Tissue Differential Microarray Analysis of Dexamethasone Induction Reveals Potential Mechanisms of Steroid Glaucoma

Abstract

purpose. To identify myocilin (TIGR/MYOC) properties that are specific to the human trabecular meshwork (HTM). To search for genes highly expressed in dexamethasone (DEX)-induced HTM cells that are barely expressed or absent in DEX-induced cells from other tissues.

methods. TIGR/MYOC induction by DEX (10⁻⁷ M for 8–10 days) was analyzed by Northern and Western blot analyses in HTM, human umbilical vein endothelial cells, HeLa cells, and human embryonic skeletal muscle cells and optic nerve head (ONH) astrocytes at confluence. Processing and secretion were analyzed after the cells were infected with adenoviruses overexpressing wild-type and mutant forms of TIGR/MYOC. Affymetrix U95Av2 GeneChips (n = 6) and software were used to compare paired expression profiles of HTM, HTM-DEX, ONH astrocytes, and ONH astrocytes-DEX. Identification of HTM-DEX–specific genes (compared with ONH astrocytes-DEX) was performed by selecting genes with the highest fold change values (≥20). Genes with fold change values of four or more were matched with loci linked to glaucoma, by using gene databases.

results. TIGR/MYOC induction by DEX occurred only in HTM cells. Secretory and glycosylation characteristics remained the same across cell types. Expression profile analysis revealed multiple genes differentially upregulated in HTM-DEX including, in addition to TIGR/MYOC, a serine protease inhibitor (α1-antichymotrypsin), a neuroprotective factor (pigment epithelium-derived factor), an antiangiogenesis factor (cornea-derived transcript 6), and a prostaglandin synthase (prostaglandin D₂ synthase). Fifteen of the 249 genes with fold change values of four or more mapped to glaucoma-linked loci.

conclusions. The induction of TIGR/MYOC by DEX is HTM-specific, whereas its secretory and glycosylation characteristics are ubiquitous. The known functions of HTM-DEX–specific genes reveal the presence of protective and damaging mechanisms for regulation of IOP during DEX treatment. Besides TIGR/MYOC, other HTM-DEX–specific genes may be good candidates for linkage to glaucoma.

The glaucomas are a group of diseases characterized by progressive degeneration of the optic nerve, usually accompanied by elevated intraocular pressure (IOP). They affect 70 million people worldwide and are the leading cause of blindness among African Americans. ¹ Recently, mutations in TIGR/MYOC, also known as trabecular meshwork inducible glucocorticoid response protein (TIGR), have been linked to several types of glaucoma. ^{2 3 4} Although the function of this gene is currently unknown, genetic linkage analysis has demonstrated the presence of more than 25 distinct mutations in families with the most prevalent form of glaucoma, primary open-angle glaucoma (POAG), and in those with a rare, early-onset form, juvenile open-angle glaucoma (JOAG). ⁵  

TIGR/MYOC was identified as a protein upregulated in the HTM after long-term treatment with DEX ^{6 7 8} and was independently discovered in ciliary body ⁹ and normal human retina. ¹⁰ Some investigators believe its induction by DEX correlates with the time course and development of steroid-induced glaucoma, ^{8 11} a condition occurring in 30% to 40% of patients who are treated with steroids. ¹² Because steroid-induced glaucoma mimics many aspects of POAG, it may be an important model for study of the disease. Mutations in TIGR/MYOC, however, are not implicated in causing or increasing susceptibility to steroid-induced glaucoma. ^{13 14}  

TIGR/MYOC is a 504-amino-acid secretory protein translated from three exons. The amino terminal domain contains N- and O-glycosylation sites and a leucine zipper, and it is 25% homologous to the heavy chain of myosin. TIGR/MYOC is glycosylated ^{8 15} and the glycosylated and nonglycosylated protein forms migrate at approximately 57 and 55 kDa, respectively, on SDS-PAGE. ¹⁵  

Most glaucoma-linked mutations in TIGR/MYOC are located at its C-terminal end, in the region known as the olfactomedin (OLF) domain because of its 40% homology to olfactomedin, an extracellular matrix (ECM) protein first described in bullfrog olfactory epithelium. ¹⁶ Recent studies have demonstrated that many missense and nonsense mutations affecting the OLF domain impair secretion of the TIGR/MYOC mutant protein. ^{17 18} The impairment of secretion appears to be due to an inefficient processing of the mutant protein that results in misfolding and accumulation of aggregates inside the cell. ¹⁵ These results have lead us to postulate that the association of TIGR/MYOC mutants with glaucoma may be through a gain of deleterious function, ¹⁵ an idea supported by the findings that heterozygous and homozygous Myoc-knockout mice are viable and fertile, have normal IOP, and exhibit normal ocular morphology. ¹⁹  

The expression of TIGR/MYOC may be under the control of its 5-kb promoter region, which contains 13 predicted hormone-response elements and includes several glucocorticoid-response elements (GREs). ^{6 8 20} It has also been reported that multiple GRE half-sites exist within the 1900-bp region upstream of the putative translation start site. ²¹ Some of these half-sites have been described in systems in which there is delayed induction of glucocorticoid-induced gene expression. ²² It has been proposed that the delayed induction of TIGR/MYOC could be the result of classic GREs far upstream, the nearer GRE half-sites, or a secondary response to steroids. ²¹ More recent studies examining the promoter’s effects in luciferase reporter gene assay systems show that certain regions of putative GREs are not responsive to treatment with DEX ²⁰ and that TIGR/MYOC is a delayed glucocorticoid-responsive gene in which the glucocorticoid receptor indirectly affects expression. ²³ Thus, the described GREs in the promoter may not be directly responsible for upregulation of DEX, and distinct factors specifically present in DEX-induced HTM may first be required to stimulate expression of TIGR/MYOC. To date, these factors have not been described. 

TIGR/MYOC is highly expressed and secreted in the TM but is also found in many other tissues of the body, including skeletal muscle, retina, ⁹ and ONH astrocytes. ^{24 25} However, mutations in TIGR/MYOC are only linked to glaucoma and do not seem to affect the function of any other organ. This fact may be an indication that TIGR/MYOC serves TM-specific functions and/or has TM-specific properties that could be integral to glaucoma pathophysiology. In this study, we compared TIGR/MYOC expression, induction, secretion, and posttranslational processing in several selected human cell types. In an effort to identify TM glucocorticoid factors that either directly modulate IOP or affect TIGR/MYOC levels only in the TM, we used gene array technology. To select against genes upregulated by DEX in non-TM tissues, we compared the expression profile of HTM-DEX cells with another DEX-induced cell line, ONH astrocytes. Our choice of ONH astrocytes was based on their relevance in degeneration of the ONH in glaucoma, ²⁶ the presence of detectable levels of TIGR/MYOC in vivo, ²⁴ and our own result of the absence of DEX-induced TIGR/MYOC in these cells. In addition, because of the delayed glucocorticoid response of TIGR/MYOC, we chose to compare DEX induction after longer periods of drug treatment. 

Materials and Methods

Cell Culture

To examine the properties of TIGR/MYOC in different cell types, we chose several human cell lines based on their characteristics. HTM cells were used as a positive point of reference. Human umbilical vein endothelial cells (HUVECs) were chosen for their endothelial nature and because they are well described in the literature. Human embryonic skeletal muscle cells (SkMCs) and ONH astrocytes were chosen as representatives of other tissues with previously reported TIGR/MYOC expression. ^{9 27 28} HeLa cells were used as an unrelated, transformed cell line. 

