Glucosamine inhibits epithelial-to-mesenchymal transition and migration of retinal pigment epithelium cells in culture and morphologic changes in a mouse model of proliferative vitreoretinopathy

Chang‐Min Liang, Ming‐Cheng Tai, Yun‐Hsiang Chang, Yi‐Hao Chen, Ching‐Long Chen, Da‐Wen Lu, Jiann‐Torng Chen

Acta Ophthalmologica · 2011 · 46 citations · 29 references

DOIFull text

Open access

Concepts

Abstract

Purpose: To explore the effect of glucosamine (GlcN) on transforming growth factor (TGF)-β signalling and several processes involved in proliferative vitreoretinopathy (PVR). Methods: We evaluated the surface levels of TGF-β receptor and its binding of TGF-β in ARPE-19 cells. Release of cytokines and collagen, and expression of signalling intermediates were quantified. Migration was qualitatively and quantitatively examined. The morphology of cells undergoing PVR in vitro and in a mouse PVR model was observed. Results: Glucosamine reduced the surface levels of TGF-β receptor and the ability of ARPE-19 cells to bind TGF-β. In ARPE-19 cells, TGF-β1 plus epidermal growth factor (EGF) or TGF-β2 increased the expression of alpha-smooth muscle actin (α-SMA) and decreased the expression of zona occludens protein (ZO-1). Transforming growth factor-(β2) also caused the release of platelet-derived growth factor (PDGF), connective tissue growth factor (CTGF) and type 1 collagen and increased the phosphorylation of SMAD2 and SMAD3. Platelet-derived growth factor and CTGF stimulated cell migration, and TGF-β2 stimulated wound closure, contraction of collagen and changes in cell morphology. Conclusions: Treatment with GlcN counteracted all of these effects, and its administration in the mouse model reduced the morphologic appearance of PVR. Glucosamine could inhibit the TGF-β signalling pathway in retinal pigment epithelium cells and several of the downstream events associated with epithelial–mesenchymal transition and PVR. Proliferative vitreoretinopathy (PVR) is defined as the growth and contraction of cellular membranes within the vitreous cavity and on both surfaces of the retina after rhegmatogenous retinal detachment (Committee TRST 1983; Pastor 1998). Despite recent progress in surgical techniques, PVR is the most common cause of treatment failure in rhegmatogenous retinal detachment (Machemer 1988). Although antiproliferative and anti-inflammatory agents are the major candidates for the pharmacologic suppression of PVR (Esser et al. 1998; Yang et al. 1998; Asaria et al. 2001; Charteris et al. 2002; Eibl et al. 2006), their poor efficacy and multiple side-effects limit their use as a routine treatment strategy (Kirchhof 2004; Scheer et al. 2004). The pathogenesis of PVR is dependent on growth factors and cytokines, including transforming growth factor (TGF)-β (Cassidy et al. 1998; Kon et al. 1999; Hinton et al. 2002; He et al. 2008). We hypothesized that PVR might be affected by agents that inhibit the synthesis and processing of N-linked oligosaccharides, on which the functioning of membrane-bound receptors for growth factors and cytokines is dependent (Partridge et al. 2004; Takahashi et al. 2004). Previously, we showed that glucosamine (GlcN) can reduce the N-glycans of surface receptors, such as EGFR (Liang et al. 2010). Therefore, in seeking novel therapeutic approaches for PVR, it seemed appropriate to also explore the effects of this agent on the TGF signalling pathway. In this study, we examined the effect of GlcN on the TGF-β receptor and on several of the processes known to be involved in PVR and set in motion by TGF. To examine the biochemical events associated with PVR, we used cultured ARPE-19 cells, and to confirm that GlcN affected the processes involved in PVR, we used an in vivo mouse model system. The human retinal pigment epithelial cell line (ARPE-19) was obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA) and maintained in Dulbecco’s modified Eagle’s medium (DMEM)–F-12 (Gibco, Carlsbad, CA, USA) supplemented