retinal pigment epithelium pigment granules stimulate the photo-oxidation of unsaturated fatty acids

Original Contribution

RETINAL PIGMENT EPITHELIUM PIGMENT GRANULES STIMULATE THEPHOTO-OXIDATION OF UNSATURATED FATTY ACIDS

ALEXANDER E. DONTSOV,* RANDOLPH D. GLICKMAN ,† and MIKHAIL A. OSTROVSKY**Institute of Bio-Chemical Physics of the Russian Academy of Sciences, Moscow, Russia and†Department of Ophthalmology,

The University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

(Received28 July 1998;Revised16 December1998;Accepted16 December1998)

Abstract—The cellular pigments of the retinal pigment epithelium (RPE) have been shown to catalyze free radicalactivity, especially when illuminated with visible or ultraviolet light. This activity is sufficient to cause photooxidationof several major cellular components. The present investigation determined the relative ability of melanin, lipofuscin,and melanolipofuscin granules isolated from human and bovine eyes to oxidize polyunsaturated fatty acids, specificallylinoleic and docosahexaenoic acids. The dark reactivity as well as the light-stimulated reactions were determined. Theproduction of hydroperoxide derivatives of the linoleic and docosahexaenoic acids were determined by NADPHoxidation coupled to the activity of glutathione peroxidase, and also by production of thiobarbituric acid reactivesubstances. All RPE pigment granules stimulated fatty acid oxidation when irradiated with short wavelength (,550 nm)visible light, with the melanosomes exhibiting the greatest light-induced activity. Only lipofuscin granules, however,caused peroxidation of fatty acids in the dark. These findings provide additional support for the role of RPE pigmentsin “blue light toxicity” as well as indicating that accumulation of lipofuscin may contribute to increased photooxidationin the aging RPE. © 1999 Elsevier Science Inc.

Keywords—Retinal pigment epithelium, Melanin, Melanolipofuscin, Lipofuscin, Polyunsaturated fatty acids, Photooxi-dative stress, Retinal degeneration, Free radicals

INTRODUCTION

The retinal pigment epithelium (RPE) of the eye con-tains several pigments that possess photoactive prop-erties but are not specifically involved with visualtransduction. Melanin, the heteropolymeric oxidationproduct of tyrosine and DOPA or its derivatives [1], isa broadband absorber and is generally thought to pro-tect ocular tissues against excess light. The protectionpresumably offered by melanin could derive from itsability to absorb and screen light from reaching sen-sitive tissues [2], to sequester heavy metals that mightotherwise catalyze oxidative reactions [3], or to trapfree radicals produced by photochemical reactions orionizing radiation [4,5].

Melanin, however, is also capable of inducing oxida-tive changes in physiologic substrates such as ascorbic

acid, fatty acids, and proteins during irradiation withvisible light [6–10]. Although the significance of thesefindings for photo-oxidative stress in vivo is not entirelyclear, they do raise the possibility that pigments in theRPE cells may have a dual-edged role in both protectingagainst and promoting light damage, depending on thecellular environment and the nature of the oxidizingstressor. On the protective side, melanin has been shownto interact with reactive oxygen species such as hydroxylradical, and has been proposed to have a role in scav-enging free radicals in pigmented tissue [11,12]. TheRPE and choroidal pigments also absorb up to 60% of alllight entering the eye [13], that otherwise would degradeimage contrast because of intraocular scatter. Scatteredlight presumably also contributes to general photochem-ical damage to ocular tissues. Supporting the phototoxicview, however, are the observations that melanin is aphotoinducible free radical [14,15], produces reactiveoxygen species when irradiated by UV and visiblelight [16,17], and promotes photochemical oxidations[9,18,19].

Address correspondence to: Randolph D. Glickman, PhD, Departmentof Ophthalmology, The University of Texas Health Science Center, 7703Floyd Curl Drive, San Antonio, TX 78284-6230, USA; Tel: (210) 567-8420; Fax: (210) 567-8413; E-Mail: [email protected].

Free Radical Biology & Medicine, Vol. 26, Nos. 11/12, pp. 1436–1446, 1999Copyright © 1999 Elsevier Science Inc.Printed in the USA. All rights reserved

0891-5849/99/$–see front matter

PII S0891-5849(99)00003-9

1436

The other pigment of note in the RPE is lipofuscin.Long considered an “aging” pigment [20 –22], lipo-fuscin has at least two UV-induced chromophores, oneabsorbing maximally at 430 nm and another absorbingat 580 nm [23]. These transient species are reactiveand can initiate photo-oxidative reactions, includingthe generation of singlet oxygen [23], superoxide an-ions [24], and lipid hydroperoxides [25]. The declinein melanin content in RPE and choroid with age isoften associated with an increase in lipofuscin content[20 –22], so that an increase in lipofuscin content inRPE cells has been considered a potential cause ofincreased oxidative stress in the eye [26].

In view of these observations, we compared thephotoinduced as well as constitutive (dark) reactionsof RPE melanin, lipofuscin, and complex (melanoli-pofuscin) granules in a model system with the produc-tion of fatty acid hydroperoxides as the endpoint. Theexperiments showed that although all the RPE pig-ments were capable of mediating the photooxidationof fatty acids, lipofuscin had the highest constitutivereaction rate, and thus may be responsible for increas-ing the basal level of oxidative stress in the RPE andpossibly the choroid. These findings further supportthe role of pigment-mediated photochemical reactionsin the RPE in the development of age-related retinaldegenerations.

MATERIALS AND METHODS

Preparation of pigment granules

Melanosomes (MS) were prepared from human andbovine eyes. Human eyes were donated by the Eye Bankof The Moscow Institute of Eye Microsurgery (Moscow,Russia), and were from donors aged 40 to 65 years whowere free of any ophthalmological disorders. The lipo-fuscin (LF) granules and melanolipofuscin (MLF) gran-ules were obtained only from human eyes. Human eyeswere processed for MS, LF, and MLF by a methodmodified from that described by Boulton and Marshall[27]. Retinal pigment epithelium cells obtained from theeyes were sonicated for 60 s at a frequency of 22 kHz andmaximal resonance. The retinal pigment epithelium cel-lular debris was removed by centrifugation at 603 g for10 min. The resulting supernatant was then centrifuged at60003 g for 10 min. The second centrifugation precip-itated the pigment granules, that were resuspended in 0.3M sucrose, layered on a sucrose density gradient (mo-larity range: 2.00:1.80:1.60:1.55:1.50:1.40:1.20), andcentrifuged at 103,0003 g for 1 h. The band containingthe LF granules was localized between the 1.20 M and1.40 M sucrose layers, the MLF granules band wasbetween the 1.55 M and 1.60 M layers, and the precip-

itate contained the melanosomes. The fractions wereisolated and centrifuged once more in the sucrose densitygradient. The resulting purified granules were washed in0.1 M phosphate buffer several times to remove sucrose.The granules were stored at220°C until use.

