Molecular Vision 2003; 9:473-481 <http://www.molvis.org/molvis/v9/a60/> Received 29 July 2003 | Accepted 25 September 2003 | Published 3 October 2003

Download Reprint

Effect of quiescence on integrin α5β1 expression in human retinal pigment epithelium

Stephanie Proulx,^1,2 Sylvain L. Guérin,³ Christian Salesse^1,2
 
 

¹Université du Québec à Trois-Rivières, Département Chimie-Biologie, Trois-Rivières, Québec, Canada; ²Unité de Recherche en Ophtalmologie, Centre de Recherche du Centre Hospitalier de l'Université Laval, Ste-Foy, Québec, Canada; ³Oncology and Molecular Endocrinology Research Center, Centre Hospitalier de l'Université Laval, Ste-Foy, Québec, Canada

Correspondence to: Christian Salesse, Unité de Recherche en Ophtalmologie, Centre de Recherche du CHUL, Pavillon CHUL, 2705 Boul. Laurier, Ste-Foy, Québec, Canada, G1V 4G2; Phone: (418) 656-4141, ext. 7243; FAX: (418) 654-2131; email: christian.salesse@crchul.ulaval.ca

Abstract

Purpose: The retinal pigment epithelium (RPE) is differentiated and mitotically inactive in the normal eye, but several pathologies such as proliferative vitreoretinopathy (PVR) cause RPE cells to dedifferentiate and resume proliferation. Integrins, a family of cell surface glycoproteins that mediate cell proliferation and differentiation, are thought to play fundamental roles in PVR. The aim of this study was to evaluate protein expression and gene regulation of the integrin α5 subunit in proliferating and quiescent RPE cells.

Methods: Protein expression was studied in situ by immunohistochemistry and in vitro at different cell confluences by immunoprecipitation. Semi-quantitative RT-PCR and transient transfections were used to determine whether increasing cell confluence also affected α5 subunit mRNA levels and promoter activity, respectively.

Results: We demonstrated that the integrin α5 subunit is present at the RPE cell surface both in situ and in vitro, and that α5 protein level is influenced by confluence. Levels of integrin α5 transcripts are similar for sub-confluent and confluent cells, and a small increase in the promoter activity was observed between sub-confluent and confluent cells. However, both the integrin α5 subunit transcript and the α5 promoter activity decreased when cells reached post-confluence.

Conclusions: We demonstrated that cell confluence affected protein and gene expression of the integrin α5 subunit. Proliferating RPE cells expressed high levels of both the α5 protein and mRNA transcripts and showed a high promoter activity. However, when cells reached quiescence, α5 gene expression was substantially reduced and RPE cells expressed little α5 protein at their cell surface.

Introduction

The retinal pigment epithelium (RPE) is a monolayer of polarized epithelial cells located between the photoreceptor outer segments of the neural retina and the choroidal blood supply [1]. RPE cells have many metabolic and biochemical functions essential for maintaining retinal integrity, such as phagocytosis of rod and cone outer segments, transport of metabolites to and from the visual cells and choroid, maintenance of the blood-ocular barrier and adhesion of the retina [1]. At their basal surface, the RPE basement membrane is separated from the basement membrane of the choroidal capillaries by the Bruch's membrane. The Bruch's membrane is composed of a number of extracellular matrix (ECM) macromolecules such as type I, III and IV collagen, laminin and fibronectin (FN), which are also synthesized by RPE cells in vitro [2]. Studies of cultured RPE cells have shown that they attach to and spread on individual matrix molecules such as collagen [3], laminin [4,5] and fibronectin [5], as well as on ECM produced by various cell lines [3-6].

The interactions between a cell and its ECM are, in part, mediated by a family of cell surface glycoproteins called integrins [7,8]. Signals transduced by integrins can influence cell adhesion, migration, differentiation and proliferation [7,8]. Integrins are heterodimers consisting of variable αand βsubunits. The different combinations of these subunits determine the receptor specificity for the different ECM molecules [7,8]. Although integrins often bind to more than one ligand, some show selectivity. It is the case for the α5β1 integrin which binds specifically to FN [7-9]. The α5β1 integrin has been shown to mediate adhesion to FN for a number of cell types, including leukemic cells [10], mammary carcinoma cells [11], and RPE cells [12]. Cell spreading and proliferation are inhibited in α5-deficient NRK fibroblasts [13], whereas mammary carcinoma cell proliferation is inhibited by antibodies to α5β1 [11]. Furthermore, α5β1 has been shown to mediate migration of aortic endothelial cells [14], fibroblasts [15] and RPE cells [16] on FN, a further evidence that α5β1 also promotes cell proliferation and migration.

The FN integrin α5β1 also appears to play pathogenic roles in a variety of proliferative diseases. Indeed, changes in α5β1 expression have been correlated with malignancy of uveal melanoma [17]. Robbins et al. [18] showed the presence of the α5 subunit (amongst others) on pigmented cells (presumably of RPE origin) from proliferative vitreoretinopathy (PVR) membranes and concluded that its presence was abnormal when compared to the lack of α5 staining on normal retina [19].

Integrins have been postulated to play a major role in PVR [7,18], a pathology in which normally quiescent and differentiated RPE cells detach from the underlying Bruch's membrane, migrate and proliferate [20-22]. When RPE cells are cultured, they dedifferentiate, migrate and proliferate very much the same as those in PVR. Studying integrin expression of cultured RPE cells, and comparing integrin expression on proliferating and quiescent cells, could thus help understand the involvement of integrins in PVR.

