Purpose
To evaluate whether selective laser trabeculoplasty and prostaglandin analogs regulate the permeability of cultured human Schlemm canal cells by inducing intercellular junction disassembly.
Design
Laboratory investigation.
Methods
Intercellular junctions were made visible in living cells by making them fluoresce after transfection with a plasmid expressing the zonula occludens 1 protein tagged with green fluorescent protein. Schlemm canal cells were treated by direct laser irradiation; by exposure to media conditioned by either lasered Schlemm canal cells or trabecular meshwork cells; by exposure to the prostaglandin analogs latanoprost, bimatoprost, and travoprost; or by the addition of the nonprostaglandin agents brimonidine, timolol, and dorzolamide. Junction disassembly was monitored using fluorescence microscopy, and permeability alterations were measured as changes in conductivity using flow meters.
Results
The direct laser irradiation of Schlemm canal cells caused a 3-fold increase in conductivity. Exposure of the cells to media conditioned by lasered Schlemm canal cells or trabecular meshwork cells induced junction disassembly and a 2- to 4-fold increase in conductivity. Exposure to prostaglandin analogs also induced junction disassembly and a 4- to 16-fold increase in conductivity, whereas the 3 nonprostaglandin agents tested were ineffective in both regards.
Conclusions
Exposure to factors secreted by lasered Schlemm canal cells and lasered trabecular meshwork cells and the application of prostaglandin analogs induced junction disassembly while increasing the permeability of Schlemm canal cells. These findings support our hypothesis that selective laser trabeculoplasty and prostaglandin analogs share a common mechanism that likely mediates their pressure-lowering effects.
In our initial experience with selective laser trabeculoplasty (SLT), we were perplexed by the highly variable success rates and modest intraocular pressure (IOP) lowering observed in some patients. Most of our patients were using multiple topical glaucoma medications at the time of SLT, which made us wonder whether the laser irradiation procedure and certain of these medications competed with each other because of the existence of common receptors or shared mechanisms of action. Among the topical medications in common use were some, such as timolol or dorzolamide, that work by decreasing the secretion of aqueous humor. These medications were considered unlikely to compete with SLT, because this procedure does not seem to affect aqueous production. In contrast, and as has been suggested already, the prostaglandin analogs latanoprost, travoprost, and bimatoprost were particularly deserving of investigation, since prostaglandin analogs and SLT are both outflow agents that may share a common mechanism of action.
Soon after the introduction of prostaglandin analogs for the treatment of glaucoma, they were proposed to lower the IOP by inducing the expression of matrix metalloproteinases and by increasing aqueous outflow through the extracellular matrix of the uveoscleral pathway. However, the mechanisms of action of prostaglandin analogs and SLT are the subject of current research, and some investigators have proposed alternative pathways. For instance, the study by Lim and associates reported tonography measurements in humans demonstrating that prostaglandin analogs increase the facility of aqueous outflow by acting both on the conventional aqueous outflow pathway as well as the previously identified uveoscleral pathway. Furthermore, in another study using cultured human anterior segments in which the uveoscleral pathway had been surgically removed, Bahler and associates demonstrated that the significant majority of the increase in aqueous outflow induced by prostaglandin analogs occurs across the conventional aqueous outflow pathway. In addition, Goyal and associates also reported that SLT-treated patients exhibit an increase in the facility of aqueous outflow. Thus, data from these 3 studies provide support for our hypothesis that both prostaglandin analogs and SLT share a common mechanism of action, with the involved cellular or molecular events yet to be elucidated.
Herein we report the findings of in vitro assays designed to determine whether SLT and prostaglandin analogs share a common mechanism of action whereby the permeability of Schlemm canal cells is regulated by controlling the intercellular junction disassembly process. We implemented a novel technique described herein to visualize the intercellular junctions forming the Schlemm canal cellular barrier under living conditions. Using this technique, we could observe the dynamics of the intercellular junction assembly and disassembly process in living Schlemm canal cells while monitoring permeability changes occurring in response to various laser and pharmacologic treatment protocols. Our findings support the hypothesis that at least part of the mechanism of lowering of the IOP by SLT and prostaglandin analog therapies involves distinct cellular and molecular events that we have detected in vitro and that could mediate an increase in the egress of fluid across the conventional aqueous outflow pathway in vivo. These in vitro findings support future studies using in situ methods and human ocular tissues seeking support for the notion that SLT and prostaglandin analogs have similar effects on the barrier function of Schlemm canal cells in vitro and in vivo.
