is a well-known antioxidant. ²⁴ Structure-activity relationships of flavonoids with opioid receptor ligands show binding activity in vitro. ^25,26 The purpose of this study was to evaluate the mechanism by which TJs are decreased in ARPE-19 cells under HG conditions and to determine whether cocoa powder, through its EC content, could prevent this effect. Our data revealed that HG promotes an increase in TNF-α levels and iNOS expression, accompanied by an increased production of NO and a resulting nitrosative imbalance. Consequently, CAV-1 is S-nitrosylated, thereby modulating the claudin-1/occludin CAV-1 interactions and CAV-1 traffic. Epicatechin, through its opioid receptor binding capacity, activates DOR and decreases TNF-α–induced CAV-1 endocytosis. As a result, the claudin-1 and occludin levels are restored in the TJs, leading to reestablishment of paracellular permeability and restoration of the resistance of ARPE-19 monolayers. 

Materials and Methods

Characteristics of Low- and High-Polyphenol Content Cocoa Powder

Cocoa powders with different amounts of polyphenol were provided by Barry Callebau. The composition of cocoa was the same in both preparations, with the only difference being in the amounts of polyphenol: 0.5% for low-polyphenol cocoa (LPC) and 60.5% for high-polyphenol cocoa (HPC). The quantitative analysis of HPC composition is shown in Supplementary Figure S1. 

The ARPE-19 Cell Line Culture

The human RPE cell line (ARPE-19) was obtained from the Federal University of Rio de Janeiro (RJCB Collection) at passage 28. Cell cultures at 70% to 90% confluence were serum starved with a fetal bovine serum (FBS) concentration of 1%, and then were exposed to normal glucose (5.5 mM = NG ²⁷ ) or HG (30 mM = HG); or HG plus 100 ng/mL LPC (HG+LPC) or HPC (HG+HPC) for 24 hours in the presence or absence of the NOS nonselective inhibitor L-NAME and the iNOS-specific inhibitor aminoguanidine (AG) (2 μM–2 mM), EC (0.38 nM–380 nM), naltrindole (Nalt) (10 nM–100 μM) (Sigma-Aldrich, St. Louis, MO, USA); S-nitrosoglutathione (GSNO) (10 nM–10μM) (synthesized at the Chemistry Institute, State University of Campinas [UNICAMP] as previously described ²⁸ ), and TNF-α (10-100ng/ml) (Calbiochem-Novabiochem, La Jolla, CA, USA). The potential interference of glucose osmolarity in ARPE-19 cells was checked by treating cells with 30 mM mannitol for 24 hours. The cytotoxicity of treatments on ARPE-19 cells was determined by a thiazolyl blue tetrazolium bromide (MTT) colorimetric assay. ²⁹ Concentrations that caused less than 10% cell toxicity were chosen for the experimental treatments (Supplementary Fig. S2). 

Immunofluorescence Assays

The immunofluorescence assays in ARPE-19 cells were performed as previously published. ³⁰ The cover glasses with fixed cells were incubated with the appropriate primary antibodies: anti-claudin-1, occludin, ZO-1 (Zymed Lab Gibco; Invitrogen, Carlsbad, CA, USA); FN (Calbiochem-Novabiochem); Col-IV (Southern Biotech, Birmingham, AL, USA), CAV-1 (Santa Cruz Biotechnologies, Santa Cruz, CA, USA), and Cellular retinaldehyde-binding protein (CRALBP; Thermo Scientific, Waltham, MA, USA) overnight at 4°C, and the appropriate secondary antibodies were applied for 1 hour at room temperature. 

Western Blotting Analysis

Western blotting was performed as previously described. ¹² Membranes were incubated overnight at 4°C with the appropriate primary antibodies: anti-claudin-1, occludin-1, ZO-1 (Zymed Lab Gibco; Invitrogen); FN (Calbiochem-Novabiochem); Col-IV (Southern Biotech) and iNOS (Cell Signaling Technology, USA) and the appropriate secondary antibodies were applied for 1 hour at room temperature. Equal loading and transfer were ascertained by reprobing the membranes for ß-actin. Exposed films were scanned with a densitometer (Bio-Rad, Hercules, CA, USA) and analyzed quantitatively with Multi-Analyst Macintosh Software for Image Analysis Systems (Hercules, CA, USA). The arbitrary unit of densitometry was transformed to fold increment. 

Measurement of Permeability to Dextran and Paracellular Epithelial Electrical Resistance

The integrity of TJs in cell culture is generally measured using transepithelial electrical resistance (TER) and/or paracellular tracer flux, ³⁰ as previously described with some modifications. The ARPE-19 cells were placed on a Transwell-Clear Polyester Membrane Insert (HTS, Costar; Corning, Inc., Corning, NY, USA). At day 5, a monolayer structure was observed and the complete medium was replaced by a medium with the treatments. The permeability of the RPE cells was determined by measuring the apical-to-basolateral movements of FITC dextran (40 kDa; 100 μg/mL) (Sigma-Aldrich, St. Louis, MO, USA). Sample fluorescence was measured at an excitation wavelength of 485 nm and emission wavelength of 528 nm with a microplate fluorescence reader (SynergyMx; Biotek, Winooski, VT, USA). For dextran permeability, at least three samples of each condition in four independent experiments were considered for area under the curve (AUC) analysis at 30, 60, 120, and 240 minutes. The AUC values were calculated by adding the areas under the graph between each pair of consecutive observations, as follows: (T₂ − T_l)(Y₁ + Y₂)/2. ³¹ To compare independent experiments, AUC values were calculated and expressed by fold increment. 

