Conflict/competing interest: No stated conflict of interest.
Regulation of cell-mediated collagen gel contraction in human retinal pigment epithelium cells by vascular endothelial growth factor compared with transforming growth factor-β2
Article first published online: 18 AUG 2011
© 2011 The Authors. Clinical and Experimental Ophthalmology © 2011 Royal Australian and New Zealand College of Ophthalmologists
Clinical & Experimental Ophthalmology
Volume 40, Issue 1, pages e76–e86, January/February 2012
How to Cite
Ma, J., Zhang, Q., Moe, M. C. and Zhu, T. (2012), Regulation of cell-mediated collagen gel contraction in human retinal pigment epithelium cells by vascular endothelial growth factor compared with transforming growth factor-β2. Clinical & Experimental Ophthalmology, 40: e76–e86. doi: 10.1111/j.1442-9071.2011.02618.x
Funding sources: National Natural Science Foundation of China (J20080843), Zhejiang Provincial Natural Science Foundation of China (J20050897), and Medical Scientific Research Foundation of Zhejiang Province, China (2004A047).
- Issue published online: 5 FEB 2012
- Article first published online: 18 AUG 2011
- Accepted manuscript online: 13 JUN 2011 11:52AM EST
- Received 23 March 2011; accepted 19 May 2011.
- collagen gel;
- RPE cell;
Background: To explore the potential role of vascular endothelial growth factor compared with transforming growth factor-β2 in the regulation of human retinal pigment epithelium cell-mediated collagen gel contraction.
Methods: The retinal pigment epithelium cell mediated type I collagen gel contraction assay was performed to evaluate and compare the effect of vascular endothelial growth factor and transforming growth factor-β2. The number of viable retinal pigment epithelium cells in the gel and the expression of α-smooth muscle actin were analysed.
Results: Both vascular endothelial growth factor and transforming growth factor-β2 caused a time dependent gel contraction, associated with over expression of α-smooth muscle actin in retinal pigment epithelium cells undergoing a fibroblast like transformation. The decrease in volume of the collagen gel and increase in α-smooth muscle actin expression were more significant in the transforming growth factor-β2-treated group than in vascular endothelial growth factor-treated group beginning at day 2, and the growth of retinal pigment epithelium cells was significantly more inhibited in the transforming growth factor-β2-treated group compared with the vascular endothelial growth factor-treated group after day 1 (P < 0.05). Transforming growth factor-β2 stimulation increased both vascular endothelial growth factor mRNA expression and secretion. The α-smooth muscle actin expression and the change in volume of collagen gel were significantly positively correlated in both experimental groups.
Conclusions: Both vascular endothelial growth factor and transforming growth factor-β2 can cause induction of retinal pigment epithelium cell-mediated collagen gel contraction in vitro via partial upregulation of α-smooth muscle actin expression.
The formation of proliferative cellular membranes is a typical medical sign in the advanced stages of proliferative vitreoretinopathy (PVR), requiring additional surgical removal.1 However, in the early stages of PVR, before the formation of the proliferative membrane, cell-mediated vitreous gel contraction occurs and induces pathological processes such as tractional retinal detachment, traction macular oedema or retinal tears.2–4 Retinal pigment epithelial (RPE) cells are thought to play a central role in the pathophysiology of PVR.4 The interaction between RPE cells and extracellular matrix involves a number of cellular activities including migration and reorganization of the extracellular matrix.5 The processes of RPE-mediated vitreous gel contraction is influenced by the expression and function of several cytokines that are found to be elevated in the vitreous of patients with PVR, including basic fibroblast growth factor,6 platelet derived growth factor7 and transforming growth factor (TGF)-β2.8 Among these, the TGF-β2 is thought to play a major role in the cell-mediated matrix contraction process and in the proliferative retinal response during PVR formation.9,10 Another growth factors that also is elevated in vitreous samples of human PVR, and especially at early stages, is vascular endothelial growth factor (VEGF).11,12 VEGF is known to be one of the main factors regulating ocular angiogenesis.13 However, until now this growth factor has achieved significantly less attention than TGF-β2 in the process of PVR formation, and its role in cell-mediated vitreous gel contraction remains elusive. In addition to increased vitreous levels, high levels of both VEGF and TGF-β2 have been found in epiretinal membranes of patients with PVR, further suggesting a possible association with both factors in PVR formation.14,15 It has also been demonstrated that both VEGF and pigment epithelium-derived factor are imbalanced in eyes with PVR, and that VEGF levels were substantially higher in early PVR stages.16,17 Furthermore, it has been shown that the anti-VEGF-A monoclonal antibody bevacizumab can inhibit collagen gel contraction ability and fibrosis activities in fibroblast culture, underpinning the proposed antifibrotic effect of VEGF in vivo.18 As anti-VEGF therapy now is widely used in the clinic, a correlation between VEGF and RPE-mediated gel contraction could give an easy accessible therapeutic target for early PVR. In order to further explore the role of VEGF compared with TGF-β2 in cell-mediated vitreous gel contraction during early PVR formation, we wanted to compare the effect of the two growth factors on RPE-mediated collagen gel contraction in vitro.