Given the importance of the α5β1 integrin in cell adhesion, proliferation and migration, we sought to investigate the α5β1 protein expression and gene regulation in proliferating and quiescent human RPE cells. We determined if the integrin α5 subunit was present at the RPE cell surface in situ and in vitro and examined the α5 mRNA transcribed in human RPE cells in vitro at different cell confluences. We also assessed whether the changes in the amounts of α5 mRNA in vitro found with increasing cell confluence were reflected by differences in the transcriptional activity directed by the α5 gene promoter.

Our results demonstrate that the integrin α5 subunit is present at the RPE cell surface both in vitro and in situ, and that its expression is modulated with cell quiescence. Proliferating (sub-confluent and confluent) RPE cells have similar amounts of α5 protein and α5 mRNA transcripts, and both protein and mRNA expression are downregulated when RPE cells become quiescent. A small increase in α5 promoter activity is however observed between sub-confluent and confluent RPE cells, although promoter activity also decreases when cells become quiescent, providing further evidences that cellular quiescence dictates the level to which the α5 integrin is expressed in RPE cells.

Methods

Cell cultures

Cultures of primary RPE cells were established from eyes of 14 human donors (age range: 6 days old to 48 years old, mean=22 years old) obtained from the National Eyebank Inc. (Québec, Qc, Canada) within 24 h of death. This research was conducted in accordance with the tenets of the Declaration of Helsinki. Dissection was done by a modification of a previously published method [23]. We did not use eyes from donors older than 50 years old because we usually harvested more cells from younger adult eyes that also attached and spread faster than cells from older adult eyes. Cells were cultured with Keratinocyte-SFM medium (Canadian Life Technologies, Burlington, Ontario, Canada) supplemented with 5% bovine calf serum (Hyclone, Logan, Utah, USA) and 200 mg/ml albumax, 45 mg/ml ascorbic acid, 1 mg/ml carnitine, 500 mg/ml glucose, 112 mg/ml fructose, 5 mg/ml glutathione, 6 mg/ml hypoxanthine, 67 mg/ml oxalacetic acid, 0.15 mg/ml retinol acetate, 5 mg/ml taurine, 0.025 mg/ml, D-α-tocopherol, 50 mg/ml transferrin and 0.3 mg/ml uridine. Medium was changed every 2-3 days and cultures were maintained in an incubator at 37 °C in 5% CO₂/95% air atmosphere. For all experiments, cultures were used at passage 2 and cells were plated at different densities as follows: non-confluent cells: 2.5x10⁴ cells/cm², 2 days in culture; confluent cells: 1.5x10⁵ cells/cm², 2 days in culture; post-confluent cells: 1.5x10⁵ cells/cm², 2 weeks in culture.

Bromo-deoxyuridine labeling

DNA synthesis of RPE cells at different confluence was monitored by the addition of the thymidine analogue bromo-deoxyuridine (BrdU; Sigma, Oakville, Ontario, Canada) to the culture medium (final concentration 10 μM). Labeling was allowed for 24 h upon which time cells were washed and fixed with an ethanol fixative (70% ethanol, 50 mM glycine, pH 2.0) for 30 min at -20 °C. Incorporated BrdU was detected by immunofluorescence. Briefly, RPE cells were first incubated with a mouse anti-BrdU antibody (clone BRD.2; Medicorp, Montreal, Quebec, Canada) diluted 1:50 in PBS containing 3000 U/ml of Exonuclease III (Sigma) and 150 U/ml of restriction endonuclease SAU 3A I (Sigma) for 30 min at 37 °C, followed by incubation with a secondary antibody (anti-mouse IgG-TRITC conjugated; Sigma) for 30 min at 37 °C. After 3 other washes in PBS and 2 final washes in distilled water, the coverslips were mounted on a microscope slide and fixed with nailpolish. The fluorescence was observed with an epifluorescence microscope (Nikon Diaphot 300). For negative controls, the primary antibody was replaced by PBS-BSA1%. The results shown are representative of a minimum of 3 independent experiments conducted on separate RPE populations derived from different donors.

Immunohistochemistry

Immunohistochemistry was performed on 2 different human donor eyes (a 21 year old donor and a 81 year old donor) which were obtained, respectively, 15.5 h and 4 h post-mortem. They were immediately fixed with 4% paraformaldehyde in PBS for 24 h at 4 °C. Following incubation in 20% sucrose for 2 h at room temperature and 30% sucrose for 24 h at 4 °C, eyes were embedded in OCT and frozen in liquid nitrogen. Sections of 10 μm were cut using a Leitz 1720 digital cryostat, collected on superfrost/plus Fisherbrand slides (Fisher scientific, Nepea, Ontario, Canada) and stored at -80 °C until use. Sections were incubated for 20 min at room temperature in 95% ethanol, and then treated with 0.25% potassium permanganate for 15 min at room temperature [24] to reduce autofluorescence arising from lipofuscin present in RPE [25]. Blocking buffer and antibodies were diluted in 0.3% BSA in PBS. Sections were blocked overnight at 4 °C in 5% goat serum (Vector Laboratories, Burlington, Ontario, Canada), then incubated for 90 min at 37 °C with rabbit anti-integrin α5 polyclonal antibody (Chemicon, Temecula, CA), or with normal rabbit serum (Sigma) for negative control, both at a concentration of 103 μg/ml. Sections were blocked again in 5% goat serum for 60 min at 37 °C, and incubated with secondary antibody labeled with Alexa Fluor 488 at a 1:100 dilution (Molecular Probes, Eugene, OR) for 90 min at 37 °C. Sections were mounted with Prolong antifade mounting medium (Molecular Probes). Images were obtained using a MRC 1024 confocal system (Bio-Rad) mounted on a Nikon Diaphot TMD inverted microscope equipped with 60x 1.4 N.A. immersion optics. Confocal 8 bit images were captured using Bio-Rad's Lasersharp acquisition software running under OS2.