The procedures in this study adhered to the provisions of the Declaration of Helsinki for research in human tissue. HTM cell lines used in these experiments were derived from two different donors with no history of glaucoma (HTM40 and HTM41). Primary cultures were established from donor eyes obtained from national eye banks within 48 hours of death. The TM from each of the donor eyes was isolated from surrounding tissue with incisions both anterior and posterior to the meshwork and treated with collagenase, as previously described. ²⁹ Briefly, the explanted tissue was placed on 2% gelatin-coated 35-mm dishes and improved minimal essential medium (IMEM; Biofluids, Rockville, MD) supplemented with 20% fetal bovine serum (FBS) (Gibco BRL, Rockville, MD) and 50 μg/mL gentamicin (Gibco BRL). Cells that grew from digested tissue were first-passaged 1:4 at 2 to 3 weeks after explantation and subsequently propagated in the same medium with 10% serum. Each cell line survived for approximately 8 to 10 passages. For the current experiments, HTM cells were used at passages 4 to 6. HUVECs and HeLa cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). Human embryonic SkMCs were obtained from Clonetics (Walkersville, MD). HUVECs were maintained in endothelial cell medium EBM-2 basal medium with EGM-2 SingleQuots supplements and growth factors from Clonetics (Walkersville, MD). HeLa cells were maintained in IMEM supplemented with 10% FBS, 50 μg/mL gentamicin, and 100 μM nonessential amino acid solution (Gibco BRL). SkMCs were maintained in the provider’s medium (Clonetics). 

ONH astrocytes (line 00-1LLS) were grown and characterized from two normal eyes (aged 15 years) with no history of eye disease, diabetes, or neurodegenerative disease, as previously described in detail. ^{30 31} Briefly, four explants from each human lamina cribrosa were dissected and freed from sclera and surrounding tissues. Each explant was placed into a 25-cm² tissue culture flask (Primaria; Falcon-BD Biosciences, Lincoln Park, NJ) and maintained in Dulbecco’s modified Eagle’s medium (DMEM)/Ham’s F-12 (Washington University Medical School Tissue Culture Support Center, St. Louis, MO) supplemented with 10% FBS (BioWhittaker, Walkersville, MD) and PSFM (10,000 U/mL penicillin, 10,000 μg/mL streptomycin, and 25 μg/mL amphotericin B; Gibco BRL). Cells were kept in a 37°C, 5% CO₂ incubator. After 2 to 4 weeks, using a modified immunopanning procedure described by Mi and Barres, ³² we obtained purified primary cultures. To select ONH (type 1B) astrocytes, cell suspensions from the primary cultures were placed on a P100 panning dish coated with C5 anti-neuroepithelial antibody (1:20, gift from Ben Barres, Stanford University, Stanford, CA) and incubated at room temperature for 30 minutes. After the nonadherent cells were removed, the attached cells were grown for 1 to 2 weeks in DMEM/Ham’s F-12 with 10% FBS in the incubator. Cultured ONH astrocytes were characterized by positive immunostaining for both glial fibrillary acidic protein (GFAP) and neural cell adhesion molecule (NCAM), as described. ³⁰ For these experiments, third-passage cells were grown and maintained in DMEM/Ham’s F-12 (Gibco BRL), 10% FBS, and 50-μg/mL gentamicin. 

RNA Extraction and Northern Blot Hybridization

Cell cultures were grown to confluence in 10-cm plates and treated with DEX (Sigma, St. Louis, MO) for 10 days at a concentration of 10⁻⁷ M. DEX was prepared in absolute ethanol at 0.1 mM and diluted 1000-fold into fresh complete IMEM every other day for the duration of the experiment. At the end of 10 days, both control and treated sample RNAs were extracted with a kit (RNeasy Mini Kit; Qiagen, Valencia, CA). RNA molecules selectively bound to the silica gel base were eluted with 50 μL RNase-free water. RNA from each sample was lyophilized to dryness, denatured in 50% formamide, and separated by 2.2 M formaldehyde, 1.25% agarose, 0.05 M 3-[N-morpholino] propane sulfonic acid (MOPS), and 1 mM EDTA gel electrophoresis. After electrophoresis, gels were washed with dH₂0 for 30 minutes and transferred to a nitrocellulose membrane (Schleicher & Schuell, Keene, NH) overnight by capillary action with 10× sodium saline citrate (SSC). After the transfer, UV cross-linked blots were prehybridized at 42°C for 6 to 12 hours in a buffer containing 50% formamide, 5× SSC, 5× Denhardt’s, 50 mM NaP0₄ (pH 7.4), 0.1% sodium dodecyl sulfate (SDS) with 50 μg/mL sheared salmon sperm DNA (Research Genetics, Huntsville, AL). The filters were then placed in fresh buffer, hybridized overnight to 1 to 5 × 10⁶ cpm/mL of random-primer–labeled TIGR/MYOC cDNA (Roche Molecular Biosystems, Indianapolis, IN) at the same temperature. The TIGR/MYOC cDNA probe used in the hybridization experiments contained 1603 base pairs and was obtained by PCR amplification of pMC2 ¹⁵ with primers 5′-AGCTTTGTTTAAACGCCTCACCAAGCCTCTGCAA-3′ (59–78 nucleotides [nt] TIGR/MYOC cDNA) and 5′-GGCGGATCCTGCCATTGCCTGTACAGCTTGGAG-3′ (1597–1621 nt TIGR/MYOC cDNA) followed by gel purification of the fragment with a PCR purification kit (QIAquick; Qiagen). After hybridization, filters were washed five times (15–20 minutes each): four times in 2× SSC-1% SDS (two at room temperature [RT] and two at 52°C) and once in 2× SSC at 52°C. Exposure was conducted using x-ray film (BioMax MR; Scientific Imaging Systems, Eastman Kodak, New Haven, CT) at −80°C with intensifying screens. To monitor RNA degradation and loading, filters were subsequently rehybridized to 2 × 10⁶ cpm/mL of 28S oligonucleotide at 42°C for 2 hours, washed twice in 2× SSC-1% SDS (one at RT, one at 37°C) and exposed at RT. ³³ TIGR/MYOC nucleotide number is according to Adam et al. ³⁴  

Protein Extraction and Western Blot Analysis

Cell cultures were grown to confluence in 35-mm wells and treated with fresh 10⁻⁷ M DEX every other day. Control and DEX-treated cultures were collected at time points ranging from 1 to 10 days. For each cell culture, medium and cells from each 35-mm well were collected separately. The medium was centrifuged at 41,000g for 12 minutes to remove cellular debris. Cleared medium containing secreted cell products was then concentrated 20× on a filter (Centricon C-30; Amicon-Millipore, Bedford, MA) followed by buffer exchange to 0.01 M Tris-HCl [pH 7.4] and storage at −20°C. After the medium was removed, the cell monolayer from each well was washed twice with phosphate-buffered saline (PBS; Gibco BRL) followed by the addition of 2× loading buffer (100 mM Tris-HCl, 200 mM dithiothreitol [DTT], 4% SDS, 20% glycerol, and 0.02% bromophenol blue) to a concentration of 3 × 10⁷ cells/mL. Proteins were extracted by boiling samples for 10 minutes, followed by centrifugation at 16,000g and storage at −20°C. 

