The Dark Side of Light: Rhodopsin and the Silent Death of Vision The Proctor Lecture

In the following a surprising effect of light upon the retina of albino and pigmented rats will be described. It was discovered when “normal” unanesthetized and unrestrained rats were maintained continuously for 24 hours in an environment illuminated by ordinary fluorescent light bulbs.—Noell et al., 1966 

Paracelsus, the medieval alchemist and physician, noted that it is the dose of a substance that renders it a poison or a remedy. This is also true of light, which obviously is essential for most life on earth, but in an overdose, can threaten the health of vision. 

Such an overdose was observed in the seminal studies of Noell et al. ¹ on damage caused by “normal” laboratory lights in the rat, a nocturnal species. The human visual system can adapt to an illuminance range of 10 to 11 log units. Why then should the human retina be endangered by light? However, epidemiologic studies and numerous case reports demonstrate that light can indeed injure the human retina (for a review, see Ref. ² ). 

Light damage and retinal degenerative diseases in humans and in animal models have an important feature in common: cell loss by apoptosis of photoreceptors and pigment epithelium (PE). ^{2 3 4 5} This crucial feature may render light damage a suitable model system to investigate cellular and molecular mechanisms of apoptosis with the aim of understanding and eventually treating or preventing, if only in part, retinal degenerative diseases in humans. 

There are several model systems suitable for the investigation of cellular and molecular mechanisms of light-induced damage to the retina:

•  
  Chronic, low-level exposure in the range of weeks to white fluorescent light in unrestrained animals.
•  
  Acute high-level exposure in the range of 30 minutes to several hours to white fluorescent light in unrestrained animals.
•  
  Spectral exposures in the range of 1 to 30 minutes to light of defined wavelengths, in unrestrained or restrained animals.

In our laboratory, we use acute, short-term exposure of unrestrained mice and rats to high levels of white fluorescent light and acute exposure of anesthetized mice and rats to defined wavelengths (blue and green). 

Is Rhodopsin an Essential Mediator of White Light–Induced Apoptosis?

In our model, we create a dose–response from threshold lesions to cell death. Threshold lesions are confined to photoreceptor outer segments and consist of dilations and disruptions of disc membranes. Disc dilations and disruptions increase from the tip toward the base, with increasing doses of light. Damage confined to outer segments is removed in the course of physiological renewal and can thus be considered reversible. ^{6 7} Shedding of injured membranes bearing oxidized proteins and lipids imposes a metabolic burden on the digestive apparatus including lysosomal function in PE cells. Undigestible remnants accumulate throughout life in PE-lipofuscin, such as the well-known retinoid fluorophore A2E ⁸ and perhaps its precursors A2PE ⁹ and A2-Rh, ¹⁰ which are initially formed in photoreceptors. Above threshold, light exposure induces apoptotic death of photoreceptors and, in rats, of PE cells. Dying cells then evoke conspicuous phagocytic activity by PE cells and mobile invading phagocytes (Fig. 1) . 

Even though there are several pigments in retina and PE that absorb visible light, rhodopsin and cone pigments have been proposed to represent the death-mediating chromophores since the work of Noell et al. ¹ in rats. By using a mouse model lacking measurable amounts of functional rhodopsin, the Rpe ^−/− mouse, ¹¹ we showed that no light-induced apoptotic cell loss and no PE lesions occur after exposure to damaging light. ¹² Therefore, rhodopsin constitutes the major mediator of light-induced apoptosis in the mouse and presumably other species. 

Earlier work in our laboratory has shown that rat retinas devoid of the polyunsaturated fatty acid docosahexaenoic acid (22:6 n-3; DHA) were protected against light-induced damage. ¹³ Notably, retinas of those rats contained higher rhodopsin levels than did those of untreated control rats, but the regeneration of rhodopsin after light exposure was slowed. ¹⁴ When levels of unsaturated fatty acids were increased by means of a fish oil diet, light-induced lesions were not increased, indicating that oxidative stress promoted by those higher levels of polyunsaturated fatty acids was not a crucial damage determinant. ¹⁵  

In the mouse, significant strain differences in susceptibility to damage have been described earlier ¹⁶ ; however, the underlying molecular mechanism was unclear. The PE protein RPE65 was discovered to bear two genetic variants coding for leucine or methionine at position 450. ¹⁷ Depending on the genetic variant, RPE65 levels were found to be low (methionine variant, M/M) or high (leucine variant, L/L), perhaps because of instability or reduced activity of the M/M-bearing protein. ¹⁸ Retinyl esters in the PE bind to soluble RPE65 protein which promotes the flow of retinyl esters to the isomerohydrolase for conversion into 11-cis retinal. ¹⁹ By correlating levels of RPE65 with rhodopsin regeneration rates and damage thresholds, a clear correlation appeared: mice with low levels of RPE65 and a slow regeneration rate showed high damage thresholds, whereas a fast regeneration conferred a low damage threshold. ¹⁸ Thus, any light damage study performed in mice should carefully investigate the genetic background with respect to the RPE65 variant. A recent study shows that this can also apply to animal models of inherited degenerations, especially those that are accelerated by light. ²⁰  

