Novel molecular insights into RhoA GTPase-induced resistance to aqueous humor outflow through the trabecular meshwork

Abstract

Impaired drainage of aqueous humor through the trabecular meshwork (TM) culminating in increased intraocular pressure is a major risk factor for glaucoma, a leading cause of blindness worldwide. Regulation of aqueous humor drainage through the TM, however, is poorly understood. The role of RhoA GTPase-mediated actomyosin organization, cell adhesive interactions, and gene expression in regulation of aqueous humor outflow was investigated using adenoviral vector-driven expression of constitutively active mutant of RhoA (RhoAV14). Organ-cultured anterior segments from porcine eyes expressing RhoAV14 exhibited significant reduction of aqueous humor outflow. Cultured TM cells expressing RhoAV14 exhibited a pronounced contractile morphology, increased actin stress fibers, and focal adhesions and increased levels of phosphorylated myosin light chain (MLC), collagen IV, fibronectin, and laminin. cDNA microarray analysis of RNA extracted from RhoAV14-expressing human TM cells revealed a significant increase in the expression of genes encoding extracellular matrix (ECM) proteins, cytokines, integrins, cytoskeletal proteins, and signaling proteins. Conversely, various ECM proteins stimulated robust increases in phosphorylation of MLC, paxillin, and focal adhesion kinase and activated Rho GTPase and actin stress fiber formation in TM cells, indicating a potential regulatory feedback interaction between ECM-induced mechanical strain and Rho GTPase-induced isometric tension in TM cells. Collectively, these data demonstrate that sustained activation of Rho GTPase signaling in the aqueous humor outflow pathway increases resistance to aqueous humor outflow through the trabecular pathway by influencing the actomyosin assembly, cell adhesive interactions, and the expression of ECM proteins and cytokines in TM cells.

glaucoma is a leading cause of blindness worldwide, being commonly associated with elevated intraocular pressure (IOP) and characterized by optic nerve degeneration and progressive visual field loss. In primary open-angle glaucoma, the most common form of the disease, elevated IOP occurs as a result of pathologically increased resistance to drainage of aqueous humor through the pressure-dependent trabecular or conventional outflow system (46). Therefore, elevated IOP is considered as a major risk factor for glaucoma, and lowering IOP is the only course of treatment available for glaucoma (46). The conventional outflow pathway is composed of the trabecular meshwork (TM), juxtacanalicular region (JCT), and Schlemm's canal (SC), and in humans, this pathway represents a predominant route of aqueous humor drainage (20, 38). Aqueous humor is secreted by the ciliary epithelium into the anterior chamber, which then drains through the TM into the SC and the episcleral veins on a continuous basis. Glaucoma is an age-related disease in which ciliary secretion of aqueous humor usually remains normal (12). Therefore, abnormal accumulation of extracellular material/matrix (ECM), which increases resistance to drainage of aqueous humor through the conventional pathway, and changes in contractile activity and cell adhesive interactions of the cells of aqueous outflow pathway is believed to be partly responsible for the elevated IOP and primary open-angle glaucoma (12, 20, 21, 38, 48). However, the cellular mechanisms that regulate the production and turnover of ECM, the contractile activity within the outflow pathway, and how these cellular changes are linked to regulation of aqueous humor outflow through the TM, are far from clearly understood.

The TM beams have been characterized as connective tissue containing elastic and collagen fibers surrounded by endothelial-like trabecular cells resting on a basement membrane (20). The TM cells exhibit a smooth muscle-like phenotype, based on their electromechanical characteristics and expression of various smooth muscle-specific proteins (39, 48). Studies using cytoskeletal-disrupting agents, such as actin depolymerizing agents, inhibitors of myosin light chain (MLC) kinase, myosin II, protein kinase C, Rho GTPase, and Rho kinase, and data derived from both perfusion and in vivo model systems, have indicated a link between cytoskeletal integrity within the TM and aqueous outflow through the TM (12, 31, 38, 39). Agents that increase actin depolymerization and decrease cell-ECM interactions and myosin II phosphorylation in the TM increase aqueous outflow presumably by cellular relaxation and altering the geometry of outflow pathway and fluid flow through the inner wall of SC (6, 12, 31, 39). In contrast to the effects of inhibiting Rho and Rho-kinase activity, myosin II, and MLC kinase, which lead to increased aqueous outflow facility, physiological agonists that activate Rho/Rho kinase signaling, such as endothelin-1, transforming growth factor (TGF)-β, thrombin, and lysophospholipids, reduce aqueous humor outflow in the perfused eye model (21, 22, 48, 50). Additionally, agonists of Rho GTPase activation elicit increases in MLC phosphorylation, actin stress fiber, and focal adhesion formation in cultured TM cells (22, 29). More importantly, the levels of some of these Rho GTPase agonists, including endothelin-1 and TGF-β, have been reported to be elevated in the aqueous humor of glaucomatous patients (21, 23, 40). Although these different lines of evidence implicate the activation status of the Rho/Rho-kinase pathway and the contractile status of the TM in regulation of aqueous outflow and homeostasis of IOP, the precise role of Rho GTPase in the regulation of aqueous outflow is yet to be determined (31). Furthermore, while there is a substantial amount of phenomenological data in the literature regarding the possible role of the ECM and cytoskeletal integrity in modulation of aqueous outflow, an understanding of the mechanistic basis for these interactions has lagged behind (20, 38, 39). Thus identifying the molecular basis by which the ECM, cytosketal integrity, and cellular tension modulate aqueous humor outflow facility through the TM is critical for understanding the homeostasis of intraocular pressure.

Smooth muscle contraction is regulated predominantly by the phosphorylation status of MLC (34). MLC is a regulatory subunit of myosin II, and myosin II ATPase activity and myosin II contraction are stimulated by actin as a function of MLC phosphorylation (34). MLC is phosphorylated by Ca²⁺/calmodulin-dependent MLC kinase and dephosphorylated by Ca²⁺-independent MLC phosphatase, and the balance between these two enzyme activities is a critical determinant of MLC phosphorylation (14, 25). Modulation of MLC phosphorylation can thus occur independently of alterations in cytosolic Ca²⁺ concentration (Ca sensitization) through signaling pathways that regulate MLC phosphatase activity, such as the Rho/Rho kinase pathway (34). When activated by RhoA, Rho kinase inhibits MLC phosphatase activity by phosphorylating the myosin-binding regulatory subunit of MLC phosphatase (11, 41). Rho/Rho kinase has been shown to be involved in the regulation of force or basal tone of smooth muscle, as well as in mediating response to contractile agonists (34). Myosin II cross bridging with actin not only generates force but also influences cell shape, cell adhesion, and cell-cell junctions (11, 47). The Rho GTPases act as molecular switches by cycling between an active GTP-bound and an inactive GDP-bound form. In the GTP-bound form, the Rho GTPase interact with specific downstream effector proteins, which include Rho kinase, regulators of actin polymerization, and adaptor proteins (8). The activity of Rho GTPase is regulated by signaling inputs originating from different classes of cell surface receptors, including the heterotrimeric G protein-coupled receptors tyrosine kinase receptors, cytokine receptors, frizzled receptors, and adhesion receptors (8, 43). Abnormal activity of Rho/Rho kinase signaling has been thought to be involved in the pathophysiology of various age- and smooth muscle-related diseases, including pulmonary hypertension, cardiovascular problems, erectile dysfunction, urinary bladder dysfunction, and stroke (4, 17, 33, 47). Therefore, to test the hypothesis that sustained activation of Rho/Rho kinase signaling in aqueous humor outflow pathway decreases aqueous humor outflow by altering cell/tissue contractile activity, cell adhesive interactions, and gene expression, in this study we investigated the effects of expression of constitutively active RhoA (RhoAV14) on aqueous humor outflow. The data from this study demonstrate that RhoA activation in the aqueous humor outflow pathway decreases aqueous humor outflow by altering the actin stress fibers, cell adhesive interactions, and expression of ECM components and cytokines.

