Proteomic and Transcriptomic Analyses of Retinal Pigment Epithelial Cells Exposed to REF-1/TFPI-2

Abstract

purpose. The authors previously reported a growth-promoting factor, REF-1/TFPI-2, that is specific to retinal pigment epithelial (RPE) cells. The purpose of this study was to determine the genes and proteins of human RPE cells that are altered by exposure to TFPI-2.

methods. Human primary RPE cells were cultured with or without TFPI-2. Cell extracts and isolated RNA were subjected to proteomic and transcriptomic analyses, respectively. Proteins were separated by two-dimensional gel electrophoresis followed by gel staining and ion spray tandem mass spectrometry analyses. Transcriptomic analysis was performed using a DNA microarray to detect 27,868 gene expressions.

results. Proteomic analysis revealed c-Myc binding proteins and ribosomal proteins L11 preferentially induced by TFPI-2 in human RPE cells. Transcriptomic analysis detected 10,773 of 33,096 probes in the TFPI-2 treated samples, whereas only 2186 probes were detected in the nontreated samples. Among the genes up-regulated by TFPI-2 at the protein level were c-myc, Mdm2, transcription factor E2F3, retinoblastoma binding protein, and the p21 gene, which is associated with the c-myc binding protein and ribosomal protein L11.

conclusions. The mechanisms by which TFPI-2 promotes the proliferation of RPE cells may be associated with augmented c-myc synthesis and the activation of E2F in the retinoblastoma protein (Rb)/E2F pathway at the G1 phase of the RPE cells. Activation of ribosomal protein L11 and the Mdm2 complex of the p53 pathway may be counterbalanced by the hyperproliferative conditions.

Retinal pigment epithelial (RPE) cells play important roles in maintaining the homeostasis of the retina. RPE cells, located between the sensory retina and the choroidal blood supply, form a diffusion barrier controlling access to the subretinal space, with the RPE membrane regulating the transport of proteins and controlling the hydration and ionic composition of the subretinal space. The sensitivity and viability of the photoreceptors thus depend on RPE-catalyzed transport activity. Proteins in the RPE cells that function in ionic, sugar, peptide, and water transport have been identified. ¹ Damage to RPE cells generally leads to degeneration of the neural retina, as occurs in retinitis pigmentosa and age-related macular degeneration. Transplantation of the healthy retinal pigment cells or embryonic stem cells differentiating into RPE cells would be an ideal therapeutic approach to treat such diseases, and such attempts have been made. ²  

An alternative approach to treat these retinal diseases would be the use of a growth factor that promotes proliferation of the remaining RPE cells in a damaged retina or one that stimulates the regeneration of damaged RPE cells. To find such factor(s), the proteins expressed in human fibroblast cells were fractionated and assayed, leading to the isolation of RPE cell factor-1 (REF-1), which selectively promoted the proliferation of primary human RPE cells. ³  

Subsequently, the cDNA of REF-1 was cloned using information from the N-terminal amino acid sequences, which was identical with the tissue factor pathway inhibitor-2 (TFPI-2). ³ Earlier studies have shown that TFPI-2 is a Kunitz-type serine protease inhibitor ^{4 5 6} involved in the regulation of extrinsic blood coagulation ^{4 7} and in the proliferation, invasion, and metastasis of various types of malignant cells. ^{4 8 9 10 11 12 13} Extensive studies on the physiological roles of TEPI-2 have revealed that the ERK/MAPK pathway ¹³ may be associated with the up-regulation of the TFPI-2 gene and that DNA methylation ^{9 10} in certain tumor cell lines may be related to the downregulation of the TEPI-2 gene. When TFPI-2 is added to the culture medium of vascular smooth muscle cells, it promotes cell proliferation. ¹⁴  

Our initial finding that TFPI-2 enhanced RPE proliferation prompted us to question how this was achieved. We applied proteomic and transcriptomic analyses to screen the changes in the expression of the RNAs and proteins in RPE cells and will show that the proliferation promoting activity of TFPI-2 on RPE cells is associated with the regulation of an oncogene product, c-myc, and representative cancer repressor proteins retinoblastoma protein (Rb)/E2F and p53. 

