Protein expression, biochemical pharmacology of signal transduction, and relation to intraocular pressure modulation by bradykinin B₂ receptors in ciliary muscle

Immunohistochemical determination of bradykinin receptors in ocular tissues

Since the human ciliary body contained a relatively high level of the mRNA for the B₂ receptor [19], it was necessary to determine whether the cells in this tissue contained the respective B₂-receptor protein. Thus, three human donor eyes obtained from a local eye bank and two cynomolgus monkey eyes (Covance Research Products; Madison, WI) were fixed in 4% alcoholic-zinc-paraformaldehyde fixative, processed into paraffin (CPI 09,013), and sectioned at 4 microns. Sections were antigen-retrieved and incubated with rabbit antihuman B₂ receptor (Novus NLS797) or control rabbit immunoglobulin G (IgG; Jackson ImmunoResearch, West Grove, PA) overnight. Primary antibody labeling was detected with biotinylated donkey antirabbit IgG (Jackson ImmunoResearch) and streptavidin-horseradish peroxidase conjugate (DAKO, Carpinteria, CA). Labeling was developed with 3-amino-9-ethylcarbazole (AEC), a high sensitivity peroxidase substrate (DAKO). Sections were counter-stained with hematoxylin. Unfortunately, suitable antibodies for the rabbit B₂ receptor are not commercially available; thus, similar studies could not be conducted on rabbit eye sections to correlate with the IOP studies described in the following sections.

[³H]-BK binding to human cloned B₂ receptor

To first define the B₂-receptor binding affinity of the key BK-related peptides for subsequent concentration selection for cell-based experiments, we first decided to conduct radioligand binding experiments. However, in view of the difficulty of obtaining sufficient h-CM cells for these initial studies, we decided to use cell membranes of Chinese hamster ovary cells containing the human cloned B₂ receptor (CHO-B₂; Euroscreen, Gosselies, Belgium). Thus, CHO-B₂ cell homogenates (2.5 µg protein) were incubated for 60 min at 23 °C with 0.2 nM [³H]-BK (95.5 Ci/mmole; PerkinElmer, Cambridge, MA) in the absence or presence of the test compound in a buffer containing 50 mM Tris/HCl (pH 7.4), 0.2 g/l 1–10-phenanthroline, and 0.1% bovine serum albumin [5,21]. Non-specific binding was determined in the presence of 1 µM BK. Following this incubation, the samples were filtered rapidly under vacuum through glass fiber filters (PerkinElmer) presoaked with 0.3% polyethyleneimine and rinsed several times with an ice-cold 50 mM Tris-HCl buffer using a 96-sample cell harvester (TomTec; Gaithersburg, MD). The filters were air-dried and the radioactivity counted in a beta-scintillation counter (TopCount, PerkinElmer). Data were analyzed using a sigmoidal-fit, iterative algorithm of a computer program designed to automatically fit the new data (Activity base; IDBS, Surrey, UK) [5,21]. The equilibrium inhibition constants (K_i, drug concentration required to inhibit the binding by 50%) were calculated and the mean±standard error of the mean (SEM) values determined thereafter.

Functional assay measuring [Ca²⁺]_i mobilization in cultured human ciliary muscle cells

