Oxidation Causes Melanin Fluorescence
Peter Kayatz, Gabriele Thumann, Thomas T. Luther, Jens F. Jordan,
Karl Ulrich Bartz–Schmidt, Peter J. Esser, and Ulrich Schraermeyer
PURPOSE. The goal of this study is the characterization of the
strong yellow fluorescence of oxidized melanin in the retinal
pigment epithelium (RPE) and the choroid.
METHODS. Naturally occurring melanin in the human retina and
choroid was oxidized by exposing fixed and plastic-embedded
sections of a human eye to light and hydrogen peroxide.
Synthetic melanin was also oxidized in vitro by exposure to
light and hydrogen peroxide. The fluorescence of oxidized
melanin was examined by absorption spectroscopy, fluores-
cence spectroscopy, and fluorescence microscopy.
RESULTS. Naturally occurring melanin oxidized in situ exhibited
a lipofuscin-like yellow fluorescence. Oxidation of melanin in
vitro degraded the melanin polymer, resulting in a fluorescent
solution. Fluorescence spectroscopy gave an excitation maxi-
mum at approximately 470 nm and an emission maximum at
approximately 540 nm for both natural and synthetic melanin.
Increasing the time of exposure to light or hydrogen peroxide
increased melanin fluorescence.
CONCLUSIONS. The results indicate that the strong yellow fluo-
rescence of melanin in the RPE and choroid in situ is a property
of oxidized melanin and is not due to contamination of the
melanin by proteinaceous or lipid materials. The data pre-
sented allow a reinterpretation of the results obtained from
fluorescence investigations of melanin-containing tissue and
suggest a link between melanin degradation and lipofuscin
formation. (Invest Ophthalmol Vis Sci. 2001;42:241–246)
melanin and lipofuscin.1Lipofuscin is a highly fluorescent
mixture of different fluorophores, which reflects the heteroge-
neous nature of this cellular pigment.2,3It is derived, at least in
part, from product of phagocytized and degraded photorecep-
tor outer segments (POSs).4Emphasis has been put on the
fluorescence properties of lipofuscin as a means to identify
these granules in situ and to understand the mechanisms of
lipofuscin formation.5Particularly, autofluorescence is used to
distinguish lipofuscin from melanin in situ.6
In contrast to lipofuscin autofluorescence, only a very faint
fluorescent signal has been thoroughly investigated7and dif-
ferentially characterized for melanin isolated from human RPE
from donors of different ages.5Additionally, statements can be
found in published reports that even deny melanin fluores-
cence.8The large increase of melanin fluorescence after oxi-
dative conditions is generally ignored in ophthalmic research.
he retinal pigment epithelium (RPE) of normal adult hu-
man eyes contains two populations of pigment granules:
Because in clinical ophthalmology retinal fluorescence, in par-
ticular the autofluorescence associated with the RPE and
Bruch’s membrane, is used for diagnostic purposes, it is essen-
tial to understand whether natural and/or artificial conditions
can lead to the development of fluorescence. In this study we
characterized the fluorescence of melanin in situ after oxida-
tive conditions were established and compared it with the
fluorescence of synthetic melanin subjected to the same oxi-
Fluorescence of Oxidized Melanin
Embedded in Spurr’s Resin
An eye of a 52-year-old healthy male human donor was fixed in
glutaraldehyde, as previously described.9After fixation, the vitreous
body was removed, and specimens from the macular area were ex-
cised, washed in 0.1 M cacodylate buffer (pH 7.4) dehydrated in a
graded series of ethanol and embedded in Spurr’s resin.
Additionally, two types of melanin were embedded directly into
Spurr’s resin: synthetic melanin prepared by oxidation of tyrosine with
hydrogen peroxide purchased from Sigma (Deisenhofen, Germany)
and isolated bovine melanosomes from RPE and choroid, prepared as
previously described10(referred to as bovine melanosomes).
