Targeting the Fibronectin Type III Repeats in Tenascin-C Inhibits Epithelial–Mesenchymal Transition in the Context of Posterior Capsular Opacification

Abstract

Purpose.: Posterior capsular opacification (PCO) is a common complication following extracapsular surgery, associated with fibrosis, opacification, and contraction of the posterior lens capsule. It is characterized by increased expression of extracellular matrix proteins such as tenascin-C, fibronectin, collagens, and proteoglycans. Tenascin-C is known to be critical for injury-induced epithelial–mesenchymal transition (EMT) in the lens epithelium. We aimed to target fibronectin type III repeats 1-5 within tenascin-C (TNfnIII 1-5) using an scFv (single-chain variable fragment) antibody, and to evaluate its effectiveness in the context of lens epithelial cells.

Methods.: Phage display library screening was used to generate an antibody against TNfnIII 1-5. Lens epithelial cells were cultured in the presence of the scFv antibodies to evaluate the effects on cell proliferation, migration, fibronectin polymerization and deposition, matrix metalloprotease (MMP) regulation, actin stress fiber distribution, and expression of EMT markers. The effect on SMAD-dependent and SMAD-independent pathways was also examined.

Results.: The scFv TN64 was found to be effective in regulating the proliferation, migration, and expression of MMP-2 and MMP-9, fibronectin polymerization and deposition, and expression of EMT markers. TN64 did not interfere with SMAD3 phosphorylation. Altered localization of β-catenin, as well as downregulation of phosphorylation of mitogen-activated protein (MAP) kinases and focal adhesion kinase (FAK), was involved.

Conclusions.: Our data suggest that the TNfnIII 1-5 repeats play an important role in PCO pathology. The inhibition of EMT by TN64 is mediated by SMAD-independent, integrin–β-catenin–FAK signaling pathway, and is therefore proposed as a novel antifibrotic therapeutic candidate.

Introduction

Extracapsular cataract surgery involves removal of the lens anterior capsule to create a capsular bag (formed by the remaining anterior capsule and the entire posterior capsule), followed by phacoemulsification of the cataractous lens and implantation of an intraocular lens into the bag. The remnant lens epithelial cells (LECs) left behind in the capsular bag respond by undergoing a wound healing response (aided by the presence of cytokines such as TGF-β present in high concentration in the aqueous humor^1,2) either leading to formation of myofibroblasts or by differentiating into lens fiber cells in an attempt to regenerate the damaged lens.³ While the former response leads to the proliferation, migration, and epithelial–mesenchymal transition (EMT) of LECs,^4,5 in the latter response, the lens fibers undergo an aberrant organization^6,7; so together, the effects of EMT and fiber cell differentiation result in loss of visual acuity and opacification of the posterior capsule and/or fibrosis and contraction of the capsule bag (called posterior capsular opacification or PCO).⁸ 

Posterior capsular opacification is the most common problem post cataract surgery.^8,9 Although improvements in intraocular lens material and surgical techniques have reduced the incidence of PCO, it still remains a significant problem.¹⁰ The incidence rate of PCO is approximately 50% in adults, while in children the rates are higher.^11,12 Currently, the only effective treatment is neodymium:YAG capsulotomy; however, it is associated with several complications (e.g., retinal detachment, corneal edema), and the procedure is not always available in developing countries.¹³ Therefore other methods to inhibit PCO, such as improvements in intraocular design and materials^14,15 and design of better pharmacological interventions,^16–19 as well as improvements in surgical procedures, are being explored in an effort to reduce the incidence of PCO.^20,21 The latter involve the use of reagents capable of selectively depleting the residual LECs. 

Remodeling of the extracellular matrix is an important event in PCO pathology. Increased expression of extracellular matrix (ECM) proteins such as fibronectin, tenascin, and type I collagen has been documented in LECs post EMT^5,22–24; loss of tenascin-C has been in fact shown to attenuate EMT in the lens epithelium post injury.²⁵ Fibronectin and tenascin-C are important ECM proteins that interact with each other²⁶; the type III modules 3 through 5 on tenascin are known to bind to a cryptic site in the N-terminal 29-kDa domain of fibronectin.²⁷ Both proteins modulate the ECM by interacting with other ECM proteins, including collagen and other heparan sulfate proteoglycans. Fibronectin and tenascin-C are composed of multiple repeating domains of fibronectin type III repeats (TNfnIII, made up of ~90 amino acids). It is now well understood that proteins with such repeats bind to activated integrins on cell surfaces, leading to cell signaling events and thereby influencing the behavior of cells with respect to cell adhesion, differentiation, growth and migration, and cell–matrix communication.^28–30 

Previous studies have shown that the TNfnIII domain and its spliced variant are overexpressed under many pathological conditions.^31,32 Further, given the fact that there is a requirement of tenascin-C for the induction of EMT in the injured lens,²⁵ along with the fact that there is increased expression of tenascin-C post EMT in human PCO,³³ we explored targeting of the TNfnIII domain using an scFv (single-chain variable fragment) antibody to assess the contribution of this domain in causation of the PCO pathology. 

The human epithelial cell line HLE-B3 and primary culture of the LECs from explants of anterior capsules of cataract patients were used to show that TNfnIII plays an important role; blocking the domain with an scFv antibody resulted in controlling the key processes of PCO. In addition, candidate signaling pathways associated with TGF-β-induced EMT were evaluated. 

