Merlin Regulates Epithelial-to-Mesenchymal Transition of ARPE-19 Cells via TAK1-p38MAPK-Mediated Activation

Abstract

Purpose.: The purpose of this study was to investigate the function of merlin, which is a binding partner of the hyaluronan receptor CD44, during the epithelial-to-mesenchymal transition (EMT) of human retinal pigment epithelium (ARPE-19) cells.

Methods.: Human retinal pigment epithelium cells were stimulated with TNF-α and treated using epithelial-to-mesenchymal-transition inhibitors (a dynamin inhibitor or transforming growth factor-β-activated kinase 1 [TAK1] inhibitor). Levels of protein expression were assessed by immunoblot analysis, and localization of the relevant proteins was determined by immunofluorescence microscopy. Cell proliferation was evaluated by bromodeoxyuridine (BrdU) incorporation assays. All experiments were performed in serum-free medium.

Results.: Tumor necrosis factor-α treatments downregulated the expression of merlin and led to the dissociation of CD44 and merlin. The ezrin-radixin-moesin (ERM) proteins were phosphorylated, and hyaluronan endocytosis was accelerated in merlin small interfering RNA (siMerlin)-transfected cells. Treatment with the endocytosis inhibitor dynasore blocked hyaluronan endocytosis, whereas treatment with TNF-α induced mesenchymal phenotypes and downregulation of merlin. Additionally, siMerlin transfection promoted p38MAPK phosphorylation, which was inhibited not only by TAK1 inhibitor treatment but also by TAK1 small interfering RNA (siRNA, siTAK1) transfection. The increased level of BrdU incorporation in siMerlin cells was reduced by additional siTAK1 transfection. Furthermore, TNF-α–induced mesenchymal differentiation and high motility were also inhibited by TAK1 inhibitor treatment and by siTAK1 transfection.

Conclusions.: Our findings demonstrated that merlin exerts inhibitory effects on TNF-α–induced EMT by regulating hyaluronan endocytosis and the TAK1-p38MAPK signaling pathway. The proliferative and mesenchymal characteristics of RPE cells play important roles in the development of intraocular fibrotic disorders, such as proliferative vitreoretinopathy (PVR), and our findings provide new therapeutic strategies to prevent the development of PVR.

Proliferative vitreoretinopathy (PVR) is a severe type of intraocular fibrosis caused by traumatic injury or by rhegmatogenous retinal detachment, which often results in serious vision deterioration.^1,2 Retinal pigment epithelial (RPE) cells, as well as glial cells, fibroblasts, macrophages, and astrocytes, are known to contribute to the development of this proliferative disorder.^3,4 Abnormally activated RPE cells not only contribute cellular component to PVR but also affect expression of cytokines and growth factors, including TGF-β and TNF-α.^1,3–5 Previous studies found that TGF-β and TNF-α induced the epithelial-to-mesenchymal transition (EMT) of RPE cells.⁵ 

Epithelial-to-mesenchymal transition is thought to be associated with fibrosis, wound healing, and cancer cell metastasis. During the EMT process, epithelial cells lose their epithelial characteristics and acquire the characteristics of a mesenchymal phenotype, namely, a morphological change to the spindle shape and a highly motile capacity, as well as the overexpression of extracellular matrix components and the expression of mesenchymal cell markers.⁶ During chronic inflammation, epithelial cells are continuously exposed to various cytokines (including TNF-α), growth factors, and reactive oxygen species, which leads to the induction of EMT.⁷ These transdifferentiated epithelial cells produce excess amounts of extracellular matrix components, which contribute to the development of fibrosis.⁸ 

We previously reported that TNF-α– and TGF-β–induced EMT are essential for the production of hyaluronan and anchor proteins and for the activation of the ezrin-radixin-moesin (ERM) proteins following the formation of hyaluronan-CD44-moesin complexes.⁵ Recently, merlin, which is a member of the ERM protein family, was shown to inhibit TGF-β signaling.⁹ Moreover, TGF-β is a well-known contributor to the development of PVR due to the proliferation of RPE cells.^10,11 Thus, we investigated the interactions between merlin and associated molecules in human RPE cells to elucidate the role of merlin in the inhibition and/or elimination of the intraocular proliferation of RPE cells. 

However, the role of merlin in EMT remains unclear. The present study demonstrated that merlin-regulated mechanisms affect the interaction between CD44 and phosphorylated ERM, as well as transforming growth factor-β-activated kinase 1 (TAK1)-dependent phenomena such as EMT and proliferation. 

Materials and Methods

Cell Culture

Human retinal pigment epithelium (ARPE-19) cells were obtained from American Type Culture Collection (Manassas, VA, USA). The cells were cultured in 5% CO₂ at 37°C in Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 nutrient mixture (Sigma-Aldrich Corp., St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS). All experiments were performed in serum-free medium. 

