Effects of TGF-β2 on Cadherins and β-Catenin in Human Trabecular Meshwork Cells

Abstract

Purpose.: The effects of TGF-β2 on cadherin-mediated cell–cell adhesion in human trabecular meshwork (HTM) cells were characterized, since TGF-β–induced changes in the cytoskeleton, cell–cell, and cell–matrix interactions of HTM cells are suggested to have a significant role in primary open angle glaucoma.

Methods.: The HTM cells were derived from donor cornea rings and treated with TGF-β2 or vehicle, and protein expression was studied by Western Blot, while protein localization was studied by fractionation of lysates and by confocal immunofluorescence microscopy. Cell–cell adhesion was assessed functionally in dissociation experiments and N-cadherin–mediated cell contact formation in cell spreading experiments on cadherin-coated substrates. A rho-associated protein kinase (ROCK) inhibitor was used to evaluate the contribution of cytoskeletal tension to TGF-β2–induced changes in protein expression and localization.

Results.: TGF-β2 activated Smad-2/3, serine–threonine kinase (AKT), and extracellular signal-regulated kinase (ERK) signaling, and enhanced expression of β-catenin as well as N- and OB-cadherin. The nuclear fraction of β-catenin was not enhanced by TGF-β2. Immunofluorescence microscopy revealed an increased localization of N-cadherin and β-catenin to cell–cell adhesions, and an increase in F-actin. The TGF-β2 increased cell–cell adhesion strength and enhanced N-cadherin–mediated cell contact formation. This effect was blocked by inhibition of mitogen-activated protein kinase kinase (MEK) or AKT. Cytoskeletal relaxation by a ROCK inhibitor did not prevent a TGF-β2–induced increase in cadherin and β-catenin expression.

Conclusions.: The cytokine TGF-β2 enhances cadherin-mediated cell–cell adhesion and β-catenin expression in HTM cells. Increased cell–cell adhesion may contribute to biomechanical alterations in glaucomatous trabecular meshwork (TM), and changes in β-catenin levels and its possible sequestration to cell adhesion sites may affect Wnt signaling. Thus, the crosstalk of TGF-β2 and Wnt signaling in TM may deserve further investigation.

Introduction

The cytokine TGF-β2 is present at elevated levels in glaucomatous eyes ^1,2 and is known to induce a plethora of changes in the cytoskeleton, as well as in extracellular matrix (ECM) protein expression and ECM processing of human trabecular meshwork (HTM) cells. ^3,4 These changes may have an essential role in primary open-angle glaucoma. ^5,6 The TGF-β cytokine also is known to induce significant changes in cadherin-mediated cell–cell interactions, but the effect of TGF-β on cadherin-mediated cell junctions in trabecular meshwork (TM) cells is not fully elucidated. In epithelial cells, TGF-β induces a switch of cadherin expression from E- to N-cadherin. ⁷ This switch is the hallmark of epithelio-mesenchymal transition, a key step in cancer formation. In mesenchymal cells, an increase in N- and OB-cadherin expression also has been reported as a characteristic element of myofibroblast transdifferentiation, ⁸ which has an essential role in wound healing. OB-cadherin was shown to enhance intercellular adhesion strength in fibroblasts developing a profibrotic phenotype. ⁹ Cell–cell adhesions and the actin cytoskeleton are intimately coupled interdependent structures. Recent data indicate that cadherin-mediated adhesions are mechanosensitive transmitters of intercellular forces. ¹⁰ Furthermore, studies on fibronectin matrix remodeling in vivo revealed that cadherin-mediated cell–cell contact modulates tissue tension and integrin-mediated cell–matrix interactions. ¹¹ Since the TM is a mechanosensitive structure and a major site of ocular outflow resistance, alterations in cadherin-mediated cell–cell adhesion may have strong implications in IOP regulation. 