For TER measurement, 5-day monolayer ARPE-19 cells were obtained in the same Transwell membrane insert as described above. The TER values were obtained by using an epithelial voltmeter (MILLICELLERS; Millipore, Billerica, MA, USA) with an STX100C (for 24-well format) electrode (World Precision Instruments, Sarasota, FL, USA) according to the manufacturer's instructions. The TER measurements were calculated by subtracting the resistance of the filter without cell (background) from the values obtained from the filters containing RPE cells under different conditions expressed in ohms-cm² and corrected by the insert area. These values were expressed in fold increment. For TER measurements, three different samples with three reproducible measurements in each condition in four independent experiments were used for analysis after 24 hours of treatment. 

Measurement of Intracellular Reactive Oxygen Species (ROS) Production in Cells by Dichlorodihydrofluorescein Diacetate (DCF) and NO^• Formation by Diaminofluorescein Diacetate (DAF)

Cells were grown on a 96-well plate and incubated for 30 minutes in Hank's buffer containing 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA; 10 μM) for total intracellular ROS production, or DAF (DAF-2DA; 10 μM) for intracellular NO ^• levels. The fluorescence of DCF or DAF was measured using a fluorescence microplate reader (SynergyMx; Biotek) at excitation and emission wavelengths of 485 and 528 nm, respectively. The relative fluorescence values were corrected by the number of cells in each treatment. 

Immunoprecipitation

Cells were washed after treatments with ice-cold PBS and lysed directly in a buffer containing 100 mM Tris base, 10 mM sodium pyrophosphate, 100 mM sodium fluoride, 10 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 10 mM sodium orthovanadate, and 0.1 mg/mL aprotinin. After centrifugation at 14,481g for 15 minutes at 4°C, the supernatants were collected, Protein A-Sepharose (GE Healthcare Life Sciences) was added (10% of total volume), and the samples were incubated for 15 minutes at 4°C to remove nonspecific bindings. The samples were centrifuged again at 14,481g for 10 minutes at 4°C and the protein content was determined by the Bradford method. ³² Samples corresponding to 500-μg protein were incubated overnight with mouse anti-caveolin-1 (Santa Cruz Biotechnologies) or anti-S-Nitroso-Cysteine antibody (Sigma-Aldrich), followed by the addition of protein A-Sepharose for 2 hours. After centrifugation, the pellets were washed with PBS containing 1 mM EDTA. The immunoprecipitated samples were prepared under reducing or nonreducing conditions as necessary. Western blotting was performed as described above by using rabbit anti-claudin-1 or occludin for immunoprecipitated samples and subsequent incubation with appropriate secondary antibodies. 

Albumin Endocytosis Assay

The ARPE-19 cells were treated with BSA conjugated with Alexa 594 (50 μg/mL; Life Technologies, Gaithersburg, MD) as previously described, with some modifications. ³³ Cells on coverslips were washed with Hank's buffer and fixed with cold methanol (−20°C) for 10 minutes at room temperature. Cells were then blocked and incubated overnight at 4°C with rabbit polyclonal caveolin-1 antibody, and then incubated for 1 hour with the appropriate secondary antibody. 

Human TNF-α Levels

The quantitative levels of human TNF-α in the cell culture supernatant of ARPE-19 cells after 24 hours of treatments were determined with an Invitrogen Human Tumor Necrosis Factor-Alpha Ultra Sensitive (Hu TNF-α US) colorimetric ELISA Kit (Invitrogen), according to the manufacturer's instructions. 

Fluorometric Measurement of CAV-1 S-Nitrosylation by the Diaminonaphthalene (DAN) Assay

S-Nitrosylation of CAV-1 was measured as previously described, with some modifications. ³⁴ Samples were immunoprecipitated with mouse anti-caveolin-1. The immunoprecipitates were washed with PBS containing 1 mM EDTA. The pellets were resuspended in 500 μL PBS+EDTA and incubated with 200 μM HgCl₂ and 200 μM DAN for 30 minutes in the dark at room temperature, followed by the addition of NaOH at a final concentration of 0.1 N. In the DAN assay, NO from S-NO bonds on S-nitrosylated proteins was displaced by HgCl₂. A fluorescent triazole generated from the reaction between DAN and NO released from S-nitrosylated CAV-1 was then measured using a spectrofluorometer (model F-4500; Hitachi, Tokyo, Japan) operating at excitation and emission wavelengths of 375 and 450 nm, respectively, with slit widths of 2.5 nm. 