Isolation of Mouse RPE

All animal experiments were conducted in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animal in Ophthalmic and Vision Research and were approved by the experimental animal ethics committee at Kumamoto University. We used a protocol modified from previous studies^12,13 for mouse RPE isolation. Briefly, mice were killed between 3 and 4 days of age. Enucleated eyes were washed twice in DMEM containing high glucose (Sigma-Aldrich Corp.) supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 0.5 μg/mL amphotericin B growth medium (GM) in 5% CO₂ at 37°C. Eye globes were incubated for 45 minutes at 37°C in 2% dispase (BD Biosciences, Bedford, MA, USA) in GM. After the globes were washed and corneas removed, the posterior eyecups containing the lens and the retina were placed in fresh GM and incubated for 20 minutes at 37°C. The lens was then removed, and the neural retina was peeled off. The RPE layer was gently brushed from Bruch's membrane with a sterile brush, centrifuged at 1500g for 5 minutes and resuspended in GM. The isolated RPE sheets were seeded onto laminin-coated plates (Wako, Osaka, Japan) and incubated in 5% CO₂ at 37°C. The cells were cultured in antibiotic-free medium for 24 hours before experiments were conducted. Primary murine RPE cells were used during passages 3 to 5, and experiments were replicated in at least three independent cell lines. 

Antibodies and Chemicals

We used the commercially available antibodies and chemicals human recombinant TNF-α and TGF-β2 and mouse recombinant TNF-α (R&D Systems, Minneapolis, MN, USA); interleukin (IL)-6 and IL-8 (Gibco, Grand Island, NY, USA), and monocyte chemoattractant protein (MCP)-1 (Wako). Antibodies directed against merlin, TAK1, nuclear factor κB (NF-κB) and Na⁺/K⁺-ATPase α were purchased from Santa Cruz Biotechnology (Dallas, TX, USA); anti-CD44 (F10-44-2) for immunoprecipitation assay was purchased from Abcam (Cambridge, UK); and anti-CD44 (IM-7.8.1) for immunoblot analysis and for immunofluorescence staining was purchased from BioLegend (San Diego, CA, USA). Anti-ezrin, anti-moesin, and anti-fibronectin antibodies were obtained from Abcam; anti-phospho-ERM antibody was purchased from Millipore (Billerica, MA, USA); antibodies directed against phospho-p38MAPK, p38MAPK, phospho-p65, and p65 were purchased from Cell Signaling Technology (Danvers, MA, USA); and antibodies directed against α-smooth muscle actin (α-SMA) and β-actin were obtained from Sigma-Aldrich Corp. Dynasore hydrate was purchased from Sigma-Aldrich Corp., and (5Z)-7-oxozeaenol was purchased from Tocris Bioscience (Bristol, UK). 

Enzyme-Linked Immunosorbent Assay (ELISA)

Cells were either stimulated with or without TNF-α for 4 days in serum-free medium or initially stimulated with TNF-α for 2 hours. Next, the culture medium was replaced with new serum-free medium without TNF-α for 4 days. A 2-hour treatment with TNF-α was used as a positive control. The culture medium was collected, and the concentration of TNF-α was measured according to the manufacturer's protocol (R&D Systems). 

Reverse Transcription-Polymerase Chain Reaction (RT-PCR)

Reverse transcription–PCR was performed as previously described⁵ using the primers and PCR cycle protocol for TNF-α reported by Bates et al.¹⁴ Forward primer 5′-CGAGTGACAAGCCTGTAGCC-3′ and reverse primer 5′-GTTGACCTTGGTCTGGTAGG-3′ were used for TNF-α; and forward primer 5′-TCCCTGGAGAAGAGCTACGAGC-3′ and reverse primer 5′-GTAGTTTCGTGGATGCCACAGG-3′ were used for β-actin. 

Small Interfering RNA (siRNA) Transfection

We designed the following two human merlin siRNA sequences 5′-CUACUUUGCAAUCCGGAAUdTdT-3′ (sequence 1) and 5′-GGAGUUUACUAUUAAACCAdTdT-3′ (sequence 2). For mouse small interfering RNA (siRNA), we used sequences 5′-UACCGAGCUUCGACAUUAUUG-3′ (Ms1) and 5′-GGAGUUUACUAUUAAACCATT-3′ (Ms2). The control siRNA sequence used for mouse was 5′-AAUCCGGUUGCAUAGUUCAUG-3′ (siRNA scrambled M1).¹⁵ The human TAK1 siRNA sequence was 5′-UGGCUUAUCUUACACUGGGAdTdT-3′.¹⁶ Small interfering RNA duplexes were synthesized by Japan BioServices (Saitama, Japan). Cells were transfected with the annealed siRNAs by using Lipofectamine RNAiMAX reagent (Life Technologies, Grand Island, NY, USA) according to the manufacturer's protocol. Double-stranded RNA targeting the luciferase gene (GL-2)⁵ was used as the control. 