MATERIALS AND METHODS

Amplification of adenoviral vectors.

Well-characterized replication defective recombinant adenoviral vectors (pAdTrack-CMV vectors) encoding either enhanced green fluorescent protein (GFP) alone or GFP and mutant RhoAV14 (Gly14 →Val) were provided by Dr. Patrick J. Casey, Dept. of Pharmacology and Cancer Biology, Duke University School of Medicine. Viral vectors were amplified as described earlier (28).

Organ culture perfusion.

Anterior segments of porcine eyes were prepared as described earlier (28). Anterior segments were perfused at a constant flow (3 μl/min) under 5% CO2 at 37°C, using Harvard microinfusion pumps (Harvard Bioscience, South Natick, MA). Dulbecco's modified Eagle medium (DMEM) containing 0.1% fetal bovine serum (FBS), penicillin (100 U/ml), streptomycin (100 μg/ml), gentamicin (36 μg/ml), and amphotericin (0.25 μg/ml) was used as the perfusion medium. Intraocular pressure was monitored continuously with a pressure transducer connected to the dish's second cannula and recorded with an automated computerized system. Anterior segments were initially perfused with culture medium for 40 h during which the initial baseline pressure was monitored on a continuous basis. After this, adenoviral vectors expressing either GFP alone or RhoA14V/GFP were injected intracamerally in a 100-μl volume containing 5 × 10⁷ plaque-forming units (pfu). After viral vector injections, perfusion was continued at a constant flow (3 μl/min) for an additional 96 h. The effects of RhoAV14 on outflow facility were expressed as the percentage change in outflow facility (compared with baseline values), and a paired two-tailed Student's t-test was applied to determine the significance of differences in outflow facility between control (GFP) and RhoAV14/GFP-expressing anterior segments. After perfusion was completed, anterior eye segments were detached from the perfusion chambers and fixed either for histological analysis or fluorescence analysis, or dissected to obtain TM tissue, which was subsequently stored at −80°C for biochemical analysis.

Cryosectioning and histological analysis.

To determine the distribution and expression of RhoAV14 in aqueous outflow pathway, organ-cultured porcine eye anterior segments infected with adenoviral vectors expressing either the RhoAV14/GFP or GFP alone were fixed in 4% paraformaldehyde prepared in phosphate-buffered saline (PBS) for 24 h, dissected into quadrants, and then stored sequentially (24 h) at 4°C in 5% sucrose-PBS and 30% sucrose-PBS in preparation for cryosectioning. Tissue was then embedded in Tissue-Tek Optimum Cutting Temperature compound (Sakura, Torrance, CA) at −20°C and sectioned (7 μm thickness) using a cryomicrotome (Microm HM 550, Walldorf, Germany). Cryosections were placed on glass slides and stained for cell nucleus and F-actin. Briefly, the cryosections derived from RhoAV14/GFP or GFP-expressing specimens were incubated with 0.3% Triton X-100 in PBS buffer (PBST) for 30 min and blocked with 5% goat serum in PBST buffer, for 30 min at room temperature. The preblocked sections were labeled for 1 h at room temperature with tetrarhodamine isothiocyanate-conjugated (TRITC, 500 ng/ml; Sigma-Aldrich, St. Louis, MO) phalloidin, washed three times with PBST, and treated with a 1:500 dilution of 4′,6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich) for 10 min at room temperature for evaluating cellularity of the trabecular meshwork. After this, tissue sections were rinsed twice with PBST, immersed in 70% ethanol for 5 min, and incubated with auofluorescence eliminator reagent (Chemicon, Temecula, CA) for 5 min to avoid tissue autofluorescence. Tissue sections were then rinsed three times with 70% ethanol (1 min per wash), coverslipped, and mounted using Vectamount (Vector Laboratories, Burlingame, CA) for viewing with a Zeiss Axioplan II fluorescence microscope. The same tissue sections were thus monitored for GFP fluorescence, DAPI staining, and F-actin distribution.

Some of the perfused tissue specimens were also fixed with 2.5% glutaraldehyde and 2% formaldehyde at room temperature, as described previously (28), for light microscopy-based evaluation of histological changes in the outflow pathway.

Cell cultures.

Porcine primary trabecular meshwork cells were isolated from freshly obtained cadaver eyes using collagenase IV digestion, as described previously by us (30). Cells were cultured at 37°C under 5% CO₂, in DMEM containing 10% FBS and penicillin (100 U/ml)-streptomycin (100 μg/ml). All experiments were conducted using confluent cell cultures between 3–5 passages. Human TM cells were isolated from TM tissue extracted from donor eyes (from subjects aged 2 mo, 26, and 54 years) obtained from local eye banks as described above.

Immunofluorescence staining.

The trabecular meshwork cells grown on gelatin-coated glass coverslips were infected with adenoviral vectors encoding either GFP alone or RhoAV14/GFP at 30 multiplicity of infection (MOI). Changes in cell shape were recorded using phase-contrast microscopy (Zeiss IM 35, Thornwood, NY). After an adequate proportion (up to 80% of the population) of cells exhibited positivity for GFP expression (24–48 h after infection), cells were serum-starved for 24 h and then fixed for immunofluoresence analysis of actin cytoskeletal organization and focal adhesions as described previously (30). Actin was stained with rhodamine-phalloidin (Sigma-Aldrich). Focal adhesions were stained with an anti-vinculin primary antibody (Sigma-Aldrich) and TRITC-conjugated secondary antibody (Sigma-Aldrich). Representative micrographs (recorded using a Zeiss Axioplan-II fluorescence microscope) from replicate experiments are discussed in results.

Immunoblot analysis.

Protein concentration of total cell lysates and Triton-X-100 insoluble fractions (27) was estimated using the Bio-Rad protein assay reagent. Samples containing equal amounts of protein were mixed with Laemmli buffer and separated by SDS-polyacrylamide gel electrophoresis (10% or 12.5% acrylamide), followed by transfer of resolved proteins to nitrocellulose membranes. Membranes were then blocked for 2 h at room temperature in Tris-buffered saline containing 0.1% Tween-20 and 3% (wt/vol) nonfat dry milk. Membranes were then probed with anti-vinculin monoclonal antibody (Sigma-Aldrich) and anti-RhoA (Santa Cruz Biotechnology, Santa Cruz, CA) in conjunction with a horseradish peroxidase-conjugated secondary antibodies. Detection of immunoreactivity was performed by enhanced chemiluminescence (ECL).