Samples of the two types of melanin were embedded into Spurr’s
resin without previous incubation with glutaraldehyde to test the
influence of glutaraldehyde incubation, the presence of associated
proteins and lipids in natural melanin, and the cytoplasmic environ-
ment of the tissue on the fluorescence properties of the oxidized
Semithin sections (approximately 700 nm) were prepared without
staining. Oxidation of semithin sections of specimens embedded in
Spurr’s resin was performed using a solution of hydrogen peroxide
dissolved in polypropylene glycol-2000 (Fluka, Buchs, Switzerland), a
minimally hydrophilic medium, so that only traces of hydrogen perox-
ide were dissolved in the medium. A volume of 500 ?l polypropylene
glycol-2000 was mixed thoroughly with 100 ?l of 30% hydrogen
peroxide centrifuged to separate the phases, and the aqueous phase
was discarded. Semithin sections were placed on a microscope slide, 5
to 10 ?l hydrogen peroxide dissolved in polypropylene glycol was
added to the section, and the slide was coverslipped. Fluorescence
observations were begun immediately (before any significant oxidation
had occurred) and continued for 8 minutes under a fluorescence
microscope (400-nm excitation, 520-nm barrier; Axioplan; Zeiss,
Oberkochen, Germany) at a magnification of ?63 in combination with
a microscope camera (MC 100; Zeiss). After 8 minutes, the microscope
was switched to bright-field illumination, the illuminated field dia-
phragm was opened maximally, and maximum brightness was
achieved by selecting 12 V (100-W halogen lamp), resulting in a
fluorescent illumination of approximately 500,000 lux, as measured
with a photometer (Colormaster 3F; Gossen, Erlangen, Germany). The
illumination of the specimens was continued for an additional 8 min-
utes after oxidation and repeatedly examined by fluorescence micros-
Fluorescence of In Vitro Oxidized Melanin
Synthetic melanin (2.5 mg/ml) and of isolated bovine melanosomes
(2.5 mg/ml) were incubated with the organic oxidizing medium de-
From the Department of Vitreoretinal Surgery, Division of Oph-
thalmology, University of Cologne, Germany.
Supported in part by the Deutsche Forschungsgemeinschaft He/
840/5-3, Es 82/5-3, Th 603/4-1, the Retinovit Foundation, the Propter
Homines Foundation, and the Koeln Fortune Program.
Submitted for publication March 21, 2000; revised July 25, 2000;
accepted August 22, 2000.
Commercial relationships policy: N.
Corresponding author: Ulrich Schraermeyer, University of Co-
logne, Ophthalmology, Department of Vitreoretinal Surgery, Joseph-
Stelzmann-Strasse 9, D-50931 Cologne, Germany.
Investigative Ophthalmology & Visual Science, January 2001, Vol. 42, No. 1
Copyright © Association for Research in Vision and Ophthalmology
scribed by Frangioni and Borgioli,11which contains benzyl alcohol (20
ml), acetone (10 ml), hydrogen peroxide 10% (5 ml), and ammonia
solution 25% (4 drops). Oxidation was allowed to proceed overnight at
room temperature. The oxidized melanin formed an emulsion that
separated into a hydrophilic phase. The oxidized melanin was exam-
ined by phase-contrast microscopy and by fluorescence microscopy
(400-nm excitation, 520-nm barrier) using the fluorescence micro-
scope (Axioplan; Zeiss) with a magnification of ?63.
For complete oxidation, 2.5 mg/ml of synthetic melanin and 2.5
mg/ml of isolated bovine melanosomes were incubated with a 30%
aqueous hydrogen peroxide solution for 4 hours.
Absorption Spectroscopy. Absorption spectroscopy (300–
800 nm) of the resultant yellow solution was performed using a
spectrophotometer (DU-6; Beckman, Irvine, CA) at a scanning speed of
150 nm/min against air. The absorption spectrum of the oxidized
melanin was compared with the spectrum of a 2.5 mg/ml solution of
non-oxidized synthetic melanin in water.
Fluorescence Spectroscopy. Fluorescence spectroscopy of
the oxidized melanin was performed using a spectrophotometer (F-
2000; Hitachi, San Jose, CA), at a scanning speed of 240 nm/min, a
bandpass of 10 nm, a photomultiplier (PM) voltage of 400 V and a list
interval of 5 nm. The scanning data were corrected for the Raman
signal of water. Excitation and emission maxima were determined by
the spectrophotometer’s autoscan function. The excitation wave-
length for the emission spectrum was 470 nm, and the emission
wavelength for the excitation spectra was 540 nm.
Time-Response Study of Fluorescence
To analyze the large increase in fluorescence by the oxidized melanin,
2.5 mg/ml synthetic melanin was incubated with a 30% aqueous
solution of hydrogen peroxide and sonicated for rapid dissolution.
Aliquots of 1.5 ml were added to each of six petri dishes (diameter, 3.5
cm) and the increase in fluorescence (450-nm excitation, 580-nm
emission) was measured using a fluorescence multiwell plate reader
(CytoFluor Series 4000; PerSeptive Biosystems–Applied Biosystems,
Foster City, CA) for up to 90 minutes. A solution of 30% hydrogen
peroxide was used as control. The first time point measured was at 2.5
minutes from the beginning of the oxidation process. Fluorescence of
nonoxidized melanin was measured against a background of water.