Immunoprecipitation and Immunoblot Analysis

The analysis of protein expression by immunoprecipitation followed by immunoblot analysis was performed as previously described.⁵ Briefly, 48 hours after TNF-α stimulation, the cells were lysed on ice for 20 minutes, and then lysates were incubated for 2 hours with an antibody directed against CD44 or a mouse immunoglobulin G (IgG; isotype control; Medical and Biological Laboratories, Nagoya, Japan). Protein A-Sepharose beads (GE Healthcare, Little Chalfont, Buckinghamshire, UK) were added to the lysates, and the mixtures were incubated for an additional 1.5 hours. The beads were isolated by centrifugation, and the bead-bound proteins were subjected to immunoblot analysis. Gel electrophoresis was performed, and the proteins were transferred onto polyvinylidene difluoride membranes (Bio-Rad, Hercules, CA, USA). The membranes were incubated with primary antibodies overnight at 4°C. After being washed, the membranes were incubated with horseradish peroxidase–conjugated secondary antibodies for 60 minutes at room temperature, and the labeled bands were visualized using enhanced chemiluminescence (Amersham Biosciences/GE Healthcare, Tokyo, Japan). 

Cell Surface Protein Isolation

We used the Pierce cell surface protein isolation kit (ThermoFisher Scientific, Rockford, IL, USA) to isolate cell surface proteins. Briefly, after biotinylation of cell surface proteins for 30 minutes at 4°C, cells were lysed and incubated with NeutrAvidin agarose (ThermoFisher Scientific) for 60 minutes at room temperature. The samples were then incubated with SDS-PAGE sample buffer for 60 minutes at room temperature, and the biotinylated proteins were eluted to the buffer by heating. The samples were then subjected to Western blot analysis. 

Immunofluorescence Microscopy

Cells were fixed in 4% paraformaldehyde, permeabilized with 0.2% Triton X-100, and incubated for 60 minutes with the primary antibodies at room temperature. After being washed, the cells were incubated for 60 minutes with Alexa Fluor 488- or Alexa Fluor 594-conjugated secondary antibodies (Invitrogen-Molecular Probes, Eugene, OR, USA). The cells were washed and mounted using Vectashield (Vector Laboratories, Burlingame, CA, USA) or Vectashield containing 4′,6-diamidino-2-phenylindole (DAPI) for nuclear staining. Immunohistochemical analysis was performed using a confocal microscope (FluoView; Olympus, Tokyo, Japan) or a fluorescence microscope (BZ-X710; Keyence, Osaka, Japan). 

Evaluating Internalization of Fluorescein-Labeled Hyaluronan (Fl-HA)

The internalization of Fl-HA was evaluated as previously described,¹⁷ with some modifications. Briefly, cells were seeded on poly-l-lysine-coated coverslips (Matsunami Glass Ind., Ltd., Osaka, Japan) and transfected with siRNAs. After 24 hours of serum starvation, the cells were incubated with 20 μg/mL Fl-HA (Calbiochem, San Diego, CA, USA) at 37°C for 3 hours in serum-free medium. After the cells were washed and fixed, the Fl-HA-positive vesicles were imaged using a fluorescence microscope (BZ-X710) and counted using BZ-X Analyzer software (Keyence). 

Cell Growth

For the bromodeoxyuridine (BrdU; Sigma-Aldrich Corp.) incorporation assays, cells were cultured in the presence or absence of TNF-α with or without (5Z)-7-oxozeaenol for 24 hours. Bromodeoxyuridine was added to a final concentration of 10 μM, and the cells were incubated for an additional 4 hours. The cells were fixed in 4% paraformaldehyde for 15 minutes, permeabilized with 0.2% Triton X-100 for 5 minutes, and then treated with 2 N HCl for 90 minutes. Subsequently, the cells were incubated with an anti-BrdU antibody for 60 minutes and then washed with PBS before incubation for 60 minutes with an Alexa 488-conjugated secondary antibody. The nuclei were visualized by DAPI staining. The cells were subjected to fluorescence microscopic analysis (Fluoview [Olympus] or BZ-X710 [Keyence]). 

Wound-Healing Assay

The in vitro wound-healing assay was performed as previously described.⁵ Briefly, cells were cultured with or without TNF-α and with or without (5Z)-7-oxozeaenol for 24 hours in serum-free medium, after which the confluent monolayer was scratched using a sterile plastic tip in serum-free medium. The migration of cells into the wound was imaged at each indicated time point using light microscopy (Olympus). 

