Molecular Vision 2007; 13:1363-1374 <http://www.molvis.org/molvis/v13/a150/> Received 8 April 2007 | Accepted 27 July 2007 | Published 1 August 2007

Download Reprint

Functional delivery of synthetic naked siRNA to the human trabecular meshwork in perfused organ cultures

Núria Comes, Teresa Borrás
 
 

Department of Ophthalmology, University of North Carolina, Chapel Hill, NC

Correspondence to: Teresa Borrás, Ph.D., Department of Ophthalmology, University of North Carolina School of Medicine, 6109 Neuroscience Research Building CB 7041, 115 Mason Farm Road, Chapel Hill, NC, 27599-7041; Phone: (919) 843-0184; FAX: (919) 843-0749; email: tborras@med.unc.edu

Abstract

Purpose: To investigate whether naked short-interfering RNA (siRNA) molecules could be directly delivered to perfused intact human trabecular meshwork (TM) tissue, whether this siRNA could silence a trabecular meshwork preferred gene, and whether it could counteract the downstream effect of a deleterious agent (dexamethasone, DEX) by silencing its receptor.

Methods: Anterior segments from post-mortem normal human donors were perfused at 3.4±0.3 μl/min-constant flow or 15 mmHg-constant pressure to stable baseline (outflow facility, C=0.22±0.19 μl/min/mmHg; n=14). Commercial siRNAs were diluted in DMEM (Dulbecco's Modified Eagle's Medium) perfusion medium and used without coupling to transfection reagents ("naked"). Perfusion of Cy3-labeled siRNA was performed at 100 nM for 48 h followed by 24 h with DMEM medium (two pairs). Perfusions of Matrix GLA protein (MGP) siRNA (100 nM; right eye [Oculus Dexter]; OD) and scramble-siRNA (control; left eye [Oculus Sinster]; OS) were performed for 48 h (two pairs). Perfusions of glucocorticoid receptor (GR)-siRNA (OD) and scramble-control (OS) were performed for 48 h and continued by adding 100 nM DEX to the perfusion media for an additional 24 h (two pairs). Frozen sections of labeled anterior segments were analyzed by confocal fluorescence microscopy. Differential expression of GR, MGP, myocilin (MYOC), cornea-derived transcript 6 (CDT6), and 18S genes was determined by reverse-transcriptase TaqMan polymerase chain reacion (RT-TaqMan PCR) on RNA extracted from dissected trabecular meshwork. Primary human trabecular meshwork cells were generated from single individuals and transfected using the nucleofector electroporator with program T-23. Levels of secreted MYOC in the effluents were analyzed by western blot.

Results: Histological evaluation of anterior segments perfused with Cy3 labeled siRNA followed by unlabeled medium showed intense fluorescence in the trabecular meshwork region. MGP gene expression was silenced in the trabecular meshwork perfused with naked MGP siRNA. MGP transcripts were reduced 94.7% ± 0.62 (individual 3) and 93.6% ± 0.13 (individual 4) from those present in the contralateral eye perfused with the scramble control. Pretreatment of GR siRNA followed by DEX treatment caused a reduction of the MYOC and CDT6 gene expressions when compared with eyes pretreated with scramble-control (percent silencing: 99.3% ± 0.005 and 97.3% ± 0.25, respectively, for individual 5 and 98.2% ± 0.06 and 85.6% ± 0.88, respectively, for individual 6). Western blots revealed the decrease of MYOC secreted by GR siRNA-treated cell and organ cultures.

Conclusions: Readily available siRNA can be delivered to the intact human trabecular meshwork by intracameral perfusion. The delivered naked siRNA is functional, inhibiting not only the targeted gene but also their downstream effectors. This functional intracameral delivery might be of use to protect the trabecular meshwork from unwanted insults and could have important therapeutic applications.

Introduction

RNA interference (RNAi) is emerging as a highly effective tool for specific gene silencing. RNAi is an evolutionarily conserved phenomenon of posttranscriptional gene silencing which is triggered by the presence of 21-23 nucleotides, double stranded (ds) RNA molecules called short interfering RNAs (siRNAs). In the cytoplasm, siRNAs from endogenous or exogenous origins interact with a nuclease-containing multiprotein complex called RISC (RNA-induced silencing complex). Upon binding to RISC, the ds siRNAs unwind, pair with their complementary target mRNA, and allow the RISC complex to cleave the mRNA (messenger RNA) strand within the target site. This initial cleavage results in a rapid degradation of the mRNA molecule, which prevents its translation into protein [1].

While RNAi is greatly helping to better understand gene function, it also has great potential as a novel therapeutic strategy to silence unwanted disease-relevant genes. The biggest challenge in the use of siRNA is delivery. In order to become effective and induce silencing, the siRNA must reach the cytoplasm of the target cell. Currently, several delivery systems have been used for the direct application of siRNAs, which induce RNAi in vitro and in vivo [2-4]. In vitro, silencing has been accomplished in most cells by coupling the siRNAs to cationic complexes and using standard transfection techniques [5,6]. For example, in the eye, the effective silencing of transforming growth factor-β (TGF-β) type II receptor (TβRII) inhibited fibronectin production and cell migration in human cultured corneal fibroblasts [5]. Nucleofector transfection of the siRNA that silences the inhibitor of calcification gene, MGP, into nontransformed HTM cells inhibited 96% of its mRNA and reduced the activity of the calcification marker, alkaline phosphatase [7]. In vivo, examples of siRNA therapy for the management of a number of diseases are beginning to be reported (recently review in [2,3]). Development of sepsis in mice following a lethal dose of lipopolysaccharides was significantly inhibited by pretreatment of the animals with an intraperitoneal injection of anti-TNFα siRNA formulated in liposomes [8]. Using the so-called hydrodynamic method (targeting the liver by injection of a large volume via the tail vein) [9], successful silencing of marker genes in hepatocytes was achieved after injection of unmodified, naked siRNA [10].

The unique situation and conditions of the eye appear ideal for the use of therapeutic siRNA in vivo. Delivery of naked siRNA to ocular tissues by local injection has already proven successful in different mouse models. In two studies, intraocular delivery by subretinal injection of siRNA designed for vascular endothelial growth factor (VEGF) as well as for intravitreal or periocular injection of its receptor (VEGFR1) reduced ocular neovascularization [11,12]. In a another example, subconjunctival delivery of specific TβRII siRNA suppressed the TGF-β action on collagen deposition and ocular inflammation [5]. Our goal in the present manuscript was to investigate whether unmodified, naked siRNA could be locally delivered to the human trabecular meshwork tissue in perfused organ cultures for the potential treatment of glaucoma.