In contrast to the mouse, susceptibility to damage in rats is not primarily determined by the levels of RPE65, and RPE65 levels may not be a limiting factor for rhodopsin regeneration. ²¹ It is important to identify determinants of susceptibility in rats, apart from those that already have been described, such as the modification of phospholipid fatty acid composition, rhodopsin content, antioxidant levels, and time of day. ^{22 23 24 25}  

If a slow rhodopsin regeneration rate increases the threshold for light-induced apoptosis, could inhibition of rhodopsin regeneration confer protection against light- induced cell death in mice? The volatile anesthetic halothane is known to compete for the chromophore binding site in the opsin molecule. ²⁶ Thus, rhodopsin regeneration is blocked at the step of 11-cis retinal insertion into the opsin molecule. When dark-adapted mouse retinas are bleached and animals are placed into darkness for rhodopsin regeneration under halothane anesthesia, no regeneration occurs. Accordingly, when dark-adapted mice are exposed to damaging light under halothane, their retinas are completely protected against light-induced apoptosis. ²⁷ These data show that one strong bleaching of rhodopsin without its regeneration, as was applied under halothane anesthesia, is insufficient to induce damage by light. We made similar observations by exposing mice to green light of 550 nm. Rhodopsin, of course, was rapidly bleached, but no detectable lesions occurred (Remé CE, et al. IOVS 2003;43:ARVO E-Abstract 5132). We conclude that continuous absorptions by rhodopsin molecules or, in other words, availability of unbleached rhodopsin during light exposure are needed to inflict damage and cell death. 

Blocking the regeneration of 11–cis retinal by pharmacological means also confers protection. Application of 13-cis retinoic acid in vivo is likely to inhibit the PE-retinol dehydrogenase and RPE65, so that 11-cis retinal cannot be regenerated. ^{28 29}  

The Role of Phototransduction and Shut Off in White Light–Induced Apoptosis

Because rhodopsin is crucial in mediating light-induced apoptosis, the investigation of steps downstream of the initial bleaching event is essential. Activated rhodopsin (Rho* or metarhodopsin II [M II]) combines with the α subunit of the G-protein transducin with an exchange of GDP for GTP. This complex activates the αβ subunit of the phosphodiesterase so that on hydrolysis of cGMP, a fraction of the light-sensitive channel closes. The hyperpolarized photoreceptor cell then initiates a visual signal. Deactivation of Rho* (M II) or downregulation of its capacity to bind transducin occurs through phosphorylation by rhodopsin kinase (RhoK) and subsequent “capping” by arrestin (Arr). ³⁰  

To investigate the role of phototransduction in our light-damage system we used mice without functional transducin (Gnat 1α^−/−). In these mice, rhodopsin can absorb photons, but no binding to transducin-α occurs. Gnat 1α^−/− retinas were damaged indistinguishably from those of control mice; therefore, in damage induced by acute exposure to bright light, phototransduction is not involved. ³¹ By contrast, when low-level, continuous light is applied, Gnat 1α^−/− mice are protected, ³¹ thus essentially distinguishing mechanisms of short-term bright light and long-term, low-level light exposures. 

Studying the role of rhodopsin inactivation in light damage produced puzzling data. We used mice without RhoK and Arr (Rhok^−/−/Sag^−/−) and exposed them for different periods, to acute bright light. As little as 2 minutes of light exposure had devastating effects on the retina, with abundant cell death and increased activity of the proapoptotic transcription factor activator protein (AP)-1 (described later). ³¹ Rhodopsin regeneration was completed within 1 hour (Fig. 2) , which corresponds to the effect in wild-type mice of the L/L genotype for the RPE65 protein. ¹⁸  

Regulative Mechanisms in Bright Light–Induced Apoptosis

The immediate early gene c-fos responds rapidly to a variety of stress situations in many organs and tissues. ^{32 33} The c-Fos protein is a component of the transcription AP-1, which is involved in a multitude of cellular processes, such as proliferation, differentiation, cell death, and tumor cell growth. ³⁴ Furthermore, AP-1 is proapoptotic in many tissues. ³³ It is induced by a vast number of stimuli, including cytokines and UV irradiation, and couples such signals to cellular responses by regulating target genes. ³² In the retina, c-fos activity follows a diurnal rhythm and also responds to light pulses. ³⁵ Similarly, c-fos activity in the suprachiasmatic nuclei, the circadian master clock, responds to retinal light exposure. ³⁶ In the rd1 mouse, a model of human retinitis pigmentosa, abnormal c-fos expression occurs during the period of apoptotic cell death of photoreceptors. ³⁷  