Protein extracts were separated by 10% SDS-PAGE and electrotransferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad, Hercules, CA). After blocking with 5% nonfat dry milk in PBS-0.2% Tween 20 (Sigma) for 2 hours, the membranes were incubated overnight at 4°C with rabbit polyclonal TIGR/MYOC antibodies. ³⁵ Immunoreactive bands were visualized by chemiluminescence (ECL Plus; Amersham Pharmacia Biotech, Piscataway, NJ) after treatment with anti-rabbit IgG secondary antibodies conjugated to horseradish peroxidase (1:8000; Pierce, Rockford, IL). For the loading control, blots were reprobed with a mouse monoclonal β-tubulin antibody (clone TUB 2.1, 1:1000; Sigma) and visualized with conjugated horseradish peroxidase under the same conditions. Gels were run with equivalent volumes of the total cell extract samples and their corresponding supernatants. 

Recombinant Adenoviruses and their Delivery to Tissue Culture Cells

Replication-deficient recombinant adenoviral expression vectors encoding full-length TIGR/MYOC (AdhTIG3) and a truncated mutant TIGR/MYOC (AdhTIG1, Thr345STOP) were constructed and purified as previously described. ¹⁵ For gene delivery, cell cultures were grown to confluence in 35-mm wells, washed twice with PBS, and exposed to 25 to 100 plaque forming units (pfu)/cell of either AdhTIG3 or AdhTIG1 for 1 hour in 1 mL serum-free medium. After exposure to the virus, 1 mL IMEM-5% FBS was added to the medium and incubation continued for 48 hours. Proteins from the medium and cell monolayer were then extracted and analyzed as described earlier. Control dishes were treated with the same volume of viral vehicle under identical conditions. 

Gene Microarray Analysis

Early passages of HTM cell lines HTM-40 and HTM-41 and two separate dishes of 00-1LLS ONH astrocytes were treated with 10⁻⁷ M DEX for 6 to 8 days. Total RNA from each of the four experiments was extracted as described earlier. The targets for the DNA microarray analysis were prepared according to the manufacturer’s instructions (Human Genome U95Av2 GeneChip microarrays, n = 6; Affymetrix, Santa Clara, CA) at the Duke University Microarray Facility. These microarrays contain oligonucleotide probes representing 12,627 genes, in which each gene is represented by a probe set consisting of 16 to 20 probes. Arrays were hybridized to the targets at 45°C for 16 hours and then washed and stained, by using the staining station according to the manufacturer’s instructions (GeneChip Fluidics; Affymetrix). DNA chips were scanned (GeneChip Scanner; Affymetrix), and signals obtained by scanning were processed by the accompanying software (Microarray Suite, ver.4.0; Affymetrix). The absolute analysis determines whether transcripts represented on the array are detected in the sample (transcripts present, marginal, or absent). It uses a variety of metrics, including background and noise calculations, based on the hybridization intensities measured by the scanner. The comparison analysis determines the relative change in abundance for each transcript between a baseline and an experimental sample (fold change). It also uses a variety of metrics, including normalization and scaling factors, based on the intensity differences between each baseline and experimental probe set. The program normalizes expression values against the internal control and eliminates genes absent in both chips from the comparisons. Listed fold change values of ±2 are significant. 

To search for genes located at chromosomal regions linked to glaucoma, each of the selected genes was chromosome mapped using the Locus Link database (http://www.ncbi.nlm.nih.gov/LocusLink/). Each of the obtained chromosomal regions was then manually matched for glaucoma by using known glaucoma-associated loci acquired in the OMIM Morbid Map database (http://www.ncbi.nlm.nih.gov/Omim/searchmorbid/, provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). 

Relative Quantitative Reverse Transcription–Polymerase Chain Reaction (RQ RT-PCR)

Reverse transcription (RT) reactions were performed with a kit (Retroscript; Ambion). One microgram total RNA was mixed with 2 μL of 50 μM random decamers and nuclease-free water in a total of 12 μL, heated at 75°C for 2 minutes, and cooled on ice. The reaction was continued in a 20-μL reaction mixture containing 10 μL of 10× RT buffer (500 mM Tris-HCl [pH8.3], 750 mM KCl, 30 mM MgCl₂, and 50 mM DTT), 4 μL dNTPs (2.5 mM each), 1 μL RNase inhibitor (10 U), and 1 μL Moloney murine leukemia virus (M-MuLV) reverse transcriptase (100 U). The RT reactions were incubated at 42°C for 60 minutes and terminated at 92°C for 2 minutes. A 2.5-μL aliquot of RT reaction was used for amplification by PCR, which was performed in a 50-μL reaction mixture containing 5 μL 10× PCR buffer (400 mM Tricine-KOH [pH 9.2], 150 mM KOAc, 35 mM Mg(OAc)₂, and 37 μg/mL bovine serum albumin), 4 μL dNTPs (2.5 mM each), 4 μL of the gene primer-specific mix (5 μM each forward and reverse primer) and 1 μL cDNA polymerase mix (Advantage; Clontech, Palo Alto, CA). The PCR mixture was split into four tubes, and amplification was performed as follows: 94°C, 4 minutes and cycles of 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 60 seconds. For each primer pair, the linear range of the PCR reaction was determined by varying the number of cycles in each of the four tubes. RT-PCR products were electrophoresed on 2% agarose gels containing 25 ng/mL ethidium bromide. To eliminate the possibility of amplifying genomic DNA, primer pairs were chosen to span intron–exon splice boundaries, RT reactions were conducted in parallel omitting the reverse transcriptase enzyme, or both. An additional control was performed by incubating an aliquot of the PCR mixture without the RT template. 

Quantification of the PCR products was accomplished by the use of 18S rRNA internal standards. 18S rRNA was amplified with a primer-competimer set (QuantumRNA; Ambion) yielding a band of 489 bp. Because 18S rRNA is far more abundant than the mRNA of any gene under study, the 18S amplification reaction is reduced by the addition of competimers. These primers are modified to block extension by DNA polymerase. When combined with the functional primers, they compete for annealing to the 18S rRNA and allow amplification of a reduced number of 18S rRNA molecules. To determine the correct ratio of primers to competimers to use with a given gene, a 2.5-μL aliquot of the RT reaction was used in a 40-μL reaction mixture containing 5 μL 10× PCR buffer (400 mM Tricine-KOH [pH 9.2], 150 mM KOAc, 35 mM Mg(OAc)₂, and 37 μg/mL bovine serum albumin), 4 μL dNTPs (2.5 mM each), and 1 μL cDNA polymerase mix (Advantage; Clontech). The mixture was split into five 8-μL tubes each containing 1 μL of gene-specific primers (5 μM each) and 1 μL of different 18S primer-competimer ratios (usually 1:9, 2:8, and 3:7) plus controls. PCR amplification was performed at the determined linear cycle for the gene under study in the same conditions, to yield multiplex PCR products and to allow selection of the linear range of the 18S rRNA. 

For relative quantification, reactions for each gene were repeated in triplicate in conditions in which the multiplex PCR products were all in the linear range. Band intensities were captured (ChemiDoc System, including a Chemi-cooled charge-coupled device camera, PCI digitizing image acquisition board, EpiChemi II Darkroom with transilluminator, and LabWorks Software; all from UVP, Upland, CA), and transferred to a spreadsheet (Excel, Microsoft, Redmond, WA) for calculation of averages and standard errors. 