In order to distinguish residual autofluorescence from specific immunoreactivity, images were obtained as described in Kennedy et al. [26]. Briefly, sections were subjected to sequential excitation with the 488 nm line (green) and the 568 nm line (red) of the krypton/argon laser. Autofluorescent materials have broad spectral properties and are visible at both 488 nm and 568 nm excitation. By contrast, the Alexa Fluor green fluorescence is visible only with the 488 nm excitation. By this method, autofluorescent components appear orange/yellow and specific immunoreactivity appears green.

Immunoprecipitation

RPE cells were harvested, cell density was adjusted to 5x10⁶ cells/ml and labeled with 1 mg/ml soluble Sulfo-NHS-LC-Biotin (Pierce, Rockford, IL) in PBS for 60 min at room temperature. Cells (5x10⁶ cells) were then lysed in lysis buffer (50 mM Tris-HCl, pH 7.4 containing 1% Igepal CA-630, 120 mM NaCl, 1 mM CaCl₂.2H₂O, 1 mM MgCl₂, 0.7 μg/ml pepstatin, 0.5 μg/ml leupeptin, 1 mM pefabloc). Insoluble materials were removed by centrifugation at 16 000 g for 10 min at 4 °C. Protein concentration was measured with the BCA method (Pierce) to ensure that equal amounts of protein were loaded in each lane. Each ml of supernatant was pre-cleared by a 1 h incubation with 50 μl of unconjugated protein-A-Sepharose CL-4B beads (Amersham Pharmacia Biotech, Baie d'Urfé, Québec, Canada). Approximately 50 μl of goat anti-mouse IgG coated sepharose beads (Zymed Laboratories, San Francisco, CA) was incubated for 3 h at room temperature with 10 μg primary antibody (mouse anti-human integrin α5, clone IIA1, Pharmingen), then rinsed with wash buffer (50 mM Tris-HCl pH 7.5, 0.5 M NaCl, 1 mM CaCl₂, 1 mM MgCl₂, 0.1% Tween 20) to remove unbound antibodies. Pre-cleared (500 μl) samples were precipitated for 3 h at room temperature with 50 μl of protein-sepharose/anti-mouse IgG with bound primary antibody. The Sepharose beads with bound immune complexes were washed 6 times in wash buffer. Then, 50 μl of 2X Laemmli buffer containing 5% 2-mercaptoethanol was added to the 50 μl of rinsed beads, boiled for 5 min at 85 °C, cooled, and 20 μl was loaded per lane (corresponding to 7.5x10⁵ immunoprecipitated cells) on a 6% polyacrylamide gel and transferred to nitrocellulose (Bio-Rad Laboratories, Mississauga, Ontario, Canada). The membrane was blocked for at least 1 h with 3% BSA, incubated for 1 h with streptavidin-HRP (Pierce) and processed for autoradiography. Band density was evaluated using the QuantityOne 1-D image analysis software (Bio-Rad Laboratories).

Semi-quantitative Reverse Transcription-Polymerase Chain Reaction

RNA from RPE cells at different confluences was isolated using the Tri-Reagent (Sigma) according to the manufacturer's instruction. Reverse transcription (RT) was performed using 5μg total RNA and 0.2μg of random hexamer primers (Fermentas; Burlington, Ontario, Canada) following the manufacturer's protocol for synthesis of first strand cDNA. Polymerase chain reaction (PCR) was performed using the QuantumRNA 18S Internal standards protocol for semi-quantitative RT-PCR (Ambion, Austin, Texas) according to the manufacturer's instructions. Briefly, the linear range of the α5 PCR product was determined using RNA from sub-confluent, confluent and post-confluent RPE. PCR products were in the linear range between 23 and 33 cycles for sub-confluent and confluent cells, and between 25 and 31 for post-confluent cells. The total number of cycles used for semi-quantitative RT-PCR was 28. The QuantumRNA 18S Internal standard (Ambion) used comes with 18S primers and 18S competimers that competes one another to decrease 18S PCR product in order to co-amplify and detect both 18S and gene specific PCR products. The optimal ratio of 18S primers: competimers was determined and the ratio used was 2:8 (18S primers:competimers). Bands were analyzed with the GelDoc2000 gel documentation system (Bio-Rad Laboratories) and the QuantityOne 1-D image analysis software (Bio-Rad Laboratories). Semi-quantitative RT-PCR was done twice in duplicate on 4 different donors. The DNA sequence of the primers used for the amplification of the human α5 transcript were: forward primer: 5'-GGC AGC TAT GGC GTC CCA CTG TGG-3', reverse primer: 5'-GGC ATC AGA GGT GGC TGG AGG CTT-3' (171-bp PCR product) and cycle parameters were 94 °C for 30 s, 60 °C for 30 s and 72 °C for 30 s. The 18S primers (Ambion) gave a PCR product of 489 bp.