Considering the known effects of light exposure on c-fos expression, could c-fos be a mediator of bright light–induced apoptosis in the retina ? We used c-fos ^−/− mice and exposed them and corresponding wild-type control animals to acute bright light. The knockout animals were completely protected against cell death, whereas the control mice exhibited distinct apoptosis, with DNA strand breaks, as confirmed by in situ nick end labeling (TUNEL staining), and internucleosomal DNA fragmentation, as visualized by a “ladder” after gel electrophoresis. ³⁸ Subsequent studies revealed a distinct pattern of AP-1 activation after damaging light exposure. Beginning at 30 minutes after the end of exposure, AP-1 activity increased to 6 hours after the end of exposure and declined at 12 hours and thereafter (Fig. 3) . ³⁹  

In view of our data on strain differences in the susceptibility to damage of mice, which is determined, at least in part, by genetic variants of the PE protein RPE65, might the sensitive genotype (L/L) overcome the protective effect of the c-fos gene knockout? In fact, c-fos ^−/− mice with leucine at codon 450 (L/L) had a significantly lower damage threshold than did mice with the methionine genotype. Other members of the AP-1 family replaced c-Fos—namely, Fra-2 and FosB—and apparently could function as the death mediators. ^{39 40} Thus, it appears that the activity of the transcription factor AP-1, rather than its component c-Fos, is necessary for the induction of apoptosis, at least in our experimental conditions. Support for this notion comes from earlier work that showed that Fra-1 could replace some c-fos-dependent functions in the retina and in bone development. ⁴¹  

We also created rd1 mice without the c-fos gene (rd1/rd1/c-fos ^−/−) to test the potential protective effect of the absence of c-Fos function on photoreceptor degeneration. ⁴² No rescue was observed in double-mutant mice, indicating an independence of the degenerative apoptosis from c-fos. Because the degenerative cell death occurs at the same time as the developmental, histogenetic apoptosis, it is conceivable that regulatory mechanisms of the latter mask any potential rescue effect provided by the absence of the c-fos gene. Alternatively, other members of the AP-1 family replaced the c-Fos protein, as was observed in the light-damage experiments. 

Are there pharmacological tools to block the activity of AP-1? When mice receive injections of dexamethasone (52 mg/kg intraperitoneally) before light exposure, their retinas are completely protected against cell death: the morphology shows regular features, AP-1 is not activated, and the electroretinogram (ERG) reveals normal function. ⁴³ Dexamethasone is known to bind the cytoplasmic glucocorticoid receptor. This complex translocates into the nucleus and inhibits AP-1 binding. ⁴⁴ Even though effects of corticosteroid hormones are multiple, dexamethasone application provides a tool to test the AP-1 dependence of apoptosis in our experimental conditions. For example, short-wavelength blue light readily induces severe apoptotic cell death in the retina and PE and an expression of AP-1; however, dexamethasone does not confer protection, as it does against white light, indicating that blue light–mediated cell death is AP-1 independent (Remé CE, et al. IOVS 2003;44:ARVO E-Abstract 5132). 

Similarly, AP-1 independent apoptosis was observed after application of the potent apoptosis inducer N-methyl-N-nitrosourea (MNU). c-Fos ^−/− and wild-type mice received intraperitoneal injections of MNU and were kept in darkness. AP-1 was upregulated in wild-type mice but not in c-fos ^−/− mice, whereas distinct photoreceptor apoptosis occurred in both groups. ⁴⁵ Thus, there are apoptotic pathways in the retina not primarily regulated by AP-1. 

Blue Light–Induced Lesions: Another Pathway of Light-Induced Apoptosis?

Lesions induced by high-energy blue light have been observed in primate cones (e.g., Ref. ⁴⁶ ), in rodent photoreceptors (e.g., Ref. ⁴⁷ ), and in several studies of the PE in vivo and in vitro ⁴⁸ (reviewed in Ref. ² ). Furthermore, there is ongoing discussion of whether life-long blue-light exposure, for example, as an important component of sunlight or of bright artificial light sources, may contribute to pathogenic steps in age-related macular degeneration (AMD). ⁴⁹ There are several chromophores in the retina and PE that absorb in the blue range and thus could mediate such lesions. 

