The Role of the Prostaglandin EP₄ Receptor in the Regulation of Human Outflow Facility

Abstract

Purpose.: Activation of prostaglandin (PG)-EP₄ receptors by 3,7-dithiaPGE₁ robustly lowers intraocular pressure in nonhuman primate eyes, which increases outflow facility but has no effect on aqueous secretion or uveoscleral outflow. Because of differences in PG efficacy in outflow function between nonhuman primates and humans, we tested the impact of 3,7-dithiaPGE₁ on conventional outflow function in human donor eyes.

Methods.: The expression pattern of PG-EP₄ receptors was determined in corneoscleral tissues of human donor eyes and in cultures of human outflow cells by immunofluorescence microscopy and Western blot, respectively. The efficacy of 3,7-dithiaPGE₁ was determined by assaying agonist-stimulated cAMP accumulation and β-arrestin mobilization in cultured human cells. Agonist effects on outflow facility were examined in paired human donor eyes that were perfused at 8 mm Hg of constant pressure, equivalent to 15 mm Hg in vivo.

Results.: The trabecular meshwork (TM) and Schlemm's canal (SC) cells expressed PG-EP₄ receptors. Agonist-mediated effects on the PG-EP₄ receptors were detected in SC (EC₅₀ = 6.3 × 10⁻⁹ M, n = 4), but not TM (EC₅₀ = 1.7 × 10⁻⁷ M, n = 5) cells. Effects in SC cells were blocked by the PG-EP₄ receptor–selective antagonist GW627368 (EC₅₀ = 1.09 × 10⁻² M, n = 4), but not the PG-EP₂ receptor–selective antagonist AH6809 (EC₅₀ = 4.10 × 10⁻⁹ M, n = 5). Perfused into human eyes at a concentration that selectively activates PG-EP₄ receptors, 3,7-dithiaPGE₁ (10 nM) increased outflow facility by 51% ± 18% over baseline levels in individual drug-treated eyes after drug exchange (n = 6 eyes; P = 0.05) and by 69% ± 23% (P < 0.01) compared with that in contralateral eyes.

Conclusions.: Activation of PG-EP₄ receptors expressed by SC cells of the human conventional outflow pathway appears to contribute to PG regulation of outflow facility.

Blindness due to glaucoma affects almost 70 million people worldwide. ¹ It has been established that maintaining and significantly lowering intraocular pressure in individuals with primary open-angle glaucoma (POAG) prevents disease progression and vision loss. ² Currently, the most effective ocular antihypertensive agents used to medically treat POAG are the prostaglandin (PG)-F2_α analogues. ^3,4 Other classes of endogenous PGs, such as PGE₂, PGD₂, and PGI₂ also efficaciously lower IOP in animals, human anterior segments, and whole-eye organ perfusions, although side effects have prevented clinical use. ^{5 –10} Therefore, novel, more selective PG analogues are currently under development as potential therapies. For example, a novel and selective PG-EP₄ agonist lowers IOP in canines by 30% to 45%. ⁸  

PGs are locally acting, signaling molecules that are ubiquitously expressed, but have diverse, tissue-specific effects throughout the body. These differential effects are exerted when PGs bind to specific G-protein-coupled receptors (GPCRs). For example, differentially, PGE₂ activates all four PG-EP receptor subtypes (EP₁–EP₄). However, the effects of PGE₂ on a particular tissue are dependent on the expression pattern of PG-EP receptors. Thus, to achieve a desired functional response of tissue behavior with minimal side effects, it is imperative to develop compounds that can selectively activate or block an individual receptor subtype or a restricted combination of specific receptor subtypes. 

PG-FP analogues selectively bind to PG-FP receptor isoforms, lowering IOP in humans and nonhuman primates primarily by increasing uveoscleral outflow. ^3,4,11,12 Of interest, recent studies suggest that these compounds also have significant activity on the pressure-dependent pathway in humans, an action that was not initially detected in nonhuman primates. ^{13 –15} A new PG-EP₄ selective agonist, 3,7-dithiaPGE₁, effectively lowers IOP by ∼40% in nonhuman primates, increasing conventional outflow facility but not uveoscleral outflow. ⁹ In the present study, to investigate the contribution of PG-EP₄ receptors to conventional outflow regulation in human eyes, we used complementary models of the human conventional outflow pathway. We observed that human SC and TM cells express PG-EP₄ receptors and are differentially activated by 3,7-dithiaPGE₁. In addition, we found that activation of PG-EP₄ receptors in the human conventional pathway rapidly and significantly increases outflow facility in situ. 