Materials and Methods

Ethics Statement

The capsules were obtained for research use with the written consent of all the participants enrolled in a protocol approved by the Institutional Ethics Committee of the Postgraduate Institute of Medical Education and Research; the tenets of the Declaration of Helsinki were adhered to. 

Maintenance and Culture of the Cell Lines

Human lens epithelial cell line HLE-B3 was maintained in minimal essential medium (MEM) supplemented with 50 μg/mL gentamycin and 20% FBS.³⁴ The cell line was used each time at the third passage. 

Culturing of Anterior Capsule Explants Derived From Cataract Patients

Lens capsule of cataract patients was collected in a sterile vial containing MEM. Capsules were allowed to attach to a 24-well plate precoated with poly-L-lysine in the presence of a minimal amount of complete medium (MEM + 20% FBS) for approximately 2 to 3 hours, followed by addition of complete medium, and incubated at 37°C and 5% CO₂. Medium was changed after every 3 days. Cells were passaged at 60% confluency, and cells were used after every third passage. 

Cloning and Expression of Tenascin Constructs

For the construction of TNfnIII 1-5, tenascin-C clone TN1-5 (gifted by Erickson, Duke University) was re-engineered to include a His tag at its N-terminus. The gene was amplified using a set of primers with SphI site incorporated in the forward primer and HindIII site in the reverse primer. The amplified gene was cloned into vector pQE-30 for incorporating a hexa histidine tag (6xHis-tag) at its N-terminus to ease downstream purification. Polymerase chain reaction amplification of the tenascin (1-5) gene was carried out using the following primers: forward, 5′GCGGGATCGCATGCTCAGAGGTGTCTCCTCCC3′; reverse, 5′CGCCAAGCTTCTAAGTGGATGCCTTCACATGTGCG3′. These include the SphI and HindIII sites, respectively. The PCR conditions used for amplification were as follows: initial denaturation at 94°C, 5 minutes, followed by 30 cycles of denaturation at 94°C, 1 minute; annealing at 59°C, 1 minute; extension at 72°C, 3 minutes; and a final extension at 72°C for 10 minutes. The amplified product (1400 bp) was cut with SphI and HindIII, respectively, and ligated into the correspondingly digested plasmid, pQE-30 vector. Following transformation into Escherichia coli strain XL1-Blue, colonies were screened, and confirmation of the incorporation of gene into the vector was done by DNA sequencing. Expression of TNfnIII 1-5 was carried out by transformation of the plasmid into M15 (pREP4) cells, and purification of the protein was done under native conditions by Ni-NTA column chromatography (Qiagen, Valencia, CA, USA). 

From the TNfnIII 1-5 construct, the other variants TNfnIII 1-3, TNfnIII 3-5, and TNfnIII 3 were derived using appropriate primers; all the constructs were similarly expressed and purified. 

Generation and Purification of Anti-TNfnIII 1-5 Antibody

Fibronectin type III-like repeat of tenascin-C region 1-5 (TNfnIII 1-5) was used as an antigen for screening the phage display antibody library (Tomlinson I+J; Medical Research Council, Centre for Protein Engineering, Cambridge, UK) by the usual methods of biopanning and amplification.³⁵ Three rounds of biopanning were done against immobilized antigen TNfnIII 1-5. Eluted phages at the end of the third round were transformed into exponentially growing (optical density = 0.4) HB2151 E. coli cells to produce soluble scFv. The phagemid was transformed into E.coli HB2151 strain (a nonsuppressor strain); the cells were cultured overnight at 30°C in the presence of 0.1% glucose to produce soluble scFv, and purification of the antibody was carried out as described earlier.³⁵ In all experiments, scFv O52 (generated against platelet cell surface proteins) was used as an irrelevant negative control, since it is a structurally matched antibody to TN64. 

Screening of scFv Clones for Binding to TNfnIII 1-5

TNfnIII 1-5 was coated on 96-well plates, and the supernatant containing soluble scFv was used as primary antibody in the ELISA assay. After washing with 1× PBS and incubation for 2 hours at room temperature, horse radish peroxidase (HRP)–conjugated Protein A was added to the wells at a dilution of 1:10,000 and incubated for 1 hour. Color was developed with 3,3′,5,5′-tetramethylbenzidine (TMB); the reaction was stopped with 1 N sulfuric acid, and the plates were read at 450 nm (Supplementary Fig. S1A). The selected scFv clone TN64 was further expressed and purified using Ni-NTA agarose column; Western blotting of TN64 using mouse anti-His-HRP confirmed the expression of the ~35-kDa band corresponding to TN64 expression (Supplementary Fig. S1B). When the various tenascin variants derived from TNfnIII 1-5 were checked for binding to TN64, it was observed that while TN1-3 and TN3-5 bound to TN64, there was only a very weak signal with the variant TN3 (Supplementary Fig. S1C). We tested the possible cross-reactivity of TN64 against matricellular/ECM proteins, fibronectin, osteopontin, and fibrinogen by ELISA (Supplementary Fig. S2A) and immunoblotting (Supplementary Fig. S2B). Single-chain Fv O52 was also tested alongside. The scFv TN64 was observed to bind to TNfnIII 1-5 alone and did not bind to any of the other proteins tested. O52 showed no positivity with any protein, including TNfnIII 1-5. Immunofluorescence staining of LECs grown from patient capsules showed that TN64 was able to bind to the LECs, colocalizing with anti-tenascin-C antibody (Sigma-Aldrich Chemical Co., St. Louis, MO, USA); in another such capsule, O52 showed no binding to the LECs (Supplementary Fig. S3). 

