The expression of native and cultured RPE grown on different matrices

Jane Tian, Kazuki Ishibashi and James T. Handa

Michael Panitch Macular Degeneration Research Laboratory, Wilmer Eye Institute, Johns Hopkins Medical Institutes, Baltimore, Maryland 21287


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
The purpose of this work was to determine the expression profiles of retinal pigment epithelial (RPE) cells grown on different matrices and to assess the degree of culture-induced artifact by comparing the profiles to native RPE. Visually confluent ARPE-19 cells were grown on plastic, Matrigel, collagen I, collagen IV, laminin, and fibronectin for 1 wk, and serum was withdrawn for 3 days. Morphologically normal, macular RPE cells were laser-capture microdissected from three human eye globes. Total RNA was extracted from 5,000 cells and reverse transcribed, and radiolabeled cDNA probes were hybridized to an array containing 4,325 known genes. Arrays were assessed by cluster analysis and significance analysis of microarrays (SAM). Real-time RT-PCR was used to validate differentially expressed genes. Despite similar morphology, ARPE-19 demonstrated different expression profiles when grown on different matrices. Cluster analysis showed that cells grown on collagen IV, laminin, and fibronectin had similar profiles that were distinct from cells grown on collagen I. Cells grown on plastic clustered closest to native RPE. This expression pattern was confirmed with supervised cluster analyses. The number of differentially expressed genes, function of differentially expressed genes, and profile of expressed and unexpressed genes suggest that the overall expression profile of cultured cells is significantly different from native RPE. RPE cells grown on collagen IV, laminin, and fibronectin have profiles more similar than cells grown on plastic, Matrigel, or collagen I. The overall mRNA phenotype, however, is different from morphologically normal, native macular RPE.

ARPE-19; Bruch’s membrane; laser-capture microdissection


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
THE BASEMENT MEMBRANE is a dynamic, supramolecular scaffold of structural molecules. Interaction of the basement membrane with transmembrane molecules provides environmental information that can modulate cell behavior by inducing intracellular signaling. The most abundant basement membrane components are collagen IV and laminin, which form a polymeric network (3). Interaction of the cell with specific components of the basement membrane is an important factor that contributes to a cell’s phenotype (24).

Our laboratory studies how basement membrane changes influence the retinal pigment epithelial (RPE) cell mRNA phenotype. Alteration in basement membrane composition, such as the deposition of lipids, formation of advanced glycation end products, or oxidative stress-related changes, could influence the RPE cell. Likewise, an increase in matrix proteins such as fibronectin and collagen I has been observed with aging in the RPE basement membrane (20, 21), which could also change the cellular phenotype. Before the impact of these alterations can be understood, the contribution of normal matrix proteins on the RPE mRNA phenotype must first be established. A number of different matrices, such as Matrigel, collagen IV, or laminin, induce a morphologic phenotype that simulates cells in vivo (10, 25) and are a logical starting point to assess the effect of matrix proteins on the mRNA phenotype of cells.

We hypothesize that normal constituents of the basement membrane induce an mRNA phenotype of cultured RPE cells that approximates native cells. Culture conditions, however, can be a potential confounding influence the cell’s phenotype. The effect of culture conditions on the mRNA phenotype is essentially unknown. In this study, we measured the mRNA phenotype of RPE cells grown on different matrices to determine the matrix condition that most closely simulated native RPE, and we assessed the degree of artifact induced by in vitro conditions from this analysis.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

Cell culture.
The routine maintenance of the established, nonimmortalized human RPE cell line ARPE-19 has been previously described (14). For experiments, 35-mm culture dishes were coated with different matrix preparations. Human fibronectin (25 µg; BD Biosciences, Bedford MA) was incubated at room temperature for 1 h, human collagen, type I (10 µg in 10 mM acetic acid; Calbiochem, San Diego, CA), was incubated at room temperature for 2 h, mouse collagen IV (50 µg in 0.05 N HCl; BD Biosciences) was incubated at room temperature for 1 h, and mouse laminin (50 µg in serum-free DMEM; BD Biosciences) was incubated for 1 h, all as recommended by the manufacturer. Growth factor reduced Matrigel (1:16 dilution in DMEM; BD Biosciences) was incubated for 1 h at 37°C using our previously published protocol (15). ARPE-19 cells were seeded at 100,000 cells/cm2 in 35-mm dishes in DMEM with 15 mM HEPES buffer (BioWhittaker, Walkersville, MD) and 10% fetal bovine serum (FBS; UBI Upstate, Lake Placid, NY) and 2 mM L-glutamine solution (GIBCO-BRL, Life Technologies, Gaithersburg, MD) on the chosen matrix at 37°C for 1 wk, and serum was withdrawn in DMEM+1% bovine serum albumin (BSA) for 3 days to make ARPE-19 cells quiescent (14).