Trichloroacetic acid protein precipitates derived from TM tissue and TM cells were dissolved in 8 M urea buffer containing 20 mM Tris, 23 mM glycine, 10 mM dithiothreitol, and saturated sucrose by using a sonicator (22). Equal amounts of protein from the urea-solubilized samples were used to determine the levels of phospho-MLC, phospho-paxillin (Tyr118), and phospho-focal adhesion kinase (PY397) by immunoblot analysis using the respective monoclonal or polyclonal antibodies obtained from Cell Signaling (Danvers, MA) and BD Transduction Laboratories (San Jose, CA) as we described earlier (22). Blots were developed using peroxidase-conjugated goat anti-rabbit IgG or anti-mouse IgG and an ECL detection system. Densitometry of scanned films was performed using NIH Image software (National Institutes of Health, Bethesda, MD). Data were normalized to the loading controls. Statistical significance was evaluated using a Student's t-test based on three independent experiments. Glyceraldehyde 3-phosphate dehydrogenase (G3PDH) or β-actin was probed by immunoblotting to confirm equality of protein loading.

Rho GTPase activation assay.

Porcine TM cells infected with adenoviral vectors (30 MOI for 24 h) were serum starved for 24 h, and cell lysates were prepared using 50 mM Tris buffer, pH 7.5, containing 10 mM MgCl₂, 0.5 M NaCl, 1% Triton X-100, aprotinin (10 μg/ml), leupeptin (10 μg/ml), and 500 μg/ml tosyl arginine methyl ester. Cell lysates were then subjected to Rho-GTP pull-down assays to quantify activated Rho GTPase using rhotekin-Rho binding domain (RBD) beads (Cytoskeleton, Denver, CO) and immunoblot analysis, as we described earlier (22).

ECM coating.

ECM-coated plates (OPTILUXTM Petri-dish, Lincoln Park, NJ) or coverslips were prepared by incubating them in polylysine solution (0.01% solution from Sigma-Aldrich) for 2 h at room temperature. After removal of the polylysine solution, the plates were dried and incubated overnight at 4°C with either collagen IV (100 μg/ml), fibronectin (20 μg/ml), or laminin (10 μg/ml, Sigma-Aldrich). Excess ECM solution was removed the next day and the plates were dried at room temperature and rinsed twice with PBS before use.

cDNA microarray analysis.

Human TM cells derived from 2-mo-old donor eyes (passage 3) were cultured to confluence and infected with viral vectors expressing either GFP alone or RhoAV14/GFP for 24 h as described earlier, followed by serum starvation for 24 h. Total RNA was extracted using RNeasy Micro Kit from Qiagen (Valencia, CA). The quality and concentration of RNA was determined using the Nanodrop chip (Wilmington, DL) and the Agilent bioanalyzer (Agilent Technologies, Santa Clara, CA), respectively, and gene-expression data were obtained. All steps involved in RNA processing, probe preparation, microarray hybridization, and data processing were based on Minimal Information about a Microarray Experiment (MIAME), the guidelines established by the Microarray Gene Expression Data Society. Analysis was carried out at the Duke University Genome Analysis Core facility. Oligonucleotide arrays were printed at the Duke Microarray Facility using the Operon Human Genome Oligo Set, Version 3.0 (Operon, Huntsville, AL), which includes 35,000 gene transcripts (optimized 70 mers). GeneSpring GX 7.3.1 software (Agilent Technology, Foster City, CA) was used to perform data analysis. Based on duplicate analyses per sample per condition evaluated, a threshold of twofold change in expression relative to control was considered significant. Data files were submitted to NCBI-GEO and the accession number for these files is GSE9288 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE9288).

RT-PCR and Q-PCR.

Total RNA extracted (RNeasy Mini kit; Qiagen, Valencia, CA) from GFP or RhoAV14/GFP expressing human TM cells was treated with DNAse1 to eliminate contamination from genomic DNA. Purified RNA was quantitated using the Ribogreen RNA assay kit (Invitrogen, Carlsbad, CA). Equal amounts of RNA were then reverse transcribed (Advantage RT-for-PCR kit; Clontech, Mountain View, CA) according to the manufacturer's instructions. Controls lacking reverse transcriptase (RT) were also set up to confirm the absence of genomic DNA-generated signals. PCR amplification was then performed on the resultant TM RT-derived complementary DNA libraries using the sequence-specific forward and reverse oligonucleotide primers for the indicated genes (supplemental material Table 1). PCR cycle numbers were chosen such that the amplified DNA product was within the linear range. The housekeeping gene G3PDH (464 bp) was amplified as an internal control for normalizing the cDNA content of control and RhoAV14-expressing samples.

To obtain further confirmation of differentially expressed genes identified by the cDNA microarray analysis, we performed quantitative real-time polymerase reactions (Q-PCR) for selected genes following the protocol described earlier by us and using the iCycler iQ detection system; Bio-Rad, Philadelphia, PA (5). Supplemental Table 1 lists the primer sets of different genes used in RT-PCR and quantitative PCR. cDNA content in samples was normalized based on expression of an endogenous housekeeping gene G3PDH.

To confirm the reproducibility of differential gene expression induced by RhoAV14 in human TM cells, two additional human TM cell lines derived from 26- and 54-year-old donor eyes were infected with adenoviral vectors encoding GFP or GFP/RhoAV14. Total RNA extracted from these cells was subjected to semiquantitative RT-PCR analysis for selected genes, as described above.

RESULTS

RhoAV14-induced changes in aqueous humor outflow.

The aqueous outflow facility in organ-cultured anterior segments of porcine eyes expressing the RhoAV14 (RhoAV14/GFP) was found to be much lower when compared with fellow-paired control eyes expressing GFP alone (Fig. 1). In control eyes expressing GFP alone, outflow facility (percent change from baseline value) exhibited a gradual increase, to a value of ≈20% over the baseline facility at 50 h, thereafter staying steady at around 15% over the baseline value based on the mean value of 13 individual samples. On the other hand, in the fellow paired-eyes expressing RhoAV14/GFP, aqueous outflow facility was found to stay below the baseline facility throughout the perfusion period. The difference in outflow facility between the RhoAV14/GFP-expressing eyes and GFP-expressing eyes was significant (P < 0.05), starting at 36 h postinfection and continuing up to 96 h of perfusion as shown in Fig. 1.

Fig. 1.