Since the mRNA and protein immunoreactivity for the B₂ receptor was confirmed to be present in human h-CM, it was deemed important to demonstrate that functionally coupled and pharmacologically sensitive B₂ receptors existed in this tissue. Accordingly, we isolated h-CM cells and studied BK-induced [Ca²⁺]_i mobilization in these cells using the Fluorescence Imaging Plate Reader (FLIPR; Molecular Devices, Menlo Park, CA) instrument as previously described [24-27]. Low passage h-CM cells were seeded at a density of about 20,000 cells/well in a black-wall, 96-well tissue culture plates and grown for 2 days. On the day of the experiment, one vial of FLIPR Calcium Assay Kit dye (Molecular Devices) was resuspended in 50 ml of a FLIPR buffer consisting of Hank’s Balanced Salt Solution (HBSS), 20 mM HEPES, and 2.5 mM probenecid, pH 7.4. Cells were loaded with the calcium-sensitive dye (proprietary to Molecular Devices) by adding an equal volume (50 µl) to each well of the 96-well plate and incubated with dye for 1 h at 23 °C. Then the test compound plate and the cell plate were placed in the FLIPR instrument, and an aliquot (25 μl) of the BK solution (positive control) or the test compound was added to the existing 100 μl dye-loaded cells at a dispensing speed of 50 μl/sec. Fluorescence data were collected in real time over a period of time. Responses were measured as peak fluorescence intensity minus basal and where appropriate are expressed as a percentage of a maximum BK-induced response [E_max %]. When antagonist studies were performed, the latter were incubated with the cells for 15 min before BK was added. When the effects of various cellular signaling inhibitors/treatments were investigated, they were added to the cells 5 min before BK was added. The potencies (EC₅₀ or IC₅₀ values) of agonists and antagonists were determined from concentration-response curves using a sigmoidal-fit, iterative algorithm of a computer program designed to automatically fit the raw data (ActivityBase, IDBS, Surrey, UK). The relative intrinsic activities (E_max values) of the agonists were determined relative to reference agents where the maximal response was set to 100%. Concentration-response curves and other plots for the manuscript were constructed using Origin® software package (Microcal® Software Inc., Northampton, MA) [24-27].

Measurement of prostaglandins in human ciliary muscle and Chinese hamster ovary cells expressing B₂ receptors

Since functionally coupled B₂ receptors were demonstrated in h-CM cells, we decided to study a downstream event from the [Ca²⁺]_i mobilization that BK and related peptides may induce. This involved stimulating h-CM cells and CHO-B₂ cells with various concentrations of test peptides and measuring the release of various PGs in the extracellular medium. The concentration-dependent effects of these peptides on total PGs, and specifically PGE₂, PGF_2α, and PGD₂ release, were studied. Cells were grown as a monolayer in a 24-well plate (precoated with fibronectin) to confluence and subsequently serum-starved overnight before the experiment. On the day of the experiment, just before the drug was added, the wells were aspirated completely, and 990 µl basal medium and 10 µl 100× series of concentrations of the different peptide agonists were added to achieve the desired final concentrations of each compound (e.g., 0.1 nM to 10 µM). Plates were incubated with BK agonists for 1 h at 37 °C in the incubator. At the end of the incubation period, the cellular supernatants were collected and stored at −80 °C or used immediately to determine the PG content. Cell supernatants were tested for total PG levels and sub-species of PGs (PGE₂, PGF_2α, and PGD₂) using specific competitive enzyme immunoassays (Cayman Chemical Co., Ann Arbor, MI) as directed in the manufacturer’s recommended protocols using the supplied standards and reagents. The limits of detection for various PGs were as follows: total PGs=29 pg/ml; PGE₂=15 pg/ml; PGF_2α=9 pg/ml; PGD₂=200 pg/ml.

Measurement of nitric oxide production

In certain cell types, BK elevates intracellular levels of nitric oxide (NO), which in turn causes the generation of cGMP via the soluble guanylate cyclase in the cell cytosol [3,28]. In the present studies, we determined whether the B₂ receptors in h-CM cells were coupled to NO production using a specific kit (Promega Corp., Madison, WI) and according to the manufacturer’s instructions.

Measurement of extracellular signal-regulated kinase-1/2 phosphorylation and pro-matrix metalloproteinase production

Since BK has been shown to stimulate ERK1/2 phosphorylation in h-TM cells as an early step in the release of matrix metalloproteinases (MMPs) [23], we sought to ascertain whether a similar phenomenon was operable in h-CM cells. In the current assay, an homogeneous-time-resolved fluorescence (HTRF) assay (CisBio; Bedford, MA) was used. The HTRF assay is based on a sandwich immunoassay using an antiphospho-ERK1/2 antibody labeled with D₂ and an anti-ERK1/2 antibody labeled with Eu³⁺-Cryptate. In brief, h-CM cells (passage 7) were cultured in 96-well half-area white plates and incubated overnight, at 37 °C in CO₂ atmosphere. Cells were then incubated in a serum-free culture medium overnight before being exposed to BK or vehicle at 23 °C. Lysis buffer provided in the HTRF kit was used to lyse the cells at 23 °C with shaking for 30 min after the agonist treatment. HTRF conjugates containing phospho-ERK1/2 antibodies were then added to the lysates and incubated for an additional 2 h at 23 °C. The fluorescence signal was recorded using a Tecan M1000 fluorescence plate reader (Molecular Devices) at 620 nm for the donor and 665 nm for the acceptor.