The β-catenin is an important component of cadherin-mediated cell adhesions, as it links cadherins to the actin cytoskeleton via the adaptor protein α-catenin. ¹² Besides this essential adhesive function, β-catenin acts in the Wnt signaling pathway as a nuclear transcriptional activator. In epithelial-mesenchymal transition, a loss of cadherin-mediated cell–cell adhesion is associated with increased transcription of Wnt target genes. ¹³ Thus, it appears that adhesive and transcriptional functions of β-catenin are coupled in a counteracting balance. ¹³ Interestingly, mounting evidence indicates a role of Wnt signaling in IOP regulation. The canonical Wnt pathway can be activated in human TM cells in vitro, and Wnt antagonists increased IOP in mice and in a human anterior chamber perfusion model. ^14,15  

In light of these data, we studied the effects of TGF-β2 on the expression and localization of cadherins and β-catenin, and possible subsequent changes in cell–cell adhesion in HTM cells. 

Methods

Reagents

Antibodies raised against the following proteins were used: α–smooth muscle actin (α-SMA; Sigma, Taufkirchen, Germany); focal adhesion kinase (FAK), GSK3β, p-GSK3β, serine–threonine kinase (AKT), p-AKT, N-cadherin, and β-catenin (Cell Signaling Technology/NEB, Frankfurt, Germany); FAK pY397 (Biosource, Nivelles, Belgium); OB-cadherin (EMD Millipore, Billerica, MA); extracellular signal-regulated kinase (ERK; Santa Cruz Biotechnology, Inc., Santa Cruz, CA); pERK (Promega, Mannehim, Germany); pSmad2/3 (Zymed, Zytomed, Berlin, Germany); glyceraldehyde-3-phosphate dehydrogenase (GAPDH; R&D Systems, Wiesbaden, Germany); Alexa-488–conjugated goat anti-mouse (Molecular Probes, Eugene, OR); and horseradish peroxidase (HRP)–conjugated secondary antibodies (Jackson/Dianova, Hamburg, Germany). Phalloidin-TRITC (Sigma) was used to stain filamentous actin. For stimulation experiments, Dulbecco's modified Eagle's medium (DMEM, 3% fetal calf serum, see below) was used as vehicle and recombinant human TGF-β2 (PeproTech, Hamburg, Germany) was added as indicated. Inhibitors for mitogen-activated protein kinase kinase 1/2 (MEK1/2, U0126; Calbiochem, San Diego, CA), AKT (AKT inhibitor X; Calbiochem), and rho-associated protein kinase 1/2 (ROCK1/2, H-1152; Calbiochem) were used in 10 μM concentrations. 

Cell Culture

HTM tissue from donor cornea rings was used with informed consent and cells were cultivated according to methods published previously with slight modifications. The tenets of the Declaration of Helsinki were followed in all procedures. In brief, donor rings were transferred from the storage medium and kept in DMEM (PAA Laboratories GmbH, Pasching, Austria) supplemented with 10% heat inactivated fetal calf serum (FCS; Biochrom, Berlin, Germany), 100 U/mL penicillin and 100 μg/mL streptomycin (both from PAA Laboratories GmbH) for 24 hours. Under microscopic guidance, anterior and posterior incisions were placed to isolate the TM, which then was removed using forceps and cut into smaller sections. The tissue sections were placed in 24-well plates, covered with a glass coverslip to avoid floating, and incubated in growth medium (as above). Confluent cell layers were passaged by trypsinization. From the second passage, FCS concentration was reduced to 3%. Cells were characterized by assessing baseline α-B–crystallin expression and increased myocilin expression after 7 days of dexamethasone treatment. The myocilin response to dexamethasone treatment has been reported as specific for TM cells. ¹⁶ All experiments were performed at least three times using cells derived from different preparations with similar results. 