Results

TIGR/MYOC Induction by DEX Is HTM Specific

Total RNAs from control and DEX-treated cells corresponding to the different tissue types were analyzed by Northern blot hybridization with full-coding TIGR/MYOC cDNA and 28S control probes. Figure 1A shows the photograph of an autoradiogram from representative Northern blots containing control and DEX-treated samples. Hybridization to control, 10-day confluent HTM cells revealed the presence of a single TIGR/MYOC mRNA band of approximately 2.0 kb (Fig. 1A) . HTM treatment with 10⁻⁷ M DEX for the same length of time showed a marked increase in TIGR/MYOC signal in the two previously described mRNA bands. ³⁶ However, hybridization of total RNA from embryonic SkMCs, HeLa cells, and HUVECs revealed no detectable levels of TIGR/MYOC mRNA, either with or without DEX treatment. Reprobing of the blots with 28S showed that the lack of detection of TIGR/MYOC mRNA was not due to a breakdown of total RNA. Because of the poor RNA recovery of the DEX-treated HeLa cells, the noninduction of TIGR/MYOC in this cell line was reconfirmed using RT-PCR (not shown). TIGR/MYOC RNA was not detected in primary ONH astrocytes by microarray analysis. The absolute values for this gene in both uninduced and DEX-treated ONH astrocyte GeneChips resulted in “absent” calls. 

Western blot analyses of protein extracts from control and DEX-treated cells were performed using TIGR/MYOC and β-tubulin antibodies. Results are shown in Figure 1B . With TIGR/MYOC antibody, blots from HTM protein extracts showed a 55-kDa doublet protein, corresponding to glycosylated and nonglycosylated forms of TIGR/MYOC. ¹⁵ Treatment with 10⁻⁷ M DEX for 10 days resulted in a significant increase of TIGR/MYOC in HTM cells. However, analysis of protein extracts from control and DEX induced HeLa cells, HUVECs, and SkMCs, as well as ONH astrocytes, revealed faint TIGR/MYOC cross-reacting bands. In ONH astrocytes, the TIGR/MYOC protein band appears to be less intense in the DEX-treated cells. The cross-reacting band in the SkMC extracts had a slightly faster electrophoretic mobility than TIGR/MYOC. Although it is possible that some of the faint bands observed in the nonHTM cell extracts are nonspecific, a positive control HTM-DEX sample that was run in all gels showed only TIGR/MYOC at the corresponding size. Reprobing of the blots with β-tubulin antibody controlled the loading of the samples. These observations demonstrate that the TIGR/MYOC gene is induced with DEX only in the HTM cells. 

Processing and Secretion of TIGR/MYOC in the Different Cell Types

Regardless of a tissue’s ability to produce TIGR/MYOC natively, we wanted to determine whether the cell lines in which TIGR/MYOC was not induced by DEX possessed the intrinsic processing and trafficking machinery necessary for the posttranslational modification and secretion of TIGR/MYOC. HeLa cells, HUVECs, and SkMCs were grown to 70% to 80% confluence and infected with replication-deficient adenoviral expression vectors for both wild-type TIGR/MYOC (AdhTIG3) ¹⁵ and a truncated mutant TIGR/MYOC (AdhTIG1) ¹⁷ previously shown to be retained intracellularly in HTM cells. ¹⁷ Western blot analysis of equivalent fractions of infected pellets and supernatants revealed that in all cases, the wild-type form of TIGR/MYOC was secreted, whereas the mutant was not (Fig. 2) . As seen previously in HTM cells, ^{15 17 18} the wild-type infected cultures showed a doublet of 55 kDa in all cell types, both intracellularly and secreted into the supernatant. Similarly, when infected with truncated mutant TIGR/MYOC expression vector, all cell lines retained the protein intracellularly. The observation that the mutant TIGR/MYOC is not secreted may indicate that this pathologic mutation interrupts or cripples a common processing pathway. ^{17 18} Together, these results indicate that although the DEX induction of TIGR/MYOC is not present in other cell types, the glycosylation and secretion mechanisms for this protein are ubiquitous among the different body tissues studied. 

Microarray Comparison of Genes Present in DEX-Induced ONH Astrocytes and HTM

From our RNA and protein studies we found that the DEX response of TIGR/MYOC appears to be TM tissue-specific. To identify other TM factors that could play a role in the TIGR/MYOC secondary glucocorticoid response and/or be involved in the DEX-triggered increase of resistance, we used gene microarrays. We compared expression profiles of DEX-induced ONH astrocytes (baseline) versus DEX-induced HTM (experimental) using microarrays (U95Av2 GeneChips, n = 6; Affymetrix). We searched for genes overexpressed in the HTM-DEX in comparison with ONH astrocytes-DEX. These HTM-overexpressed genes included genes from three groups: those upregulated by DEX in HTM and not upregulated (or less upregulated) in ONH astrocytes; those that were present in HTM and not in ONH astrocytes (tissue specific genes), and genes that were downregulated by DEX in ONH astrocytes and not downregulated (or less downregulated) in HTM cells. 

Four different experiments—two using one human ONH astrocytes cell-cultured line (00-1LLS; denoted A1DEX and A2DEX) and two using two HTM cell-cultured lines (HTM40 and HTM41; denoted H1DEX and H2DEX)—were treated with 10⁻⁷ M DEX for 6 to 8 days and their RNA extracted. Each of the four treated RNAs and two untreated controls (A1 and H1) were then hybridized to individual microarrays (n = 6) and analyzed as described in the Materials and Methods section. Individual outputs from each of the gene chips were compared in sets of two. To identify genes that were specifically overexpressed in the HTM-DEX sample, ONH-DEX astrocytes were used as a baseline. A total of four paired comparisons were performed: A1DEX versus H1DEX, A1DEX versus H2DEX, A2DEX versus H1DEX, and A2DEX versus H2DEX. Additional comparisons using the untreated control (H1 versus H1DEX and A1 versus A1DEX) were also performed. 

The number of genes with intensities in the chips that resulted in absolute values classified as present (P) by the manufacturer’s algorithms (Affymetrix) were calculated for the A1DEX and H1DEX chips. From the 12,627 genes measured on each microarray, we found that 6099 genes were expressed in the A1DEX cell line, whereas 4751 genes were expressed in the H1DEX. Within the set of genes expressed in H1DEX-treated cells, 458 genes were not expressed at any detectable levels in A1DEX-treated cells. We called these genes “turned on” in H1DEX. Similarly, 1806 genes that showed expression in the A1DEX astrocytes were not detected in H1DEX and were designated “turned off” in H1DEX. A total of 4292 genes showed expression in both induced cell types. 

Genes with Expression Levels Changing More Than 20-fold in Comparing ONH Astrocytes-DEX with HTM-DEX

To identify those genes with expression levels that were the most changed (either most increased or most decreased) in HTM-DEX compared with ONH astrocytes-DEX cells, we used fold change analysis. Fold change values were calculated as the ratio of the mean expression levels of each gene between HTM and ONH astrocyte DEX-treated cell cultures. We identified 503 genes that were overexpressed and 1227 genes that were underexpressed in the HTM-DEX cells. Those genes with expression levels that changed by more than 20-fold in the HTM-DEX cells compared with ONH astrocyte-DEX are shown in Table 1 . All but one of the 40 genes showing a more than 20-fold increase were genes of known function. Four of the 18 genes with expression decreased more than 20-fold were unknown expressed sequence tags (ESTs). Very few if any of these genes were housekeeping genes and at least one fourth of them seemed to be involved in extracellular functions. 