Human RPE cell isolation and culture
The RPE cell cultures were prepared using post mortem human eyes from eight Chinese donors (five men, three women) ranging in age from 26 to 65 years (49.5 ± 10.6 years), and with no known ocular pathology. The study was approved by the ethics committee of the University of Zhejiang, China. The method was described by Hu et al.,19 with slight modification. In brief, the isolated human RPE cells from the whole retina, including the macular and peripheral area, were cultured in the Ham's F12 nutrient medium supplemented with 10% Fetal Bovine Serum (FBS), 4 mmol/L glutamine, 100 IU/mL penicillin, 100 mg/mL streptomycin, and 250 ng/mL amphotericin (ICN Biomedicals, Inc., Aurora, OH, USA). The RPE cultures were maintained at 37°C in the presence of 5% CO2 and air, and reached confluence within 3–5 days. Confirmation of human RPE purity was routinely performed by immunohistochemical labelling of the cells with a wide-spectrum anti-cytokeratin monoclonal antibody (clone K8·13, ICN Biomedicals Ltd, High Wycombe, UK) that has been shown to stain human RPE cells.20 Passages 4–6 of cultured human RPE cells were used in the present study. The cells were plated at a density of 2.0 × 103 cells/cm2 on culture dishes and grown in growth media (1:1 (v/v) mixture of Dulbecco's Modified Eagle's Medium (DMEM) and F12), supplemented with 10% FBS. One day in front of all experiments, the media was replaced with a non-serum media. RPE cells from the same donor were used for one set of experiments.
Preparation of collagen gel contraction assay
The contraction assay was performed according to the method described by Noda et al.21 with slight modification. In brief, type I collagen (Vitrogen 100; Cohesion Technology, Palo Alto, CA), a reconstitution buffer, RPE cell suspension, and distilled water were mixed on ice at a ratio of 7:1:1:1:1 (final concentration of type I collagen gel, 1.5 mg/mL; final cell density, 1 × 105 cells/mL). The resultant mixture (1 mL) was added to a 24-multiwell plate (Sigma, St. Louis, MO, USA), and the formation of collagen gel was induced by incubation at 37°C in 5% CO2 for 60 min. Polymerized gels incubated at 37°C with 5% CO2 were treated with buffer (control group), 10 ng/mL VEGF or 3 ng/mL TGF-β2 (experimental groups), and the volume of the collagen gel was measured at indicated time points (day 1, 2, 3, 4, 5) after stimulation. For quantitative purposes, contraction data is presented as the reduction in collagen gel volume. The volume of collagen gel inside the culture well was calculated using the following formula based on the volume of fluid required to fill the well: Volume of collagen gel (mL) = (Volume of each well in the 24-well plate [3.5 mL]) − (Volume of fluid required to fill the well [mL]). The more fluid that can be added to the well, the lower the volume of collagen gel that can be measured (Change in volume of collagen gel = [Initial volume of collagen gel] − [Final volume of collagen gel]). Thirty wells were established for each treatment or control group for each experimental time (days 1–5), and data from three repeated measurements entered final analysis. Each experimental condition was assayed in triplicate, and at least three independent experiments were performed.