RhoAV14-induced decrease in aqueous humor outflow facility in organ-cultured anterior segments of porcine eyes. Organ-cultured anterior segments of porcine eyes infected with an Ad-RhoAV14/GFP viral vector (5 × 10⁷ pfu) exhibited significant decreases in aqueous outflow facility starting at 36 h postinfection and continuing until 96 h. *A paired two-tailed Student's t-test was used to determine the significance of differences in outflow facility between GFP control and RhoAV14/GFP test samples. Vertical bars represent the standard error associated with the mean values.

To confirm the expression of the recombinant RhoAV14 and GFP in the viral vector-infected perfused eyes, cryosections of organ-cultured anterior segments of the eyes were viewed directly under a fluorescence microscope for GFP fluorescence, following completion of the perfusion period. Intense GFP fluorescence was noted in the aqueous outflow pathway, including the TM tissue from both control (GFP) and RhoAV14/GFP-expressing samples (Fig. 2A). To evaluate aqueous outflow pathway cellularity in organ-cultured perfused eyes expressing GFP or RhoAV14/GFP, tissue cryosections from perfused eyes were labeled with the nucleus-specific fluorescent dye DAPI. The distribution and intensity of DAPI staining was found to be very similar between the GFP- and RhoAV14/GFP-expressing specimens indicating no obvious difference in cell number between the two sets of samples (Fig. 2A, middle). Intense staining for F-actin was also evident in the aqueous outflow pathway including the TM tissue of the RhoAV14-expressing specimens compared with GFP-expressing control specimens, indicating the influence of RhoAV14 on actin cytoskeletal organization (Fig. 2A, bottom). The results shown in Fig. 2 are representative of multiple analyses based on the use of tissue sections derived from four different quadrants of the eye derived from a minimum of three independent eyes, from both GFP- and GFP/RhoAV14-expressing eyes.

Fig. 2.

Expression of RhoAV14 and its influence on actin filament organization and myosin II phosphorylation in the aqueous outflow pathway of organ-cultured porcine eye anterior segments. Organ-cultured anterior segments of porcine eyes infected with adenoviral vectors (Ad-RhoAV14/GFP or Ad-GFP), confirmed expression of GFP in the aqueous outflow pathway. A: both control and RhoAV14-expressing specimens exhibited expression of GFP (green fluorescence) in the aqueous outflow pathway including trabecular meshwork (TM). These data were obtained from perfused anterior segments, at 96 h after viral infection. Nucleus-specific staining using 4′,6′-diamidino-2-phenylindole (DAPI) did not reveal significant differences between the two sets of samples. The RhoAV14-expressing tissue specimens demonstrate increased actin filament staining (rhodamine phalliodin) in the trabecular meshwork compared with the GFP-expressing controls. B: to determine the effect of RhoAV14 expression on the contractile status of trabecular meshwork, TM tissue extracted from the RhoAV14- or GFP-expressing organ-cultured eyes was evaluated for changes in MLC phosphorylation by immunoblot analysis. TM tissue pooled from three perfused eyes showed increased MLC phosphorylation, and data were based on 3 individual samples. In the same samples, the levels of total RhoA were determined by immunoblot analysis and were found to be significantly (P < 0.05) increased in the RhoAV14-expressing specimens compared with the GFP-expressing specimens. Histograms depict the densitometric analysis of immunoblots of MLC phosphorylation and total RhoA. To confirm equality of protein loading, G3PDH was simultaneously analyzed by immunoblotting and the densitometric values were normalized to the GAPDH. Values represent means ± SE from 3 independent observations.

To determine the effects of expression of RhoAV14 on MLC phosphorylation in TM tissue, TM lysates (TM tissue from three perfused eyes were pooled per sample) from eyes infected with either GFP alone or with RhoAV14/GFP were pooled and analyzed for changes in MLC phosphorylation by Western blot analysis using a phosphospecific MLC antibody. RhoAV14 expressing TM tissue exhibited an increase of 25% in MLC phosphorylation compared with GFP expressing control tissue (Fig. 2B) and significantly increased total RhoA protein levels, as evident by immunoblot analysis of the same tissue lysates (Fig. 2B). Data were based on three individual samples and represent the means ± SE (Fig. 2B).

Histological examination of anterior segments from perfused porcine eyes transfected with GFP and RhoAV14/GFP by light microscopy (×20 magnification) revealed only subtle differences in the integrity of trabecular beams between the two types of samples (Fig. 3, top). However, analysis of tissue sections at higher magnification (×100) revealed higher deposition of extracellular material in the TM of the RhoAV14-expressing specimens, relative to that noted in the GFP-expressing controls (Fig. 3, bottom, indicated with arrows). These data were based on using the multiple tissue sections derived from a minimum of three independent specimens from each group.

Fig. 3.

Histological integrity of aqueous humor outflow pathway expressing RhoAV14. Light microscope-based morphological changes in the aqueous humor outflow pathway of porcine eye anterior segments expressing either GFP alone or RhoAV14/GFP. Outflow tissue integrity including that of the TM, appeared to be comparable with no obvious distinction between specimens expressing GFP alone (control) or RhoAV14 (top). However, specimens viewed under higher magnification (×100) exhibited accumulation of some extracellular material (indicated with arrows) in the TM of RhoAV14-expressing samples relative to the corresponding GFP controls (bottom). Representative images (from multiple sections derived from 4 different eye quadrants) obtained from 3 independent specimens from each group are shown here.

RhoAV14-induced changes in actin cytoskeletal and contractile properties of TM cells.

To better understand the cellular basis for the effects of Rho GTPase activation on aqueous outflow facility, we evaluated cellular changes in cultured primary porcine TM cells expressing RhoAV14. Porcine TM cells infected with adenoviral vectors (30 MOI) expressing RhoAV14/GFP revealed obvious changes in cell shape. Twenty-four hours postinfection, more than 80% of the TM cells were found to express GFP in both GFP-expressing control and RhoA14/GFP-expressing test samples. Under conditions of serum starvation (24 h), cells expressing RhoAV14/GFP exhibited stiffened and rounded cell morphology with cellular retraction being evident (Fig. 4A). Additionally, these RhoAV14-induced morphological changes were associated with markedly increased formation of actin stress fibers (phalloidin staining) and focal adhesions (vinculin staining, Fig. 4A). In addition to immunofluorescence evaluation, we also confirmed the effects of RhoAV14 on focal adhesion formation by determining vinculin content in the Triton X-100 detergent-insoluble fraction by immunobloting (27). RhoAV14-expressing cells exhibited increased levels of vinculin in the detergent-insoluble fraction confirming increased vinulin-dependent cell adhesive interactions in TM cells harboring a constitutively active RhoA GTPase (Fig. 4B). The cell lysates derived from the viral vector-infected TM cells were subjected to immunoblot analysis to determine the levels of total RhoA and the levels of activated RhoA by RhoA-GTP pulldown assays using Rhotekin-RBD agarose beads as described in materials and methods. In the RhoAV14-expressing cells, the levels of total RhoA and activated RhoA-GTP were significantly higher than the levels found in GFP-expressing controls (Fig. 4B). Similarly, in the RhoAV14-expressing TM cells, the levels of phosphorylated MLC, the regulatory subunit of myosin II, were markedly increased compared with the GFP-expressing controls (Fig. 4B). All the above described results were based on three individual samples and values were expressed as the means ± SE.