Western Blot

Cells plated on tissue culture plastic were used for Western Blot analysis. Cells were rinsed with ice cold PBS and total cell protein extracts were prepared using a radio-immunoprecipitation assay (RIPA) lysis buffer (20 mM TRIS, 150 mM NaCl, 0.1 mM EDTA, 1% Triton X-100, 1% deoxycholate, 0.1% SDS) containing phosphatase and protease inhibitors (Phosphatase Inhibitor Cocktail III; Calbiochem/Merck, Bad Soden, Germany; and Complete Protease Inhibitor; Roche, Mannheim, Germany). Protein concentrations were measured using a BCA assay (KMF, Lohmar, Germany). Then, 10 μg protein extracts were boiled in Laemmli sample buffer and subjected to SDS polyacrylamide gel electrophoresis. Proteins were transferred onto a polyvinyl difluoride (PVDF) membrane (Amersham, Braunschweig, Germany) using a Bio-Rad gel blotting apparatus (Bio-Rad Laboratories, Inc., Hercules, CA). Membranes were blocked in 3% BSA in Tris-buffered saline with Tween (TBST; 10 mM Tris HCl, 150 mM NaCl, 0.1% Tween 20) for 1 hour. Membranes were incubated with primary antibody overnight at 4°C and with a peroxidase-conjugated secondary antibody for 45 minutes at room temperature. After each incubation step, membranes were washed in TBST for 30 minutes. Peroxidase was visualized by enhanced chemoluminescence and exposure to Hyperfilm ECL films (both from Amersham) for appropriate times. 

Fractionation Assay

The fractionation assay was performed as described previously ¹⁷ with minor modifications. The cell layer was scraped off in cold Hepes buffer (4°C, 10 mM HEPES, pH 8.0, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA), and allowed to swell on ice for 15 minutes. Then, Triton X-100 (final concentration 0.6%) was added, and the tubes were vortexed, and then centrifuged (20,000g, 4°C) for 30 seconds. The cytoplasmic supernatant was retrieved and frozen for subsequent protein analysis. The nuclear pellet was washed in Hepes buffer without detergent (as above) and resuspended in nuclear lysis buffer (340 mM Tris pH7.4, 1% Triton X-100, 1% sodium deoxycholate, 1% SDS, 0.1 mM EGTA, 10 mM sodium fluoride). The lysates were incubated on ice for 15 minutes, vortexed, and centrifuged (20,000g, 5 minutes), and the supernatant was retrieved as a nuclear protein fraction. 

Dissociation Assay

The dissociation assay was performed as reported previously ¹⁸ with slight modifications. Confluent cell layers were treated with vehicle or TGF-β2 (10 ng/ml) for 3 days, washed with PBS, scraped in PBS, and triturated in a standardized fashion using a 200 μL pipet tip. Aliquots were assessed in a Neubauer chamber (Carl Roth, Karlsruhe, Germany) and particles were counted in 5 random fields. Next, EGTA (final concentration 3 mM) was added to the cell suspension to dissolve the cells fully, and total cells were counted as above. 

Spreading Assay

Confluent HTM cells were treated with vehicle or TGF-β2 (2 ng/ml) in the presence or absence of AKT or MEK inhibitors for 2 days. Next, the cells were trypsinized, kept in suspension in a cell culture incubator for 1 hour in the presence or absence of AKT or MEK inhibitors as indicated to allow for equal retraction of all cells, and subsequently plated on glass coverslips coated with N-cadherin-Fc chimeric protein (Invitrogen, Carlsbad, CA), and blocked with heat-denatured BSA. ¹⁹ At 40 minutes after plating, the cells were fixed in 2% paraformaldehyde (Merck, Mannheim, Germany), permeabilized with 0.1% Triton X-100, and F-actin was stained with Phalloidin-TRITC (Sigma). After washing in PBS, the stained samples were mounted in Vectashield (Vector Laboratories, Inc., Burlingame, CA) and viewed with a fluorescence microscope (Axiophot; Zeiss, Oberkochen, Germany). To assess cell spreading, slide labels were masked and cells were counted in randomly chosen fields until at least 100 cells were counted. Spreading cells were identified by their formation of lamellipodia and scored as “spreading,” all other cells were scored as “nonspreading.” 

Statistics

To assess the two dissociation assay groups, a 2-tailed t-test was used. Spreading assay data were analyzed by Dunnett contrast post hoc ANOVA using R software (available in the public domain at http://www.R-project.org; The R Foundation for Statistical Computing). 

