Characterization of Genetically Modified Human Retinal Pigment Epithelial Cells Developed for In Vitro and Transplantation Studies

Abstract

purpose. To develop, by specific genetic modification, a differentiated human retinal pigment epithelial (RPE) cell line with an extended life span that can be used for investigating their function in vitro and for in vivo transplantation studies.

methods. Primary human RPE cells were genetically modified by transfecting with a plasmid encoding the simian virus (SV)40 large T antigen. After characterization, two cell lines, designated h1RPE-7 and h1RPE-116, were chosen for further investigation, along with the spontaneously derived RPE cell line ARPE-19. Factors reported to be important in RPE and photoreceptor cell function and survival in vivo were examined.

results. Both h1RPE-7 and h1RPE-116 cells exhibited epithelial morphology, expressed cytokeratins, and displayed junctional distribution of ZO-1, p100-p120 and β-catenin. The cells expressed mRNA for RPE65 and cellular retinaldehyde-binding protein (CRALBP) and the trophic and growth factors brain-derived neurotropic factor (BDNF), ciliary neurotrophic factor (CNTF), basic fibroblast growth factor (bFGF), pigment epithelium–derived factor (PEDF), nerve growth factor (NGF), platelet-derived growth factor (PDGF)-α, insulin-like growth factor (IGF)-1, and vascular endothelial growth factor (VEGF). Secreted BDNF, bFGF, and VEGF, but not CNTF, were identified in cell supernatants. The cell lines constitutively expressed HLA-ABC, CD54, CD58, and CD59. After activation with IFN-γ both HLA-ABC and CD54 were upregulated, and the expression of HLA-DR was induced. Both cell lines failed to express CD80, CD86, CD40, or CD48 in vitro and in a mixed lymphocyte reaction were unable to induce T-cell proliferation. Fas ligand (CD95L) was not detected in vitro by RT-PCR. Similar results were obtained with the ARPE-19 cell line.

conclusions. RPE lines h1RPE-7 and h1RPE-116 retain many of the morphologic and biochemical characteristics of RPE cells in vivo and may serve as a source of cells for in vitro analysis of RPE cell function, as well as for orthotopic transplantation studies.

Retinal pigment epithelial (RPE) cells perform a complex array of functions that are necessary in maintaining retinal photoreceptor homeostasis. Positioned strategically between the neural retina and the vascular choriocapillaris, the RPE constitutes the outer aspect of the blood–retinal barrier and controls the passage of metabolites to and from the circulation. Dysfunction of the RPE cell layer can lead to devastating effects on retinal function and in some diseases, such as age-related macular degeneration (AMD), ^{1 2 3} can lead to photoreceptor cell death. Our detailed understanding of the cellular function of human RPE cells remains limited, particularly with respect to their role in ocular disease. Much of our understanding has arisen from in vitro investigations of primary or early-passage cells harvested from donor eyes and more recently from the development of a number of human RPE cell lines that are phenotypically more stable in long-term culture. The limited life span of normal human RPE cells represents a substantial obstacle for cell and molecular analysis of these cells, and therefore such cell lines are considered to be an important resource. 

One cell line in particular, spontaneously arising ARPE-19, has been a valuable source of human RPE cells, despite the unknown nature of the immortalizing event conferring immortality. However, the exact profile of RPE-specific characteristics of this cell line remain unknown, and there is a strong case for the development of further RPE cell lines that may express specific characteristics of interest. Furthermore, in in vivo RPE transplantation studies, particularly if these are ultimately intended for treatment of patients, a known genetic modification may be more desirable than an undefined genetic alteration. 

We have previously reported that the subretinal transplantation of RPE cells into the Royal College of Surgeons (RCS) rat delays photoreceptor cell death ^{4 5} and, along with others, ^{6 7 8 9} have proposed that this may be an effective treatment strategy for diseases such as AMD. To overcome some of the limitations imposed by using primary cultures harvested from donor eyes, such as inadequate supply and difficulty in screening for pathogens, the use of cell lines as a source of RPE cells for orthotopic transplantation may prove advantageous. Despite some major potential problems associated with this approach, such as the cells’ safety and ability to function effectively in vivo, it remains a promising strategy. In this study we therefore set out to develop additional human RPE cell lines with extended life span by transfection with the simian virus (SV)40 large T antigen. The generation of a range of cell phenotypes after expression of SV40 large T antigen allows for the systematic selection of lines that exhibit both normal RPE characteristics and a nontransformed phenotype. Furthermore, this approach enables us to generate a pathogen-free cell line generated from an individual with a recorded lineage. The cell lines generated in this study were then tested alongside ARPE-19 cells for RPE-specific and epithelial characteristics and for factors that are believed to be important in cell survival and in the support of photoreceptor cells in vivo. 