Fig. 4.

RhoAV14-induced effects on cell morphology, actin stress fibers, and focal adhesions in cultured porcine TM cells. A: porcine primary TM cells expressing RhoAV14/GFP (for 24 h) under serum-free conditions (for 24 h) exhibited stiffened, rounded cell morphology compared with the flattened and elongated morphology noted for the GFP-expressing controls. Furthermore, under similar conditions, the RhoAV14-expressing cells revealed increased actin stress fibers (stained with rhodamine phalloidin) and focal adhesions (stained for vinculin) compared with GFP-expressing control cells. Representative images are shown. B: under similar conditions described in A, RhoAV14-expressing cells exhibited increases in membrane localization of vinculin (Triton X-100 insoluble fraction) and MLC phosphorylation based on immunoblot analysis. These changes were correlated with increased levels of activated RhoA as determined by Rho-GTP pull-down assays and total RhoA based on immunoblot analysis. GAPDH or total actin was also immunoblotted to confirm equal protein loading between the GFP control and RhoAV14-expressing samples. Histograms depict the densitometric analysis of immunoblots of vinculin, active RhoA, total Rho, and phospho-MLC between GFP control and RhoAV14-expressing cells, based on 3 individual samples. Statistical differences are the following: *P < 0.05 and **P < 0.01 based on a Student's t-test. Values are means ± SE (n = 3).

RhoAV14-induced differential gene expression profiles in TM cells.

Since Rho GTPase is recognized to regulate many cellular activities besides its well-understood role in actin cytoskeletal reorganization and cell adhesive interactions, we profiled gene expression in RhoAV14/GFP- and GFP-expressing human TM cells by cDNA microarray analysis to obtain further molecular insights into RhoAV14-induced resistance to aqueous humor outflow (Fig. 1). Confluent cultures of human TM cells derived from 2-mo-old donor eyes expressing RhoAV14/GFP or GFP alone for 24 h were serum starved for 24 h, and the total RNA extracted from such samples was subjected to cDNA microarray analysis using spotted arrays containing immobilized oligonucleotide-based probes for a total of ≈35,000 genes. Data presented here were based on a single set of samples. Based on a threefold difference in the gene expression profile between the RhoAV14/GFP-expressing and GFP-expressing control cells, there was an approximate total of 1,632 genes that were upregulated in the RhoAV14-expressing cells (complete list of upregulated genes was provided in supplemental material; Supplemental Table 2). Among the known upregulated genes, we found a significant difference (>3-fold increase) in genes related to five different functional groups: 1) ECM proteins; 2) ECM-regulatory proteins; 3) actin- and cell adhesion-related proteins; 4) integrins; and 5) signaling proteins. Some of the notable genes representing these different functional categories include the following: Perlecan-2, Fibulin-2, brevican, asporin, collagen, fibronectin, laminin, periostin, myocilin, agrin, decorin, lysyl oxidase-like protein-2, microfibrillar-associated protein-5, versican, secreted protein acidic and cysteine-rich (SPARC), and matrix Gla-protein among the ECM-related genes (Table 1). Among the ECM-regulatory genes are the following: TGF-β₂, gremlin-2, BMP-7, latent-TGF-β, IL-11, IL-1β, connective tissue growth factor (CTGF), endoglin, and CTGF-related CNN1 (Table 2). Cytoskeletal and cell adhesion-related genes include the following: contactin-associated protein, ZO-3, heat shock protein-27 (Hsp27), Pinch, contactin 3, ARP2/3 protein, actin, leiomodin 1, talin, tensin, caldesmon, calponin, cofilin, merlin, and α-parvin (supplemental Table 3, A and B). The integrin class includes the following: integrin α-7, E, 3, 2, 1 and β-like, and β-5 (Table 2), whereas the signaling group of genes included the prostacyclin receptor, WNT-11, phosphodiesterase (PDE5A5A), WNK4, phosphodiesterase 4A (PDE4A), guanylate cyclase, myosin light chain kinase (MLCK), RhoGEF, myosin phosphatase, phospholipase-2, regulator of G protein (RGS-5), RGS-3, Abi-2, VAV-3, semaphorin-3, plexin A3, and acidic FGF (supplemental Table 4). In addition to these genes, many genes encoding trafficking proteins were found to be upregulated in RhoAV14 expressing TM cells (see supplemental Table 2). To validate the microarray data, we have also performed semiquantitative and quantitative RT-PCR analysis for representative genes from the different functional groups described above, using total RNA derived from the RhoAV14/GFP- and GFP-expressing cells similar to the cDNA microarray analysis. Figure 5 shows the increased expression of PDE5A, fibronectin-1, collagen-1, myocilin, TGF-β2, Perlecan-2, and CTGF both by semiquantitative (Fig. 5, DNA gels) and quantitative PCR analysis (Fig. 5, fluorescence traces), independently confirming the cDNA microarray results. The quantitative and semiquantitative RT-PCR results were confirmed in duplicates. Expression of G3PDH, a housekeeping gene used as an internal control, showed no detectable difference between the RhoAV14/GFP-expressing and GFP-expressing cells. RhoAV14 expression in human TM cells was also caused downregulation of several genes (≈1,416) by threefold. The downregulated genes, however, unlike the upregulated genes, did not cluster into any particular functional groups (data not shown).

Table 1.

RhoA-induced upregulation of ECM and related gene expression in human TM cells

GenBank Accession Number	Fold Change (Increase)	Description
AF479675	230	Perlecan-2
X82494	71	Fibulin-2
L38956	41	Collagen-α
BC027971	28	Brevican
M55603	22	Collagen Alpha 1(III)
AF117949	15	Lysyl oxidase-like (LOXL2)
AK027359	16	Asporin
D13665	15	Periostin
M11718	14.8	Collagen-α 2
J04217	12.7	Collagen-α2 (IV)
U70313	12.7	EGF-like repeat discoidn rel
AJ005580	11.9	ADAM 23
BC005901	11.7	MAGP-2/ MAGP-5
M99425	9.9	Thrombospondin
M11315	9.7	Collagen-α1
NM054034	8.9	Fibronectin 1
M60832	8.5	Collagen-α2
M22817	8.5	Collagen-α2
AF117949	8.3	Lysyl oxidase-like (LOXL2)
AK027269	8.2	Matrix-remodelling
J04217	8	Collagen-α2
BC020740	7.8	Sarcoglycan
BC005901	7.6	MAGP-2/MFAP-5
BC010456	7.6	EGF containing Fibulin 4
AK097310	7.5	Myocilin
S77512	6.7	Laminin β-2
BC010444	6	Matrilin-2
U60068	5.7	Fibronectin
BC032725	5.6	Fibronectin type III
X62008	5	Fibrillin 1
U84406	5	Agrin
U60068	5	Fibronectin
U60068	4.6	Fibronectin
M27447	4.6	Collagen-α1
AB023177	4.6	Thrombospondin
BC005322	4.5	Decorin
BC013581	4.2	Collagen Alpha 1
X75546	4.1	Fibromodulin
BC008011	3.9	SPARC
BC024189	3.9	Fibronectin
BC018786	3.8	Thrombospondin
AF084545	3.7	Versican
U43328	3.7	Cartilage-link protein
S59334	3.6	Collagen-α5
BC005272	3.3	Matrix GLA-protein
X13939	3.0	Laminin-γ1

ECM, excellular matrix; TM, trabecular meshwork.