Results

Effects of TGF-β2 on Cadherins and β-Catenin

We initially studied the expression of cell–cell adhesion proteins. Western blot analysis revealed that TGF-β2 induced an increase in N-cadherin, OB-cadherin, and β-catenin expression within 12 hours, and a concomitant increase in the myofibroblast marker α-SMA (Fig. 1A). These changes were preceded by a rapid and transient activation of the Smad signaling pathway, and subsequent activation of AKT and ERK non-Smad signaling. Since β-catenin serves a dual role as a linker in cadherin-mediated cell adhesion and as a nuclear transcription cofactor in Wnt signaling, we also studied the localization of β-catenin. Biochemical fractionation experiments revealed that TGF-β2 stimulation for 2 days raised cytosolic β-catenin levels, but failed to increase β-catenin in the nuclear compartment (Fig. 1B). The p-Smad2/3 and GAPDH were assessed to control for appropriate fractionation and treatment. Confocal immunofluorescence microscopy showed enhanced localization of N-cadherin and β-catenin to cell–cell borders, and an increase in F-actin stress fibers after 2 days of TGF-β2 treatment (Figs. 1C–J). Only very weak nuclear staining of β-catenin was observed and this was not changed by TGF-β (Figs. 1E, 1F). 

Figure 1

 

(A) Time course of TGF-β2–induced (2 ng/ml) changes in protein expression and signaling protein phosphorylation. (B) Distribution of proteins in nuclear and cytosolic fractions of whole cell lysates. Cells were treated with vehicle or TGF-β2 (2 ng/ml) for 2 days. The GAPDH and p-Smad2/3 were assessed as fractionation and treatment controls. (C–J) Confocal immunofluorescence micrographs of HTM cells treated with vehicle or TGF-β2 (2 ng/ml) for 2 days. To improve signal visibility, localization of N-cadherin and β-catenin is presented in inverted grayscale. Arrowheads indicate enhanced actin stress fiber bundles in (H).

TGF-β2 Enhances Cell–Cell Adhesion

To elucidate whether changes in adhesion protein expression would confer functional changes in cell–cell adhesion, we performed cell dissociation assays. The HTM cells were treated with TGF-β2 (10 ng/mL) or vehicle for 2 days, and the cell layer was scraped off and dissolved by a fixed trituration scheme. The resulting cell suspension was assessed by microscopy in a masked fashion and particles (single cells or cell clusters)/volume were counted. In a second step, cells were dissociated fully by EGTA binding of calcium, which is essential for cadherin-mediated adhesion, and particles/volume were assessed again. The ratio of particles/total cells reflected cell cluster formation. Our data indicated that TGF-β2 treatment rendered the cells significantly more resistant to dissociation (Figs. 2A, 2B), with ratios of 0.50 ± 0.08 in vehicle and 0.26 ± 0.03 in TGF-β2-treated cells (P = 0.04). 

Figure 2

 

The TGF-β2–induced changes in cell–cell adhesion. (A) Phase contrast images of dissociated cell cultures after treatment with vehicle or TGF-β2 (10 ng/ml) for 2 days. Cell clusters are visible in TGF-β2–treated cells (arrows). (B) Quantitative assessment of cell dissociation. Particles (single cells and clusters) were counted after standardized trituration, then cells were dissociated fully using EGTA and total cells were counted. Means ± SEM. *P < 0.05.

TGF-β2 Enhances N-Cadherin–Mediated Cell Contact Formation

To address N-cadherin–mediated binding specifically, we performed cell spreading assays using N-cadherin–coated cell culture substrata, a method established in the literature. ¹⁹ Following 2 days of TGF-β2 pretreatment (2 ng/mL) N-cadherin–mediated cell spreading was enhanced robustly (Figs. 3A, 3B). The presence of the MEK-ERK inhibitor U0126 or an AKT inhibitor during TGF-β2 pretreatment abolished the TGF-β2–induced enhancement. To control for more general effects of the inhibitors on cell spreading, short-term (1 hour) inhibition was assessed in parallel. Spreading was similar to control conditions in the presence of U0126. In contrast, short-term AKT inhibition also affected cell spreading. 

Figure 3

 

N-cadherin–mediated cell contact formation. The HTM cells were treated with vehicle or TGF-β2 (2 ng/ml) for 2 days in the presence or absence of inhibitors as indicated (both 10 μM). Cells then were plated on N-cadherin–coated substrates to allow for N-cadherin–mediated cell spreading. (A) Cells were fixed after 40 minutes and F-actin was stained with phalloidin to assess lamellipodia formation as an indication of cell spreading. (B) Quantitation of spreading cells. Means ± SEM, asterisks indicate significance of difference from control (***P < 0.001, **P < 0.01, *P < 0.05).