Materials and Methods

Reagents

Ham’s F-10, Myoclone heat-inactivated fetal calf serum (FCS; sourced from the United States) and penicillin-streptomycin were purchased from Gibco Life Technologies (Paisley, Scotland, UK). FuGene 6 transfection reagent was obtained from Roche Molecular Biochemicals (Mannheim, Germany). Antibodies for immunocytochemistry were hamster polyclonal anti-SV40 large T antigen (a gift from Mireille Viguier, Institut Cochin de Génétique Moléculaire, Paris, France); rabbit polyclonal ZO-1 (Zymed, Cambridge Bioscience, Cambridge, UK); monoclonal antibodies RGE-53, specific for cytokeratin 18, and RCK-102, specific for cytokeratins 5 and 8 (Eurodiagnostica, Arnhem, The Netherlands); and NCL-5D3, specific for cytokeratins 8 and 18 (Novocastra Laboratories, Newcastle-upon-Tyne, UK); β-catenin (clone 14) and p100-p120 (clone 98; Transduction Laboratories, Nottingham, UK); and rabbit polyclonal von Willebrand factor (Dako, High Wycombe, UK). Monoclonal antibodies used for flow cytometry were anti-human HLA-ABC (clone W6-32), CD48 (clone MEM-102), CD58 (clone BRIC-5), CD59 (clone MEM-43), CD80 (B7-1; clone DAL-1) and CD86 (B7-2; clone BU63), all from Serotec (Oxford, UK); HLA DR (clone G46-6) and CD40 (clone 5C3) from BD Pharmingen (Oxford, UK); and CD54 (ICAM-1; clone BBIG-11) from R&D Systems (Oxford, UK). Tetrarhodamine isothiocyanate (TRITC) and FITC-conjugated goat anti-mouse and goat anti-rabbit secondary antibodies were obtained from Jackson ImmunoResearch Laboratories (West Grove, PA). Primers for RT-PCR were purchased from Oswel (Southampton, UK), and other PCR reagents were from Perkin Elmer (Beaconsfield, UK). Unless specified, all other reagents were obtained from Sigma (Poole, UK) and were of the highest grade available. 

Isolation and Culture of Human RPE Cells

After local ethics committee approval and according to national guidelines and in compliance with the tenets of the Declaration of Helsinki, a human eye was obtained from a 50-year-old female white donor, who underwent exenteration due to conjunctival disease, and the anterior segment, iris, and lens were removed. The posterior segment was placed in sterile culture medium consisting of Ham’s F-10, 20% FCS, 2 mM l-glutamine, 100 IU/mL penicillin, 100 μg/mL streptomycin, and 1 μg/mL amphotericin-B and was stored overnight in the dark at 4°C to facilitate separation of the retina from the RPE. Vitreous and neural retina were gently teased away from the optic nerve, and the eye cup was dissected into three segments. Cloning rings were placed on the exposed RPE cell layer, and cells were rinsed with 0.02% EDTA for 5 minutes. Trypsin-EDTA (0.25%/0.02%) was pipetted into the cloning rings and incubated at 37°C for 45 minutes to detach RPE from Bruch’s membrane. RPE cells, released by gentle trituration, were plated onto FCS-coated tissue culture flasks, and after 10 minutes, culture medium was added. Medium was changed every 2 to 3 days, and all cells were cultured at 37°C in 5% CO₂. For comparative studies, primary cultures of human RPE cells were generated from additional donor eyes, using the same procedure, and the spontaneously arising human RPE cell line, ARPE-19 was obtained from ATCC (CRL-2302; Rockville, MD) and grown as previously reported. ¹⁰  

Genetic Modification of Human RPE Cells

The primary human RPE culture, derived from the single donor described earlier, was transfected with RSV puro (Neurotech S.A., Evry, France; which encodes a puromycin-selectable marker) and either a construct encoding SV40 large T antigen (pVim Twt) or SV40 large T antigen that was deleted for the small T antigen (pVim TΔt). The constructs were generously provided by Denise Paulin, University of Paris, Paris, France). ¹¹ This construct was chosen, because expression of the SV40 large T antigen is known to result in a stable cell phenotype and an associated extension of life span, but without induction of cellular transformation. Transfection was performed in six-well plates seeded with 2 × 10⁴ cells/cm². pVimTwt or pVimTΔt (0.9 μg) was combined with 0.1 μg of RSV puro in 100 μL serum-free medium containing 3 μL transfection reagent (FuGene; Roche). After a 48-hour incubation, cells were trypsinized and replated into Petri dishes and cultured in medium containing 1 μg/mL puromycin to select for transfected cells. Surviving colonies were removed from Petri dishes with cloning rings and cultured in 20% FCS, 2 mM l-glutamine, 100 IU penicillin with 100 μg/mL streptomycin, and 1 μg/mL puromycin. One hundred twenty-six cell lines with large TΔt and 42 cell lines with wild-type large T antigen were generated. Each cell line was assessed by phase-contrast microscopy for ability to form contact-inhibited monolayers of an epithelial cobblestone phenotype and, using morphologic criteria, 18 cell lines were selected for further characterization. 

Characterization of RPE Cells

Immunocytochemistry.

Primary human RPE cultures, the ARPE-19 cell line, and the 18 selected genetically modified cell lines were cultured to confluence on eight-well chamber slides (Gibco) for immunocytochemical characterization. Cells were fixed in either 3.7% paraformaldehyde for 10 minutes or, for cytokeratin detection, ice-cold methanol-acetone (1:1). Cells were permeabilized with 0.25% Triton X-100 and PBS for 10 minutes and blocked for 15 minutes with 10% goat serum and PBS. Primary antibodies were diluted in blocking solution, added to the cells and incubated for 1 hour at room temperature. Detection was achieved using FITC- or TRITC-conjugated secondary antibodies (1:50) added for 1 hour in the dark. After washing, cells were mounted, and the cell expression and distribution recorded by epifluorescence and confocal laser scanning microscopy. Omission of the primary or secondary antibody in each case served as the negative control. On the basis of morphologic appearance and the expression of cytokeratins and SV40 large T and junctional molecules, two cell lines, generated as described earlier, were selected for further analysis. One cell line, designated h1RPE-7, originated from the SV40 large T antigen deleted for the small T antigen–transfected parent line, and another cell line, designated h1RPE-116, originated from transfection with the SV40 large T wild-type antigen. All subsequent analyses described in the following sections were performed on these two cell lines between passages 13 and 22. 