Glaucoma is the leading cause of irreversible blindness and it is characterized by the degeneration of retinal ganglion cells, progressive optic nerve head degeneration, and visual field loss [13,14]. The major risk factor for the most common type of glaucoma (primary open angle glaucoma, POAG) is elevated intraocular pressure (IOP) [15,16]. Elevated IOP results from increased resistance to the aqueous humor flow offered by the outflow pathway [17]. In humans, the main route of aqueous humor outflow is through the trabecular meshwork and Schlemm's canal tissue [18-20], referred to here also as the trabecular meshwork. Disruptions to the trabecular meshwork function lead to an increase in resistance to the flow and to the generation of the elevated IOP commonly associated with POAG [17,21,22]. The trabecular meshwork tissue is easily accessed by injection into the anterior chamber of the eye and it is located within the natural pathway of the flow of aqueous humor. These characteristics make the outflow pathway tissue an easy target for gene transfer.

Another clinical important risk factor associated with elevated IOP and glaucoma is treatment with steroids. Corticosteroids such as dexamethasone (DEX) are known to cause an increased resistance to aqueous humor outflow and elevated IOP in 30%-40% of normal patients and 90% of POAG patients (steroid responders) [23-26]. The molecular mechanisms as to why DEX induces high IOP are not fully understood. In their quest, a number of studies have addressed the search for trabecular meshwork genes which are altered by prolonged treatment with DEX [27-30]. Numerous genetic studies have corroborated that one of their identified genes, trabecular meshwork-inducible glucocorticoid response gene (TIGR, later renamed myocilin) [31,32], is genetically link to 4.6% of POAG and 36% juvenile OAG (JOAG) patients [33-35].

Because of the easy accessibility of the IOP/DEX response tissue and with the intent of addressing a new therapeutic approach for the management of glaucoma, this present study investigated the direct delivery of unmodified, naked siRNA into the intact human trabecular meshwork. For an initial proof of concept, we first perfused post-mortem anterior segments with fluorescently labeled RISC-free siRNA. To demonstrate functional gene silencing in the trabecular meshwork, we then tested the silencing of the highly abundant trabecular meshwork gene, Matrix Gla protein (MGP), by perfusing its commercially available siRNA. Finally, we determined the possibility of regulating the response of the outflow pathway to DEX by perfusing the siRNA and silencing the glucocorticoid receptor (GR) followed by measuring its downstream effects on two well characterized trabecular meshwork-relevant genes, myocilin (MYOC) and cornea-derived transcript factor 6 (CDT6). Our results showed for the first time that under organ culture's perfused conditions, genes of the human trabecular meshwork tissue can be regulated by the direct treatment of commercial, unmodified siRNA.

Methods

Materials

Fluorescently labeled siGLO^TM RISC-free siRNA with the fluorophore Cy3^TM was obtained from Dharmacon (Lafayette, CO). The Silencer® predesigned human MGP siRNA (ID number 122163), the Silencer® predesigned human glucocorticoid receptor NR3C1 (Nuclear Receptor, subfamily 3, group C) GR siRNA (ID number 3909), and non targeted scramble Silencer® Negative Control number 1 siRNA (catalog number 4611) were synthesized by Ambion (Austin, TX). The siRNA sequences targeting MGP (GenBank NM_000900) and NR3C1 (GenBank U01351) were 5'-GUG AGG GUC AAA GGA GAG Utt-3' and 5'-GGA GCU ACU GUG AAG GUU Utt-3', respectively. MGP and NR3C1 siRNAs were reconstituted to a final concentration of 100 μM in 200 μl of Rnase-free water and stored in aliquots at -20 °C. The negative control siRNA was comprised of a 19 bp (base pair) scrambled sequence and was reconstituted in the same manner but to a final concentration of 50 μM. Dexamethasone (DEX; Sigma, St Louis, MO) stock solution was prepared by adding absolute ethanol to the commercial vial obtaining a final concentration of 0.1 mM and kept at 4 °C.

Anterior segment perfused organ cultures

Seven pairs of eyes from nonglaucomatous human donors ranging from 61 to 79 years of age were obtained within 39-45 h from National Disease Research Interchange (NDRI, Philadelphia, PA) following signed consent of the patients' families (one pair, individual 2, 70 years old was obtained at 55 h). All procedures were in accordance with the Tenets of the Declaration of Helsinki. Whole eye globes were dissected at the equator, cleaned away from the vitreous, iris, and lens, then mounted on the perfusion chambers as described previously [36-38]. Perfusion was conducted at constant flow at an average of 3.4±0.3 μl/min using serum-free high glucose Dulbecco's modified Eagle's medium (DMEM; Invitrogen-Gibco, Carlsbad, CA). One pair of eyes was perfused at 15 mmHg constant pressure. The average outflow facility value at baseline was 0.22±0.19 μl/min/mmHg (n=14). All eyes attained a stable baseline for over 20 h before being used for these experiments.

Primary culture of outflow facility cells

Cell line HTM-55 was generated from a pair of a 25-year-old nonglaucomatous human donor obtained 36 h post-mortem from NDRI following same procurement procedures as for the organ culture eye. The trabecular meshwork (TM) was isolated from surrounding tissue by making incisions both anterior and posterior to the meshwork then removing it using forceps. The tissue was subsequently cut into small pieces, treated with 1 mg/ml collagenase type IV (Worthington, Lakewood, NJ) in phosphate buffered saline (PBS), and incubated at 37 °C in a shaker water bath for one h. Incubation was followed by low speed centrifugation for five min and pellets resuspended in 5 ml of Modified Improved Minimal Essential Medium (IMEM; Biofluids, Rockville, MD) supplemented with 20% fetal bovine serum (FBS; Invitrogen-Gibco, Carlsbad, CA) and 50 mg/ml gentamicin (IMEM high serum, Invitrogen-Gibco). Resuspended tissue was plated on a single, 2% gelatin coated 35 mm well and was maintained in a 37 °C, 7% CO₂ incubator. The medium was changed every other day and once confluent (2-3 weeks), cells were passed to a T-25 flask and labeled as passage 1. Subsequently, cells were passed 1:4 at confluency and maintained in the same medium with 10% serum.

Cell line HTM-69 was generated from the trabecular meshwork dissected from the residual cornea rim of a 19-year-old donor after a surgical corneal transplant at the University of North Carolina Eye Clinic. Specimens were cut in small pieces, carefully attached to the bottom of a 2% gelatin-coated 35 mm dish, and covered with a drop of medium (IMEM high serum) and a cover slip. Dishes were then incubated and treated as described above.

All cells were used at passages 4-6. These outflow pathway cultures comprise all cell types involved in maintaining resistance to flow. These include cells from the three distinct regions of the trabecular meshwork and cells lining Schlemm's canal. Because most of the cells in these cultures come from the trabecular meshwork, they are commonly referred to as "trabecular meshwork cells".