Bovine eyes were obtained freshly in San Antonio(TX, USA) from local slaughterhouses. Only melano-somes were prepared from the bovine material; fewlipofuscin granules were present in these eyes. Afterdissecting away the anterior segments, and removingthe retinae, the remaining eyecups were rinsed withtris-buffered saline. The RPE and choroid were re-moved from the eyecups, placed in 0.25 M sucrose inplastic centrifuge tubes, and vigorously vortexed for30 s to dislodge RPE cells. The large pieces of choroidwere removed from the centrifuge tubes and the ma-terial left behind was sonicated for 30 s at maximalpower. The tubes were placed in swinging-bucketrotors and centrifuged at 25003 g for at least 10 min.The resulting pellet was resuspended in 10 ml of 0.25M sucrose, leaving behind any large clumps of tissue,and spun again at 25003 g for 10 min. This secondpellet was resuspended in about 10 ml of 0.1 Msucrose, layered on top of 2 M sucrose, and centri-fuged at 25003 g for 30 min. The resulting separationresulted in a clear, upper fraction and a “fuzzy” inter-face, that were both discarded. The lower fraction, thatcontained “light” (round) melanosomes was removedand saved. The pellet contained the “heavy” (ellip-soid) melanosomes. The lower fraction containing thelight melanosomes was further purified by diluting it1:1 with distilled water and centrifuging at 80003 gfor 20 min. The melanosomes were resuspended in0.25 M sucrose and stored at220°C. Before use, themelanosomes were spun and resuspended twice in 0.1M potassium phosphate buffer to remove sucrose. Allof the experiments reported here with bovine melano-somes used the “light” fraction; however, in previouswork using NADPH oxidation as an assay [9], wefound no difference in the photochemical reactivity ofthe “heavy” and “light” fractions.

Measurement of fatty acid peroxidation

The peroxidation of fatty acids was determined by thereaction of NADPH-dependent glutathione peroxidasewith the hydroperoxide products, and the extent of thereaction was measured by the decrease of NADPH ab-sorption at 340 nm [28]. The production of thiobarbituricacid reactive substances (TBARS) [29,30] was also usedto assay the hydroperoxides. Linoleic acid (LA) andcis-4,7,10,13,16,19-docosahexaenoic acid (DHA) (bothfrom Sigma Chemical, St. Louis, MO, USA) were usedas the fatty acid substrates.

1437Photooxidative reactions of RPE pigments

Linoleic acid reactions

For LA reactions, 0.72 mg/ml of LA was mixed with0.1 M potassium phosphate buffer, pH 7.44, and varyingdensities of pigment granules. Photooxidation was initi-ated by exposure to a light source that was, in differentexperiments, either a Xenon arc lamp, or an Argon ioncontinuous wave laser. After light exposure, an aliquotcontaining approximately 140–150mg LA was added to0.1 M potassium phosphate buffer, pH 7.44, 1 mMethylenediaminetetraacetic acid (EDTA), 1 mM GSH(reduced form of glutathione), 200mM NADPH, 1.25units of glutathione reductase, and 0.6 units of glutathi-one peroxidase. To check the enzymatic activities,40–50mg tert-butyl hydroperoxide was added at the endof the measurement period.

Decosahexaenoic acid reactions

For DHA reactions, 70–80mg/ml DHA was mixedwith 0.1 M potassium phosphate buffer, pH 7.44, 1 mMEDTA, 1 mM GSH, 0.7 units GSH reductase, 0.4 units ofGSH peroxidase, 200mM NADPH, and varying densi-ties of pigment granules. The sample was divided intotwo equal parts, one for light exposure, and one for darkcontrol.

Thiobarbituric acid reactive substances reactions

Thiobarbituric acid reactive substances was deter-mined in DHA samples prepared in 0.1 M potassiumphosphate buffer, pH 7.44, 110–130mg/ml DHA, andvarying densities of pigment granules. The sample wasdivided into equal parts for laser exposure and darkcontrol. After the exposure period, the samples weremixed with 15% trichloroacetic acid and 30 mM thio-barbituric acid and boiled at 100°C for 15 min. After thesamples cooled to room temperature, the absorbance at532 nm was measured.

“White light” exposure

Samples were exposed for 10 to 20 min to the broad-band output of a 150 W xenon arc lamp. The sampleirradiance achieved with this system was about 112 mW/cm2. Samples were mixed continuously during exposure.

Laser exposure

A continuous wave, Argon ion laser (a CoherentModel 920 photocoagulator) was used to generate blue-green light to excite pigment reactions. The laser wasused in a mixed output mode so that both 488.1 nm and514.5 nm wavelengths (in a ratio of;45:55%) reached

the sample. The beam was expanded into an approxi-mately 0.5 cm diameter beam to irradiate the entiresample. Sample irradiance was adjusted as required from50 to 1500 mW/cm2, and was monitored continuouslyduring the exposure by means of a prism beamsplitterthat deflected a portion of the beam into a power detec-tor. Laser power measurements were made with a Molec-tron EPM-1000 meter fitted with a PowerMax 3 head(Molectron Detectors, Inc, Portland, OR, USA).

Experiments with a superoxide anion generatingphotosensitized system

Superoxide anions were generated by a photosensi-tized system [31] containing 14 mM methionine, 12mMriboflavin, 110–120mg/ml DHA, varying densities ofpigment granules, in 0.1 M potassium phosphate buffer,pH 7.44. The system was excited by exposure to bluelight derived from a 150 W xenon arc lamp with a 400nm peak bandpass filter (Oriel 59820) inserted into itsoutput. The sample irradiance was 5.3 mW/cm2. ForTBARS assay, aliquots were taken at 0, 10, and 20 minof light exposure.

RESULTS

Photooxidation in the absence of pigment granules

The fatty acids, LA and DHA, have negligible opticalabsorbance in the visible spectrum; therefore, when anexposure source with output restricted to the visiblewavelengths was used, such as the Argon ion laser, therewas very little photooxidation of the fatty acids in theabsence of pigment granules. In the experiment pre-sented in Fig. 1, the peroxidation rate of DHA alone,exposed to the blue-green output of the Argon laser, wasvirtually identical to that of unexposed DHA kept in thedark for an equal amount of time. In contrast, whenhuman melanosomes were present at a density of 2.23107 gran/ml in the reaction mixture, the amount of pho-tooxidized DHA increased linearly with the duration ofthe laser exposure (Fig. 1). LA was also photooxidizedby the laser when melanosomes were present in thereaction mixture (data not shown), but the extent of thereaction was considerably less than with DHA.