The release of pro-MMP-1/2/3 in the extracellular medium of h-CM cells exposed to kinin’s, for 24 h at 37 °C [29,30] was studied next. Thus, h-CM cells isolated from multiple human donor eyes were cultured as described above and then serum-starved for 24 h prior to the addition of any test compound. Cells were incubated with the vehicle or test compound for 24 hours at 37 ºC after which the medium was collected and concentrated ten-fold using Centricon concentrators (Millipore Corp., Bellireca, MA). Equal volumes of this concentrate were loaded on 10% SDS-polyacrylamide gels (SDS-PAGE) and the proteins separated and transferred to nitrocellulose membranes. The membranes were then blocked with 5% non-fat dry milk (Bio-Rad Laboratories, Hercules, CA) followed by incubation with antibodies to pro-MMP-1, pro-MMP-2, or pro-MMP-3 for 12 hours at 4ºC with gentle shaking. After washing, these membranes were incubated with appropriate secondary antibodies (HRP-conjugated; dilution 1:3000) for 1 hour at 20 ºC. The pre-stained proteins and molecular weight markers were run in parallel. For chemiluminescent detection, the membranes were treated with enhanced chemiluminescent reagent (Amersham Pharmacia, Piscataway, NJ) and the signal was quantified using a Bio-Rad Versadoc imaging system (Bio-Rad Laboratories, Hercules, CA) [29,30].

Intraocular pressure measurements

To extend our in vitro observations of the high density of B₂-receptor protein in human and cynomolgus monkey ciliary muscle and non-pigmented ciliary epithelium, we studied the effects of topically administered BK (100 µg total dose in a sterile vehicle) on IOP in normotensive and TM-ablated hypertensive eyes of these animals [30,31]. Since no remarkable changes in IOP were observed, we concluded that most likely the peptide did not penetrate the cornea to reach the anterior chamber. However, since we could not subject the monkeys to additional studies involving BK delivery to the eye interior, it was decided that intravitreal (ivt) injection of BK into pigmented Dutch-Belt rabbits would be worth pursuing. Thus, after light corneal anesthesia with 0.1% proparacaine, basal IOPs were determined in up to 10 Dutch-Belt rabbits using an Alcon Pneumotonometer (Alcon Laboratories Inc., Fort Worth, TX) [30,31]. After this time, 50 µg of BK (B₂ agonist) or Des-Arg⁹-BK (B₁ agonist) in 20 µl of sterile basal salt solution (BSS plus), or just BSS plus vehicle, was injected ivt into one eye of each rabbit and the IOP determined at various times thereafter. Authors confirm adherence to the ARVO Statement for Animals in Ophthalmic and Vision Research, with all animal studies conducted according the guidelines for Animal Care and Use at Alcon Research Ltd.

B₂-receptor immunohistochemistry

The protein encoded by the mRNA for the B₂-receptor subtype was examined with immunohistochemical procedures conducted on thin sections of postmortem human and cynomolgus monkey eyes. A high level of B₂-receptor protein immunoreactivity was expressed by human (Figure 2A,B) and monkey ciliary bodies (ciliary muscle and non-pigmented ciliary epithelium; Figure 2C,D). The negative control experiments conducted on adjacent sections of the same eyes using IgG control showed no specific labeling of the B₂-receptor protein (i.e. sections B and D for each species).