with 4 mm l-glutamine, 10% fetal bovine serum (FBS; Gibco), 100 U/ml penicillin and 100 μg/ml streptomycin (Sigma-Aldrich, St. Louis, MO, USA). Cells were cultured at 37°C in the presence of 5% CO2. The duration of culture was 1–2 days (until they reached 80% confluence) for the experiments that involved ELISA of platelet-derived growth factor (PDGF)-BB and connective tissue growth factor (CTGF), flow cytometric examination of surface levels of the TGF-β receptor and ligand–receptor binding, the type I collagen ELISA, and the western blots for expression and phosphorylation of SMAD2/3. The ARPE-19 cells were cultured 3–4 days (until they reached monolayer confluence) for the evaluation of wound closure and immunostaining and western blots to detect alpha-smooth muscle actin (αSMA) and zona occludens protein (ZO-1). The culture medium was replaced twice weekly. Cells were induced to undergo epithelial–mesenchymal transition (EMT) with 10 ng/ml TGF-β2 (PeproTech, New York, NY, USA) for 24, 48 and 72 hrs, as previously described (Okada et al. 1997). ARPE-19 cells were cultured to 80% confluence. The medium was changed, and the cells were cultured with the addition of 2 μg/ml tunicamycin, 30 mm N-acetyl-glucosamine (GlcNAc), 30 mm glucose, 2.5 mm GlcN or 5 mm GlcN in medium containing 10% FBS for a further 24 hrs. The cells were harvested with PBS-based enzyme-free dissociation buffer (Invitrogen, Gibco BRL, Gaithersburg, MD, USA), a membrane-filtered, isotonic and enzyme-free aqueous formulation of salts, chelating agents (EDTA) and cell-conditioning agents constituted in Ca2+- and Mg2+-free PBS. The cells were incubated either with anti–TGF-β receptor type II antibody (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and FITC-conjugated secondary antibody (1:200; Jackson ImmunoResearch Laboratories, West Grove, PA, USA), or they were incubated in the presence of biotinylated TGF-β1 (Fluorokine biotinylated human TGF-β1 kit; R&D Systems, Minneapolis, MN, USA) for 1 hr and avidin-FITC was added. The fluorescence was measured by flow cytometry. Confluent monolayers of ARPE-19 cells were serum starved for 24 hrs and pretreated with 10 μg mitomycin-C for 2 hrs before inflicting a scratch wound on the monolayer with a p20 pipette tip. The cells were treated with 10 ng/ml TGF-β2 in the absence and presence of GlcN (Sigma-Aldrich) either at 2.5 or 5.0 mm, and the cells’ ability to migrate and close the wound space was assessed by light microscopy at 24, 48 and 72 hrs after the application of the scratch. Migration was also measured with a modified Boyden chamber assay, as previously described (He et al. 1998). Briefly, ARPE-19 cells were seeded at a density of 5 × 104/well in the upper chamber of a fibronectin-coated 24-well plate with an 8-μm transwell pore (Corning, Corning, NY, USA). The lower chamber was filled with 0.1% FBS-DMEM-F12 medium containing 10 ng/ml rhCTGF (PeproTech) or 10 ng/ml PDGF-BB (PeproTech). After a 5-hr incubation, the inserts were washed with PBS, fixed with cold methanol (4°C) for 10 min and counterstained with haematoxylin for 20 min. The number of migrated cells was counted by phase-contrast microscopy. Four randomly chosen fields were counted per insert. Subconfluent ARPE-19 cells in a 24-well plate were serum starved for 24 hrs and stimulated with 10 ng/ml TGF-β2 with or without GlcN for 48 hrs. Cell culture supernatants were collected, and PDGF-BB was quantified with the commercially available Quantikine® Human PDGF-BB Immunoassay kit (R&D Systems). Connective tissue growth factor was quantified with a sandwich ELISA with CTGF-specific monoclonal antibodies (anti-human CTGF N-terminal peptide antibody; Santa Cruz Biotechnology, anti-human CTGF C-terminal peptide antibody; R&D Systems). Ninety-six-well plates were coated overnight at 4°C with anti-human CTGF N-terminal peptide antibody [1:200 dilution in coating buffer (0.05 m sodium bicarbonate, pH 9.6)]. The plates were blocked with 1% bovine serum albumin