The β-Adrenergic Receptor Antagonist Metipranolol Blunts Zinc-Induced Photoreceptor and RPE Apoptosis

Abstract

purpose. To determine the effect of zinc on retinal cells at concentrations at which it is known to cause oxidative stress. Furthermore, the effects of metipranolol, known to prevent retinal damage, and of other antiglaucoma drugs were determined on zinc-injured retinal cells.

methods. Lipid peroxidation assays were conducted on rat brain and bovine retina–retinal pigment epithelial (RPE) membrane preparations. Immunohistochemistry, immunoblot analysis and the terminal-deoxynucleotidyl transferase dUTP-linked nick-end labeling (TUNEL) procedure determined the effects of zinc with or without trolox or metipranolol on photoreceptor death in situ. The effect of treatments on cultured RPE cells was analyzed using cell viability assays, immunoblot analysis, and the TUNEL procedure.

results. Zinc-induced lipid peroxidation of rat brain and bovine retina–RPE membranes, although the effect of the latter was of a (twofold) greater magnitude. Both effects, however, were similarly attenuated by metipranolol, desacetylmetipranolol, and trolox. Antiglaucoma drugs other than metipranolol had no effect. Intraocular injection of 150 μM zinc and treatment of cultured RPE cells with zinc led to mainly photoreceptor apoptosis and apoptotic death of RPE cells (50% death at 18 μM rising to 10% at 50 μM), respectively. Zinc-induced apoptosis of cultured RPE cells and photoreceptors were attenuated only by metipranolol and trolox.

conclusions. The combined data suggest that oxidative injury to RPE cells and photoreceptors may be caused by elevated levels of zinc in diseases such as age-related macular degeneration (AMD) and that metipranolol may act as an efficacious antioxidant to blunt this process.

The essential trace element, zinc, plays an integral role in the activation of enzymes involved in cell signaling, proliferation, and differentiation, as well as in DNA binding of many nuclear regulatory elements. ^{1 2} In the central nervous system (CNS), zinc also functions as a neurosecretory cofactor and, as such, is highly concentrated in the synaptic vesicles of specific glutamatergic neurons, the zinc-containing neurons. ³ These are mainly found in the forebrain, where, in mammals they have evolved into a complex associational network connecting cortical and limbic structures. ^{4 5} Much zinc is also found in the retina (see Ugarte and Osborne, for review ⁶ ), primarily associated with photoreceptors and the retinal pigmented epithelium (RPE) ^{7 8} and in glutamatergic synaptic vesicles in both plexiform layers. ⁹ It is thought that zinc is coreleased with glutamate from photoreceptors, playing an important role in modulating postsynaptic processes. ^{6 10}  

In the CNS, zinc can alleviate the effects of focal cerebral ischemia in vivo ^{11 12} and, on depletion, can induce cortical neuronal death. ¹³ In contrast, zinc has been reported to exacerbate ischemic or excitotoxic death in the cortex, ^{14 15 16} cerebellum, ¹⁷ and hippocampus ^{18 19 20 21 22} and in the nigrostriatal dopaminergic system. ²³ It is believed that zinc-induced neurotoxicity is primarily manifest at high concentrations through oxidative stress, ^{23 24 25 26} although inhibition of glycolysis, ²⁷ gene induction, ²⁸ activation of PKC ²⁹ and opening of mitochondrial permeability transition pores ³⁰ have also been suggested to be involved. It is obvious from previous studies, then, that zinc can either be protective or destructive according to the particular concentration and experimental system used. 

Oxidative stress has been suggested to be the cause of photoreceptor and RPE cell loss in age-related macular degeneration (AMD). ^{31 32 33} However, the cause of this oxidative stress remains unknown. One possibility is an increase in the extracellular zinc concentration. Aging may result in cell membranes functioning less efficiently, with free zinc leaking out of photoreceptors and RPE cells. Previous studies have shown that zinc induces oxidative damage to RPE cells, ^{25 26} but there are no data describing whether photoreceptors are likewise affected. The purpose of the present study was to investigate this possibility. Moreover, it is known that only one of the many presently used antiglaucoma substances, metipranolol, has antioxidant properties, ³⁴ and this compound has been shown to attenuate photoreceptor apoptosis induced by sodium nitroprusside (SNP). ³⁵ Studies were therefore conducted to determine whether metipranolol was also able to counteract insults to RPE cells and/or photoreceptors caused by zinc. 

Materials and Methods

Experimental Design

All experiments conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Home Office in the United Kingdom (as outlined in the Scientific Procedures Act, 1986). Adult Wistar rats (200–250 g; Physiology Department, University of Oxford, UK) were housed in a temperature- and humidity-controlled room with a 12-hour light–dark cycle and provided with food and water ad libitum. 

