Regulation of Glucocorticoid Responsiveness in Glaucomatous Trabecular Meshwork Cells by Glucocorticoid Receptor-β

Abstract

purpose. Glucocorticoid administration can lead to increased intraocular pressure in greater than 90% of patients with primary open-angle glaucoma (POAG), compared with 30% to 40% of the general population. The molecular mechanisms for increased steroid responsiveness among patients with glaucoma are unknown. An alternative splicing variant of the human glucocorticoid receptor GRβ has dominant negative activity and has been implicated in a variety of steroid-resistant diseases. GRβ also may play a role in glucocorticoid hyperresponsiveness in glaucoma.

methods. Western blot analysis was performed to detect the expression of GRα and GRβ in TM cells and its regulation by dexamethasone (DEX). Immunocytochemistry was used to compare the subcellular expression of GRβ between normal and glaucomatous TM cell lines. DEX transgene induction in a luciferase reporter was performed to investigate the differential glucocorticoid responsiveness between multiple normal and glaucomatous TM cell lines. Overexpression of GRβ was conducted in glaucomatous TM cell lines, and the regulation of GRβ in the Dex-induced reporter gene luciferase or endogenous myocilin and fibronectin expression were determined.

results. Trabecular meshwork (TM) cell lines derived from normal individuals expressed higher levels of GRβ than did glaucomatous TM cells. Glaucomatous TM cells were more susceptible to DEX induction of a luciferase reporter gene than were TM cells derived from normal donors. Overexpression of GRβ in glaucomatous TM cells inhibited DEX induction of a luciferase reporter gene as well as the endogenous genes MYOC and fibronectin.

conclusions. The decreased amount of GRβ in glaucomatous TM cells could result in enhanced glucocorticoid responsiveness and ocular hypertension.

Glaucoma is a heterogeneous group of optic neuropathies and a leading cause of blindness in the world. ¹ Elevated intraocular pressure (IOP) is a major risk factor for the development and progression of glaucomatous damage to the eye. ^{2 3} This elevated IOP is due to increased aqueous outflow resistance in the trabecular meshwork (TM), ^{4 5} a reticulated tissue located at the corneal–iridial junction that regulates aqueous outflow resistance. 

For several years, glucocorticoids (GCs) have been implicated in the development of glaucomatous ocular hypertension and a secondary open-angle glaucoma that is clinically similar to primary open-angle glaucoma (POAG). ⁶ In both clinical conditions, the elevated IOP is due to increased aqueous humor outflow resistance and is associated with morphologic and biochemical changes in the TM. GCs have a wide variety of effects on TM cells—inhibition of TM cell functions, an increase in the deposition of extracellular matrix material, and alteration of the actin cytoskeleton—many of which may be responsible for increased outflow resistance due to GC treatment. ⁷ Evaluation of the effects of GCs on TM gene expression lead to the discovery of the first glaucoma gene MYOC. Polansky et al. ⁸ and Nguyen et al. ⁹ identified a major secreted glycoprotein that was induced by GCs in cultured human TM cells. This gene mapped to the POAG gene locus GLC1A, and disease-associated mutations were found in MYOC, ¹⁰ although the precise function of MYOC is unknown. It appears that the GC-induction of MYOC expression in TM cells is indirect and not mediated by a specific glucocorticoid-responsive element in the MYOC promoter. ¹¹ Although it has been suggested that increased levels of myocilin are responsible for steroid-induced ocular hypertension and glaucoma, ^{8 9} results in other studies do not support this hypothesis. ¹² In addition to myocilin, a clear association between enhanced deposition of the extracellular matrix glycoprotein fibronectin with POAG and glucocorticoid-induced glaucoma has been documented in the literature. ^{13 14 15}  

GC therapy can lead to elevated IOP, but there are differences in steroid sensitivity among the population. Whereas topical ocular administration of GCs causes measurably increased IOP in approximately 30% to 40% of the general population, ¹⁶ a greater percentage of patients with POAG ^{17 18} and their descendants ^{16 19} have elevated IOP. The molecular basis of the increase in intraocular pressure experienced by patients with glaucoma and individuals receiving glucocorticoids is not well understood. 

Several previous studies have shown that TM cells and tissues are GC responsive and express the ligand-binding isoform of the glucocorticoid receptor GRα. ²⁰ An additional splice variant of the human GR has been identified (GRβ) in which the C-terminal 50 amino acids encoding the GC ligand-binding domain are replaced with 15 different amino acids that do not bind ligand. ^{21 22 23} The alternative splicing of GRβ appears to be regulated by the splicesome protein SRp30c ²⁴ and GRβ mRNA stability determined by a 3′ untranslated region (UTR) single nucleotide polymorphism (SNP). ²⁵ Most of the physiological and pharmacologic effects of GCs are mediated by the major glucocorticoid receptor transcript GRα. ²⁶ However, in vitro experiments have demonstrated that GRβ acts as a dominant negative regulator of GRα function, ^{23 27 28} although there is some controversy. ²⁹ In addition, GRβ may be involved in a variety of GC-resistant diseases, including rheumatoid arthritis, asthma, and inflammatory bowel diseases. ^{25 30 31}  

The purpose of the present study was to determine whether GRβ is expressed in human TM cells and whether altered levels of GRβ explain the differential GC sensitivity in glaucoma. In the study, glaucomatous TM cells had lower levels of GRβ and were more sensitive to GC than were normal TM cells, and increased GRβ suppressed the expression of GC-induced genes in glaucomatous TM cells. 