Ultrastructure.

RPE cells were grown to confluence on 24-well tissue culture plates and fixed with one-half strength Karnovsky’s fixative (2% formaldehyde, 2% glutaraldehyde, 0.2 M sodium cacodylate buffer, and 6.5 mM calcium chloride). The monolayer was postfixed in 1% osmium tetroxide for 1 hour, washed, and dehydrated through ascending concentrations of ethanol. For transmission electron microscopy (TEM), the monolayers were embedded in resin and ultrathin sections cut and counterstained with uranyl acetate and lead citrate. The sections were viewed in a transmission electron microscope (model 1010; JEOL, Herts, UK). For scanning electron microscopy (SEM), the monolayers were critical-point dried with CO₂ and sputter coated with 20 nm of gold and then viewed on a scanning electron microscope (model 6100; JEOL). 

Reverse Transcription–Polymerase Chain Reaction.

Expression of mRNAs encoding for discriminating markers, for growth and trophic factors deemed to be important in photoreceptor cell function and survival, and for Fas ligand (FasL; CD95L) were examined using reverse transcription–polymerase chain reaction (RT-PCR). The three cell lines—h1RPE-7, h1RPE-116, and ARPE-19—were investigated for the expression of mRNA for the discriminative cellular markers RPE65 and cellular retinaldehyde-binding protein (CRALBP). The growth and trophic factors investigated were pigment epithelium–derived factor (PEDF), brain-derived neurotrophic factor (BDNF), basic fibroblast growth factor (bFGF), ciliary neurotrophic factor (CNTF), platelet-derived growth factor (PDGF)-α, insulin-like growth factor (IGF-1), nerve growth factor (NGF), and vascular endothelial growth factor (VEGF). Oligonucleotide primers complimentary to the 5′ and 3′ ends were used in RT-PCR studies (Table 1) . 

For identification of discriminating markers and growth and trophic factors, total RNA was extracted from 2-week postconfluent monolayers of h1RPE-7, h1RPE-116, and ARPE-19 cells. For identification of FasL, total RNA was extracted from confluent cell monolayers of h1RPE-7, h1RPE-116, and ARPE-19 cell lines and primary RPE cells, as well as from human T cells. In all cases, total RNA was extracted using a kit (RNeasy; Qiagen, West Sussex, UK). Oligo (dT)_12-18 (1 μg ) primed total RNA was reverse transcribed using avian myoblastosis virus reverse transcriptase (AMV-RT) supplied with a first-strand synthesis system for RT-PCR (SuperScript; Life Technologies). PCR reactions contained 1 μL cDNA, 67 mM Tris-HCl (pH 8.8), 16 mM (NH₄)₂SO₄, 1.5 mM MgCl₂, 0.2 mM dNTPs, 10 pmol sense primer, 10 pmol antisense primer, 1.25 U DNA polymerase (BioTaq; Bioline, London, UK) in a 50-μL reaction. PCR cycle parameters were conducted with a preamplification denaturation at 94°C for 4 minutes, followed by 35 cycles of denaturation at 94°C for 30 seconds, annealing at the primer pair–specific annealing temperature (Table 1) for 30 seconds, and extension at 72°C for 30 to 90 seconds, depending on the amplicon size expected. PCR amplification of β-actin was routinely used as a control to assess integrity of RNA and cDNA. Ten-microliter samples of the amplification reactions were resolved on 1% to 2% TAE agarose gels and the products visualized by ethidium bromide staining. To verify authenticity, amplicons were excised from gels, repurified, and subjected to DNA sequence analysis. Amplicons were cycle sequenced, using sense primers and analyzed on an automated DNA sequencer (ABI 377 Prism; Perkin Elmer). 

Secreted Growth and Trophic Factors.

Cell culture supernatants were collected from either RPE cells derived from the original donor or from h1RPE-7, h1RPE-116, and ARPE-19 cell lines 2 weeks after reaching confluence. Cell supernatants were collected after 5 days of conditioning and centrifuged to remove particulates. The concentrations of bFGF, BDNF, CNTF, and VEGF (R&D Systems), human interferon (hIFN)-α (Amersham PLC, Amersham, UK) and human interferon (hIFN)-β (Biosource International, Camarillo, CA) were measured using commercially available enzyme-linked immunosorbent assays (ELISA) in accordance with the manufacturer’s protocol. 

Flow Cytometry.