Materials and Methods

TFPI-2 Treatment of Human RPE Cell Culture

Human primary RPE cells (passage 5) were seeded at a density of 2.5 × 10⁴ cells/0.5 mL per well in 24-well plastic plates (BD Biosciences, Franklin Lakes, NJ) with Dulbecco modified MEM (DMEM; Invitrogen Japan, Tokyo, Japan) containing 15% fetal calf serum (FCS, Invitrogen). TFPI-2 was added to 20 wells with the RPE cells at 10 ng/mL concentrations and was incubated at 37°C for 24 hours for the proteomic samples, and for 6 hours, 12 hours, and 24 hours for the transcriptomic samples. An equal amount of saline was added to 20 wells containing RPE cells for controls. TFPI-2 was donated by Toray Industries, Inc., Tokyo, Japan. 

Protein Sample Preparation

To isolate whole cellular protein extracts from cultured RPE cells, the cells were rinsed 3 times with 1× PBS (pH 7.4) and were lysed in a denaturing lysis buffer containing 7 M urea, 2 M thiourea, 4% CHAPS, 40 mM Tris, 0.2% purifier (Bio-Lyte, pH range 3–10; Bio-Rad, Hercules, CA), and 50 mM dithiothreitol (DTT). The collected lysate was then centrifuged at 14,000g for 15 minutes at 4°C. Proteins in the supernatant were repeatedly concentrated and precipitated and finally desalinated (Readyprep 2-D Cleanup kit; Bio-Rad). The protein concentration in the RPE samples was determined by a modified Lowry method adapted for use with the lysis buffer. 

Two-Dimensional Electrophoresis

Protein samples were separated by a two-dimensional electrophoresis method. A 300-μg protein sample was loaded on immobilized pH gradient (IPG) strips (pH 3–10, 7 cm; pH 4–7, 17 cm; Bio-Rad) by in-gel rehydration at 20°C overnight. For the 7-cm strip, isoelectric focusing (IEF) was used for the first dimension at an initial voltage of 250 V for 15 minutes, increased to 4000 V for 2 hours, and held until 20,000 V/h was reached. For the 17-cm strip, the initial voltage was set at 250 V, as for the 7-cm strip. Then the voltage was increased to 10,000 V for 3 hours and was held until 60,000 V/h was reached. Immediately after IEF, the IPG strips were equilibrated for 20 minutes in buffer containing 6 M urea, 2% SDS, 0.375 M Tris (pH 8.8), and 20% glycerol under a reduced condition with 2% DTT (Bio-Rad), followed by another incubation for 10 minutes in the same buffer under alkylating conditions with 2.5% iodoacetamide (Bio-Rad). ¹⁵  

Equilibrated IPG strips were then electophoresed by SDS-PAGE for the second dimension. Images of the chemiluminescent signals were captured and merged with those of protein spots made visible by protein gel stain (Sypro Ruby; Bio-Rad), and the spots corresponding to the immunoreactivity were cut out. To test reproducibility, the experiment was performed twice. 