Materials and Methods

Cell Culture

Eleven human TM cell lines were generated as previously described. ^{13 32 33} Four primary normal TM cell lines (SNTM, SNTM153-00, SNTM302-00, and NTM334-02) and a stable transformed TM cell line (NTM-5) were derived from donors (ages 79, 58, 77, and 18 years, respectively) with no reported history of glaucoma. The five primary glaucomatous TM cell lines (GTM956-99, SGTM152-99, GTM602-02, GTM626-02, and GTM554-99) and a stable transformed TM cell line (GTM-3) were derived from donors (ages 75, 79, 94, 78, 80, and 72 years, respectively) with a documented history of glaucoma. Early passages of TM cells were grown in 37°C and 5% CO₂ in DMEM supplemented with 10% FBS and penicillin, streptomycin, and glutamate (Invitrogen-Gibco, Grand Island, NY). The transformed cell lines GTM-3 and NTM-5 were initially were grown in high-glucose DMEM, whereas primary TM cell lines were grown in low-glucose DMEM. To be certain that the glucose level in the media did not influence the results, the primary TM cells also were tested in high-glucose medium and the transformed TM cells in low-glucose medium. The results were the same, regardless of the type of medium used. 

Quantitative Polymerase Chain Reaction

Total cellular RNA isolation and quantitative PCR (QPCR) were performed as previously described. ^{34 35} QPCR primers for human myocilin, ¹¹ fibronectin, ³⁶ and S15 are shown in Table 1 . The constitutively expressed “housekeeping” gene S15, a small ribosomal subunit protein, served as an internal control. Quantification of relative RNA concentrations was achieved by using the comparative C_T method. ³⁷  

Western Blot Analysis

The cytoplasmic and nuclear fractions were isolated as described previously. ³⁸ Whole cell lysates were prepared in RIPA buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1% Triton X-100, 1% NaDOC, 0.1% SDS, 1 mM EDTA). SDS-PAGE was performed on 4% to 15% gradient gels (Bio-Rad Laboratories, Hercules, CA). A polyclonal anti-GR antibody (SC-1003; Santa Cruz Biotechnology, Santa Cruz, CA), which recognizes the 95-kDa molecule of GRα; the GRβ-specific antibody PA3-514 (Affinity Bioreagents, Golden, CO); anti-myocilin (Ab129) ^{33 39} ; monoclonal anti-fibronectin (Sigma-Aldrich, St. Louis, MO); and the secondary antibodies horseradish peroxidase-conjugated anti-mouse and anti-rabbit IgG (GE Healthcare, Piscataway, NJ) were used. Immunoreactivity was detected by enhanced chemiluminescence (GE Healthcare). Histone1 was used as a control for separation of cytoplasmic and nuclear fractions. 

Immunocytochemistry

TM cells were grown on glass coverslips to confluence, fixed in 4% paraformaldehyde for 30 minutes, permeabilized in 0.2% Triton X-100 for 15 minutes, incubated in 0.2M glycine for 30 minutes, and blocked with 5% bovine serum albumin + 5% normal goat serum for 20 minutes. For immunostaining, the cells were incubated overnight at 4°C with anti- GRβ PΑ3-514 and subsequently incubated with Alexa Fluor 594 goat anti-rabbit IgG (Molecular Probes, Eugene, OR) for 1 hour. DAPI was used to stain nuclear regions. Images were viewed with a fluorescence microscope (Diaphot; Nikon, Melville, NY), and image analysis was performed on computer (IPLab; Scanalytics, Billerica, MA). For immunostaining of GRβ and myocilin, cells were incubated with rabbit anti-GRβ PA3-514 and sheep anti-myocilin, and then incubated with Alexa Flour 633 goat anti-rabbit IgG and Alexa Flour 488 donkey anti-sheep IgG. Confocal immunofluorescent microscopy was performed on a confocal scanning laser microscope system (model LSM-410; Carl Zeiss Meditec, Inc., Thornwood, NY). 