Fluorescence of Oxidized Melanin Embedded in
Human Ocular Tissue In Situ. Untreated specimens from
the macular area of a human donor eye showed an accumula-
tion of lipofuscin in the RPE, identified by its strong autofluo-
rescence, whereas the choroid was free of lipofuscin-like
autofluorescence (Fig. 1B). Nonoxidized melanin (Fig. 1A) did
not fluoresce before oxidation (Fig. 1B); however, after 8
minutes of oxidation by hydrogen peroxide, melanin granules
in the RPE and choroid exhibited a strong fluorescence remi-
niscent of lipofuscin in apparent color and intensity (Fig. 1D).
Double exposure of phase contrast and fluorescence allowed
the simultaneous localization of lipofuscin and melanin gran-
ules (Fig. 1C). Comparison of Figures 1C and 1D reveals that
the melanin granules (arrows) were nonfluorescent (or only
slightly) at the beginning of the oxidation process (Fig. 1C) but
became strongly fluorescent after 8 minutes of oxidation (Fig.
1D). In contrast to melanin, lipofuscin autofluorescence de-
creased after treatment of the specimen with hydrogen perox-
ide and/or light (Figs. 1B, 1D).
Synthetic and Isolated Bovine Melanosomes Embed-
ded in Spurr’s Resin. Isolated bovine melanosomes and syn-
thetic melanin were embedded into Spurr’s resin to confirm
the hypothesis that the induced fluorescence is a property of
the oxidized melanin and is independent of the fixation pro-
cedure, the environment of the tissue, and the associated
proteins and lipids.
Isolated bovine melanosomes exhibited a strong yellow
fluorescence after oxidation (Fig. 2A). Oxidized synthetic mel-
anin, which does not contain proteins or lipids, exhibited a
fluorescence similar to that of isolated bovine melanosomes
(Fig. 2B). Nonoxidized isolated melanosomes and nonoxidized
synthetic melanin showed no fluorescence.
fuscin-like fluorescence after oxida-
tive stress in RPE and choroid of the
human macula. (A) Phase contrast-mi-
crograph of POSs, RPE, and choroid of
a human macula demonstrates the
presence of melanin granules (visible
as dark grains) both in RPE and cho-
roid. (B) Fluorescence micrograph
(400 nm excitation, 520 nm barrier)
of the same section at the beginning
of the oxidation reveals the intense
autofluorescence of lipofuscin gran-
ules in the RPE. Melanin granules are
not visible in the fluorescence micro-
graph. (C) Double exposure of phase-
contrast and fluorescence microscopy
allowed the simultaneous localization
of lipofuscin granules (yellow) and
melanin granules (black). A slight flu-
orescence of melanin in RPE and cho-
roid resulted from initial oxidation.
(D) After 8 minutes of oxidation, mel-
anin granules in the choroid displayed
an intense lipofuscin-like fluorescence.
Melanin exhibited a lipo-
A comparison of (C) and (D) reveals that the melanin granules (arrows) were nonfluorescent (or only slightly) at the beginning of the oxidation
(C) and became intensely fluorescent after 8 minutes of oxidation (D). Autofluorescence in the basal part of the RPE of (D) resulted from lipofuscin,
whereas autofluorescence in the apical part of the RPE was due to the oxidized melanin (compare C and D). Autofluorescence of lipofuscin and
oxidized melanin was almost indistinguishable in color and intensity (D). Lipofuscin autofluorescence decreased slightly with illumination
(compare B and D). All micrographs show the same area of the same section. Magnification, ?630.
242Kayatz et al.
IOVS, January 2001, Vol. 42, No. 1
Fluorescence of In Vitro Oxidized Melanin
To exclude the possibility that embedding in Spurr’s resin
causes melanin fluorescence after oxidation, synthetic and
isolated bovine melanosomes were oxidized in vitro and ex-
amined for fluorescence.
Oxidation of both synthetic and isolated bovine melano-
somes by 30% hydrogen peroxide (according to Frangioni and
Borgioli11) resulted in a yellow-ocher hydrophilic liquid. The
solid black oxidized melanin liquefied almost completely on
oxidation, creating an aqueous phase underneath the organic
phase of the oxidizing medium (data not shown).
Oxidized synthetic melanin appeared in phase-contrast mi-
croscopy as a clear aqueous liquid with no apparent impurities
(Fig. 3A). However, oxidized isolated bovine melanosomes
formed a heterogenous mixture of oxidized liquid melanin and
solid residual particles (Fig. 3C).