Materials and Methods

Cell Culture

Human donor eyes were obtained from Life Legacy Foundation (Tucson, AZ), National Disease Research Interchange (Philadelphia, PA), and Sun Health Research Institute (Sun City, AZ). They were managed in compliance with the Declaration of Helsinki's guidelines for research involving human tissue. Human trabecular meshwork (TM) and Schlemm's canal (SC) cells were isolated, characterized, and cultured by methods developed in our laboratory. Briefly, TM cells were isolated from human donor eyes by using a blunt dissection procedure followed by an extracellular matrix digestion, ¹⁶ and SC cells were isolated from conventional outflow tissues with a cannulation technique. ¹⁷ The cells were maintained in Dulbecco's modified Eagle's medium (DMEM, low glucose; Invitrogen-Gibco, Carlsbad, CA), supplemented with 10% fetal bovine serum, penicillin (100 U/mL), streptomycin (0.1 mg/mL), and glutamine (0.29 mg/mL), until confluence when the TM cells were switched to DMEM containing 1% FBS. 

Human embryonic kidney (HEK-293-EBNA) cells stably expressing the PG-EP₄ and -EP₂ receptors were a gift generated by John Regan (University of Arizona, Tucson). ¹⁸ The cells were maintained in culture with DMEM containing 10% fetal bovine serum, geneticin (250 μg/mL), gentamicin (100 μg/mL), and hygromycin B (200 μg/mL). 

Cell Surface Biotinylation

Monolayers of TM and SC cells were grown in T25 flasks and allowed to mature at confluence for at least 2 weeks before experimentation. HEK 293 cells stably expressing PG-EP₂ or -EP₄ receptors were grown to confluence in T25 flasks and used immediately. Each flask was rinsed three times with reaction buffer (100 mM NaCl and 50 mM NaHCO₃ [pH 8.0]) followed by two 30-minute incubations with sulfo-NHS-LC-biotin (1 mg/mL; Pierce Rockford IL). The reactions were attenuated with Tris/glycine buffer (25 mM Tris, 192 mM glycine [pH 8.3]) and then scraped in lysis buffer (2 mM EDTA, 1% Triton X, and 1% Tween-20 in TBS) in the presence of protease inhibitor cocktail (Pierce). Next, the preparations were centrifuged at 14,000 rpm for 15 minutes, to remove any remaining intact cells, and then were incubated overnight (rotating at 4°C) with streptavidin-conjugated beads, to capture biotinylated surface proteins. After incubation, the proteins that bound to beads were separated from the unbound proteins by centrifugation at 2500 rpm and were washed with lysis buffer. Reducing sample buffer (40% glycerol, 0.5 M Tris [pH 6.8], 10% SDS, and 5% β-mercaptoethanol) was added to bound and unbound fractions, and constituent proteins were fractionated by SDS-PAGE. The proteins were transferred electrophoretically to nitrocellulose and probed for PG-EP₄ receptors or β-actin, by using affinity-purified polyclonal rabbit IgGs (Cayman Chemical, Ann Arbor, MI). The proteins of interest were visualized with x-ray film in contact with blots that had been incubated with goat anti-rabbit IgG conjugated with horseradish peroxidase followed by chemiluminescence reagent (GE Health Care Amersham, Piscataway, NJ). 