Western Blotting

HLE-B3 or human lens epithelium primary cells were seeded in the presence of the complete medium (MEM + 20% FBS) and incubated overnight in a CO₂ incubator. Cells were serum starved for 8 to 10 hours, followed by the addition of the test reagent (scFv) along with TGF-β2 (2 ng/mL), and incubated for 24 hours. Cells were washed with PBS, and cell lysate was boiled in 1× SDS-PAGE sample loading buffer, followed by electrophoresis and Western blotting. The signal was visualized by enhanced chemiluminescence (ECL) reagent (Amersham Biosciences, Piscataway, NJ, USA) and was captured on X-ray films. 

Immunofluorescence

HLE-B3 cells or human lens epithelium primary cells were seeded in a manner similar to that used in the Western blotting experiments. At the end of the period of treatment with TGF-β2 and scFv antibodies, cells were fixed with 4% paraformaldehyde, permeabilized using 0.5% Triton X-100 and 0.05% Tween-20, and blocked with 5% BSA in PBS. This was followed by incubation with primary antibody and secondary FITC/tetramethylrhodamine (TRITC)/phycoerythrin (PE)–labeled antibody for 1 hour each. Images were acquired using Olympus confocal microscopy (Tokyo, Japan). 

Cell Viability and Proliferation

Lens epithelial cells were seeded at a density of 5000 per well in 96-well plates in complete medium and incubated overnight. Cells were serum starved for 8 to 10 hours, followed by addition of the scFv antibody in a dose-dependent manner, and incubated for 24 hours. An MTT (3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazoliumbromide) assay was carried out as described earlier.³⁵ For proliferation, 5-bromo-2-deoxyuridine (BrdU) was added at a concentration of 10 μM per well 6 to 8 hours before the end of the incubation, following which media were discarded and the cells were fixed using methanol/H₂O₂ for 1 hour, denatured, and processed as described earlier.³⁵ 

Scratch Assay to Assess Cellular Migration

For evaluating the efficacy of the scFv antibody for cell migration, wound healing/scratch assay was performed.³⁶ Briefly, LECs were cultured to achieve 90% confluency and serum starved, and a vertical wound was created using a 200-μL pipette tip, followed by washing. Single-chain Fv antibody (25 μg/mL) was added along with TGF-β2 (2 ng/mL). Imaging was done to note the status at 0- and 24-hour periods. Extent of migration was evaluated by analyzing the images using t-scratch software (CSE Lab, ETH Zurich, Switzerland). Data were represented as percent closure of wound; control (with no scFv added) was assigned a value of 100%. 

Matrix Metalloprotease Expression

To evaluate the effect of the scFv antibodies on the expression of matrix metalloprotease MMP2 and MMP9, the culture medium was collected from the cell migration assays, and expression of the MMP was evaluated by gelatin zymography using a previously described protocol.³⁷ 

DOC Insolubility Assay

The extent of deposition of insoluble fibronectin in the ECM was evaluated by performing deoxycholic acid (DOC) insolubility assay as described previously.³⁸ Briefly, 1 × 10⁶ cells were seeded per well of a six-well plate in the presence of complete medium and incubated overnight, serum starved for 8 to 10 hours, treated with the respective scFv antibodies along with TGF-β2 (2 ng/mL), and further incubated for 24 hours. At the end of the incubation, medium was removed and cells were lysed in DOC lysis buffer according to the described protocol.³⁸ The blot was developed by chemiluminescence (Bio-Rad, Hercules, CA, USA) and captured on X-ray film. ImageJ software (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA) was used for densitometry quantitation. To ensure equal loading, DOC soluble fraction was separated on 10% SDS-PAGE in similar fashion and immunoblotted with anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) antibody (1:5000; Sigma-Aldrich Chemical Co.). 

Measurement of EMT-Specific Markers

The expression of mesenchymal markers N-cadherin, collagen type I, β1 integrin, and α-smooth muscle actin (SMA) was measured in the presence and absence of the scFv antibodies³⁹ to evaluate the efficacy of the antibody. For assessment of expression of α-SMA and N-cadherin, LECs were seeded in 96-well-plates or on coverslips, and the same protocol was used as for immunofluorescence (described above). Cells were probed with either anti-N cadherin or anti-α-SMA antibody, followed by FITC-conjugated secondary antibody. DAPI (4′,6-diamidino-2-phenylindole) was used as nuclear dye. Fluorescence intensity was quantified using a PerkinElmer ELISA plate reader (PerkinElmer, Norwalk, CT, USA), and DAPI was used to ensure equal cell density. Data were plotted as percent fluorescence intensity. For imaging of the cells, confocal microscopy was carried out using the Olympus microscope. For evaluation of levels of collagen type I and β1 integrin, LECs were cultured in complete medium followed by serum starvation for 8 to 10 hours. Cells were treated with and without scFv along with TGF-β2 (2 ng/mL). At the end of the treatment, cells were detached and washed with PBS. For collagen type I, cells were fixed, permeabilized, blocked, and incubated with mouse anti-collagen type I antibody (Sigma-Aldrich Chemical Co.). For β1 integrin, cells were incubated in blocking buffer, followed by incubation with primary mouse anti-β1 integrin antibody (Sigma-Aldrich Chemical Co.). Fluorescein isothiocyanate–conjugated anti-mouse antibody was used as secondary antibody. Mean fluorescent intensity of the cells was acquired on the BD FACSCanto system (BD, San Jose, CA, USA). Mean fluorescence intensity of the control was taken as 100%. 