RPE cells in collagen gels
Changes in cell morphology were observed using an inverted microscope (Olympus IX70, Olympus Corporation, Tokyo, Japan). The viable cell number in each collagen gel excluded the effect of cell growth or cytotoxicity on the collagen gel contraction or inhibition of contraction. At each time point (day 1–5), the medium was replaced with 0.5 mL serum-free DMEM containing 3 mg/mL collagenase (collagenase type II, Worthington Biochemical Corporation, Lakewood, NJ, USA). The RPE cells were centrifuged and the supernatant was decanted. The viable cell number was determined using a haemocytometer with trypan blue staining. The number of cells per experiment is expressed as the percentage of the number of cells in control cultures.
α-Smooth muscle actin (SMA) immunofluorescence staining
Cells were incubated in phosphate-buffered saline (PBS) and 2% bovine serum albumin for 1 h in a moist chamber at room temperature. A 1:150 dilution of mouse anti-α-SMA monoclonal antibody (R&D systems, Minneapolis, MN, USA) in PBS was added and incubated for 1 h at room temperature. The primary antibody was removed and the cells were incubated for 60 min at room temperature with fluorescein isothiocyanate-conjugated anti-mouse IgG (Molecular Probes, Eugene, OR, USA) diluted 1:50 with 1% normal goat serum-PBS. Then the cells were washed three times with PBS for 5 min. The cover slips were mounted on the slides using 50% glycerin and micrographs were taken using a confocal microscope (LSM 410, Zeiss, Oberkochen, Germany).
α-SMA expression by Western blot
To lyse cells, the culture medium was removed from the cells, and the cells were washed twice with ice cold PBS and then resuspended in radio immunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl [pH 7.4], 1% NP-40, 0.25% Na-deoxycholate, 150 mM NaCl, and 1 mM ethylene-diamine-tetraacetic acid (EDTA)) containing a protease inhibitor cocktail (Sigma-Aldrich Chemical, St. Louis, MO, USA). The cell suspension was transferred to a centrifuge tube, left on ice, and vortexed every 5 min for 15 min to lyse the cells. The lysate was centrifuged at 14 000 g in a chilled centrifuge for 15 min. The total protein in the cell lysates and the concentrated medium was measured using a standard bicinchoninic acid (BCA) assay (Pierce, Rockford, IL, USA). Protein samples (10 µg) were dissolved in sample buffer (100 mM Tris [pH 6.8], 4% sodium dodecyl sulfate (SDS), 20% glycerol, 0.02% bromophenol blue, and 50 µL/mL β-mercaptoethanol) at a concentration of 1:1. Samples were resolved using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electroblotted onto a polyvinylidene difluoride membrane. The blot was placed in blocking solution (5% wt/vol non-fat dried milk in TBST) for 60 to 120 min at room temperature on a shaking table. The blots were incubated with the primary antibody, a mouse monoclonal anti-α-smooth muscle actin (Sigma-Aldrich Chemical, St. Louis, MO, USA) 1:2000, overnight at 4°C. Bound antibody was visualized using a secondary peroxidase-conjugate anti-mouse IgG antibody, enhanced chemiluminescence (Amersham Biosciences, Buckinghamshire, UK), and a Typhoon Scanner (Amersham Biosciences). Densitometry was used to quantify the expression of α-SMA using a National Institutes of Health (NIH) image. The signal density was measured and analysed using analysis software, and all graphs are shown as the α-SMA/(GAPDH) Glyceraldehyde 3-Phosphate Dehydrogenase ratio versus each control.