Measurement of Lipid Peroxidation in Rat Brain and Bovine Retina-RPE Extracts

Adult Wistar rats were stunned and killed by decapitation, which avoided the use of any anesthetic, and cerebral tissue was rapidly dissected. Retina-RPE was rapidly dissected from several bovine eyes received from the abattoir. Tissues were homogenized in 10 volumes of ice-cold 0.9% saline (pH 7.0) using a motor-driven polytetrafluoroethylene-glass homogenizer. The homogenate in each case was centrifuged at 1000g for 10 minutes at 4°C and the resultant low-speed supernatant was used for lipid peroxidation assays. The rate of membrane lipid peroxidation was determined from the amount of formed thiobarbituric acid reactive species (TBARS). Aliquots of supernatant (0.5 mL) were preincubated at 37°C for 5 minutes with 0.3 mL of 0.9% saline (pH 7.0) that in some instances contained 100 μM trolox (final concentration; Sigma-Aldrich, Poole, UK) or metipranolol (Dr. Mann Pharma, Berlin, Germany). Lipid peroxidation was initiated by the addition of 0.1 mL zinc sulfate (0–350 μM, final concentration; or 0–200 μM if testing with drugs), SNP (0–100 μM, final concentration; or 0–50 μM if testing with drugs), or buffer vehicle. After a 45-minute incubation at 37°C, lipid peroxidation was stopped by placing tubes on ice. TBARS were determined as described previously, ³⁶ with some modifications, ³⁵ by measuring color products resulting from the reaction of thiobarbituric acid with compounds formed during lipid peroxidation such as malonaldehyde. Briefly, the color reaction was developed by the sequential addition of 0.2 mL 8.1% (wt/vol) sodium dodecyl sulfate (SDS), 1.5 mL 20% (vol/vol) acetic acid (pH 3.5), and 1.5 mL 0.8% (wt/vol) thiobarbituric acid. This mixture was incubated for 30 minutes in a boiling water bath. After cooling with tap water, 1.5 mL of n-butanol-pyridine (15:1 vol/vol) was added and the reaction mixture centrifuged at 2500g for 10 minutes. Absorbance of the organic layer was measured at a wavelength of 532 nm and the amount of TBARS determined using a standard curve constructed using the malondialdehyde derivative 1,1,3,3,-tetraethoxypropane (1–30 nmol). The protein concentrations in brain and retina/RPE supernatants were determined with a bicinchoninic acid protein assay kit (Sigma-Aldrich) using bovine serum albumin as the standard. 

Injection of Zinc into the Eye

Adult Wistar rats (females; 200–250 g; eight separate experiments comprising six to eight animals each time) were anesthetized by an intramuscular injection of a combination of diazepam (0.4 mL/kg; Janssen, Grove, UK) and Hypnorm (0.3 mg/kg; Janssen). Pupils were dilated by topical application of 1% tropicamide drops. Zinc sulfate was dissolved in normal saline and a 5-μL volume injected as a bolus into the superotemporal region of the vitreous humor, just behind the lens (final concentration in vitreous was usually 150 μM) with a syringe (Hamilton, Reno, NV) approximately 5 mm from the optic nerve (eye assumed to be of approximate volume, 100 μL). Control eyes were injected with saline. Rats were subsequently replaced in their cages and housed under normal lighting conditions (12-hour light–dark). Three days after injection the animals were killed (all within 20 minutes of midday for any single experiment) and their retinas were either processed for analysis by Western blot analysis or fixed in 2% paraformaldehyde and 10 μm frozen sections cut for immunohistochemistry and staining for breakdown of DNA (using the terminal-deoxynucleotidyl transferase dUTP-linked nick-end labeling, or TUNEL procedure). 

Immunohistochemistry

Frozen retinal sections were incubated overnight at 4°C with mouse anti-Bcl2 monoclonal antibody (2 μg/mL clone C-2; 1:100; sc-7382; Santa Cruz Biotechnology, Santa Cruz, CA) mouse anti-poly-ADP-ribose-polymerase (PARP; 0.25 μg/mL 1:400; recognizes only the cleaved form of PARP by immunohistochemistry; ASP214; BD-Pharmingen, Oxford, UK), mouse anti-rhodopsin kinase (1 μg/mL MAI-720; 1:1000; Affinity BioReagents, Cambridge BioScience, Cambridge, UK) or mouse anti-caspase-3 monoclonal antibody (1 μg/mL Clone 46, 1:100; recognizes the cleaved form of caspase-3; 611048; BD-Pharmingen). They were then developed with appropriate secondary antibodies conjugated with fluorescein (Sigma-Aldrich). 