Construction of Human GRβ Vector and Transfection

A pCMX-hGRα expression vector was converted into pCMX-hGRβ by the following steps: (1) generation of a C-terminal PCR fragment that overlaps an EcoRI restriction site on the common GRα-GRβ sequence on exon 9 (forward primer) and has the unique GRβ carboxyl terminal sequence containing a 3′ BamHI restriction site, (2) restriction digestion (EcoRI/BamHI) of the GRβ PCR product and purification of the fragment, (3) restriction digestion (EcoRI/BamHI) of the pCMX-hGRα plasmid and purification of the digested plasmid, (4) ligation of restriction-digested pCMX-hGRα with the GRβ PCR fragment, and (5) DNA sequencing to confirm the GRβ sequence in the plasmid. This GRβ plasmid was transfected into Escherichia coli DH5α to amplify the GRβ plasmid DNA. Confluent TM cells were transfected using a transfection reagent (Lipofectamine) according to the manufacturer’s protocol (BD Biosciences, San Jose, CA). Cells were switched to serum-free medium 24 hours after transfection and treated with DEX for another 24 hours. Approximately 40% of the cells were effectively transfected with the GRβ expression vector, and 5% of the cells died after transfection. 

Luciferase Assays

Cells were transfected with a Mercury luciferase reporter pGRE-Luc (BD Clontech, Palo Alto, CA), as described earlier and in the appropriate figure legends. After DEX treatment, cell lysates were prepared, and luciferase assays were performed with a luminometer. Luciferase activity was normalized with 1 μg of protein for each sample. 

Fibronectin ELISA

Cells were transfected with empty vector pCMX or the GRβ expression vector pCMX-hGRβ and treated with 100 nM DEX in serum-free DMEM for 12, 24, and 36 hours, respectively. Culture medium was collected and secreted fibronectin was determined with a fibronectin ELISA kit (Chemicon International, Temecula, CA), according to the assay protocol. Fibronectin in the culture medium was normalized with 1 μg of cellular protein for each sample. 

Specificity of GRβ Antibody

A number of other investigators have reported results with the same GRβ antibody used in our studies. ^{23 28} Further evidence of the specificity of this antibody in our studies include that the GRβ recognized a lower-molecular-mass isoform of the GR (90 kDa) than did GRα (95 kDa); cells transfected with the GRβ expression vector showed greater GRβ immunostaining and immunoreactivity; and unlike GRα, GRβ did not downregulate expression or translocate to the nucleus after DEX treatment. 

Results

Subcellular Expression of GRα and GRβ and Their Regulation by DEX in Cultured TM Cell Lines

Western immunoblot analyses of the cytoplasmic and nuclear fraction lysates was used to examine the subcellular expression of GRα (Fig. 1A)and GRβ (Fig. 1B)isoforms in cultured primary and transformed TM cell lines in the presence or absence of DEX treatment. Immunoblot analysis for histone 1 was used as an internal control for the separation of cytoplasmic and nuclear fractions. The GRα protein band was detected at approximately 95 kDa in both normal and glaucomatous TM cell lines. Treatment with DEX (100 nM) induced the translocation of GRα from the cytoplasm to the nuclear fractions and also caused a time-dependent downregulation of GRα expression in normal SNTM302-00 and glaucomatous SGTM152-99 cells (Fig. 1A) , as well as all the other normal and glaucomatous TM cell lines studied (SNTM, NTM-5, SGTM956-99, GTM-3, GTM626-02, GTM602-02; data was not shown). 

We detected GRβ doublets in human TM cells by Western Blot analysis (Fig. 1B) . The GRβ identified in TM cells had the same molecular mass of 90 kDa observed in HeLa cell lysates, which were used as a positive control. Downstream in-frame alternative translation initiation sites for the human GR gene have been reported, identifying to two isoforms (GR-A and GR-B). ^{40 41 42} Frequently, our Western blot analyses identified both A and B receptor isoforms for GRα and GRβ. GRβ was present in both the cytoplasm and nuclear fractions. However, unlike GRα, DEX treatment for 3 days did not cause a shift of GRβ between the cytoplasmic and nuclear compartments, nor did it alter the amount of GRβ protein in either normal SNTM302-00 or glaucomatous SGTM152-99 cell lines (Fig. 1B) , as well as all other normal and glaucomatous TM cell lines studied (data not shown). These data demonstrate that the GRβ isoform expression was not susceptible to GCs and persisted in the TM cells despite persistent GC administration. 

Differential Expression of GRβ between Normal and Glaucomatous TM Cell Lines

Patients with POAG are more sensitive to the development of GC-induced ocular hypertension compared with normal subjects. ^{17 18 43} To investigate the potential role of GRβ in the regulation of glucocorticoid sensitivity in glaucoma, the expression and subcellular distribution of GRβ between normal and glaucomatous TM cell lines were compared. Immunocytochemistry detected GRβ staining in both the cytoplasm and the nuclear regions (Figs. 2A 2B) , consistent with our GRβ Western blot data. In addition, 100 nM DEX treatment for 3 days apparently had no effect on the expression or intracellular localization of GRβ in any of these normal and glaucomatous TM cell lines (Fig. 2) . These data confirm that GRβ was located in both the cytoplasm and the nucleus and does not undergo nuclear translocation or downregulation with steroid administration. GRβ immunostaining was relatively high and more concentrated in the nucleus, in four of the five normal TM cell lines (SNTM, NTM-5, SNTM302-00, and SNTM153-00; Fig. 2A ). One normal TM cell line (NTM334-02, from a 6-day-old donor) had low immunoreactivity of GRβ in the nucleus (Fig. 2A) , which may be the result of the very young age of the donor tissue used to derive this cell line. However, in all six glaucomatous TM cell lines (GTM956-99, GTM-3, SGTM152-99, GTM602-02, GTM626-02, and GTM554-99), the amount of GRβ was much lower than in the normal TM cell lines, with GRβ evenly distributed in the cytoplasm and nucleus (Fig. 2B) . DEX treatment for 3 days did not change the amount or distribution of GRβ. 