Table 2.

Upregulation of ECM regulatory genes in the RhoA V14 expressing human TM cells

GenBank Accession Number	Fold Change (Increase)	Description
AK024848	958	Gremlin 2
AB070218	247	WNT-11 precursor
BC008678	37	IL-1β
M87843	17	TGF-β2
BC004248	17	BMP7
AF034833	15.5	Integrin-α7
Z29665	12.8	Complement Factor H
U70313	12.7	EGF-like repeats
AJ005580	11.9	Disintegrin
NM002208	10.2	IntegrinαE
M99425	9.9	Thrombospondin
S82451	8.8	Latent TGF-binding protein
AF168787	7	Integrin-αE
BC003110	6.7	IL-11
AK024477	4.9	Latent TGF-binding protein
AB008375	4.8	Integrin, β-like
BC001749	4.8	WNT-5B
U60805	4.7	IL-6 family protein
AB023177	4.6	Thromospondin
BC012479	4.5	TNF receptor
BC026262	4.5	TNF receptor
DC01038	4.4	Integrin α-3
BC013316	4.3	IL-1 receptor kinase
AB008375	4.2	Integrin, β-like
BC010241	4.2	TNFR
AB008375	4	Integrin, β-like
AL354866	4	CTGF
BC006541	3.9	Integrin, β-5
BC008011	3.9	SPARC
BC018786	3.8	Thrombospondin
AF512556	3.7	Integrin α-2
BC000097	3.7	TGF-beta-induced βIG-H3
BC034975	3.6	ADAM metallopeptidase
BC013690	3.6	Synaptotagmin X1
L20861	3.6	WNT-5A
U37441	3.4	Endoglin (TGF-related)
BC016952	3.2	CNN1 (CTGF-related)
BC024267	3.1	TNF receptor-associated
BC016952	3.1	CNN1 (CTGF-related)
AF531102	3.1	IL-1 receptor
M38449	3.0	TGF- β1

Fig. 5.

Confirmation of RhoAV14-induced differential gene expression profile in human TM cells by RT-PCR and quantitative real-time PCR. For an independent confirmation of the cDNA microarray-based differential gene expression profiles observed for RhoAV14-expressing human TM cells, total RNA extracted from the human TM cells expressing the RhoAV14/GFP or GFP (24 h postinfection and after a 24-h period of serum starvation) was subjected to RT-PCR and Q-PCR analyses as described in methods using specific oligonucleotide primers for representative genes. Both RT-PCR (ethidium bromide-stained DNA gels) and Q-PCR (fluorescence recording traces) results were found to be consistent with the cDNA microarray results on the selected genes as shown in this figure. Under similar conditions, expression of the housekeeping gene “G3PDH” expression was compared between the RhoAV14/GFP and GFP samples using the same methods. These results were based on duplicate analyses.

In this analysis we have also included a no-treatment control to evaluate the specificity of RhoAV14-induced effects on gene expression compared with the effects induced by GFP expression. When compared with the no-treatment controls, expression of GFP in human TM cells also induced expression of various genes; many of which belong to the stress-related category and encode heat shock proteins (HSP70, HSP105, HSP90 and HSP40), calpains, glutathione S-transferase, β- and γ-crystallins, histones, histone deacetylase, caspase-2, granzymes, mucin, metallothionein, synaptotagmin III, cytochrome P450, p38 mitogen-activated protein kinase (MAPK), JUN3, cathepsin, amyloid-βA4, and ubiquitin. Other genes exhibiting upregulation include those encoding transmembrane helix receptors, sorting nexins, ADAMs, ephrin-A3, α-tubulin, myosin phosphatase, syndecan-3, collagen, laminin b2, WNT3A, WASP-related proteins, Rac GAP, N-CAM, myosin 1A, α-tubulin, fibrinogen gamma chain, BMP-6, MLC2, Matrilin-4, spectrin, and cadherin-3. However, very few genes were commonly induced in cells expressing GFP versus those expressing RhoAV14/GFP.

To confirm the reproducibility of the RhoAV14-induced gene expression profile in TM cells, we also tested the effects of RhoAV14 on selected genes in HTM cells derived from 26- and 54-yr-old donor eyes, by semiquantitative RT-PCR analysis using total RNA. Both of these cell lines exhibited a largely similar response as it relates to the expression profile of selected genes including Il-1, PDE5α, fibronectin, collagen, myocilin, CTGF, TGF-β2, Gremlin, and BMP7 (data not shown).

RhoA-induced ECM accumulation in TM cells.

Among the various causative factors proposed to account for the pathophysiology of glaucoma, accumulation of plaque material and ECM in the aqueous outflow pathway is believed to cause increased resistance to aqueous outflow (3, 12, 20, 38, 50, 51). In support of this, perfusion of matrix matalloproteases in enucleated eyes has been reported to increase aqueous outflow (3). Hence, from our observations related to the RhoAV14-induced upregulation of genes encoding various ECM proteins as shown in Table 1, we directly examined for the effects of Rho activation on the accumulation of selected ECM proteins in cultured human TM cells. Confluent cultures of HTM cells expressing either RhoAV14/GFP or GFP alone and cultured on gelatin-coated glass coverslips for 48 h were serum starved (24 h), fixed, and immunolabeled for collagen IV, laminin, and fiberonectin without permeabilizing the cells, using monoclonal antibodies raised against the respective ECM proteins in conjunction with TRITC-conjugated secondary antibody. Representative immunofluorescence images derived from multiple analyses for each of the specified ECM proteins revealed a marked increase in immunofluorescence specific for collagen IV, laminin, and fibronectin in the RhoAV14-expressing cells compared with GFP controls, indicating increased production of ECM in the presence of increased levels of activated RhoA (Fig. 6). Immunofluorescence images representing the RhoAV14/GFP and GFP controls were captured simultaneously under identical magnification settings.

Fig. 6.

RhoAV14-induced ECM protein accumulation in cultured human TM cells. Human TM cells cultured on gelatin-coated glass coverslips were infected with adenoviral vectors expressing RhoAV14/GFP or GFP alone for 48 h and serum starved for 24 h. After this, cells were fixed with 4% paraformaldehyde and immunostained for collagen IV, fibronectin, and laminin using the respective monoclonal antibodies in conjunction with TRITC-conjugated secondary antibody, as described in materials and methods and in the absence of permeabilization buffer. Representative fluorescence images (based on triplicates) recoded under identical magnification for each of the ECM proteins indicated in the figure are shown for both the RhoAV14/GFP- and GFP-expressing cells.