By fluorescence microscopy nonfixed, nonembedded, oxi-
dized synthetic melanin or isolated melanosomes fluorescence
was green (Figs. 3B, 3D), whereas the fluorescence of oxidized
melanin in situ (Fig. 1) and embedded directly into Spurr’s
resin was yellow (Fig. 2). The impurities of the oxidized bovine
melanosomes did not appear to be fluorescent or were only
weakly fluorescent (Fig. 3D).
Absorption Spectroscopy. The absorbance spectrum of
nonoxidized melanin is characterized by uniform absorbance
from 800 to 350 nm, followed by a steep increase at 350 nm.
In contrast, the absorbance of the yellow solution of oxidized
melanin was also uniform but lower than in nonoxidized mel-
anin from 800 to 600 nm than increase, reached the level of
nonoxidized melanin at 425 nm, and remained at this level
until the steep increase in absorbance at 350 nm that was
observed for nonoxidized melanin (Fig. 4).
Fluorescence Spectroscopy. Examination of the oxidized
melanin species by fluorescence spectroscopy resulted in an
excitation peak at 471 nm for synthetic melanin and at 469 for
bovine melanosomes. The emission peak of synthetic melanin
was at 548 nm and for bovine melanosomes at 543 nm (Fig. 5).
Excitation and emission spectra of oxidized synthetic and bo-
vine melanosomes were similar, except for the presence of
two additional excitation peaks for oxidized bovine melano-
somes at 400 and 420 nm, forming a broad excitation plateau
Time Response of Melanin Fluorescence
Fluorescence as a function of oxidation time was examined
during a period of 90 minutes. Figure 6 shows a first-order
linear relationship between time of oxidation and fluorescence
reaching a maximum at 20 minutes. Because of the technical
limitation of the procedure, the earliest time point examined
was 2.5 minutes. The portion of the curve from 0 to 2.5
minutes was extrapolated. Nonoxidized melanin was not fluo-
rescent (data not shown).
Synthetic melanin as well as isolated bovine melanosomes do
not fluoresce; however, if these melanins are subjected to
harsh oxidative conditions (i.e., light exposure and hydrogen
peroxide), they acquire a fluorescence similar to the autofluo-
rescence of lipofuscin (Fig. 1). The fluorescence of the oxi-
dized melanins are associated with an almost complete degra-
dation of melanin, and it may be considered as a model for the
thetic melanin oxidized after embedding in Spurr’s resin. Isolated
bovine melanosomes (A) and synthetic melanin (B) were embedded in
Spurr’s resin. Semithin sections were prepared, incubated with the
oxidizing medium, and irradiated with visible light for 5 minutes.
Yellow melanin fluorescence (400 nm excitation, 520 nm barrier)
became apparent in the isolated bovine melanosomes (A) and in the
synthetic melanin prepared by oxidation of tyrosine with hydrogen
peroxide (B). Melanin liquefies after oxidation and disperses in the
oxidizing medium, and therefore the micrographs appear out of focus.
Magnification, (A) ?630; (B) ?3000.
Fluorescence of isolated bovine melanosomes and syn-
rescence. Emulsions of the liquid oxidized melanin (synthetic melanin:
A, B; bovine melanosomes: C, D) in the organic oxidizing medium
were examined by phase-contrast (A, C) and fluorescence microscopy
(B, D). The oxidation product of synthetic melanin prepared from
tyrosine (A, B) was a clear hydrophilic liquid containing no apparent
impurities, as demonstrated by phase-contrast microscopy (A). The
yellow circle was due to light reflection of the border between the
organic (organic oxidizing medium, outside) and the hydrophilic
phases (liquid oxidized melanin, inside). Fluorescence microscopy
revealed an intense greenish fluorescence of the oxidation product of
synthetic melanin (B) and of isolated bovine melanosomes (D). Phase-
contrast microscopy detected apparently solid impurities in the liquid
oxidation product of bovine melanosomes (C). These impurities ap-
peared as darker spots in the fluorescence micrograph (D) and were
less fluorescent than the liquid oxidized melanin. Magnification, ?630.