Western immunoblot analyses of the cytoplasmic and the nuclear fraction lysates were also performed to compare the subcellular expression of GRβ between several primary normal and glaucomatous TM cell lines (Fig. 2C) . With loading of equal amounts of cell lysates, the GRβ protein bands were relatively more intensive in primary normal TM cell lines (SNTM, SNTM302-00, and SNTM153-00), than they were in the primary glaucomatous TM cell lines (GTM956-99, SGTM152-99, and SGTM554-99). This result is consistent with the results of GRβ immunofluorescent staining. 

Differential Responses to DEX Transgene Induction between Normal and Glaucomatous TM Cell Lines

We compared the sensitivities of TM cell lines to GC induction of a GRE-luciferase reporter. Normal and glaucomatous TM cell lines were transfected with a pGRE-Luc reporter construct, and luciferase activity was measuring after treatment with or without DEX for 24 hours. The pSV-β-galactosidase vector was cotransfected as an internal control for monitoring transfection efficiencies. There was no significant difference in β-galactosidase activity among the TM cell lines (data not shown), indicating that there was similar transfection efficiency among the cell lines. However, DEX caused no or a slight (3–4.5-fold) induction of luciferase activity in normal TM cell lines (Fig. 3) , whereas DEX-induced luciferase activity was 8- to 30-fold higher in glaucomatous TM cell lines (Fig. 3) . The greater induction of luciferase activity by DEX in glaucomatous TM cell lines is consistent with the high prevalence of glucocorticoid responsiveness among patients with POAG, and appears to correlate with the lower expression of GRβ in glaucomatous TM cell lines. If GRβ attenuates GRα activity in TM cells, the decreased expression of GRβ in glaucomatous TM cells could be the mechanism responsible for enhanced glucocorticoid responsiveness. 

Effect of GRβ on DEX-Induced Luciferase Activity in Transformed GTM-3 Cells

GRβ can inhibit the transcriptional activity of GRα. ^{23 28} To determine whether GRβ alters GC activity in TM cells, glaucomatous GTM-3 cells were transfected with a pGRE-Luc reporter gene construct and various amounts of GRβ expression vector pCMX-hGRβ. After transfection, the cells were treated with or without DEX for 24 hours and analyzed for luciferase activity. GRβ protein expression was increased in pCMX-hGRβ-transfected GTM-3 cells (Fig. 4B)compared with TM cells transfected with the pCMX empty vector (Fig. 4A) . DEX treatment significantly increased the luciferase activity of pGRE-Luc-transfected TM cells (Fig. 4C) , but increasing amounts of pCMX-hGRβ inhibited this DEX luciferase induction in a dose-dependent manner (Fig. 4C) . Transfection with 0.8 μg of pCMX-hGRβ inhibited luciferase activity by 30%, whereas transfection with 1.5 or 2.0 μg of pCMX-hGRβ totally inhibited DEX-induced luciferase activity. In contrast, DEX treatment did not alter luciferase activity in GTM-3 cells transfected with control vector pTAL-Luc, which lacked the glucocorticoid response element, and overexpression of GRβ had no effect on pTAL-Luc activity (Fig. 4C) . 

Effect of GRβ on DEX-Induced Myocilin Expression in Primary Glaucomatous TM Cells

MYOC is a GC-regulated gene in the TM ^{8 9 33} and was the first glaucoma gene identified. ^{10 44} Because glaucomatous TM cells expressed low levels of GRβ, consistent with an enhanced responsiveness to glucocorticoids in patients with POAG, we tested the effect of GRβ on TM cell myocilin expression by using immunofluorescent confocal microscopy and Western immunoblot analysis. Primary glaucomatous TM cells (SGTM152-99) transfected with pCMX-hGRβ overexpressed GRβ compared with the empty vector pCMX-transfected cells (Fig. 5A) . DEX treatment increased myocilin expression in glaucomatous TM cells transfected with the control vector pCMX. However, the effect of DEX on the expression of myocilin was significantly diminished in GTM cells that overexpressed GRβ (Fig. 5A) . Western Blot analysis also demonstrated increased GRβ in transfected SGTM152-99 cells (Fig. 5B , left, lanes 3 and 4), and DEX treatment (24 hours) had no effect on the amount of GRβ detected (Fig. 5B , left lane 2). DEX treatment (24 hours) significantly increased the expression of myocilin (57 kDa; Fig. 5B , middle, lane 2), however, in the cells that overexpressed GRβ, DEX treatment did not increase the expression of myocilin (Fig. 5B , middle, lanes 3 and 4). QPCR detected that DEX also significantly increased the mRNA expression of myocilin in empty vector pCMX transfected SGTM152-99 cells, whereas overexpression of GRβ blocked the upregulation of myocilin mRNA (Fig. 5C) , indicating that higher expression of GRβ blocked the DEX-induced gene transcription of myocilin. 