We investigated whether rhodopsin and its bleaching products may play a role in blue light–induced lesions. To achieve fast and strong bleaching of rhodopsin in anesthetized rats, notably by a light source that does not emit any blue components, we used exposure to green light of 550 ± 10 nm. This exposure was rapidly followed by exposure to blue light of 403 ± 10 nm. We observed the distinct death of photoreceptor cells and PE, and spectrophotometry indicated that blue light induced a molecule maximally absorbing at 500 nm. ^{50 51} Other studies have shown the formation of such a molecule in vitro. ^{52 53} We also observed that blue light alone can inflict apoptotic death of visual cells, indicating to us, not only that rhodopsin readily absorbs blue light and is bleached, but also that this same blue light is absorbed by a photochemically active bleaching product that inflicts cellular damage (Remé CE, et al. IOVS 2003;43:ARVO E-Abstract 5132). ^{50 51}  

When anesthetized mice were exposed to blue light of 408 ± 10 nm (in previous studies in rats a 403 ± 10-nm blue light was used), distinct photoreceptor apoptosis occurred (Fig. 4) . In notable contrast, exposure to green light of 550 ± 10 nm did not induce any structural lesions in the retina and PE, similar to our observations in rats (Remé CE, et al. IOVS 2003;43:ARVO E-Abstract 5132). These findings further support the notion that one strong bleaching event of rhodopsin is not sufficient to induce apoptosis but rather, a further molecular mechanism is required. During exposure to white, green, or blue light, bleaching of rhodopsin is supposed to generate a molecule that, by itself, is a strong absorber of blue light and thus a potential mediator of cell death. Inhibition of metabolic rhodopsin regeneration by halothane (as described earlier) did not prevent blue light–induced cell death, suggesting the rapid formation of such a detrimental molecule by blue light independent of metabolic regeneration. 

AP-1, which is a crucial mediator of white light–induced apoptosis, was upregulated after blue but also after green light exposure. Dexamethasone, which inhibits AP-1 expression and completely protects against white light–induced lesions, did not prevent cell death after blue-light exposure (Remé CE, et al. IOVS 2003;43:ARVO E-Abstract 5132). Rpe65 knockout mice crucially demonstrated the role of rhodopsin also for blue light lesions. There was no apoptotic cell death in knockout mice but distinct apoptosis in wild-type control mice. ⁵¹ Thus, all lesions described so far in our light-damage studies appear to depend on rhodopsin. However, consecutive steps in the death cascade may differ. Current studies in our laboratory in knockout mouse models lacking phototransduction (Gnat 1α^−/−) or rhodopsin inactivation (Rhok^−/−/Sag^−/−) or lacking a functional c-fos gene (c-fos ^−/−) in both genotypes for RPE65 indicate that none of these conditions can prevent blue light–induced lesions (Remé CE, unpublished observations, 2004). 

PE chromophores may also absorb blue light and generate photo-oxidative lesions in PE cells and photoreceptors. An important molecule has been demonstrated as a strong PE photosensitizer: the lipofuscin fluorophore A2E, a pyridinium bis-retinoid generated from bleached rhodopsin (two molecules of all-trans retinal) and disc membrane phospholipid (1 molecule of phosphatidylethanolamine). ^{8 54} Apart from its light-independent toxic effect on PE functions, ^{55 56} A2E induces free radicals by photo-oxidation in lipofuscin ⁵⁷ and apoptosis when PE cells are irradiated with blue light. ⁴⁸  

In summary, we have shown that the threshold for white light–induced apoptosis critically depends on the rate of metabolic rhodopsin regeneration in mice. The regeneration rate, in turn, depends on the level of the PE protein RPE65. The genetic variants leucine or methionine determine availability of RPE65. In rats, by contrast, damage susceptibility and rhodopsin regeneration rate do not critically depend on RPE65. Other determinants, such as time of day, photoreceptor outer segment lipid and protein composition, antioxidant state, and, to date, unknown factors, contribute to the damage threshold. ^{21 22 23 24 25} In mice, the proapoptotic transcription factor AP-1 is essential in white light–mediated cell death. To date, it is unclear, which signals transduce the light–induced message to transcriptional activity of AP-1, and further, whether AP-1 transactivates death genes or transsuppresses protective ones or whether there may be any other more indirect of effect of AP-1. Phenomenologically, we have observed that apoptotic bodies are removed by abundant invading phagocytic cells and by the PE. Signaling and execution of phagocytic activity remain to be elucidated. 

Concerning blue light–induced apoptosis, we have shown that, similar to white light, the photon absorption by rhodopsin plays a critical role. A special feature of blue light, however, comes into play by the fact that rhodopsin bleaching product(s) strongly absorb in the blue (and near UV) range and possibly induce detrimental photochemical lesions in photoreceptors. The determination of such damaging molecules and their absorption characteristics is a logical continuation of studies on blue light–induced damage. Such studies would have practical implications. Manufacturers of therapy lamps, sunglasses, and intraocular lenses would know at which point of the visible spectrum a meaningful cutoff or reduction of blue light should be located. 