VEGF and TGF-β2 secretion assay
At each time point, all media surrounding the gels were collected from wells and stored frozen at −70°C until assayed. In order to explore whether the TGF-β2 and VEGF affect synthesis and secretion of each other, the level of VEGF in TGF-β2-treated groups and the level of TGF-β2 in VEGF-treated groups in the culture medium (200 µL) was determined by ELISA. The assay was performed according to the manufacturer's instructions (Human VEGF and TGF-β2 ELISA Kit, R&D Systems, Minneapolis, MN, USA). The lower limit of ELISA was 3.0 pg/mL. These experiments were repeated three times.
(RT-PCR) Reverse transcription polymerase chain reaction analysis of VEGF and TGF-β2 mRNA
Human RPE cells grown to confluence in 60 or 100 mm culture dishes were incubated in serum-free medium or serum-free medium containing the experimental growth factors. After 12 or 24 h, total cellular RNA was prepared by using RNA Stat-60 (Tel-Test, Friendswood, TX, USA), and purity and integrity of RNA assessed by ultraviolet spectrum analysis and agarose gel electrophoresis. PCR primers 5'-CCA TGA ACT TTC TGC TGT CTT-3' and 5'-TCG ATC GTT CTG TAT CAG TCT-3', that amplify all four alternatively spliced mRNA species of VEGF, were used.22 RT-PCR was performed by using RNA PCR kit and Gene Amp PCR system 9600 (Perkin Elmer, Foster City, CA, USA) according to manufacturers instructions as described earlier.23 PCR products were separated on an ethidium bromide containing agarose gel (FMC, Portland, ME, USA) photographed and integrated using Eagle Eye system (Stratagene, San Diego, CA, USA).
Data are expressed as mean ± standard deviation (SD). The values between experiment and control groups were compared using a paired t-test. The multiple comparisons among the volume of collagen gel as well as the expression of α-SMA on each experimental time were made by Student–Newman–Keuls test. Correlation between the expression of α-SMA and the change in the volume of collagen gel was determined using the Spearman rank correlation. Results were considered significant if P < 0.05.
Identification of RPE cells
The RPE cells in the primary culture were polygonal cells with various degrees of pigmentation, and partly became flattened and gradually lost their pigment granules during culture (Fig. 2a). We confirmed that more than 90% of primary culture RPE cells were strongly cytokeratin positive.
RPE cell-mediated collagen contraction
Mean volumes of collagen gel (±SD) in both cultures are shown in Figure 1. The volume of collagen gel in both experimental groups decreased significantly more than the control group beginning at day 2 (P < 0.01, respectively). The multiple comparisons of the volume of collagen gel at each time point indicated that the volume of collagen gel was significantly decreased on day 3–5 compared with day 1–2 in both VEGF- and TGF-β2-treated groups (P < 0.01, respectively). There were also a statistical difference in volume between day 1 and day 2 in both experimental groups (P < 0.05 for VEGF and P < 0.01 for TGF-β2). No statistically significant differences were found among the day 3–5 in both VEGF- and TGF-β2-treated groups. This demonstrates that both VEGF (10 ng/mL) and TGF-β2 (3 ng/mL) cause a time-dependent gel contraction. Comparison of the VEGF- and TGF-β2-treated group demonstrated that the decrease in collagen gel volume was more significant in the TGF-β2-treated group than in the VEGF-treated group beginning at day 2 (P < 0.05) (Fig. 1).
Cell counting and morphology studies
The RPE cells in collagen gel de-differentiated to a more mesenchymal cell-like phenotype, which was associated with cell spreading and the development of cytoplasmic extensions with some cells acquiring a more fusiform, fibroblastic morphology (Fig. 2). Compared with the VEGF-treated group, the growth of RPE cells was significantly inhibited in the TGF-β2-treated group beginning at day 1 (P < 0.05) (Fig. 2). There was no significant correlation between the number of RPE cells and the change in collagen gel volume at all time points in either the VEGF-treated group (r = 0.437, P > 0.05) and TGF-β2-treated group (r = 0.189, P > 0.05).
α-SMA immunofluorescence staining
After RPE cell fixation, the α-SMA expression was observed at each time point using immunofluorescence staining. Treatment with VEGF or TGF-β2 resulted in transformation of RPE cells in to a more fibroblast-like phenotype, including a high expression of α-SMA (Fig. 3a–c).
