Retinal Ischemia and Oxidative Stress

Figure 1 Terminology.

Figure 2 Ocular diseases where retinal ischaemia is implicated (retina blood supplies: retinal, choroidal, optic nerve).

not ischemic. Similarly, anaemia (generally a reduction, rather than complete absence of haemoglobin) is always a component of ischemia, but not vice versa.

Ischemia deprives a tissue of three requirements: oxygen, metabolic substrates, and removal of waste products. The loss of these requirements will initially lower homeostatic responses and with time will induce injury to the tissue. If withheld for a sufficiently long time the tissue will die (an infarct).

It is convenient to think of the retina as having three blood supplies: the retinal blood supply, the choroidal blood supply and the optic nerve head blood supply. Ischemia cause by one or the other of these blood supplies being affected has been implicated in a variety of ocular diseases (Figure 2).

OXIDATIVE STRESS

Oxidative stress is imposed on cells as a result of one of three factors: (1) an increase in oxidant generation, (2) a decrease in antioxidant protection, or (3) a failure to repair oxidative damage (Figure 1). Cell damage is induced by reactive oxygen species (ROS). ROS are either free radicals, reactive anions containing oxygen atoms, or molecules containing oxygen atoms that can either produce free radicals or are chemically activated by them. Examples are hydroxyl radical, superoxide, hydrogen peroxide, and peroxynitrite. The main source of ROS in vivo is aerobic respiration, although ROS are also produced by peroxisomal b-oxidation of fatty acids, microsomal cytochrome P450 metabolism of xenobiotic compounds, stimulation of phagocytosis by pathogens or lipopolysaccharides, arginine metabolism, and tissue specific enzymes. Under normal conditions, ROS are cleared from the cell by the action of superoxide dismutase (SOD), catalase, or glutathione (GSH) peroxidase. The main damage to cells results from the ROSinduced alteration of macromolecules such as polyunsaturated fatty acids in membrane lipids, essential proteins, and DNA. Additionally, oxidative stress and ROS have been implicated in retinal ischemic disease states, such as occurs in diabetic retinopathy, ophthalmic artery occlusion, central retinal artery occlusion and glaucoma.

RELATIVE RESISTANCE OF THE RETINA TO ISCHEMIA COMPARED WITH THE BRAIN

One striking difference between the retina and brain is the relative resistance of the retina to an ischemic insult. There is universal agreement that the retina survives considerably longer than the brain (Table 1). A few minutes of cerebral ischemia in the human results in widespread injury and death, but the primate retina can suffer up to 100 minutes of central retinal artery occlusion without permanent injury.1 Furthermore, the retina exhibits a regionalised sensitivity to ischemia, with the outer layers less sensitive than the inner layers. Significantly neuroglobin, a neurone-specific respiratory protein distantly related to haemoglobin and neuroglobin, is present in high amounts in the retina. The estimated concentration is 100-fold greater than in the brain and particularly located to the retinal photoreceptors.2 In ischemic models on holangiotic retinae that obstruct both retinal and choroidal, such as the rat elevated pressure-induced model,3 the photoreceptors incur less functional and structural injury than the inner retina. In contrast, in the merangiotic rabbit retina the reverse situation occurs, indicating fundamental inter-species variation in pathophysiology.4,5 The explanation for this phenomenon in vascularised retinae remains obscure, but may relate to the exceptional ability of the photoreceptors to extract energy from available sources anaerobically.6,7 Similarly, this explanation may extend to the other retinal layers

Table 1 Retinal Tolerance Time (Minimum Period of Ischemia to Cause Irreversible Damage) in Experimental Models of Retinal Ischemia(a)

(a) Adapted in part from.1 (b) Retinal tolerance time is the minimum period of ischemia required to cause irreversible damage.

when survival times are compared with the brain. This possible explanation is based on three facts:

(1) local energy: substrates are present (the vitreous contains considerable amounts of glucose, and there is a species-dependent store of glycogen in the retina)8; (2) these energy substrates are depleted during periods of retinal ischemia,9,10 (3) the isolated retina can efficiently extract adenosine triphosphate (ATP) from glycolysis as long as glucose is plentiful, even in the complete absence of oxygen.6,7 Alternatively, it is possible that retinal neurones are intrinsically more resistant to ischemia than cerebral neurones; however this ‘‘explanation’’ tells us nothing of any underlying mechanisms. Finally, the so called no-reflow phenomenon may explain the discrepancy.11,12 This term describes the situation in the oedematous brain after a period of ischemia: the rigid cranium cannot accommodate the swollen brain, which compresses the microvasculature producing ongoing ischemia despite restoration of macroscopic blood flow. In contrast, upon cessation of retinal ischemia, oedematous retina does not obstruct the microvasculature, as the thin retinal tissue has space (the vitreous cavity) to expand into.

Any contribution of retinal metabolism to its relative insensitivity to ischemia is an area that could conceivably be clinically exploited, but to date has received little attention. Buchi et al. (1991)13 using a pressure ischemia model in rats showed that if 5% dextrose was used as the infusate instead of 0.9% saline, then the histological injury to the retina for a given duration of ischemia was

markedly reduced. Romano et al. (1993)14 using a photothrombosis retinal ischemia model in rats, showed that a single intravitreal injection of glucose immediately prior to the ischemic insult attenuated the early histological changes compared to the saline-injected eyes. Similarly, using an isolated retina model, Romano et al. (1998)15 showed that glucose markedly reduced N-methyl-D- aspartate (NMDA)-induced excitotoxic injury. Results from our own laboratory support the findings of Buchi et al. (1991)13 and demonstrate remarkable functional (by electroretinography) and structural protection when glucose is used as the infusate.16 Furthermore, when hypoglycaemia is induced or a nonmetabolisable analogue of glucose is used (2-deoxyglucose), the retinal injury is greater than that observed with saline17 supporting the notion that the protective mechanism is metabolic. This area of research awaits further study.

RETINAL ISCHEMIA AND NEUROLOGICAL MECHANISMS

Mammalian retinal ischemia results in irreversible morphological and functional changes. These are the consequence of depleted ATP stores, due to deprivation of both glucose and oxygen, though transient loss of these substrates is not immediately lethal. The cell death is the result of an extremely complex (not completely understood) cascade of biochemical responses initiated by energy failure. The tissue damage and functional deficits that follow periods of transient ischemia reflect the combined effects of several, often interrelated pathophysiological pathways. These result in drastic changes in ion movements, neurotransmitter levels and metabolites.

Reperfusion may also damage cells that, until that moment, had sustained only reversible injury and were potentially salvable. This concept of oxygen restoration to ischemic tissue amplifying injury sustained during oxygen deprivation has its origins in experimental myocardial studies from the 1960s.18 Leakage of LDH, a marker of cell death caused by retinal ischemia, increases after reintroduction of oxygen in the continued glucose absence.19

It is useful to consider that stroke lesions consist of a densely ischemic focus, the ischemic core, surrounded by a better perfused area, the ischemic penumbra.20 Cells in the focus are usually doomed unless reperfusion is quickly instituted. In contrast, penumbral cells may remain viable for several hours and can be saved by reperfusion or by drugs that prevent the infarction extending into the penumbral zone.20

During the past decade a considerable amount of experimental work has been devoted to the elucidation of the mechanisms of ischemic neuronal injury, but there is still much debate over the underlying processes causing the injury. One area of interest is the role of glutamate and aspartate, whose extracellular concentrations increase markedly during ischemia.21 It is generally accepted that glutamate release during the early phase of brain ischemia triggers events leading to irreversible injury, not only in those areas in which oxygen supply is critically reduced, but also in regions of seemingly less disturbed energy

metabolism i.e. the penumbra of focal ischemia. However there is still much controversy over the extent of their role in the pathophysiology of neural ischemia. Moreover, neuronal bodies (grey matter) and axons (white matter) are not affected in precisely the same way by ischemia.22

It is important to remember that much of the work studying ischemic neuronal and axonal damage has been carried out on tissues derived from defined brain regions. So care is important when drawing comparisons with what happens in the retina. In particular, the way photoreceptors respond to ischemia in light and dark conditions may be unique. Moreover, the thinness of the retina, its well defined blood systems and the large glycogen supply associated with the Mu¨ller cells provides the tissue with a unique energy supply when compared with the brain, so making the response of tissues to ischemia not the same.