The TUNEL Procedure for Determination of DNA Breakdown

The method used was essentially the same as that of Gavrieli et al. ³⁷ Thawed, frozen sections or coverslips containing fixed RPE cells ^{38 39} were incubated for 5 minutes at room temperature in 30 mM Tris-HCl (pH 7.8). Incubation was continued at 37°C for 10 minutes in the presence of pepsin (800 U/mL) to expose the free DNA ends. After the pepsin was washed away with Tris-HCl, sections and coverslips were preincubated for 10 minutes with buffer A (30 mM Tris-HCl [pH 7.2] 140 mM sodium cacodylate, 1 mM cobalt chloride) and then placed in a humid chamber for 60 minutes at 37°C with buffer A containing 0.2 U/μL terminal deoxynucleotidyl transferase (TdT; Roche Diagnostics, Livingston, Scotland, UK) and 15 μM biotin-16-dUTP (Roche Diagnostics). The reaction was stopped by a 15-minute wash in sodium citrate buffer (300 mM NaCl, 30 mM sodium citrate [pH 7.3]) at room temperature. Sections and coverslips were then rinsed for 10 minutes in Tris-buffered saline (TBS: 20 mM Tris and 150 mM NaCl [pH 7.3]) and for a further 10 minutes in TBS solution containing 1% bovine serum albumin. Finally, sections were exposed for 10 to 20 minutes to Cy3-conjugated streptavidin, whereas coverslips were developed using the Vector Laboratories (Cambridge, UK) avidin-biotin conjugate kit. A final washing step for each was in TBS buffer. Sections were then mounted in buffered glycerol containing phenylendiamine to reduce fluorescence fading. RPE cells on coverslips were counterstained with 30% Gill’s hematoxylin in distilled water for 7 seconds before washing in water and mounting. In the case of cultured RPE cells, quantification was achieved by counting the number of TUNEL-labeled nuclei and the total number of cells on five randomly selected fields per coverslip (by a single observer), and expressing data as a percentage of TUNEL-labeled cells. Obviously, some dead cells lost adherence to coverslips after treatments and this was reflected in decreased cell numbers. Only cells with intensely labeled nuclei, combined with a shrunken cytoplasmic area were considered apoptotic. 

To generate a positive control for the TUNEL-labeling methodology, fixed RPE cells on coverslips were treated with 1 U/μL DNase (bovine pancreas, type II, Sigma-Aldrich) in buffer A for the 10-minute preincubation period before treatment with TdT, as mentioned earlier. 

Electrophoresis and Western Blot Analysis

Retinas or RPE cells were homogenized in freshly prepared 20 mM Tris-HCl buffer (pH 7.4) containing 2 mM EDTA, 0.5 mM ethylene-glycol-tetracetic acid (EGTA), and the protease inhibitors, phenylmethylsulfonyl fluoride (PMSF; 0.1 mM), leupeptin (50 μg/mL), aprotinin (50 μg/mL), and pepstatin A (50 μg/mL). An equal volume of sample buffer (62.5 mM Tris-HCl [pH 7.4], containing 4% sodium dodecyl sulfate, 10% glycerol, 10% β-mercaptoethanol, and 0.002% bromphenol blue) was added, and the samples were boiled for 3 minutes. An aliquot was taken for determination of protein content. Electrophoresis of samples was performed using 10% polyacrylamide gels containing 0.1% sodium dodecyl sulfate, as described by Laemmli. ⁴⁰ Samples were then transferred onto nitrocellulose according to the method of Towbin et al. ⁴¹ The nitrocellulose blots were then incubated with mouse monoclonal anti-caspase-3 (clone 46; 1:1000; recognizes the cleaved form of caspase-3), mouse anti-PARP (1:1000; recognizes both the cleaved and noncleaved forms by immunoblot), mouse monoclonal anti-Bcl2 antibody (Clone C-2; 1:200), mouse anti-rhodopsin kinase (1:1000), or mouse anti-actin (1:2000; Chemicon, Chandler’s Ford, UK) antibodies for 3 hours at room temperature and appropriate secondary antibodies conjugated to horseradish peroxidase subsequently used. Nitrocellulose blots were developed with a 0.016% solution of 3-amino-9-ethylcarbazole in 50 mM sodium acetate (pH 5) containing 0.05% Tween-20 and 0.03% H₂O₂. 

Cultured Human RPE Experiments

All culture medium components except for gentamicin (Sigma-Aldrich) were obtained from Invitrogen (Paisley, Scotland, UK). Culture plasticware was from Fahrenheit (Milton Keynes, UK), and all other chemicals were obtained from Sigma-Aldrich. RPE cells were cultured from donor human eyes obtained within 48 hours of enucleation from the Bristol Eye Bank (Bristol, UK). Cultures were prepared from three separate donors (two male and one female) with different ages (28, 46, and 54 years). All culture work was conducted in a sterile laminar flow hood, and the RPE cells were characterized by labeling for cytokeratin (K_G 8.13), as described previously. ³⁹ Culture medium consisted of nutrient mixture Ham’s-F10 supplemented with 10% fetal calf serum (FCS), 0.4% glucose (final concentration, 25 mM), 2 mM glutamine, amphotericin B (25 μg/mL), and gentamicin (100 μg/mL). Cultures were maintained in a humidified incubator at 35.5°C with an atmosphere containing 5% carbon dioxide in normal air. Confluent cultures in flasks were passaged at a ratio of 1:3, and after the third passage some cells were transferred to either 96-well plates or to 13-mm borosilicate glass coverslips in 24-well plates at the approximate seeding density of 2.0 × 10⁴ cells per well. Confluent RPE cells in 96-well plates (passage 4) were then subjected to various experimental conditions as described in the figure legends. 