Possible Links of Apoptosis to Intracellular Degradation: Autophagy, Lysosomes, and Proteasomes

In earlier years apoptosis was understood as induced by a multitude of stimuli, including external components such as toxins, UV, and visible light, as well as endogenous factors such as cytokine signaling via cytokine receptors. Death messages were thought to be conveyed through mitochondrial release of cytochrome c to one major pathway of initiation and execution of apoptosis by caspases, the endopeptidases that cleave proteins at aspartate residues. At present, evidence is accumulating that caspase-independent proteolysis may play an important role in apoptosis, either in concert with or without caspases. Proteolytic enzymes include cathepsins, serine proteases, calpains, and granzymes A and B. ^{58 59}  

Apoptosis further includes internucleosomal cleavage of DNA ⁶⁰ and externalization of membrane phosphatidylserine as a signal for phagocytosis of dead cells. ⁶¹ Finally, the formerly strict distinction between apoptosis and necrosis no longer applies, because morphologic features of both types of cell death are observed in dying cells, ⁶² and the same proteolytic machinery may be used by both apoptosis and necrosis. ⁶³  

Further intracellular degradative pathways comprise different cytoplasmic functional units: autophagosomes, lysosomes, and proteasomes. These pathways may operate independently or in concert with caspases to initiate and execute apoptosis. ^{64 65 66 67 68 69 70} These proteolytic units may interact at different points in the hierarchy of death pathways, with or without involvement of caspases. Neither autophagy nor lysosomal or proteasomal degradation kills the cells; rather, they subserve physiological responses to external or internal stimuli in homeostatic conditions. In this context, proteolysis is viewed as adaptational, and only in “deadly” situations would it activate apoptotic pathways. 

In the retina and PE, such alternative enzyme and organelle systems and their relation to caspases and to apoptotic pathways are as yet unexplored. 

Autophagic degradation implies the bulk degradation of cytoplasmic organelles by lysosomes to remove dangerous molecules or for the purpose of recycling—for example, during starvation in yeast ⁷¹ —and for reuse of important substrates in mammals. ⁷² Autophagy genes are highly conserved from yeast to mammals. ⁷³ In neurons, autophagy can be activated by apoptosis signals. ⁷⁴  

We have extensively studied autophagy in photoreceptors and PE under various conditions, including different light regimens (e.g. Ref. ⁷⁵ ). In early stages of light–induced photoreceptor apoptosis, there are abundant autophagic vacuoles in rod inner segments, suggesting possible links between autophagy and apoptosis (Fig. 5) . Of interest, rather scarce autophagy was seen in photoreceptors from animals protected against light-induced apoptosis, such as the c-fos ^−/− mouse with the genetic variant methionine in the RPE65 protein and the Rpe65 ^−/− mouse (Fig. 6 ; Remé CE, unpublished observations, 2004). In an earlier study we interpreted a significant increase of autophagy after acutely changing illuminance levels from low to bright as an adaptive measure to prevent induction of cell death. Adaptation was construed to consist of removal of opsin in rod inner segments to reduce visual pigment content and, with that, sensitivity to light. ⁷⁶ In view of potential apoptotic signaling by autophagy, there may be a point of no return beyond which autophagy confers a death message. 

Ubiquitin-dependent proteolysis by proteasomes also has been observed in the retina. Light exposure induces ubiquitin conjugation, and degradative activity in the rat retina, ⁷⁷ the βγ subunits of the G-protein transducin, and rhodopsin, have been shown to be a substrate of ubiquitin-mediated degradation, ^{78 79} and ubiquitin was observed in rod inner and outer segments in the rat retina after illuminance levels were acutely changed from low to high. ⁷⁶ It is important to note that oxidized proteins in the retina are removed by proteasomes. ⁸⁰ In other tissues, misfolded or mutated proteins have been shown to be targets of proteasomes and, in some instances, also of autophagy. ⁶⁹  

Some of the lysosomal proteases can also function at neutral pH, that is, in the cytosol, and lysosomal proteases can trigger apoptosis by multiple molecular pathways including caspase activation and the Bcl-2 family. ⁶⁴ In the retinal PE, lysosome-triggered apoptosis has been amply documented, assigning an important role to the lipofuscin fluorophore A2E (N-retinylidene-N-retinylethanolamine). ⁸ Lysosomal dysfunction may be induced by A2E, which is a lysosomotropic compound ⁵⁶ or by photodamage. ⁴⁸ Alternatively, A2E may be released from lysosomes and taken up into mitochondria, leading to mitochondrial damage and release of proapoptotic factors. ⁵⁵  

Deficient Degradation in Retinal Degenerations: A Pathway to Cell Death?