Mobilization of [Ca²⁺]_i

To determine whether the B₂-receptor subtype immunoreactivity observed in human CM tissue (Figure 2A) related to functionally active receptors, we isolated h-CM cells from multiple human donor eyes and performed cell-based functional assays to monitor increases in [Ca²⁺]_i in response to BK and the related family of peptides. Accordingly, h-CM cells exhibited elevated levels of mobilized [Ca²⁺]_i when challenged with different concentrations of these peptides, for example, BK itself (Figure 3A). The collated concentration-response data obtained from many such studies (e.g., Figure 3B) indicated the following rank order to potency of these compounds: Hyp³-BK EC₅₀=2.2±0.2 nM=BK EC₅₀=2.4±0.2 nM > Lys-BK EC₅₀=3.2±0.8 nM=RMP-7 EC₅₀=3.7±1.2 nM > Met-Lys-BK EC₅₀=16.1 nM >> Des-Arg⁹-BK EC₅₀=4.2 µM (Table 1). The two BK receptor antagonists, HOE-140 (IC₅₀=1.4±0.1 nM; Figure 3C) and WIN-64338 (IC₅₀=174±18 nM; Figure 3D), abrogated the BK-induced responses in h-CM cells with potencies (Table 1, second column) that matched their affinities for the B₂ receptor determined by [³H]-BK binding to the cloned human B₂ receptor expressed in the cell membranes of CHO-B₂ cells (Table 1, last column).

In subsequent experiments, we used various inhibitors of cell signaling and other treatments to determine the source of the [Ca²⁺]_i in h-CM cells detected in the FLIPR Tetra experiments. Preincubating h-CM cells with ethylene glycol tetraacetic acid (EGTA) for 5 min to chelate extracellular Ca²⁺ before BK was added caused a diminution of the [Ca²⁺]_i response, with complete abolition of the response to BK with 1 mM EGTA (Figure 4A). Exposing h-CM cells to a PLC inhibitor (U73122; 0.1 nM to 10 µM) before BK (100 nM) was added resulted in a concentration-dependent reduction in the Ca²⁺-mobilizing effects of BK (Figure 4B), indicating that the Ca²⁺ monitored was originating, at least partially, from the endoplasmic reticulum (ER) and was being mobilized by the IP₃-generation pathway [7]. Indeed, when the ER pool of Ca²⁺ was depleted by preincubating h-CM cells with thapsigargin (0.1 or 30 µM), the [Ca²⁺]_i levels mobilized by BK were reduced by >50% compared to the control cells that had not been exposed to thapsigargin (Figure 4C), thus suggesting an alternative source of Ca²⁺ in addition to the ER possibly involving entry via Ca²⁺ channels on the cell membrane. Interestingly, when extracellular Ca²⁺ was depleted when 1 mM EGTA was included, BK continued to raise [Ca²⁺]_i by a small amount, and this was also concentration-dependently reduced by thapsigargin (Figure 4D). Since the ER also contains a ryanodine-sensitive Ca²⁺ channel, in addition to the IP₃-sensitive one [32], we were interested to determine whether adding ryanodine to h-CM cells enhanced the mobilization of [Ca²⁺]_i above and beyond that could be induced by BK itself. However, ryanodine (1–30 µM) did not enhance but instead slightly reduced the effect of BK on the [Ca²⁺]_i levels (data not shown).

Production and release of prostaglandins

Since the human TM cells exposed to various receptor agonists, including BK, produced various PGs [33], we decided to explore whether a similar phenomenon occurred in h-CM cells. The basal levels of total PGs released were lower in h-CM cells than in the CHO-B₂ cells (mean of 56 versus 165 pg/ml of total PGs, respectively). BK and its analogs stimulated the generation and secretion of total PGs into the extracellular medium in a concentration-dependent manner in h-CM cells (Figure 5A; Table 2) and in CHO-B₂ (Figure 5B; Table 2) cells, with potencies and rank order of activity similar to their [Ca²⁺]_i-mobilization effects (Table 1). However, the peptides were generally less potent in the PG release assays than in the [Ca²⁺]_i mobilization assays. BK induced up to a sevenfold increase in the levels of total PGs relative to basal levels in h-CM cells depending on the donor cells, while a 4.9-fold elevation in extracellular PGs was observed in the CHO-B₂ cells. BK-induced total PGs production in the h-CM cells was, for example, 393.3±24.2 pg/ml (n=8) at concentrations ranging from 0.1 to 10 µM. In CHO-B₂ cells, the amounts secreted were 913.6±47.6 pg/ml (n=6). PGE₂ levels were the most elevated followed by PGF_2α in response to BK, and no PGD₂ was detected in h-CM cells with or without BK treatment. The potencies of BK and BK-related peptides at promoting PGE₂ and PGF_2α release in h-CM cells were similar (Table 3) and compared well with their effects on total PG release (Table 2). CHO-B₂ cells were much more responsive to the peptides than h-CM cells since the agonist potencies for initiating secretion of PGE₂, PGF_2α, and PGD₂ were 100–1000 fold greater (Tables 3).