Comparison of the photooxidizing capacity of the RPEpigment granules

The three types of pigment granules, MS, LF, andMLF, extracted from human RPE cells were comparedwith respect to their ability to photooxidize DHA. Reac-tion mixtures were made up with DHA and either: hu-man MS (1.953 107 gran/ml), LF (1.723 107 gran/ml),

1438 A. E. DONTSOV et al.

or MLF (1.573 107 gran/ml). Aliquots were exposed tothe argon laser for 10 min at 0.8 W/cm2, or kept in thedark as a control. The samples were then analyzed forhydroperoxides by the NADPH-glutathione peroxidase

assay. The results are shown in Fig. 2A, with the dataexpressed as the difference between the laser-exposedand dark control samples. The samples with the MSgranules contained the largest increase in hydroperoxidesafter laser exposure, although LF- and MLF-contain-ing samples had lesser increases in hydroperoxides,respectively.

The results of this experiment suggested that MSgranules may have the highest photoactivated reactivitytowards DHA; indeed, this conclusion is emphasized ifthe data are replotted as a light/dark ratio (Fig. 2B). Twoprincipal conclusions may be drawn from this analysis ofthe data. The first is that the difference between the lightand dark levels of oxidizing activity is the greatest formelanin, whereas LF and MLF granules show lowerratios of light-to-dark activity. Of the three types of RPEpigment granules, the photoinduced activity of MS gran-ules grew most rapidly with increasing irradiance. Witha sample irradiance of 800 mW/cm2, the light activity ofMS rose to nine times its dark level of activity, increas-ing from a ratio of about 4 with a sample irradiance of200 mW/cm2. LF and MLF granules, in comparison,only increased from a ratio of 2 up to about 3.5 over thisrange of irradiance. The second conclusion is that, al-though the MS granules exhibited the greatest increase inlight-induced oxidative reactivity, LF granules possessedthe greatest total reactivity towards DHA, i.e., the sum ofthe DHA hydroperoxides produced during dark and light.This may be appreciated from Fig. 2C, that shows totalDHA oxidized by the RPE pigment granules during darkand light. The activity of LF and MS granules wasapproximately equal during laser irradiation, but during

Fig. 1. Thekinetics of DHA peroxidation by argon laser exposure in thepresence and absence of human melanosomes (MSH). When present,human melanosomes were added at a density of 2.23 107 gran/ml. Thereaction was monitored by the oxidation of NADPH coupled to GSHperoxidase. Results were converted into nmol ROOH per mg initial DHA.See text for additional experimental details. Filled squares: autoxidation ofDHA in the absence of melanosomes or laser; filled diamonds: DHAsolution exposed to laser but without melanosomes; filled circles: DHAsolution with melanosomes and exposed to laser output for the durationsindicated. At 30 m,t-butylOOH was added to reaction mixture to testactivity of GSH peroxidase assay system.

Fig. 2. Comparison of DHA peroxidation produced by photoactivation of the RPE pigment granules. Argon laser exposure was for aduration of 10 min at the indicated sample irradiances. Each point is the average of two measurements, and the error bars indicate thedata range. (A) DHA hydroperoxides produced by photoactivated pigment granules, i.e., the light minus the dark activity. (B) The ratioof light to dark activity. (C) Total DHA hydroperoxides produced, i.e., the sum of the dark plus light activity. (A) and (C)Decosahexaenoic acid peroxidation is expressed as nmol of ROOH produced per mg of initial DHA. All reactions were carried out inthe presence of 1 mM EDTA. MSH5 human melanosomes at 1.953 107 gran/ml; MLFH 5 human melanolipofuscin granules at1.573 107gran/ml; LFH5 human lipofuscin granules at 1.723 107 gran/ml.


the dark LF granules remained reactive although MSgranules became relatively quiescent. MLF granuleswere less reactive overall than were the other two typesof pigment granules.

Dependence of photooxidizing activity of human andbovine melanosomes on irradiance

The amount of DHA hydroperoxides formed by pho-toactivated human and bovine melanosomes dependedon their irradiance by a visible light source. Expressed asnmol of hydroperoxide produced per mg DHA, bothhuman and bovine MS showed photooxidizing activityproportional to the output of the irradiating argon blue-green laser emission (Fig. 3). Bovine MS were somewhatmore reactive than human MS, but because the reactionextent also depended on granule density (see below), thisdifference may have been partly due to uncertainty in thegranule counts. We estimated the error in counting gran-ules at620%.

Dependence of photostimulating activity on granuledensity

Photostimulation of DHA hydroperoxide formationwas measured for granule densities up to;1010 gran/mlfor bovine MS (Fig. 4A), and up to;109 gran/ml forhuman MS (Fig. 4B). At lower densities, the photooxi-dation of DHA was negligible, and at higher densities,saturation and even a reduction in the reaction extent wasapparent. As noted above, some of the difference in theresults with human and bovine MS may be attributable toinaccuracies in granule counts. Nevertheless, the patternof increasing photooxidation of DHA with increasinggranule density up to a maximum density, followed by areduction of effect at high granule densities was similarfor both human and bovine granules.

Photooxidation of linoleic acid

Linoleic acid was also photooxidized by light-stimu-lated pigment granules, although the reaction extent waslower than with DHA. In the data shown in Fig. 5, theproduction of linoleic hydroperoxides is expressed as thedifference between the light-stimulated and dark (con-trol) reactions, normalized to a granule density of 107

gran/ml. Photostimulation was with the Argon laser blue-green emission. Human lipofuscin was the most reactivetoward LA, followed by human MS and bovine MSgranules, in that order. A limited supply of LF granulesprevented the examination of the dependence of pho-tooxidation of LA on LF granule density and irradiance,as was done with MS granules.