ECM-induced cytoskeletal and contractile properties in the TM cells.

Since RhoA activation appears to increase the expression and protein levels of various ECM proteins, to understand functional crosstalk between the signaling pathways initiated by the ECM proteins and Rho GTPase activity, we determined the effects of ECM proteins on TM cell shape, cell spreading, actin cytoskeletal organization, cell adhesions, MLC phosphorylation, and Rho activation. Serum-starved porcine TM cells were seeded onto collagen IV, fibronectin, or laminin coated-glass coverslips or plastic Petri plates for 8 h and evaluated for changes in cell shape and cell spreading by viewing them under a phase-contrast microscope. Whereas control cells seeded onto polylysine-coated plates exhibited a strong phase-bright appearance and possessed rounded cell morphology and clustering interactions, cells seeded on the fibronectin, laminin, or collagen IV-coated coverslips exhibited a stretched morphology with a distinctly lesser phase brightness as shown in Fig. 7A. Under these conditions, cells seeded onto ECM (e.g., fibronectin, laminin, and collagen-IV)-coated plates exhibited a dramatic increase in filamentous actin and in focal adhesions when compared with control cells and as determined by phalloidin-labeling or vinculin immunofluorescence labeling (Fig. 7A). Figure 7A depicts representative data from multiple analyses (n = 3). Additionally, cell lysates derived from the individual ECM protein-stimulated TM cells exhibited significantly increased levels of phospho-focal adhesion kinase (Tyr397) and phospho-paxillin (Tyr118) compared with the polylysine-treated control cells (Fig. 7B). These results were based on triplicate analysis. Immunoblotting analysis for phosphorylated MLC, which regulates actin-myosin interactions revealed a significant increase in the levels of phospho-MLC in cells treated with fibronectin, laminin, and collagen, with a much stronger response elicited by laminin and fibronectin relative to collagen IV (Fig. 7C). These changes were also associated with increased levels of activated RhoA in the fibronectin- and laminin-stimulated cells compared with polylysine-treated cells (Fig. 7D).

Fig. 7.

Effects of ECM protein on cell spreading, actin stress fibers, and focal adhesion formation in porcine TM cells. A: to determine the influence of ECM proteins on TM cell shape, actin cytoskeletal organization, and cell adhesion properties, serum-starved porcine TM cells were seeded onto glass coverslips that were precoated with collagen IV, fibronectin, or laminin and cultured for 8 h in the absence of serum. Control cells were seeded onto polylysine-coated glass coverslips. Time-dependent changes in cell morphology and spreading were recorded by phase-contrast microscope. Changes in actin cytoskeletal organization and cell adhesions were determined by staining with rhodamine phalloidin and anti-vinculin, respectively, in both fixed and permeabilized cells. Representative images derived from multiple analyses are shown in this figure. B: ECM-induced increases in focal adhesion kinase and paxillin phosphorylation in porcine TM cells. As described in A, cells seeded onto ECM protein-coated Petri plates were extracted, and cell lysates were immunoblotted for changes in phosphorylation of focal adhesion kinase and paxillin using the respective phospho-specific antibodies. Immunoblots were densitometrically scanned for quantifying changes (histograms). GAPDH was immunolabeled as a loading control. C: ECM-induced increases in myosin light chain phosphorylation in porcine TM cells. As described in B, cell lysates derived from fibronectin-, laminin-, and collagen IV-stimulated porcine TM cells showed significant increase in MLC phosphorylation based on immunoblot analysis and densitometric quantification (histogram). Actin was immunoblotted to confirm loading equivalence between the control and ECM-treated samples. D: ECM-induced increases in Rho GTPase activation in porcine TM cells. As described in B, cell lysates derived from serum-starved TM cells seeded onto fibronectin- and laminin-coated plastic Petri plates revealed marked increases in the levels of Rho-GTP based on pull-down assays, with subsequent immunoblot and densitometric analyses for RhoA. Actin was immunoblotted as a housekeeping protein to confirm loading equivalence between control and ECM-treated samples. Statistical differences are the following: *P < 0.05 and **P < 0.01 based on a Student's t-test. Values were means ± SE (n = 3).

Furthermore, fibronectin, collagen IV, and laminin-induced MLC phosphorylation in TM cells was suppressed significantly by Rho kinase inhibitor-Y-27632, indicating the involvement of Rho/Rho kinase in ECM-induced contractile activity of TM cells (Fig. 8A). In addition to Rho kinase, Src kinase and protein kinase C also appear to participate in ECM-induced MLC phosphorylation as shown in Fig. 8. Inhibitors of Src kinase (PP2) and protein kinase C (GF109203X) both decreased laminin and collagen IV-induced MLC-phosphorylation (Fig. 8, B and c), indicating the participation of more than one protein kinases in mediating ECM-induced MLC phosphorylation. The densitometric values shown in Fig. 8 were based on an average of three independent samples each.

Fig. 8.

Effects of inhibitors of Rho kinase, protein kinase C, and Src kinase on ECM-induced MLC phosphorylation. Serum-starved TM cells cultured on fibronectin, collagen, and laminin coated-glass coverslips were treated with an inhibitor of Rho kinase (10 μM Y-27632), protein kinase C (20 μM GF109203X), or Src kinase (20 μM PP2) for 8 h. Cell lysates derived from these cells were analyzed by immunoblotting analysis for changes in MLC phosphorylation. Densitometric analysis was performed to quantify the changes in MLC phosphorylation (A, B, and C), and results were subjected to a Student t-test for significance. Actin levels were also assessed in the same samples to confirm equality of protein loading for the different samples. Statistically significant differences are the following: *P < 0.05 and **P < 0.01. Values represent the means ± SE from 3 independent observations.

DISCUSSION

The aim of this work was to evaluate the role played by Rho GTPase activity in regulation of aqueous outflow facility through the TM, and the results of this study demonstrate that sustained activation of Rho GTPase activity in the cells of aqueous humor outflow pathway increases resistance to aqueous humor drainage. Furthermore, this increased resistance to aqueous humor outflow by activated RhoA appears to be associated with increased actin stress fibers and cell adhesive interactions, increased expression of various ECM proteins, and cytokines involved in regulating ECM synthesis. Moreover, various ECM proteins were found to stimulate Rho GTPase activity, cell adhesions, and actin stress fibers in TM cells. These observations uncover a potential interaction among ECM protein expression, actomyosin contraction, and Rho GTPase activity, and their feedback influence on aqueous humor drainage through the trabecular meshwork and homeostasis of intraocular pressure.

