Effect of Robo1 on Retinal Pigment Epithelial Cells and Experimental Proliferative Vitreoretinopathy

Abstract

Purpose.: The Roundabout (Robo) family of proteins is related to the transmembrane receptors and plays a major role in neurogenesis. However, the role of the Robo proteins in proliferative retinopathy has not yet been defined. This study was conducted to determine whether Robo1 is expressed in the retina of patients with proliferative retinal disease and whether it has a pathobiological role in the disease.

Methods.: Immunohistochemistry was used to determine the presence and distribution of Robo1 in the pathologic membranes in proliferative retinopathy. Small interfering (si)RNA technology was used to knockdown Robo1 expression and to study its effects on retinal pigment epithelial (RPE) cells in vitro. The impact on PVR development of blocking Robo1 expression was determined by applying specific siRNA in a PVR rabbit model. The prevalences of PVR and retinal detachment were determined by indirect ophthalmoscope on days 1, 3, 7, 14, 21, and 28 after the injection of RPE cells into the vitreous.

Results.: Immunohistochemistry showed that Robo1 expression was detected in GFAP-labeled glial cells and cytokeratin-labeled RPE cells in proliferative membranes. Robo1 expression was also detected in CD31-labeled vascular endothelial cells. Knockdown of Robo1 expression not only reduced human RPE cell proliferation in vitro but also effectively suppressed the development of PVR in a rabbit model.

Conclusions.: Robo1 is present in the extracellular matrix of proliferative membranes and may be derived from dedifferentiated RPE cells. Silencing the expression of Robo1 in RPE cells inhibited cell proliferation and suppressed the development of PVR in an animal model, indicating a potential therapeutic usefulness in treating PVR.

The pathologic proliferation of intraocular tissues and cells, such as retinal pigment epithelial (RPE) cells in proliferative vitreoretinopathy (PVR) and vascular cells in eyes with ischemic retinopathies, such as proliferative diabetic retinopathy (PDR), is an important problem in clinical ophthalmology. The term proliferative retinopathy refers to distinct conditions characterized by cellular proliferation and matrix deposition within the retina. ^1–3 Although modern vitreoretinal microsurgery has improved treatment in PVR and PDR, in many cases there are technical failures or the visual results are poor. ^4–6 It is therefore important to obtain a better understanding of the pathogenesis of this disorder, to devise strategies for its prevention and for improved treatments. 

Robo1 was first identified in Drosophila during a comprehensive screening for genes controlling central nervous system (CNS) midline crossing. In Robo1 mutants, ipsilateral axons that normally avoid the midline cross it, and commissural axons are seen crossing and recrossing the midline repeatedly. ⁷ Robo1 has been seen to be expressed in most mouse and rat organs, such as the developing spinal cord, thymus, teeth, spleen, lymph nodes, liver, trigeminal ganglia and hippocampus, and kidney, as well as in scattered cells in the developing retinal ganglion cell (RGC) layer of the eye. ^8–15 Homozygous Robo1/Dutt1-knockout mice are embryonically lethal, but heterozygous mice have been found to develop lymphomas and lung adenocarcinomas at a high frequency. ¹⁶ The human ROBO1/DUTT1 locus has been found to be deleted in a child that showed a developmental delay and congenital anomalies but did not develop cancer. ¹⁷ These observations seem to pose a dilemma for understanding the functions of ROBO1 in different species. Overexpression of ROBO1 has been reported in hepatocellular carcinoma and colorectal cancer and is localized in tumor endothelial cells especially in tumor cells. ^18,19 In addition, the soluble Robo1 was detected in the culture medium of hepatocellular carcinoma cancer cell lines and in sera from hepatocellular carcinoma patients. ¹⁸ Recently, Wang et al. ²⁰ demonstrated a relation between Robo1 and tumor angiogenesis in human melanoma: Robo1 was detected on vessels of malignant melanoma but not on vessels of normal tissue. Furthermore, microvessel density and tumor volume could be reduced by applying a soluble form of Robo1 as well as a blocking antibody in a xenograft model. These studies demonstrate the importance of Robo1 signaling in tumor angiogenesis and growth. Furthermore, Robo1 has been shown to be involved in embryogenesis: Robo1 expression has been detected in all tissues examined at embryonic days 12.5 and 15.5, whereas in adults, it has been found only in the brain, kidney, and eye. ²¹  

Although there have been many studies of Robo1 expression and its role in embryogenesis and tumorigenesis, there have been no reports about it in relation to human retinal diseases. In this study, we characterized the expression of Robo1 in the proliferative retinal diseases and explored its possible role in proliferative retinopathy. 