Short-interfering RNA and drug treatments of perfused organ cultures

For the organ cultures, continuous treatment was carried out by perfusion of the siRNAs and/or combined DEX-siRNA diluted in a DMEM perfusion medium. Final concentrations of the siRNAs and DEX were 100 nM each. Fresh siRNA media were changed from the syringes every 24 h. For the single treatments, eyes were injected by the use of a MX7900-000 HPLC pump (Rheodyne, Rohnert Park, CA) with 20 μl of siRNAs diluted in perfusion medium to a final concentration of 200 nM. A DEX reagent was prepared in absolute ethanol as a stock solution of 0.1 mM and diluted 1000x into the media before use. Media syringes were kept at room temperature while about half of perfusing tubing as well as perfusion chambers were maintained in a 37 °C, 7% CO₂ atmosphere. Perfusions were stopped 48 h after siRNA treatments and processed for histological or gene expression analyses.

Short-interfering RNA and drug treatments of nontransformed human trabecular meshwork cells

siRNAs were transfected to HTM cells using Nucleofector (Amaxa Biosystems, Gaithersburg, MD) technology and their basic kit from Primary Mammalian Endothelial Cells according to manufacturer's directions. Pilot experiments with 4x10⁵ nontransformed HTM cells and 2 μg of Amaxa's pMax. Green fluorescent protein (GFP) plasmid was used to determine conditions for maximum transfection efficiency. Of all their built-in programs, Program T-23 was found to result in an approximate 80% efficiency/80% viability of HTM cells thus it was selected for consecutive studies. For this transfection procedure, cells in 10 cm dishes were split 24 h before transfection to obtain a 70-80% confluency. Cells were then trypsinized, counted by hemocytometer (Hausser Scientific, Horsham, PA), and centrifuged at 100 xg for 10 min. Cell pellets were resuspended in full strength proprietary mammalian endothelial solution provided in the kit (Amaxa) at a concentration of 4x10⁵ cells/100 μl. Upon the addition of 1.2 μl or 2.4 μl from 100 μM GR siRNA or 50 μM scramble siRNAs, respectively, the cell-siRNA mix was electroporated on the nucleofector apparatus using program T-23 and following Amaxa's recommendations. Immediately after electroporation, cell-siRNA solution was allowed to recover for 15 min in 0.5 ml of pre-warmed IMEM, which was supplemented with 10% FBS and 50 μg/ml gentamicin kept in the CO₂ incubator in an eppendorf tube. Following the recovery period, cells were transferred to warm medium-containing 3 cm dishes where the final siRNA concentration was 27 nM. Twenty four h after transfection, media was changed and subconfluent HTM cells were treated with DEX for 48 h in the presence of serum followed by an additional 24 h serum-free medium plus DEX. The final DEX concentration was 100 nM. Parallel dishes received IMEM medium containing a drug vehicle under the same conditions. At the end of each treatment, cells were washed two times with PBS and lysed using 350 μl guanidine thiocyanate buffer (RLT, QIAGEN, Valencia, CA). HTM cells were then scraped from the plate and processed for RNA extraction as indicated below.

Histology

Meridional wedge shape specimens from opposite quadrants containing the angle region with the trabecular meshwork were fixed for one h in fresh 4% paraformaldehyde (Fisher Scientific, Fair Lawn, NJ) in PBS and immersed consecutively in 10% sucrose for approximately six h and 30% sucrose overnight (Sigma). Tissues were then immersed in OCT compound (Sakura Finetek USA, Torrance, CA), frozen in liquid nitrogen, and sectioned at 10 μm with a 2800 FrigoCut Reichert-Jung cryostat (Leica Microsystems Inc., Bannockburn, IL) at -20 °C. Sections were mounted with Fluoromount G (Southern Biotechnology Associates, Birmingham, AL). High-resolution fluorescence imaging was performed with a Zeiss 510 Meta confocal laser scanning microscope (Carl Zeiss, Inc., Thornwood, NY), was operated with a HeNe 533 nm laser and LP585 emission filter for detection of Cy3 siRNA at the Michael Hooker Microscope Facility University of North Carolina. Confocal images were collected by using 10x0.30 numerical aperture (NA) and 20x0.45 NA objectives and the pinhole size was set to 1 Airy disc unit, which yielded optical slices of approximately 15 μm and 7 μm thickness, respectively. Digital images were arranged with image-analysis software (Photoshop CS; Adobe, Mountain View, CA).

RNA extraction and reverse transcription

Anterior segments were either frozen at -80 °C or placed in RNAlater (Ambion, Austin, TX) immediately after perfusion. RNA extraction was performed by homogenizing dissected trabecular meshwork tissues or cellular pellets in 300 μl of guanidine thiocyanate buffer and loading the solution onto a QIAshredder^TM column (QIAGEN, Valencia, CA). The extraction continued by the use of the RNeasy Mini kit with RNase-free DNase digestion on the column provided according to manufacturer's recommendations (QIAGEN). Total RNA recovery averaged approximately 0.75 μg per individual TM and 2.6 μg per 3 cm dish.

Reverse transcription (RT) reactions were conducted with 1 μg (HTM cells) of spectrophotometrically measured total RNA in a total volume of 25 μl of proprietary RT buffer containing random primers, dNTPs, and 62.5 U of Multiscribe MuLV RT enzyme with RNAse inhibitor (High Capacity cDNA kit, ABI) following manufacturer's recommendations (25 °C for 10 min, 37 °C for two h). For the TM tissue, starting RNA was 0.5 μg and the volume of reagents was divided by two.

Differential expression by TaqMan real-time polymerase chain reaction

Fluorescently labeled TaqMan probe/primer sets for selected genes were purchased from the Applied Biosystems TaqMan Gene Expression collection. The GC receptor probe (NR3C1) corresponded to sequences from exons 2 and 3 (Hs00230813_m1), the MYOC probe corresponded to sequences from exons 2 and 3 (Hs00165345_m1), the CDT6 probe corresponded to sequences from exons 1 and 2 (Hs00221727_m1), the MGP probe corresponded to sequences from exons 1 and 2 (Hs00179899_m1), and the 18S RNA probe corresponded to sequences surrounding position nucleotide 609 (Hs99999901_s1). Reactions were performed in 20 μl aliquots using TaqMan Universal PCR Master mix No AmpErase UNG (Applied Biosystems), run on an Applied Biosystems 7500 Real-Time PCR System, and analyzed by 7500 System SDS software (Applied Biosystems). Relative Quantification (RQ) values between treated and untreated samples were calculated by the formula 2^-ΔΔC_T where C_T is the cycle at threshold (automatic measurement), ΔC_T is C_T of the assayed gene minus C_T of the endogenous control (18S), and ΔΔC_T is the ΔC_T of the normalized assayed gene in the treated sample minus the ΔC_T of the same gene in the untreated one (calibrator). Because of the high abundance of the 18S rRNA used as the endogenous control and in order to get a linear amplification, RT reactions from control and experimental samples were diluted 10⁴ times prior to their hybridization of the 18S TaqMan probe. Relative quantification values that are greater than or equal to 1 corresponded to increased fold changes. Relative quantification values that are less than or equal to 1 corresponded to a fraction of the gene expression and were converted to percentage of silencing. For example, an RQ of 4 corresponds to a four-fold increase in expression (+4) and an RQ of 0.25 corresponds to 0.25 of gene expression or a 75% silencing.