Primary Porcine RPE Isolation and Cell Culture

Primary porcine RPE (pRPE) cells were isolated and cultured as previously described, with some modifications. ^35–37 Two freshly slaughtered pig eyes (pig body weight approximately 20 kg, age ranging from 2 to 3 months) were cleaned of adjacent tissue and immersed briefly in antiseptic solution. The anterior part of the eye was removed, as were the lens, vitreous, and retina. Trypsin-EDTA was added to each eyecup and incubated for 1 hour at 37°C. The RPE cells were gently pipetted off the choroid. The cells were further purified by density centrifugation on a cushion composed of a single-density gradient (Percoll 40%; Sigma-Aldrich) prepared with 0.01 M Na₂PO₄ and 0.15 M NaCl (pH 7.4). After centrifugation at 500g for 3 minutes, the RPE cells were recovered as a pellet. Cells were seeded in a T25 flask and cultivated in a medium composed of Dulbecco's minimum essential medium supplemented with penicillin/streptomycin (1%), L-glutamine (300 μg/mL), amphotericin B (2.5 μg/mL), HEPES (11 mM), sodium bicarbonate (24 mM), and 15% FBS. Confluent RPE cells at passage 3 were used for all experiments. 

Statistical Analysis

Results are presented as fold increments to allow independent comparisons of experiments and are expressed as the means ± SD. The groups were compared by one-way ANOVA, followed by the Fisher protected least-significant difference test. StatView statistics software (SAS Institute, Inc., NC, USA) was used for all comparisons, with a significance value of P ≤ 0.05. 

Results

The presence of CRALBP can be used as a marker for RPE cells of the retina. The CRALBP plays an important role in the regeneration of 11-cis-retinal for use in rod visual pigments, such as opsin and rhodopsin; CRALBP immunofluorescence confirmed the cell phenotypes of ARPE-19 and pRPE cell cultures (Supplementary Fig. S3). 

High-Polyphenol Cocoa Preserved the Integrity of TJs of ARPE-19 Cells Exposed to Diabetic (DM) Milieu Conditions

We investigated the integrity of the monolayer of ARPE-19 cells by evaluating the expression of the TJ proteins claudin-1, occludin, and ZO-1. The ARPE-19 cells exposed to HG conditions for 24 hours showed a decrease in the expression of claudin-1 and occludin proteins when compared with the NG control (P ≤ 0.03). Immunofluorescence assays revealed less staining of these TJ proteins on the cell membrane under HG conditions. This decrease was prevented by HPC (P ≤ 0.03) but not by LPC (P > 0.2). Immunofluorescence measurements for cells exposed to mannitol, used as an osmotic control, were not significantly different from the NG values (P = 0.7) (Figs. 1A, 1B). 

Figure 1

 

Expression of claudin-1, occludin, and ZO-1 TJ proteins in ARPE-19 and primary pRPE cells under the HG condition and the effects of LPC and HPC treatments. The expression of TJs was evaluated after 24 hours in NG, HG, HG+HPC, or LPC (100 ng/mL) conditions. Mannitol (Man.) was used as an osmotic control. (A, B, C) Western blot for claudin-1, occludin, and ZO-1 expressions, respectively, in total ARPE-19 cell lysates. Equal loading and transfer were confirmed by reprobing the membranes for ß-actin. The arbitrary unit of densitometry was transformed to a fold increment. The bars represent mean ± SD. *P ≤ 0.03 versus NG; ^# P ≤ 0.03 versus HG. Confocal immunofluorescence images showing claudin-1, occludin, and ZO-1 immunolocalization. The marked TJs are shown in green (located on cell membrane) and the nucleus is indicated in red in the confocal microscopic field (magnification ×630). At least three independent experiments were performed for each assay. (D) Western blot and immunofluorescence for claudin-1 in pRPE cells. Equal loading and transfer were confirmed by reprobing the membranes for ß-actin.

In contrast, ZO-1 expression was increased in response to HG, compared with NG conditions (P = 0.03); the treatment with HPC significantly prevented this increase compared with HG (P = 0.001) (Fig. 1C). The LPC treatment showed a tendency toward a decrease compared with HG, but the difference did not reach statistical significance (P = 0.09). 

In addition, the expression of claudin-1 in pRPE cells (Fig. 1D) showed significant decreases compared with NG medium (P = 0.01) and in the presence of HPC, this alteration was prevented (P = 0.03 versus HG). 

High-Polyphenol Cocoa Prevented Extracellular Matrix Accumulation in Cells Exposed to HG Conditions

The RPE cells play a crucial role in the survival of photoreceptors, the choriocapillaris, and the choroid through the release of growth factors and production of extracellular matrix (ECM). ^4,38 The changes and disorganization of TJ proteins observed under HG conditions means that the ARPE-19 monolayer barrier must depend on ECM accumulation. After 24 hours under HG, the cells showed upregulated production of collagen IV and fibronectin proteins when compared with NG conditions (P ≤ 0.02), and the treatment with HPC prevented this alteration (P ≤ 0.05) (Fig. 2). Thus, the ARPE-19 cells exposed to HG for 24 hours showed a profound disturbance of TJ protein expression accompanied by accumulation of ECM proteins, but HPC prevented these abnormalities. Therefore, we assessed the functional features (permeability and resistance) of this monolayer barrier. 