Protein Identification by Mass Spectrometry

Excised gel pieces were rinsed with water and then with acetonitrile and were completely dried for the reduction-alkylation step. They were incubated with 10 mM DTT in 100 mM ammonium bicarbonate for 45 minutes at 56°C, then with 55 mM iodoacetamide in 100 mM ammonium bicarbonate for 30 minutes at room temperature in the dark. The supernatant was removed, and the washing procedure was repeated three times. Finally, the gel pieces were again completely dried before trypsin digestion and were rehydrated in a solution of trypsin (12.5 ng/μL; Promega, Madison, WI) in 50 mM ammonium bicarbonate. The digestion was continued for 16 hours at 37°C, and the extraction step was performed once with 25 mM ammonium bicarbonate, then twice with 5% formic acid, and finally with water. After resuspension in 40 μL solution of aqueous 0.1% trifluoroacetic acid/2% acetonitrile, the samples were analyzed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). For analysis by LC-MS/MS, the tryptic digests were injected by an automatic sampler (HTS-PAL, CTC Analytics, Zwingen, Switzerland) onto a 0.2 × 50-mm capillary reversed-phase column (Magic C18, 3 μm; Michrom BioResources, Inc., Auburn, CA) using an HPLC (Paradigm MS4; Michrom BioResources). Peptides were eluted with a gradient (95% solvent A consisting of 98% H₂O/2% acetonitrile/0.1% formic acid)/5% solvent B (10% H₂O/90% acetonitrile/0.1% formic acid; 0 minute)/35% solvent A/65% solvent B (20 minutes)/5% solvent A/95% solvent B (21 minutes)/5% solvent A/95% solvent B (23 minutes)/95% solvent A/5% solvent B (30 minutes) for 30 minutes at a flow rate of 1.5 μL/min. Peptides were eluted directly into an ion trap mass spectrometer (ESI; Finnigan LTQ; Thermo Electron Corporation, Waltham, MA) capable of data-dependent acquisition. Each full MS scan was followed by an MS/MS scan of the most intense peak in the full MS spectrum with the dynamic exclusion enabled to allow detection of less-abundant peptide ions. Mass spectrometric scan events and HPLC solvent gradients were controlled with the use of a computer program (Paradigm Home; Michrom BioResources). 

Total RNA Isolation from RPE Cells

Total RNA was isolated from the cultured RPE cells after 6 hours, 12 hours, and 24 hours with TFPI-2 using a total RNA isolation kit (RNA-Bee-RNA Isolation Reagent; Tel-Test, Friendswood, TX). Total RNA samples were treated with RNase-free DNase (Roche Diagnostics Japan) to minimize genomic DNA contamination. 

DNA Microarray Analysis

DNA microarray analysis was performed (AB1700 Chemiluminescent Microarray Analyzer; Applied Biosystems, Foster City, CA). The survey array used (Human Genome Survey Array; Applied Biosystems) contained 33,096 60-mer oligonucleotide probes representing a set of 27,868 individual human genes and more than 1000 control probes. Sequences used for the microarray probe were obtained from curated transcripts (Celera Genomics Human Genome Database), RefSeq transcripts that had been structurally curated from the LocusLink public database, high-quality cDNA sequences from the Mammalian Gene Collection (MGC; http://mgc.nci.nih.gov), and transcripts that were experimentally validated (Applied Biosystems). The 60-mer oligo probes were synthesized using standard phosphoramidite chemistry and solid-phase synthesis and underwent quality control by mass spectrometry. The probes were deposited and covalently bound to a derivatized nylon substrate (2.5 × 3 inches) that was backed by a glass slide by contact spotting with a feature diameter of 180 μm and more than 45 μm between each feature. A 24-mer oligo internal control probe (ICP) was cospotted at every feature with 60-mer gene expression probe on the microarray. Digoxigenin-UTP labeled cRNA was generated and linearly amplified from 1 μg total RNA (Chemiluminescent RT-IVT Labeling Kit, version 2.0; Applied Biosystems) according to the manufacturer’s protocol. Array hybridization (two arrays per sample), chemiluminescence detection, image acquisition, and analysis were performed (Chemiluminescence Detection Kit and AB1700 Chemiluminescent Microarray Analyzer; Applied Biosystems) according to the manufacturer’s protocol. 

Briefly, each microarray was first prehybridized at 55°C for 1 hour in hybridization buffer with blocking reagent. Sixteen micrograms labeled cRNA targets were first fragmented into 100 to 400 bases by incubation with fragmentation buffer at 60°C for 30 minutes, mixed with internal control target (ICT; 24-mer oligo labeled with LIZR fluorescent dye), and hybridized to each prehybrid microarray in 1.5 mL vol at 55°C for 16 hours. After hybridization, the arrays were washed with hybridization wash buffer and chemiluminescence rinse buffer. Enhanced chemiluminescent signals were generated by first incubating the arrays with anti–digoxigenin alkaline phosphatase and enhanced with chemiluminescence enhancing solution and chemiluminescence substrate. 