α-SMA expression in Western blot
Mean expression of α-SMA (±SD) in both cultures are shown in Figure 3d, and blots are shown in Figure 4. Both VEGF and TGF-β2 promoted a time-dependent overexpression of α-SMA in the RPE cells. The expression of α-SMA in both groups increased significantly more than in the control groups beginning at day 1 (P < 0.01). Multiple comparisons of the expression of α-SMA for each day indicated that, in the VEGF- or TGF-β2-treated groups, α-SMA expression increased in a time-dependent manner; the changes were significant on day 2 and day 3 in the VEGF-treated group and on day 2–4 in the TGF-β2-treated group (P < 0.01, respectively). Comparison of the VEGF- and TGF-β2-treated group demonstrated that the increase in α-SMA expression was more significant in the TGF-β2-treated group than in the VEGF-treated group after day 2 (P < 0.05). The expression of α-SMA significantly correlated with the decrease in collagen gel volume at all time points for both the VEGF- (r = 0.799, P < 0.01) and TGF-β2-treated group (r = 0.836, P < 0.01) (Fig. 3d).
VEGF and TGF-β2 expression by RPE cells
In order to explore whether the stimulation by VEGF and TGF-β2 can cause synthesis and secretion of the other growth factor, we detected the VEGF or TGF-β2 secretion by RPE cells during cell-mediated collagen gel contraction. Mean secretion and mRNA expression of VEGF and TGF-β2 (±SD) in both cultures are shown in Figure 5. In the ELISA assay, the VEGF secretion in TGF-β2-treated group increased significantly more than the control groups beginning at day 2 (P < 0.01, respectively), with significant increase in VEGF mRNA expression by RT-PCR examination beginning at day 1 (P < 0.01, respectively). There were no statistical significant changes in the TGF-β2 secretion – and mRNA expression in the VEGF-treated groups (P > 0.05) (Fig. 5).
By constructing a three-dimensional model of the mixture of RPE cells and collagen gel, we were able to mimic the microenvironment of early PVR in which no proliferative membrane has been formed. VEGF was evaluated as a stimulating factor in this cell-mediated collagen contraction model, and compared with TGF-β2 that is known to have a major role in the cell-mediated matrix contraction process. We found that VEGF, like TGF-β2, induces RPE cell-mediated collagen gel contraction, which is thought to be related to the morphological changes and biological activity of the RPE cells. RPE cells lost their immature epithelial morphology and acquired a more mesenchymal cell-like phenotype. In addition, the expression of α-SMA in RPE cells was significantly enhanced, and this was considered to be a reflection of mesenchymal transdifferentiation.24 Gamulescu et al.25 reported that RPE cells possessed the potential to transdifferentiate into myofibroblasts after stimulation with certain cytokines, and this process were implicated in the pathogenesis of PVR is characterized by a high expression of α-SMA in RPE cells and fibration of the extracellular matrix. If RPE cells migrates to the vitreous and mixed with or adhered to the extracellular matrix, contraction in RPE cells could induce fibration and traction of the collagen gel or surrounding tissue. Although it has been suggested that the presence of α-SMA does not necessarily reflect the contractile potential of a cell,8 our results indicates that the increase in α-SMA expression may be a shared indicator of pathogenesis for both VEGF and TGF-β2-dependent collagen gel contraction by RPE cells. The significant correlation between collagen gel contraction and α-SMA expression in RPE cells indicate that both VEGF and TGF-β2 promoted the time-dependent gel contraction associated with overexpression of α-SMA in RPE cells that have undergone a fibroblast-like transformation.