Liquid in vitro oxidized melanin exhibited a greenish fluo-
Absorption spectra of nonoxidized and oxidized synthetic
IOVS, January 2001, Vol. 42, No. 1
Autofluorescence of Oxidized Melanin243
slower and partial physiological oxidation of melanin that oc-
curs during a lifetime. Melanins and RPE cells have been shown
to produce hydrogen peroxide and hydroxyl radicals12,13,14
when exposed to light, suggesting the fluorescence of the RPE
melanin in vivo may be the results of oxidation resulting from
the constant irradiation by light during a lifetime. Melanin also
undergoes spontaneous autoxidation without the addition of
exogenous hydrogen peroxide.15Additionally, some diagnos-
tics and research devices such as fundus photometry involve
artificial light exposure.16Comparison of the fluorescence in-
tensities of oxidized melanin and the autofluorescence of lipo-
fuscin (Fig. 1) justifies the attribution of strong autofluores-
cence of oxidized melanin.
It has been suggested that the fluorescence induced by
oxidative stress in the melanin of human, rat, and bovine RPE
cells and melanocytes in situ and in vitro17is the result of
contaminating proteins and lipids. The results of our studies,
which show that synthetic melanin, which does not contain
contaminating proteins or lipids, becomes fluorescent after
oxidation, suggests that these impurities do not contribute, or
may only partially contribute to the fluorescence observed in
vivo. In the present study, this explanation was ruled out by
comparing the fluorescence properties of natural melanin with
the fluorescence properties of synthetic melanin. The melano-
proteins tended to reduce the photoreactivity of melanin to-
ward oxidizable substrates, such as reduced nicotinamide ad-
enine dinucleotide phosphate (NADPH). When these are
removed or disrupted, the melanosomes are much more pho-
The excitation and emission maxima (approximately 470
and 540, respectively) were the same for synthetic and isolated
bovine melanosomes (Fig. 5). Differences in the spectra related
only to the excitation spectra. Bovine melanosomes showed an
additional excitation plateau at approximately 400 nm that was
missing in the spectrum of synthetic melanin and probably is
evidence of contamination in the phase-contrast micrograph
(Fig. 3C). However, the possibility that the excitation spectrum
of pure natural melanin without contamination differs from the
one of synthetic melanin cannot be excluded.
The apparent color and intensity of the fluorescence in-
duced in melanin by oxidation as observed in histologic sec-
tions (Fig. 1) was very similar to the intensity and color of
lipofuscin (Figs. 1B, 1D). In fact, it is difficult to distinguish
between RPE melanin and RPE lipofuscin in Figure 1D after
oxidation. However, comparing Figure 1C (RPE and choroid at
the beginning of oxidation process) and Figure 1D (same view
of the RPE and choroid after 8 minutes of oxidation) makes the
clear differentiation of the two types of pigment granules
Fluorescence spectroscopy of melanin oxidized in vitro
reveals an emission maximum at 540 nm, and this wavelength
appears as greenish yellow.4Similarly, a greenish fluorescence
was observed by fluorescence microscopy for melanin oxi-
dized in vitro (Fig. 3). However, oxidized isolated bovine mela-
nosomes and synthetic melanin embedded in Spurr’s resin (Fig.
2) exhibited an orange-yellow fluorescence similar to that of
melanin oxidized in situ in histologic sections (Fig. 1).
The difference in the apparent color of in situ oxidized
melanin (Fig. 1) and melanin embedded in Spurr’s resin (Fig. 2;
orange-yellow) and of in vitro oxidized melanin (greenish yel-
low; Figs. 3, 5) may be due to the different physicochemical
environments or the different concentration of the fluoro-
phore.19Gallas and Eisner7observed that melanin samples
irradiated with UV laser light exhibit a blue-green lumines-
cence, whereas when solid samples of similar melanin are
irradiated by the same light, the luminescence appears yellow.
Additionally, Boulton et al.5reported different emission spectra
for melanin samples suspended in saline and melanin prepared
The blue shift observed for liquid samples of melanin ana-
lyzed by spectrofluorometry is also observed for lipofuscin. In
situ lipofuscin analyzed by microfluorometry generally shows
yellowish autofluorescence; however, the fluorescence max-
ima of lipofuscin extracts analyzed by spectrofluorometry are
reported to be in the blue region (400–500 nm).4,20
(Ex470) of in vitro oxidized melanin. Excitation and emission spectra of
both oxidized synthetic melanin (top) and oxidized isolated bovine
melanosomes (bottom) peaked at approximately 470 and 540 nm. The
excitation spectrum of bovine melanosomes shows two additional
minor peaks at 400 and 420 nm. Fluorescence intensity is shown in
Fluorescence excitation (Em540) and emission spectra
of melanin fluorescence after oxidation, a time response study was
performed using hydrogen peroxide–mediated oxidation of synthetic
Time response study. To demonstrate the marked increase
244Kayatz et al.