TGF-β2–Induced Increase in N-Cadherin and β-Catenin Is Unaffected by ROCK-Inhibition

Cadherin-mediated cell adhesions are load-dependent, ²⁰ and an increase in cytoskeletal tension due to TGF-β–induced Rho activation may have a role in increased expression of cadherins and β-catenin, and in the localization of β-catenin to cell–cell junctions. Inhibitors of the Rho-dependent kinase ROCK release cytoskeletal tension and increase aqueous humor outflow. ²¹ Therefore, we studied the effects of a ROCK inhibitor on TGF-β2–induced changes in cadherin and β-catenin expression. Both were not affected by ROCK inhibition, but FAK phosphorylation, α-SMA expression, and phosphorylation of GSK3β were reduced (Fig. 4A). Confocal immunofluorescence microscopy revealed an increase of F-actin stress fibers by TGF-β2 (Figs. 4B, 4C) and a global loss of F-actin and cell confluence in the presence of H-1152, a potent ROCK inhibitor (Figs. 4D, 4E). However, a TGF-β2–induced increase in junctional β-catenin occurred (Figs. 4D, 4E). 

Figure 4

 

Effects of the ROCK inhibitor H-1152 on TGF-β2–induced changes in protein expression and activation. (A) Western blot of cells treated with vehicle or TGF-β2 in the presence or absence of H-1152 (10 μM). (B–E) Confocal micrographs depicting F-actin and β-catenin localization. Arrowheads indicate discontinuity of cell–cell junctions in (D). Scale bar: 40 μm.

Discussion

Little is known about TGF-β–induced changes in TM cell–cell adhesion. In a TGF-β–dependent rat glaucoma model, overexpressed TGF-β was shown to induce corneal endothelial cell transdifferentiation, ²² with subsequent closure of the iridocorneal angle. In this study, only faint N-cadherin expression was detected in the TM, irrespective of treatment. Data on TGF-β–induced alterations in OB-cadherin expression and β-catenin expression or localization in the TM have been lacking. Our results indicated that TGF-β2 induces myofibroblast transdifferentiation in TM cells as characterized by a robust increase in α-SMA, β-catenin, and N- and OB-cadherin expression (Fig. 1A). These observations are supported by earlier reports of an OB-cadherin shift and concomitant α-SMA expression in mesenchymal cells in a model of dermal wound healing. ⁸  

β-Catenin may translocate to the nucleus to drive Wnt-dependent gene expression, bind to cadherin-mediated cell adhesions to modulate actin coupling, or be degraded following ubiquitination. To this end, our experiments revealed that the nuclear fraction of β-catenin is not enhanced by TGF-β2 (Fig. 1B). In contrast an increase in the cytosolic fraction was detected. Along these lines, confocal immunofluorescence micrographs indicated a preferred recruitment of N-cadherin and β-catenin to cell–cell junctions, with only a very faint nuclear β-catenin signal (Figs. 1C–F, 1I, 1J). These data suggested that TGF-β2 enhances cadherin-mediated cell–cell adhesions. 

Next, we performed cell dissociation assays to study cell–cell adhesion in a robust functional assay. The TGF-β2–pretreated cells were more difficult to dissociate than vehicle-treated cells, indicating enhanced cell–cell adhesion (Fig. 2). An increase in intercellular adhesion strength mediated by OB-cadherin has been reported in fibrogenic fibroblasts. ⁹ However, a mere increase in cadherin binding sites due to increased cadherin expression levels may suffice to enhance cell–cell adhesion. 

Spreading assays allowed us to gain further insight into N-cadherin–mediated cell binding. These experiments revealed that TGF-β2 enhanced N-cadherin–mediated cell spreading, further supporting our previous results. The N-cadherin–mediated cell spreading process appears to depend on PI3K-AKT signaling, since an AKT-inhibitor blocked spreading irrespective of cell pretreatment. A role of PI3K-AKT signaling in N-cadherin–mediated cell spreading has been reported previously. ²³ On the other hand, MEK-ERK inhibition by U0126 did not affect the spreading process, but blocked the effect of TGF-β2 treatment. These observations argue for a role of non-Smad signaling pathways in TGF-β2–induced HTM transdifferentiation. We had detected a role of non-Smad signaling in TGF-β2–induced α-SMA expression in earlier studies. ²⁴ AKT inhibition also blocked the TGF-β2–induced increase in N-cadherin and β-catenin expression as determined by Western blot (data not shown). 