Methods
Cell Culture and Transfection
Primary human Schlemm canal and trabecular meshwork cell cultures were established separately using our previously described methods and preserved for future use as frozen stocks. To begin our experiments, vials containing the desired cell types were defrosted and the cells were plated on collagen type IV-coated 10-cm culture dishes (BD, Franklin Lakes, New Jersey, USA) and maintained in Dulbecco’s Modified Eagle’s Medium, which was supplemented with 15% fetal bovine serum , 2 mM L-glutamine, 50 μg/mL gentamicin, and 2.5 μg/mL amphotericin B (all from Mediatech, Manassas, Virginia, USA) and fed to the cultured cells every 48 hours. Madin-Darby canine kidney (MDCK) epithelial cells were used only to compare the junction morphologic features with those in Schlemm canal cells. For transfection of the plasmid construct containing the gene for the junction protein zonula occludens-1 (ZO-1) tagged with green fluorescent protein (GFP), (GFP–ZO-1), Schlemm canal cells and MDCK cells underwent nucleofection using methods described previoulsy. After transfection, the cells were seeded on Lab-Tek chamber slides (Thermo Fisher Scientific, Rochester, New York, USA) and were cultured for an additional 48 hours before being exposed to treatment conditions.
Application of Laser Irradiation and Treatment with Prostaglandin Analogs and Nonprostaglandin Agents
The frequency-doubled Q-switched nedymium:yttrium–aluminum–garnet laser was used to treat dishes measuring 10 cm in diameter containing nontransfected Schlemm canal cells or trabecular meshwork cells. Twenty evenly distributed laser shots per square centimeter were delivered onto the monolayer of cells. The laser shots consisted of a burst of light with a wavelength of 532 nm, a duration of approximately 3 nanoseconds, a power of approximately 0.66 mJ, and a diameter measuring 400 μm. The irradiated Schlemm canal cells and trabecular meshwork cells were allowed to condition the media for 12 hours before collecting the media and adding these conditioned media onto cultures of Schlemm canal cells to be tested. In addition, nontransfected Schlemm canal cells also were irradiated directly to measure any direct effects on the conductivity of these cells. Controls consisted of cells exposed to regular medium.
For the prostaglandin analog treatment experiments, we used the free acid forms of bimatoprost, latanoprost, and travoprost, which were obtained from Cayman Chemical (Ann Arbor, Michigan, USA). Schlemm canal cells were exposed to a standard 5-μM concentration of latanoprost (dissolved in methyl acetate), travoprost (dissolved in ethanol), and bimatoprost (dissolved in ethanol). This concentration is well above the peak for these drugs measured in the aqueous humor of humans or rabbits after the topical administration of a single drop. Negative controls consisted of Schlemm canal cells exposed to either media containing the corresponding solvent diluted in regular medium to a concentration of 0.1% volume/volume or less or to a standard culture medium without additives. The duration of treatment used was chosen based on in vitro optimization experiments for the particular assay.
For the nonprostaglandin agent treatment experiments, we used the pure chemical forms of brimonidine, timolol, and dorzolamide. Brimonidine (Sigma-Aldrich, St. Louis, Missouri, USA) dissolved in dimethyl sulfoxide was used at a 20-μM concentration. Timolol (Sigma-Aldrich) was dissolved in cell culture growth medium and was used at a concentration of 10 μM. Dorzolamide (USP, Rockville, Maryland, USA) was dissolved in cell culture growth medium and was used at a concentration of 5 μM. These agents were applied at above peak concentrations of these drugs measured in the aqueous humor of humans after the topical administration of a single drop of each agent. Controls consisted of either regular medium or the appropriate volume of solvent diluted in regular medium to a concentration of 0.1% volume/volume or less. The duration of treatment used was selected based on in vitro optimization experiments for the particular assay.