The expression of potentially important immunologic molecules was evaluated in h1RPE-7, h1RPE-116, and ARPE-19 cell lines and in third-passage human RPE cells by flow cytometry. Cells were grown in six-well tissue culture dishes and were left either untreated or treated with 100 to 1000 U/mL IFN-γ (R&D Systems) for 48 or 72 hours. After treatment, the cells were prepared as a single-cell suspension by treating with 0.5 mL of 1 mg/mL collagenase for 10 to 15 minutes at 37°C. The cells were pelleted by centrifugation and washed by resuspension in 1 mL sterile PBS. Each cell suspension was then pelleted and resuspended in 100 μL PBS, with or without the appropriate primary antibody and incubated on ice for 1 hour. Two PBS washes were performed and the cells were incubated with a secondary FITC-conjugated antibody for 1 hour on ice in the dark, washed with PBS, and resuspended in 200 to 500 μL PBS for flow cytometric analysis. The cell surface expression of major histocompatibility complex (MHC) class I (HLA-ABC) and class II (HLA-DR) antigens; the costimulatory molecules intercellular adhesion molecule (ICAM)-1 (CD54), CD40, B7-1 (CD80), and B7-2 (CD86); and the potential CD2 ligands CD48, CD58, and CD59 were evaluated by flow cytometry (FACScan; BD Biosciences, Oxford, UK). The data were analyzed on computer (CellQuest software; BD Biosciences). 

Assessment of Tumorigenicity

To assess the tumorigenic capacity of h1RPE-7 and h1RPE-116 cells in vivo, 10 × 10⁶ cells suspended in 200 μL PBS and 10 mM glucose were injected subcutaneously (SC) into irradiated athymic Swiss nude mice (n = 10 animals for each clone). As a positive control, U87 human glioblastoma cells were also injected SC at the same cell density (n = 5 animals). Animals were assessed by daily gross examination and after 3 and 15 weeks by histologic analysis of the graft site. At 3 or 15 weeks after injection the nude mice were perfusion fixed with 4% paraformaldehyde and PBS. The tissue was then postfixed in 4% paraformaldehyde and PBS for 2 hours at 4°C, followed by cryoprotection in 20% sucrose and PBS overnight at 4°C and finally embedded in optimal cutting temperature compound (TissueTek; Miles Laboratories, Elkhart, IN). Tissue was frozen for 30 seconds at− 40°C in isopentane and liquid nitrogen and stored at −20°C until sectioned. Sections were cut (14 μm) and processed for hematoxylin and eosin staining. Tissue sections were finally dehydrated and mounted in resin (Eukitt; Agar Scientific, Essex, UK). Animal care was in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Home Office (United Kingdom) regulations for the care and use of laboratory animals and the provisions of the UK Animals (Scientific Procedures) Act (1986) were observed at all times. 

Anchorage-independent growth, which is a function of cell transformation and tumorigenicity, was assessed using a soft agar colony assay. Low melting temperature agarose (0.5% in culture medium; Seaplaque; Novara Group, Ltd, Staffs, UK) was plated onto 24-well plates and left to gel. The cell lines h1RPE-7 and h1RPE-116 and the human glioblastoma cell line U87 were then suspended in a second layer of 0.33% agarose (7 × 10⁴ cells per well), plated into individual wells, and placed at 4°C for 5 minutes to set. Media were then added to each well and incubated at 37°C with regular changes. The wells were examined by phase-contrast microscopy every 2 to 3 days for up to 40 days for the presence of cell aggregates or foci. 

T-Cell Proliferation Assay: Mixed-Lymphocyte Reaction

Human peripheral blood mononuclear cells (PBMCs) were harvested from a healthy human donor by density-gradient centrifugation (Ficoll-Paque; Pharmacia Biotech AB, Uppsala, Sweden). The PBMCs were then plated onto plastic Petri dishes for 45 minutes at 37°C-5%CO₂ to allow monocytes to adhere. The nonadherent cells were then removed, and the enriched population of T lymphocytes was cocultured (2 × 10⁵ cells/well) with either stimulated (IFN-γ at 100 U/mL for 48 hours) or unstimulated human RPE cell lines that had previously been irradiated (240 Gy). The RPE cell lines, h1RPE-7, h1RPE-116, and ARPE-19 were plated at various concentrations, 2 × 10⁴, 1 × 10⁴, 0.5 × 10⁴, and 0.25 × 10⁴ cells/well in triplicate in 96-well plates in 200 μL RPMI 1640 (Life Technologies) supplemented with 10% FCS, penicillin 100 U/mL, and streptomycin 100 μg/mL. Two irradiated (120 Gy), allogeneic human B-cell lines (MOU, 9050; SA, 9001; from the Tenth International Histocompatibility Workshop, New York, 1987) were used as a positive control at the same concentrations as the RPE cells. The allogeneic T-cells were harvested, and proliferation was assessed on day 6, after an 18-hour pulse with[ ³H]-thymidine (0.5 μCi/well), by measuring incorporated [³H]-thymidine by liquid scintillation spectroscopy. Background uptake of[ ³H]-thymidine into the irradiated cells alone was subtracted from the harvested T-cell data. 

Results

Morphology

Primary cultures of human RPE cells grew as contact-inhibited monolayers and exhibited a cobblestone epithelial morphology with areas of dense pigmentation that were gradually lost as the cells divided in culture (Fig. 1A) . The two human RPE cell lines, h1RPE-7 (Fig. 1B) and h1RPE-116 (Fig. 1C) , also grew as contact-inhibited monolayers and exhibited an epithelial morphology but were devoid of pigmentation. Ultrastructurally, the two human RPE cell lines produced their own basal lamina, had a heterogeneous expression of apical microvilli (Figs. 2A 2B) , and exhibited junctional specialization suggestive of tight junctions (Figs. 2C 2D 2E) . 