Western blotting

HTM media from DEX-treated cultures pre-transfected with either negative control or GR siRNAs and effluents from DEX perfused organ cultures pre-perfused with either negative control or GR siRNAs were assayed for levels of myocilin. Media and effluents were centrifuged at 16000 xg for 30 min at 4 °C to remove cellular debris. Cleared media containing secreted cell products were applied to NAP-10 columns (Amersham Biosciences, Piscataway, NJ) to exchange the buffer to 0.1 M Tris-HCl pH 7.4 (Sigma) and then concentrated three fold on a Amicon Ultra-4 Centrifugal Filter Device (Millipore, Billerica, MA). After the addition of 1X protease inhibitor cocktail (Roche, Indianapolis, IN), samples were boiled for 10 min in 1X Laemmli buffer (Bio-Rad, Hercules, CA) with 5% 2-Mercaptoethanol followed by a short centrifugation at 16000 xg. Equivalent volumes of total protein extracts were separated by SDS-PAGE in 4-15% Tris-HCl 50 μl polyacrylamide gels (Bio-Rad) and electro-transferred to a PVDF membrane (Bio-Rad). After blocking with 5% blotting grade blocker non-fat dry milk (Bio-Rad) in PBS-0.2% Tween 20 (Sigma) for two h, membranes were incubated overnight at 4 °C with MYOC goat polyclonal IgG antibodies (Santa Cruz Biotechnology, Santa Cruz, CA), diluted 1:200 to 1:50. After treatment with anti-goat IgG secondary antibodies conjugated to horseradish peroxidase (1:5000; Pierce Biotechnology, Inc, Rockford, IL) for one h, immunoreactive bands were visualized by chemiluminescence ECL Plus western blotting detection system (Amersham Biosciences, Piscataway, NJ) and membranes were exposed to light film (BioMax MR; Kodak, Rochester, NY).

MYOC bands from HTM cell media and perfused organ culture effluents were captured using a ChemiDoc System equipped with a Chemi-cooled CCD camera, PCI digitizing image acquisition board, EpiChemi II Darkroom with transilluminator and LabWorks Software (UVP, Upland, CA). Normalization for the distinct MYOC band intensities was subsequently performed using the total protein concentration of each sample, which was measured using a Bradford Protein Assay (Bio-Rad).

Results

Direct delivery of labeled short-interfering RNA to intact human trabecular meshwork tissue in perfused organ cultures

As a first approach to investigate the potential of in vivo siRNA delivery to an undisturbed human TM tissue, we used perfused human anterior segment cultures from post-mortem donors and perfused them with Cy3-labeled RISC-free synthetic siRNA. Two pairs of anterior segments (individual 1 and individual 2) were either continuously perfused with 100 nM siRNA for two days (OD) or injected twice (24 h apart) with 200 nM siRNA during the two-day perfusion regimen (OS). To distinguish between the detection of labeled siRNA just passing through the tissue and/or siRNA remaining attached/inside the trabecular meshwork cells, we "washed" the labeled siRNA treatment by post-perfusing with fresh standard medium for 24 h before the end of the experiment. Evaluation of fluorescence localization on the tissue was conducted by histochemistry.

Confocal fluorescence analysis showed extensive localization of labeled siRNA to the anterior segment tissues facing the surrogate anterior chamber of the organ culture. The CY3 red fluorescence was considerably more intense on the region of the human trabecular meshwork (Figure 1). The cornea endothelium had a significant, though lower, intensity labeling and no fluorescence was observed on the deepest regions of the cornea (Figure 1). No significant differences were found between anterior segments perfused for two days with the labeled siRNA (100 nM) and those that were injected (200 nM) during the same time period (not shown). These results strongly suggested a positive and efficient transfection of a naked siRNA molecule to the TM intact tissue.

Specific gene silencing in the intact human trabecular meshwork tissue by perfusion of a synthetic Matrix Gla siRNA

In view of the above results, experiments were performed to demonstrate the ability of naked siRNA to silence a specific trabecular meshwork gene. For these perfusions, we selected the siRNA for MGP, a highly expressed gene in the human trabecular meshwork tissue and cells [39,40]. We had previously shown that this commercially available MGP siRNA efficiently silenced MGP expression in nontransformed HTM cells [7]. Two pairs of post-mortem eyes (individual 3 and individual 4) were set in organ culture and perfused with 100 nM of MGP siRNA (OD) and with the same concentration of a scrambled negative control (OS) for three days. To measure the siRNA-mediated effect (that is, silencing) we used the reverse transcription and quantitative TaqMan PCR method in the dissected trabecular meshwork tissue of the perfused eyes. Normalizations were done with 18S cDNA and results were plotted relative to the effect of the negative control, the scrambled siRNA-perfused eye. As shown in Figure 2, the mRNA level from individual 3 treated with MGP siRNA had an RQ of 0.053±0.006 corresponding to a silencing of 94.7%±0.62 (n=6, p=0.00004) from scramble controls. In individual 4, the MGP mRNA of the treated eye showed a similar RQ of 0.064±0.001 corresponding to a silencing of 93.6%±0.13 (n=6, p=0.000002) from the scrambled control-perfused eye. These data suggest that the perfused siRNA did enter the cells of the intact trabecular meshwork tissue in organ culture and induced a marked degradation of its targeted mRNA.

Modulation of dexamethasone-induced gene expression by short-interfering RNA transfection of the glucocorticoid receptor in nontransformed human trabecular meshwork cells

To extend our findings of direct gene silencing in the trabecular meshwork and to investigate the usefulness of synthetic siRNA to inhibit the expression of pathogenic proteins, we next studied the downstream effect of silencing the receptor of glucocorticoids, which are known to induce elevated IOP in responding individuals. Because of the importance of potentially using this technology in the presence of the steroid inducer, we first tested the effects of GR siRNA in nontransformed HTM cells exposed to either DEX or vehicle.

Two nontransformed cell lines, originating from two different donors (HTM-55 and HTM-69), were nucleofector transfected in parallel dishes with same concentrations of GR siRNA and scramble control. Twenty four h later, cultures were stimulated with either DEX or vehicle for three additional days. First set of measurements indicated that statistically, the DEX-receptor GR mRNA was significantly reduced in the absence and presence of DEX as compared to scramble controls (each normalized to their internal 18S cDNAs; Figure 3A). In the absence of DEX, GR mRNA in HTM-55 had an RQ of 0.29±0.02, which corresponds to a decrease of 71.6%±2.4 (n=6, p=0.001) from scramble controls whereas in HTM-69 the RQ was 0.41±0.07 corresponding to a silencing of 59.3%±7.2 (n=6, p=0.015). In the presence of DEX, the reductions were slightly lower but still marked and statistically significant. In line HTM-55, levels of GR mRNA had an RQ of 0.40±0.03 or silencing of 60%±3.0 (n=6, p=0.003) and in HTM-69, the RQ was 0.55±0.06 corresponding to 44.7%±5.6 (n=6, p=0.015) silencing (Figure 3A). This data indicated that transfected GR siRNA can specifically inhibit the expression of the glucocorticoid receptor in HTM cells.