Results

Western Blot Analysis of Merlin Expression in Cytokine-Stimulated ARPE-19 Cells

First, we used Western blots to analyze merlin expression in human ARPE-19 cells stimulated with TNF-α, TGF-β2, IL-6, IL-8, or MCP-1; elevated levels of these factors have been observed in vitreous samples obtained from patients with PVR or proliferative diabetic retinopathy.^18,19 Interestingly, TNF-α was the only cytokine to reduce the merlin protein level, whereas the other cytokines did not change the merlin protein level (Fig. 1A). We conducted a Western blot analysis of merlin expression to elucidate time-dependent changes in merlin protein levels after TNF-α stimulation; TGF-β2 was used as the control. Merlin expression was reduced in a time-dependent manner during the 4 days following the addition of TNF-α (at 10 ng/mL) alone to the culture medium (Fig. 1B). In addition, TNF-α reduced merlin expression in primary murine RPE cells in addition to ARPE-19 cells (Fig. 1C). Furthermore, withdrawal of TNF-α did not reduce merlin expression (Fig. 1D). Next, we assessed the activation of the TNF-α signaling pathway by phosphorylation and nuclear translocation of the p65 subunit of NF-κB. Western blot analysis showed that p65 was phosphorylated (Fig. 1D), and fluorescence microscopic analysis also showed the nuclear translocation of p65 (Fig. 1E) after 4 days of TNF-α stimulation. Neither the phosphorylation nor the nuclear translocation of p65 were observed after withdrawal of TNF-α (Figs. 1D, 1E). ELISA revealed that TNF-α was present in the culture medium of cells that had been treated for 4 days; however, the concentration of TNF-α was not detectable after withdrawal of TNF-α (Fig. 1F). Reverse transcription–PCR showed that TNF-α message production was increased by TNF-α stimulation (Fig. 1G). 

Figure 1

 

Tumor necrosis factor-α treatment suppresses merlin expression in ARPE-19 cells. (A) Human retinal pigment epithelium cells were stimulated with TNF-α (10 ng/mL), TGF-β2 (5 ng/mL), IL-6 (10 ng/mL), IL-8 (50 ng/mL), or MCP-1 (100 ng/mL) in serum-free medium for 3 days. Merlin expression was evaluated using an anti-merlin antibody. β-Actin was used as the loading control. (B) Human retinal pigment epithelium cells were stimulated with TNF-α (10 ng/mL) or TGF-β2 (5 ng/mL) in serum-free medium for the indicated times. (C) Primary murine RPE cells (Ms RPE cells) were stimulated with recombinant mouse TNF-α (10 ng/mL) in serum-free medium for 4 days. Merlin expression was assessed using an anti-merlin antibody, and β-actin was used as a loading control. (D) Human retinal pigment epithelium cells were stimulated with TNF-α for 2 hours, and the medium was subsequently removed. The cells were further incubated for 4 days in serum-free medium without TNF-α (Withdrawal). The cells were continuously stimulated with TNF-α without a change of medium for 4 days (4 days). A 2-hour stimulation of TNF-α was used as a positive control (2 hours). The cell lysates were subjected to Western blot analysis with the indicated antibodies. Phosphorylation of p65 (p-p65) indicates the activation of the TNF-α signaling pathway. (E) Cells were treated as in (D) and then fixed, permeabilized, and stained with an antibody against p65. Nuclei were stained with DAPI. Scale bars: 50 μm. (F) Supernatant from cells treated as in (D) was collected, and the TNF-α concentration in the culture medium was evaluated using ELISA. Data are means ± standard deviations (SD) from three independent experiments. ***P < 0.01 versus control (Student's t-test). (G) Reverse transcription–PCR was performed using RNA from the cells stimulated with or without TNF-α in serum-free medium for 24 hours. Control reactions were performed using β-actin primers with the same PCR cycle protocol as for TNF-α.