Cultures were assessed for viability of RPE cells by using the 3,[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) reduction analysis method, modified from that of Mosmann. ⁴² Briefly, cells were subjected to appropriate treatments for 23 hours and then MTT was added to wells at a final concentration of 0.5 mg/mL for a further hour at 37°C. After this time, medium was removed from the cultures and reduced MTT (a blue formazan product) was solubilized by adding 100 μL dimethyl sulfoxide (DMSO) to each well. After agitation of plates for 15 minutes, the optical density of the solubilized formazan product in each well was measured with an automatic microplate reader (Titertek Plus MS212; ICN Flow, Thame, UK) with a 570-nm test wavelength and a 690-nm reference wavelength. 

Statistical Analysis

All experiments were performed with appropriate controls performed in the same plates, with at least four determinations for each data point. Data were analyzed for significance using a one-way analysis of variance (ANOVA) followed by a Tukey multiple-comparison test and are expressed as a mean percentage of the control value plus SEM. P < 0.05 was considered significant. 

Results

Zinc stimulated TBARS formation in rat brain membranes in a dose-dependent manner (Fig. 1A) . Zinc and, for comparative purposes, SNP, both induced maximum stimulation of TBARS at concentrations of 200 and 100 μM, respectively. Moreover, at these maximum stimulations of TBARS, zinc was approximately 50% less effective when compared with the influence of SNP. Similar results were obtained for bovine retina/RPE samples (Fig. 1B) , but in this case, full dose-dependent responses were not undertaken due to limited tissue quantities. In addition, of all the substances tested, only metipranolol and its active metabolite, desacetylmetipranolol and trolox (vitamin E derivative) significantly attenuated both zinc and SNP-induced formation of TBARS both in brain (Table 1)and retina/RPE (Fig. 1B) . Thus, of all the antiglaucoma drugs tested, only metipranolol attenuated the oxidant properties of zinc, as has been reported previously for the alternative oxidant regimen, iron-ascorbate, ³⁴ and SNP. ³⁵  

Injection of zinc into the vitreous humor (concentration estimated to be 150 μM) showed TUNEL-positive cells to be primarily associated with the photoreceptor cell bodies (Fig. 2) . Very few TUNEL-positive cells were associated with other cell types in the inner retina and no labeling was seen in control sections. No attempt was made to quantify the number of TUNEL-positive cells because the numbers varied in different retinal sections (Fig. 2) . Moreover, the total number of TUNEL-positive cells per retina (estimated to be approximately 40% of photoreceptors) appeared optimum after 3 days, but this was not investigated in detail. Immunohistochemical analysis of sections showing labeling of TUNEL-positive cells also revealed clear expression of Bcl-2, caspase-3, and PARP immunoreactivities (Figs. 3 and 4) . Sections in which TUNEL-positive cells were absent did not show outer retinal staining for these antigens. Moreover, a correlation existed between the location of TUNEL-positive cells and where intense staining for Bcl-2, caspase-3, and PARP occurred. This correlation is shown, for example, in Figure 2for TUNEL-positive cells and caspase-3 immunoreactivity. 

When injected into the vitreous humor, zinc caused a dose-dependent reduction in retinal levels of rhodopsin kinase with the maximum effect being noted at the concentration of 150 μM (Fig. 5E) . To examine in a more detailed manner whether trolox and/or metipranolol attenuated zinc-induced photoreceptor apoptosis, analyses were particularly performed on whole retinal extracts where 150 μM zinc had been injected into the vitreous humor 3 days earlier. On its own, zinc significantly altered the amount of caspase-3, rhodopsin kinase, Bcl-2, and PARP relative to treatment with saline vehicle (Figs. 5A 5B 5C 5D) . Moreover, the effect of zinc was in all cases clearly counteracted (Fig. 5)when co-injected with metipranolol (100 μM) or trolox (50 μM). 