To reduce the number of genes for further analysis, we focused on those with the most increased expression. To do so, we searched for those genes showing the highest fold change values not only in one, but in all four ONH-DEX (baseline) versus HTM-DEX (experimental) comparisons (as described earlier). Because of the wide variability of fold change values among the four comparisons, we established a stringent criterion. We selected only those genes with fold change values of five or higher in at least three of the four comparisons (fold change values of 2 are already significant). By implementing this filter, we identified 15 genes. These 15 genes along with the fold change values in each of the four DEX comparisons were designated The Most Overexpressed HTM-DEX–Specific Genes and are listed in Table 2 . 

Classification of the Most Overexpressed HTM-DEX–Specific Genes

The genes identified on Table 2 were categorized into three groups defined as follows: genes induced by DEX in HTM and not induced (or less induced) in ONH astrocytes (group a), genes present in HTM and not present in ONH astrocytes (group b, tissue specific genes) and genes downregulated by DEX in ONH astrocytes and not downregulated (or less downregulated) in HTM (group c). 

To categorize the genes from Table 2 into these three groups, we made use of the comparison analysis obtained with the microarrays from HTM uninduced versus HTM-DEX, ONH astrocytes uninduced versus ONH-DEX, and HTM uninduced versus ONH astrocytes uninduced (H1 versus H1DEX; A1 versus A1DEX; A1 versus HTM). We found that most genes (12 out of 15) belonged to group a. Of these, nine were induced by DEX in HTM but not in ONH astrocytes, and three (IGFBP2, AKR1C3, and AP1G1) were induced by DEX in both cell lines (higher in HTM). Two genes belonged to group b. Of these, one (AKR1C1), was overexpressed in uninduced HTM versus uninduced ONH astrocytes (tissue specific) and further overexpressed by DEX in the HTM cells but not in ONH astrocytes; the second gene (ASS) was only HTM tissue–specific and its expression was not induced by DEX. Only one gene belonged to group c. This gene, PCOLCE, was severely downregulated in ONH astrocytes treated with DEX but showed no change with DEX in HTM cells. 

In addition to TIGR/MYOC, 8 of the 15 genes differentially overexpressed in HTM-DEX encode secreted glycoproteins. These include: α1-antichymotrypsin (ACT) a serine protease inhibitor (serpin) involved in neurodegenerative diseases ³⁷ ; pigment epithelium-derived factor (PEDF), an RPE-secreted glycoprotein with a neuroprotective role ³⁸ ; type I procollagen C-proteinase enhancer (PCOLCE), a procollagen binding protein that affects synthesis of collagen type I ³⁹ ; human cartilage protein GP-39, a secretory product of chondrocytes, highly abundant in cartilage and a marker for joint disease ⁴⁰ ; cornea-derived transcript 6 (CDT6), involved in increased ECM deposition and antiangiogenesis ⁴¹ ; prostaglandin D2 synthase (PGD2S), responsible for the conversion of prostaglandin H₂ into prostaglandin D₂ (PGD₂), a hormone that inhibits release of nitric oxide (NO) and induces vasodilatation ⁴² ; aldoketoreductase 3 (AKR1C3), an enzyme that catalyzes the conversion of PGD₂ to prostaglandin F₂ (PGF₂); ⁴³ and finally, apolipoprotein D (APOD), a lipocalin shown to accumulate in the cerebrospinal fluid of patients with neurologic diseases, such as Alzheimer, multiple sclerosis, and schizophrenia. ^{44 45}  

The nonsecreted proteins identified by our criteria include metabolic enzymes, (AKR1C1 and ASS), an actin (ACTG2), and proteins involved in mitochondrial transport (CTP) and trafficking (AP1G1). 

RQ RT-PCR Confirmation of Differentially Expressed Genes

Five selected genes identified as being differentially expressed according to the parameters of the microarray system (Affymetrix) were further analyzed by RQ RT-PCR, using H1DEX and A1DEX total RNA. The primers used for the amplifications are shown in Table 3 , and the amplified products in Figure 3 . For each of the genes measured by RQ RT-PCR, the mRNA levels were shown to change in the same direction as the changes measured by the microarray detection. However, the fold change values showed a higher dynamic range, a phenomenon also known to occur between relative quantitative PCR and microarrays. ⁴⁶ The expression of TIGR/MYOC and PGD2S was not detected by PCR in ONH astrocytes-DEX, agreeing with the absence call obtained in the arrays. In the TIGR/MYOC panel, the TIGR/MYOC reaction loaded in the gel was diluted 100×. Assuming a minimal number for densitometry detection of the A1DEX band, the change calculation would result in a TIGR/MYOC H1DEX overexpression of approximately 315-fold. Overexpression of α1-antichymotrypsin was calculated to be 54 ± 1.1 higher. In the PEDF panel, the PEDF reaction loaded in the gel was diluted 100×, resulting in a calculated change multiple of overexpression in the HTM of 111 ± 3.5. Last, in the PCOLCE panel, the overexpression of this gene by this method was found to be 1.5 ± 0.1. 

HTM-DEX–Specific Genes Located in Regions Linked to Glaucoma

Because glaucoma develops in some patients treated with corticosteroids, we were interested to know how many of the specific HTM-DEX–upregulated genes would map to chromosomal regions linked to glaucoma. Comparison of A1DEX versus H1DEX revealed that 249 genes were differentially overexpressed in HTM-DEX with fold change values higher than 4. Locus Link database mapping of each of those genes, together with examination of known glaucoma-linked regions listed in OMIM (as of January 24, 2002), produced an overlap of 15 genes (Table 4) . Of these 15 genes, only the TIGR/MYOC gene has thus far been linked to glaucoma. ² Two additional genes, angiopoietin like-factor CTD6 and type I membrane associated glycoprotein T1A-2 are significantly induced by DEX only in the HTM. Optineurin, (labeled as FIP2 in the microarray chip; Affymetrix) a gene recently linked to glaucoma, was not overexpressed in H1DEX versus A1DEX. 

Discussion

Although TIGR/MYOC is expressed in many different tissues in the body, mutations in TIGR/MYOC have thus far been associated only with glaucoma. How mutations in this widely expressed protein cause disease only in the eye is not yet understood. 

Our results showing that altered processing and secretion of the mutant occurring across cell types are an indication that these characteristics, per se, are not responsible for the disease. Current hypotheses on the pathologic nature of TIGR/MYOC focus on the abnormal intracellular accumulation of the mutant protein in the TM cells ¹⁵ and thus in a gain of function. ^{15 19} An intracellular accumulation of TIGR/MYOC mutant protein resulting in a pathologic condition in the TM and not in other tissues could be caused by a tissue-specific overexpression of a gene in response to certain insults. Similarly, the disease could be triggered by the specific induction of a particular cellular trait that would make these cells less tolerable to defective protein accumulation. Studying the differential induction of TIGR/MYOC in the TM could then provide an insight as to what causes glaucoma to develop in patients with mutated TIGR/MYOC. 

In this study, we report that the distinguishing feature of TIGR/MYOC in the TM is its marked induction with glucocorticoid treatment. TIGR/MYOC RNA and protein levels increase in HTM after 7 to 10 days of DEX treatment but do not increase in any of the other tested cells. TIGR/MYOC was also reported not to be induced by DEX on porcine astrocytes. ²⁵ We believe this TIGR/MYOC tissue-specific property may be able to elucidate the linkage of this protein to glaucoma only. 