Immunohistochemistry

Human eyes received within 24 hours after death were perfusion fixed with 10% paraformaldehyde (PFA) in PBS under 8 mm Hg of pressure followed by immersion fixation at 4°C in 5% PFA (in PBS). The eyes were bisected at the equator and stored in 70% ethanol until they were embedded in paraffin. Sections (4 μm) were cut and kept overnight at 37°C to dry. They were deparaffinized with 100% xylene (three 10-minute incubations), rehydrated with 100% ethanol (three 3-minute incubations), followed by 1 minute sequential submersion in 90%, 80%, and 70% ethanol. The slides were then rinsed in tap water for 30 minutes followed by PBS. Antigens were retrieved by heating sections in citrate buffer (10 mM citric acid and 0.05% Tween 20; pH 6.0) until the solution boiled (repeated at a lower heat) and then were placed on ice for 30 minutes. Endogenous peroxidase activity in sections was blocked with 0.3% H₂O₂ in methanol for 30 minutes at room temperature (RT). After the sections were washed in PBS (three 5-minute washes), they were incubated in blocking buffer (10% goat serum in PBS) for 1 hour at RT. PG-EP₄ receptors were visualized by using polyclonal rabbit IgGs (Cayman Chemical) overnight at 4°C in a humidified chamber. Negative controls were exposed to goat serum alone. The next day, sections were given three 10-minute rinses in PBS and then exposed to HRP-anti-rabbit (EnVision+System; Dako, Carpinteria, CA) for 1 hour at RT. Staining was visualized with a DAB chromogen solution (0.2 mg/mL in 0.05 M Tris-HCl, [pH 7.6], 6 μL/mL 3% H₂O₂) by incubating tissue sections for 15 minutes at RT followed by washing in tap water for 5 minutes. The sections were rinsed three times with PBS and cleared by dehydration in 100% ethanol (three times for 3 minutes each) followed by 100% xylene (three times for 3 minutes each). The sections then were coverslipped with mounting medium (Cytoseal; Richard-Allan Scientific, Kalamazoo, MI). Hematoxylin and eosin (H&E) staining was conducted by the Histology Service Laboratory at the University of Arizona Health Sciences Center. 

β-Arrestin Mobilization

HEK cells stably expressing PG-EP₄ receptors were plated on eight-well glass slides and treated with 10 nM 3,7-dithiaPGE₁ for 0 to 210 seconds (drug provided by Allergan, Inc., Irvine, CA). The cells were fixed with ice-cold methanol three times for 10 minutes each and then incubated overnight at 4°C with polyclonal antibodies against β-arrestin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and anti-EP₄ (Cayman Chemical) receptor IgGs. The following day, the cells were rinsed three times with TBS-t (19.2 mM Tris, 0.15 M NaCl [pH up to 7.4], and 0.2% Tween-20) before incubation with rhodamine (TRITC)-conjugated donkey anti-goat and fluorescein (FITC)-conjugated donkey anti-rabbit IgG secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 hour at room temperature (RT). The cells were rinsed three times with TBSt and then coverslipped with a 50% glycerol solution. Immunostaining was analyzed with a laser scanning microscope (LSM 700; Carl Zeiss Meditec, Dublin, CA) coupled to a confocal inverted research microscope (Axio Observer.Z.1; Carl Zeiss Meditec). Three-dimensional analyses of the EP₄ receptor and β-arrestin distribution were performed with ImageJ 1.42 (developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html). 

cAMP Accumulation Assay

Intracellular cAMP was measured with a protein kinase A binding assay, as described elsewhere. ¹⁹ Briefly, HEK 293 cells stably expressing the PG-EP₄ or -EP₂ receptors, primary SC cells, along with TM cells were seeded onto 24-well plates at a density of 100,000 cells per well. EP₄- and EP₂-HEK cells were maintained until confluent in high-glucose DMEM containing 10% FBS, geneticin (250 μg/mL), gentamicin (100 μg/mL), and hygromycin B (200 μg/mL). Primary SC and TM cells were maintained as previously described. Twelve to 16 hours before treatment, the EP₄-HEK, EP₂-HEK, and TM cells were transferred to serum-free DMEM, and the SC cells were transferred to low-glucose DMEM containing 1% FBS. 

Agonist Alone.

Before treatment, cells were incubated with 0.5 mM 3-isobutyl 1-methyl-xanthine (IBMX) for 2 minutes at 37°C. The medium was aspirated, and increasing concentrations of a PG-EP₄ agonist (3,7-dithiPGE₁) plus IBMX (0.5 mM) were added for 15 minutes and incubated at 37°C. 

Agonist+Antagonist.

Before treatment, the cells were incubated with either the PG-EP₂ antagonist AH6809 (5 μM; 33458-93-4; Cayman Chemical) or the PG-EP₄ antagonist GW627368 (1 μM; 439288-66-1; Cayman Chemical) for 15 minutes at 37°C in the presence of IBMX. The medium was aspirated, and increasing concentrations of 3,7-dithiPGE₁ plus IBMX (0.5 mM) in the presence of 5 μM AH6809 or 1 μM GW627368 were added for 20 minutes and incubated at 37°C. 