for 1 hr at room temperature and washed with TBST. Cell culture supernatants (100 μl) were added to each well and incubated for 1 hr at 37°C. Plates were washed with TBST and incubated for 1 hr at 37°C with monoclonal anti-human CTGF C-terminal peptide antibody [1:200 dilution in assay buffer containing 0.05 m Tris (pH 7.8), 0.1% BSA, 4 mm MgCl2, 0.2 m ZnCl2, 0.1% sodium azide, 50 mg/l sodium heparin and 0.1% Triton X-100]. Plates were washed with TBST and incubated for 1 hr at 37°C with alkaline phosphatase–conjugated secondary antibody (1:200; Jackson ImmunoResearch Laboratories). Plates were washed and developed with p-nitrophenyl phosphate (1.5 mg/ml; Sigma-Aldrich) in diethanolamine buffer (1 m diethanolamine, 0.5 mm MgCl2, and 0.02% sodium azide), and the optical density was read at 405 nm. Purified rhCTGF was used as a standard. For western blot analysis, the cells were treated with 10 ng/ml TGF-β2 or TGR-β1 + EGF for 72 hr in the absence and presence of GlcN (Sigma-Aldrich) either at 2.5 or 5.0 mm. Treated and untreated cells were washed with PBS, harvested by scraping and centrifuged at 1000 × g. Cell pellets were resuspended and sonicated in cold lysis buffer comprising 50 mm Tris–HCl (pH 7.5), 2% SDS, 1 mm phenylmethylsulfonyl fluoride and 10 μl/ml protease inhibitors (Sigma-Aldrich). Lysates were centrifuged at 12 000 × g for 10 min, and the protein concentration in the clear supernatant was determined with the BCA protein assay kit (Pierce, Rockford, IL, USA). Lysates (20 μg) were resolved by 10% SDS/PAGE and transferred to PVDF membranes. Membranes were blocked with 5% (w/v) milk for 1 hr at room temperature and then incubated at room temperature for 60 min with 1:1000 dilution of antibodies against β actin (Santa Cruz Biotechnology), αSMA (Sigma-Aldrich), ZO-1 (Zymed Laboratories, South San Francisco, CA, USA) or 1:2000 dilutions of anti-SMAD2/3 and anti-phosphorylated SMAD2/3 antibodies (Cell Signaling Technology, Danvers, MA, USA). Membranes were washed and incubated with horseradish-peroxidase-conjugated secondary antibody (1:1000; Jackson ImmunoResearch Laboratories) for 1 hr at room temperature, and protein was visualized using an enhanced chemiluminescence procedure (enhanced chemiluminescence reagent; Millipore, Billerica, MA, USA). For immunocytochemistry, the cells were treated with 10 ng/ml TGF-β2 for 72 hr in the absence and presence of GlcN (Sigma-Aldrich) at 5.0 mm. Cells were fixed with paraformaldehyde and stained with antibodies against ZO-1 (Zymed Laboratories) and vimentin (Santa Cruz Biotechnology). Nuclei were counterstained with PI. We analysed cell-populated three-dimensional collagen matrix contraction using a modification of a previously described method (Mazure & Grierson 1992). Rat tail type I collagen (Sigma-Aldrich) was dissolved in 0.1% acetic acid and at 4°C of a 24-well plate were overnight with 2% FBS to ARPE-19 cells were resuspended at a density of 1 × in The cell was with 5.0 of collagen (Sigma-Aldrich) and of medium containing (100 U/ml penicillin and 100 μg/ml and 1 m The was seeded in a 24-well plate and incubated at 37°C for 1 hr to The collagen were from the of the after hrs, and the were in 1 containing 10% After 24 hrs, the medium was and the were washed with and incubated at 37°C for days in containing 10 ng/ml TGF-β2 with or without The medium was without cells were used to The surface of each matrix was and measured with the The of contraction was as at 1 at at × The was in Subconfluent ARPE-19 cells were starved of serum for 24 hrs and stimulated with 10 ng/ml TGF-β2 with or without GlcN for 48 hrs. The collagen in the was with and then with at pH to collagen to were after and ELISA was with the human collagen kit USA). The was The PVR model was in the of each mouse as described et al. The was used as the The were with an of and detachment was induced in the with the of a surgical after application of and was in the the The was with and the vitreous was after which the