Even though various B₂ antagonists (HOE-140; (S)-WIN-64338; (R)-WIN-64338) and inhibitors of cyclooxygenases (COX; flurbiprofen; bromfenac) on their own failed to influence PG production in h-CM and CHO-B₂ cells (e.g., Figure 6A; Table 2), the antagonists and inhibitors blocked the BK-induced total PGs production in both cell types (Figure 6B, Figure 7A,B).

Measurement of nitric oxide

In limited experiments, we showed that while the NO donor sodium nitroprusside (1 mM) increased NO in h-CM cells by ninefold above basal levels, BK (0.1 and 1 µM) failed to have any effect on NO levels (data not shown).

Extracellular signal-regulated kinase-1/2 phosphorylation and pro-matrix metalloproteinase production

Initial assay development included a cell number and concentration-response titration of BK stimulation in the h-CM and CHO-B₂ cells with subsequent evaluation of phospho-ERK1/2 using the HTRF kit. The basal background levels of ERK1/2 phosphorylation were much higher in the CHO-B₂ cells, and we did not obtain a meaningful response to BK in these cells. However, the h-CM cells yielded a better signal-to-noise response ratio and lower basal background levels of phosphorylation. The stimulation of ERK1/2 phosphorylation was 1.9±0.3 fold (n=3) in the presence of 100 nM BK (or RMP-7) compared to the basal ERK1/2 phosphorylation level using 50k h-CM cells/well and a 10 min incubation period. In subsequent concentration-response experiments, we found the potency of BK ranged between 20 nM and 80 nM, with a maximal effect at 1 µM, and HOE-140 (100 nM) reduced the ERK1/2 phosphorylation induced by BK down to basal levels.

Since BK and related peptides released PGs from h-CM cells (see above), we expected that the generation of MMPs would follow as has been previously demonstrated for FP-receptor agonists PGF_2α and PHXA85 in h-CM cells [29,33]. Accordingly, BK (10 nM to 1 µM for 24 h at 37 ^○C) stimulated release of pro-MMP-1 (3.12±0.46-fold above basal level, n=6), pro-MMP-2 (1.67±0.25-fold above basal level, n=6), and pro-MMP-3 (1.41±0.09-fold above basal level, n=5). Likewise, RMP-7 (1–10 µM) also increased the secretion of pro-MMP-1 and pro-MMP-2 1.4–2.1-fold above basal levels in h-CM cells following a 24 h incubation. In contrast, the peptide B₂-receptor antagonist, HOE-140 (10 nM to 1 µM) alone, failed to alter levels of pro-MMP-1 (0.94±0.13-fold of basal level), pro-MMP-2 (0.87±0.1-fold of basal level), and pro-MMP-3 (0.95±0.1-fold of basal level; all n=6) in the h-CM cells.

Intraocular pressure changes induced by kinins

We studied the effects of topical ocularly administered BK (100 µg) in the ocular normotensive and hypertensive eyes of conscious cynomolgus monkeys to link the B₂-receptor immunoreactivity observations in the CM and non-pigmented ciliary epithelium (NPCE) tissues in this species. However, due to the poor ocular penetration of BK across the monkey cornea and conjunctiva, and the metabolic instability of this peptide, no alteration in IOP was observed in the monkey eyes (Table 4). In contrast, when BK (50 µg) was injected into the vitreous of the Dutch-Belt rabbits, a time-dependent reduction in IOP was observed, with a maximal 37.0±5.6% reduction 8 h post-injection (Figure 8). However, neither the vehicle nor Des-Arg⁹-BK (50 µg) injected ivt influenced rabbit IOP (Figure 8). Thus, the IOP-lowering effect of BK in the rabbits was specific and mediated by the B₂ receptors.