Immunocytochemical Characterization

Cells of epithelial origin express cytokeratin intermediate filaments, which are considered a characteristic marker of the epithelial phenotype. Primary RPE cells, and both h1RPE-7 and h1RPE-116 cells showed positive immunoreactivity for a panel of anti-human cytokeratin antibodies including clone RCK102 directed against cytokeratins 5 and 8 (all epithelia) and clones RGE53 (cytokeratin 18) and NCL5D3 (cytokeratins 8 and 18; Figs. 1D 1E 1F ), which recognize nonstratified epithelium. ARPE-19 were also found to be positive for these cytokeratins, whereas all RPE cells tested were negative for the endothelial cell marker, von Willebrand factor (data not shown). The cell lines h1RPE-7 and h1RPE-116 cells were also confirmed as positive for the SV40 large T antigen (data not shown). 

The barrier function and polarization of RPE cells is considered to be highly dependent on the integrity of specialized junctions. Analysis of junctional proteins revealed strong peripheral staining of ZO-1 in both primary RPE and the cell lines h1RPE-7 and h1RPE-116 (Figs. 1G 1H 1I) . The staining pattern of ZO-1 at the point of cell–cell contact was largely continuous but exhibited occasional discontinuities. The adherens junctional proteins, β-catenin (Fig. 1J 1K 1L) and p100-p120 (data not shown) were also expressed at the cell border, but in a more diffuse pattern characteristic of these molecules. Similar results were obtained for the cell line ARPE-19 (data not shown). 

Reverse Transcription–Polymerase Chain Reaction

The discriminatory molecules RPE-65, an RPE cell–specific molecule that is thought to play an important role in the RPE-photoreceptor vitamin A cycle, and CRALBP, which is involved in the regeneration of visual pigment, were both detected by RT-PCR in h1RPE-7, h1RPE-116, and ARPE-19 cell lines producing amplicons of the predicted size (Fig. 3) and correct sequence. 

A panel of growth and trophic factors with potential to affect both RPE and photoreceptor cell function and survival were also investigated by RT-PCR. Positive mRNA expression was observed for CNTF, BDNF, bFGF, NGF, PDGF-α, PEDF, IGF-1, and VEGF (Fig. 3) . Sequence analysis of all RT-PCR products showed amplified sequences were 100% concordant with published human sequences (data not shown). 

Fas and FasL are a receptor-ligand pair involved in the induction of apoptotic cell death. FasL is a type II transmembrane protein and a member of the TNF family that binds to Fas (CD95) and induces apoptosis of the Fas-expressing cells. The expression of FasL by grafted RPE cells, therefore, may be a critical determinant in their ability to attenuate the rejection process through FasL-mediated apoptosis of host T cells after transplantation. The cell lines h1RPE-7, h1RPE-116, and ARPE-19 as well as a third-passage human RPE cell culture, were all found to be negative for FasL mRNA. However, T cells assessed under identical stringent conditions were clearly positive for FasL (Fig. 4) , demonstrating that in vitro, these RPE cell lines do not express FasL. 

Secretion of Growth and Trophic Factors

ELISAs were performed on serial dilutions of cell culture supernatants, conditioned for 5 days, from a minimum of three independent experiments. BDNF, bFGF, and VEGF were positively detected in primary RPE, h1RPE-7, h1RPE-116, and ARPE-19 cells (Fig. 5) . Secretion of bFGF by both h1RPE-7 and h1RPE-116 was 30-fold higher than in primary RPE and ARPE-19 cells (P < 0.001; Fig. 5B ), whereas secretion of VEGF, was much lower than in primary RPE cells (P < 0.001; Fig. 5C ). CNTF, hIFN-α, and hIFN-β were not detected above the limit of the assays. 

Tumorigenicity Testing

After subcutaneous transplantation of h1RPE-7 or h1RPE-116 cells into irradiated athymic mice, neither cell line formed tumors over a 15-week period. Conversely, the human U87 glioblastoma cell line formed large tumors by 3 weeks after transplantation (data not shown). 

The anchorage-independent growth of both h1RPE-7 and h1RPE-116 was determined by growing cells in soft agar. These cell lines failed to form aggregates or colony foci over a 40-day culture period (data not shown), indicating that they are both nontransformed and nontumorigenic. 

Expression of Molecules of Potential Immunologic Importance

The expression of surface-expressed molecules of potential immunologic importance was determined in resting or IFN-γ–activated (100–1000 U/mL for 48 or 72 hours) cells by flow cytometry. RPE cell lines h1RPE-7, h1RPE-116, and ARPE-19 were all found to constitutively express MHC class I (HLA-ABC). After activation with IFN-γ, class I was upregulated, and the expression of MHC class II (HLA-DR) was induced (Fig. 6) . Similar results were obtained for third-passage human RPE cells. 

Under resting conditions the cell adhesion–costimulatory molecule ICAM-1 (CD54) was expressed constitutively in all cells studied and was upregulated after activation with IFN-γ. The costimulatory molecules B7-1 (CD80) and B7-2 (CD86) were not expressed on the three cell lines, nor were they expressed on early-passage human RPE cells under either resting conditions or after activation by IFN-γ (Fig. 6) . Furthermore, CD40 expression was negligible on the three cell lines but was found on a proportion of the third-passage human RPE cells. No differences were observed between groups activated with 100 or 1000 U/mL IFN-γ for either 48 or 72 hours. 