To then determine whether reducing GR gene expression had an effect on the expression of DEX-induced genes, we measured levels of MYOC, a gene linked to glaucoma and a marker of DEX induction in HTM cells. The DEX induction of MYOC in cells transfected with scramble control was 6.2±0.6 fold (n=6, p=0.014) and 3.9±0.1 fold (n=6, p=0.001) in HTM-55 and HTM-69, respectively, as compared to vehicle-treated cells (Figure 3B). In contrast, in the same cell lines, the three-day DEX induction of MYOC in cells transfected with GR siRNA was very much reduced. In this case, MYOC increases were 2.0±0.2 fold (n=6, p=0.02) and 2.0±0.1 fold (n=6, p=0.009) in HTM-55 and HTM-69, respectively, (Figure 3B) over vehicle-treated cells. Moreover, in cells treated with DEX, expression levels of MYOC showed a decreased expression in those transfected with GR siRNA as compared to those transfected with scrambled control with RQs of 0.51±0.02 and 0.80±0.02 corresponding to silencing values of 48.9%±2.0 (n=6, p=0.002) and 19.9%±2.1 (n=6, p=0.01) in HTM-55 and HTM-69, respectively (Figure 3C). Taken together, these results showed that inhibiting the GR receptor by siRNA reduces the expression of MYOC thus suggest that it might be feasible to use siRNA to hamper the harmful effects of DEX in HTM cells.

Effects of perfusing glucocorticoid receptor short-interfering RNA on the expression of dexamethasone-induced genes in post-mortem human organ cultures

To further examine the ability of inhibiting downstream gene expression in a closer to natural experimental system, we investigate whether administering synthetic GR siRNA in flowing conditions could silence specific induced genes in the human trabecular meshwork tissue. To optimize the effect of the siRNA, the experimental design included a perfusion of siRNA pre-treatment for two days followed by the addition of DEX for another 24 h.

Two pairs of eyes (individual 5 and individual 6) were perfused and used in this study. In individual 5, the GR mRNA level of the perfused GR siRNA eye (OD) had an RQ of 0.005±0.0005, which amounts to a GR silencing in the TM tissue of 99.5%±0.05 (n=6, p=0.0000003) compared to GR expression levels in the scrambled perfused eye (OS). Once there was confirmed reduction of the receptor expression, we examined levels of two DEX-induced genes in the same RNA samples. In addition to MYOC, we selected cornea-derived transcript 6 (CDT6, alias ANGPTL7), a recently discovered to be the most specifically DEX-induced gene in the human TM [29,30]. Expression of both genes was reduced in the GR siRNA perfused eye. The RQ of MYOC was 0.007±0.00005 or a percent MYOC reduction of 99.3%±0.005 (n=6, p=0.00000001) and that of CDT6 was 0.027±0.0025 or a percent silencing of 97.3%±0.25 (n=6, p=0.00001; Figure 4, top panel). Reduction values and trends for individual 6 were very similar. The RQ of the GR receptor gene was 0.02±0.0006 corresponding to a percent silencing of 98.2%±0.06 (p=0.000003), which in turn resulted in RQs of 0.10±0.002 of MYOC mRNA levels and 0.14±0.009 of CDT6 mRNA levels, which corresponds to decreases of 90.0%±0.22 (p=0.000006) and 85.6%±0.88 (p=0.0002; n=6 each), respectively. A third pair of eyes (individual 7), perfused and induced with DEX for three days, did not show the effect. Under these conditions, the receptor had an RQ of 0.80 (20%) which was not enough to induce the silencing of MYOC and CDT6, which were slightly increased 2.9- and 3.8 fold, respectively, relative to the scramble perfused controls (not shown). These results confirm that synthetic, naked siRNA can specifically inhibit the expression of an endogenous trabecular meshwork gene by its direct delivery to the intact human tissue. Furthermore, they show that under proper conditions, the silencing of the target gene can functionally modulate the expression of other downstream affected genes.

Changes in the myocilin protein by silencing the glucocorticoid receptor in the presence of dexamethasone

The effect of silencing the GR mRNA on the mRNA of the MYOC gene would be expected to result in a corresponding reduction of the MYOC secreted protein. Western immunoblot analysis, performed with anti-myocilin antibody on equivalent aliquots of the perfused organ culture effluents, as well as on media from HTM cells, transfected with GR siRNA and scrambled control in the presence of DEX, showed the doublet myocilin 55 kDa band. In the HTM-55 cell line, after normalization to total protein, secreted myocilin was reduced 18.8%±6.6 (n=5, p=0.02) compared to scramble controls while the reduction of myocilin in the effluent of Individual 5, perfused with GR siRNA, was 33.6%±2.0 (n=5, p=0.0000002) from the contralateral control (Figure 5). The reduction also occurred on the second cell line, HTM-69 (25.5%±8.6, n=2, p=0.1) and on Individual 6 (16.1%±6.0, n=4, p=0.04). These results indicate that silencing of the receptor was functional and reduced not only the mRNA but also the secreted protein of one of its affected genes.

Discussion

In the present study, we show that unmodified, naked siRNA delivered locally to the human anterior segment in perfused organ culture is able to silence the expression of specific genes in the trabecular meshwork tissue. The observed silencing extended not only to the targeted gene but also to the downstream transcription and translation of genes induced by a ligand after knock down of its receptor. Thus, we found that after continuous perfusion or repeated individual dose of Cy3-labeled unspecific siRNA molecules, fluorescence was detected in the TM region by histology of frozen specimens. More significant, we found that treatment with MGP siRNA at 100 nM for 72 h effectively downregulated the expression of this gene in the trabecular meshwork tissue and that pretreatment with the GR receptor siRNA was sufficient to significantly reduce the expression of two highly DEX-induced relevant genes, MYOC, and CDT6.

The perfused organ culture utilizes anterior segments from post-mortem donors under parameters close to physiological IOP conditions [36-38]. Over the years, we have shown that TM tissue originating from these cultures preserves all viable physiological and molecular characteristics. Cells in the TM of these cultures conserve their morphology, generate reproducible gene expression libraries, and support efficient viral gene transfer [37,39,41]. Although the perfusion syringes are kept at room temperature, the chambers containing the anterior segments are maintained in an incubator at 37 °C. Whether the silencing effect of the siRNAs observed in this study was affected by the temperature is not currently known. It has been reported that similar to plants, hypothermic temperatures have a diminishing effect on the efficacy of siRNA in cultured mammalian cells [42,43]. A mathematical model based on the bioheat transfer equation has determined that the temperature of the human aqueous humor is about 34 °C [44], which would imply that the temperature of the trabecular meshwork might be lower than 37 °C [42]. The influence of lower temperatures on myocilin silencing in cell culture is dependent on the siRNA sequence and the gene region targeted [42]. Some siRNAs are not affected at all. The commercial siRNAs used here are trademark designed for optimal specificity and potency. Thus, due to their potential use in vivo, it would be interesting to investigate their efficacy in organ cultures at slightly lower temperatures.