Recent work based on the use of Rho kinase-specific inhibitors and their ocular hypotensive effects in both in vivo and perfused model systems has offered a new therapeutic possibility to lower intraocular pressure in glaucoma patients by targeting the Rho kinase or Rho GTPase in the conventional aqueous outflow pathway (15, 31, 38, 45). Whereas there is a great deal of enthusiasm to develop potent and safe Rho kinase and Rho GTPase-specific inhibitors for the treatment of glaucoma (37), very little is known about the role of Rho GTPase signaling in aqueous outflow both in normal and glaucoma patients (31). Toward this, in this study, the expression of constitutively active Rho GTPase (RhoAV14) in the aqueous outflow pathway of organ-cultured anterior segments of porcine eyes led to decreases in aqueous humor outflow. Whereas the aqueous outflow facility was increased in a time-dependent manner in GFP-expressing control eyes, reaching up to 20% over the baseline facility and accounting for the washout response in perfused eyes (30), aqueous outflow facility was constantly below the baseline value by a maximum of 10% (Fig. 1) in the RhoAV14/GFP-expressing specimens. One possibility for the aqueous outflow facility not decreasing by more than 10% from the baseline in the RhoAV14-expressing eyes could be that in the normal and healthy porcine eyes used in this study, the outflow facility might not decrease any further with RhoAV14 overexpression or there may be a feedback mechanism that counters the RhoAV14-induced decreases in outflow facility. The other possibility is that the time-dependent increase in aqueous outflow facility that is known to occur in many non-human perfused eyes (washout phenomena), due to either gradual loss of extracellular material or decreased adhesive interactions between innerwall SC and JCT, is probably impaired in RhoAV14-expressing eyes (7, 24). In support of such a possibility, the RhoAV14-expressing eyes exhibit accumulation of extracellular material in the TM as shown in Fig. 3 (bottom, arrows) compared with the GFP-expressing specimens. Further biochemical analyses are required to determine the nature of the extracellular material noted in these specimens. However, the response to Rho activation in aging eyes could be different and severe since age-related diseases have been reported to be associated with increased Rho/Rho kinase activity in tissues with smooth muscle-like properties (4, 17, 33, 47). The RhoAV14-induced resistance to aqueous outflow observed in this study also provide significant molecular insight into the mechanisms harnessed by endothelin-1, TGF-β, and lysophospholipids in decreasing aqueous humor outflow, since these physiological agonists are known to activate the Rho/Rho kinase signaling pathway (10, 22, 31, 48).

Though the molecular basis for increased resistance to aqueous humor drainage in glaucoma is not clear at present, increased cell-cell junction formation between SC cells and accumulation of ECM and its interaction with cells of the juxtacanalicular region and TM are thought to underlie the increased resistance to aqueous outflow (12, 20, 30, 38, 39, 42, 51). It is possible that the constitutively increased Rho GTPase activity in our study might potentially lead to increases in intercellular junctions in the innerwall of SC and thus impair aqueous outflow. However, the expression of activated Rho GTPase in the aqueous outflow pathway stimulated an increase in myosin II phosphorylation and filamentous actin in the cells of the outflow pathway, indicating increased cellular contraction (Fig. 2). Additionally, in TM cell-based studies, we noted increased cell-cell separation, together with increased myosin II phosphorylation in response to RhoAV14 expression (Fig. 4). Activation of Rho GTPase in endothelial cells has been reported to disrupt the permeability barrier due to increased contraction (49), and under such a scenario the aqueous outflow would be expected to increase instead of the opposite response noted in this study. We did not determine the status of cell-cell junctions in the innerwall of SC in the RhoAV14-expressing eyes. However, since SC cells are known to exhibit properties similar to endothelial cells (35), we speculate that cellular mechanisms other than, or in addition to, an altered permeability barrier might play a significant role in RhoAV14-induced resistance to aqueous outflow facility.

To obtain additional molecular insights into Rho GTPase-induced resistance to aqueous outflow, we examined for differential gene expression profile induced by activated Rho GTPase in cultured human TM cells. Interestingly, human TM cells expressing a constitutively active RhoA (RhoAV14) exhibited upregulation of expression of genes encoding different ECM proteins, extracellular proteoglycans, and genes involved in ECM synthesis, cytosketal proteins, integrins, and signaling. The nature of several individual ECM and ECM-regulating genes and other cytoskeletal-regulating genes (Tables 1 and 2, and supplemental material) that were upregulated in the RhoAV14-expressing TM cells strongly suggests that Rho GTPase activation influences ECM synthesis, contractile properties, and cytoskeletal interactions in TM cells. Interestingly, some of the same genes have been reported to have an upregulated expression in the TM and lamina cribrosa cells treated with TGF-β, dexamethasone, mechanical stretch, or increased ocular pressure (2, 18, 21, 38, 44, 50). Additionally, perfusion with, or overexpression of some of these proteins (TGF-β2, fibronectin, Gremlin, and CTGF) in the aqueous outflow pathway has been reported to increase resistance to aqueous outflow (12, 19, 21, 38, 50). Myocilin, a newly identified candidate gene for juvenile primary open angle glaucoma (36), was also found to exhibit upregulated expression upon RhoA activation (Table 1). Although the cellular activity of myocilin is not known, perfusion of myocilin has been reported to increase aqueous humor outflow resistance and alter cell adhesion properties (9, 38).

TM cells binding to ECM components revealed increased actin stress fibers, focal adhesions formation, myosin II phosphorylation, and Rho GTPase activation (Fig. 7), indicating TM cell sensitivity to ECM-induced mechanical strain leading to increased contractile force through Rho GTPase activation. Our data also suggest that the activation status of Rho GTPase influences ECM synthesis and organization in the mechanosensing TM cells (Table 1). In support of this, TM cells expressing the RhoAV14 exhibited accumulation of collagen, fibronectin, and laminin as shown in Fig. 6. As has been reported (13), it is possible that the actomyosin-induced tension/force against the ECM, which is proportional to ECM stiffness, might influence gene expression in TM cells. Conversely, loss of force has been shown to trigger the disassembly of stress fibers and adhesion sites, suggesting the existence of homeostatic regulatory feedback mechanisms between ECM and actin cytoskeletal tension via Rho GTPase activation, which ultimately influence gene expression of ECM proteins (1, 13, 16, 26, 32). Similarly, Rho-mediated actomyosin contraction has been shown to promote assembly of fibronectin into a fibrillar matrix (52). These observations support the involvement of common signaling pathways in modulation of cytoskeletal tension and of ECM synthesis and remodeling (26). ECM-induced MLC phosphorylation appears to be regulated not only by Rho kinase but also via Src kinase and protein kinase C (Fig. 8), indicating an involvement of different signaling pathways acting down stream of ECM and integrins in modulating the contractile properties of TM cells.

In summary, the increased actin stress fibers and cell adhesive interaction via augmented Rho GTPase activation appears to trigger the gene expression of ECM and ECM-regulating proteins in TM cells, and these changes in turn partly influence the resistance of aqueous humor outflow in a feedback response. Further studies involving monitoring intraocular pressure changes and determining the changes in ECM components in aqueous humor outflow pathway in a live animal model overexpressing RhoAV14 might provide significant insights into the regulation of aqueous humor drainage in normal and glaucomatous eyes.

GRANTS

This work was supported by National Institutes of Health R01 Grants EY12201, EY018590 (to P. V. Rao) and P30EY005722 and a grant from the Research to Prevent Blindness (to P. V. Rao) organization.