IOVS, January 2001, Vol. 42, No. 1
In this study the organic hydrophobic oxidizing medium of
Frangioni and Borgioli11was used to oxidize melanin in vitro,
and it dissolved the melanin and formed a yellow, highly
fluorescent hydrophilic liquid. Oxidation of eumelanins by
hydrogen peroxide also results in a yellow aqueous liquid.15
The organic hydrophobic oxidizing medium of Frangioni and
Borgioli11was originally designed to eliminate melanin from
heavily pigmented specimens prepared for histologic examina-
tion. (The term “bleaching” has been used in publications for
at least three different concepts: 1) decreasing fluorescence36;
2) decreasing light absorbance37; and 3) complete elimination
of pigment38. In the present article we avoid the term “bleach-
ing” by actually describing what we mean.)
By phase-contrast microscopy it is possible to observe the
dissolution of melanin in situ exposed to oxidative conditions.
Single melanin granules, for example, in the RPE, become
translucent and finally merge with adjacent melanin drops
(data not shown).
In 1943, Sachs8was the first to report that pigments of
various origin become fluorescent after oxidation with hydro-
gen peroxide. He reported the appearance of a yellow fluores-
cence in human RPE cells after incubation with hydrogen
peroxide. In contrast to our findings, Sachs could not demon-
strate the same effect for choroidal melanin. He termed the
choroidal melanin “real melanin” and RPE melanin “fuscin,”
because “real melanin” did not show any fluorescence, even
after incubation with hydrogen peroxide. Contrary to Sachs,
our data presented in Figure 1 demonstrate the strong autofluo-
rescence of choroidal melanin after incubation with hydrogen
peroxide. Fluorescence of neuromelanin after incubation with
hydrogen peroxide was described by Barden21in 1969, but the
spectroscopic characterization remained controversial.22–24
Katz et al.25reported the appearance of yellow-orange fluores-
cence in melanocytes of the choroid after incubation with
permanganate and speculated that the choroidal fluorescence
was due to melanin.
The fluorescence spectra (Fig. 5) of melanin fluorescence
we observed after exposure to oxidative stress are similar to
the data for the weak autofluorescence of nonoxidized melanin
described by Gallas and Eisner,7which suggests that the weak
autofluorescence of melanin may simply reflect the beginning
of its oxidation. This hypothesis is supported by the observa-
tion of the same authors that fluorescence of melanin is asso-
ciated with structural defects of the melanin polymer. Incuba-
tion of melanin with hydrogen peroxide leads to an oxidative
degradation of the melanin polymer. Thus, detection of mela-
nin autofluorescence in skin,26hair,27and blood28may also be
explained by oxidation of melanin in these tissues. We also
observed photoinduced melanin fluorescence of histologic sec-
tions without the addition of exogenous hydrogen peroxide,
but this was an infrequent observation (data not shown). Sim-
ilar observations have been reported by the groups of Cathy K.
Dorey (personal communication, May 1999) and Jan Borovan-
Our findings have significant implications for a wide range
of investigations dealing with the detection and quantification
of lipofuscin-like autofluorescence in melanin-containing tis-
sue. Because RPE cells, as well as other melanin-containing
cells are frequently exposed to laboratory light and variation in
oxygen concentration in culture, the appearance of fluorescent
material can be explained by the oxidation of melanin due to
nonphysiological conditions during culture.
Another important issue is the potential role of oxidized
reactive melanosomes in producing photograph-oxidative
stress to ocular tissues. Because of its low oxidation potential30
damage to melanosomes would likely lead to melanin oxida-
tion, increase its fluorescent signal (possibly being confounded
with lipofuscin), and increase photographic–oxidative stress
to RPE cells through the photochemical reactions of melanin.31
Boulton et al.5report single excitation and emission peaks
for human fetal melanin granules and the appearance of a
second excitation and emission peak in melanin granules in
adult donor eyes, with a progressive increase of melanin fluo-
rescence intensity with increasing age.32These additional ex-
citation and emission peaks are similar to those of oxidized
melanin we report herein and may be due to the accumulation
of oxidized melanin in the adult eye, which is exposed to a
high intracellular oxygen concentration33and light over a
The melanin fluorescence observed after oxidative stress
and irradiation with light also affects investigations of in vivo
fluorescence of the ocular fundus, as performed by Delori et
al.16The authors report that intrinsic fundus fluorescence
results from at least two fluorophores: a dominant fluorophore
with peak emission at 630 nm and a minor fluorophore with
peak emission at 540 nm when excited at 470 nm. The spectral
properties of this minor fluorophore (emission peak at 540
nm) correspond exactly to the spectral properties of oxidized
melanin (Fig. 5). However, the authors rejected the interpre-
tation that melanin was the minor fluorophore, because of the
assumption that the melanin autofluorescence was too faint to
cause this effect. However, our studies have shown that mel-
anin autofluorescence is not weak but can be very intense.