Figure 2

 

High-polyphenol cocoa prevents the ECM accumulation in ARPE-19 cells under HG conditions. The expression of ECM materials was evaluated after treating with NG, HG, HG+HPC, or LPC (100 ng/mL) for 24 hours. (A, C) Western blot for fibronectin and collagen-IV expression, respectively, in total cell lysates. Equal loading and transfer were ascertained by reprobing the membranes for ß-actin. The arbitrary unit of densitometry was transformed to fold increments in relation to the NG treatment in each experiment to compare independent experiments. The bars represent mean ± SD. *P ≤ 0.02 versus NG; ^# P ≤ 0.05 versus HG. (B, D) Immunofluorescence images showing collagen-IV and fibronectin expression and localization. Collagen-IV is marked in red and the nuclei by 4′,6-diamidino-2-phenylindole (DAPI), and fibronectin is marked in red green and the nuclei with propidium iodide (PI). Both are localized on the membrane under a microscopic field (magnification ×630). At least three independent experiments were performed for each assay.

High-Polyphenol Cocoa Protected the Barrier Features of the ARPE-19 Cell Monolayer Under HG Conditions

We investigated whether the observed changes could influence the functional features of the ARPE-19 monolayer barrier by measuring its paracellular permeability (by assessing apical-basolateral movements of FITC-dextran) and its TER. The cells grown under HG conditions showed significantly lower dextran diffusion accompanied by higher TER, when compared with the NG condition (P = 0.05 and P = 0.008, respectively). Similar to the previous results, only HPC significantly protected the normal barrier function (P = 0.02 and P = 0.01 for dextran permeability and TER, respectively) (Fig. 3). The ARPE-19 cells showed no differences from the NG control when grown in mannitol (P = 0.7). 

Figure 3

 

High-polyphenol cocoa protection against RPE dysfunction in ARPE-19 cells under HG conditions. The permeability and TER of the RPE cells were determined after 24 hours in NG, Man., HG, HG+HPC, or LPC (100 μg/mL) conditions. Mannitol was used as an osmotic control. (A) Permeability was measured by the apical-to-basolateral movements of FITC dextran (40 kDa). Samples (200 μL) were collected from the basolateral side at 30, 60, 120, and 240 minutes after adding the molecules. (B) The bars represent fold increment (mean ± SD) of AUC of ARPE-19 cells' monolayer permeability at each condition. *P = 0.05 versus NG; ^# P = 0.02 versus HG. (C) Transepithelial electrical resistance measurements expressed in fold increment of resistance values (mean ± SD). *P = 0.008 versus NG; ^# P = 0.01 versus HG.

One possible mechanism by which the TJ claudin-1 and occludin expressions are decreased in the HG condition could involve a CAV-1 endocytosis phenomenon. In vitro studies have reported occludin endocytosis via macropinocytosis, clathrin-coated pits, and caveolae. ^39–41 We tested the potential involvement of endocytosis by examining the formation of CAV-1/claudin-1 and CAV-1/occludin complexes by immunocytochemistry. 

High-Polyphenol Cocoa Prevented the Formation of CAV-1/Claudin-1 and CAV-1/Occludin Complexes in ARPE-19 Cells Under HG Conditions

In the HG condition, a clear augmentation of CAV-1/claudin-1 and CAV-1/occludin binding occurred when compared with the NG condition (P ≤ 0.04); the presence of HPC suppressed these interactions when compared with the HG condition (P ≤ 0.04) (Figs. 4A, 4B). Immunolocalization of the CAV-1 protein on the ARPE-19 monolayer revealed a uniform distribution of CAV-1 in all parts of the cell, membrane, cytoplasm, and nucleus in the NG condition. Exposure of the cells to HG caused a massive internalization of CAV-1, which formed roundish structures in the cytoplasm compatible with caveosomes. These structures are associated with endocytosis. The LPC seemed to attenuate this process, and HPC clearly prevented this endocytosis (Fig. 4C). In agreement with this finding, pRPE cells also displayed a similar translocation of CAV-1 from the membrane to the nucleus and HPC treatment prevented this change (Fig. 4D). 

Figure 4

 

High-polyphenol cocoa prevented CAV-1/claudin-1 and occludin complexes and CAV-1 internalization. The expressions of CAV-1/claudin or occludin complexes and CAV-1 were evaluated after a 24-hour treatment with NG, HG, HG+HPC, or HG+LPC (100 ng/mL). (A, B) Immunoprecipitation of cell lysate with CAV-1 antibody incubated with claudin-1 or occludin antibodies, respectively. The CAV-1/claudin or occludin complex expressions were measured by Western blotting. Equal loading and transfer were ascertained by reprobing the membranes for CAV-1.The arbitrary unit of densitometry was transformed to fold increments in relation to the NG data in each experiment to compare independent experiments. The bars represent mean ± SD in both experiments. *P ≤ 0.04 versus NG; ^# P ≤ 0.04 versus HG. (C) Immunofluorescence images showing CAV-1 expression and localization. Caveolin-1 was marked in green and the nuclei with PI under the microscopic field (magnification ×630). At least three independent experiments were performed for each assay. (D) Immunofluorescence images showing CAV-1 expression and localization in pRPE cells marked in green (magnification ×630).