Changes in ERM Phosphorylation Caused By Inhibiting Merlin Expression in ARPE-19 Cells

Immunoprecipitation analysis revealed that merlin interacted with CD44 in ARPE-19 cells in the absence of TNF-α stimulation (Fig. 2A). In contrast, the interaction between merlin and CD44 was reduced following TNF-α stimulation (Fig. 1), which demonstrated that the merlin-CD44 complex decreases in the presence of TNF-α. Next, we used two human merlin-specific siRNAs (number 1 and 2 siRNAs) and two mouse merlin-specific siRNAs (Ms1 and Ms2) to examine the effect of merlin inhibition on ARPE-19 cells. Western blot analysis demonstrated that these siRNAs inhibited merlin expression in both ARPE-19 cells (Fig. 2B, left) and primary murine RPE cells (Fig. 2B, right). Further analysis revealed that merlin depletion increased the localization of CD44 to the cell surface (Fig. 2C). Subsequent Western blot analysis showed increased levels of phosphorylated ERM proteins but no changes in CD44, ezrin, or moesin expression in merlin siRNA-transfected cells (Fig. 2D). Merlin small interfering RNA (siMerlin)-transfected cells also increased phosphorylation of ERM proteins in murine RPE cells (Fig. 2E). Confocal immunofluorescence analysis demonstrated the colocalization of CD44 and phospho-ERM proteins in siMerlin-transfected cells (Fig. 2F). 

Figure 2

 

Merlin prevents ERM phosphorylation and the CD44-ERM interaction. (A) Interaction of CD44 with merlin. Cells were treated with or without TNF-α (10 ng/mL) for 48 hours in serum-free medium. Cells were lysed and subjected to immunoprecipitation using an antibody directed against CD44 or a mouse IgG (isotype control). Whole-cell lysates (the input for immunoprecipitation) and immunoprecipitates were subjected to immunoblot analysis using antibodies directed against merlin or CD44. (B) At 24 hours after siRNA transfection, cells were incubated for 48 hours in serum-free medium. Protein expression was evaluated using an anti-merlin antibody. β-Actin was used as a loading control. Left panels show the results of immunoblot analysis using cell lysates from ARPE-19 cells. Right panels show the same experiment in primary murine RPE cells (Ms RPE cells). (C) After treatment as in (B), cell surface proteins were isolated using the Pierce cell surface protein isolation kit according to the manufacturer's protocol. Whole-cell lysates and isolated proteins were subjected to Western blot analysis. Na⁺/K⁺-ATPase α was used as a control for isolated cell surface proteins. (D) Cells were treated as in (B), and cell lysates were prepared and subjected to immunoblot analysis with antibodies against phospho-ERM (pERM) or CD44. β-Actin, ezrin, and moesin were used as loading controls. (E) Primary murine RPE cells were transfected with siRNA targeting mouse merlin sequences and incubated for 48 hours in serum-free medium. Whole-cell lysates were subjected to immunoblot analysis using the indicated antibodies. (F) Colocalization of CD44 and phospho-ERM (pERM) in siMerlin-transfected cells. Cells stained with antibodies directed against CD44 (green) and phospho-ERM (red) were examined by fluorescence microscopy. Scale bars: 20 μm.

Changes in Hyaluronan Endocytosis Caused By Merlin Knockdown in ARPE-19 Cells

We used Fl-HA to detect hyaluronan in the cytosol of ARPE-19 cells transfected with siMerlin 1 or 2 or siControl (siCTL) to elucidate the role of merlin in hyaluronan endocytosis. We observed increased levels of Fl-HA-containing vesicles in siMerlin-transfected cells. In contrast, internalized Fl-HA was rarely observed in the siCTL-transfected cells (Fig. 3A). Dynasore is an inhibitor of the dynamin GTPase, which is essential for clathrin-dependent endocytosis.²⁰ Dynasore treatment inhibited the merlin siRNA–induced Fl-HA endocytosis (Fig. 3B), and the Fl-HA-positive endocytotic vesicles colocalized with clathrin and the hyaluronan receptor CD44 (Fig. 3C). 

Figure 3

 

Effect of merlin depletion on HA-CD44 endocytosis. (A) Small interfering-RNA–transfected cells were treated with 20 μg/mL Fl-HA at 37°C for 3 hours. Cells were fixed and examined with fluorescence microscopy (green, Fl-HA; blue, DAPI). Fluorescein-labeled hyaluronan–positive vesicles were counted using BZ-X710 software. Scale bars: 50 μm. (B) At 24 hours after transfection, the cells were treated with dynasore (Dyn) or dimethyl sulfoxide (DMSO) as the control (Vehicle) in serum-free medium for 48 hours, and then Fl-HA endocytosis was evaluated. Data are means ± SD for the cells in nine representative fields. **P < 0.01, ***P < 0.001 (Student's t-test). Scale bars: 50 μm. (C) Cells were examined using fluorescence microscopy to detect Fl-HA (green) and immunostained clathrin (red) or CD44 (red). Scale bars: 20 μm.