Methods

Cell Culture and Reagents

Human RPE cells (D407 cell line) were obtained from the American Tissue Culture Collection (Manassas, VA) and were cultured in DMEM with 10% fetal bovine serum (FBS; Invitrogen-Gibco, Grand Island, NY), 100 U/mL penicillin, 100 μg/mL streptomycin at 37°C under 5% CO₂, and 95% humidified air. The transfection reagent (HiPerfect) was purchased from Qiagen (Hilden, Germany). A first-strand cDNA synthesis kit and real-time SYBR green RT-PCR mastermix were purchased from Toyobo (Osaka, Japan). For Western blot analysis, anti-Robo1 and anti-cyclin D1 antibodies were purchased from Abcam (Cambridge, UK), and an HRP-goat anti-rabbit secondary antibody was obtained from Rockland Immunochemical, Inc. (Gilbertsville, PA) at a dilution of 1:4000. For immunochemistry, the Robo1 antibody was used at a dilution of 1:100. Monoclonal antibodies specific for human CD31 (1:100 dilution), glial fibrillary acidic protein (GFAP, 1:100 dilution), and pan-cytokeratin (1:100 dilution) were purchased from Chemicon (Temecula, CA). Rhodamine phalloidin (Cat. No. PHDR1) was purchased from Cytoskeleton (Denver, CO). Goat anti-mouse FITC and mouse anti-rabbit FITC-conjugated secondary antibodies were obtained from Sigma-Aldrich (St. Louis, MO) and used at a dilution of 1:100. A nuclear staining kit for detecting the distribution of cells in the various cell cycle phases (Cycletest Plus DNA Reagent Kit) was obtained from BD Biosciences (Bedford, MA). All other reagents were purchased from Sigma-Aldrich. 

Tissue Samples

Epiretinal or subretinal membranes were surgically excised from patients with PDR and PVR (Table 1). Human tissue was obtained after proper consent and approval, in compliance with the Declaration of Helsinki. Membrane tissues were put into phosphate-buffered saline (PBS; pH 7.4) and mounted in optimal cutting temperature (OCT) medium (Merck, Darmstadt, Germany), and 6-μm sections were cut. Thawed tissue sections were air dried, fixed in 4% PFA (20 minutes), washed with phosphate-buffered saline (PBS), and blocked with 10% normal goat serum for 1 hour at 37°C. Anti-Robo1 polyclonal antibody with anti-GFAP, anti-CD31, or anti-cytokeratin was applied to the tissue sections at 4°C overnight and incubated for 1 hour at 37°C with TRITC-conjugated goat anti-rabbit, FITC-conjugated mouse anti-goat, and goat anti-mouse secondary antibodies. After incubation, the slides were washed and the cell nuclei were stained with DAPI (4′, 6′-diamino-2-phenylindole). Images were acquired with a fluorescence microscope equipped with a digital camera. For each of the immunostaining procedures, negative controls included omission of the primary antibody and use of an irrelevant polyclonal or isotype-matched monoclonal primary antibody; in all cases, negative controls showed only faint, insignificant staining. 

Table 1.

 

Clinical Characteristics for Individual Proliferative Retinal Membranes

Patient No.	Age (y)	Sex	Diagnosis	Duration
1	58	M	PDR epiretinal membrane	15 y
2	60	M	PDR subretinal membrane	20 y
3	65	M	PDR subretinal membrane	18 y
4	55	F	PDR epiretinal membrane	6 y
5	62	F	PDR epiretinal membrane	11 y
6	70	F	PDR subretinal membrane	13 y
7	8	M	PVR epiretinal membrane	2 mo
8	10	M	PVR epiretinal membrane	1 y
9	15	M	PVR subretinal membrane	1 mo
10	22	M	PVR subretinal membrane	8 mo
11	59	M	PVR epiretinal membrane	2 wk
12	65	M	PVR epiretinal membrane	3 wk
13	12	F	PVR subretinal membrane	2 y
14	45	F	PVR epiretinal membrane	2 mo

siRNA and Transfection Assays

Robo1 (GenBank accession no. NM002941)-specific small interfering (si)RNAs (Robo1-siRNA: forward: 5′-GGA UGU AUU UGC AAC AAG ATT-3′; reverse: 5′-UCU UGU UGC AAA UAC AUC CTT-3′) were chemically synthesized by Qiagen. D407 RPE cells were transfected with siRNA (Hiperfect Transfection Reagent; Qiagen) according to the manufacturer's instructions. Briefly, the original stock of the siRNA was suspended in the siRNA suspension buffer provided by the manufacturer and stored at −20°C until used. On the day of transfection, the cells were seeded in plates at the recommended densities. siRNA was then gently introduced into the cells by mixing with the required amount of transfection reagent, at a final concentration of 10 nM siRNA. Nonsilencing siRNA (NS siRNA forward: 5′-UUC UCC GAA CGU GUC ACG UTT-3′; reverse: 5′-ACG UGA CAC GUU CGG AGA ATT-3′) was used to control for any effects of the transfection reagent and siRNA. The in vitro assays described herein were performed 48 hours after transfection. 