Preparation of tissue.
Three eye globes from donors aged 63, 71, and 74 yr old were obtained from National Disease Research Interchange (NDRI, Philadelphia, PA) with a death-to-enucleation time within 4.5 h (Table 1). Based on the report by Johnston et al. (18), which found that premorbid conditions, such as rapidity of death, were the greatest influencing factor in RNA quality, we used globes where the donors were on life support for <24 h. The anterior segment was removed, and inspection with a dissecting microscope revealed a normal posterior pole for each globe. Using RNase-free conditions, a 6 x 6-mm macular calotte that included the optic nerve was cryoprotected using the technique of Barthel and Raymond (4) with slight modification. Briefly, calottes were rinsed three times in 1x PBS containing 5% sucrose (wt/vol) for 10 min at 4°C. Calottes were progressively infiltrated with sucrose by 30-min incubations each at 4°C in PBS containing 10% and 20% sucrose (wt/vol), before being infiltrated in 2:1 of 20% sucrose (wt/vol):OCT compound (VWR International, Bridgeport, NJ) for 30 min. Calottes were embedded in fresh 2:1 of 20% sucrose (wt/vol):OCT mixture and frozen by immersion in isopentane (Fisher-Aldrich Chemical, Milwaukee, WI) chilled with dry ice. All tissue blocks were stored at –80°C for later use.


View this table:
[in this window]
[in a new window]
 
Table 1. Eye globe donor characteristics

 
Tissue sectioning and staining.
Tissue blocks were sectioned at 7 µm thickness on a Leica 1850 cryotome (Leica Microsystems, Bannockburn, IL). Only macular sections were used for this study. Individual sections were mounted on uncharged glass slides (Fisher Scientific, Fair Lawn, NJ) and immediately stained. The sections were fixed in 70% ethanol for 30 s and stained with hematoxylin and eosin Y (Fisher Scientific), each for 15 s. Sections were dehydrated in 100% ethanol twice for 60 s, then placed in xylene twice for 5 min. Sections were air dried for 5 min before being used immediately for laser-capture microdissection.

Laser-capture microdissection.
Native RPE cells were removed from cryosections using an Arcturus PixCell II laser-capture microdissector (LCM; Arcturus Engineering, Mountain View, CA) with LCM transfer film (Cap-Sure TF-100; Arcturus Engineering) using our previously published protocol (6). Typically, a 7.5-µm spot size, 50-µs pulse duration, and 75-mW power was necessary to cleanly dissect the cell of interest off of the tissue section. After dissection, the transfer cap was inspected with the microscope for contaminating tissue before being placed in 200 µl denaturing buffer that contained 4 M guanidine isothiocyanate, 0.02 M sodium citrate, 0.5% sarcosyl, and 1.6 µl ß-mercaptoethanol (14.5 M; Qiagen, Valencia, CA).

RNA extraction.
Total RNA was extracted from laser-captured native RPE cells using the RNeasy Mini-kit (Qiagen, Valencia, CA) according to the manufacturer’s recommendations. Total RNA, which was treated with DNase I (Qiagen) during RNA extraction, was eluted from the column in 30 µl RNase-free water. For laser-captured cells, before synthesizing probe, RNA quality was assessed by the expression of GAPDH from 100 cells using real-time RT-PCR with primers designed at the 5' end of gene.

Probe synthesis.
Total RNA from 5,000 ARPE-19 or native RPE cells was reverse transcribed in the presence of 50 µCi [33P]dCTP and 50 µCi [33P]dATP with 0.5 µg oligo dT according to a modified method of Sgroi et al. (26). To facilitate dissociation of reverse transcriptase from the newly synthesized first-strand cDNA, 1.5 µg T4 gene 32 protein (Ambion, Austin, TX) was added to the reaction mixture with 1 µl Sensiscript RT (Qiagen) (6) and incubated at 37°C for 90 min. The second strand was synthesized in the presence of 50 µCi [33P]dCTP and 50 µCi [33P]dATP, 500 ng random hexamers, and 20 U Klenow fragment (GIBCO-BRL) at room temperature for 90 min. The probe was purified by passage through a Bio-Spin 6 chromatography column (Bio-Rad Laboratories, Hercules, CA).

Microarray analysis.
The labeled, double-stranded cDNA was denatured and hybridized to the cDNA GeneFilter "Named genes" human array (4,325 genes; Invitrogen/Research Genetics, Huntsville, AL) using the manufacturer’s protocol. This array contains known genes that are an insert DNA from a sequence-verified IMAGE/LLNL clone from the 3' end of the gene. Arrays were exposed for 3 days to a high-density phosphor imaging screen (Bio-Rad Laboratories) and scanned at 50 µm resolution in a phosphor imaging instrument (FX Pro-Plus, Bio-Rad Laboratories). Lee et al. (19) have demonstrated that at least three replicate experiments are necessary to obtain consistent and reliable findings using cDNA microarrays. Thus, for each condition, three independent experiments were performed.