Localization of β-Catenin

The levels of β-catenin associated with the membrane and cytosol was assessed by immunofluorescence and Western blotting. Immunofluorescence was carried out as described above, using rabbit anti-β-catenin antibody (1:400; Sigma-Aldrich Chemical Co.), followed by anti-rabbit-FITC-labeled secondary antibody. Slides were visualized under the Olympus confocal microscope. For Western blotting, cytosolic and membrane fractions were prepared.⁴⁰ Following transfer of proteins to nitrocellulose membrane, the membrane was blocked, incubated with rabbit anti-β-catenin antibody (1:2000; Sigma-Aldrich Chemical Co.) for 1 hour, and then incubated with HRP-conjugated anti-rabbit antibody; membrane was developed using Bio-Rad ECL kit. 

Evaluation of Signaling Pathways

Lens epithelial cells were seeded at a density of 5 × 10⁵ cells per well of a six-well plate and incubated overnight. Cells were serum starved at 80% confluency, followed by treatment with the scFv antibodies or the respective inhibitors of individual pathways being studied: SMAD3(SIS3), ERK1(PD98059), p38(SB203580), and JNK(SP600125). Transforming growth factor-β2 (2 ng/mL) was added after 1 hour, and cells were further incubated for the time periods specified in Results. At the end of the incubation, cells were washed twice with PBS followed by cell lysis in 200 μL ice-cold RIPA buffer (25 mM Tris-HCl [pH 7.6], 150 mM NaCl, 1 mM Na₃VO₄, 1% Nonidet [N]P-40, 1% sodium deoxycholate, 0.1% SDS) along with protease inhibitor cocktail (Roche, Mannheim, Germany). Cell lysate was electrophoresed on 10% SDS-PAGE, followed by transfer to the nitrocellulose membrane. The membrane was blocked and probed with primary anti-pFAK (Y397, 1:500; Sigma-Aldrich Chemical Co.) antibody, anti-pSMAD3 (Y208; Sigma-Aldrich Chemical Co.), anti-pERK (T202/Y204; Cell Signaling, Beverly, MA, USA), pJNK (T183/Y185; Cell Signaling), or pp38 (T180/Y182; Cell Signaling) and secondary anti-rabbit-HRP conjugate (1:10,000). The membrane was developed using ECL. The same membrane was stripped again and probed with the corresponding antibodies for visualizing focal adhesion kinase (FAK), SMAD, p38, JNK, and ERK proteins, followed by probing with secondary HRP-conjugated anti-rabbit antibody or anti-mouse antibody (1:10,000). The membrane was developed using the ECL kit. Membrane was also probed with anti-beta-actin antibody as a loading control. 

In addition to Western blotting, nuclear phosphorylation of SMAD3 was also examined by immunostaining using the confocal microscope (Olympus). 

Statistical Analysis

All data are expressed as mean ± SD unless otherwise stated. Statistical analysis was performed by one-way ANOVA. A P value less than 0.05 was considered significant. 

Results

TN64 Reduces the Proliferation of Lens Epithelial Cells Without Significantly Affecting the Cell Viability

As described in the Methods section, TN64 was selected for binding against TNfnIII 1-5. The extent of MTT metabolism into formazone crystals was used to assess the viability of the LECs (HLE-B3) in the presence of TN64. There was no significant decrease in the viability of the cells in the presence of TN64 in the range of concentration of 0 to 50 μg/mL (Supplementary Fig. S4A). The BrdU incorporation assay showed that there was a significant decrease in proliferation starting from a concentration of 25 μg/mL in the presence of TN64; no such effect was observed in the presence of any concentration of the scFv O52 (irrelevant control antibody). Proliferation of the LECs in the absence of any scFv antibody was assigned the value 100% (Supplementary Fig. S4B). A concentration of 25 μg/mL for each scFv was chosen for the rest of the experiments. 

The scFv TN64 Has an Inhibitory Effect on the Migration of Lens Epithelial Cells

Significant reduction in migration of the HLE-B3 cell line and primary cultures of LECs grown from anterior lens capsule explants from patients was observed in the presence of TN64, as observed in the wound healing assay when performed over a time period of 24 hours (Fig. 1). In the control set, that is, LECs seeded in the presence of TGF-β2 alone, closure of the wound due to cell migration was assigned a value of 100%. In the presence of TN64, the closure of the wound was approximately 58% and 64% using HLE-B3 cells and primary epithelial cells, respectively, while 95% and 98% wound closure was seen in the presence of the irrelevant scFv O52 using HLE-B3 cells and primary HLE cells, respectively (Figs. 1A, 1B, respectively). 