Transient transfections and CAT assays

Plasmids containing the chloramphenicol acetyltransferase (CAT) reporter gene from the plasmid pSKCAT fused to various 5' deletions of the α5 gene and stripped of their multicloning site (MCS; -954α5CATΔMCS, -178 α5CATΔMCS, -132α5CATΔMCS, -92α5CATΔMCS, -41α5CATΔMCS) have been previously described [27]. RPE cells at different confluences were transiently transfected using the calcium phosphate precipitation method as described by Graham and Van der Eb (1973) [28]. Each individual plate received 15 μg of the test plasmid and 5 μg of human growth hormone (hGH)-encoding plasmid pXGH5 [29] for 4 to 6 h before being washed with PBS and incubated an additional 48 h in fresh medium. Levels of CAT activity for the transfected cells were determined as described [30] and normalized to the amount of hGH secreted into the culture media and assayed using a kit for quantitative measurement of hGH (Immunocorp, Montréal, Québec, Canada). The value presented for each individual test plasmid transfected corresponds to the mean of at least three separate transfections done with RPE cells from three different donors, each in triplicate. To be considered significant, each individual value needed to be at least three times over the background level. Standard deviation is also provided for each transfected CAT plasmid. Statistical analyses were performed using a single factor ANOVA (Excel, Microsoft, Bellvue, WA). A p value of less than 0.001 was considered statistically significant.

Results

The α5 integrin subunit is expressed on RPE cells in situ

Integrin α5 subunit expression in situ with normal RPE cells has been studied previously with contradictory results. Indeed, the lack of α5 subunit has been reported in situ on human [19] and chicken RPE cells [31]. On the other hand, Anderson et al. [32] demonstrated its presence on monkey RPE cells. To verify if the α5 integrin subunit is indeed expressed in vivo on RPE cells, a 21 year old human donor eye was analyzed by immunohistochemistry using a monoclonal antibody directed against the α5 subunit (Figure 1). As revealed on Figure 1A, expression of α5 was readily detected in the RPE monolayer, and staining was primarily concentrated along the basal surface of the cells. Staining for the integrin α5 subunit was also detected in the choroid. The labeling pattern for the 81 year old retina was the same, although much more lipofuscin autofluorescence could be observed (data not shown).

Proliferating and quiescent RPE cells in vitro

To assess whether cultured RPE cells grown at the cell densities selected for this study were in a quiescent or actively proliferating state, incorporation of BrdU was performed. RPE cell cultures that are left confluent for 2 days (confluent cells) incorporated BrdU into their nuclei indicating their high proliferative activity (Figure 2A). On the other hand, RPE cell cultures that remained confluent for 2 weeks (post-confluent cells) previous to BrdU labeling showed a markedly reduced BrdU staining indicating that most cells at this stage are out of the cell cycle and have reached cellular quiescence (Figure 2B). These results are consistent with the study of Kaida et al. [33] who also demonstrated that RPE cells were proliferating when cells were left for 2 to 3 days at confluence, and that proliferation of RPE cells was infrequent when cells were maintained at confluence for 8 weeks.

Decrease of the α5 protein at the RPE cell surface with cellular quiescence

The α5 integrin subunit has already been observed in vitro on RPE cells [12,16,32,34-39], and some of these studies reported that the staining pattern differed depending on the culture state reached by RPE cells [32,34,35,38]. Indeed, immunofluorescence studies showed that proliferating (undifferentiated) RPE cells had a heterogeneous α5 staining with fluorescence clusters at the cellular surface and fluorescence on the border of the cell membrane [32,38]. When RPE cells reached confluence or remained confluent for 1 to 2 weeks, the α5 staining yielded a nearly continuous linear band around the cells [32,34,35]. We also found, through immunofluorescence analysis, that confluence alters the α5 staining of RPE cells (data not shown) and wished to determine whether confluence also affected α5 protein levels at the RPE cell surface. Cell surface levels of the α5 subunit was thus examined in RPE cells by immunoprecipitation of biotinylated cell surface proteins (Figure 3). A protein with a molecular mass corresponding to that reported for the α5 integrin subunit (~140 kDa) [40,41] was observed upon immunoprecipitation with a monoclonal antibody raised against the α5 subunit. A smaller band at ~120 kDa that corresponds to the co-precipitated integrin β1 subunit [40,41] was also observed (Figure 3). Figure 3 also shows that non-confluent and confluent RPE cells express similar amounts of the biotinylated α5 subunit for the same number of cells at the cell surface. In contrast, post-confluent cells exibited a marked decrease in cell surface expression of α5 when compared to proliferating cells.

Transcription of the α5 gene varies with increasing cell confluence

To investigate whether the difference in the expression of the α5 subunit protein at the RPE cell surface is the consequence of a reduced transcription of the α5 subunit gene, total RNA from cultured RPE cells was isolated and semi-quantitative RT-PCR was performed to evaluate the amount of α5 mRNA transcript present at different stages of cell confluence. When normalized to the amount of transcripts encoded by the ribosomal 18S gene, both non-confluent and confluent RPE cells expressed similar levels of α5 transcripts whereas post-confluent cells yielded a 41% decrease in α5 mRNA (Figure 4).