Having identified the specific TIGR/MYOC induction, we set out to find other TM factors that were present in DEX-induced HTM cells and barely present or absent in other tested DEX-induced cells, including ONH astrocytes, a key cell type involved in ONH remodeling in glaucoma. ²⁶ We reasoned that their specific presence in HTM-DEX could be involved in overexpression of TIGR/MYOC or be independently relevant in steroid-induced elevated IOP. Using microarray technology (GeneChip; Affymetrix), we compared expression profiles of HTM cells and ONH astrocytes induced with DEX. 

In addition to confirming our TIGR/MYOC findings, the microarray studies further identified approximately 600 genes with expression levels in HTM-DEX in comparison with ONH astrocytes-DEX had fold change values higher than plus four or lower than minus four. To narrow our selection, we implemented filters on the fold change expression. Our focus was, rather than to define an actual fold change value, to select those genes that exhibited the highest overexpression. Because of the variability of fold change values observed among the compared cell lines, we applied a stringent criterion and selected only genes with values higher that fivefold (values of twofold are significant) in three of four comparisons. We think this observed variability is mostly due to the distinct response of a given gene to insults in different individuals. ³⁶ Our criteria identified 15 genes, most of which encode secreted glycoproteins. Although we do not know the particular role of these proteins in the TM, their reported known functions appear to have potential relevance to the regulation and maintenance of IOP. 

Because of the important role of the abundance and composition of ECM in the physiology of TM, ^{47 48} an induction of functional serine protease inhibitors such as α1-antichymotrypsin would result in increased ECM deposition and decreased degradation in the TM. Furthermore, α1-antichymotrypsin is associated with formation of amyloid plaques ³⁷ in aged and Alzheimer-affected brains. ^{49 50} It seems possible that the presence of plaques or extracellular aggregates observed in the HTM of steroid-induced glaucoma ⁵¹ is influenced by the HTM DEX-induced expression of this powerful serpin. 

Because of their clear role in other tissues, two of the other glycoproteins identified, CDT6 and human cartilage GP39, may also influence the composition of the ECM in the DEX-induced TM. It is well established that the TM from human anterior segments perfused with DEX, as well as that of patients diagnosed with corticosteroid-induced glaucoma, has a distinct morphology, with thickened trabecular beams, decreased intratrabecular spaces and an increased amount of extracellular materials. ^{51 52} It is further intriguing that CDT6, a relatively new corneal protein, maps to chromosomal region 1p36, where the GLC3B region has been associated with a recessive form of congenital glaucoma. ⁵³ We had previously reported GP-39 as one of the most abundant clones in an HTM library from a normal individual. ⁵⁴ Its specific induction with DEX reinforces the relevance of this protein in the TM. 

The specific induction of PEDF in the TM is intriguing. Its reported protection of neurons from apoptosis, glutamate toxicity, and especially hydrogen peroxide ^{38 55 56} leads us to speculate that perhaps this protein serves a similar protective role in the TM. 

One of the two reported functions of PGD₂ synthase is the formation of PGD₂. A synthetic analogue of PGD₂ that has been shown to reduce IOP in humans. ⁵⁷ PGD₂ synthase has been detected in several ocular tissues and in aqueous humor. ⁵⁸ Recently, the same enzyme has been reported to be overexpressed in ONH astrocytes cultured from glaucomatous eyes. ⁵⁹ It is interesting that the expression of this gene is altered in two glaucoma-implicated cell types and under two distinct glaucomatous conditions. 

In our study, two other genes encoding enzymes that regulate steroid hormone metabolism, AKR1C1 and AKR1C3, were also markedly overexpressed in HTM-DEX cultures. The role of these enzymes, also known as 3α-hydrosteroid dehydrogenases (3α-HSD), ⁶⁰ in the HTM have been studied in detail in relation to steroid-induced elevated IOP. ⁶¹ Abnormal activity of 3α-HSD in the HTM results in abnormal accumulation of 5α- and 5β-dihydrocortisol, which causes elevated IOP in rabbits, whereas their normal product, tetrahydrocortisol, was inactive. ^{62 63 64} More recently, it has been shown that primary cultures of glaucomatous ONH astrocytes exhibit extraordinary upregulation of AKR1C1 and AKR1C3 mRNA, suggesting active steroid metabolism by glaucomatous astrocytes. ⁵⁹ Our current results in HTM-DEX, together with these findings, suggest for the first time a potential association between abnormal steroid metabolism in the HTM, elevated IOP, and glaucomatous optic neuropathy. 

Although this study focused primarily on genes overexpressed in the TM, we have also listed those genes with expression that is most reduced in HTM-DEX, compared with ONH astrocytes-DEX. Of particular interest is the lower expression of elastin in the HTM, further confirmed by the presence of two probe sets in the microarray. This difference appears to be due to a higher abundance of the elastin gene in ONH astrocytes rather than a down- or upregulation of the gene by DEX. Using immunohistochemistry, earlier studies reported no detection of elastin in the HTM cells, with only a moderate induction of DEX. ⁶⁵ In contrast, ONH astrocytes express detectable elastin mRNA in vitro and in vivo. ⁶⁶ Although it has been reported that DEX suppresses elastin gene expression in skin fibroblasts in culture, ⁶⁷ it has also been shown that the elastin promoter contains glucocorticoid response elements that upregulate the expression of reporter genes in transgenic mice. ^{68 69} Other genes that are downregulated in DEX-HTM are associated with the central nervous system; therefore, they are more abundant in astrocytes. Among these genes are STAT1, P311, Pentraxin 1, and Serpin B2. 

Most of the proteins identified in this study are secreted. As such, they have the potential to affect significantly the ECM of the TM. It is interesting to note that the known functions of these proteins point to mechanisms that could have different, possibly opposite, effects on outflow facility. The overexpression of proteins, such as α1-antichymotrypsin, CDT6, and cartilage GP39, seems to suggest a direct increase in ECM deposition and decreased degradation—thus, a risk for potential obstruction of outflow channels. Stimulation of other types of proteins such a PEDF, with an extensive history of its protective role; TIGR/MYOC, a protein potentially involved in protecting TM against elevated IOP ³⁶ ; and perhaps PGD2S, an enzyme involved in the synthesis of an IOP-lowering prostaglandin, points to the induction of different types of defense mechanisms against increased resistance. 

TIGR/MYOC was first identified as a gene upregulated in HTM cells with long-term steroid treatment in culture. The initial excitement surrounding this discovery was the prospect that this gene could be responsible for steroid-induced glaucoma. Establishing the connection between TIGR/MYOC function and pressure regulation has been difficult and remains the object of intense research and controversy. Our data show that in the cell types studied, TIGR/MYOC was significantly induced by DEX in HTM only. By comparing expression profiles, we have identified additional genes that share this tissue-specific induction. This set of genes, though far from complete, confirms at the molecular level the primacy of the role of the ECM in the development of steroid-induced glaucoma. Most important, these genes reveal the presence of new, potentially protective mechanisms that may also become activated in the cells of the HTM after treatment with DEX. Elucidation of which of those mechanisms TIGR/MYOC serves in the meshwork and of the relevance of the genes identified in this analysis remains an important challenge. 

 Supported by the Research to Prevent Blindness, Howard Hughes Medical Institute, National Institutes of Health Grants EY11906 (TB), EY13126 (TB), EY06416 (MRH), and Core Grant P30EY05722 to Duke University Eye Center. TB is a Jules and Doris Stein Research to Prevent Blindness Professor Awardee.

 Submitted for publication May 6, 2002; revised July 23, 2002; accepted August 9, 2002.

 Commercial relationships policy: N.

 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.