Images were collected from each microarray using the 1700 analyzer equipped with a high-resolution, large-format CCD camera, including 2 “short” chemiluminescent images (5-second exposure length each) and 2 “long” chemiluminescent images (25-second exposure length each) for gene expression analysis, two fluorescent images for feature finding and spot normalization, and two quality control images for spectrum cross-talk correction. Images were quantified, corrected for background and spot, and spatially normalized. 

Data Analysis

MS data were identified with the use of a protein search program (BioWorks 3.2; Thermo Electron Corporation, Waltham, MA). For protein database searches, the same program was used to create centroid peak lists from the raw spectra. These peak lists were then submitted for database searching (BioWorks). The identity of the samples was searched from databases (nrNCBI [www.ncbi.nlm.nih.gov]) that extracted proteins and were restructured; search terms included human and Homo sapiens. Differentially expressed proteins were further analyzed for related genes and proteins using natural language processing software (Pubgene database; PubGene Inc., Boston, MA) and data mining software of gene expression (OmniViz; OmniViz, Inc., Maynard, MA). 

Results

Proteome Analysis of RPE Cells Treated with TFPI-2

To determine the mechanisms responsible for the proliferation-promoting activity of TFPI-2 on RPE cells, protein synthesis and RNA expression were determined before and after TFPI-2 exposure. Differentially expressed proteins in the primary human RPE cells in response to TFPI-2 were identified by two-dimensional electrophoresis (Fig. 1) . Samples were initially separated using IPG at a pH range of 3 to 10 to observe the full distribution of protein spots. The pH range was then narrowed to 4 to 7 to obtain higher resolution for spot picking. Consequently, approximately 480 spots were identified in the whole gel. We then focused on molecular weight less than 25 kDa, which is easy to check for changes. Ten spots considered differentially expressed in the two-dimensional gel were collected and subjected to LC-MS/MS analysis. Among the identified proteins, ribosomal protein L11 (RPL11; Fig. 1-1 ) and c-Myc binding protein (MYCBP; Fig. 1-3 ), known for regulating cell proliferation, were identified. ¹⁶ These two proteins, identified by LC-MS/MS analysis and data analysis software (Bioworks 3.2), were consistent with those estimated from the results of two-dimensional electrophoresis (Table 1) . 

Transcriptomic Analysis of RPE Cells Treated with TFPI-2

The expression of 8134 genes in RPE cells was analyzed using DNA microarray with and without TFPI-2 exposure for 6 hours, 12 hours, and 24 hours. Signal normalization was performed for six independent DNA microarray chips according to the manufacturer’s protocol. Genes differentially expressed by more than threefold were considered significant and were selected for further analysis. Among the 33,096 possible probes, 10,773 probes were detected in the RPE cells incubated with TFPI-2, whereas only 2186 probes were detected without TFPI-2. Based on expression levels at the three time points (6 hours, 12 hours, and 24 hours), the time-dependent expression pattern of each gene was calculated and clustered with other genes with similar expression patterns using data mining software (OmniViz). Data analysis resulted in 38 clusters of genes that either increased or decreased their expression levels by more than twofold after TFPI-2 (Fig. 2) . Nineteen genes were upregulated in 5 clusters, 108 genes in 16 clusters, and 717 genes in 22 clusters at 6 hours, 12 hours, and 24 hours, respectively. For downregulated genes, 30 genes in 16 clusters, 119 genes in 19 clusters, and 3 genes in 19 clusters were observed after 6 hours, 12 hours, and 24 hours, respectively. Transcriptomic analysis revealed significantly more genes differentially expressed at the transcriptional level than at the proteome level. 

Discussion

Proteins and genes whose expression was upregulated or downregulated after exposure to TFPI-2 were analyzed in human RPE cells to study the proteomic and transcriptomic changes. Protein and gene expression profiles for human RPE cells have been reported by West et al., ¹⁷ who identified 278 proteins, and Cai et al., ¹⁸ who reported 5580 ± 84 genes expressed in adult human RPE and ARPE19 cell lines using a DNA chip with 12,600 probes (Human U95Av2; Affymetrix, Santa Clara, CA). Our study showed changes in the expression of 8134 of 27,868 genes. DNA microarray analyses were simultaneously performed at three time points (6 hours, 12 hours, and 24 hours) to monitor the course of expression of the possible 27,868 genes in human RPE cells exposed and not exposed to TFPI-2. This study was conducted at the translational and the transcriptional levels to complement the disadvantages of each method. 