Our first experiments perfusing a labeled, generic siRNA showed intense fluorescence in the trabecular meshwork-angle region. Although, to a considerably lower extent, the fluorescence was also observed in the cornea endothelium and it persisted after perfusion with a fresh, unlabeled medium. Besides confirming our ongoing observations that molecules delivered to the anterior chamber are preferentially targeted to the trabecular meshwork tissue, this result also suggested that trabecular meshwork cells may have intrinsic properties that make them more receptive to small siRNAs. The fact that no labeling was observed in the deepest region of the cornea was an indication that the molecule was not just diffusing freely but rather retained by the cells.

In order to discern whether the localization of the siRNA fluorescent label in the trabecular meshwork was due to internalization of the molecule into the cells, to extracellular bindings, or simply to trappings in between the extracellular matrix (ECM), we tested the function of a siRNA for a specific trabecular meshwork expressed gene. We perfused the synthetic siRNA for the MGP gene and measured its expression in the dissected trabecular meshwork upon normalization by cell number (represented by expression of 18S RNA). The knock down of MGP expression was very efficient, indicating that siRNA molecules entered the trabecular meshwork cell and performed their silencing function.

At the present time, we do not know the mechanism as to how the unmodified siRNA entered the trabecular meshwork cells while in their original architectural conformation. Most in vivo targeting reports involve the complex of siRNA to different types of reagents like polyethylenimine. These polymers are believed to facilitate cell entry by creating high cationic charges and favoring endocytosis [3]. However, a few studies report relatively successful systemic targeting of naked siRNA to the liver or lung by injection of high volumes at high pressures and over a short period of time (hydrodynamic method) [2,45]. In such cases, the authors hypothesize that the mechanism of uptake might involve formation of transient pores in the targeted cells and that injection at high pressure might play a role in cell entry. One could speculate that the intact organ culture conditions used here, which implied delivery of large volumes at pressure, could also influence entry of siRNA into the trabecular meshwork cell.

The eye presents unique advantages for the delivery of naked siRNAs. To reach the eye, siRNA does not need to enter in contact with serum, which is highly rich in RNases. However, stability of siRNA in the eye tissues is not known. Other short nucleic acids such oligonucleotides or ribozymes have been shown to be degraded fairly rapidly in the presence of eye tissues especially those of the posterior segment [4]. However, in one report, direct application of naked siRNA was able to correct physiological effects in a living animal model [3,4,46]. In such a study, repetitive subconjunctival injections of VEGF siRNAs at concentration of 10 μg per injection inhibited corneal angiogenesis in mouse models of neovascularization [11]. This example demonstrates that even if degradation had occurred, the remaining levels of unmodified siRNA were able to directly enter the cell and exert their function.

In our experiments, the anterior chamber surrogate does not contain aqueous humor and is instead filled with DMEM medium. We do not know whether RNases are present in the human aqueous humor or whether its protein content would interfere with siRNA uptake. But even if attacked by degradation enzymes, a remaining small amount of siRNA could be enough to silence the gene. It is worth noting that the calculated total siRNA amount perfused through the TM in these experiments was 13 ng, which is about 1,000 times less than when siRNA was used on the subconjunctival injection in vivo [11]. Experiments in Wistar rats are in progress to test whether the siRNA functional delivery is maintained in a living animal.

Another potential factor favoring the entry of the siRNA into the trabecular meshwork tissue is the high phagocytic activity of the trabecular meshwork cells. HTM cells are able to uptake foreign material such as nonmetabolizable microspheres and pigment granules [47] and ingest erythrocytes and pseudoexfoliation material [48-51]. HTM cells also phagocytize digested ECM material produced by the action of secreted enzymes [52]. It is then possible that foreign material like a siRNA molecule wandering in between the spongiform tissue (and thus increasing contact time) would be amenable to phagocytosis by the cells of the TM. Interestingly, it is known that DEX diminishes the phagocytic activity of the trabecular meshwork [53,54], which would mean that the siRNA delivery seen here could be further enhanced by the use of agents that do not affect phagocytosis.

To move a step toward the potential application of siRNA for elevated IOP treatment, we tested whether the downstream effects of glucocorticoids could be regulated by silencing its receptor (GR siRNA). It is well established that the trabecular meshwork is a highly glucocorticoid (GC)-responsive tissue and that GC administration leads to steroid-induced ocular hypertension [23-26]. Of the two GR receptors, α and β, it has been shown that the ligand-binding isoform GRα mediates most of the GC physiological effects [55] while the alternatively spliced GRβ acts as a dominant negative regulator of GRα [54,56,57]. The commercial GR receptor siRNA used in this study, NR3C1, targets a sequence which silences both receptors. This indicates that the transcriptional response of the downstream genes observed might have been further enhanced if the GRβ receptor mRNA had not also been silenced.

Experiments to evaluate the effect of GR silencing on GC-induced gene transcription were first conducted in two cell lines of HTM cells generated from two different donors. We have previously shown that gene expression can be specifically silenced in cultured HTM cells by nucleofector electroporation [7]. Studies in other systems have also shown that downregulation of GR by specific siRNA restores the cell proliferation suppressed by DEX in murine macrophages [58]. To measure the transcription effect of GR siRNA treatment, we chose myocilin, a known HTM gene specifically induced by DEX in the HTM [27,29,30,59]. We found that pretransfection at 24 h with siRNA followed by three days of DEX treatment did significantly reduce the induced transcription of this gene. In addition, not only mRNA but also induction of the secreted protein was also reduced. In comparison, the reduction of the protein, albeit significant in three out of the four cases, was lower than that of the MYOC RNA, a result that could be in part due to the large sensitivity differences between the technologies of TaqMan PCR and western blots. Whether the reduction observed would be sufficient to affect a physiological outcome or whether larger reductions could be detected by the use of more sensitive procedures is not yet known.

Perfusion of anterior segments from two individuals with GR siRNA did also reduce induction of myocilin plus the induction of an additional HTM-DEX-induced gene [29,30]. Interestingly, there was a difference in the extent of reduction between both individuals, which is perhaps a reflection of the GC responsiveness level of each donor. Our strategy here included a short DEX treatment of 24 h. In a third individual, where the treatment of DEX was prolonged three times longer, the silencing effect was not observed, suggesting that the concentrations of siRNA used were not high enough to silence the extra gene induction that resulted after three days of DEX treatment. Compared to other reports [11], our siRNA concentrations in these experiments are about 1,000 times lower.