After incubations in dose–response experiments with the agonist and with the agonist in the presence of the antagonist, the same method was used. The drug solutions were aspirated and the plates put on ice, and 150 μL of cold Tris/EDTA buffer (50 mM Tris/4 mM EDTA) was added to each well. The cells were scraped from culture plates, transferred into 1.5 mL tubes (Eppendorf, Fullerton, CA), boiled for 8 minutes, and centrifuged at 4°C for 15 minutes at 15,000 rpm. Supernatant (50 μL) was added to 50 μL of [³H]cAMP (0.5 μCi/mL) plus 100 μL of cold protein kinase-A solution (180 μg/mL PKA in Tris/EDTA buffer). After incubation at 4°C for 2 hours, 100 μL of activated charcoal slurry (26 mg/mL activated charcoal in Tris/EDTA containing 2% BSA) was added. The mixture was vortexed briefly and then centrifuged for 1 minute to pellet the charcoal. The supernatant (200 μL) was added to a scintillation cocktail (9.8 mL of Ecolite; Fisher Scientific, Pittsburgh, PA), and radioactivity was analyzed by an automated scintillation counter (LS 6500; Beckman Coulter, Fullerton, CA). Data were analyzed with commercial statistical software (Prism; GraphPad, San Diego, CA). 

Whole-Eye Organ Perfusion

Dulbecco's PBS containing 5.5 mM glucose (DBG) served as mock aqueous humor for the perfusion studies. DBG was perfused into the eyes at a constant pressure of 8 mm Hg via a needle inserted through the cornea into the posterior chamber, as described elsewhere. ^20,21 After a stable baseline outflow facility was reached for 30 minutes, one eye underwent an anterior chamber exchange (5 mL over 10 minutes) of 10 nM 3,7-dithiaPGE₁ in DBG for the PBS/DBG and the contralateral eye and exchange of DBG only. A two-chamber, constant-perfusion exchange through a second needle inserted into the anterior chamber ensured that an IOP of 8 mm Hg was maintained during the exchange. We used a constant pressure of 8 mm Hg equivalent to 15 mm Hg in vivo, because of the absence of episcleral venous pressure in the enucleated eyes. We calculated the percentage increase in outflow facility (C) as 100(C _final/C _baseline − 1) and the net change in C as the percentage increase in experimental eye minus the percentage increase in the control eye. C _baseline was measured as the volume per minute per mm Hg (μL/min/mm Hg) necessary to maintain a constant pressure of 8 mm Hg within the eye for at least 30 minutes before the exchange. C _final was measured as the volume per minute per mm Hg (μL/min/mm Hg) necessary to maintain a constant pressure of 8 mm Hg after the exchange of either medium or drug treatment and analyzed during the 180 minutes after the exchange. Average net changes were analyzed by a two-tailed, paired Student's t-test, assuming unequal variance. Differences were considered significant at P < 0.05. 

Perfusion experiments concluded with an anterior chamber exchange and perfusion with 4% PFA at 8 mm Hg for 30 minutes. The eyes were removed from the system, submersion fixed for 2 weeks in 4% PFA, and hemisected by cutting the outflow tissue into wedges for use in immunohistochemical analysis. 

Results

To determine the expression pattern of PG-EP₄ receptors in the conventional outflow pathway of human eyes, we probed tissue sections taken from human donor eyes with IgGs that specifically recognize PG-EP₄ receptors (Fig. 1). The results showed that PG-EP₄ receptors were expressed by both the TM and SC cells. Of note, the SC cells demonstrated more robust staining than did the TM cells, comparable to the strong staining observed in the ciliary muscle and nonpigmented cells of the ciliary epithelium. Expression of PG-EP₄ receptors in TM and SC cells was confirmed with a cell surface biotinylation strategy that serves to concentrate low-density surface proteins in cultures of human cells. We observed that both the TM and SC cells expressed a protein at the plasma membrane that was immunoreactive with antibodies to the PG-EP₄ receptor and that corresponded to the molecular weight and banding pattern of a glycosylated PG-EP₄ receptor (Fig. 2). 

Figure 1.

 

PG-EP₄ receptor expression in human conventional outflow pathway. Immunohistochemical analysis of angle structures including TM, SC, CM, and CP. Sagittal sections of eye tissue were probed with anti-EP₄ IgG followed by visualization with DAB (A, D, G, J) at low (A–C) and high (D–L) magnifications. The TM showed diffuse (light brown) positive staining (D; arrows). The SC outer wall stained dark brown (D, arrowheads). Strong positive staining was also found in the CM (G, arrows), along with the nonpigmented cells of the CP (J, arrowheads). (B, E, H, K) The negative controls showed background labeling without incubation with the primary antibody. (C, F, I, L) Hematoxylin and eosin staining showed the orientation and structure of the tissues. The images are from one human donor eye (age 80) and are representative of four eyes that were examined. Scale bar, 100 μm.