 

Induction of TIGR/MYOC by dexamethasone (DEX) in different cell types after 10 days of no treatment (lane C) or 10⁻⁷ M DEX treatment (lane D). (A) Northern blot analysis of total RNA from the indicated cells. Each blot was hybridized sequentially to TIGR/MYOC and 28S probes to control for RNA integrity and equal loading. (B) Western blot analysis of proteins from the indicated cells. Protein extracts were separated on 10% SDS-polyacrylamide gels and transferred to PVDF membrane. Specific proteins were detected with a TIGR/MYOC polyclonal antibody ²⁹ and visualized by chemiluminescence. Blots were reprobed with a β-tubulin antibody for control loading. TIGR/MYOC was induced with DEX only in HTM cells. Induction occurred at transcription and translation levels. ONH-Astr, human ONH astrocytes.

Figure 2.

 

Processing and secretion of TIGR/MYOC in the different cell types. Cell lines were infected with either wild-type (AdhTIG3) or mutant (AdhTIG1, Thr345STOP) TIGR/MYOC adenovirus expression vectors at 25 to 200 pfu/cell. Proteins were extracted from supernatants (Sup) and cellular fractions (Cell) at 48 hours after infection. Equivalent volumes of cell extracts and their corresponding supernatants were run on 10% SDS-PAGE and transferred to PVDF membrane. Specific proteins were detected with a TIGR/MYOC polyclonal antibody ²⁹ and visualized by chemiluminescence. Wild-type TIGR/MYOC (molecular mass doublet at 55–57 kDa) was secreted in every cell type. Mutant truncated TIGR/MYOC (mass ∼43 kDa) was found to be retained intracellularly in every cell type.

Table 1.

 

Genes with Expression Levels That Changed More Than 20-fold, when A1DEX Astrocytes were compared with H1DEX Trabecular Meshwork Cells

GenBank Accession No.	UniGene	Gene Name	Fold Change*	Gene Function
Overexpressed in H1DEX
AF089747	Hs.234726	α1-antichymotrypsin (SERPINA3)	478	Serine protease inhibitor
U05861	Hs.201967	Hepatic dihydrodiol dehydrogenase (AKR1C1)	354	NADH/NADPH aldoketoreductase
U37100	Hs.116724	Aldose reductase-like peptide (AKR1B10)	175	NADPH reductase
Z97171	Hs.78454	Trabecular meshwork inducible gluco. resp. (MYOC)	148	Secreted glycoprotein
D17793	Hs.78183	Aldo-ketoreductase (AKR1C3)	132	Prostaglandin D2 metabolism
M28130	Hs.624	Interleukin 8 (IL8)	100	Chemokine, cell motility, signaling
U29953	Hs.173594	Pigment epithelium-derived factor (PEDF)	87	Serine protease inhibitor, serpin
AF037335	Hs.5338	Carbonic anhydrase XII (CA12)	85	Aqueous humor formation/calcification
L33799	Hs.202097	Procollagen C-proteinase enhancer protein (PCOLCE)	79	Collagen processing
AL050025	Hs.5344	γ1-adaptin (AP1G1)	55	Formation of clathrin pits and vesicles
J02611	Hs.75736	Apolipoprotein D (APOD)	50	Lipocalin, secr glycpr binds hydroph mol
M98539	Hs.8272	Prostaglandin D2 synthase (PTGDS)	49	Prostaglandin biosynthesis
D31628	Hs.2899	4-Hydroxyphenylpyruvic acid dioxygenase (HPD)	49	Tyrosine catabolism
L25879	Hs.89649	Epoxide hydrolase 1 (EPHX1)	43	Detoxification exogenous chemicals
AF017786	Hs.239756	Phosphatidic acid phosphatase (PPAP2B)	41	Phosphatidylinositol (PI) turnover
D00654	Hs.78045	γ2-actin, enteric smooth muscle (ACTG2)	40	Cytoskeletal network
AF015926	Hs.184276	Solute carrier family 9 (SLC9A3R1)	39	Sodium/hydrog exchange/actin organization
A1445461	Hs.351316	Transmembrane 4 superfamily member 1 (TM4SF1)	39	Tetraspanin, regulates basic cell processes
M90657	Hs.351316	Tumor-associated antigen L6 (TAAL6)	39	Tetraspanin, regulates basic cell processes
AL049653	Hs.146559	Angiopoietin-like factor (CDT6)	37	Anti-angiogenesis/ECM deposition
M17017	Hs.624	Interleukin 8 (IL8)	32	Chemokine, cell motility
U09937	Hs.179657	Plasminogen activator urokinase receptor (PLAUR)	31	Signal transduction/ECM remodeling
Y18483	Hs.22891	Solute carrier family 7 (SLC7A8)	31	Amino acid transporter
M25915	Hs.75106	Clusterin (CLU)	29	Plasma glycoprotein
M76665	Hs.37012	Hydroxysteroid dehydrogenase (HSD11B1)	27	Conversion of cortisol to cortisone
L26232	Hs.154854	Phospholipid transfer protein (PLTP)	26	Plasma lipid transfer
X01630	Hs.160786	Argininosuccinate synthetase (ASS)	26	Nitric oxide synthesis/urea cycle
L38486	Hs.118223	Microfibril-associated glycoprotein 4 (MFAP4)	26	Integrin ligand, extracellular matrix
AI936826	Hs.85339	G protein-coupled receptor 39 (GPR39)	25	Signal transduction
AF096870	Hs.194540	Estrogen-responsive B box protein (EBBP)	24	Zinc finger protein
L13698	Hs.65029	Growth arrest-specific 1 (GAS1)	24	Cell growth inhibition
J04164	Hs.146360	Interferon induced transmembrane protein 1 (IFITM1)	22	Cell growth inhibition
U96750	Hs.194695	Ras homologue gene family member 1 (ARH1)	22	Cell growth inhibition
AB003184	Hs.102171	Immunoglobulin superfamily leucine-rich (ISLR)	22	Protein binding, cell adhesion
X14885	Hs.2025	Transforming growth factor β3 (TGFB3)	21	Cell proliferation and differenciation
X04470	Hs.251754	Secretory leukocyte protease inhibitor (SLPI)	21	Protease (chymotrypsin) inhibitor
AF060567	Hs.126782	Sushi repeat protein (SRPUL)	21	Deleted in RP, unknown function
L13698	Hs.65029	Growth arrest-specific 1 (GAS1)	21	Cell growth inhibition
AF027734	Hs.6090	Deleted in bladder cancer chromosomal region candidate 1 (DBCCR1)	21	Tumor suppressor
AI381790	Hs.74120	Adipose-specific 2 (APM2)	21	Unknown/specific to adipose tissue
Underexpressed in H1DEX
AB018301	Hs.22039	KIAA0758 protein (KIAA0758)	−112	Unknown
X52896	Hs.9295	Elastin (ELN)	−110	Elastic fiber component/connective tissue
U61849	Hs.84154	Neuronal pentraxin 1 (NPTX1)	−68	Snake venom-binding protein
Z24680	Hs.151641	Glycoprotein A repetitions predominant (GARP)	−63	Unknown
M18533	Hs.169470	Dystrophin (DMD)	−42	Anchoring cytoskeleton to membrane
U30521	Hs.142827	P311 protein (P311)	−41	Cell transformation and motility
AF045800	Hs.40098	Gremlin/bone morphogenetic protein antagonist1 (CKTSF1B1)	−39	Secreted/blocks BMP signaling
AB023194	Hs.300855	KIAA0977 protein (KIAA0977)	−37	Unknown
D83402	Hs.302085	Prostacyclin synthase (PTGIS)	−35	Potent vasodilator/P450 superfamily
M26326	Hs.65114	Keratin 18 (KRT18)	−35	Intermediate filament component
W27472	Hs.24391	Hypothetical protein FLJ13612 (FLJ13612)	−33	Unknown
Y00630	Hs.75716	Plasminogen activator-inhibitor (SERPINB2)	−30	Inhibitor extracellular proteolysis
M97936	Hs.21486	Transcription factor (STAT1) (ISGF-3)	−29	Activator of transcription
X15998	Hs.81800	Chondroitin sulfate proteoglycan 2 (CSPG2)	−29	ECM glycoprotein/binds to hyaluronan
AL031058	Hs.74316	Desmoplakin I (DPI)	−26	Cytoskeleton/intercellular adhesion
L06139	Hs.89640	Receptor tyrosine kinase (TEK)	−25	Cells of endothelial lineage
X15998	Hs.81800	Chondroitin sulphate proteoglycan 2 (CSPG2)	−25	ECM glycoprotein/binds to hyaluronan
M36860	Hs.9295	Elastin (ELN)	−23	Elastic fiber component/connective tissue