Assessment of the Junction Disassembly Process and Associated Conductivity Changes
To examine the junction disassembly process, transfected cells were treated by exposure to media conditioned by laser-irradiated trabecular meshwork cells and Schlemm canal cells, prostaglandin analogs, and nonprostaglandin agents. In additional studies, transfected Schlemm canal cells were also exposed to 0.8 mM ethacrynic acid (Sigma-Aldrich). The treatment responses were photographed using the Zeiss LSM 5 Pascal confocal microscope (Carl Zeiss, Thornwood, New York, USA) to acquire real-time video photography. The confocal microscope experiments were repeated twice, and the data presented herein is representative of the results of the replicate experiments.
For conductivity studies, nontransfected Schlemm canal cells were grown as monolayers on Millipore filter supports (Millipore, Billerica, Massachusetts, USA), and the conductivity was assessed using a computer-driven hydraulic conductivity-measuring apparatus equipped with sensitive flow meters described previously. Briefly, cultured cells grown over porous filter supports were kept in a humidified 8% CO 2 incubator and fed every 48 hours with media containing 15% fetal bovine serum until they reached confluence (approximately 10 to 14 days), and the fetal bovine serum concentration subsequently was reduced from 15% to 10%. The monolayers then were exposed to the experimental and control conditions described above. The conductivity was measured in microliters per minute per millimeter of mercury per square centimeter at a constant perfusion pressure of 4.5 mm Hg at baseline and at the end of the treatment period. Conductivity experiments were repeated 3 times with multiple replicates in each sample.
Statistical Analyses
The statistical significance of the difference between 2 means for the conductivity experiments was determined using a Student 2-tailed t test for samples with unequal variance.
Results
Visualization of Intercellular Junctions in Schlemm Canal Cells
Schlemm canal cells and MDCK cells transfected with the GFP–ZO-1 plasmid construct displayed brightly fluorescent intercellular junctions for approximately 5 days. In Figure 1 , we contrast the intercellular junction protein localization most commonly encountered in epithelial and endothelial cells, as demonstrated in our experiments by the MDCK cell type ( Figure 1 , Left), with the unique distribution present in Schlemm canal cells ( Figure 1 , Right). Please note that the junction proteins in the MDCK cells are located strictly within the paracellular pathway outlining the cell perimeter. In a sharp contrast, the Schlemm canal cell junction proteins are disposed along fine finger-like structures known as filopodia. The filopodia extend atop the cell monolayer alongside the apical-most portion of the paracellular pathway. The filopodia are numerous and are closely grouped together forming a palisade arrangement, as previously demonstrated. The apposing surfaces between each filopodium and the underlying apical cell surface are densely coated with junction proteins, providing the binding material holding the filopodium and the cell onto each other.
Dynamics of the Junction Disassembly Process in Schlemm Canal Cells
We first examined the junction disassembly process in Schlemm canal cells treated with ethacrynic acid (EA). We used EA because it induces junction disassembly rapidly by interacting with actin and other components of the cytoskeletal apparatus and by inducing cellular contraction. Several snapshots in Figure 2 highlight the major steps of the junction-disassembly process when Schlemm canal cells were exposed to EA. At the outset of the EA experiment ( Figure 2 , Top left panel), palisades of fully formed filopodia bind neighboring cells together. Twenty-two minutes later ( Figure 2 , Top right), as the filopodia are stretched with contraction of the cytoskeleton under the influence of EA, the intercellular space begins to widen (asterisk). At 32 minutes ( Figure 2 , Bottom left), only some filopodia remain intact and can be seen to span across the width of the widened intercellular space, whereas most have undergone complete retraction and shortening. The disassembly process is nearly completed at 37 minutes ( Figure 2 , Bottom right), when the intercellular space has widened 5 to 10 fold (asterisk). At this stage, the basal margin of 1 cell (arrows) rests over the surface of the culture dish, and its apical-most margin is observed in the opposite location (lower arrowheads). The full extension of the paracellular pathway (i.e., height) is well appreciated in this panel, and it is pointed out between the arrows and arrowheads in the lower part of this image. This panel also shows that the ZO-1 junctional protein is localized along the apical-most margin of the paracellular pathway of the Schlemm canal cells.