Image and statistical analysis.
The tiff images acquired from the phosphor imaging instrument were imported into the image analysis software (Pathways 3, Invitrogen/Research Genetics). This software aligns the images, quantifies a signal intensity for each gene, and when comparing images, normalizes the different hybridization signals on the basis of the 75% of the average signal intensity of the entire array. An individual gene was "expressed" if the signal intensity was >=1.4 fold above background (1) in at least two of three experiments.

To allow statistical comparison of the arrays, the gene expression signals were scaled according to the method of Tusher et al. (30, 32) that we have previously used. A reference data set was generated using the average expression of each gene from all arrays. The data for each hybridization were compared with this reference data set in a cube root scatter plot, and intensity values for each gene were then corrected with this "scaling factor."

Hierarchical cluster analysis of native RPE cells and the ARPE-19 cells under different culture conditions was determined with Cluster using average linkage clustering and visualized with TreeView (11). Significance analysis of microarrays (SAM; version 1.12) was used to determine gene expression differences by culture condition compared with native cells (30).

The microarray data appears on the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) web site (http://www.ncbi.nlm.nih.gov/geo/) using the platform number GPL538 and the series accession number GSE741.

Real-time RT-PCR.
An aliquot (1 ng) of total RNA used for microarray analysis was mixed in a volume of 20 µl that included 0.5 µl (0.5 µg/µl) oligo dT (Qiagen) and 0.75 µl T4 gene 32 protein (1.5 µg; Ambion, Austin, TX) at 70°C for 10 min, chilled on ice for 2 min, and reverse transcribed with Sensiscript (1 µl; Qiagen) in the presence of dNTP, RNase inhibitor, DTT inhibitor at 37°C for 60 min, 93°C for 5 min, and then 4°C. The first-strand cDNA was assayed using the LightCycler apparatus (Roche Diagnostics, Nutley, NJ). The primer sequences were designed to span consecutive exons using Primer3 (Whitehead Institute/MIT, Cambridge, MA) or the LightCycler Probe Design software. Sequences were verified using NCBI UniGene (Table 2). The standard curve consisted of PCR products (serial dilutions of 1–10–6 pg) using total RNA from ARPE-19 cells that were amplified from the same primer set for the specific gene of interest. The PCR products were separated on a TAE agarose gel, removed, and purified using the QIAEX II gel extraction kit (Qiagen). Thermocycling of each reaction was performed in a final volume of 20 µl containing SYBR Green PCR Master Mix (10 µl; Qiagen), primers A and B (10 µM each), and 2 µl template DNA in a concentration of2.5 mM MgCl2. PCR products were quantified using the second derivate maximum values calculated by the Light-Cycler analysis software. Negative controls without template were produced for each run. Expression levels of all genes were normalized to acidic ribosomal phosphoprotein mRNA levels (28). All PCR products were checked by melting point analysis. The Student’s t-test was used to compare the differential gene expression between conditions. P < 0.05 was considered significant.


View this table:
[in this window]
[in a new window]
 
Table 2. Real-time RT-PCR primers

 

    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

Expression profile of morphologically normal RPE cells overlying unthickened Bruch’s membrane and ARPE-19 cells.
Figure 1 demonstrates the typical appearance of macular RPE cells that were laser-capture microdissected. The cells appeared morphologically normal and were attached to an unthickened Bruch’s membrane. Confluent ARPE-19 cells, which had uniform morphology, did not vary in appearance with matrix condition (Fig. 2). Figure 3 shows the scatter plots of microarray analysis for pairwise comparison of native RPE from three donors. The R2 values ranged from 0.944 to 0.966 for all comparisons. Similar agreement was seen for the cultured cells grown on the different matrix proteins (data not shown).



View larger version (91K):
[in this window]
[in a new window]
 
Fig. 1. Cryosections that were stained with hematoxylin and eosin demonstrated normal morphology of the retinal pigment epithelial (RPE) cell and Bruch’s membrane from the macula of a 63-yr-old donor. A: the cryosection before laser microdissection. B: the sections after dissection with an area that is absent of RPE cells. C: microdissected RPE cells that are adherent to the transfer cap. Bar = 10 µm.

 


View larger version (175K):
[in this window]
[in a new window]
 
Fig. 2. Confluent ARPE-19 cells that were grown on plastic in DMEM+10% FBS and 5% CO2. Cell morphology did not vary by matrix protein. Bar = 10 µm.

 


View larger version (33K):
[in this window]
[in a new window]
 
Fig. 3. A pairwise comparison of microarray analysis scatter plots of laser-capture microdissected RPE from 3 donors. The R2 values, which ranged from 0.944 to 0.966, indicate a high degree of reproducibility for all comparisons.