RNA Isolation and Real-Time RT-PCR

Total RNA was isolated (Trizol; Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Two micrograms of retinal RNA were converted into cDNA. ²² The single-stranded cDNA was amplified in a polymerase chain reaction (PCR), with sequence-specific primers for GAPDH (forward primer: 5′-GAG TCC ACT GGC GTC TTC AC-3′; reverse primer: 5′-GTT CAC ACC CAT GAC GAA CA-3′, or for human Robo1 (forward primer: 5′-AAA TAT GGT GGG CAA AGC TG-3′; reverse primer: 5′-CTG GAT GAC TGT GGT GGT TG-3′). The real-time PCR assays were performed according to the manufacturer's instructions. Robo1 was normalized to GAPDH expression and calculated using the equation: change (x-fold) = 2^−ΔΔCt. 

Western Blot Analysis

Cells were washed with ice-cold PBS, prepared by using the protein extraction and protease inhibitor kits (Pierce), and cleared by centrifugation at 12,000g at 4°C. The supernatant was collected, and the protein content of each lysate was measured by using a BCA protein assay kit (Pierce), according to the manufacturer's instructions. Equal amounts of protein were loaded and analyzed by immunoblot. The proteins were visualized with enhanced chemiluminescence Western blot detection reagents (Pierce). Band densities of the Robo1 and cyclin-D1 proteins were normalized to β-actin. Western blot analyses were repeated three times, and qualitatively similar results were obtained. 

Immunocytochemistry

D407 RPE cells grown on glass coverslips were washed and fixed with 4% PFA in PBS and then permeabilized with 0.1% Triton X-100 before blocking with 10% goat serum. The slides were incubated with Robo1 at 4°C overnight and then were washed with PBS and incubated for 1 hour at 37°C with TRITC-conjugated mouse anti-rabbit secondary antibody. After the incubation, the slides were washed, and the cell nuclei were stained with DAPI. Images were acquired with a fluorescence microscope equipped with a digital camera. In each case, preimmune IgG and secondary control incubations were conducted to determine the staining specificity. 

Cytotoxicity and Cell Proliferation Assays

Cell lines were plated at 1 × 10⁴ per well in 96-well plates and were allowed to adhere overnight. Cell survival was determined by MTT test. ²³ Three controls were used: one without transfection, one with the transfection reagent only, and one with control siRNA. After incubation for 48 hours, MTT was added and the cells were incubated for a further 4 hours. Absorbance was measured with an ELISA plate reader at a wavelength of 570 nm. To detect the toxicity of the specific siRNA, we also evaluated apoptosis. Cells (1 × 10⁶) were seeded in six-well plates with NS siRNA and Robo1-siRNA transfection for 48 hours. Then, the cells were detached with EDTA, washed in cold PBS, and stained with propidium iodide and annexin-V-FITC. The samples were analyzed on a flow cytometer (FACSCaliber; BD Biosciences, Franklin Lakes, NJ). 

Cell proliferation was measured by a modified MTT assay on days 1, 2, 3, 4, and 5. Briefly, the cells were washed twice with PBS and MTT was added to each well. The plates were incubated at 37°C for 4 hours, 100 μL DMSO was added, and the plates were shaken gently for 15 minutes, to solubilize the formazan blue crystals. Absorbance at 570 nm was measured with an ELISA plate reader (Dynatech Medica, Guernsey, UK). Each experiment was performed in three wells and was duplicated at least three times. 

Cell-Attachment Assay

The 96-well plates were coated with 1.25 g/mL of fibronectin in 100 μL PBS overnight at 4°C. Transfected cells (1 × 10⁴) were trypsinized and added to each well and allowed to attach for 60 minutes. The cells were washed gently with PBS twice, and fresh medium (150 μL) containing MTT was added to each well. Absorbance was measured by ELISA plate reader at 570 nm. The experiments were performed in triplicate and repeated three times. 

Cell-Spreading Assay

After transfection, 1 × 10⁴ cells were trypsinized and added to fibronectin-coated coverslips in DMEM and 10% serum at 37°C for 2 hours. After they were washed with PBS three times, the cells were fixed with 4% PFA in PBS and then were stained with vimentin and DAPI. For quantification of cell spreading, four separate fields were photodocumented with a microscope (Leica, Deerfield, IL). Cell spreading was characterized by the formation of a clearly defined cytoplasm halo around the cell nucleus. Quantitation was performed by measuring the ratio of cytoplasm area to nucleus area of cells in each field (Image-Pro Plus software; Media Cybernetics, Inc., Bethesda, MD). Experiments were performed in triplicate and repeated three times. 