Effect of Overexpression of GRβ on DEX-Induced Expression of Endogenous Gene Fibronectin in Transformed GTM-3 Cells

Fibronectin is a major glycoprotein secreted by the TM and incorporated into the extracellular matrix. Accumulation of fibronectin in TM tissue is associated with POAG and glucocorticoid treatment. ^{13 14} In the current study, the effect of GRβ on DEX-induced expression of secreted fibronectin was studied in GTM-3 cells by using a fibronectin ELISA kit (Table 2) . Time-course studies showed significant argumentation of fibronectin in the culture medium after 36 hours of DEX treatment in pCMX-transfected cells. In pCMX-hGRβ-transfected cells, DEX did not change the secretion of fibronectin at any time studied. The effect of DEX for 36 hours on the expression of fibronectin was significantly diminished in pCMX-hGRβ-transfected cells, compared with pCMX empty vector–transfected cells. 

Using Western blot analysis, we also detected the fibronectin in culture medium with a typical molecular mass of 200 kDa (Fig. 6A) . Consistently, we detected that DEX increased the secretion of fibronectin in pCMX transfected GTM-3 cells (Fig. 6A , lanes 1 and 2), whereas in pCMX-hGRβ transfected cells, DEX treatment had no effect on fibronectin secretion, compared with control vehicle treatment (Fig. 6A , lanes 3 and 4). Similarly, QPCR detected that DEX significantly increased the mRNA levels of fibronectin in empty vector–transfected GTM-3 cells (Fig. 6B , lanes 1 and 2). However, with the overexpression of GRβ, there was no significant difference of mRNA levels of fibronectin between DEX and vehicle treatments (Fig. 6B , lanes 3 and 4). This finding indicates that GRβ inhibits the gene transcription of fibronectin induced by DEX. 

Discussion

In the present study, we found decreased expression of GRβ and increased responsiveness to GC in TM cells from patients with glaucoma compared with those of normal individuals. Increased expression of GRβ suppressed the GC induction of a transiently expressed GRE-luciferase construct and decreased the expression of myocilin and fibronectin, which are endogenous DEX-induced, glaucoma-associated genes in TM cells. The low expression of GRβ in glaucomatous TM cells could contribute to the increased responsiveness of patients with glaucoma to glucocorticoids. 

For the past 50 years, there have been several reports implying that glucocorticoids are associated with glaucoma. Elevated cortisol levels in plasma and aqueous humor of patients with POAG have been reported, ⁴⁵ although other studies failed to corroborate these initial findings. ⁴⁶ Certain individuals have significant ocular hypertension when treated ocularly or systemically with anti-inflammatory glucocorticoid therapy. ^{6 7 16 17 18 19} This elevated IOP is due to increased aqueous outflow resistance and is associated with morphologic and biochemical changes in the trabecular meshwork by glucocorticoids. There are differences in ocular steroid responsiveness among the general population, with 4% to 6% showing significantly elevated IOP and approximately 30% with more modestly elevated IOP. ¹⁶ In contrast, almost all patients with POAG have ocular hypertension caused by GC therapy. ^{17 18} In addition, patients with POAG have a greater cutaneous sensitivity to GCs, ⁴⁶ suggesting a more generalized susceptibility to GC therapy. Steroid-responsive nonglaucomatous individuals are at higher risk of development of POAG compared with steroid nonresponders. ⁴³ However, the molecular basis of the increase in IOP experienced by patients with glaucoma and individuals receiving glucocorticoids is not well understood. 

There is considerable controversy about the expression and physiological significance of GRβ. The dominant negative effect of GRβ on GRα’s transactivational activity has been reported, ^{23 27 28 47} although others have questioned the relevance of GRβ based on its low expression relative to GRα. ^{29 48} Northern blot, RT-PCR, and Western blot analyses have been used to measure the ratios of GRα to GRβ in T-cells, HeLa cells, and peripheral blood mononuclear cells. There also are conflicting data on the relative expression of GRα and GRβ proteins in various cells and tissues. Although GRα has been reported to be the predominant isoform in HeLa cells and lymphocytes, ^{29 41} others have shown that the amount of GRβ is equal or higher than that of GRα in HeLa cells, various human tissues, and lymphocytes from GC-resistant patients. ^{49 50} The use of anti-GRα and -GRβ antibodies with different specificity as well as the use of different cells, tissues, and cell fractions may be partially responsible for these discrepant reports on GRβ expression. In some cells, the relative expression of GRβ protein is higher than predicted by the mRNA expression level, ⁵¹ perhaps because of to the longer half-life and stability of nuclear GRβ protein. In addition, GRβ inhibition of GRα transcriptional activity may occur in the nucleus through the formation of transcriptionally impaired GRα-GRβ heterocomplexes or by competing with GRα for GRE binding. ²⁸ The subcellular distribution of GRβ may be more critical to its inhibitory activity. In the present study, GRβ protein immunoreactivity was detectable in both normal and glaucomatous TM cells, whereas the expression of GRβ was higher, particularly in the nucleus, in normal TM cells. Increased expression of GRβ has been reported in several GC-resistant diseases including asthma, ³⁰ rheumatoid arthritis, ²⁵ and inflammatory bowel diseases. ³¹ Higher expression of GRβ in normal TM cells should make them more resistant to the ocular hypertensive effects of GCs, and, conversely, glaucomatous TM cells should be more responsive. 