In Figure 2 , Bottom right panel, the widened intercellular space is pointed out by the 4 arrowheads. Although the intercellular junctions have undergone disassembly, the ZO-1 protein is still visible as punctate structures arranged in a linear disposition. This linear disposition, or linearization, of the ZO-1 protein signals the fact that the intercellular junctions have become disassembled as the ZO-1 protein is no longer distributed along filopodia. However, the ZO-1 protein, which is actually located internally along the cytoplasmic aspect of the cell membrane, remains visible. The space extending between the 4 arrowheads in Figure 2 , Bottom right, represents the width of the markedly widened paracellular pathway, providing an illustration of the potential capacity of the cellular barrier in Schlemm canal cells to accommodate a substantial volume of transendothelail fluid flow. The junction disassembly process is reversible, and on removal of the EA, the intercellular space narrows down, and concurrently the filopopia reform and span across the intercellular space to reach and bind neighboring cells together once more (data not shown).
Laser Irradiation Induces the Disassembly of Junctions and an Increase in Conductivity
Figure 3 , Top left, shows the typical appearance of control Schlemm canal cells examined after 48 hours of exposure to standard nonconditioned media in a control preparation. The filopodia are fully formed and extend across the intercellular space of adjacent cells. In contrast, Schlemm canal cells exposed for 24 hours to media conditioned by lasered trabecular meshwork cells ( Figure 3 , Top right) show less prominent filopodia, which at 48 hrs largely have disappeared and have assumed a linear disposition ( Figure 3 , Middle left). Figure 3 , Middle right, shows at a higher magnification (×63) the portion of the intercellular junction depicted by a rectangle in Figure 3 , Middle left. Note that the junctions have assumed a linear appearance, which we propose signals the termination of the junction disassembly process. These changes were seen in most of the cells examined during replicate experiments. Similar findings were obtained when Schlemm canal cells were exposed to media conditioned by lasered Schlemm canal cells (data not shown).
The mean baseline conductivity of Schlemm canal cell monolayers exposed to standard medium was 1.42 μL/minute/mm Hg/cm 2 ( Figure 3 , Bottom). After 36 hours of incubation, this conductivity was increased to a mean of 4.68 μL/minute/mm Hg/cm 2 in lasered Schlemm canal cells, to a mean of 3.17 μL/minute/mm Hg/cm 2 in Schlemm canal cell monolayers exposed to medium conditioned by lasered Schlemm canal cells and to a mean of 5.54 μL/minute/mm Hg/cm 2 in Schlemm canal cell monolayers exposed to medium conditioned by lasered trabecular meshwork cells ( P < .001, compared with the control samples for all 3 conditions). Compared with controls, these changes correspond to a 3-fold increase in conductivity induced by lasering the Schlemm canal cells directly and to a 2-fold and a 4-fold increase in conductivity, respectively, induced by adding media onto Schlemm canal cells conditioned by lasered Schlemm canal cells and lasered trabecular meshwork cells.
Exposure to Prostaglandin Analogs Also Induces the Disassembly of Junctions and an Increase in Conductivity
In Figure 4 , we present snapshots from an experiment in which Schlemm canal cells were exposed for 30 hours to a standard 5-μM dose of the free acid forms of latanoprost, travoprost, or bimatoprost and were compared with untreated controls. Figure 4 , Top left, shows the typical appearance of Schlemm canal cells in untreated controls, where the filopodial palisades are easily discernable at the end of the incubation. In Figure 4 , Top right, a group of cells is shown after treatment with latanoprost. Note that no filopodial palisades are apparent, and instead the junctions have a blunted appearance having assumed the so-called linearized disposition similar to the appearance of the junctions shown in Figure 2 after they have undergone disassembly. This appearance is characteristic of junctions that have undergone disassembly and have become more permeable. Also note that compared with untreated cells ( Figure 4 , Top left panel), the Schlemm canal cells have become markedly elongated. The results after exposure to travoprost and bimatoprost are shown, respectively, in Figure 4 , Middle left and Middle right, which again demonstrate disappearance of filopodial palisades and the presence of cell elongation. These changes were seen in most of the cells examined during replicate experiments. Similar results were obtained when pharmacologic preparations of the commercially available products used clinically by patients (i.e., Xalatan [Pfizer/Pharmacia & Upjohn Company, Kalamazoo, Michigan, USA], Travatan [Alcon Laboratories Inc, Fort Worth, Texas, USA], and Lumigan [Allergan Inc, Irvine, California, USA]) were applied.