Cell Migration

Migration was assayed in a modification of the Boyden chamber method with polycarbonate filters (Nucleopore, Karlsruhe, Germany) with a pore size of 8.0 μm, as described before. ²⁴ Briefly, 2 × 10⁴ cells were placed in the top part of a modified Boyden chamber in a volume of 200 μL of serum-free medium. Serum (containing 10% FBS) was placed in the bottom chamber in a final volume of 600 μL. All migration assays were conducted at 37°C for 4 hours. At the end of the assay, the cells were fixed in 4% PFA and stained with DAPI for 15 minutes. The cells that had not migrated were removed with a cotton swab, and the membrane was imaged. The number of cells from five random fields of view was counted. 

Immunofluorescence Analysis of Microfilaments

Microfilament organization of RPE cells was assessed by a modification of the immunofluorescence analysis protocol with rhodamine-phalloidin, as described elsewhere. ^25,26 After transfection, the cells were trypsinized and seeded onto coverslips for 4 hours at 37°C in 5% CO₂. The medium was aspirated, and adherent cells were fixed with 4% PFA in PBS for 20 minutes. After they were washed with PBS (pH 7.4) three times, the cells were permeabilized with 0.1% Triton X-100 for 20 minutes and blocked with 1% BSA in PBS for 5 minutes. The cells were then incubated with rhodamine-phalloidin (200 U/mL) for 30 minutes and DAPI (0.1 μg/mL) for 1 minute in the dark. PBS was used as the base of all solutions and intervening rinses, and incubations were performed at room temperature. After the cells were mounted, the slides were examined by epifluorescence microscope (Leica), with appropriate excitation and emission filters under 100× magnification. Images of observed fields were recorded digitally. The experiments were performed in triplicate and repeated three times. 

Enzyme-Linked Immunosorbent Assay

siRNA-transfected cells were seeded in six-well plates (3 × 10⁵ cells/well) and incubated at 37°C. After 48 hours, the cell culture supernatant was harvested, centrifuged to remove cellular debris, and stored at −80°C until used. VEGF protein secreted by RPE cells in the culture medium was measured by ELISA, according to the manufacturer's instruction (Dynatech Medica). Experiments were performed in triplicate and repeated three times. 

Flow Cytometry

RPE cell lines (1 × 10⁶) were seeded in each well of six-well plates with NS siRNA, and Robo1-siRNA transfection proceeded for 48 hours. The cells were harvested after siRNA transfection. Cell cycle analysis was performed (Cycletest Plus DNA Reagent Kit; BD Biosciences) according to the manufacturer's instructions. ²⁷ Samples were analyzed by flow cytometry (FACSCaliber; BD Biosciences). The experiments were performed in triplicate and repeated three times. 

Induction of PVR and Injection of Robo1-siRNA

Adult pigmented rabbits (2.0–2.5 kg) were anesthetized with an intramuscular injection of ketamine (50 mg/kg) and promethazine (25 mg/kg). The animals were managed according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The pupils were dilated with one drop of 1% atropine sulfate and tropicamide. The animals were divided into two groups. PVR induction was performed by injecting 2.5 × 10⁵ RPE cells in 0.1 mL balanced salt solution into the vitreous cavity through a 30-gauge needle, 4 mm posterior to the limbus. Before the injection, with a 30-gauge needle and a tuberculin syringe, an anterior chamber paracentesis was made and approximately 0.1 mL aqueous humor was drained. RPE cells for control animals (n = 10) had been transfected with chemically modified nonsilencing siRNA (0.3 μg) by using 6 μL transfection reagent (HiPerfect; Qiagen). The treatment group (n = 10) received RPE cells transfected with chemically modified Robo1-specific siRNA (0.3 μg) by 6 μL of the transfection reagent. The fundus of each rabbit was examined on days 7, 14, 21, and 28 in a masked manner. A retinal drawing was made at the time of each examination. PVR was classified in six stages according to published criteria ²⁸ (Table 2). At the end of the follow-up, the animals were killed by an intravenous overdose of anesthesia (pentobarbital 750 mg/kg). 

Table 2.

 

Criteria for Stages of PVR

Stage	Characteristics
0	Normal eye
1	Intravitreal membrane
2	Focal traction, localized vascular changes, hyphema, engorgement, dilation, blood vessel elevation
3	Localized detachment of medullary ray
4	Extensive retinal detachment, total medullary ray detachment, peripapillary retinal detachment
5	Total retinal detachment, retinal folds and holes

 Adapted from Fastenberg et al. ²⁸

Statistical Evaluation

All data are presented as the mean ± SD and have been evaluated for normality of distribution. Differences were evaluated with ANOVA followed by Student-Newman-Keuls test for multiple comparisons and the Student's t-test for pair-wise comparisons. P < 0.05 was considered significant. 