Melanin autofluorescence in general may simply reflect a
partial oxidation of the polymer. Therefore, the increasing
melanin fluorescence after oxidative stress may be useful for
designing a fluorescence spectroscopic method for determin-
ing or integrating oxidative stress in the ocular fundus, al-
though it may be difficult to distinguish the fluorescence pro-
duced by lipofuscin versus the fluorescence induced by
Finally, the participation of fluorescent oxidized melanin
compounds in the formation of lipofuscin should be consid-
ered. Lipofuscin is likely to be heterogenous, because it is
probably derived from autophagy and the degradation of
POSs.4In fact, 10 different fluorescent fractions have been
identified in lipofuscin extracted from RPE.3Whether any of
these fluorescent fractions are derived from melanin remains to
be investigated. In RPE cells of elderly humans, pigment gran-
ules consist almost completely of melanin cores with lipofus-
cin margins and are called melanolipofuscin granules.34,35The
origin of these granules is still unknown1; however, it is tempt-
ing to speculate that oxidized melanin contributes to their
formation. Additionally, the hypothesis that degradation of
melanin contributes to lipofuscin formation is supported by
the pathologic course of diseases such as age-related macular
degeneration (ARMD) in which there is a correlation between
loss of melanin and the formation of lipofuscin and melanoli-
In conclusion, the results show that melanin fluorescence
increases after oxidation and that the intense fluorescence of
melanin subjected to oxidative stress is a characteristic of the
oxidized melanin and is independent of the fixation protocol
(embedding in Spurr’s resin) or proteinaceous or lipid contam-
ination of the isolated melanin.
The authors thank Andrea Bieker and Hanna Janicki for excellent
1. Feeney–Burns L. The pigments of the retinal pigment epithelium.
Curr Top Eye Res. 1980;2:119–178.
IOVS, January 2001, Vol. 42, No. 1
Autofluorescence of Oxidized Melanin245
2. Kennedy CJ, Rakoczy PE, Constable IJ. Lipofuscin of the retinal
pigment epithelium: a review. Eye. 1995;9:763–771.
3. Eldred GE, Katz M.L. Fluorophores of the human RPE: separation
and spectral characterization. Exp Eye Res. 1988;47:71–81.
4. Yin D. Biochemical basis of lipofuscin, ceroid, and age pigment-
like fluorophores. Free Radic Biol Med. 1996;21:871–888.
5. Boulton M, Docchio F, Dayhaw–Barker P, Ramponi R, Cubeddu R.
Age-related changes in the morphology, absorption and fluores-
cence of melanosomes and lipofuscin granules of the retinal pig-
ment epithelium. Vision Res. 1990;30:1291–1303.
6. Pearse AGE. Histochemistry, Theoretical and Applied. 4th ed.
London: Churchill Livingstone; 1985:874–928.
7. Gallas JM, Eisner M. Fluorescence of melanin-dependence upon
excitation wavelength and concentration. Photochem Photobiol.
8. Sachs HW. U¨ber die autogenen Pigmente, besonders das Lipofus-
cin und seine Abrenzung von Melanin. Beitr Pathol Anat. 1943;
9. Kayatz P, Heimann K, Schraermeyer U. Ultrastructural localization
of light-induced lipid peroxides in the rat retina. Invest Ophthal-
mol Vis Sci. 1999;40:2314–2321.
10. Schraermeyer U, Dohms M. Detection of a fine lamellar gridwork
after degradation of ocular melanin granules by cultured peritoneal
macrophages. Pigment Cell Res. 1996;9:248–254.
11. Frangioni G, Borgioli G. Rapid bleach for melanin. Stain Technol.
12. Korytowski W, Pilas B, Sarna T, Kalyanaraman B. Photoinduced
generation of hydrogen peroxide and hydroxyl radicals in mela-
nins. Photochem Photobiol. 1987;45:185–190.
13. Dorey CK, Khouri GG, Syniuta LA, Curran SA, Weiter JJ. Superox-
ide production by porcine retinal pigment epithelium in vitro.
Invest Ophthalmol Vis Sci. 1989;30:1047–1054.
14. Miceli MV, Liles MR, Newsome DA. Evaluation of oxidative pro-
cesses in human pigment epithelial cells associated with retinal
outer segment phagocytosis. Exp Cell Res. 1994;214:242–249.