Figure 1

 

Effect of TN64 antibody on the migration of the HLE-B3 cells (A) and primary culture of LECs (B) by the scratch assay. Images were recorded at 0 and 24 hours after addition of the antibody. ([A, B], right) The corresponding quantified data using t-scratch software. The control culture (minus any scFv antibody) was assigned the value of 100%. P < 0.05 was considered statistically significant and is represented by an asterisk. (C) Zymogram of culture medium supernatants obtained from primary culture of cells in (B) to estimate MMP2 and MMP9 activity; lane 1: no scFv; lane 2: scFv O52; lane 3: scFv TN64. M, molecular weight markers stained with Coomassie blue.

Single-Chain Fv TN64 Leads to Decrease in Expression of Matrix Metalloproteases

Since the migration of the cells was significantly reduced in the presence of TN64, we next evaluated the effect of this antibody on the expression of MMP-2 and MMP-9, which are directly related to the migration of the cells. TN64 was able to reduce the levels of both, MMP-2 as well as MMP-9, as compared to the control primary LECs (without scFv) and cultures of these cells treated with the irrelevant scFv O52 (Fig. 1C). 

TN64 Causes Changes in Actin Stress Fiber Formation

When HLE-B3 cells or primary cultures of LECs (cultured from capsule explants obtained from patients undergoing cataract surgery) were cultured in the presence of TGF-β2 (control) or O52 (irrelevant control antibody), extensive formation of F-actin stress fibers was observed (Fig. 2A). However, in the presence of TN64, both types of cultures showed a cortical pattern and significant attenuation of actin stress fiber formation. 

Figure 2

 

(A) Effect of TN64 on actin stress fiber formation. HLE-B3 cells and primary culture of LECs (derived from capsules from patients) were grown to confluency on gelatin-coated coverslips and incubated in the absence (control) or in the presence of scFv antibody, O52 (irrelevant control antibody), or the test antibody, TN64 (25 μg/mL each). The cells were fixed, detergent permeabilized, and treated with TRITC-phalloidin for actin stress fibers. Confocal images indicate the formation of extensive actin stress fibers running throughout the cell cytoplasm in the untreated set (control) and in the presence of the scFv O52 (irrelevant scFv). Rearrangement of the actin fibrils was observed in the presence of the TN64 antibody. The stress fibers localized more toward the periphery in the presence of the scFv antibody TN64, particularly in the case of primary cultures of lens epithelial cells. (B) Effect of TN64 on fibronectin assembly and deposition. Confocal images of fibronectin polymerization in HLE-B3 cells and primary culture LECs with and without scFv antibodies. Fibronectin was visualized using rabbit anti-fibronectin antibody, followed by FITC-conjugated anti-rabbit antibody. (C) Quantitation of extent of fibronectin deposition in the extracellular matrix estimated by deoxycholic acid (DOC) insolubility assay (left). DOC insoluble fraction, generated starting with the same number of cells in each case, was separated on SDS-PAGE and transferred to nitrocellulose membrane. The blots were developed using rabbit anti-fibronectin antibody by chemiluminescence (Amersham Biosciences). The figure is a representative image of three separate experiments. The lower blot in (C) shows the corresponding signals obtained using mouse anti-GAPDH antibody (loading control). (D) Densitometric quantitation corresponding to the Western blot in (C), analyzed by ImageJ software. The figure represents mean ± SD calculated from three separate experiments.

Fibronectin Polymerization Is Significantly Decreased in the Presence of TN64

Post cataract surgery, the remaining LECs undergo a wound healing response leading to increased expression of α-SMA-expressing myofibroblasts, as well as deposition of ECM proteins such as fibronectin and tenascin-C, as well as α-SMA and β1 integrin.⁴¹ The fibronectin type III repeats 1-5 of tenascin-C bind not only to a mature fibronectin matrix, but also to newly synthesized fibronectin fibrils.⁴² Further, it is also well understood that polymerized fibronectin is necessary for the induction of the myofibroblastic phenotype seen in PCO.²² Considering all of the above, it can be inferred that it is possible for an antibody directed against the TNfnIII 1-5 repeats to be able to block tenascin–fibronectin interactions, leading to inhibition of fibronectin polymerization, and thereby inhibit the transformation of LECs to a myofibroblastic phenotype. From Figure 2B, it is evident that TN64 is able to decrease the fibronectin assembly when compared with the situation in cells cultured in the presence of TGF-β2 alone (control, no scFv added). Thick, extensive, and continuous fibronectin fibrils are observed in the control population, as well as in the presence of O52 antibody, while the fibril formation is significantly affected in the presence of TN64. The extent of fibronectin deposition in the absence and presence of the scFv antibodies was evaluated by DOC insolubility assay (Fig. 2C); the level of fibronectin deposition was found to be significantly lower (68%) in the presence of scFv TN64 compared to the control (100%) or the irrelevant scFv O52 (90%) (Fig. 2D). 