THE ROLE OF FREE RADICALS IN RETINAL ISCHEMIA

Many cascades generated by glutamate and glucose/oxygen deprivation result in the formation of free radicals23 and it has been proposed that free radicals are important mediators in damage caused by retinal ischemia.24,25 Reperfusion injury after ischemia appears paradoxical, but oxygen-derived and other free radicals are principally formed when reduced compounds, which accumulate during ischemia, are reoxidized (Figures 3, 4). There is evidence that this free radical burst, produced during the early stage of reperfusion, overwhelms normal cellular antioxidant defense mechanisms, causing oxidative stress and a variety of types of tissue injury.26

There are many ways in which free radicals can be formed during ische- mia-reperfusion, but the burst of superoxide radicals (·O2 ) which occurs during the early stage of reperfusion is thought to occur by the following pathway. During ischemia, degradation of ATP leads to the formation of hypoxanthine, and increases in intracellular calcium in neurones activate the Ca2þ-dependent protease calpain. Calpain converts xanthine dehydrogenase into xanthine oxidase and upon reperfusion the latter enzyme oxidizes the accumulated hypoxanthine to uric acid resulting in the release of O2 . These two molecules react by the Haber-Weiss mechanism to yield the highly toxic hydroxyl radical (·OH). This reaction is catalyzed by iron, which is released from its protein-bound stores at the low pH generated during ischemia. In addition, O2 interacts with nitric oxide (NO·), which is produced in considerable amounts following ischemia, leading to the formation of peroxynitrite, nitrosyl radical and eventually OH.26 It is not just from the mitochondrial systems of neuronal cells that free radicals are generated, activation of glial cells and infiltrating leukocytes release inflammatory mediators, such as arachidonic acid, nitric oxide and cytokines, which all play major roles in the formation of free radicals following ischemia.

An early indication that excessive free radical formation may be detrimental to the retina was the finding that iron-ascorbate perfusion rapidly attenuates the isolated rat retina’s b-wave and causes lipid peroxidation.27 However, the first studies to provide evidence for the involvement of free

Figure 3 Formation of reactive oxygen species and antioxidant mechanisms. Oxygen is converted to superoxide (O2· ) by oxidative enzymes in the endoplasmic reticulum (ER), mitochondria, plasma membrane, peroxisomes and cytosol. Superoxide is converted to H2O2 by dismutation and then to OH· by the copper and iron catalysed Fenton reaction. H2O2 is also derived directly from oxidases and peroxisomes. Free radicals cause lipid peroxidation, protein damage and DNA damage. Superoxide catalyses the reduction of iron so enhancing OH generation by the Fenton reaction. The major antioxidant enzymes are superoxide desmutase, catalase and glutathione peroxidase. GSH – reduced glutathione; GSSG – oxidised glutathione; NADPH – reduced nicotinamide adenine dinucleotide phosphate.

radicals in retinal damage after ischemia were performed by Szabo and colleagues who showed that administration of superoxide dismutase (SOD) affords protection against ischemia-induced histological damage28 and the ionic imbalance that occurs in the reperfusion period.29 The work of Szabo and colleagues showed, albeit indirectly, that the superoxide free radical is generated in such quantities during retinal ischemia-reperfusion that the normal endogenous levels of superoxide dismutase expressed by the retina are overwhelmed and are unable to protect the tissue from oxidative damage caused by this radical. Subsequent studies supported the involvement not only of superoxide in ischemic injury to the retina but also of hydrogen peroxide and the hydroxyl radi- cal.30–32 Furthermore, the inability of endogenous free radical quenching mechanisms to cope with the demand posed following ischemia is illustrated by the capacity of a variety of free radical scavengers, such as extract of Ginkgo biloba,28 a-lipoic acid,33 vitamin E,34,35 thioredoxin,36 an ascorbic acid

Figure 4 Production and reactions of nitrogen-derived radicals. (A) Nitric oxide synthase (NOS) produces nitric oxide radicals in the conversion of L-arginine to L-citrulline. This compound acts as an important homeostatic modulating agent under physiological conditions via its vasodilator, antioxidant, antiplatelet and antineutrophil actions. In the event that superoxide radicals are present (·O2 ), usually as a result of rapid tissue reoxygenation subsequent to an ischemic event, peroxynitrite is formed which rapidly decomposes to highly reactive oxidant species that can cause tissue injury. Under physiological conditions, there is a critical balance between cellular concentrations of NO, ·O2 and superoxide dismutase activity which favour NO production. In pathological conditions such as reperfusion following an ischemic event, the formation of ONOO is favoured. The latter compound can be rapidly detoxified if it is combined with reduced glutathione (GSH) to form S-nitrosoglutathione (GSNO), but this depends upon the cellular antioxidant defence system being functional and this is generally overwhelmed during tissue reperfusion. (B) Reactions and formation of nitric oxide radicals as a result of ischemia-reperfusion. The combined production of nitric oxide derived radicals and failure of cellular antioxidant defence will lead to widespread macromolecular damage and cell death.

derivative,37 mannitol38 and the iron chelator desferrioxamine39 to protect the retina from ischemia-reperfusion injury.

Although it can be inferred from all of these studies that there is an elevated level of free radicals in the retina after ischemia, direct measurement of free radical formation has only been performed by Muller and co-workers,30 who showed increased free radical formation both during and after ischemia, and by Szabo et al (1997)40 who subjected diabetic retinas to ischemia and documented increased levels of free radicals during the reperfusion phase.

As previously mentioned, xanthine oxidase has been established as an important source of oxygen free radicals in certain ischemia-reperfusion injuries and there is evidence for a similar action in the retina. Xanthine oxidase activity has been reported to increase 5-fold within 10 minutes of reperfusing the ischemic rat retina, while concentrations of hypoxanthine and xanthine, respectively the substrate and product of xanthine oxidase, increase in a time-related fashion following ischemia-reperfusion32,41 and enhance free radical formation.42 Further-

more, administration of allopurinol or oxypurinol, blockers of xanthine oxidase, both result in significant improvement of the ERG after ischemia.41,43 These

combined data suggest that xanthine oxidase-mediated processes contribute to the functional anomalies of retinal ischemia. However, there is limited evidence to suggest that xanthine oxidase may not be an important source of free radicals after retinal ischemia. Faberowski et al (1989)44 reported that allopurinol provided no significant protection against ischemia produced by transient ligation of the optic nerve, while Szabo et al (1993)45 found that allopurinol was only neuroprotective when administered in combination with extract of Ginkgo biloba. The reported variations may be due to either the allopurinol doses used, or the manner of its administration, or to the methods of evaluating injury.

One potential source of free radicals in the retina following ischemia are polymorphonuclear leukocytes. In the brain, neutrophils oxidize NADPH to generate superoxide and are important free radical donors during and after focal ischemia.46 Agents that prevent the accumulation or activation of neutrophils are protective.47 In the retina, infiltration of neutrophils occurs during the early phase of reperfusion, probably in response to increased levels of cytokines and free radical formation. Although there is no direct evidence for increased free radical formation from neutrophils, blocking leukocyte accumulation has been shown to afford protection to the ischemic retina.48

Oxygen-derived free radicals cause extensive cellular damage in the brain. One of the main mechanisms by which this occurs is by attacking unsaturated fatty acids, which leads to lipid peroxidation of membranes. This will result in loss of membrane fluidity, cell swelling, oedema and feed forward production of more free radicals. Many additional mechanisms of damage have been ascribed to free radicals generated during ischemia-reperfusion, including attacking sulphhydryl protein bonds, which leads to the destruction of amino acids and polypeptide chains, fragmentation of DNA molecules, which leads to activation of poly(ADP-ribose) polymerase, activation of cytokines and NF-kB, which are likely to be instrumental in upregulations of iNOS and COX-2 and subsequent release of glutamate, and effects on Ca2þ homeostasis.47 Interestingly, a clear association between neuronal cell death and formation of free radicals and lipid peroxides in the retina subjected to ischemia-reperfusion has only recently been demonstrated (Celebi et al., 2002).35 In all of these studies, ischemia-reperfusion induced free radical formation, lipid peroxidation and neuronal injury and were largely preventable by administration of free radical scavengers35,36 and a novel metal chelate.32

THE FREE RADICAL NITRIC OXIDE IN RETINAL ISCHEMIA

Nitric oxide is an important neuromediator throughout the CNS, and is implicated in many physiological processes in the retina.49 It is synthesised from L-arginine via the action of nitric oxide synthetase (NOS) and three distinct isoforms of NOS have been characterised. Neuronal NOS (nNOS) and endothelial NOS (eNOS) are Ca2þ-dependent and are constitutively expressed by a variety of nervous tissues and by endothelial cells of blood vessels, respectively. Inducible or immunologic NOS (iNOS) is Ca2þ-independent and is not generally found under normal physiological conditions but, as the name suggests, is induced by certain stimuli. In the normal retina, nNOS is synthesised by a variety of neurones,50 eNOS is only found in retinal vessels,51 while iNOS has been detected at low levels in Mu¨ller cells and the RPE.52