In the current study, TGF-β2 induced a significantly higher increase in collagen gel contraction and α-SMA expression beginning at day 1 compared with VEGF. This might be due to differences in biological activity and intracellular signalling pathway for these two cytokines. TGF-β2 is one of the multi-functional cytokines known to mediate RPE production of collagen and fibronectin, and is also known to promote cell migration and RPE transition into a fibroblastic phenotype.26 In this study, the VEGF expression/secretion in RPE cells increased significantly during TGF-β2 stimulation. These findings support that TGF-β2 not only plays a major role in every step of the cell-mediated matrix contraction process in PVR, but can also directly or indirectly induce the synthesis and secretion of other cytokines such as VEGF, which would exert a synergistic effect on its biological functions.27 This might also explain why TGF-β2 demonstrates a stronger effect than VEGF on the regulation of RPE cell-mediated collagen gel contraction.
Although collagen gel contraction increased after VEGF and TGF-β2 treatment, the growth of RPE cells was significantly more inhibited in the TGF-β2-treated group than in VEGF-treated group. Because the pathogenesis of PVR is based on the proliferation of RPE cells and the formation of extracellular matrix, it could be considered a controversy that TGF-β2 can inhibit RPE cell proliferation. However, it has been demonstrated that epithelial cell proliferation can either be inhibited, or unaffected by TGF.28 In the present study, we demonstrated that the proliferation of RPE cells increased in the VEGF-treated group compared with control. Several previous studies have also suggested that the VEGF-induced signalling network was considered to be a major factor that stimulates the proliferation and migration of RPE cells.29 In addition, we verified that the contraction of the collagen gel was significantly correlated with α-SMA expression in RPE cells, and not with growth activity, in both VEGF and TGF-β2-treated groups. This indicates that the stimulatory effect of VEGF or TGF-β2 on the gel contraction was not due to a growth-promoting effect on RPE cells. This indicates that the pathogenesis of early PVR might be more attributed to the mesenchymal transformation of RPE cells rather than RPE cell proliferation.
The selection of the concentrations of VEGF and TGF-β2 used in this study was based on the reported concentration that can influence the activity of RPE cells. Miura et al.8 investigated the signalling mechanisms of TGF-β2-dependent collagen gel contraction by RPE cells using 3 ng/mL TGF-β2. Hoffmann et al.5 also examined the ability of VEGF, and other cytokines, to stimulate RPE cell migration. They found that VEGF induced a migratory response in the RPE cells at a concentration of 10 ng/mL. Thus the concentrations of VEGF and TGF-β2 used in the present study should be effective in activation of the RPE cells. However, further investigations are needed in order to determine the dose–effect relationship between the VEGF and the level of RPE cell-mediated gel contraction.
Many studies have focused on exploring the ideal adjuvant pharmacological treatment of cell-mediated gel contraction during PVR formation. However, at present there is no such drug available for clinical use. The results in this study showed that VEGF had a similar effect as TGF-β2 in inducing RPE cell-mediated collagen gel contraction in vitro, which may highlight a new potential therapeutic target for the control of the early stages of proliferative vitreoretinal diseases. Anti-VEGF drugs are widely used in the clinic for the treatment of many vitreoretinopathies.30 By utilizing VEGF inhibitors in patients risk factors for development PVR one might be able to pharmacologically block one of the molecular targets for RPE-mediated gel contraction at an early stage. It has been demonstrated that bevacizumab possesses potential antifibrotic activity, which underpins the antifibrosis effect proposed for PVR treatment.18 Therefore, further examinations of anti-VEGF therapy in both in vitro and in vivo models of early PVR are needed.
In conclusion, we show that both VEGF and TGF-β2 can cause induction of human RPE cell-mediated collagen gel contraction in vitro via partial upregulation of α-SMA expression.
This work was supported by National Natural Science Foundation of China (J20080843), Zhejiang Provincial Natural Science Foundation of China (J20050897) and Medical Scientific Research Foundation of Zhejiang Province, China (2004A047). Also supported by the Research Council of Norway, the Faculty of Medicine University of Oslo and Oslo University Hospital.
- 23Transforming growth factor-beta expression in human retinal pigment epithelial cells is enhanced by Toxoplasma gondii: a possible role in the immunopathogenesis of retinochoroiditis. Clin Exp Immunol 2002; 128: 372–8., , .