Inhibition of Dynamin Activity Increased the Level of Merlin in TNF-α–Treated Cells

Tumor necrosis factor-α–treated cells were treated with dynasore, which is an inhibitor of the clathrin-dependent endocytic pathway, to elucidate the nature of the interaction between the endocytic pathway and merlin degradation (or ERM phosphorylation/dephosphorylation). Dynasore treatment rescued the TNF-α–induced degradation of merlin in a dose-dependent manner (Fig. 4, upper lane). In contrast, dynasore treatment led to the accumulation of the phosphorylated form of ERM in TNF-α–stimulated cells relative to that of moesin, which was used as the control (Fig. 4, middle and lower lanes). 

Figure 4

 

Dynamin-mediated endocytosis is associated with TNF-α–induced merlin degradation. Human retinal pigment epithelium cells were treated with dynamin inhibitor dynasore (Dyn) at the indicated concentrations for 48 hours in the presence or absence of TNF-α. Cells were subjected to Western blot analysis using antibodies directed against merlin, phospho-ERM (pERM), or moesin (loading control).

Effects of Dynamin Inhibitor on Induction of a Mesenchymal Phenotype in TNF-α–Treated Cells

We performed Western blot analysis to examine the dose-dependent effects of dynasore on the expression of mesenchymal markers such as α-SMA and fibronectin. Our Western blot analysis revealed that a higher dose (80 μM) of dynasore inhibited the levels of induced α-SMA and fibronectin expression in TNF-α–stimulated cells (Fig. 5A). Upregulation of the expression of mesenchymal markers (α-SMA and fibronectin) was also inhibited by dynasore treatment in a time-dependent manner (Fig. 5B). In addition, the morphology of the ARPE-19 cells changed from an epithelial to a fibroblast-like phenotype after TNF-α treatment, and this appearance changed to an epithelial phenotype after the addition of dynasore (dynamin inhibitor) to the culture media containing TNF-α (Fig. 5C). 

Figure 5

 

Dynamin-mediated endocytosis was associated with TNF-α–induced mesenchymal phenotypes. (A) Cells were treated with TNF-α and dynasore (Dyn) at the indicated concentrations for 48 hours. Whole-cell lysates were used in an immunoblot assay using antibodies directed against fibronectin or α-SMA. Dynasore inhibited the TNF-α-induced mesenchymal changes. (B, C) Cells were treated with TNF-α and dynasore (80 μM) for the indicated times (days), subjected to an immunoblot assay (B) and imaged (C). Moesin and β-actin were used as the loading controls. Scale bars: 200 μm.

Phosphorylation of p38 MAPK in Merlin siRNA–Transfected Cells

Western blot analysis showed that higher concentrations of dynasore inhibited p38MAPK phosphorylation in the cells treated with TNF-α (Fig. 5A) and that cells treated with siMerlin had increased levels of p38MAPK phosphorylation (Fig. 6A). Treatment with (5Z)-7-oxozeaenol, which is a TAK1 inhibitor, inhibited p38MAPK phosphorylation in a dose-dependent manner (Fig. 6A). Phosphorylation of p38MAPK was also increased in siMerlin-transfected murine RPE cells, and (5Z)-7-oxozeaenol blocked the increase in p38MAPK phosphorylation induced by merlin depletion (Fig. 6B). As shown in Figure 6C, TAK1 siRNA strongly inhibited TAK1 protein expression in ARPE-19 cells. Subsequent experiments showed that p38MAPK phosphorylation was inhibited in cells that were co-transfected with siMerlin and siTAK1 (Fig. 6D). 

Figure 6

 

Merlin controls the phosphorylation of p38MAPK via TAK1. (A) At 24 hours after siRNA transfection, ARPE-19 cells were treated with the TAK1 inhibitor (5Z)-7-oxozeaenol (Oxo) at the indicated concentrations for 48 hours in serum-free medium. Whole-cell lysates were analyzed by immunoblot analysis using antibodies directed against phospho-p38MAPK (p-p38MAPK), p38MAPK, merlin, or β-actin (loading control). (B) Primary murine RPE cells were transfected with siRNA and incubated with vehicle or 3 μM (5Z)-7-oxozeaenol (Oxo) for 48 hours in serum-free medium. Whole-cell lysates were subjected to immunoblot analysis using the indicated antibodies. (C) Effect of siTAK1. At 24 hours after transfection, the cells were cultured for 48 hours, and TAK1 expression was evaluated using anti-TAK1 antibody. (D) Merlin suppressed p38 MAPK phosphorylation via TAK1 activity. After siRNA transfection, the cells were incubated in serum-free medium for 48 hours and subjected to immunoblot analysis using an antibody directed against p-p38MAPK or p38MAPK. We confirmed the effect of the merlin and TAK1 siRNAs using the appropriate antibodies. β-Actin was used as the loading control.