In the last decade, a significant body of evidence has indicated an involvement of NO in the pathogenesis of ischemic damage in the brain and retina. In models of focal and global ischemia in the brain, all three NOS isoforms are induced in the post-ischemic period, leading to a sustained production of NO. Protein levels of nNOS and eNOS are elevated shortly after the onset of ischemia, presumably due to the rise in intracellular Ca2þ, while the induction of iNOS is delayed by several hours. Many studies have attempted to determine the impact of NOS modulation on brain ischemia and, in general, the results indicate that nNOS and iNOS are detrimental but eNOS is beneficial to neurones.53

Increases in expression of all of the NOS isoforms have also been reported in the retina post-ischemia, although inevitably there is some variation between studies. This is likely the consequence of the different models used and the different durations of the ischemic insults. Increases in nNOS, generally in cells of the inner retina, have been shown following high IOP-induced ischemia54 and in the 2-vessel occlusion model of ischemia,55 although surprisingly, not after optic nerve bundle occlusion.56 In fact, levels of nNOS were seen to decline appreciably in this model of ischemia.57 Increases in eNOS have been detected after high IOP ischemia51 and optic nerve ligation.57 Substantial elevations of iNOS in the retina have been shown using the high IOP model of ischemia,58 in the 2-vessel occlusion model of ischemia,55 after optic nerve bundle occlusion,56 in a murine model of ischemic proliferative retinopathy,59 and in patients with diabetic retinopathy and ocular ischemic disease.60 The source of NO produced by iNOS are activated macrophages and other invading inflammatory cells, as well as retinal astrocytes and Mu¨ller cells.58

The question as to whether the post-ischemia increase in NO production is beneficial or detrimental to the retina has proved difficult to answer. Some studies suggest that activation of NOS causes cell death, whereas others have reported the opposite. This is partly due to the lack of specificity of the majority of pharmacological tools employed to date, partly due to the variety of different

experimental protocols, and partly to the complexity of the NO system in the retina: all three isoforms of NOS are present and inducible in different cell types of the retina and the different NOS isoforms will be preferentially involved in the different stages of ischemia-reperfusion.

The case for a positive effect of NO on the post-ischemic retina was first advanced by Veriac et al. (1993)61 who showed that recovery of the b-wave of the ERG in rabbits was more rapid after administration of a nitric oxide donor, yet delayed after administration of the NOS inhibitor L-NNA. These results were duplicated in rats by Hangai et al. (1999)57 who showed additionally that histological as well as functional damage to the retina was aggravated by L-NNA. The authors suggested that the detrimental effect of L-NNA results from delaying the onset of retinal reperfusion. It would seem logical to suppose, therefore, that activation of eNOS shortly after ischemia increases retinal perfusion and contributes to neuronal survival, as has been shown in the brain.62 However, L-NNA is also detrimental to the retina in an isolated retinal preparation of ischemia,63 and moreover, unlike in the brain, eNOS-deficient mice are not more susceptible to retinal ischemic damage.64 The group of Lipton have conducted a number of detailed studies over recent years which demonstrate, also, that charged cellular NO-derived equivalents (NOþ or NO ) have the ability to S-nitrosylate (transfer the NO group to) the sulfhydryl group of a cysteine residue, specifically in the active site of the ‘‘apoptotic executioner’’ enzyme, caspase-3 and on the external face of the NMDA receptor. In both cases, activity of the target proteins are decreased and neuroprotection is afforded. NO itself, however, shows no such activity, preferentially acting on superoxide species to produce peroxynitrite which can cause neurodegeneration.47

Notwithstanding these reports, the majority of studies, as in the brain, have shown that inhibition of NOS leads to histological and functional protection of the ischemic retina. The nonspecific inhibitors aminoguanidine,58 L-NAME,65 L-NNA66 and L-NMMA66 have all been shown to reduce neuronal injury caused by ischemia/reperfusion. The involvement of the iNOS isoform in the injury process can be inferred from the study of Hangai et al. (1996),56 who showed that administration of the relatively specific inhibitor of iNOS, L-NIO, partly ameliorated damage caused by CRAO. Clearer proof was provided by the study of Neufeld et al. (2002),58 who found that chronic administration of the potent and selective iNOS inhibitor, SC-51, afforded substantial protection against retinal ischemia, and a greater degree of protection than aminoguanidine. It is also of significance that in elevated IOP models of glaucoma there is an activation of iNOS in rat retinal non-neuronal cells and the formed NO causes degeneration of ganglion cells.67 It should also be borne in mind that activation of the NMDA receptor during an ischemic insult will lead to an increase in the levels of intracellular calcium and this itself is known to induce expression of and activate each NOS isoform, either directly (nNOS) or via the activation of calcium-dependent protein kinase C (PKC) isoforms.47

PROBABLE EVENTS ASSOCIATED WITH RETINAL ISCHEMIA

The complex pathophysiology of retinal ischemia underscores the dynamic relationship of the retina and its vascular supply. Incipient reduction of choroidal, retinal or optic nerve head blood supply alone will cause different retinal pathologies to that occurring when all supplies are affected at the same time. When blood flow is disturbed, the normal homeostatic mechanisms linking metabolic demands and haemodynamics are altered, affecting the different retinal cell-types in some way. Depletion of cellular energy stores beyond a critical threshold (that may vary for every retinal cell-type) triggers waves of depolarisation and a series of molecular events known as the ‘‘ischemic cascade’’ (that may vary for every retinal cell-type), irrespective of the initial mechanism of vascular compromise. Thus different types of vascular changes in the optic nerve head blood supply are suggested to lead to similar retinal pathologies, as proposed to be the case in glaucoma.68

The complex pathophysiology of retinal ischemia underscores the dynamic relationship of the retina and its vascular supply. Incipient reduction of choroidal, retinal or optic nerve head blood supply alone will cause different retinal pathologies to that occurring when all supplies are affected at the same time. When blood flow is disturbed, the normal homeostatic mechanisms linking metabolic demands and haemodynamics are altered, affecting the different retinal cell-types in some way. Depletion of cellular energy stores beyond a critical threshold (that may vary for every retinal cell-type) triggers waves of depolarisation and a series of molecular events known as the ‘‘ischemic cascade’’ (that may vary for every retinal cell-type), irrespective of the initial mechanism of vascular compromise. Thus different types of vascular changes in the optic nerve head blood supply are suggested to lead to similar retinal pathologies, as proposed to be the case in glaucoma.68 Although early perfusion may prevent permanent retinal injury, the limited tolerance of neurones to hypoxic stress imposes a restricted time window on thrombolytic therapy, although this is the most effective way to treat stroke.69 A time-defined therapeutic window could be wider for retinal than for brain neurones. The possible causes of retinal ischemia suggest a number ofobvious targets for therapy (Figure 6), one being simply to restore the nutrient supply. However, such an approach is always associated with a temporal delay, during which time the retina undergoes ischemic damage. Therefore, specific pharmacological strategies need to be developed to arrest any of the many putative cascades generated during ischemia. The ‘‘ischemic cascade’’ is a complex succession of interrelated pathological changes at the cel-

lular and molecular level whose knowledge derives mainly from studies on brain tissues.47,70,71 This cascade may be summarised as follows (Figure 5). Reduction

of energy caused by loss of normal blood flow results in cell membrane depolarisation and excessive release of neurotransmitters. These include excitatory amino acids such as glutamate that activate specific receptors, leading to deranged ion fluxes and the detrimental effects of intracellular Ca2þ

Figure 5 Cascade of events thought to occur in retinal ischemia. An interruption in the supply of blood to the retina leads to tissue ischemia which causes rapid failure of energy production and oxidative stress. This causes a number of biochemical events as outlined in the figure. Key steps include the failure of the Naþ/Kþ-ATPase pump, membrane depolarisation, cytoplasmic accumulation of sodium and calcium ions and the formation of destructive free radical species. The summed cellular response of these processes, if left unchecked, is cell death. This can occur by the classical and rapid necrotic process or by longer-duration apoptosis.

accumulation. Ca2þ interacts with intracellular proteins causing, amongst other events, degradation of structural cellular elements. Impairment of normal homeostasis then leads to cytotoxic oedema and anaerobic glycolysis will lead to progressive acidosis especially during prolonged hypoxia, compounding oedema formation and causing mitochondrial dysfunction. An accumulation of NO and the degradation of membrane phospholipids will contribute to the generation of neurotoxic free radical species. Inflammatory mediators will be activated to cause secondary neuronal injury through release of cytokines, phospholipases and chemokines. Superimposed waves of gene expression and growth factors will eventually induce cell death, mediated by a family of cysteine proteases (caspases). It should be emphasised, however, that precise knowledge of the temporal relationship and pathological relevance of these myriad processes remains unclear.