The CD2 ligand, CD58 (lymphocyte function-associated antigen-3), was also found to be constitutively expressed by RPE cells and remained unchanged after cytokine activation (Fig. 6) . The expression of the two potential low-affinity CD2 ligands, CD48 and CD59, was also investigated. Both h1RPE-7 and h1RPE-116 failed to express CD48 in vitro, whereas both ARPE-19 and third-passage human RPE cells expressed low levels. Conversely, CD59 was strongly expressed by all RPE cells (Fig. 6) . 

The haplotype of h1RPE-7 and h1RPE-116, as determined by PCR–sequence specific priming (SSP), was found to be HLA-A 3;32, -B 44;62, -C 9;5, -DR 1;4, and -DQ 5;7. 

T-Cell Proliferation Assay

To evaluate the potential of the RPE cell lines to induce T-cell proliferation, a mixed-lymphocyte reaction assay was performed. The three RPE cell lines—h1RPE-7, h1RPE-116, and ARPE-19—did not induce significant T-cell proliferation (Fig. 7) irrespective of IFN-γ activation or the number of RPE cells used (data not shown). As expected, the two B-cell lines used as positive controls were able to induce significant T-cell proliferation in the absence of other costimulatory factors, particularly at a concentration of 2 × 10⁴ cells/well (Fig. 7) . 

Discussion

In this study, we generated two new human RPE cell lines and, together with the previously described spontaneously immortalized cell line ARPE-19, investigated in vitro various properties that may be considered important in studying RPE cell biology. By comparing the properties of these cell lines, we have endeavored to identify those characteristics that would demonstrate their suitability for use in vitro in cell biological studies and in vivo in proof-of-principle transplantation studies in the RCS rat. 

Transfection of a primary culture of human RPE cells with SV40 large T antigen gave rise to a large number of cell lines exhibiting a wide range of characteristics. On the basis of morphologic phenotype and preliminary screening for characteristic markers, two lines were selected. Using established assays for determining tumorigenicity, we demonstrated that the two new cell lines generated here did not proliferate in irradiated athymic nude mice and did not show anchorage-independent growth. Moreover, in subretinal grafting studies using the h1RPE-7 and ARPE-19 cell lines in the RCS rat we did not observe tumor formation up to 6 months after grafting. ¹²  

For RPE cells to exhibit barrier function and to undergo apical and basolateral polarization, they must be capable of the successful assembly of adherence junctions and the subsequent formation of functional tight junctions, the latter of which are composed of a complex of transmembrane and intracellular proteins including ZO-1. It has previously been shown that a continuous unbroken expression of ZO-1 correlates with functional junctions, whereas a loss or reduction in ZO-1 expression is associated with breakdown of barrier function. Both h1RPE-7 and h1RPE-116 cell lines expressed ZO-1 as a near-continuous belt at the point of cell–cell contact in a pattern identical with primary cultures of RPE and to ARPE-19 (data not shown). The pattern of expression, however, showed occasional flaws or discontinuities, and this was consistent with our inability to record transmonolayer electrical resistances above 30 to 40Ω /cm² under normal conditions (Kanuga et al., unpublished observations, 2000). These results, however, are consistent with previously reported transmonolayer electrical resistances of ARPE-19 cells in the absence of specialized media. ¹⁰ The adherens junction components, p100-p120 andβ -catenin, were also found to be strongly expressed in vitro at the point of cell–cell contact in the two cell lines in a manner identical with that observed in primary RPE cells and ARPE-19. Although the junctional expression of these molecules is a prerequisite for the formation of fully functional junctions, their appearance per se does not necessarily indicate functional junctions, because this is dependent on the correct assembly of the junctional complex as a whole. Although we did not investigate the associated question of cell polarization in this study, it is interesting to note that when grafted subretinally into RCS rats, both h1RPE-7 and ARPE-19 cell lines are able to integrate with the host RPE, synthesize pigment granules, and express them at their apical face, which suggests some degree of polarization. ¹² Clearly, further work is required to evaluate the capacity of these cells to exhibit functional polarity and the factors controlling this process. 

The participation of RPE cells in the visual cycle is well documented. These cells are engaged in the synthesis and storage of retinyl esters, isomerization of all-trans-retinoids to 11-cis-retinoids and the eventual conversion of retinol to retinal. The expression of genes implicated in retinal vitamin A metabolism is therefore an essential property for any RPE cell line. RPE65, a tissue-specific, highly conserved protein, is present at high levels in vivo and is thought to play an important role in the RPE-photoreceptor vitamin-A cycle. ^{13 14} In this study, we demonstrated mRNA expression of both the full-length and nested transcript of RPE65 in h1RPE-7 and h1RPE-116 cells. We also demonstrated the presence of mRNA for CRALBP in the two cell lines. This important protein is involved in the regeneration of visual pigment and has been implicated in autosomal recessive retinitis pigmentosa. ¹⁵ The finding that these genes are also expressed in ARPE-19 is consistent with a previous report. ¹⁰  

Recent observations have dramatically illustrated the effects of either a single growth and trophic factor or a combination of these on photoreceptor viability in dystrophic animal models. In particular bFGF, BDNF, and CNTF have been shown to be neuroprotective in the retina. ^{16 17 18 19 20 21} The capacity of the RPE cells to produce these factors was therefore investigated and bFGF, BDNF, and CNTF were all expressed at the transcriptional level, although only bFGF and BDNF were detected in conditioned supernatant. bFGF secretion by h1RPE-7 and h1RPE-116 cells was found to be significantly higher than that produced by primary RPE or ARPE-19 cells and in concentrations that are physiologically relevant. ^{22 23} These studies demonstrate the capacity of these cells to produce potentially important neurotrophic factors that may protect photoreceptors and enhance survival, although it remains to be shown whether these factors are produced once the cells are grafted into the retina. 