Results

Immunohistochemistry

Robo1 expression was detected in all the tissues of PDR and PVR patients. Robo1 expression was in the vessels (Fig. 1A) and fibrous-like tissues (Fig. 1E, arrow) in PDR membranes. Staining for CD31 (Fig. 1B) and GFAP (Fig. 1F) was also present in the PDR membranes. Colocalization of Robo1 and CD31 (Fig. 1D) was observed in the vessels. Robo1 was also detected in GFAP-labeled glial cells (Fig. 2D) and pan-cytokeratin-labeled RPE cells (Fig. 2H, arrows) in PVR membranes. The nucleus was stained with DAPI (Figs. 1C, 1G, 2C, 2G). We found no visible staining of CD31 in the PVR membranes; furthermore, GFAP and pan-cytokeratin were not detected in all PVR membranes (data not shown). 

Figure 1.

 

Micrographs showing immunohistochemical double staining of an epiretinal PDR membrane section for Robo1 (A, E), CD31 (B), and GFAP (F). Double staining revealed colocalization (yellow) of Robo1 and CD31 (D). Robo1 expression was detected in fibrous-like tissue (E, short arrow) as well as in GFAP-labeled glial cells (H). Nuclei were stained with DAPI (C, G). Bar, 100 μm.

Figure 2.

 

Micrographs showing immunohistochemical double staining of an epiretinal PVR membrane section for Robo1 (A, E), GFAP (B), and cytokeratin-pan (F). Double staining revealed that Robo1 protein was detected in glial cells (D) and pan-cytokeratin-labeled RPE cells (H, long arrow). Bar, 100 μm.

Knockdown of of Robo1 RNA and Protein Expression

The effectiveness of Robo1-siRNA transfection into human D407 RPE cells was determined by real-time RT-PCR and Western blot assay: Robo1 expression in RPE cells was significantly reduced at both the mRNA and protein levels (Fig. 3). Real-time RT-PCR demonstrated that siRNA specifically reduced Robo1 mRNA levels in RPE cells (P < 0.01; Fig. 3A). Immunoblot analysis for Robo1 revealed a single band of approximately 150 kDa (Fig. 3B), in the control siRNA-treated cells (NS) and the untreated (NC) cells, whereas in the Robo1-siRNA–treated group, only a faint band was detected, consistent with the strong depletion of Robo1 protein expression in Figure 3C. In contrast, there was no significant difference between NS and NC cells (P > 0.05). Immunocytochemical imaging of the expression of Robo1 in RPE cells further confirmed the knockdown of Robo1 (Fig. 4). The fluorescence intensity representing the expression of Robo1 in NC (Fig. 4A) and NS (Fig. 4B) were very strong, but that in Robo1-siRNA-transfected cells (Fig. 4C) was barely detectable. These results show that we have generated specific knockdown reagents that selectively target Robo1 in RPE cells. 

Figure 3.

 

Robo1-siRNA specifically knocked down Robo1 mRNA and protein level. (A) Robo1 expression in human RPE cells was significantly knocked down at the mRNA level, measured by real-time RT-PCR 48 hours after transfection. The protein expression of Robo1 in human RPE cells was measured by immunoblot analysis, normalized to β-actin expression in RPE cells. (B) A photograph of a representative Western blot of Robo1 in RPE. (C) The data of the relative Robo1 protein in the NC, NS, and Robo1-siRNA–treated cells. Data are expressed as the mean ± SD of results in at least three independent experiments. *P < 0.01 vs. control. The pixel intensity of NC was set to 100%.

Figure 4.

 

Immunocytochemical assays for Robo1 in human D407 RPE cells. The fluorescence, representing the expression of Robo1 in RPE cells, were very strong in NC (A) and NS (B) cells, but was barely detectable in the Robo1-siRNA–treated cells (C). Bar, 100 μm.

Effect of Transfection of Robo1-Specific siRNA on Cytotoxicity of RPE Cells

We performed an MTT assay to detect any changes in transfection-induced cell survival. There was no significant difference in cell viability in the NC, HF (cells treated with HiPerFect Transfection Reagent only; Qiagen), and NS groups (P > 0.05; Fig. 5C). We also determined cell apoptosis of NS (Fig. 5A) and Robo1-siRNA–treated group (Fig. 5B) to look for any transfection-induced cell death, but observed no significant difference between those two groups (P > 0.05). 

Figure 5.

 

Cytotoxicity induced by transfection of RPE cells. Cells were transiently transfected with transfection reagent (HiPerfect [HF]; Qiagen) and control siRNA (NS) and incubated for 48 hours. The effect of Robo1 on apoptosis of human RPE cells in the NS (A) and RPE Robo1-siRNA–treated (B) cells. The normal living cells (bottom left quadrants) showed low annexin V and propidium iodide staining. The early apoptotic cells (bottom right quadrants) showed high annexin V staining, but low propidium iodide staining. The late apoptotic cells (top right quadrants) showed intense annexin V and propidium iodide staining. The percentages of cells in the quadrants are indicated within the quadrant. Representative results of three separate experiments are shown. The extent of inhibition of cellular viability was measured by MTT assay (C). Data are the mean ± SD of results in at least three independent experiments.