Figure 7 Reports supporting the view that reduction of oxidative stress attenuates retinal ischemia. HIOP – high intraocular pressure, ONBL – optic nerve bundle, 2VO – two vessel occlusion, 4VO – four vessel occlusion.

compound dimethylthiourea or other antioxidant treatments. Protection of the retina after ischemia/reperfusion has also been demonstrated with the following antioxidant strategies: vitamin E, a-lipoic acid, superoxide dismutase, catalase and EGB-761, CV-3611, desferrioxamine, mannitol and allopurinol. Antioxidant properties could also account for the ameliorative effects of calcium dobesilate, flupirtine, and trimetazidine in retinal ischemia.

Antioxidants which can readily reach the intracellular compartment and are well tolerated in long-term therapy, such as EGB-761, a Ginkgo biloba extract, may theoretically be prescribed prophylactically as an oral therapy to patients at a high risk of developing retinal ischemia. Furthermore, several of the above-mentioned antioxidants, such as superoxide dismutase, catalase, manitol and desferrioxamine, are protective when administered just before reperfusion following retinal ischemia, which is of major interest for potential clinical use. However, the rapid clearance of superoxide dismutase and catalase by the kidneys could make the therapeutic use of both enzymes difficult. It should be noted that, apart from directly counteracting the molecular damage caused by free radicals, antioxidant compounds may also attenuate ischemic insults by decreasing the release of glutamate evoked by ischemia.

CONCLUSION

Retinal ischemia is a common cause of visual impairment and blindness. At the cellular level, ischemic retinal injury consists of a number of self-reinforcing destructive cascades, initiated by energy failure, to cause neuronal depolarisation,

calcium influx and inevitably oxidative stress. The resulting cell death is by apoptosis and necrosis, pending on the intensity of the ischemic insult. Combating mild ischemic insults with appropriate antioxidants is one therapeutic strategy worthy of consideration.

REFERENCES

1.Hayreh SS, Weingeist TA. Experimental occlusion of the central artery of the retina. IV: retinal tolerance time to acute ischaemia. Br J Ophthalmol 1980; 64:818–825.

2.Schmidt M, Giessl A, Laufs T, et al. How does the eye breathe? Evidence for neuroglobin-mediated oxygen supply in the mammalian retina. J Biol Chem 2003; 278:1932–1935.

3.Osborne NN, Ugarte M, Chao M, et al. Neuroprotection in relation to retinal ischemia and relevance to glaucoma. Surv Ophthalmol 1999; 43(suppl 1):S102–S128.

4.Osborne NN, Larsen AK. Antigens associated with specific retinal cells are affected by ischaemia caused by raised intraocular pressure: effect of glutamate antagonists. Neurochem Int 1996; 29:263–270.

5.Osborne NN, Schwarz M, Pergande G. Protection of rabbit retina from ischemic injury by flupirtine. Invest Ophthalmol Vis Sci 1996; 37:274–280.

6.Winkler BS. The electroretinogram of the isolated rat retina. Vision Res 1972; 12:1183–1198.

7.Stone J, Maslim J, Valter-Kocsi K, et al. Mechanisms of photoreceptor death and survival in mammalian retina. Prog Retin Eye Res 1999; 18:689–735.

8.Kuwabara T, Cogan D. Retinal glycogen. Arch Ophthalmol 1961; 66:680–688.

9.Weiss H. The carbohydrate reserve in the vitreous body and retina of the rabbit’s eye during and after pressure ischaemia and insulin hypoglycaemia. Ophthal Res 1972; 3:360–371.

10.Johnson NF. Retinal glycogen content during ischaemia. Albrecht von Graefes Arch Klin Exp Ophthalmol 1977; 203:271–282.

11.Ames A III, Wright RL, Kowada M, et al. Cerebral ischemia. II. The no-reflow phenomenon. Am J Pathol 1968; 52:437–453.

12.Fischer EG, Ames A III, Hedley-Whyte ET, et al. Reassessment of cerebral capillary changes in acute global ischemia and their relationship to the ‘‘no-reflow phenomenon’’. Stroke 1977; 8:36–39.

13.Buchi ER, Suivaizdis I, Fu J. Pressure-induced retinal ischemia in rats: an experimental model for quantitative study. Ophthalmologica 1991; 203:138–147.

14.Romano C, Price M, Bai HY, et al. Neuroprotectants in Honghua: glucose attenuates retinal ischemic damage. Invest Ophthalmol Vis Sci 1993; 34:72–80.

15.Romano C, Price MT, Almli T, et al. Excitotoxic neurodegeneration induced by deprivation of oxygen and glucose in isolated retina. Invest Ophthalmol Vis Sci 1998; 39:416–423.

16.Casson RJ, Wood JPM, Melena J, et al. The effect of ischemic preconditioning on light-induced photoreceptor injury. Invest Ophthalmol Vis Sci 2003; 44:1348–1354.

17.Casson RJ, Chidlow G, Wood JPM, et al. The effect of hyperglycemia on experimental retinal ischemia. Arch Ophthalmol 2004; 122:361–366.

18.Jennings RB, Sommers HM, Smyth GA, et al. Myocardial necrosis induced by temporary occlusion of a coronary artery in the dog. Arch Pathol 1960; 70:68–78.

19.Sims NR. Energy metabolism and selective neuronal vulnerability following global cerebral ischemia. Neurochem Res 1992; 17:923–931.

20.Leker RR, Shohami E. Cerebral ischemia and trauma. Different etiologies yet similar mechanisms: neuroprotective opportunities. Brain Res Rev 2002; 39:55–73.

21.Davalos A, Castillo J, Serena J, et al. Duration of glutamate release after acute ischemic stroke. Stroke 1997; 28:708–710.

22.Petty MA, Wettstein JG. White matter ischaemia. Brain Res Rev 1999; 31:58–64.

23.Pellegrini-Giampietro DE, Cherici G, Alesiani M, et al. Excitatory amino acid release and free radical formation may cooperate in the genesis of ischemia-induced neuronal damage. J Neurosci 1990; 10:1035–1041.

24.Bonne C, Muller A, Villain M. Free radicals in retinal ischemia. Gen Pharmacol 1998; 30:275–280.

25.Osborne NN, Casson RJ, Wood JP, et al. Retinal ischemia: mechanisms of damage and potential therapeutic strategies. Prog Retin Eye Res 2004; 23:91–147.

26.Gilgun-Sherki Y, Rosenbaum Z, Melamed E, et al. Antioxidant therapy in acute central nervous system injury: current state. Pharmacol Rev 2002; 54:271–284.

27.Doly M, Braquet P, Bonhomme B, et al. Effects of lipid peroxidation on the isolated rat retina. Ophthalmic Res 1984; 16:292–296.

28.Szabo ME, Droy-Lefaix MT, Doly M, et al. Ischemia and reperfusion-induced histologic changes in the rat retina. Demonstration of a free radical-mediated mechanism. Invest Ophthalmol Vis Sci 1991; 32:1471–1478.

29.Szabo ME, Droy-Lefaix MT, Doly M, et al. Ischaemiaand reperfusion-induced Naþ, Kþ, Ca2þ and Mg2þ shifts in rat retina: effects of two free radical scavengers,

SOD and EGB 761. Exp Eye Res 1992; 55:39–45.

30.Muller A, Pietri S, Villain M, et al. Free radicals in rabbit retina under ocular hyperpressure and functional consequences. Exp Eye Res 1997; 64:637–643.

31.Rios L, Cluzel J, Vennat JC, et al. Comparison of intraocular treatment of DMTU and SOD following retinal ischemia in rats. J Ocul Pharmacol Ther 1999; 15: 547–556.

32.Banin E, Berenshtein E, Kitrossky N, et al. Gallium-desferrioxamine protects the cat retina against injury after ischemia and reperfusion. Free Radic Biol Med 2000; 28:315–323.

33.Chidlow G, Schmidt KG, Wood JP et al. Alpha-lipoic acid protects the retina against ischemia-reperfusion. Neuropharmacology 2002; 43:1015–1025.

34.Block F, Schwarz M. Effects of antioxidants on ischemic retinal dysfunction. Exp Eye Res 1997; 64:559–564.

35.Celebi S, Dilsiz N, Yilmaz T, et al. Effects of melatonin, vitamin E and octreotide on lipid peroxidation during ischemia-reperfusion in the guinea pig retina. Eur J Ophthalmol 2002; 12:77–83.

36.Shibuki H, Katai N, Yodoi J, et al. Lipid peroxidation and peroxynitrite in retinal ischemia-reperfusion injury. Invest Ophthalmol Vis Sci 2000; 41:3607–3614.

37.Kuriyama H, Waki M, Nakagawa M, et al. Involvement of oxygen free radicals in experimental retinal ischemia and the selective vulnerability of retinal damage. Ophthalmic Res 2001; 33:196–202.

38.Gupta LY, Marmor MF. Mannitol, dextromethorphan, and catalase minimize ischemic damage to retinal pigment epithelium and retina. Arch Ophthalmol 1993; 111:384–388.