In summary, our findings reveal that direct delivery of siRNA to the intact human trabecular meshwork is functional. The results open a new avenue for studying trabecular meshwork cell regulation and for using siRNA as a potential protection treatment against unwanted glaucomatous insults. Many mechanistic questions still remain. Whether the positive result is due to the pressure used in the system, the increased contact time of the molecule wandering through the intact trabecular meshwork tissue, the avid phagocytic activity of the trabecular meshwork cells, or the lack of siRNA degradation in the perfusion medium is not yet known. Future studies will continue searching for answers to these questions. However, currently, we think our findings provide evidence that the use of transient, unmodified siRNA might be used to silence trabecular meshwork genes and could potentially serve as an important therapeutic alternative in the management of glaucoma risk factors.

Acknowledgements

References

1. Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE, Mello CC. Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 1998; 391:806-11.

2. Li CX, Parker A, Menocal E, Xiang S, Borodyansky L, Fruehauf JH. Delivery of RNA interference. Cell Cycle 2006; 5:2103-09.

3. Aigner A. Delivery Systems for the Direct Application of siRNAs to Induce RNA Interference (RNAi) In Vivo. J Biomed Biotechnol 2006; 2006:71659.

4. Andrieu-Soler C, Bejjani RA, de Bizemont T, Normand N, BenEzra D, Behar-Cohen F. Ocular gene therapy: a review of nonviral strategies. Mol Vis 2006; 12:1334-47 <http://www.molvis.org/molvis/v12/a150/>.

5. Nakamura H, Siddiqui SS, Shen X, Malik AB, Pulido JS, Kumar NM, Yue BY. RNA interference targeting transforming growth factor-beta type II receptor suppresses ocular inflammation and fibrosis. Mol Vis 2004; 10:703-11 <http://www.molvis.org/molvis/v10/a85/>.

6. Kim SH, Jeong JH, Lee SH, Kim SW, Park TG. PEG conjugated VEGF siRNA for anti-angiogenic gene therapy. J Control Release 2006; 116:123-9.

7. Xue W, Comes N, Borras T. Presence of an established calcification marker in trabecular meshwork tissue of glaucoma donors. Invest Ophthalmol Vis Sci 2007; 48:3184-94.

8. Sorensen DR, Leirdal M, Sioud M. Gene silencing by systemic delivery of synthetic siRNAs in adult mice. J Mol Biol 2003; 327:761-6.

9. Liu F, Song Y, Liu D. Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther 1999; 6:1258-66.

10. McCaffrey AP, Meuse L, Pham TT, Conklin DS, Hannon GJ, Kay MA. RNA interference in adult mice. Nature 2002; 418:38-9.

11. Kim B, Tang Q, Biswas PS, Xu J, Schiffelers RM, Xie FY, Ansari AM, Scaria PV, Woodle MC, Lu P, Rouse BT. Inhibition of ocular angiogenesis by siRNA targeting vascular endothelial growth factor pathway genes: therapeutic strategy for herpetic stromal keratitis. Am J Pathol 2004; 165:2177-85.

12. Shen J, Samul R, Silva RL, Akiyama H, Liu H, Saishin Y, Hackett SF, Zinnen S, Kossen K, Fosnaugh K, Vargeese C, Gomez A, Bouhana K, Aitchison R, Pavco P, Campochiaro PA. Suppression of ocular neovascularization with siRNA targeting VEGF receptor 1. Gene Ther 2006; 13:225-34.

13. Quigley HA. Number of people with glaucoma worldwide. Br J Ophthalmol 1996; 80:389-93.

14. Quigley HA. Open-angle glaucoma. N Engl J Med 1993; 328:1097-106.

15. Sommer A, Tielsch JM, Katz J, Quigley HA, Gottsch JD, Javitt J, Singh K. Relationship between intraocular pressure and primary open angle glaucoma among white and black Americans. The Baltimore Eye Survey. Arch Ophthalmol 1991; 109:1090-5.

16. Kass MA, Heuer DK, Higginbotham EJ, Johnson CA, Keltner JL, Miller JP, Parrish RK 2nd, Wilson MR, Gordon MO. The Ocular Hypertension Treatment Study: a randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma. Arch Ophthalmol 2002; 120:701-13; discussion829-30.

17. GRANT WM. Experimental aqueous perfusion in enucleated human eyes. Arch Ophthalmol 1963; 69:783-801.

18. Bill A, Maepea O, Hamanaka T. [Aspects of aqueous humor drainage through Schlemm's canal]. Klin Monatsbl Augenheilkd 1989; 195:277-80.

19. Bill A, Phillips CI. Uveoscleral drainage of aqueous humour in human eyes. Exp Eye Res 1971; 12:275-81.

20. Bill A, Maepea O. Mechanisms and routes of aqueous humor drainage. In: Albert DM, Jacobiec FA, editors. Principles and practice of ophthalmology. Philadelphia: WB Saunders; 1995. p. 206-26.

21. Lutjen-Drecoll E. Structural factors influencing outflow facility and its changeability under drugs. A study in Macaca arctoides. Invest Ophthalmol 1973; 12:280-94.

22. Maepea O, Bill A. Pressures in the juxtacanalicular tissue and Schlemm's canal in monkeys. Exp Eye Res 1992; 54:879-83.

23. Armaly MF, Becker B. Intraocular pressure response to topical corticosteroids. Fed Proc 1965; 24:1274-8.

24. Armaly MF. Effect of corticosteroids on intraocular pressure and fluid dynamics. II. The effect of dexamethasone in the glaucomatous eye. Arch Ophthalmol 1963; 70:492-9.

25. Becker B, Hahn KA. Topical corticosteroids and heredity in primary open-angle glaucoma. Am J Ophthalmol 1964; 57:543-51.

26. Weinreb RN, Polansky JR, Kramer SG, Baxter JD. Acute effects of dexamethasone on intraocular pressure in glaucoma. Invest Ophthalmol Vis Sci 1985; 26:170-5.

27. Polansky JR. HTM cell culture model for steroid effects on IOP: overview. In: Lutjen-Drecoll E, editor. Basic aspects of glaucoma research III. Stuttgart, Germany: Schattauer; 1993. p. 307-18.

28. Ishibashi T, Takagi Y, Mori K, Naruse S, Nishino H, Yue BY, Kinoshita S. cDNA microarray analysis of gene expression changes induced by dexamethasone in cultured human trabecular meshwork cells. Invest Ophthalmol Vis Sci 2002; 43:3691-7.

29. Lo WR, Rowlette LL, Caballero M, Yang P, Hernandez MR, Borras T. Tissue differential microarray analysis of dexamethasone induction reveals potential mechanisms of steroid glaucoma. Invest Ophthalmol Vis Sci 2003; 44:473-85.

30. Rozsa FW, Reed DM, Scott KM, Pawar H, Moroi SE, Kijek TG, Krafchak CM, Othman MI, Vollrath D, Elner VM, Richards JE. Gene expression profile of human trabecular meshwork cells in response to long-term dexamethasone exposure. Mol Vis 2006; 12:125-41 <http://www.molvis.org/molvis/v12/a14/>.