 
Using our criterion for expression, we found that the number of genes expressed by native RPE was 2,996 genes (69%). The number of genes expressed by ARPE-19 cells varied with the matrix condition. The number of genes expressed by cultured RPE cells ranged from 2,219 (51%; Matrigel), 2,224 (51%; plastic), 2,502 (58%; collagen I), 3,139 (73%; fibronectin), 3,599 (83%; laminin), and 3,850 (89%; collagen IV).

Table 3 shows the 50 most highly expressed genes by morphologically normal native RPE. The majority of these genes are functionally related to cell shape/cytoskeleton, cell cycle/proliferation/apoptosis, and protein homeostasis (synthesis, processing and degradation). Of the 50 genes most highly expressed by native RPE, 66–84% of these genes were in the 50 most highly expressed genes by RPE cells grown on different matrices. All of the 50 most highly expressed genes, however, were expressed within the top 135 most highly expressed genes of cultured RPE cells, regardless of culture condition. Figure 4 shows the within group expression differences of the top 50 expressed genes for each condition. Table 4 lists the 50 genes with the lowest expression by native RPE cells. With this gene set, RPE cells grown on laminin (94%), fibronectin (96%), and collagen IV (86%) had a higher percentage of genes expressed than cells grown on Matrigel (68%), plastic (44%), or collagen I (48%).


View this table:
[in this window]
[in a new window]
 
Table 3. Fifty most abundant genes expressed by native RPE cells

 


View larger version (77K):
[in this window]
[in a new window]
 
Fig. 4. Plots of the relative signal intensity of the top 50 expressed genes for each condition (n = 3 experiments for each condition). Pink square = experiment 1; yellow triangle = experiment 2; and blue diamond = experiment 3.

 

View this table:
[in this window]
[in a new window]
 
Table 4. Fifty lowest abundant genes expressed by native RPE

 
Proximity of cultured cells to native RPE by cluster analysis.
Unsupervised cluster analysis was performed to determine the effect of different matrix proteins on the mRNA phenotype of ARPE-19 cells and to determine which condition most closely approximates morphologically normal, native RPE cells. Figure 5 shows that the expression profile of ARPE-19 cells grown on fibronectin, laminin, and collagen IV were separated from cells grown on collagen I, plastic, and Matrigel. The expression profile of RPE cells grown on plastic clustered closest to native RPE cells.



View larger version (57K):
[in this window]
[in a new window]
 
Fig. 5. Unsupervised cluster analysis identified two distinct groups. The first group consisted of ARPE-19 cells grown on plastic, collagen I, and Matrigel, while the second group comprised cells grown on fibronectin, laminin, and collagen IV. Cells grown on plastic clustered closest to laser-capture microdissected RPE cells. Green is log2 > 1; red is log2 < 1; black is log2 = 1; and gray indicates not expressed.

 
Several supervised cluster analysis approaches were utilized to determine whether the culture condition that most closely produced that of native RPE changed. Using unpaired two-class SAM analysis with a moderately stringent false discovery rate (FDR) of 5% and a twofold differential expression ratio threshold, 950 genes were identified as potentially differentially expressed between native RPE and cultured cells (all matrix conditions combined). As shown in Figure 6, supervised cluster analysis of this gene set was similar to the unsupervised approach with cells grown on plastic again, clustering closest to native RPE cells. A dendrite now shows that the mRNA phenotypes of cells grown on collagen IV and laminin were more similar than cells grown on fibronectin. Supervised cluster analysis using 510 genes involved in differentiation (Fig. 7), 308 genes related to cell-matrix interaction (data not shown), and 167 genes from the native RPE library of Buraczynska et al. (7) (data not shown) yielded a similar clustering pattern for the arrays as the gene set from the unpaired two-class SAM analysis. Cells grown on plastic had the closest expression profile to native macular RPE cells. Cells grown on collagen IV and laminin displayed similar expression profiles and were different from cells grown on collagen I.



View larger version (53K):
[in this window]
[in a new window]
 
Fig. 6. Supervised cluster analysis using 950 differentially expressed genes by SAM. Green is log2 > 1; red is log2 < 1; black is log2 = 1; and gray indicates not expressed.

 


View larger version (45K):
[in this window]
[in a new window]
 
Fig. 7. Supervised cluster analysis using 510 genes involved in differentiation. Green is log2 > 1; red is log2 < 1; black is log2 = 1; and gray indicates not expressed.

 
Proximity of cultured RPE cells to native macular RPE by SAM analysis.
Several methods were used to assess the similarity of expression profiles between native RPE and RPE cells in vitro grown on different matrix proteins. The expression profiles of RPE cells grown on plastic and collagen IV were used in this assessment, since their expression profiles were distinct from each other, and they had relatively similar and dissimilar profiles, respectively, to native RPE. First, the number of genes that native RPE expressed was determined for RPE cells grown on plastic and collagen IV. RPE cells grown on plastic and collagen IV expressed 91% and 97% of the genes, respectively, which were expressed by native RPE.