39.Ugarte M, Osborne NN. Zinc in the retina. Prog Neurobiol 2001; 64:219–249.

40.Szabo ME, Droy-Lefaix MT, Doly M. Direct measurement of free radicals in ischemic/reperfused diabetic rat retina. Clin Neurosci 1997; 4:240–245.

41.Roth S, Park SS, Sikorski CW, et al. Concentrations of adenosine and its metabolites in the rat retina/choroid during reperfusion after ischemia. Curr Eye Res 1997; 16:875–885.

42.Zhang H, Agardh CD, Agardh E. Increased catalase levels and hypoxanthineenhanced nitro-blue tetrazolium staining in rat retina after ischemia followed by recirculation. Curr Eye Res 1995; 14:47–54.

43.Peachey NS, Green DJ, Ripps H. Ocular ischemia and the effects of allopurinol on functional recovery in the retina of the arterially perfused cat eye. Invest Ophthalmol Vis Sci 1993; 34:58–65.

44.Faberowski N, Stefansson E, Davidson RC. Local hypothermia protects the retina from ischemia. A quantitative study in the rat. Invest Ophthalmol Vis Sci 1989; 30:2309–2313.

45.Szabo ME, Droy-Lefaix MT, Doly M, et al. Modification of ischemia/reperfusioninduced ion shifts (Naþ, Kþ, Ca2þ and Mg2þ) by free radical scavengers in the rat

retina. Ophthalmic Res 1993; 25:1–9.

46.Matsuo Y, Kihara T, Ikeda M, et al. Role of neutrophils in radical production during ischemia and reperfusion of the rat brain: effect of neutrophil depletion on extracellular ascorbyl radical formation. J Cereb Blood Flow Metab 1995; 15:941–947.

47.Lipton P. Ischemic cell death in brain neurons. Physiol Rev 1999; 79:1431–1568.

48.Tsujikawa A, Ogura Y, Hiroshiba N, et al. Retinal ischemia-reperfusion injury attenuated by blocking of adhesion molecules of vascular endothelium. Invest Ophthalmol Vis Sci 1999; 40:1183–1190.

49.Goldstein IM, Ostwald P, Roth S. Nitric oxide: a review of its role in retinal function and disease. Vision Res 1996; 36:2979–2994.

50.Shin DH, Lee HY, Kim HJ, et al. In situ localization of neuronal nitric oxide synthase (nNOS) mRNA in the rat retina. Neurosci Lett 1999; 270:53–55.

51.Cheon EW, Park CH, Kang SS, et al. Change in endothelial nitric oxide synthase in the rat retina following transient ischemia. Neuroreport 2003; 14:329–333.

52.Lopez-Costa JJ, Goldstein J, Saavedra JP. Neuronal and macrophagic nitric oxide synthase isoforms distribution in normal rat retina. Neurosci Lett 1997; 232:155–158.

53.Iadecola C. Bright and dark sides of nitric oxide in ischemic brain injury. Trends Neurosci 1997; 20:132–139.

54.Cheon EW, Park CH, Kang SS, et al. Nitric oxide synthase expression in the transient ischemic rat retina: neuroprotection of betaxolol. Neurosci Lett 2002; 330:265–269.

55.Kobayashi M, Kuroiwa T, Shimokawa R, et al. Nitric oxide synthase expression in ischemic rat retinas. Jpn J Ophthalmol 2000; 44:235–244.

56.Hangai M, Yoshimura N, Hiroi K, et al. Inducible nitric oxide synthase in retinal ischemia-reperfusion injury. Exp Eye Res 1996; 63:501–509.

57.Hangai M, Miyamoto K, Hiroi K, et al. Roles of constitutive nitric oxide synthase in postischemic rat retina. Invest Ophthalmol Vis Sci 1999; 40:450–458.

58.Neufeld AH, Kawai S, Das S, et al. Loss of retinal ganglion cells following retinal ischemia: the role of inducible nitric oxide synthase. Exp Eye Res 2002; 75:521–528.

59.Sennlaub F, Courtois Y, Goureau O. Inducible nitric oxide synthase mediates the change from retinal to vitreal neovascularization in ischemic retinopathy. J Clin Invest 2001; 107:717–725.

60.Abu El-Asrar AM, Desmet S, Meersschaert A, et al. Expression of the inducible isoform of nitric oxide synthase in the retinas of human subjects with diabetes mellitus. Am J Ophthalmol 2001; 132:551–556.

61.Veriac S, Tissie G, Bonne C. Oxygen free radicals adversely affect the regulation of vascular tone by nitric oxide in the rabbit retina under high intraocular pressure. Exp Eye Res 1993; 56:85–88.

62.Lo EH, Hara H, Rogowska J, et al. Temporal correlation mapping analysis of the hemodynamic penumbra in mutant mice deficient in endothelial nitric oxide synthase gene expression. Stroke 1996; 27:1381–1385.

63.Maynard KI, Chen D, Arango PM, et al. Nitric oxide produced during ischemia improves functional recovery in the rabbit retina. Neuroreport 1996; 8:81–85.

64.Vorwerk CK, Hyman BT, Miller JW, et al. The role of neuronal and endothelial nitric oxide synthase in retinal excitotoxicity. Invest Ophthalmol Vis Sci 1997; 38:2038–2044.

65.Adachi K, Fujita Y, Morizane C, et al. Inhibition of NMDA receptors and nitric oxide synthase reduces ischemic injury of the retina. Eur J Pharmacol 1998; 350:53–57.

66.Lam TT, Tso MO. Nitric oxide synthase (NOS) inhibitors ameliorate retinal damage induced by ischemia in rats. Res Commun Mol Pathol Pharmacol 1996; 92:329–340.

67.Neufeld AH, Sawada A, Becker B. Inhibition of nitric-oxide synthase 2 by aminoguanidine provides neuroprotection of retinal ganglion cells in a rat model of chronic glaucoma. Proc Natl Acad Sci U S A 1999; 96:9944–9948.

68.Osborne NN, Melena J, Chidlow G, et al. A hypothesis to explain ganglion cell death caused by vascular insults at the optic nerve head: possible implication for the treatment of glaucoma. Br J Ophthalmol 2001; 85:1252–1259.

69.Stapf C, Mohr JP. Ischemic stroke therapy. Annu Rev Med 2002; 53:453–475.

70.Kristian T, Siesjo BK. Calcium in ischemic cell death. Stroke 1998; 29:705–718.

71.Nishizawa Y. Glutamate release and neuronal damage in ischemia. Life Sci 2001; 69:369–381.

albinism (OCA). Tyrosinase is the rate limiting enzyme of melanin biosynthesis and catalyses the first two steps of melanin synthesis: the hydroxylation of L-tyrosine to L-3,4-dihydroxyphenylalanine (L-DOPA), and the oxidation of L-DOPA to DOPAquinone.4 Furthermore, a third enzymatic reaction was assigned to tyrosinase: the oxidation of 5,6-dihydroxyindole to 5,6-dihydroxyquinone.5 It has been postulated that melanogenesis in the RPE is restricted to prenatal periods, since tyrosinase, the key enzyme in melanin biosynthesis, was detected in early stage human embryos only and was absent long before gestation ends.6–9 This is the actual state of knowledge up to the present day.

However, premelanosomes and early-stage-melanosomes have been found in adult RPE10–13 which led to this hypothesis being questioned. Tyrosinase

activity has also been demonstrated in adult cultured bovine,14,15 porcine,11 mouse,16 rabbit,17 rat18 and human RPE cells.19 Additionally, tyrosinase activity was found in adult bovine RPE.20

Tyrosinase promoter activity was significantly up-regulated in in cultured human RPE cells treated with bFGF, PEDF, verapamil, CT and tyrosine compared with control cells. In conclusion, the tyrosinase gene is not only expressed but can be regulated in response to different chemicals in cultured human RPE cells. However, in that study tyrosinase enzymic activity was not found.21 Recent studies show that phagocytosis of ROS induces gene expression in RPE cells.22

Our study was performed to examine the presence of tyrosinase protein as well as its enzymatic activity in adult mammalian RPE. Phagocytosis of shed photoreceptor outer segment distal discs is one of the most important functions of the RPE. Cellular functions of RPE cells such as disc shedding and subsequent phagocytosis are controlled by circadian rhythm.23 Therefore we investigated whether phagocytosis of ROS can increase tyrosinase expression in vitro.