Figure 6Ashows the effect of zinc sulfate on both cell viability and induction of TUNEL-positive nuclei in cultured human RPE cells. It can be seen that zinc at concentrations of 10 μM or less had no obvious effect on the survival of the cells. However, slightly greater concentrations of zinc significantly affected the survival of the cells. Only approximately 50% of cells survived when the concentration of zinc was elevated to 18 μM. Cell survival was even more reduced as zinc concentrations increased, with only approximately 10% of cells surviving at a concentration of 50 μM. Evidence suggests that the zinc-induced death to RPE cells was by apoptosis, as it caused breakdown of DNA (after staining for TUNEL; Figs. 6B 6C ) and also an elevation in the cleaved form of PARP (89 kDa) and the active (∼17 kDa) form of caspase-3 (see Fig. 8 ). Metipranolol and trolox clearly reduced the elevation in the active form of caspase-3, whereas the effect was less obvious with PARP, in three separate experiments. Moreover, metipranolol or trolox significantly blunted zinc-induced apoptosis of RPE cells (Figs. 7 and 8) , whereas timolol, betaxolol, carteolol, and levobutanol were ineffective (Fig. 7A) . 

Discussion

A reduced oxygen supply to cells, as occurs in ischemia or possibly aging, results in an alteration in the distribution of intracellular and extracellular ions across membranes. Furthermore, ischemia has been implicated in the pathogenesis of AMD. ³¹ Thus, the possibility exists that at later stages of AMD, free zinc, which may be present at very high concentrations within the photoreceptors and RPE, may be released into the extracellular space. If this process occurs, then the present data suggest that the elevated level of zinc has the potential to cause oxidative stress to photoreceptors and RPE cells. In addition, evidence is provided to show that antioxidants like trolox or metipranolol would be beneficial to such cells. 

The role of zinc in AMD has been much discussed, especially in relation to the beneficial use of zinc supplementation, which has been generally supported by both laboratory and clinical data. ^{43 44 45 46 47 48 49} Also, low concentrations of zinc protect against oxidative stress of RPE cells. ^{25 50} Furthermore, important antioxidant RPE enzymes such as the metallothioneins are zinc-dependent, and their activities are affected by intake of dietary zinc. ⁵¹ However, the Eye Disease Case Control Study (EDCC) ⁵² and the Beaver Dam Eye Study (BDES) ⁵³ found no significant relationship between serum zinc levels and the risk of AMD, in agreement with Stur et al. ⁴⁵ We interpret these findings to suggest that in the very initial stages of AMD when perhaps nutrient deprivation plays a part, zinc supplementation may be beneficial, as this element has an intricate involvement in cellular antioxidant defense. However, as the disease progresses, zinc may be released into the extracellular space from the RPE and photoreceptors, and at this stage the positive effect of zinc supplementation may become redundant and even contribute to the disease process. This hypothesis would also predict that antioxidants are likely to have a positive effect at all stages of the AMD, provided they are present in sufficient concentrations. 

In the present study it was clearly shown that zinc can act as an oxidant and can also induce death of RPE and photoreceptors. Evidence for the pro-oxidant property of zinc in comparison with SNP is documented by the lipid peroxidation studies (see Fig. 1 ). The measurement of TBARS is generally considered to provide an index for lipid peroxidation. From these studies it can be concluded that zinc is a much weaker oxidant than SNP, but that it does work in a similar manner to promote lipid peroxidation. Previous studies have shown SNP to be as effective an oxidant in our system as iron/ascorbate. ^{34 35} EC₅₀ values for stimulation of TBARS in the present study for zinc and SNP were determined to be 100 and 10 μM, respectively. Moreover, maximum stimulation of TBARS formation for SNP was approximately twice that of zinc. The mechanisms by which zinc acts as an oxidant probably involves several events that eventually result in a generation of free radicals. ⁵⁴ The effects on brain and retina-RPE membranes, shown in Figure 1 , clearly support the view that zinc can directly interact with membrane constituents to generate free radicals. In addition, the effect of zinc on TBARS formation was, like SNP (Table 1) , attenuated by metipranolol and trolox with other β-blockers used for the treatment of glaucoma having no effect. 

Intravitreal injection of 150 μM zinc clearly caused mainly photoreceptor death with approximately 40% of the cells staining positively for TUNEL and caspase-3-/Bcl-2. Photoreceptor labeling was not uniform throughout the retina and, in this way, it resembled data previously reported that described the injection of SNP into the vitreous. ³⁵ We suggest that this is related to the injection site. As found for SNP, ³⁵ the maximum amount of photoreceptor apoptosis after intravitreal injection of zinc was detected after approximately 3 days. Moreover, the effect of zinc (150 μM vitreous concentration) on photoreceptor apoptosis was less than that of SNP (50 μM). ³⁵ Thus, the effectiveness of zinc in comparison with SNP in inducing photoreceptor apoptosis in vivo resembled the in vitro lipid peroxidation studies on brain membranes. It is important to note that in preliminary experiments, zinc was injected into the vitreous humor at a concentration range of between 25 and 150 μM. It was clear from these studies that when zinc was present at the concentration of 150 μM, the appearance of photoreceptor apoptosis and its counteraction with metipranolol or trolox was optimal. Significantly, even when the concentration of zinc in the vitreous was at its greatest, no obvious detrimental effect appeared to be associated with the inner retina. Thus, we conclude that zinc, at an intravitreal concentration of 150 μM does not cause an acute, nonspecific lethal effect on the retina but rather that some photoreceptors are affected. 