31. Tamm ER. Myocilin and glaucoma: facts and ideas. Prog Retin Eye Res 2002; 21:395-428.

32. Rozsa FW, Shimizu S, Lichter PR, Johnson AT, Othman MI, Scott K, Downs CA, Nguyen TD, Polansky J, Richards JE. GLC1A mutations point to regions of potential functional importance on the TIGR/MYOC protein. Mol Vis 1998; 4:20 <http://www.molvis.org/molvis/v4/a20/>.

33. Stone EM, Fingert JH, Alward WL, Nguyen TD, Polansky JR, Sunden SL, Nishimura D, Clark AF, Nystuen A, Nichols BE, Mackey DA, Ritch R, Kalenak JW, Craven ER, Sheffield VC. Identification of a gene that causes primary open angle glaucoma. Science 1997; 275:668-70.

34. Alward WL, Fingert JH, Coote MA, Johnson AT, Lerner SF, Junqua D, Durcan FJ, McCartney PJ, Mackey DA, Sheffield VC, Stone EM. Clinical features associated with mutations in the chromosome 1 open-angle glaucoma gene (GLC1A). N Engl J Med 1998; 338:1022-7.

35. Wiggs JL, Allingham RR, Vollrath D, Jones KH, De La Paz M, Kern J, Patterson K, Babb VL, Del Bono EA, Broomer BW, Pericak-Vance MA, Haines JL. Prevalence of mutations in TIGR/Myocilin in patients with adult and juvenile primary open-angle glaucoma. Am J Hum Genet 1998; 63:1549-52.

36. Johnson DH, Tschumper RC. Human trabecular meshwork organ culture. A new method. Invest Ophthalmol Vis Sci 1987; 28:945-53.

37. Borras T, Rowlette LL, Erzurum SC, Epstein DL. Adenoviral reporter gene transfer to the human trabecular meshwork does not alter aqueous humor outflow. Relevance for potential gene therapy of glaucoma. Gene Ther 1999; 6:515-24.

38. Borras T, Rowlette LL, Tamm ER, Gottanka J, Epstein DL. Effects of elevated intraocular pressure on outflow facility and TIGR/MYOC expression in perfused human anterior segments. Invest Ophthalmol Vis Sci 2002; 43:33-40.

39. Gonzalez P, Epstein DL, Borras T. Characterization of gene expression in human trabecular meshwork using single-pass sequencing of 1060 clones. Invest Ophthalmol Vis Sci 2000; 41:3678-93.

40. Xue W, Wallin R, Olmsted-Davis EA, Borras T. Matrix GLA protein function in human trabecular meshwork cells: inhibition of BMP2-induced calcification process. Invest Ophthalmol Vis Sci 2006; 47:997-1007.

41. Borras T, Matsumoto Y, Epstein DL, Johnson DH. Gene transfer to the human trabecular meshwork by anterior segment perfusion. Invest Ophthalmol Vis Sci 1998; 39:1503-7. Erratum in: Invest Ophthalmol Vis Sci 1998; 39:1984.

42. Russell P, Walsh E, Chen W, Goldwich A, Tamm ER. The effect of temperature on gene silencing by siRNAs: implications for silencing in the anterior chamber of the eye. Exp Eye Res 2006; 82:1011-6.

43. Kameda T, Ikegami K, Liu Y, Terada K, Sugiyama T. A hypothermic-temperature-sensitive gene silencing by the mammalian RNAi. Biochem Biophys Res Commun 2004; 315:599-602.

44. Scott JA. A finite element model of heat transport in the human eye. Phys Med Biol 1988; 33:227-41.

45. Tompkins SM, Lo CY, Tumpey TM, Epstein SL. Protection against lethal influenza virus challenge by RNA interference in vivo. Proc Natl Acad Sci U S A 2004; 101:8682-6.

46. Aigner A. Gene silencing through RNA interference (RNAi) in vivo: strategies based on the direct application of siRNAs. J Biotechnol 2006; 124:12-25.

47. Shirato S, Murphy CG, Bloom E, Franse-Carman L, Maglio MT, Polansky JR, Alvarado JA. Kinetics of phagocytosis in trabecular meshwork cells. Flow cytometry and morphometry. Invest Ophthalmol Vis Sci 1989; 30:2499-511.

48. Buller C, Johnson DH, Tschumper RC. Human trabecular meshwork phagocytosis. Observations in an organ culture system. Invest Ophthalmol Vis Sci 1990; 31:2156-63.

49. Grierson I, Lee WR. Erythrocyte phagocytosis in the human trabecular meshwork. Br J Ophthalmol 1973; 57:400-15.

50. Chisholm IA, Grierson I. Particulate phagocytosis by trabecular meshwork endothelium. Can J Ophthalmol 1977; 12:293-9.

51. Schlotzer-Schrehardt U, von der Mark K, Sakai LY, Naumann GO. Increased extracellular deposition of fibrillin-containing fibrils in pseudoexfoliation syndrome. Invest Ophthalmol Vis Sci 1997; 38:970-84.

52. Acott TS. Biochemistry of aqueous humor outflow. In: Kaufman PL, Mittag TW, editors. Textbook of Ophthalmology. Vol 7. Glaucoma. London: Mosby; 1994.p. 1.47-1.78

53. Matsumoto Y, Johnson DH. Dexamethasone decreases phagocytosis by human trabecular meshwork cells in situ. Invest Ophthalmol Vis Sci 1997; 38:1902-7.

54. Zhang X, Ognibene CM, Clark AF, Yorio T. Dexamethasone inhibition of trabecular meshwork cell phagocytosis and its modulation by glucocorticoid receptor beta. Exp Eye Res 2007; 84:275-84.

55. Evans RM. The steroid and thyroid hormone receptor superfamily. Science 1988; 240:889-95.

56. Bamberger CM, Bamberger AM, de Castro M, Chrousos GP. Glucocorticoid receptor beta, a potential endogenous inhibitor of glucocorticoid action in humans. J Clin Invest 1995; 95:2435-41.

57. Oakley RH, Sar M, Cidlowski JA. The human glucocorticoid receptor beta isoform. Expression, biochemical properties, and putative function. J Biol Chem 1996; 271:9550-9.

58. Zhu XY, Liu YJ, Lu J, Xu RB. Knockdown of glucocorticoid receptor expression by RNA interference promotes cell proliferation in murine macrophage RAW264.7 cells. J Steroid Biochem Mol Biol 2004; 92:375-82.

59. Polansky JR, Fauss DJ, Chen P, Chen H, Lutjen-Drecoll E, Johnson D, Kurtz RM, Ma ZD, Bloom E, Nguyen TD. Cellular pharmacology and molecular biology of the trabecular meshwork inducible glucocorticoid response gene product. Ophthalmologica 1997; 211:126-39.