MATERIALS AND METHODS

Organ Culture of Bovine RPE-Choroid Complexes

Pigmented RPE-choroid complexes were isolated according to Schraermeyer and Stieve.15 In brief, cattle eyes from two year-old animals were transported on ice from a slaughter house to the laboratory. The eyes were washed once with Hanks’ balanced salt solution (HBSS) containing penicillin (100 U/ml) and streptomycin (0.1 mg/ml). All tissue culture media or solutions were purchased from Sigma (Deisenhofen, Germany). The anterior half of the eye was removed and discarded. The vitreous was removed, and the retina was gently floated off the RPE by pipetting HBSS into the subretinal space. The retinae were removed after cutting the optic nerve and used for isolation of the rod outer segments as described below. Specimens (1 mm3) of the RPE-choroid complex were kept in 24-well tissue culture plates (Cluster,24 Costar, Cambridge, UK). The complete tissue culture medium contained Dulbecco’s Modified Eagle Medium with 4500 mg/l glucose supplemented with 15% fetal calf serum, 25 mM HEPES buffer, 100 U/ml penicillin, 0.1 mg/ml streptomycin, 3.7 g/l

sodium bicarbonate and 4 mM L-glutamine. The cultures were gassed with 5% CO2 and the medium was changed completely once a week. In a separate set of experiments for immunodetection of tyrosinase RPE cells were cultured as monolayers after trypsinisation in the same manner.

Isolation of Rod Outer Segments

The method was performed as described by Schraermeyer and Stieve.15 In detail, isolated retinae were agitated for 2 min in KCl buffer (0.3 M KCl, 10 mM HEPES, 0.5 mM CaCl2, 1 mM MgCl2 and 48% sucrose) at pH 7.0 and then centrifuged at 2000 rpm in a tabletop centrifuge (type UJ 1, Christ, Germany) for 5 min. The supernatant was filtered through a tube gauze fingerling and then diluted with KCl-buffer (1:1) and centrifuged at 2500 rpm for 10 min. The pellet, containing the rod outer segments, was washed and centrifuged in complete culture medium before feeding to the RPE.

Assay for Phagocytosis

RPE cell monolayers or RPE-choroid complexes, adapted to the culture medium conditions for 2 weeks, were exposed to rod outer segments (rods from one eye/ 4 wells) that had been isolated and marked as described above. After 4 hrs the non-phagocytosed rod fragments were washed out and replaced by fresh medium. The medium was changed daily during the experiments.

Electron-Microscopical Localization of Tyrosinase

The enzyme tyrosinase was localized by electron-microscopical histochemistry.24,25 In brief, for localisation of tyrosinase, RPE cells that had been

exposed to fragments of rod outer segments as described above were fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 6.8) for 1h. RPE cells without feeding were treated in the same manner. Specimens were washed twice in sodium cacodylate buffer and kept at 48C overnight in this buffer, containing 5 mM L-dihydroxphenylalanine (L-DOPA) or, as a control, 5 mM D-DOPA (Sigma, Deisenhofen, Germany). Thereafter, these solutions were renewed, and the tissue pieces were incubated for a further 5hrs at 378C. The reacted RPE cells were washed in sodium cacodylate buffer and immersed for 1 h at room temperature in the same buffer containing a mixture of osmium tetroxide (1%) and potassium ferrocyanide (1.5%). Finally, the tissue pieces were dehydrated and embedded in Spurr’s resin for routine electron microscopy. Ultrathin sections were stained with uranyl acetate and lead citrate and observed examined with a Zeiss EM 902A electron microscope.

Immunodetection of Tyrosinase

The mouse anti-tyrosinase monoclonal antibody 2G10 was purchased from Chemicon Int. Inc. (Temecula, CA, USA) and is described by the manufacturer

as an antibody to human tyrosinase.26 The monoclonal anti-tyrosine hydroxylase antibody, clone TH-2 and IgG1 isotype, was from Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany and was used as a control. This antibody was raised against an epitope of rat tyrosine hydroxylase, also present in human tyrosine hydroxylase. The secondary purified goat Cy3-conjugated anti-mouse IgG antiserum was obtained from Rockland Immunochemicals Inc., Gilbertsville, PA, USA. This antiserum had been raised against mouse IgG. Antibodies (3-8 mg/ml) were used on paraformaldehyde fixed RPE monolayers as described elsewhere.27 First antibodies were diluted at 1:200 and incubated in buffer containing 0.01 M sodium phosphate and 0.25 M NaCl for 30 minutes at room temperature. Second antibodies were used diluted 1:1000 and incubated in the same buffer for 1 hour. Cells were photographed under an Axioplan 2 mikroskop (Zeiss, Oberkochem, Germany). Image processing was performed with a Orca camera (Hamamatsu Photonics, Germany) and Openlab Software (Improvision, Tu¨bingen, Germany.

Tyrosine Hydroxylase Activity of Tyrosinase

Samples were homogenized (using a Potter-Elvehjem homogenizer at 10 strokes up-and-down at 1,200 rpm) in 100 mM potassium phosphate buffer, pH 7.4. Then 25 ml of the homogenate were mixed with 25 ml assay buffer containing 100 mM potassium phosphate, pH 7.4, 18.5 MBq L-[3,5-3H]tyrosine (specific radioactivity 1.83 GBq/mmol; Moravek Biochemicals, Brea, CA) and 1 mM L-DOPA (Sigma-Aldrich Chemie, Taufkirchen, Germany) and incubated for 24 hours at 378C. The reaction was stopped by adding 100 mg Celite 545 (Merck KGaA, Darmstadt, Germany) and 100 mg activated charcoal suspended in 1 ml 0.1 N HCl solution. After 1 hour shaking at room temperature, the samples were centrifuged (15,000 g, 5 min), 500 ml of the resulting supernatant were mixed with 10 ml of Ultima GoldTM scintillation cocktail and radioactivity was determined by a Packard TRI-CARB 2900TR liquid scintillation counter.

Protein concentration of the samples was measured using the method of Lowry et al.28 applying bovine serum albumin as standard.

Statistical Analysis

Data are given as mean standard deviation. Statistical evaluation was based on Student’s t-test for two populations. A double-sided p-value of less than 0.05 was considered statistically significant.

RESULTS

Electron Microscopic Localization of Tyrosinase by DOPA Histochemistry

To optimize the in vitro conditions, we used organ cultures consisting of melanotic RPE-choroid complexes from bovine eyes. The RPE layer remains on Bruch’s membrane and under these physiological conditions the apical microvilli

are preserved in vitro. ROS isolated from bovine eyes were phagocytized and immediately after exposure to ROS many phagosomes containing ROS were present within the RPE cells (Figure 1A).

Tyrosinase activity was investigated by electron microscopic (EM) histochemistry. DOPA oxidase activity of tyrosinase was determined.

Five hours after feeding with ROS, tyrosinase activity was demonstrated in RPE by DOPA-oxidase assay. DOPA-positive electron-dense vesicles, indicating the presence of tyrosinase, were present throughout the cytoplasm as shown in electron micrographs (Figure 1B, C). Without feeding DOPA positive vesicles were not found in RPE cells (Figure 1D). DOPA positive Golgi bodies were regularly seen 5 hours (Figure 1E) after feeding. Occasionally DOPA-positive vesicles were observed in close vicinity to melanosomes (Figure 1F).

Twenty-four hours after feeding with ROS, DOPA-positive vesicles and Golgi bodies disappeared from the cytoplasm (Figure 1G). Lipid like droplets, 1-3 mm in diameter, were present in many RPE cells (Figure 1G, H). The lipid droplets were surrounded by membrane stacks that contained DOPA-positive material (Figure 1G, H). These organelles corresponded exactly to those found after immunocytochemistry with anti-tyrosinase antibodies as shown in Fig. 2D. DOPA oxidase activity was not detected in controls incubated with 5 mM D DOPA (data not shown).

Antibody Staining of Tyrosinase in Bovine RPE Cells

In a monolayer of cultured bovine RPE cells, control immunocytochemical experiments performed using antibodies against L-tyrosine 3-hydroxylase (EC 1.14.16.2) were negative (Figure 2A). Without feeding with ROS, using antityrosinase antibodies, no specific staining different from background levels was detected (Figure 2B). However, five hours after feeding slight specific staining was found (Figure 2C) corresponding to the electron microscopic DOPA staining shown in Fig. 1B, C. Twenty-four hours after feeding with ROS by immunocytochemistry with a monoclonal antibody, tyrosinase protein was predominately detected in ellipsoid organelles (Figure 3D).