Figure 2.

 

PG-EP₄ receptors at the surface of cultured SC and TM cell monolayers. Western blot analysis of bound (B) and unbound (U) protein captured by cell surface biotinylation and streptavidin chromatography of cultured TM and SC cell monolayers. Blots were probed with anti-EP₄ IgG and β-actin, used as a control for cytosolic protein contamination. PG-EP₂- and -EP₄-transfected HEK 293 cells were used as positive controls. Shown is a representative experiment of four total, executed on three different TM and three different SC cell strains.

In addition to PG-EP₄ receptors, cells of the conventional outflow pathway also express PG-EP₂ receptors. ²² To determine the concentration of the PG-EP₄ agonist, 3,7-dithiPGE₁ necessary to activate PG-EP₄ but not PG-EP₂ receptors, we used HEK 293 cells stably expressing PG-EP₄ or -EP₂ receptors in cAMP accumulation assays. The results showed that 3,7-dithiaPGE₁ stimulated cAMP accumulation in a dose-dependent manner in cells expressing either receptor subtype (Fig. 3A). Significantly, 3,7-dithiaPGE₁ was more potent at the PG-EP₄ receptor (EC₅₀ = 4.2 × 10⁻¹⁰ M; n = 4) than the PG-EP₂ (EC₅₀ = 1.4 × 10⁻⁷ M; n = 4), exhibiting ∼1000-fold greater selectivity. To test the functionality of the PG-EP₄ receptors present in both TM and SC primary cells, we stimulated the cells with increasing concentrations of 3,7-dithiaPGE₁ and measured the accumulation of cAMP (Fig. 3B). We found differential responses between TM (EC₅₀ = 1.7 × 10⁻⁷; n = 5) and SC (EC₅₀ = 6.3 × 10⁻⁹; n = 4) cells, particularly at 10 nM, a dose selective for PG-EP₄ receptors. 

Figure 3.

 

Concentration–response relationship for 3,7-dithiaPGE₁ in (A) PG-EP₄ and PG-EP₂ receptor stably transfected HEK 293 cells in addition to (B) primary cultures of TM and SC cell monolayers. The accumulation of cAMP was used as an indicator of drug efficacy. Shown are combined results of four experiments in HEK cells, five total in TM and four total in SC, using two different TM and SC cell strains conducted on different days.

To determine whether the cAMP accumulation was due to a GPCR-mediated event, we used a second assay to test a concentration of 3,7-dithiaPGE₁ (10 nM), which activates PG-EP₄ but not PG-EP₂ receptors. In HEK cells stably expressing PG-EP₄ receptors, we monitored the translocation of β-arrestin, a protein that binds to GPCR receptors on activation. After 60 seconds of treatment with 10 nM 3,7-dithiPGE₁, we observed that PG-EP₄ receptors and β-arrestin co-localized at the plasma membrane, with β-arrestin mobilizing to the surface from intracellular locations (Fig. 4). We also noted the internalization of the PG-EP₄ receptors after 60 seconds of treatment. 

Figure 4.

 

PG-EP₄ receptor-mediated translocation of β-arrestin. HEK 293 cells stably expressing PG-EP₄ receptors were treated with 10 nM 3,7-dithiPGE₁ for 60 seconds (D–F) or remained untreated (A–C). Localization of PG-EP₄ receptor (green) and β-arrestin (red) was monitored by immunofluorescence confocal microscopy (1-μm optical sections) which were digitally merged (C, F) to determine co-localization (yellow). Blue arrows: PG-EP₄ receptors localized to the membrane. White arrowheads: (B, C) localization of β-arrestin in the cytoplasm; (D–F) internalization of receptor. White arrows: β-arrestin distributed in the cytoplasm in cells where PG-EP₄ receptors are not present. Shown is a representative experiment of four total performed on different days. Bar, 5 μm.