Table 2.

 

The Most Overexpressed HTM-DEX–Specific Genes

GenBank Accession No.	UniGene	Fold Change									Gene Name	Gene Group*	Gene Function
				A1DEX vs. H1DEX	A1DEX vs. H2DEX	A2DEX vs. H1DEX	A2DEX vs. H2DEX	Average
AF089747	Hs.234726	478	125	42	13	165	α1-antichymotrypsin (SERPINA3)	a	Serine protease inhibitor
U05861	Hs.201967	354	69	66	13	126	Hepatic dihydrodiol dehydrogenase (AKR1C1)	b	NADH/NADPH aldoketoreductase
BC004312	Hs.162	6	5	113	102	57	Insulin-like growth factor binding protein-2 (IGFBP2)	a	Modulates IGF function
Z97171	Hs.78454	148	18	30	5	50	Trabecular Meshwork inducible gluco. resp. (MYOC)	a	Secreted glycoprotein
U29953	Hs.173594	87	59	20	18	46	Pigment epithelium-derived factor (PEDF)	a	Neuron survival factor
L33799	Hs.202097	79	60	18	17	44	Procollagen C-proteinase enhancer (PCOLCE)	c	Collagen processing
D17793	Hs.78183	132	27	12	3	43	Aldo-keto reductase (AKR1C3)	a	Prostaglandin D2 metabolism
M98539	Hs.8272	49	65	13	25	38	Prostaglandin D2 synthase (PTGDS)	a	Prostaglandin biosynthesis
Y08374	Hs.75184	12	91	3	24	33	Cartilage GP-39 (CHI3L1)	a	Extracellular matrix remodeling
D00654	Hs.78045	40	57	9	15	30	γ2-actin, enteric smooth muscle (ACTG2)	a	Cytoskeletal network
AL049653	Hs.146559	37	49	10	17	28	Angiopoietin-like factor (CDT6)	a	Antiangiogenesis/ECM deposition
X01630	Hs.160786	26	58	8	18	28	Argininosuccinate synthetase (ASS)	b	Nitric oxide synthesis/urea cycle enzyme
X96924	Hs.111024	31	42	8	12	23	Citrate transport protein (CTP)	a	Citrate mitochondrial transmembrane transport
J02611	Hs.75736	50	20	12	5	22	Apolipoprotein D (APOD)	a	Lipocalin/secreted glycoprotein binds hydrophobic molecules
AL050025	Hs.5344	55	7	21	3	21	γ1-adaptin (AP1G1)	a	Formation of clathrin pits and vesicles, protein sorting, and traffic

Table 3.

 

Primers and Amplification Products Used in RQ RT-PCR Analysis

Gene Name	Accession No.	Sense Primer	Antisense Primer	Size (bp)
TIGR/MYOC	NM_000261	5′-CTG GAG GAA GAG AAG AAG CGA CTA A-3′	5′-CTG TGT CAT AAG CAA AGT TGA CGG TA-3′	912
α1-antichymotrypsin	AF089747	5′-AAG CTC ATC AAC GAC TAC GTG A-3′	5′-ATT CCA AGT TCC TTA CTG AGA GCC-3′	969
PEDF	M76979	5′-GGC TGT CTC CAA CTT CGG CTA-3′	5′-TGG TGA CTT CGC CTT CGT AAC T-3′	811
Prostaglandin D2 synthase	BC005939	5′-CCA ACT TCC AGC AGG ACA AGT TCC-3′	5′-CTT CCG GAG TTT ATT GTG CAB GAT-3′	633
PCOLCE	L33799	5′-CTA ATA AGG AGT GCA TCT GGA CCA TAA-3′	5′-CTG AAG GAG ATT CCT CTG TTT TCT C-3′	691

Figure 3.

 

Confirmation of the differential expression of selected genes in DEX-treated HTM cells and ONH astrocytes by RQ RT-PCR. Amplification of the five selected genes was performed with their specific primers (Table 3) and their predetermined exponential number of cycles (shown underneath each panel). Multiplex PCR reactions for each gene were performed together with an 1:15 primer/competimer mix of 18S RNA in conditions under which the multiplex products were in the linear range. Band intensities were then captured and analyzed. Amplification of each gene was repeated three times. H1DEX cDNA in the first and third panels was diluted 100× because of the high abundance of the target gene. Lane M: ΦX174 RF DNA/HaeIII fragments. The level of expression of all selected genes changed in the same direction as the changes measured by the microarrays.

Table 4.

 

H1DEX Overexpressed Specific Genes Located in Chromosomal Regions Linked to Glaucoma

GenBank Accession No.	UniGene	Gene Name	Fold Change	Glaucoma-Linked Chromosomal Region
Z97171	Hs.78454	Trab Mesh inducible gluco. resp. (MYOC)*	148	1q24.3-q25.2
AI445461	Hs.351316	Transmembr 4 superfamily1 (TM4SF1)	39	3q21-q25
M90657	Hs.351316	Tumor-associated antigen L6 (TAAL6)	39	3q21-q25
AL049653	Hs.146559	Angiopoietin-like factor (CTD6)*	37	1p36
U30255	Hs.75888	Phosphogluconate dehydrogenase (PGD)	19	1p36.2-p36.13
AI660929	Hs.135150	Type-I membrane-asso glycoprotein (T1A-2)*	12	1p36
AF030428	Hs.135150	Type-I membrane-asso glycoprotein (T1A-2)*	11	1p36
D83492	Hs.3796	Ephrin (kinase) receptor protein (EPHB6)	9	7q33-q35
X15414	Hs.75313	Aldose reductase (AR)	9	7q35
D28124	Hs.76307	Human mRNA for unknown product	6	1p36.3-p36.2
M94345	Hs.82422	Capping protein (actin filament) (CAPG)	5	2cen-q24
AF044253	Hs.298184	Potassium channel (KCNAB2)	5	1p36.3
U50648	Hs.274382	Protein kinase, interferon inducible (PRKR)	5	2p22-p21
M64231	Hs.76244	Spermidine synthase (SRM)	5	1p36-p22
M55914	Hs.254105	Enolase 1 (ENO1)	4	1p36.3-p36.2

The authors thank Holly Dressman, Duke University Microarray Facility, and Eric Porterfield for invaluable assistance with the computer analysis of the expression profiles and Jason Vittitow for critical evaluation of the manuscript.