Tyrosine Hydroxylase Activity of Tyrosinase

Using a highly specific radioactive assay, L-tyrosine 3-hydroxylase activity was determined in the same adult bovine melanotic RPE-choroid complexes after feeding with ROS (Figure 3A). By measuring the release of tritium from tritiated L-tyrosine in homogenized samples, we demonstrated that feeding with ROS increases tyrosinase enzyme activity in comparison to non-challenged RPE (15.8 2.5 to 725.0 70.9 fmol/mg/h, Figure 3) by a factor of 40. The difference between the ROS-fed and non-fed RPE was statistically significant ( p ¼ 0.003;

Figure 1 A. Organ culture consisting of melanotic RPE-choroid complex from bovine eyes. Electron micrograph of a cross section immediately after feeding with ROS for 4 hrs. The RPE cell with many phagosomes (P) containing ROS. Under these culture conditions the apical microvilli (arrow) of an RPE cell are preserved in vitro since the RPE layer is cultivated on Bruch’s membrane (B). This natural environment may improve the functions of RPE in vitro. B, C. Phagocytosis of ROS induces tyrosinase activity in RPE. DOPAoxidase activity of tyrosinase demonstrated by EM histochemistry in RPE cells 5 hours after feeding with ROS. Many DOPA positive vesicles (arrows) are scattered throughout the cytoplasm (M = mitochondrion, N = nucleus). D. Electron micrograph taken from control cultures that were not exposed to ROS. DOPA-oxidase activity was not present in these cultures. E, F. DOPA-positive Golgi (G) bodies are seen 5 hrs after exposure to ROS. Golgi derived vesicles are close to melanosomes and suggest classical biosynthesis of this enzyme 5 hrs after feeding with ROS (arrow in 1F). G. Twenty-four hours after feeding tyrosinase activity surrounds large lysosomes (arrows) filled with lipid-like material (L) but is no longer seen in cytoplasmic vesicles. 1H. Tyrosinase is localized within membrane stacks (arrow) surrounding these lipid droplets (L). These droplets correspond to organelles staining positive for anti-tyrosinase antibodies in Fig. 2D.

Figure 2 (See color insert.) For tyrosinase immunocytochemistry, the RPE monolayer was prepared and exposed to ROS for 4 hrs. The expression of tyrosinase was investigated before feeding with ROS, as well as 5 and 24 hours afterwards. 2A Five hours after feeding with ROS, no staining was visible with anti-tyrosine hydroxylase antibodies. 2B Without feeding with ROS no staining was found with anti-tyrosinase antibodies. 2C Five hours after feeding with ROS faint staining was observed with anti-tyrosinase antibodies corresponding to DOPA positive vesicles in Fig. 1B, 2D. Twenty-four hours after feeding with ROS intense staining of lysosome-like organelles (arrows) was found with antityrosinase antibodies. These organelles correspond to those shown in Fig. 1G, 1B.

3 measurements), which were not significantly increased after exposure to ROS (19.9 fmol/mg/h, n ¼ 1, mean of 3 measurements.

DISCUSSION

These results clearly demonstrate that phagocytosis of ROS can induce tyrosinase expression in RPE. Here, we provide EM histological and biochemical evidence for the presence and induction of tyrosinase activity after phagocytosis of ROS in adult mammalian RPE.

Figure 3 Phagocytosis of ROS increased the activity of tyrosinase approximately 40-fold. Organ cultures consisting of melanotic bovine RPE-choroid complexes were exposed to ROS for 4 hrs. Twenty-four hours after feeding with ROS the cultures were homogenized and used for determination of tyrosinase activity. Without feeding with ROS tyrosinase activity was almost not detectable (*, p – 0.003). These results clearly demonstrate that tyrosinase activity was induced only after feeding with ROS.

The role of phagocytosis-induced tyrosinase expression is not yet clear.

Several studies suggest that melanin biosynthesis can take place within lysosomes15,16,25,29,30 because melanosomes belong to the same lineage of organ-

elles. Moreover, melanosomes contain most lysosomal enzymes31 and it was stated that the melanosome is a specialized lysosome.32

It is not surprising that synthesis of tyrosinase was induced by phagocytosis, because ROS uptake is associated with regulation of many genes in the RPE.22 Furthermore, lipids, particularly sphingolipids, induce melanogenesis by increasing the expression of tyrosinase and its related proteins in vitro.33 Therefore, we intend to investigate in future studies whether ROS lipids induce expression of tyrosinase and melanogenesis. Whereas the electron microscopic findings of tyrosinase activation early after feeding with ROS correspond to the classical scheme of melanogenesis with Golgi derived DOPA positive vesicles scattered throughout the cytoplasm of a pigment cell, there are unexpected findings after 24 hours: At this time point DOPA positive vesicles were not present, but tyrosinase was seen in membrane stacks surrounding lipid droplets (Figure 1G, F). As this was seen in all RPE cultures of this study with DOPA histochemistry and immunocytochemistry using anti-tyrosinase antibodies it is not an artifact but a real finding. The meaning of this observation remains unknown but may be in part explained by the following hypothesis: culture conditions decrease the ability of RPE cells to degrade the high amounts of ROS (Figure 1A) within 24 hours and lipid-like material accumulates in large lysosomes. This may be partially because in vivo entire rods are taken up, instead of

shed tips. In vivo tyrosinase containing vesicles may fuse with phagosomes. Inside phagosomes it may react with superoxide as an radical trapping enzyme,34 may simply be targeted for degradation as an unfunctional enzyme or may have other yet unknown functions. In this study the content of lysosomes is so densely packed that the uptake of protein seems to be blocked, and the vesicles may fuse forming membrane stacks around the lysosomes. This would explain why the margins of the lipid vacuoles stain positive for tyrosinase whereas the inner layers do not (Figure 1G, F). This hypothesis is in agreement with findings of this and earlier studies, in which tyrosinase was detected in incompletely degraded phagosomes (not shown).15 Moreover, there is a connection between phagosomal degradation pathways and melanosomes in RPE cells, as it was found that material from phagosomes is transported to the melanosomes or melanogenesis taking place inside phagosomes.35

As phagosomes are related to endosomes, the hypothesis, earlier findings as well as the observations of this study fit into the general scheme of tyrosinase trafficking in melanocytes. In melanocytes tyrosinase is transported from transGolgi network to early endosomes and reaches via late endosomes lysosomes and premelanosomes.36

Tyrosinase has many variable functions, the mechanisms of which are still unknown. It is directly involved in light adaptation in zebra fish.37 Moreover,

tyrosinase plays a role in glaucoma of zebra fish38 and mouse models.39 Tyrosinase is also present in most neurons of the murine brain40,41 without being

melanogenic active and it promotes survival of catecholaminergic neurons.42

Tyrosinase or its metabolites are responsible for development of a fully functional macula39,43 and the normal crossing of optic nerve fibers.44 In addition

immature children are more protected from retinopathy of prematurity (ROP) if they have a darker fundus resulting from melanin pigmentation.45,46 It is not known whether tyrosinase or its metabolites protect against oxidative stress in this disease. Moreover, black people are more protected from ARMD and a

lower amount of Drusen, crosslinked proteins in Bruch’s membrane and lipofuscin in the RPE than white Caucasians.47–49 In ARMD it is not known

whether this is caused by tyrosinase itself or its metabolites. The functional role of tyrosinase expression induced by phagocytosis is not known. The phagocytosis of ROS is associated with generation of reactive oxygen species such as superoxide anion radicals (O2 ), hydrogen peroxide (H2O2) and hydroxyl radicals (OH ).50 Hydrogen peroxide was shown to increase tyrosinase activity in cultured human melanoma cells,51 to be a competitive inhibitor of Tyrosinase,52 and to induce tyrosinase.51,53 Furthermore, superoxide anions increase human tyrosinase activity.52

Tyrosinase may protect against oxidative stress by several mechanisms: firstly, by oxidation of tetrahydroisoquinones, when tyrosinase may promote their sinking into insoluble inert polymers; secondly, by utilizing superoxide anion as substrate, when it could act as a free radical trapping enzyme.34

CONCLUSION

Although the function of tyrosinase, or whether it is melanogenetically active or not in RPE cells remains unknown, here it is shown for the first time with three independent methods that tyrosinase can be induced by phagocytosis in RPE cells of postnatal vertebrates.

REFERENCES

1.Marks MS, Seabra MC. The melanosome: membrane dynamics in black and white. Nat Rev Mol Cell Biol 2001; 2:738–748.

2.Seiji M, Fitzpatrick TB, Simpson RT, et al. Chemical composition and terminology of specialized organelles (melanosomes and melanin granules) in mammalian melanocytes. Nature 1963; 197:1082–1084.

3.Boulton M, Dayhaw-Barker P. The role of the retinal pigment epithelium: topographical variation and ageing changes. Eye 2001; 15:384–389.

4.Lerch K. Monophenol monooxygenase from Neurospora crassa. Methods Enzymol 1987; 142:165–169.

5.Korner A, Pawelek J. Mammalian tyrosinase catalyzes three reactions in the biosynthesis of melanin. Science 1982; 217:1163–1165.

6.Miyamoto M, Fitzpatrick TB. On the nature of pigment in retinal pigment epithelium. Science 1957; 126:449–450.

7.Carr RE, Siegel IM. The retinal pigment epithelium in ocular albinism. In: Zinn KM, Marmor MF, eds. The Retinal Pigment Epithelium. Cambridge, Massachusetts, London: Harvard University Press, 1979:413–423.