Second, 2,080 genes on the array were not expressed by native RPE. Of these genes, 337 (16.2%) genes were expressed by RPE cells grown on plastic. The functional categories of the genes expressed by cells grown on plastic, but not native RPE, include signal transduction (20%), metabolism (13.6%), transcription factors/DNA binding (10%), protein processing (9%), and cytoskeleton/cell shape (6%). Of the genes not expressed by native RPE, 1,081 genes (52.8%) were expressed by cells grown on collagen IV. The functional categories of genes expressed by cells grown on collagen IV, but not native RPE, included signal transduction (20%), metabolism (14.5%), protein processing (12.8%), transcription factors/DNA binding (12.6%), and cytoskeleton/structural (6.3%).

Third, the number of differentially expressed genes between native and RPE cells in vitro was used to assess the proximity of the expression profiles. An FDR <10% has been recommended as a reliable indicator of statistical precision for studies using human tissue (16). Two-class unpaired SAM analysis between native RPE and cultured RPE cells (all conditions combined) with an FDR of 9.6% identified 2,161 genes (50% of genes on the array) that were differentially expressed. When native RPE cells were compared with each individual culture condition by SAM, the number of differentially expressed genes ranged from 120 (RPE cells grown on plastic; FDR 9.7%), 433 (cells grown on collagen I; FDR 9.8%), 1,950 (cells grown on collagen IV; FDR 9.7%), 2,405 (cells grown on fibronectin; FDR 9.9%), and 2,588 (cells grown on laminin; FDR 7.3%). Cells grown on Matrigel had significant variation with the lowest FDR of 18%, which identified 581 differentially expressed genes. These results suggest that a significant number of genes are differentially expressed between native and cultured RPE cells.

The function of differentially expressed genes between native RPE and individual culture conditions was a fourth factor used to evaluate the similarity of expression profiles. With SAM, the molecular functional annotation of the 120 genes (FDR 9.7%) that were differentially expressed between native RPE and cells grown on plastic included intracellular signal transduction (22.6%), metabolism (16.5%), cytoskeleton (11.3%), protein processing (11.3%), and transcription factors/DNA binding (11.3%). Highly stringent conditions using an FDR threshold of 1.5% was performed to identify individual genes that might contribute to differences in mRNA phenotype between these two conditions. This analysis shows that 32 genes were upregulated (Table 5) and 18 genes were downregulated (Table 6) by native RPE compared with cells grown on plastic. Of these genes, 45% had biological function related to cell cycle/proliferation/apoptosis, and 25% were related to cell shape/differentiation.


View this table:
[in this window]
[in a new window]
 
Table 5. Differential gene expression (upregulated genes) between RPE cells grown on plastic and native RPE sorted by SAM with an FDR of 1.5%

 

View this table:
[in this window]
[in a new window]
 
Table 6. Differential gene expression (downregulated genes) between RPE cells grown on plastic and native RPE sorted by SAM with an FDR of 1.5%

 
With an FDR of 9.7%, SAM identified 1,950 potential differentially expressed genes between native RPE and cells grown on collagen IV. The most common molecular functional categories were intracellular signal transduction (17.4%), metabolism (13.4%), protein processing (12.4%), transcription factors (10.2%), and cytoskeleton (6.1%). High-stringency (1% FDR), two-class unpaired SAM, which was used to determine individual differentially expressed genes, found 25 upregulated genes (Table 7) and 23 downregulated genes (Table 8) by native RPE cells compared with cells grown on collagen IV. Twenty-nine genes from this list overlapped with differentially expressed genes from those grown on plastic compared with native RPE. As with cells grown on plastic, a large percentage of genes had biological function related to cell cycle/proliferation/apoptosis (44%) and cell shape/differentiation (19%).


View this table:
[in this window]
[in a new window]
 
Table 7. Differential gene expression between RPE cells grown on collagen IV and native RPE sorted by SAM with an FDR of 1.0%

 

View this table:
[in this window]
[in a new window]
 
Table 8. Differential gene expression (downregulated genes) between RPE cells grown on collagen IV and native RPE sorted by SAM with an FDR of 1.0%

 
Real-time RT-PCR validation.
Real-time RT-PCR (n = 3 experiments for each condition) was performed on differentially expressed genes randomly chosen from genes identified by SAM analysis comparing native RPE to cultured RPE (all conditions combined). As seen in Table 9, the differential expression pattern was similar to the arrays for all five genes and statistically validated for four of the five genes.