Because of the activity of 3,7-dithiPGE₁ at two receptor subtypes and the potential differences in G-protein coupling, we used selective PG-EP₄ and -EP₂ antagonists to determine receptor activation. To initially characterize the specificity and selectivity of both the PG-EP₄ receptor antagonist (GW627368) and the PG-EP₂ receptor antagonist (AH6809), we measured the cAMP accumulation in the EP₄- and EP₂-HEK cells treated with 3,7-dithiPGE₁ in the presence of each antagonist separately (Fig. 5). Our results in the EP₄-HEK cells treated with 3,7-dithiaPGE₁ in the presence of 1 μM GW627368 (PG-EP₄R antagonist) showed inhibition of cAMP accumulation and a rightward shift in EC₅₀ (EC₅₀ = 4.2 × 10⁻¹⁰ M; n = 4 vs. EC₅₀ = 2.6 × 10⁻⁷ M; n = 4), with no significant effect on the EP₂-HEK cells (EC₅₀ = 8.7 × 10⁻⁸ M; n = 3; Figs. 5A, 5C). Similarly, the PG-EP₂ antagonist AH6809 had no significant effect on cAMP accumulation in the EP₄-HEK cells (EC₅₀ = 2.3 × 10⁻¹⁰ M; n = 3); however, it exhibited a rightward shift in EC₅₀ in the EP₂-HEK cells (EC₅₀ = 1.4 × 10⁻⁷ M; n = 4 vs. EC₅₀ = 1.2 × 10⁻⁶ M; n = 3). In primary cultures of SC cell monolayers, we found that 3,7-dithiPGE₁-stimulated cAMP accumulation shifted the dose–response curve rightward in the presence of 1 μM GW627368 (EC₅₀ = 6.3 × 10⁻⁹ M; n = 4 vs. EC₅₀ = 1.09 × 10⁻² M; n = 4). Co-incubation of the agonist with 5 μM AH6809 (PG-EP₂ receptor antagonist) exhibited a reduced efficacy suggestive of noncompetitive antagonist activity, although, there was no significant shift in the EC₅₀ (EC₅₀ = 4.1 × 10⁻⁹ M; n = 5; Figs. 6A, 6B). 

Figure 5.

 

Selectivity of PG-EP₄ and -EP₂ receptor antagonists. EP₄-HEK (A) and EP₂-HEK (C) cells were treated with increasing concentrations of 3,7-dithiaPGE₁ in the presence of either 1 μM PG-EP₄ receptor antagonist GW627368 (blue) or 5 μM PG-EP₂ antagonist AH6809 (green). Concentration–response curves for agonist alone are shown in black (A) and gray (C). Antagonist- and agonist-alone controls are shown in (B) for EP₄-HEK and in (D) for EP₂-HEK. Shown are combined results of four experiments for each cell type done in triplicate.

Figure 6.

 

Effects of PG-EP₄ and PG-EP₂ receptor-selective antagonist on 3,7-dithiaPGE₁-mediated cAMP accumulation in TM and SC cell monolayers. cAMP accumulation was measured in human SC (A) and TM (C) cells in response to increasing concentrations of 3,7-dithiaPGE₁ in the presence or absence of 1 μM PG-EP₄ receptor antagonist (GW627368) or 5 μM PG-EP₂ receptor antagonist (AH6809). Antagonist- and agonist-alone controls for these experiments are shown in (B; SC cells) and (D; TM cells). These data are cumulative of four experiments done with each antagonist in at least two different cell strains for each cell type.

In the TM cells, we observed no change in cAMP accumulation in the presence of 3,7-dithiaPGE₁ and 1 μM GW627368 (EC₅₀ = 5.9 × 10⁻⁷ M; n = 5). However, when the TM cells were treated with 3,7-dithiaPGE₁ and 5 μM AH6809, a partial decrease in the efficacy consistent with PG-EP₂ receptor occupation was observed in the cells, with no change in the EC₅₀ (EC₅₀ = 7.3 × 10⁻⁸ M; n = 3). Cells treated with either the PG-EP₄ antagonist or the PG-EP₂ antagonist alone showed no significant activity (Figs. 6B, 6D). 

To determine the role of PG-EP₄ receptors in conventional outflow regulation, we perfused human eyes with 3,7-dithiaPGE₁ and measured the infusion flow rate needed to maintain a constant pressure (8 mm Hg) over time. The contralateral eyes served as paired controls. Six pairs of donor eyes were used with a mean age of 83 years at time of death (Table 1). Baseline facility measurements were similar between the control (0.18 ± 0.03 μL/min/mm Hg) and drug-treated eyes (0.21 ± 0.03 μL/min/mm Hg; Table 1) and were similar to values previously reported by our group. ²⁰ After the chamber fluid exchange, the results showed that 3,7-dithiPGE₁ significantly and rapidly (within 20 minutes) increased outflow facility above baseline by 51% ± 18% (Fig. 7A; P = 0.05). When compared to the contralateral eyes, outflow facility in the drug-treated eyes increased by 69% ± 23% (Fig. 7B, P < 0.01, n = 6). Analyses of conventional outflow tissues in sagittal sections (Figs. 7C, 7D) showed no discernable differences in gross morphology, number of cells or appearance between the drug-treated and contralateral control eyes. 