Similar to integrin-dependent focal adhesions, cadherin-mediated adhesions were shown to be force-dependent. ²⁰ An increase in contractile force due to TGF-β2–induced F-actin formation and α-SMA expression, thus, could suffice to enhance cadherin-mediated adhesion in HTM cells. To test this possibility, we used H-1152, a specific ROCK inhibitor or vehicle in preincubation experiments. The TGF-β2 increased the phosphorylation of FAK and GSK-3β as an indication for increased focal adhesion and AKT signaling, respectively. ROCK inhibition failed to prevent a TGF-β2–induced increase in the levels of N-cadherin and β-catenin, but FAK and GSK3β phosphorylation were suppressed as an indication of decreased cytoskeletal tension (Fig. 4A). Confocal immunofluorescence also revealed typical signs of decreased cellular tension in the presence of H-1152, with dissolution of actin stress fibers and discontinuity of cell–cell junctions (Fig. 4B). However, β-catenin staining at cell borders was increased by TGF-β2 treatment. These data suggested that the TGF-β2–induced changes in cadherin and β-catenin expression levels are independent of cytoskeletal tension. Interestingly, earlier reports suggested that a localized suppression of Rho-kinase signaling is essential for formation of cadherin-mediated cell–cell adhesions. ²⁵  

Since β-catenin serves a dual role as cytoskeletal mediator in cadherin-dependent adherent junctions and as a second messenger in Wnt signaling (Fig. 5), changes in β-catenin expression and localization may have additional consequences for TM signaling and outflow regulation. Recent studies indicate that Wnt signaling is active in the TM, ¹⁵ and inhibition of Wnt by overexpression of secreted frizzled-related protein-1 (sFRP-1) has been shown to increase outflow resistance in mice. ¹⁴ Wnt signaling depends on the availability of cytosolic β-catenin for nuclear translocation. Biochemical experiments and observations in colon carcinoma cells have shown that cadherin binding can prevent the nuclear localization and transactivation activity of β-catenin. ²⁶ Thus, a sequestration of β-catenin to adherent junctions may limit Wnt-induced signal transduction. By the same token, TGF-β could enhance Wnt-signal transduction by an increase in total cytosolic β-catenin. It appears that possible effects on outflow depend on a balance of cadherin-mediated adhesion, cytosolic β-catenin levels, and Wnt receptor activation levels. Disruption of adherent junctions by ROCK inhibitors, thus, may enhance Wnt signaling as it may boost availability of cytosolic β-catenin. 

Figure 5

 

Simplified model of signaling pathways mentioned in the study. β-Catenin serves dual roles as a Wnt signaling transcription factor and as an essential component of cadherin-mediated cell–cell adhesions. The TGF-β activates Smad, MEK-ERK, Pi3K-AKT, and RhoA-ROCK signaling, and alters cytoskeletal structures, as well as β-catenin and cadherin expression.

In summary, TGF-β2 induces HTM transdifferentiation, and alters cell–cell adhesion by changing cadherin and β-catenin expression. Enhanced cell–cell adhesion could alter tissue biomechanics, ²⁷ and altered cytosolic levels of β-catenin have the potential to affect Wnt signaling. ¹³ Both aspects have a role in TM outflow regulation. Possible crosstalk of TGF-β2 and Wnt signaling pathways in the TM deserves further investigation. 

Acknowledgments

Supported by Interdisziplinäres Zentrum für Klinische Forschung Würzburg (IZKF Seed Grant Z 4-81; GS) and Deutsche Forschungsgemeinschaft (DFG Schl 563/3; GS). The authors alone are responsible for the content and writing of the paper. 

Disclosure: T. Wecker, None; H. Han, None; J. Börner, None; F. Grehn, None; G. Schlunck, None