Quantification of the degree of photoreceptor death from analysis of retinal sections was difficult due to the variability associated with injection both in relation to the site and the depth into the vitreous humor. It was concluded that it would be necessary to analyze every section throughout a single retina to obtain meaningful results from these analyses. Quantification was therefore restricted to an analysis of Western blot data derived from whole retinal extracts. This showed that zinc clearly caused a reduction in rhodopsin kinase, an enzyme associated exclusively with photoreceptors in the eye. ⁵⁵ Moreover, the active form of caspase-3, the cleaved form of PARP, and Bcl-2 were all increased. The increase in the cleaved forms of both caspase-3 and PARP were consistent with what is expected to occur in apoptosis, but this was not the case for Bcl-2, where it would be anticipated to decrease rather than increase. One possible explanation is that those photoreceptors that did not label positively for TUNEL were mildly injured, resulting in an increase in their Bcl-2 protein levels, in an attempt to prevent cell death. This being the case, then the overall retinal Bcl-2 level would appear to be elevated with respect to untreated samples. Significantly, zinc effects were counteracted when metipranolol or trolox were co-injected. These in vivo studies on photoreceptor apoptosis were therefore consistent with the in vitro lipid peroxidation studies conducted on brain membranes. 

In all experimental cases, retinal regions that showed no photoreceptor damage or apoptosis remained in contact with the RPE and appeared normal when analyzed by light microscopy. Retinal detachment was observed only at sites of photoreceptor damage. Furthermore, the RPE cell layer generally appeared normal throughout the eye, even in areas where photoreceptors were significantly damaged. Moreover, the effect of zinc injection on photoreceptors was localized in an obvious manner that reflected the distance from the injection site. Because there was a clear delineation in the toxic effect of zinc on retinal photoreceptors, and because no RPE cells appeared to have been induced to die by this ion after 3 days, in situ, at least when analyzed at the light microscopic level, then we could only conclude that retinal detachment occurred as a result of photoreceptor damage and did not precede this event. The fact that RPE cells appeared at a gross level to be unaffected by intraocular injections of zinc, in situ, at the time points analyzed in this study, implied that this ion did not penetrate as far as this epithelial monolayer in sufficient quantities to induce apoptosis. It is possible and, indeed, likely that zinc affected RPE cells in a less obvious, or sublethal manner and that this may have led to the death of these cells after longer periods of time (i.e., >3 days). This would especially be the case after photoreceptor loss and subsequent retinal detachment, because penetration of the ion into these cells would therefore have been greatly increased under these circumstances. Our studies, however, stopped short of describing the effects of intraocular injections of zinc on RPE cells in situ—merely concentrating on the neural retina. This is the reason that analysis of the RPE was restricted to cells in culture, since these remained a pure and homogeneous population, uncontaminated by other cell types. 

Studies of RPE cells revealed that zinc above a certain concentration also induced apoptotic cell death that was counteracted by metipranolol or trolox but not by other ophthalmic β-adrenergic receptor blockers. The question that arises from these observations is why RPE cells and photoreceptors are more prone to the oxidant effect of zinc than are other retinal cell-types. The answer may relate to common membrane characteristics of the two cell types. It is known that photoreceptors are particularly susceptible to free radical damage ⁵⁶ because of their high content of polyunsaturated fatty acids (e.g., DHA). ^{57 58} Whether RPE membranes contain such membrane components is unknown. The specific effect of zinc on photoreceptors after injection may also relate to its penetrance into the retina. Thus, even after injection of a vitreous concentration of 150 μM zinc, retinal levels may remain much lower, which could be sufficient only to cause death of photoreceptors because of their high DHA content. This notion would be compatible with the finding that 10 μM zinc has little effect on RPE cells as well as causing negligible lipid peroxidation to brain membranes. 