Robo1 Regulation of Cell Attachment, Proliferation, Migration, and Spreading

As Robo1 is associated with cancer formation in some models, we wondered how it affects cell adhesion, proliferation, and migration, and we determined this by blocking Robo1 function with specific siRNA. In the cell adhesion assay, Robo1-siRNA treatment reduced the adhesive capacity of RPE cells after 1 hour compared with that of the NS group (P < 0.01; Fig. 6). Cell attachment in the NC and NS groups was not significantly different (P > 0.05). Blocking Robo1 reduced cell proliferation in RPE cells at day 3 compared with that in the NS cells (P < 0.01; Fig. 7), and the suppression peaked on day 4. 

Figure 6.

 

Effect of Robo1 on attachment of human D407 cells. Cell attachment was assessed after a 1-hour incubation and subsequent MTT testing. Data are the mean ± SD of results in at least three independent experiments. *P < 0.01 vs. the control. NS was set to 100%.

Figure 7.

 

Effect of Robo1 on the proliferation of human D407 RPE cells. Cell proliferation was measured with an MTT assay at days 1, 2, 3, 4, and 5. Data are the mean ± SD of results in at least three independent experiments. *P < 0.01 vs. the control.

Next, we explored the role of Robo1 in the migration of RPE cells by using a modified Boyden chamber. As shown in Figure 8, the mean number of migrating cells among the Robo1-siRNA–treated RPE cells (Fig. 8C) was significantly lower than the number of migrating NC (Fig. 8A) and NS (Fig. 8B) cells (P < 0.01), although the number of migrating cells in the NS and NC groups did not differ significantly (P > 0.05; Fig. 8D). Robo1-siRNA-transfected cells displayed different F-actin assembly and cell morphologies when compared with the NS group (Fig. 9). F-actin assembly in the Robo1-siRNA group was significantly disturbed after transfection. There was less lamellipodia and filopodia formation in Robo1-knockdown cells (Fig. 9A). However, the cells in the NS group displayed a well-organized actin skeleton, with fibers extending throughout the cytoplasm into the cell membrane (Fig. 9B). NS siRNA-treated cells also showed a higher degree of cell spreading when compared with Robo1-siRNA–treated cells (Fig. 10). 

Figure 8.

 

Effect of Robo1 on the migration of human D407 RPE cells. The migratory activity of the NC (A), NS (B), and R1 siRNA (C) groups was estimated based on the number of cells that had migrated through the filter of the chamber (D). Data are the mean ± SD of results in at least three independent experiments. *P < 0.01 vs. the control. NC was set to 100%.

Figure 9.

 

Effect of Robo1 on microfilament dynamics of RPE cells. Microfilament organization was assessed by immunofluorescence analysis with rhodamine-phalloidin. NS cells (B) displayed an elaborate network of precisely organized F-actin filaments. The F-actin filament architecture became significantly disturbed after transfection with less lamellipodia and filopodia formation in Robo1-siRNA–treated cells (A). Experiments were repeated three times. Bar, 50 μm.

Figure 10.

 

Effect of Robo1 on spreading of human RPE cells. Cell-spreading quantitation was performed by measuring the ratio of the cytoplasm to the nucleus area of cells in each field. NS cells (B) displayed a higher degree of cell spreading when compared with Robo1-siRNA–treated cells (A). Bar, 100 μm.

As shown in Figure 11, depletion of Robo1 levels in RPE cells caused a significant accumulation of cells in the G₀/G₁-phase of the cell cycle and a marked reduction in the accumulation of cells in the S-phase compared with the NS group (Figs. 11A, 11B). Of the Robo1-siRNA–treated RPE cells, 72.44% were in the G₁-phase, compared with only 61.17% in the NS group (P < 0.01). Of the Robo1-siRNA–treated RPE cells, 19.56% were in the S-phase, compared with 30.98% in the NS group (P < 0.01; Fig. 11C). These results show that knockdown of Robo1 induced cellular arrest in the G₁- and S-phases of the cell cycle. Because cyclin-D1 has an important role in the G₁- to-S transition in the cell cycle, we measured the expression of cyclin-D1 in our study. The expression of cyclin-D1 in Robo1 knockdown cells was inhibited compared with that in the NS control cells (Fig. 12). 

Figure 11.

 

Effect of Robo1 on the cell cycle of human RPE cells. (A) RPE NS and (B) RPE Robo1-siRNA. (C) The data of RPE cell cycle distribution of NS and Robo1-siRNA group cells. Data are the mean ± SD of results in at least three independent experiments. *P < 0.01 vs. the control.