The eye has been described as an immune-privileged site, with its immunity attributed in part to the expression of FasL, particularly at the site of the blood–ocular barriers. ²⁴ FasL is a type II integral membrane protein homologous with the TNF receptor and is important for protection against inflammatory processes. In the eye it has been reported to be constitutively expressed on RPE cells where it promotes apoptosis of invading Fas-positive immune cells. ^{25 26} Thus, FasL expression by RPE cells may play an essential role in local immune regulation and in a transplantation setting in the prevention of immune rejection of transplanted cell lines. ²⁷ The in vitro detection of FasL on RPE, however, has been problematic, because of reported problems associated with certain FasL antibodies. ^{28 29} In immunocytochemical assays we were also unable to show the presence of surface expressed FasL on RPE cells in vitro (data not shown), and this prompted us to restrict our investigation to the use of RT-PCR for the detection of FasL. The three cell lines that we studied, as well as the early-passage human RPE cells, were all found to be negative for FasL at the transcriptional level. It is likely, therefore, that, as previously reported, ²⁹ FasL is expressed only on RPE cells in vivo. It remains to be determined, however, whether these RPE cell lines are able to express this important immunomodulatory molecule when reintroduced subretinally. 

The expression of MHC class II molecules on RPE has also been the subject of much conjecture in relation to their ability to act as antigen-presenting cells. Furthermore, in transplantation studies it has been suggested that RPE allografts are vulnerable to rejection as a result of the host immune response being directed against transplantation autoantigens. ^{30 31} Using immunocytochemical (not shown) and flow cytometric analysis, we have shown that the RPE cell lines did not express MHC class II constitutively but that it was induced after cytokine activation. These results are consistent with previous reports describing MHC expression in human RPE cells. ³² The expression of costimulatory molecules would also determine whether these cells are potentially capable of inducing T-cell activation. We found that, under in vitro conditions, the three cell lines expressed the adhesion–costimulatory molecule ICAM-1 (CD54) but did not express B7-1 (CD80) or B7-2 (CD86) in either resting or IFN-γ–activated cells. These results are generally consistent with other studies performed on primary or early-passage cultures of RPE cells, in which CD54 has been shown to be expressed, ³³ but not CD80 and CD86. ^{33 34 35} However, unlike a previous study in human fetal RPE cells, ³⁴ we were unable to confirm the expression of CD40 on the RPE cell lines, although we observed a significant population of positive cells in the early-passage human RPE cells. This may indicate that the cell lines have lost the capacity to express CD40. 

The alternative CD2-mediated pathway for T-cell activation has been suggested as a possible mechanism by which RPE cells may also activate T cells. ³⁵ We have shown that the three cell lines, h1RRPE-7, h1RPE-116, and ARPE-19, express similar levels of the CD2 ligand, CD58, under both resting conditions and after IFN-γ activation. Both CD48 ^{36 37} and CD59 ^{38 39} have also been proposed as possible low-affinity ligands for human CD2, although their functional relevance in humans has been questioned. ^{35 40} However, in the rat, both CD48 and CD59 may act as major ligands for CD2 ³⁵ and, in that these cell lines have been grafted into rat retina, ¹² we investigated the expression of these ligands on the RPE cells. Both CD58 and CD59 were expressed constitutively, whereas CD48 was expressed only weakly on ARPE-19 and early-passage RPE cells. It remains unclear whether in the in vivo setting either CD58 or CD59 is able to activate T cells through this alternative pathway. 

Although the cell lines were found to express MHC class II, the absence of costimulatory molecule expression suggests that they would not, at least in vitro, be able to induce T-cell proliferation. This was confirmed in a mixed lymphocyte reaction where the cell lines were unable to induce T-cell proliferation unlike a B-cell line. Should these properties be retained after transplantation of these cells into the subretinal space, it would suggest that in an allogeneic setting there would be a limited immune response, thus improving the longevity and viability of grafted cells. 

In this study, the cell lines h1RPE-7 and h1RPE-116 retained many of the phenotypic features of RPE cells in vivo. The suitability of these newly generated cell lines, along with that of ARPE-19, for in vitro research of RPE function are currently under way. ⁴¹ In addition, we have recently reported that subretinal transplantation of these cells in the RCS rat is able to preserve vision as assessed using histologic, behavioral, and physiological criteria. ^{12 42} These observations highlight the value of generating carefully characterized human RPE cell lines for both in vitro and in vivo studies. 

 Supported by Neurotech SA, Bâtiment Génopole-Industries, Evry, France.

 Submitted for publication July 9, 2001; revised October 11, 2001; accepted October 24, 2001.

 Commercial relationships policy: E (POC, LB, AK, AZ); F (RDL, PC, PA, JG); P (PA, RDL, JG); N (NK, HLW, SH, DK).