vitreous cavity was filled with sodium and the retina was using without the pigment After the addition of sodium to the the was with at was the and the were to for of to The were 1000 GlcN or the an each 24 hrs after and 1 before the were and was chosen on the of in which we a of GlcN as as 1000 in et al. 2008). this GlcN the of in the and or from each were examined at 10 and days after the surgical The were by and was fixed in paraformaldehyde in m phosphate buffer and in The were and stained with haematoxylin and were by the and of the and were in with the for in and for the of in and were by of a was multiple of were with the are as is of at were and evaluated at the 0.05 of were with IL, USA). To with GlcN could reduce the levels of TGF-β on the surface of ARPE-19 cells, we examined the binding of antibody against the TGF-β receptor by flow cytometry. in the addition of GlcN reduced the surface levels of TGF-β receptor in a The levels of TGF-β receptor in the presence of GlcN were to was blocked by the addition of The ability of the cells to bind TGF was also reduced by the presence of GlcN and levels of or glucose, on the increased the levels of TGF-β receptor and its binding and as a for to examine GlcN effects on downstream processes that PVR and were stimulated by TGF-β. of glucosamine (GlcN) on the surface levels of transforming growth factor (TGF)-β receptor and ligand–receptor binding Cells were cultured with the agents for 24 and cells were incubated with receptor type II antibody and FITC-conjugated secondary antibody and analysed by flow cytometry. The cells were incubated in the presence of biotinylated avidin-FITC was and the fluorescence was measured by flow cytometry. The are of at The for the surface levels of TGF-β receptor are as 0.05 GlcN 2.5 The for ligand–receptor binding ability are as in of ARPE-19 cells with TGF-β1 plus EGF or TGF-β2 increased the expression of and and decreased the expression of ZO-1 and effects were counteracted by GlcN in a of transforming growth factor plus EGF or TGF-β2 on expression of alpha-smooth muscle actin (α-SMA) and zona occludens protein and blot of from ARPE-19 cells, treated for 72 hrs with TGF-β1 + EGF and of expression and of ZO-1 expression were after treatment with TGF-β1 + EGF and Glucosamine the of expression and maintained the expression of ZO-1 in a for and ZO-1 expression are as 0.05 of ARPE-19 cells after treatment with TGF-β2 for 72 hrs. Cells were fixed with paraformaldehyde and stained with antibodies against ZO-1 and Nuclei were counterstained with PI. The are of at Treatment of ARPE-19 cells with TGF-β2 caused the release of levels of and CTGF the medium with untreated cells addition of either 2.5 or 5 mm GlcN caused a in the release of and CTGF in these cells. Release of platelet-derived growth factor and connective tissue growth factor Treatment with TGF-β2 induced the release of and CTGF from ARPE-19 cells the as determined by Treatment with 2.5 or 5 mm glucosamine (GlcN) this and phosphorylation of SMAD2 and SMAD3. ARPE-19 cells were pretreated with 2.5 or 5.0 mm GlcN for 24 hrs. Cells were treated with TGF-β2 for 60 min, and cell were with anti-phosphorylated and antibodies to SMAD2 and SMAD3. was with a The are of at of type 1 collagen levels in cells treated with TGF-β2 was with cells treated with TGF-β2 in the presence of 2.5 or 5 mm as determined by blot with antibodies as well as antibodies to SMAD2 and showed that treatment of ARPE-19 cells with TGF-β2 in an in the levels of the of SMAD2 and and which was by of the cells with GlcN Treatment of ARPE-19 cells with TGF-β2 in a in the levels of the type I collagen in the medium to cells The of 2.5 or 5 mm GlcN counteracted the and the expression of type 1 collagen to a to that of the To examine the ability of ARPE-19 cells to close the space after a wound cells were starved of serum for 24 hrs, pretreated with 10 μg mitomycin-C for 2 hrs cell with a scratch and at 24, 48 and 72 hrs in the absence and