GCs also appear to differentially regulate the subcellular localization and level of expression of GRα and GRβ. In the absence of ligand, GRα is complexed with heat shock proteins in the cytoplasm. On binding GC, GRα and Hsp90 are translocated to the nucleus along microtubules through dynein motor proteins. ⁵² GRα expression is downregulated after several days of exposure to GC. ⁵³ As in many other cells and tissues, our results showed that DEX treatment of TM cells caused GRα to translocate from the cytoplasm to the nucleus, and several days of DEX treatment also decreased TM cell GRα expression. In contrast to GRα, there are conflicting reports on the subcellular localization of GRβ, which has been reported to be located exclusively in the nucleus ^{23 41} or in the cytoplasm, where it translocates to the nucleus on exposure to GCs. ⁴⁹ We showed that GRβ was located in both the cytoplasm and the nucleus in TM cells, but DEX treatment had no effect on intracellular location of GRβ. It has been reported that Hsp90 can also complex with GRβ. ^{28 29} We have found that Hsp90, immunophilin FKBP51, and the microtubule motor protein dynein are involved in the transport of GRβ from the cytoplasm to the nucleus in TM cells (Zhang X, et al. IOVS 2005;46:ARVO E-Abstract 3687). Unlike GRα, GRβ expression was not downregulated by DEX treatment. 

Normal TM cells showed increased expression of GRβ and were less susceptible to GC induction of the GRE reporter gene. Glaucomatous TM cells, which had an overexpression of GRβ by transfection, mimicked normal TM cells with less responsiveness to GC induction of the GRE reporter gene as well as the endogenous genes fibronectin and myocilin. The higher expression of endogenous GRβ in normal TM cells compared with glaucomatous TM cells may make the normal TM cells more resistant, and conversely the glaucomatous TM cells more responsive, to the physiological and pharmacologic effects of glucocorticoids. This appears to be a unique mechanism for causing disease. In several other diseases, increased GRβ expression is associated not with disease pathogenesis, but instead with resistance to glucocorticoid therapy. ^{25 30 31} Glucocorticoid-induced ocular hypertension in susceptible individuals receiving exogenous GC therapy may be due to acute (weeks to months) effects of glucocorticoids on TM cells that lead to compromised aqueous outflow. Susceptibility of GC-induced ocular hypertension could be due to the levels of GRβ in an individual’s TM cells. Patients with POAG have shown greater ocular response to GCs (GC-induced IOP elevation) in other studies, ^{6 17 18} which is consistent with our current findings. In addition, the enhanced cutaneous GC vascular responsiveness shown in patients with POAG ⁴⁶ may also be due to altered GRβ expression in skin microvascular cells. Even in the absence of exogenous GC therapy, POAG may be at least partially due to the increased susceptibility of the glaucomatous TM cells to chronic exposure (years) to endogenous cortisol levels. 

Several mechanisms may be responsible for higher levels of GRβ in normal than in glaucomatous TM cells. Expression of GRβ could be regulated by alternative splicing efficiency to change the ratio of GRβ to GRα mRNA through variations in splice sites, altered expression of splicing factors, ²⁴ or the presence of functional polymorphisms in splicing factors. GRβ mRNA stability appears to be controlled by a 3′UTR SNP, ²⁵ and this SNP may vary between normal subjects and patients with glaucoma. Genotyping studies are currently in progress to determine whether there are disease-associated polymorphisms in any of these sites in patients with glaucoma. Alternatively, there may be differences in GRβ protein stability in such patients. Differences in the nuclear translocation of GRβ may also explain the differential expression in normal versus glaucomatous TM cells. We have shown that immunophilin FKBP51, but not FKBP52, is involved in the nuclear transport of GRβ (Zhang X, et al. IOVS 2005;46:ARVO E-Abstract 3687). It appears that overexpression of FKBP51 is associated with glucocorticoid resistance. ^{54 55} FKBP51 may differentially regulate the nuclear transport of GRβ between normal and glaucomatous TM cells and lead to the nuclear accumulation of GRβ and glucocorticoid resistance in normal TM cells. 

In conclusion, we have, for the first time, demonstrated that the expression level of GRβ regulates cellular responses to glucocorticoids in TM cells. The low expression of GRβ in the nucleus in glaucomatous TM cells results in the enhanced transcriptional activity of GRα and may contribute to enhanced GC responsiveness and increased IOP in patients with glaucoma. 