Given that the levels of integrin α5 subunit mRNA was influenced by cell confluence, we next asked whether this was due to differences in the transcriptional activity normally driven by the α5 gene promoter. The identification of the regions that are important for the regulation of the α5 promoter was examined by comparing the activity directed by different α5 promoter-bearing recombinant constructs following their transient transfections into cultured RPE cells grown at different cell confluences. The recombinant constructs selected for these assays contained the CAT reporter gene fused to different lengths from the human α5 gene promoter that extend to 5' positions -954 bp (in plasmid -954α5CATΔMCS), -178 bp (in plasmid -178α5CATΔMCS), -132 bp (in plasmid -132α5CATΔMCS), -92 bp (in plasmid -92α5CATΔMCS) and -41 bp (in plasmid -41α5CATΔMCS) relative to the α5 mRNA start site. As indicated in Figure 5, no significant CAT activity was obtained for plasmid -41α5CATΔMCS under any of the cell culture conditions selected. Extending further the α5 promoter up to position -92 resulted in a dramatic increase in CAT activity (232-fold, 255-fold and 51-fold increase for sub-confluent, confluent and post-confluent cells, respectively) that reached maximal level of activity at confluence and then slightly decreased when RPE cells reached post-confluence. These results clearly suggested the presence of one (or a few) strong positive regulatory element(s) on the α5 promoter segment is comprised between positions -41 to -92. Extending the α5 promoter from -92 to -132 and -178 resulted in a marked decrease in promoter activity (3-fold, 3-fold and 7-fold decrease between -92 and -178 for sub-confluent, confluent and post-confluent cells, respectively), suggesting that negative regulatory elements functional in RPE cells are located between these positions. Further extension up to position -954 resulted in no statistically significant alteration in the α5 promoter activity relative to the level directed by the -178α5CATΔMCS plasmid (2.0-fold, 1.3-fold and 1.6-fold increase for sub-confluent, confluent and post-confluent cells, respectively).

Figure 5 also shows that all of the α5 promoter constructs used in this study directed maximal transcriptional activity at confluence, which then considerably decreased when RPE cells became quiescent (post-confluence), especially for the -132α5CATΔMCS, -178α5CATΔMCS and -954α5CATΔMCS plasmids. The -92α5CATΔMCS, -132 α5CATΔMCS, -178 α5CATΔMCS and -954α5CATΔMCS plasmids yielded 1.3-fold, 1.4-fold, 1.5-fold and 1.0-fold increases, in CAT activity, respectively, between sub-confluent and confluent cells, although this increase was not found to be statistically significant, except for the -132α5CATΔMCS plasmid (p=0.0005). All but one construct (the -92α5CATΔMCS) were found to have a statistically significant decrease between confluent and post-confluent cells. This reduction between confluent and post-confluent cells was 1.3-fold (p=0.02), 1.9-fold (p=0.0001), 3.3-fold (p=0.0003) and 2.6-fold (p=1.5x10^-8) for the -92α5CATΔMCS, -132 α5CATΔMCS, -178 α5CATΔMCS and -954α5CATΔMCS plasmids, respectively. We therefore conclude that α5 promoter sequences that determine basal promoter activity are located between α5 positions -41 to -92, and that cell quiescence does affect α5 promoter function in vitro.

Discussion

PVR is a disease in which normally quiescent RPE cells start to dedifferentiate, detach from their basement membrane and resume growth [20-22]. They migrate, proliferate and secrete ECM molecules that form fibrocellular membranes on the retina or in the vitreous [20-22]. As the disease progresses, the membranes contract and cause tractional retinal detachment [20-22]. Migration, proliferation and dedifferentiation of RPE cells should involve membrane proteins like integrins since they have been shown to play a predominant role in such processes [7,8]. Moreover since FN was clearly shown to be present in epiretinal membranes in PVR [42-44], it has been postulated that its membrane receptor, the integrin α5β1, could be involved in PVR [45]. We therefore used cultures of RPE cells as a model for dedifferentiated cells that are actively proliferating and studied the expression of the α5 integrin subunit at the protein, mRNA and gene levels.

In order to establish if the integrin α5 subunit is differently regulated on recently dedifferentiated RPE cells that are actively proliferating, we first needed to establish whether this subunit is normally expressed in vivo. We found that the integrin α5 subunit is indeed expressed on human RPE cells in situ. This contrasts with the data of Brem et al. [19] who found no α5 in human RPE cells. The presence of α5 on RPE cells in vitro is less unsettled, and its presence has been well documented [12,16,32,34-39]. Indeed, indirect evidence that RPE cells do express the α5 integrin subunit has been provided in a recent study by Zhao et al. [39]. They demonstrated that phagocytosis of FN-coated beads by human RPE cells was markedly inhibited by exposure of such cells to antibodies directed against the α5 integrin subunit. Direct evidence that human RPE cells do indeed express the α5 integrin subunit was also provided by the same group through both FACS and RT-PCR analyses [16]. In that study, α5 integrin subunit expression was shown to be substantially increased through the activation of the MAPK signaling pathway upon exposure of RPE cells to TNF-α[16]. Although these studies provided evidence that the α5 integrin subunit is expressed in RPE cells, none examined whether cell confluence has any influence on this integrin subunit expression and whether such alterations also translate into altered α5 promoter function as well.

It has been shown that RPE cells in culture demonstrate different patterns of expression for the same protein depending on how much time cells are left at confluence [35]. For example, adhesion molecules of the integrin and CAM families showed a similar diffuse distribution of varying intensity among individual cells early in confluence, then each accumulated in a distinct pattern at the cell periphery when cell reached full confluence. The time required for each molecule to achieve its characteristic distribution differed, ranging from 1-2 weeks post-confluence for the integrin α5β1, 5 weeks post-confluence for ICAM and PECAM, and 2 months post-confluence for N-cadherin and NCAM [35]. Our immunofluorescence data (not shown) are consistent with these results. However, immunofluorescence does not provide quantitative information on the amount of protein expressed at the cell surface. We thus evaluated if this differing pattern of expression was reflected with a change in the amount of protein at the cell surface. In our experiments with cultured RPE cells, immunoprecipitation studies showed that proliferating cells express much more α5 at the cell surface than quiescent cells. Thus, unlike proliferating cells, quiescent cells may need different integrins at their cell surface and the presence of α5β1 may not be required in large amount. This is in accordance with other studies [41,46-48] which demonstrated that terminal differentiation of keratinocytes in culture also involves loss of α5β1 at the cell surface. Also, Sastry et al. [49] demonstrated that α5-transfected myoblasts remained in the proliferative phase and that differentiation was inhibited even in confluent cultures, suggesting that loss of integrin α5 expression may be required in order for differentiation to take place.