Collectively, these results suggest that the HG condition promotes a decrease in expression of TJ claudin-1 and occludin in ARPE-19 cells that is associated with caveosome formation. The protective effect of cocoa powder was at least partly dependent on the amounts of polyphenols present in the cocoa treatment. 

Epicatechin, the Main Polyphenol in Cocoa, Counteracts Nitrosative Stress and Prevents CAV-1 S-Nitrosylation

Tumor necrosis factor-α has been shown to stimulate occludin endocytosis via caveolin in intestinal epithelial cells, ⁴² but the mechanism by which it increases internalization and trafficking of CAV-1 is unclear. Post-translational modifications of CAV-1, such as ubiquitination ⁴³ and phosphorylation ³³ at the N-terminal near the scaffolding domain, are involved in the increased trafficking. Tumor necrosis factor-α is also known to increase NO production by regulating iNOS; therefore, we examined the direct modification of the assembly and mobility of CAV-1 by S-nitrosylation as a potential mechanism for regulation of vesicular trafficking in ARPE-19 cells. 

We first evaluated the levels of oxidative and nitrosative stress in our system. Cells exposed to HG showed an increase in total ROS production compared with the NG condition (P ≤ 0.0001). This effect was suppressed by LPC treatment (P = 0.01) and was strongly suppressed by HPC treatment when compared with HG (P ≤ 0.0001) or LPC conditions (P = 0.01). We tested the effect of EC, the most abundant polyphenol present in cocoa, using the corresponding percentages found in the LPC and HPC powders (0.15% and 12%, respectively). Only the higher EC concentration was effective in suppressing the HG-induced increase in ROS production (P = 0.0005) (Fig. 5A). This demonstrates that the high polyphenol levels present in the HPC counteracted the increase in ROS production in ARPE-19 cells exposed to HG. 

Figure 5

 

Epicatechin, the main phenolic compound in HPC, counteracts nitrosative stress. Measurements of intracellular ROS and NO production after treatment with NG, HG, HG+HPC, or LPC (100 ng/mL) for 24 hours; the EC content in the HG+EC treatment corresponds to the percentage of EC found in the LPC and HPC (0.25% and12%, respectively). (A, B) Total intracellular ROS production determined by the H₂DCFDA fluorescence method. (A) The effects of EC amounts on ROS production. Bars represent the mean ± SD of fluorescence units obtained with an ELISA reader and corrected by the number of cells at the end of each treatment. Mannitol was used as an osmotic control. *P ≤ 0.0001 versus NG; ^# P ≤ 0.01 versus HG group. (B) The NOS nonselective inhibitor L-NAME (2 mM) was used to test the role of EC in downregulating ROS production via the NO system. Bars represent the mean ± SD of fluorescence units obtained with an ELISA reader and corrected by the number of cells at the end of each treatment. *P ≤ 0.03 versus NG; ^# P ≤ 0.02 versus HG; ^† P = 0.05 versus HG+L-NAME+EC. (C) Intracellular NO production measured by the DAF-2DA method. Bars represent the mean ± SD of fluorescence units obtained in ELISA reader and corrected by the number of cells at the end of each treatment. *P = 0.008 versus NG; ^# P ≤ 0.005 versus HG. (D) Western blot of iNOS expression in total cell lysates. Equal loading and transfer were confirmed by reprobing the membranes for ß-actin. The arbitrary unit of densitometry was transformed to fold increments in relation to the NG treatment in each experiment to compare independent experiments. The bars represent mean ± SD. *P = 0.003 versus NG; ^# P = 0.007 versus HG. At least three independent experiments were performed for each assay.

The NO pathways potentially involved in ROS production in HG conditions were then investigated by the colorimetric method for detection of intracellular ROS, H₂DCFDA. The presence of HPC or the unspecific NOS blocker, Nω-Nitro-L-arginine methyl ester hydrochloride (L-NAME), prevented the increase in ROS levels seen with HG (P ≤ 0.02). Concomitant treatments of L-NAME+HPC or L-NAME+EC under HG conditions showed no additive effects. The significant difference observed between HG+L-NAME+HPC and HG+L-NAME+EC (P = 0.05) suggests that other cocoa compounds might exert a protective effect through different pathways (Fig. 5B). 

We then evaluated the production of intracellular NO by the DAF-2DA method and the specific effect of EC on this production. We observed an increase in NO production in cells exposed to HG compared with the NG condition (P = 0.008). Treatment with HPC or the corresponding EC percentage prevented this increase (P ≤ 0.005) (Fig. 5C), suggesting a specific effect of EC on the NO system. We also confirmed the source of NO by assessing iNOS expression. As expected, iNOS was upregulated under HG conditions (P = 0.003), and this upregulation was abolished in the presence of HPC (P = 0.007) (Fig. 5D). 

The presence of nitrosative stress induced by iNOS in our system prompted us to investigate whether CAV-1 was nitrosylated by the excessive amounts of NO. The nitrosylation of CAV-1 was assessed by immunoprecipitation experiments and analyzed by spectrofluorometry with the DAN method and Western blots (Figs. 6A, 6B). Figure 6A shows the NO release from S-nitrosylated CAV-1 detected by DAN. Cells incubated with an NO donor (NG+GSNO) as a positive control or in HG showed a clear induction of S-nitrosylation of the CAV-1 molecule (P < 0.0001). Treatment with HPC, EC, or the iNOS inhibitor AG equally prevented this process (P ≤ 0.002). The combination of AG and HPC showed an additional effect compared with HPC alone (P = 0.05), suggesting that other cocoa compounds might act in this pathway. However, the HG+AG+EC treatment did not show any additional effects compared with HG+EC (P = 0.6), indicating that the main action of EC is in the prevention of CAV-1 S-nitrosylation via iNOS downregulation (Fig. 6A). 