In a prior study, we suggested that the reason that only metipranolol, of the β-blockers used to treat glaucoma, acts as an antioxidant relates to its structure, ³⁵ as is the case with other nonophthalmic β-adrenoceptor antagonists. ^{59 60 61} Given that evidence exists from animal studies to show that topically applied drugs can reach the retina, ^{62 63} the possibility of employing metipranolol in this way to attenuate photoreceptor/RPE death in AMD is worthy of consideration. Metipranolol is completely converted in vivo to its active metabolite, desacetylmetipranolol, ⁶⁴ and this is an even more potent antioxidant than metipranolol (Table 1) . It is possible that the half-life of desacetylmetipranolol is greater than antioxidants such as vitamin E. If this is the case, then desacetylmetipranolol may accumulate in the eye with constant topical use, allowing for a more sustained antioxidant effect. It is theoretically possible that metipranolol merely binds to zinc and prevents its destructive action. However, the clear antioxidant activity of metipranolol and the fact that it will prevent photoreceptor destruction induced by other insults (e.g., nitric oxide). ³⁵  

In summary, the present study shows that above a certain concentration, zinc can cause apoptotic death of RPE cells in culture, as well as retinal photoreceptors after intravitreal injection. Furthermore, the nonselective β-blocker, metipranolol and its active metabolite, desacetylmetipranolol, can both partially attenuate zinc-induced lipid peroxidation, in a manner similar to trolox, and prevent zinc-induced retinal cell death. Given the proposed role of free radicals in the pathogenic destruction of RPE cells and photoreceptors in AMD, the remarkable antioxidant activity of metipranolol and desacetylmetipranolol may be of clinical importance. 

 Supported by European Community PRO-AGE-RET Grant QLK6-CT-2001-00385).

 Submitted for publication October 19, 2005; revised January 31 and March 2, 2006; accepted May 1, 2006.

 Disclosure: N.N. Osborne, None; J.P.M. Woods, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) The comparative effects of different concentrations of zinc and SNP in stimulating TBARS formation in rat brain membranes. Maximum stimulation of TBARS was achieved with approximately 50 to 100 μM and 100 to 200 μM SNP and zinc, respectively. Moreover, the maximum stimulation of TBARS for SNP was about twice that for zinc. Results are expressed as the mean ± SEM; n = 5 determinations for each. (B) Comparative effect of SNP (50 μM) and different concentrations of zinc in stimulating TBARS formation in bovine retina/RPE membranes and the blunting of the zinc responses with metipranolol. Results are expressed as the mean ± SEM; n = 4 determinations for each. *P < 0.05, comparing ZnSO₄ with ZnSO₄ in the presence of metipranolol, by one-way ANOVA followed by a Tukey multiple-comparison test (n = 4–8).

Table 1.

 

Effect of Antiglaucoma Drugs and Other Compounds on the Stimulation of Lipid Peroxidation (TBARS) Caused by SNP or ZnSO₄ in Rat Brain Homogenates

Drug	100 μM SNP	250 μM ZnSO4
10 μM Metipranolol	51 ± 7*	44 ± 6*
100 μM Metipranolol	15 ± 4*	18 ± 3*
10 μM Desacetylmetipranolol	20 ± 3*	29 ± 4*
100 μM Desacetylmetipranolol	10 ± 4*	12 ± 4*
100 μM Carteolol	102 ± 9	98 ± 8
100 μM Levobetaxolol	94 ± 8	104 ± 9
100 μM Levobunolol	98 ± 8	98 ± 10
100 μM Timolol	89 ± 9	94 ± 7
100 μM Propranolol	97 ± 7	105 ± 12
10 μM Brimonidine	89 ± 8	94 ± 8
100 μM Brimonidine	97 ± 11	89 ± 7
100 μM Dorzolamide	107 ± 112	100 ± 7
100 μM Pilocarpine	99 ± 9	97 ± 9
10 μM Latanoprost	98 ± 7	89 ± 8
10 μM Travoprost	104 ± 9	93 ± 7
10 μM Unoprostone	107 ± 9	99 ± 9
10 μM Trolox	17 ± 5*	21 ± 6*

 Results are expressed as %TBARS, relative to the effect of each stimulating substance (SNP, ZnSO₄; 100% reference value).

*   P < 0.01, by one-way ANOVA followed by a Tukey multiple-comparison test (n = 4–8 in each case).

Figure 2.

 

Typical variability in the number of TUNEL-positive cells and caspase-3-labeled cells in five pairs of retinal sections, after ZnSO₄ (vitreous concentration of 150 μM) had been injected into the vitreous humor of a single eye. There was a correlation between the number of TUNEL-positive cells (top images) and caspase-3 staining (bottom images). The intensity and numbers of cells labeled in each case probably reflects the distance from the injection site, as defined in the schematic. It can be seen that in areas close to the injection site, almost all cell bodies in the outer nuclear layer (ONL) were TUNEL-positive and were also intensely labeled for the presence of active caspase-3 (A, B). In other areas of the ONL, farther away from the site of injection, a few TUNEL-positive cells and caspase-3-immunoreactive cells were detected (C, D). In yet other areas of the ONL, at the greatest distance from the injection site, very few TUNEL-positive cells and caspase-3-immunoreactive cells were evident (E). Scale bars, 40 μm.

Figure 3.