15. Prota G. Eumelanins. Melanins and Melanogenesis. London: Aca-
demic Press; 1997:88–118.
16. Delori FC, Dorey CK, Staurenghi G, Arend O, Goger DG, Weiter
JJ. In vivo fluorescence of the ocular fundus exhibits retinal pig-
ment epithelium lipofuscin characteristics. Invest Ophthalmol Vis
17. Kayatz P, Heimann K, Schraermeyer U. Induction of lipofuscin-like
autofluorescence in melanin of RPE cells and in melanocytes after
exposure to oxidative stress [ARVO Abstract]. Invest Ophthalmol
Vis Sci. 1999;40(4):S927. Abstract nr 4891.
18. Gan EV, Haberman HF, Menon IA. Oxidation of NADH by melanin
and melanoproteins. Biochim Biophys Acta. 1974;370:62–69.
19. Yin DZ, Brunk UT. Microfluorometric and fluorometric lipofuscin
spectral discrepancies: a concentration-dependent metachromatic
effect? Mech Ageing Dev. 1991;59:95–109.
20. Hammer C, Braum E. Quantification of age pigments (lipofuscin).
Comp Biochem Physiol B Biochem Mol Biol. 1988;90:7–17.
21. Barden H. The histochemical relationship of neuromelanin and
lipofuscin. J Neuropathol Exp Neurol. 1969;28:419–441.
22. Barden H. Interference filter microfluorometry of neuromelanin
and lipofuscin in human brain. J Neuropathol Exp Neurol. 1980;
23. Dowson JH. A comparison of autofluorescence emission spectra of
bleached neuromelanin in non-diseased substantia nigra with spec-
tra of other intraneuronal pigments in non-diseased and diseased
tissue. Acta Neuropathol (Berl). 1983;61:196–200.
24. Hack MH, Helmy FM. The melanins and lipofuscin. Comp Biochem
Physiol B Biochem Mol Biol. 1983;76:399–407.
25. Katz ML, Stone WL, Dratz EA. Fluorescent pigment accumulation
in retinal pigment epithelium of antioxidant-deficient rats. Invest
Ophthalmol Vis Sci. 1978;17:1049–1058.
26. Fellner MJ. Green autofluorescence in human epidermal cells.
Arch Dermatol. 1976;112:667–670.
27. Fellner MJ, Chen AS, Mont M, McCabe J, Baden M. Patterns and
intensity of autofluorescence and its relation to melanin in human
epidermis and hair. Int J Dermatol. 1979;18:722–730.
28. Hegedus ZL, Kuttab SH, Altschule MD. Studies on rheomelanins,
Part VI: the apparent lipofuscin characteristics of rheomelanins.
Arch Int Physiol Biochim. 1980;88:265–271.
29. Elleder M, Borovansky J. UV induced autofluorescence of melanin
and melanin-like pigments in tissue sections and pigment samples
(abstract). Pigment Cell Res. 1998;11:S234.
30. Jacobson ES, Hong JD. Redox buffering by melanin and Fe(II) in
Cryptococcus neoformans. J Bacteriol. 1997;179:5340–5346.
31. Dontsov AE, Glickman RD, Ostrovsky MA. Retinal pigment epithe-
lium pigment granules stimulate the oxidation of unsaturated fatty
acids. Free Radic Biol Med. 1999;26:1436–1446.
32. Docchio F, Boulton M, Cubeddu R, Ramponi R, Barker PD. Age-
related changes in the fluorescence of melanin and lipofuscin
granules of the retinal pigment epithelium: a time-resolved fluo-
rescence spectroscopy study. Photochem Photobiol. 1991;54:
33. Sarna T. Properties and function of the ocular melanin: a photo-
biophysical view. J Photochem Photobiol B. 1992;12:215–258.
34. Feeney L. Lipofuscin and melanin of human retinal pigment
epithelium: fluorescence, enzyme cytochemical, and ultrastruc-
tural studies. Invest Ophthalmol Vis Sci. 1978;17:583–600.
35. Schraermeyer U, Heimann K. Current understanding on the role of
retinal pigment epithelium and its pigmentation. Pigment Cell Res.
36. Bock G, Hilchenbach M, Schauenstein K, Wick G. Photometric
analysis of antifading reagents for immunofluorescence with laser
and conventional illumination sources. J Histochem Cytochem.
37. Makino CL, Howard LN, Williams TP. Intracellular topography of
rhodopsin bleaching. Science. 1987;238:1716–1717.
38. Frangioni G, Borgioli G. One hour bleach for melanin. Stain
246 Kayatz et al.
IOVS, January 2001, Vol. 42, No. 1