Our results provided evidence that cell quiescence also affected expression of integrin α5 subunit at the gene level. Both mRNA levels and promoter strength decreased between confluent and post-confluent cells. This indicates that promoter activity is at its highest when proliferating cells reach confluence and then there is a general down regulation of the α5 promoter when cells become quiescent. These results are consistent with those from other studies [48,50]. Indeed, the reduced transcription of integrin α5 subunit in keratinocytes has been shown to correlate with the expression of differentiation markers [48]. In addition, when human fibrosarcoma HT-1080 cells are released from quiescence, an increase in promoter activity, mRNA level as well as protein levels has been observed for the integrin α5 subunit [50]. In a more recent study conducted by our group, expression directed by the human α5 integrin subunit gene promoter was found to be strongly affected by cell density in primary cultures of rabbit corneal epithelial cells (RCECs) [51]. Indeed, the α5 promoter activity was found to be 9 to 13 times lower in confluent RCECs when compared to the level measured in non-confluent cells, indicating that the α5 promoter is regulated differently in different cell types. It must be pointed out that RCECs reach quiescence after only a few days at post-confluence.

When we look more closely at the regulation of the α5 promoter in RPE cells, we found that the sequence located between positions -41 to -92 functions as a powerful transcriptional activator. Previous studies demonstrated that the -62 to -132 α5 promoter region indeed functions as a positive regulatory region in the fibrosarcoma HT-1080 cell line [50]. The -41 to -92 α5 promoter segment was also shown to positively modulate α5 promoter activity in the promonocytic U937 cell line [52]. The α5 -41/-92 promoter sequence bears target sequences for the transcription factors AP-1 (between positions -45 and -51 bp) [53] and Sp1 (between positions -61 and -77 bp) [54]. Sp1 has recently been shown to bind a target site within this α5 promoter segment designated the fibronectin responsive element (FRE) and located between positions -56 to -82, which confers responsiveness of the α5 promoter toward FN [54]. Whether AP-1 or Sp1 are involved in the -41/-92 mediated up regulation of the α5 promoter activity in RPE cells remains to be elucidated.

The present study shows that when RPE cells are dedifferentiated and proliferate, promoter activity, mRNA transcripts and α5 protein levels are highly increased compared to quiescent cells. This finding is important in view of the possible involvement of this integrin subunit in PVR. Indeed, given that FN was shown to be expressed in epiretinal membranes in PVR [42-44], that the α5β1 integrin binds specifically to FN [7-9] and that α5β1 promotes cell proliferation and migration [14,16], it can be postulated that the α5 integrin subunit plays a major role in PVR in which RPE cells detach from the underlying Bruch's membrane, migrate and proliferate [20-22]. This integrin subunit could thus represent a therapeutic target for PVR if the mechanism of its upregulation can be identified. In this regard, a recent study [36] demonstrated that human RPE cells grown in the presence of vitreous resulted in an elevation of expression of the α5 subunit. In PVR, RPE cells come in contact with vitreous through a tear in the neural retina [55]. Therefore, a detailed study on how the transcriptional activity directed by the α5 promoter is modulated following exposure of RPE cells to the vitreous would most certainly contribute to the understanding of why this integrin subunit is upregulated during PVR.

Acknowledgements

The authors would like to thank the National Eyebank inc. (Québec, Canada) for providing us with human globes. We wish to thank Dr. Bruce A. Pfeffer (from Bausch & Lomb, Rochester, New York) for his help in starting RPE cell cultures. We are also grateful for the technical assistance of Marcelle Giasson and Steeve Leclerc (both form the CHUL Research Center) in cell cultures, transient transfections and RT-PCR. Supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). SP held a scholarship from the NSERC and was supported by a joint scholarship from the Canadian Institutes of Health Research (CIHR) and the E.A. Baker Foundation from the Canadian National Institute for the Blind (CNIB). CS and SLG are Chercheur Boursier National from the Fonds de la Recherche en Santé du Québec (FRSQ).

References

1. Marmor MF. Structure, function and disease of the retinal pigment epithelium. In: Marmor MF, Wolfensberger TJ, editors. The retinal pigment epithelium: function and disease. New York: Oxford University Press; 1998. p. 3-9.

2. Campochiaro PA, Jerdon JA, Glaser BM. The extracellular matrix of human retinal pigment epithelial cells in vivo and its synthesis in vitro. Invest Ophthalmol Vis Sci 1986; 27:1615-21.

3. Dutt K, Scott MM, Del Monte M, Brennan M, Harris-Hooker S, Kaplan HJ, Verly G. Extracellular matrix mediated growth and differentiation in human pigment epithelial cell line 0041. Curr Eye Res 1991; 10:1089-100.

4. Campochiaro PA, Hackett SF. Corneal endothelial cell matrix promotes expression of differentiated features of retinal pigmented epithelial cells: implication of laminin and basic fibroblast growth factor as active components. Exp Eye Res 1993; 57:539-47.

5. Miceli MV, Newsome DA. Effects of extracellular matrix and Bruch's membrane on retinal outer segment phagocytosis by cultured human retinal pigment epithelium. Curr Eye Res 1996; 15:17-26.