View this table:
[in this window]
[in a new window]
 
Table 9. Differential expression by real-time RT-PCR

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
The expression profile of uniform, morphologically normal native macular RPE cells that were attached to unthickened Bruch’s membrane from human elderly donors illustrate its functional diversity. For example, highly abundant genes included D-dopachrome tautomerase (melanin biosynthesis), cytochrome c oxidase subunit VIIc (respiratory chain/energy production), phosphodiesterase 6H, cGMP-specific, cone, gamma (visual transduction), carbonic anhydrase VA (fluid transport), and glutathione peroxidase 3 (antioxidant defense system). Like Sharon et al. (27), who used SAGE of native RPE from an 88-yr-old donor, we found a high degree of the most abundant genes that were related to protein synthesis and degradation, possibly due to the high phagocytic activity of RPE. A complete comparison with other libraries is difficult because of array design, small sample size, technique differences, and possible differences in RPE cell heterogeneity. In addition, the impact of the elderly donor age on the expression profile is unknown. Our laboratory is currently evaluating the impact of age on the expression profile of morphologically normal, native RPE.

Our interpretations are based on native RPE dissected from three eyes. This data set had reasonable precision, as evidenced by the pairwise comparison of scatter plots for native RPE (Fig. 3) and SAM analysis, which had FDR values well below the recommended <10% level for estimating reliable precision on studies using human tissue (16). The statistical precision allowed us to identify statistically distinct gene sets between different groups. We acknowledge that a larger sample size would provide more power that could expand the statistically relevant genes, and we understand that these results may not be generalizable to all native RPE.

Unamplified RNA was used to synthesize the probes for microarray analysis. We adapted our RNA extraction and reverse transcription protocol of laser-captured cells for RT-PCR to maximize the RNA yield for microarray analysis (6). We optimized a "double labeling" 33P-labeling system from Sgroi et al. (26) that reduced from 20,000 to 5,000 the cells necessary for microarray analysis. Although an unamplified approach reduces bias that concerns any amplification, radiolabeling is necessary because fluorescent labeling used with cDNA or oligonucleotide arrays requires significantly more probe than what was used with this protocol. This approach also reduces the number of genes that are available to study on the arrays compared with oligonucleotide or cDNA arrays.

ARPE-19 cells grown to confluence are a commonly utilized in vitro approach to study RPE cell behavior because these can have differentiated morphology that simulates RPE cells in vivo (2, 5, 12, 13, 17, 22, 23, 29). To develop highly differentiated morphology, a long culture time (i.e., 2.5 mo) can be required (2). We seeded cells at a relatively high density, grew cells until they reached visual confluence, and serum withdrew cells using our previously published protocol to reduce the proliferative stimulus (14). We wanted a short culture time to limit matrix production by cells so we could assess the effect of the intended matrix protein on the mRNA phenotype. Although cells appeared uniform across different matrix conditions, the expression profiles were different. Cluster analysis identified cells grown on collagen IV, laminin, and fibronectin as similar and distinct from cells grown on collagen I. These results seem logical since collagen IV, laminin, and fibronectin are major components of the RPE basement membrane and since collagen I, which is not normal component of the basement membrane, is typically found during aging or pathological conditions involving the RPE basement membrane (21).

We were surprised that RPE cells grown on plastic had the closest expression profile to native RPE by cluster analysis, SAM, and the functional profile of differentially expressed genes. With the culture conditions used in this study, we hypothesize that plastic stimulates RPE cells to produce a matrix that is more similar to the basement membrane than the individual matrix proteins used in this study. Campochiaro et al. (8) have shown that cultured RPE cells grown on plastic synthesize a matrix that is similar to its in vivo basement membrane. It is probable that the supramolecular structure of the basement membrane influences the cell’s phenotype, which of course would be absent when utilizing individual matrix proteins. A composite of matrix proteins, Matrigel is considered a close in vitro approximation to the basement membrane in vivo (25). Our experiments produced more expression variability when cells were grown on (growth factor reduced) Matrigel than the other matrix conditions. One explanation for the inconsistent expression profile is the variability in the supramolecular structure from experiment to experiment, which induced a different mRNA phenotype.

Although RPE cells grown on plastic clustered closest to native macular RPE, the criterion that we used suggested that the expression profiles of cultured cells were different from native cells. For example, a significant proportion of genes not expressed by native RPE cells were expressed by cells grown on plastic (16.2%) and collagen IV (52%). Many genes that were differentially expressed between native RPE and cells grown on plastic and collagen IV are involved in signal transduction, transcription factors, and protein processing. The large number of genes related to these molecular functions makes it difficult to predict a specific phenotype, because of the redundancy and complexity of intracellular signaling pathways. Many of these differentially expressed genes are involved in cell cycle/proliferation/apoptosis and cell shape/differentiation. A number of laboratories use RPE cells in vitro to investigate aging and age-related macular degeneration changes. In the macula, RPE cells undergo apoptosis and morphological deterioration with aging (9, 31). The overlay of in vitro-induced alterations in the expression of genes related to cell cycle and morphology would make age-related changes difficult to interpret in an in vitro system such as what was used in this study. It is possible to reduce the proliferative stimulus or improve morphology by lengthening the time in culture, but this approach could change the influence of the matrix on the mRNA phenotype. In addition, the stability of the matrix can change over long culture duration. For example, the stability of reconstituted Matrigel is 14 days at 37°C. Other culture conditions are likely to have an influence on the expression profile. In a separate group of experiments, we are currently assessing the effects of culture medium and duration on the expression profile of RPE cells grown on plastic in an attempt to provide a better culture approximation of native RPE. Although our data show that typical basement proteins such as collagen IV, laminin, and fibronectin induce similar mRNA RPE phenotypes, our results also suggest that these phenotypes are different from native RPE. Hence, care must be utilized when interpreting and extrapolating gene expression changes from cultured cells to native RPE.


    GRANTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
We thank the National Disease Research Interchange for the donor eyes. This work was supported by National Institutes of Health Grant EY-14055 (to J. T. Handa), the Michael Panitch Macular Degeneration Research Fund, a gift from Aleda Wright, and an unrestricted award from the Research to Prevent Blindness (RPB) to the Wilmer Eye Institute. J. T. Handa is the recipient of a Lew R. Wasserman Merit Award from the RPB.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: J. T. Handa, M.D. 3-109 Jefferson St. Bldg.; 600 N. Wolfe St; Johns Hopkins Medical Institutes; Baltimore, MD 21287 (E-mail: jthanda{at}jhmi.edu).

10.1152/physiolgenomics.00179.2003.


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, Boldrick JC, Sabet H, Tran T, Yu X, Powell JI, Yang L, Marti GE, Moore T, Hudson J Jr, Lu L, Lewis DB, Tibshirani R, Sherlock G, Chan WC, Greiner TC, Weisenburger DD, Armitage JO, Warnke R, Levy R, Wilson W, Grever MR, Byrd JC, Botstein D, Brown PO, and Staudt LM. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403: 503–511, 2000.[CrossRef][ISI][Medline]
  2. Alizadeh M, Wada M, Gelfman CM, Handa JT, and Hjelmeland LM. Downregulation of differentiation specific gene expression by oxidative stress in ARPE-19 cells. Invest Ophthalmol Vis Sci 42: 2706–2713, 2001.[Abstract/Free Full Text]
  3. Aumailley M and Gayraud B. Structure and biological activity of the extracellular matrix. J Mol Med 76: 253–265, 1998.[CrossRef][ISI][Medline]
  4. Barthel LK and Raymond PA. Improved method for obtaining 3-microns cryosections for immunocytochemistry. J Histochem Cytochem 38: 1383–1388, 1990.[Abstract]
  5. Bevitt DJ, Mohamed J, Catterall JB, Li Z, Arris CE, Hiscott P, Sheridan C, Langton KP, Barker MD, Clarke MP, and McKie N. Expression of ADAMTS metalloproteinases in the retinal pigment epithelium derived cell line ARPE-19: transcriptional regulation by TNF alpha. Biochim Biophys Acta 1626: 83–91, 2003.[ISI][Medline]
  6. Boylan S, Honda S, Hjelmeland LM, and Handa JT. An optimized protocol for first strand cDNA synthesis from laser capture microdissected tissue. Lab Invest 81: 1167–1169, 2001.[ISI][Medline]
  7. Buraczynska M, Mears AJ, Zareparsi S, Farjo R, Filippova E, Yuan Y, MacNee SP, Hughes B, and Swaroop A. Gene expression profile of native human retinal pigment epithelium. Invest Ophthalmol Vis Sci 43: 603–607, 2002.[Abstract/Free Full Text]
  8. Campochiaro PA, Jerdon JA, and Glaser BM. The extracellular matrix of human retinal pigment epithelial cells in vivo and its synthesis in vitro. Invest Ophthalmol Vis Sci 27: 1615–1621, 1986.[Abstract]
  9. Del Priore LV, Kuo YH, and Tezel TH. Age-related changes in human RPE cell density and apoptosis proportion in situ. Invest Ophthalmol Vis Sci 43: 3312–3318, 2002.[Abstract/Free Full Text]
  10. Dunn KC, Aotaki-Keen AE, Putkey FR, and Hjelmeland LM. ARPE-19, a human retinal pigment epithelial cell line with differentiated properties. Exp Eye Res 62: 155–169, 1996.[CrossRef][ISI][Medline]
  11. Eisen MB, Spellman PT, Brown PO, and Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95: 14863–14868, 1998.[Abstract/Free Full Text]
  12. Fan W, Zheng JJ, Peiper SC, and McLaughlin BJ. Changes in gene expression of ARPE-19 cells in response to vitreous treatment. Ophthalmic Res 34: 357–365, 2002.[ISI][Medline]
  13. Fukuoka Y and Medof EM. C5a receptor-mediated production of IL-8 by the human retinal pigment epithelial cell line, ARPE-19. Curr Eye Res 23: 320–325, 2001.[CrossRef][ISI][Medline]
  14. Handa JT, Reiser KM, Matsunaga H, and Hjelmeland LM. The advanced glycation endproduct pentosidine induces the expression of PDGF-B in human retinal pigment epithelial cells. Exp Eye Res 66: 411–419, 1998.[CrossRef][ISI][Medline]