TN64 Causes Marked Reduction in Expression of Other Markers of Epithelial–Mesenchymal Transition

Flow cytometry was done to evaluate the changes in expression levels of collagen and β-1 integrin (Supplementary Fig. S5). TN64 causes decreased expression of collagen type I (24%, Supplementary Fig. S5A) as well as β-1 integrin levels (29%, Supplementary Fig. S5B) as compared to cells cultured either in the absence of any scFv or in the presence of the irrelevant antibody, O52 (Figs. 3A, 3B). Further, TN64 was effective in reducing the expression of both N-cadherin and α-SMA in LECs (Figs. 3C, 3D). While the α-SMA expression was decreased to ~69% and 73% (Fig. 3C), the N-cadherin level was reduced to ~66% and ~72% (Fig. 3D) in HLE-B3 cells and primary cultures of LECs, respectively. Control cultures of both types of cells (with no scFv added) were assigned the value of 100%, and there was no significant decrease in the levels of N-cadherin and α-SMA in the presence of the irrelevant scFv O52. Similar results were seen in the immunostaining data shown for the two markers in Figure 3E. 

Figure 3

 

Patient-derived lens epithelial cells were cultured in complete medium followed by serum starvation for 8 to 10 hours. Cells were treated with or without scFv in the presence of TGF-β2 (2 ng/mL). (A, B) Mouse anti-collagen type I (1:50) and mouse anti-β1 integrin (1:100) antibodies were used, respectively, followed by incubation with FITC-conjugated anti-mouse antibody (1:200). Mean fluorescent intensity of the cells was acquired on BD FACSCanto system. Mean fluorescence intensity of the control was taken as 100%. (C, D) Mouse anti-N-cadherin (1:100) and mouse anti-α-SMA (1:200) antibodies were used, respectively, followed by incubation with FITC-conjugated anti-mouse antibody (1:200). Fluorescence intensity was quantified on VICTOR X3 multilabel ELISA reader (PerkinElmer). Data were plotted as percent fluorescence intensity. The figure represents mean ± SD calculated from three separate experiments. P < 0.05 was considered statistically significant. (E) Confocal images corresponding to (C, D).

Blocking of TNfnIII 1-5 Using TN64 Causes Altered Localization of β-Catenin

Since the β-catenin signaling pathway is involved in the pathogenesis of several fibrotic diseases through its role in the induction of EMT, we explored the possibility that TN64 may be influencing this pathway. During EMT, β-catenin accumulates in the cytoplasm, where it forms a complex with cytosolic transcription factors and is then shuttled to the nucleus, leading to activation of target genes associated with EMT. We examined the localization of β-catenin in TN64-treated and untreated cells. Immunostaining results showed that while the cells not treated with scFv (i.e., treated with TGF-β2 alone), as well as cells treated with scFv O52, showed mostly diffuse staining of β-catenin in the cytoplasm, β-catenin expression in the TN64-treated cells was mainly restricted to the cell membrane (Fig. 4A). To further confirm this finding, we lysed the cells, and the cytosolic and membrane fractions were examined for β-catenin by Western blotting using anti-β-catenin antibody. In line with the immunofluorescence data, Western blot also showed increased β-catenin in the membrane fraction of TN64-treated cells (Fig. 4B). The increased membrane localization of β-catenin is expected to strengthen the cadherin–catenin interactions.⁴³ We therefore examined the expression of E-cadherin in the cultures, and as expected, we observed colocalization of β-catenin and E-cadherin (Fig. 4A; colocalization in yellow in merged images) in cultures grown in the presence of TN64, unlike what was seen in the control cells (with no added scFv) or in the control cells grown in the presence of O52 antibody. 

Figure 4

 

(A) Immunostaining to study the colocalization of E-cadherin and β-catenin in the presence and absence of the scFv antibodies upon treatment with TGF-β2 (2 ng/mL). Lens epithelial cells derived from the patients' capsules were cultured and treated with scFv TN64 and O52 in the presence of TGF-β2 (2 ng/mL) for 24 hours. Control set refers to the condition in which cells were treated with TGF-β2 (2 ng/mL) alone. Rabbit anti-β-catenin and rat anti-E-cadherin antibody were used as primary antibodies, followed by FITC-labeled anti-rabbit and TRITC-labeled anti-rat antibodies used as secondary antibodies, respectively. DAPI was used as a nuclear dye. In the presence of TN64, colocalization of β-catenin and E-cadherin is seen in yellow along the cell membrane. (B) Western blot for examination of β-catenin in cytosolic and membrane fractions, with anti-β-catenin antibody. In the control set, β-catenin appears to be more cytoplasmic as compared to the cells treated with TN64, where β-catenin is mostly cell membrane–associated. The cells treated with the irrelevant antibody, O52, appear to be similar to the control cells.

TN64 Does Not Act Through the SMAD Signaling Pathway

Transforming growth factor-β receptor signaling can act through two possible pathways, SMAD-dependent or SMAD-independent (but adhesion-dependent) pathways.⁴⁴ In order to examine whether the SMAD signaling pathway is involved in the action of TN64, we cultured the primary LECs in the presence of the scFv antibodies as well as independently in the presence of SIS3, a SMAD3 inhibitor (positive control) that blocks the phosphorylation of SMAD3. As seen from Figure 5A, phosphorylation of SMAD3 was unaffected in all the cultures, ranging from time periods of 15 minutes to 24 hours, except in the positive control (SIS3), which showed inhibition of phosphorylation 30 minutes post TGF-β addition. These data were further confirmed by immunostaining, which showed that TN64 had no effect on nuclear translocation of pSMAD3 (Fig. 5B). It can be concluded from the above data that TN64 does not interfere with the phosphorylation of SMAD3. 