Figure 6

 

Epicatechin counteracts nitrosative stress and prevents S-nitrosylation of CAV-1. Measurement of CAV-1 S-nitrosylation after treatment with NG, HG, HG+HPC (100 ng/mL), or HG+EC (12 ng/mL) for 24 hours. The specific iNOS blocker aminoguanidine (AG) was used at a concentration of 2 mM to assess the protective effects of EC via downregulation of iNOS. (A) Fluorometric measurement of CAV-1 S-nitrosylation using the diaminonaphthalene (DAN) assay. S-nitrosoglutathione (1 μM) was used as a positive control for the induction of CAV-1 S-nitrosylation. The fluorescence units obtained via ELISA reader were transformed to fold increments in relation to the media of NG in each experiment to compare independent experiments. Bars represent the mean ± SD. *P < 0.0001 versus NG conditions; ^# P ≤ 0.002 versus HG treatment; ^† P = 0.05 versus HG+HPC. (B) Expression of S-nitrosylated CAV-1. Western blot of cell lysate immune precipitated with S-Nitroso-Cysteine (SNO-Cys) antibody and immunoblotted for anti-CAV-1. Equal loading and transfer were ascertained by reprobing the membranes for SNO-Cys antibody. The arbitrary unit of densitometry was transformed to fold increments in relation to the NG treatment in each experiment to compare independent experiments. The bars represent mean ± SD in both experiments. *P < 0.0001 versus NG; ^# P ≤ 0.002 versus HG; ^† P = 0.04 versus HG+HPC. At least three independent experiments were performed for each assay.

In line with the fluorometric assays, HG treatment caused a significantly increased expression of S-nitrosylated CAV-1 compared with the NG condition (P < 0.0001). The HPC and EC were equally effective at suppressing this expression when compared with HG (P = 0.5). The combined HG+AG+HPC treatment again indicated an additional effect when compared with HG+HPC (P = 0.04), and no difference was noted between the HG+EC and HG+AG+EC treatments (P = 0.2) (Fig. 6B). 

We next investigated whether the nitrosylation of the CAV-1 molecule interferes in the interaction between CAV-1 and the TJs and the CAV-1 internalization by studying the endocytosis of CAV-1. 

Epicatechin Prevents CAV-1 Endocytosis and This Effect is Dependent on TNF-α–iNOS Upregulation Via the δ-Opioid Receptor (DOR)

We measured the TNF-α levels in the supernatant of these cells and observed an increase in cells exposed to HG compared with NG (P = 0.002), and both HPC and EC were equally effective at preventing this increase (P ≤ 0.03) (Fig. 7A). Activation of opioid receptors is known to reduce TNF-α production in the retina model of ischemic/reperfusion ²³ ; therefore, we used the DOR blocker naltrindole (Nalt) to test whether opioid receptors were involved in the HG effects on ARPE-19 cells. The addition of Nalt abolished the effects of HPC or EC compared with controls (P ≤ 0.0008) (Fig. 7A), indicating that EC modulates TNF-α via DOR receptor. 

Figure 7

 

Epicatechin prevents TNFα-iNOS upregulation through DOR. (A) Measurement of TNF-α levels in the supernatant of ARPE-19 cells by ELISA. We used a specific DOR blocker, naltrindole (Nalt), at a concentration of 10 μM to investigate whether the effects of EC are dependent on the DOR receptor. The absorbance values were corrected for protein concentration and expressed as pg/mL/μg. *P = 0.002 versus NG; ^# P ≤ 0.03 versus HG; ^† P ≤ 0.0008 versus HG+HPC and HG+EC, respectively. (B) Western blot for iNOS expression in total cell lysates. Tumor necrosis factor-α, at a concentration of 40 ng/mL, was used to induce iNOS expression. Equal loading and transfer were confirmed by reprobing the membranes for ß-actin. The arbitrary units of densitometry were transformed to fold increments in relation to the NG treatment in each experiment to compare independent experiments. The bars represent mean ± SD. *P ≤ 0.01 versus NG; ^# P ≤ 0.02 versus; ^† P = 0.04 versus HG+EC. At least three independent experiments were performed for each assay.

We then assessed whether TNF-α could induce iNOS upregulation in the ARPE-19 cells and if the DOR was involved in this modulation. As expected, TNF-α treatment increased iNOS expression compared with the NG condition (P = 0.0006) and treatment with EC prevented this stimulation (P = 0.001). The addition of Nalt abrogated this effect when compared with the NG+TNF-α+EC treatment (P = 0.0002). As expected, ARPE-19 cells under HG conditions showed increased iNOS expression (P = 0.008 versus NG conditions), and the addition of EC suppressed this alteration, restoring normal iNOS levels (P = 0.02). The DOR blocker Nalt abrogated the protective effect of EC when compared with the HG+TNF-α+EC condition (P = 0.04) (Fig. 7B). 