 

Retinal sections obtained close to the injection site of zinc into the vitreous humor and processed for the localization of TUNEL-positive cells (A) and cleaved (active) caspase-3 immunoreactivity (B). It can be seen TUNEL-positive cells and strong caspase-3 labeling appeared to be exclusively located in the ONL. Also shown are sections of retina close to the site where vehicle was injected into the vitreous humor and processed for breakdown of DNA (TUNEL; C) and localization of active caspase-3 immunoreactivity (D), respectively. In these instances, the ONL contained no TUNEL-positive cells (C) or caspase-3 immunoreactivity (D). (B, D, arrows): blood vessels that were revealed by the secondary antibody used to localize caspase-3. OS, outer segments; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Scale bars, 40 μm.

Figure 4.

 

Retinal sections taken close to the injection site of zinc into the vitreous humor and processed for the localization of cleaved (inactive) PARP (A) and Bcl-2 (B) immunoreactivities. PARP and Bcl-2 immunoreactivities were both particularly associated with the ONL. (C, D) Sections of retina close to the site where vehicle was injected into the vitreous humor and processed for cleaved PARP and Bcl-2 localization, respectively. The ONL contained small amounts of cleaved PARP (C) and no Bcl-2 (D) immunoreactivity. (B, D, filled arrows): blood vessels that were revealed by the secondary antibody used to localize Bcl-2. Some PARP-immunoreactivity was associated with cell bodies (open arrows) located in the inner nuclear and ganglion cell layers in the normal retina (C). Scale bars, 40 μm.

Figure 5.

 

The effect of ZnSO₄ alone (intravitreal concentration of 150 μM) and ZnSO₄ in combination with metipranolol (100 μM) or trolox (50 μM) on retinal protein expression for Bcl-2 (A), cleaved caspase-3 (B), native PARP (C), and rhodopsin kinase (D). In all cases, ZnSO₄ caused an alteration in control protein levels that was partially corrected by the presence of either metipranolol or trolox. (E) The effect of different intravitreal zinc concentrations on retinal rhodopsin kinase content (photoreceptor loss) 3 days after injection. Neither metipranolol or trolox had any effect on levels of retinal proteins when added without ZnSO₄. *P < 0.05, comparing ZnSO₄-treated eyes with control eyes (n = 6–8). **P < 0.05, comparing ZnSO₄-treated eyes with those also having the test compound co-injected (n = 6–8). In both cases, analyses were performed by one-way ANOVA followed by a Tukey multiple-comparison test.

Figure 6.

 

Showing the influence of ZnSO₄ on cultured human RPE cells. (A) Cell viability was drastically decreased by ZnSO₄ in a dose-dependent manner. (B) Concurrent with loss of cell viability, ZnSO₄ also caused a dose-dependent increase in detectable TUNEL-labeled cell nuclei, as a marker of DNA breakdown. This is exemplified in (C) where control cells (in absence or presence of DNase, as a positive marker for DNA cleavage) are compared with those cells insulted with 18 μM ZnSO₄ for 24 hours in serum-free medium. Data shown in (A) and (B) are representative of mean ± SEM; n = 6 experiments comprising quadruplicate determinations.

Figure 7.

 

(A) The protective effect of metipranolol (MET) and trolox against 18 μM ZnSO₄-induced toxicity (CON) of cultured human RPE cells (24 hours, passage 4 cells, serum-free medium). Both compounds were most effective at 100 μM (A), whereas all other antiglaucoma drugs at 100 μM (betaxolol, BET; timolol, TIM; carteolol, CART; levobunolol, LEVO) were ineffective. (B) A more detailed influence of metipranolol and trolox over a range of ZnSO₄ concentrations. (A) *P < 0.05, comparing 18 μM ZnSO₄-treated cultures with control eyes; †P < 0.05, comparing metipranolol (100 μM) or 100 μM trolox-treated cultures with those exposed to ZnSO₄. (B) *P < 0.05, comparing metipranolol (100 μM) or 100 μM trolox-treated cultures with those exposed to ZnSO₄. Data shown in (A) and (B) are representative of the mean ± SEM; n = 6 experiments comprising quadruplicate determinations.

Figure 8.

 

Cultured human RPE cells exposed to 18 μM ZnSO₄ for 24 hours exhibited TUNEL-labeling that can be detected as a dark nuclear stain (A), which was absent when cells are also treated with 100 μM metipranolol (B) or 100 μM trolox (C). (D) Induction of DNA breakdown by exposure to ZnSO₄ was concurrent with PARP (loss of 116-kDa protein and increase in 89-kDa species) and caspase-3 (17 kDa) cleavage as detected by electrophoresis/blot analyses (lane 2). The cleavage of caspase-3 was clearly attenuated by the presence of metipranolol (lane 3) or trolox (lane 4) in three separate experiments. Lane 1: Control, untreated cell extracts shown for comparative purposes.

The authors thank Nigel Swietalski for his expert technical assistance.