Role of the Merlin-TAK1 Pathway in Proliferation of ARPE-19 Cells

We used BrdU incorporation assays to investigate the role of the merlin-TAK1 pathway in ARPE-19 cell proliferation. Our results showed that siMerlin treatment promoted ARPE-19 cell proliferation. Additionally, cotransfection of siMerlin and siTAK1 significantly reduced the level of BrdU incorporation in treated ARPE-19 cells (Fig. 7). 

Figure 7

 

Merlin regulates cell proliferation via TAK1. After being transfected with siRNA, cells were incubated with BrdU (10 μM) for 4 hours, after which the cells were fixed and subjected to fluorescence analysis using an antibody directed against BrdU. Nuclei were stained with DAPI. The number of proliferating cells (cellular incorporation of BrdU) and the total number of nuclei were counted using BZ-X Analyzer software. Data are means ± SD for the cells in three different fields in each of three independent experiments. *P < 0.05 (Student's t-test). Scale bars: 200 μm.

TAK1 Affected the TNF-α–Induced EMT of ARPE-19 Cells

The TAK1 inhibitor (5Z)-7-oxozeaenol was administered to TNF-α–treated cells to investigate the effect of TAK1 on TNF-α–induced EMT of ARPE-19 cells. We observed that (5Z)-7-oxozeaenol treatment inhibited alterations in cell shape from the epithelial to the mesenchymal phenotype (Fig. 8A, upper panels), the dissociation of cell–cell contacts (Fig. 8A, lower panels), and the increased motility of cells caused by EMT (Fig. 8B). The TNF-α–induced phosphorylation of p38MAPK was inhibited by siTAK1 in ARPE-19 cells (Fig. 8C), which is consistent with the results of a previous study.¹⁶ Additionally, the high level of expression of mesenchymal markers such as fibronectin and α-SMA that was caused by TNF-α treatment was reduced by siTAK1 transfection (Fig. 8D). 

Figure 8

 

Transforming growth factor-β–activated kinase 1 is required for TNF-α–induced mesenchymal phenotypes. (A) Cells were stimulated with TNF-α and (5Z)-7-oxozeaenol (3 μM Oxo) or with DMSO as the control (Vehicle). Cells were observed after 48 hours of treatment by light microscopy ([A] upper panels) or subjected to immunofluorescence analysis using an antibody directed against N-cadherin after 24 hours of treatment and examined using fluorescence microscopy ([A] lower panels). (B) Cells that were treated as in (A) for 24 hours were evaluated using a wound-healing assay. The monolayer was scratched, and the migration of cells into the wounded area was imaged at 0 and 20 hours by light microscopy. Yellow dotted lines indicate the initial margins of the wound. Scale bars: 200 μm. (C) After siRNA transfection, cells were stimulated with TNF-α for 48 hours, and then the cell lysates were subjected to immunoblot analysis using antibodies directed against phospho-p38MAPK (p-p38MAPK) or p38MAPK. (D) Cells that were treated as in (C) were subjected to immunoblot analysis using antibodies directed against fibronectin, α-SMA, or β-actin.

Discussion

Retinal pigment epithelium cells are important cellular components of intraocular proliferative diseases. Proliferative vitreoretinopathy is a representative intraocular fibrotic disease involving RPE proliferation and the subsequent onset of EMT. Moreover, inducing EMT in RPE cells contributes to the development of fibrosis in AMD.²¹ Tumor necrosis factor-α plays important roles in the process of EMT and the disruption of cell–cell contacts, which is the initial step triggered by TNF-α and IL-17.^5,22 Additionally, TNF-α promotes the phosphorylation of p38MAPK, extracellular signal-regulated kinase (ERK), and c-Jun N-terminal kinase (JNK) in ARPE-19 cells.²³ 

Merlin, which is a CD44-binding protein, plays a role in various cellular behaviors, such as proliferation, migration, and invasion. Merlin is encoded by the tumor-suppressor gene neurofibromatosis (NF) type 2 on chromosome 22q12.²⁴ The hyaluronan-CD44-merlin complex mediates the contact-dependent inhibition of proliferation and acts as a molecular switch for cell signaling and proliferation.²⁵ In our previous study, microarray analysis showed that the level of merlin mRNA expression was reduced by stimulation with TNF-α combined with TGF-β2, which induced EMT in RPE cells.⁵ Taken together, these results led us to hypothesize that merlin plays an important role in regulating the proliferation and EMT of RPE cells and that this concept might lead to the development of a new inhibitory therapeutic modality. The present study elucidated the roles of merlin and its signal transduction pathway in regulating the proliferation and EMT of RPE cells. 

We found that TNF-α stimulation is involved in the maintenance of TNF-α in culture medium by an autocrine, positive feedback loop and in the sustained activation of the TNF-α signaling pathway following a decrease in merlin expression. Tumor necrosis factor-α alone can induce the downregulation of merlin protein expression (Fig. 1). No other cytokines had similar effects on merlin. 