A multitude of mutations are known (www.sph.uth.tme.edu/Retnet/ provided in the public domain by the University of Texas Houston Health Science Center, Houston, TX) that lead to retinal degenerations with the endpoint of apoptotic cell death. However, much less data exist on the signaling and molecular pathways that comprise the road to cell death. Increasing evidence indicates that insufficient intracellular degradation may be an important contributory factor. Altered gene products, such as those resulting from rhodopsin gene mutations, can lead to misfolded proteins that cannot be degraded by the proteolytic machinery and thus accumulate. ^{81 82 83 84} Those “aggresomes” may eventually lead to cell death by inhibiting vital metabolic functions. Comparable mechanisms have been described for several neurodegenerative diseases of the central nervous system. ⁸⁵ It is unclear to date how visual cells deal with the abnormal proteins resulting from different gene mutations in retinal degenerations and how abnormal proteins convey the message for apoptotic cell death. In such cases, an efficient rescue strategy would include the removal of aggregated and undigestible proteins. ⁸⁵  

An important role in retinal degenerations is ascribed to molecular chaperones, which cannot properly handle misfolded or otherwise altered molecules that therefore accumulate and disturb cellular function. ⁸⁶ Another level of complexity is reached when proteasome systems or molecular chaperones themselves bear mutations alone or in addition to a mutation leading to an abnormal protein of important cellular function. ⁸⁷  

Proteasomes are known to remove oxidized proteins. In view of light-induced apoptosis it is important to note that impaired proteasome activity can potentiate the effects of oxidative stress. ^{87 88} Oxidative stress is readily encountered after damaging light exposure (for review, see Refs. ² , ⁸⁹ ). Thus, the severity of light-induced degeneration may be modulated by the functional capacity of the proteasome system. 

The role of autophagic degradation in retinal degenerations remains to be investigated. Activation of autophagy has been observed in those neurodegenerative diseases that are characterized by aggregations of abnormal proteins. ⁷⁴ Autophagy was suggested to be increased as a compensatory mechanism for a deficient proteasome pathway. ⁹⁰ One may envision that failing proteasome and autophagy function eventually constitutes a death signal in those degenerative diseases associated with abnormal protein (and lipid) aggregation. 

Light and AMD: The Discussion Continues

Recent studies have been focused on a role of inflammation and immune responses in the pathogenesis of AMD. A key role is ascribed to injured PE cells, which may activate microglia-derived dendritic cells that in turn constitute a core for drusen formation. Recent analyses of drusen indicate the likely origin of many of their components in PE cells. The necrotic and apoptotic cells that are not rapidly removed by phagocytes are known to generate stimuli for autoimmune responses. ⁹¹ In contrast, clearance of apoptotic cells can also induce immune responses and inflammation—for example, by triggering caspase-1-mediated release of IL-1β from dying cells. ⁹² Drusen proteome analysis has revealed, among several other elements, protein and lipid oxidation products, which also arise after bright light exposure and may represent molecules damaging the PE (Reganathan K, et al. IOVS 2004;45:ARVO E-Abstract 1795). ⁹³ Thus, numerous pathogenic stimuli emerge that eventually lead to drusen formation and AMD. ^{94 95 96} However, it has been known for a long time by clinicians that not all eyes that accumulate lipofuscin and develop drusen will have AMD. A potential missing link was recently described by using mice deficient for monocyte chemoattractant protein-1 (Ccl-2/MCP-1) or its chemokine receptor. ⁹⁷ Those mice exhibited several features similar to those observed in human AMD. Thus, a macrophage dysfunction leading to the accumulation of cellular debris and drusen formation was implicated as an important factor in the pathogenesis of AMD. 

In an earlier review, we have summarized our observations on light exposure eliciting inflammatory responses on a cellular and molecular level. ² Macrophages removing apoptotic photoreceptor cells are distinct in our models. In some species such as the rabbit, ⁹⁸ PE cells appear to proliferate, besides removing cellular debris. We have also observed light-evoked release of eicosanoid mediators derived from retinal phospholipids, which in turn may activate cytokine responses. Those effects were suggested to be contributory to the pathogenesis of AMD. ²  

Light can also release polyunsaturated fatty acids in the retina, such as arachidonic acid (AA) and DHA. ^{99 100} Dietary deprivation of DHA, by contrast, protects against light-induced damage. ¹³ The carboxyethyl pyrrole adducts demonstrated in drusen are exclusively formed from the oxidation of docosahexaenoate-containing lipids, ⁹³ suggesting exposure to light as a possible factor in drusen formation. 

Light-Induced Apoptosis: A Good or a Bad Model for Learning about Human Disease?