presence of TGF-β2 We that cells were to close the wound space at 48 and 72 hrs of treatment with with cells. the presence of 2.5 or 5 mm of GlcN the closure of the wound space in a The closure of the wound space in the absence of cell that the cells migrated in to TGF-β. of transforming growth factor and glucosamine (GlcN) on closure of wound space after ARPE-19 cells were of ARPE-19 cells were at 24, 48 and 72 hrs after of a scratch. The untreated cells. Cells were treated with TGF-β2 in the absence or presence of The the and were added by a to the of × The are of at The closure of the wound space in the that the cells migrated in to TGF-β. We used a modification of the Boyden chamber method to cell in ARPE-19 cells treated with or which also of cells and in induced by TGF-β. We that 2.5 and 5 mm of GlcN caused a of cell in ARPE-19 cells stimulated with or CTGF in a of cell ARPE-19 cells were treated with platelet-derived growth factor or connective tissue growth factor in the absence and presence of glucosamine and their quantified with a modified Boyden chamber contraction was measured with collagen containing ARPE-19 cells, which were incubated with TGF-β2 in the absence or presence of We a in the of the by the cells treated with TGF-β2 with the both The contraction in to TGF-β2 was counteracted by the presence of GlcN in a Glucosamine at 5.0 mm caused a in the by cells. of glucosamine (GlcN) on transforming growth factor collagen ARPE-19 cells were treated with TGF-β2 in the absence or presence of 2.5 or 5 mm or was determined as a of the Treatment of ARPE-19 cells with TGF-β1 and EGF for 72 hr caused the appearance of the cells to from an epithelial monolayer to a morphology Cells treated with GlcN in an monolayer The morphology of ARPE-19 cells after 72 hrs of with transforming growth factor and EGF in the presence or absence of glucosamine the in morphology of ARPE-19 cells from an epithelial monolayer to a morphology after 72 hrs of treatment with TGF-β1 and Cells treated with GlcN in an monolayer × We used to that the morphologic effects in the mouse model of PVR were to in the cultured cells. in of from the and treatment were stained with and examined at days 10 and after retinal we that the cells a in the PVR at 10 which increased in at In treatment with GlcN caused the cells to in a monolayer at 10 and a cellular at examination of tissue from and treatment with of the mouse proliferative vitreoretinopathy of were examined at 10 and after retinal 20 We that treatment with GlcN reduced the levels of the TGF-β receptor and the binding of TGF-β on the surface of ARPE-19 cells. with an of TGF-β signalling by we also that the presence of GlcN counteracted of the effects of TGF-β on expression or release of associated with and the of PVR, for type I collagen and the of SMAD2 and SMAD3. In GlcN counteracted several of cultured cells, such as the ability of ARPE-19 cells to close a wound space in the absence of the cellular caused by and and the contraction of collagen induced by GlcN the morphologic effects of TGF-β2 on cells both in culture and in an in vivo mouse model of PVR. Proliferative which is a major of rhegmatogenous retinal detachment is by of retinal pigment epithelial cells a of et al. 2008). Transforming growth factor-(β2) is in the vitreous and of with PVR, and it a in the pathogenesis of PVR et al. et al. TGF-β2 previously to in cells et al. et al. et al. 2004; et al. Therefore, in seeking novel therapeutic approaches for PVR, it seemed appropriate to explore agents that with the TGF signalling pathway. Glucosamine the and processing of N-linked et al. & which are for the functioning of membrane-bound receptors for growth factors and cytokines (Partridge et 2004; Takahashi et al. 2004). in this study, we the effect of GlcN on the TGF-β receptor and on several of the processes known to be involved in PVR and set in motion by TGF. We that the levels of the TGF-β receptor and the binding of TGF-β to the surface