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Primer Sequences used for PCR

Gene	Primer Sequence (5′-3′) Position in mRNA	Annealing Temp (°C)	Product Size (bp)
RPE-65 5′	ATG TCT ATC CAG GTT GAG
RPE-65 3′	TCA AGA TTT TTT GAA CAG	52	1601
RPE-65 5′	CCT TTC TTC ATG GAG TCT TTG
RPE-65 3′	ATT GCA GTG GCA GTT GTA TTG	52	390
CRALBP 5′	ATG TCA GAA GGG GTG GG
CRALBP 3′	TCA GAA GGC TGT GTT CTC A	60	953
PEDF 5′	GGA CGC TGG ATT AGA AGG CAG
PEDF 3′	TTG TAT GCA TTG AAA CCT TAC AGG	65	1490
BDNF 5′	ATG ACC ATC CTT TTC CTT ACT ATG GT
BDNF 3′	TCT TCC CCT TTT AAT GGT CAA TGT AC	52	741
bFGF 5′	GCC TTC CCG CCC GGC CAC TTC AAG G
bFGF 3′	GCA CAC ACT CCT TTG ATA GAC ACA A	55	179
CNTF 5′	TGG CTA GCA AGG AAG ATT CGT
CNTF 3′	ACG AAG GTC ATG GAT GGA CCT	65	468
PDGFα 5′	CCT GCC CAT TCG GAG GAA GAG
PDGFα 3′	TTG GCC ACC TTG ACG CTG CG	65	225
IGF-1 5′	ATG TCC TCC TCG CAT CTC TTC
IGF-1 3′	CCT GTA GTT CTT GTT TCC TGC	65	337
NGF 5′	TCG GCA TAC AGG CGC AAC CA
NGF 3′	CCT GCT TGC CAT CCA TGG TC	55	612
VEGF 5′	TTG CCT TGC TGC TCT ACC TC
VEGF 3′	AAA TGC TTT CTC CGC TCT GA	65	424
FasL 5′	GGA TTG GGC CTG GGG ATG TTT CA
FasL 3′	TTG TGG CTC AGG GGC AGG TTG TTG	67	344
β-Actin 5′	GAG CAC AGA GCC TCG CCT TTG C
β-Actin 3′	GGA TCT TCA TGA GGT AGT CAG TCA GG	65	620

Figure 1.

 

Phase-contrast photomicrographs of (A) a primary-donor RPE cell culture (day 3) and the derived cell lines, (B) h1RPE-7 (passage 15) and (C) h1RPE-116 (passage 17). Cytokeratin-8 and -18 expression in third-passage donor RPE cells (D), h1RPE-7 (E), and h1RPE-116 (F). ZO-1 andβ -catenin expression in third-passage donor RPE cells (G and J, respectively), passage 16 h1RPE-7 cells (H and K, respectively), and passage 17 h1RPE-116 (I and L, respectively). Bars, (A–C) 100μ m; (D–F) 50 μm; (G–L) 20 μm.

Figure 2.

 

Electron micrographs of h1RPE-7 cells. (A) Scanning electron micrograph showing microvilli on apical surface. (B) Transmission electron micrograph showing microvilli (arrows). (C–E) Transmission electron micrographs depicting junctional specialization (arrowheads) and actin filaments radiating from the junctional region (arrows). Bars, (A) 1 μm; (B–E) 500 nm.

Figure 3.

 

RT-PCR analysis of discriminatory molecules and growth and trophic factors in RPE cells. RNA was extracted from confluent flasks of h1RPE-7, h1RPE-116, and ARPE-19 cells. RT was performed on 1 μg RNA. Lane 1: commercially supplied marker; lane 2: h1RPE-7; lane 3: h1RPE-116; lane 4: ARPE-19. cDNA (1 μL) was used for all PCR reactions andβ -actin was used as a positive control throughout. Sequencing was undertaken to verify authenticity of all PCR products.

Figure 4.

 

RT-PCR analysis of FasL (CD95L) in RPE cells. RNA was extracted from confluent flasks of h1RPE-7, h1RPE-116, and ARPE-19 cell lines and from primary human RPE cells and T cells. RT was performed on 1 μg RNA. Sequencing was undertaken to verify authenticity of all PCR products.

Figure 5.

 

ELISA for RPE secretion of (A) BDNF, (B) bFGF, and (C) VEGF in primary RPE, h1RPE-7, h1RPE-116, and ARPE-19 cell culture supernatants conditioned for 5 days. Data are the mean ± SD from a minimum of three independent experiments. Significant differences from primary RPE cultures (*P < 0.001; **P < 0.05).

Figure 6.

 

Flow cytometric analysis of surface-expressed antigens of immunologic importance. Expression of HLA-ABC, HLA-DR, CD54, CD80, CD86, CD40, CD48, CD58, and CD59 in h1RPE-7, h1RPE-116, ARPE-19, and third-passage RPE cells. Histograms depict relative fluorescence intensity (log scale) of negative control (gray line), nonactivated (black line), and IFN-γ–activated (shaded histogram) RPE cells.

Figure 7.

 

Example of a mixed-lymphocyte reaction. h1RPE-7, h1RPE-116, ARPE-19, or the B-cell lines 7 and 29 (2 × 10⁴ cells/well) were cultured in the presence of human T cells (2 × 10⁵ cells/well). RPE cells did not induce significant T-cell proliferation (no significant difference between h1RPE-7, h1RPE-116, and ARPE-19), whereas significant proliferation was observed when cells were cocultured with the B7 and B29 B cells. Significant difference from RPE cell lines (*P < 0.0005).

The authors thank Peter Munro for generous help with the electron microscopy; Pascal Chaux for valuable critical comment; and John Hungerford, Phillip Luthert, and Alan Bird (Moorfields Eye Hospital and Institute of Ophthalmology, London, UK) for their efforts in the provision of human donor tissue.