Many models are born with the “genetic defect” of limited applicability to human disease. An obvious but unavoidable obstacle is the fact that hardly any animal model reaches the lifespan of humans. Thus, this very simple condition enables many more environmental stress factors to interact with different genes or affect directly the human eye than can ever be simulated in the long term in animal models. Animal models, therefore, often have to be manipulated in a time-lapse fashion. This manipulation is essentially what our laboratory has developed with short-term high illuminance levels to induce apoptosis. This model has the advantage of synchronized cell death and, therefore, an easier way to interact with apoptosis. Furthermore, short-term exposures to light reduce the probability of masking by secondary phenomena after the initial event of cell death. The knowledge from this artificial situation does help in the study of how the apoptotic machinery works in the retina and which rescue strategies are effective. With this gain, we can then move to animal models of retinal degenerations and analyze differences and commonalities of both conditions. 

An essential difference between light- and mutation-induced apoptosis needs consideration in the context of cell rescue. Light evokes apoptosis in an otherwise healthy cell, and rescue would restore this health. This is in contrast to mutation-induced apoptosis, in which abnormal proteins are synthesized from birth, and rescue may prevent cell death, but in most cases does not repair the metabolic defect resulting from the mutation. Nevertheless, there are several animal models in which rescue works in both light-damage–and mutation-induced apoptosis, perhaps indicating similar mechanisms of apoptosis. ⁵  

Apart from representing a model system, light-induced apoptosis is directly linked to human life. Light-induced retinopathies are observed in the clinic and the laboratory, and findings in epidemiologic studies have suggested that exposure to light can contribute to AMD. Therefore, the development of preventive measures is important, as is the elucidation of pathogenic mechanisms through which light promotes retinal degenerations. 

 Supported by Swiss National Science Foundation, the German Research Council and numerous private foundations in Switzerland and Germany.

 Disclosure: C.E. Remé, None

 

Figure 1.

 

Light micrographs of albino rat retinas exposed to different intensities of white fluorescent light: (A) 1000 lux for 30 minutes, death 24 hours after the end of exposure to light. Dilations and disruptions of rod outer segment disks are visible (arrow). (B) At higher light doses (1000–3000 lux for 1 hour) distinct swelling of PE cells and numerous newly shed phagosomes are visible (star). Rod inner segments are condensed (arrowhead), and several nuclei are pyknotic (arrow), indicating apoptosis. (C) At 48 hours after the end of light exposure, most rod nuclei are dissolved, inner segments have disappeared, and outer segments are fragmented. Large phagocytic cells (arrow) invade the outer retina. Note the thinning of PE cells. (D) Representation of cellular changes. PE, pigment epithelium; OS, outer segments; IS, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer.

Figure 2.

 

Exposure of mice without rhodopsin kinase and arrestin (Rhok^−/−/Sag^−/−) and wild-type controls. (A) As little as 2 minutes of exposure to white fluorescent light of 5000 lux causes detrimental lesions with apoptotic cell death in double-knockout mice, photoreceptor nuclei are pyknotic (arrow), outer segments are destroyed, and PE cells are distinctly swollen (star). (B) Wild-type control mice showed no structural lesions after 20 minutes of the same lighting regimen. (C) The transcription factor AP-1 in double-knockout mice was activated at 20 minutes and 2 hours after exposure. (D) Wild-type control animals show AP-1 activation at 6 hours after damaging light exposure. (E) Rhodopsin regeneration in double-knockout mice is completed at 1 hour after strong bleaching. The same kinetic was observed in wild-type mice of identical genotype for the RPE65 protein. Abbreviations as in Figure 1 . (A–D) Reprinted, with permission, from Hao W, Wenzel A, Obin MS, et al. Evidence for two apoptotic pathways in light-induced retinal degeneration. Nat Genet. 2002;32:254–260. © The Nature Publishing Group.

Figure 3.

 

Time course of activation of the proapoptotic transcription factor AP-1. (A) After light exposures leading to photoreceptor apoptosis, AP-1 activity increases for up to 6 hours after the end of light exposure. The increase begins at approximately 30 minutes and gradually declines at 12 hours. (B) In mice without a functional c-fos gene, no increase in AP-1 activity is seen (M/M genotype for RPE65 protein, but see section Regulative Mechanisms in Bright Light-Induced Apoptosis). Reprinted, with permission, from Wenzel A, Grimm C, Marti A, et al. c-fos controls the “private pathway” of light-induced apoptosis of retinal photoreceptors. J Neurosci. 2000;20:81–88. © Society for Neuroscience.

Figure 4.

 

Blue light–induced lesions in comparison to no lesions after exposure to green light. (A) Exposure to short-wavelength blue light (408 ± 10 nm) caused a distinct “hot spot” in the central retina. (B) At 48 hours after the end of light exposure, the photoreceptor layer within the hot spot have disappeared, and giant phagocytic cells invade the subretinal space (arrow). (C) Exposure to green light (550 ± 10 nm) inflicts no hot spot and (D) no structural lesions on the light microscopic level. (E) Notably, the proapoptotic transcription factor is upregulated at 6 hours after both blue- and green light exposures of 1 minute and 15 minutes, respectively. Abbreviations as in Figure 1 .