Raw gene expression data were further analyzed with data mining software (OmniViz) to obtain an overall picture of the transcriptional changes induced by TFPI-2 in human primary RPE cells. Genes whose expressions were changed by more than twofold were clustered into 38 groups showing a change of expression at each time point (Fig. 2) . The number of genes upregulated at each time point was considerably higher than the number that was downregulated. A small number of genes was triggered by TFPI-2 treatment at 6 hours, before the major changes occurred at 24 hours. Among the initially upregulated genes were reticulon 4 interacting protein 1, phospholipase C, delta 1, granzyme M (lymphocyte met-ase 1; GZMM), and mitochondrial ribosomal protein L41 (MRPL41). 

Proteomics analysis simultaneously performed at 24 hours identified two differentially expressed proteins, the c-myc binding protein (MYCBP) and the ribosomal protein L11 (RPL11). MYCBP and RPL11 (Fig. 3)are well known to regulate cell cycling through the Rb/E2F pathway and the p53 pathway, respectively. MYCBP stimulates c-myc transcription through the retinoblastoma protein (Rb)/E2F pathway (see 4 Fig. 5 ). Sears et al. ¹⁹ reported that activation of Myc increased the signal transduction of the cyclin D/cdk4 and cyclin E/cdk2 pathways. Activation of these pathways inactivates Rb after phosphorylation and E2F dissociation, which then promotes RPE cells to go into the S-phase of the cell cycle. The twofold transcriptional increase of Rb and E2F3 in TFPI-2 exposed cells compared with control at 24 hours supports this hypothesis (Figs. 4C 4F) . 

Concomitantly, the expressions of Rb and Mdm2 were upregulated twofold in growth-stimulated cells compared with control cells. Because Rb is associated with the negative regulation of the G₁-phase of the cell cycle, the enhanced expression of Mdm2 might have been involved in the augmented degradation of Rb through the ubiquitin/proteasome-dependent pathway. Recently, Uchida et al. ²⁰ suggested that Mdm2 regulates the function of RB through the ubiquitin-dependent degradation of RB. 

The Rb gene was the first identified tumor-suppressor gene, ²¹ and it was recognized as a central component of a signaling pathway that controlled cell proliferation. Specifically, the D-type G₁ cyclins, together with their associated cyclin-dependent kinases (CKDs) Cdk4 and Cdk6, initiated the phosphorylation of Rb and Rb family members, inactivating their capacity to interact with the E2F transcription factors (Fig. 5) . ¹⁹ This phosphorylation leads to an accumulation of E2F1, E2F2, and E2F3a, which activate the transcription of a large number of genes essential for DNA replication and further cell cycle progression. ^{22 23 24 25 26} Among the E2F targets are genes encoding a second class of G₁ cyclins, cyclin E, and the associated kinase Cdk2 (Fig. 5) . ¹⁹ The activation of cyclin E/Cdk2 kinase activity by E2F leads to further phosphorylation and inactivation of Rb, further enhancing E2F activity and increasing the accumulation of cyclin E/Cdk2 (Fig. 5) . ¹⁹ This feedback loop, which leads to a continual inactivation of Rb independent of the action of cyclin D/Cdk4—defined as a junction in cell proliferation response when passaged through the cell cycle—becomes growth factor independent. ^{25 26} The activity of the G₁ Cdks is negatively regulated by a family of cyclin-dependent kinase inhibitors (CKIs), including p21 ^WAF1 , p27 ^Kip1 , and the p16 ^INK4a family. ²⁷ The three upregulated E2Fs associate exclusively with Rb and appear to play a positive role in cell cycle progression. ¹⁹  

RPL11 binds the mouse double-minute 2 (Mdm2 is the mouse homologue of Hdm2 in humans) protein with other ribosomal proteins (L23 and L5) to form a complex to inhibit ubiquitin-dependent degradation of p53. ^{28 29 30} The RPL11 protein is expressed in ARPE-19 cells. ³¹ Inhibition of p53 degradation leads to p21 signaling, which participates in the G₁ arrest of the cell cycle but also negatively regulates cell proliferation (Fig. 5) . ^{30 32 33 34} In support of this hypothesis, p21 transcription was increased by twofold after 24 hours by TFPI-2. 