Effect of broadband light

All of the previous experiments were carried out withnarrowband or laser sources. The effect of broadband(white) light in photoactivating pigment granules wasalso determined. In general, broadband light, delivered ata similar irradiance as the laser emission, was much lesseffective in exciting the photooxidative reactions. In fact,when a long-pass filter with a cut-on wavelength of 420nm (Oriel No. 59480) was used in front of the xenon arc

Fig. 3. Dependence of DHA photooxidation by photoactivated mela-nosomes on sample irradiance. Samples were exposed for 10 min atvarious laser intensities. DHA peroxidation was proportional to sampleirradiance up to about 1 W/cm2. DHA peroxidation expressed as nmolof ROOH produced per mg of initial DHA. (A) Bovine melanosomesat a density of 9.53 107 granule/ml. (B) Human melanosomes at adensity of 2.53 108 granule/ml.


lamp, there was negligible photooxidation; thereforemost of the effects described here for the broadbandsource were likely due to UVA emissions from the lamp(i.e., the wavelengths not passed by the 420 nm longpassfilter; sample irradiance due to UVA was estimated atabout 20 mW/cm2 for this lamp). For these measure-ments, the TBARS assay was used to measure peroxida-tive breakdown products of DHA. There was stimulationof TBARS products when the pigment granules were

present at a density up to 109 gran/mg DHA for bovineMS (Fig. 6A) and up to about 108 gran/mg DHA forhuman MS (Fig. 6B).

Activity of melanosomes in the presence of aphotosensitizer

To determine the relative reactivity of MS granulescompared to exogenous photosensitizers known to pro-duce reactive oxygen species, the TBARS from DHAphotooxidation were determined in the presence of amixture of methionine and riboflavin illuminated by bluelight. For these experiments, the light source was theXenon arc lamp with the output filtered with a 400 nmbandpass filter (Oriel No. 59820). With the methionine-riboflavin photosensitizer present, this weak blue lightsource was able to produce an increase in TBARS com-pared to the nonirradiated control, indicating the produc-tion of reactive oxygen species, probably superoxideanion [31]. When bovine MS were present, however, theamount of TBARS produced decreased in an amountproportional to the density of the MS granules in thereaction mixture (Fig. 7). The effect of human MS in thisreaction was not determined because of insufficient ex-perimental material.

DISCUSSION

Photo-oxidizing activity of RPE pigment granules

The RPE of normal human and animal eyes containspigmented inclusions, that are related to three types of

Fig. 4. Dependence of DHA photooxidation on density of melano-somes. DHA reaction mixtures were prepared with varying densities ofbovine (A) or human (B) melanosomes and exposed to the argon laserat a sample irradiance of 800 mW/cm2 for 10 min. ROOH assayed byNADPH coupled to GSH peroxidase as described in text. DHA per-oxidation expressed as nmol of ROOH produced per mg of initialDHA; (inset in (B) plot of relative log(V) vs. log(granule density).Slope of best-fit line for human MS is 0.77 (r 5 .995) and for bovineMS is 0.70 (r 5 .995). See Discussion for more details.

Fig. 5. Photooxidation of linoleic acid by laser-stimulated RPE pigmentgranules. Reaction mixtures were made up with linoleic acid andlipofuscin or melanosomes and exposed to the Argon laser for 10 minat a sample irradiance of 1.58 W/cm2. Assay for ROOH as described intext. LFH5 human lipofuscin granules; MSH5 human melanosomes;MSB 5 bovine melanosomes.


pigment granules: melanosomes, melanolipofuscin, andlipofuscin. RPE melanosomes contain the black or brownscreening pigment melanin. The “age pigment,” lipofus-cin, is a heterogeneous aggregation of damaged mole-cules, rather than a genetically programmed, native prod-uct. Complex, or melanolipofuscin, granules arepresumably the concretion of melanin and lipofuscinpigments due to the effects of degenerative or aging

processes. Although it is generally thought that RPEmelanosomes protect ocular tissue against light damageby passively absorbing excess light [32,33] or inhibitingfree radical reactions [34], it is known that melanosomesare able to generate toxic superoxide radicals and hydro-gen peroxide under illumination [17,35,36] and to pro-mote photooxidation processes in vitro [7,8,37]. Accord-ing to some authors these processes may intensify thedamaging effect of illumination on RPE and neural retina[38,39], especially with blue light [40,41]. Nevertheless,it is not clear under what particular conditions the mela-nosomes are able to stimulate vs. inhibit free-radicalreactions.

Lipofuscin, a heterogeneous aggregation of damagedmolecules, accumulates in the RPE so that in elderlypersons, the amount of lipofuscin may exceed that ofmelanin by five to ten times [20]. There is also anage-related increase in the quantity of the hybrid com-pound, melanolipofuscin, contained in granules withinRPE cells. Lipofuscin consists of molecules so seriouslyaltered that they cannot be degraded by native enzymes[21,42,43]. Retinal pigment epithelium lipofuscin appar-ently represents the lifelong accumulation of lysosomalresidual bodies containing the end products of the phago-cytosis of photoreceptor outer segments [44–46]. Highlevels of lipofuscin are thought to exacerbate the agingprocess and have been associated with a number ofpathologies that include both ocular and systemic disor-

Fig. 6. Photo-oxidation of DHA by broadband light. Reaction mixtureswere prepared containing DHA and bovine (A) or human (B) melano-somes at the indicated densities, and exposed to the broadband (unfil-tered) output of a 150 W xenon arc lamp. Sample irradiance was 112mW/cm2. Data shown are the average of 10 and 20 min exposures.DHA hydroperoxides were assayed as TBARS. Results are shown asnmol of TBARS produced per mg initial DHA.

Fig. 7. Effect of increasing bovine melanosome density on DHAperoxidation by a photosensitized system of riboflavin-methionine.Photoactivation was with 400 nm light at a sample irradiance of 5.3mW/cm2. See text for additional experimental details. Peroxidation wasassayed as TBARS. Addition of bovine MS inhibited production ofTBARS.


ders. Indeed, in somecases of retinal degenerations, it hasbeen shown that the greatest lipofuscin accumulation in theRPE was associated with the greatest photoreceptor loss inthe overlying retina [47]. Lipofuscin granules are able tophotosensitize the generation of superoxide radicals [24]and singlet oxygen [23]. Although UV irradiation effec-tively excites free radical activity of LF in vitro [23,48],there is little information about in vivo conditions that causelipofuscin granules to stimulate free-radical oxidative reac-tions, namely intraocular wavelength and intensity of light,the critical concentration of lipofuscin, the effect of theconcentration of oxidizable substrates, oxygen saturation,etc. The exact mechanisms, too, by which lipofuscin mightproduce damage in situ are not well understood [25,26,48].

The work reported here has confirmed the previousfindings of the present investigators that all of the RPEpigments have photooxidizing activity [6–9,36,37,49], aswell as earlier reports of the photochemical activity ofmelanins [10,19,50–52]. The present work has extended theprevious findings by focusing on the photochemical activityof the RPE pigments towards fatty acids, especially doco-sahexaenoic acid, that in the retina is an essential fatty acidcomprising 25% or more of the polyunsaturated fatty acidsin the rod photoreceptors [53,54] and has been implicated innormal visual development and function [55,56].