8.Sarna T. Properties and function of the ocular melanin—a photobiophysical view. J Photochem Photobiol B 1992; 12:215–258.

9.Smith-Thomas L, Richardson P, Thody AJ, et al. Human ocular melanocytes and retinal pigment epithelial cells differ in their melanogenic properties in vivo and in vitro. Curr Eye Res 1996; 15:1079–1091.

10.Young RW. The daily rhythm of shedding and degradation of rod and cone outer segment membranes in the chick retina. Invest Ophthalmol Vis Sci 1978; 17: 105–116.

11.Dorey CK, Torres X, Swart T. Evidence of melanogenesis in porcine retinal pigment epithelial cells in vitro. Exp Eye Res 1990; 50:1–10.

12.Schraermeyer U. Does melanin turnover occur in the eyes of adult vertebrates? Pigment Cell Res 1993; 6:193–204.

13.Schraermeyer U, Heimann K. Current understanding on the role of retinal pigment epithelium and its pigmentation. Pigment Cell Res 1999; 12:219–236.

14.Basu PK, Sarkar P, Menon I, et al. Bovine retinal pigment epithelial cells cultured in vitro: growth characteristics, morphology, chromosomes, phagocytosis ability, tyrosinase activity and effect of freezing. Exp Eye Res 1983; 36:671–683.

15.Schraermeyer U, Stieve H. A newly discovered pathway of melanin formation in cultured retinal pigment epithelium of cattle. Cell Tissue Res 1994; 276:273–279.

16.Novikoff AB, Leuenberger PM, Novikoff PM, et al. Retinal pigment epithelium. Interrelations of endoplasmic reticulum and melanolysosomes in the black mouse and its beige mutant. Lab Invest 1979; 40:155–165.

17.Varela JM, Stempels NA, Vanden Berghe DA, et al. Isoenzymic patterns of tyrosinase in the rabbit choroid and retina/retinal pigment epithelium. Exp Eye Res 1995; 60:621–629.

18.Weisse I. Changes in the aging rat retina. Ophthalmic Res 1995; 27(suppl 1): 154–163.

19.Aronson JF. Human retinal pigment cell culture. In Vitro 1983; 19:642–650.

20.Dryja TP, O’Neil-Dryja M, Pawelek JM, et al. Demonstration of tyrosinase in the adult bovine uveal tract and retinal pigment epithelium. Invest Ophthalmol Vis Sci 1978; 17:511–514.

21.Abul-Hassan K, Walmsley R, Tombran-Tink J, et al. Regulation of tyrosinase expression and activity in cultured human retinal pigment epithelial cells. Pigment Cell Res 2000; 13:436–441.

22.Chowers I, Kim Y, Farkas RH, et al. Changes in retinal pigment epithelial gene expression induced by rod outer segment uptake. Invest Ophthalmol Vis Sci 2004; 45:2098–2106.

23.LaVail MM. Circadian nature of rod outer segment disc shedding in the rat. Invest Ophthalmol Vis Sci 1980; 19:407–411.

24.Schraermeyer U. Localization of peroxidase activity in the retina and the retinal pigment epithelium of the Syrian golden hamster (Mesocricetus auratus). Comp Biochem Physiol B 1992; 103:139–145.

25.Schraermeyer U. Evidence for melanogenesis in the retinal pigment epithelium of adult cattle and golden hamster. Comp Biochem Physiol B 1992; 103:435–442.

26.Cuomo M, Nicotra MR, Apollonj C, et al. Production and characterization of the murine monoclonal antibody 2G10 to a human T4-tyrosinase epitope. J Invest Dermatol 1991; 96:446–451.

27.Schraermeyer U, Enzmann V, Kohen L, et al. Porcine iris pigment epithelial cells can take up retinal outer segments. Exp Eye Res 1997; 65:277–287.

28.Lowry OH, Rosebrough NJ, Farr AL, et al. Protein measurement with the Folin phenol reagent. J Biol Chem 1951; 193:265–275.

29.Nakagawa H, Rhodes AR, Momtaz TK, et al. Morphologic alterations of epidermal melanocytes and melanosomes in PUVA lentigines: a comparative ultrastructural investigation of lentigines induced by PUVA and sunlight. J Invest Dermatol 1984; 82:101–107.

30.Schraermeyer U. Transport of endocytosed material into melanin granules in cultured choroidal melanocytes of cattle—new insights into the relationship of melanosomes with lysosomes. Pigment Cell Res 1995; 8:209–214.

31.Diment S, Eidelman M, Rodriguez GM, et al. Lysosomal hydrolases are present in melanosomes and are elevated in melanizing cells. J Biol Chem 1995; 270: 4213–4215.

32.Orlow SJ. Melanosomes are specialized members of the lysosomal lineage of organelles. J Invest Dermatol 1995; 105:3–7.

33.Mallick S, Singh SK, Sarkar C, et al. Human placental lipid induces melanogenesis by increasing the expression of tyrosinase and its related proteins in vitro. Pigment Cell Res 2005; 18:25–33.

34.Valverde P, Manning P, McNeil CJ, et al. Activation of tyrosinase reduces the cytotoxic effects of the superoxide anion in B16 mouse melanoma cells. Pigment Cell Res 1996; 9:77–84.

35.Schraermeyer U, Peters S, Thumann G, et al. Melanin granules of retinal pigment epithelium are connected with the lysosomal degradation pathway. Exp Eye Res 1999; 68:237–245.

36.Hearing VJ. Biogenesis of pigment granules: a sensitive way to regulate melanocyte function. J Dermatol Sci 2005; 37:3–14.

37.Page-McCaw PS, Chung SC, Muto A, et al. Retinal network adaptation to bright light requires tyrosinase. Nat Neurosci 2004; 7:1329–1336.

38.Link BA, Gray MP, Smith RS, et al. Intraocular pressure in zebrafish: comparison of inbred strains and identification of a reduced melanin mutant with raised IOP. Invest Ophthalmol Vis Sci 2004; 45:4415–4422.

39.Libby RT, Smith RS, Savinova OV, et al. Modification of ocular defects in mouse developmental glaucoma models by tyrosinase. Science 2003; 299:1578–1581.

40.Tief K, Hahne M, Schmidt A, et al. Tyrosinase, the key enzyme in melanin synthesis, is expressed in murine brain. Eur J Biochem 1996; 241:12–16.

41.Tief K, Schmidt A, Beermann F. New evidence for presence of tyrosinase in substantia nigra, forebrain and midbrain. Mol Brain Res 1998; 53:307–310.

42.Higashi Y, Asanuma M, Miyazaki I, et al. Inhibition of tyrosinase reduces cell viability in catecholaminergic neuronal cells. J Neurochem 2000; 75:1771–1774.

43.Summers CG. Vision in albinism. Trans Am Ophthalmol Soc 1996; 94:1095–1155.

44.Rachel RA, Mason CA, Beermann F. Influence of tyrosinase levels on pigment accumulation in the retinal pigment epithelium and on the uncrossed retinal projection. Pigment Cell Res 2002; 15:273–281.

45.Tadesse M, Dhanireddy R, Mittal M, et al. Race, Candida sepsis, and retinopathy of prematurity. Biol Neonate 2002; 81:86–90.

46.Saunders RA, Donahue ML, Christmann LM, et al. Racial variation in retinopathy of prematurity. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1997; 115:604–608.

47.Gregor Z, Joffe L. Senile macular changes in the black African. Br J Ophthalmol 1978; 62:547–550.

48.Friedman DS, Katz J, Bressler NM, et al. Racial differences in the prevalence of age-related macular degeneration: the Baltimore Eye Survey. Ophthalmology 1999; 106:1049–1055.

49.Hollyfield JG. IOVS 2004; 45:2289 (ARVO E-Abstract).

50.Dorey CK, Khouri GG, Syniuta LA, et al. Superoxide production by porcine retinal pigment epithelium in vitro. Invest Ophthalmol Vis Sci 1989; 30:1047–1054.

51.Karg E, Odh G, Wittbjer A, et al. Hydrogen peroxide as an inducer of elevated tyrosinase level in melanoma cells. J Invest Dermatol 1993; 100(2 suppl): 209S–213S.

52.Wood JM, Schallreuter KU. Studies on the reactions between human tyrosinase, superoxide anion, hydrogen peroxide and thiols. Biochim Biophys Acta 1991; 1074:378–385.

53.Gomez-Sarosi LA, Rieber MS, Rieber M. Hydrogen peroxide increases a 55-kDa tyrosinase concomitantly with induction of p53-dependent p21 waf1 expression and a greater Bax/Bcl-2 ratio in pigmented melanoma. Biochem Biophys Res Commun 2003; 312:355–359.