 Supported by National Eye Institute Grants EY11979 and EY016242, Alcon Research, Ltd., and the Advanced Technology Program, Texas Higher Education Coordinating Board.

 Submitted for publication May 10, 2005; revised July 20, 2005; accepted October 14, 2005.

 Disclosure: X. Zhang, Alcon Research, Ltd. (F); A. F. Clark, Alcon Research, Ltd. (E, F); T. Yorio, Alcon Research, Ltd. (F)

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

 

Table 1.

 

Quantitative PCR Primer Sequences and Expected Product Sizes

Gene	Sense	Antisense	Size (bp)
Fibronectin	AGAGTGGAAGTGTGAGAG	TTGTAGGTGAATGGTAAGAC	303
Myocilin	GCCCATCTGGCTATCTCAGG	CTCAGCGTGAGAGGCTCTCC	82
S15	TTCCGCAAGTTCACCTACC	CGGGCCGGCCATGCTTTACG	361

Figure 1.

 

Effects of dexamethasone on GRα and GRβ expression in normal and glaucomatous TM cell lines. TM cells were cultured in 10% FBS-DMEM to complete confluence, shifted to 5% FBS-DMEM, and subjected to DEX treatments. (A) Glaucomatous TM cells SGTM152-99 and normal TM cells SNTM302-00 were treated with control vehicle (ethanol) or 100 nM DEX for 1, 2, 3, 4, or 5 days, as labeled. The cytoplasmic (Cyto) and nuclear (Nuc) fractions were isolated. One hundred micrograms of cytoplasmic and 50 μg of nuclear proteins were resolved in SDS-PAGE on 4% to 15% gradient gels. Western blot analysis of GRα receptor was performed by using a polyclonal antibody against glucocorticoid receptor. GRα often was detected as a 95-kDa protein band. Western blot analysis of histone 1 was used as an internal control for identifying separation of cytoplasmic and nuclear fractions and also as a control for equal loading. (B) Glaucomatous TM cell line SGTM152-99 and normal TM cells SNTM302-00 were treated with control vehicle (ethanol) or 100 nM DEX for 3 days. The cytoplasmic and nuclear fractions were isolated. One hundred micrograms of cytoplasmic and 50 μg of nuclear proteins were resolved in 4% to 15% gradient gels. Immunoblot analysis of GRβ isoform was performed by using a specific GRβ antibody. HeLa cell lysate was used as a GRβ-positive control. GRβ often was detected as a 90-kDa doublet.

Figure 2.

 

Differential expression and distribution of GRβ between normal and glaucomatous TM cell lines. (A, B) Immunocytochemistry: normal TM cell lines (A) and glaucomatous TM cell lines (B) were grown on 35-mm coverslips in 10% FBS-DMEM, and then shifted to 5% FBS-DMEM and subjected to control vehicle (ethanol) or 100 nM DEX for 3 days, fixed, and subjected to immunofluorescent staining for GRβ. Cells were incubated with primary anti-GRβ antibody and followed by incubation with secondary Alexa Fluor 594 goat anti-rabbit IgG. Conventional immunofluorescence microscopy was used to detect the GRβ staining. Red: GRβ staining; blue: DAPI staining of the nucleus. The negative control staining for GRβ was performed with the rabbit preimmune serum instead of primary anti-GRβ antibody. Scale bars, 50 μm. (C) Western blot analysis of GRβ for primary normal (left) and glaucomatous (right) TM cell lines. One hundred micrograms of cytoplasmic and 50 μg of nuclear proteins were resolved in 4% to 15% gradient gels. Immunoblot analysis of GRβ was performed. TM cells were treated for 3 days with ethanol control (lane 1: cytoplasmic fraction; lane 2: nuclear fraction) or 100 nM DEX (lane 3: cytoplasmic fraction; lane 4: nuclear fraction). HeLa cell lysate was used as a positive control (lane 5).

Figure 3.

 

Differential responses to DEX in the induction of reporter gene luciferase activity between normal and glaucomatous TM cell lines. Normal (NTM-5, SNTM, SNTM302-00, and SNTM153-00) and glaucomatous (GTM-3, SNTM152-99, GTM626-02, and GTM956-99) TM cell lines were transfected with 0.2 μg of glucocorticoid-responsive mercury luciferase reporter pGRE-Luc in 12-well culture plates. After posttransfection incubation, cells were treated with control vehicle (ethanol) or 100 mM DEX in serum-free DMEM for 24 hours. Cell lysates were then prepared, and luciferase activity was determined. Data are plotted as x-fold change from each basal activation of control vehicle treatment and represent the mean ± SE of results in three independent experiments (*P < 0.05 DEX versus vehicle control, t-test).

Figure 4.