6. Ho TC, Del Priore LV, Kaplan HJ. En bloc transfer of extracellular matrix in vitro. Curr Eye Res 1996; 15:991-7.

7. Elner SG, Elner VM. The integrin superfamily and the eye. Invest Ophthalmol Vis Sci 1996; 37:696-701.

8. Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 1992; 69:11-25.

9. Aota S, Yamada KM. Fibronectin and cell adhesion: specificity of integrin-ligand interaction. Adv Enzymol Relat Areas Mol Biol 1995; 70:1-21.

10. Yin Z, Gabriele E, Leprini A, Perris R, Colombatti A. Differential cation regulation of the alpha 5 beta 1 integrin-mediated adhesion of leukemic cells to the central cell-binding domain of fibronectin. Cell Growth Differ 1997; 8:1339-47.

11. Nista A, Leonetti C, Bernardini G, Mattioni M, Santoni A. Functional role of alpha4beta1 and alpha5beta1 integrin fibronectin receptors expressed on adriamycin-resistant MCF-7 human mammary carcinoma cells. Int J Cancer 1997; 72:133-41.

12. Kim KS, Tezel TH, Del Priore LV. The reattachment of adult human retinal pigment epithelium (RPE) to extracellular matrix (ECM) is mediated mainly by beta1 integrin subunit. Invest Ophthalmol Vis Sci 1998; 39:S98.

13. Davey G, Buzzai M, Assoian RK. Reduced expression of (alpha)5(beta)1 integrin prevents spreading-dependent cell proliferation. J Cell Sci 1999; 112:4663-72.

14. Chon JH, Netzel R, Rock BM, Chaikof EL. Alpha4beta1 and alpha5beta1 control cell migration on fibronectin by differentially regulating cell speed and motile cell phenotype. Ann Biomed Eng 1998; 26:1091-101.

15. Hakkinen L, Heino J, Koivisto L, Larjava H. Altered interaction of human granulation-tissue fibroblasts with fibronectin is regulated by alpha 5 beta 1 integrin. Biochim Biophys Acta 1994; 1224:33-42.

16. Jin M, He S, Worpel V, Ryan SJ, Hinton DR. Promotion of adhesion and migration of RPE cells to provisional extracellular matrices by TNF-alpha. Invest Ophthalmol Vis Sci 2000; 41:4324-32.

17. Beliveau A, Berube M, Rousseau A, Pelletier G, Guerin SL. Expression of integrin alpha5beta1 and MMPs associated with epithelioid morphology and malignancy of uveal melanoma. Invest Ophthalmol Vis Sci 2000; 41:2363-72.

18. Robbins SG, Brem RB, Wilson DJ, O'Rourke LM, Robertson JE, Westra I, Planck SR, Rosenbaum JT. Immunolocalization of integrins in proliferative retinal membranes. Invest Ophthalmol Vis Sci 1994; 35:3475-85.

19. Brem RB, Robbins SG, Wilson DJ, O'Rourke LM, Mixon RN, Robertson JE, Planck SR, Rosenbaum JT. Immunolocalization of integrins in the human retina. Invest Ophthalmol Vis Sci 1994; 35:3466-74.

20. Kirchhof B, Sorgente N. Pathogenesis of proliferative vitreoretinopathy. Modulation of retinal pigment epithelial cell functions by vitreous and macrophages. Dev Ophthalmol 1989; 16:1-53.

21. Nagasaki H, Shinagawa K, Mochizuki M. Risk factors for proliferative vitreoretinopathy. Prog Retin Eye Res 1998; 17:77-98.

22. Pastor JC. Proliferative vitreoretinopathy: an overview. Surv Ophthalmol 1998; 43:3-18.

23. Pfeffer BA. Improved methodology for cell culture of human and monkey retinal pigment epithelium. In: Osborne NN, Chader J, editors. Progress in retinal Research. Vol 10. Oxford: Pergamon Press; 1991. p. 251-91.

24. Guntern R, Vallet PG, Bouras C, Constantinidis J. An improved immunohistostaining procedure for peptides in human brain. Experientia 1989; 45:159-61.

25. Feeney L. Lipofuscin and melanin of human retinal pigment epithelium. Fluorescence, enzyme cytochemical, and ultrastructural studies. Invest Ophthalmol Vis Sci 1978; 17:583-600.

26. Kennedy BG, Haley BE, Mangini NJ. Creatine kinase in human retinal pigment epithelium. Exp Eye Res 2000; 70:183-90.

27. Beliveau A, Leclerc S, Rouleau M, Guerin SL. Multiple cloning sites from mammalian expression vectors interfere with gene promoter studies in vitro. Eur J Biochem 1999; 261:585-90.

28. Graham FL, van der Eb AJ. A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 1973; 52:456-67.

29. Selden RF, Howie KB, Rowe ME, Goodman HM, Moore DD. Human growth hormone as a reporter gene in regulation studies employing transient gene expression. Mol Cell Biol 1986; 6:3173-9.

30. Pothier F, Ouellet M, Julien JP, Guerin SL. An improved CAT assay for promoter analysis in either transgenic mice or tissue culture cells. DNA Cell Biol 1992; 11:83-90.

31. Rizzolo LJ, Zhou S, Li ZQ. The neural retina maintains integrins in the apical membrane of the RPE early in development. Invest Ophthalmol Vis Sci 1994; 35:2567-76.

32. Anderson DH, Guerin CJ, Matsumoto B, Pfeffer BA. Identification and localization of a beta-1 receptor from the integrin family in mammalian retinal pigment epithelial cells. Invest Ophthalmol Vis Sci 1990; 31:81-93.