were decreased in a by treatment of ARPE-19 cells with that GlcN also the processes that as a of the binding of TGF-β to its Transforming growth to a in and in a of epithelial cell and in cells et al. et al. 1999; et al. et al. 2004; et al. In this study, we that GlcN counteracted the increased expression of and the decreased expression of ZO-1 induced by TGF-β. muscle actin is a that the transition of cells as they the wound of PVR & on the is an epithelial the of which also that cells are in the of & that the of the TGF-β signalling pathway by GlcN could the that is of PVR. cells both and to and which in the and of cells et We that treatment of ARPE-19 cells with TGF-β2 increased the release of and CTGF the a that was also by the presence of that of the TGF-β signalling pathway by GlcN could also the release of these and which previously to be associated with et al. et al. et al. 2004; et al. downstream processes in the TGF-β signalling pathway that also to in the of including phosphorylation of the of and release of type 1 collagen, were also by the treatment of ARPE-19 cells with GlcN in and the tissue contraction are dependent on the signalling pathway et al. are factors that by the from the TGF-β receptor to the to et al. 2004). and were all dependent on signalling in a mouse model of retinal detachment et al. TGF-β signalling PDGF-BB to an secondary proliferative PVR et al. 2004). the of phosphorylation of could the effects of GlcN on the release of and type 1 collagen in that GlcN the phosphorylation of SMAD2 and that its suppression of TGF-β signalling at an of the to the which also the that GlcN the TGF-β the changes we as a of the suppression of TGF-β signalling by GlcN a in the of PVR, we treatment with GlcN to of the morphologic and changes that in cells PVR. In this we that GlcN the closure of the wound space caused ARPE-19 cells were the cells were to the of this we that the wound closure was by In of this we that the of ARPE-19 after treatment with and CTGF was by GlcN as determined by the of the assay that the of the receptors for and CTGF was also by which further that this suppression of from in of surface receptors caused by processes that were affected by GlcN in collagen contraction and the morphologic changes of the all of the experiments were in cultured cells, it was to also an effect of GlcN in For we used a mouse model for PVR, in which retinal detachment was in the and we examined the effects of GlcN on the at 10 and days after The in of the showed morphologic changes of PVR, the of the treated with GlcN was a to an of the for GlcN to the of PVR. experiments are to the of the by which GlcN the TGF-β signalling pathway and the processes involved in PVR. be that the of GlcN used for the cells in culture were chosen on the of the in these that which are obtained in is to the effects of on the effects by GlcN to the of the GlcN the in vivo of the mouse that is with the we obtained in it is to that the mouse model of PVR we used is for which it is to the Therefore, and in the in vivo to the growth factors and cytokines that in the vitreous and after of the all the of PVR Although the in vivo the effect of GlcN on treatment of PVR, it to and these for to examine the effects of GlcN on the contraction of collagen associated with PVR in the also be to explore the of signalling the and in the pathogenesis of PVR and the effect of GlcN on these In we used GlcN to inhibit phosphorylation of SMAD2 and cell and tissue contraction in which are all events in the pathogenesis of PVR. Glucosamine also the morphologic effects of TGF-β2 on cells both in culture and in an in vivo mouse model of PVR. the of the of this study, we that GlcN could inhibit the TGF-β signalling pathway in cells and several of the downstream events that are associated with the and the of PVR. The of the to be PVR is an of for rhegmatogenous retinal pharmacologic for PVR and with its as a therapeutic of further was in by and from the

References

29