A major finding reported here is that although all of theRPE pigment granules have photoinducible oxidizing ac-tivity towards the polyunsaturated fatty acids DHA and LA,there are differences in the dark activity levels of the pig-ments. Specifically, although RPE melanosomes are themost active in the photooxidation of DHA, in the dark thisreaction is essentially quiescent. Lipofuscin granules incontrast are less photoinducible than are MS granules, buthave a much higher dark reactivity. Melanolipofuscin(complex) granules have activity characteristics similar toLF granules, i.e., they exhibit some photoactivated activity,but also a relatively high dark reactivity. As shown in Fig.2C, the total reactivities of these pigments relative to eachother, i.e., the dark- plus light-activated peroxidation ofDHA, indicate that LF and light-activated MS granules areapproximately equally reactive, although MLF granules areless reactive by half. The effect of visible light in stimulat-ing the reactivity of MS granules is also clearly shown inFig. 2C, with the greatest increase in reactivity occurringover the physiologically relevant irradiance range from zero(dark) to 400 mW/cm2.

Effect of RPE pigment granule density and lightexposure parameters

The production of DHA hydroperoxides was stimulatedby increasing the sample irradiance by the laser as well asincreasing the density of pigment granules, at least up to asaturation point with MS. The saturation observed with

increased MS granule density may have been due to self-screening by the high optical density of the MS suspen-sions, and was similar to the saturation observed in a pre-vious report on the photooxidation of linoleic acid by laser-excited MS granules [8]. The lowest radiant exposure(irradiance3 time) used in these experiments to producedetectable peroxidation of DHA was about 120 J/cm2 (0.2W/cm2 for 10 min). For the 488 nm and 514.5 nm emis-sions of the Argon laser, this was visually a very brightlight, but was about 1/8th of the threshold for thermaldamage calculated as the equivalent retinal radiant exposureproduced by the incident corneal threshold limit allowed bythe ANSI laser safety standard [57]. Therefore it was un-likely fatty acid reactions observed in these experimentswere due to thermal processes.

Reaction mechanisms

The rate dependence of the production of DHA hy-droperoxides on the density of MS granules before thesaturation of the reaction (c.f., inset in Fig. 4B) was foundto have a slope close to one, suggesting that the reactionwas first order with respect to MS granules. Based on thisand on the assumption that molecular oxygen is present inexcess, it is likely that the reaction of photoexcited melaninproceeds according the mechanism(s) proposed by Roza-nowska et al. [10]. The probable primary reactive speciesare the quinone (MQ), semiquinone (MSQ•2), and hydro-quinone (MQH2) groups of the hydroxyindole subunits ofthe melanin heteropolymer. The effect of visible light is topush the equilibrium between the quinone, hydroquinone,and semiquinone species to the right so that the occurrenceof the semiquinone form is favored (Eq. 1):

MQ 1 MQH2 L|;hn

2 MSQ•2 1 2H1 (1)

This is presumably the only light-induced reaction. Thesemiquinones may react directly with DHA in a Type I(free radical) reaction or through a Type II reaction involv-ing an oxygen radical intermediate [58,59] as follows(Eq. 2):

MSQ•2 1 DHA-H 1 H1 º DHA•

1 MQH2 ~Type I! (2)

where DHA-H represents the polyunsaturated fatty acidand DHA• the alkyl radical, or (Eq. 3):

MSQ•2 1 O2 º MQ 1 O2•2 ~Type II! (3)

where the reaction of the semiquinone radical with oxy-gen produces superoxide anion. After this(ese) reac-


tion(s), hydroperoxides of the fatty acids may be pro-duced in the following reactions (Eq. 4):

DHA-H 1 HOO• 3 H2O2 1 DHA• (4)

where DHA• is formed by abstraction of a hydrogenatom by the perhydroxyl radical, HOO•. Although thisconjugate acid may only represent 1% of the total super-oxide at physiological pH, it is relatively efficient atproducing fatty acid hydroperoxides, especially if hy-droperoxides are already present [60]. Interaction of thealkyl radical with oxygen forms a peroxyl radical (Eq. 5):

DHA• 1 O2 3 DHA-OO• (5)

A chain reaction between the peroxyl radical and thePUFA ensues, producing the hydroperoxide products(Eq. 6):

DHA-OO• 1 DHA-H 3 DHA-OOH 1 DHA• (6)

The dienes produced by these processes can be con-jugated or non-conjugated, depending on whether a TypeI or Type II reaction occurred [61], but our methodologydid not include such an analysis. Based on energetic andrate considerations, it is likely that a majority of thehydroperoxides were produced by a Type II reaction.

Interaction of melanin with a photosensitizer

When superoxide was generated by the photosensi-tized system of riboflavin-methionine, we observed thatthe peroxidation of DHA was diminished by the presenceof melanin. This result could have been due to thescreening of the riboflavin from photic stimulation. An-other possibility is that the superoxide generated by thephotosensitizer reacted with the melanosomes instead ofthe DHA. That the reaction of superoxide with melanin isfavored is indicated by the inhibition of nitroblue tetra-zolium reduction by superoxide in the presence of mel-anin [4], and also by its high rate constant (.104 M21-s21) [12,62]. Melanins generally exist in an oxidizedform [63], probably because of interaction with oxygen.Generally, the oxidized forms of melanin (the quinoneand semiquinone forms) react with superoxide as shownin Eqs. 7 and 8:

MQ 1 O2•2 º MSQ•2 1 O2 (7)

MSQ•2 1 O2•2 1 2 H1 º MQH2 1 O2 (8)

If a large excess of O2•2 is produced, for example, in

the photosensitized riboflavin-methionine experiment,then reactions (Eq. 7) and (Eq. 8) will be pushed to theright, consuming the superoxide in the process and re-

ducing melanin. The reaction of melanin hydroquinoneswith continuously produced superoxide may lead to re-generation of the melanin semiquinone radical, as pro-posed by Jarabak et al. [64] for the redox cycling ofo-quinones in the presence of superoxide. In addition, ifthe oxidation potential of eumelanin is similar to that ofsynthetic dopa-melanin, determined by cyclic voltametryto be less than1100 mV [65], then the redox cycling ofmelanin would be thermodynamically favored over oxi-dation of PUFA, that typically occurs at an oxidationpotential of 1600 mV [66]. Nevertheless, the precisemechanism of the interaction of melanin and exogenousphotosensitizers remains to be determined.