  15. Honda S, Farboud B, Hjelmeland LM, and Handa JT. Induction of an aging mRNA retinal pigment epithelial cell phenotype by matrix-containing advanced glycation end products in vitro. Invest Ophthalmol Vis Sci 42: 2419–2425, 2001.[Abstract/Free Full Text]
  16. Hwang D, Schmitt WA, and Stephanopoulos G. Determination of minimum sample size and discriminatory expression patterns in microarray data. Bioinformatics 18: 1184–1193, 2002.[Abstract/Free Full Text]
  17. Janssen JJ, Kuhlmann ED, van Vugt AH, Winkens HJ, Janssen BP, Deutman AF, and Driessen CA. Retinoic acid delays transcription of human retinal pigment neuroepithelium marker genes in ARPE-19 cells. Neuroreport 11: 1571–1579, 2000.[ISI][Medline]
  18. Johnston NL, Cervenak J, Shore AD, Torrey EF, Yolken RH, and Cerevnak J. Multivariate analysis of RNA levels from postmortem human brains as measured by three different methods of RT-PCR. Stanley Neuropathology Consortium. J Neurosci Methods 77: 83–92, 1997.[CrossRef][ISI][Medline]
  19. Lee ML and Whitmore GA. Power and sample size for DNA microarray studies. Stat Med 21: 3543–3570, 2002.[CrossRef][ISI][Medline]
  20. Loffler KU and Lee WR. Basal linear deposit in the human macula. Graefes Arch Clin Exp Ophthalmol 224: 493–501, 1986.[ISI][Medline]
  21. Newsome DA, Hewitt AT, Huh W, Robey PG, and Hassell JR. Detection of specific extracellular matrix molecules in drusen, Bruch’s membrane, and ciliary body. Am J Ophthalmol 104: 373–381, 1987.[ISI][Medline]
  22. Ong JM, Aoki AM, Seigel GM, Sacerio I, Castellon R, Nesburn AB, and Kenney MC. Oxysterol-induced toxicity in R28 and ARPE-19 cells. Neurochem Res 28: 883–891, 2003.[CrossRef][ISI][Medline]
  23. Philp NJ, Wang D, Yoon H, and Hjelmeland LM. Polarized expression of monocarboxylate transporters in human retinal pigment epithelium and ARPE-19 cells. Invest Ophthalmol Vis Sci 44: 1716–1721, 2003.[Abstract/Free Full Text]
  24. Schock F and Perrimon N. Molecular mechanisms of epithelial morphogenesis. Annu Rev Cell Dev Biol 18: 463–493, 2002.[CrossRef][ISI][Medline]
  25. Schuetz EG, Li D, Omiecinski CJ, Muller-Eberhard U, Kleinman HK, Elswick B, and Guzelian PS. Regulation of gene expression in adult rat hepatocytes cultured on a basement membrane matrix. J Cell Physiol 134: 309–323, 1988.[ISI][Medline]
  26. Sgroi DC, Teng S, Robinson G, LeVangie R, Hudson JR Jr, and Elkahloun AG. In vivo gene expression profile analysis of human breast cancer progression. Cancer Res 59: 5656–5661, 1999.[Abstract/Free Full Text]
  27. Sharon D, Blackshaw S, Cepko CL, and Dryja TP. Profile of the genes expressed in the human peripheral retina, macula, and retinal pigment epithelium determined through serial analysis of gene expression (SAGE). Proc Natl Acad Sci USA 99: 315–320, 2002.[Abstract/Free Full Text]
  28. Simpson DA, Feeney S, Boyle C, and Stitt AW. Retinal VEGF mRNA measured by SYBR green I fluorescence: A versatile approach to quantitative PCR. Mol Vis 6: 178–183, 2000.[ISI][Medline]
  29. Singh S, Zheng JJ, Peiper SC, and McLaughlin BJ. Gene expression profile of ARPE-19 during repair of the monolayer. Graefes Arch Clin Exp Ophthalmol 239: 946–951, 2001.[ISI][Medline]
  30. Tusher VG, Tibshirani R, and Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98: 5116–5121, 2001.[Abstract/Free Full Text]
  31. Watzke RC, Soldevilla JD, and Trune DR. Morphometric analysis of human retinal pigment epithelium: correlation with age and location. Curr Eye Res 12: 133–142, 1993.[ISI][Medline]
  32. Weigel AL, Handa JT, and Hjelmeland LM. Microarray analysis of H2O2-, HNE-, or tBH-treated ARPE-19 cells. Free Radic Biol Med 33: 1419–1432, 2002.[CrossRef][ISI][Medline]