33. Kaida M, Cao F, Skumatz CM, Irving PE, Burke JM. Time at confluence for human RPE cells: effects on the adherens junction and in vitro wound closure. Invest Ophthalmol Vis Sci 2000; 41:3215-24.

34. Hunt RC, Pakalnis VA, Choudhury P, Black EP. Cytokines and serum cause alpha 2 beta 1 integrin-mediated contraction of collagen gels by cultured retinal pigment epithelial cells. Invest Ophthalmol Vis Sci 1994; 35:955-63.

35. McKay BS, Irving PE, Skumatz CM, Burke JM. Cell-cell adhesion molecules and the development of an epithelial phenotype in cultured human retinal pigment epithelial cells. Exp Eye Res 1997; 65:661-71.

36. Meitinger D, Hunt DM, Shih DT, Fox JC, Hunt RC. Vitreous-induced modulation of integrins in retinal pigment epithelial cells: effects of fibroblast growth factor-2. Exp Eye Res 2001; 73:681-92.

37. Mousa SA, Lorelli W, Campochiaro PA. Role of hypoxia and extracellular matrix-integrin binding in the modulation of angiogenic growth factors secretion by retinal pigmented epithelial cells. J Cell Biochem 1999; 74:135-43.

38. Yang CH, Huang TF, Liu KR, Chen MS, Hung PT. Inhibition of retinal pigment epithelial cell-induced tractional retinal detachment by disintegrins, a group of Arg-Gly-Asp-containing peptides from viper venom. Invest Ophthalmol Vis Sci 1996; 37:843-54.

39. Zhao MW, Jin ML, He S, Spee C, Ryan SJ, Hinton DR. A distinct integrin-mediated phagocytic pathway for extracellular matrix remodeling by RPE cells. Invest Ophthalmol Vis Sci 1999; 40:2713-23.

40. Akiyama SK, Yamada KM. Biosynthesis and acquisition of biological activity of the fibronectin receptor. J Biol Chem 1987; 262:17536-42.

41. Hotchin NA, Watt FM. Transcriptional and post-translational regulation of beta 1 integrin expression during keratinocyte terminal differentiation. J Biol Chem 1992; 267:14852-8.

42. Immonen I, Tervo K, Virtanen I, Laatikainen L, Tervo T. Immunohistochemical demonstration of cellular fibronectin and tenascin in human epiretinal membranes. Acta Ophthalmol (Copenh) 1991; 69:466-71.

43. Sramek SJ, Wallow IH, Stevens TS, Nork TM. Immunostaining of preretinal membranes for actin, fibronectin, and glial fibrillary acidic protein. Ophthalmology 1989; 96:835-41.

44. Yeo JH, Sadeghi J, Campochiaro PA, Green WR, Glaser BM. Intravitreous fibronectin and platelet-derived growth factor. New model for traction retinal detachment. Arch Ophthalmol 1986; 104:417-21.

45. Goldschild M, Baudouin C. [Role of integrins in ocular physiology and diseases]. J Fr Ophtalmol 1997; 20:311-25.

46. Adams JC, Watt FM. Changes in keratinocyte adhesion during terminal differentiation: reduction in fibronectin binding precedes alpha 5 beta 1 integrin loss from the cell surface. Cell 1990; 63:425-35.

47. Adams JC, Watt FM. Expression of beta 1, beta 3, beta 4, and beta 5 integrins by human epidermal keratinocytes and non-differentiating keratinocytes. J Cell Biol 1991; 115:829-41.

48. Nicholson LJ, Watt FM. Decreased expression of fibronectin and the alpha 5 beta 1 integrin during terminal differentiation of human keratinocytes. J Cell Sci 1991; 98:225-32.

49. Sastry SK, Lakonishok M, Thomas DA, Muschler J, Horwitz AF. Integrin alpha subunit ratios, cytoplasmic domains, and growth factor synergy regulate muscle proliferation and differentiation. J Cell Biol 1996; 133:169-84.

50. Wang D, Birkenmeier TM, Yang J, Venkateswarlu S, Humphrey L, Brattain MG, Sun L. Release from quiescence stimulates the expression of integrin alpha 5 beta 1 which regulates DNA synthesis in human fibrosarcoma HT1080 cells. J Cell Physiol 1995; 164:499-508.

51. Gingras ME, Larouche K, Larouche N, Leclerc S, Salesse C, Guerin SL. Regulation of the integrin subunit alpha5 gene promoter by the transcription factors Sp1/Sp3 is influenced by the cell density in rabbit corneal epithelial cells. Invest Ophthalmol Vis Sci 2003; 44:3742-55.

52. Boles BK, Ritzenthaler J, Birkenmeier T, Roman J. Phorbol ester-induced U-937 differentiation: effects on integrin alpha(5) gene transcription. Am J Physiol Lung Cell Mol Physiol 2000; 278:L703-12.

53. Birkenmeier TM, McQuillan JJ, Boedeker ED, Argraves WS, Ruoslahti E, Dean DC. The alpha 5 beta 1 fibronectin receptor. Characterization of the alpha 5 gene promoter. J Biol Chem 1991; 266:20544-9.

54. Larouche K, Leclerc S, Salesse C, Guerin SL. Expression of the alpha 5 integrin subunit gene promoter is positively regulated by the extracellular matrix component fibronectin through the transcription factor Sp1 in corneal epithelial cells in vitro. J Biol Chem 2000; 275:39182-92.

55. Glaser BM, Lemor M. Pathobiology of proliferative vitreoretinopathy. In: Ryan SJ, editor. Retina. St. Louis: Mosby; 1994. p. 2249-63.