We tested whether EC prevents CAV-1 endocytosis via the DOR receptor by using the albumin endocytosis assay. A low level of endocytosis of CAV-1 was observed in the NG condition, but endocytosis was greatly increased under the HG condition. Treatment with HPC or EC decreased CAV-1 endocytosis, and addition of Nalt prevented this decrease. These findings indicate that the protective effect of EC occurs at least in part via the DOR receptor, thus preventing transcellular caveolin trafficking (Fig. 8). 

Figure 8

 

Epicatechin prevents CAV-1 endocytosis by blocking the DOR. Albumin (BSA) endocytosis assay. Immunofluorescence images showing endocytosis of BSA conjugated with Alexa 594 (red), CAV-1 localization (green), and the nuclei with DAPI (blue) (magnification ×630). At least three independent experiments were performed in each assay. HG+EC (%LPC), high glucose + epicatechin 0.25 ng/mL; HG+EC (%HPC), high glucose + epicatechin 12 ng/mL.

Epicatechin Prevented TJ Downregulation and CAV-1 Internalization in ARPE-19 and pRPE Cells

We confirmed that the protective effects of HPC in cells exposed to HG conditions occurred mainly through EC by treating pRPE and ARPE-19 cells with EC (12 ng/mL) under the HG condition and looking for end points (tight junctions and CAV-1). Supplementary Figure S4 shows that EC treatment prevented the decrease in claudin-1 and occludin expressions and CAV-1 internalization in ARPE-19 cells (P ≤ 0.02) (Supplementary Figs. S4A–C) and claudin-1 expression and CAV-1 translocation in pRPE cells (P = 0.05) (Supplementary Figs. S4D–F). 

Collectively, these data strongly suggest that the beneficial effects of polyphenols, and especially EC, on ARPE-19 cells in the HG condition occur via DOR binding and stimulation, which decreases iNOS–TNF-α–dependent upregulation. As a consequence, the intracellular NO production is ameliorated and CAV-1 nitrosylation is prevented, thereby restoring normal CAV-1 trafficking. 

Discussion

This study provides evidence that the process by which TJ claudin-1 and occludin decrease in ARPE-19 cell and primary pRPE monolayers exposed to DM milieu conditions is through the CAV-1 endocytosis associated with a profound imbalance in paracellular resistance and permeability. An increase in ZO-1 expression and ECM accumulation also was observed in the HG condition. The CAV-1 endocytosis process is dependent on stimulation by TNF-α, which in turn upregulates iNOS expression. The increased NO production then results in S-nitrosylation of CAV-1. Treatment with HPC or EC abolished all these effects through direct interaction with the DOR of the ARPE-19 cells. These results are the first, to the best of our knowledge, to show a protective role for EC in preventing nitrosative posttranslational modification of CAV-1 in ARPE-19 and pRPE cells and maintenance of the RPE barrier properties following exposure to HG conditions. 

Most of the research on the physiopathology of DR has focused on the impairment of the neuroretina and the breakdown of the inner BRB. By contrast, the effects of DM on the RPE, and in particular on its secretory and transport activity, have received less attention. In one direction, the RPE transports electrolytes and water from the subretinal space to the choroid, and in the opposite direction, it transports glucose, retinol, ascorbic acid, and fatty acids from the choriocapillaris to the photoreceptors. Transport in both directions was changed under HG conditions. ² Glucose transport by the RPE is conducted via the large numbers of glucose transporters in the apical and basolateral membranes. Both GLUT1 and GLUT3 are highly expressed in the RPE. ^44–46 Recently, HG has been shown to promote alterations in transport by downregulation of GLUT-1 ⁴⁷ ; alterations in transport of retinol due to a downregulation of the interstitial retinol binding protein; and impairment of the transport of ascorbic acid, thereby limiting the antioxidant defenses of the RPE. ^48,49  

A large amount of water is produced in the retinal tissue, mainly as a consequence of the large metabolic turnover in neurons and photoreceptors. The Na⁺-K⁺-ATPase, located in the apical membrane, provides energy for the transepithelial transport. ⁵⁰ Constant elimination of water from the subretinal space produces an adhesive force between the retina and the RPE that is lost by inhibition of Na⁺-K⁺-ATPase (e.g., by ouabain). ⁵¹ Cultured bovine RPE cells exposed to high glucose conditions show a loss of Na⁺/K⁺ ATPase function, which in turn decreases the permeability. ⁵²  