Figure 5.

 

Autophagic vacuoles in photoreceptor inner segments. (A) Schematic drawing depicting the process of autophagic vacuole formation and degradation. Cytoplasmic organelles (e.g., mitochondria, are surrounded by double membranes, forming autophagic vacuoles which then fuse with primary lysosomes and thus constitute secondary lysosomes or phagolysosomes). (B) Electron micrograph of a newly formed autophagic vacuole containing mitochondria (arrow) in a frog ellipsoid. (C) Electron micrograph of degraded contents of an autophagic vacuole (arrow) in a frog myoid, with acid phosphatase staining for secondary lysosomes.

Figure 6.

 

Electron micrographs of mouse rod inner segments. (A) Light-induced early apoptotic stages in wild-type mice show condensed cytoplasm of inner segments with abundant autophagic vacuoles (arrow). Note the relatively well-preserved mitochondria in apoptotic and nonapoptotic inner segments. Arrowhead: outer limiting membrane. (B) Mice that are protected against light-induced apoptotic cell death such as the Rpe65 ^−/− mice and c-fos ^−/− mice of the methionine-450 genotype for the RPE65 protein reveal only few autophagic vacuoles after exposure to damaging light doses.

My very sincere and wholehearted thanks go to ARVO—the Awards Committee, the Board of Trustees, and the society as a whole—for bestowing this great honor on me. 

Special thanks go to my friend and colleague François Delori for introducing me in such an elegant and humorous way; to Christian Grimm and Andreas Wenzel for being my innovative, gifted, and loyal colleagues in the laboratory; and to Farhad Hafezi for having been an original and loyal coworker while he was a postdoctoral fellow in our laboratory. 

I will always appreciate past and present collaborators for essential input in our work: Anna Wirz-Justice, Michael and Jiuan Terman, Gunter Niemeyer, Armand Malnoe, Joseph Pfeilschifter, Graig Eldred, Pascal Rol, Andreas Marti, Jim Dillon, and Uwe Wolfrum, Edward N. Pugh, and Trevor Lamb. 

Our group owes fruitful collaborations to Matthias Seeliger, Eberhard Zrenner, Deborah Farber, Don Fox, Michael Redmond, Michael Danciger, Pete Humphries, Janis Lem, Martin Obin, Christoph Richter, Mel Simon, Wolfgang Berger, Françis Munier, Ivan Arsenijevic, John Crabb, Theo van Veen, Elisabeth Rungger, Rex Martin, members of the program “age-related macular degeneration” of the German Research Council. 

Several postdoctoral fellows and visiting scientists did great work in our laboratory: Thomas Hoppeler, Philip Hendrikson, Enrica Strettoi, Eveline Federspiel, Ursula Urner, Hans Jung, Urs Braschler, Ron Bush, Jörg Reinboth, Matthias Clausen, Marijana Samardzia, Charlotte Keller, Rico Frigg, and Hans-Peter Iseli. 

Our work over many years could not have been done without the excellent technical help of Maja Sulser, Sylvia Hoffmann, Esther Bossuge, Barbara Aeberhard, Astrid Rhyner, Kurt Munz, Cornelia Imsand, Gabi Hoegger, and Dora Greuter. 

Among colleagues, there are my friends Dan Organisciak and Barry Winkler, with whom I shared many stimulating and refreshing conversations and hearty laughs. Highly respected and longstanding colleagues are Alan Bird, Dean Bok, Matt LaVail, Gene Anderson, and Joe Hollyfield. 

I will never forget Ted Williams, not only as the man who, among many other skills, taught us to count photons, but also as an amiable, generous, and gentle friend, and his widow Ruth Anne, for years of warm and lively friendship. 

Throughout life, it was not easy to become a reasonable human being. Without the help of my parents and my partner Heinrich Schmid, I would not have made it. 

A similarly long distance on the road of professional life could not have been covered without my great mentor Richard W. Young, who introduced me to the world of visual cells and provided vital encouragement to go ahead in science. Rudolf Witmer was a pioneer in founding a research laboratory within a clinical setting. He helped in my decision to move from clinical work to research and was my role model for generosity and tolerance as head of the Ophthalmology Department in Zurich. Balder Gloor, as subsequent head of the Department of Ophthalmology, essentially facilitated and continuously encouraged the growth of our laboratory and my professional career. 

I thank Christina Fasser, president of Retina International and Retina Suisse, for many years of friendship. It is with great and unceasing admiration that I regard her and the incredible courage and good spirit she demands of herself day after day and passes on to her fellow beings. 

Last but not least, I thank the Swiss National Science Foundation, the German Research Council and numerous private foundations in Switzerland and Germany for continuous and essential support of our work.