Reactivity of LF granules in the dark

Several recent studies have indicated that human LFgranules contain free radical and other transient speciesinducible by UVA exposure that produce reactive oxy-gen species and induce lipid peroxidation [23,25,48]. Wenow add the finding that even in the dark, human LFgranules are reactive towards polyunsaturated fatty acids.Although one component of human LF has been chem-ically characterized [67], LF granules isolated from RPEcells are not homogeneous and contain several other,uncharacterized components [21,23]. Their dark reactiv-ity evidently arises from other than a photochemicalprocess, possibly involving reactions catalyzed by labilehydroperoxides accumulated during the degradation pro-cess that gives rise to the LF granules. These labilehydroperoxides readily undergo thermally-dependent,homolytic decomposition to free radical products, thatcatalyze in the dark the DHA peroxidation. Anotherpossibility is that metal ions associated with the LFgranules catalyze oxidative reactions. Metal ions, partic-ularly Fe12, may be involved in the lysosomal produc-tion of LF precursors (reviewed by Yin [21]), and areprobably complexed in the LF granules. Although all ofthe reactions described in this study were performed inthe presence of 1 mM EDTA, it is possible that metalions not accessible to the chelator, e.g. within the granuleinterior, or not effectively chelated by EDTA [68], maysupport the dark reactions with DHA described in thepresent work. In any event, the constant reactivity of LFgranules, if it occurs in vivo, would be a major factorcontributing to oxidative stress in the RPE.

Implications for origins of oxidative stress in theretina-RPE complex

Currently, we can only speculate on what the signif-icance of the reactions, observed with isolated RPE pig-ment granules, is for the generation of oxidative stress in


the RPE and choroid. If oxidative reactions with poly-unsaturated fatty acids were initiated by photoactivatedRPE melanin and lipofuscin granules, then damage couldoccur to the RPE cell membranes, and possibly to thephotoreceptors with their large content of membranousdisks in the outer segments. The relatively high oxygentension in the retina and choroid is likely to amplify therisk of lipid peroxidation in these tissues.

The densities of pigment granules used in these ex-periments are consistent with the physiological rangefound in the RPE. Based on the cellular pigment granulecontent reported for the RPE of the human eye [20], andusing 10 to 20mm as the range of values for the humanRPE cell diameter [69] (with calculated cellular volumefrom 0.5 3 1029 to 4 3 1029 ml), estimates of maxi-mum pigment granule densities in RPE cells may bemade. For MS the density range is 6.33 109 to 53 1010

gran/ml; for LF it is 8.83 109 to 7 3 1010 gran/ml; andfor MLF it is 4 3 109 to 3 3 1010 gran/ml. Our exper-iments utilized granule densities up to about 1011 gran/mlfor MS, and over 107 gran/ml for LF and MLF granules.Therefore, our experimental conditions overlap the phys-iological situation to a considerable extent, and if thephotoactivated reactions reported here also occur in vivo,then photooxidative damage to cellular componentswould be likely. Moreover, although lipofuscin granulesare not nearly as photoexcitable as are the melanosomes,we have demonstrated that they have higher dark, orbasal, reactivity towards fatty acids. The age-related in-crease of lipofuscin content in the RPE makes it likelythat there is a progressive increase in oxidative stress inthe RPE associated with this accumulation. Unless thereis a corresponding increase in the antioxidant capacity ofthe cell, or sequestration of the LF granules, accumula-tion of LF may be a specific factor contributing toage-related retinal degenerations.

Acknowledgements— Dr. A. E. Dontsov was a Research to PreventBlindness International Scholar at the UTHSCSA during the perfor-mance of this research. Additional research support was provided byAFOSR grant F49629-95-1-0332 (to R.D.G.), grant No. 96-04-49819from the Russian Foundation for Basic Research (to MAO), the SanAntonio Area Foundation, the Helen Freeborn Kerr Foundation, anEnrichment Subgrant from the Howard Hughes Medical Institute Re-search Resources Program grant to the UTHSCSA, and an unrestrictedgrant from Research to Prevent Blindness (RPB) to the Department ofOphthalmology of the UTHSCSA. We thank Ms. Neeru Kumar andMr. Steven Stubblefield for technical assistance during this project.

REFERENCES

[1] Koch, W. H.; Chedekel, M. R. Photochemistry and photobiologyof melanogenic metabolites: formation of free radicals.Photo-chem. Photobiol.46:229–238; 1987.

[2] Wolbarsht, M. L.; Walsh, A. W.; George, G. Melanin, a uniquebiological absorber.Appl. Opt.20:2184–2186; 1981.

[3] Sarna, T. Properties and function of the ocular melanin—a pho-tobiophysical view.J. Photochem. Photobiol. B12:215–258;1992.

[4] Porebska-Budny, M.; Sakina, N. L.; Stepien, K. B.; Dontsov, A. E.;Wilczok, T. Antioxidative activity of synthetic melanins. Cardiolipinliposome model.Biochim. Biophys. Acta1116:11–16; 1992.

[5] Sakina, N. L.; Dontsov, A. E.; Afanas’ev, G. G.; Ostrovski,M. A.; Pelevina, I. I. The accumulation of lipid peroxidationproducts in the eye structures of mice under whole-body x-rayirradiation. [Russian]Radiobiologiia30:28–31; 1990.

[6] Glickman, R. D.; Lam, K.-W. Oxidation of ascorbic acid as anindicator of photooxidative stress in the eye.Photochem. Photo-biol. 55:191–196; 1992.

[7] Glickman, R. D.; Sowell, R.; Lam, K.-W. Kinetic properties oflight-dependent ascorbic acid oxidation by melanin.Free Radic.Biol. Med.15:453–457; 1993.

[8] Glickman, R. D.; Lam, K.-W. Melanin may promote photooxidationof linoleic acid. In: Jacques, S. L., ed.Laser-tissue interaction VI,Proc. SPIE. Vol. 2391. Bellingham, WA: SPIE; 1995:254–261.

[9] Glickman, R. D.; Rockwell, B. A.; Jacques, S. L. Action spectrumof oxidative reactions mediated by light-activated melanin. In:Jacques, S. L., ed.Laser-tissue interaction VIII, Proc. SPIE. Vol.2975. Bellingham, WA: SPIE; 1997:138–145.

[10] Rozanowska, M.; Bober, A.; Burke, J. M.; Sarna, T. The role ofretinal pigment epithelium melanin in photoinduced oxidation ofascorbate.Photochem. Photobiol.65:472–479; 1997.