 

Overexpression of GRβ inhibited GRα-mediated activation of the reporter gene luciferase. (A, B) Transformed glaucomatous GTM-3 cells were transfected with 0.4 μg of glucocorticoid-responsive mercury luciferase reporter pGRE-Luc or 1.5 μg of empty vector pCMX or GRβ expression vector pCMX-hGRβ, in 12-well culture slides. Confocal microscopy was used to visualize the increased expression of GRβ after transfection of pCMX-hGRβ. (C) GTM-3 cells were grown in 12-well culture plates and transfected with 0.4 μg of the pGRE-Luc or 0.8, 1.5, or 2.0 μg pCMX or pCMX-hGRβ, as indicated. After incubation with control vehicle (Con) or 100 nM DEX in serum-free DMEM for 24 hours, cells were harvested, and luciferase activity was determined. The TATA-like promoter-luciferase reporter pTAL-Luc does not respond to DEX because of the absence of a GRE and was used as a control vector. Cells were cotransfected with two of the vectors pGRE-Luc, pTAL-Luc, pCMX-hGRβ, or pCMX. Cells were transfected (1) with (+) or without (−) the pGRE-Luc contruct; (2) with (+) or without (−) pTAL-Luc; (3) without (−) or with (in micrograms) the indicated amount of pCMX-hGRβ construct; and (4) without (−) or with (in micrograms) the indicated amount of pCMX vector. Data are plotted as the x-fold change from each basal activation of vehicle treatment and represent the mean ± SE of results in three independent experiments (*P < 0.05 pCMX+DEX versus pCMX+vehicle control; †P < 0.05 pCMX-hGRβ+DEX versus pCMX-hGRβ+vehicle control; t-test).

Figure 5.

 

Overexpression of GRβ inhibited DEX-induced expression of myocilin in primary glaucomatous TM cells. SGTM152-99 cells were transiently transfected with a control vector pCMX or GRβ expression vector pCMX-hGRβ. After transfection, cells were incubated with either vehicle control (ethanol) or 100 nM DEX in serum-free DMEM for 24 hours. The effects of increased GRβ on DEX-induced expression of myocilin were examined by confocal immunofluorescence microscopy, Western blot analysis, and QPCR. (A) Confocal immunofluorescent microscopy. Cells were incubated with primary rabbit anti-GRβ and sheep anti-myocilin antibodies and subsequently with the secondary antibodies Alexa Flour 633 goat anti-rabbit IgG to label GRβ (red) and Alexa Flour 488 donkey anti-sheep IgG to label myocilin (green). (B) Western Blot analysis: SGTM152-99 cells were transiently transfected with a control vector pCMX or GRβ expression vector pCMX-hGRβ. After transfection, cells were incubated with either thevehicle control or 100 nM DEX for 24 hours. Either whole-cell lysates were then subjected to Western blot analysis (B) with anti-GRβ and anti-myocilin or (C) total cellular mRNA was isolated and subjected to quantitative PCR to detect myocilin expression. Lane 1: pCMX+Con; lane 2: pCMX+DEX; lane 3: pCMX-hGRβ+Con; and lane 4: pCMX-hGRβ+DEX. (*P < 0.05 pCMX+DEX versus pCMX+vehicle control; t-test).

Table 2.

 

Fibronectin Released in Culture Medium from GTM-3 Cells

Time Courses of Treatment (h)	pCMX Transfection			pCMX-hGRβ Transfection
		Vehicle	DEX	Vehicle	DEX
12	41.4 ± 1.1	39.7 ± 1.6	39.1 ± 0.16	40.0 ± 2.34
24	53.5 ± 1.96	58.0 ± 0.97	53.5 ± 1.99	52.9 ± 0.82
36	76.5 ± 1.77	87.2 ± 0.92*	76.8 ± 2.36	79.3 ± 1.79, †

 DEX was applied at 100 nM. Data are expressed as the mean ± SEM nanograms per microgram of cellular protein (n = 3). Statistical significance was determined by one-way ANOVA.

*   P < 0.001 pCMX + control versus pCMX + DEX.

†   P = 0.002 pCMX + Dex versus pCMX-hGRβ + DEX.

Figure 6.

 

Overexpression of GRβ inhibited DEX-induced secretion of fibronectin in glaucomatous TM cells. GTM-3 cells were transfected with empty vector pCMX or pCMX-hGRβ and treated with ethanol or 100 nM DEX in SF-DMEM for 36 hours. (A) Culture medium (40 μg) was subjected to Western blot analysis to detect the secretion of fibronectin protein, and (B) total cellular mRNA was isolated to study the mRNA expression using QPCR. Lane 1: pCMX+Con; lane 2: pCMX+DEX; lane 3: pCMX-hGRβ+Con; and lane 4: pCMX-hGRβ+DEX. (*P < 0.05 pCMX+DEX versus pCMX+vehicle control; t-test).

The authors thank Allen Shepard (Alcon Research, Ltd.), for help in generating the GRβ construct pCMX-hGRβ; Shaoyou Chu, Larry Oakford, and I-fen Chen Chang for technical support in confocal microscopy; and Ganesh Prasanna, Raghu Krishnamoorthy, and Santosh Narayan for useful discussions.