Table 1.

 

Whole-Eye Organ-Perfusion Donor Information

Donor ID	Age (y)	Sex	Time to Enucleation (h)	Time to Perfusion (h)	Baseline C (μL/min/mm Hg)	Ending C (μL/min/mm Hg)	Increase in C (%)	Net Change in C (%)
01-C	91	F	2.3	17.3	0.239	0.129	−46.02	161.23
01-D					0.250	0.538	115.20
04-C	93	F	3.1	13.5	0.233	0.102	−56.22	96.22
04-D					0.260	0.364	40.00
18-C	81	M	2.4	20.3	0.110	0.160	45.45	10.42
18-D					0.170	0.265	55.88
22-C	92	F	NA	18.0	0.164	0.100	−39.02	76.27
22-D					0.247	0.339	37.25
29-C	86	M	2.3	16.0	0.218	0.239	9.63	3.34
29-D					0.239	0.270	12.97
43-C	55	M	3.5	12.0	0.089	0.103	15.73	13.79
43-D					0.105	0.136	29.52
Mean-C	83	3F/3M	2.7	16.2	0.193	0.146	−17.24	60.21
Mean-D					0.233	0.355	52.26

 C, control; D, drug-treated.

Figure 7.

 

Effect of 3,7-dithiPGE₁ on outflow facility in perfused enucleated human eyes. (A) Summary of perfusion data (mean ± SEM for six eyes). Measured outflow facility was normalized to baseline readings (30 minutes) before treatment. Anterior chamber contents were exchanged with normal media (gray) or normal media containing 10 nM 3,7-dithiaPGE₁ (black). *Time points when facility between drug-treated and contralateral eyes are significantly different (P < 0.05). (B) A comparison of mean (±SEM) outflow facilities for drug-treated and media-perfused eyes 30 minutes before (solid black, solid gray) and after (striped black, striped gray) drug exchange. Significance between before and after the exchange in drug-treated and medium-perfused eyes was analyzed by one-way ANOVA followed by the Bonferroni multiple comparison test (P < 0.05). (C, D) Histologic sections (sagittal) through angle tissues of perfused human eyes (Donor #22) examined by H&E staining.

Discussion

This study demonstrates that PG-EP₄ receptors are expressed by both SC and TM cells of the human conventional outflow pathway and that activation of these receptors, using a selective agonist, results in an immediate and substantial increase in outflow facility. Our observations suggest that PG-EP₄ receptor activation via PGE₂-mediated paracrine signaling between resident cells of the human conventional outflow pathway participates in the regulation of aqueous humor dynamics. 

Our expression results are consistent with previous studies that found PG-EP₄ receptors expressed throughout the conventional outflow pathway. ^{22 –24} Our study extended these initial findings by validating the presence of PG-EP₄ receptors in cultured human TM and SC cells. In a significant finding, we demonstrated that PG-EP₄ receptors were located at the plasma membrane by using a cell surface biotinylation technique. This observation is consistent with previous studies showing functional PG-EP₄ receptors in immortalized human TM cells (TM-3). ²⁵ However, our data provide the first evidence of PG-EP₄ expression in primary cultures of human TM and SC cells. 

Currently, there are two compounds under development that target PG-EP₄ receptors and efficaciously lower IOP in both canines and nonhuman primates. ^8,9 Studies in nonhuman primates used the same compound tested in the present study, 3,7-dithiaPGE₁, and showed that IOP lowering was due to effects on conventional outflow and not aqueous humor production or uveoscleral outflow. ⁹ Of note, in a recent study by Toris et al. (unpublished data, 2010), the PG-EP₄ receptor agonist lowered IOP in monkeys, apparently via a uveoscleral-dominant mechanism. Because of species differences in the effects of the PG-FP and -EP₄ receptor agonists on conventional versus unconventional outflow, we examined 3,7-dithiaPGE₁ in perfused human eyes. In contrast to results with PG-FP receptor agonists, we observed that 3,7-dithiaPGE₁ increased outflow facility in humans, similar to the previous report in nonhuman primates. Using histologic analysis, we examined the anterior angle—specifically, the TM and SC tissues. Although we identified no morphologic changes with light microscopy, the morphologic correlations with increases in outflow facility may be detectable at the electron microscopic level. 