[11] McGinness, J.; Proctor, P. The importance of the fact that melaninis black.J. Theoret. Biol.39:677–678; 1973.

[12] Sarna, T.; Pilas, B.; Land, E. J.; Truscott, T. G. Interaction ofradicals from water radiolysis with melanin.Biochim. Biophys.Acta 833:162–167; 1986.

[13] Boettner, E. A.; Wolter, J. R. Transmission of the ocular media.Invest. Ophthalmol. Vis. Sci.1:776–783; 1962.

[14] Mason, H. S.; Ingram, D. J. E.; Allen, B. The free radical propertyof melanins.Arch. Biochem. Biophys.86:225–230; 1960.

[15] Cope, F. W.; Sever, R. J.; Polis, B. D. Reversible free radicalgeneration in the melanin granules of the eye by visible light.Arch. Biochem.100:171–177; 1963.

[16] Sarna, T.; Sealy, R. C. Free radicals from eumelanin: quantumyields and wavelength dependence.Arch. Biochem. Biophys.232:574–578; 1984.

[17] Korytowski, W.; Pilas, B.; Sarna, T.; Kalyanaraman, B. Photoin-duced generation of hydrogen peroxide and hydroxyl radicals inmelanins.Photochem. Photobiol.45:185–190; 1987.

[18] van Woert, M. H. Reduced nicotinamide-adenine oxidation bymelanin: inhibition by phenothiazines.Proc. Soc. Exp. Biol. Med.129:165–171; 1968.

[19] Menon, I. A.; Haberman, H. F. Mechanisms of action of melaninin photosensitized reactions. In: Klaus, S. N., ed.Pigment cell.Basel: Karger; 1979:345–351.

[20] Feeney-Burns, L.; Hilderbrand, E. S.; Eldridge, S. Aging humanRPE: morphometric analysis of macular, equatorial and peripheralcells. Invest. Ophthalmol. Vis. Sci.25:195–200; 1984.

[21] Yin, D. Biochemical basis of lipofuscin, ceroid, and age pigment-like fluorophores.Free Radic. Biol. Med.21:871–888; 1996.

[22] Weiter, J. J.; Delori, F. C.; Wing, G. L.; Fitch, K. A. Retinalpigment epithelial lipofuscin and melanin and choroidal melaninin human eyes.Invest. Ophthalmol. Vis. Sci.27:145–152; 1986.

[23] Gaillard, E. R.; Atherton, S. J.; Eldred, G.; Dillon, J. Photophysi-cal studies on human retinal lipofuscin.Photochem. Photobiol.61:448–453; 1995.

[24] Boulton, M.; Dontsov, A.; Jarvis-Evans, J.; Ostrovsky, M.; Svis-tunenko, D. Lipofuscin is a photoinducible free radical generator.J. Photochem. Photobiol. B19:201–204; 1993.

[25] Rozanowska, M.; Jarvis-Evans, J.; Korytowski, W.; Boulton,M. E.; Burke, J. M.; Sarna, T. Blue light-induced reactivity ofretinal age pigment. In vitro generation of oxygen-reactive spe-cies.J. Biol. Chem.270:18825–18830; 1995.

[26] Wihlmark, U.; Wrigstad, A.; Roberg, K.; Nilsson, S. E.; Brunk,U. T. Lipofuscin accumulation in cultured retinal pigment epithe-lial cells causes enhanced sensitivity to blue light irradiation.FreeRadic. Biol. Med.22:1229–1234; 1997.

[27] Boulton, M.; Marshall, J. Re-pigmentation of human retinal pig-ment epithelial cells in vitro.Exp. Eye Res.41:209–218; 1985.


[28] Organisciak, D. T.; Noell, W. K. Hereditary retinal dystrophy inthe rat: lipid composition of debris.Exp. Eye Res.22:101–113;1976.

[29] Buege, J. A.; Aust, S. D. Microsomal lipid peroxidation.MethodsEnzymol.52:302–310; 1978.

[30] Ottolenghi, A. Interaction of ascorbic acid and mitochondriallipids. Arch. Biochem. Biophys.79:355–363; 1959.

[31] Beauchamp, C.; Fridovich, I. Superoxide dismutase: improvedassays and an assay applicable to acrylamide gels.Anal. Biochem.44:276–287; 1971.

[32] Walls, G. L. The vertebrate eye and its adaptive radiation.Bloomfield Hills, MI: The Cranbrook Institute of Science; 1942.

[33] Young, R. W. Solar radiation and age-related macular degenera-tion. Surv. Ophthalmol.32:252–269; 1988.

[34] Ostrovsky, M. A.; Sakina, N. L.; Dontsov, A. E. An antioxidativerole of ocular screening pigments.Vision Res.27:893–899; 1987.

[35] Felix, C. C.; Hyde, J. S.; Sarna, T.; Sealy, R. C. Melanin photo-reactions in aerated media: electron spin resonance evidence forproduction of superoxide and hydrogen peroxide.Biochem. Bio-phys. Res. Commun.84:335–341; 1978.

[36] Lapina, V. A.; Dontsov, A. E.; Ostrovsky, M. A. Superoxidegeneration via melanin interaction with oxygen. [Russian]Bio-khimiia 50:1712–1718; 1985.

[37] Glickman, R. D.; Jacques, S. L.; Schwartz, J. A.; Rodriguez, T.;Lam, K.-W.; Buhr, G. Photodisruption increases the free radicalreactivity of melanosomes isolated from retinal pigment epithe-lium. In: Jacques, S. L., ed.Laser-tissue interaction VII, Proc.SPIE. Vol. 2681. Bellingham, WA: SPIE; 1996:460–467.

[38] Ham, W. T., Jr.; Ruffolo, J. J., Jr.; Mueller, H. A.; Clarke, A. M.;Moon, M. E. Histologic analysis of photochemical lesions pro-duced in rhesus retina by short-wave-length light.Invest. Oph-thalmol. Vis. Sci.17:1029–1035; 1978.

[39] Ham, W. T., Jr.; Ruffolo, J. J., Jr.; Mueller, H. A.; Guerry, D., III.The nature of retinal radiation damage: dependence on wave-length, power level and exposure time.Vision Res.20:1105–1111; 1980.

[40] Pautler, E. L.; Morita, M.; Beezley, D. Hemoprotein(s) mediateblue light damage in the retinal pigment epithelium.Photochem.Photobiol.51:599–605; 1990.

[41] Dorey, C. K.; Delori, F. C.; Akeo, K. Growth of cultured RPE andendothelial cells is inhibited by blue light but not green or redlight. Curr. Eye Res.9:549–559; 1990.