Figure 5

 

Effect of TN64 on TGF-β2-induced signaling to monitor SMAD3 signaling (A, B) and FAK (C), p38 (D), ERK (E), and JNK phosphorylation (F) on the human lens epithelial cells. Primary cultures of LECs were grown to 70% to 80% confluency in the presence of complete medium overnight, followed by serum starvation for 8 to 10 hours and treatment with scFv antibodies or the respective inhibitors (INH) as detailed in Materials and Methods. Control set indicates absence of scFv treatment. (A) Immunoblots for time-dependent SMAD3 signaling assessed by probing with phospho-SMAD3, SMAD3, and beta-actin (internal loading control) antibodies, respectively. SIS3, a specific SMAD3 inhibitor, was used as a positive control. (B) Confocal images of LECs treated as above and incubated with anti-phospho-SMAD3 antibody followed by FITC-conjugated secondary antibody. (C–F) Immunoblots showing reduction in phosphorylation of FAK, p38, ERK, and JNK with TN64 as compared to the control. The level of total proteins in each case was unaltered in either the presence or the absence of the scFv antibody. Beta-actin was used as loading control.

TN64 Operates Through Inhibition of the Noncanonical Signaling Pathways

Since tenascin-C can induce EMT-like changes through phosphorylation of FAK,⁴⁵ we wished to investigate whether blocking of TNfnIII 1-5 repeats by TN64 can cause downregulation of p-FAK. As can be seen from the immunoblot in Figure 5C, the level of FAK remained constant when cells were cultured in the presence of TGF-β2 alone or in the presence of TN64 or O52; however, cells treated with TN64 in the presence of TGF-β2 showed reduction in phosphorylation of FAK (at Y397). We also examined whether the inhibition of EMT-like changes by TN64 was mediated by the SMAD-independent, mitogen-activated protein kinase (MAP kinase) pathway. The phosphorylation status of all three kinases, p38, JNK (c-Jun N-terminal kinase 1-3), and ERK (extracellular signal regulated-kinase), was assessed in the presence and absence of TN64 (Fig. 5); as seen in Figures 5D through 5F, the phosphorylation of all these kinases was inhibited when cultured in the presence of TN64, unlike the irrelevant antibody, O52. 

Discussion

Transforming growth factor-β is a major inducer of EMT in the lens,³³ which is responsible for myofibroblast formation in the wound healing process operating post cataract surgery. In this paper, we show the EMT-inhibiting property of an antibody (TN64) directed against the fibronectin type III-like repeat (1-5) in tenascin. Tenascin has been shown to play key roles in cell adhesion, spreading, and migration of varied kinds of cells.⁴⁶ We generated TN64 through phage display antibody library screening methods and found that it specifically recognizes the TNfnIII 1-5 region. More accurately, it binds independently to TN1-3 and TN3-5, but binding to TN3 is very weak. We chose this target since it has been shown previously that (1) tenascin-C expression is upregulated upon EMT of LECs in human PCO³³; (2) in the absence of tenascin-C, injury-induced EMT of the mouse lens epithelium is attenuated²⁵; (3) multiple growth factors bind with high affinity to this very repeat region⁴⁷; and (4) TNfnIII 1-5 has the ability to bind to different domains within fibronectin, with potential to cross-link the ECM by interacting with neighboring fibronectin fibrils, thereby promoting fibronectin polymerization.⁴² Increased fibronectin polymerization induces the expression of MMP-9 and MMP-2,^48,49 which can cause cleavage of collagen type IV (of the basement membrane) on the one hand, resulting in increased cell migration; on the other hand, this can also cause cleavage of TGF-β-binding protein, thereby releasing TGF-β from its latent form and causing its activation (Fig. 6). It must be noted, however, that this pathway would be operational in in vivo situations; in our system (ex vivo), active TGF-β has been used to induce EMT. Nevertheless, this highlights the importance of the repeat region of tenascin-C in modulating the ECM and also cell–ECM interactions during the wound healing process leading to PCO (Fig. 6). 

Figure 6

 

Schematic diagram highlighting the importance of fibronectin type III-like repeats 1-5 of tenascin-C in the context of fibrosis. Repeats 1-5 are known to bind to different domains within fibronectin, as well as growth factors such as platelet-derived growth factor (PDGF), FGF, TGF-β, and neurotrophins, resulting in fibronectin polymerization and ECM modulation. Increased fibronectin polymerization results in adhesion-dependent cell growth and increased MMP expression. The latter causes cleavage of collagen IV in the basement membrane and thereby facilitates cell migration. In addition, the increased MMP-2 expression causes cleavage of TGF-β-binding protein, resulting in activation of TGF-β, which is available to further bind to tenascin-C and also facilitate the EMT process.

It is known that TGF-β2 levels increase dramatically after cataract surgery, potentially triggering PCO,⁷ with TGF-β2 known to be 10 times more effective than TGF-β1 in causing changes associated with human subcapsular cataract.⁵⁰ We tested the effects of TN64 on two types of LECs exposed to TGF-β2: the lens epithelium cell line HLE-B3 and primary LECs grown from explants obtained from patients undergoing cataract surgery. It may be noted that results showing TN64-mediated decreased migration of LECs (Fig. 1), actin stress fiber formation (Fig. 2), and fibronectin polymerization (Fig. 2) were all more pronounced with primary cultures of cells (compared to HLE-B3 cell line–derived cells), emphasizing the importance of using such cells since they are more akin to the physiological (pathologic) setting. 