Figure 12.

 

The effect of Robo1 on the expression of cyclin D1 of RPE cells. The expression of cyclin D1 was decreased after Robo1 knockdown.

Effect of Robo1 Depletion on VEGF Secretion in RPE Cells

It is reported that Slit-Robo signaling induces and maintains tumor angiogenesis in vivo model, independent of VEGF signaling. ²⁹ To explore whether Robo1-siRNA reduces VEGF activity in RPE cells, we used ELISA to examine the levels of VEGF secreted into the culture medium. We found that there was no significant difference in the level of VEGF secretion in the culture medium between the Robo1-siRNA–treated cells and the NS cells (P > 0.05; Fig. 13). 

Figure 13.

 

Effect of Robo1 on VEGF secretion of RPE cells. After transfection, the culture medium was harvested. VEGF released into the culture supernatant was measured by ELISA. There was no significant difference in the level of VEGF secretion in the culture medium between the NC, NS, and Robo1-siRNA–treated cells (P > 0.05). Data are the mean ± SD of results in three independent experiments. NS cells set to 100%.

Effect of Robo1-Specific siRNA on the Inhibition of Experimental PVR

Having shown that Robo1 plays a role the cell proliferation, adhesion, and migration of cultured RPE cells, we hypothesized that it would be involved in the development of PVR. To test this notion, we used an animal model of experimentally induced PVR. In both the Robo1-siRNA–treated group and the NS control group, the stage of PVR progressed steadily over time. A fibrous membrane without tractional retinal detachment was formed on the retina in all eyes in both the treated and the control groups at 3 days after injection. On day 7, four (40%) eyes in the control group showed stage 3 PVR, which was not seen in the treated eyes, but the difference was not significant (P > 0.05). By day 28, six (60%) eyes in the control group had stage 5 PVR, and the other four (40%) eyes displayed stage 4. In eyes treated with Robo1-siRNA, no eye developed stage 5, three (30%) eyes showed stage 4, four (40%) eyes displayed stage 3, and the remaining two eyes did not have a retinal detachment (P < 0.01, Figs. 14 and 15). 

Figure 14.

 

Twenty-eight days after injection of NS and Robo1-siRNA–treated RPE cells. Fundus photographs of the normal retina (A) and the retinal detachment with fibrous proliferation (B). Gross pathology of (C) a Robo1-siRNA–treated eye or (D) an NS siRNA–treated eye. The NS siRNA–treated eye showed extensive formation of membrane, whereas the Robo1-siRNA–treated eye showed a lesser degree of fiber proliferation.

Figure 15.

 

The stages of PVR in each group. The stage of PVR was evaluated at the indicated time points for each rabbit injected with Robo1-siRNA or NS. Day 7 after injection (A) and day 28 after injection (B). R1 siRNA: Robo1-siRNA–treated cells.

Discussion

In other studies, histopathology and immunohistochemistry have demonstrated that membranes in proliferative retinopathy consist of complex fibrocellular tissue, mainly formed of RPE, macrophages, glial cells, fibroblast-like cells, and various amounts of extracellular matrix components and vascular elements. ^28,30 The mechanisms that regulate fibrosis and angiogenesis in PVR disorders are incompletely characterized. 

In recent years, four groups of neuronal guidance molecules—namely, the ephrin/Eph receptor, delta/Notch receptor, netrin/Unc receptor, and Slit/Roundabout protein/receptor—have been found in endothelial cells. ^31,32 Previous studies of Robo1 have mainly concentrated on organ development and tumor tissues and until now, there has been no report of Robo1 expression in human eye tissue. In our previous study, we found Robo1 to be expressed in neural tissues including nerve fiber layer, ganglion cell layer, inner and outer plexiform layers, and retinal vessels during mouse retinal development. ²² These findings are consistent with the results of the present study, as we detected doublestaining of CD31/GFAP and Robo1 in the proliferative membranes. We also, for the first time, found Robo1 expression in both human RPE cells in PVR membranes in vitro and cultured RPE cells in vivo. Proliferative retinal membranes are heterogeneous, with their content varying with the type and age of the membrane and probably other variables. RPE and glial cells can undergo dedifferentiation, with the result that they may appear as fibroblasts and change their expression of cytokeratins or GFAP. ³³ This phenomenon may in part explain why GFAP and cytokeratin were not detected in all the proliferative membranes. 