Together, the variety of PG receptor subtypes in the conventional outflow pathway and the observation that independently activating these subtypes results in increased outflow facility suggest that receptor and G-protein coordination is complicated. For example, both PG-FP and -EP₄ agonists increase outflow facility but traditionally couple to different signal transduction systems. PG-FP receptors trigger intracellular calcium mobilization and therefore affect the contractile machinery within cells through G_q. Interestingly, data from our laboratory and others have shown that activation of PG-FP receptors relaxes both TM and ciliary muscle (CM) cells. ²⁶ On the other hand, PG-EP₄ receptors stimulate adenylate cyclase to increase cAMP, which is commonly known for its role in relaxation by stimulating PKA and by inhibiting the activation of myosin-light chain kinase (MLCK). In the present study, activation of the PG-EP₄ receptor stimulated cAMP production and mobilized β-arrestin, indicating G_s coupling. Some studies have shown that activation of PG-EP₄ receptors drive several pathways, including the inhibition of cAMP through G_i. ²⁷ The results of the present study are consistent with G_s activation in PG-EP₄-transfected HEK 293 and SC primary cells; however, even though we identified the PG-EP₄ receptor in TM primary cells to be at the membrane surface, we were not able to detect receptor activation by monitoring cAMP accumulation. It is possible, that the PG-EP₄ receptor couples to another G-protein in TM cells. Therefore, further experiments must be performed to determine the exact coupling and specific cellular effects. 

In this study, we characterized both PG-EP₄ and -EP₂ receptors by using selective PG-EP₂ and -EP₄ antagonists. We showed that the 3,7-dithiPGE₁ dose-dependent effect on cAMP seen in the transfected and primary systems were subject to blockade by corresponding receptor antagonists. We selectively blocked the accumulation of cAMP in both EP₄-HEK and SC cells with a PG-EP₄ antagonist that had no significant effect at the PG-EP₂ receptor. That this antagonist did not have any effect on the accumulation of cAMP in TM cells is consistent with the PG-EP₄ receptor's acting through an alternative G-protein. In addition, we were able to inhibit the accumulation of cAMP in EP₂-HEK cells by using a selective PG-EP₂ antagonist that appeared to have no significant activity at the PG-EP₄ receptor. However, we did observe a decrease in the maximum cAMP response, consistent with noncompetitive antagonist activity, although previous studies have classified AH6809 as a simple competitive antagonist by using a linear Schild plot analysis. ²⁸ There was also a dramatically greater shift in EC₅₀ for the SC cells treated with increasing concentrations of 3,7-dithiaPGE₁ in the presence of 1 μM GW627368, most likely because of the blocking of additional PG receptors, which may contribute to the effect of PG-EP₄ in these endogenous cells. For example, a study in which GW627368 was used showed that this compound also antagonizes the PG-DP receptors. ²⁹  

Clearly, the PGs play an important role in controlling the removal of aqueous humor from the eye. We know that PGF_2α, -E₂, and -D₂ all affect outflow by increasing fluid flow through a combination of primary and secondary routes. In addition, other members of the prostanoid family, such as prostacyclin and thromboxane, have exhibited potential to be used as IOP-lowering agents. ³⁰ Although PG analogs efficaciously lower IOP by affecting both uveoscleral and trabecular outflow, presently, we do not have a daily therapeutic available that primarily targets the conventional outflow pathway, the diseased tissue in ocular hypertension. Perhaps the development of PG-EP₄ analogues will finally provide a drug that selectively, safely, and efficaciously activates receptors in the conventional pathway to lower pressure in those with POAG, thus offering patients a complementary medical treatment to those currently available. 

Footnotes

 Supported by Allergan and Research to Prevent Blindness Foundation.

Footnotes

 Disclosure: L.H. Millard, None; D.F. Woodward, Allergan (E); W.D. Stamer, Allergan (F, R)

The authors thank John W. Regan and Brian S. McKay for helpful discussions during the planning, troubleshooting, and interpretation of experiments and the Lion's Clubs of Arizona for the donation of their time in the transportation of the human eyes.