Transforming growth factor-β is known to induce EMT through the canonical SMAD-dependent (adhesion-independent) or the noncanonical SMAD-independent (adhesion-dependent) pathways. Our data clearly show that the TN64-mediated inhibition of EMT status is SMAD independent. Importantly, it has been reported that suppression of injury-induced EMT in tenascin-C null mice is associated with attenuation of phosphorylation of SMAD signaling.²⁵ This is not in conflict with our results; while the previous work involved mice lacking the whole molecule of tenascin-C, the present work is different on two counts. First, unlike the case with the previous work, this is an ex vivo system in which active TGF-β2 was used; however, in vivo, TGF-β2 is largely present in its latent form, whose activation is supported by reactive oxygen species (ROS), specific integrins, thrombospondin-1 (TSP-1), or proteases.^51–54 Secondly, unlike the previous work, in which tenascin was knocked out completely, in our study, tenascin-C was targeted by an antibody directed at only the first five fibronectin type III repeats (TNfnIII 1-5). In light of these differences, it may be speculated that the two sets of experiments could lead to different cellular effects of tenascin. It is therefore conceivable that an inhibitor such as TN64 causes downregulation of EMT, leaving the SMAD signaling pathway intact. 

It has been previously demonstrated that TGF-β-induced EMT can occur through the noncanonical integrin-dependent autophosphorylation and activation of FAK⁵⁵ or through the MAP kinase pathway.⁵⁶ Several of our results also indicate the involvement of the FAK pathway in the inhibition by TN64 (Fig. 7): (1) There is an inhibition of Y-397 phosphorylation of FAK by TN64 in the presence of TGF-β; (2) phosphorylation of FAK is associated with the upregulation of integrin receptors and fibronectin,⁵⁷ changes in actin stress fiber formation, organization of fibronectin matrix, and cell migration,⁵⁸ all of which are downregulated in the presence of TN64; (3) FAK signaling is known to be required for delocalization of membrane-bound E-cadherin,⁵⁹ and our results show that TN64 causes downregulation of the mesenchymal marker N-cadherin and causes re-expression of the cell–cell junction adhesion protein E-cadherin; (4) FAK signaling is also implicated in fibronectin-dependent upregulation of MMPs mediated through integrins⁶⁰—our data show that TN64 is capable of inhibiting the expression of both MMP2 and MMP9. Since FAK can link integrin-initiated signals to the mitogen-activated protein kinase (MAPK) pathway,⁶¹ we examined whether TN64 has any effect on the phosphorylation of these kinases, and, as expected, phosphorylation of p38, ERK, and JNK was inhibited by the antibody. Further, since EMT initiates in the lens through the activation of Wnt-responsive genes following the nuclear translocation of β-catenin in LECs,⁶² and FAK is also required downstream of Wnt signaling,⁶³ we examined whether TN64 also operates through the Wnt signaling pathway involving β-catenin. We found that TN64 supports membrane localization of β-catenin where it localizes with the epithelial (membrane-associated, E-cadherin-associated) phenotype. 

Figure 7

 

Transforming growth factor-β is a key mediator of fibrosis. On the one hand, it regulates the integrin/FAK pathway; increased phosphorylation of FAK causes increased levels of expression of integrin and fibronectin, as well as disruption of E-cadherin and β-catenin on cell–cell junctions, along with increased expression of MMP levels resulting in increased cell migration. Further, FAK can link integrin-initiated signals to the mitogen-activated protein kinase (MAPK) pathway. The figure illustrates the various key steps being regulated by the TN64 antibody.

In the context of fibrosis, blockade of TGF-β would seem attractive; however, this cytokine plays an important role in control of cell proliferation and tumor suppression.⁶⁴ Moreover, a population-based cohort study has shown increased incidence of epithelial cell malignancies in fibrotic diseases⁶⁵; therefore, it would be desirable to design an antifibrotic strategy that does not adversely affect a pathway involved in tumor suppression. Importantly, targeted deletion of FAK in mammary epithelium has been shown to suppress mammary tumorigenesis.⁶⁶ The integrin/FAK pathway has therefore been proposed to be a safer and more effective therapeutic strategy for fibrotic diseases.⁵⁵ 

In summary, our results suggest that (1) the fibronectin type III-like repeats in tenascin-C may be playing an important role in PCO; (2) TN64 is capable of negating the effects of TGF-β2 in LECs; (3) TN64 acts by inhibiting the TGF-β-induced phosphorylation of FAK and nuclear translocation of β-catenin; and (4) TN64 is capable of inhibiting the MAP kinases. Therefore, our results are in support of the need to target the codependent Wnt signaling and FAK signaling pathways, independent of the SMAD pathway, in that TN64 clearly operates through the former and not the latter while inhibiting EMT-associated changes. 

Acknowledgments

Supported by the Department of Science and Technology and the Council of Scientific and Industrial Research, Government of India (MLG), and a Council of Scientific and Industrial Research fellowship (AT). 

Disclosure: A. Tiwari, None; J. Ram, None; M. Luthra-Guptasarma, None