In healthy eyes, RPE cells form a polarized monolayer adjacent to the photoreceptors and are involved in diverse activities that are essential to retinal homeostasis and visual function. ³⁴ In normal circumstances, RPE cells are stationary and mitotically feeble. In contrast, they become activated and mobilized in PVR. Indeed, dissemination of migratory, proliferating RPE cells from their normal site on Bruch's membrane to multiple loci on the detached neuroretina is thought to be a key pathologic event in the genesis of the complex epiretinal membranes associated with the development of PVR. ^34–38  

RNA interference (RNAi) is a recently developed technique for silencing proteins in a sequence-specific manner by inhibiting mRNA and consequently reducing protein expression. The high efficiency and specificity of RNAi has made it a powerful and widely used tool for gene therapy. The functional mediator of RNAi is a short double-stranded RNA (dsRNA) oligonucleotide called siRNA. ³⁹ In the present study, the expression of Robo1 was knocked down in the human D407 RPE cell line with an siRNA, and we observed significant differences between the Robo1-specific siRNA–treated group and the control siRNA-treated group, which confirmed the specificity of Robo1-siRNA. We performed MTT assays to investigate the cytotoxicity of transfection and found that there was no significant difference in cell viability among NC, HF, and NS groups. To further exclude the cell cytotoxicity, we also used flow cytometry to detect cell apoptosis in the untreated control and Robo1-siRNA–treated groups and found no difference between the two groups. All these experimental results indicate that the siRNA application performed well in our study. 

As cell attachment and spreading were key cellular events in early PVR, it is reasonable to predict that antiadhesive and antispreading therapy hold promise for the treatment of PVR. In our study, downregulation of Robo1 inhibited cell spreading and attachment. 

Cell migration is another important process in the development of PVR. Without migration, cells would not gain access to the ectopic sites and form membranes. Hence, it is reasonable that inhibition of cell migration would be an ideal treatment of proliferative retinopathy. In our study, downregulation of Robo1 markedly inhibited the migration of RPE cells, consistent with its effects on the migration of human umbilical vein endothelial cells (HUVECs), but opposite its role as a repellent and inhibitor in neuronal and leukocyte cells. ^13,20,40,41 We also found that the formation of the actin cytoskeleton and motility structures in Robo1 knockdown cells was severely affected. Our data suggest that targeting Robo1 affects microfilament assembly, resulting in decreased RPE to attachment, spread, and migration. Robo1 may have a preset effect on migration according to the cell type and may be controlled by the expression of other Roundabout members subsequently. ⁴⁰  

Cell proliferation is a late-occurring process in the development of proliferative retinal diseases. In this study, we investigated whether Robo1 affects the proliferation of RPE cells. The inhibitory effect on cell growth was apparent in the Robo1-siRNA–treated cells after 48 hours and induced cell cycle arrest in the G₁-phase through the reduction of cyclin-D1 expression in the RPE cells. Targeting the expression of Robo1 may be a novel approach for antiangiogenic therapy. 

Growth factors including bFGF and VEGF regulate proliferation, migration, and invasion of the RPE cells and are expressed in PVR membranes. ⁴² The failure of treatment targeting Robo1 to alter the expression profile of VEGF in vitro implies that Robo1 signaling inhibits RPE proliferation and invasion, independent of VEGF signaling. 

Experimental models of PVR were developed to evaluate intraocular proliferation. ^43–45 Our results suggest that treatment with the Robo1-specific siRNA not only inhibits cell proliferation in vitro but also delays the onset of PVR in a rabbit model. As PVR is an overreaction to injury repair in eyes, including the inflammatory stage, proliferative stage, and chronic scarring (cicatrix) stage, it is important to choose different therapies at different stages of disease. In the PVR model induced by fibroblasts, cell proliferation often reaches a peak on day 3 after injection. ⁴⁶ Although, in the PVR model induced by macrophages, the inflammatory stage persists for 1 week after injection, cell proliferation reaches a peak nearly 2 weeks after injection. ⁴⁷ Our results showed that the inhibitory effect of Robo1-siRNA was sustained at least for 5 days, and inhibition of RPE cell proliferation could prevent the later stage of disease, slow down the progression of lesions, and improve the prognosis. As we know, although D407 cells cloned from human RPE cells possess many characteristics such as phagocytic activity similar to human RPE cells, they may not accurately reflect primary RPE activities. Further studies are needed to confirm the role of Robo1 in the proliferation of RPE cells. 

In summary, our results for the first time confirm the Robo1 expression pattern in the membranes of PDR and PVR. Robo1 may play a role in cell proliferation, migration, and attachment, independent of VEGF signaling, and thus Robo1 may be a proper target for treatment of retinal proliferative diseases. Further studies to characterize the precise physiologic function of Robo1 in proliferative retinopathy are currently under way in our laboratory. 

Footnotes

 Supported by National Basic Research Program of China 973 Program, 2005CB724307; Research Fund for the Doctoral Program of Higher Education of China Program 20090001110083; and People's Hospital Research and Development Fund Program RDB2009-43.

Footnotes

 Disclosure: L. Huang, None; Y. Xu, None; W. Yu, None; Y. Li, None; L. Chu, None; J. Dong, None; X. Li, None

The authors thank Martine J. Jager for assistance in writing the article.