The TJs constitute the barrier between the subretinal space and the choriocapillaris. The literature contains contradictory data concerning the TER of ARPE-19 cells under high glucose conditions. For example, Villarroel and colleagues ⁵³ reported that 3 weeks of exposure to high glucose increased the TER and decreased the permeability of ARPE-19 cells. In addition, this reduction in permeability was unrelated to claudin-1 mRNA overexpression. On the other hand, Trudeau et al. ⁵⁴ showed increased permeability of ARPE-19 cells exposed for 18 days to 25 mM glucose when treated with IL-1β. This in vitro model indicates that inflammation is playing a pivotal role in the increasing permeability of the RPE barrier. However, ARPE-19 cells exposed to endoplasmic reticulum stress induced by tunicamycin or thapsigargin showed a significant increase in ZO-1, occludin, and claudin-1 associated with an increase in TER. ⁵⁵ In addition, hyperglycemia could impair the transport of water from the subretinal space to the choriocapillaris and, consequently, might contribute to diabetic macular edema development. ² On the other hand, the direct visualization of macromolecules leaked through the outer BRB by a microscopic imaging assay demonstrated the importance of outer BRB breakdown in rats made diabetic by streptozotocin treatment. ³  

Nitrosative stress is an early event in the pathogenesis of DR. ⁵⁶ In this present work, the iNOS activation increased NO levels and posttranslationally altered CAV-1 molecules, thereby increasing the communication with claudin-1 and occludin. Recently, a pivotal role for CAV-1–dependent occludin endocytosis induced by TNF-α has been proposed for the regulation of TJs in intestine epithelial cells. ¹³ Other work showed that phosphorylation of CAV-1 increases its association with vascular endothelial (VE)-cadherin/catenin complexes in response to the proinflammatory mediator thrombin. This weakens the association of catenin with VE-cadherin and the junction-associated actin filaments are lost, thereby compromising the barrier function. ⁵⁷  

Abnormalities in gene expression of CAV-1 have been linked to DR, ⁵⁸ but the possible involvement of the CAV-1/caveolae in the outer retina needs to be better understood. The caveolae of RPE cells have a unique bipolar distribution, ⁵⁹ but their functions in either the apical or basolateral RPE membrane domains have not been elucidated. Previous work showed that ablation of CAV-1 resulted in reduced inner and outer retinal functions, whereas Cav-1 knockout retinas also displayed unusually tight adhesion with the RPE, suggesting alterations in outer retinal fluid homeostasis. These findings demonstrate that the reduced retinal function resulting from CAV-1 ablation involves impairment of subretinal and/or RPE ion/fluid homeostasis. ⁶⁰ Posttranslational modifications of CAV-1, such as ubiquitination ⁶¹ and phosphorylation ³³ of the N-terminal near the scaffolding domain, resulted in exacerbated trafficking. In line with our present data, these authors demonstrated that CAV-1 SNO is an important regulatory mechanism controlling caveolar trafficking in endothelial cells. 

The opioid receptor family comprises three members, the μ-, δ-, and k-opioid receptors, which respond to classical opioid alkaloids, such as morphine and heroin, as well as to endogenous peptide ligands like endorphins. They belong to the G-protein-coupled receptor superfamily, and are recognized as excellent therapeutic targets for pain control. ⁶² Activation of one or more opioid receptors by morphine can reduce ischemic/reperfusion injury by the suppression of TNF-α production in the retina. Naloxone, an opioid antagonist, can reverse the morphine-induced suppression of TNF-α production in vitro. ²³ Epicatechin, the predominant flavonoid component present in cocoa and dark chocolate, is a well-known antioxidant associated with a lower risk of stroke and heart failure. ^63–65 Moreover, EC-induced cardiac protection has shown a dependence on DOR stimulation. ²⁶ In the present study, we demonstrated that the increase in TNF-α levels in ARPE-19 cells exposed to HG is abolished when the cells were treated with HPC or the corresponding amount of EC, which effectively neutralized the HG effect; this action was abrogated in the presence of Nalt, a DOR blocker. This set of experiments clearly demonstrated that EC protects the ARPE-19 monolayer barrier/permeability through stimulation of DOR, thereby modulating TNF-α action. The crystal structure of the mouse DOR, bound to the subtype-selective antagonist Nalt, ⁶⁶ has indicated that blocking the DOR with oral administration of Nalt resulted in a decrease of the cardiac protective effect of EC on mitochondrial structure in mice. ⁶⁷  

In conclusion, we identified EC as a negative regulator of the CAV-1 nitrosylation that occurs in retinal pigmented epithelium cells under DM milieu conditions due to activation of DOR. Caveolin-1 plays an important role in major diseases, such as cancer, ⁶⁸ atherosclerosis, ⁶⁹ DM complications, ⁷⁰ and inflammation ⁷¹ ; therefore, our findings might provide insights into the regulation of claudin-1 and occludin by CAV-1 internalization in RPE cells exposed to HG conditions, as well as in other pathological conditions. 

Supplementary Materials

pdf S1 

Acknowledgments

The authors thank the staff of the Life Sciences Core Facility from the University of Campinas for support with confocal microscopy. The authors thank the personnel from the Renal Pathophysiology Laboratory, School of Medical Sciences, and University of Campinas for their invaluable help with this work. 

Supported by the Fundação de Amparo à Pesquisa do Estado de São Paulo (Grants 2008/57560-0 and 2011/06719-1). MABR received a scholarship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior. 

Disclosure: M.A.B. Rosales, None; K.C. Silva, None; D.A. Duarte, None; F.A. Rossato, None; J.B. Lopes de Faria, None; J.M. Lopes de Faria, None 

Footnotes

 MABR and KCS contributed equally to the work presented here and should therefore be regarded as equivalent authors.