[42] Taubold, R. D. Studies on chemical nature of lipofusion (agepigment) isolated from normal human brain.Lipids 10:383–390;1975.

[43] Eldred, G. E.; Miller, G. V.; Stark, W. S.; Feeney-Burns, L.Lipofuscin: resolution of discrepant fluorescence data.Science216:757–759; 1982.

[44] Feeney, L. Lipofuscin and melanin of human retinal pigmentepithelium. Fluorescence, enzyme cytochemical, and ultrastruc-tural studies.Invest. Ophthalmol. Vis. Sci.17:583–600; 1978.

[45] Katz, M. L.; Drea, C. M.; Eldred, G. E.; Hess, H. H.; Robison,W. G., Jr. Influence of early photoreceptor degeneration on lipo-fuscin in the retinal pigment epithelium.Exp. Eye Res.43:561–573; 1986.

[46] Boulton, M.; McKechnie, N. M.; Breda, J.; Bayly, M.; Marshall,J. The formation of autofluorescent granules in cultured humanRPE.Invest. Ophthalmol. Vis. Sci.30:82–89; 1989.

[47] Dorey, C. K.; Wu, G.; Ebenstein, D.; Garsd, A.; Weiter, J. J. Cellloss in the aging retina. Relationship to lipofuscin accumulationand macular degeneration.Invest. Ophthalmol. Vis. Sci.30:1691–1699; 1989.

[48] Reszka, K.; Eldred, G. E.; Wang, R. H.; Chignell, C.; Dillon, J.The photochemistry of human retinal lipofuscin as studied byEPR.Photochem. Photobiol.62:1005–1008; 1995.

[49] Lam, K.-W.; Glickman, R. D. Prevention of light-induced freeradical production from melanin granules by ascorbic acid. In:Yagi, K.; Niki, E.; Yoshikawa, T., eds.Oxygen radicals. Amster-dam: Elsevier Science Publishers; 1992:633–636.

[50] van Woert, M. H. Oxidation of reduced nicotinamide adeninedinucleotide by melanin.Life Sci.6:2605–2612; 1967.

[51] Sarna, T.; Menon, I. A.; Sealy, R. C. Photosensitization of mel-anins: a comparative study.Photochem. Photobiol.42:529–532;1985.

[52] Gan, E. V.; Haberman, H. F.; Menon, I. A. Oxidation of NADHby melanin and melanoproteins.Biochim. Biophys. Acta370:62–69; 1974.

[53] Stephens, R. J.; Negi, D. S.; Short, S. M.; van Kuijk, F. J. G. M.;Dratz, E. A.; Thomas, D. W. Lipid peroxidation and retinalphototoxic degeneration.Basic Life Sci.49:283–289; 1988.

[54] van Kuijk, F. J. G. M.; Buck, P. Fatty acid composition of thehuman macula and peripheral retina.Invest. Ophthalmol. Vis. Sci.33:3493–3496; 1992.

[55] Neuringer, M.; Connor, W. E.; Van Petten, C.; Barstad, L. Dietaryomega-3 fatty acid deficiency and visual loss in infant rhesusmonkeys.J. Clin. Invest.73:272–276; 1984.

[56] Makrides, M.; Neumann, M.; Simmer, K.; Pater, J.; Gibson, R.Are long-chain polyunsaturated fatty acids essential nutrients ininfancy?Lancet345:1463–1468; 1995.

[57] American National Standards Institute. American National Standardfor Safe Use of Lasers. ANSI Z136.1-1993, American NationalStandards Institute, 11 W. 42nd St., New York, NY 10036; 1993.

[58] Girotti, A. W. Photodynamic lipid peroxidation in biologicalsystems.Photochem. Photobiol.51:497–509; 1990.

[59] Holte, L. L.; van Kuijk, F. J. G. M.; Dratz, E. A. Preparativehigh-performance liquid chromatography purification of polyun-saturated phospholipids and characterization using ultraviolet de-rivative spectroscopy.Anal. Biochem.188:136–141; 1990.

[60] Aikens, J.; Dix, T. A. Perhydroxyl radical (HOO•) initiated lipidperoxidation.J. Biol. Chem.266:15091–15098; 1991.

[61] van Kuijk, F. J. G. M.; Thomas, D. W.; Stephens, R. J.; Dratz,E. A. Gas chromatography-mass spectrometry method for deter-mination of phospholipid peroxides; I. Transesterification to formmethyl esters.J. Free Radic. Biol. Med.1:215–225; 1985.

[62] Lapina, V. A.; Dontsov, A. E.; Ostrovsky, M. A.; Emanuel, N. M.Interaction of superoxide radicals with eye melanin and ommo-chrome. [Russian]Dokl. Akad. Nauk. SSSR280:1463–1465; 1985.

[63] Horak, V.; Gillette, J. R. A study of the oxidation-reduction stateof synthetic 3,4-dihydroxy-DL-phenylalanine melanin.Mol. Phar-macol.7:429–433; 1971.

[64] Jarabak, R.; Harvey, R. G.; Jarabak, J. Redox cycling of polycy-clic aromatic hydrocarbon o-quinones: reversal of superoxidedismutase inhibition by ascorbate.Arch. Biochem. Biophys.339:92–98; 1997.

[65] Jacobson, E. S.; Hong, J. D. Redox buffering by melanin andFe(II) in Cryptococcus neoformans.J. Bacteriol.179:5340–5346;1997.

[66] Koppenol, W. H. Oxyradical reactions: from bond-dissociationenergies to reduction potentials.FEBS Lett.264:165–167; 1990.

[67] Eldred, G. E.; Lasky, M. R. Retinal age pigments generated byself-assembling lysosomotropic detergents.Nature361:724–726;1993.

[68] Buettner, G. R. Ascorbate autoxidation in the presence of iron andcopper chelates.Free Radic. Res. Comm.1:349–353; 1986.

[69] Feeney-Burns, L.; Burns, R. P.; Gao, C.-L. Age-related macularchanges in humans over 90 years old.Am. J. Ophthalmol.109:265–278; 1990.

ABBREVIATIONS

DHA—decosahexaenoic acidLA—linoleic acidLF—lipofuscin granulesMLF—melanolipofuscin (complex) granulesMS—melanosomes of human (H) or bovine (B) originPUFA—polyunsaturated unsaturated fatty acidRPE—retinal pigment epitheliumTBARS—thiobarbituric acid reactive substances


retinal pigment epithelium pigment granules stimulate the photo-oxidation of unsaturated fatty acids

Documents