Whether merlin and the ERM proteins competitively bind to the cytosolic domain of CD44 remains controversial.²⁴ Several studies have demonstrated that activated (phosphorylated) ERM proteins anchor CD44 to the actin cytoskeleton and that this interaction strengthened hyaluronan-CD44 binding and its endocytosis.²⁶ In our present investigation, Merlin interacted with CD44 in ARPE-19 cells that were in a static state. The TNF-α–induced downregulation of merlin expression contributed to ERM phosphorylation and to CD44-phosphoERM binding, leading to hyaluronan endocytosis (Figs. 2, 3). The hyaluronan-CD44 interaction and endocytosis facilitate the association of intracellular mediators of signal transduction through the role of CD44 as a coreceptor for several receptors, such as the TGF-β receptor, epidermal growth factor receptor, and c-Met.^5,27,28 In addition, hyaluronan-CD44 endocytosis occurs via clathrin-dependent and -independent mechanisms,^29,30 and we previously reported that the hyaluronan-CD44–TGF-β receptor complex is internalized via clathrin-coated vesicles in the presence of TNF-α.⁵ The present study demonstrated that the rate of hyaluronan-CD44 endocytosis was accelerated by depleting merlin expression. Moreover, hyaluronan-containing endocytotic vesicles were coated with clathrin heavy chain and were internalized in a dynamin-dependent manner (Fig. 3). Additionally, the expression of EMT markers (such as fibronectin and α-SMA) and the phosphorylation of p38MAPK were downregulated after the inhibition of dynamin-dependent endocytosis as determined by Western blot analysis. These findings clearly showed that merlin plays an important role in the regulation of clathrin-dynamin–dependent hyaluronan endocytosis in RPE cells, which may modify the extracellular components of proliferative tissues. 

During the process of clathrin-dependent endocytosis, dynamin forms a ring-shaped structure around the neck of budding vesicles, resulting in vesicle internalization.³¹ Our previous study⁵ showed that the hyaluronan-CD44–phospho-ERM (moesin) complex is essential for the induction of EMT in RPE cells. In addition, because treatment with the dynamin inhibitor dynasore inhibited EMT in the TNF-α–treated RPE cells, ERM protein phosphorylation might be inhibited by dynasore treatment. However, the levels of phosphorylated ERM proteins and merlin protein increased in the TNF-α–treated RPE cells after treatment with dynasore in the present study (Fig. 4). In contrast, siMerlin treatment increased the phosphorylation of ERM proteins. Taken together, these findings suggested the possibility that endocytosis is associated with the regulation of merlin expression and with the phosphorylation of ERM proteins (Fig. 9). Further studies will be required to determine the exact molecular mechanisms underlying these interactions. 

Figure 9

 

Model of the merlin-mediated regulation of the signaling pathway. CD44 interacts with merlin but not with the ERM proteins in static (epithelial) cells. In contrast, the level of merlin decreases, and merlin dissociates from CD44 in the presence of TNF-α. Merlin downregulation accelerated the rates of CD44-pERM binding and of hyaluronan endocytosis, leading to activation of the TAK1-p38MAPK signaling pathway and the degradation of the merlin protein.

Transforming growth factor-β–activated kinase 1 was identified as a TGF-β–activated mitogen-activated protein kinase kinase kinase in 1995,³² and TNF-α and IL-1 are reported to activate TAK1.^33,34 Recently, studies reported that TAK1 modulates the synergistic effects of TNF-α and TGF-β to drive EMT in bronchial epithelial cells³⁵ and contributes to the mesenchymal phenotypes of mesothelial cells during EMT.³⁶ Our results also showed that TNF-α activates p38MAPK and induces EMT through TAK1 and that merlin regulates the phosphorylation of p38MAPK via TAK1 (Fig. 6). Moreover, BrdU incorporation assays showed that merlin regulated TAK1-dependent proliferation of RPE cells (Fig. 7). These observations suggest that a pathway involving merlin, TAK1 and p38MAPK plays a role in the proliferation and EMT of RPE cells. 

Conclusions

Our present investigation demonstrated that complex interactions between the merlin-TAK1-p38MAPK pathway and dynamin-dependent endocytosis are involved in the proliferation and EMT of RPE cells. These findings may lead to a breakthrough in the development of novel therapeutics to inhibit the proliferation and EMT of RPE cells that cause fibrosis during the pathogenesis of PVR and AMD. 

Acknowledgments

Supported by JSPS KAKENHI Grants 26861457 and 25462721. 

Disclosure: E. Takahashi, None; A. Haga, None; H. Tanihara, None 

Footnotes

 Open Access Article: No