The p53 gene mediates a major tumor-suppression pathway in mammalian cells and is frequently altered in human tumors. ³⁰ Its function is kept at a low level during normal cell growth and is activated in response to various cellular stresses by acting as a sequence-specific transcription factor. ³⁰ The p53 protein induces cell cycle arrest or apoptosis. ³⁰  

Shinoda et al. ¹⁴ reported cell growth proliferation of vascular smooth muscle endothelial cells by a purified mitogenic substance from human umbilical vein endothelial cells, later identified as TFPI-2. These authors showed the rapid activation of mitogen-activated protein kinase (MAPK) by TFPI-2 and the induced activation of proto-oncogene c-fos mRNA in smooth muscle cells. ¹⁴ They concluded that c-fos activation was initiated by MAPK based on MAPK inhibitor PD098059 suppression. 

In conclusion, the results of proteomic and transcriptomic analyses suggest that the proliferation of RPE cells induced by TFPI-2 is regulated through the Rb/E2F, p53, and Ras/Raf/MAPK pathways. We and others ^{3 35} have reported a transcript of TFPI-2 in the mRNA of RPE cells. It is now reasonable to expect that RPE cells are able to self-proliferate by generating TFPI-2. Additional studies are needed to determine whether TFPI-2 can act as such an autocrine factor and can be modified for future treatment of the dry-type age-related macular degeneration and of retinitis pigmentosa. 

 Supported in part by a grant-in-aid from the policy-based Medical Services Foundation.

 Submitted for publication April 18, 2006; revised July 17, 2006; accepted December 4, 2006.

 Disclosure: M. Shibuya, None; H. Okamoto, None; T. Nozawa, AMR Inc. (F); J. Utsumi, Toray Industries, Inc. (F); V.N. Reddy, None; H. Echizen, None; Y. Tanaka, None; T. Iwata, None

 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.

 

Figure 1.

 

Two-dimensional gel electrophoresis of human RPE cells culture with (A) and without (B) TFPI-2. Spots corresponding to proteins whose expression is dependent on the presence of TFPI-2 in the culture medium are indicated by the arrows (insets). Proteins were detected by SYPRO Ruby staining. Spots corresponding to the differentially expressed proteins indicated by arrows (1 vs. 2 and 3 vs. 4) were subsequently subject to the LC-MS/MS analysis so that proteins could be identified.

Table 1.

 

Two-Dimensional Gel Spots Identified by Mass Spectrometry

Protein	Number of AA	Peptide Residues	Identified Peptide from Database	MW	Score	Accession Number
c-Myc binding protein	167	108–117	TAEDAKDFFK	18642.6	10.13	1731809
Ribosomal protein L11	177	88–94	VREYELR	20125.1	20.21	14719845

Figure 2.

 

Differentially expressed genes detected by DNA array are plotted as clusters. Differentially expressed genes whose expression level was increased by more than twofold (A–C) or was reduced by more than 0.5-fold (D–F) in RPE cells treated with TFPI-2 at incubation times of 6 hours, 12 hours, and 24 hours compared with the control cells are shown. Expression profile analysis revealed different gene expression patterns at each incubation time.

Figure 3.

 

Time course of gene expression for TFPI-2 (A), c-myc binding protein (B), and ribosomal protein L11 (C) in the cultured human RPE cells after exposure to TFPI-2.

Figure 5.

 

Hypothetical network of various genes and proteins associated with the growth-promoting effect of TFPI-2 on the human RPE cells. Arrows: stimulatory signals. Straight and dotted lines: inhibitory effects.

Figure 4.

 

Time courses of protein expression patterns for c-myc (A), Mdm2 (B), retinoblastoma binding protein 4 (C), p-53 inducible protein (D), p21 (E), and transcription factor E2F3 in the cultured human RPE cells after exposure to TFPI-2.